Schottky barrier inhomogeneity can complicate the electrical characterization of metal-semiconductor devices, including SnO2 nanowire structures. When a junction is treated as having one uniform barrier height, measured current-voltage behavior may not match the model—particularly across a broad temperature range. Analysis of Au-Ni/SnO2/Au-Ni devices shows how a distribution of local barrier heights can help explain changes in both apparent barrier height and ideality factor.
For engineers and technical teams evaluating semiconductor components, the practical lesson is methodological: electrical parameters extracted from I-V curves depend on the assumptions used to interpret the junction. A non-ideal result does not, by itself, identify a single cause. Considering spatial variation at the contact interface can provide a more complete picture.
Why a uniform Schottky barrier may be an incomplete model
A Schottky contact forms where a metal meets a semiconductor. In the idealized thermionic-emission picture, carriers with sufficient energy pass over a well-defined energy barrier. The Schottky barrier height (SBH) describes that energy barrier, while the ideality factor indicates how closely the measured diode behavior follows the ideal model. An ideality factor near one is expected in the simplest case; higher values indicate that the observed behavior departs from it.
Several effects may contribute to non-ideal behavior, including interface states, tunneling, image-charge effects, and recombination or generation in the space-charge region. The study of SnO2 nanowire devices also considers another possibility: the barrier height may vary across the metal-semiconductor interface. In that case, the contact is better represented as a collection of regions with different local barriers than as one perfectly uniform junction.
This distinction matters because current does not necessarily reflect the average barrier in a straightforward way. Regions with different barriers contribute differently to carrier transport. As temperature changes, the balance of those contributions can change, making the barrier height and ideality factor extracted from I-V measurements appear temperature-dependent.
What temperature-dependent I-V measurements revealed
The devices used Au-Ni alloy contacts on SnO2 nanostructures, forming two nominally identical metal-semiconductor interfaces. Current-voltage characteristics were measured over temperatures from 10 to 300 K. The researchers observed symmetric behavior in the two voltage polarities, consistent with similar barriers at the two contacts.
However, the extracted parameters did not remain constant with temperature. The apparent Schottky barrier height increased as temperature rose, while the ideality factor decreased. This trend is not predicted by the simplest Schottky-Mott treatment, which assumes a uniform barrier and temperature-independent junction parameters.
Richardson plots—often used to estimate barrier height from the temperature dependence of saturation current—also showed curvature rather than a single straight-line trend. That curvature indicates that a single activation energy does not adequately describe the measurements across the tested range. For device characterization, this is a signal to question whether the uniform-barrier assumption is appropriate before relying on one extracted value.
Modeling Schottky barrier inhomogeneity
To represent local variation, the analysis treated the barrier heights as a Gaussian distribution. The distribution is described by a mean barrier height and a standard deviation, which indicates how widely local values vary around the mean. Rather than assigning one barrier to the entire contact, this approach accounts for contributions from a range of barriers.
At 300 K, 130 measurements were used to examine the distribution. The measured values were fitted with a Gaussian profile, yielding a mean barrier height of 0.30 eV and a standard deviation of 60 meV. These results support the interpretation that spatial variation in the junction barrier contributes to the observed electrical behavior.
Within the Gaussian-distribution framework, the apparent barrier height and ideality factor can vary with temperature and applied voltage. This provides a way to interpret both the changing I-V parameters and the curvature in the conventional Richardson plot. The model does not mean that every non-ideal contact has the same cause; rather, it offers a specific explanation consistent with the measurements in these SnO2 devices.
Implications for semiconductor device characterization
When evaluating a metal-semiconductor junction, engineers and researchers should interpret extracted parameters in the context of the model used. If barrier height or ideality factor changes with temperature, or if a Richardson plot is curved, a single uniform-barrier interpretation may be insufficient. Assessing a distribution of barrier heights can help distinguish interface-level variation from behavior predicted by ideal thermionic emission alone.
The analysis also reported a modified Richardson constant of 70 A cm-2 K-2, corresponding to an estimated electron effective mass of 0.58 times the free-electron mass. These are model-derived results for the studied devices, not universal values for SnO2 contacts. Their relevance depends on the device structure and the assumptions used to extract them.
For practical measurement workflows, the findings reinforce the value of collecting I-V data across multiple temperatures and checking whether standard plots support a consistent activation energy. Such checks can reveal when additional junction physics may need to be considered. They also help prevent temperature-dependent apparent parameters from being mistaken for a direct change in a single, uniform barrier.
Conclusion
In Au-Ni/SnO2 nanowire devices, temperature-dependent barrier height and ideality factor were interpreted through spatial Schottky barrier inhomogeneity. A Gaussian distribution of barrier heights provided a framework for explaining the measured trends and the curvature in conventional Richardson analysis. For engineers working with semiconductor contacts, the central takeaway is to treat extracted I-V parameters as model-dependent and to consider interface variation when ideal-junction assumptions do not fit the temperature-dependent data.