Journal of Materials Science: Materials in Electronics (2019) 30:15705–15714

https://doi.org/10.1007/s10854-019-01955-0

Design and fabrication of SOI technology based MEMS differential

capacitive accelerometer structure
Nidhi Gupta1,2 · Shankar Dutta1 · Abha Panchal1 · Isha Yadav1 · Surender Kumar1 · Yashoda Parmar1 ·
Siva Rama Krishna Vanjari2 · K. K. Jain1 · D. K. Bhattacharya1

Received: 17 April 2019 / Accepted: 30 July 2019 / Published online: 2 August 2019
© Springer Science+Business Media, LLC, part of Springer Nature 2019

Abstract
This paper discusses the design and fabrication of MEMS differential capacitive accelerometer (z-axis sensitive) structure.
The accelerometer structure consists of one each movable and reference capacitors in the single accelerometer die fabricated
using highly conductive (p-type, resistivity: 0.001 Ω cm) SOI substrate. Resonant frequencies of the designed movable and
reference capacitive structures were found to be 9.6 kHz and 150 kHz respectively. Corresponding rest capacitance (at 0 g)
of both the capacitors was 2.21 pF. The movable and reference structures showed a deflection of 0.14 µm and 0.6 nm respec-
tively at 50 g applied acceleration. Corresponding changes in capacitances of the movable and reference capacitors were
82.3 fF and < 0.33 fF respectively. The designed accelerometer showed a scale factor sensitivity of the movable capacitor was
of ~ 1.65 fF/g. The device demonstrated a dynamic range of in − 17 g to 42 g with a full-scale non-linearity of ~ 3%. Cor-
responding measured scale factor sensitivity in the centrifuge test was found to be ~ 47 mV/g with an acceleration resolution
of ~ 17 mg. The device exhibited cross-axis sensitivity of ~ 2% in the full-scale range. Measured 3 dB bandwidth (380 Hz)
of the device matches reasonably with the simulated value (~ 400 Hz).

1 Introduction drift, low noise, high sensitivity, excellent linearity and low
power dissipation characteristics [1, 3, 21, 26, 29, 33, 34].
In recent times, accelerometers based on micro-electro- In MEMS capacitive accelerometer structure, change
mechanical system (MEMS) fabrication technology are in capacitance arises due to either change in overlap area
being employed in wide variety of defence applications like between capacitor plates [19, 21, 27, 30, 31, 36, 37] or
navigation and guidance of missiles, rockets and artillery change in gap (1 µm to 7 µm) between them [21, 25, 27, 32,
projectiles, unmanned aerial vehicles (UAVs) [1–6], explo- 34, 38, 39]. The capacitive accelerometer structures can have
sive impact detection [7]. MEMS accelerometers in micro-g higher scale factor sensitivity either by raising the overlap
acceleration range are also becoming crucial in micro and area by fabricating large number of inter-digitated comb-like
nano satellites [8]. Accelerometers are also being extensively capacitive electrodes; or by decreasing the air gap between
used in industrial [9, 10] and biomedical [11] applications. the capacitor plates. Nevertheless, the comb-type capaci-
Acceleration can be sensed by employing different sens- tive structure experiences higher fringing field due to low
ing principles: piezoelectric [5, 11–13], piezoresistive [3, area to gap ratio in individual comb-electrodes [21, 30, 37];
5, 14], resonant [15–17] and capacitive [3, 6, 8, 18–35], and dust/chipping elements often stick in between the comb-
tunneling [3, 12] etc. In navigational applications, capaci- fingers during MEMS devices dicing and get it electrically
tive sensing principle is favored due to its low temperature shorted [21, 31, 37]. On the other hand, decreasing the elec-
trode gap yield poorer dynamic range, as-well-as reduced
operating bandwidth due to higher squeeze film damping [5,
* Shankar Dutta 8, 10, 12, 21, 22, 29, 34]. Therefore, selection of the device
[email protected] dimensions and the control over associated fabrication pro-
cesses are very crucial to realize highly reproducible quality
1
Solid State Physics Laboratory, DRDO, Lucknow Road, of capacitive MEMS accelerometer structure.
Timarpur, Delhi 110054, India
In past, several researchers have developed MEMS
2
Department of Electrical Engineering, Indian Institute capacitive accelerometer structure by both surface
of Technology, Hyderabad 502285, India

