Materials Science and Engineering Studies to Support Advanced Compound Semiconductor Devices
 J. Woodall, P.I., 6 students, 2 senior research engineers

We have a number of ongoing materials science and engineering projects that are designed to facilitate the development of technology enablers for new or vastly improved photonic and high speed electronic devices made of III-V compound semiconductor materials. These include, but are not limited to:

Aqueous Solution Stable, GaP Based J-FET Biosensor Arrays with Single Charge State Discrimination - J. Woodall, P.I., Ahn Chen - graduate student

Arrays with chemical and biological selectivity at the individual pixel level are basic to continued progress in the field of genometrics. A straight forward approach to biomedical sensor arrays is to develop specially modified semiconductor gate arrays. The obvious candidate is Si MOS CCDs, or related devices. Unfortunately, Si MOS devices are unstable in aqueous solutions. Therefore, this approach was abandoned early on. Fortunately, alternative approaches have been develop in which, for example, molecules with unique florescent properties bind to specific bio-materials of interest. Then a multi-component mixture of "tagged" materials are spread onto chip with an array of photodetectors integrated with appropriate electronic circuits. In spite of the success of this approach, the ability to fabricate semiconductor gate arrays that are stable in aqueous solutions and can detect single charge state events is still desirable.

We propose to realize such a chip using GaP as the semiconductor material. GaP with its large band gap energy of 2.26 eV can theoretically yield FETs with a low enough noise floor to be able to detect single charge state changes. Since Prof. Woodall has already shown that GaP can be made to be stable in acid or basic aqueous solutions, even under illumination, we propose to develop low noise GaP J-FETs gate arrays that are stable in aqueous solution sensitive to changes in single charge state events. Since we have a world class GaP epitaxial facility, we have already fabricated test structures to characterize noise in such devices.

Power Schottky Diodes using InAs as the Metal to a GaP/InGaP Schottky Diode Structure
J. Woodall, P.I., Ahn Chen -student

There is much activity on developing new materials for power electronics. Even though Si has served the community well, it has certain limitations with respect to simultaneously achieving both high blocking voltages and low on state power densities. The coordinated search for improve power devices is known as the Power Electronics Building Block (PEBBs) program. In this program newer and more exotic materials including SiC and GaN, are being studying because of their high theoretical performance figure of merit. Since SiC substrates are still very expensive and GaN is not yet a mature technology for this application, we have undertaken a program to develop a viable "interim" technology that should be significantly better than the current Si based technology.

We have developed a reliable and robust InAs/GaP Schottky diode device. The novel and technologically important feature is the use of InAs as the "metal" part of the Schottky diode. Since InAs is a zincblende semiconductor that covalently bonds to GaP and has an 11 percent lattice constant mismatch to GaP, its interface to GaP is thermodynamically stable against thermal or electrical field driven diffusion (the maximum E-field is at the metal/semiconductor interface for Schottky diodes). Next, since InGaP lattice matched to GaAs has 20x higher power diode figure of merit compared to Si, our ultimate optimal device will be InAs/GaP/InGaP.

Gravimetric Efficient/Radiation Hard InP Extraterrestrial Solar Cells
 J. Woodall, J. Pan co-P.I.'s, Yanning Sun -student.

Solar Cells used to power satellites should have a high power to weight ratio, i.e. high gravimetric efficiency (GE), and, if used in the Van Allen Belt (VAB) for optimal stationary orbit communication links, should be robust against high energy particle radiation damage, i.e. be Rad Hard. Unprotected Si cells placed in the VAB will degrade instantly. Adding protection covers can retard radiation damage but at the expense of significantly degrading gravimetric efficiency. InP is a material which in its thin film form is predicted to be sufficiently Rad Hard and yet with high GE to be a useful candidate for solar cells operating in conditions like the VAB for 5-10 years. We have designed and are building a n-i-p type thin film InP cell. In our innovation the normal thin n-type InP surface layer is replaced with a thin n-type InAs layer for even higher performance.

An InAs Based Transistor Approach to Terahertz Electronics including Improved GaAlAs/GaAs HBTs J. Woodall, P.I., H.Tsukamoto, visiting scientist, and T. Boone and Qingyu Yin - students

In the global quest for ever higher bandwidth for telecommunications, the teraherz (THz) frequency regime is under intense investigation. Many physical phenomena are being studied, including photonic arrays, for either WDM or TDM, Spintronics, quantum size effects, e.g. Bloch oscillators, etc. Not too much thought has been given to a transistor based approach to emitter and detectors, since Si technology, even at the 10 nm gate length regime, is unlikely to work. The reason is fairly simple: to make practical THz devices and chips we need to use materials with fast electrons and use practical lithographic processing techniques. For Si the saturation drift velocity is only 1x10⁷ cm/sec. For InAs however, this velocity is predicted to be in excess of 1x10⁸ cm/sec. Thus if we have a viable InAs epilayer technology (as we do), there are three possible transistor structures that can be made with both fts and fmaxs above 1 THz. Our current focus is on an MOSFET structure with a 0.1 micron gate, with a predicted fmax of 3 THz.

In addition since we can Be dope the base layer of a AlGaAs/GaAs HBT to 1x10²⁰/cc we have designed a structure that is predicted to have fts and fmaxs in the 0.5 THz range.

Fast, Efficient and Cheap LEDs for 2-40+ GHz Applications
J. Woodall, P.I., H. Tsukamoto, visiting scientist, T. Boone, R. Koudelka -students

Fast optical emitters and detectors are important components of future high bandwidth telecommunications system. The Vertical Cavity Surface Emitting Laser Diode (VCSELD) is a leading contender for massively parallel arrays for THz systems. However, there is some uncertainty about current and projected price-performance issues for this approach.

We have initiated two different fast LED projects here at Yale. The first is a continuation of Prof. Woodall's previous work on high doping effects in GaAs. To date we have an enabling technology to fabricate LEDs with fts of nearly 2 GHz and with external quantum efficiencies of 10%, driven with currents up to 100 mA. This is limited by both radiative recombination rates and non radiative recombination. We are now determined the growth conditions for optimal performance of this device.

Second, we have invented a new kind of fast light emission device with operation principles that are different than "traditional" LEDs. It can be driven both electrically and optically and it can be modulated both electrically and optically.

Currently the Si Avalanche Photo Diode (APD) is highly developed and these APDs are currently capable of single photon detection. However, they are blind to eye safe lasers which operate in the 1.3-1.5 micron spectral range. There is great market demand for low noise detectors for eye safe radiation. InGaAs based APD technology is being studied. However, to our knowledge this technology cannot effect single photon detection. We are developing a novel Si base APD with a structural component that will allow the Si APD to "see" 1.3-1.5 micron radiation.