CZTS PL data JC talk
Download as PPTX, PDF0 likes248 views
The document summarizes two recent studies on photoluminescence (PL) spectroscopy of copper zinc tin sulfide (CZTS) materials for solar cell applications. The first study examines PL and internal quantum efficiency measurements of CZTSSe and CIGSSe solar cell devices, finding evidence that band tailing from electrostatic potential fluctuations is more severe in CZTSSe. The second study analyzes temperature- and intensity-dependent PL of CZTS polycrystals with varying compositions, observing different emission behaviors correlated with composition and attributing results to a model of fluctuating electrostatic potentials. Both papers provide support for CZTS being a highly disordered system affected by charged defects and potential fluctuations.
![IQE & PL On CZTSSe & CIGSSe Devices
• Estimate band gap from inflection of IQE curve
• PL peaks shifts to lower E than Eg (red shift) due to band
tailing from
• Band gap fluctuations (different phases)
• And/ or electrostatic potential fluctuations (charged defects, e.g.
[CuZn- + ZnCu+] complex with lowest formation E)](https://image.slidesharecdn.com/cztspljc-170916154310/85/CZTS-PL-data-JC-talk-15-320.jpg)
![Main Conclusions From 2 Studies
1. Gokmen, Mitzi et al
• Existence of tail-states due to electrostatic potential fluctuations
are an important factor for CZTS devices
• Do not discount band gap fluctuations but believe effect of potential
fluctuations is more pronounced in CZTS than CIGS due to formation of
charged [CuZn- + ZnCu+] defects and low dielectric constant
2. Halliday et al
• Believe fluctuating potential model is the most suitable for their PL
data
• PL more sensitive than XRD to impact of structural differences and
secondary phases
• More evidence for CZTS being a highly disordered system
• Possibly merge discussion of two effects (band gap and potential
fluctations) considered separately in the Gokmen study?](https://image.slidesharecdn.com/cztspljc-170916154310/85/CZTS-PL-data-JC-talk-29-320.jpg)
CZTS PL data JC talk
- 1. WMD JOURNAL CLUB Overview of Recent CZTS Photoluminescence Studies 13th October 2015 Suzanne Wallace
- 2. Why Would We Care About PL Spectroscopy? • Easy measurement Contactless method does not require a full PV device to be constructed (…can be time consuming) • Non-destructive optical technique • Particularly relevant to PV sub-group (or anyone who cares about defects in materials!) • Detecting point defects and bulk impurities in semiconductors
- 3. Why Would We Care About PL Spectroscopy? • Easy measurement Contactless method does not require a full PV device to be constructed (…can be time consuming) • Non-destructive optical technique • Particularly relevant to PV sub-group (or anyone who cares about defects in materials!) • Detecting point defects and bulk impurities in semiconductors
- 4. useful! +
- 6. What Is Luminescence? • ‘Cold light’ (as opposed to incandescence from e.g. a bulb or flame) • Light-emission from an excited semiconductor • Excited by… • Electric current electroluminescence (LEDs) • Electron beam cathodoluminescence • Mechanical action triboluminescence • Nuclear radiation radioluminescence • Optical absorption photoluminescence • (and even more types of luminescence!)
- 7. 2 Types of Photoluminescence • Fast fluorescence (~ 10-8 s) • Small dye molecules and nano-inorganic materials • Slower phosphorescence (s, mins, …hours) • Relatively large-scale inorganic materials Main difference between measuring fluorescence and phosphorescence delay between excitation and measurement
- 8. What is PL Spectroscopy? • Optical excitation of a sample to create e--h+ pairs (excitons) • Excitons eventually deactivate either radiatively (give off a phonon) or non-radiatively • Radiative recombinations are observable with high sensitivity • Lifetime of exciton depends on crystal lattice, type and concentration of impurities
- 10. Common PL Transitions Rarely seen in materials with small effective masses
- 11. Common PL Transitions Free exciton recombination (materials with strong excitonic binding show decrease in gap)
- 12. Common PL Transitions Excitons bound to shallow donors or free electron to shallow acceptors
- 13. Band Tails In Heavily-Doped Semiconductors • Band structure characterizes material • Doping deforms the band structure • Each impurity ion (or intrinsic defect) locally introduces distinct level in band gap At high densities local levels interact to form a band • For high n-type doping, impurity band merges with CB ‘rigid shift’ of the CB towards VB
- 14. 1) IQE & PL on CZTSSe Devices: Gokmen, Mitzi, et al • IQE & PL on hydrazine process CZTSSe and CIGSSe solar cells (finished cells and bare films) • Same behaviours observed for each • IQE estimate of band gap • IQE & PL band-edge tail states twice as severe in CZTSSe • Low temp time-resolved PL tailing primarily from fluctuations in electrostatic potential
- 15. IQE & PL On CZTSSe & CIGSSe Devices • Estimate band gap from inflection of IQE curve • PL peaks shifts to lower E than Eg (red shift) due to band tailing from • Band gap fluctuations (different phases) • And/ or electrostatic potential fluctuations (charged defects, e.g. [CuZn- + ZnCu+] complex with lowest formation E)
- 16. Low Temperature Time-Resolved PL (TRPL) • Larger difference at low T for CZTSSe than CIGSSe • Conclude more severe tailing in CZTSSe • Larger amplitude in electrostatic potential fluctuations due to lower dielectric constant
- 17. Band Gap & Electrostatic Potential Fluctuations • Don’t rule out band gap fluctuations • Link TRPL data to electrostatic potential fluctuations • Diagram a bit misleading? Appearance of photon- assisted transition?
