1.

Introduction

Until now, the development of IR imaging systems has been of growing interest due to the wide range of application from military, civilian night vision, mine detection, surveillance, and medical imaging.^1–4 The recent advances in micromachining technology have made it possible to fabricate highly sensitive thermal IR microbolometer detection that can be operated at room temperature.⁵

V_xO_y thin film is one of the most commercially sensitive materials for microbolometer,^6–19 which is counted as a smart material. Besides the most popular thin films of vanadium oxide, other materials have been studied as thermal-sensitive element for bolometer, such as metal films,²⁰ yttrium barium copper oxide (Yba Cuo),^21–23 polycrystalline silicon germanium (SiGe),²⁴ and amorphous silicon (α-Si).^18,25

The amorphous silicon can be used as a unique sandwich structure which consists of the segregation layer located between the metal and the active layers. It requires high annealing temperature of about 1000°C to achieve stability of microstructure. The temperature coefficient of resistance (TCR) value of A-Si microbolometer is about −2.8%/K. It was suggested that the A-Si bolometer has higher temperature sensitivity and a detectivity that depends strongly on frequency.²⁵

At room temperature, material temperature coefficient of resistance (TCR) is an important factor for infrared sensors performance,^26–28 and the bolometer with high TCR has reached detectivity of ($1.{0\times 10}^9\mathrm{cm}{\mathrm{Hz}}^{1/2}/\mathrm{W}$).²⁸

Due to the strong effect of deposition parameters on the composition and microstructure of vanadium oxides, different phases can be obtained by changing the deposition process parameters. They undergo transition from an insulator or semiconductor to a metal phase at a specific temperature.

In our previous works, we have determined electrical, thermal, and optical properties of different multilayer of vanadium oxide mixed phase.^9,29,30 Achieving a desirable structure with higher TCR and lower resistivity is the main motivation for using different layer structures. The target total thickness of the structure was between 80 and 120 nm, which is suitable for standard bolometer thickness. It seems that, as the number of layers is increased, the TCR increases while the resistivity decreases. The aim of using a different number of layers is to increase the percentage of V in the structure to obtain different oxides during annealing.³¹ The nine alternate layers of ${\mathrm{V}}_2{\mathrm{O}}_5/\mathrm{V}$ deposited on silicon substrate have proven successful in realizing higher TCR and lower resistivity than the five alternate layers of ${\mathrm{V}}_2{\mathrm{O}}_5/\mathrm{V}$ deposited on quartz substrate.³¹ We have also proposed the layer structure of the air-bridge microbolometer which presented the fabrication and design of vanadium oxide microbolometer.³²

However, to improve the efficiency of such device, in this work, we report on the microstructural, optical, and thermal properties of the nine alternative layers of ${\mathrm{V}}_2{\mathrm{O}}_5/\mathrm{V}$ by Raman spectroscopy and photothermal deflection techniques (PDS and PDT). The results prove that changing the annealing atmospheres cause a change in Raman spectra which presents the formation of multiple phases such as ${\mathrm{VO}}_2$, ${\mathrm{V}}_2{\mathrm{O}}_5$, and ${\mathrm{V}}_6{\mathrm{O}}_{13}$. This latter introduces a variation in the most sensitive parameters to the temperature coefficient of resistance, such as optical absorption spectra, optical band gap, and thermal conductivity.

2.

Experiment

The layer structure proposed for the air-bridge microbolometer based upon the fabrication and design of vanadium oxide microbolometer is represented in Fig. 1.

Fig. 1

The layer structure of the proposed air-bridge microbolometer.

The thin film thermometer material with a 95-nm-thick multilayer structure composed of nine alternating layers of ${\mathrm{V}}_2{\mathrm{O}}_5$ and V with thickness of 15 and 5 nm, respectively, was deposited at room temperature and further annealed at 300°C for 30 min in ${\mathrm{N}}_2$ and ${\mathrm{O}}_2$ atmospheres. As shown in Fig. 2 shows a two-dimensional schematic of the multilayer thin-film structure.

Fig. 2

Schematic diagram of the multilayer structure.

