Access Official Technical PDF

Share Summary:LinkedInEmail

WSe2 electroluminescence offers a way to investigate light emission from atomically thin semiconductors, but stable operation can be difficult when electrical current changes over time. In a 2019 study of monolayer tungsten diselenide (WSe2) p-n diodes, researchers identified current decay as a central obstacle to consistent emission in ambient conditions. Their solution—pulsed voltage bias—maintained light output while retaining a relatively simple, back-gated transistor structure.

The findings are relevant to engineers and technical decision-makers following emerging optoelectronic devices. They illustrate how electrical behavior at the material and device level can determine whether an emitter remains stable, not just how efficiently it produces light at a particular moment.

Why stable WSe2 electroluminescence is challenging

Monolayer transition metal dichalcogenides (TMDCs) are atomically thin semiconductors with direct band gaps. WSe2 is also ambipolar: depending on the applied gate and contact voltages, it can conduct through electrons or holes. That property makes it suitable for a p-n diode, where electrons and holes are injected into the same channel and recombine to produce light.

In this device, electrostatic gating establishes the p-type and n-type regions rather than relying on chemical doping. However, TMDC field-effect transistors can exhibit hysteresis: their electrical response depends on the device’s voltage history. At a fixed gate bias, this behavior can appear as a decline in current over time. The study links that decline to reduced electroluminescence, because the light output tracked the falling current.

In ambient air, adsorbed molecules—particularly water—can contribute to hysteresis and current instability. The supplementary measurements also indicate that hole current can remain unstable even after adsorbates are reduced, pointing to additional traps, such as interface traps or material defects. The result is a practical device-level issue: a diode may emit under continuous bias, yet its current and brightness can diminish within seconds.

How pulsed voltage bias stabilizes emission

Rather than holding the device at a constant voltage, the researchers applied voltage in repeated on-and-off cycles. During the on period, the two contacts were biased relative to the gate to inject electrons and holes. During the off period, the contact voltages returned to the gate voltage. Light was produced during the on period, when both carrier types were present.

Under direct-current bias, emission and current fell by orders of magnitude within a few seconds. With pulsed bias, both remained substantially more stable over the measurement period. Most tested devices showed no decrease or only a small initial reduction in emission, followed by stable output. The longest reported stability test extended beyond three hours.

The pulse timing matters. In measurements using a 50% duty cycle, emission was relatively stable above approximately 1 kHz. At lower frequencies, current decay caused the light output to decline more quickly. The study also notes that stability depends on duty cycle, so frequency alone does not determine performance. At low frequencies, a sufficiently long off period could help the device recover between pulses.

Balancing electron and hole injection

Stable operation is only part of diode performance. Efficient light generation also depends on the relative amounts of electrons and holes reaching the channel. If one carrier type dominates, those carriers can travel across the channel without recombining efficiently to form excitons—the bound electron-hole pairs associated with the observed light emission.

The researchers varied the gate voltage while keeping the voltage difference between the p and n contacts constant. The brightest emission occurred near the condition where both carrier types were injected. The highest internal quantum efficiency (IQE) occurred at a different point in the sweep, illustrating why brightness and efficiency are not interchangeable measures. IQE compares emitted photons with injected carriers, while measured brightness also depends on operating conditions and collection.

At the optimized bias condition, the reported electroluminescence IQE reached about 1% in the study’s summary, with a peak near 1.3% in the detailed measurements. This was close to the photoluminescence efficiency, which was approximately 1.6%. The comparison suggests that, in these devices, converting injected carriers into excitons was not substantially less efficient than producing excitons through optical excitation. At equal p- and n-contact bias, electroluminescence efficiency was lower.

Nanosecond-scale modulation in a lateral diode

The device also demonstrated fast changes in light output in response to electrical pulses. Time-resolved measurements found a rise time of about 12 nanoseconds and a fall time of about 18 nanoseconds. The emission occurred when the p- and n-side voltage pulses overlapped, supporting the interpretation that light resulted from bipolar carrier injection.

The authors attributed the fast response, relative to the vertical tunnel-injection devices discussed in their paper, likely to the lower capacitance of the lateral injection arrangement. The measured response establishes a device-level modulation result; it does not, by itself, demonstrate readiness for a particular industrial system or application.

Practical implications for optoelectronic device design

For engineers evaluating emerging semiconductor emitters, the study highlights several design considerations:

  • Assess time-dependent current: A stable initial output does not guarantee stable operation under sustained bias.
  • Consider pulsed drive: Pulse frequency, on/off timing, and duty cycle can affect current recovery and emission stability.
  • Optimize carrier balance: Gate and contact biases influence whether electrons and holes recombine in the channel or one carrier type dominates.
  • Separate brightness from efficiency: Output intensity and IQE provide different information about device behavior.

The demonstrated structure used a back-gated WSe2 transistor without dual gates, a heterostructure, contact optimization, or chemical doping. The authors identify contact improvements and better material quality as possible routes to higher efficiency and lower-voltage operation, while optical-cavity coupling could support higher speed.

Conclusion

Pulsed injection addresses a key limitation in WSe2 electroluminescence: current decay that also reduces light output under continuous bias. In the studied monolayer p-n diodes, pulsing supported stable ambient-condition emission, efficiency close to photoluminescence, and nanosecond-scale modulation. The practical lesson is that drive strategy and carrier balance are integral to optoelectronic device performance. Further work on contacts and material quality would be needed to advance the device characteristics described in the study.

Access Official Technical PDF

About The Author