Search

A system is presented that is capable of measuring subnanosecond reverse recovery times of diodes in wide-bandgap materials over a wide range of forward biases (0 – 1 A) and reverse voltages (0 – 10 kV). The system utilizes the step recovery technique and comprises a cable pulser based on a silicon (Si) Photoconductive Semiconductor Switch (PCSS) triggered with an Ultrashort Pulse Laser, a pulse charging circuit, a diode biasing circuit, and resistive and capacitive voltage monitors. The PCSS-based cable pulser transmits a 130 ps rise time pulse down a transmission line to a capacitively coupled diode, which acts as the terminating element of the transmission line. The temporal nature of the pulse reflected by the diode provides the reverse recovery characteristics of the diode, measured with a high bandwidth capacitive probe integrated into the cable pulser. This system was used to measure the reverse recovery times (including the creation and charging of the depletion region) for two Avogy gallium nitride diodes; the initial reverse recovery time was found to be 4 ns and varied minimally over reverse biases of 50–100 V and forward current of 1–100 mA.

Wide-bandgap GaN optically-controlled switches have the potential for driving down the cost and size and improving the efficiency and capabilities of high voltage pulsed-power applications. Key to these applications is the understanding of the high-field photo-conductive properties of this material to determine if it will operate in a high-gain and/or sub-bandgap triggering mode, such as has been observed in GaAs. Photoconductive semiconductor switches (PCSS) were fabricated and tested from GaN wafers from Kyma Technologies and Ammono. With relatively low voltages applied the switch, linear photoconductive currents are measured in response to exposure to laser pulses of sub-bandgap wavelength (532 nm). Above a threshold voltage applied to the PCSS, persistent conductivity is measured that lasts well beyond the duration of the laser pulse and discharges the charged transmission line in the system. The sustaining field for this switching mode is approximately 3kV/cm. Another known distinguishing characteristic of high-gain switching is the formation of filamentary current channels that can be imaged due to the emission of recombination radiation from the plasma within the filaments. These filaments have been also observed in the GaN switches. High-gain switching has been initiated in GaN devices with as little as 2.5 μJ laser energy. This persistent photoconductivity is a distinguishing characteristic of what is known as "lock-on" effect most notably known in GaAs-based photoconductive switches, and is the basis for highly-efficient photoconductive switching requiring relatively little laser energy.

The demand for greater flexibility and increased energy density in pulsed-power systems is moving highly interactive components closer together. The development of compact technologies for less complex and more robust system designs is critical. A key system component that can impact these goals is the trigger generator (TG). Inexpensive, compact, and fiber-optically controlled TGs that deliver trigger pulses with subnanosecond jitter have been created with photoconductive semiconductor switches (PCSSs). However, high-voltage (HV) GaAs PCSSs are typically pulsed charged for less than 100 s so that they can hold off 60-100 kV/cm without self-triggering into high-gain (lock-on) switching or initiating surface flashover. Since many new pulsed-power system designs are based on dc-charged HV switches, pulse charging the trigger system is an additional complication requiring space, HV switching components, and HV cables. A further improvement in PCSS-based TG is to move from pulsed to dc-charged PCSSs. This paper reports results from dc-charged GaAs PCSSs with 0.25-1.0 cm gaps, extending previously reported results on smaller devices at 3 kV to a new regime of 100 kV. To hold off high fields for longer periods and to extend GaAs PCSSs to dc applications, we have utilized neutron-irradiated GaAs (n-GaAs). Neutron irradiation in GaAs increases the defect density, shortens the carrier recombination time, and (for devices with large insulating regions) reduces the dark current, which improves the dc hold-off strength. PCSS contacts in this research were created using rapid thermal annealing (RTA) to produce high adhesion and low contact resistance. However, this can reduce the defect density near the contacts by annealing some of the n-induced defects. Hence, a range of RTA temperatures and neutron doses was studied to understand the tradeoff space for contact adhesion and dc hold-off. This paper presents results from I-V characterization and dc hold-off on irradiated and nonirradiated GaAs PCSSs. These PCSS devices were demonstrated to hold off fields of 39-61 kVDC/cm, respectively. Irradiation doses over a range of 3×1013 - 1×1015 (1 MeV Si equivalent) were explored in search of the optimal performance. Additionally, the impact of the fabrication processes on the benefits of irradiation is explored, and the observation of unusual low-frequency oscillations during GaAs I-V testing is discussed.

Prior to this research, we have developed high-gain GaAs photoconductive semiconductor switches (PCSSs) to trigger 50-300 kV high-voltage switches (HVSs). We have demonstrated that PCSSs can trigger a variety of pulsed-power switches operating at 50-300 kV by locating the trigger generator (TG) directly at the HVS. This was demonstrated for two types of dc-charged trigatrons and two types of field distortion midplane switches, including a ±100 kVDC switch produced by the High Current Electronics Institute used in the linear transformer driver. The lowest rms jitter obtained from triggering an HVS with a PCSS was 100 ps from a 300 kV pulse-charged trigatron. PCSSs are the key component in these independently timed fiber-optically controlled low jitter TGs for HVSs. TGs are critical subsystems for reliable and efficient pulsed-power facilities because they control the timing synchronization and amplitude variation of multiple pulse-forming lines that combine to produce the total system output. Future facility-scale pulsed-power systems are even more dependent on triggering, as they are composed of many more triggered HVSs, and they produce shaped pulses by independent timing of the HVSs. As pulsed-power systems become more complex, the complexity of the associated trigger systems also increases. One of the means to reduce this complexity is to allow the trigger system to be charged directly from the voltage appearing across the HVS. However, for slow or dc-charged pulsed-power systems, this can be particularly challenging as the dc hold-off of the PCSS dramatically declines. This paper presents results that are seeking to address HVS performance requirements over large operating ranges by triggering using a pulsed-charged PCSS-based TG. Switch operating conditions that are as low as 45% of the self-break were achieved. A dc-charged PCSS-based TG is also introduced and demonstrated over a 39-61 kV operating range. DC-charged PCSS allows the TG to be directly charged from slow or dc-charged pulsed-power systems. GaAs and neutron-irradiated GaAs (n-GaAs) PCSSs were used to investigate the dc-charged operation.

