Room-temperature terahertz emission from N-polar GaN grating-gate high-electron-mobility transistors

Abstract

Compact, electrically driven terahertz (THz) sources remain a major bottleneck for chip-scale spectroscopy, imaging and high-speed wireless links. AlGaN/GaN two-dimensional electron gas (2DEG) systems have been investigated for THz emission, but experimental demonstrations have almost exclusively used Ga-polar platforms. Here we report a room-temperature THz signal from N-polar GaN grating-gate high-electron-mobility transistors (HEMTs) under pulsed electrical excitation. In the reversed-polarization heterostructure, the GaN channel lies above the AlGaN barrier, a geometry that places the channel close to the surface grating. The detected THz power increases with lateral electric field and square-wave duty cycle, whereas the apparent conversion efficiency falls at high duty cycle, consistent with increased Joule heating. Spectral measurements reveal a broad low-frequency feature and a gated/ungated spectral difference. The long-pulse room-temperature response is consistent with predominantly thermal emission, and the gated/ungated difference may reflect geometry-dependent thermal emissivity. These measurements do not establish a separate collective 2DEG contribution or a plasmonic resonance. These results establish an experimental N-polar GaN HEMT platform for studying electrically driven THz radiation and motivate calibrated emissivity, polarization-resolved and gate-tunable measurements.

Graphical abstract

Keywords

N-polar GaN / terahertz emitter / AlGaN/GaN high-electron-mobility transistor / grating-gate coupling / two-dimensional electron gas

1 Introduction

Terahertz radiation bridges electronics and photonics, enabling probes of low-energy excitations, imaging of non-metallic objects and ultrafast wireless communication. Practical systems, however, are still constrained by the lack of compact, efficient and electrically addressable terahertz (THz) sources [1–4]. Field-effect semiconductor emitters are attractive because they can be integrated with electronic circuits and driven using simple electrical biases.

GaN-based heterostructures are well suited to this task. Polarization-induced AlGaN/GaN two-dimensional electron gases (2DEGs) combine high sheet density, high breakdown field, high saturation velocity and thermal robustness [5,6]. The plasma-wave framework predicted current-driven oscillations in field-effect transistor channels [7], while related studies analyzed THz detection and mixing and demonstrated a resonant detector response [8,9]. THz emission has been reported in high-electron-mobility transistors (HEMTs) without grating-gate structures [10,11]. Grating-gate coupling between THz radiation and two-dimensional plasmons has been investigated separately in theoretical and experimental studies [12,13]. Electrical THz emission from Ga-polar AlGaN/GaN structures has been examined using periodic contacts or grating coupling [14–16]. Measurements of biased HEMTs have also identified thermal and trap-related contributions [17]. Related work has examined two-dimensional plasmon spectra, electrical pumping, modulation and device concepts [18–22]. Under electrical pumping, these background channels can coexist with collective effects. Separating these channels requires mechanism-specific controls rather than spectral structure alone.

Lisauskas et al. [17] specifically concluded that radiation from biased AlGaN/GaN HEMTs could not be directly attributed to plasmonic instability and was better described as a combination of thermal emission from heated material, thermally excited plasmons and trap states.

Most experimental GaN plasmonic THz emitters have used Ga-polar AlGaN/GaN structures. In the N-polar geometry studied here, the GaN channel lies above the AlGaN back-barrier and closer to the surface grating. This arrangement motivates examining how the reversed-polarity platform affects electrically driven THz radiation. The present study does not independently quantify improved confinement or stronger grating coupling relative to a Ga-polar control.

Our previous numerical study examined N-polar AlGaN/GaN HEMTs with asymmetric boundaries [23]. Prior N-polar work established layer-transferred N-face devices [24], analyzed polarity-dependent short-channel effects [25], and reported high channel conductivity and low-resistance contacts [26,27]. The present work moves from that design study to an experimental platform. We fabricate N-polar GaN grating-gate HEMTs and measure their room-temperature THz power and spectra. Because the measurements use one grating period, a fixed gate bias and millisecond electrical pulses, they do not establish the gate-tuning or instability criteria needed for a unique collective-mode assignment.

