Electrochimica Acta 69 (2012) 152–159

Contents lists available at SciVerse ScienceDirect

Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta

Differential dynamic potentiometry with ion selective electrodes: A tool for drug
fingerprinting
María Cuartero a , Joaquín A. Ortuño a,∗ , Ma Soledad García a , Francisco Martínez-Ortiz b
a
Department of Analytical Chemistry, Faculty of Chemistry, University of Murcia, 30071 Murcia, Spain
b
Department of Physical Chemistry, Faculty of Chemistry, University of Murcia, 30071 Murcia, Spain

a r t i c l e i n f o a b s t r a c t

Article history: Differential dynamic potentiometry (DDP) is a novel potentiometric technique, which consists of record-
Received 27 September 2011 ing the dynamic potential difference between two ion-selective electrodes (ISEs). In this article, we
Received in revised form 21 February 2012 describe its application to studying the dynamic contribution of beta-cyclodextrin (␤-CD) to the poten-
Accepted 24 February 2012
tial response of drug-selective electrodes. Surprisingly, DDP responses in serial calibration mode were
Available online 3 March 2012
characteristic for each individual drug, and might serve as drug-fingerprints. Also, non-monotonic DDP
responses were obtained using single concentration step experiments, and these were used for quantita-
Keywords:
tive drug analysis. Theoretical simulations of the DDP responses based on a reported dynamic diffusion
Differential dynamic potentiometry
Ion-selective electrodes
model allowed both types of the experimentally obtained signals to be predicted.
Cyclodextrin © 2012 Elsevier Ltd. All rights reserved.
Drug-fingerprints
Non-monotonic signals

1. Introduction
Dynamic response studies of ion-selective electrodes (ISEs) are
important at three levels at least. First, they permit the response
time of the sensor to be quantified [1], which is important for
practical applications; second, they give insight into the working
mechanism, which is particularly relevant in the presence of inter-
fering ions [2,3]; and third, they supply a data matrix (potential
and time) that can be exploited for qualitative and quantitative
purposes [4].
In 1970, Brand and Rechnitz reported a new instrument which
made it possible to measure the differential potential between two
high impedance ISEs [5]. The potential difference was obtained by
means of an analogue circuit. Here, we use a homemade instrument
and software which can simultaneously measure the potentials of
two ISEs with respect to a common reference electrode and mon-
itor the potential difference between both ISEs. In this paper, we
define differential dynamic potentiometry (DDP) as the recording of
the potential difference between both ISEs versus time.
The initial aim of the present work was to study the dynamic
contribution of a cyclodextrin to the potential response of
cyclodextrin-based drug-selective electrodes. This was done by
measuring the DDP response between a membrane electrode con-
taining PVC, plasticizer, ionic additive and ␤-CD, and another with a
∗ Corresponding author. Tel.: +34 868 887 414; fax: +34 868 887 682.
E-mail address: jortuno@um.es (J.A. Ortuño).
blank membrane without ␤-CD, when both electrodes are simulta-
neously exposed to different drugs. If the dynamic response of the
electrodes for different drugs is to be compared, it is necessary to
achieve a reproducible initial stage of the corresponding membrane
systems. Interestingly, when this was done and the electrodes were
calibrated in serial calibration mode, the DDP responses obtained
for all the drugs showed unusual shapes characterized by a poten-
tial inversion. Moreover, a different shape for the different drugs
assayed was obtained. These signals are here proposed as drug fin-
gerprints that may be useful for drug identification purposes. Also,
when discrete calibration mode was followed, the non-monotonic
DDP responses obtained for each concentration were used for quan-
titative analysis.
This paper is proposed as new contribution to modern poten-
tiometry, which had its starting point in the mid-1960s [6]. By the
early 1990s, potentiometry seemed to have reached a standstill
because no major developments seemed to be possible. In con-
trast, recent novel developments show that progress in the field
of potentiometric sensors is livelier than it was fifteen years ago
[7–12]. Many of these achievements, including those described
here, rely on dynamic processes involving ion fluxes through the ISE
membrane. Key findings were the influence of these fluxes on the
detection limit of the ISEs and on their selectivity coefficients [13].
Two types of theoretical model have been developed to explain
these findings and to describe the dynamic responses. The first,
the phase boundary model, relies on the assumption of local equi-
librium at the aqueous solution/membrane interfaces and on the
absence of diffusion potential within the membrane [14]. The other,

0013-4686/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2012.02.090
 M. Cuartero et al. / Electrochimica Acta 69 (2012) 152–159 153

Table 1
Composition of the membranes assayed.