Semiconductor materials with wide bandgaps, including GaN and AlxGa1-xN, offer many performance advantages for power electronic devices compared to conventional Si-based devices. These include larger critical electric fields enabling higher reverse breakdown voltages, fast turn-off and turn-on transitions allowing higher-frequency operation, and high thermal conductivity making higher power density device operation possible. Various device architectures have been demonstrated in these materials, with laterally-oriented AlxGa1-xN/GaN high electron mobility transistors (HEMTs) maturing first for radio-frequency applications. More recently, lateral architectures are also starting to be seen in power switching applications. Vertical GaN PN diodes have been demonstrated with excellent forward and reverse electrical behavior including high current capacity, low forward specific on-resistance, high reverse breakdown voltage, and fast reverse recovery time. Selective-area doping control in GaN would enable traditional vertical power devices including merged PN-Schottky (MPS) diodes, junction field effect transistors (JFETs), and more advanced device architectures common in Si and SiC, thereby greatly extending the operating functionality and capability of GaN power devices. This talk will report on our work for two advanced GaN device technologies, including fundamental studies of selective-area doping control and the demonstration of a photoconductive semiconductor switch (PCSS). Both areas of work share the common goal of increasing the power density of switching power converters. Selective-area doping control using epitaxial regrowth processes by metal-organic chemical-vapor deposition (MOCVD) will be presented. Literature reports of GaN PN diodes where the p-layer is formed by epitaxial regrowth have, to date, indicated lower breakdown voltages and/or higher reverse leakage currents compared to continuously-grown PN junctions. The measurement of a “Si spike” at the regrown interface is commonly observed for GaN using characterization methods such as secondary ion mass spectroscopy (SIMS). Our initial studies to examine possible Si sources and the effects of interfacial Si on electrical performance have indicated that the MOCVD reactor is not a major source of Si. Additionally, intentional Si doping at the interface, with similar levels as measured by SIMS in regrown structures ([Si] = 6E16 cm-3 and [Si] = 5E17 cm-3), shows no noticeable degradation in forward- or reverse-bias electrical performance in planar c-plane devices fabricated on GaN substrates (Figure 1). However, regrown GaN diodes formed on m-plane GaN substrates have shown worse electrical performance than c-plane diodes. The increased chemical reactivity of the m-plane and the resulting increase in impurity uptake is likely a contributing factor, and studies are underway to identify the specific impurities responsible. Mitigation strategies to manage the challenges of the m-plane are also being investigated. Further, fabrication-induced damage due to dry-etching processes was shown to increase reverse leakage currents by at least an order of magnitude in planar, c-plane diodes. We will report on these results and the study of various surface treatment methods to reduce these leakage currents in planar, c-plane diodes. The details of a laterally-oriented, GaN PCSS switch will also be presented. PCSS technology is of interest for applications including high-voltage pulsed-power, electrical grid protection, and increased electrification through new build-outs of high-efficiency/performance power systems. PCSS devices have been demonstrated in several semiconductor materials, but GaN promises an improvement in switch capability due to its advantageous material properties. The device is operated by optically triggering photocarriers in a high-field area generated between two electrodes. Using commercially-available, semi-insulating GaN wafers, surface electrodes were fabricated with a 600 µm gap, and the device was optically triggered using a sub-bandgap 532 nm laser source. Low-voltage (low field) operation of the device demonstrated a linear photoconductive response. Increasing the field over a threshold value of 10-15 kV/cm demonstrated a “high-gain” or “lock-on” operating mode where current continues to flow after the laser source is turned off, if a sufficient field exists (Figure 2). This operating mode has been observed for other semiconductors, including GaAs, and is desirable for high-efficiency switch operation because of the low optical energy needed to turn on the switch. High-gain switching has been demonstrated with PCSS powers of 150 kW (1250 V, 120 A) triggered by optical energies as low as 35 µJ. Further increases in switch power are expected with device improvements. The information, data, or work presented herein was funded in part by the Advanced Research Projects Agency – Energy (ARPA-E), U.S. Department of Energy under the IDEAS and PNDIODES programs directed by Dr. Isik Kizilyalli. Sandia National Laboratories is a multi-mission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA-0003525. Figure 1

The development and testing of fiber-optically controlled trigger generators (TGs) based on high gain photoconductive semiconductor switches (PCSSs), constructed from high resistivity GaAs, are described in this paper. The TGs are optimized to trigger the high voltage switches (HVSs) in pulsed power systems, where they control the timing synchronization and amplitude variation of multiple pulse forming lines that combine to produce the total system output. Future pulsed power systems are even more dependent on triggering, as they consist of many more HVS and, in some cases, produce shaped pulses by independent timing of the HVS. The goal of the PCSS TG is to improve timing precision and replace high voltage trigger cables or line-of-sight optics with fiber-optic trigger control. The PCSS trigger has independent EMP-free timing control via 200-mum-diameter optical fibers. This design is simpler than other TG because optical isolation allows PCSS triggers to be remotely located near the HVS at any voltage. PCSS can improve the performance of prime power HVS, diverters, and diagnostics by supplying trigger pulses with subnanosecond jitter and rise time that are more precise and easily adjusted than the conventional TG. For pulse-charged HVS, the PCSS TG can generally derive their trigger energy from the stray fields of the HVS. High gain PCSS capabilities for producing pulsed power TG have been demonstrated previously (not all simultaneously): 220 kV, 8 kA, 350-ps rise time, 100-ns pulsewidth, 50-ps rms jitter, and 10-kHz repetition rate. Furthermore, PCSS has previously triggered a 300-kV trigatron with 100-ps rms jitter.

In a submerged environment, power cables may experience accelerated insulation degradation due to water - related aging mechanisms . Direct contact with water or moisture intrusion in the cable insulation s ystem has been identified in the literature as a significant aging stressor that can affect performance and lifetime of electric cables . Progressive reduction of the dielectric strength is commonly a result of water treeing which involves the development of permanent hydrophilic structures in the insulation coinciding with the absorption of water into the cable . Water treeing is a phenomenon in which dendritic microvoids are formed in electric cable insulation due to electrochemic al reactions , electromechanical forces , and diffusion of contaminants over time . These reactions are caused by the combined effect s of water presence and high electrical stress es in the material . Water tree growth follow s a tree - like branching pattern , i ncreasing in volume and length over time . Although these cables can be "dried out," water tree degradation , specifically the growth of hydrophilic regions, is believed to be permanent and typically worsens over time. Based on established research , water treeing or water induced damage can occurmore » in a variety of electric cables including XLPE, TR - XLPE and other insulating materials, such as EPR and butyl rubber . Once water trees or water induced damage form, the dielectric strength of an insulation materia l will decrease gradually with time as the water trees grow in length, which could eventually result in failure of the insulating material . Under wet conditions or i n submerged environments , several environmental and operational parameters can influence w ater tree initiation and affect water tree growth . These parameters include voltage cycling, field frequency, temperature, ion concentration and chemistry, type of insula tion material , and the characteristics of its defects. In this effort, a review of academic and industrial literature was performed to identify : 1) findings regarding the degradation mechanisms of submerged cabling and 2) condition monitoring methods that may prove useful in predict ing the remaining lifetime of submerged medium voltage p ower cables . The re search was conducted by a multi - disciplinary team , and s ources includ ed official NRC reports, n ational l aboratory reports , IEEE standards, conference and journal proceedings , magazine articles , PhD dissertations , and discussions with experts . The purpose of this work was to establish the current state - of - the - art in material degradation modeling and cable condition monitoring techniques and to identify research gaps . Subsequently, future areas of focus are recommended to address these research gaps and thus strengthen the efficacy of the NRC's developing cable condition monitoring program . Results of this literature review and details of the test ing recommendations are presented in this report . FOREWORD To ensure the safe, re liable, and cost - effective long - term operation of nuclear power plants, many systems, structures, and components must be continuously evaluated. The Nuclear Regulatory Commission (NRC) has identified that cables in submerged environments are of concern, particularly as plants are seeking license renewal. To date, there is a lack of consensus on aging and degradation mechanisms even though the area of submerged cables has been extensively studied. Consequently, the ability to make lifetime predictions for submerged cable does not yet exist. The NRC has engaged Sandia National Laboratories (SNL) to lead a coordinated effort to help elucidate the aging and degradation of cables in submerged environments by collaborating with cable manufacturers, utilities, universities, and other government agencies. A team of SNL experts was assembled from the laboratories including electrical condition monitoring, mat erial science, polymer degradation, plasma physics, nuclear systems, and statistics. An objective of this research program is to perform a l iterature r eview to gather a body of knowledge on prior research projects, technical papers, and literature related to cable degradation in a submerged environment. In addition, the information gathered from the literature review will be employed to gain insights for developing an aging coefficient, and to determine which condition monitoring techniques are capable of tracking cable degradation in a submerged environment. Moreover, the information gathered from the l iterature r eview will also be used to determine which approach or approaches are best suited to develop test methods for accelerated aging and condition m onitoring of medium voltage cables. In summary of this initial effort, s ignificant work has been performed on submerged cable insulation degradation; however, there is a lack of uniform theories and acceptance of chemical and physical pathways. This lack of fundamental understanding is coupled with the inability to make predictive statements about material performance in wet or submerged environments. S elect condition monitoring methods known to the industry are discussed in this report and a dditional co ndition monitoring methods were added in this effort based on recommendations from the Nuclear Energy Standards Coordinating Collaborative and available literature. This NUREG review provides additional clarity on the use of condition monitoring methods t o detect water - related damage to medium voltage cable and new methods and approaches proposed in academia and industry. In order t o ensure continued improvement in the efficacy of a cable condition monitoring program, continued research and development (R&D) efforts are necessary. R&D efforts should complement operations, iteratively improving condition monitoring policies, procedures and outcomes. Ideally, field and laboratory data enable improved understanding of material science which in turn inform s the development of new or improved condition monitoring methods and lifetime models. Finally, these improved methods and models aid in the refinement of condition monitoring policies and procedures.« less

