betaFeedback

WorksPhysics and Astronomy

Voltage-Current-Emission Framework of Time-Resolved Electroluminescence

Abstract

Time-resolved electroluminescence has been a central methodology in the development of LED devices. However, it still lacks quantitative theoretical descriptions of time-resolved electroluminescence. By combining the equivalent circuit models of semiconductor devices and the carrier dynamics of electroluminescence processes, we here developed a Voltage-Current-Emission framework to illustrate time-resolved electroluminescence from the pulse voltage as input, to injection current and finally generate electroluminescence as output. The features of different LED devices can be quantitatively interpreted, confirming the universal applicability across LEDs with different materials and device structures. Importantly, the curves of time-resolved electroluminescence can be simulated to quantitatively describe the dynamics of current and carriers of LED operation. In all, the Voltage-Current-Emission framework offers theoretical guidance for achieving advanced LED devices for AI technology.

Main

Light emitting diodes (LED) have been central devices to revolute lighting and display technology[1, 2, 3], providing products of GaN-LED[4, 5], organic light emitting diode (OLED)[6], and demos of quantum dots light emitting diode (QLED)[7]. With the revolution of AI technology, there are emerging demands to develop advanced LED devices with high resolution, high brightness, and/or high response speed for AR display[8, 9, 10, 11], optical communication[12], and optical computer[13]. Time-resolved electroluminescence (TREL) measures the electroluminescence (EL) signals over time under an applied pulse voltage, while time-resolved current (TRC) can record the evolution of current with time prolonging[14, 15, 16, 17, 18, 19]. Previous analysis of TREL curves have tried to illustrate the carrier dynamics of LED operation from the microscopic viewpoint including recombination, transport and injection[17, 20, 21, 22, 23, 24, 25]. Equivalent circuit models have been widely used in the analysis of semiconductor devices but have been neglected in TREL interpretations[25, 26, 27, 28, 29]. Consequently, it still lacks a comprehensive picture to quantitatively describe the TREL curves.

Typically, the TREL curves of LED devices usually contain four stages including delay, rising, balance, and decay[16, 25]. Delay stage shows immediate current flow under voltage application while no detectable luminescence is observed. During rising stage, the time-resolved current approaches to a balance state, meanwhile the EL signals rises and approach to maximum value with time prolonging. Balance stage is similar with operation state under constant voltage. Decay stage reveals the features of luminescence decay after pulse voltage. It is worth noting that there are also conflicts in describing TREL stages. For example, the rising edge has been attributed to the rise of electron concentration[23], hole concentration[24], or both[25]. The delay stage was correlated to the transport time of low-mobility carriers[30, 31, 32], or the charging time of device capacitance[25], or their combination[33, 22]. These limitations have led to many misunderstandings in the interpretation of TREL.

We here discover the operational principle of "voltage drives current, current drives emission" in LED devices. In response to an applied pulse voltage, LED devices immediately generate current, and then the injected current leads to the formation of excitons with following emission. Based on this principle, the TREL dynamics can be correlated with time-resolved current using equivalent circuit models. To verify this point, we varied external resistors (variable control 1), low-level voltages (variable control 2), and layer thicknesses of hole transport layer, emissive layer and electron transport layer (variable control 3,4,5) to experimentally connect the equivalent circuit model and carrier dynamics of LEDs using QLED as a platform (Figure 1a), and then the conclusions are verified in GaN-LED and OLED.

Six-panel schematic and TREL response plots

Figure 1. a) Schematic diagram of the TREL test circuit and variable control. b) TREL response for QLED with external resistances of 100 Ω\Omega, 450 Ω\Omega, and 1100 Ω\Omega. c) Delay phase characteristics of TREL curves under different VLV_{\mathrm{L}} voltages. d) Variation of TREL curve delay time with VLV_{\mathrm{L}} voltage in QLEDs. e) TREL curves of QLED devices with different quantum dot layer thicknesses. f) Decay time of TREL curves versus equilibrium current for QLEDs with different ETL thicknesses.

Figure 1b shows the typical TREL curves of QLED under different external resistors. To ensure that the voltage drop across the LED device remained constant, the applied pulse voltage was adjusted accordingly when the external resistor was varied. With external resistors increasing from 100 Ω\Omega to 1100 Ω\Omega, delay time (TdT_{\mathrm{d}}) increases from 0.85 μs\mu\mathrm{s} to 1.70 μs\mu\mathrm{s} and decay time (TcT_{\mathrm{c}}) increases from 0.44 μs\mu\mathrm{s} to 0.81 μs\mu\mathrm{s}. As shown in the inset of Figure 1a, VLV_{\mathrm{L}} represents the voltage level of the pulse signal at its low state and VHV_{\mathrm{H}} represents the voltage level of the pulse signal at its high state. Figure 1c shows the relationship between delay time (TdT_{\mathrm{d}}) and the low-level voltage (VLV_{\mathrm{L}}) of the pulse signal, with its high-level voltage (VHV_{\mathrm{H}}) maintained constant. Delay time (TdT_{\mathrm{d}}) decreases from 1.0 μs\mu\mathrm{s} to near zero with the VLV_{\mathrm{L}} increasing from 0V to 2.0V. As shown in Figure 1d, when the VLV_{\mathrm{L}} gets close to the device's turn-on voltage, TdT_{\mathrm{d}} almost approaches to zero. Figure 1e shows the delay stage of TREL curves for QLED with QD layers of different thickness. When the QD layer becomes thicker, delay time (TdT_{\mathrm{d}}) decreases from 0.73 μs\mu\mathrm{s} to 0.51 μs\mu\mathrm{s}. Similarly, TdT_{\mathrm{d}} decreases from 0.91 μs\mu\mathrm{s} to 0.49 μs\mu\mathrm{s} with the thicknesses of hole transport layer (HTL) increasing (Figure S1). Figure 1f and Figure S1 show the decay time for different electron transport layer (ETL) and QD thicknesses. The decay time decreases with increasing ETL or QD thickness. These trends were also observed in OLED and GaN-LED devices (Figure S2, Figure S3). In all, the delay and decay time are not only related with material properties, but also correlated with external circuit, low-level voltage and device structure. Based on these results, both of the carrier dynamics (microscopic) and equivalent circuit (macroscopic) can vary the TREL behaviors.

Delay stage and decay time of TREL curves for QLEDs

Figure S1. (a) Delay stage of TREL curves for QLEDs with different HTL thicknesses. (b) Decay time of TREL curves versus balance current for QLEDs with different QD thicknesses.

TREL curves and delay times of OLED devices

Figure S2. (a) TREL curves of OLED devices under different external resistances. (b) Decay time of TREL curves versus balance current for OLEDs with different external resistances. (c) TREL curves in delay stage for OLED devices at varying VLV_{\mathrm{L}} voltages. (d) Delay time versus VLV_{\mathrm{L}} for OLED devices.