Membrane Percentage (w/w) of components in membrane

DOS NPOE TCP FNDPE PVC TFMPB ␤-CD

A 66 33 1
B 60 30 1 9
C 66 33 1
D 60 30 1 9
E 66 33 1
F 60 30 1 9
G 66 33 1
H 60 30 1 9

non-equilibrium response model, makes less restrictive assump-

tions at the expense of more elaborated mathematical treatments
[15]. Both types of model have succeeded in explaining experimen-
tal findings. The approach of Radu et al. [16], which falls within the
first type, permitted us to simulate the dynamic potential response
of both ISEs and therefore the DDP response upon addition of incre-
mental concentrations of analyte to the sample solution.

2. Experimental
2.1. Apparatus and electrodes
A homemade dual channel potentiometer, a fast multiplexer,
an analog to digital converter, a PC connected in series and home-
made software were used to simultaneously monitor the potential
of any of two ISEs immersed in the same solution with respect to a
common reference electrode and the corresponding DDP response.
Fig. S1 shows the scheme of the potentiometric device.
Fluka ISE bodies and an Orion Ag/AgCl double-junction reference
electrode (Orion 90-02) containing a 10−4 M KCl solution in the
outer compartment were also used.
2.2. Reagents
All chemicals were of analytical reagent grade and Milli-Q water
was used throughout. Polyvinyl chloride (PVC) of high molecular
weight, dioctyl sebacate (DOS), 2-nitrophenyl octyl ether (NPOE),
tricresyl phosphate (TCP), 2-fluoro-2 -nitrodiphenyl ether (FNDPE),
potassium tetrakis[3,5-bis-(trifluoromethyl)phenyl]borate
(TFMPB), tetrahydrofuran (THF) and papaverine hydrochloride
were purchased from Fluka. Heptakis(2,3,5-tri-O-benzoyl)-␤-
cyclodextrin (␤-CD), procainamide, procaine, lidocaine, tetracaine,
quinidine, quinine, clomipramine, chlorpromazine and bupiva-
caine in theirs hydrochloride forms were purchased from Sigma.
The pharmaceuticals used were 2% Lidocaine injectable (B. Braun
Medical S.A., Barcelona, Spain), Biocoryl injectable (J. Uriach and
Cia. S.A., Barcelona, Spain), Anafranil (Defiante Farmaceutica,
Madeira, Portugal) and Largactil (Aventis Pharma S.A., Madrid,
Spain).
2.3. Membranes preparation and conditioning
The compositions of the membranes assayed are resumed in
Table 1. A composition of about 60% of one of four different plasti-
cizers (DOS, TCP, NPOE and FNDPE), 30% PVC, 1% ionic additive and
9% ␤-CD was used for membranes B, D, F and H. The corresponding
blank membranes (A, C, E and G) without ␤-CD were also prepared.
The membranes were prepared by dissolving 100 mg PVC, 200 mg
of plasticizer, 3 mg of TFMPB, and 30 mg of ␤-CD for membranes
containing cyclodextrin, in 3 mL of THF. This solution was poured
into a Fluka glass ring (inner diameter 28 mm, height 30 mm) on a
Fluka glass plate, and allowed to settle overnight until total evapo-
ration of THF had occurred, thus obtaining a thin plastic membrane.
Fig. 1. Individual dynamic responses of NPOE plasticized electrodes with (red line)
and without (blue line) ␤-CD, and the corresponding DDP (green line), for lidocaine
concentrations, (1) 0, (2) 1 × 10−6 , (3) 5 × 10−6 , (4) 1 × 10−5 , (5) 5 × 10−5 , (6) 1 × 10−4 ,
(7) 5 × 10−4 , and (8) 1 × 10−3 mol L−1 . (For interpretation of the references to colour
in the figure legend and in the text, the reader is referred to the web version of this
article.)
A 6-mm-diameter piece was cut out with a punch and incorporated
into an ISE electrode body containing 10−4 mol L−1 KCl as internal
filling solution. The electrodes were first conditioned by immers-
ing in water for one day. Before assaying any of the drugs tested the
electrodes were conditioned by immersing both together in 50 mL
of water, stirring the solution with a magnetic stirrer at 700 rpm and
waiting until both electrodes reached constant potential values. The
conditioning step took several minutes. When the electrode had
previously been exposed to a very lipophilic drug, the conditioning
step took longer and it was necessary to discard and renew once
the conditioning water to achieve the original baselines. When not
in use, the electrode was kept immersed in water.
2.4. Differential dynamic potentiometric measurements of drugs
Two different drug concentration steps schemes were used.
The first was a serial calibration, in which drug standard solu-
tions were added in a series of consecutive aliquots to 50 mL water,
while the second consisted of a discrete calibration in which sin-
gle additions of the corresponding standard drug solutions were
made to 50 mL of water. In the first scheme, consecutive volumes
of 5 ␮L, 20 ␮L and 25 ␮L of 10−2 mol L−1 and 20 ␮L, 25 ␮L, 200 ␮L
and 250 ␮L of 10−1 mol L−1 standard solutions of the corresponding
drug were injected with micropipettes, with a constant magnetic
stirring rate of 700 rpm, to obtain the final concentration range
of 1 × 10−6 –1 × 10−3 mol L−1 . The injections were made far from
the ISEs membranes to avoid high local transient concentrations of
