This new protocol effectivelytaps the full spectrum, and I presume wee are talking about the visiblespectrum. We live in an ocean ofinfrared radiation that would be neat to convert into brake horsepower. Obviously this has to now be made efficientin order to be useful.
The good news is that what theyhave so far will be easy to manufacture and we could have a fairly efficientsolar cell able to across the available spectrum. This means a greater energy gain per squarefoot.
In the best of all worlds,capturing around thirty per cent of the available spectrum would pretty wellend the hunt because everyone could agree that diminishing returns thenrules. Up to that point, we know that smallincremental advances have big consequences. This is one such.
Commercial equipment makers claimas high as fifteen percent efficiency, but no one bothers to explain quite howthat is actually measured. In thisexample, we are using three bands. Ifone band is good for 12% and the other two are good for 5% then we end up withan output equivalent to 22% using the measure equivalent for the single bands.
This will be worth tracking tosee what is made of it all.
A Step Closer toPractical Full Spectrum Solar Cells
JANUARY 25, 2011
A solar cell’s ability to convert sunlight to electric current islimited by the band gaps of the semiconductors fromwhich it is made. For example, semiconductors with wide band gaps respond toshorter wavelengths withhigher energies (lower left). A semiconductor with an intermediate band hasmultiple band gaps and can respond to a range of energies(lowerright) 
Using the unique features of the electronic band structure of GaNxAs1-xalloys, we have designed, fabricated and tested a multiband photovoltaicdevice. The device demonstrates an optical activity of three energy bands thatabsorb, and convert into electrical current, the crucial part of the solarspectrum. The performance of the device and measurements ofelectroluminescence, quantum efficiency and photomodulated reflectivity areanalyzed in terms of the band anticrossing model of the electronic structure ofhighly mismatched alloys. The results demonstrate the feasibility of using highlymismatched alloys to engineer the semiconductor energy band structure forspecific device applications.
How to make a full-spectrum solar cell
“Since no one material is sensitive to all wavelengths, the underlyingprinciple of a successful full-spectrum solar cell is to combine differentsemiconductors with different energy gaps,” says Walukiewicz.
One way to combine different band gaps is to stack layers of differentsemiconductors and wire them in series. This is the principle of currenthigh-efficiency solar cell technology thatuses three different semiconductor alloys with different energy gaps. In 2002,Walukiewicz and Kin Man Yu of Berkeley Lab’s MSD found that by adjusting theamounts of indium and gallium in the same alloy, indium gallium nitride, eachdifferent mixture in effect became a different kind of semiconductor thatresponded to different wavelengths. By stacking several of the crystallinelayers, all closely matched but with different indium content, they made aphotovoltaic device that was sensitive to the full solar spectrum.
However, says Walukiewicz, “Even when the different layers are well matched,these structures are still complex – and so is the process of manufacturingthem. Another way to make a full-spectrum cell is to make a single alloy withmore than one band gap.”
In 2004 Walukiewicz and Yu made an alloy of highly mismatched semiconductorsbased on a common alloy, zinc (plus manganese) and tellurium. By doping thisalloy with oxygen, they added a third distinct energy band between the existingtwo – thus creating three different band gaps that spanned the solar spectrum.Unfortunately, says Walukiewicz, “to manufacture this alloy is complex andtime-consuming, and these solar cells are also expensive to produce inquantity.”
The new solar cell material from Walukiewicz and Yu and their colleagues in Berkeley Lab’s MSD and RoseStreet Labs Energy, workingwith Sumika Electronics Materials in Phoenix, Arizona, is another multibandsemiconductor made from a highly mismatched alloy. In this case the alloy isgallium arsenide nitride, similar in composition to one of the most familiarsemiconductors, gallium arsenide. By replacing some of the arsenic atoms withnitrogen, a third, intermediate energy band is created. The good news is thatthe alloy can be made by metalorganic chemical vapor deposition (MOCVD), one ofthe most common methods of fabricating compound semiconductors.
How band gaps work
Band gaps arise because semiconductors are insulators at a temperature ofabsolute zero but inch closer to conductivity as they warm up. To conductelectricity, some of the electrons normally bound to atoms (those in thevalence band) must gain enough energy to flow freely – that is, move into theconduction band. The band gap is the energy needed to do this.
When an electron moves into the conduction band it leaves behind a “hole” inthe valence band, which also carries charge, just as the electrons in theconduction band; holes are positive instead of negative.
A large band gap means high energy, and thus a wide-band-gap material respondsonly to the more energetic segments of the solar spectrum, such as ultravioletlight. By introducing a third band, intermediate between the valence band andthe conduction band, the same basic semiconductor can respond to lower andmiddle-energy wavelengths as well.
