micromachines

Article
Research of a Novel Ultra-High Pressure Sensor with
High-Temperature Resistance †
Guo-Dong Zhang, Yu-Long Zhao *, Yun Zhao, Xin-Chen Wang and Xue-Yong Wei ID

State Key Laboratory for Manufacturing Systems Engineering, Xi’an Jiaotong University, Xi’an 710049, China;
zhangguodong@stu.xjtu.edu.cn (G.-D.Z.); zhaoyun@stu.xjtu.edu.cn (Y.Z.);
wangxinchen@stu.xjtu.edu.cn (X.-C.W.); seanwei@mail.xjtu.edu.cn (X.-Y.W.)
* Correspondence: zhaoyulong@mail.xjtu.edu.cn; Tel.: +86-029-8339-5073
† This paper is an extended version of our paper accepted in the 13th International Symposium on
Measurement Technology and Intelligent Instruments, Xi’an, China, 22–25 September 2017.

Received: 20 November 2017; Accepted: 21 December 2017; Published: 25 December 2017

Abstract: Ultra-high pressure measurement has significant applications in various fields such as high
pressure synthesis of new materials and ultra-high pressure vessel monitoring. This paper proposes
a novel ultra-high pressure sensor combining a truncated-cone structure and a silicon-on-insulator
(SOI) piezoresistive element for measuring the pressure up to 1.6 GPa. The truncated-cone structure
attenuates the measured pressure to a level that can be detected by the SOI piezoresistive element.
Four piezoresistors of the SOI piezoresistive element are placed along specific crystal orientation
and configured as a Wheatstone bridge to obtain voltage signals. The sensor has an advantage of
high-temperature resistance, in that the structure of the piezoresistive element can avoid the leakage
current at high temperature and the truncated-cone structure separates the piezoresistive element
from the heat environment. Furthermore, the upper surface diameter of the truncated-cone structure
is designed to be 2 mm for the application of small scale. The results of static calibration show that
the sensor exhibits a good performance in hysteresis and repeatability. The temperature experiment
indicates that the sensor can work steadily at high temperature. This study would provide a better
insight to the research of ultra-high pressure sensors with larger range and smaller size.

Keywords: ultra-high pressure sensor; silicon-on-insulator (SOI) piezoresistive element; small size;
high-temperature resistance; microelectromechanical systems (MEMS) technology

1. Introduction
Ultra-high pressure sensors are of great importance in many special civil and military applications,
such as high pressure synthesis of new materials, pressure calibration of diamond anvil cell, shock wave
propagation, explosive detonation, and high-speed impact. To date, numerous ultra-high pressure
sensors based on various sensing principles—including piezoresistive, piezoelectric, capacitive, and
optical fiber—have been developed [1]. The piezoelectric pressure sensor must be combined with
a charge amplifier because of its high output impedance and it can only measure the dynamic
pressure [2]. The capacitive sensor possesses large nonlinearity, for the restriction of sense principle [3].
Optical-fiber-grating sensor’s structure is complicated, which makes output signal conditioning
complex and sensor packaging difficult [4]. The piezoresistive sensor has been mainly used to measure
the ultra-high pressure because of low output impedance, large output signal, excellent accuracy
and good dynamic property. In literature, there are many papers proposing numerous ultra-high
pressure piezoresistive sensors with different structures. The design concepts of those structures
can be divided into two main broad categories such as (a) the measured pressure is directly applied
to the sensitive element; and (b) the sensitive element is separated from the measured pressure by

Micromachines 2018, 9, 5; doi:10.3390/mi9010005 www.mdpi.com/journal/micromachines

Micromachines 2018, 9, 5 2 of 13

elastic elements. As for the first category, piezoresistive materials including carbon, ytterbium, and
manganin are made into sensitive films that can subject the ultra-high pressure uniformly without
any drastic stress concentration. Nevertheless, the piezoresistive curve of the carbon pressure gage is
nonlinear [5]. Ytterbium is a kind of material in which phase transformation will take place at high
pressure. This limits the measuring range of the ytterbium sensor [6]. The manganin gage offers a low
piezoresistive coefficient and its accuracy is affected by the transverse tensile strain effect [7]. In the
second category, many elastic element structures with different materials are designed. Among those
sensors, sensitive elements and elastic elements of some sensors are integrated together which allows
them to perform with high sensitivity and precision, such as the silicon piezoresistive sensor [8–12].
However, since the silicon structure is weak in strength, their measuring range is often no more than
300 MPa. Therefore, using high strength material as an elastic element can effectively increase the
measuring range of the sensor. Fu [13] considered a piezoresistive pressure transduce which used
high intensity alloy film as initial sensing element to endure the measured pressure and silicon oil to
translate the pressure to piezoresistive sensing element. Its measuring range is 0–400 MPa. Zhao [14]
explored an ultra-high pressure sensor based on autofrettaged cylinder structure made of the spring
steel of 60Si2CrVA and silicon flat chip with an inverted cup structure. Its measuring range is 0–1 GPa.
Zhao [15] studied an ultra-high pressure sensor combining a cylindrical elastic body made of the
stainless steel of 17-4PH and a micro silicon-on-insulator (SOI) solid piezoresistive chip. Its measuring
range is 0–2 GPa. All these sensors mentioned above lay a foundation for the research of ultra-high
pressure sensors with larger range and better accuracy.
In general, ultra-high pressure environments are often accompanied by high temperature, such as
explosive detonation [16], high pressure synthesis of new materials [17], and ultra-high pressure
vessels [18]. High temperatures will affect the measurement accuracy of the sensor. However, many
researchers focus more on the measuring range than on high-temperature resistance of the sensor.
This research proposes an ultra-high pressure sensor combining a truncated-cone structure and an SOI
piezoresistive element. It exhibits a large measuring range up to 1.6 GPa, high-temperature resistance,
small size, and good static performance. The sensor reported here could be beneficial to the research of
increasing the measuring range and further improving the performance of high-temperature resistance
of ultra-high pressure sensors.