- 18. 2) PL on CZTS Polycrystals of Differing Compositions: Halliday, Durose et al • Temperature- and intensity-dependent PL • CZTS polycrystals with a range of different compositions • Reports of Cu-poor compositions giving better PV performance • Confirm crystalline CZTS using XRD • Determine compositions from initial weight of elements • PL peak positions and intensity consistent with model of fluctuating potential for a heavily doped semiconductor
- 19. General Overview • Very good introduction covering key issues for CZTS • TW scale energy production • Record efficiencies for devices synthesized by different methods • Secondary phases • Ambiguity in interpretation of PL data (DAP emission vs. model of fluctuating potentials) • Band tailing introduced nicely
- 20. General Overview • Very good introduction covering key issues for CZTS: • TW scale energy production • Record efficiencies for devices synthesized by different methods • Secondary phases • Ambiguity in interpretation of PL data (DAP emission vs. model of fluctuating potentials) • Band tailing introduced nicely
- 21. General Overview • Very good introduction covering key issues for CZTS: • TW scale energy production • Record efficiencies for devices synthesized by different methods • Secondary phases • Ambiguity in interpretation of PL data (DAP emission vs. model of fluctuating potentials) • Band tailing introduced nicely • Mix the two effects discussed separately by Mitzi & Gokmen? • Model of fluctuating potentials related to compositional differences (phases) • In Mitzi paper this is related to band gap fluctuations
- 23. XRD On 8 CZTS Polycrystals • Reflections are those expected for kesterite CZTS • Stannite and kesterite phases differ in Cu and Zn ordering • XRD cannot distinguish between Cu & Zn • Also considerable overlap with other phases (e.g. ZnS) • Additional peak for sample C8 attributed to unidentified secondary phase
- 24. PL Data Categorized Into 3 Groups • Group a: C4, C5, C6 single dominant CZTS feature at 1.4 eV • Group b: C1, C2, C7 dominant CZTS feature and series of less intense features • Group c: C3, C8 emission peaks across full PL range Results suggest PL provides more sensitive measure of the impact of structural differences and presence of secondary phases in CZTS than XRD N.B. Differences in PL data for C1, C2 with C3 which all had same metal composition but differing excess of S
- 25. Compositions of CZTS Polycrystals • C1, C2 and C3 all considered stoichiometric • Group a: C4, C5, C6 • Group b: C1, C2, C7 • Group c: C3, C8
- 26. Temperature-Dependent PL To excite electron or hole out of local fluctuating potential well
- 27. Temperature-Dependent PL To excite electron or hole out of local fluctuating potential well It’s all a mystery!
- 28. Temperature-Dependent PL To excite electron or hole out of local fluctuating potential well It’s all a mystery! • Discussion about validity of model • Changes in capture cross section of one of the levels provides alternative explanation (see ref 47)
- 29. Main Conclusions From 2 Studies 1. Gokmen, Mitzi et al • Existence of tail-states due to electrostatic potential fluctuations are an important factor for CZTS devices • Do not discount band gap fluctuations but believe effect of potential fluctuations is more pronounced in CZTS than CIGS due to formation of charged [CuZn- + ZnCu+] defects and low dielectric constant 2. Halliday et al • Believe fluctuating potential model is the most suitable for their PL data • PL more sensitive than XRD to impact of structural differences and secondary phases • More evidence for CZTS being a highly disordered system • Possibly merge discussion of two effects (band gap and potential fluctations) considered separately in the Gokmen study?
- 31. Original Figure For Radiative Recombination? Local distortions of band edges due to fluctuations of impurity potential Wavy lines show: band-tail (BT), band- band (BB) and band- impurity (BI) transitions
- 32. … The Verdict!
- 33. … The Verdict!
- 34. … The Verdict! • Studies seemed to be well- conducted • Although I don’t know enough about IQE to know how much to trust this measurement! • Interpretation of PL data seems to be a very subjective process! • Undecided if I agree with interpretations in these papers • Studies are well presented, arguments clearly discussed and cite very useful original works used in their analysis
- 35. … The Verdict! • Studies seemed to be well- conducted • Although I don’t know enough about IQE to know how much to trust this measurement! • Interpretation of PL data seems to be a very subjective process! • Undecided if I agree with interpretations in these papers • Studies are well presented, arguments clearly discussed and cite very useful original works used in their analysis
- 36. Resources • Paper on PL of semiconductors: http://ac.els-cdn.com/004060908190465X/1-s2.0-004060908190465X- main.pdf?_tid=ec24517c-6c46-11e5-871b- 00000aacb35e&acdnat=1444148937_409d6ef7bc4d654e1841c5627b3b2da6 • Paper on band tailing: http://journals.aps.org/rmp/abstract/10.1103/RevModPhys.64.755 • ‘Edge luminescence of direct gap semiconductors’ paper referred to by Mitzi and Halliday in their interpretation of PL data: http://iopscience.iop.org/article/10.1070/PU1981v024n03ABEH004770/meta • Very good general background on band structure (perhaps not as much PL spec as I would have liked!): http://memo.cgu.edu.tw/sykuo/Mat-8.pdf • Slides on electrical and optical characterization of semiconductors from NREL (including various PL measurements): http://www.nrel.gov/docs/gen/fy04/36831k.pdf