www.nature.

com/scientificreports

OPEN Reduced order modelling

and experimental validation
of a MEMS gyroscope
test‑structure exhibiting 1:2
internal resonance
Giorgio Gobat1, Valentina Zega1*, Patrick Fedeli2, Luca Guerinoni2, Cyril Touzé3 &
Attilio Frangi1

Micro-Electro-Mechanical Systems revolutionized the consumer market for their small dimensions,
high performances and low costs. In recent years, the evolution of the Internet of Things is posing new
challenges to MEMS designers that have to deal with complex multiphysics systems experiencing
highly nonlinear dynamic responses. To be able to simulate a priori and in real-time the behavior of
such systems it is thus becoming mandatory to understand the sources of nonlinearities and avoid
them when harmful or exploit them for the design of innovative devices. In this work, we present
the first numerical tool able to estimate a priori and in real-time the complex nonlinear responses of
MEMS devices without resorting to simplified theories. Moreover, the proposed tool predicts different
working conditions without the need of ad-hoc calibration procedures. It consists in a nonlinear Model
Order Reduction Technique based on the Implicit Static Condensation that allows to condense the high
fidelity FEM models into few degrees of freedom, thus greatly speeding-up the solution phase and
improving the design process of MEMS devices. In particular, the 1:2 internal resonance experienced
in a MEMS gyroscope test-structure fabricated with a commercial process is numerically investigated
and an excellent agreement with experiments is found.

The spread of Micro-Electro-Mechanical Systems (MEMS) in the consumer world triggered a revolution in
gaming, mobile phones and navigation. Similarly, in the near future, the evolution of the Internet of Things in its
different declinations will require new generations of sensors and actuators with improved performances, smaller
dimensions and innovative working principles. As a consequence, MEMS designers more and more frequently
will have to deal with complex mechanical structures exhibiting nonlinear dynamic ­behaviors1,2.
Among others, MEMS gyroscopes represent a meaningful example of such a trend. They are electro-mechan-
ical systems able to measure the angular rate by exploiting the Coriolis force. To guarantee a correct function-
ing, where at least two modes are coupled through the Coriolis force, and to satisfy the strict requirements on
the footprint, the mechanical structure is usually very complex, i.e. made by folded springs and rigid masses.
Nonlinear phenomena often a­ rise3–8 but are difficult to decipher.
Among the wide variety of nonlinear phenomena arising in MEMS devices, internal resonance, i.e. when two
or more modes get nonlinearly coupled and exchange energy, is attracting increasing interest for its potential
benefits on the performances of MEMS d ­ evices2,9–12. It has been demonstrated that thanks to internal resonance
it is indeed possible (i) to stabilize the oscillation frequency of non-linear self-sustaining micromechanical
­resonators13, (ii) to redistribute and store mechanical energy among vibrational modes and coherently transfer
it back to the principal one when the external excitation is ­off14 and (iii) to tune the quality factor Q of the driven
mode over a wide r­ ange15,16. Moreover, internal resonance has been recently employed in MEMS gyroscopes as a
new and very promising detection technique of angular rate ­signals17 and to design innovative MEMS bandpass
­filters18.

1
Politecnico di Milano, Civil and Environmental Engineering, Milan, Italy. 2Analog and MEMS Group,
STMicroelectronics, Cornaredo, Italy. 3IMSIA, ENSTA Paris, Institut Polytechnique de Paris, Palaiseau,
France. *email: [email protected]

Scientific Reports | (2021) 11:16390 | https://doi.org/10.1038/s41598-021-95793-y 1

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol.:(0123456789)
 www.nature.com/scientificreports/

