Chain length effects on the vibrational structure and molecular interactions in the liquid normal alkyl alcohols

Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 189 (2018) 57–65

Contents lists available at ScienceDirect

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

Chain length effects on the vibrational structure and molecular interactions in the liquid normal alkyl alcohols Johannes Kiefer a,b,c,d,⁎, Sabine Wagenfeld a, Daniela Kerlé a,b a

Technische Thermodynamik, Universität Bremen, Badgasteiner Str. 1, 28359 Bremen, Germany MAPEX Center of Materials and Processes, Universität Bremen, Germany School of Engineering, University of Aberdeen, Fraser Noble Building, Aberdeen AB24 3UE, UK d Erlangen Graduate School in Advanced Optical Technologies (SAOT), Universität Erlangen-Nürnberg, Germany b c

a r t i c l e

i n f o

Article history: Received 10 April 2017 Received in revised form 3 July 2017 Accepted 30 July 2017 Available online 31 July 2017 Keywords: Alkan-1-ols n-Alkanols FTIR Vibrational structure Hydrogen bonding

a b s t r a c t Alkyl alcohols are widely used in academia, industry, and our everyday lives, e.g. as cleaning agents and solvents. Vibrational spectroscopy is commonly used to identify and quantify these compounds, but also to study their structure and behavior. However, a comprehensive investigation and comparison of all normal alkanols that are liquid at room temperature has not been performed, surprisingly. This study aims at bridging this gap with a combined experimental and computational effort. For this purpose, the alkyl alcohols from methanol to undecan-1-ol have been analyzed using infrared and Raman spectroscopy. A detailed assignment of the individual peaks is presented and the influence of the alkyl chain length on the hydrogen bonding network is discussed. A 2D vibrational mapping allows a straightforward visualization of the effects. The conclusions drawn from the experimental data are backed up with results from Monte Carlo simulations using the simulation package Cassandra. © 2017 Elsevier B.V. All rights reserved.

1. Introduction The normal alcohols are a highly important class of chemicals. Their representatives find numerous applications in all industrial areas and in our everyday lives. For example, alcohols are used as reactants and solvents in the chemical industry [1–4], as (bio)fuels [5–8], as cleaning agents [9,10], and as germicides in the medical and health sector [11, 12]. In fundamental research, alcohols are commonly used as reference materials, e.g. to study molecular interactions [13,14] and phase behavior [15,16]. In this context, the thermophysical and chemical properties of alcohols are beneficial. Changing the length of the alkyl chain in alkanols allows a systematic variation of these properties as can be seen in Table 1. At the same time, the general chemical nature remains the same in a sense that all alkanols have a polar hydroxyl group to allow dipole-dipole interactions and hydrogen bonding, and a nonpolar alkyl chain. Consequently, at room temperature the normal alkanols span a wide range of the dielectric constant, see Table 1. In the past two decades, spectroscopic methods entered new fields including process monitoring [17,18], environmental monitoring [19, 20], material and fluid characterization [21,22], medicine [23,24], and security [25,26]. As alcohols find applications in all these areas as well,

it is not surprising that the common n-alcohols have been extensively studied by all kinds of spectroscopy [27–34]. However, a look at the literature reveals that the efforts in a systematic investigation into the vibrational spectroscopy of n-alkanols are very limited. Most studies analyzed the individual short chain compounds methanol [28,35–39], ethanol [28,40], propanol [28], and butanol [29]. Some studies covered higher alkanols [30,31] or made comparisons between two or more different alcohol compounds [41–50]. A comparative investigation of all n-alkanols that are liquid at room temperature, i.e. from methanol (C1) to undecan-1-ol (C11), has not been performed to the best of our knowledge. In the present study, we aim at closing this gap in the literature by performing a detailed vibrational analysis of all liquid n-alcohols. For this purpose, the experimental Raman and Infrared (IR) spectra are presented, analyzed, and interpreted. Conclusions on the chemical structure and intermolecular interactions are drawn and backed up by Monte Carlo simulations using the simulation package Cassandra [51]. The paper is organized as follows: the next section introduces the experimental and computational methods; Section 3 presents the results and discusses them; Section 4 concludes. 2. Materials and Methods

⁎ Corresponding author at: Technische Thermodynamik, Universität Bremen, Badgasteiner Str. 1, 28359 Bremen, Germany. E-mail address: [email protected] (J. Kiefer).

http://dx.doi.org/10.1016/j.saa.2017.07.061 1386-1425/© 2017 Elsevier B.V. All rights reserved.

