-
Bridging the gap between photovoltaics R&D and manufacturing with data-driven optimization
Authors:
Felipe Oviedo,
Zekun Ren,
Xue Hansong,
Siyu Isaac Parker Tian,
Kaicheng Zhang,
Mariya Layurova,
Thomas Heumueller,
Ning Li,
Erik Birgersson,
Shi**g Sun,
Benji Mayurama,
Ian Marius Peters,
Christoph J. Brabec,
John Fisher III,
Tonio Buonassisi
Abstract:
Novel photovoltaics, such as perovskites and perovskite-inspired materials, have shown great promise due to high efficiency and potentially low manufacturing cost. So far, solar cell R&D has mostly focused on achieving record efficiencies, a process that often results in small batches, large variance, and limited understanding of the physical causes of underperformance. This approach is intensive…
▽ More
Novel photovoltaics, such as perovskites and perovskite-inspired materials, have shown great promise due to high efficiency and potentially low manufacturing cost. So far, solar cell R&D has mostly focused on achieving record efficiencies, a process that often results in small batches, large variance, and limited understanding of the physical causes of underperformance. This approach is intensive in time and resources, and ignores many relevant factors for industrial production, particularly the need for high reproducibility and high manufacturing yield, and the accompanying need of physical insights. The record-efficiency paradigm is effective in early-stage R&D, but becomes unsuitable for industrial translation, requiring a repetition of the optimization procedure in the industrial setting. This mismatch between optimization objectives, combined with the complexity of physical root-cause analysis, contributes to decade-long timelines to transfer new technologies into the market. Based on recent machine learning and technoeconomic advances, our perspective articulates a data-driven optimization framework to bridge R&D and manufacturing optimization approaches. We extend the maximum-efficiency optimization paradigm by considering two additional dimensions: a technoeconomic figure of merit and scalable physical inference. Our framework naturally aligns different stages of technology development with shared optimization objectives, and accelerates the optimization process by providing physical insights.
△ Less
Submitted 28 April, 2020;
originally announced April 2020.
-
Embedding Physics Domain Knowledge into a Bayesian Network Enables Layer-by-Layer Process Innovation for Photovoltaics
Authors:
Zekun Ren,
Felipe Oviedo,
Muang Thway,
Siyu I. P. Tian,
Yue Wang,
Hansong Xue,
Jose Dario Perea,
Mariya Layurova,
Thomas Heumueller,
Erik Birgersson,
Armin Aberle,
Christoph J. Brabec,
Rolf Stangl,
Shi**g Sun,
Qianxiao Li,
Fen Lin,
Ian Marius Peters,
Tonio Buonassisi
Abstract:
Process optimization of photovoltaic devices is a time-intensive, trial and error endeavor, without full transparency of the underlying physics, and with user-imposed constraints that may or may not lead to a global optimum. Herein, we demonstrate that embedding physics domain knowledge into a Bayesian network enables an optimization approach that identifies the root cause(s) of underperformance w…
▽ More
Process optimization of photovoltaic devices is a time-intensive, trial and error endeavor, without full transparency of the underlying physics, and with user-imposed constraints that may or may not lead to a global optimum. Herein, we demonstrate that embedding physics domain knowledge into a Bayesian network enables an optimization approach that identifies the root cause(s) of underperformance with layer by-layer resolution and reveals alternative optimal process windows beyond global black-box optimization. Our Bayesian-network approach links process conditions to materials descriptors (bulk and interface properties, e.g., bulk lifetime, do**, and surface recombination) and device performance parameters (e.g., cell efficiency), using a Bayesian inference framework with an autoencoder-based surrogate device-physics model that is 100x faster than numerical solvers. With the trained surrogate model, our approach is robust and reduces significantly the time consuming experimentalist intervention, even with small numbers of fabricated samples. To demonstrate our method, we perform layer-by-layer optimization of GaAs solar cells. In a single cycle of learning, we find an improved growth temperature for the GaAs solar cells without any secondary measurements, and demonstrate a 6.5% relative AM1.5G efficiency improvement above baseline and traditional black-box optimization methods.
