-
Accelerating computational materials discovery with artificial intelligence and cloud high-performance computing: from large-scale screening to experimental validation
Authors:
Chi Chen,
Dan Thien Nguyen,
Shannon J. Lee,
Nathan A. Baker,
Ajay S. Karakoti,
Linda Lauw,
Craig Owen,
Karl T. Mueller,
Brian A. Bilodeau,
Vijayakumar Murugesan,
Matthias Troyer
Abstract:
High-throughput computational materials discovery has promised significant acceleration of the design and discovery of new materials for many years. Despite a surge in interest and activity, the constraints imposed by large-scale computational resources present a significant bottleneck. Furthermore, examples of large-scale computational discovery carried through experimental validation remain scar…
▽ More
High-throughput computational materials discovery has promised significant acceleration of the design and discovery of new materials for many years. Despite a surge in interest and activity, the constraints imposed by large-scale computational resources present a significant bottleneck. Furthermore, examples of large-scale computational discovery carried through experimental validation remain scarce, especially for materials with product applicability. Here we demonstrate how this vision became reality by first combining state-of-the-art artificial intelligence (AI) models and traditional physics-based models on cloud high-performance computing (HPC) resources to quickly navigate through more than 32 million candidates and predict around half a million potentially stable materials. By focusing on solid-state electrolytes for battery applications, our discovery pipeline further identified 18 promising candidates with new compositions and rediscovered a decade's worth of collective knowledge in the field as a byproduct. By employing around one thousand virtual machines (VMs) in the cloud, this process took less than 80 hours. We then synthesized and experimentally characterized the structures and conductivities of our top candidates, the Na$_x$Li$_{3-x}$YCl$_6$ ($0 < x < 3$) series, demonstrating the potential of these compounds to serve as solid electrolytes. Additional candidate materials that are currently under experimental investigation could offer more examples of the computational discovery of new phases of Li- and Na-conducting solid electrolytes. We believe that this unprecedented approach of synergistically integrating AI models and cloud HPC not only accelerates materials discovery but also showcases the potency of AI-guided experimentation in unlocking transformative scientific breakthroughs with real-world applications.
△ Less
Submitted 8 January, 2024;
originally announced January 2024.
-
Geometry dependence of TLS noise and loss in a-SiC:H parallel plate capacitors for superconducting microwave resonators
Authors:
K. Kouwenhoven,
G. P. J. van Doorn,
B. T. Buijtendorp,
S. A. H. de Rooij,
D. Lamers,
D. J. Thoen,
V. Murugesan,
J. J. A. Baselmans,
P. J. de Visser
Abstract:
Parallel plate capacitors (PPC) significantly reduce the size of superconducting microwave resonators, reducing the pixel pitch for arrays of single photon energy-resolving kinetic inductance detectors (KIDs). The frequency noise of KIDs is typically limited by tunneling Two-Level Systems (TLS), which originate from lattice defects in the dielectric materials required for PPCs. How the frequency n…
▽ More
Parallel plate capacitors (PPC) significantly reduce the size of superconducting microwave resonators, reducing the pixel pitch for arrays of single photon energy-resolving kinetic inductance detectors (KIDs). The frequency noise of KIDs is typically limited by tunneling Two-Level Systems (TLS), which originate from lattice defects in the dielectric materials required for PPCs. How the frequency noise level depends on the PPC's dimensions has not been experimentally addressed. We measure the frequency noise of 56 resonators with a-SiC:H PPCs, which cover a factor 44 in PPC area and a factor 4 in dielectric thickness. To support the noise analysis, we measure the TLS-induced, power-dependent, intrinsic loss and temperature-dependent resonance frequency shift of the resonators. From the TLS models, we expect a geometry-independent microwave loss and resonance frequency shift, set by the TLS properties of the dielectric. However, we observe a thickness-dependent microwave loss and resonance frequency shift, explained by surface layers that limit the performance of PPC-based resonators. For a uniform dielectric, the frequency noise level should scale directly inversely with the PPC area and thickness. We observe that an increase in PPC size reduces the frequency noise, but the exact scaling is, in some cases, weaker than expected. Finally, we derive an engineering guideline for the design of KIDs based on PPC-based resonators.