13
Vol.:(0123456789)
15706 Journal of Materials Science: Materials in Electronics (2019) 30:15705–15714

micromachining [4–6, 24, 29, 32, 35, 38, 39] as well as bulk capacitive structure comprises of a 1000 µm × 1000 µm sili-
micromachining [5, 6, 19–28, 30, 31, 33, 36, 37] techniques. con proof mass over-hanged by four L shaped beams on a
The bulk micromachining technique predominantly uses sin- 4 µm glass cavity as shown schematically in Fig. 1a. Whereas,
gle crystal silicon as the mechanical structure, which is hav- Fig. 1b shows the schematic of the acceleration in-sensitive
ing much superior mechanical properties than that of poly- capacitor consisting of a rigid double-sided clamped beam
silicon or other metal structures being utilized in surface (1200 µm × 1000 µm) suspended over a 4 µm glass cavity.
micromachining techniques. Thus, for high performance, Thickness of both the capacitive structures is defined by the
low-drift MEMS capacitive accelerometers are still being active layer thickness (30 µm) of the SOI substrate. The glass
fabricated out of single crystal silicon using bulk microma- cavities have Au bottom electrodes to form the capacitor
chining technique. Among different bulk micromachining structure.
techniques, dissolved wafer process (DWP) is often being
used to fabricate different MEMS structures including accel- 3 Design and simulations
erometer. In this technique, thickness of the MEMS device is
being defined by fabricating regions of deep boron diffusion The deflection of movable (acceleration sensitive) capacitor
­p++ Si etch-stop layer [3, 21, 22, 25, 30, 31, 36, 37]. How- proof-mass due to the applied acceleration can be written as:
ever, fabrication of p­ ++ Si etch-stop layer beyond 12–15 µm
thicknesses is realistically difficult. It also generates high
Mm az = kmov zmov , (1)
level of residual stress (450–900 MPa), which might be det- where, kmov be the total stiffness constants of the four L
rimental for MEMS device operation [37, 40, 41]. shaped beam defined as [42]:
Alternatively, silicon on insulator (SOI) wafer based
4
DWP technique can be employed for fabrication of MEMS kmov = 1 ( 3 ) 1
,
(2)
devices [33, 34]. Here, the top silicon layers (Active layer) Ewm t3
L1 + L23 + L2 L
2𝛽Gtw3m 1 2
can be chosen for the exact thickness, crystal orientation,
and conductivity for the required MEMS device application; where, ­Mm = mass of the proof-mass; az = applied accel-
and the buried oxide layer of SOI substrate provides a bril- eration; zmov = displacement of the proof-mass; E = Young’s
liant etch stop characteristic during the DWP. modulus; wm = beam width; ­L1 and ­L2 are length of L shaped
In order to sense minute capacitance change (in femto- beams (see Fig. 1(a)); β = shape constant; G = shear modu-
farad), differential capacitance measurement principle lus; t = thickness of the device structure.
is being employed. Generally, this is done by creating Deflection of the movable capacitor proof-mass can be
push–pull type capacitive structure fabricated by three expressed as:
wafer bonding technology [34, 38, 39]. In this paper, we Mm
have presented the design of MEMS capacitive accelerom- zmov = a. (3)
kmov z
eter (z axis) structure operated in differential capacitance
sensing mode using two wafer bonding technology. Here,
the differential capacitive sensing mode is conceptualized by
creating one acceleration sensitive and another acceleration
in-sensitive MEMS capacitive structures. The acceleration
in-sensitive capacitive structure is required for the reference
capacitance. The device was fabricated by SOI wafer tech-
nology based dissolved wafer process involving glass-silicon
wafer bonding. The fabricated accelerometer was tested by
hybrid integration of the sensor along with commercial read-
out integrated circuit (ROIC) in a lead-less ceramic (LCC)
package.

