Thin Film Solar PV vs Silicon Wafer - Which is Better?

A guest article by Dr KT Tan cuts through the marketing to find out



Figure 1 (Source: Jethro Betcke, Oldenburg University, Germany)

Thin film solar PV was hailed as the next big thing in solar nearly a decade ago. Then, crystalline silicon wafer (c-Si) cells occupied more than 80% of the market share compared to thin film PV (1). There was a high anticipation in the industry for thin film PV to position itself for a run at c-Si and dominate the market for the near future. However, 10 years on, history shows that not only did thin film fail to conquer the market, but its market share has subsequently declined to only 7% (2).

Obviously, one major factor was due to the collapse of the price for c-Si cells, which quickly wiped off the cost advantages of thin film technologies. This blog is not going to discuss the reasons for this distorted market competition, caused mainly by the exponential expansion of production in c-Si cells, but to question and compare the technical merits of thin film PV versus c-Si.

Do thin film PV technologies have an arsenal of special features to outperform c-Si cells? 



Low Light Performance

The first common belief is that thin film solar PV performs better in low light conditions or diffuse sunlight (for example on a cloudy day). But is this true? The fact that this has been heavily promoted by the marketing guys is because these two technologies do have different spectrum responses to solar light. In other words, their ability to convert solar energy to electricity varies at different wavelengths. In general, the average wavelength in diffuse sunlight is shorter (i.e. more blue) that of direct sunlight – so if you have a spectral response peaking at short wavelengths, e.g. thin film amorphous silicon (a-Si), then you would perform better under diffuse conditions than clear sky conditions.

Figure 1 for shows the different spectrum responses of different solar technologies against the power of sunlight of different wavelengths at sea level at mid-lattitudes of Earth (called AM1.5).  Crystalline monocrystalline silicon (labelled m-Si) is compared against different thin film solar technologies based on amorphous silicon (a-Si), Copper Indium Gallium Selenide (CIGS) and Cadmium Telluride (CdTe).

If you look at Figure 1, you probably would have noticed that not all thin film technologies have the same performance response to differing light wavelengths. Thin film CIGS solar panels, for example, have a broad spectrum response akin to mono-crystalline wafer cells (m-Si), so based on this their performance in diffuse lighting conditions would be little different to m-Si.

Amorphous Silicon has a quite significantly different spectral response to crystalline silicon, with a greater response to low wavelength light.  So how do they compare in field trials? Figure 2 illustrates the results of a comparative study between a-Si and c-Si on a cloudy day. On average, the tests show an increase in energy generated of 15% for a-Si at low irradiance levels below 260 W/m2.  (Note: the tests were published by NexPower, a manufacturer of amorphous Silicon panels)

Figure 2 (Source: NexPower In-house test report)

 

However, performing better on cloudy days is of little benefit if it is combined with performing less well on sunny days (when more energy can be collected). If this were the case then the advantage of thin film PV under diffuse conditions might be a complete red-herring created by the marketing gurus.

A recent research project (3) supported by the Deutsche Bunderstiftung Umwelt (German Federal Foundation for the Environment) compared several solar module types (including thin film and c-Si) under North German Climatological conditions in a side by side trial for a year, and it turned out that no significant difference between the performances of the different type of modules could be found .

Shading

Let’s move on to the second common claim, that thin film PV are more immune to shading effects. There is no magic physics in thin film technologies that make them less tolerant to PV’s number one enemy – partial shading, except that the cells in thin film panels are usually very long and narrow (5 to 10mm wide and the whole length of the panel). In this case, the likelihood of total cell shading is diminished, provided that the installer has correctly oriented the solar modules. Most modern thin film solar modules have further split the narrow cell into multiple sections and incorporated by-pass diodes (4). Nevertheless, if they are not oriented wisely to avoid potential shadows, then it is back to square one (See figure 3).

