On the change of the NMR isotropic shielding in blue- and red-shifted complexes of F3CH

Chemical Physics Letters 441 (2007) 194–197 www.elsevier.com/locate/cplett

On the change of the NMR isotropic shielding in blue- and red-shifted complexes of F3CH Sean A.C. McDowell

*

Department of Biological and Chemical Sciences, The University of the West Indies, Cave Hill Campus, Barbados Received 10 March 2007; in final form 5 May 2007 Available online 10 May 2007

Abstract The change in the 1H and 13C isotropic magnetic shielding constant of F3CH on formation of the H-bonded F3CH  Y complexes, Y = F, Cl, NH3, N(CD3)3, FH, ClH, OH2, SH2, N2, CO, OCO, Ar, NCH and O(CH3)2, were determined by GIAO MP2 ab initio and DFT computations. A good correlation between the sign of frequency shift and the change of the proton magnetic shielding was obtained, i.e., red-shifted complexes show the characteristic downfield shift of the isotropic proton magnetic resonance while in blueshifted complexes the resonance is shifted upfield or shifted downfield but to a much smaller extent than in red-shifted complexes. Similar results were obtained for the 13C shifts. Counterpoise correction and basis set variation of the magnetic shielding were also considered.  2007 Elsevier B.V. All rights reserved.

1. Introduction Blue-shifting hydrogen bonding has been a subject of considerable interest in recent years since it has different features to those normally associated with hydrogen bonds. A hydrogen bond may be represented symbolically by X–H  Y where X is an element or fragment with a high electron affinity and Y is a molecule containing a region of high electron density. In other words, X–H represents the proton donor and Y the proton acceptor. Normal features of hydrogen bonding include a diminution of the X–H stretch (a red shift), extension of the X–H bond and an increase in the infrared intensity of the X–H stretch [1]. These features are usually reversed for blue-shifted hydrogen bonds, i.e., blue shift, X–H bond contraction and infrared intensity decrease are observed; see Ref. [2] for a review. In this Letter, significant differences in 1H and 13C NMR isotropic shielding constants of blue-shifted complexes relative to red-shifted complexes are investigated since the features of the former species are quite different from those of the latter. A downfield shift of the proton magnetic res*

Fax: +1 246 417 4325. E-mail address: [email protected]

0009-2614/$ - see front matter  2007 Elsevier B.V. All rights reserved. doi:10.1016/j.cplett.2007.05.029

onance, ranging up to about 5 ppm for an X–H molecule going from the gas to liquid phase, is typical [3,4]. However, in a recent study of the five linear ClH  Y dimers (Y = N2, CO, OC, BF, FB) it was noticed that the two blue-shifting species ClH  OC and ClH  FB have much smaller downfield shifts on complexation [5]. Therefore, it is natural to enquire whether this observation is more general and whether it holds for all blue-shifted complexes. We undertook a study of some complexes of trifluoromethane CF3H, which is a prototypical proton donor for blue-shifted hydrogen bonds but which also forms some red-shifted complexes. The proton acceptor in these complexes was varied to see what effect this has on the magnetic shielding with respect to the uncomplexed CF3H molecule. We studied fourteen CF3H  Y complexes, where Y = F, Cl, NH3, N(CD3)3, FH, ClH, OH2, SH2, N2, CO, OCO, Ar, NCH and O(CH3)2, with the first four complexes being red-shifted and the remaining ten blue-shifted. Experimental data for the frequency shifts of some of these complexes are available for comparison with our computational results. For example, experimental FTIR studies of the spectra of CF3H dissolved in liquid Ar, N2, CO and CO2 show that the C–H stretching vibration of trifluoromethane is blue-shifted on complexation and is accompanied by a

S.A.C. McDowell / Chemical Physics Letters 441 (2007) 194–197

decrease in the corresponding v1 band [6,7]. Also, the C–H stretch of the complex of trifluoromethane and dimethyl ether in liquid argon appeared 17.7 cm1 above that of the monomer in the IR spectrum [8]. The main aim of this study was to determine any correlation between the sign and magnitude of the computed frequency shift and the sign and magnitude of the change in the magnetic properties. We also looked specifically at the performance of the B3LYP functional of density functional theory (DFT) in the computation of the magnetic properties since this approach allows us to use larger basis sets which would be too computationally demanding for MP2 with the resources available to the author. Previous ab initio computations of the proton isotropic chemical shielding using the B3LYP functional with modest basis sets (like 6-31+G** and 6-311G**) for C–H  O complexes [9] and in N-methyl maleimide-DMSO [10] were found to be in good agreement with MP2 and experiment. The proton shielding tensor in weakly bound X–H  Y complexes can be approximated by Hð0Þ