the drug. In the second scheme, an analogous procedure was fol-
lowed using single additions of 5 ␮L, 25 ␮L and 50 ␮L of 10−2 M and
25 ␮L, 50 ␮L and 250 ␮L of 10−1 mol L−1 standard solutions of the
corresponding drug.
2.5. Differential dynamic potentiometric responses of
pharmaceuticals
DDP responses of pharmaceuticals were obtained by serial
calibration following the procedure described above, adding
154 M. Cuartero et al. / Electrochimica Acta 69 (2012) 152–159
Fig. 2. DDP responses for lidocaine concentrations, (1) 0, (2) 1 × 10−6 , (3) 5 × 10−6 , (4) 1 × 10−5 , (5) 5 × 10−5 , (6) 1 × 10−4 , (7) 5 × 10−4 , and (8) 1 × 10−3 mol L−1 , using
membranes with four different plasticizers: DOS (a), TCP (b), NPOE (c), and FNDPE (d).
appropriate volumes of the corresponding pharmaceutical solu-
tions to obtain the same concentration profiles as achieved with
pure drug solutions.
To determine the lidocaine content in pharmaceuticals, a sin-
gle addition of 50 ␮L of the pharmaceutical to 50 mL of water
was made and the corresponding analytical signal (obtained as
described below) from the corresponding DDP response was com-
pared with a calibration graph obtained by the discrete mode using
pure drug solutions.
3. Results and discussion
We describe the application of DDP for studying the response of
ionophore-based ISEs to organic ions, since not only the membrane
containing the ionophore but also the corresponding blank with
no ionophore provide a significant potential response. The latter
response is partially due to the relatively low standard Gibbs energy
values for the direct transfer of many organic ions from water to
plasticized PVC membranes [17]. In particular, the case of ionis-
able drugs is interesting. In this work, we chose a cyclodextrin as
ionophore because several cyclodextrin-based drug ISEs have been
proposed [18–22] and, in most of these papers, the response of the
blank membrane was not reported.
For the studies described here, we selected a benzoyl
beta-cyclodextrin derivative and the following drugs: the antiar-
rhytmics procainamide and quinidine; quinine, an antimalarial;
clomipramine, an antidepressant; the local anaesthetics bupi-
vacaine, lidocaine, procaine, and tetracaine; the antispasmodic
papaverine and chlorpromazine, an antipsychotic. These drugs con-
tain a lipophilic aromatic ring system and a nitrogen that can be
protonated to provide a cationic drug. These kinds of drug are prone
to interact with ␤-CD since they manifest recognition through
three types of interaction: conventional hydrophobic bonding,
N H. . .O and N C H. . .O, hydrogen bonding and van der Waals’
forces.
Prior to exposure of the ISEs to the different cationic drugs,
the ISEs were conditioned in water, as described in Section 2. As
in similar membrane systems [7,13], during this process a flux
of potassium ions from the inner solution to the sample solution
would occur until a steady-state concentration profile inside the
membrane is reached, as manifested by a constant electrode poten-
tial.
The DDP response obtained for lidocaine by the serial calibra-
tion procedure, using membranes containing NPOE as plasticizer,
is shown in Fig. 1 (green line). As can be seen, the DDP response
increased with the lidocaine concentration up to a concentration
value at which the response changed direction during exposure.
Above this concentration, the response decreased with increas-
ing concentrations. In order to gain insight into the origin of
this phenomenon, the dynamic responses for the individual elec-
trodes are included in Fig. 1 (red and blue lines). As can be seen,
the electrodes exhibited different dynamic response behaviours
at the lower lidocaine concentrations, 1 × 10−6 , 5 × 10−6 , 1 × 10−5
and 5 × 10−5 mol L−1 . The electrode containing ␤-CD (red line)
showed a faster response at these concentrations. The steady-state
potential was not reached for the first three concentrations with
either electrode within the time used, while it was reached for
5 × 10−5 mol L−1 with the electrode containing ␤-CD. All these facts
combined to produce a potential inversion in the corresponding
DDP.
3.1. Influence of the membrane plasticizer
The DDP responses obtained for lidocaine using membranes
containing four different plasticizers, FNDPE, DOS, NPOE and TCP,
using the serial calibration procedure, are shown in Fig. 2. As can
be seen, the lidocaine concentration at which the potential inver-
sion occurred depended on the plasticizer used. Interestingly, the
potential corresponding to the inversion was roughly the same
for the plasticizers DOS, NPOE and FNDPE. One drawback of the
membranes constructed with these three plasticizers was that
the response was not reproducible, because the initial potential
difference in water could not be restored. In contrast, the mem-
brane constructed with TCP (Fig. 2b) displayed a much lower DDP
 M. Cuartero et al. / Electrochimica Acta 69 (2012) 152–159 155