This is because, in a multiband semiconductor, there is a narrow band gap thatresponds to low energies between the valence band and the intermediate band.Between the intermediate band and the conduction band is another relativelynarrow band gap, one that responds to intermediate energies. And finally, theoriginal wide band gap is still there to take care of high energies.
“The major issue in creating a full-spectrum solar cell is finding the rightmaterial,” says Kin Man Yu. “The challenge is to balance the proper compositionwith the proper doping.”
In solar cells made of some highly mismatched alloys, a third band ofelectronic states can be created inside the band gap of the host material byreplacing atoms of one component with a small amount of oxygen or nitrogen. Inso—called II-VI semiconductors (which combine elements from these two groups ofMendeleev’s original periodic table), replacing some group VI atoms with oxygenproduces an intermediate band whose width and location can be controlled byvarying the amount of oxygen. Walukiewicz and Yu’s original multiband solarcell was a II-VI compound that replaced group VI tellurium atoms with oxygenatoms. Their current solar cell material is a III-V alloy. The intermediatethird band is made by replacing some of the group V component’s atoms –arsenic, in this case – with nitrogen atoms.
Finding the right combination of alloys, and determining the right dopinglevels to put an intermediate band right where it’s needed, is mostly based ontheory, using the band anticrossing model developed at Berkeley Lab over thepast 10 years.
“We knew that two-percent nitrogen ought to do the job,”says Yu. “We knew where the intermediate band ought to be and what to expect.The challenge was designing the actual device.”
Passing the test
A test device of the new multiband solar cell was arranged to block currentfrom the intermediate band; this allowed a wide range of wavelengths found inthe solar spectrum to stimulate current that flowed from both conduction andvalence bands (electrons and holes, respectively). In a comparison device thecurrent from the intermediate band was not blocked, and it interfered withcurrent from the conduction band, limiting the device’s response. (For bestresolution, click on image.)
At top, a test device of the new multiband solar cell was arranged to blockcurrent from the intermediate band; this allowed a wide range of wavelengthsfound in the solar spectrum to stimulate current that flowed from bothconduction and valence bands (electrons and holes, respectively). In acomparison device, at bottom, the current from the intermediate band was notblocked, and it interfered with current from the conduction band, limiting thedevice’s response. (For best resolution, click on image.)
Using their new multiband material as the core of a test cell, the researchersilluminated it with the full spectrum of sunlight to measure how much currentwas produced by different colors of light. The key to making a multiband cellwork is to make sure the intermediate band is isolated from the contacts wherecurrent is collected.
“The intermediate band must absorb light, but it acts only as a stepping stoneand must not be allowed to conduct charge, or else it basically shorts out thedevice,” Walukiewicz explains.
The test device had negatively doped semiconductor contacts on the substrate tocollect electrons from the conduction band, and positively doped semiconductorcontacts on the surface to collect holes from the valence band. Current from theintermediate band was blocked by additional layers on top and bottom.
For comparison purposes, the researchers built a cell that was almost identicalbut not blocked at the bottom, allowing current to flow directly from theintermediate band to the substrate.
The results of the test showed that light penetrating the blocked deviceefficiently yielded current from all three energy bands – valence tointermediate, intermediate to conduction, and valence to conduction – andresponded strongly to all parts of the spectrum, from infrared with an energyof about 1.1 electron volts (1.1 eV), to over 3.2 eV, well into theultraviolet.
By comparison, the unblocked device responded well only in the near infrared,declining sharply in the visible part of the spectrum and missing thehighest-energy sunlight. Because it was unblocked, the intermediate band hadessentially usurped the conduction band, intercepting low-energy electrons fromthe valence band and shuttling them directly to the contact layer.
Further support for the success of the multiband device and its method ofoperation came from tests “in reverse” – operating the device as a lightemitting diode (LED). At low voltage, the device emitted four peaks in theinfrared and visible light regions of the spectrum. Primarily intended as asolar cell material, this performance as an LED may suggest additionalpossibilities for gallium arsenide nitride, since it is a dilute nitride verysimilar to the dilute nitride, indium gallium arsenide nitride, used in commercial“vertical cavity surface-emitting lasers” (VCSELs), which have found wide usebecause of their many advantages over other semiconductor lasers.
With the new, multiband photovoltaic device based on gallium arsenide nitride,the research team has demonstrated a simple solar cell that responds tovirtually the entire solar spectrum – and can readily be made using one of thesemiconductor industry’s most common manufacturing techniques. The resultspromise highly efficient solar cells that are practical to produce.
