Daniel Abou-Ras

Characterization of Grain Boundaries in Cu(In,Ga)Se

This paper discusses the advantages of pulsed laser atom-probe tomography (APT) to analyze Cu(In,Ga)Se2-based solar cells. Electron backscatter diffraction (EBSD) was exploited for site-specific preparation of APT samples at selected Cu(In,Ga)Se2 grain boundaries. This approach is very helpful not only to determine the location of grain boundaries but also to classify them as well. We demonstrate that correlative transmission electron microscopy (TEM) analyses on atom-probe specimens enable the atom-probe datasets to be reconstructed with high accuracy. Moreover, EBSD and TEM can be very useful to obtain complementary information about the crystal structure in addition to the compositional analyses. The local chemical compositions at grain boundaries of a solar grade Cu(In,Ga)Se2 film are presented here. Na, K, and O impurities are found to be segregated at grain boundaries. These impurities most likely diffuse from the soda lime glass substrate into the absorber layer during cell fabrication and processing. Based on the experimental results, we propose that Na, K, and O play an important role in the electrical properties of grain boundaries in Cu(In,Ga)Se2 thin films for solar cells.

_{\bf 2}

Microstructural and chemical properties of the interfaces between Cu(In,Ga)Se2 (CIGS) and In2S3 layers in dependence on the In2S3 deposition temperature and Na concentration were investigated. The In2S3 layers were deposited by atomic layer deposition on CIGS layers at substrate temperatures ranging from 140°C to 240°C. Interfaces were investigated by means of scanning electron microscopy, bright-field and high-resolution transmission electron microscopy, electron diffraction, and energy-dispersive x-ray spectrometry. An orientation relationship between CIGS {112) and In2S3 {103) planes was found for the sample deposited at 210°C, whereas no orientation relationship was detected for the 240°C sample. Cu diffusion from CIGS into In2S3 was detected, as well as Cu depletion and In enrichment on the CIGS side of the interface. All three effects are enhanced with increasing deposition temperature. These results indicate the formation of a buried junction in the CIGS layer. In addition, a Na-free solar cell was...

Films Using Atom-Probe Tomography

In a recent publication by Abou-Ras et al., various techniques for the analysis of elemental distribution in thin films were compared, using the example of a 2-µm thick Cu(In,Ga)Se2 thin film applied as an absorber material in a solar cell. The authors of this work found that similar relative Ga distributions perpendicular to the substrate across the Cu(In,Ga)Se2 thin film were determined by 18 different techniques, applied on samples from the same identical deposition run. Their spatial and depth resolutions, their measuring speeds, their availabilities, as well as their detection limits were discussed. The present work adds two further techniques to this comparison: laser-induced breakdown spectroscopy and grazing-incidence X-ray fluorescence analysis.The present work shows results on elemental distribution analyses in Cu(In,Ga)Se2 thin films for solar cells performed by use of wavelength-dispersive and energy-dispersive X-ray spectrometry (EDX) in a scanning electron microscope, EDX in a transmission electron microscope, X-ray photoelectron, angle-dependent soft X-ray emission, secondary ion-mass (SIMS), time-of-flight SIMS, sputtered neutral mass, glow-discharge optical emission and glow-discharge mass, Auger electron, and Rutherford backscattering spectrometry, by use of scanning Auger electron microscopy, Raman depth profiling, and Raman mapping, as well as by use of elastic recoil detection analysis, grazing-incidence X-ray and electron backscatter diffraction, and grazing-incidence X-ray fluorescence analysis. The Cu(In,Ga)Se2 thin films used for the present comparison were produced during the same identical deposition run and exhibit thicknesses of about 2 μm. The analysis techniques were compared with respect to their spatial and depth resolutions, measuring speeds, availabilities, and detection limits.

Microstructural and chemical studies of interfaces between Cu(In,Ga)Se2 and In2S3 layers

CuInSe2, CuGaSe2, Cu(In,Ga)Se2 and CuInS2 thin-film solar absorbers in completed solar cells were studied in cross section by means of electron-backscatter diffraction. From the data acquired, grain-size distributions were extracted, and also the most frequent grain boundaries were determined. The grain-size distributions of all chalcopyrite-type thin films studied can be described well by lognormal distribution functions. The most frequent grain-boundary types in these thin films are 60°−〈221〉tet and 71°−〈110〉tet (near) Σ3 twin boundaries. These results can be related directly to the importance of {112}tet planes during the topotactical growth of chalcopyrite-type thin films. Based on energetic considerations, it is assumed that the most frequent twin boundaries exhibit a 180°−〈221〉tet constellation.

