Quantasol exploits quantum effects

Rising concerns over global warming and energy security are fueling a ramp in solar cell manufacture. Global production is increasing at 47% a year in terms of power, and independent market analyst Peter Lynch claims that this industry is now growing faster than any other.

This rapid expansion has led to a temporary shortage of polycrystalline silicon, the dominant material used for making solar cells. Supply of this semiconductor has struggled to keep up with recent demand and this has opened the door for thin-film technologies, such as amorphous silicon-on-glass. These alternatives offer a lower cost-per-unit-area and a faster production process, but have the drawback of low efficiencies. While traditional silicon solar cells can typically deliver efficiencies of 15%, thin-film versions can only yield single-digit values.

Martin Green from the Photovoltaics Research Center at the University of New South Wales, Australia, claims that any further reduction in the solar cell s key metric – the dollar-per-peak-watt cost ($/Wp) – will be restricted by the large, fixed, balance-of-systems costs of a photovoltaic power system, such as the inverter that converts the electricity from DC to AC form. So, the best way to reduce the $/Wp ratio is to use more-efficient cells.

The high-efficiency triple-junction solar cells produced by the likes of Emcore and Spectrolab offer one solution. These devices were mainly developed for space applications where high costs are not a concern, but they are now starting to be deployed into high-concentration systems that focus sunlight onto small cells with cheap lenses and mirrors.

The market for these systems is being stimulated by feed-in tariffs for renewable electricity. These have recently been introduced in a number of countries, with Germany and other parts of Europe leading the way. However, the booming market is also attracting other forms of innovative concentrator technologies, such as large dish systems and novel solar combined-heat-and-power systems.

At Quantasol – a UK spin-out of Imperial College London, based in Richmond upon Thames – we are also planning to enter this market with our patented high-efficiency concentrator cells. These devices share the hallmarks of the triple-junction cell – high efficiencies at high concentrations. However, we avoid problems with material dislocations by strain balancing the entire epistructure. As a result, we can offer unique features that boost efficiency, such as photon recycling and hot-electron effects.

One of the primary advantages of our quantum well-based solar cells is the wider spectral absorption range. This extension occurs in the infrared, with absorption of sunlight that would pass straight through a conventional GaAs cell.

If our cell material quality is sufficiently high, photons with energy greater than the host bandgap will "see" an epistructure that behaves like a conventional GaAs cell (figure 1a). For radiation with that energy, incident light is absorbed and an output current results from the separation of electron-hole pairs by the junction s built-in electric field.

However, unlike a conventional cell, our design also absorbs photons that have an energy that is less than the GaAs bandgap (figure 1b). The additional electrons and holes that result have a very high probability of escaping from the wells and boosting the output current.

For our design to be effective, the material quality must be high, while the structure must contain a sufficient number of quantum wells. That is because the additional current must be delivered with a minimal reduction in voltage, so that the output power, the product of these two quantities, increases.

This growth challenge is compounded by a lack of low-bandgap alloys that are lattice matched to GaAs. InGaAs is not suitable, even though it is used in the quantum wells of the low threshold-current lasers, because the larger atomic spacing of InGaAs leads to compressive strain in the active region. This works for a laser requiring a few quantum wells, but is unsuitable for solar cells that demand 50 wells to provide sufficient absorption in the undoped region.

Our approach avoids this issue and draws instead on the work carried out by ourselves, co-founder Massimo Mazzer, colleagues at Imperial College, and John Roberts, a co-founder of Quantasol who is also a researcher at the EPSRC National Centre for III-V Technologies at the University of Sheffield, UK. By sandwiching GaAsP barriers between InGaAs wells, we can form structures that balance the tensile and compressive strain (figure 2). Characterization techniques, such as X-ray diffraction and electron microscopy, reveal that the material quality is excellent and free from crystal dislocations. Up to 65 quantum wells can be grown in an intrinsic region using patent-pending growth recipes.