2 Device structure

The present MEMS capacitive accelerometer consists of two

capacitive structures: one is acceleration sensitive and other
one is acceleration in-sensitive. This is required to implement
differential capacitive sensing without any three-wafer stack Fig. 1  Schematic of capacitive accelerometer: a acceleration sensitive
based push–pull type configuration. The acceleration sensitive capacitor structure and b acceleration in-sensitive capacitor structure

13
Journal of Materials Science: Materials in Electronics (2019) 30:15705–15714 15707

On the other hand, the reference capacitor (acceleration Table 1  Optimized dimensions of the MEMS differential capacitive
in-sensitive) has a double-sided clamped rigid beam struc- accelerometer structure
ture having stiffness constant (kref) [24]: Parameters Dimensions (µm)
( )3 Movable capacitor
t
kref = 32Ewf . (4) Proof-mass area 1000 × 1000
Lf
Beam length (L1) 1150
Beam length (L2) 100
Displacement of the double-sided clamped rigid beam
Beam width (wm) 30
structure (under applied acceleration: az) varies along its
Thickness (t) 30
length as follows [12, 24]:
Area of bottom Au electrode 1000 × 1000
Mf Gap between top and bottom electrodes 4
( )2
zref = x 2 Lf − x a z . (5) Reference capacitor (double-sided clamped rigid beam)
2ELf wf t3
Beam length (Lf) 1200
Beam width (wf) 1000
With maximum displacement at the centre of the beam
Thickness (t) 30
(x = Lf/2):
Area of bottom Au electrode 1000 × 1000
Gap between top and bottom electrodes 4
| Mf Lf3 M
zref | = a = ba , (6)
|max 32Ewf t3 z kref z

where, ­Mb = mass of the double-sided clamped rigid beam After estimating stiffness constants of both the mov-
structure; az = applied acceleration; zref = displacement of able and reference MEMS capacitive structures by using
the beam structure; E = Young’s modulus; wb = beam width; Eqs. (2) and (4), the corresponding resonant frequencies
­Lf = length of the rigid beam; t = thickness of the device were estimated by using Eq. (7). Estimated stiffness con-
structure. stants of the movable and reference structures were 237 N/m
Resonant frequencies (f) of both the movable and rest and 6.5 × 104 N/m respectively. The corresponding resonant
capacitive structures can be estimated from the following frequencies of the movable and rest MEMS capacitive struc-
equation [12, 21]: tures are found to be 9.26 kHz and 140 kHz respectively.
√ From Eqs. (3) and (6), the calculated values of deflec-
1 k tions under the 50 g applied acceleration were found to be
f = , (7)
2𝜋 M 0.145 µm and 0.63 nm corresponding to the movable and
reference structures. Deflection of the movable structure is
where, k be the stiffness constants and M be the mass of the
230 times more than the deflection of the reference structure.
respective structures.
Thus, deflection of the double-sided clamped beam can be
The minimum detectable acceleration (noise equivalent
ignored in comparison to the movable structure.
acceleration) can be determined by mechanical-thermal
Calculated values of rest capacitance of the reference and
noise, known as Brownian noise:
movable structure were ~ 2.21 pF. At 50 g applied accel-
√ eration, the change in capacitance of the movable capaci-
4kB T𝜔0
amin = , (8) tor would be 83.2 fF; whereas, the change in capacitance
MQ of the reference capacitor would be < 0.33 fF from its rest
capacitance value. The movable capacitor has a scale factor
where, amin = the minimum detectable acceleration when sig- of 1.664 fF/g. In the present accelerometer structure, the
nal to noise ratio (SNR) is unity; ­kB = Boltzman’s constant; calculated minimum detectable acceleration is found to be
T = temperature in Kelvin scale; M = mass of the acceler- 38.65 µg.
ometer structure (movable capacitor); Q = Quality factor ≈
1/2ζ; ζ = damping factor. 3.1 FEM simulations
The device dimensions are reference after optimization of
the proof-mass deflections and spring constants of the mov- The performances of the accelerometer structure are evalu-
able and rest capacitor structures. For the present study, the ated by using finite element method (FEM) analysis using
optimized dimensions are tabulated in Table 1. Coventorware software (version 10.0). Manhattan brick