△ Less
Submitted 8 May, 2024; v1 submitted 21 November, 2023;
originally announced November 2023.
-
Resolving Power of Visible to Near-Infrared Hybrid $β$-Ta/NbTiN Kinetic Inductance Detectors
Authors:
Kevin Kouwenhoven,
Daniel Fan,
Enrico Biancalani,
Steven A. H. de Rooij,
Tawab Karim,
Carlas S. Smith,
Vignesh Murugesan,
David J. Thoen,
Jochem J. A. Baselmans,
Pieter J. de Visser
Abstract:
Kinetic Inductance Detectors (KIDs) are superconducting energy-resolving detectors, sensitive to single photons from the near-infrared to ultraviolet. We study a hybrid KID design consisting of a beta phase tantalum ($β$-Ta) inductor and a NbTiN interdigitated capacitor (IDC). The devices show an average intrinsic quality factor $Q_i$ of 4.3$\times10^5$ $\pm$ 1.3 $\times10^5$. To increase the powe…
▽ More
Kinetic Inductance Detectors (KIDs) are superconducting energy-resolving detectors, sensitive to single photons from the near-infrared to ultraviolet. We study a hybrid KID design consisting of a beta phase tantalum ($β$-Ta) inductor and a NbTiN interdigitated capacitor (IDC). The devices show an average intrinsic quality factor $Q_i$ of 4.3$\times10^5$ $\pm$ 1.3 $\times10^5$. To increase the power captured by the light sensitive inductor, we 3D-print an array of 150$\times$150 $μ$m resin micro lenses on the backside of the sapphire substrate. The shape deviation between design and printed lenses is smaller than 1$μ$m, and the alignment accuracy of this process is $δ_x = +5.8 \pm 0.5$ $μ$m and $δ_y = +8.3 \pm 3.3$ $μ$m. We measure a resolving power for 1545-402 nm that is limited to 4.9 by saturation in the KID's phase response. We can model the saturation in the phase response with the evolution of the number of quasiparticles generated by a photon event. An alternative coordinate system that has a linear response raises the resolving power to 5.9 at 402 nm. We verify the measured resolving power with a two-line measurement using a laser source and a monochromator. We discuss several improvements that can be made to the devices on a route towards KID arrays with high resolving powers.
△ Less
Submitted 13 February, 2023; v1 submitted 12 July, 2022;
originally announced July 2022.
-
Model and Measurements of an Optical Stack for Broadband Visible to Near-IR Absorption in TiN KIDs
Authors:
K. Kouwenhoven,
I. Elwakil,
J. van Wingerden,
V. Murugesan,
D. J. Thoen,
J. J. A. Baselmans,
P. J. de Visser
Abstract:
Typical materials for optical Kinetic Inductance Detetectors (KIDs) are metals with a natural absorption of 30-50% in the visible and near-infrared. To reach high absorption efficiencies (90-100%) the KID must be embedded in an optical stack. We show an optical stack design for a 60 nm TiN film. The optical stack is modeled as sections of transmission lines, where the parameters for each section a…
▽ More
Typical materials for optical Kinetic Inductance Detetectors (KIDs) are metals with a natural absorption of 30-50% in the visible and near-infrared. To reach high absorption efficiencies (90-100%) the KID must be embedded in an optical stack. We show an optical stack design for a 60 nm TiN film. The optical stack is modeled as sections of transmission lines, where the parameters for each section are related to the optical properties of each layer. We derive the complex permittivity of the TiN film from a spectral ellipsometry measurement. The designed optical stack is optimised for broadband absorption and consists of, from top (illumination side) to bottom: 85 nm SiOx, 60 nm TiN, 23 nm of SiOx, and a 100 nm thick Al mirror. We show the modeled absorption and reflection of this stack, which has >80% absorption from 400 nm to 1550 nm and near-unity absorption for 500 nm to 800 nm. We measure transmission and reflection of this stack with a commercial spectrophotometer. The results are in good agreement with the model.
△ Less
Submitted 15 August, 2022; v1 submitted 12 October, 2021;
originally announced October 2021.