H rH ab ¼ rab þ drab

ð1Þ

Hð0Þ rab

where is the shielding tensor for the uncomplexed X–H molecule and drH ab is the change in the shielding tensor due to complexation. The isotropic shielding, defined as r = (rxx + ryy + rzz)/3, is the observable of interest here. Eq. (1) allows us to evaluate changes in the isotropic shielding in red- and blue-shifted complexes by taking the difference between the ab initio values for the proton shielding tensors for X–H  Y and X–H. The 13C shielding tensors were computed in similar fashion. The GIAO (gaugeinvariant atomic orbitals) method [11–13] was used to calculate the magnetic properties. The significant contributions to drH ab come from the change in the tensor due to the non-uniform electric field, including those due to high-frequency   fluctuations, arising HðelÞ from the proton acceptor Y rab , and from the magnetization  induced in Y by the external magnetic field HðmagÞ . Therefore, drH rab ab can also be approximated by [14], drH ab ¼

HðelÞ rab

þ

HðmagÞ rab

ð2Þ

For the monoatomic (F, Cl, Ar) and the linear proton acceptors (like N2, CO and OCO), the magnetic contribuHðmagÞ tion rab can be estimated by computing the magnetizability, n, of Y. Assuming that the magnetization of Y can be represented as n times the magnetic field then, in the long-range limit [4] that ignores contributions from higher induced magnetic multipole moments (like magnetic quadrupoles, etc.) [15], HðmagÞ

rab

¼ ðl0 =4pÞT ac ncb

ð3Þ

where l0 is the permeability of a vacuum (l0/ 4p = 107 JA2 m1) and Tac = (3RaRc  R2dac)/R5. R is the distance from an origin in Y (either the nucleus of the monoatom or the centre of the linear molecule) to the H nu-

195

cleus. For the linear molecules, there are two independent components of the magnetizability, nk and n^, so the proton magnetic shielding is given by rHðmagÞ ¼ ðl0 =4pÞð2nk =R3 Þ zz 3 HðmagÞ HðmagÞ and rxx ¼ ryy = (l0/4p)(n^/R ) [4]. From the comHðmagÞ HðelÞ , rab can be estimated puted values of drH ab and rab using Eq. (2). 2. Computational procedure All calculations were performed with the GAUSSIAN 03 suite of programs [16]. The uncomplexed CF3H molecule and its complexes were first optimized to stable structures at the MP2(full)/6-31+G(d,p) level of theory and the harmonic vibrational frequencies of the C–H stretch calculated, from which the frequency shift (Dx) was approximated as Dx = xcomplex  xmonomer. The NMR computations were performed using the GIAO method with the same basis set. The computed changes in the C–H bond length, harmonic vibrational frequency and 1 H and 13C NMR isotropic shielding are shown in Table 1. The changes in molecular properties on complexation may be affected by the basis set superposition error (BSSE), which is usually corrected by the Boys–Bernardi counterpoise correction [17]. The BSSE is diminished by increasing the size of the basis set and should tend to zero in the limit of an infinite basis set. We examined the basis set variation of the magnetic shielding in one of the complexes, CF3H  FH, to assess the suitability of the 6-31+G(d,p) basis set for computing magnetic properties. The results for five basis sets of systematically increasing size are shown in Table 2. We also applied the counterpoise correction to the MP2 shielding with the 6-31+G(d,p) basis. Table 1 The change in the GIAO 1H and 13C isotropic NMR shielding constants ˚ ) and the C–H (drH, drC, in ppm), the C–H bond-length (drCH, in A harmonic vibrational stretch (dx, in cm1) for CF3H on formation of CF3H  Y at the MP2(full)/6-31+G(d,p) level drH