Fig. 3. DDP responses for procaine (a), lidocaine (b), quinidine (c), quinine (d), procainamide (e), papaverine (f), tetracaine (g), bupivacaine (h), clomipramine (i), and
chlorpromazine (j) at concentrations of (1) 0, (2) 1 × 10−6 , (3) 5 × 10−6 , (4) 1 × 10−5 , (5) 5 × 10−5 , (6) 1 × 10−4 , (7) 5 × 10−4 , and (8) 1 × 10−3 mol L−1 .

response but was reproducible since the same DDP responses were
obtained for consecutive measurements. We found no plausible
explanation for the different results obtained with TCP with respect
to the other plasticizers, although a possible clue could be the
results reported for the use of TCP as ionophore in the construction
of ISEs for cationic amines [23]. Taking into account the repro-
ducibility of the DDP responses, the membranes plasticized with
TCP (membranes C and D) were selected for further studies.
3.2. Differential dynamic potentiometric responses for different
drugs
Preliminary DDP experiments testing drugs of different
lipophilicity were carried out to observe the behaviour of each
drug. We found that the more lipophilic the ionic drug, the lower
the inversion concentration and the higher the inversion potential.
Therefore, all the drugs tested were grouped into three groups of
low, medium and high lipophilicity, depending on whether the
inversion concentration was 1 × 10−5 mol L−1 , 5 × 10−6 mol L−1 or
1 × 10−6 mol L−1 , respectively.
In order to know whether the different drugs are mainly in their
ionic forms, the pH of their corresponding solutions was measured.
In all cases the pH values were slightly below 5.0. The reported
pKa values of these drugs are between 7.5 and 9.5 [24]. In all the
solutions used, pH < (pKa − 2), which implies that more than 99%
of every drug is in its protonated form. In previous papers of our
group and others [25–28], it was found that the response of drug-
selective electrodes based on ion-exchangers for some of the drugs
assayed in this work and other drugs with similar pKa values, were
not affected by pH within intervals of several pH units. The pH
value of about 5 used in the present work falls within all these
intervals.
156 M. Cuartero et al. / Electrochimica Acta 69 (2012) 152–159

Fig. 4. DDP responses of lidocaine (a), procainamide (b), clomipramine (c), and chlorpromazine (d) (black curves) corresponding pharmaceuticals (blue curves) at concen-
trations of (1) 0, (2) 1 × 10−6 , (3) 5 × 10−6 , (4) 1 × 10−5 , (5) 5 × 10−5 , (6) 1 × 10−4 , (7) 5 × 10−4 , and (8) 1 × 10−3 mol L−1 . (For interpretation of the references to colour in this
figure legend, the reader is referred to the web version of this article.)

