One Wavelength Fits All
Structuring of Thin-film Solar Cells with a Single Laser Wavelength
Structuring of a PV module into a number of cells is necessary to lower the current and to increase the voltage, and is typically accomplished with nanosecond laser pulses of different wavelengths. Due to the many available laser sources, complex and expensive scribing setups are necessary. To overcome this a concept for laser structuring of thin-film PV modules using a single wavelength allows prediction of the ablation behaviour for a given laser pulse energy.
Fig. 1: Structuring a PV module into a number of cells is necessary in order to lower the current and to increase the voltage. This structuring involves the complete and selective ablation of very fine lines within the thin-film layers and is typically accomplished with nanosecond laser pulses - Picture: © Yang MingQi Fotolia.com
Economically viable photovoltaic (PV) power generation requires low-cost solar cell devices. Thin-film PV cells are being developed as a means of substantially reducing material and energy consumption costs during manufacture. A thin-film PV module consists of an active (absorber) layer sandwiched between two electrically conductive layers on a substrate, which is typically glass. Depending on the type of PV module, the active layer can consist of hydrogenated amorphous and microcrystalline silicon (e.g. a-Si:H/µc-Si:H tandem structures) or of polycrystalline compounds like Cu(Inx,Ga1-x)(Sey,S1-y)2, i.e. CIGS, and CdTe.
Structuring the module into a number of cells is necessary in order to lower the current (which is proportional to the area of the module and can easily reach values as high as 100 A) and to increase the voltage (which is not area dependent but rather limited by the splitting of the quasi-Fermi levels). This structuring involves the complete and selective ablation of very fine lines within the thin-film layers to form serial interconnections between the front and back contact while electrically isolating the individual solar cells. As shown in Fig. 2, three scribing steps P1, P2, and P3 alternating with layer deposition are necessary to separate the solar cells and to perform so-called monolithic interconnection.
Fig. 2: Three scribing steps P1, P2, and P3 alternating with layer deposition are necessary to separate the solar cells and to perform so-called monolithic interconnection
Typically, in industrial production of Si and CdTe modules these scribes are performed by nanosecond laser pulses of different wavelengths. In the case of CIGS only the P1 scribe is done by a laser, whereas P2 and P3 are still done mechanically by needles. Though lasers are considered as a rather expensive solution for structuring the modules, their application has turned out to be highly advantageous, since it has been proven to enable very precise and thin scribes with well-defined edges and thus to minimize the dead area zone of the modules.
Essential prerequisite for successful material removal is the proper adaptation of the accessible laser parameters, such as wavelength, power, pulse duration, and repetition rate, with respect to the material properties of the particular layer. In general, it has to be considered that the different absorptive properties of the individual layers require laser pulses of different wavelengths. Accordingly, when structuring Si thin-film tandem modules, green laser pulses (532 nm) are generally used for scribing of P2 and P3 (cf. Fig. 2), since this wavelength provides optimal absorption in the active Si layer as well as in the metallic contact layer. In contrast, for scribing the transparent conductive oxide (TCO) contact layer red laser pulses (1 064 nm) for absorption by free electrons are used. Also, blue laser pulses (355 nm) for band-band absorption are successfully applied when hazardous absorption in the glass substrate can be suppressed. Though a variety of suitable laser sources for industrial use is commercially available, complex and expensive scribing setups are necessary to provide laser light in the appropriate spectral range for complete cell structuring.
Fig. 3: Single laser pulses of different energy are directed onto the material until visually apparent material damage or ablation occurs
A single wavelength does it all
In order to circumvent these constraints we propose a simple concept for complete laser structuring of thin-film PV modules using just a single wavelength of 532 nm which allows a reliable prediction of the ablation behaviour for a given laser pulse energy. This prediction is based on an understanding of the interaction between laser pulses and the material, which is based on the laser parameters and the material properties. The basic idea is that each layer material has a certain ablation threshold, which has to be exceeded in order to remove the material. The corresponding fundamental laser parameter is the minimum spatial energy density that is required to achieve ablation, termed the threshold fluence. In addition, a knowledge of the material behaviour on repeated laser pulse exposure (i.e., incubation) allows prediction of the influence of the laser pulse overlap. Thus, as described in the following procedure for the case of laser structuring of TCO with green laser light, the desired processing parameters for scribing a line can be derived if the threshold fluence and the incubation of the material are known.
