John I B Wilson and Robert R Mather, Power Textiles Limited, Upland House,Ettrick Road,Selkirk,TD7 5AJ,UK.
A Helena N Lind and Adel G Diyaf, School of Engineering and Physical Sciences,Heriot-Watt University,Edinburgh,EH14 4AS, UK.
Present day photovoltaic (PV) modules are mostly rigid panels of standardised sizes that are not readily integrated into buildings or other large structures without compromising the architect’s design. Although thin-film solar cells have been available for many years, they are often based on metal or glass sheets and do not provide any adaptability in shape. We are developing thin-film solar cells that are fabricated directly on woven polyester fabric in an effort to address these limitations of conventional PV modules. After a brief explanation of how photovoltaic cells operate, we describe how the required layers are produced with plasma enhanced chemical vapour deposition (PECVD) and other coating technologies. Finally we envisage how the performance of these innovative renewable energy devices is anticipated to improve.
Solar cells use the photovoltaic effect in some form of P-N junction diode to generate electricity from sunlight, without moving parts or chemical reactions. These renewable energy converters are now widely used but mainly in the form of rigid glass and metal modules that enclose crystalline silicon cells (See Figure 1), interconnected in a series/parallel arrangement to provide the current and voltage required. These silicon cells are based on the same semiconductor material that is used for integrated microelectronics and therefore use costly, high purity, single-crystal wafers; a common alternative is multi-crystalline silicon which uses similar material but cast into large blocks rather than grown as a large cylindrical ingot of silicon.
Each of these processes needs high temperatures and consequently a great amount of thermal energy, and each also wastes some of the pure material during slicing into thin wafers. If instead, the silicon is formed directly into a thin film from a gaseous chemical source, the thermal energy and waste are considerably reduced, which should improve the energy payback period for solar cells. Of course, by taking this route there are not only gains but also a shortcoming, namely that the performance of thin film silicon cells is lower than that of crystalline silicon cells.
2. Thin film solar cells
At present the most common thin film cells use silicon, in a form which is actually a compound of silicon and hydrogen known as amorphous silicon (a-Si:H). This is usually deposited on to glass (see Figure 2) or thin metal substrates and the cells have an efficiency for converting sunlight to electricity of around half that of crystalline silicon cells . Alternative semiconducting materials are available but they also have some drawbacks, including the less developed state of their technology and some concerns about abundance of raw material. Compound semiconductor thin films cells include cadmium telluride (CdTe), gallium arsenide (GaAs) and copper indium gallium selenide (CuInGaSe2 known as “CIGS”). These approach similar conversion efficiencies to crystalline silicon or may even surpass this, but they generally require much higher processing temperatures than amorphous silicon. Finally, organic cells using polymer films have a similar efficiency to amorphous silicon cells and the capability of a much simpler and cheaper liquid-based process, but at present they are only made in very small sizes and are highly susceptible to both oxygen and moisture.
Fig.1 Conventional roof-mounted PV modules. Fig.2 Rigid 10W a-Si:H PV module on glass.
3. Thin film amorphous silicon
So what is the biggest difference between all the methods used to synthesise these semiconducting materials? The answer is that the necessary reactions usually require heat for melting or evaporating the source materials, but by using an electrical discharge or “plasma” some of this thermal energy may be replaced by energetic electrons. Such a “plasma enhanced chemical vapour deposition” (PECVD) process may be made to work well for a-Si:H, eliminating the need to form crystalline ingots and subsequent slicing into wafers. PECVD uses a gaseous source of silicon, silane SiH4, and will operate at around 200-250oC whereas forming silicon into ingots requires a temperature of more than 1400oC. The compounds mentioned above are commonly formed as thin films by vaporising their elements, each of which has a relatively high melting point.
