(International Cryogenics Monograph Series) Guglielmo Ventura, Mauro Perfetti (Auth.) - Thermal Properties of Solids at Room and Cryogenic Temperatures-Springer Netherlands (2014)

International Cryogenics Monograph Series
Guglielmo Ventura
Mauro Perfetti
Thermal
Properties of
Solids at Room
and Cryogenic
Temperatures
International Cryogenics Monograph Series
Series editor
Steven W. Van Sciver, Florida State University, Tallahassee, FL, USA
For further volumes:

http://www.springer.com/series/6086
The International Cryogenics Monograph Series was established in the early 1960s
to present an opportunity for active researchers in various areas associated with
cryogenic engineering to cover their area of expertise by thoroughly covering its
past development and its present status. These high level reviews assist young
researchers to initiate research programs of their own in these key areas of
cryogenic engineering without an extensive search of literature.
Guglielmo Ventura Mauro Perfetti
•
Thermal Properties of Solids

at Room and Cryogenic
Temperatures
123
Guglielmo Ventura Mauro Perfetti
INFN Dipartimento di Chimica
Roma Università di Firenze
Italy Sesto Fiorentino
Italy
ISSN 0538-7051 ISSN 2199-3084 (electronic)

ISBN 978-94-017-8968-4 ISBN 978-94-017-8969-1 (eBook)
DOI 10.1007/978-94-017-8969-1
Springer Dordrecht Heidelberg New York London
Library of Congress Control Number: 2014941685
Springer Science+Business Media Dordrecht 2014

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or
information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed. Exempted from this legal reservation are brief
excerpts in connection with reviews or scholarly analysis or material supplied specifically for the
purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the
work. Duplication of this publication or parts thereof is permitted only under the provisions of
the Copyright Law of the Publisher’s location, in its current version, and permission for use must
always be obtained from Springer. Permissions for use may be obtained through RightsLink at the
Copyright Clearance Center. Violations are liable to prosecution under the respective Copyright Law.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
While the advice and information in this book are believed to be true and accurate at the date of
publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for
any errors or omissions that may be made. The publisher makes no warranty, express or implied, with
respect to the material contained herein.
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)

To Eva and other demons
Guglielmo Ventura
If I were a metal, I would definitely say

that I am a heavy lanthanide.
When the temperature is lowered enough
with a powerful refrigerator (Guglielmo),
and a strong magnetic field is applied
(Roberta) my domains (Sergio, Antonella,
Sandro) cooperatively align themselves
to create a strong ferromagnet.
When another magnet (Eva) is sufficiently
near me, we can interact and cooperate
with other magnets (the LaMM staff) to
create an even stronger material
Mauro Perfetti
Preface
This book addresses the needs of researchers in physics and chemistry, project
engineers and students interested in cryogenics and thermal properties of matter.
Using a musical analogy it is piano four hands, not a work for two different
instruments. The book’s three parts, each of which is devoted to a respective
thermal property, are written with the same guiding philosophy: (1) to describe
theories on the propagation of heat in solids in a format that is concise but suf-
ficiently detailed to understand the three thermal phenomena; (2) to review the
main experimental techniques with some examples taken from the literature; and
(3) to present experimental data in the form of tables and graphs.
A rich bibliography is provided at the end of each chapter.
Scientists will be particularly interested in the measurements methods, which
describe some important details in set-ups at cryogenic temperatures. In addition,
data on the thermal properties of several materials at the low (4–300 K) and very
low (\4 K) temperature range are provided at the end of each Part.
For Project Engineers data on the three thermal properties and the integrated
data in the form of tables will offer an essential and time-saving resource.
Students will be provided with the basics for performing measurements at low
temperatures, and with a general, concise guide to the theory involved, focusing on
the most important formulas and concepts necessary for understanding the thermal
properties of solids at low temperatures.
For the sake of conciseness, the words ‘‘materials solid at standard temperature
and pressure (stp)’’ were omitted from the title. Of course any material will
become solid when the temperature is lowered and/or pressure is increased: for
example 4He becomes solid below T & 2 K under a pressure C25 bar.
Only a few materials not solid at stp (e.g. noble gases) are examined, in Part I.
Though data on these materials is often of considerable interest (consider e.g. the
importance of solid nitrogen enthalpy), it would go beyond the scope of this book.
Guglielmo Ventura
Mauro Perfetti
vii
Contents
Part I Heat Capacity
1 Heat Capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Lattice Specific Heat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Electronic Specific Heat . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4 Electronic Specific Heat in Superconducting Materials . . . . . . . 11
1.5 Specific Heat Contributions from Transitions and Defects. . . . . 14
1.6 Magnetic Specific Heat. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.6.1 Magnetic Order and Specific Heat . . . . . . . . . . . . . . . . 17
1.6.2 The Schottky Anomaly. . . . . . . . . . . . . . . . . . . . . . . . 20
1.6.3 Materials Used for Magnetic Refrigeration . . . . . . . . . . 24
1.6.4 Heat Capacity of Regenerators for Cryocoolers . . . . . . . 25
1.7 Specific Heat Due to the Amorphous State . . . . . . . . . . . . . . . 28
1.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2 How to Measure Heat Capacity at Low Temperatures . . . . . . . . . 39

2.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.2 Calorimeters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.3 Heat Pulse Calorimetry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.3.1 Example 1: Heat Pulse Calorimeter
for a Small Sample at Temperatures Below 3 K ...... 44
2.3.2 Example 2: Heat Pulse Calorimetry
for the Measurement of the Specific Heat
of Liquid 4He Near its Superfluid Transition. . . ...... 47
2.4 Relaxation Calorimetry . . . . . . . . . . . . . . . . . . . . . . . ...... 48
2.4.1 Example: Measurement of Specific Heat
of Heavily Doped (NTD) Ge. . . . . . . . . . . . . . . . . . . . 50
2.5 Dual Slope Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.6 AC Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.7 Differential Scanning Calorimetry . . . . . . . . . . . . . . . . . . . . . 57
2.8 Other Methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
2.9 Industrial Calorimeters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
ix
x Contents
2.10 Small Sample Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . 61

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3 Data of Specific Heat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

3.1 Presentation of DATA of Specific Heat . . . . . . . . . . . . . . . . . 69
3.2 Very-Low Temperature DATA (Below About 4 K) . . . . . . . . . 69
3.2.1 Metals and Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.2.2 Dielectrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.3 Low-Temperature Specific Heat DATA
(Approximately 4–300 K) . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.3.1 Metals and Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.3.2 Dielectrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Part II Thermal Expansion
4 Thermal Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.2 Thermal Expansion Theory . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.3 Negative Thermal Expansion. . . . . . . . . . . . . . . . . . . . . . . . . 86
4.3.1 Application of NTE . . . . . . . . . . . . . . . . . . . . . . . . . . 86
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
5 How to Measure the Thermal Expansion Coefficient

at Low Temperatures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.1 Capacitive Dilatometers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.1.1 Principles of Capacitive Techniques. . . . . . . . . . . . . . . 94
5.1.2 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
5.2 Interferometric Dilatometers . . . . . . . . . . . . . . . . . . . . . . . . . 102
5.2.1 Principles of Interferometric Dilatometry . . . . . . . . . . . 103
5.2.2 Homodyne Dilatometer: Example . . . . . . . . . . . . . . . . 106
5.2.3 Heterodyne Dilatometer with Cryogenic Liquids:
Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 107
5.2.4 Heterodyne Interferometric Dilatometer: Example . .... 111
5.2.5 Heterodyne Dilatometer with Mechanical
Coolers: Examples . . . . . . . . . . . . . . . . . . . . . . . .... 116
5.3 Very Low Temperature Thermal Expansion . . . . . . . . . . .... 117
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 118
6 Data of Thermal Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Contents xi
Part III Thermal Conductivity
7 Electrical and Thermal Conductivity . . . . . . . . . . . . .......... 131

7.1 Electrical Conductivity . . . . . . . . . . . . . . . . . . . .......... 131
7.1.1 Relation Between Thermal and Electrical
Conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
7.1.2 Electrical Resistivity of Metals . . . . . . . . . . . . . . . . . . 134
7.1.3 Electrical Conductivity of Semiconductors . . . . . . . . . . 137
7.2 Magnetic and Dielectric Losses . . . . . . . . . . . . . . . . . . . . . . . 139
7.2.1 Losses in Dielectric Materials . . . . . . . . . . . . . . . . . . . 139
7.3 Thermal Conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
7.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
7.3.2 Lattice Thermal Conductivity . . . . . . . . . . . . . . . . . . . 145
7.3.3 Thermal Conductivity of Dielectrics . . . . . . . . . . . . . . 146
7.3.4 Thermal Conductivity of Nanocomposites . . . . . . . . . . 152
7.3.5 Composite Materials . . . . . . . . . . . . . . . . . . . . . . . . . 155
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
8 How to Measure Thermal Conductivity . . . . . . . . . . . . . . . . . . . . 169

8.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
8.2 Steady State Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
8.2.1 Longitudinal Flux Method . . . . . . . . . . . . . . . . . . . . . 171
8.2.2 Radial Flux Method or Cylinder Method . . . . . . . . . . . 173
8.3 Transient Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
8.3.1 The 3x Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
8.3.2 Pulse Power Method . . . . . . . . . . . . . . . . . . . . . . . . . 177
8.4 Thermal Diffusivity Measurements. . . . . . . . . . . . . . . . . . . . . 179
8.4.1 Laser Flash Method . . . . . . . . . . . . . . . . . . . . . . . . . . 179
8.4.2 Temperature Wave Method. . . . . . . . . . . . . . . . . . . . . 180
8.5 Examples of Measurements of Electrical
and Thermal Conductivity. . . . . . . . . . . . . . . . . . . . . . . .... 181
8.5.1 Measurement of Electrical Resistivity of Heavily
Doped NTD 31 Germanium at Very Low
Temperatures, and Calculation of Electron-phonon
Decoupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 181
8.5.2 Measurement of the Thermal Conductivity
of Torlon in the 0.08–300 K Temperature Range . .... 185
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 190
9 Data of Thermal Conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

9.1 Very Low Temperature Data . . . . . . . . . . . . . . . . . . . . . . . . . 195
9.2 Low Temperature Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
9.3 Crystalline Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Part I
Heat Capacity
Main Symbols
Q Heat flow
C Heat capacity
c Specific heat
T Temperature
E Internal energy
V Volume/molar volume
S Entropy/Total spin momentum
p Pressure
cV Constant volume specific heat
cp Constant pressure specific heat
b Coefficient of volume thermal expansion
cG Grüneisen parameter
fT Isothermal compressibility coefficient
fS Adiabatic compressibility coefficient
hD Debye temperature
NA Avogadro’s constant
kB Boltzmann’s constant
h Plank’s constant
R Ideal gas constant/resistance
q Density
Mm Molar mass
v Velocity
g Volumetric density of states
EF Fermi energy
c Sommerfeld’s constant
r Number of atoms per molecule
L Latent heat/orbital angular momentum
H Magnetic field/enthalpy/heater
G Gibbs free energy
v Magnetic susceptibility
2 Part I Heat Capacity
P Power
t Time
j Thermal conductivity
R Thermal resistance
F de Gennes factor
J Total angular momentum
a Radial distance from a given ion
Chapter 1
Heat Capacity
Abstract Specific heat provides a link among the many solid state theories; vice
versa these theories can also be used to estimate the specific heat of materials.
From a practical point of view, the knowledge of the specific heat of technically
important materials is often fundamental in the design of instruments and systems
which have to work in the low temperature regime. Since cryogenics is presently
used in research, aerospace, industry and energy production and storage, specific
heat data for commonly used materials are mandatory. In this chapter theories
about contributions to specific heat are reported: lattice specific heat (Sect. 1.2),
electronic specific heat in normal (Sect. 1.3) and superconducting (Sect. 1.4)
materials, contributions from transitions and defects (Sect. 1.5), magnetic specific
heat (Sect. 1.6), contributions present in amorphous materials (Sect. 1.7).
1.1 Introduction
Specific heat, c, is an intensive variable that relates the heat per unit mass supplied
to a system to the resultant absolute temperature change of the system itself.
Several needs have stimulated the study of this thermal property of matter. In fact,
specific heat experiments verified Nernst’s statement of the third law of thermo-
dynamics, allowed the calorimetric determination of the energy states of sub-
stances and were the basis for Greywall’s study about 3He melting pressure
thermometry.
Specific heat provides a link among the many solid state theories; conversely,
these theories can be used to estimate the specific heat of materials. From a practical
point of view, the knowledge of the specific heat of technically important materials
is often fundamental in the design of instruments and systems which have to work
in the low temperature regime. Since cryogenics is presently used in research,
aerospace, industry and energy production and storage, specific heat data for
commonly used materials are mandatory. Most low temperature specific heat data
refer to very pure samples (difficult to obtain) or exotic compounds; hence, needed
G. Ventura and M. Perfetti, Thermal Properties of Solids at Room 3

and Cryogenic Temperatures, International Cryogenics Monograph Series,
DOI: 10.1007/978-94-017-8969-1_1, Springer Science+Business Media Dordrecht 2014
4 1 Heat Capacity
values must be computed from very limited data. Calculation and estimation of
specific heats are sometimes possible by using accepted models [1]. Predictions on
the specific heat of particular materials, e.g., have been reported [2–4].
Luckily, when the temperature is reduced, the thermal properties of materials
often become easier to describe (solid 3He is one of the few exceptions). In
particular, the various contributions to the specific heat of a material can be studied
independently: for example, the contribution which originates from the nuclear
spin can be considered independent from the electron and phonon systems.
Moreover, the electron and phonon contribution of a metal can be simply added to
obtain the total specific heat. However, this is not always true; in fact it is usually
impossible to calculate the specific heat of superconducting alloys from the
knowledge of the various components’ specific heat. Rigorous developments of
theories about the various contributions to the specific heat can be found in many
books on solid-state physics [5–13]. Here, we present the principal models which
explain the underlying physics of each contribution in a concise form.
Heat capacity, C, is defined as the amount of heat, dQ, required to raise the
temperature of a material by a unit of temperature:

dQ
Cx ¼ ð1:1Þ
dT x
where x denotes any of the several constraints which can be imposed. The most
common constraints are pressure (p) and volume (V). Contrarily to specific heat,
heat capacity is an extensive quantity, which means that it depends on the size of
the system.
As many other physical properties, the heat capacity can be defined in terms of
other thermodynamic state variables, in particular, it can be written as the deriv-
ative of either the entropy, S, or internal energy, E, as

oE oS
CV ¼ ¼T : ð1:2Þ
oT V oT V
Equation (1.2) indicates that when the temperature of a system is increased but
the system is forced to maintain a constant volume, all the heat is stored as internal
energy. Likewise, constant pressure heat capacity may be written as

oH oE oV
CP ¼ ¼ þp ð1:3Þ
oT p oT p oT p
where H is the enthalpy. It is also possible to define C with other external variables
held constant. For example, in systems with relevant magnetic properties, CH or
CM may be used to designate the heat capacity at a constant applied magnetic field
or magnetization, respectively. This topic is of particular interest in magnetic
cooling systems (see Sect. 1.6) [14]. Among the several systems of units often
used to express specific heat, we shall adopt J mol-1 K-1, except in the case of
1.1 Introduction 5
alloys or nonstoichiometric compounds, where mass is more appropriate than

molar basis. Hence,

1 dQ
cx ¼ ð1:4Þ
n dT x
where n is the number of moles. Let us note that in some applications, it is useful
to define a heat capacity per unit volume. In order to avoid misunderstanding, we
will remark ‘‘per unit volume’’ every time that this quantity is present in equations.
To evaluate the specific heat, the various excitations that take place in the
examined material are to be considered. To do that, an accurate chemical analysis
of the studied sample itself is mandatory because impurities and defects can play a
crucial role at very low temperatures. The chemical structure of the materials is
strictly connected to the ways in which the heat is propagated through the mate-
rials. Thus, as mentioned, the specific heat gives plenty of information about the
specimen.
Note that experimental data usually give the specific heat at constant pressure cp
because of the practical difficulty of keeping the volume constant. Instead, theories
usually refer to the specific heat at the constant volume cV parameter which does
not take into account the thermal expansion (see Part II). This is why the specific
heat cp is always greater than cV, by a factor (1 + bcGT), where b is the volumetric
coefficient of thermal expansion and cG is the so-called Grüeneisen parameter
defined as
V V
cG ¼ b ¼b ð1:5Þ
fT c V fS c P
where V is the molar volume and fT and fS are the isothermal and adiabatic
compressibility coefficients (see Sect. 1.5 and Part II, Sect. 4.2) [5–13].
For most solids below room temperature, cp exceeds cV by less than 2 % (about
1 % at T = hD/2, with hD the Debye temperature of the material). This little dif-
ference is due to the fact that b is extremely small below hD (see Part II, Sect. 4.2
for further details). The latter consideration allows one to discuss how cV, studied in
theoretical models, [5–13] depends on temperature, and to compare it with
experimental data of cp. In fact, a fairly extensive amount of experimental data of cp
exists for solids at room and low-temperatures. For simple solids such as metals and
crystalline insulators, there is a good match between experimental and theoretical
data near room temperature. The classical model of harmonic oscillators developed
by Dulong and Petit [15] works well around room temperature. Adopting this
model, the heat capacity (independently from temperature and type of material) is
25 J/mol K for a monoatomic solid. At low temperatures, cp is temperature
dependent and there is a markedly different behavior according to the type of solid,
as we shall see in the following sections, where the phenomena which contribute to
the specific heat are separately considered. Note that to get the specific heat of a
material, it is not always necessary to take all of these contributions into account
6 1 Heat Capacity
Fig. 1.1 Specific heat as a

function of temperature of
silicon (Si), [16] copper (Cu)
and polyvinyl chloride (PVC)
[17]
because, in many cases (depending on the chemical nature of the material and on
temperature range), some of them can be neglected.
Specific heats of materials in the low temperature range are strong functions of
temperature, as shown in Fig. 1.1; note, for example, that silicon specific heat
changes by more than six orders of magnitude in the 1–200 K range.
1.2 Lattice Specific Heat
In this section, we will describe the lattice contribution to specific heat based on
Debye’s model because in many cases, it gives a reasonable agreement with
experimental results. However, other models of lattice vibration of a solid have
also been proposed [7, 9, 10, 12].
At high temperatures (T [ 100 K), cV is essentially due to the phonon contri-
bution cph approaching the classical Dulong and Petit value, namely,

J
cV ¼ 3rNA kB ¼ 3rR ffi 25r ; ð1:6Þ
mol K
where r is the number of atoms per molecule, NA is Avogadro’s number

(6.022 9 1023 mol-1), kB is Boltzmann’s constant (1.38 9 10-23 J K-1) and R is
the constant of gases (8.31 J/mol K) [18].
For cph, Debye’s elastic continuum model for solids [5–13] gives
3 hZD =T
T x4 ex
cph ðTÞ ¼ 9rNA kB dx ð1:7Þ
hD ðex 1Þ2
0
1.2 Lattice Specific Heat 7
Fig. 1.2 Debye temperature

versus temperature for some
materials. The values of
Debye temperatures are
obtained from heat capacity
measurements [5–13]
where x : hx/(2pkBT) (x is the frequency, h is the Planck’s constant). The

integral can be solved by noting that for T \ hD, the upper limit of (1.7) can be
extended to infinity with a small error [5, 8].
3 3
12 4 T T
cph ðTÞ ¼ p rNA kB ¼ 1944r ½J=mol K ð1:8Þ
5 hD hD
for temperatures T \ hD/10. The parameter hD can be identified as the temperature

above which all modes of the atoms are excited, and thus is a function of the
temperature (see Fig. 1.2). In Table 1.1, we report the low temperature limit
values of hD for some elements.
Deviations from (1.8) depend on the chemical nature of the material as shown
in Fig. 1.2. Note that for applications in thermometry, one of the requirements is
that the hD of the element used as a temperature sensor must remain constant in the
temperature range of interest. Therefore, e.g., Platinum-based thermometers can be
used in the 30–80 K range.
The cubic dependence of cph on temperature accounts for the small specific heat
at a low temperature of crystalline insulators. The specific heat of noble gas
crystals is shown in Fig. 1.3a (the dashed line represents the Dulong and Petit limit
value), while Fig. 1.3b displays the T3 dependence of the specific heat of Ar below
2 K. Fitting the experimental data, it is possible to obtain hD using (1.8). For
example, the fit performed in Fig. 1.3b demonstrates that the measured specific
heat of Ar is in good agreement with the Debye law, with a low temperature limit
value of hD equal to 92 K.
We wish to note that Debye theory is only applicable at low temperatures.
When using (1.6) at temperature T [ hD, it must be recognized that the results are
rough approximations.
hD can also be calculated by other methods [22, 23] (e.g., from the measured
ultrasonic velocities) and data may be compared to those obtained from specific
8 1 Heat Capacity
Table 1.1 Low temperature limit of Debye temperatures for some elements [10, 19]
Fig. 1.3 a Specific heat of Ar, Kr and Xe. The horizontal dashed line is the classical Dulong–
Petit value [20]. b Specific heat of Ar as a function of T3 [10, 21]
heat measurements. The Debye’s temperature evaluated from data of ultrasonic

velocity can be achieved by starting from the definition of hD as (see, e.g., [22, 24])
1 1
h 3N 3 h 3NA q 3
hD ¼ mm ¼ mm ð1:9Þ
kB 4pV kB 4pMm
where N is the number of atoms for the cell, Mm is the molar mass, q is the density
and mm is the mean velocity given by
1.2 Lattice Specific Heat 9
0 113
3
mm ¼ @ A ð1:10Þ
1= 3 þ 2= 3
mL mT
where mL and mT are the velocities of the longitudinal and transverse waves,
respectively [5]. Keeping in mind that the product kBhD indicates the minimum
energy required to excite all the modes, [5] it is easy to understand that Debye’s
temperature is generally high for lattices made by strongly bound light atoms, as
diamond (hD = 2230 K), and low for lattices made of weakly bound heavy atoms,
as lead (hD = 105 K). Table 1.1 reports a lot of data referred to hD of various
elements; however, when two or more elements are combined to form a com-
pound, other effects may become relevant (intermolecular weak forces, magnetic
coupling etc.). Therefore, the resultant hD can be difficult to predict.
1.3 Electronic Specific Heat
When Sommerfeld applied Fermi–Dirac statistics to free electrons in metals, it

became clear why the electronic contribution to the specific heat was much smaller
than the classically predicted value (1.6); in fact, by writing the internal energy in
terms of density of states, we obtain
oE p2
ce ðTÞ ¼ ¼ gðEF ÞkB2 T ð1:11Þ
oT 3
where g(EF) = (3/2)(NA/EF) is the volumetric density of states calculated at the

Fermi energy (EF). Substituting in (1.11) the definition of g(EF), [5] a linear
dependence on temperature for the electronic specific heat is obtained, that is,
p2 NA 2
ce ðTÞ ¼ k T ¼cT ð1:12Þ
2 EF B
where c is called the Sommerfeld constant. Knowing that EF is of the order of

some eV (1 eV = 1.6 9 10-19 J) for almost all metals, it is clear that c will be of
the order of 1 mJ/mol K2. The experimental values of c for some metals and alloys
are reported in Table 1.2. It is worth noting that the observed values can differ
from the calculated ones [10].
Thus, the total specific heat of a metal is a sum of two terms: a cubic (phononic)
contribution and a linear (electronic) contribution, becoming relevant only for
T B 10 K for most materials:
10 1 Heat Capacity
Table 1.2 Observed Sommerfeld constant of some elements in mJ/(mol K2). All data are taken
from [10] except for the rare earths metals collected from [25–34]
Fig. 1.4 Specific heat c of

copper divided by the
temperature T fitted as a
function of T2 [11]
1944r 3
cðTÞ ¼ c T þ T : ð1:13Þ
h3D
Dividing c by T, it is then possible to obtain, by a simple fit, the value of c. The

fit reported in Fig. 1.4 for c of copper gives hD = 339 K and c = 0.684 mJ/mol K,
in good agreement with the values reported in Tables 1.1 and 1.2, respectively.
1.4 Electronic Specific Heat in Superconducting Materials 11
Table 1.3 Approximate transition temperature of some elements in zero magnetic field and
critical fields at T = 0 K [10, 45–58]
*Superconducting only under pressure. For some elements which can assume different crystallo-
graphic structures (e.g., La, Sn), we chose one of them and invoke the appropriate references for
further details. For Pt and S, the sample was a compacted powder, Pd was measured under
irradiation with He+ ions and Cr was measured as a thin film
1.4 Electronic Specific Heat in Superconducting Materials
Materials which exhibit the phenomenon of superconductivity enter into a new

state below a critical temperature, Tc (see Table 1.3). In this ‘‘superconducting
state,’’ the electrical resistivity of the material becomes zero and its thermal
properties also change. Tc is strongly affected by the applied magnetic field as well
as by the purity of the material [35]. By applying a magnetic field stronger than the
critical field (Hc), the material does not enter the superconducting state as shown in
Fig. 1.5, where the specific heat of Hg was measured without applied field and in a
field of 1 kG [36]. We notice that the lattice specific heat cph is not modified by the
superconducting transition, while the electronic contribution vanishes as temper-
ature decreases. Using (1.8), it is possible to extract hD = 69 K, a value close to
the one reported in Table 1.1. Calculations carried out for Hg can only be done
with low hD materials which present a strong phonon contribution even at low
temperatures. For other materials, the behavior may be quite different if the
dominant contribution to the specific heat is electronic at temperatures close to Tc.
In Fig. 1.6a, we report cs and cn for Al around Tc where we observe a jump, Dce,
in the electronic specific heat due to the superconducting transition. For ‘‘simple’’
superconductors, such as Al and Sn, the Bardeen–Cooper–Schrieffer theory (BCS)
[37–41] gives
12 1 Heat Capacity
Fig. 1.5 Specific heat of Hg

in the superconducting state
(cs) and in the normal state
(cn) [36]

cs cn
Dce ¼ c TC ¼ 1:43 c TC : ð1:14Þ
c n Tc
In Fig. 1.6a, Dce at Tc is about 2.12 mJ/mol K, giving c = 1.30 mJ/mol K, in

good agreement with the value reported in Table 1.2. Generally, the predictions of
BCS theory are in good agreement with the experimental values, except for strong-
coupling superconductors as Pb or Hg [5].
Below Tc, the electronic specific heat of a superconductor, ces, decreases with
the temperature as (see Fig. 1.6b)
TC
ces ¼ a eb T ð1:15Þ
where a and b are constants. The quantities a, TC, c, and b are essentially related to
the zero-temperature energy gap [5]. In particular, the value of b is similar for
many metals and is about 1.34 [43]. Thermodynamic arguments [44] indicate that
the transition from the normal to the superconducting state at zero magnetic field
does not involve a latent heat and therefore must be a higher order transition.
Experimental evidence indicates that it is a second order transition.
For example, the observed specific heat of single-crystal vanadium, reported in
Fig. 1.7 at H = 0, shows the form of the theoretically predicted curve. The latent
heat is

VHc T dHc
L ¼ TðSn Ss Þ ¼ ð1:16Þ
4p dT
1.4 Electronic Specific Heat in Superconducting Materials 13
Fig. 1.6 a Specific heat of Al in the superconducting state (cs) and in the normal state (cn) [42].
b Data of specific heat of (a) divided by T. Solid lines represent the fits
Fig. 1.7 Specific heat of

metal vanadium [59]. The
solid line represents data
taken at H = 0 and illustrates
the second order transition.
The dashed curve for normal
vanadium was obtained by
applying a magnetic field
H [ Hc
which gives

VHc dHc
ðSn Ss Þ ¼ ð1:17Þ
4p dT
where Hc is the temperature-dependent critical field, V is the volume, and Sn and Ss

are the entropies in the normal and superconducting states, respectively. The
temperature dependence of Hc is not easily predictable, being strongly dependent
on the characteristics of the superconductor; it is often expressed in terms of the
deviation from the empirical law [5]:
2
Hc ðTÞ T
1 : ð1:18Þ
Hc ð0Þ Tc
14 1 Heat Capacity
From
dS
C¼T ð1:19Þ
dT
we get

VT d dHc VHc T d 2 Hc VT dHc 2
ðCs Cn Þ ¼ Hc ¼ þ : ð1:20Þ
4p dT dT 4p dT 2 p dT
At T = Tc and Hc = 0, the first term in (1.20) vanishes, giving

VTC dHc 2
ðCs Cn Þ ¼ : ð1:21Þ
4p dT
At very low temperatures, the second derivative term in (1.20) dominates

(d2HC/dT2 \ 0), so Cs - Cn \ 0. The difference (Cs - Cn) at TC has been care-
fully determined: it is 9.75, 10.6, and 41.5 mJ/mol K for In, Sn, and Ta, respec-
tively. Corresponding values computed from (1.20) using values of dHc/dT from
[60] are 9.52, 10.56, and 41.6 mJ/mol K, respectively. This gap between Cs and Cn
can be significant (e.g., 44 % for Sn) and has to be considered in any experimental
setup containing superconductive elements.
Finally, note that for some elements (e.g., Y), the superconductivity can be
observed either with or without applied pressure; the critical temperature is gen-
erally strongly dependent on the applied pressure. Therefore, data in Table 1.3
refer to a particular value of external pressure which can be found in the references
listed in the caption. In addition, Tc and Hc can also significantly vary as a function
of the composition of alloys; e.g., we can cite that Sn–Pb (50–50 wt%) alloy has a
Tc = 7.75 K and Hc = 2000 G while Sn–Pb (60–40 wt%) has a Tc = 7.05 K and
Hc = 800 G [61].
1.5 Specific Heat Contributions from Transitions

and Defects
In the case of electrical conductors, the main specific heat contributions come from
phonons and free electrons, as described by (1.13). The specific heat for these
materials has characteristic temperature dependence, i.e., it always decreases with
decreasing temperature, as shown in Fig. 1.1. In some materials, however, exper-
imental data show irregularities resulting from various transitions in the material.
Only first-order, second-order, and k-type transitions are experimentally
observed, although theories allow for the existence of higher order transitions.
Each of these transitions produces a characteristic effect on the specific heat.
1.5 Specific Heat Contributions from Transitions and Defects 15
Fig. 1.8 Examples of: a first-order phase-transition in SbSI from Ref. [63], b second-order
phase-transition in Nb from Ref. [64], c k-type phase-transition in NiCl2 6H2O from Ref. [65]
Before discussing these effects, the nature of the transitions will be briefly
reviewed. More detailed discussions of the thermodynamic basis of these transi-
tions can be found in most texts on thermodynamics, e.g., Ref. [44].
As mentioned in Sect. 1.1, the most common thermodynamic variables that can
be maintained constant are T and p. The Gibbs free energy, G = H - TS, must not
change when temperature and pressure are kept constant. However, according to
Ehrenfest [62], an i-order transition is characterized by a discontinuity of the ith
derivative of G, while all the i-1 derivatives are continuous.
Starting from the Ehrenfest classification, a first-order transition is character-
ized by a discontinuous change of the first derivatives of the Gibbs free energy, the
entropy, S = -(qG/qT)p, and the volume, V = (qG/qP)T, while cp = T(qS/qT)p
goes to infinity. This condition is shown schematically in Fig. 1.8a. The specific
heat is finite in both phases until the transition is reached, that is, there is no
anticipation of the transition. Common transitions which fall into this category are
melting, vaporization, and sublimation. Some transitions exist in which S, V, T, p,
and G are unchanged while specific heat, coefficient of volume expansion, b, and
isothermal compressibility, fT, show finite changes. This behavior requires that the
first derivative of G has to be continuous, whereas the second derivative must be
16 1 Heat Capacity
discontinuous: transitions of this type are known as second-order transitions. b and

fT are defined as

1 oV
b¼ ð1:22Þ
V oT p

1 oV
fT ¼ : ð1:23Þ
V op T
The effect of a second-order transition on specific heat is shown in Fig. 1.8b.

Like in the case of first-order transitions, there is no anticipation of the transition.
Superconducting transition in zero magnetic field is an example of a second-order
transition (see Sect. 1.5).
Another phase transition frequently observed in specific heat data is the so-
called k transition. It is similar to second-order transitions in that S, V, T, p, and
G remain constant; however, k transitions produce infinite changes in Cp, b, and
fT. There is, in this case, a transition anticipation which leads to the k shape shown
in Fig. 1.8c.
This k-shaped behavior in solids is observed for order-disorder transformations
in alloys, magnetic and electric transformations, the onset of molecular rotations,
and ordering of axis orientation of molecular rotations. Contributions to specific
heat at or near these transitions may be attributed to either noncooperative (par-
ticles change energy states independently) or cooperative interactions (with strong
interactions among particles). The phenomena mentioned above in connection
with second-order and k transitions are examples of cooperative behavior. Two
cooperative transitions which are particularly relevant to the specific heat of solids
are magnetic spin alignment and normal-to-superconducting transition; noncoop-
erative components of specific heat include the normal lattice and electronic terms,
and the Schottky effect (see Sect. 1.6).
Since specific heat depends strongly on the lattice configuration, imperfections
such as dislocations and strain should play a role in the determination of specific
heat. Theoretical analysis of pinned dislocations and small-angle boundaries
indicates that the temperature dependence is proportional to T and T 2, respectively
[66]. Reasonable values of dislocation densities (*1010 cm-2) and boundary
surface areas (*106 cm-1) result in extremely small values of specific heat
contributions. Contributions from dislocations may be detected experimentally
only in insulators and superconducting materials at T \ 0.1 K, where the elec-
tronic component is very small. The small-angle boundary effect is generally too
small to be experimentally observed.
Martin [67] measured the specific heat of annealed and cold-worked 99.999 %
pure copper and found that it was about 0.15 % higher than the one of the annealed
specimen (20 K \ T \ 300 K). This small difference is in the direction expected
when the effect of strain on the lattice constants is considered: in fact strains lower
the lattice constants. Note that Collings and co-workers [68] determined the specific
1.5 Specific Heat Contributions from Transitions and Defects 17
heat of several structural alloys with different heat treatments for T \ 20 K,

T = 80 K, and T = 300 K. The specific heat differences for similar specimens with
different thermal histories were generally within the experimental uncertainty [68].
1.6 Magnetic Specific Heat
1.6.1 Magnetic Order and Specific Heat
If a material contains atoms which can be permanently or temporary magnetized

under an external magnetic field, an additional term MdH has to be added to the
thermodynamic expression for the change in total energy of the system:
dQ ¼ TdS ¼ dE þ pdV þ MdH: ð1:24Þ
In this formula, M is the modulus of magnetization, defined as the vector sum of

all the magnetic dipole moments, and H is the magnetic field strength. The term
MdH is conceived as an external energy that comes from magnetic field [8].
For the correct understanding of the equations, another formalism, which
0
includes the magnetic energy E = E + MH into the internal energy of the
material, is preferred. Thus, (1.24) becomes
dQ ¼ TdS ¼ dE0 þ pdV HdM ð1:25Þ
where pdV can usually be neglected when compared to -HdM, giving the relation
dQ ¼ TdS ¼ dE0 HdM: ð1:26Þ
Fundamental definitions for the two main magnetic heat capacities are [8]

dQ dS
CH ¼ ¼T ð1:27Þ
dT H dT H

dQ dS
CM ¼ ¼T : ð1:28Þ
dT M dT M
The heat capacity at constant magnetization is independent on an external

magnetic field and, as was the case for CV, appears in theories. Heat capacity at
constant field strength, CH, is the quantity determined experimentally (corresponds
to Cp in the nonmagnetic case). Comparing (1.27) and (1.28) to (1.2) and (1.3), we
clearly note that the correspondence is magnetic M ? -V and H ? p. As in the
18 1 Heat Capacity
case of Cp and CV, a useful estimation of the difference CH - CM has to be

calculated in order to know how strong the deviation of experimental data from the
computed data can be. Reference [69] gives
2 1
oM oM
CH CM ¼ T : ð1:29Þ
oT H oH T
Equation (1.29) can be demonstrated by analogy with

op oV
Cp CV ¼ T : ð1:30Þ
oT V oT p
Substituting M = -V, H = p,
2
oM oH oM oH oT
CH CM ¼ T ¼ T : ð1:31Þ
oT H oT M oT H oT M oM H
Since, in this case,

oH oT oH
¼ : ð1:32Þ
oT M oM H oM T
By substituting (1.32) into (1.31), one obtains (1.29).

All the magnetic behaviors, except diamagnetism and Pauli paramagnetism,
[70] which are independent of temperature, influence the specific heat. Before
discussing the changes in the specific heat connected to magnetic molecules or
atoms in detail, we briefly recall the different magnetic behaviors of the materials.
When a material composed by atoms with unpaired electrons, which have net
electronic magnetic moments, is subjected to an external magnetic field, all the
spins tend to align in parallel to the field direction. Since the energy provided by
the field is in contrast to the thermal energy, which tends to randomly orient the
moments, the total magnetization is proportional to the field strength through the
magnetic susceptibility tensor, defined as vij ¼ oMi =oHj .
If a material is paramagnetic and an external field is applied, the susceptibility
is positive (this means that the spins tend to align in parallel to the magnetic field)
and, when the field is switched off, the thermal energy is higher compared to the
magnetic coupling energy, and thus the system returns into a state with no net
magnetization (see Fig. 1.9a). The relaxation time of paramagnets is generally
short (\10-9 s), but can be observed with particular techniques like pulsed-EPR
[71].
The magnetic moments of ferromagnetic materials spontaneously and cooper-
atively assume a local regular order (magnetic domain) when the temperature is
lower than a characteristic temperature called Curie temperature (TC). If a strong
enough external field is applied and then removed, this class of materials remains
1.6 Magnetic Specific Heat 19
Fig. 1.9 Alignment of spins

without applied field in
paramagnets (a),
ferromagnets (b),
antiferromagnets (c),
ferrimagnets (d)
permanently magnetized due to the strong coupling of magnetic moments

(Fig. 1.9b); this is possible because there is a singularity in the susceptibility at TC
[8]. Above TC, the spontaneous parallel alignment disappears, leaving the disor-
dered paramagnetic state. The magnetic susceptibility of these materials is extre-
mely high compared to the paramagnets and increases rapidly with increasing H.
Some materials minimize their energy if the alignment of magnetic moments is
antiparallel. These substances, formed by atoms with magnetic moments of the
same intensity, are called antiferromagnets and their spin structure is often
decomposed into two spin lattices, each one composed by spins aligned in a
ferromagnetic way. In analogy to ferromagnets, antiferromagnets can exhibit
magnetic order only under a characteristic temperature, called Néel temperature
(TN). Since the two spin lattices have the same intensity but opposite orientation,
the resultant magnetization of an antiferromagnet is zero (Fig. 1.9c).
If the material tends to order in an antiferromagnetic way but it is formed by
magnetic moments of different strength, the spin lattices are not compensated and
a net magnetization is observed (Fig. 1.9d). Also, this phenomenon, known as
ferrimagnetism, can be observed under a critical temperature that has been labeled
either TN or TC. Both terms have some logical merit: TN reminds the antiparallel
alignment of spins, while TC recalls the spontaneous magnetization observed.
It can be shown [73, 74] that for ferromagnetic and ferrimagnetic materials
3
CM / T 2 ð1:33Þ
20 1 Heat Capacity

NpCo2 illustrates the general
effect of a magnetic
transition. The Néel
temperature of this compound
is 12.5 K [72]
and for antiferromagnetic compounds
CM / T 3 : ð1:34Þ
An example of the behavior of the specific heat near a magnetic transition is

shown in Fig. 1.10 for NpCo2. This alloy undergoes a para-to-antiferromagnetic
transition at 12.5 K, while at low temperatures, the persistence of spin fluctuations
in the antiferromagnetic state is the reason for the increase of specific heat at
T \ 6 K [72]. These phenomena are theoretically described by several models
which differ for both the type of interaction and for the ‘‘geometry’’ and dimen-
sionality of the magnetic lattice (1, 2 or 3 dimensions, see [73] ). Changes in the
dimensionality of the magnetic lattice have a dramatic effect upon the thermo-
dynamic properties. It is interesting to notice that the peak shown in Fig. 1.10 can
be suppressed by applying a magnetic field stronger than 4.3 T because the
compound undergoes a metamagnetic transition that can be observed when the
external field is stronger than the antiferro- or ferromagnetic interactions between
the atoms [72].
1.6.2 The Schottky Anomaly
In some systems, one more magnetic effect, which is due to noncooperative

changes in order, may occur. This phenomenon is known as the anomalous
Schottky effect and is observed in paramagnetic salts, in some ferromagnetic
metals and also in native minerals [75]. The Schottky anomaly is a result of
electronic energy degeneracies in the paramagnetic materials and nuclear energy
degeneracies in the ferromagnetic materials. The observed Schottky specific heat
contribution is due to changes in internal energy which occur when nearly adjacent
Fig. 1.11 Specific heat c of a

two-level system with energy
separation DE
energy levels are occupied. The degenerate energy levels may be caused by
external or internal magnetic fields. If atoms having magnetic moments are
immersed in a magnetic field, the crystal field produces a set of degenerate orbital
levels for the single magnetic ion. An atom with a magnetic moment l has
(2l + 1) possible orientations with respect to the magnetic field.
For the Boltzmann distribution, at a given temperature, only the ground state
and the excited states lower than kBT are occupied. In the case of the magnetic
properties, only the levels with energies slightly larger than kBTc must be con-
sidered. An additional contribution to the specific heat exists: in the simplest case
of nuclear spin I = 1/2, there are two possible spin orientations with equal
degeneracy and energetic splitting DE. The contribution to the specific heat is
given by (see Fig. 1.11) [5, 12, 76]
ffi
2 DE
kB T
DE e
cM ¼ kB N A ffi !2 : ð1:35Þ
kB T DE
kB T
1þe
From (1.35), we see that the temperature at which the maximum of the mag-
netic contribution to the specific heat occurs is fixed by the energy splitting DE of
the levels
DE ffi
DE DE k T
þ2 ¼ 2 e B : ð1:36Þ
kB T kB T
In most cases, DE is small in comparison to the thermal energy kBT. In this case,
from (1.35), we obtain
22 1 Heat Capacity

Stainless Steel 304 [77],
Constantan (57 % Cu, 43 %
Ni), Manganin (87 % Cu,
13 % Mn), 9 % W, 91 % Pt
alloy [78] and Copper [11]
2
DE
cM ¼ kB N A ; ð1:37Þ
2kBT
corresponding to the right part of Fig. 1.11, well beyond the maximum.
In the case of a pure metal, for DE kBT,
cM ¼ cT þ dT 2 : ð1:38Þ
An important example is the specific heat of stainless steel 304 [77] which,
between 70 and 700 mK, can be fitted as
cM ¼ ð465T þ 0:56T 2 Þ½lJ=gK: ð1:39Þ
This result, together with the magnetic specific heats of some alloys containing
paramagnetic atoms and copper for comparison, is shown in Fig. 1.12. Note that
below 0.1 K, magnetic materials as manganin have a specific heat about two
orders of magnitude higher than copper. Data of Fig. 1.12 evidence that the use of
manganin or constantan leads is to be avoided in some applications as wiring of
very low temperature detectors.
In general, the spin may be larger than 1/2 and more than two levels can be
available, but results are qualitatively similar [79] as shown in Fig. 1.13. It is
worth noting that the temperature at which the maximum of the magnetic con-
tribution to the specific heat occurs is determined by the energy splitting DE of the
levels. Hence, for nuclear magnetic moments, which are about a factor of 103
smaller than electronic magnetic moments, this maximum occurs at much lower
temperatures than for the electronic magnetic moments. For example, an electronic
magnetic moment of 1 lB in a field of 1 T leads to a maximum in CM at about 1 K,
whereas a nuclear magnetic moment in the same field gives a maximum in CM at
only about 1 mK. Note also that the maximum value of the specific heat is
Fig. 1.13 Molar specific

heat divided by the gas
constant R as a function of
x-1 = kBT = glB for a
different value of nuclear spin
I (see Table 1.4)
Table 1.4 Position xmax of xe = gelBB = kBT or xn = gnlnB = kBT and value (Cmax = R) of
the maximum of the magnetic specific heat divided by the gas constant as a function of nuclear
spin I (see Fig. 1.13)
I 1/2 3/2 5/2 7/2 9/2
ðkB TÞ1
max
2.399 1.566 1.193 0.976 0.831
Cmax/R 0.439 0.743 0.849 0.899 0.927

FeCl24H2O drawn from data
obtained by Friedberg et al.
[81] and Raquet and
Friedberg [80]. The peak near
1 K is only partly shown, the
highest value of the specific
heat measured being above
25 J/mol K [73]
independent of the energy splitting, being a function of the number of degrees of

freedom (2I + 1), as reported in Fig. 1.13. This means that an electronic para-
magnet with a spin 1/2 in an arbitrary external magnetic field will have CM,
(max) = 0.439 R (depending on the magnitude of its moment and on the magnetic
field this moment is exposed to). On the other hand, a nuclear magnetic moment,
24 1 Heat Capacity
again with a spin 1/2, has the same maximum value of the specific heat, but occurs
at a much lower temperature.
The above considerations on specific heat contributions resulting from inter-
actions between a magnetic moment and a magnetic field can be applied analo-
gously to the specific heat resulting from interactions of an electric quadrupole
moment with an electric field gradient.
An example of magnetic contributions to the specific heat is reported in
Fig. 1.14 and shows the specific heat of FeCl24H2O [80, 81]. Here, the Schottky
anomaly, having its maximum at 3 K, is clearly resolved from the lattice specific
heat as well as from the sharp peak at *1 K, which is due to a transition to
antiferromagnetic order (k peak). As we can see from Fig. 1.14, the magnitude of
the Schottky effect may be several orders of magnitude greater than that due to
lattice and electronic contributions even at liquid helium temperatures. This is
especially relevant in the case of atoms containing a high number of f-unpaired
electrons (e.g., lanthanides or actinides); in fact, in addition to the strong exchange
coupling between electronic spins, one has to take into account the interactions of
nuclear spins with the intense magnetic field generated by the electrons near the
nucleus [82, 83]. For example, this effect is observed in several neptunium inter-
metallic compounds as a consequence of the nuclear hyperfine Schottky term due to
the splitting of the 237Np ground state (I = 5/2) by the hyperfine field [84, 85].
1.6.3 Materials Used for Magnetic Refrigeration
Magnetic refrigeration is a process which can be realized by employing materials

with unusual low temperature heat capacities, in particular, systems which undergo
magnetic ordering transitions at low temperatures (usually rare earths compounds).
Briefly, the magnetocaloric effect is due to the interaction between the electronic
spins and the magnetic field. Just as with the compression of a gas, the isothermal
magnetizing of a paramagnetic or ferromagnetic sample reduces the entropy due to
the onset of magnetic ordering. When the field is switched off, the system restores
the zero field entropy (just as in the expansion of a gas). The four thermodynamic
variables involved in this process (T, S, H, M) are related by Maxwell’s relation

oSðT; HÞ oMðT; HÞ
¼ ð1:40Þ
oH T oT H
which, when integrated for an isothermal-isobaric process, yields
ZH2
oMðT; HÞ
DSM ðT; DHÞ ¼ dH; ð1:41Þ
oT H
H1
Fig. 1.15 Volumetric heat

capacity of rare-earth-based
materials used as
regenerators. To facilitate the
reader, all the Gd compounds
are drawn as solid lines, all
the Ho compounds as dashed
lines, all the Er compounds as
dotted lines, DySb as empty
squares and Nd as empty
circles [86]
where DSM is the variation of the magnetic contribution to the entropy.

In a similar way, it is possible to obtain
ZH2
T oMðT; HÞ
DTad ðT; DHÞ ¼ dH: ð1:42Þ
CðT; HÞ H oT H
H1
It is easy to see that if the enhancement of the magnetic field increases the order
of the system, DSM is negative while DTad is positive and the solid heats up. The
signs of DSM and DTad are opposite if the magnetic field is reduced.
Ferromagnetic materials have maximum (qM/qT)H at T = TC, while para-
magnets only have significant DTad at low temperatures due to the extremely small
heat capacity.
By using these materials, it is possible to refrigerate cryostats to temperatures
below 4 K. For further information about the use for refrigeration of rare-earth-
based compounds, see Sect. 1.6.4.
Finally, the low temperature specific heat of some paramagnetic materials used
in Adiabatic Demagnetization Refrigerators (ADR) can be found in [14, 87, 88]. It
is to be pointed out that cp of these materials usually show a strong dependence on
the magnetic field (see Fig. 1.15).
1.6.4 Heat Capacity of Regenerators for Cryocoolers
The use of lanthanide-based materials as low temperature ‘‘regenerators’’ has led

to some significant advances in low temperature (\20 K) cryogenics since 1990
when Er3Ni replaced Pb in the low temperature stage of a two-stage Gifford–
McMahon (G–M) cryocooler, thus allowing one to lower the limit temperature
26 1 Heat Capacity
Fig. 1.16 Low temperature

specific heat of rare earths
metals (except the radioactive
Pm). References: Er [94], Ce
[95], Eu Yb [96], La Pr Nd
Sm [83], Gd Tb Dy Ho Tm
[82]
from 10 to 4 K [89, 90]. Subsequently, several other lanthanide materials, in

particular Nd [91] and HoCu2 [92], have been utilized for cooling down to *4 K.
Rare earth’s f orbitals contain up to even seven unpaired electrons (Gd3+) that
provide large magnetic moments, thus giving large values of cp near the ordering
transition; in fact, the total orbital quantum number (J) is proportional to the
theoretical magnetic entropy (SM)
SM ¼ R lnð2J þ 1Þ ð1:43Þ
where R is the gas constant and J is the ground state total angular momentum. For
Hund’s rule, J can assume two values J = L + S or J = |L - S| if the external
configuration of the ion is more or less than half-filled, respectively. From (1.43), it
is clear that the larger SM are provided by ions belonging to the second half of the
lanthanide series (from Dy to Yb). An interesting comparison between Er3Ni and
HoCu refrigeration power is reported in Ref. [93].
However, SM is not the only parameter which influences the refrigeration
process. In fact most of the metallic and intermetallic lanthanide compounds show
a peak in the specific heat at 5–15 K (see Fig. 1.16), that is undesirable for
refrigeration below 4 K. The presence of this peak is also the reason why for
samples containing lanthanides it is mandatory to reach temperatures as low as
*380 mK [25–33] in order to estimate (for measurements) the electronic con-
tribution to specific heat (see Table 1.2).
The relative magnetic ordering temperature for lanthanides in a given chemical
environment (e.g., for a series of isomorphous compounds where only the lan-
thanide ion changes) tends to follow the de Gennes factor, F, see Fig. 1.17, [93]
defined as
F ¼ JðJ þ 1Þðg 1Þ2 ð1:44Þ

Fig. 1.17 De Gennes factors

for the whole series of
tripositive lanthanides ions
where g is the Landè factor [97] that for the ground state of lanthanides can be
easily calculated from angular orbital momentum (L) and the total spin (S) as
JðJ þ 1Þ þ SðS þ 1Þ LðL þ 1Þ

g¼1þ : ð1:45Þ
2JðJ þ 1Þ
In addition, the large angular orbital momentum (L) is not quenched as in the
case of light transition metals (e.g., Fe and Cu), thus providing a large coupling
with the total spin (S), namely, the spin-orbit coupling, which can remove the
degeneration of the ground state splitting it in states with different J (recall that
J can assume values from L + S to |L - S|). This is not true in the case of Gd3+
which has an L = 0 and, of course, only one possible value of J (J = S).
Another aspect that has to be considered is the zero field splitting (due to the
electrical charges of the atoms surrounding the metal ion, known in chemistry as
ligand atoms). This parameter is strictly dependent on the symmetry of the mol-
ecule and agrees well with the fact that the de Gennes factor of complexes with
different symmetry cannot be directly compared, so it is mandatory to have access
to a detailed crystallographic characterization of the studied compound (or series
of compounds) [98].
The last parameter to be taken into account is the RKKY interaction, able to
predict the ferro or antiferromagnetic interaction between lanthanides in a metallic
compound as a function of the distance of the metal center. In this model, [93] the
6 s conduction electrons can provide a spin polarization following an oscillatory
behavior
ðx cos x sin xÞ
f ðxÞ / ð1:46Þ
x4
function of x = 2akF (kF is the Fermi wave vector and a is the radial distance from
a given lanthanide ion). From the oscillatory form of this function, it is clear that a
28 1 Heat Capacity
Table 1.5 Temperature dependence of specific heat in rare-earths-based compounds in 2–22 K

temperature range, expressed in kJ/(m3 K) [56, 100–102]
Temperature (K) 2 K 3 K 4 K 5 K 6 K 7 K 8 K 9 K 10 K 12 K 15 K 20 K 22 K
HoTiO3 – – – 32.7 29.3 31.1 30.3 30.2 32.2 36.2 49.7 83.3 –
DyTiO3 – – – 11.8 9.5 8.4 8.4 9.3 10.3 14.4 26.9 56 –
TbTiO3 – – – 40.7 45.8 51.2 50.3 50.2 48.5 48.2 54.8 72.5 –
GdTiO3 – – – 42.6 34.8 29.3 24.6 23.1 23 23.9 33.3 55.7 –
HoMnO3 – – – 136 139 144 150 154 157 162 174 – –
DyMnO3 – – – 68.3 86.8 93.6 97 98.5 100 109 134 186 –
TbMnO3 – – – 54.9 76.5 92.3 89.6 84.9 82.4 84.8 103 150 –
GdMnO3 – – – 161 170 169 165 157 153 145 145 170 –
HoCoO3 – – – 74.6 78.8 79.9 80 78 77 72.8 71.7 77.6 –
DyCoO3 – – – 21.6 17.3 14.8 16.7 18.6 21.8 30 47.7 90.6 –
TbCoO3 – – – 126 152 184 162 146 139 127 113 114 –
GdCoO3 – – – 66.8 44.8 33.5 28.2 24.9 21 19.9 23.7 42.8 –
Er3Ni 31 66 141 227 320 412 364 305 308 355 444 622 683
HoCu2 – – 252 342 430 486 359 413 383 232 247 322 367
Er3NiH6 80 75 50 40 50 50 55 60 75 100 200 320 378
Er3NiH8 60 55 50 45 45 45 50 55 63 90 136 235 295
Er96.8O2.7N0.3C0.2 – – 21 26 38 55 82 100 148 240 359 810 1000
Gd2SO2 – – 491 899 232 158 122 110 96 – – – –
given series of isomorphous compounds can have either ferro or antiferromagnetic

coupling as a function of the atomic radius of the lanthanide [99].
More recently, Er and Er–Pr alloys (up to 50 % Pr) have been suggested as a
replacement for Pb as the intermediate temperature (*10 – *60 K) range
regenerators. Today, research continues for determining improved regenerator
materials (especially below 10 K).
In Table 1.5, data of specific heat of several materials candidate as regenerators
in 2–25 K temperature range are reported.
1.7 Specific Heat Due to the Amorphous State
In the noncrystalline (amorphous) state, the solids (glasses, polymers…) do not

present a regular and periodic arrangement of atoms in a lattice. As shown in
Fig. 1.18, the specific heat of both an amorphous and a crystalline polymer has a
steep dependence on temperature, though in the case of the amorphous material,
the behavior is more complex.
For temperatures below about 80 K, the shape of the c(T) curves for the various
amorphous materials do not differ very much and are, at a first approximation,
independent of the chemical composition. Before the 1970s, an explanation of how
the difference in the microscopic structure could produce such a difference in the
material properties did not exist. It was believed that Debye’s theory should also
1.7 Specific Heat Due to the Amorphous State 29
Fig. 1.18 Typical

temperature dependence of
the specific heat for an
amorphous (straight line) and
a crystalline polymer (dashed
line)
explain the properties of the amorphous solids since at low temperature, the
wavelength and mean free path of phonons increase; as a consequence, defects and
irregularity in the structure of the solid should not play an important role. The first
systematic work of measure and collection of data both of specific heat and
thermal conductivity of noncrystalline solids was due to Zeller and Pohl [103].
Their revolutionary hypothesis was that the excess in specific heat observed below
1 K in amorphous polymers and glasses was to be attributed to low energy
localized excitations (e.g., vibrations of atoms or of groups of atoms) capable of
producing scattering centers for phonons. Only the measurements of thermal
conductivity carried out by Zaitlin and Anderson in 1975 instead demonstrated that
below 1 K, the acoustic phonons are mainly responsible for the heat transfer [104].
The excitations which produce the excess (the quasilinear contribution, see (1.47))
of specific heat cannot carry thermal energy because they are to be considered as
localized excitations. These excitations may be represented by two level systems
(TLS) or, more generally, by strongly anharmonic oscillators. The linear contri-
bution to the specific heat must be attributed to the thermal excitation of the TLS.
A physical explanation is given in the frame of the tunneling model proposed
independently by Anderson et al. [105] and Phillips [106] in 1972 with the aim of
explaining the measured thermal and acoustic properties of amorphous materials.
According to this theory, because of the structural disorder, groups of atoms have
more than one possible position which corresponds to a small energy difference
(see Fig. 1.19). The typical excitation energy of TLS is of the order of 10-4 eV
and the quantum tunneling transition between the two levels can only take place
with absorption or emission of phonons in order to conserve the energy.
As an example, in Fig. 1.19, a schematic two-dimension representation of the
structure of cristobalite (a crystalline form of SiO2) and of vitreous SiO2 is shown.
Three cases of double possible equilibrium positions for the atoms of the material
in the amorphous state are drawn [107]. Atoms can tunnel from one position to
another. The thermal excitation of TLS is responsible for the linear contribution to
the specific heat of amorphous solids.
30 1 Heat Capacity
Fig. 1.19 Schematic two-dimension representation of the structure of cristobalite (a crystalline

form of SiO2, left) and of vitreous SiO2 (right). Si atoms are represented by full circles, and
oxygen by open circles
The model proposed by Anderson and Phillips provides a phenomenological

explanation of the properties of amorphous materials without supplying a detailed
microscopic description [108]. Low-temperature measurements of the specific heat
of amorphous solids have, however, shown that instead of a linear contribution as
expected from the TLS theory, the best fit of data is obtained with an over-linear
term of the type [109]
c ¼ aT 1þd with 0:1\d\0:5: ð1:47Þ
This does not mean that the TLS theory is wrong, but only that some
approximations are to be revised. In fact, the TLS theory does not take into
account the processes of absorption and emission of a phonon by a TLS which lead
to relaxation phenomena in the tunneling levels. The speed and time by which the
perturbed TLS systems relax to the equilibrium thermal populations depend on the
TLS characteristics and, in particular, on the coupling energy of tunneling states.
Anderson et al. [105] also pointed out that an important consequence of the
tunneling model was the logarithmic dependence of the measured specific heat on
the time needed for the measurement of C. This phenomenon is due to the large
energy spread and relaxation time of TLS. In 1978, Black [110], by a critical
revision of the tunneling theory, was able to explain the time dependence of the
low-temperature specific heat. We also wish to mention the discovery of an
apparently absurd magnetic-field dependence of specific heat measured in some
multicomponent glasses in the 0.3–4 K range [111]. An explanation of the phe-
nomenon can be found in [112]. In the case of polymers, the complex chain
structure allows for a variety of vibrational motions which give a contribution to
the specific heat and characterize other properties of the material. Vibrations
taking place along a chain or among the chains are called stretching or longitudinal
vibrations, whereas the bending or transversal vibrations, characteristic of the
specific polymer, depend on the intrinsic bending stiffness of the polymeric chains
[113]. The latter are responsible for peculiar temperature dependences of the
specific heat and of the thermal expansion of some polymeric materials (e.g., they
1.7 Specific Heat Due to the Amorphous State 31

TORLON in the range
0.15–1 K and the related
fitting curve [118]
may produce negative expansion coefficients). Even these types of vibrations can
be described as phonons whose density distribution depends on the vibration
mode. For each vibration mode, a state density function can be obtained together
with the respective contribution to the specific heat. Each i-mode has its own
Debye temperature hDi, above which the specific heat contribution tends to the
Dulong–Petit limit. In fact, for each mode, there are two limiting cases [114–117]:
T hDi ! ci ¼ constant ð1:48Þ
T hDi ! ci ¼ aT 1þd with 0\d\2: ð1:49Þ
d depends on the vibrational modes which contribute to the specific heat.

