Download presentation
Published byTyrone Walker Modified over 9 years ago
1
Emerging High-Efficiency Low-Cost Solar Cell Technologies
Mike McGehee Materials Science and Engineering Center for Advanced Molecular Photovoltaics Bay Area Photovoltaic Consortium Precourt Institute for Energy Stanford University
2
DOE’s Sunshot Goal: $1/W by 2017
A module efficiency of 25 % is wanted to enable the installation cost reductions. Source: US DOE report “$1/W Photovoltaic Systems,” August 2010.
3
Last Week’s Lecture by Anshu Sahoo and Stefan Reichelstein
Silicon is currently competitive in favorable locations with the current subsidies. In 2017 the cost of silicon cells will probably be $0.65/W.
4
There are many approaches to making PV cells and experts do not agree on which one is the best
20x-100x 500x Cu(In,Ga)Se2 ~ 1-2 um c-Si ~ 180 um
5
The term ‘III-Vs’ refers to one or more of the group 13 elements on the periodic table (Boron, Aluminum, Gallium, Indium, Thallium, and Ununtrium) combining in some combination with one or more of the group 15 elements (Nitrogen, Phosphorous, Arsenic, Antimony, Bismuth, and Ununpentium). The nomenclature with the Roman Numerals III-V is actually historic, as the group numbering for those elements was formerly III (now group 13) and V (now group 15) under the old International Union of Pure and Applied Chemistry (IUPAC) nomenclature. In this materials set, the Group 13-to-Group 15 atomic ratios total 1:1. Examples include GaAs, Ga0.5In0.5P, and GaAsxP(1-x). Efficiency Records Chart available at: The chart above was downloaded on 2/27/2014.
6
Silicon PV Silicon Feedstock Ingot Growth Slicing Wafers
Cell Fabrication Photovoltaic System Module Encapsulation
7
19.6% efficient planar cells on silicon
Source: J-H Lai, IEEE PVSC, June 2011
8
Cost analysis of Si Modules
LOS is line of sight Tonio Buonassisi et al., Energy and Env. Sci. 5 (2012) p
9
Tonio Buonassisi et al., Energy and Env. Sci. 5 (2012) p. 5874.
10
Tonio Buonassisi et al., Energy and Env. Sci. 5 (2012) p. 5874.
11
Cost analysis of Si Modules
Tonio Buonassisi et al., Energy and Env. Sci. 5 (2012) p
12
Thin Crystalline Silicon Cells
14
Using nanocones to enable complete light absorption in thin Si
Sangmoo Jeung, Mike McGehee, Yi Cui, Nature Comm. in press.
15
Gallium Arsenide The 1.4 eV band gap is ideal for solar cells.
High quality films are grown on single crystal substrates with MOCVD.
17
Photon recycling GaAs Si
Eli Yablonovitch et al., IEEE J. of Photovoltaics, 2 (2012) p. 303.
18
Why thin film GaAs is better
Remitted photons are weakly absorbed and can easily travel more than a carrier diffusion length away from the junction in a wafer-based device. In a thin cell, a mirror keeps photons near the pn junction. Eli Yablonovitch et al., IEEE J. of Photovoltaics, 2 (2012) p. 303.
19
Theoretical limits Eli Yablonovitch et al., IEEE J. of Photovoltaics, 2 (2012) p. 303.
20
A Manufacturing Cost Analysis
Relevant to Photovoltaic Cells Fabricated with III-Vs The National Center for Photovoltaics Aaron Ptak David Young Scott Ward Myles Steiner Pauls Stradins John Geisz Daniel Friedman Sarah Kurtz Graphics and Communications Alfred Hicks Kendra Palmer Nicole Harrison Michael Woodhouse Alan Goodrich Additional NREL Contributors: September 30, 2013 Publication Number: NREL/PR-6A Contract Number: NREL: DE-AC36-08GO28308 Strategic Energy Analysis Center Ted L. James David Feldman Robert Margolis Hydrogen Technologies Center (Photoelectrolysis Interest) Todd Deutsch Henning Doscher John Turner
21
An Example Process Flow for Making Single-Junction III-V Devices by ELO
Ongoing NREL Analysis 9/13/2013 The model process flow that we use for our single junction GaAs cost models is shown above, and was conceived in consultation with NREL researchers (including those listed on the Title slide), industry collaborators, and after an extensive literature survey.