drC

CF3H  F CF3H  Cl CF3H  NH3 CF3H  N(CD3)3

5.982 2.142 0.387 0.022

10.325 3.887 1.988 2.118

CF3H  FH CF3H  ClH CF3H  OH2 CF3H  SH2 CF3H  N2 CF3H  CO CF3H  OCO CF3H  Ar CF3H  NCH CF3H  O(CH3)2

1.776 1.858 0.594 0.841 2.189 2.182 1.924 1.96 1.637 0.872

0.628 0.08 1.405 0.759 0.244 0.121 0.417 0.043 1.024 1.627

CF3H  Y 

dx

dxexpt a

0.0482 0.0078 0.0009 0.0034

712.7 123.4 14.5 60.9

12 22

0.0009 0.0007 0.0016 0.0012 0.001 0.001 0.0015 0.0007 0.0016 0.0009

+18.4 +13.1 +33.1 +19.3 +19.1 +20.8 +30.3 +14.8 +32.6 +18.9

drCH

b

+14.1 +19.1 +23.6 +2.8 +17.7

c

Experimental frequency shifts (dxexpt) are included where available. For the uncomplexed CF3H molecule: rH = 23.3513 ppm, rC = 83.8808 ppm, ˚ , xe(C–H) = 3272.5 cm1. re(C–H) = 1.0839 A a Ref. [6]. b Ref. [19]. c Ref. [8].

196

S.A.C. McDowell / Chemical Physics Letters 441 (2007) 194–197

Table 2 Basis set variation of drH and drC for CF3H  FH at the MP2(full) level of theory Basis set

drH (ppm)

drC (ppm)

6-31+G(d,p) 6-31++G(d,p) 6-311++G(d,p) 6-311++G(2d,2p) 6-311++G(2df,2pd)

1.776 0.226 2.749 0.241 0.247

0.628 0.652 0.696 0.814 0.800

Table 4 Comparison of the isotropic electric (rH(el)) and magnetic (rH(mag)) components of the isotropic proton shielding, and the change in the proton isotropic shielding (drH) for CF3H  Y complexes with a linear C– H  Y H-bond and with Y being either an atom or a linear molecule CF3H  Y 

B3LYP shielding values with this basis set and the much larger 6-311++G(2df,2pd) basis were also computed. These results are in Table 3. Finally, a comparison of the electric and magnetic components of the isotropic shielding for selected CF3H complexes are displayed in Table 4 and may shed light on the trends that are evident from the preceding tables.

CF3H  F CF3H  Cl CF3H  Ar CF3H  N2 CF3H  CO CF3H  OCO

rH(el)

rH

(mag)

8.433 5.166 0.126 0.049 0.119 0.179

0 0 0 0.302 0.31 0.155

drH 8.433 5.166 0.126 0.253 0.191 0.024

3. Discussion

The following approximation was used, drH = rH(el) + rH(mag); drH is taken from the B3LYP/6-311++G(2df,2pd) computed values in Table 3, rH ðmagÞ ¼ ð2rHðmagÞ þ rHðmagÞ Þ=3 is obtained from computed magnetizxx zz abilities and rH(el) is the difference between drH and rH(mag). The B3LYP/ 6-311++G(2df,2pd) values for the magnetizabilities are: niso (F) = 2.7649 a.u.; niso (Cl) = 6.5187 a.u.; niso (Ar) = 4.3697 a.u.; nxx (N2) = 1.9851 a.u., nzz (N2) = 3.8385 a.u., niso (N2) = 2.6029 a.u.; nxx (CO) = 2.0598 a.u., nzz (CO) = 3.8376 a.u., niso (CO) = 2.6524 a.u., nxx (OCO) = 4.1988 a.u., nzz (OCO) = 5.5348 a.u., niso (OCO) = 4.6441 a.u.

