# Array of Resonant Electromechanical Nanosystems: A Technological Breakthrough for Uncooled Infrared Imaging

^{1}

^{2}

^{*}

## Abstract

**:**

## 1. Introduction

^{−24}g—mass of a proton) has even been demonstrated with single-wall carbon nanotubes (CNT) [11,12]. More generally, nanoresonators can extend proteomics to high mass biomolecules (such as complexes of proteins or viruses) [3,13]. There is also intense activity in the NEMS community related to the study of oscillators in their fundamental quantum mode using the reciprocal interaction of an optical micro cavity with a mechanical resonator [14,15,16]. Examples of mechanical resonator applications are plethoric: particle counting in a fluid medium [17,18], magnetometry [19,20,21,22,23], actuators [24], RF filters [25], and in biology [26,27].

## 2. Design and Fabrication of Electromechanical Resonator Arrays

_{0}~ 376 Ω), in order to obtain a direct absorption rate that was close to 50%. The λ/4-optical cavity (2 µm thick) between the aluminum–copper electrodes and the TiN layer allowed an absorption efficiency of 80% to be reached over the 8–14 µm wavelength range. Figure 2g shows the spectral absorption of such a cavity with this specific SiN/TiN/SiN/a-Si stack in this wavelength range.

_{B}is the bias voltage applied on the paddle (through the studs), and V

_{AC}is the sinusoidal polarization applied on the actuation electrode through the capacitance Ca. This signal can be applied with an external RF-source, in particular for the first electromechanical characterizations, but can come from the feedback loop in the case of a closed loop. The actuation frequency f is swept to measure the electromechanical response and the resonance frequency f

_{0}.

_{inc}into a resonance frequency shift ∆f according to a sensitivity Rf, which depends on both the thermal conductance of the paddle insulation (through insulation legs between the torsional rods and the plate) and the temperature coefficient of frequency:

_{th}= C/G is the thermal time constant of the sensor, $C={\left(\frac{\partial U}{\partial T}\right)}_{V}$, the thermal capacitance at constant volume, $G$ is the thermal conductance, ${\alpha}_{T}$ is the temperature coefficient of frequency (TCF) (typically −60 ppm/°C for silicon), β is the pixel fill factor, η is the bolometer absorption, f

_{0}is the resonance frequency, and v the frame rate of the electronic readout. The thermal conductance is mainly due to the thermal conductance of heat through the legs. The other sources of thermal leaks—radiative and heat conductance through air—are negligible.

_{0}and θ

_{max}are respectively the capacitance value at rest, and the maximum deflection angle. The deflection angle is directly computed from the dynamic equation:

_{e}is the electrostatic torque. J is the moment of inertia of the paddle, assuming that the inertia moment of rods is negligible. k, G, and I

_{r}are respectively the rod torsional stiffness, the shear modulus, and the torsional quadratic moment of the rectangular suspended rods (${I}_{r}={w}_{r}{t}_{r}^{3}\left(\frac{1}{3}-0.21\frac{{t}_{r}}{{w}_{r}}\left(1-\frac{1}{12}\frac{{t}_{r}^{4}}{{w}_{r}^{4}}\right)\right),{w}_{r}{t}_{r}$.). E and v are the equivalent Young modulus and Poisson ratio of the stack.

^{6}. This attenuation of the signal can be deleterious for obtaining a high-enough SBR to initiate a self-oscillation within a closed-loop. A way to address this issue can be through the use of a semitone-actuation (at $\raisebox{1ex}{${f}_{0}$}\!\left/ \!\raisebox{-1ex}{$2$}\right.$): ${V}_{pol}={V}_{AC}\mathrm{cos}\left(\frac{2\pi ft}{2}\right)-{V}_{B}$. In this case, the electrostatic torque is proportional to ${V}_{AC}^{2}/2$, which reduces the coupling between the actuation signal and the output signal. A differential measurement can also be added to further improve the SBR. In this scheme, two identical pixels are used to cancel out the common modes. A more complex approach based on the down-mixing method [45] can be used to get rid of the parasitic capacitances. In particular, the bias voltage is no more constant and is modulated: ${V}_{B}={V}_{B0}\mathrm{cos}\left(2\pi ft+\Delta f\right)$ where $\Delta f\ll f$. A comparison between different readout modes is shown in Table 2. We notice a quite strong improvement of the SBR. However, in the best cases, the signal-to noise-ratio (SNR) was lower than 20 dB, which did not guarantee a functional closed-loop.

## 3. Results

#### 3.1. Electromechanical Characterizations

- ${f}_{0}=\left[1.05\u20131.2\mathrm{MHz}\right]$;
- $Q=\left[1600\u20132500\right]$;
- ${V}_{out}=\left[100\u2013350\mathrm{\mu}\mathrm{V}\right]$

#### 3.2. Thermal Characterizations

#### 3.2.1. TCF & G

#### 3.2.2. Thermal Response

_{f}= 1050 W

^{−1}were extracted with the best devices (${f}_{0}=1.15\mathrm{MHz})$. Assuming a fill factor β = 0.8, an efficiency η = 0.8 in the 8–12 µm window, and considering the measured TCF, α

_{T}= −76 ppm/°C, a theoretical thermal response R

_{f}= 950 W

^{−1}was expected, which was very close to the observed sensitivities. In a second experiment, the incident IR flux on a pixel was changed by varying the distance between the window and the IR-source. The frequency shift was then the measured for optical powers varying from 2.5 to 16 nW. The experimental results and their linear fit are presented in Figure 8b. A thermal response of R

_{f}=1350 W

^{−1}was extracted from the slope, considering the resonance frequency mentioned above. Above 8 nW, the relationship between the IR-flux and the thermal frequency shift was no more linear. To increase the incident power, the source was moved closer to the window, which caused it to heat up. This effect lowered its transmittance, resulting in an incorrect estimation of the thermal response (namely, R

_{f}=1050 W

^{−1}).

