Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma
a r t i c l e i n f o a b s t r a c t
Article history: Cloud-point extraction (CPE) draws increasing interest in a number of analytical fields including bioanal-
Received 26 May 2015 ysis, but combining CPE and LC–MS with electrospray ionization (ESI) in the determination of drugs in
Received in revised form 21 October 2015 biological fluids such as plasma, serum or blood has not been reported so far. Bisoprolol was deter-
Accepted 25 October 2015
mined in human plasma by CPE using Trition X-114 as a surfactant and metoprolol as the internal
Available online 29 October 2015
standard. NaOH concentration, temperature and Trition X-114 concentration were optimized. All anal-
yses were performed using liquid chromatography–electrospray ionization-tandem mass spectrometry
Keywords:
(LC–ESI–MS/MS). All validation experiments met international acceptance criteria and no significant
Cloud-point extraction
Pharmacokinetics matrix effect was observed. The compatibility of CPE and LC–ESI–MS/MS was confirmed using clinical
Sample preparation plasma samples and appropriate statistical tests. The determination of bisoprolol concentration in human
Bioanalysis plasma in the range 1.0–70 ng mL−1 by the CPE method leads to the results which are equivalent to those
Liquid chromatography coupled to mass obtained by the widely used liquid–liquid extraction method. The results revealed that a structural ana-
spectrometry logue may be an appropriate internal standard when CPE is used as the extraction technique. CPE offers
significant practical advantages over the classical extraction methods, including a positive impact on the
environment, therefore its wider application in future pharmacokinetic studies is justifiable.
© 2015 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.chroma.2015.10.076
0021-9673/© 2015 Elsevier B.V. All rights reserved.
40 J. Giebułtowicz et al. / J. Chromatogr. A 1423 (2015) 39–46
[16] and the only use of CPE and LC–MS combination was reported purification system (Milli-Q, Billerica, US). Blank plasma was
[3] for APCI as the ionization source. LC–ESI–MS/MS is the state- obtained from the Regional Center of Blood Donation and Treat-
of-the-art bioanalytical technique and is more frequently used in ment (Warsaw, Poland).
the bioanalysis of drugs than LC–APCI–MS/MS. However, the sur-
factants introduced to the sample may cause matrix effects and ESI 2.2. Chromatographic and mass spectrometric conditions
is more sensitive to this phenomena than APCI [17]. Therefore, as
regards further applications of CPE in the bioanalysis of drugs, it is The instrumental analysis was performed using Agilent 1260
of critical importance that the possibility of its compatibility with Infinity (Agilent Technologies, Santa Clara, CA, US), equipped with
LC–ESI–MS/MS is evaluated. a degasser, autosampler, binary pump coupled to a hybrid triple
Bisoprolol is a selective -1-adrenoceptor antagonist bereft of quadrupole/linear ion trap mass spectrometer QTRAP 4000 (AB
any partial agonist effect [18]. Taking into account the social impor- Sciex, Framingham, MA, US). The Turbo Ion Spray source was oper-
tance of bisoprolol, novel methods for its rapid, convenient and ated in the positive mode. The curtain gas, ion source gas 1, ion
environmentally safe determination in human plasma are still source gas 2 and collision gas (all high purity nitrogen) were set
needed. Previously reported methods of LC–MS/MS detection of at 345 kPa, 207 kPa, 276 kPa and “high” instrument units, respec-
bisoprolol [19–28] did not use CPE as serum or plasma prepara- tively. The ion spray voltage and source temperature were 5000 V
tion technique. The most common one was protein precipitation and 600 ◦ C, respectively. The target compounds were analyzed in
by the addition of organic solvents (e.g. acetonitrile) [20,21,24–26]. the MRM mode (Table 1).
Although this technique is rapid and simple, its disadvantages Chromatographic separation was achieved with a Symmetry
result from insufficient sample clean-up and include a shorter HPLC C18 column (150 mm × 4.6 mm, 3.5 m, Waters, Milford, US). The
column life, increased risk of significant matrix effects and con- column was maintained at 40 ◦ C at the flow rate of 0.5 mL min−1 .
tamination of a mass spectrometer [29]. The SPE procedures using The mobile phases consisted of HPLC grade water with 0.1% formic
Oasis polymeric mix sorbent cartridges were also applied for the acid as the eluent A and methanol with 0.1% formic acid as the elu-
determination of bisoprolol alone or among other beta-antagonists ent B. The gradient (%B) was as follows: 0 min. 10%; 1 min. 10%;
and beta-agonists [28]. SPE based on the hydrophilic-lipophilic bal- 8 min. 90%, 16 min. 90%. The re-equilibration of the column to the
anced sorbent beds is successfully used in the multi-analysis of initial conditions lasted 4 min. The injection volume was 10 L.
compounds of different polarity [30]. However, this technique may
be considered time consuming and rather expensive, especially if 2.3. Standard solutions, calibration standards and quality control
there is a need for analyzing large numbers of samples in a pharma- samples
cokinetic study. LLE was also successfully applied in the bioanalysis
of bisoprolol [19,22,23]. This technique is often used in bioanaly- The standard stock solutions of bisoprolol and IS (1 mg mL−1 )
sis as it is fast and inexpensive. Our previous experience confirms were prepared by dissolving an appropriate amount of the sub-
that its application allows to avoid the contamination of an MS stance in methanol. The working solutions (10 ng mL−1 ) were
ion source and extends the duration of an HPLC column [29,31,32]. prepared by diluting standard stock solutions with water. The solu-
However, LLE requires the use of organic solvents which are toxic tions were stored at 4 ◦ C. The calibration standards for bisoprolol
to laboratory staff and the environment, as well as expensive to dis- were made at the following concentrations: 0.3, 0.9, 3.0, 10, 20, 30,
pose of. CPE is an alternative technique with all the advantages of 45, 60 and 70 ng mL−1 . The quality control samples were prepared
LLE. Additionally, CPE is regarded safer and more environmentally at the following levels: 0.9, 30 and 60 ng mL−1 . The calibration stan-
friendly due to the extraction being based on surfactants which are dards and quality control samples were stored at −80 ◦ C until use.
less toxic and less expensive to dispose of [33]. An extensive search Sodium citrate was used as an anticoagulant.
did not return any reports on the application of CPE to the biso-
prolol determination nor the statistical comparison of CPE and the 2.4. Plasma samples from the pharmacokinetic study
well-established sample preparation techniques.
In this paper we introduce a novel combination of sample prepa- Plasma samples from the pharmacokinetic study in humans
ration with the separation and detection techniques. The aim of the were obtained from the Pharmaceutical Research Institute [34].
study was to assess the compatibility of CPE and LC–ESI–MS/MS and All samples were previously analyzed by LLE and the LC–MS/MS
the applicability of CPE to the bisoprolol determination in spiked method adapted from Ding et al. [35]. Briefly, plasma sample (1 mL)
human plasma and clinical samples. was mixed with the IS (metoprolol, 50 L), 1 M NaOH (0.1 mL) and
shaken for 5 min with ethyl acetate (5 mL). After centrifugation the
2. Materials and methods
organic layer was evaporated to dryness. The residue was recon-
2.1. Chemicals stituted in 60% aqueous methanol (200 L) and 30 L aliquot was
injected onto the LC–MS/MS consisting of Alliance 2695 liquid chro-
The reference standard of bisoprolol fumarate was a kind gift matograph (Waters, Milford, MA, USA) coupled to a Quattro micro
from ICN Polfa Rzeszów S.A. (Poland). Metoprolol tartrate (internal API triple quadrupole mass spectrometer equipped with an electro-
standard, IS) and Trition X-114 was purchased from Sigma-Aldrich spray ion source (Waters, Manchester, UK). The method was fully
(St. Louis, US). The solvents, HPLC gradient grade methanol, acetoni- validated according to the international guidelines [36] and all the
trile and formic acid 98% were purchased from Merck (Darmstadt, analysis were performed in accordance with the Good Laboratory
Germany). Ultrapure water was obtained from the Millipore water Practice (GLP).
Table 1
Optimized MRM conditions and ion transitions.
Substance [M + H] + ion DP [V] EP [V] Quantitative CE [V] CXP [V] Qualitative CE [V] CXP [V]
product ion product ion
2.5. Sample preparation post-extraction spiked samples. Various lots of matrix used for the
test included haemolyzed and hyperlipidaemic plasma samples.
2.5.1. CPE
50 L of the IS working solution, 300 L of 7% Triton X-114 2.7. Statistical analysis
solution and 400 L of 0.2 M NaOH were added to 250 L plasma
sample and vortexed for 5 min. Next, the sample was incubated in The statistical evaluation of the results was performed with the
a water bath at 55 ◦ C for 20 min and centrifuged at 10 000 rpm at SAS software version 9.4. for Windows. The Shapiro–Wilk test was
25 ◦ C for 5 min. Then the aqueous phase was separated by decant- used to evaluate normal distribution of QC results. The hypothesis
ing. After the CPE 700 L of acetonitrile was added to the micellar on the equality of variances was verified by the F-test (p ≥ 0.05).
phase. Alternatively to create more environmental friendly method The linear regression for bisoprolol concentration measured using
ethanol can be used in that step as well. The sample was incubated the CPE and LLE methods was calculated using the least squares
at −20 ◦ C for 20 min and centrifuged at 10 000 rpm at 25 ◦ C for method, the 95% confidence limit band of the mean response at
5 min. The supernatant (500 L) was diluted with an equal volume particular LLE points and the 95% prediction interval limits.
of water (500 L) and transferred to the vial. Ten L aliquot was
injected onto the LC–MS/MS. All CPE experiments were performed
3. Results and discussion
at the Bioanalysis and Drugs Analysis Department of the Medical
University of Warsaw.
3.1. Optimization of the CPE procedure
The method previously developed at the Pharmacology Depart- 3.1.1. Effect of the equilibrium temperature
ment of the Pharmaceutical Research Institute was fully validated The CPE is based on the separation of two phases, when the
and the stability of bisoprolol was confirmed in various conditions aqueous solutions of the surfactant are heated above the cloud-
[34]. Therefore, only partial validation was performed for this study point temperature. Theoretically, the optimal temperature for the
according to the European Medicines Agency (EMA) [37] and US extraction is 15–20 ◦ C greater than the cloud-point of the surfac-
Food and Drug Administration (FDA) [36] guidelines. Briefly, the lin- tant. With the increase of the equilibrium temperature, the volume
earity range was selected as 0.3–70 ng mL−1 as described previously of the surfactant-rich phase decreases due to the disruption of the
[34]. The accuracy and precision of the method were determined hydrogen bonds and the dehydration of the phase [16]. However,
within-run and between-run using LLOQ (0.3 ng mL−1 ) and qual- too high temperature may reduce the recovery—because of both the
ity control (QC) samples (0.9, 30 and 60 ng mL−1 ). Carry over and decomposition of the heat labile compounds as well as the thermal
recovery was also studied. instability of the surfactant aggregates [8]. The range of 40–55 ◦ C
As the aim of the study was to evaluate the compatibility of was selected for the optimization experiments. Due to a possi-
CPE and LC–ESI–MS/MS, particular attention was paid to the matrix ble analyte instability, thermal stability problems of the surfactant
effect assessment. Ion suppression or enhancement was initially aggregates and denaturation of plasma proteins it was decided
studied by the steady post-column infusion of the analyte, i.e. biso- not to perform experiments in higher temperatures. The analyti-
prolol or the IS at the concentration 60 ng mL−1 and the injection cal signal of bisoprolol increased with increasing the equilibrium
of the extracted blank plasma sample. The second approach was temperature, thus 55 ◦ C was selected as the working equilibrium
based on the method by Matuszewski et al. and allowed the quan- temperature (Fig. 1A). One can observe that standard deviation
titative assessment of the absolute and relative matrix effect by the for selected temperature was the highest one, still during method
comparison of the analyte peak areas in the standard solutions and validation acceptance criteria for precision were met.
Fig. 1. Optimization of the CPE method (n = 5). Experimental conditions: (A) 7% Triton X-114, 0.5 M NaOH; (B) extraction temperature at 55 ◦ C, 0.5 M NaOH; (C) extraction
temperature at 55 ◦ C, 7% Triton X-114.
42 J. Giebułtowicz et al. / J. Chromatogr. A 1423 (2015) 39–46
XIC of +MRM (7 pairs): 268.133/116.100 Da ID: metoprolol 8 from Sample 3 (matryca i 1' 60ng/ml bisopr i metoprolol) of 2014-04-30bisoprolol i m... Max. 3341.7 cps.
18.98
3342
0
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0
Time, min
XIC of +MRM (7 pairs): 326.259/116.000 Da ID: bisoprolol 1 from Sample 3 (matryca i 1' 60ng/ml bisopr i metoprolol) of 2014-04-30bisoprolol i met... Max. 4.8e4 cps.
OH 13.87 19.24
4.8e4 H 12.84 18.98
4.5e4 O N
18.19
4.0e4 11.77
+
3.5e4 H
3.0e4
O
Bisoprolol 11.32
17.25
Intensity, cps
15.05
2.5e4 O 11.13 16.99
5000.0
0.0
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0
Time, min
Fig. 2. Chromatograms recorded during the evaluation of the matrix effect by the steady post-column infusion and the injection of the extracted blank plasma sample for
metoprolol (internal standard, top) and bisoprolol (bottom).
These results indicate that CPE is a suitable technique for sample results. The statistical evaluation by the F-test indicated that the
preparation prior to the LC–ESI–MS/MS analysis as no significant hypothesis of the equality of variances was rejected for the low QC
matrix effect was observed. However, the chromatographic sep- (p = 0.003), as opposed to the medium and high QCs (p = 0.647 and
aration is probably crucial to reduce the matrix effect caused by p = 0.106, respectively). Therefore, one can conclude that for lower
the surfactant. Similar results of a negligible matrix effect in the concentrations, around 0.9 ng mL−1 , the precision of a bioanalytical
CPE–LC–MS method were observed by Liu et al. [3] using the atmo- method is significantly lower when CPE is applied instead of LLE.
spheric pressure chemical ionization (APCI) source. The results of To confirm that hypothesis we performed additional analysis
the validation also revealed that the structural analogue may be an using LLE method in the same laboratory, where CPE experiments
appropriate internal standard when CPE is used as the extraction were conducted. Due to instrumental differences we modified LLE
technique. method (LLE2) as described in Section 2.5.2. The results confirmed
our previous observation that in lower concentration (low QC) CPE
has significantly lower precision than LLE2 method (p = 0.042).
3.4. Comparison of the CPE and LLE methods
XIC of +MRM (7 pairs): 326.259/116.000 Da ID: bisoprolol 1 from Sample 16 (18/15) of 2014-04-17bisoprolol pacjencji.wiff (Turbo Spray) Max. 2016.7 cps.
8.36
2000
1900 Bisoprolol
1800
1700
1600
1500
1400
1300
1200
In te ns ity , c p s
1100
1000
900
800
700
600
500
400
300
200
100
0
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0
Time, min
XIC of +MRM (7 pairs): 268.133/116.100 Da ID: metoprolol 8 from Sample 16 (18/15) of 2014-04-17bisoprolol pacjencji.wiff (Turbo Spray) Max. 7.6e4 cps.
7.67
7.5e4
7.0e4 IS
6.5e4
6.0e4
5.5e4
5.0e4
4.5e4
In te ns ity , c p s
4.0e4
3.5e4
3.0e4
2.5e4
2.0e4
1.5e4
1.0e4
5000.0
0.0
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0
Time, min
Fig. 3. Chromatogram of sample obtained from healthy volunteer after 5 h of oral administration of bisoprolol fumarate. The determined concentration was 2.4 ng mL−1 . The
blue line represents the main ion transitions m/z 326 → 116 and m/z 268 → 116 for bisoprolol (top) and metoprolol (internal standard, bottom). The red line represents the
second ion transition specific m/z 326 → 222 and m/z 268 → 133 for bisoprolol and metoprolol, respectively.
J. Giebułtowicz et al. / J. Chromatogr. A 1423 (2015) 39–46 45
60 60%
y = 0.9646x + 1.2289
R² = 0.9608 40%
50
% difference
CPE result [ng/ml]
20%
40
0%
30
-20%
0 10 20 30 40 50 60
20 LLE result [ng/ml]
Fig. 5. The %difference, i.e. (CPE result—average result)/average result × 100%, plot-
10 ted against bisoprolol concentrations measured using LLE.
4. Conclusion
%difference decreased below 15%. The positive value of the mean
%difference may indicate that the application of CPE may lead to
It has been experimentally proved that cloud-point extraction
higher concentrations measured than in the case of LLE; however,
is compatible with LC–ESI–MS/MS. This novel sample prepara-
this difference seems insignificant, as in both data sets, i.e. n = 28
tion technique was successfully applied for the first time for the
and n = 22, the mean %difference was below 15%.
bisoprolol determination in human plasma. The bisoprolol plasma
concentrations in the clinical samples obtained by the CPE and LLE
methods were comparable in the range 1.0–70 ng mL−1 . The results
Table 3
Comparison of LLE and CPE results for clinical plasma samples.
indicate that further research on combining CPE and LC–MS is rec-
ommended and may lead to significant advances in the bioanalysis
Sample no. LLE result CPE result %Difference of drugs. The study provides the first report on the statistical com-
[ng/ml] [ng/ml]
parison of CPE and the conventional extraction method. CPE offers
1 1.51 2.60 27% significant practical advantages and can have a positive impact on
2 0.30 0.54 28%
the environment, therefore its wider application in future pharma-
3 0.30 0.56 30%
4 38.82 45.10 7% cokinetic studies is justifiable.
5 33.71 42.90 12%
6 33.71 36.10 3%
Conflict of interest statement
7 30.04 30.60 1%
8 3.26 4.14 12%
9 1.96 2.14 4% The authors declare no conflict of interest regarding the content
10 0.80 2.00 43% of this article.
11 24.31 23.10 −3%
12 36.44 36.30 0%
13 53.07 47.40 −6% Acknowledgments
14 43.43 47.50 4%
15 2.49 4.60 30% The authors are grateful to all investigators of the former biso-
16 6.90 7.02 1%
17 2.60 2.37 −5%
prolol bioequivalence study [34] for kindly providing their clinical
18 0.70 2.13 51% samples for the additional analysis.
19 46.44 42.60 −4%
20 42.73 47.50 5%
21 44.71 39.20 −7%
References
22 2.97 3.18 3%
23 0.89 1.30 19% [1] K. Madej, K. Persona, M. Wandas, E. Gomółka, Sequential cloud-point extraction
24 0.35 0.59 26% for toxicological screening analysis of medicaments in human plasma by high
pressure liquid chromatography with diode array detector, J. Chromatogr. A
25 13.34 15.10 6%
1312 (2013) 42–48.
26 28.11 33.00 8%
[2] K. Madej, K. Persona, Drug screening in human plasma by cloud-point extrac-
27 44.20 33.80 −13%
tion and HPLC, Cent. Eur. J. Chem. 11 (2013) 94–100.
28 37.95 36.70 −2% [3] W. Liu, K. Bi, X. Liu, J. Zhao, X. Chen, Cloud-point extraction combined with
All data n= 28 LC–MS for analysis of memantine in rat plasma, Chromatographia 69 (2009)
837–842.
% of samples meeting acceptance criteria 75%
[4] N. Gao, H. Wu, Y. Chang, X. Guo, L. Zhang, L. Du, Y. Fu, Mixed micelle cloud
mean difference 10%
point-magnetic dispersive -solid phase extraction of doxazosin and alfuzosin,
SD of difference 16%
Spectrochim. Acta A: Mol. Biomol. Spectrosc. 134 (2015) 10–16.
CPE result over 1 ng/ml n= 22 [5] R. Heydari, N.S. Elyasi, Ion-pair cloud-point extraction: a new method for the
% of samples meeting acceptance criteria 91% determination of water-soluble vitamins in plasma and urine, J. Sep. Sci. 37
(2014) 2724–2731.
Mean difference 4%
[6] F. Nazari Serenjeh, P. Hashemi, M. Safdarian, Z. Kheirollahi, Semi-automated
SD of difference 10%
cloud point extraction with cold column trapping of surfactant-rich phase for
46 J. Giebułtowicz et al. / J. Chromatogr. A 1423 (2015) 39–46
phenazopyridine determination in human serum, J. Iran. Chem. Soc. 11 (2014) [25] M. Liu, D. Zhang, Y. Sun, Y. Wang, Z. Liu, J. Gu, Liquid
733–739. chromatographic–electrospray tandem mass spectrometric determination of
[7] Z. Pourghobadi, R. Heydari, R. Pourghobadi, M. Rashidipour, Determination of bisoprolol in human plasma, Biomed. Chromatogr. 21 (2007) 508–513.
gabapentin in human plasma using simultaneous cloud point extraction and [26] S. Li, G. Liu, J. Jia, Y. Liu, C. Pan, C. Yu, Y. Cai, J. Ren, Simultaneous determina-
precolumn derivatization by HPLC, Monatsh. Chem. Chem. Mon. 144 (2013) tion of ten antiarrhythic drugs and a metabolite in human plasma by liquid
773–779. chromatography—tandem mass spectrometry, J. Chromatogr. B 847 (2007)
[8] H. Ma, J. You, Y. Liu, Cloud-point extraction combined with HPLC for deter- 174–181.
mination of larotaxel in rat plasma: a pharmacokinetic study of liposome [27] J. Bhatt, G. Subbaiah, S. Kambli, B. Shah, M. Patel, A. Saxena, A. Baliga,
formulation, J. Sep. Sci. 35 (2012) 1539–1546. S. Nigam, H. Parekh, G. Yadav, A high throughput and sensitive liquid
[9] W.-J. Zhao, W. Liu, J.-B. Chen, Z.-M. Zhou, M.-M. Yang, Use of cloud point extrac- chromatography–tandem mass spectrometry (LC–MS/MS) method for the
tion with derivatizing reagent for the extraction and determination of isoniazid, estimation of bisoprolol in human plasma using multiplexing technique, J.
J. Chromatogr. Sci. 49 (2011) 154–158. Chromatogr. B 852 (2007) 374–381.
[10] F. Han, R. Yin, X.-L. Shi, Q. Jia, H.-Z. Liu, H.-M. Yao, L. Xu, S.-M. Li, Cloud point [28] M. Josefsson, A. Sabanovic, Sample preparation on polymeric solid phase
extraction-HPLC method for determination and pharmacokinetic study of flur- extraction sorbents for liquid chromatographic-tandem mass spectrometric
biprofen in rat plasma after oral and transdermal administration, J. Chromatogr. analysis of human whole blood—a study on a number of beta-agonists and
B 868 (2008) 64–69. beta-antagonists, J. Chromatogr. A 1120 (2006) 1–12.
[11] H. Zhang, H.-K. Choi, Analysis of meloxicam by high-performance liquid chro- [29] J. Musijowski, M. Filist, P.J. Rudzki, Sensitive single quadrupole LC/MS method
matography with cloud-point extraction, Anal. Bioanal. Chem. 392 (2008) for determination of lapatinib in human plasma, Acta Pol. Pharm. 71 (2014)
947–953. p7.
[12] X. Liu, X.-H. Chen, Y.-Y. Zhang, W.-T. Liu, K.-S. Bi, Determination of arbidol in rat ˛
[30] J. Giebułtowicz, G. Nałecz-Jawecki, Occurrence of antidepressant residues in
plasma by HPLC–UV using cloud-point extraction, J. Chromatogr. B 856 (2007) the sewage-impacted Vistula and Utrata rivers and in tap water in Warsaw
273–277. (Poland), Ecotoxicol. Environ. Safe. 104 (2014) 103–109.
[13] M.D. Rukhadze, S.K. Tsagareli, N.S. Sidamonidze, V.R. Meyer, Cloud-point [31] J. Musijowski, E. Piórkowska, P.J. Rudzki, Determination of sunitinib in human
extraction for the determination of the free fraction of antiepileptic drugs in plasma using liquid chromatography coupled with mass spectrometry, J. Sep.
blood plasma and saliva, Anal. Biochem. 287 (2000) 279–283. Sci. 37 (2014) 2652–2658.
[14] A.B. Tabrizi, S. Naini, K. Parnian, S. Mohammadi, F.E. zad, S.P. Anvarian, A. Abdol- [32] M. Filist, K. Buś-Kwaśnik, H. Ksycińska, P.J. Rudzki, Simplified LC–MS/MS
lahi, Determination of triamterene in human plasma and urine after its cloud method enabling the determination of azithromycin in human plasma after
point extraction, Quím. Nova 37 (2014) 1182–1187. a low 100 mg dose administration, J. Pharm. Biomed. Anal. 100 (2014)
[15] H. Wu, G.-Y. Zhao, L.-M. Du, Determination of ofloxacin and gatifloxacin 184–189.
by mixed micelle-mediated cloud point extraction-fluorimetry combined [33] H. Filik, S.D. Çekiç, Cloud point extraction of pesticide residues, in: M.
methodology, Spectrochim. Acta, A: Mol. Biomol. Spectrosc. 75 (2010) Stoytcheva (Ed.), Pesticides in the Modern World—Trends in Pesticides Analy-
1624–1628. sis, InTech, Rijeka, Croatia, 2011.
