Comparative Study on the Properties of Hydration Water of Na-and K

Article pubs.acs.org/JPCA

Comparative Study on the Properties of Hydration Water of Na- and K‑Halide Ions by Raman OH/OD-stretching Spectroscopy and Dielectric Relaxation Data Yuichiro Okazaki,† Tetsuo Taniuchi,‡,† George Mogami,† Nobuyuki Matubayasi,#,⊥ and Makoto Suzuki*,† †

Graduate School of Engineering and ‡Frontier Research Institute for Interdisciplinary Sciences, Tohoku University, Sendai 980-8579, Japan # Division of Chemical Engineering, Graduate School of Engineering Science, Osaka University, Toyonaka, Osaka 560-8531, Japan ⊥ Elements Strategy Initiative for Catalysts and Batteries, Kyoto University, Katsura, Kyoto 615-8520, Japan S Supporting Information *

ABSTRACT: Properties of hypermobile water (HMW) were studied by Raman OH-stretching spectroscopy. Hydration water properties measured by Raman OH-stretching spectra of NaX/KX (X: Cl, Br, I) solutions (0.05−0.2 M) were comparatively analyzed with the data by dielectric relaxation spectroscopy (DRS), NMR, and statistical mechanical studies. The Raman OH-stretching spectra were well-fitted with linear combinations of the spectra of pure water both at the same and the higher temperatures. The fitting analysis determined the “structure temperature” Tstr and mole fraction of the high Tstr water region, giving the hydration number Nhyd, for each electrolyte solution. The determined Tstr was much higher than the solution temperature of 293 K for each tested salt and was higher for larger halide ions, consistent with commonly known “structure-breaking” order Cl < Br < I. No significant differences in Nhyd were observed between NaX and KX and among even halide ion species within the experimental errors. Measured Nhyd values of 25−27 were much greater than the reported numbers by NMR chemical shift and 17O NMR relaxation studies and comparable to the numbers of hypermobile water reported in the previous DRS studies. The results indicated that the hydration region around NaX or KX measured by the present Raman study was nearly overlapped with the region of HMW by DRS. It was also suggested that differences in the ion size effects on Tstr and the DR frequency resulted from the sensitivity difference to long-range many-body interactions among water molecules. High structure−temperature regions were also detected by the analysis of OH-stretching and OD-stretching bands for 0.2 M NaI in H2O/D2O mixed solvent of 50 mol %, and we found that both OH-stretching and OD-stretching bands have almost equivalent Tstr ≈ 330 K and mole fractions with each other. (F−, Cl−, Br−, I−) and the cation series (Na+, K+, Rb+, Cs+).11 The molecular origin of HMW detected by DRS was explained by the angle-dependent integral equation theory, which accounted for many-body interactions, giving a higher rotational entropy region of water molecules surrounding ions than that of the bulk.12 A recent study using a statistical mechanical approach and MD simulation of water molecules around a monatomic ion (Z = 0, ±1, ±2) indicated that the cross-correlation term of molecular polarization was essential to cause a faster dielectric relaxation (DR) mode than that of bulk water, while the selfcorrelation term caused only a slower relaxation mode.13 Those results imply that the dynamical responses in electrolyte solutions faster than those of liquid H2O are caused from many-body interactions, or cooperative interactions, among ions and surrounding water molecules.

1. INTRODUCTION Structure and dynamics of the solvent water in alkali halide solutions have been extensively studied.1−5 The strength of those ionic effects on the structure of water is often discussed with the Hofmeister series, which indicates the orders of sedimentation ability of salt ions for aqueous protein solutions.6 Commonly, the viscosity B coefficient is often used as a measure of the water-structuring effect; ions with positive B coefficients are classified as water structure makers, and ions with negative B are classified as water structure breakers.3,7 Thermodynamic hydration properties of ions were summarized by Marcus.8 Recent studies using dielectric relaxation spectroscopy (DRS) revealed that the modified water by salt ions exhibited a single Debye component with a higher relaxation frequency other than γ dispersion of bulk water. Those results indicated the existence of fast-response water, called hypermobile water (HMW), in NaX and KX (X:Cl, Br, I) solutions.9,10 A theoretical study using a statistical mechanical method of RISM-type analysis revealed the ion size effects on the translational friction coefficient for the anion series © 2014 American Chemical Society

Received: December 31, 2013 Revised: March 25, 2014 Published: March 27, 2014 2922

dx.doi.org/10.1021/jp412804d | J. Phys. Chem. A 2014, 118, 2922−2930

The Journal of Physical Chemistry A

Article

OH concentration with D2O, where τrc was measured by timeresolved pump−probe laser spectroscopy. Such resonant intermolecular energy transfer of the OH-stretching excitation was also discussed in the study of KCl and KI aqueous solutions using Raman anisotropy spectroscopy, and it suggested the importance of three-body interactions.33 Thus, it is quite interesting to investigate the ionic hydration water from the viewpoint of the many-body interaction phenomena. In this paper, the properties and mole fraction of the hydration water of sodium and potassium halide (X: Cl, Br, I) solutions are analyzed from Raman spectra by assuming the existence of bulk water as a continuous phase dissolving hydrated solute ions and are compared with the results on hydration water given by the previous DRS studies.9,10 The differences between the DRS result and Raman result for the same object solution should reveal the phenomena due to many-body interactions among water molecules rather than two-body interactions between the water molecule and solute molecule.