_{f}varied from 700 to 1350 W

^{−1}, showing some dispersion attributed to the fabrication process.

#### 3.2.3. Response Time

#### 3.3. Noises and Temperature Sensitivity

## 4. Discussion

^{2}/sr/K [53].

- Frequency stability and matrix readout strategy: A 50 Hz integration bandwidth requires an improvement of the noise amplitude of our buffer electronics close to the pixel. Lower amplitude noise levels can be reached by using self-oscillating electronics requiring only a few transistors, unlike PLL circuits. Moreover, our electronics was realized close to the pixel but this was not done through an application-specific integrated circuit (ASIC) fabricated underneath the electromechanical pixels. The low-temperature fabrication process presented above has already been used to manufacture resistive bolometer imagers on top of CMOS circuits (ROIC) by post-processing [37,55], and it should be straightforward to reuse this approach in our case. As mentioned in the introductive section of the paper, a co-integration of the readout electronics at the pixel level will reduce the parasitic capacitance down to a few fF, and will decrease the electrical noise down to a theoretical level of $10\mathrm{nV}/\sqrt{\mathrm{Hz}}$, or even $5\mathrm{nV}/\sqrt{\mathrm{Hz}}$. This approach makes the down-mixing detection scheme unnecessary, leading to a much simpler measurement chain than the strategy presented here. ${\sigma}_{y}$ will be decreased by a factor of 8 with a self-oscillating IC (a gain of a factor 4 on the absolute noise, and a gain of a factor 2 on the output voltage with a the direct detection (see Equation (9))). Thus, the electronics noise will become lower than the fundamental APN (${\sigma}_{y}=1.5\times {10}^{-7}$) for a 700 Hz integration bandwidth. This conclusion leads us to suggest a new readout scheme consisting of reading 700/50 = 14 pixels during a 50 Hz frame rate, which allows for a larger area for the co-integrated readout. These two straightforward improvements allow us to obtain a FOM that is close to 0.75 for a ${f}_{BW}=50\mathrm{Hz}$ (global shutter approach), which is an encouraging element.
- Thermal response: At the end, the noise floor level will be set by the APN, whatever the electronics and the readout strategy. An improvement of the signal through the thermal insulation $1/G$ is much trickier in our case. Indeed, this would require long and thin rods/insulations legs, and this would lower the onset of nonlinearity of ${\theta}_{c}$ (see Equation (8)), leading to a degradation of the SNRs and therefore the frequency stability ${\sigma}_{y}$.

_{2}materials, on top of our pixel. This material was deposited in its amorphous state by reactive deposition (Ion Beam Deposition) and annealed at 400 °C to obtain the crystalline state. The process temperature is kept low enough to be used in a post-process of a CMOS circuit. The Raman characterizations were done to verify the crystallization obtained with this method. The resonators were designed to keep the mechanical features of our current typical pixel (Figure 3c). Nano-indentation measurements were performed on a full layer to extract the Young’s modulus of our VO

_{2}layer (177 GPa for the crystalline state and 80 GPa for the amorphous state). A thickness of 80 nm was then chosen with 1.5 µm long and 300 nm wide torsional rods. In an initial version, both the rods and the plate are covered with the VO

_{2}layer, and in a second version, the VO

_{2}layer is only left on the rods. An example of the fabricated devices is shown in Figure 11. The TCF measurements and then the frequency stability are ongoing. We expect an improvement of one order of magnitude in the thermal response (the mechanical features and the thermal insulation being kept constant compared to the standard pixel).