The overall specific heat of a polymer is given by the sum of the various
contributions to the specific heat of longitudinal and transversal phonons. At
temperatures below 1 K, the linear contribution due to the TLS must be added.
All amorphous materials, in summary, show a specific heat with a cubic and an
overlinear contribution:
c ¼ bT 3 þ aT 1þd : ð1:50Þ
The almost linear contribution and the heat release from the sample must be
carefully considered when an amorphous material is used at very low temperatures.
In most cases, the c(T) of amorphous materials can be interpreted as being due
to a few contributions previously described in this section. For example, specific
heat data [118] of TORLON (a polyamide imide polymer with excellent
mechanical properties) in the 0.16–1 K temperature range, shown in Fig. 1.20, can
be represented by (1.45) where: b = (2.82 + 0.03) 9 10-5 J K-4 g-1,
a = (5.41 + 0.08) 9 10-6 J K-(2+d) g-1 and d = 0.28 + 0.01. The fit curve
obtained for TORLON is in the shape predicted by the tunneling theory for the
amorphous materials [106, 113] and d is within the range of values obtained for
other disordered solids [113].
32 1 Heat Capacity
Fig. 1.21 Specific heat for

technical materials from 1 to
300 K [16, 17, 119–122]
1.8 Conclusion
Figure 1.21 is a plot of the specific heat of a variety of materials used in

cryogenics.
Pure metals (Fe, Cu, Al, Be) show a linear dependence at low temperatures
(T \ 10 K) due to the electron contribution followed by a transition range where
cp is proportional to T3 (phonon contribution) and finally approaches a near con-
stant value for T [ 100 K (Dulong and Petit value).
The metallic alloys (stainless steel, brass) generally don’t show the linear region
due to a smaller contribution by free electrons; otherwise, their behavior is similar
to that of pure metals.
Nonmetals (Pyrex, glass resin, Teflon) show a dependence on T3 at low tem-
peratures due to the dominance of the phonon excitations, added to a quasilinear
contribution below *1 K.
Finally, superconducting materials (not reported in Fig. 1.21) display a phase
transition with a discontinuity in the specific heat which, below Tc, decreases
rapidly to values close to zero.
References
1. Reed, R.P., Clark, A.F.: Materials at Low Temperatures. American Society for Metals,
Metals Park, US (1983)
2. DiMarzio, E., Dowell, F.: Theoretical prediction of the specific heat of polymer glasses.
J. Appl. Phys. 50(10), 6061–6066 (1979)
3. Sakiadis, B.C., Coates, J.: Prediction of specific heat of organic liquids. AlChE J. 2(1),
88–93 (1956)
4. Turney, J.E.: Predicting Phonon Properties and Thermal Conductivity Using Anharmonic
Lattice Dynamics Calculations. Carnegie Mellon University, Pittsburgh (2009)
References 33
5. Ashcroft, N.W., Mermin, N.D.: Solid State Physics. Holt, Rinehart and Winston, New York
(1976)
6. Rosenberg, H.M. (ed.): The Solid State. Clarendon Press, Oxford (1984)
7. Cezairliyan, A., Ho, C.Y.O. (eds.): Specific Heat of Solids, ed. by C.Y.O. Ho and A.
Cezairliyan. Hemisphere Publishing Corp., New York (1988)
8. Gopal, E.S.R. (ed.): Specific Heats at Low Temperatures. Plenum Press, New York (1966)
9. Barron, T.H.K., White, G.K. (eds.): Heat Capacity and Thermal Expansion at Low
Temperatures. Plenum Press, New York (1999)
10. Kittel, C. (ed.): Introduction to Solid State Physics, 8th edn. Wiley, New York (2005)
11. Ibach, H., Lüth, H. (eds.): Solid-State Physics, an Introduction to Theory and Experiments.
Springer, Berlin (1991)
12. Ziman, J. (ed.): Electrons and Phonons. Clarendon Press, Oxford (1972)
13. McClintock, P.V.E., Meredith, D.J., Wigmore, J.K. (eds.): Matter at Low Temperatures.
Blackie, London (1984)
14. Ventura, G., Risegari, L.: The Art of Cryogenics: Low-Temperature Experimental
Techniques. Elsevier, Amsterdam (2007)
15. Petit, A.T., Dulong, P.L.: Recherches sur quelques points importants de la Théorie de la
Chaleur. Annales de Chimie et de Physique 10, 395–413 (1819)
16. Corruccini, R.J., Gniewek, J.J.: Specific Heats and Enthalpies of Technical Solids at Low
Temperatures: A Compilation from the Literature. National Bureau of Standards,
Washington, DC (1960)
17. Chang, S.S.: Heat capacity and thermodynamic properties of polyvinylchloride. J. Res. Natl.
Bur. Stand. 82, 9–17 (1977)
18. Sirdeshmukh, D.B., Sirdeshmukh, L., Subhadra, K.: Atomistic Properties of Solids, vol.
147. Springer, Berlin (2011)
19. Rosen, M., Kalir, D., Klimker, H.: Single crystal elastic constants and magnetoelasticity of
holmium from 4.2 to 300 K. J. Phys. Chem. Solids 35(9), 1333–1338 (1974)
20. Klein, M., Goldman, V., Horton, G.: Thermodynamic properties of solid Ar, Kr and Xe
based upon a short range central force and the improved self-consistent phonon scheme.
J. Phys. Chem. Solids 31(11), 2441–2452 (1970)
21. Finegold, L., Phillips, N.E.: Low-temperature heat capacities of solid argon and krypton.
Phys. Rev. 177(3), 1383 (1969)
22. Anderson, O.L.: A simplified method for calculating the Debye temperature from elastic
constants. J. Phys. Chem. Solids 24(7), 909–917 (1963)
23. Somasundari, C., Pillai, N.N.: Debye temperature calculation from various experimental
methods for—grown from aqueous solution. IOSR J. Appl. Phys. (IOSR-JAP) 3(5), 1–7
(2013). http:\\www.iosrjournals.org
24. Phillips, N.E.: Low-temperature heat capacity of metals. Crit. Rev. Solid State Mater. Sci.
2(4), 467–553 (1971)
25. Lounasmaa, O.: Specific heat of holmium metal between 0.38 and 4.2 K. Phys. Rev. 128(3),
1136 (1962)
26. Lounasmaa, O.: Specific heat of lutetium metal between 0.38 and 4 K. Phys. Rev. 133(1A),
A219 (1964)
27. Lounasmaa, O.: Specific heat of gadolinium and ytterbium metals between 0.4 and 4 K.
Phys. Rev. 129(6), 2460 (1963)
28. Lounasmaa, O.: Specific heat of samarium metal between 0.4 and 4 K. Phys. Rev. 126(4),
1352 (1962)
29. Lounasmaa, O.: Specific heat of thulium metal between 0.38 and 3.9 K. Phys. Rev.
134(6A), A1620 (1964)
30. Lounasmaa, O.: Specific heat of praseodymium and neodymium metals between 0.4 and
4 K. Phys. Rev. 133(1A), A211 (1964)
31. Lounasmaa, O.: Specific heat of cerium and europium metals between 0.4 and 4 K. Phys.
Rev. 133(2A), A502 (1964)
34 1 Heat Capacity
32. Lounasmaa, O., Guenther, R.: Specific heat of dysprosium metal between 0.4 and 4 K. Phys.
Rev. 126(4), 1357 (1962)
33. Lounasmaa, O., Roach, P.R.: Specific heat of terbium metal between 0.37 and 4.2 K. Phys.
Rev. 128(2), 622 (1962)
34. Dreyfus, B., Goodman, B., Lacaze, A., Trolliet, G.: The specific heat of rare earth metals
between 0.5 and 4 K. Compt. Rend. 253, 1764–1766 (1961)
35. Peruzzi, A., Gottardi, E., Pavese, F., Peroni, I., Ventura, G.: Investigation of the titanium
superconducting transition as a temperature reference point below 0.65 K. Metrologia
37(3), 229 (2000)
36. Van der Hoeven Jr, B., Keesom, P.: Specific heat of mercury and thallium between 0.35 and
4.2 K. Phys. Rev. 135(3A), A631 (1964)
37. Tilley, D.R., Tilley, J.: Superfluidity and Superconductivity. CRC Press, Boca Raton (1990)
38. Biondi, M.A., Forrester, A.T., Garfunkel, M.P., Satterthwaite, C.B.: Experimental evidence
for an energy gap in superconductors. Rev. Mod. Phys. 30(4), 1109–1136 (1958)
39. Phillips, N.E., Lambert, M.H., Gardner, W.R.: Lattice heat capacity of superconducting
mercury and lead. Rev. Mod. Phys. 36(1), 131–134 (1964)
40. Tinkham, M. (ed.): Introduction to Superconductivity. McGraw-Hill, New York (1975)
41. Rose-Innes, A.C., Rhoderick, E.H. (eds.): Introduction to Superconductivity. Pergamon
Press, London (1977)
42. Phillips, N.E.: Heat capacity of aluminium between 0.1 and 4.0 K. Phys. Rev. 114(3),
676–685 (1959)
43. Pobell, F.: Matter and methods at low temperatures. Springer, Berlin (2007)
44. Zemansky, M.W. (ed.): Heat and Thermodynamics. McGraw-Hill, New York (1968)
45. Rosenberg, H.M.: The thermal conductivity of metals at low temperatures. Philos. Trans.
R. Soc. Lond. Ser. A Math. Phys. Sci. 245, 441–497 (1955)
46. Smith, J.L., Haire, R.G.: Superconductivity of americium. Science 200(4341), 535–537
(1978)
47. Myers, H.P. (ed.): Introductory Solid State Physics, 2nd edn. CRC Press, Boca Raton (1997)
48. Debessai, M., Matsuoka, T., Hamlin, J., Schilling, J., Shimizu, K.: Pressure-induced
superconducting state of europium metal at low temperatures. Phys. Rev. Lett. 102(19),
197002 (2009)
49. Dunn, K., Bundy, F.: Pressure-induced superconductivity in strontium and barium. Phys.
Rev. B 25(1), 194 (1982)
50. Wittig, J., Probst, C., Schmidt, F., Gschneidner Jr, K.: Superconductivity in a new high-
pressure phase of scandium. Phys. Rev. Lett. 42(7), 469 (1979)
51. König, R., Schindler, A., Herrmannsdörfer, T.: Superconductivity of compacted platinum
powder at very low temperatures. Phys. Rev. Lett. 82(22), 4528 (1999)
52. Shimizu, K., Kimura, T., Furomoto, S., Takeda, K., Kontani, K., Onuki, Y., Amaya, K.:
Superconductivity in the non-magnetic state of iron under pressure. Nature 412(6844),
316–318 (2001)
53. Eremets, M.I., Struzhkin, V.V., Mao, H.-K., Hemley, R.J.: Superconductivity in boron.
Science 293(5528), 272–274 (2001)
54. Struzhkin, V.V., Hemley, R.J., Mao, H.-K., Timofeev, Y.A.: Superconductivity at 10–17 K
in compressed sulphur. Nature 390(6658), 382–384 (1997)
55. Shimizu, K., Suhara, K., Ikumo, M., Eremets, M., Amaya, K.: Superconductivity in oxygen.
Nature 393(6687), 767–769 (1998)
56. Gschneidner, K.A., Bünzli, J.-C., Pecharsky, V.K.: Handbook on the physics and chemistry
of rare earths, vol. 34. (Access Online via Elsevier, Amsterdam, 2004)
57. Buzea, C., Robbie, K.: Assembling the puzzle of superconducting elements: a review.
Supercond. Sci. Technol. 18(1), R1 (2005)
58. Duan, D., Meng, X., Tian, F., Chen, C., Wang, L., Ma, Y., Cui, T., Liu, B., He, Z., Zou, G.:
The crystal structure and superconducting properties of monatomic bromine. J. Phys.:
Condens. Matter 22(1), 015702 (2010)
References 35
59. Corak, W.S., Goodman, B.B., Satterthwaite, C.B., Wexler, A.: Atomic heats of normal and
superconducting vanadium. Phys. Rev. 102(3), 656–661 (1956)
60. Schoenberg, D. (ed.): Superconductivity. Cambridge University Press, Cambridge (1952)
61. Ekin, J. (ed.): Experimental Techniques for Low Temperature Measurements. Oxford
University Press, Oxford (2006)
62. Ehrenfest, P., Klein, M.J., Casimir, H.: Collected Scientific Papers. North-Holland,
Amsterdam (1959)
63. Stokka, S., Fossheim, K., Ziolkiewicz, S.: Specific heat at a first-order phase transition:
SbSI. Phys. Rev. B 24(5), 2807 (1981)
64. Leupold, H., Boorse, H.: Superconducting and normal specific heats of a single crystal of
niobium. Phys. Rev. 134(5A), A1322 (1964)
65. Robinson, W., Friedberg, S.: Specific heats of NiCl26H2O and CoCl26H2O between 1.4
and 20 K. Phys. Rev. 117(2), 402 (1960)
66. Granato, A.: Thermal properties of mobile defects. Phys. Rev. 111(3), 740–746 (1958)
67. Martin, D.L.: The specific heat of copper from 20 to 300 K. Can. J. Phys. 38(1), 17–24
(1960). doi:10.1139/p60-003
68. Collings, E., Jelinek, F., Ho, J., Mathur, M., Timmerhaus, K., Reed, R., Clark, A.: Advances
in Cryogenic Engineering, vol. 22, pp. 159. Plenum Press, New York (1977)
69. Gopal, E.: Magnetic contribution to specific heats. In: Specific Heats at Low Temperatures,
pp. 84–111. Springer, New York (1966)
70. Kahn, O.: Molecular Magnetism. VCH, Weinheim (1993)
71. Schweiger, A., Jeschke, G.: Principles of Pulse Electron Paramagnetic Resonance
Spectroscopy. Oxford University Press, Oxford (2001)
72. Sanchez, J., Griveau, J.-C., Javorsky, P., Colineau, E., Eloirdi, R., Boulet, P., Rebizant, J.,
Wastin, F., Shick, A., Caciuffo, R.: Magnetic and electronic properties of NpCo2: evidence
for long-range magnetic order. Phys. Rev. B 87(13), 134410 (2013)
73. De Jongh, L., Miedema, A.: Experiments on simple magnetic model systems. Adv. Phys.
50(8), 947–1170 (2001)
74. Van Kranendonk, J., Van Vleck, J.H.: Spin waves. Rev. Mod. Phys. 30(1), 1–23 (1958)
75. Hemingway, B.S., Robie, R.A.: Heat capacity and thermodynamic functions for gehlenite
and staurolite: with comments on the Schoitky anomaly in the heat capacity of staurolite.
Am. Mineral. 69, 307–318 (1984)
76. Berman, R. (ed.): Thermal Conduction in Solids. Clarendon Press, Oxford (1976)
77. Hagmann, C., Richards, P.: Adiabatic demagnetization refrigerators for small laboratory
experiments and space astronomy. Cryogenics 35(5), 303–309 (1995)
78. Ho, J.C., Phillips, N.E.: Tungsten-Platinum alloy for heater wire in calorimetry below
0.1 K. Rev. Sci. Instrum. 36(9), 1382 (1965)
79. Pobell, F. (ed.) Matter and Methods at Low Temperature, 2nd edn. Springer, Berlin (1991)
80. Friedberg, S., Raquet, C.: The heat capacity of Cu(NO3)22.5H2O at low temperatures.
J. Appl. Phys. 39(2), 1132–1134 (1968)
81. Friedberg, S., Cohen, A., Schelleng, J.: The specific heat of FeCl24H20 between 1.1 and
20 K. In: DTIC Document (1961)
82. Lounasmaa, O., Sundström, L.J.: Specific heat of gadolinium, terbium, dysprosium,
holmium, and thulium metals between 3 and 25 K. Phys. Rev. 150(2), 399 (1966)
83. Lounasmaa, O., Sundström, L.J.: Specific heat of lanthanum, praseodymium, neodymium,
and samarium metals between 3 and 25 K. Phys. Rev. 158(3), 591 (1967)
84. Colineau, E., Javorský, P., Boulet, P., Wastin, F., Griveau, J., Rebizant, J., Sanchez, J.,
Stewart, G.: Magnetic and electronic properties of the antiferromagnet NpCoGa5. Phys.
Rev. B 69(18), 184411 (2004)
85. Sanchez, J., Aoki, D., Eloirdi, R., Gaczyński, P., Griveau, J., Colineau, E., Caciuffo, R.:
Magnetic and electronic properties of NpFeGa5. J. Phys. Condens. Matter 23(29), 295601
(2011)
86. srdata.nist.gov
36 1 Heat Capacity
87. Bromiley, P.A.: Development of an Adiabatic Demagnetisation Refrigerator for Use in

Space. University of London, London (2000)
88. Wikus, P., Burghart, G., Figueroa-Feliciano, E.: Optimum Operating Regimes of Common
Paramagnetic Refrigerants. Cryogenics 51(9), 555–558 (2011)
89. Kuriyama, T., Hakamada, R., Nakagome, H., Tokai, Y., Sahashi, M., Li, R., Yoshida, O.,
Matsumoto, K., Hashimoto, T.: High efficient two-stage GM refrigerator with magnetic
material in the liquid helium temperature region. In: Proceedings of the 1989 Cryogenic
Engineering Conference. Advances in Cryogenic Engineering, vol. 35B, pp. 1261–1269
(1990)
90. Sahashi, M., Tokai, Y., Kuriyama, T., Nakagome, H., Li, R., Ogawa, M., Hashimoto, T.:
New magnetic material R3T system with extremely large heat capacities used as heat
regenerators. Adv. Cryog. Eng. 35(Part B) (1990)
91. Ackermann, R.A.: Cryogenic Regenerative Heat Exchangers. Springer, Berlin (1997)
92. Satoh, T., Onishi, A., Umehara, I., Adachi, Y., Sato, K., Minehara, E.: A Gifford-McMahon
cycle cryocooler below 2 K. In: Cryocoolers 11, pp. 381–386. Springer, New York (2002)
93. Gschneidner, K.A., Jr., Pecharsky, A.O., Pecharsky, V.K.: Low temperature cryocooler
regenerator materials. In: Ross, R., Jr. (ed.) Cryocoolers 12, pp. 457–465. Springer, New
York (2002)
94. Hill, R., Cosier, J., Hukin, D.: The specific heat of erbium from 0.4 to 23 K. J. Phys. F Met.
Phys. 14(5), 1267 (1984)
95. Parkinson, D., Roberts, L.: The atomic heat of cerium between 1.5 and 20 K. Proc. Phys.
Soc. Sect. B 70(5), 471 (1957)
96. Lounasmaa, O.: Specific heat of europium and ytterbium metals between 3 and 25 K. Phys.
Rev. 143(2), 399 (1966)
97. Gatteschi, D., Sessoli, R., Villain, J.: Molecular Nanomagnets. Oxford University Press,
Oxford (2006)
98. Boulon, M.-E., Cucinotta, G., Luzon, J., Degl’innocenti, C., Perfetti, M., Bernot, K., Calvez,
G., Caneschi, A., Sessoli, R.: Magnetic anisotropy and spin-parity effect along the series of
lanthanide complexes with DOTA. Angewandte Chemie (International ed. in English)
52(1), 350–354 (2013). doi:10.1002/anie.201205938
99. Kirchmayr, H.R., Poldy, C.A., Groessinger, R., Haferl, R., Hilscher, G., Steiner, W.,
Wiesinger, G.: Magnetic properties of intermetallic compounds of rare earth metals. In:
Handb. Phys. Chem. Rare Earths 2, 55–230 (1979)
100. Konings, R., van Miltenburg, J., Van Genderen, A.: Heat capacity and entropy of
monoclinic Gd2O3. J. Chem. Thermodyn. 37(11), 1219–1225 (2005)
101. Jin, T., Li, C., Tang, K., Chen, L., Xu, B., Chen, G.: Hydrogenation-induced variation in
crystal structure and heat capacity of magnetic regenerative material Er3 Ni. Cryogenics
51(5), 214–217 (2011)
102. Alekseev, P., Znamenskiı̆, N., Lazukov, V., Keı̆lin, V., Kovalev, I., Kruglov, S., Nefedova,
E., Sadikov, I.: Microscopic nature of the extremely high specific heat of rare earth
intermetallic compounds at low temperatures and the possibility of its application in
technical superconductivity. Crystallogr. Rep. 51(1), S79–S84 (2006)
103. Zeller, R., Pohl, R.: Thermal conductivity and specific heat of noncrystalline solids. Phys.
Rev. B 4(6), 2029 (1971)
104. Zaitlin, M.P., Anderson, A.: Phonon thermal transport in noncrystalline materials. Phys.
Rev. B 12(10), 4475 (1975)
105. Anderson, P.W., Halperin, B., Varma, C.M.: Anomalous low-temperature thermal
properties of glasses and spin glasses. Phil. Mag. 25(1), 1–9 (1972)
106. Phillips, W.: Tunneling states in amorphous solids. J. Low Temp. Phys. 7(3–4), 351–360
(1972)
107. Hunklinger, S.: Ultrasonics Symposium Proceedings. In: IEEE, New York, pp. 443 (1974)
108. Hunklinger, S.: Tunneling in amorphous solids. Cryogenics 28(4), 224–229 (1988)
References 37
109. Lasjaunias, J., Ravex, A., Vandorpe, M., Hunklinger, S.: The density of low energy states in
vitreous silica: specific heat and thermal conductivity down to 25 mK. Solid State Commun.
17(9), 1045–1049 (1975)
110. Black, J.L.: Relationship between the time-dependent specific heat and the ultrasonic
properties of glasses at low temperatures. Phys. Rev. B 17(6), 2740–2761 (1978)
111. Meissner, M., Abens, S., Strelow, P.: Hahn-Meitner Institute Report, Berlin (2000)
112. Jug, G.: Theory of the thermal magnetocapacitance of multicomponent silicate glasses at
low temperature. Phil. Mag. 84(33), 3599–3615 (2004)
113. Hartwig, G.: Polymer Properties at Room and Cryogenic Temperatures. Springer, Berlin
(1994)
114. Engeln, I., Meissner, M., Hartwig, G., Evans, D.: Non-metallic Materials and Composites at
Low Temperatures, vol. 2. Plenum, New York (1982)
115. Baur, H.: Über die Wärmekapazität des kristallinen Polyäthylens. Colloid Polym. Sci.
241(1), 1057–1070 (1970)
116. Baur, H.: Bemerkungen zur Wärmeleitfähigkeit und Visko-Elastizität von Polymer-
Festkörpern. Kolloid-Zeitschrift und Zeitschrift für Polymere 247(1–2), 753–762 (1971)
117. Baur, H.: Einfluß der Valenzwinkelsteifigkeit auf die thermischen Schwankungen und den
Debye-Waller-Faktor von Polymer-Kristallen. Kolloid-Zeitschrift und Zeitschrift für
Polymere 250(4), 289–297 (1972)
118. Barucci, M., Di Renzone, S., Olivieri, E., Risegari, L., Ventura, G.: Very-low temperature
specific heat of Torlon. Cryogenics 46(11), 767–770 (2006)
119. Touloukian, Y., Powell, R., Ho, C., Klernens, P.: Thermal conductivity, Thermophysical
Properties of Matter. IFI/Plenum, New York (1970)
120. Johnson, V.J.: A compendium of the properties of materials at low temperature (phase I).
Part II. Properties of solids. In: DTIC Document (1960)
121. White, G.K., Meeson, P.: Experimental techniques in low-temperature physics. In:
Monographs on the Physics and Chemistry of Materials, vol. 59 (2002)
122. Marquardt, E., Le, J., Radebaugh, R.: 11th international cryocooler conference Keystone,
Co. Cryogenic Material Properties Database, National Institute of Standards and
Technology, Boulder, 20–22 June 2000
Chapter 2
How to Measure Heat Capacity at Low
Temperatures
Abstract This chapter is devoted to the description of calorimetric techniques

used to measure heat capacity of solids: pulse heat calorimetry (Sect. 2.3),
relaxation calorimetry (Sect. 2.4), dual slope calorimetry (Sect. 2.5), a.c. calo-
rimetry (Sect. 2.6), differential scanning calorimetry (Sect. 2.7). Examples of
measurements of heat capacity are reported in Sects. 2.3 and 2.4.
2.1 Introduction
Specific heat defined by (1.4) is useful only if the material is homogeneous. In this
chapter, the heat capacity of the sample under measurement will always be con-
sidered in order to also include data about inhomogeneous devices of cryogenic
interest (see, e.g., Ref. [1]).
When a power, P(t), is supplied to an isothermal sample of heat capacity
CS(T) in adiabatic conditions, the sample heating is described by
PðtÞdt ¼ CS ðTÞdT: ð2:1Þ
If the initial temperature (at t = t0) of the sample is T0, at the time t, the sample
temperature will be found by integration of (2.1)
Zt ZT
PðtÞdt ¼ Q ¼ CS ðTÞdT ð2:2Þ
t0 T0
where Q is the total heat supplied to the sample in the time interval (t - t0).
Equation (2.2) finds two basic applications:
(a) Evaluation of Q if CS(T) is known and T is measured (detectors)
(b) Evaluation of CS(T) if both Q and T are measured.

40 2 How to Measure Heat Capacity at Low Temperatures
We are interested in the second application. The apparatus which measure

CS(T) at any temperature is called the ‘‘calorimeter’’.
To measure the heat capacity of a sample at low temperature, we must
refrigerate the material of mass m to the starting temperature T0, isolate it ther-
mally from its environment (for example, by opening a heat switch, [2] and supply
an amount of heat Q to reach the final temperature T. The result is often shown in
the form CS = Q/(T - T0) at the intermediate temperature Ti = (T + T0)/2.
In most cases, a low temperature calorimeter is fabricated following the scheme
of Fig. 2.1. It usually consists of a platform (the sample holder) to which a sample
of heat capacity CS, a thermometer TSH, and a heater H are mechanically and
thermally connected often by glue or vacuum grease. A thermal resistance RTb
links the platform to the thermal bath, while RSH is the thermal resistance between
the sample and the sample holder. Depending on the shape and size of the sample
and on the used experimental method, the thermometer and the heater can be
connected to the sample holder as indicated in Fig. 1.22 or can be directly glued on
the sample. In both cases, a good thermal contact between the sample and the
thermometer has to be reached. If the product RSHCS is much smaller than RTbCSH
(neglecting the time constants associated to heater and thermometer), the tem-
perature of the sample TS is equal to TSH during the measurement. The thermal
contact resistance between the thermometer and the sample holder, and between
the heater and the sample holder are labeled as RCT and RCH, respectively.
As we shall see in Sect. 2.2, a basic distinction can be made between an
adiabatic (RTb = ?) and nonadiabatic (RTb 6¼ ?) situation. The former condition
can only be approximated: no calorimeter is perfectly adiabatic. The more tradi-
tional adiabatic methods are based on a good thermal isolation of the sample, and
the use of a heat switch to connect and disconnect the calorimeter to the bath.
However, heat switching may give rise to experimental problems since, especially
with small samples at very low temperatures, the influence of parasitic heat leaks
may become dominant. Therefore, scientists have developed several techniques in
which there is no need of a complete thermal isolation, and in which the sample is
linked to a heat sink by a thermal conductance RTb 6¼ ?.
In general, the heat capacities of the addenda (sample holder, thermometer,
glues, leads) are small compared to that of the sample; otherwise, addenda heat
capacities have to be known with sufficient accuracy from an additional mea-
surement without a sample, or evaluated by the exact knowledge of their mass and
specific heats (for subtracting it from the total measured value). The leads to the
thermometer and heater must be of low thermal conductance (for measurements at
T ffi TC, the best are thin superconducting wires) and have to be carefully heat-
sunk at a temperature close to the temperature of the sample to avoid heat flow into
it. The heat capacities of wires contribute to the addendum [3]. The thermometer,
heater and the sample should, of course, be thermally well-coupled to the platform
in order to avoid unknown temperature differences. The power supplied to the
thermometer should be small enough to avoid overheating [2]. Parasitic heat losses
or heat inflow by radiation must be reduced by a thermal shield at a temperature
very close to the temperature of the sample. When a heat switch is present, heat
2.1 Introduction 41
Fig. 2.1 Scheme of the main elements of a calorimeter for measurements of heat capacity
produced by opening and/or closing should be small. If an exchange gas is used to

the cool down the calorimeter, it has to be carefully pumped away before carrying
out the measurement. Also, the possibility of adsorption and desorption of residual
gas when the temperature is changed should be taken into account since it involves
heat of adsorption/desorption.
As a result of these problems, heat-capacity data rarely have accuracy better
than 1 %, though more often it is 3–5 %. If high accuracy is needed or the
parameters of the setup are not well known, the calorimeter accuracy can be
validated by measuring the heat capacity of a well-known reference sample [4–9].
Continuous improvements in calorimetry have been achieved due to advances
in electronics, thermometers, microfabrication techniques, and computer auto-
mation. In particular, one has to keep in mind that the accuracy of the thermometer
is a critical parameter in this type of measurement.
2.2 Calorimeters
Calorimetry started in the 18th century with the pioneering studies of Joseph Black
[10] who first introduced the concepts of latent heat and heat capacity. The term
calorimeter is used for the description of an instrument devised to determine heat
and the rate of heat exchange or, vice versa, heat capacity if the first two quantities
are measured, following (2.2).
The design of calorimeters has been modified and adapted for plenty of purposes,
e.g., microcalorimeters and nanocalorimeters are intended to designate calorimeters
in which heat capacities of the order of lJ/K and nJ/K, respectively, can be detected
(see Sect. 2.10). These instruments prompted the study of thermal properties of
layers of molecules (generally in the gas phase) adsorbed on a surface.
Depending on the heat transfer conditions between the sample holder and the
thermal bath, calorimeters can be classified by isothermal, isoperibol, and adia-
batic types. A possible classification and standard nomenclature of calorimeters is
reported in [11, 12].
Isothermal calorimeters have both calorimeter and thermal bath at constant TTb.
If the surroundings are only isothermal, the mode of operation is called isoperibol
[13]. In adiabatic calorimeters, the exchange of heat between the calorimeter and
the shield is kept close to zero by making the thermal conductance as small as
possible. Nevertheless, the thermal insulation of the device can never be perfect as
long as there is a temperature difference between calorimeter and shield. If the
temperature of the shield changes following the temperature of the internally
heated calorimeter, there will be no heat flux by radiation or conduction along the
supporting elements. This heat compensation becomes particularly important
above 100 K, when the radiation heat transfer becomes relevant. The first adiabatic
calorimeter was described in 1911 by Nernst [14], who recognized the necessity of
thermal insulation for low temperature measurements. Adiabatic conditions
become more and more difficult to be fulfilled when the temperature and dimen-
sions of the sample decrease. Semiadiabatic conditions are typically met for
samples with masses between 10 mg and 1 g [15]. Nonadiabatic or isoperibol
conditions exist when the measured heat capacities are so small that the thermal
conductance along the electrical leads cause the sample temperature to decay
exponentially towards the shield temperature.
The use of a sample holder, an external thermometer and an electric heater is a
common feature of these methods. This kind of setup requires the knowledge of
addendum heat capacity, and thus the accuracy of the measurements is limited by
the calibration errors.
Within the three groups, several techniques have been used which often mimic
the methods used to measure the electrical capacitance according to the equiva-
lence Table 2.1.
Table 2.1 must be used with great caution. In fact, it is possible to use the
standard Kirchhoff laws to describe the thermal systems and to solve circuit
equations for T(t) or P(t); however, the thermal quantities, such as the thermal
resistance and the heat capacity, often have properties that rapidly change with
temperature, whereas the electrical quantities, such as capacitance and electrical
resistance, are usually almost independent on the voltage. It is worth pointing out
that there is no correspondence between the electrical inductance L and the kinetic
inductance Lk [16].
The well-known techniques used to solve electric circuit problems can only be
employed for ‘‘small signals’’ (see, e.g., [2]). Note that also the equivalence
between thermal grounding and electrical grounding only holds for small signals.
2.2 Calorimeters 43
Table 2.1 Equivalence Thermal parameter Electrical parameter

between some electrical and
thermal parameters V (voltage) T (temperature)
P (power) I (current)
R (thermal resistance) R (electrical resistance)
C (heat capacity) C (capacitance)
Thermal grounding Electrical grounding
Moreover, the approximation with ‘‘lumped elements,’’ which is an excellent

approximation in electrical circuits at low frequency, fails or is a rough approxi-
mation in ‘‘thermal circuits’’ even if the latter only involves a frequency range of a
few Hz.
Finally, the thermal bath temperature, which is formally equivalent to the
electrical ground, is kept at temperature TTb with fluctuations larger than those of
electric V or I supply. Special care should be devoted to the problem of the
temperature stability of the bath since the refrigerator has a finite cooling power,
and the thermal bath represents a ground (to a good approximation) only in the
case that the incoming power on it does not substantially change its temperature.
In analogy with the electrical I(t), the waveform of P(t) appearing in (2.2) is not
restricted to sinusoidal oscillations, but can have any other waveform, e.g.,
impulsive, rectangular or triangular waveforms have been used [17–19]. The
modulation was also indirectly induced to the sample by giving a modulated power
to the heat shield [20].
Since the pioneering work of Eucken [21] and Nernst [22] in the early 20th
century, adiabatic calorimetry has provided the most accurate means of obtaining
specific heat data. The high accuracy arises from the simplicity of the measure-
ment principle. The adiabatic measurement approach directly comes from the
definition of heat capacity:

DQ
Cp ¼ lim : ð2:3Þ
DT!0 DT p
Due to the general applicability independent of the sample thermal conduc-

tivity, this method is the most favored choice for heat capacity measurements of
condensable gases which have poor thermal conductivity in their low temperature
solid phase [23–25].
Adiabatic calorimetry is a very precise technique and can be used to determine
the latent heat at strong first order transitions. However, it usually lacks in
achieving the resolution needed to characterize the temperature dependence of
Cp(T) close to the critical temperature Tc for a second-order transition. Also,
because of the inherent limitations on getting the ideal adiabatic conditions and the
long time required to cover a few tens of K range with reasonable number of data
points, nonadiabatic techniques (e.g., the AC calorimetry) are often preferred at
low temperatures.
Because of the large quantity of the existing calorimetric methods for the
measurements of the heat capacity, we only selected some of them, describing the
experimental setup and giving some examples of their applications:
• Heat pulse calorimetry (Sect. 2.3)
• Relaxation calorimetry (Sect. 2.4)
• Dual slope calorimetry (Sect. 2.5)
• AC calorimetry (Sect. 2.6)
• Differential scanning calorimetry (Sect. 2.7).
2.3 Heat Pulse Calorimetry
In the heat pulse technique, we can either be in an adiabatic or nonadiabatic

situation. In the former case, (2.3) is applied; in the latter case, the sample is
usually connected to the bath through a weak thermal link. Following a heat pulse
of energy DQ, which is commonly supplied by an electrical heater, the temperature
of the sample first rises and then decays to its initial value with a time constant
s = RTbC, where RTb is the thermal resistance of the link and C is the total heat
capacity of the sample plus addenda (C = CS + CSH + Cadd,). The heat capacity
is obtained through C = DQ/DT where DT is a suitable extrapolation of the
temperature step (see Fig. 2.2). Note that s should not be too small, even at the
lowest temperatures (s C 1 s, in most cases) because of the time response of
measuring instruments. The DT(t) curve after a heat pulse has the shape shown in
Fig. 2.2. The temperature difference DT is also obtained by extrapolating the log
plot of the DT(t) curve to the zero time (the time of the end of the heat pulse). Heat
pulse calorimetry has been used, e.g., in measurements reported in [26–32].
2.3.1 Example 1: Heat Pulse Calorimeter for a Small Sample

at Temperatures Below 3 K
As a typical example of the heat pulse method, we will describe the measurement
of the specific heat of Cu and amorphous Zr65Cu35 reported in [33]. Figure 2.3
shows the experimental setup for the measurement of heat capacity: the sample is
glued onto a thin Si support slab. The thermometer is a doped silicon chip and the
heater is made by a gold deposition pattern (*60 nm thickness) on the Si slab.
Electrical wiring to the connecting terminals are made of superconductor (NbTi).
The thermal conductance to the thermal bath (i.e., mixing chamber of a dilution
refrigerator) is due to four thin nylon threads. The silicon slab, the thermometer
and the heater represent the ‘‘addendum’’ whose heat capacity CA(T) must be
measured in a preliminary run.
2.3 Heat Pulse Calorimetry 45
Fig. 2.2 Typical DT(t) curve. The inset shows the linear fit applied to the exponential decay after
the peak
Fig. 2.3 Sample holder for the measurement of heat capacity [33]
When a sample of heat capacity C(T) is added, a second run of measurements

gives CA(T) + C(T). It is obvious that, if possible, the condition CA ffi C should
be fulfilled. For the heat pulse technique, the sample is thermally connected to the
cold source through a weak link. Following a heat pulse of energy Q, which is
delivered by means of the electrical heater, the temperature of the sample first rises
and then decays to its initial value with a time constant r = RLCT. Here, RL is the
thermal resistance of the link and CT is the total heat capacity of the sample and
addenda (sample holder, heater, thermometer, etc.). CT is obtained through
CT = DQ/DT where DT is a suitable extrapolation of the temperature step.
In this experimental setup, the nylon threads fixing the sample holder provide a
sufficient thermal coupling between the sample holder and mixing chamber at low
temperatures. Therefore, the sample holder is precooled down to about 20 K with
H2 as the exchange gas. It was checked experimentally that the thermal coupling
occurs through nylon threads and not through NbTi wires.
The choice of the thermal link is a compromise between two conflicting
requirements. In fact, the value of RL must be rather large since CT is very small
and, as we noticed, a suitably large time constant is needed; small values of RL
would be necessary to avoid an excessively large temperature drop DT = R P
between mixing chamber and sample holder to prevent parasitic pick up.
In the case of this example, a heat leak of P = 1 nW resulted in DT & 50 mK.
Heat pulses (with a duration of sH & 10 ms at 0.1 K and 0.1 s at 1 K) were
applied with a conventional pulse generator.
The energy input DQ = (V2/RH) sH was determined by a measurement of V, RH
and sH. The power dissipation in the Si thermometer could be kept below
10-14 W.
The expression [3, 34]
h i
lnðlnðR=R0 ÞÞ
A0
T ¼ T0 e ð2:4Þ
was used to fit the data points, resistance R versus temperature T, with the con-
stants T0, R0 and A0 determined by the fit [35].
From the curve DT(t) = Ti - T(t) (where Ti is the stationary initial temperature
value before the heat pulse), the temperature difference DT can be obtained by
extrapolating the DT(t) curve to zero time (see Fig. 2.2).
The heat capacity of the empty sample holder (addendum) was
Cadd = aT + bT3 with a = 5.6 10-8 J K-2 and b = 9.9 10-8 J K-4. The T3
contribution to Cadd is explained quantitatively with the Debye heat capacity of the
Si plate plus a small contribution arising from the Au wires, grease and contri-
bution of one-third of heat capacity of the nylon threads. The linear term of Cadd
arises instead from the conductors (Si, Au), from insulators (grease and nylon) and
the remaining contribution (&2 10-8 J K-2) probably stems from the degener-
ate n pads of the Si thermometer. A contribution of the same size has been
previously observed in Si thermometers [36].
The resulting measurements carried out by this apparatus are shown in Fig. 2.4.
2.3 Heat Pulse Calorimetry 47
Fig. 2.4 a Specific heat C of Cu as a function of temperature T (log–log), of 40 mg Cu. The solid
line indicates the Cu standard reference as determined between 0.4 and 3 K. The inset shows the
observed additional contribution DC = C - cT - bT3. b Specific heat C of amorphous Zr65Cu35
as a function of temperature T (log–log). The arrow marked TC indicates the superconductive
transition as determined resistively [33]
2.3.2 Example 2: Heat Pulse Calorimetry

for the Measurement of the Specific Heat of Liquid 4He
Near its Superfluid Transition
One of the most interesting measurements using heat pulse calorimetry was carried
out onboard the Space Shuttle (October 1992) [32]. The objective of the mission
was to measure the specific heat at a constant pressure of liquid 4He near its
superfluid transition with the effect of gravity removed [37, 38]. In these experi-
ments, C was measured with sub-nanokelvin resolution at temperatures within one
nanokelvin of the transition temperature Tk = 2.177 K. Such an extreme tem-
perature resolution is only meaningful for the investigation of a phase transition of
liquid helium because purity is high enough only in this substance, and thus the
phase transition shows the required sharpness. In all other materials, the phase
transitions are smeared by impurities and by imperfections of the structure. In
addition, these measurements had to be carried out in reduced gravity in order to
decrease the rounding of the transition caused by gravitationally induced pressure
gradients and therefore spreading the transition temperature over the liquid sample
of finite height. The high-resolution magnetic susceptibility thermometers devel-
oped for these experiments are described in [39]. In these experiments, the tem-
perature stability was extremely important: in the experimental setup, four thermal
control stages in series with the calorimeter were actively regulated in tempera-
ture: a stability of less than 0.1 nK/h was reached. Besides this thermal regulation,
the experiment required a very careful magnetic shielding, in particular, of the
electric leads, as well as extremely low electric noise levels. Figure 2.5 shows
averaged data of the heat capacity close to the 4He transition.
Fig. 2.5 Averaged data close

to the transition. The
continuous line shows the
best-fit function [38]
2.4 Relaxation Calorimetry
A relaxation calorimeter (isoperibol) measures the total heat capacity (sample and
addenda) by using a simple relation
C ¼js ð2:5Þ
where j is the thermal conductance of the weak link between the platform and the
thermal reservoir and s is the constant of the temperature relaxation time of the
platform.
Referring again to Fig. 1.22, a sample of heat capacity CS and temperature TS is
fixed on a sample holder of heat capacity CSH and temperature TSH. Initially, for
sake of simplicity, the sample and the sample holder are supposed to be isother-
mal. RSH is the thermal resistance between the sample and the sample holder. The
sample holder, whose temperature is measured by a thermometer TSH, is connected
to a heat bath at TTb by a link of thermal conductance RTb and negligible heat
capacity. A constant power P0 is applied to the sample holder until thermal
equilibrium is achieved. At t = t1, the power is switched off and the sample
temperature TS relaxes toward TTb. In the hypothesis that RTb RSH, the sample
temperature follows
TS ðtÞ TTb ¼ P0 RTb eðt=ðCS þCSH ÞRTb Þ ¼ DTeðt=sÞ : ð2:6Þ
Changing TTb and repeating the measurement, a set of points for

C(T) = (CS + CSH)(T) is obtained. The temperature difference DT must be kept as
small as possible, usually a few percent of TTb, in order to ensure that s can be
considered as a constant. Practical values of s range between about 1 and 1,000 s.
At very low temperatures, the thermal resistance RSH between the sample and
the holder can no longer be neglected because its temperature dependence usually
becomes steeper than that of RTb. This introduces a second time constant s2 and the
decay is described by
2.4 Relaxation Calorimetry 49
TS ðtÞ TTb ¼ A1 eðt=s1 Þ þ A2 eðt=s2 Þ ð2:7Þ
where
A1 þ A2 ¼ DT ¼ P0 RTb : ð2:8Þ
Equation (2.7) can be solved [40] to give

A1 s 1 þ A2 s 2
CS þ CSH ¼ K : ð2:9Þ
A 1 þ A2
In realistic situations, s2 is much smaller than s1 and cannot be measured with

enough accuracy to use Eq. (2.9). Reference [40] gives the useful approximation

CS þ CSH A1 s1=DT RTb ð2:10Þ
which is accurate in most cases within a few percent and avoids the need of
calculating s2. It is worth noting that the above-described ‘‘lumped s2 effect’’ is not
the so-called ‘‘distributed s2 effect’’ due to low thermal conductivity of the sample
itself. This latter case is discussed in [3, 34].
A variation of the relaxation method (see Sect. 2.5) was proposed by Riegel and
Weber [41]. They describe a long (about 10 h) cycle to measure C over several
degrees. In this method, they use an extremely weak thermal link to the heat sink
and record the temperature of the sample while heating at constant power for one-
half of the cycle, then allow the sample to relax while recording the temperature
during the second half of the cycle with zero power input. The heat loss to the bath
and surrounding can be eliminated from the calculation of C using this technique,
provided the bath temperature can be held constant over the 10 h cycle. Note that
this procedure is a particular case of the dual slope method of Sect. 2.5.
In [42, 43], the relaxation method is used with an amorphous silicon-nitride
membrane, supported by a silicon frame, onto which thin-film heaters and ther-
mometers (Pt for T [ 50 K, amorphous NbSi or B doped Si for lower tempera-
tures) are patterned. The heat capacity of this addendum is \l nJ K-1 at 2 K and
only 6 lJ K-1 at 300 K. This calorimeter was used to investigate microgram
samples or thin films in steady fields up to 8T; according to the authors, it should
also be usable in pulsed fields up to 60T. This is the result of the rather weak
dependence of the properties of the calorimeter parts on magnetic field.
Finally, in [44], the heat capacity of holes in heavily doped Ge samples was
measured using the relaxation method, an approximated ‘‘addendum free’’ con-
figuration was obtained using a Ge thermometer with the same doping as the
sample and extremely low capacity addendum components. We report the
description of this experiment in some detail in Sect. 2.4.1.
The limitations of the thermal-relaxation method in properly measuring sharp
features in the specific heat are illustrated, e.g., by the measurements of the
specific heat in the proximity of the first-order antiferromagnetic transition at

T = 14 K in Sm2IrIn8 [45].
The relaxation method has been used for measurement of the specific heat of
several materials as reported in [46, 47] (15–300 K, based on a closed cycle
cryocooler) and also in [3, 34, 35, 38, 44, 48–59].
2.4.1 Example: Measurement of Specific Heat of Heavily

Doped (NTD) Ge
The heat capacity of a NTD (Neutron Transmutation Doped, Ge 34B) Ge sample

[60], 3 mm thick and with a diameter of about 3 cm (12.043 g), was measured in
the 24–80 mK temperature range using the relaxation method [44].
In the realization of the experiment, authors approximated an ‘‘addendum free’’
configuration (see Table 1.7). The experimental setup is shown in Fig. 2.6.
The Ge wafer sample was glued with small spots of GE-varnish onto a Cu
holder in good thermal contact with the mixing chamber of a dilution refrigerator.
Three Kapton foils (2 9 2 9 0.01 mm3 each) electrically isolated the Ge wafer
samples from the holder and realized the thermal conductance G(T) between the
samples and the heat sink.
For the two runs (described later), a calibrated NTD Ge #34B thermistor (same
material of the sample, 3 9 3 9 1 mm3) and a Si heater were used. Electrical
connections were made by means of superconducting NbTi wires 25 lm in
diameter. The connections between the gold wires of both thermistor and heater,
and the NbTi leads were done by crimping the wires in a short Al tube (0.1 mg).
At the ends of the NbTi wires, a four lead connection was adopted. An AVS47 AC
resistance bridge was used for the thermometry, while a four-wire I–V source-
meter (Keithley 236) supplied the current for the Si heater.
The addendum (represented by heater, glue spots, and Al tubes) gave a negli-
gible contribution to the total heat capacity (see Table 2.2). The thermistor heat
capacity was instead considered as part of the sample. The whole experiment was
surrounded by a Cu shield at the mixing chamber temperature, measured by a
calibrated RuO2 thermometer.
Two measurements of heat capacity were carried out at different temperature
range: (1) from 24 to 40 mK, (2) from 40 to 80 mK. For the two runs, the thermal
conductance between the Ge wafers and the heat sink was measured by a standard
integral method (see Ref. [2]).
The best fits of the values obtained in the two runs were
(
G1 ðTÞ ¼ 1:22 104 T 2:52 ½WK 1
: ð2:11Þ
G2 ðTÞ ¼ 3:05 105 T 2:45 ½WK 1
2.4 Relaxation Calorimetry 51
Fig. 2.6 Experimental setup of Ref. [44]
Table 2.2 Estimated heat capacity contributions. Specific heat data references are in [61]
Material Volume C (50 mK) C (40 mK) C (30 mK)
(mm3) (J K-1) (J K-1) (J K-1)
NTD Ge 1950 10-7 8 9 10-8 6 9 10-8
(electrons)
NTD Ge (phonons) 1950 6.8 9 10-10 3.6 9 10-10 1.5 9 10-10
GE-varnish 0.52 1.8 9 10-10 1.45 9 10-10 1.1 9 10-10
Al tubes 0.157 2.65 9 10-13 1.36 9 10-13 6 9 10-14
NbTi wires 0.12 10-12 5.4 9 10-13 2.3 9 10-13
The value of the heat capacity was calculated from equation C = s G, where
the thermal time constant s is obtained from the fit to the exponential relaxation of
the wafer temperature.
Using the known thermal conductivity data of the wafer, the internal thermal
relaxation time was estimated to be less than 1 ms, i.e., much shorter than C/G.
Fig. 2.7 Heat capacity per

gram of the NTD Ge wafer.
The black line represents the
linear fitting
Such an estimate was confirmed by the fact that within the experimental errors, a
single discharge time constant s was always observed [61].
Since in the measured temperature range the Debye temperature of Ge is
*370 K, the phonon contribution to the heat capacity can be neglected [62].
Hence, the heat capacity of the samples is expected to be substantially influenced by
only electron contribution, thus giving a linear dependence from T (see Sect. 1.2).
Let us consider data from the two measurements (see Fig. 2.7). Data can be
well represented by a linear fit which crosses the origin within the experimental
errors. The heat capacity per unit volume of the wafer sample is expressed by the
following formula in the measured temperature range of 24–80 mK:
cðTÞ ¼ ð1:22 0:01Þ 107 ½JK 1 g1 : ð2:12Þ
We can express specific heat in terms of volume:

cðT Þ ¼ c T ¼ ð7:52 0:08Þ 107 T J K1 cm3 : ð2:13Þ
The value of c is close to most of the Sommerfeld constant values reported in the
literature for the NTD Ge of similar doping [63–66]. Note that the theoretical
dependence of c on the compensated dopant concentration is p1/3
c [67].
2.5 Dual Slope Method
In this method (isoperibol), Cp is evaluated by directly comparing the heating and

cooling rates of the sample temperature without need of measuring the thermal
conductance between sample and bath.
The heat capacity can be measured continuously through an extended tem-
perature range, making use of both the heating and the cooling curves. This so-
2.5 Dual Slope Method 53
Fig. 2.8 Example of the

charge–discharge of heat
power in the dual slope
method
called Dual Slope (DS) method was proposed by Riegel and Weber [41], and
Marcenat [68]. It consists of applying a heating power P(t) to the sample holder
(see Fig. 2.8) while continuously monitoring the sample temperature. For exper-
imental setup in which the resistance between sample and holder is almost zero
(RSH & 0), the equations describing the heating and cooling curves, respectively,
are
dTh ðTÞ
CðTÞ ¼ Ph ðTÞ Pl ðTÞ þ Pp ðTÞ ð2:14Þ
dt
oTc ðTÞ
CðTÞ ¼ Pl ðTÞ þ Pp ðTÞ: ð2:15Þ
ot
If we assume that in all the experiments the parasitic power (Pp(T )) and the power
loss via heat link (Pl(T)) only depend on T and T0, it is possible to obtain C(T ) as
Ph ðTÞ
CðTÞ ¼ oT ðTÞ oT ðTÞ : ð2:16Þ
ot ot
h c
Thus, the heat capacity of the sample at a certain temperature T can be obtained
from the slope of the heating and cooling curves measured at T. It is worth noting
that in (2.8), (2.9), and (2.10), the notation C(T ) indicates the total heat capacity
that has to be split in two contributions: the first from the holder (CSH(T )) and the
second from the sample (CS(T )).
In Fig. 2.8, an example of power charge and discharge is reported.
The method is very useful for making a quick scan through a large temperature
range when the shape of the heat capacity curve is unknown, and considerably
speeds up further measurements. As can be seen in (2.10), the dual slope method is
self-correcting regarding parasitic heat leaks. Moreover, it is not necessary to
explicitly know the thermal conduction of the heat link, although in most cases, the
measurement of 1/RTb = Po/DT is easily performed and this quantity can provide
useful additional information for the data analysis. The frequencies with which the
data points are taken will have to be adjusted to meet the requirement that when
using this method, the derivatives with respect to the time of the sample tem-
perature must be determined with great accuracy.
Also, in this method, poor thermal contact between the sample and sample
holder can introduce a second time constant. In that case, it is still possible to
retrieve C(T) if an accurate determination of the second derivatives to the time of
Th(t) and Tc(t) can be made. The latter determination is often very difficult and one
usually has to restrict the use of the DS method to those temperatures at which the
influence of RSH can be neglected.
Although the DS method is very elegant and easy to implement, the technique
has usually been employed for small samples (typically less than 0.5 g) with good
thermal conductivity and at temperatures lower than 20 K. The success of this
method heavily depends on achieving an excellent thermal equilibrium between
the sample, sample holder, and the thermometer, and for large samples with poor
thermal conductivity (i.e., large s2), this method may also fail when we only
consider the first-order approximation of the heat balance equations. Nevertheless,
the DS method has been successfully used around 1 K, reducing to a minimum
CSH [49].
A slightly less sensitive method (variation of the dual slope) involving a fixed
heat input followed by a temperature decay measurement is described in Ref. [47].
In this method, the sample is raised to an equilibrium temperature above the
thermal bath and then allowed to relax to the bath temperature with no heat input.
The method requires extensive calibration or the heat losses of the sample as a
function of temperature between the reservoir and final sample temperature and
relies on accurate, smooth temperature calibrations of thermometers because the
time derivative of the temperature during the decay process is necessary to extract
the specific heat. A data analysis method designed to eliminate the calculation of
the time derivative of the temperature during the decay was also described in Ref.
[47]. The technique requires a determination of many equilibrium heat losses
during the heating portion of the relaxation cycle to obtain good temperature
resolution of the specific heat changes.
A modification of the DS technique which is a hybrid between the AC method
and the DS method and reduces the duration of the measuring cycle has been
proposed in Ref. [69]. The method which was devised to measure specific heats of
small samples has good sensitivity and can give many points on the heat capacity
versus temperature curve during one cycle of about a 2 h duration covering a
temperature range of several K.
As stated, the DS method is time consuming both in performing measurements
and for data analysis. For these reasons, it has been used less than other methods,
despite the inherent advantages hereafter described. Some examples are reported
in [41, 49, 69–71].
2.6 AC Calorimetry 55
2.6 AC Calorimetry
AC calorimetry (isoperibol, also known as modulation calorimetry, TMC) consists

of generating a periodic oscillation of power P(t) with the frequency 2x that heats
the sample of heat capacity CS and in recording the resulting temperature oscil-
lations TS(t) as a function of time. The measured temperature oscillates with the
same frequency and amplitude DTS(t) around a mean temperature TM. A phase
shift u develops between P(t) and TS(t) due to the finite thermal resistance Rtb
between the sample and thermal bath, TTb. The scheme of such a calorimeter is the
same as in Fig. 1.22 with some differences and is shown in Fig. 2.9.
The basic relations for a modulation calorimeter result from the following
common equation that describes, in its simplest form, any calorimetric system

dTS
PðtÞ ¼ CS þ KTb ðTS TTb Þ ð2:17Þ
dt
where KTb = 1/RTb.

If one drives the heater with a current I = I0 cos(xt), Joule heating occurs at a
frequency 2x at the heater. Equation (2.11) becomes
PðtÞ ¼ PAC ei2xt ¼ 2ixCS DT ðtÞe2ixt þ KTb DT ðtÞe2ixt ð2:18Þ
where DT* denotes the complex amplitude of the temperature oscillation.

