Вы находитесь на странице: 1из 7

Composites: Part A 35 (2004) 423429

www.elsevier.com/locate/compositesa

Thermal conductivity, thermal diffusivity, and specific heat capacity


of particle filled polypropylene
Bernd Weidenfellera,b,*, Michael Hoferc, Frank R. Schillingc
a
Institute of Polymer Science and Plastics Processing, Technical University of Clausthal, Agricolastrasse 6, D-38678 Clausthal-Zellerfeld, Germany
b
Institute of Metallurgy, Robert-Koch-Str. 42, D-38678 Clausthal-Zellerfeld, Germany
c
GeoForschungsZentrum Potsdam, Section 4.1 Experimental Geochemistry and Mineral Physics, Telegrafenberg, D-14473 Potsdam, Germany
Received 22 July 2003; revised 19 November 2003; accepted 20 November 2003

Abstract
By the addition of metal and oxide particles to plastics, thermal transport properties, heat capacity, and density of polymers can be varied
systematically. Composites samples of polypropylene (PP) with various fillers in different fractions (up to 50 vol%) were prepared with an
injection moulding process to study the evolution of the properties as a function of filler content. Standard filler materials like magnetite,
barite, talc, copper, strontium ferrite and glass fibres were used. Thermal diffusivities, specific heat capacities and densities of the prepared
composite samples were measured, and thermal conductivities were derived. Thermal conductivity of the polypropylene is increased from
0.27 up to 2.5 W/(m K) with 30 vol% talc in the polypropylene matrix. Thermal conductivities of the filled polypropylene samples are
compared with the modelled values according to Hashin and Shtrikman. The interconnectivity of the particles in the polypropylene matrix is
derived from a comparison between modelled and measured thermal conductivity values. For higher talc and glass fibre content in PP
plastics, a complete interconnectivity is achieved, while copper particles in PP show a very poor interconnectivity. Specific heat capacities
and thermal diffusivities of magnetite and barite filled polypropylene were measured in the temperature range from 300 to 395 K.
q 2004 Elsevier Ltd. All rights reserved.
Keywords: A. Polymer-matric Composites (PMCs); B. Thermal Properities; D. Physical Methods of testing; E. Injection Moulding

1. Introduction graphite [4], aluminum nitrides [3,5] or magnetite particles


[6]. Furthermore, the cooling behaviour in the mould of the
Commonly used plastics, such as polypropylene (PP) or injection moulding machine is influenced by the thermal
polyamide are electrical insulators with a low thermal properties of the used filler, where a high thermal diffusivity
conductivity. New applications, like heat sinks in electronic is responsible for a high efficiency of the moulding process.
packaging, require new composites with higher thermal A high thermal diffusivity is achieved for samples with a
conductivity. By the addition of fillers to plastics the thermal high thermal conductivity and low density and low heat
behaviour of polymers can be increased significantly. Such capacity.
filled polymers with higher thermal conductivities than Published values of thermal conductivities of the same
unfilled ones become more and more an important area of filler materials in different polymer matrices vary drastically
study because of the wide range of applications, e.g. in and a comparison of different materials is difficult [6], as
electronic packaging in applications with decreasing most of these studies do not vary systematically the filler
geometric dimensions and increasing output of power, like content and do not describe detailed enough the filler and
in computer chips or in electronic packaging [1 3]. The the microstructure of the composite samples. Therefore,
higher thermal conductivity can be achieved by the use of a polypropylene samples with different commercially avail-
suitable filler such as aluminum [1], carbon fibres and able fillers were prepared by extrusion and injection
moulding using various volume fractions of filler content
* Corresponding author. Present address: Institute of Metallurgy, Robert-
Koch-Str. 42, D-38678 Clausthal-Zellerfeld, Germany. Tel.: 49-5323-
to systematically vary density and thermal transport
723708; fax: 49-5323-723184. properties of these composites. The thermal properties,
E-mail address: bernd.weidenfeller@tu-clausthal.de (B. Weidenfeller). heat capacity and density of these samples were investigated
1359-835X/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.compositesa.2003.11.005
424 B. Weidenfeller et al. / Composites: Part A 35 (2004) 423429

