J. Phys. Chem. 1987, 91, 6464-6466
6464
2H "IR Line-Shape Studies of D,O,+CF,SO,C. I. Ratcliffe Division of Chemistry, National Research Council of Canada, Ottawa, Ontario, Canada K l A OR6 (Received: August 18, 1987)
2H NMR line shapes of D502+CF3S03have been obtained as a function of temperature. The line shapes very clearly distinguish the central, strongly hydrogen-bonded 2H atom (e2qQ/h = 43.92 kHz, 7 = 0) of the DSO2' ion from the four peripheral 2H atoms (e2qQ/h= 210.0 kHz, 7 = 0.124), and confirm the very short 0---0distance determined from X-ray diffraction studies. The first motion detected as the temperature increases is exchange of the deuterons within the two "D20" units at each end of the ion. These observations point to the integrity of the DSO2+ ion in the compound rather than a more loosely hydrogen-bonded oxonium hydrate, D30f.D20.
Introduction Although a number of X-ray and neutron diffraction structural studies have been performed on solids containing nominally H502+ groups,' little work has so far been done to characterize this species using other techniques. The principal features of a true HS02+ ion are a very short, almost linear, 0--H--0 hydrogen bond, in which the H atom may or may not be centered, connecting two "H20" units, and a squat pyramidal configuration of H atoms around each oxygen. There is considerable flexibility in conformation from trans to cis and gauche forms involving the relative orientations of the planes of the H 2 0 units at each end of the molecule. There is also a good deal of variation in the 0--H- -0 distances and indeed those cases where 0-- -0> 2.48 A and 0-H < 1.10 A are better described as H30+-H20.It is clear that the particular conformation of H502+in any one material is largely determined by interactions with neighboring anions or, in some cases, by additional H 2 0 molecules. The short, and therefore strong, hydrogen bond might also be e x p t e d to strongly influence other properties of the ion. The example studied here by means of 2H N M R line shapes is a good example of H502+with a very short central hydrogen bond. The X-ray diffraction studies2 at 85 K have obtained an 0-- -0distance of 2.409 A for the central hydrogen bond. The conformation is gauche and the four peripheral H s are hydrogen bonded to 0 atoms of the CF3S03- ions with 0-- -0distances 2.652-2.776 A. (Attempts to refine the positions of the H atoms placed the central H atom at 1.11 A from the nearest 0 atom, though not much confidence was placed in this result.) The present study was undertaken in order to verify the presence of the very short hydrogen bond and also to determine the nature of any motions of the HS02+ion which might be occurring. 2H N M R line shapes should be a sensitive probe of both properties. (Preliminary results of these experiments were mentioned in a review.') Previous N M R studies of systems thought to contain H5O2' (or D5O2+) are few in number (these were reviewed in ref 1 ) and the evidence concerning motions is in most cases rather scant or ambiguous. The only *H N M R results to date are for materials containing water molecules in addition to the D5O2+, Le., DAuCl4.4D,O3 and DU02P04.4D20(DUP) and its arsenate analogue ( D U A S ) . ~ These results were principally of the highest temperature phases and all indicated equivalence of all deuterons due to their rapid diffusion among all the types of lattice site. In all three materials the D 2 0 molecules and D502+are interlinked into a continuous hydrogen-bonded network. Since this situation does not occur in D502CF3S03,rapid 2H diffusion would not be expected. Experimental Section D502+CF3S03-was prepared by mixing appropriate amounts of D 2 0 (MSD Isotopes) and the acid anhydride (CF3S02)20 (Aldrich). 2HN M R line shapes were obtained with a Bruker CXP 'Issued as NRCC No. 28404.
0022-3654/87/2091-6464$01 .50/0
TABLE I: Quadrupole Coupling Constants and Asymmetry Parameters
e2qQ/h,kHz
7
210.0 43.92 111.5 47.87
0.124 0 1
150 K
240 K
0
180 spectrometer operating at a frequency of 27.63 MHz and a variable-temperature N2 gas-flow probe with a Bruker B-VT-1000 temperature controller. The phase-alternated quadrupole echo technique5 was used with pulse widths of 2.6 ps (-78O flip angle) and echo pulse spacing of 35 ps. Dwell times of 0.5 or 1.0 ps were used to record 500 point quadrupole echoes which were then zero filled to 4 K before Fourier transformation. The cycle repetition rates varied from 1 to 200 s depending on the spin-lattice relaxation time.
