Frequency-Resolved Forward Capacitance in GaN-based LEDs

Bo Shen; Boyan Dong; Chuhan Deng; Daqi Wang; Fei Jiao; Guoyi Zhang; Haodong Zhang; Jingxin Nie; Qi Wang; Weihua Chen

REVIEW 4 major objections 6 minor 27 references

Frequency-Resolved Forward Capacitance in GaN-based LEDs

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Forward-bias negative capacitance in GaN LEDs is an intrinsic trap-emission effect, not a measurement artifact.

desk verdict Broadband C-f data and a trap-inductance model that reads plausibly, but the key attribution to intrinsic trap emission rests on fitting Im(Z) alone. read the letter →

arxiv 2411.16626 v3 pith:46OXBDPV submitted 2024-11-25 physics.app-ph

Yuchen Li , Zhizhong Chen , Chuhan Deng , Boyan Dong , Daqi Wang , Zuojian Pan , Haodong Zhang , Jingxin Nie

show 7 more authors

Weihua Chen Fei Jiao Xiangning Kang Qi Wang Guoyi Zhang Bo Shen Wenji Liang

This is my paper · ORCID

classification physics.app-ph

keywords negativecapacitanceGaNLEDimpedancespectroscopytrapstatesequivalentcircuitcarrieremissioncapacitance-voltage-frequencyinductive

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to settle a long-running dispute over why forward-biased GaN LEDs show negative capacitance at MHz frequencies: is it a real device property or a measurement artifact? Its answer is that the negative capacitance is real and originates in delayed emission of carriers from deep-level trap states, not in series resistance or parasitic inductance. Broadband capacitance-frequency measurements between 40 Hz and 6.4 MHz separate three regimes—interfacial response below 1 kHz, diffusion and recombination near 100 kHz, and trap-mediated inductive response above 3 MHz. The unified equivalent circuit reproduces the entire spectrum with sub-1% fitting errors (R² > 0.99) and extracts a trap inductance around 5 µH, three orders of magnitude larger than packaging parasitics. If right, this turns the negative-capacitance dip into a measurable signature of trap dynamics and a design handle for fast-switching LEDs.

What carries the argument

The central object is a three-branch equivalent circuit: a contact-interface branch with capacitance and conductance, an active-region branch with capacitance C_a and conductance g_a, and a high-frequency branch containing a series resistance and an inductance L_trap that models delayed carrier emission from trap states. The identity that carries the argument is Eq. (6), $$\mathrm{Im}(Z) = -\frac{\omega C_a}{$g_a^{2}$+\$omega^{2}$ $C_a^{2}$}+\omega L_{\mathrm{trap}},$$ which turns measured imaginary-impedance curves into device parameters plus one inductive term. On the physics side, a Shockley-Read-type occupation-rate equation for interface states, including capture and emission at traps below the Fermi level, predicts that this effective inductance saturates at high injection; the observed saturation of L_trap near 5 µH is the main evidence that the circuit element is physical. The low-frequency regime is handled separately by integrating Debye-like trap responses over a logarithmic lifetime distribution, which yields the observed C ∝ 1/f scaling.

What would settle it

Perform a time-domain small-signal step response on the same LEDs and look for a delayed current overshoot whose time constant corresponds to the fitted L_trap ≈ 5 µH; alternatively, remeasure with a four-terminal on-wafer calibration to eliminate contact and lead parasitics. If the MHz negative capacitance vanishes under four-terminal measurement, or the transient shows no trap-delayed inductive component, the intrinsic-trap-emission claim is falsified.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that the negative capacitance seen in forward-biased InGaN/GaN LEDs is caused by delayed carrier emission from trap and localized states below the Fermi level, not by extrinsic measurement artifacts. The authors support this by fitting a hybrid equivalent circuit to broadband impedance data and finding that the high-frequency branch requires a series inductance L_trap that grows with injection current and saturates near 5 µH, far above any package inductance. They also show that the sub-kHz capacitance follows a 1/f law that matches a continuum of trap lifetimes, and that quantum-well cap thickness changes the active-region capacitance and conductance. The result unifies previously separate explanations of low-frequency and high-frequency capacitance anomalies in one model.