and correlated to the amount and the kind of the filler r and specific heat capacity cp according to
material.
l acp r 4
while thermal diffusivity a quantifies the thermal equili-
2. Theoretical considerations bration process and the thermal conductivity l describes the
ability of a body to transport heat.
The thermal conductivity of a homogeneous compo-
site can be derived according to the model of Hashin
Shtrikman (HS). The thermal conductivity of a composite 3. Experimental
consisting of a homogenous matrix M with a thermal
conductivity lM ; spherical particles P with a thermal con- 3.1. Materials and sample preparation
ductivity of lP ; and a volume fraction of spheres xP relates
to [6] Test materials were supplied by Minelco B.V. (The
Netherlands). Minelco B.V. prepared in cooperation with
2lM lP 2 2xP lM 2 lP
lComp;l lHS2 lM 1 RTP s.a.r.l (France) several polypropylene (PP) compounds
2lM lP xP lM 2 lP with various fillers (Fe3O4, BaSO4, Cu, glass fibres, talc,
with sHS2 the HS lower boundary [7] for a matrix with a SrFe12O19, cf. Table 1) in an extrusion process similar to
low thermal conductivity lM , lP : that one described in Ref. [6]. The filler fraction was varied
If the matrix consists of a highly thermally conducting from 0 to 50 vol%.
phase surrounding particles of low thermal conductivity, the With an injection moulding machine (Allrounder 320C
HS upper boundary sHS would describe the conductivity 600-250, Arburg, Germany) standard samples for measuring
behaviour lM . lP ; i.e. a high conducting talc matrix tensile properties together with a rod for thermal measure-
would surround polymer particles ments of 10 mm diameter and 130 mm length were
prepared in one mould. Selected injection moulding
parameters are listed in Table 2. For thermal diffusivity
2ltalc lPP 2 2xPP ltalc 2 lPP measurements, small cylinders of 10 mm diameter and
lComp;l lHS ltalc 2
2ltalc lPP xPP ltalc 2 lPP 5 6 mm height were cut out of the injection-moulded rods.
Eqs. (1) and (2) only depend on the volume fraction of
3.2. Experimental methods
the materials and do not take into account the micro-
geometry and morphology of filler and matrix, while using a
3.2.1. Specific density
spherical approximation for particle shape. For a compari-
The specific density rC of the composite materials was
son of measurements of particle filled polymer composites
determined using Archimedes principle, where the volume
with Eqs. (1) and (2) the particles have to be randomly
is measured by the buoyancy in a fluid with known density.
distributed in the matrix.
The weight and buoyancy measurements were performed
Using the model conductivities after Hashin and Shtrik-
with an electromagnetic balance (LA203S, Sartorius,
man [7], lHS and lHS2 ; the interconnectivity of the
Germany). Isopropyl was chosen as fluid. The measurement
conducting phase Xinterconnected can be revealed according to
conditions as proposed in standard DIN 53479 [13] were
Schilling and Partzsch [12]
applied. The accuracy of density measurements is # 0.5%.
lmeasured 2 lHS2
Xinterconnected 3 3.2.2. Specific heat measurements
lHS 2 lHS2
Specific heat capacity measurements were performed
The as determined interconnectivity is a relative measure using a modulated differential scanning calorimeter (DSC
to an ideally interconnected network of the high thermally 2920 modulated, TA Instruments). Measurements were
conducting phase. As the conductivity is dominated by the carried out in the temperature range from 2 10 to 50 8C with
smallest junctions of the path of high conductivity, the a heating rate of 5 K/min. The DSC was calibrated in the
resultant interconnectivity is dominated by these junctions. same temperature region before each experiment, using a
If, for example, the high conducting phase is interconnected sapphire sample as standard, with a well-known specific
by point contacts, most of the high conducting phase will heat capacity. This leads to an accuracy of # 2% for the
not directly contribute to the overall conductivitynot presented data. Heat capacity values at 23 8C are presented.
ideally interconnectedand the determined interconnectiv-
ity is dominated by point contacts. This especially holds 3.2.3. Thermal diffusivity measurements
true, if the difference between the conductivities of the The thermal diffusivity of the polymers is measured by a
different phases is huge. transient method, closely related to laser-flash experiments
In an homogenous body, thermal diffusivity a and [14]. The used transient technique is especially optimised
thermal conductivity l are inter-related by specific density for measurements of polyphase aggregates. A temperature
B. Weidenfeller et al. / Composites: Part A 35 (2004) 423429 425

