Invited Paper

MEMS-Based Uncooled Infrared Bolometer Arrays – A Review
Frank Niklaus
KTH - Royal Institute of Technology, Microsystem Technology Lab, 10044 Stockholm, Sweden and Faun AB, SE-18361, Sweden Tel: +46 8 76 216 73 49, Fax: +46 8 10 08 58,

Christian Vieider
Acreo AB, Electrum 236, 164 40 Kista, Sweden Tel: +46 8 632 78 45, Fax: +46 8 750 54 30,

Henrik Jakobsen
Vestfold University College, Institute for Microsystem Technology, 3103 Tønsberg, Norway Tel: +47 330 37 711, Fax: +47 330 31 03,

Uncooled infrared bolometer arrays have become the technology of choice for low-cost infrared imaging systems used in applications such as thermography, firefighting, driver night vision, security and surveillance. Uncooled infrared bolometer arrays are reaching performance levels which previously only were possible with cooled infrared photon detectors. With a continuously increasing market volume (> 100 000 units per year to date), the cost for uncooled infrared imaging chips are decreasing accordingly. In this paper we give an overview of the historical development of uncooled infrared bolometer technology and present the most important bolometer performance parameters. The different technology concepts, bolometer design approaches and bolometer materials (including vanadium oxide, amorphous silicon, silicon diodes, silicon-germanium and metals) are discussed in detail. This is followed by an analysis of the current state-of-the-art infrared bolometer technologies, the status of the infrared industry and the latest technology trends. Keywords: uncooled infrared bolometer arrays, microbolometer, MEMS, infrared detector, focal plane array, IR, FPA

1. Introduction
The word bolometer originates from the Greek word bolē and means ray-meter [1]. Bolometers are thermal infrared sensors that absorb electromagnetic radiation and thus increase their temperature. The resulting temperature increase is a function of the radiant energy striking the bolometer and is measured with e.g. the thermoelectric, pyroelectric, resistive or other temperature sensing principles. In the context of uncooled infrared imaging technologies, the term “infrared bolometer” usually refers to resistive microbolometers in which the temperature increase is measured by a resistance change. This review paper focuses on resistive microbolometers and does not include other thermal infrared detectors such as e.g. thermocouples [2], pyroelectric [3] and ferroelectric [4] uncooled infrared detectors. In the first bolometer, invented by the American scientist Samuel P. Langley in 1880 [5], a Wheatstone bridge was used along with a galvanometer that produced a deflection proportional to the intensity of radiation for small deflections. A later bolometer [6] consists of four platinum gratings, each of which is made of a series of strips, inserted in the arms of a resistance bridge. A number of other resistive thin film bolometers have been proposed from 1947 to 1980 [7-11]. The first thin film resistive micobolometers were proposed by R. Hartmann [12] and K.C. Liddiard [13, 14] in 1982 and 1984 respectively. Uncooled infrared microbolometers and focal plane array technology developed at Honeywell has been published since the late 80’s [15-22]. A number of books, book chapters and review articles on infrared technology, including uncooled infrared bolometers, have been published in recent years [23-29]. Uncooled infrared bolometers have become the dominating technology for the majority of commercial and military infrared imaging applications. Some of the most common infrared imaging applications are thermography, night vision (military, commercial and automotive),
MEMS/MOEMS Technologies and Applications III, edited by Jung-Chih Chiao, Xuyuan Chen, Zhaoying Zhou, Xinxin Li, Proc. of SPIE Vol. 6836, 68360D, (2007) · 0277-786X/07/$18 · doi: 10.1117/12.755128 Proc. of SPIE Vol. 6836 68360D-1

For microbolometers. predictive maintenance and industrial process control.mine detection. depending on the emissivity of the surface of the object. The infrared spectrum where the infrared transmission is allowed by the atmosphere is in the 3-5 µm wavelength region (mid wave infrared. a temperature increase of 1 K in the object typically results in a temperature increase in the bolometer membrane on the order of 4 mK [25]. MWIR) and in the 8-14 µm wavelength region (long wave infrared.75 µm to 1000 µm. Bolometers absorb the incident radiation and cause the thermally isolated bolometer membrane to increase its temperature. The radiation flux from real objects varies. y•pa CeLt Design Infrared Lens System Pitch Bolometer FPA Infrared Rays (a) (b) Figure 2: (a) Schematic drawing of a monolithically integrated infrared bolometer focal plane array (FPA) [32] and (b) placement of a focal plane array with an infrared lens system. fire fighting. 6836 68360D-2 . The temperature change correlates to the energy of the absorbed radiation and is measured by a change of the electrical resistance of the bolometer thermistor material. surveillance. 2. ranging from 0. The emitted radiation of a black body with a temperature of 300 K has an intensity peak at a wavelength of about 10 µm as can be seen in Figure 1b. Figure 2a shows a typical infrared bolometer pixel and a two-dimensional bolometer focal plane array. LWIR) as can be seen in Figure 1a. Ensembles of infrared detectors in two-dimensional arrays are called focal plane arrays (FPAs). There are also bolometers that are optimized for broad-band detection in both the 3-5 µm and the 8-14 µm wavelength region at the same time [33]. 4x0 lx 10'' Wv 2>110 I >110>1 'a (a) +4+4+++++ H±O Wavelength (microns) ca° caa 4 H±O Ca +4 4 Ca1 ID Absorbino Molecule 20 30 40 30 (b) Wavelength (Din) Figure 1: (a) Plot of atmospheric transmittance in part of the infrared region [30] and (b) radiant flux versus infrared wavelength for a black body with a temperature of 300 K [31]. of SPIE Vol. Proc. Infrared Bolometer Operation and Figures of Merit Infrared radiation is part of the electromagnetic spectrum with wavelengths above the visible spectrum. For the majority of infrared imaging applications the bolometers are optimized to detect radiation in the 8-14 µm wavelength region. reconnaissance. Figure 2b shows a schematic of a typical assembly of an infrared imaging system with a FPA and an infrared lens system. medical imaging.

