Differential Pulse Voltammetry for Ion Transfer at Liquid Membranes

Anal. Chem. 2009, 81, 4220–4225

Differential Pulse Voltammetry for Ion Transfer at Liquid Membranes with Two Polarized Interfaces A. Molina,*,† C. Serna,† J. A. Ortun˜o,‡ J. Gonzalez,† E. Torralba,† and A. Gil‡ Departamento de Quı´mica Fı´sica, and Departamento de Quı´mica Analı´tica, Facultad de Quı´mica, Universidad de Murcia, 30100 Murcia, Spain A simple analytical expression for the response of the double-pulse technique differential pulse voltammetry (DPV) corresponding to ion transfer processes in systems with two liquid/liquid polarizable interfaces has been deduced. This expression predicts lower and wider curves than those obtained with a membrane system with a single polarizable interface. Moreover, the peak potential of these systems is shifted 13 mV from the half-wave membrane potential. We have applied this expression to study the ion transfer of drugs with different pharmacological activities (verapamil, clomipramine, tacrine, and imipramine), at a solvent polymeric membrane ion sensor. Ion transfer across the interface between two immiscible electrolyte solutions (ITIES) is a fundamental charge-transfer process that has applications in several fields, such as electroanalysis, liquid-liquid extraction, electroassisted extraction, phase transfer catalysis, and electrocatalysis.1-3 In addition, it serves as a simplified model for ion transfer in biomembranes.4,5 Interestingly, it provides a direct way to quantify the ion lipophilicity through the electrochemical determination of the standard ion transfer potential.6 Many of the systems used in electrochemical studies on ion transfer at ITIES consist of a membrane separating two aqueous solutions, i.e., the sample solution and an inner solution. In these systems the polarization is usually only effective at the sample solution/membrane interface (outer interface), since the potential drop through the inner solution/membrane interface can be kept constant by using a sufficiently high concentration of a common ion in both adjacent phases. To achieve this, two salts of this common ion with hydrophilic and lipophilic counterions are dissolved in the inner aqueous and organic phases, respectively.

A particular case arises when it is not possible to find two salts containing a common ion with the above characteristics and enough solubility in the respective phases to achieve sufficiently high concentrations in the membrane and in the inner aqueous phase. In this situation, both liquid/liquid interfaces in the membrane system are polarized.3,7-10 In a previous paper we obtained easy, analytical, explicit expressions for the I-E response and concentration profiles corresponding to the ion transfer through a membrane system with two polarizable interfaces, which are valid for any multipulse potential technique.11 The aim of this paper is to give the steps for the application of the double-pulse technique differential pulse voltammetry (DPV) to these systems, because this technique is one of the most suitable for quantitative analysis and determination of the characteristic parameters of the system, such as the standard ion transfer potential, owing to its high sensitivity, resolving power, and also to the fact that the undesirable effects on the currents can be minimized.12,13 It therefore enables small differences in the standard ion transfer potentials of species with very similar chemical structure to be distinguished. For this purpose, we have deduced a simple, explicit, and analytical expression of the I-E curve in DPV which has allowed us to fully characterize this response for reversible systems. From this expression, we have obtained those corresponding to the peak parameters and we have observed that the peak potential of these systems is shifted 13 mV from the half-wave membrane potential. Moreover, the DPV curves obtained in this system are lower and wider than those obtained at a membrane system with a single polarizable interface (i.e., the half-peak width is around 130 mV vs the 90 mV which is obtained when only one interface is considered). We have applied the expressions deduced for DPV response to study the ion transfer of several drugs with different pharmacological activities at a solvent polymeric membrane ion sensor in order to determine the standard potentials corre-