The main advance is the experimental observation of an electrically driven THz signal from an N-polar GaN HEMT. The measurements establish the platform and its field-, duty-cycle- and geometry-dependent response, while leaving the microscopic emission mechanism open.

2 Device fabrication and direct-current (DC) characterization

Figure 1 shows the layer-transferred N-polar GaN/AlGaN structure used in this study. The chemical-mechanical-polishing (CMP)-incorporated layer-transfer process exposes the N-polar surface and places the GaN channel above the AlGaN back-barrier, close to the surface grating. The source-drain field drives current through the channel, while the patterned surface can modify the frequency-dependent radiative response.

The CMP step was used as part of the layer-transfer process to thin the transferred stack and expose the N-polar surface; the target layer thicknesses were verified by the cross-sectional focused-ion-beam scanning electron microscopy (FIB-SEM) image in Fig. 1b. Residual CMP-related surface damage was not independently quantified in this study. The layer-transfer route avoids the need for a dedicated N-polar growth substrate, whereas a direct-growth comparison was not performed here.

The N-polar GaN HEMT material was prepared using a CMP-incorporated layer-transfer process, following the platform reported by Guo et al. [28]. The transferred stack consisted of a 10 nm Si₃N₄ surface-passivation layer, a 100 nm GaN channel, a 23 nm AlGaN back-barrier, a 200 nm low-pressure chemical vapor deposition (LPCVD) silicon-nitride diffusion-barrier layer, Ti/Cu/Au bonding metals with thicknesses of 50/800/20 nm, a second 200 nm silicon-nitride layer and a Si(100) support substrate. Van der Pauw measurements gave a Hall sheet density of 2.0 × 10¹³ cm⁻², an electron mobility of 1230 cm² V⁻¹ s⁻¹ and a sheet resistance of 190 Ω sq⁻¹. The specific contact resistivity was 2.2 × 10⁻⁵ Ω cm².

The 600 μm × 600 μm grating-coupler area was patterned by electron-beam lithography, followed by deposition and lift-off of Ni/Au (10/50 nm). The grating-line width was 150 nm and the period was 500 nm. Source and drain Ohmic electrodes were formed around the grating using Ti/Au (50/150 nm), with a source-drain separation of 600 μm. The electrode layout aligned the lateral current with the grating wave vector. The processed sample was diced into 5 mm × 5 mm dies and wire-bonded to a CSB02491 package for electrical and THz measurements.

The geometric fill factor of the grating is 150/500 = 0.30; a 600 μm grating length therefore corresponds to approximately 1200 periods. The active grating-coupler footprint is 600 μm × 600 μm, and the source-drain separation is 600 μm. The device geometry and grating dimensions are shown in Fig. 2.

The completed devices were characterized using a 1505 A high-voltage, high-current probe system. The transfer curve in Fig. 3a, measured at V_DS = 1 V, gives a threshold voltage of approximately −33 V and an on-state current of about 2.25 mA at V_GS = 5 V. The large negative threshold voltage is primarily associated with the approximately 100 nm gate-to-channel separation. Surface or buffer traps, contact resistance and distributed gate electrostatics may also contribute to the non-ideal transfer shape, but the present data do not isolate these effects. Figure 3b shows output curves for V_GS from −45 to −5 V and V_DS up to 20 V. Only one completed wire-bonded device with functional gate control was available for the reported THz measurements (n = 1), so device-to-device statistics are not available.

