effects of ions on the oh stretching band of water as revealed … · 2017. 8. 23. · raman...

Click here to load reader

Upload: others

Post on 14-Oct-2020

0 views

Category:

Documents


0 download

TRANSCRIPT

  • Effects of Ions on the OH Stretching Band of Wateras Revealed by ATR-IR Spectroscopy

    Norio Kitadai • Takashi Sawai • Ryota Tonoue • Satoru Nakashima •

    Makoto Katsura • Keisuke Fukushi

    Received: 16 December 2013 / Accepted: 12 March 2014 / Published online: 26 June 2014� Springer Science+Business Media New York 2014

    Abstract The effects of various cations (Li?, Na?, K?, Rb?, Cs?, Mg2?, Ca2?, Sr2?,

    Ba2?, Mn2?, Co2?, and Ni2?) and anions (Cl-, Br-, I-, NO�3 , ClO�4 , HCO

    �3 , and CO

    2�3 )

    on the molar absorptivity of water in the OH stretching band region (2,600–3,800 cm-1)

    were ascertained from attenuated total reflection infrared spectra of aqueous electrolyte

    solutions (22 in all). The OH stretching band mainly changes linearly with ion concen-

    trations up to 2 mol�L-1, but several specific combinations of cations and anions (Cs2SO4,Li2SO4, and MgSO4) present different trends. That deviation is attributed to ion pair

    formation and cooperativity in ion hydration, which indicates that the extent of the ion–

    water interaction reflected by the OH stretching band of water is beyond the first solvation

    shell of water molecules directly surrounding the ion. The obtained dataset was then

    N. Kitadai (&)Earth-Life Science Institute, Tokyo Institute of Technology, 2-12-1-IE-1, Ookayama, Megoro-ku,Tokyo 152-8550, Japane-mail: [email protected]

    T. SawaiNakatsugawa Factory Manufacturing, Mitsubishi Electric Corporation, 1213 Matsuoshirota, Iida,Nagano 395-0812, Japane-mail: [email protected]

    R. Tonoue � S. Nakashima � M. KatsuraDepartment of Earth and Space Science, Graduate School of Science, Osaka University, 1-1Machikaneyama, Toyonaka, Osaka 560-0043, Japane-mail: [email protected]

    S. Nakashimae-mail: [email protected]

    M. Katsurae-mail: [email protected]

    K. FukushiInstitute of Nature and Environmental Technology, Kanazawa University, Kakuma, Kanazawa,Ishikawa 920-1192, Japane-mail: [email protected]

    123

    J Solution Chem (2014) 43:1055–1077DOI 10.1007/s10953-014-0193-0

  • correlated with several quantitative parameters representing structural and dynamic

    properties of water molecules around ions: DGHB, the structural entropy (Sstr), the viscosityB-coefficient (Bg), and the ionic B-coefficient of NMR relaxation (BNMR). Results show

    that modification of the OH stretching band of water caused by ions has quasi-linear

    relations with all of these parameters. Vibrational spectroscopy can be a useful means for

    evaluating ion–water interaction in aqueous solutions.

    Keywords ATR-IR � Water � OH stretching band � Ion–water interaction

    1 Introduction

    Ascertaining the physicochemical properties of water in aqueous electrolyte solutions is an

    extremely attractive objective in natural sciences because of water’s ubiquitous presence in

    daily life and its importance in technical, chemical, and biological processes. Vibrational

    spectroscopy, i.e. IR and Raman spectroscopy, provides microscopic information related to

    the behaviors of water molecules in aqueous electrolyte solutions because the OH

    stretching band of water is sensitive to interactions between ions and hydration water

    molecules as well as hydrogen bonds between water molecules. To date, several experi-

    mental works have been performed to evaluate structural and dynamic properties of water

    molecules around ions from shifts and intensity changes of the OH stretching band of water

    caused by the ions [1–12]. The OH stretching band is generally regarded as comprising

    several components that are attributed to water molecules embedded in different hydrogen-

    bonded environments (Table 1). Components located at the lower frequency region are

    attributed to water molecules forming stronger hydrogen bonds, whereas those at the

    higher frequency region are attributed to water molecules forming weaker ones. Conse-

    quently, ions that shift the OH stretching band toward lower frequencies are typically

    interpreted as strengthening the hydrogen bonds between water molecules in the hydration

    layer because of tight ion–water interaction, whereas others are interpreted as weakening

    the hydrogen bonds. As one example of related research, Masuda et al. [1] performed curve

    fitting of the OH stretching bands of NaCl, NaHCO3, and Na2CO3 solutions with four

    Gaussian components that were attributed to water molecules having different mean

    hydrogen bond distances ranging from 0.273 to 0.295 nm. From the increase of higher- (or

    lower-) wavenumber components with increasing NaCl (or Na2CO3) concentration, they

    interpreted this to mean that a NaCl solution has longer H-bond distance characteristics,

    whereas a Na2CO3 solution has shorter ones.

    It is noteworthy that, as shown in Table 1, different researchers use different definitions

    for the number, position, and assignment of respective components. Consequently, the

    results of their studies are not always consistent. For instance, Li et al. [2] concluded

    through Gaussian fitting analyses of the Raman OH stretching band of MgCl2 and CaCl2solutions that Ca2?, Mg2?, and Cl- destroy the tetrahedral water structure through mutual

    interaction of water molecules and ions, whereas Chen et al. [5] interpreted that to mean

    the structure-making capability of Mg2? is much higher than those of Li? and Na? based

    on spectral comparisons among the ATR-IR spectra of Mg(ClO4)2, LiClO4, and NaClO4solutions.

    One reason for the lack of a definition of the OH stretching band of water is its strong

    dependence on experimental conditions and techniques. For instance, the parallel polarized

    1056 J Solution Chem (2014) 43:1055–1077

    123

  • Raman spectrum of water shows an OH stretching band having its maximum at

    3,400 cm-1 with a broad shoulder band around 3,250 cm-1, whereas the perpendicular

    polarized Raman spectrum of water shows a symmetrical one centered at 3,430 cm-1 [13–

    15]. As a response to dissolving KBr in water, the former band exhibits a shift toward

    higher frequency of the maximum, whereas the latter changes its maximum position only

    slightly, although the width at half maximum decreases concomitantly with increasing KBr

    concentration [14]. Extreme care should also be taken in interpreting the attenuated total

    reflection infrared (ATR-IR) spectrum of water because it is distorted strongly by the

    anomalous dispersion of aqueous solutions [16].

    Another cause arises from the fact that both cations and anions present in aqueous

    solutions influence the OH stretching band of water. Their influences generally strongly

    overlap and consequently are difficult to separate. Therefore, most reports of experimental

    studies have described the overall effect of cations and anions instead of their respective

    contributions. That overlap often complicates interpretation of the data and renders con-

    clusions ambiguous. Additionally, it should be considered that the OH stretching band of

    water has different sensitivity to cations and anions: anions, which interact directly with

    Table 1 Reported empirical definitions on the OH stretching band of water

    Position(cm-1)

    Assignment Technique Reference

    3,051 Fully four-hydrogen-bonded water molecules Raman Li et al. [2–4]

    3,233

    3,393 Partly hydrogen bonded water molecules

    3,511

    3,628 Free water molecules or free OH

    3,014 Single donor–double acceptor (DAA) Raman Sun [11]

    3,220 Double donor–double acceptor (DDAA)

    3,430 Single donor–single acceptor (DA)

    3,572 Double donor–single acceptor (DDA)

    3,636 Free OH symmetric stretching vibration

    3,080 Largest cluster with the shortest mean H-bond distance ATR-IR Masuda et at. [1]

    3,230 Larger cluster with shorter mean H-bond distance

    3,400 Smaller cluster with longer mean H-bond distance

    3,550 Smallest cluster with the longest mean H-bond distance

    3,230 An ice-like component ATR-IR Chen et al. [5], Liuet al. [7],Wei et al. [6],Guo et al. [12]

    3,420 An ice-like liquid component

    3,540 A liquid-like amorphous phase

    3,620 Monomeric water molecules

    3,248.9 Strongly hydrogen-bonded components Raman Dong et al. [9]

    3,468.4 Weekly hydrogen-bonded components

    3,628.8 Slightly hydrogen-bonded components

    3,242 Shorter H-bond component ATR-IR Kataoka et al. [10]

    3,428 Medium H-bond component

    3,562 Longer H-bond component

    3,251 Strong hydrogen bond TransmissionIR

    Zhao et al. [8]

    3,371 Weak hydrogen bond

    J Solution Chem (2014) 43:1055–1077 1057

    123

  • the hydrogen atoms of water molecules, more strongly affect the OH stretching band of

    water than cations do. The latter interact with the oxygen atoms of water molecules.

    Consequently, a direct interpretation of ion–water interactions based on the total contri-

    butions of cations and anions on the OH stretching band can engender misunderstandings.

    Raman analysis of Na2SO4 aqueous solutions revealed that SO2�4 influences the OH

    stretching band of water only slightly [17]. That fact was confirmed more recently by

    Wei et al. [6] through ATR-IR analysis of (NH4)2SO4 solutions. Based on their

    observations, Wei et al. [6] regarded SO2�4 as a ‘‘blank anion’’ that has no influence onthe OH stretching band of water. The effects of cations (Na?, Mg2? and Zn2?) were

    obtained by subtracting the appropriate amount of pure water spectrum from the spectra

    of aqueous sulfate solutions (i.e., Na2SO4, MgSO4, and ZnSO4) [6]. They also obtained

    the effect of ClO�4 by calculating the difference spectrum between perchlorate solutionsand sulfate solutions having the same cations at the same concentrations (e.g., NaClO4vs. Na2SO4).

