Issue
Eur. Phys. J. Appl. Phys.
Volume 89, Number 3, March 2020
Disordered Semiconductors: Physics and Applications
Article Number 30101
Number of page(s) 8
Section Semiconductors and Devices
DOI https://doi.org/10.1051/epjap/2020190245
Published online 23 April 2020

© EDP Sciences, 2020

1 Introduction

High-resolution infrared images require the reduction of the pixel area of the microbolometers used in Focal Plane Arrays (FPAs), especially in portable applications [1,2]. This means that the design should consider narrower support arms [3]. In addition, structural stability requires materials with minimum residual stress for successful device fabrication [4]. In both cases, a film with high stress will deform the structure affecting the overall performance. The maximum absorption at a certain wavelength is achieved by having high planarity in the bolometer's membrane [5], also, if it touches the surface of the substrate the thermal insulation will be severely reduced. In addition, long-term stability is a very important parameter for infrared cameras, manufacturers need to ensure proper operation during the products lifetime [6,7]. In this sense, the study of residual stress in the thermosensing film becomes a subject that must be analyzed.

It is known that residual stress is affected by the deposition technique, composition in the solid phase, thermal budget, thickness, among many others [810]. Taking this into account, the TCR of a material and its electrical resistivity are two important parameters for the design of microbolometers with high performance, however, they are directly related to the composition of the material and therefore to the stress of the film. Another way to adjust the tension is by changing the thickness of the film or having a multilayer structure, however, this complicates the design of the device, affecting the response time and the thermal conductance. Finite element simulations are very useful for design and predict the structural behavior of devices, however, the scarce information in the literature on mechanical properties in amorphous semiconductors makes analysis difficult.

Polymorphous semiconductors consist of an amorphous matrix with embedded nanocrystals, which are formed in the plasma and incorporated into the film during growth. Polymophous silicon is a material known for its excellent transport properties, unlike amorphous silicon, the density of defects is lower, and also its resistance to radiation-induced degradation (Staebler–Wronski effect) is lower as evidenced by “light-soaked” results of solar cells [11,12].

However, little or nothing is reported about its mechanical properties, that is why in this work its properties are studied to be used as an active film and at the same time as a structural material. In particular, three compositions of thermosensing material based on a hydrogenated polymorphous silicon-germanium (pm-SixGe1–x:H) alloy are studied. Stress reduction was performed with thermal treatments at low temperature, which resulted in a simple method with acceptable results. One of the optimized films was implemented in devices and its performance was characterized with infrared radiation.

2 Experimental details

2.1 Electrical properties

Three Corning 2947 glass substrates were initially cleaned in acetone, trichlorethylene, rinsed in deionized water and dried by centrifugation. After cleaning they were immediately metalized with 300 nm of titanium to create strips separated by 2 mm for electrical characterization. After achieving a base pressure of 5 mTorr for one hour, the films of pm-SixGe1–x:H were deposited using an Applied Materials AMP-3300 plasma deposition system working at 110 kHz according to Table 1.

For dark conductivity σ(T) measurements, the samples were placed in a vacuum thermostat from Janis at 13 mTorr. Current–voltage measurements were made with a Keithley 6517A electrometer. A first temperature sweep of 27–127 °C was carried out in steps of 10 °C with a 5 °C/min ramp; the sample was stabilized for 3 min before taking the value. After that, it was allowed to cool slowly to room temperature without breaking the vacuum; finally, it was reheated up to 127 °C as described above to guarantee reproducibility of the data. Conductivity σ(T) of pm-SixGe1–x:H shows a thermally activated behavior as described in equation (1), where Ea is the activation energy and k the Boltzmann constant [13](1)

Solving for Ea, so that the data can be adjusted by a straight line, gives equation (2):(2)

From equation (2) it can be seen that the slope of the line that adjusts the experimental data corresponds to Ea. TCR is related to Ea by means of equation (3):(3)

Table 1

Pm-SixGe1–x:H deposition parameters.