Room-temperature THz power measurements were carried out under square-wave source-drain excitation. Unless otherwise stated, the gate bias was held at 0 V. The drain modulation frequency was set to 5 Hz to match the high-responsivity range of the Golay detector, and the detector output was read using lock-in detection referenced to the electrical modulation. The lateral electric field was calculated from the applied source-drain voltage using the measured 600 μm electrode separation; fields of 417, 500 and 583 V cm⁻¹ therefore correspond to voltage amplitudes of 25, 30 and 35 V, respectively. The duty cycle was varied to tune the electrical input energy and thermal accumulation. Emission spectra were measured with a far-infrared Fourier-transform spectroscopy (FTIR) system under the same electrical-pumping scheme. Additional details on the layer-transfer process, THz power-measurement system and far-infrared FTIR spectral-measurement setup are provided in the Supporting Information.

For a rectangular source-drain pulse, the average electrical input power used in Fig. 4b is

P_{\text{in}}=\frac{1}{T}{\int }_0^T{V}_{\text{DS}}\left(t\right)I_{\text{DS}}\left(t\right)\text{d}t={D⟨V_{\text{DS}}{I}_{\text{DS}}⟩}_{\text{on}},

where the voltage and current terms are instantaneous drain-source values, T is the modulation period, D is the duty cycle, and angle brackets denote the on-state average. The slope of THz power versus P_in is therefore an apparent conversion efficiency, not a complete wall-plug efficiency, because it does not include all electrical, thermal, optical and collection losses. The Golay detector was a GC-1P detector with a high-density polyethylene (HDPE) window. At the 5 Hz modulation used here, its responsivity was 140 kV W⁻¹; the lock-in output was converted to power using this responsivity. The FTIR system was checked with a 660 GHz source, which produced a spectral feature near 0.67 THz. The moving mirror was scanned from 0 to 400 μm relative to the zero-path-difference position in 1 μm steps. This single-sided scan corresponds to an optical-path-difference range of 800 μm and a nominal scan-based spectral resolution of approximately 0.375 THz. The fast Fourier transform (FFT) frequency-bin spacing depends on the retained sample count, as specified below. No nitrogen purge was used during the spectral measurements reported here. DU and MD in the figure annotations denote the signal duty cycle and modulation frequency, respectively. The spectra are reported as processed FFT amplitudes, not calibrated emissivity spectra. The HDPE window and atmospheric water-vapor absorption may affect the 1–3 THz region; neither effect was corrected. For each spectrum, one recorded interferogram (one monotonic single-step or continuous scan) was processed; no arithmetic averaging of multiple scans was performed. No separate background or reference spectrum was subtracted. Each interferogram was detrended with a fifth-order polynomial and filtered using a Butterworth band-pass filter. The nominal lower and upper optical-frequency cutoffs were 0.16 and 20 THz, respectively. The MATLAB order parameter was 4, giving an eighth-order band-pass transfer function that was applied once in the forward direction using filter. With the frequency normalization used here, the implemented cutoffs equal the nominal values multiplied by N₀/(N₀ − 2), where N₀ denotes the sample count before truncation. After filtering, the last N_FFT samples were retained, with N_FFT equal to the largest power of two not exceeding N₀. The mean of the retained samples was subtracted before the FFT. The frequency-bin spacing is c/(2ΔxN_FFT), where Δx is the mirror-displacement step and c is the speed of light. The FFT magnitudes were divided by N_FFT and doubled to obtain one-sided amplitudes. The individual spectra were not normalized to their respective maxima. Both detrending and filtering can change the low-frequency spectral shape and amplitude. Neither operation removes background radiation or corrects the instrument response. Before the room-temperature measurements, the THz power-measurement chain was calibrated with a standard THz source; the detector responsivity and 660 GHz system check described above were then used for the power conversion. A separate raw-data repeatability check on records from the same packaged device at V_GS = 2 V gave relative standard deviations of 12.8%, 9.6% and 15.0% for the 2.0–2.5 THz-band peak FFT amplitude at V_DS = 2, 3 and 4 V, respectively (n = 2, 3 and 2). These separate records were not used to define the plotted spectra; they describe scan-to-scan repeatability of the processed FFT signal rather than a complete absolute-power uncertainty budget.