## 5. Conclusions

## Author Contributions

## Funding

## Acknowledgments

## Conflicts of Interest

## References

- Rugar, D.; Budakian, R.; Mamin, H.J.; Chui, B.W. Single spin detection by magnetic resonance force microscopy. Nature
**2004**, 430, 329. [Google Scholar] [CrossRef] [PubMed] - Arlett, J.L.; Maloney, J.R.; Gudlewski, B.; Muluneh, M.; Roukes, M.L. Self-sensing micro- and nanocantilevers with attonewton-scale force resolution. Nano Lett.
**2006**, 6, 1000–1006. [Google Scholar] [CrossRef] - Sage, E.; Brenac, A.; Alava, T.; Morel, R.; Dupré, C.; Hanay, M.S.; Roukes, L.M.; Duraffourg, L.; Masselon, C.; Hentz, S. Neutral particle mass spectrometry with nanomechanical systems. Nat. Commun.
**2015**, 6, 6482. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Hanay, M.S.; Kelber, S.; Naik, A.K.; Chi, D.; Hentz, S.; Bullard, E.C.; Colinet, L.; Duraffourg, E.; Roukes, M.L. Single-protein nanomechanical mass spectrometry in real time. Nat. Nanotechnol.
**2012**, 7, 602–608. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Bargatin, I.; Myers, E.B.; Aldridge, J.S.; Marcoux, C.; Brianceau, P.; Duraffourg, L.; Colinet, E.; Hentz, S.; Andreucci, S.; Roukes, M.L. Large-scale integration of nanoelectromechanical systems for gas sensing applications. Nano Lett.
**2012**, 12, 1269–1274. [Google Scholar] [CrossRef] [PubMed] - Fahad, H.M.; Shiraki, H.; Amani, M.; Zhang, C.; Hebbar, V.S.; Gao, W.; Ota, H.; Hettick, M.; Kiriya, D.; Chen, Y.-Z.; et al. Room temperature multiplexed gas sensing using chemical-sensitive 3.5-nm-thin silicon transistors. Sci. Adv.
**2017**, 3, e1602557. [Google Scholar] [CrossRef] [PubMed] - Ralph, J.E.; King, R.C.; Curran, J.E.; Page, J.S. Miniature quartz resonator thermal detector. In Proceedings of the IEEE Ultrasonics Symposium, San Francisco, CA, USA, 16–18 October 1985; p. 362. [Google Scholar]
- Cabuz, C.; Shoji, S.; Fukatsu, K.; Cabuz, E.; Minami, K.; Esashi, M. Fabrication and packaging of a resonant infrared sensor integrated in silicon. Sens. Actuators A
**1994**, 43, 92–99. [Google Scholar] [CrossRef] - Zhang, X.C.; Myers, E.B.; Sader, J.E.; Roukes, M.L. Nanomechanical torsional resonators for frequency-shift infrared thermal sensing. Nano Lett.
**2013**, 13, 1528–1534. [Google Scholar] [CrossRef] [PubMed] - Yamazaki, T.; Ogawa, S.; Kumagai, S.; Sasaki, M. A novel infrared detector using highly nonlinear twisting vibration. Sens. Actuators A
**2014**, 212, 165–172. [Google Scholar] [CrossRef] [Green Version] - Jensen, K.; Kim, K.; Zettl, A. An atomic-resolution nanomechanical mass sensor. Nat. Nanotechnol.
**2008**, 3, 533–537. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Chaste, J.; Eichler, A.; Moser, J.; Ceballos, G.; Rurali, R.; Bachtold, A. A nanomechanical mass sensor with yoctogram resolution. Nat. Nanotechnol.
**2012**, 7, 301–304. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Ruz, J.J.; Tamayo, J.; Pini, V.; Kosaka, P.M.; Calleja, M. Physics of nanomechanical spectrometry of viruses. Sci. Rep.
**2014**, 4, 6051. [Google Scholar] [CrossRef] [PubMed] - Aspelmeyer, M.; Kippenberg, T.J.; Marquardt, F. Cavity optomechanics. Rev. Mod. Phys.
**2014**, 86, 1391. [Google Scholar] [CrossRef] - Metcalfe, M. Applications of cavity optomechanics. Appl. Phys. Rev.
**2014**, 1, 031105. [Google Scholar] [CrossRef] - Li, B.B.; Bulla, D.; Bilek, J.; Prakash, V.; Forstner, S.; Sheridan, E.; Madsen, L.; Rubinsztein-Dunlop, H.; Foster, S.; Schäfermeier, C.; et al. Ultrasensitive and broadband magnetometry with cavity optomechanics. In Proceedings of the IEEE Conference on Lasers and Electro-Optics (CLEO), San Jose, CA, USA, 14–19 May 2017. [Google Scholar]
- Godin, M.; Bryan, A.K.; Burg, T.P.; Babcock, K.; Manalis, S.R. Measuring the mass, density, and size of particles and cells using a suspended microchannel resonator. Appl. Phys. Lett.