TREL curves and decay and delay times for GaN LED devices

Figure S3. (a) TREL curves of GaN LED devices under different external resistances. (b) Decay time of TREL curves versus balance current for GaN LEDs with different external resistances. (c) TREL curves in delay stage for GaN LED at varying VLV_{\mathrm{L}} voltages. (d) Delay time versus VLV_{\mathrm{L}} for GaN LED devices.

To gain a comprehensive physical picture, we propose physical framework of Voltage-Current-Emission (VCE) to describe the evolution from voltage input to current injection and ultimately light emission. As shown in Figure 2, it considers time-resolved injection current by introducing a diode into RC equivalent circuit model, which connects the equivalent circuit model and carrier dynamics.

Schematic of the Voltage-Current-Emission (VCE) framework

Figure 2. Schematic of the Voltage-Current-Emission (VCE) framework. a) Device structure and pulse voltage input. b) Diode-based RC equivalent circuit. c) Simulated capacitor current (ICI_C) as a function of time. d) Simulated diode current (IRI_R) as a function of time. e) Schematical figure of carrier dynamics in the emissive layer. f) Simulated diode current and corresponding TREL curve.

For a LED device, the injection current exhibits an exponential relationship with voltage, which can be described using Shockley equation[29, 34, 35, 36]. As shown in Figure 2b, we used a diode-based RC equivalent circuit model for transient current analysis. Compared with the linear RC parallel model, the key improvement of diode-based RC model lies in its accurate characterization of the nonlinear nature of LED. The dynamics of the improved equivalent circuit system can be described by the following set of equations.

1C∫0tICdt=QC=VR(1)\frac{1}{C}\int_{0}^{t} I_C\mathrm{d}t=\frac{Q}{C}=V_R \tag*{(1)}
IR+IC=Itotal(2)I_R+I_C=I_{total} \tag*{(2)}
VR+Rs∗Itotal=V(3)V_R+R_s*I_{total}=V \tag*{(3)}
IR=Isat(eqVRnkT−1)(4)I_R=I_{sat}\left(e^{\frac{qV_R}{nkT}}-1\right) \tag*{(4)}

Here, IRI_R, ICI_C, ItotalI_{total}, VV, QQ, and VRV_R represent the PN junction current, capacitor current, total current, applied voltage, accumulated charge on the capacitor, and voltage across parallel components respectively, with IsatI_{sat} and nn denoting the saturation current and ideality factor. Since these coupled differential equations lack an analytical solution, we use numerical simulation to obtain the transient response of this equivalent circuit. Figure 2c and Figure 2d show the simulated time-dependent curves of IRI_R and ICI_C (Isat=5.52×10−6mAI_{sat}=5.52\times10^{-6}mA, n=20n=20, C=1.2nfC=1.2nf, Rs=300ΩR_s=300\Omega).

Under voltage pulse application, ICI_C instantly reaches peak values and then decays to balance values due to the capacitor charging process, while IRI_R exhibits a delayed process (defined as current delay time) and then increases to a balance value. After pulse termination, ICI_C immediately reverses to peak negative values before decaying to zero due to the capacitor discharging process, whereas IRI_R decays from its original value to zero (defined as current decay time). Owing to the PN junction features of LED device, PN junction current (IRI_R) serves as the injection current for both electrons and holes in subsequent TREL simulations.

TREL curves represent the time evolution of carrier concentration in the emissive layer, whose quantitative description is derived from carrier continuity equations, as shown in Figure 2e. TREL measurements only detect temporal dynamics, allowing us to ignore spatial terms and thus simplify the continuity equations to following form.

dn(t)dt=In(t)−krn(t)p(t)−knn(t)(5)\frac{dn(t)}{dt}=I_n(t)-k_r n(t)p(t)-k_n n(t) \tag*{(5)}
dp(t)dt=Ip(t)−krn(t)p(t)−kpp(t)(6)\frac{dp(t)}{dt}=I_p(t)-k_r n(t)p(t)-k_p p(t) \tag*{(6)}
TRELnorm=n(t)p(t)(np)max⁡(7)TREL_{\mathrm{norm}}=\frac{n(t)p(t)}{(np)_{\max}} \tag*{(7)}

Here, n(t)n(t) and p(t)p(t) represent the total carrier counts in the quantum dot layer, converted from carrier concentrations through volume of QD layer. In(t)I_n(t) and Ip(t)I_p(t) represent electron and hole injection rates (units: carriers/μ\mus), defined as In(t)=A∗IR(t)/qI_n(t)=A*I_R(t)/q and Ip(t)=B∗IR(t)/qI_p(t)=B*I_R(t)/q, where A and B (0<A,B<10<A,B<1) represent the ratios of electron and hole injection currents relative to the total diode current IR(t)I_R(t). The second order recombination term krn(t)p(t)k_r n(t)p(t) describes bimolecular recombination kinetics, where krk_r is the bimolecular recombination coefficient. First order loss terms knn(t)k_n n(t), kpp(t)k_p p(t) include deep-trap-assisted non-radiative recombination and carrier leakage effects, both exhibiting first order kinetics. It is noted that shallow defect-induced non-radiative recombination is also a bimolecular process. Thus, bimolecular recombination coefficient cannot be considered as the radiative recombination coefficient. Figure 2f presents the simulated diode current and its corresponding TREL simulation results. When the diode current is substituted into equations (5) and (7) for simulation, a time delay between diode current and TREL can be noticed. This observation can be explained to the information masking effects of normalization at the initial injection stage.

Based on VCE framework, we qualitatively illustrate the evolution of current and emission under different stages. In delay stage, capacitor charging occurs without diode current, while the diode voltages gradually increasing to turn-on voltage. In rising stage, the capacitor continues to be charged until it reaches saturated value, while the diode's current increases along with EL rising. In decay stage, the fully charged capacitor discharges, while the diode's current decays along with EL decay. In the following, we tried to explain the experimental results through numerical simulation and fitting.

Figure 3a and Figure 3b simulated the influences of capacitance and external resistance on the diode current. Either the increase of capacitance or external resistance significantly prolong the delay time of the diode injection current, which also increase the decay time of diode current. Figure 3c and Figure 3d show the simulated TREL curves. With the capacitance or external resistance increasing, both of the delay and decay time of TREL become longer. Figure 3e presents simulated TREL curves under varying VLV_{\mathrm{L}}. The increase of VLV_{\mathrm{L}} reduces the delay time. Moreover, the delay time starts to disappear when the VLV_{\mathrm{L}} approaches the turn-on voltage (Figure 3f). The good agreements between simulations (Figure 3) and experiments (Figure 1) demonstrate that the experimental TRC and TREL results can be well described using the VCE framework.