Load-bearing premise

The argument depends on assuming that the measured impedance spectrum has one correct circuit explanation—a single active-region capacitance in parallel with a conductance, plus a series inductance from traps—and that no other arrangement of resistances and capacitances could reproduce the same curves.

Editorial extensions

If this is right

The MHz negative capacitance in GaN LEDs can be treated as an intrinsic trap signature; it should persist under careful calibration and scale with injection and temperature as the model predicts.
Quantum-well cap thickness becomes a design lever: switching from GaN to AlGaN capping reduces mid-frequency capacitance by about 30% and lowers the extracted active-region capacitance and conductance by about 10%.
The extracted L_trap provides a quantitative, frequency-domain metric for trap-mediated carrier delay, complementing capacitance-voltage and deep-level transient spectroscopy.
Broadband capacitance-frequency spectroscopy, rather than a few discrete frequency points, is required to separate interfacial, diffusion-recombination, and trap-emission contributions in LED impedance data.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A consequence the authors leave implicit: if L_trap is real, MHz-rate micro-LED driving must account for a trap-induced inductive transient in addition to the usual RC charging time.
A testable extension: devices with intentionally different deep-level densities—for example, grown under different V/III ratios—should show proportionally larger or smaller L_trap and correspondingly stronger or weaker negative capacitance.
The logarithmic trap-lifetime continuum used for the sub-kHz 1/f tail could be checked by variable-temperature measurements: the extracted lifetime distribution should shift in a way that reveals the activation energy of the states.
A time-domain pulse experiment would provide a direct cross-check of the inductance picture: a small forward-voltage step should produce a delayed current overshoot whose time constant tracks L_trap, while a purely parasitic explanation would show no such overshoot.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

arXiv:2411.16626 is a useful addition to the LED impedance literature, but the headline claim needs another round of work. The new data are real: broadband C-f sweeps from 40 Hz to 6.4 MHz on two LED variants and an n-GaN control, with a clear cap-layer comparison. The low-frequency 1/f dispersion and its distributed-lifetime explanation are well argued, and the series-RL-plus-parallel-C-G framework does tie the three regimes together in a single picture.

The soft spot is exactly where the stress-test lands. The negative-capacitance-to-inductance mapping in Eq. (4) is definitional—measuring Im(Z) > 0 at high frequency is equivalent to a positive series reactance. Attributing that reactance specifically to trap emission requires more than a good fit of Eq. (6) to Im(Z). The paper never checks whether the extracted Ca, ga, and Ltrap reproduce Re(Z); it never compares the proposed circuit against an alternative with, say, a frequency-dependent resistance or a distributed interface; and Eq. (9) is stated without the steps connecting it to the fitted 5 μH. The current dependence of Ltrap does make a constant package parasitic unlikely, but it leaves open other current-dependent delay mechanisms.

There are also internal inconsistencies that a referee would want fixed: the abstract says the cap thickness reduces mid-frequency capacitance by 30% while the conclusion says ~10%, and the abstract and text disagree about whether the model covers 10 kHz–6.4 MHz or 40 Hz–6.4 MHz. No error bars are given, and the parameter count per bias point is high enough that a sub-1% fit is not surprising.

That said, I don't think the paper is wrong in its general direction. The temperature dependence and the cap-layer effect are consistent with a carrier-mediated origin, and the data are worth having. The central attribution is plausible but not established.

I would send this to review. A good referee can push for the missing Re(Z) validation, an alternative-model test, and a clean derivation of Eq. (9). The authors appear to have the equipment and samples to do the extra experiments. I wouldn't cite the trap-inductance value yet, but I would cite the broadband C-f data if I were working on GaN LED impedance.

