We study the sequential fabrication of Cu(In,Ga)Se-2 (CIGSe) absorber layers by using an atmospheric pressure selenization with a process duration of only a few minutes and the utilization of elemental selenium vapor from independent Se sources. This technology could proof to be an industrially relevant technology for the fabrication of thin-film solar cells. Controlling the amount of Se provided during the selenization of metal precursors is shown to be an effective measure to adjust the Ga in-depth distribution. A reduced Se supply for CIGSe formation leads to a more homogeneous Ga distribution within the absorber. The underlying growth dynamics is investigated by interrupting the selenization at different times. At first, CIGSe formation occurs in accordance with previously suggested growth paths and Ga segregates at the Mo back contact. Between 520 and 580 degrees C, the growth dynamics differs distinctly, and In and Ga distribute far more uniformly within the absorber depth. We also studied the impact of the precursor architecture. The best performing precursor in terms of efficiency of the respective solar cells was a multilayer with 22 In/CuGa/In triple layers. Simple bilayers stacks lead to films of higher roughness and correlated shunting. By optimizing the precursor architecture and the Ga in-depth distribution in the CIGSe layer, a conversion efficiency of up to 15.5% (active area) could be achieved. To our knowledge, this is the highest reported efficiency for sulfur free CIGSe-based solar cells utilizing fast (few minutes) atmospheric processes and elemental Se vapor. Copyright (C) 2017 John Wiley & Sons, Ltd.

A remarkable discrepancy between the optically active bandgap of Cu(In,Ga)(S,Se)(2) absorber layers for thin film solar cells and the minimum bandgap as determined via elemental depth profiling has been observed in this study. This behavior occurs in absorbers sequentially grown by sulfurization after selenization of metal precursors and is demonstrated and explored in the following, using glow discharge optical emission spectroscopy, external quantum efficiency and Raman scattering. Furthermore this mismatch is explained by investigating the microscopic elemental distributions using transmission electron microscopy. It turns out, that sulfur is-on a microscopic scale-inhomogeneously distributed in the bulk of the absorber and solely present in areas near the absorber surface itself and at inner surfaces, e.g. in voids in the bulk. (C) 2017 Elsevier B.V. All rights reserved.

Sequential Cu(In,Ga)Se2 fabrication with a thermally activated reaction of Cu-In-Ga metal precursor layers in chalcogen atmosphere is an industrially attractive route for preparation of Cu(In,Ga)Se2 absorber based thin film solar cells. Recent results show that controlling the selenium supply during rapid thermal processing has a huge impact on absorber growth. Especially a two stage process applying a first annealing step with or without Se at temperatures up to 400°C was shown to have a positive effect on the elemental in-depth distribution. However, during this annealing, lateral phase separation, dewetting and coarsening may occur in the metal phase, leading to lateral non-uniformity of the absorber. In this study we show how the dewetting can be strongly decreased by adjusting the precursor architecture, applying faster heating rates and NaF addition on top of a precursor. In contrast, NaF deposited underneath the precursor increases the dewetting effect. Further we show that lateral phase separation during annealing increases with temperature and leads to phase domain sizes of several micrometers at 580°C.

Many applications in biology, medicine, and material science have a demand for a time and spatially resolved detection of luminescence, particularly in the near infrared spectral region. For example, observation of singlet oxygen phosphorescence is important in understanding cell-cell interactions as well as for research involving photo-dynamic therapy. Furthermore, investigating the temporal behavior of semiconductor luminescence has become an important quality indicator for the quality of wafer material, especially many kinds of solar cell research and production.

