Hydrogen Bonding in Pyridine - American Chemical Society

Article pubs.acs.org/JPCA

Hydrogen Bonding in Pyridine N‑Oxide/Acid Systems: Proton Transfer and Fine Details Revealed by FTIR, NMR, and X‑ray Diffraction Vytautas Balevicius,*,† Kęstutis Aidas,† Ingrid Svoboda,‡ and Hartmut Fuess‡ †

Department of General Physics and Spectroscopy, Vilnius University, Sauletekio 9-3, LT-10222 Vilnius, Lithuania Institute for Materials Science, Darmstadt University of Technology, Petersen Str. 23, D-64287 Darmstadt, Germany

‡

ABSTRACT: The H-bonded complexes of pyridine N-oxide (PyO) with H2O, acetic, cyanoacetic, propiolic, tribromoacetic, trichloroacetic, trifluoroacetic, hydrochloric, and methanesulfonic acids have been studied by FTIR and NMR spectroscopy, X-ray diffraction, and quantum chemical DFT calculations. Correlations between vibrational frequencies of the NO stretching and PyO ring modes and geometric parameters of the H-bond have been established. FTIR experiments show and DFT calculations confirm that definite discontinuity is present in the vicinity of the midpoint in the proton transfer pathway. The established correlations significantly aid in the understanding of fine effects such as the isotope (deuteration) effect, crystal-to-solution transition, or criticality of aqueous solutions induced by ionic pairs. Geometric isotope effect in the ionic H-bond aggregate of PyO·H(D)Cl was found to be extraordinary large. Measured FTIR, CP/MAS, and high-resolution 13C NMR spectra indicate that H-bond in the PyO·HCl complex in polar solvent can potentially be more ionic than in the crystal. Vibrational modes of ionic pairs originating via proton transfer in H-bond complexes can provide new information concerning the interionic interaction and its role in the phase separation and mezo-structuring processes. The results are compared to the relevant data for PyO·HCl complex in argon matrix.

■

INTRODUCTION Recent developments in the hydrogen bond (H-bond) research show persisting or even boosting interest in this type of interaction having a great importance in many physical or chemical processes.1 A rich variety of H-bond related phenomena arises from the interplay between stronger and weaker bonds and their evolution upon media effects, clustering, bulk/surface effects, etc.2 Because of its abundance and importance in the molecular world, precise definition of Hbond, detection and evaluation of energy, determination of proton transfer pathways, and monitoring of proton dynamics are subjects to continuous debate and research.1,3,4 Aromatic heterocyclic compounds, such as derivatives of pyridine (Py) and pyridine N-oxide (PyO), occupy in the field of H-bond research a very particular place.5,6 Their proton accepting abilities can be varied in a broad range by proper substitution at definite positions of the π-electron carcass. Hence, one can manipulate the H-bond strength and thus gain or dampen proton dynamics. In addition, the complexes between PyO or its derivatives and various acids are considered as promising benchmark systems to improve the understanding of nontrivial physical features of short O−H···O hydrogen bonding.7 The most comprehensive information concerning geometric H-bond parameters are provided by methods of the direct structural research, that is, neutron and X-ray diffraction (XRD) techniques applied to monocrystalline samples, but neutrons are not easily available and are fairly expensive. Furthermore, © 2012 American Chemical Society

the growth of single crystals is sometimes extremely difficult or not at all feasible. In this light, empirical correlations between the geometric H-bond parameters and easily measurable quantities are of high practical value. Most of the spectral observables such as frequencies and intensities of the vibrational bands in the FTIR absorption and Raman spectra or NMR chemical shifts can be indeed considered as being easy to measure. Some exceptions are, however, possible. For example, very strong H-bonds with short distance between heavy atoms typically lead to broad asymmetric and complexshaped FTIR bands of the stretching mode, and stretching band parameters are thus very difficult to extract. NMR chemical shifts of nuclei of low natural abundance or NMR in partially ordered mesophases and solids are to be added to the list too. In the present work, continuing the studies of the proton shared H-bonding with incipient proton transfer and already transferred proton in liquid and crystalline phases,8−10 we report new FTIR, 13C CP/MAS, XRD, and DFT data concerning fine details of the H-bonding in PyO complexes with different acids of increasing acidity: (i) correlations between vibrational frequencies of the NO stretching and PyO ring modes and geometric parameters of the H-bond; (ii) discontinuity in the vicinity of the midpoint in the proton transfer pathway; (iii) extraordinary large geometric isotope Received: June 4, 2012 Revised: August 6, 2012 Published: August 9, 2012 8753

dx.doi.org/10.1021/jp305446n | J. Phys. Chem. A 2012, 116, 8753−8761

The Journal of Physical Chemistry A

Article

X-ray Diffraction Study. Diffraction data have been collected with rotation method data acquisition using ω scans on the Oxford Diffraction Xcalibur (TM) Single Crystal X-ray Diffractometer with Sapphire CCD Detector. Type of absorption correction: multiscan (Empirical absorption) correction using spherical harmonics implemented in SCALE3 ABSPACK scaling algorithm. Structure solutions and refinements were performed using SHELX 18 and SHELXL,19 respectively. The hydrogen atoms at the benzene ring were positioned with idealized geometry using a riding model with C−H = 0.93 Å. H atoms of the OH group were located in the difference map. Isotropic displacement parameters for all H atoms were set equal to 1.2Ueq of the parent atoms. Details of data acquisition, refinement, lattice constants, and hydrogen bonding scheme are collected in Tables 1 and 2 (CCDC reference number 879632).

effect in the ionic H-bond aggregate of PyO·H(D)Cl, etc. Considered complexes of PyO with acetic, cyanoacetic, propiolic, tribromoacetic, trichloroacetic, and trifluoroacetic acids cover a very broad range of H-bonding features and allow to follow proton dynamics in the wide region close to the middle of the H-bond, culminating in the complexes with hydrochloric (HCl) and methanesulfonic acids where complete proton transfer is definitely observed. The results are compared to analogous data on H-bonding in Py/acid systems.9,10 Presented results contribute also the series of works initiated by Hadži and continued by the Ljubljana group (see refs 7 and 11−14) on the systems with short and very short H-bonds and large-amplitude proton dynamics.

