The Rise of Solar Cell Efficiency

Solarplaza tells us that the National Center for Photovoltaics (NCPV), part of the American National Renewable Energy Lab (NREL) has published an impressive chart that shows all solar cell record efficiencies since 1975.


Over on Wiki-P is an excellent timeline for the history of PV cells. It begins in 1839 when Alexandre Edmond Becquerel observed the photovoltaic effect via an electrode in a conductive solution exposed to light. It an NCPV’s chart are good resources. Much has happened since 1963 when the Sharp Corporation produced a viable photovoltaic module of silicon solar cells. I’ll pick up the story from 1976 and go straight to the dark blue lines that represent the development of crystalline silicon PV cells known as first generation PV cells.

First Generation PV Cells

These use relatively simple and robust technologies to coat silicon crystals onto glass. Their basic structure and mechanism has not changed much and they have always had a good efficiency.


  1. Photons in sunlight hit the solar panel and are absorbed by semiconducting materials, such as silicon.
  2. Electrons (negatively charged) are knocked loose from their atoms, causing an electric potential difference. Current starts flowing through the material to cancel the potential and this electricity is captured. Due to the special composition of solar cells, the electrons are only allowed to move in a single direction.
  3. An array of solar cells converts solar energy into a usable amount of direct current (DC) electricity.

There’s a bit more to it than this, but this is sufficient for now. Crystalline silicon solar cells remain the dominant type because of the availability and price of silicon. This silicon is known as “solar grade” silicon because its quality isn’t as high as that used for semiconductors even though its price still fluctuates according to supply and demand.

The efficiency of a solar cell can be broken down into reflectance efficiency, thermodynamic efficiency, charge carrier separation efficiency and conductive efficiency. The overall efficiency is the product of each of these individual efficiencies. Laboratory efficiencies are always higher than commercial efficiencies.

The Chart shows how the efficiency of solar cells varies according to the silicon crystal size and the type of crystal formed. It also shows that (funding for) research into the various types of silicon crystal structure stops and starts according to what looks most promising. Some research is discontinuous. In 1997 and 2005 were two isolated attempts to use thick Si films, for example, and that Panasonic achieved a 25% efficiency in 2012 with a silicon heterostructure which they are marketing as HIT (Heterojunction with Intrinsic Thin-layer) solar cells.

img02 You can always tell a mono-crystalline silicon solar cell by the octagonal corners that result when the wafer is cut from the cylindrical ingots manufactured by the Czochralski process. This is wasteful.


The International Space Station has about 27,000 sq.ft (2,500 sq.m) of old-school, monocrystalline silicon solar cells on its PV arrays so that when one side receives light from the Sun, the other side picks up light reflected from Earth.

international space station solar cells

Second Generation PV Cells 

Instead of having crystals grown on glass, the second generation of PV cells shown by the green lines use photolithography to deposit layers of materials on a base material known as a substrate. Photolithography is like stencilling and is also used to make semiconductor devices. In 1976, David Carlson and Christopher Wronski of RCA Laboratories created the first amorphous silicon PV cell with an efficiency of 1.1%.

PVeff(rev130307) The main advantage of amorphous silicon PV cells is not efficiency but cost. Amorphous silicon PV cells are thin and flexible and used for devices such as calculators that don’t need much power. Their efficiency can be improved by layering but this increases the cost/area and makes them unsuitable for large-area applications such as solar arrays.

