HIGH-POWER microwave (HPM) sources are being developed

Transcription

1 238 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 41, NO. 1, JANUARY 2013 Frequency Variation of a Reflex-Triode Virtual Cathode Oscillator Amitava Roy, Archana Sharma, Vishnu Sharma, Ankur Patel, and D. P. Chakravarthy Abstract We present a study of the shot-to-shot variation in frequency of a reflex-triode virtual cathode oscillator with respect to peak diode voltage, current, impedance, and perveance. The typical electron-beam parameters were 185 kv, 7 ka, and 300 ns, with a current density of a few hundreds of amperes per square centimeter. The one-way analysis of variance was employed to examine the statistical correlation of the diode voltage, current, impedance, and perveance with the emitted microwave frequency. It was shown that the microwave dominant frequency variation is statistically correlated only with the peak diode current. However, the secondary emitted frequency is not statistically correlated with any of the electron-beam diode parameters. Index Terms Electron-beam diode, microwave frequency, reflex-triode virtual cathode oscillator (RT vircator), statistical analysis. I. INTRODUCTION HIGH-POWER microwave (HPM) sources are being developed for various applications like plasma heating, particle acceleration, high-power radar, and many other industrial and military fields [1], [2]. The reflex-triode virtual cathode oscillator (RT vircator) is one among several types of pulsed HPM sources in use today. The vircator is considered to be very attractive due to its high-power capability, frequency tunability, and device simplicity while facing difficult problems in microwave efficiency and frequency stability [1] [3]. Another important issue related to these types of pulsed HPM devices is the shot-to-shot variation in the emitted microwave power, frequency, and pulsewidth [4] [6]. HPM devices are highly dependent on high-current-density explosive field emission cathodes [7]. The electron beam is generated using the graphite cathode from explosive emission plasma which is formed when the strong electric field E> 10 7 V/cm is applied to the anode cathode (AK) gap. It was shown that, in a vircator with cold explosive field emission cathodes, the beam-to-microwave efficiency can be increased by improving the electron-beam quality with the use of a carbon fiber cathode [8]. It was also shown that there is a statistical correlation between emission uniformity and the shot-to-shot Manuscript received July 2, 2012; revised September 25, 2012; accepted October 15, Date of publication November 16, 2012; date of current version January 4, The authors are with the Accelerator and Pulse Power Division, Bhabha Atomic Research Centre, Mumbai , India ( aroy@barc.gov.in; arsharma@barc.gov.in; vishnu@barc.gov.in; ankur_05ec@yahoo.co.in; dpc@ barc.gov.in). Color versions of one or more of the figures in this paper are available online at Digital Object Identifier /TPS variation in diode current [9]. The variation of the microwave peak power and the pulsewidth from shot to shot might be due to the variation of the effective cathode emission area and the plasma expansion velocity [10]. The statistical variation of the microwave pulse characteristics on shot-to-shot basis may be due to the variation in the diode voltage and current. The shot-to-shot variation of the microwave pulsewidth for an axial vircator was studied for various AK gaps and two different cathode materials (velvet and graphite) [4]. Results indicated that the average microwave pulsewidth for a diode with a graphite cathode increased as the AK gap was increased. The one-way analysis of variance (ANOVA) was employed to examine the statistical correlation of the diode voltage, current, and perveance with the microwave pulsewidth for various diode gaps and for two different cathode materials. It was shown that the microwave pulsewidth variations are not statistically correlated with the diode voltage, current, and perveance [4]. In our previous experiments, the time-dependent behavior of the RT vircator oscillation frequency has been studied for various AK-gap distances [5]. The typical electron-beam parameters were 200 kv, 4 ka, and 300-ns FWHM, with a current density of a few hundreds of amperes per square centimeter. It was found that the measured HPM power and the E-field are highest for a 15-mm AK-gap distance. The RT vircator emits several frequency components, and maximum power is emitted when all the power goes into a single frequency. This dominant frequency does not significantly vary during shot-to-shot fluctuations of diode voltage and current and is less affected by these fluctuations than the other (secondary) frequency components. It was observed that the rate of increase (over time) in the dominant frequency does not substantially vary on a shot-toshot basis; however, the time of occurrence of the dominant and secondary frequency peaks does vary on a shot-to-shot basis. In addition, the time duration of the occurrence of the dominant frequency varies on a shot-to-shot basis [5]. In this paper, we address the shot-to-shot variation in frequency of oscillation for an RT vircator for fixed charging voltage and AK gap. The one-way ANOVA was employed to examine the statistical correlation of the peak diode voltage, current, impedance, and perveance with the microwave frequency. We show results from experiments using a 1-kJ Marx generator to drive an RT vircator to generate HPMs. The 1-kJ Marx generator is capable of producing a maximum output voltage of 300 kv into a matched load of 25 Ω with a pulse duration of 300-ns FWHM /$ IEEE