The solution DT(t) can be written as
DTAC ðtÞ ¼ DT eixt ¼ j jeiu eixt : ð2:19Þ
Introducing the complex heat capacity as C*S = C0 - iC00 , the temperature

oscillation and the phase shift (between power and temperature) are given by the
following equations:
0 PAC
C ¼ sin u ð2:20Þ
2xjDT j
00 PAC KB
C ¼
cos u ð2:21Þ
2xjDT j 2x
PAC
jDT j
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð2:22Þ
ð2xCS Þ2 þ KTb 2
Fig. 2.9 Scheme of a calorimetric measuring cell. TTb temperature of the bath, Tshield
temperature of heat shield, TS temperature of the sample, RTb thermal resistance towards
thermal bath, and CS sample heat capacity. Note that RH is neglected
with
KTb ¼ ½PAC =jDT j cos u ð2:23Þ
u ¼ arctanf2xCS =KTb g: ð2:24Þ
If the thermometer is connected to the sample without thermal resistance and

the sample is thought of as isothermal, amplitude of the thermal oscillations is the
same through the sample and C00 = 0.
Let us note that the modulation method requires either a heat loss by conduction
to the thermal bath through RTb or by radiative cooling toward the shield. As a
consequence, AC calorimetry can never operate under adiabatic conditions and
there always exists a heat flow from the sample to the thermal bath through a
radiation resistance RTb. If RTb xCS, (2.16) yields DTAC & PAC/xCS and the
phase approaches the value u = -(p/2). These are the usual working conditions of
a traditional AC calorimeter and in this case, the system works under quasiadia-
batic conditions [72]. For all other cases, heat losses and phase shifts between
input power and sample temperature must be taken into account using the full
formula (2.16).
Finally, according to all of the above-mentioned equations, the oscillations are
superimposed by a temperature increase DTDC = PAC RTb/2. We note that in con-
trast to other calorimetric methods, in the AC calorimetry, the sample can be located
in vacuum or in exchange gas. Also, modulation is not restricted to sinusoidal
2.6 AC Calorimetry 57
oscillations, but can have any other waveform, e.g., rectangular or triangular wave
forms have been used [18, 19, 73]. Modulation can also be induced indirectly to the
sample by giving a modulated power to the heat shield [20].
A modification of the TMC technique (the 3x-method) is based on the original
work by Corbino [74]. Later, Rosenthal [75] and Filippov [76] used the bridge
technique to measure third harmonic signals. In the original work in 1910–1911,
Corbino [74] used the resistance of electrically conducting samples to determine
the temperature oscillations with a method known as the third-harmonic (3x)
method. In this kind of experiment, the same metal resistor element is used as both
a heater and thermometer. The heater, with resistance R, is driven by a current at
frequency x which results in a power of 2x frequency that causes diffusive
thermal waves which perturb the sensor resistance. The combined effect of driving
current and resistance oscillations gives a voltage across the resistor in a form
containing a first term which is the normal AC voltage at the drive frequency,
while the second and third terms, which derive the current and resistance oscil-
lations from mixing, are dependent on the DT, (the temperature oscillation
amplitude, which, in turn, is related to the sample heat capacity) [77].
Although the 3x signal is relatively small (see, e.g., [72]) compared to the two
terms (oscillating with x and 2x), it can be well separated by the lock-in tech-
nique. For more experimental details, refer to [20, 78–82]. Measurements using
AC calorimetry are reported, e.g., [83–88].
2.7 Differential Scanning Calorimetry
A precious tool for investigating the thermodynamics of chemical reactions and

phase transitions is the differential thermal-analysis (DTA) technique, which is
widely used in chemical and material sciences. In a typical experiment, the
specimen and a reference material of similar heat capacity are heated simulta-
neously. If the two samples are sufficiently thermally insulated from each other,
changes in the temperature difference between the sample and the reference
material reflect heat capacity variations or indicate the occurrence of chemical
reactions. Typical heating rates are of the order of several degrees per minute.
Under such conditions, however, the samples are not always in thermal equilib-
rium. The resulting problems related to the geometry of the samples and the
sample holders make a calculation of absolute heat-capacity data rather compli-
cated, and it is therefore rarely done in practice [3].
The use of high-precision (magnetic field independent) electronic components
makes it possible to achieve a relative accuracy DC/C \ 0.02 % on samples of
milligram weight. This accuracy is at least of the same order of magnitude as that
reached with the frequently used continuous-heating technique where significantly
larger sample masses are usually required. Moreover, the method is time saving
and very simple to apply since neither calibration nor a very precise temperature
regulation is necessary in principle. Heat-capacity measurements can be done upon

heating or cooling the samples.
We shall describe the principle of the measurement [89] referring to the con-
figuration shown in Fig. 2.10. The sample (with heat capacity CS at a temperature
TS) and a reference sample (at a temperature TR, with known heat capacity CR) are
thermally connected to a sample holder (temperature TSH) directly anchored to a
heat reservoir (usually TTb = TSH, and hence the heater H acts on the sample
holder and on thermal bath) via the heat links RSH and RRH, respectively. We might
also include a thermal connection RSR between the sample and the reference
object, but it is possible to thermally isolate the two specimens from each other in
a real experiment, i.e., RSR RSH, RRH. Results of a calculation taking the effect
or a nonzero thermal conductance between the sample and the reference into
account are discussed in an Appendix to [89].
If the temperature of the sample holder TSH varies with time, it is possible to
describe the system by
CS T ¼ kS ðTSH TS Þ ð2:25Þ
CR TR ¼ kRH ðTSH TR Þ ð2:26Þ
where kSH = 1/RSH and kRH = 1/RRH. It is instructive to first consider steady-state
solutions of (2.25) and (2.26) when CS and CR are assumed to be temperature
independent, and TSH(t) is a linear function of time. In order to avoid the thermal
equilibrium problems, we choose the characteristic time constants of the system.
Tj = Cj/kj (with j = S or R) to be larger than the internal equilibrium times of the
samples, sint which are typically less than l s at T = 100 K. In a straightforward
manner, we obtain TS = TR = TSH and TSH - Tj = CjTSH/kj. Assuming identical
heat links (ks = kr), we find that the quantity
DC0 ¼ CR ½ðTS TR Þ=ðTR TSH Þ ð2:27Þ
asymptotically approaches the heat-capacity difference DC = CS - CR at times

t sj after switching on the linear temperature ramp TSH(t). In a more realistic
contest, the heat capacities Cs and CR are temperature dependent, and TSH(t) may
deviate from linearity. It may then be argued that DC0 still approaches DC for
t sj as long as heat-capacity changes and variations of TSH in time are suffi-
ciently slow, i.e., occur on a time scale much larger than sj. The DTA technique
has been used, e.g., in [90, 91].
A remarkable advantage of DSC is that it is easy to measure the temperatures
TS, TR and TSH in an experiment; thus, if kSH = kRH, then DC0 calculated according
to (2.27) represents an excellent approximation for the heat-capacity difference,
and can be determined without knowledge of the absolute values of kSH and kRH.
Note, however, that a sharp feature in the quantity CS and thus in DC will be
smeared out due to the finite thermal relaxation of the system. This occurs on the
time scale ss = Cs/ks. Within this time, the sample will be heated approximately
2.7 Differential Scanning Calorimetry 59
Fig. 2.10 Schematic of DTA configuration. The sample (heat capacity CS, temperature TS) and
the reference sample (heat capacity CR, temperature TR) are thermally connected to the sample
holder (temperature TTb) through the heat links RSH, RRH. The effect of an unwanted link RSR is
discussed in Ref. [89]. RTb is usually neglected
by the quantity sS dTS/dt = sb dTTb/dt = TTb - Ts, which is therefore a measure

for the resulting broadening effect on DC0(T) on the temperature axis. According
to (2.27), DC/CR can be determined within this instrumental uncertainty simply by
simultaneously monitoring the three temperatures TR, TS and TSH, without calcu-
lating any derivative in time or in temperature. The aforementioned limit (TTb -
Ts) in the instrumental temperature resolution for DC0(T) can be reduced by
slowing down the variation TSH(t) of the heat reservoir since TSH - TS = (dTSH/
dt)CS/kSH.
Another remarkable achievement reported in [89] is said to be the substantial
improvement of the conventional differential thermal-analysis (DTA) method by
means of using high-precision electronics and careful temperature control. This
method was used to measure the heat capacity of milligram samples at low tem-
peratures and in magnetic fields up to 7 T with a relative accuracy of about 10-4.
References about DSC are [92–94].
2.8 Other Methods
Besides the methods for measuring the heat capacity described in the previous
sections, other methods have been proposed and used. Among then, it is worth
citing:
(a) Nonadiabatic measurements of the heat capacity involving sample-inherent

thermometry proposed in [95]. The method is realized with a superconducting
quantum interference magnetometry device and applied to FeBr2 single
crystals by using the magnetization for both thermometry and relaxation
calorimetry.
(b) Modulated-bath calorimetry [96] is a variation of the AC calorimetric tech-
nique for measuring the absolute heat capacity of extremely small samples.
The method uses a thermocouple as the weak link to the bath and modulates
the temperature of the bath in time. This eliminates the need for a separate
thermometer and heater on the sample while retaining the ability to make
absolute measurements with minimal addenda.
2.9 Industrial Calorimeters
We also wish to mention the automated heat-capacity measurement system (for

samples weighing 10–500 mg) manufactured by Quantum Design [97] which
employs a thermal-relaxation calorimeter and operates in the temperature range of
1.8–395 K. Examples of measurements carried out by means of this instrument are
reported in [98, 99]. The system also allows one to perform very sensitive electric
and magnetic measurements (e.g., AC susceptibility and DC magnetization). It
employs the thermal relaxation method in the temperature range 1.8–395 K
(optional 0.35–350 K with a continuously operating closed-cycle 3He system). As
an option, it can be equipped for measurements in magnetic field up to 16 T
longitudinal or 7 T transverse. Its calorimeter platform consists of a thin alumina
square of 3 9 3 mm2, backed by a thin-film heater and a bare Cernox (cryogenic
thermometers fabricated from sputtered zirconium oxynitride thin films, com-
mercially available from Lake Shore Cryotronics, Inc. under the trademark Cer-
noxTM). A heat pulse is applied and the platform temperature is recorded. With
known values for the conductance of the thermal link to the bath, of the heat
capacity of the addendum, and of the applied heat, the heat capacity of the sample
and the internal time constant of the calorimeter are determined analytically from
the T(t) data by numerically integrating the relevant differential equations. The
curve fitting is improved by carrying out a number of decay sweeps at each
temperature and averaging the results. Two Cernox thermometers are used over
the full temperature range and their calibration is based on the lTS-90 temperature
scale [100]. The resolution of the system is 10 nJK-l at 2 K. According to an
examination of the system, the accuracy is 1 % at 10–300 K, which decreases to
about *5 % at T \ 5 K. The system is quite adequate to describe broad second-
order phase transitions; however, sharp first-order transitions cannot be investi-
gated properly, mainly because the applied software cannot describe nonexpo-
nential decay curves. This drawback can be removed by using an alternate analytic
approach [97]. The system has recently been equipped with a simple, fully-
2.9 Industrial Calorimeters 61
automated dilution refrigerator for heat-capacity measurements from 55 mK to

4 K and in fields up to 9 T.
Recommendations for the use of quantum devices are reported in Ref. [101].
2.10 Small Sample Calorimetry
With reference to (2.2), cryogenic calorimeters are not only used to measure heat
capacities of liquids and solids, but also in a variety of other applications like the
detection of weakly interacting massive particles, of x-rays and c-rays, and in
astrophysics, or as bolometers for detection of phonons, particles or electromag-
netic waves, and particularly for detection of infrared radiation. These applications
have emerged from the very high sensitivity of recent microcalorimeters capable
of measuring C in the range of nJ/K (corresponding to the heat capacity of a
monolayer of 4He, see, e.g., [101]) or even less. For a review on these devices and
their applications, see, e.g., [2, 102].
The study of novel materials, many of which can be obtained in only small
amounts, has benefited from small-sample calorimetric measurements. These types
of measurements are usually made using the AC method developed by Handler
and coworkers in 1967 [82] and Sullivan and Seidel in 1968 [87], or by a thermal
relaxation method developed by Bachmann et al. in 1972, [3] as amply discussed
in the preceding sections.
Independently of the adopted technique, the absolute accuracy of any mea-
surement of heat capacity is limited by the fraction of the total C which is not due
to the sample, i.e., the addenda. The last 20 years has seen a continuous decrease
in the mass of the measurable sample, mostly because of decreases in the addenda
contribution (particularly of thermometers).
The silicon bolometer described in Ref. [103] enabled one to measure a 1 mg
sample in 1972 [36]; it consisted of approximately 25 mg of silicon divided into
two sections. Both sections had phosphorous diffused into the surface and then
etched to give a concentration versus depth profile such that one section had a
resistance of several kilo-ohms at 4.2 K and the other a resistance of several tens
of kilo-ohms at 4.2 K. Thus, the combination of two thermometers enabled a wider
temperature range; the thermometer which was not in use at a given temperature
was used as a heater. The next improvement in thermometer design was the use of
a small chip of doped germanium attached to a thin sapphire disk which had
sufficient area and strength to serve as a platform.
In the thermal relaxation method, [42, 43, 97, 104] the sensitivity is determined
by the quality of the thermometer and by the (as small as possible) heat capacity of
the addenda. This method can have rather high absolute accuracy; however, its
relative accuracy is limited. On the contrary, the AC method can detect very small
changes in the heat capacity [104–108]. As we saw in Sect. 2.6, the heating power
waveform is usually sinusoidal and the resulting temperature oscillation at fre-
quency x is determined. It is
P
DT ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð2:28Þ
xC l þ x21s2 þ x21s2
1 2
where s2 is the relaxation time within the calorimeter assembly and s1 is the
relaxation time of the calorimeter to the bath. In the usual limit xs1 1 xs2,
the heat capacity can be obtained as C = P/DT. To check whether this limit has
been achieved, the temperature response has to be measured at various frequencies
(typically between 10 and 200 Hz) to determine below which frequency heat leaks
through the thermal link to the bath (within the measuring period) and above
which frequency the calorimeter can no follow the heat modulation. In other
words, the pass-band of the calorimeter is measured.
A relaxation calorimeter for use in a top-loading 3He–4He dilution refrigerator
in high magnetic field has been described in Ref. [34]. It has been used for
milligram samples from 34 mK to 3 K and in magnetic fields up to 18 T. In order
to keep thermal time constants in the magnetic field reasonably short, most of the
addenda, like the thermal reservoir, were made from Ag, which has a very low
nuclear heat capacity.
Very sensitive microcalorimeters are described in [107–109]. The AC calo-
rimeter of Ref. [86], for the Kelvin temperature range consists of a 2–10-lm-thick
monocrystalline silicon membrane substrate produced by etching, taking advan-
tage of the high thermal conductivity and low heat capacity at low temperatures of
this material. A 150 nm CoNi heater (with temperature independent resistivity
between l and 20 K) and a 150 nm NbN thermometer are deposited onto the
substrate. The addenda of this calorimeter varied between less than 0.1 nJ/K at
T \ l K and some nJ/K at 4 K. The device was used to measure heat capacities of
systems of deposited thin films or multilayers of microgram single crystals
and eventually of mesoscopic superconducting loops. The achieved resolution of
DC/C \ 5 9 10-5 allowed measurements of variations of C as small fJ/K [108].
A further improvement in the thermometer design was the silicon on sapphire
(SOS) thermometer technique [110] in which phosphorous was ion-implanted into
a 0.6 micron-thick Si layer on a 0.005-in.-thick sapphire substrate. Ion implanta-
tion allowed the concentration versus depth profile to be as desired, eliminating the
inexact etching step. A third section had been added to the original bolometer
design to serve as a heater. The SOS design is able to measure sample as light as
0.1 mg, to be compared to the 1 mg limit typical of the old silicon bolometer
design.
Another innovation in sample platform thermometers that was quite similar to
the SOS design is the use of a flash-evaporated thin Au–Ge layer (with low heat
capacity) as a thermometer which adheres to the platform without the need for glue
required in the Ge chip design. The aforementioned designs use wires which, even
if the diameter is as small as 25 micron, make a significant addendum contribution
[3]; all use a relatively massive sample platform, at least 10 mg, which, even if it is
sapphire or diamond (with large Debye temperature, see Table 1.1), contributes a
relatively large addendum.
2.10 Small Sample Calorimetry 63
For the development of small sample calorimetry, see, e.g., Ref. [15, 106, 111],
where a combination of AC and heat-pulse calorimetry is used to measure the
specific heat of the ceramic superconductor YBa2Cu3O7-d near the transition
temperature Tc = 90 K.
As seen in Sect. 2.2, small sample calorimetry has now attained the nanogram
range [112, 113].
References
1. Ventura, G., Lanzi, L., Peroni, I., Peruzzi, A., Ponti, G.: Low temperature thermal
characteristics of thin-film Ni–Cr surface mount resistors. Cryogenics 38(4), 453–454
(1998)
2. Ventura, G., Risegari, L.: The art of cryogenics: low-temperature experimental techniques.
Elsevier, Amsterdam (2007)
3. Bachmann, R., DiSalvao, F.J., Geballe, T.H., Greene, R.L., Howard, R.E., King, C.N.,
Kirsch, H.C., Lee, K.N., Schwall, R.E., Thomas, H.U., Zubeck, R.B.: Heat capacity
measurements on small samples at low temperatures. Rev. Sci. Instr. 51, 205 (1972)
4. Martin, D.L.: Use of pure copper as a standard substance for low temperature calorimetry.
Rev. Sci. Instrum. 38(12), 1738–1740 (1967)
5. Martin, D.L.: Specific heats below 3 K of pure copper, silver, and gold, and of extremely
dilute gold-transition-metal alloys. Phys. Rev. 170(3), 650–655 (1968)
6. Martin, D.L.: Tray type calorimeter for the 15–300 K temperature range: copper as a
specific heat standard in this range. Rev. Sci. Instrum. 58(4), 639–646 (1987)
7. Cetas, T.C., Tilford, C.R., Swenson, C.A.: Specific heats of Cu, GaAs, GaSb, InAs, and
InSb from 1 to 30 K. Phys. Rev. 174(3), 835–844 (1968)
8. Ahlers, G.: Heat capacity of copper. Rev. Sci. Instrum. 37(4), 477–480 (1966)
9. Holste, J.C., Cetas, T.C., Swenson, C.A.: Effects of temperature scale differences on the
analysis of heat capacity data: the specific heat of copper from 1 to 30 K. Rev. Sci. Instrum.
43(4), 670–676 (1972)
10. Black, J., Robinson, J., Chemist, P., Britain, G.: Lectures on the Elements of Chemistry,
Delivered in the University of Edinburgh. Mundell and Son for Longman and Rees, London,
and William Creech, Edinburgh (1803)
11. Hansen, L.D.: Toward a standard nomenclature for calorimetry. Thermochim. Acta 371(1),
19–22 (2001)
12. Zielenkiewicz, W.: Towards classification of calorimeters. J. Therm. Anal. Calorim. 91(2),
663–671 (2008)
13. Kubaschewski, O., Alcock, C., Spencer, P.: Materials Thermochemistry, vol. 6. Pergamon
Press, Oxford (1993)
14. Nernst, W.: The energy content of solids. Ann Physik 36, 395–439 (1911)
15. Schnelle, W., Gmelin, E.: Critical review of small sample calorimetry: improvement by
auto-adaptive thermal shield control. Thermochim. Acta 391(1), 41–49 (2002)
16. Cardwell, D.A., Ginley, D.S.: Handbook of superconducting materials, vol. 1. CRC Press,
Boca Raton (2003)
17. Kishi, A., Kato, R., Azumi, T., Okamoto, H., Maesono, A., Ishikawa, M., Hatta, I.,
Ikushima, A.: Measurement of specific heat anomaly and characterization of high TC
ceramic superconductors by AC calorimetry. Thermochim. Acta 133, 39–42 (1988)
18. Jin, X.C., Hor, P.H., Wu, M.K., Chu, C.W.: Modified high-pressure ac calorimetric
technique. Rev. Sci. Instrum. 55(6), 993–995 (1984)
19. Machado, F.L.A., Clark, W.G.: Ripple method: an application of the square-wave excitation
method for heat-capacity measurements. Rev. Sci. Instrum. 59(7), 1176–1181 (1988)
20. Kraftmakher, Y.A., Cherepanov, V.Y.: Compensation of heat losses in modulation

measurements of specific heat. Teplofiz. Vys. Temp. 16, 647–649 (1978)
21. Euken, A.: The determination of specific heats at low temperatures. Physik. Z 10, 586–589
(1909)
22. Nernst, W.: Sitzungsbericht der K. Preuss. Akad. Wiss 12, 261 (1910)
23. Bagatskii, M.I., Minchina, I.Y., Manzhelii, V.G.: Specific heat of solid para-hydrogen.
Soviet J. Low Temp. Phys. 10, 542 (1984)
24. Ward, L.G., Saleh, A.M., Haase, D.G.: Specific heat of solid nitrogen-argon mixtures: 50 to
100 mol% N2. Phys. Rev. B 27(3), 1832–1838 (1983)
25. Alkhafaji, M.T., Migone, A.D.: Heat-capacity study of butane on graphite. Phys. Rev. B
53(16), 11152–11158 (1996)
26. Sellers, G.J., Anderson, A.C.: Calorimetry below 1 K: the specific heat of copper. Rev. Sci.
Instrum. 45(10), 1256–1259 (1974)
27. Filler, R.L., Lindenfeld, P., Deutscher, G.: Specific heat and thermal conductivity
measurements on thin films with a pulse method. Rev. Sci. Instrum. 46(4), 439–442 (1975)
28. Harrison, J.P.: Cryostat for the measurement of thermal conductivity and specific heat
between 0.05 and 2 K. Rev. Sci. Instrum. 39(2), 145–152 (1968)
29. Morin, F.J., Maita, J.P.: Specific heats of transition metal superconductors. Phys. Rev.
129(3), 1115–1120 (1963)
30. Al-Shibani, K.M., Sacli, O.A.: Low temperature specific heats of AgSb alloys. Phys. Status
Solidi B 163(1), 99–105 (1991). doi:10.1002/pssb.2221630108
31. Osborne, D.W., Flotow, H.E., Schreiner, F.: Calibration and use of germanium resistance
thermometers for precise heat capacity measurements from 1 to 25 k. high purity copper for
interlaboratory heat capacity comparisons. Rev. Sci. Instrum. 38(2), 159–168 (1967)
32. Hiroo, O., Toshiaki, E., Nobuhiko, W.: Heat capacity anomaly of Ag ultrafine particles at
low temperatures. J. Phys. Soc. Jpn. 59, 1695 (1990)
33. Albert, K., Löhneysen, H., Sander, W., Schink, H.: A calorimeter for small samples in the
temperature range from 0.06 K to 3 K. Cryogenics 22(8), 417–420 (1982)
34. Tsujii, H., Andraka, B., Muttalib, K., Takano, Y.: Distributed s2 effect in relaxation
calorimetry. Phys. B 329, 1552–1553 (2003)
35. Gutsmiedl, P., Probst, C., Andres, K.: Low temperature calorimetry using an optical heating
method. Cryogenics 31(1), 54–57 (1991)
36. Greene, R.L., King, C.N., Zubeck, R.B., Hauser, J.J.: Specific heat of granular aluminium
films. Phys. Rev. B 6(9), 3297–3305 (1972)
37. Lipa, J., Swanson, D., Nissen, J., Chui, T.: Lambda point experiment in microgravity.
Cryogenics 34(5), 341–347 (1994)
38. Lipa, J., Nissen, J., Stricker, D., Swanson, D., Chui, T.: Specific heat of liquid helium in
zero gravity very near the lambda point. Phys. Rev. B 68(17), 174518 (2003)
39. Chui, T., Day, P., Hahn, I., Nash, A., Swanson, D., Nissen, J., Williamson, P., Lipa, J.: High
resolution thermometers for ground and space applications. Cryogenics 34, 417–420 (1994)
40. Shepherd, J.P.: Analysis of the lumped s2 effect in relaxation calorimetry. Rev. Sci. Instrum.
56(2), 273–277 (1985)
41. Riegel, S., Weber, G.: A dual-slope method for specific heat measurements. J. Phys. E: Sci.
Instrum. 19(10), 790 (1986)
42. Zink, B.L., Revaz, B., Sappey, R., Hellman, F.: Thin film microcalorimeter for heat capacity
measurements in high magnetic fields. Rev. Sci. Instrum. 73(4), 1841–1844 (2002)
43. Denlinger, D.W., Abarra, E.N., Allen, K., Rooney, P.W., Messer, M.T., Watson, S.K.,
Hellman, F.: Thin film microcalorimeter for heat capacity measurements from 1.5 to 800 K.
Rev. Sci. Instrum. 65(4), 946–959 (1994)
44. Olivieri, E., Barucci, M., Beeman, J., Risegari, L., Ventura, G.: Excess heat capacity in
NTD ge thermistors. J. Low Temp. Phys. 143(3–4), 153–162 (2006)
45. Pagliuso, P.G., Thompson, J.D., Hundley, M.F., Sarrao, J.L., Fisk, Z.: Crystal structure and
low-temperature magnetic properties of R_{m}MIn_{3 m + 2} compounds (M = Rh or Ir;
m = 1,2; R = Sm or Gd). Phys. Rev. B 63(5), 054426 (2001)
References 65
46. Catarino, I., Bonfait, G.: A simple calorimeter for fast adiabatic heat capacity measurements
from 15 to 300 K based on closed cycle cryocooler. Cryogenics 40(7), 425–430 (2000)
47. Forgan, E.M., Nedjat, S.: Heat capacity cryostat and novel methods of analysis for small
specimens in the 1.5–10 K range. Rev. Sci. Instrum. 51(4), 411–417 (1980)
49. Willekers, R., Meijer, H., Mathu, F., Postma, H.: Calorimetry by means of the relaxation
and dual-slope methods below 1 K: application to some high Tc superconductors.
Cryogenics 31(3), 168–173 (1991)
50. Drulis, M.: Low temperature heat capacity measurements of U6FeH15 hydride. J. Alloys
Compd. 219(1), 41–44 (1995)
51. Barucci, M., Brofferio, C., Giuliani, A., Gottardi, E., Peroni, I., Ventura, G.: Measurement
of low temperature specific heat of crystalline TeO2 for the optimization of bolometric
detectors. J. Low Temp. Phys. 123(5–6), 303–314 (2001). doi:10.1023/a:1017555615150
52. Kim, J.S., Stewart, G.R., Bauer, E.D., Ronning, F.: Unusual temperature dependence in the
low-temperature specific heat of U3Ni3Al19. Phys. Rev. B 78(15), 153108 (2008)
53. Cinti, F., Affronte, M., Lascialfari, A., Barucci, M., Olivieri, E., Pasca, E., Rettori, A.,
Risegari, L., Ventura, G., Pini, M.G., Cuccoli, A., Roscilde, T., Caneschi, A., Gatteschi, D.,
Rovai, D.: Chiral and helical phase transitions in quasi-1d molecular magnets. Polyhedron
24(16–17), 2568–2572 (2005)
54. Nakajima, Y., Li, G., Tamegai, T.: Specific heat study of ternary iron-silicide
superconductor Lu2Fe3Si5: evidence for two-gap superconductivity. Physica C 468(15),
1138–1140 (2008)
55. Kasahara, S., Fujii, H., Mochiku, T., Takeya, H., Hirata, K.: Specific heat of novel ternary
superconductors La3Ni4X4 (X = Si and Ge). Physica C 468(15), 1231–1233 (2008)
56. Kasahara, S., Fujii, H., Mochiku, T., Takeya, H., Hirata, K.: Low temperature specific heat
of ternary germanide superconductor La3Pd4Ge4. Phys. B 403(5), 1119–1121 (2008)
57. Kasahara, S., Fujii, H., Takeya, H., Mochiku, T., Thakur, A., Hirata, K.: Low temperature
specific heat of superconducting ternary intermetallics La3Pd4Ge4, La3Ni4Si4, and
La3Ni4Ge4 with U3Ni4Si4-type structure. J. Phys.: Condens. Matter 20(38), 385204 (2008)
58. Fanelli, V., Christianson, A.D., Jaime, M., Thompson, J., Suzuki, H., Lawrence, J.:
Magnetic order in the induced magnetic moment system Pr3In. Phys. B 403(5), 1368–1370
(2008)
59. Suzuki, H., Inaba, A., Meingast, C.: Accurate heat capacity data at phase transitions from
relaxation calorimetry. Cryogenics 50(10), 693–699 (2010)
60. Haller, E.: Advanced far-infrared detectors. Infrared Phys. Technol. 35(2), 127–146 (1994)
61. Lounasmaa, O.V. (ed.): Experimental principles and methods below 1 K. Academic Press,
London (1974)
62. Keesom, P., Seidel, G.: Specific heat of germanium and silicon at low temperatures. Phys.
Rev. 113(1), 33 (1959)
63. Wang, N., Wellstood, F.C., Sadoulet, B., Haller, E.E., Beeman, J.: Electrical and thermal
properties of neutron-transmutation-doped Ge at 20 mK. Phys. Rev. B 41(6), 3761–3768
(1990)
64. Aubourg, É., Cummings, A., Shutt, T., Stockwell, W., Barnes Jr, P., Silva, A., Emes, J.,
Haller, E., Lange, A., Ross, R., Sadoulet, B., Smith, G., Wang, N., White, S., Young, B.,
Yvon, D.: Measurement of electron-phonon decoupling time in neutron-transmutation
doped germanium at 20 mK. J. Low Temp. Phys. 93(3–4), 289–294 (1993)
65. Alessandrello, A., Brofferio, C., Camin, D.V., Cremonesi, O., Giuliani, A., Pavan, M.,
Pessina, G., Previtali, E.: Signal modelling for TeO2 bolometric detectors. J. Low Temp.
Phys. 93(3–4), 207–212 (1993)
66. Stefanyi, P., Zammit, C., Rentzsch, R., Fozooni, P., Saunders, J., Lea, M.: Development of a
Si bolometer for dark matter detection. Phys. B 194, 161–162 (1994)
67. Efros, A., Shklovskii, B.: Electronic Properties of Doped Semiconductors. Springer Series
in Solid-State Sciences. Springer, Berlin (1984)
68. Marcenat, C.: Etudes calorimetrique sous champ magnetique des phases basses temperature
des composes Kondo (1986)
69. Xu, Jc, Watson, C.H., Goodrich, R.G.: A method for measuring the specific heat of small
samples. Rev. Sci. Instrum. 61(2), 814–821 (1990)
70. Flachbart, K., Gabáni, S., Gloos, K., Meissner, M., Opel, M., Paderno, Y., Pavlík, V.,
Samuely, P., Schuberth, E., Shitsevalova, N., Siemensmeyer, K., Szabó, P.: Low
temperature properties and superconductivity of LuB12. J. Low Temp. Phys. 140(5–6),
339–353 (2005)
71. Pilla, S., Hamida, J., Sullivan, N.: A modified dual-slope method for heat capacity
measurements of condensable gases. Rev. Sci. Instrum. 71(10), 3841–3845 (2000)
72. Gmelin, E.: Classical temperature-modulated calorimetry: a review. Thermoch. Acta 305,
1–26 (1997)
73. Castro, M., Puértolas, J.: Simple and accurate ac calorimeter for liquid crystals and solid
samples. J. Therm. Anal. 41(6), 1245–1252 (1994)
74. Corbino, O.M.: Specific heat. Phys. Z. 12, 292 (1911)
75. Rosenthal, L.A.: Thermal response of bridgewires used in electro explosive devices. Rev.
Sci. Instrum. 32(9), 1033–1036 (1961)
76. Filippov, L.: Procedure of measuring liquid thermal activity. Inzh.-Fiz. Zh. 3(7), 121–123
(1960)
77. Birge, N.O., Nagel, S.R.: Wide-frequency specific heat spectrometer. Rev. Sci. Instrum.
58(8), 1464–1470 (1987)
78. Jeong, Y.H., Bae, D.J., Kwon, T.W., Moon, I.K.: Dynamic specific heat near the Curie point
of Gd. J. Appl. Phys. 70(10), 6166–6168 (1991)
79. Moon, I.K., Jeong, Y.H., Kwun, S.I.: The 3x technique for measuring dynamic specific heat
and thermal conductivity of a liquid or solid. Rev. Sci. Instrum. 67(1), 29–35 (1996)
80. Jewett, D.M.: Electrical heating with polyimide-insulated magnet wire. Rev. Sci. Instrum.
58(10), 1964–1967 (1987)
81. Cahill, D.G.: Thermal conductivity measurement from 30 to 750 K: the 3x method. Rev.
Sci. Instrum. 61(2), 802–808 (1990)
82. Handler, P., Mapother, D.E., Rayl, M.: AC measurement of the heat capacity of nickel near
its critical point. Phys. Rev. Lett. 19(7), 356–358 (1967)
83. Hatta, I.: History repeats itself: progress in ac calorimetry. Thermochim. Acta 300(1), 7–13
(1997)
84. Pradhan, N., Duan, H., Liang, J., Iannacchione, G.: Specific heat and thermal conductivity
measurements for anisotropic and random macroscopic composites of cobalt nanowires.
Nanotechnology 19(48), 485712 (2008)
85. Hashimoto, M., Tomioka, F., Umehara, I., Fujiwara, T., Hedo, M., Uwatoko, Y.: Heat
capacity measurement of CePd2Si2 under high pressure. Phys. B 378, 815–816 (2006)
86. Hemminger, W., Höhne, G.: Grundlagen der Kalorimetrie. Verlag Chemie, Weinheim
(1979)
87. Sullivan, P.F., Seidel, G.: Steady-state, ac-temperature calorimetry. Phys. Rev. 173(3), 679
(1968)
88. Maglic, K., Cezairliyan, A., Peletsky, V.: Compendium of Thermophysical Property
Measurement Methods: Vol. 1, Survey of Measurement Techniques. Plenum Press, New
York (1984)
89. Schilling, A., Jeandupeux, O.: High-accuracy differential thermal analysis: a tool for
calorimetric investigations on small high-temperature-superconductor specimens. Phys.
Rev. B 52(13), 9714–9723 (1995)
90. Budaguan, B., Aivazov, A., Meytin, M., Sazonov, A.Y., Metselaar, J.: Relaxation processes
and metastability in amorphous hydrogenated silicon investigated with differential scanning
calorimetry. Phys. B 252(3), 198–206 (1998)
91. Sturtevant, J.M.: Biochemical applications of differential scanning calorimetry. Annu. Rev.
Phys. Chem. 38(1), 463–488 (1987)
References 67
92. Rahm, U., Gmelin, E.: Low temperature micro-calorimetry by differential scanning.
J. Therm. Anal. 38(3), 335–344 (1992)
93. Junod, A.: An automated calorimeter for the temperature range 80–320 K without the use of
a computer. J. Phys. E: Sci. Instrum. 12(10), 945 (1979)
94. Junod, A., Bonjour, E., Calemczuk, R., Henry, J., Muller, J., Triscone, G., Vallier, J.:
Specific heat of an YBa2Cu3O7 single crystal in fields up to 20 T. Physica C 211(3),
304–318 (1993)
95. Kharkovski, A., Binek, C., Kleemann, W.: Nonadiabatic heat-capacity measurements using
a superconducting quantum interference device magnetometer. Appl. Phys. Lett. 77(15),
2409–2411 (2000)
96. Graebner, J.: Modulated-bath calorimetry. Rev. Sci. Instrum. 60(6), 1123–1128 (1989)
97. Lashley, J., Hundley, M., Migliori, A., Sarrao, J., Pagliuso, P., Darling, T., Jaime, M.,
Cooley, J., Hults, W., Morales, L.: Critical examination of heat capacity measurements
made on a quantum design physical property measurement system. Cryogenics 43(6),
369–378 (2003)
98. Newsome Jr, R., Park, S., Cheong, S.-W., Andrei, E.: Low-temperature measurements of the
specific heat capacity of a thick ferroelectric copolymer film of vinylidene fluoride and
trifluoroethylene. Phys. Rev. B 77(9), 094103 (2008)
99. Javorský, P., Wastin, F., Colineau, E., Rebizant, J., Boulet, P., Stewart, G.: Low-
temperature heat capacity measurements on encapsulated transuranium samples. J. Nucl.
Mater. 344(1), 50–55 (2005)
100. Preston-Thomas, H.: The international temperature scale of 1990(ITS-90). Metrologia
27(1), 3–10 (1990)
101. Kennedy, C.A., Stancescu, M., Marriott, R.A., White, M.A.: Recommendations for accurate
heat capacity measurements using a quantum design physical property measurement system.
Cryogenics 47(2), 107–112 (2007)
102. Giazotto, F., Heikkilä, T.T., Luukanen, A., Savin, A.M., Pekola, J.P.: Opportunities for
mesoscopics in thermometry and refrigeration: physics and applications. Rev. Mod. Phys.
78(1), 217 (2006)
103. Bachmann, R., Kirsch, H.C., Geballe, T.H.: Low temperature silicon thermometer and
bolometer. Rev. Sci. Instrum. 41(4), 547–549 (1970)
104. Doettinger-Zech, S., Uhl, M., Sisson, D., Kapitulnik, A.: Simple microcalorimeter for
measuring microgram samples at low temperatures. Rev. Sci. Instrum. 72(5), 2398–2406
(2001)
105. Schwall, R., Howard, R., Stewart, G.: Automated small sample calorimeter. Rev. Sci.
Instrum. 46(8), 1054–1059 (1975)
106. Stewart, G.R.: Measurement of low-temperature specific heat. Rev. Sci. Instrum. 54(1),
1–11 (1983)
107. Bourgeois, O., Skipetrov, S., Ong, F., Chaussy, J.: Attojoule calorimetry of mesoscopic
superconducting loops. Phys. Rev. Lett. 94(5), 057007 (2005)
108. Riou, O., Gandit, P., Charalambous, M., Chaussy, J.: A very sensitive microcalorimetry
technique for measuring specific heat of lg single crystals. Rev. Sci. Instrum. 68(3),
1501–1509 (1997)
109. Fominaya, F., Fournier, T., Gandit, P., Chaussy, J.: Nanocalorimeter for high resolution
measurements of low temperature heat capacities of thin films and single crystals. Rev. Sci.
Instrum. 68(11), 4191–4195 (1997)
110. Early, S., Hellman, F., Marshall, J., Geballe, T.: A silicon on sapphire thermometer for
small sample low temperature calorimetry. Physica B + C 107(1), 327–328 (1981)
111. Wilhelm, H., Lühmann, T., Rus, T., Steglich, F.: A compensated heat-pulse calorimeter for
low temperatures. Rev. Sci. Instrum. 75(8), 2700–2705 (2004)
112. Tagliati, S., Rydh, A.: Absolute accuracy in membrane-based ac nanocalorimetry.
Thermochim. Acta 522(1), 66–71 (2011)
113. Tagliati, S., Rydh, A., Xie, R., Welp, U., Kwok, W.: Membrane-based calorimetry for
studies of sub-microgram samples. J. Phys.: Conf. Ser. 052256 (2009) (IOP Publishing)
Chapter 3
Data of Specific Heat
Abstract In this chapter experimental data for materials of common use in

cryogenics are reported, in particular, approximate integrated data useful for a
rough estimation of the power necessary to cool down a system. Specific heat data
are divided into two groups of tables: (1) very-low temperature data of cp (below
about 4 K) given by a fit when possible (Sect. 3.2). (2) low temperature of cp
(approximately 4–300 K) of specific heat cp and H*(integral of cp) (Sect. 3.3).
3.1 Presentation of DATA of Specific Heat
In the following sections we wish to present experimental data for materials of

common use in cryogenics. In particular, approximate integrated data are useful
for a rough estimation of the power necessary to cool down a system.
We have split specific heat data into two groups:
(1) very-low temperature data of cp (below about 4 K) given by a fit when
possible (Sect. 3.2).
(2) low temperature (approximately 4–300 K) of cp and H*(integral of cp),
presented in the form of table (Sect. 3.3).
3.2 Very-Low Temperature DATA (Below About 4 K)
3.2.1 Metals and Alloys
(See Table 3.1).

70 3 Data of Specific Heat
Table 3.1 Very-low temperature data of metals

Material Reference Temperature Temperature Fit (lJ/gK)
range of data range of fit
Ag [1] 0.4–4.2 K 5.9 T + 1.6 T3
AgSb (0.89 %Sb) [2] 0.4–4.2 K 6 T + 1.6 T3
AgSb (2.67 %Sb) [2] 0.4–4.2 K 6.2 T + 1.7 T3
AgSb (3.74 %Sb) [2] 0.4–4.2 K 6.3 T + 1.7 T3
AgSb (4.72 %Sb) [2] 0.4–4.2 K 6.6 T + 1.7 T3
Al 5056 (wt%:95.0 % [3] 0.1–3.6 K T \ 0.8 K 43.8 T + 5.2 T3
Al, 0.12 % Mn,
5.0 % Mg, 0.12 %
Cr)
Cu [4] T\4 K 0–4.0 K 10.88 T + 7.310-7T3 + 1.32
9 103 T5
Cu [5] 1–30 K 1.6–4.2 K 693 T + 47.88 T3
Be–Cu [6] 0.5–20 mK 0.5–5 mK 3.58 9 10-3 T-1.64
GaSb [5] 1–30 K 1.4–4.2 K 5.77 T + 98.63 T3 + 0.720
T5
GaAs [5] 1–30 K 1.0–4.2 K 2.48 T + 46.73 T3 + 0.143
T5
Gold [4] T\4 K 0–3.0 K 3.5 T + 2.3 9 10–6- 4.89
9 10-3T5
Indium [7] 2.7–4.2 K 15.5 T3
Indium (normal state) [8] 0.08–4.2 K T \ 0.8 K 8.8 9 10-3 T-2+ 14.7
T + 12.4 T3
Indium [8] 0.08–4.2 K T \ 0.350 K 10.6 T3
(superconducting
state)
InSb [5] 1–30 K 0.93–3.0 K 7.54T + 220.36T3 + 3.045T5
InSb [5] 1–30 K 0.93–4.2 K 7.54 + 220.17T3 + 3.285T5
– 0.0753T7 + 0.00614T9
InAs [5] 1–30 K 0.95–4.2 K 4.8T + 123.02 T3 + 0.987 T5
Invar (Fe-64 %, [9] 2–20 K 1–4.5 K 11348 T + 41 T3
Ni-36 %)
Tin (normal state) [8] 0.1–1.1 K T \ 0.45 K 14.99 T + 2.072 T3
Tin (superconducting [8] 0.1–1.1 K T \ 0.45 K 2.1 T3
state)
Constantan [10] 0.15–4.2 K 0.15–0.3 K 205 T + 2.8 T-2
Manganin [10] 0.2–4 K 0.2–2.5 K 59.5 T + 2.94 T3 + 11.5 T-2
Pt91–W9 heater wire [11] 0.07–1.2 K 0.07–1.2 K 17.6 T + 1.4 T3 + 0.12 T-2
Stainless steel 304 [12] 0.07–0.6 K 0.07–0.6 K 465 T + 0.56 T-2
Stainless steel 304 [13] T \ 10 K 460 T + 0.38 T3
3.2.2 Dielectrics
(See Table 3.2).

3.2 Very-Low Temperature Data (Below About 4 K) 71
Table 3.2 Very-low temperature data of dielectrics materials

Material Reference Temperature Temperature Fit (lJ/gK)
range of data range of fit
Apiezon N grease [14] 0.1–2.5 K 0.32 T + 25.8 T3
+ 0.0044 T-2
Apiezon N grease [15] 1 K\T\4 K 29 T3
GE 7031 varnish [16] T\1 K 6.5 T + 19 T3
GE 7031 varnish [17] 2 K\T\4 K 35T3
Stycast 1266 epoxy [18] 0.1–1 K 2.91
T + 15.7 T3 + 8.98 T5
Polypropylene [19] 0.06–1 K 6.15 T1.33 + 20 T3
PMMA [20] 0.07–0.2 K 3.0 T + 77 T3
PMMA [16] T\1 K 4.6 T + 29 T3
Polystyrene [20] 0.07–0.2 K 4.6 T + 93 T3
Silica [16] T\1 K 1.1 T + 2.0 T3
Silica [21] 2 K\T\4 K 2.2 T3
3SiO2ffiNa2O [16] T\1 K 2.0 T + 3.4 T3
3SiO2ffiNa2O [21] 2 K\T\4 K 2.3 T3
Araldite (epoxy) [22] 1 K\T\4 K 18 T3
Nylon [22] 1 K\T\4 K 20 T3
Zerodur [23] 1.5 K \ T \ 5.4 T + 0.8 T3
5.5 K + 0.03 T5
Torlon 4203 [24] 0.15–4.2 K T\1 K 5.41 + T1.28
+ 28.2 T3
Torlon 4203 [24] 0.15–4.2 K 1 K \ T \ 4.2 K 2.68 T3.32
TiO2 [25] 0.3–20 K T\1 K 1.55 T3
Silicate glass (60 % Mol [26] 0.1–10 K 0.1–10 K 4.5 T + 0.8 T3
SiO2, 27.5 % LiO2, 10 % + 0.004 T5
CaO, 2.5 % Al2O3)
Rare-earth-doped silicate [26] 0.1–10 K 0.1–10 K 9.6 T1.29 + 1.1 T3
glass (59.9 %Mol SiO2, + 0.001 T5
27.5 % LiO2, 10 % CaO,
2.5 % Al2O3, 0.1 % Pr2O3)
Rare-earth-doped silicate [26] 0.1–10 K 0.1–10 K 5.6 T + 0.7 T3
glass (57 %Mol SiO2, + 0.003 T5
27.5 % LiO2, 10 % CaO,
2.5 % Al2O3, 3 % Eu2O3)
Rare-earth-doped silicate [26] 0.1–10 K 0.1–10 K 6 T1.05 + 1.2 T3
glass (68.3 %Mol SiO2, + 0.005 T5
15 % LiO2, 11 % CaO,
4.7 % Al2O3, 1 % Pr2O3)
Rare-earth-doped silicate [26] 0.1–10 K 0.1–10 K 3.5 T + 1.2 T3
glass (74.75 %Mol SiO2, + 0.03 T5
15 % Na2O, 5 % BaO, 5 %
ZnO, 0.25 % Eu2O3)
3.3 Low-Temperature Specific Heat DATA

(Approximately 4–300 K)
Data of specific heat of technical solids below 4 K are reported in Tables 3.3 and
3.5, while values of integrated cp:
ZT

H ðTÞ ¼ cp ðTÞdT ð3:1Þ
0
are reported in Tables 3.4 and 3.6.

The integral in (3.1) can be easily solved in the case of linear dependence of cp
in temperature e.g. for metals and alloys. In the other cases the cp(T) values from
literature were integrated over small DT interval:
ZT1 ZT2 ZTn

H ðTn Þ ¼ cp ðTÞdT þ cp ðTÞdT þ . . . cp ðTÞdT ð3:2Þ
0 T1 Tn1
approximated as:
ðcp ðTn Þ þ cp ðTn1 ÞÞ

H ðTn Þ H ðTn1 Þ þ ðTn Tn1 Þ ð3:3Þ
2
that we used for the calculation of the values reported in Table 3.3 for metals and
alloys. In the case of insulators, the main contribution to the specific heat is
proportional to T3. Unfortunately, a pure cubic function does not fit well the
experimental data; thus we fitted data with the polynomial:
cðTÞ ¼ AT þ BT 2 þ CT 3 þ DT 4 ð3:4Þ
Integration of (3.4) gives the values of H* reported in Table 3.6.
3.3.1 Metals and Alloys
(See Tables 3.3and 3.4).
3.3.2 Dielectrics
(See Tables 3.5 and 3.6).

Table 3.3 Specific heat cp (J g-1 K-1) of metals and alloys
Material References 4K 25 K 50 K 75 K 100 K 150 K 200 K 293 K
Metals
Ag [27] 1.2 9 10-4 2.58 9 10-2 1.1 9 10-1 1.6 9 10-1 1.9 9 10-1 2.1 9 10-1 2.2 9 10-1 2.3 9 10-1
Al [28] 3 9 10-4 1.7 9 10-2 1.4 9 10-1 3.2 9 10-1 4.8 9 10-1 6.8 9 10-1 8 9 10-1 9 9 10-1
Cu [28] – 1.5 9 10-2 9.7 9 10-2 1.87 9 10-1 2.52 9 10-1 3.22 9 10-1 3.55 9 10-1 3.83 9 10-1
OFCH Cu [29] 1 9 10-4 1.52 9 10-2 9.6 9 10-2 1.88 9 10-1 2.53 9 10-1 3.21 9 10-1 3.55 9 10-1 3.84 9 10-1
Fe [28] – 8.0 9 10-3 5.1 9 10-2 1.36 9 10-1 2.16 9 10-1 3.24 9 10-1 3.84 9 10-1 4.44 9 10-1
Nb [28] – 2.0 9 10-2 8.5 9 10-2 1.47 9 10-1 1.88 9 10-1 2.30 9 10-1 2.48 9 10-1 2.62 9 10-1
Ni [28] – 9.8 9 10-3 6.9 9 10-2 1.56 9 10-1 2.32 9 10-1 3.29 9 10-1 3.83 9 10-1 4.35 9 10-1
Si [28] – 8.5 9 10-3 7.8 9 10-2 1.70 9 10-1 2.60 9 10-1 4.25 9 10-1 5.57 9 10-1 6.94 9 10-1
Ti [28] – 1.37 9 10-2 9.8 9 10-2 2.10 9 10-1 3.00 9 10-1 4.08 9 10-1 4.66 9 10-1 5.18 9 10-1
W [28] – 4.1 9 10-3 3.2 9 10-2 6.4 9 10-2 8.7 9 10-2 1.12 9 10-1 1.23 9 10-1 1.33 9 10-1
Alloys
Al 2024 [28] – – – – 4.6 9 10-1 6.5 9 10-1 7.3 9 10-1 8.4 9 10-1
CuZn (65/35) [28] – 2.2 9 10-2 1.18 9 10-1 2.1 9 10-1 2.7 9 10-1 3.3 9 10-1 3.6 9 10-1 3.77 9 10-1
Constantan [28] – 1.3 9 10-2 8 9 10-2 1.7 9 10-1 2.4 9 10-1 3.2 9 10-1 3.6 9 10-1 4.1 9 10-1
Inconel 718 [28] – – 7 9 10-2 1.6 9 10-1 2.7 9 10-1 3.6 9 10-1 4.0 9 10-1 4.3 9 10-1
Nb-38Ti [28] – 3 9 10-2 1.1 9 10-1 2.4 9 10-1 – – – –
SnPb (50/50) [28] – 6.2 9 10-2 1.16 9 10-1 1.40 9 10-1 1.52 9 10-1 1.63 9 10-1 1.70 9 10-1 1.78 9 10-1
S.S. 304/316 [28] – 1.9 9 10-2 9.2 9 10-2 1.9 9 10-1 2.8 9 10-1 3.5 9 10-1 4.2 9 10-1 4.7 9 10-1
Ti-6Al-4 V [28] – – – 2.1 9 10-1 – 4.0 9 10-1 4.9 9 10-1 5.5 9 10-1
3.3 Low-Temperature Specific Heat DATA (Approximately 4–300 K)
6061-T6 Al [29] 3 9 10-4 1.8 9 10-2 1.49 9 10-1 3.34 9 10-1 4.92 9 10-1 7.13 9 10-1 8.35 9 10-1 9.43 9 10-1
304 SS [29] 2 9 10-3 2.1 9 10-2 9.6 9 10-2 1.97 9 10-1 2.75 9 10-1 3.62 9 10-1 4.16 9 10-1 4.71 9 10-1
UCu3.5Fe1.5Al7 [30] 1.6 9 10-3 2.4 9 10-2 9.8 9 10-2 – – – – –
UCu3Cr2Al7 [30] 1.4 9 10-3 2.1 9 10-2 8.9 9 10-2 – – – – –
UCu3Mn2Al7 [30] 1.2 9 10-3 2.1 9 10-2 8.8 9 10-2 – – – – –
Data from ref. [29] are calculated from fits
73
74
Table 3.4 Integrated specific heat H* (J g-1) for some metals

Material H* 4 K H* 10 K H* 20 K H* 30 K H* 50 K H* 77 K H* 100 K H* 150K ) H* H*
200(K) 300(K)
Metals
Al 5 9 10-4 6 9 10-3 6 9 10-2 2.6 9 10-1 2 8.5 18 47 84 169
Cu 2 9 10-4 3 9 10-3 4 9 10-2 2.1 9 10-1 1.5 5.4 10 25 42 79
Fe 8 9 10-4 6 9 10-3 3 9 10-2 1.1 9 10-1 7.8 9 10-1 3.5 8 21 39 80
In 2 9 10-3 5 9 10-2 4.3 9 10-1 1.3 4 8.7 13 24 35 58
Nb 8 9 10-4 9 9 10-3 8 9 10-2 3 9 10-1 1.6 5.2 9 21 33 59
Ni 1 9 10-3 7 9 10-3 4 9 10-2 1.6 9 10-1 1 4.1 9 23 40 82
Si 3 9 10-5 1 9 10-3 2 9 10-2 1.2 9 10-1 1.1 4.5 10 27 51 115
Ti 6 9 10-4 5 9 10-3 5 9 10-2 2 9 10-1 1.4 5.7 12 29 51 101
W 8 9 10-5 9 9 10-4 1 9 10-2 6 9 10-2 3.5 17 35 60 66 79
Alloys
Al 6061 6 9 10-4 6 9 10-3 6 9 10-2 2.7 9 10-1 2.1 8.8 18 49 87 177
Costantan 1 9 10-3 8 9 10-3 5 9 10-2 1.9 9 10-1 1.2 4.7 9 23 41 79
Stainless steel 310 4 9 10-3 3 9 10-2 1.4 9 10-1 2.7 9 10-1 1.4 5.4 11 26 44 88
For refs. see Table 3.2. The starting data are taken from ref. [31]
3 Data of Specific Heat
Table 3.5 Specific heat cp (J g-1 K-1) of insulators
Material References 4K 25 K 50 K 75 K 100 K 150 K 200 K 293 K
Non-metals
Bi2Ti2O7 [32] 3 9 10-4 3.5 9 10-2 9.6 9 10-2 1.48 9 10-1 2.2 9 10-1 – – –
Bi2InNbO7 [32] 3 9 10-4 3 9 10-2 9.3 9 10-2 1.45 9 10-1 1.9 9 10-1 – – –
Bi4Ti3O12 [32] 3 9 10-4 2.7 9 10-2 7.2 9 10-2 1.05 9 10-1 – – – –
GaP [33] – 2.29 9 10-2 8.99 9 10-2 1.57 9 10-1 2.13 9 10-1 – – –
MgO [white] [28] – 1.9 9 10-3 2.07 9 10-2 8.5 9 10-2 1.95 9 10-1 4.49 9 10-2 6.61 9 10-1 9.16 9 10-1
PbTiO3 [32] 2 9 10-4 3 9 10-2 8.5 9 10-2 1.08 9 10-1 – – – –
Pyrex [28] – 4.3 9 10-2 – – 2.8 9 10-1 4.06 9 10-1 5.33 9 10-1 7.2 9 10-1
Sapphire [28] – 1.4 9 10-3 1.48 9 10-2 5.58 9 10-2 1.26 9 10-1 3.13 9 10-1 5.01 9 10-1 7.63 9 10-1
Silica [28] – 3.8 9 10-2 1.11 9 10-1 1.88 9 10-1 2.68 9 10-1 4.20 9 10-1 5.46 9 10-1 7.28 9 10-1
Y2Ti2O7 [32] 1 9 10-4 1.4 9 10-2 5.6 9 10-2 9.1 9 10-2 – – – –
ZnSe [33] – 4.61 9 10-2 1.17 9 10-1 1.81 9 10-1 2.25 9 10-1 – – –
ZnTe [33] – 4.88 9 10-2 1.09 9 10-1 1.6 9 10-1 1.91 9 10-1 – – –
ZrO2 [28] – 9 9 10-3 4.1 9 10-2 9.5 9 10-2 1.5 9 10-1 2.6 9 10-1 3.5 9 10-1 4.5 9 10-1
Polymers
Epoxy [28] – 1.3 9 10-1 2.7 9 10-1 3.9 9 10-1 4.8 9 10-1 – 1.0 1.3
G-10 [29] 2 9 10-3 6.4 9 10-2 1.49 9 10-1 2.32 9 10-1 3.17 9 10-1 4.89 9 10-1 6.64 9 10-1 9.77 9 10-1
G10 (GFRP) [28] – – 3 9 10-1 4 9 10-1 5 9 10-1 – 1.0 1.5
Nylon 6 [28] – – – 4.7 9 10-1 – 8.1 9 10-1 1.01 1.5
Stycast [28] – 3.2 9 10-2 8.8 9 10-2 1.5 9 10-1 2.2 9 10-1 – – –
3.3 Low-Temperature Specific Heat DATA (Approximately 4–300 K)
Teflon [28] – 1.0 9 10-1 2.1 9 10-1 2.9 9 10-1 3.9 9 10-1 5.6 9 10-1 7.2 9 10-1 1.0
Teflon [29] – 1.5 9 10-2 9.6 9 10-1 1.88 9 10-1 2.53 9 10-1 3.21 9 10-1 3.55 9 10-1 3.84 9 10-1
Composites
Glass/resin [34] 2.8 9 10-2 1.20 9 10-1 5.22 9 10-1 9.20 9 10-1 1.14 1.44 1.63 1.89
Boron/aluminum [34] 4 9 10-3 3.2 9 10-2 1.5 9 10-1 2.8 9 10-1 4.0 9 10-1 5.9 9 10-1 7.6 9 10-1 9.5 9 10-1
Boron/epoxy [34] 5 9 10-3 5.1 9 10-2 1.2 9 10-1 2.0 9 10-1 3.0 9 10-1 5.1 9 10-1 7.6 9 10-1 1.22
Data from reference [29] are calculated from fits
75
76
Table 3.6 Integrated specific heat H* (J g-1) for some insulators

H* (4 K) H* (10 K) H* (20 K) H* (30 K) H* (50 K) H* (77 K) H* (100 K) H* (150 K) H* (200 K) H* (300 K)
Apiezon 2.9 9 10-2 2.0 9 10-1 9.0 9 10-1 2.2 7.0 18 32 72 124 351
Cy 221 4.4 9 10-2 2.7 9 10-1 1.1 2.4 6.5 15 25 56 100 219
Diamond 5.2 9 10-4 2.6 9 10-3 7.0 9 10-3 1.1 9 10-2 2.3 9 10-2 1.3 9 10-1 4.4 9 10-1 2.8 9.6 45
G10 1.9 9 10-2 1.2 9 10-1 5 9 10-1 1.2 3.4 8.5 15 35 64 147
Glass 1.0 9 10-2 6.6 9 10-2 2.8 9 10-1 6.5 9 10-1 2 5.4 10 30 68 213
Ice 3.5 9 10-2 2.4 9 10-1 1.1 2.8 9.1 24 42 96 165 447
Kapton 2.8 9 10-2 1.8 9 10-1 7.4 9 10-1 1.7 5 12 21 45 75 141
MgO 4.0 9 10-4 6.4 9 10-3 5.2 9 10-2 1.8 9 10-1 8.3 9 10-1 3.1 6.7 22 51 138
Nylon 4.7 9 10-2 3 9 10-1 1.2 2.9 8.4 21 35 79 134 272
Plexiglass 2.2 9 10-2 1.6 9 10-1 7.2 9 10-1 1.8 5.9 15 27 58 97 310
Quartz 2.7 9 10-3 2.5 9 10-2 2.0 9 10-1 4.4 9 10-1 1.7 5.3 10 28 52 115
Shappire 1.7 9 10-4 2.8 9 10-3 2.4 9 10-2 8.4 9 10-2 4.2 9 10-1 1.7 3.9 14 34 103
SrTiO3 1.3 9 10-2 8.4 9 10-2 3.5 9 10-1 8.4 9 10-1 2.5 6.5 11 27 47 95
Stycast 3.4 9 10-3 3 9 10-2 1.7 9 10-1 4.8 9 10-1 1.7 4.9 9.3 31 103 830
Teflon 3.3 9 10-2 2 9 10-1 8.2 9 10-1 1.8 5.1 12 20 44 75 152
The starting data are taken from ref. [31]
3 Data of Specific Heat
References 77
References
1. Lounasmaa, O.V. (ed.): Experimental Principles and Methods Below 1 K. Academic Press,
London (1974)
2. Al-Shibani, K.M., Sacli, O.A.: Low Temperature Specific Heats of AgSb Alloys. phys. Status
Solidi (b) 163(1), 99–105 (1991). doi:10.1002/pssb.2221630108
3. Barucci, M., Ligi, C., Lolli, L., Marini, A., Martelli, V., Risegari, L., Ventura, G.: Very low
temperature specific heat of Al 5056. Phys. B 405(6), 1452–1454 (2010)
4. Martin, D.L.: Specific heats below 3 K of pure copper, silver, and gold, and of extremely
dilute gold-transition-metal alloys. Phys. Rev. 170(3), 650–655 (1968)
5. Cetas, T.C., Tilford, C.R., Swenson, C.A.: Specific heats of Cu, GaAs, GaSb, InAs, and InSb
from 1 to 30 K. Phys. Rev. 174(3), 835–844 (1968)
6. Karaki, Y., Koike, Y., Kubota, M., Ishimoto, H.: Specific heat of beryllium–copper alloy at
very low temperature. Cryogenics 37(3), 171–172 (1997)
7. Xu, J.C., Watson, C.H., Goodrich, R.G.: A method for measuring the specific heat of small
samples. Rev. Sci. Instrum. 61(2), 814–821 (1990)
8. O’neal, H., Phillips, N.E.: Low-temperature heat capacities of indium and tin. Phys. Rev.
137(3A), A748 (1965)
9. Collocott, S.: A simple microcomputer-controlled calorimeter: the heat capacity of copper,
invar and RbNiCl3 in the range 2? 20 K. Aust. J. Phys. 36(4), 573–582 (1983)
10. Ho, J.C., O’Neal, H., Phillips, N.E.: Low temperature heat capacities of constantan and
manganin. Rev. Sci. Instrum. 34, 782 (1963)
11. Ho, J.C., Phillips, N.E.: Tungsten–platinum alloy for heater wire in calorimetry below 0.1 K.
Rev. Sci. Instrum. 36(9), 1382 (1965)
12. Hagmann, C., Richards, P.: Specific heat of stainless steel below T = 1 K. Cryogenics 35(5),
345 (1995)
13. Du Chatenier, F., Boerstoel, B., De Nobel, J.: Specific heat capacity of a stainless steel.
Physica 31(7), 1061–1062 (1965)
14. Schink, H., Lohneysen, H.: Specific heat of Apiezon N grease at very low temperatures.
Cryogenics 21(10), 591–592 (1981)
15. Wun, M., Phillips, N.: Low temperature specific heat of Apiezon N grease. Cryogenics 15(1),
36–37 (1975)
16. Stephens, R.: Low-temperature specific heat and thermal conductivity of noncrystalline
dielectric solids. Phys. Rev. B 8(6), 2896 (1973)
17. Jayasuriya, K., Stewart, A., Campbell, S.: The specific heat capacity of GE varnish (200-
400 K). J. Phys. E: Sci. Instrum. 15(9), 885 (1982)
18. Pobell, F.: Matter and methods at low temperatures. Springer, New York (2007)
19. Barucci, M., Gottardi, E., Olivieri, E., Pasca, E., Risegari, L., Ventura, G.: Low-temperature
thermal properties of polypropylene. Cryogenics 42(9), 551–555 (2002)
20. Nittke, A., Scherl, M., Esquinazi, P., Lorenz, W., Li, J., Pobell, F.: Low temperature heat
release, sound velocity and attenuation, specific heat and thermal conductivity in polymers.
J. Low Temp. Phys. 98(5–6), 517–547 (1995)
21. White, G., Birch, J., Manghnani, M.H.: Thermal properties of sodium silicate glasses at low
temperatures. J. Non-Cryst. Solids 23(1), 99–110 (1977)
22. Brewer, D., Edwards, D., Howe, D., Whall, T.: A simple helium-3 cryostat and the specific
heats of nylon and an epoxy resin below 4 2 K. Cryogenics 6(1), 49–51 (1966)
23. Collocott, S., White, G.: Heat capacity and thermal expansion of Zerodur and Zerodur M at
low temperatures. Cryogenics 31(2), 102–104 (1991)
25. Sandin, T., Keesom, P.: Specific heat and paramagnetic susceptibility of stoichiometric and
reduced rutile (TiO2) from 0.3 to 20 K. Phys. Rev. 177(3), 1370 (1969)
26. van de Straat, D., Baak, J., Brom, H., Schmidt, T., Völker, S.: Low-temperature specific heat
of rare-earth-doped silicate glasses. Phys. Rev. B 53(5), 2179 (1996)
27. Smith, D.R., Fickett, F.: Low-temperature properties of silver. J. Res. Natl. Inst. Stand. Tech.
100, 119 (1995)
28. White, G.K., Meeson, P.: Experimental Techniques in Low-Temperature Physics (Monographs
on the Physics and Chemistry of Materials, 59) (2002)
29. Marquardt, E., Le, J., Radebaugh, R.: 11th international cryocooler conference 20–22 June
2000 keystone, Co. Cryogenic material properties database, National Institute of Standards
and Technology Boulder, CO 80303
30. Suski, W., Gofryk, K., Hackemer, A., Wochowski, K.: Low temperature specific heat and
thermoelectric power of UCu3M2Al7 alloys. J. Alloys Compd. 423(1), 37–39 (2006)
32. Melot, B.C., Tackett, R., O’Brien, J., Hector, A.L., Lawes, G., Seshadri, R., Ramirez, A.P.:
Large low-temperature specific heat in pyrochlore Bi2Ti2O7. Phys. Rev. B 79(22), 224111
(2009)
33. Irwin, J., LaCombe, J.: Specific heats of ZnTe, ZnSe, and GaP. J. Appl. Phys. 45(2), 567–573
(1974)
34. Reed, R., Schramm, R., Clark, A.: Mechanical, thermal, and electrical properties of selected
polymers. Cryogenics 13(2), 67–82 (1973)
Part II
Thermal Expansion
Main Symbols
b Coefficient of volumetric expansion

a Coefficient of linear expansion
C Heat capacity/capacitance
c Specific heat
T Temperature
E Internal energy
V Volume
S Entropy
cG Grüneisen coefficient
hD Debye Temperature
xD Cutoff frequency
kB Boltzmann constant
h Plank constant

h h/2p
a Average linear expansion coefficient
S Area
e Dielectric constant
d Distance
r Radius
E Electric field
A Amplitude
u Phase
m Frequency
x 2pm
t Time
k Wave vector
Chapter 4
Thermal Expansion
Abstract All solid materials, when cooled to low temperatures experience a

change in physical dimensions which called ‘‘thermal contraction’’ and is typically
lower than 1 % in volume in the 4–300 K temperature range. Although the effect
is small, it can have a heavy impact on the design of cryogenic devices. The
thermal contraction of different materials may vary by as much as an order of
magnitude: since cryogenic devices are constructed at room temperature with a lot
of different materials, one of the major concerns is the effect of the different
thermal contraction and the resulting thermal stress that may occur when two
dissimilar materials are bonded together. In this chapter, theory of thermal con-
traction is reported in Sect. 1.2. Section 1.3 is devoted to the phenomenon of
negative thermal expansion and its applications.
4.1 Introduction
All materials experience a change in physical dimensions when cooled to low

temperatures. This effect, normally called thermal contraction, is typically lower
than 1 % in volume for most materials in the 4–300 K temperature range.
Although the effect is small, it can have a massive impact on the design of
cryogenic devices. The thermal contraction of different materials may vary by as
much as an order of magnitude. Since cryogenic devices are assembled at room
temperature with a variety of materials, one of the major concerns is the effect of
the different thermal contraction and the resulting thermal stress that may occur
when two dissimilar materials are bonded together. Differential contraction is
particularly important for the design of low temperature vacuum seals, structural
supports, and electrical insulation systems. Thus, it is of paramount importance to
know the thermal expansion of technical materials.
Possibly, the most known device that exploits the basic principle of thermal
expansion is the mercury thermometer. In -38–356 C temperature range, mercury
has a coefficient of thermal expansion much higher compared to that of glass, and

82 4 Thermal Expansion
thus, when the temperature is increased (decreased), the liquid metal expands
(contracts) much more than the glass pipe, allowing the measurement of temperature.
4.2 Thermal Expansion Theory
If energy is supplied to a solid, two phenomena take place: an increase in tem-

perature and a change in volume; both of these are directly related to the increased
vibrational energy of the molecules. The former phenomenon is described in terms
of heat capacity, while the latter leads to the concept of thermal expansion caused
by anharmonic terms in the restoring potential between the single molecules [1].
The Debye model assumes that a solid is described by a set of harmonic
oscillators. In this scenario, the oscillator frequencies in a perfect crystal (normal
mode frequencies) are unaffected by change in volume, so no thermal expansion is
predicted. From measurements, we know that this model is not correct. In fact, the
introduction of anharmonic terms in the interaction potential is needed to get a
positive volumetric coefficient of thermal expansivity, function of pressure, and
temperature defined as

o ln V 1 oV
bffi ¼ : ð4:1Þ
oT V oT p
Not all materials show an isotropic expansion; instead, the extent of expansion
may depend on the particular measuring direction. To quantify anisotropic thermal
expansion (e.g., in a single crystal), it is more useful to define the thermal
expansion in some particular (crystallographic) direction and in such cases, the
linear expansion coefficient, a, is used. This is defined as

o ln L 1 oL
affi ¼ : ð4:2Þ
oT p L oT p
Note that b, for a definite phase of the solid, is always positive, whereas a in
anisotropic material may be also negative (see Sect. 4.3). These latter substances,
commonly called Negative Thermal Expansion (NTE) materials, can be mixed
with materials that possess positive a materials to form composites with a desired
overall thermal expansion coefficient.
For isotropic materials, the expansion in the three directions x, y and z is equal,
and thus,
b ¼ 3a: ð4:3Þ
Sometimes, especially in cases where the composition of the material is well

known, a may be defined in terms of density (q):
4.2 Thermal Expansion Theory 83

1 oq
aq ¼ : ð4:4Þ
q oT p
If instead the material is anisotropic, the quantity a becomes a tensor, and can
be correlated to the strain tensor (e) as
eij ¼ aij dT; ð4:5Þ
with i, j = 1, 2, 3. According to this notation, the diagonal components of the

tensor e are the strains along the three orthogonal axis and the transverse com-
ponents are half the shear strains in all possible planes [2].
For isotropic solids, (4.3) holds, and thus we can write

1 oV fT op
a¼ ¼ ð4:6Þ
3V oT p 3 oT V
where fT is the isothermal compressibility

1 oV
fT ¼ : ð4:7Þ
V op T
For a solid, the anharmonic terms can be represented by the Grüneisen coef-
ficient for the normal mode k, s (k is the wave vector and s the branch index):
oðln xD Þ
cks ffi ð4:8Þ
oðln VÞ
where xD is the cut-off frequency (maximum phonon frequency) [1].

