Friday, November 10, 2006

Investment in Photovoltaic Complexity

As my last two posts illustrate, I'm in a bit of a rut on the potential for photovoltaics to solve all of our problems. It would be overly simplistic to boil this argument down to "Roddenberrys" vs. "Doomers," because those terms (the former of which I just coined, so bear with me) seem to only address irrational behavior by individuals at either extreme end of the techno-optimism spectrum. There is a very genuine debate underway in what I consider the more moderate middle of this spectrum. Two people whom I enjoy reading, and who's writing I respect, are on either side of my personal view of this, but both within the "reasonable" zone: Big Gav from Peak Energy (leaning towards Roddenberry), and Ted Heistman from Freerange Organic Human (leaning towards Doomer).

At the end of the day, the information available suggests to me that the Energy Return on Energy Invested (EROEI) for photovoltaics is less than or about equal to 1:1. If I'm wrong, and it is more like 10:1 and will steadily rise indefinitely with futher research, then a strong case for "Star Trek" optimism (and hence "Roddenberrys") can be made. There is no doubt in my mind that improvements in photovoltaics will be made--the real question is whether the return on these investments in technology (in complexity) will provide linear returns, or whether they will be subject to diminishing marginal returns. Here is a recent project from the Solar 2006 convention in San Jose last month:

SunPower is approaching a 23% efficient PV. This helps it take business from typical 17% efficient PV. Dr. Richard Swanson, CEO, SunPower gave the conference good reason to expect continued high growth. He pointed out that in 1975 solar modules cost $100/watt. By 2002, the cost had fallen to $3 per watt. The industry learning curve of 30 years has been consistent – each time that production doubles, cost drops 81%. Dr. Swanson expects $1.40 per watt by 2013 and 65 cents per watt by 2023.

This Dr. Swanson of SunPower is making the case for a kind of "Moores Law" for improving solar panel efficiency. Is this really a linear decrease in cost? Right now there is about 5300 MegaWatts of installed PV capacity worldwide. The US alone currently generates on average over 1,000,000 MegaWatts of electricity (just electricity...this won't power a hydrogen fuel scheme). So existing PV would need to double eight times in order to just match the current US electrical generation. If, per Dr. Swanson's linear decrease in cost projection is true, after eight doublings in capacity PV cells would cost 0.005 cents per watt. You could build enough PV to power the entire United States for $5,095. Something tells me that the reality is not linear! No, the reality is most likely best expressed by some form of logistics curve, such as the diminishing marginal return curve suggested by Joseph Tainter:

The rapid increase in efficiency of photovoltaics coupled by the decrease in cost per watt from the '70s to the present is represented by the return on investmen tin complexity, which rises rapidly from 0,0 to C1,B1. The salient question is: what point on the curve represents the return on current marginal investment in PV complexity? Probably somewhere between C1,B1 and C2,B2. Projections, like Dr. Swanson's that assume linearity in cost decrease per watt are basing this assumption on the roughly linear increase represented by the curve between 0,0 and C1,B1. But the reality is that the benefit from each marginal investment in photovoltaics at this point will return less and less. This technology cannot save us from the ultimate ramifications of diminishing marginal returns.

There is, in fact, some evidence that PV technology is already at the peak of the diminishing marginal return curve (C2,B2). Sunpower, the same company where Dr. Swanson extolls the historical decreasing cost of photovoltaics, recently made this press release:

...Overall, these changes result in a 43 percent increase in power, said Julie Blunden, vice president of external affairs at SunPower. Each panel can generate 315 watts of electricity and will have roughly the same cost per watt as the existing line, she said.
When you improve "efficiency," but the cost of doing so keeps the cost per watt stagnant, then you have peaked on the diminishing marginal returns curve. Future increases in efficiency are most likely possible, but they will become so costly as to actually increase the cost per watt. Investment in complexity is inelegant, and will always run into exactly this problem...


Big Gav said...

Interesting analysis Jeff - hopefully Tainter's theory isn't an immutable law of technological development.