**2007**, 91, 123121. [Google Scholar] [CrossRef] - Agache, V.; Blanco-Gomez, G.; Baleras, F.; Caillat, P. An embedded microchannel in a MEMS plate resonator for ultrasensitive mass sensing in liquid. Lab Chip
**2011**, 11, 2598–2603. [Google Scholar] [CrossRef] [PubMed] - Min, D.H.; Moreland, J. Quantitative measurement of magnetic moments with a torsional resonator: Proposal for an ultralow moment reference material. J. Appl. Phys.
**2005**, 97, 10R504. [Google Scholar] [CrossRef] - Antonio, D.; Dolz, M.I.; Pastoriza, H. Micromechanical magnetometer using an all-silicon nonlinear torsional resonator. Appl. Phys. Lett.
**2009**, 95, 133505. [Google Scholar] [CrossRef] - Davis, J.P.; Vick, D.; Li, P.; Portillo, S.K.N.; Fraser, A.E.; Burgess, J.A.J.; Fortin, D.C.; Hiebert, W.K.; Freeman, M.R. Nanomechanical torsional resonator torque magnetometry. J. Appl. Phys.
**2011**, 109, 07D309. [Google Scholar] [CrossRef] - Perez-Morelo, D.; Pastoriza, H. Torque magnetometry in YBa
_{2}Cu_{3}O_{7}-δ single crystals using high sensitive micromechanical torsional oscillator. J. Phys. Conf. Ser.**2014**, 568, 22038. [Google Scholar] [CrossRef] - Degen, C.L.; Poggio, M.; Mamin, H.J.; Rettner, C.T.; Rugar, D. Nanoscale magnetic resonance imaging. Proc. Natl. Acad. Sci. USA
**2009**, 106, 1313–1317. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Loh, O.Y.; Espinosa, H.D. Nanoelectromechanical contact switches. Nat. Nanotechnol.
**2012**, 7, 283–295. [Google Scholar] [CrossRef] [PubMed] - Wong, A.-C.; Nguyen, C.T.C. Micromechanical mixer-filters (‘mixlers’). J. Microelectromech Syst.
**2004**, 13, 100–112. [Google Scholar] [CrossRef] - Arlett, J.L.; Myers, E.B.; Roukes, M.L. Comparative advantages of mechanical biosensors. Nat. Nanotechnol.
**2011**, 6, 203–215. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Tamayo, J.; Kosaka, P.M.; Ruz, J.J.; San Paulo, Á.; Calleja, M. Biosensors based on nanomechanical systems. Chem. Soc. Rev.
**2013**, 42, 1287–1311. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Duraffourg, L.; Arcamone, J. From MEMS to NEMS. In Nanoelectromechanical Systems, 1st ed.; Baptist, R., Duraffourg, L., Eds.; ISTE Ltd.: London, UK, 2015; Chapter 1; p. 5. ISBN 978-1-84821-669-3. [Google Scholar]
- Mile, E.; Jourdan, G.; Bargatin, I.; Labarthe, S.; Marcoux, C.; Andreucci, P.; Hentz, S.; Kharrat, C.; Colinet, E.; Duraffourg, L. In-plane nanoelectromechanical resonators based on silicon nanowire piezoresistive detection. Nanotechnology
**2010**, 21, 165504. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Colinet, E.; Duraffourg, L.; Labarthe, S.; Hentz, S.; Robert, P.; Andreucci, P. Self-oscillation conditions of a resonant nanoelectromechanical mass sensor. J. Appl. Phys.
**2009**, 105, 124908. [Google Scholar] [CrossRef] [Green Version] - Sage, E.; Sansa, M.; Fostner, S.; Defoort, M.; Gély, M.; Naik, A.K.; Morel, R.; Duraffourg, L.; Roukes, M.L.; Alava, T.; et al. Single-particle mass spectrometry with arrays of frequency-addressed nanomechanical resonators. arXiv, 2017arXiv:1712.06694. [CrossRef]
- Endoh, T.; Tohyama, S.; Yamazaki, T.; Tanaka, Y.; Okuyama, K.; Kurashina, S.; Miyoshi, M.; Katoh, K.; Yamamoto, T.; Okuda, Y.; et al. Uncooled infrared detector with 12 μm pixel pitch video graphics array. In Proceedings of the SPIE Conference on Defense, Security, and Sensing, Baltimore, MD, USA, 11 June 2013; Volume 8704, p. 87041G. [Google Scholar]
- FLIR Boson® | FLIR Systems. Available online: http://www.flir.com/cores/boson/ (accessed on 24 February 2017).
- TWV640 Thermal Camera Core | BAE Systems | International. Available online: http://www.baesystems.com/en/product/twv640-thermal-camera-core (accessed on 24 February 2017).
- Jo, Y.; Kwon, I.-W.; Kim, D.S.; Shim, H.B.; Lee, H.C. A self-protecting uncooled microbolometer structure for uncooled microbolometer. In Proceedings of the SPIE Conference on Defense, Security, and Sensing, Orlando, FL, USA, 25–29 April 2011; Volume 8012, p. 80121O. [Google Scholar]
- Dorn, D.A.; Herrera, O.; Tesdahl, C.; Shumard, E.; Wang, Y.-W. Impacts and mitigation strategies of sun exposure on uncooled microbolometer image sensors. In Proceedings of the SPIE Conference on Defense, Security, and Sensing, Orlando, FL, USA, 25–29 April 2011; Volume 8012, p. 80123Z. [Google Scholar]
- Tissot, J.L.; Durand, A.; Garret, T.; Minassian, C.; Robert, P.; Tinnes, S.; Vilain, M. High performance uncooled amorphous silicon VGA IRFPA with 17 μm pixel-pitch. In Proceedings of the SPIE Conference on Defense, Security, and Sensing, Orlando, FL, USA, 5–9 April 2010; Volume 7660, p. 76600T. [Google Scholar]
- Fraenkel, A.; Mizrahi, U.; Bikov, L.; Adin, A.; Malkinson, E.; Giladi, A.; Seter, D.; Kopolovich, Z. VOx-based uncooled microbolometric detectors: Recent developments at SCD. In Proceedings of the SPIE Conference on Defense, Security, and Sensing, Orlando, FL, USA, 17 May 2006; Volume 6206, p. 62061C. [Google Scholar]
- Truitt, P.A.; Hertzberg, J.B.; Huang, C.C.; Ekinci, K.L.; Schwab, K.C. Efficient and Sensitive Capacitive Readout of Nanomechanical Resonator Arrays. Nano Lett.
