Spectra-structure correlations in NIR region: Spectroscopic and anharmonic DFT study of n-hexanol, cyclohexanol and phenol

SAA-15760; No of Pages 9 Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2018) xxx–xxx

Contents lists available at ScienceDirect

Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa

Spectra-structure correlations in NIR region: Spectroscopic and anharmonic DFT study of n-hexanol, cyclohexanol and phenol Krzysztof B. Beć a,⁎, Justyna Grabska a, Mirosław A. Czarnecki b a b

Department of Chemistry, School of Science and Technology, Kwansei Gakuin University, Sanda, Hyogo 669-1337, Japan Faculty of Chemistry, University of Wrocław, F. Joliot-Curie 14, 50-383 Wrocław, Poland

a r t i c l e

i n f o

Article history: Received 3 October 2017 Received in revised form 8 January 2018 Accepted 13 January 2018 Available online xxxx Keywords: Near-infrared spectroscopy Alcohols Phenols Quantum chemical calculation Overtones Combination modes Time-independent vibrational Schrödinger equation Spectra-structure correlation Anharmonic spectra

a b s t r a c t We investigated near-infrared (7500–4000 cm−1) spectra of n-hexanol, cyclohexanol and phenol in CCl4 (0.2 M) by using anharmonic quantum calculations. These molecules represent three major kinds of alcohols; linear and cyclic aliphatic, and aromatic ones. Vibrational second-order perturbation theory (VPT2) was employed to calculate the first overtones and binary combination modes and to reproduce the experimental NIR spectra. The level of conformational flexibility of these three alcohols varies from one stable conformer of phenol through four conformers of cyclohexanol to few hundreds conformers in the case of n-hexanol. To take into account the most relevant conformational population of n-hexanol, a systematic conformational search was performed. Accurate reproduction of the experimental NIR spectra was achieved and detailed spectra-structure correlations were obtained for these three alcohols. VPT2 approach provides less reliable description of highly anharmonic modes, i.e. OH stretching. In the present work this limitation was manifested in erroneous results yielded by VPT2 for 2νOH mode of cyclohexanol. To study the anharmonicity of this mode we solved the corresponding time-independent Schrödinger equation based on a dense-grid probing of the relevant vibrational potential. These results allowed for significant improvement of the agreement between the calculated and experimental 2νOH band of cyclohexanol. Various important biomolecules include similar structural units to the systems investigated here. A detailed knowledge on spectral properties of these three types of alcohols is therefore essential for advancing our understanding of NIR spectroscopy of biomolecules. © 2018 Elsevier B.V. All rights reserved.

1. Introduction1 Near-infrared (NIR) spectroscopy (12,500–4000 cm−1; 800–2500 nm) has been growing in importance in science and technology over the last two decades [1–3]. In relation to other vibrational spectroscopies (mid-IR or Raman), NIR spectroscopy (NIRS) offers advantages such as simpler instrumentation and general versatility [1,4]. Numerous physicochemical studies i.e. on anharmonicity of molecular vibrations [3,5], molecular structure [3,6], intermolecular interactions [3,7,8], and hydrogen-bonding [3,9–13], solution chemistry [14] and microheterogeneity [15], solvent effects [16,17], etc., provide a good basis for applications of NIRS in analytical chemistry. NIRS appears to be very useful for qualitative and quantitative analysis of i.e. natural products [18], food [19], pharmaceuticals [20], medical samples [21] and medical tools [22]. Therefore, NIRS has become an important tool in a scientific or industrial laboratory.

⁎ Corresponding author. E-mail address: [email protected] (K.B. Beć). This article is dedicated to Professor Yukihiro Ozaki from Kwansei Gakuin University on the occasion of his retirement, to honour his work which significantly advanced our knowledge on near infrared spectroscopy. 1

Unfortunately, NIR spectra remain difficult for direct interpretation. The spectral information is intrinsically complex as a result of significant overlapping of overtones and numerous combination bands [23]. Therefore, applications of NIRS strongly rely on statistical data analysis [3]. As a result, NIRS is often used as a black-box tool. However, the field of applications of NIRS could be extended if the correlations between the molecular properties of a sample and its NIR spectrum were known in detail. Quantum mechanical calculations are one of primary sources of an independent insight in the case of IR and Raman spectroscopies. The analysis of fundamental bands can be carried out by ordinary harmonic calculations [24]. On the other hand, the case of NIR spectra unequivocally requires an anharmonic approach. This imposes considerable requirements on the accuracy and computational affordability of calculations [2,25]. For a long time only the simplest molecules could be satisfactorily treated with fully anharmonic calculations. Therefore, our understanding of NIR spectra remains insufficient [1–3,26]. Advances in the theory of anharmonic methods and a rapid growth of available computational power over the years have allowed to push the limits in theoretical NIRS. Recent development of deperturbed and generalized VPT2 schemes [27], which combine good accuracy with relatively modest computational complexity and high versatility, should be noted. As a result, a growing number of NIR spectroscopic studies

https://doi.org/10.1016/j.saa.2018.01.041 1386-1425/© 2018 Elsevier B.V. All rights reserved.

Please cite this article as: K.B. Beć, et al., Spectra-structure correlations in NIR region: Spectroscopic and anharmonic DFT study..., Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2018), https://doi.org/10.1016/j.saa.2018.01.041