2.2 Mechanical properties

Three silicon wafers with a single polished face were cleaned in the same way as the glass substrates; in addition, the native oxide was etched with immersion in hydrofluoric acid solution. After that pm-SixGe1–x:H thin films were deposited according to Table 1 for each wafer. Wafer curvature measurement was made using a profilometer KLA-Tencor P-7 and stress was calculated using the Stoney equation [14]:(4)

where E is Young's modulus, ν is Poisson ratio and ts the thickness of the silicon wafer; tf the thickness of the pm-SixGe1–x:H film, R and R0 the final and initial curvature radius of the wafer respectively. The thickness obtained for the films 1–3 was: 300.6 nm, 407.6 nm and 486.3 nm. For each sample, the annealing temperature was carried out with slow heating of 8 °C/min until reaching 200 °C in a nitrogen environment, the temperature was maintained at that value for 2 hours, and then it was allowed to cool slowly to room temperature. Once the stress measurement was done, it was reheated as described above until reaching 200 °C for 1.5 hours to accumulate a total of 3.5 hours. This procedure was used also to measure stress at different temperatures and annealing times. It is important to mention that the temperature of this annealing process must not exceed the deposition temperature of the film in order to do not affect its properties as TCR and resistivity [15].

2.3 Devices

Having studied the best conditions to reduce residual stress, one film with high TCR (sample 2) was chosen to manufacture test structures and evaluate their performance. The process started with an oxidized silicon substrate, where SiO2 serves as an electrical insulator between devices, the process is depicted in Figure 1. Then, an aluminum layer of 600 nm was evaporated by e-beam and patterned through wet etching as the first electrode. After that, a 180 nm-thick spin-on-glass (Filmtronics 700B) layer was deposited and cured in air at 250 °C followed by a layer of amorphous silicon nitride (a-SiNx:H) deposited by PECVD at 200 °C with thickness of 120 nm. With both layers, coverage of pinholes is obtained and the planarity is improved. Next, another aluminum layer of 600 nm was e-beam evaporated and patterned through lift-off process to form the second electrode and the reflecting mirror. Then a 2.5 µm thick PI-2610 polyimide sacrificial layer was deposited and cured at 250 °C to form a Fabry-Perot quarter-wave cavity tuned at 10 μm wavelength. Infrared absorption is enhanced when cavity depth d satisfies the relation d = λ/4n + Nλ/2n, where N = 0, 1, 2, ... and n is the refractive index of vacuum after polyimide removal. Radiation that is not absorbed initially by the membrane travels through the cavity then is reflected back by the aluminum mirror interfering destructively with the incident wave at the membrane giving an enhanced absorption at 10 μm wavelength. After that, windows were opened through the polyimide to form the membrane supports using oxygen plasma, a thin layer of aluminum was employed as a hard mask and then removed by wet etching. Afterward, CF4 plasma was used to etch the remaining bilayer of SOG/SiNx at the bottom of the windows. Then the thermosensing film of pm-SixGe1–x:H with thickness of 407.6 nm was deposited by PECVD at 200 °C (same conditions as sample 2) followed by 3.5 hours of thermal treatment in a nitrogen environment at 200 °C in order to reduce stress. This layer is patterned by CF4 plasma to form the microbolometer membrane. Next, a 600 nm-thick titanium layer was also evaporated by e-beam and patterned by lift-off to create the microbolometer supports; the choice of titanium was convenient due to its low thermal conductivity. Finally, the wafer was diced and the polyimide sacrifice film was ashed in oxygen plasma until the structures were completely released.

thumbnail Fig. 1

Microbolometer structure.

3 Results and discussion

Figure 2 shows a High-Resolution Transmission Electron Microscopy (HRTEM) image for pm-SixGe1–x:H deposited with the same conditions of sample 2 without any annealing. Small nanocrystals with sizes of 3–4 nm are distinguished which are randomly distributed in the amorphous matrix.

Figure 3 shows an Arrhenius plot of conductivity as a function of temperature for the films under study, there is a strong relationship between the film composition and conductivity with an increase of two orders of magnitude, which is achieved by modifying the proportion of precursor gases SiH4 and GeH4. Sample 1 presented a TCR = 4.26% K−1 and σRT = 9.1 × 10−6 S cm−1, sample 2 presented a TCR = 4.08% K−1 and σRT = 1.5 × 10−5 S cm−1, finally, sample 3 presented a TCR = 2.25% K−1 and σRT = 3.7 × 10−3 S cm−1. A microbolometer requires an active material with high TCR, which means that small changes in the temperature of the thermosensing material translate into large changes in resistance. On the other hand, the high conductivity of the material decreases the noise of the devices, thus increasing the signal-to-noise ratio [16].