Numerical simulation results based on the diode-based RC equivalent circuit model and TREL differential equations

Figure 3. Numerical simulation results based on the diode-based RC equivalent circuit model and TREL differential equations in Figure 2. (a, b) Simulated diode current under different capacitances and series resistances. (c, d) Simulated normalized TREL curves under different capacitances and series resistances. (e) Simulated delay phase characteristics of TREL curves with varying VLV_{\mathrm{L}} voltage. (f) Simulated relationship between TREL curve delay time and VLV_{\mathrm{L}} voltage. (g) TREL curves under different injection imbalance coefficients (k). (h) Normalized TREL curves with only first-order recombination processes at various saturation current densities. (i) Normalized TREL curves incorporating second-order recombination processes at different saturation current densities.

Specifically, the delay time of TREL can be explained to the current delay of diode with turning-on and capacitor charging, while the decay time of TREL is correlated with the current decay of diode with capacitor discharging. Before the turn-on voltage, no current can be observed in the delay stage of diode current due to capacitor charging, which leads to the delay of TREL. Therefore, in the aforementioned experimental results, when reducing the thickness of the transport layer or increase the external series resistance, the increased charging time of the capacitance leads to a longer delay in the diode current, consequently increasing the TREL delay time. Similarly, as VLV_{\mathrm{L}} increases, the initial voltage division across the parallel capacitor increases, thus reduces the time required for the capacitor to charge to the turn-on voltage. Consequently, the delay in diode current is shortened, leading to a decrease in the TREL delay time (Figure 1c, Figure 3e). Notably, when VLV_{\mathrm{L}} exceeds the turn-on voltage, no charging time is required, and delay time drops to zero (Figure 1d, Figure 3f). In comparison, the decay stage of TREL is correlated with diode current decaying due to capacitor discharging.

In the following, we use numerical simulations to discuss the relationship between carrier dynamics and TREL curves, including the balance of electron and hole injection, the features of first- and second-order recombination, the quantitative relationships between TREL response time and the dynamic parameters of TRC and TREL. Mathematically, we used the following approximate function to describe the feature of the diode current curve.

Iinj={0,t<t0I0∗(1−e−t−t0τ),t0<t≤tendI0∗e−t−tendτ,t≥tend(8)I_{inj} = \begin{cases} 0, & t < t_0 \\ I_0 * \left(1-e^{-\frac{t-t_0}{\tau}}\right), & t_0 < t \le t_{end} \\ I_0 * e^{-\frac{t-t_{end}}{\tau}}, & t \ge t_{end} \end{cases} \tag*{(8)}

Here, I0I_0 is the saturated carrier injection rate (units: 1/μ\mus), τ\tau represents the diode current's rise or fall time, t0t_0 is the delay time of diode current, and tendt_{end} is the end time of the square-wave voltage signal, set to 10 μ\mus in our simulation. Specifically, I0I_0 depends on the diode's I-V characteristics, while τ\tau and t0t_0 are related with charging time and diode turn-on voltage. Because t0t_0 is only related with the time shift of IinjI_{inj} and TREL curves, we set t0t_0 to 0 for simplifying the simulation.

Overshoot is a phenomenon where the luminance initially rises, peaks at a maximum value, and subsequently decays to a steady-state level under voltage pulse application[37, 38, 39]. In the following, equation (7) was used to discuss the relationship between injection imbalance and overshoot. A coefficient kk of In/IpI_n/I_p was introduced to describe injection imbalance, where IpI_p represents for the saturated carrier injection rate of holes and k∗Ipk * I_p represents for the saturated carrier injection rate of electrons. As shown in Figure 3g, the imbalance of carrier injection (k>1k > 1) is related with the overshoot of TREL curves, and the overshoot become obviously with kk increasing. As simulated in Figs. S4a-c, the overshoot can be enhanced by increasing IpI_p, or reducing τ\tau, or increasing krk_r, or their combinations under imbalanced injection. As simulated in Figs. S5a-c, no overshoot can be observed for the devices under balance injection (k=1k = 1) regardless the variation of parameters (IpI_p, τ\tau, krk_r). Thus, the imbalance injection will lead to accumulation of electron or hole and then lead to overshoot. These results support previous conclusion that overshoot of TREL curves can be correlated with injection imbalance[37, 38].

Mathematically, the lifetime in first-order kinetics is concentration-independent, while the second-order kinetics is inversely related to the initial concentration[40]. To qualitatively understand the influence of recombination dynamics on TREL curves, we examined the kinetic characteristics of first and second order recombination processes. If second order recombination is present, both of rising time and decay time decrease with saturation injection current (I0I_0) increasing, as shown in the normalized TREL curves of Figure 3h. In contrast, normalized TREL curves remain unchanged with the variation of I0I_0 if only first-order recombination exists as shown in Figure 3i. This result proves the presence of second-order recombination in operated LED devices because both rising and decay time decrease with increasing pulse voltage, as seen in Figure S6.

TREL measurement curves of QLED devices under different applied voltages

Figure S6. TREL measurement curves of QLED devices under different applied voltages.

Response time of TREL (including rise and fall time) is a critical parameter for optical communication applications[41]. We therefore conducted numerical simulations to quantitatively analyze the correlation between above physical parameters (I0I_0, knk_n, krk_r, τ\tau) and response times. Here, we define the rise time (τrise\tau_{rise}) as the time for normalized TREL intensity to increase from 0.1 to 0.9, and the fall time (τfall\tau_{fall}) as the time for normalized intensity to decrease from 0.9 to 0.1 after voltage Turn-off. For the device under balanced electron-hole injection with equal first-order recombination coefficients (kn=kpk_n = k_p), the response time decrease with τ\tau decreasing while it decreases with I0I_0, knk_n, and krk_r increasing, as shown in Figure S7-Figure S7. Furthermore, we derived Equations (S1)-(S5) through combined numerical simulation and theoretical formulation to illustrate the relationships between response time and the combination of I0I_0, knk_n, krk_r τ\tau (see part 1 of SI for details). Based on the analysis in SI, the response time can be shortened by increasing the injection current density (I0I_0) and the recombination coefficients (krk_r and knk_n). Reducing the current time constant (τ\tau) is also effective, as it determines the speed of the current response and sets a fundamental lower bound on the device response time. In practice, τ\tau can be reduced by lowering the RC time constant of the equivalent circuit through device structure optimization, as demonstrated in recent experimental work on high-speed QLEDs[41].

Simulated results of response time at varying parameters

Figure S7. Simulated results of response time at varying (a) KnK_{\mathrm{n}}, (b) krk_{\mathrm{r}}, (c) I0I_{0}, and (d) τ\tau.