Referee Report

4 major / 6 minor

Summary. The manuscript reports forward-bias capacitance-frequency measurements on two InGaN/GaN LED structures with different cap layers (GaN vs AlGaN) over 40 Hz–6.4 MHz. A hybrid equivalent circuit with series RL and parallel C-G elements is used to decompose the impedance into contact, active-region, and trap-mediated components. The authors identify three frequency regimes: sub-kHz interfacial C∝1/f behavior, mid-frequency diffusion-recombination response, and high-frequency negative capacitance attributed to deep-level trap emission. They also report a cap-layer effect on mid-frequency capacitance and a fitted trap inductance of ~5 μH.

Significance. If the central attribution is correct, the work would provide a unified explanation of negative capacitance in GaN LEDs and a practical design rule (cap-layer tuning) for high-speed display devices. The paper's experimental strengths include broadband characterization, a control structure (metal/n-GaN/metal) that supports the low-frequency interface interpretation, and temperature-dependent Im(Z) data. However, the load-bearing claim that MHz-range negative capacitance originates from intrinsic trap emission is underdetermined by the current analysis, and several quantitative claims are internally inconsistent.

major comments (4)

[Abstract vs. Conclusion] The abstract states that cap thickness modulation reduces mid-frequency capacitance by 30% and that the model covers 10 kHz–6.4 MHz, while the conclusion states a ~10% reduction and the methods section states sweeps from 40 Hz. These discrepancies are load-bearing because the cap-thickness effect and frequency coverage are central quantitative claims. Please correct the numbers and specify which value is measured.
[Eq. (6) and fitting procedure] The central attribution of the high-frequency negative capacitance to trap emission rests on fitting Im(Z) with three free parameters (C_a, g_a, L_trap) per bias point, using Eq. (6). The paper does not validate the fit against the measured Re(Z), nor does it test alternative circuit topologies without L_trap (e.g., residual series inductance, distributed RC delay, contact transit-time effects). Since Eq. (4) defines Im(Z) = ωL_s by construction, any positive increasing Im(Z) component is automatically absorbed into the ωL_trap term, so the fit alone cannot discriminate intrinsic trap emission from measurement artifacts. Please fit the full complex impedance, test a model without L_trap, and report the stability of the extracted parameters across candidate topologies.
[Eq. (9) and the trap model] Equation (9) is introduced without a derivation and uses symbols (C_i, ΔV/δj, σ_i) that are not all defined. The paper never evaluates Eq. (9) with independent estimates of the capture cross sections, lifetimes, and interface capacitance to show that the trap model actually predicts L ≈ 5 μH. This missing quantitative link leaves the proposed mechanism disconnected from the fitted inductance. Please provide the full derivation or a precise citation and a numerical estimate with physically plausible parameter values.
[Error bars and statistical significance] No error bars or measurement repeatability are reported, and the 'sub-1% fitting errors (R² > 0.99)' is not defined quantitatively. The cap-thickness effect is quoted as 30% in the abstract and 10% in the conclusion, and without uncertainty estimates or multiple devices it is impossible to judge whether the difference is significant. Please report standard deviations or confidence intervals for the extracted parameters and for the cap-thickness comparison.

minor comments (6)

[References] Reference 12 is duplicated: both Shim et al. and Han et al. are assigned the number 12. Please renumber the reference list.
[Low-frequency section] The sentence 'This The model matches experimental data' contains a typo; it should read 'This model matches the experimental data.'
[Eq. (2)] The integral leading to C(ω) ∝ π/(2ω) is stated without the normalization constant a; please define a and show the integration step explicitly.
[Introduction] The phrase 'verse-bias capacitance' appears in the discussion of prior trap studies; it should be 'reverse-bias capacitance.'
[Figure 11] The temperature-dependent Im(Z) data are presented qualitatively; extracting L(T) quantitatively and comparing it with a thermal activation model would strengthen the trap-emission interpretation.
[Methods] The text says the system was calibrated with open/short/load compensation, but the residual inductance after compensation is not quantified; reporting the residual L would help exclude extrinsic contributions to the ωL_trap term.