Interactions among water molecules in liquid H2O have been studied by vibrational spectroscopy.14 The Raman band of aqueous solutions of electrolytes in the range from 2700 to 3800 cm−1 contains information on the water structure. The peak frequency is much lower than that of OH-stretching modes of water vapor (3657, 3756 cm−1) and much higher than that of ice (∼3200 cm−1).15 The band was often decomposed into three or more components of Gaussian, Lorentzian, or their combinations with major peak frequencies of ∼3230, ∼3450, and ∼3620 cm−1.15,16 The ∼3230 cm−1 component was assigned as OH-stretching modes of correlated hydrogen-bond clusters,17 such as a cage-hexamer containing AAD-water18 and a cyclic pentamer.19 A and D denote the acceptor and donor of hydrogen for hydrogen bonds, respectively. The ∼3450 cm−1 component was assigned as an assembly of hydrogen-bonded OH-stretching modes such as a cyclic trimer/tetramer/ pentamer,19 cage-hexamer/prism-hexamer containing AADwater,18 and a bifurcated hydrogen bond (BHB) based on the neutron diffraction data.20 The ∼3620 cm−1 component was assigned as an assembly of nonbonded OH-stretching modes such as symmetric OH-stretching modes of the monomer and A-water, OH-stretching modes of the trimer containing ADwater,21 and cage/prism hexamers containing ADD-water.18 A recent quantum chemical study determined optimum geometries and harmonic vibrational frequencies for cyclic water clusters at the coupled cluster including single, double, and full perturbative triple excitations (CCSD(T))/aug-cc-pVDZ level of theory.19 It showed a clear proportional relationship, Δν/ΔR = 20 cm−1/ 0.001 Å up to Δν = 700 cm−1, between the red shift Δν from the monomer’s average value of R ≈ 3700 cm−1 (not clearly shown) and the elongation (ΔR) of the hydrogen-bonded OH bond from the monomer’s OH bond length R for the cyclic clusters of n = 2−6. Optimum geometries of those water clusters and the proportional relationship lead us to understand the previous works consistently.18,20,22 For alkali halide aqueous solutions, the number of measurements of Raman spectra of the OH-stretching band was determined in the frequency range of 2500−4000 cm−1.22−26 Li et al. reported that the Gaussian component with a peak at 3233 cm−1 decreased and that the Gaussian components of 3393 and 3511 cm−1 increased with increasing NaX concentration, deriving that hydrogen-bond-breaking actions of halogen ions were in the order of F < Cl < Br < I, while the cation effect was undetectable. Many scientists noticed that the OH-stretching bands of alkali halide solutions were similar to those of heated liquid water,27,28 even though some differences (3051, 3625 cm−1 by Li et al.) were seen in those spectral profiles.25,26,29 Aliotta et al.30 found that the supercooled water and supercooled LiCl aqueous solution could be fitted with a two-state model of more-structured water and less-structured water when these were above some critical temperature. They also wrote that the Raman OH-stretching spectrum at T was expressed with a linear combination of the bulk water spectrum at T and the spectrum of eutectic composition at T. In this paper, we examine whether the water surrounding ions were classified into water at the same temperature and highertemperature water over six alkali halide species based on quantitative examination instead of using the spectrum of eutectic composition. Bakker et al.31,32 reported a strong resonance coupling of the OH-stretching vibration among water molecules in liquid water based on the fact that the excitation relaxation time τrc was in the subpicosecond range and increased by diluting the

2. MATERIALS AND METHOD 2.1. Preparation of Solutions. NaCl (purity 99.5%), NaBr (purity 99.0%), NaI (99.5%), KCl (purity 99.5%), KBr (purity 99.0%), and KI (99.5%) were purchased from Wako Chemical and dissolved in deionized water by Milli-Q (Millipore) at concentrations from 0.05 to 0.2 mol/L(M). Deuterium oxide (purity 99.9%, DLM-4-25) was purchased from Cambridge Isotope Laboratories. H2O/D2O (molar ratio 1:1) mixed solvent and 0.2 M NaI in H2O/D2O (molar ratio 1:1) mixed solvent were prepared. The concentration of each solution was confirmed by the solution density measured with a DMA5000M density meter (Anton Paar, Graz, Austria) and literature values34 to ensure three significant digits. Each solution was degassed and introduced into the measurement glass cell. 2.2. Raman Spectroscopy. 2.2.1. Measurement. All Raman spectra were recorded using a Nd:YAG laser (Elforlight) operated at 100 mW with an excitation wavelength at 532 nm and a spectrometer (Lambda Vision Inc.) with an entrance slit of 25 μm. The resolution was about 5 cm−1. The laser beam was focused on the sample using a 5× objective lens. Each measurement took 0.1 s from 750 to 4000 cm−1. Each Raman spectrum I(ν) was obtained as the average of five sets of 200 sweeps for each salt solution at 293 K and for water (H2O) and the H2O/D2O mixed solvent at every 10 K in the temperature range from 293 to 373 K using a thermo-controlled holder (Shimadzu) within less than 0.1 K. Each Raman spectrum was corrected using the straight baseline connected between 2700 and 4000 cm−1 points and then normalized before analysis as the integrated area of the spectrum in the range of 2700−4000 cm−1 to be equal to that of liquid H2O at 293 K. OD-stretching Raman spectra of 0.2 M NaI in H2O/D2O mixed solvent at 293 K were corrected using the straight baseline connected between 1980 and 2850 cm−1 points and then normalized before analysis as the integrated area of each spectrum is equal to that of H2O/D2O mixed solvent (molar ratio 1:1) in the range of 1980−2850 cm−1 at 293 K. OH-stretching Raman spectra of 0.2 M NaI in H2O/ D2O mixed solvent at 293 K were corrected using the straight baseline connected between 2850 and 4000 cm−1 points. 2.2.2. Analysis. For the OH-stretching band of alkali halide H2O solutions at 293 K, the solution spectrum Isol (v) is divided into two terms, assuming the existence of the bulk water phase in dilute electrolyte solutions 2923