While I'm still going to sit on the fence regarding solar (I quite like the Roddenbury tag), I'd be interested to see the result of the same sort of analysis applied to thin film solar (which seems to be much less mature in its development than PV) - the NREL estimates that you've criticised in your earlier post are significantly more optimistic about thin film.

Ditto for solar concentrator style solutions for that matter - my enthusiasm for solar power isn't based purely on PV - and the technological complexity of focussing a large group of mirrors onto a central point seems pretty low.

The same goes for Stirling engine based solutions.

(Sorry - I know I'm lazy and should go and work out the maths myself but I won't have time tonight and you're on a roll with this stuff).

Anonymous said...

Hi Jeff,

My gut feeling is that your er/ei figure of 1:1 for PV is about right. To really get these figures right you have to start thinking about things like the pavement and tires that trucks ride on. There is a whole industrial support infrastructure that things like PV relie on. This entire infrastructure must be accounted for to arrive at an accurate calculation.

As an aside, in responce to my earlier comment, you seemed to indicate that an er/ei calculation is essentially equivalent to a CC type calculation. Could you elaborate or explain your reasoning on this sometime?


Jeff Vail said...

Dave: Sorry if I suggested earlier that EROEI is the same as carrying capacity. I didn't mean to say that. What I meant to say is that EROEI is the critical value within the context of hierarchal societies, because they depend on surpluses to support the overhead of hierarchy. Within a decentralized system, carrying capacity may be the more important number, and in such cases it may be viable to employ a technology with an EROEI of less than 1. For example, more human energy may go in to building an adobe solar cooling tower for a home than it returns, but that is viable because it is not possible to directly convert human energy into cooling air...

Big Gav: I'm with you in that I'm more optimistic with regard to solar concentrators or Stirling engines--I think that they can potentially be very low tech, and may have an EROEI of well above 1. Ideally we would begin rearching with EROEI as one of our measures of success. Along these lines, I'm even more optimistic about passive solar design. Sure, some applications require electricity (communications is perhaps the best example), but most of what we actually use electricity for, especially heating and cooling, can more easily be accomplished by direct conversion of solar energy into heat, without the loss in efficiency by converting energy first to an electicity middleman. In these applications--home, food, and water heating and cooling--I think that decentralized, passive solar has extraordinary promise. Technologies like annualized geo-solar need serious work, but have great promise. I've long seen solar-box ovens, but I'm surprised that we don't use solar heated oil coils for both stove and oven uses--as has been in use for decades at the Gaviotas development in Colombia. If we all only used electricity for things like communications, we could get by with a DC-only system that would run off a small PV panel, or a very simple wind generator. This is what leads me to be cautiously optimisitc about solar in general, but to be pessimistic specificaly with regards to photovoltaics. I think that the tendency of our society to focus on PV, because it provides the illusion that we can just keep motoring along as we have become accustomed, is symptomatic of our greater problem that we (as a society) are generally unwilling to reconsider our basic assumptions. If we can't get past that hurdle, I don't think that we have much chance...

And just a side note, I received a mailer today from Real Goods Solar, offering to install a photovoltaic system on my roof. The flyer advertises a 3.15KW grid-intertie system installed for $28,350. That's $9/watt installed (no batteries). Electricity only costs about 7 cents per KW-hour here in Denver, and we are in a pretty good solar zone, but this still only works out to a price-estimated EROEI of 0.5:1. I'm not including the pretty generous local and federal tax credits, which total over $17,000 (making the actual cost to me out of pocket not bad at all... I may consider it in a year or two after I get a new roof).

halcyon said...

Good stuff.

Some pointers (sorry, if old news):

Prelim. Protocol for determining EROI (ASPO USA 2006)

(I know it's difficult, but I also think that economics based ballpark estimates can be several orders of magnitudes wrong, as it has been shown that markets/futures do not follow physical reality in terms of availability/usefulness of energy sources).

Recent Technology Review had also some new info on concentrating photovoltaics:

I've been trying to find sources that discuss the following, but so far have found none:

1) areal scaling for pv, concentrated pv and Stirling engine: how much solar energy plant efficiency varies as a function of installed pv area. I'd be surprised if it remains flat. All sorts of practical issues may arise when one starts to scale to bigger (or smaller) installed surface area. Is there an "optimal" areal size for current and foreseeable technology?