The majority of the optimized CF3H  Y complexes, Y = F, Cl, NH3, N(CD3)3, FH, ClH, OH2, SH2, N2, CO, OCO, Ar, NCH and O(CH3)2 have linear or nearly linear C–H  Y bonds, i.e., with \C–H  Y . 180 and Y represents the atom of the proton acceptor closest to the H atom of CF3H. The C–H  Y bonds that are significantly nonlinear are in CF3H  N(CD3)3 (163), CF3H  O(CH3)2 (158), CF3H  ClH (118) and CF3H  FH (113); in the last two complexes there are two H-bonded interactions, C–H  F/Cl and F–H  F, which are in the same plane and form a 5-membered ring structure. The intermolecular H  Y separations range in ˚ (for CF3H  F) to 3.0377 A ˚ magnitude from 1.5372 A (for CF3H  ClH). Table 1 shows that the modest 6-31+G(d,p) basis set gives reasonably good agreement between experimental and theoretical frequency shifts, except for CF3H  N(CD3)3

and CF3H  Ar, where the computed shifts were overestimated. The expected correlation of C–H red shift/bond extension and blue shift/bond contraction is evident throughout, with the largest changes obtained for CF3H  F and the smallest for CF3H  ClH. However, a remarkable correlation between the change in the proton isotropic NMR shielding constant and the type of frequency shift is noticeable. The four red-shifted complexes show a downfield shift in the proton magnetic resonances, while the blue-shifted complexes show an upfield shift (except for CF3H  O(CH3)2). The 13C resonances, on the other hand, all show a downfield shift on complexation, but here also there is a distinction between the red- and blue-shifted species since the downfield shift for the latter species is generally smaller in magnitude than for the former.

Table 3 Changes in the GIAO 1H and 13C isotropic NMR shielding constants (drH, drC, in ppm) for CF3H on formation of CF3H  Y at the MP2(full)/631+G(d,p) level with counterpoise-correction (I), at the B3LYP/6-31+G(d,p) (II) and B3LYP/6-311++G(2df,2pd) (III) levels without counterpoise correction CF3H  Y

drH(I)

drH(II)

drH(III)

drC(I)

drC(II)

drC(III)

CF3H  F CF3H  Cl CF3H  NH3 CF3H  N(CD3)3 CF3H  FH CF3H  ClH CF3H  OH2 CF3H  SH2 CF3H  N2 CF3H  CO CF3H  OCO CF3H  Ar CF3H  NCH CF3H  O(CH3)2

6.687 3.794 2.304 1.881 0.213 0.14 1.386 1.159 0.178 0.168 0.086 0.061 0.357 1.108

7.864 4.493 2.465 2.097 0.236 0.166 1.416 1.221 0.294 0.193 0.045 0.066 0.239 1.172

8.433 5.166 2.38 2.09 0.258 0.209 1.342 1.209 0.253 0.191 0.024 0.126 0.37 1.159

2.757 0.329 0.456 0.664 0.133 0.203 0.346 0.201 0.158 0.149 0.072 0.042 0.004 0.405

9.951 3.266 2.015 2.149 0.694 0.029 1.478 0.72 0.119 0.023 0.304 0.028 0.919 1.633

10.318 3.322 1.806 2.538 0.897 0.207 1.406 0.611 0.026 0.043 0.32 0.004 0.89 1.78

The MP2(full)/6-31+G(d,p) optimized geometries were used.

S.A.C. McDowell / Chemical Physics Letters 441 (2007) 194–197

The basis set variation of the magnetic shielding shown in Table 2 for CF3H  FH prompted us to apply the counterpoise correction to the computation of dr, as well as using a larger basis set, so that a reasonable level of confidence can be established for the observed trends in the magnetic properties. The results show that the 1H resonance is very sensitive to basis set while the 13C resonance is less so. For example, drH changes sign and drops by an order of magnitude in going from the 6-31+G(d,p) basis to the largest 6-311++G(2df,2pd) basis. By contrast, drC goes from 0.628 ppm to 0.800 ppm. It is interesting to note that a previous theoretical study of the gas-phase magnetic properties of fluoromethanes, including CF3H, found that the proton shieldings were relatively insensitive to basis set size [18]. Table 3 shows that counterpoise correction of the MP2/ 6-31+G(d,p) proton shielding gives good agreement with the uncorrected B3LYP/6-31+G(d,p) and B3LYP/ 6-311++G(2df,2pd) calculations; the latter calculation was considered to be relatively more accurate since a much larger basis set was used. The correlation between the change in vibrational frequency and magnetic shielding is still evident but now only CF3H  N2 and CF3H  CO show upfield shifts while the other complexes show downfield shifts. Nonetheless, there is still a clear distinction between the magnitude of drH and the sign of dx. The red-shifted complexes show a decrease in rH by more than about 1.8 ppm, whereas for all of the blue-shifted complexes, rH is never decreased by more than 1.4 ppm (and in most complexes by much less) or a small increase is predicted. Similar, though less pronounced, trends for the 13C resonances are also noticeable. Beyond the general trend elucidated above it is difficult to find an all-embracing reason for the variation in the magnitude of drH. For example, the C–H  N interaction yields B3LYP/6-311++G(2df,2pd) drH values of 2.38 and 2.09 ppm, respectively, for the red-shifted CF3H  NH3 and CF3H  N(CD3)3 species, and 0.253 and 0.370 ppm, respectively, for the blue-shifted CF3H  N2 and CF3H  NCH species. Partitioning of the isotropic shielding into its electric and magnetic components for complexes with a linear C–H  Y contact (Table 4) shows that in the blue-shifted complexes, the magnetic component is positive and either larger than or comparable in size to the negative electric component, thereby yielding a change in the shielding which is either positive or negative but small in magnitude. By contrast, for the red-shifted CF3H  F and CF3H  Cl species, the magnetic compo-