2 ROY et al.: FREQUENCY VARIATION OF A REFLEX-TRIODE VIRTUAL CATHODE OSCILLATOR 239 Fig. 1. Schematic of the experimental setup. II. EXPERIMENTAL ARRANGEMENT Fig. 1 shows the schematic of the 1-kJ Marx generator along with the RT vircator. The 1-kJ Marx system (300 kv, 300 ns, and 25 Ω) is a six-stage bipolar Marx generator [11]. The reflex triode is housed inside an 8-in six-way stainless steel (SS) vacuum chamber (40-cm length and 40-cm diameter) with one of the ports bolted onto the vacuum system. A vacuum explosive electron emission diode was used to generate an intense relativistic electron beam. The high-voltage pulse generated from the pulsed-power system is applied to the anode. The diode consists of a planar cylindrical graphite cathode (77-mm diameter) and an SS anode mesh (150-mm diameter) at various AK gaps and various voltage levels. The AK-gap separation can be adjusted by screwing the cathode brass stock inward or outward. The outer edge of the cathode was rounded with a 10-mm radius. Prior to use, the graphite cathode surface was roughened by sanding it with an emri paper. The anode is made of SS wire with a diameter of 0.5 mm and a 2-mm 2 mesh. A resistive CuSO 4 voltage divider and a self-integrating Rogowski coil were used to measure the diode voltage and current pulses, respectively. The voltage divider was placed parallel to the cathode plane and touched the anode holder. An SS vacuum chamber (50-cm length and 63.5-cm diameter) was also connected between the reflex triode and the Marx generator in order to house the voltage divider. An open-ended waveguide was used to radiate the output signal into the atmosphere. A vacuum level on the order of < mbar was maintained in the vircator chamber by a diffusion pump backed by a rotary pump. The typical electron-beam parameters were 185 kv, 7 ka, and 300 ns. A shielded room was situated approximately 20 m away from the RT vircator. All oscilloscope measurements were carried out inside the shielded room. For each shot, the beam parameters were recorded using a 500-MHz 2-GS/s oscilloscope. Various components used in the diagnostics were calibrated using a standard modulated (few milliseconds to nanoseconds) RF source. All experiments were carried out with fixed 15-mm AK gap and 26-kV Marx generator charging voltage. A microwave sensor consisting of a B-dot probe was used to detect the microwave signal over a wide frequency range (< 6 GHz) and send it to a high-speed digital storage oscilloscope (with a sampling rate of 40 GS/s) through a coaxial RF cable. The B-dot probe consists of a pair of magnetic sensors Fig. 2. Electron-beam diode voltage and current waveforms for 15-mm- AK-gap 77-mm-diameter graphite cathode at 26-kV Marx generator charging voltage. Fig. 3. (a) B-dot probe signal for (b) FFT of the B-dot probe signal captured for that measure the rate of the field change. The B-dot probe was situated at 4.5 m from the source window. The B-dot probe oscilloscope measurements were also carried out inside the shielded room. For the free-space radiation pattern measurement, the B-dot probe was positioned at a distance of r =4.5mfrom the radiation window. In the measurements, the B-dot probe was moved from the bore sight to positions 10,20,30,40,50,60, and 90 with respect to the axial direction. The microwave power generated from the reflex triode is approximately described by [12] ( ) E 2 P = (4πr 2 ) (1) ZG T where G T = π 2 π 2 P m Pθ 2 sin θdθ

3 240 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 41, NO. 1, JANUARY 2013 Fig. 4. Histogram showing the shot-to-shot variation in peak diode voltage for Fig. 6. Histogram showing the shot-to-shot variation in diode impedance for Fig. 5. Histogram showing the shot-to-shot variation in peak diode current for where E is the free-space electric field measured by the B-dot probe, Z is the free-space wave impedance, G T is the gain of the reflex-triode open-ended waveguide, P m is the microwave power in the main radiation lobe, and P θ is the microwave power at direction angle θ with respect to the axial direction or main lobe. Fig. 7. Histogram showing the shot-to-shot variation in RT vircator dominant frequency for III. EXPERIMENTAL RESULTS The diode voltage and current waveforms for a 15-mm AK gap are shown in Fig. 2. The peak diode voltage and current are 180 kv and 7 ka, respectively. The electron beam will establish a virtual cathode if the injected current exceeds the space charge limiting current beyond the anode mesh. A typical HPM waveform recorded by the B-dot probe is shown in Fig. 3(a). One can see from Fig. 3(a) that the HPM pulse duration is only 150 ns. Fig. 3(b) shows the Fourier transform of the B-dot waveform. From Fig. 3(b), the microwave frequency components can be Fig. 8. Histogram showing the shot-to-shot variation in RT vircator second dominant frequency for