2) physical realities/limits in current pv technology scaling. As you may know, Moore's law is dead and never was much more than a marketing/engineering plan. That is, it was NOT a physical law per se. It was finally and completely killed by tunneling and ensuing leakage current in transistors of diminishing size. That is a physical limit. No amount of dollars/marketing will break through that physical limit, although exotic materials and various bias currents may mitigate the effects a bit. PV efficiency scaling probably has foreseeable physical scaling limits also, due to the technology used.

3) Systemic calculations for the effect of mass scale pv/technical solar energy installation.

Solar radiation is not "free to use for anything". Most of suns's radiation is currently used by inefficient biomass convertors (e.g. chlorophyll), which are about 2% efficient - meaning they require a lot of surface area to get their solar need. Earth's ocean & atmospheric processes are primarily solar energy driven (through thermal conversion). There are probably dozens of other big scale processes I'm not even aware of.

Are there any calculations as to what size/efficiency of pv plants start to have an effect on surrounding natural solar driven cycles?

technician said...

Yes, perhaps Roddenberries and Doomers are both right. It’s just they have discrepancies in there timelines. Doomers might just be proven right first.
When your hear banker starting to say “human interference”, “human computation”, and that they are working on removing themselves out of the flux.
The diminishing of marginal return deeply transforms us already - at the very root of our perception of information.
The result of such transition would give a golden ground 0 for Machines and their collaboz. In a total collapse of human meaning, in an overwhelming victory of silicon there will be no more room for undemonstrable languages.
With no more frontier in-between words like biology and machines, it would be the perfect time for nano and bio technologies to merge, thus propelling marginal return in another dimension. As you said there is a fine line in-between Intel and Monsanto.
If this fine line dies out I betcha solar techs will kick asses!
Literally speaking.

norway calling said...

Analyzed right to the point.. Whats more is that PVs are subject to depletion . prone to be substituted by . say 290 years or so - PVs will be a costly business from concumers point of view...
Think Again

norway calling said...

... yeah rigth - beerspill on keyboard - 290 years is "the dream" - but 20 years is the reality..

Theo_musher said...


I really appreciate your point of view and your brilliant insights. I've been reading your book again. I am serious about trying to overcome a doomesday mindset.

So I have been really challenging myself to shift my thinking. Its leading me in some new directions. One direction, is for me to be a bit more of a realist than an idealist. I mean, sure it would be nice to live naked in the woods as a hunter gatherer in a tight community living in harmony with the Earth. Would this involve illiteracy and mathematics dying out in a few generations along with global communications and free inquiry in general?

These are some of my dilemmas. What do I owe as human being to my fellow 6.5 billion human beings I share the planet with? What do I owr to Western Culture of whom I am a product?

How long is this crash going to take? How will it unfold? What if many aspects of our society remain the same for many years?

As I think about these things I am coming to appreciate more and more many of your economic insights and mathematical eguations.

It seems to be part and parcel of a more universal approach.

After all whether we like it or not this is a global community we live in, what effects one effects all. I just thought I would share that.

Anonymous said...

Theo: What does the broiler owe to Western Culture of ehich it is a product?

piers said...

Anonymous: the broiler is screwed, yes, but we, for example, are all using the tools of western enlightenment culture in having this discussion. Not just the internet, but the very objective intellectual approaches we're taking. The ability to consider different cultures, to take this huge overview of history and prehistory, to consider options in ways of life, and so on and so on - all these are the products of western civilisation. They may not outweigh its huge bad things, but could be considered to be good in themselves. True knowledge about the universe (eg, why planets move as they do) is good in itself, if you ask me. The cost-benefit analysis is another question.

So pointing out the crappy results of western civilisation doesn't actually negate its achievements, which are unique, and many are valuable in themselves. They are still valuable in themselves even if the overall cost-benefit analysis concludes that it wasn't worth it (which it probably wasn't). But for that, you have to make value judgements, like happiness is more important than true knowledge.

Theo_musher said...

