Analytics-statistics mixed training and its fitness to semisupervised manufacturing.

While there have been many studies using machine learning (ML) algorithms to predict process outcomes and device performance in semiconductor manufacturing, the extensively developed technology computer-aided design (TCAD) physical models should play a more significant role in conjunction with ML. While TCAD models have been effective in predicting the trends of experiments, a machine learning statistical model is more capable of predicting the anomalous effects that can be dependent on the chambers, machines, fabrication environment, and specific layouts. In this paper, we use an analytics-statistics mixed training (ASMT) approach using TCAD. Under this method, the TCAD models are incorporated into the machine learning training procedure. The mixed dataset with the experimental and TCAD results improved the prediction in terms of accuracy. With the application of ASMT to the BOSCH process, we show that the mean square error (MSE) can be effectively decreased when the analytics-statistics mixed training (ASMT) scheme is used instead of the classic neural network (NN) used in the baseline study. In this method, statistical induction and analytical deduction can be combined to increase the prediction accuracy of future intelligent semiconductor manufacturing.

An ultra-compact blackbody using electrophoretic deposited carbon nanotube films.

Carbon nanotubes (CNTs) possesses decent optical properties and thus can be considered as a candidate for perfect absorbers due to their close-to-air refractive index and minimal extinction. However, weak absorption in porous materials, due to the low extinction coefficients, requires an inevitably thick absorption layer (∼100 µm) for the perfect opaque absorbers. Thus, the requirement of large thicknesses of CNTs prohibits them from being used as miniaturized integrated photonic devices. Here, we propose an electrophoretic deposited (EPD) CNT resonant cavity structure on tantalum (Ta) to enhance optical absorption. Efficient random light scattering along with the resonant cavity structure using Ti/SiO2 stacking enhances the absorption in our proposed EPD-CNT film while maintaining the total device thickness to <1 µm. The experiment results reveal that the absorption band covers the entire UV-VIS-NIR spectrum (λ = 0.3-2.6 µm), using resonant-cavity EPD-CNT design. The EPD deposition process is done at relatively low temperature < 120 °C. We believe that this proposal is very promising for sensing, antenna, and thermophotovoltaics (TPV), in terms of bandwidth, compactness and cost.

Polarization-selective ultra-broadband super absorber.

While a broadband metamaterial perfect absorber (MPA) has been implemented and proposed intensively in recent years, an ultra-broadband perfect absorber with polarization selectivity has not been realized in literature. In this work, we propose a configuration of polarization-selective (PS) MPA with ultra-wide absorption bandwidth. The aluminum wire grid is integrated on top of the ultrathin-metal-dielectric stacking. The transverse electric (TE) wave is blocked due to the requirement of zero tangential electric field at the metal surface. The transverse magnetic field can pass the aluminum wire-grids because the normal electric field can be supported by the surface charge density at the metal surface, and full absorption of the TM wave is accomplished by the metal-dielectric stacking beneath. Theoretical calculation using rigorously coupled wave analysis demonstrates the wavelength selectivity from λ = 1.98µm to λ = 11.74µm where the TE absorption is <0.04 while TM absorption is >0.95, using 300 nm thick aluminum (Al) wire grid with 16-pair SiO2/Ti stacking. Additionally, the design is wavelength scalable by adjusting the dielectric thickness (tSiO2) and the wire grid period (P) and height (t). The experimental result is demonstrated using Al grids and Ti/SiO2, and the measured result fully supports the calculated prediction.

A fully functionalized metamaterial perfect absorber with simple design and implementation.

