2nd Generation (2G) PV: CdTe, CIS, and a-Si Thin FilmsAnd some reflections on the Federal R&D Program: 2nd Generation (2G) PV: CdTe, CIS, and a-Si Thin Films And some reflections on the Federal R&D Program Ken Zweibel NREL


Listeners & Messages: Listeners & Messages For the “congenital skeptics”: Great potential Good track record; more difficulties behind us than ahead Clear path forward; Take another look! For the “true believers”: Many setbacks Many ongoing, even multiplying risks Don’t take success for granted; we still face challenges of technological readiness For the PV insiders: We know that if we do not see substantially higher public and private funding, we will still need some good luck to succeed in a timely manner


Basics: Basics The solar resource is large and ubiquitous Large enough to meet global energy needs (assuming the proper economics) Universal enough that almost any locale would improve its energy independence by adopting solar Solar is diffuse & intermittent, so components must be cheap: Conversion Transmission Storage Innovations to make it usable for transportation The DOE funding in PV since the energy crisis of the 1970s has been about $50-$100M per year (now about $80M) 2G Thin film R&D has been about 30% of that The least understood fact about PV progress to-date seems to be that its apparent slowness has mostly been a function of lack of money and not a lack of good ideas


2G Thin Films: 2G Thin Films 2G thin films (CdTe, CIS, a-Si) are on a slow, risky, but very clear path to reach economic viability for many purposes (possibly including dispatchable electricity through transportation) Risk and timeliness are purely a function of money – more money: faster, surer results (but ‘not enough money’, high risk – today’s reality) For almost all reasonable cost goals (down to about 3 ¢/kWh DC electricity) today’s 2G thin film technologies are almost certainly capable of meeting them This is what I am here to establish But low cost, possibly long distance (even hemispheric) transmission (north/south, east/west) and low cost storage capabilities need to be developed to optimally use of solar PV for dispatchable electricity and transportation (once PV has reached proper economic levels) This is what I am here to learn about


Two Ways to Consider Future PV Costs: Top Down and Bottom-up: Two Ways to Consider Future PV Costs: Top Down and Bottom-up Learning curve based on current sales and sales projections Possible new learning curve for thin films Bottom-up assessment of components for key technologies & Large-scale manufacturing economies of scale for key technologies All 2G options examined are in early manufacturing right now Which does not mean they don’t need important technical advances and don’t still face major hurdles


First Observations: First Observations X-Si can get there; it is just a matter of time and markets; this is good news and has yet to be absorbed by those who dismiss PV as too costly or unreal New technology is about getting their faster and cheaper


The Point about New Technologies: The Point about New Technologies New technologies starting on a lower cost, lower-sales volume learning curve may get to lower module prices an-order-of-magnitude in sales volume earlier than current x-Si technology Not only would this mean that PV would be cheaper sooner, it means any government subsidies for early PV markets could be an order-of-magnitude less if new technologies can be the vehicle for such subsidies (since fewer modules would have to be subsidized) I.e., there is a great payoff for research progress!


What’s Actually Happening?: What’s Actually Happening? It all sounds nice, but will it be done? Depends critically on ongoing subsidy programs (to create demand) like those in Europe and Japan Flat plate silicon X-Si wafers and ribbons (90% market share) Flat-plate thin films (10%, but growing rapidly) A-Si and thin x-Si; multijunctions of same CdTe CuInSe2 alloys Concentrators – not commercial Si cells (low concentration) GaAs alloy multijunctions (high concentration) 3G concepts


2 Rules of Thumb: 2 Rules of Thumb Two key parameters drive module cost: module efficiency and module manufacturing cost To calculate module cost, combine them: so, {($50/m2) / (100 Wp/m2)} = $0.5/Wp At the system level, $1/Wp is about equivalent to 5-6 ¢/kWh levelized cost electricity in an average US solar location (Kansas City)


