Characterizing Electronic States that Control the Performance of CdTe Solar Cells

1
Characterizing Electronic States that Control the
Performance of CdTe Solar Cells

• Fred H. Seymour
• December 7, 2005

2
Contents

• Introduction and research goals
• Selected research results
• Admittance spectroscopy (AS) technique
• Computer simulations of AS
• Identification of deep electronic states
  
• Preliminary defect correlation to performance
• Conclusions

3
Solar Energy

• Earth irradiated with 125,000 TW of solar energy
• 1 TW 1012 W
• 20 MW per person
• Average total global industrial energy
  consumption 13 TW
• 2 kW per person

4
MW

• Exponential Growth
• Elastic demand for PV
• Decreasing costs
• Economies of scale
• Increase energy conversion efficiency

5
Colorado School of Mines

• PV program since early 1990s
• Cell characterization, CdTe cell growth
• 15 Ph.D.s/Collaborations with NREL
• PV ties into CSM 10 yr plan for all 4 focus areas
• Development of Earths resources
• Energy
• Advanced Materials
• Environment
• Training future engineers

6
Polycrystalline CdTe thin film solar cells

• Low cost among simplest to manufacture
• Successful commercial production, First Solar
  scaling up to 75 MW/year capacity
• Production conversion efficiencies of 10 leave
  plenty of room for improvement
• Must better understand electronic structures to
  optimize cell performance

7
CdTe Solar Cell Structure
9) Back contact (Cathode)
8) Anneal with Copper
7) Etch with NP or BrM
6) Anneal with CdCl2
5) CdTe absorber (3 to 8 µm)
4) CdS window ( 0.1 µm)
3) HRT
2) TCO ( 0.05 µm) (- Anode)
1) Glass superstrate (1000 µm)
Light
8
PN Junction

• External load for power extraction creates
  voltage potential
• Decreases band bending and depletion width
• Leads to recombination and current loss across PN
  junction

Recombination
V (EQfn EQfp)/q
Photo generated current
Voltage potential, V IR Power extracted
PVI
9
Deep Electronic States as Recombination Centers

• Semiconductor band gap
• Defects, impurities, and GBs cause allowed DES
• These states capture and emit electrons and holes
  from CB and VB
• Shockley-Read-Hall recombination degrades solar
  cell performance

10
Research

• Goals
• Detect, identify, and characterize deep
  electronic states (DES) in polycrystalline CdTe
  solar cells
• Correlate DES to cell performance
• Recombination degrades performance
• Doping enhances performance
• Presentation of Selected Results
• Admittance Spectroscopy (AS) technique
• SCAPS simulations variable cell thickness
• Experimental yes/no Cu and yes/no CdCl2
• Effects of cell stressing on DES
• JSC performance versus T

11
Admittance Spectroscopytechnique for
characterizing deep electronic states (DES)
12
DES Emission and Capture Rates

• Assumes no degeneracy and near equilibrium
  conditions
• Capture rate depends exponentially on Ef and T
• Emission rate depends exponentially on Et and T
• If Et Ef, emission rate capture rate, DES
  half full

13
Band Diagram with DES
E conduction
At equilibrium DES are charged just up to the
fermi level
E DES
E Fermi
One Sided PN Junction
E acceptor
E valence
Depletion Region
P type
14
Band Diagram with Small AC Signal
E conduction
Small AC signal Oscillates the Fermi level
DES fill/empty in response to oscillating Fermi
level
E DES
Oscillating Fermi level
One Sided PN Junction
E acceptor
E valence
Depletion Region
P type
15
Determining Characteristic Frequency

• If AC freq lt capture/emission rate, DES charge
  near Fermi level crossing point follows the
  oscillation.
• If AC freq gt capture/emission rate, DES charge
  does not follow the oscillation.
• Measure over a range of AC frequencies to
  determine capture/emission rate characteristic
  frequency
• Measure over range of T to determine DES as
  a function of T