    The aim of this study is to re-investigate whether and how the OH stretching band of

    water reflects structural and dynamic properties of the ion–water interaction in an aqueous

    solution. To this end, we first obtained ATR-IR spectra of many aqueous electrolyte

    solutions (22 in total). The electrolytes consist of the sulfate and chloride salts of mono-

    and divalent cations plus other electrolytes such as carbonates and nitrates. To avoid

    distortions of the ATR-IR spectra attributable to the optical effects, we extracted the molar

    absorptivity of water from these ATR-IR spectra using the methodology reported by Bertie

    and Eysel [18]. We then determined the effects of individual cations (Li?, Na?, K?, Rb?,

    Cs?, Mg2?, Ca2?, Sr2?, Ba2?, Mn2?, Co2?, and Ni2?) and anions (Cl-, Br-, I-, NO�3 ,

    ClO�4 , HCO�3 , and CO

    2�3 ) on the molar absorptivity of water using the effect of SO

    2�4 as a

    benchmark (i.e., the effects of all ions are presented as differences from that of SO2�4 ).Finally the obtained dataset was correlated with several quantitative parameters repre-

    senting structural and dynamic properties of water molecules around ions: DGHB, structuralentropy (Sstr), the viscosity B-coefficient (Bg), and the ionic B-coefficient of NMR relax-

    ation (BNMR). It will be demonstrated herein that modification of the OH stretching band of

    water caused by ions has quasi-linear relations with all of these parameters. The OH

    stretching band of water can thereby be used as a good indicator to evaluate the ion–water

    interactions occurring in an aqueous solution.

    2 Experimental

    2.1 Materials

    The electrolytes used for this study, presented in Table 2, were of analytical reagent

    quality, with purities greater than 99.0 % (except for NaClO4, of which the purity was

    99 %). They were used without further purification. Stock solutions (1 mol�L-1 for sulfatesolutions and 2 mol�L-1 for the others) were prepared by dissolving weighed amounts ofrespective solids in distilled deionized water. Lower-concentration solutions with con-

    centrations of 0.05 and 1.5 mol�L-1 were prepared by diluting the stock solutions withdistilled deionized water in a 5 mL volumetric flask. The resulting total mass was mea-

    sured so that the solution densities were obtained (Table 2). The obtained densities are

    similar to the values reported in the relevant literature: within 0.3 % [19].

    1058 J Solution Chem (2014) 43:1055–1077

    123

  • 2.2 ATR-IR Measurements

    ATR-IR measurements were performed using an FTIR spectrometer (FTIR-4200; Jasco

    Corp.) equipped with an MCT detector. A 45� ZnSe crystal (six reflections) attached to anATR plate (Benchmark ATR trough top plate; Specac Ltd.) with a horizontal ATR

    accessory was used for measurements. All spectra were obtained by collecting 1,024 scans

    with a spectral range of 400–7,800 cm-1 with 4 cm-1 resolution. The background spec-

    trum was first measured on the ATR plate without a sample. The ATR-IR spectra of sample

    solutions were then recorded as absorbance, -log10(I/I0) (where I0 is the background

    spectrum intensity), and defined here as pATR. A rubber lid was placed over the ATR plate

    during the measurements to prevent evaporation of the solutions (one measurement took

    about 11 min). For NaI and MnCl2 solutions, the ATR spectra were measured within 24 h

    after dissolving the respective solids in distilled deionized water to prevent the possible

    oxidation of I- and Mn2? by oxygen in the air. No change was observed in spectra

    measured several days later. Thus the influence of oxidation was negligible. The ATR-IR

    measurements were replicated several times to confirm that the obtained spectra were

    reproducible in terms of peak position and intensity. The maximum error range in the OH

    stretching band region (2,600–3,800 cm-1) of the obtained spectra was estimated as

    ±0.0015 in pATR units (dimensionless).

    All experiments presented above were carried out at 20 �C, which was controlledby an air-conditioner. The accuracy of temperature is supported by the good consis-

    tency of solution densities measured in this study (Table 2) with the reported coun-

    terparts [19].

    2.3 Extraction of the Molar Absorptivity of Water from the ATR-IR Spectra

    of Aqueous Electrolyte Solutions

    The molar absorptivity of water was extracted from the ATR-IR spectra of aqueous

    electrolyte solutions using the procedure reported by Bertie and Eysel [18]. The procedure,

    which uses Fresnel equations for reflection, calculates the real (n(m)) and imaginary (k(m))refractive index of the sample.

    First, k(m) is calculated from the pATR spectra on the assumption that the penetrationdepth of the evanescent wave at each reflection is calculable from the standard formula for

    the distance for the evanescent wave to decrease in magnitude by 1/e in a non-absorbing

    sample. This is:

    d ¼ k2pn1

    ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

    sin2h� n2n2r

    q ; ð1Þ

    where k is the wavelength of radiation in a vacuum, n and nr respectively denote the realrefractive indices of a sample solution and the ZnSe crystal, and h is the angle of incidence,45�. The real refractive index of ZnSe as a function of the wavenumber was taken from theliterature [20]. The real refractive indices of sample solutions were assumed to be constant.

    Their values at the sodium D line (589 nm), nD, are used (Table 2).

    Based on the assumptions presented above, the path length at each reflection is 2d. If m

    is the effective number of reflections, then:

    k mð Þ ¼ ln104pmm 2dð Þ pATRðmÞ; ð2Þ

    J Solution Chem (2014) 43:1055–1077 1059

    123

  • where m is not simply the actual number of reflections because it contains contributions

    from the approximations in Eq. 1 and experimental factors. We set m = 4.87 here so that

    the maximum of the OH stretching band of water becomes the same as that reported by

    Bertie and Lan [21] (Fig. 1). The approximate values of k(m) were then replaced by:

    k mð Þ ¼ k mð Þ 1þffiffiffiffiffiffiffiffiffi

    k mð Þp

    � �

    ð3Þ

    to improve the values calculated from Eq. 2 [18, 22].

    Next, n(m) was calculated using the Kramers–Kronig transformation:

    n mað Þ ¼ n1 þ2

    pP r1

    0

    mkðmÞm2 � m2a

    dm ð4Þ

    Therein, m denotes the wavenumber of radiation in a vacuum and n? is the realrefractive index of sample solutions at infinite wavenumber. The values of k(m) outside ofthe experimental spectral region (400–7,800 cm-1) were set to zero. The values of nD were

    used as n? because the sodium D line (about 17,000 cm-1) can be regarded as that at near

    infinite frequency because it is far from any vibrational or electronic absorption band. The

    nD values that were not available in the literature [19, 23, 24] were obtained by extrapo-

    lation on the basis that nD is additive for ions in water [25–27]. The nD values of Li2SO4,

    Na2SO4, and K2SO4 solutions obtained using this method (Table 2) agree with the liter-

    ature values within 0.2 % [28, 29].

    In the following step k(m) is refined. The refractivity for light polarized parallel to theinterface between the ZnSe crystal and a sample solution, Rs, is:

    Rs ¼cosh�

    ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

    nþiknr

    � �2

    �sin2hr

    coshþffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

    nþiknr

    � �2

    �sin2hr

    2

    ð5Þ

    and the reflectivity for the perpendicular polarization, Rp, is:

    Rp ¼nþik

    nr

    � �2

    cosh�ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

    nþikn1

    � �2

    �sin2hr

    nþiknr

    � �2

    coshþffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

    nþiknr

    � �2

    �sin2hr

    2

    ð6Þ

    Mol

    ar a

    bsor

    ptiv

    ity [m

    ol–1

    ·cm

    –1]

    Wavenumber [cm–1]

    8001600240032004000

    20

    40

    60

    80

    100

    0

    This studyBertie and Lan (1996)

    Fig. 1 Comparison between themolar absorptivity of watercalculated in this study (solidline) and that reported by Bertieand Len [21] (dashed line) in thespectral range of800–4,000 cm-1

    1060 J Solution Chem (2014) 43:1055–1077

    123

  • They were used to calculate the pATR spectrum as presented below:

    pATRcalc mð Þ ¼ �log101

    2Rmp þ Rms� �

    � �

    ð7Þ

    The calculated pATR spectrum was compared with the experimental one. k(m) wasadjusted as:

    k mð Þ ¼ k mð Þ � pATRðmÞpATRcalcðmÞ

    � �

    ð8Þ

    We repeated the calculations from Eqs. 4–8 up to 20 cycles so that the square of the

    difference between calculated and observed pATRs,P

    j

    pATR mj�

    � pATRcalc mj� �2

    ,

    became less than 1 9 10-4.

    Finally, the molar absorptivity of water, e(m), was calculated using the refined k(m) as

    eðmÞ ¼ 4pkðmÞkcln10

    ; ð9Þ

    where c is the concentration (mol�L-1) of water in a sample solution calculated using theexperimentally determined densities of sample solutions. The obtained molar absoptivity

    of water in distilled deionized water is presented in Fig. 1 together with that reported by

    Bertie and Lan [21]. Their spectra are similar except for the region below 1,000 cm-1. The

    difference below 1,000 cm-1 arises because the low-frequency cutoff of the ZnSe crystal

    prevents one from obtaining pATR spectra below 700 cm-1. Bertie and Lan [21] overcame

    the limitation by adding the known k(m) spectrum of water below 700 cm-1 to that cal-culated from the pATR spectrum. Its influence on the OH stretching band region appears to

    be negligible (Fig. 1). Therefore, we use the calculated e(m) spectrum without the furtherrefinements performed by Bertie and Lan [21].

    2.4 Extraction of the Effects of Individual Ions on the OH Stretching Band of Water

    The effects of individual ions on the OH stretching band of water, which we designate as

    ‘‘effect’’, were extracted from the e(m) spectra as follows. First, assuming the effect of SO2�4as a benchmark, the effects of Li?, Na?, K?, Rb?, Cs?, and Mg2? were ascertained merely

    by calculating the difference spectrum between their sulfate solutions and distilled

    deionized water (subtraction factor = 1) (Fig. 2a–e, i).