Article

Isol(ν) = (1 − ϕ)Ibulk(ν) + ϕIhyd(ν)

(1)

where Ibulk(ν) and Ihyd(ν) are the liquid H2O spectrum at 293 K and the spectrum of the water region modified by ions, which is here denoted as the hydration shell, respectively. ϕ is the mole fraction of the modified water by ions in the solution. In this study, Ihyd(ν) is approximated with the liquid H2O spectrum at temperature T, following the high salt concentration results22 and neglecting small differences.25 Ihyd(ν) is Ihyd(ν) ≈ Ibulk(ν , T ) ⎧ ⎡ ⎪ ⎛ν ⎪ Ibulk(ν , T ) ≈ ∑ a0, n⎨αn exp⎢ −4 ln 2⎜⎜ ⎢ ⎪ ⎝ n=1 ⎣ ⎪ ⎩ ⎛ ⎜ 1 + (1 − αn)⎜ ⎜ ν − a1, n ⎜1 + 4 a 2, n ⎝ 3

(

2⎤ − a1, n ⎞ ⎥ ⎟⎟ a 2, n ⎠ ⎥⎦

⎞⎫ ⎟⎪ ⎪ ⎟⎬ 2 ⎟⎪ ⎟⎪ ⎠⎭

)

(2)

Temperature dependences of parameters used in eq 2 are given by fitting the experimental spectra of liquid H2O obtained between 293 and 373 K. Each component had a peak as follows; component 1 (n = 1) at 3230−3250 cm−1, component 2 (n = 2) at 3440−3460 cm−1, and component 3 (n = 3) at 3610−3625 cm−1. Parameters αn, a0,n, a1,n, and a2,n are temperature-dependent. For 0.2 M NaI in H2O/D2O mixed solvent at 293 K, the OD-stretching band in the frequency range of 1980−2850 cm−1 and the OH-stretching band in the frequency range of 2850− 4000 cm−1 are analyzed. The OD-stretching band spectrum Isol(ν) is divided into two terms, assuming the existence of the H2O/D2O mixed solvent phase is not affected by ions in dilute electrolyte solutions in the same way as noted above using eqs 1 and 2 with three components, component 1 with a peak at ∼2400 cm−1, component 2 at 2500−2550 cm−1, and component 3 at ∼2670 cm−1. 2.3. DRS of the H2O/D2O Mixed Solvent System. Here, we have to stress that the word “hypermobile” means faster DR than that of bulk water, and it does not mean higher rotational mobility of water molecules according to our previous theoretical studies.12,13 In a H2O/D2O mixed solvent system, moderately fast cross-correlation decay among water dipoles could be detected by DR time measurement because the number of cross interactions among H2O dipoles should be reduced. DRS measurements of H2O/D2O mixed solvent and 0.2 M NaI in H2O/D2O mixed solvent were carried out by the method used in the previous paper.10

Figure 1. (a) Temperature dependence of the OH-stretching band of liquid H2O. (b) Decomposition of the OH-stretching band of liquid H2O at 293 K. (c−f) Temperature dependences of the fitting parameters, (c) αn, (d) a0,n, (e) a1,n, and (f) a2,n, in eq 2. α1 = 2 × 10−6T 2 − 0.0032T + 1.3818

(R2 = 0.9286)

α2 = 1 α3 = 1 a0,1 = 6 × 10−6T 4 − 0.0084T 3 + 4.3099T 2 − 986.37T + 85703 (R2 = 0.9996) a0,2 = 2.0751T + 401.43

(R2 = 0.9992)

a0,3 = − 0.0004T 3 + 0.3869T 2 − 129.95T + 14541

(R2 = 0.9973)

a1,1 = 2 × 10−6T 4 − 0.0021T 3 + 1.0835T 2 − 242.78T + 23529 (R2 = 0.996) a1,2 = 2 × 10−6T 4 − 0.0022T 3 + 1.113T 2 − 244.35T + 23468 (R2 = 0.9936) a1,3 = − 0.0004T 2 + 0.107T + 3632.2 a 2,1 = − 0.0021T 2 + 1.3482T + 14.705

(R2 = 0.990) (R2 = 0.8159)

a 2,2 = 1 × 10−4T 3 − 0.1003T 2 + 34.034T − 3582.4

(R2 = 0.974)

a 2,3 = − 3 × 10−8T 4 − 6 × 10−5T 3 + 0.0857T 2 − 31.246T + 3721.6 (R2 = 0.9958)

(3)

3.2. OH-Stretching Band of NaX and KX Solutions. The OH-stretching band of sodium halide (NaX) and potassium halide (KX) solutions (X: Cl, Br, I) at 293 K are shown in Figure 2. A proportional relation was seen between ΔI(ν) [=Isol(ν) − Ibulk(ν)] and the NaI concentration in Figure 2c, corresponding to Figure 2a. As noted previously, the spectral difference between Na and K was very small. The differences among Cl, Br, and I were small but systematic based on the small measurement errors, as in Figure 2d−f, which were consistent with the previous reports.25

3. RESULTS 3.1. OH-Stretching Band of Liquid H2O. The temperature dependence of the OH-stretching band of liquid H2O was measured every 10 K within the range of 293−373 K, as shown in Figure 1a. Each OH-stretching band of liquid H2O at absolute temperature T was fitted with eq 2 (Figure 1b). Temperature dependences of the fitting parameters αn, a0,n, a1,n, and a2,n are indicated in Figure 1c−f, respectively. Then, these parameters at different temperatures were fitted with polynominal functions of T as follows. 2924



Article

simultaneously determined. Then, the water number Nhyd in the hydration shell per NaX or KX was calculated by eq 4, using salt concentration c in M, solution density ρsolution in g/mL, and mass weight of salt Msalt and mass weight of solvent Msolvent in g/mol. Nhyd =

