Quantitative determination and experimental validation of cathode control characteristics for the Gunn effect under imperfect contacts

Kroemer's 1968 theory of the Gunn effect under imperfect cathode boundary conditions explains the diversity of observed transferred-electron device behaviors through the cathode control characteristic $j_c(F)$, the current-field characteristic of the contact interface itself, but it left the actual characteristics of real contacts undetermined: they were neither measured for identified contact processes nor computed from contact microphysics, and the framework's quantitative predictions (nonlinear oscillation amplitudes and spectra, full frequency-dependent admittance, refined stability thresholds) were developed only in small-signal or idealized-model approximations. Later literature substantially advanced the theory side: stability criteria were generalized to injection-limiting cathodes (Zetsche 1974), cathode-boundary-field models were validated experimentally with assigned cathode fields (Solomon, Shaw, Grubin and Kaul 1975), cathode notches and heterojunction cathodes became standard engineering practice, and the Bonilla school (1995-1997) derived contact boundary conditions from thermionic-emission contact physics and completed an asymptotic nonlinear theory of Gunn oscillations under them. The open core of this problem is the quantitative determination of $j_c(F)$ for real, independently characterized contact processes - experimentally and from contact microphysics - together with the end-to-end demonstration that the same characteristic predicts measured device current-voltage characteristics, stability thresholds, small-signal admittance, and nonlinear oscillation waveforms.

Background

In semiconductors such as n-type GaAs and n-type InP, the electron drift velocity $v(F)$ first rises with electric field $F$, reaches a peak, and then falls as carriers are scattered into satellite valleys of the conduction band with lower mobility. This negative differential mobility (NDM) makes a uniformly doped, voltage-biased bar electrically unstable: space-charge fluctuations grow into high-field domains that travel from cathode to anode, producing periodic current oscillations at microwave frequencies. Engineered in this way, the device is a transferred-electron device (Gunn diode), long used as a microwave and millimeter-wave source.

A one-dimensional theory (drift-diffusion current plus Poisson's equation) needs a boundary condition at the injecting cathode contact, and the early treatments assumed a well-behaved cathode: the field climbs into the NDM region entirely within the bulk, and the negative space charge that nucleates a domain consists of accumulated free electrons. Kroemer (1968) pointed out that real contacts often violate both assumptions - thin resistive interface layers, shallow Schottky barriers (metal-semiconductor contacts with a potential barrier), heterojunctions (junctions between dissimilar semiconductors), embedded charged sheets, or alloyed regrown contact regions all perturb the near-cathode field and current. He replaced the assumption with a general boundary-condition framework expressed through the control characteristic $j_c(F)$: the relation between the current density crossing the contact plane and the electric field $F$ at that plane. Because an imperfect interface impedes current, $j_c(F)$ lies at least partly below the bulk neutral characteristic $j_n(F) = qNv(F)$, where $q$ is the electron charge and $N$ the donor density; the primary space charge nucleated at the cathode is then a depletion layer (a region stripped of electrons, its charge made up by ionized donors) rather than an accumulation layer, and the crossover field $F_x$ at which $j_c(F_x) = j_n(F_x)$ separates operating regimes. Depending on the shape of $j_c(F)$ and the position of the crossover, the framework explains Gunn's static preoscillation cathode fall (a stationary high-field layer at the cathode preceding oscillations), Kennedy's prethreshold small-signal amplification (amplification of small microwave signals below the oscillation threshold), delayed or suppressed oscillation onset with excessive threshold fields, hysteresis and current saturation, and oscillations of lower amplitude and higher frequency than in the usual transit-time mode (oscillation at the period set by a domain's transit across the device length).

Kroemer supported the framework with depletion-layer electrostatics giving the device current-voltage characteristic parametrically, a linearized small-signal impedance analysis under a constant-parameter approximation, numerical drift-diffusion simulations for two idealized contact models (a shallow cathode barrier fixing the electron density at the contact at $0.95N$, and an embedded p-type layer modeled as a fixed negatively charged sheet), and domain sustaining-voltage estimates from Butcher's equal-areas rule (the condition fixing the internal field of a stationary domain at a given current). The paper itself acknowledged what it could not do: the actual $j_c(F)$ of real contacts was not known from available data; the small-signal, constant-parameter treatment did not yield the nonlinear oscillation evolution, saturation amplitudes, spectral properties, or the full frequency dependence of the device admittance (the small-signal alternating current response per applied alternating voltage); some stability thresholds contained uncertain numerical factors, such as an effective $\alpha$ entering the instability criterion and the appropriate value of $d\ln v/dF$ averaged over strongly spatially varying depletion layers; and the shape of $j_c(F)$ away from the crossover had not been computed for many physically plausible contact models, including the effects of interface-state kinetics (charge trapped at interface defects with finite capture and emission rates), barrier lowering, and diffusion within thin contact sections. Experiments correlating measured device current-voltage and high-frequency behavior, and direct probes of the near-cathode field and charge distributions, with independently characterized contact fabrication and process parameters, were called for to test the framework.