13
15708 Journal of Materials Science: Materials in Electronics (2019) 30:15705–15714

mesh is chosen for the FEM simulation and mesh size is the movable capacitor proof-mass (0.14 µm), which matches
optimized to 15:15:1 corresponding to x:y:z directions. quite well with the calculated values. Figure 4 shows the
The modal, static and dynamic behaviours of the capaci- proof mass displacement of both the structures at 50 g.
tive MEMS accelerometer structure are presented in this The capacitances of the movable ­(Cm) and reference ­(Cr)
section. The modal patterns of the movable and reference capacitors at different g were estimated from the simulation
capacitive structures were simulated to envisage it’s favora- data and plotted in Fig. 4a. The resultant change in capaci-
ble deflection modes. The first three modal patterns of the tance between movable and reference capacitors (­ Cm − Cr)
movable structure are shown in Fig. 2a–c respectively. Cor- was varying linearly with applied acceleration as shown in
responding modal frequencies of the movable structure are Fig. 4b. The simulated change in capacitance of the movable
found to be 9.67 kHz, 17.05 kHz and 17.05 kHz respectively. capacitor was found to be 80.24 fF at 50 g. The correspond-
The first three modal frequencies of the reference capaci- ing scale factor was found to be ~ 1.605 fF/g. However, since
tive structure were found to be 156.8 kHz, 196.0 kHz and the reference capacitor also changes marginally (< 0.37 fF
387.9 kHz as shown in Fig. 2d–f respectively. Deflections at 50 g), the error generated during the acceleration sensing
of the accelerometer structure are simulated with the input also need to be estimated. The simulation results showed a
acceleration (0–50 g). Figure 3 shows the displacement and minimum detectable acceleration of 39.25 µg. The simu-
Von-misses stress contours of the movable and reference lated results were matching well to the theoretically calcu-
capacitor structures at 50 g acceleration. From Fig. 3a and lated data. The full-scale non-linearity (0–50 g range) of the
b, it is quite clear that at 50 g the reference capacitor has simulated structure is found to be ~ 3% due to the change in
negligibly small deflection (0.67 nm) compared to that of gap approach. Figure 4 (b) shows the error in sensing the

Fig. 2  Modal analysis: movable MEMS capacitor a mode 1: 9.67 kHz; b mode 2: 17.05 kHz; c mode 3: 17.05 kHz; and Reference MEMS
capacitor d mode 1: 156.8 kHz; e mode 2: 196.0 kHz; f mode 3: 387.9 kHz

13
Journal of Materials Science: Materials in Electronics (2019) 30:15705–15714 15709

Fig. 3  FEM simulated response

of the differential capacitive
accelerometer structure under
50 g acceleration: movable
capacitor a deflection, b von-
Mises stress; and reference
capacitor c deflection, d von-
Mises stress

acceleration in the current design. The simulated device per- the accelerometer structure is assessed using the damping
formance showed an error of output capacitance of 232 mg solver module of Coventorware software. Due to the large
over 50 g acceleration (7.412 aF/g). The error value is found ratio of capacitor plate area and its gap, the extent of squeeze
to be linearly dependent on acceleration sensing range as film damping would be quite severe. The estimated damping
also shown in Fig. 4b. Thus, fairly accurate accelerom- factor (the ratio of damping coefficient arise due to squeeze
eter output can be achieved by adding the error term into film effect and critical damping) of the accelerometer struc-
it. Therefore, one can write the output of the differential ture is found to be > 100 (hugely over-damped scenario). To
capacitive accelerometer as: minimize the squeeze film damping, a large number of holes
( ) ( ) are introduced in the proof-mass. After the introduction of
ΔC = Cr − Cm + 7.412 × 10−3 az . (9) holes, the damping factor reduced to 5.22. The 3 dB band-
Frequency response of the accelerometer greatly depends width of the accelerometer is estimated from the normalized
on the magnitude of damping force developed due to the displacement curve as shown in Fig. 5. Using the FEM simu-
squeeze film effect of the trapped air (or other fluidic lation study, the estimated bandwidth of the accelerometer
medium) between the movable capacitor proof mass and structure is found to be ~ 400 Hz.
bottom plate/package surface. The damping coefficient of