4 ROY et al.: FREQUENCY VARIATION OF A REFLEX-TRIODE VIRTUAL CATHODE OSCILLATOR 241 TABLE I ONE-WAY ANOVA OF DOMINANT FREQUENCY AND THE SECONDARY FREQUENCY WITH PEAK DIODE VOLTAGE, CURRENT, IMPEDANCE, AND PERVEANCE identified. One can see that the spectra contain a dominant (largest magnitude) peak frequency and a secondary (next largest magnitude) peak frequency. It was observed that the pulsewidth, the frequency, and the peak of the microwave signal varied on a shot-to-shot basis. The peak diode voltage and current also vary on a shot-toshot basis. Figs. 4 6 show the shot-to-shot variation in the peak diode voltage, current, and impedance at the 15-mm AK-gap distance and a constant 26-kV charging voltage for 50 shots. One can see from Fig. 5 that the mean diode impedance is 25 Ω, which is the matched impedance of the pulsed-power system. The frequencies of the two main peaks also vary significantly on a shot-to-shot basis. Figs. 7 and 8 show the shot-to-shot variation in the dominant frequency and the second dominant frequency, respectively, for 42 shots. The dominant peak has a mean value of 2.92 GHz with a standard deviation of 0.8 GHz, whereas the secondary peak has a mean value of 3.32 GHz with a standard deviation of 0.99 GHz. Therefore, the frequency of the dominant peak is less affected due to shot-to-shot fluctuations of diode voltage and current than the frequency of the secondary peak [5]. The one-way ANOVA [13] was employed to examine the statistical correlation of the peak diode voltage, current, impedance, and perveance with the microwave frequency for a fixed AK gap. The one-way ANOVA can be employed to test whether two or more populations have the same mean. One-way ANOVA assumes that the sample data sets have been drawn from populations that follow a normal distribution with a constant variance. For the ANOVA calculations, the confidence level P<0.05 has been used. Table I shows the one-way ANOVA of the peak diode voltage, current, impedance, and perveance, for the dominant and the secondary RT vircator frequency components. One can see from Table I that only the peak diode current variance is not significantly different from the dominant frequency variance. However, all other electron-beam diode parameter variances under test are significantly different from the dominant and the secondary peak frequency variances. Therefore, Fig. 9. Measured power density at various angles with respect to the axial direction or main lobe. one can conclude that the RT vircator dominant frequency variation is statistically correlated only with the peak diode current. In order to get an idea about the peak power generated by the RT vircator at 15-mm AK gap and 26-kV charging voltage, the emitted radiation pattern has been measured by placing the B-dot probe at various angles with respect to the axial direction or main lobe. Fig. 9 shows the measured power density at various angles with respect to the axial direction or main lobe. At each location, five shots were taken, and the average value has been plotted in Fig. 9. The main radiation lobe in the radiation region appeared at the axial direction because of TE 11 mode operation. The HPM electric fields measured using the B-dot probe at 4.5-m distance from the reflex-triode window are about 9 and 14 kv/m at 26- and 35-kV Marx charging voltages, respectively, at a The experimentally established peak HPM power generated at the RT vircator was found to be 48 MW at 35-kV Marx charging voltage and

5 242 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 41, NO. 1, JANUARY 2013 IV. CONCLUSION The shot-to-shot variation of the RT vircator oscillation frequency has been studied with respect to electron-beam diode voltage, current, impedance, and perveance. The typical electron-beam parameters were 185 kv, 7 ka, and 300-ns FWHM, with a current density of a few hundreds of amperes per square centimeter. It was found that all the diode parameters vary significantly on shot-to-shot basis as do the RT vircator power and frequency. The one-way ANOVA was employed to examine the statistical correlation of the peak diode voltage, current, impedance, and perveance with the emitted microwave frequency. It was found that only the peak diode current is statistically correlated with the RT vircator dominant frequency variation. However, the secondary peak frequency is not statistically correlated with the electron-beam diode parameters. The measured radiation pattern indicates the TE 11 mode operation of the RT vircator. The present RT vircator