[16] X.Y. Qin, J. Meng, X.Y. Li, J. Zhou, X.L. Sun, A.D. Wen, Determination of ven- [34] K. Buś-Kwaśnik, H. Ksycińska, A. Leś, K. Serafin-Byczak, P.J. Rudzki, J. Raszek, T.
lafaxine in human plasma by high-performance liquid chromatography using Łazowski, A. Bielak, A. Wybraniec, Bioequivalence and pharmacokinetics of two
cloud-point extraction and spectrofluorimetric detection, J. Chromatogr. B 872 10-mg bisoprolol formulations as film-coated tablets in healthy white volun-
(2008) 38–42. teers: a randomized, crossover, open-label, 2-period, single-dose, fasting study,
[17] B.K. Matuszewski, M.L. Constanzer, C.M. Chavez-Eng, Strategies for the Int. J. Clin. Pharmacol. Ther. 50 (2012) 909–919.
assessment of matrix effect in quantitative bioanalytical methods based on [35] L. Ding, X. Zhou, X. Guo, Q. Song, J. He, G. Xu, LC–ESI–MS method for the deter-
HPLC−MS/MS, Anal. Chem. 75 (2003) 3019–3030. mination of bisoprolol in human plasma, J. Pharm. Biomed. Anal. 44 (2007)
[18] G.-F. Li, K. Wang, R. Chen, H.-R. Zhao, J. Yang, Q.-S. Zheng, Simulation of the phar- 520–525.
macokinetics of bisoprolol in healthy adults and patients with impaired renal [36] FDA Guidance for Industry, Center for Drug Evaluation and Research
function using whole-body physiologically based pharmacokinetic modeling, (CDER).Bioanalytical Method Validation, Food and Drug Administration,
Acta Pharmacol. Sin. 33 (2012) 1359–1371. Rockville, MD,U.S, 2001.
[19] H. Chang, J. Li, J. Li, X. Guan, F. Sun, Z. Qian, K. Bi, G. Fan, Simulta- [37] Guideline on bioanalytical method validation, EMEA/CHMP/EWP/192217/
neous determination of amlodipine and bisoprolol in rat plasma by a liquid 2009, European Medicines Agency. Committee for Medicinal Products for
chromatography/tandem mass spectrometry method and its application in Human Use (CHMP), 2011.
pharmacokinetic study, J. Pharm. Biomed. Anal. 71 (2012) 104–110. [38] Z.-M. Zhou, D.-Y. Zhao, J. Wang, W.-J. Zhao, M.-M. Yang, Study of cloud point
[20] G.-Y. Liu, W. Wang, J.-Y. Jia, C. Lu, Y.-M. Liu, M.-Q. Zhang, Y. Liu, S.-J. Li, C. Yu, extraction and high-performance liquid chromatographic determination of iso-
Liquid chromatography tandem mass spectrometry method for determination niazid based on the formation of isonicotinylhydrazone, J. Chromatogr. A 1216
of bisoprolol in human plasma using d5-bisoprolol as the internal standard, (2009) 30–35.
Biomed. Chromatogr. 24 (2010) 574–580. [39] G. Ren, Q. Huang, J. Wu, J. Yuan, G. Yang, Z. Yan, S. Yao, Cloud point
[21] O. Gonzalez, G. Iriarte, E. Rico, N. Ferreirós, M.I. Maguregui, R.M. Alonso, R.M. extraction—HPLC method for the determination and pharmacokinetic study
Jiménez, LC–MS/MS method for the determination of several drugs used in of aristolochic acids in rat plasma after oral administration of Aristolochiae
combined cardiovascular therapy in human plasma, J. Chromatogr. B 878 Fructus, J. Chromatogr. B 953–954 (2014) 73–79.
(2010) 2685–2692. [40] J. Zhou, P. Zeng, H.H. Tu, F.Q. Wang, Development and application of
[22] G. Peste, N. Bibire, M. Apostu, A. Vlase, C. Oniscu, A New Liquid high-performance liquid chromatography for the study of ampelopsin phar-
Chromatography-Tandem Mass Spectrometry Method for Determination of macokinetics in rat plasma using cloud-point extraction, J. Sep. Sci. 34 (2011)
Bisoprolol in Human Plasma Samples, J. Biomed. Biotechnol. 2009 (2009), 160–168.
http://dx.doi.org/10.1155/2009/736327, Article ID 736327, 8 pages. [41] A. 43 Nordström, J.V. Apon, W. Uritboonthai, E.P. Go, G. Siuzdak, Surfactant-
[23] G. Peste, C. Oniscu, A. Vlase, Experimental research for determination of enhanced desorption/ionization on silicon mass spectrometry, Anal. Chem. 78
bisoprolol fumarate in human plasma samples using liquid chromatography- (2005) 272–278.
tandem mass spectrometry (LC–MS/MS) technique, Rom. Biotechnol. Lett. 15 [42] D. Gansser, Experience with a cross-validation approach, Chromatographia 55
(2010) 5141–5145. (2002) S71–S74.
[24] M.F. Tutunji, H.M. Ibrahim, M.H. Khabbas, L.F. Tutunji, Simultaneous determina- [43] M.T. Gilbert, I. Barinov-Colligon, J.R. Miksic, Cross-validation of bioana-
tion of bisoprolol and hydrochlorothiazide in human plasma by HPLC coupled lytical methods between laboratories, J. Pharm. Biomed. Anal. 13 (1995)
with tandem mass spectrometry, J. Chromatogr. B 877 (2009) 1689–1697. 385–394.
Talanta 83 (2011) 980–987
Talanta
journal homepage: www.elsevier.com/locate/talanta
a r t i c l e i n f o a b s t r a c t
Article history: The immobilization of tyrosinase onto glutaraldehyde activated streptavidine magnetic particles and
Received 22 September 2010 subsequent retention onto a magnetized carbon paste electrode for the amperometric assay of tyrosi-
Received in revised form 26 October 2010 nase inhibitors is described. Tyrosine was used as substrate as it is the first substrate in the melanogenesis
Accepted 1 November 2010
process. The sensing mode is based on monitoring the decrease of the amperometric signal correspond-
Available online 10 November 2010
ing to the electrochemical reduction of dopaquinone enzymatically generated. This current decrease is
due to the presence of inhibitors acting directly on the enzyme or inhibitors acting on the product of the
Keywords:
enzymatic reaction, i.e. dopaquinone. The methodology is designed for the evaluation of the inhibitory
Magnetic nanoparticles
Tyrosinase
potency of the most frequently used active substances in cosmetic marketed products against hyper-
Skin whitening agent pigmentation such as kojic acid, azelaic acid and benzoic acid. These compounds bind to the tyrosinase
Biosensor active center. Ascorbic acid is also investigated as it interrupts the synthesis pathway of melanin by reduc-
ing the melanin intermediate dopaquinone back to l-dopa. By comparing the obtained IC50 , under the
same experimental conditions, the order of their inhibitory potency was: kojic acid (IC50 = 3.7 × 10−6 M,
Ki = 8.6 × 10−7 M), ascorbic acid (IC50 = 1.2 × 10−5 M), benzoic acid (IC50 = 7.2 × 10−5 M, Ki = 2.0 × 10−5 M)
and azelaic acid (IC50 = 1.3 × 10−4 M, Ki = 4.2 × 10−5 M) in close agreement with literature spectrophoto-
metric inhibition data using the soluble tyrosinase.
© 2010 Elsevier B.V. All rights reserved.
0039-9140/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.talanta.2010.11.005
V.H. Sima et al. / Talanta 83 (2011) 980–987 981
chelator of the copper of the enzyme active site [6], arbutin is a Different phenolic compounds can be enzymatically oxidized
pro-drug of hydroquinone that is an alternative substrate, azelaic by tyrosinase and several substrates have been proposed in the
acid blocks the tyrosine access to the active site and ascorbic acid literature for the evaluation of the enzyme inhibitors. Dopamine
chemically reduces the dopaquinone back to l-dopa, thus avoid- was used as a substrate in the construction of an amperometric
ing dopachrome and melanin formation [7]. Also benzoic acid, a biosensor for the detection of tyrosinase inhibitors such as kojic
preservative very frequently used in cosmetic and food industry [8], acid, benzoic acid and SCN− ion [14]. Catechol, p-cresol, m-cresol,
has been demonstrated to have inhibitory properties for tyrosinase phenol and p-chlorophenol were used as enzyme substrates for
[9]. the determination of benzoic acid based on its inhibition prop-
The need for the quantification and comparative analysis at high erties, using an amperometric biosensor device [19–21]. l-Dopa
throughput of the inhibitory activity of possible new products and was used as enzyme substrate in spectrophotometric evaluation of
of the existing ones is becoming more and more of interest for the skin-whitening agents [10]. Also l-tyrosine was used as a substrate
evaluation and development of efficient inhibitors of the hyperpig- in spectrophotometric evaluation of different inhibitors [22,23].
mentation process. Cellular tests with human melanocytes were used for the trans-
Until now, several methods for the screening of skin-whitening formation rate study of the tyrosinase substrate l-dopa using a
agents have been proposed. The method most frequently applied spectrophotometric method [24].
exploited the spectrophotometric quantification of dopaquinone, Recently, biosensors based on the inhibition of tyrosinase activ-
before and after the inhibition of soluble tyrosinase [6,10]. Another ity in the presence of a monophenol or o-diphenol substrate have
method with tyrosinase in solution relied on electron spin reso- been proposed for the determination of triazine pesticides [25], car-
nance spectrometry that measured the accumulation of eumelanin bamate and organophosphate pesticides [16], diuron and atrazine
radicals in melanin culture cells [11]. In vivo studies evaluated the [26], herbicides such as alachlor, diazinon and carbaryl [27], propyl
inhibitors potency by photometry [12]. Usually in each of the cur- gallate [28], 2,4-dichlorophenol [29], fluoride [30], benzoic acid
rently developed method, the inhibitory strength of the studied [19], cinnamic acid and sorbic acid [31].
compounds was compared to that of a standard inhibitor, namely Besides choosing the adequate transducer and enzyme in the
kojic acid [6]. Spectrophotometry is not a primary choice analyti- construction of a biosensor, an important role is played by the
cal method for the quantification of an on going high ratio reaction enzyme immobilization method. The use of magnetic particles
which is the case of the enzymatic formation of dopaquinone due (beads) in biomedical and pharmaceutical sciences has increased
to the low reproducibility of the obtained results. The use of spec- significantly in recent years [32]. The masterbeads are specially
trometry has the disadvantage of the need of culture cells and the designed for allowing simple and effective immobilization of lig-
photometric one is applied to in vivo studies so it can be used just for ands such as enzyme through activation with bi-functionalized
substances that have been approved for human use and the imple- reactants [33], like glutaraldehyde. Immobilization of biomolecules
mentation of such study involves ethic problems. Thus they are not onto the surface of magnetic particles can result in a number of
suitable for rapid assays in scientific research of novel potential skin additional functionalities: on one hand the nanostructure of the
whitening agents. The use of biosensors would be an interesting and magnetic particles permits a high enzyme loading without affect-
valuable alternative for the rapid screening and comparison of the ing its natural structure and consequently its activity and on the
inhibitory potency of the existing compounds and the novel ones other hand the beads can be attracted by a magnet and be strongly
due to its relative simplicity in use and the high reproducibility of retained in close proximity to the electrode surface and read-
the analytical data. ily washed away if needed [34]. The streptavidine masterbeads
To the best of our knowledge there have been very few tyrosi- (NMPS) are mainly designed for immunological analyses [35] but
nase based biosensors fabricated for the evaluation of the inhibitory they can very well be used for enzyme immobilization. Based on
potency of kojic acid. A Clark-type oxygen electrode that monitored the IUPAC definition [36] such a configuration is strictly speaking
oxygen consumption during catechol conversion in the presence not a biosensor, i.e. a self-contained integrated device, but since it
of kojic acid inhibitor was reported [13]. Another biosensor for has the enzyme, during the assays, in close spatial contact with the
kojic acid quantification used dopamine as substrate and it mon- electrode we may consider it as a biosensor.
itored the oxygen consumption with a Clark type electrode [14]. In the present work, a novel method for the evaluation of the
There have been a couple of tyrosinase biosensors developed for potency of the tyrosinase and melanogenesis inhibitors used as
the evaluation of the inhibitory potency of benzoic acid [8,15] skin whitening agents is proposed. The biosensor was fabricated
but none of them used l-tyrosine as enzyme substrate. Tyrosinase using tyrosinase immobilized via glutaraldehyde activated strep-
based biosensors usually used phenol or its derivatives as sub- tavidine magnetic nanoparticles and retained onto a “magnetized”
strate with the enzyme being combined with an electrochemical carbon paste electrode, mCPE. The interest in using a carbon paste
transducer to sense either the oxygen consumption of the overall electrode lies in its low background current, allowing for highly
enzymatic reaction [14] or the enzymatic production of the elec- sensitive assay to be performed in the applied potential range
troactive quinone species [16]. There was one article reported of close to 0.0 V versus an Ag/AgCl reference electrode. The employed
a tyrosinase based biosensor for quantitative analysis of phenols. tyrosinase was the mushroom tyrosinase as it is highly homology
Tyrosinase was immobilized to core–shell magnetic particles and with the human enzyme. The substrate used was the l-tyrosine,
subsequently attached to the surface of a carbon paste electrode the enzyme’s natural substrate in the skin and the investigated
with the help of a permanent magnet [17]. inhibitors were kojic acid, azelaic acid, ascorbic acid, the most
Tyrosinase is widely distributed in microorganisms, animals, frequently used active substances in marketed products for hyper-
humans and plants. The enzyme extracted from the mushroom pigmentation, and benzoic acid, a well known preservative in the
Agaricus bisporus is highly homologous with the mammalian one cosmetic industry. The obtained configuration was used to moni-
making it well suited as a model for studies on melanogenesis. tor the reduction current of the enzymatically generated oxidized
Practically all studies of inhibitors conducted so far have used species of l-tyrosine, i.e. the dopaquinone, in the presence of molec-
the mushroom tyrosinase likely because of its readily commercial ular oxygen (Fig. 1). The biosensor was applied to the screening
availability [3]. Yet alternative systems for testing can be obtained of inhibitors which induced a decrease of the reduction current
like mammalian tyrosinase, melanocytic cultures, co-cultures of proportional to the inhibitor concentration and strength.
keratinocytes and melanocytes, skin culture to finally in vivo appli- In practice, the steady state current for l-tyrosine was recorded
cation to animal skin [5,18]. as I0 . The biosensor response in the case of the addition of an
982 V.H. Sima et al. / Talanta 83 (2011) 980–987
I (A)
1.5E-06
it has been demonstrated that the intensity of inhibition varied
b
considerably depending on the characteristics of the substrate [19].
1.0E-06
The biosensor response was tested comparatively both in the pres-
5.0E-07 a ence of l-tyrosine and l-dopa using 10 L of the tyrosinase-NMPS
0.0E+00
suspension (10 mg/mL) deposited onto the surface of the mCPE. It
-1.0 -0.5 0.0 0.5 1.0 1.5 was observed that, in the concentration range 4.98 × 10−6 M and
-5.0E-07 2.91 × 10−5 M, the sensitivity to l-dopa (−5.56 × 10−4 A/M) was
E (V) vs Ag/AgCl
-1.0E-06 almost 6 times higher comparing to l-tyrosine (−9.37 × 10−5 A/M)
and the response time (the time necessary for reaching 95% of the
-1.5E-06
maximum response) to l-dopa (56 s) was faster comparing to l-
Fig. 2. Repetitive CV for l-tyrosine 4.76 × 10−5 M, 0.1 M phosphate buffer pH 6.5, tyrosine (86 s). l-Tyrosine, however, was chosen as model substrate
mCPE, scan rate, a: 10 mV s−1 , b: 100 mV s−1 . for subsequent experiments. This decision was based on the fact
that, in vivo, l-tyrosine is the natural substrate of tyrosinase and
After each amperometric inhibition assay, the tyrosinase- also the starting point of melanin synthesis and its oxidation, to
NMPSs were readily removed by flushing a water burst over the l-dopa, is the rate-limiting step of the whole process [3]. The sensi-
biosensor surface. The surface of the mCPE was renewed after each tivity to l-tyrosine was considered high enough for assuring reliable
inhibition assay. results and the lag period of its first enzymatic oxidizing step to l-
dopa was sufficiently short for further readily implementation of
the testing conditions.
3. Results and discussions
3.1.3. Selection of the pH
3.1. Working conditions
The biosensor response was studied in the pH range comprised
between 5.5 and 7. It is well known that enzyme activity is highly pH
The working conditions for an optimal response to the sub-
dependent and that the optimum pH for an enzymatic assay must
strate of the tyrosinase based biosensor have been tested. Since
be determined empirically [39]. By decreasing the pH from 7.0 to
the biosensor was designed for testing inhibitors of the skin pig-
5.5, the amperometric signal increased but no stable steady-state
mentation process it was taken into account to operate as close as
plateau was obtained at pH 6.0 and 5.5. For further experiments,
possible to the natural environment of the skin tyrosinase which is
the pH chosen was 6.5. At this pH, the current and the steady-state
a transmembrane protein of melanosomes [37].
of the biosensor response were considered optimal, in agreement
with different literature data where the working pH of the tyrosi-
3.1.1. Selection of the applied potential nase based assays vary between 6.0 and 7.0. [9,14,22,23]. The
It is well known that l-tyrosine enzymatic oxidation pro- selected pH corresponded to the Sigma declared tyrosinase stabil-
duces dopaquinone and it was expected that this product could ity pH and is also close to the normal pH range of the skin which is
be detected by electroreduction at the mCPE. Likewise the elec- considered to vary between 4 and 6.5.
trochemical oxidation of l-tyrosine was supposed to generate
dopaquinone. Cyclic voltammetry (CV) performed at 10 and 3.1.4. Amount of tyrosinase-NMPS spiked onto the mCPE
100 mV/s (from 0.0 mV to 1.1 V, and back to −0.6 V) in 0.1 M phos- The amperometric response was determined as a function of l-
phate buffer pH 6.5, showed the irreversible oxidation peak of tyrosine, in the concentration range 4.98 × 10−6 and 2.91 × 10−5 M,
l-tyrosine (1 × 10−3 M and 4.76 × 10−5 M) since no reduction peak using different amounts (5 L, 10 L and 20 L) of the 10 mg/mL
was observed in the investigated potential domain (Fig. 2). Same tyrosinase-NMPS suspension spiked onto the surface of the mCPE.
experiments were performed also for l-dopa which gave an oxida- The response increased proportionally by raising the amount of
tion peak at 0.550 V but still no reduction peak was observed. The immobilized enzyme onto the surface of the electrode. The sen-
absence of a reduction current and the progressive decrease of the sitivities obtained were: −6.82 × 10−5 A/M, −9.37 × 10−4 A/M
l-tyrosine peak during multiple scanning suggested that the gen- and −1.68 × 10−4 A/M for 5, 10 and 20 L spiked suspension,
erated dopaquinone escaped rapidly from the electrode solution respectively. The repeatability of 5 L deposited beads was poor
interface likely due its high reactivity giving rise to polymer-like comparing to 20 L and 10 L deposits. The latter was selected
structures fouling the electrode surface [38]. Amperometry of l- in subsequent work since it offered good sensitivity and repeata-
tyrosine (2.44 × 10−5 M) at the tyrosinase-NMPS mCPE, thanks to bility (RSD = 2.8%, n = 3) and it consumed less beads then a 20 L
the low background current compared to CV, permitted, however, deposit. For the subsequent inhibition assays it was decided to use
to detect a reduction current at potentials near 0 V versus Ag/AgCl a diluted suspension of tyrosinase-NMPS (1.25 mg/mL) since it gave
likely corresponding to the electroreduction of some enzymatically a good response to l-tyrosine (3.33 × 10−4 M) and allowed to reduce
generated dopaquinone remaining at the electrode solution inter- substantially the cost of the assays.
face. Amperometric experiments were realized at 0, −50, −100,
−150, −200 and −250 mV. By decreasing the applied potential the 3.1.5. Selection of the substrate concentration for the inhibition
reduction current increased but the response steady-state became studies
less stable. An applied potential of −100 mV was considered opti- The opinions about the substrate concentration to be consid-
mal taking into account the magnitude of the reduction current, ered in inhibition assays are quite contradictory. Kok et al. [40]
the ratio between signal and background current, the steady-state concluded, when measuring the inhibition potency of a competi-
of the response plateau and the possible interfering species at the tive inhibitor with an acetylcholinesterase and a choline oxidase
applied potential, in agreement with other tyrosinase based biosen- biosensor, that the inhibition percentage increased by raising the
sors described in the literature [21]. substrate concentration. Therefore they worked in enzyme satu-
rated substrate conditions. Shan et al. [19] demonstrated that for
3.1.2. Selection of the substrate a tyrosinase biosensor for the determination of benzoic acid, the
To the best of our knowledge there is no tyrosinase based amper- concentration of the catechol substrate did not affect the maxi-
ometric biosensor for the evaluation of skin-whitening agents that mum inhibition percentage but it affected the sensitivity of the
984 V.H. Sima et al. / Talanta 83 (2011) 980–987
method. It must be noted, however, that the use of a high sub- Table 1
IC50 for kojic acid, benzoic acid, azelaic acid and ascorbic acid.
strate concentration would not yield sensitive inhibition responses
when the quantification of a competitive inhibitor is performed by IC50 (M) Kojic acid Benzoic acid Azelaic acid Ascorbic acid
simultaneous addition of the inhibitor and the substrate. Because IC50 1 3.66 72.0 125 12.4
in competitive inhibition the substrate competes with the inhibitor IC50 2 3.57 72.0 127 11.6
for the enzyme active site, and the inhibition, especially at low IC50 3 3.71 72.6 128 11.0
inhibitor concentrations, would likely not be detected [40]. IC50 4 3.72 70.9 119 12.2
IC50 5 3.63 70.5 127 11.7
Two different l-tyrosine concentrations were studied;
IC50average 3.66 71.6 125 11.8
3.33 × 10−4 M and 4.76 × 10−5 M, and the obtained IC50 for
kojic acid were found to be equal within the experimental error
(RSD = 2.3%). The percentage of inhibition was evaluated from The proposed biosensor was intended to test inhibitors that
the response of the uninhibited form of the enzyme versus the have two different mechanisms of action, namely: true enzyme
response of the enzyme partially inhibited. Taking into account that inhibitors which bind to the active site of the enzyme and inhibitors
our device was not designed for the quantification of inhibitors but which consume the dopaquinone intermediate of the melanin syn-
for the comparative screening and quantification of their inhibitory thesis pathway.
potency we considered that all the immobilized enzyme molecules It was checked that the NMPS mCPE (i.e. without tyrosinase)
must take part in the reaction and this could only be possible in gave no response under the selected experimental conditions and
substrate concentration corresponding to the saturation portion in the presence of the studied compounds. It was found that l-
of the activity versus substrate curve [40]. Enzyme saturated tyrosine, kojic acid, benzoic acid, azelaic acid and ascorbic acid gave
substrate concentration, namely 3.33 × 10−4 M l-tyrosine, was no amperometric signal.
then selected for further inhibition studies. Also blank experiments at the tyrosinase-NMPS-mCPE without
Since the concentration of substrate was quite high, it had to be addition of the l-tyrosine substrate were performed. No ampero-
established if the quantity of the dissolved oxygen in the solution metric response was observed for the tested inhibitors.
was not a limiting factor for a 3.33 × 10−4 M l-tyrosine concentra-
tion. After 30 min of air bubbling into the working buffer solution 3.2.1. “True” enzyme inhibitors
the biosensor response did not change, but after 30 min of nitrogen Kojic acid, benzoic acid and azelaic acid inhibit the melanin for-
bubbling the biosensor response decreased dramatically. The latter mation by competing with tyrosinase natural substrate, l-tyrosine,
was also a further indirect evidence of the generation of an elec- for the binding to the enzyme active site. As illustrated in Fig. 3a
troactive species by the immobilized tyrosinase. It was concluded typical time-dependent response was obtained at the tyrosinase-
that in the selected experimental conditions the quantity of oxy- NMPSs biosensor by l-tyrosine injection followed by stepwise
gen was not a limiting factor and that the enzymatic process was additions of kojic acid. It can be seen that the biosensor baseline
saturated by the substrate. was reached within 600 s allowing the experiment to be per-
formed. Once the steady state amperometric response of l-tyrosine
3.1.6. Michaelis–Menten curve (3.33 × 10−4 M) was observed, successive additions of inhibitor
A typical Michaelis–Menten plot and its linearization by the produced a diminution of the reduction current. This clearly indi-
Lineweaver–Burk plot (y(A−1 ) = −7.91 × 103 × (M−1 )−8.53 × 107 , cated that kojic acid interfered with the production of dopaquinone
R2 = 1.000, RSD slope = 3.3%, RSD intercept = 7.7%, n = 3) was at the electrode surface. The inhibition percentage (In %) increased
obtained for l-tyrosine between 2.49 × 10−6 M and 4.41 × 10−4 M with the inhibitor concentration. Kojic acid was tested in the con-
in the above described optimal working conditions of the tyrosinase centration range 2.50 × 10−6 to 1.24 × 10−5 M where it inhibited
based biosensor, namely: applied potential (Eapp ) = −100 mV, 0.1 M between 33% and 88% of the response to l-tyrosine. Same experi-
phosphate buffer pH 6.5, 10 L of 1.25 mg/mL tyrosinase-NMPS ments were performed for benzoic acid in the concentration range
suspension spiked onto the mCPE surface. The kinetic parameters 3.98 × 10−5 to 1.96 × 10−4 M where it inhibited between 34% and
Km app and Imax were calculated as the average of 3 consecu- 81% of the response and also for azelaic acid in the concentration
tive determinations: Km app = 9.7 × 10−5 M (RSD = 3.1%, n = 3) and range 9.90 × 10−5 to 4.76 × 10−4 M where it inhibited between 43%
Imax = −1.2 × 10−8 A (RSD = 7.5%, n = 3). and 88% of the response.