Based on VCE framework, we performed a global fitting analysis by combining TRC and TREL experimental data to extract dynamics parameters for QLED device analysis. As shown in Figure 4a, a pre-processing method was employed to fit the transient current using peak value (IpeakI_{\mathrm{peak}}) and balance value (IssI_{\mathrm{ss}}). Firstly, we calculated the equivalent parallel resistance and series resistance under different voltages using Rp(V)=(V−RsIss(V))/Iss(V)R_{\mathrm{p}}(V) = (V - R_{\mathrm{s}}I_{\mathrm{ss}}(V))/I_{\mathrm{ss}}(V) and Rs=V/IpeakR_{\mathrm{s}}=V/I_{\mathrm{peak}}. As shown in Figure S9-Figure S9, the series resistance is voltage-independent, while the parallel resistance exhibits exponential decrease with increasing voltage. The diode’s IV curve was then reconstructed using Rp(V)R_{\mathrm{p}}(V). The RsR_{\mathrm{s}} and reconstructed IV curve were used to fit the TRC data and extract the capacitance. This pre-processing method offers universal fitting capability for RC parallel circuits with various nonlinear parallel resistances including diode-based RC circuits. Figure 4b shows the fitting results of TRC data for QLED. All TRC fittings achieve R2^2 > 0.99, confirming the reliability of above fitting. Figure S9-Figure S9 displays typical circuit parameters extracted from TRC fitting, including the diode IV curve, series resistance, and capacitance. Using the fitted parameters, we simulated the diode current as a function of time. Figure 4c compares the fitted diode current curve with the measured TREL curve. The diode current exhibits similar shapes like TREL, which include delay, rise, balance, and decay stages, indicating a correlation between TREL behavior and the diode current. Notably, the fitted delay time of diode current is matched with the measured delay time of TREL curves, although most of the TREL delay times are slightly longer. This phenomenon has been explained in above simulation.

Four-panel transient current and TREL curve comparison

Figure 4. a) Transient current (TRC) curve with peak current IpeakI_{\mathrm{peak}} and steady-state current IssI_{\mathrm{ss}} annotation. b) TRC data fitting results using preprocessing method. c) Diode current from TRC fitting and experimental TREL curve comparison. d) Comparison of global fitting results and experimental TREL curves.

Rp, Rs, and C versus voltage from TRC fitting, with reconstructed diode I-V curve

Figure S9. (a) RpR_{\mathrm{p}}, (b) RsR_{\mathrm{s}}, and (c) CC versus voltage from TRC fitting, with (d) reconstructed diode II-VV curve using RpR_{\mathrm{p}}.

Subsequently, the diode current IRI_{\mathrm{R}} obtained from TRC fitting was converted into carrier injection rates for TREL curve fitting. The electron and hole injection components are defined as In(t)=A∗IR(t)/qI_{\mathrm{n}}(t) = A * I_{\mathrm{R}}(t)/q and Ip(t)=B∗IR(t)/qI_{\mathrm{p}}(t) = B * I_{\mathrm{R}}(t)/q, respectively, where AA and BB represent the ratios of electron and hole injection currents to the total diode current and serve as fitting parameters. Since TREL curves are symmetric with respect to electrons and holes, we set A>BA > B in the fitting process.

Based on the carrier dynamics equations (Equations (5)-(7)), A global fitting was performed to simultaneously describe the delay, rising, steady, and decay stages of the experimental TREL data. Figure 4d and Figure S10-Figure S10 display the fitting curves for typical TREL data, and Figure S11-Figure S11 shows the dynamic parameters extracted from TREL data of a typical QLED using aforementioned global fitting methods. The high coefficients of determination (R2=0.9971±0.0015R^2 = 0.9971 \pm0.0015) support the effectiveness of global fitting and VCE framework in describing the measured TREL and TRC curves. Furthermore, the good consistency indicates that all TREL stages are governed by a single physical mechanism. Beyond this, the global TRC-TREL fitting provides a reasonable and comprehensive extraction of carrier dynamics parameter.

Global fitting results of QLEDs’ TREL curves at (a) 6.0 V, (b) 6.8 V, (c) 7.6 V, and (d) 8.4 V

Figure S10. Global fitting results of QLEDs’ TREL curves at (a) 6.0 V, (b) 6.8 V, (c) 7.6 V, and (d) 8.4 V.

Carrier dynamics parameters versus voltage extracted from global fitting of QLEDs’ TREL curves.

Figure S11. Carrier dynamics parameters versus voltage extracted from global fitting of QLEDs’ TREL curves. (a) A/BA/B, (b) krk_{\mathrm{r}}, (c) Kn/KpK_{\mathrm{n}}/K_{\mathrm{p}}, and (d) coefficient of determination (R2R^2).

We further tried to fit OLED and GaN-LED using the VCE framework. As shown in Figure S12–Figure S13. The TREL results of OLED can be well fitted with an average R2R^2 of 0.9906 ±\pm 0.0016. For GaN-LED, we considered the barrier capacitance in the equivalent circuit due to their pronounced PN junction features[42]. Equation (9) describes the modified capacitance as the sum of geometric capacitance and barrier capacitance. With these modifications, The TRC and TREL curves of GaN-LED can be well fitted with an average R2R^2 of 0.9995 ±\pm 0.0002, as shown in Figure S14 and Figure S15. These results further demonstrate the applicability of VCE framework in quantitatively describing the carrier dynamics of electroluminescent devices.

Global fitting results of OLEDs’ TREL curves at (a) 5.0 V, (b) 5.4 V, (c) 5.8 V, and (d) 6.2 V

Figure S12. Global fitting results of OLEDs’ TREL curves at (a) 5.0 V, (b) 5.4 V, (c) 5.8 V, and (d) 6.2 V.

Carrier dynamics parameters versus voltage extracted from global fitting of OLEDs’ TREL curves. (a) $A/B$, (b) $k_{\mathrm{r}}$, (c) $K_{\mathrm{n}}/K_{\mathrm{p}}$, and (d) coefficient of determination ($R^2$)

Figure S13. Carrier dynamics parameters versus voltage extracted from global fitting of OLEDs’ TREL curves. (a) A/BA/B, (b) krk_{\mathrm{r}}, (c) Kn/KpK_{\mathrm{n}}/K_{\mathrm{p}}, and (d) coefficient of determination (R2R^2).

Global fitting results of GaN-LEDs’ TREL curves at four voltages

Figure S14. Global fitting results of GaN-LEDs’ TREL curves at (a) 5.0 V, (b) 5.4 V, (c) 5.8 V, and (d) 6.2 V.

Carrier dynamics parameters versus voltage extracted from global fitting of GaN-LEDs’ TREL curves

Figure S15. Carrier dynamics parameters versus voltage extracted from global fitting of GaN-LEDs’ TREL curves. (a) A/BA/B, (b) krk_{\mathrm{r}}, (c) Kn/KpK_{\mathrm{n}}/K_{\mathrm{p}}, and (d) coefficient of determination (R2R^2).

C(V)=Cg+Cj01−VVbi(9)C(V) = C_g + \frac{C_{j0}}{\sqrt{1-\frac{V}{V_{bi}}}} \tag*{(9)}

To establish a comprehensive understanding of VCE framework, we summarize the characteristic features and key factors of the TREL measurement based on the aforementioned analysis.