3 Results

Figure 4a shows the dependence of THz power on lateral field at different duty cycles. At low field, the signals remain close to the measurement background. Above approximately 80 V cm⁻¹, the curves separate and rise more rapidly. At 333 V cm⁻¹, the emitted powers are approximately 16, 48, 74, 94 and 105 nW for duty cycles of 5%, 15%, 25%, 35% and 45%, respectively. The systematic increase at a fixed field is consistent with the greater electrical energy delivered during longer voltage pulses.

Figure 4b plots the THz output power against the average electrical input power. The approximately linear trend within each series indicates that the detected THz output increases with the electrical energy supplied to the channel. The 15% and 25% series have steeper slopes than the 35% and 45% series. Under the present drive range, the 45% duty cycle gives the highest absolute power because it supplies more average energy, but it does not give the highest apparent conversion efficiency. The reduced slope at longer duty cycles is consistent with stronger Joule heating and increased non-radiative energy loss.

In Fig. 5, the four spectra first rise from the low-frequency baseline and form a broad maximum near 2–2.5 THz. This low-frequency feature becomes progressively stronger as the duty cycle increases from 5% to 45%, while its central frequency remains nearly unchanged. Preprocessing and the instrument response can shape this feature, so its position cannot be assigned to an intrinsic mode frequency.

At 45% duty cycle, the spectrum is elevated over much of the measured band and contains broad maxima near 5.5–6 and 9.5–10 THz. The lower-duty-cycle spectra show weaker modulation above 3 THz. Because the spectra are processed FFT amplitudes without full instrument-response correction, these broad features cannot be assigned to separate resonances.

Figure 6 compares the gated and ungated devices at two duty cycles. At 25% duty cycle, the grating-gate device shows a stronger broad maximum near 2.2 THz and a higher response over much of the measured band, whereas the ungated maximum is near 1.4 THz. At 45%, approximate maxima near 2.4 and 2.8 THz differ by only 0.4 THz. Because the features are broad, their positions are affected by preprocessing and the instrument response, and the devices were not current- or input-power matched, these values are descriptive and are not treated as a resolved mode shift. The frequency-dependent difference can arise from geometry-dependent emissivity; a separate collective contribution cannot be resolved.

Figure 7 compares source-drain fields at fixed modulation conditions. Increasing the field strengthens the low-frequency feature and broadband response, while the spectral position changes only modestly. The broad features and background fluctuations prevent a reliable determination of any field-induced frequency shift.

4 Discussion

Taken together, Figs. 4–7 establish that the electrically driven THz response increases with field and pulse duration and that the gated and ungated devices have different broad spectra. These observations do not identify a unique microscopic mechanism. Given the long electrical pulses, the response is consistent with predominantly thermal emission. The gated/ungated difference may reflect geometry-dependent thermal emissivity, while a separate collective 2DEG contribution cannot be resolved by the present measurements.

The broadband response is consistent with electrical heating followed by carrier-phonon energy relaxation. Related III-nitride studies reported impurity electroluminescence [29] and geometry-controlled thermal emission from n-GaN gratings [30]. Earlier room-temperature HEMT experiments reported broad, device-dependent spectra that were not fully described by the one-dimensional Dyakonov-Shur model [31]. Other AlGaN/GaN studies separated plasmon-related features from blackbody-like and impurity-related emission, and showed that accumulated heat can suppress spectral selectivity [15,16]. In the present device, longer 10–90 ms pulses increase the total output but reduce the apparent conversion slope, consistent with stronger lattice heating and nonradiative loss.

Drain-modulated reflection of the approximately 300 K surroundings could also contribute a lock-in signal at the electrical modulation frequency. The setup did not include independent reflection calibration, cryogenic optical isolation or nitrogen purging. This possible contribution, together with detector-window and atmospheric effects, limits quantitative interpretation of the processed spectra.