16
Capacitance vs Frequency
CdTe Solar Cell from NREL
Decreasing Cp with frequency
dCp/dln(?) minima indicate ?t
17
Capacitance vs Temperature
8
Each line is at a fixed AC signal frequency
1kHz line
Decreasing Cp with T indicates steep DES
concentration gradient
6
1MHz line
4
Cp (nF/cm2)
Majority carrier DES ?Cp indicates trap
concentration
2
0
Temperature ( C )
18
SCAPS Simulation

• SCAPS software models 1D solar cell based on
  Poissons equation, continuity equations, and
  boundary conditions
• Examples illustrate
• Cell thickness impact on capacitance amplitude
  (8µm vs 3µm)
• Fermi level pinning from high concentration
  defects distorts measured Ea values
• Concentration gradient can cause decreasing
  capacitance with increasing temperature.

19
SCAPS Simulated CdTe Solar Cell, -170oC
Simulated DES, (Ea0.15 eV, sa2 X 10-14 cm2 ,
Nt3 X 1014 cm-3) Negligible shallow level
acceptor concentration Na1012 cm-3
eV
eV
Cell A 8µm CdTe layer
EC
Cell B 3µm CdTe layer
Ef
Lateral extent of Fermi level (blue line) pinning
by trap level (red/green line) is greater with
thick CdTe layer
EV
Distance from back contact µm
20
SCAPS Simulated CdTe Solar Cell, -170oC
eV 0.15eV DES Relative to Fermi
s-1 Hole Capture/Emission Rates
106
104
cp 8µm cell ep 8µm cell cp 3µm cell ep 3µm
cell
Fermi 8 µm cell 3 µm cell
102
1
Distance from back contact µm
21
SCAPS Simulated CdTe Solar Cell, -170oC
cm-3 Free Carrier Density (holes)
cm-3 Ionized acceptor DES Density nt
1013
1015
1012
1014
1011
p 8µm cell p 3µm cell
1013
1010
nt 8µm cell nt 3µm cell
1012
Distance from back contact µm
22
SCAPS Simulated CdTe Solar Cell, -170oC
DES Characteristic Frequency ft
Capture/emission rate per unit volume
s-1 cm-3
s-1
106
1019
105
104
1018
8µm cell 3µm cell
103
ft 8µm cell ft 3µm cell
102
1017
Distance from back contact µm
23
SCAPS Simulated CdTe Solar Cell, -170oC
nF cm-2 Capacitance
nF cm-2 dCp/dln(?)
3
0.1
Non-exponential behavior
0.0
2
-0.1
-0.2
Cp 8µm cell Cp 3µm cell
1
-0.3
8µm cell 3µm cell
-0.4
ft
ft
0
-0.5
Simulated LCR meter frequency (Hz)
24
SCAPS Simulated CdTe Solar Cell
1/tT2 (s-1 K-2) Arrhenius Regression,
Simulated DES, (Ea0.15 eV,
sa2 X 10-14 cm2 , Nt3 X 1014 cm-3)
1000.0
8 µm cell Ea0.10 eV sa2
X 10-15 cm2 ß0.45
100.0
10.0
3 µm cell Ea0.15 eV sa2
X 10-14 cm2 ß0.94
1.0
-170 oC
0.1
25
SCAPS Simulated CdTe Solar Cell

• Fermi level pinning can cause AS signal intensity
  dependency on cell thickness
• Anything that impacts Fermi/trap proximity
  impacts signal intensity
• Back contact Schottky barrier height
• Shallow acceptor concentration
• Because ?t depends on Et and Ef, band bending can
  distort Ea and sa.
• Thus, ranges of Ea values can represent the same
  discrete energy defect level.