Circularity Check

2 steps flagged · score 6.0 of 10

High-frequency 'trap-mediated inductance' is a fitted relabeling of negative capacitance, and the low-frequency 1/f 'explanation' assumes the distribution that produces 1/f.

self definitional [Results and Discussion, low-frequency C-f section, Eqs. (1)-(2)]
"Whereas individual traps with characteristic lifetime τ produce: 𝐶(𝜔) = 𝑎 ∫ exp (− 𝑡/𝜏) cos𝜔𝑡 𝑑𝑡 ... a logarithmic lifetime distribution (g(τ)∝1/τ) yields: 𝐶(𝜔) ∝ ∫ 𝜏/(1 + 𝜔2𝜏2) 1/𝜏 𝑑𝜏 = 𝜋/2𝜔. This model matches experimental data, suggesting long-lifetime traps (τ ∼ 1 ms) influence low-frequency capacitance."

The observed regime is C ∝ 1/f, and the paper 'resolves the paradox' by positing g(τ) ∝ 1/τ. That distribution is the unique choice whose Debye integral evaluates to 1/ω; the derivation is therefore an ansatz selected to reproduce the measured frequency dependence. No independent measurement of the trap-lifetime distribution is provided, so the model does not predict the 1/f law: it encodes it in the assumed g(τ).
fitted input called prediction [Results and Discussion, Eqs. (4)-(6) and Fig. 8 fitting discussion]
"𝐼𝑚(𝑍̃) = −𝜔𝐶𝑝/(𝑔2+𝜔2𝐶𝑝2) = 𝜔𝐿𝑠 (4) ... Using Equation: 𝐼𝑚(𝑍) = −𝜔𝐶𝑎/(𝑔𝑎2+𝜔2𝐶𝑎2) + 𝜔𝐿𝑡𝑟𝑎𝑝 (6) the experimental Im(Z)-f data (Figure 6) are fitted to extract 𝐶𝑎, 𝑔𝑎, and 𝐿𝑡𝑟𝑎𝑝 ... Key findings include: 𝐿𝑡𝑟𝑎𝑝 increases with injection current, saturating at ~5μH for I>10 mA."

Equation (4) already defines any negative parallel capacitance as a positive series inductance, so a fitted ωL_trap term in Eq. (6) is mathematically forced to absorb any positive high-frequency Im(Z) slope. The extracted L_trap ≈ 5 µH is thus a fitted parameter relabeled as 'trap-mediated inductance' and then cited as evidence that negative capacitance is intrinsic. The paper does not evaluate Eq. (9) with independently measured capture/emission parameters to show trap kinetics quantitatively produce ~5 µH, nor does it validate against Re(Z) or alternative circuit topologies. The current-dependence support is the same fitted trend used to build the model.

full rationale

This paper is not built on a load-bearing self-citation chain: the trap-kinetics equations [26,27] and occupancy response [23,24] are external works, and the authors' own prior contribution [15] appears only in the historical survey. The circularity is interpretive rather than citational. At low frequency, the observed C ∝ 1/f is 'explained' by assuming g(τ) ∝ 1/τ, an assumption chosen precisely so the Debye integral evaluates to 1/ω; this is a self-definitional derivation rather than an independent prediction. At high frequency, Eq. (4) states that any negative parallel capacitance is equivalent to a positive series inductance, and Eq. (6) then fits an extra ωL_trap term to the same Im(Z)-f data. The extracted L_trap ≈ 5 µH is therefore a fitted parameter relabeled as a trap-emission signature. The paper does not independently evaluate Eq. (9) with measured cross-sections or trap densities to show that the kinetics yield 5 µH, nor does it fit the real part or test alternative equivalent circuits that could produce the same spectra. Temperature dependence supplies some independent phenomenology, but the central 'intrinsic trap origin' conclusion rests on the fitted L_trap term. Score 6 reflects partial circularity: the low-frequency 1/f derivation is an ansatz tailored to the data, and the high-frequency attribution reduces largely to a mathematical identity plus a fitted inductance.