The longevity of high gain GaAs photoconductive semiconductor switches (PCSSs) has been extended to well over ten million pulses by reducing the density of carriers at the semiconductor to metal interface. This was achieved by reducing the density in the vertical and lateral directions. The latter was achieved by varying the spatial distribution of the trigger light thereby widening the current filaments that are characteristic of the high gain switches. We reduced the carrier density in the vertical direction by using ion implantation. These results were obtained for currents of about 10 A, current duration of 3.5 ns, and switched voltage of /spl sim/2 kV. At currents of /spl sim/70 A, the switches last for 0.6 million pulses. In order to improve the performance at high currents, new processes such as deep diffusion and epitaxial growth of contacts are being pursued. To guide this effort we recorded open shutter, infra-red images, and time-resolved Schlieren images of the current filaments, which form during high gain switching. We measured, under varying conditions, a carrier (electrons or holes) density that ranges from 3/spl times/10/sup 17/ cm/sup -3/ to 6/spl times/10/sup 18/ cm/sup -3/.

Optically activated GaAs switches operated in their high-gain mode are being used or tested for pulsed power applications as diverse as low-impedance, high-current firing sets in munitions; high impedance, low-current Pockels cell or Q-switch drivers for lasers; high-voltage drivers for laser diode arrays; high-voltage, high-current, compact accelerators; and pulsers for ground penetrating radar. This paper will describe the properties of high-gain photoconductive semiconductor switches (PCSS), and how they are used in a variety of pulsed power applications. For firing sets, we have switched up to 7 kA in a very compact package. For driving Q switches, the load is the small (30 pF) capacitance of the Q switch which is charged to 6 kV. We have demonstrated that we can modulate a laser beam with a subnanosecond rise time. Using PCSS, we have demonstrated gain switching a series-connected laser diode array, obtaining an optical output with a peak power of 50 kW and a pulse duration of 100 ps. For accelerators, we are using PCSS to switch a 260 kV, 60 kA Blumlein. A pulser suitable for use in ground-penetrating radar has been demonstrated at 100 kV, 1.3 kA. This paper will describe the specific project requirements and switch parameters in all of these applications, and emphasize the switch research and development that is being pursued to address the important issues.

n/a

Summary form only given. Pulsed power gas switches operate under extreme conditions holding off voltages at millions of volts and greater, conducting currents at 100's of kiloamps, and can reliably trigger with less than 10ns of jitter at optimal operating conditions. However as Martin and Braginskii had noted the impedance of these switches is dynamic and a function of many parameters such as gas pressure, current, and the number of channels. In systems such as PROTOGEN or URSA MINOR many gas switches are utilized to enable optimal performance and pulse shaping. A challenge with either system is that, if the switches are tightly coupled and one switch triggers early, then the fields on the second switch change dramatically and can move the second switch into a new operating regime that will result in a different dynamic resistance. This paper presents a least squares approach to extracting the Martin and Braginskii dynamic resistance from measured data. Results from this approach are compared to previous approaches utilizing piecewise nonlinear least squares.

Summary form only given. Methods of producing multiple current-sharing filaments (MCSF) in GaAs photoconductive semiconductor switches (PCSSs) have produced as many as 30 filaments per switch with a spacing of 330 μm. As the approaches to triggering and isolating MCSF mature, the replacement of high current, conventional pulsed power switches with banks of MCSF PCSSs capable of switching tens of kiloamps become feasible and economical. Multiple banks of MCSF PCSS can eventually produce optically controlled pulsed power systems with faster rise-times, shorter pulses, and higher peak powers using more economical switches with device lifetimes of over one million pulses.This paper will report results from three types of MCSF approaches: (1) line-of-sight optics focused with cylindrical micro-lens arrays into bright narrow lines of light on the surface of the PCSS; (2) high-reflectivity dielectric optical masks which produce shadows of bright narrow lines of light on the surface of the PCSS; and (3) etched-steps in the surface of the PCSS to divide up the illuminating light and isolate the filaments. These approaches are being tested to switch 2-3 kA with three 1cm square GaAs PCSS. Switching parameters, approximate device lifetimes, and current-sharing capability are being measured and evaluated for each of these triggering techniques. Many other approaches have been considered previously, and their potential for other specific applications will be discussed. Increasing PCSS current density through the development of longer-lived PCSS with higher currents per filament is research that has continued in parallel with the multi-filament triggering work. This however is primarily an interface issue where the contact metal meets the GaAs. We will discuss new electrical contact geometries capable of reducing current crowding. Thereby lowering the peak current density to enable higher total filament current. Theoretical bases for improving contacts will also be discussed.