The duration of each concentration step was studied because
this parameter was seen to affect the DDP response obtained. It
was found that a relatively long exposure to the lower drug con-
centrations was necessary to observe the inversion phenomena.
A time of 17 min was selected for each of the first three concen-
tration steps and 3 min for each of the last four concentration
steps.
Fig. 3 depicts the DDP responses for all the drugs assayed in these
conditions. A striking observation was that each drug exhibited a
characteristic DDP response which differed from the rest, a char-
acteristic that can be used for qualitative purposes as a fingerprint
for the recognition of the corresponding drug.
better reproducibility was obtained if the potential inversion was
referred to the base line obtained in each experiment. In this case
the values were 31.4 ± 0.6 and 32.9 ± 1.8 mV between days and
membranes, respectively.
3.4. Pharmaceutical analysis
In order to demonstrate the use of such fingerprints for phar-
maceutical analysis, the fingerprints obtained for pure solutions of
lidocaine, procainamide, clomipramine and chlorpromazine were

3.3. Reproducibility

We proceeded to evaluate the reproducibility of the fingerprints.

Three consecutive fingerprints obtained for procaine, lidocaine
and clomipramine were recorded and evaluated (see Fig. S2 in
Supplementary Material). These three drugs were chosen to study
the reproducibility of the fingerprints because they were represen-
tative of the three levels of lipophilicity mentioned above. Very
good agreement was found between the fingerprints obtained
for each drug. The mean ± standard deviation of the inversion
potential values were 17.8 ± 0.3, 20.4 ± 0.6 and 65.8 ± 0.5 mV for
procaine, lidocaine and clomipramine, respectively, while the
mean ± standard deviation of the inversion time values were
38.5 ± 0.3, 20.9 ± 0.4, 6.1 ± 0.4 min for the same drugs.
The reproducibility of the fingerprint between-days and
between membranes was studied using lidocaine. Fingerprints
were obtained with the same membrane on different days (n = 4)
over a period of 30 days, and for different membranes (see Fig. S3
in Suplemmentary Material). The mean ± standard deviation of the
inversion potential and the inversion time were 17.3 ± 1.8 mV and
Fig. 5. Individual dynamic responses of electrodes with (red line) and without (blue
22.2 ± 1.0 min between days, respectively, and 20.8 ± 3.1 mV and
line) ␤-CD, and corresponding DDP (green line) for a 1 × 10−4 mol L−1 lidocaine con-
20.7 ± 0.2 min between membranes. Although these values indi- centration step. (For interpretation of the references to colour in this figure legend,
cated good reproducibility between days and membranes, even the reader is referred to the web version of this article.)
 M. Cuartero et al. / Electrochimica Acta 69 (2012) 152–159 157

Fig. 6. DDP response for lidocaine concentrations (1) 1 × 10−6 , (2) 5 × 10−6 , (3) 1 × 10−5 , (4) 5 × 10−5 , (5) 1 × 10−4 , and (6) 5 × 10−4 mol L−1 , using a discrete calibration mode.

compared with those obtained for pharmaceutical preparations

containing these drugs. The results obtained are shown in Fig. 4.
Identical fingerprints were obtained for the pure solutions and the
corresponding pharmaceuticals.
In order to widen the scope of DDP application to the field of
quantitative analysis, the DDP responses of lidocaine were obtained
using the other type of concentration step scheme described in
Section 2, corresponding to the discrete calibration. The individ-
ual dynamic responses of both electrodes for a 1 × 10−4 mol L−1
lidocaine concentration step, together with the corresponding DDP
response are shown in Fig. 5. As can be seen, the DDP response
obtained was a non-monotonic signal, i.e. it exhibited a change
in the direction of the signal. This type of signal has already been
reported for the dynamic response of several types of single ISEs in
the presence of interfering ions [29,30]. It is interesting to mention
that, although the individual dynamic responses of both single elec-
trodes in our system were monotonic, the DDP responses obtained
from their difference were non-monotonic as a result of the differ-
ent dynamic behaviours of the individual signals of both electrodes
used.
The corresponding DDP responses obtained for increasing con-
centrations of lidocaine are shown in Fig. 6. All the DDP responses
obtained were non-monotonic signals.
To ascertain whether these signals might be useful for quantita-
tive purposes, the difference between the potential portion under
and over the base line, in absolute values, was plotted against
Fig. 7. Calibration graph for lidocaine and analytical signal used (inserted graph).
the logarithmic concentration of lidocaine (Fig. 7). Data obtained
158 M. Cuartero et al. / Electrochimica Acta 69 (2012) 152–159

