Jurgen Michel

With semiconductor chips and solar cells, the ongoing challenge is the same: make them better, stronger, faster, and cheaper. It’s easy to say but hard to do, because often the key is finding substrate materials that perform well without increasing costs.

Jurgen Michel, a senior research associate in the MIT Material Processing Center, says he’s found potential solutions by using germanium for light emitters and less expensive germanium on silicon for solar cell substrates. The result would be computer chips that are more powerful, solar cells that are more efficient, and ultimately technology that’s available to more people. “Especially when it comes to solar cells, higher efficiency and lower cost will open access to markets in both urban and remote areas,” Michel says.

Directing light

Michel’s focus is on photonic and solar cell materials and devices. On one front, he’s been working on designing a new kind of semiconductor chip. The old version used metal wires to route information. He wants to use light to do the same thing. These photonic interconnects between cores on a chip are much faster and less costly in terms of power consumption. This improvement is due to the use of germanium, which Michel says is the ideal material as it has a band gap at 1,550 nanometers and can tap into communication wavelengths.

“You can use the material to detect, to modulate, and to emit the light that you want to route on the chip,” he says. “In principle, several thousand cores could be put on a microprocessor.”

With his work on high-efficiency solar cells, Michel uses a lateral cell design, in which layers of material are put on top of each other, in this case, III-V semiconductors on a germanium-on-silicon substrate, and an optical element spreads the wavelengths of the solar spectrum laterally. The advantage is that, as the sun shines, different wavelengths are absorbed and upwards of 50 percent of the sunshine will be converted into electricity, as opposed to 20 percent with silicon cells and 40 percent with expensive tandem cells, he says.

In particular, the substrate material drives the cost. At first, gallium arsenide was used. The problem was that it’s expensive, Michel says. After that, germanium was used to grow gallium arsenide on top of it, which reduced the cost by half. Michel says that by using silicon with a thin film of germanium, he’s lowered the cost by at least a factor of 4. This reduction, he adds, is largely because germanium is abundant, only a small amount is needed, and silicon is inexpensive and provides a large size to work on, unlike the other wafers.

One specific application Michel is looking at is combining his solar cells with thermal storage, an ongoing need with solar energy. Sunlight is unavoidably intermittent; the ability to store heat that can be converted to electricity on demand would bring a consistency in times of both low sun and high demand and help the renewable energy become more widespread. Michel has proposed to split the solar spectrum and use the visible part for immediate power generation while the infrared part is used for thermal storage. “You can extend the power generation in such a system,” he says.

Closing the gap

Michel says that he’s focusing on building the actual solar cell and photonic systems. One advantage of germanium is that it accesses the entire wavelength range, but in order to take advantage of it, and reach the mid-infrared range in particular, a reliable light source is needed. Lasers would be ideal, but their tuning range is limited, requiring too many to be a realistic solution. He’s working on frequency combs — light sources that emit at multiple wavelengths.

Michel says that the significance of getting into the mid-infrared range is that this is where there are molecular vibration bands. “It’s what’s called the fingerprint region for molecules,” he says. Since chip-integrated sensors would be disposable and light, along with being sensitive, they could be used for commercial and military applications in identifying pollutants in the air and toxins in liquids.

Germanium has allowed this to happen, and it’s an example of a challenge in this kind of work — utilizing a material in different ways than before. Germanium is an indirect band gap material, and conventional understanding would say that it’s not possible to lase, Michel says. But, on second look, the indirect and direct gaps are near to each other, and by ignoring the indirect band and accessing the direct one, there’s the possibility to produce a laser. Michel has done this by straining and highly doping the germanium with phosphorus. The process moves the band gaps closer together, causing a spillover of electrons from the indirect band gap into the direct one and opening a path for the light emission.

Along with its flexibility, using germanium opens up commercial possibilities. “For microprocessors, it improves performance and reduces power consumption; for solar cells, it reduces cost significantly, opening new markets,” he says. With chips, the design means they can be run with a regular cell battery. With sensors, they’re much smaller and no longer need to be put on cars and trucks for mobility. “All together, the new design means the work gets out of the lab and into more urban and remote places that were once inaccessible, but now can benefit from the technology,” Michel says.