All chemicals had a purity of N 99.5% and were used as received. The water content was determined by Karl-Fischer titration and the

58

J. Kiefer et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 189 (2018) 57–65

Table 1 Properties of the n-alkanols. The density and relative permittivity values are given for a temperature of 293.2 K. M = molar mass, ρ = density, Tm = melting temperature at atmospheric pressure, Tb = boiling temperature at atmospheric pressure, Tc = critical temperature, pc = critical pressure, εr = relative permittivity (dielectric constant). M

C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11

ρa,b

g/mol

kg/dm

32.04 46.07 60.10 74.12 88.15 102.18 116.20 130.23 144.26 158.28 172.31

0.79 0.79 0.80 0.81 0.81 0.82 0.82 0.83 0.83 0.83 0.83

3

Tmb

Tbb

Tcb

pcb

εra,c

K

K

K

bar

–

175.6 158.8 146.7 188.0 194.7 226.0 238.3 257.0 268.0d 279.6 288.5

337.8 351.5 370.3 390.6 411 430 448 468 485 505 516e

513 514 536.9 562 580 610.5 633 655 672 690 704

81 63 52 45 39 34.2 30.6 27.0 25.3 23.2 21.5

33 25.3 20.8 17.8 15.3 13.0 11.8 10.3 8.8 7.9 6.6f

a

At T = 293.2 K. Source: NIST Standard Reference Database Number 69. Source: CRC Handbook of Chemistry and Physics, 95th Edition. d Source: Handbook of Data on Organic Compounds. Volume I. 3rd ed. CRC Press, Inc. Boca Raton, FL. 1994, p. V4: 3651. e Source: Entry for CAS 112-42-5 in GESTIS-Stoffdatenbank of the IFA. f Estimated from empirical equation of lower alcohols based on the value 5.98 given for 313.2 K. b c

Table 2 Water content determined by Karl-Fischer titration. C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11

348.9 ppm 255.7 ppm 404.7 ppm 603.8 ppm 386.8 ppm 3254.3 ppm 3324.1 ppm 528.9 ppm 5768.0 ppm 331.8 ppm 874.6 ppm

individual values are given in Table 2. All experiments were carried out at room-temperature (296 K). The infrared spectra were recorded on an Agilent Cary 630 instrument equipped with a ZnSe ATR module (5 reflections at 45°). The spectra were recorded in the range 650–4000 cm− 1 with a nominal resolution of 2 cm− 1. 32 scans were averaged in order to obtain an

Fig. 2. (a) Infrared and (b) Raman spectra of the eleven n-alcohols. The IR spectra are normalized with respect to the OH stretching band. The Raman spectra are normalized with respect to the strongest peak.

Fig. 1. View into a simulation box of (a) methanol and (b) n-propanol. The yellow lines represent hydrogen bonds.

J. Kiefer et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 189 (2018) 57–65

appropriate signal to noise ratio. The raw data are provided as a data set in the supplementary material available online. The Raman spectra in the range 650–4000 cm−1 were recorded with a custom-made experimental setup using a continuous-wave frequency-doubled Nd:YAG laser (532 nm, 80 mW). The scattered light was collected and collimated with a 100-mm achromatic lens in direction perpendicular to the laser beam. Elastically scattered light was blocked with a dielectric long-pass filter (cut-off wavelength 538 nm). A 250-mm achromatic lens focused the signal onto the slit of an imaging spectrograph (Princeton Instruments, 300 mm focal length, 600 grooves/mm grating). The spectrally dispersed signals were detected with an intensified CCD camera (Andor). The acquisition time was 1 s and 10 spectra were averaged. The spectral resolution was ~2 cm−1. The Raman spectrum in the range 100–3700 cm− 1 was recorded with an Avantes instrument (AvaSpec-ULS2048L) equipped with a 785 nm diode laser and a fiber-optical Raman probe (backscattering). The acquisition time was 1 s and 10 spectra were averaged. In the later analysis, we only use the range 100–700 cm−1 from these spectra as the reduced sensitivity of the detector at wavelengths beyond 1 μm results in a weak CH and OH stretching region. The full spectra, however, are provided as a data set in the supplementary material available online. To provide insights at the molecular level, Monte Carlo simulations were performed with the simulation package Cassandra [51] using the

59

well-established TraPPE-UA parameters [52]. The simulations were performed on some selected alkanol systems at T = 298 K and p = 1 bar. As an example, Fig. 1 shows the view into a simulation box of (a) methanol and (b) n-propanol. The CH3 and CH2 groups were simulated as one interaction center (united atom) and therefore they are shown as spheres. The hydrogen bonds, which were defined by geometric criteria are illustrated as yellow lines. 3. Results and Discussion In this section, we start with an overview of the vibrational spectra. The subsequent subsections will then provide a detailed discussion of the different characteristic spectral regions. The last subsection (Section 3.5) will eventually present the spectra again, but in the form of two-dimensional contour plots that provide further insights into the effects of changing the alkyl chain length. 3.1. Overview Fig. 2 displays the IR and Raman spectra over the spectral range 650– 3750 cm−1. The IR spectra (Fig. 2a) are normalized with respect to the OH stretching band. This representation is useful as all alcohols carry one OH group and hence the normalized spectra allow observing the increasing CH band with incrementing alkyl chain length. The Raman

Table 3 Observed Infrared peak positions in cm−1 and tentative vibrational assignments; str. = stretching, def. = deformation, sciss. = scissoring, rock. = rocking, wagg. = wagging, tors. = torsion, s = symmetric, as = antisymmetric. C1

