Electrochemistry Communications 11 (2009) 808–811

Contents lists available at ScienceDirect

Electrochemistry Communications

journal homepage: www.elsevier .com/locate /e lecom

Voltammetric determination of water with inner potential referenceand variable linear range based on structure- and redox-controllablehydrogen-bonding interaction between water and quinones

Limin Zhang 1, Haojie Zhou 1, Xianchan Li 1, Yuqing Lin, Ping Yu, Lei Su, Lanqun Mao *

Beijing National Laboratory for Molecular Sciences, Institute of Chemistry, The Chinese Academy of Sciences (CAS), Beijing 100190, China

a r t i c l e i n f o

Article history:Received 30 December 2008Received in revised form 29 January 2009Accepted 29 January 2009Available online 6 February 2009

Keywords:Voltammetric methodWater determinationQuinonesHydrogen-bonding interaction

1388-2481/$ - see front matter � 2009 Elsevier B.V. Adoi:10.1016/j.elecom.2009.01.037

* Corresponding author. Tel.: +86 10 62646525; faxE-mail address: lqmao@iccas.ac.cn (L. Mao).

1 Also in the Graduate School of CAS, Beijing 100049

a b s t r a c t

This communication describes a new voltammetric method for the determination of water in nonaque-ous solvent by taking advantage of the structure- and redox-controllable hydrogen-bonding interactionbetween quinone species and water. Three kinds of quinones, i.e., tetrachloro-p-benzoquinone (TCBQ),benzoquinone (BQ), and tetramethyl-p-benzoquinone (TMBQ), are employed in this study in terms oftheir different structures and thereby different basicities and hydrogen-bonding interaction activitieswith water. The hydrogen-bonding interaction activities of the quinone species with water actuallydepend on the structures and the species of quinones, where the interaction activity between quinonedianion and water remains remarkably greater than that between quinone monoanion and water. Theformer interaction activity eventually leads to the positive shift of the half-wave potential of quinonemonoanion/dianion couple, which can be essentially used for the voltammetric determination of water.The structure- and redox-controllable hydrogen-bonding interaction activities of quinones and watersubstantially make it possible to determine trace amount of water in the nonaqueous solution with innerreference potential and variable dynamic linear range.

� 2009 Elsevier B.V. All rights reserved.

1. Introduction

Determination of water is of great industrial and environmentalimportance and has been one of the commonest routine proce-dures in many research and industrial processes since water is of-ten used in the preparation of many materials and remains as oneof the most common contaminants in the organic solvents and inthe industrial products [1]. So far, many methods, such as gravi-metric, spectroscopic, and amperometric methods have been em-ployed for water determination, of which Karl Fischer methodhas been used most frequently [2–4]. In spite of their applicationsin the determination of water content in moisture, most of thosemethods are still experimentally complicated and may not be suit-able for the determination of water content in organic solutions[1,5]. Although voltammetric methods based on the redox proper-ties of the analytes have been proved to be both theoretically andexperimentally simple and, as such, have been widely used forelectroanalytical purposes, the poor redox property of water itselfunder the conditions generally employed for electrochemical mea-surements substantially makes the voltammetric determination ofwater a challenge in electrochemical studies.

ll rights reserved.

: +86 10 62559373.

, China.

This communication describes a new voltammetric method forthe determination of water in nonaqueous solution. The strategyis essentially based on the structure- and redox-controllablehydrogen-bonding interaction between water and quinones, asshown in reactions (1)–(4). In aprotic media, quinone undergoestwo successive one-electron electrochemical redox processes toform radical monoanion (Q��) (reaction (1)) and dianion (Q2�)(reaction (3)), which have different hydrogen-bonding interactionactivity with water (reactions (2) and (4)). The hydrogen-bondinginteraction stabilizes the radical monoanion and dianion andeventually leads to the positive shift of the half-wave potentialsof the quinone species [6,7], which can be employed for the deter-mination of water content in the organic solvent. Although thehydrogen-bonding interaction has been known to be ubiquitouslyinvolved in many biological processes and has been previouslyinvestigated with quinones as the model receptors [8–11], suchan interaction has not been used for the voltammetric determina-tion of water in organic solutions so far. This study essentially of-fers a simple and straightforward voltammetric method for waterdetermination.

