Effective integration of photovoltaics into the built environment

Electrical energy can be generated from solar radiation using the photovoltaic (PV) effect. In the drive for an energetically sustainable world, PV systems provide the most promising form of sustainable energy and the only type suitable for use on large scales. Two main complementary visions exist for harnessing its potential. The first involves the construction of big, centralized PV power plants to replace conventional fossil-fuel or nuclear-power installations. The second makes commercial and residential buildings self-sufficient by means of PV modules on roofs or in façades. This decentralized and more democratic option of ‘building-integrated photovoltaics’ has grown rapidly in recent years to become a significant part of the PV market. The technology's long-term economic viability requires that it meet the needs of both building construction and energy generation.

At present, PV systems are usually mounted on the roof, requiring some small adjustments to the building structure. This is expensive, inefficient, and unsightly. It is more effective to use a PV module that can both produce solar energy and work as building element. Some steps have already been taken toward this goal by fabricating alternative roof tiles¹ or shingles² with embedded solar cells. We wanted to improve on this and integrate the two elements more fully by depositing a thin-film solar cell directly onto a conventional building element, such as a ceramic tile.

The main advantage of thin-film compared with conventional crystalline-silicon solar cells is that the PV device, a few microns in thickness, can be deposited using a low-temperature process (e.g., plasma-enhanced chemical-vapor deposition) onto a variety of substrates, such as glass, steel, or ceramics. It is therefore possible to design a complete industrial production process to turn a standard building material into a working PV device, such as a PV ceramic tile.

Figure 1. PV mini-module tiles with areas of 5×5cm² (left) and 10×10cm² (right). Solar-cell devices are defined by a violet-colored zone with metal contacts on the top.

Clearly, commercial ceramic tiles are unusual substrates for PV-energy generation. High-quality ceramics, such as mulllite or alumina, have already successfully been used as substrates for thin-film solar cells.^3,4 In contrast, we used commercial tiles ‘as received’ from industry. There are problems with contamination, high roughness, and surface irregularities, such as pores and holes of 1–50μm across. The deposition of homogeneous and high-electronic-quality thin films on this kind of material could be very difficult.

Figure 2. Current-voltage diagram for PV mini-module tiles of 10×10cm²under illumination. η represents the conversion efficiency.

Both chemical contamination and surface irregularities could lead to severe problems with ‘shunting,’ i.e., unwanted short circuiting between the front- and back-surface contacts of the solar cell. Nevertheless, we have fabricated working prototypes of 1cm²amorphous-silicon (a-Si) solar cells with a conversion efficiency of over 4%.⁵ Layers were deposited onto 5×5cm² tiles: tile/conductive layer/a-Si p-i-n/indium-tin oxide (ITO)/silver grid, where p-i-n refers to the doping of an active a-Si layer between the p- and n-type substrates of the diode. Figure 1 (left) shows a tile with nine devices, six of which have a silver grid contact on top. We anticipated that larger devices would run into problems with shunting, because the substrate would be more significant. Figure 1 (right) shows recently fabricated larger-area devices on 10×10cm² tiles. A mini-module is formed from four series-connected 7×1cm² cells. As anticipated, many devices were affected by shunting.

To overcome this, we tested many types of back contact layers and found that an ITO back contact was superior to a layer of metal or a combination of ITO/metal or metal/ITO as back layers. We reasoned that ITO worked better alone because it has a higher sheet resistivity than metal. If we assume that there are many local shunting paths randomly distributed all over the device area, the high sheet resistivity of ITO limits the effect of each path on the device's shunt resistance. We therefore developed a device structure with a screen-printed silver back contact, where the metal front bus bar does not overlap with the metal back contact, avoiding metal-to-metal direct shunting paths.⁶ This ‘trick’ made larger working devices with initial conversion efficiency of (2.0±0.2)% possible. Figure 2 shows current-voltage curves under illumination for some modules with the new contact arrangement onto different tiles.

It may be possible to increase the conversion efficiency to 5% for the mini-module tiles shown in Figure 1 with just an optimization of back- and front-contact design. We hope to increase the efficiency further by improving the deposition conditions. The next step would be to fabricate much larger-area (>1m²) PV tiles ready for use as ‘PV bricks’ in the building industry.