Results and Discussion Deuterium powder line shapes show either three or two pairs of features separated by the frequencies6 Avzz = vq (1) Avyy = f/zv,(l
+ 7)
(2)
(3) where the quadrupole coupling constant (e2qQ/h)= 2 / 3 v q ,and the asymmetry parameter 7 = (Avyy- Avxx)/Avzz.Av,, and Avyy become equal for 7 = 0. Avyyand Avzz become equal and Av,, = 0 for 7 = 1. See Figure 1. The 2H N M R line shapes of D502+CF3S03-at different temperatures are shown in Figure 2. At 150 K the line shape must correspond to an essentially rigid D5O2' group since there are no changes in shape below this temperature and the relaxation time is clearly very long (long recovery times were required between acquisition cycles). One can immediately see that the line shape consists of two components; a broad 7 # 0 component and a narrow, = 0 component. This is exactly what one would anticipate on the basis of correlations which show that quadrupole coupling constants of 2Hin hydrogen bonds decrease dramatically as the bond length becomes very short (f0r.a review see ref 7). Figure 3a,b shows the observed line shape and a simulation based ~
~~~
~~~
~~
~
~
~
(1) Ratcliffe, C. I.; Irish, D. E. In WaferScience Reviews; Franks, F., Ed.; Cambridge University Press: Cambridge, U.K., 1986; Vol. 2, p 149, and references cited therein. (2) Delaplane, R. G.; Lundgren, J.-0.; Olovsson, I. Acta Crysfallogr.,Sect. B. 1975, 831, 2202. (3) O'Reilly, D. E.; Peterson, E. M.; Scheie, C. E.; Williams, J. M. J . Chem. Phys. 1971, 55, 5629. (4) Halstead, T. K.; Boden, N.; Clark, L. D.; Clarke, C. G. J . Solid Stare Chem. 1983, 47, 225. (5) Davis, J. H.; Jeffrey, K. R.; Bloom, M.; Valic, M. I.; Higgs, T. P. Chem. Phys. Left. 1976, 42, 390. (6) Barnes, R. G. Advances in Nuclear Quadrupole Resonance; Smith, J. A. S., Ed.; Heyden: London, 1974; Vol. 1, p 335. (7) Weiss A.; Weiden, N. Advances in Nuclear Quadrupole Resonance; Smith, J. A. S., Ed.; Heyden: London, 1980; Vol. 4, p 226.
Published 1987 by the American Chemical Society
The Journal of Physical Chemistry, Vol. 91, No. 26, 1987 6465
Letters
JWI
Figure 1. 2H NMR line shapes for (a) y = 0, (b) y = 0.2, and (c) y = 1, indicating the frequency separations AvXx,AU,, and A s r . I I
d)
200 kHz
Figure 3. Comparison of experimental and calculated 2H NMR line shapes of D502+CF3S03in the static (a, b) and fast motion (c, d) limits. Standard powder line-shape function^^^'^ were used to generate patterns, with features of the appropriatewidths, which were then convoluted with a Gaussian broadening function of 4.92 kHz half-width to produce (b) and (d). A simple Fortran program was written and implemented on a VAX 11/780. Other details of the calculations are given in the text.