Summary form only given. Material interfaces often reveal complicated and interesting phenomena and physics during electrical breakdown. In this talk, the physics at interfaces in two specific applications will be discussed. One involves the insulator-electrode interface of photoconductive semiconductor switch (PCSS) devices. Another is the air-dielectric interface found in many electrical breakdown phenomena, such as lightning arresters. In this case, the role of high dielectric permittivity in initiating breakdown will be a focus. The physics to be described will be examined in a series of transport calculations in which avalanche breakdown of solids and gases is included. These continuum calculations produce information such as the time-resolved electron, hole and ion densities following switch triggering by an overvoltage or radiation. The results will be used to examine the initial breakdown phenomena, such as thermionic and field emission at the interface, during the start of breakdown caused by an overvoltage. In addition, phenomena, such as Auger recombination, that are involved in charge collection and electrode will be included in these calculations. Finally, thermal breakdown along an air-insulator interface will be described in relation to experiments that reveal breakdown that "hugs" such an interface. These illustrative calculations will be compared with data.

Summary form only given. Ferroelectric materials exhibit a hysteretic response similar to the B-H curve of magnetic materials. With ferroelectric materials the polarization-electric field (P-E) hysteretic loop is typically measured at low frequencies. However, the material behavior at high frequencies is less understood. To address this information gap, a test bed has been created to characterize non-linear material behavior at high frequencies and high voltages. This paper will report testbed goals in addition to design, assembly, analysis and issues. Preliminary results will also be presented from testing commercially available ferroelectric capacitors and materials, including 10 nf non-linear BaTiO-based capacitors and 3 mm thick lead zirconate titanate (PZT)-based materials. The main goal of the test bed is to drive a 1-50 nF ferroelectric sample around its polarization-electric field (P-E) loop with a low inductance circuit (10 nH) and high bandwidth (1 GHz) diagnostics at voltages up to 20 kV and currents up to 10 kA. A secondary goal is to drive the sample through P-E sub-loops to measure the sample response at high frequencies within the P-E loop. Polarization of the sample is obtained by integrating the current passing through the sample and dividing by its cross-sectional area. The electric field is determined by measuring the voltage across the sample and dividing by its thickness. Simulation of the circuit and analysis of the data require modeling of the sample behavior internal to the P-E loop and along its boundaries. A hyperbolic tangent (Fermi function) model that depends on the initial values of P and E and the polarity of the current provides a reasonable model for low bandwidth measurements of most materials. However P-E loops with sharper corners and/or a strong variation with frequency may require a model with a higher order exponential dependence, and a free parameter to characterize the sharpness of the corners independent of the slope of the loop through zero polarization. Simulation and analysis of our preliminary results will be discussed.

Dynamic materials properties research at Sandia National Laboratories has resulted in research that is advancing capabilities in precision programmable pulsed power systems operating in multi-mega amp regimes. Programmable pulse shaping capabilities require the gas switches in these systems to perform over a large range of dynamic operating conditions. Runtime, jitter, and the number of channels formed are all impacted by the conditions of these switches at the time of trigger. This paper provides a model and analysis of a 200 kV gas switch designed for linear transformer drivers operating at percentages of self break ranging from 45% to 100%. This work expands on the research performed by T.H. Martin and S.I. Braginskii.

High current switching is the most critical challenge remaining for photoconductive semiconductor switch (PCSS) applications in Pulsed Power. Many authors have described the advantageous properties of high gain PCSS such as, low optical trigger energy and inductance, sub-nanosecond risetime and jitter, optical isolation and control, pulsed or DC charging, and long device lifetime, provided the current per filament is limited to 20–30A for short pulse (10–20ns) applications [1,2]. Low energy optical triggering, long device lifetime, and current filaments are related features of high gain PCSS that make high current switching a challenge. Since the location and number of current filaments can be controlled with parallel “lines” of optical pulses across the insulating gap, the problem of high current, multi-filament PCSS switching is essentially the problem of producing a reliable, efficient, multi-line, optical delivery system [3].

As the demand for greater flexibility and increased energy density moves highly interactive components closer together in pulsed power systems, the development of technologies for less complex and more robust system designs is critical. A key system component that impacts these goals is the trigger generator (TG). Inexpensive, compact, fiber-optically controlled TGs that deliver trigger pulses with sub-nanosecond jitter have been created with photoconductive semiconductor switches (PCSSs) [1]. A further simplification in pulsed power design is to move from pulsed to DC charged components. This paper reports results from DC charged GaAs PCSSs with 0.25–1.0 cm gaps extending previously reported results on smaller devices at 3 kV [2] to a new regime of 100 kV.

Prior to this research, we have developed high-gain, GaAs, photoconductive semiconductor switches (PCSSs) to trigger 50–300 kV high voltage switches (HVSs). We have demonstrated that PCSSs can trigger a variety of pulsed power switches operating at 50–300kV by locating the trigger generator directly at the HVS. This was demonstrated for two types of DC-charged trigatrons and two types of field distortion mid-plane switches, including a ±100 kVDC switch produced by the High Current Electronics Institute (HCEI) used in the linear transformer driver. The lowest rms jitter obtained from triggering a HVS with a PCSS was 100 ps from a 300 kV pulse-charged trigatron. PCSSs are the key component in these independently timed, fiber-optically controlled, low jitter trigger generators (TGs) for HVSs. TGs are critical sub-systems for reliable, efficient pulsed power facilities because they control the timing synchronization and amplitude variation of multiple pulse forming lines that combine to produce the total system output. Future facility scale pulsed power systems are even more dependent on triggering, as they consist of many more triggered HVSs and produce shaped-pulses by independent timing of the HVSs. As pulsed power systems become more complex, the complexity of the associated trigger systems also increases. One means to reduce this complexity is to allow the trigger system to be charged directly from the voltage appearing across the HVS. However, for slow or DC charged pulsed power systems this can be particularly challenging as the DC hold off of the PCSS dramatically declines. This paper presents results seeking to address HVS performance requirements over large operating ranges by triggering using a pulsed charged PCSS based TG. Switch operating conditions as low as 45% of self break were achieved. A DC charged PCSS based TG is also introduced and demonstrated over a 39 kV – 61 kV operating range. DC charged PCSS allow the TG to be directly charged from slow or DC charged pulsed power systems. GaAs PCSSs and neutron irradiated GaAs (n-GaAs) PCSSs were used to investigate the DC charged operation.

Three categories of circuits were explored to induce fast recovery, after lock-on by temporarily reducing the field across the switch. Measurements of recovery times from 35 to 80 ns, multiple monopolar and bipolar bursts at 5-40 MHz, and hold-off fields ranging from 5-44 kV/cm (corresponding to 15-66 kV across individual switches) are presented. A comparison is made of the different circuits used to induce recovery from lock-on, and the various factors that influence the recovery are discussed. >

Progress toward the triggering of high-power photoconductive semiconductor switches (PCSSs) with laser diode arrays, is reported. An 850-W optical pulse from a laser diode array was used to trigger a 1.5-cm-long switch that delivered 8.5 MW to a 38.3- Omega load. Using 166-W arrays, it was possible to trigger a 2.5-mm-long switch delivering 1.2 MW with 600-ps rise-times at pulse repetition frequencies of 1 kHz. These 2.5-mm-long switches survived 10/sup 5/ pulses at 1.0 MW levels. In single-pulse operation, up to 600 A was switched with laser diode arrays. The goal is to switch up to 5 kA in a single-shot mode and up to 100 MW repetitively at up to 10 kHz. At electric fields below 3 kV/cm GaAs switches are activated by creation of one electron-hole pair per photon. This linear mode demands high laser power and, after the light pulse, the carriers recombine in nanoseconds. At higher electric fields GaAs acts as a light-activated Zener diode. The laser light generates carriers as before, but the field induces gain such that the amount of light required to trigger the switch is reduced by a factor of up to 500. The gain continues until the field across the sample drops to a material-dependent lock-on field. The gain in the switch allows for the use of laser diodes. >

