University of Houston • University of Houston-Clear Lake • ISSO Annual Report Y2003 • 16-21
Miniature Optical Sensors for Detection of Water and Air Contaminants
ISSO RESEARCHERS HAVE INVESTIGATED THE DEVELOPMENT of optoelectronic chemical sensors based on group III nitride materials. Until recently, compounds of GaN, AlN, InN, and their alloys were known to be optically active from 650 nm (InN) to 200 nm (AlN). Recent results on InN growth and characterization indicate that the bandgap of this material is not 1.9 eV as was previously thought, but 0.7 eV, extending optical applications for these materials to the 1770 nm range. Nitride alloys become thus ideally suited for use in UV-VIS-IR chemical sensors.
Emission and detection devices can be separately tailored to specific wavelengths and grown on the same chip (Fig. 1). Integrated devices made of AlInGaN materials could offer many advantages over current optical chemical sensors, among them high chemical and thermal stability, smaller size, and higher sensitivity.
Figure 1. Simplified Structure of a Nano-Integrated Multi-Wavelength Sensor
The objective of this two-year project was to develop and fabricate a working prototype of a nitride-based optoelectronic chemical sensor. The sensor was tested with various concentrations of a known contaminant in water. Towards this goal, the first part of the project has focused on growth of the various materials necessary to fabricate the sensor. In 2000, the ISSO Post-Doctoral Fellow was hired and started working on materials growth issues related to sensor fabrication. Specifically, the growth of high-quality GaN, AlGaN, InGaN, and InN layers on sapphire and Si have been studied. The basic materials research and fabrication and testing of the prototypes were nearly completed in this project with future work directed toward the growth of multi-layer structures and the processing of these samples into devices for testing.
In order to characterize finished sensors, we have recently assembled and tested a portable prototype device that includes "macro" components made from discrete LEDs used as emission sources and photodetectors.
The nitride layers in our investigation were grown by radiofrequency gas source molecular beam epitaxy (RFMBE). This method used an EPI Uni-Bulb plasma source to generate active nitrogen species while standard effusion cells supplied the group III metals. As part of our preliminary work, we grew several layers on commercial grade Si(III) substrates which are significantly less costly than sapphire wafers of the same size. Later experiments were performed on sapphire wafers when growth conditions had been narrowed down.
For experiments done on Si substrates during the year 2000, a 200Å thick AlN buffer layer was deposited between 750°C and 800°C prior to growth of GaN, InGaN, or AlGaN films.
Experiments on sapphire began with either direct deposition of GaN followed by subsequent layers or by the same AlN buffer layer used for silicon substrates. Since we had previously demonstrated n- and p-type GaN, this work focused on two main objectives: (1) growth of InxGa1-xN layers with varying values of x for emission and detection windows and (2) growth of AlxGa1-xN layers for emission wavelengths less than 363 nm. Layers were characterized by photoluminescence (PL), cathodo-luminescence (CL), secondary ion mass spectroscopy (SIMS), and x-ray diffraction (XRD).
A range of substrate temperatures and In/Ga flux ratios were explored to study the effect of growth conditions on InxGa1-xN layers deposited on Si. Substrate temperature was varied between 600°C and 650°C with a profound impact on the indium incorporation in the film. At 650°C, no indium was found in the layers for any indium flux as determined from PL and SIMS. Only by lowering the growth temperature to 600°C was a substantial amount of indium incorporated into the film. This reaction is caused by the higher re-evaporation rate of In as compared to Ga at these temperatures. By adjusting the ratio of In to Ga during growth, mole fractions of up to 42 percent In were achieved. However, there were problems with uniform indium incorporation. For growths on Si, the indium tended to separate out into two or more distinct compositions of InxGa1-xN.
For experiments performed on sapphire, the growth temperature of 600°C was fixed and the In/Ga ratio was adjusted. Layers grown in this manner showed much less InxGa1-xN phase separation, as illustrated by the PL data shown in Fig. 2. By changing the relative fluxes, compositions of up to 50 percent indium mole fraction have been achieved without phase separation. The corresponding optical emission for this particular layer is around 520 nm, which is roughly the lower energy (longer wavelength) limit that we should need for our chemical sensors.