C2

C3

C4

C5

C6

C7

C8

C9

C10

C11

Assignment

3325 2977s 2946

3323 2975

3322 2964

3325 2961

3324 2959

3327 2959

3327 2959

3328 2958

3331 2960

3330 2958

3330 2958

2930 2900s 2882

2938

2934

2931

2930

2876

2867

2873 2862

2923 2904s 2874 2855

2874s 2856

2924 2906s 2872s 2855

2744

2745

2741

2741

2748

2925 2904s 2873s 2860s 2856 2745

2924

2878

2929 2925 2875s 2862s 2858 2742

1926 1482s 1456

1940 1457

1463

1468 1458s

1469 1461

1468 1459

1419 1380

1384

1435 1380

1434s 1380

1380

1436s 1379

1482s 1468 1456s 1438s 1379

1338 1294

1343 1294 1275

1342 1301 1263 1244 1216

1467 1456s 1438s 1379 1354 1343 1306

1236

1344 1302 1277 1257 1224

1467 1458s 1438s 1379 1353 1343 1306

1200

1190

1182

1116 1076

1118 1071s

OH str. CH3 as str. CH3 as str. CH3 s str. CH2 s str., FR CH2 s str. Overtone Overtone Combination Combination Combination Overtone Overtone CH3 as def. CH3 as def. CH3 as def., CH2 sciss. CH2 sciss. CH3 s def. CH2 wagg. COH def. CH2 wagg., COH def. CH2 wagg., COH def. CH3 rock., CH2 rock. CH3 rock., CH2 rock. CH2 tors. CH2 rock., CH2 tors. CO str., CH2 rock. CO str., CH2 rock. CO str., CC str.

1054 1036s 1017s 1005 981

1057 1034 1020

2919s 2833

2595 2526 2229 2045 1449 1415

1329 1275

1347 1295 1272 1236

1144s 1088

1125s 1100 1069

1046

1054

881 803

906 888 858

1134s 1114 1072 1061 1045 1039s 1029 1011 992 964s 952 901 880 847

(526)

755 (583)

738 (590)

1114

1021 1017 967

(562)

1251 1215

910s 889 841 784 730 (621)

1347 1304 1275 1252 1235 1209 1176

1198 1173

1118

1120 1072s

1121 1072s

1122 1073s

1123 1071s

1056 1036

1057 1045s 1029

1056 1037

1057 1041s

1057 1040s

1008

1010s 987s

1014s

973 929s 912

932

946

874

877

722 (661)

721 (654)

1014 992 964 921 892

1343 1308

990 968 938 905 867

761 726 (562)

1467 1457s 1438s 1379

724 (641)

956 926 890 818 768 723 (650)

890

721 (658)

CO str., CC str. CC str. CO str., CH2 tors. CC str. CO str., CC str. CO str., CC str. CO str., CC str., CH3 rock. CC str., CH3 rock. CC str. CH3 rock., CH2 rock. CH3 rock., CH2 rock. CH2 rock. COH tors.

60

J. Kiefer et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 189 (2018) 57–65

spectra in Fig. 2b are normalized with respect to the strongest signal peak, which is a CH stretching mode for all the alkanols under investigation. Due to the high scattering cross section of the CH stretching and the low cross section of the OH vibrations, normalization with respect to the OH band is not sensible here. Generally speaking, between 3100 and 3600 cm−1, the characteristically broad band of the OH stretching vibrations can be found. The CH stretching modes are located in the range 2700–3100 cm−1. Below 1600 cm−1, the fingerprint region can be found. It contains a multitude of peaks originating from deformation (bending), scissoring, wagging, rocking, and torsional modes of all molecular moieties. In addition, the CC and CO stretching modes contribute to this part of the spectrum. Therefore, the analysis and interpretation of the fingerprint region is not straightforward. An overview of the vibrational assignments is given in Tables 3 and 4 for the IR and Raman spectra, respectively. 3.2. OH Stretching Region Fig. 3 shows the enlarged OH stretching bands in the IR (panel a) and the Raman (panel b) spectra. The broad OH peak is frequently utilized as a marker for intermolecular interactions [30,31,35,42, 43] as the covalent bond between oxygen and hydrogen is significantly polarized. Hence, the OH group can act as donor and acceptor for hydrogen bonds. Such interactions are accompanied by a charge transfer between the participating molecules and lead to alterations

in the strengths of the covalent bonds. A consequence of this modification is that the vibrational properties change [53]. When the bond becomes stronger, it becomes more rigid and thus the vibrational frequency will increase leading to a blue-shift of the corresponding peak in the IR or Raman spectrum. On the other hand, a reduction of the bond strength will result in a red-shift. The latter is commonly observed as a result of hydrogen bonding and has been included in the IUPAC list of criteria for hydrogen bond identification [54]. The hydrogen bonding interactions in alcohols give rise to the broad OH stretching band. The molecules in the liquid alcohol can be in a number of different bonding states. The most common ones are illustrated in Fig. 4a using a snapshot of the Monte Carlo simulation of propan-1-ol as an example. Molecule A acts as hydrogen bond acceptor, B acts a hydrogen bond donor, C is hydrogen bond acceptor and donor at the same time, and D does not participate in hydrogen bonding interactions. Even the formation of ring structures is possible as shown in Fig. 4b. Moreover, the Monte Carlo simulation allows the visualization of hydrogen bonding via the spatial distribution functions of the relevant atoms around a given molecule. As an example, Fig. 5 illustrates the spatial distribution functions of the oxygen and hydrogen atoms around a methanol molecule at 298 K indicating the local structure resulting from hydrogen bonding interactions. The highest occurrence density near the polar head group is observed for all alcohols with the typical double cone form.