Assumptions & free parameters 4 free parameters · 3 assumptions · 1 invented entities

The central measurement is a fit of an equivalent circuit with multiple free parameters per bias condition. The physical mechanisms (trap emission, interface states) are borrowed from prior models, and the paper does not provide independent measurements of the trap parameters. The invented entity is an effective inductance used to represent negative capacitance, which by construction is equivalent to a negative capacitance in the series model.

free parameters (4)

Ca (active-region capacitance) = per bias point, shown in Fig. 9
Extracted by fitting Eq. (6) to the measured Im(Z) spectra; not independently predicted.
ga (active-region conductance) = per bias point, shown in Fig. 9
Extracted by fitting Eq. (6) to the measured Im(Z) spectra; not independently predicted.
Ltrap (trap-mediated inductance) = saturates near 5 uH for I > 10 mA
Extracted by fitting Eq. (6); the physical interpretation as trap emission is asserted from the model's fit.
Effective trap lifetimes tau_eff1, tau_eff2 and capture cross sections sigma_n, sigma_i = not explicitly reported
These parameters enter the trap occupation model leading to Eq. (9) but are not independently measured; they are used to explain the current dependence of Ltrap.

assumptions (3)

domain assumption The low-frequency 1/f capacitance arises from a logarithmic distribution of trap lifetimes g(tau) proportional to 1/tau.
Invoked in Eq. (2) to derive C proportional to 1/f; no independent evidence for the distribution shape is given.
ad hoc to paper The device impedance can be decomposed into three independent parallel/series blocks (contact, active region, trap) with no cross-coupling.
This is the core modeling assumption behind Eq. (5) and the fitting in Fig. 8; the separability is not tested against alternative circuit topologies.
domain assumption The trap occupation kinetics follow the Wu-Yang collision model, with electrons crossing the barrier ejecting trapped carriers.
Taken from references [26-27] and applied to GaN LEDs without verification for this material system.

invented entities (1)

Effective trap inductance L_trap
purpose: Model delayed carrier emission from deep-level states as a series inductance in the equivalent circuit.
It is an extracted fitting parameter, not a directly measured physical quantity. Its existence is inferred from the current and temperature dependence of the fitted values.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Frequency-Resolved Forward Capacitance in GaN-based LEDs." pith.science (2026). https://pith.science/paper/46OXBDPV

@misc{pith2026241116626,
  author       = {Pith},
  title        = {Pith review of: Frequency-Resolved Forward Capacitance in GaN-based LEDs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/46OXBDPV}},
  note         = {Machine review of arXiv:2411.16626}
}

read the original abstract

This study establishes a unified framework for interpreting dynamic capacitive responses in InGaN-based light-emitting diodes (LEDs) through forward-bias capacitance-voltage-frequency spectroscopy. A hybrid impedance model integrating series RL components and parallel C-G networks was developed to resolve distinct frequency-dependent capacitive regimes. The low-frequency regime (<1 kHz) is governed by interfacial capacitance with characteristic reciprocal frequency dependence, while the mid-frequency range(10 kHz-6.4 MHz) demonstrates carrier diffusion and recombination dynamics. At MHz frequencies, negative capacitance manifests due to delayed carrier emission mediated by deep-level traps. The model achieved sub-1% fitting errors (R^2 > 0.99)across a broad bandwidth(10 kHz-6.4 MHz) , conclusively attributing negative capacitance to intrinsic trap processes rather than extrinsic artifacts. Critical advances include quantum well cap thickness modulation reducing mid-frequency capacitance by 30% and the dominance of trap-mediated inductance over parasitic contributions by three orders of magnitude. This framework resolves persistent controversies in LED impedance interpretation. By bridging semiconductor physics with device engineering, this methodology provides essential tools for designing next-generation optoelectronic systems requiring ultralow-latency operation and precise charge-state control.