Independent thermal-radiation studies of AlGaN/GaN heterostructures emphasize the importance of lattice temperature [32]. Lisauskas et al. [17] likewise cautioned that radiation from biased HEMTs cannot be assigned to plasmonic instability without mechanism-specific controls.

A periodic grating can in principle supply in-plane momentum for coupling 2DEG excitations to radiation [12]. Gate-tunable plasmon absorption has been observed in grating-gate AlGaN/GaN structures under suitable conditions [13]. The present experiment, however, uses one grating period, fixed gate bias and broad room-temperature spectra. Consequently, momentum matching alone cannot assign the observed low-frequency feature to a collective mode.

The gated/ungated comparison does not isolate a collective-2DEG contribution from geometry-dependent thermal emissivity. A grating can make thermal emission frequency selective and partially polarized [30], so the observed spectral difference may be of geometric emissivity origin. The present data therefore do not establish that the gated enhancement arises from grating-coupled 2DEG emission. Calibrated emissivity and polarization-resolved spectra would be required to separate these possibilities.

An order-of-magnitude estimate places the gated-plasmon frequency above the observed 2–2.5 THz feature. Using q = 2π/P with P = 500 nm, n_s = 2.0 × 10¹³ cm⁻², d = 100 nm, m* = 0.2m₀ and ε_r = 9.5 in the gated-plasmon estimate gives a frequency near 11 THz. This is well above the observed broad low-frequency feature. This estimate does not support assigning the observed low-frequency feature to a gated-plasmon resonance.

The hydrodynamic interpretation of gated-2DEG emission depends on the device geometry and damping. Dyakonov and Shur showed that a gated 2DEG can behave hydrodynamically when boundary amplification exceeds collisional and viscous damping [7]. By contrast, the present N-polar device has a 600 μm source-drain separation, a distributed grating, broad spectra and a fixed gate bias. The field-dependent amplitude alone is therefore insufficient evidence for a classical Dyakonov–Shur instability.

The mobility-derived momentum-relaxation time is approximately 0.14 ps. At 2.2 THz, this gives ω_τ ≈ 1.9, which is consistent with a strongly damped broad response rather than a narrow high-Q plasmon line. The broad spectra do not allow a reliable extraction of a plasmon quality factor or mode frequency.

Other III-nitride emission channels remain relevant to the measured background. Shallow O and Si donors have been reported to produce THz electroluminescence under specific temperature and excitation conditions [29], whereas patterned n-GaN gratings can produce directional thermal emission controlled by geometry and angle [30]. The present room-temperature data do not establish a donor-related or collective-plasmon contribution. Future measurements should combine calibrated emissivity, polarization-resolved spectra using a wire-grid polarizer, shorter electrical pulses, temperature dependence, gate-voltage tuning and a second grating period. These controls would help assess whether a collective 2DEG contribution is needed in addition to thermally dominated emission and geometry-dependent emissivity.

The fabrication route is compatible with electron-beam-defined grating geometries and wafer-level layer transfer. A scalable process will require control of grating uniformity, contact resistance, thermal extraction and device-to-device reproducibility, together with mechanism-resolving optical measurements.

5 Conclusions

We report a room-temperature THz signal from an N-polar GaN grating-gate HEMT under pulsed electrical excitation. This work establishes an experimental N-polar GaN HEMT platform for electrically driven THz studies.

The detected THz power increases with lateral electric field and duty cycle, while the apparent conversion efficiency decreases at high duty cycle. The gated and ungated devices show different broad spectra. These observations demonstrate a field- and duty-cycle-dependent electrical THz response but do not identify a single resonance or a unique microscopic emission mechanism.

The response is consistent with predominantly thermal emission, while the gated/ungated spectral difference may reflect geometry-dependent thermal emissivity. A possible collective 2DEG contribution cannot be established from the present long-pulse room-temperature measurements. Further emissivity and polarization measurements, together with shorter pulses, gate-voltage tuning and a second grating period, are needed to test for a collective contribution.