■

EXPERIMENTAL SECTION Samples. Commercial pyridine N-oxide (PyO), acetonitrile (AN), and acids, acetic (AA), cyanoacetic (CyA), propiolic (PA), tribromoacetic (TBA), trichloroacetic (TCA), trifluoroacetic (TFA), hydrochloric (HCl), and methanesulfonic (MSA), were preliminary purified by standard methods.15 Water from TFA was removed by adding trifluoroacetic acid anhydride and subsequent distillation. The nonaqueous solutions were prepared in a nitrogen filled drybox. Other remedies, such as molecular sieves, degassing, argon bubbling, vacuum flushing, etc., were applied for drying the chemicals. Samples of PyO·HCl were prepared by direct dissolving of PyO·HCl crystal in AN. Limited solubility allowed to achieve only 0.1 M concentration of PyO·HCl and 0.5 M for PyO complex with MSA. The solutions were prepared in a drybox by weighing (±0.1 mg) the components. The solutions containing water were mixed in open air. PyO·HCl and PyO·DCl·HCl crystals have been obtained from water (H2O or D2O) as well as from methanol solutions in the National Institute of Chemistry (Ljubljana, Slovenia). The degree of deuteration of PyO·DCl achieved was not less than 50%. A colorless transparent single-crystal PyO·HCl with dimensions ca. 0.4 × 0.4 × 0.1 mm was selected for the X-ray diffraction study. Unfortunately, no single crystal of PyO·DCl·HCl suitable for Xray diffraction was obtained. FTIR Spectra. FTIR spectra were recorded on a PerkinElmer 2000 FTIR spectrometer on line with a standard computer. Solutions were contained in a cell between NaCl windows with Teflon spacer (0.05 mm). FTIR spectra of crystals were recorded using KBr pellets or Nujol mulls. The instrument was purged with dry nitrogen to avoid CO2 absorption. The spectral resolution was 1−2 cm−1, and a total of 16 interferograms were added and averaged for each spectrum. Difference spectra were obtained by interactive subtraction using the PE Spectrum for Windows program package. Additionally, some spectra were processed using Microcal Origin.16 NMR Measurements. Details of NMR studies in the liquid state have been given in earlier works.10,17 13C CP/MAS (crosspolarization/magic angle spinning) spectra of solids were obtained using a Doty dynamic angle-spinning (DAS) probe. Crystals of PyO·HCl were ground and put into a ceramic rotor of DAS in a drybox. For MAS spectra, the sample-spinning rate was 4 kHz, and the error in the spinning rate was less than ±10 Hz. All NMR measurements were performed at 293 K. The temperature was controlled with an accuracy of ±0.5 K. The NMR spectra contours were processed digitally using the Microcal Origin program package.

Table 1. X-ray Diffraction Results and Structure Refinement on PyO·HCl Crystal empirical formula formula weight temperature (K) wavelength (Å) crystal system space group unit cell dimensions a (Å) b (Å) c (Å) α, β, γ (deg) volume (Ǻ 3) Z calculated density (mg m−3) absorption coefficient (mm−1) F(000) crystal description and size (mm) Θ range for data collection (deg) limiting indices reflections collected independent reflections refinement method data/restraints/parameters goodness-of-fit on F2 final R indices (I > 2σ(I)) R indices (all data) largest diff. peak and hole (e·Ǻ −3)

C5H6ClNO 131.56 293(2) 0.71073 orthorhombic P212121 7.2871(3) 7.6283(3) 11.0251(5) 90.00 612.87(4) 4 1.426 0.516 272 prism, colorless, 0.40 × 0.40 × 0.10 3.25 to 26.36 −9 < h < 9; −9 < k < 9; −13 < l < 13 4204 1248, [R(int) = 0.0163] full-matrix least-squares on F2 1248/0/74 1.121 R1 = 0.0232, wR2 = 0.0566 R1 = 0.0258, wR2 = 0.0575 0.148 and −0.127

DFT Calculations. All quantum chemical calculations in this work were performed using the Gaussian09 program package.20 Harmonic frequency analysis was based on the geometryoptimized molecular systems utilizing the B3LYP exchangecorrelation functional21 along with the 6-311++G(3df,3pd) basis set22 to ensure convergence. No scaling was applied for the calculated vibrational frequencies. For the strongly hydrogen bonded molecular complexes with double-well potential, theoretical approaches beyond harmonic approximation are generally mandatory, see, for example, ref 23. Our focus in this work is, however, on vibrational modes where the bridge proton is not directly involved; thus, harmonic approximation is expected to be reasonably valid in these cases. The PBE0 functional24 and the 6-311++G** basis22 were used to geometry-optimize relevant molecular aggregates and 8754



Article

nonbonded PyO and PyO·acid complexes are presented in Table 3.