-
Superconducting Microstrip Losses at Microwave and Sub-millimeter wavelengths
Authors:
S. Hähnle,
K. Kouwenhoven,
B. Buijtendorp,
A. Endo,
K. Karatsu,
D. J. Thoen,
V. Murugesan,
J. J. A. Baselmans
Abstract:
We present a lab-on-chip experiment to accurately measure losses of superconducting microstrip lines at microwave and sub-mm wavelengths. The microstrips are fabricated from NbTiN, which is deposited using reactive magnetron sputtering, and amorphous silicon which is deposited using plasma-enhanced chemical vapor deposition (PECVD). Sub-mm wave losses are measured using on-chip Fabry-P{é}rot reson…
▽ More
We present a lab-on-chip experiment to accurately measure losses of superconducting microstrip lines at microwave and sub-mm wavelengths. The microstrips are fabricated from NbTiN, which is deposited using reactive magnetron sputtering, and amorphous silicon which is deposited using plasma-enhanced chemical vapor deposition (PECVD). Sub-mm wave losses are measured using on-chip Fabry-P{é}rot resonators (FPR) operating around $350\ $GHz. Microwave losses are measured using shunted half-wave resonators with an identical geometry and fabricated on the same chip. We measure a loss tangent of the amorphous silicon at single-photon energies of $\tanδ=3.7\pm0.5\times10^{-5}$ at $6\ $GHz and $\tanδ= 2.1\pm 0.1\times10^{-4}$ at $350\ $GHz. These results represent very low losses for deposited dielectrics, but the sub-mm wave losses are significantly higher than the microwave losses, which cannot be understood using the standard two-level system loss model.
△ Less
Submitted 25 August, 2021;
originally announced August 2021.
-
Graphical Gaussian Process Regression Model for Aqueous Solvation Free Energy Prediction of Organic Molecules in Redox Flow Battery
Authors:
Peiyuan Gao,
Xiu Yang,
Yu-Hang Tang,
Muqing Zheng,
Amity Anderson,
Vijayakumar Murugesan,
Aaron Hollas,
Wei Wang
Abstract:
The solvation free energy of organic molecules is a critical parameter in determining emergent properties such as solubility, liquid-phase equilibrium constants, and pKa and redox potentials in an organic redox flow battery. In this work, we present a machine learning (ML) model that can learn and predict the aqueous solvation free energy of an organic molecule using Gaussian process regression me…
▽ More
The solvation free energy of organic molecules is a critical parameter in determining emergent properties such as solubility, liquid-phase equilibrium constants, and pKa and redox potentials in an organic redox flow battery. In this work, we present a machine learning (ML) model that can learn and predict the aqueous solvation free energy of an organic molecule using Gaussian process regression method based on a new molecular graph kernel. To investigate the performance of the ML model on electrostatic interaction, the nonpolar interaction contribution of solvent and the conformational entropy of solute in solvation free energy, three data sets with implicit or explicit water solvent models, and contribution of conformational entropy of solute are tested. We demonstrate that our ML model can predict the solvation free energy of molecules at chemical accuracy with a mean absolute error of less than 1 kcal/mol for subsets of the QM9 dataset and the Freesolv database. To solve the general data scarcity problem for a graph-based ML model, we propose a dimension reduction algorithm based on the distance between molecular graphs, which can be used to examine the diversity of the molecular data set. It provides a promising way to build a minimum training set to improve prediction for certain test sets where the space of molecular structures is predetermined.
△ Less
Submitted 15 June, 2021;
originally announced June 2021.
