2.1.

Principle of the Bimaterial Microcantilever FPA

Figure 1(a) shows the structure of a bimaterial microcantilever FPA pixel; Fig. 1(b) is an FPA photo captured by a scanning electron microscope. As shown in Fig. 1(a), the pixel structure consists of an absorber/reflector, two bimaterial deflection arms, and two thermal isolation arms. The absorber/reflector of a microcantilever pixel has one layer made of an IR absorber made by SiNx and the other layer as a visible light reflector made by Au. Figure 1(c) shows deflection of the bimaterial microcantilever under IR radiation. When IR radiation is incident on the FPA, the absorber converts the radiation to heat and transfers the heat to the connected bimaterial deflection arms. The arms are made of two kinds of materials: SiNx and Au. Since these two materials have different thermal expansion coefficients, the cantilevers bend with the change of temperature and drive the connected reflectors to change their deflection angles. The target infrared radiation can then be detected by reading out the deflection angles. The isolation arms are a layer of SiNx, whose extremely low thermal conductance thermally isolates the pixels from the supporting substrate.

Fig. 1

(a) Structure of a microcantilever FPA pixel, (b) an FPA photo captured by a scanning electron microscope, and (c) deflection of the bimaterial cantilever under IR radiation.

2.2.

Fabrication of the Bimaterial Microcantilever FPA

A bulk silicon micromachining process, which is schematically shown in Fig. 2, has been developed to fabricate the bimaterial microcantilever FPA. The fabrication starts with a double side polished silicon wafer with a thickness of 300 μm. The detailed micromachining steps are introduced as follows: first, an ${\mathrm{SiO}}_2$ film of 200 nm is deposited to form the etching buffer layer between the structure layer and the substrate, and a layer of low stress SiNx of 1 μm is deposited with the low pressure chemical vapor deposition method. Second, a 500 nm Au film is sputtered and patterned to form the bimaterial arms with a wet etching technique. Third, in order to decrease the bending and the surface roughness of the cantilever reflector, only 50 nm Au film is sputtered and patterned as the reflecting mirror on the SiNx absorber with a lift off technique. Next, the SiNx cantilever structure is patterned and etched with reactive ion etching (RIE) technique. Fifth, a glass wafer is patterned and etched to form a 60 μm deep cavity with the cavity dimension the same as the FPA. Then the glass and silicon wafers are anodic bonded together. Here, the glass wafer is used as a supporting wafer to avoid the thin FPA structure to be broken. Next, the silicon wafer is thinned down to 100 μm in KOH solution to increase etching uniformity and decrease footing effect in the following inductively coupled plasmas (ICP) etching. In the penultimate stage, double-side lithography is used to pattern the backside silicon. At last, the silicon substrate is anisotropically etched in ICP system and the ${\mathrm{SiO}}_2$ buffer layer is etched in RIE system to release the cantilevers.

Fig. 2

Fabrication process of the FPA.

In this process, silicon-glass anodic bonding and deep reactive ion etching (DRIE) techniques are introduced to selectively remove the substrate silicon at the area where the cantilevers are located and form a silicon frame for each of the cantilever pixels. Figure 1(b) shows an FPA photo captured by a scanning electron microscope.

2.3.

Optical-Thermal Transducing Theory of SWIR, MWIR, and LWIR

We define the wavelength range of SWIR from 1 to 3 μm, the range of MWIR from 3 to 8 μm, and the range of LWIR as 8 to 14 μm. It is known that the temperature of bimaterial microcantilever will change owing to the absorption of infrared radiation. The optical-thermal transducing calculation theory based on different wavebands can be given by the following two steps. The first step is to calculate the change of energy to temperature per unit area in the SWIR, MWIR, and LWIR wavebands respectively. The second is to get the microcantilever’s temperature change caused by the IR target’s temperature change according to the first step’s results.