In view of its high potentiality for the design of innovative and high-end MEMS devices, internal resonance
has been analyzed ­theoretically19 and experimentally verified on a variety of simple MEMS ­structures20–24, ranging
from arch r­ esonators25–27 to micro-mirrors28. In most cases, the coefficients of Reduced Order Models (ROMs)
are obtained from simplified electro-structural t­ heories26,29 or are calibrated on experimental d ­ ata23. Despite the
great interest of the topic, a general a priori simulation tool that could predict in real-time the nonlinear dynamic
behavior of complex MEMS structures like e.g. gyroscopes under different actuation conditions, is still miss-
ing. Such a tool would also dramatically improve the design process and pave the way to a new class of sensors/
actuators experiencing complex nonlinear dynamic phenomena.
Numerical methods able to simulate the Full Order Model (FOM) have been proposed as a generalization of
simplified ­appoaches30, but their computational cost remains a major issue especially if complex MEMS structures
are considered. Dedicated Harmonic Balance techniques or shooting procedures are indeed overwhelmingly
complex and time c­ onsuming31,32.
As a consequence, the focus has been set on the generation of nonlinear ROMs starting e.g. from large FEM
models that might reshape the governing equations into a nonlinear, dynamical system featuring a much lower
dimensionality, yet able to capture the physical features of the p ­ roblem29,33–35. The Stiffness Evaluation Procedure
(STEP) in its various v­ ariants33 assumes a trial subspace spanned by a set of linear modes which however must
also include all the coupled high-frequency modes that are often difficult to i­ dentify36,37. The Proper Orthogonal
Decomposition (POD)38 is also based on a linear trial space but this is generated from a set of FOM snapshots
employing Singular Value Decomposition, thus allowing to identify all the relevant contributions automatically.
A different approach is taken by the implicit condensation and expansion (ICE) ­method39–41 which defines a small
set of master modes and assumes a quasi-static coupling with the high frequency contributions (slave modes).
Also modal derivatives (MD)42–44 have been introduced with the aim of accounting for the amplitude depend-
ence of modes. ICE and MD are indeed very accurate when a slow/fast separation between the frequencies of
the master and slave modes e­ xist45,46. Recently, Nonlinear Normal Modes (NNMs) have received considerable
attention as a technique for generating ROMs. Initially defined as a vibration in unison of the ­system47–49, they
have been later extended by the notion of invariant ­manifold50 and of spectral submanifold (SSM)51,52. However,
only very recently efficient approaches have been proposed for the computation of invariant manifolds for large
FEM ­models53,54, but applications have been limited so far to mechanical structures with geometrical nonlineari-
ties and no multiphysics coupling.
In this work, we elaborate on the Implicit Condensation approach based on static modal loadings recently
tailored by the authors for simple MEMS s­ tructures41. In particular, the ICE applies to structures which undergo
transformations which are no-longer infinitesimal, but still moderate. The approach has been verified on a
double-ended tuning fork resonator experiencing both geometric, electrostatic and damping n ­ onlinearities55,
and represents a fast a priori multiphysics simulation tool able to reproduce the nonlinear dynamics caused
by the interaction of two modes of a complex MEMS gyroscope test-structure without the need of calibration
procedures. To the authors best knowledge this represents the first fast numerical predictive tool able to simu-
late a priori the internal resonance phenomenon including bifurcations of the periodic response in a complex
structure and in general, the nonlinear dynamic behavior of MEMS devices. Numerical results are compared
with experimental data and an excellent agreement is achieved for different actuation conditions, thus proving
the versatility and the predictivity of the proposed tool.

Results
MEMS gyroscope test‑structure. A schematic view of the MEMS gyroscope test-structure employed
in this work is reported in Fig. 1a, close-up views and geometrical dimensions are also reported in the Supple-
mentary Information for the sake of clarity. The mechanical structure is constituted by four masses and several
folded springs that provide the suspension of the device and the coupling of the masses with a central auxiliary
component. The gyroscope test-structure is fabricated through the Thelma process of STMicroelectronics in
polysilicon (E = 167 GPa, υ = 0.22, ρ = 2330 kg/m3) and has an overall footprint of 1.5 mm × 1.3 mm × 24 µm.
Comb fingers and parallel plate electrodes allow for the in-plane actuation/readout, while electrodes located on
the substrate are employed for the out-of-plane actuation/readout. In Fig. 1b,c, two modes of the MEMS gyro-
scope test-structure are reported: they will be referred to in the following as roll mode (Fig. 1b) and spurious roll
mode (Fig. 1c). Their natural frequencies are computed through a FEM modal analysis and read f1 = 22,522 Hz
and f2 = 43,386 Hz, respectively.
1:2 internal resonance between the two modes can be triggered by driving the roll mode through the elec-
trodes on the substrate. The two linear natural frequencies have an initial ratio of 1.926, which evolves to almost
exactly 2 as the applied electrostatic bias increases and due to the electrostatic nonlinearities given by the parallel-
plate electrostatic s­ cheme55.