△ Less
Submitted 3 November, 2019; v1 submitted 25 July, 2019;
originally announced July 2019.
-
Nuclear deformation and neutron excess as competing effects for pygmy dipole strength
Authors:
R. Massarczyk,
R. Schwengner,
F. Dönau,
S. Frauendorf,
M. Anders,
D. Bemmerer,
R. Beyer,
C. Bhatia,
E. Birgersson,
M. Butterling,
Z. Elekes,
A. Ferrari,
M. E. Gooden,
R. Hannaske,
A. R. Junghans,
M. Kempe,
J. H. Kelley,
T. Kögler,
A. Matic,
M. L. Menzel,
S. Müller,
T. P. Reinhardt,
M. Röder,
G. Rusev,
K. D. Schilling
, et al. (5 additional authors not shown)
Abstract:
The electromagnetic dipole strength below the neutron-separation energy has been studied for the xenon isotopes with mass numbers A = 124, 128, 132, and 134 in nuclear resonance fluorescence experiments using the ELBE bremsstrahlung facility at Helmholtz-Zentrum Dresden-Rossendorf and the HIgS facility at Triangle Universities Nuclear Laboratory Durham. The systematic study gained new information…
▽ More
The electromagnetic dipole strength below the neutron-separation energy has been studied for the xenon isotopes with mass numbers A = 124, 128, 132, and 134 in nuclear resonance fluorescence experiments using the ELBE bremsstrahlung facility at Helmholtz-Zentrum Dresden-Rossendorf and the HIgS facility at Triangle Universities Nuclear Laboratory Durham. The systematic study gained new information about the influence of the neutron excess as well as of nuclear deformation on the strength in the region of the pygmy dipole resonance. The results are compared with those obtained for the chain of molybdenum isotopes and with predictions of a random-phase approximation in a deformed basis. It turned out that the effect of nuclear deformation plays a minor role compared with the one caused by neutron excess. A global parametrization of the strength in terms of neutron and proton numbers allowed us to derive a formula capable of predicting the summed E1 strengths in the pygmy region for a wide mass range of nuclides.
△ Less
Submitted 8 November, 2013;
originally announced November 2013.
-
Neutron total cross section measurements of gold and tantalum at the nELBE photoneutron source
Authors:
Roland Hannaske,
Zoltan Elekes,
Roland Beyer,
Arnd Junghans,
Daniel Bemmerer,
Evert Birgersson,
Anna Ferrari,
Eckart Grosse,
Mathias Kempe,
Toni Kögler,
Michele Marta,
Ralph Massarczyk,
Andrija Matic,
Georg Schramm,
Ronald Schwengner,
Andreas Wagner
Abstract:
Neutron total cross sections of $^{197}$Au and $^\text{nat}$Ta have been measured at the nELBE photoneutron source in the energy range from 0.1 - 10 MeV with a statistical uncertainty of up to 2 % and a total systematic uncertainty of 1 %. This facility is optimized for the fast neutron energy range and combines an excellent time structure of the neutron pulses (electron bunch width 5 ps) with a s…
▽ More
Neutron total cross sections of $^{197}$Au and $^\text{nat}$Ta have been measured at the nELBE photoneutron source in the energy range from 0.1 - 10 MeV with a statistical uncertainty of up to 2 % and a total systematic uncertainty of 1 %. This facility is optimized for the fast neutron energy range and combines an excellent time structure of the neutron pulses (electron bunch width 5 ps) with a short flight path of 7 m. Because of the low instantaneous neutron flux transmission measurements of neutron total cross sections are possible, that exhibit very different beam and background conditions than found at other neutron sources.
△ Less
Submitted 5 November, 2013;
originally announced November 2013.
