Think a record-beating solar cell is what it takes to maximise energy from sunlight?
Marc Vermeersch, head of photovoltaic R&D at energy company Total, has a question for the solar industry. For years, he says. the focus has been on how many photons a cell will convert into electricity: 'why do you see this incredible amount invested in the efficiency of the cells? When all those cells are connected together in a monolithic module and then connected to the grid, there is an incredible amount of energy wasted'.
Companies such as Total have begun to focus more on how to control the energy that comes out of photovoltaics, although it has shareholdings in a number of cell manufacturers. The corporation has a 50 per cent shareholding in Tenesol - the other half is owned by EDF Energy - which, as well as panels, makes inverters to convert the direct current (DC) electricity that comes direct from a module into grid-compatible alternating current (AC).
'This area is probably the most profitable to work in when you calculate cost per kilowatt-hour,' says Vermeersch.
The problem for the industry is there is no one-size-fits-all answer for maximising the conversion potential of a solar cell or panel: it depends on where the panel will be used, how it is made and even the size of the entire installation. Add in the tendency to not analyse all the components in the chain and you have a recipe for confusion. So, it seems easier to focus on ultimate conversion efficiency in the photovoltaic cells.
'What has been surprisingly lacking in the industry is a high degree of accuracy in the modelling. The way to work through the arguments is to ask not just what the capital cost is, but also the net price of electricity and the avoided energy cost: how much are they saving?' says Chris Edgette, director of Stratagen Consulting.
The lack of modelling goes right back to the core cell itself. Tom Tibbits, director of product marketing at PV cell maker QuantaSol, says: 'Everyone in the industry uses the AM1.5D spectrum.'
The AM1.5D spectrum developed by the American Society for Testing and Materials is meant to approximate the type of direct sunlight that falls on the 48 contiguous States of the US and has significant notches from oxygen, ozone and water absorption. AM stands for 'air mass'.
'ASTM decided they needed a uniform test spectrum,' says Tibbits. Even in the US, 'those conditions only exist once or twice a year if you are lucky', he adds. 'If you are interested in kilowatt hours, you want to maximise energy over the year. So we have invested in people, such as Alison Pearson, an atmospheric scientist, to look at light and ambient temperature and then optimise design for custom systems rather than making world-record solar cells according to the AM1.5D test,' Tibbits explains (see boxout 'Tune your cells', p46).
The problem is in convincing module makers that tuning for local conditions versus something that looks good on a datasheet is what they want to do. 'Some companies really get it and understand they make money out of selling kilowatt hours. Then there are others who want the best solar cell on the market. And that is defined by the laboratory conditions. The market is very bifurcated in that respect,' Tibbits says.
Efficiency over cost
How the cells come out of the factory can play into efficiency. There is a tolerance in manufacture that leads to mismatches between cells and the harvesting electronics.
David Joseph, chief strategy officer at semiconductor process-modelling specialist PDF Solutions, argues: 'The focus in PV manufacturing from the beginning has been on cost. But we think the biggest leverage is in controlling the manufacturing process to improve efficiency.' Part of the problem for solar-cell manufacturers is that many use turnkey fabs and do not typically spend a lot of time tuning the process to cut variability between wafers, says Joseph. 'A lot depends on the technology but the variability can be huge,' he adds.
'Silicon cost dominates in crystalline panels. So this is a heresy in the PV world, but if you increase cost by 5 per cent and get 5 per cent better efficiency, you can get 3.75 per cent drop in the overall cost per watt. A trade-off for more complex processing will give you an improvement in overall efficiency and improve system cost. But, in PV, we are definitely in evangelistic mode trying to promote this.' Even if the PV cells are well-matched and optimised for typical lighting conditions of a location, those conditions are not always going to be present, particularly with changing weather and climate.
Most installations today use a central inverter to deliver AC electricity to the grid. This puts most of the complex electronic control needed to synchronise with the grid into one unit.
Leo Casey, chief technology officer for inverter maker Satcon, claims the economics of inverters favour large, centralised units because most of the cost lies not in the electronics but the transformer. 'The only way to get the cost down is through the iron and copper. The cost of magnetics drops precipitously with size and weight. Large transformers are a factor of a penny per watt,' he says.
The problem for the central inverter is that it cannot react to changes in shading and temperature that may occur across an installation. Utility-scale farms can avoid the issue through good layout. But installers working on domestic and commercial systems work around the problem by simply not using as much roof space, according to Ralf Münster, director of the renewable energy business unit at National Semiconductor, quoting a survey by Greenberg Quinlan Rosner Research.
A little over half of the 150 installers surveyed said any shade on installations was unacceptable and simply designed around it. Münster reckons installations could, on average, be up to 20 per cent bigger if they could generate power reliably when portions of the array are shaded.
The trouble is, shade has a disproportionate effect on generating efficiency. As little as 2.6 per cent shading can lead to a total loss of 16.7 per cent in actual energy conversion. According to the US National Renewable Energy Laboratory a shadow can represent a reduction in power more than 30 times its physical size because of the way that cells are wired up and energy harvested. Rising temperature has an effect too, pushing efficiency down.
UK startup Naked Energy aims to capitalise on the situation by encapsulating recycled PV wafers inside vacuum tubes that use fluid-filled copper pipes behind the cells to transfer excess heat to a heat exchanger to provide a hot-water supply on top of electrical generation. Managing director Christophe Williams claims the approach 'provides optimum efficiency for longer and you get hot water, which is 80 per cent of a home's energy demand'.
Michael Curran, applications engineer with Microchip Technology, explains the problem for solar installers: 'Current output is affected by light and temperature affects voltage. For each of these conditions, there is a unique operating point at which maximum power is delivered.'
The trick is finding the right operating point. But, when you have panels at different temperatures and with different levels of incident light, there may be no common maximum power point (MPP) that delivers the highest efficiency. 'We can't have the maximum if our MPP tracking is centralised,' says Francesco Pulvirenti, director of STMicroelectronics' photovoltaic business unit.
The answer, argue inverter companies such as Enphase Energy and semiconductor makers Microchip, National Semiconductor and STMicrolectronic, is to distribute the control to individual panels, even individual cells. Each module can convert to a common voltage level using its own inverter tuned for its local MPP.
'Any time you have a lot of shading you want module-level components,' says Edgette. 'They deal better with mismatch and have better resilience to shade and debris,' says Raghu Belur, vice president of Enphase Energy.
In larger systems, a central power converter may still be needed to coordinate with the grid, although Enphase claims its technology can do away with that element. However, EIQ Energy favours using high-voltage DC as the means to deliver electricity with fewer losses to a central inverter that can be optimised for a more consistent source in terms of voltage. Components made by National and ST are designed to act as DC/DC converters for this kind of architecture.
'Our expectation is that this approach can obtain up to 20 per cent more energy but it can be much lower, between 2 and 10 per cent. It depends on weather conditions. In ideal conditions, the extra energy harvested will be very small. But, in most conditions where we have shadow or different thermal gradients, then we can reach 20 per cent,' claims Pulvirenti.
The problem with the distributed approach is that it potentially increases cost: you now have two sets of power-conversion stages and, with a central inverter to maintain grid compatibility, little opportunity to reduce the cost because it depends on materials prices. But backers of the distributed approach believe the quest for cost efficiency over the lifetime of a PV array will win out - and move attention away from cells that win records based on a synthetic benchmark.