2. Materials and Methods

2.1. Structure of the Sensor

As shown in Figure 1a, the ultra-high pressure sensor is composed of an elastic body, an SOI
piezoresistive element, and two patch boards. The elastic body has two sections: one is called
truncated-cone section and the other is cylinder section. The truncated-cone section was designed
to reduce the ultra-high pressure to a value that can be detected by the SOI piezoresistive element.
The SOI piezoresistive element and patch boards were attached to the same side of the cylinder section.
The piezoresistive element was fabricated on an SOI wafer whose buried SiO2 layer can avoid
the leakage current at high temperature [19]. Four equivalent piezoresistors with sheet resistances of
10 Ω/ were located in the (100) crystal plane which has a small crystal alignment deviation. In order
to obtain the maximum and equivalent piezoresistance coefficent, we let piezoresistors be parallel
to specific orientation by microelectromechanical systems (MEMS) technology, as Figure 1b shows.
The beam lead method of the pad stacks was used to solve the problem of connection between the
inside and the outside of the piezoresistive element at high temperature. Titanium was chosen to form
ohmic contact with p-type silicon because of its low specific contact resistivity (SCR). The theoretical
contact resistance of a single piezoresistor is about 8.89 mΩ, which is beneficial to the work stability of
the piezoresistive element. Aurum was chosen as the third layer of the beam lead to contact with the
external leads. Platinum was used as second layer of the beam lead to prevent the formation of Ti-Au
intermetallic compound. The patch boards were connected with the pad stacks through gold wires.
Micromachines 2017,
2018, 8,
9, 55 3 of 13
Micromachines 2017, 8, 5 3 of 13
wires. A full Wheatstone bridge widely used in MEMS sensors was formed in an external circuit for
A full
signal Wheatstone
wires. A full
read bridge
outWheatstone
yielded bywidely
bridge used inused
widely MEMS
the piezoresistors sensors
in MEMS
[20,21]. was formed
sensors in anin
was formed external circuit
an external for signal
circuit for
read out yielded by the piezoresistors [20,21].
signal read out yielded by the piezoresistors [20,21].

Figure
Figure1. 1.Schematic
Schematicofofthe
thesensor:
sensor: (a)
(a) structure of the
structure of the sensor;
sensor;(b)
(b)layout
layoutofofthe
thesilicon-on-insulator
silicon-on-insulator
Figure 1. Schematic of the sensor: (a) structure of the sensor; (b) layout of the silicon-on-insulator (SOI)
(SOI) piezoresistive
(SOI) piezoresistiveelement.
element.
piezoresistive element.

2.2.2.2. Working
Working Principleofofthe
Principle theSensor
Sensor
2.2. Working Principle of the Sensor
The
The ultra-highpressure
ultra-high pressuresensor
sensorwas
was developed
developed basedbasedononthetheequivalent
equivalenttransfer
transfer ofof
a force
a forcein in
thethe
The
elastic ultra-high
limit rangeandpressure
andthe sensor was
thepiezoresistive developed
piezoresistive effect
effect of based on the equivalent transfer of a force in the
elastic limit range of silicon.
silicon.The
Thework
workprinciple
principleisisdescribed
described as as
follows.
follows.
elastic limit range and the piezoresistive effect of silicon. The work principle is described as follows.
AsAsshown
shownininFigure
Figure2,2, when
when an an ultra-high
ultra-high pressure
pressure isis applied
appliedon onthe
theupper
uppersurface
surface of of
thethe
As shown insection,
truncated-cone Figure the 2, when an section
cylinder ultra-highwill pressure
produce is
a appliedAccording
strain. on the upperto thesurface of the
equivalent
truncated-cone section, the cylinder section will produce a strain. According to the equivalent
truncated-cone
transfer section, thethe
cylinder
elasticsection will produce a strain. According to the equivalent transfer
transfer ofof a force
a force within
within the elastic limit range,
limit range, we
we have
havethetheequation
equation
of a force within the elastic limit range, we have the equation
εε1 = P ⋅⋅ AA00
P
(1)
= P·⋅AE0
ε 11 = A
AA111⋅·E1
(1)
(1)
E11
where ε1 is the strain of the place where the sensitive element was attached, P is an ultra-high pressure
where εε11 is
where the strain of the
the place where the
the sensitive element was attached, P is
is an
P of an ultra-high pressure
appliedisonthe strain
the ofsurface
upper place where
of the sensitivesection,
truncated-cone elementA0was
and attached,
A1 are area ultra-high
upper surface pressure
of the
applied
applied on the upper
on the upper
truncated-cone surface
surface
section of the
of the
and cross truncated-cone
truncated-cone
sectional section,
section,
area of the A 0
place whereand
A0 and A are area
theAsensitive of
1 are areaelement
1 upper
of upper surface
surface
was of
of the
attached, the
truncated-cone
truncated-cone section
respectively, Esection and
and
1 denotes
cross sectional
crossmodulus
elastic area
sectionalofarea of the place
of the place
the elastic where the sensitive element was attached,
body. where the sensitive element was attached,
respectively,
respectively, EE11 denotes
denotes elastic
elastic modulus
modulus of of the
the elastic
elastic body.
body.

Figure 2. Work principle of the sensor.