Table 2. Hydrogen Bonding Scheme of the PyO·HCl Complex (in Å and deg)a

■

D−H···A

D−H

H···A

D···A

∠D−H···A

O1 H1A CL1

0.95

1.91

2.8604(14)

174.3

RESULTS AND DISCUSSION We have previously reported8 that the 13C NMR chemical shifts of the PyO ring are very sensitive to the strength and state of the H-bond in the systems of PyO·acid. These spectral parameters are indeed easily obtained from routine highresolution 13C NMR measurements and thus are ideal to be used for the correlations of the type structure ⇔ easily measurable quantity. Regrettably, such application of the highresolution NMR spectroscopy is restricted to the liquid samples only. It is, however, rather easy to characterize H-bonding in the solid state by using FTIR spectroscopic techniques. In this way, H-bonding in liquid crystalline gels, matrix isolated clusters, and other systems of technological or life sciences interest can be studied. For example, FTIR spectroscopy was successfully used to monitor proton transfer in some polymers and supramolecular nanostructures.27,28 Some H-bond-sensitive vibrational modes, namely, O−H and CO stretching vibrations, can not always serve the purpose, as it was discussed in our previous work,9 where Py·acid complexes were studied. The FTIR absorption bands of these modes for those complexes are typically located in the region of ca. 1000− 1500 cm−1 and exhibit broad asymmetric shapes, sometimes along with Evans holes.29 The νO−H bands of some selected H-bond complexes of PyO, namely, with CyA, PA, TFA, and MSA, are illustrated in Figure 1. It is clear that it is barely possible to determine reliable spectra parameters like centers of gravity, widths, or integrated intensities of these bands. In the case of PyO as H-bond acceptor, the stretching vibration of the N → O moiety (νNO) is expected to be also very sensitive to the state of H-bond. Indeed, this band is seen at 1252 cm−1 in the non-H-bonded PyO, referred to as free PyO in the following, and is red-shifted by ∼18 cm−1 even in rather weak complex with H2O (Figure 2). In the strongest (being precise, the strongest among those investigated)

a

Neutron diffraction data for PyO·HCl crystal was communicated by D. Hadži:30 D−H 0.994(4); H···A 1.771 (4); D···A 2.7633 (18) (T = 125 K).

predict their 13C NMR isotropic shielding constants. London atomic orbitals were used in the NMR calculations as implemented in Gaussian09. London atomic orbitals are perturbation dependent, and they ensure form invariance of the total Hamiltonian under gauge transformation,25 therefore leading to gauge origin invariance of molecular properties defined as response to external magnetic field, e.g., NMR shielding. To model bulk effects of the acetonitrile solvent and argon matrix, we have relied on the polarizable continuum model (PCM).26 We have modified the PCM built-in parameters ofac and rmin to possess values of 0.8 and 0.5, respectively, in order to simplify the constructed molecular voids within the dielectric medium. Default PCM settings adopted in Gaussian09 were used otherwise. Calculated structural parameters of the H-bond and harmonic vibrational frequencies of the relevant modes of

Table 3. Theoretically Predicted Structural Parameters and Frequencies of Relevant Vibrational Modes of Nonbonded PyO (nb PyO) and PyO·Acid H-Bond Complexes in AN Solutiona 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

acid

R(A···B) (Å)

nb PyO H2O AA CyA PA TBA TCA TFA PT-TCAb PT-TFAb MSA H2SO4 HCl + H nb PyO/Ar HCl/Ar nb PyO/in vacuo HCl/

∞ 2.743 2.604 2.541 2.540 2.492 2.486 2.476 2.453 2.460 2.502 2.541 2.881 ∞ ∞ 2.865 ∞ in vacuo2.927

rA−H (Å) 0.982 1.010 1.029 1.030 1.052 1.056 1.064 1.378 1.384 1.463 1.518 1.844 ∞ 1.380 1.351

R(H···B) (Å)

∠AHB (deg)

∞ 1.762 1.609 1.522 1.521 1.447 1.438 1.420 1.082 1.079 1.041 1.024 1.037 0.971 ∞ 1.486 ∞ 1.578

177.7 167.3 169.6 169.1 170.8 170.5 171.3 173.5 173.5 175.0 176.1 178.3

176.1 174.6

rA−H/ R(A···B)

νNO (cm−1)

ν1 (cm−1)

ν2 (cm−1)

0 0.358 0.388 0.405 0.406 0.422 0.425 0.430 0.561 0.563 0.585 0.598 0.640 1 0 0.482 0 0.462

1267.1 1252.7 1241.4 1230.2 1233.4 1232.7 1233.3 1216.6 1225.7 1235.3 1233.8 1232.6 1224.6 1199.6 1315.9 1238.0 1328.5 1270.0

1496.7 1500.4 1505.0 1505.3 1505.3 1505.1 1505.3 1505.1 1533.5 1539.5 1516.1 1516.8 1513.1 1513.9 1497.0 1498.9 1498.0 1503.0

1506.5 1507.0 1515.9 1514.5 1515.1 1515.0 1515.5 1515.7 1490.8 1499.0 1497.9 1495.0 1483.6 1422.9 1508.4 1511.5 1508.9 1511.6

a

For PyO and PyO·HCl, the values in Ar matrix and in vacuo are given in addition (pos. 15−18). See Samples section for labeling of molecular species. bH-bond complex with the proton transfer. 8755



Article

Figure 3. Dependence of the PyO νNO frequency on the proton position in the H-bond: black circle, experimental data (all 1 M in AN; lower concentrations of PyO·HCl and PyO·MSA due to limited solubility); red square, DFT based data (see Table 3). Experimental shift for PyO·HCl in Ar matrix (blue diamond) was evaluated taking the data from ref 31 (see Table 4); H-bond geometry of crystalline PyO·HCl (green pentagon) was used as determined by neutron and X-ray diffraction; both are given in Table 2. Larger error bars for some experimental frequencies are due to the overlap of the νNO band with intensive bands of the acids in the region of 1170−1270 cm−1 (Figure 1).

Figure 1. FTIR spectra of some PyO·acid solutions (all 1 M in AN, except 0.5 M for PyO·MSA due to limited solubility): 1, PyO·MSA; 2, PyO·TFA; 3, PyO·PA; 4, PyO·CyA. Shaded areas depict approximate positions of the O−H stretching bands.