PECVD amorphous silicon is deposited on to any substrate that can withstand heating to ~200oC whilst being unaffected by the flux of free electrons and activated chemical species formed by the plasma. This all occurs at low pressure, requiring a set of vacuum pumps to remove the air before the source gases are admitted to the reaction chamber (see Figure 3). A radio frequency (RF) voltage applied to two flat parallel plates drives the reaction that produces a thin film of silicon (on any surface exposed to the low pressure plasma). The substrate is placed on the grounded plate and heated whilst the RF supply is connected to the opposite plate. When the RF supply is switched on the gas breaks down and there is a visible glow between the plates. In addition to silane, very small amounts of either a phosphorus or a boron containing gas are added to form the P-N junction required for a solar cell. All solar cells use an inbuilt electric field to separate the pairs of negatively and positively charged electrons and holes that are generated by light, and it is this structure that provides the cell’s output voltage.
Fig.3 Schematic diagram of plasma enhanced chemical vapour deposition system for a-Si:H films .
Because the pairs of photo-generated charges are very prone to recombining in a-Si:H, it is essential to have a built-in field that extends across most of the cell, which is usually achieved by using a three-layer P-I-N structure. Of course, if we could manufacture thin film crystalline silicon this would have better electrical properties than amorphous silicon, but this can only be approached and is not possible at low temperature: by using a different form of plasma and a different gas mixture, we can make nanocrystalline silicon (nc-Si:H) which contains very small crystallites in an amorphous matrix: this has the high optical absorption of a-Si:H and some of the electrical advantages of crystalline Si.
4. Flexible solar cells
In addition to increasing the performance of solar cells or reducing their cost, there is an interest in alternative constructions. The market would be open to foldable or rollable modules, or to otherwise flexible devices, not necessarily to compete with existing types. There is a limited range of such modules at present, generally using Si or CdTe as the active material and stainless steel foil or polyimide as the substrate (see Figure 4). In development are polymer cells on polyester but as noted above, these require excellent hermetic sealing to prevent rapid degradation.
Fig.4 Non-rigid 7W a-Si:H panel.
We are attempting to put a-Si:H cells on to woven fabric, not by bonding or laminating conventional crystalline cells as is commonly done, but instead by directly depositing the cell layers on to the fabric itself (see Figure 5). By using PECVD we are able to use polyester, perhaps the most widespread polymer fabric, as it is stable to above the deposition temperature of a-Si:H and is not affected by the plasma itself. The cell is a multi-layer device based on making the fabric electrically conducting whilst maintaining its conductivity when flexed  (see Figure 6). We have developed procedures to deposit a conducting polymer layer from a liquid precursor, over which we add a thin evaporated layer of aluminium. This pair of films enables small breaks in the less flexible, but more conducting aluminium, to be bridged by the more electrically resistive, but still conducting polymer.
Fig.5 The different processes in manufacturing thin film a-Si:H cells on textile substrates .
Fig.6 Vigorous bending test of a sample of Al/conducting polymer/polyester fabric showing retention of electrical conductivity after repeated cycles of rough handling. 
Over this we add the three P-I-N a-Si:H layers by PECVD and complete the basic cell by a sputtered transparent conducting oxide layer, at present of indium tin oxide (ITO) but in future this could be aluminium doped zinc oxide (AZO). Finally the cells are encapsulated in a polymer coating by one of several alternative techniques, according to the end use and required life span. (See Figure 7.)
Fig.7 Schematic diagram of the multiple layers of a flexible cell on woven polyester (upper picture); the magnified appearance of a woven polyester fabric after coating successively with the conducting and semiconducting films (lower left picture); small area flexible a-Si:H cells on polyester fabric (lower right picture).
These cells are of much lower conversion efficiency than their rigid counterparts but they target different applications such as shade awnings, tensile architectural membranes, tents and horticultural uses. We presently use a small area batch process to demonstrate the feasibility of the fabrication but roll-to-roll coating is an obvious way forward for manufacturing larger areas.