Figure
Figure 2. Work
Work principle of the sensor.
Micromachines 2018, 9, 5 4 of 13

As for piezoresistors of the piezoresistive element, the relative change ratio of resistivity along the
resistor stripe can be written as [22]

∆ρ
 
= τl σl + τt σt (2)
ρ0 l

where τ l and σl are longitudinal piezoresistive coefficient and longitudinal stress along the direction
parallel to the resistor stripe, respectively, τ t and σt denote transverse piezoresistive coefficient and
transverse stress along the direction perpendicular to the resistor stripe, respectively.
In this research, equivalent piezoresistors were arranged in (100) crystal plane and along [110]
and [110] crystallographic orientations. According to Equation (2), the relative change ratio of R1 and
R4 can be expressed as
∆R1 ∆R4 1
= = − π44 σ2 (3)
R1 R4 2
where the minus sign indicates that resistances of R1 and R4 decrease when there is only a principal
stress along the [110] crystallographic orientation, π 44 is shear piezoresistive coefficient of p-type
silicon, σ2 denotes the principal stress along the [110] crystallographic orientation. Similarly, we can
obtain the relative change ratio of R2 and R3

∆R2 ∆R3 1
= = + π44 σ2 (4)
R2 R3 2

where the plus sign indicates that resistances of R2 and R3 increase when there is only a principal stress
along the [110] crystallographic orientation.
Based on the principle of Wheatstone bridge, output voltage of the bridge can be written as

∆R
V= ·V0 (5)
R
where V 0 is the voltage of constant voltage source. On substituting Equation (4) into Equation (5) and
solving for σ2 ,
2V
σ2 = (6)
V0 ·π44
Then we have
2V
ε2 = (7)
V0 ·π44 · E2
where ε2 is the principal strain along the [110] crystallographic orientation. E2 denotes the elastic
modulus of silicon.
In this study, the piezoresistive element was attached to the elastic body with M-Bond 610 and
[110] crystallographic orientation of the piezoresistive element is parallel to axial direction of the elastic
body. Suppose that
ε1 = ε2 (8)

Under above situation. On substituting Equations (1) and (7) into Equation (8) and solving for P,

2 A1 · E1 ·V
P= · (9)
π44 A0 · E2 ·V0

It can be noted from the Equation (9) that the measured pressure is proportional to the output
voltage of Wheatstone bridge. That is to say, as long as we know the output voltage of Wheatstone
bridge, we can work out the measured pressure. However, in practical applications, we calculate the
measured pressure through the calibration curve of the sensor.
Micromachines 2018, 9, 5 5 of 13

2.3. Design Details of the Sensor

As we know, silicon has a strain limit of 0.003 [19]. Using this restrictive condition and the
hypothesis shown in Equation (8), measuring range of the sensor is limited by following inequation
 2
r1 εm
P≤ · E1 · (10)
r0 k

where r0 and r1 are radii of upper and lower surface of the truncated-cone section, respectively, εm is
the strain limit of silicon, k is safety factor of the strain limit. It can be observed that the measuring
range of the sensor depends on the radius ratio of lower and upper surface of the truncated-cone
section and the elastic modulus of the elastic body. Moreover, the sensor must work in the elastic range.
Therefore, the following inequation needs to be met

P ≤ σs (11)

where σs is yield strength of the elastic body. Only when Equations (10) and (11) are satisfied at the
same time can the sensor work properly. In other words, the measuring range of the sensor depends
on the radius ratio of lower and upper surface of the truncated-cone section and elastic modulus and
yield strength of the elastic body.
Tungsten alloy was used to produce the elastic body. As listed in Table 1 [23], it offers high
Young’s modulus and yield strength, which can help to increase the measuring range of the sensor.
In addition, tungsten alloy also has a high melting point and low coefficient of thermal expansion
(CTE), which can help the sensor work at high temperature.

Table 1. Material parameters of tungsten alloy.

Young’s Yield Strength Melting Point Coefficient of Thermal Expansion

Modulus (GPa) (GPa) (◦ C) (1 × 10–6 /◦ C)
345 1.6 3410 4.5

Many factors should be considered to design dimensions of the sensor. According to our processing
capacity and practical applications, the minimum value of r0 can be set to 1 mm. In order to take full
advantage of high yield strength of the elastic body, we consider that
 2
r1 εm
σs ≤ · E1 · (12)
r0 k

Given that k = 5, we obtain

r1 ≥ 2.78 mm (13)

In order to obtain a small size sensor, we let r1 be equal to 3 mm. The rest sizes of the sensor are
shown in Figure 3. Under the above conditions, the sensor proposed in this study has a measuring
range of up to 1.6 GPa.
Micromachines 2018, 9, 5 6 of 13
Micromachines 2017, 8, 5 6 of 13

Micromachines 2017, 8, 5 6 of 13

Figure
Figure 3. 3. Threeviews
Three viewsofofthe
theelastic
elasticbody:
body: (a)
(a) top view;
view; (b)
(b)front
frontview;
view;(c)
(c)left
leftview.
view.
Figure 3. Three views of the elastic body: (a) top view; (b) front view; (c) left view.
2.4.2.4. Analysis
Analysis of High-TemperatureResistance
of High-Temperature Resistance
High-temperature
2.4. Analysis resistance
of High-Temperature of the sensor can be attributed to the following two aspects. First,
Resistance
High-temperature resistance of the sensor can be attributed to the following two aspects.
the piezoresistive element was fabricated on an SOI wafer by MEMS technology. Second, high
First, the piezoresistive resistance
High-temperature element was of thefabricated
sensor can onbean SOI wafer
attributed byfollowing
to the MEMS technology.
two aspects. Second,
First,
temperature of the measured environment was dissipated mostly by the elastic body.
thetemperature
high piezoresistive of element
the measuredwas fabricated
environment on anwasSOI wafer bymostly
dissipated MEMSbytechnology. Second, high
the elastic body.
The SOI technology introduces a buried oxide layer between the top silicon and the silicon
temperature of the measured
The SOI technology environment
introduces a buriedwas oxide
dissipated
layermostly
betweenby the
theelastic body. and the silicon
top silicon
substrate. The piezoresistors of the piezoresistive element is separated from the silicon substrate by
The
substrate. SOI
The technology
piezoresistors introduces a buried oxide layer between the top silicon and the silicon by
the oxide layer to avoid theof the piezoresistive
leakage current between element is separated from
the piezoresistors and thethesilicon
silicon substrate
substrate at
thesubstrate.
high
Thetopiezoresistors
oxidetemperature,
layer avoid of thecurrent
the leakage
thereby improving
piezoresistive
betweenelement
high-temperature
is separated from
the piezoresistors
resistance ofand the the
the silicon
silicon
piezoresistive
substrate
substrate
element.atby
high
In
the oxide layer
temperature, to avoid the leakage
thereby current between the piezoresistors and the silicon substrate at
addition, the high improving
temperaturehigh-temperature
experiment showed resistance
that theofresistance
the piezoresistive
values ofelement. In addition,
the piezoresistors
high
thecan
high temperature,
temperature thereby improving
experiment high-temperature resistance of the piezoresistive element. In
keep relatively stable for two showed
hours at 250 that°Cthe resistance values of the piezoresistors can keep
[19].
addition, the high temperature experiment ◦ showed that the resistance values of the piezoresistors
relativelyThe stable
heat for two hours
conduction at 250
of the elasticC [19].
body was analyzed by ANSYS 15 (Ansys, Inc., Canonsburg,
can keep relatively stable for two hours at 250 °C [19].
The
PA, heatAconduction
USA). thermal load of the elastic
of 800 °C was body was analyzed
applied on the upperby ANSYS
surface 15 of (Ansys,
the elastic Inc., Canonsburg,
body, and the
The heat conduction of the elastic body was analyzed by ANSYS 15 (Ansys, Inc., Canonsburg,
PA,steady-state
USA). A thermaltemperature of 800 ◦ C was
load distribution of theapplied
elastic on
bodytheisupper surface
presented of the4.elastic
in Figure body,the
Especially, and
PA, USA). A thermal load of 800 °C was applied on the upper surface of the elastic body, and the
thetemperature
steady-stateoftemperature
the place where the piezoresistive
distribution of the element
elastic bodywas attached is approximately
is presented in Figure 4. 250Especially,
°C. This
steady-state temperature distribution of the elastic body is presented in Figure 4. Especially, the◦
theresult indicatesofthat
temperature the working
the place where thetemperature of theelement
piezoresistive sensor iswasas high
attachedas about 800 °C (the upper
is approximately 250 C.
temperature of the place where the piezoresistive element was attached is approximately 250 °C. This
Thissurface
result of the elastic
indicates body
that the isworking
exposedtemperature
to the measured of theenvironment).
sensor is as high as about 800 ◦ C (the upper
result indicates that the working temperature of the sensor is as high as about 800 °C (the upper
surface
surfaceof of
thetheelastic
elasticbody
bodyisisexposed
exposedto tothe
the measured environment).
measured environment).