DFT calculations. We have measured FTIR spectra of free PyO as well as PyO/acid H-bond complexes in AN solutions. The relevant experimental FTIR data are given in Table 4. Since Table 4. Relevant Experimental FTIR Data on Nonbonded PyO (nb PyO) and PyO·Acid H-Bond Complexes (All 1 M in AN, except 0.5 M for PyO·MSA and 0.1 M for PyO·HCl, Due to Limited Solubility)a 1 2 3 4 5 6 7, 9 8, 10 11 13

Figure 2. Experimental and calculated FTIR spectra of PyO in the top panel; ring and NO stretching modes in the bottom panel. The shifts of these modes upon PyO interaction with water are shown in the two lower spectra, using the samples prepared in drybox (black), open-air (red), and after the addition of 1 M (green) and 2 M (blue) H2O to solution, respectively.

15 16

acid

νNO (cm−1)

ν1, ν2 (cm−1)

nb PyO H2O AA CyA PA TBA TCA TFA MSA HCl HCl/crystal nb PyO/Ar HCl/Ar

1252.0 1234.0 ovb 1202.6 1203.2 ov ov 1197 1205 1197.8 1195.8c 1291, 1289d 1253d

1462.6 1467.0 1468.8 1469.7 1470.2 1473.4 1475.4 1475.9 1477.2 1474.6 1433.0, 1446.3c

a

The subject numbering used for convenience is the same as that n Table 3. bOverlapped with intensive bands of the acids in the region of 1170−1270 cm−1. cCrystalline sample. dIn Ar matrix, taken from ref 31. The values of νNO for nb PyO (pos. 1 and 15) have been used by calculating the mode shift upon media effects (AN solution and Ar matrix) that was later used to compare the mode shifts upon Hbonding (see Figure 3).

complex of PyO·HCl, this band is shifted to 1198 cm−1 in AN solution and to 1196 cm−1 in the crystalline phase (Figure 3). However, the νNO band can be hidden under more intense bands of H-bond partners, as, for example, seen in the region of 1170−1270 cm−1 of the spectra in Figure 1. To complement information provided by the νNO vibration, we have also considered vibrational ring modes of PyO as probes of the Hbond state. We have studied the behavior of νNO and two PyO ring modes (designated by ν1 and ν2, see Figure 2) with changing H-bond strength using FTIR measurements and theoretical

there are no experimental techniques for direct estimation of the geometry of H-bonds in the liquid state, the calculated values of rA−H and R(A···B) using DFT at the level of theory given above were used for the relative proton positioning in the A−H···B bridge in this series. Nevertheless, checking the reliability, one true experimentally obtained point was successfully added to the plots: the value of rA−H/R(A···B) 8756



Article

of the argon atoms around the complex. Detailed modeling like this is a study in its own, and we thus have decided not to undertake it. As FTIR experiments show and DFT calculations confirm (Figure 3), definite discontinuity is present in the vicinity of the midpoint rA−H/R(A···B) = 0.5. The shift of the νNO mode remains to be red even when the proton transfer occurs and ionic structures of NO+H···O− are formed in the complexes of PyO with MSA and HCl. This rather unexpected behavior of the νNO shift vs rA−H/R(A···B) resembles singularities usually seen for certain physical parameters in the vicinity of the second-order phase transition, i.e., when the symmetry of the system is being changed. Not only discontinuity but also conversion of the type of shift from blue to red was observed for PyO ring modes ν1 and ν2 crossing the midpoint (Figure 4). It is interesting and useful to compare this behavior with that of the corresponding ring modes of Py and Py derivatives in a series of their H-bond complexes with some acids.9,32 In particular, detailed FTIR and DFT studies were performed for the 4-cyanopyridine (CyPy) complexes with H2O, AA, TCA, and MSA, all in AN solution.9 First of all, the difference between the frequncies of the ν1 and ν2 modes is ∼40 cm−1 in Py and only ∼10 cm−1 in PyO. Therefore, these two modes are easily seen in the FTIR spectrum of CyPy,9 but they overlap in the spectrum of PyO (Figure 2). Sensitivity of these modes to the H-bond strength is also rather different. For example, the shift of the ν1 mode of 17−22 cm−1 (depending on the solvent9) in CyPy·TCA was observed, whereas this shift is only 12.8 cm−1 in the complexes of PyO with TCA and TFA (Figure 4). In contrast to PyO·acid systems, the shift of the ring modes in Py·acid series of complexes retains its sign in the whole range of rA−H/R(A···B) from 0 to 1. Bands at 1635 cm−1 and 1639 cm−1 in the spectra of Py·TFA32 and CyPy·TCA,9 respectively, were assigned to the vibrational ring mode of pyridinium (PyH+). This vibrational mode is insensitive to different substitutions of Py but suffers sizable blue shift of around 40 cm−1 upon protonation of Py. Therefore, this particular vibrational mode can be used as a universal probe of proton transfer in the entire family of pyridines. A special comment concerning complexes of PyO with TCA and TFA is necessary. Both neutral and proton transfer H-bond structures of these complexes were found using DFT calculations along with the PCM description of AN solvent; see Table 3 for geometrical parameters. Corresponding points are also placed in Figures 3 and 4. However, experimental points fit better to the smooth curve on the left-hand branch in the ring mode frequency dependence on the positioning parameter, i.e., at rA−H/R(A···B) < 0.5 (see the block arrow in Figure 4). This corresponds to the neutral H-bond complexes and indicates that neutral structures of PyO·TCA and PyO·TFA are, nevertheless, more stable at the present conditions (room temperature and AN solution) than the ionic pairs. It means that the thermodynamic equilibrium is significantly shifted toward the neutral species. However, note that the proton transfer in the crystalline PyO·TCA has been found at very low temperatures (10 K) using INS technique.7 Obtained dependencies of νNO and PyO ring mode frequencies on the positioning parameter rA−H/R(A···B) can be applied to study some fine H-bonding effects. Isotope (Deuteration) Effect. Geometric isotope effects in H-bonded systems have been intensively studied since the 1980s; see refs 33−36 and refs cited therein. Several quite

= 0.668 was obtained directly from the X-ray diffraction data analysis of PyO·HCl single crystal in the present work (Table 2) and 0.641 was found using neutron diffraction.30 The dependencies of νNO and PyO ring modes on the positioning parameter rA−H/R(A···B) are shown in Figures 3 and 4. A

