Shedding light on new frontiers of solar cell semiconductors

IBM sets world record for photovoltaic energy conversion efficiency with earth-abundant materials

Editor’s note: This article is by Teodor Todorov and David Mitzi, IBM Research photovoltaic scientists.

by David Mitzi, Teodor K. Todorov, Jiang Tang, Santanu Bag, Oki Gunawan, Tayfun Gokmen, Yu Zhu, David B. Mitzi

Energy from the sun reaching the earth’s surface amounts to several thousand times our global consumption of electricity. Yet electricity from photovoltaic (PV) solar cells currently contributes significantly less than one percent of worldwide production. Of the numerous existing PV technologies, none so far have combined the virtues of being highly efficient, cheaply scalable and made with abundantly available materials.

IBM’s Materials Science team has partnered with Solar Frontier, Tokyo Ohka Kogyo (TOK) and DelSolar to develop an efficient and affordable PV cell made of abundant natural materials. So far, the tests of our Cu2ZnSn(S,Se)4 (made of readily available copper, zinc, and tin, and referred to as CZTS) thin-film devices have achieved a world-record PV solar-to-electric power conversion efficiency of 11.1 percent (10 percent better than any previous reports) for this class of semiconductors. And it can be manufactured by simple ink-based techniques such as printing or casting.

What makes CZTS better

Currently, the most widespread PV semiconductors, made of crystalline silicon, are abundant and highly efficient. They’re in panels used for everything from home electricity to the International Space Station. However, they have extremely high material purity requirements (>99.9999 percent!), and the wafers are typically cut from large solid ingots and wired in series to form PV modules – making it expensive and difficult to upscale. 

Photos of IBM's CZTS Solar Cell Device.

Other thin-film chalcogenide materials used in PV cells, such as Cu(In,Ga)(SSe)2 (CIGS) and CdTe, have been developed to a performance level close to that of silicon, with inherently more scalable processing. They are directly deposited on large-area, low-cost substrates such as glass, metal or plastic foil. While CIGS and CdTe are easy to integrate into buildings and consumer products, their compounds contain rare and expensive elements that increase cost and limit their manufacturing levels to less than 100 Gigawatts per year (worldwide continuous electricity consumption is 15 Terawatts – 150 times greater than the level of what these CIGS can produce).

Our CZTS PV cells could potentially yield up to 500 GW/year – getting closer to the Terawatt levels of renewable electricity the planet needs. 

The focus of our joint-development team remains to further increase this device efficiency and transfer the technology to environmentally-friendly, high-throughput industrial manufacturing. The hope is that within several years this new class of photovoltaic materials will begin to contribute to the wider availability of lower-cost solar electricity. 


  1. aterials Science team has partnered with Solar Frontier, Tokyo Ohka Kogyo (TOK) and DelSolar to develop an efficient and affordable PV cell made of abundant natural materials. So far, the tests of our Cu2ZnSn(S,Se)4 (

  2. Conversion and motion amidst the power electronics industry has increased demand for lighter/smaller, cheaper and more efficient systems. This shift begins at the semiconductor level.Friday, August 17, 2012: Four technologies are best suited to handle ...

  3. solar cell is the next big thing as power generation is becoming one of the best investment now a days. I love all IBM and lenovo parts

  4. This is important work - keep up the great progress as we look for PV technologies made with abundant natural materials. These milestones will lead us to a Smarter Planet! Proud that IBM is part of this work.

  5. I am proud to be an IBMer. How could access to this technology?, Is actually real?

  6. When you say 500 GW/year do you mean 500GWH/year (total energy) or, over the course of a year, an average of 500GW of continuous power?

    1. I think they mean neither of those, but a production level of 500 GWp per year.
      So the amount of solar panels that can be produced in a year, capable of producing peak power of 500 GW.

      Zwerius Kriegsman

  7. Semiconductor materials are so useful because their actions can be easily controlled by the addition of impurities, known as doping.

  8. well that a massive step taken by a big company. i think this is a great inspiration to lot of people.


  9. IBM really cares! Soon enough solar electricity installation is possible and even developing countries can afford. I'm not that technical and all that. All I know this company is doing its best to make solar technology very affordable to all. Good job IBM!

  10. I had gone through your blog.I too think it is inspiration and many have to learn from this.