Innovative Characterization of Amorphous and Thin-Film Silicon for Improved Module Performance
1Department of Physics and Material Science Institute
J. David CohenDepartment of Physics and Materials Science Institute University of Oregon, Eugene, OR 97403
Support the development of two thin-film technologies
Relevance/Objective
1. Thin Film Silicon-Based Devices• Characterize the electronic properties of nanocrystalline silicon to
help optimize for bottom cell component in multijuction cells.
• Aid in developing better high deposition rate materials, particularly the amorphous silicon-germanium alloys
• Understand mechanisms of light-induced degradation
2. Copper Indium Di-Selenide Based Devices• Identify factors limiting performance of commercial devices
relative to high performance laboratory cells
Summary of Activities
1. Evaluate the effects of oxygen contamination on the electronic properties of high quality NREL hot-wire CVD amorphous silicon-germanium alloys
2. Examine the electronic properties of United Solar nanocrystalline silicon, their relationship with cell performance, and their light-induced degradation.
3. Examine correlations between the electronic properties of high performance NREL CIGS materials and associated cell performance.*
* Not to be included in this presentation due to lack of time
A. Technical Focus of Recent Activities
Activities: Participants
B. Participants in Recent Activities University of Oregon Personnel:Name Primary Activity in Current PeriodJ. David Cohen Program ManagerShouvik DattaResearch Associate
Studies of NREL HWCVD amorphous silicon-germanium alloys
Peter HuggerResearch Assistant
Evaluation of USOC nanocrystalline silicon materials
Pete ErslevResearch Assistant (part-time)
Electronic properties of Cu(In,Ga)Se2materials in NREL laboratory cells
Thin-Film Partnership Collaborators:
Organization Primary Personnel InvolvedNREL Yueqin Xu, A.H. Mahan, Howard Branz,
Miguel Contreras, Rommel NoufiUnited Solar Ovonic Baojie Yan, Jeffrey Yang, Subhendu Guha
C. Experimental Methods Employed
I. Transient Photocapacitance SpectroscopyA type of sub-band-gap optical absorption spectroscopy
Discloses: Band-tail distributions, deep defect distributions and properties, minority carrier transport
II. Admittance/Drive-level capacitance profilingDiscloses: Majority carrier density, majority carrier mobility,
deep defect densities and their capture cross sections
Activities: Methods
Comparison of Sub-band-gap Optical Spectroscopies
Method Detected Quantity DisadvantagesDirect Transmitted/Reflected Light Not Sensitive enough for thin films
Photothermal Deflection (PDS)
Absorbed Energy Dominated by surface defects for low bulk defect samples
Constant Photo-current Method(CPM)
Carriers Photoexcited into Majority Band
Carrier mobility influences spectra; Dominant current path may not reflect internal bulk properties
Photoemission Electrons ejected into vacuum Very surface sensitive
PhotocapacitanceSpectroscopy
Photoexcited released charge from depletion region
Requires reasonably good barrier junction, moderate conductivities
This last method is thus naturally suited for examining sub-band-gap optically induced defect transitions within photovoltaic devices.
Activities: Methods
Photocapacitance Spectra for Disordered Semiconductors
Two Primary Features:
(1) Exponential “Urbach” Tail.g(E) ∝ exp(E/EU) where EU
indicates degree of disorder
(2) Dominant Deep defect band.Usually well fit to a Gaussiandistribution of transitions
EU=42meV
Intrinsic a-Si:H
Gelatos, Mahavadi, Cohen, and Harbison, Appl. Phys. Lett (1988)
Activities: Methods
Drive level capacitance profiling (DLCP)Michelson, Gelatos, and Cohen, Appl. Phys. Lett. 47, 412 (1985).
•Junction capacitance is determined as a function of the frequency and the amplitude ofthe applied oscillatory voltage.
•One obtains a direct integral over defect states down to an emission cutoff energy, Ee:
•Procedure is then repeated at different dc biasto produce spatial profile.