    The effects of anions (Cl-, Br-, I-, NO�3 , ClO�4 , HCO

    �3 , and CO

    2�3 ) were then

    obtained by subtracting the effects of Na? or K? from the spectra of their respective

    sodium or potassium salt solutions (e.g., NaCl, KHCO3) after subtracting the spectrum of

    distilled deionized water (Fig. 2n–s). The effects of 1 mol�L-1 Na? or 1 mol�L-1 K?were used here because the bands observed in the effects of Na? and K? showed

    increasing linear intensity with ion concentration within the range of error (±0.1)

    (Fig. 2b, c). Complete subtraction was achieved by multiplying the effects of 1 mol�L-1Na? (or 1 mol�L-1 K?) by a subtraction factor (e.g., the effect of 1.5 mol�L-1 Cl- = thespectrum of 1.5 mol�L-1 NaCl—the spectrum of distilled deionized water—(the effect of1 mol�L-1 Na?) 9 1.5).

    Regarding Ca2?, Sr2?, and Ba2?, the solubilities of their sulfates in water are low

    (1.5 9 10-2, 6.2 9 10-4, 1.0 9 10-5 mol�L-1 at 20–30 �C, respectively) [30]. Theireffects were therefore obtained from the spectra of their chloride solutions by subtracting

    J Solution Chem (2014) 43:1055–1077 1061

    123

  • the effect of Cl- (Fig. 2j–l). The effect of 1 mol�L-1 Cl- was used. It was multiplied by asubtraction factor so that the effect of Cl- was subtracted completely (e.g., the effect of

    1.5 mol�L-1 Ca2? = the spectrum of 1.5 mol�L-1 CaCl2—the spectrum of distilleddeionized water—(the effect of 1 mol�L-1 Cl-) 9 3). The error in the obtained spectraincreases as the subtraction factor of the effect of Cl- increases. The range of error

    21.5

    –1.8

    –1

    –0.2

    0.6

    1.4

    3335

    3580

    3665

    +

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    2900320035003800

    –1.2

    –0.6

    0

    0.6

    1.2

    3335

    3560

    3665

    3120

    +

    0.53(±0.03)

    –0.56(±0.01)

    –1.07(±0.04)

    3580 cm–13665 cm–1

    3335 cm–1

    3310 cm–1

    3665 cm–13550 cm–1

    0.78(±0.06)

    –0.42(±0.004)

    –1.06(±0.06)

    –1.3

    –0.6

    0.1

    0.8

    1.5

    3310

    3550

    3665

    +

    –1.7

    –0.8

    0.1

    1

    1.9

    3280

    3535

    3665

    +

    3535 cm–13665 cm–1

    3280 cm–1

    0.80(±0.01)

    –0.42(±0.01)

    –1.01(±0.02)

    0 0.5 1

    3335 cm–1

    3665 cm–13560 cm–1

    0.47(±0.01)

    –0.50(±0.03)

    –0.70(±0.01)

    3120 cm–1

    –0.9

    0.1

    1.1

    2.1

    3.1

    3655

    3170+

    3170 cm–1

    3655 cm–1

    –0.71(±0.01)

    Wavenumber [cm–1] Conc. [mol/L]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    3645

    3180

    3535

    2+

    –3

    1

    5

    9

    133180 cm–1

    3645 cm–1

    –2.09(±0.02)

    3535 cm–1

    –3

    –0.5

    2

    4.5

    7

    3655

    3175

    3515

    2+3175 cm–1

    3515 cm–1

    3655 cm–1

    3.24(±0.04)

    –0.42(±0.01)

    –1.84(±0.02)

    –2.5

    –1

    0.5

    2

    3.53145 cm–1

    3555 cm–1

    3655 cm–1

    1.61(±0.04)

    0.68(±0.02)

    –1.58(±0.02)

    3655

    3145

    3555

    2+

    3660

    3135

    3570

    3390

    2+ 1.14(±0.01)

    0.55(±0.02)

    –1.72(±0.02)

    –1.60(±0.02)

    3135 cm–13390 cm–13570 cm–13660 cm–1

    –5

    1

    7

    13

    19

    3645

    3175

    3505

    2+

    9.9(±0.2)

    3175 cm–1

    3505 cm–1

    3645 cm–1

    –3.04(±0.03)

    –3.38(±0.02)

    –4

    3

    10

    17

    24

    3640

    3165

    3510

    2+3165 cm–1

    3510 cm–1

    3640 cm–1

    10.5(±0.1)

    –3.27(±0.02)

    –3.60(±0.06)

    –6

    1

    8

    15

    22

    3640

    3165

    3505

    2+3165 cm–1

    3505 cm–1

    3640 cm–1

    10.2(±0.1)

    –3.56(±0.01)

    –4.57(±0.10)

    –2.3

    –1.1

    0.1

    1.3

    2.5

    –3.52900320035003800 0 0.5 1 21.5

    Wavenumber [cm–1] Conc. [mol/L]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    –5.5

    –3

    –0.5

    2

    4.5

    3215

    3475 – 3215 cm–1

    3475 cm–1

    2.33(±0.3)

    –3.18(±0.03)

    –7

    –3

    1

    5

    9

    3255

    3500–

    3255 cm–1

    3500 cm–1

    3.30(±0.02)

    –4.20(±0.07)

    –10

    –5

    0

    5

    10 3525

    3285

    –3285 cm–1

    3525 cm–1

    –5.89(±0.07)

    4.43(±0.02)

    –6.5

    –3

    0.5

    4

    7.5 3580

    3265

    3–

    3265 cm–1

    3580 cm–1

    –3.75(±0.09)

    3.26(±0.04)

    –10

    –3

    4

    11

    183315 cm–1

    3600 cm–1

    –6.13(±0.14)

    8.58(±0.09)

    3600

    3315

    4–

    0.5

    1.5

    2.5

    3.5

    4.53585 cm–1

    3075 cm–1

    2.09(±0.02)

    0.75(±0.02)

    0.67(±0.03)

    2890 cm–13075

    35852890

    3460

    3045

    3–

    (d) Rb

    (e) Cs

    (c) K

    (b) Na

    (a) Li

    (i) Mg

    (j) Ca

    (k) Sr

    (l) Ba

    (f) Mn

    (g) Co

    (h) Ni

    (n) Cl

    (m) Br

    (o) I

    (p) NO

    (q) ClO

    (r) HCO

    (s) CO32–3460 cm–1

    3045 cm–1

    7.64(±0.14)

    –5.36(±0.04)

    –9

    –3

    3

    9

    15

    –152900320035003800 0 0.5 1 21.5

    Wavenumber [cm–1] Conc. [mol/L]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Mol

    . Abs

    . [m

    ol–1

    ·cm

    –1]

    Fig. 2 Effects of univalent cations (a–e), divalent cations (f–l) and various anions (n–s) on the molarabsorptivity of water in the OH stretching band region (2,600–3,800 cm-1). The right sides of the respectivefigures show band intensities observed in effects as a function of the ion concentration. Vertical scales differfor each effect of ions

    1062 J Solution Chem (2014) 43:1055–1077

    123

  • estimated in the effect of Ca2?, Sr2?, and Ba2? consequently increased from about ±0.1 at

    0.1 mol�L-1 up to ±0.4 at 2 mol�L-1.The effects of Mn2?, Co2? and Ni2? were obtained using the same procedure as that

    used for Ca2?, Sr2?, and Ba2? (Fig. 2f–h), although the solubilities of their sulfates in

    water are sufficiently high to be analyzed because their effects might be distorted by

    interactions between these cations and SO2�4 in aqueous solutions (see Sect. 3).

    3 Results and Discussion

    3.1 Effects of Ions on the OH Stretching Band of Water

    Figure 2 shows the effects of univalent cations (Li?, Na?, K?, Rb? and Cs?; Fig. 2a–e),

    divalent cations (Mn2?, Co2?, Ni2?, Mg2?, Ca2?, Sr2? and Ba2?; Fig. 2f–l) and various

    anions (Cl-, Br-, I-, NO�3 , ClO�4 , HCO

    �3 and CO

    2�3 ; Fig. 2m–s) on the OH stretching band

    of water. The concentration dependences of the band intensities observed in each effect are

    also shown at the right side of each figure. The band intensities were measured after

    drawing a linear baseline from 2,600 to 3,800 cm-1, except for the effects of HCO�3 and

    CO2�3 for which the linear baselines were drawn, respectively, from 2,150 to 3,800 cm-1

    and from 2,050 to 3,800 cm-1. In most cases, the band intensities increased linearly with

    ion concentration in the investigated concentration range (0.1–2 mol�L-1). This linearrelation shows that the effects of ions are additive. Therefore, the extraction procedures to

    obtain effects of individual ions used for this study are valid. However, several band

    intensities observed in the effects of Cs?, Li? and Mg2? (Fig. 2a, e, i) departed downward

    from the linear relation with concentration. The deviations became greater as their con-

    centrations increased. The result suggests that the effects of these three cations are non-

    additive. However, because these effects were all derived from sulfate solutions, and

    because the effects derived from the other salt solutions (i.e., chloride) showed linear

    increases of band intensities, another possibility is that SO2�4 selectively influenced theeffects of Li?, Cs? and Mg2?, and that it does not influence the other cations (i.e., Na?,

    K?, and Rb?).

    To obtain more evidence of the possible influence of SO2�4 , we extracted the effects ofMg2? and Li? from the spectra of their chloride solutions by subtracting the effect of Cl-.

