Predicting LED parameters from electroluminescent semiconductor wafer testing

摘要

A diode model and conductive-probe measurements taken at the wafer lever are used to predict the characterization parameters of a semiconductor device manufactured from the wafer. A current-voltage curve (I-V) model that expresses a current-voltage relationship as a function of resistance, ideality factor, and reverse saturation current is fitted to a number of conductive-probe measurement data. The current-voltage curve (I-Vd) for the device is then estimated by subtracting from the (I-V) model the product of current times the resistance produced by fitting the (I-V) model.

主权项

1. A method of characterizing a semiconductor device from conductive-probe measurements performed on a semiconductor wafer from which the device is manufactured, said wafer including a p-type semiconductor layer and an n-type semiconductor layer defining a p-n junction, the method comprising the following steps:
applying electrical stimuli to the wafer through a probe in direct contact with one of said semiconductor layers to produce conductive-probe measurement data; fitting a current-voltage curve (I-V) model to said conductive-probe measurement data, said model expressing a current-voltage relationship as a function of resistance; and estimating a current-voltage curve (I-Vd) for the device by removing from the (I-V) model resistance effects calculated from data generated by the fitting step; wherein the resistance is assumed variable with current and the resistance is calculated from measurement data points from the equation:R=Δ⁢⁢VΔ⁢⁢Iwhere R is resistance, V is voltage, and I is current;
wherein the device's forward voltage, ideality factor, resistance and reverse saturation current are calculated from the following set of equations:Ri=αIiβ+CI⁡(i)=I0⁢exp⁡(q⁡(Vi-Ii⁢Ri)nkT)Vd⁡(i)=Ii⁢Ri+kTq⁢n⁢⁢ln⁡(IiI0+1)where Vd is the device's forward voltage, I0 is reverse saturation current, n is ideality factor, and α, β, and C are constant parameters estimated by fitting N current-voltage measurement data points, (Ii,Vi), i=1K N, with a multi-parameter minimization calculation; and
wherein said multi-parameter minimization calculation is carried out with the following estimator:Σ⁡(α,β,C,n,I0)=∑i=1N⁢[wi⁡(Ii-IMi)2+1wi⁢(Vi-VMi)2]where Mi refers to said measurement data points, and wi is an empirical weight factor that accounts for the increased sensitivity of the current-voltage curve for the device to current at high current values and to voltage at low current values.