Q þ e�Q �� ð1Þ

Q �� þ xH2O�Q ��ðH2OÞx ð2Þ

-30

-20

-10

0

10

20 O

O

Cl

ClCl

Cl

I / µ

A

TCBQ

-30

-20

-10

0

10

20 O

O

I / µ

A

BQ

-1.5 -1.0 -0.5 0.0

-30

-20

-10

0

10

20 O

O

CH3

CH3CH3

CH3

TMBQ I / µ

A

E / V vs . Ag/AgCl

A

B

C

Fig. 1. CVs obtained at GC electrodes in acetonitrile containing TCBQ (A), BQ (B), orTMBQ (C) in the absence (solid curves) and presence (dashed curves) of water(VH2 O/VACN = 0.001). The concentrations of TCBQ, BQ, and TMBQ were 1.0 mM. A0.1 M Bu4NClO4 was used as the electrolyte. Potential scan rate was 50 mV s�1.

L. Zhang et al. / Electrochemistry Communications 11 (2009) 808–811 809

Q ��ðH2OÞx þ e�Q 2�ðH2OÞx ð3Þ

Q 2�ðH2OÞx þ ðy� xÞH2O�Q 2�ðH2OÞy ð4Þ

Table 1Parameters involved in the redox processes of quinones and in the voltammetric determi

x Kð1Þeq y Kð2Þeq

TCBQ 0.60 2 4.1 3.3 � 104

BQ 1.4 29 5.7 1.4 � 109

TMBQ 1.5 32 5.8 2.1 � 109

2. Experimental

2.1. Chemicals

Tetrachloro-p-benzoquinone (TCBQ), benzoquinone (BQ), tetra-methyl-p-benzoquinone (TMBQ) and tetrabutylammonium per-chlorate (Bu4NClO4) were purchased from Aldrich. Acetonitrile(ACN) was obtained from Tianjin Siyou Chem. Co., Ltd. (Tianjin,China) and was distilled from CaH2 prior to use. Doubly distilledwater was used throughout the experiments.

2.2. Apparatus and electrochemical measurements

Glassy carbon (GC, 3 mm diameter) electrodes were used in thisstudy. The electrodes were polished with 0.3 and 0.05 lm aluminaslurry on a polishing cloth, cleaned under bath sonication for 5 minin acetone and distilled water, and thoroughly rinsed with doublydistilled water. Cyclic voltammetry was performed on an electro-chemical analyzer (CHI 660A, CH Instruments) in ACN solutioncontaining 0.10 M Bu4NClO4 with a three-electrode configurationwith GC electrodes as working electrode and a Pt wire as counterelectrode. All potentials were biased versus Ag/AgCl electrode witha glassy salt bridge with two compartments. One compartmentconnected to the ACN solution was filled with ACN containing0.1 M Bu4NClO4 and the other one connected to the KCl-saturatedaqueous solution was filled with saturated KCl solution. Prior toelectrochemical measurements, the electrolyte was bubbled withN2 gas for more than 30 min and the measurements were con-ducted under N2 atmosphere. All electrochemical measurementswere performed at ambient temperature.

3. Results and discussion

Fig. 1 compares cyclic voltammograms (CVs) of three kinds ofquinones in ACN solution in the absence and presence of water.In dry, neutral and proton-deficient media, such as ACN, all qui-nones exhibit two pairs of well-defined redox waves (solid curves),which were attributed to two successive one-electron redox pro-cesses to give quinone monoanions (Q��) and dianions (Q2�)[6,7]. The small peak-to-peak separation and the linear relation-ship between the peak currents and the square root of potentialscan rate in a range from 50 to 400 mVs�1 (data not shown) essen-tially suggest that the redox processes are diffusion-controlled fastelectron transfer processes. In addition, the half-wave potentials(E1/2) of both redox waves essentially vary with quinone structures,of which TCBQ with four electron-withdrawing chloro groups hasthe most positive E1/2 and, in contrast, TMBQ bearing four elec-tron-donating methyl groups has the most negative one, amongall quinones used here. The structurally dependent half-wavepotentials of quinones actually reflect the variable basicity andthe capability of quinones as the receptors in the hydrogen-bond-ing interaction with water and such dependency substantiallymakes it possible to tune the analytical properties for the voltam-metric determination of water, as described below.

The addition of water in solution results in a clear positive shiftof the half-wave potential for Q��/Q2� (Eð2Þ1=2) (dashed curves, Fig. 1),which could be elucidated in terms of the hydrogen-bonding

nation of water.