Experiments have been performed to develop photoconductive semiconductor switches (PCSSs) for pulsed power applications that cannot be implemented with traditional high-power switching technology. A scalable lateral PCSS configuration has been tested and has demonstrated a potential for faster risetimes, less jitter, lower inductance, faster recovery, and optical triggering for new pulse power projects. These switches have been used as both closing and toggling switches at repetition rates up to 40 MHz. A high-field, gain mechanism (lock-on) was explored and tested which may eliminate the major disadvantage of this type of switch, its requirement for large optical trigger energies. >

Summary form only given. Pulsed power systems designed with many switches or many different switch trigger times are heavily dependent upon the triggering systems. Conceptual pulsed power systems for fusion and isentropic compression are being designed with hundreds to hundreds of thousands of switches. Some have to be triggered simultaneously and some will need to be programmable and staggered in time. Trigger system cost, performance, and lifetime will be driving factors in these systems. For example, designers of repetitive pulsed power systems considered for fusion are seeking switch lifetimes >107 shots. A solution that addresses the performance requirements is the use of optically triggered photoconductive semiconductor switches (PCSS), which excel at delivering low jitter, fast-rise-time pulses. Lifetimes for these devices have been demonstrated to be in the >> 106 shot range at lower currents (tens of amps), but drop to the 10~ shot range when the current per filament is hundreds of amps. To increase the switch reliability at higher current levels necessitates further investigation of device reliability and failure mechanisms. Results from lifetime studies on GaAs PCSS devices indicate that the majority of failures occur near the anode. However, switch fabrication variables can shift this failure mechanism to a short-life, cathode-limited condition. A statistical sample population of PCSS switches in a circuit designed to deliver 500V/100A, 10ns pulses is shown to have a single log-normal distribution of shot life over a range of testing temperatures and voltages. Switches for higher operating voltage (70 kV) were fabricated in multiple split lots to identify the influence of design and processing parameters on the switch lifetime. It was found that these larger-gap 70kV switches exhibit the same anode-limited lifetime as the 500V devices. This paper discuss results from the lifetime studies, including variables in the fabrication process (annealing temperature, pad-metal thickness, length of contact-metal extension beyond the pad-metal, and inclusion of a silicon nitride passivation/anti-reflection coating) that affect reliability, and provide strategies for improving shot life.

The cost and complexity of pulsed power systems continue to increase as researchers strive for more precise control and flexibility. Systems with large numbers of switches or large numbers of switch trigger times result in the need for more cost effective high performance triggering systems. This need can be addressed by photoconductive semiconductor switches (PCSS) that have demonstrated sub-nanosecond jitter with up to 300kV switches. Previous work has motivated continued research into the trigger system parameters that allow for reliable cost effective operation.

This paper describes results from a 3 year project to develop fiber-optically controlled trigger generators (TG) from high gain photoconductive semiconductor switches (PCSSs) for high voltage switches (HVSs) in pulsed power applications. Triggers for HVSs are critical components for reliable, efficient pulsed power systems because they control the timing synchronization and amplitude variation of multiple pulse forming lines that combine to produce the total output. Proposed pulse power systems are even more dependent on triggering as they consist of many triggered HVSs and produce shaped-pulses by independent timing of the HVSs.

Trigger systems are becoming increasingly important in pulsed power systems with large numbers of switches or large numbers of different switching times. Performance can be critical with demands for fast rise-times, sub-nanosecond jitter, and long lifetimes . In particular, component lifetimes affect maintenance costs and the available operational time of the system. High gain photoconductive semiconductor switches (PCSSs) deliver many of the desired properties including optical-isolation, 350 ps risetime, 100 ps r-m-s jitter, scalability to high power (220 kV and 6 kA demonstrated), and device lifetimes up to 10b shots with 20 A 5 ns wide pulses. However, higher current and longer pulse applications can drastically reduce device lifetime. For typical single shot pulsed power applications, lifetimes of several thousand shots are required and much longer-lived switches are required for repetitive pulsed power applications. The key parameters that impact PCSS lifetime are voltage, current, and pulse width. Voltage affects the lifetime of a lateral switch when the electric field near the surface of the switch approaches the surface breakdown limit, which is approximately 100 kV/cm for pulse charged switches under transformer oil or Fluorinert (a liquid dielectric). The current in high-gain GaAs PCSS always forms filaments, so this lifetime dependence can be considered in terms of the current per filament, and most of our lifetime testing is done with switches producing a single main filament. Since PCSS can have sub-nanosecond risetime and jitter, most of our interest in lifetime is for switches that produce 1-100 ns long pulses. These requirements lead us to testing high gain PCSS in high speed, 50 ohm transmission line, discharge circuits that can deliver up to 300 A. Control of the PCSS is achieved with a fiber-coupled laser directed between the PCSS metal contact pads resulting in randomly formed filaments (one primary). Devices demonstrating long lifetimes are operated at up to a few kilohertz, whereas higher current and longer pulse tests are operated at lower repetition rates. In all cases, rep-rates are below the limit where bubbles form in the liquid dielectric. New switches are always tested at 20 A for direct comparison to the lifetime data that we have accumulated over the last 20 years. Fligher current tests are performed to predict switch lifetimes for specific applications that don't require such long device lifetimes. This paper will discuss testing procedures, circuits, and dramatic changes in PCSS component lifetime due to contact methods (e.g. soldering versus ribbon bonding).

We are developing advanced optically-activated solid-state switch technology for Firing Sets. Advanced switch development at Sandia has demonstrated multi-kA/kV switching and the path for scalability to even higher current/power, resulting in good prospects for sprytron replacement and other even higher current pulsed power switching applications. Realization of this potential requires development of new optical sources/switches based on key Sandia photonic device technologies: vertical-cavity surface-emitting lasers (VCSELs) and photoconductive semiconductor switch (PCSS) devices. The key to increasing the switching capacity of PCSS devices to 5kV/5kA and higher has been to distribute the current in multiple parallel line filaments triggered by an array of high-brightness line-shaped illuminators [Mar, A., et al., 2001]. This was limited by commercial mechanically-stacked edge-emitting lasers, which are difficult to scale and manufacture with the required uniformity. In VCSEL arrays, adjacent lasers utilize identical semiconductor material and are lithographically patterned to the required aspect ratio. However, we have demonstrated that good optical uniformity in rectangular-aperture (e.g. 5-by-500 mum) VCSELs is difficult to achieve due to the lack of optical confinement in the long dimension. We have demonstrated line filament triggering using 1-D VCSEL arrays to approximate line generation. These arrays of uncoupled circular-aperture VCSELs have fill factors ranging from 2% to 50%. Using these arrays, we are developing a better understanding of the illumination requirements for stable triggering of multiple-filament PCSS devices. In particular, we are examining the dependence of filament formation versus the illumination fill factor and spatial brightness along the length of the filament. Ultimately, we will apply effective index techniques, pioneered at Sandia for leaky-mode VCSELs, to create a lateral photonic lattice that selects a single transverse mode with high brightness and uniformity for even higher fill factors and illumination unformity [Zhou, D., et al., 2000]. These sources will be developed and tested with complementary PCSS designs employing interdigitated multifilament contacts for high-power switching.