the NETDthermal from the thermal fluctuation noise of the bolometers including temperature fluctuation noise from radiation heat exchange. when viewed by the infrared imaging system. 6836 68360D-3 . The noise equivalent temperature difference (NETD) is one of the most important performance parameters for infrared imaging systems and is defined as the difference in temperature between two side-by-side blackbodies of large lateral extent which. 34-40]. gives rise to a difference in signal-to-noise-ratio of 1 in the electrical outputs of the two halves of the array. the responsivity ( ℜ ). the NETDJohnson from the Johnson noise of the bolometers. For conventional infrared imaging systems based on uncooled bolometer arrays with column based read-out designs and integrated analogue-to-digital conversion (ADC). and the NETDROIC from the read-out integrated circuit (ROIC) related noise. The responsivity ( ℜ ) is defined as the output signal voltage or output signal current of an infrared bolometer pixel per incident radiant power on the pixel. Infrared imaging systems based on uncooled bolometer arrays can reach NETDs of below 25 mK with a F-number of the infrared optics of F = 1 [29. G is the thermal conduction between each bolometer and its surroundings.6) where F is the F-number of the infrared optics. The detectivity provides information that is equivalent to the noise equivalent power (NEP). of SPIE Vol. resulting in a signal to noise ratio of 1). 32. but with the possibility to compare bolometer pixels of the same type but with different pixel areas. the noise equivalent power (NEP) and the detecitivity (D*) [24.1) NETD 2 = NETD + NETD + NETD + NETD 1 f Johnson thermal ROIC where the total NETD consists of the NETD1/f from the 1/f-noise of the bolometers.e. φ is the transmission of the infrared optics in the wavelength interval from λ1 to λ2.5) VROIC ⎛ R ⋅R = Vamp ( xl ⋅ f i ) + + ⎜ ROIC bol ⎜R 12 ⎝ ROIC + Rbol λ 1−λ 2 VQ2 ⎞ ⎟ ⋅ I ROIC ( xl ⋅ f i ) 2 ⎟ ⎠ 2 (Eq. 42]. The noise equivalent power (NEP) is the incident infrared power on an infrared bolometer pixel that generates a signal output that is equal to the root-mean-square (RMS) noise output (i.4) NETDROIC ⎞ ⎛C⎞ ( RROIC + Rbol ) ⋅ 1 + ω 2 ⋅ ⎜ ⎟ ⎟ 2 F G ⎝G⎠ ⎟⋅ ⋅ ⋅ VROIC ⎟ β ⋅ A⋅ε ∆P ⎞ TCR ⋅U bias ⋅ RROIC ⎛ λ 1− λ 2 ⋅⎜ ⎟ ⎟ ⎝ ∆T ⎠ λ1−λ 2 ⎠ 2 2 (Eq.Important figures of merit for uncooled infrared bolometer and system performances are the noise equivalent temperature difference (NETD). viewing the two blackbodies. For the assumption that the bolometer temperature and the surrounding (background) temperature are equal. 25.3) NETDthermal ⎛ ⎜ = ⎜4⋅ ⎜ ⎜ φ λ 1− λ 2 ⎝ ⎛ ⎜ = ⎜4⋅ ⎜ ⎜ φ λ 1−λ 2 ⎝ ⎞ ⎟ T F G ⎟⋅ k ⋅ ⋅ 1 ⎟ β ⋅ A ⋅ ε λ 1− λ 2 C ⎛ ∆P ⎞ ⋅⎜ ⎟ ⎟ ⎝ ∆T ⎠ λ1−λ 2 ⎠ 2 (Eq. The detecitivity (D*) is defined as the root-mean-square signal-to-noise ratio of a 1 Hz bandwidth per unit root-mean-square incident radiant power per square root of bolometer area.2) NETDJohnson ⎞ ⎛ ⎛C ⎞ T1 ⋅ Rbol ⋅ xl ⋅ f i ⋅ 1 + ω 2 ⋅ ⎜ ⎟ ⎟ ⎜ 2 F G ⎝G⎠ ⎟⋅2⋅ k ⋅ ⋅ = ⎜4⋅ ⎟ ⎜ ∆P ⎞ TCR ⋅U bias β ⋅ A ⋅ ε λ1−λ 2 ⎛ ⎟ ⎟ ⎜ φ λ 1−λ 2 ⋅ ⎜ ⎝ ∆T ⎠ λ1−λ 2 ⎠ ⎝ 2 (Eq. 39] 2 2 2 2 (Eq. the NETD including the contributing NETD parts can be expressed by [38. 37. C is the heat capacity of the Proc. 41. the contributing NETD parts can be estimated by equations 2 to 6 ⎛x ⋅f ⎞ K ⎛C ⎞ ⎛ ⎞ ⋅ ln⎜ l i ⎟ ⋅ 1 + ω 2 ⋅ ⎜ ⎟ ⎜ ⎟ ⎜ f ⎟ 2 υ ⎝G⎠ F G ⎝ s ⎠ ⎟⋅ NETD 1 = ⎜ 4 ⋅ ⋅ ⎜ ⎟ β ⋅ A⋅ε TCR ⎛ ∆P ⎞ f λ 1− λ 2 ⎟ ⎜ φ λ 1− λ 2 ⋅ ⎜ ⎟ ⎝ ∆T ⎠ λ1−λ 2 ⎠ ⎝ 2 (Eq.

K is the 1/f noise constant of the resistive bolometer material. have a small cross-sectional area and consist of materials with a low thermal conductivity. a bolometer temperature sensing material with a high temperature coefficient of resistance (TCR) and low 1/f noise properties and a sufficiently low bolometer thermal time constant. The legs typically contain a thin metal layer to provide electrical contact between the bolometer material and the read-out electronics. VQ is the input reference analogue-to-digital quantization interval. 43. and the vias that connect the bolometer and the read-out integrated circuit wafer as can be seen in Figure 3a. Rbol is the bolometer resistance at the temperature T1. Ubias is the bolometer bias voltage. ε λ 1−λ 2 bolometer membrane in the wavelength interval from λ1 to λ2. that the bolometer pixels are as small as possible with reported pixel pitches being as small as 17 µm x 17 µm [33. typically a function of xl ⋅ f i ). a number of design features and tradeoffs have to be considered as can be seen from equations 1 to 9. and k is the Boltzmann constant. K In addition. T1 is the bolometer membrane temperature. To increase the bolometer pixel fill factor. The remaining area of the bolometer pixel is consumed by the bolometer legs. TCR is the temperature coefficient of resistance of the resistive bolometer material (dependent on the bolometer temperature). Bolometer Design and Manufacturing 3. IROIC is the current noise from the ROIC input including the bolometer bias is the temperature current source (depending on the ROIC architecture. 44]. Conventional single-level infrared bolometer arrays typically have a fill factor between 60 % and 70 % [32. Small bolometer pixels allow the implementation of high-resolution focal plane arrays at acceptable cost. ⎛ ∆P ⎞ ⎜ ⎟ ⎝ ∆T ⎠ λ1−λ 2 contrast in the wavelength interval from λ1 to λ2. the space between the bolometer and the neighboring bolometer membranes. Thermal conduction through the bolometer legs can be as low as 3. two-layer bolometer designs (umbrella Proc. a high absorption of the infrared radiation including a large absorbing area. At the same time it is important for commercial infrared imaging applications. To obtain a small thermal conduction between the bolometer and its surroundings.9) τ= G where τ is the thermal time constant of a bolometer pixel. υ is the volume of the resistive bolometer material. fs is the shutter or uniformity correction frequency for the bolometer array. The bolometer pixel fill factor defines the portion of the bolometer pixel area that is used to absorb the incident infrared radiation. f bol = 3. the bolometer legs are long. VROIC is the total noise voltage of the ROIC. 6836 68360D-4 . β is the infrared absorption rate of the is the bolometer pixel fill factor. Vamp is the input reference noise voltage of the ROIC (depending on the ROIC architecture.8) 4 ⋅τ C (Eq. The above NETD equations are valid for the assumption that the bolometer readout integration frequency fr is much larger than the thermal bolometer integration frequency fbol with (Eq. typically a function of xl ⋅ f i ).7) f r = xl ⋅ f i 1 (Eq. fr is the bolometer read-out integration frequency and fbol is the thermal bolometer integration frequency. Some of the most important bolometer design parameters are a low thermal conductance between the bolometer and its surrounding.01 mbar [45]. xl is the amount of bolometer pixels per column (number of bolometers that are read-out during one imaging frame). of SPIE Vol. RROIC is the input impedance (resistance) of the ROIC. A is the bolometer pixel area (pixel pitch). The vacuum atmosphere in which the bolometers are operated is typically on the order of 0. Figure 3a shows details of a typical bolometer design. fi is the image read-out frequency (imaging frame rate). 37].bolometer pixel.5 ⋅ 10 −8 W [32]. conventional bolometers are operated in a vacuum package to minimize the thermal conduction between the bolometers and their surroundings through the surrounding gas. The cost for both the FPA chip and for the infrared optics is reduces by reducing the active area of the FPA.1 Geometrical and optical bolometer design To design uncooled infrared bolometer arrays with a high sensitivity (low NETD). ω is the modulation frequency of the infrared signal from the image scene.

