Infrared imaging at high temperatures

Although IR focal-plane arrays (FPAs) have been widely used in many imaging systems, the demand for cameras and linear arrays sensitive to the mid-wave IR (MWIR) spectral band (3–5μm) has drawn considerable interest for a variety of applications, such as pollution control, remote detection of chemicals, and developing imaging systems, including cameras and linear arrays.^1,2 Most conventional detectors, such as those with mercury cadmium telluride (HgCdTe) sensors, can be classified as electrical detectors because they produce electrical signals. However, HgCdTe detectors are costly and bulky, and the electrical signals mean that they must operate at very low temperatures (∼77K).

To overcome these drawbacks, many studies have focused on using other semiconductors, such as indium gallium arsenide (InGaAs) and indium arsenide (InAs), to fabricate MWIR detectors. Despite considerable efforts, cooling remains a major limitation of IR detectors, with the exception of thermal detectors, which generally have very low detectivity. Recent attention has focused on making uncooled MWIR detectors using alternative operating mechanisms and other materials. For example, microcantilever arrays have been studied for developing optomechanical IR-imaging systems.^{3, 4} However, these types of systems have a number of drawbacks, including relatively long response times.

Silicon carbide (SiC) is a wide bandgap semiconductor with a higher radiation damage threshold and better thermal stability than conventional semiconductors. SiC's wide bandgap means that it can be used to fabricate detectors for different wavelengths by adding appropriate dopants.

In our SiC-based detector, incident MWIR photons excite electrons from the valence band to the acceptor energy level, resulting in different electron densities in the two energy levels. Since the complex refractive index depends on the electron density and the reflectivity in turn depends on the refractive index, this movement of electrons changes the detector's reflectance. These changes in reflectance can be considered an optical signal, in contrast to conventional electrical detectors where electrical signals, such as changes in voltage, are produced. Also, the optical photodetectors can be integrated with a CCD camera to generate images directly. Read-out integrated circuits (ROICs), which are commonly used in traditional electrical detectors, are not necessary for the optical detectors.

Our detector essentially acts as an image exchange device (IED) where the optical wavefront, which carries the image information from the object of interest, hits the detector and transfers the photonic image information to electrons to create an electronic image pattern in the form of electron density distribution. This electronic image pattern is picked up by the wavefront of a probe light emitted by an LED. Finally, the optical wavefront travels to a CCD camera to reproduce the image of the object of interest. Thus, the performance of the IED depends on the interactions of these three wavefronts: the optical wavefront from the object of interest, the electronic wave pattern in the photodetector, and the optical wavefront of the probe light to recreate the image.

We chose to fabricate a detector for the MWIR wavelength of 4.21μm. Many chemicals emit radiation with characteristic wavelengths in this spectral range, enabling our device to be used as a remote gas sensor. To make such a detector we doped an n-type 4H-SiC substrate (1×1cm) with Ga using a laser doping technique with triethylgallium—(C₂H₅)₃Ga—as the metallo-organic precursor.^5,6 Ga atoms create an acceptor energy level corresponding to photons of wavelength 4.21μm, enabling photoexcitation of electrons and consequently photon detection at this wavelength (see Figure 1).

Figure 1. Operating principle of the uncooled silicon carbide (SiC) optical photodetector. The reflectance of the detector changes as mid-wave IR (MWIR) photons excite electrons from one level to another. E_s(λ_s): Energy of a photon of wavelength λ_s emitted by the radiation source. E_v: Energy of an electron at the top of the valence band. E_c: Energy of an electron at the bottom of the conduction band. E_a: Acceptor (p-type dopant) energy level. E_g: Energy gap between the acceptor level and E_v, i.e., E_g=E_a- E_v.

We positioned the Ga-doped SiC sample between a red (633nm) LED source and a CMOS camera, each directed toward the sample at a 45° downward tilted angle (see Figure 2). The Ga-doped regions in the sample were centered in each of the four quadrants and characterized by different dopant concentrations. We directed radiation from the MWIR source up through a narrow bandpass filter, CaF₂ (calcium fluoride) source, and detector lenses.

Figure 2. (a) Schematic diagram showing the components of the experimental setup for optical signal measurement and optical readout imaging. Ga: Gallium. 4H-SiC: Hexagonal unit cell. r_d: Radius of the detector. CaF₂: Calcium fluoride. 2r_rs: Diameter of the IR radiation source. (b) Photograph of the optical readout measurement setup in the laboratory.

First, the CMOS camera captured images of an undoped sample illuminated by the LED source. We then replaced this sample with the Ga-doped sample, and ramped up the MWIR source, before maintaining it at a fixed temperature in intervals of 100°C over the range of 100–600°C. We collected images for each quadrant at each temperature and used image-processing techniques to subtract the image of the doped sample from that of the original sample for each exposed quadrant at all six MWIR source temperature levels (see Figure 3). For each processed image, the average pixel value in the portion of the subtracted image representing the doped region in the MWIR-exposed quadrant is computed (black boxes). Our results showed a significant change in the average pixel contrast as temperature increased, demonstrating a significant change in reflectance and thus a change in the refractive index between the air and doped sample interface with increasing exposure to MWIR radiation.

Figure 3. Contrast between undoped and doped regions exposed to the MWIR source over the temperature range visualized through image subtraction.

Our new uncooled SiC-based detector with optical readout overcomes obstacles associated with conventional detectors, including cooling issues, high cost, low damage threshold, and thermal stability. In tests we have measured distinct changes in the refractive index of doped SiC detectors in the presence of carbon dioxide (CO₂) and nitrogen monoxide (NO) gases, enabling gas-sensing applications.⁵ A potential next step is to dope SiC using different elements to fabricate detectors for other MWIR or even terahertz wavelengths based on the optical properties of the detectors.

This research was sponsored by In-house Laboratory Independent Research FY11/12.