Figure 3a: Correct orientation to shading           Figure 3b: Incorrect orientation to shading
(Source: Technical Note – Optimising Thin-Film Module PV Systems by SolarEdge)

High Temperature

Finally, how about the claims for superior heat resistance of thin film PV? This is perhaps the only undisputable advantage of thin film technologies – intrinsically, they all have a better temperature coefficient compared to s-Ci (5). In other words, their performance does not degrade as quickly as s-Ci when cell temperatures increase above 25oC.  However, as figure 4 shows, different thin film technologies display a wide variation in temperature response.  Amorphous Silicon (a-Si) is least affected by temperature, whereas CIGS solar panels are very similar in performance to crystalline Silicon.

Fig. 4 Variation of Power Output with Temperature for Different Solar Technologies
 Source: Virtuani. A, Pavanello. D and Friesen. G. Overview of Temperature Coefficient of Different Thin Film Photovoltaic Technologies, 25th European Photovoltaic Solar Energy Conference and Exhibition. 2010, Spain.

A comparative study between amorphous silicon and crystalline silicon suggests the benefit can be up to 20% more output on a hot day with an average ambient temperature of 34oC. See Figure 5. (Note: the tests were published by NexPower, a manufacturer of amorphous Silicon panels).

Figure 4 (Source: NexPower In-house test report)

Although the above result may sound impressive, you may be wondering which parts of the world regularly has an average ambient temperature above 30oC. Unsurprisingly, some research bodies in countries likes, Thailand (6) and India (7), have recommended thin film PV for precisely this reason.

In Summary

Bringing all these factors together, a collaborative research project carried out by Universities of Stuttgart and Cyprus compared thin film PV and c-Si by measuring actual performance over many years in Cyprus (8). The data has obviously taken into account all the differences in spectrum responses and temperature coefficients, the results are summarised in Figure 5.  Data for four years is presented from 2007 (labelled a) to 2010 (labelled d). The clear conclusion from this multi-year side by side test is that thin film modules do not outperform crystalline silicon modules.

Figure 5 Muli-Year Comparison of Solar Energy Yield from Different Technologies
(Source: Reference 8 – page 222)

There appears to be no clear technological advantage for thin-film PV against c-Si at present. In order for thin-film PV to experience a revival, there must be other factors involved which would make thin film PV more attractive than crystalline silicon solar PV. 

For example the homogenous appearance of thin film panels may make them look more appealing.
Thin film solar can be printed on any thickness of substrate and combine with other materials to form see-through graphics, stained glass, company logos, and blinds. With the ability of being semi-transparent, they could even mimic the appearance of natural materials, for example wood or marble.

Needless to say, apart from such niche applications, thin film PV also needs to gain more headroom in cost advantage against c-Si to offset a lower overall efficiency. Until then, it seems like c-Si will stay on top for now.

References:

(1) http://www.marketsandmarkets.com/Market-Reports/thin-film-pv-31.html
Renewable Energy Sources and Climate Change Mitigation: Special Report of the Intergovernmental Panel on Climate Change
(2) Photovoltaics Report by Fraunhoher Institute for Solar Energy System. 6 June 2016www.ise.fraunhofer.de
(3) FLINS Project (www.flins-projekt.de) hosted by Universitat Oldenburg, Germany (http://www.uni-oldenburg.de/en/physics/research/ehf/energiemeteorology/research/former-projects/flins/).
(4) Correspondence with NexPower (www.nexpw.com )
(5) Overview of Temperature Coefficients of Different Thin Film Photovoltaics Technologies by Alessandro Virtuani, Diego Pavanello, Gabi Friesen at 5th World Conefrence on Photovoltaic Energy Conversion, Spain (https://www.researchgate.net/publication/256080289)
(6) Investigation on Temperature Coefficients of three types Photovoltaic Module Technologies under Thailand Operating Condition by P. Kamkird, N. Ketjoy, W. Rakwichian and S. Sukchai. Published on Procedia Engineering 32 (2012) 376 – 383.
(7) Variation of Temperature Coefficient of different technology Photovoltaic modules with respect to irradiance by P. Dash and N. Gupta. Published on International Journal of Current Engineering and Technology, Vol. 5, No. 1 (Feb 2015).
(8) Performance of Photovoltaics under Actual Operating Conditions by G. Makrides, B. Zinsser, M. Norton and G. Georghiou (pages 201 to 232). Third Generation Photovoltaics ISBN 978-953-51-0304-2. March 2012.

