Title: How thin film PV can be different
1How thin film PV can be different
- Presented by Victor Karpov
- Contributions by
- D. Shvydka, Y. Roussillon,
- A. D. Compaan, D. M. Giolando
- University of Toledo
2x-PV standard interpretation
- 1-Dimensional p-n junction
- Defect dominated loss
- Material degradation due to defect generation
3Electronic processes in crystalline PV
Defects cause uniform recombination and loss
Cast
electrons
holes
E
defects
field
4Some motivations for revisiting
- So imperfect, how come it works?
- Optically inactive back contact hits Voc
- Strong buffer layer effects
- Super-additive front and back
- Variations between nominally identical devices
5Conceptual Motivations
- Lack of crystallinity disorder
- Small thickness
It follows then
6Lack of transversal self-averagingLateral
non-uniformities
Random micro-diodes
Voc1
Local diode characteristics vary
7Recombination pathways weak diodes
Low Voc micro-diode Runs exponentially strong
current In the wrong direction
Nonlinear shunt No slope in J(V) _at_ V0
Signature effects Low Voc, FF
L 1cm
8Microscopic nature of weak diodes
Main junction
Back contact
Strong back barrier
Weak SPV
Recombination path through CdS or depletion region
Poor doping Contamination Metal delamination
Either front or back junction, or both (more in
detail described in our other talk)
9Thin films tend to be insulators
Depletion width
- Lltltl,
- Semiconductor
- High carrier concentration
- Large thickness
- Lgtgtl ,
- Insulator
- Low carrier concentration
- Small thickness
10MIS rather than P-N junction
- P-N junction
- Higher N beneficial
- M does not affect VBI
- MIS junction
- N does not matter (I)
- M affects VBI
11Strong interfacial effects
High equivalent doping l - film thickness,
Ns (cm-2) - surface state density
CdS example l10-5 cm, NS1012 cm-2, ND
1017 cm-3
Interfacial states can act stronger than bulk
doping (Too few bulk states in thin films)
12MIS model facts and venues
- Evidence
- M-side having effect on VBI
- Highly resistive CdS, CdSO
- Photoconductive CdS
- Piezo-effect in CdS
- Cu depleting CdS
- Venues
- Work on CdS and interfaces
- CdS/TCO(buffer) and CdS/CdTe stresses and states
matter - Cu is not necessary for SPV when deposition is
right
13Short drift time
-- Typical uniform lifetime in poor
photoconductors
Charge carriers do not have time to
recombine (possibly facilitated by GB)
14In poly-crystalline PV
- Almost no uniform recombination
- Nonuniform recombination in weak diodes
- Nonuniformities and contacts matter
Cast
GB
E
Weak diode
electrons
holes
field
15Degradation mechanisms
- Increase in nonuniformity
- Metal shunting
- Cu path shunting through CdS
- Diode weakening under localized stress
- Back barrier deterioration
- Cu diffusion from back surface
- Adhesion failure and metal delamination
- In all cases - structure related degradation
rather than material (CdTe is quite stable)
16Conclusions
- MIS type of device
- Lateral nonuniformities
- Nonlinear shunts (weak diodes, etc.)
- Strong interfacial effects
- Uniform recombination irrelevant
- Degradation - nonuniformity and back contact
failure - Qualitatively consistent with the observations