The results of stress reduction with annealing are shown in Figure 4, where is observed that effectively the thermal treatment helps to reduce the film stress and the larger reduction was achieved after 3.5 hours. It has been found in amorphous silicon (a-Si) films that crystallization at high temperature (above 500 °C) changes stress from compressive to tensile as a consequence of volume contraction [17]. However, in this work, the annealing was done at the same deposition temperature (200 °C), which indicates that this reduction may be related to other factors. In hydrogenated amorphous carbon films (a-C:H) it has been found that the hydrogen content is related to the microstructure of the material and therefore to the residual stress [18]. In this sense, annealing at 200 °C could help hydrogen to diffuse through the amorphous matrix of pm-SixGe1–x:H releasing stress. In hydrogenated amorphous germanium films (a-Ge:H) a strong relationship was found between the amount of unbonded hydrogen (interstitial or inside small voids H2 molecules) and the compressive stress of the films, suggesting that this applies pressure on the walls of the voids, giving a compressive stress contribution [19]. Similar results were found in silicon carbide films (a-SiC:H(F)) in annealing temperature ranges of 200 °C [20].

After reaching values close to zero, it was decided to study the influence of larger thermal treatments but at low temperature. During the last stage of polyimide ashing with oxygen plasma, the process can last between 2 and 5 hours, in that sense, the maximum temperature was limited to 120 °C. From Figure 4 it can be seen that the stress in sample 2 increases from 39.8 MPa to 47.3 MPa, which is still a low value. In sample 3 is observed a larger reduction of stress; however this film has lower TCR, which is not interesting for a further study. In the same way, sample 1 presented high stress and electrical resistance that was too high for its implementation in devices.

Besides, σ(T) was measured again after the thermal treatments, having found no significant variations (below 0.6%) in both TCR and σRT. This means that the ramp of 8 °C/min during the annealing process which was slightly higher than the ramp of 5 °C/min during the electrical characterization did not significantly affect the electrical properties.

Before device characterization, microbolometers were analyzed by scanning electron microscopy (SEM). Figure 5 shows a comparative analysis of the devices without thermal treatment and devices after thermal treatment, the devices area is of 50 × 50 μm2. Figure 5a shows devices with coplanar titanium electrodes on top of the pm-SixGe1–x:H film, in order to reduce the resistance of devices. However, the imbalance between the residual stress of titanium and pm-SixGe1–x:H caused a severe bending of the suspended membranes. Additionally, those devices were not subjected to any thermal treatment. To eliminate the stress in the membranes, they were thermally treated and the electrodes on top of the pm-SixGe1–x:H film were suppressed. Figure 5b and c show the devices after the thermal treatment process, where is observed a significant improvement on the reduction in the membranes stress and good uniformity across the array.

Silicon-based thermosensing materials, as it is the case for pm-SixGe1–x:H, allow monolithic integration of sensors with the readout integrated circuit (ROIC), this is a key advantage over other sensitive materials. However, as mentioned before, only test structures were fabricated in the present work. In order to quickly assess arrays with different thermosensing materials we are developing in our group an FPGA-based test platform. In this sense, the methods to characterize individual pixels are described here. Electrical characterization of microbolometers was carried out in a vacuum chamber from LakeShore model MTD-150 at 20 mTorr while the temperature was kept at 27 ± 0.01 °C by a temperature controller from LakeShore model DR-93CA. For the responsivity measurements infrared radiation was provided by a silicon carbide (SiC) bar from Kanthal Globar, the radiation was modulated using a mechanical chopper from 5 to 350 Hz and filtered using a flat zinc selenide (ZnSe) window with transmittance of 70% in the range of 0.6–20 μm followed by a 260 µm-thick silicon lid. The infrared power was calibrated using a 71968 thermopile from Oriel. Figure 6 shows the assembled installation. A constant current of 2.5 µA was passed through a metallic resistance of 2 MΩ and then through the microbolometer.

The load resistance value was very similar to that of the microbolometer, while the output voltage was measured with an oscilloscope. For noise measurements, the characterization was also performed in vacuum and at room temperature, but in dark conditions. Figure 7 shows the experimental arrangement; the voltage signal was fed to a lock-in amplifier to make the measurements.