45. titanium nitride) of the Fabry-Perot cavity usually has a target sheet resistance of 377 Ω/sq to reach optimum performance of the FabryPerot cavity in the wavelength interval of 8 µm to 14 µm [48]. 6836 68360D-5 . — 10 05 0 (a) (b) Figure 3: (a) Details of a typical single-level bolometer design [24] and (b) example of a two-level (umbrella) design for improved pixel fill factor [33]. The membrane thickness is typically set to λ x . which in turn can be used to minimize the thermal bolometer conductance and the resulting NETD. The mirror of this type of resonant optical cavity is placed at the lower surface of the bolometer membrane and the thickness of the bolometer membrane defines the resonant optical cavity. where λ x is the wavelength of the targeted infrared 4 radiation in the bolometer membrane material(s) [48]. Antireflection Layer Bolometer Membrane Free Space Infrared Mirror λ 4 To provide a high absorption of the radiation in the bolometer membrane. Most commercial bolometer FPAs make use of the optical cavity design shown in Figure 4a since it allows thinner bolometer membranes with a lower heat capacity. In monolithic integration. 49] as shown in Figure 5. in which the resonant optical cavity is part of the bolometer membrane. conventional bolometers contain resonant optical cavity (Fabry-Perot) structures that are optimized for the targeted wavelength interval as depicted in Figure 4.designs) as shown in Figure 3b have been reported that reach fill factors of up to 90 % [46. The most commonly used resonant optical cavity design is shown in Figure 4a. 42.g. For a targeted wavelength interval of 8 µm to 14 µm. the ROIC is pre-manufactured and the bolometer Proc. of SPIE Vol. 42]. 3. 41. The top metal layer (e. the distance between the bolometer membrane and the mirror on the substrate is typically about 2 µm to 2. Thus. a high fraction of the 4 incident infrared radiation at a specific wavelength λ is absorbed in the bolometer membrane [37]. and in some cases the temperature sensing material [33] are placed underneath the absorbing bolometer membrane. Such umbrella designs have been implemented in bolometer FPAs with very small pixel sizes [33. 47].2 Bolometer focal plane array manufacturing techniques The most commonly used manufacturing approach for uncooled infrared bolometer FPAs is monolithic integration [37. A second type of resonant optical cavity design is shown in Figure 4b.5 µm [37. bolometer membrane is placed at a distance of λ from the mirror surface on the substrate. in which the infrared mirror (typically aluminium) of the resonant optical cavity is placed on the surface of the underlying substrate (the ROIC) and the Via Antireflection Layer Bolometer Membrane Infrared Mirror λx 4 Via d d Substrate Wafer (ROIC) Substrate Wafer (ROIC) (a) (b) Figure 4: Cross-sectional image of two bolometer designs with resonant optical cavities for high absorption of the incident radiation. The bolometer legs. 46].

Bulk micromachining is a second alternative to manufacture uncooled infrared bolometers [50] as shown in Figure 6. One potential disadvantage of monolithic integration is that the deposition process for the temperature sensing bolometer material is limited to about 450ºC and does not allow the deposition of monocrystalline materials. of SPIE Vol. (b) deposition of bolometer materials. Thin-Film Resistive Thin-Film Thermistor Bolometer Material Material Handle Wafer Polymer Adhesive Etch-Stop Layer Handle Wafer Via Landing Pads Bolometer Vias ROIC Wafer (a) ROIC Wafer (b) ROIC Wafer (c) ROIC Wafer (d) ROIC Wafer (e) ROIC Wafer (f) Figure 7: Heterogeneous 3D integration for uncooled infrared bolometer arrays: (a) separate fabrication of ROIC wafer and handle wafer with resistive bolometer material. Heterogeneous three dimensional (3D) bolometer integration has been proposed for the integration of high-performance mono-crystalline temperature sensing bolometer materials on ROICs [48. This can make the optimization of the temperature sensing bolometer material difficult. Process temperatures higher than 450ºC risk damaging the ROIC. Subsequently. (b) selective etching of the bulk material underneath the bolometer membrane [50]. the polyimide layer is sacrificially removed in an oxygen plasma to obtain the free-standing. Monolithic integration is a cost-efficient and well-established post CMOS process in which the electronics for the signal read-out can be efficiently placed underneath and beside the bolometer membranes. a high-temperature stable polyimide is used as the sacrificial layer. the substrate is selectively etched underneath the bolometers to thermally separate them from the rest of the substrate. in between or after processing the wafers e. In bulk micromachining.materials are subsequently deposited and patterned on the ROIC wafer. This usually reduces the bolometer fill factor. but has to be placed beside the bolometers. (b) adhesive wafer bonding. Commercial infrared bolometer arrays that are manufactures using bulk micromachining techniques are diode bolometers. the bolometers are formed in the substrate surface of a wafer. (a) (b) Figure 6: Bulk micromachining for uncooled infrared bolometer arrays: (a) Formation of the bolometer and the electronics for signal read-out (typically side-by-side). The advantage of bulk micromachining is that the electronics and the bolometers can typically be manufactured in a standard CMOS line. Sacrificial Layer Bolometer Materials Via Landing pads Bolometer Vias Bolometer Membrane ROIC Wafer (a) ROIC Wafer (b) ROIC Wafer (c) ROIC Wafer (d) ROIC Wafer (e) Figure 5: Monolithic integration for uncooled infrared bolometer arrays: (a) deposition of sacrificial layer on ROIC wafer. The materials are then transferred from the handle wafer to Proc. The free-etching of the bolometers can then be done in a single post CMOS process step. (c) patterning of bolometer materials. Typically. All commercially available vanadium oxide and amorphous silicon bolometer FPAs are monolithically integrated on top of the ROICs. (d) bolometer definition. In the final step. A disadvantage of bulk-micromachining is that the electronics for the signal readout can not be placed underneath the bolometer membranes. (c) thinning of handle wafer. (e) via formation. In 3D bolometer integration.t i a e n-well. 6836 68360D-6 . thermally isolated bolometer membranes. the bolometer materials are deposited on a separate handle wafer. 51-56]. (f) sacrificial etching of polymer adhesive.g. The bulk micromaching processes can be implement before. in a CMOS line to implement the necessary electronic components. (f) etching of sacrificial layer. 0. (d) via formation.

however.3. the most common bolometer temperature sensing materials are vanadium oxide (VOx). A high temperature coefficient of resistance (TCR) and a small 1/f noise constant are desirable material properties as can be seen from equations 1 to 9. Bolometers made of α-Si can consist of very thin memebranes. TCR up to about 3 %/K at room temperature have been reported [69]. of SPIE Vol. Singlecrystal VO2 and V2O5 have TCR above 4 %/K. Syllaios et. Because high electrical resistance of a device results in a high level of noise. 3. Today. They undergo transition from an insulator or semiconductor to a metal phase at a specific temperature. in 2000 [68].3 Bolometer temperature sensing materials The selected bolometer temperature sensing material has a large influence on the sensitivity (NETD) of the bolometer. it must be possible to integrate the temperature sensing material together with signal read-out electronics (e. V and V2O5 by a conventional sputter method. 36.3. The hydrogenated amorphous silicon (α-Si) microbolometer arrays that have been developed take advantage of the high TCR. al. There are many phases in vanadium oxides. relatively high optical absorption coefficient and the advantage that they can be manufactured using silicon fabrication compatible processes. mono-crystalline materials can have a significantly lower 1/f noise constant as compared to amorphous or poly-crystalline materials. Several groups are active to find processes to obtain high TCR in combination with sufficient low sheet resistance for good noise performance. VITO. A 3D bolometer integration process is currently being implemented in a project with consortium partners Autoliv.the ROIC wafer using low-temperature adhesive wafer bonding in combination with sacrificial removing of the handle wafer as shown in Figure 7c. but its resistance at room temperature is very high. Umicore. For most materials 1/f noise constant is not very well documented in literature. The performance limits of uncooled VOx microbolometer focal plane arrays are reported in [32]. including the annealing and oxidation of evaporated vanadium under controlled conditions.g. Lv et. Y. Y-H. M. V2O5 can be formed by ion beam from vanadium metal target in high O2 partial pressure. the use of the V2O3 phase showing low resistance is important to the fabrication of low-noise microbolometers. Soltani et. It is reported [64-66] that bolometer characteristics can be improved by using vanadium-tungsten oxide made by low temperature oxidation of vanadium-tungsten metal films. 58] and is today used in a variety of bolometer products. Infineon Technologies SensoNor. [62] have shown that through the formation of a sandwich structure of V2O5. al. 6836 68360D-7 . such as VO2. 60] and pulsed laser deposition [61]. amorphous silicon (α-Si) and silicon diodes. The advantage of 3D bolometer integration is that it allows the use of highperformance mono-crystalline temperature sensing bolometer materials on top of standard ROICs. 38. Vanadium oxide films can be prepared by a variety of methods.12 %/K. 3. but are difficult to make.1 Vanadium oxide (VOx) The vanadium oxide thin films that are used have TCR in the range 2 %/K and 3 %/K at room temperature [57. Proc. [63] have obtained TCR value about 4. 36]. Thus. They achieved films with TCR 5. for bolometers with a low thermal conductance while maintaining a fixed bolometer time constant. one way to reduce the NETD of a bolometer FPA may be the use of mono-crystalline temperature sensing bolometer materials with a low 1/f noise constant [32. The achievable bolometer dimensions are identical to the ones made with conventional monolithic micromachining techniques. a CMOS wafer) in a cost efficient way. A description of amorphous silicon microbolometers technology was published by A. J. so its resistance is very low at room temperature. small area solar powered photovoltaic devices for consumer products and large area power solar cells. Commonly used thin film depositions techniques such as evaporation and sputtering will give amorphous or polycrystalline films. 39]. Han et. 3.4 %/K and low sheet resistance <20 kΩ/square by using facing target sputtering and heat treatment. At the same time. KTH-MST and Acreo [53-55]. which allows for a low thermal mass and consequently. V2O5 and V2O3. al. [67] report a strong effect of Ti–W co-doping on both the optical and electrical properties of vanadium dioxide thin films grown by a reactive pulsed laser deposition. The NETD of state-of-the-art uncooled infrared bolometer FPAs is typically limited by the 1/f noise from the bolometer temperature sensing material [32.2 Amorphous silicon (α-Si) Amorphous silicon (α-Si) is extensively described and used in a large variety of products such as active layer in thin-film transistors for liquid crystal displays. V2O3 has low formation energy and undergoes a transition from semiconductor to metal phase at low temperature. a controlled mixed phase of VOx with good electrical properties by can be formed by post-annealing in oxygen. al. reactive RF sputtering [49. The 1/f noise constant is a material parameter that can vary several orders of magnitude for different materials and even small variations of the material composition can dramatically change the 1/f noise constant [35].

