What's the problem with solar energy? What "breakthrough" from a theoretical point of view is needed for solar energy to become feasible energy alternative?
 A: Although I am no big friend of the solar panels today - because they are subsidized - I think it is correct to say that from a theoretical point of view, we are already "there".
Solar panels are typically guaranteed to work for 10 years at their 90% capacity and/or for 25 years at their 80% capacity. 
The price of a piece of solar panel producing 1 Watt may be as small as USD 1.30 today, see
http://www.ecobusinesslinks.com/solar_panels.htm
However, this is a Watt when it's actually running and in average, it's only running at the maximum capacity for 2 hours a day in average. Clouds, nights, and bad angles are a problem. So a Watt panel only produces 2 Wh per day or 730 Wh per year or 10 kWh per 15 years or so which is the estimated lifetime.
So assuming no extra expenses when the panel is already installed, and no expenses for their liquidation, you pay the minimum of USD 1.30 per 10 kWh, or USD 0.13 per kWh. That's actually acceptably cheap - a few times cheaper than the electricity sold in many European countries and pretty much identical to the consumer electricity price in the U.S. No profit.
In reality, countries such as my homeland introduced lots of subsidies and lots of solar panels were installed that were not yet cheap enough; and they were installed in places with suboptimal inflow of solar radiation. The counting implied that the electricity was five times more expensive than what it was sold for - the difference had to be paid from the subsidies.
But I think it's a purely technological question and we're on the edge of having solar panels that are profitable at the current market prices. Of course, you must also find the place where they're installed - on the arable land, you should also add the rent and/or opportunity costs for the food that could be grown up there. It's a barbarianism to use arable land for solar panels.
However, there's a lot of useless space where the panels could work - e.g. Sahara - and in a couple of years or a decade, the solar panels may become OK economically. It's now a job for the engineers to improve the materials and other things. of course, the main problem for solar panels today remain the fossil fuel (especially coal) competitors that are simply cheaper, so offer the energy companies to guarantee a safe and healthy profit, too.
But I want to say that if we ran out of the fossil fuels in some time, there would be no "physics obstacle" and almost no "economic obstacle" preventing us from switching to electricity that could be produced by solar panels. It's plausible that some totally new kind of a material will replace the materials used today - graphene (a layer of graphite, carbon) seems pretty promising at this moment (most existing solar panels contain silicon); it's conceivable that the evolution will be gradual and less dramatic. At any rate, there's no longer any reason to subsidize the solar panels. Quite on the contrary, the subsidies should be removed everywhere in the world to encourage the engineers to make the last steps needed for profitability.
A: The simple answer is price. Solar panels are still highly cost inefficient (particularly in low-sunlight regions) compared to conventional methods of energy production.
To be fair, there is really very little physics to be said about this. Efficiency, while far from 100%, is still pretty decent (typically from 10% to 30%). One interesting area just appearing is organic solar panels, which are significantly cheaper to produce. If the issues of efficiency (already getting to decent levels I hear) and lifetime are improved, then this could be a real boost for solar panels.
A: I don't want this comment to be taken the wrong way, but I would argue that solar energy is already a successful energy alternative in many applications.  I think in large scale commercial markets in industrialized countries, I would agree that solar still has not reduced its costs to the point where it is a "natural" alternative.  However, solar definitely has viability in niche markets, starting with space and trickling all the way down to certain classes of consumer electronics (solar powered calculators and watches for instance).  
Long term cost effectiveness of solar is really a question centered around expected price of fuel in the future.  This problem gets tricky due to relationships in supply and demand and prices people are willing to pay for a commodity. This becomes very difficult to estimate, and puts industrial solar continuously on the threshold of profitability.  However, one economic model is to introduce solar to people who are in extreme environments, who can benefit from having access to electricity, but lack access to existing infrastructure.  Although that market is exceedingly small in industrialized countries, global markets are potentially much larger. 
A: Even in some less extreme environments, solar energy may make sense today. I have some panels on my roof in California, and the fact that peak solar production is highly correlated with peak demand, probably makes them beneficial for the grid as a whole.
But, there are breakthroughs needed to make them a slamdunk. One area is efficiency, the best currently is around 40%, but this are very pricy multijunction cells, which only make economic sense for satelites, or for systems where sunlight is optically concentrated to circa a thousand times the unfocused intensity. I've seen some sciencedaily reports about projects to use transparent conductors to allow severalway junction cells, with perhaps 65% efficiency. If achievable this might make highly concentrated (optically focused) systems cost effective, as well as significantly reducing the landarea needed. We also have more standard (10% to 20%) silicon cells coming down in price -but this is heavily modulated by supply/demand, and currently demand is low. Nevertheless if medium efficiency panels (and balance of system components such as mounting and inverters) can continue to come down in price, the number of areas/applications where they make economic (as opposed to simple feelgood) sense should be on the rise. 
A: Doped layers of graphene on a plastic substrate, or other thin film technologies, should permit one to buy solar cells in rolled sheets.  One would cut out sheets of the stuff, as if one were cutting pieces of wall paper, and then “plug & play” from there  These would be put on roof tops and other places.  Further, these would be lithographically produced, as opposed to using silicon melt or vapor deposition.  This would reduce the cost to maybe as low as \$10US per square meter.  This could be used to solarize lots of “dead” areas on top of buildings and other places.  About 5% of the Earth is covered by human structures, and even if 1% were solarized there could be a considerable solar energy input.  
The energy future has to involve a matrixed approach with a diversity of sources.  Wind energy is playing a growing role.  I have thought that tethered submersible turbines in the ocean conveyer could generate a fair amount of energy.  Then nuclear power will have to be employed to fill in the down times on renewable sources.  By 2030 we will be well down the post peak on oil, and by 2050 the same with coal may hold true.  So we need to get our butts in gear over the energy issue.
A: Since this came up to first page by an edit, I will offer my two cents of the euro.
1) Solar energy has the problem of most renewable energy resources: lack of storage capacity for the energy. Though one could devise a water storage and recovery for off time use, this is not different than having a grid based on stable sources, oil, coal, gas, nuclear, hydroelectric,  and conserving those sources when sun or wind or tides are available. Lack of good long term electric batteries have "solar cars" possible only as demonstrations whereas in countries on the sun belt they could be  a good alternative.
2)It is very expensive to install, and is only for a decade or so. I calculated that to power a house in Greece, where sunshine is plenty, at the level of kilowatt/hrs  that a normal household uses would cost about 40.000euro for the panels at the prices available here.
Both 1 and 2 can be overcome if good cheap batteries are invented, which is a matter of materials science and not theory. If a viable storage exists there will be a demand and the prices will drop.