13
15710 Journal of Materials Science: Materials in Electronics (2019) 30:15705–15714

4 Fabrication and testing

The differential capacitive accelerometer structure (the

movable and reference capacitors) was realized by using
silicon-on-insulator (SOI) substrate. For this fabrication,
high conductivity (resistivity: 0.001 Ω cm) p type SOI
wafer, having 30 μm active layer and 1 μm buried oxide
layer, was chosen. A three mask dissolved wafer process
(DWP) was used to release the accelerometer structure.
The cross-section and top views of the fabrication process
steps of the differential capacitive accelerometer struc-
ture are shown schematically in Fig. 6. Thickness of the

Fig. 4  Simulated variation of capacitance versus acceleration: a ref-

erence and movable capacitances; and b change in capacitance and
associated error

1.2
Normalized displacement

1.0

0.8

0.6

0.4

0.2

0.0

1 2 3 4
10 10 10 10
Frequency (Hz)

Fig. 5  Simulated frequency response of the accelerometer structure

Fig. 6  Fabrication process steps followed for fabrication of the differ-

ential MEMS capacitive accelerometer structure

13
Journal of Materials Science: Materials in Electronics (2019) 30:15705–15714 15711

accelerometer structure was defined by the active layer tested to gauge their performance under applied accelera-
thickness (~ 30 μm) of the SOI wafer. The accelerometer tion. Details of the packaging and testing methodology and
structure (movable and reference capacitor structures) associated results are presented elsewhere [43].
was patterned on the highly conductive active layer of the
SOI substrate using 2 μm thick Shipley S1818 photoresist
(positive) layer (mask 1). Thereafter, the pattern was trans- 5 Results and discussions
ferred into the SOI substrate by deep reactive ion etching
technique using (SPTS Pegasus DRIE system) as shown The reference and acceleration sensitive capacitors of the
in Fig. 6b. The DRIE experiment was performed by using accelerometer structures were fabricated by following the
Bosch process with etching (330 sccm S ­ F6 and 33 sccm above process sequence. Optical and SEM images of the
­O 2 flow rates) cycle RF power of 2500 W and polymer fabricated differential capacitive accelerometer structure are
deposition (150 sccm C ­ 4F 8 flow rate) cycle at 2200 W. shown in Figs. 7 and 8 respectively. Figure 8 shows SEM
The etching and deposition cycle times were kept at 7 s images of the batch fabricated reference and movable capaci-
and 4 s respectively. tors of the accelerometer structures. Each accelerometer die
Simultaneously, a Pyrex glass (7740) substrate was is comprised of one reference and one acceleration sensi-
processed for handling the delicate 30 μm thick MEMS tive capacitor. Figure 9 shows the isometric view of the sus-
structures. The glass substrate was coated with Cr–Au lay- pended movable capacitive structure along with the metal
ers (30 nm Cr/100 nm Au) by thermal evaporation process contact pads. Figure 10 shows the etched cavity in the glass
and subsequently Au electroplated (~ 1 μm) by electroplat- substrate along with the anodically bonded silicon active
ing as shown in Fig. 6c. This Au layer was then patterned layer of the SOI substrate.
with masking layer (mask 2) to fabricate cavities (4.5 μm Rest capacitance of the movable and reference capacitor
depth) in the glass substrate using wet chemical etching in structures were tested at the wafer level. Both the capaci-
­HNO3and HF acidic mixture (3:7) as shown in step (d) and tors of the accelerometer sensor die were found to have
(e) of Fig. 6. After the etch-pit formation, the Cr–Au layer capacitance in the range of 2.15–2.25 pF. The variation
was removed and then the substrate was coated with Ti–Au in the rest capacitance seems to be due to the variation of
layer (30 nm Ti/100 nm Au) using sputtering technique as etch pit depth over the large 150 mm diameter substrates.
shown in step (f) and (g) of Fig. 6. Thereafter, the selected accelerometer dies were packaged
The Au layer was thickened (0.5 μm) by electroplating at in a LCC package. Figure 11a shows the MEMS based dif-
the bottom electrode, contact lines and electrical bond-pad ferential capacitive accelerometer die packaged in a LCC
areas (using mask 3). The Ti–Au layer was removed from package along with MS3110 ROIC die. The ROIC gain
the areas other than the bottom electrode, contact lines and was kept at 30 mV/fF for the capacitance to voltage output.
electrical bond-pad areas as shown schematically in Fig. 6h.
This makes an effective gap of 4 μm between the bottom Au
electrode and the reference/movable capacitor proof masses.
The patterned SOI and glass substrates were then aligned
together and bonded (anodic bonding) as shown in step (i) of
Fig. 6. Thereafter, bulk of the SOI substrate was etched away
in 35 wt% KOH solution. The etching easily gets stopped
after reaching the 1 μm oxide layer. Finally, the accelerom-
eter structure (active layer) was released by removing the
buried oxide layer (1 μm) by using vapor phase etching in
HF vapour as shown in Fig. 6j. Different parts of the released
reference and movable capacitive structures were studied
by optical microscopy (Make: Carl Zeiss Axio-imazer Z1)
and field-emission scanning electron microscopy (FE-SEM)
(Make: SUPRA-50).
Rest capacitance of the reference and acceleration-sen-
sitive capacitors of the fabricated accelerometer structure
was measured at the wafer level (before chip dicing). The
devices were then assembled in lead less ceramic (LCC)
package along with commercial read-out integrated circuit
(ROIC) chip (IRVINE SENSORS MS3110) and then sealed Fig. 7  Optical image of the fabricated differential capacitive acceler-
in nitrogen environment. Packaged accelerometers were then ometer with movable and reference capacitors

