Chemical Physics Letters 379 (2003) 427–431

www.elsevier.com/locate/cplett

Raman spectroscopic studies of the stretching bandfrom water up to 6 kbar at 290 K

Qiang Sun a,*, Haifei Zheng a, Ji-an Xu b, E. Hines c

a School of Earth and Space Science, Peking University, Beijing 100871, Chinab Geophysical Laboratory, Carnegie Institute of Washington, Washington 20008, USA

c Anvil Department, Charles and Colvard Ltd., Morrisville 27560, USA

Received 30 September 2002; in final form 15 July 2003

Published online: 17 September 2003

Abstract

Raman scattering studies of the stretching band from liquid water have been conducted up to 6 kbar at 290 K. It

shows that the ðv1Þmax decreases with increasing pressure initially and reaches the minimum at about 2 kbar, and in-

creases with higher pressure up to about 4 kbar, then decreases with increasing pressure up to 6 kbar. This is accordance

with the behavior of rOO at high pressure. Additionally, the influence of pressure on water structure is also discussed.

� 2003 Elsevier B.V. All rights reserved.

Water is the most ubiquitous and intriguing

fluid in nature. A comprehensive molecular theory

for water is needed for two reasons. First, this

substance is a major chemical constituent of ourplanet�s surface and as such it may have been in-

dispensable for the genesis of life. Second, it ex-

hibits a fascinating array of unusual properties

both in pure form and as a solvent [1]. Therefore,

water has been the subject of numerous experi-

mental and theoretical investigations. However, in

comparison with its simplicity at the molecular

level, water is a complex and poorly understoodliquid.

* Correspondence author.

E-mail address: sunqiang168@sina.com (Q. Sun).

0009-2614/$ - see front matter � 2003 Elsevier B.V. All rights reserv

doi:10.1016/j.cplett.2003.07.028

In contrast with studies of water at high tem-

perature, especially in supercritical water, there

have not been many experimental works of the

effect of high hydrostatic pressure on hydrogenbonding in water [2–8]. From the energy-dispersive

X-ray diffraction technique (EDXD), Okhulkov

et al. [3] have found that the average separation

between nearest molecules rOO decreases with the

pressure rise up to �2 kbar, but at higher pressures

it begins to grow and reach the initial value at

4–5 kbar, then decreases with increasing pressure.

Although Bellissent-Funel and Bosio [9] seem toconfirm qualitatively the behavior of rOO, some

authors [10] reckoned this result as debatable. In

Raman scattering experiments, because of the

large scatter of experimental points, Walrafen and

Abebe [2] approximated their data with a linear

fit. Cavaille and Combes [4] definitely indicate a

ed.

428 Q. Sun et al. / Chemical Physics Letters 379 (2003) 427–431

singularity in the pressure range 2–3 kbar, but

their data points are scarce and seem to contra-

dictory to the behavior of rOO because the

stretching vibration increases up to 2 kbar then

decreases with increasing pressure.

In this letter, experiments were conducted inMoissanite anvil cells [11] to study the change of

the stretching vibration band of water up to 6 kbar

at 290 K. In order to control the increase of ex-

perimental pressure and obtain dense data points,

1-mm thick Cr1Ni18Ti9 stainless steel was applied

as gasket and the sample chamber was 600 lm in

diameter. Experimental pressure was calculated

according to Raman shift of the 464 cm�1 peak ofquartz. This is because the Raman shift was more

obvious ((9� 0.5) cm�1/GPa [12]), the pressure

could be determined more accurately. H2O used in

experiments was ion-distilled water and tempera-

ture was 290 K. During experiments, pressure was

applied and maintained for three minutes before

Raman spectra were measured, in order to attain

the hydrostatic pressure distribution of the system.Raman spectra were obtained using a confocusal

micro-Raman system Reshaw1000. The excitation

wavelength was the 514.5 nm line of an Arþ

ion laser operating at 25 mW. The spectra were

Fig. 1. (a) Raman spectra of the H2O stretching vibrations up to 6 kba

quartz.

recorded with scan times 1, accumulation times of

10 s, slit of 50 lm and ocular of 50. The resolution

was �1 cm�1.

