In view of the complexity of thin-film solar cells, which are comprised of a multitude of layers, interfaces, surfaces, elements, impurities, etc., it is crucial to characterize and understand the chemical and electronic structure of these components. Because of the high complexity of the Cu2ZnSn(S,Se)4 compound semiconductor absorber material alone, this is particularly true for kesterite-based devices. Hence, this paper reviews our recent progress in the characterization of Cu2ZnSnS4 (CZTS) thin films. It is demonstrated that a combination of different soft x-ray spectroscopies is an extraordinarily powerful method for illuminating the chemical and electronic material characteristics from many different perspectives, ultimately resulting in a comprehensive picture of these properties. The focus of the article will be on secondary impurity phases, electronic structure, native oxidation, and the CZTS surface composition.

In this work, we find that the pressure-induced phase transition of InP from III-V semiconductor phase having zincblende (ZB) crystal structure to metallic phase having rocksalt (RS) structure occurs at a pressure of 8.56 GPa accompanied by an 18% volume collapse. It is found that the nearest In and P atoms bonded as covalent bond. Crystal space of ZB is just occupied by In-P tetrahedrons partly with many interstices, but that of RS is fulfilled by close-packed octahedrons entirely. With pressures, broadened energy band of antibonding state and the reduced density of states (DOS) of bonding state cause the weakening of tetrahedral In-P covalent bonds. And then, ZB is destroyed and rebuilt to RS structure. Some In-5s, In-5p, P-3p and a few P-3s move to unoccupied high energy level, across Fermi level, and migrate from valence band to conduction band, and then generate metallic properties. Furthermore, changes of covalent bond would cause evident variation of elastic properties on the {100} and {110} planes.

In this article, single-phase CuIn(SxSe1-x)2 thin films with different x have been prepared using a new two-step method consisting of partly selenizing Cu–In precursors in Se ambient, and then exposing the partly selenized films to H2S/Ar under different conditions. The x-ray diffraction indicated that the CuIn(SxSe1-x)2 films exhibited the homogeneous chalcopyrite structure. With the increase of the sulfurization temperature from 425 to 525 °C, x increases from 0 to 1, Eg increases from 0.95 to 1.42 eV, the Raman shift of the A1[Se–Se] mode increases from 173 to 197 cm−1, and the resistivity increases from 101 to 103 Ω.cm.

Carbon spheres (CSs) with regular shapes (Ø 300–1200 nm) and numerous oxygen groups (–OH, C6H5–C=O, and C=O) were prepared by a simple glucose hydrothermal process. CSs were studied by x-ray powder diffraction, scanning electron microscopy, Fourier-transform infrared spectroscopy, x-ray photoelectron spectra, and elemental analysis. Their size was directly proportional to temperature and glucose concentration. Their shape could be controlled by temperature and reaction time. To prepare CSs (Ø 300–800 nm) with regular shapes and smooth surfaces, temperature and reaction time in the range of 180–230 °C and 3–4 h, respectively, were optimal. The chemical properties of the CSs were affected by temperature. A phase transformation from amorphous to turbostratic structure took place at T > 230 °C. The number of oxygen groups decreased as the temperature increased, and T ≤ 230 °C were optimal to prepare oxygen-rich CSs. Comparison of oxygen contents and O/C ratios indicated a further carbonization, and the degree was directly related to temperature. A possible formation mechanism for the CSs is proposed.

Under the COMPASS (condensed-phase optimized molecular potentials for atomistic simulation studies) force field, the molecular dynamics (MD) simulation was applied to first-to-third generation nanosize amine-based and butanediamine-based graphite/dendrimers composites. In this paper, we briefly introduced the constructive process of the composite system by means of MD simulation. The stability and mechanism of six intercalation composites were studied with microcosmic figure and variational energy under the invariable NVT ensemble. The energy variety was analyzed using the radial distribution function. The results indicate that the bulk of the dendrimer is small, the graphite layer is easy to bend and its systematic total energy is higher, which lead to the instability of the composite system. Therefore, the 3G dendrimer is the most stable system.