Figures

Figures reproduced from arXiv: 2411.16626 by the authors.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 27 canonical work pages

[1]

J., Chen, Z

Pan, Z. J., Chen, Z. Z., Jiao, F., Zhan, J. L., Chen, Y . Y ., Chen, Y . F., Nie, J. X., Zhao, T. Y ., Deng, C. H., Kang, X. N., Li, S. F., Wang, Q., Zhang, G. Y ., & Shen, B. (2020). A review of key technologies for epitaxy and chip process of micro light-emitting diodes in display application. Wuli xuebao, 69(19), 64–87

work page 2020
[2]

Solid State Sci

Wong M S, Nakamura S, DenBaars S 2020 ECS J. Solid State Sci. Technol. 9 015012

work page 2020
[3]

Gong, Y ., Zhang, L., Lin, P., Yuan, Z., Peng, L.-M., & Kang, J. (2024). P ‐140: The Miniaturization of InGaN/GaN Micro ‐LEDs for Micro ‐Displays – Size Effects, Frequency Dispersion and Compact Modeling. SID International Symposium Digest of technical papers, 55(1), 1928–1931

work page 2024
[4]

Solid-State Electronics,1982,25(5):355-357

Sandoval, F., Lopez, C., Munoz, E., Forward-bias impedance of GaAs1 -xPxLED's [J]. Solid-State Electronics,1982,25(5):355-357

work page 1982
[5]

Bozkurt, K., Özdemir, O., Ayarcı Kuruoğlu, N., Alshehri, B., Dogheche, K., Gaimard, Q., Ramdane, A., & Dogheche, E. (2019). Impact of trap states on inductive phenomena in 30% InGaN/GaN MQW LED devices. Journal of Physics D: Applied Physics, 52(10), 105102

work page 2019
[6]

Xiao, Y ., Feng, X., Meng, Y ., He, L., Zhang, P., Zhang, D., Gao, S., Shields, P., Tian, H., & Liu, H. (2025). Negative Capacitance Analysis of Multi-Quantum-Well Light-Emitting Diodes. Electronics, 14(3), 413

work page 2025
[7]

Sarcan, F., Masoumi, A., Kalyon, G., & Erol, A. (2025). Negative capacitance in a novel bidirectional p-i- n diode for near-infrared LED and photodetector applications. Optics & Laser Technology, 190, 113207

work page 2025
[8]

Mock, J., Kallergi, M., Groß, E., Golibrzuch, M., Rieger, B., & Becherer, M. (2022). Revealing the Negative Capacitance Effect in Silicon Quantum Dot Light -Emitting Diodes via Temperature -Dependent Capacitance-V oltage Characterization. IEEE Photonics Journal, 14(4), 8237309

work page 2022

Show all 27 references

[9]

Bourim, E.-M., & Han, J. I. (2016). Size effect on negative capacitance at forward bias in InGaN/GaN multiple quantum well-based blue LED. Electronic Materials Letters, 12(1), 67-75. 10.Bao, X., Xu, J., Li, C., Qiao, H., Zhang, Y ., & Li, X. (2013). Temperature and frequency d...