Rigorously, this coefficient can be defined for all the modes, but for our pur-
poses, it is better to write a weighted average Grüneisen coefficient
P
cks cVs ðkÞ
k;s
cG ¼ P ð4:9Þ
cVs ðkÞ
k;s
where cVs(k) is the volumetric heat capacity for the normal mode k, s. It is
important to notice that cG is related to the crystal packing, so if a material can be
obtained in different crystal structures, different values of cG must be considered.
In Table 4.1, some values of cG are reported; for the related crystal packing and
other information, refer to [3–10].
From (4.6) and (4.9), we can rewrite a as
cG c V fT
a¼ : ð4:10Þ
3
At low temperatures, a is far from linear and approaches absolute zero with zero
slope, a fact that can be understood in terms of thermodynamics. The difference
Table 4.1 Grüneisen parameter (dimensionless) for some elements [3–10]
between the constant volume and constant pressure heat capacity is shown [1] to
be proportional to the square of the volume expansivity b, that is,
2
oV op b2
cp cV ¼ T ¼ TV: ð4:11Þ
oT p oV T fT
Because of the Third Law of Thermodynamics, the quantity (cp – cV) must go to
zero as T ? 0; it follows that b must also do the same. This effect has a physical
sense because the harmonic terms should dominate the interatomic potential at
such low temperatures. These considerations are in agreement with experimental
results [11, 12].
According to the Debye model, all the (phonon) frequencies of the crystal scale
linearly with the cut-off frequency following the relation
kB hD ¼ hxD : ð4:12Þ
Taking into account the derivation rules, we can write (4.8) as
oðln hD Þ
cG ¼ : ð4:13Þ
oðln VÞ
For design purposes, it is often more useful to know the integrated thermal
contraction DL/L = (LT - L293K)/L293K (see Fig. 4.1) instead of a.
Now, if we consider the change in length of a sample between two temperatures
T1 and T2, from (4.2), we get
4.2 Thermal Expansion Theory 85
Fig. 4.1 Thermal linear expansion DL/L = (LT - L293K)/L293K of materials. a Polymers [13–17]
(TFE tetrafluoroethylene, CTFE polychlorotrifluoroethylene, a type of Teflon). b Metals [14, 18],
(c) composites [14, 16, 19–21] (For glass–epoxy composites, G-10CR (warp) indicates the
expansion along the glass–fiber direction, whereas G-10CR (normal) is transverse to the fiber
direction), (d) alloys [14, 19, 22]
ZT1
LðT2 Þ
aðTÞdT ¼ ln ð4:14Þ
LðT1 Þ
T2
" #
RT2
aðTÞdT
LðT2 Þ ¼ LðT1 Þe T1
: ð4:15Þ
If the coefficient a is constant or presents a smooth dependence on temperature,

we can define an average thermal contraction coefficient a. If aðT2 T1 Þ 1, at
the first order,
LðT2 Þ ffi LðT1 Þ½1 þ aðT2 T1 Þ ð4:16Þ
or
LðT2 Þ LðT1 Þ
ffi aðT2 T1 Þ ¼ aDT: ð4:17Þ
LðT1 Þ
In Fig. 4.1, note that metals typically have total contractions in the range of
0.5 % or less with the lowest value being for Invar, which is a special alloy
designed for that purpose. Polymers, such as epoxy or Teflon, can have a total
contraction as high as 2 %. An exception is Torlon, which contracts less than most
of Al alloys. Some amorphous materials, particularly Pyrex, have nearly zero a.
Composite materials can often have their thermal contraction predicted by a
linear combination of the two individual materials, taking into account the elastic
modulus of each constituent. This approach to estimate the thermal contraction of a
composite is referred to as ‘‘the rule of mixtures’’ [23]. However, composite
materials are frequently anisotropic by design, which makes their a dependent on
the internal structure and orientation of the included materials. A clear example of
this behavior can be seen in the structural material, G-10, which is a composite of
epoxy and fiber glass. In this case, the thermal contraction of the composite depends
both on the volume ratios of the two materials and on the orientation of the fibers
within the composite. For example, the integrated DL/L from 300 to 4.2 K is about
0.25 % in the fiber direction (wrap) and about 0.75 % normal to the fiber direction.
4.3 Negative Thermal Expansion
The discovery of volume reduction upon heating (NTE) is usually dated as

occurring in 1933 [24]. Nevertheless, examples of systems which exhibit anom-
alous thermal expansion behavior were known before, as is the case of the fer-
romagnetic Invar alloy Fe0.64Ni0.36 and alloys of similar composition [25]. Later,
this area captured the interest of several authors [26–33]. Other systems which
show NTE include the orthorhombic antiferromagnetic CuCl2H2O [34], CeNiSn
[35] and the metal Holmium [36]. Materials such as Lu2Fe17 and Y2Fe17 also show
negative thermal expansion below approximately 400 K [37].
Tino and Iguchi [38] investigated the possibility of NTE in Fe-Pd alloys and
correlated their thermal expansion with that of Invar. The structure and magnetic
properties of Y2Al3Fe14-xMnx compounds [39] were investigated by means of X-
ray diffraction and NTE behavior was found and attributed to magnetostriction.
Also, a lot of lanthanide-based compounds exhibit similar behaviors [40–43]. The
magnetic effects on NTE molecule-based magnets M[N(CN)2]2 (M = Co, Ni) [44]
were also investigated.
4.3.1 Application of NTE
Materials exhibiting NTE are not only of interest because of the fact that this
behavior is highly anomalous, but also because this anomalous behavior can been
used in many practical applications. For example:
4.3 Negative Thermal Expansion 87
(1) First of all, in the composites industry, negative thermal expansion

materials are used as components of composites to adjust the overall ther-
mal expansion of composites to some particular value. The thermal
expansion of composites based on component properties is complex since
properties other than the thermal expansion of the starting materials are
important, in particular, bulk or elastic moduli. Note, e.g., that systems
made of carbon fiber composites and metals, with an extremely negative
value of CTE, have been developed (with values about three times that of
steel, but with a negative sign) [45]. A further advantage of these systems is
their very low thermal conductivity and high compressive strength [23].
(2) In the production of symmetric laminated with useful thermal defor-
mation properties; Wetherhold and Wang [46, 47] suggest that by com-
bining laminate having a positive coefficient of thermal expansion with
laminate with NTE, one can tailor the laminate thermal curvature.
A three-layer, e.g., can be used either to control laminate CTE or eliminate
thermal curvature. A five-layer laminate can be used to control both CTE
and eliminate curvature.
(3) In the construction of mechanically enhanced capillary columns; through
the deposition of NTE TiO2-doped silica on the inside and outside of the
silica tube, a compression of the tube takes place as it is heated, reducing the
propagation of surface flaws [48].
(4) In the electronics area, where substrates and heat sinks that match the
thermal expansion of Si are needed, there are currently several efforts in this
area that use ZrW2O8 to reduce thermal expansion. In the heat sink appli-
cation, Cu/ZrW2O8 composites have successfully matched the thermal
expansion of Si over some hundred degree temperature range [49].
(5) In the production of advanced high-field superconducting solenoid
magnets, such magnets sometimes quench by wire motion induced by
electromagnetic force. It was suggested that the quick wire motions could be
constrained by a high strength polyethylene fiber-reinforced plastic (DFRP)
bobbin with a NTE coefficient and a low frictional coefficient [50].
In fact, Dyneema–glass hybrid composite fiber-reinforced plastic (DGFRP)
has NTE coefficient, low frictional coefficient and high thermal conduc-
tivity. Its use as a material of a coil bobbin has also been described in
literature by Takeo et al. [51].
(6) In the making of interfaces exhibiting good adhesion properties, it has
been reported that additional NTE reinforcing fibers, such as Kevlar fibers,
are helpful to strengthen the reliability of the interface and enhance the
actuating ability of SMA (shape memory alloy) hybrid composites [52].
(7) In medical applications: Another interesting application for NTE materials
is that of adjusting the thermal expansion of the white composites used in
teeth fillings. It is suggested that the thermal expansion of teeth and con-
ventional fillings mismatch and this may result in failure [53].
(8) In the production of materials which do not change shape when heated.
Zero thermal expansion is, of course, of profound interest. One of the
biggest uses of such materials is as substrate materials for mirrors in various
telescope and satellite applications [54].
(9) In the photonics sector, the use of NTE materials has been suggested in
chirped fiber gratings. Wei and coworkers [55] theoretically analyzed and
experimentally demonstrated a simple method for adjusting the chirp of
chirped fiber gratings by means of a temperature method, while the central
wavelength is temperature insensitive. In this work, chirped fiber grating
with a tapered cross section area was mounted under tension in NTE
material. Similar work has been carried out in this field by other workers,
e.g., Mavoori et al. [56] and Ngo et al. [57] who also reported that channel
waveguides with Bragg gratings have been fabricated on glass ceramic
substrates with the NTE coefficient.
In optical fiber reflective grating devices, high-precision optical mirrors,
printed circuit boards and catalyst supports [2, 39, 58–62].
(10) In cryogenic engineering, such as low heat leak cryogenic valves or piston/
piston sleeves of refrigerators [63, 64].
NTE behavior has been studied experimentally and theoretically [65–68]. NTE
has been reviewed by Barrera et al. in Ref. [61]. For more information regarding
the CTE prediction of composite based on association of materials with positive
CTE and NTE, see [69–71].
References
1. Ashcroft, N.W., Mermin, N.D.: Solid State Physics. Holt. Rinehart and Winston, New York
(1976)
2. Grima, J.N., Zammit, V., Gatt, R.: Negative thermal expansion. Xjenza 11, 17–29 (2006)
3. Wallace, D.C.: Melting of elements. Proc. R. Soc. Lond. A 433(1889), 631–661 (1991)
4. Parshukov, A.: Measurement of the Gruneisen coefficients and their dependence on the
volume of certain metals. Fiz. Tverd. Tela 27(4), 1228–1232 (1985)
5. Magomedov, M.: Atomic interaction parameters for lanthanides and actinides. Russ. J. Inorg.
Chem. 52(12), 1953–1962 (2007)
6. Moruzzi, V., Janak, J., Schwarz, K.: Calculated thermal properties of metals. Phys. Rev. B
37(2), 790 (1988)
7. Singh, H.: Determination of thermal expansion of germanium, rhodium and iridium by X-
rays. Acta Crystallogr. A Cryst. Phys. Diffr. Theoret. Gen. Crystallogr. 24(4), 469–471
(1968)
8. Manghnani, M.H., Katahara, K., Fisher, E.S.: Ultrasonic equation of state of rhenium. Phys.
Rev. B 9(4), 1421 (1974)
9. Rao, R.R., Ramanand, A.: Thermal expansion and bulk modulus of cobalt. J. Low Temp.
Phys. 26(3–4), 365–377 (1977)
10. Hamlin, J., Tissen, V., Schilling, J.: Superconductivity at 20 K in yttrium metal at pressures
exceeding 1Mbar. Phys. C 451(2), 82–85 (2007)
11. White, G.K., Collins, J.G.: Thermal expansion of copper, silver, and gold at low
temperatures. J. Low Temp. Phys. 7(1–2), 43–75 (1972). doi:10.1007/bf00629120
References 89
12. White, G., Meeson, P.: Experimental Techniques in Low-Temperature Physics. Clarendon
13. Corruccini, R.J., Gniewek, J.J.: Thermal expansion of technical solids at low temperatures: a
compilation from the literature. National Bureau of Standards, US Department of Commerce
(1961)
14. Reed, R.P., Clark, A.F.: Materials at Low Temperatures. American Society for Metals, Ohio,
(1983)
15. Kirby, R.K.: Thermal expansion of polytetrafluoroethylene (Teflon) from -190 to + 300 C.
J. Res. Nat. Bur. Stan. 57(2), 91–94 (1956)
16. Laquer, H.L.: Low temperature thermal expansion of various materials. Technical
information Service (AEC), Oak Ridge, TN (1952)
17. Marquardt, E., Le, J., Radebaugh, R.: 11th International Cryocooler Conference June 20–22,
2000 Keystone, Co. Cryogenic Material Properties Database, National Institute of Standards
and Technology Boulder, CO
18. Hahn, T.A.: Thermal expansion of copper from 20 to 800 K—Standard reference material
736. J. Appl. Phys. 41(13), 5096–5101 (1970)
19. Clark, A.: Low temperature thermal expansion of some metallic alloys. Cryogenics 8(5),
282–289 (1968)
20. Dahlerup-Petersen, K., Perrot, A.: Properties of Organic Composite Materials at Cryogenic
Temperatures. CERN, Geneva (1979)
21. Clark, A., Fujii, G., Ranney, M.: The thermal expansion of several materials for
superconducting magnets. IEEE Trans. Magn. 17(5), 2316–2319 (1981)
22. Arp, V., Wilson, J., Winrich, L., Sikora, P.: Thermal expansion of some engineering
materials from 20 to 293 K. Cryogenics 2(4), 230–235 (1962)
23. Hartwig, G.: Low-temperature properties of epoxy resins and composites. In: Timmerhaus,
K.D., Reed, R.P., Clark, A.F. (eds.) Advances in Cryogenic Engineering, pp. 17–36.
Springer, New York (1978)
24. Guillaume, C.: Open innovation challenges. Nature 131, 658 (1933)
25. Guillaume, C.É.: Recherches sur les aciers au nickel. Dilatations aux temperatures elevees;
resistance electrique. CR Acad. Sci 125(235), 18 (1897)
26. Chikazumi, S.: Invar anomalies. J. Magn. Magn. Mater. 10(2), 113–119 (1979)
27. Schlosser, W., Graham, G., Meincke, P.: The temperature and magnetic field dependence of
the forced magnetostriction and thermal expansion of Invar. J. Phys. Chem. Solids 32(5),
927–938 (1971)
28. Manosa, L., Saunders, G., Rahdi, H., Kawald, U., Pelzl, J., Bach, H.: Longitudinal acoustic
mode softening and Invar behaviour in Fe72Pt28. J. Phys.: Condens. Matter 3(14), 2273
(1991)
29. Manosa, L., Saunders, G., Rahdi, H., Kawald, U., Pelzl, J., Bach, H.: Acoustic-mode
vibrational anharmonicity related to the anomalous thermal expansion of Invar iron alloys.
Phys. Rev. B 45(5), 2224 (1992)
30. Saunders, G., Senin, H., Sidek, H., Pelzl, J.: Third-order elastic constants, vibrational
anharmonicity, and the Invar behavior of the Fe72Pt28 alloy. Phys. Rev. B 48(21), 15801
(1993)
31. van Schilfgaarde, M., Abrikosov, I., Johansson, B.: Origin of the Invar effect in iron-nickel
alloys. Nature 400(6739), 46–49 (1999)
32. Kainuma, R., Wang, J., Omori, T., Sutou, Y., Ishida, K.: Invar-type effect induced by cold-
rolling deformation in shape memory alloys. Appl. Phys. Lett. 80(23), 4348–4350 (2002)
33. Collocott, S., White, G.: Thermal expansion and heat capacity of some stainless steels and
FeNi alloys. Cryogenics 26(7), 402–405 (1986)
34. Harding, G., Lanchester, P., Street, R.: The low temperature magnetic thermal expansion of
CuCl2. 2H2O. J. Phys. C: Solid State Phys. 4(17), 2923 (1971)
35. Aliev, F., Villar, R., Vieira, S., de la Torre, M.L., Scolozdra, R., Maple, M.: Energy gap of
the ground state of CeNiSn caused by local and long-range magnetic-moment interactions.
Phys. Rev. B 47(2), 769 (1993)
36. White, G.K.: Phase transitions and the thermal expansion of holmium. J. Phys.: Condens.
Matter 1(39), 6987 (1989)
37. Gignoux, D., Givord, D., Givord, F., Lemaire, R.: Invar properties in the rare earth-3d
transition metal alloys. J. Magn. Magn. Mater. 10(2), 288–293 (1979)
38. Tino, Y., Iguchi, Y.: Zero or negative thermal expansion of Fe-Pd alloys and the cause of
Invar characteristic. J. Magn. Magn. Mater. 31, 117–118 (1983)
39. Hao, Y., Gao, Y., Wang, B., Qu, J., Li, Y., Hu, J., Deng, J.: Negative thermal expansion and
magnetic properties of Y2Al3Fe14-xMnx. Appl. Phys. Lett. 78(21), 3277–3279 (2001)
40. Hao, Y., Zhao, M., Zhou, Y., Hu, J.: Negative thermal expansion and spontaneous volume
magnetostriction of Tb2Fe16Cr compound. Scripta Mater. 53(3), 357–360 (2005)
41. Hao, Y., Zhou, Y., Zhao, M.: Negative thermal expansion and spontaneous magnetostriction
of Dy2AlFe10Mn6 compound. Adv. Eng. Mater. 7(6), 517–520 (2005)
42. Yan-Ming, H., Yan, Z., Miao, Z.: Spontaneous magnetostriction of Dy2AlFe13Mn3
compound. Chin. Phys. 14(7), 1449 (2005)
43. Yan-Ming, H., Miao, Z., Yan, Z.: Spontaneous magnetostriction of Y2Fe16Al compound.
Chin. Phys. 14(4), 818 (2005)
44. Kmety, C.R., Manson, J.L., Huang, Q., Lynn, J.W., Erwin, R.W., Miller, J.S., Epstein, A.J.:
Collinear ferromagnetism and spin orientation in the molecule-based magnets M[N(CN)2]2
(M = Co, Ni). Phys. Rev. B 60(1), 60 (1999)
45. Hartwig, G.: Polymer Properties at Room and Cryogenic Temperatures. Springer, New York
(1994)
46. Wetherhold, R.C., Wang, J.: Tailoring thermal deformation by using layered beams. Compos.
Sci. Technol. 53(1), 1–6 (1995)
47. Wetherhold, R.C., Wang, J.: Controlling thermal deformation by using laminated plates.
Compos. B Eng. 27(1), 51–57 (1996)
48. Berthou, H., Neumann, V., Dan, J., Hintermann, H.: Mechanically enhanced capillary
columns. Surf. Coat. Technol. 61(1), 93–96 (1993)
49. Holzer, H., Dunand, D.: Processing, structure and thermal expansion of metal matrix
composites containing zirconium tungstate. In: 4th International Conference on Composite
Engineering, Hawaii 1997
50. Yamanaka, A., Kashima, T., Nago, S., Hosoyama, K., Takao, T., Sato, S., Takeo, M.: Coil
bobbin composed of high strength polyethylene fiber reinforced plastics for a stable high field
superconducting magnet. Phys. C 372, 1447–1450 (2002)
51. Takeo, M., Sato, S., Matsuo, M., Kiss, T., Takao, T., Yamanaka, A., Kashima, T., Mito, T.,
Minamizato, K.: Dependence on winding tensions for stability of a superconducting coil.
Cryogenics 43(10), 649–658 (2003)
52. Cui, L.S., Schrooten, J., Zheng, Y.J.: Effects of additional reinforcing fibers on the interface
quality of SMA wire/epoxy composites. In: Materials Science Forum 2005, pp. 2047–2050.
Trans Tech Publications (2005)
53. Versluis, A., Douglas, W.H., Sakaguchi, R.L.: Thermal expansion coefficient of dental
composites measured with strain gauges. Dent. Mater. 12(5), 290–294 (1996)
54. Collins, E.G., Richter, S.: Linear-quadratic-gaussian-based controller design for Hubble
space telescope. J. Guid. Control Dyn. 18(2), 208–213 (1995)
55. Wei, Z., Yu, Y., Xing, H., Zhuo, Z., Wu, Y., Zhang, L., Zheng, W., Zhang, Y.: Fabrication of
chirped fiber grating with adjustable chirp and fixed central wavelength. IEEE Photonics
Technol. Lett. 13(8), 821–823 (2001)
56. Mavoori, H., Jin, S., Espindola, R., Strasser, T.: Enhanced thermal and magnetic actuations
for broad-range tuning of fiber Bragg grating based reconfigurable add drop devices. Opt.
Lett. 24(11), 714–716 (1999)
57. Ngo, N., Li, S., Zheng, R., Tjin, S., Shum, P.: Electrically tunable dispersion compensator
with fixed center wavelength using fiber Bragg grating. J. Lightwave Technol. 21(6), 1568
(2003)
References 91
58. Evans, J., Hanson, P., Ibberson, R., Duan, N., Kameswari, U., Sleight, A.: Low-temperature
oxygen migration and negative thermal expansion in ZrW2-xMoxO8. J. Am. Chem. Soc.
122(36), 8694–8699 (2000)
59. Evans, J.S., Hu, Z., Jorgensen, J., Argyriou, D., Short, S., Sleight, A.: Compressibility, phase
transitions, and oxygen migration in zirconium tungstate, ZrW2O8. Science 275(5296), 61–65
(1997)
60. Kintaka, K., Nishii, J., Kawamoto, Y., Sakamoto, A., Kazansky, P.G.: Temperature
sensitivity of Ge-B-SiO2 waveguide Bragg gratings on a crystallized glass substrate. Opt.
Lett. 27(16), 1394–1396 (2002)
61. Barrera, G., Bruno, J., Barron, T., Allan, N.: Negative thermal expansion. J. Phys.: Condens.
Matter 17(4), R217 (2005)
62. Sleight, A.: Compounds that contract on heating. Inorg. Chem. 37(12), 2854–2860 (1998)
63. Huang, R., Xu, W., Xu, X., Li, L., Pan, X., Evans, D.: Negative thermal expansion and
electrical properties of Mn3(Cu0.6NbxGe0.4- x)N (x = 0.05–0.25) compounds. Mater. Lett.
62(16), 2381–2384 (2008)
64. Huang, R., Wu, Z., Yang, H., Chen, Z., Chu, X., Li, L.: Mechanical and transport properties
of low-temperature negative thermal expansion material Mn3CuN co-doped with Ge and Si.
Cryogenics 50(11), 750–753 (2010)
65. Baughman, R., Galvao, D.: Negative volumetric thermal expansion for proposed hinged
phases. Chem. Phys. Lett. 240(1), 180–184 (1995)
66. Evans, J.O.: Negative thermal expansion materials. J. Chem. Soc., Dalton Trans. 19,
3317–3326 (1999)
67. Sleight, A.W.: Negative thermal expansion materials. Curr. Opin. Solid State Mater. Sci.
3(2), 128–131 (1998)
68. Tao, J., Sleight, A.: The role of rigid unit modes in negative thermal expansion. J. Solid State
Chem. 173(2), 442–448 (2003)
69. Miller, W., Mackenzie, D., Smith, C., Evans, K.: A generalised scale-independent
mechanism for tailoring of thermal expansivity: Positive and negative. Mech. Mater.
40(4), 351–361 (2008)
70. Palumbo, N., Smith, C., Miller, W., Evans, K.: Near-zero thermal expansivity 2-D lattice
structures: performance in terms of mass and mechanical properties. Acta Mater. 59(6),
2392–2403 (2011)
71. Steeves, C.A., Dos Santos E Lucato, S.L., He, M., Antinucci, E., Hutchinson, J.W., Evans,
A.G.: Concepts for structurally robust materials that combine low thermal expansion with
high stiffness. J. Mech. Phys. Solids 55(9), 1803–1822 (2007)
Chapter 5
How to Measure the Thermal Expansion
Coefficient at Low Temperatures
Abstract Thermal expansion measurements in the high temperature range have

been thoroughly explored, and various experimental methods are available even as
commercial instrumentation, measurements at cryogenic temperatures have been
confined to the field of high-precision laboratory experiments, needing large
experimental efforts and expenses, and often also suffering from intrinsic limita-
tions. All techniques used for the measurements of thermal expansion can be
divided into two categories, namely: absolute methods and relative methods.
While in the former the linear changes of dimension of the sample are directly
measured at various temperature, in the latter the coefficient of thermal expansion
is determined through comparison with a reference materials of known thermal
expansion. A lot of experimental set-ups are described in Sect. 2.1, while Sect. 2.2
some examples of measurements performed at very low temperatures are listed.
Everything you always wanted to know about the measurement of thermal

expansion, but were afraid to ask is not the name of a Woody Allen movie, though
it might as well be the title of the paper of Kanagaraj and Pattanayak [1] where
several methods of how to measure the Coefficient of Thermal Expansion (CTE)
are described; 14 of them are suited for low temperature measurements. However,
while thermal expansion measurements in the high-temperature range have been
thoroughly explored, and various experimental methods are available, even as
commercial instrumentation, measurements at cryogenic temperatures have been
confined to the field of high-precision laboratory experiments, needing large
experimental efforts and expenses, and often also suffering from intrinsic limita-
tions [2]. On the other hand, we shall see that new methods (see, e.g., Sect. 5.2.4)
have been recently proposed besides those cited in Ref. [1].
All techniques used for the measurements of thermal expansion can be divided
into two categories, namely, absolute methods and relative methods. While in the
former, the linear changes of dimension of the sample are directly measured at
various temperatures, in the latter, CTE is determined through comparison with
reference materials of known thermal expansion. A lot of experimental setups
allow one to measure CTE by both absolute and relative methods as described in
the next sections.

94 5 How to Measure the Thermal Expansion Coefficient
Fig. 5.1 Scheme of a three-terminal capacitance dilatometer and its equivalent circuit
5.1 Capacitive Dilatometers
5.1.1 Principles of Capacitive Techniques
In this method, CTE is measured by the change of capacitance of a capacitor due

to a change of sample dimension. Capacitance change is measured by a two-
terminal or three-terminal method.
The two-terminal set-up essentially consists of a capacitor whose capacity is
determined by the distance between the plates: the first one is in a fixed position,
while the second one is movable and rests on the top of specimen [3]. The use of
this technique is limited because of edge effects, thermal instability and other
parasitic contributions as the change of the effective area of plates. It was applied
mainly in high-temperature measurements [4–6].
The three-terminal set-up presents high sensitivity and is particularly appro-
priate for low-temperature measurements, where dimensional changes of speci-
mens are small for small temperature changes. It allows one to overcome the
limitation due to edge effects and parasitic contributions; in Fig. 5.1, a three-
terminal capacitor with its equivalent circuit is schematically shown.
In this configuration, the electric field between the A and B plates has a very
small edge distortion when the guard ring C surrounds the electrodes leaving a
small gap between the high potential and low potential electrode, smaller than the
distance between the electrodes A and B.
The three-terminal capacitive method, first developed by White [7], has been
applied in a lot of experiments for measurements at low temperatures CTE on
specimens of different size and composition:
• metallic materials: Cu [7–9], Al and Ge [10], Nb [11];
• alloys: DyCu2 [12], SBN [13], YbPd [14];
• polymers: TTF-TCNQ [15], Poly (dl propylene oxide) [16], Teflon [17];
• composite materials: E-glass reinforced epoxy resin [1];
• ceramic materials [18].
5.1 Capacitive Dilatometers 95
Authors adopted capacitive or other techniques depending on the type of

material and temperature range. For details on each particular configuration, refer
to original work. In the following sections, various realizations of this technique
are reported.
The capacitance of an ideal plate capacitor with area S and distance between
plates of d is
S
C¼e ð5:1Þ
d
with e = e0 er where er is the dielectric constant of the medium (gas filling the
expansion cell) and e0 = 8.854 * 10-12 Fm-1. The gas medium is usually 3He at a
pressure of about 10 Pa: we assume e = 1 because even at atmospheric pressure,
Helium presents a dielectric constant very close to 1 (1.00007 at 0 C [19]).
Moreover, most of measurements were carried out under vacuum. In case of
circular plates of radius r, we obtain
r2
d ¼ e0 p ð5:2Þ
C
which implies a length change
r 2 ðC2 ffi C1 Þ
Dd ¼ d2 ffi d1 ¼ ffie0 p : ð5:3Þ
C1 C2
Any parallel plate capacitor presents a distortion of electric field, even if it is

provided by a guard ring. Therefore, (5.1) must be corrected, taking into account
the increase in the effective area of the central electrode from S to
S + DS (DS being an additional strip extending over half the width of the gap, g,
between the electrode and guard ring). This effect gives a contribution to C which
can be calculated with a formula proposed by Maxwell [20] and discussed by
Hartshorn [21]:
r2 rg g
C¼p þp 1þ : ð5:4Þ
d d þ 0:22g 2r
Neglecting terms of the order (g/r)2 and higher,

DC
C ¼ C0 1 þ ð5:5Þ
C0
where C0 = epr2/d and

DC g d
¼ : ð5:6Þ
C0 r d þ 0:22g
With typical value of g, r and d, we obtain a DC/C0 of the order of 0.5 % [7, 8].
Since this correction is a weak function of d, it can be set to a constant value, d.
Another correction is due to the thermal expansion of capacitor plates them-
selves; for a circular plate, including the correction for edge effect (5.2), (5.3)
becomes

C2 ffi C1 Dr
Dd ¼ ffie0 pð1 þ dÞr 2 ffi2 : ð5:7Þ
C1 C2 r1
In Ref. [8], the effect of tilting and the nonflatness of capacitor plates is
evaluated.
Let us consider how length measurement Dd is connected to the thermal
expansion of the sample and how it is influenced by the thermal expansion DLcell
of cell
ffiDd ¼ DLspecimen ffi DLcell : ð5:8Þ
The thermal expansion of the cell can be determined by measuring a sample

(with the same length of the investigated one) with the well-known CTE (e.g., for
copper [9]).
Capacitive-based methods take advantage of the availability of off-the-shelf
capacitance bridges with a resolution of about 1 part per billion (ppb), which
means a change-length resolution of better than 0.1 Å. Most expansion cells are
quite small, allowing measurements on samples of millimeter lengths. Although
this method provides the highest sensitivity, it needs a very good calibration of the
cell. In fact, one of the main disadvantages is that the cell material must have well-
known thermal properties in order to accurately subtract its contribution from raw
data. Furthermore, most expansion cells are made of copper (with a considerable
CTE). For this reason, although this technique is very sensitive, it may not be very
accurate.
5.1.2 Examples
Hereafter, five examples of setups of a capacitive dilatometer are reported.

(a) White’s setup (1961)
The first capacitance cell for measurements of thermal expansion is described
by White [7]. With this instrument, the change of capacitance in a three-terminal
configuration is detected by a Thompson bridge. A sensitivity of 10-7 pF corre-
sponding to a length change of 1 pm was achieved. In this experiment, two
Fig. 5.2 a Cryostat with expansion cell: (1) first invar electrode (2) invar guard ring, (3) second
invar electrode, B2 support ring (A) specimen. b Three-terminal reference capacitor, filled with
dry nitrogen and kept at constant temperature. Reprinted with permission [7]
different cells were built: the first for the measurement of sample CTE, the second
(see Fig. 5.2) for the absolute measurement of the copper used in the first one.
The cell is entirely made of high-conductivity copper, apart from brass screws,
spring washers, and mica-insulating washers. (2) is the central electrode, (3) is the
guard ring and (B2) is the support ring which is assembled with insulating washers
and screws into a unit whose lower face has been machined and lapped for a final
flatness of about 10-8 m. The same grounding and lapping was done to the cyl-
inder B and the end plate B1. End faces of specimens (A), of length about 10 cm,
had to be machined with a precision of 10-6 m.
Equation (5.7) was used to obtain displacement from capacity, taking into
account the distortion due to the edge effect. In the original work, White reported
the correction in case the cylinders are not perfectly coaxial [7].
Drawbacks of this setup are: the precision required in machining the sample and
the instability of the expanding cell in periods of 30–40 min. These limitations are
reduced or overcome in the next examples.
Fig. 5.3 a Scheme of the cryogenic part of the dilatometer, LN liquid nitrogen; 1 Indium gasket;
T Teflon ring; C cell; LHe liquid helium, V1 inner dewar, V2 outer dewar; H heater HS copper
heat shield; C1 outer vacuum can; C2 inner vacuum can, S stainless steel support; P1 P2 pumping
lines. b Construction of the cell: 1 frame; 2 movable part; 3 screw; 4 piston; 5 sample; 6 bayonet
fitting; 7a guard ring; 7b capacitor plate; 8a guard ring; 8b capacitor plate; 9 washers; 10 springs;
11, 12 electrical leads; 13 platinum resistor; 14 silicon diode; 15 germanium resistor. Inset 16
capacitor plate; 17 guard ring; 18 spring; 19 screw; 20 araldite; 21 sapphire spacer; 22 araldite
washer. Reprinted with permission [8]
(b) Pott’s setup (1982)

Pott et al. [8] tried to overcome the limitation by request of sample flatness.
This is an important achievement since some materials are hard or impossible to
machine.
The expansion cell is shown in Fig. 5.3a; its overall dimensions are 46 mm in
length and 32 mm in diameter. All parts are made of high-purity copper. To get a
fast relaxation time, the mass of the cell is small (about 180 g). The high ratio
surface area/total mass ensures a good contact with the 3He exchange gas.
As shown in Fig. 5.3b, the cell is built in a way which allows an easy insertion
of the sample. Note that the cell does not have to be dismounted between two
measurements and the capacitor plates remain in the same position: this avoids
run-to-run nonreproducible changes. Samples of arbitrary shape and of wide range
of length (2–10 mm) can be used. This is a very important point since for materials
with large CTE, the capacitor gap becomes too large upon cooling, thus reducing
the sensitivity of the dilatometer. Some details: the capacitor plates 7b and 8b are
surrounded by guard rings 7a and 8a to avoid stray electric fields; the electrode 8 is
fastened by six screws to the frame 1, whereas the electrode 7 is fixed to the
movable part 2, which is fixed to the frame by two spring washers 10; the upper
electrode 7 can only move vertically. The hollow cylindrical frame 1 was milled
out in a way that allows a free access to the sample. The sample 5 is fixed between
the upper plane of part 2 and the lower plane of piston 4 by the slight force of the
springs 10 (about 1 N) without use of glues. In this configuration, a length change
of the sample causes an equal displacement of the upper electrode with respect to
the lower. The gap between the electrodes can be adjusted by screw 3 to the
desired length (0.1 mm). By means of the coupling 6 and the stainless steel
support, it is possible to turn the screw from the top of the cryostat, even when the
cell is at He temperatures. The piston 4 fits exactly in the screw 3 and can only
move axially. Piston 4 cannot rotate together with screw 3 because it is fixed by
two pins: this prevents brittle samples from being damaged.
After mounting the sample, both vacuum cans were evacuated to a pressure of
about 10-5 Pa and then flooded with 3He exchange gas (*10 Pa). The outer dewar
was filled with liquid nitrogen, and the sample and cell were kept cooled at about
80 K overnight. The lowest temperature of 1.5 K was reached by pumping the
liquid 4He in the inner dewar. Before increasing the temperature, the outer vacuum
can was evacuated, which means that the inner vacuum can was coupled to the 4He
bath only by radiation. The temperature of the heat shield, cell and sample was
then increased by applying power (a few mW) to the heat shield. The rate of the
power input was increased continuously. Owing to the 3He exchange gas, the
sample and cell were in good thermal equilibrium (as long as the heating rate is not
too high). A copper sample (99.999 %, l0 = 7, 5 mm diameter) was used to cal-
ibrate the instrument.
Corrections to values of capacitance are due to the edge effect of the electric
field, expansion of capacitor plates, nonparallelism of plates, and cell effect. With
this configuration, the relative error is of the order of 6 % in the intermediate-
temperature range (25 K \ T \ 50 K); for T [ 50 K, the relative error is smaller
than 2 %.
(c) Roth’s setup, a capacitive dilatometer with elastic diaphragm (1991)
Capacitive setups usually require a high-precision adjustment (in the case of
parallel plates or in the case in which plates have to be fixed to a guide mecha-
nism). Cells based on the elastic deformation of a diaphragm avoid difficult
mechanical adjustments.
Roth et al. [13] developed an instrument based on a displacement sensor with an
elastic diaphragm, working on the capacitance method. This configuration allows
one to avoid mechanical adjustment problems, retaining the sensitivity typical of
the capacitive technique.
In Fig. 5.4, the section of the capacitance displacement sensor is depicted. The
capacitor plates are the diaphragm and the ring electrode made of oxygen-free
high-conductivity copper (OFHC) polished and covered with a gold layer. The
Fig. 5.4 Section of the capacitance sensor [13]
Cu0.98Be0.02 diaphragm is 38 mm in diameter and 150 lm thick (it is also polished

and covered with a gold layer). The thickness of diaphragm was chosen so that a
central force of 0.04 N does not produce a displacement more than 5 lm. The ring
electrode is electrically isolated by mylar foil and Araldite epoxy. The screening
box (connected with a guard electrode for the three-terminal capacitance mea-
surement) and clamp ring are both made of copper to eliminate possible thermal
gradients. This screening box has a central 3 mm bordering within which a copper
piston can slide; on the piston, a polished Al2O3–SiO2 ceramic ball isolates the
piston from the diaphragm and allows one to produce a force exactly at the center
of the diaphragm. To obtain a displacement measurement of a sample, the sensor is
coupled to a sample holder. When the specimen expands, the piston is shifted and
the ball presses against the diaphragm, deforming it.
The measured C(DL) capacitance was calculated from
ZD=2
rdr
CðDlÞ ¼ 2pe0 : ð5:9Þ
b þ Dl½1 ffi ðr=RÞ þ 2ðr=RÞ2 lnðr=RÞ
2
d=2
For vanishing displacements, this formula describes a parallel plate capacitor;

for displacements higher than 50 lm, the membrane is plastically deformed, thus
limiting the range of displacements to less than 50 lm.
Taking into account the resolution of the capacitance bridge, mechanical
vibrations and electrical noise, this method achieves a sensitivity of 0.1 nm.
Fig. 5.5 Scheme of the capacitance dilatometer developed by Rotter et al. [12]
(d) Rotter’s setup, a miniature dilatometer for thermal expansion and

magnetostriction of single crystals (1998)
Rotter et al. [12] developed a miniature capacitance dilatometer to measure the
thermal expansion of small and irregularly shaped samples and, in particular, for
studying phase transitions of intermetallic rare earth compounds. Since for these
compounds only small single crystals (*1 mm3) are available and for such
investigations a wide range of physical parameters is necessary, it was necessary to
design a small and compact dilatometer for a wide temperature range and high
magnetic fields, combining most advantages of the existing capacitive dilatome-
ters, but avoiding their drawbacks. To obtain a reasonable accuracy, the active
length of the sample must be larger than 0.5 mm. The sample can nearly have any
shape; only the base surface should be flat, giving a stable sample position.
The working principle is reported in Fig. 5.5. The lower part consists of a plate
holder (Ag), which includes the ring-like lower capacitor (Ag) and the sample
support (Ag). The upper part consists of the upper plate holder (Ag) and the
(differently shaped) upper capacitor plate. It is separated from the lower one by
two needle bearings (brass) and the sample to obtain a well-defined support on
three points. The capacitor plates, as well as the sample support, are insulated from
the holders by sapphire washers. The needle bearings define an exact pivot point
and avoid any transversal shift between upper and lower plate holder. Although
this sensor is more difficult to calibrate when compared to a normal parallel plate
dilatometer, it gave data with a maximum deviation from literature data of *1 %
in DL/L.
(e) Neumeier’s setup, a capacitive dilatometer with the cell made of fused
quartz (2008)
Neumeier et al. [11] constructed a dilatometer cell that can detect sub-angstrom
changes in the length of solid specimens in the 5 K \ T \ 350 K temperature
range. It is constructed entirely from fused quartz.
CTE of a 10.818 mm long single crystal of Nb near the superconducting
transition temperature was measured, evidencing the jump in CTE expected for a
second-order phase transition.
This setup provides the highest sensitivity (0.1 Å) among capacitive methods,
but it needs an accurate calibration of the cell. In fact, one of the main drawbacks
Fig. 5.6 a Exploded view of the fused quartz dilatometer; b Assembled view of the cell [11]
of this method is that the properties of cell material must be well known in order to
subtract its contribution to raw data.
Neumeier et al. [11] tried to overcome this limitation by making the expansion
cell by fused quartz (see Fig. 5.6) which has a CTE much smaller than copper.
Nevertheless, the problem of the calibration is inherent to this method and cannot
be bypassed. On the other hand, an advantage is the very low size of the cell,
which allows the measurements of samples smaller than 1 mm.
As shown in Fig. 5.6a, the cell is made of five pieces. A moveable L-shaped
piece and a stationary L-shaped piece form the capacitor plates. On the vertical
faces of the L-shaped pieces, a 100–1000 Å Cr–Au film is deposited to form the
metallic capacitor plates. The L-shaped pieces are joined with two fused quartz
plates, which act as springs (glued to the L-shaped pieces using a mixture of
talcum powder and sodium silicate solution). The stationary L-shaped piece has a
cutout on the top with a 3 angle relative to the capacitor plate. The sample is
placed on top of the movable plate and a wedge, also with a 3 angle, is pushed
parallel to the capacitor plates to wedge the sample between it and the movable
plate. This applies tension to the springs in order to establish the desired capacitor
gap at room temperature. For the thermal expansion measurements, Neumeier
et al. used a helium cryostat with a custom designed insert that allowed mea-
surements in the 5–350 K range. The calibration was made by a comparison with
standard of copper.
5.2 Interferometric Dilatometers
Several mechanical–optical thermal expansion measurement methods have also

been developed, for example, the optical comparator [22] or the twin telemicro-
scope [23] technique. However, in these kinds of measurements, optical techniques
5.2 Interferometric Dilatometers 103
are only used as a method to amplify the sample displacement, while calibration
problems, typical of mechanical methods, are still present. A purely optical
method, e.g., the interferometric technique, permits one to overcome such
limitations.
An example of a typical application is a Fabry–Perot setup, where the spacer
between the reflectors is made of the sample material. [24] There is a lot of work
based on such dilatometers, also at room temperature and at high temperatures,
[23, 25–33] but in this book, we will restrict our analysis only to low-temperature
setups.
In the following sections, homodyne and heterodyne dilatometers will be dis-
cussed: the former type of dilatometer is based on a single-frequency laser source
(see Sect. 5.2.2), while the latter works by using a two-frequency laser (see
Sect. 5.2.3). A further classification is based on the cryogenic system: a cryoliquid
cryostat or a mechanical refrigerator cryostat.
5.2.1 Principles of Interferometric Dilatometry
In this section, the basic principles of interferometry as applied to length mea-

surements are revisited, starting from the electromagnetic wave representation of a
light source.
The electric field component, E, of an electromagnetic wave can be described
by
Eðx; y; z; tÞ ¼ ~
~ Aðx; y; z; tÞeiuðx;y;z;tÞ ð5:10Þ
or
~ r; tÞ ¼ ~
Eð~ r; tÞeiuð~r;tÞ
Að~ ð5:11Þ
where A represents the amplitude of the field, the complex part contains the
information about the phase of the wave and r is the position vector. Both
amplitude A and phase u are functions of the spatial coordinates and time (the
polarization state of the field being contained in the temporal variations in the
amplitude vector).
For a monochromatic light,
~ r; tÞ ¼ ~
Eð~ rÞei½xtffiuð~r;tÞ
Að~ ð5:12Þ
where x is the angular frequency.

The time dependence has been eliminated from the amplitude term to indicate a
constant linear polarization, while the phase has been split into spatial and tem-
poral terms.
In the measurements, laser sources will be described as plane waves. The

complex amplitude of a linearly polarized plane wave is
~
~ r; tÞ ¼ ~
Eð~ Aei½xtffik~r ð5:13Þ
where k is the wave vector.

If the direction of propagation is parallel to the z axis, the expression for the
complex amplitude of the plane wave becomes
~ r; tÞ ¼ ~
Eð~ Aei½xtffikz : ð5:14Þ
Keep in mind that wavefronts represent surfaces of constant phase for the
electromagnetic field; they are normally used to show the spatial variations of the
field, and are drawn or computed at a fixed time. A wavefront also represents a
surface of constant optical path length (OPL) from the source
ZP
OPL ¼ nðsÞds ð5:15Þ
S
where S is the source position, P is the observation point and n(s) is the refraction
index along the path. The local normal to the wavefront defines the propagation
direction of the field.
The net complex amplitude of the field is the sum of all field components
X
~ r; tÞ ¼
Eð~ ~
Ei ð~
r; tÞ; ð5:16Þ
i
and the resulting field intensity (time averaged over a period much longer than 1/m)
is
2
r; tÞ ¼ ~
Ið~ r; tÞ :
Eð~ ð5:17Þ
In the case of two interfering waves E1 and E2,

e0 c 2 ffi
r; tÞ ¼
Ið~ E1 þ E12 þ E1 E2 þ E1 E2 ð5:18Þ
2
or
e0 c ffi
r; tÞ ¼ I1 þ I2 þ
Ið~ E1 E2 þ E1 E2 ð5:19Þ
2
where I1 and I2 are the intensities due to the two beams individually.
This general result can be greatly simplified if we assume linearly polarized
monochromatic waves of the form (5.10)
~ r; tÞ ¼ ~
Ei ð~ rÞei½xi tffiui ðx;y;zÞ :
Ai ð~ ð5:20Þ
The resulting field intensity is

e0 c n ~ ~ o
r; tÞ ¼ I1 þ I2 þ
Ið~ 2A1 A2 cos½ðx1 ffi x2 Þt ffi ðuð~
rÞ ffi uð~
rÞÞ : ð5:21Þ
2
The third term of (5.21) represents interference. Note that if the two interfering
waves are orthogonally polarized, the interference term does not exist. Also, if the
frequencies of the two waves are different, the interference effects will be mod-
ulated at a beat frequency equal to the difference frequency.
Let us suppose that the two linear polarizations are parallel and that the two
waves are at the same frequency. Equation (5.19) becomes
pffiffiffiffiffiffiffi
rÞ ¼ I1 þ I2 þ 2 I1 I2 cosðDuð~
Ið~ rÞÞ ð5:22Þ
where Du is the phase difference which is due to the difference in the optical path
lengths between the source and the observation point for the two waves

2p
OPD ¼ OPL1 ffi OPL2 ¼ Du ð5:23Þ
k
or

k
Du ¼ OPD: ð5:24Þ
2p
Each fringe period corresponds to a change in the OPD of one wavelength.