Response to infrared radiation of 576 nW without modulation is shown in Figure 8. A high response is produced due to the elevated TCR of the thin film employed and the enhanced infrared absorption by the resonant cavity. In several works it is common to use an infrared absorber layer; however, management of residual stress in a multilayer structure is more complicated. Voltage responsivity (ℜV) is calculated through equation (5) where η is the absorption coefficient of pm-SixGe1–x:H, β is pixel fill factor, I is bias current, α is TCR, Ref stands for an effective resistance resulting from the equivalent parallel resistance between bolometer and readout circuit impedance, Gth is thermal conductance, ω is modulation frequency and τth is the thermal time constant. Another way to experimentally calculate ℜV is through equation (6) where ΔV is the voltage difference from dark to infrared and Pin is the infrared power which was 576 nW(5) (6)

According to equation (5) the response to radiation increases proportionally to the bias current, Figure 9 shows that the behavior of the devices is as expected. In order to avoid thermal runaway and burn out the microbolometers, a bias current of 2.5 μA was chosen. Another common way to increase ℜV is by reducing Gth, however, this instead makes the microbolometer more vulnerable to self-heating effect by Joule power dissipation. Self-heating can induce a temperature increase much higher than that caused by the absorption of weak infrared signals. The techniques most used to avoid this problem are pulsed biasing of detectors with duration much shorter than the thermal response time and using a self-heating cancellation circuit. In the latter, there is a Wheatstone bridge where one of the branches is the microbolometer to be measured, while in the other a reference resistor is used, especially one constructed with the same characteristics and dimensions as the microbolometers but protected from radiation. This works pretty well to compensate for variations in the temperature of the system. Although the measurement was not performed in that way, the microbolometer temperature was controlled at 27 ± 0.01 °C.

Using the installation of Figures 6 and 7, the responsivity and noise in the range of 5 to 350 Hz were measured; the results are shown in Figure 10. According to data of voltage responsivity, the cutoff frequency (fc) was calculated when the response dropped to ‑3 dB from its maximum [21], fc was found to be 76.25 Hz. After that, thermal response time (τth) was calculated by equation (7) giving a value of 2.08 ms, which is an improvement of our previous results with a cutoff frequency of 17 Hz [22].(7)

By eliminating the titanium electrodes used in the previous designs it was reduced the thermal capacitance and the response time. The noise level was tens of μV in the frequency range studied, which is well above the Jhonson noise floor of about 1.7 × 10−7 V/Hz1/2. The high resistivity in pm-SixGe1–x:H has a negative impact on the noise level, other factors that also affect this are the non-ohmic contacts, the geometry of the membrane, bias sources, the resistance to the load, among others. So that many parameters must be optimized to improve the overall performance of the system.

Noise equivalent power (NEP) gives the input power necessary to produce a signal to noise ratio equal to one, this figure of merit is described by equation (8). According to the previous results, a NEP value of 1 × 10−11 W/Hz1/2 at 30 Hz was found.(8)

Furthermore, in order to compare the performance of the microbolometers taking into account the signal to noise ratio normalized in area, the figure of merit detectivity (D*) was employed and is described by equation (9), where A is the effective area of the microbolometer and Vn is the noise voltage over the frequency bandwidth Δf, here considered equal to 1 Hz. The results are shown in Figure 11, where D* reaches a value of 4.4 × 108 cm Hz1/2/W at 30 Hz. D* at 30 Hz had a moderate value, this was due to the fact that the frequency at which the maximum of ℜV occurs does not coincide with the minimum noise frequency [23].(9)

As mentioned earlier, by having a membrane composed of a single layer of pm-SixGe1–x:H with optimized stress, we enhanced mechanical stability reducing the bending of the structure. The latter improved thermal isolation by lowering Gth, which is evidenced by D*. Furthermore, a simple structure working both as support and infrared absorber reduces the thermal mass which translates to better response time.