**2007**, 7, 120–126. [Google Scholar] [CrossRef] [PubMed] - Younis, M.I. MEMS Linear and Nonlinear Statics and Dynamics; Springer: Boston, MA, USA, 2011; Volume 20. [Google Scholar]
- Verbridge, S.S.; Parpia, J.M.; Reichenbach, R.B.; Bellan, L.M.; Craighead, H.G. High quality factor resonance at room temperature with nanostrings under high tensile stress. J. Appl. Phys.
**2006**, 99, 124304. [Google Scholar] [CrossRef] - Kacem, N.; Hentz, S.; Pinto, D.; Reig, B.; Nguyen, V. Nonlinear dynamics of nanomechanical beam resonators: Improving the performance of NEMS-based sensors. Nanotechnology
**2009**, 20, 275501. [Google Scholar] [CrossRef] [PubMed] - Kozinsky, I.; Postma, H.W.C.; Bargatin, I.; Roukes, M.L. Tuning nonlinearity, dynamic range, and frequency of nanomechanical resonators. Appl. Phys. Lett.
**2006**, 88, 253101. [Google Scholar] [CrossRef] [Green Version] - Analog Devices, Op Amp Total Output Noise Calculations for Second-Order System. Available online: http://www.analog.com/media/en/training-seminars/tutorials/MT-050.pdf (accessed on 1 June 2016).
- Bargatin, I.; Myers, E.B.; Arlett, J.; Gudlewski, B.; Roukes, M.L. Sensitive detection of nanomechanical motion using piezoresistive signal downmixing. Appl. Phys. Lett.
**2005**, 86, 133109. [Google Scholar] [CrossRef] [Green Version] - Ziegler, J.G.; Nichols, N.B. Optimum Settings for Automatic Controllers. J. Dyn. Syst. Meas. Control
**1993**, 115, 220–222. [Google Scholar] [CrossRef] - Laurent, L.; Yon, J.J.; Moulet, J.S.; Imperinetti, P.; Duraffourg, L. Compensation of nonlinear hardening effect in a nanoelectromechanical torsional resonator. Sensors Actuators A
**2017**, 263, 326–331. [Google Scholar] [CrossRef] - Wood, R.A. Chapter 3—Monolithic Silicon Microbolometer Arrays. In Uncooled Infrared Imaging Arrays and Systems; Semiconductors and Semimetals Series; Kruse, W., Skatrud, D.D., Eds.; Elsevier: New York, NY, USA, 1997; Volume 47, p. 43. [Google Scholar]
- Robins, W.P. Phase Noise in Signal Sources: Theory and Applications; IET: London, UK, 1984. [Google Scholar]
- Cleland, A.N.; Roukes, M.L. Noise processes in nanomechanical resonators. J. Appl. Phys.
**2002**, 92, 2758–2769. [Google Scholar] [CrossRef] [Green Version] - Allan, D.W. Time and Frequency (Time-Domain) Characterization, Estimation, and Prediction of Precision Clocks and Oscillators. IEEE Trans. Ultrason. Ferroelectr. Freq. Control
**1987**, 34, 647–654. [Google Scholar] [CrossRef] [PubMed] - Sansa, M.; Sage, E.; Bullard, E.C.; Gély, M.; Alava, T.; Colinet, E.; Naik, A.K.; Villanueva, L.G.; Duraffourg, L.; Roukes, M.L.; et al. Frequency fluctuations in silicon nanoresonators. Nat. Nanotechnol.
**2016**, 11, 552–558. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Kruse, P.W.; Skatrud, D.D. Uncooled Infrared Imaging Arrays and Systems; Academic Press: Cambridge, MA, USA, 1997; Volume 47. [Google Scholar]
- Skidmore, G.D.; Han, C.J.; Li, C. Uncooled microbolometers at DRS and elsewhere through 2013. In Proceedings of the SPIE Conference on Image Sensing Technologies: Materials, Devices, Systems, and Applications, Baltimore, MD, USA, 21 May 2014; Volume 9100, p. 910003. [Google Scholar]
- Tissot, J.-L.; Mottin, E.; Martin, J.-L.; Yon, J.-J.; Vilain, M. Advanced uncooled infrared focal plane development at CEA/LETI. In Proceedings of the International Conference on Space Optics, Rhodes Island, Greece, 4–8 October 2010; p. 1056902. [Google Scholar] [CrossRef]
- Schodowski, S.S. Resonator self-temperature-sensing using a dual-harmonic-mode crystal oscillator. In Proceedings of the 43rd Annual Symposium on Frequency Control, Denver, CO, USA, 31 May–2 June 1989; pp. 2–7. [Google Scholar] [CrossRef]
- Li, M.-H.; Chen, C.-Y.; Li, C.-S.; Chin, C.-H.; Li, S.-S. Design and Characterization of a Dual-Mode CMOS-MEMS Resonator for TCF Manipulation. J. Microelectromech. Syst.
**2015**, 24, 446–457. [Google Scholar] [CrossRef] - Holsteen, A.; Kim, I.S.; Lauhon, L.J. Extraordinary Dynamic Mechanical Response of Vanadium Dioxide Nanowires around the Insulator to Metal Phase Transition. Nano Lett.
**2014**, 14, 1898. [Google Scholar] [CrossRef] [PubMed] - Rúa, A.; Cabrera, R.; Coy, H.; Merced, E.; Sepúlveda, N.; Fernández, F.E. Phase transition behavior in microcantilevers coated with M1-phase VO
_{2}and M_{2}-phase VO_{2}:Cr thin films. J. Appl. Phys.**2012**, 111, 104502. [Google Scholar] [CrossRef]