Laser-based patterning for serial interconnection of chalcopyrite (i.e., Cu(In ${}_{\rm{x}},\text{Ga}_{\rm{1-x}}$)Se2 or CIGSe) solar cells was obtained by 1) laser ablation using picosecond (ps) pulses, 2) local phase transformation using nanosecond (ns) laser pulses, and 3) conventional needle-based patterning. All three patterning approaches cause a modification of the material properties in the vicinity of the actual P2 scribing lines, which affects and limits the electrical functionality of the interconnection, and thus has to be considered for positioning the P3 scribe. Thus, the extension and the properties of the affected zone aside the P2 scribe was investigated through spectral and spatial photoluminescence (PL). From the depletion of the PL intensity when approaching the scribing line and a peak shift analysis it is concluded that the laser-affected zone is distinctively larger than visual inspections suggest. Even putatively ultrashort, nonthermal ps pulses cause material modifications which might be facilitating recombination losses and thus limiting solar cell efficiencies. For ps laser patterning the affected area is even larger than for the ns laser patterning, due to a modification of the band structure and to thermal decomposition. Evolving subpeaks at the low energy tails are found to originate from Cu-related flat defect levels, i.e., Cu vacancies ($V_{{\rm{Cu}}}$) and antisites (Cu In), created upon laser impact. These findings provide insights into laser-based material modification and provide beneficial information for minimizing the dead area resulting from laser-based monolithic interconnection.

The change of electrical conductivity in chalcopyrite (i.e., Cu(Inx, Ga1−x)Se2 or CIGSe) solar cells induced by nanosecond laser pulses is investigated as a function of the elemental composition and its spatial distribution. The underlying laser induced phase transformation process, which results in a decomposition of the CIGSe semiconductor and a modification of its elemental composition, is utilized to form the monolithic series interconnection between front and back contact in CIGSe based thin film solar cells. The results show a dependence of the composition of the CIGSe layer and the resulting series resistance on the applied laser fluence. Lower series resistance is primarily related to an enhanced fraction of copper, gallium and zinc in the laser transformed zone resulting from selective vaporization of absorber elements. For intermediate laser fluences (~0.36J/cm2) a patterning process is established that allows reliable and high-quality series interconnection. Both, lower and higher laser fluences result in high series resistances due to incomplete phase transformation or damages of the back contact, respectively.

We report a route to deposit In2S3 thin films from air-stable, low-cost molecular precursor inks for Cd-free buffer layers in chalcopyrite-based thin film solar cells. Different precursor compositions and processing conditions were studied to define a reproducible and robust process. By adjusting the ink properties, this method can be applied in different printing and coating techniques. Here we report on two techniques, namely spin-coating and inkjet printing. Active area efficiencies of 12.8% and 12.2% have been achieved for In2S3-buffered solar cells respectively, matching the performance of CdS-buffered cells prepared with the same batch of absorbers.

Biology, medicine, and material science, especially many kinds of solar cell research and production demand for time- and spatially resolved detection in the near infrared spectral region. Observation of singlet oxygen phosphorescence is, for example, important for understanding cell-cell interactions and in the field of photo-dynamic therapy. Whereas the temporal behavior of the luminescence from semiconductors is an agreed indicator for the quality of wafer material. Single photon counting based data acquisition has proven to yield the best sensitivity and a very high dynamic range - it is therefore the ideal method for measuring weak luminescence in the near infrared spectral region. Based on this principle, we present the methodology for state-of-the-art luminescence measurements. Our hardware is designed to detect spectrally- and spatially-resolved weak emission in steady-state and time-resolved manner. Depending on the samples size and aggregation state, large excitation volume spectrometer or diffraction limited small excitation spot microscope systems are employed. A combination of both types of set-ups allows us to record even spectrally-resolved images. The high sensitivity of these systems was proven by detecting singlet oxygen phosphorescence from aqueous samples of photo-sensitizers. With the reached detection sensitivity level it is feasible to record even singlet oxygen phosphorescence images.

Electrical and structural functionality of monolithic interconnection of CIGSe solar cells by P1-P3 picosecond laser ablation has been successfully established and evaluated as being competitive to conventional needle scribing. In all three patterning steps the material is selectively and completely removed yielding structurally well-defined trenches at high scribing speeds up to 1 m/s. P1 and P2 ps laser scribing clearly improves the solar cell efficiencies whereas P3 scribing is still challenging due to shunts resulting from laser-induced alteration of the absorber material.

