Title: NREL CdTe Team Meeting May 56, 2005
1INFLUENCE OF BAND PROFILES ON TRANSPORT AND
RECOMBINATION
Alan Fahrenbruch Colorado State University, Dept.
of Physics, Fort Collins, CO 83052
The author acknowledges the support by NREL
through Colorado State University, and the
encouragement of Jim Sites and the CdTe Team
members. Modeling results obtained using AMPS-1D
(Version 1,0,0,1), written by S. Fonash,
Pennsylvania State Univ., with EPRI support.
NREL CdTe Team Meeting May 5-6, 2005
22
3High net space charge (Na - Nd) gt thin
depletion layer
Similar to CIGS cells High relative Voc, but
requires long Ld
3
4Low or negative net space charge (Na - Nd) gt
depletion layer fills CdTe layer
Change Nd and sn so that Ndsn const.
Øbc subtracts directly from Voc Collection field
depends directly on bias gt low J at Pmax,
low ff
4
5Intermediate case
C-V meas. gt 3 - 4 µm cells near full depletion
5
6DEPLETION LAYERS
Charge neutrality no longer holds Carrier
densities are very different than net space
charge density Band profiles are sensitive to
small differences in space charge density Band
profiles are likely to change with illumination
6
7At forward bias, the band shapes are influenced
by back-contact barrier height
Standard diode equations are not a good
representation, more complex equations are needed
7
8CdTe is highly compensated -- The net space
charge density is the difference of 2 or more
large numbers. Large densities of Cl, Cu, native
defects, and all their combinations determine the
electronic properties. The combination of the
properties above make it almost impossible to
predict the band profile
8
9Defect levels for CdTe are too complex to be
useful for simulations
9
10A possible simplification is to group them by
energy level
10
11Defects in relation to the band profiles
11
12Effect of illumination at zero bias
12
13Effect of bias at Vmax
V 0, LIGHT
13
14Light and dark J-V curves are much different
Label curves
14
15SCAPS simulation of C-V data
15
16Simulated QE curves
16
17Sensitivity plot tests the effect of model
parameters
Fairly large densities of compensating pairs of
shallow defects, Na Nd 1017, don't
significantly change result
17
18IN CONCLUSION
Its important to look at the combined effects of
defects in the depletion layer of a model
cell Defect grouping may be useful in reducing
complexity Strong compensation in CdTe may
actually simplify the problem
18