Search

We investigate magnetron-sputtered In2(OxS1−x)3 compounds acting as an alternative buffer system to the solution-grown CdS or Zn(O,S) buffer layers in Cu(In,Ga)Se2 (CIGS) thin-film solar cells. The influence of the oxygen content on the solar cell performance, microstructure of the mixed systems, bandgap, and band offsets to CIGS is investigated experimentally and also characterized by calculations based on density functional theory. Samples in a series with different chemical compositions ranging from In2S3 to In2O3 are either directly deposited from ceramic targets or from a pure In2S3 target by reactive sputtering by adding O2 in the Ar sputtering gas. The binary compounds In2S3 and In2O3 sputtered at 220 °C substrate temperature from ceramic targets exhibit a crystalline structure, whereas the ternary In2(O,S)3 compounds are either nanocrystalline in the case of In2(O0.25S0.75)3 or amorphous for In2(O0.5S0.5)3 and In2(O0.75S0.25)3. For [O]/([O] + [S]) ratios above 0.25, the cell efficiencies decrease drastically, mainly due to lower open-circuit voltages (VOC). This behavior can be explained by an increase of the negative conduction band offset between the CIGS absorber and the oxygen-rich In2(OxS1−x)3 or In2O3 buffer, resulting in pronounced VOC losses. Adding oxygen to In2S3 with optical bandgap energies of around 2 eV results in a bowing of the values to below 2 eV and finally reaching values of around 2.7 eV for In2O3 if an indirect band transition is assumed. In summary, our results reveal that pronounced oxygen incorporation in In2S3 is not beneficial in terms of CIGS device efficiency because oxygen is electronically inactive and poorly miscible. [ABSTRACT FROM AUTHOR]

Structural health monitoring (SHM) is the process of detecting damage in structures. Typically, the damage-sensitive feature identification is based on fitting some models to the measured system response data. The parameters of these models or the prediction errors associated with these models then become the desired damage-sensitive features. Real-world structures usually exhibit significant nonlinear behaviors. Data-driven nonlinear system identification (NSI) is essential to obtain accurate models of structural dynamics for reliable damage detection. However, on the one hand, the data acquisition of real-world structures for damage detection is usually difficult or scarce. On the other hand, NSI generally needs to solve a nonlinear optimization problem that requires more computational time for convergence than linear system identification. Data efficiency and computational efficiency become two critical challenges in data-driven NSI. To address these challenges, this work presents a proof-of-principle study on developing a meta learning-based method for efficient data-driven NSI. Specifically, the meta learning algorithm leverages a previously collected database of similar but different systems to learn the meta-knowledge about how to identify a new system, enabling a fast identification/modeling with limited data. Numerical simulations are performed to validate the method on fundamental nonlinear dynamical systems widely seen in civil engineering. It is observed that the presented meta learning-based method outperforms the conventional method in terms of both data and computational efficiency. Limitations of this work are further discussed. [ABSTRACT FROM AUTHOR]

The design of a Cd‐free and wider‐bandgap buffer layer is stringent for future Cu(In,Ga)Se2 (CIGSe) thin‐film solar cell applications. For that, an In2S3 buffer layer alloyed with a limited amount of O (well below 25 mol%) has been proposed as a pertinent alternative solution to CdS or Zn(O,S) buffers. However, the chemical stability of the In2S3/CIGSe heterointerface when O is added is not completely clear. Therefore, in this work, the buffer/absorber interface for a series of sputter‐deposited In2S3 buffers with and without O is investigated. It is found that the solar cell with the highest open‐circuit voltage is obtained for the O‐free In2S3 buffer sputtered at 220 °C. This improved open‐circuit voltage could be explained by the presence of a 20 nm‐thick ordered vacancy compound (OVC) at the absorber surface. A much thinner OVC layer (5 nm) or even the absence of this layer is found for the cell with In2(O0.25S0.75)3 buffer layer where O is inserted. The volume fraction of the OVC layer is directly linked with the magnitude of Cu diffusion from the CIGSe surface into the In2(OxS1−x)3 buffer layer. The O addition strongly reduces the Cu diffusion inside the buffer layer up to complete suppression for very high O contents in the buffer. Finally, it is discussed that the presence of the OVC layer may lower the valence band maximum, thereby forming a hole barrier, suppressing charge carrier recombination at the In2(OxS1−x)3/CIGSe interface, which could result in an increased open‐circuit voltage. [ABSTRACT FROM AUTHOR]