-
Phonon-trap** enhanced energy resolution in superconducting single photon detectors
Authors:
Pieter J. de Visser,
Steven A. H. de Rooij,
Vignesh Murugesan,
David J. Thoen,
Jochem J. A. Baselmans
Abstract:
A noiseless, photon counting detector, which resolves the energy of each photon, could radically change astronomy, biophysics and quantum optics. Superconducting detectors promise an intrinsic resolving power at visible wavelengths of $R=E/δE\approx100$ due to their low excitation energy. We study superconducting energy-resolving Microwave Kinetic Inductance Detectors (MKIDs), which hold particula…
▽ More
A noiseless, photon counting detector, which resolves the energy of each photon, could radically change astronomy, biophysics and quantum optics. Superconducting detectors promise an intrinsic resolving power at visible wavelengths of $R=E/δE\approx100$ due to their low excitation energy. We study superconducting energy-resolving Microwave Kinetic Inductance Detectors (MKIDs), which hold particular promise for larger cameras. A visible/near-infrared photon absorbed in the superconductor creates a few thousand quasiparticles through several stages of electron-phonon interaction. Here we demonstrate experimentally that the resolving power of MKIDs at visible to near-infrared wavelengths is limited by the loss of hot phonons during this process. We measure the resolving power of our aluminum-based detector as a function of photon energy using four lasers with wavelengths between $1545-402$ nm. For detectors on thick SiN/Si and sapphire substrates the resolving power is limited to $10-21$ for the respective wavelengths, consistent with the loss of hot phonons. When we suspend the sensitive part of the detector on a 110 nm thick SiN membrane, the measured resolving power improves to $19-52$ respectively. The improvement is equivalent to a factor $8\pm2$ stronger phonon trap** on the membrane, which is consistent with a geometrical phonon propagation model for these hot phonons. We discuss a route towards the Fano limit by phonon engineering.
△ Less
Submitted 11 March, 2021;
originally announced March 2021.
-
Suppression of radiation loss in high kinetic inductance superconducting co-planar waveguides
Authors:
S. Hähnle,
N. v. Marrewijk,
A. Endo,
K. Karatsu,
D. J. Thoen,
V. Murugesan,
J. J. A. Baselmans
Abstract:
We present a novel lab-on-chip technique to measure the very low losses in superconducting transmission lines at (sub-) mm wavelengths. The chips consist of a 100 nm thick NbTiN Coplanar Waveguide (CPW) Fabry Perot (FP) resonator, coupled on one side to an antenna and on the other side to a Microwave Kinetic Inductance detector. Using a single frequency radiation source allows us to measure the fr…
▽ More
We present a novel lab-on-chip technique to measure the very low losses in superconducting transmission lines at (sub-) mm wavelengths. The chips consist of a 100 nm thick NbTiN Coplanar Waveguide (CPW) Fabry Perot (FP) resonator, coupled on one side to an antenna and on the other side to a Microwave Kinetic Inductance detector. Using a single frequency radiation source allows us to measure the frequency response of the FP around 350 GHz and deduce its losses. We show that the loss is dominated by radiation loss inside the CPW line that forms the FP and that it decreases with decreasing line width and increasing kinetic inductance as expected. The results can be quantitatively understood using SONNET simulations. The lowest loss is observed for a CPW with a total width of $6\ \mathrm{μm}$ and corresponds to a Q-factor of $\approx15,000$.
△ Less
Submitted 28 April, 2020; v1 submitted 23 March, 2020;
originally announced March 2020.
-
An Ultra-Wideband Leaky Lens Antenna for Broadband Spectroscopic Imaging Applications
Authors:
Sebastian Hähnle,
Ozan Yurduseven,
Sven van Berkel,
Nuria Llombart,
Juan Bueno,
Stephen J. C. Yates,
Vignesh Murugesan,
David J. Thoen,
Andrea Neto,
Jochem J. A. Baselmans
Abstract:
We present the design, fabrication and characterisation of a broadband leaky lens antenna for broadband, spectroscopic imaging applications. The antenna is designed for operation in the 300-900 GHz band. We integrate the antenna directly into an Al-NbTiN hybrid MKID to measure the beam pattern and absolute coupling efficiency at three frequency bands centred around 350, 650 and 850 GHz, covering t…
▽ More
We present the design, fabrication and characterisation of a broadband leaky lens antenna for broadband, spectroscopic imaging applications. The antenna is designed for operation in the 300-900 GHz band. We integrate the antenna directly into an Al-NbTiN hybrid MKID to measure the beam pattern and absolute coupling efficiency at three frequency bands centred around 350, 650 and 850 GHz, covering the full antenna band. We find an aperture efficiency $η_{ap} \approx 0.4$ over the whole frequency band, limited by lens reflections. We find a good match with simulations for both the patterns and efficiency, demonstrating a 1:3 bandwidth in the sub-mm wavelength range for future on-chip spectrometers.