Pulsed power systems are continually being designed to increase performance and flexibility in order to meet new needs. Drivers for fusion, equation of state research and radiography are just a few of the scientific areas driving accelerated pulsed power development. Improved system timing and jitter requirements, along with pulse shaping capabilities through independent timing of individual switches, are placing higher demands upon high voltage (HV) switch performance, triggering system performance and cost. Conventional trigger systems require high voltage trigger cables or line-of-sight optics that complicate design, demand space, require extensive maintenance, and significantly impact system cost. With electrical triggers, large diameter, high voltage transmission line cables must be fed through high field regions. With optical triggers, line-of-sight optics must focus high energy laser beams to the interior of the switches with clean, rigidly-mounted, shock-withstanding optics. This paper reports on progress in the development of fiber optically triggered photoconductive semiconductor switches (PCSS) for triggering high voltage pulsed power switches with improved precision. This technology can eliminate the need for large-diameter trigger cables and line-of-sight optics. It also has the potential of significantly reducing the cost of trigger generation systems. It has the potential to improve the performance of prime power switches, diverters, and diagnostics because of the low-jitter sub-nanosecond rise times. Test results will be presented that demonstrate sub-nanosecond jitter from a PCSS, less than 1.

We have presented experimental and simulation results for a SI GaAs and discussed the results in terms of the dominance of the trap levels in determining the I–V characteristics of the devices. Experimental results and simulation analyses indicate that, when subjected to neutron irradiation, there is an improvement in the hold-off characteristics of the devices, which may lead to higher voltage operation and thus an improvement in the rise time of the device. The maximum energy transferred to the load would increase, thereby improving the overall device characteristics. The filamentary conduction processes of the PCSS are more susceptible to failure if a large number of defect sites are formed in the device during a radiation transient. This is mainly because defects are expected to bring about inhomogeneous conduction.

Attempts to use photoconductive semiconductors as high-power (10-100 kV, 0.1-2 kA) toggling switches with recovery times of 5-100 ns have stimulated the exploration of their recovery mechanisms. We have observed that optically triggered GaAs switches exhibit "lock-on," i.e., when triggered, they do not recover as long as they are holding more than 4-8 kV/cm. Experiments are being performed to determine the minimum recovery time of these switches after lock-on. by immediately reducing their fields or currents. As an alternative to GaAs, Si is a semiconductor that does not exhibit lock-on. It has a very long recovery time (> 100 /spl mu/s) that can be shortened to less than 100 ns with highly concentrated gold doping (/spl ges/10/sup 15//cm/sup 3/). Device models that predict the behavior of PCSS and experiments on the recovery of GaAs and Au:Si switches are presented.

A high peak power impulse pulser that is controlled with high gain, optically triggered GaAs Photoconductive Semiconductor Switches (PCSS) has been constructed and tested. The system has a short 50 ohm line that is charged to 100 kV and discharged through the switch when the switch is triggered with as little as 90 nJ of laser energy. The authors have demonstrated that the GaAs switches can be used to produce either a monocycle or a monopulse with a period of total duration of about 3 ns. For the monopulse, the voltage switched was above 100 kV, producing a peak power of about 48 MW to the 30 ohm load at a burst repetition rate of 1 kHz. The laser that is used is a small laser diode array whose output is delivered through a fiber to the switch. The current in the system has rise times of 430 ps and a pulse width of 1.4 ns when two laser diode arrays are used to trigger the switch. The small trigger energy and switch jitter are due to a high gain switching mechanism in GaAs. This experiment also shows a relationship between the rise time of the voltage across the switchmore » and the required trigger energy and switch jitter. The time evolution of the current filaments in an optically triggered, high gain GaAs switch was studied by recording the infrared photoluminescence from the filaments. When the system is triggered with two laser diode arrays that are activated within 1 ns of each other, two current filaments are observed. By delaying one laser with respect to the other, the evolution of the filament was recorded in a time resolved fashion. The filament that is triggered first crosses the switch, the voltage drops and the other filament ceases to grow. By varying the delay between the trigger lasers, the tip velocity is measured to be up to 5.9 {+-} 1 {times} 10{sup 9} cm/s.« less

Switching properties are reported for high gain photoconductive semiconductor switches (PCSS). A 200 ps pulse width laser was used in tests to examine the relations between etectric field, rise time, delay, and minimum optical trigger energy for switches which reached 80 kV in a 50 /spl Omega/ transmission line with rise times as short as 600 ps. Infrared photoluminescence was imaged during high gain switching providing direct evidence for current filamentation. Implications of these measurements for the theoretical understanding and practical development of these switches are discussed.

One of the most important limitations of Photoconductive Semiconductor Switches (PCSS) for pulsed power applications is the high laser powers required to activate the switches. In this paper, we discuss recent developments on two different aspects of GaAs PCSS that result in reductions in laser power by a factor of nearly 1000. The advantages of using GaAs over Si are many. First of all, the resistivity of GaAs can be orders of magnitude higher than that of the highest resistivity Si material, thus allowing GaAs switches to withstand dc voltages without thermal runaway. Secondly, GaAs has a higher carrier mobility than Si and, thus, is more efficient (per carrier). Finally, GaAs switches can have naturally fast (ns) opening times at room temperature and low fields, microsecond opening times at liquid nitrogen temperature of 77 K, or, on demand, closing and opening at high fields and room temperature by a mechanism called lock-on (see Ref. 1). By contrast, Si switches typically opening times of milliseconds. The amount of laser light required to trigger GaAs for lock-on, or at 77 K, is about three orders of magnitude lower than at room temperature. In this paper we describe the study of lock-on inmore » GaAs and InP, as well as switching of GaAs at 77 K. We shall show that when GaAs is switched at 77 K, the carrier lifetime is about three orders of magnitude longer than it is at room temperature. We shall explain the change in lifetime in terms of the change in electron capture cross section of the deep levels in GaAs (these are defect or impurity levels in the band gap). In the second section, we describe the lock-on effect, now seen in GaAs and InP, and at fields as high as 70 kV/cm. We show how lock-on can be tailored by changing the GaAs temperature or by neutron bombardment. In the third section, we discuss possible lock-on mechanisms. 5 refs., 5 figs.« less

Characteristics of current filamentation are reported for high gain photoconductive semiconductor switches (PCSS). Infrared photoluminescence is used to monitor carrier recombination radiation during fast initiation of high gain switching in large (1.5 cm gap) lateral GaAs PCSS. Spatial modulation of the optical trigger, a 200--300 ps pulse width laser, is examined. Effects on the location and number of current filaments, rise time, and delay to high gain switching, minimum trigger energy, and degradation of switch contacts are presented. Implications of these measurements for the theoretical understanding and practical development of these switches are discussed. Efforts to increase current density and reduce switch size and optical trigger energy requirements are described. Results from contact development and device lifetime testing are presented and the impact of these results on practical device applications is discussed.