For the delay stage, the delaying originates from the initial capacitor charging process. During this process, the injection current of diode is nearly zero. EL signals cannot be observed until the voltage on the diode reach to a threshold. The delay time is mainly related with the diode turn-on voltage, equivalent circuit charging time, and normalization processing, while turn-on voltage is mainly related with the diode characteristics of the device and charging time are correlated with dielectric constant, conductivity, and thickness of functional layers. Therefore, the mobility of functional layer is not directly related with the delay time of LED, but it can affect the charging time and the LED’s turn-on voltage of LED due to its strong relationship with conductivity. The correlations between mobility and delay time are complex to be solved. The above conclusions are conflict with calculation of mobility using the delay time of TREL measurements[30, 31, 32]. For the rising stage, the rising process is correlated with carrier recombination dynamics and the features of transient injection current. For the overshoot in rising stage, numerical simulation illustrates its correlation with electron-hole injection imbalance. For the decay stage, diode current continues to discharge the capacitor at the earlier stage of decay stage. Because of the second order recombination, the decay time decreases with increasing current of balance stage. Therefore, the decay stage of TREL is mainly determined by the decay of diode current and the carrier recombination dynamics.

Based on the principle that voltage drives current, current drives emission of LED operation, we develop the VCE framework for TREL and TRC analysis by combining the diode-based RC equivalent circuit model and the carrier dynamics of electroluminescence processes. Furthermore, it enables global fitting of TREL curves, where a single set of physical parameters can quantitatively fit the complete evolution of the delay, rising, steady, and decay stages with a coefficient of determination exceeding 0.99 across QLED, OLED, and GaN-LED devices. This framework can not only explain the abnormal experimental phenomena that contradict conventional understanding, but also is universally applicable across other LEDs with different materials and device structures. Importantly, this work provides a fundamental physical tool for understanding electroluminescence transient behaviors in LEDs. It offers design guideline for developing high speed LEDs for AI technology, especially optical communication, optical computing.

Materials and Methods

Experimental details concerning the device/material fabrication, TREL characterization, and data processing are provided in the Supporting Information.

Acknowledgements

This work is granted by the National Natural Science Foundation of China (52525309 and U23A20683). We thank the MIIT Key Laboratory for Low-Dimensional Quantum Structure and Devices, Experimental Center of Advanced Materials of Beijing Institute of Technology, for technical assistance.

*Correspondence to Haizheng Zhong

E-mail: [email protected]

Conflict of interest

The authors declare no competing interests.

Contributions

H.Z. conceived the project and provided guidance. H.B. conceived the VCE framework, performed the experiments, analyzed the data, and developed the numerical fitting code. Y.W. fabricated the OLED devices and conducted the JVL measurements. Y.G. and J.Y. fabricated the GaN-LED devices and conducted the JVL measurements. H.B. and H.Z. wrote the manuscript with input from all authors.

Support Information for Voltage-Current-Emission Framework of Time-Resolved Electroluminescence

Hui Bao; Longjia Wu; Yanan Guo; Jianchang Yan; Ying Wang; Haizheng Zhong

Materials and methods

Material and Device fabrication

QLEDs were fabricated on ITO glass substrates with a sheet resistance of ∼20 Ω sq−1\sim20\ \Omega\ \mathrm{sq}^{-1}. The substrates were cleaned with deionized water, acetone and isopropanol, consecutively, for 15 min each, and then treated for 15 min with ozone generated by ultraviolet light in air. These substrates were spin-coated with PEDOT:PSS (AI 4803) and baked at 150 °C for 15 min in air. The coated substrates were then transferred to a N2\mathrm{N}_2-filled glove box for spin-coating of the TFB, CdSe-ZnS QD and ZnO nanoparticle layers. The TFB layer was spin-coated using 1.5 wt% in chlorobenzene (2,000 rpm. for 30 s), followed by baking at 110 °C for 30 min. This was followed by spin-coating of CdSe–ZnS QDs (20 mg ml−1^{-1}, toluene) and ZnO nanoparticles (30 mg ml−1^{-1}, ethanol) layers followed by baking at 145 °C for 30 min. The spin concentration of TFB varied from 4 mg mL−1^{-1} to 16 mg mL−1^{-1} to control TFB thickness. The spin concentration of QDs varied from 10 mg mL−1^{-1} to 30 mg mL−1^{-1} to control QD thickness. The spin concentration of ZnO varied from 15 mg mL−1^{-1} to 50 mg mL−1^{-1} to control ZnO thickness. These multilayer samples were then loaded into a custom high vacuum deposition chamber (background pressure, ∼3×10−7\sim3 \times10^{-7} torr) to deposit the top Al cathode (100 nm thick) patterned by an in situ shadow mask to form an active device area of 4 mm2^2.

For GaN-LED, the blue LED wafer was grown on c-plane sapphire substrate by metal-organic chemical vapor deposition (MOCVD). The epilayer structure comprises a 2-μm-thick undoped GaN template, a 1-μm-thick Si-doped n-type GaN layer, 10 periods of In0.12_{0.12}Ga0.88_{0.88}N/GaN (3 nm/10 nm) multiple quantum wells (MQWs), a 30-nm-thick Mg-doped Al0.15_{0.15}Ga0.85_{0.85}N electron blocking layer and a 150-nm-thick Mg-doped p-GaN contact layer. After the epitaxy, standard chip fabrication process in a face-up configuration was performed. Indium tin oxide (ITO) was deposited on the wafer to serve as the transparent conductive layer. The chip mesa structures were then defined via inductively coupled plasma (ICP) etching to expose the n-GaN layer. Subsequently, Cr/Pt/Au metal stacks were deposited by e-beam evaporation to form the n-type and p-type electrodes, respectively. The wafer was then back grinded and thinned to a thickness of 150 μm, and individual chips with an area of 45 × 45 mil2^2 were separated using the nanosecond (ns) pulse laser scribing. The LED chips were further mounted onto lead frames and wire-bonded with gold wires. The electroluminescence (EL) spectra and light output power (LOP) of the uncapsulated LEDs were measured at room temperature in a calibrated integrating-sphere system equipped with an EVERFINE® HAAS-2000 high-accuracy array spectroradiometer.