Table 1
Selected properties of filler materials

Magnetite Barite Talc Strontium ferrite Copper Glass fibres


(Fe3O4) (BaSO4) (Mg3[Si4O10][OH]2) (SrFe12O19) (Cu)

Thermal conductivity (W/(m K)) a: 4.61 ^ 0.42 : 2.07 ^ 0.02 : 1.76 ^ 0.00 .3 401 1.21.5
a: 5.10 k: 2.92 ^ 0.07 k: 10.69 ^ 1.35
l11 : 9.7 a: 1.72 ^ 0.04 a: 2.97
a: 3.00 ^ 0.10
a: 6.10 ^ 0.90
Ref. [8] [8] [8] [9] [10] [11]
Mean particle diameter (mm) 9 1.5 2.0 1.5 15 11
Particle shape Irregular Irregular Platelet Irregular Fibre
Density (g/cm3) 5.1 4.48 2.78 5.11 8.94 2.58

a denotes measurements on monomineralic aggregates. Directions of anisotropy are specified by the thermal conductivity tensor l11 ; l22 ; l33 ; where l33
is parallel to the crystals optical c-axis; and the optical c-axis lies within the plane defined by l11 and l22 or by the thermal conductivity components normal or
parallel to the direction of maximum thermal conductivity ( , k).

signal is transferred to the upper side of the sample and glass fibre and barite filled polypropylene composites is
registered by a thermocouple. The transferred temperature relative poor, whereas magnetite and strontium ferrite
signal starts a thermal equilibration process in the specimen, increases the thermal conductivity of the PP significantly.
which is recorded by a difference thermocouple on the rear The enhancement of thermal conductivity of polymers by
surface and used for the evaluation of thermal diffusivity. the use of magnetite or strontium ferrite is in the same range
The thermal diffusivity of the sample is evaluated from the as if standard fillers like aluminum oxide (Al2O3) or
temperature time evolution recorded at the two thermo- aluminum nitride (AlN) are used to increase thermal
couples using a one-dimensional finite difference scheme, conductivity of polymers [6].
which allows consideration of heat losses from the sample To enhance the density of polymers, both barite and
surfaces [15]. The finite difference scheme is based on a magnetite are used. However, the competing material to
discrete formulation of diffusivity equation magnetite for the increase of density of polymersbarite
  ! has an apparent lower thermal conductivity (lbarite <
T 2 T
a 5 2 W/(m K)) than magnetite (lmagnetite < 5 W/(m K)) and
t x 2
strontium ferrite (lstrontium ferrite . 3 W/(m K) approximated
where t is time, a thermal diffusivity, x distance, and T from its structure and chemical composition according to
temperature. Slack [9]) (Table 1).
The temperature time curve recorded by difference Copper (lcopper < 400 W/(m K)) has a much higher
thermocouple is sensitive to the thermal diffusivity of the thermal conductivity than talc (ltalc < 10:6 W/(m K)) and
sample. Similarly to laser-flash experiments, the evaluation the other used filler materials. Surprisingly, the thermal
procedure does not require the absolute heat transfer. A least
square algorithm is utilized to determine the thermal
diffusivity, while varying systematically the thermal
diffusivity value in the finite difference scheme. A detailed
description of the apparatus is given by Schilling [15]. The
accuracy of the measurements of the polyphase aggregates
is 3 5%.

4. Results

The thermal conductivities of the investigated samples


are plotted in Fig. 1. The increase in thermal conductivity of

Table 2
Selected injection moulding parameters during preparation of sample rods
for measurements of thermal diffusivity

Mass temperature 200 8C


Mould temperature 20 8C
Cycle time 70.8 s
Injection pressure 6 107 Pa Fig. 1. Thermal conductivity of injection-moulded polypropylene-particle-
composites with different fillers and various filler fractions.
426 B. Weidenfeller et al. / Composites: Part A 35 (2004) 423429

Fig. 2. Comparison of specific heat capacities at room temperature of several injection-moulded particle filled polypropylene samples for several filler
fractions.