emitted a peak HPM power of 48 MW at 35-kV Marx generator charging voltage with 2.4-GHz dominant frequency. ACKNOWLEDGMENT The authors would like to thank Dr. L. M. Gantayet, the Director of the BTD Group, Bhabha Atomic Research Centre (BARC), and Dr. A. K. Ray, a Raja Ramanna Fellow, of BARC for providing useful guidance, facilities, and efficient manpower and S. R. Raul, S. Patil, S. Vaity, and N. K. Lawangare for the technical help. REFERENCES [1] J. Benford, J. Swegle, and E. Schamiloglu, High Power Microwaves, 2nd ed. New York: Taylor & Francis, [2] D. J. Sullivan, J. E. Walsh, and E. A. Coutsias, Virtual cathode oscillator (vircator) theory, in High Power Microwave Sources, V. Granastein and I. Alexeff, Eds. Norwood, MA: Artech House, 1987, p [3] L. E. Thode, Virtual cathode microwave device research: Experiment and simulation, in High Power Microwave Sources, V. Granastein and I. Alexeff, Eds. Norwood, MA: Artech House, 1987, p [4] A. Roy, S. K. Singh, R. Menon, D. S. Kumar, S. Khandekar, V. B. Somu, S. Chottray, P. C. Saroj, K. V. Nagesh, K. C. Mittal, and D. P. Chakravarthy, Pulsewidth variation of an axial vircator, IEEE Trans. Plasma Sci., vol. 38, no. 7, pp , Jul [5] A. Roy, A. Sharma, S. Mitra, R. Menon, V. Sharma, K. V. Nagesh, and D. P. Chakravarthy, Oscillation frequency of a reflex-triode virtual cathode oscillator, IEEE Trans. Electron Devices, vol.58,no.2, pp , Feb [6] M. Elfsberg, T. Hurtig, A. Larsson, C. Möller, and S. E. Nyholm, Experimental studies of anode and cathode materials in a repetitive driven axial vircator, IEEE Trans. Plasma Sci., vol. 36,no.3,pp ,Jun [7]G.A.MesyatsandD.I.Proskurovsky,Pulsed Electrical Discharge in Vacuum. Berlin, Germany: Springer-Verlag, [8] L. Liu, L.-M. Li, X.-P. Zhang, J.-C. Wen, H. Wan, and Y.-Z. Zhang, Efficiency enhancement of reflex triode virtual cathode oscillator using the carbon fiber cathode, IEEE Trans. Plasma Sci., vol. 35, no. 2, pp , Apr [9] D. A. Shiffler, J. Luginsland, M. Ruebush, M. LaCour, K. Golby, K. Cartwright, M. Haworth, and T. Spencer, Emission uniformity and shot-to-shot variation in cold field emission cathodes, IEEE Trans. Plasma Sci., vol. 32, no. 3, pp , Jun [10] A. Roy, R. Menon, S. K. Singh, M. R. Kulkarni, P. C. Saroj, K. V. Nagesh, K. C. Mittal, and D. P. Chakravarthy, Shot to shot variation in perveance of the explosive emission electron beam diode, Phys. Plasmas, vol. 16, no. 3, pp , Mar [11] A. Sharma, S. Kumar, S. Mitra, V. Sharma, A. Patel, A. Roy, R. Menon, K. V. Nagesh, and D. P. Chakravarthy, Development and characterization of repetitive 1-kJ Marx-generator-driven reflex triode system for highpower microwave generation, IEEE Trans. Plasma Sci., vol. 39, no. 5, pp , May [12]L.Li,L.Liu,G.Cheng,Q.Xu,H.Wan,L.Chang,andJ.Wen, The dependence of vircator oscillation mode on cathode material, J. Appl. Phys., vol. 105, no. 12, pp , Jun [13] J. Neter, M. Kutner, C. Nachtsheim, and W. Wasserman, Applied Linear Statistical Models. Boston, MA: McGraw-Hill, Amitava Roy received the B.Sc. degree (with honors) in physics from the University of North Bengal, Darjeeling, India, in 1995, the M.Sc. degree in physics from Jawaharlal Nehru University, New Delhi, India, in 1997, and the Ph.D. degree (focusing on intense relativistic electron-beam generation and applications) from the Homi Bhabha National Institute, Mumbai, India, in Since 1998, he has been with the Accelerator and Pulse Power Division, BARC, where he is currently a Scientific Officer (F). His research interests include particle accelerators, pulsed-power technology, intense relativistic electron beams, generation of high-power microwaves, and Flash X-ray. He has more than 50 scientific publications/presentations in international/national journals/conferences. Dr. Roy was a recipient of a university medal for scoring second position in the B.Sc. examination. Archana Sharma, photograph and biography not available at the time of Vishnu Sharma, photograph and biography not available at the time of Ankur Patel, photograph and biography not available at the time of D. P. Chakravarthy received the B.E. degree in electronics and communication from Osmania University, Hyderabad, India, in He completed the oneyear orientation course on nuclear sciences from Bhabha Atomic Research Centre (BARC), Mumbai, India, in In 1974, he joined the Plasma Physics Division, BARC, where he is currently the Head of the Accelerator and Pulse Power Division. He has worked on plasma diagnostics, instrumentation and controls for magnetohydrodynamic generators, and high-power arc and plasma torches. His other main activities have been on the development of high-power radio-frequency (RF) power supplies, capacitively coupled discharges for CO 2 lasers, and inductively coupled plasma discharges. His current area of work includes development in the field of pulsed-power electron accelerators for high-power microwave and Flash X-ray generation, RF and direct-current industrial electron accelerators, and electron cyclotron resonance ion sources.