Figure 4. Steady-state thermal analysis of elastic body.

Figure 4. Steady-state thermal analysis of elastic body.

Figure 4. Steady-state thermal analysis of elastic body.
Micromachines 2018, 9, 5 7 of 13

Micromachines 2017, 8, 5 7 of 13
Micromachines 2017, 8, 5 7 of 13
The above design considerations can not only increase the working temperature of the sensor,
but alsoThe above
reduce thedesign
impact considerations
of temperature canon notthe
only increase
accuracy ofthe
theworking
sensor. temperature of the sensor,
The above design considerations can not only increase the working temperature of the sensor,
but also reduce the impact of temperature on the accuracy of the sensor.
but also reduce the impact of temperature on the accuracy of the sensor.
2.5. Fabrication of the Sensor
2.5. Fabrication of the Sensor
2.5.The
Fabrication of the Sensor element was fabricated by MEMS technology. Its fabrication process is
SOI piezoresistive
illustratedThe inSOI piezoresistive
Figure 5: (1) clean element
the SOI was fabricated by MEMS technology. Its fabrication process is
The SOI piezoresistive element waswafer withby
fabricated dilute
MEMS HFtechnology.
solution toIts remove the native
fabrication process oxide;
is
illustrated
(2)illustrated
increase the in Figure
thickness5: (1) clean the SOI wafer with dilute HF solution to remove the native oxide; (2)
in Figure 5: (1)of top the
clean silicon layer by
SOI wafer withlow pressure
dilute chemical
HF solution vapor the
to remove deposition
native oxide; (LPCVD)
(2)
increase
to increase
ensure the the sensitivity
thickness ofoftop silicon layer element;
piezoresistive by low pressure
(3) dope chemical
boron vaporinto
ions deposition
top (LPCVD)
silicon layer to
and
the thickness of top silicon layer by low pressure chemical vapor deposition (LPCVD) to
makeensure the
annealingsensitivity
treatment of piezoresistive element;
for formingelement; (3)
p-type silicon dope boron ions into top silicon layer and make
ensure the sensitivity of piezoresistive (3) dopewhichboron offers high
ions into toppiezoresistive
silicon layer and effect
makeand
goodannealing treatment
temperature for forming
resistance; (4) form p-type silicon
SiO2silicon
layer by which offers high piezoresistive
thermal effect and good
annealing treatment for forming p-type which offersoxidation for improving
high piezoresistive effecttheand
stability
good of
temperature resistance; (4) form SiO2 layer by thermal oxidation for improving the stability of
measurement
temperature circuit layer(4)
resistance; and produce
form Si3 N4by
SiO2 layer layer as a stress
thermal oxidation matching layer by the
for improving LPCVD; (5) of
stability etch
measurement circuit layer and produce Si3N4 layer as a stress matching layer by LPCVD; (5) etch
piezoresistors
measurement bycircuit
reactivelayerionand
etchproduce
(RIE); (6) Si3etch contact
N4 layer as aholes
stressfor exposing
matching p-type
layer silicon; (7)
by LPCVD; (5)sputter
etch
piezoresistors by reactive ion etch (RIE); (6) etch contact holes for exposing p-type silicon; (7) sputter
Ti, piezoresistors by reactive ion
Pt, and Au successively; (8) etch (RIE);
corrode the(6)extra
etch contact
metal layerholesforforforming
exposingelectrodes;
p-type silicon; (7) sputter
(9) keep 30 min
Ti, Pt, and Au successively; (8) corrode the extra metal layer for forming electrodes; (9) keep 30 min
◦ C Au
Ti, Pt,
under 550and successively;
in vacuum (8) corrode
for good the extra
ohmic contact metal layer
between Ti andforp-type
forming electrodes;
silicon. (9)a keep
Finally, single30piece
min of
under 550 °C in vacuum for good ohmic contact between Ti and p-type silicon. Finally, a single piece
under 550
piezoresistive °C in vacuum for good ohmic contact between Ti and p-type
element was obtained after dicing. Figure 6 shows a SEM photograph of the fabricatedsilicon. Finally, a single piece
of piezoresistive element was obtained after dicing. Figure 6 shows a SEM photograph of the
ofpiezoresistive
piezoresistive element
element with was a obtained after dicing. Figure 3 6 showsare a SEM photograph of the
SOI fabricated SOI piezoresistive size of
element with 1.8 a×size × 1.8
1.6 of 0.2 ×mm1.6 .× There two groups
0.2 mm33. There are twoofgroups
measuringof
fabricated
circuits SOI piezoresistive
with resistances element
of resistances
1 KΩ and 2ofKΩ, with a size of
respectively, 1.8 × 1.6 × 0.2 mm . There are two groups Inof
measuring circuits with 1 KΩ and 2 KΩ,on every SOI piezoresistive
respectively, on every SOI element.
piezoresistive this
measuring
study, the circuits
resistors of with
2 KΩ resistances
were used of
to 1form
KΩ aand 2 KΩ, respectively,
Wheatstone bridge. on every SOI piezoresistive
element. In this study, the resistors of 2 KΩ were used to form a Wheatstone bridge.
element. In this study, the resistors of 2 KΩ were used to form a Wheatstone bridge.