Interferometers can be configured to measure small variations in distance, index,
or wavelength. This is the basis of homodyne interferometry.
From Eq. (5.19), assuming that the waves are at different frequencies (equal
intensity parallel-polarized beams), the interference term is modulated at a beat
frequency
Iðx; y; tÞ ¼ I0 f1 þ cos½2pDvt ffi Duðx; yÞg ð5:25Þ
where Dm is the beat frequency. The phase difference Du has the effect of a
spatially varying phase shift of the beat frequency. This is the basis of the het-
erodyne technique used in the distance-measuring interferometers. There must be a
phase relationship between the two sources even though they are at different
frequencies, for example, it is possible to start with a single source, to split it into
two beams, and frequency-shift one beam by Doppler effect. The system can also
work in reverse; the interferometric beat frequency is measured to determine the
velocity of the object producing the Doppler shift.
Fig. 5.7 Experimental setup of Bianchini et al. [34]
5.2.2 Homodyne Dilatometer: Example
For the measurement of the thermal expansion of materials with quite high
monotonic CTE, a very simple instrument may be used. Bianchini et al. [34]
developed a technique based on a Michelson interferometer in which the sample
behaves as a spacer between two identical retroreflectors.
This technique presents the advantages of purely optical measurements. If
corner cube retroreflectors are used, it becomes independent of the mechanical
properties of the sample due to the tilt-compensation of the beams because of the
given sample deformations. This setup can measure CTE of materials with very
different thermal, mechanical and electrical properties.
The experimental setup is shown in Fig. 5.7. The variation of sample length
determines an optical path difference and thus an interference signal (see (5.22,
5.24)).
Corner cube prisms make the reflected beam parallel to the incident beam, and
thus the interferometer is tilt-independent. To minimize systematic errors due to a
change of room temperature, the interferometer plate is stabilized to a temperature
slightly higher than room temperature.
Measurements are possible from 4 to 300 K by using a two-stage 4He cryostat
in which the sample holder is thermally anchored to the liquid helium reservoir.
An optical window enables laser light to enter the cryostat vacuum chamber.
Temperature is monitored by a calibrated carbon thermometer put on the sample
holder.
The interference signal is read by a current to the voltage amplifier. Data are
recorded during the warm-up cycle in two steps: the first, from 4 to 77 K, starts
when liquid helium has evaporated, the process slowed down by the presence of
the 77 K shield; the second step, from 77 K to room temperature, starts when even
liquid nitrogen ends. In these conditions, the system reaches room temperature in
more than 6 h. A slow warming cycle is fundamental to ensure thermal homo-
geneity of the sample and holder (but specific considerations should be taken as a
shape function and thermal conductivity of the sample).
Since the distance between signal zeroes, in terms of sample length variation, k/4,
where k is the laser wavelength, the total expansion of the sample is DL = N k/4,
where N is the number of measured zeros. Bianchini et al. assume k/4 as accuracy of
this system (the system is not able to estimate fractional fringe variations). The total
error can be evaluated of the order of 0.3 % of the total measured dilatation plus
2.4 lm. [34].
This measurement setup has been successfully applied to measure metallic,
amorphous plastic and fiber-reinforced plastic samples in the 4–300 K temperature
range. [35, 36].
This instrument is not able to detect an inversion of the fringe counting, as
happens in the case of a material characterized by nonmonotonic thermal
expansion.
This limitation could be overcome, by an improved interferometer setup pro-
viding two quadrature outputs, either by using polarizing components [37] or a
semi-absorptive beam-splitter. [38] In this way, a bidirectional fringe counting
algorithm could be implemented permitting nanometric accuracy [32], and
removing ambiguity in the sign of CTE. This achievement is possible in a simpler
way with a heterodyne laser system. This particular setup is described in
Sect. (5.2.4).
5.2.3 Heterodyne Dilatometer with Cryogenic Liquids:

Examples
As we saw in the previous paragraph, optical interferometry is an absolute method

to measure length difference, but is not recognized as a highly sensitive technique.
However, the use of a heterodyne method allows one to achieve nanometer range
resolution. The two next paragraphs discuss the setups developed by Okaji et al. in
1995 [39] and by Martelli et al. in 2013 [17].
Fig. 5.8 Diagram of the double-path interferometer [39]
5.2.3.1 Okaji et al.’s Instrument
Okaji et al. developed a cryogenic setup starting from the instrument they made for
middle- and high-temperature dilatometry [28, 33, 40].
Figure 5.8 consists of two polarized beam splitters (PBS-l and PBS-2), a half-
wave plate (HWP), a quarter-wave plate (QWP), a corner cube prism (CCP), a
front reflector (FR) and a reflector (BR). An incident beam from a Zeeman He–Ne
laser (frequency stability of laser a few parts in 10-9) had two slightly different
frequency components, f1 and f2, characterized by linearly polarizing planes
orthogonal to each other (specified by the letters s and p, vertical and horizontal to
the drawing plane, respectively).
The frequency difference between the two components was =100 kHz, which
was used as the beat frequency for the heterodyne interferometer.
The actual optical paths are shown in Fig. 5.9. The optical configuration was
modified from the old version (single plane configuration of the light beams) to a
new one as in Fig. 5.9 (square configuration of the beams) in order to obtain a
Fig. 5.9 Optical paths of the

interferometer, divided into
three steps: [39] a Front
reflector (FR) and the back
reflector (BR). b Corner cube
prism (CCP). c Beam splitters
(PBS-l, PBS-2)
smaller cell. The two components were divided and made parallel by the polarized
beam splitters (PBS-l, PBS-2) and separated by the half-wave plate (HWP). Each
beam was reflected by the front reflector (FR) and the back reflector (BR),
respectively (Fig. 5.9a). The quarter-wave plate (QWP) was oriented in such a way
that the polarized plane of each beam was rotated 90 before and after one
reflection. Accordingly, the two parallel beams were transmitted and reflected
between the original beam splitters (PBS-l, PBS-2), and the front and back
reflectors without losing intensity. These two beams were folded by 180 by the
corner cube prism (CCP), then traced the second transmission/reflection paths
(Fig. 5.9b) and were finally recombined by the same polarized beam splitters
(PBS-l, PBS-2) (Fig. 5.9c). The interferometer guaranteed a self-compensation
mechanism for optical misalignment such as tilt of the specimen system thanks to
a combination of the CCP and the double-path-configuration. The optical align-
ment was easy and a wide tolerance of nonparallelism between the two reflectors
was tolerated.
Equation (5.25) illustrates that in optical heterodyne interferometry, the inter-
ference intensity between the two electric fields depends on frequency and phase
difference. In this experiment, the interference beat signal (from the measure-
Fig. 5.10 Scheme of cryostat and interferometer [39]
beam) and the reference signal (directly from the laser) are both fed to a digital
lock-in amplifier and length change is detected as a relative phase change between
signals.
As in Fig. 5.10, the cryostat is a Helium continuous flow unit. The continuous
flow cryostat is adopted here to give rapid temperature cooling/heating of the
specimen system, very low cryogen consumption and easy optical alignment of the
interferometer. The sample cell is shown in Fig. 5.11.
A silicon diode thermometer is used for controlling the temperature of the cold
head in this system and a Rh-Fe thermometer measures the temperature of the
sample; it is calibrated with an accuracy of better than 5 mK in the temperature
range 5–300 K.
The error on CTE is of the order of 5 %.
Okaji et al. applied this instrument to CTE measurements of fused silica [39]
and copper [41].
Fig. 5.11 Sample cell [39]
5.2.4 Heterodyne Interferometric Dilatometer: Example
5.2.4.1 Martelli et al.’s Dilatometer
As we saw in Sect. 5.2.3, an interferometric dilatometer, simply based on fringe

counting, can be successfully used to measure the CTE of metallic, amorphous
plastic and fiber-reinforced plastic samples [36, 42, 43] in the 4–300 K tempera-
ture range in the case of monotonic thermal expansion.
This heterodyne interferometer not only overcomes such limitations, but
achieves much better resolution.
Here we will describe a recent heterodyne interferometric dilatometer devel-
oped by Martelli et al. in 2013 [2].
The optical setup does not require that the sample surfaces are optical-grade
machined. Corner cube retroreflectors also provide tilt-compensation, thus making
the measurement of CTE possible along a defined axis, even for samples which
deform under thermal stress, like some fiber-reinforced polymers. The sample and
Fig. 5.12 a Schematic view of the optical setup of the interferometer. The two arms, parallel and
displaced by 10 mm, are directed toward the two corner cube retroreflectors (CC-1 and CC-2),
inside the cryostat (CC-1 on top of the sample and CC-2 flush with its bottom, on the support
plate). The two recombined beams travel 3 mm above the incoming beams and are intercepted by
M3 and directed toward the optical pickup. b Side view of the interferometer. The M3 mirror
intercepts the shifted beams. Reprinted with permission [2]
retroreflectors are kept in position by a system that does not require that the sample
should be threaded or bored, thus permitting the characterization of poorly-
machinable materials. The only required constraints on the sample geometry is that
it is available in a rod/bar shape with flat and parallel ends, with length less than
100 mm and diameter at about 5–10 mm.
The dilatometer consists of two subsystems: (1) The room temperature optical
part of the system with a He–Ne laser based parallel-arm heterodyne interfer-
ometer in which the laser, beam splitter and steering optics operate in a thermally
insulated enclosure in order to keep the air inside still and the temperature uni-
form; (2) the sample holder and the retroreflectors are enclosed in a 4He dewar.
The part of the instrument optics that operates at room temperature is shown in
Fig. 2.13: the laser source, a two-frequency Zeeman-stabilized He–Ne laser, and
the electronics are provided by Zygo Corporation [44]. Three plane mirrors (M1,
M2, M3) and a polarizing beam splitter [45] cause beam steering in order to have
the two arms of the interferometer parallel and displaced by exactly 10 mm. The
beam splitter prism is thermally stabilized to better than 10 mK through a Pel-
tier-based closed loop system. The two parallel beams impinge onto the two corner
cube retroreflectors inside of the cryogenic section of the optical system, one on
top of the sample rod, the other, flush with the base of the sample (see Fig. 5.12).
The incoming beams travel *1.5 mm below the center of the retroreflectors, and
are reflected *3 mm above, as shown in Fig. 5.13 (a knife-edge mirror (M3)).
The M3 mirror can intercept the recombined return beams and send them to the
optical pick-up.
The corner cube retroreflectors are mounted in such a way that the corner of the
prism is in contact with the sample rod, as shown in Fig. 5.12. The corner cube
Fig. 5.13 Assembly procedure of the corner cube retroreflector. The prism is glued at its bottom
and inside of a hole in the support in order to have the corner in contact with the sample
prism is glued to the support at its corner in order to minimize systematic errors
due to the thermal expansion of the prism itself, the support and the glue. To avoid
the detachment of the prism from its support because of the large thermal
excursions, a low CTE adhesive is used to assemble the retroreflectors.
The interferometer electronics are interfaced with a personal computer on
which a LabView program directly provides the measurement of the length var-
iation of the sample with a resolution of 2.47 nm (k/256). The longterm stability of
the optical system, at room temperature, is better than 5 nm/24 h.
A detailed schematic of the cryogenic part of the optical setup is shown in
Fig. 2.15. The laser beam enters the cryostat through a 20 mm antireflection
coated optical window which is positioned on a 40 mm diameter flange and also
supports the sample holder (Fig. 5.14, number 3). The sample holder (Fig. 5.14,
number 5) is mounted on a structure supported by the external wall of the 4He
dewar (Fig. 5.14, number 1). This support (Fig. 5.14, number 4) consists of a
100 mm cylindrical tube made of a 50 lm stainless steel sheet. Each end of the
sheet steel tube is clamped on a SS306 flange, without the use of any kind of
adhesive.
This configuration decouples the sample holder from the movements (in par-
ticular, the rotation, which gives a systematic effect on the measurement) of the
low-temperature plate during the cooling or warming of the sample. The thermal
connection of the sample holder to the cold plate is obtained by ultra-thin, highly-
flexible copper wires (12 bundles, of about 50 lm wires each, supplied by
Elschukom) [46].
The sample holder is made of copper and is easily removable from its support in
order to simplify the insertion of the sample. In Fig. 5.15, a detailed schematic of
this part of the apparatus is shown. The sample and the retroreflectors are sup-
ported by nylon threads tensioned by a screw. An aluminum thermal shield, bolted
onto the copper base of the inner sample holder, surrounds the sample, as depicted
in Fig. 5.14, number 4.
CERNOXTM sensors are used as thermometers. A calibrated sensor (CX-1050-
SD-HT-1.4M, provided by Lakeshore) was used to calibrate other four sensors
Fig. 5.14 Exploded view of the cryostat system: (1) cryostat case (2) dewar cold plate (3) flange
bolted on the cryostat case and optical window holder (4) external sample holder (5) internal
sample holder with aluminum thermal shield installed (6) system for fixing the external sample
holder to the window. Reprinted with permission [2]
Fig. 5.15 Scheme of the internal sample holder structure. The nylon threads make the
positioning of the sample possible without using glues or screws. Reprinted with permission [2]
placed in the experimental setup. The thermometers determine temperature with an

uncertainty better than 1 %.
An encapsulated sensor has been placed on the backside of the sample holder,
outside the thermal shield, in order to monitor the temperature of the sample holder
Fig. 5.16 a Measured thermal expansion of the SRM-731 Borosilicate sample. b Difference
between measured and NIST certified data. Dashed line is before applying the correction coming
from the calibration procedure, and the solid line is after the correction. Dotted lines represent 1-
sigma uncertainty
itself; two thermometers have been placed on the sample, respectively close to the
sample holder and close to the top retroreflector. This configuration allows a more
precise determination of the average temperature of the sample and of the thermal
gradient during the measurement. An additional encapsulated sensor has been
placed on the cold plate of the cryostat in order to check the good thermal contact
between the sample holder and the plate through the copper wire thermal link.
A typical CTE measurement is carried out first by cooling the system down to
the lowest temperature required (liquid helium or nitrogen), letting it thermalize
for several hours. Thermal expansion data is then acquired during the warm-up
phase, thus obtaining sample length variation as a function of the increasing
temperature. In this way, a slow and regular temperature variation is easily
obtained. The duration of the warm-up phase is about 24 h.
In order to validate the instrument, the thermal expansion coefficient of a SRM-
731-L1 Borosilicate Glass standard reference (77–300 K, srdata.nist.gov) material
was carried out.
The CTE as a function of T is then obtained as a third degree polynomial fit to
the experimental data points. DL/L values were obtained from CTE through
integration. 1-sigma uncertainties are respectively 3 9 10-8 for CTE and
8 9 10-6 for DL = L [47].
In Fig. 5.16a, the measured thermal expansion curve of the SRM-731 sample is
shown. In Fig. 5.16b, the difference between measured and NIST data is shown,
with and without the correction provided by the calibration measurement
(respectively, solid and dashed lines), along with the NIST 1-sigma uncertainty
(dotted lines). The uncertainty on data is 4 9 10-6.
The decrease in the difference between measured and NIST data near room
temperature is due to the fact that approaching room temperature, the warm-up rate
slows considerably.
By means of this instrument, the CTE of AISI 420, between 20 and 293 K, was
recently measured.
5.2.5 Heterodyne Dilatometer with Mechanical Coolers:

Examples
In the two next paragraphs, we describe two measurement setups developed by

JPL laboratory [48] and the Okaji group [49]. The main difference with the
instrument described in Sect. 5.2.4 is the use of a mechanical cooler instead of a
nitrogen-helium bath. The use of a mechanical cooler allows for a gradual cooling
and warming, and a simple assembling procedure between optical and cryogenic
parts, resulting in a compact setup.
5.2.5.1 Instrument of Karlmann et al.
The James Webb Space Telescope (JWST) materials working group at NASA Jet
Propulsion Laboratory developed a new cryogenic dilatometer to carry out mea-
surements of CTE on the candidate materials for manufacturing the mirror sub-
strate and back structure of the telescope (6.5 m diameter orbiting infrared
telescope, located in L2 orbit). The thermomechanical properties of the mirror
substrate and the composite back structure must be known with an extremely high
precision because the requirement of primary mirror total WFE (wavefront error)
is about 60 nm and the working temperature is about 30 K.
The components of the system are: YAG laser source (532 nm with a power of
100 mW), and the AOMs (Acoustic optics modulators) and signal processing
electronics and data acquisition system.
The output beam from the laser is split in two beams, namely, the measurement
beam and local oscillator beam. Then, AOMs shift the measurement beam by
80.016 MHz and the local oscillator beam by 80 MHz. The frequency difference
of 16 kHz is the heterodyne frequency. A Gifford–McMahon closed-cycle helium
cryocooler is used. The main problem of using this cryocooler is the vibration
isolation since the vacuum chamber is mounted directly in contact with the cooler.
For this reason, commercially available vibration isolation bellows were installed
between the cooler and the vacuum chamber.
The scheme of the interferometer is shown in Fig. 5.17.
5.2.5.2 Okaji Group Setup
As we have seen in Sect. 5.2.3.1, the Okaji group developed a high-sensitivity

interferometric dilatometer (working in the 4–300 K temperature range). The
cooling of the instrument was due to a liquid helium continuous flow cryostat. The
improved dilatometer used a mechanical cooler (V204SC, Daikin Industries).
The schematic design is reported in Fig. 5.18. To reduce the effects of mechanical
vibration, the cryogenic refrigerator is connected to the cryostat by a vacuum
bellow. The optical components are arranged as in their previous work [39].
Fig. 5.17 Setup of the cryogenic interferometer
In this configuration, the experimental error in a measurements was reduced to

1 10-6 K-1.
5.3 Very Low Temperature Thermal Expansion
As we have seen in Sect. 5.2, CTE becomes very low below 4 K. The measure-
ment of CTE is possible with an extremely high-resolution apparatus and large
samples. An example of this type of measurement is reported in [50] where a
SQUID is used as a sensing element of the contraction. In this instrument, a
resolution of 2 10-5 nm is achieved and data for Cu were obtained in the
0.2–1.9 K temperature range.
Another example is reported in Ref. [51] where the measurement of the thermal
expansion of a 2 m long Al–Mg alloy bar is described. The measured values were:
a = [(10.9 + 0.4)T + (1.3 + 0.1)T3] 9 10-10 K-1 for the normal state of con-
duction in the temperature interval 0.9 \ T \ 2 K and a = [(-2.45 ± 0.60) +
Fig. 5.18 Scheme of the dilatometer with cryogenic refrigerator [49]
(-10.68 ± 1.24)T + (0.13 ± 0.01)T3] 9 10-9K-1 for the superconducting state

in the interval 0.3 \ T \ 0.8 K. Also, in this case, a SQUID system was used [52].
The knowledge of the specific heat of the alloy was needed [53] to evaluate CTE.
References
1. Kanagaraj, S., Pattanayak, S.: Measurement of the thermal expansion of metal and FRPs.
Cryogenics 43(7), 399–424 (2003)
2. Martelli, V., Bianchini, G., Natale, E., Scarpellini, D., Ventura, G.: A novel interferometric
dilatometer in the 4–300 K temperature range: thermal expansion coefficient of SRM-731
borosilicate glass and stainless steel-304. Meas. Sci. Technol. 24(10), 105203 (2013)
3. Bijl, D., Pullan, H.: A new method for measuring the thermal expansion of solids at low
temperatures; the thermal expansion of copper and aluminium and the Grüneisen rule.
Physica 21(1), 285–298 (1954)
4. Sao, G., Tiwary, H.: Thermal expansion of poly (vinylidene fluoride) films. J. Appl. Phys.
53(4), 3040–3043 (1982)
5. Rao, K., Jeyasri, M.: Measurement of linear thermal expansion of solids by a capacitance
method. Indian J. Pure Appl. Phys. 15, 437–440 (1977)
6. Tong, H., Hsuen, H., Saenger, K., Su, G.: Thickness-direction coefficient of thermal
expansion measurement of thin polymer films. Rev. Sci. Instrum. 62(2), 422–430 (1991)
7. White, G.: Measurement of thermal expansion at low temperatures. Cryogenics 1(3),
151–158 (1961)
References 119
8. Pott, R., Schefzyk, R.: Apparatus for measuring the thermal expansion of solids between 1.5
and 380 K. J. Phys. E: Sci. Instrum. 16(5), 444 (1983)
9. Kroeger, F., Swenson, C.: Absolute linear thermal-expansion measurements on copper and
aluminum from 5 to 320 K. J. Appl. Phys. 48(3), 853–864 (1977)
10. Subrahmanyam, H., Subramanyam, S.: Accurate measurement of thermal expansion of solids
between 77 K and 350 K by 3-terminal capacitance method. Pramana 27(5), 647–660 (1986)
11. Neumeier, J., Bollinger, R., Timmins, G., Lane, C., Krogstad, R., Macaluso, J.: Capacitive-
based dilatometer cell constructed of fused quartz for measuring the thermal expansion of
solids. Rev. Sci. Instrum. 79(3), 033903–033908 (2008)
12. Rotter, M., Muller, H., Gratz, E., Doerr, M., Loewenhaupt, M.: A miniature capacitance
dilatometer for thermal expansion and magnetostriction. Rev. Sci. Instrum. 69(7), 2742–2746
(1998)
13. Roth, P., Gmelin, E.: A capacitance displacement sensor with elastic diaphragm. Rev. Sci.
Instrum. 63(3), 2051–2053 (1992)
14. Tokiwa, Y., Grüheit, S., Jeevan, H., Stingl, C., Gegenwart, P.: Low-temperature
antiferromagnetic ordering in the heavy-fermion metal YbPd. J. Phys: Conf. Ser. 273,
012062 (IOP Publishing) (2011)
15. Schafer, D., Thomas, G., Wudl, F.: High-resolution thermal-expansion measurements of
tetrathiafulvalenetetracyanoquinodimethane (TTF-TCNQ). Phys. Rev. B 12(12), 5532 (1975)
16. McCammon, R., Work, R.: Measurement of the dielectric properties and thermal expansion
of polymers from ambient to liquid helium temperatures. Rev. Sci. Instrum. 36(8),
1169–1173 (1965)
17. Kanagaraj, S., Pattanayak, S.: Simultaneous measurements of thermal expansion and thermal
conductivity of FRPs by employing a hybrid measuring head on a GM refrigerator.
Cryogenics 43(8), 451–458 (2003)
18. García-Moreno, O., Fernández, A., Khainakov, S., Torrecillas, R.: Negative thermal
expansion of lithium aluminosilicate ceramics at cryogenic temperatures. Scripta Mater.
63(2), 170–173 (2010)
19. Keesom, W.H., Andronikashvili, E., Lifshits, E.M.: Helium. Elsevier, Amsterdam (1942)
20. Maxwell, J.C.: Lehrbuch der Electricität und des Magnetismus, vol. 1. J. Springer, Berlin
(1883)
21. Hartshorn, L.: Radio-Frequency Measurements by Bridge and Resonance Methods, vol. 10.
Chapman and Hall, London (1940)
22. Jones, R.: Some developments and applications of the optical lever. J. Sci. Instrum. 38(2), 37
(1961)
23. James, J., Spittle, J., Brown, S., Evans, R.: A review of measurement techniques for the
thermal expansion coefficient of metals and alloys at elevated temperatures. Meas. Sci.
Technol. 12(3), R1 (2001)
24. Bennett, S.: An absolute interferometric dilatometer. J. Phys. E: Sci. Instrum. 10(5), 525
(1977)
25. Imai, H., Bates, W.: Measurement of the linear thermal expansion coefficient of thin
specimens. J. Phys. E: Sci. Instrum. 14(7), 883 (1981)
26. Schödel, R.: Ultra-high accuracy thermal expansion measurements with PTB’s precision
interferometer. Meas. Sci. Technol. 19(8), 084003 (2008)
27. Cordero, J., Heinrich, T., Schuldt, T., Gohlke, M., Lucarelli, S., Weise, D., Johann, U.,
Braxmaier, C.: Interferometry based high-precision dilatometry for dimensional
characterization of highly stable materials. Meas. Sci. Technol. 20(9), 095301 (2009)
28. Okaji, M., Imai, H.: A practical measurement system for the accurate determination of linear
thermal expansion coefficients. J. Phys. E: Sci. Instrum. 17(8), 669 (1984)
29. Uchil, J., Mohanchandra, K., Ganesh Kumara, K., Mahesh, K., P Murali, T.: Thermal
expansion in various phases of nitinol using TMA. Phys. B: Condens. Matter 270(3),
289–297 (1999)
30. Vijay, A.: Temperature dependence of elastic constants and volume expansion for cubic and
non-cubic minerals. Phys. B 349(1), 62–70 (2004)
31. Singh, K., Gupta, B.: A simple approach to analyse the thermal expansion in minerals under
the effect of high temperature. Phys. B 334(3), 266–271 (2003)
32. Birch, K.: An automatic absolute interferometric dilatometer. J. Phys. E: Sci. Instrum. 20(11),
1387 (1987)
33. Okaji, M., Inai, H.: A high-temperature dilatometer using optical heterodyne interferometry.
J. Phys. E: Sci. Instrum. 20, 887–891 (1987)
34. Bianchini, G., Barucci, M., Del Rosso, T., Pasca, E., Ventura, G.: Interferometric dilatometer
for thermal expansion coefficient determination in the 4–300 K range. Meas. Sci. Technol.
17(4), 689 (2006)
35. Ventura, G., Bianchini, G., Gottardi, E., Peroni, I., Peruzzi, A.: Thermal expansion and
thermal conductivity of Torlon at low temperatures. Cryogenics 39(5), 481–484 (1999)
36. Barucci, M., Bianchini, G., Del Rosso, T., Gottardi, E., Peroni, I., Ventura, G.: Thermal
expansion and thermal conductivity of glass-fibre reinforced nylon at low temperature.
Cryogenics 40(7), 465–467 (2000)
37. Greco, V., Molesini, G., Quercioli, F.: Accurate polarization interferometer. Rev. Sci.
Instrum. 66(7), 3729–3734 (1995)
38. Raine, K., Downs, M.: Beam-splitter coatings for producing phase quadrature interferometer
outputs. J. Mod. Opt. 25(7), 549–558 (1978)
39. Okaji, M., Yamada, N., Nara, K., Kato, H.: Laser interferometric dilatometer at low
temperatures: application to fused silica SRM 739. Cryogenics 35(12), 887–891 (1995)
40. Okaji, M., Birch, K.: Intercomparison of Interferometric Dilatometers at NRLM and NPL.
Metrologia 28(1), 27 (1991)
41. Okaji, M., Yamada, N., Kato, H., Nara, K.: Measurements of linear thermal expansion
coefficients of copper SRM 736 and some commercially available coppers in the temperature
range 20–300 K by means of an absolute interferometric dilatometer. Cryogenics 37(5),
251–254 (1997)
43. Gottardi, E., Bianchini, G., Peroni, I., Peruzzi, A., Ventura, G.: Thermal conductivity of
polyetheretherketone at low temperatures. In: Proceedings of Tempmeco, Berlin (2001)
44. Corporation, Z.: Laurel Brook Road, Middlefield, Connecticut 06455–0448
45. GmbH, E.O.: Zur Giesserei 19–27, 76227 Karlsruhe, Germany
46. Elschukom, E., GmbH, Gewerbestrasse 87, D-98669 Veilsdorf
47. Hahn, T.A.: Thermal expansion of copper from 20 to 800 K—standard reference material
736. J. Appl. Phys. 41(13), 5096–5101 (1970)
48. Karlmann, P.B., Dudik, M.J., Halverson, P.G., Levin, M., Marcin, M., Peters, R.D., Shaklan,
S., van Buren, D.: JLP Technical Report. California Institute of Technology, (1992)
49. Yamada, N., Okaji, M.: Development of a low-temperature laser interferometric dilatometer
using a cryogenic refrigerator. High Temp. High Pressures 32(2), 199–206 (2000)
50. Ackerman, D., Anderson, A.: Dilatometry at low temperatures. Rev. Sci. Instrum. 53(11),
1657–1660 (1982)
51. Bassan, M., Buonomo, B., Cavallari, G., Coccia, E., D’Antonio, S., Fafone, V., Foggetta, L.,
Ligi, C., Marini, A., Mazzitelli, G.: Measurement of the thermal expansion coefficient of AN
Al-Mg alloy at ultra-low temperatures. Int. J. Mod. Phys. B 27(22), 1350119–1350131 (2013)
52. Barucci, M., Bassan, M., Buonomo, B., Cavallari, G., Coccia, E., D’Antonio, S., Fafone, V.,
Ligi, C., Lolli, L., Marini, A.: Experimental study of high energy electron interactions in a
superconducting aluminum alloy resonant bar. Phys. Lett. A 373(21), 1801–1806 (2009)
53. Barucci, M., Ligi, C., Lolli, L., Marini, A., Martelli, V., Risegari, L., Ventura, G.: Very low
temperature specific heat of Al 5056. Phys. B 405(6), 1452–1454 (2010)
Chapter 6
Data of Thermal Expansion
Abstract In this chapter useful tables containing data of thermal contraction

below room temperature of several materials, most of them used in cryogenic
applications, are reported together with the coefficient of thermal expansion at
293 K. The values of thermal contraction are referred at the length of the sample at
293 K.
Here we present a table containing data of thermal contraction of several materials,

most of them used in cryogenic applications. The values of thermal contraction are
referred at L293 i.e. the length of the sample at 293 K (Table 6.1).

Table 6.1 Linear thermal contraction (DL/L, %) from 293 K and a (10-6K-1) at 293 K
122
DL/L (%, referred References 4 (K) 40 (K) 77 (K) 100 (K) 150 (K) 200 (K) 250 (K) a (10-6
to 293 K) J-1)
293 (K)
Metals
Ag [1] 4.13 9 10-1 4.05 9 10-1 3.7 9 10-1 3.39 9 10-1 2.59 9 10-1 1.73 9 10-1 8.2 9 10-2 18.5
Al [1] 4.15 9 10-1 4.13 9 10-1 3.93 9 10-1 3.7 9 10-1 2.95 9 10-1 2.01 9 10-1 9.7 9 10-2 23.1
Au [1] 3.24 9 10-1 3.13 9 10-1 2.81 9 10-1 2.56 9 10-1 1.95 9 10-1 1.29 9 10-1 6.1 9 10-2 14.1
Be [1] 1.31 9 10-1 1.31 9 10-1 1.3 9 10-1 1.28 9 10-1 1.15 9 10-1 8.7 9 10-2 4.5 9 10-2 11.3
Cu [1] 3.24 9 10-1 3.22 9 10-1 3.02 9 10-1 2.82 9 10-1 2.21 9 10-1 1.48 9 10-1 7 9 10-2 16.7
Fe [1] 1.98 9 10-1 1.97 9 10-1 1.9 9 10-1 1.81 9 10-1 1.48 9 10-1 1.02 9 10-1 4.9 9 10-2 11.6
Hg (data [1] 8.43 9 10-1 7.88 9 10-1 7.88 9 10-1 5.92 9 10-1 3.96 9 10-1 1.76 9 10-1 – –
referred to 234 K)
In [1] 7.06 9 10-1 6.76 9 10-1 6.02 9 10-1 5.49 9 10-1 4.21 9 10-1 2.82 9 10-1 1.35 9 10-1 32
Mo [1] 9.5 9 10-2 9.4 9 10-2 9 9 10-2 8.4 9 10-2 6.7 9 10-2 4.6 9 10-2 2.2 9 10-2 4.8
Nb [2] – 1.49 9 10-1 1.38 9 10-1 1.28 9 10-1 9.90 9 10-2 6.63 9 10-2 3.11 9 10-1 7.2
Nb [1] 1.43 9 10-1 1.41 9 10-1 1.3 9 10-1 1.21 9 10-1 9.4 9 10-2 6.3 9 10-2 3 9 10-2 7.3
Ni [1] 2.24 9 10-1 2.23 9 10-1 2.12 9 10-1 2.01 9 10-1 1.62 9 10-1 0.111 0.053 13.4
Pd [3] – 4.0 9 10-1 3.7 9 10-1 3.4 9 10-1 2.6 9 10-1 1.7 9 10-1 8.2 9 10-2 19
Pb [1] 7.08 9 10-1 6.67 9 10-1 5.78 9 10-1 5.28 9 10-1 3.98 9 10-1 2.63 9 10-1 1.24 9 10-1 29
Ta [1] 1.43 9 10-1 1.41 9 10-1 1.28 9 10-1 1.17 9 10-1 8.9 9 10-2 5.9 9 10-2 2.8 9 10-2 6.6
Sn [1] 4.47 9 10-1 4.33 9 10-1 3.89 9 10-1 3.56 9 10-1 2.72 9 10-1 1.83 9 10-1 8.6 9 10-2 20.5
Ti [1] 1.51 9 10-1 1.5 9 10-1 1.43 9 10-1 1.34 9 10-1 1.07 9 10-1 7.3 9 10-2 3.5 9 10-2 8.3
W [1] 8.6 9 10-2 8.5 9 10-2 8 9 10-2 7.5 9 10-2 5.9 9 10-2 4 9 10-2 1.9 9 10-2 4.5
Alloys
Al 2024 [4] 3.96 9 10-1 3.94 9 10-1 3.72 9 10-1 3.51 9 10-1 2.78 9 10-1 1.90 9 10-1 – 21.2
Al 5083 [4] 4.15 9 10-1 4.13 9 10-1 3.90 9 10-1 3.68 9 10-1 2.94 9 10-1 2.01 9 10-1 – 22.8
Al-6061-T6 [1] 4.14 9 10-1 4.12 9 10-1 3.89 9 10-1 3.65 9 10-1 2.95 9 10-1 2.03 9 10-1 9.7 9 10-2 22.5
(continued)
6 Data of Thermal Expansion
Table 6.1 (continued)
to 293 K) J-1)
293 (K)
BeCu [2] – 3.1 9 10-1 2.9 9 10-1 2.7 9 10-1 2.1 9 10-1 1.5 9 10-1 6.9 9 10-2 16
Brass (65 %Cu–35 %Zn) [1] 3.84 9 10-1 3.8 9 10-1 3.53 9 10-1 3.26 9 10-1 2.53 9 10-1 1.69 9 10-1 8 9 10-2 19.1
(yellow brass)
Constantan (50 %Cu–50 %Ni) [1] – 2.64 9 10-1 2.49 9 10-1 2.32 9 10-1 1.83 9 10-1 1.24 9 10-1 4.3 9 10-2 13.8
Berylco 25 [1] 3.16 9 10-1 3.15 9 10-1 2.98 9 10-1 2.77 9 10-1 2.19 9 10-1 1.51 9 10-1 7.4 9 10-2 18.1
(Cu + 2 %Be + 0.3 %Co)
Fe1.5Mn1.5Si [5] – – 2.3 9 10-1 2.1 9 10-1 1.7 9 10-1 1.2 9 10-1 5.7 9 10-2 13
Fe–9 %Ni [1] 1.95 9 10-1 1.93 9 10-1 1.88 9 10-1 1.8 9 10-1 1.46 9 10-1 1 9 10-1 4.9 9 10-2 11.5
Fe64Ni36 [4] 4.5 9 10-2 4.8 9 10-2 4.8 9 10-2 4.5 9 10-2 3.0 9 10-2 2.0 9 10-2 – 1
Hastelloy [1] 2.18 9 10-1 2.16 9 10-1 2.04 9 10-1 1.93 9 10-1 1.5 9 10-1 0.105 4.7 9 10-2 10.9
Inconel [1] 2.38 9 10-1 2.36 9 10-1 2.24 9 10-1 2.11 9 10-1 1.67 9 10-1 1.14 9 10-1 5.5 9 10-2 13
Invar [1] – 4 9 10-2 3.8 9 10-2 3.6 9 10-2 2.5 9 10-2 1.6 9 10-2 9 9 10-3 3
Invar 36 [6] – 3.89 9 10-2 3.85 9 10-2 3.59 9 10-2 2.6 9 10-2 1.60 9 10-2 8.81 9 10-3 3
LaRu4P12 [7] – 1.0 9 10-1 9.6 9 10-2 9.1 9 10-2 7.6 9 10-2 5.4 9 10-2 2.6 9 10-2 5.7
Mn steel JK2LB [8] – 2.0 9 10-1 2.0 9 10-1 1.9 9 10-1 1.6 9 10-1 1.1 9 10-1 6.2 9 10-2 17
TD Nichel (Ni0.98(ThO2)0.02) [4] – 2.2 9 10-1 2.2 9 10-1 2.04 9 10-1 1.63 9 10-1 1.11 9 10-1 5.39 9 10-2 13.2
PrRu4P12 [7] – 1.3 9 10-1 1.2 9 10-1 1.1 9 10-1 8.9 9 10-2 6.3 9 10-2 3.1 9 10-2 7.3
50 %Pb–50 %Sn solder [1] 5.14 9 10-1 5.1 9 10-1 4.8 9 10-1 4.47 9 10-1 3.43 9 10-1 2.29 9 10-1 1.08 9 10-1 23.4
Stainless steel (AISI 304) [1] 2.96 9 10-1 2.96 9 10-1 2.81 9 10-1 2.61 9 10-1 2.06 9 10-1 1.39 9 10-1 6.6 9 10-2 15.1
Stainless steel (AISI 310) [1] – – – 2.37 9 10-1 1.87 9 10-1 1.27 9 10-1 6.1 9 10-2 14.5
Stainless (AISI 306) [1] 2.97 9 10-1 2.96 9 10-1 2.79 9 10-1 2.59 9 10-1 2.01 9 10-1 21.36 6.5 9 10-2 15.2
9 10-1
Ti–6 % Al–4 %V [1] 1.73 9 10-1 1.71 9 10-1 1.63 9 10-1 1.54 9 10-1 1.18 9 10-1 7.8 9 10-2 3.6 9 10-2 8
(continued)
123
124
to 293 K) J-1)
293 (K)
Superconductors
Bi-2212 a, b-axes [1] 1.52 9 10-1 1.5 9 10-1 1.39 9 10-1 1.32 9 10-1 1.06 9 10-1 7.4 9 10-2 3.6 9 10-2 8.3
Bi-2212 c-axis [1] 2.95 9 10-1 2.89 9 10-1 2.66 9 10-1 2.5 9 10-1 1.99 9 10-1 1.36 9 10-1 6.4 9 10-2 15.1
Bi-2223 a, b-axes [1] 1.5 9 10-1 1.5 9 10-1 1.4 9 10-1 1.3 9 10-1 1.1 9 10-1 7 9 10-2 4 9 10-2 8.3
Bi-2223 c-axis [1] 3.0 9 10-1 2.9 9 10-1 2.7 9 10-1 2.5 9 10-1 2.0 9 10-1 1.4 9 10-1 6 9 10-2 15
Bi (2223)/Ag tape ([1 cool [1] – 3.1 9 10-1 3.0 9 10-1 2.8 9 10-1 2.2 9 10-1 1.5 9 10-1 7 9 10-2 13
down)
Bi-2223/61 % Ag-alloy tape [1] – – 2.4 9 10-1 – – – – –
Nb3Sn [1] 1.6 9 10-1 1.6 9 10-1 1.4 9 10-1 1.3 9 10-1 9.5 9 10-2 6.5 9 10-2 3 9 10-2 7.6
Nb3Sn(10vol%)/Cu wire [1] 3 9 10-1 2.8 9 10-1 – – – – – –
Nb–45 %Ti [1] 1.88 9 10-1 1.84 9 10-1 1.69 9 10-1 1.56 9 10-1 1.17 9 10-1 7.8 9 10-2 3.8 9 10-2 8.2
Nb–Ti/Cu wire [1] 2.65 9 10-1 2.62 9 10-1 2.47 9 10-1 2.31 9 10-1 1.79 9 10-1 1.17 9 10-1 5.4 9 10-2 12.5
YBCO a-axis [1] – – 1.2 9 10-1 1.2 9 10-1 1 9 10-1 7 9 10-2 4 9 10-2 7.4
YBCO b-axis [1] – – 1.6 9 10-1 1.5 9 10-1 1.3 9 10-1 1 9 10-1 5 9 10-2 9.6
YBCO c-axis [1] – – 3.4 9 10-1 3.3 9 10-1 0.25 0.17 0.09 17.7
YBCO-123(ab) [4] 1.5 9 10-1 1.5 9 10-1 – 1.4 9 10-1 1.15 9 10-1 8 9 10-2 – 10
YBCO-123(c) [4] 3.6 9 10-1 3.5 9 10-1 3.3 9 10-1 3.1 9 10-1 2.5 9 10-1 1.7 9 10-1 – 17
Polymers
Araldite [1] 1.06 1.02 9.35 9 10-1 8.8 9 10-1 7.1 9 10-1 5.05 9 10-1 – 60
Epoxy [1] 1.16 1.11 1.028 9.59 9 10-1 7.78 9 10-1 5.5 9 10-1 2.77 9 10-1 66
Epoxy EPON 815 [9] 1.19 1.16 1.07 1.00 8.3 9 10-1 6.1 9 10-1 3.2 9 10-1 83
Epoxy (Stycast 2850FTTM) [1] 4.4 9 10-1 4.3 9 10-1 4 9 10-1 3.8 9 10-1 3.2 9 10-1 2.25 9 10-1 1.2 9 10-1 28
CTFE (TeflonTM) [1] 1.135 1.07 9.71 9 10-1 9 9 10-1 7.25 9 10-1 5.17 9 10-1 2.69 9 10-1 67
TFE (TeflonTM) [1] 2.14 2.06 1.941 1.85 1.6 1.24 7.5 9 10-1 250
PMMA (PlexiglasTM) [1] 1.22 1.16 1.059 9.9 9 10-1 8.2 9 10-1 5.9 9 10-1 3.05 9 10-1 75
(continued)
to 293 K) J-1)
293 (K)
Polyamide (NylonTM) [1] 1.389 1.352 1.256 1.172 9.46 9 10-1 6.73 9 10-1 3.39 9 10-1 80
Polyimide (KaptonTM) [1] 4.4 9 10-1 4.4 9 10-1 4.3 9 10-1 4.1 9 10-1 3.6 9 10-1 2.9 9 10-1 1.6 9 10-1 46
Polyamide-imide (Torlon) [10] 4.5 9 10-1 4.4 9 10-1 4.0 9 10-1 3.6 9 10-1 2.8 9 10-1 2.0 9 10-1 9.8 9 10-2 24
Polyamide-imide (Torlon)* [1] 4.48 9 10-1 4.34 9 10-1 3.87 9 10-1 3.58 9 10-1 2.79 9 10-1 1.91 9 10-1 – 24
Polypropilene [11] 1.13 1.10 1.02 9.6 9 10-1 8.0 9 10-1 6.0 9 10-1 3.3 9 10-1 96
Composites, ceramics
and non-metals
Al N (parallel a-axis) [1] – – 3.2 9 10-2 3.1 9 10-2 2.8 9 10-2 2 9 10-2 1.1 9 10-2 3.7
Al N (parallel c-axis) [1] – – 2.5 9 10-2 2.5 9 10-2 2.2 9 10-2 1.7 9 10-2 9 9 10-3 3
Diamond [1] 2.4 9 10-2 2.4 9 10-2 2.4 9 10-2 2.4 9 10-2 2.3 9 10-2 1.9 9 10-2 1.1 9 10-2 1
G-10CR epoxy/glass (parallel [1] 2.41 9 10-1 2.34 9 10-1 2.13 9 10-1 1.97 9 10-1 1.57 9 10-1 1.08 9 10-1 5.2 9 10-2 12.5
glass fibers)
G10CR warp [12] 2.5 9 10-1 2.4 9 10-1 2.1 9 10-1 2.0 9 10-1 1.6 9 10-1 1.1 9 10-1 5.4 9 10-2 13
G-10CR epoxy/glass (normal) [1] 7.06 9 10-1 6.9 9 10-1 6.42 9 10-1 6.03 9 10-1 4.91 9 10-1 3.46 9 10-1 1.71 9 10-1 41
G-10CR (normal) [12] 7.1 9 10-1 6.9 9 10-1 6.4 9 10-1 6.0 9 10-1 4.9 9 10-1 3.5 9 10-1 1.7 9 10-1 43
Glass (PyrexTM) [1] 5.5 9 10-2 5.7 9 10-2 5.4 9 10-2 5 9 10-2 4 9 10-2 2.7 9 10-2 1.3 9 10-2 3
Glass fiber reinforced nylon [13] 5.9 9 10-1 5.7 9 10-1 5.3 9 10-1 4.9 9 10-1 3.8 9 10-1 2.5 9 10-1 1.2 9 10-1 28
Graphite (parallel to extrusion [14] – 7.3 9 10-2 6.6 9 10-2 6.0 9 10-2 4.7 9 10-2 3.3 9 10-2 1.6 9 10-2 3.4
direction)
Graphite (perpendicular to [14] – 2.9 9 10-2 2.5 9 10-2 2.3 9 10-2 1.7 9 10-2 1.1 9 10-2 5.5 9 10-3 1.4
extrusion direction)
MgO [1] 1.39 9 10-1 1.39 9 10-1 1.37 9 10-1 1.33 9 10-1 1.14 9 10-1 8.3 9 10-2 4.2 9 10-2 10.2
Quartz (parallel optic axis) [1] – – – 1.04 9 10-1 8.5 9 10-2 6.1 9 10-2 3 9 10-2 7.5
Sapphire [14] – 7.9 9 10-2 7.7 9 10-2 7.5 9 10-2 6.5 9 10-2 4.8 9 10-2 2.4 9 10-2 6.4
Sapphire (parallel c-axis) [1] – 7.9 9 10-2 7.8 9 10-2 7.5 9 10-2 6.6 9 10-2 4.8 9 10-2 2.5 9 10-2 5.4
Si [1] 2.2 9 10-2 2.2 9 10-2 2.3 9 10-2 2.4 9 10-2 2.4 9 10-2 1.9 9 10-2 1 9 10-2 2.32
125
(continued)
126
to 293 K) J-1)
293 (K)
SiC100 (sintered) [15] – 2.3 9 10-2 2.3 9 10-2 2.3 9 10-2 2.1 9 10-2 1.6 9 10-2 9.3 9 10-3 2.7
a-SiC (polycrystalline) [1] – – 3 9 10-2 3 9 10-2 2.9 9 10-2 2.4 9 10-2 1.3 9 10-2 3.7
Silica glass [1] -8 9 10-3 -5 9 10-3 -2 9 10-3 -1 9 10-4 2 9 10-3 2 9 10-3 2 9 10-3 0.4
Stycast [16] 4.4 9 10-1 4.3 9 10-1 4.0 9 10-1 3.7 9 10-1 3.1 9 10-1 2.2 9 10-1 1.1 9 10-1 24
ZrO2 [4] 1.31 9 10-1 1.30 9 10-1 1.24 9 10-1 1.18 9 10-1 9.8 9 10-2 6.8 9 10-2 – 8.0
References 127
References

2. Radcliffe, W., Gallop, J., Dominique, J.: A microwave method for thermal expansion
measurement. J. Phys. E: Sci. Instrum. 16(12), 1200 (1983)
3. Waterhouse, N., Yates, B.: The interferometric measurement of the thermal expansion of
silver and palladium at low temperatures. Cryogenics 8(5), 267–271 (1968)
5. Mukherjee, G., Bansal, C., Chatterjee, A.: Thermal expansion study of Fe–Mn–Si alloys.
Phys. B 254(3), 223–233 (1998)
6. Clark, A.: Low temperature thermal expansion of some metallic alloys. Cryogenics 8(5),
282–289 (1968)
7. Matsuhira, K., Takikawa, T., Sakakibara, T., Sekine, C., Shirotani, I.: Thermal expansion of
PrRu4P12. Phys. B 281, 298–299 (2000)
8. Lu, J., Walsh, R., Han, K.: Low temperature physical properties of a high Mn austenitic steel
JK2LB. Cryogenics 49(3), 133–137 (2009)
9. Hamilton, W., Greene, D., Davidson, D.: Thermal expansion of epoxies between 2 and
300 K. Rev. Sci. Instrum. 39(5), 645–648 (1968)
12. Clark, A., Fujii, G., Ranney, M.: The thermal expansion of several materials for
superconducting magnets. IEEE Trans. Magn. 17(5), 2316–2319 (1981)
Cryogenics 40(7), 465–467 (2000)
14. Arp, V., Wilson, J., Winrich, L., Sikora, P.: Thermal expansion of some engineering
materials from 20 to 293 K. Cryogenics 2(4), 230–235 (1962)
15. Enya, K., Yamada, N., Onaka, T., Nakagawa, T., Kaneda, H., Hirabayashi, M., Toulemont,
Y., Castel, D., Kanai, Y., Fujishiro, N.: High-precision CTE measurement of SiC-100 for
cryogenic space telescopes. Publ. Astron. Soc. Pac. 119(855), 583–589 (2007)
16. Swenson, C.: Linear thermal expansivity (1.5–300 K) and heat capacity (1.2–90 K) of
Stycast 2850FT. Rev. Sci. Instrum. 68(2), 1312–1315 (1997)
Part III
Thermal Conductivity
Main Symbols
kF Radius of the fermi sphere

EF Fermi energy
vF Fermi velocity
TF Fermi temperature
kB Boltzmann constant
h Plank constant
h h/2p
e Electron charge
nC Number of electrical carriers
ne Number of conduction electrons
H Magnetic field
Eg Energy gap of semiconductors
j Thermal conductivity
w Thermal diffusivity
g Mass density
\k[ Mean free path
c Specific heat
A Area
Q Heat
\s[ Mean scattering time
r Electrical conductivity
\m[ Mean atomic mass
a Mean atomic space
hD Debye temperature
L0 Free electron lorentz number
je Electron thermal conductivity
fc Grade of crystallinity
e Dielectric constant
e’ Real part of the dielectric constant
e’’ Imaginary part of the dielectric constant
130 Part III Thermal Conductivity
G Thermal conductance
B Magnetic induction
D Electric displacement
J Current density
Chapter 7
Electrical and Thermal Conductivity
Abstract After a Sect. 1.1 devoted to electrical conductivity and a section that
deals with magnetic and dielectric losses (1.2), this chapter explores the theory of
thermal conduction in solids. The examined categories of solids are: metals Sect.
1.3.2, Dielectrics Sects. 1.3.3 and 1.3.4 and Nanocomposites Sect. 1.3.5. In Sect.
1.3.6 the problem of thermal and electrical contact between materials is considered
because contact resistance occurring at conductor joints in magnets or other high
power applications can lead to undesirable electrical losses. At low temperature,
thermal contact is also critical in the mounting of temperature sensors, where bad
contacts can lead to erroneous results, in particular when superconductivity phe-
nomena are involved.
7.1 Electrical Conductivity
7.1.1 Relation Between Thermal and Electrical Conductivity
It is well known that when applying a gradient of temperature to a metal, a heat

flow is observed. Moreover, an electric field applied to a metal produces both a
current flow and a heat flow. This evidence indicates that thermal conductivity and
electrical conductivity are strongly related, and thus they can be treated in parallel
(see Sect. 7.3.3.2).
In general, the electrical and thermal conductivities of pure metals are higher
than those of alloys, see Table 7.1; this is due to the presence of defects which act
as scattering centers for electrons and phonons. On the other hand, insulating
materials and most composites have extremely low thermal conductivities. Some
special crystalline insulators, such as quartz, diamond and sapphire, have high
thermal conductivities. They are useful for electrical insulating connections that
require good thermal contact. Several books deal with these items in detail, see,
e.g., Refs. [1, 2].

Table 7.1 Electrical resistivity of some metals and alloys (lX cm). All data are from Ref. [146]
132
10 K 20 K 50 K 77 K 100 K 150 K 200 K 250 K 295 K

Metals
Ag (RRR 1800) 1 10-4 3 10-3 1.03 10-1 2.7 10-1 4.2 10-1 7.2 10-1 1.03 1.39 1.60
Al (RRR 3500) – 7 10-4 4.7 10-2 2.2 10-1 4.4 10-1 1.01 1.59 2.28 2.68
Au (RRR 3400) 6 10-4 1.2 10-2 2.0 10-1 4.2 10-1 6.2 10-1 1.03 1.44 1.92 2.20
Cu (RRR 3400) – 1.0 10-3 4.9 10-2 1.9 10-1 3.4 10-1 7.0 10-1 1.05 1.38 1.69
Cu (OFHC) (RRR 100) 1.5 10-2 1.7 10-2 8.4 10-2 2.1 10-1 3.4 10-1 7.0 10-1 1.07 1.41 1.70
Cu (OFHC) (60 cold drawn) 3.0 10-2 3.2 10-2 1.0 10-1 2.3 10-1 3.7 10-1 7.2 10-1 1.09 1.43 1.73
Fe (RRR 100) 1.5 10-3 7 10-3 1.35 10-1 5.7 10-1 1.24 3.14 5.3 7.55 9.8
In (RRR 5000) 1.8 10-2 1.6 10-1 9.2 10-1 1.67 2.33 3.80 5.40 7.13 8.83
Nb (RRR 213) – 6.2 10-2 8.9 10-1 2.37 3.82 6.82 9.55 12.12 14.33
Ni (RRR 310) – 9 10-3 1.5 10-1 5.0 10-1 1.00 2.25 3.72 5.40 7.04
Pb (RRR 14000) – 0.53 2.85 4.78 – – – – –
Pb (RRR 100000) – – – – 6.35 9.95 13.64 17.43 20.95
Pt (RRR 600) 2.9 10-3 3.6 10-2 7.2 10-1 1.78 2.742 4.78 6.76 8.70 10.42
Ta (RRR 77) 3.2 10-3 5.1 10-2 9.5 10-1 2.34 3.55 6.13 8.6 11.0 13.1
Ti (RRR 20) – 2.0 10-2 1.4 4.45 7.9 16.7 25.7 34.8 43.1
W (RRR 100) 2 10-4 4.1 10-3 1.50 10-1 5.6 10-1 1.03 2.11 3.20 4.33 5.36
Alloys
Al 1100-0 8 10-2 8 10-2 1.6 10-1 3.2 10-1 5.1 10-1 1.07 1.72 2.37 2.96
Al 5083-0 3.03 3.03 3.13 3.33 3.55 4.15 4.79 5.39 5.92
Al 6061-T6 1.38 1.39 1.48 1.67 1.88 2.46 3.09 3.68 4.19
Berylco (Cu97.7 Be0.02 Co0.003) 6.92 6.92 7.04 7.25 7.46 7.96 8.48 8.98 9.43
Cartridge brass (70 %Cu 30 %Zn) 4.22 4.22 4.39 4.66 4.90 5.42 5.93 6.42 6.87
Hastelloy C 123 123 123 124 – – 126 – 127
Inconel 625 124 124 125 125 – – 127 – 128
Inconel 718 108 108 108 109 – – 114 134 156
(continued)
7 Electrical and Thermal Conductivity
10 K 20 K 50 K 77 K 100 K 150 K 200 K 250 K 295 K
Invar (Fe0.64 Ni0.36) 50.3 50.5 52.1 54.5 57.0 63.3 70.0 76.5 82.3
Monel CuNi30 36.4 36.5 36.6 36.7 36.9 37.4 37.9 38.3 38.5
Phosphor bronze A 8.58 8.58 8.69 8.89 9.07 9.48 9.89 10.3 10.7
Stainless Steel (304 L) 49.5 49.4 50.0 51.5 53.3 58.4 63.8 68.4 73.3
Stainless Steel (310) 68.6 68.8 70.4 72.5 74.4 78.4 82.3 85.7 88.8
Stainless Steel (316) 53.9 53.9 54.9 56.8 58.8 63.8 68.9 73.3 77.1
– 147 148 150 152 157 162 166 169
7.1 Electrical Conductivity
V0.9 Ti0.06 Al0.04

133
134 7 Electrical and Thermal Conductivity
As we mentioned in other parts of this book, the knowledge of the thermal

properties of matter is mandatory to carry out cryogenic experiments. In this
section and in Chap. 9, useful values of electrical and thermal conductivities for
many technical materials are given over the cryogenic range of temperatures.
7.1.2 Electrical Resistivity of Metals
Electric charge is transported through metals by electrons, belonging to the con-

duction band, which are free to move within the crystal lattice of the solid.
Electron motion is generally described by a collective wave model in which the
electron clouds move through the material as waves [3]. It is useful, however, to
explain the temperature dependence of the electrical resistivity and heat conduc-
tion due to electrons by using the elementary kinetic theory of transport in metals
[3, 4].
The free electron model describes the electrons like a ‘‘Fermi gas’’ (a gas made
of free fermions). Let us assume that the thermal energy (kBT) is sufficiently low
compared to the Fermi energy (EF, the energy of the electrons which occupy the
highest-occupied state at T = 0 K). This is, at cryogenic temperatures, a very good
approximation since EF is about 104 K in most metals [5].
Only electrons with energy very close to EF can contribute to the conduction
because they are the only ones that receive a thermal energy sufficient to jump in
the conduction band. Their average velocity (\v[) is
hkF
\v [ ffi vF ¼ ð7:1Þ
me
where kF is the radius of the Fermi sphere and me is the mass of the electron [5].
This velocity is commonly written as
4:20 6 m
vF ¼ 10 ð7:2Þ
n s
where n is a dimensionless parameter ranging between 2 and 6 for most metals [5].
Considering that n is of the order of unit, vF is about 1 % of the velocity of light
and is thus several orders of magnitude higher compared to the velocity of the
particles which form a classical gas.
We can further define the relation between the mean scattering time (the time
between two collisions) and the mean free path (\k[) as
\k [
\s [ ¼ : ð7:3Þ
vF
7.1 Electrical Conductivity 135
The mean scattering time is related to the electrical resistivity q, and thus to the
electrical conductivity r by the equation
1 nc e 2 nc e2
¼r¼ \s [ ¼ \k [ ð7:4Þ
q me me vF
where nc is the number of electrical carriers (only electrons for metals) and e is the
charge of electron. In metals, n is approximately constant; in fact, nc can be
obtained knowing the valence of the metal, and so the only temperature-dependent
term that can influence the electrical resistivity is \k[.
The mean free path between collisions is dominated by two different scattering
mechanisms:
1. At very low temperatures, only a few phonons are present. Thus,\k[is mainly
limited by scattering processes due to chemical or physical crystal-lattice
imperfections (impurities, vacancies, interstitials) and, therefore, is indepen-
dent on temperature. Thus, near liquid helium temperature, q approaches a
constant value which is referred to as ‘‘residual resistivity’’ q0.
2. Near room temperature, the electrical resistivity of most pure metals decreases
monotonically with temperature following an approximately linear relation-
ship. This trend is the result of electron-phonon scattering and is the dominant
temperature-dependent contribution to the resistivity.
It can be useful to introduce Matthiessen’s rule. This empirical rule assumes
that if more than one scattering source is present (e.g., electron-phonon and
electron-impurities), the total q is simply the sum of the resistivities one would
have if each scattering process was present alone [6].
Taking into account the electron-impurities and electron-phonon scattering
contributions, we obtain
qtot ¼ q0 þ qðTÞ: ð7:5Þ
As an example of the behavior of electrical resistivity, Fig. 7.1 shows a plot of

q (T) for various purities of copper defined in terms of the residual resistivity ratio
(RRR = q (273 K)/q (4.2 K), see, e.g., Refs. [1, 7]). The more pure and defect
free the metal, the higher its RRR value. It should also be noted that the tem-
perature at which a near constant resistivity is obtained decreases with increasing
purity. This is obviously due to the fact that, for small amount of impurities, the
defect contribution to scattering became dominant only at very low temperatures.
At very low temperature in high purity samples,\k[may become very large, even
approaching the sample size, such that scattering off the surface of the sample can
cause a ‘‘size effect’’ dependence of the q [8]. Note that at high temperatures, q
curves of all grades of purity collapse in one curve (dashed), representing the
electron-phonon scattering dependence.
In Fig. 7.2., the electrical resistivity of some metals of comparable purity is
reported. A linear behavior is observed from about 50 K to room temperature. It is
Fig. 7.1 Electrical resistivity

versus temperature for copper
of differing purities [12]
Fig. 7.2 Electrical resistivity

of some metals of comparable
purity. Plot from data of
Table 7.1
worth noting that for metallic elements, a concentration of impurities of about

1 ppm can have a significant effect on electron transport, as can the amount of
cold-worked generated imperfections [9].
The universal form for the q of pure metals makes them very useful as tem-
perature sensors, such as platinum resistance thermometers which are used for
precise measurements in the intermediate temperature regime (30–100 K) where
their sensitivity, dR/dT, is roughly constant [10].
The electrical resistivity is often (but not always (see, e.g., [11] and the example
of Sect. 8.5.1) one of the easiest properties to measure and, as a result, q (T) is
known and tabulated for many elements and alloys of interest [12–17].
Electrical resistivities of metals, technical alloys and common solders are
reported in Table 7.1.
7.1 Electrical Conductivity 137
7.1.3 Electrical Conductivity of Semiconductors
Pure semiconductors are a class of materials which have an energy gap (Eg)
between the valence and conduction band of the order of 1 eV at room
temperature.
In (7.4), we found that for metals, the only nonconstant parameter controlling q
was\k[. However, in the case of semiconductors, nc (the number of carriers) also
varies drastically with temperature. At T = 0 K, semiconductors are perfect
insulators: in fact, the conduction band is entirely empty, while the valence band is
full. The number nc of carriers in the conduction band increases with increasing
thermal energy because the fraction of electrons which can ‘‘jump’’ to the con-
duction band is higher at higher temperatures; in particular, it is possible to find an
approximate dependence:
nc eEG=kB T : ð7:6Þ
Pure (or ‘‘intrinsic’’) semiconductors can be distinguished from insulators

simply by the value of the energy gap between conduction and valence band, e.g.,
diamond (EG & 6.3 104 K at 300 K) is also an insulator at high temperatures. It is
worth noting that the energy gap is temperature-dependent. In fact, it can vary by
about 10 % from 0 to 300 K. This behavior is due to the thermal expansion which
modifies the periodic potential experienced by electrons and to the temperature-
dependent phonon distribution.
The resistivity of pure semiconductors covers a large range of values (from
10-4 to 107 Xm) depending on the chemical nature of the material. These values
are orders of magnitude higher than that of most metals (q * 10-8 Xm).
The conductivity of semiconductors can be increased by ‘‘doping’’ them with
impurities which introduce energy levels inside the gap, and hence charge carriers
whose number increases with temperature. Even small concentrations of impurities
can change the conductivity of a semiconductor by several orders of magnitude at
room temperature, as reported in Fig. 7.7.
The typical dopants are elements from the 13th group (called ‘‘acceptors,’’ e.g.,
B, Al, Ga) and from the 15th group (called ‘‘donors,’’ e.g., P, As, Sb) of the
periodic table. This choice is obviously related to the external electronic config-
uration of those elements, which have one electron less (the acceptors) or more
(the donors) compared to the semiconductors belonging to the 14th group (e.g., Si,
Ge). Generally (but not always!), the doping element differs from the semicon-
ductor in terms of total electronic configuration by only one electron: Ge (atomic
number 32) is often doped with Ga (atomic number 31) or with As (atomic number
33) because a similar atomic radius favors the creation of a uniform structure
which minimizes the increased scattering due to the insertion of impurities that can
act as scattering centers.
Fig. 7.3 Resistivity of Ge

doped with Sb versus 1/T.
The number refers to the
donor concentration (cm-3)
[136]
Note that in Fig. 7.3, all curves tend to collapse into one at high temperatures.
The temperature at which curves become one increases with increasing donor
concentration. One might think that this phenomenon is due to the increase of
impurity concentration, but this is not true. In fact, we have to remark that the
range of doping reported in Fig. 7.3 is really limited. If we consider the density
and the atomic weight of pure Ge (gGe = 5.323 g/cm3 and p.a. = 72.64 gmol-1,
respectively) and calculate the number of Ge atoms in a volume of 1 cm3, we get
gGe V NA
nGe ¼ ¼ 4:3 1022 atoms: ð7:7Þ
p:a:
Comparing this number with the doping range (1014–1016 atoms/cm3), we

deduce that in the most doped sample, the ratio between Ge atoms and Sb atoms is
greater than 5 105! The main reason for this behavior is instead due to the different
number of carriers (see (7.4)) in the different samples; this ‘‘extrinsic’’ effect is
present even if the impurity concentration only changes by a factor 102.
Due to the strong temperature dependence of their resistivity, semiconductors
are most commonly encountered in cryogenic applications as temperature sensors
with high negative temperature coefficients of q. For example, a high sensitivity at
T \ 1 K can be achieved using doped Ge as a sensor [18]. Electrical character-
istics of these sensors (frequently called thermistors) often show a strong depen-
dence on magnetic field (magnetoresistance) which can be either a useful or
harmful characteristic, depending on the type of measurement that one wants to
perform [7].
7.2 Magnetic and Dielectric Losses 139
7.2 Magnetic and Dielectric Losses
To understand the problem of losses, we start reminding the reader of the for-
mulation of two of Maxwell’s equations, namely,
!
! oB
r E ¼ ð7:8Þ
ot
!
! ! oD
r H ¼ J þ ; ð7:9Þ
ot
with E = electrical field, B = magnetic induction, H = magnetic field,

J = current density, D = electric displacement field.
It is easy to see that a change of B or D can produce a change in the electrical
properties of the material that are related to the thermal conductivity, thus causing
a generation of heat. For example, the power dissipated in a cylinder of radius
r length L is
pr 4 LðdB=dtÞ2
Pe ¼ : ð7:10Þ
8q
The factor 1/q in (7.10) leads to the choice of low-conductivity materials for the
mixing chamber of dilution refrigerators if high magnetic fields and vibrations are
present [7].
7.2.1 Losses in Dielectric Materials
Losses in dielectric materials are seldom considered in cryogenics. In steady

operating conditions, ‘‘dc’’ losses are extremely small at cryogenic temperatures,
but ‘‘ac’’ losses cannot be neglected because of the low values of c at low tem-
peratures. As with a capacitor, losses are described either in terms of a complex
form of the dielectric constant e(x) = e0 + e00 , whose real part (e0 ) is responsible
for heating, or by tand = e0 /e. The simple lumped constant model which sche-
matizes the phenomenon by a pure capacitance Ce paralleled with a pure resistance
R is physically unsatisfactory. Fortunately, when the capacitor impedance is
measured, a good bridge (e.g., Andeen–Hagerling [19]) supplies the correct value
of both e0 and e00 .
In the case of polymers, losses are due to dynamic mechanical relaxation
caused by heat transfer between the intermolecular mode (strain-sensitive mode)
and the intramolecular mode (strain-insensitive mode) [20]. Since heat is trans-
ferred into the intramolecular modes with a characteristic relaxation time, that is, a
Fig. 7.4 Dielectric loss of

poly(ethy1 methacrylate) at
three frequencies plotted
against temperature [28]
function of temperature, the physical properties of polymer materials heavily

depend on the frequency of the excitation.
If a wide range of temperatures for polymer materials is considered, the exis-
tence of various transitions is very important. The simplest is a phase transition
which occurs in the crystalline region and is a first order transition. The physical
properties of polymeric materials change significantly before and after this phase
transition point. A polymeric substance has many subtransitions originated in
molecular motions beside first order transitions. Since the molecular structure of
the polymer is very complex, many degrees of freedom exist.
The most important transition is the amorphous region in the glass transition
that occurs at a characteristic temperature called Tg. It is well recognized that
polymers possess considerable molecular mobility, even below their glass transi-
tion temperature. Hereafter, molecular mechanisms for relaxations below Tg are
classified by:
(1) internal rotation of an end group in the side chain, as shown in Fig. 7.4 for
poly(ethy1 methacrylate);
(2) proton tunneling, as shown in Fig. 7.5 for polyethylene;
(3) motion of a methyl group, see Refs. [21–23];
(4) molecular motion at defect regions, see Refs. [24–28];
(5) effects due to impurities or additives [29] as shown in Fig. 7.6.
A comparison between the contribution of dielectric and thermal conduction
losses can be found for Upilex R in Refs. [7, 30].
7.3 Thermal Conductivity 141
Fig. 7.5 Temperature

dependence of the dielectric
loss tangent for high-density
polyethylene at 1 kHz [27]
Fig. 7.6 Dielectric loss

tangent of high-density
polyethylene with 0.2 %
Ionox B; high-density
polyethylene with 0.2 %
BHT D; polystyrene with
0.2 % Ionox C together with
the loss for oxidized high-
density polyethylene A. Full
symbols refer to
measurements at 4.2 K, while
empty ones refer to 1.56 K
[29]
7.3 Thermal Conductivity
7.3.1 Introduction
The heat flow through a material is the energy transport phenomenon due to a
thermal gradient. The thermal current density is defined as heat flow qQ/qt per area
(A), and can be expressed (in the x direction) as
1 oQ oTðx; y; z; tÞ
jx ðx; y; z; tÞ ¼ ¼ jðTÞ ð7:11Þ
A ot ox
where j is the thermal conductivity and the minus sign accounts for the fact that
heat Q moves from warmer to colder zones. This formula is valid for other
directions and both in the stationary and nonstationary case. It is important to note
that the thermal gradient is generally time- and direction-dependent, and (7.11) is
valid, point by point, since carriers do not follow a simple linear path, but can
diffuse in all directions.
In a steady situation, the time dependence disappears; when heat conduction is
mainly in x direction, (7.11) becomes
oQ dT
¼ jðTÞA ¼ 0: ð7:12Þ
ot dx
If T is time-dependent, differentiation of (7.12) gives
oT o2 T
¼w 2 ð7:13Þ
ot ox
where w, called thermal diffusivity, is related to thermal conductivity by

j
w¼ ð7:14Þ
gc
where g is the mass density and c is the specific heat of material.