ROM based on implicit static condensation. A numerical FOM made by discretizing the geometry of
the MEMS gyroscopes test-structure with quadratic pentahedrons, properly distributed so as to have at least two
elements in the spring thickness, consists of around 2.5 millions of degrees of freedom which become 17 mil-
lions if electro-mechanical coupling is considered. This makes any direct numerical simulation computationally
unaffordable especially if complex nonlinear dynamic phenomena such as the 1:2 internal resonance are inves-
tigated. As an example, we estimate that the simulation of the nonlinear dynamic response of the MEMS quad-
mass structure through a fully coupled time domain analyses in COMSOL Multiphysics v.5.6, would require
approximatively one year on a standard workstation (AMD Ryzen 9 5950X, 16 Cores, 128 Gb RAM).
The implicit static condensation method validated by the authors on simple s­ tructures41,55 is here applied
to reduce the system to two degrees of freedom (i.e. the amplitudes of the two master roll and spurious-roll

Scientific Reports | (2021) 11:16390 | https://doi.org/10.1038/s41598-021-95793-y 2

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol:.(1234567890)
www.nature.com/scientificreports/

a proof masses
springs
auxiliary component
in-plane electrodes
out-of-plane electrodes
anchors

b c

Figure 1.  (a) Schematic view of the MEMS gyroscope test-structure. (b) Roll (f1 = 22,522 Hz) and (c) spurious
roll (f2 = 43,386 Hz) modes. The contour plot of the displacement field is shown in color.

modes), thus dramatically reducing the computational effort without losing accuracy and physical meaning
(see “Methods” section).
By considering a constant Direct Current (DC) voltage VDC on the MEMS gyroscope test-structure proof
masses and an Alternate Current (AC) signal VAC ≪ VDC at an angular frequency ω close to the one of the roll
mode on the electrodes on the substrate, the resulting nonlinear system describing the dynamics of the device
reads:
ω01 2
q̈1 +
Q1
q̇1 + β1 (q1 , q2 ) − F̌e11 (q1 , q2 )ǫ0 VDC = 2ǫ0 VDC VAC F̌e21 (q1 , q2 ) sin ωt, (1)

ω02 2
q̈2 +
Q2
q̇2 + β2 (q1 , q2 ) − F̌e12 (q1 , q2 )ǫ0 VDC = 0, (2)

where ǫ0 is the vacuum permittivity, qi is the modal coordinate, Qi is the quality factor, ω0i = 2πfi is the natural
pulsation, βi (q1 , q2 ) is the nonlinear mechanical force and F̌e1i (q1 , q2 ) is a time independent nonlinear electro-
static force for the i-th mode, with i = 1 for the roll mode and i = 2 for the spurious roll mode. F̌e21 (q1 , q2 ) is the
time dependent nonlinear electrostatic force that acts on the driven roll mode. In the following, Q1 = 2400 and
Q2 = 3480 according to the simplified numerical ­tool56 proposed by the authors to compute fluid damping in
MEMS resonant structures working in low pressure conditions such as in this case. Note that a nonlinear quality
factor can in principle be also considered in the case of very large displacements of the proof mass with respect
to the air gap between it and the fixed ­electrodes55.
For the sake of simplicity, we approximate βi (q1 , q2 ), F̌e1i (q1 , q2 ) and F̌e21 (q1 , q2 ) with a complete third order
polynomial whose coefficients are reported in the Supplementary Information. Once the voltage levels are fixed,
Eqs. (1) and (2) are solved through numerical continuation, using the package M ­ ANLAB57 that implements a
combination of Harmonic Balance (HB) with an asymptotic numerical method (ANM) for path-following.
The nonlinear frequency response of the roll mode in terms of amplitude and phase is reported in continuous
light blue line in Fig. 2 for a ­VDC = 4.28 V and a V
­ AC = 3.16 mV. For this actuation condition, the model correctly
reproduces the activated 1:2 internal resonance as demonstrated by the characteristic shape of the frequency
response made with two peaks and by the presence of a quasi-periodic/chaotic region (see green path in Fig. 2a of
Supplementary Information) delimited by Neimark–Sacker bifurcations (dark blue stars in Fig. 2) in the central
region of the s­ pectrum58. Red stars represent the Saddle–Node bifurcations predicted by the ROM model and
delimit the unstable part of the solution branch (see red path in Fig. 2a of Supplementary Information). To fur-
ther validate the adequacy of the proposed ROM, in the Supplementary Information we report the comparison
between the curves obtained through the full ROM here proposed and the ones analytically derived through

Scientific Reports | (2021) 11:16390 | https://doi.org/10.1038/s41598-021-95793-y 3

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol.:(0123456789)
 www.nature.com/scientificreports/

experimental data (upward sweep) Neimark-Sacker bifurcation (quasi-periodic regime)

numerical prediction (ROM model) Saddle node bifurcation (delimitating the unstable portion of the
branch)
a b
1.2
0.55

Normalized Current [-]

Normalized Current [-]