(1000ρsolution − Msaltc)/Msolvent c

ϕ

(4)

The results are summarized in Table 1. The volume and water number of the hydration shell per anion−cation pair were found to be independent of c for NaI, as shown in Table 1 and Figure 4 and for other salts tested in this study (see Figure S1 of the Supporting Information). The structure temperature Tstr was found to increase with the halide ion size, while no significant difference in Nhyd of the hydration shell per anion−cation pair between Na+ and K+ was found. The halide ion effect on Nhyd was also insignificant. Thus, every OH-stretching band spectrum of the hydration shells of six alkali halides was well-fitted with the higher-temperature spectrum of the solvent, and the salt species dependences of the mole fraction and Tstr were determined. 3.3. OH-Stretching Band of 0.2 M NaI in H2O/D2O Mixed Solvent. The Raman spectrum of H2O/D2O mixed solvent where the molar ratio was 1:1 was taken as noted in section 2.2.1 and shown in Figure 5. The temperature dependence of the OH-stretching band of the H2O/D2O mixed solvent was measured every 10 K in the range of 293− 373 K, as shown in Figure 6a. Each OH-stretching band at T was fitted with eq 2 (Figure 6b). Temperature dependences of the fitting parameters αn, a0,n, a1,n, and a2,n are indicated in Figure 6c−f, respectively. Then, these parameters at different temperatures were fitted with polynominal functions of T as follows.

Figure 2. Raman OH-stretching spectra of alkali halide solutions. (a) Isol(ν) of NaI solutions at 293 K, (b) Isol(ν) of alkali halide solutions at 0.2 M, (c) ΔI(ν) (=Isol(ν) − Ibulk(ν)) for NaI solutions at 0.05, 0.10, 0.15, and 0.20 M, (d) ΔI(ν) for alkali halide solutions at 0.20 M (solid lines for Na and broken lines for K), (e,f) standard errors of ΔI(ν) corresponding to (c) and (d), respectively.

By minimizing the sum of fitting errors over 2700−3800 cm−1 with eqs 1 and 2, as shown in Figure 3, each spectrum was decomposed into the bulk water component of 293 K and the hydration shell component, which was given by the liquid H2O spectrum at the “structure temperature”28 Tstr using eq 3, where the mole fraction ϕ of the hydration shell in total water was

Figure 3. Decompositoin of the Raman OH-stretching spectrum. Isol(ν) of (a) NaI 0.05 M, (b) NaI 0.10 M, (c) NaI 0.15 M, (d) NaI 0.20 M, (e) NaBr 0.20 M, (f) NaCl 0.20 M, (g) KI 0.20 M, (h) KBr 0.20 M, and (i) KCl 0.20 M. For each solution, fitting was made with eqs 1−3. 2925



Article

Table 1. Structure Temperature and Hydration Numbers by the Raman OH-Stretching Spectrum

a

salt

c (M)

NaI NaI NaI NaI NaBr NaCl KI KBr KCl

0.05 0.1 0.15 0.2 0.2 0.2 0.2 0.2 0.2

Tstr/T0 (T0 = 293 K)

structure temperature Tstr (K) 372.2 371.7 371.8 371.8 353 344 371.9 355.5 346.3

± ± ± ± ± ± ± ± ±

4.6 2.4 2.5 2.3 3.1 3.9 1.9 3.0 2.9

1.269 1.268 1.268 1.268 1.204 1.173 1.268 1.212 1.181

ϕ 0.022 0.050 0.073 0.097 0.099 0.090 0.093 0.096 0.095

± ± ± ± ± ± ± ± ±

DRa Nhyper

Nhyd 0.003 0.003 0.004 0.004 0.003 0.005 0.004 0.004 0.007

24.4 27.7 26.9 26.7 27.4 24.9 25.6 26.5 26.2

± ± ± ± ± ± ± ± ±

3.4 1.6 1.5 1.1 1.1 1.4 1.1 1.1 2.0

27.9 24.7 21.5 27.1 22.0 17.1

± ± ± ± ± ±

1.6 2.0 1.7 0.9 1.2 0.7

DRa fc (GHz)

NMRb N17O_tau

NMRc NH_shift

IRd NIR

± ± ± ± ± ±

−1.1 1.1 2.8 −5.0 −2.8 −1.1

3.1 3.1 3.1 2.1 2.1 2.1

3.5 5 5 4 5 5

17.9 17.9 17.8 18.9 19.7 20.4

0.2 0.3 0.3 0.2 0.1 0.2

From ref 10 (283 K). bFrom refs37 and 38 (298 K). (According to ref 38, negative values correspond to negative hydration.) cFrom ref 35 (298 K). From ref 36 (300 K).

d

Figure 5. Raman spectrum of a H2O/D2O mixed solvent at 293 K (molar ratio 1:1).