I
i
1
200 kHz
200 kHz
Figure 2. *H NMR line shapes (27.63 MHz) of D5O2+CF3SOC as a
function of temperature. on a 4:l intensity ratio of the broad:narrow components with their features adjusted for best agreement with the experimental line shape. The narrow component was assumed to be axial, 7 = 0. The match is remarkably good. The quadrupole coupling constants and 7's are given in Table I. If one inserts the (e2qQ/h)values into the correlations given by Berglund et a1.8
(e2qQ/h) = 271 - 8.63
X
lo5 e ~ p ( - 3 . 4 8 R ~ . . - ~ ) (4)
(e2qQ/h) = 303 - 521/R30..D
(5)
(where the ( g q Q / h ) is in kHz and the R in A), one obtains values of RS-& = 2.369 A, and &--D = 1.262 A, for the central strong hydrogen bond (cf. X-ray2 2.409 and 1.30 A, respectively) and Ro-.o = 2.746 A (and Ro. .D = 1.776 A) for the average of the four peripheral hydrogen bonds (cf. X-ray Ro.. .o = 2.652-2.776 A, Ro. .D not given in ref 2). This comparison is also very reasonable, considering the scatter of points used to derive the empirical relationships. This also helps to confirm that the 150 K line shape represents the static D502+group. Note that it was not possible to resolve individual components for the four peripheral deuterons of the D 2 0 end units from the powder line shape. Between 150 and 230 K the broad component of the line shape narrows considerably due to the onset of some motion while the narrow component is hardly changed. At 235 K the line shape (8) Berglund, B.; Lindgren, J.; Tegenfeldt, J. J . Mol.Srruct. 1978,43, 179. (9) Mehring, M. In NMR Basic Principles and Progress; Diehl, P., Fluck, E., Kosfeld, R., Eds.; Springer-Verlag: New York, 1976; Vol. 11, p 21. (10) Bloembergen, N.; Rowland, T. J. Acta Metallurgica 1953, 1 , 7 3 1 .
appears to be in the fast motion limit and can be simulated very well by the superimposition of a broad line shape with 7 = 1 and a narrow line shape (7 = 0) in the ratio 4:l (see Figure 3c,d). The (e2qQ/h)values required to match the observed spectrum are also given in Table I. This is particularly interesting because it clearly shows that the two D 2 0 units are undergoing similar motions which do not involve the central hydrogen bond; Le., the molecule does not appear to be behaving as D30+.D20. A comparison of the rigid (1 50 K) and motionally averaged (235 K) line shapes shows that Au,, for the rigid case is almost the same as Ad,, (=Adzz)for the averaged case. This immediately suggests that each D 2 0 unit is undergoing a flip which exchanges the positions of its two deuterons, a motion which is commonly found in crystalline hydrate^.^ For individual D 2 0 molecules it is normally safe to assume that for each 2H the principal qrZ component of the tensor is along the 0 - D bond and that the qy, component is at right angles to the molecular plane. This second assumption is likely not valid in the present case since the D 2 0 unit is moderately strongly bound to another 2H atom forming a pyramidal configuration around the oxygen. This complicates the equations describing the motionally averaged line shape considerably. A full analysis is given in the Appendix and from this one can estimate the bond angle of the D 2 0 units as 111.2O. This falls within the range of values found from neutron diffraction studies of other HS02+compounds containing the very short central hydrogen bond, 105.7-1 12.0° (see summary in ref 1). Returning to the narrow line shape component for the central 2H one notices from a comparison of the spectra at 150 and 235 K and from the simulations (Table I) that in fact there is a small increase in (e2qQ/h)a t the higher temperature. This can be rationalized in terms of an increase in the length of the hydrogen bond. The change in (e2qQ/h)would correspond to an increase in the Ro.. .o distance of 0.005 A, from Berglund et al.'s formula, eq 5. The effect may be as a result of the motions of the D 2 0 end units or perhaps is due to increased vibrational amplitudes at the higher temperature. At temperatures approaching the melting point of the sample (H5O2+CF3SO3melts at 267 K2) all the fine structure of the line
6466 The Journal of Physical Chemistry, Vol. 91, No. 26, 198 7
shape begins to collapse and a strong sharp central line appears.
Conclusion The results presented above provide very clear evidence of the existence of a very strong and short hydrogen bond in D502+CF3S0,/h = f167.25 and eQV’Jh = 0 kHz, that y’= 129.98’ and @ = 5 5 . 6 0 O . The calculated bond angle is thus 111.2O. Note that the absolute signs of the tensor components are not known from the 2H NMR, but this information is not necessary for the calculations. It should again be emphasized that the experimental static tensor values derive from the average quadrupole coupling constants and asymmetry parameters for four slightly inequivalent deuteron sites and this must introduce some error into the derived angles.