Table 4 Observed Raman peak positions in cm−1 and tentative vibrational assignments; str. = stretching, def. = deformation, sciss. = scissoring, rock. = rocking, wagg. = wagging, tors. = torsion, s = symmetric, as = antisymmetric. C1

C2

C3

C4

C5

C6

C7

C8

C9

C10

C11

Assignment

3363 3279 2985s

3647s 3357 3266 2974

3647s 3349 3265

3647 3348 3255

3647 3348 3255

3647 3348 3255

3647 3348 3255

3647 3348 3255

3647 3348 3255

3647 3348 3255

3647 3348 3255

2944

2929

2961 2934 2915

2959 2935 2908

2835

2880

2878 2737 2675 1473s 1450

2957 2933s 2915s 2905 2875 2860s 2731

2957 2930s 2913 2905 2877 2862 2731

2957 2930s 2911s 2903 2875 2860 2730

2957 2931s 2915s 2903 2873 2858 2728

2957 2930s 2915s 2903 2880 2858 2728

2957 2930s 2913s 2903 2875 2857 2726

1473s 1453

2748 2717 1474s 1449

2876 2843s 2736 (2721s) 1471s 1445

2958 2934 2913 2905 2875 2862s 2731 (2674) 1473s 1442

1474s 1452 1439

1473s 1451 1439

1340 1292 1272 1237s

1295 1248 1216

1298 1265s 1221 1190 1135s 1116 1072 1058s 1023 1000s

1298 1261s 1210 1179 1135s 1118 1074 1060s 1035 1012s 963 923s

1473s 1450 1439 1366 1298 1273s 1250s 1173 1133s 1118 1074 1058s 1027s 1010 959 919

1473s 1450 1438 1364 1298 1271s 1238s 1167 1130s 1119 1076 1062s 1035s 1006s 970 915s

1473s 1448 1437 1364 1298 1273s

1473s 1447 1437 1363 1298 1277s

1165 1133s 1120 1076 1062s 1037s 1010

1160 1132 1120 1078 1060s 1037s 1010s

915s

915s

886 871s 836 795s

451 400 369

886s 872 852s 814 769 530s 483 461 401 346

524s 480s 453 406 361

888s 869 848 814s 777s 527s 486 447 403 336

887s 871 846s 829s 797s 534s 494 450 403 352

286

269

255

243

240 212s

free OH str. OH str. OH str. CH3 as str. CH2 as str. CH3 s str. CH2 s str. CH2 s str. CH2 s str. Overtone Combination Combination CH2 sciss., CH3 def. CH2 sciss. CH2 sciss. CH2 wagg. CH2 wagg., COH def. CH2 wagg., COH def. CH2 rock., CH3 rock. CH2 rock. CO str., CH2 rock. CH3 rock. CO str., CC str. CC str. CO str., CH2 tors. CC str. CC str., CH3 rock., CO str. CC str., CH3 rock. CC str., CH3 rock. CC str. CH3 rock., CH2 rock. CH3 rock., CH2 rock. CC str. CH2 rock. CCO def., CCC def. CCO def., CCC def. CCO def., CCC def. CCC def. CCO def., CCC def. CCC def. CO tors. CO tors.

1272 1147s 1107

1119s 1091 1048

1125s 1100 1064s 1054

1031 1016s 966

880

884 856

1131s 1110 1066 1051s 1028 1006s 961 948s 896 877 844s 825 808s

767 465 433

515 488 453 398 355

326 430s

260s

272

1297 1273s 1232 1199 1133s 1113 1072 1058s 1021s 1010 981 910s 886 861 838 774 505 436 401 366 329

916 887 861 816 759 511 453 408 362 315 287s

892 872 837 790 510

J. Kiefer et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 189 (2018) 57–65

Fig. 3. Enlarged (a) Infrared and (b) Raman spectra of the eleven n-alcohols in the OH stretching region. The black arrows highlight the systematic change in absorbance/ intensity with increasing aliphatic chain length. The red arrow in (b) indicates the vibrational band of free OH groups.

The molecules in Fig. 4a represent four different bonding states and thus they would contribute sub-peaks to the overall OH band. When a molecule is not taking part in hydrogen bonding, the OH bond will exhibit the highest possible strength and thus the highest vibrational frequency. The band of such “free” OH groups is commonly observed in alcohol vapors due to the very limited hydrogen bonding in the gas phase [41], but also the liquid spectra of alcohols contain signatures of free OH groups [40,44]. The liquid IR spectra in Fig. 3a do not show any signs of such OH groups, but the Raman spectra in panel b do. This observation is in perfect agreement with the work of Eysel and Bertie [48]. From the IR spectra, they concluded that the fraction of non-hydrogen-bonded OH groups is very small. The free OH peak in the Raman spectra was explained by the fact that the scattering activity is not affected by hydrogen bonding in contrast to the IR activity making the detection via Raman spectroscopy more sensitive [48–50]. The corresponding Raman band at 3647 cm−1 is marked by a red arrow Fig. 3. This band is clearly visible for the alcohols C2–C11. In the methanol spectrum, there is a weak shoulder band at the same spectral position, which may indicate a small number of free OH groups. It must be kept in mind that the methanol molecule is highly polar and relatively small. Therefore, in the liquid state, only a few molecules will be present without forming hydrogen bonds to neighboring molecules. The period such a situation without a hydrogen bond lasts will be short. This is similar to liquid water, where a weak shoulder band in the IR spectrum suggests the presence of very weakly bonded OH groups as well [55]. By contrast, when the alkyl chain in the alcohol becomes longer, the molecules will form polar and nonpolar domains in the liquid. In such a micro-structured environment, the probability of having