* To whom correspondence should be addressed. Phone: +34 968 367524. Fax: +34 968 364148. E-mail: [email protected]. † Departamento de Quı´mica Fı´sica. ‡ Departamento de Quı´mica Analı´tica. (1) Samec, Z. Pure Appl. Chem. 2004, 76, 2147–2180. (2) Reymond, F.; Fermı´n, D.; Lee, H. J.; Girault, H. H. Electrochim. Acta 2000, 45, 2647–2662. (3) Vany´sek, P.; Basa´ez Ramı´rez, L. J. Chil. Chem. Soc. 2008, 53, 1455–1463. (4) Koryta, J. Ions, Electrodes and Membranes, 2nd ed.; Wiley: Chichester, U.K., 1991, Chapter 3. (5) Bender, C. J. Chem. Soc. Rev. 1988, 17, 317–346. (6) Reymond, F.; Chopineaux-Courtois, V.; Steyaert, G.; Bouchard, G.; Carrupt, P.; Testa, B.; Girault, H. H. J. Electroanal. Chem. 1999, 462, 235–250.

(7) Shirai, O.; Kihara, S.; Yoshida, Y.; Matsui, M. J. Electroanal. Chem. 1995, 389, 61–70. (8) Samec, Z.; Trojanek, A.; Langmaier, J.; Samcova´, E. J. Electroanal. Chem. 2000, 481, 1–6. (9) Kakiuchi, T. Electrochim. Acta 1998, 44, 171–179. (10) Ulmeanu, S. M.; Jensen, H.; Samec, Z.; Carrupt, P.-A.; Girault, H. H. J. Electroanal. Chem. 2002, 530, 10–15. (11) Molina, A.; Serna, C.; Gonza´lez, J.; Ortun ˜o, J. A.; Torralba, E. Phys. Chem. Chem. Phys. 2009, 11, 1159–1166. (12) Ortun ˜o, J. A.; Serna, C.; Molina, A.; Gil, A. Anal. Chem. 2006, 78, 8129– 8133. (13) Ortun ˜o, J. A.; Gil, A.; Serna, C.; Molina, A. J. Electroanal. Chem. 2007, 605, 157–161.

4220

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

10.1021/ac802503b CCC: $40.75  2009 American Chemical Society Published on Web 04/29/2009

1 1/2 1 Figure 1. Potential waveform corresponding to the DPV technique for three values of the potentials applied (solid lines): EM < EM (A), EM = 1/2 1 1/2 EM (B), and EM . EM (C). Black circles indicate the times at which the current is measured. Dotted and dashed lines show the calculated distribution of the potentials applied (eq 10) between the outer and the inner interfaces, respectively.

sponding to the ion transfer from water to the solvent polymeric membrane. The ion transfer standard potential of an ionizable drug allows the quantifying of its lipophilicity,13 which is one of the most significant physicochemical properties related with its pharmacological activity.6 The drugs studied here are verapamil, an antiarrhythmic drug also used in treating hypertension and angina pectoris, tacrine, the first centrally acting cholinesterase inhibitor approved for the treatment of Alzheimer’s disease, imipramine, a tricyclic antidepressant, and clomipramine, a tricyclic drug with both antidepressant and antiobsessional properties.14 EXPERIMENTAL SECTION Apparatus. The design of the voltammetric ion sensor has been previously reported.12 Briefly, a Pt wire counter electrode was accommodated inside the inner solution compartment of a Fluka ion-selective electrode (ISE) body. This permits the use of a four-electrode potentiostat like that described in ref 15. A glass ring of 28 mm i.d. and 30 mm height, glass plate, vial, and punch were purchased from Fluka for the construction of the membranes. Reagents and Solutions. Poly(vinyl chloride) (PVC) high molecular mass, 2-nitrophenyl octyl ether (NPOE), and tetrahydrofuran (THF) were Selectophore products from Fluka. Tetradodecylammonium tetrakis-(4-chlorophenyl) borate (TDDATClPB), tetrabutylammonium chloride (TBAC), and (10,11-dihydroN,N-dimethyl-5H-dibenz[b,f]azepine-5-propanamine) (imipramine, Im), (3-chloro-10,11-dihydro-N,N-dimethyl-5H-dibenz[b,f]azepine5-propanamine) (clomipramine, Cm), 9-amino-1,2,3,4-tetrahydroacridina hydrochloride (tacrine, Tc), and 5-[N-(3,4-dimethoxyphenylethyl)methylamino]-2-(3,4-dimethoxy-phenyl)-2isopropylvaleronitrile (verapamil, Vr) were purchased from Sigma in the form of hydrochlorides. All the other reagents used were of analytical reagent grade. Nanopure water (18 MΩ) prepared with a Milli-Q (Millipore) system was used throughout. Membrane Preparation. The membranes were prepared by dissolving 200 mg of NPOE, 100 mg of PVC, and 8.4 mg of TDDA-TClPB in 3 mL of THF. This solution was poured into the glass ring resting on the glass plate and was left overnight (14) Brunton, L.; Lazo, J.; Parker, K., Eds. Goodman and Gilman’s The Pharmacological Basis of Therapeutics, 11th ed.; McGraw-Hill: New York, 2005. (15) Sa´nchez-Pedren ˜o, C.; Ortun ˜o, J. A.; Herna´ndez, J. Anal. Chim. Acta 2002, 459, 11–17.