Problem Statement

Determine, quantitatively, by direct near-cathode measurement and/or quantitative contact-physics calculation, the cathode control characteristics of real imperfect contacts in Gunn-effect (transferred-electron) devices, and validate Kroemer's control-characteristic framework quantitatively against direct experiment. The single scientific objective is to turn the 1968 boundary-condition framework from a phenomenological classification into a quantitatively calibrated and experimentally validated predictive theory, in which the contact's own characteristic - once fixed by independent measurement or contact-physics calculation - determines everything the device does. Concretely, for the real contact processes that occur in fabricated devices (thin resistive interface layers, shallow Schottky barriers, heterojunctions (junctions between dissimilar semiconductors), embedded charged sheets, alloyed regrown contact regions, and interface-state-dominated contacts):

1. Establish the control characteristic $j_c(F)$ - the current density through the contact plane as a function of the electric field at that plane, or an equivalent boundary condition with physically defined contact parameters - by direct near-cathode measurement and/or by quantitative contact-physics calculation (thermionic emission - thermally activated emission of carriers over the contact barrier, including its field-assisted lowering - tunneling, interface-state kinetics, and diffusion within thin control sections), over the full field range of device operation and not only near the crossover $j_c = j_n$ with the bulk neutral characteristic, with the contact's fabrication and process parameters independently characterized - 'independently' meaning that neither the process parameters nor $j_c(F)$ may be fitted to the device responses that item 2 requires $j_c(F)$ to predict.

2. Demonstrate that the same, independently fixed characteristic quantitatively predicts the measured device behavior: the dc current-voltage characteristic including saturation, hysteresis and any quasi-static negative conductance; the stability thresholds and oscillation-onset conditions, including the numerical factors entering the instability criterion for realistic velocity-field curves $v(F)$ and diffusion; the frequency-dependent small-signal admittance; and the nonlinear oscillation waveforms, saturation amplitudes and spectral content, including the interaction with the external circuit.