Soft magnetic materials are a core element of power electronics and electrical machines. However, none of the soft magnetic materials is able to exploit the full potential of wide bandgap semiconductors, which operate above MHz frequency for efficient energy conversion in power electronic systems. Here, a high‐performance Fe–Si soft magnetic composites with a 2D ordered domain structure are reported, enabling efficient energy conversion at MHz frequencies. By transforming spherical particles into flakes and arranging them in layers, 2D magnetic domains are created within the composite. This leads to a 90% increase in permeability and a tenfold decrease in loss at 3 MHz compared to composite made with spherical particles. The significantly increased cut‐off frequency indicates that the ordered flaky particles are suitable for MHz applications, unlike the disordered spherical particles. These findings provide an effective approach for fabricating high performance soft magnetic composites from traditional spherical particles and miniaturizing magnetic devices for efficient power electronics operation. [ABSTRACT FROM AUTHOR]

We propose a spatiotemporal coupling deep neural network approach for time-resolved reconstruction of the velocity field around a circular cylinder. The neural network leverages two distinct data types: (1) non-time-resolved velocity field around the cylinder, consisting of fixed frequency sampling and variable frequency sampling velocity field, and (2) the time-resolved surface pressure sequence around the cylinder. The deep neural network comprises two sub-networks: a convolutional autoencoder (CAE) for nonlinear mode extraction and a Transformer for sequence-to-sequence learning. We refer to this architecture as CTNet (CAE-Transformer Network). The encoder in the CAE maps non-time-resolved velocity field to a latent vector, enabling the extraction of nonlinear modal coefficients. An appropriate time window length for the surface pressure sequence is then selected to establish a Transformer sequence learning model, using the chosen sequence as input to predict the corresponding nonlinear modal coefficients. Once the Transformer is well trained, the time-resolved nonlinear modal coefficients of velocity field can be achieved. Along with the well-trained decoder in the CAE, the time-resolved velocity field can be reconstructed from the output of the Transformer. We verify the performance of CTNet by a simulated dataset at a representative Reynolds number of 3900. The results show a relative reconstruction error of just 6.3% for the time-resolved velocity field, demonstrating high reliability in the reconstruction. We further compare the reconstructed velocity field obtained with and without the utilization of variable frequency sampling velocity field. Notably, the inclusion of variable frequency sampling velocity field significantly improves the reconstruction quality. [ABSTRACT FROM AUTHOR]

Thin-film solar cells with Cu(In,Ga)Se2 (CIGS) absorber layers have been intensively studied due to their high power conversion efficiencies. CIGS solar cells with Zn(O,S) buffer layers achieved record efficiencies due to their reduced parasitic absorption compared with the more commonly used CdS buffer. Accordingly, we have studied solution-grown Zn(O,S) buffer layers on polycrystalline CIGS absorber layers by complementary techniques. A bandgap energy Eg of 2.9 eV is detected by means of angle-resolved electroreflectance spectroscopy corresponding to Zn(O,S), whereas an additional Eg of 2.3 eV clearly appeared for a post-annealed CIGS solar cell (250 °C in air) compared with the as-grown state. To identify the chemical phase that contributes to the Eg of 2.3 eV, the microstructure and microchemistry of the Zn(O,S) buffer layers in the as-grown state and after annealing were analyzed by different transmission electron microscopic techniques on the submicrometer scale and energy-dispersive x-ray spectroscopy. We demonstrate that the combination of these methods facilitates a comprehensive analysis of the complex phase constitution of nanoscaled buffer layers. The results show that after annealing, the Cu concentration in Zn(O,S) is increased. This observation indicates the existence of an additional Cu-containing phase with Eg close to 2.3 eV, such as Cu2Se (2.23 eV) or CuS (2.36 eV), which could be one possible origin of the low power conversion efficiency and low fill factor of the solar cell under investigation. [ABSTRACT FROM AUTHOR]