61

Fig. 4. Snapshots of the main configurations of OH groups in a liquid alcohol using npropanol as an example. Panel (a); A: hydrogen bond acceptor, B: hydrogen bond donor, C: hydrogen bond acceptor and donor, D: free OH. Panel (b) illustrates a ring structure formed by four propanol molecules.

individual OH groups surrounded by aliphatic chains is higher and hence the observation of a distinct peak at high wavenumber in the Raman spectrum. The ratio between this free OH band and the overall OH stretching increases with alkyl chain length. However, it should be noted that the free OH groups are still influenced by the molecular environment. Such cooperative effects commonly play a role in alcohols, see for example Ohno et al. [56]. This means that the neighboring molecules affect the hydrogen bond formation and hence may influence the resulting vibrational frequency.

Fig. 5. Spatial distribution function of the oxygen (red) and hydrogen (white) atoms around a methanol molecule at 298 K from the Monte Carlo simulations. The same molecule is shown from two perspectives for clarity.

62

J. Kiefer et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 189 (2018) 57–65

To emphasize this, the spectra of the liquids can be compared to the corresponding gas phase spectra (the gas phase data were taken from the NIST Standard Reference Database [57]). For example, the OH peak in the IR spectrum of gaseous heptanol is observed at 3670 cm− 1. This means it is blue-shifted by about 23 cm−1 compared to the liquid due to the neighboring molecules. 3.3. CH Stretching Region The CH stretching region is typically the spectral range 2700– 3100 cm−1, see Fig. 6. Like the OH stretching region, it is frequently utilized as an indicator of the molecular behavior [36–38,45,46,58]. At first glance, the spectral interpretation and assignment seems straightforward. However, the multitude of possible modes is rather large, in particular for the higher alcohols. Even for the apparently simplest case of methanol, which contains only a methyl group, the CH stretching band remains an active field of research [36–38,46]. Intuitively, the CH stretching band of methanol would essentially contain two peaks, the antisymmetric and the symmetric CH3 stretch. The experimental spectra in Fig. 6, however, show a number of side and shoulder bands. These additional bands originate from the fact that there are different possible vibrational geometries in terms of an in-plane (ip) and an out-of-plane (op) motion. Consequently, there are two pairs of subpeaks, which slightly differ in wavenumber. Another interesting phenomenon that contributes to the complex shape of the band is Fermi resonance. Fermi resonance can take place when the vibrational frequencies of an overtone or combination band and a fundamental mode are nearly degenerate. In this situation, the two modes may couple with each other, which leads to vibrational energy being transferred from the normal vibration to the combination or overtone band. As a consequence, the normal mode becomes weaker and the combination/overtone band becomes stronger. In addition, there is a frequency effect, which leads to a spectral separation of the initially nearly degenerate modes. The mode at the slightly higher frequency is blue-shifted and the mode at lower frequency is red-shifted [59]. In methanol and the higher alcohols, such Fermi resonance can happen between the overtone and combination bands of the CH bending modes and the CH stretching modes [36,38]. In the methanol CH stretching spectrum, the IR/Raman bands at 2977/2985 and 2946/2944 cm−1 can be assigned to antisymmetric ip and op CH3 vibrations, while the 2833/2835 cm−1 mode is due to the symmetric CH3 stretching [38]. The shoulder band at 2919 cm−1 in the

IR spectrum is a Fermi resonance [38,39]. It probably results from a coupling of the overtone of the CH bending mode at 1473 cm−1 (×2 = 2946 cm−1) with the fundamental CH stretching mode at 2977 cm−1. Originally, this fundamental mode may be located close to 2946 cm−1, and the Fermi resonance leads to a spectral separation as described above so that the observed frequencies are 2919 and 2977 cm−1. Similar effects take place between other overtone/combination bands located in the low wavenumber part of the CH stretching region and the fundamental modes. A detailed theoretical analysis of the Fermi resonances in methanol was carried out by Halonen [39]. An analogue discussion could be made for the other alkanols. 3.4. Fingerprint Region Fig. 7a and b displays the fingerprint region of the IR and Raman spectrum, respectively. While some peaks are contained in basically every alcohol spectrum, there is quite some variation in other spectral regions. Similar observations were made in the normal alkanes at cryogenic conditions [60]. For example, the spectral window around 1000 cm−1 in the IR spectrum shows a multitude of peaks and it is difficult to identify any signatures that change systematically with alkyl chain length. At the high wavenumber end (N 1300 cm−1), the fingerprint region is dominated by antisymmetric and symmetric bending/deformation vibrations of the methyl, methylene, and hydroxyl groups. They are accompanied by CH2 scissoring and wagging modes. At lower wavenumber, the torsional and rocking modes of CH2 and CH3 can be found. The range between 900 and 1100 cm−1 is dominated by CC and CO stretching vibrations. These skeletal vibrations are rather unique for each alcohol and hence the strong variations in the spectrum. The peaks in this range of the spectrum of alkanes have been utilized as markers for the conformational order in the alkyl chain [61,62]. For example, in liquid octadecane the Raman bands at 1065 and 1081 cm−1 were assigned to the trans and gauche conformations of CC groups, respectively [61]. The spectra of the alkyl alkanols show a similar behavior, see Tables 3 and 4. An unambiguous assignment of the observed peaks to different conformers, however, is more difficult than in the alkanes as the CO stretching modes appear in this part of the spectrum as well. Fig. 8 illustrates the Raman spectra at the low wavenumber end of the fingerprint region. This part of the spectrum mainly contains deformation vibrations of CCO and CCC. In agreement with the above statement on the individualistic CC and CO stretching modes, the corresponding deformations also give rise to a very characteristic vibrational signature for each

Fig. 6. Enlarged (a) Infrared and (b) Raman spectra of the eleven n-alcohols in the CH stretching region.

J. Kiefer et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 189 (2018) 57–65

63

Fig. 7. Enlarged (a) Infrared and (b) Raman spectra of the eleven n-alcohols in the fingerprint region.

alcohol. These bands can, for instance, be utilized to distinguish between different conformers in terms of alkyl chain orientation as it was shown by Pitsevich and co-workers [47]. They analyzed experimental Raman spectra with the aid of density functional theory. This allowed assigning trans and gauche configurations to the carbon skeleton along the alkyl chain in alcohols ranging from pentane-1-ol to decane-1-ol.

3.5. Vibrational Mapping As could be seen above, there are several regions in the vibrational spectra that appear highly complicated and it is difficult to identify

Fig. 8. Enlarged Raman spectra of the eleven n-alcohols in the low-wavenumber range of the fingerprint region.

trends. In order to provide a clearer picture of the systematic changes with alkyl chain length, this section presents the spectra as function of the number of carbon atoms as two-dimensional contour plots. Fig. 9a and b illustrates these plots for the IR and Raman spectra, respectively, over the range from 650 to 3700 cm−1. The color maps are arranged such that they are linear between zero and the maximum absorbance/ Raman intensity. In both panels, the decrease of the OH stretching band and the increase of the CH stretching band with increasing chain length are clearly visible. Starting at methanol, the CH stretching at around 3000 cm−1 shows some wavenumber jumps and shifts with chain length up to pentan-1-ol. The higher alcohols exhibit a rather constant CH band, which only changes in intensity. A rather constant signal position can also be observed for the CH bending modes between 1450 and 1470 cm−1. In order to enhance the fingerprint region in the Raman spectrum, Fig. 9d shows the contour plot with a modified color map. The CH stretching region is no longer color resolved for the benefit of a better visualization of the low wavenumber bands. Again, several of these bands show significant changes from one alcohol to the other, but appear constant for pentan-1-ol and higher alcohols. Fig. 9c illustrates the contour plot of the low wavenumber Raman spectrum. Looking at the individual spectra in Fig. 8, this spectral region appeared rather complicated and did not allow systematic trends to be identified unambiguously. The contour plot, however, suggests systematic changes in the wavenumber of peaks as a function alkyl chain length and hence simplifies the assignment. As aforesaid, the low wavenumber region mainly contains deformation vibrations of CCO and CCC. In case of methanol, there are no CCC or CCO moieties so that the weak contributions can be assigned to torsional modes of the CO bond. In ethanol, the peak at 433 cm−1 must originate from CCO bending, and the signals at lower wavenumber are again torsion modes. The spectrum of propane-1-ol exhibits two distinct peaks, one is red- and one is blue-shifted with respect to the ethanol peak. Considering the different masses involved in the vibrating groups, the mode at 465 cm−1 is due to CCC bending and the mode at 326 cm−1 is due to CCO bending. As another CH2 group is added in butane-1-ol, the CCC mode is shifted to 398 cm−1. The assignment can be continued following this line of argument. However, it should be noted that it becomes increasingly difficult to assign peaks to individual modes as the chain length increases, because the frequencies of different vibrations can become very similar; hence, the multiple assignments in Tables 3 and 4.

64

J. Kiefer et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 189 (2018) 57–65

Fig. 9. Spectral maps of the eleven n-alcohols: (a) Infrared, (b) Raman with intensity full range, (c) Raman in low wavenumber range, (d) Raman with modified color map.

4. Conclusion The vibrational spectra of all normal alcohols that are liquid at room temperature have been analyzed and interpreted. The OH stretching region in the Raman spectrum of ethanol and the higher alcohols disclosed vibrational signatures of OH groups not participating in hydrogen bonding interactions. Monte Carlo simulations using the Cassandra simulation package supported the experimental analysis. The CH stretching region showed a rather complex mix of normal vibrations, overtone and combination bands, and Fermi resonances. The fingerprint region contains a multitude of CH and skeletal deformation, stretching, scissoring, wagging and rocking modes resulting in unique spectra for the individual alcohols. Plotting the spectra as two-dimensional maps provides an enhanced clarity as systematic changes with alkyl chain length can be observed despite the spectral complexity. Acknowledgment The authors would like to thank Avantes BV (Apeldoorn, The Netherlands) for the loan of the Raman instrument. Furthermore, the authors gratefully acknowledge financial support from the German Research Foundation (DFG) through grant KI1396/4-1. The authors also acknowledge the financial support from the Institutional Strategy of Universität Bremen funded by the Federal Government's and the Federal States' Excellence Initiative. Appendix A. Supplementary Data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.saa.2017.07.061. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]

A. Chowdhury, D. Sarkar, D. Mitra, Chem. Eng. Technol. 39 (2016) 730–740. L.V. Mattos, G. Jacobs, B.H. Davis, F.B. Noronha, Chem. Rev. 112 (2012) 4094–4123. E.R. Baumler, M.E. Carrin, A.A. Carelli, J. Food Eng. 178 (2016) 190–197. E. Reverchon, R. Adami, J. Supercrit. Fluids 37 (2006) 1–22. A.K. Agarwal, Prog. Energy Combust. Sci. 33 (2007) 233–271. A. Demirbas, Prog. Energy Combust. Sci. 33 (2007) 1–18. F.M. Zehentbauer, J. Kiefer, ChemPhysChem 16 (2015) 3846–3858. B.R. Kumar, S. Saravanan, Renew. Sust. Energ. Rev. 60 (2016) 84–115. M. Horgnies, E. Darque-Ceretti, Prog. Org. Coat. 55 (2006) 27–34. M. Horgnies, E. Darque-Ceretti, R. Combarieu, Prog. Org. Coat. 47 (2003) 154–163. M.M. Ribeiro, V.A. Neumann, M.C. Padoveze, K.U. Graziano, Revista Latino-Americana De Enfermagem 23 (2015) 741–752.

[12] D.J.M. Macdonald, E.C.A. Mckillop, S. Trotter, A.J.R. Gray, Int. J. Qual. Health Care 18 (2006) 120–122. [13] K. Noack, A. Leipertz, J. Kiefer, J. Mol. Struct. 1018 (2012) 45–53. [14] G.P. Dubey, P. Kaur, J. Chem. Eng. Data 60 (2015) 2232–2239. [15] V.R. Vale, S. Will, W. Schröer, B. Rathke, ChemPhysChem 13 (2012) 1860–1867. [16] C. Leal, A. Bilalov, B. Lindman, J. Phys. Chem. B 110 (2006) 17221–17229. [17] L.L. Simon, H. Pataki, G. Marosi, F. Meemken, K. Hungerbuhler, A. Baiker, S. Tummala, B. Glennon, M. Kuentz, G. Steele, H.J.M. Kramer, J.W. Rydzak, Z.P. Chen, J. Morris, F. Kjell, R. Singh, R. Gani, K.V. Gernaey, M. Louhi-Kultanen, J. O'Reilly, N. Sandler, O. Antikainen, J. Yliruusi, P. Frohberg, J. Ulrich, R.D. Braatz, T. Leyssens, M. von Stosch, R. Oliveira, R.B.H. Tan, H.Q. Wu, M. Khan, D. O'Grady, A. Pandey, R. Westra, E. Delle-Case, D. Pape, D. Angelosante, Y. Maret, O. Steiger, M. Lenner, K. AbbouOucherif, Z.K. Nagy, J.D. Litster, V.K. Kamaraju, M.S. Chiu, Org. Process. Res. Dev. 19 (2015) 3–62. [18] K.A. Bakeev, Process Analytical Technology, Wiley-Blackwell, 2010. [19] P.A. Martin, Chem. Soc. Rev. 31 (2002) 201–210. [20] C.B. Ojeda, F.S. Rojas, Appl. Spectrosc. Rev. 44 (2009) 245–265. [21] S. Prati, E. Joseph, G. Sciutto, R. Mazzeo, Acc. Chem. Res. 43 (2010) 792–801. [22] J. Kiefer, Energies 8 (2015) 3165–3197. [23] V. Turzhitsky, L. Qiu, I. Itzkan, A.A. Novikov, M.S. Kotelev, M. Getmanskiy, V.A. Vinokurov, A.V. Muradov, L.T. Perelman, Appl. Spectrosc. 68 (2014) 133–154. [24] T.W. Bocklitz, S.X. Guo, O. Ryabchykov, N. Vogler, J. Popp, Anal. Chem. 88 (2016) 133–151. [25] E.L. Izake, Forensic Sci. Int. 202 (2010) 1–8. [26] E.M.A. Ali, H.G.M. Edwards, I.J. Scowen, J. Raman Spectrosc. 40 (2009) 2009–2014. [27] T. Tsukahara, M. Harada, H. Tomiyasu, Y. Ikeda, J. Phys. Chem. A 112 (2008) 9657–9664. [28] E.K. Plyler, J. Natl. Bur. Stand. 48 (1952) 281–286. [29] K. Ohno, H. Yoshida, H. Watanabe, T. Fujita, H. Matsuura, J. Phys. Chem. 98 (1994) 6924–6930. [30] M. Paolantoni, P. Sassi, A. Morresi, R.S. Cataliotti, Chem. Phys. 310 (2005) 169–178. [31] P. Sassi, A. Morresi, M. Paolantoni, R.S. Cataliotti, J. Mol. Liq. 96-97 (2002) 363–377. [32] S.J. Barlow, G.V. Bondarenko, Y.E. Gorbaty, T. Yamaguchi, M. Poliakoff, J. Phys. Chem. A 106 (2002) 10452–10460. [33] M. Tomsic, G. Fritz-Popovski, L. Vlcek, A. Jamnik, Acta Chim. Slov. 54 (2007) 484–491. [34] M. Tomsic, A. Jamnik, G. Fritz-Popovski, O. Glatter, L. Vlcek, J. Phys. Chem. B 111 (2007) 1738–1751. [35] R. Zheng, Y. Sun, Q. Shi, Phys. Chem. Chem. Phys. 13 (2011) 2027–2035. [36] T. Ishiyama, V.V. Sokolov, A. Morita, J. Chem. Phys. 134 (2011) 024509. [37] T. Ishiyama, V.V. Sokolov, A. Morita, J. Chem. Phys. 134 (2011) 024510. [38] Y. Yu, Y. Wang, K. Lin, X. Zhou, S. Liu, J. Phys. Chem. A 117 (2013) 4377–4384. [39] L. Halonen, J. Chem. Phys. 106 (1997) 7931–7945. [40] L.M. Levering, C.J. Hayes, K.M. Callahan, C.M. Hadad, H.C. Allen, J. Phys. Chem. B 110 (2006) 6325–6331. [41] T.N. Wassermann, P. Zielke, J.J. Lee, C. Cezard, M.A. Suhm, J. Phys. Chem. A 111 (2007) 7437–7448. [42] M. Paolantoni, P. Sassi, A. Morresi, R.S. Cataliotti, J. Raman Spectrosc. 37 (2006) 528–537. [43] V. Crupi, S. Magazu, G. Maisano, D. Majolino, P. Migliardo, S. Venuto, Phys. Chem. Liq. 26 (1994) 263–272. [44] F. Palombo, M. Paolantoni, P. Sassi, A. Morresi, R.S. Cataliotti, J. Mol. Liq. 125 (2006) 139–146.