Figure 4. Dependencies of the PyO ring mode frequencies on the proton position in the H-bond. The possible proton transfer in PyO complexes with TCA and TFA is shown as the block arrow.

very large red shift of ∼55 cm−1 of the νNO mode is seen when going from the free PyO to its neutral complex with TFA (Figure 3). The rA−H/R(A···B) ratio of 0.430 was found for the PyO·TFA complex and is the largest among all neutral complexes we have considered. It is always interesting to monitor the changes in proton positioning in the H-bond upon media effects, e.g., as it was succeeded in NMR and FTIR spectra going from less to more polar solvents.8,9 The matrix isolation experiments are therefore of particular importance. This is because noble gas matrixes are often considered as intermediate media between solutions and the gas phase, where proton transfer has never been observed. The FTIR data are available for one of the complexes from the current series, viz. PyO·HCl in argon matrix. 31 The experimentally determined shift of the νNO mode in argon matrix with respect to the nonbonded PyO 37 cm−1 (see Table 4) is thus much smaller as compared to that of ca. 50 cm−1 seen for the complexes of PyO with different acids in solution or PyO·HCl complex in the crystalline phase (Figure 3). The ratio rA−H/R(A···B) of 0.482 in argon matrix was obtained based on our DFT calculations (Table 3). However, experimental and calculated νNO frequencies perfectly fit to the line on the righthand branch in the frequency dependence on the positioning parameter, i.e., at rA−H/R(A···B) > 0.5 (Figure 3) that corresponds to the H-bond complexes with proton transfer. Thus, our experimental measurements indicate the presence of the zwitterionic PyO·HCl complex in the argon matrix. From the computational side, we have applied a very simplistic dielectric continuum approach to model argon environment, which did not lead to full proton transfer in this system. Dispersion and short-range repulsion interactions dominate in the noble gas matrixes, which are completely neglected by the continuum model. We could account for short-range repulsion by including argon atoms explicitly in the calculation, yet dispersion interactions require electronic structure methods beyond DFT. Such an approach would also introduce issues related to dynamical effects as well as to structural distribution 8757



Article

cm−1 using DFT based frequencies and 100 ± 20 cm−1 using measured frequencies, respectively. It follows that Δ(r/R) = (r/ R)H − (r/R)D can reach ∼0.08. Since the value of (r/R)H = 0.668 was obtained directly from the X-ray diffraction data analysis (Table 2), the corresponding value for deuterated complex (r/R)D drops by ∼0.08 to ∼0.59. Expecting that the deuteration effect on the O···Cl distance is also confined within ca. 0.03 Å, the value of rA−D could be within 1.69 Å (in the case ΔR = 0) and 1.71 Å (the increase ΔR by 0.03 Å), whereas rA−H = 1.91 Å was found for PyO·HCl (Table 2). This means that D is shifted by 0.2−0.22 Å toward the midpoint of the O···D+···Cl− bridge with respect to the H position in O···H + ···Cl−. Hence, the contraction Δr is extraordinary large. Much lower values, Δ(r/R) ≈ 0.04 and, consequently, Δr ≈ −0.1 Å, are obtained by processing the data for the PyO ring modes. The experimentally determined shifts of the ring modes upon deuteration are much smaller, ca. 2 cm−1 (Figure 5). However, the slopes of the ν1 and ν2 vs rA−H/R(A···B) are also significantly lower (∼40 cm−1, Figure 4). A giant deuteron migration, anomalous spectroscopic isotopic ratio, very large vibrational frequency shifts, etc., often occur in H-bonded crystals due to structural phase transitions.36−38 These effects can be interpreted as a consequence of different dimensions and geometries of the H-bonds that stabilize the crystalline structure.38 The structural phase transitions in related H-bond systems usually occur somewhere ∼100 K below the room temperature, e.g., at 175 K for 3,5-pyridinedicarboxylic acid37 and at 170 K for alanine crystal.38 In the present work, the effect of deuteration has been studied using PyO·DCl·HCl (H/D ≈ 1:1) cocrystal at 293 K. The H-bond structure in the investigated PyO·HCl crystal is definitely ionic already at 293(2) K (rA−H/R(A···B) = 0.668, Table 2). The above evaluated shift of D by ∼0.2 Å toward the midpoint of the H-bond certifies that PyO·DCl complex at this temperature should also be ionic. The cooling can only deepen the proton/deuteron migration toward PyO, and thus, the structural transition related with H/D transfer across the midpoint of the H-bond can be excluded in the present case. It is obvious that a certain error in the scaling of the H-bond geometry may arise because the calculated values of rA−H and R(A···B) have been used to calibrate vibrational frequency vs H-bond geometry curves (Figures 3 and 4). Nevertheless, considering evaluations above, one can expect that geometric isotope effects in certain ionic H-bond structures can be quite large, e.g., of the order of ∼0.1 Å. Whether there exists the compounds that exhibit the extraordinary large isotope effect was pointed out in ref 34 as one of the most burning questions that still remains to be settled. Crystal-to-Solution Transition. The importance of media effects on the strength of H-bonding and proton transfer has been stressed almost in all studies on this topic. Some interesting aspects have been discussed in ref 39. Namely, the driving force of proton transfer is the electric field created by the electric dipoles of solvent molecules at the H-bond site. The strength of solvent field depends on the ordering of individual dipoles. The question of the extent of this dipole ordering, i.e., whether it is an overall effect or it covers only a part of the molecules in the closest shell surrounding the H-bond complex, is still open. Therefore, we have attempted to compare the patterns of H-bond in the PyO·HCl complex in the crystalline and liquid phases. Crystalline and liquid samples were studied using FTIR, 13C CP/MAS, and conventional high-resolution 13 C NMR. Recorded spectra are shown in Figures 5 and 6.

general rules have been established during these studies. However, new X-ray and neutron diffraction data enforce some of them to be revised from time to time.34 Essential findings that are still valid and of importance in the context of this work are listed as follows: (i) the magnitude of deuteration effect on the O···O distance (the increase) is confined within 0.03 Å, being close to 0, for weak, longer than 2.65 Å, and for strong, shorter than ∼2.45 Å bonds;36 (ii) O−H decreases by about 0.005 Å; and (iii) H···O increases by about 0.02 Å.34,35 The effect of deuteration in PyO·HCl and PyO·DCl·HCl cocrystals has been studied by FTIR spectroscopy (Figure 5).