The performance values in Table 1 are representative literature or present-device values and are not directly comparable because the device geometries, collection efficiencies and measurement conditions differ.

References

[1] Dhillon , S.S. , Vitiello , M.S. , Linfield , E.H. , Davies , A.G. , Hoffmann , M.C. , Booske , J. , Paoloni , C. , Gensch , M. , Weightman , P. , Williams , G.P. , Castro-Camus , E. , Cumming , D.R.S. , Simoens , F. , Escorcia-Carranza , I. , Grant , J. , Lucyszyn , S. , Kuwata-Gonokami , M. , Konishi , K. , Koch , M. , Schmuttenmaer , C.A. , Cocker , T.L. , Huber , R. , Markelz , A.G. , Taylor , Z.D. , Wallace , V.P. , Axel Zeitler, J. , Sibik , J. , Korter , T.M. , Ellison , B. , Rea , S. , Goldsmith , P. , Cooper , K.B. , Appleby , R. , Pardo , D. , Huggard , P.G. , Krozer , V. , Shams , H. , Fice , M. , Renaud , C. , Seeds , A. , Stöhr , A. , Naftaly , M. , Ridler , N. , Clarke , R. , Cunningham , J.E. , Johnston , M.B. : The 2017 terahertz science and technology roadmap. J. Phys. D Appl. Phys 50(4), 043001(2017)

[2] Siegel , P.H. : Terahertz technology. IEEE Trans. Microw. Theory Tech 50(3), 910–928(2002)

[3] Cooper , K.B. , Dengler , R.J. , Llombart , N. , Thomas , B. , Chattopadhyay , G. , Siegel , P.H. : THz imaging radar for standoff personnel screening. IEEE Trans. Terahertz Sci. Technol 1(1), 169–182(2011)

[4] Koenig , S. , Lopez-Diaz , D. , Antes , J. , Boes , F. , Henneberger , R. , Leuther , A. , Tessmann , A. , Schmogrow , R. , Hillerkuss , D. , Palmer , R. , Zwick , T. , Koos , C. , Freude , W. , Ambacher , O. , Leuthold , J. , Kallfass , I. : Wireless sub-THz communication system with high data rate. Nat. Photonics 7(12), 977–981(2013)

[5] Mishra , U.K. , Parikh , P. , Wu , Y.F. : AlGaN/GaN HEMTs - an overview of device operation and applications. Proc. IEEE 90(6), 1022–1031(2002)

[6] Ambacher , O. , Smart , J. , Shealy , J.R. , Weimann , N.G. , Chu , K. , Murphy , M. , Schaff , W.J. , Eastman , L.F. , Dimitrov , R. , Wittmer , L. , Stutzmann , M. , Rieger , W. , Hilsenbeck , J. : Two-dimensional electron gases induced by spontaneous and piezoelectric polarization charges in N- and Ga-face AlGaN/GaN heterostructures. J. Appl. Phys 85(6), 3222–3233(1999)

[7] Dyakonov , M. , Shur , M. : Shallow water analogy for a ballistic field effect transistor: new mechanism of plasma wave generation by dc current. Phys. Rev. Lett 71(15), 2465–2468(1993)

[8] Dyakonov , M. , Shur , M.S. : Detection, mixing, and frequency multiplication of terahertz radiation by two-dimensional electronic fluid. IEEE Trans. Electron Dev 43(3), 380–387(1996)

[9] Knap , W. , Deng , Y. , Rumyantsev , S. , Shur , M.S. : Resonant detection of subterahertz and terahertz radiation by plasma waves in submicron field-effect transistors. Appl. Phys. Lett 81(24), 4637–4639(2002)

[10] Knap , W. , Lusakowski , J. , Parenty , T. , Bollaert , S. , Cappy , A. , Popov , V.V. , Shur , M.S. : Terahertz emission by plasma waves in 60 nm gate high electron mobility transistors. Appl. Phys. Lett 84(13), 2331–2333(2004)