1.0
0.50
0.8
0.45
0.6

0.4 0.40

0.2 0.35

21.4 21.45 21.5 21.55 21.46 21.465 21.47

f [kHz] f [kHz]

c d
50
0

0
Phase [°]

Phase [°]
-20

-50 -40

-60
-100

21.4 21.45 21.5 21.55 21.46 21.464 21.468

f [kHz] f [kHz]

Figure 2.  Frequency response of the MEMS gyroscope test-structure for a V ­ AC = 3.16 mV and a ­VDC = 4.28 V
in terms of (a) amplitude and (c) phase. Close-up views of the quasi-periodic region in the frequency response
in terms of amplitude and phase are shown in (b)–(d). Numerical predictions are plotted with continuous blue
lines, and experimental data with orange dashed lines. Stars mark theoretical Neimark–Sacker (dark blue) and
Saddle–Node (red) bifurcations that delimit the quasi-periodic regime region and the unstable paths of the
frequency response, respectively.

the Multiple Scale ­Method58 from a simplified ROM based on the coefficients numerically extracted through
the Implicit Condensation Method.
In Fig. 2, numerical curves in terms of displacements obtained by integrating Eqs. (1) and (2) are converted
in terms of current as detailed in Zega et al.55 to simplify the comparison with experimental data.

Experimental results. In order to validate the proposed simulation tool, the experimental frequency
response of the MEMS gyroscope test-structure is measured in the same actuation condition previously con-
sidered for the theoretical model: ­VDC = 4.28 V and ­VAC = 3.16 mV. Experimental curves are reported in orange
dashed lines in Fig. 2 and well agree with the numerical predictions. Note that the jumps of the experimental
upward frequency sweep shown in Fig. 2a,c are in a satisfactory agreement with the Saddle–Node bifurcations
(red stars) predicted by the models (see Fig. 2a of Supplementary Information). Moreover, in the close-up views
of Fig. 2b,d, it is evident that the Neimark–Sacker bifurcations predicted by the ROM correctly delimit the
experimental quasi-periodic region, thus further proving the accuracy of the proposed a priori simulation tool.
Additional experimental curves measured for a ­VAC = 3.16 mV and different levels of ­VDC are reported in
dashed-lines in Fig. 3 together with corresponding numerical predictions. Only experimental upward frequency
sweeps are reported for the sake of clarity and all the curves, both numerical and experimental, are normalized
with the maximum amplitude of the hardening peak of the experimental curve obtained for V ­ DC = 4.28 V and
­VAC = 3.16 mV. This value corresponds to an out-of-plane maximum displacement of the proof masses of 71 nm.
The maximum displacement experienced by the proof masses in this experimental campaign is then in the order
of a couple of hundreds of nanometers (i.e. orange curve in Fig. 3a), which is fully compatible with a stable
operation of the device far from pull-in instabilities (the gap between the masses and the underlying electrodes
is of 1.2 µm) and with the assumption of moderate transformations required by the proposed ICE method. In
the inset of Fig. 3b, the evolution of the resonant frequency of the roll mode for different ­VDC is reported and

Scientific Reports | (2021) 11:16390 | https://doi.org/10.1038/s41598-021-95793-y 4

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol:.(1234567890)
www.nature.com/scientificreports/

a
1.8 V DC = 4.145 V

1.6 V DC = 4.205 V

V DC = 4.240 V
1.4

Normalized Current [-]

V DC = 4.280 V
1.2
V DC = 4.335 V
1.0 V DC = 4.360 V

0.8 V DC = 4.410 V

V DC = 4.470 V
0.6
MOR
0.4
experiments
0.2

21.35 21.4 21.45 21.5 21.55 21.6 21.65

f [kHz]

b
21.60
f 1 (roll)
50 f2 /2 (spurious roll)
21.55

21.50

f [kHz]
21.45
0
Phase [°]

21.40

21.35
4.1 4.2 4.3 4.4 4.5
-50 VDC [V]

-100

21.35 21.4 21.45 21.5 21.55 21.6 21.65

f [kHz]

Figure 3.  Frequency responses of the MEMS gyroscope test-structure for V ­ AC = 3.16 mV and different levels
of ­VDC in terms of (a) amplitude and (b) phase. Numerical predictions are reported in continuous lines while
experiments are in dotted lines. Stars of the same color of the curves mark the Saddle–Node bifurcations
estimated through the ROM model, while black stars mark the Neimark–Sacker bifurcations that delimit the
quasi-periodic region of the curves at ­VDC = 4.28 V. The resonant frequency of the roll mode (f1) and half of the
resonant frequency of the spurious roll mode (f2/2) are reported in the inset for different ­VDC.