Comprehensive comparison of various techniques for the analysis of elemental distributions in thin films

Cu(In,Ga)Se2 (CIGS) thin-film solar cells with InxSy buffer layers deposited by physical vapor deposition yield efficiencies of up to 14.8%. For substrate temperatures during the InxSy deposition ranging from 23to200°C, air annealing of the completed solar cells leads to an improvement of the photovoltaic performance. However, at substrate temperatures of 300°C, the efficiencies are practically zero, and air annealing does not improve this value. To understand the effects of substrate temperature and air annealing on the CIGS/InxSy interfaces of the solar cells, these interfaces have been studied by means of bright-field and high-resolution transmission electron microscopy, selected-area electron diffraction (SAED), and energy-dispersive x-ray spectrometry (EDX). It is shown that air annealing leads to a substantial Cu depletion on the CIGS side of the CIGS/InxSy interface, probably inducing the formation of a compositionally graded interface between the buffer and CIGS. For the 300°C sample, CuIn5S8 form...

Grain-size distributions and grain boundaries of chalcopyrite-type thin films

The symmetry-dependence of electronic grain boundary (GB) properties in polycrystalline CuInSe2 thin films was investigated in a combined study applying scanning electron microscopy, electron backscatter diffraction, and Kelvin probe force microscopy. We find that highly symmetric Σ3 GBs have a higher probability to be charge neutral than lower symmetric non-Σ3 GBs. This symmetry-dependence can help to explain the large variations of electronic properties found for GBs in Cu(In,Ga)Se2.

Interfacial layer formations between Cu(In,Ga)Se2 and InxSy layers

The present contribution discusses buffer growth by chemical bath deposition (CBD) on polycrystalline Cu(In,Ga)Se2 (CIGS) films deposited by in-line co-evaporation with an integral [Ga]/([Ga]+[In]) ratio of 0.3. We report a correlation of the coverage of CBD Zn(O,S) and CdS films with the CIGS grain orientation as determined by electron backscatter diffraction. 〈221〉-oriented CIGS grains are sparsely covered with the CBD films, whereas on 〈100〉/〈001〉- and 〈110〉/〈201〉-oriented CIGS grains, we found very dense coverage of the CIGS surfaces. This result may be explained by lower energies of CIGS {112} surfaces compared with those of {100}/{001} and {110}/{102}.

Symmetry-dependence of electronic grain boundary properties in polycrystalline CuInSe2 thin films

The growth mechanism of Cu2ZnSnS4 thin films is studied starting from highly textured ZnS precursor films. These precursors were converted to Cu2ZnSnS4 by subsequent deposition of Cu, Sn, and S at high temperatures. Orientation measurements revealed that the ⟨111⟩ texture of the ZnS precursor is inherited by the Cu2ZnSnS4 layer. On the basis of texture and transmission electron microscopy measurements, a growth model is proposed. According to this model, the initial formation of Cu2ZnSnS4 nuclei is controlled by a topotactic or epitactic mechanism with respect to the ZnS precursor. The further growth of the Cu2ZnSnS4 grains appears to be independent of the precursor lattice.

Chemical bath deposition of Zn(O,S) and CdS buffers: Influence of Cu(In,Ga)Se2 grain orientation

Compound semiconductors belong to the most important materials for optoelectronic applications. Many of them exhibit favorable optical properties, such as a direct energy band gap (in contrast to silicon) and high-absorption coefficients over a wide spectral range. Moreover, varying the composition of the compound or substituting some of its elements often allows for controlled band gap engineering and optimization for specific applications. Because many compound semiconductors enable efficient conversion of light into electricity and vice versa, they are commonly used materials for optoelectronic devices.

Texture inheritance in thin-film growth of Cu2ZnSnS4

Thin-film solar cells based on Cu(In,Ga)Se2 (CIGSe) reach high power-conversion efficiencies in spite of large dislocation densities of up to 1010–1011 cm−2. The present work gives insight into the structural and compositional properties of dislocations in CIGSe thin films, which are embedded in a complete solar cell stack. These properties are related to the average electrical potential distributions obtained by means of inline electron holography. At a part of the dislocations studied, the average electrostatic potential shows local minima, all with depths of about −1.4 V. The measured average electrostatic potential distributions were modeled in order to reveal possible influences from strain fields, excess charge, and also compositional changes at the dislocation core. Cu depletion around the dislocation core, as evidenced by atom-probe tomography, explains best the measured potential wells. Their influences of the strain field around the dislocation core and of excess charge at the dislocation core are small. A structural model of dislocations in CIGSe thin films is provided which includes a Cu-depleted region around the dislocation core and gives a possible explanation for why decent photovoltaic performances are possible in the presence of rather large dislocation densities.