C = C0+C1δV+C2δV2+…
∫−
+=−=F
eC
E
EEDL dxxEgn
CAqCN ),(
2 12
30
ε
Ee = kBT log(ν/ω)
Activities: Methods
Schematic showing portion of the defect band that is detected by the drive-level profiling method (shaded region). Ndl corresponds to an integral over g(E):
where Ee = kBT ln(ν/ω)
Drive-level Profiles for Amorphous Silicon Sample
∫∫−−
≈+=−=F
eC
F
eC
E
EE
E
EEDL dxxEgdxxEgn
CAqC
N ),(),(2 1
2
30
ε
We infer a broad deep defect about 0.8eV below EC of density ~ 1.5 x 1016 cm-3
Activities: Methods
Budget History
ProjectTask(s)
Total Value
High growth rate a-Si,Ge:H alloys
Nanocrystalline Si films and devices
CIGS electronic properties vs. devices
Grand Total over 16 months
40%(~$90K)
40%(~$90K)
20%(~$40K)
~$220K
Expenditures: October 2005 to January 2007
(1) HWCVD Amorphous Silicon-Germanium Alloys
• Background: Need for higher deposition rate a-Si,Ge:H material
• Improved Hot-Wire CVD Deposited a-Si,Ge:H seems promising
With Yueqin Xu, Harv Mahan, and Howard Branz
Accomplishments
• Loss of Optimum Properties Due to Oxygen Contamination?
• Results on 2006 Sample Series with Controlled Oxygen Leak
Critical Need: a-SiGe:H Cells at Higher Growth Rates
• United Solar devices – All in Light Degraded State
• Specular Stainless Steel -- No back reflector
Accomplishments: Background
Electronic Properties of High Growth Rate Materials
Cell performance deteriorates as growth rate is increased, and this appears to be uncorrelated with deep defect densities
Higher Urbach energies (bandtail widths) appear to be strongly correlated with higher growth rates and poorer cell performance
Higher frequencies for PECVD growth yield higher performance (and narrower bandtails). Not as successful for a-Si,Ge:H.
Possible solution explored at NREL: Hot-wire CVD of a-Si,Ge:H
Accomplishments: Background
Urbach Energies for the Best PECVD a-SiGe:H
Accomplishments: Background
The Best PECVD vs. HWCVD a-Si,Ge:H
Accomplishments: Background
Spectra for NREL HWCVD a-Si,Ge:H
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Photon energy (eV)
10-142
10-132
10-122
10-112
10-102
10-9
Phot
ocap
acita
nce
sign
al (a
rb. u
nits
)
0at.% Ge5at. % Ge20at. % Ge
55meV
45meV45meV
SS substrates
2000oC Tungsten Filament
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.01x10-14
1x10-13
1x10-12
1x10-11
1x10-10
1x10-9
15at.% Ge 29at.% Ge 47at.% Ge
Phot
ocap
acita
nce s
igna
l (ar
b.un
its)
Photon Energy (eV)
1800oC Tantalum Filament
42meV43meV
45meV
Accomplishments: Background
The Best PECVD vs. HWCVD a-Si,Ge:H
Accomplishments: Background
2005 Samples Exhibited Poorer Properties
0.50 0.75 1.00 1.25 1.50 1.75 2.0010-7
10-6
1x10-5
1x10-4
10-3
10-2
10-1
100
101
29% Ge,More O2,E
U~66 meV
29% Ge,Less O2,EU~43 meV
Phot
ocap
acita
nce
Sign
al (a
rb.u
nits
)
Energy (eV)
Photocapacitance Spectra: 29at.% Ge
10 15
10 16
1017
1018
10 19
10 20
1021
10 22
2.01.51.00.50.0
Depth: [µm]
L1306_O_level L1430_O_level
Oxygen SIMS Profiles
43meV
66meV
Accomplishments: Background
Oxygen Contamination Problem Resolved
Controlled leak valve added
New sample series deposited by Yueqin Xu in early 2006 with varying air leak contamination levels
Fall 2005: Leak source identified and eliminated
0.50 0.75 1.00 1.25 1.50 1.75 2.0010-7
10-6
1x10-5
1x10-4
10-3
10-2
10-1
100
101
29% Ge,More O2,E
U~66 meV
29% Ge,Less O2,EU~43 meV
Phot
ocap
acita
nce
Sign
al (a
rb.u
nits
)
Energy (eV)
Photocapacitance Spectra: 29at.% Ge
43meV
66meV
Accomplishments
HWCVD a-Si,Ge:H Sample Series with Controlled Air Leak
SIMS Analysis:
Germanium content 29-32at.% for all samples
Systematic oxygen level variation
Sample Air leak Flow rate (sccm)
Oxygen Content ( cm-3)
Thickness (μm)
Initial Substrate Temp (K)
Final Substrate
Temp (K)
Gas Ratio
GeH4/(GeH4+SiH4)
A 0.00 ~8×1018 1.80 204 289 0.19 B 0.02 ~3×1019 1.60 204 289 0.19 C 0.06 ~1×1020 1.75 204 284 0.19 D 0.20 ~5×1020 2.20 204 275 0.19
(oC)(oC)
0.01 0.1
1019
1020
Accomplishments
0.1
1
Rat
io o
f 74G
e/30
Si (a
rb.u
nits
)
Oxy
gen
Con
tent
(cm
-3)
Air leak Flow Rate (sccm)
0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.810-15
1x10-14
1x10-13
1x10-12
1x10-11
1x10-10
1x10-9
1x10-8
350 K1.1 kHz
TPC
(arb
.uni
ts)
Incident Photon Energy (eV)
A, 0.00 sccm C, 0.06 sccm D, 0.20 sccm
Photocapacitance Spectra for Oxygen Series
Sample A exhibits EU = 43.5meV
Sample B exhibits EU = 38 meV (?)