13
15712 Journal of Materials Science: Materials in Electronics (2019) 30:15705–15714

Fig. 8  SEM image of the batch fabricated capacitive accelerometer

Fig. 11  Packaged differential capacitive accelerometer: a in a LCC

package, b reference in an evaluation board

The interconnections between the accelerometer bond pads,

ROIC bond pads and package bond pads are done by wire
bonding (2 mil Au wire). Figure 11b shows the packaged
Fig. 9  SEM image of the movable capacitor along with Au bond pads
accelerometer reference in the evaluation board for further
testing.
Packaged accelerometer along with the evaluation board
was mounted on Bruel & Kjear LDS shaker system (model
no v201-PA 25E) for testing. Figure 12 shows the typical
response of the device under a shock pulse of 30 g for 5 ms
duration. Frequency response of the device also measured
in the same shaker system with 1 g acceleration and sweep-
ing frequency (5–1000 Hz). The fabricated accelerometer
exhibited a 3 dB bandwidth of 380 Hz which matches well
with the simulated FEM based result (~ 400 Hz). Flip test-
ing of the packaged accelerometer yielded a scale factor of
47.5 mV/g under gravitational acceleration (± 1 g). In the
centrifuge test, the device showed a dynamic range of in
– 17 g to 42 g acceleration with a full-scale non-linearity
of ~ 3%. Measured scale factor sensitivity in this acceler-
ation range was found to be ~ 47 mV/g (which is equiva-
lent to 1.57 fF/g). The scale factor sensitivity depends on
the dynamic range of the accelerometer. Lower dynamic
Fig. 10  SEM image of the glass cavity and anodically bonded anchor ranges (3–6 g) correspond to higher scale factor sensitivity
region (205–330 mV/g) [4, 7] and vice versa. The scale factor of

13
Journal of Materials Science: Materials in Electronics (2019) 30:15705–15714 15713

Acknowledgements The authors would like to thank Dr. Seema

Vinayak, Director, Solid State Physics Laboratory (DRDO) for her
guidance and for the permission to publish this work. The authors also
sincerely thank fabrication team of STARC, Bangalore for their help
and support. Help from other colleagues of MEMS and Characteriza-
tion divisions of SSPL are also acknowledged.