References

1. H. Kroemer, The Gunn effect under imperfect cathode boundary conditions, IEEE Transactions on Electron Devices 15(11), 819-837 (1968). DOI: 10.1109/T-ED.1968.16523, https://doi.org/10.1109/T-ED.1968.16523
2. W. J. Williamson, Effect of bias and cathode boundary conditions on GaAs space-charge amplifiers, Applied Physics Letters 18(9), 389-390 (1971). DOI: 10.1063/1.1653711, https://doi.org/10.1063/1.1653711
3. H. Zetsche, Stability criterion for Gunn diodes with injection-limiting cathodes, IEEE Transactions on Electron Devices 21(2), 142-146 (1974). DOI: 10.1109/T-ED.1974.17883, https://doi.org/10.1109/T-ED.1974.17883
4. P. R. Solomon, M. P. Shaw, H. L. Grubin, R. Kaul, An experimental study of the influence of boundary conditions on the Gunn effect, IEEE Transactions on Electron Devices 22(3), 127-139 (1975). DOI: 10.1109/T-ED.1975.18091, https://doi.org/10.1109/T-ED.1975.18091
5. H. L. Grubin, Transferred electron devices with emphasis on the role of the contacts - a review, CRC Critical Reviews in Solid State Sciences 5(3), 409-417 (1975). DOI: 10.1080/10408437508243503, https://doi.org/10.1080/10408437508243503
6. K. Ohta, M. Kawashima, S. Kataoka, Electro-optic modulation of a laser beam by a traveling high-field domain in a GaAs Gunn diode, Journal of Applied Physics 46(3), 1318-1321 (1975). DOI: 10.1063/1.321699, https://doi.org/10.1063/1.321699
7. H. L. Grubin, Dynamic cathode boundary fields and transferred electron oscillators, IEEE Transactions on Electron Devices 23(9), 1012-1015 (1976). DOI: 10.1109/T-ED.1976.18528, https://doi.org/10.1109/T-ED.1976.18528
8. R. M. Raymond, H. Kroemer, R. E. Hayes, Design of cathode doping notches to achieve uniform fields in transferred-electron devices, IEEE Transactions on Electron Devices 24(3), 192-195 (1977). DOI: 10.1109/T-ED.1977.18708, https://doi.org/10.1109/T-ED.1977.18708
9. M. P. Shaw, H. L. Grubin, P. R. Solomon, The Gunn-Hilsum Effect, Academic Press / Elsevier (1979), ISBN 9780126383508; chapter The NDM Element's Environment: Circuits and Boundaries, pp. 22-33, DOI: 10.1016/B978-0-12-638350-8.50007-6, https://doi.org/10.1016/B978-0-12-638350-8.50007-6
10. H. L. Grubin, J. P. Kreskovsky, The Role of Boundary Conditions in Near- and Submicrometer-Length Gallium Arsenide Structures, in VLSI Electronics Microstructure Science, Vol. 10: Surface and Interface Effects in VLSI (N. G. Einspruch, R. S. Bauer, eds.), Elsevier (1985), pp. 237-321. DOI: 10.1016/B978-0-12-234110-6.50011-9, https://doi.org/10.1016/B978-0-12-234110-6.50011-9
11. M. R. Friscourt, P. A. Rolland, M. Pernisek, Heterojunction cathode contact transferred-electron oscillators, IEEE Electron Device Letters 6(10), 497-499 (1985). DOI: 10.1109/EDL.1985.26207, https://doi.org/10.1109/EDL.1985.26207
12. J. Czekaj, M. P. Shaw, The influence of cathode contacts on near-micron InP transferred electron devices, Solid-State Electronics 31(3-4), 599-602 (1988). DOI: 10.1016/0038-1101(88)90350-4, https://doi.org/10.1016/0038-1101(88)90350-4
13. K. F. Wu, J. Czekaj, M. P. Shaw, Study of submicron InP transferred electron devices, Journal of Applied Physics 74(1), 315-326 (1993). DOI: 10.1063/1.354110, https://doi.org/10.1063/1.354110
14. L. L. Bonilla, F. J. Higuera, The Onset and End of the Gunn Effect in Extrinsic Semiconductors, SIAM Journal on Applied Mathematics 55(6), 1625-1649 (1995). DOI: 10.1137/S0036139991199456, https://doi.org/10.1137/S0036139991199456
15. G. Gomila, J. M. Rubi, I. R. Cantalapiedra, L. L. Bonilla, Stationary states and phase diagram for a model of the Gunn effect under realistic boundary conditions, Physical Review E 56(2), 1490-1499 (1997). DOI: 10.1103/PhysRevE.56.1490, https://doi.org/10.1103/PhysRevE.56.1490 (preprint: arXiv:cond-mat/9705262, https://arxiv.org/abs/cond-mat/9705262)
16. L. L. Bonilla, I. R. Cantalapiedra, G. Gomila, J. M. Rubi, Asymptotic analysis of the Gunn effect with realistic boundary conditions, Physical Review E 56(2), 1500-1510 (1997). DOI: 10.1103/PhysRevE.56.1500, https://doi.org/10.1103/PhysRevE.56.1500 (preprint: arXiv:cond-mat/9705264, https://arxiv.org/abs/cond-mat/9705264)
17. L. L. Bonilla, I. R. Cantalapiedra, Universality of the Gunn effect: Self-sustained oscillations mediated by solitary waves, Physical Review E 56(3), 3628-3632 (1997). DOI: 10.1103/PhysRevE.56.3628, https://doi.org/10.1103/PhysRevE.56.3628
18. L. L. Bonilla, Theory of nonlinear charge transport, wave propagation, and self-oscillations in semiconductor superlattices, Journal of Physics: Condensed Matter 14, R341-R381 (2002). DOI: 10.1088/0953-8984/14/14/201, https://doi.org/10.1088/0953-8984/14/14/201
19. L. L. Bonilla, S. W. Teitsworth, Nonlinear Wave Methods for Charge Transport, Wiley-VCH (2010), ISBN 9783527406951. DOI: 10.1002/9783527628674, https://doi.org/10.1002/9783527628674
20. Y. Wang, L.-A. Yang, Z. Wang, Q. Chen, Y. Huang, Y. Dai, H. Chen, H. Zhao, Y. Hao, Ultra-short channel GaN high electron mobility transistor-like Gunn diode with composite contact, Journal of Applied Physics 116(9), 094502 (2014). DOI: 10.1063/1.4894627, https://doi.org/10.1063/1.4894627
21. B. Li, Y. Alimi, G. L. Ma, Investigation on multi-frequency oscillations in InGaAs planar Gunn diode with multiple anode-cathode spacings, Solid State Communications 247, 1-5 (2016). DOI: 10.1016/j.ssc.2016.08.006, https://doi.org/10.1016/j.ssc.2016.08.006
22. A. Rozhkov, M. Ivanov, P. Rodin, The lock-on effect and collapsing bipolar Gunn domains in high-voltage GaAs avalanche p-n junction diode, Solid State Communications 354, 114895 (2022). DOI: 10.1016/j.ssc.2022.114895, https://doi.org/10.1016/j.ssc.2022.114895 (preprint: arXiv:2206.03940, https://arxiv.org/abs/2206.03940)