Does this mean low levels of oxygen actually improves the electronic properties of HWCVD a-SiGe:H materials?
(NO, but it doesn’t hurt that much either)
Accomplishments
Defect Densities from DLCP Measurements
Mid-gap defect densities are low and vary roughly by a factor of two over range of oxygen
Larger density in higher leak samples may be due to effects of oxygen weak donor level
0.3 0.4 0.5 0.6 0.71015
1016
1017
Accomplishments
A B C D
ND
L (cm
-3)
Distance from the Barrier (μm)
State B
0.00 sccm0.02 sccm0.06 sccm0.20 sccm
Light-Degraded State
Sensitivity to Minority Carrier Collection
• If both escape depletion region there is no net change in charge density
⇒ Small photocapacitance signal, but larger junction photocurrent signal
EF
EC
EV
Photocapacitance signal ∝ n - p while Photocurrent signal ∝ n + p
• For optical energies near the gap, electrons and holes generated equally
Accomplishments
Measuring Fraction of Collected Holes in a-Si,Ge:H
United Solar RF35at.% Ge
Compare transient photocapacitance and photocurrent spectra
Ratio of signals in bandtail region⇒ the hole collection fraction is ≈ 95%
under our experimental conditions
(c) (d) (a) (b)
EC
EF
EV
Toward Barrier Junction
(a) (b)(c) (d)
Ratio is nearly a constant:
(1+p/n)/(1-p/n)
Accomplishments
Measuring Fraction of Collected Holes
United Solar RF35at.% Ge
United Solar RF PECVD 35at.% Ge Sample
Hole Collection: 95% 66%
Accomplishments
Measuring Fraction of Collected Holes
Hole Collection: 98%
NREL Hot-Wire CVD 29at.% Ge, No air-leak
No air-leakState A
99%
No air-leakState B
Accomplishments
Measuring Fraction of Collected Holes
Hole Collection: 95%
NREL Hot-Wire CVD 29at.% Ge, 0.06sccm air-leak
0.06 sccmair-leakState A
97%
0.06 sccmair-leakState B
Accomplishments
Effect of Extra Oxygen: New Defect Transition
NREL Hot-Wire CVD 29at.% Ge, 0.06sccm air-leak
0.06 sccmair-leakState BIntroducing this defect transition (in addition
to usual mid-gap deep defect) allows excellent fits to both TPC and TPI (thin lines shown)
Urbach is actually 47meV, not the < 40meV value that seemed to fit TPC by itself
Moderate oxygen levels do not appear to greatly impair hole collection, or the sus-ceptibility to light-induced degradation
Dip in TPC Spectrum is due to a 1.4eV transition from VB into empty defect level
Accomplishments
HWCVD a-Si,Ge:H Summary
Bad News: Oxygen wasn’t the problem, so we really don’t know why the samples produced during 2005 were so poor!
Good News: HWCVD a-Si,Ge:H Samples again exhibit excellent electronic properties! (in some respects, even betterthan the best PECVD samples)
Interesting: Oxygen introduces new defect transition into gap: The oxygen donor level predicted as long as 24 years ago?
Accomplishments
(2) Nanocrystalline Silicon Materials Properties
TPC Spectra reveal mixed phase nature of nanocrystalline SiTemperature variation shows degree of minority carrier collection
New Sample Series StudiedHigh efficiency working n-i-p devices examined directly
Hole Collection Determined via TPC vs. Device Performance Is there a correlation?