References
1. D.K. Shaeffer, IEEE Commun. Mag. 51, 100–109 (2013)
2. N.F. Watson, U.S. Patent 4601206 A (1986)
3. N. Yazdi, F. Ayazi, K. Najafi, Proc. IEEE 86, 1640–1659 (1998)
4. S.H. Tseng, P.C. Wu, H.H. Tsai, Y.Z. Juang, Microelectron. Eng.
119, 178–182 (2014)
5. V. Narasimhan, H. Li, M. Jianmin, J. Micromech. Microeng. 25,
033001 (2015)
6. M. Benmessaoud, M.M. Nasreddine, Microsyst. Technol. 19, 713
(2013)
Fig. 12  Accelerometer response in vibratory shaker measurement 7. P. Ragam, D.S. Nimaje, IEEE Sens. Lett. 2, 2501304 (2018)
8. R. Abdolvand, B.V. Amini, F. Ayazi, IEEE MEMS 16, 1036
(2007)
the current design (47 mV/g) is at par with the other reported 9. A.L. Gesing, E.D.P. Alves, S. Paul, J.A. Cordioli, Sci. Rep. 8,
values (50–60 mV/g) [21, 37]. Corresponding cross-axis 3920 (2018)
sensitivity of the differential capacitive accelerometer 10. W. Wu, D. Liu, W. Qiu, H. Liu, F. Hu, J. Fan, C. Hu, L. Tu, J.
was ~ 2% in the full-scale range. Resolution of the device Micromech. Microeng. 28, 125016 (2018)
11. G.S. Maruthi, V. Hegde, IEEE Sens. 16, 145–152 (2016)
was estimated using inclinometer measurement in near ver- 12. M. Bao, Analysis and Design Principles of MEMS Devices (Else-
tical tilt condition. The estimated resolution of the device vier, Amsterdam, 2005)
was found to be ~ 17 mg. Details of these testing results are 13. N. Gupta, A. Pandey, S.R.K. Vanjari, S. Dutta, Microsyst. Tech-
reported elsewhere [43]. nol. (2019). https​://doi.org/10.1007/s0054​2-019-04334​-1
14. M. Messina, J. Njuguna, C. Palas, Sensors 18, 289 (2018)
15. C.M. Yu, A. Ali, S. Ghosh, J.E.Y. Lee, IEEE Transducers, 607–
610 (2017)
6 Conclusions 16. S.A. Zotov, B.R. Simon, A.A. Trusov, A.M. Shkel, IEEE Sens.
15, 5045–5052 (2015)
17. H.J. DeLos Santos, Introduction to Micro-electro-mechanical
High performance capacitive accelerometers are being used (MEMS) Microwave Systems, 2nd edn. (Artech House, Boston,
in wide variety of defense applications. This paper presents 2002)
design, simulation and fabrication of MEMS differential 18. X. Zhou, L. Che, S. Liang, Y. Lin, X. Li, Y. Wang, Microelectron.
capacitive accelerometer (z-axis sensitive) structure. The Eng. 131, 51–57 (2015)
19. S. Tez, U. Aykutlu, M.M. Torunbalci, T. Akin, IEEE MEMS 24,
differential capacitance concept is implemented by design- 1264–1274 (2015)
ing movable and reference capacitors in the single accel- 20. A. Utz, C. Walk, A. Stanitzki, M. Mokhtari, M. Kraft, R. Kokoz-
erometer die. The rest capacitance of both the capacitors inski, IEEE Sens. 18, 6533–6539 (2018)
was 2.21 pF (at 0 g). At 50 g acceleration, the changes in 21. R.K. Bhan, Shaveta, Md Imran, R. Pal, S. Dutta, I. Yadav, Sens.
Transducers 203, 8–15 (2016)
capacitances of the movable and reference capacitors were 22. R.K. Bhan, Shaveta, A. Panchal, Y. Parmar, C. Sharma, R. Pal, S.
82.3 fF and < 0.33 fF respectively. Corresponding scale fac- Dutta, Sens. Transducers 192, 44–52 (2015)
tor sensitivity of the movable capacitor was 1.65 fF/g. Esti- 23. W. Li, Z. Song, X. Li, L. Che, Y. Wang, Microelectron. Eng. 115,
mated minimum detectable acceleration of the accelerometer 32–38 (2014)
24. S. Franssila, Introduction to Microfabrication, 2nd edn. (Wiley,
structure was of ~ 39 µg. The accelerometer structure was Hoboken, 2010)
fabricated by SOI wafer based DWP technology. Fabricated 25. R.K. Bhan, Shaveta, M.D. Imran, R. Pal, S. Dutta, Sens. Transduc-
accelerometer was assembled in a LCC package along with ers 189, 128–136 (2015)
readout circuitry. The device showed a dynamic range of 26. A. Aydemir, Y. Terzioglu, T. Akin, Sens. Acta A 244, 324–333
(2016)
− 17 g to 42 g in the centrifuge with a full-scale non-linear- 27. D. Felnhofer, K. Khazeni, M. Mignard, Y.J. Tung, J.R. Webster,
ity of ~ 3%. Corresponding scale factor sensitivity was found C. Chui, E.P. Gusev, Microelectron. Eng. 84, 2158–2164 (2007)
to be ~ 47 mV/g. The device exhibited cross-axis sensitivity 28. D. Xiao, Q. Li, Z. Hou, D. Xia, X. Xu, X. Wu, Sens. Acta A258,
of < 2% in the full-scale range. The device showed operat- 182–192 (2017)
29. M.G. Guney, X. Li, V.P.J. Chung, J. Paramesh, T. Mukherjee, G.K.
ing bandwidth (3 dB) of 380 Hz which matches satisfacto- Fedder, IEEE MEMS Conf. 2018, Belfast, Northern Ireland, UK
rily with the simulated result (~ 400 Hz). Resolution of the (2018)
device was found to be ~ 17 mg.

