Table of Contents
Advertisement
enabling
innovative electronics
Data analysis: Data analysis options
Computation Method specifies the algorithm used to calculate the mobility.
Important: Before computing the mobility, the drain current is 'smoothed' through recursive
application of the Savitzky-Golay filter on the raw (measured) current data. Any current-based
figures of merit (minimum and OFF current, ON current etc) are then calculated using the filtered
data. Consequently, the value of the reported drain current figures of merit may not correspond to
the values read directly from the raw data.
Important: The user is strongly recommended to familiarise with the caveats associated with each
mobility computation algorithm, see
The three mobility algorithms currently available with SuperFACT are:
1. Partition Method. The experimental data I
of M data points. M can be any number ranging from 5 to 10 as specified by the user on the
Data Analysis: Advanced Settings panel of the Advanced SMU and Acquisition Settings
Advanced UI. For each data subset, the linear fit is calculated and the slope of this fit
inserted in Eqs. 4 and 9 to compute the mobility. Therefore, for each subset, the partition
method will provide a gate voltage–dependent mobility μ
gate voltage of the subset i.
The x-axis intercept
The partition method will return any μ
the user on the Data Analysis: Advanced Settings section of the Advanced SMU and
Acquisition Settings Advanced UI. The maximum mobility μ
flagged as the reference mobility of the DUT.
Tip: Partition Method can be especially useful to calculate gate-voltage dependent mobility
in presence of noise/poor quality experimental data.
2. Derivative Method. The mobility is computed by differentiating the drain current I
respect to the gate voltage V
20
the Cartesian x-axis is the voltage axis
Ossila Ltd
Mobility
Computation.
DS
20
of the linear fits gives the gate dependent threshold voltage, V
complying with the R-squared condition specified by
i
, i.e.
GS
Copyright © 2009-2015
(V
) are grouped (partitioned) in subsets made
GS
(
), where
i
GS,i
(
) satisfying this condition is
i
GS,i
is the median
GS,i
Th,i.
with
DS
68
Advertisement
Table of Contents
