Integrated Raman intensities of the spectral contour arising from the intermolecular librational motions of pure water have been obtained in the temperature range of ∼10°—95°C. In addition, integrated intensities of nearly symmetric librational components centered near ∼475 and ∼710 cm−1 were obtained from manual contour analysis according to two components. However, contour analysis was also accomplished by means of a special‐purpose analog computer, and three Gaussian librational components having average frequencies of 439, 538, and 717 cm−1 were thus revealed. The total contour intensity, the manually determined component intensities, and the Gaussian component intensities were found to have the same temperature dependence, and that dependence was found to be in excellent quantitative agreement with the previously reported temperature dependence of the hydrogen‐bond‐stretching intensity [J. Chem. Phys. 44, 1546 (1966)]. Integrated Raman intensities of pure water were also obtained in the temperature range of 10°—90°C for the intramolecular valence and deformation contours in the spectral region of ∼2800–3900 cm−1, and near 1645 cm−1, respectively. The integrated intensity of the deformation contour was found to be nearly independent of temperature, but the total integrated intensity of the intramolecular valence contour was found to decrease with increasing temperature. However, heights of the high‐frequency portion of the intramolecular valence contour were observed to increase, whereas heights of the low‐frequency portion were observed to decrease at nearly the same rate, with increasing temperature. An isosbestic point was also found at approximately 3460 cm−1. Further, computer analysis revealed the existence of four Gaussian components having opposite temperature dependences in pairs—two intense valence components at ∼3247 and ∼3435 cm−1 were found to decrease in intensity with increasing temperature, and two weak components at ∼3535 and ∼3622 cm−1 were found to increase in intensity. Computer analysis of infrared absorbance spectra also revealed four Gaussian components at approximately 3240, 3435, 3540, and 3620 cm−1. The quantitative agreements involving temperature dependences of the intermolecular hydrogen‐bond‐stretching and librational intensities, as well as the intramolecular valence data, would appear to preclude models of water structure involving consecutive hydrogen‐bond breakage. Continuum models of water structure are also precluded by the inter‐ and intramolecular intensity dependences, and particularly by the isosbestic point in the intramolecular valence region, but a model involving an equilibrium between two forms of water is consistent with all of the data. The two forms refer to water molecules which have or have not surmounted a barrier arising from a partially covalent hydrogen‐bond potential of C2v symmetry, and they may be described as nonhydrogen‐bonded monomeric water, and as lattice water, respectively. Polarized argon‐ion‐laser—Raman spectra were also obtained in the intermolecular frequency region of the water spectrum, and the depolarization ratios of the intermolecular Raman bands were found to be in complete agreement with predictions from intermolecular C2v symmetry. Studies of the intramolecular valence region were also made with polarized mercury excitation, and the spectra were analyzed by the analog method. Short‐lived CS intramolecular perturbations were indicated by the observed depolarization ratios of the four Gaussian valence components. Accordingly, CS intramolecular valence perturbations occur in the lattice water, as well as in the nonhydrogen‐bonded water, but the perturbations are of little importance on the intermolecular time scale.

1.
G. E.
Walrafen
,
J. Chem. Phys.
40
,
3249
(
1964
).
2.
G. E.
Walrafen
,
J. Chem. Phys.
44
,
1546
(
1966
).
3.
G. E.
Walrafen
,
J. Chem. Phys.
43
,
479
(
1965
).
4.
G. E.
Walrafen
,
J. Chem. Phys.
36
,
1035
(
1962
).
5.
G. E.
Walrafen
,
J. Chem. Phys.
46
,
1870
(
1967
).
6.
In previous work,2 uncorrected as well as corrected inter‐molecular hydrogen‐bond‐stretching intensities were reported, and least‐squares equations were obtained for each of the cases. The least‐squares equation resulting from the uncorrected data is log10{I(t)/[0.320−I(t)]} = (1116.2/T)−3.7107. The least‐squares equation resulting from the corrected data is log10[fB/(1−fB)] = (1212.2/T)−4.2192. The effect of the corrections is small as evidenced by the small differences between the quantities on the right. Accordingly, the effect on the thermodynamic quantities is also small, e.g., the uncorrected ΔH° and ΔS° values are −5.1 kcal/mole and −17 cal/degmole, and the corrected ΔH° and ΔS° values are −5.6 kcal/mole and −19 cal/degmole. The librational intensity data, and in particular the librational frequency data, are not of sufficiently high accuracy to warrant the small corrections applied to the hydrogen‐bond‐stretching intensities. Therefore, the librational intensities can only be compared with the uncorrected least‐squares hydrogen‐bond‐stretching intensities.
7.
That the agreements are significant, and not the result of refractive index affecting the bands in the same way is confirmed later by observations indicating no significant temperature dependence in the intramolecular deformation band intensity, and opposite dependences in the intramolecular valence component intensities.
8.
K. E.
Larsson
and
U.
Dahlborg
,
Reactor Sci. Tech. (J. Nucl. Energy)
16
,
81
(
1962
).
9.
D. A.
Draegert
,
N. W. B.
Stone
,
B.
Curnutte
, and
D.
Williams
,
J. Opt. Soc. Am.
56
,
64
(
1966
).
10.