J. Kiefer et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 189 (2018) 57–65 [45] C.D. Stanners, Q. Du, R.P. Chin, P. Cremer, G.A. Somorjai, Y.R. Shen, Chem. Phys. Lett. 232 (1995) 407–413. [46] R. Lu, W. Gan, B.H. Wu, Z. Zhang, Y. Guo, H.F. Wang, J. Phys. Chem. B 109 (2005) 14118–14129. [47] G.A. Pitsevich, I.Y. Doroshenko, V.Y. Pogorelov, E.N. Kozlovskaya, T. Borzda, V. Sablinskas, V. Balevicius, Vib. Spectrosc. 72 (2014) 26–32. [48] H.H. Eysel, J.E. Bertie, J. Mol. Struct. 142 (1986) 227–230. [49] C. Perchard, J.P. Perchard, J. Raman Spectrosc. 7 (1978) 173–177. [50] C. Perchard, J.P. Perchard, J. Raman Spectrosc. 3 (1975) 277–302. [51] J.K. Shah, E.J. Maginn, J. Chem. Phys. 135 (2011) 134121. [52] B. Chen, J.J. Potoff, J.I. Siepmann, J. Phys. Chem. B 105 (2001) 3093–3104. [53] J. Joseph, E.D. Jemmis, J. Am, Chem. Soc. 129 (2007) 4620–4632. [54] G.R. Desiraju, Angew. Chem. Int. Ed. 50 (2011) 52–59. [55] V.M. Wallace, N.R. Dhumal, F.M. Zehentbauer, H.J. Kim, J. Kiefer, J. Phys. Chem. B 119 (2015) 14780–14789.

65

[56] K. Ohno, T. Shimoaka, N. Akai, Y. Katsumoto, J. Phys. Chem. A 112 (2008) 7342–7348. [57] NIST Standard Reference Database 69: NIST Chemistry Webbook: http://Webbook. NIST.Gov/Chemistry 2017. [58] Y.Q. Yu, K. Lin, X.G. Zhou, H. Wang, S.L. Liu, X.X. Ma, J. Phys. Chem. C 111 (2007) 8971–8978. [59] J. Kiefer, K. Noack, T.C. Penna, M.C.C. Ribeiro, H. Weber, B. Kirchner, Vib. Spectrosc. 91 (2017) 141–146. [60] R.G. Snyder, J.H. Schachtschneider, Spectrochim. Acta 19 (1963) 85–116. [61] C.J. Orendorff, M.W. Ducey Jr., J.E. Pemberton, J. Phys. Chem. A 106 (2002) 6991–6998. [62] G. Zerbi, R. Magni, M. Gussoni, J. Mol. Struct. 73 (1981) 235–237.