22
Step 1: Unpack and Clean GaAs Parent Epi-Substrate—3
(The Reference Case Scenario in the Bar Chart Assumes 50 Reuses and 70% Yields) 1. The cost analysis represented on this slide shows the sensitivity of manufacturing costs to the number of substrate reuses and to a fixed CMP cost of $8/repolish. Note that the translated $3.4/W repolishing cost is based upon an assumption of a CMP step in-between each growth cycle and an assumption of 70% yields. Again, if there can be multiple growth cycles between each CMP, that cost will scale proportionally. 2. Also included in the bar chart are representative costs for visual inspection and microcrack detection of incoming wafers, which are based upon cost-of-ownership inputs borrowed from our c-Si PV manufacturing cost models. For a description of those costs, the reader is referred to: A Goodrich, P Hacke, Q Wang, B Sopori, R Margolis, T L James and M Woodhouse “A Wafer-Based Monocrystalline Silicon Photovoltaics Roadmap: Utilizing Known Technology Improvement Opportunities for Further Reductions in Manufacturing Costs.” (2013). Solar Energy Materials and Solar Cells 114; pp. 110–135. Please see the annotated notes below this slide for more details.
23
Cost Summary, by Step, for the Reference Case.
(20 substrate reusages, precursor utilizations of 30% for the III- source and 20% for the V- source, 15 mm/hr GaAs, 70% effective cell yield ) This slide shows the costs for each step in the ‘Reference Case’ scenario for single-junction GaAs. For steps 2-8 of the model process flow, the times for MOVPE reactor chamber atmosphere preparation, temperature ramp up, temperature ramp down, and maintenance are all included in the GaAs base layer step only. Please see slide 16 for the specific times assumed. Of course, the total calculated depreciation expense for the GaAs base layer step also depends upon the deposition rate and thickness. The total average cycle time (TACT) is, therefore, highest in that step and is inclusive of more factors than what is included in steps , 7 and 8 of the model process flow. Please see the annotated notes below this slide for more details.
24
Technology Roadmap Simulations for Single-Junction III-V’s (GaAs Base)
This slide again shows the sum of costs for the ‘Reference Case’ scenario, and also highlights select technology pathways to lower manufacturing costs from that scenario. In the Mid-Term and Long-Term cases, we highlight what are currently believed to be practically achievable targets in cell processing—based upon numerous conversations with NREL researchers and industry collaborators—for solid III-V films deposited by MOVPE.
26
What can be done to bring the costs down?
Huge breakthrough in reducing materials deposition cost. Light trapping to reduce film thickness. See Jim Harris’s 2013 GCEP talk at Use concentrators. With epitaxial liftoff, 500 X concentrators might not be necessary. Trackers for 10 X concentrators are relatively cheap.
27
Cadmium Telluride Solar Cells
CdS/CdTe glass Direct bandgap, Eg=1.45eV High module production speed Very inexpensive 20.4 % efficiency Image from Rommel Noufi Schematic from Bulent Basol
28
CdTe: Industrial Status
First Solar is the leader. It takes them 2.5 hours to make a 13.4 % module. Average Manufacturing Cost 2006: $1.40/watt 2007: $1.23/watt 2008: $1.08/watt 2009: $0.87/watt 2010: $0.77/watt 2011: $0.74/watt 2012: $0.64/watt 2013: $0.53/watt The energy payback time is 0.8 years. greentech media
29
Efficiency limits Sources of energy loss
Thermalization of excess energy Efficiency limits CB Below band gap photons not absorbed VB Increasing VOC and decreasing JSC
30
There are lots of 3rd Generation ideas to beat the Shockley-Quiesser limit, but only one that works.
31
Multijunctions: The Road to Higher Efficiencies
Higher-efficiency MJ cells require new materials that divide the solar spectrum equally to provide current match Ge provides lattice match but the bandgap is too small GaInNAs 1.0 eV Ge 0.7 eV The cells are in series; current is passed through device The current is limited by the layers that produces the least current. The voltages of the cells add The higher band gap must see the light first. New Solar Junction MJ Cell ConventionalMJ Cell
32
4-junction cell with 44.7 % efficiency at 297 suns
33
Multijunction Cells are Very Expensive
These complex structures are grown very slowly under high vacuum. 37 % cells can be purchased for $50,000/m2 Concentrating the light is essential. Ga0.50In0.50P: Top Cell Ga0.99In0.01As: Middle Cell Ge substrate: Bottom Cell R.R. King; Spectrolab Inc., AVS 54th International Symposium, Seattle 2007
34
Concentrating Light Dish Shape
It is possible to track the sun and concentrate the light by 500X Dish Shape Sol Focus 34
35
Hybrid Tandems Are Intended to be a High-Performance Low-Cost Option