but plastic as well. The TCR and sheet resistance depends on material properties such as doping concentration. reduced-pressure chemical vapor deposition (CVD). e. or ultra-high vacuum vapor deposition [80].3. allows standard metallization annealing temperatures [54. The Ge content is in the order of 85%. Titanium up to 0. 3. [70. Wide-band gap materials have also been investigated. Other reported alternatives for poly-SiGe film are molecular-beam deposition [79]. Another type of thermistor materials are semiconducting YBaCuO. VOx due to the higher 1/f noise of the poly crystalline material. The detector performance that can be obtained is up to now lower than e. 3. The TFT used had a high temperature coefficient of its drain current in the range 1.Hydrogenated amorphous silicon (α-Si) is a thin–film amorphous material that can only be produced by a nonequilibrium process. but the relatively high 1/f noise lower the potential detector performance. 85]. of SPIE Vol. Important requirements for the packaging of bolometer arrays are: good and reliable hermetic seal. TCR values up to 5. Due to the high deposition and annealing temperatures has the membrane however to be manufactured before metallization [78].4 Other Bolometer Materials A simple alternative for temperature sensing material are thin film metals. and high yield low cost packaging.7 V for silicon pndiodes and about 0. 75. as low as 75 °C. α-Si can be deposited at very low temperatures. 76]. TCR can be controlled with the deposition parameters and e. and TCR values between 2.7 and 2. Dong et. the high quality crystalline material provides very low 1/f-noise characteristics. Deposition can be made at low temperature with magnetron sputtering.4 Bolometer vacuum package Bolometers need to be packaged in vacuum for best performance. conventional bolometers operate with vacuum levels < 0. A recently reported material with performance claimed to be compatible with VOx is thin-film carbon. With a forward voltage of about 0.g. The use of amorphous silicon thin film transistors (α-Si TFT) as the active elements has been reported by L. The TCR can thereby be high (3. 6836 68360D-8 . Perovskite metal-oxide manganites with colossal magneto resistance effect (CMR) have also been proposed for thermal imaging.3 Silicon diodes The temperature coefficient of the forward voltage of a pn-junction or Schottky barrier junction can be used as sensing principle for the temperature caused by absorption of IR radiation. high-volume applications. Titanium is preferred due to its low thermal conductance. Another alternative is poly crystalline silicon germanium (SiGe) with approximately 30% Ge. TCR is unfortunately also very low (e. e.3. At room temperature has a TCR of 4.g. The properties are therefore a direct consequence of the deposition process. For MEMS based Proc. Very high TCR values can be obtained around the phase transition temperature when changing from semiconductor to metal. It is made by depositing a parylene layer and pyrolyze it to carbon at 800°C [87].35 %/K [72]).g.g. The relatively lower electronic performance of low-temperature α-Si devices could be compensated by the cheaper production. making it a candidate for a roll-to-roll processing technique. an order of magnitude lower than the best resistive devices.4 %/K been obtained [86]. The noise level increase however with higher TCR and a trade-off has to be made to achieve maximum performance.5 to 6. which allows for deposition on not only silicon and glass. 71]. Advantageous is also the stability of the quantum well thermistor material.g. Both reliability and cost of MEMS is heavily related to the encapsulation techniques chosen.6 V for silicon Schottky-diodes and a voltage change with temperature in the range 1-2 mV/K give a temperature coefficient of about 0.g. An alternative to α-Si are different types of amorphous germanium-silicon-oxygen compounds (GexSi1-xOy) grown by reactive sputtering in an Ar or Ar:O2 environment [74.01 mbar [47]. The advantage is lower thermal conductance. They are easy to integrate with CMOS ROIC process and the 1/f noise is low. which e. 3. which results in low performance detectors [73]. or by plasma enhanced chemical vapor deposition [77]. Very thin membranes (100nm) can be produced using standard CMOS processes. PECVD deposited amorphous silicon carbide [82] and photochemical vapor deposition of nanocrystalline silicon carbide [83]. 81]. deposition temperature and annealing. An advantage of using diodes may be the possibility this gives for smaller pixel size because diodes can normally be made on smaller areas than resistors. Therefore.2 %/K.8 and 4 %/K have been reported [84. al. The material can also be designed with pyroelectric properties depending on the oxygen stoichiometry. integration of IR window material with good infrared transmission. The bolometer diodes can potentially be manufactured in standard CMOS lines. ultra-low-cost. be varied between 0. By growing very thin (nm) Si/SiGe multi-layers epitaxially is it possible to create a single crystalline material with integrated valence barriers. for future.3 %/K) and simultaneously.5 %/K at room temperature. such as plasma-enhanced chemical vapor deposition (PECVD) or sputtering.1 %/K have been reported.3 %/K.