Fig. 8. Theoretical simulation of individual dynamic responses of electrodes with (red line) and without (blue line) ␤-CD, and corresponding DDP (green line) using (a) serial
calibration mode for lidocaine concentrations of (1) 0, (2) 1 × 10−6 , (3) 5 × 10−6 , (4) 1 × 10−5 , (5) 5 × 10−5 , (6) 1 × 10−4 , (7) 5 × 10−4 , and (8) 1 × 10−3 mol L−1 ; (b) discrete
calibration mode for a 1 × 10−4 M concentration step. Experimental DDP response obtained for 1 × 10−4 mol L−1 lidocaine concentration (inserted graph). (For interpretation
of the references to colour in this figure legend, the reader is referred to the web version of this article.)

were adjusted by non-linear curve fitting to the Nicolsky–Eisenman
equation [10]:
0
E = Eapp + S log(Clidocaine + LD)
where E is the analytical signal, Eapp 0 is an apparent standard
potential, S is the calibration slope and LD the limit of detection,
respectively.
As can be seen in Fig. 7, a good calibration graph is obtained.
Other combinations of the above mentioned portions of the signals
were also tested but they provided worse results.
The repeatability and reproducibility of the calibration param-
eters were studied by making successive calibration graphs with
the same membrane on the same day (n = 3), with the same mem-
brane on different days (n = 3) over a period of 20 days, and with
different membranes (n = 2). The results obtained for the limit
of detection were 8.9 × 10−6 ± 1.1 × 10−7 , 1.0 × 10−5 ± 2.2 × 10−6
and 9.6 × 10−6 ± 3.7 × 10−7 mol L−1 , respectively, and for the slopes
150.1 ± 0.2, 137.3 ± 4.5 and 148.8 ± 1.1 mV per decade of con-
centration, respectively. The linear range was from 1 × 10−5 to
5 × 10−4 mol L−1 , the highest lidocaine concentration assayed.
This new calibration method was used to determine lidocaine
in a pharmaceutical preparation, directly adding a suitable volume
of the injectable to water and recording the DDP response. This
was made in quintuplicate. Good agreement was obtained between
the mean value obtained applying our method, 200.2 ± 0.4 mg of
lidocaine per 10 mL of injectable, and the labelled value in the
pharmaceutical, 200 mg of lidocaine per 10 mL of injectable. The
proposed method could be applied to determine any of the drugs
tested in this work.
3.5. Theoretical simulation of the DDP responses
First, we simulated a DDP response with a potential inversion
by using the above model to trace the corresponding dynamic
responses of two electrodes upon addition of incremental concen-
(1) trations of the analyte to the sample. In this model the predicted
pot
dynamic response is influenced by the selectivity coefficient, KIJ .
The simulated dynamic responses together with the corre-
sponding DDP response are shown in Fig. 8a. The parameter
values used for the numerical simulation were Dorg = 10−8 cm2 s−1 ,
Daq = 10−5 cm2 s−1 , ıorg = 200 ␮m, ıaq = 100 ␮m, Corg,b = 0 for both
electrodes and CRorg = 12.4 mmol kg−1 and Kpot CJ,aq = 10−13 for
the membrane with ionophore and CRorg = 12.4 mmol kg−1 and
Kpot CJ,aq = 10−10 for the corresponding blank membrane, respec-
tively. In our case, the difference between the values of the product
Kpot CJ,aq arises from the difference between the Kpot values of the
membranes with and without ionophore.
As can be seen, a potential inversion similar to that obtained for
the membrane plasticized with TCP and with other plasticizers (see
below Fig. 2) is predicted.
We have also used the theoretical model to simulate the
non-monotonic DDP signal obtained in the discrete calibration
mode. The predicted dynamic signals of both electrodes for a
1 × 10−4 mol L−1 concentration step, together with the correspond-
ing DDP are shown in Fig. 8b. The parameters used in the simulation
are the same as those used in the previous simulation. The exper-
imental DDP response obtained for 1 × 10−4 mol L−1 lidocaine
concentration is shown in the inserted graph of Fig. 8b. Inter-
estingly, the simulation predicts the fast non-monotonic portion
obtained experimentally.
Taking into account all the results obtained, it can be concluded
that the model used nicely predicts the potential inversion and the
non-monotonic transient signals obtained in the DDP responses
using both calibration modes.