    Subsequently, we compared them with the corresponding ones derived from sulfate

    solutions (Fig. 3). The band maxima are located at similar positions irrespective of the

    extraction procedure, whereas the effects derived from chloride solutions show linear

    increases of the band intensities with concentration (Fig. 3c, g). To evaluate the differences

    in band intensity for sulfate and chloride solutions quantitatively, we calculated the

    intensity ratios (sulfate/chloride) of the 3,180 and 3,530 cm-1 bands in the effect of Mg2?,

    and that for the 3,170 cm-1 band in the effect of Li? at the same Mg2? (or Li?) con-

    centration (Fig. 3d, h). The ratios are near one at the lowest concentration (0.1 mol�L-1).They decrease gradually as the cation concentrations increase. Additionally, for the effect

    of Li?, a broad shoulder band around 3,500 cm-1 is not significant when the effect was

    extracted from the chloride solution (Fig. 3f).

    It is noteworthy that these differences are not solely attributable to the overlapping of

    the effect of SO2�4 , which was ignored in the extraction procedure of the effects of Mg2?

    and Li?. Even if SO2�4 has some influence on the OH stretching band profile of water, itscontribution to the effects of Mg2? (and Li?) obtained using the two extraction procedures

    J Solution Chem (2014) 43:1055–1077 1063

    123

  • is the same. For instance, the spectrum of 1 mol�L-1 MgSO4 is regarded as including theeffect of 1 mol�L-1 Mg2? and 1 mol�L-1 SO2�4 , if SO2�4 is assumed to have an effect. Thespectral procedure to extract the effect of 1 mol�L-1 Mg2? from the spectrum of 1 mol�L-1MgCl2 yields the following:

    1 mol � L�1MgCl2� 1 mol � L�1NaCl � 0:5 mol � L�1Na2SO4�

    � 2¼ the effect of 1 mol � L�1Mg2þ þ the effect of 1 mol � L�1SO2�4

    ð10Þ

    The resultant spectra therefore consist of the same contributions of 1 mol�L-1 Mg2? and1 mol�L-1 SO2�4 . The spectral difference indicates that the effects of Mg

    2? and Li? are

    distorted in the presence of SO2�4 . In other words, the effects of ions are non-additive forsome specific combinations of cations and anions.

    A possible cause of the non-additivity is the formation of ion pairs. Mg2? is known to

    form a strong ion pair [contact ion pair (CIP)] with SO2�4 in aqueous solutions. Theconcentration dependence of the fraction of Mg2? present as CIPs in MgSO4 solutions was

    determined quantitatively by Hefter and co-authors [31–33]. They revealed that the frac-

    tion increases rapidly to about 10 % as the MgSO4 concentration increases from 0 to

    0.5 mol�L-1, then it increases moderately to about 13 % at 2.0 mol�L-1. They alsodemonstrated the existence of a triple or more aggregated ion pair (e.g., Mg2SO

    2þ4 ) at high

    MgSO4 concentration ([1 mol�L-1). However, the CIP between Mg2? and Cl- wasevaluated theoretically as energetically unfavorable and was formed only slightly in

    aqueous solutions [34, 35]. It is expected that the penetrations of SO2�4 into the hydrationsphere of Mg2? through the CIP formation partly pushes out the hydration waters of Mg2?

    0.9

    1

    1.1

    0.8

    0.7

    0.8

    0.85

    0.9

    0.95

    1

    0.75

    0.71.510.50 2

    0.5

    0.6

    0.7

    0.8

    1

    0.4

    0.3

    0.9

    Mg2+ Conc. [mol/L]

    Mol

    ar a

    bsor

    ptiv

    ity [m

    ol–1

    ·cm

    –1]

    1

    6

    11

    16

    21

    –4

    –9300034003800

    Wavenumber [cm–1]

    300034003800

    (a) Mg2+ (in MgSO4) (b) Mg2+ (in MgCl2)

    0 0.5 1 21.5

    Mg2+ Conc. [mol/L]

    0.5

    2

    3.5

    5

    –1

    –2.5

    (d) Li+ (in Li2SO4) (e) Li + (in LiCl)

    Li+ Conc. [mol/L]

    0 0.5 1 21.5300034003800

    Wavenumber [cm–1]

    300034003800

    Mol

    ar a

    bsor

    ptiv

    ity [m

    ol–1

    ·cm

    –1]

    1.510.50 2

    Li+ Conc. [mol/L]In

    tens

    ity r

    atio

    (I M

    gSO

    4 / I

    MgC

    l2)

    at 3

    180

    cm-1

    Intensity ratio (IMgS

    O4 / IM

    gCl2 )

    at 3535 cm-1

    Inte

    nsity

    rat

    io (

    I Li2

    SO

    4 / I

    LiC

    l)at

    317

    0 cm

    -1

    (c)3180 (MgSO4)3180 (MgCl2)

    3535 (MgSO4)3535 (MgCl2)

    (f)3170 (Li2SO4)3170 (LiCl)

    (g)3180 cm-1

    3535 cm-1

    (h)

    Fig. 3 a, b Spectral comparisons of the effects of Mg2? extracted from the spectra of MgSO4 solutions(a) and those of MgCl2 solutions (b). c, g Intensities (c) and the intensity ratios (g) of the band maxima at3,180 cm-1 and at 3,535 cm-1 observed in a and b as a function of the Mg2? concentration. d, e Spectralcomparisons of the effects of Li? extracted from the spectra of Li2SO4 solutions (d) and those of LiClsolutions (e). f, h intensities (f) and the intensity ratios (h) of the band maxima at 3,170 cm-1 observed ind and e as a function of the Li? concentration. Dashed lines in g and h are included only as a visual aid

    1064 J Solution Chem (2014) 43:1055–1077

    123

  • and consequently induces a decrease of the effect of Mg2?. Decreases of the intensity ratios

    (sulfate/chloride) with increasing Mg2? concentration (Fig. 3d) therefore probably reflect a

    decrease in the number of hydrated water molecules around Mg2? in MgSO4 solutions,

    compared to those in MgCl2 solutions, caused by the increase of the fraction of CIP.

    The formation of ion pairs alone, however, cannot fully explain the observed spectral

    difference because Li? does not form the CIP with SO2�4 as well as it does with Cl- [36].

    Additionally, results show that the fraction of Li? present as weaker ion pairs [i.e., double-

    solvent-separated ion pair (2SIP) and solvent-shared ion pair (SIP)] are not significantly

    different between those in LiCl solution and those in Li2SO4 solution at the same Li?

    concentration [36].

    Another possible cause is the cooperativity in ion hydration, which is observed only

    when the cations and anions in aqueous solutions are both strongly hydrated [37].

    Cooperativity occurs because the cation and anion lock in different degrees of freedom

    of water molecules. The local electric field around the cation causes the dipole vector of

    water molecules in the solvation shell to point radially away from the cation, whereas for

    an anion one OH group of a hydrogen-bonded water molecule points linearly toward the

    anion. The nearby presence of the strongly hydrated cation (e.g., Mg2?) and anion (e.g.,

    SO2�4 ) can thereby engender a locking in of both directions of the hydrogen bondstructure of several intervening water layers. However, if the counter ion is weakly

    hydrated, then the strongly hydrated ion is surrounded by a semi-rigid solvation shell,

    where the dynamics of water molecules are restricted only in a certain direction but is

    unrestricted in other directions. Such differences in hydration nature probably cause the

    different spectral profiles of Li? between a Li2SO4 solution and a LiCl solution. The

    extent of ion–water interactions reflected by the OH stretching band of water is therefore

    expected to be beyond the first solvation shell of water molecules directly surrounding

    the ion.

    Tielrooij et al. [37] also observed cooperativity even in the combination of the mod-

    erately hydrated cation Na? with the strong anion SO2�4 . The effect of Na? derived from

    Na2SO4 (Fig. 2b) might therefore also be modified by SO2�4 . In fact, the negative broad

    band around 3,280 cm-1 in the effect of 2 mol�L-1 Na? appears to be somewhat asym-metric compared to that in the effect of 1 mol�L-1 Na?. It is noteworthy, however, that thedifference is near the error level and that its resultant influence on the effects of the other

    ions that were extracted using the effect of 1 mol�L-1 Na? (e.g., Cl- and Br-) is expectedto be slight. Evidence supports the expectation that the effect of 1 mol�L-1 Cl- obtainedusing the spectral combination of 1 mol�L-1 KCl and 0.5 mol�L-1 K2SO4 coincides withthat obtained using the spectra of 1 mol�L-1 NaCl and 0.5 mol�L-1 Na2SO4, within therange of error (data not shown). Further detailed experiments might provide useful

    information to elucidate the cooperative interaction between Na? and SO2�4 in aqueoussolutions. That discussion is beyond the scope of this study. Therefore, we will not

    investigate the slight influence of SO2�4 on the effect of Na? further.

    3.2 Correlations Between the OH Stretching Band of Water and Structural

    and Dynamic Properties of the Ion–Water Interaction in Aqueous Solution

    This section presents an examination of whether and how the OH stretching band of water

    reflects structural and dynamic properties of the ion–water interaction in aqueous solution.

    Figure 4 presents a comparison of the effects of cations (Fig. 4a) and anions (Fig. 4b) at

    the concentration of 1 mol�L-1. We also show the effects of these ions on the OH bending

    J Solution Chem (2014) 43:1055–1077 1065

    123

  • band of water in the Appendix (Fig. 7). The effects of Mg2? and Li? presented in these

    figures were extracted from the spectra of their chloride solutions.