2016
[11]

Arslan, E., Şafak, Y ., Altındal, Ş., Kelekçi, Ö., & Özbay, E. (2010). Temperature dependent negative capacitance behavior in (Ni/Au)/AlGaN/AlN/GaN heterostructures. Journal of Non -Crystalline Solids, 356(20), 1006–1011

2010
[12]

Bisquert, J., Garcia-Belmonte, G., Pitarch, Á., & Bolink, H. J. (2006). Negative capacitance caused by electron injection through interfacial states in organic light-emitting diodes. Chemical physics letters, 422(1), 184–191

2006
[13]

Han, D.-P., Kim, Y .-J., Shim, J.-I., & Shin, D. -S. (2016). Forward-Capacitance Measurement on Wide- Bandgap Light-Emitting Diodes. IEEE photonics technology letters, 28(21), 2407–2410

2016
[14]

F., Li, Y ., Zhu, C

Feng, L. F., Li, Y ., Zhu, C. Y ., Cong, H. X., & Wang, C. D. (2010). Negative Terminal Capacitance of Light Emitting Diodes at Alternating Current (AC) Biases. IEEE journal of quantum electronics, 46(7), 1072– 1075

2010
[15]

Feng, L., Wang, S., Li, Y ., Yang, X., Li, D., & Wang, C. (2018). Opposite change trend of electrical behavior curves near the threshold between GaAs - and GaN -multi-quantum-well laser diodes. Applied Physics B, 124(3)

2018
[16]

Y ., Feng, L

Zhu, C. Y ., Feng, L. F., Wang, C. D., Cong, H. X., Zhang, G. Y ., Yang, Z. J., & Chen, Z. Z. (2009). Negative capacitance in light-emitting devices. Solid-State Electronics, 53(3), 324–328

2009
[17]

Yang, W., Zhang, S., McKendry, J. J. D., Herrnsdorf, J., Tian, P., Gong, Z., Ji, Q., Watson, I. M., Gu, E., Dawson, M. D., Feng, L., Wang, C., & Hu, X. (2014). Size -dependent capacitance study on InGaN -based micro-light-emitting diodes. Journal of Applied Physics, 116(4)

2014
[18]

C., Li, L., Buchanan, M., Wasilewski, Z

Ershov, M., Liu, H. C., Li, L., Buchanan, M., Wasilewski, Z. R., & Jonscher, A. K. (1998). Negative capacitance effect in semiconductor devices. IEEE transactions on electron devices, 45(10), 2196–2206

1998
[19]

Irokawa, Y ., Matsuki, N., Sumiya, M., Sakuma, Y ., Sekiguchi, T., Chikyo, T., Sumida, Y ., & Nakano, Y . (2010). Anomalous capacitance -voltage characteristics of Pt -AlGaN/GaN Schottky diodes exposed to hydrogen. Physica Status Solidi C, 7(7-8), 1928-1930

2010
[20]

Doğan, H., Yıldırım, N., Orak, İ., Elagöz, S., & Turut, A. (2015). Capacitance -conductance-frequency characteristics of Au/Ni/n-GaN/undoped GaN Structures. Physica B: Condensed Matter, 457, 48–53

2015
[21]

Janardhanam, V ., Kim, J.-H., Jyothi, I., Kang, M.-S., Lee, S.-K., & Choi, C.-J. (2023). Carrier transport across PtSe₂/n-type GaN heterojunction. Vacuum, 218, 112597

2023
[22]

A., & Maaref, H

Charfeddine, M., Jabli, F., Zaidi, M. A., & Maaref, H. (2014). Deep traps responsible for capacitance hysteresis in AlGaN/GaN FAT-HEMT’s studied under the temperature effects. Journals of Optoelectronics and Advanced Materials, 16(7–8), 820–825

2014
[23]

Reddy, M. S. P., Lee, J. -H., & Jang, J. -S. (2013). Frequency dependent series resistance and interface states in Au/bio-organic/n-GaN Schottky structures based on DNA biopolymer. Synthetic Metals, 185–186, 167–171

2013
[24]

S.-S., Khasanah, R

Chien, F. S.-S., Khasanah, R. A. N., Lin, P.-T., Lin, Y .-F., & Suen, Y .-W. (2019). Impedance Elements of Significant Junctions in InGaN Light -Emitting Diodes Studied by Electric Modulus Spectroscopy. IEEE Transactions on Electron Devices, 66(8), 3393–3398