Thermal carriers determining thermal conductivity are lattice vibrations (pho-
nons) and electric-charge carriers (electrons or holes). To estimate the temperature
dependence of thermal conductivity, a very simplified model which considers
thermal carriers as particles of a gas diffusing through a material is often used.
For phonons, the thermal conductivity of an isotropic material can be expressed
as [5]
ZxD
1X
jðTÞ ¼ cix ðx; TÞvi ðx; TÞ\k [ ðx; TÞdx ð7:15Þ
3
0
where the sum runs over the modes ‘‘i’’ up to xD, v is the velocity of phonons,\k[
is the mean free path and cix (differential specific heat) is the contribution of
phonons of frequency between x and x + dx. We can simplify this expression in
the ‘‘dominant phonon’’ approximation [3–5], obtaining
1
j ¼ c \v [ \k [ ð7:16Þ
3
where c is the specific heat per unit volume, \v[ is the average velocity of
particles and \k[ is the mean free path of a carrier inside a material. The velocity
of phonons is the velocity of sound in the material with typical values of about
(3–5) 105 cm/s.
Fig. 7.7 Thermal

conductivity of various
materials at
2 K \ T \ 300 K [7]
This simple formalism can also be used in the case of electrons; however, for
electrons, the velocity can be assumed to be the Fermi velocity because only
electrons near the Fermi energy can contribute to thermal transport (they are the
only ones which can give rise to a transition to higher energy levels, as explained
in Sect. 7.1.1 for electrical conductivity). The typical value of Fermi velocity is
107–108 cm/s.
Note that in (7.16), thermal carriers do not move in a ballistic path: hence,
thermal conductivity is determined by scattering processes between carriers and
point-defects, dislocations, other thermal carriers, boundaries, and crystallites
boundaries. Each scattering event gives a thermal resistance contribution Ri. We
obtain the total thermal conductivity by Matthienssen’s rule [6]
1 X
¼ Ri : ð7:17Þ
j i
Figure 7.7 shows the trend of j for various materials for T [ 2 K. The highest
j is registered for metals and ‘‘special’’ insulators (diamond, sapphire, quartz),
while the lowest can be found in polymers like nylon and polystyrol.
7.3.1.1 Thermal Conductivity of Metals
In analogy with the process of electrical conductivity, the behavior of j can be

understood in terms of a kinetic theory model for gases of electrons and phonons
[31]. In the frame of such a simple model, j is in the form (7.16).
Remembering that for free electrons, the expressions of C and vF,
Fig. 7.8 Thermal conductivity versus temperature: a Of some metals with comparable purity (Al
99.994 %, Ir = 99.995 %, Sn = Zn = 99.997 %, Cu = Ag = 99.999 %) [137]. b Of Cu at
differing purities [12]
p2 kB2 ne
C¼ T ð7:18Þ
2EF
rffiffiffiffiffiffiffiffi
2EF
vF ¼ ; ð7:19Þ
me
are the electronic contribution to the thermal conductivity, and je can be easily
calculated by inserting (7.18) and (7.19) into (7.16) as
p2 kB2 ne \s [
je ¼ T: ð7:20Þ
3me
At high temperatures (T [ hD), the main free path (and thus also \s[, as
described in (7.4)) is proportional to T-1, due to the increase in the lattice
vibrations, and j approaches a constant value. At low temperatures, \s[ is
approximately constant since impurity scattering dominates; hence, the thermal
conductivity should be proportional to T.
As in the case of q, j depends on the chemical nature of the metal and on the
grade of purity of the sample. In Fig. 7.8a, we report the values of j for some
metals with comparable purity, while in Fig. 7.8b, j of copper specimens with
different purities. The asymptotic value near room temperature in Fig. 7.8b gives a
near constant j = 4 W/cm K. With decreasing temperature, the thermal conduc-
tivity passes through a maximum that is typical of almost all metals, which
depends on the purity of the sample, followed by a linear region at the lowest
temperatures, a behavior also visible for all metals in Fig. 7.8a.
As mentioned, the electronic thermal and electrical conductivities in pure
metals have similar scattering processes, thus a correspondence clearly should
exist between these two properties.
Fig. 7.9 Electronic Lorentz

ratio for pure metals and
defect-free metals [35]
The empirical formula which relates these two quantities is known as the
Wiedemann–Franz Law [32]. For the free-electron model, the ratio between
electron thermal and electrical conductivities is given by
je p2 kB2
¼ T ¼ L0 T ð7:21Þ
r 3e2
where e is the electron charge (e & 1.6 10-19 C).

The quantity L0 & 2.45 10-8 V2/K2 is the free electron Lorenz number which
is almost independent of material properties and temperature.
Experiments have shown that the value of L0 is not exactly the same for all
materials. Kittel [33] gives some values of L0 ranging from L = 2.23 9 10-8 W X
K-2 for copper at 273 K to L = 7.2 9 10-8 W X K-2 for tungsten at 373 K. The
Wiedemann–Franz law is generally valid near room temperature and for low
temperatures (T \\ hD), but may not hold at intermediate temperatures [5]. In
certain materials (such as Ag or Al), however, the value of L0 also may decrease
with increasing temperature. In the purest samples of Ag and at very low tem-
peratures, L0 can drop by as much as an order of magnitude [34]. The overall
behavior of the Lorenz ratios with sample purity are plotted in Fig. 7.9 [35].
7.3.2 Lattice Thermal Conductivity
Also, the lattice contribution to the thermal conductivity of metals, semiconductors

and insulators may be explained in terms of kinetic theory, although the thermal
carriers in this case are phonons. It is still possible to apply (7.16). Note that
sometimes, \k[ becomes as large as the specimen size (i.e., the ‘‘size effect’’).
Most insulators and intrinsic semiconductors have thermal conductivities sev-

eral orders of magnitude lower than common pure metals. At temperatures above
*20 K, generally the thermal conductivity decreases monotonically with tem-
perature. At low temperatures, below *10 K, where the scattering becomes
approximately independent of temperature, the thermal conductivity decreases
more rapidly, approaching zero as T7. For more details, see, e.g., Ref. [36].
7.3.3 Thermal Conductivity of Dielectrics
Materials presenting very low electric conductivity (e0 B 10-2 at room tempera-
ture, frequency = 1 kHz, see Sect. 7.2.1) and low thermal conductivity
(B1 Wm-1 K-1) are called dielectric materials; in this case, no electronic con-
tribution to thermal conductivity exists, so only phonon scattering phenomena are
to be considered. We shall divide materials in two different subclasses: pure
crystals and amorphous materials. This choice is due to the fact that different
peculiar scattering events occur in the two cases.
7.3.3.1 Pure Crystals
The most important scattering mechanisms which have been observed are:
(a) Phonon-phonon scattering (umklapp processes).
(b) Phonon-boundaries of specimen (or crystallites).
(c) Phonon-point defects scattering.
(d) Phonon-dislocations scattering.
Detailed and formal treatment of phonon scattering processes may be found,
e.g., in Ref. [3].
At high temperatures (T [ hD), the main contribution is due to the phonon-
phonon scattering. In fact, in this range of temperatures, the number of phonons is
large enough to give rise to umklapp processes (u-processes, [3]) since the
probability of this kind of events is proportional to the number of phonons. When
two phonons interact, a wave vector which falls outside the first zone of bound-
aries is obtained: the total effect is a reduction of heat flow or, in other words, a
thermal resistance [4]. This phenomenon gives a resistance contribution, for
T C hD ,
T
Ru / ð7:22Þ
\m [ ahD
where \m[ and a are the mean atomic mass and spacing, respectively.
Instead, for T hD,
Fig. 7.10 Qualitative representation of the conductivity of a pure crystalline material (a) and of a
crystal with isotopes or impurities (b)
Ru / ehD=T : ð7:23Þ
We obtain the boundary contribution from (7.15)
RB / T 3 : ð7:24Þ
For materials with point defects, such as vacant lattice sites, interstitial atoms,
impurity atoms, or isotopes of the specimen, starting from (7.14), we obtain a
linear dependence on temperature. Callaway [37] gives a more rigorous treatment,
obtaining
RD / T 3=2 : ð7:25Þ
Dislocations, i.e., imperfections in the crystal lattice with one dimension

extension, give a contribution with a temperature dependence
RD / T 2 : ð7:26Þ
Summing all these contributions, the effect on thermal conductivity leads to the
graphs of Fig. 7.10.
In Fig. 7.11, we report thermal conductivity data of some pure crystals.
A special material is graphite for the thermal behavior due to its peculiar
structure. In fact, graphite behaves as a very good conductor at high temperature,
but below 1 K is a very good insulator [38].
This fact is very interesting from a cryogenic point of view because graphite
can be employed as a thermal switch. In Fig. 7.7, we report the comparison among
data of some types of graphite (Figs. 7.12, 7.13).
Fig. 7.11 Thermal

conductivities for some pure
crystals: LiF, KCl, TiO2
[138]; KBr, KBr0.53I0.47
[139]; Li3N [140]; Al2O3
[141]; La2CuO4,
La1.9Sr0.1CuO4 [142]
Fig. 7.12 Comparison of the

conductivity of graphites [38]
Fig. 7.13 Comparison

between the thermal
conductivity of an
amorphous, semi-crystalline
and crystalline material
7.3.3.2 Amorphous and Semi-crystalline Materials
Amorphous materials have a peculiar temperature dependence of thermal con-

ductivity. This behavior is shown in Fig. 7.17.
We observe a T 2 dependence at very low temperature, a plateau between about
5–15 K and a weak positive slope above 20 K. A semi-crystalline sample shows a
higher conductivity compared to amorphous above 15 K, and lower j at low
temperature. We shall now concisely analyze contributions to j of the scattering
processes. For further information, see, e.g., Ref. [39].
(a) Amorphous materials
T B 1 K: Tunneling Processes
The measurements of thermal conductivity carried out by Zaitlin and Anderson
in 1975 demonstrated for the first time that below 1 K, the acoustic phonons are
the main responsibility of the heat transfer [40]. Instead, the excitations which
produce the excess (the almost linear contribution) of specific heat cannot carry
thermal energy because they are to be considered as localized excitations. The
measurements of Zaitlin and Anderson confirmed the tunneling model proposed
independently by Anderson et al. [41] and Phillips [42] in 1972 with the aim of
explaining the measured thermal and acoustic properties of amorphous materials.
According to this ‘‘two-level state’’ (TLS) theory, because of the structural
disorder, groups of atoms have more than one possible position, each differing
from the other for a very small energy, of the order of E B 10-4 eV (see also
Sect. 1.7). The quantum tunneling transition between the two levels can only take
place with absorption or emission of phonons in order to conserve energy. From
this theory, the mean free path is a function of frequency and temperature
hx
\k [ / x1 coth : ð7:27Þ
2kB T
The two-level systems present separation energy arranged in a wide range; thus,
phonons involved in this kind of scattering can present a wide range of frequency.
In the ‘‘dominant phonon’’ approximation, an estimation of the mean free path in
the function of temperature gives
\k [ ðxÞ / x1 ð7:28Þ
\k [ ðTÞ / T 1 : ð7:29Þ
As we discussed in Sect. 1.7, the specific heat contribution of TLS is propor-

tional to T. Recalling that the mean free path is proportional to j-1, we obtain
j / T 2: ð7:30Þ
4 K \ T \ 15 K: Plateau Region
This plateau is typical of amorphous materials in this indicative temperature
range. There is not a really satisfactory explanation of this behavior. A formal
solution is given by choosing an opportune\k[(x) in order to obtain a constant j.
The specific heat can be calculated as
o
cx / ð hxDðxÞ \ f ðx; TÞ [Þ: ð7:31Þ
oT
The value is determined by the product D(x)k(x) which can be expressed by an

exponential law xd. In the acoustic approximation (T \\ hD), we obtain
j / T 1þd : ð7:32Þ
In this zone, polymer materials act as a low-pass filter for phonon, and domi-
nant-phonon approximation is not applicable. For further information, see Refs. [3,
43–47].
T [ 30 K
For T [ 30 K, the mean free path does not depend on temperature or frequency:
in fact, \k[ is the order of a few atomic spaces. In this case, the term ‘‘phonon’’
has no significance. In this range, we experimentally observe a weak dependence
of j on temperature. For polymers,
j / Td ð7:33Þ
with d = 0.3–0.5, depending on the chemical composition of the material.

Figure 7.14 shows the thermal conductivity of some amorphous solids.
(b) Semi-crystalline polymeric materials
The thermal conductivity of semi-crystalline materials show quite a different
dependence on temperature compared with amorphous materials because it is
strictly dependent on the quantity and size of crystalline inclusions (crystallites)
[48]. Normally, compared to pure amorphous samples, semi-crystalline materials
show a lower conductivity below 30 K, and a higher conductivity above. A
parameter of paramount importance is the crystallinity (fc) [49] of the sample,
defined as the weight percentage of crystal phase over the total weight of the
sample because a low degree of order in the polymer (disordered chains) can
drastically decrease \k[, thus lowering j, according to (7.16).
T \ 20 K: Interface Scattering
Semi-crystalline materials do not show a plateau region because they have an
additional resistive term due to discontinuity between amorphous and crystalline
zones [50]. They show a temperature dependence Tc with 0.5 \ c \ 3. Interface
scattering resistivity increases with fc, leading to a decrease of j with temperature
Fig. 7.14 Thermal

conductivity of some
amorphous solids [143]
Fig. 7.15 Thermal

conductivity of (BaF2)1-
x(LaF3)x [51]
and to an increase of j(T) slope. For temperatures below 1 K, thermal conductivity

decreases and converges to a T2 dependence for temperatures low enough
(T \ 0.1 K). At such temperatures, in fact, phonons have a wavelength long
enough to see the crystalline zones as point defects; hence, scattering processes
due to TLS remain dominant. Above *30 K, thermal conductivity increases with
fc and it sometimes can show a peak around 100 K [49].
In Fig. 7.15, we report the thermal conductivity of (BaF2)1-x(LaF3)x as a typical
example. The pure BaF2 (x = 0) and the pure LaF3 (x = 1) have a thermal con-
ductivity typical of perfect crystals. By increasing doping, the thermal conductivity
is lower, slowly approaching a minimum value (jmin, the straight line in Fig. 7.15)
for x = 0.33 at high temperatures. It is worth noting that for this value of doping,
the conductivity assumes a trend similar to that of a-SiO2 (dashed line in Fig. 7.15);
however, a complete amorphous-like behavior is never observed [51].
In Table 7.2, we report a summary of thermal behavior for amorphous poly-
meric and semi-crystalline materials.
Table 7.2 Summary of thermal behavior for amorphous polymeric and semi-crystalline
materials
Temperature range Temperature dependence Scattering processes
Amorphous polymers
TB1K j / T2 TLS scattering
4 B T B 15 K j ¼ kos Rayleigh scattering
T [ 30 K j / T n n = 0.5–0.3 Structural defects scattering
Semi-crystalline polymers
TB1K j / T2 TLS scattering
1BTB2K j / T n n = 1–2 TLS scattering and crystalline zones
2 B T B 20 K j / T n n = 0.5–3 Crystalline zones scattering
T [ 30 K j / Tn Depends on degree of crystallinity
7.3.4 Thermal Conductivity of Nanocomposites
When reduced to nanoscale, a lot of materials radically change their physical and
chemical properties. This is principally due to the fact that the smaller is the
dimension of the nanomaterial, the bigger is the ratio between surface and volume
atoms, thus providing high and sometimes unexpected reactivity. To emphasize
that atoms at the surface possess high reactivity (due to their unsaturated coor-
dination sphere) compared to bulk atoms, and also behaviors very difficult to
rationalize, we can cite the famous sentence by Wolfgang Pauli, that is, ‘‘God -
made the bulk; surfaces were invented by the devil.’’ The nanoparticles and
nanomaterials have unique mechanical, electronic, magnetic, thermal, optical, and
chemical properties, thus providing a wide spectrum of new possibilities of
engineered nanostructures and nanocomposites for communications, biotechnol-
ogy and medicine, photonics and electronics. For all of these reasons, nanotech-
nology is a research field of growing importance [52].
The first remarkable talk about nanotechnology was given by Richard Feynman
in 1959 [53]. However, even if the terms nanomaterial and nanocomposite were
introduced in the 20th century, such materials have actually been used for cen-
turies and have always existed in nature [54]. One of the first and most famous
examples is the Lycurgus cup, made in the 4th century AD. The opaque green cup
turns to a glowing translucent red if illuminated. Chemical analysis of this
extraordinary artwork indicates that the glass contains approximately 330 ppm of
silver and 40 ppm of gold with an average particle size of approximately 70 nm.
However, it is not the presence of these particular elements that is responsible for
the effect, but rather the way the initial glass composite was produced [55–57].
Another example where enhanced properties were not obtained from under-
standing but from empirical experiments is the Damascus steel. The swords that
were made of these alloys were very flexible, sharp and stiff. Many centuries later,
it was discovered that ancient Muslim smiths, in the 17th century, were inadver-
tently using carbon nanotubes within the metallic matrix of the blade [58]. Other
primitive nanocomposites were created in the 1860s. Experiments with vulcanized
rubber and carbon black led to significant enhancements of the mechanical

properties of rubber tires [59].
A nanocomposite (NC) may be defined as a composite system consisting of a
polymer matrix and homogeneously dispersed filler particles which have at least
one dimension below 100 nm. Polymers are the most common materials that are
used for NCs fabrication. During the past decades, polymer NCs have attracted
considerable interest both in science and industry [60].
The first nanoclay composite, which was produced to reinforce the macroscopic
properties of an elastomer, was described in a patent from the National Lead
Company in 1950 [61]. The discovery of carbon nanotubes (CNT) by Iijima in
1991 [62] and Buckminsterfullerene (C60) by R. F. Curl, Sir H. W. Kroto and R.
E. Smalley in 1995 (Nobel Prices in Chemistry in 1996) were the first steps
towards a production of single- and multi-walled carbon nanotubes and new
nanoscale materials and devices based on CNT [63]. For the great versatility of
chemical and physical behaviors that characterize the nanomaterials, the fields of
application range from agriculture and food production to space science and
medicine [64].
There is no satisfactory explanation for the origin of the change of the prop-
erties of polymer NCs. It is generally accepted that the large surface-to-volume
ratio of the nanoscale inclusions plays a significant role [65, 66]. Smaller particles
display a much larger surface area for interaction with the polymer for the same
microscopic volume fraction than larger particles, so it is generally better to
minimize the dimension, even providing the desired properties [55]. It is currently
thought [66] that many of the characteristics of NCs are determined by the
interactions that occur at nanoparticle-matrix interfaces. The creation of a
homogeneous distribution of nanoparticles is not an easy task because particles
have a strong tendency to agglomerate: in fact, almost all the nanomaterials are
kinetically and thermodynamically unstable objects. To prevent aggregation, it is
common to cover the surface of these objects with single molecules or polymers,
thus limiting the interaction because of hindrance and/or electrostatic repulsion
[59, 67]. The formation of chemical bonds between the inorganic and organic
components is of great importance for a homogeneous dispersion of the filler in
host polymers [68–70]. A coupling agent is a chemical substance that is applied to
the surface of a material that has to be modified to make it compatible with another
material of a different nature. The molecular structure enables the coupling agent
to work as an intermediary in bonding organic and inorganic materials [71]. A
variety of coupling agents, such as silanes, zirconates, titanates and zircoalumi-
nates have been introduced to the market since then in order to improve the
interface between the polymer and the filler [72].
One of the critical aspects of nanotechnology research is how to modify the
surface of different nanoparticles to make them compatible with polymer matrices
and more useful for different applications [73]. The most important changes in
properties of NC are not caused by the order of magnitude in size reduction, but by
the phenomena such as size confinement, predominance of interfacial phenomena
and quantum mechanisms [74–76].
Fillers may be classified as inorganic or organic substances, and are further

subdivided according to their chemical family [77]. Stable dispersion of filler in
the final composite is necessary to eliminate filler agglomerates that would act as
weak points that might induce electrical or mechanical failure.
Nanoparticle composite properties at room temperature have received enor-
mous attention in the last decade, but studies at low temperature are very rare.
Room temperature thermal conductivities of insulating polymer materials are
usually 1–3 orders of magnitude lower than those of ceramics and metals. Due to
the chain-like structure of polymers, the heat capacity consists of the contribution
of both lattice vibrations and other type of vibrations, characteristic of the con-
sidered material, which originate from internal motions of the repeating unit. The
lattice (skeleton) vibrations are acoustic vibrations which give the main contri-
bution to the thermal conductivity at low temperatures. The characteristic vibra-
tions of the side groups of the polymer chains are instead optical vibrations which
become visible at temperatures above 100 K [78]. As we have seen in
Sect. 7.3.3.2, generally, the thermal conductivity of amorphous polymers increases
with increasing temperature if the temperature is in the glassy region and decreases
slowly or remains constant in the rubbery region.
Numerous applications in the field of electrical engineering require high ther-
mal conductivity, such as insulating materials for power equipment, electronic
packaging and encapsulations, computer chips, satellite devices and other areas
where good heat dissipation is needed. For polymers reinforced with different
types of fillers, this is even more important. Improved thermal conductivity in
polymers may be achieved either by molecular orientation or by the addition of
highly heat-conductive fillers [79, 80].
Temperature, pressure, density of the polymer, orientation of chain segments,
crystal structure, crystallinity and many other factors may affect the thermal
conductivity of polymers [7, 81].
Figure 7.16 shows the schematic representation of the higher-order structure of
a resin to achieve macroscopic isotropy and high thermal conductivity. The pro-
posed resin has three characteristic features:
(a) microscopic anisotropic crystal-like structures obtained via local alignment,
e.g., via oriented mesogens (see, e.g., [82]);
(b) macroscopic isotropy of the epoxy due to disorder of the domains of the
crystal-like structures;
(c) the oriented mesogens are connected with the amorphous structure via
covalent bonds.
The thermal conductivity values of the new developed resin were up to five
times higher than those of conventional epoxy resins because the mesogens form
highly ordered crystal-like structures which suppress phonon scattering.
To improve the thermal conductivity of the polymer composites, Ekstrand and
co-authors [83] proposed three approaches that can also be realized in parallel:
Fig. 7.16 Schematic

representation of a
macroscopically isotropic
epoxy resin [144]
(a) decreasing the number of thermally resistant junctions;

(b) forming conducting networks by suitable packing;
(c) minimize filler-matrix interfacial defects.
To enhance the thermal conductivity of polymeric structures the main scientific
approach is to fill them with particles of materials with high thermal conductivity
such as a-Al2O3, b-SiO2, SiC, diamond, SiN and BN [84–86]. In particular, boron
nitride [87–89] led to the best candidate to effectively improve the thermal con-
ductivity of epoxy-based composites. A study published by Han et al. [89] proved
that the size of dispersed particles is not crucial until a high doping ratio is
reached, thus allowing for an easier composites preparation. Industrial companies
specializing in the production of polymer-based insulating materials use a fill-
grade up to 60 wt% of SiO2 or Al2O7. The thermal conductivity of these materials
is not significantly higher than that of pure polymers, but the very low price
justifies their production [84, 85].
7.3.5 Composite Materials
Different theoretical and empirical approaches are available to predict and fit the
thermal conductivity of two-phase systems. Here, we present a simple overview of
the principal theories about the thermal conductivity of composite materials.
The simplest three are the rule of mixture (parallel model, arithmetic mean):
jC ¼ f jF þ ð1 f Þjm ; ð7:34Þ
the inverse rule of mixture (series model, harmonic mean):
1 f ð1 f Þ
¼ þ ; ð7:35Þ
j C jF jm
and the geometric mean
jC ¼ ðjF Þf ðjm Þð1f Þ : ð7:36Þ
In all formulas, jc, jF and jm are the thermal conductivities of composite, filler
material and polymer matrix, respectively, and f is the volume fraction of the filler
[90, 91].
The upper or lower boundaries of the thermal conductivity are given when filler
particles are arranged either parallel to or in series with the heat flow. As soon as
the particles have a random distribution and are not aligned in the direction of the
heat flow in the polymer, the parallel and series model do not give a good pre-
diction of the thermal conductivity of the composites. The parallel model typically
overestimates the thermal conductivity of a composite and thus shows the upper
limit, while the series model tends to predict the lower limit of the thermal con-
ductivity of a two-component system [92, 93].
Maxwell obtained a formula for the electrical conductivity of randomly dis-
tributed and noninteracting homogeneous spheres in a homogeneous medium [94].
Eucken adapted the electrical conductivity equation to thermal conductivity [95].
Frieke extended Maxwell’s model and derived an equation for ellipsoidal particles
in a continuous phase [96]. Using different assumptions for permeability and field
strength than Maxwell, Bruggeman derived the theoretical model for a dilute
suspension of noninteracting spheres dispersed in a homogeneous medium [97].
However, most of the experimental results show that Maxwell–Eucken and
Bruggeman models as well as the Frieke model do not predict the thermal con-
ductivity of a composite correctly [86, 98–100].
Tsao developed a model relating the thermal conductivity of a composite to two
experimentally determined parameters which describe the spatial distribution of
the two phases [101]. Cheng and Vachon extended Tsao’s model by assuming the
discrete phase in the continuous matrix [102].
Sundstrom and Lee reported that the Cheng–Vachon model shows a reasonable
agreement with experimental data obtained from polystyrene or polyethylene
systems filled with glass, calcium oxide (CaO), aluminum oxide (Al2O3) and
magnesium oxide (MgO) [103]. Contrary to Sundstrom and Lee, Hill and Supancic
showed that the results predicted by the Cheng–Vachon model have much lower
thermal conductivity values compared to experimental results [99].
Table 7.3 List of thermal conductivity models and when they can be successfully applied
Model Terms of use
Series, parallel Particles are aligned either parallel or perpendicular to heat flow
Maxwell–Eucken, Frieke, Ideal system, noninteracting spherical/spheroidal particles in
Bruggeman homogeneous medium
Geometric mean, Cheng– Discrete phase in continuous matrix, only taking into account
Vachon filler loading
Hamilton–Crosser Sphericity is taken into consideration
Hatta–Taya For the systems filled with high-aspect ratio particles
Meredith–Tobias For high-loaded composites
Lewis–Nielsen Size, geometry and manner of particle packing is taken into
account
Agari–Uno Fitting function with adjustable constants
Russell, Topper For porous composites, containing voids of gas
Aforementioned models are based on the amount of filler loading and do not
take into account the geometry of the particles and the size of filler particles. The
work of Hamilton and Crosser [104, 105] is based on Maxwell’s and Frieke’s
theoretical models. They take into consideration the sphericity of particles (the
sphericity is defined as the surface area of a sphere with the same volume as the
particle divided by the surface area of the particle). Hatta and Taya [106] proposed
a model which can be applied to systems filled with particles having a high aspect
ratio or unidirectional fillers.
Meredith and Tobias suggested a model for high-loaded systems [107]. Lewis
and Nielsen [108–110] adopted the Halpin–Tsai [111] mechanical model to obtain
a model for the thermal conductivity. Agari and Uno [112–114] proposed a model
which is based on the generalization of both parallel and series models for filled
composites. Generally, the Agari and Uno semi-empirical model fits experimental
data well. However, it does not predict the thermal conductivity, but is basically a
fit function. To extend the overview of the thermal conductivity modeling, many
different models can be mentioned, for example, Russell [115], Topper [116],
Jefferson–Witzell–Sibitt [117], Springer–Tsai [118], Budiansky [119], Baschirow
and Selenew [120], McCullough [90], and McGee [121], and many others,
including mathematical numerical methods [122, 123].
Summarizing, we can conclude that the thermal properties become more
complicated with the addition of fillers to polymers, and thus no single theory or
technique accurately predicts the thermal conductivity for all types of composites,
but all of them can be successfully applied in particular cases.
Table 7.3 shows the list of models which can be used to describe the thermal
conductivity of composite systems and when the particular model can be applied.
As mentioned, the number of references to measurements of thermal conduc-
tivity at room temperature is huge. We only wish to cite two low temperature
examples: Ref. [124] reports on the measurements of thermal properties of
nanosystems at very low temperatures by the 3x method. Authors discuss the
Fig. 7.17 Low-temperature

thermal conductivity of
Nylon-6/Cu obtained by
Martelli et al. [126] compared
to the one obtained by Scott
et al. for pure Nylon-6 [145]
intrinsic limitations of these methods when the thermal properties of nano-objects

are studied at temperatures below 1.2 K. In Ref. [125], SiO2/epoxy NCs room
temperature tensile properties are reported. The effects of silica nanoparticle
content is studied on the cryogenic thermal properties of the NCs.
Reference [126] reports the measurement of the thermal conductivity of a NC
material made of a Nylon-6 matrix in which metallic copper nanoparticle (5 % in
weight) are uniformly dispersed. Nevertheless, data measured can differ substan-
tially from the one obtained for pure polymers, also showing interesting features,
in particular, a sharp dip at 1.4 K as shown in Fig. 7.17.
This is a unique features in thermal conductivity interpreted as a resonant
scattering of phonons by copper nanoparticles. The temperature at which phonon
frequency equals nanoparticle resonant frequency is
sffiffiffiffiffiffiffiffiffiffiffiffiffi
h EþG
T¼ ð7:37Þ
2pkB L 2q
where L is the peak of size distribution, E is longitudinal elastic modulus (Young

modulus), G is the shear modulus for tangential strain and q is the (Cu) density.
Hence, relying on (a), the temperature of the negative notch in conductivity can be
modulated by changing the parameters of the chemical synthesis. The practical
applications of notch in the conductivity of composite materials have not yet been
explored and are beyond the goals of this book.
7.3.5.1 Contact Resistance
Thermal and electrical contact between materials (also two pieces of the same
material) is an important subject in cryogenics (particularly at very low temper-
atures), and yet it is still only qualitatively understood. Contract resistance
occurring at conductor joints in magnets or in other instruments that require high

power can lead to undesirable electrical losses. At low temperature, thermal
contact is also critical in the mounting of temperature sensors, where bad contacts
can lead to erroneous results, in particular, when superconductivity phenomena are
involved.
When two materials are joined together for the purpose of transporting heat or
electrical current, a localized resistance appears at the boundary. The magnitude of
this resistance depends on a number of factors, including the properties of the bulk
materials, the preparation of the interface between the two materials, whether there
are bonding or interface agents present, and external factors such as the applied
pressure.
The electrical contact resistance is of greatest interest in the production of joints
between high purity metals such as copper, where its value can contribute or even
dominate the overall resistance of an electrical circuit. Generally, the contact
resistance in pure metals has a temperature dependence that scales with the
properties of the bulk material. For electrical contacts between pure metals without
bonding materials like solder, the value of the electrical contact resistance
decreases with applied pressure normal to the joint interface. This tendency is due
to an increase with pressure in the effective contact area between the two bulk
samples. In fact, the two surfaces have microscale roughness due to how the
surfaces were prepared: as the pressure is increased normal to the surfaces, the
asperities tend to mechanically yield and deform, increasing the effective area of
contact. As the bulk material has high conductivity, the contact resistance is
mostly due to the constriction of current or heat flow that occurs at the small
contact points [127]. By increasing the contact pressure, the amount of constriction
for current flow decreases, thus reducing the contact resistance. At very low
temperatures, the aforementioned phenomenon should be investigated more
thoroughly since mechanical stress in the contact zone may change the bulk
properties of materials (see, e.g., Ref. [128]).
A summary of the measured electrical contact resistivity for various unbonded
samples as a function of applied pressure can be found in Ref. [129].
Values of contact resistance can be obtained by, RB = qB/A, where A is the
apparent contact area. Note that at a particular contact pressure, there is still a wide
variation in the contact resistivity, a result that is probably due to variations in
sample preparation, treatment and oxidation. The contact resistance generally
decreases with applied pressure as
n
qB ð7:38Þ
p
where the pressure is expressed in Pa and the resistivity in Xm2. The parameter n is
experimentally determined and is often about 3 104 [Kg m s-2 X] for metals [1],
but can vary significantly for insulators [129].
Fig. 7.18 Thermal contact

conductance as a function of
temperature for a variety of
contact preparations and
conditions. The contact area
is assumed to be of 1 cm2.
The apex s indicates a solder.
Dashed lines are estimates of
the conductance in
temperature regions where no
data were available. Data
were taken from [132]
For thermal contact resistance, there are two cases:

(a) the thermal contact resistance between metals, which is expected to correlate
with the electrical contact resistance as much as with bulk metals. This
correlation is approximately correct for contacts between identical metals.
This means that Wiedemann–Franz law must be used with great caution. If
the contact is between dissimilar metals or if there are solders or other
interface metals involved, the thermal contact resistance can no longer be
scaled with qB. This latter point is particularly significant at low temperatures
where many alloys are superconducting (see Fig. 7.18);
(b) for thermal contact resistance between nonconducting materials, the funda-
mental limit, even for ideal contacts, is the mismatch in the phonon transport
across the interface [7, 130]. Since the phonon spectra for the two types of
materials are not the same, there is an impedance mismatch that leads to a
resistance occurring within roughly one phonon wavelength at the interface.
This effect is known as Kapitza conductance which initially referred to the heat
transfer between liquid helium and metals (see, e.g., Ref. [7]). The theory of
Kapitza conductance predicts [131] a Kapitza conductivity (jK)
jK const T 3 : ð7:39Þ
For most solids, const is on the order of 1 kW/m2 K4 [129]. Equation (7.39)
puts an upper limit on the magnitude of the thermal contact conductance for
insulating contacts; real contacts between nonideal surfaces are more complex and
their understanding is still qualitative.
For joints between real materials, the interface is irregular with random points
of contact. In this case, the thermal contact conductance mostly depends on the
constriction resistance at the asperities similarly to the electrical contact resistance
in metals. Thus, particularly for deformable materials, the thermal contact con-
ductance increases with interface pressure.
It is important to pay attention because for metals at very low temperature, the
thermal contact resistance may be higher than that of a low-resistivity thermal
connection [128].
Experimentally, the correlation between contact conductivity and pressure can
be written as
jK w p n ð7:40Þ
where n & 1 and w is an empirical coefficient [129].

Thermal contact conductance varies over a wide range, depending on whether
the contact is insulating or conducting. Figure 7.17 shows data for low temperature
thermal contacts [132]. Some general features can be observed:
(a) the thermal contact conductance values at low temperatures can range over
six orders of magnitude, depending on materials and surface preparation;
(b) solder-bonded contacts,with solder of similar agents that fill the asperities
generally have higher thermal conductance than bare contacts. However, the
bonding agents can also contribute to the interface resistance, particularly if
the bond region is thick. In the low temperature region (T \ 5 K), most of the
data agree with a power law, q * T-n, but there are two distinct charac-
teristic behaviors:
(1) pure metal-metal contacts have a temperature dependence that correlates
with that of the bulk metal, hence, at low temperature q * T-1, with the
coefficient of proportionality being mainly determined by sample purity
and contact pressure, but varying between 10-1 and 10-3 W/cm2K2;
(2) if the contact is bonded with solder or indium, the conductance can be
much higher, but at low temperatures, such contacts may become
superconducting.
Finally, if the interface is between two nonconducting materials, the thermal
conductance is generally lower, following the correlation scaling with the bulk
thermal conductivity, q * T-n, where n * 7.
In addition to Fig. 7.18, Ref. [133] reports the experimental values of the
thermal boundary resistance occurring at interfaces between two solids at su-
bambient temperatures. Data are in the 4–300 K range and report the thermal
resistance between different metals (Cu, stainless steel), interlayered by various
cryogenic bonding agents (Apiezon-N, Cryocon grease, In and InGa), or
mechanically connected (dry) contacts.
In Ref. [134], the thermal contact conductance of several demountable copper
joints below 1 K is reported. Joints were made by bolting together either two gold
flat surfaces or by a clamp around a rod. A linear dependence on temperature was
seen. Most of the measured conductance values fall into a narrow range:
0.1–0.2 WK-1 at 1 K. Results in the literature for similar joints consist of
predictions based on electrical resistance measurements using the Wiedemann–

Franz law. However, there is little evidence of the validity of this law in the case of
joints. Nevertheless, the results are in agreement with the literature predictions,
suggesting that such predictions are a reasonable approximation.
In Ref. [135], thermal conductance measurements of different types of bolted
joint at sub-Kelvin temperatures are presented. Joints containing sapphire surfaces
provided good thermal isolation in the 100 mK and 4 K temperature range. The
best joint contained sapphire discs separated by diamond powder and had a con-
ductance of 0.26 lWK-1. A mechanical support structure constructed from similar
joints, but using alumina powder, had a measured heat leak of 2.57 lW between
80 mK and 1.1 K and was capable of supporting a mass of over 10 kg. Joints
between metal surfaces provided good thermal conduction; a bolted joint between
copper and a beryllium-copper alloy (C17510 TF00) had a measured conductance
of 46 mW K-1 at 100 mK, increasing linearly with temperature. The paper also
reports measurements made on a copper-copper compression joint using differ-
ential thermal contraction to provide the clamping force: the performance is about
an order of magnitude worse than for the bolted joint.
References
1. Van Sciver, S.W.: Helium Cryogenics. Springer, New York (2012)

2. Tritt, T.M.: Thermal Conductivity: Theory, Properties, and Applications. Springer, New
York (2004)
3. Ziman, J. (ed.): Electrons and Phonons. Clarendon Press, Oxford (1972)
4. Rosenberg, H.M. (ed.): The Solid State. Clarendon Press, Oxford (1984)
5. Ashcroft, N.W., Mermin, N.D.: Solid State Physics Holt. Rinehart and Winston, New York
(1976)
6. Matthiessen, A., Vogt, C.: On the influence of temperature on the electric conducting-power
of alloys. Philos. Trans. R. Soc. Lond. 154, 167–200 (1864)
8. Olson, J.: Thermal conductivity of some common cryostat materials between 0.05 and 2 K.
Cryogenics 33(7), 729–731 (1993)
9. DeGarmo, E.P., Black, J.T., Kohser, R.A., Klamecki, B.E.: Materials and Process in
Manufacturing. Macmillan Publishing Company, New York (1984)
10. Moiseeva, N.: Methods of constructing an individual calibration characteristic for working
platinum resistance thermometers. Meas. Tech. 44(5), 502–507 (2001)
11. Woodcraft, A.L.: Zirconium copper—a new material for use at low temperatures? In: AIP
Conference Proceedings 2006, p. 1691 (2006)
12. Powell, R., Fickett, F.: Cryogenic properties of copper vol. 1. In: Proceedings of INCRA
REP (1979)
13. Clark, A., Childs, G., Wallace, G.: Electrical resistivity of some engineering alloys at low
temperatures. Cryogenics 10(4), 295–305 (1970)
14. Ledbetter, H., Reed, R., Clark, A.: Materials at Low Temperatures, vol. 1. American Society
for Metals, Metals Park, OH (1983)
15. Meaden, G.T.: Electrical Resistance of Metals, vol. 2. Plenum press, New York (1965)
16. Hall, L.: Survey of Electrical Resistivity Measurements on 16 Pure Metals in the
Temperature Range 0 to 273 K (1968). http://www.getcited.org/pub/101292840
References 163
17. Matula, R.A.: Electrical resistivity of copper, gold, palladium, and silver. J. Phys. Chem.
Ref. Data 8, 1147 (1979)
18. Haller, E.: Advanced far-infrared detectors. Infrared Phys. Technol. 35(2), 127–146 (1994)
19. Andeen, C.G., Hagerling, C.W.: High Precision Capacitance Bridge. In. Andeen-Hagerling
Inc., Ohio (1988)
20. Hayakawa, R., Tanabe, Y., Wada, Y.: A thermodynamic theory of mechanical relaxation
due to energy transfer between strain-sensitive and strain-insensitive modes in polymers.
J. Macromol. Sci. Part B Phys. 8(3–4), 445–461 (1973)
21. Powles, J.: Nuclear magnetic resonance absorption in polymethyl methacrylate and
polymethyl a-chloroacrylate. J. Polym. Sci. 22(100), 79–93 (1956)
22. Odajima, A., Woodward, A., Sauer, J.: Proton magnetic resonance of some a-methyl group-
containing polymers and their monomers. J. Polym. Sci. 55(161), 181–196 (1961)
23. Tanabe, Y., Hirose, J., Okano, K., Wada, Y.: Methyl group relaxations in the glassy phase of
polymers. Polymer J 1, 107–115 (1970)
24. Armeniades, C., Baer, E.: Structural origin of the cryogenic relaxations in poly (ethylene
terephthalate). J. Polym. Sci. Part A-2: Polym. Phys. 9(8), 1345–1369 (1971)
25. Hiltner, A., Baer, E.: A dislocation mechanism for cryogenic relaxations in crystalline
polymers. Polym. J. 3(3), 378–388 (1972)
26. Arisawa, H., Yano, O., Wada, Y.: Dielectric loss of poly (vinylidene fluoride) at low
temperatures and effect of poling on the low temperature loss. Ferroelectrics 32(1), 39–41
(1981)
27. Yano, O., Wada, Y.: Dynamic mechanical and dielectric relaxations of polystyrene below
the glass temperature. J. Polym. Sci. Part A-2: Polym. Phys. 9(4), 669–686 (1971)
28. Shimizu, K., Yano, O., Wada, Y.: Dielectric relaxations in polymers with pendent phenyl or
pyridine groups at temperatures from 4 K to 80 K. J. Polym. Sci. Polym. Phys. Ed. 13(12),
2357–2368 (1975)
29. Yano, O., Yamaoka, H.: Cryogenic properties of polymers. Prog. Polym. Sci. 20(4),
585–613 (1995)
30. Barucci, M., Gottardi, E., Peroni, I., Ventura, G.: Low temperature thermal conductivity of
Kapton and Upilex. Cryogenics 40(2), 145–147 (2000)
31. Berman, R. (ed.) Thermal Conduction in Solids. Oxford University Press, Oxford (1976)
32. Franz, R., Wiedemann, G.: Ueber die Wärme-Leitungsfähigkeit der Metalle. Ann. Phys.
165(8), 497–531 (1853)
33. Kittel, C. (ed.): Introduction to Solid State Physics, 8th edn. Wiley, New York (2005)
34. Gloos, K., Mitschka, C., Pobell, F., Smeibidl, P.: Thermal conductivity of normal and
superconducting metals. Cryogenics 30(1), 14–18 (1990)
35. Hust, J., Sparks, L.: Lorenz Ratios of Technically Important Metals and Alloys, vol. 634.
US Government printing office, Washington (1973)
36. Pobell, F.: Matter and Methods at Low Temperatures. Springer, Berlin (2007)
37. Callaway, J., Wang, C.: Energy bands in ferromagnetic iron. Phys. Rev. B 16(5), 2095
(1977)
38. Woodcraft, A.L., Barucci, M., Hastings, P.R., Lolli, L., Martelli, V., Risegari, L., Ventura,
G.: Thermal conductivity measurements of pitch-bonded graphites at millikelvin
temperatures: finding a replacement for AGOT graphite. Cryogenics 49(5), 159–164 (2009)
39. White, G., Meeson, P.: Experimental Techniques in Low-Temperature Physics. Clarendon
40. Zaitlin, M.P., Anderson, A.: Phonon thermal transport in noncrystalline materials. Phys.
Rev. B 12(10), 4475 (1975)
41. Anderson, P.W., Halperin, B., Varma, C.M.: Anomalous low-temperature thermal
properties of glasses and spin glasses. Phil. Mag. 25(1), 1–9 (1972)
(1972)
43. Lubchenko, V., Wolynes, P.G.: Intrinsic quantum excitations of low temperature glasses.
Phys. Rev. Lett. 87(19), 195901 (2001)
44. Lubchenko, V., Wolynes, P.G.: The origin of the boson peak and thermal conductivity
plateau in low-temperature glasses. Proc. Natl. Acad. Sci. 100(4), 1515–1518 (2003)
45. Talon, C., Zou, Q., Ramos, M., Villar, R., Vieira, S.: Low-temperature specific heat and
thermal conductivity of glycerol. Phys. Rev. B 65(1), 012203 (2001)
46. Parshin, D.: Interactions of soft atomic potentials and universality of low-temperature
properties of glasses. Phys. Rev. B 49(14), 9400 (1994)
47. Buchenau, U., Galperin, Y.M., Gurevich, V., Parshin, D., Ramos, M., Schober, H.:
Interaction of soft modes and sound waves in glasses. Phys. Rev. B 46(5), 2798 (1992)
48. Reese, W.: Thermal properties of polymers at low temperatures. J. Macromol. Sci. Chem.
3(7), 1257–1295 (1969)
49. Choy, C., Greig, D.: The low temperature thermal conductivity of isotropic and oriented
polymers. J. Phys. C: Solid State Phys. 10(2), 169 (1977)
50. Pobell, F. (ed.) Matter and Methods at Low Temperature, 2nd edn. Springer, Berlin (1991)
51. Cahill, D.G., Watson, S.K., Pohl, R.O.: Lower limit to the thermal conductivity of
disordered crystals. Phys. Rev. B 46(10), 6131 (1992)
52. Roy, R., Komarneni, S., Parker, J., Thomas, G.: Nanophase and Nanocomposite Materials.
Mater. Res. Soci. 241 (1984)
53. Feynman, R.P.: There’s plenty of room at the bottom. Eng. Sci. 23(5), 22–36 (1960)
54. Advani, S.G.: Processing and Properties of Nanocomposites. World Scientific, Singapore
(2007)
55. Green, C., Vaughan, A.: Nanodielectrics-How Much Do We Really Understand?[Feature
Article]. IEEE Electr. Insul. Mag. 24(4), 6–16 (2008)
56. Tait, H.: 5 Thousand Years of Glass. University of Pennsylvania Press, Philadelphia (2004)
57. Vaughan, A.: Raman nanotechnology-the Lycurgus Cup-letter to the editor. IEEE Electr.
Insul. Mag. 24(6), 4 (2008)
58. Andritsch, T.: Epoxy based nanocomposites for high voltage DC applications. Synthesis,
dielectric properties and space charge dynamics. PhD thesis, Delft University of
Technology (2010)
59. Gupta, R.K., Kennel, E., Kim, K.-J.: Polymer Nanocomposites Handbook. CRC Press
(2010)
60. Mark, J., Wen, J.: Inorganic-organic composites containing mixed-oxide phases. In:
Macromolecular Symposia 1995, pp. 89–96. Wiley Online Library (1995)
61. Carter, L.W., Hendricks, J.G., Bolley, D.S.: U.S. Patent 2.531.396. USA Patent
62. Iijima, S.: Helical microtubules of graphitic carbon. Nature 354(6348), 56–58 (1991)
63. Tans, S.J., Devoret, M.H., Dai, H.J., Thess, A., Smalley, R.E., Geerligs, L.J., Dekker, C.:
Individual single-wall carbon nanotubes as quantum wires. Nature 386(6624), 474–477
(1997)
64. Twardowski, T.E., Twardowski, T.A.: Introduction to Nanocomposite Materials: Properties,
Processing, Characterization. DEStech publications, Inc U.S.A (2007)
65. Liao, J., Ren, Y., Xiao, T., Mai, Y., Yu, Z.: Polymer Nanocomposites. Woodhead
Publishing Limited, Cambridge (2006)
66. Pissis, P., Kotsilkova, R.: Thermoset Nanocomposites for Engineering Applications. Rapra
Technology, UK (2007)
67. Kango, S., Kalia, S., Celli, A., Njuguna, J., Habibi, Y., Kumar, R.: Surface modification of
inorganic nanoparticles for development of organic-inorganic nanocomposites–a review.
Prog. Polym, Sci (2013)
68. Kurimoto, M., Watanabe, H., Kato, K., Hanai, M., Hoshina, Y., Takei, M., Okubo, H.:
Dielectric properties of epoxy/alumina nanocomposite influenced by particle dispersibility.
In: Electrical Insulation and Dielectric Phenomena, 2008. CEIDP 2008. IEEE Annual
Report Conference on 2008, pp. 706–709 (2008)
69. Nelson, J.K.: Dielectric Polymer Nanocomposites. Springer, New York (2010). http://link.
springer.com/book/10.1007/978-1-4419-1591-7
References 165
70. Singha, S., Thomas, M.J.: Polymer composite/nanocomposite processing and its effect on
the electrical properties. In: Electrical Insulation and Dielectric Phenomena, 2006 IEEE
Conference on 2006, pp. 557–560 (2006)
71. Yung, K., Wang, J., Yue, T.: Thermal management for boron nitride filled metal core
printed circuit board. J. Compos. Mater. 42(24), 2615–2627 (2008)
72. Levering, A.W.: Interphases in Zirconium Silicate Filled High Density Polyethylene and
Polypropylene (1995)
73. Schulz, M.J., Kelkar, A.D., Sundaresan, M.J.: Nanoengineering of Structural, Functional
and Smart Materials. CRC Press, Boca Raton (2005)
74. Capek, I.: Nanocomposite Structures and Dispersions, vol. 27. Access Online via Elsevier
(2006)
75. Roy, M., Nelson, J., MacCrone, R., Schadler, L., Reed, C., Keefe, R.: Polymer nanocomposite
dielectrics-the role of the interface. IEEE Trans. Dielectr. Electr. Insul. 12(4), 629–643
(2005). http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1511089&tag=1
76. Smith, R., Liang, C., Landry, M., Nelson, J., Schadler, L.: The mechanisms leading to the
useful electrical properties of polymer nanodielectrics. IEEE Trans. Dielectr. Electr. Insul.
15(1), 187–196 (2008)
77. Xanthos, M.: Polymers and Polymer Composites. Functional Fillers for Plastics, pp. 1–16.
Wiley, New York (2005)
78. Godovsky, Y.K.: Thermodynamic Behavior of Solid Polymers in Plastic Deformation and
Cold Drawing. Springer, Berlin (1992)
79. Tekce, H.S., Kumlutas, D., Tavman, I.H.: Effect of particle shape on thermal conductivity of
copper reinforced polymer composites. J. Reinf. Plast. Compos. 26(1), 113–121 (2007)
80. Hansen, D., Bernier, G.: Thermal conductivity of polyethylene: the effects of crystal size,
density and orientation on the thermal conductivity. Polym. Eng. Sci. 12(3), 204–208 (1972)
81. Mark, J.E.: Physical Properties of Polymers Handbook. Springer, Berlin (2007). http://link.
springer.com/book/10.1007/978-0-387-69002-5
82. Kesava Reddy, M., Subramanyam Reddy, K., Yoga, K., Prakash, M., Narasimhaswamy, T.,
Mandal, A., Lobo, N.P., Ramanathan, K., Rao, D.S., Krishna Prasad, S.: Structural
characterization and molecular order of rodlike mesogens with three-and four-ring core by
XRD and 13C NMR spectroscopy. J. Phys. Chem. B 117(18), 5718–5729 (2013)
83. Ekstrand, L., Kristiansen, H., Liu, J.: Characterization of thermally conductive epoxy nano
composites. In: Electronics Technology: Meeting the Challenges of Electronics Technology
Progress, 2005. 28th International Spring Seminar on 2005, pp. 35–39 (2005)
84. Kim, W., Bae, J.W., Choi, I.D., Kim, Y.S.: Thermally conductive EMC (Epoxy Molding
Compound) for microelectronic encapsulation. Polym. Eng. Sci. 39(4), 756–766 (1999)
85. Hsieh, C.Y., Chung, S.L.: High thermal conductivity epoxy molding compound filled with a
combustion synthesized AlN powder. J. Appl. Polym. Sci. 102(5), 4734–4740 (2006)
86. Wong, C., Bollampally, R.S.: Comparative study of thermally conductive fillers for use in
liquid encapsulants for electronic packaging. IEEE Trans. Adv. Packag. 22(1), 54–59 (1999)
87. Okamoto, T., Sawa, F., Tomimura, T., Tanimoto, N., Hishida, M., Nakamura, S.: Properties
of high-thermal conductive composite with two kinds of fillers. In: Properties and
Applications of Dielectric Materials, 2007. Proceedings of the 7th International Conference
on 2003, pp. 1142–1145 (2007)
88. Xu, Y., Chung, D.: Increasing the thermal conductivity of boron nitride and aluminum
nitride particle epoxy-matrix composites by particle surface treatments. Compos. Interfaces
7(4), 243–256 (2000)
89. Han, Z., Wood, J., Herman, H., Zhang, C., Stevens, G.: Thermal properties of composites
filled with different fillers. In: Electrical Insulation, 2008. ISEI 2008. Conference Record of
the 2008 IEEE International Symposium on 2008, pp. 497–501 (2008)
90. McCullough, R.L.: Generalized combining rules for predicting transport properties of
composite materials. Compos. Sci. Technol. 22(1), 3–21 (1985)
91. Progelhof, R., Throne, J., Ruetsch, R.: Methods for predicting the thermal conductivity of
composite systems: a review. Polym. Eng. Sci. 16(9), 615–625 (1976)
92. Tavman, I.: Effective thermal conductivity of isotropic polymer composites. Int. Commun.
Heat Mass Transf. 25(5), 723–732 (1998)
93. Agarwal, S., Khan, M.M.K., Gupta, R.K.: Thermal conductivity of polymer nanocomposites
made with carbon nanofibers. Polym. Eng. Sci. 48(12), 2474–2481 (2008)
94. Maxwell, J.C.: A Treatise on Electricity and Magnetism, vol. 1. Clarendon Press, Oxford
(1881)
95. Pal, R.: On the Lewis-Nielsen model for thermal/electrical conductivity of composites.
Compos. A Appl. Sci. Manuf. 39(5), 718–726 (2008)
96. Fricke, H.: A mathematical treatment of the electric conductivity and capacity of disperse
systems I. The electric conductivity of a suspension of homogeneous spheroids. Phys. Rev.
24(5), 575 (1924)
97. Bruggeman, V.D.: Berechnung verschiedener physikalischer Konstanten von heterogenen
Substanzen. I. Dielektrizitätskonstanten und Leitfähigkeiten der Mischkörper aus isotropen
Substanzen. Ann. Phys. 416(7), 636–664 (1935)
98. Lee, E.S., Lee, S.M., Shanefield, D.J., Cannon, W.R.: Enhanced thermal conductivity of
polymer matrix composite via high solids loading of aluminum nitride in epoxy resin.
J. Am. Ceram. Soc. 91(4), 1169–1174 (2008)
99. Hill, R.F., Supancic, P.H.: Thermal conductivity of platelet-filled polymer composites.
J. Am. Ceram. Soc. 85(4), 851–857 (2002)
100. Stevens, G., Herman, H., Han, J., Wood, J., Mitchell, A., Thomas, J.: The role of nano and
micro fillers in high thermal conductivity electrical insulation systems. In: 11th Insucon
Conference, Birmingham, UK 2009, pp. 286–291
101. Tsao, G.T.-N.: Thermal conductivity of two-phase materials. Ind. Eng. Chem. 53(5),
395–397 (1961)
102. Cheng, S., Vachon, R.: The prediction of the thermal conductivity of two and three phase
solid heterogeneous mixtures. Int. J. Heat Mass Transf. 12(3), 249–264 (1969)
103. Sundstrom, D.W., Lee, Y.D.: Thermal conductivity of polymers filled with particulate
solids. J. Appl. Polym. Sci. 16(12), 3159–3167 (1972)
104. Hamilton, R.: Thermal conductivity of two phase materials. Dissertation, University of
Oklahoma (1960)
105. Hamilton, R., Crosser, O.: Thermal conductivity of heterogeneous two-component systems.
Ind. Eng. Chem. Fundam. 1(3), 187–191 (1962)
106. Hatta, H., Taya, M.: Effective thermal conductivity of a misoriented short fiber composite.
J. Appl. Phys. 58(7), 2478–2486 (1985)
107. Meredith, R.E., Tobias, C.W.: Conduction in heterogeneous systems. Advances in
electrochemistry and electrochemical engineering 2(II), 15–47 (1962)
108. Nielsen, L.E.: Mechanical properties of particulate-filled systems. J. Compos. Mater. 1(1),
100–119 (1967)
109. Lewis, T., Nielsen, L.: Dynamic mechanical properties of particulate-filled composites.
J. Appl. Polym. Sci. 14(6), 1449–1471 (1970)
110. Landel, R.F.: Mechanical Properties of Polymers and Composites, vol. 90. CRC Press,
(1994)
111. Halpin, J.: Stiffness and expansion estimates for oriented short fiber composites. J. Compos.
Mater. 3(4), 732–734 (1969)
112. Agari, Y., Uno, T.: Thermal conductivity of polymer filled with carbon materials: effect of
conductive particle chains on thermal conductivity. J. Appl. Polym. Sci. 30(5), 2225–2235
(1985)
113. Agari, Y., Uno, T.: Estimation on thermal conductivities of filled polymers. J. Appl. Polym.
Sci. 32(7), 5705–5712 (1986)
114. Agari, Y., Ueda, A., Nagai, S.: Thermal conductivity of a polymer composite. J. Appl.
Polym. Sci. 49(9), 1625–1634 (1993)
115. Russell, H.: Principles of heat flow in porous insulators*. J. Am. Ceram. Soc. 18(1–12), 1–5
(1935)
References 167
116. Topper, L.: Industrial design data—analysis of porous thermal insulating materials. Ind.
Eng. Chem. 47(7), 1377–1379 (1955)
117. Jefferson, T., Witzell, O., Sibbitt, W.: Thermal conductivity of graphite—silicone oil and
graphite-water suspensions. Ind. Eng. Chem. 50(10), 1589–1592 (1958)
118. Springer, G.S., Tsai, S.W.: Thermal conductivities of unidirectional materials. J. Compos.
Mater. 1(2), 166–173 (1967)
119. Budiansky, B.: Thermal and thermoelastic properties of isotropic composites. J. Compos.
Mater. 4(3), 286–295 (1970)
120. Baschirow, A., Selenew, J.: Thermal conductivity of composites. Plaste Kaut 23, 656 (1976)
121. McGee, S., McGullough, R.: Combining rules for predicting the thermoelastic properties of
particulate filled polymers, polymers, polyblends, and foams. Polym. Compos. 2(4),
149–161 (1981)
122. Privalko, V., Novikov, V.: Model treatments of the heat conductivity of heterogeneous
polymers. In: Thermal and Electrical Conductivity of Polymer Materials, pp. 31–77.
Springer (1995)
123. Dul’Nev, G., Novikov, V.: Conductivity determination for a filled heterogeneous system.
J. Eng. Phys. Thermophys. 37(4), 1184–1187 (1979)
124. Heron, J.-S., Souche, G.M., Ong, F.R., Gandit, P., Fournier, T., Bourgeois, O.: Temperature
modulation measurements of the thermal properties of nanosystems at low temperatures.
J. Low Temp. Phys. 154(5–6), 150–160 (2009)
125. Huang, C., Fu, S., Zhang, Y., Lauke, B., Li, L., Ye, L.: Cryogenic properties of SiO2/epoxy
nanocomposites. Cryogenics 45(6), 450–454 (2005)
126. Martelli, V., Toccafondi, N., Ventura, G.: Low-temperature thermal conductivity of Nylon-
6/Cu nanoparticles. Physica B 405(20), 4247–4249 (2010)
127. Batchelor, G., O’Brien, R.: Thermal or electrical conduction through a granular material.
Proc. R. Soc. Lond. A Math. Phys. Sci. 355(1682), 313–333 (1977)
128. Risegari, L., Barucci, M., Bucci, C., Fafone, V., Gorla, P., Giuliani, A., Olivieri, E., Pasca,
E., Pirro, S., Quintieri, L.: Use of good copper for the optimization of the cooling down
procedure of large masses. Cryogenics 44(3), 167–170 (2004)
129. Van Sciver, S.W., Nellis, M.N., Pfotenhauer, J.: Thermal and electrical contact
Conductance between metals at low temperatures. In: Proceedings Space Cryogenics
Workshop 1984, Berlin (DE) (1984)
130. Peterson, R., Anderson, A.: The Kapitza thermal boundary resistance. J. Low Temp. Phys.
11(5–6), 639–665 (1973)
131. Little, W.: The transport of heat between dissimilar solids at low temperatures. Can. J. Phys.
37(3), 334–349 (1959)
132. Radebaugh, R.: Thermal conductance of indium solder joints at low temperatures. Rev. Sci.
Instrum. 48, 93 (1977)
133. Gmelin, E., Asen-Palmer, M., Reuther, M., Villar, R.: Thermal boundary resistance of
mechanical contacts between solids at sub-ambient temperatures. J. Phys. D Appl. Phys.
32(6), R19 (1999)
134. Didschuns, I., Woodcraft, A., Bintley, D., Hargrave, P.: Thermal conductance
measurements of bolted copper to copper joints at sub-Kelvin temperatures. Cryogenics
44(5), 293–299 (2004)
135. Bintley, D., Woodcraft, A.L., Gannaway, F.C.: Millikelvin thermal conductance
measurements of compact rigid thermal isolation joints using sapphire–sapphire contacts,
and of copper and beryllium–copper demountable thermal contacts. Cryogenics 47(5),
333–342 (2007)
136. Fritzsche, H.: Resistivity and hall coefficient of antimony-doped germanium at low
temperatures. J. Phys. Chem. Solids 6(1), 69–80 (1958)
137. Rosenberg, H.M.: The thermal conductivity of metals at low temperatures. Philos. Trans. of
the R. Soc. A Math. Phys. Sci. 247(933), 441–497 (1955)
138. Berman, R., Foster, E., Ziman, J.: The thermal conductivity of dielectric crystals: the effect
of isotopes. Proc. R. Soc. Lond. A 237(1210), 344–354 (1956)
139. Nathan, B., Lou, L., Tait, R.: Low temperature thermal properties of mixed crystal KBr KI.
Solid State Commun. 19(7), 615–617 (1976)
140. Guckel, H.: Silicon microsensors: construction, design and performance. Microelectron.
Eng. 15(1), 387–398 (1991)
141. Locatelli, M., Arnaud, D., Routin, M.: Thermal conductivity of some insulating materials
materials below 1 K. Cryogenics 16(6), 374–375 (1976)
142. Morelli, D., Doll, G., Heremans, J., Peacor, S., Uher, C., Dresselhaus, M., Cassanho, A.,
Gabbe, D., Jenssen, H.: Thermal conductivity of single crystal lanthanum cuprates at very
low temperature. Solid State Commun. 77(10), 773–776 (1991)
144. Fukushima, K., Takahashi, H., Takezawa, Y., Hattori, M., Itoh, M., Yonekura, M.: High
thermal conductive epoxy resins with controlled high-order structure [electrical insulation
applications]. In: Electrical Insulation and Dielectric Phenomena, 2004. CEIDP’04. 2004
Annual Report Conference on 2004, pp. 340–347 (2004)
145. Scott, T.A., de Bruin, J., Giles, M.M., Terry, C.: Low-temperature thermal properties of
nylon and polyethylene. J. Appl. Phys. 44(3), 1212–1216 (1973)
146. Ekin, J. (ed.) Experimental Techniques for Low Temperature Measurements. Oxford
Chapter 8
How to Measure Thermal Conductivity
Abstract The methods to measure the thermal conductivity at low temperature are
described: the steady-state techniques, (Sect. 2.2); the 3x technique (Sect. 2.3);
and the thermal diffusivity measurement (Sect. 2.4). Each of these techniques has
its own advantages as well as its inherent limitations, with some techniques more
appropriate to specific sample geometry, such as the 3x technique for thin films
which is discussed in detail in Sect. 2.4.2. The radial flux method is reported in
Sect. 2.2.4, the laser flash diffusivity method in Sect. 2.4.1 and the ‘‘pulsed power
or Maldonado technique’’ in Sect. 2.3.2.
8.1 Introduction
The methods to measure electrical conductivity at low temperatures do not differ

from those at room temperature, except that in the former case, much lower
powers are involved. As a consequence, the signal-to-noise ratio may become
quite small and filtering techniques are often needed (see, e.g., [1]). The latter are
also used when measuring thermal conductivity. In Sect. 8.5.1, we will report a
particular example of measurement of electrical conductivity in which all the
aforementioned problems are involved. However, electrical measurements are
usually simpler than thermal ones. In fact, a lot of thermal conductivity data for
metals have been obtained from electrical conductivity values by applying the
Wiedemann–Franz law (see Sect. 7.6.2). An illuminating confirmation of that can
be found in [2, 3] where the electrical origin for most of thermal data is explicit.
Thus, it is evident that the accurate measurement of the thermal conductivity of
bulk materials may be more complex. For instance, loss terms of the heat input
intended to flow through the sample usually exist and can be very difficult to
quantify. This chapter provides an overview of the more typical measurement
techniques used to determine the thermal conductivity of bulk materials and thin
films. This overview is not intended to be a complete description of all the
available measurement techniques, but it should provide an introduction and