Figure 4. Volume fraction of the hydration shell and structure temperature versus the salt concentration c, corresponding to Figure 3d (see also Figure S1 of the Supporting Information). α1 = 6 × 10−5T 2 − 0.00331T + 5.427

(R2 = 0.9827)

α2 = 1 α3 = 1 a0,1 = 5 × 10−6T 4 − 0.0068T 3 + 3.3199T 2 − 721.59T + 59419 (R2 = 0.9192) a0,2 = − 1 × 10−5T 4 − 0.0156T 3 − 7.5444T 2 + 1620.3T − 129888 (R2 = 0.9424) a0,3 = 1.039T − 183.62

(R2 = 0.9969)

a1,1 = 0.0146T 2 − 8.9402T + 4640.5

(R2 = 0.9781)

a1,2 = 0.6519T + 3248.8

(R2 = 0.984)

a1,3 = 0.0327T + 3608.5

(R2 = 0.7177)

a 2,1 = 0.0002T 3 − 0.1675T 2 + 52.038T − 5160.9 a 2,2 = 1 × 10−5T 3 − 0.0144T 2 + 6.0065T − 526.65 a 2,3 = 0.0125T + 108.76

(R2 = 0.9771) (R2 = 0.9892)

(R2 = 0.1766)

(5)

Using these parameters, the Raman spectrum of 0.2 M NaI in the H2O/D2O mixed solvent in the range of 2850−3800 cm−1 was well-fitted with the same and higher-temperature spectra of the H2O/D2O mixed solvent, as shown in Figure 7. The resulting Tstr, mole fraction ϕ, and Nhyd were found to be 323.7 ± 1.8 K, 0.177 ± 0.011, and 48.4 ± 3.0, respectively. 3.4. OD-Stretching Band of 0.2 M NaI in H2O/D2O Mixed Solvent. The temperature dependence of the ODstretching band of the H2O/D2O mixed solvent corresponding to the previous section was analyzed every 10 K in the range of 293−373 K, as shown in Figure 8a. Each OD-stretching band at

Figure 6. OH-stretching band of a H2O/D2O mixed solvent (molar ratio 1:1). (a) Temperature dependence of the OH-stretching band of the H2O/D2O mixed solvent. (b) Decomposition of the OH-stretching band of the H2O/D2O mixed solvent at 293 K by eqs 1 and 2. (c−f) Temperature dependences of the fitting parameters, (c) αn, (d) a0,n, (e) a1,n, and (f) a2,n, in eq 2.

T was fitted with eq 2 (Figure 8b). Temperature dependences of the fitting parameters αn, a0,n, a1,n, and a2,n are indicated in Figure 8c−f, respectively. Then, these parameters at different 2926



Article

Figure 7. Decompositoin of the Raman OH-stretching spectrum of 0.2 M NaI in the H2O/D2O mixed solvent at 293 K by eqs 1−3.

temperatures were fitted with polynominal functions of T as follows. α1 = − 2 × 10−6T 2 − 0.0017T + 0.2797

(R2 = 0.266)

α2 = 1 α3 = 1 × 10−6T 3 − 0.001T 2 + 0.3192T − 34.138 a0,1 = 0.0246T + 200.67

(R2 = 0.0154)

a0,2 = 1.2584T − 962.86

(R2 = 0.9958)

a0,3 = 1.0099T − 112.31

(R2 = 0.9216)

a1,1 = 0.0024T 2 − 1.7163T + 2694.5 a1,2 = 0.4779T + 2372

(R2 = 0.4918)

Figure 8. OD-stretching band of the H2O/D2O mixed solvent (molar ratio 1:1). (a) Temperature dependence of the OD-stretching band of the H2O/D2O mixed solvent. (b) Decomposition of the ODstretching band of the H2O/D2O mixed solvent at 293 K by eqs 1 and 2. (c−f) Temperature dependences of the fitting parameters, (c) αn, (d) a0,n, (e) a1,n, and (f) a2,n, in eq 2.

(R2 = 0.9859)

(R2 = 0.9991)

a1,3 = 0.0367T + 2659.9

(R2 = 0.8621)

a 2,1 = − 0.0038T 2 + 2.4299T − 251.3

(R2 = 0.8769)

a 2,2 = − 0.2179T + 271.35

(R2 = 0.9667)

a 2,3 = − 0.1131T + 131.42

(R2 = 0.841)

(6)

Using these parameters, the Raman spectrum of 0.2 M NaI in the H2O/D2O mixed solvent in the range of 1980−2850 cm−1 was well-fitted with the same and higher-temperature spectra of the H2O/D2O mixed solvent, as shown in Figure 9. The resulting Tstr, mole fraction ϕ, and Nhyd were found to be 328.7 ± 1.3 K, 0.180 ± 0.006, and 49.1 ± 1.6, respectively. Experimental errors of the OD and OH-stretching Raman bands were less than 0.06%, as shown in Figure S2 of the Supporting Information. Thus, both the OH-stretching and OD-stretching bands of 0.2 M NaI in the H2O/D2O mixed solvent were successfully fitted with a linear combination of the same and highertemperature spectra of the H2O/D2O mixed solvent. Evaluated values of Tstr and Nhyd from the OD-stretching were found to be close to those from the OH-stretching bands, indicating the existence of cooperative intramolecular or intermolecular interactions among OH- and OD-stretching vibrations of H2O, HDO, and D2O molecules around ions. Tstr for the 0.2 M NaI H2O/D2O system were still higher than the bulk temperature but lower than that in the H2O system. Nhyd for the 0.2 M NaI H2O/D2O system was about twice that for the 0.2 M NaI H2O system. 3.5. Results of DRS Measurements of 0.1 M NaI in H2O/D2O Mixed Solvent at 283 K. DR spectra of the H2O/ D2O (molar ratio 1:1) mixed solvent and 0.1 M NaI in the H2O/D2O (molar ratio 1:1) mixed solvent were recorded for the frequency range of 0.2−26 GHz using the same method as the previous study.10 As a result, the DR frequency fc and the

Figure 9. Decompositoin of the Raman OD-stretching spectrum of 0.2 M NaI in the H2O/D2O mixed solvent at 293 K by eqs 1−3.

number of HMWs per anion−cation pair Nhyper were determined for the solvent and 0.1 M NaI dissolved in the H2O/D2O mixed solvent, as shown in Table 2, where f w is the DR frequency of the solvent. It was found that fc was 15.6 GHz, which was 13% lower than fc of the H2O system and 43% higher than f w of the H2O/D2O mixed solvent, and the number of HMWs Nhyper was 29, slightly larger than that in the H2O system. Thus, the order of the DR frequency was (H2O/D2O mixed solvent) < (H2O) < (NaI in H2O/D2O mixed solvent) < (NaI in H2O). Nhyper for NaI in H2O was slightly greater or equal to that for NaI in H2O/D2O mixed solvent. The order of fc was the same as that of Tstr, while the order of Nhyper showed some difference from that of Nhyd by Raman spectroscopy.

