Article pubs.acs.org/JPCB
Coextraction of Water into Nitrobenzene with Organic Ions Yasuhiro Naito,† Wataru Murakami,† Kazuo Eda,† Masahiro Yamamoto,‡ and Toshiyuki Osakai*,† †
Department of Chemistry, Graduate School of Science, Kobe University, Nada, Kobe 657-8501, Japan Department of Chemistry of Functional Molecules, Faculty of Science and Engineering, Konan University, Higashinada, Kobe 658-8501, Japan
‡
S Supporting Information *
ABSTRACT: Various organic anions (sulfonates (RSO3−), carboxylates (RCO2−), and phenolates (RO−)) and ammonium cations (RNH3+, R2NH2+, and R3NH+) were distributed in the nitrobenzene (NB)−water system by using Crystal Violet and dipicrylaminate, respectively. The number of water molecules (n) being coextracted into NB with an ion was then determined by the Karl Fischer method. The n values determined and those reported previously showed the variation from 0.51 to 3.4, depending on not only the charged groups but also the noncharged R-groups. In this study, we focused our attention to the strong electric field on the charged group and its facilitation effect for binding water molecules in NB. The local electric field (Ei) on the surface of an organic ion was evaluated by using Gaussian09 program with a subprogram developed in our recent study. It was found that the n values showed a clear dependence on the average value of Ei on oxygen or hydrogen atoms, respectively, of an anionic or cationic group.
S
interface to those in bulk water. The water molecules, being strongly bound to the metal ions in W, appear to be undissociated even in the transfer across the O/W interface. In this study, we focused our attention to organic ions with charged groups and their ability to coextract water molecules to organic solvent. Previously, our group used a wellestablished and conventional technique, i.e., the Karl Fischer method to determine the n values for carboxylates18 (RCO2−) and also for primary, secondary, and tertiary ammonium ions19 (RNH3+, R2NH2+, and R3NH+) in the NB−W system. In each case, the n values are not significantly dependent on the chemical structures of noncharged groups (i.e., R) of the ions, and their respective charged groups have characteristic values of n (=2.4, 1.64, 1.04, and 0.66, respectively, for RCO2−, RNH3+, R2NH2+, and R3NH+). However, it is not clear what factors actually determine the n values for such charged groups. In this study, we have determined n values for a variety of organic ions, including sulfonates, carboxylates, phenolates, and ammonium ions. It has been found that the electric field strength on a charged group can be used as a useful index to evaluate the n value for organic ions.
ince about a half century ago, much attention has been paid to the coextraction of water (W) to water-immiscible organic solvents (O) with hydrophilic ions including Li+, Na+, Ca2+, and Cl−.1−9 Even water-immiscible solvents such as nitrobenzene (NB) dissolve a considerable amount of water (e.g., 0.168 M H2O in NB7); therefore, such phenomena can be elucidated in terms of selective (or preferential) hydration of ions in mixed solvents.10 This is a key concept intimately related to important ion-transfer processes in separation/ detection systems, including solvent extraction, membrane transport, emulsion, and ion-selective electrodes. Also, it has a fundamental significance for understanding the role of water in biological systems, including lipid membranes and proteins (enzymes, ion channels, etc.). In previous theories11−14 on the Gibbs energy of ion transfer at the O/W interface (ΔG°tr,W→O), it is assumed that hydrophilic ions are transferred into the O phase with some of hydrated water molecules. In the so-called non-Bornian analyses,11−13 ΔGtr°,W→O values at the NB/W interface have successfully been evaluated for hydrophilic inorganic ions, by considering the number (n) of water molecules coextracted into NB with an ion. Furthermore, molecular dynamic simulations have also confirmed the cotransport of water with hydrophilic ions at the O/W interface.15−17 Regarding spherical inorganic ions, it is generally known that the more hydrophilic the ion is, the larger its n value. An excellent correlation was seen between the Stokes radius and the hydrated radius (estimated from the n value) for alkali and alkaline earth metal ions.7 This suggests the similarity of transfer processes of the hydrophilic ions at the O/W © 2015 American Chemical Society
■
EXPERIMENTAL SECTION Materials. Eleven sulfonates (RSO3− with R = (1) Ph, (2) p-(CH3)Ph, (3) o,o,p-(CH3)3Ph, (4) m-(NO2)Ph, (5) o,pReceived: February 8, 2015 Revised: April 22, 2015 Published: April 22, 2015 6010