Laser parameters in detail
Solid-state lasers with pulse durations in the nanosecond range are preferred for industrial use. In the present investigation a Gen5 high speed motion system for high-precision laser processing (Spectra Physics) was used. This system consists of a linear motor drive and an air-mounted laser head panel that is equipped with six laser heads which can be moved at a speed of up to 1.6 m/s. PV modules under investigation are Si thin-film modules, which were prepared at the new facilities of the Competence Center for Thin-Film and Nanotechnology for Photovoltaics Berlin (PVcomB). For structuring with green laser light an Explorer was used operating at 532 nm and 2 W maximum. The laser pulse duration was less than 15 ns and depends on the laser pulse repetition with a maximum of 150 kHz. The laser is a Q-switched, diode-pumped Nd:YVO4 solid state laser with a pulse-to-pulse stability of < 3 %. To allow sufficiently large distances between the individual spots (up to 50 µm) the laser was operated at a 20 kHz laser pulse repetition rate and the feeding speed of the panel was set to 1 m/s, resulting in pulse durations of 5 ns.
Fig. 4: Scribing a complete line requires the complete removal of the material along the line by appropriate overlap of individual laser pulses with a certain pulse energy, repetition rate, and feeding speed
Determination of the threshold fluence
Scribing a complete line requires complete removal of the material along the line by appropriate overlap of individual laser pulses with a certain pulse energy, repetition rate, and feeding speed (see Fig. 4). As the spot overlap leads to an accumulation of the pulse energy per spot the ablation threshold decreases. Thus the ablation threshold shifts, since the total energy of multiple pulses needed for ablation is less than the ablation energy for a single pulse.
The incubation coefficient provides a characteristic value for the threshold behaviour of the material under such multi-pulse treatment. To determine the incubation coefficient the threshold fluences for a varying number of laser pulses per spot have to be determined. Technically it is not possible to control the precise number of laser pulses. Alternatively, the feeding speed can be varied in order to control the pulse-to-pulse interval and hence also the pulse overlap for a given repetition rate and a constant beam radius.
Determination of the incubation coefficient
The threshold fluence is the characteristic laser pulse energy density at which the material layer is either modified (for layer-side processing) or ablated (for glass-side processing). Determination of the value of the threshold fluence can be performed in different ways.
Rather straightforward is the visual determination. Single laser pulses of different energy are directed to the material until a visually apparent material damage or ablation occurs (see Fig. 2). The corresponding fluence can be considered as threshold fluence. This value can be either a melting threshold, which indicates slight thermal melting effects, or an ablation threshold, which corresponds to the complete removal of layer. Based on this method the international standard ISO 11 254 was developed. According to this standard the threshold is obtained from the onset value of the linear interpolation of the frequency of the observed effects versus the incident laser fluence.
More sophisticated methods of threshold fluence determination take into account the Gaussian profile and the diameter of the laser beam and relate the measured crater diameter to the laser pulse energy. By taking further into account a narrowing of the effective beam radius for glass-side laser treatment a good agreement of both methods is found.
Results for P1 laser structuring of TCO at 532 nm
Based on these considerations the successful removal of the TCO layer by laser treatment at a wavelength of 532 nm was achieved from the glass side as well as from the layer side by systematic variation of the ablation spot overlap and the laser beam defocusing for different wavelengths (cf. Fig. 4). Laser light absorption occurs via nonlinear effects. Glass-side treatment requires far less laser pulse energy and results in sharp edges and high-quality morphologies. Owing to the small spot size generated by the laser pulses, scribing a line requires high spot overlap of > 90 %. This high spot overlap allows for narrow scribing lines with a width of ~ 20 µm, which is sufficient for reliable isolation and assures low dead area losses. Thus, successful P1 scribing at 532 nm wavelength completes single wavelength laser scribing of Si thin-film modules.
Monolithic interconnection of thin-film PV modules can be achieved by laser structuring using a single wavelength provided that the threshold fluence and incubation coefficient are known. These two experimentally accessible parameters allow prediction and quantitation of the result of laser treatment, since they include all relevant information about the material behaviour and the laser parameters. This general approach can be extended to further solar cell thin-film materials and laser sources with different pulse durations and wavelengths. Optimal scribing parameters for each material can be found by systematic variation of the ablation spot overlap and the laser beam defocusing for different wavelengths. This method clearly provides further simplification and flexibility in thin-film solar cell production and thus contributes to cheaper thin-film solar cells.
Contact www.htw-berlin.de
Pictures: HTW Berlin