Presentation is loading. Please wait.

Presentation is loading. Please wait.

Routes for Rapid Synthesis of Photovoltaic Absorber Materials: The Need for Diffusion Data Ranga Krishnan, Gabriel Tong, Chris Muzzillo, Woo Kyoung Kim,

Similar presentations


Presentation on theme: "Routes for Rapid Synthesis of Photovoltaic Absorber Materials: The Need for Diffusion Data Ranga Krishnan, Gabriel Tong, Chris Muzzillo, Woo Kyoung Kim,"— Presentation transcript:

1 Routes for Rapid Synthesis of Photovoltaic Absorber Materials: The Need for Diffusion Data Ranga Krishnan, Gabriel Tong, Chris Muzzillo, Woo Kyoung Kim, Tim Anderson (UF), Andrew Payzant (ORNL), Carelyn Campbell, Ursala Kattner (NIST), and Jianyun Shen (Beijing Inst. Met.)

2 Jevons Paradox William Stanley Jevons Technological progress that increases the efficiency with which a resource is used, tends to increase (rather than decrease) the rate of consumption of that resource.

3 Energy (eV) Simple Diode Analysis for a 24%- efficient Si solar cell Intensity (mW/m 2 -  m) Wavelength (nm) Photons used Maximum energy collected = E gap Usable power 24% Unused Photons 19% 31% Loss for Energy above E gap V oc < E gap 16% Fill Factor 5% Other Losses 5% Other losses: Absorption Collection Reflection Series R Shunts

4 Photovoltaic Technologies First generation – –Single crystal and multicrystalline Si – –a-Si:H single junctions Second generation – –Inorganic thin films (CuInSe 2 & CdTe) – –Multijunction III-V & a-Si:H Third generation – –Nanostructured & quantum dot PV – –Photoelectrochemical cells – –Organic photovoltaics – –Intermediate band concepts Solarmer OPV module (PV-Tech.org) CdTe nanorods N. Lewis, Caltech Shell Solar United solar

5 Routes to Increased Efficiency www1.eere.energy.gov/solar/solar_america/pdfs/d_horwitz_os_bes.pdf

6

7 Substrate Sputtered Mo p-CIGS absorber n-CdS ZnO AR coating Ni/Al grid CIGS solar cell structure Most Promising Thin Film Absorber Material Direct band gap (Eg ~ 1.2 eV) High optical absorption coefficient: ~ 2  m High radiation resistance High reliability Lower cost per Watt installed High conversion efficiency: cell: 20% and module: 13% Efficient in low-angle & low-light conditions Flexible substrates possible (BIPV, cheaper substrates?) Positive response under concentration Cu(In 1-x Ga x )Se 2 Solar Cells Cu In or Ga Se a a c Chalcopyrite structure

8 NREL 3-stage Process: Champion Cell NREL 3-stage Process: Champion Cell Metals rates (Å/s) Se rate (Å/s) Se In Ga Cu In 1st stage 2nd stage 3rd stage Run time (min) Ga 400 C 600 C Open shutter

9 Approaches to CIGS Synthesis 9

10 CIGS Manufacturing Costs Module costs must be considered in the context of module efficiency (impact on installation costs) Courtesy of NREL

11 The Value Proposition for High Efficiency CIGS 11 At no added cost ($/m 2 ), 17.5% CIGS module = ~$0.50/Wp module ASP target New champion lab cell efficiency (≥22%), BoS improvements are required Courtesy of NREL

12 Device structure varies between monolithically & mechanically integrated modules Future Challenges and Opportunities Device structure varies between monolithically & mechanically integrated modules Today Forward ZnO, ITO (2500 Å) Sputter Hardened TCO (moisture barrier) CdS (700 Å) Chemical Bath Deposition Sputter Cd-free; dry, eliminate CIGS (1-2.5 µm) Multiple methods (coevaporation, sputtering, printing, electrodeposition) Increase Ga-%, Reduce thickness, Rapid deposition Uniformity (composition, temp., thickness) Mo (0.5-1 µm) Sputter Na dosing Glass, Metal Foil, Plastics High temp. glass Metal foils: smooth, flex-dielectric (monolith.) Screen Print Ag Reduce shadowing Faster application Front Contact Grid

13 DICTRA Atomic Mobilities Chemical Potentials Seek CIGS Pathway High-rate High-quality Low temperature Seek CIGS Pathway High-rate High-quality Low temperature Phase Diagrams Thermodynamic Properties CIS Formation Pathways & Kinetics HT-XRD Pathways Rates Mobilities Pathway Prediction Approach

14 Approach to Developing Phase Diagrams DATA Thermochemical & Equilibrium Structure Literature Measurements EMF, Equilibration First Principles Estimation Critical Evaluation, Model Selection & Assessment Bulk & Point Defects Cu-Ga-In-Se System 4 unary sub-systems 6 binary sub-systems 4 ternary sub-systems Equilibrium Phase Diagrams