of SPIE Vol. This technology allows the pixels to be packaged in a hermetic cavity with the electrical contacts outside the cavity at low cost. automotive and environmental monitoring. industry. X. we develop a wafer scale package to form lids with silicon IRwindow over the bolometer pixel [55]. As a result. Hereby vacuum pressure < 0. Silicon lid wafer with IR coating Silicon lid wafer with IR coating Wire bonding pads Pixels Pixels Substrate wafer (ROIC) Substrate wafer (ROIC) (a) (b) Figure 8: Principle cross-sectional of wafer level packaging of silicon lid wafer with window areas onto ROIC wafer with bolometer pixels (a) after wafer level sealing and (b) after release of wire bonding areas by sawing. packaging may be done at chip level or wafer level. The window areas are made by deep reactive-ion etching. Establishing manufacturing schemes by using accelerated testing for checking the actual gas pressure and bonding consistency will hereby be made possible at low cost. The need for packaging of microbolometers in vacuum is related to the fact that heat loss due to thermal conduction from the bolometer structure through the gas gap to the substrate underneath will increase the NETD as indicated in equations 1-9. 6836 68360D-9 . The cap-wafer is then aligned and attached to the ROIC wafer with the pixels and sealing take place in vacuum. He et al. Several alternative methods for wafer level packaging are already developed by the MEMS industry. Hermetic methods include low temperature fusion bonding. Figure 8 shows the principles of the wafer scale packaging technology that is being developed. Good descriptions of these methods will be available in a new handbook for MEMS [89]. Standard antireflection coating is deposited on both sides of the window areas to obtain windows without any significant degradation of the final effective NETD.1 mbar for a device with pixel area 50 µm x 50 µm and an air-gap of 20 µm. The inherently good detector performance developed allows us to use a lower vacuum without compromising overall performance too much. eutectic bonding and soldering. the areas on the cap-wafer around the pixel areas are removed by standard wafer dicing saw in order to release the wire bonding pads. hereby reducing the system cost substantially. anodic bonding. The cap-wafer material (window) is made in single-crystal silicon wafers with low oxygen concentration to obtain good IR transmission over the whole wavelength range 8-14 µm. high volume applications such as home. For a specific new volume automotive application. Proc. [88] have measured the typical thermal conductance and heat capacitance of a microbolometer as a function of pressure at room temperature on test structures. and metal based bonding that includes thermo-compression bonding. A seal ring is added on the frame areas of the ROIC wafer. Their results shows that thermal conduction through the gas starts to have an effect from a pressure in the range 0. At higher pressure levels the thermal conductance Gair through the air-gap can be estimated to be: Gair = kair A/d (Eq.bolometers. Wafer-level vacuum packaging may be an important factor in bringing commercialization of micro-bolometer arrays into low-cost. Development of specific methods to measure the quality of the vacuum both on wafer level and in the final devices are investigated. this allows the bolometer chip to be integrated directly to the camera housing.01 mbar can be obtained and open up for applications requiring better NETD. The process sequence for the wafer scale packaging allows building of additional thin-film getter structures and Pirani gauge vacuum sensors into each device. The most common way of packaging is to build the bolometer chip into a hermetic metal or ceramic package with an IR-transmitting lid build into the package cap. A is the area of the bolometer plate and d is the distance between the plate and the substrate. This enables good long term stability and makes the use of getter materials and custom processes obsolete. 10) where kair is the thermal conduction for the air. After sealing. glass frit bonding. An important advantage for MEMS based bolometer arrays is the possibility this technology offers for low cost wafer level vacuum packaging with methods that has become available the last few years.

of SPIE Vol. Bolometer Read-Out Integrated Circuits (ROICs) Bolometer read-out integrated circuits allow measuring the very small resistance changes of individual bolometers that are part of large bolometer arrays. Commercial and State-of-the-Art Infrared Bolometer Arrays Today. USA ULIS. Commercially available bolometer arrays are either made from vanadium oxide (VOx). Figure 9a shows scanning electron microscope (SEM) images of a commercial VOx bolometer [32] and Figure 9b shows a VOx umbrella type bolometer [33] that is in the R&D stage. with VOx being the dominating technology.320x240 320x240 . Highperformance uncooled infrared bolometer FPAs are subject to export control under the Wassenaar agreement (www.640x480 320x240 320x240 320x240 . To measure the resistance change of a bolometer.5 30 28 R&D: 17 25 R&D: 17 25 R&D: 17 25-50 25 23. today many infrared cameras do not require a TEC for temperature stabilization for the FPA chip.640x480 320x240 160x120 . Column parallel read-out architectures with integrated AD conversion are commonly used in commercial FPAs [32. Other suppliers include Raytheon in USA. 90-92]. some of the FPAs with very small bolometer pixels that are in the R&D stage. Table 1: Commercial and state-of-the-art R&D uncooled infrared bolometer arrays. Japan SCD. As can be seen from equations 1 to 9.4. modern FPAs have implemented design features that tolerate and compensate for a floating chip temperature in the specified temperature range [92]. Mitsubishi in Japan. L-3 in USA. However. BAE Systems in USA and DRS in USA. the dominating suppliers of high-performance uncooled infrared bolometer arrays are Flir Systems in USA. the ROIC is required to handle a higher dynamic range. 640x480 320x240 320x240 384x288 Pixel pitch (µm) 25 37. Company FLIR. 6836 68360D-10 . 5.640x480 640x512 160x120. Thus. Early FPAs used thermo-electric-coolers (TEC) to stabilize the FPA chip at a fixed temperature and to provide a low noise voltage. USA France Mitsubishi. the ROIC applies a bias voltage which causes a temperature increase (offset) in the bolometer membrane. the noise voltage from the ROIC should be sufficiently low to not be the domination noise contribution to the NETD. USA BAE. Ulis in France. However. NEC in Japan and SCD in Israel. the bolometer technology and specifications for existing products and for bolometer arrays that are in the R&D stage.5 25 Detector NETD (F=1. Japan NEC. USA L-3. As a result of the bolometer self-heating. 45]. USA Bolometer type VOx bolometer VOx bolometer α-Si bolometer VOx bolometer VOx bolometer (standard design) VOx bolometer (umbrella design) VOx bolometer (umbrella design) VOx bolometer VOx bolometer (umbrella design) α-Si bolometer Si diode bolometer VOx bolometer VOx bolometer Array format (pixels) 160x120 .640x480 160x120 . 20-60 Hz) 35 mK 50 mK 50 mK 30-50 mK 50 mK 35 mK 50 mK 30-40 mK 50 mK 35-100 mK 50 mK 75 mK 50 mK DRS. Table 1 shows an overview of the main suppliers. Most commercially available bolometer FPA make use of standard single-level bolometer designs.wassenaar. consist of two-layer (umbrella-type) bolometer designs [33. Israel Proc. amorphous silicon (α-Si) or silicon diodes.

80x30 pixel arrays with a NETD of about 150 mK @F1. the cost for uncooled infrared imaging chips are decreasing accordingly. 44] to low cost production [54. With a continuously increasing market volume (>100 000 units per year to date). Conclusions Bolometer arrays have become the technology of choice for low-cost infrared imaging systems used in both civil and military applications. new temperature sensor materials with low 1/f noise and high Proc. The main development issues are: 1. Only standard materials and fabrication processes that can be used in CMOS and MEMS foundry fabrication. State-off-the-art performance is for many applications good enough. and the production yield and testing requirements will be of larger importance for the final production cost. of SPIE Vol. First level packaging needs to be integrated in the chip fabrication process using wafer-level technologies and integrated getter materials. e. For the future we can divide the technology requirement and thus development efforts towards the following main market segments: • High resolution. machine vision and consumables. 3. This will open up for many producers and in long term also lower manufacturing cost for smaller series production. Outlook and Future Developments Development of micro-bolometer technology has the last 5 years continued to dramatically improve performance and resolution. 94]. This puts additional requirement on high temperature stable detector material. microbolometer detectors are now produced in larger volumes than all other IR array technologies together. 6. Camera module integration.g. low performance image based sensing for automotive safety. Mass-market application will become the technology driver. Low resolution arrays will be completely integrated with driving electronics and optics. 43. 93. 6836 68360D-11 . High-end products for niche markets will in the long run benefit from the achievements and manufacturing infrastructure resulting from the development of mass-market products. • Low resolution. To increase detector sensitivity.(a) TJ — —j s_-a -. Uncooled infrared bolometer arrays are reaching performance levels which previously only were possible with cooled infrared photon detectors. 2. The demonstrated performance is getting closer to the theoretical limit and the gap to photonic detectors is getting smaller. Initial development cost and complexity of the manufacturing will be of less importance for large volume production. power consumption and cost of uncooled technology. This changes the development focus from performance and resolution [33. medium performance imaging for automotive and security – 320x240 pixel arrays with a NETD of about 35 mK @F1. • Medium resolution. s__a '1 (b) Figure 9: (a) Commercial VOx bolometer design with 28 µm x 28 µm pitch from BAE [32] and (b) R&D VOx bolometer with umbrella design and 17 µm x 17 µm pitch from DRS [33]. Vacuum packaging. similar as digital CMOS cameras are produced today. With the advantages regarding weight. high performance imaging for surveillance and military – mega pixel arrays with lowest possible NETD. Foundry manufacturing. The next generation detectors will have a pixel pitch of 17 µm.