The obtained Km app values (Table 2) can be compared to The inhibition calibration curve of kojic acid is shown in
those reported in the literature for the enzyme in solution, i.e. the insert of Fig. 3. IC50 , the concentration of inhibitor which
Km app = 6.2 × 10−4 M [23] or 3.3 × 10−4 M [41] at pH 6.5, illustrating inhibited 50% of the l-tyrosine signal, was calculated from 5
the fact that the applied immobilization technique maintained the different curves by plotting 1/In(%) versus 1/inhibitor concen-
enzyme affinity for its substrate. tration (Table 1): kojic acid IC50 = 3.7 × 10−6 M, RSD = 1.6%, n = 5
(y = 4.54 × 10−8 x + 7.61 × 10−3 , R2 = 0.9999, RSD slope = 3.1%, RSD
3.2. Inhibitors assay intercept = 4.1%), benzoic acid IC50 = 7.2 × 10−5 M, RSD = 1.2%, n = 5
(y = 7.88 × 10−7 x + 8.99 × 10−3 , R2 = 0.9990, RSD slope = 5.1%, RSD
The developed biosensor was designed for testing inhibitors of intercept = 6.3%) and azelaic acid IC50 = 1.3 × 10−4 M, RSD = 3.0%,
the skin pigmentation process. When testing their potency, there n = 5 (y = 1.60 × 10−6 x + 7.76 × 10−3 , R2 = 0.9996, RSD slope = 9.3%,
are many factors which contribute to the final results: type of RSD intercept = 7.6%).
enzyme, quantity of immobilized enzyme, immobilization method, The study of the kinetics and of the mechanism of inhibition
type of substrate, substrate concentration, time of contact between of kojic acid, benzoic acid and azelaic acid was also carried out.
the enzyme, substrate and inhibitor, pH, temperature, applied The response of the tyrosinase-NMPSs biosensor to l-tyrosine was
potential, rate of solution stirring. In order to obtain a correct com- studied in the absence and in the presence of different concentra-
parison of the inhibition potency of the studied compounds, it was tions of inhibitor (Fig. 4). Table 2 data show that approximately
worked under the same experimental conditions [19]. the same maximum current was obtained but different values
As the developed device was a novel analytical tool, in order to for the apparent Michaelis–Menten constant (Km app ), determined
test its efficiency, it was applied to the study of compounds that following a Lineweaver–Burk plot, were calculated when various
have already demonstrated their activity in clinical practice or by amounts of kojic acid was added in the initial testing solution. The
using other analytical methods. same experiment was performed and the same conclusions were
V.H. Sima et al. / Talanta 83 (2011) 980–987 985
I (A)
Response 1 Kojic acid
-4.00E-09 0.025
Response 2 Kojic acid
1/In (%)
0.020
-5.00E-09 Response 3 Kojic acid
0.015 Response 4 Kojic acid
-6.00E-09 Response 5 Kojic acid
0.010 y = 5E-08x + 0.0076
2 Average
-7.00E-09 0.005 R = 0.9999 Linear (Average)
-8.00E-09 0.000
0.0E+00 2.0E+05 4.0E+05 6.0E+05
-1
1/conc (mol *L)
Fig. 3. Tyrosinase based biosensor amperometric response. l-Tyrosine 3.33 × 10−4 M, kojic acid inhibitor between 2.50 × 10−6 and 1.24 × 10−5 M, 0.1 M phosphate buffer pH
6.5, Eapp = −100 mV, 10 L of a 1.25 mg/mL tyrosinase-NMPS suspension spiked onto the mCPE. Insert: inhibition calibration curve.
0.0E+00 5.0E+04 1.0E+05 1.5E+05 2.0E+05 2.5E+05 3.0E+05 3.5E+05 4.0E+05 4.5E+05
0.00E+00
-2.00E+09 1/L-tyrosine (M-1)
-8.00E+09
1/I (A-1)
-1.60E+10
y = -44490x - 1E+08
-1.80E+10 c
R2 = 0.9994
-2.00E+10
Fig. 4. Tyrosinase based biosensor amperometric data: 1/signal versus 1/[l-tyrosine]. l-Tyrosine between 2.48 × 10−6 and 4.39 × 10−4 M, a: without kojic acid, b: in the
presence of 2.09 × 10−6 M kojic acid, c: in the presence of 4.18 × 10−6 M kojic acid, 0.1 M phosphate buffer pH 6.5, Eapp = −100 mV, 10 L of a 1.25 mg/mL tyrosinase-NMPS
suspension deposited onto the mCPE.
dropped out also for benzoic acid and azelaic acid, data shown bition constant, Ki (Fig. 5). The obtained Ki were 8.6 × 10−7 M,
in Table 2. For the three inhibitors, a different value of Imax was 2.0 × 10−5 M and 4.2 × 10−5 M for kojic acid, benzoic acid and aze-
obtained (−1.1 × 10−8 A, RSD = 9.2%, −1.2 × 10−8 A, RSD = 21.1% and laic acid, respectively.
−5.3 × 10−9 A, RSD = 5.0%, n = 3 for kojic acid, benzoic acid and aze-
laic acid respectively). This was not due to the nature of the studied 3.2.2. Inhibitors of the melanogenesis process
molecule but to the fact that the inhibitors were tested at different Another inhibitor tested was ascorbic acid. Unlike kojic acid,
periods of time after the preparation of the tyrosinase-NMPS sus- benzoic acid or azelaic acid, ascorbic acid does not interact with the
pension. However, as it can be seen from the stability tests, this enzyme. It is known to inhibit the formation of melanin by reducing
does not influence the inhibition parameters (see below). the dopaquinone back to l-dopa and thus it interrupts the melanin
Since, for the same inhibitor Imax did not change and Km app synthesis pathway.
increased proportionally with the inhibitor concentration, for The inhibition experiments were conducted in the same manner
kojic acid (y(A−1 ) = 8.13 × 10+1 x(M−1 ) + 9.93 × 10−5 , R2 = 0.9994), as in the case of true tyrosinase inhibitors. A time-dependent curve
benzoic acid (y(A−1 ) = 8.42x(M−1 ) + 6.95 × 10−5 , R2 = 0.9941) and of the tyrosinase-NMPSs biosensor response to l-tyrosine substrate
azelaic acid (y(A−1 ) = 2.04x(M−1 ) + 7.55 × 10−5 , R2 = 0.9946), a com- followed by successive additions of ascorbic acid was realized. The
petitive inhibition process was inferred for the three studied amperometric response increased as l-tyrosine was added into the
molecules. This conclusion is in agreement with literature data electrochemical cell. Then, due to successive additions of ascorbic
for kojic acid [14], azelaic acid [7] and benzoic acid [19]. From acid into the l-tyrosine solution (3.33 × 10−4 M) the reduction cur-
the primary Lineweaver–Burk data, secondary plots were gener- rent decreased clearly indicating that ascorbic acid interfered with
ated, by plotting the slopes from the primary plots versus inhibitor the reduction of dopaquinone at the electrode surface. The inhi-
concentration (Table 2) in order to determine the apparent inhi- bition percentage (In %) of ascorbic acid increased by raising its
Table 2
Kojic acid, benzoic acid and azelaic acid inhibition data.
Kojic acid Without kojic acid 2.09 × 10−6 M kojic acid 4.18 × 10−6 M kojic acid
Km (M) 9.7 × 10−5 2.7 × 10−4 4.4 × 10−4
Imax (A) −1.2 × 10−8 −1.1 × 10−8 −9.8 × 10−9
Benzoic acid Without benzoic acid 3.33 × 10−5 M benzoic acid 6.67 × 10−5 benzoic acid
Km (M) 8.2 × 10−5 3.3 × 10−4 6.4 × 10−4
Imax (A) −1.1 × 10−8 −1.5 × 10−8 −1.0 × 10−8
Azelaic acid Without azelaic acid 8.32 × 10−5 M azelaic acid 1.64 × 10−4 M azelaic acid
Km (M) 8.2 × 10−5 2.3 × 10−4 4.2 × 10−4
Imax (A) −5.0 × 10−9 −5.4 × 10−9 −5.6 × 10−9
986 V.H. Sima et al. / Talanta 83 (2011) 980–987
-2.00E+04 60.0
-3.00E+04 40.0
Slope
y = -4E+08x - 8085.5
b
-4.00E+04 R2 = 0.9978 20.0
0.0
-5.00E+04 a
0 5 10 15 20 25 30 35
y = -9E+09x - 7507.1
Time (days)
-6.00E+04 R2 = 0.9991
Fig. 5. Tyrosinase based biosensor data: slope of the data versus [In] (Table 2). ence of l-tyrosine (3.33 × 10−4 M). The biosensor retained about
a: kojic acid, b: benzoic acid, c: azelaic acid, 0.1 M phosphate buffer pH 6.5, 66% (RSD = 1.1%, n = 3) of the original response after one week, 46%
Eapp = −100 mV, 10 L of a 1.25 mg/mL tyrosinase-NMPS suspension deposited onto (RSD = 2.6%, n = 3) after the second week, 23% (RSD = 2.1%, n = 3)
the mCPE.
after the third week and 5% (RSD = 3.2%, n = 3) after four weeks
(Fig. 6). When not in use the tyrosinase-NMPSs suspension was
concentration. Ascorbic acid was tested in the concentration range maintained at 4 ◦ C. IC50 of kojic acid was determined at different
6.62 × 10−6 to 3.23 × 10−5 M where it was found to inhibit between periods of time following enzyme immobilization. Interestingly it
33% and 82% of the biosensor response. The inhibitory potency, was observed that IC50 was not affected by the loss of enzyme activ-
IC50 , of ascorbic acid was calculated: IC50average = 1.2 × 10−5 M, ity: the day of suspension preparation IC50 = 3.8 × 10−6 M, after one
RSD = 4.8%, n = 5 (y = 1.60 × 10−6 x + 7.76 × 10−3 , R2 = 0.9996, RSD week IC50 = 3.6 × 10−6 M, after two weeks IC50 = 3.7 × 10−6 M and
slope = 6.1%, RSD intercept = 7.2%) (Table 1). after three weeks IC50 = 3.7 × 10−6 M, RSD = 1.5%. After four weeks
By comparing the IC50 obtained in the same experimental the IC50 was not anymore determined because the amperometric
conditions for all the four tested inhibitors, the order of their signal to l-tyrosine was considered too small, around 1 nA, to give
inhibitory potency was: kojic acid (IC50 = 3.7 × 10−6 M), ascorbic reliable results.
acid (IC50 = 1.2 × 10−5 M), benzoic acid (IC50 = 7.2 × 10−5 M) and The IC50 for 5 consecutive determinations gave RSD values of
azelaic acid (IC50 = 1.3 × 10−4 M). 1.6%, 1.2%, 3.0% and 4.8% for kojic acid, benzoic acid, azelaic acid and
Even if the obtained values of IC50 and Ki of the studied com- ascorbic acid, respectively. The good reproducibility of the results
pounds cannot be directly compared with literature data due to can be explained by the fact that the inhibition signal was normal-
the various factors that may contribute to the final outcomes, the ized to the initial signal of the substrate for each curve. Thus, the
obtained results are in agreement with our expectations, based on final results were not influenced by the variations of substrate’s sig-
their practical utilization and on the general conclusions from the nal which can occur due to the loss of the enzyme activity or due
literature. Both the IC50 and the Ki show that kojic acid was the to random errors that can happen during biosensor’s preparation.
most potent inhibitor in comparison with ascorbic, benzoic and
azelaic acid. Kojic acid is considered the standard inhibitor due to 4. Conclusion
its high inhibitory potency [6] and is known as the most potent
active ingredient in the commercialized skin-whitening cosmetic A tyrosinase-streptavidine magnetic beads-magnetized carbon
products. Ascorbic acid is well known for its medium inhibitory paste biosensor for the screening of tyrosinase “true inhibitors”
property therefore it is found as adjuvant in most of the cosmetic and of the melanogenesis process inhibitors has been realized. The
creams for hyperpigmentation. Azelaic acid is mainly used as an biosensor operation mode is based on monitoring the decrease
anti-acne agent but also as adjuvant in skin-whitening creams of the amperometric signal of the electrochemical reduction of
due to its inhibitory properties along with other more powerful dopaquinone, enzymatically generated from l-tyrosine substrate,
inhibitors. Benzoic acid is above all an antibacterial agent but due to due to the presence of tyrosinase inhibitors that bind to the enzyme
its inhibitory characteristics it is considered a preferential preserva- active site such as kojic acid, benzoic acid or azelaic acid or due to
tive in cosmetic creams for hyperpigmentation treatments. Benzoic the presence of compounds that reduce the dopaquinone interme-
acid is generally used as a tyrosinase inhibitor in food industry. diate back to l-dopa interrupting the synthesis process like ascorbic
It should be outlined that the developed device allowed the acid. The enzyme immobilization is performed via glutaraldehyde
quantification and the characterization of inhibitors from the point activated streptavidine magnetic particles retained onto the sur-
of view of their interaction with tyrosinase and the enzymatic reac- face of a magnetized carbon paste electrode. To the best of our
tion in the presence of the natural substrate l-tyrosine. In the skin knowledge this is the first electrochemical tyrosinase based biosen-
pigmentation process these compounds may interference also in sor designed for the screening of the inhibitory potency of different
other biochemical steps having as final result a decrease in melanin skin-whitening agents that uses the enzyme’s skin natural sub-
formation. For example, azelaic acid inhibits melanin formation strate, l-tyrosine. The use of l-tyrosine as substrate in the design
through other mechanism besides its interaction with the tyrosi- of the biosensor is the main novelty of the proposed device being
nase active site. It may also interfere with DNA synthesis and the first electrochemical biosensor that mimics the two enzymatic
mitochondria activity in hyperactive and abnormal melanocytes steps of melanin formation in skin. The methodology is simple
[7]. in use, suitable for rapid analysis in the evaluation of inhibitors
potency.
3.2.3. Stability and reproducibility The obtained results are in agreement with the expected rank-
The long-term stability of the immobilized tyrosinase-NMPS ing of the inhibitory potency of the studied skin whitening agents.
was evaluated by measuring the biosensor response in the pres- The reproducibility of the results is very good and is not influenced
V.H. Sima et al. / Talanta 83 (2011) 980–987 987
neither by the loss of the enzyme activity nor by other variations [15] D. Shan, Q. Shi, D. Zhu, H. Xue, Talanta 72 (2007) 1767–1772.
that can occur during the immobilization process or which can be [16] Y.D. Tanimoto de Albuquerque, L.F. Ferreira, Anal. Chim. Acta 596 (2007)
210–221.
due to electrode sensitivity fluctuation. In this respect the use of [17] Z. Liu, Y. Liu, H. Yang, Y. Yang, G. Shen, R. Yu, Anal. Chim. Acta 533 (2005) 3–9.
magnetic particles for tyrosinase immobilization can be regarded [18] N. Smit, J. Vicanova, S. Pavel, Int. J. Mol. Sci. 10 (2009) 5326–5349.
as an interesting approach since it allows their use “on demand” [19] D. Shan, Q. Li, H. Xue, S. Cosnier, Sens. Actuators B 134 (2008) 1016–1021.
[20] S. Li, Y. Tan, P. Wang, J. Kan, Sens. Actuators B Chem. 144 (2010) 18–22.
and since the sensing electrode is not affected by an enzymatic [21] M.D. Morales, S. Morante, A. Escarpa, M.C. Gonzalez, A.J. Reviejo, J.M. Pingarron,
immobilization process. Talanta 57 (2002) 1189–1198.
Other substances acting through one of the above mentioned [22] Z.P. Zheng, K.W. Cheng, J. Chao, J. Wu, M. Wang, Food Chem. 106 (2008)
529–535.
mechanisms could be evaluated using the developed device. Tak-
[23] O. Nerya, R. Musa, S. Khatib, S. Tamir, J. Vaya, Phytochemistry 65 (2004)
ing into consideration the relative simplicity of fabrication and use 1389–1395.
of this biosensor, it may represent an interesting and valuable ana- [24] S. Okombi, D. Rival, S. Bonnet, A.M. Mariotte, E. Perrier, A. Boumendjel, Bioorg.
Med. Chem. Lett. 16 (2006) 2252–2255.
lytical tool especially for the evaluation of tyrosinase – inhibitors.
[25] L. Campanella, A. Bonanni, E. Martini, N. Todini, M. Tomassetti, Sens. Actuators
B 111–112 (2005) 505–514.
Acknowledgements [26] T.M. Anh, S.V. Dzyadevych, M.C. Van, N.J. Renault, C.N. Duc, J.M. Chovelon,
Talanta 63 (2004) 365–370.
[27] X. Wang, L. Chen, S. Xia, Z. Zh, J. Zhao, J.M. Chovelon, N.J. Renault, Int. J. Elec-
Thanks are expressed to WBI and AUF for fellowship to V.S. and trochem. Sci. 1 (2006) 55–61.
TUBIKA (Turkey) for fellowship to Z.A. [28] M.D. Morales, M.C. Gonzalez, B. Serra, A.J. Reviejo, J.M. Pingarron, Sens. Actua-
tors B 106 (2005) 572–579.
[29] L. Kong, S. Huang, Z. Yue, B. Peng, M. Li, J. Zhang, Microchim. Acta 165 (2009)
References 203–209.
[30] E. Asav, E. Yorganci, E. Akyilmaz, Talanta 78 (2009) 553–556.
[1] Y.J. Kim, H. Uyama, Cell. Mol. Life Sci. 62 (2005) 1707–1723. [31] I. Narli, S. Kiralp, L. Toppare, Anal. Chim. Acta 572 (2006) 25–31.
[2] S. Parvez, M. Kang, H.S. Chung, H. Bae, Phytother. Res. 21 (2007) 805–816. [32] B. Rittich, A. Spanova, D. Horak, M.J. Benes, L. Klesnilova, K. Petrova, A. Rybnikar,
[3] T.S. Chang, Int. J. Mol. Sci. 10 (2009) 2440–2475. Colloids Surf. B 52 (2006) 143–148.
[4] A. Rescigno, F. Sollai, B. Pisu, A. Rinaldi, E. Sanjust, J. Enzyme Inhib. Med. Chem. [33] Ademtech (Bordeaux-France), Product Specification Leaflet,
17 (2002) 207–218. http://www.ademtech.com/products.aspx?id p=23.
[5] F. Solano, S. Briganti, M. Picardo, G. Ghanem, Pigment Cell Res. 19 (2006) [34] D. Yu, O. Dominguez Renedo, B. Blankert, V. Sima, R. Sandulescu, J. Arcos, J.M.
550–571. Kauffmann, Electroanalysis 18 (2006) 1637–1642.
[6] J.M. Noh, S.Y. Kwak, H.S. Seo, J.H. Seo, B.G. Kim, Y.S. Lee, Bioorg. Med. Chem. [35] S. Patris, C. De Vriese, F. Prohoroff, E. Burgoa Calvo, J. Arcos Martinez, J.M.
Lett. 19 (2009) 5586–5589. Kauffmann, Electroanalysis 22 (2010) 41–48.
[7] S. Briganti, E. Camera, M. Picardo, Pigment Cell Res. 16 (2003) 101–110. [36] D.R. Thevenot, K. Toth, R.A. Durst, G.S. Wilson, Biosens. Bioelectron. 16 (2001)
[8] S. Li, Y. Tan, P. Wang, J. Kan, Sens. Actuators B 144 (2010) 18–22. 121–131.
[9] S.E. Stanca, I.C. Popescu, J. Mol. Catal. B Enzyme 27 (2004) 221–225. [37] A. Slominski, D.J. Tobin, S. Shibahara, J. Wortsman, Physiol. Rev. 84 (2004)
[10] S.H. Jeon, K.H. Kim, J.U. Koh, K.H. Kong, Bull. Korean Chem. Soc. 26 (2005) 1155–1228.
1135–1137. [38] E. Sanjust, G. Cecchini, F. Sollai, N. Curreli, A. Rescigno, Arch. Biochem. Biophys.
[11] K. Maeda, M. Fukuda, J. Pharmacol. Exp. Ther. 276 (1996) 2765–2769. 412 (2003) 272–278.
[12] S.S.K. Tai, C.G. Lin, M.H. Wu, T.S. Chang, Int. J. Mol. Sci. 10 (2009) 4257–4266. [39] A. Amine, H. Mohammadi, I. Bourais, G. Palleschi, Biosens. Bioelectron. 21
[13] K. Streffer, H. Kaatz, C.G. Bauer, A. Makower, T. Schulmeister, et al., Anal. Chim. (2006) 1405–1423.
Acta 362 (1998) 81–90. [40] F.N. Kok, F. Bozoglu, V. Hasirci, Biosens. Bioelectron. 17 (2002) 531–539.
[14] Y. Hasebe, K. Oshima, O. Takise, S. Uchiyama, Talanta 42 (1995) 2079–2085. [41] B. Lejczak, P. Kafarski, E. Makowiecka, Biochem. J. 242 (1987) 81–88.
Talanta 186 (2018) 279–285
Talanta
journal homepage: www.elsevier.com/locate/talanta
A R T I C LE I N FO A B S T R A C T
Keywords: We report here the fabrication of solid-contact calcium-selective electrodes (Ca2+-SCISEs) made of a poly-
Tetramethacrylate poly(3 urethane acrylate ion-selective membrane (ISM) that was covalently attached to the underlying ion-to-electron
4-ethylenedioxythiophene) transducer (solid-contact). Methacrylate-functionalized poly(3,4-ethylenedioxythiophene) (Meth-PEDOT) and
MWCNT Meth-PEDOT films containing either multiwalled carbon nanotubes (MWCNT) or carboxylated MWCNT
All-solid-state ion-selective electrode
(cMWCNT) were used as solid contacts. The solid contacts were deposited by drop-casting on screen-printed
Calcium
electrodes and characterized by cyclic voltammetry (CV), electrochemical impedance spectroscopy (EIS) and
Polyurethane membrane
potentiometry. Covalent binding between the solid contact and the ISM was obtained via photopolymerization in
order to increase the robustness of the Ca2+-SCISEs. The performance of the Ca2+-SCISEs was studied by
measuring their potentiometric response and their sensitivity to light, oxygen and carbon dioxide. Meth-PEDOT
was found to be a promising solid-contact material to develop low-cost and easy to prepare ISEs.
⁎
Corresponding author.
E-mail addresses: andrei.bratov@imb-cnm.csic.es (A. Bratov), jbobacka@abo.fi (J. Bobacka).
https://doi.org/10.1016/j.talanta.2018.04.056
Received 23 February 2018; Received in revised form 17 April 2018; Accepted 19 April 2018
Available online 25 April 2018
0039-9140/ © 2018 Elsevier B.V. All rights reserved.
C. Ocaña et al. Talanta 186 (2018) 279–285
not the first time that Meth-PEDOT and ISM matrix are copolymerized. 500 (2.0 wt%).
Rzewuska et al. reported SCISEs based on the copolymerization of ISMs were deposited on the screen-printed electrode substrates (Ag
polyacrylate-based membrane and Meth-PEDOT on a glassy carbon and Au) by applying the ISM solution with a microsyringe on the Meth-
electrode surface forming a single phase membrane matrix [36]. PEDOT, Meth-PEDOT-MWCNT and Meth-PEDOT-cMWCNT solid-con-
However, in this approach the polymerization time was 5 min and in tact layers. After the ISM deposition, the electrodes were exposed to
our previous studies we observed that when the polymerization time 365 nm UV light for 20 s using a UV Curing Light Lamp 8141 (Düren,
increases the selectivity of the SCISEs becomes worse due to the sodium Germany) to form covalent bonds between the ISMs and the acrylate
tetrakis[3,5-bis-(trifluoromethyl)phenyl]borate photobleaching [37]. groups of SCs. A scheme of the copolymerization process is shown in
The main aim of this work is electrochemical characterization of Fig. S1. Crosslinking between the PU-ISM and polyaniline functiona-
Meth-PEDOT, Meth-PEDOT-MWCNT and Meth-PEDOT-cMWCNT and lized with methacrylate groups was confirmed by FTIR-ATR analysis in
their application as ion-to-electron transducers in polyurethane (PU) previous studies [32]. Therefore, we assume that crosslinking occurs
based Ca2+-SCISEs. In this paper, we present a simple, low-cost and also between the PU-ISM and the Meth-PEDOT used in this work. The
robust method for preparing Ca2+-SCISEs, which might be beneficial thicknesses of the ISM was ca. 200 µm, as measured with a micrometer
for different fields such as ranching where they require easy to use and screw gauge (resolution: 1 µm) after the polymerization process.
semiautomatic analytical systems with very low costs per analysis. The
end-capped methacrylate groups of Meth-PEDOT functioned as reactive 2.4. Cyclic voltammetry (CV) and electrochemical impedance spectroscopy
sites for obtaining better bonding and stronger adhesion to the PU-ac- (EIS)
rylate ISM during the copolymerization process. This was expected to
enhance the durability of the Ca2+-SCISEs. Cyclic voltammograms and impedance spectra were recorded by
using a one-compartment three-electrode cell connected to the Autolab
2. Experimental potentiostat equipped with a frequency response analyzer (AUT20.
FRA2-AUTOLAB, Eco Chemie, B.V., The Netherlands). The screen-
2.1. Chemicals printed electrodes, a glassy carbon rod and a double junction Ag/AgCl/
3 M KCl//0.1 M LiOAc served as the working (WE), auxiliary (AE) and
Tetramethacrylate end-capped poly(3,4-ethylenedioxythiophene) reference electrode (RE), respectively. CVs were recorded between
solution (Meth-PEDOT, 0.5 wt% dispersion in propylene carbonate − 0.5 V and + 0.4 V in 0.1 M KNO3 at a scan rate ν = 0.01 Vs−1. The
carbonate and p-toluenesulfonate as charge compensating ion), KCl, impedance spectra were measured at the open circuit potential (Edc) in
KNO3, CaCl2, Ca(NO3)2 and carboxylated MWCNTs were obtained from 0.1 M KNO3 within the frequency range (f = 10 kHz – 0.01 Hz) using an
Sigma Aldrich. MWCNTs were obtained from DropSens (Oviedo, excitation signal amplitude ΔEac = 10 mV. The impedance spectra were
Spain). Calcium Ionophore II (ETH 129), bis(2-ethylhexyl) sebacate fitted to equivalent circuits with the ZView software (Scribner
(DOS), potassium tetrakis(p-chlorophenyl)borate (KTpClPB), ETH500 Associates Inc., USA).
and hexafluorobutyl acrylate (HFBuA) were obtained from Fluka.