to allow the solvent to evaporate slowly. A 7 mm diameter piece was cut out with the punch and incorporated into the modified ISE body described above. Electrochemical Measurements. The electrochemical cell used can be expressed as Ag|AgCl|5 × 10-2 M LiCl|4 × 10-2 M TDDA-TClPB|5 × 10-2 M LiCl, xM X+Cl- |AgCl|Ag with X+ ) Im+, Cm+, Tc+, Vr+, TBA+. The applied potential, EM, is defined as the potential difference between the right- and left-hand terminals. EM is maintained at the preset value by means of the four-electrode potentiostat that applies the necessary potential between the right and left counter electrodes. A positive current corresponds to the transfer of positive charge from the outer aqueous solution to the organic phase, i.e., through the working interface, which implies the same current sign for the coupled transfer of positive charge from the organic phase to the inner aqueous solution. The potential-time waveform used in the DPV and the general procedure to obtain the DPV recordings were described in ref 1 2 12. In our study, consecutive potential steps (EM and EM ) with a 2 1 pulse amplitude (∆E ) EM - EM) of 50 mV were applied on a base potential (E0) of 75 mV, during times t1 ) 12.5 s and t2 ) 0.25 s, respectively, being the current measured at the end of 1 each potential step (see Figure 1). The magnitude of EM was linearly changed between consecutive double pulses. The delay time between each pair of pulses, td, was 50 s. THEORY Let us consider the ion transfer through a liquid membrane system like that described in ref 11, consisting of a planar organic phase (the membrane, M), separating two aqueous solutions, w1 and w2 phases, with the former containing the X+ ion whose transfer across a liquid/liquid interface (w1/M) is going to be studied. The cell used for this study can be described by the following scheme: B+A-(w2)|R+Y-(M)|B+A-, X+A-(w1)

(1)

where B+A- is the supporting electrolyte for the aqueous phases w1 and w2, and R+Y- is the supporting electrolyte for the organic phase. Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

4221

This type of membrane system in which there is no common ion in the liquid membrane and aqueous phases implies two polarizable liquid/liquid interfaces. So, when an external polarization is applied, it is distributed between the two interfaces and the potential drop through each interface cannot be controlled individually, although the electrochemical processes occurring at each interface are linked to each other by virtue of the same intensity of electrical current. We assume that all the phases, w1, w2, and M, contain sufficient concentrations of electrolytes such that the different ohmic drops can be neglected. In these conditions the applied potential, EM (membrane potential), can be written as the difference of the two interfacial potential differences, EM ) Eout - Einn, due to the transfer of w X+ through the outer interface w1/M (Eout ) ∆M φX+) coupled with the transfer of the membrane electrolyte cation, R+, w through the inner interface w2/M (Einn ) ∆M φR+). 11 In a previous paper we developed the theory for the response corresponding to the reversible ion transfer through a membrane system with any multipotential technique. This theory provides an explicit general expression for the current corresponding to the pth potential, Ip, of any sequence of consecutive potential steps (eq 29 in ref 11). From this equation, the following particular expressions for the cur1 rents corresponding to two consecutive potential steps, EM 2 and EM, are obtained:




I1 ) FA




I2 ) FA

DXw+1 π

c*X+

[√

1

t1 + t2

DXw+1 πt1

1 c*X+g(ηM )