△ Less
Submitted 16 December, 2019;
originally announced December 2019.
-
Mitigation of Cosmic Ray Effect on Microwave Kinetic Inductance Detector Arrays
Authors:
K. Karatsu,
A. Endo,
J. Bueno,
P. J. de Visser,
R. Barends,
D. J. Thoen,
V. Murugesan,
N. Tomita,
J. J. A. Baselmans
Abstract:
For space observatories, the glitches caused by high energy phonons created by the interaction of cosmic ray particles with the detector substrate lead to dead time during observation. Mitigating the impact of cosmic rays is therefore an important requirement for detectors to be used in future space missions. In order to investigate possible solutions, we carry out a systematic study by testing fo…
▽ More
For space observatories, the glitches caused by high energy phonons created by the interaction of cosmic ray particles with the detector substrate lead to dead time during observation. Mitigating the impact of cosmic rays is therefore an important requirement for detectors to be used in future space missions. In order to investigate possible solutions, we carry out a systematic study by testing four large arrays of Microwave Kinetic Inductance Detectors (MKIDs), each consisting of $\sim$960 pixels and fabricated on monolithic 55 mm $\times$ 55 mm $\times$ 0.35 mm Si substrates. We compare the response to cosmic ray interactions in our laboratory for different detector arrays: A standard array with only the MKID array as reference; an array with a low $T_c$ superconducting film as phonon absorber on the opposite side of the substrate; and arrays with MKIDs on membranes. The idea is that the low $T_c$ layer down-converts the phonon energy to values below the pair breaking threshold of the MKIDs, and the membranes isolate the sensitive part of the MKIDs from phonons created in the substrate. We find that the dead time can be reduced up to a factor of 40 when compared to the reference array. Simulations show that the dead time can be reduced to below 1 % for the tested detector arrays when operated in a spacecraft in an L2 or a similar far-Earth orbit. The technique described here is also applicable and important for large superconducting qubit arrays for future quantum computers.
△ Less
Submitted 10 January, 2019; v1 submitted 8 January, 2019;
originally announced January 2019.
-
Surface wave control for large arrays of microwave kinetic inductance detectors
Authors:
Stephen J. C. Yates,
Andrey M. Baryshev,
Ozan Yurduseven,
Juan Bueno,
Kristina K. Davis,
Lorenza Ferrari,
Willem Jellema,
Nuria Llombart,
Vignesh Murugesan,
David J. Thoen,
Jochem J. A. Baselmans
Abstract:
Large ultra-sensitive detector arrays are needed for present and future observatories for far infra-red, submillimeter wave (THz), and millimeter wave astronomy. With increasing array size, it is increasingly important to control stray radiation inside the detector chips themselves, the surface wave. We demonstrate this effect with focal plane arrays of 880 lens-antenna coupled Microwave Kinetic I…
▽ More
Large ultra-sensitive detector arrays are needed for present and future observatories for far infra-red, submillimeter wave (THz), and millimeter wave astronomy. With increasing array size, it is increasingly important to control stray radiation inside the detector chips themselves, the surface wave. We demonstrate this effect with focal plane arrays of 880 lens-antenna coupled Microwave Kinetic Inductance Detectors (MKIDs). Presented here are near field measurements of the MKID optical response versus the position on the array of a reimaged optical source. We demonstrate that the optical response of a detector in these arrays saturates off-pixel at the $\sim-30$ dB level compared to the peak pixel response. The result is that the power detected from a point source at the pixel position is almost identical to the stray response integrated over the chip area. With such a contribution, it would be impossible to measure extended sources, while the point source sensitivity is degraded due to an increase of the stray loading. However, we show that by incorporating an on-chip stray light absorber, the surface wave contribution is reduced by a factor $>$10. With the on-chip stray light absorber the point source response is close to simulations down to the $\sim-35$ dB level, the simulation based on an ideal Gaussian illumination of the optics. In addition, as a crosscheck we show that the extended source response of a single pixel in the array with the absorbing grid is in agreement with the integral of the point source measurements.
△ Less
Submitted 7 July, 2017;
originally announced July 2017.