Figure 2. Photoluminescence Spectrum of a In0.5Ga0.5 N Layer Grown on Sapphire
The initial investigation on the growth of AlxGa1-xN for higher energy (lower wavelength) applications was performed on Si(III) substrates. Prior to AlGaN growth at 750°C, AlN was deposited at 800°C followed by GaN at 800°C. We explored a range of Ga/Al flux ratios to determine the compositional dependence on the group III fluxes.
In the case of AlxGa1-xN, it is the higher sticking coefficient of the Al that strongly determines the film composition. Due to the lattice and thermal expansion mismatches between the layers and the Si substrate, cracking of the films proved a problem. Therefore, layer thickness had to be kept below about 4000Å in order to prevent cracking during post-growth cooling. Composition of the layers on silicon ranged from 7% Al to 42%, as determined by CL emission peaks, shown in Fig. 3.
Figure 3. Cathodoluminescence Spectra of AlxGa1-xN Layers with Compositions Ranging from 15% to 42% Al
Work has recently begun on the growth of AlGaN on sapphire substrates. Because the substrate is transparent, measurement of the transmission of the film as a function of wavelength can be used to determine the bandgap, and, hence, the Al mole fraction of the layer. Only a few growths have been performed to date, but Al compositions of up to 71 percent have been achieved as shown in Fig. 4.
Figure 4. Transmission Spectra from Three Layers Grown on Single-Side Polished Sapphire: GaN, AlGaN, and AlN. The adsorption edge of the AlGaN layer is at 239 nm (5.19 eV) which corresponds to a Al mole fraction of about 71%.
We have made good progress on the development of nitride-based integrated optoelectronic chemical sensors. We have demonstrated growth of InxGa1-xN and AlxGa1-xN layers by RFMBE on sapphire substrates, with indium mole fractions up to 50% for InxGa1-xN, and AlxGa1-xN, films with up to 71% Al. Currently, we can fabricate layers that are optically active from 200 nm (AlN) up to 520 nm (InGaN). The next step in this project will integrate these layers into device structures and fabricate the samples into integrated optoelectronic chemical sensors. That effort will require working closely with personnel at the Johnson Space Center (JSC) in order to test and optimize sensor performance with the goal of being able to accurately detect the concentration of known contaminants in water.
During 2001-2002, the materials fabrication aspect of this project focused on two main areas. The first area was the growth of InGaN, which is the material used as the light-emitting region in our device structure. The purpose was to vary the InGaN growth parameters in order to produce spectral emission in different wavelength regions. Room temperature photoluminescence of the InGaN layers was used to determine the wavelengths of the emission.
The second area of interest was the improvement of the optical and electrical properties of our base GaN material. In order to produce samples with improved properties, film growth parameters, including initialization and deposition conditions, were varied and their effects measured. The resulting films were analyzed by room temperature photoluminescence, low temperature PL, and Hall effect measurements.
Researchers met specific challenges related to each of the two aspects of the project. The growth of InGaN is typically more difficult than growth of GaN because of the greater vapor pressure of the indium species relative to gallium and aluminum at the film growth temperature. This difficulty necessitates lowering the deposition temperature so that indium can incorporate into the InGaN film. The drawback is that the gallium species have less surface mobility at this lower temperature and thus require a lower growth rate. In addition, when growing InGaN layers that are several hundred angstroms thick, the indium tends to segregate to the surface of the film, thus leading to uneven indium composition within the layer.
To overcome these problems, we employed a quasi-superlattice structure where the InGaN was grown by alternately opening and closing the indium shutter with the gallium and nitrogen shutters always open. This would allow any excess indium left over from the direct growth of InGaN to be consumed during the stage when the indium shutter was closed. Control of the indium composition could be obtained by varying both the ratio of indium to gallium and the ratio of time the indium shutter was opened and closed.
One of the issues in GaN growth is the lack of latticematched substrates on which to grow epilayers. This lack of substrates results in GaN layers with large numbers of dislocations which can reduce the optical quality of the material as well as the mobility of the free carriers. In order to address these problems, we optimized the initiation of film growth on the substrate, the deposition of the AlN buffer layers, and the growth parameters of the GaN overlayers. Mainly, the substrate preparation, buffer layer growth temperature, and buffer layer thickness were varied. The effects on the optical quality were gauged by the full width at half maximum (FWHM) of the room temperature photoluminescence and the electrical quality judged by measuring the free carrier mobility using the Hall effect.
