Access Official Technical PDF

Share Summary:LinkedInEmail

Accurately predicting current is essential when evaluating a transistor for integration into a sensor circuit. For a GaN Schottky barrier MOSFET, the drain current depends not only on gate and drain bias but also on how electrons cross the source-side metal-semiconductor barrier. A mathematical model of accumulation-mode operation combines thermionic emission and tunneling to describe these contributions and helps explain how device structure influences current in GaN-based UV sensor designs.

Why model a GaN Schottky barrier MOSFET?

Gallium nitride (GaN) is used in electronic devices because of material properties that include a wide bandgap, high breakdown voltage, high carrier mobility, and high velocity saturation. A GaN Schottky barrier MOSFET also offers low off-state current, associated with a high hole barrier that inhibits leakage.

One proposed application is an active-pixel-sensor UV imager that combines a GaN transistor with a Schottky-electrode GaN UV photodiode. Integrating the sensing and transistor functions can support a more compact sensor architecture than one built around a silicon circuit and a separate GaN metal-semiconductor-metal UV sensor. For design and evaluation, a useful current model must represent the transport mechanisms at the Schottky contact as operating conditions change.

Two current paths at the Schottky contact

The model describes source-junction current as the sum of two mechanisms: thermionic emission and tunneling. In thermionic emission, electrons have sufficient energy to pass over the effective Schottky barrier and enter the semiconductor. The model accounts for the transmission of these electrons and for image-force lowering, which reduces the effective barrier under an electric field.

Tunneling provides a second path through the barrier. Its contribution depends strongly on barrier width: a wider barrier suppresses tunneling, while a narrower barrier makes tunneling more probable. The analysis approximates the conduction-band profile as triangular to represent the barrier for tunneling electrons.

These mechanisms respond differently to device bias. Near equilibrium, the Schottky barrier is relatively thick and tunneling is suppressed. Applying a positive gate voltage increases electron concentration in the channel. The associated change in the energy-band profile narrows the barrier; with drain bias applied, electrons can then tunnel through it. The model therefore links channel charge to the barrier width rather than treating tunneling as independent of gate operation.

How channel charge shapes the model

In accumulation mode, gate voltage controls the concentration of electrons near the semiconductor surface. The model uses this channel electron concentration to estimate the depletion-region concentration relevant to tunneling. This substitution matters because, in the accumulation region, the channel charge is the key factor determining the changing Schottky barrier width.

The analysis also relates gate voltage to channel charge through the MOS structure, including substrate charge and the voltage across the gate oxide. For the n-type GaN accumulation case considered, the substrate-related contribution dominates, allowing the channel-charge relationship to be simplified. The resulting concentration estimate is then used to calculate barrier width and tunneling current, alongside the thermionic component.

At higher drain bias, the model considers pinch-off. When the drain voltage exceeds the pinch-off condition, the electron supply at the drain end of the channel becomes insufficient to sustain a continuing increase in current. Including this behavior allows the model to represent the transition toward current saturation rather than treating current as increasing indefinitely with drain bias.

What the model indicates about device parameters

The calculations for an n-type GaN device with indium tin oxide (ITO) source and drain electrodes used a Schottky barrier height of 0.6 eV and a 30 nm SiO2 gate oxide. The calculated threshold voltage was 3.5 V, compared with a measured value of 3.75 V. The calculated drain current was 1.2 times the measured value. These results indicate a close threshold-voltage estimate, while also showing a difference between calculated and measured current that should be considered when using the model for device evaluation.

The parameter analysis highlights different levels of sensitivity:

  • Schottky barrier height: Increasing the barrier height reduces drain current and raises threshold voltage.
  • Gate oxide thickness: Reducing oxide thickness increases channel electron concentration in the modeled device, narrows the Schottky barrier, and increases tunneling current. The calculations therefore show higher drain current with a thinner oxide.
  • Doping concentration: Drain current changes little with doping concentration in the on-state accumulation region. Doping affects barrier width and can influence off-state current, but channel electron concentration is more important to barrier width once the device is on.

This distinction is useful for interpreting design trade-offs: changing a parameter can have a different effect before and after the transistor enters accumulation. The findings apply to the modeled accumulation-mode n-type GaN structure and should not be read as a general statement about every GaN transistor or operating regime.

Practical implications for UV sensor design

For engineers assessing GaN transistors in integrated UV sensors, the model provides a way to connect gate bias and device structure with the current mechanisms at the source contact. It can help explain why oxide thickness and Schottky barrier height are influential, why doping has a comparatively small effect on on-state current in the modeled case, and why pinch-off must be included when predicting output behavior.

The calculated and measured results are not identical, so the model is best treated as an analytical tool for understanding trends and estimating device behavior—not as a substitute for validating a specific design against measurements. Its practical value lies in making the relationship between channel charge, barrier width, and current explicit during evaluation of accumulation-mode GaN Schottky barrier MOSFETs.

Conclusion

Accumulation-mode current in a GaN Schottky barrier MOSFET reflects both electron transport over the Schottky barrier and tunneling through it. Accounting for image-force lowering, channel-dependent barrier width, and pinch-off gives a more complete picture of how the device responds to bias. For the modeled UV-sensor-related structure, oxide thickness and barrier height strongly affect drain current, while doping concentration has little on-state influence. These relationships can guide device analysis, with measured validation remaining important for design decisions.

Access Official Technical PDF

About The Author