Linear regression R Linear range

DE1/2 = �0.090 ln(VH2 O/V) + 0.30 0.9972 0.6–60%DE1/2 = �0.080 ln(VH2 O/V) + 0.03 0.9989 0.04–6%DE1/2 = �0.074 ln(VH2 O/V)�0.01 0.9902 0.03–3%

- 0 . 8 - 0 .4 0 .0 0 .4

-30

-15

0

15A

I / µ

A

E / V vs Ag/AgCl

0.0 0.2 0.4 0 .6

- 0.5

0.0

B

E 1/2 /

V

VH2O /VACN

-7.5 -5.0 - 2.5

0.2

0.4

0.6

0.8

C

E 1/2 /

V

ln (VH2O/V )

Δ

Fig. 2. (A) CVs obtained at GC electrodes in acetonitrile containing 1.0 mM TCBQ inthe presence of water. The volume ratio of water to acetonitrile (VH2 O/VACN) was(from left to right) 0, 0.001, 0.003, 0.012, and 0.020. Scan rate was 50 mV s�1. (B)Plots of E1/2 of the first (d) and second (�) redox waves of TCBQ against VH2 O/VACN.(C) Plots of DE1/2 of TCBQ (N), BQ (d), and TMBQ (j) against ln(VH2 O/VACN).

810 L. Zhang et al. / Electrochemistry Communications 11 (2009) 808–811

interaction between Q2� and water (reactions (2) and (4)), as re-ported previously [7,12]. In contrast, such a procedure caused anegligible shift of the half-wave potential for Q/Q�� (Eð1Þ1=2), indica-tive of a relatively weak interaction between Q�� and water. Thesedemonstrations essentially suggest that the hydrogen-bondinginteraction between quinone species, such as Q2� and Q��, isredox-dependent. For the first redox process, i.e., reactions (1)and (2) [7],

Eð1Þ1=2 ¼ E01=2ð1Þ þ RT=F ln 1þ Kð1Þeq Cx

H2O

� �ð5Þ

where, E01=2ð1Þ

is the half-wave potential of Q/Q�� redox couple in theabsence of water in solution, x is the number of water moleculehydrogen-bonded to Q��, Kð1Þeq is the equilibrium constant of reaction(2), and CH2O is the concentration of water. If Kð1Þeq Cx

H2O � 1, a plot ofEð1Þ1=2 vs. logCH2O gives a straight line with a slope of 2.3xRT/F, fromwhich the value of x could be estimated.

For the second redox process, i.e., reactions (3) and (4),

Eð2Þ1=2 ¼ E01=2ð2Þ þ RT=F ln 1þ Kð2Þeq

� �Cy

H2O= 1þ Kð1Þeq

� �Cx

H2O ð6Þ

where, y and Kð2Þeq are the number of water molecules hydrogen-bonded to Q2� and the equilibrium constant of reaction (4), respec-tively. For a strong hydrogen-bonding, 1 in the denominator andnumerator in Eq. (6) could be neglected and thus the value of(y–x) could be obtained from the plot of Eð2Þ1=2 vs. log CH2O. The valuesof y for three kinds of quinones were estimated by taking x valuesfrom the first redox wave and the values of Kð2Þeq were thus calculatedfrom the intercept of the plot of Eð2Þ1=2 vs. log CH2O with the known x, y,

and Kð1Þeq . The parameters of x, y, Kð1Þeq and Kð2Þeq for three kinds of struc-turally different quinones used here were summarized in Table 1.

Consistent with the experimental observation on the negligibleshifts of Eð1Þ1=2 caused by the addition of water (Fig. 1, dashed curves),both the numbers of water molecules and the equilibrium con-stants of the hydrogen-bonding interaction between Q�� and waterfor three quinones are smaller, as compared with those of thehydrogen-bonding interaction between Q2� and water for threequinones. This demonstrates that the former interactions are muchweaker than the latter ones and further confirm the structure- andredox-dependent hydrogen-bonding interactions between qui-nones and water. The large shifts of Eð2Þ1=2 in response to the additionof water essentially offer a basis for the voltammetric determina-tion of water in the nonaqueous solution. Meanwhile, the negligibleshifts of Eð1Þ1=2 caused by the addition of water substantially enablesthe first half-wave potential to be used as the inner potential refer-ence for the voltammetric determination of water. This is remark-able especially for the cases when the determination is based onthe potential shift since the reference potential must be stable forthe accurate measurements. In most cases, the Fc/Fc+ couple wasused as the reference in organic media [13]. However, this generallyrequired a post-calibration. Moreover, the most commonly used Ag/AgCl electrode may not be used in organic media because of thehigh liquid junction potential existing on the interface between ref-erence electrode in aqueous phase and the organic solvent.