Mono vs Polycrystalline Solar cells - Myths Busted

Customers often ask what's the difference, but the old certainties have gone. 

Monocrystalline have missing corners, polycrystalline cells are square : Myth

Monocrystalline solar cells are cut from a large single crystal of silicon. The process by which this crystal is grown is remarkable. It is drawn from a molten crucible of liquid silicon by dipping in a 'seed' crystal and then slowly pulling this away from the liquid surface and rotating it.  By carefully controlling the temperature gradient in the crucible and the speed of withdrawal it is possible to create a solidified single crystal with the same atomic orientation as the seed.

If this cylindrical crystal were sliced to produce silicon wafers, they would be round and this would leave gaps when you tried to assemble them together into a solar panel.  So the cylinder is first cut along its length on four sides to make its shape closer to a square in cross-section.

There's a compromise here. The more you slice off, the closer to a square shape you get, and the more working area you can squeeze into your monocrystalline PV panel. The less you slice off, the less material you waste and the cheaper are the cells to manufacture.  The compromise that most manufacturers have reached is to make a shape that was a square with rounded corners (pseudo-square).

By contrast, a polycrystalline silicon wafer is made by melting the silicon feed stock, pouring it into a cube shaped mould and letting it cool and solidify.  The resulting block of silicon is sliced into pillars and these are in turn sliced into perfectly square cells.

So one difference between mono and poly is the characteristic shape of each; Poly are square and mono have missing corners.

Not any more!

The trimmings from cutting and slicing the silicon are no longer wasted; they are re-cycled as a material input for polycrystalline cell production. Some manufacturers now offer mono crystalline panels with full square cells.
  

Monocrystalline cells have an even black colour, polycrystalline are patterned and blue: Myth

When the polycrystalline ingots solidify in their mould, crystals start to form in many, many different places (nucleation sites) and grow until they meet up with other crystals.  The orientation of the atomic structure in each crystal is random and is different from its neighbours. When you slice though the ingot to make the wafer this creates a characteristic pattern, a kind of metal flake effect, on the surface of the solar cell because each crystal reflects the light differently. The cells also have a bluish colour. By contrast, mono crystalline cells have a homogeneous atomic structure throughout and have an even black colour.

Not any more!

High performance solar cells are now treated during processing to create pyramidal micro structures on the surface which improves light absorption.  Anti-reflective coatings are added to reduce light reflection from the surface. Both polycrystalline and monocrystalline cells can be made to look matt black with an even colour.

Monocrystalline panels are more efficient : True - well, sort of

The boundaries between the crystals in a polycrystalline cell (grain boundaries) can impede the flow of electricity, so mono crystalline cells (which have no grain boundaries) have always had higher efficiency. However, polycrystalline  cells have been closing the gap in recent years and the point has  just about been reached where the additional active surface area from the square cell shape in a polycrystalline panel makes up for the lower efficiency in the cell itself.

Check out this table.

It shows the product range from one of the world’s largest manufacturers.  Power is given in Watt-peak (Wp), the power output under standard test conditions. 

If you compare the standard mono and poly products (code 6/60 models), you can see the range of peak power output runs from 250 to 270Wp for the mono panel and from 245 to 265Wp  for the poly panel.  The difference is 5Wp, or 2% less power for the polycrystalline.

Monocrystalline/Polycrystalline  panels work better in low light conditions : No evidence

I have read many claims that one type of panel works better than the other in low light conditions, and writers on other websites seem to be evenly split in whether it is monocrystalline or polycrystalline that is best (presumably depending on which they sell).