2019
[25]

S., Janardhanam, V ., Reddy, V

Reddy, D. S., Janardhanam, V ., Reddy, V . R., & Choi, C.-J. (2023). Modification of interface properties of Au/n -GaN Schottky junction by rare -earth oxide Nd₂O₃ as an interlayer and its microstructural characterization. Vacuum, 215, 112300

2023
[26]

A., Brown, D

Schaake, C. A., Brown, D. F., Swenson, B. L., Keller, S., Speck, J. S., & Mishra, U. K. (2013). A donor- like trap at the InGaN/GaN interface with net negative polarization and its possible consequence on internal quantum efficiency. Semiconductor Science and Technology, 28(10...

2013
[27]

Wu, X., & Yang, E. S. (1989). Interface capacitance in metal-semiconductor junctions. Journal of applied physics, 65(9), 3560–3567

1989
[28]

S., & E vans, H

Wu, X., Yang, E. S., & E vans, H. L. (1990). Negative capacitance at metal -semiconductor interfaces. Journal of applied physics, 68(6), 2845–2848

1990

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

J., Chen, Z

Pan, Z. J., Chen, Z. Z., Jiao, F., Zhan, J. L., Chen, Y . Y ., Chen, Y . F., Nie, J. X., Zhao, T. Y ., Deng, C. H., Kang, X. N., Li, S. F., Wang, Q., Zhang, G. Y ., & Shen, B. (2020). A review of key technologies for epitaxy and chip process of micro light-emitting diodes in display application. Wuli xuebao, 69(19), 64–87

work page 2020

[2] [2]

Solid State Sci

Wong M S, Nakamura S, DenBaars S 2020 ECS J. Solid State Sci. Technol. 9 015012

work page 2020

[3] [3]

Gong, Y ., Zhang, L., Lin, P., Yuan, Z., Peng, L.-M., & Kang, J. (2024). P ‐140: The Miniaturization of InGaN/GaN Micro ‐LEDs for Micro ‐Displays – Size Effects, Frequency Dispersion and Compact Modeling. SID International Symposium Digest of technical papers, 55(1), 1928–1931

work page 2024

[4] [4]

Solid-State Electronics,1982,25(5):355-357

Sandoval, F., Lopez, C., Munoz, E., Forward-bias impedance of GaAs1 -xPxLED's [J]. Solid-State Electronics,1982,25(5):355-357

work page 1982

[5] [5]

Bozkurt, K., Özdemir, O., Ayarcı Kuruoğlu, N., Alshehri, B., Dogheche, K., Gaimard, Q., Ramdane, A., & Dogheche, E. (2019). Impact of trap states on inductive phenomena in 30% InGaN/GaN MQW LED devices. Journal of Physics D: Applied Physics, 52(10), 105102

work page 2019

[6] [6]

Xiao, Y ., Feng, X., Meng, Y ., He, L., Zhang, P., Zhang, D., Gao, S., Shields, P., Tian, H., & Liu, H. (2025). Negative Capacitance Analysis of Multi-Quantum-Well Light-Emitting Diodes. Electronics, 14(3), 413

work page 2025

[7] [7]

Sarcan, F., Masoumi, A., Kalyon, G., & Erol, A. (2025). Negative capacitance in a novel bidirectional p-i- n diode for near-infrared LED and photodetector applications. Optics & Laser Technology, 190, 113207

work page 2025

[8] [8]

Mock, J., Kallergi, M., Groß, E., Golibrzuch, M., Rieger, B., & Becherer, M. (2022). Revealing the Negative Capacitance Effect in Silicon Quantum Dot Light -Emitting Diodes via Temperature -Dependent Capacitance-V oltage Characterization. IEEE Photonics Journal, 14(4), 8237309

work page 2022