15 Comparison of Calculated Cu-Se Phase Diagram with Experimental Data Phase Model Liquid Ionic two sub-lattice model (Cu+1,Cu+2)p(Se-2,Va,Se)q α-Cu 2-x Se Sub-lattice model (3 sub-lattices) (Cu,Va) 1 (Se,Va) 1 (Cu) 1 β-Cu 2-x Se Sub-lattice model (3 sub-lattices) (Cu,Va) 1 (Se,Va) 1 (Cu) 1 Fcc (Cu) Regular solution model Se 2 Partial Pressure V Cu in Cu 2-x Se W/Cu,Cu 2 O//YSZ//Cu 2-x Se, Cu 2 O/C/W [Cu] ↔ Cu [Cu 2-x Se] ln a Cu = -FE/RT

16 Phase Equilibria in Cu-In-Ga-Se System

17 RegionEquilibrium phases 1  -CISe 2 +  -Cu + β-Cu 2 Se 2  -CISe 2 +  -Cu + Cu 7 In 3 3  -CISe 2 + Cu2In + Cu 7 In 3 4  -CISe 2 + Cu2In + In 4 Se 3 5  -CISe 2 + InSe + In 4 Se 3 6  -CISe 2 + InSe + δ-CuInSe 2 7  -CISe 2 + β-CuIn 3 Se 5 + δ-CuInSe 2 8  -CISe 2 + β-CuIn 3 Se 5 + Liquid 9  -CISe 2 + β-Cu 2 Se + Liquid Isothermal section at 500  C (18 phases) Isothermal section at 500  C (18 phases) Phase Diagram of Cu-In-Se

18 x(-CIS, ) Chemical Potential (kJ/mol) Cu In Chemical Potential Diagram Cu InSe / Se Reg.Equilibrium phases 1  -CISe 2 +  -Cu + β-Cu 2 Se 2  -CISe 2 +  -Cu + Cu 7 In 3 3  -CISe 2 + Cu 2 In + Cu 7 In 3 4  -CISe 2 + Cu 2 In + In 4 Se 3 5  -CISe 2 + InSe + In 4 Se 3 6  -CISe 2 + InSe + δ-CuInSe 2 7  -CISe 2 + β-CuIn 3 Se 5 + δ-CuInSe 2 8  -CISe 2 + β-CuIn 3 Se 5 + Liquid 9  -CISe 2 + β-Cu 2 Se + Liquid -CuInSe 2 7 T = 500 C InCu Se Wide molecularity range yields 18 to >20% devices Cu/(In+Ga): 0.95 to 0.82 and Ga/(In+Ga): 0.26 to 0.35

19 Graphite Dome Precursor sample Selenium powder X-ray tube PSD Chamber In Out (N 2 ) Kapton/Be window CW in out Sample holder Panalytical Philips X’pert System Surrounding Heater Precursor sample Selenium powder HT-XRD System

20 BELLJAR LOAD LOCK Sorption pump Ventury pump Mechanical pump Mechanical pump TMP Diffusion pump  Ultra high vacuum system  Operating pressure : ~ Torr Ga Cu In Heater Zone MetalDepositionZone Load Lock Zone Chalcogen Zone (Se) Plasma cracker Thermal cracker Na  Rotating platen with 9 substrates (22 inches)  Sequential deposition Schematic top view of MEE reactor UF PMEE Reactor

21 Glass/Mo/-In 2 Se 3 /Cu 2 Se: RTXRD (100) (102) (101) (202) (201) (121) (122) (205) (300) (305) (209) (220) (110) (211) (221) (060) (215)  - In 2 Se 3  - Cu 2 Se Se Mo (110) Intensity (a.u.) 22 (006) (200) 22

22 Temperature ( o C) InSe CIS(112)  -CuSe 22 CuSe 2  -In 2 Se 3 ( )  -Cu 2 Se ( )  - In 2 Se 3  -CuSe ll ll Mo MoSe 2 + Mo 3 Se 4 25 Glass/Mo/  -In 2 Se 3 /Cu 2 Se/Se: Temperature Ramp Annealing Selenium Overpressure ~ 1 Torr  CuSe+ Se (s) CuSe 2 (s) (180 o C) CuSe 2  - CuSe + Se(g) (315 o C)  CuSe  -Cu 2 Se+ Se(g) (375 o C) SS box

23 Cu-Se System Phase Equilibria log P Se /T (1/K) CuSe 2  -CuSe  -Cu 2-x Se L1L1 L2L2 L3L3 CuSe 2  -CuSe T ( ο C)  -Cu 2-x Se T-x Diagram P-T Diagram L1L1 L2L2 L3L Torr