In this work we investigate the degradation behavior of solar cells with a copper front side metallization based on a fine-line screen-printed silver seed-layer, a plated nickel diffusion barrier, a plated copper conductive layer and a silver capping. The results reveal that the cell degradation depends, besides the effectiveness of the nickel diffusion barrier, on the used seed-layer and the firing temperature of the seed-layer. The degradation behavior of cells, produced with different Ag seed-layer pastes resulting in different contact finger geometries after screen-printing and firing, was evaluated. With similar nickel diffusion barrier masses the seed-layer that generates the biggest metallized area at the cell surface shows the fastest degradation. Apart from that, the composition of the seed layer also has an impact on the degradation. The analysis of different firing temperatures shows that higher set-peak temperatures result in faster cell degradation due to copper migration.

The thermal impact of nanosecond laser pulses was beneficially employed and well-controlled for the preparation of the P2 interconnect by local phase transformation (i.e., by drawing conductive lines rather than removing the material) in CIGSe mini-modules, which were demonstrated to outperform their conventionally needle-patterned counterparts. Conductivity and elemental composition of the scribed lines as well as the extent of the heat-affected area were analyzed, quantified and taken into account for achieving optimal CIGSe solar module performances. This approach opens new prospects for significant simplification of the serial interconnection, since the P2 und the P3 can be scribed simultaneously after deposition of both the CIGSe and the TCO layer.

Detection sensitivity from the ultraviolet to the near infrared spectral region is a key parameter to meet today's demand for handling smallest analyte amounts and short measurement times in the optical evaluation of miscellaneous samples. The introduction of single photon counting based data acquisition has proven to yield a major sensitivity increase and very high dynamic range - it is the ideal method for measuring weak luminescence. We present the hardware and handling optimization of a state of the art spectrometer for steady-state and time-resolved fluorescence measurements. The high sensitivity of the spectrometer was shown by measurements of popular fluorescent dyes as well as the Raman spectrum of water under well defined and reproducible conditions. The achieved sensitivity allows us to quantify singlet oxygen generation and to characterize the singlet oxygen phosphorescence decay, a prerequisite when studying photosensitisers like porphyrins and phthalocyanines used for example in photodynamic therapy (PDT). Moreover, with the help of an integrating sphere fluorescence quantum yields of low fluorescent samples like Ru(bpy)3 in water can be determined very precisely. The fibre connection of the spectrometer to a time-resolved fluorescence microscope (MicroTime100/200) was also realized. The combination of the advantages of both setups makes it e.g. possible to perform 2D-lifetime imaging with a freely tunable detection window for low luminescent samples even far into the near infrared region. The measurements with such a combination give not only the spectral and lifetime information of a luminescent sample but also the spatial information which is especially important for hetergeneous samples.

In this work the long term stability of silicon solar cells with a copper front side metallization based on a fine-line screen-printed silver seed-layer, a plated nickel diffusion barrier, a plated copper conductive layer and a silver capping is investigated in detail. Silicon nitride layers deposited by PECVD or sputtering effectively hinder copper diffusion, which might occur e.g. at contact edges. To investigate the plated nickel diffusion barrier, fast degradation of full size 156×156 mm² cells at elevated temperatures on hotplates and module degradation in the climate chamber executing a damp heat test were performed. Plated nickel thickness influences the degradation speed on cell level. On module level, only the cell without diffusion barrier shows a degradation of 1.2% rel. in efficiency (IEC criteria passed) after 1500 h damp heat test (85 °C, 85% r.h.). The cells with diffusion barrier do not show any degradation. Comparison of cell and module results indicate that fast degradation on hotplates at cell level gives a reasonable first estimate regarding cell degradation due to copper diffusion.