Aliphatic urethane diacrylate (oligomer Ebecryl 270), cross-linker and
hexanediol diacrylate (HDDA) were from UCB Chemicals and the 2.5. Chronopotentiometric measurements
photoinitiator 2,2- dimethoxy-2-phenylacetophenone (IRG 651) from
Ciba-Geigy. All other chemicals used were analytical reagent grade and Constant-current chronopotentiograms were recorded in 0.1 M
all solutions were prepared using ultrapure deionized water with the CaCl2 with the Autolab potentiostat described above by passing a
resistivity of 18.2 MΩ cm (Milli-Q system, Millipore, Billerica, MA). constant current of ± 1 nA through the Ca2+-SCISEs for 60 s [39].
Screen-printed electrodes (SPEs) with silver ink and flash gold
(d=2 mm) prepared on PET polymer substrates were supplied by FAE 2.6. Potentiometric measurements
(FAE S.A., Spain).
The potentiometric response of the Ca2+-SCISEs was measured in a
2.2. Meth-PEDOT, Meth-PEDOT-MWCNT/cMWCNT deposition Faraday's cage with a 16-channel potentiometer (Lawson Labs, Inc.,
input impedance: 1015 Ω) and by using the double junction Ag/AgCl/
5.0 µL (2 ×2.5 µL) of Meth-PEDOT solution containing either 3 M KCl//0.1 M LiOAc as the RE. The Ca2+-SCISEs were preconditioned
0.2–0.5 wt% MWCNT or cMWCNT solution was drop-cast on the screen- under stirring in 10−3 M CaCl2 for 3 days prior to the potentiometric
printed electrodes to form SC layers of Meth-PEDOT-MWCNT and measurements. Calibration plots were obtained in CaCl2 solutions from
Meth-PEDOT-cMWCNT. After that, the electrodes were placed on a 10−7 M to 10−2 M by stirring the solutions for 2 min at each con-
heating plate at 50 °C to increase the evaporation rate of the propylene centration. The activity coefficients were calculated according to the
carbonate solvent. All further experiments were performed at room Debye-Hückel equation.
temperature. The selectivity coefficients towards Na+, K+, Mg2+, NH4+ and Li+
ions were determined at 0.1 M concentrations of interfering ions with
2.3. Ca2+-selective ISM deposition the mixed solution method recommended by IUPAC [40]. The po-
tentiometric aqueous layer tests were performed by recording the Ca2+-
The photocurable ISM composition was prepared as presented pre- SCISE potentials in 0.1 M CaCl2, 0.1 M NaCl and 0.1 M CaCl2 solutions
viously [38]. Briefly, the main polymer composition was prepared by (in this order). The CO2 and O2 sensitivities of the Ca2+-SCISEs were
mixing the aliphatic urethane diacrylate oligomer Ebecryl 270 with the obtained by measuring the potential of the Ca2+-SCISEs in a 0.1 M
reactive diluent HDDA and the photoinitiator Irgacure 651 in the ratio CaCl2 solution that was purged with pure O2, N2 and CO2 gas in the
(w/w) of 81:17:2. In the next step, 0.3 g of the main polymer compo- following sequence: N2 (1 h), O2 (1 h), N2 (1 h), CO2 (1 h), N2 (1 h). The
sition was dissolved in 0.2 ml THF and the plasticizers DOS and HFBuA, light sensitivity was measured in 0.1 M CaCl2 solution by monitoring
Ca2+-ionophore II and lipophilic salts were then added to this solution. the Ca2+-SCISE potentials as the illumination of the electrode surfaces
This ISM solution was placed in an ultrasonic bath until it was homo- changed in the following order: darkness (30 min), ambient room light
geneous and then left unstirred for several hours to allow evaporation (30 min), cold light (30 min), ambient room light (30 min) and darkness
of THF from the solution. This resulted in the following ISM composi- (30 min). The Leica CLS 150 XE cold light source (150 W, 21 V) was
tion: main polymer mixture (56.5 wt%), DOS (20.0 wt%), HFBuA directed through the electrolyte solution towards the Ca2+-SCISEs
(20.0 wt%), Ca2+-ionophore II (1.0%), KTpClPB (0.5 wt%) and ETH surfaces.
280
C. Ocaña et al. Talanta 186 (2018) 279–285
281
C. Ocaña et al. Talanta 186 (2018) 279–285
Table 1
Results of the chronopotentiometry measurements with the Ca2+-SCISEs where
R, ΔE/Δt and CL are the bulk resistance of the ISM, potential drift of the Ca2+-
SCISEs and the redox capacitance of the SC. We also show for comparison the
redox capacitances of the bare Meth-PEDOT and Meth-PEDOT-MWCNT/
cMWCNT solid-contacts (prepared on screen-printed electrodes, without ISM)
obtained for from CV (CCV) and EIS (CEIS) measurements.
wt% wt% R (MΩ) ΔE/Δt CL (µF) CCV (µF) CEIS (µF)
MWCNT cMWCNT (µV/s)
a
Value not available, because EIS data were not fitted to the equivalent
circuit.
282
C. Ocaña et al. Talanta 186 (2018) 279–285
Table 2
Reproducibility of the standard potential (E0), potentiometric slopes (S) of the
linear part of the calibration curves and the detection limit (LOD) for the Ca2+-
SCISEs (n = 6).
wt% MWCNT wt% cMWCNT E0 (mV) S (mV/pCa) LOD (M)
= − 3.3, log KCa, K = − 3.4, log KCa, Li = − 3.1 and log KCa, Mg =
2++ 2++ 2+ 2+
283
C. Ocaña et al. Talanta 186 (2018) 279–285
the illumination conditions changed from darkness, room light to cold References
light. Fig. S5 shows stable electrode potentials for the Ca2+-SCISEs
prepared with Meth-PEDOT and Meth-PEDOT containing 0.2 wt% [1] J.D. Stenger-Smith, Intrinsically electrically conducting polymers, Synth., Charact.,
cMWCNT, whereas the electrodes prepared with Meth-PEDOT-MWCNT their Appl. Progress. Polym. Sci. 23 (1998) 57–79.
[2] A.O. Patil, A.J. Heeger, F. Wudl, Optical-properties of conducting polymers, Chem.
showed a minor potential drift during the first hour after the illumi- Rev. 88 (1988) 183–200.
nation was changed from dark to room light. We conclude that the [3] B.P. Jelle, G. Hagen, Electrochemical multilayer deposition of polyaniline and
Ca2+-SCISEs based on Meth-PEDOT and Meth-PEDOT containing Prussian blue and their application in solid state electrochromic windows, J. Appl.
Electrochem. 28 (1998) 1061–1065.
cMWCNT are not light sensitive under normal measurement conditions [4] Q. Pei, G. Zuccarello, M. Ahlskog, O. Inganäs, Electrochromic and highly stable poly
in room light. (3,4-ethylenedioxythiophene) switches between opaque blue-black and transparent
sky blue, Polymer 35 (1994) 1347–1351.
[5] J.L. Reddinger, G.A. Sotzing, J.R. Reynolds, Muticoloured electrochromic polymers
derived from easily oxidized bis[2-(3,4-ethylenedioxy)thienyl]carbazoles, Chem.
3.7. Durability of the Ca2+-SCISEs
Commun. (1996) 1777–1778.
[6] Y. Qiu, L. Duan, L.D. Wang, Flexible organic light-emitting diodes with poly-3,4-
To check the quality of the membrane/conducting polymer inter- ethylenedioxythiophene as transparent anode, Chin. Sci. Bull. 47 (2002)
1979–1982.
face formed by copolymerization in prevention of water penetration
[7] L.S. Yu, S.A. Chen, High efficiency polymer light emitting diodes based on poly
Ca2+-SCISEs were kept in constant contact with 10−3 M CaCl2 solution (phenylene vinylene)s with balanced electron and hole fluxes, Synth. Met 132
during three months. After this period of time it was found that all (2002) 81–86.
electrodes except those prepared with Meth-PEDOT containing 0.2 wt% [8] K.L. Paik, N.S. Baek, H.K. Kim, J.H. Lee, Y. Lee, White light-emitting diodes from
novel silicon-based copolymers containing both electron-transport oxadiazole and
MWCNT (no response) and 0.5 wt% cMWCNT show close to Nernstian hole-transport carbazole moieties in the main chain, Macromolecules 35 (2002)
slopes of 26 ± 1 mV/pCa2+. The Meth-PEDOT-0.2% MWCNT lost it 6782–6791.
response and sensors with 0.5% cMWCNT showed sensitivity less than [9] W.H. Meyer, Polymer electrolytes for lithium-ion batteries, Adv. Mater. 10 (1998)
439.
20 mV/decade. The average E0 drift during this time was ca. 2 mV per [10] K. Gurunathan, A.V. Murugan, R. Marimuthu, U.P. Mulik, D.P. Amalnerkar,
day and detection limits remained as low as 5 × 10−6 M. These data Electrochemically synthesised conducting polymeric materials for applications to-
confirm that the cross-linking between the conductive Meth-PEDOT-SC wards technology in electronics, optoelectronics and energy storage devices, Mater.
Chem. Phys. 61 (1999) 173–191.
and the PU-ISM gives SCISEs with a relatively long lifetime compared [11] B.E. Conway, Transition from supercapacitor to battery behavior in electrochemical
with other developed SCISES which have lifetime between 2 and 6 energy-storage, J. Electrochem Soc. 138 (1991) 1539–1548.
month [48–50]. These results also show that the adhesion between the [12] A. Kros, R.J.M. Nolte, N. Sommerdijk, Conducting polymers with confined dimen-
sions: track-etch membranes for amperometric biosensor applications, Adv. Mater.
ISM and the screen-printed substrate is excellent.
14 (2002) 1779–1782.
[13] L.H. Dall'Antonia, M.E. Vidotti, S.I.C. de Torresi, R.M. Torresi, A new sensor for
ammonia determination based on polypyrrole films doped with dodecylbenzene-
4. Conclusions sulfonate (DBSA) ions, Electroanalysis 14 (2002) 1577–1586.
[14] A. Chaubey, M. Gerard, R. Singhal, V.S. Singh, B.D. Malhotra, Immobilization of
lactate dehydrogenase on electrochemically prepared polypyrrole-poly-
We report here a low-cost and simple fabrication method of dis- vinylsulphonate composite films for application to lactate biosensors, Electrochim.
posable Ca-ion sensors based on ISE with solid contact formed by Meth- Acta 46 (2000) 723–729.
PEDOT and Meth-PEDOT-MWCNT/cMWCNT conducting polymer [15] T. Lindfors, S. Ervela, A. Ivaska, Polyaniline as pH-sensitive component in plasti-
cized PVC membranes, J. Electroanal. Chem. 560 (2003) 69–78.
films. Electrochemical characterization was performed using cyclic [16] R.E. Gyurcsanyi, A.S. Nyback, K. Toth, G. Nagy, A. Ivaska, Novel polypyrrole based
voltammetry, electrochemical impedance measurements and potentio- all-solid-state potassium-selective microelectrodes, Analyst 123 (1998) 1339–1344.
metric measurements. The results of CV and EIS experiments reveal that [17] A. Michalska, A. Hulanicki, A. Lewenstam, All-solid-state potentiometric sensors for
potassium and sodium based on poly(pyrrole) solid contact, Microchem. J. 57
the Meth-PEDOT-cMWCNT has a higher redox capacitance than Meth- (1997) 59–64.
PEDOT-MWCNT and Meth-PEDOT. However, after coating the solid [18] A. Cadogan, Z. Gao, A. Lewenstam, A. Ivaska, D. Diamond, All-solid-state sodium-
contacts with a photo-cured polyurethane-based Ca2+-selective mem- selective electrode based on a calixarene ionophore in a poly(vinyl chloride)
membrane with a polypyrrole solid contact, Anal. Chem. 64 (1992) 2496–2501.
brane (ISM) the redox capacitance of the solid contact was essentially [19] A. Peramo, M.G. Urbanchek, S.A. Spanninga, L.K. Povlich, P. Cederna, D.C. Martin,
limited by the high bulk resistance of the ISM. Consequently, no sig- In situ polymerization of a conductive polymer in acellular muscle tissue constructs,
nificant improvements in potentiometric response of the Ca2+-SCISEs Tissue Eng. Part A 14 (2008) 423–432.
[20] S.M. Richardson-Burns, J.L. Hendricks, D.C. Martin, Electrochemical polymeriza-
were obtained by adding MWCNT into the conducting polymer films in
tion of conducting polymers in living neural tissue, J. Neural Eng. 4 (2007) L6–L13.
terms of sensitivity, selectivity, LOD, lifetime as well as robustness to [21] A. Elschner, S. Kirchmeyer, W. Lövenich, U. Merker, K. Reuter, PEDOT, Princ. Appl.
the influence of light, O2 and CO2. No aqueous layer formation was Intrinsically Conduct. Polym. (2011).
[22] A.S. Alshammari, M. Shkunov, S.R.P. Silva, Correlation between wetting properties
detected at the interfaces between the solid-contacts and the mem-
and electrical performance of solution processed PEDOT:PSS/CNT nano-composite
branes of the Ca2+-SCISEs by using the water layer test. Based on the thin films, Colloid Polym. Sci. 292 (2014) 661–668.
results of this work it can be concluded that Meth-PEDOT is a promising [23] T. Lindfors, R.-M. Latonen, Improved charging/discharging behavior of electro-
solid-contact material to develop low-cost and easy to prepare ISEs. polymerized nanostructured composite films of polyaniline and electrochemically
reduced graphene oxide, Carbon 69 (2014) 122–131.
[24] F. Sundfors, R. Bereczki, J. Bobacka, K. Toth, A. Ivaska, R.E. Gyurcsanyi,
Microcavity based solid-contact ion-selective microelectrodes, Electroanalysis 18
Acknowledgments (2006) 1372–1378.
[25] R.E. Gyurcsanyi, N. Rangisetty, S. Clifton, B.D. Pendley, E. Lindner, Microfabricated
ISEs: critical comparison of inherently conducting polymer and hydrogel based
This work was supported by the People Programme (Marie Curie inner contacts, Talanta 63 (2004) 89–99.
Actions) of the 7th Framework Programme of the European Union [26] J.X. Zhang, Y.X. Guo, S.J. Li, H. Xu, A solid-contact pH-selective electrode based on
(FP7/2007–2013) under REA grant agreement no. 600388 tridodecylamine as hydrogen neutral ionophore, Meas. Sci. Technol. 27 (2016) 9.
[27] P. Sjoberg, A. Määttänen, U. Vanamo, M. Novell, P. Ihalainen, F.J. Andrade, et al.,
(TECNIOspring programme) and from the Agency for Business Paper-based potentiometric ion sensors constructed on ink-jet printed gold elec-
Competitiveness of the Government of Catalonia (ACCIÓ). N.A. also trodes, Sens Actuator B-Chem. 224 (2016) 325–332.
acknowledges the Government of Russian Federation (Grant 074-U01) [28] M. Apostu, N. Bibire, G. Tantaru, M. Vieriu, A.D. Panainte, L. Agoroaei, Ion-selec-
tive membrane electrodes for the determination of heavy metals Construction
characterization and applications, Rev. Chim. 66 (2015) 657–659.
[29] N. He, T. Lindfors, Determination of water uptake of polymeric ion-selective
Appendix A. Supplementary material membranes with the coulometric karl fischer and FT-IR-attenuated total reflection
techniques, Anal. Chem. 85 (2013) 1006–1012.
Supplementary data associated with this article can be found in the [30] N. Abramova, A. Bratov, Application of photocured polymer ion selective mem-
branes for solid-state chemical sensors, Chemosensors 3 (2015) 190–199.
online version at http://dx.doi.org/10.1016/j.talanta.2018.04.056.
284
C. Ocaña et al. Talanta 186 (2018) 279–285
[31] A. Beltran, J. Artigas, C. Jimenez, R. Mas, J. Bartroli, J. Alonso, Development of Appl. Chem. 46 (1976) 127–132.
durable nitrate-selective membranes for all-solid state ISE and ISFET sensors based [41] J. Bobacka, Conducting polymer-based solid-state ion-selective electrodes,
on photocurable compositions, Electroanalysis 14 (2002) 213–220. Electroanalysis 18 (2006) 7–18.
[32] N. Abramova, J. Moral-Vico, J. Soley, C. Ocana, A. Bratov, Solid contact ion sensor [42] J. Bobacka, A. Lewenstam, A. Ivaska, Electrochemical impedance spectroscopy of
with conducting polymer layer copolymerized with the ion-selective membrane for oxidized poly(3,4-ethylenedioxythiophene) film electrodes in aqueous solutions, J.
determination of calcium in blood serum, Anal. Chim. Acta 943 (2016) 50–57. Electroanal. Chem. 489 (2000) 17–27.
[33] D.N. Reinhoudt, J.F.J. Engbersen, Z. Brzozka, H.H. van der Vlekkert, G.W.N. Honig, [43] Z. Mousavi, J. Bobacka, A. Lewenstam, A. Ivaska, Poly(3,4-ethylenediox-
H.A.J. Holterman, et al., Development of durable K+-selective chemically modified ythiophene) (PEDOT) doped with carbon nanotubes as ion-to-electron transducer in
field effect transistors with functionalized polysiloxane membranes, Anal. Chem. 66 polymer membrane-based potassium ion-selective electrodes, J. Electroanal. Chem.
(1994) 3618–3623. 633 (2009) 246–252.
[34] N. Abramova, A. Bratov, Photocurable pH-sensitive membrane for ion-selective [44] J.R. McDonald, Impedance Spectroscopy, John Wiley, New York, 1987.
field effect transistors, Talanta 81 (2010) 208–212. [45] Z. Mousavi, A. Teter, A. Lewenstam, M. Maj-Zurawska, A. Ivaska, J. Bobacka,
[35] G.S. Cha, D. Liu, M.E. Meyerhoff, H.C. Cantor, A.R. Midgley, H.D. Goldberg, et al., Comparison of multi-walled carbon nanotubes and Poly(3-octylthiophene) as ion-
Electrochemical performance, Biocompatibility, and adhesion of new polymer to-electron transducers in all-solid-state potassium ion-selective electrodes,
matrices for solid-state ion sensors, Anal. Chem. 63 (1991) 1666–1672. Electroanalysis 23 (2011) 1352–1358.
[36] A. Rzewuska, M. Wojciechowski, E. Bulska, E.A.H. Hall, K. Maksymiuk, [46] U. Schefer, D. Ammann, E. Pretsch, U. Oesch, W. Simon, Neutral carrier based
A. Michalska, Composite Polyacrylate−Poly(3,4- ethylenedioxythiophene) mem- calcium(2+)-selective electrode with detection limit in the sub-nanomolar range,
branes for improved all-solid-state ion-selective sensors, Anal. Chem. 80 (2008) Anal. Chem. 58 (1986) 2282–2285.
321–327. [47] M. Fibbioli, W.E. Morf, M. Badertscher, N.F. de Rooij, E. Pretsch, Potential drifts of
[37] A. Bratov, N. Abramova, J. Mu¤oz, C. Dominguez, S. Alegret, J. Bartroli, solid-contacted ion-selective electrodes due to zero-current ion fluxes through the
Photocurable polymer matrices for potassium-sensitive ion- selective electrode sensor membrane, Electroanalysis 12 (2000) 1286–1292.
membranes, AnalChem 67 (1995) 3589–3595. [48] M. Wagner, G. Lisak, A. Ivaska, J. Bobacka, Durable PEDOT:PSS films obtained from
[38] N. Abramova, A. Bratov, A. Ipatov, S. Levichev, A. Aris, E.M. Rodriguez, New flow- modified water-based inks for electrochemical sensors, Sens. Actuators B: Chem.
through analytical system based on ion-selective field effect transistors with opti- 181 (2013) 694–701.
mised calcium selective photocurable membrane for bovine serum analysis, Talanta [49] H. Xu, Y. Wang, Z.Y. Luo, Y.W. Pan, A miniature all-solid-state calcium electrode
113 (2013) 31–35. applied to in situ seawater measurement, Meas. Sci. Technol. 24 (2013) 6.
[39] J. Bobacka, Potential stability of all-solid-state ion-selective electrodes using con- [50] A. Matusevich, M. Pietrzak, E. Malinowska, Miniaturized F–selective all-solid-state
ducting polymers as ion-to-electron transducers, Anal. Chem. 71 (1999) potentiometric sensors with conductive polymer as an intermediate layer, Sens
4932–4937. Actuator B-Chem. 168 (2012) 62–73.
[40] G. Guilbault, R.A. Durst, M.S. Frant, H. Freiser, E.H. Hansen, T.S. Light, et al., Pure
285
Academic Press is an imprint of Elsevier
225 Wyman Street, Waltham, MA 02451, USA
525 B Street, Suite 1800, San Diego, CA 92101–4495, USA
125 London Wall, London, EC2Y 5AS, UK
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK
No part of this publication may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, recording, or any information storage and
retrieval system, without permission in writing from the publisher. Details on how to seek
permission, further information about the Publisher’s permissions policies and our
arrangements with organizations such as the Copyright Clearance Center and the Copyright
Licensing Agency, can be found at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by
the Publisher (other than as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and
experience broaden our understanding, changes in research methods, professional practices,
or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in
evaluating and using any information, methods, compounds, or experiments described
herein. In using such information or methods they should be mindful of their own safety and
the safety of others, including parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors,
assume any liability for any injury and/or damage to persons or property as a matter of
products liability, negligence or otherwise, or from any use or operation of any methods,
products, instructions, or ideas contained in the material herein.
ISBN: 978-0-12-803300-5
ISSN: 1871-5125 (Series)
Polyhedron
journal homepage: www.elsevier.com/locate/poly
a r t i c l e i n f o a b s t r a c t
Article history: Two novel iron(III) complexes involving pentadentate Schiff-base ligands, [Fe(LBr)(L1)], show temperature
Received 22 August 2014 induced incomplete spin crossover of a gradual nature. While for the L1 = NCSe coligand the transition
Accepted 6 November 2014 temperature (326 K or 317 K) lies above room temperature, the N 3 coligand causes its drop down to
Available online 21 November 2014
143 K or 140 K (two values were obtained by different models). This shift is associated with a significant
decrease of the enthalpy of the transition from about DH = 6.1 kJ mol1 to ca. DH = 1.7 kJ mol1, while the
Keywords: entropy of the transition is about DS = 19 J K1 mol1 and DS = 12 J K1 mol1, respectively. Two analo-
Spin crossover
gous complexes with Cl and NCS coligands remain high-spin over the whole temperature range.
Iron(III) complexes
Pseudohalides
Ó 2014 Elsevier Ltd. All rights reserved.
Schiff-base ligand
Ising-like model
http://dx.doi.org/10.1016/j.poly.2014.11.014
0277-5387/Ó 2014 Elsevier Ltd. All rights reserved.
C. Krüger et al. / Polyhedron 87 (2015) 194–201 195
Scheme 1. Synthetic route for the preparation of the complexes [Fe(LBr)Cl]0.5H2O (1), [Fe(LBr)N3]CH3OH (2), [Fe(LBr)NCS] (3) and [Fe(LBr)NCSe] (4) using the Schiff-base
ligand H2LBr.
1,6-diamino-4-azahexane (Scheme 1). The complexes [Fe(LBr)(- (Cl)] molecules non-covalently bridged (O–H O hydrogen bonding)
Cl)]0.5H2O (1), [Fe(LBr)(N3)]CH3OH (2), [Fe(LBr)(NCS)] (3), and by two water molecules and N–H Br (Fig. 1). The water molecules
[Fe(LBr)(NCSe)] (4) have been characterized by single-crystal X- provide hydrogen bonding of significant strength with
ray diffraction, elemental analysis, IR spectroscopy and magnetic d(O O) = 2.657(4) and 2.868(6) Å, while N–H Br contacts are
measurements. In our previous work we reported on two isostruc- much weaker: d(N Br) = 3.499(2) Å.
tural compounds with a general formula [Fe(LCl)(NCS/Se)], which The bond lengths in the chromophore of [Fe(LBr)(N3)] CH3OH (2)
exhibit rather gradual SCO near room temperature with TC = 280 correspond to the HS state of the metal center (Fig. 2): the Fe–Nam
(NCS) and 293 K (NCSe). In this study, the effect of the peripheral bonds are the longest with d(Fe–Nam) = 2.183(4) Å, while the Fe–
substitution of Cl by the Br atom is examined in terms of the crystal Nim bonds are significantly shorter (2.109(4), 2.117(4) Å) and the
packing and intermolecular interactions, and magnetic properties Fe–Nterm bond is the shortest one (2.048(3) Å). The Fe–O bond
of the compounds lie in the center of interest. lengths do not differ significantly from those observed for the LS
structure of 1 and they amount to 1.916(3) and 1.968(3) Å.
2. Results and discussion Remarkably, the dihedral angle (previously correlated to the occur-
rence of SCO in Fe(III) compounds with hexadentate ligands [41])
2.1. Structural data between the least-square planes of the aromatic rings is very nar-
row (a = 35.0°) in comparison with the other compounds reported
Cell parameters of the investigated compounds are listed in in this article (Table 2). Also the angular distortion parameter
Table 1; the remaining crystallographic data are deposited in Sup- (R(2) = 72.3°) is very large for the HS structure of this type of com-
plementary material and they can be retrieved from the Cambridge pounds (typical values for [Fe(X-L5)L1] HS compounds are close to
Crystallographic Data Centre. 60°).
The molecular structures of the compounds 1–4 are very similar The structure of 2 is formed of the [Fe(LBr)(N3)] molecules aligned
to the structures of the [Fe(LX)(L1)] compounds (where L1 = Cl, N in supramolecular chains and held by the N–H N hydrogen bonds
3,
NCS, NCSe; X marks the substitutions by functional groups on between the nitrogen atoms from the amine and azido groups of the
the aromatic rings of the primary ligand H2L = N,N0 -bis(1- adjacent complex molecules with d(N N) = 3.036(6) Å. The crystal
hydroxy-2-benzyliden)-1,6-diamino-4-azahexane) reported previ- structure contains methanol molecules coupled with the [Fe(LBr)
ously [9–11,14]. The pentadentate Schiff base ligand (LBr)2 chelates (N3)] molecules by a relatively short hydrogen bond between the
the central iron(III) atom by three nitrogen atoms (one amine methanol oxygen atom and phenolato group of the complex mole-
(Nam), two imine (Nim)) and two oxygen atoms. cule (Fig. 2): d(O O) = 2.781(6) Å.
The nitrogen atoms are arranged in a fac manner while the oxy- The compound [Fe(LBr)(NCS)] (3) crystallizes in the monoclinic
gen atoms are in the cis positions. The sixth coordination site of the space group Pn. The chromophore bond lengths and angular distor-
central atom is occupied by the monodentate chlorido or pseud- tion parameter (R(3) = 56.3°) refer to the HS state (Table 2).
ohalido ligand with the donor atom (Cl or Nterm) in the trans posi- The crystal structure is formed of the chains of the [Fe(LBr)(NCS)]
tion to the oxygen atom from the more flexible ‘‘propyl’’ part of the molecules interconnected by weak NH S hydrogen bonds between
pentadentate ligand. The metal–ligand bond lengths of the com- the nitrogen atoms from the amine groups and sulfur atoms of the
pounds under study are summarized in Table 2. thiocyanato ligand from the adjacent complex molecules
The compound [Fe(LBr)(Cl)]0.5H2O (1) is isostructural with the (d(N S) = 3.323(7) Å). The thiocyanato sulfur atoms are further
previously reported [Fe(LCl)(Cl)]0.5H2O and [Fe(LCl)(CN)]0.5H2O involved in other non-covalent interactions – the supramolecular
complexes [9]. The asymmetric unit contains two complex mole- [Fe(LBr)(NCS)]n chains are mutually interlocked by weak Br S
cules and one molecule of the crystal water. The chromophore non-covalent contacts (Fig. 3) with d(BrS) = 3.481(3) and
bond lengths and angular distortion parameter, R,1 correspond to 3.564(4) Å (sum of the van der Waals radii of the Br S pair is
the HS state of the complex molecule (R(1) = 62.3 and 52.3°) and 3.650 Å).
they are similar to those observed for other iron(III) chlorido com- The bond lengths (Fig. 4) within the chromophore of [Fe(LBr)
plexes with pentadentate Schiff base ligands with the longest Fe– (NCSe)] (4) at T = 193 K are close to the typical LS values with the
Cl bond (2.3458(6) and 2.3569(6) Å). The Fe–Nam bond lengths are longest bond between the iron atom and the amine nitrogen atom:
a bit shorter (2.178(2) and 2.198(3) Å) and the Fe–Nim and Fe–O d(Fe–Nam) = 2.013(7) Å. The Fe–Nim, Fe–Nterm and Fe–O bond
are the shortest bond lengths (Fig. 1 and Table 2). The crystal water lengths are significantly shorter: d(Fe–Nim) = 1.937(7), 1.967(7);
is involved in the hydrogen bonding pattern including four [Fe(LBr) d(Fe–Nterm) = 1.944(6); d(Fe–O) = 1.896(6), 1.872(5) Å. The N–Fe–
N and N–Fe–O bond angles of the trans bonds in the chromophore
are close to linearity and their values range from 174.9° to 177.6°.
1
The angular distortion from the ideal octahedron R is calculated as sum of the Also, the cis chromophore angles adopt values close to the
absolute values of the deviations of the twelve cis chromophore angles from the ideal
value 90°.
right angle (ideal octahedron): from 84.4° to 94.4°. This causes a
196 C. Krüger et al. / Polyhedron 87 (2015) 194–201
Table 1
Crystallographic data for compounds 1–4.
1 2 3 4
Br Br Br
[Fe(L )Cl]0.5H2O [Fe(L )(N3)]CH3OH [Fe(L )(NCS)] [Fe(LBr)(NCSe)]
M/g mol1 581.49 611.09 595.11 642.01
T/K 90(1) 293(1) 100(1) 193(1)
Crystal system triclinic monoclinic monoclinic monoclinic
Space group P1 P21/c Pn P21/c
a (Å) 12.1643(7) 12.6077(3) 12.1647(6) 8.31180(10)
b (Å) 12.9750(7) 15.7666(4) 7.9347(4) 12.99640(10)
c (Å) 14.8456(8) 12.7549(4) 12.4122(6) 20.23030(10)
a (°) 77.065(3) 90 90 90
b (°) 72.484(2) 112.623(3) 115.540(2) 97.503(3)
c (°) 66.470(2) 90 90 90
V (Å)3 2034.22(19) 2340.34(11) 1081.00(9) 2166.64(3)
Z 2 4 2 4
Rint/R1/wR2(I > 2r(I)) 0.0272/0.0259/0.0613 0.0281/0.0500/0.1102 0.0717/0.0472/0.1257 0.0629/0.0782/0.2352
CCD deposit number 998903 998904 998905 998906
Table 2
Selected structural parameters for the title compounds.
Compound Behavior Fe–Nam (Å) Fe–X (Å)a Fe–Nim (Å)b Fe–O (Å)c Rd(°) ae(°) Ref.
Br f f
[Fe(L )(Cl)]0.5H2O (1) HS 2.188 2.352 2.088 1.959 62.3/52.3 63.1/54.2 This work
[Fe(LBr)(NCS)] (3) HS 2.207 2.111 2.099 1.927 56.3 67.7 This work
[Fe(LBr)(N3)]CH3OH (2) SCO 2.183 2.048 2.113 1.942 72.3 35.0 This work
[Fe(LBr)(NCSe)] (4) SCO 2.013 1.944 1.952 1.884 27.8 85.9 This work
[Fe(LCl)(Cl)]0.25H2O HS 2.199f 2.361f 2.095 1.958 66.9/53.1 62.2/53.4 [9]
[Fe(LCl)(NCO)] HS 2.205f 1.996f 2.093 1.954 55.4/57.3 69.0/61.6 [9]
[Fe(LCl)(NCS)] SCO 2.000 1.944 1.932 1.879 27.1 84.5 [9]
[Fe(LCl)(NCSe)] SCO 2.001 1.952 1.932 1.881 26.9 85.5 [9]
[Fe(LCl)(CN)]H2O LS 2.012f 1.959f 1.934 1.903 19.3/28.8 65.9/56.0 [9]
[Fe(L)(CN)]CH3OH LS 2.000 1.962 1.922 1.880 24.6 76.0 [10]
[Fe(3tBu5Me-L)(NCS)] HS 2.191 2.100 2.096 1.916 56.0 75.6 [11]
[Fe(3MeO-L)(NCO)]CH3OH HS 2.218f 2.081f 2.078 1.946 54.0f 73.3/82.8 [11]
[Fe(3MeO-L)(N3)] HS 2.211 2.085 2.085 1.936 57.4 58.1 [11]
[Fe(3MeO-L)(CN)]CH3OH LS 2.028f 1.971f 1.935 1.904 26.0f 69.1/71.5 [11]
[Fe(3EtO-L)(CN)]H2O LS 2.010 1.975 1.944 1.904 20.8 66.4 [11]
a
The bond length between the iron atom and the donor atom of the monodentate ligand.
b
The average value calculated from the iron-imino nitrogen bond lengths.
c
The average value calculated from the Fe–O bond lengths.
d
The angular distortion from the ideal octahedron R is calculated as sum of the absolute values of the deviations of the twelve cis chromophore angles from the ideal value
90°.
e
The dihedral angle (a) between the least-square planes of the aromatic rings.
f
The average value calculated from two bond lengths/parameters. Ligand abbreviations: H2L = N,N0 -bis(1-hydroxy-2-benzyliden)-1,6-diamino-4-azahexane, H2LBr =
N,N0 -bis(1-hydroxy-4-bromo-2-benzyliden)-1,6-diamino-4-azahexane, H2LCl = N,N0 -bis(1-hydroxy-4-chloro-2-benzyliden)-1,6-diamino-4-azahexane, H23MeO-L = N,N0 -
bis(1-hydroxy-6-methoxy-2-benzyliden)-1,6-diamino-4-azahexane, H23EtO-L = N,N0 -bis(1-hydroxy-6-ethoxy-2-benzyliden)-1,6-diamino-4-azahexane, H23tBu5Me-L = N,N0 -
bis(1-hydroxy-4-methyl-6-tert-buthyl-2-benzyliden)-1,6-diamino-4-azahexane.
Fig. 1. Molecular structure of 1 (ORTEP drawing with thermal ellipsoids at the 50% probability level). Selected bond lengths (Å): Fe1–O2 = 1.8986(19), Fe1–O1 = 2.0246(18),
Fe1–N3 = 2.087(2), Fe1–N1 = 2.091(3), Fe1–N2 = 2.178(2), Fe2–O4 = 1.9097(19), Fe2–O3 = 2.0012(18), Fe2–N4 2.081(2), Fe2–N6 = 2.094(2), Fe2–N5 = 2.198(3), Fe2–
Cl2 = 2.3569(6). Right – the crystal packing in 1. Hydrogen atoms are omitted for clarity except for those involved in the hydrogen bonds (dashed lines).
low-value of the spin dependent parameter of the angular distor- consists of supramolecular dimers linked by rather weak non-
tion from the ideal octahedron: R(4) = 27.8°. The compound 4 is covalent N–H Br interactions with d(N Br) = 3.506(6) Å.
isostructural with [Fe(LCl)(NCR)] complexes [9] (R = S and Se) and This alignment is supported by p–p stacking interactions of the
C. Krüger et al. / Polyhedron 87 (2015) 194–201 197
Fig. 2. Molecular structure of 2 (ORTEP drawing with thermal ellipsoids at the 50% probability level. Selected bond lengths (Å): Fe1–N1 = 2.117(4), Fe1–N3 = 2.109(4), Fe1–
N2 = 2.183(4), Fe1–N4 = 2.048(3), Fe1–O1 = 1.968(3), Fe1–O2 = 1.916(3). Right – the crystal packing in 2. Hydrogen atoms are omitted for clarity except for those involved in
the hydrogen bonds (dashed lines).
Fig. 3. The supramolecular chains [Fe(LBr)(NCS)]n in 3. Right: Interconnection of the chain substructures by Br S non-covalent contacts (dashed lines) in 3. Hydrogen atoms
are omitted for clarity (except for those involved in the hydrogen bonds). Selected bond lengths (Å): Fe1–N1 = 2.099(5), Fe1–N2 = 2.207(9), Fe1–N3 = 2.099(5), Fe1–
N4 = 2.111(7), Fe1–O1 = 1.905(7), Fe1–O2 = 1.948(5).
Fig. 4. Molecular structure of 4 (ORTEP drawing with thermal ellipsoids at the 50% probability level). Right: The crystal packing in 4. Hydrogen atoms are omitted for clarity
except for those involved in the hydrogen bonds (dashed lines). Selected bond lengths (Å) in 4: Fe1–O1 = 1.896(6), Fe1–O2 = 1.872(5), Fe1–N1 = 1.937(7), Fe1–N4 = 1.944(6),
Fe1–N3 = 1.967(7), Fe1–N2 = 2.013(7).
neighboring complex molecules with the centroid–centroid dis- coupled dimers combined with the zero field splitting (abbr. ZFS)
tance of 3.700(5) Å and by short C Br contacts with was employed in the data analysis (for more details, see ESI)
d(C Br) = 3.474(8) Å (sum of the van der Waals radii of the C Br [42]. The temperature dependence of the magnetic susceptibility
pair is 3.650 Å, Fig. 4). and the field dependence of the magnetization (ESI, Fig. S1) are fit-
ted simultaneously thus allowing more reliable determination of
2.2. Magnetic data the sign and value of independent parameters, namely the isotro-
pic exchange coupling constant J, the axial zero-field splitting
The magnetic functions of complexes 1–4 are shown in Fig. 5. parameter D, the isotropic gyromagnetic factor g, the Weiss con-
The complex 1, [Fe(LBr)Cl]2 H2O, is high-spin in the whole temper- stant H and the temperature-independent susceptibility aTIM
ature interval (S = 5/2), however, the structural data show that (not to be confused with the dihedral angle where no lower index
there are two units bridged by a rather strong hydrogen bond (vide is used). Their optimum values were found as: J/hc = 0.145 cm1,
supra) therefore the value of the magnetic moment is higher than g = 1.984, aTIM = 4.76 109 m3 mol1, D/hc = 1.02 cm1,
in the case of the monomeric complex 2. The model of exchange H = 0.130 K; the corresponding discrepancy factors are
198 C. Krüger et al. / Polyhedron 87 (2015) 194–201
10 7
B = 0.1 T
B = 0.1 T
6
8
5
25 25
6
4
χmol/(10-6 m3 mol-1)
20
μeff/μB
μeff/μB
χmol/(10-6 m3 mol-1)
20
15 3 15
4
10 10
2
5 5
2
1 0
0
0 10 20 30 40
0 10 20 30 40
0 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
T/K T/K
6 6 6 6
B = 0.1 T B = 0.1 T
5 5 5 5
χmolT / cm3.mol-1.K
χmolT / cm3.mol-1.K
4 4 4 4
μeff/μB
μeff/μB
3 3 3 3
2 2 2 2
1 1 1 1
0 0 0 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300 350 400
T/K T/K
Fig. 5. Effective magnetic moment for 1 (up left), 2 (bottom left), 3 (up right) and 4 (bottom right). The inset for 1 and 3 shows the low temperature fit of the magnetic
susceptibility. The empty circles for 2 and 4 show the vmT product which is directionally proportional to the high spin fraction. Circles – experimental data, solid lines – fitted.
R(v) = 0.039 and R(M) = 0.042 for the susceptibility and magnetiza-
tion, respectively. Here; J is too small in order to cause a maximum
at the susceptibility curve above 1.9 K, nevertheless, it is responsi-
6
ble for a rapid drop of the effective magnetic moment at the lowest
temperatures (Fig. 5).
Also the complex 3, [Fe(LBr)(NCS)], is high-spin in the whole
5
temperature range and the same approach can be used as for 1
except that no magnetic exchange is present. The data fitting gave
the following set of optimum parameters: g = 1.926, aTIM = 4
2.98 109 m3 mol1, D/hc = 0.53 cm1, H = 0.130 K;
R(v) = 0.011, R(M) = 0.015. In this case the drop of the effective
μeff/μB
Table 3
Review of spin crossover parameters in Fe(III) complexes with pentadentate Schiff-base ligands.
cooling down, not all molecules switch their state (ESI, Fig. S2). The [Fe(LBr)(NCS)] complex molecule, which would lead to lowering
obtained value of the transition enthalpy is DH = 6.03 kJ mol1 and of the overall ligand-field strength resulting in stabilization of
6.18 kJ mol1, the transition entropy DS = 18.5 J K1 mol1 and the HS state. Only a very slight difference between the ligand field
19.5 J K1 mol1 and the characteristic temperature Tc = 326 K strengths of the LCl and LBr ligands can be deduced from the tran-
and 317 K for the model A and B, respectively. In both models C/ sition temperatures of the SCO compounds: Tc([Fe(LCl)
kB < Tc so that a thermal hysteresis is absent. The optimum param- (NCSe)]) = 293 K and Tc([Fe(LBr)(NCSe)]) 320 K. The alteration of
eters resulting from the fitting procedure are collected in Table 3 the L1 ligand from NCSe to NCS does not lead to a sizable
and the fitted data are displayed as solid lines in Fig. 5 for the case decrease of the increment to the crystal-field strength (D = gM fL);
of model B. fL(NCS)=1.02 while fL(NCSe) = 1.05 (fL(Cl) = 0.78).
Substitution of the NCSe ligand for N3 led to significant lower- When inspecting the topology of the complex molecule, a sig-
ing of the transition temperature, but the complex 2 still shows nificant difference can be found: the a parameter describing the
SCO (Fig. 5). The same fitting procedure was applied with the dihedral angle between the planes of the aromatic rings differs sig-
results DH = 1.64 kJ mol1 and 1.69 kJ mol1, DS = 11.5 J K1 mol1 nificantly when comparing three SCO compounds (a = 84.5–85.9°)
and 12.1 J K1 mol1 and Tc = 143 K and 140 K, for models A and B, with compound 3 (a = 67.7°, Table 2). This difference opens a ques-
respectively. The cooperativeness is again below the limiting value tion if the molecular shape can play a dominant role in hindering
for the hysteresis. the SCO behavior in 3 as it was observed previously for a group
Thermal evolution of the effective magnetic moment for [Fe(LCl/Br) of [FeIII(R-L6)]+ complexes [41] (H2R-L6 = variously substituted H2
X] complexes is compared in Fig. 6. The quantitative characteris- sal2trien). In these compounds the purely LS or HS complexes pre-
tics, as they result from the fitting procedure are listed in Table 3. fer different molecular shapes (a is significantly larger for HS com-
It can be concluded that SCO centered near the room temperature pounds) [44]. However, this might not be the case for 3 and it can
is associated with a considerable transition enthalpy of the order of be documented by the example of compound 2, which has a very
DH 5.8–6.5 kJ mol1. The transition entropy varies as DS = 18.5– low value of the a parameter (a = 35.0°, Table 2) but still exhibits
22.3 J K1 mol1, which means that a considerable contribution SCO (Fig. 6). Furthermore, there are several examples of purely
from molecular vibrations assists. HS [Fe(La)(NCS)] compounds[12] (H2La stands for 2-{3-[2-
(1-hydroxynaphthalen-2-yl)methylenaminoethylaminopropylimi-
no}methyl)naphthalene-1-ol) with very low a values ranging from
3. Discussion 17° to 25°; this is in a clear contradiction with the magneto-struc-
tural correlation found for the [FeIII(R-L6)]+ complexes.
A detailed analysis of the crystal and molecular structures Remarkably, the crystal structure of 2 consists of the supramo-
reveals that the occurrence of SCO in the series of the [Fe(LCl/Br) lecular chains of the [Fe(LBr)(N3)] molecules held by N–H N
(L1)] compounds is influenced by a number of factors. First, the hydrogen bonds. It must be noted that such structural motif (i.e.
expected isostructurality of the isothiocyanato and isoselenocy- assembly to chain supramolecular structure by NH N/O hydrogen
anato compounds is not preserved within these two series: only bonding) was observed for similar SCO compounds [Fe(Ln)(NCO)]
the compounds [Fe(LCl)(NCS)], [Fe(LCl)(NCSe)] and [Fe(LBr)(NCSe)] and [Fe(Ln)(N3)] previously [11] (H2Ln stands for N,N0 -bis(2-
(4) are isostructural (P21/c, Fig. 6) and they exhibit SCO. The hydroxy-naphthylidene)-1,6-diamino-4-azahexane); however, in
[Fe(LBr)(NCS)] compound (3) possesses different crystal packing the case of 2 an additional methanol molecule is found in the crys-
(Pn, Fig. 4) and it stays HS over the whole temperature range tal structure (Fig. 2). On the other hand, the previously reported
(Fig. 5). compound with similar composition, [Fe(3MeO-L)(N3)], does not
When inspecting their molecular structures at the chromophore possess the above mentioned structural motif and it does not exhi-
level (compounds 3 and 4), it is apparent that they do not differ in bit SCO.[11] This brings discussion back to the importance of the
bond lengths or bond angles significantly from the typical LS/HS isostructurality, or in other words, of preservation of the SCO
forms of the SCO compounds, or from pure HS compounds molecular packing motifs and it underlines the importance of
reported for this type of complexes previously [9,12]. Thus there molecular organization in the solid state for the occurrence and
is no evidence that there exists any structural difference in the character of SCO [45].
200 C. Krüger et al. / Polyhedron 87 (2015) 194–201
5.1.1. [Fe(LBr)Cl]0.5H2O, 1 The magnetic data were taken with the SQUID apparatus
The Schiff-condensation of 5-bromo-2-hydroxybenzaldehyde (MPMS-XL7, Quantum Design) using the RSO mode of detection
(2.0 mmol) with an asymmetric triamine, 1,6-diamino-4-azahex- with ca 20 mg of the sample encapsulated in a gelatin-made sam-
ane, H2N(CH2)3NH(CH2)2NH2 in the ratio of 2:1 in 30 cm3 of metha- ple holder. The magnetic susceptibility taken at B = 0.1 T was cor-
nol resulted in a pentadentate Schiff-base ligand (bright yellow rected for the underlying diamagnetism and converted to the
solution) according to Scheme 1. The reaction mixture was stirred effective magnetic moment. The magnetization was measured at
at 55 °C for 30 min. This mixture was combined with a methanol two temperatures: T = 2.0 and T = 4.6 K.
solution of FeCl36H2O (2.0 mmol in 10 cm3) at 55 °C for 15 min.
Then, the solution of triethylamine (3.8 mmol in 5 cm3 of methanol) Acknowledgments
was added drop wise. After 40 min of stirring at the boiling temper-
ature, the black powder was separated by filtration on a fritted fun- Grant Agencies (Slovakia: VEGA-1/0233/12, APVV-0014-11,
nel, washed with cold methanol and diethyl ether (yield 36%). The APVV-0132-11; Germany: DAAD SK/DE; Czech Republic:
filtrate was left to evaporate spontaneously. A few days later, black PrF_2014_09) are acknowledged for the financial support.
crystals were separated. Yield: 37%. C19H20Br2FeClN3O2.5
(M = 581.49). Anal. Calc.: C, 39.25; H, 3.47; N, 7.23. Found: C,
Appendix A. Supplementary data
39.06; H, 3.57; N, 7.32%.
[1] L. Cambi, L. Szegö, Ber. Dtsch. Chem. Ges. (A and B Series) 64 (1931) 2591.
5.1.3. [Fe(LBr)(NCS)], 3 [2] L. Cambi, L. Szegö, Ber. Dtsch. Chem. Ges. (A and B Series) 66 (1933) 656.
Compound 1 (0.2 g, 0.4 mmol) was dissolved in 80 cm3 metha- [3] L. Cambi, L. Malatesta, Ber. Dtsch. Chem. Ges. (A and B Series) 70 (1937) 2067.
nol and combined with 0.041 g (0.42 mmol) KNCS in 20 cm3 of [4] (a) P.J. van Koningsbruggen, Y. Maeda, H. Oshio, Top. Curr. Chem. 233 (2004)
259;
methanol. Then, the solution was stirred for 60 min and filtered.
(b) M. Nihei, T. Shiga, Y. Maeda, H. Oshio, Coord. Chem. Rev. 251 (2007) 2606;
Black single crystals were obtained by slow evaporation at room (c) J. Olguin, S. Brooker, Spin-Crossover in Discrete Polynuclear Complexes, in:
temperature. C20H19FeN4O2Br2S1 (M = 595.11). Anal. Calc.: C, Malcolm A. Halcrow (Ed.), Spin-Crossover Materials: Properties and
40.37; H, 3.22; N, 9.41. Found: C, 39.98; H, 3.21; N, 9.30; IR Applications, Wiley, 2013;
(d) K.S. Murray, The Development of Spin-Crossover Research, in: Malcolm A.
(ATR): m~(N–H) = 3166 cm1, m~(C–H) = 2913 cm1 (m), Halcrow (Ed.), Spin-Crossover Materials: Properties and Applications, Wiley,
1 1
m~(NCS) = 2052 cm , m~(C@N) = 1616 cm . 2013;
C. Krüger et al. / Polyhedron 87 (2015) 194–201 201
(e) P.N. Martinho, C. Rajnák, M. Ruben, Nanoparticles, Thin Films and Surface [23] F. Renz, P. Kerep, D. Hill, R. Müller-Seipel, M. Klein, Hyperfine Interact. 168
Patterns From Spin Crossover Materials and Electrical Spin State Control, in: (2006) 1051.