The stretching vibration bands of water at

various pressure and 290 K are listed in Fig. 1(a)

and (b), shows the Raman behaviors of the 464cm�1 peak of quartz. The 464 cm�1 peak of quartz

is related to bending vibrations of the intra-tetra-

hedral O–Si–O angles [13]. According to Schmidt

and Ziemann�s studies [12], the pressure can be

calculated by

P ðMPaÞ ¼ 0:36079 � ½ðDvpÞ464�2 þ 110:86 � ðDvpÞ464:

The pressure calculated using the equation has anuncertainty of �50 MPa. The peak maximum is

determined by fitting the spectra using JANDELANDEL

SCIENTIFICCIENTIFIC PEAKFITEAKFIT v4.04 computer program. In

the process of determining ðv1Þmax, in order to

eliminating error, the symmetric stretching band of

water is all taken from 2500 to 4000 cm�1, and the

same fitted parameters are used in determining

ðv1Þmax. Fig. 2 shows the changes of the stretchingvibration maximum of water ðv1Þ with increasing

pressure, and the dashed lines in the figure refer to

linear square treatment of the data. It means that

higher pressure can make the ðv1Þmax shift to lower

r at 290 K. (b) The pressure dependence of the 464 cm�1 peak of

Fig. 2. Pressure dependence of the stretching vibration maxi-

mum ððv1ÞmaxÞ of H2O at 290 K. The uncertainty of pressure is

±50 MPa [12]. The determination of ððv1ÞmaxÞ is described in the

text. The uncertainties of pressure and wavenumber are all re-

flected in the data symbols. The dashed lines are the least square

fitted lines.

Q. Sun et al. / Chemical Physics Letters 379 (2003) 427–431 429

wavenumber initially and reaches the minimum at

about 2 kbar, and ðv1Þmax increases with increasing

pressure up to about 4 kbar, then decreases with

increasing pressure up to 6 kbar. In other words,

there exists discontinuity in liquid water at about 2

and 4 kbar. These are different from Walrafen andAbebe�s studies [2] and Cavaille and Combes�s re-sults [4]. We attribute the reason to the scarce data

points in their experiments.

As for water molecules, besides the covalent

bonds in H2O produced by orbital overlap be-

tween inharmonic sp3 hybrid orbits of oxygen

atom and 1s orbit of hydrogen atom, there also

exists strong hydrogen bond interactions betweenwater molecules. It is well known that many of the

unique properties of water are attributed to the

result of three-dimensional hydrogen bonding

network formed between water molecules. There-

fore, the OH stretching vibration of water is also a

strong function of the hydrogen bond strength.

The behavior of ðv1Þmax reflects the change of hy-

drogen bond. So, from the change of ðv1Þmax up to6 kbar at 290 K, we can conclude that the hy-

drogen bond energy decreases up to 2 kbar, and

increases up to 4 kbar, then decreases with in-

creasing pressure up to 6 kbar. This is accordance

with the behavior of rOO measured by Okhulkov

et al. [3].

The structure of liquid water has been the

subject of numerous investigations and remains

controversial. Up to now, most of models can be

divided into two categories: (a) the mixture/inter-

stitial and (b) the distorted hydrogen bond (con-

tinuum) categories [14]. In the former, the mixturemodels postulate the simultaneous existence of two

or more relatively long-lived structures in the li-

quid, such as the �flickering-cluster� model pro-

posed by Frank and Wen [15]. Different and

discrete combinations of hydrogen-bonded mole-

cules are assumed to coexist as evidenced by the

existence of isosbestic points which are well known

in the spectroscopy of reversible chemical reac-tions [16]. The second and currently the most fa-

vored model, is based on the assumption that the

structure relaxes on a time scale that is similar to

that observed in other liquids, and water is

thought to exist as a continuous network of mol-

ecules interconnected by somewhat distorted hy-

drogen bonds [17,18]. Recently, an outer structure

two-state model was put forward and applied toexplain the anomalous properties of the liquid

[19,20]. Very simply speaking, the outer two-state

model is a mixture of ice-Ih- and ice-II-type

bonding, locally rearranging on picosecond time-

scales with average compositions that depend on

the temperature and pressure.