Figure 5. FTIR data for PyO·H(D)Cl to illustrate fine details of the Hbonding: 1, PyO·HCl in AN solution; 2, PyO·HCl crystal; 3, PyO·HCl·DCl (H/D ≈ 1:1) crystal. Spectrum of PyO in AN solution (4) has been added to identify the traces of the nonbonded PyO in the spectra of PyO·HCl.

The PyO ring modes at 1474−1478 cm−1 are barely altered by the deuteration. In contrast, the νNO mode is expected to be quite sensitive. Indeed, two distinct bands are seen in the FTIR spectrum of the PyO·DCl·HCl cocrystal (Figure 5). By comparing the latter spectrum to that recorded on PyO·HCl crystal, we can safely assign these two bands to the νNO mode perturbed by the interaction with HCl (at 1196.8 cm−1) and DCl (at 1204.4 cm−1). Interestingly, no difference in the frequencies of the νNO mode has been detected in the spectra of PyO·HCl and PyO·DCl complexes in the argon matrix.31 We can estimate the magnitude of the geometrical changes of the H-bonding induced by the deuteration. The deuteration induced shift of the νNO mode is Δν ≈ 8 cm−1 (Figure 5). The slope Δν/Δ(r/R) evaluated by the linear fit of the νNO frequency vs rA−H/R(A···B) dependence (Figure 3) is 75 ± 10 8758



Article

Ionic character and proton sharing properties of the H-bond in the family of PyO·acid systems can be very effectively monitored using the 13C NMR chemical shifts.8 It has been shown experimentally and confirmed theoretically by DFT calculations that the 13C NMR chemical shifts of C2, C6 and C3, C5 carbon atoms of PyO (see Figure 6 for atom labeling) are only very little affected by the changing H-bond strength and polarity of the solvent. In contrast, chemical shift of the C4 atom shows a clear tendency to be shifted to higher values with increasing H-bond strength, reaching maximum in the proton transfer complexes of PyO·H+·MSA− and PyO·H+·Cl−.8 The 13 C NMR spectra of free PyO and PyO·HCl complex in AN solution together with CP/MAS spectrum are shown in Figure 6. Both experimental data and theoretical predictions (Table 5) Table 5. Experimental and Calculated 13C Chemical Shifts (all in ppm) of nb PyO and PyO·HCl Complex in AN Solution and in the Crystalline Phasea Calculatedb nb PyO PyO·HCl Experimental nb PyO PyO·HCl PyO·HCl/crystal (CP/MAS data)

C2,6

C3,5

C4

146.2 146.6

133.2 134.8

131.9 149.8

139.5 139.9 142.3

127.0 128.8 130.9

125.5 141.0 142.4

a For carbon numbering, see Figure 6. bAll calculated chemical shifts are given in the δ-scale as the difference between the calculated isotropic part of magnetic shielding tensor and that of tetramethylsilane (TMS) as the reference molecule (see ref 17).

show that δ(C4) is larger than δ(C2,6) in the PyO·HCl complex in AN solution. In the CP/MAS spectrum of the crystalline sample, signals of C4 and C2,6 carbons are completely overlapped (Figure 6). The spectral contour was processed by applying a nonlinear curve fitting procedure using two Lorentz functions. It was found that δ(C4) in this case is slightly (∼0.1 ppm) higher than δ(C2,6). A smaller difference between chemical shifts δ(C4) and δ(C2,6) in the solid phase than in the liquid phase indicates that H-bond in the PyO·HCl complex in AN solution can potentially be more ionic than in the PyO·HCl crystal. Added Water and the Critical Aqueous Solutions. Binary solvent mixtures of water and organic components like AN, tetrahydrofuran (THF), etc., are often used in modern bioand nanotechnological applications and various processes of mezo-structuring.40,41 Solutions of AN and THF in H2O or D2O are homogeneous for any molar ratio of the components at room temperature. However, even a very small amount of dissolved ions can significantly change their phase diagrams.42−44 This phenomenon has been intensively studied and interpreted to occur due to the competition between entropic and energetic effects.43,45 Multicritical behavior and crossover of critical regimes have also been scrutinized.46,47 As it was indicated in the experimental section, solubility of the PyO·HCl crystal in AN is rather low, ca. 0.1 M at room temperature. Solubility can be improved by adding H2O or D2O. For example, with added water in an amount that corresponds to the stoichiometric ratio, approximately PyO·HCl·4H2O, this conglomerate can be dissolved in AN already up to concentrations of 1 M. By adding water, we produce, however, a rather inhomogeneous composition of

Figure 6. 13C NMR CP/MAS of crystalline PyO·HCl (top panel) and high-resolution (HR) spectra of PyO·HCl and PyO in AN solution (bottom panel). Theoretically predicted 13C NMR stick spectra are included for comparison. The numbers on peaks indicate positions of the carbon atoms in the PyO molecule.