[11] El Fatimy, A. , Dyakonova , N. , Meziani , Y. , Otsuji , T. , Knap , W. , Vandenbrouk , S. , Madjour , K. , Théron , D. , Gaquiere , C. , Poisson , M.A. , Delage , S. , Prystawko , P. , Skierbiszewski , C. : AlGaN/GaN high electron mobility transistors as voltage-tunable room-temperature terahertz sources. J. Appl. Phys 107(2), 024504(2010)

[12] Popov , V.V. , Fateev , D.V. , Polischuk , O.V. , Shur , M.S. : Enhanced electromagnetic coupling between terahertz radiation and plasmons in a grating-gate transistor structure on membrane substrate. Opt. Express 18(16), 16771–16776(2010)

[13] Muravjov , A.V. , Veksler , D.B. , Popov , V.V. , Polischuk , O.V. , Pala , N. , Hu , X. , Gaska , R. , Saxena , H. , Peale , R.E. , Shur , M.S. : Temperature dependence of plasmonic terahertz absorption in grating-gate gallium-nitride transistor structures. Appl. Phys. Lett 96(4), 042105(2010)

[14] Onishi , T. , Tanigawa , T. , Takigawa , S. : High power terahertz emission from a single-gate AlGaN/GaN field effect transistor with periodic Ohmic contacts for plasmon coupling. Appl. Phys. Lett 97(9), 092117(2010)

[15] Jakštas , V. , Grigelionis , I. , Janonis , V. , Valušis , G. , Kašalynas , I. , Seniutinas , G. , Juodkazis , S. , Prystawko , P. , Leszczyński , M. : Electrically driven terahertz radiation of 2DEG plasmons in AlGaN/GaN structures at 110 K temperature. Appl. Phys. Lett 110(20), 202101(2017)

[16] Shalygin , V.A. , Moldavskaya , M.D. , Vinnichenko , M.Y. , Maremyanin , K.V. , Artemyev , A.A. , Panevin , V.Y. , Vorobjev , L.E. , Firsov , D.A. , Korotyeyev , V.V. , Sakharov , A.V. , Zavarin , E.E. , Arteev , D.S. , Lundin , W.V. , Tsatsulnikov , A.F. , Suihkonen , S. , Kauppinen , C. : Selective terahertz emission due to electrically excited 2D plasmons in AlGaN/GaN heterostructure. J. Appl. Phys 126(18), 183104(2019)

[17] Lisauskas , A. , Ramer , A. , Burakevic , M. , Chevtchenko , S. , Krozer , V. , Heinrich , W. , Roskos , H.G. : Terahertz emission from biased AlGaN/GaN high-electron-mobility transistors. J. Appl. Phys 125(15), 151614(2019)

[18] Pashnev , D. , Kaplas , T. , Korotyeyev , V. , Janonis , V. , Urbanowicz , A. , Jorudas , J. , Kašalynas , I. : Terahertz time-domain spectroscopy of two-dimensional plasmons in AlGaN/GaN heterostructures. Appl. Phys. Lett 117(5), 051105(2020)

[19] Janonis , V. , Pashnev , D. , Grigelionis , I. , Korotieiev , V. , Balagula , R.M. , Minkevicius , L. , Jorudas , J. , Alexeeva , N. , Subacius , L. , Valušis , G. , Kasalynas , I. : Electrically pumped THz emitters based on plasma waves excitation in III-nitride structures. Proc. SPIE 11499, 1149909(2020)

[20] Huang , Y.D. , Yu , Y. , Qin , H. , Sun , J.D. , Zhang , Z.P. , Li , X.X. , Huang , J.J. , Cai , Y. : Plasmonic terahertz modulator based on a grating-coupled two-dimensional electron system. Appl. Phys. Lett 109(20), 201110(2016)