1 g(ηM )+

1

√t2

(2)

2 1 (g(ηM ) - g(ηM ))

]

(3)

where t1 and t2 are the duration of the first and second potential steps, respectively, c*X+ is the initial concentration of the studied w ion, X+, in the outer aqueous solution (w1 phase), DX+1 is the corresponding diffusion coefficient, F and A have their usual j meaning, and the g(ηM ) function (j ) 1 or 2) is given by

j g(ηM ))

√(λ+eη

j

M

)2 + 8λ+eη M - λ+eη M 4 j

j

j ) 1, 2

(4)

with F j 0′ (E - EM ) RT M j j j EM ) Eout - Einn

j ηM )

j ) 1 or 2

0′ w 0′ w 0′ ) ∆M φX+ - ∆M φR+ EM

(5)

(6)

and

λ+ )

2√DRw+2 DXM+ c*R+ c*X+ Dw+1 X

4222

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

(7)

In these last two equations, ∆wMφ0′X+ and ∆wMφ0′R+ are the formal ion transfer potentials for the target ion X+ and cation R+ of the membrane electrolyte, respectively, c*R+ is the concentraM tion of the membrane electrolyte cation, R+, and DX+ and w2 + DR+ are the diffusion coefficients of X in the membrane (M phase) and R+ in the inner aqueous solution (w2 phase), respectively. 1/2 The half-wave potential, EM is given by11 1/2 0′ EM ) EM -

RT ln λ+ F

(8)

When the coupled ion transfer at the inner interface is that of the anion of the supporting electrolyte of the aqueous phase w2 (A-), similar equations are obtained, but by changing λ+ for λ- in the above equations:

λ- )

2√DAM-DXM+ c*Ac*X+ Dw+1

(9)

X

The individual potential drops at each interface caused by the j j application of the first and second potential steps, Eout and Einn with j ) 1 or 2, respectively, can be expressed as a function of the corresponding applied potential EjM through g(ηjM) function, as follows (see eq 36 in ref 11):

j w 0′ Eout ) ∆M φX+ +

j Einn

)

w 0′ ∆M φR+

(
(
)

RT ln F

RT ln + F

DXw+1 DXM+

+

DRw+2 c*R+ DXw+1 c*X+

j g(ηM ) RT ln j F 1 - g(ηM )

)

j)

RT j ln(g(ηM )) F 1, 2

(10)

where we have assumed that the coupled ion transfer at the inner interface is that of the cation of the membrane electrolyte (R+). RESULTS AND DISCUSSION In the DPV technique two consecutive potential steps, E1M and 2 EM, are applied over times t1 and t2, with t1 . t2. The signal is built by subtracting the currents obtained at the end of both pulses, IDPV ) I2 - I1. The magnitude of E1M is linearly changed between consecutive double pulses, while the pulse amplitude, 2 1 ∆E ) EM - EM , is kept constant. Figure 1 shows the potential-time waveform used in this 1 technique for three different values of the potentials applied EM 2 and EM (solid lines) and its distribution between the outer (dotted lines) and the inner (dashed lines) interfaces calculated from eq 10. As can be seen, the inner interface is quite sensitive to the external polarization at those potential values corresponding 1 1/2 to the beginning of the dc wave (EM , EM ). However, as the applied potential is scanned toward the limiting current region 1 1/2 (EM . EM ), the potential drop through this inner interface becomes constant (compare Figure 1, part A with parts B and C). This behavior also can be observed in Figure 2, where we have plotted the current corresponding to the first potential pulse (curves a), to the second pulse (curves b) and to the DPV response (curves c) versus the membrane potential (solid

Figure 2. Normalized IN1 - E (a), IN2 - E (b), and IN,DPV - E (c) curves corresponding to a system with two polarizable liquid/liquid interfaces, w 1 1 1 calculated from eqs 2, 3, 11, and 10, being (IN ) I/(FAc*X+(DX+1 t2/π)1/2 ). The potential Ewis EM (solid lines), Eout (dotted lines), and Einn (dashed w2 1 M w 0′ w 0′ -5 2 -1 -8 2 -1 + + + + + lines). ∆MφX ) -200 mV, ∆MφR ) -350 mV, ∆E ) 40 mV, t1 ) 12.5 s, t2 ) 0.25 s, DX ) DR ) 10 cm s , DX ) 10 cm s , c*X+ ) 0.1 mM, c*R+ ) 50 mM, T ) 298.15 K.