conductivity of talc filled PP are comparably high polypropylene samples are very low. For filler fractions
(< 2.5 W/(m K) for 30 vol% filler) and the copper filled higher than 30 vol% the specific heat capacities of barite
PP shows much lower conductivity values (< 1.25 W/(m K) and copper filled PP is significant lower than the specific
for 30 vol% filler) compared to the values of the other heat capacities of the other investigated materials. While the
composite materials (Fig. 1). specific heat capacity of talc and magnetite filled PP are
about the same for 15 and 30 vol% filler fraction, the
5. Discussion thermal diffusivity of talc filled PP is considerably higher
(cf. Fig. 3) than the diffusivity of the magnetite filled PP and
The high thermal conductivity of the talc filled PP can be the other investigated composites. As expected from the
explained by the relative high thermal conductivity of talc high thermal conductivity of copper, the thermal diffusivity
ltalc 10:69 W/(m K) parallel to the direction of maximum of copper filled PP at 30 vol% filler fraction is clearly higher
thermal conductivity (cf. Table 1). As the thermal than the diffusivities of the other investigated composites
conductivity of copper is approximately 40 times higher except talc filled PP.
than the thermal conductivity of talc one might expect that
the thermal conductivity of copper filled polymers is much
higher than the thermal conductivity of the talc filled PP.
However, the opposite is observed. The talc filled PP shows
a higher thermal conductivity ltalc;PP < 2:5 W/(m K) that is
twice as high as the copper filled PP with a thermal
conductivity of approximately lCu;PP < 1:25 W/(m K) at a
filler fraction of 30 vol% for both composites. One might
regard that the density of copper is approximately three
times higher than the density of talc, and relates the large
difference in the thermal conductivity between the different
composites as a result of the differences of heat capacities
and density (Eq. (4)). However, the difference in density
(3 ) and heat capacity (, 2 ) (Fig. 2) of the copper and
talc filled PP cannot explain the high discrepancy in the
observed thermal conductivity and thermal diffusivity
behaviour.
Figs. 2 and 3 show the specific heat capacities and the
thermal diffusivities of the investigated materials. It can be Fig. 3. Comparison of thermal diffusivity at room temperature of injection-
seen, that the heat capacities of the copper filled moulded particle filled polypropylene samples for different filler fractions.
B. Weidenfeller et al. / Composites: Part A 35 (2004) 423429 427

fibres and talc form a complete interconnected network.


This high value of interconnectivity might be due to an
alignment of the glass fibres and the talc particles in the
polypropylene matrix (cf. Fig. 5) into the flow direction
during the injection moulding process of the sample rods.
This behaviour simulates quite well polypropylene particles
in a glass matrix or polypropylene particles in a talc matrix,
respectively. However, the model conductivities after
Hashin and Shtrikman require spherical particles. This
condition is not fulfilled for the talc and glass fibre fillers.
However, the isotropic approach by HS gives a first
approximation for the interconnectivity behaviour of these
composites, as the upper and lower HS boundaries describe
maximum and minimum estimates for macroscopic iso-
Fig. 4. Interconnectivity of filler particles in polypropylene at a filler tropic two phase compositesregardless of the detailed
fraction of 30 vol% at room temperature (for details see text).
microstructure.
In Fig. 4 the very low interconnectivity of the copper
The thermal transport properties cannot be explained filled polypropylene is remarkable. This interconnectivity is
solely by the differences in the properties of the filler the result of poor contacts between the copper particles and
materials. Therefore, the microstructure and texture has to indicates that the copper particles are well wetted by the PP
be taken into account. As described by Eqs. (1) and (2), the matrix. The differences may be the result of a different
thermal conductivity of composites strongly depends on the wetting behaviour of filler and matrix or due to a clustering
thermal conductivity of the matrix and the interconnectivity of barite and magnetite particles. Therefore, the clear
of the higher conducting phase (Eq. (3)). The observed differences in the interconnectivity between magnetite or
different values of thermal conductivities in talc and copper barite filled PP and copper filled PP cannot be explained by
filled PP can be related to the interconnectivity of the the differences in particle sizes alone and therefore further
particles using Eq. (3). research is required to clarify the experimental results.
Fig. 4 shows the interconnectivity of the investigated As can be seen in Figs. 1 3 the thermal conductivity, the
materials for a filler fraction of 30 vol%. For the specific heat capacity and the thermal diffusivity of
calculations of the upper and lower HS boundary the bold- magnetite filled polypropylene samples are higher than the
faced values of thermal conductivity of Table 1 are used. corresponding values of polypropylene with the competing
The high interconnectivity around Xinterconnect < 1 of the filler material barite. This also holds true at temperatures
glass fibre and talc filled polypropylene is conspicuous. For higher than room temperature. Fig. 6 shows the dependence
interconnectivity close to 1 the measured values reach the of the specific heat capacity of magnetite and barite in the
upper HS boundary (Eq. (2)). This means that the glass temperature range from 295 to 385 K. The heat capacities