Cu(In, Ga)Se 2 solar cell absorbers are prepared on a Mo-coated glass substrate by using a sequential process consisting of a sputter deposition of an In/CuGa/In metal precursor, subsequent PVD deposition of a Se layer and annealing in N 2 atmosphere. The Se concentration in the final layer stack was found to be relatively up to 20 % higher than expected. The excess Se is bound in a MoSe 2 layer with laterally varying thickness, between the absorber and the Mo back contact. Such a layer can lead to an increase in the series resistance of the completed solar cells. By pumping at a specific time, we were able to reduce the Se partial pressure selectively during the selenization. For a constant annealing time, we find that the MoSe 2 thickness increases with the time in which a high Se partial pressure is maintained, i.e., the time before the selective pumping. A significant reduction of the Se partial pressure after half the annealing time led to solar cells with the smallest series resistance and overall best conversion efficiency. We further found that the addition of NaF before the annealing led to comparatively thin MoSe 2 layers. This suggests that the Na incorporation from the glass substrate in our process is too small to hinder the MoSe 2 growth. A more specific control of the Na supply is required in our process to manipulate the MoSe 2 growth and the doping density in the absorber.

In this work we present results of different strategies of how CIGSe solar cells can be laser scribed focusing on P2 and P3. Therefore a laser source with a wavelength of 532 nm and pulse duration of 13 ns, as well as a laser, with pulse durations of 10 ps for both wavelengths of 532 nm and 1064 nm are used. The ablation mechanisms and the results ablation process (behaviors) due to different wavelengths and pulse durations are studied. Different parameters like process speed, scribe quality and solar cell performance are discussed. Successful laser patterning of all layers is demonstrated on a minimodule reaching 10% efficiency which is slightly better than the needle scribed reference.

Distinguishing the scattering contributions of isoelectronic atomic species by means of conventional x-ray- and/or electron diffraction techniques is a difficult task. Such a problem occurs when determining the crystal structure of compounds containing different types of atoms with equal number of electrons. We propose a new structural model of Cu(InxGa1−x)3Se5 which is valid for the entire compositional range of the CuIn3Se5–CuGa3Se5 solid solution. Our model is based on neutron and anomalous x-ray diffraction experiments. These complementary techniques allow the separation of scattering contributions of the isoelectronic species Cu+ and Ga3+, contributing nearly identically in monoenergetic x-ray diffraction experiments. We have found that CuIII3Se5 (III=In,Ga) in its room temperature near-equilibrium modification exhibits a modified stannite structure (space group I4¯2m). Different occupation factors of the species involved, Cu+, In3+, Ga3+, and vacancies have been found at three different cationic positions of the structure (Wyckoff sites 2a, 2b, and 4d) depending on the composition of the compound. Significantly, Cu+ does not occupy the 2b site for the In-free compound, but does for the In-containing case. Structural parameters, including lattice constants, tetragonal distortions, and occupation factors are given for samples covering the entire range of the CuIn3Se5–CuGa3Se5 solid solution. At the light of the result, the denotation of Cu-poor 1:3:5 compounds as chalcopyrite-related materials is only valid in reference to their composition.

Polycrystalline samples of CuGaSe2-related defect compounds (DC) have been prepared by Chemical Close-Spaced Vapour Transport (CCSVT) (thin films) and elemental synthesis (powder) respectively. In the latter case a homogenisation step was introduced during the preparation, including mechanical intermixing and adjacent-heat treatment after the main reaction. Following this route we assured conditions "close to" thermodynamic equilibrium. The influence of the annealing temperature on the lattice parameters a(0) and c(0), as well as the cation distribution, was investigated. By means of X-ray and neutron diffraction analysis the structural properties of the Cu-Ga-Se based DCs have been determined and a new structural model has been derived. Finally the structural parameters of CCSVT-grown thin-film material were correlated with findings for bulk material samples which were intentionally prepared off-equilibrium. (C) 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim