Statistics within UV–Visible Absorption Spectrum of Ethanolic Azobenzene

Eemeli A. Eronen, Johannes Niskanen · 2025

Read Statistics within UV–Visible Absorption Spectrum of Ethanolic Azobenzene by Eemeli A. Eronen, Johannes Niskanen (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-02610.txt.

Introduction

Eemeli Eronen and Johannes Niskanen pick apart light spectra of a dye in alcohol in this 2025 paper. The dye, azobenzene, shifts shape under light, and its color bands vary a lot from one molecule pose to the next. The team asks which pose traits drive that spread.

A targeted stats method finds a small set of pose axes that explain most of the color spread. Poses that soak up light at a set color turn out to be over-common in that slice. That skew may shape what the dye does next after a flash. The authors say the method filters noise from the traits that truly count in liquid spectra.

Catalog record

CREATOREemeli A. Eronen, Johannes Niskanen
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE28 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-02610

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Statistics within UV–Visible Absorption Spectrum of Ethanolic Azobenzene Eemeli A. Eronen, Johannes Niskanen Abstract We report a statistical simulation of the UV–visible absorption spectrum of trans -azobenzene in ethanol solution. Due to intermolecular interactions, the used explicit solvent environment necessitates accounting for numerous transitions for a spectrum covering the two energetically lowest lines, S 1 and S 2 . Furthermore, the spectrum manifests vast variation as a function of the underlying local structure, in conjunction with previous observations for spectra of liquids in the X-ray regime. We disentangle the complex structure–spectrum relationship using a machine learning-based method known as the emulator-based component analysis. This structural decomposition outperforms commonly used principal component analysis in explained spectral variation and reveals a small subset of latent structural variables responsible for the total spectral variance. Among other structural characteristics, blueshifting of the S 2 line occurs with fewer hydrogen bonds with the ethanol solvent, and a contracted N=N bond within the C–N=N–C bridge. The observed structural dependence of the absorption spectrum thus implies an overrepresentation of certain structural classes after a photoexcitation, potentially significant for the subsequent nuclear dynamics, photophysics, and photochemistry. A liquid system exhibits a wealth of local atomistic structures, which is reflected as strong variation in their X-ray Wernet et al. 2004 ; Ottosson et al. 2011 ; Niskanen et al. 2016 ; Niskanen et al. 2017 ; Eronen et al. 2024a and UV–visible Timrov et al. 2016 ; Bononi et al. 2020 ; Chen et al. 2022 spectra. In general, sampling of the ensemble is required, as the effect is not necessarily captured by just a few representative structures and their respective spectra Gómez et al. 2024 . On one hand, statistical variation of the spectra implies pre-selection of structures for, e.g. , absorption at a given wavelength. On the other hand, this dependence allows for interpreting the details of spectra for the purposes of characterization if the spectrally relevant structural degrees of freedom can be identified. In a liquid system, however, this is not completely straightforward. Fortunately, extensive simulations enable an in-depth study of the phenomenon when combined with analysis techniques capable of transforming numerical results into human-interpretable forms Niskanen et al. 2022 ; Vladyka et al. 2023 ; Eronen et al. 2024b ; Eronen et al. 2025 . We investigate statistical variation in the UV–visible spectrum of trans -azobenzene in ethanol solution using machine learning-based analysis of data based on sampling by molecular dynamics (MD). Photoisomerization between the trans and the cis forms Hartley 1937 of the molecule is an archetypal topic of photochemistry and a mechanism utilized in applications Bandara and Burdette 2012 ; Jerca et al. 2022 . We sampled 30 000 structures from extended tight binding MD (xTBMD) using the GFN1-xTB method Grimme et al. 2017 ; Bannwarth et al. 2021 implemented in the CP2K software Kühne et al. 2020 v. 2024.2. In the simulations, a trans -azobenzene molecule was dissolved in 159 ethanol molecules in a cubic