The frequencies of the ν1 and ν2 modes are expected to be very sensitive diagnostic probes of the H-bond structure of the PyO·HCl complex. This expectation is based on the considerably increased difference between frequencies of the ν1 and ν2 modes in the FTIR spectra due to the proton transfer as predicted by our theoretical calculations (Figure 4). These two bands are resolved in the spectra of crystalline PyO·HCl and PyO·DCl·HCl (Figure 5), but they are found to overlap in the spectra of free PyO (Figure 2) and its complexes with weaker acids. The DFT/PCM calculations indicate proton transfer to occur in the PyO·HCl complex in the AN solution. Thereby, the presence of the absorption band at around 1443 cm−1 was also expected in the measured spectrum of PyO·HCl in AN solution, and yet, it was absent. Most probably this band is hidden under intense bands of AN at ∼1445 cm−1, and it was lost during the solvent band subtraction procedure. The frequencies of the νNO mode, 1198 cm−1 in the 0.1 M AN solution, 1196 cm−1 in the crystal from methanol, and 1198 cm−1 in the crystal from H2O, imply high similarity between Hbond patterns in these three samples. Thereby, FTIR measurements in general, and the νNO, ν1, and ν2 modes, in particular, do not provide definite answers concerning differences of the H-bond structures of the PyO·HCl complexes in liquid and solid phases. 8759



Article

AN/H2O/PyO/HCl where the aggregate of PyO·H+·Cl− acts as an ionic ingredient. This inhomogeneity of the sample is not always visually noticeable. However, even a small variation of temperature, i.e., cooling in the present case, renders the sample turbid, and consequently, the phase separation is provoked. Coexisting AN-rich and water-rich phases appear, in some cases with precipitation of the third phase (solid or gel-like). Similar behavior was observed also in some other aqueous solutions.43,44 FTIR spectra of the samples with critical compositions of PyO·HCl + 4H2O and PyO·HCl + 4D2O in 0.5−0.85 M AN solutions have been measured and analyzed. The ring and νNO modes were observed at 1478 cm−1 and 1201 cm−1, respectively, i.e., both of them are blue-shifted by ca. 3 cm−1 with respect to the corresponding signals in the water-free sample of PyO·HCl in AN solution (Figure 5). Theoretical modeling of the Py·MSA system in ref 48 indicates that water molecules in the vicinity of the PyO·MSA H-bond should stimulate proton transfer. Dependencies in Figures 3 and 4 in the region where rA−H/R(A···B) is larger than 0.5 imply, in fact, the opposite trend; that is, interionic distance at the composition and conditions of the sample close to critical can decrease. The changes in interionic interactions produce a supplementary contribution to the effect on the H-bond network of water molecules upon interaction with ions, thus inducing the phase separation processes in aqueous solutions.

Vibrational modes of ionic pairs originated via proton transfer in H-bond complexes can provide new information concerning the interionic interaction and its role in the phase separation and mezo-structuring processes.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: vytautas.balevicius@ff.vu.lt. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS One of us (V.B.) is thankful for the hospitality of the National Institute of Chemistry and Slovenian NMR Centre (Ljubljana, Slovenia), where a part of the experimental work was carried out. A special thanks is given to Professor Dušan Hadži for many guiding ideas, useful data, and discussions. Postdoctoral fellowship to K.A. from Research Council of Lithuania is kindly acknowledged.

■

REFERENCES

(1) XIX. International Conference on Horizons in Hydrogen Bond Research, September 12−17, 2011, Göttingen, Germany; see http:// www.hbond.de. (2) Herrebout, W. A.; Suhm, M. A. Phys. Chem. Chem. Phys. 2011, 13, 13858−13859. (3) Arunan, E.; Desiraju, G. R.; Klein, R. A.; Sadlej, J.; Scheiner, S.; Alkorta, I.; Clary, D. C.; Crabtree, R. H.; Dannenberg, J. J.; Hobza, P.; et al. Pure Appl. Chem. 2011, 83, 1619−1636. (4) Koeppe, B.; Tolstoy, P. M.; Limbach, H. H. J. Am. Chem. Soc. 2011, 133, 7897−7908. (5) Pyridine and Its Derivatives; Abramovich, R. A., Ed.; WileyInterscience: New York, 1974; Vol. 2. (6) Szafran, M. J. Mol. Struct. 1996, 381, 39−64. (7) Stare, J.; Hartl, M.; Daemen, L.; Eckert, J. Acta Chim. Slov. 2011, 58, 521−527. (8) Balevicius, V.; Gdaniec, Z.; Aidas, K. Phys. Chem. Chem. Phys. 2009, 11, 8592−8600. (9) Balevicius, V.; Bariseviciute, R.; Aidas, K.; Svoboda, I.; Ehrenberg, H.; Fuess, H. Phys. Chem. Chem. Phys. 2007, 9, 3181−3189. (10) Balevicius, V.; Fuess, H. Lith. J. Phys. 2005, 45, 241−247. (11) Stare, J.; Jezerska, A.; Ambrozic, G.; Kosir, I.; Kidric, J.; Koll, A.; Mavri, J.; Hadzi, D. J. Am. Chem. Soc. 2004, 126, 4437−4443. (12) Panek, J.; Stare, J.; Hadzi, D. J. Phys. Chem. A 2004, 108, 7417− 7423. (13) Pirc, G.; Stare, J.; Mavri, J. J. Chem. Phys. 2010, 132, 224506(7). (14) Stare, J.; Panek, J.; Eckert, J.; Grdadolnik, J.; Mavri, J.; Hadzi, D. J. Phys. Chem. A 2008, 112, 1576−1586. (15) Furniss, B. S.; Hannaford, A. J.; Smith, P. W. G.; Tatchell, A. R. Vogel‘s Textbook of Practical Organic Chemistry; Pearson Education Ltd: Singapore, 1989. (16) OriginLab Corporation. http://www.OriginLab.com. (17) Aidas, K.; Marsalka, A.; Gdaniec, Z.; Balevicius, V. Lith. J. Phys. 2007, 47, 443−449. (18) Sheldrick, G. M. SHELX: A Program for Automatic Solution of Crystal Structures; University of Göttingen: Göttingen, Germany, 1997. (19) Sheldrick, G. M. SHELXL: A Program for Crystal Structure Refinement; University of Göttingen: Göttingen, Germany, 1997. (20) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A. et al. Gaussian 09, revision C.01; Gaussian, Inc.: Wallingford, CT, 2009. (21) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652. (22) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980, 72, 650−654. (23) Pirc, G.; Mavri, J.; Stare, J. Vib. Spectrosc. 2012, 58, 153−162.

■

CONCLUDING REMARKS Established correlations between FTIR frequencies of the PyO vibrational ring and N → O stretching modes, and geometric parameters of the H-bond are efficient tools for studying the fine effects in the systems of PyO acting as a base. Validity of these correlations for other derivatives of PyO is yet to be investigated. Compared to related systems of Py·acid,8,10 certain differences have been observed. Both the FTIR and the 13C NMR data indicate that differences between the πelectronic structures of the two molecules may mainly arise due to the presence of the N → O group in PyO. In particular, the electronic structure of the π-electronic ring in Py is under direct influence of the H-bonding proton attacking the nitrogen atom. In contrast, the effect of the H-bonding on the properties of the PyO ring is mediated by the presence of the NO moiety. This can be also related to phenomena where N-oxide groups exhibit remarkable acidifying effects on the neighboring protons.7 Geometric isotope effect on the ionic H-bond structure of PyO·H(D)Cl was found to be extremely prominent. Since PyO·HCl and PyO·DCl·HCl crystals have been obtained either from H2O or D2O or methanol solutions, the amount and the distribution of water can be rather different in these cases. The role of the residual water in these crystals should be examined more carefully. FTIR, CP/MAS, and high-resolution 13C NMR data implies a high degree of similarity between the PyO·HCl complexes in the crystalline phase and in very polar solvents. The H-bond in solution, however, could be more ionic than in the crystal. This in turn indicates that requirements for close molecular packing get softer and that more degrees of freedom appear for molecules and molecular complexes in solution. Thus, the proton transfer can be deeper. Those more ionic structures can be stabilized by the reaction field created by solvent molecules in the first solvation shell although the long-range ordering of their dipole moments is absent. 8760



Article

(24) Adamo, C.; Barone, V. J. Chem. Phys. 1999, 110, 6158−6170. (25) Helgaker, T.; Jaszuński, M.; Ruud, K. Chem. Rev. 1999, 99, 293− 352. (26) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105, 2999−3093. (27) Ruokolainen, J.; Mäkinen, R.; Torkkeli, M.; Mäkelä, T.; Serimaa, R.; ten Brinke, G.; Ikkala, O. Science 1998, 280, 557−560. (28) Ruokolainen, J.; Torkkeli, M.; Serimaa, R.; Vahvaaselkä, S.; Saariaho, M.; ten Brinke, G.; Ikkala, O. Macromolecules 1996, 29, 6621−6628. (29) Schreiber, V. M.; Shchepkin, D. N.; Sokornova, T. V. J. Mol. Struct. 1994, 322, 217−221. (30) Hadži, D. Private communication. (31) Mielke, Z. J. Phys. Chem. 1984, 88, 3288−3292. (32) Langner, R.; Zundel, G. J. Chem. Soc., Faraday Trans. 1995, 91, 3831−3838. (33) Olovsson, I.; Jönsson, P. G. In The Hydrogen Bond, Recent Developments in Theory and Experiments; Schuster, P., Zundel, G., Sandorfy, C., Eds.; North-Holland: Amsterdam, The Netherlands, 1976; pp 393−456. (34) Ichikawa, M. J. Mol. Struct. 2000, 552, 63−70. (35) Katrusiak, A. Cryst. Rev. 1996, 5, 133−180. (36) Majerz, I.; Malarski, Z.; Lis, T. J. Mol. Struct. 1990, 240, 47−58. (37) Ford, S. J.; Delamore, O. J.; Evans, J. S. O.; McIntyre, G. J.; Johnson, M. R.; Radosavljević Evans, I. Chem.Eur. J. 2011, 17, 14942−14951. (38) de Souza, J. M.; Freire, P. T. C.; Bordallo, H. N.; Argyriou, D. N. J. Phys. Chem. B 2007, 111, 5034−5039. (39) Golubev, N. S.; Shenderovich, I. G.; Smirnov, S. N.; Denisov, G. S.; Limbach, H. H. Chem.Eur. J. 1999, 5, 492−497. (40) Kazlauskas, K.; Miasojedovas, A.; Dobrovolskis, D.; Arbačiauskienė, E.; Getautis, V.; Šačkus, A.; Juršeṅ as, S. J. Nanopart. Res. 2012, 14, 877. (41) Hao, E.; Meng, T.; Zhang, M.; Pang, W.; Zhou, Y.; Jiao, L. J. Phys. Chem. A 2011, 115, 8234−8241. (42) Pfenning, D.; Woermann, D. J. Membr. Sci. 1987, 32, 105−116. (43) Balevicius, V.; Fuess, H. Phys. Chem. Chem. Phys. 1999, 1, 1507− 1510. (44) Jacob, J.; Anisimov, M. A.; Sengers, J. V.; Oleinikova, A.; Weingärtner, H.; Kumar, A. Phys. Chem. Chem. Phys. 2001, 3, 829− 831. (45) Balevicius, V.; Weiden, N.; Weiss, A. Ber. Bunsenges. Phys. Chem. 1994, 98, 785−792. (46) Wagner, M.; Stanga, O.; Schröer, W. Phys. Chem. Chem. Phys. 2004, 6, 580−589. (47) Kostko, A. F.; Anisimov, M. A.; Sengers, J. V. Phys. Rev. E 2004, 70, 026118(11). (48) Lehtonen, O.; Hartikainen, J.; Rissanen, K.; Ikkala, O.; Pietilä, L. O. J. Chem. Phys. 2002, 116, 2417−2424.

8761


Hydrogen Bonding in Pyridine - American Chemical Society

Recommend Documents