As for water molecules, apart from the uni-

versal van der Waals� interaction, a specific inter-action-hydrogen bonding also exists, and many of

the unique properties of water are attributed to the

result of three-dimensional hydrogen bonding

network formed between water molecules. It is

well known that the hydrogen bond in liquid water

arises as a result of electrostatic interaction be-

tween a hydrogen atom and some excess of nega-

tive charge on a neighboring oxygen atombelonging to another molecule. Such a bond is

much weaker than the usual chemical bonds but,

like the latter, it reveals quite a noticeable orien-

tation correlation. In order to maximize hydrogen

bond, each water molecule tends to interact with

its around molecules, then on time scales less than

the lifetime of a hydrogen bond, more complicated

molecular structural unit can be formed. And thisstructural unit can be called water molecular

concentration �cluster�. However, it should be

430 Q. Sun et al. / Chemical Physics Letters 379 (2003) 427–431

noted that the molecular cluster of liquid water

can only be termed as �V-structure� [21]. In other

words, it is the time averaging of the molecular

coordinates. At this point, the structure of liquid

water is different from that of ice.

In liquid water, the distribution of electriccharges on the oxygen atom in the water molecule

allows it to form two �legal� bonds with hydrogen

atoms belonging to the same molecule, and two

�illegal� bonds with the hydrogen atoms of other

molecules. Each pair of the bonds forms a nearly

tetrahedral angle (105�–109�). Thus, each molecule

of water can join four water molecules, thereby

forming a tetrahedron around it. Ohtomo et al.[22] have shown that a combined analysis of X-ray

and neutron diffraction data suggests the presence

of tetrahedral pentamer clusters. Such a molecular

configuration is a primary element of the structure

of ice and an inescapable attribute of any reason-

able model of the water structure.

On the other hand, according to theoretical

calculation for water cluster (H2O)n (n ¼ 6, 5), themost stable structure should be quasi-planar mo-

lecular cyclic hexamer and pentamer [23,24]. The

reason for the dominance of pentagons and

hexagons in bulk water systems is that these are

the smallest polygons that can produce O–O–O

angles near the optimum (tetrahedral) value,

which maximizes the hydrogen bond energy [25]. It

has been found that near a melting line, local orderof liquid phase structure is like the local order of

the solid phase [26,27]. Studies have shown that

normal ice (Ih) consists of regular arrays of

hexagons. From this, it can be deduced that there

should exist cyclic water hexamer in liquid water

near melting point. In order to maximize hydrogen

bond, each water molecule should be tetrahedrally

hydrogen bonded to their neighbors. At this time,the local structure of liquid water just resemble the

local structure of ice Ih [16]. From these, we con-

clude that there also should exist phase transition

in liquid water just like ice (Ih), but the changes of

physical and chemical properties in water phase

transition should be much weaker than those in ice

phase transition. In fact, the phase transitions of

ice (Ih)! ice (III) and ice (III)! ice (V) respec-tively occurs at about 2 and 3.7 kbar. Studies have

suggested that the minimum of viscosity [28] and

the maximum of self-diffusion [29,30] are observed

around 2 kbar. From the above discussion, We

conclude that, just like ice phase transition ice

(Ih)!ice (III)! ice (V), there also exists water

phase transition water (Ih)!water (III)!water

(V). It can be foreseen that the longitudinal waveof liquid water should attain minimum at 2 kbar.

In mathematical character, it can be termed as

�inflexion�, which is different from the �discontinu-ity� from ice Ih to ice III phase transition.