In this article, we investigated the effect of Sn grain structure on the electromigration (EM) reliability of Sn–2.5Ag (wt%) solder joints used in flip-chip packages. The electron backscattering diffraction technique was applied to characterize the Sn grain size and orientation of the solder joints. Failure analyses on Sn–2.5Ag solder joints after EM tests showed that the Sn grain structure was important in controlling the kinetics of the intermetallic compound growth and void formation under EM. Further microstructural analysis revealed that the grain sizes and orientations of the solder joints after multiple solder reflows were statistically different from those with a single solder reflow and resulted in an improved EM reliability. Thermal annealing effect was also investigated to separate the thermal effect from the EM-induced effect. Results obtained in this study demonstrated that EM reliability of Pb-free solder joints could be improved by optimization of the Sn grain structure.

Reflow behavior of a Sn–8.5Zn–0.5Ag–0.01Al–0.1Ga (five-element) solder on the Ni/Cu substrate was investigated under different heating rates. Reflowed samples show decreased Zn and increased AgZn3 in the solder with a reduction in the heating rate. The Zn at the solder/substrate interface was found to be much lower than that in the Sn–Zn solder systems. Cu was observed to be diffused through the electroplated Ni layer and noticed only with the Ag–Zn compound in the solder. Ga was spotted at the interface in the Ag–Zn matrix, whereas Al was detected with the Zn at the interface. Small intermetallic compound (IMC) layer was formed at the interface; however, its amount enhanced with the reduction in the heating rate. Present study relates the reflow behavior of the five-element solder with the reactivity of different elements in the system and its influence on the formation of IMCs in the solder and at the solder/substrate interface.

A bismuth telluride alloy-based thermoelectric generator with high-aspect ratio, free-standing legs was fabricated. Such legs are desirable for efficient generator performance from low-grade heat sources but are difficult to assemble because they are fragile and difficult to handle and position. Plunge and wire electro-discharge machining (EDM) were used to produce 150 μm × 300 μm legs, approximately 6-mm long, with high fidelity. Removal of recast material from EDM was necessary for good adhesion of metallization, but sputter etching was found to deteriorate the mechanical strength of the contacts. A wet chemical cleaning process was developed instead that resulted in good adhesion under test conditions. Au was preferred for designs where interconnects could be patterned directly on the module. Module figure of merit (ZT) was 0.72, close to the 0.85 value expected from bulk material property measurements. Impedance spectroscopy and the Harman technique were shown to significantly underestimate module ZT in the present test configuration. Shear and fatigue testing were performed on arrays of high-aspect ratio legs. Legs survived over 104 cycles of shear loading at 90% of the load to failure.

The thermoelectric parameters have been investigated depending on the thickness of the layer of nanostructures IV–VI (PbS, PbSe, PbTe, and SnTe). Based on the theoretical model of quantum well (QW) with infinitely high walls, it is demonstrated that this model explains nonmonotonous behavior of the Seebeck coefficient S with the change of the well width. On the basis of oscillation period Δdexp, we have approached the theoretical d-dependence of the coefficient S to the experimental one and defined the value of the Fermi energy in the corresponding nanostructures. It has been established that the minimum QW width dmin, where the first energy level coincides with the Fermi energy, is equal to the oscillation period of the Seebeck coefficient in this structure.

Atomic structures of the Zr48Cu45Al7 as-prepared and annealed metallic glasses (MGs) were investigated by performing the reverse Monte Carlo simulation on the synchrotron radiation-based experiments. It was found that although the annealed sample remains completely amorphous, the volumes of the Al-centered clusters evidently expand, which is attributed to the relatively longer Al–Zr bonds. As a result, the role of Al atoms as the glue atoms to connect and fix the Zr- and Cu-centered large clusters is accordingly weakened, which leads to the ease of the rearrangement of atoms and clusters in the glass state. This study provides an insight into the microstructures of MGs, which extends understanding of the structural evolution in the glass alloys during annealing prior to the precipitation of nanocrystals.

Ni-based under-bump metallization (UBM) has attracted wide attention due to its low reaction rate with Sn, compared with Cu and Cu alloy. In this study, the interfacial reactions between eutectic Sn–3.5Ag solder and Ni-based UBM, including electroplated Ni (EP-Ni) and electroless Ni (EL-Ni) are investigated. Morphology and growth kinetics of Ni3Sn4 intermetallic compounds are studied at different reflow temperatures and durations. The growth rate and the growth activation energy of Ni3Sn4 were measured for the two sets of samples. The activation energies are measured to be 25 kJ/mol and 38 kJ/mol for the Ni3Sn4 growth on EP-Ni and EL-Ni, respectively. The Ni3Sn4 on EP-Ni UBMs shows a slower growth rate and the Ni3Sn4/solder interface is void free even after 20-min reflow at 240 °C. On the other hand, the interface of Ni3Sn4/EL-Ni has a lot of microvoids after reflowing at 240 °C for 20 min.