24 25 Temperature ( o C)  - In 2 Se 3  -CuSe CuSe 2 CIS  -In 2 Se 3 ( )  -Cu 2 Se ( ) (Cu 2 Se) x (In 2 Se 3 ) 1-x InSe CIS(112) CIS(220) CIS(312) Glass/Mo/  -In 2 Se 3 /Cu 2 Se/Se: Temperature Ramp Annealing Graphite dome  -Cu 2 Se + Se (s)  CuSe (140 o C)  CuSe+ Se (s) CuSe 2 (180 o C) CuSe 2  - CuSe + Se(g) (260 o C)  -In 2 Se 3 2  nSe+ Se(g) (275 o C)  -CuSe+ InSe CIS (275 o C)  -CIS  -CIS (485 o C) 2  CuSe  -Cu 2 Se+ Se(g) (275 o C)

25 Time(min) T=500 ( o C) In 2 Se 3 + InSe CuSe 22 CIS T=260 ( o C) Time(min) T=270 ( o C) T=290 ( o C) T=280 ( o C) Time(min) Time(min) Isothermal Annealing Glass/Mo/  -In 2 Se 3 /Cu 2 Se/Se

26 B C A B Diffusion thru product & reaction (ex. D BC > D AC ) Solid-State Growth Models A B Nucleation at A-B interface A B Before reaction y y 2 = k p ·t D·k=dydty Parabolic growth model B A Avrami growth model Phase A in matrix B NucleationDiffusion & growth x = 1 - exp[-(kt) n ] ln[-ln(1-x)] = nlnt + nlnk = nlnt + nlnk 0.5 < n < 1.5 (1-D diffusion)

27 ln t ln(-ln(1-  )) α2α2 t (min) Avrami Model Parabolic Model Fractional Conversion Glass Mo  -In 2 Se 3 β-Cu 2 Se Se Cap

28 Glass Mo  -In 2 Se 3  -Cu 2 Se E a =202 (±2) kJ/mole ~ o C E a =194 (±16) kJ/mole ~ o C ln k (1/s) 1000/T (1/K) ln k (1/s) 1000/T (1/K) Avrami ModelParabolic Model Rate Constants Glass CIS Mo 2 minutes

29 Pathways for Binary Precursor Structures In 2 Se 3 (slow) CuSe/CuSe 2 /Cu 2-x Se (fast) Cu 7 Se 4 /CuSe/Cu 2-x Se (slow) In 4 Se 4 /In 2 Se 3 (slow) Ga 2 Se 3 (fast) GaSe (fast)

30 Bilayer Precursor Structure s Glass Mo  -In 2 Se 3  -Cu 2 Se Glass Mo  -In 2 Se 3 Cu Se y Glass GaSe Cu Se Glass Mo InSe Cu Se Glass Mo In-Se Cu Se Glass Mo CuSe In-Se Glass Mo In-Se Ga-Se Cu Glass Mo InSe  -Cu 2 Se Slow Fast

31 Comparison of Interdiffusion in Various Intermetallics in Cu-In-Ga-Se

32 Metal Selenization Precursor Sets SS Mo (0.6µm) Cu (0.1 µm) In (0.216 µm) Ga (0.13 µm) Se (1 µm) Stacked Elemental Layer SS Mo (0.6µm) Ga (0.13 µm) In (0.216 µm) Cu (0.1 µm) Se (1 µm) 6 times 3 times Glass Mo Cu(30 nm) In (70 nm) Ga (5 nm) Glass Mo Cu(10 nm) In (25 nm) Ga (2 nm) *Modulated Structure Slow

33 Temperature ( o C) CuIn Cu 11 In 9 In Cu 9 Ga 4 CuIn x Ga 0.44-x Cu 11 In 9 Mo 22 Glass/Mo/CuIn/CuGa: Temperature Ramp Annealing Indium Melting CuIn Melting Cu 11 In 9 Formation Formation of CuIn x Ga 0.44-x Diffusion of In & Ga Indium textured in

34 Cu-In-Ga Phase Diagram T= 350 o C NoAtomic FractionEquilibrium phases CuInGa Cu 16 (In,Ga) 9, In Cu 16 (In,Ga) 9, In Cu 16 (In,Ga) 9, In, Cu 9 (In,Ga) Cu 16 (In,Ga) 9, In, Cu 9 (In,Ga) Cu 16 (In,Ga) 9, In, Cu 9 (In,Ga) In, Cu 9 (In,Ga) Cu 16 (In,Ga) 9, Cu 9 (In,Ga) Cu 9 (In,Ga) 4 Phase Relationship at 350 o C

35 No CGS formation Temperature ( o C) Mo 22 Aluminum Se In CuIn CIGS CuSe CIGS Glass/Mo/CuIn/CuGa/Se: Temperature Ramp Annealing Indium Melting CuIn Melting Selenium crystallization & melting CIG solid solution formation CIGS formation CuIn x Ga 1-x