The “Idea of Thin Films” Carries the Weight: The “Idea of Thin Films” Carries the Weight The primary idea is a tiny amount of expensive material (1 micron or so) and lots of cheap glass and wire and metal and plastic But even these 2ndary materials add up… Semiconductor choice mainly comes in, in terms of Light-to-electricity conversion efficiency Ease of manufacturing, including yield and reproducibility Thickness, materials use and thus materials costs Outdoor reliability (intrinsic stability and sensitivity) And possibly in terms of substrate and packaging choices (e.g., temperature extremes) (All of which isn’t exactly chicken feed…)


Let’s Look at Potential: Let’s Look at Potential For lowest module cost For highest module efficiency


Visualize a Cheap Thin Film Module: Visualize a Cheap Thin Film Module A piece of window glass on top A piece of glass or plastic on the bottom A top contact, which may be a conductive oxide A thin metal back contact A wire and two strips of metal bus bars to get the electricity out A micron-thick pair of semiconductor layers in the middle (this is the only thing people ever talk about!) Some adhesive to stick things together A mounting scheme (rails, frames, glue, etc)


Potential Low Cost Module (NOT now) Given Current Perspectives: Potential Low Cost Module (NOT now) Given Current Perspectives


Observations about efficiency: Observations about efficiency We ‘know’ efficiency – we know the challenges, and we know many of the solutions that work We have dropped numerous other thin films that didn’t work as well There are reasons each of these work well If someone has found a new thin film that is easier to improve, we don’t know about it (tho it’d be something special if true!) These 2G thin films work almost as well as wafer-based x-Si; that matters! Remember, you have to have both pieces, cost & efficiency


Potential High Efficiency Module: Potential High Efficiency Module Direct gap semiconductors can absorb 90% sunlight in 2 absorption lengths – about 0.2 microns (or 0.6 g/m2 of material; @100/kg = $0.12/m2 of material) Due to direct gap, good cells are easy to make, since so much current is produced by e-field drift (within the depletion region) Single junction theoretical cell efficiencies 25% Losses in uniformity, resistance losses, and lost area for modules about 10% - so 22.5% Losses due to average performance at high yield 10% - so 20% commercial module (practical maximum) Today’s best cells 16%-20%, so a more conservative potential might be 20% less than that – 13%-16% (best modules already at 11%-13%, yet several years behind in terms of optimal designs; so maybe this isn’t high enough…)


Where does that put our potential?: Where does that put our potential? Combining best efficiencies (13%-16%) with best case costs ($27 to $35/m2): 27 ¢/W to 18 ¢/W modules If system adds 25 $/m2 area-related and $0.2/W power related, total system cost is then about 50 to 70 ¢/W installed, or 2.5-3.5 ¢/kWh (AC). That’s all based on reasonable, physically defensible projections of today’s technologies. And it’s underway, now…


2G Commercial Status: 2G Commercial Status Three thin films (a-Si, CdTe, CIS), either in commercial or pilot production Commercial module efficiencies marginal-to-good (6%-11%) Costs are not optimized at all (no volume, and short-cuts taken to start production)! Critical technical challenges remain because exponential volume increases depend on solving numerous problems on the fly Lack of science and technology base hinder success – but where’s the money to do this?


Amorphous Silicon: Amorphous Silicon


Cadmium Telluride: Cadmium Telluride


Copper Indium Diselenide: Copper Indium Diselenide


Strengths and Weaknesses: Strengths and Weaknesses


More Aggressive 2G Research (than we currently have money for): More Aggressive 2G Research (than we currently have money for) Doing anything more than as a proof-of-concept (i.e., repeatability) Like crossing a stream on slippery rocks – we’ve done it once, but… Making ultra-thin layers to reduce materials, energy, capital & maintenance costs Finding new ways to deposit layers with minimal equipment costs Reducing deposition temperatures to allow for cheaper packaging; experimenting with breathable packages that use ultra-low-cost materials Developing low-cost, flexible, nonconductive substrates capable of enduring high temperatures and outside conditions Minimizing module-level efficiency losses Adding inexpensive (“paint-on”) bottom cells Building up the science base to prevent stupid failures in manufacturing and reliability If we could afford to do these: Greatly reduce risk of transition to large-scale manufacturing and the risk of unknown module failures outdoors Take out a large fraction of semiconductor materials costs, equipment costs, energy costs, and packaging costs and reduce total system cost even below best projections given earlier in talk