At the same time RCA was developing their amorphous silicon PV cell In 1976, Matsushita (better known as National or Panasonic outside Japan) developed a cadmium telluride PV cell with an efficiency of 9%. Efficiencies of CdTe PVs have always been about 5% higher that those of amorphous-Si PVs ever since the 12% achieved by United Solar in 2005. The highest CdTe efficiency to date is about 18.5% by First Solar who seem to be leaders in the technology. They say:

  • CdTe-based PVs have a superior temperature coefficient and lower temperature-related losses.
  • Because of this, their PV plants produce more energy than competing solar systems with the same power rating.
  • CdTe also enables high-volume, low-cost production of solar modules using a continuous, automated process that requires less energy and water and utilises 98% less semiconductor materials compared to conventional crystalline silicon production.
  • The carbon footprint of CdTe systems compares favourably to that of other energy sources, including other PV technology. (The asterisk in the graph refers to RENEWABLE & SUSTAINABLE ENERGY REVIEWS”, Vol 8, 2004, pp 303–334, V M Fthenakis, “Life Cycle Impact Analysis of Cadmium in CdTe PV Production”, with permission from Elsevier.) 


  • They also have the fastest energy payback time. In less than a year, CdTe PV power plants produce more energy than was required to create them. Sounds good.

infographic-energy-payback-time Problems of CdTe PV cells include the availability of tellurium which is neither a particularly abundant nor easily found element. Undersea ridges appear to have a lot of it. Uh-oh. Wait, there’s more. Tellurium’s partner-in-compound, cadmium, is nastily toxic and you will die a painful death if it and you are in the same food chain. First Solar ensures their PV panels are fully recycled and all cadmium recovered and reused.

In 1977 the University of Main developed the first copper indium gallium telluride PV cells. CuIn1-xGaxSe2 (CIGS) has the highest efficiency of all of the thin-film PV technolgies because the material has a high absorption coefficient and strongly absorbs sunlight. This makes it possible to have a thinner film.

Know your elements!

  1. Gallium is a byproduct of producing aluminium and zinc and about 98% of annual production (216 metric tons in 2011) is used in the manufacture of semiconductor devices.
  2. Indium is an atomic neighbour of gallium, shares many properties with it, and is also produced during the production of zinc. Its 2002 price was US$94 per kilogram but 2006 to 2009 prices ranged from US$382/kg to US$918/kg. It’s estimated that there’s less 

    than 20 years’ supply of indium left so no long-term bets on indium.

  3. Selenium is one of those nice, mid-table elements that can do a lot of things and one of those things is that its electrical resistance changes when light falls on it – a property that was first noticed in 1873.

The band gap

We can’t move on without first talking about the band gap. The absorption coefficient of any material is determined by its band gap which is an energy range in a solid where no electron states can exist. 


It generally refers to the energy difference (in electron volts) between the top of the valence band and the bottom of the conduction band in insulators and semiconductors. Substances with large band gaps are generally insulators, those with smaller band gaps are semiconductors (hey – we’ve heard of you!), and conductors either have very small band gaps or none, because the valence and conduction bands overlap.  Every material has a characteristic band gap and most PV and solar cell research goes into finding materials with better ones. 

How well a PV or solar cell can absorb photons and convert them into a potential difference depends on how easily the photon can split and send the electron (-) to the conduction band and the hole (+) to the valence band. (The hole is called that because it’s where the electron was. Electron-hole pairs are called excitons.)  The relationship between the band gap of a material and the theoretical maximum efficiency at which that material can convert light into energy is determined by the Shockley-Queisser limit. It’s about 1.5eV.

The Shockley-Queisser limit for the efficiency...

CdTe solar cell research looked promising because the band gap of CdTe is about 1.5 eV and almost perfectly matches the distribution of photons in the solar spectrum. The energy level of photos reaching our planet is one of those things that can’t be changed.

CIGS solar cell research also looked promising because CIGS has a direct band gap. A direct band gap is just that – the shortest path. A direct band gap makes it easier and quicker for the exciton components to go their separate ways.

CIGS is the third mainstream thin-film PV technology. CIGS PVs have had better efficiencies than CdTe PVs since 1995. The best is currently around 20% but their commercial development hasn’t been as fast as the other two types. 

Multi-junction solar cells/tandem cells 


Earlier, I wrote that the photon energy level of 1.5eV was just something that had to be accepted, but this 1.5eV s the average energy level of visible light. Multi-junction solar cells stack two or more solar cells optimised for light of different wavelengths. Clever.

multijunction cell structure This overcomes one of largest inherent sources of losses and produces a considerable increase in efficiency. Traditional single-junction cells have a maximum theoretical efficiency of 34% but a theoretical “infinite-junction” cell would achieve 87% under highly concentrated sunlight. As you’d expect, multi-junction solar cells are difficult and expensive to manufacture. If semiconductor research is anything to go by, we can be sure that scientists and researchers are working very hard to find ways to

  1. use less materials,
  2. use less expensive materials,
  3. use simpler processes,
  4. use fewer processes, and
  5. use fewer processes that require high degrees of purity or precision.

These are good things to do because it means more people can benefit. So far, the higher price and lower price-to-performance ratio of MJ solar cells has limited their use to special roles, notably in aerospace where their high power-to-weight ratio is desirable. Here’s some of Mars Rover’s MJ solar cells in action on Mars. This is not a render.

mars_spiritcolor_PIA05015_c1 Spectrolab Inc., a wholly-owned subsidiary of The Boeing Company, is the world’s leading merchant supplier of High-Efficiency Multijunction Solar Cells for CPV and spacecraft power systems, with proven reliability drawn from six decades of space heritage. 

6_decades_panel.eps Third-generation solar cells

Third-generation solar cells use materials other than silicon. They are shown in red on The Chart and include solar inks using conventional printing press technologies, solar dyes, and conductive plastics.


The Chart shows that research into dye sensitised solids began in 1991 but quickly reached its theoretical maximum. Research into organic cells began in 2001. Mitsubishi Chemical’s 2006 thin-film organic solar cell had an efficiency of 4.9% and, unsurprisingly, it’s green.

Energy vs. crystal momentum for a semiconductor with a direct band gap, showing that an electron can shift from the lowest-energy state in the conduction band (green) to the highest-energy state in the valence band (red) without a change in crystal momentum. Depicted is a transition in which a photon excites an electron from the valence band to the conduction band.

Energy vs. crystal momentum for a semiconductor with a direct band gap, showing that an electron can shift from the lowest-energy state in the conduction band (green) to the highest-energy state in the valence band (red) without a change in crystal momentum. Depicted is a transition in which a photon excites an electron from the valence band to the conduction band.


By 2012, Mitsubishi had increased the efficiency to 11%. They’re currently working on how to improve efficiency by creating organic cells with multiple layers.

Quantum dot solar cells

mitsubishi thin film organic pv

Quantum dot solar cells have been around since 2010. Instead of any of the materials mentioned so far, they use quantum dots as the photovoltaic material.

quantum-dot-solar-cellA quantum dot is a particle of semiconductor material that has been smaller than the distance that defines the exciton. I don’t fully understand why this makes them behave like artificial atoms but what it means is that it is possible to tailor their size to produce any required energy level. This is exciting and very important.  

Quantum dots can can be made to have any desired band gap without changing the basic material or construction. THE FUNDAMENTAL ARCHITECTURE DOES NOT CHANGE. 

Moreover, quantum dot solar cells do not require complicated manufacturing processes. Once commercialised, they will be very inexpensive.

Lead sulfide (PbS) quantum dots can even have their band gaps tuned into the far infrared. This is also fantastic news because half of all the solar energy reaching Earth is infrared and most of that in the near infrared region.

Final efficiencies of 65% are expected for quantum dot cells but it’s difficult to compare a 65% conversion rate for electromagnetic radiation against, say, an 85% in-your-dreams conversion rate for the visible spectrum only. It becomes meaningless to talk of efficiencies and comparisons without a common reference. 

There is a common reference, of course. In the future, the only way to compare the efficiencies of solar cells will be to compare the cost of the energy produced vs. the cost of producing it. We may never get to know all the embodied costs, hidden costs or the true life-cycle cost, but it’s still the best way to compare any two different energy sources – which, after all, is really what it’s all about. 

solar energy cost

• • •