36 Temperature Ramp : Glass/Mo/CI/CG+ selenium vapor Temperature ( o C) Mo 22 570 No CGS formation Aluminum In CuIn CIGS MoSe 2 Cu 11 In 9 CuSe CIGS Indium Melting CuIn Melting CIG solid solution formation CIGS Growth CIGS & CuSe formation MoSe 2 Formation

37 Ga Distribution CIS(112) Cu In 0.5 Ga 0.5 Se 2 CGS(112) Cu In 0.7 Ga 0.30 Se 2 22 Intensity (a.u.) CuIn/CuGa/Se CIS(112) Cu In 0.5 Ga 0.5 Se 2 CGS(112) Cu In 0.7 Ga 0.30 Se 2 22 Intensity (a.u.) CuGa/CuIn/Se CIS(112) Cu In 0.5 Ga 0.5 Se 2 CGS(112) Cu In 0.7 Ga 0.30 Se 2 22 Intensity (a.u.) CuIn/CuGa +Se Vapor 22 Intensity (a.u.) CIS(112) Cu In 0.5 Ga 0.5 Se 2 CGS(112) Cu In 0.7 Ga 0.30 Se 2 CuGa/CuIn+ Se Vapor x(Ga) = 46-50% x(Ga) = 44-45% x(Ga) = 44-53% x(Ga) = 39-53%

38 TEM-EDS Glass Mo MoSe 2 CIGS Pt 200 nm Intensity (a.u.) Distance (  m) MoSe 2 MoSe 2 identified (300 nm) and same as previous sample (amorphous) Interface nature could not be determined Ga and In distributed non-uniformly across the thickness direction. More Ga towards the back contact. Gallium Dark Field Image

39 ln k (1/s) 1000/T (1/K) T ( ο C) Metal-Selenization Kinetics minutes

40 Role of MoSe 2 in CIGS Literature suggests MoSe 2 - CIGS interfacial properties sensitive to processing conditions Modifies the contact resistance – – Schottky contact without MoSe 2 – –Ohmic with but high resistance if MoSe 2 too thick or wrong orientation Influences adhesion characteristics – –Poor adhesion without MoSe 2 – –Good adhesion depending on orientation Need better understanding of growth habit o R.Wurz et al., “Formation of interfacial MoSe 2 layer in CVD grown CuGaSe 2 based thin film solar cells”, Thin solid films , 2003, pp o P.E. Russell et al., “Properties of the Mo-CuInSe 2 interface”, Appl. Phys. Lett. 40(11), 1984, pp

41 MoSe 2 Crystal Structure Hexagonal Crystal Structure – –Lattice constants: a= 3.28 Å, c=12.9 Å Strong Mo-Se covalent bonds in Se-Mo-Se layers Weak Van der Waals force between Se-Mo-Se layers Mo Se Mo

42 Temperature ( ο C) 2 MoSe 2 (100) Mo (110) MoSe 2 (110) Temperature Ramp Selenization of Molybdenum in N 2 /H 2 Atmosphere Glass Mo N 2 /H 2 Se

43 ln k 1000/T Avrami Model ln k 1000/T Parabolic Model Kinetic Analysis k=4.08×10 5 exp(-101.2/RT)k=1.5×10 5 exp(-101.7/RT)

44 Conclusions Pathways are dependent on precursor structure – –In phase particularly important Most paths are diffusion limited High-rate processes are possible – –Film quality needs assessed – –Liquid phase assisted growth Point defect chemistry helpful (low disordering energy) – –Enhance diffusivity, defect compensation, type- inversion, impurity passivation

45 Raw material sourceMany but depletingElectrical ground Number of species10 2 or more2 (electron, hole) TransportPipe (10 inch O.D.)Wire, metal interconnect (10 -5 inch O.D.) StorageTank (10 6 moles)Capacitor ( moles) Pump10 hp10 -9 hp (bipolar transistor) ControlGate valveFET On-off valveTransistor Check valveDiode ReactionsManyRecombination/generation Flow Rates10 3 moles/s moles/s Unit operations10 4 /mi /mi 2 Cost$10 8 ($10 9 /mi 2 )$10 2 ($10 9 /mi 2 ) Diffusion coefficient10 -2 to cm 2 /s10 to 10 3 cm 2 /s Reaction Rate10 6 1/moles/s /moles/s Comparison between a Chemical Processing Plant and an Integrated Circuit TYPICAL CHEMICAL PLANT TYPICAL INTEGRATED CIRCUIT


Download ppt "Routes for Rapid Synthesis of Photovoltaic Absorber Materials: The Need for Diffusion Data Ranga Krishnan, Gabriel Tong, Chris Muzzillo, Woo Kyoung Kim,"

Similar presentations


Ads by Google