OLEDs were fabricated on glass substrates coated with indium tin oxide (ITO) having a sheet resistance of approximately 20 Ω\Omega sq−1^{-1}. The substrates were cleaned sequentially with deionized water, acetone, and isopropanol for 15 minutes each under ultrasonic agitation. Subsequently, the substrates were treated with ozone generated by ultraviolet light in air for 15 minutes to enhance surface hydrophilicity. After cleaning, the substrates were immediately transferred into a high-vacuum deposition chamber (base pressure ∼3×10−7\sim3 \times10^{-7} torr) without exposure to air, and all organic layers as well as the aluminum cathode were thermally evaporated in the same vacuum run, with thicknesses monitored in situ by a quartz crystal microbalance. The layers were deposited in the following order. A 10nm-thick layer of HAT-CN (hexaazatriphenylenehexacarbonitrile), a 160 nm layer of TAPC (1,1-bis[(di-4-tolylamino)phenyl]cyclohexane), a 10nm layer of TCTA (4,4′,4″-tris(carbazol-9-yl)-triphenylamine), a 20 nm co-evaporated emissive layer of CBP (4,4′-bis(N-carbazolyl)-1,1′-biphenyl) doped with 4 wt % of an iridium complex, a 35 nm layer of TmPyPB (1,3,5-tri(m-pyrid-3-yl-phenyl)benzene), a 2 nm layer of Liq, and a 100 nm aluminum (Al) cathode were deposited sequentially.

Characterizations

TREL and TRC measurements were performed using a home-built setup consisting of a digital oscilloscope (DPO 7104, Tektronix, Beaverton, OR, USA), a photodetector (HAMAMATSU 012702-11, Hamamatsu Photonics, Hamamatsu, Japan), a voltage pulse generator (SDG 5162, Siglent, Shenzhen, China), a variable resistor, and BNC coaxial cables.

LED device characterization was performed at ambient temperature utilizing a J-V-L (current density-voltage-luminance) measurement system comprising: a photodetector (Thorlabs FDS1010), fiber-optic spectrometer (Avantes HSC-TEC), 6-inch integrating sphere, and source measurement unit (Keithley 2400). External quantum efficiency (EQE) values were calculated through the Lambertian emission model following established methods. Figure S16 displays the IVL characteristics of one representative device from each of the three LED types (GaN-LED, OLED, and QLED) used in this work.

IVL curves of a typical (a) QLED, (b) OLED and (c) GaN-LED

Figure S16. IVL curves of a typical (a) QLED, (b) OLED and (c) GaN-LED

Data process methods

In this paper, the data processing workflow of the global TREL fitting code is as follows. 1) Time, TRC, and TREL signals are extracted from CSV files at each bias voltage. The series resistance is calculated by extrapolating the current peak, and an interpolation function of the diode I-V curve is reconstructed from steady-state current-voltage data. This data preprocessing method can be applied to various types of nonlinear parallel resistance, not just diodes. 2) For each voltage, a fitting is carried out using an RLC transient model. The inductance in the model comes from the wires of the measurement system, while the LED device itself can still be treated as an RC circuit. A differential evolution algorithm is used to simultaneously optimize the capacitance-related parameters and the wire inductance, yielding an accurate diode current waveform. 3) The fitted diode current serves as the carrier injection term and is fed into a rate equation model for a two-stage fitting of the normalized TREL decay curve. Differential evolution first performs a global optimization, and then the trust-region reflective method refines the result and provides the covariance matrix, from which confidence intervals and correlations are computed. During the TREL fitting process, we impose a constraint to keep the carrier injection rate essentially non-decreasing as the bias voltage increases. All fitting results and intermediate variables are finally summarized and exported to Excel and image files.

Part 1: response time analysis

At first, we find that the rise time and fall time of the TREL curve exhibit the following relationship with I0I_0, knk_n, and krk_r under pulse current drive.

τrise∝1I0∗kr+(kn2)2(10)\tau_{\mathrm{rise}} \propto\frac{1}{\sqrt{I_0 * k_r + \left(\frac{k_n}{2}\right)^2}} \tag*{(10)}
τfall∝1kn∗ln⁡kn0.316+I0∗krkn0.948+I0∗kr(11)\tau_{\mathrm{fall}} \propto\frac{1}{k_n} * \ln\frac{\frac{k_n}{0.316} + \sqrt{I_0 * k_r}}{\frac{k_n}{0.948} + \sqrt{I_0 * k_r}} \tag*{(11)}

When the injection current follows the approximate form given in Equation (8), characterized by a rise/fall time constant τ\tau, the response time is not only bounded by τ\tau but also depends on the interplay between first- and second-order recombination. Through a number of numerical simulations, we derive the empirically modified expressions, defined as follows.

T1=1I0∗kr+(kn3.6)2(12)T_1 = \frac{1}{\sqrt{I_0 * k_r + \left(\frac{k_n}{3.6}\right)^2}} \tag*{(12)}
T2=1kn∗ln⁡kn1.8+I0∗krkn0.6+I0∗kr(13)T_2 = \frac{1}{k_n} * \ln\frac{\frac{k_n}{1.8} + \sqrt{I_0 * k_r}}{\frac{k_n}{0.6} + \sqrt{I_0 * k_r}} \tag*{(13)}

Then we simulated how variations in I0I_0, knk_n, krk_r affect the response time of the TREL curve at different τ\tau and generated the corresponding τrise\tau_{\mathrm{rise}}-T1T_1 and τfall\tau_{\mathrm{fall}}-T2T_2 plots. As shown in Figure S8, we fitted the functional relationships between τrise\tau_{\mathrm{rise}} and T1T_1 as well as between τfall\tau_{\mathrm{fall}} and T2T_2 using Equation (14).

Simulated results and fitting curves for TREL rise time and fall time

Figure S8. Simulated results and fitting curves using Eqs. (12)–(14) for TREL (a) rise time and (b) fall time.

y=ax+b+c∗e−dx(14)y = ax + b + c * e^{-dx} \tag*{(14)}

Here a, b, c, and d are fitting coefficients, where x is τrise\tau_{\mathrm{rise}} (or τfall\tau_{\mathrm{fall}}) and y is the corresponding rise time (or fall time). Using Equation (14), we fitted the curves in Figure S8. Equation (S5) provides a good description of the simulated data across the entire range of τ\tau values.

The parameter ranges used in the numerical simulations are as follows. 0.1<τ<1.00.1 < \tau< 1.0 (μs\mathrm{\mu s}), 1×109<I0<1×10101 \times10^9 < I_0 < 1 \times10^{10} (1μs\frac{1}{\mathrm{\mu s}}), 1×10−10<kr<1×10−81 \times10^{-10} < k_r < 1 \times10^{-8} (1μs\frac{1}{\mathrm{\mu s}}), 0.1<kn<10.1 < k_n < 1 (1μs\frac{1}{\mathrm{\mu s}}). The simulation parameter ranges were determined based on experimentally fitted values. This empirical formula approximates the relationship between physical parameters and response times in the model, serving as a reference framework for related research.

Based on the above analysis, the response time of TREL can be effectively shortened by increasing the injection current density (I0I_0), the bimolecular recombination coefficient (krk_r), and the first-order loss coefficient (knk_n), while reducing the current rise/fall time constant (τ\tau). Among these factors, τ\tau sets a fundamental lower bound on the response time—no response can be faster than the current injection itself. The effects of krk_r and knk_n are distinct: increasing krk_r accelerates radiative recombination and thus shortens both rise and fall times, while increasing kn/kpk_n/k_p accelerates carrier depletion and primarily shortens the fall time, albeit at the cost of reduced efficiency. These results provide clear guidelines for device design: to achieve faster TREL response without sacrificing efficiency, the priority should be to reduce the RC time constant (thereby reducing τ\tau) and to enhance the radiative recombination coefficient (krk_r), while minimizing trap-assisted recombination (knk_n) to maintain high efficiency.