Our technology promises significant efficiency enhancements in both single-junction and tandem solar cells. Carrier loss mechanisms are kept to a minimum by producing good-quality material and employing high-bandgap semiconductors in the p-type, n-type and barrier regions (figure 3).

The only loss mechanism that cannot be avoided is quantum-well radiative recombination, which is the dominant loss mechanism at the high current levels employed in concentrator systems. This loss results from the photons generated by the recombination of conduction band electrons in the bottom of the quantum well with valence band holes in the top of this well (figure 4). Fortunately, the unique feature of our particular design means that these photons will only be absorbed by the substrate. So if a mirror is inserted between the quantum wells and the substrate (figure 4), a high proportion of the photons generated can be reabsorbed in the wells and contribute to the output current.

The way that we exploit this in our first product is to grow the device on top of a Bragg reflector (DBR). This mirror can reflect one-third of the photons that would be lost and convert them into current-generating carriers.

We think that this feature gives us the greenest photovoltaic cells around – we even recycle our waste photons! This photon recycling has enabled us to produce single-junction cells delivering more than 27% efficiency at a concentration factor of up to 500. This is just shy of the record for any concentrator single-junction solar cell, which stands at 27.8% for a concentration factor of 216.

We are also working on another type of mirror, a "chirped" DBR, which will improve the proportion of reflected photons. Calculations show that this could lead to single-junction efficiencies close to 30% at a concentration factor of 320 (figure 5).

Our strain-balanced design was developed on research reactors at the University of Sheffield with a maximum capacity of three 75 mm wafers. As this throughput is insufficient for high-volume manufacturing, we have also scaled up our process at a UK-based epiwafer foundry. Subsequent device processing is also carried out off-site, alongside important lifetime qualification tests at the Centre for Integrated Photonics, Ipswich, UK.

Our single-junction cell s epistructure is relatively complicated, as it consists of a thick quantum-well stack and a DBR region, so we were very encouraged by the good material quality produced on the first attempt in an 8 × 4 inch reactor platform. This material had a similar quality to that grown on the research reactor and an excellent quantum-well uniformity across the wafer (figure 6).

We will now develop production processes for tandem cells that can deliver even higher efficiencies. They feature an InGaP-based top cell epitaxially grown on a GaAs-based cell with a lower bandgap. Both cells contain quantum wells that extend the spectral response. In a tandem cell the two subcells are connected in series and their current output can be tuned by adjusting the quantum wells.

The Quantum Photovoltaic group at Imperial has demonstrated that quantum wells in the GaAs cell alone can produce tandem efficiencies above 30%. We are targeting a production efficiency of 35% at 500-sun concentration by mid-2008. This is comparable to the performance of triple-junction cells.

Our development of single and tandem cells with quantum wells is backed by £1.35 million ($2.7 million) of seed funding from investors, including Imperial Innovations and Low Carbon Accelerator. We aim to raise a significantly larger series A round in spring 2008 to finance high-volume manufacturing capability and capacity development, alongside significant additions to product development, operations and global customer support. We are planning to continue to pursue a "fab-lite" business model to keep costs low and enable a quick ramp up in production, which will involve the outsourcing of epigrowth and cell processing. Although we are currently working with UK partners to fulfill these roles, we are keen to develop overseas partnerships to provide additional production capacity.

We believe that our approach will allow us to tailor our designs to meet the precise needs of our customers. We are also free from any restrictions that face terrestrial concentrator solar-cell manufacturers that also serve the space and/or military sectors. Alpha-sampling of our single-junction cells will begin soon, with tandem cells to follow in the first half of this year.

Further reading
www.Solarbuzz.com/
M A Green 2001 Progress in Photovoltaics: Research and Applications 9 123.
K W J Barnham et al. 2006 Nature Materials 5 161.
D C Johnson et al. 2007 Appl. Phys. Lett. 90 213505.