For the growth of InGaN, we were able to demonstrate films that were optically active from 371 nm to 456 nm, as shown in the room temperature photoluminescence in Fig. 5. The emission from the films was typically single-peaked, which indicates that segregation of the indium during the growth was mostly avoided. The layers were grown with indium/gallium flux ratios from 0.72 to 1.11 and In + Ga / Ga shutter ratios from 1 to 0.72.
Figure 5. Photoluminescence of Various InGaN Layers Emitting at Different Wavelengths
In the case of the base GaN materials, we were able to achieve layers with improved electrical and optical properties. As the left axis of Fig. 6 illustrates, the electron mobility of similarly doped GaN films was increased from 88.5 cm2/Vs to over 300 cm2/Vs. At the same time, the optical quality was improved, as shown on the right axis of Fig. 6, by the reduction of room temperature PL FWHM values from 59 meV to 53 meV. In addition to room temperature measurements, our better layers exhibited excellent photoluminescence at 10 K. Figure 7 is an LT PL example of a layer with very bright near-band-edge emission at 354.6 nm with a narrow FWHM of 16 meV. The low temperature result is further indication of the optical quality of the recent films.
Figure 6. Progression of Si-Doped GaN Films Showing Increased Carrier Mobility and Reduced Photoluminescence FWHM Values
Figure 7. Photoluminescence at 10 K of Typical Improved Si-Doped GaN Film Showing Single Peak and Narrow Line Width
The main focus of the nitride material growth during the year 2003 was investigation of conditions for the growth of semiconductor quality InN layers. The ultimate goal of this work is to extend further into the IR range the spectral range of the optical components based on III nitrides.
InN has been much less studied than other III-Nitride materials, such as GaN and AlN. The reason is the difficulty of growing highly crystalline InN. InN has a large lattice mismatch with GaN (11%) and AlN (14%), which are typically used as buffer or template layers, and with the typical substrate materials Al2O3 (26-29%) and Si(III) (8%). In addition, InN is less thermally stable than the other nitrides and usually has a high background n-type doping level attributed to likely nitrogen vacancies.
Recent results in the growth of InN by RF-MBE1-4 have brought a better perspective to this material and highlighted some of its potential advantages. InN has a very high room temperature electron mobility (>2000 cm2/Vs), has higher polarization and piezoelectric constants than GaN, and, most important, has been found to have a much lower bandgap than was predicted and measured earlier from lower quality films. The new bandgap measuring ~0.7-0.9 eV for InN has not been universally accepted. There is however strong evidence from the multiple literature reports and our laboratory that the bandgap is not the previously-thought 1.9 eV. Thus, alloys between GaN, AlN, and InN have the theoretical potential to cover a very large spectral range, from 200 nm up to 1.38-1.77 µm.
The films detailed in these recent reports have all been fabricated by RF-MBE employing InN grown on templates of GaN and/or LT-InN in a growth temperature range between 370ºC-600ºC. In these reports, the intrinsic background carrier concentration has been reduced from typical values of 1020 cm-3 down the low 1018 cm-3 and even 1017 cm-3 ranges. Both room and low temperature band edge-related luminescence have been observed. Factors such as buffer layer type, growth temperature, and template polarity play major roles in the optical and electrical properties of the layers.
Our preliminary work on the growth of InN has demonstrated that we are able to grow smooth, two-dimensional InN films on GaN buffer layers deposited on sapphire substrates. We have seen a very strong growth temperature and buffer layer polarity dependence in our growths, with the best results so far employing a growth temperature around 500ºC. We have observed 14 K photoluminescence in the energy range 0.71 to 0.73 eV (un-calibrated response) from nonoptimized, 2000Å thick InN films (Fig. 8).
Figure 8. InN Grown on Sapphire: SEM View of the Surface (l.) and Low Temperature Photoluminescence (r.)
Future work has been directed toward further improvement of the InN growth by employment of a large variation of growth substrates, substrate preparation, and growth parameters in order to investigate their applicability for Schottky barrier diode fabrication.