4. DISCUSSION In this study, an increase of structure temperature with increasing halide ion size was reconfirmed as in the order Cl− < Br− < I−. High structure temperature was caused by a decrease of the 3230 cm−1 component of OH-stretching of an ice-like water cluster 2927



Article

Table 2. Hydration Properties of NaI Deduced from Raman OH/OD-Stretching Bands and DRS

a

NaI solution

Tstr (K)

ϕ/c (M−1)

Nhyd

OH-stretching band in H2O at 293 K OH-stretching band in H2O/D2O at 293 K OD-stretching band in H2O/D2O at 293 K DRS: 0.1 M NaI in H2O at 283 K DRS: 0.1 M NaI in H2O/D2O at 283 K

371.8 ± 2.3 323.7 ± 1.8 328.7 ± 1.3

0.485 ± 0.020 0.885 ± 0.055 0.900 ± 0.030

26.9 ± 1.1 48.4 ± 3.0 49.1 ± 1.6

Nhyper

DR fc (GHz)/f w (GHz)

27.9 ± 1.6a 29.0 ± 1.6b

17.9 ± 0.2/12.6a 15.6 ± 0.2/10.9b

From ref 10. bThe values were measured using the method according to ref 10.

accompanied with an increase of the 3450 cm−1 (hydrogen-bonded OH) and 3620 cm−1 (free OH of the monomer and A-water) components. Thus, the structure temperature was derived from the population ratio of three different states of water. On the other hand, an increase of HMW, which was defined as water having a higher DR frequency fc than that of bulk water in electrolyte solutions, was accompanied by a decrease in bulk water and restrained water.10 Tstr and fc are quantities of the water property within the hydration shell of an ion, and Nhyd and Nhyper count the numbers of water molecules per anion−cation pair influenced by the ion. Both Tstr and fc were much higher than those of bulk water. However, the ion size effects on Tstr and fc were different. Although it is not simple to explain this fact, such a difference should be caused from the difference of observed quantities Tstr and fc. DRS provides information of dipole−dipole correlation functions involving all water molecules, as long-range interactions, in the electrolyte solution. Such correlation functions consist of a self-correlation term and a cross-correlation term among water molecules, as described in the previous paper.13 Thus, the significant difference between Raman spectroscopy and DRS of solutions is in the sensitivity for many-body interaction dynamics of solvent water molecules. Nevertheless, numbers Nhyd and Nhyper were in the similar range of 17−28. Moreover, the modified water fraction observed by Raman was proportional to c, suggesting that such high structure temperature water regions were dispersed in the bulk water (Figure 10). As discussed in the

Except for the ion size effect, the water region of high Tstr was found to be nearly overlapped with the HMW region, indicating the distance of how far the ionic structure-breaking effect extended. Nhyd and Nhyper were much greater than the reported numbers by proton NMR chemical shift,35 IR,36 or 17O NMR relaxation,37 as indicated in Table 1. The reason could be as follows. The former two technologies measure only short-range ionic interactions of water molecules; the proton NMR chemical shift method can detect protons with modified electron density by an ion,35 and the ATR-IR method can detect tightly or loosely bound water molecules around an ion.36 The 17O NMR relaxation method can measure the rotational self-correlation time of water molecules averaging over the whole solution,34 giving small hydration numbers, which were calculated using 55.5B′, given by ref 38. Therefore, the difference between Nhyper and NO17_tau is believed to be caused from the many-body interactions based on the cross-correlation function over all water molecules.13 High Tstr corresponded to shortening of the average OH distance of water molecules within the second water layer around an ion according to the previous report.19 High fc of the HMW around an ion was caused by fast relaxation of cross-correlation with water molecules distant from the ion,13 so that high fc does not mean fast rotation of water molecules. The HMW region was caused by high surface charge density of the ion and limited within the second water layer according to the previous study,13 which could explain the relatively lower fc of HMW around a larger halogen ion, as indicated in Table 1. Taken together, it was derived that the hydration water region of the ellipsoidal volume in NaX and KX (X: Cl, Br, I) solutions could be composed of a high structure temperature region (green), which was almost overlapped with the HMW region (cyan) by DRS,10 and a bound water region (brown) by H NMR chemical shift,35 as drawn in Figure 10. Here, we stress that the word “hypermobile” means faster cross-correlation decay among water dipoles than that of bulk water, and it does not mean higher rotational mobility of water molecules according to our previous theoretical studies.12,13 In a H2O/D2O mixed solvent system, moderately fast crosscorrelation decay among water dipoles was expected by DR time measurement because the number of cross interactions among H2O dipoles should be reduced, and the DR fc of D2O is lower than that of H2O. From the analysis of the 0.2 M NaI H2O/D2O system, Tstr from the OH-stretching band was higher than the bulk temperature but lower than that for the 0.2 M NaI in H2O system, as expected from suppressed resonance coupling and coupled oscillations among vibration modes of water molecules. Moreover, Tstr evaluated from the OD-stretching band was almost equal to that from the OH-stretching band. This fact indicates that the OH- and OD-stretching bands are not independent, and there would be strong interactions among vibration modes of water molecules probably due to a combination of resonance coupling and coupled oscillations.1,32 The greater number Nhyd observed in the 0.2 M NaI in H2O/D2O