Fig. 5. (a) SEM micrograph of a fracture of the glass fibre filled PP sample with glass fibres aligned parallel to the flow direction of the polypropylene in the
mould of the injection moulding machine, and (b) SEM micrograph of a fracture of the talc filled PP with talc plates aligned parallel to the flow direction in the
mould.
428 B. Weidenfeller et al. / Composites: Part A 35 (2004) 423429

If a material with a high thermal conductivity but


considerable low density is requested, talc filled PP will be a
good candidate. Similarly to the reduced cooling time for
magnetite filled PP, the talc filled PPwhich has the highest
thermal diffusivity of the examined sampleswill have the
lowest cooling time and therefore enhance the productivity
in the moulding process.

6. Conclusions

Thermal properties of injection-moulded polypropylene


samples with various fillers and filler fractions were
prepared and investigated. As filler materials copper, talc,
magnetite, strontium ferrite, barite and glass fibres were
selected to vary over a broad range of different filler
properties, ranging from 2.58 (glass fibres) to 8.94 g/cm3
Fig. 6. Temperature dependence of the specific heat capacity of PP filled (copper) in density, and from nearly 1.5 (glass fibres)
with 45 vol% magnetite and barite. The symbols are measured values, the to approximately 400 W/(m K) (copper) in thermal
lines represent linear fits. conductivity.
Surprisingly, the highest thermal conductivity of a
of magnetite and barite increase in the measured temp- composite was achieved with talc which has a moderate
erature range from cp Fe3 O4 ; 300 K 1:20 J/(g K) and thermal conductivity (ltalc < 10:6 W/(m K)) compared to
cp BaSO4 ; 300 K 0:77 J/(g K) up to cp Fe3 O4 ; 392 K the high thermal conductivity of copper ( lcopper <
1:87 J/(g K) and cp BaSO4 ; 395 K 1:29 J/(g K). In the 400 W/(m K)). The thermal conductivity of the talc filled
same temperature range the thermal diffusivity (Fig. 7) polymer was increased from 0.27 up to 2.50 W/(m K) with
decreases from aFe3 O4 ; 300 K 0:443 mm2 /s and 30 vol% talc filler. This behaviour is a consequence of the
microstructure of talc filled PP showing a high intercon-
aBaSO4 ; 300K 0:353 mm2/s down to aFe3 O4 ; 392 K
nectivity of the talc crystals. On the other hand, the thermal
0:322 mm2/s and aBaSO4 ; 395 K 0:244 mm2/s.
conductivity of copper filled polypropylene is in spite of
Analogous to the differences in the measured values
the very high conductivity of pure copper lower than talc
between magnetite and barite filled polypropylene it was
filled PP composites due to a poor interconnectivity of the
proven in our injection moulding experiments that the
Cu-particles.
cooling time in injection moulding process for magnetite
The talc platelets and the glass fibres were aligned in the
filled PP is shorter than for barite filled PP, this is a direct
test samples by the injection moulding process. This leads to
consequence of the higher thermal diffusivity of the
a thermal conductivity close to the HS upper boundary
magnetite filled composite.
and simulates a talc matrix and a glass matrix, respectively,
which surrounds polypropylene particles. Also the specific
heat capacities and thermal diffusivities of talc filled
polypropylene are much higher than that of copper filled
polypropylene.
A comparison of the two competing materials, barite and
magnetite, for increasing the weight of polymers, shows for
magnetite filled PP a higher thermal conductivity than for
barite filled PP. This results not only from small differences
in specific density but also from lower specific heat capacity
and thermal diffusivity of barite filled PP compared with
magnetite filled PP. The significant lower thermal conduc-
tivity of the barite filled PP compared to the magnetite filled
one can be directly related to the substantially lower thermal
conductivity of barite < 2 W/(m K) compared to the higher
conductivity of magnetite < 5 W/(m K). Furthermore, the
higher interconnectivity of magnetite filled PP increases
the thermal conductivity of the composite compared to the
Fig. 7. Temperature dependence of the thermal diffusivity of PP filled with
barite filled one.
45 vol% magnetite and barite. The symbols are measured values, the lines In the temperature range from 300 to 385 K the specific
represent linear fits. heat capacity of the composites is increased for PP with
B. Weidenfeller et al. / Composites: Part A 35 (2004) 423429 429