Nature of light-induced degradationPossible model is proposed
UNITED SOLAR OVONICwith Baojie Yan, Jeffrey Yang, and Subhendu Guha
Accomplishments
A.F. Halverson, J.J. Gutierrez, J.D. Cohen, B. Yan, J. Yang, and S. Guha, Appl. Phys. Lett. 88, 71920 (2006)
Comparing CPM and Photocapacitance Spectrum for nc-Si:H
FTPS w/o Si FilterFTPS w Si FilterCPM
Fourier-Transform Photo Current Spectroscopy(FTIR) and CPM Spectra
United Solar nanocrystalline Si spectrum from TPC spectroscopy
Accomplishments: Background
M. Vaněček, A. Poruba, Z. Remeš, N. Beck, and M.Nesládek, J. Non-Cryst. Solids 227-230, 967 (1998).
Very temperature dependent !
EC
EV
EF
Junction Interface
In a-Si:H component many holes remain trapped ⇒ mostly electrons escape ⇒ large photocapacitance signal
Electron Energy
a-Si:H
nanocrystallites
y
At photon energies greater than 1.1eV, free electronsand holes excited in nano-crystallites.
At moderately high temper-atures both carriers escape⇒ little change in charge⇒ small photocapacitance signal
x
Hole Collection Inhibited in a-Si:H Component
Accomplishments: Background
Minority Carrier Collection Increases with T
0.8 1.0 1.2 1.4 1.6 1.8 2.0
Incident Light Energy (eV)
1015
1016
1017
1018
1019
1020
1021
1022
Pho
toca
paci
tanc
e si
gnal
(a.u
.)
170K300K
(a) USOC 14661State A
Transient Photocapacitance
0.8 1.0 1.2 1.4 1.6 1.8 2.0
Incident Light Energy (eV)
1015
1016
1017
1018
1019
1020
1021
1022
1023
Pho
tocu
rren
t Sig
nal (
a.u.
)
170K300K
(b) USOC 14661State A
Transient Junction Photocurrent
Photocurrent signal dominated by nanocrystallite component at all T
Increased minority carrer collection suppresses signal in nanocrystallites
Accomplishments: Background
Types of Nanocrystalline Si Devices StudiedDeposited at : United Solar Ovonic Corp.
SANDWICHSS/n+/a-Si:H/nc-Si:H/a-Si:H
Advantages:- Purely intrinsic layers- Blocks diffusion of O2
Disadvantages:- Contains a-Si:H layers- Differs from working device
~1.3 μm .7 μm
Advantages:- Purely nc-Si:H- Working device allows compar-
ison with cell peformanceDisadvantages:
- Contains thin p+ and n+ layers
n-i-pSS/n+/i/p+/TCO
~1.1 μm
Most sample devices examined prior to Fall, 2005.
Sample devices studied since late 2005
Accomplishments: Background
Sample No Substrate Jsc Voc FF Eff (%) H2 Dilution
14027 Ag/ZnO 26.5 0.469 0.581 7.22 Ramping14036 SS 17.38 0.448 0.568 4.42 Ramping14037 Ag/ZnO 25.99 0.429 0.512 5.71 Constant14038 SS 17.36 0.451 0.531 4.16 Constant13993 Ag/ZnO 24.39 0.548 0.641 8.57 Ramping
Early 2006 United Solar nc-Si:H Device Series
All are n-i-p devices with either Specular Stainless (SS) substrates, or with textured Ag/ZnO back reflectors. All i-layers were more than 1 micron thick.
Samples were deposited either using a constant level of hydrogen dilution, or with a linear ramp variation in hydrogen dilution to control crystallite size
Final sample in list is close to the current best nc-Si:H United Solar device
Accomplishments
Crystalline Fractions Revealed by Raman Spectroscopy
300 400 500 600
300 400 500 600
Raman Shift (cm-1)
USOC 1399353vol.%
USOC 1403663vol.%
780n
m R
aman
Sig
nal
Low crystalline fractions yield highest VOC’s
No clear correlations for fractions above 60vol.%
Accomplishments
DLC Profiles for Constant vs. Ramped H-dilution
0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
1E15
1E16
1E17 USOC RF14036DLCP 1.1kHz
375K 350K 325K 300K
Driv
e Le
vel D
ensi
ty (c
m-3
)
Profile Distance <x> (μm)0.1 0.2 0.3 0.4 0.5 0.6 0.7
1E15
1E16
1E17USOC RF14038DLCP 1.1kHz
350K 325K 300K 275K 250K
Profile Distance <x> (μm)
Constant H-dilution Linearly Ramped H-dilution
Consequences of employing varying H-dilution to limit crystallite size:
Large defect density near n-i junction can be decreasedLower defect density maintained through most of bulk film region
Accomplishments
Spectra for 8.57% Textured n-i-p Device
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2102
103
104
105
106
107
TPI TPC, 285K
USOC RF13993
Pho
totra
nsie
nt S
igna
l (a.
u.)