13
15714 Journal of Materials Science: Materials in Electronics (2019) 30:15705–15714

30. S. Dutta, Shaveta, R. Pal, D.K. Bhattacharya, P. Datta, R. Chat- 38. MdS Mahmood, Z.C. Butler, D.P. Butler, Sens. Acta A263, 530–
terjee, Sens. Transducers 91, 31–38 (2008) 541 (2017)
31. S. Dutta, R. Pal, P. Kumar, O.P. Hooda, J. Singh, Shaveta, G. 39. X. Zhou, L. Che, Y. Lin, X. Li, J. Wu, Y. Wang, Microsyst. Tech-
Saxena, P. Datta, R. Chatterjee, Sens. Transducers 111, 18–24 nol. 20, 1365–1372 (2014)
(2009) 40. S. Dutta, A. Pandey, M. Singh, R. Pal, Mater. Lett. 164, 316–319
32. M.H. Tsai, Y.C. Liu, W. Fang, IEEE MEMS 21, 1329–1337 (2016)
(2012) 41. S. Dutta, A. Pandey, M. Singh, R. Pal, Mater. Lett. 170, 76–79
33. K.K. Mistry, K.B.M. Swamy, S. Sen, Microsyst. Technol. 16, (2016)
2057–2066 (2010) 42. G.M. Rebeiz, RF MEMS: Theory, Design, and Technology (Chap-
34. S. Dutta, P. Saxena, A. Panchal, R. Pal, K.K. Jain, D.K. Bhat- ter 2) (Wiley, New York, 2003)
tacharya, Microsyst. Technol. 24, 4855–4862 (2018) 43. Y. Parmar, N. Gupta, V. Gond, S.S. Lamba, S.R.K. Vanjari, S.
35. M. Keshavarzi, J.Y. Hasani, Microsyst. Technol. (2018). https​:// Dutta, K.K. Jain, D.K. Bhattacharya, Microsyst. Technol. (2019).
doi.org/10.1007/s0054​2-018-4187-5 https​://doi.org/10.1007/s0054​2-019-04561​-6
36. S. Dutta, M. Kumar, S. Kumar, Md Imran, I. Yadav, A. Kumar, P.
Kumar, R. Pal, J. Mater. Sci. 25, 1984–1990 (2014) Publisher’s Note Springer Nature remains neutral with regard to
37. S. Dutta, Shaveta, Md Imran, R. Pal, R.K. Bhan, J. Mater. Sci. 25, jurisdictional claims in published maps and institutional affiliations.
3828–3832 (2014)

13