Efficiency Cost Organic 12% efficient $30/m2 Hybrid 30% efficient $100/m2 Epitaxial Crystalline 45 % efficient $40,000/m2 Established Technology: Silicon or CIGS Eg ~ 1.1 eV Low Cost Defect-Tolerant Technology: Perovskite, Organic, Nanowires or II-VI Eg ~ 1.9 eV
36
Surface recombination velocity = 170 cm/s
Science 339 (2013) p
37
Stion, Khosla-Funded PV Startup, Hits 23
Stion, Khosla-Funded PV Startup, Hits 23.2% Efficiency With Tandem CIGS Greentech Media February 24, 2014
38
‘Perovskite’ Describes a Crystal Structure Class
Generic formula: ABX3 CH3NH3PbI3 CH3NH Pb I- A B X “Perovskite” is the name of the mineral CaTiO3 but this term is being used for all compounds with the general formula ABX3 that have the same crystal structure as CaTiO3 or are derived from this structure. These materials consist of two cations, the cation A is 12 fold coordinated by the anions X and the cation B 6-fold where X can either be oxygen or a halide. The most prominent perovskite in solar cells right now is Methyl ammonium lead iodide. It is worth noting that the structure of methyl ammonium lead iodide deviates from the ideal cubic perovskite structure as the octrahedra become slightly tilted and the structure consequently tetragonal. Methylammonium-lead-iodide
39
Perovskite Solar Cells are Soaring
mesop. TiO2 2-step solution deposition MSSC, Al2O3 solution-processed 15.9% 15% 15.4% evaporated thin film PIN Another jump in efficiency was reported by the Gratzel group last month. By modifying the perovskite absorber deposition procedure, into a 2-step deposition route, Julian Burschka and co-workers were able to increase the light harvesting efficiency of the devices while decreasing the thickness of the mesoporous TiO2 layer to only 350 nm. Year Snaith et al., En. Env Sci. Jan 2014 Snaith et al., Nature Sep 2013 Grätzel et al., Nature Jul 2013
40
Perovskite Solar Cells Evolved From the Dye-Sensitized Solar Cell
41
Perovskites Are Compatible With A Planar P-I-N Architecture
This device was thermally evaporated by the co-deposition of PbCl2 and CH3NH3I. Jsc = 21.5 mA/cm2 Voc = 1.07 V FF = 0.68 η = 15.4%
42
Low Bandgap – q·Voc Loss in Perovskite Solar Cells
Material Bandgap (eV) q·Voc (eV) Energy loss (eV) GaAs 1.43 1.12 0.31 Silicon 0.75 0.37 CIGS ~1.15 0.74 0.41 Perovskite (CH3NH3PbI3) 1.55 1.07 0.48 CdTe 1.49 0.90 0.59 a-Silicon 0.89 0.66 M. Green et al. Solar cell efficiency tables (version 42) July 2013
43
The traps are shallow Yanfa Yan et al., Appl. Phys. Lett. 104 (2014)
44
The Perovskite is a Strongly-Absorbing Direct Band Gap Semiconductor
45
The Perovskite Bandgap can be tuned by Chemical Substitution
The band gap can be tuned from 1.57 eV to 2.23 eV by substituting bromine for iodine in CH3NH3Pb(BrxI1-x)3 For hybrid tandem with CIGS eV We thought about strategies to tune the bandgap. Already in the first report on the methylammonium lead halide perovskites, an approximately linear relationship between the ion radius of the halide and the bandgap was observed with the pure Br compound having a bandgap of 2.23 eV and the pure iodide a bandgap of 1.57 eV. This inspired us to tune the bandgap by making mixed Br/I perovskite compounds…. and found beautiful relations between composition and bandgap. We were very excited of course and slightly disappointed, when our work was published in March by the Seok group. For hybrid tandem on CIGS of Si, we modeled that the ideal bandgap of the topcell should be 1.76 eV. Using this relation between perovskite bandgap and composition, we can now extrapolate that the mixed halide perovskite compounds should have an approximate composition of MAPbBrI2. Eva Unger data by McGehee group and Noh et al., Nano Lett. 2013
46
Hybrid Tandem Architectures
Bottom Cell Rear Contact Transparent Electrode Top Cell Glass Bottom Cell Rear Contact Top Cell Tunnel Junction/ Recomb Layer Transparent Electrode 4 Terminal Easier prototyping No current matching required No tunnel junction or recombination layer required 2 Terminal Fewer layers that parasitically absorb Module fabrication easier
47
Our Semitransparent Perovskite Cells
Colin Bailie, Grey Christoforo
48
Preliminary Cost Estimates
Today’s Silicon Silicon-Perovskite Efficiency 19.4 % 25 % Cost/Area $153/m2 $167/m2 Cost/Watt $0.79/W $0.67/W Expected improvements in silicon technology will take the cost below $0.5/W!
49
Conclusions Conventional silicon leads the solar cell race, but will not take us where we need to go. Several technologies could take over in the next 10 years. We are still discovering new materials with substantially better properties. I think multijunction solar cells will be thin, light, cheap and > 30 % efficient.
50
Final thoughts We have to solve the energy problem.
Any technology that has good potential to cut carbon emissions by > 10 % needs to be explored aggressively. Researchers should not be deterred by the struggles some companies are having. Someone needs to invest in scaling up promising solar cell technologies.
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.