The oxidation behavior of Cu50Zr50 and Cu46Zr46Al8 glasses during continuous heating up to 1073 K has been investigated, with special emphasis on the oxidation resistance in the supercooled liquid (SCL) state. For Cu50Zr50, the oxide layer mostly consists of monoclinic ZrO2 (m-ZrO2), while for Cu46Zr46Al8, the oxide layer consists of two different layers: an outer layer consisting of tetragonal ZrO2 (t-ZrO2) + Al2O3 + metallic Cu (oxidation product from the SCL state of the glass matrix) and inner layer comprised of m-ZrO2 + metallic Cu islands (oxidation product from the crystallized matrix). Cu-enriched regions consisting of Cu51Zr14 (in Cu50Zr50) or AlCu2Zr + Cu70Zr15Al15 + Cu51Zr14 (in Cu46Zr46Al8) are present below the oxide layer. The present study shows that the addition of Al (8 at.%) in Cu50Zr50 results in a significant deterioration of the oxidation resistance in the SCL state since the solutionizing of Al in t-ZrO2 leads to a higher oxygen ion vacancy concentration, thus providing a higher activity of oxygen ions.

Bulk metallic glasses (BMGs) are a promising class of engineering materials, but they can be difficult to machine due to high hardness and a metastable structure. Crystallization due to machining can have negative effects, such as a decreased load-bearing capacity of fabricated parts, and thus should be avoided. Here, a Zr-based BMG was machined using abrasive waterjet (AWJ), electrical discharge, ns-pulsed laser engraving, and conventional dry-milling techniques. Characterization of the processed material indicated that AWJ preserves the amorphous phase and provides the combination of speed and flexibility required to rapidly fabricate small three-dimensional parts, while the other techniques did not achieve these goals. As proof-of-principle, a screw, similar to an orthopedic implant, was rapidly machined from the BMG using AWJ.

Silica-titania sub-close-packed single layers were deposited by spin coating titanium alkoxide sols containing inert 0.5 micrometer silica particles to study the process of reactive sintering more closely than has been done before. The sub-close-packed single layers were designed to achieve a coating density such that pairs or chains of silica particles were placed on the flat substrate and held together by the reactive titania thin films overlaid on the surface and in the neck regions of these essentially 2-D particle networks. Because of the low density of silica particles, all of the two-particle junctions and neck regions were aligned for geometrically direct viewing; as a result scanning electron microscopy was useful for observing the morphology of these neck regions. Image analysis was used to quantify the neck diameter for varying titania/silica precursor concentration ratios. Geometrical calculations that relate the change in neck volume to the neck radius are presented. Implications for design of reactive sintering systems are discussed.

A single crystal aluminum nitride (AlN) wafer surface was investigated via the use of a novel software-based, Charge-based Deep Level Transient Spectroscopy (Q-DLTS) apparatus, both before and after surface bond termination with hydrogen plasma. The sample was cleaned and metalized with a thermoresistive evaporator to create electrical contacts and then annealed in a helium atmosphere at 825 °C. Current-voltage (I-V) measurements were performed to investigate the nature of the metal/substrate contacts. The effect of hydrogen termination was investigated and Arrhenius plots were produced from Q-DLTS spectra at temperatures ranging from −15.9 °C to 136.0 °C. Activation energies and capture cross-section values were calculated from the Q-DLTS spectra for traps existing in the AlN substrate surface. Prior to hydrogen termination, four charge traps were observed with activation energies of 0.31 eV, 0.61 eV, 0.56 eV, and 0.18 eV and capture cross sections 5.6 × 10−21 cm2, 1.1 × 10−16 cm2, 3.5 × 10−19 cm2, and 1.3 × 10−21 cm2, respectively After hydrogen termination, five charge traps were observed with activation energies of 0.31 eV, 0.61 eV, 0.52 eV, 0.19 eV, and 0.40 eV, and capture cross sections 4.9 × 10−21 cm2, 1.3 × 10−16 cm2, 2.9 × 10−19 cm2, 3.1 × 10−19 cm2, and 4.7 × 10−19 cm2, respectively. Four of these peaks after termination are matched with the peaks prior to termination and the fifth peak appears to be the result of the hydrogen termination.