Figure 5. Fabricationprocess
process of the
the SOI piezoresistive element.
Figure5.5.Fabrication
Figure Fabrication process of SOI piezoresistive
of the SOI piezoresistiveelement.
element.

Figure 6. SEM photograph of the SOI piezoresistive element.

Figure 6. SEM photograph of the SOI piezoresistive element.
Figure 6. SEM photograph of the SOI piezoresistive element.
Micromachines 2018, 9, 5 8 of 13
Micromachines 2017, 8, 5 8 of 13
Micromachines 2017, 8, 5 8 of 13

fabricated with 93
The elastic body was fabricated 93 tungsten
tungsten alloy
alloy by
by mechanical
mechanical processing.
processing. Its outline
The elastic body was fabricated with 93 tungsten alloy by mechanical processing. Its outline
dimensions are presented in Figure 3. The roughness of the cutting face on both sides of the elastic
dimensions
body are presented
was required to be 3.2inµm
μmFigure 3. Theattaching
for better
better roughness theofpiezoresistive
the cutting face
piezoresistive on both sides of the elastic
element.
for attaching the element.
body was required to be 3.2 μm for better attaching the piezoresistive element. _
The
The SOI
SOI piezoresistive element was
piezoresistive element was attached
attached on on the
the cutting
cutting face
face with
with M-Bond
M-Bond 610 610 and
and its
its [110]
[11_ 0]
The SOI piezoresistive
crystallographic
crystallographic orientationelement
orientation is was
is parallel
parallel to attached
to axial on theof
axial direction
direction cutting
of face with
the elastic
the elastic body.M-Bond
body. Then,
Then, two 610patch
two and its
patch [110]
boards
boards
crystallographic
were orientation
symmetrically arranged
were symmetrically is parallel
arrangedatatboth to
bothsides axial
sidesof ofdirection
thethe of the elastic
piezoresistive
piezoresistive body.
element.
element. Then,
The The two patch
measuring
measuring boards
circuit
circuit with
were
with symmetrically
2 KΩ on the arranged
piezoresistive at both sides
element wasof the piezoresistive
connected with element.
the proximal The measuring
side
2 KΩ on the piezoresistive element was connected with the proximal side of the patch boards through of the circuit
patch with
boards
2 KΩ on
through the
gold
gold wires. piezoresistive element
wires. A Wheatstone
A Wheatstone was
bridge wasbridge connected
formed was with the
formed wires
by leading proximal
by leading side
wires
from the of the
from
other patch
the
side boards
of other through
sideboards.
the patch of the
gold
patch wires. A
boards. Wheatstone
The SOI bridge
piezoresistivewas formed
element, by leading
gold wires, wires
and from the
welding other
points side
The SOI piezoresistive element, gold wires, and welding points were coated with high temperature-wereof the patch
coated withboards.
high
The SOI piezoresistive
temperature-resistance element,
pastern. gold
Figurewires,
7 and welding
illustrates a points
photograph
resistance pastern. Figure 7 illustrates a photograph of the fabricated sensor. were
of thecoated with
fabricated high
sensor.temperature-
resistance pastern. Figure 7 illustrates a photograph of the fabricated sensor.

Figure 7. Photograph of the fabricated sensor.

Figure 7.
Figure Photograph of
7. Photograph of the
the fabricated
fabricated sensor.
sensor.

2.6. Static Calibration

2.6. Static Calibration
2.6. Static Calibration
Static calibration test was performed to determine the static characteristics of the sensor. In this
Static
Static calibration
calibration test
test was performed to determine the
the static characteristics ofof the sensor. In
In this
study, measuring range of was performed
the sensor is 0–1.6to GPa.
determine static
Accordingly, thecharacteristics
force exerted on the
the sensor.
upper surfacethis
study,
study, measuring
measuring range of the sensor is 0–1.6 GPa. Accordingly, the force exerted on the upper surface
of the elastic bodyrange
should of the
be sensor
0–5 KN. is Static
0–1.6 GPa. Accordingly,
calibration test wastheoperated
force exerted
on anonelectro-mechanical
the upper surface
of
of the
the elastic
elastic body
body should
should be
be 0–5
0–5 KN.
KN. Static
Static calibration
calibration test
test was
was operated
operated on
on an
an electro-mechanical
electro-mechanical
universal testing machine which could apply standard force for a certain time. During the test, the
universal
universal testing machine
testing machine whichcould couldapplyapplystandard
standard forceforfor a certain time. During the test,
upper surface of the elasticwhich
body was applied from 0 N force a certain
to 5 KN with time.
an interval During
of 500 the test,each
N and the
the
upperupper surface
surface of of
the the elastic
elastic body
body waswas applied
applied from
from 0 0
N N
toto
5 5
KNKN with
with anan interval
interval of
of 500
500 N
N and
and each
each
interval would be maintained for 30 s. The measuring circuit was excited by 5V DC and the output
interval would
intervalwere
would be
be maintained
maintained for 30
30 s. The measuring circuit was
was excited
excited by
by 5V
5V DC
DC and the
the output
signals recorded by highfor s. The
accuracy andmeasuring
resolutioncircuit
precision digital multimeter. and output
A diagram of
signals
signals were recorded
were recorded by high accuracy and resolution precision digital multimeter. A diagram of the
the experimental setup by high accuracy
is shown in Figureand 8. resolution precision digital multimeter. A diagram of
experimental
the experimental setup is shown
setup in Figure
is shown 8. 8.
in Figure

Figure 8. Experimental setup for static calibration.