**Figure 1.**The principle of mass measurement (adapted from [28]). (

**a**) Example of a nanoresonator on which particles have landed; (

**b**) Shift in the frequency caused by the arrival of particles. Monitoring in real-time the resonance frequency allows us to deduce the amount of accreted mass; (

**c**) From the spectral perspective: a shift in the spectrum toward low frequencies.

**Figure 2.**Synopsis of the fabrication process along the cross section AA’: (

**a**) Deposition of a 300 nm thick AlCu layer on a silicon substrate/strip lines/reflector wet etching; (

**b**) 2 µm deposition of a polyimide layer; (

**c**) Etching of the polyimide layer to build up the metal studs, and deposition of the plate material; (

**d**) Definition of the plate; (

**e**) Dry etching O

_{2}plasma release; (

**f**) 3D artist view of a pixel; (

**g**) Absorption spectrum of the 2 µm thick micro cavity.

**Figure 3.**Scanning electron microscopy (SEM) pictures of an array of electromechanical pixels fabricated with a low-temperature process: (

**a**) Large field view of an array; only the central 96 × 96 array is connected to electrical pads; pixels above the connection wires have been removed to avoid any cross-talk; (

**b**) Zoom-in on the center of the array; (

**c**) SEM picture of a typical H-shape pixel; Nanorod length = 1.5 µm, width = 250 nm and thickness = 180 nm (insulation arm length = 8.6 µm).

**Figure 4.**SEM pictures of alternative versions derived from the nominal design: (

**a**) Butterfly-shaped pixel with longer rods; (

**b**) Simple pixel without insulating legs; (

**c**) H-shape pixel with thinner nanorods for enhancing the thermal insulation; rod-thickness = 30 nm; (

**d**) Zoom of the legs (Figure 4c) attached to a stud that acts as a mechanical anchor and that provides electrical contact with the lines underneath.

**Figure 5.**Synopsis of the open-loop measurement chain: The red box corresponds to a single electromechanical pixel that translates the incident IR-radiation P

_{inc}on the scene into a resonance frequency shift; the blue box corresponds to the close-by electronics that convert the mechanical oscillations into an electrical signal; V

_{pol}is the polarization of the pixel; θ(f) is the angular oscillation of the paddle around the rods; Cd(f) is the induced capacitance variation used to read out the signal; V

_{out}(f) is the output signal supplied by the buffer. C

_{p}is the total capacitance due to amplifier input capacitance, and parasitic capacitances between the electrical connections and the ground.

**Figure 6.**Enhancement provided by the buffer circuit: (

**a**) Synopsis showing the set-up to characterize the pixels in a down-mixed readout scheme (in open loop or in a closed-loop: red part); (

**b**) Comparison of the output signal between a semitone down-mixing approach with the buffer circuit and without the buffer circuit (the polarization voltages are explained in Table 2); (

**c**) Typical output signals for the three down-mixing approaches with the buffer circuit. At resonance, the capacitance variation is close to 10 aF. The SNR is larger than 40 dB for the three cases.

**Figure 7.**Electromechanical response when the frequency is swept around the resonance for different actuation schemes: (

**a**) Amplitude versus $f$ and ${V}_{AC}\text{}$ for a semitone actuation $\left({V}_{B}=10\text{}\mathrm{V}\right)$; (

**b**) Amplitude versus $f$ and ${V}_{B}\text{}$ for a semitone actuation $\left({V}_{AC}=6.5\text{}\mathrm{V}\right)$; (

**c**) Amplitude versus $f$ and ${V}_{AC}\text{}$ for a 2f-actuation $\left({V}_{B}\text{}=10\text{}\mathrm{V}\right)$; (

**d**) Amplitude versus $f$ and ${V}_{B}\text{}$ for a 2f-actuation $\left({V}_{AC}=1.8\text{}\mathrm{V}\right)$.

**Figure 8.**Measurement of the thermal response: (

**a**) Resonance frequency shift induced by incident IR pulses (peaks of 17 nW); (

**b**) Frequency shift when the incident power is varied from 2 to 16 nW.

**Figure 9.**Resonance frequency according to the acquisition time. A 1 mW red laser is focalized onto the pixel under test (typical Figure 3c). Insert: Full data from our measurement of response time. The average frequency jump is estimated as 110 Hz. Then, the response time is extracted from one event fall time (red).

**Figure 10.**Noise characterization achieved on the typical electromechanical pixel—the amplitude at resonance is set at 320 µV: (

**a**) Allan deviation measurement; the red hexagon indicates the frequency deviation for ${f}_{BW}=7\text{}\mathrm{kHz}$ (${\sigma}_{A}=3.5\times {10}^{-6}$ ) and the red disk is the one for ${f}_{BW}=50\text{}\mathrm{Hz}$ (${\sigma}_{A}=3\times {10}^{-7})$; a plateau is reached between 50 and 200 ms integration time (${\sigma}_{A}=1.5\times {10}^{-7}$ ); beyond 200 ms, a strong drift effect can be observed; (

**b**) spectral power density measurement achieved on the same device. The Allan deviation has a ${\tau}^{-\frac{1}{2}}\text{}$ stop at a short integration time corresponding to the signature of a White amplitude noise. Beyond 50 ms integration time, the Allan deviation presents a plateau, which shows a 1/f frequency noise. These two noises can also be distinguished on the power spectral density: the slope of $-10\text{}\mathrm{dB}/\mathrm{decade}$ corresponds to this plateau.