Malcolm A. Halcrow (Ed.), Spin-Crossover Materials: Properties and [24] M. Gembický, R. Boča, F. Renz, Inorg. Chem. Commun. 3 (2000) 662.
Applications, Wiley, 2013. [25] F. Renz, V. Martinez, M. Klein, Hyperfine Interact. 184 (2008) 251.
[5] (a) P. Gütlich, A.B. Gaspar, Y. Garcia Beilstein, J. Org. Chem. 9 (2013) 342; [26] F. Renz, V. Martinez, M. Klein, M. Schott, T. Hoffmann, M. Blumers, I. Fleischer,
(b) A. Bousseksou, G. Molnár, L. Salmon, W. Nicolazzi, Chem. Soc. Rev. 40 G. Klingelhöfer, R. Boča, M. Menzel, Hyperfine Interact. 184 (2008) 259.
(2011) 3313; [27] S. Hayami, Z.Z. Gu, H. Yoshiki, A. Fujishima, O. Sato, J. Am. Chem. Soc. 123
(c) G. Molnár, L. Salmon, W. Nicolazzi, F. Terki, A. Bousseksou, J. Mater. Chem. (2001) 11644.
C 2 (2014) 1360. [28] J.K. Tang, J.S. Costa, S. Smulders, G. Molnár, A. Bousseksou, S.J. Teat, Y.G. Li, G.A.
[6] R. Boča, Y. Fukuda, M. Gembický, R. Herchel, R. Jaroščiak, W. Linert, F. Renz, J. van Albada, P. Gamez, J. Reedijk, Inorg. Chem. 48 (2009) 2128.
Yuzurihara, Chem. Phys. Lett. 325 (2000) 411. [29] S. Dorbes, L. Valade, J.A. Real, C. Faulmann, Chem. Commun. (2005) 69.
[7] F. Renz, P. Kerep, D. Hill, M. Klein, Hyperfine Interact. 168 (2007) 981. [30] M.S. Haddad, W.D. Federer, M.W. Lynch, D.N. Hendrickson, Inorg. Chem. 20
[8] R. Boča, I. Nemec, I. Šalitroš, J. Pavlik, R. Herchel, F. Renz, Pure Appl. Chem. 81 (1981) 131.
(2009) 1357. [31] P. Gütlich, Eur. J. Inorg. Chem. (2013) 581.
[9] C. Krüger, P. Augustín, I. Nemec, Z. Trávníček, H. Oshio, R. Boča, F. Renz, Eur. J. [32] H.A. Goodwin, Coord. Chem. Rev. 18 (1976) 293.
Inorg. Chem. 5–6 (2013) 902. [33] P. Gütlich, Struct. Bonding 44 (1981) 83.
[10] I. Šalitroš, R. Boča, Ľ. Dlháň, M. Gembický, J. Kožíšek, J. Linares, J. Moncoľ, I. [34] E. König, Struct. Bonding 76 (1991) 51.
Nemec, L. Perašínová, F. Renz, I. Svoboda, H. Fuess, Eur. J. Inorg. Chem. (2009) [35] P. Gütlich, A. Hauser, H. Spiering, Angew. Chem. 106 (1994) 2109.
3141. [36] P. Gütlich, Y. Garcia, H.A. Goodwin, Chem. Soc. Rev. 29 (2000) 419.
[11] I. Nemec, R. Herchel, R. Boča, Z. Trávníček, I. Svoboda, H. Fuess, W. Linert, [37] P. Gütlich, Y. Garcia, P.J. van Koningsbruggen, F. Renz, in: F. Palacio, J.
Dalton Trans. 40 (2011) 10090. Schweizer, E. Ressouche (Eds.), Introduction to Physical Techniques in
[12] S. Mossin, B.L. Tran, D. Adhikari, M. Pink, F.W. Heinemann, J. Sutter, R.K. Molecular Magnetism, Part 1 – Structural and Magnetic Methods, University
Szilagyi, K. Meyer, D.J. Mindiola, J. Am. Chem. Soc. 134 (2012) 13651. Press, Zaragoza, 2000.
[13] I. Nemec, R. Boča, R. Herchel, Z. Trávníček, M. Gembický, W. Linert, Monatsh. [38] A. Hauser, J. Jeftić, H. Romstedt, R. Hinek, H. Spering, Coord. Chem. Rev. 190–
Chem. 140 (2009) 815. 192 (1999) 471.
[14] I. Šalitroš, R. Boča, R. Herchel, J. Moncoľ, I. Nemec, M. Ruben, F. Renz, Inorg. [39] Spin crossover in transition metal compounds I, II, IIIP. Gütlich, H.A. Goodwin
Chem. 51 (2012) 12755. (Eds.), Topics in Current Chemistry, 233, Springer, Berlin, 2004.
[15] R. Herchel, R. Boča, M. Gembický, J. Kožíšek, F. Renz, Inorg. Chem. 43 (2004) [40] P. Gütlich, P. van Koningsbruggen, F. Renz, Struct. Bonding 107 (2004) 27.
4103. [41] M.A. Halcrow, Chem. Soc. Rev. 40 (2011) 4119.
[16] F. Renz, P. Kerep, Polyhedron 24 (2005) 2849. [42] R. Boča, Handbook of Magnetochemical Formulae, Elsevier, Amsterdam, 2012.
[17] F. Renz, P. Kerep, Hyperfine Interact. 156/157 (2004) 371. [44] I. Nemec, R. Herchel, I. Šalitroš, Z. Trávníček, J. Moncoľ, H. Fuess, M. Ruben, W.
[18] F. Renz, M. Klein, S. Jung, P. Homenya, R. Saadat, D. Nariaki, Unimagazin Linert, CrystEngComm 14 (2012) 7015.
(Leibniz Universität Hannover) 1–2 (2011) 24. [45] J. Tao, R.-J. Wei, R.-B. Huang, L.-S. Zheng, Chem. Soc. Rev. 41 (2012) 703.
[19] F. Renz, D. Hill, M. Klein, J. Hefner, Polyhedron 26 (2007) 2325. [46] G.M. Sheldrick, Acta Crystallogr., Sect. A 64 (2008) 112.
[20] R. Boča, I. Šalitroš, J. Kožíšek, J. Linares, J. Moncoľ, F. Renz, Dalton Trans. 39 [47] L.J. Farrugia, J. Appl. Crystallogr. 45 (2012) 849.
(2010) 2198. [48] (a) A.L. Speck, Acta Crystallogr., Sect. D. Biol. Crystallogr. 65 (2009) 148;
[21] F. Renz, S. Jung, M. Klein, M. Menzel, A.F. Thünemann, Polyhedron 2009 (1818) (b) Y. Le Page, MISSYM1.1 – a flexible new release, J. Appl. Crystallogr. 21
28. (1988) 983.
[22] F. Renz, C. Zaba, L. Rossberg, S. Jung, M. Klein, G. Klingelhöfer, A. Wünsche, S. [49] C.F. Macrae, P.R. Edgington, P. McCabe, E. Pidcock, G.P. Shields, R. Taylor, M.
Reinhardt, M. Menzel, Polyhedron 28 (2009) 2036. Towler, J. van de Streek, J. Appl. Crystallogr. 39 (2006) 453.
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 151 (2015) 945–955
h i g h l i g h t s g r a p h i c a l a b s t r a c t
a r t i c l e i n f o a b s t r a c t
Article history: A comparative study was established between two signal processing techniques showing the theoretical
Received 21 April 2015 algorithm for each method and making a comparison between them to indicate the advantages and
Received in revised form 20 June 2015 limitations. The methods under study are Numerical Differentiation (ND) and Continuous Wavelet
Accepted 25 June 2015
Transform (CWT). These methods were studied as spectrophotometric resolution tools for simultaneous
Available online 30 June 2015
analysis of binary and ternary mixtures. To present the comparison, the two methods were applied for
the resolution of Bisoprolol (BIS) and Hydrochlorothiazide (HCT) in their binary mixture and for the
Keywords:
analysis of Amlodipine (AML), Aliskiren (ALI) and Hydrochlorothiazide (HCT) as an example for ternary
Derivative Spectrophotometry
Continuous Wavelet Transform
mixtures. By comparing the results in laboratory prepared mixtures, it was proven that CWT technique
Amlodipine is more efficient and advantageous in analysis of mixtures with severe overlapped spectra than ND.
Aliskiren The CWT was applied for quantitative determination of the drugs in their pharmaceutical formulations
Bisoprolol and validated according to the ICH guidelines where accuracy, precision, repeatability and robustness
Hydrochlorothiazide were found to be within the acceptable limit.
Numerical Differentiation Ó 2015 Elsevier B.V. All rights reserved.
1. Introduction
http://dx.doi.org/10.1016/j.saa.2015.06.100
1386-1425/Ó 2015 Elsevier B.V. All rights reserved.
946 E.S. Elzanfaly et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 151 (2015) 945–955
commonly used signal processing technique is Derivative AmturnideÒ is a combination of Amlodipine besylate (AML),
Spectrophotometry (DS). Aliskiren hemifumarate (ALI) and Hydrochlorothiazide (HCT) indi-
Rutherford [1] was the first to apply derivative methods suc- cated for the treatment of hypertension to reduce the risk of
cessfully to experimental data, he suggested that the sensitivity strokes and myocardial infarctions [56]. Many reported methods
of mass spectroscopy would be increased on applying the first were reported for the determination of AML, ALI or HCT in different
derivative technique. Derivative methods were first applied to dosage forms [57–62]. HPLC and capillary electrophoresis methods
UV–VIS and IR spectroscopy in 1953 by Hammond and Price [2], were reported for simultaneous determination of this ternary mix-
followed by Morrison [3] and in 1955 Giese and French [4] ture [63,64]. No spectrophotometric methods were developed for
described the theory of the derivative technique. In 1959, Martin the determination of this mixture.
[5] introduced the theoretical basis for second and higher orders In this work, ND and CWT were applied for determination of
of derivative, and in 1968, Morrey [6] generated first to fourth Bisoprolol (BIS) and Hydrochlorothiazide (HCT) in their binary
order derivatives with a digital computer. In 1964, Savitzky and mixture and for determination of Amlodipine (AML), Aliskiren
Golay [7] introduced least-squares procedures for the smoothing (ALI) and Hydrochlorothiazide (HCT) in their ternary mixture as
and differentiation of fluctuating data. examples for complex spectra. A comparative study between the
Derivative calculation is a powerful technique used in analytical two methods was performed to compare their resolution power
chemistry to resolve spectra, sharpen peaks, remove background in analysis of overlapped spectra. Then the CWT was used for the
interference, and carry out quantitative analysis. The most impor- analysis of the cited mixtures in pharmaceutical formulations,
tant field is its use in resolution enhancement technique to facili- namely ConcorÒ Plus tablets and AmturnideÒ tablets.
tate the detection and location of poorly resolved components in
complicated experimental signals [8]. It was applied successfully
to chromatography [9], electrochemistry [10–12], flow-injection 2. Theoretical background
analysis [13] and in all fields of spectroscopy [14–16]. In the region
of UV–VIS spectroscopy, DS played an important role in resolution 2.1. Numerical Differentiation (ND) [65]
of overlapped spectra [17,18]. The zero-crossing derivative spectra
method was the most common procedure for the simultaneous For a discrete spectrum xi (i = 1, . . ., n), suppose that wi
determination of single components as well as binary and ternary (i = 1, . . ., n) are its sampling wavelengths, the direct-difference
mixtures from their overlapping spectra [19–21]. DS was also method can be expressed as follows:
applied to ratio spectra for analysis of binary and ternary mixtures
X iþ1 X i
[22–24]. yi ¼
W iþ1 W i
In the spectral studies, the application of the usual DS to the
original absorption spectra possesses several disadvantages such
Here yi (i = 1, . . ., n 1) represent the series of derivatives of spec-
as: peak intensity diminishes with higher order derivation, it
trum xi. If the sampling wavelength is equal, this equation can be
required the additional smooth function as well as the scaling
rewritten as yi = xi+1 xi (i = 1, . . ., n 1).
factor processes. These parameters may deform the obtained
derivative spectra from the original ones. As a result, the usual
derivative method produces several errors in the quantitative anal- 2.2. Wavelet Transform (WT): [27]
ysis [25]. These drawbacks suggested using other signal processing
techniques as powerful alternatives for derivative calculation such Wavelet Transform (WT) involves decomposition of a signal
as Wavelet Transform [26,27]. function or vector (spectrum) into simpler, fixed building blocks
A Wavelet Transform (WT) involves the decomposition of a sig- at different scales and positions. In WT treatment, all basis func-
nal function or vector e.g., a spectrum of a chemical species into tions Wa;b ðxÞ can be derived from a mother wavelet WðxÞ through
simpler, fixed building blocks at different scales and positions the following dilation and translation processes:
[28]. Different algorithms were proposed to achieve this purpose.
In 1977, Coifman and Weiss [29] developed the automatic decom- 1 xb
Wa;b ðxÞ ¼ a2 W a; b 2 R and a–0
position method, based on work of Calderón and Zygmund [30], a
which had a very strong influence on the development of wavelet
theory. In 1984, Grossmann and Morlet [31] proposed the where a and b are the scale and position parameters, respectively,
Continuous Wavelet Transform (CWT). In 1989, Mallat [32] intro- expressed in real number R.
duced the multi-resolution signal decomposition MRSD algorithm. The basic idea of WT is to represent any arbitrary function f ðxÞ
Daubechies [26,33,34] adopted this approach to construct families as a superposition of wavelets. Decomposition of x with respect to
of compactly supported wavelets and coupled it to quadrature mir- the wavelet function series {Wj;k ðxÞ} is described by the following
ror filtering. formula:
WT has been applied successfully for signal processing in differ-
X
þ1 X
þ1
ent fields of analytical chemistry. It was applied to de-noising of f ðxÞ ¼
ðjÞ
C k Wj;k
data [35], compression of signals [36], analysis of electrochemical j¼1 k¼1
data [37], flow-injection analysis [38], multivariate calibration
[39] and quantitative analysis of near infrared spectra [40]. In the Therefore, the signal is represented by a set of coefficients {C j;k } in
field of UV–VIS spectrophotometry, Continuous Wavelet the wavelet domain.
Transform (CWT) combined either with a zero-crossing technique The approximate first derivative of X is assigned to be equal to
[41] or ratio spectra [13] was used for simultaneous determination the difference between two scale coefficients at the first resolution
of chemical species in binary and ternary mixtures [18,42–45]. level as
ConcorÒ Plus tablets contain both Bisoprolol hemifumarate
(BIS) and Hydrochlorothiazide (HCT) for the treatment of high xð1Þ C 1;D2m C 1;D2m~ ~
m–m
blood pressure [46]. There are reported methods for the determi-
nation of BIS or HCT in different dosage forms [47–50] and in their where, D2m and D2m~ represent any two Daubechies wavelet
binary mixtures [45,51–55]. ~ being any positive integer.
functions, with m and m
E.S. Elzanfaly et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 151 (2015) 945–955 947
SHIMADZU dual beam UV–visible spectrophotometer For determination of BIS, the 3rd and 4th derivative spectra were
(Kyoto/Japan), model 1650 UV-PC, with matched 1-cm quartz cells obtained by ND method with Dk = 8 nm and scaling factor = 1000.
was used. The bundled software, UV-Probe personal spectroscopy The peak amplitude was measured at 283.6 and 287.5 nm,
software version 2.21 (SHIMADZU) is used. The spectral bandwidth respectively.
is 2.0 nm and wavelength scanning speed is 2800 nm/min with For determination of HCT, the first derivative spectra were
0.1 nm interval. obtained by ND method with Dk = 4 nm and scaling factor = 100.
The peak amplitude was measured at 272.3 nm.
3.2. Chemicals and reagents The calibration curves were constructed relating the peak
amplitudes at the specified wavelengths to the corresponding
3.2.1. Pure samples concentrations of BIS and HCT.
Ternary mixture
- BIS, HCT and AML; kindly supplied by Al-Hekma pharmaceuti-
cal Company, Cairo, Egypt, their purity was certified to be
For determination of AML, the first derivative spectra were
99.7 ± 0.8%, 100.6 ± 0.6% and 99.9 ± 0.5%, respectively.
obtained by ND method with Dk = 4 nm and scaling factor = 10.
- ALI; kindly supplied by Novartis Pharmaceuticals Corporation,
The peak amplitude was measured at 386.8 nm.
USA, its purity was certified to be 100.6 ± 0.3%.
For determination of ALI, the fourth derivative spectra were
obtained by ND method with Dk = 8 nm and scaling factor = 1000.
3.2.2. Market samples
The peak amplitude was measured at 287.5 nm.
For determination of HCT, the second derivative spectra were
- ConcorÒ 5 Plus labeled to contain 5(BIS)/12.5(HCT) mg, batch
obtained by ND method with Dk = 8 nm and scaling factor = 100.
number 1030039 and ConcorÒ 10 Plus labeled to contain
The peak amplitude was measured at 332.1 nm.
10(BIS)/25(HCT) mg, batch number 0795049, manufactured
The calibration curves were constructed relating the peak
by Pfizer Ltd., Cairo, Egypt.
amplitudes at the specified wavelengths to the corresponding
- AmturnideÒ tablet dosage forms; labeled to contain 10
concentrations of AML, ALI and HCT.
(AML)/300 (ALI)/25 (HCT) mg batch number F0005, manufac-
tured by Novartis Pharmaceuticals Corporation, USA.
3.3.3.2. Continuous Wavelet Transform method (CWT).
3.2.3. Methanol
Binary mixture
Spectroscopy grade (El-NASR Pharmaceutical Chemicals Co.,
Abu-Zaabal, Cairo, Egypt).
The first derivative spectra were obtained by CWT method with
Morlet family (morl) and scaling = 200. The calibration curves were
3.3. Procedures constructed relating the peak amplitudes at 278.3 and 283.6 nm to
the corresponding concentrations of BIS and HCT, respectively.
3.3.1. Standard solutions
Ternary mixture
- BIS, HCT, AML and ALI standard stock solutions; 1000 lg /mL in
methanol. For determination of AML, the first derivative spectra were
- BIS, HCT and AML standard working solutions; 50 lg/mL in obtained by CWT method with Gaussian-1 family (Gaus-1) and
methanol. scaling = 200. The peak amplitude was measured at 387.2 nm.
- ALI standard working solution; 500 lg/mL in methanol. For determination of ALI, the first derivative spectra were
- Preparation of laboratory prepared mixtures; aliquots of stan- obtained by CWT method with Gaussian-8 family (Gaus-8) and
dard working solutions were transferred into two series of scaling = 200. The peak amplitude was measured at 315.6 nm.
10-mL measuring flasks, completed to volume with methanol For determination of HCT, the first derivative spectra were
to prepare binary mixture of BIS and HCT and ternary mixture obtained by CWT method with Coiflet-3 family (Coif-3) and
of AML, ALI and HCT containing different ratios of the drugs. scaling = 200. The peak amplitude was measured at 289.1 nm.
The calibration curves were constructed relating the peak
3.3.2. Spectral characteristics of the drugs amplitudes at the specified wavelengths to the corresponding
The zero-order (D0) absorption spectra of 4 lg/mL BIS and concentrations of AML, ALI and HCT.
10 lg/mL HCT were recorded against methanol as a blank over
the range of 200–400 nm. 3.3.4. Application of the proposed methods for the determination of
The zero-order (D0) absorption spectra of 4, 120 and 10 lg/mL laboratory-prepared mixtures
solutions of AML, ALI and HCT, respectively, were recorded against The spectra of laboratory prepared mixtures were scanned from
methanol as a blank over the range of 200–400 nm. 200 to 400 nm and stored in the computer. The same procedure
under construction of calibration curves was applied and the con-
3.3.3. Construction of calibration curves centrations of the drugs were calculated from the corresponding
Aliquots equivalent to 20–400 lg BIS, 5–250 lg HCT, 10–350 lg regression equations.
AML and 150–2500 lg ALI were accurately transferred from their
respective standard working solutions into separate series of 3.3.5. Application of the proposed methods for the determination of the
10-mL volumetric flasks then completed to volume with methanol. drugs in dosage forms
The spectra of the prepared standard solutions were scanned from Ten tablets of both ConcorÒ 5 Plus, ConcorÒ 10 Plus were accu-
200 to 400 nm and stored in the computer. rately weighed, finely powdered and an amount of the powder
948 E.S. Elzanfaly et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 151 (2015) 945–955
equivalent to 10 mg BIS was weighed. Five tablets of AmturnideÒ ALI and HCT show highly overlapped spectral band in the region
were accurately weighed, finely powdered and amount of the pow- 200–350 nm (Fig. 2). The difficulty of BIS determination in the bin-
der equivalent to 300 mg ALI was weighed. The powder was dis- ary mixture and the severe overlap in the ternary mixture nomi-
solved in methanol by shaking in ultrasonic bath for about nated these combinations to be ideal mixtures presenting the
30 min, then they were filtered and transferred quantitatively into comparative study to show the advantages and disadvantages of
separate 100-ml volumetric flasks and the volume was completed both algorithms in analysis of overlapping spectra. HCT in the bin-
to the mark with methanol. Ten ml aliquots were transferred into ary mixture and AML in the ternary mixture can be directly deter-
separate 50-ml volumetric flasks and the volume was completed to mined in the zero order spectra at their kmax, 316 and 360 nm
the mark with methanol. Then necessary dilutions were made to respectively, without interference from other components, but
reach concentrations in the linearity range. The spectra of these they were determined by ND and CWT for investigational
solutions were scanned from 200 to 400 nm and stored in the com- purposes.
puter. Spectra obtained were analyzed by the proposed methods. For using ND method, the parameters of derivative calculation,
wavelength increment over which the derivative is obtained (Dk)
and the scaling factor, were optimized. Different Dk (2, 4 and 8)
4. Results and discussion
and scaling factors of 10, 100 and 1000 were tested, and the best
parameters, in terms of spectral shapes, linearity and recovery,
Signal processing techniques use different algorithms in trans-
are mentioned under procedure.
forming the spectra into new graphs where the spectra of different
For analysis of chosen mixtures using CWT algorithm, different
compounds are resolved. Derivative Spectrophotometry (DS) is the
wavelet families were tested such as Haar, Daubechies (db),
most commonly used processing technique for resolution of over-
Symlets (sym), Coiflets (coif), Meyer (meyr), Gaussian (gaus),
lapped spectra. Lately, Continuous Wavelet Transform algorithm
Mexican hat (mexh) and Morlet (morl). The only parameter that
(CWT) started to be used as an alternative to Numerical
should be optimized after choosing the proper wavelet is the scal-
Differentiation (ND) for calculation of DS.
ing parameter. No need for scaling factor as CWT has the advantage
The aim of this work was to establish a comparative study
of signal amplification compared to ND.
between the two signal processing techniques. For proper presen-
tation of the comparison, two different mixtures were chosen to
show the resolution power of the methods. The first mixture was 4.1. Binary mixture
a combination of BIS and HCT as an example of binary mixtures.
The spectra of BIS and HCT (Fig. 1) show severe overlap as the peak In this section, we will focus on determination of BIS being a
of HCT (kmax 226 nm) completely coincide with that of BIS (kmax challenge in presence of highly absorptive HCT. Thus, the choice
225 nm) showing the same features. This overlap, together with of derivative order and optimum parameters is mainly considered
the high absorptivity of HCT which is also found in higher level for determination of BIS.
in ConcorÒ Plus tablets, prevents direct determination of BIS. In For analysis of BIS in this mixture using ND algorithm, the 1st,
previous work, the transformed ratio spectra had to be combined 2 , 3rd and 4th order derivatives were obtained (Fig. 3). The 1st
nd
with zero-crossing technique for analysis of BIS in this mixture order did not show any possible zero-crossing points that can be
[45]. The second mixture composed of AML, ALI and HCT was used used for BIS determination (Fig. 3a), while the wavelengths 241.0
as an example of ternary mixtures. The absorption spectra of AML, and 270.0 nm did not show a real zero-crossing. Possible
Absorbance
Wavelength (nm)
Fig. 1. Zero order absorption spectra of 4 lg/mL BIS (—) and 10 lg/mL HCT (- - -) using methanol as blank.