26
Decreasing Cp with increasing T

• Three criteria for decreasing Cp with increasing
  T
• High concentration DES (1014 cm-3)
• Back contact barrier (0.5eV)
• Concentration gradient near back contact (1019
  cm-3 to 1014 cm-3)

27
Decreasing Cp with increasing T
Cp (nF cm-2)
Decreasing Cp with increasing T
4
1 kHz
3
2
1 MHz
1
0
-190
-170
-150
-130
-110
-90
Temperature (oC)
28
Identification of DES
29
NREL cells Cu/CdCl2 study
30
NREL CdTe cell Yes Cu, Yes CdCl2 Before stress
NREL CdTe cell No Cu, Yes CdCl2 Before stress
H1
H3
H1
H2
Top line 1KHz, Bottom line 1MHz
Top line 1KHz, Bottom line 1MHz
Temperature (Celsius)
Temperature (Celsius)
NREL CdTe cell Yes Cu, No CdCl2 Before stress
NREL CdTe cell No Cu, No CdCl2 Before stress
H3
H1
H1
H2
Top line 1KHz, Bottom line 1MHz
Top line 1KHz, Bottom line 1MHz
31
Arrhenius Plot
1/(t T2)
H1 Ea0.13eV
103
H2 Ea0.30eV
H3 Ea0.47eV
102
101
100
10-1
10-2
32
H1 Signature

• H1 attributed to and A-center
• Ea0.13eV consistent with reported values
• Higher concentration with addition of CdCl2
• Lower concentration with addition of Cu because
  of displacing
• NREL cells had a nitric phosphoric acid (NP) etch
  that creates a tellurium rich layer at the back
  contact and a concentration gradient.

33
Cells from IEC

• IEC (Institute for Energy Conversion) cells had a
  bromine methanol etch that has less cadmium
  leaching and creates fewer VCd.
• No VCd gradient (no decreasing Cp with increasing
  temperature)
• H1 not detected in non-CdCl2 treated cells.

H3
H2
H1
H1
H3
H2
34
H2, H3 Signature

• H2 Attributed to
• Ea0.30eV consistent with reported values
• Only detected in Cu treated cells
• H3 possibly or
• Only detected with non-Cu treated cells
• Assignment is tentative

35
NREL CELL Yes Cu, Yes CdCl2Stressing

• VCd concentration decreases
• CuCd concentration increases
• VCd gradient essentially not detectable

36
correlation to
37
NREL CELL J-V-T Analysis
yes-Cu/yes-CdCl2 20oC to -40oC -50oC to
-110oC -120oC to -180oC
38
Jsc versus T for NREL cells

• Sharp carrier transport freeze out for CdCl2
  cells
• Corresponds to Fermi energy Ef0.16eV
• Indicates that A-center
  impacts Jsc current collection

JSC
39
Jsc versus T for IEC cells

• Sharp carrier transport freeze out for CdCl2
  cells
• Corresponds to Fermi energy Ef0.14eV
• Indicates that A-center
  impacts Jsc current collection

40
Conclusions

• Applied admittance spectroscopy to detect,
  identify, and characterize DES in CdTe solar
  cells
• Identified , A-center and
  defects
• Observed concentration gradient in NP
  etched cells (NREL) but not in BrM etched cells
  (IEC)
• Detected migration/mutation of DES with stressing
• SCAPS simulations have shown how high
  concentration DES can distort AS results
• Indications that enhances
  current collection in CdCl2 treated cells.

41
Future Work

• Determine if really is the
  dopant in CdCl2 treated cells. Ramifications for
  cell optimization.
• Continue characterizing cells from different
  sources/process treatments to better understand
  their DES
• Design and develop higher efficiency cell growth
  processes based on results from admittance
  spectroscopy and other characterization
  techniques

42
Acknowledgements

• Thesis Advisors - Tim Ohno, Victor Kaydanov
• Dave Albin for supplying NREL cells and
  stressing, and Brian McCandless for supplying IEC
  cells
• CSM/NREL/CdTe National Team for numerous
  discussions
• Dave Albin, Joe Beach, Reuben Collins, Sam
  Demtsu, Al Enzenroth, Alan Fahrenbruch, Scott
  Feldman, Steve Hegedus, Steve Johnston, Tim
  Gessert, Brian McCandless, Dennis Readey, Jim
  Sites, Xuanzhi Wu
• NREL/DOE sponsors for funding this research
• NREL subcontract ADJ-2-30630-05

43
Questions ?