Experiments were performed to determine the applicability of photoconductive semiconductor switches (PCSS) for use as output switches in subnanosecond pulse for EMP simulators. Lateral switches made of both Gallium Arsenide and Silicon with 1.5 cm long insulating regions immersed in Fluorinert were tested in a 50 ohm tri-plate transmission line geometry. Mode locked and Q- switched lasers were used to trigger both a gas switched Marx generator which pulse-charged the transmission line in 100- 150 ns and to illuminate the PCSS via an optical delay line. Illuminating beam energies and electric field strengths at switchout were varied to determine minimum risetimes and light energies required for triggering. The GaAs switches were operated in the high gain (lock- on) mode. Risetimes as fast as 600 ps were observed using a mode locked laser and 700 ps using a Q-switched laser. The minimum light energy required to trigger GaAs was 22/spl mu/J and the highest switched fields for both GaAs and Si is about 60 kV/ cm.

The discovery of current filaments in GaAs photoconductive semiconductor switches (PCSS) raised concerns about the location and density of the current distribution during high gain switching. This paper describes experiments using fiber-optic coupled laser diode arrays (LDA) to control the location and number of current filaments in GaAs PCSS. Infrared (IR) images of the recombination radiation, that is emitted from the surface of the PCSS, show precisely where the current is concentrated. These were recorded as the fiber-optic configuration and switch test parameters were varied. Up to twenty individually-coupled laser diodes and fibers were used to initiate multiple, simultaneous filaments. Fibers were configured in a row both parallel and perpendicular to the direction of current. Two types of experiments were performed: high current and high voltage. System impedances of 0.25 {Omega} and 50 {Omega} were used with switch currents up to 5.2 kA and initial charging voltages up to 100 kV, respectively. Switch sizes were 1{times}5 mm{sup 2} and 15{times}38 mm{sup 2} (parallel {times} perpendicular to the current). With the 0.25 {Omega} system charged to 4 kV (40 kV/cm across the PCSS), individual filaments were triggered with 150 nJ of optical trigger energy. In this case, filaments were triggered atmore » both ends near the contacts with pairs of 250 {mu}m fibers, each of which delivered 75 nJ in 140 ps. The use of multiple fiber pairs to distribute the current more uniformly across the switch improved switch performance. It allowed higher total currents to be switched with narrower switches. With the 50 {Omega} system charged to 100 kV (67 kV/cm across the PCSS), individual filaments were triggered with as little as 90 nJ of optical trigger energy. The trigger pulse was delivered in 2 ns through one fiber which was 400 {mu}m in diameter.« less

The lock-on effect is a high gain, high field switching mechanism that has been observed in GaAs and InP. This switching mode is exciting because the amount of light required to trigger it is small when compared to triggering the same switch at low fields. For this reason we can use laser diode arrays to trigger high voltages, currents and power. This paper will describe the lock-on effect, and our recent experiments to understand the effect. We will show that impact ionization from deep levels cannot account for the observed current densities, delays, and rise times unless a second mechanism is invoked. We will also describe our applications for laser diode array triggered lock-on switches, the best results that illustrate our potential for the application, and the studies carried out to improve the lifetime and current carrying capability of the switches.

Photoconductive semiconductor switches (PCSS's), such as optically-triggered GaAs switches, have been developed for a variety of applications. Such switches exhibit unique properties as a consequence of lock-on, a phenomenon associated with the bistable switching of these devices. In this paper, lock-on is explained in terms of collective impact ionization. Furthermore, the effect of defects on the performance of these devices is investigated.

Terahertz radiation from optically-induced plasmas on metal, semiconductor, and dielectric surfaces is compared to electron-hole plasma radiation from GaAs and Ge. Electro-optic sampling and electric-field probes measure radiated field waveforms and distributions to 0.350 THz.

Summary form only given. Short pulse lasers can be used to generate two types of short-lived, radiating plasmas. When absorbed by a semiconductor, a sudden optical pulse creates a charge-neutral, electron-hole plasma that radiates as the carriers relax in local crystal, surface, or applied electric fields. When focused onto the surface of an insulator or metal, the electric fields in an intense optical pulse break down the solid surface and surrounding gas. The charged particles in this surface-breakdown plasma radiate as they are created at the solid-gas (air) interface. The sudden creation of either of these types of plasmas with a sub-picosecond optical pulse produces a non-equilibrium distribution of charges that radiates with an extremely wide bandwidth in the THz regime. This paper reports the radiated electric field intensities, spectra, and associated radiation patterns for electron-hole plasmas from gallium arsenide and silicon, and for surface-breakdown plasmas from aluminum, glass (BK-7), gallium arsenide, and silicon. The plasmas are created and diagnosed with a chirped-pulse regenerative amplifier laser that produces a 1 kHz train of 120 fs wide pulses with up to 2 mJ of energy per pulse at a wavelength of 800 nm. To explore the wide bandwidth of these plasma radiation sources, two electric field diagnostics were used. A conventional electric field probe (D-dot) with a 40 GHz bandwidth and a 70 GHz sampling oscilloscope measured the low frequency tail of the radiation spectrum. To measure the high frequency range of the spectrum, electro-optic crystals (zinc telluride) sampled the radiated electric field as it co-propagated with an optical probe beam that was split off the excitation (pump) beam. The bandwidth of the electro-optic sampling is limited by the thickness of the electro-optic crystals, so a range of thicknesses was employed to demonstrate source-limited measurements. The fundamental differences between electron-hole and surface-breakdown plasmas will be discussed. Similarities and contrasts in radiation patterns and efficiencies will be described for all of these plasma radiation sources. Simple models for the picosecond carrier motion in these sources will be presented and compared with measurements. The advantages and challenges of using a kilohertz, millijoule, short-pulse laser for plasma generation and electro-optic sampling will also be mentioned.

We have studied the feasibility of an innovative device to sample 1ns low-power single current transients with a time resolution better than 10 ps. The new concept explored here is to close photoconductive semiconductor switches (PCSS) with a Laser for a period of 10 ps. The PCSSs are in a series along a Transmission Line (TL). The transient propagates along the TL allowing one to carry out a spatially resolved sampling of charge at a fixed time instead of the usual timesampling of the current. The fabrication of such a digitizer was proven to be feasible but very difficult.