In order to explain the origin of the DDP responses obtained 4. Conclusions

under the serial and discrete calibration modes, we used the
dynamic diffusion model proposed by Radu et al. [16]. This model
is based on the boundary potential model, taking into account the
fluxes of primary and interfering ions across the membrane. This
dynamic diffusion model allows potential dynamic responses for
any sequence of primary ion concentration steps to be simulated.
The basis of this model and the mathematical numerical treatment
used in the present work to obtain the simulated dynamic response
are shown in Appendix A.
The DDP technique which is introduced here provides pecu-
liar signals when it is used with a cyclodextrin-based electrode
and its corresponding blank electrode exposed to different drugs.
The signals are useful for quantitative and qualitative analysis.
With respect to the former, the DDP provided reproducible drug-
fingerprints, while for quantitative purposes the DDP provided a
new calibration method that can be used to determine drugs in
pharmaceuticals. Both types of DDP response obtained arise from
 M. Cuartero et al. / Electrochimica Acta 69 (2012) 152–159 159

the different dynamic behaviours displayed by two electrodes,
which is corroborated by a theoretical model. It is suggested that
the technique could be extended to other systems.
Acknowledgements
We gratefully acknowledge the Ministerio de Ciencia e Inno-
vación, Subdirección General de Proyectos de Investigación, Spain
(project CTQ2011-27049). M.C.B thanks the University of Murcia
for a grant.
Appendix A. Model and mathematical treatment for
dynamic response simulation
A finite differences procedure with three-point discretization
scheme was used for the resolution of Eq. (A.1). The condition (A.2)
was discretized using five-point expressions [31]. The numerical
integration was performed for each concentration pulse using the
Crank–Nicolson method that was preceded by four extrapolation
steps as is suggested in the literature [32]. Grids of 250 points were
used in the discretization of the distances ıaq and ıorg , and each
concentration step was divided into 500 time-intervals.
The value of Caq∗ was changed to simulate each concentration
step and the potentiometric response was simulated from the con-
centration profiles generated in the previous step.
Appendix B. Supplementary data