Above all, ${(dP/dT)}_{\lambda 1-\lambda 2}$ is defined as the change of energy to temperature per unit area in the specific IR waveband. The calculation formula of $dP/dT$ can be expressed as¹⁶

Here $v_{\lambda 1-\lambda 2}$ represents the infrared radiation conversion efficiency (absorptivity) of the bimaterial microcantilever pixel, $\lambda 1-\lambda 2$ stands for the specific IR waveband, $\lambda 1$ is the minimum wavelength and $\lambda 2$ is the maximum wavelength. The different values of $v$ in different IR wavebands can be gotten by the measurement results of Nicolet Magna 750 Fourier Transform Infrared Spectrometer. The values of $v$ are about 35%, 38%, and 61% in the SWIR, MWIR, and LWIR wavebands.

Next, the cantilever temperature change caused by the temperature change of the IR target source can be expressed as¹⁷

Here, $A_{\text{pixel}}$ represents the total area of a microcantilever cell, $\epsilon$ is the emission efficiency (${\epsilon }_{\mathrm{SiNx}}=0.8$, ${\epsilon }_{\mathrm{Au}}=0.01$), $\sigma$ is Stefan-Boltzmann constant ($\sigma =5.67\times {10}^{-8}\mathrm{W}·{\mathrm{m}}^{-2}·{\mathrm{K}}^4$), $T_b$ stands for the room temperature (300 K). As is shown in Fig. 1(a), $A_{\text{pixel}}$ is $146\times 110{\mu \mathrm{m}}^2$ in this paper. The leg’s conductance $G_{\mathrm{arm}}$ can be calculated using

2.4.

Thermal-Mechanical Response Theory of the Bimaterial Microcantilever

It is known that the bimaterial microcantilever will deflect an angle with the change of its temperature and the thermal-mechanical response theory is the same in different IR wavebands. The following two steps can be used to express the thermal-mechanical response calculation method: (1) get the microcantilever deflection $\delta$ according to its temperature change $\mathrm{\Delta }T$; (2) calculate the deflection angle $\theta$ corresponding to the deflection $\delta$.

The bimaterial microcantilever deflection $\delta$ cased by the temperature change $\mathrm{\Delta }T$ in a microcantilever can be expressed as⁵

Here, $n=d_{\mathrm{Au}}/d_{\mathrm{SiN}}$, $K=4+6n+4n^2+{\varphi n}^3+{(\varphi n)}^{-1}$, $\varphi =E_{\mathrm{Au}}/E_{\mathrm{SiN}}$, $d$ is the thickness of the film ($d_{\mathrm{SiN}}=750\mathrm{nm}$, $d_{\mathrm{Au}}=500\mathrm{nm}$), $\alpha$ stands for the coefficient of thermal expansion (${\alpha }_{\mathrm{SiN}}$${\alpha }_{\mathrm{SiN}}=0.8\times {10}^{-6}{\mathrm{K}}^{-1}$, ${\alpha }_{\mathrm{Au}}=14.2\times {10}^{-6}{\mathrm{K}}^{-1}$), $L_m$ is the length of the bimaterial arm ($L_m=80\mu \mathrm{m}$) and $E$ represents the elastic modulus ($E_{\mathrm{SiN}}=180\mathrm{GN}/\mathrm{m}$, $E_{\mathrm{Au}}=73\mathrm{GN}/\mathrm{m}$).

Then, the microcantilever deflection $\delta$ can be expressed as $\delta =L_{\mathrm{arm}}\times \sin (\theta )$ where $L_{\mathrm{arm}}$ is the length of the bimaterial arm ($L_{\mathrm{arm}}=80\mu \mathrm{m}$), so the deflection angle $\theta$ can be obtained by $\theta =\mathrm{arc\; sin}(\delta /L_{\mathrm{arm}})$.

3.1.