    Spectral profiles of the effects of Mn2?, Co2?, Ni2?, and Mg2? in Fig. 4a appear to

    comprise a large positive band centered around 3,170 cm-1, a broad shoulder band around

    3,310 cm-1, and two negative bands around 3,640 and 3,510 cm-1. Based on previously

    reported interpretations of the OH stretching band of water (Table 1), the former two bands

    Mol

    ar a

    bsor

    ptiv

    ity [m

    ol–1

    ·cm

    –1]

    3

    8

    13

    18

    23

    –2

    28

    33

    38

    43

    48

    53

    Mg2+

    Ca2+

    Sr2+

    Ba2+

    Li+

    Na+

    K+

    Rb+

    Cs+

    Mn2+

    Ni2+

    Co2+

    25

    50

    75

    100

    0

    2

    12

    22

    32

    42

    –2

    52

    Wavenumber [cm–1]

    2600300034003800 3600 3200 2800

    25

    50

    75

    100

    0

    Wavenumber [cm–1]

    2600300034003800 3600 3200 2800

    Mol

    ar a

    bsor

    ptiv

    ity [m

    ol–1

    ·cm

    –1]

    NO3–

    ClO4–

    HCO3–

    CO32–

    Cl–

    I–

    Br–

    (a) (b)

    Fig. 4 Effects of 1 mol�L-1 cations (a) and 1 mol�L-1 anions (b) on the molar absorptivity of water in theOH stretching band region (2,600–3,800 cm-1); the OH stretching band of water is shown at the top of eachfigure for comparison

    1066 J Solution Chem (2014) 43:1055–1077

    123

  • are attributed to water molecules forming stronger hydrogen bonds, whereas the latter are

    attributed to water molecules forming weaker ones. Consequently, because of the increase

    of the stronger hydrogen bond components and the decrease of the weaker ones, interac-

    tions of the four cations with surrounding water molecules are undoubtedly attributable to

    strengthening of the water–water hydrogen bonds in their hydration layers. For the other

    cations, the lowest frequency band (about 3,170 cm-1) diminishes gradually in the order of

    Ca2? [ Sr2? * Li? [ Ba2? [ Cs? and was not observed in the effects of Na?, K?, andRb? (Fig. 4a). The highest frequency band (about 3,640 cm-1), however, was observed in

    all effects of cations investigated. The water–Ca2? (or Li?) interaction can still be

    attributable to strengthening of the water–water hydrogen bonds, although for the other

    cations, particularly Cs? and Ba2?, the interpretation is difficult because these cations

    show positive (or negative) bands on both regions of the stronger and weaker hydrogen

    bond components.

    The effects of anions showed simple spectral profiles compared with those of cations.

    The effects of Cl-, Br-, I-, NO�3 , and ClO�4 showed, respectively, a negative band and a

    positive band on the region of the stronger and weaker hydrogen bond components

    (Fig. 4b). Interactions of the five anions with surrounding water molecules are therefore

    interpreted as the result of weakening of the water–water hydrogen bonds in their

    hydration layers. The effects of HCO�3 and CO2�3 present spectral features distinct from

    those of the other anions: the effect of HCO�3 showed a broad positive band around3,075 cm-1. Its intensity increases strongly as the proton is dissociated

    (HCO�3 ! CO2�3 ). Therefore, CO2�3 is expected to cause considerably stronger water–water hydrogen bonds.

    The discussion presented above is only qualitative. To correlate the spectral profiles

    with structural and dynamic properties of the ion–water interaction in aqueous solution

    quantitatively, some clear definition is necessary to quantify the effects of ions. Curve-

    fitting analysis might be an appropriate technique to establish the definition. It is note-

    worthy, however, that curve fittings with several Gaussians imply that water molecules are

    embedded in several distinct hydrogen bond environments. The mixture model of water is

    supported by isosbestic points observed in temperature-dependent spectra of water [13].

    However, Monte Carlo simulations by Smith et al. [38] show that a continuous distribution

    of hydrogen bond geometries and energies in water also generates temperature-indepen-

    dent isosbestic points. An ultrafast IR investigation of dilute HDO in D2O solutions also

    engendered the conclusion that a continuum, rather than mixture models, is more appro-

    priate [39, 40]. Therefore, we did not conduct any curve fitting analyses in the present

    study.

    As an alternative, we divided the spectral profiles into two regions of 2,600–3,420 and

    of 3,420–3,800 cm-1, and then calculated the difference in area between the lower and

    higher frequency regions after drawing a linear baseline from 2,600 to 3,800 cm-1

    (DMAL–H). For this study, 3,420 cm-1 was chosen because it is near the center of the OH

    stretching band of water (Fig. 4) and because the position is near the boundary between the

    two regions of the stronger and weaker hydrogen bond components defined qualitatively

    above. The value for each ion (at 1 mol�L-1 concentration) is presented in Table 3 withtheir maximum errors estimated from several independent measurements. Based on the

    interpretation presented above, ions having positive DMAL–H values are expected tostrengthen the water–water hydrogen bonds in their hydration layers, whereas those ions

    having negative DMAL–H weaken them. In fact, Mn2?, Co2?, Ni2?, and Mg2? show large

    positive DMAL–H values whereas Cl-, Br-, I-, NO�3 , and ClO

    �4 showed negative ones.

    J Solution Chem (2014) 43:1055–1077 1067

    123

  • Ca2? and Li? show small but definitely positive DMAL–H values, but those of Ba2? and

    Cs? are nearly zero.

    As quantitative parameters representing structural and dynamic properties of water

    molecules around ions, we selected the following four parameters: DGHB, structuralentropy (Sstr), the viscosity B-coefficient (Bg), and the ionic B-coefficient of NMR relax-

    ation (BNMR).

    DGHB is a dimensionless parameter representing the effect of ions on the averagenumber of hydrogen bonds in which a water molecule participates [41]. DGHB is calculatedusing DtrG (solute, H2O ? D2O), the standard molar Gibbs energy of transfer of the solutefrom light to heavy water:

    DGHB ¼ DtrG solute; H2O! D2Oð Þ = ð�929Þ ð11Þ

    where -929 J�mol-1 is the molar difference in the hydrogen bonding energies of light andheavy water.

    The structural entropy, Sstr, was defined by Marcus [42] to represent the effects of

    ions on the structure of water. The value of Sstr is calculated using the standard molar

    entropy of ion hydration (DhydS) by subtracting the contribution of the ionic hydrationshell formation:

    DSstr ¼ DhydS� DSnt þ DSel1 þ DSel2ð Þ ð12Þ

    Therein, DSnt represents the formation of the cavity in water for the accommodation ofthe ion as well as the dispersion interactions of a neutral entity having the same size as the

    ion (e.g., a rare gas atom) with surrounding water molecules. DSel1 and DSel2, respectively,represent the electrostatic interactions of the ion with water molecules in the hydration

    shell and beyond it. The Sstr values are often used as a quantitative measure to classify ions

    into structure makers and breakers. The ions having negative Sstr values are classified as

    structure makers and vice versa [41, 42].

    The viscosity B-coefficient (Bg) is the slope of the viscosity of an aqueous electrolyte

    solution (g) against concentration:

    g ¼ gw 1þ Ac1=2 þ Bcþ � � �� �

    ð13Þ

    where gw stands for the viscosity of water. Also, c represents the molar concentration ofelectrolyte. Coefficient A represents interionic (electrostatic) forces that maintain a space

    lattice structure of the electrolyte. The coefficient B is representative of the retardation of

    solution flow doe to hydration on ions. In this study, the values of Bg calculated using the

    assumption of Marcus [42] that B(Rb?) = B(Br-).

    The ionic B-coefficient of NMR relaxation (BNMR) was defined by Engel and Hertz [43]

    in order to correlate the NMR longitudinal proton relaxation times in aqueous electrolyte

    solution (T1) with those in neat water (T1*), using an expression analogous to Eq. 13:

    1=T1ð Þ= 1=T1�ð Þ�1½ � ¼ BNMRc þ � � � ð14ÞThe convention that BNMR(K

    ?) = BNMR(Cl-) is used to obtain the ionic values. The

    rotational correlation times of water molecules (s) are given similarly as:

    sion=sW ¼ 1þ 55:51=nionð ÞBNMR ð15ÞSubscripts ion and W, respectively, denote hydration and bulk water. In addition, nion is

    the hydration number of the ion. The NMR measurements of longitudinal proton relaxation

    1068 J Solution Chem (2014) 43:1055–1077

    123

  • times, T1, are confined to diamagnetic ions. Therefore, the BNMR values for transition metal

    cations (i.e., Mn2?, Co2?, and Ni2?) were not reported in Engel and Hertz [43].

    Figure 5 presents correlations between DMAL–H and the four parameters (DGHB, Sstr,Bg, and BNMR). It is readily apparent that DMAL–H has quasi-linear relations with all ofthese parameters. The values of DMAL–H increase concomitantly with increasing DGHB,Bg, and BNMR (Fig. 5a, c, d), whereas they decrease concomitantly with increasing Sstr(Fig. 5b). In the former correlations, most of the ions occupy the two regions of the upper

    right quadrant (DMAL–H [ 0 and DGHB, Bg and BNMR [ 0) and the lower left quadrant(DMAL–H \ 0 and DGHB, Bg and BNMR \ 0) of the diagram, whereas in the DMAL–H -Sstr correlation, the ions mainly occupy the upper left quadrant (DMAL–H \ 0 and Sstr [ 0)and the lower right quadrant (DMAL–H [ 0 and Sstr \ 0). These observations are consistentwith the qualitative interpretation presented above: as the ion–water interaction becomes

    stronger, water molecules around the ion are constrained more tightly (Bg and BNMR [ 0)and the strengths and energies of water–water hydrogen bonds increase (DMAL–H andDGHB [ 0). Consequently, the entropy of water decreases (Sstr \ 0). Conversely, as theion–water interactions sufficiently weaken to become comparable to those for water–water

    hydrogen bonds, water molecules around the ions become freer to move than bulk water

    (Bg or BNMR \ 0). Then the hydrogen bond strengths and energies decrease (DMAL–H \ 0and DGHB \ 0). Consequently, the entropy of water increases (Sstr [ 0). Although furthertheoretical and/or experimental works are necessary to explain the observed correlations

    quantitatively, the quasi-linear relations (Fig. 5) are strong evidence that the OH stretching

    band of water does reflect the structural and dynamic behaviors of water molecules around