Incident Photon Energy (eV)
Here an a-Si:H componentis clearly in evidence
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2102
103
104
105
106
107
TPI TPC, 275K TPC, 285K
USOC RF13993
Pho
totra
nsie
nt S
igna
l (a.
u.)
Incident Photon Energy (eV)
Reducing the temperature increases nanocrystallite signal
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2102
103
104
105
106
107
TPI TPC, 275K TPC, 285K TPC, 295K TPC<0
USOC RF13993
Pho
totra
nsie
nt S
igna
l (a.
u.)
Incident Photon Energy (eV)
Increasing the temperatureactually yields a negativeTPC signal near 1.5eV !
Accomplishments
Temperature Dependence of Hole Collection in nc-Si:H
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2102
103
104
105
106
107
TPC, 285K TPC, 275K TPC, 295K TPC<0
USOC RF13993
Pho
totra
nsie
nt S
igna
l (a.
u.)
Incident Photon Energy (eV)
Two activated hole trap processes seem to be evident in this sample
3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8 5.0103
104
105
106
TPC spectra from 8.57% n-i-p sample device
signal>0 signal<0
USOC RF13993
Ea=.15 eV
Ea=.72 eV
TP
C S
igna
l
1000/T (K-1 )
Increasing hole collection
Accomplishments
Comparison of the three Ag/ZnO Devices plus one from previous sample series
Correlation with Cell Performance
Accomplishments
Minority carrier collection determined from TPI/TPC ratio at 300K near 1.5eV
Correlation looks promising, but data from more sample devices need to be added
Effects of Prolonged Light Exposure
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2102
103
104
105
106
107
TPC, 285K TPC, 275K TPC, 295K TPC<0
USOC RF13993
Pho
totra
nsie
nt S
igna
l (a.
u.)
Incident Photon Energy (eV)
Effect of varying temperature
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2102
103
104
105
106
107
State B State A
USOC RF13993295 K
Pho
toca
paci
tanc
e S
igna
l (a.
u.)
Incident Photon Energy (eV)
Annealed vs. light-soaked state
50 hours of light- soaking of 610nm filtered ELH light at 500mW/cm2
Accomplishments
Effect of Light Soaking on TPC Spectra
Light-soaking significantly reduces hole collection
Effect of Light Soaking on Capacitance Profiles
a)
USOC RF14036
350 K325 K300 K
0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1
1E15
USOCRF14036State B
350 K325 K300 K
Distance FromJunction Interface (μm)
b)
Accomplishments
Does light soaking actually reduce the density of deep defects?
No, we believe it increases the an activation barrier that inhibits its response
Annealed State Light-soaked State
Proposed Model to Explain Changes with Light-Soaking
Accomplishments
E F E F
As illustrated, positive charge collects at the a-Si:H/crystallite interfaces, while compensated by negative charge in the a-Si:H region, and the crystallites nearby.
Holes have larger potential barriers at the phase boundary. Also, the activation energy for exchanging electrons between deep defects and conduction band increases.
In light-soaked State B, defects shift down in energy with respect to the bulk conduction band.
Major Events/Milestones
Key Findings and Conclusions
Our measurements continue to show superior electronic properties for lower filament temperature NREL HWCVD a-Si,Ge:H alloys.
The minority hole collection appears nearly as good as the best PECVD alloys with possibly less deterioration after light soaking.
Previous problems with air contamination were solved; moreover, a distinct oxygen related defect state has been identified.
We have successfully carried out TPC and TPI sub-band-gap spectroscopy on nc-Si:H n-i-p devices employing both SS and textured Ag/ZnO substrates.
A correlation appears to exist between the minority carrier collection fraction deduced by the TPC measurements and the device performance.
Light-induced degradation exists and reduces minority carrier collection. A possible mechanism to explain our observations has been formulated.
Future Directions
Broad Future Plans
NREL HWCVD a-Si,Ge:H
Work with NREL as they seek to incorporate and optimize performance of their HWCVD a-Si,Ge:H materials within photovoltaic devices.