2

K.B. Beć et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2018) xxx–xxx

aided by VPT2 calculations could be observed over the last two or three years [10,18,28–31]. Nowadays, the theoretical NIR spectra of fairly complex molecules, i.e. rosmarinic acid [18] or medium-chain fatty acids [32] can be successfully obtained. The OH group strongly affects physicochemical properties of alcohols. The OH group can be attached to various types of molecular structures, i.e. normal or cyclic aliphatic chain or aromatic ring. The properties of this group, in particular its vibrational frequencies, strongly depend on the environment. For this reason, it is of high interest to establish detailed correlations between the structural factors of the three types of alcohols and their specific NIR spectral features. The first overtones of the free or weakly bonded OH are very pronounced in NIR spectra; therefore their spectral parameters (intensity, position, half-width) provide rich information on self-association and interactions with other molecules. It is relatively easy to study the OH stretching of the first overtone bands and so far many investigations focused on these bands [1,3,6,8–11,13]. In contrast, the other bands appearing in NIR region have not been used so often, despite carrying plentiful information on the structure and properties of these molecules [3,7] The complexity of NIR spectra has been the major hindrance in such studies. To overcome this problem it is necessary to establish detailed band assignments and obtain comprehensive spectra-structure correlations in a broad range of NIR (7500–3800 cm−1). Here, we attempt to deeply analyze NIR spectra of medium-size alcohols which bear distinct structural differences. The OH group appears in variety of molecules, i.e. carbohydrates, nucleic acids; aliphatic OH group plays a role in the adsorption of proteins or nucleic acids [33,34], cyclic alcohols appear in metabolic paths [35,36], polyphenols in natural products act as radical scavengers [18], etc. The aim of the present work is to elucidate the differences between the NIR spectra for three major kinds of alcohols represented by n-hexanol, cyclohexanol and phenol, on the basis of anharmonic calculations. To achieve this aim we employ two different approaches. Firstly, second-order vibrational theory (VPT2) computations are used to simulate NIR spectra from 7500 to 4000 cm−1. An analysis of the contributions from overtone and combination bands in various spectral subregions provides reliable band assignments. The limitations of VPT2 in describing highly anharmonic modes, such as OH stretching mode, clearly appears in the case of cyclohexanol. Secondly, in the latter case we perform a detailed study of the vibrational potential, vibrational levels and corresponding transition frequencies on the ground of numerical solving of time-independent Schrödinger equation.

3.1 for detailed explanation), geometry optimization and anharmonic vibrational analysis were performed with the use of Density Functional Theory (DFT). Single-hybrid B3LYP density functional coupled with 631G(d,p) basis set for conformational searches, and SNST basis set for the subsequent computations were employed. The chosen methods have been reported to be very efficient and accurate in similar studies [10,25,28,30–32]. Superfine grids for integration and solving CPHF equations, and very tight convergence of geometry optimization were applied; with the exception of conformational searches (for which standard options were used). The calculations were carried out with conductor-like polarizable continuum model (CPCM) [38] of CCl4 solvent and third formulation of Grimme's empirical correction for dispersion with Becke-Johnson damping (GD3BJ) [39]. To simulate NIR spectra of studied alcohols we performed fully anharmonic vibrational analysis by means of deperturbed/generalized vibrational second-order perturbation theory (DVPT2/GVPT2); [27] tightly coupled modes were not subjected to variational treatment. This computational procedure enabled us to obtain information on the first overtones and binary combinations. As shown [28,30–32], this approach is sufficient to capture the major features of NIR spectra without extensive calculations. The vibrational frequencies and intensities were used for reconstruction of NIR bands for particular model structures in the same way as previously described [30,45]. The final theoretical NIR spectra were constructed as weighted-sum of calculated spectra of different conformational isomers (ref. Section 3.1) with respect to the calculated Boltzmann coefficients. The Boltzmann coefficients were derived from Gibbs free energies (B3LYP/SNST) corresponding to 298 K, additionally corrected by anharmonic zero-point energy (ZPE) values. This approach worked well with the exception of cyclohexanol case, which will be discussed in detail in Section 3.6. The band assignments were performed with an aid of potential distribution analysis carried out in Gar2Ped software [40], after defining a non-redundant set of natural internal coordinates in accordance with Pulay et al. [41] To enable accurate calculation of anharmonicity of OH stretching vibration, in the case where VPT2 calculation scheme gave erroneous results (cyclohexanol), an independent approach of numerical solving of time-independent Schrödinger equation (Eq. (1)) was employed: 2

∂ ΨðQ Þ ∂Q

2

¼

  2μ ð V ð Q Þ−E Þ ΨðQ Þ ℏ2

ð1Þ

2. Materials and Methods 2.1. Experimental All samples were purchased from Wako Pure Chemical Industries Japan (n-hexanol, min. 97%; cyclohexanol, min. 98%; phenol, min. 99%) and dried by freshly activated molecular sieves (Wako Pure Chemical Industries Japan, 4 Å pore size). The NIR spectra were measured in 10,000–3700 cm−1 range on a Perkin Elmer Spectrum One NTS FT-NIR spectrometer operating in a transmittance mode. The solutions of 0.2 M in CCl4 (Infinity Pure, min. 99.9%; Wako Pure Chemical Industries Japan; dried similar as above) were placed in rectangular quartz cell of optical path of 10 mm. Spectral measurements were performed at resolution of 4 cm−1, resulting in an interpolated data spacing of 1 cm−1, and 64 scans were accumulated. Each spectrum was recorded 3 times, preceded by a background collection (the spectrum of the solvent). The spectra were measured at a controlled temperature of 298 K. Baseline correction was performed by using the software which operates the spectrometer. No other spectral pre-treatment was applied. 2.2. Computational Details All quantum mechanical calculations were carried out using Gaussian 09 Rev. E.01 software [37]. Conformational search (refer to Section

In Eq. (1) Q denotes the respective normal coordinate, Ψ the wave function, μ the reduced mass of the corresponding oscillator, V the potential energy, and E the energy eigenvalue. The scan of the potential energy over the OH stretching normal coordinate was performed from − 0.4 to 2.0 Å with 0.005 Å/step. The harmonic analysis to determine the normal coordinate was carried out at B3LYP/6-311++G(3df,3pd) level, with preliminary geometry optimization using very tight convergence criteria, 10−12 SCF convergence level, superfine integration and CPHF grids and CPCM solvent model of CCl4. The following grid-based energies were obtained with the use of 6-311G(d,p) basis set, all other parameters being equal to those listed above. The solution of the corresponding Schrödinger equation was performed by means of generalized Numerov's method, with seven-point numerical differentiation [42]. 3. Results and Discussion 3.1. The Conformational Flexibility of Alcohols The studied alcohols (Fig. 1a–c) differ notably in the levels of conformational flexibility. Phenol is the simplest case as it features only one stable conformer, with O\\H bond in plane of the aromatic ring (Fig. S1 in Supplementary Material). Six-membered saturated ring of cyclohexanol takes four stable conformations, in which the OH group

Please cite this article as: K.B. Beć, et al., Spectra-structure correlations in NIR region: Spectroscopic and anharmonic DFT study..., Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2018), https://doi.org/10.1016/j.saa.2018.01.041

K.B. Beć et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2018) xxx–xxx

is either in gauche or trans position (with the respective values of OHCH torsional angle: 64 and 180°). All conformers of cyclohexanol were included in the computations of NIR spectrum, although the axial conformers remain in minority (with total abundance of around 5%). The energetic relations between equatorial gauche and trans conformers appears to be highly dependent on the computational method (see Section 3.6 and Tables S2–S3 in Supplementary Material). n-Hexanol required a different approach as it features a significant level of conformational flexibility. Due to 5 possible rotations (one about C\\O and four about C\\C bonds) the total number of stable conformations of n-hexanol (assuming two gauche and one trans configurations, and regarding the symmetry properties) is 243. For this reason we performed a six-dimensional rigid scan of potential energy surface (PES) of n-hexanol molecule at B3LYP/6-31G(d,p) level. Afterwards, we selected the representative population of the most stable conformers, by selecting those within 2 kcal/mol (versus the most stable one) of the relative B3LYP/6-31G(d,p) single point energy as resulted from the PES scan. At this step the structures redundant due to symmetry operations were excluded. The selected 46 non-redundant conformers were subjected to geometry optimization (B3LYP/SNST) which reduced the number of unique, spectroscopically distinguishable conformers of n-hexanol to 32 (for detailed information refer to Table S1 in Supplementary Materials) which were the basis for the subsequent spectroscopic computations, as described in Section 2.2.

3

Fig. 2. Experimental (0.2 M, CCl4) and calculated NIR spectra of n-hexanol together with proposed band assignments. Band numbers correspond to those presented in Table 1. For better view of details refer to Fig. S1 in Supplementary Material.

3.3. Analysis of NIR Spectrum of Cyclohexanol 3.2. Analysis of NIR Spectrum of n-Hexanol In the NIR spectrum of n-hexanol (Fig. 2, Table 1) four major spectral subregions can be noticed. In the range of 7150–7000 cm−1 a moderately intense first overtone band of OH stretching mode appears, with Amax reached at 7103 cm−1. Between 6000 and 5500 cm−1 a number of overlapped bands due to stretching modes of the CH2 and CH3 groups appear. This region was not reproduced satisfactorily by the theoretical calculations due to overestimation of the intensities of the overtone bands (Fig. 1). The region between 5000 and 4600 cm−1 shows a broadened structure of weak intensity due to the combination bands of the OH stretching and CH2 deformation modes. The band intensities in the low frequency fragment (~4700 cm−1) were underestimated in the calculated spectrum. Probably this results from the contribution of minor conformers (Section 3.1) and also higher order modes (i.e. second overtones, ternary combinations), which were not included in our calculations. The dominant region of NIR spectrum of n-hexanol appears between 4350 and 4000 cm−1, with the major peak at 4336 cm−1. It is the most complex region, in which multiple combination bands overlap. The most important contributions are due to combinations of deformation and stretching modes of the methylene groups. The NIR spectrum of n-hexanol systematically deviates from the spectra of lower weight aliphatic alcohols due to an increased contribution of the bands involving CH2 group vibrations (Fig. 2, Table 1) [28,45,43].

In comparison to n-hexanol, the NIR spectrum of cyclohexanol reveals similar four spectral subregions with comparable relative peak intensities (Fig. 3, Table 2). In the experimental spectrum 2νOH peak appears at 7073 cm−1; VPT2 calculations failed to predict this value correctly as will be discussed in Section 3.6. The region of 6000–5500 cm−1 is populated by overtone and combination bands of CH2 stretching modes. Between 5000 and 4500 cm−1 a number of peaks arise from the combinations of OH stretching and CH2 deformations. The 4400– 4000 cm−1 region is populated by intense peaks due to combinations of CH2 stretching and deformation (scissoring, wagging and twisting) modes. Since modeling of the NIR spectrum of cyclohexanol involved its entire conformational population (instead of selected representation as in the case of n-hexanol), the agreement with experimental spectrum is higher than that in the previous case (except the mentioned 2νOH band). Besides, n-hexanol also features a CH3 group in its structure; this significantly increases a number of the binary combinations in the relevant region. As can be seen from Table 2, the calculated combinations involving stretching OH mode reveal a considerably lower agreement with the experimental wavenumbers. The relatively poor accuracy of prediction of highly anharmonic modes, such as OH stretching, is a known limitation of VPT2 approach; the case will be discussed in Section 3.6. It is noteworthy that this effect propagates onto combination modes involving νOH vibration as well.

Fig. 1. Molecular structure of major conformers of (A) n-hexanol, (B) cyclohexanol, (C) phenol.

Please cite this article as: K.B. Beć, et al., Spectra-structure correlations in NIR region: Spectroscopic and anharmonic DFT study..., Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2018), https://doi.org/10.1016/j.saa.2018.01.041

4

K.B. Beć et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2018) xxx–xxx

Table 1 Experimental and calculated NIR bands of n-hexanol and proposed band assignments. Band number

Wavenumber/cm−1

Band assignment

Experimental

Calculated

Diff.

1 2 3 4

7103 6514 5912 5865

7077 6472 5855 5827, 5806

−26 −42 −57 –

5 6

5807 5682

5777 5760

−30 78

7

5620–5400

5740–5500

–

8 9 10

5010 4962 4900–4700

4990 4951 4900–4700

−20 −11 –

11 12

4401 4336

4385 4333

−16 −3

13 14

4265 4200–4100

4281 4200–4100

16 –

15

4068

4054

−14

3.4. Analysis of NIR Spectrum of Phenol NIR spectrum of phenol (Fig. 4 and Table 3) can be roughly divided into three subregions. The bands appearing between 6100–5900 cm−1 arise mainly from CH stretching modes. The lower frequency fragment, 5200–4000 cm−1, is highly characteristic with numerous sharp and intense peaks. These peaks mainly originate from the combinations of ring deformation and CH stretching modes (Fig. 4, Table 3). Note, that due to lower symmetry compared to benzene molecule, the normal modes of phenol ring are different (refer to Supplementary Material 2 for animated presentation of all normal vibrations of phenol molecule). The differences between NIR spectra of aliphatic and aromatic alcohols

Fig. 3. Experimental and calculated NIR spectra of cyclohexanol (0.2 M, CCl4) and proposed band assignments. Band numbers correspond to those presented in Table 2. For better view of details refer to Fig. S2 in Supplementary Material.

2νOH νasCH2 + νOH νas'CH3 + νasCH3 νasCH2 + νas′CH3 2νas′CH3 2νasCH2 2νsCH2 νasCH2 + νasCH2 νasCH2 + νsCH2 νasCH2 + νasCH2 νasCH2 + νsCH2 2νasCH2 δwaggCH2 + νOH [δtwistCH2, δCOH] + νOH δwaggCH2 + νOH, [δtwistCH2, δCOH] + νOH [δtwistCH2, δrockCH2] + νOH [δrockCH2, νCC] + νOH [δtwistCH2, δrockCH2, δrock′CH2] + νOH [νasCH2, νas′CH2] + νas′CH3 [δscissCH2, δas′CH3] + νasCH2 δscissCH2 + νsCH2 [δsCH3, δwaggCH2] + νsCH3 δtwistCH2 + νasCH2 δwaggCH2 + νsCH2 [δrockCH2, δrock'CH2] + νasCH2 [δrockCH2, δrock'CH2] + νsCH2 δtwistCH2 + νsCH2 [δtwistCH2, δCOH] + νasCH2

will be discussed in detail in Section 3.7. Similar to the case of cyclohexanol, the accuracy of calculated wavenumbers of the combinations involving νOH mode is relatively lower than that from the CH modes (Fig. 4, Table 3). A lower agreement between the simulated and experimental spectra in the vicinity of 6000 cm−1 region, particularly noticeable in the Table 2 Experimental and calculated NIR bands of cyclohexanol and proposed band assignments. Band number

Wavenumber/cm−1

Band assignment

Experimental Calculated

Difference

1

7073

–

2νOH

2 3

5873 5803

7045, 6783a 5838 5776

−35 −27

4

5710

5721

11

5

5674

5690

16

6

5650–5380

5660–5500 –

7 8 9 10 11

5011 4951 4884 4849 4752

4919 4855 4794 4763 4668

−92 −96 −90 −86 −84

12 13 14 15 16 17 18 19

4691 4361 4338 4267 4199 4175 4112 4069

4583 4343 4314 4255 4197 4163 4130 4070

−108 −18 −24 −12 −2 −12 18 1

2νasCH2 2νasCH2 νasCH2 + νasCH2 2νasCH2 νCH + νasCH2 νasCH2 + νasCH2 2νasCH2 νasCH2 + νasCH2 νasCH2 + νsCH2 2νasCH2 2νsCH2 2νCH δCH + νOH δCH + νOH δwaggCH2 + νOH δtwistCH2 + νOH [δCOH, δrockCH2, δtwistCH2] + νOH [δrockCH2, δCOH] + νOH νasCH2 + δscissCH2 νasCH2 + δscissCH2 νasCH2 + δscissCH2 δwaggCH2 + νCH [δtwistCH2, δCOH] + νCH [δtwistCH2, δCOH] + νasCH2 δtwistCH2 + νsCH2

a

This discrepancy will be discussed in detail in Section 3.6.

Please cite this article as: K.B. Beć, et al., Spectra-structure correlations in NIR region: Spectroscopic and anharmonic DFT study..., Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2018), https://doi.org/10.1016/j.saa.2018.01.041

K.B. Beć et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2018) xxx–xxx

5

Supplementary Material). The genesis of NIR peaks appearing in the spectra of studied alcohols is in general similar to that of simpler alcohols [28,45]. The relative influence of the overtone bands rapidly decreases toward lower NIR wavenumbers. The 6100–5500 cm−1 region features comparable contributions from the first overtones and binary combinations. However, the binary combinations are the most

Fig. 4. Experimental and calculated NIR spectra of phenol (0.2 M, CCl4) and proposed band assignments. Band numbers correspond to those presented in Table 3. For better view of details refer to Fig. S3 in Supplementary Material.

case of alcohol molecules, is related to the lower quality of prediction of the first overtones of C\\H stretching vibrations. Although further studies would be needed to bring full explanation of this discrepancy, it can be expected that inter-mode anharmonicity is the key factor standing behind it. The fundamental C\\H stretching vibrational levels have similar energy to the combinations of the stretching and bending modes of the same groups; this effect can be followed by comparing the spectra of the three alcohols (Figs. 2–4 and Tables 1–3), for which the positions of νCH modes clearly differ. The resulting degeneracy affects the involved states, as illustrated by us before [44]. The reproduction of these degeneracies is more challenging, particularly for VPT2 computational scheme, thus resulting in relatively lower accuracy of the simulated spectra in the region around 6000 cm−1. 3.5. Spectral Contributions of Overtones, Binary Combinations and Conformational Isomers An analysis of the simulated peaks makes it possible to obtain insight into the origin of NIR spectra (Fig. 5A–C and Figs. S5–S7 in

Table 3 Experimental and calculated NIR bands of phenol and proposed band assignments. Band number Wavenumber/cm−1 Experimental Calculated

Band assignment Difference

1 2

7052 5999

6970 −82 6073,6041 –

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

5940 5208 5079 4948 4782 4677 4645 4617 4551 4384 4309 4226 4131 4061 4050

5906 5164 5038 4844 4726 4663 4639 4598 4546 4391 4288 4244 4088 4064 4044

−34 −44 −41 −104 −56 −14 −6 −19 −5 7 −21 18 −43 3 −6

2νOH 2νCH νCH + νCH 2νCH [νCC, δCH] + νOH [δCH, νCC] + νOH [δCH, δCOH] + νOH [δCOH, νCC] + νOH [νCC, δCH] + νCH [νCC, δCH] + νCH [νCC] + νCH [δCH, νCC] + νCH(ip)a νCC + νCH [νCO, ring δtrigonal] + νCH [νCO, ring δtrigonal] + νCH νCC + νCH(ip)a [ring δtrigonal, νCC] + νCH [ring δtrigonal, νCC] + νCH

a νCH(ip) denotes in-phase CH stretching mode; for the clarity all other νCH modes in the table refer to opposite-phase stretching.

Fig. 5. Contributions of the first overtone and binary combination bands to NIR spectra of (A) n-hexanol, (B) cyclohexanol, (C) phenol according to the results of GVPT2//B3LYP/ SNST calculations. The intensities of calculated bands are in common scale with the final theoretical spectra. For better view of details refer to Figs. S4–S6 in Supplementary Material.

Please cite this article as: K.B. Beć, et al., Spectra-structure correlations in NIR region: Spectroscopic and anharmonic DFT study..., Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2018), https://doi.org/10.1016/j.saa.2018.01.041

6

K.B. Beć et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2018) xxx–xxx

Fig. 6. Contributions of the conformational isomers (narrow lines) into NIR spectra of (A) n-hexanol, (B) cyclohexanol, according to the results of GVPT2//B3LYP/SNST calculations. For better view of details refer to Figs. S8–S9 in Supplementary Material.

important species below 5500 cm−1. As can be seen, this region reveals a large number of overlapped combination modes. Thus, the 5500– 4000 cm−1 region is very characteristic and sensitive to structural differences; it resembles a “fingerprint” region known in IR spectroscopy. The influence of conformational flexibility on the NIR spectrum is shown in Fig. 6a–b. The spectrum of n-hexanol (Fig. 6a) contains broad bands resulting from extensive overlap of the conformational contributions. In the case of cyclohexanol, where there is one major conformer (Fig. 6b), the band separation is higher. This fact remains consistent with the case of phenol (Fig. 4), which only has one stable conformational isomer. The general trend, as one would expect, is that the increase in conformational flexibility increases the experimental bandwidths and decreases the overall “sharpness” of the peaks. Higher number of contributing modes due to higher number of the isomers is the key factor here, although their relative abundances should also be considered. The spectral regions in which a larger number of

contributions appear are affected stronger. For studying subtle effects the most informative cases are cyclohexanol and phenol, as the difference in the number of their relevant conformers is moderate. The lower NIR region (4500–4000 cm−1) is affected the most, as evidenced by the calculated spectra (Fig. 3–4). The case of n-hexanol demonstrates that high conformational flexibility leads to a broadening of the bands in the entire NIR region (Fig. 2). 3.6. Analysis of the 2νOH Band The first overtone band of the OH stretching mode is highly specific in NIR spectra and has been rich source of information about the structure and interactions of alcohols. The position of this band is different for each of the three alcohols. For n-hexanol it is located at 7103 cm−1, for cyclohexanol at 7073 cm−1 and for phenol at 7052 cm−1. The neighbouring of the cyclic structure (and particularly an aromatic

Fig. 7. NIR and second derivative spectra of studied alcohols (0.2 M; CCl4) in the region of the first overtone of the OH stretching band.

Please cite this article as: K.B. Beć, et al., Spectra-structure correlations in NIR region: Spectroscopic and anharmonic DFT study..., Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2018), https://doi.org/10.1016/j.saa.2018.01.041

K.B. Beć et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2018) xxx–xxx

7

Fig. 8. Vibrational potential and vibrational states [B3LYP/6-311G(d,p)] of the OH stretching mode of the main (equatorial-gauche) conformer of cyclohexanol.

ring) shifts the peak position to lower wavenumbers as a result of a lower electron density located at the oxygen due to the ring effect. As it has been demonstrated, the shape of this peak is influenced by the contributions from different conformational isomers [45]. In the case when only one conformer exist, i.e. methanol [28] or tert-butanol [30,45] the shape of the 2νOH is symmetric, if the associated forms (i.e. dimers, trimers, etc.) do not contribute. A possible broadening of the red-tail should be accounted for the higher order overtone and combination bands of weak intensity [45]. Fig. 7 compares the envelopes of the 2νOH band for all studied alcohols. Clearly, the fact that phenol has only one conformer can be recognized by its NIR spectrum. The broadening of the band of n-hexanol (Fig. 7) is due to the presence of two major components located at 7109 and 7081 cm−1 (Δν = 28 cm−1), as evidenced in the second derivative spectrum. This splitting is assumed to originate from frequency grouping of gauche and trans conformers with respect to C\\O bond (OH group) [46]. The 2νOH bandshape of cyclohexanol reveals a similar splitting, but the difference in the relative intensity of both rotational conformers is significantly higher (Fig. 7). Again, this observation remains consistent with previous report [46]. The major component appears at 7074 cm−1 and the minor one at 7047 cm−1 (Δν of 27 cm−1). The presence of two significant equatorial conformers (and two axial ones with negligible contribution) was predicted by the calculations discussed in Sections 3.1 and 3.2. However, VPT2 calculations at B3LYP/SNST level provided a completely wrong picture of the 2νOH band, by predicting a unrealistic position of the leading conformer (Fig. 3).

3.7. Anharmonicity of the OH Stretching Mode and Conformational Analysis of Cyclohexanol The VPT2 calculation at B3LYP/SNST level does not yield reliable position of the 2νOH band. To get more reliable insight into the conformational contributions to the NIR spectrum, and to improve the agreement with the experimental spectrum, we employed a detailed study on the anharmonic potential of OH stretching vibration of the two major conformational isomers of cyclohexanol. A scan of the potential energy over a dense grid along the OH stretching coordinate and subsequent solving of the time-independent Schrödinger equation predicts vibrational levels with an accuracy of less than 0.1 cm−1. Therefore, the only factor affecting the agreement with experimental frequencies results from the accuracy of energy calculations (Fig. 8). This approach captures vastly higher amount of anharmonicity compared to VPT2 calculations at the expense of longer computational time. In order to obtain highly accurate νOH normal coordinate of cyclohexanol, the harmonic analysis was performed at B3LYP/6-311+ +G(3df,3pd) level of theory. These calculations also resulted in corrected order of the conformational stability (Table S2–S3 in Supplementary Material). Previously, the inaccuracy of DVPT2-B3LYP/SNST vibrational frequencies with anharmonic ZPE corrections was propagated into inaccurate Gibbs free energies and thus Boltzmann coefficients. By applying these two improvements we obtained much better agreement between the calculated and experimental 2νOH bands of cyclohexanol. First of all, the frequency of the major conformer (equatorial-gauche) is higher (7092 cm−1) than the following one (equatorial-

Table 4 The comparison of 2νOH vibrational frequencies in [cm−1] of two main conformers of cyclohexanol determined with the use of VPT2 approach and numerical solving of Schrödinger equation based on scanning of the potential energy along the 2νOH normal coordinate.

Equat. gauche Equat. trans Δν

Exp.

Calc. (VPT2//B3LYP/SNST)

Calc. (V(Q) probing//B3LYP/6-311G(d,p))

7074 7047 27

7043 6783 260

7092 7062 30

Please cite this article as: K.B. Beć, et al., Spectra-structure correlations in NIR region: Spectroscopic and anharmonic DFT study..., Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2018), https://doi.org/10.1016/j.saa.2018.01.041

8

K.B. Beć et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2018) xxx–xxx

trans, 7062 cm−1); this fully corresponds to the conclusion drawn from the second derivative spectrum (Fig. 7). Moreover, the calculated splitting (Δν = 30 cm−1) between the conformers is similar to the experimental value (Δν = 27 cm−1); VPT2 calculations completely failed in this case (Δν = 260 cm−1), as evidenced in Table 4. The 2νOH frequencies obtained by solving Schrödinger equation were overestimated by 18 cm−1 and 15 cm−1 for gauche and trans conformers, respectively. However, these values correspond only to 0.25% and 0.21% of the relative error (Table 4).