Finally, we investigated the noise equivalent temperature difference (NETD) as an additional performance parameter, it is given by equation (10). This gives the minimum temperature difference that the microbolometer can detect and most manufacturers of imaging cameras use it to specify performance.(10)

where F and τ0 are the focal ratio and the transmittance of the optics, respectively [24]; and (ΔPT)λ1-λ2 is the change in power density from a blackbody at temperature T with respect to ΔT, measured within the spectral band of λ1 to λ2. Here we consider F = 1, τ0 = 0.5, (ΔP/ΔT)8–12 = 1.972 × 10−4 W/cm2 K and according to the previous results, calculations predict a NETD value of about 18 mK.

thumbnail Fig. 2

HRTEM image of pm-SixGe1–x:H where small nanocrystals of sizes about 3–4 nm can be distinguished.

thumbnail Fig. 3

Temperature dependence of conductivity for samples of pm-SixGe1–x:H.

thumbnail Fig. 4

Reduction of pm-SixGe1–x:H stress in a nitrogen environment at two temperatures as a function of time.

thumbnail Fig. 5

SEM image of fabricated microbolometers with area of 50 × 50 μm2. (A) Devices without thermal treatment. (B) and (C) devices after 3.5 hours of thermal treatment.

thumbnail Fig. 6

Experiment setup for responsivity measurement.

thumbnail Fig. 7

Experiment setup for noise measurement.

thumbnail Fig. 8

Infrared response of a microbolometer.

thumbnail Fig. 9

Dependence of infrared response on bolometer bias current.

thumbnail Fig. 10

Voltage responsivity and noise measurements for a microbolometer with bias current of 2.5 μA.

thumbnail Fig. 11

Dependence of detectivity on chopper frequencies.

4 Conclusions

Compressive residual stress in pm-SixGe1–x:H films deposited by PECVD was found around 500 MPa. By means of thermal treatments for 3.5 hours in a nitrogen environment at 200 °C, it was possible to reduce it to almost zero. A slight change to tensile stress occurs if the annealing is prolonged for a longer time. This process did not affect the TCR value and the electrical resistivity of the films. The microbolometers analyzed by SEM showed good mechanical stability and the figures of merit studied showed that the thermal response time is improved with respect to the previous results. Finally, the analysis of the results shows that the method for stress reduction is suitable for the development of microbolometer arrays, where fabrication of homogeneous devices is mandatory.

Author contribution statement

A.B.C.E. conceived the experiment of a silicon-based semiconductor for its application in microbolometers; A.B.C designed, planned and performed the experiment on the reduction of residual stress; B.C.D.F. supervised, reviewed and evaluated the results of the characterization of microbolometers; A.B.C. wrote the manuscript with support from D.F.

Acknowledgments

This research was partially supported by FOMIX Puebla − CONACYT project No. PUE-2018-03-02-84557. As well Ricardo Jimenez would like to thank CONACYT for the PhD scholarship No.-364498.