DOI: 10.1021/acs.jpcb.5b01316 J. Phys. Chem. B 2015, 119, 6010−6017
Article
The Journal of Physical Chemistry B (NO2)2Ph, (6) o,o,p-(NO2)3Ph, (7) 2-Naphthyl, (8) n(C3H7), (9) n-(C6H13), (10) Br(CH2)3, and (11) cyc(C6H11)), four carboxylates (RCO2− with R = (12) Ph, (16) m-(NO2)Ph, (20) n-(C5H11), and (21) n-(C7H15)), and three phenolates (RO− with R = (23) o-(NO2)Ph, (24) o,o,p(NO2)3Ph, and (25) o,o,p-Cl3Ph) were purchased as sodium salts from Aldrich or Tokyo Chemical Industries (TCI). In addition, seven RCO2− ions with R = (13) o-(CH3)Ph, (14) m-(CH3)Ph, (15) p-(CH3)Ph, (17) o,p-(NO2)2PhCH2, (18) o,o,p-F3Ph, (19) Ph2CH, and (22) cyc-(C6H11)CH2 and one RO− ion with R = (26) o,o,p-Br3PhO− were purchased as the acid forms (i.e., RCO2H or ROH) from Aldrich or TCI. Four ammonium ions, (27) cyc-(C6H11)CH2NH3+, (28) N,NMeEtNH2+, (29) N,N,N-Me2EtNH+, and (30) N,N,NMeEt2NH+, were also purchased as the deprotonated amine forms (e.g., cyc-(C6H11)CH2NH2 or N,N-MeEtNH) from Aldrich or TCI. All the above reagents were used as received. See also the Supporting Information for the other reagents, which include Crystal Violet (tris(4-dimetylaminophenyl)methyl chloride), NaDPA (dipicrylamine sodium salt), and bis(triphenylphosphoranylidene)ammonium chloride (BTPPACl) and tetrakis(4-chlorophenyl)borate 2 0 (BTPPATClPB). Distribution Experiments. In a similar manner as described previously,7,18,19 the above-listed organic anions and cations were distributed between W and NB by using Crystal Violet cation (CV+) and DPA−, respectively, as the counterion. Ten mL of the W and NB phases were put in a glass bottle with a Teflon cap, shaken rigorously, and then allowed to stand overnight in a water bath thermostated at 25.0 ± 0.1 °C. An aliquot (500 μL) of the NB phase in the bottle was slowly drawn off using a microsyringe and then subjected to the determination of water concentration by using a Karl Fischer coulometer (Kishida Chemical CA-20). In the distribution experiments of sulfonates (1−11), carboxylates (12, 16, 20, 21), and phenolates (23−25), each sodium salt was added to the W phase at the initial concentration of x0 = 5−20 mM (M = mol dm−3), while Crystal Violet (CV+Cl−) was added to the NB phase at the same concentration. For seven carboxylates (13−15, 17−19, 22) and one phenolate (26), the sodium salts were not commercially available, so their acid forms were dissolved in the W phase at 5−20 mM by equimolar addition of sodium hydroxide. After confirming that the pH jumped to 7 or above (i.e., complete dissociation of the acids), similar distribution experiments as above were performed. Because CV+ is very hydrophobic, it can extract a significant portion of RSO3−, RCO2−, and RO− ions to NB by releasing Cl− into W. The equilibrium concentration of Cl− in the W phase, [Cl−]W, was determined by potentiometric titration with a standard silver nitrate solution. Spectrophotometric determination of CV+ in the W phase18 (with ε = 9.745 × 104 M−1 cm−1 at 589 nm) showed that practically 100% of CV+ remained in the NB phase. In accordance, the equilibrium concentration of an organic anion in NB, [A−]NB, was obtained by equating to [Cl−]W. The equilibrium concentration of Cl− in NB was determined from the difference between [Cl−]W and the added concentration of CV+Cl− to the NB phase (i.e., x0). In the distribution experiments of ammonium ions (27− 30) with DPA−, each deprotonated amine (denoted here by B; oily) was dissolved in W by equimolar addition of hydrochloric acid, so that the concentration became y0 = 5− 25 mM. After confirming that the pH dropped to 7 or below
(i.e., complete protonation of the amine), NaDPA powder was added to the W phase at the same concentration as the ammonium ion (i.e., y0). The resultant reddish suspension was subjected to the distribution experiment with NB as described above. Because DPA− is very hydrophobic, it can efficiently extract the amines to the NB phase as the protonated forms (i.e., ammonium ions, BH+). The equilibrium concentration of DPA− that remained in the W phase, [DPA−]W, was determined spectrophotometrically7 (ε = 2.61 × 104 M−1 cm−1 at 428 nm). Then, the concentration of BH+ extracted to NB, being equal to that of DPA−, was determined as [BH+]NB = y0 − [DPA−]W. Under the present experimental conditions, the extraction of Na+ with DPA− to NB can be regarded as negligible, based on a theoretical consideration described below. Ion-Transfer Voltammetry. For theoretical consideration of the distribution behaviors of organic ions, we required their standard ion-transfer potentials (ΔOWϕ°) at the NB/W interface. Since the ΔW O ϕ° values were not reported for about half of the ions studied, we have employed ion-transfer voltammetry to determine the Δ OW ϕ° values. In the voltammetric measurements, the transfer of these ions at a micro O/W interface21−23 was observed with two-electrode electrochemical cells:
The double bar in each cell represents the NB/W interface to be tested. The Galvani potential difference across the NB/ W W interface (ΔW − ϕO with ϕW and ϕO being the Oϕ = ϕ internal potential of the W and O phases) was controlled by means of a potentiostat (Hokuto Denko Co., HA1010mM1A) with a laboratory-constructed, computer-controlled system.24 The potential E applied to cell A or B is related to the ΔW Oϕ of the test NB/W interface as E = ΔW ϕ + ΔE , where ΔE O ref ref is a constant determined only by the reference electrode system used. In this study, the ΔEref value was not determined, because its reproducibility was not very satisfactory (2σ ≈ 18 mV). Alternatively, the ΔW O ϕ° values of ions were determined by using the transfer of a reference ion, i.e., ClO4− (for anions; added as the potassium salt) or Me4N+ (for cations; added as the chloride salt); the ΔW O ϕ° values at the NB/W interface were reported to be −0.082 V for ClO4− and −0.035 V for Me4N+.11 All voltammetric measurements were done at room temperature (25 ± 3 °C). For the other details, see the Supporting Information. Quantum Chemical Calculation. The Gaussian09 program package25,26 with the B3LYP hybrid density functional theory27−30 was used to optimize the groundstate geometry of organic ions in vacuum. Though a variety of quantum chemical studies have been performed on the solvent effect on ion solvation,31−34 we carried out quantum chemical calculations for ions isolated in vacuum. This is for 6011
DOI: 10.1021/acs.jpcb.5b01316 J. Phys. Chem. B 2015, 119, 6010−6017
Article
The Journal of Physical Chemistry B figuring out their n values from intrinsic properties of ions (i.e., their size, shape, charge distribution, etc.). At the B3LYP/6-311++G(2d,p) level, three different quantum chemical charges were computed, i.e., Mulliken,35 Merz− Kollman (MK),36,37 and natural population analysis (NPA)38 charges. Using a previously developed program,39 the distribution of the surface field strength (Ei with i being a finite positive integer) on an ion was then calculated from the atomic coordinate data and partial atomic charges for the optimized structure of the ion. The program was written in Microsoft Visual Basic 6.0, and the source code has been released as the Supporting Information in ref 39. In the program, Ei can be obtained as the local values at the van der Waals (vdW) surface or the solvent-accessible surface40,41 (SAS) being apart from the vdW surface by 0.14 nm (=the radius of water molecule). Then the averaged value of Ei on a charged group
values for cations or anions, by referring to the peak potential of a reference ion, i.e., Me4N+ or ClO4−, respectively. In this determination, we simply assumed that the difference in the peak potential corresponds to the difference in ΔW O ϕ°. This assumption is valid for the present system, because the ion to be tested and the reference ion have the same charge number (+1 or −1) and also the transfer of these ions across the NB/ W interface is fast enough to be diffusion-limited. Thus, we could determine the ΔW O ϕ° values for several ions, which are shown in Table 1. As also shown in the table, some relatively hydrophilic anions (e.g., 8, 13, 23) gave no clear transfer peak in the potential window; their ΔW O ϕ° values have been evaluated as < −0.23 V. Distribution Behaviors of Ions. In organic solvents such as NB having a relatively high permittivity, the ion association is generally of less significance. Previous conductometric measurements19 showed that primary to tertiary ammonium ions formed no significant ion pairs with DPA− in watersaturated NB; the association constants were less than 3.7 M−1. As described previously,43−46 the distribution equilibrium in such a system can be described in terms of the standard ion-transfer potential (ΔW O ϕ° i ) for each ion (i) participating in the distribution equilibrium. Here, ΔW O ϕi° is given by ΔW °,W→O/ziF (F, the Faraday constant), O ϕi° = ΔGtr,i where ΔG°tr,i,W→O and zi are the Gibbs energy of the transfer from W to O and the charge number (including the sign) of ion i, respectively. We can use ΔW °,W→O) as O ϕi° (and thus ΔGtr,i a useful measure of extractability of individual ions. When j kinds of ions are distributed in an O−W system, the equilibrium potential difference (ΔW O ϕeq) across the O/W interface is given by43−46
was obtained.