**Figure 11.**Torsional resonator design: (

**a**) Schematics of the design; (

**b**) SEM image of a typical pixel with a VO

_{2}layer on top of both the plate and torsional rods, plus partially on the insulation legs.

**Table 1.**Key parameters presented in the Equations (1)–(5) for our device compared with an advanced resistive bolometer and microelectromechanical systems (MEMS) bolometer: temperature sensitivity corresponds to $\frac{1}{f}\times \partial f\u2044\partial T$ for a resonant thermal sensor and $\frac{1}{R}\times \partial R\u2044\partial T$ for a resistive one.

Electromechanical & Thermal Features | This Work (Figure 3c) | Bolometer [32] | Resonant MEMS [9] |
---|---|---|---|

Maximal Angle ${\theta}_{max}$ (°) | 21 | $N.A$ | - |

Inertial Moment $J$ (kg.m^{2}) | $3.9\times {10}^{-25}$ | $N.A$ | $1.5\times {10}^{-27}$ |

Torsional stiffness $\kappa $ (N.m) | $1.8\times {10}^{-11}$ | $N.A$ | $6.8\times {10}^{-13}$ |

Resonant Frequency (MHz) | 1.1 | $N.A$ | - |

Onset of Nonlinearity ${\theta}_{c}$(°) (This value is computed by solving a nonlinear dynamic equation [44].) | 13.5 | $N.A$ | - |

Quality Factor Q | $1800$ | $N.A$ | $1555$ |

Capacitance at Rest ${C}_{0}$ (fF) | 0.185 | $N.A$ | $N.A$ |

Pitch (µm) | $12$ | $12$ | $5$ |

Thermal Conductance G (W/K) | $5\times {10}^{-8}$ | $5\times {10}^{-9}$ | $1.5\times {10}^{-8}$ |

Thermal Capacity C (J/K) | $26\times {10}^{-12}$ | $80\times {10}^{-12}$ | $3\times {10}^{-12}$ |

Thermal Constant ${\tau}_{th}$ (ms) | $0.5$ | $16$ | $0.2$ |

Temperature Sensitivity (/°C) | $0.01\%$ | $3.6\%$ | $0.0092\%$ |

**Table 2.**Preliminary measurement of the signal-to-background ratio (SBR) for different transduction strategies (direct semitone, direct 1f, differential and down-mixed); ${V}_{B0}=10\mathrm{V}$ , ${V}_{DC}=10\mathrm{V}$ and ${f}_{0}=1\mathrm{MHz},\Delta f=10\mathrm{kHz}$; ${V}_{AC0}=4.2\mathrm{V}$ for semitone actuation and ${V}_{AC0}=0.5\text{}\mathrm{V}$ for 1f and 2f actuations.

Transduction Method | Voltages | SBR (dB) | |
---|---|---|---|

- | ${V}_{AC}$ | ${V}_{B}$ | - |

$1f$-actuation | ${V}_{AC0}\mathrm{cos}\left(2\pi ft\right)$ | ${V}_{B0}$ | −33 |

$f/2$-actuation | ${V}_{AC0}\mathrm{cos}\left(\frac{2\pi ft}{2}\right)$ | ${V}_{B0}$ | −13 |

$f/2$-actuation/differential mode | ${V}_{AC0}\mathrm{cos}\left(\frac{2\pi ft}{2}\right)$ | ${V}_{B0}$ | 2 |

$f/2$-actuation/down-mixing mode | ${V}_{AC0}\mathrm{cos}\left(\frac{2\pi ft}{2}\right)$ | ${V}_{B0}\mathrm{cos}\left(2\pi ft+\Delta f\right)$ | 22 |

$f$-actuation/down-mixing mode | ${V}_{AC0}\mathrm{cos}\left(2\pi ft\right)+{V}_{DC}$ | ${V}_{B0}\mathrm{cos}\left(2\pi ft+\Delta f\right)$ | 20 |

$2f$-actuation/down-mixing mode | ${V}_{AC0}\mathrm{cos}\left(4\pi ft+\Delta f\right)$ | ${V}_{B0}\mathrm{cos}\left(2\pi ft+\Delta f\right)$ | 22 |

**Table 3.**Temperature coefficient of frequency (TCF) measured on different types of pixels: mean and standard deviation per wafer; thermal conductance of rods & legs $G$ (computed from material properties and geometry measured by SEM).

Pixel Types | $\langle {\mathit{\alpha}}_{\mathit{T}}\rangle \text{}(\mathbf{ppm}/\xb0\mathbf{C})$ | ${\mathit{\sigma}}_{{\mathit{\alpha}}_{\mathit{T}}}$ | $\mathit{G}$$(\mathit{W}/\mathit{K})$ | $\langle {\mathit{\alpha}}_{\mathit{T}}\rangle /\mathit{G}$ |
---|---|---|---|---|

Typical (Figure 3c) | 55.4 | 14.6 | $5\times {10}^{-8}$ | $1.11\times {10}^{9}$ |

Butterfly (Figure 4a) | 45.2 | 3.6 | $3.10\times {10}^{-8}$ | $1.46\times {10}^{9}$ |

Typical with Thin Nano-Rod (Figure 4c) | 86.2 | 16.4 | $1.8\times {10}^{-8}$ | $4.79\times {10}^{9}$ |

**Table 4.**Theoretical frequency stability and NEP for the nominal pixel with our electronics and 2f-down-mixing readout scheme: ${R}_{f}=1050/\mathrm{W}$, ${\theta}_{C}=13.5\xb0$, ${V}_{n}=10\mathrm{nV}/\sqrt{\mathrm{Hz}}$, $Q=2500$ and ${V}_{out}=320\mathrm{\mu}\mathrm{V}$.