Figure 10. Hydration components of the hydration region in NaX and KX solutions. X: Cl, Br, and I. The high structure temperature region (green) is given by the present Raman study. The HMW region (cyan) is according to ref 10 using DRS. The bound water region (brown) is according to ref 35 using the H NMR chemical shift.

previous paper,10 the dispersed region can be approximated with an ellipsoid including a pair of an anion and cation. For the cation effect, no significant difference was observed in the OH-stretching spectra between Na+ and K+, while ion size dependence was observed; the larger cation had less Nhyper by the previous DRS study, as indicated in Table 1. As for the anion, the ion size effect on Nhyd by Raman was again not obvious, while the larger anion was found to have more Nhyper by DRS. Thus, Nhyd by Raman was irrespective of X in the range from 25 to 27 per NaX or KX. 2928



Article

solvent than Nhyper by DRS as in Table 2 must be further investigated in the future.

Program, the Strategic Programs for Innovative Research of the Next-Generation Supercomputing Project, and by Core Research of Evolutional Science and Technology of Japan Science and Technology Agency.

5. CONCLUSION By the present Raman spectroscopy, it was confirmed that the Raman OH-stretching spectra of NaX/KX (X: Cl, Br, I) solutions (0.05−0.2 M) were almost perfectly fitted with linear combinations of the spectra of liquid H2O both at the same temperature and higher temperatures. The fitting analysis determined the structure temperature and mole fraction of the high-temperature component for each electrolyte solution. The determined structure temperature was higher for the larger halide ion, consistent with commonly known structurebreaking order Cl− < Br− < I−. The mole fraction of the high-temperature component was proportional to the salt concentration in the measured concentrations up to 0.2 M, giving the number of modified water molecules, the hydration water number Nhyd, per NaX or KX molecule. No significant differences in Nhyd were observed between NaX and KX and among even halide ion species within the experimental error. These results (Nhyd = 25−27) were comparable to the numbers of HMWs. Thus, it was derived that the high Tstr region around an ion was not beyond the second water layer for NaX and KX and exhibited faster dielectric response than bulk water. Although the appearance of HMW properties was different between Raman and DRS, Raman OH-stretching spectroscopy could provide information of hydration properties of ions including HMW observed by DRS. In addition, the Raman spectra of 0.2 M NaI in the H2O/ D2O mixed solvent were examined. Both the OH-stretching and OD-stretching bands of 0.2 M NaI in the H2O/D2O mixed solvent were successfully fitted with a linear combination of the same and higher-temperature spectra of the H2O/D2O mixed solvent. Evaluated values of Tstr and Nhyd from the OHstretching coincided with those from the OD-stretching bands, suggesting the cooperative interactions among OH- and ODstretching vibrations of water molecules around ions.

■

■

(1) Bakker, H. J. Structural Dynamics of Aqueous Salt Solutions. Chem. Rev. 2008, 108, 1456−1473. (2) Jungwirth, P.; Tobias, D. J. Specific Ion Effects at the Air/Water Interface. Chem. Rev. 2006, 106, 1259−1281. (3) Marcus, Y. Effect of Ions on the Structure of Water: Structure Making and Breaking. Chem. Rev. 2009, 109, 1346−1370. (4) Ohtaki, H.; Radnai, T. Structure and Dynamics of Hydrated Ions. Chem. Rev. 1993, 93, 1157−1204. (5) Robinson, R. A.; Stokes, R. H. Electrolyte Solutions, 2nd revised ed.; Dover Publications: Mineola, NY, 1959. (6) Lo Nostro, P.; Ninham, B. W. Hofmeister Phenomena: An Update on Ion Specificity in Biology. Chem. Rev. 2012, 112, 2286− 2322. (7) Gurney, R. W. Ionic Processes in Solution; McGraw-Hill: New York, 1953. (8) Marcus, Y. Ion Properties; Marcel Dekker: New York, 1997; pp 177−185. (9) Miyazaki, T.; Mogami, G.; Wazawa, T.; Kodama, T.; Suzuki, M. Measurement of the Dielectric Relaxation Property of Water−Ion Loose Complex in Aqueous Solutions of Salt at Low Concentrations. J. Phys. Chem. A 2008, 112, 10801−10806. (10) Mogami, G.; Miyazaki, T.; Wazawa, T.; Matubayasi, N.; Suzuki, M. Anion-Dependence of Fast Relaxation Component in Na-, KHalide Solutions at Low Concentrations Measured by HighResolution Microwave Dielectric Spectroscopy. J. Phys. Chem. A 2013, 117, 4851−4862. (11) Chong, S.-H.; Hirata, F. Dynamics of Ions in Liquid Water: An Interaction-Site-Model Description. J. Chem. Phys. 1999, 111, 3654− 3667. (12) Kinoshita, M.; Suzuki, M. A Statistical-Mechanical Analysis on the Hypermobile Water around a Large Solute with High Surface Charge Density. J. Chem. Phys. 2009, 130, 014707/1−014707/11. (13) Kubota, Y.; Yoshimori, A.; Matubayasi, N.; Suzuki, M.; Akiyama, R. Molecular Dynamics Study of Fast Dielectric Relaxation of Water around a Molecular-Sized Ion. J. Chem. Phys. 2012, 137, 224502/1− 224502/4. (14) Lilley, T. H. In Water, A Comprehensive Treatise; Franks, F., Ed.; Plenum: New York, 1973; Vol. 3, p 265. (15) Eisenberg, D.; Kauzmann, W. The Structure and Properties of Water; Oxford University Press: Oxford, U.K., 1969; p 229. (16) Ratcliffe, C. I.; Irish, D. E. Vibrational Spectral Studies of Solutions at Elevated Temperatures and Pressures. 5. Raman Studies of Liquid Water up to 300°C. J. Phys. Chem. 1982, 86, 4897−4905. (17) Walrafen, G. E.; Chu, Y. C. Linearity between Structural Correlation Length and Correlated-Proton Raman Intensity from Amorphous Ice and Supercooled Water up to Dense Supercritical Steam. J. Phys. Chem. 1995, 99, 11225−11229. (18) Kim, K.; Jordan, K. D.; Zwier, T. S. Low-Energy Structures and Vibrational Frequencies of the Water Hexamer: Comparison with Benzene−(H2O)6. J. Am. Chem. Soc. 1994, 116, 11568−11569. (19) Miliordos, E.; Aprà, E.; Xantheas, S. S. Optimal Geometries and Harmonic Vibrational Frequencies of the Global Minima of Water Clusters (H2O)n, n = 2−6, and Several Hexamer Local Minima at the CCSD(T) Level of Theory. J. Chem. Phys. 2013, 139, 114302/1− 114302/13. (20) Giguere, P. A. The Bifurcated Hydrogen-Bond Model of Water and Amorphous Ice. J. Chem. Phys. 1987, 87, 4835−4839. (21) Reimer, J. R.; Watts, R. O. The Structure and Vibrational Spectra of Small Clusters of Water Molecules. Chem. Phys. 1984, 85, 83−112.