I also wonder if knowledge can be seperated from a desire to dominate. Seems like the larger the sphere of knowledge the larger the sphere of dominance over nature.

I think primitive man seeks to dominate nature as well, but takes maybe a different path and a more unconscious one.

Maybe every culture is a "taker" culture maybe some cultures are just mor effective at taking or maybe I should say "efficient" But then knowledge spreads until we are really able to take stock of the true implications of our actions.

To me these are some of the implications of "A Theory of Power"

jallabo said...

Your assumed eroi of 1:1 seems very pessimistic. With your method using only prices you can certainly establish an credible estimate for a lower boundary, but unless you are going to argue that almost 100% of our economic activities are consisting of energy transactions, the eroi for already existing PV modules should be already considerably higher.
For a quick and dirty estimate of energy input into the general economy: The world GDP is around 40e12$, but the world oil bill priced at around 60$/bbl would be only around 1.8e12$, the gas bill at 10$/mBTU 1e12$, the coal bill at 60$/ton around 300e9$ (All assumed prices much too high if you look at world averages). Add on top of it hydro and nuclear and you are still looking at an energy contribution to World GDP of less than 10%. And much of it is spent on activities not related in any ways to manufacturing. So if you assume an energy input of maybe 25%-50% in your price calculation your are already looking at erois in the range of 1:2 - 1:4.
Considering that the profit margins of the producers are right now in range of 30%-40% due to demand far outpacing supply on the world market (mainly due to demand from Germany, where you get 0.4-0.5€ for every kwh generated, but France, Spain and Italy have recently enacted similar laws), the true number is probably even higher.

Paraclete said...

I have a very simple question: Can someone provide me with a solar industry accepted standard math formula for calculating energy conversion efficiency for solar cells?

Thank you.

Jeff Vail said...


I am arguing precisely that 'almost 100% of our economic activities consist of energy transactions.'

Your $40e12 figure is misleading. $25e12 of that consist of financial derivatives--magic money--that result from the production of absolutely nothing. Another unknown but significant portion of that is real-estate--when we speak specifically of the land itself, not the built environment, this is also an entirely fabricated value. And your energy calculations are not accounting for human energy input. I'm not talking about the Watts of energy input by humans, because as your own analysis above shows, different energy inputs have differential value. Human energy input, because it is so finely controlled and channeled compared to, say, gas, is an exceptionally large chunk of that $40e12. Our market economy reflects this in terms of "labor costs." So, with this taken in to accoun, yes...I think it is almost all energy.

fdg said...

I like your blog. Thank you. They are really great . Ermunterung ++ .
Some new style Puma Speed is in fashion this year.
chaussure puma is Puma shoes in french . Many Franzose like seach “chaussure sport” by the internet when they need buy the Puma Shoes Or nike max shoes. The information age is really convenient .

By the way ,the nike max ltd is really good NIKE air shoes ,don’t forget buy the puma mens shoes and nike air max ltd by the internet when you need them . Do you know Nike Air Shoes is a best Air Shoes . another kinds of Nike shoes is better . For example , Nike Air Rift is good and Cheap Nike Shoes .the nike shox shoes is fitting to running.

Spring is coming, Do you think this season is not for Ugg Boots? maybe yes .but this season is best time that can buy the cheap ugg boots. Many sellers are selling discounted. Do not miss . Please view my fc2 blog and hair straighteners blog.
.thank you .

I like orange converse shoes ,I like to buy the cheap converse shoes by the internet shop . the puma shoes and the adidas shoes (or addidas shoes) are more on internet shop .i can buy the cheap nike shoes and cheap puma shoes online. It’s really convenient.
Many persons more like Puma basket shoes than nike air rift shoes . the Puma Cat shoes is a kind of Cheap Puma Shoes .
If you want to buy the Cheap Nike Air shoes ,you can buy them online. They are same as the Nike Air shoes authorized shop. Very high-caliber Air shoes and puma cat shoes . the cheap puma shoes as same as other.

polo shirts

ralph lauren polo shirts
chaussure puma

chaussure sport

chaussures puma

puma CAT

ed hardy clothing

ed hardy clothes

ed hardy womens

ed hardy sunglasses

fdg said...