K. Proc. pp. C. pp. pp. Review of Scientific Instruments.29. H.6.42.74-76.19. 377-385. Proc. Higashi.195-206. Heterogeneous 3D integration processes will be developed to solve this issue. pp. R. pp. W.S. Liddiard. M. N. Vol.E.A. R. Higashi.A. Paul.2.E. making post-processing deposition impossible. Vol. F.J.S. pp. Schneeberger. 1956. A. “Thin-film resistance bolometer IR detectors”. Clarendon Press. Han.F. “Ferroelectric IR bolometers . Waelti.H. 1991. “HIDAD – a monolithic silicon uncooled infrared imaging focal plane”. R. Higashi.26. Han. R.C. Infrared Physics. Electronics. Oxford. R. Vol. Proc.W. The manufacturing needs to be aligned with standard CMOS and MEMS foundry processing. The most promising materials are fabricated at relatively high temperature. N. 1989. Infrared Physics. E. Munch. Muralt.10.43-49.57-64. Gosch. 1957. R. Yosihara. 1984. Proc. N. 1990.TCR are required. J.184. Liddiard.A.18.E. Vol. B. C. No.1339-1388. M.J. [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] Proc. O. “Advances in uncooled silicon monolithic IR 2D arrays”. Ferroelectrics. pp. Wood. No. Philips. Holmen. T. This requires trade-off optimization of the performance related to cost for each part in the system. W. E. Weinstein.24. R. No. Selders. D. IRIS DSG. pp. Doering. C. Vol. 2001. Science of Light. Hyde. P. IRIS DSG. IRIS DSG. 1980. USA.132-135. Smith. "An investigation of the properties of bolometers made by vacuum evaporation". “Large-area bolometers of evaporated gold". 1991.4.141-150. Review of Scientific Instruments.1. Lubke. Proc. Cole. R. Han. B. R.J.E.429-435. This is required for large volume fabrication and the technology development progress will even faster benefit from the achievements continuously gained in the whole semiconductor field. No. Menolfi.503-510.E.E. Essen. Ohnstein.A. Nielsen. Brand. Billings. Barr. Wood. Cole. Hilton Head Island. Proc. “Micromachined infrared detectors based on pyroelectric thin films”. R. Vol. B. J. Wood. Vol. Vol. C. Cole. W. West Germany. “Uncooled low-cost thermal imager based on micromachined CMOS integrated sensor array”. Solid-State Sensor and Actuator Workshop 1992. Journal of Scientific Instruments. Reports on Progress in Physics. Loepfe. Acknowledgments The authors thank all co-workers in the Eurimus-project “Pedestrian Injury Mitigation System . “The Infrared Pioneers – III. R.E. Qiuting. Vol. IRIS DSG. C.from ceramic hybrid arrays to direct thin film integration”. 28. “The production of film type bolometers with rapid response”. pp. Feb.A. ”Construction and characteristics of evaporated nickel bolometers”.A. “Thin-film resistance bolometer IR detectors – II”. B.E. Kruse. C. R. 1986. Barr. 1947. 1946. Proc. R. 6836 68360D-12 .A. References [1] [2] Websters Dictionary. Samuel Pierpont Langley”.B. Baltes. Brown. Aiken. Hartmann. Blevin.A.E. 2001.3. of SPIE Vol. R. K. Vol. Carney. Carter. Schaufelbuhl. Special thanks to John Franks and Jan Van Nylen from Umicore for providing the illustration in Figure 2b and to Jan-Erik Källhammer from Autoliv for valuable feed-back for improvements of the paper. pp. Journal of Microelectromechanical Systems. M. R.L. Infrared Physics.E. Jones.17. O. Chasmar. Carlson. Foss. K. 1965.C. Wood. R. R. P.J. Mass-market applications will focus the development on low-cost manufacturing and integration of the whole IR camera module.H. pp.A. pp.102-116. Wood.PIMS” for the fruitful cooperation. “Monolithic silicon uncooled focal planes for the HIDAD program”. Han. Wood. 1988.J. Foss. Watton.64. Sensor 1982 – Sensor Technology and Temperature Measurement. 1996. 1963. “Monolithic silicon uncooled focal plane development for high density array development program (HIDAD)”. No. As example can the requirement on a very low vacuum pressure be exchanged with a better performing detector material. J. “Progress in silicon monolithic uncooled IR focal plane arrays”. “High-sensitivity thin-film bolometers”. R. U. “Thin-film enhances bolometer's sensitivity". Higashi. The Detection and Measurement of Infrared Radiation. pp.10. GOMAC.A. Higashi. H. “Integrated uncooled infrared detector imaging arrays”. E. W.5. A. F.

USA.J.W.6542. 2000. No.-E.200-207. Richwine. pp. Hewitt. 1999. Journal of Optical Technology. Kruse. Pettersson. USA. K. Ito. Infrared Detectors.infrared. USA. J. Geneczko.4820. Miyoshi. Leary. Soodb.153-157. Van Lue. http://dx. Russia. A.W. “A comprehensive model for bolometer element and uncooled array design and imaging sensor performance prediction”. of SPIE 2006. Proc. Hewitt. SPIE 2004. 1998. Brand. Akin. M. Infrared Physics and Technology. T. Orlando. “Uncooled microbolometer infrared sensor arrays”. Wyles. Proc. Sessler. Wood. Howard. Wyles. A. N. “CMOS-based Thermal Sensors”. [30] Source: U. J. Ponomarenko.5406. 2004. E. Germany. Vol. L.L. [32] M.2. pp. pp. Vol. “Performance Model for Uncooled Infrared Bolometer Arrays and Performance Predictions of Bolometers Operating at Atmospheric Pressure”. R. Weinheim.1M. Tissot. R. Kruse. A. Martin. Berlin. Blackwell. C. San Diego.5074. Moscow. 52-85. Orlando. “New thermally isolated pixel structure for highresolution uncooled infrared FPAs”. I. Proc. Semiconductors and Semimetals. Vol. “Infrared detectors: an overview”. Kennedy. Orlando. Pejnefors. SPIE 2007. E.L. S. SPIE 2003. Orlando USA. Skatrud. Vol. Freyvogel. Elliott. USA. SPIE 2004. 2001.45-121. R. [26] T. E. [40] R. Proc. USA.3. P. D. Vol. D. Kohin. 1993.5406. J.D. 1997.wikipedia. Infrared Physics and Technology.363-378. Kohin. Boston. Murphy. CMOS . pp. San Diego. Rogalski. USA. Tabata. Wood. Vol. M. Niklaus. Wan. E. B. Boston. Han.T.F. M.A.1-1Y. June. T. J. Proc. Foss. C.12. C. USA. KGaA.548-556. [47] D. Bisotto. I. “Performance improvements for VOx microbolometer FPAs”.6294. pp.6542. Fedder. C. Wyles. J. G. Proc. Ajisawa. Jansson. T. Korvink. USA. ”Calculational modeling of the main characteristics of an uncooled linear microbolometer array”. “Performance improvements for VOx microbolometer FPAs”. Bellingham. Orlando. Vol. Mottin. Stemme. Tissot. Kurashina.1Y. pp. [24] R.J. J.6542. Kruse. USA. Orlando. N. J. S. C. SPIE 1993. SPIE Press. V. USA. Wallgren. WILEY-VCH Verlag GmbH & Co. Infrared Detectors and Emitters: Materials and Devices. USA. [42] D. Hierold. Vol. [45] J.149-174. pp. Ouvrier-Buffet.A. D. Orlando. Gordon. [43] R. Ray.T. Ray. Clarke “Recent development of ultra small pixel uncooled focal plane array at DRS. Rapach. Murphy. pp. [22] R. Proc. Niklaus. Wood. Sasaki. Gordon.L. J. SPIE 2007. [48] F. Micromachined transducers sourcebook. J. Advanced Micro and Nanosystems. Wood. L. “Uncooled IR focal plane arrays”. 2000. S. 2007. J. Robas. “Uncooled Infrared Bolometer Arrays Operating in a Low to Medium Vacuum Atmosphere: Performance Model and Tradeoffs”. M. in press. Emmett. Orlando. M. Asbrock. Vol. Proc. Proc. D. Gordon and Breach Science Publishers. G. pp. [25] P. M. pp. “640 × 512 17 µm microbolometer FPA and sensor development”.5406. Niklaus. “Micromachined bolometer arrays achieve low-cost imaging”. C. CH. Arrays. Sessler. 5406. No. M. Khrebtov. Peysha. Malyarov. “17 µm pixel 640 x 480 microbolometer FPA development at BAE Systems”. R. [29] A. 5406. AMAA 2003. pp. Proc. 531-540. SPIE 2004. USA. J. S.7. “Uncooled amorphous silicon technology enhancement for 25 µm pixel pitch achievement”. Wyles. Van Lue. Proc. Baur. McGraw-Hill. 6836 68360D-13 . [38] F. SPIE 2004. J. pp.62940F1-62940F11.A. Balceraka.2020. Ray. Rumbaugh. Dereniak. I.Y. 447-453. “Infrared microbolometer sensors and their application in automotive safety”. [23] P. Yon. Williams. G. Vol. Biancardini.08. Seattle. Li. Decharat.A. in P. Kruse. [34] P. http://en. pp.1016/j. Kluwer Academic Publishers. pp.47. Vol. Letellier. Buttler. SPIE 2004. “Infrared focal plane arrays: state of the art and development trends”. G.65421U.-E. Zerov. [37] E. Vol. O'Donnell. D. SPIE 2003. “Performance limits of uncooled VOx microbolometer focal plane arrays”. 1-15.L. Systems. Backer. Asbrock. 5126. C.W. Yon. E. Vol. SPIE 2007. Wyles. “Characterization of transferbonded silicon bolometer arrays”.A. J. C. L.[21] R. Van Lue. C. D. pp.MEMS. USA. “Monolithic silicon microbolometer arrays”. G. [41] B. Uncooled Thermal Imaging.G. Ushakova. E.D. pp.437-446. Filachev.J. [39] F. P. 5406. Dainese. Jansson. Marasco. pp. Sessler. R. Vol. Gordon. [36] P. [27] A. A. Proc. K. 428-436. “Can the 300 K radiating background noise limit be attained by uncooled thermal imagers ?”. M.J. pp. S. G.S. pp. USA. Baur.2007. Joswick. Trujillo. Taubkin. USA. H.001. Kennedy.521-530. O. [31] G. Hewitt.L. Proc. Decharat. 2005. Laser Focus World.6542. Bain. T.A Kovacs.101-106.253-258. M. pp.531-540. M. [28] A Rogalski. Tohyama. C. Källhammer.W. C. A. A. in P. pp. Baltes. Amsterdam. Proc. “Uncooled infrared sensor performance”. [44] D. Oda. USA. Bach. K. in H. N. J. J. Proc. pp. Black. C. Opto-Electronics Review. Academic Press. Baur. Stemme. Orlando. Capper. 2002. T.doi. J. Murphy. U. [46] S. Skidmore. Stemme.7. Orlando. Vol. of SPIE Vol. and Applications. Wyles. SPIE 2004. Vol. [33] C. Navy. O. N. Hellström. S. Vol. Jardine. SPIE 2003. Orlando. Proc. Almada.L. A.43. “Advances in uncooled technology at BAE systems”. Vol. Blackwell. Haggblad. The Netherlands. pp. Vol. A. V.65421Z. SPIE 2007. [35] V.1-1M.12. R.187-210.

[51] F. S. Chi-Anh. Margot. Yi.[49] P. Gopal. “Wafer-level membrane transfer bonding of polycrystalline silicon bolometers for use in infrared focal plane arrays”.1. T. M. X. R. V. F. Tanrikulu.14. D. F. Vol. VanNylen. Porter. Kruzelecky. [52] F. [67] M. pp. Franks. Part 3.44. A. H. W. “Fabrication and performance of microbolometer arrays based on nanostructured vanadium oxide thin films”. U.38-41.315. Proc. Halldin. ”Low temperature electrical properties of a boron-doped amorphous silicon bolometer”. J.1.4. Struct. Vercammen.10.55-61.312. F. R. “ Effects of Ti–W codoping on the optical and electrical switching of vanadium dioxide thin films grown by a reactive pulsed laser deposition”. Shannon. pp. I. Eriksson. No. No. J. E. Edwards.102-113. A. “Far infrared low-cost uncooled bolometer for automotive use”. Kvisterøy. F. No.031101/1-031101/28. Stemme. Mansi. A14. “AUTHENTIC: a very lowcost infrared detector and camera system”. pp. SPIE 2001. Chen. Shin.11. G. Källhammer.4969-4972.1. Vol. T. pp. Vol.Y. Surface and Coating Technology. G. C. C. Vol. L. Bi-feng. [59] N. J. J. J. 2007. [69] A. Källhammer. 2004. Vol. AMAA 2007. [61] R. pp. H-J. Strasbourg. P. “Infrared responsivity of uncooled VO~2-based thin films bolometer”. Savage. USA. Stemme.254101-3. Liu. Vol. “Fabrication and characterization of integrated uncooled infrared sensor arrays using a-Si thin-film transistors as active elements”.109. Orlando. “Fulfilling the pedestrian protection directive using a long-wavelength infrared camera designed to meet the performance and cost targets”.22. H. [62] Y-H. K. 3. Smart Mater. J. S. NEC Research & Development. H-J. 6836 68360D-14 . Junique. pp.86. G. [58] C. 2005. K-T.1. pp. G. Vol. J. Yue. Hu. Eriksson. No. M. Manoravi. Moon. Niklaus. R. Choi.1-6.1167-1177.6198. R. E.397-404. [66] N. 2001. pp. Thin Solid Films. U. Sensors and Actuators A: Physical. [72] M. [63] Y. “Amorphous silicon microbolometer technology”.425. B. Vol. al. pp. Vol. Sasaki.1958-1960.. [68] A. Gooch. Mangalaraj. No. Wissmar. Zhang et al. The Electrochemical Society 2003. Sylaios. “Study of a pulsed laser deposited vanadium oxide based microbolometer array”. Proc. [70] L. Proc. Karunagaran. Vol.Y. [65] N. K. pp. Lu. Torres. X. VanNylen. 2003. Forsberg. Journal of Microelectromechanical Systems.c. reactive sputtering and annealing process”. Oda.-Q.1-2.16. Kälvesten. Chen. [53] J. Akin. Acta physica Sinica.188-192. [64] Y-H. VanHulsel. J. pp. Kim.-E. Chang-hong.6. [56] F. Jaramillo. Eminoglu. D.609. Vol. X. R. Pettersson. 2006. Vol. D. Vol. H. Proc. 2001..-E. Yue.87-92. J. “Fabrication of vanadium oxide thin film with high-temperature coefficient of resistance using V2O5/V/V2O5 multi-layers for uncooled microbolometers”. “A high performance single-chip uncooled a-Si TFT infrared sensor”. L. [60] S. J. Vieider. A. F.1L. M. S. Halldin. J.-H. MRS 2000. S. Niklaus. J. Stemme. Kälvesten. A.50. Proc.227-238. VanHulsel. H. Schimert. Xin-jian. Joseph and V. 2007.12. 2005. Y. Abs. Proc. Infrared Physics & Technology. Dong.F. Stemme. Vol. 1997. Vol. Proc. P. Källhammer. R. Proc. Applied Physics Letters. “Characterization of uncooled bolometer with vanadium tungsten oxide infrared active layer”.J. M. Vieider. H. “A new concept for CMOS-compatible fabrication of uncooled infrared focal plane arrays using wafer-scale device transfer bonding”. Gutmann. SPIE 2006.123. B. Vol. Jakobsen.4369. Stemme.6542. Andersson. P. Ritchey. USA..260-264. Niklaus. Shin. Kvisterøy. T. S. [50] S. Pettersson. 2003.-E. Han.. “Micromachined uncooled IR bolometer linear array using VO2 thin films”. H. “Adhesive wafer bonding”. S.11. Chaker. Pettersson. Jakobsen. G. Vol. Yi. A. 2003. Ericsson. Struct. Eriksson. J. S. Liu. Orlando.-Y. “Thermal characterization of surface-micromachined silicon nitride membranes for thermal infrared detectors”. T. Eriksson. M. A. ”Enhanced characteristics of an uncooled microbolometer using vanadium-tungsten oxide as a thermometric material”. Lambkin. G. SPIE 2003. Brookfield. Soltani. Berlin. Proc. “Low-cost uncooled infrared detectors in CMOS process”. S. pp. Tregilgas. Wu and Z. E.509-513. M. Chi-Anh. P. Niklaus. H.T. pp.. Journal of Micromechanics and Microengineering. Franks. Stemme. pp. Andersson. 99. W. Y. Ericsson. Niklaus. Kawahara. Bold. J. Shin and S. S. Kurashina. “Performance of 320 x 240 bolometer-type uncooled infrared detector”. G. 201. D. I. A. Moon: “Excess noise in vanadium tungsten oxide bolometric material”. Vercammen. Niklaus. pp. Kang. S. H-K. Vol. 2003.4820. A. [57] C. McCardel. 2005.450-452. T. 2007. Han. J.53-58. France. 2001. Park. Mathewson. Sensors & Actuators A. Applied Physics Letters. R. Vol. Ajisawa. SPIE 2007. Narayandass. pp.881. Journal of Applied Physics: Applied Physics Reviews..696-700.50.2. Dong. H. pp. De la Hidalga. “Preparation of vanadium oxide thin films with high temperature coefficient of resistance by facing targets d.1-1L. Transducers 2003.5. pp. Stemme. Ma. International Journal of Infrared and Millimeter Waves . pp. Vol. Xiang. [71] L. Kim. Germany. Haddad. [55] T. Wissmar. No.J. Lv. Vol. Journal of Microelectromechanical Systems. K-T. H. of SPIE Vol. “Low-cost far infrared bolometer camera for automotive use”. Kumar. pp. Kim et.85. Sabuncuoglu Tezcan. Heredia. L.74-84. Tanaka. [54] C. No. Vol. Smart Mater. No. D. H-J.Y. B. Liu. Moon. Vieider.