The present model gives a solution to the diffusion equation in Supplementary data associated with this article can be found, in
the aqueous (sample) and organic (membrane) phase for a primary the online version, at doi:10.1016/j.electacta.2012.02.090.
ion using a numerical method based on finite difference discretiza-
tion in time and in space. References
The diffusion is governed by the equation
[1] E. Lindner, K. Toth, E. Pungor, Pure Appl. Chem. 58 (1968) 469.
∂Ci (x, t) ∂2 Ci (x, t) [2] A. Lewenstam, A. Hulanicki, T. Sokalski, Anal. Chem. 59 (1987) 1539.
= Di (A.1) [3] E. Lindner, K. Toth, E. Pungor, Dynamic Characteristics of Ion-selective Elec-
∂t ∂x2
trodes, CRC Press, Boca Raton, FL, 1988.
where x is the distance from the interface, t the time, Ci the con- [4] D. Calvo, A. Durán, M. Del Valle, Anal. Chim. Acta 600 (2007) 97.
centration of the primary ion in aqueous and organic phase and Di [5] M.J.D. Brand, G.A. Rechnitz, Anal. Chem. 42 (1970) 616.
[6] R.P. Buck, E. Lindner, Anal. Chem. 73 (2001) 88A.
the diffusion coefficient of the primary ion in each phase. [7] E. Bakker, E. Pretsch, Trends Anal. Chem. 24 (2005) 199.
A non-accumulative condition is imposed in x = 0: [8] E. Pretsch, Trends Anal. Chem. 26 (2007) 46.
[9] E. Bakker, V. Bhakthavatsalam, K.L. Gemene, Talanta 75 (2008) 629.
∂Caq (0, t) ∂Corg (0, t) [10] J. Bobacka, A. Ivaska, A. Lewenstam, Chem. Rev. 108 (2008) 329.
Daq = Dorg (A.2)
∂x ∂x [11] T. Sokalski, W. Kuzca, M. Danielewski, A. Lewenstam, Anal. Chem. 81 (2009)
5016.
The concentration of the primary ion at the limit of each diffusion [12] M. Pietrzak, M.E. Meyerhoff, Anal. Chem. 81 (2009) 5961.
layer (x = −ıaq and x = ıorg ) is fixed: [13] E. Bakker, P. Bühlmann, E. Pretsch, Talanta 63 (2004) 3.
[14] S. Mathinson, E. Bakker, Anal. Chem. 70 (1998) 303.
∗
C(−ıaq , t) = Caq (t) (A.3) [15] A. Lewenstam, J. Solid State Electrochem. 15 (2011) 15.
[16] A. Radu, A.J. Meir, E. Bakker, Anal. Chem. 76 (2004) 6402.
∗
C(ıorg , t) = Corg (t) (A.4) [17] A. Molina, C. Serna, J.A. Ortuño, J. González, E. Torralba, A. Gil, Anal. Chem. 81
(2009) 4220.
where Ci∗
are the fixed values for the primary ion concentrations in [18] K. Odashima, H. Hashimoto, Y. Umezawa, Microchim. Acta 113 (1994) 223.
[19] R. Kataky, S. Palmer, Electroanalysis 8 (1996) 585.
the limit of each diffusion layer. [20] A. Ferancová, J. Labuda, Fresenius J. Anal. Chem. 370 (2001) 1.
An additional condition in the sample-membrane interface is [21] Y.S. El-Saharty, F.H. Metwaly, M. Refaat, S.Z. El-Khateeb, Talanta 72 (2007) 675.
also needed. The interface condition is based on current ion- [22] C.G. Amorim, A.N. Araújo, A.N. Montenegro, V.L. Silva, J. Pharm. Biomed. Anal.
48 (2008) 1064.
exchange equilibrium across the interface and ISE selectivity theory [23] T. Katsu, D. Xu, K. Tsuji, T. Nagamatsu, Anal. Chim. Acta 354 (1997) 301.
as is suggested by Radu et al. [16]: [24] A.C. Moffat, J.V. Jackson, M.S. Moss, B. Widdop (Eds.), Clarke’s Isolation and Iden-
tification of Drugs in Pharmaceuticals, Body Fluids, and Post-Mortem Material,
pot Caq (0, t)[CRorg − Corg (0, t)] second ed., The Pharmaceutical Press, London, 1986.
KIJ (A.5)
Corg (0, t)CJaq [25] J.A. Ortuño, J. Hernández, C. Sánchez-Pedreño, Sens. Actuators B: Chem. 119
(2006) 282.
pot [26] M.S. Lonescu, A.A. Abrutis, N. Ridulescu, G.E. Baiulescu, V.V. Cosofret, Analyst
where KIJ is the selectivity coefficient (which here is equivalent
110 (1985) 929.
to an ion-exchange constant), and CJaq and CRorg are the concentra- [27] A.F. Shoukry, Y.M. Issa, H. Ibrahim, O.A. El-rashiedy, Anal. Lett. 24 (1991) 1861.
tions of interfering ion J in the aqueous phase and of ion-exchanger [28] A.A. Bouklouze, A. El-Jammal, G.J. Patriarche, G.D. Christian, J. Pharm. Biomed.
Anal. 9 (1991) 393.
in the organic phase, respectively.
[29] M. Gratzl, E. Lindler, E. Pungor, Anal. Chem. 57 (1985) 1506.
Moreover, the relative potential was calculated as in the phase [30] J.A. Ortuño, C. Sánchez-Pedreño, D. Martinez, Electroanalysis 15 (2003) 1536.
boundary model: [31] D. Britz, Digital Simulation in Electrochemistry, third ed., Springer, Berlin,
2005.
RT Caq (0, t) [32] F. Martínez-Ortiz, A. Molina, E. Laborda, Electrochim. Acta 56 (2011)
E= ln (A.6) 5707.
zF Corg (0, t)