Broadband perfect metamaterial absorbers have been drawing significant attention in recent years. A close-to-unity absorption over a broad spectral range is established and this facilitates many photonic applications. A more challenging goal is to construct a broadband absorber with a tailored spectral absorption. The spectral absorption control and spectral shaping are very critical in many applications, such as thermal-photovoltaic, thermal emitters, spectrum imaging system, biomedical and extraterrestrial sensing, and refractive index sensor. In this work, one-dimensional (1D) planar stacking structure is designed to achieve the ultimate goal of a functionalized absorber with a fully tailorable spectral absorption. The lithography and etching process are totally eliminated in this proposed structure, and the fabrication is fully compatible with the regular silicon IC processing. By using ~2 nm ultra-thin metallic layers with a 10-pair (10X) SiO2/Si3N4 integrated dielectric filter, we can achieve decent spectral response shaping. The planar configuration of the ultra-thin-metal metamaterial perfect absorber (MPA) is the key to the easy design/integration of the dielectric filters on top of the MPA. Specifically, band-rejected, high-pass, low-pass and band-pass structure are constructed successfully. Finally, experimental evidence to support our simulation result is also provided, which proves the feasibility of our proposal.

Omnidirectional, polarization-independent, ultra-broadband metamaterial perfect absorber using field-penetration and reflected-wave-cancellation.

In this work, we present the result of nickel (Ni)-based metamaterial perfect absorbers (MPA) with ultra-broadband close-to-one absorbance. The experimental broadband characteristic is significantly improved over the past effort on metamaterial perfect absorbers. An in-depth physical picture and quantitative analysis is presented to reveal the physical origin of its ultrabroadband nature. The key constituent is the cancellation of the reflected wave using ultra-thin, moderate-extinction metallic films. The ultra-thin metal thickness can reduce the reflection as the optical field penetrates through the metallic films. This leads to minimal reflection at each ultra-thin metal layer, and light is penetrating into the Ni/SiO2 stacking. More intuitively, when the layer thickness is much smaller than the photon wavelength, the layer is essentially invisible to the photons. This results in absorption in the metal thin-film through penetration while there is minimal reflection by the metal film. More importantly, the experimental evidence for omni-directionality and polarization-insensitivity are established for the proposed design. Detailed measurement is conducted. Due to the ultrathin metal layers and the satisfactory tolerance in dielectric thickness, the broadband absorption has minimal degradation at oblique incidence. Such a wide angle, polarization-insensitive, ultra-broadband MPA can be very promising in the future, and the optical physics using sub-skin-depth metal film can also facilitate miniaturized high-performance nano-photonic devices.

Experimentally-implemented genetic algorithm (Exp-GA): toward fully optimal photovoltaics.

The geometry and dimension design is the most critical part for the success in nano-photonic devices. The choices of the geometrical parameters dramatically affect the device performance. Most of the time, simulation is conducted to locate the suitable geometry, but in many cases simulation can be ineffective. The most pronounced examples are large-area randomized patterns for solar cells, light emitting diode (LED), and thermophtovoltaics (TPV). The large random pattern is nearly impossible to calculate and optimize due to the extended CPU runtime and the memory limitation. Other scenarios that numerical simulations become ineffective include three-dimensional complex structures with anisotropic dielectric response. This leads to extended simulation time especially for the repeated runs during its geometry optimization. In this paper, we show that by incorporating genetic algorithm (GA) into real-world experiments, shortened trial-and-error time can be achieved. More importantly, this scheme can be used for many photonic design problems that are unsuitable for simulation-based optimizations. Moreover, the experimentally implemented genetic algorithm (Exp-GA) has the additional advantage that the resultant objective value is a real one rather than a theoretical one. This prevents the gaps between the modeling and the fabrication due to the process variation or inaccurate numerical models. Using TPV emitters as an example, 22% enhancement in the mean objective value is achieved.

Approaching conversion limit with all-dielectric solar cell reflectors.

Metallic back reflectors has been used for thin-film and wafer-based solar cells for very long time. Nonetheless, the metallic mirrors might not be the best choices for photovoltaics. In this work, we show that solar cells with all-dielectric reflectors can surpass the best-configured metal-backed devices. Theoretical and experimental results all show that superior large-angle light scattering capability can be achieved by the diffuse medium reflectors, and the solar cell J-V enhancement is higher for solar cells using all-dielectric reflectors. Specifically, the measured diffused scattering efficiency (D.S.E.) of a diffuse medium reflector is >0.8 for the light trapping spectral range (600nm-1000nm), and the measured reflectance of a diffuse medium can be as high as silver if the geometry of embedded titanium oxide(TiO(2)) nanoparticles is optimized. Moreover, the diffuse medium reflectors have the additional advantage of room-temperature processing, low cost, and very high throughput. We believe that using all-dielectric solar cell reflectors is a way to approach the thermodynamic conversion limit by completely excluding metallic dissipation.