The ${\mathrm{V}}_2{\mathrm{O}}_5$ layers were deposited using RF sputtering at chamber base pressure of ${2\times 10}^{-6}\mathrm{Torr}$ at 3 mTorr of Ar pressure and 150 W of RF power using 99.5% pure ${\mathrm{V}}_2{\mathrm{O}}_5$ target. The V layers were deposited using DC sputtering at chamber base pressure of ${2\times 10}^{-6}\mathrm{Torr}$ at 3 mTorr of Ar pressure and 150 W of DC power using 99.8% pure V target. Deposited structures were then annealed in a tube furnace at 300°C for 30 min. The annealing was performed at ${\mathrm{O}}_2$ and ${\mathrm{N}}_2$ at flow rates of 120 ml/min.

3.

Results and Discussion

3.1.

Raman Spectroscopy

Raman spectroscopy is also used, which is a sensitive tool for further investigation of the structure of a material.^33,34

Figure 3 shows the room temperature Raman spectra using 632.8 nm excitation, of the vanadium oxide films deposited in different annealing atmosphere. The set is composed of three samples, the first annealed in ${\mathrm{O}}_2$, annealed in ${\mathrm{N}}_2$ and the third is not annealed for comparison.

Fig. 3

Room temperature Raman spectra of the as-grown sample and samples subjected to different annealing time at 300°C for 30 min.

The spectra were recorded in the range of 200 to $1200{\mathrm{cm}}^{-1}$. The spectra show a broad band centered at $528{\mathrm{cm}}^{-1}$ which refers to the short-range order in the multiple layer sample. The appearance of the other peaks from the samples annealed in ${\mathrm{O}}_2$ and ${\mathrm{N}}_2$ atmosphere caused by the convolution with a vibration mode from the vanadium oxide.

The samples spectra annealed in ${\mathrm{O}}_2$ and ${\mathrm{N}}_2$ are dominated by emerging peaks at 169, 229, 310, 403, 528, 706, and $988{\mathrm{cm}}^{-1}$.

These results suggested the formation of multiple phases of vanadium oxide. The high frequency vibration at $1000{\mathrm{cm}}^{-1}$ can be attributed to the terminal oxygen (VO) stretching. The vibration at about 625 and $762{\mathrm{cm}}^{-1}$ corresponds to the stretching vibration of ${\mathrm{V}}_3\mathrm{O}$.

The spectra of annealed sample in ${\mathrm{O}}_2$ atmosphere show more peaks than the sample annealed in ${\mathrm{N}}_2$ atmosphere and the as grown sample.

This is due to the kinetic reaction of vanadium with oxygen. The peak at $528{\mathrm{cm}}^{-1}$ appears in Raman spectra of all the samples which indicate the presence of ${\mathrm{V}}_2{\mathrm{O}}_5$. The peak at $310{\mathrm{cm}}^{-1}$ shown in the spectra of the as-grown sample and the sample annealed in ${\mathrm{N}}_2$ atmosphere indicates the presence of ${\mathrm{VO}}_2$. This peak disappears for an annealing in ${\mathrm{O}}_2$ atmosphere and the presence of ${\mathrm{VO}}_2$ is indicated in the peak at $229{\mathrm{cm}}^{-1}$. We acknowledge the presence of ${\mathrm{V}}_6{\mathrm{O}}_{13}$ in all Raman spectra at $988{\mathrm{cm}}^{-1}$, however for the sample annealed in ${\mathrm{O}}_2$ spectra the peaks corresponded to ${\mathrm{V}}_6{\mathrm{O}}_{13}$ appears at 167 and $887{\mathrm{cm}}^{-1}$.

These results are in good agreement with those obtained previously.³⁰

3.2.

Photothermal Deflection Measurements

The photothermal deflection technique PDT or “mirage effect” technique was first introduced in the early 1980s by Bocarra, Fournier, Baldoz³⁵ and subsequently developed by Aamodt,³⁶ Murphy,³⁷ and by Jackson et al.³⁸ The biggest advantage of this noncontactless, nondestructive, and highly sensitive technique is the great precision for optical and thermal properties materials evaluation.^39–41

The principle of the photothermal deflection technique or “mirage effect” consists of heating the sample (placed in a gas or liquid) with a modulated light. This heating will generate a temperature gradient therefore a refractive index gradient in the fluid in contact with the heated sample’s surface, which will cause the deflection of a laser probe beam skimming the sample surface (Fig. 4).

