OS100: A Benchmark Set of 100 Digitized UV–Visible Spectra and Derived Experimental Oscillator Strengths
OS100: A Benchmark Set of 100 Digitized UV–Visible Spectra and Derived Experimental Oscillator Strengths
复制标题
OS100：100 个数字化紫外可见光谱和衍生实验振荡器强度的基准集
DOI：
10.1021/acs.jpca.1c08988
复制
发表时间：
2022
期刊：
影响因子：
--
通讯作者：
Gozem, Samer
中科院分区：
文献类型：
--
作者：
Tarleton, Astrid S.;Garcia-Alvarez, Jorge C.;Wynn, Anah;Awbrey, Cade M.;Roberts, Tomas P.;Gozem, Samer
激发态量子化学计算通常报告每个电子跃迁的激发能和振子强度f。另一方面，紫外-可见光分光光度实验测量与能量相关的摩尔消光/衰减系数ε(V)，当绘制时，它给出吸收带线的形状。ε(V)和Fare相关，但这种关系因展宽和溶剂化效应而变得复杂。我们对100个实验紫外-可见光谱进行了拟合和积分，得到了这些光谱中出现的吸收带的164 fexp值。100个紫外可见光谱属于溶剂化有机分子，大小从6-34个原子不等。我们估计了拟合中的不确定度，以表明报告fexp值的置信度。用含时密度泛函理论和可极化连续介质溶剂模型计算了相应的振子强度(Fcomp)。通过用一个共同的量来表示实验和计算的吸收强度，我们直接比较了Component和Fexp。尽管Compare和fExp有很好的相关性(线性回归R2=0.921)，但fCompont在大多数情况下高估了fexp(回归斜率=1.34)。通过考虑溶剂折射率系数，化合物绝对值和fexp值之间的一致性得到了显著改善，正如文献中的一些推导所建议的那样。100个数字化的UV-可见光光谱以纯文本文件的形式包含在支持信息中，以帮助对旨在模拟真实UV-可见光吸收光谱的计算或机器学习方法进行基准测试。
Excited-state quantum chemical calculations usually report excitation energies and oscillator strengths,f, for each electronic transition. On the other hand, UV–visible spectrophotometric experiments measure energy-dependent molar extinction/attenuation coefficients, ε(v), that give absorption band line shapes when plotted. ε(v) andfare related, but this relation is complicated by broadening and solvation effects. We fitted and integrated 100 experimental UV–visible spectra to obtain 164fexpvalues for absorption bands appearing in these spectra. The 100 UV–visible spectra belong to solvated organic molecules ranging in size from 6–34 atoms. We estimated uncertainties in the fitting to indicate confidence level in the reportedfexpvalues. The corresponding computed oscillator strengths (fcomp) were obtained with time-dependent density functional theory and a polarizable continuum solvent model. By expressing experimental and computed absorption strengths using a common quantity, we directly comparedfcompandfexp. Althoughfcompandfexpare well correlated (linear regressionR2= 0.921),fcompin most cases overestimatedfexp(regression slope = 1.34). The agreement between absolutefcompandfexpvalues was substantially improved by accounting for a solvent refractive index factor, as suggested in some derivations in the literature. The 100 digitized UV–visible spectra are included as plain text files in the Supporting Information to aid in benchmarking computational or machine learning methods that aim to simulate realistic UV–visible absorption spectra.