    –4000 –2000 0 2000 4000 6000

    ΔMAL–H

    ΔGH

    B

    1.2

    0.8

    0.4

    0

    –1.2

    –0.8

    –0.4

    Co2+Ni2+

    Mg2+

    Mn2+

    CO32–Ca2+Li

    +Sr2+

    HCO3–

    Ba2+

    Na+

    SO42–

    Cs+Rb+K

    +Cl–NO3–

    Br–ClO4

    I–

    y = 2.53×10–4x – 0.255R2 = 0.841

    (a) vs. ΔGHB (b) vs. Sstr

    –4000 –2000 0 2000 4000 6000

    ΔMAL–H

    Sst

    r [J

    K–1

    mol

    –1]

    200

    100

    0

    –100

    –200

    y = –3.11×10–2x – 14.2R2 = 0.840

    Co2+

    Ni2+

    Mn2+

    CO32–

    Mg2+

    Ca2+

    Li+

    Sr2+

    HCO3–

    Ba2+

    SO42–

    Na+

    Cs+

    Rb+K+Cl–

    NO3–

    Br–I–

    ClO4–

    –4000 –2000 0 2000 4000 6000

    ΔMAL–H

    Vis

    cosi

    ty B

    [L m

    ol–1

    ]

    0.5

    0.4

    0.3

    0.2

    –0.1

    0

    0.1

    Co2+Ni2+

    Mn2+Mg2+

    CO32–

    Ca2+

    Li+

    HCO3–

    Sr2+

    Ba2+SO42–

    Rb+ Cs+

    Na+

    K+

    Cl–

    NO3–

    Br–ClO4

    I–

    (c) vs. Viscosity B

    y = 6.82×10–5x – 0.0999R2 = 0.801

    –4000 –2000 0 2000 4000 6000

    ΔMAL–H

    BN

    MR

    0

    0.6

    0.4

    –0.2

    0.2

    Mg2+

    CO32–

    Ca2+

    Li+

    Sr2+

    Ba2+

    HCO3–

    SO42–

    Cs+Rb+

    Na+

    K+Cl–

    NO3–

    Br–

    I–ClO4

    –y = 7.68×10–5x – 0.0849

    R2 = 0.724

    (d) vs. BNMR

    Fig. 5 Correlations between the DMAL–H (difference in area between the lower (2,600–3,420 cm-1) and

    the higher (3,420–3,800 cm-1) frequency regions of the bands portrayed in Fig. 4 and: a GHB, b structuralentropy (Sstr), c viscosity B-coefficient (Bg), and d the ionic B-coefficient of NMR relaxation (BNMR)

    J Solution Chem (2014) 43:1055–1077 1069

    123

  • ions. DMAL–H can thereby be a useful means to evaluate the ion–water interaction in anaqueous solution.

    It is noteworthy that although the effects of all ions were calculated using the effect

    of SO2�4 as a benchmark, SO2�4 shows non-zero values in the parameters correlated with

    DMAL–H (Fig. 5) except for Sstr. In fact, Bg and BNMR show positive values for SO2�4 ,

    whereas DGHB shows a negative one. These results suggest that ions located nearer theorigin of each figure (e.g., K?) are more appropriate as a base to calculate the effects of

    ions. However, the deviations of the values for SO2�4 from zero are not marked

    compared with the scatter of the plots in each figure. Consequently, the switch of SO2�4to another ion (e.g., K?) only slightly influences the general conclusions obtained in

    this study.

    It is also noteworthy that ions caused similar modifications of the pATR spectra of

    water (Fig. 6a, b) and the molar absorptivity spectra (Fig. 4a, b). The similarities are

    observed because the modification of the pATR spectrum of water is proportional to

    the modification of the k(m) spectrum [25, 27, 44]. Moreover, the difference in areabetween the lower (2,600–3,300 cm-1) and the higher (3,300–3,800 cm-1) frequency

    regions in the pATR spectra (DATRL–H; Table 3) show similar linear correlations withthe four parameters (DGHB, Sstr, Bg and BNMR) to those observed in Fig. 5 (only thecorrelation between DATRL–H and DGHB is depicted in Fig. 6c). Ions having positiveDATRL–H values showe positive DGHB values, while ions having negative DATRL–Hshow negative DGHB. ATR-IR spectroscopy is a readily accessible and cost-effectivetechnique to study aqueous solutions. Moreover, it does not require highly skilled

    operators. ATR-IR spectroscopy can therefore offer a rapid and simple means to

    estimate structural and dynamic properties of the ion–water interaction in aqueous

    solution.

    Wavenumber [cm–1]2600300034003800

    Wavenumber [cm–1]

    26003000340038000.5

    0.9

    1.1

    1.5

    –0.1

    0.7

    1.3

    0.3

    0.1

    0.5

    0.9

    1.1

    –0.1

    0.7

    1.3

    0.3

    0.1

    Mol

    ar a

    bsor

    ptiv

    ity [m

    ol–1

    ·cm

    –1]

    Mol

    ar a

    bsor

    ptiv

    ity [m

    ol–1

    ·cm

    –1]

    Mg2+

    Ca2+

    Sr2+

    Ba2+

    Li+

    Na+

    K+

    Rb+

    Cs+

    Mn2+

    Ni2+

    Co2+ NO3–

    ClO4–

    HCO3–

    CO32–

    Cl–

    I–

    Br–

    (a) (b)

    –25 25 50 100–50 0 75

    ΔATRL–H

    ΔGH

    B

    –1.2

    –0.8

    –0.4

    0

    0.4

    0.8

    1.2

    Ni2+

    Co2+

    Mn2+

    CO32–

    Mg2+

    Ca2+Sr2+

    Li+

    HCO3–

    SO42–

    Na+ Ba2+

    Cs+Rb+K+Cl

    NO3–

    Br–ClO4

    I–

    y = 1.23×10–2x – 0.298R2 = 0.823

    (c)

    Fig. 6 a, b Effects of cations (a) and anions (b) on the ATR-IR spectrum of water in the spectral range of2,600–3,800 cm-1. c Correlation between the DATRL–H (difference in area between the lower(2,600–3,300 cm-1) and the higher (3,300–3,800 cm-1) frequency regions of the bands observed ina and b and DGHB

    1070 J Solution Chem (2014) 43:1055–1077

    123

  • 4 Conclusion

    The effects of various cations (Li?, Na?, K?, Rb?, Cs?, Mg2?, Ca2?, Sr2?, Ba2?, Mn2?,

    Co2?, and Ni2?) and anions (Cl-, Br-, I-, NO�3 , ClO�4 , HCO

    �3 , and CO

    2�3 ) on the molar

    absorptivity of water, as a function of ion concentration up to 2 mol�L-1, were determined(1) by measuring the ATR-IR spectra of many electrolyte aqueous solutions (22 in all) (2)

    by converting the ATR-IR spectra into the molar absorptivity spectra, and (3) by separating

    the effects of individual cations and anions using the effect of SO2�4 as a benchmark.Spectral analyses produced the following conclusions:

    (1) Most OH stretching bands in the molar absorptivity spectra show linear intensity

    increases and decreases with ion concentration, showing additivity of each effect of

    these cations and anions.

    (2) For certain specific combinations of cations and anions (Cs2SO4, Li2SO4, and

    MgSO4), the band intensities depart downward from the linear trend at higher

    concentrations. That type of deviation was attributed to the formation of ion pairs and

    to cooperativity in ion hydration, which indicates that the extent of the ion–water

    interaction as reflected by the OH stretching band of water is beyond the first

    solvation shell of water molecules directly surrounding the ion.

    (3) The difference in areas (DMAL–H) between the lower (2,600–3,420 cm-1) and the

    higher (3,420–3,800 cm-1) frequency regions showed quasi-linear relations with

    several quantitative parameters representing structural and dynamic properties of

    water molecules around ions: DGHB, Sstr, Bg and BNMR. This observation indicatesthat the modification of the OH stretching band of water caused by ions is useful as an

    indicator to evaluate ion–water interactions in an aqueous solution.

    Acknowledgments We greatly appreciate Dr. Tadashi Yokoyama and Mr. Naoki Nishiyama of OsakaUniversity for their help with sample preparations. We also thank two anonymous referees and associatededitor Luigi Paduano for their careful reviews of this manuscript. This research was financially supported bya JSPS Research Fellowship for Young Scientists to Norio Kitadai.

    Appendix

    Although it is not the main subject of this study, we show the effects of 1 mol�L-1 cationsand 1 mol�L-1 anions on the molar absorptivity of water in the OH bending band region(1,200–2,000 cm-1) in Fig. 7. This presentation of the influences of various ions on the

    OH bending band of water is the first ever reported. We hope that these results can

    stimulate future theoretical and/or experimental studies in this area.

    We also present in Tables 2 and 3 a summary of the aqueous solutions measured in this

    study, and the numerical results on the areas of DMAL–H and DATRL–H for each ion (at1 mol�L-1 concentration), respectively.