It is worth noting that, from a purely kinetic standpoint, increasing knk_{\mathrm{n}} alone would also shorten the response time.

Part 2: figures

Numerical simulated TREL curves at K=3

Figure S4. Numerical simulated TREL curves at K=3K=3 for varying (a) I0I_0, (b) krk_{\mathrm{r}}, and (c) τ\tau.

Numerical simulated TREL curves at k=1

Figure S5. Numerical simulated TREL curves at k=1k=1 for varying (a) I0I_0, (b) krk_{\mathrm{r}}, and (c) τ\tau.

References

  1. [1]Pattison, P. M.; Tsao, J. Y.; Brainard, G. C.; Bugbee, B. LEDs for Photons, Physiology and Food. Nature 2018, 563, 493–500.DOI
  2. [2]Kang, J.; Baek, G. W.; Lee, J. Y.; Kwak, J.; Park, J.-H. Advances in Display Technology: Augmented Reality, Virtual Reality, Quantum Dot-Based Light-Emitting Diodes, and Organic Light-Emitting Diodes. J. Inf. Disp. 2024, 25, 219–234.DOI
  3. [3]Park, S.-I.; Xiong, Y.; Kim, R.-H.; Elvikis, P.; Meitl, M.; Kim, D.-H.; Wu, J.; Yoon, J.; Yu, C.-J.; Liu, Z.; Huang, Y.; Hwang, K.-C.; Ferreira, P.; Li, X.; Choquette, K.; Rogers, J. A. Printed Assemblies of Inorganic Light-Emitting Diodes for Deformable and Semitransparent Displays. Science 2009, 325, 977–981.DOI
  4. [4]Lin, C. C.; Lee, C. C. Light Emitting Diodes. Top. Appl. Phys. 2015, 129 (1), 123–150.
  5. [5]Wasisto, H. S.; Prades, J. D.; Gülink, J.; Waag, A. Beyond Solid-State Lighting: Miniaturization, Hybrid Integration, and Applications of GaN Nano- and Micro-LEDs. Appl. Phys. Rev. 2019, 6, 041315.DOI
  6. [6]Hong, G.; Gan, X.; Leonhardt, C.; Zhang, Z.; Seibert, J.; Busch, J. M.; Bräse, S. A Brief History of OLEDs—Emitter Development and Industry Milestones. Adv. Mater. 2021, 33, 2005630.DOI
  7. [7]Jang, E.; Jang, H. Review: Quantum Dot Light-Emitting Diodes. Chem. Rev. 2023, 123, 4663–4692.DOI
  8. [8]Lin, L.; Wang, J.; Hu, H.; Luo, H.; Liu, Y.; Yang, X.; Su, J.; Li, D.; Xu, Z.; Luo, C.; Yu, Y.; Guo, T.; Li, F. Nanoscale Transfer-Printed Full-Colour Ultrahigh-Resolution Quantum Dot LEDs. Nature 2026, 652, 349–358.DOI
  9. [9]Huang, Y.; Hsiang, E. L.; Deng, M. Y.; Wu, S. T. Mini-LED, Micro-LED and OLED Displays: Present Status and Future Perspectives. Light Sci. Appl. 2020, 9, 105.DOI
  10. [10]Jing, Y.; Yao, M.; Yang, M.; Li, M.; Ding, H.; Yang, G.; Zhang, R.; Han, D.; Liu, H.; Zhong, H. Photolithographic Fabrication of High-Resolution Micro-QLEDs Towards Color-Conversion Microdisplay. Light Sci. Appl. 2025, 14, 370.DOI
  11. [11]Ren, A.; Wang, H.; Zhang, W.; Wu, J.; Wang, Z.; Penty, R. V.; White, I. H. Emerging Light-Emitting Diodes for Next-Generation Data Communications. Nat. Electron. 2021, 4, 559–572.DOI
  12. [12]Yoshida, K.; Manousiadis, P. P.; Bian, R.; Chen, Z.; Murawski, C.; Gather, M. C.; Haas, H.; Turnbull, G. A.; Samuel, I. D. W. 245 MHz Bandwidth Organic Light-Emitting Diodes Used in a Gigabit Optical Wireless Data Link. Nat. Commun. 2020, 11, 1171.DOI
  13. [13]Kalinin, K. P.; Gladrow, J.; Chu, J.; Clegg, J. H.; Cletheroe, D.; Kelly, D. J.; Rahmani, B.; Ballani, H. Analog Optical Computer for AI Inference and Combinatorial Optimization. Nature 2025, 645, 354–361.DOI
  14. [14]Pommerehne, J.; Vestweber, H.; Tak, Y. H.; Bässler, H. Transient Electroluminescence from Single- and Double-Layer Light-Emitting Diodes (LEDs) Based on Polymer Blends. Synth. Met. 1996, 76, 67–70.DOI
  15. [15]Hosokawa, C.; Tokailin, H.; Higashi, H.; Kusumoto, T. Transient Behavior of Organic Thin Film Electroluminescence. Appl. Phys. Lett. 1992, 60, 1220–1222.DOI
  16. [16]Xue, X.; Zhu, B.; Kang, Z.; Chi, X.; Zhang, H.; Tang, A.; Ji, W. Transient Electroluminescence Spectroscopy in Planar Electroluminescent Devices. Laser Photonics Rev. 2024, 19, 2401166.DOI
  17. [17]Kim, J.; Hahm, D.; Bae, W. K.; Lee, H.; Kwak, J. Transient Dynamics of Charges and Excitons in Quantum Dot Light-Emitting Diodes. Small 2022, 18, 2202290.DOI
  18. [18]Olshansky, R.; Su, C.; Manning, J.; Powazinik, W. Measurement of Radiative and Nonradiative Recombination Rates in InGaAsP and AlGaAs Light Sources. IEEE J. Quantum Electron. 1984, 20, 838–854.DOI
  19. [19]Sin, Y.; Ives, N.; Presser, N.; Moss, S. C. Root Cause Investigation of Catastrophic Degradation in High Power Multi-Mode InGaAs-AlGaAs Strained Quantum Well Lasers. Proc. SPIE 2010, 7583, 758307.DOI
  20. [20]Xiao, S.; Qiao, X.; Lin, C.; Chen, L.; Guo, R.; Lu, P.; Wang, L.; Ma, D. In Situ Quantifying the Physical Parameters Determining the Efficiency of OLEDs Relying on Triplet–Triplet Annihilation Up-Conversion. Adv. Opt. Mater. 2022, 10, 2102333.DOI
  21. [21]Kang, J.; Son, J. B.; Kim, G. W.; Bae, S.; Min, K. S.; Sul, S.; Jeon, W. S.; Jang, J.; Park, G.-S.; Shin, J. K.; Kwon, J. H.; Kim, S. K. Time-Resolved Electroluminescence Study for the Effect of Charge Traps on the Luminescence Properties of Organic Light-Emitting Diodes. Phys. Status Solidi A 2020, 217, 2000081.DOI