Optically activated high voltage switches are commonly used in pulsed power systems for reliable low jitter, multichannel and multiswitch (low inductance) applications. In addition to low jitter switching, optical activation provides a high degree of electrical isolation between the triggering and switching power systems simplifying pulsed power design. The disadvantages of optical triggering for large-gap gas switches are the optical energy, line-of-sight optics, and system maintenance required to obtain reliable operation. This paper describes two technologies which can reduce or eliminate these disadvantages and provide more flexible optically activated switches for pulsed power systems. One approach is to reduce the optical trigger energy requirement for gas gap switches. Shorter optical pulses require less energy to initiate a plasma discharge. An experiment is being assembled to trigger a 50-100 kV gas gap switch with 120 fs wide optical pulses. Lower trigger energy has also been demonstrated by the introduction of metallic aerosols into a gas gap W. Frey (1997). The apparatus will be added to this experiment to reproduce these results and determine the optical energy and power density requirements over a range of wavelengths and pulse widths. The status of this experiment will be discussed. A second approach uses solid state switching in two configurations: (1) as the main switch and (2) as a trigger. High-gain photoconductive semiconductor switches (PCSS) are practical for some direct pulsed power switching applications. We have demonstrated switching up to 220 kV and 8 kA. Higher power optically activated switching can be obtained by combining solid state and gas gap switching technologies. Multimegavolt (MMV) switches can be triggered with fiber-optically triggered, remotely located PCSS. By placing the compact PCSS trigger extremely close to the trigger point, reliable, low-jitter, high power switching is achieved with low energy fiber-optic trigger systems that can easily be controlled and adjusted from a remote control center. Power for the trigger system can be derived from the electrical fields near the trigger, so all electrical cables to the trigger system are eliminated and replaced with 100 micron diameter fibers that trigger and monitor the operation of the system. Results from experiments with PCSS triggered gas gap switches and the design for a PCSS triggered multimegavolt switch will be reported. PCSS switching properties including new picosecond pulse results and fabrication procedures for improved longevity will also be described. A 120 fs wide 780 nm laser pulse was used to radiate THz bandwidth pulses with a GaAs PCSS operating in the linear mode. New approaches for PCSS contact fabrication are being developed and tested to simplify the growth procedure, increase the current per filament capability, and improve device longevity. Progress continues to make PCSS a more useful component for pulsed power applications.

Gallium arsenide (GaAs) photoconductive semiconductor switches (PCSS) have been studied as an enabling technology for a variety of applications at both the Air Force Research Laboratory and Sandia National Laboratories. High gain PCSS can be triggered with small laser diodes or laser diode arrays. The requirements of these applications require the switching of high voltage in sub-nanosecond time with low temporal jitter of the switches relative to the trigger laser. There have been several configurations and sizes of these switches studied by the Air Force Research Laboratory over the last several years. The most recent designs are with small structures where the electrical contacts are placed on opposite sides of the bulk material. This configuration allows for different electrical conditions on either side depending on the nature of the semiconductor structure; i.e., p-i-n or n-i-n. In addition to the type of structure used and geometry of the contacts, the performance of these switches (switch time, voltage, and jitter) is dependent on the thickness of the GaAs. Several thicknesses have been studied during the past year. This paper reports on the results of several studies to investigate the ultra-fast switching properties of these structures.

Photoconductive semiconductor switches (PCSSs), such as optically-triggered GaAs switches, have been developed for a variety of pulsed power applications. Such switches exhibit unique properties associated with lock-on, a phenomenon associated with bistable switching. In this paper, lock-on is explained in terms of collective impact ionization.

The authors discuss the use of PCSSs (photoconductive semiconductor switches) to create bursts of high-power pulses above 10 MHz. Three methods are considered; (1) circuit reflections which are produced when a switch closes and opens on a charged transmission line; (2) directly driven PCSSs which follow an RF modulated light source, and (3) oscillations induced by the negative-differential-conductivity characteristics of some semiconductors at high fields. Of these techniques, directly driven PCSSs which use RF-modulated light sources appear to be the most straightforward to develop. Experiments to test feasibility of PCSSs to drive a pulsed linear induction accelerator are described and preliminary results are given. >

It is shown that Si photoconductive semiconductor switches can be used to switch high voltages (up to 123 kV), high fields (up to 82 kV/cm) and high currents (2.8 kA). The ability of the samples to withstand this type of high-voltage, high-current switching depends on the way in which the current penetrates the semiconductor. The appropriate use of water or contacts greatly improves the switching capability. It is also shown that the wafers can support large currents (4.0 kA for GaAs and 2.8 kA for Si) and large linear current densities (3.2 kA/cm for GaAs and 1.4 kA/cm for Si). For GaAs this linear current density corresponds to about 1 MA/cm/sup 2/ given a penetration depth of about 10/sup -3/ cm. >

It is shown that Si photoconductive semiconductor switches (PCSSs) can be used to switch high voltages (up to 123 kV), high fields (up to 82 kV/cm) and high currents (2.8 kA). The ability of the samples to withstand this type of high-voltage, high-current switching depends on the way in which the current penetrates the semiconductor. The appropriate use of water or contacts greatly improves the switching capability. It is also shown that the wafers can support large currents (4.0 kA for GaAs and 2.8 kA for Si) and large linear current densities (3.2 kA/cm for GaAs and 1.4 kA/cm for Si). For GaAs, this linear current density corresponds to about 1 MA/cm/sup 2/, given a penetration depth of about 10/sup -3/ cm. It is determined that the lock-on phenomenon can be triggered with light of varying photon energy to reach a lock-on field that is both impurity-concentration and sample-temperature dependent. >

Pictures are presented of the time evolution of current filaments during optically triggered, high gain switching in GaAs. Two filaments are triggered with two laser diode arrays and the time delay between them is varied. When the filament that is triggered first crosses the switch the voltage drops and the other filament ceases to grow. By varying the delay between the lasers, the tip velocity is measured to be 2±1×109 cm/s, 100 times larger than the peak drift velocity of carriers in GaAs. This observation supports switching models that rely on carrier generation at the tip of the filament.

This paper describes research and development of high gain GaAs photoconductive semiconductor switches (PCSS) for two very different types of applications: compact, repetitive accelerators and short pulse, active optical sensors. The accelerator is being tested with a spark gap driven modulator. It is a short pulse, linear induction accelerator (LIA) with an electron diode. Its design goals are: 700 kV, 7 kA, 30 ns pulses at 50 Hz. After characterizing the accelerator with the spark gap modulator, it will be tested with a GaAs PCSS modulator, which is under construction. Forty-eight, 2 inch diameter PCSS will switch 70 kA in a 250 kV coaxial Blumlein to deliver 220 kV, 35 kA, 30 ns pulses to the LIA. One module (1/8/sup th/) of the PCSS modulator is being tested. Results from these tests and projections for the complete system are discussed. The short pulse sensors are for military and commercial applications in optical and electrical range sensing, 3D laser radar, and high speed photography. The highest optical power produced with PCSS-driven laser diode arrays is presently 50 kW in 75 ps wide pulses or 12 kW in 1 ns wide pulses. A variety of sizes of GaAs PCSS are being tested around voltage and current specifications of several applications. Voltages range from 2 to 100 kV, currents range from 10 to 500 A, and electrical pulse lengths range from 1 to 50 ns. This paper discusses developmental issues of GaAs PCSS, which are common to all: fundamental research in high gain GaAs, device longevity, optical triggering, circuit configuration, and switch performance.