In a few years, one of the primary components of the grid is going to become downright portable.

A number of companies are working on technologies that could replace traditional transformers -- large pieces of industrial equipment fashioned from groups of components for raising or lowering electrical voltage -- with transformers that largely consist of semiconductors on circuit boards. This process could begin to pave the way for a number of improvements in the way that power gets delivered. Integrating and managing renewable power and electrical storage could become easier. Microgrids could be deployed much more rapidly.

Grid efficiency could conceivably be increased by up to 8 percent to 10 percent because of lower conversion and transmission losses. 

"That's ten percent less power that you have to generate," said John Palmour, co-founder and Chief Technology Officer for Power and RF at Cree, which produces silicon carbide semiconductors. "We can replace an 8,000-pound transformer in a substation running at 60 hertz and replace it with one running at 20 kilohertz in a tiny design. [...] We can shrink it down to [the size of] a suitcase."

Varentec, a startup that has received funds from Khosla Ventures and others, has laid plans to develop a solid state transformer. In England, Amantys, a spin-out from Cambridge University that is currently raising funds, will show solid state prototypes later this year for converting medium and high voltages. Amantys' initial target market will be wind farms. (Solid state transformers and power electronics will be two of the major topics of discussion at The Neworked Grid taking place May 3 and 4 in San Francisco.)

On the component side, Cree, Infineon, Mitsubishi, ST Microelectronics and others have been working on silicon carbide semiconductors. Most of the efforts and commercial products to date have focused on low voltage converters. Cree's Power & RF devices group, for instance, sells silicon carbide diodes and MOSFETs, a type of switch, to solar inverter makers and server companies for computer power supplies. The current revenue run rate is around $100 million a year. Last month, Cree released a smaller version of its silicon carbide diode for inverters. 

By swapping out standard silicon components for silicon carbide ones, the efficiency of power supplies can be boosted from the high-80-percent range to the 93-percent-plus range. Some solar inverters containing silicon carbide components have even hit the 99% range.

The push now is to create components that can handle the sort of high voltages required for transformers on the grid. 

One could imagine Transphorm, which makes power converters from gallium nitride (GaN), selling components into this market. The company's first product operates at low voltages (around 400 volts), but some researchers have looked at taking GaN to higher voltages.

Relatively recent advances in material science -- reducing defect densities on test chips, boosting the yield on wafers -- and semiconductor design have effectively opened the door to high-voltage applications. In short, silicon carbide (pictured) and gallium nitride can lead to smaller, faster devices that can operate at higher temperatures more efficiently than equivalent ones based on silicon. Growing gallium nitride or silicon carbide crystals and substrates for these sort of applications, however, is no simple feat. 

Cree, in connection with DARPA and the Office of Naval Research, has built a 10-kilovolt, 100-amp power module with advanced silicon carbide MOSFETs. These power modules were integrated into equipment from General Electric to make a 1 megawatt transformer that operates at 20 kilohertz. An equivalent conventional transformer capable of handling 6.5 kilovolts might only operate at 500 hertz.

The DARPA transformer measures 16 inches high, weighs 75 pounds, and operates at 250 kilovolt amps (volts x amps). A conventional 330 kVA transformer stands 55 inches high, weighs 2,700 pounds and only operates at 60 hertz (more here).

Between now and 2013, Cree, under an ARPA-E grant, will build 15-kilovolt and 20-kilovolt insulated gate bipolar transistors (IGBTs), a switch with more capabilities than a MOSFET, out of silicon carbide. (Amantys' basic component will also be an IGBT.) If all goes well, this will lead to 50- and 100-kilowatt solid state transformers.

An existing silicon carbide 12-kilovolt IGBT achieves four times the switching speed and four times lower switching losses than an equivalent IGBT fashioned from silicon. The 20 kilovolt IGBT should have a similar switching performance, but with 10 times less loss than a silicon part. 

"It is in the silicon carbide components where the breakthroughs have come," Palmour said. "Silicon [the current material of choice for power components] is up against a wall."

Many of the advantages with solid state transformers come through reductions in size. Today, planting a transformer is not an easy task. Considerations like transportation, site preparation, installation and transmission costs all add to the budget. By contrast, smaller, cheaper solid state transformers could be relatively easily planted in small solar fields or storage pods. The faster switching speed of solid state devices would in turn make it easier for a utility to handle a multiplicity of power sources feeding into the grid because you'd have more transformers controlling and fine-tuning power quality.

"A bank of batteries for energy storage doesn't do you any good if you can't get the energy up on the line," Palmour said. 

Varentec doesn't say much about itself, but sources, and a website managed by its incubator/landlord, say that the company plans to make solid state transformers. 

Tags: cree, energy efficiency, gallium nitride, light emitting diodes, silicon, silicon carbide, smart grid, solid state transformers, transformers, varentec