I have been unable to find evidence to support these claims (in either direction…).

Until I see some evidence, I’m going to mark this one down as a myth!  Please let me know in the comments below if you know about this.

Monocrystalline panels have better high temperature performance : True – though marginal

Looking again at the table, the right hand column shows the Power Temperature Coefficient.  This is the rate at which the panel power output falls as its temperature rises.

Polycrystalline panels do indeed lose their power output more quickly, by about 0.02% more per degree C.  But what does this mean in practice? 

If, for example, a Monocrystalline solar panel were operating at 70C on a hot and sunny day, it would be producing 0.41 x (70-20) = 20.5% less power than is measured under standard test conditions (20C).  By contrast a polycrystalline solar panel could be producing 0.43 x (70-20) = 21.5% less power.

All other things being equal, polycrystalline panels would produce 1% less power at the elevated temperature.  But that is a very different thing from saying it would produce 1% less energy over a year of operation.  It’s not hot and sunny all day every day; in fact conditions to produce a 70C operating temperature are rare.  The energy penalty from choosing polycrystalline solar panels over monocrystalline would depend on climate, but will be far less than 1%.   Although there would be a penalty, it’s pretty marginal.

Polycrystalline panels are cheaper, monocrystalline are more expensive  : True, on average

The argument often goes that because the process of producing monocrystalline cells is more complex and involves more wasted material, they’re more expensive to make.

However, just because something is more expensive to make, doesn’t make it worth more to the customer.  The reason that monocrystalline panels command a price premium is that more people prefer the way they look and the panels have a higher power.  Having a higher power panel means you save money on other costs like racking and fixings for the same total energy output.  It also means that you can squeeze more energy out of situations where the area to place the panels is limited or expensive.

The PHOTON module price index report for November 2014 has average spot market prices for solar panels in Europe as follows.  (Prices are always given per watt-peak, Wp, so you can compare based on the power output).

Monocrystalline solar panels   0.65 EUR/Wp           (Range 0.48 – 0.95)
Polycrystalline solar panels     0.55 EUR/Wp           (Range 0.40 – 0.82)

So yes, on average monocrystalline solar panels are 18% more expensive on a per-watt basis, but the range of prices show that it’s perfectly possible to buy polycrystalline panels at the higher end of the market for a much higher price than the monocrystalline panels at the lower end of the market.

Conclusion

The old certainties are disappearing.  At the high end of the market, monocrystalline and polycrystalline solar panels are becoming more and more alike in aesthetics and performance.  If this trend continues, with black polycrystalline cells and square monocrystalline cells of similar performance, then average prices will converge too.

In mature solar markets, the domestic rooftop market starts to demand good looking solar panels, and has settled on solar panels with black cells and black frames with improved aesthetics.  For this market, your choice of solar panel will be far more about choosing a quality brand that you trust than worrying about whether those panels followed a polycrystalline or monocrystalline manufacturing route.

In-roof Solar PV - Hot or Not?

Results from a study into heating effects on in-roof solar modules

In-roof solar PV may look great, but what's the trade off on energy performance?
Image: Viridian Solar and Elliott Brothers

Every solar professional knows it.  The power output from a crystalline silicon PV module reduces as it gets hotter.  PV systems installed in-roof suffer from lower ventilation rates on the shade side than modules installed on a rack above the roof covering and therefore produce less energy.

But how much less?  Is it a lot or a little?

How much of the heat is actually lost from the shade-side of a panel, compared to that from the sun-side?

I was recently in a meeting at the Solar Trade Association with a group of the top technical brains from the UK PV industry.  I asked the group to estimate the increase in annual energy yield by switching a crystalline silicon PV module from a sealed in-roof installation to an installation above roof.

The answers ranged from 1% to 14%. 

Conversations with other solar professionals have produced estimates as high as 25%.

Everyone knows it has an effect. But no one seems to know by how much.