Prototype of the Microcantilever IR Imaging System

An optical readout bimaterial microcantilever FPA uncooled IR imaging system was developed for the SWIR, MWIR, and LWIR imaging experiments. Figure 3 shows the schematic of the system. The FPA is placed in a vacuum chamber with a coating IR through-window for passing specified waveband of infrared radiation (three types of IR through-window and lens were used for passing the SWIR, MWIR, and LWIR radiation respectively). The specified IR lens, which can pass the specified wavebands of IR radiation and focus the target’s temperature distribution on the FPA, is placed between the target and the vacuum chamber. Visible light from a green light emitting diode (LED) is collimated to parallel light and reflected by a reflector, then reflected by a beam splitter, and then irradiates the FPA. Lens 1 and lens 2 form an optical 4f system. A spatial filter (narrow-strip filter) is placed on the spectrum plane. The spectra of the FPA have a portion of their intensity curtained off under the action of it. A charge-coupled device (CCD) is placed on the image plane to acquire IR image. A more detailed description of the system based on a narrow-strip filter optical readout approach can be found in Ref. 14.

Fig. 3

Schematic of the microcantilever FPA uncooled IR imaging system.

Figure 4 is a top-view photograph of the integrated experimental setup. As shown in Fig. 4, the part that produces the LED collimated light beam has been integrated as an LED parallel light generator. Lens 1, lens 2, and the spatial filter have been integrated as a 4f filtering system; the IR through-window and lens can be replaced in different wavebands IR imaging experiments. In addition, the atmospheric pressure in the vacuum chamber is 1 Pa; the number of FPA pixels is $160\times 120$; the CCD is a 14-bits CCD, and its signal-to-noise ratio is 56 dB.

Fig. 4

Setup of the microcantilever FPA IR imaging system.

3.2.

Imaging Experiments for SWIR, MWIR, and LWIR

The SWIR, MWIR, and LWIR imaging experiments were conducted by changing the different wavebands of IR through-window and lens. In the short-wave infrared imaging experiment, the SWIR through-window is made of quartz and the SWIR target is an SWIR laser (the SWIR lens is not necessary because the laser beam has been focused). In the medium-wave infrared imaging experiment, the through-window and infrared lens are made of germanium glass. The glass is coated for passing medium-wave infrared radiation. The selected MWIR target is a 673.15 K electric iron. According to the Planck formula, we can draw the curve of the target object radiation. Planck formula is as follows:

Fig. 5

(a) Radiant exitance curve of MWIR target object (673.15 K electric iron); (b) Radiant exitance curve of 2197 K target (SWIR laser).

It is known that the center wavelength of the SWIR laser beam is 1.319 μm. In theory, the absolute temperature of the blackbody can be calculated by Wien’s displacement law ${\lambda }_{\max }=b/T$, where ${\lambda }_{\max }$ is the peak wavelength, $T$ is the absolute temperature of the black body, and $b$ is a constant of proportionality called Wien’s displacement constant, equal to $2.8977685\times {10}^{-3}\mathrm{m}·\mathrm{K}$. So, according to Wien’s displacement law, the absolute temperature of the blackbody corresponding to the peak wavelength 1.319 μm is about 2197 K. Figure 5(b) shows the radiant exitance curve of 2197 K target, and we can see that the absolute temperature corresponding to the peak wavelength 1.319 μm is about 2197 K.

In the long-wave infrared imaging experiment, the through-window and infrared lens are also made of germanium glass, but different from the MWIR experiment, the glass is coated for passing through LWIR infrared radiation. The selected LWIR target is a person’s head. It is well known that the human body temperature is about 310.15 K. Therefore, its radiant exitance curve can be drawn by Planck formula. As shown in Fig. 6, it can be seen that the selected target radiates LWIR infrared mainly.

Fig. 6

Radiant exitance curve of LWIR target object (a person’s head).

Next, the SWIR, MWIR, and LWIR imaging experimental results based on the optical readout bimaterial microcantilever uncooled IR imaging system are given. Figure 7(a) is the infrared image of SWIR target (SWIR laser beam); Fig. 7(b) shows the imaging result of MWIR target (673.15 K electric iron); Fig. 7(c) is the infrared image of LWIR target (a person’s head) captured by the system. As shown in Fig. 7, the clear infrared images of the SWIR, MWIR, and LWIR targets have been obtained successfully. The results prove that the microcantilever FPA responds well to SWIR, MWIR, and LWIR radiation.

Fig. 7

(a) The infrared image of SWIR target, (b) the infrared image of MWIR target, and (c) the infrared image of LWIR targets.