■
RESULTS AND DISCUSSION Ion-Transfer Voltammetry. In this study, the distribution behaviors of organic ions have been elucidated in terms of their standard ion-transfer potentials (ΔW O ϕ°). Although the ΔW ϕ° values have been reported for some ions (e.g., 1, 2, 12, O 24, etc.),11,42 those for the other ions have not been reported. For the unreported ions, we determined their ΔW O ϕ° values by means of ion-transfer voltammetry. Figure 1 shows
j
∑ i=1
zici0,W
ziF 1 + exp⎡⎣ RT (ΔOW ϕeq − ΔOW ϕi◦)⎤⎦ j
+
∑ i=1
c0,W i
zici0,O
ziF 1 + exp⎡⎣ RT (ΔOW ϕeq − ΔOW ϕi◦)⎤⎦
=0 (1)
c0,O i
where and are the initial concentrations of ion i in W and O phases, and R and T have their usual meanings. In the derivation of eq 1, it is assumed that the volume ratio of the two phases is unity. If the values of ΔW O ϕi° are known for all ions in the system of interest, the value of ΔW O ϕeq can be obtained numerically using eq 1. Then, the distribution ratio of each ion can be calculated from the Nernst equation:
Figure 1. Cyclic voltammograms for the transfer of 1.0 mM (a) Me4N+ (as an internal reference) at the micro NB/W interface (cell B). (b) 28, (c) 29, and (d) 30, Scan rate: 10 mV s−1. The inset shows schematically the two-electrode electrochemical cell.
Di ≡
cOi
ciO ciW
⎤ ⎡zF = exp⎢ i (ΔOW ϕeq − ΔOW ϕi◦)⎥ ⎦ ⎣ RT
(2)
cW i
where and are the equilibrium concentrations of ion i in O and W, respectively. It is here assumed for simplicity that the concentration of an ion is equal to the activity and that any ion pair is not formed in either O or W phase. Under the present experimental conditions, always c0,W = c0,O (vide i i W supra), so that the value of ΔO ϕeq obtained from eq 1 and thus the value of Di from eq 2 should not depend on the concentration. In the distribution experiments of organic anions (A−), because CV+ as the extractant is extremely hydrophobic, CV+ remained almost completely in the NB phase, as clearly shown by spectrophotometric measurements. On the other hand, because Na+ is very hydrophilic, we assumed that Na+ initially added to the W phase entirely remained therein. It
representative cyclic voltammograms for the transfer of some representative cations (28−30) and also of Me4N+ (reference ion) at the micro NB/W interface. For the respective cations, a well-defined anodic (positive current) peak was observed for the transfer from W to NB. On the reverse cathodic (negative current) scan, however, no current peak was observed for the transfer back to W. This is because the transfer of the cations in the NB phase is due to the spherical diffusion, which prevents effectively the back diffusion from NB to the micro interface.21−23 For the anions studied, similar voltammetric behaviors were observed, though they gave a cathodic peak due to the transfer from W to NB (Figure S1 of the Supporting Information). We then used the anodic or cathodic peak potential to determine the ΔW O ϕ° 6012
DOI: 10.1021/acs.jpcb.5b01316 J. Phys. Chem. B 2015, 119, 6010−6017
Article
The Journal of Physical Chemistry B
Table 1. Distribution Ratios Observed and Calculated for Organic Ions in the NB−W System and Their Standard IonTransfer Potentials anion
RSO3−
RCO2−
RO−
cation
RNH3+ R2NH2+ R3NH+
entry
R
log D(obs)a
log D(calc)b
c ΔW O ϕ° (V)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Ph p-(CH3)Ph o,o,p-(CH3)3Ph m-(NO2)Ph o,p-(NO2)2Ph o,o,p-(NO2)3Ph 2-Naph n-C3H7 n-C6H13 Br(CH2)3 cyc-C6H11 Ph o-(CH3)Ph m-(CH3)Ph p-(CH3)Ph m-(NO2)Ph o,p-(NO2)2PhCH2 o,o,p-F3Ph Ph2CH n-C5H11 n-C7H15 cyc-(C6H11)CH2 o-(NO2)Ph o,o,p-(NO2)3Ph o,o,p-Cl3Ph o,o,p-Br3Ph cyc-(C6H11)CH2 MeEt Me2Et MeEt2
1.4 1.9 >2.0 >2.0 >2.0 >2.0 >2.0 0.4 1.4 1.6 1.2 0.9 1.0 1.1 1.2 >2.0 >2.0 1.0 1.6 0.6 1.4 0.8 >2.0 >2.0 >2.0 >2.0 >2.0 >2.0 >2.0 >2.0
1.3 1.6 1.8 2.1 2.6 3.3 2.0