Fig. 4

Principle of photothermal deflection.

This deflection $\mathrm{\Psi }$⁴² is a function of thermal and optical properties of the sample and also of the thermal properties of fluid and substrate:

Eq. (1)

$$\psi =\frac{dz}{dx}=-\frac{l}{n}\frac{dn}{d{T}_f}\frac{d{T}_f}{dz},$$

where $T_f$ is the distribution temperature elevation in the fluid and $n$ is its refractive index.

This temperature, in the case of uniform heating, can be written

$$T_f=T_0{e}^{-{\sigma }_{f}z}{e}^{j\omega t},$$

where $T_0$ is the periodic temperature elevation at the sample surface

$${\sigma }_f=(1+j)\sqrt{\frac{\pi F}{D_f}},$$

where $F$ and $D_f$ are the modulation frequency and thermal diffusivity of the fluid, respectively.

3.2.1.

Experimental setup of photothermal deflection techniques PDS and PDT

The photothermal deflection spectroscopy mounting aims to study the variations of the photothermal signal as a function of wavelength. The experimental setup is practically the same that of Fig. 5 of PDT technique except that a monochromator was interposed between the lamp and the chopper.

Fig. 5

Experimental set-up of PDT Technique. 1: micrometric table; 2: sample; 3: laser probe; 4: attenuator; 5: mechanical chopper; 6: four-quadrant position photodetector; 7: synchronous detection; 8: halogen lamp; 9: microcomputer.

PDS is a pump-probe technique. Pump and probe beams are aligned perpendicular to each other. The sample is immersed in a liquid, such as paraffin oil which exhibits a pronounced temperature dependence of its refractive index. The sample is heated by a monochromatic light coming from a halogen lamp and a monochromator and then modulated by a mechanical chopper at a constant frequency. The absorption by the sample of the modulated pump beam will generate a thermal wave that creates a refractive index gradient in the fluid. A focused laser probe beam (He–Ne laser) crossing the liquid close to the sample surface is deflected. The deflection of the probe beam is measured due to a silicon photodetector of four quadrants (QD50T) linked to a lock-in amplifier (EG&G Model 5209) giving the amplitude and the phase of the photothermal signal. A personal computer was used to store the amplitude and the phase of the signal and draw their variations with the wavelength. However, for PDT, we study the variation of amplitude and phase as a function of square root of modulation frequency. The sample in this case is placed in air. The experimental PDT setup is described in detail elsewhere.⁴³

3.2.2.

Absorption spectrum

To investigate the gap energy with great precision and determining the absorption spectrum of ${\mathrm{V}}_x{\mathrm{O}}_y$ samples, we have used the highly sensitive technique PDS.

On the same sample, we have conducted a spectroscopic study. On Fig. 6(a), the experimental normalized amplitude variation with wavelength is reported. We note that the amplitude presents two saturated areas corresponding to high and very low absorption coefficients. To deduce the optical absorption spectrum of vanadium oxide multilayer thin films, we have to compare the experimental curves to the corresponding theoretical ones as shown in Fig. 6(b).

Fig. 6

Experimental (a) and theoretical (b) amplitude of PDS signals versus wavelength and absorption coefficient, respectively, of two samples of ${\mathrm{V}}_x{\mathrm{O}}_y$ annealed at 300°C for 30 min in ${\mathrm{N}}_2$ and ${\mathrm{O}}_2$ atmosphere and not annealed sample.

In Fig. 7 the optical absorption spectrum of the ${\mathrm{V}}_x{\mathrm{O}}_y$ samples is reported.

Fig. 7

Optical absorption spectra versus photon energy E of ${\mathrm{V}}_x{\mathrm{O}}_y$ thin films.

3.2.3.

Determination of the gap energies

Gap energy is obtained from the optical absorption spectrum using the Tauc law for energies above the gap $(\alpha E{)}^n=\beta (E-E_g)$ where $\beta$ is a constant, $E_g$ is the gap energy, $E=h\upsilon$ is photon energy, $n=2$ for semiconductors having direct gap, and $n=1/2$ for those having indirect gap. In our case, we take $n=3/2$ for direct forbidden transition.