Having a quantitative, evidence-based answer to this question is becoming more and more relevant in our industry.  As the solar market matures more and more customers for solar PV want the benefits of reduced energy bills but without compromise to the looks (and potentially re-sale value) of their properties.

In-roof systems offer an alternative that ticks the box on aesthetics for many people at a price they are willing to pay, but just how big is the trade-off on energy yield? 

Now researchers at Viridian Solar, collaborating with the Engineering Department at Cambridge University and Enphase Energy have produced an answer to this question.

The authors are aiming to publish the research in a peer-reviewed journal later this year, but a briefing document has been released summarising the experimental results.

 

Replacing Opinion with Evidence

 

The experiment is described in more detail here, but in simple terms it consisted of three steps:

 

• Build a test rig with PV modules installed in a range of situations representative of real life construction

• Understand the relationship between weather conditions and module operating temperature for each type of installation

• Use the experimentally derived temperature profiles to calculate the annual energy yield for each installation situation.

 

Test Rig

 

Clearline PV15 modules were installed in five different ways

1. Free standing on an open framework (rear fully open)

2. Above a pitched tiled roof on a metal framework (open gap between panel and tiles)

3. Integrated in a pitched tiled roof with cold-roof construction behind (batten-space ventilation)

4. Integrated in a pitched tiled roof with warm-roof construction behind (insulation between roof joists)

5. Integrated with a pitched shingle roof with plywood sarking board (module rear un-ventilated)

The images below show the roof build up for two of the pitched roof installations.

 

 

Temperature Rise

 

The graph below shows the temperature response of each of the installation types - the lines show the operating temperature above ambient as the light levels increase.  As expected, the temperature of a module with less ventilation to the shade side rise faster as light levels rise.

 

 

 

For example, at 1,000 W/m2 (a bright sunny day with sun directly onto the module) the rack-mounted module above the pitched roof was 10 degrees C warmer than the free standing module.  The integrated module in the cold roof is a further 9 degrees C warmer than the rack-mounted module.

 

Clearline PV modules have a power-temperature coefficient of -0.509 %/degree C, quite typical for a crystalline silicon module, so a reduction in temperature of 9 degrees would produce a power increase of 4.5%.

 

However it's not always sunny, and the sun isn't always directly onto the module.  In fact, in the UK irradiation levels higher than 1,000 W/m² are very much the exception and not the rule.  At lower levels of light, the temperature difference between different installation types is smaller.

 

Annual Energy

 

So, what's the answer?  What was the annual energy benefit for rack-mounted systems compared to the in-roof systems in the experiment?

 

The temperature characteristics were used with a climate file for Cambridge, UK and the power-temperature coefficient for the modules to calculate the annual energy yield for each installation.

 

It turns out that a rack-mounted module would yield 3% more energy than a roof-integrated module. 

 

 

 



 

 

 

 

 

Clearly, for some situations an extra 0.3% return on investment due to energy yield will matter, but for many domestic customers minimising the visual impact on their building will be more important. 

 

As an industry at least we now have facts to present to potential customers so that they can make an educated choice.

 

 

 

 

 

 

Consumption Assumption – The Proportion of Solar Electricity Used in a Home

The answer is 50%, now what was the question?

Three very different distributions,
each with the same average - 50%

If you have a solar PV system fitted to your home, then the amount of electricity you generate is changing all the time as the sun travels across the sky and clouds come and go.  Your electricity use also changes as you switch electrical appliances on and off.  

If solar electricity production exceeds your electricity use at any moment, power flows out of your home to supply the electricity grid (called export).  If, on the other hand, your electricity use exceeds the electricity produced by the solar panels then all of the solar electricity is used in your home and supplemented by power drawn from the grid.

Over the course of the day your electricity meter will record the flows of electricity into your home (for which you will be billed). If you can meet your electricity needs with a unit of electricity from your solar panels, then you save yourself buying that unit from your supplier.  

This saving on electricity bought represents an increasingly large part of the financial benefits of owning a solar PV installation.

So what percentage of the electricity that is generated by a solar PV system is actually used in a building?