[21] Wang , L. , Chen , X.S. , Hu , W.D. , Wang , J. , Wang , J. , Wang , X.D. , Lu , W. : The plasmonic resonant absorption in GaN double-channel high electron mobility transistors. Appl. Phys. Lett 99(6), 063502(2011)

[22] Nafari , M. , Aizin , G.R. , Jornet , J.M. : Plasmonic HEMT terahertz transmitter based on the Dyakonov−Shur instability: performance analysis and impact of nonideal boundaries. Phys. Rev. Appl 10(6), 064025(2018)

[23] Xing , R. , Guo , H. , Guo , B. , Yu , G. , Zhang , P. , Zhou , J. , Yang , A. , Li , Y. , Hao , C. , Yue , H. , Zeng , Z. , Zhang , X. , Zhang , B. : Numerical study of terahertz radiation from N-polar AlGaN/GaN HEMT under asymmetric boundaries. Front Optoelectron 18(1), 4(2025)

[24] Chung , J.W. , Piner , E.L. , Palacios , T. : N-face GaN/AlGaN HEMTs fabricated through layer transfer technology. IEEE Electron Device Lett 30(2), 113–116(2009)

[25] Park , P.S. , Rajan , S. : Simulation of short-channel effects in N- and Ga-polar AlGaN/GaN HEMTs. IEEE Trans. Electron Dev 58(3), 704–708(2011)

[26] Lu , J. , Denninghoff , D. , Yeluri , R. , Lal , S. , Gupta , G. , Laurent , M. , Keller , S. , DenBaars , S.P. , Mishra , U.K. : Very high channel conductivity in ultra-thin channel N-polar GaN/(AlN, InAlN, AlGaN) high electron mobility heterojunctions grown by metalorganic chemical vapor deposition. Appl. Phys. Lett 102(23), 232104(2013)

[27] Dasgupta , S. : Nidhi, Brown, D. F., Wu, F., Keller, S., Speck, J.S., Mishra, U.K.: Ultralow nonalloyed Ohmic contact resistance to self-aligned N-polar GaN high electron mobility transistors by In(Ga)N regrowth. Appl. Phys. Lett 96(14), 143504(2010)

[28] Guo , B.H. , Yu , G.H. , Zhang , L. , Zhou , J. , Wang , Z. , Xing , R. , Yang , A. , Li , Y. , Liu , B. , Zeng , X. , Du , Z. , Deng , X. , Zeng , Z. , Zhang , B. : High-performance N-polar GaN/AlGaN metal-insulator-semiconductor high-electron-mobility transistors with low surface roughness enabled by chemical-mechanical-polishing-incorporated layer transfer technology. Crystals (Basel) 14(3), 253(2024)

[29] Grigelionis , I. , Jakstas , V. , Janonis , V. , Kašalynas , I. , Prystawko , P. , Kruszewski , P. , Leszczyński , M. : Terahertz electroluminescence of shallow impurities in AlGaN/GaN heterostructures at temperatures above 80 K. Phys. Status Solidi, B Basic Res 255(5), 1700421(2018)

[30] Janonis , V. , Tumenas , S. , Prystawko , P. , Kacperski , J. , Kasalynas , I. : Investigation of n-type gallium nitride grating for applications in coherent thermal sources. Appl. Phys. Lett 116(11), 112103(2020)

[31] Dyakonova , N. , El Fatimy, A. , Lusakowski , J. , Knap , W. , Dyakonov , M.I. , Poisson , M.A. , Morvan , E. , Bollaert , S. , Shchepetov , A. , Roelens , Y. , Gaquiere , C. , Theron , D. , Cappy , A. : Room-temperature terahertz emission from nanometer field-effect transistors. Appl. Phys. Lett 88(14), 141906(2006)

[32] Grigelionis , I. , Kašalynas , I. : Terahertz spectroscopy of thermal radiation from AlGaN/GaN heterostructure on sapphire at low temperatures. Appl. Sci. (Basel) 10(3), 851(2020)

Rights & permissions

The author(s)