E.S. Elzanfaly et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 151 (2015) 945–955 949
Absorbance
Wavelength (nm)
Fig. 2. Zero order absorption spectra of 4 lg/mL AML (—), 120 lg/mL ALI (....) and 10 lg/mL HCT (- - -), using methanol as blank.
zero-crossing points were noticed in higher order derivatives, For analysis of the ternary mixture using CWT algorithm, differ-
which are 230.3 and 278.5 nm in 2nd order, 233.3 and 283.6 nm ent families were tried in different orders and scaling. The best
in 3rd order and 229.5, 235.3 and 287.5 nm in 4th order. In contrast results regarding presence of zero-crossing points, linearity and
to 1st order derivative, true zero-crossing points were detected in selectivity were obtained with 1st order derivatives using gaus-1,
2nd (278.5 nm), 3rd (283.6 nm) and 4th (287.5 nm) orders gaus-8 and coif-3 families and scaling = 200 for determination of
(Fig. 3b–d). The zero-crossing point in 2nd order showed poor lin- AML, ALI and HCT at 387.2, 315.6 and 289.1 nm, respectively
earity and recovery of laboratory prepared mixtures, while the (Fig. 6).
points in 3rd and 4th orders resulted in good recoveries in some The selectivity of the proposed ND and CWT methods was
laboratory-prepared mixtures, where BIS was found in higher ratio assessed by the analysis of laboratory prepared mixtures contain-
than HCT. For determination of HCT, the 1st order derivative was ing different ratios of the drugs (Tables 1 and 2). The proposed
enough to get zero-crossing point (272.3 nm) and good recoveries CWT method was applied for the determination of BIS and HCT
in all ratios of laboratory-prepared mixtures were obtained in ConcorÒ Plus tablets and AML, ALI and HCT in AmturnideÒ
Table 1. tablets and the validity was further assessed by applying the stan-
For analysis of BIS using CWT algorithm, the families were dard addition technique (Tables 3 and 4). Results obtained by CWT
tested in different orders and with different scaling parameters for analysis of the dosage forms were statistically compared to
and the best results, regarding linearity and recovery% in those obtained by the reported HPLC methods [51,63]. The results
laboratory-prepared mixtures, were obtained with 1st derivative showed no significant differences between the proposed method
calculated using morl family and scaling = 200 (Fig. 4) at and the reported ones as presented in Table 5. Validation was done
278.3 nm. Different wavelengths were detected as possible according to ICH recommendations and the validation parameters
zero-crossing points at 255.3, 265.7, 278.3, 292.2 and 306.3 nm. are shown in Table 6.
The points at 255.3 and 265.7 nm were not real zero-crossing
points and the last two wavelengths showed bad linearity and sen- 4.3. CWT versus ND
sitivity compared to 278.3 nm. HCT could be determined with the
same parameters at 283.6 nm. ND is the simplest method of derivative calculation. The major
problem is the choice of Dk. Wide Dk values can be used to sup-
4.2. Ternary mixture press noise and increase SNR, but they degrade resolution as well.
Finding the optimum Dk to compromise between the high resolu-
For analysis of this mixture using ND algorithm, the 1st order tion and good signal to noise ratio (S/N) is considered a challenge.
derivative spectra were obtained where a zero-crossing point Also the values of S/N obtained by ND are very small, so they are
(386.8 nm) was detected for AML determination (Fig. 5a). But in usually multiplied by scaling factor to be considerable.
the same order spectra, ALI and HCT were not resolved, so the The 1st order derivative fails to resolve severely overlapped
2nd order derivatives were calculated and a zero-crossing point spectra, so we use higher order derivatives. This approach can help
(332.1 nm) was noticed for determination of HCT (Fig. 5b). For in case the components to be determined are present in a reason-
determination of ALI, the 3rd and 4th order derivatives were tried. able ratio and have comparable absorptivities. In the case of BIS in
While no zero-crossing points were observed in the 3rd order spec- its binary mixture with HCT, the zero order spectrum of BIS is com-
tra (Fig. 5c), the 4th order showed a zero-crossing point (287.5 nm) pletely enclosed with that of HCT showing very similar features
that was used for ALI determination (Fig. 5d). and low absorptivity, so 1st order derivative failed to present a
950 E.S. Elzanfaly et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 151 (2015) 945–955
a b
HCT BIS
c
d
BIS
BIS
Fig. 3. Different order absorption spectra of 40 lg/mL BIS (–) and 1–22 lg/mL HCT (- - -) using ND method showing zero-crossing points for determination of BIS and HCT. (a)
1st order (b) 2nd order (c) 3rd order (d) 4th order.
Table 1
Determination of BIS and HCT in laboratory prepared mixtures by ND and CWT methods.
zero-crossing point for its determination. The 2nd, 3rd and 4th order In contrast to ND, CWT could resolve BIS from HCT in different
derivatives showed zero-crossing points for BIS determination, the ratios by 1st order derivative using morl family. Using CWT algo-
recoveries in laboratory prepared mixtures indicated that BIS can- rithm offers the flexibility of using several wavelet families, which
not be determined in some mixtures contained higher levels of results in transforming the spectra into 1st order derivative in dif-
HCT. This may be attributed to the poor absorptivity of BIS, at ferent shapes. This fact increases the possibility of finding
the wavelengths where zero-crossing points were observed, in zero-crossing points in the 1st order derivatives for compounds
addition to the fact that higher order derivatives diminish the sig- with very similar spectra. The CWT algorithm also has the advan-
nal to noise ratio (S/N). tage of signal amplification compared to ND [66]. In addition to
E.S. Elzanfaly et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 151 (2015) 945–955 951
Amplitude
HCT
BIS
Wavelength (nm)
st
Fig. 4. 1 order absorption spectra of 40 lg/mL BIS (—) and 1–22 lg/mL HCT (- - -) using CWT method (morl) showing zero-crossing points for BIS and HCT determination.
a b
AML
HCT
c d
ALI
Fig. 5. Different order absorption spectra of 35 lg/mL AML (–), 250 lg/mL ALI (....) and 25 lg/mL HCT (- - -) using ND method showing zero-crossing points for determination
of AML, ALI and HCT. (a) 1st order (b) 2nd order (c) 3rd order (d) 4th order.
the fact that calculating higher order derivatives is not necessary as zero-crossing points for determination of BIS was not observed in
the 1st order derivative in CWT is enough to get zero-crossing ND 1st order derivative, in contrast to the 1st order morl wavelet.
points. Thus CWT maintains high S/N in contrast to ND calculation, The higher S/N of CWT spectra allowed the determination of BIS
which is the reason why CWT showed higher sensitivity as shown in mixtures where HCT is found in higher ratios, while 2nd, 3rd
from LOD and LOQ values in Table 6. These facts can explain why and 4th order derivatives obtained by ND failed.
952 E.S. Elzanfaly et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 151 (2015) 945–955
AML
ALI
HCT
Fig. 6. First order absorption spectra of 35 lg/mL AML (–), 250 lg/mL ALI (....) and 25 lg/mL HCT (- - -), using CWT method showing zero-crossing point for determination of
AML, ALI and HCT. (a) gaus-1 (b) gaus-8 (c) coif-3.
The severe overlap between the spectra of AML, ALI and HCT in ALI and HCT show no interference in their original spectra. Upon
the region 200–350 nm prevented the resolution of the drugs in 1st using CWT families, gaus-1, gaus-8 and coif-3, the 1st order deriva-
order derivative in this region. So ALI and HCT were resolved only tives were enough to resolve the three drugs.
in the 4th and 2nd order derivatives, respectively. AML was excep- The advantages of CWT over ND in calculating derivatives can
tion being determined in 1st order derivative at 386.8 nm where be summarized in the following:
E.S. Elzanfaly et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 151 (2015) 945–955 953
Table 2
Determination of AML, ALI and HCT in their ternary mixture by ND and CWT methods.
Table 3
Determination of BIS and HCT in ConcorÒ Plus tablets by the proposed CWT and the reported HPLC method [51] with application of standard addition technique.
Table 4
Determination of AML, ALI and HCT in AmturnideÒ tablets by the proposed CWT and the reported HPLC method [63] with application of standard addition technique.
Table 5
Statistical comparison for the results obtained by the proposed CWT and the reported HPLC [51,63] methods for the analysis of ConcorÒ Plus and AmturnideÒ tablets.
Table 6
Assay validation sheet of ND and CWT methods for the simultaneous determination of the mixtures under study.
1. Flexibility in calculating 1st order derivatives in different shapes similar spectra without the need of higher order derivative
using different wavelet families, that allows CWT to provide calculation.
zero-crossing points for complex mixtures.
2. Signal amplification provided by the algorithm together with its
flexibility in 1st order derivatives gives CWT high sensitivity References
compared to ND. [1] E. Dymond, Math. Proc. Cambridge Philos. Soc. 22 (1924) 405–408.
[2] V. Hammond, W. Price, J. Opt. Soc. Am. 43 (1953) 924–930.
[3] J. Morrison, J. Chem. Phys. 21 (1953) 1767–1772.
[4] A.T. Giese, C.S. French, Appl. Spectrosc. 9 (1955) 78–96.
5. Conclusion [5] A. Martin, Spectrochim. Acta A 14 (1959) 97–103.
[6] J.R. Morrey, Anal. Chem. 40 (1968) 905–914.
[7] A. Savitzky, M.J. Golay, Anal. Chem. 36 (1964) 1627–1639.
Using CWT offered many advantages over traditional ND. One [8] G. Talsky, Derivative Spectrophotometry: Low and High Order, VCH, California,
advantage is that CWT algorithm maintains high S/N in contrast 1994.
to ND calculation. In addition to the fact that calculating higher [9] N. Funasaki, S. Hada, S. Neya, J. Phys. Chem. B 103 (1999) 169–172.
[10] B. Ebeshi, E. Vaikosen, D. Tuma, Int. J. Biol. Chem. Sci. 7 (2014) 2591–2599.
order derivatives is not necessary in CWT as the 1st order deriva-
[11] Y. Ni, J. Bai, Talanta 44 (1997) 105–109.
tive is enough to get zero-crossing points for severely overlapped [12] A. Molina, M. López-Tenés, C. Serna, J. Electrochem. Soc. 147 (2000) 3429–
spectra. The flexibility of CWT algorithm allowing the use of sev- 3435.
eral wavelet families in different orders, results in transforming [13] M. Blanco, J. Gene, H. Iturriaga, S. Maspoch, J. Riba, Talanta 34 (1987) 987–993.
[14] F. Singleton, G. Collier, J. Appl. Chem. 6 (1956) 495–510.
the spectra into different shapes of 1st order derivatives, which [15] W. Snelleman, T. Rains, K. Yee, H.D. Cook, O. Menis, Anal. Chem. 42 (1970)
increases the possibility of finding zero-crossing points in very 394–398.
E.S. Elzanfaly et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 151 (2015) 945–955 955
[16] T.C. O’Haver, Modulation and Derivative Techniques in Luminescence [47] A.D. Gudruman, N. Bibire, G. Tantaru, M. Apostu, M. Vieriu, V. Dorneanu, Rev.
Spectroscopy, Springer, New York, 1976. Chim. 64 (2013) 393–396.
[17] A.Y. El-Sayed, N.A. El-Salem, Anal. Sci. 21 (2005) 595. [48] H.W. Darwish, S.A. Hassan, M.Y. Salem, B.A. El-Zeany, Spectrochim. Acta A 122
[18] C.B. Ojeda, F.S. Rojas, Microchem. J. 106 (2013) 1–16. (2013) 744–750.
[19] M. Korany, A. Wahbi, M. Elsayed, S. Mandour, Farmaco Ed. Prat. 39 (1984) 243. [49] M. Emanual, J. Chromatogr. Sep. Tech. (2012).
[20] M. Korany, A. Wahbi, M. Elsayed, S. Mandour, Anal. Lett. 17 (1984) 1373–1389. [50] H.W. Darwish, S.A. Hassan, M.Y. Salem, B.A. El-Zeany, Int. J. Pharma Bio Sci. 4
[21] M.A. Korany, A.A.S. El-Din, N.A. Abael-salam, Anal. Lett. 17 (1984) 483–495. (2013) 345–356.
[22] F. Salinas, J. Nevado, A. Mansilla, Talanta 37 (1990) 347–351. [51] S.J. Joshi, P.A. Karbhari, S.I. Bhoir, K. Bindu, C. Das, J. Pharm. Biomed. Anal. 52
[23] J. Berzas, C. Guiberteau, F. Salinas, Talanta 39 (1992) 547. (2010) 362–371.
[24] E. Dinc, F. Onur, Anal. Chim. Acta 359 (1998) 93–106. [52] A. Singh, T. Sivakumar, D. Jain, U. Mishra, A. Yadav, T. Sharma, Int. J. Pharm.
[25] E. Dinç, D. Baleanu, Spectrochim. Acta A 63 (2006) 631–638. Res. Dev. 3 (2011) 11–22.
[26] I. Daubechies, Ten Lectures on Wavelets, SIAM, Pennsylvania, 1992. [53] J. Shi, Y. Chen, Y.P. Long, Cent. South Pharm. 11 (2012) 000.
[27] A.K.M. Leung, F.T. Chau, J.B. Gao, Anal. Chem. 70 (1998) 5222–5229. [54] S. Kurbanoglu, P.R. San Miguel, B. Uslu, S.A. Ozkan, Chromatographia (2013) 1–
[28] X.D. Dai, B. Joseph, R. Motard, Introduction to Wavelet Transform and Time- 7.
frequency Analysis, Springer, New York, 1994. [55] P.N. Bhoya, E.M. Patelia, J. Chromatogr. Sep. Tech. 4 (2013) 163–166.
[29] R.R. Coifman, G. Weiss, Bull. Am. Math. Soc. 83 (1977) 569–645. [56] Y. Lacourcière, S. Taddei, G. Konis, H. Fang, T. Severin, J. Zhang, J. Hypertens. 30
[30] A. Calderón, A. Zygmund, Stud. Math. 46 (1973) 297–299. (2012) 2047–2055.
[31] A. Grossmann, J. Morlet, SIAM J. Math. Anal. 15 (1984) 723–736. [57] F. Belal, M. Walash, N. El-Enany, S. Zayed, J. Chromatogr. B 933 (2013) 24–29.
[32] S.G. Mallat, IEEE Trans. Acoust. 37 (1989) 2091–2110. [58] H.W. Darwish, S.A. Hassan, M.Y. Salem, B.A. El-Zeany, Spectrochim. Acta A 113
[33] I. Daubechies, IEEE Trans. Inf. Theory 36 (1990) 961–1005. (2013) 215–223.
[34] I. Daubechies, SIAM J. Math. Anal. 24 (1993) 499–519. [59] H.W. Darwish, S.A. Hassan, M.Y. Salem, B.A. El-Zeany, Spectrochim. Acta A 104
[35] D.L. Donoho, I.M. Johnstone, C.R. Acad. Sci. I Math. 319 (1994) 1317–1322. (2013) 70–76.
[36] V. Barclay, R. Bonner, I. Hamilton, Anal. Chem. 69 (1997) 78–90. [60] M.I. Ezzeldien, E. Shokry, M.A. Fouad, R.I. Elbagary, Int. J. Adv. Chem. 1 (2013)
[37] X. Zou, J. Mo, Anal. Chim. Acta 340 (1997) 115–121. 13–20.
[38] M. Bos, E. Hoogendam, Anal. Chim. Acta 267 (1992) 73–80. [61] H.W. Darwish, S.A. Hassan, M.Y. Salem, B.A. El-Zeany, Int. J. Spectrosc. 2013
[39] H.W. Tan, S.D. Brown, J. Chemom. 16 (2002) 228–240. (2013) 8.
[40] U. Depczynski, K. Jetter, K. Molt, A. Niemöller, Chemom. Intell. Lab. Syst. 47 [62] H.W. Darwish, S.A. Hassan, M.Y. Salem, B.A. El-Zeany, Int. J. Pharma Bio Sci. 4
(1999) 179–187. (2013) 230–243.
[41] T. O’haver, G. Green, Anal. Chem. 48 (1976) 312–318. [63] G.K. Swamy, P. Sravanthi, M.L. Surekha, J.K.J. Rao, Int. J. ChemTech Res. 4
[42] E. Dinç, D. Baleanu, J. Pharm. Biomed. Anal. 31 (2003) 969–978. (2012) 1666–1673.
[43] E. Dinç, A. Özdemir, D. Baleanu, Talanta 65 (2005) 36–47. [64] M.M. Salim, W.M. Ebeid, N. El-Enany, F. Belal, M. Walash, G. Patonay, J. Sep. Sci.
[44] E. Dinç, M. Kanbur, D. Baleanu, Die Pharm. 60 (2005) 892–896. 37 (2014) 1206–1213.
[45] E.S. Elzanfaly, S.A. Hassan, M.Y. Salem, B.A. El-Zeany, Spectrochim. Acta A 140 [65] F.T. Chau, Y.Z. Liang, J. Gao, X.G. Shao, Chemometrics: From Basics to Wavelet
(2015) 334–343. Transform, John Wiley & Sons, New Jersey, 2004.
[46] W.H. Frishman, B.S. Bryzinski, L.R. Coulson, V.L. DeQuattro, N.D. Vlachakis, W.J. [66] E. Dinç, M. Kanbur, D. Baleanu, Spectrochim. Acta A 68 (2007) 225–230.
Mroczek, G. Dukart, J.D. Goldberg, D. Alemayehu, K. Koury, Arch. Internal Med. [67] F.J. Anscombe, Technometrics 2 (1960) 123–146.
154 (1994) 1461–1468.
Journal of Electroanalytical Chemistry 827 (2018) 51–57
A R T I C LE I N FO A B S T R A C T
Keywords: In this study, clopidogrel (CLP), which is an antiplatelet agent widely used in the prevention of ischemic stroke,
Clopidogrel was analyzed using functionalized multi-walled carbon nanotubes (fMWCNTs) and CdSe quantum dots (QDs)
Carbon nanotube modified glassy carbon electrode. The dependence of intensities of currents and potential on pH, concentration,
Quantum dots scan rate, nature of the buffer was further investigated. The oxidation process exhibited mixed diffusion under
Voltammetry
adsorption controlled process depending on pH. For the analytical application, adsorptive stripping differential
Determination
pulse voltammetric (AdSDPV) technique was used and all operational parameters were optimized. The plot of
peak current vs CLP concentration consisted of a linear segment in the concentration range between
2.0 × 10−6 M and 4.0 × 10−5 M and the detection limit (3σ) was obtained 7.55 × 10−8 M in pH 2.14 phosphate
buffer solution. Moreover, in order to evaluate the analytical applicability of the proposed method, it was ap-
plied to the determination of CLP in serum samples with linear range of 2.5 × 10−6–1.5 × 10−5 M and the
detection limit value of 2.99 × 10−7 M. Determination of CLP in tablet dosage form Atervix® was also proposed
with acceptable recovery values.
1. Introduction Iijima, this modifier attracted scientists' attention because of its po-
tential in biological and biomedical applications [7–11]. CNTs, im-
Clopidogrel (CLP, (S)-Methyl 2-(2-chlorophenyl)-2-(6,7-dihy- portant carbon allotropes, are all‑carbon hollow and well-ordered gra-
drothieno[3,2-c]pyridin-5(4H)-yl)acetate, Scheme 1), a thienopyridine phitic nanomaterials with a high aspect ratio, having lengths from
derivative drug, is a prodrug that prevents adenosine diphosphate-in- several hundred nanometers to several micrometers. CNTs are cylind-
duced platelet aggregation selectively and irreversibly; there is no di- rical nanostructures obtained by rolling up single or multiple sheets of
rect antiplatelet activity in the in vitro environment [1,2]. It transforms graphene to give single-walled carbon nanotubes (SWCNTs) and multi-
into active form by hepatic cytochrome P-450 3A4 and 3A5 [3]. CLP walled carbon nanotubes (MWCNTs). SWCNTs have 0.4–2 nm dia-
reduces the incidence of atherosclerotic events (myocard infarctus, meters and MWCNTs have 2–100 nm coaxial diameters [7,9,12–18].
stroke and vascular death) at people with atherosclerosis and at those They show unique properties such as electrical conductivity, large
who has the risk for having MI, stroke, acute coronary syndrome or surface area, high electrocatalytic activity and strong chemical stability,
peripheral artery disease. Especially, it is more effective than aspirin for for being an electrode modifier [19]. CNTs can be functionalized with
peripheral artery disease [2–5]. some groups such as eCOOH, eOH, eNH2 and eSH to improve the
Nano-sized materials can be used by mixing them with electrode adsorption capacities [20] and solubility [7–11]. On the other hand, in
materials or they can be used via direct application on the electrode recent years quantum dots (QDs) have also been presented as an at-
surface. Electrodes, modified with nanomaterials, can be fabricated by tractive electrode modification agent. QDs are semiconductor nano-
chemical vapor deposition, electrochemical deposition, chemical at- particles with size range of 1–100 nm nanocrystal particle
tachment with or without crosslinking agents, and mixing of nano- [14–16,21–24]. These materials having distinct electronic states are
particles with electrode materials [6]. Among all nanomaterials, carbon widely preferred recently because of their highly alterable properties.
nanotubes (CNTs) are one of the most preferred materials for electrode Having been used as electron transfer catalysts, QDs have electron hole
modification. Dated from the discovery of CNTs in 1991 by Sumio pairs on which redox reactions occur in a liquid interface. The reason
⁎
Corresponding author.
E-mail address: buslu@pharmacy.ankara.edu.tr (B. Uslu).
https://doi.org/10.1016/j.jelechem.2018.09.005
Received 24 July 2018; Received in revised form 31 August 2018; Accepted 3 September 2018
Available online 04 September 2018
1572-6657/ © 2018 Elsevier B.V. All rights reserved.
G. Ozcelikay et al. Journal of Electroanalytical Chemistry 827 (2018) 51–57
2. Experimental
Scheme 1. Chemical structure of CLP.
2.1. Apparatus
Clopidogrel bisulphate (CLP) and its dosage form, Atervix® tablets, were
kindly provided by Biofarma İlaç (Istanbul, Turkey). Required amount of
Fig. 1. AdSDPV of 1 × 10−5 M CLP at the bare GCE (1, black), CdSeQDs/GCE CLP was weighed and dissolved in methanol for preparing 1.0 × 10−3 M
(2, pink), fMWCNT/GCE (3, red) and fMWCNT/CdSeQDs/GCE (4, blue) in stock solution and it is stored in the refrigerator. 20% (v/v) methanol per-
pH 2.14 phosphate buffer solution. (For interpretation of the references to cent was kept as constant in all pH values and buffers to avoid the solubility
colour in this figure legend, the reader is referred to the web version of this problems. eCOOH group functionalized multiwalled carbon nanotubes
article.) (fMWCNT) and CdSeQDs (DRP-QDCORE550: λem = 550 nm;
concentration = 21 ± 2 μM; diameter Ø = 2.52 ± 0.03 nm) were pur-
for good electron-catalytic activity of the QDs is their small particle size chased from DropSens (Llanera, Spain).
and its quantum cavity. The band gap of the QDs determines the che- For electrochemical measurements different buffer solutions such as
mical energy which is present at the QDs under their excited status acetate (pH 3.7–5.7), phosphate (pH 2.0; 3.0; 6.0–8.0) and Britton-
[25]. As a semiconductor, the CdSe have a three-dimensional quantum- Robinson (BR) buffer (pH 2.0–11.0) were used. Reagents were of ana-
size effect which is the result of electron and hole localization over a lytical grade and buffer solutions were prepared by using distilled
confined space. The outcome of this process is the quantization of water. CH3COOH, H3PO4, methanol were obtained from
electron energy levels, which leads to an increase in the optical band Sigma–Aldrich. NaH2PO4·2H2O and Na2HPO4 were purchased from
gap of the CdSe [26]. Riedel-de Haen. N,N-dimethylformamide (DMF) was obtained from
Various methods have been developed for CLP determination, in- Merck.
cluding high performance liquid chromatography [27–37], liquid
chromatography with tandem mass spectrometry [38–41], ultra-per- 2.3. Preparation of modified fMWCNT/CdSeQDs/GCE
formance liquid chromatography with tandem mass spectrometry [42],
high performance thin layer chromatography [43], spectrophotometry Homogenous and stable fMWCNT suspension was prepared in DMF
[44–46], potentiometry [30] and voltammetry [20,31–33]. Although in the concentration of 0.5 mg/mL and sonicated for 2 h. Bare GCE
chromatographic and spectrophotometric methods have been widely surface was cleaned mechanically with alumina slurry on a polishing
performed for various analyses with high sensitivity and good se- pad, rinsed with distilled water and dried in air.
lectivity; they have some limitations, such as costly equipment, large fMWCNT/CdSeQDs suspension was prepared by mixing 200 μL
organic solvent consumption and time-consuming pre-treatment re- fMWCNT and 50 μL CdSeQDs (dispersed in chloroform by commer-
quirements. cially) under ultrasonic stirring during 30 min. The volume of
The voltammetric methods for analysis can be performed with rapid fMWCNT/CdSeQDs suspension (4:1) was dropped as 5 μL onto the
response, no pre-treatment and low organic solvent consumption [19]. cleaned GCE and the modified electrode was kept at room temperature
In addition to these superiorities, electrode modification with nano- for 2 h before use. The modified electrode was cleaned electro-
sized materials also contributes to sensitivity and selectivity of vol- chemically with cyclic voltammetry (6 cycles) between measurements
tammetric methods and these modified electrodes are widely used for and the electrode was prepared daily.
52
G. Ozcelikay et al. Journal of Electroanalytical Chemistry 827 (2018) 51–57
Fig. 2. SEM images of bare GCE (A) and CdSeQDs/GCE (B). fMWCNT/GCE (C). fMWCNT/CdSeQDs/GCE (D) Scale bars of SEM images are in 300 nm. SEM-EDX
images of fMWCNT/CdSeQDs/GCE (E).
2.4. Tablet and recovery assay procedure 3000 rpm. From this human serum stock, desired amount of CLP was
added to the pH 2.14 phosphate buffer solution while keeping the 20%
Five Atervix® tablets were taken and crushed into a fine powder in a methanol amount as constant. The calibration curve related with the
mortar. A required amount of this powder, which is enough for pre- increasing concentrations of CLP was obtained and the recovery studies
paring 1.0 × 10−3 M stock solution, was accurately weighed, dissolved were achieved using the obtained calibration data.
in methanol and sonicated. Solutions for analyzing were prepared by
taking aliquots of the clear supernatant of tablet solution and diluting 3. Results and discussion
with pH 2.14 phosphate buffer solutions. In order to investigate the
tablet excipients effect, the accuracy of the technique, known amounts 3.1. Effect of modification
of the pure drug was added into the analyzed tablet solution.