The variations of $(\alpha E{)}^n$ according to the energy $E$ are shown in Fig. 8.

Fig. 8

$(\alpha E{)}^{2/3}$ versus energy E near the band gap energy.

The extrapolation of the linear part of the curve intersects the axis of the energy ($\alpha =0$) by a value corresponding to the gap energy. The calculated values are 2.49, 2.58, and 2.7 eV for as grown and annealed samples in ${\mathrm{O}}_2$ and ${\mathrm{N}}_2$ atmosphere, respectively. Obtained results are comparable with the values reported by several authors.^44–46

3.2.4.

Thermal properties

Thermal conductivity measurements for the synthesized ${\mathrm{V}}_x{\mathrm{O}}_y$ thin films were investigated by PTD technique. Theoretical simulation consists of fitting the experimental amplitude and phase with theoretical ones obtained by the complex expression of the deflection function given by Ref. 42. The curves of Fig. 9 represent the experimental and theoretical variations of the amplitude and the phase of the photothermal signal according to the square root of frequency. The best fit of the signal amplitude and the phase variation yields an estimation of the thermal conductivity of the layered ${\mathrm{V}}_x{\mathrm{O}}_y$ film.

Fig. 9

Experimental (points) and theoretical (line) photothermal signals versus modulation frequency for ${\mathrm{V}}_x{\mathrm{O}}_y$ samples annealed at 300°C for 30 min in various annealing atmospheres.

The calculated values of thermal conductivity are $1.96\pm 0.01$, $2.55\pm 0.02$, and $3.6\pm 0.02\mathrm{W}.{\mathrm{m}}^{-1}.{\mathrm{K}}^{-1}$ for as grown and for annealed samples in ${\mathrm{O}}_2$ and ${\mathrm{N}}_2$ atmospheres, respectively. Thus, the best TCR value is found to be $-3.06\%/\mathrm{K}$ and the lower resistivity equal to 0.84 Ω.cm for the sample annealed in ${\mathrm{O}}_2$ atmosphere; however, for the sample annealed in ${\mathrm{N}}_2$ atmosphere has demonstrated high TCR $-3.55\%/\mathrm{K}$ and high resistivity $2.68\mathrm{\Omega }.\mathrm{cm}$.³¹

The increase of thermal conductivity for ${\mathrm{V}}_x{\mathrm{O}}_y$ annealed in ${\mathrm{N}}_2$ atmosphere refers to two main reasons, the first reason relies on the reduced defects density and especially oxygen-related vacancies annihilation by thermally activated atomic diffusion and/or intermixing. The second reason can be linked to the microstructural properties of the ${\mathrm{V}}_x{\mathrm{O}}_y$ films.^29,30 The observed behavior is confirmed by Raman spectroscopy as shown in Fig. 2. It suggests different phases of ${\mathrm{V}}_x{\mathrm{O}}_y$ to be formed at different annealing atmosphere.

The obtained results are in good agreement with values reported in our previous work.^9,29 However, in the present work, we have reached successful results using the nine alternating layers of $\mathrm{V}/{\mathrm{V}}_2{\mathrm{O}}_5$ deposited on silicon and annealed at 300°C for 30 min in ${\mathrm{O}}_2$ atmosphere.

4.

Conclusion

High TCR and low resistivity are the most needed parameters for vanadium oxide thin films integrated for microbolometer applications. These latter could be achieved by growing mixed phase of V and ${\mathrm{V}}_2{\mathrm{O}}_5$. Mixed phase vanadium oxide films were grown by sputter deposition in oxygen and nitrogen atmosphere at 300°C for 30 min. This study proves that changing the atmosphere of deposition will beget a changing in microstructure and so a variation in optical and thermal properties. Annealing in ${\mathrm{N}}_2$ atmosphere increases thermal conductivity and the gap energy. We have realized remarkable results with 95 nm ${\mathrm{V}}_x{\mathrm{O}}_y$ thin films annealed in ${\mathrm{O}}_2$ atmosphere, and we obtain both high TCR and low resistivity and a suitable optical and thermal properties for microbolometer application.