Another advantage of forming cells directly on a fabric is the option for integrated connections between cells. Individual solar cells deliver less than one volt and so must be connected in a series string to increase the output voltage to a more useful value. With conventional crystalline cells, this is achieved by bonding metal strips between cells mounted in a supporting frame. With thin film cells there may be the option for interconnecting them as part of the fabrication process, on a substrate that supports several cells. With flexible cells on textile we are able to do much the same integrating of cell connections with evaporated metal forming the contacts between the top of one cell and the bottom of the adjacent one, in a string of cells. Furthermore, we can alter the shape of cells by moveable masks that define the areas to be coated at each step. The output current from a cell depends on its area so in a series string of cells, each one must have the same area although they may be different shapes according to the artistic design.
5. Performance and future improvements
The challenges of making successful solar cells on a fabric are now being met. The first was to render a woven fabric electrically conducting over selected areas before attempting to deposit silicon. A dual layer overcomes any tendency for minor breaks in a single conductor to extend across the warp or weft of the fabric. PECVD of a Si P-I-N structure has been successful without major changes to the preparative conditions used for other substrate materials. By replacing the RF plasma generator with a microwave power source and making some changes to the gas mix, the silicon will take up a nanocrystalline form, which surprisingly appears to be hexagonal rather than the usual cubic crystal structure . At present we are introducing a short treatment to enhance wettability of the fabric before it is liquid-coated with conducting polymer, and adding an extra layer to prevent intermixing of aluminium and silicon.
Other changes will probably be needed as the development continues. Present performance of these solar cells is only a few per cent energy conversion efficiency, well below that of equivalent cells on glass, and the photocurrent in particular should be improved. Although a-Si:H absorbs light strongly, it is expected that the uneven texture of the fabric enhances the absorption of any light that reaches the back surface, without adding the special texturing that is needed for this purpose on smooth substrates. It may be possible to increase the efficiency by using a double or triple stacked cell, a structure that is available with PECVD silicon films by changing the gas mixture for each layer and so changing the composition of the semiconductor in each cell. Such tandem cells generate a higher voltage than a single cell but have much the same output current.
Flexible solar cells formed directly on textiles have the capability for diverse applications that are not satisfied with heavy, rigid solar panels but they may not be able to compete directly with more conventional cells as they have lower conversion efficiency than other types. There is still further improvement to be had by fine-tuning the various thin layers that form these devices and still retain the low energy inputs from thin materials and low temperature processing that lead to short energy pay-back periods.
We especially thank the Scottish Government (Scottish Enterprise) for the award of SMART: Scotland funding, the UK Engineering & Physical Sciences Research Council for supporting AHNL and the Libyan Government for supporting AGD during their PhD studies, and J & D Wilkie Limited for supplying polyester fabric.
1. M. A. Green, K. Emery, Y. Hishikawa, W. Warta and E. D. Dunlop, Solar cell efficiency tables (version 41), Prog. Photovolt: Res. Appl. 21 (2013), 1–11.
2. S. Jardine, Thin Film Silicon on Textiles by Microwave Plasma Chemical Vapour Deposition, PhD Thesis, Heriot-Watt University (2006).
3. A.H.N. Lind, Deposition and Characterisation of Silicon and Conductive Layers on Woven Polyester, PhD Thesis, Heriot-Watt University (2013).
4. A. G. Diyaf, R. R. Mather, and J. I. B. Wilson, Contacts on Polyester Textile as a Flexible Substrate for Solar Cell, PVSAT-9, Proceedings, editors M. G. Hutchins, A. Cole, T. M. Watson pp119-122 (Swansea, 2013).
5. A. H. N. Lind, J. I. B. Wilson, and R. R. Mather, Raman spectroscopy of thin-film silicon on woven polyester, physica status solidi (a), 208 (2011) 2765-2771.
The Authors: JIBW and RRM are directors of Power Textiles Limited, which they founded to develop flexible solar cells; both were previously academic members of Heriot-Watt University; AHNL and AGD are PhD students at Heriot-Watt University researching thin film solar cells.