Most of the elements needed for PV cells/frames/wiring can come from the four most common elements in Earth's crust, so there probably isn't a practical limit within the next few hundred years given the projected world population in that context. If we, instead of having our roofs absorb light from the sun and radiate it as heat, do the same with solar panels, with an intermediate step where we convert some of that energy into electricity and back to heat when we use it, we still have the same amount of energy being radiated back into space as heat and so on. What's causing the most trouble is the use of fossil fuels, which changes how much of the energy from the sun is trapped in the Earth's biosphere. Even if we somehow managed to increase the rate of light absorption/radiation as heat to the point where we're trapping as much additional energy as we are now due to GHG emissions, it still wouldn't impact the climate as much as those GHG emissions because the original (w/o GHGs) emissivity of the atmosphere would allow for the vast majority of that energy to escape. The problem with GHGs is that we have this tremendous source of energy bombarding the Earth 24/7/365 and we're changing how much energy we're letting out.

"Most of the elements" is not all of the elements and "the next few hundred years" suggests unsustainable.

I agree that, on the face of it, roofs generating electricity that can do useful work (provided that work doesn't adversely affect our habitat) that ends up as heat is preferable to simply generating the heat in the first place, I'm now very wary of us assuming anything. There are always consequences.

Lesse, we have steel for the mounting hardware, Si for the panels, Al for the larger wiring, what am I missing that isn't present in sufficient amounts? In terms of the limiting factor, it would seem to be Al/Fe for mounting the panels and the wiring since Si is present in the crust in much greater amounts. I suppose we could look at the copper/lead on any circuitry in the panels, but even weighted for relative abundance that's below the limits presented by Al/Fe. Looking at current production of Fe/Al for one year, one quarter of that limiting factor would provide enough material to construct enough panels to supply enough energy for 10 billion people to live at a European standard of living the next forty years or so. The high quality Si, as well as the Al/Fe can all be recycled, so production for just forty years can supply enough in the way of solar panels for the next 6000+ years, not including recycling, which is already at ~70-80% IIRC and would undoubtedly increase in light of scarce supply, extending that to 30,000+ years.

Given that, I find it very unlikely that solar PV will be the limiting factor in sustainability, especially compared to current human impacts. Course, predicting anything past even a hundred years is more or less impossible, so at the very best all we can do is concentrate on being more sustainable, not planning for sustainability to infinity and beyond.

I don't have a full list of materials that go into the panels and the infrastructure to support the manufacture, distribution and operation of the panels and the resultant electricity but I'm sure it is far more varied that you imply. I understand indium is used in many/most panels, for example, and steel requires a carbon source. Ore extraction facilities require resources (and can do environmental damage). Manufacturing facilities require varied resources (for construction and operation).

Solar PV is a limiting factor in sustainability, just as using any limited resource (no matter what the theoretical size of the resource) would be. But solar PV could certainly be a help in providing energy at a limited but sustainable level.

There are more materials, but the bulk of overall materials is in the infrastructure, not manufacturing, unless we're destroying and rebuilding our manufacturing line every year. ;) Transmission is almost all Al/Fe. For steel news on biocharcoal seems to be picking up a bit, although it's not receiving a lot of publicity since coking coal isn't going to be in short supply any time soon, compared to oil at least. Indium is used in thin film panel manufacturing, specifically in CIGS, but thin film is only projected to capture ~30% of the market by 2012, and has been and will represent the minority of panels produced for a while. Even then, it's not like we can't make solar panels w/o Indium, just that at the moment they appear to be the cheapest, at least in bulk, with thin film Silicon not too far behind IIRC. While Indium may present a limit to CIGS based thin film, solar panels in general do not require it, it just happens to allow for the cheapest manufacturing at the moment.

Besides, saying that solar PV is a limiting factor in sustainability because it uses limited resources is kind of pointless since the sun's useful lifespan is also limited for the same reason. I mean, when looking at sustainability, we should be looking at likely (and present in some cases) limiting factors, not trivial ones. Sure, we can't have an infinite number of solar panels, just like we can't have an infinite time with the sun as it is now, but neither one is a meaningful limiting factor in terms of sustainability, at least not yet. If we manage to get to that point that they are likely limits whenever in the future I'll happily concede the fact, but as of right now we have thousands of years worth of other crap that'll provide a limit far sooner than solar panels, or for that matter the sun's lifetime, will.

You keep saying things like "the bulk of" and "almost all", as though they mean the same as "all". If any resource becomes scarce, no matter how little is used, that will define the limit unless there is a ready substitute of at least equal quality.

Even if we don't destroy our manufacturing capability each year (with or without a smiley), to increase the capacity will take extra resources.

Even if Al and Fe are vast resources, theoretically, in practical terms they are more limited. But just because a resource will last (and be able to be extracted at constant or increasing rates) for hundreds of years or thousands of years, or more, that doesn't make it infinite.

You also use throw away terms like "isn't going to be in short supply any time soon" without defining "soon" or citing sources for such optimism.

It is not pointless to point out that PV is not unlimited and has a limit for sustainability. It IS pointless to state the Sun's useful lifespan, since solar energy, per se, is not a limiting factor for sustainability but our diversion of solar energy, and the resources needed to do so, will be limiting at some level. We should not assume that there is no practical limit because that can lead to unsustainable strategies. Why not tread carefully? I don't see what the downside would be other than having to think before acting.