compared with half the resonant frequency of the spurious roll mode, highlighting the strong link between the
nonlinear dynamic behavior of the structure under study and the ratio between the resonant frequencies of the
two coupled modes.
A good agreement in terms of amplitude and phase is found for all the DC-voltage levels, thus proving the
predictive ability of the simulation tool. It should be recalled that experiments are run in frequency control and
frequency upward sweep. As a consequence they cannot follow unstable branches and display jumps in proximity
of the Saddle–Node bifurcations predicted by the MOR (see Fig. 2a of Supplementary Information). It is worth
noting that the proposed ROM is able to catch the nonlinear dynamic response of the MEMS gyroscope test
structure under different actuating conditions without any need of ad-hoc calibration of the coefficients. This
make this simulation approach extremily versatile and general.

Discussion
The proposed ROM based on implicit static condensation is able to accurately and ab-initio reproduce the
complex nonlinear dynamics of a MEMS gyroscope test-structure undergoing 1:2 internal resonance including
bifurcations of the periodic response. The obtained two degrees of freedom model accounts for the multi-physics

Scientific Reports | (2021) 11:16390 | https://doi.org/10.1038/s41598-021-95793-y 5

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol.:(0123456789)
 www.nature.com/scientificreports/

Transfer function:
Vout / Vac

AC voltage amplitude

(MEMS gyroscope)
(closed-loop control)

(drive electrodes)
DC voltage: Vdc

AC voltage: Vac
b DC power supplier c Network Analyzer e Computer

Amplified output
voltage: Vout
Output current
(sense electrodes)

a MEMS (SEM image) + PCB d Amplifier

Figure 4.  Set-up employed to measure experimental frequency responses.

nature of the problem and does not require any calibration of the parameters: nonlinear coefficients come indeed
exclusively from numerical simulations and can be estimated without the need of experimental data. Moreover,
thanks to the reduction of the number of degrees of freedom of the system, simulations run almost real-time
and are thus very helpful for design purposes and experimental data post-processing.
This technique represents, to the authors’ best knowledge, the first tool able to estimate a priori and in real-
time the nonlinear dynamics of a complex multiphysics system like a MEMS gyroscope test-structure under
different actuation conditions.
MEMS designers and the MEMS industry in general, will strongly benefit of such tool since it will simplify
the understanding of experimental data and the design process of complex nonlinear MEMS devices.

Methods
Implicit static condensation. The implicit static condensation is based on the assumption that it is possi-
ble to describe the steady state non-linear oscillation of a resonator as a combination of few master modes (MM).
The dynamics of the ROM is described by a stress manifold obtained by implicitely condensing the effects of
higher order modes that locally modify the internal forces and thus the global stiffness of the s­ ystem42. For the
case under study where two modes interact through the internal resonance, this method allows to formulate a
ROM where the active degrees of freedom are the modal coordinates qi of the roll and the spurious roll modes
associated to the maximum out-of-plane displacement of the proof masses. Let ψi (x) denote the displacement
field of the i-th MM, mass normalized, the non-linear elastic force manifold is evaluated by statically forcing the
structure with suitable body forces F which are proportional to ψ1 (x) and ψ2 (x): F = β1 ψ1 (x) + β2 ψ2 (x). The
motivation for this choice, apart from simplicity, is that these loads are a very good approximation of inertia
forces occurring during the steady state oscillation. Once the body forces are defined, a series of static non-linear
analyses are run spanning the (β1 , β2 ) space. The range of the load-multipliers (β1 , β2 ) is prescribed so as to
cover the expected displacements of the structure, e.g. maximum out-of-plane displacements allowed by the
gap between the proof masses and the underlying substrate. Let (q1 (β1 , β2 ),q2 (β1 , β2 )) denote the solution for
a given (β1 , β2 ), we invert such relations and we obtain the terms (β1 (q1 , q2 ),β2 (q1 , q2 )) of Eqs. (1) and (2).
A similar procedure is adopted to determine the electrostatic nonlinear manifold of the ROM. This represents
a quasi-static approach which assumes that the dynamics of electromagnetic forces is much faster than the fre-
quency of oscillation of the resonators, which is verified in the case of the MEMS under consideration. We then
suppose that the gyroscope vibrates according to a combination of the two main modes, i.e. roll and spurious roll
modes, and we update the coordinates of the conductor surfaces, i.e. surfaces of the proof masses of the gyroscope
test-structure that face the underlying electrodes employed for actuation/readout, as x + ψ1 q1 + ψ2 q2, being x
the initial position of the conductor surfaces. The map of the charge surface density σ (x, q1 , q2 ) caused by the
interaction between the conductor surfaces with the underlying electrodes, is then computed as a function of
(q1 , q2 ) through integral equations accelerated with fast multipole ­methods59. Once the charge surface density is
available, the nonlinear load participation factor is computed as:
 2
σ
Fei = 2
ψni dS ≈ ǫ0 F̌e1i VDC , (3)
S 2