References

  1. S. Tohyama, T. Sasaki, T. Endoh, M. Sano, K. Katoh, S. Kurashina, M. Miyoshi, T. Yamazaki, M. Ueno, H. Katayama, T. Imai, Infrared Technol. Appl. XXXVII. 8012 , 80121M (2011) [CrossRef] [Google Scholar]
  2. R. Blackwell, G. Franks, D. Lacroix, S. Hyland, R. Murphy, Small pixel uncooled imaging FPAs and applications, in Proceedings Volume 7660, Infrared Technology and Applications XXXVI (2010), paper 76600Y [Google Scholar]
  3. D. Weiler, M. Russ, D. Würfel, R. Lerch, P. Yang, J. Bauer, H. Vogt, Technology 7660 , 76600S (2010) [Google Scholar]
  4. I.W. Kwon, J.E. Kim, C.H. Hwang, Y.S. Lee, H.C. Lee, Sens. Actuators A 147 , 391 (2008) [CrossRef] [Google Scholar]
  5. M. Ruß, J. Bauer, H. Vogt, Infrared Technol. Appl. XXXIII 6542 , 654223 (2007) [CrossRef] [Google Scholar]
  6. J.L. Tissot, S. Tinnes, A. Durand, C. Minassian, P. Robert, M. Vilain, Electro-Optical Infrared Syst. Technol. Appl. VII 7834 , 78340K (2010) [CrossRef] [Google Scholar]
  7. S. Garcia-Blanco, P. Topart, Y. Desroches, J.S. Caron, F. Williamson, C. Alain, H. Jerominek, Proc. SPIE 6884 , 68840P (2008) [CrossRef] [Google Scholar]
  8. G. Abadias, E. Chason, J. Keckes, M. Sebastiani, G.B. Thompson, E. Barthel, G.L. Doll, C.E. Murray, C.H. Stoessel, L. Martinu, J. Vacuum Sci. Technol. A 36 , 020801 (2018) [CrossRef] [Google Scholar]
  9. C. Iliescu, B. Chen, J. Micromech. Microeng. 2008 , 015024 (2008) [Google Scholar]
  10. C.-C. Lee, Opt. Eng. 49 , 063802 (2010) [Google Scholar]
  11. A.V. Kharchenko, K. Ouaras, V. Suendo, J. Ebothé, P.R. Cabarrocas, J. Appl. Phys. 125 , 163307 (2019) [Google Scholar]
  12. W. Bronner, J. Kleider, R. Brüggemann, P. Roca i Cabarrocas, D. Mencaraglia, M. Mehring, J. Non. Cryst. Solids. 299–302 , 551 (2002) [Google Scholar]
  13. S. Ajmera, J. Brady, C. Hanson, T. Schimert, A.J. Syllaios, M. Taylor, Proc. SPIE 2011 , 8012 (2011) [Google Scholar]
  14. D. Lysáček, P. Kostelník, P. Pánek, Solid State Phenom. 205–206 , 284 (2013) [CrossRef] [Google Scholar]
  15. J.Z. Liu, V. Chu, D.S. Shen, D. Slobodin, S. Wagner, Phys. Rev. B. 40 , 6424 (1989) [Google Scholar]
  16. S. Abbasi, A. Shafique, O. Ceylan, C.B. Kaynak, M. Kaynak, Y. Gurbuz, IEEE Sens. J. 18 , 6217 (2018) [Google Scholar]
  17. R.T. Howe, R.S. Muller, J. Appl. Phys. 54 , 4674 (1983) [Google Scholar]
  18. B. Racine, M. Benlahsen, K. Zellama, P. Goudeau, M. Zarrabian, G. Turban, Appl. Phys. Lett. 73 , 3226 (1998) [Google Scholar]
  19. F.C. Marques, P. Wickboldt, D. Pang, J.H. Chen, W. Paul, J. Appl. Phys. 84 , 3118 (1998) [Google Scholar]
  20. D.S. Kim, Y.H. Lee, Thin Solid Films 261 , 192 (1995) [Google Scholar]
  21. R.W. Astheimer, Proc. SPIE 0443 , 95 (1983) [CrossRef] [Google Scholar]
  22. R. Jiménez, M. Moreno, A. Torres, P. Rosales, M.T. Sanz, R. Ambrosio, Phys. Status Solidi 215 , 1700736 (2018) [CrossRef] [Google Scholar]
  23. J.E. Gray, Z. Çlik-Butler, D.P. Butler, J. Microelectromechanical Syst. 8 , 192 (1999) [CrossRef] [Google Scholar]
  24. P.W. Kruse, Semicond. Semimetals. 47 , 17 (1997) [CrossRef] [Google Scholar]

Cite this article as: Ricardo Jimenez, Mario Moreno, Alfonso Torres, Roberto Ambrosio, Aurelio Heredia, Arturo Ponce, Reduction of residual stress in polymorphous silicon germanium films and their evaluation in microbolometers, Eur. Phys. J. Appl. Phys. 89, 30101 (2020)

All Tables

Table 1

Pm-SixGe1–x:H deposition parameters.

All Figures

thumbnail Fig. 1

Microbolometer structure.

In the text
thumbnail Fig. 2

HRTEM image of pm-SixGe1–x:H where small nanocrystals of sizes about 3–4 nm can be distinguished.

In the text
thumbnail Fig. 3

Temperature dependence of conductivity for samples of pm-SixGe1–x:H.

In the text
thumbnail Fig. 4

Reduction of pm-SixGe1–x:H stress in a nitrogen environment at two temperatures as a function of time.

In the text
thumbnail Fig. 5

SEM image of fabricated microbolometers with area of 50 × 50 μm2. (A) Devices without thermal treatment. (B) and (C) devices after 3.5 hours of thermal treatment.

In the text
thumbnail Fig. 6

Experiment setup for responsivity measurement.

In the text
thumbnail Fig. 7

Experiment setup for noise measurement.

In the text
thumbnail Fig. 8

Infrared response of a microbolometer.

In the text
thumbnail Fig. 9

Dependence of infrared response on bolometer bias current.

In the text
thumbnail Fig. 10

Voltage responsivity and noise measurements for a microbolometer with bias current of 2.5 μA.

In the text
thumbnail Fig. 11

Dependence of detectivity on chopper frequencies.

In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.