Figure 8. Experimental setup for static calibration.
Figure 8. Experimental setup for static calibration.
2.7. Temperature Experiment
2.7. Temperature Experiment
2.7. Temperature Experiment
The temperature experiment is important to show the high-temperature resistance of the sensor.
The
Due The temperature
to the limitationsexperiment is important
of experimental to show
conditions, wethe high-temperature
have preliminarily resistance
tested the of the sensor.
temperature
temperature experiment is important to show the high-temperature resistance of the sensor.
Due to the limitations
characteristics of
of the sensor experimental
under no loadconditions, we
condition.we have
Accordingpreliminarily tested
to the steadytested the
state the temperature
thermal analysis
Due to the limitations of experimental conditions, have preliminarily temperature
characteristics
shown of 4,
in Figure thethe
sensor under
working no load condition.
temperature According toelement
of the piezoresistive the steady state 250
is about thermal analysis
°C when the
shown in Figure 4, the working temperature of the piezoresistive element is about 250 °C when
sensor is at the upper limit of its working temperature (it is needed to emphasize that only the upper the
sensor is at the upper limit of its working temperature (it is needed to emphasize that only the upper
Micromachines 2018, 9, 5 9 of 13

characteristics of the sensor under no load condition. According to the steady state thermal analysis
shown in Figure ◦
Micromachines 2017, 8,4,5 the working temperature of the piezoresistive element is about 250 C when the
9 of 13
sensor is at the upper limit of its working temperature (it is needed to emphasize that only the upper
surface of the elastic body is exposed to the measured environment when the sensor is in the working
state). Then,
Then, wewe tested
tested the
the temperature
temperature characteristics
characteristics of
of the
the sensor
sensor at ◦ C on the basis of our
at 200 °C
experimental conditions.
As shown in Figure 9, the sensor was placed in high temperature environmental chamber which
can provide standard temperature from 0 to ◦ C for a certain time. The measuring circuit was
to 200
200 °C for a certain time.
excited by 5V DC and the output signalssignals were
were recorded
recorded by
by aa digital
digital multimeter.
multimeter.

Figure 9. Experimental setup for temperature experiment.

Figure 9. Experimental setup for temperature experiment.

3. Results and Discussions

3. Results and Discussions
3.1. Static
3.1. Static Calibration
Calibration
The static
The static calibration
calibration test
test for
for the
the developed
developed sensor
sensor was
was performed
performed five
five times
times and
and the
the average
average
output data
output data were
were recorded
recorded inin the
the Table
Table 2.
2. As
As Figure
Figure 10a
10a shows,
shows, the
the theoretical
theoretical work
work line
line of
of the
the sensor
sensor
is obtained by the least square method with these data. It takes the
is obtained by the least square method with these data. It takes the form form

V ( mV ) = 8.933 + 7.955 P ( GPa ) (14)

V (mV) = 8.933 + 7.955P(GPa) (14)
where P is an ultra-high pressure applied on the upper surface of the truncated-cone section, and V
where P isvoltage
is output an ultra-high pressure
of the digital applied on the upper surface of the truncated-cone section, and V is
multimeter.
output voltage of the digital multimeter.
Table 2. Average output data.
Table 2. Average output data.
Force (N) Corresponding Pressure (GPa) Average Output Data (mV)
Force0 (N) Corresponding 0Pressure (GPa) Average Output
8.495 Data (mV)
500
0 0.16
0 10.125
8.495
500
1000 0.16
0.32 10.125
11.571
1000 0.32 11.571
1500 0.48 12.949
1500 0.48 12.949
2000
2000 0.64
0.64 14.234
14.234
2500
2500 0.8
0.8 15.511
15.511
3000
3000 0.96
0.96 16.712
16.712
3500
3500 1.12
1.12 17.917
17.917
4000 1.28 19.126
4000
4500
1.28
1.44
19.126
20.243
4500
5000 1.44
1.6 20.243
21.388
5000 1.6 21.388
In order to evaluate static characteristics of the sensor, four indicators are given based on the
In order to evaluate static characteristics of the sensor, four indicators are given based on the
theoretical work line.
theoretical work line.
Hysteresis error can be calculated as follows

Δymax
H= ×100% (15)
( ymax − ymin )
Micromachines 2018, 9, 5 10 of 13

Hysteresis error can be calculated as follows

∆ymax
H= × 100% (15)
Micromachines 2017, 8, 5 (ymax − ymin ) 10 of 13

where HH represents
represents hysteresis error,∆y
hysteresis error, Δymax
max is the sensor’s biggest output deviation between loading

and offloading process

process under
under aa series
series of
ofinput values,yymax
inputvalues, and yymin
max and min are maximum and minimum
minimum
output
output values of the sensor. Substituting the experimental data of Table 3 into Equationwe
values of the sensor. Substituting the experimental data of Table 3 into Equation (15), obtain
(15), we
H = 1.45%.
obtain The hysteresis
H = 1.45%. characteristic
The hysteresis curves curves
characteristic of the sensor are shown
of the sensor in Figure
are shown 10b. 10b.
in Figure

Figure 10.
Figure 10. Static
Static characteristics
characteristics of
of the
the sensor:
sensor: (a)
(a) theoretical
theoretical work
work line;
line; (b)
(b) hysteresis
hysteresis characteristics
characteristics
curve.
curve.

Table 3. Average
Table 3. Average output
output data.
data.

Pressure (GPa) Loading (mV) Unloading (mV)

Pressure (GPa) Loading (mV) Unloading (mV)
0 8.532 8.458
00.16 8.532
10.182 8.458
10.068
0.16 10.182 10.068
0.32
0.32
11.622
11.622
11.520
11.520
0.48
0.48 13.028
13.028 12.870
12.870
0.64
0.64 14.324
14.324 14.144
14.144
0.80
0.80 15.598
15.598 15.424
15.424
0.96 16.804 16.620
0.96 16.804 16.620
1.12 17.980 17.854
1.12
1.28 17.980
19.188 17.854
19.064
1.28
1.44 19.188
20.299 19.064
20.186
1.60
1.44 21.380
20.299 21.388
20.186
1.60 21.380 21.388
Repeatability can be obtained through Equations (16) and (17)
Repeatability can be obtained through Equations
v (16) and (17)
u n
u n y − y
2
u ∑ ij
( )
i 2
t j =1 y − y
Si = ij i (16)
j =1 n − 1
(16)
Si =
3Smax
n −1
R= × 100% (17)
3S− ymin
ymax
R= max
×100% (17)
max − ymin
As for a certain test cycle, assuming theynumber of measuring points is m, and for each measuring
point, the output value has been measured for n times. Thus, Si denotes standard deviation of each
As for a certain test cycle, assuming the number of measuring points is m, and for each
measuring point, the output value has been measured for n times. Thus, Si denotes standard deviation
of each measuring point, yij is the jth (j = 1 − n) measured value of the ith (i = 1 − m) measuring point,
_
yi is the average value of the measured values of ith measuring point. R stands for repeatability error,
Smax represents the biggest standard deviation among Si. On calculating R, R = 0.28%.
Micromachines 2018, 9, 5 11 of 13

measuring point, yij is the jth (j = 1 − n) measured value of the ith (i = 1 − m) measuring point, yi is
the average value of the measured values of ith measuring point. R stands for repeatability error, Smax
represents the biggest standard deviation among Si . On calculating R, R = 0.28%.
Linearity can be represented as

∆max
L= × 100% (18)
ymax − ymin

where ∆max is the biggest deviation value between the sensor’s theoretical work line and its practical
one, L denotes the linearity error. After the calculation, we have L = 3.44%.
The accuracy of the sensor is a comprehensive indicator that reflects the sensor’s static performance.
Combining the above-calculated values, we obtain
p
A= H 2 + R2 + L2 = 3.74% (19)

3.2. Temperature Experiment

Thermal zero drift test was conducted at 25 ◦ C and 200 ◦ C for one hour, respectively. The zero
output was read every 15 minutes. The output data were recorded in the Table 4.

Table 4. Zero output at 25 ◦ C and 200 ◦ C.

Time (min) Zero Output at 25 ◦ C (mV) Zero Output at 200 ◦ C (mV)

0 8.505 7.947
15 8.518 7.934
30 8.526 7.912
45 8.513 7.938
60 8.507 7.951

Thermal zero drift can be represented as

 
y ( T2 ) − y ( T1 )
0 0
γ= × 100% (20)
YFS ( T1 )( T2 − T1 )

where y0 (Ti ) (i = 1, 2) is zero average output in one hour at temperature Ti . YFS (T1 ) denotes full
scale output of the sensor at temperature T1 . Let T1 = 25 ◦ C and T2 = 200 ◦ C in Equation (20). It can
be observed from Table 2 that the full scale output of the sensor at temperature 25 ◦ C is 21.388 mV.
Then we have γ = 0.015%◦ C−1 .
Zero temperature coefficient reflects the change of zero outputs with the temperature. It can be
expressed as [17]
∆yom
α= × 100% (21)
YFS ·∆T
where ∆yom is maximum variation of zero outputs in a certain range of temperature, YFS is full scale
output of the sensor, ∆T denotes the range of temperature. Zero outputs of the sensor at different
temperatures are given in Table 5. On calculating α, α = 0.016% ◦ C−1 .
Zero time drift at a certain temperature can preliminarily characterize the working stability of the
sensor at that temperature. From Table 4, the zero time drift at 200 ◦ C is calculable as [17]

ymax − ymin
Dt = = 0.039 mV/h (22)
∆t
where Dt is zero time drift, ymax and ymin are maximum and minimum zero outputs in a certain time
range at 200 ◦ C, respectively, ∆t denotes the range of time.
Micromachines 2018, 9, 5 12 of 13

Table 5. Zero outputs at different temperatures.

Temperature (◦ C) Zero Output (mV)

25 8.513
50 8.532
75 8.545
100 8.372
125 8.121
150 7.968
175 7.942
200 7.936

4. Conclusions
This paper proposes a novel structure combining a truncated-cone section and an SOI piezoresistive
element for ultra-high pressure sensor with the measuring range of 0–1.6 GPa. The structure design,
work principle, and characteristic analysis were presented in detail. The analysis result shows that
the sensor can achieve the ultra-high pressure measurement at high temperature up to 800 ◦ C in
theory. By the calibration, we obtained the performance indexes of the sensor such as linearity
of 3.44% of full scale (FS), repeatability of 0.28% of FS, hysteresis of 1.45% of FS, and accuracy
of 3.74% of FS. These performance indexes can basically meet the general demands of ultra-high
pressure measurement. In addition, we conducted the zero temperature experiment which indicated
preliminarily that the sensor can work steadily at high temperature. This research has a positive
reference value to other researches of the ultra-high pressure sensor with higher range and smaller size.
In future work, we will continue to study high-temperature resistance of the sensor proposed in this
paper. Furthermore, we will design a new micromachined structure utilizing the principle of stress
concentration to be integrated with the piezoresistive element for measuring the ultra-high pressure
with high accuracy.

Acknowledgments: The work was supported by the National Science Fund for Distinguished Young Scholars
(No. 51325503), the National Natural Science Foundation of China (No. 51705408), and the Fundamental Research
Funds for the Central Universities (No. xjj2017017).
Author Contributions: G.-D.Z. and Y.-L.Z. conceived and designed the sensor; X.-Y.W. designed the experiments;
Y.Z. and X.-C.W. performed the experiments; G.-D.Z. fabricated the sensor, analyzed the data, and wrote the paper.
All authors have read and approved the final manuscript.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Fraden, J. Handbook of Modern Sensors, 4th ed.; Springer: New York, NY, USA; Heidelberg, Germany;
Dordrecht, The Netherlands; London, UK, 2010; pp. 375–398.
2. Morten, B.; de Cicco, G.; Prudenziati, M. Resonant pressure sensor based on piezoelectric properties of
ferroelectric thick films. Sens. Actuators A Phys. 1992, 31, 153–158. [CrossRef]
3. Mitrakos, V.; Macintyre, L.; Denison, F.; Hands, P.; Desmulliez, M. Design, manufacture and testing of
capacitive pressure sensors for low-pressure measurement ranges. Micromachines 2017, 8, 41. [CrossRef]
4. Bock, W.J.; Eftimov, T.; Molinar, G.F.; Wisniewski, R. Free active element bulk-modulus high-pressure
transducer based on fiber-optic displacement sensor. In Proceedings of the IEEE Instrumentation and
Measurement Technology Conference, Ottawa, ON, Canada, 19–21 May 1997.
5. Cunningham, B.; Vandersall, K.S.; Niles, A.M.; Greenwood, D.W.; Garcia, F.; Forbes, J.W.; Wilson, W.H.
Carbon resistor pressure gauge calibration at low stresses. In Proceedings of the 12th International
Conference of the American Physical Society Topical Group on Shock Compression of Condensed Matter,
Atlanta, GA, USA, 24–29 June 2001.
6. Grady, D.E.; Ginsberg, M.J. Piezoresistive effects in ytterbium stress transducers. J. Appl. Phys. 1977, 48,
2179–2181. [CrossRef]
Micromachines 2018, 9, 5 13 of 13

7. Du, X.S.; Yang, B.C.; Zhou, H.R. Piezoresistive response of thin film manganin gauges in the 50–100-GPa
range. Thin Solid Films 2002, 410, 167–170. [CrossRef]
8. Zhao, Y.L.; Zhao, L.B.; Jiang, Z.D. A novel high temperature pressure sensor on the basis of SOI layers.
Sens. Actuators A Phys. 2003, 108, 108–111. [CrossRef]
9. Choi, W.; Lee, J.; Yoo, Y.K.; Kang, S.; Kim, J.; Lee, J.H. Enhanced sensitivity of piezoelectric pressure sensor
with microstructured polydimethylsiloxane layer. Appl. Phys. Lett. 2014, 104, 7808–7826. [CrossRef]
10. Singh, R.; Ngo, L.L.; Seng, H.S.; Mok, F.N.C. A silicon piezoresistive pressure sensor. In Proceedings of the
IEEE International Workshop on Electronic Design, Test and Application, Christchurch, New Zealand, 29–31
January 2002.
11. Hamdana, G.; Bertke, M.; Doering, L.; Frank, T.; Brand, U.; Wasisto, H.S.; Peiner, E. Transferable micromachined
piezoresistive force sensor with integrated double-meander-spring system. J. Sens. Sens. Syst. 2017, 6,
121–133. [CrossRef]
12. Hamdana, G.; Wasisto, H.S.; Doering, L.; Yan, C.L.; Zhou, L.; Brand, U.; Peiner, E. Double-meander spring
silicon piezoresistive sensors as microforce calibration standards. Opt. Eng. 2016, 55, 091409. [CrossRef]
13. Fu, J.Q.; Fan, M.J.; Shan, Y.Y.; Qu, Y.X. 400Mpa F.S. Piezoresistive pressure transducer. J. Transducer Technol.
1995, 1, 13–16. [CrossRef]
14. Zhao, L.B.; Guo, X.; Meng, X.W.; Hebibul, R.; Zhao, Y.L.; Wang, J.Z.; Jiang, Z.D. An ultra-high pressure sensor
with cylinder structure. J. Mech. Sci. Technol. 2013, 27, 2383–2389. [CrossRef]
15. Zhao, Y.L.; Fang, X.D.; Jiang, Z.D.; Zhao, L.B. An ultra-high pressure sensor based on SOI piezoresistive
material. J. Mech. Sci. Technol. 2010, 24, 1655–1660. [CrossRef]
16. Cavanough, G.; Onederra, I. Development of pressure and temperature gauges to monitor performance of
commercial explosives. Inst. Mater. Miner. Min. 2013, 120, 74–79. [CrossRef]
17. Brazhkin, V.V. High-pressure synthesized materials: A chest of treasure and hints. Physics 2006, 27, 333–351.
[CrossRef]
18. Xin, J.; Yang, W.T.; Zhao, B.Y.; An, G.P. Design of interior off-line temperature detector for the ultrahigh
pressure vessel. Appl. Mech. Mater. 2012, 229–231, 1206–1212. [CrossRef]
19. Zhao, Y.L. Study on High Temperature Pressure Sensor. Ph.D. Thesis, Xi’an Jiaotong University, Xi’an,
China, 2003.
20. Wasisto, H.S.; Merzsch, S.; Waag, A.; Uhde, E.; Salthammer, T.; Peiner, E. Airborne engineered nanoparticle
mass sensor based on a silicon resonant cantilever. Sens. Actuators B Chem. 2013, 180, 77–89. [CrossRef]
21. Wasisto, H.S.; Merzsch, S.; Uhde, E.; Waag, A.; Peiner, E. Handheld personal airborne nanoparticle detector
based on microelectromechanical silicon resonant cantilever. Microelectron. Eng. 2015, 145, 96–103. [CrossRef]
22. Kanda, Y. Piezoresistance effect of silicon. Sens. Actuators A Phys. 1991, 28, 83–91. [CrossRef]
23. Sahin, Y. Recent progress in processing of tungsten heavy alloys. J. Powder Technol. 2014, 2014, 22. [CrossRef]

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).