170 8 How to Measure Thermal Conductivity
summary of the characterization and measurement techniques of thermal con-

ductivity and give an extensive reference set for a more deep analysis of the
concepts and techniques.
Many methods exist for the measurement of thermal conductivity of a material,
for example, the steady state technique (Sect. 8.2), the 3x technique (Sect. 8.3),
and thermal diffusivity measurement (Sect. 8.4). Each of these techniques has its
own advantages as well as its inherent limitations, with some techniques more
appropriate to specific sample geometry, such as the 3x technique for thin films
which is discussed in detail in Sect. 8.4.2. The methods presented will include the
more common steady state method (Sect. 8.2), the radial flux method (Sect. 8.2.2),
the laser flash diffusivity method (Sect. 8.4.1), and the ‘‘pulsed power’’ or
Maldonado technique (Sect. 8.3.2).
Thermal conductivity measurements are difficult to make with relatively high
accuracy, certainly better than within 5 %. Many excellent texts about techniques
discuss many of the corrections and potential errors one must consider in detail,
e.g., [4–7].
Thermal conductivity measurements can be carried out by two types of
methods: steady state methods and transient methods. Steady state conditions refer
to the constant temperature at each point of the sample, i.e., not a function of time.
The transient methods are used to record measurements during the process of
heating up or the cooling down of a sample. These methods have the advantage of
giving quicker measurements than the steady state methods.
In particular, a lot of different practical solutions are adopted because of a wide
range of thermal conductivity values of materials; it is often necessary to differ-
entiate the setup depending on the particular behavior of the investigated sample.
8.2 Steady State Techniques
In steady conditions, the power qQ/qt flowing through a sample is expressed by

(3.12).
For a constant sample of section A and length L, and a sample ending at
temperatures T0 and T1, we obtain
R T1 ZT1
T0 jðTÞdT
P ¼ ffiQ_ ¼ ffi R L dx ¼g jðTÞdT ð8:1Þ
0 AðxÞ T0
where g = A/L is the so-called geometrical factor, and P is the power that flows
through the sample, causing the temperature difference T1 – T0.
In practice, the temperature in a steady state system is maintained by a heat
source, typically an electrical heater. The temperature difference is measured
between two points with a separation distance, x, inside the test specimen.
8.2 Steady State Techniques 171
Methods are usually divided into axial (or longitudinal) and radial. Axial flow
methods have been long established and have provided some of the most con-
sistent results with the highest accuracy near room temperature by means of the so-
called guarded hot plate apparatus (Sect. 8.2.1.2), whereas the concentric cylinder
method is often used for radial systems (Sect. 8.2.2). Steady state measuring
methods provide accurate results, but they are time consuming.
8.2.1 Longitudinal Flux Method
8.2.1.1 Potentiometric Method
This method is best suited to samples with small g, like a wire or a thin rod,
because the longitudinal flux hypothesis is easier to satisfy (radiation losses
negligible compared to the power P). In the potentiometric method (Fig. 8.1), one
end of the sample is thermally connected to a thermal bath TB; at the other end of
the sample, a heater Ha is thermally connected (usually glued); in this configu-
ration, a thermal flux (mainly) along the length direction is realized. In the
hypothesis of negligible losses and uniform temperature distribution on the trans-
verse sections of the sample, knowing P and measuring T in two points, the
thermal conductivity is determined by (8.1).
This type of measurement is carried out in vacuum to avoid convection heat
exchange; a careful wiring must be done to avoid parallel thermal flux. At very low
temperatures, any mechanical connection between the sample and other parts can
give rise to thermal steps which may derail the measure (Kapitza impedance, see,
e.g., [8]).
When measuring a thermal conductor, thermal contact resistances may be
comparable or even larger than the thermal resistance of the sample. In these cases,
a method called ‘‘potentiometric’’ (as in the electrical case), shown in Fig. 8.1, is
used.
In the potentiometric method, a known power P is supplied to the sample and a
DT = T1 - T0 is measured at two points at distance L along the sample (four
probe method, see Fig. 8.1). When the two points are the ends of the sample, one
gets the simpler two probe method, often used for low conductivity samples. Two
thermometers calibrated over the measurement range are needed in the former
setup, whereas in the latter case, one thermometer calibrated over the full range
and one at a single temperature are necessary [1]. In both setups, the power
shunted through wiring to heater and thermometers must be negligible compared
to P. This goal is achieved by the use of low conductance wires (e.g., manganin or
NbTi at very low temperatures). The thermal bath temperature Tb can be varied by
an additional heater Hb (see Fig. 3.18) and the thermal conductivity can be cal-
culated by (8.1).
Fig. 8.1 Scheme of the

potentiometric method for the
measurement of thermal
conductivity
The temperature To is usually kept constant. However, when the form of the
function j(T) is known, T0 may be allowed to move, saving time in the mea-
surements (see, e.g., Ref. [9]), but complicating data analysis.
The choice of g depends on the particular material studied, the time accepted to
reach thermal equilibrium and the thermometer sensitivity. With a good thermal
conductor, g should be as high as possible in order to get a measurable DT with a
small P. Another constraint is the experimental space available in the cryostat.
This technique is quite general at very low temperatures where radiation and
conduction losses are negligible. Examples of two probe and four probe methods
are reported in Sect. 8.5.2.1 and, e.g., in Ref. [10].
8.2.1.2 Guarded Hot Plate Method
The guarded hot plate method is very versatile and commonly used for deter-
mining the thermal conductivity of low conductivity materials such as glass,
ceramics and polymers. It is used between about 80 and 800 K, and uncertainty
related to thermal conductivity measurements is about 2 % [11].
A scheme of the guarded hot plate method is shown in Fig. 8.2.
The external plates (Cu or Al) are kept at the constant temperature of a thermal
bath. A heating plate supplies a uniform and constant heating power P0 which is
transmitted to the external plates through the sample of conductivity j that we
want to measure. A guard ring around the central zone is heated by a power P1
such that the guard ring temperature is the same as that of the heating plate; thus,
all the power P0 crosses the sample. Note that P1 [ P0. The guard ring and the
heating plate are not in contact. Thermocouples are placed on the two sides of the
sample to measure DT1 and DT2.
8.2 Steady State Techniques 173
Fig. 8.2 Scheme of the

guarded hot plate method
The thermal conductivity of the sample is
LP0
jðTÞ ¼ ð8:2Þ
AðDT1 þ DT2 Þ
where A is the heating plate area and L the sample length.

The major limitation is that this method may be used only when the contact
resistances can be neglected in comparison with the thermal resistance of the
sample. This constraint is generally satisfied above 80 K.
8.2.2 Radial Flux Method or Cylinder Method
The longitudinal heat flow method can be satisfactory at very low temperatures,
but serious errors can occur at higher temperatures due to radiation losses directly
from the heater and from the sample surface. In the radial heat flow method, heat is
applied internally to the sample, minimizing radiation losses from the heat source.
The radial flow method has been applied to solids having a wide range of thermal
conductivities. Since radial flow methods are relatively more difficult to apply than
longitudinal methods, they are seldom employed at very low temperatures.
Internal sample heating has been accomplished in a variety of sample geom-
etries, including imbedding in the center of a hollow sample, and by direct elec-
trical heating of the sample itself. The symmetry of the sample geometry must
correspond to the geometry of the heater and allows for the inclusion of the heater.
Five classes of apparatus in radial methods are mentioned in Ref. [5]. We will
only describe the simplest, i.e., the cylindrical geometry with a central source (or
sink) of power where an ‘infinite’ length is assumed and therefore without end
guards.
This experimental geometry requires rather large samples, but radiation losses
are minimized, thus making it appropriate for high temperatures.
In the cylindrical symmetry (see Fig. 8.3), heat is generated along the axis of a
cylinder. In steady state conditions, the radial temperature profile is measured at
two different radii. For heat flow in a cylinder between radii r1 and r2, assuming
negligible longitudinal heat loss, thermal conductivity j is [7]
Fig. 8.3 Configuration for

measuring thermal
conductivity using a radial
flow technique [12]

ln r2=r1
jðTÞ ¼ P ð8:3Þ
2pLDT
where P is power input, L is sample length, DT is the temperature difference

between the thermocouples, and r1 and r2 are the radial positions of the inner and
outer thermocouples, respectively. The use of a cylindrical sample has the
advantage that only the ratio of inner and outer radius has to be known (see 8.3):
this fact avoids accounting for thermal contraction.
In [12], a combination of linear and radial methods were used to measure the
thermal conductivity of germanium from 3 to 1,020 K.
8.3 Transient Methods
8.3.1 The 3x Method
The 3x method was originally proposed by Corbino [13, 14] who discovered the
small third-harmonic voltage component while applying an alternating current
through a heater. Later, the method was used to measure the specific heat of the
heater itself [15, 16]. This technique became popular after it was used to measure
the specific heat of substrate materials, [17–19] where a one-dimensional heater-
on-substrate conduction model was set up for the case of an ideal, infinite, and
planar heater. A similar model was used for the simultaneous measurement of both
thermal conductivity and specific heat [20]. A further contribution is due to Cahill
8.3 Transient Methods 175
et al. [21, 22] (see later) on obtaining an analytical solution for a vanishingly thin
but finite-width heater. As we shall see, the analytical solution for a line heater on
a substrate [23] was integrated to give the solution for the heater of finite width in
integral form. An approximation for the integral solution, often used, was also
obtained for small frequencies [21]. The integral solution for finite-width heater
was investigated by a few authors and a formula was derived for heat capacity
measurement in the same frequency range as the conductivity measurement [24].
Moon et al. [25] showed that a much simpler formula exists at the high-frequency
limit which can be used for specific heat measurement. The integral solutions have
also been extended to measure the thermal properties of thin films [26–33] other
than heater-on-substrate configurations. Several authors have also developed the
3x method for suspended wires, such as nanowires [34] and nanotubes [35–38].
Few studies have solved the heater-on-substrate problem, which is the basis of
the 3x method, by starting from anything other than the original line heater
solution [21, 22]. The two-dimensional conduction problem for the 3x method
with an imposed heat flux was solved using Green’s function [39]. For the three-
dimensional conduction problem, see [40].
Note that all of these studies have neglected the thickness of the heater. Birge
and Nagel [19] state, without any analysis, that ‘‘the heater thickness can be
neglected if it is small compared to the penetration depth and if the heat stored in
the heater itself is feasible.’’ Further studies are reported in Refs. [41–44].
Although the 3x method is extensively used for thermal conductivity mea-
surement, the formulas generally used are based on the zero heater-thickness
approximation. It is therefore necessary to provide a complete analysis of the
problem to determine the limits of its applicability. In the very interesting Ref.
[45], a two-dimensional conduction model for the heater-on-substrate 3x method
is developed in the case of finite heater thickness. The analytical solution is
obtained using the method of separation of variables.
The 3x method has also been used for measuring the thermal conductivity of
dielectric solids down to 30 K [21]. An example of implementation of the elec-
tronics for a measure of thermal conductivity by the 3x method is described in
Ref. [46].
Let us now summarize Cahill’s calculations [21] to find the solution of the
integral form for thin films. This solution is commonly used to measure thermal
properties of the substrate material. In such an application, the 3x method has
several advantages over other methods, e.g., it reduces the equilibration time to
few minutes. Moreover, the effect of black body radiation is reduced due to the
small surface area of the metallic lines [21, 47].
An AC current I(t) = I0cos(xt) at angular frequency x is passed through a
metallic line resistance. The metallic line acts as both a resistive heater and a
thermometer. Due to Joule’s effect, heat will be generated in the metallic line
producing temperature oscillations at angular frequency 2x. Consequently, the
resistance of the metallic line Rml changes, following the equation [48]
Rml ¼ R0 ð1 þ qT DTÞ ð8:4Þ
where qT is the temperature coefficient of resistance in K-1, R0 is the resistance of

the metallic line at T0, and Rml is its resistance at T0 + DT.
By multiplying the small resistance fluctuations by the alternating current I(t), a
voltage at frequency 3x is generated.
Figure 8.4 shows the metallic line (deposited on the substrate) having two
contact pads through which the AC current passes.
The temperature oscillation DT measured across the metallic line is [21]
Z1
Prms sin2 ðkbÞ
DT ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi dk ð8:5Þ
pj ðkbÞ2 k2 þ q2
0
where Prms, j and b are the input power per meter of length in W/m, the thermal
conductivity of the substrate under study in W/Km, and the half width of the
metallic line, respectively. The integration variable k refers to the Fourier space.
This is useful to introduce because convolutions in the real space x can be replaced
with multiplications. In (8.5), q represents the complex wavenumber of the thermal
wave in rad/m and is given by
sffiffiffiffiffiffiffi
2ix
q¼ ð8:6Þ
w
where w is the thermal diffusivity (in m2/s) of the substrate (see 3.14).
Equation (8.5) stands on several assumptions: the substrate thickness ts is taken
as semi-infinite, the metallic line length is infinite, and the thermal conductivity is
isotropic over the specimen.
The thermal penetration depth kp is a measure of how deep thermal waves
penetrate into the substrate. It is defined as [42]
qffiffiffiffiffiffiffiffiffiffiffi
kp ¼ w=2x : ð8:7Þ
For a thermal penetration depth larger than five times the half width of the metallic
line (k [ 5b) and smaller than one-fifth of the thickness ts (k \ ts/5), Cahill [21]
determined a linear behavior for the real part of temperature oscillations as a
function of frequency. In this range of frequencies, [21] the temperature oscillation
component in-phase with current (real part) decays logarithmically with 2x,
whereas the component that is p/2 out of phase with the current (imaginary part) is
constant over the same range of frequencies.
The third harmonic voltage V3x is related to the temperature oscillation DTAC
by [29]
8.3 Transient Methods 177
Fig. 8.4 Metallic line with

two contact pads deposited on
a substrate
V3x ¼ V0 bDTAC ð8:8Þ
where V0 and b are the voltage and the temperature coefficient of resistance of the
metallic line, respectively. Plotting the in-phase temperature oscillation versus 2x,
the slope in the linear regime can be calculated. Then, the thermal conductivity of
the substrate can be computed as [22]

V03 b ln f2=f1
j¼ ð8:9Þ
4pR0 l ðV3x1 ffi V3x2 Þ
where V3x1 and V3x2 are the third harmonic voltages for frequencies f1 and f2,
respectively, and l is the length of the metallic line.
8.3.2 Pulse Power Method
Traditional methods for measuring thermal conductivity require relatively long

waiting times between measurements to enable the sample to reach steady state
conditions.
With the pulse power method, the bath temperature is slowly drifted, while the
heater current which generates the thermal gradient is pulsed with a square wave
[49]. Maldonado used this powerful technique for the simultaneous measurement
of thermal conductivity and thermoelectric power of a sample. Here, we will only
deal with the thermal conductivity measurements.
A sketch of the experimental setup is shown in Fig. 8.5. Since thermal equi-
librium is never reached, the time between measurements can be reduced. The
experimental setup does not differ from the case of steady state measurements,
except that the heating current is pulsed with a square wave of constant current,
thereby creating small thermal gradients. Figure 8.6 shows the response of the
sample temperature to the pulsed power. The heat balance equation for the heater
in Fig. 8.5 is written as the sum of the current dissipated in the heater and the heat
conducted by the sample
Fig. 8.5 Scheme of the pulse

power technique as described
in Ref. [49]
: dT1
Q ¼ CðT1 Þ ¼ RðT1 ÞI 2 ðtÞ ffi GðT1 ffi T0 Þ ð8:10Þ
dt
where C(T1) is the heat capacity of the heat source, R(T1) is the resistance of the
heater and G is the conductance of the sample.
Since G is a function of temperature, if the temperature difference T1 - T0 is
small compared to the mean sample temperature, it can be considered as a function
of the mean sample temperature. The bath temperature T0 is allowed to drift slowly
(compared to the time scale used for modulating the current) and a periodic
square-wave current with period 2s is applied through the heater, which causes T1
to vary. Maldonado gets a solution of (8.10) by making several simplifications:
since C(T), R(T), and G(T) are smooth functions of T, then T0 is used instead of T1
as the argument of C, R, and G; moreover, an adiabatic approximation is employed
by considering T0 as nearly constant. Calculations can be found in Ref. [49]. The
solution has a saw-tooth form as shown in Fig. 8.4. The difference between the
smooth curves through the maxima and minima yields a relation for the thermal
conductance

RI02 Gs
G¼ tanh ð8:11Þ
DTpp 2C
where DTpp is the temperature gradient peak-to-peak (see Fig. 8.6).

The overall accuracy is reported by Maldonado to be better than 5 %, with the
main error sources being the measurements of DT and the sample geometrical
factor, in evaluating the thermal conductivity from the thermal conductance.
An advantage of this method is that the sample temperature is slowly moved
while the measurement is performed. This saves time in the measurements since
achieving a steady state is not necessary. This technique has recently been
employed in a commercial device produced by Quantum Design [50].
8.4 Thermal Diffusivity Measurements 179
Fig. 8.6 Temperature waveform in response to pulsed power for measuring thermal conductivity
using the pulse power technique. The time dependence of the temperature difference across the
sample shows where the dashed line represents a simulation and the open circles represent
experimental data [49]
8.4 Thermal Diffusivity Measurements
8.4.1 Laser Flash Method
Another technique for measuring the thermal properties of thin film and bulk
samples is the laser flash thermal diffusivity method [51–54]. In this technique, one
face of a sample is irradiated by a laser pulse (pulse time B 1 ms). An IR detector
monitors the temperature rise of the opposite side of the sample. The thermal
diffusivity is calculated from the temperature rise versus time profile. Algorithms
exist for correcting various losses typically present in this measurement. The
thermal conductivity is related to the thermal diffusivity as written in (8.8). At high
temperatures where the heat capacity is a constant, the thermal diffusivity mea-
surement essentially yields the thermal conductivity. However, the use of this
method requires fairly stringent sample preparation requirements: there is very
little flexibility in the required sample geometry (typically thin disks or plates). In
addition, the sample surfaces must be highly emissive to maximize the amount of
thermal energy transmitted from the front surface and to maximize the signal
observed by the IR detector. Usually, this requires the application of a coating of
graphite to the sample surfaces. If good adhesion is not achieved, this coating
procedure can potentially be a source of significant error.
Commercial units are available which allow measurement of thermal diffusivity
at temperatures down to about 77 K, see, e.g., [51–54].
8.4.2 Temperature Wave Method
In this method, [23] typically an AC calorimetric technique, a thin sample is light-

irradiated. Temperature waves are due to periodic fluctuations in the heat fluxes
entering the medium, that is, to the variability of the heat sources. Temperature
waves experience strong attenuation during propagation. Significant dispersion,
that is, a frequency dependence of velocity of propagation, is characteristic of
temperature waves. The attenuation factor of temperature waves is usually
approximately 2p/k, where k is the wavelength. For a monochromatic plane
temperature wave, propagating along a thermally insulated rod of constant cross
section, the relation of k to the oscillation period s and to the thermal diffusivity w
is given by
pffiffiffiffiffiffiffiffi
k¼2 pws: ð8:12Þ
Here, the velocity v of the wave crests is

rffiffiffiffiffiffi
4pw pw
v¼ ¼2 : ð8:13Þ
k s
Thus, the shorter the oscillation period, or the wavelength, the more rapidly tem-
perature waves propagate and attenuate over short distances. The depth of pene-
tration of a plane temperature wave, defined as the distance at which temperature
fluctuations decrease by a factor equal to the Nepero’s number (e % 2.7), is
rffiffiffiffiffi
k ws
¼ ; ð8:14Þ
2p p
that is, the shorter the period, the smaller the depth of penetration. For example,
the depth of penetration of diurnal temperature fluctuations into the ground is
smaller by a factor of nearly 20 than the depth of penetration of seasonal fluctu-
ations. This method can be applied for any thickness of the sample.
The temperature wave method is especially convenient for measuring the
characteristics of pure substances at low temperatures. For example, this method
was tested for samples of nickel, silicon, stainless steel, and alumina in the range
from 50 to 300 lm in thickness [55].
8.5 Examples of Measurements of Electrical and Thermal Conductivity 181
8.5 Examples of Measurements of Electrical and Thermal

Conductivity
8.5.1 Measurement of Electrical Resistivity of Heavily Doped

NTD 31 Germanium at Very Low Temperatures,
and Calculation of Electron-phonon Decoupling
We report here about a very low temperature measurement [56] of electrical and
thermal characteristics of NTD (Neutron Transmutation Doped) 31 Ge thermistors
(used, e.g., in the CUORICINO experiment [57]) and, in particular, about the
thermal decoupling of the charge carriers from phonons (‘‘hot electron effect’’) that
occurs near the working temperature of the detectors (10 mK). The NTD pro-
duction process is described, e.g., in Sect. 2.4.1 and in Ref. [1].
The electrical resistance of a semiconductor sample as a function of the tem-
perature T follows Mott’s law [58]:
n
R ¼ R0 eðT0=T Þ ð8:15Þ
which can be explained by the Variable Range Hopping (VRH) theory [59]. In
(8.15), R0, n and T0 are constants: n depends on the density of states of the
electrons near the Fermi energy and is about 0.5 [60], T0 depends on the dopant
concentration, and R0 mainly depends on the geometry, but also on the stresses of
the sample. NTD Ge resistors obey Mott’s law below 1 K. Resistors with T0
ranging from about 1 up to 200 K have been produced (see Fig. 8.5). Thanks to its
monocrystalline structure, NTD Ge has low specific heat [61]. These two prop-
erties (strong temperature dependence and low specific heat) make NTD Ge
sensors precious as thermometers and, in particular, as sensors for bolometers both
in astronomy [62] and in nuclear physics [63].
In Fig. 8.7, a R(T) curve for a NTD 31 sample is shown in the 20–130 mK
range. A fit based on (8.17), with n = 1/2, gives: R0 = 0.164 X and T0 = 4.1 K.
Below about 100 mK, V–I and P-R curves show nonlinearities that become more
evident as the temperature decreases. This fact can be explained by the Variable
Range Hopping (VRH) theory if a Hot Electron Model (HEM) is introduced [64].
The HEM thermal model schematizes a thermistor as two systems with inde-
pendent heat capacities: the electron system and the phonon system. The electron-
phonon coupling is described by a thermal conductance Ge-ph(Te,Tph) between the
two systems, which depends on both temperatures of the two systems (Fig. 8.6)
given by
R Te
Tph jeffiph dT
Geffiph ðTe ; Tph Þ ¼ : ð8:16Þ
Te ffi Tph
Fig. 8.7 Resistivity of various NTD thermistor series versus temperature. The numbers refer to
the production series of thermistors [60]
The temperature dependence of je-ph (T) is supposed to be of the type
jeffiph ðTÞ ¼ ðb þ 1Þgeffiph T b ð8:17Þ
where ge-ph (expressed in W K-(b+1)) and b are constant.

Similarly, the thermal conductance between the phonons and the thermal bath
(Kapitza conductance, see Sect. 7.6.6) is
R Tph
Ts jphffis dT
Gphffis ðTs ; Tph Þ ¼ ð8:18Þ
Tph ffi Ts
where
jphffis ðTÞ ¼ ða þ 1Þgs T a ð8:19Þ
and a is expected to be about three, see (8.6).

Low temperature measurements were carried out in a small power dilution
refrigerator. The NTD 31 Ge sample was a thermistor 6 9 2.9 9 1.6 mm3, glued
with IMI-7031 varnish onto a copper holder in good thermal contact with the
mixing chamber. The sample was electrically insulated from the copper holder by
a thin layer of mylar. Electrical connections on two opposite surfaces
(2.9 9 1.6 mm2) were made with ultrasonically bonded 50 micron-diameter Au
Fig. 8.8 The experimental R(T) curve for an NTD 31 sample in the 20–130 mK range
wire. Since four wire resistance measurements were performed, two manganin
(100 lm in diameter) wires were soldered to each Au wire. Filters were used at
room temperature to prevent RF pick up. A calibrated RuO2 thermometer was
mounted on the copper holder and two copper shields surrounded the experiment,
Fig. 8.8 shows the dependency of the resistance versus temperature measured in
the NTD 31 Ge sample.
Referring to Fig. 8.9, a power P, flowing through the sample, can be written as
P ¼ Geffiph ðTe ; Tph ÞðTe ffi Tph Þ ð8:20Þ
P ¼ Gphffis ðTph ; Ts ÞðTph ffi Ts Þ: ð8:21Þ
From (8.20) and (8.21),
P ¼ geffiph ðTebþ1 ffi Tph

bþ1
Þ ð8:22Þ
aþ1
P ¼ gphffis ðTph ffi Tsaþ1 Þ: ð8:23Þ
It is practically impossible with this setup (see Ref. [65] for a setup capable of
measuring Tph) to directly measure the phonon temperature, Tph, however, Te, Ts
can be measured as a function of P and Ts. In particular, in the hypothesis that the
electrical resistance of the sample only depends on the electron temperature, Te, it
is possible to measure this temperature using (8.15).
Fig. 8.9 Scheme of hot

electron model
The RuO2 resistor measures the heat sink temperature, Ts. An LR 700 AC
bridge, used in the experiment as a source of power, was able to perform resistance
measurements with currents as low as 0.3 nA, up to 100 mA in 5 % steps.
The power heating the sample is given by two different contributions: the power
Pb supplied by the bridge, and the spurious power, Ps.
Measurements consisted of changing Pb at a different constant Ts. Spurious
power is supposed to be constant during each data acquisition session; this is why a
second NTD 31 sensor was used to control whether its resistance, and hence the
spurious power, remained constant during a measurement; power variation of the
order of 10-13 W could, in this way, be detected.
The temperature range studied in this experiment was 20–40 mK.
The Kapitza thermal conductance was expected to be much more relevant than
electron-phonon conductance in ‘‘high’’ temperature ranges (above 30–35 mK)
where it is supposed that Te : Tph. With this assumption, it was possible to
determine the two phonon-heat-sink thermal conductivity parameters, a and gph-s.
In the ‘‘low’’ temperature range (below 30 mK), the two thermal conductivities
are of the same order of magnitude; thus, the values already determined were used
to obtain electron-phonon decoupling parameters.
Fig. 8.10 Experimental setup for the measurement of the thermal conductivity of Torlon
The values obtained for the electron-phonon conductance and for the Kapitza
conductance for the NTD 31 Ge thermistor are reported in Ref. [56] together with
a comparison between results and data reported in the literature [66] about NTD
samples with different doping levels.
8.5.2 Measurement of the Thermal Conductivity of Torlon

in the 0.08–300 K Temperature Range
Torlon (polyamide-imide) is a thermoplastic polymer which exhibits excellent

mechanical properties and good chemical resistance. In particular, among poly-
mers, it shows a very low coefficient of thermal expansion (less than Al), so it is
often used in cryogenic applications.
It is difficult to carry out measurements of thermal conductivity in a very
extended range of temperature on the same sample (with the same g) and with the
same cryogenic apparatus. Hence, the measurements were performed in two
overlapping ranges, as reported in the following sections [67, 68].
8.5.2.1 Very Low Temperature (0.08–5 K)
In the very low temperature range (0.08–5 K), the sample was a cylinder with a
geometrical factor g = 0.106 cm (see Fig. 8.10).
Fig. 8.11 a Heating power supplied to the sample. b Thermal conductivity of Torlon 4203
The lower end of the sample was fixed onto a copper holder (heat sink) in good
thermal contact with the mixing chamber of a dilution refrigerator. A RuO2
thermometer (Rs) monitored the temperature Ts of the holder. A small block of
copper carrying another RuO2 thermometer (Ru) and also a NiCr heater (Hu) was
screwed onto the upper end of the specimen. At the ends of the sample, there are
two thermal guard rings made with a copper ribbon glued to the sample. Such rings
protrude along the sample for a length equal to that of the screws, creating an
isothermal zone around the screws and defining an effective length L of the
sample. Four bare NbTi wires, 25 lm in diameter, were used to make the electrical
connections of Ru and Hu. Their thermal resistance was about four orders of
magnitude greater than that of the sample [69].
Thermal conductivity was measured by the longitudinal steady state technique
(see Sect. 8.2.1): the measurements below 1 K were carried out with the mixing
chamber maintained at constant temperatures Ts = 70 and Ts = 300 mK by
controlling the power dissipated in a heater (Hs) glued to the copper holder.
Electrical power Ph was supplied to Hu and once the thermal equilibrium was
reached, the temperature Tu was read by Ru. The heating power Ph was then
changed in order to obtain a set of data which are shown in Fig. 8.10a. The
conductivity, obtained as the derivative of Ph/g fit curve, is shown in Fig. 8.10b
and the estimated error is less than 3 %.
Below 0.8 K, thermal conductivity can be represented by a ‘‘typical’’ power law
j(T) = aTn (see Sect. 7.3.3.2 and Fig. 8.11), and the fit gave
jðTÞ ¼ ð6:13 0:07Þ10ffi5 T 2:180:01 : ð8:24Þ
This behavior is in good agreement with the tunneling model [70, 71].
From this measurement, it turns out that Torlon has a very low thermal con-
ductivity: for instance, at 0.1 K, j is only a factor of four greater than that of
graphite AGOT [72] (which is one of the best known thermal insulators).
Fig. 8.12 Shape and size of the sample (mm) and view of the sample holder
8.5.2.2 Thermal Conductivity of Torlon 4203 Between 4.2 and 300 K
For this range of temperature, the experimental apparatus (see Fig. 8.12), the shape
and geometrical factor g of the sample and the method of measurement (mean
conductivity method) are different from those used for the very low temperature
range.
The part of the sample along which the gradient of temperature is produced has
a cylindrical form of length L = 5.75 ± 0.03 mm, and radius r = 4.00 ±
0.01 mm, g = 8.74 ± 0.09 mm at room temperature. The thermal contacts at the
end of the sample have been realized by means of two gold-plated copper screws
(Sc1, Sc2) 4 mm in diameter. The threading in the sample had a depth of 5 mm.
Since the thermal contraction of Torlon is slightly greater than that of copper (see
Sect. 3.9), the thermal contact between the screws and the threaded parts of the
sample becomes better upon cooling. On the other hand, to ensure the thermal
contacts on the two flat surfaces of the sample, two gold-plated copper blocks (B1,
B2), on which thermometers (R1, R2) were mounted, were pressed by stainless
steel springs (Sp) against the two ends of the sample.
Thermal conductivity was measured by a steady state technique. One end of the
sample was fixed (see Fig. 8.13) onto a gold-plated copper platform (Pf) whose
temperature T1 can be set by means of a heater (H1). The thermometer (R1), glued
on the copper block (B1), measured T1. The copper block (B2) held a carbon
thermometer (R2), which measured T2, and a NiCr heater (H2) was glued on the top
of the copper screw (see Fig. 8.12). Electrical connections were made of manganin
wires (diameter 50 lm, *35 cm long). A cylindrical gold-plated copper thermal
Fig. 8.13 Schematics of the

experimental setup for the
measurement of the thermal
conductivity of Torlon
between 4.2 and 300 K
shield (Sh1) enclosed the sample and an outer gold-plated copper thermal shield
(Sh2) enclosed the experiment.
A power Ph was supplied to the heater (H2) in order to create a temperature
gradient DT & 2 % T1 along the sample. The thermal conductivity at a mean
temperature T = (T1 + T2)/2 was evaluated from
Ph
jðTÞ ¼ : ð8:25Þ
gðT2 ffi T1 Þ
Three runs of measurements were carried out in a 4He Cryostat in the following
conditions:
1. 8–25 K: thermal bath at 4.2 K and dewar shield at 77 K;
2. 30–80 K: same as previous run, with a greater thermal impedance to reduce
helium consumption;
3. 80–300 K: heat sink and dewar shield at 77 K, with a greater heat sink-plat-
form thermal impedance.
The thermal conductivity of Torlon 4203 in the 4.2–300 K range is shown in
Fig. 8.14. A linear fit in the 30–350 K is proposed as
jðTÞ ¼ ð0:60 0:03Þ þ ð0:58 0:03Þ10ffi3 T : ð8:26Þ

Fig. 8.14 Thermal conductivity of Torlon 4203 in the temperature range 4.2–300 K. The line
represents (3.68)
Fig. 8.15 Thermal conductivity of Torlon in the 0.08–300 K range

In Fig. 8.15, j(T) is shown over the entire range of measure 80 mK–300 K [67,
68]. The correction due to the thermal contraction (max Dg/g = 0.4 %) was
neglected: the match of the two sets of points is really good.
The thermal conductivity data for Torlon, presented in Fig. 8.14, are typical of
amorphous polymers (see Sect. 8.2.1). In particular, they show:
1. a T2 dependence of k below 1 K, in agreement with the ‘‘tunneling model’’;
2. a plateau between 5 and 10 K, as predicted by the ‘‘soft potential model’’;
3. a steep rise of conductivity after the plateau;
4. an almost linear behavior of k(T) between 30 K and room temperature.
A similar behavior has been found for the thermal conductivity of other
polymers, e.g., polymethylmethacrylate (PMMA) [73].
References
1. Ventura, G., Risegari, L.: The art of cryogenics: low-temperature experimental techniques.
Elsevier, Amsterdam (2007)
2. Woodcraft, A.L.: Predicting the thermal conductivity of aluminium alloys in the cryogenic to
room temperature range. Cryogenics 45(6), 421–431 (2005)
3. Woodcraft, A.L.: Recommended values for the thermal conductivity of aluminium of
different purities in the cryogenic to room temperature range, and a comparison with copper.
Cryogenics 45(9), 626–636 (2005)
4. Slack, G.A.: The thermal conductivity of nonmetallic crystals. Solid State Phys. 34, 1–71
(1979)
5. Tye, R.P.: Thermal Conductivity, vol. 1. Academic Press, London (1969)
6. Johnson, V.A., Marton, L.L., Lark-Horovitz, K.: Methods of Experimental Physics. Springer,
New York (1976)
7. Berman, R. (ed.): Thermal Conduction in Solids. Oxford University Press, Oxford (1976)
8. Pobell, F.: Matter and Methods at Low Temperatures. Springer, New York (2007)
9. Wikus, P., Hertel, S.A., Leman, S.W., McCarthy, K.A., Ojeda, S.M., Figueroa-Feliciano, E.:
The electrical resistance and thermal conductivity of Ti 15 V–3Cr–3Sn–3Al at cryogenic
10. Risegari, L., Barucci, M., Olivieri, E., Pasca, E., Ventura, G.: Measurement of the thermal
conductivity of copper samples between 30 and 150 mK. Cryogenics 44(12), 875–878 (2004)
11. Buck, W.: Thermal properties. In: Czichos, H., Saito, T., Smith, L. (eds.) Springer Handbook
of Materials Measurement Methods, pp. 399–429. Springer, Berlin (2006)
12. Slack, G.A., Glassbrenner, C.: Thermal conductivity of germanium from 3 K to 1020 K.
Phys. Rev. 120(3), 782 (1960)
13. Corbino, O.: Thermal oscillations in lamps of thin fibers with alternating current flowing
through them and the resulting effect on the rectifier as a result of the presence of even-
numbered harmonics. Physikalische Zeitschrift 11, 413–417 (1910)
14. Corbino, O.: Periodic resistance changes of fine metal threads which are brought together by
alternating streams as well as deduction of their thermo characteristics at high temperatures.
Phys. Z. 12, 292–295 (1911)
15. Rosenthal, L.A.: Thermal response of bridgewires used in electroexplosive devices. Rev. Sci.
Instrum. 32(9), 1033–1036 (1961)
16. Holland, L.R.: Physical properties of titanium. III. The specific heat. J. Appl. Phys. 34(8),
2350–2357 (1963)
References 191
17. Birge, N.O., Dixon, P.K., Menon, N.: Specific heat spectroscopy: origins, status and
applications of the 3x method. Thermochim. Acta 304, 51–66 (1997)
18. Birge, N.O., Nagel, S.R.: Specific-heat spectroscopy of the glass transition. Phys. Rev. Lett.
54(25), 2674 (1985)
19. Birge, N.O., Nagel, S.R.: Wide-frequency specific heat spectrometer. Rev. Sci. Instrum.
58(8), 1464–1470 (1987)
20. Frank, R., Drach, V., Fricke, J.: Determination of thermal conductivity and specific heat by a
combined 3x/decay technique. Rev. Sci. Instrum. 64(3), 760–765 (1993)
21. Cahill, D.G.: Thermal conductivity measurement from 30 to 750 K: the 3x method. Rev. Sci.
Instrum. 61(2), 802–808 (1990)
22. Cahill, D.G., Pohl, R.O.: Thermal conductivity of amorphous solids above the plateau. Phys.
Rev. B 35(8), 4067 (1987)
23. Carslaw, H., Jaeger, J.: Conduction of Heat in Solids (paperback,). Clarendon Press, Oxford
(1959)
24. Lee, S.M., Kwun, Sl: Heat capacity measurement of dielectric solids using a linear surface
heater: application to ferroelectrics. Rev. Sci. Instrum. 65(4), 966–970 (1994)
25. Moon, I.K., Jeong, Y.H., Kwun, S.I.: The 3x technique for measuring dynamic specific heat
and thermal conductivity of a liquid or solid. Rev. Sci. Instrum. 67(1), 29–35 (1996)
26. Cahill, D.G., Fischer, H.E., Klitsner, T., Swartz, E., Pohl, R.: Thermal conductivity of thin
films: measurements and understanding. J. Vac. Sci. Tech. A: Vac. Surf. Films 7(3),
1259–1266 (1989)
27. Lee, S.-M., Cahill, D.G.: Heat transport in thin dielectric films. J. Appl. Phys. 81(6),
2590–2595 (1997)
28. Kim, J.H., Feldman, A., Novotny, D.: Application of the three omega thermal conductivity
measurement method to a film on a substrate of finite thickness. J. Appl. Phys. 86(7),
3959–3963 (1999)
29. Yamane, T., Nagai, N., Katayama, S.-I., Todoki, M.: Measurement of thermal conductivity of
silicon dioxide thin films using a 3x method. J. Appl. Phys. 91(12), 9772–9776 (2002)
30. Raudzis, C., Schatz, F., Wharam, D.: Extending the 3x method for thin-film analysis to high
frequencies. J. Appl. Phys. 93(10), 6050–6055 (2003)
31. Olson, B.W., Graham, S., Chen, K.: A practical extension of the 3x method to multilayer
structures. Rev. Sci. Instrum. 76(5), 053901–053907 (2005)
32. Tong, T., Majumdar, A.: Reexamining the 3-omega technique for thin film thermal
characterization. Rev. Sci. Instrum. 77(10), 104902–104909 (2006)
33. Alvarez-Quintana, J., Rodriguez-Viejo, J.: Extension of the 3x method to measure the
thermal conductivity of thin films without a reference sample. Sens. Actuators, A 142(1),
232–236 (2008)
34. Bourgeois, O., Fournier, T., Chaussy, J.: Measurement of the thermal conductance of silicon
nanowires at low temperature. J. Appl. Phys. 101(1), 016103 (2007)
35. Lu, L., Yi, W., Zhang, D.: 3x method for specific heat and thermal conductivity
measurements. Rev. Sci. Instrum. 72(7), 2996–3003 (2001)
36. Choi, T.Y., Poulikakos, D., Tharian, J., Sennhauser, U.: Measurement of thermal
conductivity of individual multiwalled carbon nanotubes by the 3-x method. Appl. Phys.
Lett. 87(1), 013103-013108 (2005)
37. Hu, X.J., Padilla, A.A., Xu, J., Fisher, T.S., Goodson, K.E.: 3-omega measurements of
vertically oriented carbon nanotubes on silicon. J. Heat. Trans. T. ASME 128, 1109–1113
(2006)
38. Hou, J., Wang, X., Vellelacheruvu, P., Guo, J., Liu, C., Cheng, H.-M.: Thermal
characterization of single-wall carbon nanotube bundles using the self-heating 3x
technique. J. Appl. Phys. 100(12), 124314-124319 (2006)
39. Cole, K.D.: Steady-periodic Green’s functions and thermal-measurement applications in
rectangular coordinates. J. Heat Trans. 128, 706–716 (2006)
40. Battaglia, J.-L., Wiemer, C., Fanciulli, M.: An accurate low-frequency model for the 3x
method. J. Appl. Phys. 101(10), 104510 (2007)
41. Jonsson, U.G., Andersson, O.: Investigations of the low-and high-frequency response of-
sensors used in dynamic heat capacity measurements. Meas. Sci. Technol. 9(11), 1873 (1998)
42. Jacquot, A., Lenoir, B., Dauscher, A., Stolzer, M., Meusel, J.: Numerical simulation of the 3x
method for measuring the thermal conductivity. J. Appl. Phys. 91(7), 4733–4738 (2002)
43. Borca-Tasciuc, T., Kumar, A., Chen, G.: Data reduction in 3x method for thin-film thermal
conductivity determination. Rev. Sci. Instrum. 72(4), 2139–2147 (2001)
44. Bhattacharya, P., Nara, S., Vijayan, P., Tang, T., Lai, W., Phelan, P., Prasher, R., Song, D.,
Wang, J.: Characterization of the temperature oscillation technique to measure the thermal
conductivity of fluids. Int. J. Heat Mass Transfer 49(17), 2950–2956 (2006)
45. Wang, H., Sen, M.: Analysis of the 3-omega method for thermal conductivity measurement.
Int. J. Heat Mass Transfer 52(7), 2102–2109 (2009)
46. Wang, Z.L., Tang, D.W., Zheng, X.H.: Simultaneous determination of thermal conductivities
of thin film and substrate by extending 3x-method to wide-frequency range. Appl. Surf. Sci.
253(22), 9024–9029 (2007)
47. Faghani, F.: Thermal Conductivity Measurement of PEDOT: PSS by 3-Omega Technique.
Linköping, Sweden (2010)
48. De Koninck, D.: Thermal conductivity measurements using the 3-omega technique:
application to power harvesting microsystems. In: Masters Abstracts International 2008
49. Maldonado, O.: Pulse method for simultaneous measurement of electric thermopower and
heat conductivity at low temperatures. Cryogenics 32(10), 908–912 (1992)
50. DESIGN, Q.: Sorrento Valley Rd. San Diego, CA 92121-1311 USA
51. Parker, W., Jenkins, R., Butler, C., Abbott, G.: Flash method of determining thermal
diffusivity, heat capacity, and thermal conductivity. J. Appl. Phys. 32(9), 1679–1684 (1961)
52. Donaldson, A., Taylor, R.: Thermal diffusivity measurement by a radial heat flow method.
J. Appl. Phys. 46(10), 4584–4589 (1975)
53. Vandersande, J., Pohl, R.: Simple apparatus for the measurement of thermal diffusivity
between 80–500 K using the modified Ångström method. Rev. Sci. Instrum. 51(12),
1694–1699 (1980)
54. Gembarovič, J., Vozár, L., Majerník, V.: Using the least square method for data reduction in
the flash method. Int. J. Heat Mass Transfer 33(7), 1563–1565 (1990)
55. Hatta, I., Sasuga, Y., Kato, R., Maesono, A.: Thermal diffusivity measurement of thin films
by means of an ac calorimetric method. Rev. Sci. Instrum. 56(8), 1643–1647 (1985)
56. Pasca, E., Barucci, M., Ventura, G.: Measurement of Electron-phonon decoupling in NTD31
Germanium. In: Proceedings of the ICATPP 7th 684 (2002)
57. Arnaboldi, C., Avignone, F., Beeman, J., Barucci, M., Balata, M., Brofferio, C., Bucci, C.,
Cebrian, S., Creswick, R., Capelli, S.: Physics potential and prospects for the CUORICINO
and CUORE experiments. Astropart. Phys. 20(2), 91–110 (2003)
58. Shklovskii, B.I., Efros, A.L.: Electronic properties of doped semiconductors. vol. 1. Moscow
Izdatel Nauka, Moscow (1979)
59. Mott, N.F., Physicist, G.B.: Conduction in Non-crystalline Materials. Clarendon Press,
Oxford (1987)
60. Barucci, M., Beeman, J., Olivieri, E., Pasca, E., Risegari, L., Ventura, G.: Electrical
characteristics of heavily doped NTD Ge at very low temperatures. Physica B 368(1),
139–142 (2005)
61. Keesom, P., Seidel, G.: Specific heat of germanium and silicon at low temperatures. Phys.
Rev. 113(1), 33 (1959)
62. Richards, P.: Bolometers for infrared and millimeter waves. J. Appl. Phys. 76(1), 1–24 (1994)
63. Twerenbold, D.: Cryogenic particle detectors. Rep. Prog. Phys. 59(3), 349 (1996)
64. Wang, N., Wellstood, F.C., Sadoulet, B., Haller, E.E., Beeman, J.: Electrical and thermal
properties of neutron-transmutation-doped Ge at 20 mK. Phys. Rev. B 41(6), 3761–3768
(1990)
65. Wang, N., Beeman, J., Cleland, A., Cummings, A., Haller, E., Lange, A., Ross, R., Sadoulet,
B., Steiner, H., Shutt, T.: Particle detection with semiconductor thermistors at low
temperatures. Nucl. Sci. IEEE Trans. 36(1), 852–856 (1989)
References 193
66. Soudee, J., Broszkiewicz, D., Giraud-Héraud, Y., Pari, P., Chapellier, M.: Hot electrons effect
in a# 23 NTD Ge sample. J. Low Temp. Phys. 110(5–6), 1013–1027 (1998)
68. Barucci, M., Olivieri, E., Pasca, E., Risegari, L., Ventura, G.: Thermal conductivity of Torlon
between 4.2 and 300 K. Cryogenics 45(4), 295–299 (2005)
Cryogenics 33(7), 729–731 (1993)
70. Anderson, P.W., Halperin, B., Varma, C.M.: Anomalous low-temperature thermal properties
of glasses and spin glasses. Phil. Mag. 25(1), 1–9 (1972)
(1972)
72. Woodcraft, A.L., Barucci, M., Hastings, P.R., Lolli, L., Martelli, V., Risegari, L., Ventura,
G.: Thermal conductivity measurements of pitch-bonded graphites at millikelvin
temperatures: finding a replacement for AGOT graphite. Cryogenics 49(5), 159–164 (2009)
73. Choy, C.: Thermal conductivity of polymers. Polymer 18(10), 984–1004 (1977)
Chapter 9
Data of Thermal Conductivity
Abstract In this chapter, data of thermal conductivity (j) and integrated con-
ductivity (I*) are presented in the form of tables. Data are divided into three
categories. 1. Very low temperature data (100 mK–2 K), 2. Low temperature data
(4–300 K), 3. Data about crystalline materials for which a fit in a limited range of
temperature is proposed.
In this section, we report data of thermal conductivity and the useful integrated
conductivity (I*), obtained as (see Chap. 3 for further details):
ðjðTn Þ þ jðTn1 ÞÞ
I ffi ðTn Þ ¼ I ffi ðTn1 Þ þ ðTn Tn1 Þ ð9:1Þ
2
A huge amount of data of thermal conductivity is available in literature. Here

we report data divided into three categories.
1. Very low temperature data (100 mK–2 K)
2. Low temperature data (4–300 K)
3. Crystalline materials for which a fit of a particular range of temperature is
proposed.
9.1 Very Low Temperature Data
Tables 9.1, 9.2 and 9.3.
9.2 Low Temperature Data
Tables 9.4, 9.5, 9.6 and 9.7.