Noise Sources | ${\mathit{f}}_{\mathit{B}\mathit{W}}=50\mathbf{Hz}$ | ${\mathit{f}}_{\mathit{B}\mathit{W}}=7\mathbf{kHz}$ | ${\mathit{f}}_{\mathit{B}\mathit{W}}=1\mathbf{Hz}$ | ||||
---|---|---|---|---|---|---|---|

${\mathit{X}}_{\mathit{n}}$ | ${\mathit{\sigma}}_{\mathit{y}}$ | $\mathit{N}\mathit{E}\mathit{P}$ | ${\mathit{X}}_{\mathit{n}}$ | ${\mathit{\sigma}}_{\mathit{y}}$ | $\mathit{N}\mathit{E}\mathit{P}$ | $\mathit{N}\mathit{E}\mathit{P}$ | |

Thermodynamic | $3.6\times {10}^{-6}\text{}\mathrm{rad}$ | $6.3\times {10}^{-9}$ | $6\text{}\mathrm{pW}$ | $1.2\times {10}^{-5}$ rad | $2.6\times {10}^{-8}$ | $25\text{}\mathrm{pW}$ | $0.85\text{}\mathrm{pW}$ |

Electronics | $70.7\text{}\mathrm{nV}$ | $8.9\times {10}^{-8}$ | $85\text{}\mathrm{pW}$ | $836.7\text{}\mathrm{nV}$ | $1\times {10}^{-6}$ | $1000\text{}\mathrm{pW}$ | $12\text{}\mathrm{pW}$ |

Phonon | - | $5.8\times {10}^{-9}$ | $5.5\text{}\mathrm{pW}$ | $1.8\times {10}^{-9}$ | $17\text{}\mathrm{pW}$ | $0.8\text{}\mathrm{pW}$ |

**Table 5.**Comparison between our pixels and a classical resistive bolometer for three integration bandwidths—all electromechanical devices are set at their onset of nonlinearity.

Pixel | ${\mathit{\tau}}_{\mathit{t}\mathit{h}}$$\left(\mathbf{ms}\right)$ | ${\mathit{R}}_{\mathit{f}}$$\left(/{\mathit{W}}_{\mathit{i}\mathit{n}\mathit{c}}\right)$ | $\mathbf{NEP}\text{}{\mathit{f}}_{\mathit{B}\mathit{W}}=10\mathbf{Hz}\text{}\left(\mathbf{nW}\right)$ | $\mathbf{NEP}\text{}{\mathit{f}}_{\mathit{B}\mathit{W}}=50\mathbf{Hz}\text{}\left(\mathbf{nW}\right)$ | $\mathbf{NEP}\text{}{\mathit{f}}_{\mathit{B}\mathit{W}}=7\mathbf{kHz}\text{}\left(\mathbf{nW}\right)$ | NETD $\left(\mathit{K}\right)\text{}(\mathit{FOM}\u2014\mathit{K}\xb7\mathit{ms})\text{}{\mathit{f}}_{\mathit{B}\mathit{W}}=10\mathbf{Hz}$ | NETD $\left(\mathit{K}\right)\text{}(\mathit{FOM}\u2014\mathit{K}\xb7\mathit{ms})\text{}{\mathit{f}}_{\mathit{B}\mathit{W}}=50\mathbf{Hz}$ |
---|---|---|---|---|---|---|---|

Typical | $0.5$ | $1050$ | $0.14$ | $0.19$ | $2.4$ | 1.5 (0.75) | 2 (1) |

Butterfly | $0.8$ | $1011$ | $0.47$ | $1.1$ | $12$ | 4.9 (3.96) | 11.6 (9.28) |

Thin Rod | $2.8$ | $3555$ | $1.3$ | $3$ | $35$ | 13.7 (38.3) | - |

Resistive Pixel [1] | 16 | - | - | 0.05 | - | - | 0.05 (4) |

© 2018 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/).

## Share and Cite

**MDPI and ACS Style**

Duraffourg, L.; Laurent, L.; Moulet, J.-S.; Arcamone, J.; Yon, J.-J.
Array of Resonant Electromechanical Nanosystems: A Technological Breakthrough for Uncooled Infrared Imaging. *Micromachines* **2018**, *9*, 401.
https://doi.org/10.3390/mi9080401

**AMA Style**

Duraffourg L, Laurent L, Moulet J-S, Arcamone J, Yon J-J.
Array of Resonant Electromechanical Nanosystems: A Technological Breakthrough for Uncooled Infrared Imaging. *Micromachines*. 2018; 9(8):401.
https://doi.org/10.3390/mi9080401

**Chicago/Turabian Style**

Duraffourg, Laurent, Ludovic Laurent, Jean-Sébastien Moulet, Julien Arcamone, and Jean-Jacques Yon.
2018. "Array of Resonant Electromechanical Nanosystems: A Technological Breakthrough for Uncooled Infrared Imaging" *Micromachines* 9, no. 8: 401.
https://doi.org/10.3390/mi9080401