The flow around a finite square cylinder with suction control at the side and a free-end leading edge is investigated through direct numerical simulations at a Reynolds number of 250. The absolute value of the ratio (Γ) between the suction velocity and the free-stream velocity is in the range of 0 < Γ ≤ 2. The results show that suction reduces the drag and fluctuating lift on the square cylinder. The optimal control effectiveness for reducing the fluctuating lift coefficient C l ′ and the average drag coefficient C d ¯ is achieved at Γ = 0.375 and 0.75, respectively ( C l ′ reduced by over 70% and C d ¯ reduced by nearly 20%). This is superior to the control effect achieved by active suction control only at the side leading edge. Compared to suction applied only at the side leading edge, adding suction at the free-end leading edge suppresses the flow separation on the top surface of the square cylinder. Moreover, with increasing suction ratios, the tornado-like Tip Vortex scale at the free-end of the square cylinder decreases, and the root mean square of streamwise velocity fluctuation at various spanwise planes decreases. Additionally, a data-driven balanced model-based dominant flow mode identification method is adopted to identify the dominant modes of the flow field at the z / d = 0 plane at different suction ratios. The results show that suction can suppress the influence of the square cylinder on the far wake, and as the suction ratio increases, the area of the free flow expands. [ABSTRACT FROM AUTHOR]

The flows past a twisted cylinder are studied at low Reynolds numbers using three-dimensional (3D) direct numerical simulations. For comparison, the flows over uniform circular and elliptic cylinders are also investigated. The similar transformations from the two-dimensional (2D) laminar regime to the 3D wake transition regime in the wake of a uniform cylinder are not observed for a twisted cylinder. The wakes of a twisted cylinder inherently display 3D wake patterns owing to the effects of geometrical perturbations. The 3D wakes of the twisted cylinder are still in the stable laminar regime at low Reynolds numbers, leading to continuous variations of various flow quantities with increasing Reynolds numbers. The transformations of the shedding patterns result in discontinuous changes in these flow quantities, such as aerodynamic forces, shedding frequencies, and vortex formation lengths. In addition, some novel characteristics appear in the 3D vortex structures of a twisted cylinder. The spanwise vortex wavelength presents a lock state with the structural wavelength. The shedding vortices are oblique to the cylinder axis in the near wake, and the dislocation occurs among different segments of vortices along the spanwise direction. Moreover, the effects of these vortex features on other flowing quantities are investigated. [ABSTRACT FROM AUTHOR]

The hexagonal ferrite h-YbFeO3 grown on YSZ(111) by pulsed laser deposition is foreseen as a promising single multiferroic candidate where ferroelectricity and antiferromagnetism coexist for future applications at low temperatures. We studied in detail the microstructure as well as the temperature dependence of the magnetic properties of the devices by comparing the heterostructures grown directly on YSZ(111) (i.e., YbPt_Th0nm) with h-YbFeO3 films deposited on substrates buffered with platinum Pt/YSZ(111) and in dependence on the Pt underlayer film thickness (i.e., YbPt_Th10nm, YbPt_Th40nm, YbPt_Th55nm, and YbPt_Th70nm). The goal was to deeply understand the importance of the crystal quality and morphology of the Pt underlayer for the h-YbFeO3 layer crystal quality, surface morphology, and the resulting physical properties. We demonstrate the relevance of homogeneity, continuity, and hillock formation of the Pt layer for the h-YbFeO3 microstructure in terms of crystal structure, mosaicity, grain boundaries, and defect distribution. The findings of transmission electron microscopy and X-ray diffraction reciprocal space mapping characterization enable us to conclude that an optimum film thickness for the Pt bottom electrode is ThPt = 70 nm, which improves the crystal quality of h-YbFeO3 films grown on Pt-buffered YSZ(111) in comparison with h-YbFeO3 films grown on YSZ(111) (i.e., YbPt_Th0nm). The latter shows a disturbance in the crystal structure, in the up-and-down atomic arrangement of the ferroelectric domains, as well as in the Yb–Fe exchange interactions. Therefore, an enhancement in the remanent and in the total magnetization was obtained at low temperatures below 50 K for h-YbFeO3 films deposited on Pt-buffered substrates Pt/YSZ(111) when the Pt underlayer reached ThPt = 70 nm. [ABSTRACT FROM AUTHOR]