The Real World: Things We Just Dropped Due to a 20% Shortfall: The Real World: Things We Just Dropped Due to a 20% Shortfall A CIS nanoparticle ink approach with proven cell efficiency and great potential low cost; and which was picked up by a foreign government within a month A multijunction batch process to make a-Si/nano-x-Si thin films at low cost, showing great promise, which may bankrupt the early stage development company we’d supported heretofore About 40% of the fundamental research on CIS, CdTe, and a-Si we’d supported at universities Despite great excitement and new private money entering the field (and about 40 pre-commercial companies starting work) we picked up none of them This is the real world background with which we view this much desired but rather “fantastic” (in all senses) request for a third stream renewables development program


Issues and Risks: Issues and Risks


Issues and Risks - Simplified: Issues and Risks - Simplified We are trying to affect the world’s energy infrastructure – the biggest industry on Earth We are doing it on a shoe-string and until recently with little private sector participation or marketplace demand We didn’t have enough money, so we did everything only well enough to demonstrate proof-of-concept – not robustly or reproducibly Inherent technical risks (and lack of a foundational science base) mean nature’s complexities could still destroy technologies during exponential manufacturing growth and initial, perhaps unreliable, outdoor deployments!


The Path Forward for PV: The Path Forward for PV Continued and spreading purchase subsidies Emulate the European subsidy model – ¢/kWh subsidy for actual electricity production (stimulates the greatest competition through the greatest payoff for the cheapest PV system) Realize this is a ‘global commons’ issue (environment, growing shortages, and energy security), not a normal, marketplace driven opportunity; governments must play a key introductory role, as they did with wind energy Continued and growing private investment despite the uncertainties of this unusual kind of product Stimulus from Federal R&D funding Much increased support (at every level) of PV-device science and technology Address the rest of the picture: storage, transmission, and transportation (fuel conversion or on-board electricity storage) In summary – more money, more consistency


Some Other Real Problems for PV Funding: Some Other Real Problems for PV Funding It’s hard to get people to be proactive about PV, because they can’t picture its success – ‘what is a PV?’ is still most people’s attitude Federal funding tends to follow trends, not lead them, unless experts come forward PV is not mainstream in science – universities don’t have ‘schools of PV’ – mainstream scientists don’t publish in PV journals – PV science is a step-sister PV needs subsidies and gov’t R&D funding, but Americans are ultra-conservative about gov’t ‘leading’ anything innovative that seems ‘private sector’ Europe and Japan are eating our lunch (because their energy and environmental problems are more immediate; and they are less negative about Federal support); even if we innovate, they just buy the companies because they can ‘see the money’ Leaky roof syndrome…


The “leaky roof” syndrome: The “leaky roof” syndrome People think about PV the way they think about a leaky roof! When it’s raining, they’re bothered – when energy prices are high, they care about renewables like PV – they want a solution, NOW! When the crisis goes away, they forget – when it stops raining, they don’t notice the leak (duh!) Fixing the roof is painful, irritating – easy to forget, ignore. Next time it rains, the roof’s not fixed; “Where’s the PV!” they exclaim, disconcerted – some may even think the roof can’t be fixed – “it’ll never work – they’ve been working on it for 20 years, and it’s still not ready!” We need to work hard and spend money ALL the time on developing PV as a solution to global issues, not just during an energy emergency (i.e., “when it’s raining”) – or else, next time it’s raining, the roof will still be leaking!


System Cost Definition: System Cost Definition A system cost is made up of the module cost plus an area-related system cost (like mounting, land, wiring, installation) and a power-related cost (power conditioning, inverter) and O&M So the most important parameter in a PV system is efficiency, because it drives cost via the module and the area-related costs