R. W.
Terhune
,
P. D.
Maker
, and
C. M.
Savage
,
Phys. Rev. Letters
14
,
681
(
1965
).
11.
The hyper‐Raman spectrum of water in the intermolecular librational region has recently been reinvestigated by P. D. Maker (private communication). The librational contours of spectra corresponding to octapolar or to dipolar selection rules were observed to be different, and the intensity at 450 cm−1 was observed to be large in the octapolar spectrum.
12.
S. J.
Cyvin
,
J. E.
Rauch
, and
J. C.
Decius
,
J. Chem. Phys.
43
,
4083
(
1965
).
13.
In a recent review article [
B. E.
Conway
,
Ann. Rev. Phys. Chem.
17
,
481
(
1966
)], Conway erroneously assumed ΔS° to be zero. Then from an early value for ΔH° of −12.5 kcal/mole1 he calculated a value of 10−9 for the fraction of nonhydrogen‐bonded monomeric water. Conway’s value of 10−9 is much too low, and his assumption contradicts the early Raman data, because a finite value of ΔS° was directly obtainable from them. The present value of −17 cal/degmole is more accurate than the early ΔS° value, but it is not zero. Further, the present value for the fraction of nonhydrogen‐bonded monomeric water is 0.59 at 25 °C, whereas the previous value was 0.44.
14.
G.
Némethy
and
H. A.
Scheraga
,
J. Chem. Phys.
36
,
3382
(
1962
).
15.
K.
Buijs
and
G. R.
Choppin
,
J. Chem. Phys.
39
,
2035
(
1963
).
16.
The process may also involve large deviations from hydrogen‐bond linearity, see,
J. A.
Pople
,
Proc. Roy. Soc. (London)
A205
,
163
(
1953
).
17.
A. H. Narten, M. D. Danford, and H. A. Levy, “X‐ray Diffraction Data on Liquid Water in the Temperature Range 4° to 200 °C,” ORNL‐3997, September 1966 (unpublished).
18.
J. W.
Schultz
and
D. F.
Hornig
,
J. Phys. Chem.
65
,
2131
(
1961
).
19.
Differences between the actual contour areas and the computer‐generated contour areas of this work, i.e., the residuals, were observed to be positive at all temperatures, and to be about 3%–5% of the actual areas. The major part of the residuals occurred in all cases near the low‐frequency tail of the valence contour near 2800–3000 cm−1, although another extremely small part occurred near the high‐frequency tail at approximately 3800 cm−1. Because the residual areas occurred near the contour tails, and because the residuals were positive, it is evident that the true component shapes deviated slightly from Gaussian, i.e., Gaussians slightly broadened near the tails would appear to provide better fits.
20.
E.
Fishman
and
P.
Saumagne
,
J. Phys. Chem.
69
,
3671
(
1965
).
21.
P. C.
Cross
,
J.
Burnham
, and
P. A.
Leighton
,
J. Am. Chem. Soc.
59
,
1134
(
1937
).
22.
W. R.
Busing
and
D. F.
Hornig
,
J. Phys. Chem.
65
,
284
(
1961
).
23.
T. T.
Wall
and
D. F.
Hornig
,
J. Chem. Phys.
43
,
2079
(
1965
).
24.
M.
Falk
and
T. A.
Ford
,
Can. J. Chem.
44
,
1699
(
1966
).
25.
The O‐D or O‐H stretching vibrations of (the solute) HOD decouple, because of differences in frequency, from the vibrations of (the solvents) H2O or D2O, respectively. The decoupling allows the HOD to act as a probe of the intermolecular structure of the solvent. However, in pure water the inter‐ and intramolecular vibrations are decoupled—the frequencies differ by about an order of magnitude. Consequently, the intermolecular spectrum and the uncoupled intramolecular HOD spectrum should tend to provide similar information, e.g., similar values for the fraction of nonhydrogen‐bonded molecules, and similar thermodynamic quantities.
26.
M. Falk and T. A. Ford, Symp. Mol. Structure Spectry., Columbus, 14–18 June 1965 (unpublished).
27.
J. D.
Worley
and
I. M.
Klotz
,
J. Chem. Phys.
45
,
2868
(
1966
).
28.
The shape of the infrared contour obtained in this work was found to be in excellent agreement with the contour shape reported by Terhune, Maker, and Savage.10 The infrared spectrum, Fig. 2 of Ref. 10, was enlarged photographically and compared visually with the present spectrum. No significant difference in shape could be detected. The enlargement was then analyzed into Gaussian components with the du Pont computer. The following component frequencies, half‐widths, and percentages were obtained: (1) 3240 cm−1,320 cm−1,45.5%; (2) 3420 cm−1,260 cm−1,33.0%; (3) 3530 cm−1,175 cm−1,12.0%; and (4) 3620 cm−1,150 cm−1,9.5%; (cf., the corresponding values of this work).
29.
Preliminary depolarization measurements employing 4765‐Å, argon‐ion‐laser excitation indicate that the entire contour is polarized, and that the corresponding uncorrected ρl value is significantly smaller than the value obtained with mercury excitation, e.g., <0.2.
30.
Y.
Kyogoku
,
Nippon Kagaku Zasshi
81
,
1648
(
1960
).
This content is only available via PDF.
You do not currently have access to this content.