lines) and versus the potential drops in the outer interface (dotted lines) and in the inner one (dashed lines). From eqs 2 and 3, and taking into account that in DPV the condition t1 . t2 is fulfilled (i.e., t1 + t2 = t1), the following simple expression for the response in this technique is obtained:




IDPV ) FA

DXw+1 2 1 c*+[g(ηM ) - g(ηM )] πt2 X

(11)

j with g(ηM ) given by eq 4. From these equations we have obtained by mean numerical peak peak fitting that the peak parameters, EDPV and IDPV , for ∆E values less than 50 mV, are given by

peak 1/2 EDPV (mV) = EM + 13.0




peak IDPV = FA

DXw+1 c*+(8.91 × 10-3)∆E πt2 X

(12)

(13)

and the half-peak width is 1/2 WDPV (mV) = 131 + (2.43 × 10-3)∆E

(14)

with ∆E in the above equations expressed in millivolts. Figure 3 shows the DPV curves corresponding to a membrane system with two polarizable interfaces (solid lines) and also to a system with a single polarizable interface (dashed lines), obtained for two values of the pulse amplitude ∆E. The current IDPV has been 1/2 plotted in all the cases versus the difference Eindex - EM , with 1 2 1 Eindex ) (EM + EM)/2. The use of Eindex instead of the usual EM 16 is of great interest since the IDPV - Eindex plots are centered about the half-wave potential in the case of a single polarizable interface system (see dashed curves). As can be seen, the DPV peaks obtained for the liquid membrane system are shifted 13 mV with respect to those obtained when only one polarizable interface is used, in agreement with eq 12. Moreover, in the first case the IDPV - Eindex curves are lower (around 40-45%) and wider than those obtained in the second case (W1/2 DPV = 131 mV vs the 90 mV observed when only one interface is considered). (16) Molina, A.; Morales, I. Int. J. Electrochem. Sci. 2007, 2, 386–405.

1/2 Figure 3. Normalized IN,DPV - (Eindex - EM ) curves calculated from eq 11 of this paper (solid lines) and also from eq 1 of ref 12 (dashed lines) corresponding to systems with a single polarizable interface. The values of ∆E are on the curves. Other conditions were the same as in Figure 2.

In Figure 4, we have evaluated the effect of the concentration of the target ion X+ (Figure 4a) and of the membrane electrolyte cation R+ (Figure 4b) on the DPV curves of a liquid membrane system for a pulse amplitude of 50 mV. From these figures it is clear that the increase of c*X+ has a double effectsan increase of the peak current (see eq 13) and a shift of the peak potential toward more positive values through 1/2 an increase of EM (or λ+, see eqs 7 and 8). The increase of c*R+ only causes a shift of the peak potential toward more negative values (see eq 7). Figure 5 shows the background-corrected experimental DPV curves corresponding to four different drugs obtained for a pulse amplitude of ∆E ) 50 mV. The measured values of pH for the 1 × 10-4 M solutions used of the different drugs in 0.05 M LiCl were around 5.7. The reported pKa values of imipramine, verapamil, tacrine, and clomipramine are 9.5, (17) Shalaeva, M.; Kenseth, J.; Lombardo, F.; Bastin, A. J. Pharm. Sci. 2008, 97, 2581–2606.