    J Solution Chem (2014) 43:1055–1077 1071

    123

  • Mol

    ar a

    bsor

    ptiv

    ity [m

    ol–1

    ·cm

    –1]

    Mol

    ar a

    bsor

    ptiv

    ity [m

    ol–1

    ·cm

    –1]

    Mg2+

    Ca2+

    Sr2+

    Ba2+

    Li+

    Na+

    K+

    Rb+

    Cs+

    Mn2+

    Ni2+

    Co2+

    Wavenumber [cm–1]

    120016002000 1800 1400

    3

    7

    11

    15

    19

    –1

    10

    15

    20

    25

    5

    1

    3

    5

    7

    9

    –1

    11 CO32–

    Cl–

    Br–

    I–

    ClO4–

    NO3–

    10

    15

    20

    25

    5

    Wavenumber [cm–1]

    120016002000 1800 1400

    (a) (b)

    Fig. 7 Effects of 1 mol�L-1 cations (a) and 1 mol�L-1 anions (b) on the molar absorptivity of water in theOH bending band region (1,200–2,000 cm-1). The effect of HCO�3 is not shown in this figure because of thestrong overlap with a HCO�3 band [1]. The OH bending band of water is shown at the top of each figure forcomparison

    Table 2 Summary of the aqueous solutions measured in this study

    Electrolyte Solid Purity Conc.(mol�L-1)

    Density(g�mL-1)

    Refractiveindex (nD)

    Manufacturer

    LiCl LiCl�H2O [99.9 0.1 1.000 1.334 Wakoa

    0.25 1.004 1.335

    0.5 1.008 1.338

    0.75 1.016 1.340

    1 1.021 1.342

    1.5 1.033 1.346

    2 1.044 1.350

    Li2SO4 Li2SO4�H2O [99.0 0.05 1.002 1.334 Wakoa

    0.125 1.009 1.335

    0.25 1.019 1.338

    0.375 1.030 1.340

    0.5 1.041 1.342

    1072 J Solution Chem (2014) 43:1055–1077

    123

  • Table 2 continued

    Electrolyte Solid Purity Conc.(mol�L-1)

    Density(g�mL-1)

    Refractiveindex (nD)

    Manufacturer

    0.75 1.064 1.346

    1 1.087 1.350

    NaCl NaCl [99.5 0.1 1.001 1.334 Wakoa

    0.5 1.017 1.338

    1 1.037 1.343

    1.5 1.057 1.348

    2 1.077 1.352

    NaBr NaBr [99.9 0.1 1.004 1.335 Wakoa

    0.5 1.037 1.340

    1 1.075 1.347

    1.5 1.115 1.354

    2 1.152 1.360

    NaI NaI [99.5 0.1 1.009 1.335 Wakoa

    0.5 1.054 1.344

    1 1.113 1.354

    1.5 1.171 1.365

    2 1.226 1.375

    NaNO3 NaNO3 [99.0 0.1 1.002 1.334 Wakoa

    0.5 1.026 1.338

    1 1.052 1.342

    1.5 1.079 1.346

    2 1.107 1.350

    NaClO4 NaClO4�H2O 99 0.1 1.005 1.333 Stremb

    0.5 1.037 1.336

    1 1.076 1.340

    1.5 1.113 1.344

    2 1.154 1.347

    Na2SO4 Na2SO4 [99.0 0.05 1.004 1.334 Wakoa

    0.25 1.028 1.338

    0.5 1.059 1.343

    0.75 1.088 1.348

    1 1.119 1.352

    KCl KCl [99.5 0.1 1.002 1.334 Wakoa

    0.5 1.021 1.338

    1 1.044 1.343

    1.5 1.067 1.347

    2 1.090 1.352

    KHCO3 KHCO3 [99.7 0.1 1.004 1.334 Aldrichc

    0.5 1.030 1.338

    1 1.060 1.343

    1.5 1.090 1.348

    2 1.118 1.353

    K2CO3 K2CO3 [99.5 0.1 1.010 1.335 Wakoa

    J Solution Chem (2014) 43:1055–1077 1073

    123

  • Table 2 continued

    Electrolyte Solid Purity Conc.(mol�L-1)

    Density(g�mL-1)

    Refractiveindex (nD)

    Manufacturer

    0.5 1.056 1.344

    1 1.112 1.354

    1.5 1.165 1.363

    2 1.217 1.372

    K2SO4 K2SO4 [99.0 0.05 1.004 1.334 Wakoa

    0.25 1.031 1.338

    0.5 1.065 1.343

    Rb2SO4 Rb2SO4 [99.0 0.05 1.011 1.333 Wakoa

    0.25 1.052 1.339

    0.5 1.105 1.345

    0.75 1.159 1.351

    1 1.210 1.357

    Cs2SO4 Cs2SO4 [99.9 0.05 1.012 1.334 Wakoa

    0.25 1.071 1.340

    0.5 1.147 1.346

    0.75 1.221 1.352

    1 1.296 1.358

    MgCl2 MgCl2�6H2O [99.0 0.1 1.005 1.335 Aldrichc

    0.25 1.017 1.339

    0.5 1.034 1.345

    0.75 1.055 1.350

    1 1.072 1.356

    1.5 1.107 1.366

    2 1.142 1.377

    MgSO4 MgSO4�7H2O [99.5 0.1 1.010 1.335 Wakoa

    0.25 1.027 1.339

    0.5 1.057 1.344

    0.75 1.083 1.350

    1 1.111 1.355

    1.5 1.165 1.364

    2 1.221 1.373

    CaCl2 CaCl2�2H2O [99.0 0.1 1.006 1.336 Wakoa

    0.5 1.042 1.346

    1 1.086 1.358

    1.5 1.129 1.370

    2 1.171 1.381

    SrCl2 SrCl2�6H2O [99.0 0.1 1.010 1.336 Wakoa

    0.5 1.064 1.347

    1 1.131 1.360

    1.5 1.197 1.372

    2 1.260 1.384

    BaCl2 BaCl2�2H2O [99.0 0.1 1.015 1.336 Wakoa

    0.5 1.088 1.348

    1074 J Solution Chem (2014) 43:1055–1077

    123

  • Table 2 continued

    Electrolyte Solid Purity Conc.(mol�L-1)

    Density(g�mL-1)

    Refractiveindex (nD)

    Manufacturer

    1 1.178 1.362

    1.5 1.265 1.376

    MnCl2 MnCl2�4H2O [99.0 0.1 1.007 1.336 Wakoa

    0.5 1.050 1.346

    1 1.103 1.358

    1.5 1.150 1.371

    2 1.199 1.383

    CoCl2 CoCl2�2H2O [99.0 0.1 1.008 1.336 Wakoa

    0.5 1.054 1.347

    1 1.111 1.361

    1.5 1.166 1.373

    2 1.220 1.386

    NiCl2 NiCl2�6H2O [99.9 0.1 1.009 1.336 Wakoa

    0.5 1.057 1.348

    1 1.116 1.363

    1.5 1.171 1.379

    2 1.223 1.395

    a Wako Pure Chemical Industries Ltd.b Strem Chemicals Inc.c Sigma–Aldrich Chemie GmbH

    Table 3 Left side: areas in the lower (2,600–3,420 cm-1) and the higher (3,420–3,800 cm-1) frequencyregions of the bands portrayed in Fig. 4 (MAL and MAH, respectively) and their difference (DMAL–H)calculated after drawing a linear baseline from 2,600 to 3,800 cm-1

    MAL MAH DMAL-H ATRL ATRH DATRL-H

    Li? 635 5 630 15.0 2.4 12.6

    Na? -232 92 -324 1.6 5.2 -3.7

    K? -233 102 -335 -2.1 3.1 -5.1

    Rb? -225 30 -255 -2.3 1.9 -4.2

    Cs? 6 22 -15 0.1 0.9 -0.9

    Mg2? 2,976 -545 3,522 72.3 2.0 70.3

    Ca2? 1,077 -164 1,241 39.8 11.0 28.8

    Sr2? 347 -30 377 27.7 14.3 13.4

    Ba2? 66 -108 174 23.8 14.0 9.8

    Mn2? 3,440 -742 4,183 85.2 -1.8 87.0

    Co2? 3,596 -797 4,392 93.9 0.1 93.8

    Ni2? 3,398 -914 4,312 94.7 1.1 93.6

    Cl- -871 353 -1,224 -19.6 5.4 -25.0

    Br- -1,289 515 -1,804 -24.4 8.2 -32.7

    I- -1,787 618 -2,405 -28.3 11.2 -39.6

    NO�3 -1,330 440 -1,770 -30.1 1.3 -31.4

    J Solution Chem (2014) 43:1055–1077 1075

    123

  • References

    1. Masuda, K., Haramaki, T., Nakashima, S., Habert, B., Martinez, I., Kashiwabara, S.: Structural changeof water with solutes and temperature up to 100 �C in aqueous solutions as revealed by attenuated totalreflectance infrared spectroscopy. Appl. Spectrosc. 57, 274–281 (2003)

    2. Li, R.H., Jiang, Z.P., Shi, S.Q., Yang, H.W.: Raman spectra and O-17 NMR study effects of CaCl2 andMgCl2 on water structure. J. Mol. Struct. 645, 69–75 (2003)

    3. Li, R.H., Jiang, Z.P., Chen, F.G., Yang, H.W., Guan, Y.T.: Hydrogen bonded structure of water andaqueous solutions of sodium halides: a Raman spectroscopic study. J. Mol. Struct. 707, 83–88 (2004)

    4. Li, R.H., Jiang, Z.P., Guan, Y.T., Yang, H.W., Liu, B.: Effects of metal ion on the water structurestudied by the Raman O–H stretching spectrum. J. Raman Spectrosc. 40, 1200–1204 (2009)

    5. Chen, Y., Zhang, Y.H., Zhao, L.J.: ATR-FTIR spectroscopic studies on aqueous LiClO4, NaClO4, andMg(ClO4)2 solutions. Phys. Chem. Chem. Phys. 6, 537–542 (2004)

    6. Wei, Z.F., Zhang, Y.H., Zhao, L.J., Liu, J.H., Li, X.H.: Observation of the first hydration layer ofisolated cations and anions through the FTIR-ATR difference spectra. J. Phys. Chem. A 109, 1337–1342(2005)

    7. Liu, J.H., Zhang, Y.H., Wang, L.Y., Wei, Z.F.: Drawing out the structural information of the first layerof hydrated ions: ATR-FTIR spectroscopic studies on aqueous NH4NO3, NaNO3, and Mg(NO3)2solutions. Spectrochim. Acta Part A 61, 893–899 (2005)