Scientific Reports | (2021) 11:16390 | https://doi.org/10.1038/s41598-021-95793-y 6

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol:.(1234567890)
www.nature.com/scientificreports/

where ψni = ψi · n is the projection of the modal shape function ψi along the outward unit normal vector on
the conductor surface and S is the surface portion of the proof masses that faces the underlying electrodes. An
analogous ­procedure55 allows us to determine the nonlinear amplitude of the forcing term F̌e21 (q1 , q2 ) in Eqs.
(1) and (2).

Experimental set‑up. The MEMS is bonded to a ceramic carrier and then connected to a Plastic Circuit
Board (PCB) as shown in Fig. 4a. Electrostatic actuation of the roll mode is provided through two power sup-
pliers (Fig. 4b,c): the Agilent E3631A provides the DC voltage while AC signal is generated through the Agi-
lent AG4395A. The output current measured on the electrodes on the substrate is amplified through the Signal
amplifier SRS model SR570 (Fig. 4d) and read in the frequency domain through the Agilent AG4395A (Fig. 4c).
A LabView script (Fig. 4e) acquires the output and corrects the AC signal to guarantee a close-loop control of
the circuit.

Received: 19 May 2021; Accepted: 28 July 2021

References
1. Comi, C. et al. Non-linear mechanics in resonant inertial micro sensors. Int. J. Non-Linear Mech. 120, 103386 (2020).
2. Shoshani, O. et al. Resonant modal interactions in micro/nano-mechanical structures. Nonlinear Dyn. 104, 1801–1828 (2021).
3. Nitzan, S. H. et al. Self-induced parametric amplification arising from nonlinear elastic coupling in a micromechanical resonating
disk gyroscope. Sci. Rep. 5, 9036 (2015).
4. Polunin, P. et al. Self-induced parametric amplification in ring resonating gyroscopes. Int. J. Non-Linear Mech. 94, 300–308 (2017).
5. Nitzan, S. H. et al. Countering the effects of nonlinearity in rate-integrating gyroscopes. IEEE Sensors J. 16(10), 3556–3563 (2016).
6. Nabholz, U. et al. Nonlinear dynamical system model for drive mode amplitude instabilities in MEMS gyroscopes. 2019 IEEE Int.
Symposium on Inertial Sensors and Syst. (INERTIAL), Naples, FL, USA, 1–4 (2019).
7. Zega, V. et al. Hardening, softening, and linear behavior of elastic beams in MEMS: An analytical approach. J. Microelectromech.
Syst. 28(2), 189–198 (2019).
8. Gobat, G. et al. Interpolation based reduced order modelling for non-linearities in MEMS. IEEE Sensors 2020, 1-4 (2020).
9. Shaw, A. D. et al. Periodic responses of a structure with 3:1 internal resonance. Mech. Syst. Signal Process. 81, 19–34 (2016).
10. Alfosail, F. K. et al. Theoretical and experimental investigation of two-to-one internal resonance in MEMS arch resonators. J.
Comput. Nonlinear Dyn. 14(1), 011001 (2018).
11. Qalandar, K. R. et al. Frequency division using a micromechanical resonance cascade. Appl. Phys. Lett. 105, 244103 (2014).
12. Sarrafan, A. et al. Development and characterization of an H-shaped microresonator exhibiting 2:1 internal resonance. J. Micro-
electromech. Syst. 26(5), 993–1001 (2017).
13. Antonio, D. et al. Frequency stabilization in nonlinear micromechanical oscillators. Nat. Commun. 8, 15523 (2017).
14. Chen, C. et al. Direct observation of coherent energy transfer in nonlinear micromechanical oscillators. Nat. Commun. 8, 15523
(2017).
15. Flader, I. B. et al. Tunable quality factor through 1:1 modal coupling in a disk resonator, 2015 IEEE SENSORS, Busan, Korea
