Correcting Overestimates in TLM-Based Contact Resistivity Measurements

The transfer length method (TLM) is the most common method for finding the resistance associated with contacts in semiconductor devices. In TLM, resistance measurements are made between neighbouring contact pads that are part of a test structure and a quantity called the contact resistivity, $\rho c$ is extracted. $\rho c$ contains resistance contributions from carrier transport in three different regions of the test structure: 1) in the electrode layer of the contact pads (although this is normally small and can be neglected); 2) in the films and interfaces between the electrode layer and the semiconductor substrate; and 3) in the semiconductor substrate directly beneath the contact pads (this carrier transport is mostly lateral and is needed for carriers to get across from one contact of the test structure to the neighbouring contact). Despite the fact that pc is a lumped resistance containing components from three different regions of the test structure, it is often used as if it is purely an interfacial resistivity, particularly in the characterisation of full-area solar cell contacts. Here, we show that this leads to an overestimate of the interfacial contact resistivity. Since the overestimate is consistent it is easily corrected going forward. The correction will lead to more accurate resistance loss accounting in semiconductor devices including LEDs, photodetectors, and solar cells, and will therefore help to minimise resistance losses and increase the efficiency of these devices.