box of L = 25.04 Å for the density 0.7840 g/cm 3 interpolated for pure ethanol liquid at 300 K Khattab et al. 2012 . After equilibration of 20 ps, we sampled the canonical ensemble at 300 K with a 10 fs interval from a production run of 300 ps. We used a global Nosé–Hoover thermostat Nose 1984a ; Nose 1984b with chain length 3, a time constant \tau of 1000 fs, and an MD timestep of 0.5 fs. For evaluation of spectra consisting of 100 transitions on each of the 30 000 structures, we applied time-dependent density functional theory (TD-DFT) implemented in ORCA v.6.0.0 Neese 2022 . We explicitly included ethanol within the cutoff radius of 3.0 Å, and simulated this cluster in a conductor-like polarizable continuum model Garcia-Ratés and Neese 2020 . We used the PBE exchange-correlation potential Perdew et al. 1996 , the def2-SVP basis set Weigend and Ahlrichs 2005 and def2/J auxiliary basis set Weigend 2006 in the spectral simulations. We evaluated the cross section \sigma_{i} of each transition i for a randomly oriented sample using the transition matrix element in the velocity form, and then convolved this stick spectrum to differential cross section \mathrm{d}\sigma/\mathrm{d}E in terms of energy E by a Gaussian of 0.2 eV full width at half maximum. Finally, we transformed the differential cross section into wavelength \lambda scale for \mathrm{d}\sigma/{\mathrm{d}\lambda} . We use a constant width in energy, because a constant width in wavelength makes the lines artificially broad for short wavelengths. Figure 1: Results of the spectrum calculations and variance captured by different dimensionality reduction techniques. a : Simulated ensemble average UV–visible absorption spectrum for trans -azobenzene in ethanol, together with an experiment reproduced from Ref. 28 . The blue-shaded area depicts the standard deviation of the data set, indicating significant statistical variation. The gray-shaded areas depict the two regions of interest (ROIs I and II) defined for the subsequent analysis. Computational results have been shifted -50.6 nm to match the experiment. Additionally, the figure presents an example local structure prepared with Jmol Jmol development team . b : Spectral R 2 score as a function of the rank of emulator-based component analysis (ECA) and principal component analysis (PCA), evaluated using the neural network emulator after the respective dimensionality reduction. The first two ECA components explain a large portion of the spectral ROI variance, while the higher components show rapidly diminishing R 2 score gains. Structural decomposition by PCA covers significantly less spectral variance for the same rank. c : Structural R 2 score as a function of decomposition rank. The ECA vectors cover a small fraction of the structural variance compared to structural PCA. The result indicates that explaining structural variation alone does not explain the spectral variation. Figure 1 a shows the simulated ensemble-average spectrum shifted by -50.6 nm in comparison to the experimental one by Nägele and coworkers Nägele et al. 1997 . We apply the molecular notation S 1 and S 2 for the characteristic features in this spectrum, but note that in our simulation for the gas phase in the optimized geometry, the S 2 line already consists of three transitions. In liquid, we find the photoexcited states of trans -azobenzene to be mixed with states of the ethanol, which results in a wealth of possible final states to be excited – and to be simulated with explicit solvation. In our data, the lines S 1 and S 2 typically consist of several notably intense transitions and numerous weaker ones. The S 1 feature (390 nm–520 nm) is almost unobservable, and we focus on the S 2 peak (270 nm–390 nm). We observe significant statistical variation between the spectra of individual local structures, as shown by the standard deviation \sigma in Figure 1 a. For the following analysis, we define two regions of interest (ROI), split at the maximum of the S 2 peak. Then, we study the mean differential cross section within each ROI as a function of the structure. We first train a neural network to predict the two z-score standardized ROI values for a structure, and then utilize this emulator in emulator-based component analysis (ECA) Niskanen et al. 2022 for dimensionality reduction to identify the decisive structural characteristics behind the

Read the complete text