(South), 1–4 (2015).
16. Keskekler, A. et al. Tuning nonlinear damping in graphene nanoresonators by parametric-direct internal resonance. Nat. Comm.
12, 1099 (2021).
17. Sarrafan, A. et al. A nonlinear rate microsensor utilising internal resonance. Sci. Rep. 9, 8648 (2019).
18. Hajjaj, A. Z. et al. Mode coupling and nonlinear resonances of MEMS arch resonators for bandpass filters. Sci. Rep. 7, 41820 (2017).
19. Nayfeh, A. H. Nonlinear Interactions: Analytical, Computational and Experimental Methods (Wiley, 2000).
20. Karabalin, R. B. et al. Nonlinear dynamics and chaos in two coupled nanomechanical resonators. Phys. Rev. B 79(16), 165309
(2009).
21. Ruzziconi, L. et al. Two-to-one internal resonance in the higher-order modes of a MEMS beam: Experimental investigation and
theoretical analysis via local stability theory. Int. J. Non-Linear Mech. 129, 103664 (2021).
22. Vyas, A. et al. A microresonator design based on nonlinear 1:2 internal resonance in flexural structural modes. J. Microelectromech.
Syst. 18(3), 744–762 (2009).
23. Czaplewski, D. A. Bifurcation diagram and dynamic response of a MEMS resonator with a 1:3 internal resonance. Appl. Phys. Lett.
114, 254104 (2019).
24. Sarrafan, A. et al. Analytical modeling and experimental verification of nonlinear mode coupling in a decoupled tuning fork
microresonator. J. Microelectromech. Syst. 27(3), 398–406 (2018).
25. Hajjaj, A. Z. et al. Multiple internal resonance in MEMS arch resonators. Phys. Lett. A 382, 3393–3398 (2018).
26. Hajjaj, A. et al. Two-to-one internal resonance of MEMS arch resonators. Int. J. Non-Linear Mech. 107, 64–72 (2018).
27. Bi, Q. et al. Analysis of non-linear dynamics and bifurcations of a shallow arch subjected to periodic excitation with internal
resonance. J. Sound Vib. 233(4), 553–567 (2020).
28. Nabholz, U. et al. Spontaneous parametric down-conversion induced by non-degenerate three-wave mixing in a scanning MEMS
micro mirror. Sci. Rep. 9, 3997 (2019).
29. Younis, M. et al. A reduced-order model for electrically actuated microbeam-based MEMS. J. Microelectromech. Syst. 12(5), 672–680
(2003).
30. Putnik, M. et al. Predicting the resonance frequencies in geometric nonlinear actuated MEMS. J. Microelectromech. Syst. 27(6),
954–962 (2018).
31. Kerschen, G. et al. Nonlinear normal modes, Part I: A useful framework for the structural dynamicist. Mech. Syst. Signal Process.
23(1), 170–194 (2009).
32. Renson, L. et al. Numerical computation of nonlinear normal modes in mechanical engineering. J. Sound Vib. 364, 177–206 (2016).
33. Mignolet, M. P. et al. A review of indirect/non-intrusive reduced order modeling of nonlinear geometric structures. J. Sound Vib.
332(10), 2437–2460 (2013).
34. Besselink, B. et al. A comparison of model reduction techniques from structural dynamics, numerical mathematics and systems
and control. J. Sound Vib. 332(19), 4403–4422 (2013).
35. Negri, F. et al. Efficient model reduction of parametrized systems by matrix discrete empirical interpolation. J. Comput. Phys. 303,
431–454 (2015).
36. Muravyov, A. A. et al. Determination of nonlinear stiffness with application to random vibration of geometrically nonlinear
structures. Comput. Struct. 81(15), 1513–1523 (2003).

Scientific Reports | (2021) 11:16390 | https://doi.org/10.1038/s41598-021-95793-y 7

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol.:(0123456789)
 www.nature.com/scientificreports/