196
Table 9.1 Very low temperature data of thermal conductivity (W/cm K) of metals and alloys
Metal or alloy Chemical composition References 100 mK 250 mK 500 mK 750 mK 1K 2K
Alluminium Al 99.999 % [1] – – 3.1 4.3 – –
Copper Cu high purity [2] – 3.2 6.2 – – –
Cu low purity [2] – 4 9 10-1 8 9 10-1 – – –
Cu 99.999 % [3] – – – – 3 –
Gold Au 99.999 % [1] – – 6.8 9 10-1 1.0 – –
Au 99.999 % [3] – – – –
Indium In [4] 5 9 10-2 2 9 10-1 3 9 10-1 5 9 10-1 7 9 10-1 1.3
Manganin§ Cu8.4Mn1.2Ni0.4 [5] 5.2 9 10-5 1.4 9 10-4 3.0 9 10-4 4.9 9 10-4 6 10-4 2 10-3
Nickel Ni 99.995 % [1] – – 1.2 9 10-1 1.9 9 10-1 – –
Palladium Pd 99.999 % [1] – – 1.8 9 10-1 – – –
Palladium-Alluminium-Rhenium Pd7Al2Re1 [6] – – 1.6 9 10-5 4.0 9 10-5 8 9 10-5 3 9 10-4
Platinum Pt 99.999 % [1] – – 1.4 9 10-1 2.4 9 10-1 – –
Rhenium Re 99.994 % [1] – – 4 9 10-1 7.4 9 10-1 – –
Silver Ag 99.999 % [1] – – 6.7 9 10-1 9.7 9 10-1 – –
Ag [ 99.99 % [3] – – – 4
-
Tantalum Ta 99.994 % [1] – – 7 9 10-2 1 9 10-1 1.4 9 10-1
Titanium Alluminium Ti6Al [5] – 1.6 9 10-5 1 9 10-4 2.5 9 10-4 4 9 10-4 1.1 9 10-3
-
Tungsten W [1] – – 7 9 10-2 1.1 9 10-1 1.4 9 10-1
§
The symbol refers to weight percentage while other composition are molar
9 Data of Thermal Conductivity
Table 9.2 Very low temperature data of thermal conductivity (W/cm K) of superconducting materials
Superconductor Chemical Composition References 100 mK 250 mK 500 mK 750 mK 1K 2K 4K
Aluminum Al [7] 3.0 9 10-5 3.0 9 10-2 – – – – –
Al 99.999 % [1] – – 9 9 10-1 2.9 5.4 – –
Cadmium Cd 99.999 % [1] – – 9 9 10-1 1.8 – – –
Copper-Nichel§ Cu 70 % [8] – 1 9 10-4 2.3 9 10-4 4 9 10-4 5 9 10-4 1 9 10-3 –

Ni 30 %
Indium In 99.997 % [3] – – – – – – 8.2
Lead Pb 99.99 % [9] – – – – 3 9 10-1 2 3.0
Lead-Bismuth§ Pb 99 % [9] – – – – 1.5 9 10-1 6 9 10-1 3.5 9 10-1
Bi 1 %
Niobium Nb [10] – 7 9 10-3 3 9 10-2 2 9 10-1 – – –
Nb [9] – – – – 9 9 10-3 6 9 10-2 8 9 10-2
Nb [11] 5.0 9 10-4 3.5 9 10-3 1.1 9 10-2 1.8 9 10-2 3.9 9 10-2 1 9 10-3 –
Nb 99.99 % [3] – – – – – – 3 9 10-2
Niobium-Titanium§ Nb 45 % Ti 55 % [8] – 6 9 10-6 5 9 10-5 8 9 10-5 1.6 9 10-4 3 9 10-4 1 9 10-3
Rhenium Re 99.994 % [1] – – 1 9 10-1 3.8 9 10-1 – – –
Tantalum Ta [10] 3 9 10-4 3 9 10-3 3 9 10-2 1 9 10-2 1 9 10-1 6 9 10-2 –
Ta 99.994 % [1] – – 7 9 10-2 1 9 10-1 1.4 9 10-1 – –
Ta 99.98 % [3] – – – – – – 1.5 9 10-1
Vanadium V [3] – – – – – – 1.3 9 10-2
Zinc Zn [12] – 3.1 9 10-1 2.8 – – – –
§
197
Table 9.3 Very low temperature data of thermal conductivity (W/cm K) of ceramics, polymers and other categories of materials
198
Material Chemical composition References 100 mK 250 mK 500 mK 750 mK 1K 2K

Arsenicum sulphide As2S3 [13] 1.3 9 10-5 8 9 10-5 3.9 9 10-4 5.8 9 10-4 1.0 9 10-3 1.4 9 10-3
AXM-5Q [5] 8 9 10-7 2 9 10-6 7 9 10-6 1 9 10-5 1 9 10-5 7 9 10-5
Borosilicate glass [14] 1.4 9 10-6 1.8 9 10-6 5.4 9 10-5 1 9 10-4 2.0 9 10-4 5 9 10-4
Cadmium-Germanium CdGeAs2 [13] 1.1 9 10-5 6 9 10-4 3.4 9 10-4 5.5 9 10-4 1.4 9 10-3 2.2 9 10-3
Arsenide
Calcium-Potassium Nitrate CaKNO3 [13] 2.1 9 10-6 1.2 9 10-5 6 9 10-5 1.0 9 10-4 2 9 10-4 3.9 9 10-4
Epibond 100 A [15] – 1.6 9 10-5 8.0 9 10-5 1.6 9 10-4 2.3 9 10-4 –
G-10CR [5] 2 9 10-6 1 9 10-5 5 9 10-5 9.3 9 10-5 1 9 10-4 3.8 9 10-4
Germanium oxide GeO2 [16] 5.0 9 10-6 2.5 9 10-5 9.8 9 10-5 3.0 9 10-4 3.3 9 10-4 9.4 9 10-4
HTO Ho2Ti2O7 [66] [17] 9 9 10-6 8 9 10-5 3 9 10-4 5.6 9 10-4 8.5 9 10-4 –
-
Kapton [18] 1.5 9 10-5 3.0 9 10-5 4 9 10-5 6.1 9 10-5 1 9 10-4
Kevlar 49 [19] 7 9 10-7 3 9 10-6 1 9 10-5 2.3 9 10-5 3.8 9 10-5 1.2 9 10-4
Lithium b-alumina Li2O11Al2O3 [20] – 1.9 9 10-5 6.7 9 10-5 1.5 9 10-4 2.5 9 10-4 8.2 9 10-4
Macor [5] 8 9 10-8 9 9 10-7 7 9 10-6 9.1 9 10-6 4 9 10-5 5.5 9 10-5
-6 -6 -5 -5
Nylon [21] – 2.3 9 10 7.7 9 10 1.5 9 10 2.6 9 10 6 9 10-5
PB [22] 2.0 9 10-6 1.2 9 10-5 5.0 9 10-5 1.0 9 10-4 1.6 9 10-4 4.8 9 10-4
PC [14] – 5 9 10-6 1.4 9 10-5 6 9 10-5 7 9 10-5 1 9 10-4
PET [22] 2.6 9 10-6 1.3 9 10-5 4.2 9 10-5 7.7 9 10-5 1.2 9 10-4 2.7 9 10-4
(continued)
PMMA [13] 4.6 9 10-6 2.5 9 10-5 1.1 9 10-4 1.6 9 10-4 2.8 9 10-4 4.3 9 10-4
Potassium b-alumina K2O11Al2O3 [20] – 4.7 9 10-4 4.0 9 10-3 1.3 9 10-2 2.5 9 10-2 9.0 9 10-2
Potassium bromide KBr [23] – 7.8 9 10-3 6.4 9 10-2 2.1 9 10-1 4.2 9 10-1 1.7
Potassium chloride KCl [23] – – 2.7 9 10-2 8.8 9 10-2 2.2 9 10-1 1.4
Potassium iodide KI [23] – – 1.3 9 10-1 4.1 9 10-1 9.3 9 10-1 4.0
KI0.9994(NO3)0.0006 [23] – – 1.3 9 10-2 3.8 9 10-2 7.7 9 10-2 2.6 9 10-1
PS [22] – 6.1 9 10-6 2.2 9 10-5 4.4 9 10-5 7.4 9 10-5 1.6 9 10-4
PVC [24] – 1.1 9 10-5 3.2 9 10-5 – – –
Pyrex Glass§ SiO2 80 %, [25] – 1.5 9 10-5 5 9 10-5 1.1 9 10-4 1.8 9 10-4 4.5 9 10-4
B2O3 13 %,
Na2O 4 %
Al2O3 2 %
K2O 0.4 %
Li2O 0.2 %
Silica SiO2 [13] 3.0 9 10-6 1.8 9 10-5 9 9 10-5 1.5 9 10-4 3.4 9 10-4 6.0 9 10-4
Selenium Se [13] 1.1 9 10-5 5 9 10-5 2.3 9 10-4 3.2 9 10-4 5 9 10-4 5.7 9 10-4
-5 -4 -4 -4
Se [26] – 6.0 9 10 1.7 9 10 2.9 9 10 4.0 9 10 5.3 9 10-4
(continued)
199
200

Shappire Al2O3 [1] – – 7 9 10-2 1.7 9 10-1 2.5 9 10-1 –
Silver b-alumina Ag2O11Al2O3 [20] 1.7 9 10-5 1.1 9 10-5 4.0 9 10-4 8.1 9 10-4 1.3 9 10-3 3.5 9 10-3
Sodium b-alumina Na2O11Al2O3 [20] – 1.1 9 10-5 5.4 9 10-4 1.1 9 10-3 2.0 9 10-3 6.6 9 10-3
Stycast 1266 [8] 6 9 10-6 3.8 9 10-5 1 9 10-4 2 9 10-4 3 9 10-4 7 9 10-4
Teflon [27] – 1.7 9 10-6 6.5 9 10-6 1.4 9 10-5 2.6 9 10-5 8.6 9 10-5
Torlon 4203 [28] 4 9 10-7 4 9 10-6 1 9 10-5 3 9 10-5 6 9 10-5 8 9 10-5

-6 -6 -6 -5 -5
Upilex [18] 1.2 9 10 3.6 9 10 7.7 9 10 1.5 9 10 1.7 9 10 4 9 10-5
Vespel SP1 [5] – 3 9 10-6 7 9 10-6 1 9 10-5 2 9 10-5 3 9 10-5
§
Table 9.4 Low temperature data of thermal conductivity (W/cm K) of metals and alloys
Material Chemical References 4 K 10 K 20 K 30 K 77 K 100 K 150 K 200 K 300 K
composition
Al 5083 [29] 3.3 9 10-2 8.4 9 10-2 1.7 9 10-1 2.5 9 10-1 5.5 9 10-1 6.6 9 10-1 8.5 9 10-1 9.9 9 10-1 1.18
Al 6061-T6 [29] 5.3 9 10-2 1.4 9 10-1 2.8 9 10-1 4.0 9 10-1 8.4 9 10-1 9.8 9 10-1 1.2 1.36 1.55
Aluminium Al [30] – 8.95 14.8 14.8 3.70 – – – –
Al 99.994 % [3] 17 37 43 28 – – – – –
Antimony Sb [3] 7.4 9 10-2 3.1 9 10-1 3.8 9 10-1 3.5 9 10-1 – – – – –
Beryllium Be [3] 2.7 9 10-2 6.3 9 10-2 1.4 9 10-1 2.3 9 10-1 – – – – –
Beryllium-Copper Cu0.99Be0.01 [29] 1.9 9 10-2 5.0 9 10-2 1.1 9 10-1 1.6 9 10-1 3.6 9 10-1 4.1 9 10-1 4.1 9 10-1 3.1 9 10-1 9.7 9 10-2
Brass Cu0.68Zn0.32 [31] 3.0 9 10-2 1.0 9 10-1 2.2 9 10-1 3.0 9 10-1 5.3 9 10-1 – – – –
Cadmium Cd 99.9999 % [3] – 7 3 1.9 – – – – –
Cesium Cs [32] 1.11 6.7 9 10-1 – – – – – – –
Cerium Ce 99.6 % [3] 4.7 9 10-3 1.0 9 10-2 1.9 9 10-2 – – – – – –
Cobalt Co [3] 5 9 10-1 1.3 2.2 2.7 – – – – –
Cobalt Erbium Co2Er [33] – 1.3 9 10-2 2.5 9 10-2 3.1 9 10-2 – – – – –
Copper Cu 99.999 % [3] 11 26 34 25 – – – – –
OFHC [29] 6.3 15.40 24.30 19.50 5.44 4.61 4.18 4.07 3.97
(RRR = 100)
Copper arsenic Cu99.55As0.35P0.05 [34] – 1 9 10-5 3 9 10-5 4 9 10-5 8 9 10-5 – – – –
Copper zinc Cu6.8Zn3.2 [34] – 6 9 10-6 1.5 9 10-5 4.5 9 10-5 – – – –
Dysprosium Nickel DyNi2 [33] – 2.0 9 10-2 2.7 9 10-2 3.6 9 10-2 – – – – –
Erbium Nickel ErNi [33] – 4 9 10-2 5 9 10-2 – – – – – –
Er3Ni [33] – 5 9 10-3 1 9 10-2 – – – – – –
Gallium Ga high res. [3] 8 9 10-1 1.2 7 9 10-1 4.8 9 10-1 – – – – –
Ga int. res. [3] 1.7 2.9 1.8 1.3 – – – – –
Ga low res. [3] 22 23 6.6 3.5 – – – – –
Germanium Ge [35] – 18 13 9 2.6 1.9 1.1 8 9 10-1 5 9 10-1
Gold Au 99.999 % [3] 3.3 7.3 7.5 – – – – – –
Inconel 718 [29] 4.6 9 10-3 1.5 9 10-2 3.0 9 10-2 3.9 9 10-2 6.4 9 10-2 7.1 9 10-2 8.1 9 10-1 8.7 9 10-1 9.7 9 10-1
(continued)
201
202

composition
Indium In 99.993 % [3] 8.0 1.6 8 9 10-1 7 9 10-1 – – – – –
Invar Fe0.84Ni0.36 [29] 2.4 9 10-3 7.3 9 10-3 1.7 9 10-2 2.2 9 10-2 4.2 9 10-2 6.2 9 10-2 7.6 9 10-2 1.0 9 10-1 1.2 9 10-1
Iridium Ir 99.995 % [3] 5.2 12.3 17.7 13.6 – – – – –
Iron Fe 99.99 % [3] 7 9 10-1 1.7 2.9 3.5 1.7 – – – –
Lanthanum La 99.94 % [3] 6.3 9 10-3 1.6 9 10-2 3.3 9 10-2 4.5 9 10-2 – – – – –
Lead Pb 99.998 % [3] 19 2 6 9 10-1 5.7 9 10-1 – – – – –
Lithium Li [32] 2.6 5.9 7.4 5.4 1.2 1.0 – – –
Magnesium Mg 99.95 % [3] 3.4 8.7 11.6 – – – – – –
Magnesium Tin Mg2Sn [36] – 5.9 4.9 2.5 3.8 9 10-1 2.8 9 10-1 1.7 9 10-1 1.2 9 10-1 8.0 9 10-2
Manganese Mn 99.99 % [3] 3 9 10-3 8 9 10-3 1.7 9 10-2 – – – – – –
Manganin§ Cu 84 % Mn [5] 5 9 10-3 2 9 10-2 4 9 10-2 6 9 10-2 1 9 10-1 1 9 10-1 2 9 10-1 2 9 10-1 2 9 10-1
12 % Ni 4 % [31] 4.4 9 10-3 1.4 9 10-2 3.2 9 10-2 5.0 9 10-2 1.1 9 10-1 – – – –
Molybdenum Mo 99.95 % [3] 5 9 10-1 1.4 2.8 3.6 2.1 – – – –
Nickel Ni [3] 9 9 10-1 2.2 3.6 4.0 – – – – –
Nickel Erbium Ni2Er [33] – 1.2 9 10-2 2.3 9 10-2 3.1 9 10-2 – – – – –
Niobium Nb 99.99 % [3] – 1.5 9 10-1 3.1 9 10-1 4.1 9 10-1 4.8 9 10-1 – – – –
Niobium Titanium Nb4Ti [37] – 1.5 9 10-2 3.1 9 10-2 – – – – – –
Nb4.5Ti 5.5 [37] – 9.9 9 10-3 1.7 9 10-2 – – – – – –
Nb3.5Ti6.5 [37] – 6.2 9 10-3 1.2 9 10-2 – – – – – –
Nb2.2Ti7.8 [37] – 3.2 9 10-3 6.4 9 10-3 – – – – – –
Palladium Pd 99.995 % [3] 3.4 9 10-1 8.3 9 10-1 1.3 1.3 – – – – –
Palladium Pd7Al2Re1 [6] 1.1 9 10-3 3 9 10-3 3 9 10-3 4 9 10-3 4 9 10-3 4 9 10-3 4 9 10-3 – –
Alluminium
Rhenium
Platinum Pt 99.999 % [3] 8.7 12.3 5.0 2.5 – – – – –
Potassium K [32] 5.6 4.3 1.6 1.1 – – – – –
Rhodium Rh 99.995 % [3] 2.6 6.7 11.1 – – – – – –
(continued)
composition
Rubidium Rb [32] 1.8 1.1 0.7 – – – – – –
Silicon Si [35] – 16 33 32 10 7 3 2 1
Silver Ag [ 99.99 % [3] 14 30 28 17 – – – – –
Sodium Na [32] 46.6 22.8 – – – – – – –
Solder (soft)§ Sn0.6Pb0.4 [31] 1.6 9 10-1 4.3 9 10-1 5.6 9 10-1 5.5 9 10-1 5.3 9 10-1 – – – –
Stainless Steel 304, [29] 2.7 9 10-3 9.0 9 10-3 2.2 9 10-2 3.5 9 10-2 7.9 9 10-2 9.2 9 10-2 1.1 9 10-1 1.3 9 10-1 1.5 9 10-1
316
Steel 1462 [38] – 7 9 10-3 2 9 10-2 4 9 10-2 8 9 10-2 9 9 10-2 1.0 9 10-1 1.1 9 10-1 1.4 9 10-1
Tantalum Ta 99.98 % [3] 1.6 9 10-1 3.7 9 10-1 6.4 9 10-1 6.6 9 10-1 6.1 9 10-1 – – – –
Thallium Tl 99.99 % [3] 10 4.7 1.8 1.3 – – – – –
Tin Sn 99.997 % [3] 25 9 2.3 1.4 – – – – –
Titanium Ti 99.99 % [3] 1.1 9 10-1 2.3 9 10-1 3.6 9 10-1 4.6 9 10-1 – – – – –
Titanium Alluminum Ti6Al [5] 3 9 10-3 8 9 10-3 1 9 10-2 2 9 10-2 4 9 10-2 5 9 10-2 6 9 10-2 7 9 10-2 1 9 10-1
Titanium Ti0.9Al0.06V0.04 [29] – – 8.4 9 10-3 1.4 9 10-2 3.5 9 10-2 3.8 9 10-2 4.6 9 10-2 5.8 9 10-2 7.4 9 10-2
Alluminium
Vanadium
Tungsten W 99.99 % [3] 5 9 10-1 1.7 3.1 3.6 – – – – –
Uranium U [3] 4 9 10-2 1.0 9 10-1 1.6 9 10-1 1.7 9 10-1 2.4 9 10-1 3 9 10-1 – – –
Vanadium V [3] 1.3 9 10-2 3 9 10-2 6.3 9 10-2 – – – – – –
Zinc Zn 99.997 % [3] 5 11 6.5 3.4 – – – – –
Zirconium Hafnium Zr 98 % Hf2 % [3] 1.1 9 10-1 2.9 9 10-1 4.6 9 10-1 4.6 9 10-1 – – – – –
§
203
Table 9.5 Low temperature data of integrated thermal conductivity (W/cm) of metals and alloys
204
Material Reference I*4 K I*10 K I*20 K I*30 K I*77 K I*100 K I*150 K I*200 K I*300 K
Al5083 [29] 6.6 9 10-2 4.17 9 10-1 1.687 3.787 19.05 34.175 71.925 117.925 226.425
Al6061T6 [29] 1.06 9 10-1 6.85 9 10-1 2.785 6.185 29.54 52.29 106.79 170.79 316.29
Aluminium [3] 34 196 596 951 – – – – –
Antimony [3] 1.48 9 10-1 1.3 4.75 8.4 – – – – –
Beryllium [3] 5.4 9 10-2 3.24 9 10-1 1.339 3.189 – – – – –
Beryllium-Copper [29] 3.8 9 10-2 2.45 9 10-1 1.045 2.395 12.38 22.005 42.505 60.505 124.505
Brass [31] 6 9 10-2 4.5 9 10-1 2.05 4.65 19.805 – – – –
Cerium [3] 9.4 9 10-3 5.35 9 10-2 1.985 9 10-1 – – – – – –
Cobalt [3] 1 6.4 23.9 48.4 – – – – –
Copper [3] 22 133 433 728 – – – – –
Copper [29] 12.6 77.7 276.2 495.2 605.59 731.215 950.965 1157.215 1559.215
Gallium high res. [3] 1.6 7.6 17.1 23 – – – – –
Gallium int. Res. [3] 3.4 17.2 40.7 56.2 – – – – –
Gallium low res. [3] 44 179 327 377.5 – – – – –
Gold [3] 6.6 38.4 112.4 – – – – – –
Inconel 718 [29] 9.2 9 10-3 6.8 9 10-2 2.93 9 10-1 6.38 9 10-1 2.4595 4.147 7.947 12.147 21.347
Indium [3] 16 44.8 – – – – – – –
Invar [29] 4.8 9 10-3 3.39 9 10-2 1.554 9 10-1 3.504 9 10-1 1.526 2.826 6.276 10.676 21.676
Iridium [3] 10.4 62.9 212.9 369.4 – – – – –
Iron [3] 1.4 8.6 31.6 63.6 125.7 – – – –
Lanthanum [3] 1.26 9 10-2 7.95 9 10-2 3.245 9 10-1 7.145 9 10-1 – – – – –
Lead [3] 38 101 114 119.85 – – – – –
Magnesium [3] 6.8 43.1 144.6 – – – – – –
Manganese [3] 6 9 10-3 3.9 9 10-2 1.64 9 10-1 – – – – – –
Manganin [5] 1 9 10-2 8.5 9 10-2 3.85 9 10-1 8.85 9 10-1 3.82 6.32 13.82 23.82 43.82
Manganin [31] 8.8 9 10-3 6.4 9 10-2 2.94 9 10-1 7.04 9 10-1 3.81 – – – –
(continued)
Molybdenum [3] 1 6.7 27.7 59.7 137.55 – – – –
Nickel [3] 1.8 11.1 40.1 78.1 – – – – –
Palladium [3] 6.8 9 10-1 4.19 14.84 27.84 – – – – –
Palladium [6] 2.2 9 10-2 6.4 9 10-2 9.04 9 10-2 1.29 9 10-1 1.92 9 10-1 2.92 9 10-1 4.92 9 10-1 – –
Alluminium
Rhenium
Platinum [3] 17.4 80.4 166.9 204.4 – – – – –
Rhodium [3] 5.2 33.1 122.1 – – – – – –

Silver [3] 28 160 450 675 – – – – –
Solder [31] 3.2 9 10-1 2.09 7.04 12.59 25.93 – – – –
Stainless [29] 5.4 9 10-3 4.05 9 10-2 1.955 9 10-1 4.805 9 10-1 2.714 4.8515 9.9015 15.9015 29.9015
Tantalum [3] 3.2 9 10-1 1.91 6.96 13.46 30.505 – – – –
Thallium [3] 20 64.1 96.6 112.1 – – – – –
Tin [3] 50 152 208.5 227 – – – – –
Titanium [3] 2.2 9 10-1 1.24 4.19 8.29 – – – – –
Titanium [5] 6 9 10-3 3.9 9 10-2 1.29 9 10-1 2.79 9 10-1 1.43 2.555 5.305 8.555 17.055
Aluminium
Tungsten [3] 1 7.6 31.6 65.1 – – – – –
Uranium [3] 8 9 10-2 5 9 10-1 1.8 3.45 9.805 16.555 – – –
Vanadium [3] 2.6 9 10-2 1.55 9 10-1 6.2 9 10-1 – – – – – –
Zinc [3] 10 58 145.5 195 – – – – –
Zirconium [3] 2.2 9 10-1 1.42 5.17 9.77 – – – – –
205
Table 9.6 Low temperature data of thermal conductivity (W/cm K) of various materials
206
Material Chemical composition Reference 4 K 10 K 20 K 30 K 77 K 100 K 150 K 200 K 300 K

Alumina Al2O3, sintered [31] 4.9 9 10-3 5.6 9 10-2 2.4 9 10-1 5.2 9 10-1 1.57 1.36 9.3 9 10-1 5.0 9 10-1 –
Arsenicum sulphide As2S3 [13] 1.7 9 10-3 1.5 9 10-3 1.8 9 10-3 1.7 9 10-3 2.4 9 10-3 2.2 9 10-3 – – –
AXM-5Q [5] 2 9 10-6 1 9 10-3 1 9 10-2 5 9 10-2 2 9 10-1 3 9 10-1 7 9 10-1 8 9 10-1 1
Boron nitride BN, pyrolitic oriented [39] – 2 9 10-1 7 9 10-1 1.7 4.8 5.0 5.2 4.8 3.8
BN, pyrolitic [40] – 5 9 10-3 4 9 10-2 1.2 9 10-1 1.2 1.5 2.0 2.0 –
BN, pyrolitic [40] – 6 9 10-4 2 9 10-3 5.1 9 10-3 4 9 10-2 7 9 10-2 1.3 9 10-1 1.8 9 10-1 –
Boron oxide B2O3 [22] – 1.2 9 10-3 1.3 9 10-3 1.6 9 10-3 – – – – –
Borosilicate glass [14] 1.0 9 10-3 1.6 9 10-3 2.0 9 10-3 2.2 9 10-3 – – – – –
Cadmium- CdGeAs2 [13] 4 9 10-3 4.0 9 10-3 3.7 9 10-3 3.7 9 10-3 3.8 9 10-3 3.8 9 10-3 – – –
Germanium
Arsenide
Cadmium telluride CdTe, cubic [41] – – – 1.80 4.4 9 10-1 4 9 10-1 1.84 9 10-1 1.0 9 10-1 7.5 9 10-2
Calcium Potassium CaK(NO3)3 [13] 9 9 10-4 1.5 9 10-3 2.7 9 10-3 2.6 9 10-3 1.0 9 10-2 – – – –
Nitrate
Calcium Ca0.20Co4Sb12.46 [42] – 1.8 9 10-2 3 9 10-2 4 9 10-2 5 9 10-2 5 9 10-2 4 9 10-2 4 9 10-2 4 9 10-2
skutterudite Ca0.18Co3.94Ni0.06Sb12.40 [42] – 4 9 10-2 5 9 10-2 6 9 10-2 7.3 9 10-2 7.2 9 10-2 6.7 9 10-2 6.3 9 10-2 5.7 9 10-2
Diamond C [43] 6.7 9 10-1 5.1 15 21 34 32 12 8.3 –
G-10CR Normal [29] 7.2 9 10-4 1.1 9 10-3 1.6 9 10-3 1.9 9 10-3 2.8 9 10-3 3.1 9 10-3 3.7 9 10-3 4.5 9 10-3 6.0 9 10-3
Warp [29] 7.3 9 10-4 1.4 9 10-3 2.0 9 10-3 2.3 9 10-3 3.9 9 10-3 4.5 9 10-3 5.7 9 10-3 6.7 9 10-3 8.6 9 10-3
Gadolinium Gd2Zr2O7 [44] – – 1.6 9 10-2 1.6 9 10-2 1.5 9 10-2 1.4 9 10-2 1.4 9 10-2 1.4 9 10-2 1.3 9 10-2
Zirconium
oxide
GCMCGG:Nd,Cr Gd2.57Nd0.06Ca0.39Mg [45] – 4.3 9 10-1 5.5 9 10-1 4.0 9 10-1 2 9 10-1 1.1 9 10-1 9 9 10-2 6.1 9 10-2 5.2 9 10-2
0.28Zr0.63Cr0.01
Ga0.46O12
Germanium oxide GeO2 [16] 1.7 9 10-3 1.9 9 10-3 2.2 9 10-3 2.5 9 10-3 5.5 9 10-3 7 9 10-3 – – –
GGG:Eu Gd2.91Eu0.12Ga4.97O12 [45] – 4.5 5.0 3.5 6 9 10-1 3.1 9 10-1 2 9 10-1 1.25 9 10-1 8.6 9 10-2
GGG:Nd Gd2.99Nd0.06Ga4.95O12 [45] – 4.2 3.6 1.8 5 9 10-1 2.8 9 10-1 2 9 10-1 1.24 9 10-1 8.6 9 10-2
GGG:Nd,Cr Gd2.97Nd0.07Cr0.005 [45] – 4.3 4.5 2.3 6 9 10-1 3.1 9 10-1 2 9 10-1 1.24 9 10-1 8.6 9 10-2
Ga4.955O12
(continued)
GGG:Sm Gd2.82Sm0.12Ga4.97O12 [45] – 3.9 1.8 1.4 7 9 10-1 2.1 9 10-1 1.6 9 10-1 1.09 9 10-1 8.2 9 10-2
GGG:Tb Gd2.95Tb0.06Ga4.99O12 [45] – 3.3 2.7 1.9 4 9 10-1 2.5 9 10-1 1.8 9 10-1 1.15 9 10-1 8.4 9 10-2
GSGG:Nd,Cr Gd2.89Nd0.06Sc1.90 [45] – 1.26 1.10 6.0 9 10-1 2 9 10-1 1.4 9 10-1 1 9 10-1 8.0 9 10-2 5.2 9 10-2
Cr0.05Ga3.10O12
Kapton [46] 2.3 9 10-4 5.1 9 10-4 1.1 9 10-3 2.0 9 10-3 5.2 9 10-3 5.0 9 10-3 8 9 10-3 1.1 9 10-2 1.7 9 10-2
[29] 1.1 9 10-4 2.4 9 10-4 4.8 9 10-4 6.6 9 10-4 1.3 9 10-3 1.4 9 10-3 1.6 9 10-3 1.8 9 10-3 1.9 9 10-3
Kevlar 49 [47] 5.3 9 10-4 2.0 9 10-3 4.3 9 10-3 6.2 9 10-3 1.5 9 10-2 1.9 9 10-2 2.6 9 10-2 3.0 9 10-2 3.9 9 10-2
[29] 3.0 9 10-4 1.2 9 10-3 2.9 9 10-3 4.4 9 10-3 1.0 9 10-2 1.2 9 10-2 1.5 9 10-2 1.7 9 10-2 2.0 9 10-2
Lithium b-alumina Li2O11Al2O3 [20] 2.3 9 10-3 6 9 10-3 9.3 9 10-3 1.2 9 10-2 1.7 9 10-2 2.2 9 10-2 – –
7
Lithium Fluoride LiF [48] – 104 140 56 1.5 0.9 – – –
[48] – 16 65 37 1.5 1.0 – – –
Macor [5] 8 9 10-4 3 9 10-3 6 9 10-3 6 9 10-2 1.2 9 10-2 1.3 9 10-2 1.4 9 10-2 1.4 9 10-2 1.5 9 10-2
[49] 7.5 9 10-4 2.5 9 10-3 6.0 9 10-3 – – – – – –
Magnesium oxide, MgO [31] 8.2 9 10-1 11.3 27.7 21.6 5.07 2.94 1.35 9.1 9 10-1 6.1 9 10-1
crystal
Mylar [22] 3.6 9 10-4 4.4 9 10-4 6.1 9 10-4 8.4 9 10-4 1.5 9 10-3 – – – –
[50] 3.8 9 10-4 4.8 9 10-4 7.3 9 10-4 8.5 9 10-4 1.2 9 10-3 – – – –
Nylon (6,6) [51] 1.3 9 10-4 3 9 10-4 9 9 10-4 – 3.0 9 10-3 3.1 9 10-3 3.4 9 10-3 3.5 9 10-3 3.6 9 10-3
[29] 1.2 9 10-4 3.9 9 10-4 7.0 9 10-3 2.0 9 10-3 2.9 9 10-3 3.2 9 10-3 3.4 9 10-3 3.4 9 10-3 3.4 9 10-3
PB [22] 8.3 9 10-4 9.4 9 10-4 1.0 9 10-3 1.1 9 10-3 – – – – –
PC [52] – 4.0 9 10-4 6.0 9 10-4 8 9 10-4 1.32 9 10-3 1.5 9 10-3 – – –
PE, High Density [53] 2.9 9 10-4 9.0 9 10-4 – – 4.1 9 10-3 4.5 9 10-3 – – 4.0 9 10-3
PEAI [54] – 4.0 9 10-4 6 9 10-4 8 9 10-4 1.4 9 10-3 1.5 9 10-3 1.8 9 10-3 1.9 9 10-3 2 9 10-3
(continued)
207
208

PMMA (Plexiglass) [13] 5.8 9 10-4 6.5 9 10-4 1.0 9 10-3 1.3 9 10-3 1.5 9 10-3 1.6 9 10-3 1.9 9 10-3 – –
[53] 3.3 9 10-4 6.0 9 10-4 – – – 1.6 9 10-3 1.7 9 10-3 1.8 9 10-3 2.0 9 10-3
PP [52] – 3.4 9 10-4 7.5 9 10-4 1.1 9 10-3 1.69 9 10-3 1.8 9 10-3 – – –
POM [52] – 9.3 9 10-4 2.30 9 10-3 3.2 9 10-3 5.0 9 10-3 4.8 9 10-3 – – –
Potassium b- K2O11Al2O3 [20] 2 9 10-1 5 9 10-1 5 9 10-1 4.2 9 10-1 1.6 9 10-1 1.3 9 10-1 – – –
alumina
Potassium bromide KBr [23] 3.3 1.8 7.2 9 10-1 4.0 9 10-1 – – – – –
KBr [55] – 1.8 7.4 9 10-1 4.5 9 10-1 1.4 9 10-1 1.1 9 10-1 7.4 9 10-2 –
KBr0.996(NO2)0.004 [23] – 6.8 9 10-3 2.2 9 10-2 4.0 9 10-2 4.4 9 10-2 4.3 9 10-2 – – –
K0.9996Li0.0004Br [55] – 3.3 9 10-1 3.0 9 10-1 2.7 9 10-1 1.4 9 10-1 – – – –
Potassium chloride KCl [23] – 4.4 1.9 – 3.2 9 10-1 – – – –
KCl [55] – 4.4 5.4 2.3 3.3 9 10-1 2.7 9 10-1 – – –
KCl0.9975(NO2)0.0025 [23] – 1.0 9 10-2 4.8 9 10-2 1.1 9 10-2 – – – – –
K0.9997Li0.0003Cl [55] – 2.1 9 10-1 7.4 9 10-1 6.5 9 10-1 3.4 9 10-1 2.7 9 10-1 – – –
Potassium iodide KI [23] 7.1 3.2 1.4 7.4 9 10-1 1.4 9 10-1 1.0 9 10-1 – – –
KI0.9994(NO2)0.0006 [23] 5.2 9 10-1 5.1 9 10-1 3.5 9 10-1 2.5 9 10-1 1.0 9 10-1 9.0 9 10-2 – – –
PS [22] 2.2 9 10-4 2.7 9 10-4 4.5 9 10-4 6.2 9 10-4 1.2 9 10-3 1.3 9 10-3 – – –
PVC [31] – – – – – – 1.4 9 10-3 1.5 9 10-3 1.6 9 10-3

[52] – 4.0 9 10-4 6.6 9 10-4 8.5 9 10-4 1.27 9 10-3 1.3 9 10-3 – – –
[53] 2.7 9 10-4 4.0 9 10-4 – – – – – – –
(continued)
- - -
Pyrex Glass§ SiO2 80 %, [26] 8 9 10-4 1.1 9 10-3 1.4 9 10-3 1.8 9 10-3 4.6 9 10-3 –
B2O3 13 %, [31] 1.0 9 10-3 1.2 9 10-3 1.5 9 10-3 2.0 9 10-3 4.5 9 10-3 5.8 9 10-3 7.8 9 10-3 9.2 9 10-3 1.1 9 10-2
Na2O 4 %
Al2O3 2 %
K2O 0.4 %
Li2O 0.2 %
RTV silicon [56] – – – – 1.8 9 10-1 2.3 9 10-1 2.7 9 10-1 3.0 9 10-1 2.7 9 10-1
Samarium Sm2Zr2O7 [44] – – 1.9 9 10-2 1.9 9 10-2 1.7 9 10-2 1.6 9 10-2 1.6 9 10-2 1.5 9 10-2 1.4 9 10-2
Zirconium
oxide
Selenium Se [26] 6 9 10-4 6 9 10-4 8 9 10-4 1 9 10-1 2 9 10-3 – – – –
Sapphire Al2O3, synthetic crystal [31] 2.30 29.00 157.00 138.50 11.00 4.50 1.50 8.2 9 10-1 4.7 9 10-1
Silicon carbide a-SiC, single crystal [31] 2.7 9 10-1 4.20 20.00 33.50 40.00 30.00 15.00 9.50 5.10
Silicon oxide SiO2, crystal [31] 1.85 13.45 5.45 3.40 4.3 9 10-1 3.0 9 10-1 1.8 9 10-1 1.3 9 10-1 9 9 10-2
[13] 1.1 9 10-3 1.2 9 10-3 1.9 9 10-3 2.5 9 10-3 6.4 9 10-3 7.4 9 10-3 1 9 10-2 1 9 10-2 2 9 10-2
Silver b-alumina Ag2O11Al2O3 [20] 6.0 9 10-3 7 9 10-3 5.7 9 10-3 6.7 9 10-3 1.7 9 10-2 2.1 9 10-2 – – –
Skutterudite Co4Sb12 [57] – – 4.3 9 10-1 3.9 9 10-1 3.2 9 10-1 2.6 9 10-1 1.8 9 10-1 1.4 9 10-1 8 9 10-2
Soda lime glass SiO2 70 %, [25] – 3 9 10-3 3 9 10-3 3.7 9 10-3 5 9 10-3 – – – –
(soft)§ Na2O 15 %,
CaO 10 %
Sodium b-alumina Na2O11Al2O3 [20] 1.6 9 10-3 2.7 9 10-2 2.1 9 10-2 2.0 9 10-2 2.4 9 10-2 2.5 9 10-2 – – –
Sodium Fluoride NaF, high purity [58] – 118 232 23 1.4 – – – –
NaF, low purity [58] – 40 26 13 1.4 – – – –
Teflon [29] 4.6 9 10-4 1.0 9 10-3 1.4 9 10-3 1.7 9 10-3 2.3 9 10-3 2.4 9 10-3 2.6 9 10-3 2.7 9 10-3 2.7 9 10-3
(continued)
209
210

Torlon [59] 1 9 10-4 1.6 9 10-4 6.1 9 10-4 7.2 9 10-4 1.1 9 10-3 1.1 9 10-3 1.5 9 10-3 1.7 9 10-3 2.6 9 10-3
Vespel SP1 [5] 8 9 10-5 2 9 10-4 5 9 10-4 8 9 10-4 2 9 10-3 2 9 10-3 3 9 10-3 3 9 10-3 4 9 10-3
YSGG:Er,Cr Y2.07Er0.93Sc1.45Cr0.03 [45] – 4.1 9 10-1 2.7 9 10-1 2.2 9 10-1 1 9 10-1 9 9 10-2 7 9 10-2 5.9 9 10-2 5.3 9 10-2
Ga3.52O12
YSGG:Nd,Cr Y2.95Nd0.05Sc1.72Cr0.03 [45] – 2.8 3.3 1.7 4 9 10-1 2.1 9 10-1 1 9 10-2 9.6 9 10-2 7.4 9 10-2
Ga3.25O12
Zinc oxide ZnO, hexagonal [41] – – – 5.3 2.6 2.0 1.3 7.2 9 10-1 5.4 9 10-1
Zinc selenide ZnSe, cubic [41] – – – 3.7 1.03 6.5 9 10-1 4.4 9 10-1 2.3 9 10-1 1.9 9 10-1
Zinc sulphide ZnS, cubic [41] – – – 3.6 1.55 1.02 7.0 9 10-1 3.8 9 10-1 2.7 9 10-1
Zinc telluride ZnTe, cubic [41] – – – 3.0 1.08 6.5 9 10-1 4.6 9 10-1 2.5 9 10-1 1.8 9 10-1
§
Table 9.7 Low temperature data of integrated thermal conductivity (W/cm) of various materials
Alumina [31] 9.8 9 10-3 1.925 9 10-1 1.6725 5.4725 49.635 86.26 143.51 179.26 –
Arsenicum [13] 3.4 9 10-3 1.3 9 10-2 2.95 9 10-2 4.7 9 10-2 9.805 9 10-2 1.5555 9 10-1 – – –
sulphide
AXM-5Q [5] 4 9 10-6 3.01 9 10-3 5.801 9 10-2 3.5801 9 10-1 5.925 12.175 37.175 74.675 164.675
Borosilicate [14] 2 9 10-3 9.8 9 10-3 2.78 9 10-2 4.88 9 10-2 – – – – –
Cadmium- [13] 8 9 10-3 3.2 9 10-2 5.385 9 10-2 5.755 9 10-2 9.836 9 10-2 1.9337 9 10-1 – – –
Germanium
Calcium [13] 1.8 9 10-3 9 9 10-3 3 9 10-2 5.65 9 10-2 2.987 9 10-1 – – – –
Potassium
Nitrate
Diamond [43] 1.34 18.65 119.15 299.15 1313.5 2138.5 3238.5 3746 –
G10CR Normal [29] 1.44 9 10-3 6.9 9 10-3 2.04 9 10-2 3.79 9 10-2 1.1235 9 10-1 1.861 9 10-1 3.561 9 10-1 5.611 9 10-1 1.0861
G10CR Warp [29] 1.46 9 10-3 7.85 9 10-3 2.485 9 10-2 4.635 9 10-2 1.48 9 10-1 2.53 9 10-1 5.08 9 10-1 8.18 9 10-1 1.583
Germanium [16] 3.4 9 10-3 1.42 9 10-2 3.47 9 10-2 5.82 9 10-2 1.905 9 10-1 3.4675 9 10-1 – – –
oxide
Kapton [46] 4.6 9 10-4 2.68 9 10-3 1.073 9 10-2 2.623 9 10-2 1.712 9 10-1 2.987 9 10-1 6.237 9 10-1 1.0987 2.4987
Kapton [29] 2.2 9 10-4 1.27 9 10-3 4.87 9 10-3 1.057 9 10-2 4.672 9 10-2 8.047 9 10-2 1.5547 9 10-1 2.4047 9 10-1 4.2547 9 10-1
Kevlar [47] 1.06 9 10-3 8.65 9 10-3 4.015 9 10-2 9.265 9 10-2 5.044 9 10-1 9.294 9 10-1 2.0544 3.4544 6.9044
Kevlar [29] 6 9 10-4 5.1 9 10-3 2.56 9 10-2 6.21 9 10-2 3.428 9 10-1 6.178 9 10-1 1.2928 2.0928 3.9428
Lithium beta- [20] 4.6 9 10-3 2.95 9 10-2 1.06 9 10-1 2.125 9 10-1 6.935 9 10-1 1.181 – – –
alumina
Macor [5] 1.6 9 10-3 1.3 9 10-2 5.8 9 10-2 3.88 9 10-1 1.752 2.0645 2.7395 3.4395 4.8895
Macor [49] 1.5 9 10-3 1.125 9 10-2 5.375 9 10-2 – – – – – –
Magnesium [31] 1.64 38 233 479.5 648.345 748.47 855.72 912.22 988.22
oxide
Mylar [22] 7.2 9 10-4 3.12 9 10-3 8.37 9 10-3 1.562 9 10-2 5.583 9 10-2 – – – –
Mylar [50] 7.6 9 10-4 3.34 9 10-3 9.39 9 10-3 1.729 9 10-2 4.902 9 10-2 – – – –
Nylon 6,6 [29] 2.4 9 10-4 1.77 9 10-3 3.872 9 10-2 8.372 9 10-2 1.1715 9 10-1 1.934 9 10-1 3.584 9 10-1 5.284 9 10-1 8.684 9 10-1
(continued)
211
212
PB [22] 1.66 9 10-3 6.97 9 10-3 1.667 9 10-2 2.717 9 10-2 – – – – –
PMMA ref1 [13] 1.16 9 10-3 4.85 9 10-3 1.31 9 10-2 2.46 9 10-2 6.71 9 10-2 1.0585 9 10-1 1.9335 9 10-1 – –
Potassium beta- [20] 4 9 10-1 2.5 7.5 12.1 14.05 17.675 – – –
alumina
Potassium [23] 6.6 21.9 34.5 40.1 – – – – –
bromide
Potassium [23] 14.2 45.1 68.1 78.8 21.42 24.42 – – –
iodide
PS [22] 4.4 9 10-4 1.91 9 10-3 5.51 9 10-3 1.086 9 10-2 4.34 9 10-2 7.464 9 10-2 – – –
-1
Pyrex glass ref [31] 2 9 10-3 8.6 9 10-3 2.21 9 10-2 3.96 9 10-2 1.5475 9 10 2.835 9 10-1 6.235 9 10-1 1.0485 2.0585
20
Selenium [26] 1.2 9 10-3 4.8 9 10-3 1.18 9 10-2 5.158 9 10-1 2.497 – – – –
Sapphire [31] 4.6 98.5 1028.5 2506 3651.75 3845.5 3995.5 4053.5 4118
Silicon carbide [31] 5.4 9 10-1 13.95 134.95 402.45 1760.75 2635.75 3760.75 4373.25 5103.25
Silicon oxide [31] 3.7 49.6 144.1 188.35 93.405 102.53 114.53 122.28 129.23
ref 20
Silicon oxide [13] 2.2 9 10-3 9.1 9 10-3 2.46 9 10-2 4.66 9 10-2 2.1165 9 10-1 3.8415 9 10-1 8.1915 9 10-1 1.31915 2.81915
ref 1
Silver beta- [20] 1.2 9 10-2 5.1 9 10-2 1.145 9 10-1 1.765 9 10-1 5.6365 9 10-1 1.03,865 – – –
alumina
Sodium beta- [20] 3.2 9 10-3 8.9 9 10-2 3.29 9 10-1 5.34 9 10-1 1.054 1.6665 – – –
alumina
Teflon [29] 9.2 9 10-4 5.3 9 10-3 1.73 9 10-2 3.28 9 10-2 9.57 9 10-2 1.5445 9 10-1 2.7945 9 10-1 4.1195 9 10-1 6.8195 9 10-1
Torlon [59] 2 9 10-4 9.8 9 10-4 4.83 9 10-3 1.148 9 10-2 4.349 9 10-2 7.1 9 10-2 1.36 9 10-1 2.16 9 10-1 4.31 9 10-1
-1 -1 -1
Vespel [5] 1.6 9 10-4 1 9 10-3 4.5 9 10-3 1.1 9 10-2 6.66 9 10-2 1.166 9 10 2.416 9 10 3.916 9 10 7.416 9 10-1
9.3 Crystalline Materials 213
Table 9.8 Thermal conductivity data for some crystals with original references and our simple
fit if authors do not give a fit of their data
Material Fit of k(W/cm K) Temperature range(K) Peak temperature(K) Reference
Li3N 1.7 9 10-2 T2.9 0.08-0.8 – [60]
Li3N 7.7 T-0.8 35-150 12 [60]
TiO2 4.3 9 10-2 T2.0 2-4.6 – [61]
TiO2 1.38 9 103 T-1.92 20-90 16 [61]
KCl 60 T-1.2 7-78 – [61]
LiF 4.6 9 103 T-1.8 23-85 14 [61]
KBr 0.43 T2.4 0.3-1.3 – [62]
KBr 25.6 T-1.2 10-115 5 [62]
KBr0.53I0.47 0.28 T3.0 0.05-0.3 – [62]
KBr0.53I0.47 1.4 T-1.1 9-82 4.6 [62]
La2CuO4 0.75 9 10-3 T2.5 0.1-0.8 – [63]
La2CuO4 0.76 9 10-3 T2.5 1.7-4 – [63]
La1.9Sr0.1CuO4 3.9 9 10-3 T1.9 0.07-1.1 – [63]
La1.9Sr0.1CuO4 6.8 9 10-3 T1.4 3-7.5 – [63]
Al2O3 2.7 9 10-3 T2.5 2-8 [64]
Al2O3 0.29 9 10-3 T2.7 0.1-2 – [65]
9.3 Crystalline Materials
Table 9.8.
References
1. Sharma, J.: Heat conductivities below 1 K. I. Cryogenics 7(1), 141–156 (1967)

2. Dupré, A., Van Itterbeek, A., Michiels, L.: Heat conductivity of copper below 1. K. Phys.
Lett. 8(2), 99–100 (1964)
3. Rosenberg, H.M.: The thermal conductivity of metals at low temperatures. Philos. Trans.
Royal Soc. Lond. Ser A Math. Phys. Sci. 247, 441–497 (1955)
4. Sladek, R.J.: Thermal conductivity of Indium-Thallium alloys at low temperatures. Phys.
Rev. 97(4), 902 (1955)
5. Woodcraft, A.L., Gray, A.: A low temperature thermal conductivity database. In: AIP
Conference Proceedings 2009, p. 681, 2009
6. Gianno, K., Sologubenko, A., Liechtenstein, L., Chernikov, M., Ott, H.: Low-temperature
thermoelectric power and thermal conductivity of two icosahedral Al-Pd-Re quasicrystals.
Ferroelectrics 250(1), 249–252 (2001)
7. O’Hara, S., Anderson, A.: Resonant scattering of thermal phonons by dislocations in
superconducting aluminum. Phys. Rev. B 9(9), 3730 (1974)
Cryogenics 33(7), 729–731 (1993)
9. Rowell, P.: The thermal conductivity of some superconductors. Proc. Royal Soc. Lond. Ser.
A: Math. Phys. Sci. 254(1279), 542–550 (1960)
10. O’Hara, S., Sellers, G., Anderson, A.: Influence of hydrogen on the thermal conductivities of
superconducting Nb and Ta. Phys. Rev. B 10(7), 2777 (1974)
214 9 Data of Thermal Conductivity
11. Anderson, A., Satterthwaite, C., Smith, S.: Thermal conductivity of superconducting
niobium. Phys.l Rev. B 3, 3762–3764 (1971)
12. Cotignola, J., De La Cruz, F., De La Cruz, M., Platzeck, R.: Thermal conductivity of
polycrystalline zinc below 1 k. Rev. Sci. Instrum. 38(1), 87–92 (1967)
13. Pohl, R.O., Liu, X., Thompson, E.: Low-temperature thermal conductivity and acoustic
attenuation in amorphous solids. Rev. Mod. Phys. 74(4), 991 (2002)
14. Zaitlin, M.P., Anderson, A.: Phonon thermal transport in noncrystalline materials. Phys. Rev.
B 12(10), 4475 (1975)
15. Anderson, A., Reese, W., Wheatley, J.: Thermal conductivity of some amorphous dielectric
solids below 1 K. Rev. Sci. Instrum. 34(12), 1386–1390 (1963)
17. Toews, W., Zhang, S.S., Ross, K., Dabkowska, H., Gaulin, B., Hill, R.: Thermal Conductivity
of Ho_ 2 Ti_ 2 O_ 7 along the[111] Direction. Phys. Rev. Lett. 110(21), 217209 (2013)
18. Barucci, M., Gottardi, E., Peroni, I., Ventura, G.: Low temperature thermal conductivity of
Kapton and Upilex. Cryogenics 40(2), 145–147 (2000)
19. Ventura, G., Barucci, M., Gottardi, E., Peroni, I.: Low temperature thermal conductivity of
Kevlar. Cryogenics 40(7), 489–491 (2000)
20. Anthony, P., Anderson, A.: Thermal conductivities of Li, Na, K, and Ag b-alumina below
300 K. Phys. Rev. B 14(12), 5198 (1976)
Cryogenics 40(7), 465–467 (2000)
22. Freeman, J., Anderson, A.: Thermal conductivity of amorphous solids. Phys. Rev.
B Condens. Matter 34(8), 5684–5690 (1986)
23. Narayanamurti, V., Seward, W., Pohl, R.: Rotational degrees of freedom of molecules in
solids. II. The nitrite ion in alkali halides. Phys. Rev. 148(1), 481 (1966)
24. Risegari, L., Barucci, M., Olivieri, E., Ventura, G.: Low temperature thermal conductivity of
PVC. J. Low Temp. Phys. 144(1–3), 49–59 (2006)
25. Zeller, R., Pohl, R.: Thermal conductivity and specific heat of noncrystalline solids. Phys.
Rev. B 4(6), 2029 (1971)
26. Morgan, G., Smith, D.: Thermal conduction in glasses and polymers at low temperatures.
J. Phys. C: Solid State Phys. 7(4), 649 (1974)
27. Scott, T., Giles, M.: Dislocation scattering in Teflon at low temperatures. Phys. Rev. Lett. 29,
642–643 (1972)
30. Powell, R.L., Hall, W.J., Roder, H.M.: Low-Temperature Transport Properties of
Commercial Metals and Alloys II. Aluminums. J. Appl. Phys. 31(3), 496–503 (1960)
31. Touloukian, Y., Powell, R., Ho, C., Klernens, P.: Thermophysical Properties of Matter:
Thermal Conductivity. IFI Plenum, New York (1970)
32. MacDonald, D., White, G., Woods, S.: Thermal and electrical conductivities of the alkali
metals at low temperatures. Proc. R. Soc. Lond. A 235(1202), 358–374 (1956)
33. Ogawa, M., Li, R., Hashimoto, T.: Thermal conductivities of magnetic intermetallic
compounds for cryogenic regenerator. Cryogenics 31(6), 405–410 (1991)
34. Kemp, W., Klemens, P., Tainsh, R.: The lattice thermal conductivity of copper alloys: Effect
of plastic deformation and annealing. Phil. Mag. 4(43), 845–857 (1959)
35. Glassbrenner, C., Slack, G.A.: Thermal conductivity of silicon and germanium from 3 K to
the melting point. Phys. Rev. 134(4A), A1058 (1964)
36. Kumar, A., Verma, G.: Phonon conductivity of Mg2Sn. Phys. Rev. B 1(2), 488 (1970)
37. Bychkov, Y.F., Herzog, R., Khukhareva, I.: Thermal conductivity and electrical resistivity of
Nb Ti alloys at low temperatures. Cryogenics 21(12), 741–745 (1981)
References 215
38. Maldonado, O.: Pulse method for simultaneous measurement of electric thermopower and
heat conductivity at low temperatures. Cryogenics 32(10), 908–912 (1992)
39. Sichel, E., Miller, R., Abrahams, M., Buiocchi, C.: Heat capacity and thermal conductivity of
hexagonal pyrolytic boron nitride. Phys. Rev. B 13(10), 4607 (1976)
40. Duclaux, L., Nysten, B., Issi, J., Moore, A.: Structure and low-temperature thermal
conductivity of pyrolytic boron nitride. Phys. Rev. B 46(6), 3362 (1992)
41. Slack, G.A.: Thermal conductivity of II-VI compounds and phonon scattering by Fe^{2+}
impurities. Phys. Rev. B 6(10), 3791 (1972)
42. Puyet, M., Dauscher, A., Lenoir, B., Bellouard, C., Stiewe, C., Müller, E., Hejtmanek, J.,
Tobola, J.: Influence of Ni on the thermoelectric properties of the partially filled calcium
skutterudites Ca_{y} Co_{4 - x} Ni_{x} Sb_{12}. Phys. Rev. B 75(24), 245110 (2007)
43. Berman, R., Simon, F., Ziman, J.: The thermal conductivity of diamond at low temperatures.
Proc. R. Soc. Lond. A 220(1141), 171–183 (1953)
44. Pan, W., Wan, C., Xu, Q., Wang, J., Qu, Z.: Thermal diffusivity of samarium–gadolinium
zirconate solid solutions. Thermochim. Acta 455(1), 16–20 (2007)
45. Sirota, N., Popv, P., Ivanov, I.: The Thermal Conductivity of Monocrystalline Gallium
Garnets Doped with Rare-Earth Elements and Chromium in the Range 6–300 K. Cryst. Res.
Technol. 27(4), 535–543 (1992)
46. Benford, D., Powers, T., Moseley, S.: Thermal conductivity of Kapton tape. Cryogenics
39(1), 93–95 (1999)
47. Ventura, G., Martelli, V.: Thermal conductivity of Kevlar 49 between 7 and 290 K.
Cryogenics 49(12), 735–737 (2009)
48. Thacher, P.D.: Effect of boundaries and isotopes on the thermal conductivity of LiF. Phys.
Rev. 156(3), 975 (1967)
49. Lawless, W.: Dielectric and thermal properties of a machinable glass—ceramic at low
50. Radebaugh, R.: Thermal conductance of indium solder joints at low temperatures. Rev. Sci.
Instrum. 48, 93 (1977)
51. Timmerhaus, K.D., Flynn, T.M.: Cryogenic Process Engineering, vol. 580. Plenum press,
New York (1989)
52. Choy, C., Greig, D.: The low temperature thermal conductivity of isotropic and oriented
polymers. J. Phys. C: Solid State Phys. 10(2), 169 (1977)
53. Hartwig, G.: Polymer Properties at Room and Cryogenic Temperatures. Springer, New York
(1994)
54. Hust, J., Boscardin, R.: Thermal conductivity of polyester-amide-imide film. Cryogenics
21(5), 297–298 (1981)
55. Baumann, F., Harrison, J., Pohl, R., Seward, W.: Thermal conductivity in mixed Alkali
Halides: KCl: Li and KBr: Li. Phys. Rev. 159(3), 691 (1967)
56. Baudot, A., Mazuer, J., Odin, J.: Thermal conductivity of a RTV silicone elastomer between
1.2 and 300 K. Cryogenics 38(2), 227–230 (1998)
57. Puyet, M., Candolfi, C., Chaput, L., Da Ros, V., Dauscher, A., Lenoir, B., Hejtmanek, J.:
Low-temperature thermal properties of n-type partially filled calcium skutterudites. J. Phys.:
Condens. Matter 18(49), 11301 (2006)
58. Jackson, H.E., Walker, C.T.: Thermal conductivity, second sound, and phonon-phonon
interactions in NaF. Phys. Rev. B 3(4), 1428 (1971)
59. Barucci, M., Olivieri, E., Pasca, E., Risegari, L., Ventura, G.: Thermal conductivity of Torlon
between 4.2 and 300 K. Cryogenics 45(4), 295–299 (2005)
60. Guckel, H.: Silicon microsensors: construction, design and performance. Microelectron. Eng.
15(1), 387–398 (1991)
61. Berman, R., Foster, E., Ziman, J.: The thermal conductivity of dielectric crystals: the effect of
isotopes. Proc. R. Soc. Lond. A 237(1210), 344–354 (1956)
62. Nathan, B., Lou, L., Tait, R.: Low temperature thermal properties of mixed crystal KBr KI.
Solid State Commun. 19(7), 615–617 (1976)
216 9 Data of Thermal Conductivity
63. Morelli, D., Doll, G., Heremans, J., Peacor, S., Uher, C., Dresselhaus, M., Cassanho, A.,
Gabbe, D., Jenssen, H.: Thermal conductivity of single crystal lanthanum cuprates at very
low temperature. Solid State Commun. 77(10), 773–776 (1991)
64. Gloos, K., Mitschka, C., Pobell, F., Smeibidl, P.: Thermal conductivity of normal and
superconducting metals. Cryogenics 30(1), 14–18 (1990)
65. Locatelli, M., Arnaud, D., Routin, M.: Thermal conductivity of some insulating materials
materials below 1 K. Cryogenics 16(6), 374–375 (1976)
66. Meissner, M., Abens, S., Strelow, P.: Hahn-Meitner Institute Report, Berlin (2000)

(International Cryogenics Monograph Series) Guglielmo Ventura, Mauro Perfetti (Auth.) - Thermal Properties of Solids at Room and Cryogenic Temperatures-Springer Netherlands (2014)

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

(International Cryogenics Monograph Series) Guglielmo Ventura, Mauro Perfetti (Auth.) - Thermal Properties of Solids at Room and Cryogenic Temperatures-Springer Netherlands (2014)

Загружено:

Авторское право:

Доступные форматы

International Cryogenics Monograph Series

For further volumes:

Thermal Properties of Solids

ISSN 0538-7051 ISSN 2199-3084 (electronic)

Library of Congress Control Number: 2014941685

Springer Science+Business Media Dordrecht 2014

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

If I were a metal, I would definitely say

Part I Heat Capacity

2 How to Measure Heat Capacity at Low Temperatures . . . . . . . . . 39

2.10 Small Sample Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . 61

3 Data of Specific Heat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

Part II Thermal Expansion

5 How to Measure the Thermal Expansion Coefficient

6 Data of Thermal Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

Part III Thermal Conductivity

7 Electrical and Thermal Conductivity . . . . . . . . . . . . .......... 131

8 How to Measure Thermal Conductivity . . . . . . . . . . . . . . . . . . . . 169

9 Data of Thermal Conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

G. Ventura and M. Perfetti, Thermal Properties of Solids at Room 3

alloys or nonstoichiometric compounds, where mass is more appropriate than

Fig. 1.1 Specific heat as a

1.2 Lattice Specific Heat

where r is the number of atoms per molecule, NA is Avogadro’s number

Fig. 1.2 Debye temperature

where x : hx/(2pkBT) (x is the frequency, h is the Planck’s constant). The

for temperatures T \ hD/10. The parameter hD can be identified as the temperature

heat measurements. The Debye’s temperature evaluated from data of ultrasonic

1.3 Electronic Specific Heat

When Sommerfeld applied Fermi–Dirac statistics to free electrons in metals, it

where g(EF) = (3/2)(NA/EF) is the volumetric density of states calculated at the

where c is called the Sommerfeld constant. Knowing that EF is of the order of

Fig. 1.4 Specific heat c of

Dividing c by T, it is then possible to obtain, by a simple fit, the value of c. The

1.4 Electronic Specific Heat in Superconducting Materials

Materials which exhibit the phenomenon of superconductivity enter into a new

Fig. 1.5 Specific heat of Hg

In Fig. 1.6a, Dce at Tc is about 2.12 mJ/mol K, giving c = 1.30 mJ/mol K, in

Fig. 1.7 Specific heat of

where Hc is the temperature-dependent critical field, V is the volume, and Sn and Ss

At T = Tc and Hc = 0, the first term in (1.20) vanishes, giving

At very low temperatures, the second derivative term in (1.20) dominates

1.5 Specific Heat Contributions from Transitions

discontinuous: transitions of this type are known as second-order transitions. b and

The effect of a second-order transition on specific heat is shown in Fig. 1.8b.

heat of several structural alloys with different heat treatments for T \ 20 K,

1.6 Magnetic Specific Heat

1.6.1 Magnetic Order and Specific Heat

If a material contains atoms which can be permanently or temporary magnetized

dQ ¼ TdS ¼ dE þ pdV þ MdH: ð1:24Þ

In this formula, M is the modulus of magnetization, defined as the vector sum of

dQ ¼ TdS ¼ dE0 þ pdV HdM ð1:25Þ

dQ ¼ TdS ¼ dE0 HdM: ð1:26Þ

The heat capacity at constant magnetization is independent on an external

case of Cp and CV, a useful estimation of the difference CH - CM has to be

Equation (1.29) can be demonstrated by analogy with

Since, in this case,

By substituting (1.32) into (1.31), one obtains (1.29).

Fig. 1.9 Alignment of spins

permanently magnetized due to the strong coupling of magnetic moments

Fig. 1.10 Specific heat of

and for antiferromagnetic compounds

cM ¼ ð465T þ 0:56T 2 Þ½lJ=gK: ð1:39Þ

T hDi ! ci ¼ constant ð1:48Þ

T hDi ! ci ¼ aT 1þd with 0\d\2: ð1:49Þ

cðTÞ ¼ ð1:22 0:01Þ 107 ½JK 1 g1 : ð2:12Þ