    8. Zhao, L.J., Zhang, Y.H., Wei, Z.F., Cheng, H., Li, X.H.: Magnesium sulfate aerosols studied by FTIRspectroscopy: hygroscopic properties, supersaturated structures, and implications for seawater aerosols.J. Phys. Chem. A 110, 951–958 (2006)

    9. Dong, J.L., Li, X.H., Zhao, L.J., Xiao, H.S., Wang, F., Guo, X., Zhang, Y.H.: Raman observation of the

    interactions between NHþ4 , SO2�4 , and H2O in supersaturated (NH4)2SO4 droplets. J. Phys Chem. B 111,

    12170–12176 (2007)10. Kataoka, Y., Kitadai, N., Hisatomi, O., Nakashima, S.: Nature of hydrogen bonding of water molecules

    in aqueous solutions of glycerol by attenuated total reflection (ATR) infrared spectroscopy. Appl.Spectrosc. 65, 436–441 (2011)

    11. Sun, Q.: Raman spectroscopic study of the effects of dissolved NaCl on water structure. Vib. Spectrosc.62, 110–114 (2012)

    12. Guo, Y.C., Li, X.H., Zhao, L.J., Zhang, Y.H.: Drawing out the structural information about the firsthydration layer of the isolated Cl- anion through the FTIR-ATR difference spectra. J. Solution Chem.42, 459–469 (2013)

    13. Walrafen, G.E., Hokmabadi, M.S., Yang, W.H.: Raman isosbestic points from liquid water. J. Chem.Phys. 85, 6964–6969 (1986)

    14. Terpstra, P., Combes, D., Zwick, A.: Effect of salts on dynamics of water—a Raman-spectroscopystudy. J. Chem. Phys. 92, 65–70 (1990)

    15. Pastorczak, M., Kozanecki, M., Ulanski, J.: Raman resonance effect in liquid water. J. Phys. Chem. A112, 10705–10707 (2008)

    16. Hancer, M., Sperline, R.P., Miller, J.D.: Anomalous dispersion effects in the IR-ATR spectroscopy ofwater. Appl. Spectrosc. 54, 138–143 (2000)

    Table 3 continued

    MAL MAH DMAL-H ATRL ATRH DATRL-H

    ClO�4 -2,092 492 -2,584 -46.4 -3.1 -43.3

    HCO�3 884 111 773 12.5 -4.4 16.9

    CO2�3 2,914 -914 3,828 74.6 -16.2 90.8

    SO2�4 0 0 0 0 0 0

    The errors in the DMAL–H for monovalent ions are within ±30 and those for divalent ions are within±60 cm-1, which values were estimated from several independent measurements. Right side: areas in thelower (2,600–3,300 cm-1) and the higher (3,300–3,800 cm-1) frequency regions of the bands portrayed inFig. 6a, b (ATRL and ATRH, respectively) and their difference (DATRL–H) calculated after drawing a linearbaseline from 2,600 to 3,800 cm-1. The maximum error ranges in the DATRL–H values for monovalent ionsare ±0.45 and that for divalent ions are ±0.9 cm-1. All values for SO2�4 were set to zero

    1076 J Solution Chem (2014) 43:1055–1077

    123

  • 17. Walrafen, G.E.: Raman spectral studies of effects of solutions and pressure on water structure. J. Chem.Phys. 55, 768–792 (1971)

    18. Bertie, J.E., Eysel, H.H.: Infrared intensities of liquids. 1. Determination of infrared optical anddielectric-constants by FT-IR using the circle ATR cell. Appl. Spectrosc. 39, 392–401 (1985)

    19. Weast, C.: Handbook of Chemistry and Physics, 57th edn. CRC Press, Cleveland (1977)20. Li, H.H.: Reflective-index of ZnS, ZnSe, and ZnTe and its wavelength and temperature derivatives.

    J. Phys. Chem. Ref. Data 13, 103–150 (1984)21. Bertie, J.E., Lan, Z.: Infrared intensities of liquids. 20. The intensity of the OH stretching band of liquid

    water revisited, and the best current values of the optical constants of H2O(l) at 25 �C between 15,000and 1 cm-1. Appl. Spectrosc. 50, 1047–1057 (1996)

    22. Keefe, C.D., Wilcox, T., Campbell, E.: Measurement and applications of absolute infrared intensities.J. Mol. Struct. 1009, 111–122 (2012)

    23. Ayerst, R.P., Phillips, M.I.: Solubility and refractive index of ammonium perchlorate in water. J. Chem.Eng. Data 11, 494–496 (1966)

    24. Yu, X., Zeng, Y., Yao, H., Yang, J.: Metastable phase equilibria in the aqueous ternary systemsKCl ? MgCl2 ? H2O and KCl ? RbCl ? H2O at 298.15 K. J. Chem. Eng. Data 56, 3384–3391(2011)

    25. Max, J.J., Chapados, C.: Infrared spectra of cesium chloride aqueous solutions. J. Chem. Phys. 113,6803–6814 (2000)

    26. Max, J.J., Chapados, C.: IR spectroscopy of aqueous alkali halide solutions: pure salt-solvated waterspectra and hydration numbers. J. Chem. Phys. 115, 2664–2675 (2001)

    27. Max, J.J., de Blois, S., Veilleux, A., Chapados, C.: IR spectroscopy of aqueous alkali halides. Factoranalysis. Can. J. Chem. 79, 13–21 (2001)

    28. Urrejola, S., Sanchez, A., Hervello, M.F.: Refractive indices of lithium, magnesium, and copper(II)sulfates in ethanol–water solutions. J. Chem. Eng. Data 55, 482–487 (2010)

    29. Urrejola, S., Sanchez, A., Hervello, M.F.: Refractive indices of sodium, potassium, and ammoniumsulfates in ethanol water solutions. J. Chem. Eng. Data 55, 2924–2929 (2010)

    30. Seidell, A.: Solubilities of Inorganic and Organic Compounds: A Compilation of Quantitative SolubilityData from the Periodical Literature, 2nd edn. Nostrand, New York (1919)

    31. Rudolph, W.W., Irmer, G., Hefter, G.T.: Raman spectroscopic investigation of speciation inMgSO4(aq). Phys. Chem. Chem. Phys. 5, 5253–5261 (2003)

    32. Buchner, R., Chen, T., Hefter, G.: Complexity in ‘‘simple’’ electrolyte solutions: ion pairing inMgSO4(aq). J. Phys. Chem. B 108, 2365–2375 (2004)

    33. Hefter, G.: When spectroscopy fails: the measurement of ion pairing. Pure Appl. Chem. 78, 1571–1586(2006)

    34. Larentzos, J.P., Criscenti, L.J.: A molecular dynamics study of alkaline earth metal-chloride com-plexation in aqueous solution. J. Phys. Chem. B 112, 14243–14250 (2008)

    35. Callahan, K.M., Casillas-Ituarte, N.N., Roeselova, M., Allen, H.C., Tobias, D.J.: Solvation of magne-sium dication: molecular dynamics simulation and vibrational spectroscopic study of magnesiumchloride in aqueous solutions. J. Phys. Chem. A 114, 5141–5148 (2010)

    36. Wachter, W., Fernandez, S., Buchner, R., Hefter, G.: Ion association and hydration in aqueous solutionsof LiCl and Li2SO4 by dielectric spectroscopy. J. Phys. Chem. B 111, 9010–9017 (2007)

    37. Tielrooij, K.J., Garcia-Araez, N., Boon, M., Bakker, H.J.: Cooperativity in ion hydration. Science 328,1006–1009 (2010)

    38. Smith, J.D., Cappa, C.D., Wilson, K.R., Cohen, R.C., Geissler, P.L., Saykally, R.J.: Unified descriptionof temperature-dependent hydrogen-bond rearrangements in liquid water. Proc. Natl. Acad. Sci. USA102, 14171–14174 (2005)

    39. Eaves, J.D., Loparo, J.J., Fecko, C.J., Roberts, S.T., Tokmakoff, T., Geissler, P.L.: Hydrogen bonds inliquid water are broken only fleetingly. Proc. Natl. Acad. Sci. USA 102, 13019–13022 (2005)

    40. Bakker, H.J., Skinner, J.L.: Vibrational spectroscopy as a probe of structure and dynamics of liquidwater. Chem. Rev. 110, 1498–1517 (2010)

    41. Marcus, Y.: Effect of ions on the structure of water: structure making and breaking. Chem. Rev. 109,1346–1370 (2009)

    42. Marcus, Y.: Viscosity B-coefficients, structural entropies and heat-capacities, and the effects of ions onthe structure of water. J. Solution Chem. 23, 831–848 (1994)

    43. Engel, G., Hertz, H.G.: On negative hydration. A nuclear magnetic relaxation study. Ber. Bunsenges.Phys. Chem. 72, 808–834 (1968)

    44. Max, J.J., Chapados, C.: Influence of anomalous dispersion on the ATR spectra of aqueous solutions.Appl. Spectrosc. 53, 1045–1053 (1999)

    J Solution Chem (2014) 43:1055–1077 1077

    123

    Effects of Ions on the OH Stretching Band of Water as Revealed by ATR-IR SpectroscopyAbstractIntroductionExperimentalMaterialsATR-IR MeasurementsExtraction of the Molar Absorptivity of Water from the ATR-IR Spectra of Aqueous Electrolyte SolutionsExtraction of the Effects of Individual Ions on the OH Stretching Band of Water

    Results and DiscussionEffects of Ions on the OH Stretching Band of WaterCorrelations Between the OH Stretching Band of Water and Structural and Dynamic Properties of the Ion--Water Interaction in Aqueous Solution

    ConclusionAcknowledgmentsAppendixReferences