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

4223

Table 1. Standard Ion Potentials Obtained from DPV Curves of Figure 5 drug

w 0 ∆M φX+/mV

imipramine verapamil tacrine clomipramine

-163 -157 -115 -175

drug is in its protonated form, such that the measured currents correspond to the simple ion transfer of the corresponding protonated drugs. In order to obtain values of the ion transfer potentials of these drugs, we have chosen the reference ion TBA+ due to the proximity between its own ion transfer potential and those corresponding to the different species analyzed. w 1/2 EM and A(DX+1 )1/2 have been obtained from values of the peak potentials and currents by means of eqs 12 and 13. Once we have obtained both set of values, we have calculated theoretical DPV curves by using eq 11. As can be seen, an excellent agreement was obtained in all the cases. The standard ion transfer potential of each ionic drug X+ was calculated from the difference between its half-wave potential and that corresponding to the reference ion TBA+, obtained by using the same pulse amplitude and concentrations for both ions. Thus, by applying eq 8 and taking into account the relation between formal and standard ion transfer potentials

Figure 4. Theoretical IDPV - Eindex curves calculated from eq 11 with 0′ EM ) 80.0 mV, A ) 0.15 cm2, ∆E ) 50 mV. (a) c*R+ ) 40 mM and c*X+ values are on the curves; (b) c*X+ ) 0.1 mM and c*R+ values are on the curves. Other conditions were the same as in Figure 2.

w 0 w 0 1/2 1/2 EM,X + - EM,TBA+ ) ∆MφX+ - ∆MφTBA+ +

(

)

w1 γXM+ γTBA + RT ln w M + 1 F γX+ γTBA+

[( )

M DTBA + RT ln M F DX+

1/2

DXw+1 w1 DTBA +

]

(15)

where γjp and Djp are the activity and the diffusion coefficient of ion j in the phase p, respectively. This difference is independent of the ion transfer at the inner interface, be it that of the cation of the membrane electrolyte (R+) or that of the anion of the inner aqueous solution (A-). In the above equation, the two last terms on the right-hand side are practically zero, in such a way that w 0 w 0 1/2 1/2 EM,X + - EM,TBA+ ) ∆MφX+ - ∆MφTBA+

(16)

8.8, 9.8, and 9.3, respectively.17,18 So, in all these solutions, pH < pKa - 2, which implies that more than 99% of every

In order to calculate ∆wMφ0X+, we have used a value for ∆wMφ0TBA+ ) -242 mV.19 The values obtained are shown in Table 1. From these, several conclusions can be drawn. First, this system with two polarizable interfaces allows us to determine standard ion transfer potentials of drugs which are much more lipophilic than those studied with an analogous system consisting of a single polarizable interface.13 Second, the use of DPV with this system is able to distinguish small differences between standard potentials corresponding to drugs with very similar chemical structures, such as imipramine and clomipramine. Third, the value found for tacrine is equal to that previously reported18

(18) Malkia, A.; Liljeroth, P.; Kontturi, K. Electrochem. Commun. 2003, 5, 473– 479.

(19) Langmaier, J.; Stejskalova´, K.; Samec, Z. J. Electroanal. Chem. 2001, 496, 143–147.

Figure 5. Background-subtracted DPV recordings obtained for 1 × 10-4 M solutions of (black circles) imipramine, (white triangles) clomipramine, (white squares) verapamil, and (black diamonds) tacrine. Solid lines correspond to the theoretical DPV curves calculated from eq 11 with ∆E ) 50 mV, t1 ) 12.5 s, t2 ) 0.25 s, w and the following values of A(DX+1 )1/2 in cm3 s-1/2: Vr, 0.72 × 10-4; Im, 0.93 × 10-4; Cm, 1.04 × 10-4; Tc, 1.11 × 10-4. T ) 298.15 K.

4224

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

for a membrane containing a much lower PVC content than our membrane (5% instead of 32%). This extends to higher PVC contents the observation that the presence of PVC in the membrane composition does not alter the standard ion transfer potential values with respect to the plasticizer NPOE alone.

(project nos. CTQ2005-06977/BQU, CTQ2006-12552/BQU, and CTQ2008-04806/BQU) and the Fundacio´n SENECA (Projects Number 03079/PI/05 and 08813/PI/08).

ACKNOWLEDGMENT The authors greatly appreciate the financial support provided by the Direccio´n General de Investigacio´n Cientı´fica y Te´cnica

Received for review November 26, 2008. Accepted April 4, 2009. AC802503B

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

4225