37. Vizzaccaro, A. et al. Non-intrusive reduced order modelling for the dynamics of geometrically nonlinear flat structures using
three-dimensional finite elements. Comput. Mech. 66, 1293–1319 (2020).
38. Kerschen, G. et al. The method of proper orthogonal decomposition for dynamical characterization and order reduction of
mechanical systems: An overview. Nonlinear Dyn. 41(1–3), 147–169 (2005).
39. Hollkamp, J. J. et al. Reduced-order models for non-linear response prediction: Implicit condensation and expansion. J. Sound
Vib. 318, 1139–1153 (2008).
40. Nicolaidou, E. et al. Indirect reduced-order modelling: Using nonlinear manifolds to conserve kinetic energy. Proc. R. Soc. A. 476,
20200589 (2021).
41. Frangi, A. et al. Reduced order modelling of the non-linear stiffness in MEMS resonators. Int. J. Non-Linear Mech. 116, 211–218
(2019).
42. Idelsohn, S. R. et al. A reduction method for nonlinear structural dynamic analysis. Comput. Methods Appl. Mech. Eng. 49(3),
253–279 (1985).
43. Weeger, O. et al. On the use of modal derivatives for nonlinear model order reduction. Int. J. Numer. Methods Eng. 108(13),
1579–1602 (2016).
44. Jain, S. et al. A quadratic manifold for model order reduction of nonlinear structural dynamics. Comput. Struct. 188, 80–94 (2017).
45. Vizzaccaro, A. et al. Comparison of nonlinear mappings for reduced-order modelling of vibrating structures: Normal form theory
and quadratic manifold method with modal derivatives. Nonlinear Dyn. 103, 3335–3370 (2021).
46. Shen, Y. et al. Reduced order models for geometrically nonlinear structures: Assessment of implicit condensation in comparison
with invariant manifold approach. Eur. J. Mech. A Solids 86, 104165 (2021).
47. Touzé, C. Normal form theory and nonlinear normal modes: Theoretical settings and applications. Modal Analysis of Nonlinear
Mechanical Systems 75–160 (Springer, 2014).
48. Touzé, C. et al. Nonlinear normal modes for damped geometrically nonlinear systems: Application to reduced-order modelling
of harmonically forced structures. J. Sound Vib. 298(4–5), 958–981 (2006).
49. Jiang, D. et al. Nonlinear normal modes for vibratory systems under harmonic excitation. J. Sound Vib. 288(4–5), 791–812 (2005).
50. Shaw, S. W. An invariant manifold approach to nonlinear normal modes of oscillation. J. Nonlinear Sci. 4, 419–448 (1994).
51. Haller, G. et al. Nonlinear normal modes and spectral submanifolds: Existence, uniqueness and use in model reduction. Nonlinear
Dyn. 86, 1493–1534 (2016).
52. Ponsioen, S. et al. Automated computation of autonomous spectral submanifolds for nonlinear modal analysis. J. Sound Vib. 420,
269–295 (2018).
53. Vizzaccaro, A. et al. Direct computation of nonlinear mapping via normal form for reduced-order models of finite element non-
linear structures. Comput. Methods Appl. Mech. Eng. 384, 113957 (2021).
54. Opreni, A. et al. Model order reduction based on direct normal form: Application to large finite element MEMS structures featur-
ing internal resonance. Nonlinear Dyn., arXiv preprint arXiv:​2103.​10545 (2021).
55. Zega, V. et al. Numerical modelling of non-linearities in MEMS resonators. J. Microelectromech. Syst. 29(6), 1443–1454 (2020).
56. Fedeli, P. et al. Near vacuum gas damping in MEMS: Simplified modeling. J. Microelectromech. Syst. 26(3), 632–642 (2017).
57. Guillot, L. et al. A Taylor series-based continuation method for solutions of dynamical. Syst. Nonlinear Dyn. 98(4), 2827–2845
(2019).
58. Gobat, G. et al. Backbone curves, neimark-sacker boundaries and appearance of quasi-periodicity in nonlinear oscillators: Applica-
tion to 1:2 internal resonance and frequency combs in MEMS. Meccanica https://​doi.​org/​10.​1007/​s11012-​021-​01351-1 (2021).
59. Frangi, A. et al. Multipole BEM for the evaluation of damping forces on MEMS. Comput. Mech. 37(1), 24–31 (2005).

Author contributions
G.G., V.Z., C.T. and A.F. coinceived the idea and developed the supporting theory. G.G. performed numerical
analyses, L.G. and P.F. perform experimental tests. V.Z., C.T. and A.F. oversaw the research, provide guidance
and discussed the results and implications at all stages. V.Z. wrote the manuscript and all authors edited the
manuscript.

Competing interests
The authors declare no competing interests.

Additional information
Supplementary Information The online version contains supplementary material available at https://​doi.​org/​
10.​1038/​s41598-​021-​95793-y.
Correspondence and requests for materials should be addressed to V.Z.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons licence, and indicate if changes were made. The images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.

© The Author(s) 2021

Scientific Reports | (2021) 11:16390 | https://doi.org/10.1038/s41598-021-95793-y 8

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol:.(1234567890)
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:

1. use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
2. use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
3. falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
4. use bots or other automated methods to access the content or redirect messages
5. override any security feature or exclusionary protocol; or
6. share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at

[email protected]