3.8. Generalized Spectra-Structure Correlations in n-Hexanol, Cyclohexanol and Phenol NIR spectra of three types of alcohols (open-chain aliphatic, cyclic aliphatic and aromatic ones) reflect well the structural features of these molecules. The 2νOH band (7150–7000 cm−1) was analysed in detail in Section 3.5 and its spectral parameters are strongly connected to the molecular structure. First, we elucidate the similarities and differences between n-hexanol and cyclohexanol, and afterwards we discuss the major features that distinct the aliphatic alcohols from the aromatic one. The regions of 5900–5500 cm−1 and 5100–4400 cm−1 of n-hexanol and cyclohexanol look similar (Fig. 2–3). The spectrum of open-chain alcohol shows enhanced band broadening due to high conformational flexibility and the existence of additional combination bands involving the methyl group vibrations. Apart of that, the origin of the respective bands of these two alcohols is similar. The differences in the 4500– 4000 cm−1 region are more pronounced, with much higher band separation in the case of cyclohexanol. Linear hexanol reveals a single major peak at 4336 cm−1 (combination of the stretching and bending modes of CH2 and CH3) followed by a weaker one at 4265 cm−1. This observation is consistent with that reported for butyl alcohols [30]. It seems that uniformity of the ~4336 cm−1 peak is enhanced by the conformational flexibility; the molecules with lower number of isomers tend to exhibit a separation of the major band into two peaks [30]. In the cyclic hexanol these two bands (4338 and 4267 cm−1) are more separated and the latter one is dominant. The 4200–4000 cm−1 region for n-hexanol reveals strong broadening, while that of cyclohexanol features two separated peaks. This results from a significantly lower number of conformational isomers and relatively higher separation of the fundamentals of in-ring CH2 deformation modes in cyclohexanol. The differences between the aliphatic and aromatic alcohols are more evident (Fig. 2–4). Phenol has a strong band at 5999 cm−1, while the aliphatic alcohols possess a broad structure between 5900 cm−1 and ~5500 cm−1. Further, the region of 5300-4000 cm−1 shows distinct differences between the aliphatic and aromatic alcohols. Within the entire region the band separation for phenol is notably higher. Between 5300 and 4750 cm−1 phenol features three strong, well separated peaks (combinations of νOH + νCC/δCH), while aliphatic alcohols reveal weaker and strongly overlapped bands (νOH + δCH2). Again, division of νCC, δCH fundamentals of phenol in IR region can be attributed to this distinct NIR bands separation. Further, the pattern of intensity ratio in the region of 5300–4000 cm−1 is different for aliphatic and aromatic alcohols. Between 5300 and 4500 cm−1 phenol shows medium-to-strong (5300–4900 cm−1) and strong bands (4900–4500 cm−1) followed by a significant decrease of absorbance between 4500 and 4150 cm−1, interrupted by a sharp peak at 4309 cm−1 (νCH + νCO, ring δtrigonal). Afterwards, the absorbance rises again developing a sharp band at 4061 cm−1 (νCH + νCO, ring δtrigonal). Assuming arbitrary sub-regions (in cm−1) of roughly a. 5300–4900; b. 4900–4500; c. 4500–4100; d. 4100–4000; the aliphatic alcohols feature (a-b-c-d): w-vw-vs-m (weak, very weak, very strong, medium) bands, while phenol m-s-w-vs ones. This order of relative intensities allows for reliable distinction between aliphatic and aromatic alcohols.

4. Summary NIR spectra of n-hexanol, cyclohexanol and phenol were accurately reproduced by anharmonic DFT calculations. By taking into account first overtone and binary combination bands originating from individual conformational isomers (a selection of the most relevant population was necessary in the case of n-hexanol) it was possible to achieve very good agreement between the calculated and experimental spectra. Detailed assignments of all significant bands in NIR region from 7500 to 4000 cm−1 were proposed and the spectra-structure correlations were estimated on that basis. VPT2 calculations appear to be very useful for applied spectroscopic studies, as they provide accurate reproduction of entire NIR spectra. However, this approach is somewhat limited in description of vibrational modes strongly deviating from the harmonic oscillator model, i.e. 2νOH vibration in cyclohexanol. This problem was addressed by scanning of the vibrational potential over a dense-grid and subsequent solving of the time-independent Schrödinger equation; the applied approach captures much higher amount of nahamronicity. This way, not only the details of the 2νOH band were fully elucidated, but also the predicted equilibrium between different conformations was in line with the one obtained from the experimental spectrum of cyclohexanol. The studied molecules can be considered model structures for important biomolecules, in which an OH group is attached to various kinds of chemical environments. Therefore, the obtained correlations between the structural features and the corresponding NIR spectra should provide good basis for better understanding of the spectral details for a wide range of compounds. Acknowledgement Calculations have been carried out in Wrocław Centre for Networking and Supercomputing (http://www.wcss.pl), under grant no. 375. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.saa.2018.01.041. References [1] H.W. Siesler, Y. Ozaki, S. Kawata, H.M. Heise (Eds.), Near-infrared Spectroscopy, Wiley-VCH, Weinheim, 2002. [2] Y. Ozaki, W.F. McClure, A.A. Christy (Eds.), Near-infrared Spectroscopy in Food Science and Technology, Wiley-Interscience, Hoboken, NJ, USA, 2007. [3] M.A. Czarnecki, Y. Morisawa, Y. Futami, Y. Ozaki, Advances in molecular structure and interaction studies using near-infrared spectroscopy, Chem. Rev. 115 (2015) 9707–9744. [4] C.W. Huck, Infrared Spectroscopy in Near-infrared/Infrared Bioanalysis Including Imaging, John Wiley & Sons, Encyclopedia of Analytical Chemistry, 2016. [5] Y. Futami, Y. Ozaki, Y. Hamada, M.J. Wójcik, Y. Ozaki, Frequencies and absorption intensities of fundamentals and overtones of NH stretching vibrations of pyrrole and pyrrole–pyridine complex studied by near-infrared/infrared spectroscopy and density-functional-theory calculations, Chem. Phys. Lett. 482 (2009) 320–324. [6] S. Šašić, V.H. Segtnan, Y. Ozaki, Self-modeling curve resolution study of temperature-dependent near-infrared spectra of water and the investigation of water structure, J. Phys. Chem. A 106 (2002) 760–766. [7] L. Stordrange, A.A. Christy, O.M. Kvalheim, H. Shen, Y.-Z. Liang, Study of the self-association of alcohols by near-infrared spectroscopy and multivariate 2D techniques, J. Phys. Chem. A 10 (2002) 8543–8553. [8] R. Iwamoto, T. Matsuda, H. Kusanagi, Contrast effect of hydrogen bonding on the acceptor and donor OH groups of intramolecularly hydrogen-bonded OH pairs in diols, Spectrochim. Acta A 62 (2005) 97–104. [9] M.A. Czarnecki, Near-infrared spectroscopic study of hydrogen bonding in chiral and racemic octan-2-ol, J. Phys. Chem. A 107 (2003) 1941–1944. [10] K.B. Beć, Y. Futami, M.J. Wójcik, T. Nakajima, Y. Ozaki, Spectroscopic and computational study of acetic acid and its cyclic dimer in the near-infrared region, J. Phys. Chem. A 120 (2016) 6170–6183. [11] T. Gonjo, Y. Futami, Y. Morisawa, M.J. Wojcik, Y. Ozaki, Hydrogen bonding effects on the wavenumbers and absorption intensities of the OH fundamental and the first, second, and third overtones of phenol and 2,6-dihalogenated phenols studied by visible/near-infrared/infrared spectroscopy, J. Phys. Chem. A 115 (2011) 9845–9853.