No. W. MEMS 2006. A. pp. Hijikawa. [90] C. Ahmed.The route to the mass-market thermal imager”. Kim. pp. Grishin.355-360.[73] A. Parrish. K. Part A: Vacuum. “Sangrador IR uncooled bolometers based on amorphous GeSiO on silicon micromachined structures”. “Free-standing epitaxial La1-x(Sr. Almasri. [80] R. Hoglund. T. Ringh. Jansson. K.8. Clement.11.Crastes. “Noise behavior of amorphous GeSiO for microbolometer applications”. Aizawa. D. Proc. Dupont. Orlando. T. Tranducers 1997. No.1-1V. Trouilleau. C. [83] H. Fieque. “Fabrication and performance comparison of planar and sandwich structures of micro-bolometers with Ge thermo-sensing layer”. [88] X. 2007. Proc. Moreno. [79] I.297302. . [86] J. N. Kumazawa. Vermeiren. Orlando.1-2. USA. [92] W. “High signal to noise ratio quantum well bolometer materials”. “Infrared focal plane array incorporating silicon IC process compatible bolometer”. Nakazato. T. Sridhar. P. “Performance of microbolometer focal plane arrays under varying pressure”. C. [89] V.43. S. Vol. Journal of Microelectromechanical Systems. Dong. M. SPIE 2003. Proc. Butler. Y. Vol.Ca)xMnO3 membrane on Si for uncooled infrared microbolometer”. Fomin. Phong. Vol. A.27. N.2.K. Vol. L. pp. Rana.6. Liddiard. Vol. Lim. SPIE 2003.465-472. Chistokhin.4369. USA.106-109. [85] L. Ericsson. S. [84] M. Yaradanakul. No. Vol. Woolaway. M. pp. Endoh. SPIE 2007.407-414. pp. Tanaka. Proc. pp. 2006. “Room temperature YBaCuO microbolometers” Journal of Vacuum Science and Technology. IEEE Transactions on Electron Devices. K. USA. G. [87] M. [81] S. B.1844-1850.54-63. E. U.264-275. R. P.5. Tilli. [74] E.5406. Yue. Ringh. Mei.3. A. Lindroos. pp. Vol. A.Touvignon. No. Z. Orlando. pp. Vol. No.21. pp. Cherepov. J.316. C. B. Leonov. S. Teranishi. Vol. Gillham. San Diego. pp. Okuyama. Proc.3698. Electron Device Letters. B.2. Liu. Minassian. He.264-73. “Improvements in uncooled systems using bias equalization”. pp. “Polycrystalline layers of silicon-germanium alloy for uncooled IR bolometers”. H. A. Yildiz. Proc. A. Celik-Butler. pp. [75] D. Tait.76070-7610. Parkinson. B. Kosarev. Wissmar. A. Manning. “A high performance amorphous Si1-xCx:H thermistor bolometer based on micro-machined structure”. SPIE 2006.19.4. “Semiconducting YBaCuO microbolometers for uncooled broad-band IR sensing”.4945. E. Vol. O. Vol.635-638. “Radio Frequency sputtered SiGe and SiGeO thin films for uncooled infrared detectors”. International Journal of Infrared and Millimeter Waves. Myong. P. [93] P. Yon. Vol. pp. Tal.2552. Chiba. Yoon. 2005.995-1003. Vol. Robert. Proc.J. L. [77] M. No. Proc. J. 2005. J. Andersson. No. 2000. Y-C. Journal of non-crystalline solids. Van Hoof. Verbist. of SPIE Vol. 2008. M. [91] C. J. De Moor. Du Bois. Applied Physics Letters.2. USA. Liddiard. of SPIE 2004. A.Legras.748755.644-653.5126. J. Infrared Physics and Technology. Susan. N. Grietens.6401.6542. No. Karunasiri. Herrero. K. Liger. pp. Perova. Tissot. No. Lee. “FOA/DSTO uncooled IRFPA development”. Motooka: Handbook of Silicon Based MEMS Materials and Technologies. Robinson. 2000. Arnaud. pp. Proc. pp. No.322-329. [82] T. M. U. K.7. Proc. B. SPIE 1999. S. Qiu. Matsumoto. 6836 68360D-15 . K. Ambrosio. Neuzil. Watabe. Vieider. No.514. Y. [78] V.233-235. Jansson. Vol. T. Vol. “Silicon foundry micro-bolometers . L. Gotoh. “Micromachined poly-SiGe bolometer arrays for infrared imaging and spectroscopy”. Torres. Vol.3698. T. “Uncooled amorphous silicon 160 x 120 IRFPA with 25 µm pixel-pitch for large volume applications”.1253-1256.87. J. Goessens. of the SPIE 2001. I.515. Ichihara. S. Tanaka. L. Sweden. N. 1996.1V.46. Vol. Tinnes. R. pp. Butler. Thin Solid Films. “Electrical properties of photo-CVD boron-doped hydrogenated nanocrystalline siliconcarbide (p-nc-SiC:H) films for uncooled IR bolometer applications”. pp. Tsukamoto. U. Vol.11. P. Honda. Proc. pp.33502-1-318. “Monolithic uncooled 8 × 8 bolometer arrays based on poly-SiGe thermistor”. A. Thin Solid Films. Vol. Proc. Kaushal. 2002.18. A. Vol. A.468-472. Stockholm. “A 32*32 parylene-pyrolyzed carbon bolometer imager”. pp. SPIE 1999. [76] A. Lehto. SPIE 1995. Zeng. 2003. [94] C. Iborra. Surfaces and Films.1. "Theoretical analysis of pulse bias heating of bolometers and effectiveness of pulse bias compensation". C. Michailovsky. Itoh. to be published by William Andrew Publishing. N. T. S. Y.

Sign up to vote on this title
UsefulNot useful