197

nent is isotropic (zero in magnitude) so that the large negative electric component (arising from the electric field of the halide anion) dominates the shielding. In the blueshifted CF3H  Ar, the magnetic component is also isotropic but the electric component is smaller since the electric field arising from Ar is due to the polarization induced in the atom by the dipole moment of the CF3H molecule. Therefore, it yields a smaller negative magnetic shielding. The present work suggests that there is a significant correlation between the type of frequency shift and the change in the proton (and 13C) isotropic magnetic shielding in Hbonded complexes. It is remarkable that the modest 631+G(d,p) basis set predicts striking differences between the red and blue-shifted complexes along the lines of downfield and upfield shifts, respectively. The difference manifests itself as a difference in relative magnitudes of the downfield shifts with the use of a larger basis set. It would be interesting to extend this investigation to a wider range of H-bonded species. References [1] S. Scheiner, Hydrogen Bonding, Oxford University Press, New York, 1997. [2] P. Hobza, Z. Havlas, Chem. Rev. 100 (2000) 4253, and references contained therein. [3] G.C. Pimentel, A.L. McLellan, The Hydrogen Bond, Freeman, San Francisco, 1960. [4] J.A. Pople, W.G. Schneider, H.J. Bernstein, High-resolution Nuclear Magnetic Resonance, McGraw-Hill, New York, 1959. [5] S.A.C. McDowell, A.D. Buckingham, Mol. Phys. 104 (2006) 2527. [6] S.M. Melikova, K.S. Rutkowski, P. Rodziewicz, A. Koll, Chem. Phys. Lett. 352 (2002) 301. [7] S.M. Melikova, K.S. Rutkowski, P. Rodziewicz, A. Koll, J. Mol. Struct. 705 (2004) 49. [8] S.N. Delanoye, W.A. Herrebout, B.J. van der Veken, J. Am. Chem. Soc. 124 (2002) 7490. [9] S. Scheiner, Y. Gu, T. Kar, J. Mol. Struct. (THEOCHEM) 500 (2000) 441. [10] B. Wang, J.F. Hinton, P. Pulay, J. Phys. Chem. A 107 (2003) 4683. [11] F. London, J. Phys. Radium 8 (1937) 397. [12] R. Ditchfield, Mol. Phys. 27 (1974) 789. [13] K. Wolinski, J.F. Hinton, P. Pulay, J. Am. Chem. Soc. 112 (1990) 8251. [14] A.D. Buckingham, Y. Tantirungrotechai, Mol. Phys. 96 (1999) 1217. [15] A.D. Buckingham, P.J. Stiles, Mol. Phys. 24 (1972) 99. [16] GAUSSIAN 03, Revision A.7, M.J. Frisch et al., Gaussian, Inc., Pittsburgh PA, 2003. [17] S.F. Boys, F. Bernardi, Mol. Phys. 19 (1970) 553. ˚ strand, K.V. Mikkelson, K. Ruud, T. Helgaker, J. Phys. [18] P.-O. A Chem. 100 (1996) 19771. [19] A. Kaldor, I.M. Mills, Spectrochim. Acta A 20 (1964) 523.