Please cite this article as: K.B. Beć, et al., Spectra-structure correlations in NIR region: Spectroscopic and anharmonic DFT study..., Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2018), https://doi.org/10.1016/j.saa.2018.01.041

K.B. Beć et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2018) xxx–xxx [12] Y. Hu, J. Zhang, H. Sato, Y. Futami, I. Noda, Y. Ozaki, C\ \H⋯OC hydrogen bonding and isothermal crystallization kinetics of poly(3-hydroxybutyrate) investigated by nearinfrared spectroscopy, Macromolecules 39 (2006) 3841–3847. [13] R. Iwamoto, T. Matsuda, Characterization of infrared and near-infrared absorptions of free alcoholic OH groups in hydrocarbon, Appl. Spectrosc. 58 (2004) 1001–1009. [14] A. Ikehata, C. Hashimoto, Y. Mikami, Y. Ozaki, Thermal phase behavior of triethylamine–water mixtures studied by near-infrared spectroscopy: band shift of the first overtone of the C\ \H stretching modes and the phase diagram, Chem. Phys. Lett. 393 (2004) 403–408. [15] W. Wrzeszcz, P. Tomza, M. Kwaśniewicz, S. Mazurek, R. Szostak, M.A. Czarnecki, Microheterogeneity in binary mixtures of methanol with aliphatic alcohols: ATR-IR/ NIR spectroscopic, chemometrics and DFT studies, RSC Adv. 6 (2016) 37195–37202. [16] Y. Futami, Y. Ozaki, Y. Hamada, M.J. Wójcik, Y. Ozaki, Solvent dependence of absorption intensities and wavenumbers of the fundamental and first overtone of NH stretching vibration of pyrrole studied by near-infrared/infrared spectroscopy and DFT calculations, J. Phys. Chem. A 115 (2011) 1194–1198. [17] Y. Chen, Y. Morisawa, Y. Futami, M.A. Czarnecki, H.-S. Wang, Y. Ozaki, Combined IR/ NIR and density functional theory calculations analysis of the solvent effects on frequencies and intensities of the fundamental and overtones of the C_O stretching vibrations of acetone and 2-hexanone, J. Phys. Chem. A 118 (2014) 2576–2583. [18] C.G. Kirchler, C.K. Pezzei, K.B. Beć, S. Mayr, Y. Ozaki, C.W. Huck, Critical evaluation of spectral information of benchtop vs. portable near-infrared spectrometers: quantum chemistry and two-dimensional correlation spectroscopy for a better understanding of PLS regression models of the rosmarinic acid content in Rosmarini folium, Analyst 142 (2017) 455–464. [19] B. Czarnik-Matusewicz, K. Murayama, R. Tsenkova, Y. Ozaki, Analysis of near-infrared of spectra of complicated biological fluids by two-dimensional correlation spectroscopy: protein and fat concentration-dependent spectral changes of milk, Appl. Spectrosc. 53 (1999) 1582–1594. [20] D. Ishikawa, T. Genkawa, K. Murayama, M. Komiyama, Y. Ozaki, Feasibility study of diffuse reflectance and transmittance near infrared spectroscopy for rapid analysis of ascorbic acid concentration in bilayer tablets using a high-speed polychromator-type spectrometer, J. Near Infrared Spectrosc. 22 (2014) 189–197. [21] C.W. Huck, Y. Ozaki, V.A. Huck-Pezzei, Critical review upon the role and potential of fluorescence and near-infrared imaging and absorption spectroscopy in cancer related cells, serum, saliva, urine and tissue analysis, Curr. Med. Chem. 23 (2016) 3052–3077. [22] V. Wiedemair, S. Mayr, D.S. Wimmer, E.M. Köck, S. Penner, A. Kerstan, P.-A. Steinmassl, H. Dumfahrt, C.W. Huck, Novel molecular spectroscopic multimethod approach for monitoring water absorption/desorption kinetics of CAD/CAM poly (methyl methacrylate) prosthodontics, Appl. Spectrosc. 0 (2016) 1–13. [23] C.W. Huck, K.B. Beć, J. Grabska, Y. Ozaki, Quantum chemical calculation of NIR spectra of practical materials, NIR News 28 (1) (2017) 13–20. [24] V. Barone (Ed.), Computational Strategies for Spectroscopy: From Small Molecules to Nano Systems, Wiley, 2011. [25] K.B. Beć, M.J. Wójcik, T. Nakajima, Quantum chemical calculations of basic molecules: alcohols and carboxylic acids, NIR News 27 (2016) 15–21. [26] L.G. Weyer, S.-C. Lo, Spectra structure correlation in the near infrared, in: J.M. Chalmers, P.R. Griffiths (Eds.), Handbook of Vibrational Spectroscopy, John Wiley & Sons, 2002. [27] V. Barone, Anharmonic vibrational properties by a fully automated second-order perturbative approach, J. Chem. Phys. 122 (14108) (2005) 1–10. [28] K.B. Beć, Y. Futami, M.J. Wójcik, Y. Ozaki, A spectroscopic and theoretical study in the near-infrared region of low concentration aliphatic alcohols, Phys. Chem. Chem. Phys. 18 (2016) 13666–13682. [29] J. Bloino, A. Baiardi, M. Biczysko, Aiming at an accurate prediction of vibrational and electronic spectra for medium-to-large molecules: an overview, Int. J. Quantum Chem. 116 (2016) 1543−1574. [30] J. Grabska, K.B. Beć, Y. Ozaki, C.W. Huck, Temperature drift of conformational equilibria of butyl alcohols studied by near-infrared spectroscopy and fully anharmonic DFT, J. Phys. Chem. A 121 (2017) 1950–1961. [31] J. Grabska, M. Ishigaki, K.B. Beć, M.J. Wójcik, Y. Ozaki, Correlations between structure and near-infrared spectra of saturated and unsaturated carboxylic acids. An insight from anharmonic DFT calculations, J. Phys. Chem. A 121 (2017) 3437−3451. [32] J. Grabska, K.B. Beć, M. Ishigaki, M.J. Wójcik, Y. Ozaki, Spectra-structure correlations of saturated and unsaturated medium-chain fatty acids. Near-infrared and anharmonic DFT study of hexanoic acid and sorbic acid, Spectrochim. Acta A 185 (2017) 35–44. [33] X. Chen, Q. Wang, J. Shen, H. Pan, T. Wu, Adsorption of leucine-rich amelogenin protein on hydroxyapatite (001) surface through\ \COO\ \ claws, J. Phys. Chem. C 111 (2007) 1284–1290. [34] Z.-S. Zhang, H.-H. Pan, R.-K. Tang, Molecular dynamics simulations of the adsorption of amino acids on the hydroxyapatite {100}-water interface, Front. Mater. Sci. Chin. 2 (2008) 239–245. [35] D.B. Norris, P.W. Trudgill, The metabolism of cyclohexanol by Nocardia globerula CL1, Biochem. J. 121 (1971) 363–370. [36] W. Dangel, A. Tschech, G. Fuchs, Anaerobic metabolism of cyclohexanol by denitrifying bacteria, Arch. Microbiol. 150 (1988) 358–362.