Evaluate the electronic properites of NREL HWCVD a-Si,Ge:H under higher deposition rates (greater than the ~2Å/s materials studied to date)
Examine alloys with other Ge fractions, and determine whether the effects of oxygen contamination appear similar.
United Solar nanocrystalline Si
Examine properties of nc-Si:H closer to a-Si:H transition region to help understand how far one can go toward maximizing VOC without reducing JSC
Explore the effects of hydrogen profiling over a greater range of parameter space
Test the implications of proposed light-induced degradation model; for example, by varying photon energies, or examining such effects under applied bias.
THANK YOU!
Studies of NREL Fabricated CIGS Devices using Admittance Spectroscopy and DLCP
Accomplishments
Eight NREL Devices from Three Depositions
Device#
VOC(volts)
JSC(mA/cm2
)
Fill Factor (%)
Efficiency(%)
C1919-11 Cell 3 0.630 32.95 71.43 14.833C1919-11 Cell 4 0.642 32.51 68.99 14.393C1813-21 Cell 3 0.657 33.26 76.08 16.613C1813-21 Cell 4 0.664 32.67 77.82 16.888C1813-21 Cell 6 0.667 33.19 77.52 17.168C1924-1 Cell 4 0.724 30.82 78.19 17.449C1924-1 Cell 5 0.717 30.81 77.44 17.118C1924-1 Cell 6 0.714 30.95 77.07 17.039
Sample devices provided by Miguel ContrerasThree depositions: Each such sample contained 6 devices8 devices with varying levels of performance were chosen for studyDevice areas were 0.406 cm2 in all cases
Accomplishments
Eight NREL Devices from Three DepositionsRelationship of VOC×JSC products and Fill Factors to Device Efficiencies
: C1919 devices, : C1813 devices, and : C1924 devices
Fill factor variations account for most of the differences in efficiencies
• The C1924 devices differ somewhat from the C1919 and C1813 devices, exhibiting higher VOC’s and lower JSC’s.
• The C1924 devices have a slightly higher Ga fraction.
Accomplishments
SIMS Analysis: Ga to In Ratios
0 400 800 1200
0
2
4
6
NREL Samples 1924, 1813, and 1911TOF-SIMS Depth Profi le AnalysisScaled Time on 1924
^71
Ga
/ ^11
3 In
T ime (s)
Sample 1924 Sample 1813 Sample 1911
Depositions for samples 1813 and 1911 have nearly identical composition profiles, while deposition for sample 1924 is significantly different.
Accomplishments
Zero-Field RegionDLCP Density
Depletion RegionDLCP Density
2) NREL C1919 Cell 4: 14.4%
0. 2 0 .3 0.4 0 .5 0 .6 0. 73
1 E1 6
Comparison of DLC Profiles
N R EL C 1 91 9 -1 1 C el l 4, 1 4.4 % ef f .D L C P 10 k H z 0 8 -2 3- 20 0 5A re a = .40 8 c m 2
NDL
(cm
-3)
P ro fi le D e pth ( μm )
8 0 K 1 2 0 K 1 6 0 K 2 0 0 K
0. 2 0 .3 0 .4 0 .5 0 .6 0.7
3
1 E 1 6
N R E L C 1 92 4 -1 C e ll 4 , 1 7 .4 % e ff .D L C P 10 k H z 0 8- 2 1- 20 0 5Ar e a = .40 8 c m 2
NDL
(cm
-3)
P ro fi le D e pth ( μm )
8 0 K 1 2 0 K 1 8 0 K 2 0 0 K
4) NREL C1924 Cell 4: 17.4%
0. 1 0. 2 0. 3 0. 4 0. 5 0.6 0.7 0.8 0.9 1.01 E1 5
3
1 E1 6N R E L C 18 1 3- 21 C e ll 4 , 16 .9% e f f .D LC P 1 0 k H z 08 - 28 -2 0 05A r ea = .4 1 0 c m 2
NDL
(cm
-3)
P ro fi le D e pth ( μm )
8 0 K 1 2 0 K 1 6 0 K 2 0 0 K
3) NREL C1813 Cell 4: 16.9%
Accomplishments
Correlation of Device Efficiencies with DLCP Densities in Different Absorber Regions
DLCP Densities in Region Close to Barrier DLCP Densities in Zero-Field Region
Correlation of efficiencies to DLCP density close to barrier is quite good, except for the 3 high VOC devices from sample C1924-1For those 3 devices we believe the higher value of VOC is due to a slightly higher Ga fraction, and because the carrier density is higher outside the depletion region.
Accomplishments