Moreover, a clear comparison of the separation between thehalf-wave potentials of the first and second wavesðDE1=2 ¼ Eð1Þ1=2 � Eð2Þ1=2Þ of different kind of quinones caused by theaddition of the same amount of water (Fig. 1, dashed curves) re-veals that the DE1/2 essentially varies with the structures of thequinones used in this study. Under the conditions employed here,the addition of the same amount of water into solution leads toDE1/2 in an order of TMBQ > BQ > TCBQ. Such phenomena couldbe elucidated in terms of the different basicity and reactivity ofquinone species in the hydrogen-bonding interactions, of whichthe basicity of the dianion form of TMBQ remains the strongestamong three quinones. The structurally dependent basicity and

L. Zhang et al. / Electrochemistry Communications 11 (2009) 808–811 811

reactivity eventually make it possible to tune the analytical prop-erty, such as linear range, for the voltammetric method demon-strated here for the determination of water, as described below.

From Eqs. (5 and 6), one may readily deduce that the value ofDE1/2 could be used for the determination of water content. WithTCBQ as a typical example, Fig. 2 depicts the voltammetric re-sponse of water in ACN solution. The addition of water in ACN solu-tion clearly leads to the positive shift of Eð2Þ1=2 of quinones, while Eð1Þ1=2

remains almost unchanged (panels A and B). For a simple analysis,CH2O was taken by VH2O=V ;V ¼ VH2O þ VACN. It could be seen thatthe values of DE1/2 were linear with ln(VH2O/V), as summarized inTable 1. As anticipated, the dynamic linear range for water deter-mination depends on the hydrogen-bonding interaction activityof quinones with water and is consequently structurally controlla-ble. For example, the high basicity and the good capability of BQand TMBQ receptors substantially enable the dynamic linear rangefor water determination down to a relatively low level, whereas,the low basicity and the poor capability of TCBQ receptor extendthe dynamic linear range up to a high level. This structurally tun-able linear range essentially offers new possibilities for the voltam-metric determination of water at different levels.

4. Conclusions

By taking advantages of the structure- and redox-dependenthydrogen-bonding interaction between quinone species and water,we have demonstrated a new voltammetric method for the deter-mination of water with inner potential reference and variable lin-ear range. Compared with the existing methods, the voltammetricmethod demonstrated here may be facile and sensitive and has

tunable linear range and thus could be used for the determinationof water content in nonaqueous solution. This study may offer anew and simple approach for the electrochemical determinationof water in nonaqueous media.

Acknowledgements

We gratefully acknowledge the financial support from NSF ofChina (Grant Nos. 20625515, 20721140650, 90813032 and for L.M., and 20705034 for L. S.), National Basic Research Program ofChina (973 Program, 2007CB935603), Chinese Academy of Sci-ences, and Institute of Chemistry.

References

[1] M.A. Rahman, M. Won, N. Kwon, J. Yoon, D. Park, Y. Shim, Anal. Chem. 80(2008) 5307.

[2] A. Saraullo, P.A. Martos, J. Pawliszyn, Anal. Chem. 69 (1997) 1992.[3] T. Gupta, M.E. van der Boom, J. Am. Chem. Soc. 128 (2006) 8400.[4] Q. Kuang, C. Lao, Z. Wang, Z. Xie, L. Zheng, J. Am. Chem. Soc. 129 (2007) 6070.[5] I.N. Andersson, A. Cedergren, Anal.Chem. 57 (1985) 2571.[6] S. Hayano, M. Fujihira, Bull. Chem. Soc. Jpn. 44 (1971) 2051.[7] N. Gupta, H. Linschitz, J. Am. Chem. Soc. 119 (1997) 6384.[8] J. Bu, N.D. Lilienthal, J.E. Woods, C.E. Nohrden, K.T. Hoang, D.K. Smith, J. Am.

Chem. Soc. 127 (2005) 6423.[9] J.A. Jeffrey, Hydrogen Bonding in Biological Structures, Springer-verlag, Berlin,

1991.[10] S.J. Brooks, P.R. Birkin, P.A. Gale, Electrochem. Commun. 7 (2005)

1351.[11] M. Gómez, I. González, F.J. González, R. Vargas, J. Garza, Electrochem. Commun.

5 (2003) 12.[12] M. Quan, D. Sanchez, M.F. Wasylkiw, D.K. Smith, J. Am. Chem. Soc. 129 (2007)

12847.[13] P.S. Singh, D.H. Evans, J. Phys. Chem. B 110 (2006) 637.