It seems that the UK solar industry has found a rather simple way to answer this question:

Q: Retired couple at home during the day, 1kWp system

A: 50%

Q: Working couple no-one at home during the day, 1kWp system

A: 50%

Q: Working couple no-one at home during the day, 4kWp system,

A: you guessed it, 50%

For smaller residential solar systems claiming the Feed in Tariff (FIT) , the government decided that it wouldn't measure exported solar electricity, only generated electricity.  However, the design of the scheme was to reward people with money for everything they generate, plus a little more for the energy that is exported.  So it came up with a way of calculating the exported energy. A figure of 50% is multiplied by the generated energy to calculate a value for the exported electricity. 

The industry has embraced that figure as if it represents some kind of truth written on tablets of stone. Every presentation of financial benefits from the FIT I have seen assumes that the household will use 50% of the energy it generates. 

When the UK FIT was paying people £0.43 for each unit (kWh) of solar electricity generated and electricity cost £0.14/kWh all that really mattered to the financial calculation was how much you generated, not how much of it you used yourself (called self-consumption). 

PV prices have fallen dramatically and to keep the financial benefits consistent the FIT now pays new installations much less, about £0.15/kWh generated.  In the meantime electricity costs have risen to around £0.15/kWh.  Self-consumption is becoming the greater part of the financial benefit. 

To illustrate this, let’s consider a house with a 2kWp system installed. To keep the maths easy, we'll assume its in a very good location and generates 2,000 kWh per year.  What happens to the financial return as the proportion of energy used in the home changes?

How the make-up of Feed in Tariff Benefits has Changed Since Launch
Savings on electricity bills forms a much larger proportion of the financial calculation

When the FIT launched, any inaccuracy in the self-consumption figure was much less significant than today.  The value of the savings on the electricity bill represented only 14% of the total benefits (assuming this figure of 50%).

Now the proportion of the benefit that comes through electricity savings is 30%.  Inaccuracies in the self-generation assumption have a much bigger impact on the financial returns calculated.  If solar prices continue to fall then the FIT will fall too, and this trend will continue.

As an aside, the table above is absolutely not an advert for so-called PV switches that divert solar electricity to an immersion heater in your hot water cylinder .  Once you ‘degrade’ high value electricity to heat, you reduce its value to that of the cost of the fuel would have used to heat the water.  In 80% of homes this is gas, so it’s not 15p/kWh you’re saving, but more like 5p/kWh. 

(See the outspoken guest blog from Tom Seppings of solaplug for more on this point).

Far better to use it to power electrical appliances if possible, or store it in a battery, an emerging technology already causing great excitement in the industry.

Your Guess is a Good as Mine

How good a guess is 50%?  What evidence was the government’s choice based on?  If anyone has information on this, I’d love to hear it. Please post a comment at the bottom of the article.

Let's pause to consider averages for a moment. The graphs above show three 'distributions', each of which has an average of 50%. Only the top one (1) would strongly justify using the average value for all customers. 

My hunch is that the reality is most like the bottom distribution (3), with two distinct groupings - households with people at home during the day, and households where people are out at work and school during the day. 

I also suspect that the 50% figure was arrived at when the average PV system size was much smaller than it is today. Falling prices have brought larger installations within the reach of more people (capped by the 4kWp tariff band limit). In 2011, the EST had the average system size as 2.75kWp.  Today it is 3.75kWp. The larger the system, the larger the midday energy peak, and the less likely it is that there's sufficient instantaneous electricity use in the home to mop up the solar electricity. 

Yet the industry persists with presenting financial benefits based on this 50% figure.  

Does anyone know of any work that has been done on the profile of electricity demand during the day in homes with differing types of occupancy?  How could the solar industry offer its customers a better prediction of how much solar energy they will consume themselves?

Do you think solar PV systems will become smaller to increase the proportion of self-consumption as levels of FIT support falls and electricity prices rise?

Let's hear what you think - post a comment below.