  22. [22]Keating, L. P.; Lee, H.; Rogers, S. P.; Huang, C.; Shim, M. Charging and Charged Species in Quantum Dot Light-Emitting Diodes. Nano Lett. 2022, 22, 9500–9506.DOI
  23. [23]Wang, S.; Liu, S.; Wang, T.; Bai, J.; Peng, J.; Zhang, H.; Xie, W.; Ji, W. Operando ZnO Recrystallization for Efficient Quantum-Dot Light-Emitting Diodes. Light Sci. Appl. 2025, 14, 196.DOI
  24. [24]Yan, X.; Chen, C.; Wu, B.; Sun, F.; Bao, H.; Tian, W.; Chang, S.; Zhong, H.; Jin, S. Probing the Operation of Quantum-Dot Light-Emitting Diodes Using Electrically Pumped Transient Absorption Spectroscopy. J. Phys. Chem. Lett. 2024, 15, 8593–8599.DOI
  25. [25]Bao, H.; Chen, C.; Cao, Y.; Chang, S.; Wang, S.; Zhong, H. Quantitative Determination of Charge Accumulation and Recombination in Operational Quantum Dots Light Emitting Diodes via Time-Resolved Electroluminescence Spectroscopy. J. Phys. Chem. Lett. 2023, 14, 1777–1783.DOI
  26. [26]Guerrero, A.; Bisquert, J.; Garcia-Belmonte, G. Impedance Spectroscopy of Metal Halide Perovskite Solar Cells from the Perspective of Equivalent Circuits. Chem. Rev. 2021, 121, 14430–14484.DOI
  27. [27]Bertoluzzi, L.; Bisquert, J. Equivalent Circuit of Electrons and Holes in Thin Semiconductor Films for Photoelectrochemical Water Splitting Applications. J. Phys. Chem. Lett. 2012, 3, 2517–2522.DOI
  28. [28]Xiao, H.; Wang, K.; Wang, R.; Chen, W.; Chiang, K. S. Equivalent Circuit of Quantum-Dot LED and Acquisition of Carrier Lifetime in Active Layer. IEEE Electron Device Lett. 2020, 41, 87–90.DOI
  29. [29]Shen, H.; Gao, Q.; Zhang, Y.; Lin, Y.; Lin, Q.; Li, Z.; Chen, L.; Zeng, Z.; Li, X.; Jia, Y.; Wang, S.; Du, Z.; Li, L. S.; Zhang, Z. Visible Quantum Dot Light-Emitting Diodes with Simultaneous High Brightness and Efficiency. Nat. Photon. 2019, 13, 192–197.DOI
  30. [30]Doe, T.; Kitano, K.; Yamamoto, S.; Yamamoto, M.; Goto, K.; Sakakibara, Y.; Kobashi, T.; Yamada, H.; Ueda, M.; Ryowa, T.; et al. Evaluation of Degradation Behavior in Quantum Dot Light-Emitting Diode with Different Hole Transport Materials via Transient Electroluminescence. Appl. Phys. Lett. 2021, 118, 203503.DOI
  31. [31]Niu, Q.; Blom, P. W. M.; May, F.; Heimel, P.; Zhang, M.; Eickhoff, C.; Heinemeyer, U.; Lennartz, C.; Crăciun, N. I. Transient Electroluminescence on Pristine and Degraded Phosphorescent Blue OLEDs. J. Appl. Phys. 2017, 122, 185502.
  32. [32]Xu, M.; Peng, Q.; Zou, W.; Gu, L.; Xu, L.; Cheng, L.; He, Y.; Yang, M.; Wang, N.; Huang, W.; et al. A Transient-Electroluminescence Study on Perovskite Light-Emitting Diodes. Appl. Phys. Lett. 2019, 115, 041102.
  33. [33]Shen, Q.; Hao, Y.; Ma, L.; Wang, X. Comparative Study of Red/Green/Blue Quantum-Dot Light-Emitting Diodes by Time-Resolved Transient Electroluminescence. J. Phys. Chem. Lett. 2021, 12, 7019–7025.DOI
  34. [34]Bao, H.; Sattari-Esfahlan, S. M.; Zhong, H. A Nanocrystal-Based PN Junction Model for Quantum Dot Light-Emitting Diodes. Light Sci. Appl. 2026, 15, 322.DOI
  35. [35]Zhu, B.; Wang, T.; Wang, S.; Chi, X.; Zhang, H.; Zhang, H.; Ji, W. Influence of Shells on the Charge Tunneling Behavior in Quantum-Dot Light-Emitting Diodes. J. Phys. D: Appl. Phys. 2024, 58, 075105.DOI
  36. [36]Arya, S.; Jiang, Y.; Jung, B. K.; Tang, Y.; Ng, T. N.; Oh, S. J.; Nomura, K.; Lo, Y.-H. Understanding Colloidal Quantum Dot Device Characteristics with a Physical Model. Nano Lett. 2023, 23, 9943–9952.DOI
  37. [37]Yu, R.; Yin, F.; Pu, C.; Zhou, D.; Ji, W. On the Electroluminescence Overshoot of Quantum-Dot Light-Emitting Diodes. Opt. Lett. 2023, 48, 3059–3062.
  38. [38]Hassine, L.; Bouchriha, H.; Roussel, J.; Fave, J.-L. Transient Response of a Bilayer Organic Electroluminescent Diode: Experimental and Theoretical Study of Electroluminescence Onset. Appl. Phys. Lett. 2001, 78, 1053–1055.DOI
  39. [39]Chen, J.; Zhao, X.; Tang, X.; Ning, Y.; Wu, F.; Chen, X.; Zhu, H.; Xiong, Z. An Unprecedented Spike of the Electroluminescence Turn-On Transience from Guest-Doped OLEDs with Strong Electron-Donating Abilities of Host Carbazole Groups. Mater. Horiz. 2021, 8, 2785–2796.DOI
  40. [40]Atkins, P.; de Paula, J. Physical Chemistry, 11th ed.; Oxford University Press, 2018.
  41. [41]Zhou, T.; Tian, F.; Chen, S. Nanosecond-Pulsed Electroluminescence from High Current–Driven Quantum-Dot Light-Emitting Diodes. Sci. Adv. 2025, 11, eads1388.DOI
  42. [42]Piprek, J., Ed. Nitride Semiconductor Devices: Principles and Simulation; Wiley-VCH, 2007.

Paper details

vlnc:W_7zfxmb9a
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents