American Scientist

The most important statistic for any type of photovoltaic technology is the efficiency percentage by which it converts solar energy into electricity. Current silicon-based solar cells have about a 24 percent efficiency, but improving that rate is becoming more difficult, because it requires the base material to have extremely high purity. Mercouri Kanatzidis—a Charles E. and Emma H. Morrison Professor of Chemistry and professor of materials science and engineering at Northwestern University and a senior scientist at Argonne National Laboratory—has been working for more than a decade on a class of materials that do not need a high level of purity to achieve high conversion percentages. These materials, called perovskites (so called because their structure resembles that of a mineral of the same name), are now being deployed in solar cells that are undergoing commercial testing of their stability. Kanatzidis was the recipient of Sigma Xi’s Walston Chubb Award for Innovation at the 2024 International Forum on Research Excellence (IFoRE), and he spoke with editor-in-chief Fenella Saunders after the conference about his work. (This interview has been edited for length and clarity.)

How did you get interested in solar cells and looking at these materials?

About 15 years ago, I had a general interest in solar energy conversion, but I wasn’t active in it, until I was approached by a company who wanted to develop a novel solar cell that needed a material that has certain specific properties. I thought they had a very interesting project to find such a material. It took us two and a half years to solve the problem. Unfortunately, the company didn’t survive. We were left with great results that we had obtained for that project that the company allowed us to keep.

Now, we had all this knowledge about these new materials. It was a class of materials, not just one, called the perovskites, which are very famous now, but back then, no one knew they were important for anything. We developed the chemistry, the synthesis, the crystallography. We figured out all the crystal structures, the light emission properties, the charge transport properties. We were going to publish in a chemistry journal, just put all these compounds and information together, but something happened that was totally unexpected.

I went to a seminar given by a colleague here in materials science, Robert Chang, who talked about another type of solar cell, the so-called dye-sensitized cell. A lot of people were working on it to make it more efficient and more stable. It was based on a liquid solution, but they wanted to make it a solid-state cell. They had 10 percent efficiency in conversion of solar energy to electricity, which was very high at that time, because it was low cost and easy to make. It occurred to me that one of the perovskites we worked on might actually work. You could dissolve it and deposit it and get a film that was easy to process. I suggested that to my colleague, and so we started trying it.

In the beginning, the results were not very interesting or exciting. But my colleague and I realized that the reason it wasn’t working wasn’t the material or the idea, but the device fabrication. So then we focused on making better devices. Within about a year, we had 10 percent efficiency, almost the same as the liquid dye-sensitized cell. That’s where we came out with the first paper based on perovskites supporting a solar cell. Then two other groups published two similar papers with another perovskite, with which we had also worked.

Those papers sparked what turned out to be a revolution in photovoltaics, and even science, because these semiconductors were very unconventional. The researcher community grew. The efficiency rose again and again, and it’s still rising today. Now it’s about 27 percent. Conventional semiconductors took 40 years to reach an efficiency of maybe 22 percent. It’s stunning.

What material properties are needed to make a good photovoltaic?

You need a solid film that absorbs as much of the solar radiation as possible. Generally, that’s why it makes them black, so you absorb more of the photons from the Sun. These photons excite electrons inside the material. When the electron is excited, where it bleeds off, that creates an electron hole. Now the material must be able to transport these electrons and holes away from where it happened. When you put in electrodes, they can then collect these holes and electrons on opposite sides. Now you have a voltage and a current forming. If this current doesn’t happen and the electron and hole recombine, that’s bad because there is no work being done, just heat is generated. So, the material has to be able to support this transport of electrons and holes inside it. It sounds easy, but most materials don’t do that. The electrons are trapped or scattered, and they never make it to the electrode.

Another necessary component is a band gap. Can you explain what that is?

The energy gap to excite one electron to the next level is the band gap. Between the highest level where you have some electrons and the lowest level that doesn’t have any, there’s a gap. The sunlight has energy. If that energy equals the gap, then it will excite the electron across the gap. And now you have an electron and a hole. But if the energy level is lower than the gap, the sunlight will go through the material. It will not be absorbed.

What makes a material a good conductor of electrons and holes?

That’s where the unconventionality of perovskites comes in. In a classical semiconductor, in order for the material to be a good conductor, it must be extremely high quality and pure—no impurities, no defects. Defects are atoms being in the wrong positions in the crystal structure. That means you have to work very hard to purify them, and that’s why it takes decades to raise the efficiency. However, the perovskites are full of defects, and they still work. In a classical semiconductor, if there is a defect, it will introduce a new state in the middle of the band gap. If you have an electron excited, instead of traveling, it can fall into the state and then it’s trapped. In perovskites, because of the way the chemical bonding between atoms is in this particular material, when these defects form, they don’t form these mid-gap states. They form them away inside the higher level, so they don’t play any role in trapping.

What is it about the perovskite’s structure that gives it this property?

One thing is that it’s three dimensional. So all three directions are possible to transport. Also, they have lead and tin that are bound to the halides, bromides, and iodides, and they form an octahedron. Then the octahedra share corners. That’s how they build the three-dimensional structure, which is negatively charged and takes these positive small ions inside. The tin and lead have nonbonding electrons in this valence state, and when they bond into the structure, they form antibonding states—it has a lone pair of electrons—that dominate the valence bands in the solid. When you make a defect in such a material, instead of detaching from the valence band and moving into the band gap, it’s detaching from the valence band and moving inside the band, so it’s not a trap.

Also, there’s a dynamic behavior caused by the lone pair of electrons. The actual structure fluctuates and the electronic structure is a direct result of that, so there’s a fluctuating electronic structure, which causes a delay in the recombination of the excited electron and hole after absorption of the light energy. The electron and hole are not in the exact positions they came from, and it takes a little bit of extra time to find each other, which also buys you time to collect them.

One of the remaining problems with perovskites has been their stability. How can that be increased?

The same thing that makes them work also makes them unstable, and that is that the metal halogen bonds are ionic. These devices operate under voltage created during operation, and ions can migrate and cause instability. Using two- dimensional perovskites contacting the 3D ones was one of the early strategies that we showed was effective in considerably lengthening the stability. People are now combining all the different strategies to cover everything that can go wrong. Already companies are deploying solar modules based on perovskites—in the testing stage, not in the standard commercial release stage. The potential customers are testing them to see, under real conditions, how long the modules last. This is very good news.

How do you use more than one material to increase the solar cell efficiency?

The lone pair of electrons also gets you another positive characteristic. If you have two related semiconductors and you mix them together, you can make compositions in any ratio and get band gaps in between those of the two materials. But if we have two perovskites, say, one with tin and one with lead, you can also mix them up in any ratio and make intermediate compositions. Tin has a band gap of about 1.4 electron volts, and lead has 1.55. But instead of intermediate numbers between those two, you get a bowing effect: The gaps go down, reach a minimum, and go back up. In the end you have a curve, and a composition in the middle has a lower band gap than the lowest of the two end members. We have explained it because of the same lone pair effect. A smaller band gap means it absorbs more light in the visible spectrum and especially in the infrared.

“Perovskites are full of defects, and they still work, because of the way the chemical bonding between atoms is in this material.”

• Facebook
• Twitter

But no cell can actually capture all the solar light. So, we use two cells in tandem. One cell has a wide gap to capture the high-energy light and then the rest of the light will go through to another cell that has a smaller gap to capture the low-energy light, and together we capture more than we can with the individual cells. Together you can exceed the theoretical limit of a single solar cell, going to 30 or 35 percent. And if it’s triple or quadruple tandem, people now are thinking you could go to 50 percent.

How efficient is it possible for these devices to become?

If you have only one solar cell, the limit is about 32 percent. If you go to a big number of tandem cells, the theoretical limit could be 55 or 60 percent. Right now, people are claiming tandems that have 33 to 34 percent.

Some companies are actually already marketing tandems with silicon as the bottom cell. So, in other words, it’s a perovskite and silicon hybrid. It’s difficult to dislodge silicon from its markets. If you can go to a manufacturer of silicon solar cells and say to them, “All you have to do is add one or two steps in your process, and instead of having 24 percent, you’ll have 27 percent,” the hope is that the addition would make sense to them and not seem like a big change.

What other fields could use perovskites?

They are turning out to be tremendous x-ray and gamma ray detectors, and these are new applications that will affect biomedicine, medical diagnostics, medical imaging, national security for the monitoring of nuclear materials, and so on. And there are other areas, perhaps in lasers or light-emitting diodes. The perovskites are doing to these fields what they did to photovoltaics 10 years ago.

How much of a role have you seen for serendipity in research?

You never know what you don’t know, and therefore you have to hope for serendipity. We have good ideas and good hypotheses, but nature has other things up her sleeve. Therefore, when you try something, very often something else can happen. You have to have the curiosity and the wisdom to actually look at that something else rather than dismiss it and say, that’s not what I’m looking for. Sometimes it’s a breakthrough, and it could change how you think. So, serendipity is always there in science. I don’t think it’s useful for us to pretend otherwise, to always think that we’re in control, and we always know what we’re doing. Curiosity is key. We should encourage our students and postdocs to have it, and hope that they will find that serendipity from time to time.