magnetite (from 1.20 up to 1.87 J/(g K)) and for PP with [4] King JA, Tucker KW, Vogt BD, Weber EH, Quan C. Electrically and
barite fillers (from 0.77 up to 1.29 J/(g K)). In the same thermally conductive nylon 6.6. Polym Compos 1999;20(5):643 54.
[5] Yu S, Hing P, Hu X. Thermal conductivity of polystyrene-aluminum
temperature region the thermal diffusivity decreases. nitride composite. Composites: Part A 2002;33:28992.
This study indicates that for injection moulding pro- [6] Weidenfeller B, Hofer M, Schilling F. Thermal and electrical
cesses of heavy plastic composites magnetite might be a properties of magnetite filled polymers. Composites: Part A 2002;
good choice, as it allows higher production cycles than the 33:1041 53.
barite filled composites. To increase the thermal conduc- [7] Hashin Z, Shtrikman A. A variational approach to the theory of the
effective magnetic permeability of multiphase material. J Appl Phys
tivity significantly without increasing the density very 1962;33:3125 31.
much, talc filled composites seem to be a good candidate [8] Clauser C, Huenges E. Thermal conductivity of rocks and minerals.
and can help to save cooling time during the injection In: Ahrens TJ, editor. Rock physics and phase relations. A handbook
moulding process. of physical constants. American Geophysical Union Reference; 1995.
[9] Slack GA. Nonmetallic crystals with high thermal conductivity. J Phys
Chem Solids 1973;34:32135.
[10] Klemens PG, Neuer G, Sundquist B, Uher C, White GK, In: Madelung
Acknowledgements O, White GK, editors. Numerical data and functional relationships in
science and technology, Londolt-Bornstein New Series, Group III;
The authors would like to thank P. Duifhuis and crystal and solid state physic metals: electronic transport phenomena,
R. Mangnus, Minelco B.V. (Rotterdam, The Netherlands) subvolume c: thermal and conductivity of pure metals and alloys Vol.
for the support and providing materials. 15. Berlin: Springer; 1991.
[11] Gardon R. Thermal conductivity at low and moderated temperatures.
In: Blazek A, editor. Review of thermal conductivity data in glass.
International Commission on Glass; 1983.
References [12] Schilling FR, Partzsch GM. Quantifying partial melt portion in the
crust beneath the Central Andes and the Tibetan Plateau. Phys Chem
[1] Wong CP, Bollampally RS. Thermally conductivity, elastic modulus, Earth (A) 2001;26:239 46.
and coefficient of thermal expansion of polymer composites filled [13] German standard DIN 53479, Testing of plastics and elastomers,
with ceramic particles for electronic packaging. J Appl Polym Sci determination of density. July 1976.
1999;74:3396 403. [14] Parker WJ, Jenkins RJ, Butler CP, Abbott GL. Flash method of
[2] Lu X, Xu G. Thermally conductive polymer composites for electronic determining thermal diffusivity, heat capacity, and thermal conduc-
packaging. J Appl Polym Sci 1997;65:27338. tivity. J Appl Phys 1961;32:1679 84.
[3] Xu Y, Chung DDL, Mroz C. Thermally conducting aluminum nitride [15] Schilling FR. A transient technique to measure thermal diffusivity at
polymer-matrix composites. Composites: Part A 2001;32:1749 57. elevated temperatures. Eur J Mineral 1999;11:1115 24.

Вам также может понравиться