The battles in the solar industry at present are typical of an emerging industry as companies compete for market share as well as touting their technology as the answer to the industry challenges. But what is the true state of solar technology and how does it stand up to some objective observation. St.J. Dixon-Warren and Tim White of Chipworks discuss their findings when they looked under the hood of CIGS Solar Panels.

The photovoltaic industry seems at the moment to be in one of those knock-down, drag-out contests typical of the start-up phase of a disruptive technology. Not just the business, but also the variety of technologies themselves – monocrystalline silicon versus polycrystal, bulk materials versus thin film, thin-film silicon versus CdTe (cadmium telluride) versus CIGS (copper-indium-gallium selenide).

At present, crystalline silicon panels account for the vast majority of the market; however, they remain expensive and are unlikely to ever be able to compete with electricity from the grid without subsidies. Thin-film panels have the potential for lower cost production, since the active layers are deposited directly onto molybdenum coated glass.

Vacuum deposition, including co-sputtering and coevaporation, and non-vacuum based methods are possible, including electroplating. Sputtered films can be either deposited in layers and then alloyed through thermal annealing cycle or co-sputtered using a mixed target.

Chipworks has recently completed a detailed structural analysis of two commercially available CIGS panels, namely the Wurth WSK0020 and the Avancis PowerMax 100 FB. Broadly speaking the panels have a similar structure. They are both made as a sandwich of layers between two glass sheets. The bottom sheet provides a substrate and the top provides protection from the external environment. The optically-active CIGS stack is deposited on the bottom glass substrate. The CIGS stack is comprised of a bottom molybdenum layer, a copperindium- gallium-selenide (CIGS) layer, and a top aluminumdoped ZnO transparent conductive oxide (TCO) layer.

The bottom molybdenum layer is the cell anode and helps reflect any unabsorbed light back into the active layers. The top aluminum-doped ZnO layer, as well as being the electrically conductive cell cathode, is transparent to allow the sunlight to penetrate into the CIGS layers. Scribe lines are cut into the layers of the CIGS stack to form the individual cells. A layer of transparent polymer material is applied over the CIGS stack followed by the top glass.

Figure 1 is a schematic diagram of two individual cells in a typical CIGS solar panel. It turns out that, in terms of light to electricity conversion efficiency, the optimum width of CIGS cells is around 5 to 10 mm so panels are typically formed using banks of such cells in series. Scribe lines, denoted P1, P2, and P3, are used to create the interconnect structures that form the series connections. Electrical anode and cathode connections are then formed at each end of the series of cells.

 



Figure 1 CIGS Solar Panel Structure

The PN junction apparently forms near the top surface of the CIGS layer as a result of surface segregation of indium rich materials, such as CuIn3Se5, causing the conductivity type to invert from P-type to N-type, and thus forming a PN junction.

The molybdenum film provides the interconnection and the anode back contact for each cell while the N-type, ZnO:Al TCO is the cathode contact. A layer of CdS and an undoped ZnO buffer layer are often used to reduce lattice mismatch between the CIGS layer and the TCO.

The cleaved edge of one of the CIGS solar panels recently analyzed by Chipworks is shown in Figure 2. The bottom glass is usually (cheap, easily available) soda-lime window glass and appears green in color, due to the presence of sodium. Sodium diffusion into the stack from the glass has been shown to improve cell performance.

 



Figure 2 CIGS Solar Panel Cleaved Edge

Figure 3 shows a cross-sectional view of a P1 scribe line, which provides the electrical isolation between anodes of adjacent cells, as shown schematically in Figure 1. The TCO, CIGS and Mo layers are easily distinguished in the cross-section. The Mo layer was likely laser scribed, which results in a thickening for the layer near the edges.

 



Figure 3. P1 Scribe Lane

A detailed SEM cross-section of the CIGS stack is shown in Figure 4. The top surface of the CIGS layers has a very rough morphology, which is partly re-planarized by the TCO. The CdS and ZnO buffer layer were also resolved in the analysis. The compositions were determined by SEM and TEM-based energy dispersive Xray analysis.

 



Figure 4. CIGS Stack

While the essential nature of the CIGS cells was the same, Chipworks’ analyses of the two competing panels inevitably found differences in the details of their structure, including variation in the layers used for the optically active CIGS stack, the materials used for the transparent polymer layer, and the methods of contacting the active layers. Such is the stuff of competition; as with the broader photovoltaic market, cost per installed watt will be the final arbiter.