Acknowledgements

The authors sincerely thank Charles and Col-

vard Ltd, USA to provide the Moissanite anvils

generously. This work is supported by the Na-

tional Natural Science Foundation of CHINA

(Grant Nos. 40103005 and 10032040).

References

[1] F.H. Stillinger, Science 209 (1980) 451.

[2] G.E. Walrafen, M. Abebe, J. Chem. Phys. 68 (1978) 4694.

[3] A.V. Okhulkov, Y.N. Denianets, Y.E. Gorbaty, J. Chem.

Phys. 100 (1994) 1578.

[4] D. Cavaille, D. Combes, J. Raman Spectrosc. 27 (1996)

853.

[5] M. Canpolat, F.W. Starr, A. Scala, Chem. Phys. Lett. 294

(1998) 9.

[6] Y.E. Gorbaty, R.B. Gupta, Ind. Eng. Chem. Res. 37 (1998)

3026.

[7] A.G. Kalinichev, Y.E. Gorbaty, A.V. Okhulkov, J. Mol.

Liquids 82 (1999) 52.

[8] Y.E. Gorbaty, G.V. Bondarenko, A.G. Kalinichev, Mol.

Phys. 96 (1999) 1659.

[9] M.C. Bellissent-Funel, L.A. Bosio, J. Chem. Phys. 102

(1995) 3727.

[10] H. Ohtaki, T. Radnai, T. Yamaguchi, Chem. Soc. Rev. 26

(1997) 41.

[11] J.A. Xu, H.K. Mao, Science 290 (2000) 783.

[12] C. Schmidt, C.M. Ziemann, Am. Miner. 85 (2000) 1725.

[13] J. Etchepare, M. Merian, L.S Metankine, J. Chem. Phys.

60 (1974) 1873.

[14] H.E. Stanley, J. Teixeira, J. Chem. Phys. 73 (1980) 3404.

[15] H.S. Frank, W. Wen, Discuss. Faraday Soc. 26 (1957) 133.

[16] G.E. Walrafen, M.R. Fisher, M.S. Hokmabadi, J. Chem.

Phys. 85 (1986) 6970.

[17] A. Rahman, F.H. Stillinger, J. Am. Chem. Soc. 95 (1973)

7943.

[18] F.H. Stillinger, A. Rahman, J. Chem. Phy. 60 (1974) 1545.

[19] C.H. Cho, S. Singh, G.W. Robinson, J. Chem. Phys. 107

(1997) 7979.

Q. Sun et al. / Chemical Physics Letters 379 (2003) 427–431 431

[20] J. Urquidi, S. Singh, C.H. Cho, Phys. Rev. Lett. 83 (1999)

2348.

[21] J. Frenkel, Kinetic Theory of Liquids, Clarendon, Oxford,

UK, 1946.

[22] N. Ohtomo, K. Tokiwano, K. Arakawa, Bull. Chem. Soc.

Jpn. 54 (1981) 1802.

[23] S.S. Xantheas, T.H. Dunning, J. Chem. Phys. 99 (1993)

8774.

[24] J.D. Cruzan, L.B. Braly, K. Liu, M.G. Brown, J.G.

Loesser, R.J. Saykally, Science 271 (1996) 271.

[25] R.J. Speedy, J.D. Madura, W.L. Jorgensen, J. Phys. Chem.

91 (1987) 909.

[26] F.H. Stillinger, T.A. Weber, Phys. Rev. A. 25 (1982) 978.

[27] H.J.M. Hanley, T.R. Welberry, D.J. Evans, G.P. Morris,

Phys. Rev. A. 38 (1988) 1628.

[28] D. Eisenberg, W. Kauzman, The Structure and Properties

of Water, Oxford University Press, London, 1969.

[29] D.J. Wilbur, T. Defries, J. Jonas, J. Chem. Phys. 65 (1976)

1783.

[30] J. Jonas, Y.T. Lee, J. Phys.: Cond. Mat. 4 (1992) 305.