AdSDP voltammograms of 1 × 10−5 M CLP in pH 2.14 phosphate
2.5. Determination of CLP in synthetic serum sample buffer solution for the bare GCE, CdSeQDs/GCE, fMWCNT/GCE and
fMWCNT/CdSeQDs/GCE were obtained to investigate the electro-
In order to analyze CLP in synthetic serum sample, 3.6 mL serum chemical behavior of CLP (Fig. 1). Modification by using only CdSeQDs
sample was mixed with 5.4 mL of acetonitrile and 1 mL of 1 × 10−4 M did not enhance the signal of CLP when compared to bare GCE. But
CLP standard sample. The sample was centrifuged for 20 min at modification only with fMWCNT increased the peak current of CLP and
53
G. Ozcelikay et al. Journal of Electroanalytical Chemistry 827 (2018) 51–57
54
G. Ozcelikay et al. Journal of Electroanalytical Chemistry 827 (2018) 51–57
of fMWCNT and CdSeQDs (Fig. 2D) cover the surface of the GCE
completely. Fig. 2D represents the SEM images of fMWCNT, CdSeQDs
and rough, compact morphology of fMWCNT:CdSeQDs (4:1). In this
figure, fMWCNT solution is more intense than CdSeQDs solution and
fMWCNT covers CdSeQDs surface. The prepared electrodes were also
characterized by SEM EDX analyses. According to Fig. 2E, the EDX
results confirmed that the carbon amount in the surface originates from
CNTs, Cd and Se peaks clearly demonstrate the existence of the
CdSeQDs.
AFM images fMWCNT-CdSeQDs composite films are shown in Fig. 3
which indicated that CdSeQDs were located on the fMWCNTs and the
size distribution of CdSeQDs was quite uniform.
The best peak shape and highest current value were obtained when
phosphate buffer solution at pH 2.14 was used. For further studies
pH 2.14 phosphate buffer solution was chosen as the supporting elec-
trolyte.
The relationship between peak current and scan rate was in-
vestigated to obtain information about electrochemical mechanism in-
tensely by cyclic voltammetry. The peak currents were followed within
the ranges of 5 to 1000 mV s−1 on f MWCNT/CdSeQDs/GCE in phos-
Fig. 6. Increasing concentrations of CLP for bulk solutions (A): (1) blank; (2) phate buffer solution with pH 2.14 for 1 × 10−5 M CLP.
4.00 × 10−6 M; (3) 6.00 × 10−6 M; (4) 8.00 × 10−6 M; (5) 1.00 × 10−5 M
It is found that the anodic current value related with CLP was linear
and for serum solutions (B): (1) blank; (2) 7.50 × 10−6 M; (3) 1.00 × 10−5 M;
to the scan rate applied with the equation below:
(4) 1.25 × 10−6 M in pH 2.14 phosphate buffer solution.
i p (μA) = 0.0126 v (mVs−1) + 0.4345; r = 0.997 (n = 10) at f MWCNT
Table 1 /CdSeQDs/GCE
Regression data of the calibration curves for quantitative determination of CLP
by AdSDPV. The logarithm of oxidation peak currents versus scan rates exhibited
a linear relationship with a slope of 0.751 which can be expressed by
Standard solution Serum
mixture of diffusion and adsorption controlled process [55] according
Measured Potential (V) 0.93 0.92 to following equation:
Linearity range (M) 2 × 10−6–4 × 10−5 2.50 × 10−6–1.5 × 10−5
Slope (μA M−1) 2.21 × 104 2.77 × 104 log i p (μA) = 0.7512 log v (mVs−1) − 1.1962, r
Intercept (μA) 0.01 −0.06
Correlation coefficient 0.99 0.99 = 0.990 (n = 10) at f MWCNT/CdSeQDs/GCE
LOD (M) 7.55 × 10−8 2.99 × 10−7
LOQ (M) 2.51 × 10−7 9.97 × 10−7
Within day precision of peak 0.34 1.04 3.4. Accumulation time and potential optimization
current (RSD%)a
Within day precision of peak 0.38 0.40 Accumulation potential (Eacc) and accumulation time (tacc) values
potential (RSD%)a
Between days precision of 0.94 1.19
for AdSDPV were optimized in order to obtain best conditions for the
peak current (RSD%)a analyses of CLP using fMWCNT/CdSeQDs/GCE. The effect of the Eacc on
Between days precision of 0.77 0.93 the peak current was studied for 1 × 10−5 M CLP between 0 V and
peak potential (RSD%)a 0.6 V while the tacc was 60 s. According to Fig. 5, although the highest
response was obtained at 0.15 V, Eacc value was selected as 0 V because
a
Obtained from five measurements.
of repeatability. The effect of the tacc on the peak current was studied
for 1 × 10−5 M CLP between 5 s and 90 s and the best response was
In order to modify the surface of GCE, fMWCNTs and CdSeQDs were
obtained as 60 s. Further voltammetric analysis were carried out in
mixed up in different ratios before dropping. It can be clearly seen from
pH 2.14 phosphate buffer solution using 0 V as accumulation potential
the SEM images that CdSeQDs (Fig. 2B), fMWCNT (Fig. 2C) and mixture
and 60 s as accumulation time.
55
G. Ozcelikay et al. Journal of Electroanalytical Chemistry 827 (2018) 51–57
Table 2
Comparison of voltammetric studies for the determination of CLP.
Electrode Electrochemical behavior Technique Linearity range (μM) LOD (μM) Reference
Table 3 4. Conclusion
Results obtained for CLP determination and recovery from Atervix® tablet and
spiked serum by AdSDPV. In this study, CLP was analyzed using fMWCNTs and CdSeQDs
Atervix® Serum modified GCE. The volume of fMWCNT/CdSeQDs suspension ratio was
found as 4:1, and from this suspension 5 μL was used to modify the GCE.
Labeled claim (mg) 75.00 – The obtained results show that the fMWCNTs/CdSeQDs/GCE exhibited
Amount found (mg)a 74.40 –
good electrocatalytic performance toward CLP oxidation. The oxidation
RSD% 1.09 –
Bias% 0.85 – process exhibited mixed diffusion under adsorption controlled process
Added (mg) 20.00 4.00 depending on pH. The designed nanosensor was applied for the de-
Found (mg)a 19.87 3.61 termination of CLP in human serum and pharmaceutical dosage forms,
Average recovery % 99.33 90.20 successfully. The suggested electrochemical analysis enabled to analyze
RSD% of recovery 0.57 0.26
Bias% 0.67 9.80
CLP from blood serum, although the pH of human blood is 7.4 and the
analyses can be conducted at pH 2.14 phosphate buffer. In Table 1,
a
Obtained from three measurements. obtained results were compared with the literature studies according to
electrochemical behavior, linearity range and LOD values. The designed
3.5. Analytical characterization and validation of the method fMWCNTs/CdSeQDs/GCE sensor gave 0.075 μM level LOD value which
is lower than the other published voltammetric techniques in anodic
Under the optimum conditions, quantitative evaluation of CLP was way.
achieved based on the linear correlation between the peak current and
concentration. Fig. 6A showed some of the voltammograms from linear References
range between 2.0 × 10−6 and 4.0 × 10−5 M CLP and the linearity was
obtained with the following equation; [1] S.R. Mehta, S. Yusuf, R.J. Peters, M.E. Bertrand, B.S. Lewis, M.K. Natarajan,
K. Malmberg, H. Rupprecht, F. Zhao, S. Chrolavicius, I. Copland, K.A. Fox, Effects of
i p (μA) = 2.187 × 10 4C (M) + 0.0175; r = 0.999 (n = 7) pretreatment with clopidogrel and aspirin followed by long-term therapy in patients
undergoing percutaneous coronary intervention: the PCI-CURE study, Lancet 358
Validation of the analytical procedures for the quantitative assay of the (2001) 527–533 http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&
CLP was examined via evaluation of the LOD, LOQ, repeatability, re- form=6&dopt=r&uid=11520521.
[2] D.S. Aschenbrenner, S.J. Venable, Drug Therapy in Nursing, third ed., Lippincott
producibility, recovery, precision and accuracy. LOD and LOQ were calcu- Williams & Wilkins, Philadelphia, 2009.
lated from the equations of LOD = 3 s/m and LOQ = 10s/m [56,57] using [3] W.H. Frishman, A. Cheng-Lai, J. Nawarskas, Current Cardiovascular Drugs, third
the standard deviation of the response and the slope of the calibration curve ed., Current Medicine, New York, 2000, https://doi.org/10.1007/978-1-4615-
6767-7.
as 7.55 × 10−8 M and 2.51 × 10−7 M, respectively. Characteristics and the [4] P. Savi, J.M. Pereillo, M.F. Uzabiaga, J. Combalbert, C. Picard, J.P. Maffrand,
details of the validation parameters were reported in Table 1. M. Pascal, J.M. Herbert, Identification and biological activity of the active meta-
It can be clearly seen from the Table 2 that, the LOD value of the bolite of clopidogrel, Thromb. Haemost. 84 (2000) 891–896 (doi:00110891 [pii]).
[5] H. Schühlen, A. Kastrati, J. Dirschinger, J. Hausleiter, S. Elezi, A. Wehinger,
designed fMWCNTs/CdSeQDs/GCE nanosensor is quite low, 0.075 μM J. Pache, M. Hadamitzky, A. Schömig, Intracoronary stenting and risk for major
level, lower that the other voltammetric techniques in anodic way. The adverse cardiac events during the first month, Circulation 98 (1998) 104–111,
linearity range is quite broad as it is compared with the other studies https://doi.org/10.1161/01.CIR.98.2.104.
[6] M. Pasandideh-Nadamani, A. Omrani, M.R. Sadeghi-Maleki, A. Samadi-Maybodi,
especially the study achieved by Alghamdi [37] where the nearly same Application of CdS quantum dots modified carbon paste electrode for monitoring
LOD value was obtained. the process of acetaminophen preparation, Anal. Biochem. 502 (2016) 36–42,
https://doi.org/10.1016/j.ab.2016.02.015.
[7] W. Yang, P. Thordarson, J.J. Gooding, S.P. Ringer, F. Braet, Carbon nanotubes for
3.6. Determination of CLP in tablet dosage form and spiked serum sample
biological and biomedical applications, IOP Publ. Nanotechnol. Nanotechnol. 18
using the fMWCNT/CdSeQDs/GCE (2007) 412001–412012, https://doi.org/10.1088/0957-4484/18/41/412001.
[8] R.H. Baughman, A.A. Zakhidov, W.A. de Heer, Carbon nanotubes–the route toward
applications, Science 297 (2002) 787–792, https://doi.org/10.1126/science.
The developed AdSDPV technique was applied to the direct de-
1060928.
termination of CLP in tablet dosage form Atervix® tablet and spiked [9] C. Yang, M.E. Denno, P. Pyakurel, B.J. Venton, Recent trends in carbon nanoma-
serum by AdSDPV, using the related calibration equations. terial-based electrochemical sensors for biomolecules: a review, Anal. Chim. Acta
For human serum applications a calibration graph was described 887 (2015) 17–37, https://doi.org/10.1016/j.aca.2015.05.049.
[10] J. Wang, Carbon-nanotube based electrochemical biosensors: a review,
using the serum stock solution of 1 × 10−4 M. Between 2.5 × 10−6 and Electroanalysis 17 (2005) 7–14, https://doi.org/10.1002/elan.200403113.
1.5 × 10−5 M CLP, the linearity was obtained with the equation below [11] P. Pandey, M. Datta, B.D. Malhotra, Prospects of nanomaterials in biosensors, Anal.
and selected voltammograms of different concentrations were given in Lett. 41 (2008) 159–209, https://doi.org/10.1080/00032710701792620.
[12] H. Fecht, K. Brühne, P. Gluche, Carbon-based Nanomaterials and Hybrids, (2014),
Fig. 6B. https://doi.org/10.1201/b15673.
[13] Carbon-based nanomaterials, Essentials Nanosci. Nanotechnol, John Wiley & Sons,
i p (μA) = 2.77 × 10 4C (M)–0.055; r = 0.993 (n = 6)
Inc, Hoboken, NJ, USA, 2016, pp. 189–236, , https://doi.org/10.1002/
9781119096122.ch5.
The recovery studies, which were also applied for both determina- [14] A. Chen, S. Chatterjee, Nanomaterials based electrochemical sensors for biomedical
tion of CLP in spiked serum and tablet dosage form by AdSDPV, were applications, Chem. Soc. Rev. 42 (2013) 5425, https://doi.org/10.1039/
reported in Table 3. The recovery % and Bias % were all in acceptable c3cs35518g.
[15] B. Perez-Lopez, A. Merkoçi, Nanomaterials based biosensors for food analysis
ranges.
56
G. Ozcelikay et al. Journal of Electroanalytical Chemistry 827 (2018) 51–57
applications, Trends Food Sci. Technol. 22 (2011) 625–639, https://doi.org/10. estimation of clopidogrel bisulfate in human plasma by RP-HPLC, Int. J. Appl.
1016/j.tifs.2011.04.001. Pharm. 8 (2016) 18–21, https://doi.org/10.22159/ijap.2016v8i4.13521.
[16] J. Wang, Nanomaterial-based electrochemical biosensors, Analyst 130 (2005) 421, [37] A.F. Alghamdi, Electrochemical and chromatographic studies of clopidogrel using
https://doi.org/10.1039/b414248a. cathodic stripping voltammetry and HPLC under new experimental conditions and
[17] G. Hong, S. Diao, A.L. Antaris, H. Dai, Carbon nanomaterials for biological imaging its determination in the preparation tablet, urine and plasma samples, J. Chem.
and nanomedicinal therapy, Chem. Rev. 115 (2015) 10816–10906, https://doi.org/ Pharm. Res. 7 (2015) 1023–1030 http://www.jocpr.com/articles/electrochemical-
10.1021/acs.chemrev.5b00008. and-chromatographic-studies-of-clopidogrel-using-cathodic-stripping-voltammetry-
[18] B.S. Murty, P. Shankar, B. Raj, B.B. Rath, J. Murday, Applications of Nanomaterials, and-hplc-under-new-exper.pdf (accessed January 16, 2018).
(2013), https://doi.org/10.1142/9789814340571_0009. [38] N. Bibire, M. Vieriu, G. Tantaru, M. Apostu, L. Agoroaei, A.D. Panainte,
[19] Z. Liu, M. Jin, J. Cao, J. Wang, X. Wang, G. Zhou, A. Van Den Berg, L. Shui, High- A. Znagovan, A. Vlase, A new and sensitive LC-MS/MS method for the determina-
sensitive electrochemical sensor for determination of Norfloxacin and its metabo- tion of clopidogrel in human plasma, Rev. Chim. 65 (2014) 807–810 http://www.
lism using MWCNT-CPE/pRGO-ANSA/Au, Sensors Actuators B 257 (2018) revistadechimie.ro , Accessed date: 1 February 2018.
1065–1075, https://doi.org/10.1016/j.snb.2017.11.052. [39] B. Shin, S. Yoo, Determination of clopidogrel in human plasma by liquid chroma-
[20] J. Xu, Z. Cao, Y. Zhang, Z. Yuan, Z. Lou, X. Xu, X. Wang, A review of functionalized tography/tandem mass spectrometry: application to a clinical pharmacokinetic
carbon nanotubes and graphene for heavy metal adsorption from water: prepara- study, Biomed. Chromatogr. 21 (2007) 883–889, https://doi.org/10.1002/bmc.
tion, application, and mechanism, Chemosphere 195 (2018) 351–364, https://doi. 850.
org/10.1016/J.CHEMOSPHERE.2017.12.061. [40] A. Robinson, J. Hillis, C. Neal, A.C. Leary, The validation of a bioanalytical method
[21] Y. Tu, Q. Xu, Q.-J. Zou, Z.-H. Yin, Y.-Y. Sun, Y.-D. Zhao, Electrochemical behavior of for the determination of clopidogrel in human plasma, J. Chromatogr. B 848 (2007)
levodopa at multi-wall carbon nanotubes-quantum dots modified glassy carbon 344–354, https://doi.org/10.1016/J.JCHROMB.2006.10.076.
electrodes, Anal. Sci. 23 (2007) 1321–1324 https://www.jstage.jst.go.jp/article/ [41] L. Silvestro, M.C. Gheorghe, I. Tarcomnicu, S. Savu, S.R. Savu, A. Iordachescu,
analsci/23/11/23_11_1321/_pdf , Accessed date: 29 January 2018. C. Dulea, Development and validation of an HPLC-MS/MS method to determine
[22] H. Kang, L. Wang, M. O'Donoghue, Y.C. Cao, W. Tan, Nanoparticles for biosensors, clopidogrel in human plasma. Use of incurred samples to test back-conversion, J.
Opt. Biosens, 2008, https://doi.org/10.1016/B978-044453125-4.50017-6. Chromatogr. B 878 (2010) 3134–3142, https://doi.org/10.1016/j.jchromb.2010.
[23] C.N.R. Rao, A. Mu, The Chemistry of Nanomaterials, (2004), https://doi.org/10. 09.022.
1002/352760247X. [42] X. Yang, S. Liu, J. Sun, X. Liu, Y. Sun, Z. He, UPLC for the determination of clo-
[24] X. Yao, P. Yan, Q. Tang, A. Deng, J. Li, Quantum dots based electro- pidogrel in dog plasma by tandem quadrupole MS: application to a pharmacokinetic
chemiluminescent immunosensor by coupling enzymatic amplification for ultra- study, Chromatographia 70 (2009) 259–263, https://doi.org/10.1365/s10337-009-
sensitive detection of clenbuterol, Anal. Chim. Acta 798 (2013) 82–88, https://doi. 1109-9.
org/10.1016/j.aca.2013.08.029. [43] H. Agrawal, N. Kaul, A. Paradkar, K. Mahadik, Stability indicating HPTLC de-
[25] S. Hooshmand, Z. Es'haghi, Simultaneous quantification of arginine, alanine, me- termination of clopidogrel bisulphate as bulk drug and in pharmaceutical dosage
thionine and cysteine amino acids in supplements using a novel bioelectro-nano- form, Talanta 61 (2003) 581–589, https://doi.org/10.1016/S0039-9140(03)
sensor based on CdSe quantum dot/modified carbon nanotube hollow fiber pencil 00364-3.
graphite electrode via Taguchi method, J. Pharm. Biomed. Anal. 146 (2017) [44] H.E. Zaazaa, S.S. Abbas, M. Abdelkawy, M.M. Abdelrahman, Spectrophotometric
226–235, https://doi.org/10.1016/j.jpba.2017.08.034. and spectrodensitometric determination of clopidogrel bisulfate with kinetic study
[26] Y.M. Aniskevich, M.V. Malashchonak, P.V. Chulkin, G.A. Ragoisha, E.A. Streltsov, of its alkaline degradation, Talanta 78 (2009) 874–884, https://doi.org/10.1016/J.
Cadmium underpotential deposition on CdSe and CdS quantum dot films: size de- TALANTA.2008.12.064.
pendent underpotential shift, Electrochim. Acta 220 (2016) 493–499, https://doi. [45] S. Gurav, R. Venkatamahesh, Development and validation of derivative UV-spec-
org/10.1016/j.electacta.2016.10.132. tropotometric methods for quantitative estimation of clopidogrel in bulk and
[27] V. Sivarama Krishna, D.R. Kumar, K. Balamuralikrishna, C. Rambabu, Development pharmaceutical dosage form, Int. J. ChemTech Res. 4 (2012) 497–501 http://www.
and validation of stability indicating RP-HPLC method for the determination of sphinxsai.com/2012/chemAJ/CHEM/CT=05[497-501]AJ12.pdf , Accessed date:
clopidogrel bisulphate in bulk and its dosage forms, Der Pharma Chem. 6 (2014) 29 January 2018.
366–374 http://derpharmachemica.com/archive.html , Accessed date: 29 January [46] L. Antypenko, S. Gladysheva, S. Vasyuk, Development and validation of clopidogrel
2018. bisulphate determination in bulk by UV spectrophotometric method, Scr. Sci.
[28] J. Sippel, L.L. Sfair, E.E. Schapoval, M. Steppe, New high-performance liquid Pharm. 3 (2016) 17–22.
chromatographic method for determination of clopidogrel in coated tablets, J. [47] S. Dermiş, E. Aydoǧan, Electrochemical study of the antiplatelet agent clopidogrel
AOAC Int. 91 (2008) 67–73. and its determination using differential pulse voltammetry in bulk form and
[29] M.K. Javed, Z. Iqbal, A. Khan, A. Khan, Y. Shah, L. Ahmad, Development and va- pharmaceutical preparations with a glassy carbon electrode, Pharmazie 65 (2010)
lidation of HPLC-UV method for the determination of clopidogrel in pharmaceutical 175–181, https://doi.org/10.1691/ph.2010.9123.
dosage form and human plasma, J. Liq. Chromatogr. Relat. Technol. 34 (2011) [48] A.R. Mladenović, V.M. Jovanović, S.D. Petrović, D.Ž. Mijin, S.Ž. Drmanić,
2118–2129, https://doi.org/10.1080/10826076.2011.585482. M.L. Avramov Ivić, Determination of clopidogrel using square wave voltammetry at
[30] A.L. Saber, M.A. Elmosallamy, A.A. AmIn, H.M.A. Killa, Liquid chromatographic a gold electrode, J. Serb. Chem. Soc. 78 (2013) 2131–2140, https://doi.org/10.
and potentiometric methods for determinations of clopidogrel, J. Food Drug Anal. 2298/JSC130913116M.
16 (2008) 11–18 http://search.proquest.com/openview/ [49] Z.R. Dizavandi, A. Aliakbar, M. Sheykhan, Electrocatalytic determination of clopi-
686fa68d7f1a1117bfbb55e59ee327e7/1?pq-origsite=gscholar&cbl=906352. dogrel using Bi2O3-Pp-AP/GCE by differential pulse voltammetry in pharmaceu-
[31] F. Belal, A. El-Brashy, M. Eid, J.J. Nasr, Stability-indicating micellar liquid chro- tical productions, J. Electroanal. Chem. 805 (2017) 24–31, https://doi.org/10.
matographic method for the determination of clopidogrel. Application to tablets 1016/J.JELECHEM.2017.10.013.
and content uniformity testing, J. Liq. Chromatogr. Relat. Technol. 32 (2009) [50] R. Greef, R. Peat, L.M. Peter, D. Pletcher, Instrumental Methods in Electrochemistry,
2993–3008, https://doi.org/10.1080/10826070903320764. Ellis Horwood, 1990.
[32] A. Mitakos, I. Panderi, A validated LC method for the determination of clopidogrel [51] S.A. Ozkan, B. Uslu, From mercury to nanosensors: past, present and the future
in pharmaceutical preparations, J. Pharm. Biomed. Anal. 28 (2002) 431–438, perspective of electrochemistry in pharmaceutical and biomedical analysis, J.
https://doi.org/10.1016/S0731-7085(01)00620-3. Pharm. Biomed. Anal. 130 (2016) 126–140, https://doi.org/10.1016/j.jpba.2016.
[33] U.C. Mashelkar, S.D. Renapurkar, A LCMS compatible stability-indicating HPLC 05.006.
assay method for clopidogrel bisulphate, Int. J. ChemTech Res. 2 (2010) 822–829 [52] J. Wang, Analytical Electrochemistry, Third (2006), https://doi.org/10.1002/
http://sphinxsai.com/s_v2_n2/CT_V.2No.2/ChemTech_Vol_2No.2_pdf/CT= 0471790303.
10(822-829).pdf , Accessed date: 29 January 2018. [53] C.M.A. Brett, Electrochemistry. Principles, Methods and Applications, Oxford
[34] N.A. Alarfaj, Stability-indicating liquid chromatography for determination of clo- University Press, 1993.
pidogrel bisulfate in tablets: application to content uniformity testing, J. Saudi [54] ICH Expert Working Group, ICH guideline Q1A(R2) stability testing of new drug
Chem. Soc. 16 (2012) 23–30, https://doi.org/10.1016/j.jscs.2010.10.016. substances and products, Int. Conf. Harmon. 24 (2003), https://doi.org/10.1136/
[35] S. Sahu, S.P. Sarangi, H.B. Sahoo, Development and validation methods for the bmj.333.7574.873-a.
estimation of Clopidogrel in bulk and pharmaceutical dosage forms, Int. J. Res. [55] D.K. Gosser, Cyclic Voltammetry, VCH, 1994.
Pharm. Sci. 3 (2012) 224–227 https://www.researchgate.net/profile/Dr_ [56] S.A. Ozkan, Electroanalytical Methods in Pharmaceutical Analysis and Their
Himanshu_Sahoo/publication/268187036_Development_and_validation_methods_ Validation, first ed., HNB Pub, New York, NY, 2012.
for_the_estimation_of_Clopidogrel_in_bulk_and_pharmaceutical_dosage_forms/links/ [57] S.A. Ozkan, J.-M. Kauffmann, P. Zuman, Electroanalysis in Biomedical and
547824960cf293e2da284645/Development-and-validation-methods-for , Accessed Pharmaceutical Sciences, first ed., Springer, Berlin Heidelberg, 2015, https://doi.
date: 29 January 2018. org/10.1007/978-3-662-47138-8.
[36] H. Kumar Jain, D.D. Deore, Bioanalytical method development and validation for
57
Contributing Editor
ABDULLAH A. AL-BADR
Founding Editor
KLAUS FLOREY