What It Takes to Compute Highly Accurate Rovibrational Line Lists for Use in Astrochemistry

Systematic ab initio calculation of spectroscopic constants for A-reduced rotational effective Hamiltonians of asymmetric top molecules using normal ordering of cylindrical angular momentum operators

This research paper presents a new fundamental approach for evaluating accurate ab initio quartic, sextic, and octic centrifugal distortion parameters of A-reduced rotational effective Hamiltonians of asymmetric top molecules. In this framework, the original Watson Hamiltonian, expanded up to sextic terms of kinetic and potential energies, is subjected to a series of vibrational and rotational operator unitary transformations, leading to reduced Watson effective Hamiltonians for the equilibrium configuration, ground state, and weakly perturbed vibrationally excited states. The proposed scheme is based on a numerical-analytic implementation of the sixth-order Van Vleck operator perturbation theory with the systematic normal ordering of vibrational rising and lowering operators (a†, a) and cylindrical angular momentum operators (Jz, J+, J-). The efficiency of the developed theoretical model is demonstrated by the juxtaposition of predicted centrifugal distortion parameters for several three to eight atomic molecules, including H2S, CH2O, C2H4, CH2D2, CH2F2, CH2Cl2, and B2H6, using the coupled-cluster single double triple/quadruple-ζ level of quantum chemistry. In comparison with the values derived using the customary analytic expressions, the calculated quartic and sextic parameters may improve by an order of magnitude in the fourth and sixth orders, respectively, reaching an accuracy of about 1%. Predicted octic constants can serve as an excellent starting point for fitting to experimental spectra.

High-Temperature Line List of Phosphaethyne (HCP)

Phosphaethyne (HCP) has been detected in circumstellar envelopes; its spectroscopic line list is helpful for modeling the relevant atmospheric opacity. We present the first comprehensive line list for HCP(X1Σ+) using robust first-principles methods. The analytical potential energy surface and dipole moment surface were constructed based on 26478 ab initio points from coupled-cluster calculations, along with the considerations of core-valence electron correlation and scalar relativistic effects. The variational nuclear motion program TROVE was used to obtain the ro-vibrational energy levels, Einstein A coefficients, and so on. The J-dependent Coriolis-decoupled Hamiltonian was adopted in the variational calculations with k ≤ 20, and the linear molecule treatment was applied to consider the l-type doubling of the bending vibration. The line list contains almost 0.45 billion transitions between 1.21 million levels with rotational excitation up to J = 200. It covers the wavenumber range of 0-9000 cm-1 (wavelengths above 1.11 μm) and is suitable for temperatures up to 3000 K. The millimeter wave spectra agree well with the experiments, and the Fermi resonance between 2v2 and v3 bands has been reproduced.

Picking up Good Vibrations through Quartic Force Fields and Vibrational Perturbation Theory

Quartic force fields (QFFs) define sparse potential energy surfaces (compared to semiglobal surfaces) that are the cheapest and easiest means of computing anharmonic vibrational frequencies, especially when utilized with second-order vibrational perturbation theory (VPT2). However, flat and shallow potential surfaces are exceedingly difficult for QFFs to treat through a combination of numerical noise in the often numerically computed derivatives and in competing energy factors in the composite energies often utilized to provide high-level spectroscopic predictions. While some of these issues can be alleviated with analytic derivatives, hybrid QFFs, and intelligent choices in coordinate systems, the best practice is for predicting good molecular vibrations via QFFs is to understand what they cannot do, and this manuscript documents such cases where QFFs may fail.

High-temperature spectra of the PNO molecule based on robust first-principles methods

The PNO molecule is an important species found in the interstellar medium, and its spectroscopic information is helpful for its detection. We present the first line list of PNO (X1Σ+) using robust first-principles methods. The analytical potential energy surface and the dipole moment surface were constructed based on 11 942 ab initio points. The variational nuclear motion calculation was implemented in TROVE to obtain the rovibrational energy levels, Einstein A coefficients and other parameters. The J-dependent Coriolis-decoupled Hamiltonian was adopted with k ≤ 15, and the l-type doubling was considered for the bending vibration of the linear molecule. The line list contained almost 5.87 billion transitions between 3.61 million levels with rotational excitation up to J = 200 and was used to generate the PNO spectrum below 3000 K in the wavenumber range from 0 to 6000 cm-1. The millimetre wave spectrum agrees well with available experimental benchmarks. The Fermi resonance effects in the PNO spectrum are universal and complex, resulting in significant intensity increment of the related weak transition. This line list may be helpful for the spectroscopic characterization and possible astronomical detection of PNO, especially in high-temperature environments.

Ames-2021 CO2 Dipole Moment Surface and IR Line Lists: Toward 0.1% Uncertainty for CO2 IR Intensities

A highly accurate CO₂ ab initio dipole moment surface (DMS), Ames-2021, is reported along with ¹²C¹⁶O₂ infrared (IR) intensity comparisons approaching a 1-4‰ level of agreement and uncertainty. The Ames-2021 DMS was accurately fitted from CCSD(T) finite-field dipoles computed with the aug-cc-pVXZ (X = T, Q, 5) basis for C atom and the d-aug-cc-pVXZ (X = T, Q, 5) basis for O atoms, and extrapolated to the one particle basis set limit. Fitting σ_rms is 3.8 × 10^-7 au for 4443 geometries below 15 000 cm^-1. The corresponding IR intensity, S_Ames-2021, are computed using the Ames-2 potential energy surface (PES), which is the best PES available for CO₂. Compared to high accuracy IR studies for 2001i-00001 and 3001i-00001 bands, S_Ames-2021 matches NIST experiment-based intensities [S_NIST-HIT16 or S_HIT20] to -1.0 ± 1.3‰, or matches DLR experiment-based intensities [S_{DLR-HIT16/UCL/Ames}] to 1.9 ± 3.7‰. This indicates the systematic deviations and uncertainties have been significantly reduced in S_Ames-2021. The S_UCL2015 (or S_HITRAN2016) have larger deviations (vs S_DLR) and uncertainties (vs S_DLR, S_NIST) which are attributed to the less accurate Ames-1 PES adopted in UCL-296 line list calculation. The S_Ames-2021 intensity of ¹²C¹⁶O₂ and ¹³C¹⁶O₂ is utilized to derive new absolute ¹³C/¹²C ratios for Vienna PeeDee Belemnite (VPDB) with uncertainty reduced by ¹/₃ or ²/₃. Further evaluation of S_Ames-2021 intensities are carried out on those CO₂ bands discussed in the HITRAN2020 update paper. Consistent improvements and better accuracies are found in band-by-band analysis, except for those bands strongly affected by Coriolis couplings, or very weak bands measured with relatively larger experimental uncertainties. The Ames-2021 296 K IR line lists are generated for 13 CO₂ isotopologues, with 18 000 cm^-1 and S_296 K > 1 × 10^-31 cm/molecule cutoff and then combined with CDSD line positions (except ¹⁴C¹⁶O₂). The Ames-2021 DMS and 296 K IR line lists represent a major improvement over previous CO₂ theoretical IR intensity studies, including Ames-2016, UCL-296, and recent UCL DMS 2021 update. A real 1 permille level of agreement and uncertainty will definitely require both more accurate PES and more accurate DMS.