9

[37] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, et al., Gaussian 16, Revision A.03, Gaussian, Inc., Wallingford, CT, 2016. [38] V. Barone, M. Cossi, Quantum calculation of molecular energies and energy gradients in solution by a conductor solvent model, J. Phys. Chem. A 102 (1998) 1995–2001. [39] S. Grimme, S. Ehrlich, L. Goerigk, Effect of the damping function in dispersion corrected density functional theory, J. Comput. Chem. 32 (2011) 1456–1465. [40] J.M.L. Martin, V.C. Alsenoy, GAR2PED, University of Antwerp, 1995. [41] P. Pulay, G. Fogarasi, F. Pang, J.E. Boggs, Systematic ab initio gradient calculation of molecular geometries, force constants, and dipole moment derivatives, J. Am. Chem. Soc. 101 (1979) 2550–2560. [42] M.J. Schuler, T.S. Hofer, C.W. Huck, Assessing the predictability of anharmonic vibrational modes at the example of hydroxyl groups — ad hoc construction of localised modes and the influence of structural solute–solvent motifs, Phys. Chem. Chem. Phys. 19 (2017) 11990–12001. [43] M. Kwaśniewicz, M.A. Czarnecki, The effect of the chain length on mid-infrared and near-infrared spectra of aliphatic 1-alcohols, Appl. Spectrosc. (2017)https://doi.org/ 10.1177/0003702817732253. [44] K.B. Beć, J. Grabska, Y. Ozaki, J.P. Hawranek, C.W. Huck, Influence of non-fundamental modes on mid-infrared spectra. Anharmonic DFT study of aliphatic ethers, J. Phys. Chem. A 121 (2017) 1412–1424. [45] J. Grabska, M.A. Czarnecki, K.B. Beć, Y. Ozaki, Spectroscopic and quantum mechanical calculation study of the effect of isotopic substitution on NIR spectra of methanol, J. Phys. Chem. A 121 (2017) 7925–7936. [46] M.A. Czarnecki, D. Wojtków, K. Haufa, Rotational isomerism of butanols: infrared, near-infrared and DFT study, Chem. Phys. Lett. 431 (2006) 294–299. Krzysztof B. Beć obtained his PhD degree (2014) in Physical and Theoretical Chemistry from University of Wrocław, Poland. His research involved thin-film infrared spectroscopy, optical constants and computational spectroscopy. He joined Professor Yukihiro Ozaki research team as a postdoctoral fellow at Kwansei Gakuin University, Japan (2015) where he was involved in developing applications of anharmonic methods in simulation of near-infrared spectra. He continued his work in basic and applied NIR spectroscopy in Prof. Christian W. Huck team at University of Innsbruck, Austria (2016). Afterwards, he re-joined Professor Yukihiro Ozaki team at Kwansei Gakuin University as an assistant professor, where he is currently involved in development of ATR-FUV spectroscopy and its application in studies of polymer-carbon nanostructure composites.

Justyna Grabska obtained PhD degree (2015) in Physical and Theoretical Chemistry from University of Wrocław, Poland. Her work is focused on vibrational spectroscopy, high-frequency dielectric function and scientific programming. She started her research in the field of near-infrared spectroscopy after joining Prof. Christian W. Huck team as a postdoctoral fellow at University of Innsbruck, Austria (2016). Her current research as a postdoctoral researcher in Professor Yukihiro Ozaki group focuses on computational NIR spectroscopy.

Mirosław Czarnecki obtained his M. Sc. (1981) and Ph.D. (1989) degrees in Physical and Theoretical Chemistry from the University of Wrocław. He was a postdoctoral fellow at Kwansei Gakuin University, Japan (1992–1993) and University of Essen, Germany (1995). At present, he is a professor at the Faculty of Chemistry, University of Wrocław, Poland. His current research involve hydrogen bonding, molecular structure, microheterogeneity in binary liquids, liquid crystals, MIR, NIR and Raman spectroscopy, computer aided spectroscopy, 2D correlation analysis, chemometrics and theoretical calculations.

Please cite this article as: K.B. Beć, et al., Spectra-structure correlations in NIR region: Spectroscopic and anharmonic DFT study..., Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2018), https://doi.org/10.1016/j.saa.2018.01.041