|Fig. 1: Multiple junction solar cells can absorb
different parts of the solar spectrum to avoid
The limitation of first and second generation solar cells is that there is a theoretical limit to the maximum efficiency you can get with a single junction cell. This means even if you engineer the perfect solar cell with an ideal band gap and collect all the carriers, you can only get to 30% efficiency or so.  This is a solution to the world energy problem if and only if these ideal single junction cells are cheap enough to manufacture that the efficiency limit is not a problem. This could take 20-30 years for the cost to justify this method and so the most promising technology in my opinion is the third generation tandem multi-junction solar cell with concentrator. The best use of this technology is in large solar farms in deserts that are owned by utility companies.
The theoretical efficiency limit for these cells is much higher than single junction cells. This is because we are absorbing photons that are closer in energy to the bandgap so the thermalization losses are reduced. These cells are set up with multiple (usually ~3) semiconductor materials arranged in different layers so the largest bandgap material is closest to the incident light and the next largest bandgap material is next and the smallest bandgap material is last. This arrangement allows lower energy photons to pass through larger bandgap materials and absorbed by the last material. The different semiconductors have different bandgaps which allow them to each individually absorb different parts of the solar spectrum, as seen in Fig. 1. The smaller bandgap materials absorb the lower energy photons, the higher bandgap ones absorb the higher energy photons. This structure allows for thermalization losses to be significantly lower than single junction cells. What allows the carriers to be collected at the electrodes in these cells are tunnel junctions between the different bandgap materials. Tunnel junctions are highly doped regions that allow carriers to overcome energy barriers by tunneling through very narrow sections of material. The efficiency in these cells could potentially reach 60% due to the fact that we can split up the solar spectrum into sections that correspond well with certain bandgap materials but also due to concentration of sunlight.  Cell efficiency scales with incident light intensity because both the photocurrent and voltage increase as a result of more photons hitting the cell. Current commercially available models operate at ~500 sun concentrations by using optical mirrors to focus the sunlight from a large area onto a small solar cell. Various types of concentrating optics are used for concentrator photovoltaic applications, the most common approaches use either Fresnel lenses or Cassegrain mirrors to obtain and incident beam that is hundreds of degrees higher than ambient solar flux. 
Since these multi-junction cells require several different materials on top of each other with very specific tunnel junction layers between them, the solar cell itself is very expensive. The way to make this technology cost effective is to make a very small solar cell that captures sunlight from a very large area with concentrating mirrors. A drawback of concentrators is that the angle of incidence for sunlight has to be close to normal so these systems need to have trackers that follow the path of the sun throughout the day. These trackers are expensive and add to the fixed cost of the cell as well as add an unfortunate maintenance cost to these systems due to the moving parts in the motors that turn the system. Another drawback of this system is that they do not work well at all on cloudy days. Concentrators require direct sunlight and all the sunlight is diffused through clouds on cloudy days and the efficiency of these systems plummets. One final drawback is that the temperature of the solar cells gets really high due to all the concentrated light. Dark current is directly related to temperature so the efficiency of the cell goes down as temperature increases. Temperature is the limiting factor to concentrator photovoltaic performance.  Cooling mechanisms are required to maintain high efficiency of these systems and can be passive, such as through cooling fins, or active, such as pumping a coolant over the cells. Either way these cooling mechanisms add to the cost of the overall system.
The majority of multi-junction solar cell research has primarily been conducted here at Stanford in the last and is just starting to become available at a commercial scale. Research facilities, such as the company started by Stanford researchers in this area called Solar Junction, have expanded rapidly the last 4-5 years. Many large scale (multiple MW) concentrator photovoltaic projects are currently underway in the US, a total of 1,000 MW worth of projects.  Many of these individual projects are utility-owned and in the deserts of the American southwest. This is how multi-junction cells will be utilized most efficiently, in large-scale, desolate sunny places where large solar-farms or arrays can be built. They will never be cost-effective residential use and almost never for commercial use on top of factories. Since the bulk of the cost of overall device will be in the concentrating mirrors and tracking systems, the high cost of fabricating multi-junction solar cells is offset by both the cost of the rest of the system as well as the reduced active area required due to concentration. Multi-junction solar cells allow for efficiencies far beyond the single-junction Shockley limit, concentrating mirrors take advantage of the efficiency/intensity correlation, and finally the tracking systems allow for the ideal incident angle of solar flux throughout an entire day.
© Michael Maas. The author grants permission to copy, distribute and display this work in unaltered form, with attribution to the author, for noncommercial purposes only. All other rights, including commercial rights, are reserved to the author.
 A. Luque and A. Marti, "Increasing the Efficiency of Ideal Solar Cells by Photon Induced Transitions at Intermediate Levels," Phys. Rev. Lett. 78, 5014 (1997).
 M. Yamaguchi, "III-V Compound Multi-Junction Solar Cells: Present and Future," Solar Energy Mat. Solar Cells 75, 261 (2003).
 A. W. Bett et al., "High-Concentration PV Using III-V Solar Cells," Proc. 4th IEEE Conf. on Photovoltaic Energy Conversion 1, 615 (2006).
 A. Royne, C. J. Dey and D. R. Mills, "Cooling of Photovoltaic Cells Under Concentrated Illumination: A Critical Review," Solar Energy Mat. Solar Cells 86, 451 (2005).
 "U.S. Solar Market Insight - Report - 2011 Year-In-Review - Executive Summary," Solar Energy Industries Association, 2012.