The new POS-1 device prototype was developed from vacuum-grown nitride thin-film technology at the Texas Center for Superconductivity and Advanced Materials (TCSAM). An Optical Multifunctional sensor (Fig. 9) has been integrated into a working biomedical instrument in collaboration with Integrated Micro Sensors, Inc. The instrument was operated and tested in the field as part of a Mars Society simulation called the Mars Analog Research Station (MARS) project. The second field season of the MARS project included a Mars base-like habitat located in the desert of the American southwest. In this Mars-like environment, the MARS Society has executed a program of extensive long-duration geology and biology field exploration operations conducted in the same style and under many of the same constraints that they would encounter on the Red Planet. See <http://www.mars society.org/mdrs/index.asp>.
Figure 9. Portable Optoelectronic Sensor (POS-1)
In a double-blind test, the instrument was used to analyze water at the simulation site and returned to TCSAM/IMS with the analyte water for laboratory verification. Data obtained under carefully controlled conditions at TCSAM matched samples gathered in the Utah desert. In the simulation, the sensor proved both durable and easy to use, operating optimally in extreme conditions, in the hands of technicians/"astronauts" generally unfamiliar with the system. The ultimate usefulness (and marketability) of the sensor system depends on a library of environmental chemical signatures. The water used in the simulation and subsequent in-house testing is undergoing further bio-analysis at UH, representing the first action toward building a referential database for the logic protocols of the sensor.
1A. G. Bhuiyan, A. Hashimoto, and A. Yamamoto.
"Indium Nitride (InN): A Review on Growth, Characterization, and
Properties," J. Appl. Phys. 94.5 (2003): 2779-2808.
2J. Wu, W.Walukiewicz, W. Shan, K. M. Yu, J. W. Ager III, S.
X. Li, E. E. Haller, H. Lu, and W. J. Schaff. "Temperature Dependence of the
Fundamental Band Gap of InN," J. Appl. Phys. 94.7 (2003): 4457-60.
3J. Wu, W.Walukiewicz, K. M. Yu, J. W. Ager III, S. X. Li,
E. E. Haller, H. Lu, W. J. Schaff, Yoshiki Saito, and Yasushi Nanishi. "Unusual
Properties of the Fundamental Band Gap of InN," Appl. Phys. Lett. 80.21
(2002): 3967-69.
4K. Xu and A. Yoshikawa. "Effects of Film Polarities on
InN Growth by Molecular-Beam Epitaxy," Appl. Phys. Lett. 83.2 (2003):
251-53.
Starikov, D., C. Boney, R. Pillai, and A. Bensaoula. "Dual-Band UV/IR Optical
Sensors for Fire and Flame Detection and Target Recognition," Proc., ISA/IEEE
Sensors for Industry Conference (SIcon/04), New Orleans, LA Jan. 27-29, 2004. 36-40.
Starikov, D., C. Boney, N. Medelci, R. Pillai, and A. Bensaoula. "Dual-Color UV/IR
Photodiodes Based on AlGaN Grown on Si and SOS for Advanced Fire/Flame
Detectors," Proc., 50th International AVS Symposium Baltimore, MD, Nov. 21-26,
2003. 133.
Starikov, D. "Employment of III-Nitride Materials for Microsensor
Applications," Seminar, Texas Center for Superconductivity and Advanced Materials,
University of Houston, Houston, TX, Feb. 21, 2003.
Starikov, D., C. Boney, R. Pillai, and A. Bensaoula. "Dual-Band UV/IR Optical Sensors
for Fire and Flame Detection and Target Recognition," ISA/IEEE Sensors for Industry
Conference (SIcon/04), New Orleans, LA, Jan. 27-29, 2004.
Starikov, D., C. Boney, N. Medelci, R. Pillai, and A. Bensaoula. "Dual-Color UV/IR
Photoiodes Based on AlGaN Grown Si and SOS for Advanced Fire/Flame
Detectors," Fiftieth International AVS Symposium, Baltimore, MD, Nov. 21-26, 2003.
Starikov, D. "Advanced LEDs with Temperature Control." DoD SBIR, 2003,
$70,000 (six-month project).
—. "Advanced Multi-Band UV/VIS/IR Focal Plane Arrays Based on III
Nitrides." DoD MDA STTR, 2004, $100,000 (nine-month project).
—. "Low-Cost Miniature Absorption/Scattering/Fluorescence Sensor." DoD
SBIR, 2003, $70,000 (six-month project).
PDF (495KB)
Table of Contents
Institute for Space Systems Operations - Y2003 Annual Report
Copyright © 2004
|