ASSOCIATED CONTENT

S Supporting Information *

Proportionality relationship between the mole fraction of ionic hydration shell and salt concentration for NaBr, NaCl, KI, and KBr (Figure S1). Experimental errors of the OD- and OHstretching Raman bands (Figure S2). This material is available free of charge via the Internet at http://pubs.acs.org.

■

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the Grants-in-Aid for Scientific Research on Innovative Areas (Nos. 20118001, 20118002, and 20118008), by the Grants-in-Aid for Scientific Research (Nos. 21300111 and 23651202) from the Japan Society for the Promotion of Science and the Elements Strategy Initiative for Catalysts & Batteries from the Ministry of Education, Culture, Sports, Science, and Technology, by the Nanoscience Program, the Computational Materials Science Initiative, the Theoretical and Computational Chemistry Initiative, the HPCI Strategic 2929



Article

(22) Schultz, J. W.; Hornig, D. F. The Effect of Dissolved Alkali Halides on the Raman Spectrum of Water. J. Phys. Chem. 1961, 65, 2131−2138. (23) Walrafen, G. E. Raman Spectral Studies of the Effects of Electrolytes on Water. J. Phys. Chem. 1962, 36, 1035−1042. (24) Coker, D. F.; Miller, R. E.; Watts, R. O. The Infrared Predissociation Spectra of Water Clusters. J. Chem. Phys. 1985, 82, 3554−3562. (25) Li, R.; Jiang, Z.; Chen, F.; Yang, H.; Guan, Y. Hydrogen Bonded Structure of Water and Aqueous Solutions of Sodium Halides: A Raman Spectroscopic Study. J. Mol. Struct. 2004, 707, 83−88. (26) Sun, Q. Raman Spectroscopic Study of the Effects of Dissolved NaCl on Water Structure. Vib. Spectrosc. 2012, 62, 110−114. (27) Busing, W. R.; Hornig, D. F. the Effect of Dissolved KBr, KOH, or HCl on the Raman Spectrum of Water. J. Phys. Chem. 1961, 65, 284−292. (28) Verrall, R. E. In Water, A Comprehensive Treatise; Franks, F., Ed.; Plenum: New York, 1973; Vol. 3, p 211. (29) Draegert, D. A.; Williams, D. Far-Infrared Absorption Spectra of Aqueous Solutions of Strong Electrolytes. J. Chem. Phys. 1968, 48, 401−407. (30) Aliotta, F.; Pochylski, M.; Ponterio, R. C.; Saija, F.; Salvato, G.; Vasi, C. S. Structure of Bulk Water from Raman Measurements of Supercooled Pure Liquid and LiCl Solutions. Phys. Rev. B 2012, 86, 134301/1−134301/6. (31) Woutersen, S.; Bakker, H. J. Resonant Intermolecular Transfer of Vibrational Energy in Liquid Water. Nature 1999, 402, 506−509. (32) Bakker, H. J.; Skinner, J. L. Vibrational Spectroscopy As a Probe of Structure and Dynamics in Liquid Water. Chem. Rev. 2010, 110, 1498−1517. (33) Sokolowska, A. Raman Study of the Influence of KCl and KI on Polarization Contours of OH-Stretching Band of Water. J. Raman Spectrosc. 1996, 27, 621−624. (34) Lide, D. R. CRC Handbook of Chemistry and Physics, 79th ed.; CRC Press: Boca Raton, FL, 1998. (35) Hindman, J. C. Nuclear Magnetic Resonance Effects in Aqueous Solutions of 1−1 Electrolytes. J. Chem. Phys. 1962, 36, 1000−1015. (36) Max, J.-J.; Chapados, C. IR Spectroscopy of Aqueous Alkali Halide Solutions: Pure Salt-Solvated Water Spectra and Hydration Numbers. J. Chem. Phys. 2001, 115, 2664−2675. (37) Engel, G.; Hertz, H. G. On the Negative Hydration. A Nuclear Magnetic Relaxation Study. Ber. Bunsen-Ges. Phys. Chem. 1968, 72, 808−834. (38) Ishimura, M.; Uedaira, H. Natural-Abundance Oxygen-17 Magnetic Relaxation in Aqueous Solutions of Apolar Amino Acids and Glycine Peptides. Bull. Chem. Soc. Jpn. 1990, 63, 1−5.

2930


Comparative Study on the Properties of Hydration Water of Na-and K

Recommend Documents