Aperiodic and randomized dielectric mirrors: alternatives to metallic back reflectors for solar cells.

Dielectric mirrors have recently emerged for solar cells due to the advantages of lower cost, lower temperature processing, higher throughput, and zero plasmonic absorption as compared to conventional metallic counterparts. Nonetheless, in the past, efforts for incorporating dielectric mirrors into photovoltaics were not successful due to limited bandwidth and insufficient light scattering that prevented their wide usage. In this work, it is shown that the key for ultra-broadband dielectric mirrors is aperiodicity, or randomization. In addition, it has been proven that dielectric mirrors can be widely applicable to thin-film and thick wafer-based solar cells to provide for light trapping comparable to conventional metallic back reflectors at their respective optimal geometries. Finally, the near-field angular emission plot of Poynting vectors is conducted, and it further confirms the superior light-scattering property of dielectric mirrors, especially for diffuse medium reflectors, despite the absence of surface plasmon excitation. The preliminary experimental results also confirm the high feasibility of dielectric mirrors for photovoltaics.

An optimized surface plasmon photovoltaic structure using energy transfer between discrete nano-particles.

Surface plasmon enhancement has been proposed as a way to achieve higher absorption for thin-film photovoltaics, where surface plasmon polariton(SPP) and localized surface plasmon (LSP) are shown to provide dense near field and far field light scattering. Here it is shown that controlled far-field light scattering can be achieved using successive coupling between surface plasmonic (SP) nano-particles. Through genetic algorithm (GA) optimization, energy transfer between discrete nano-particles (ETDNP) is identified, which enhances solar cell efficiency. The optimized energy transfer structure acts like lumped-element transmission line and can properly alter the direction of photon flow. Increased in-plane component of wavevector is thus achieved and photon path length is extended. In addition, Wood-Rayleigh anomaly, at which transmission minimum occurs, is avoided through GA optimization. Optimized energy transfer structure provides 46.95% improvement over baseline planar cell. It achieves larger angular scattering capability compared to conventional surface plasmon polariton back reflector structure and index-guided structure due to SP energy transfer through mode coupling. Via SP mediated energy transfer, an alternative way to control the light flow inside thin-film is proposed, which can be more efficient than conventional index-guided mode using total internal reflection (TIR).

The versatile designs and optimizations for cylindrical TiO2-based scatterers for solar cell anti-reflection coatings.

The anti-reflection coating(ARC) based on dielectric nano-particles has been recently proposed as a new way to achieve the low reflectance required for solar cell front surfaces. In this scenario, the Mie modes associated with the dielectric nano-particles are utilized to facilitate photon forward scattering. In this work, versatile designs together with systematically optimized geometry are examined, for the ARCs based on dielectric scatterers. It is found that the Si3N4-TiO2 or SiO2-TiO2 stack is capable of providing low reflectance while maintaining a flat and passivated ARC-semiconductor interface which can be beneficial for reduced interface recombination and prevent V(OC) degradation associated with topography on the active materials. It is also confirmed that the plasmonic nano-particles placed at the front side of solar cells is not a preferred scheme, even with thorough geometrical optimization. At the ultimate design based on mixed graded index(GI) Mie-scattering, the averaged reflectance can be as low as 0.25%. Such a low reflectance is currently only achievable by ultra-long silicon nano-tips, but silicon nano-tips introduce severe surface recombination. On the other hand, the mixed GI Mie design preserves a flat and passivated ARC-silicon interface, with total thickness reduced to 279.8 nm, much thinner than 1.6 µm for silicon nanotips. In addition, the light trapping capability of mixed GI Mie design is much better than silicon nanotips. In fact, when compared to the state-of-art TiO2 light trapping anti-reflection coating, the mixed GI Mie design provides same light trapping capability while providing much lower reflectance.