Analytical

Methods
View Article Online
TECHNICAL NOTE View Journal | View Issue

Simultaneous standard addition method for novel

determination of components in a single step:
Published on 14 April 2014. Downloaded by University of Sydney on 04/09/2014 06:46:20.

Cite this: Anal. Methods, 2014, 6, 6110

application in analysis of Sunset yellow and
Carmoisine by a spectrophotometric technique
Karim Asadpour-Zeynali* and Sara Manafi-Khoshmanesh

Recently a novel method for determination of an analyte in the presence of known interferences called the
net analyte signal standard addition method (NASSAM) was introduced. Although NASSAM was used for
individual standard addition, a novel generalized net analyte signal standard addition method (GNASSAM)
was used for simultaneous standard addition. By applying the proposed method, simultaneous
determination of components was done in a single step. Also, this method can be applied for linearly
dependent concentrations. Sunset yellow and Carmoisine demonstrate strong spectra overlap with
typical application of a spectrophotometry technique. GNASSAM was used for simultaneous
determination of Sunset yellow and Carmoisine in synthetic binary mixtures and real samples. The
investigated method was also used to calculate figures of merit; moreover, good limits of detection
Received 28th February 2014
Accepted 11th April 2014
(SY ¼ 0.15 and CA ¼ 0.18 mg L1) and suitable selectivity and sensitivity were determined. A HPLC
method was applied for finding the subspace in binary mixtures of real samples. The obtained results for
DOI: 10.1039/c4ay00507d
real samples were compared with the HPLC results. The obtained results were in good agreement with
www.rsc.org/methods those of the HPLC method.

vector is unique for the analyte of interest, and is orthogonal to

1. Introduction the spectra of the interferences.4 There are four net analyte
It has become increasingly common to use computer assistance signal based methods that have been reported in the literature:
for analysis of data in chemical laboratories in recent years. (1) hybrid linear analysis (HLA),5 which can be applied when a
Chemometrics methods are based on linear algebra and very accurately measured pure spectrum of the analyte is avail-
multivariate statistics, which are applied for multicomponent able; (2) the HLA method developed by Xu and Schechter (HAL/
quantitative analysis.1 XS),1 which uses all the factors for prediction in order to build a
There are two possible modes of calibration; in the classical method free from optimum factor estimation; (3) the HLA
model, the instrument responses are declared as a function of method developed by Goicoechea and Olivieri (HLA/GO), which
analyte concentration, whereas in the inverse model, the role of extracts interferent subspace and removes the analyte portion,6
the variables is reversed. Although in the inverse model the and (4) other NAS-based multivariate calibration methods in
interferences are implicitly modeled, in the classical model, it's which the vectors of the NAS of mixtures are used as an input for
important to explicitly estimate the parameters that are other multivariate calibration methods such as classical least
describing all the interferences.2 squares (NAS/CLS), principal component regression (NAS/PCR)
Despite multivariate methods having some advantages, such and partial least squares regression (NAS/PLS).7–10
as using full spectra or requiring the concentration of the If there are two distinguishable sensors with selective
analytes of interest in the calibration samples,3 the major responses for two analytes, it is possible to determine the
drawback of this approach is the need for manual intervention concentrations of both analytes of interest in one sample by
in the selection of a number of factors. preparing single standard mixture solutions, which contain
The net analyte signal (NAS) is a new type of multivariate both analytes, and carrying out a standard addition procedure.
calibration method that has been dened by Lorber as the part That is an ideal situation; however, the condition of a selective
of a mixture spectrum that is useful for model building. The NAS response for each analyte is oen not fullled in several cases.
The H-point standard addition method (HPSAM), which was
presented in 1988 by Bosch-Reig and Campins-Falco,11 is able to
Department of Analytical Chemistry, Faculty of Chemistry, University of Tabriz, Tabriz obtain an unbiased analyte concentration when both analyte
5166616471, Iran. E-mail: asadpour@tabrizu.ac.ir; k.zeynali@gmail.com; Fax: +98 and interferents are present in a sample. HPSAM uses the
411 3340191; Tel: +98 411 3393113

6110 | Anal. Methods, 2014, 6, 6110–6115 This journal is © The Royal Society of Chemistry 2014
 View Article Online

Technical Note Analytical Methods

analytical signal data at two selected wavelengths. Two straight 2. Materials and methods
lines that have a common point with coordinates H (:CH, AH)
are given, where :CH is the unknown analyte concentration and 2.1. Apparatus and soware
AH is the analytical signal for an interferent. In addition, at the Absorption measurements were carried out using a UNICO
selected wavelengths the analytical signals of the interferent (Dayton, NJ) 2100UV spectrophotometer with a 1.0 cm quartz
should be constant, while for the analyte it should be different cell. All spectra were saved in ASCII format and all data were
as much as possible. transformed to Excel and MATLAB formats. All spectropho-
Saxberg and Kowalski published a supplement of the stan- tometry data were transferred to a personal computer (PC) for
dard addition method for multivariate data, which was identi- subsequent manipulation. All pH measurements were done
ed as the generalized standard addition method (GSAM).12 The with a digital pH meter (Metrohm, Riverview, FL; Model 744).
Published on 14 April 2014. Downloaded by University of Sydney on 04/09/2014 06:46:20.

method requires that the analyte and the interferents be The data were handled by using MATLAB soware (2008
sequentially added in the sample. This relaxed the main version). All necessary programs needed for NAS calculation
restriction of analytical methods, that they must be fully selec- were written in MATLAB and run on a windows seven operating
tive to the analyte of interest. system. The HPLC measurements were obtained by using a
Recently a novel method for determination of an analyte in KNAUER high performance liquid chromatograph (KNAUER,
the presence of known interferences called the “net analyte Berlin, Germany) equipped with a binary pump, a Sphere-Image
signal standard addition method (NASSAM)” was developed.13–16 ODS column (250*4 mm, 5 mm) and a smart line UV detector.
The method uses the NAS to construct a univariate calibration
model and it could also eliminate the constraints of HPSAM. 2.2. Reagents and solutions
With application of NASSAM the calibration and prediction
steps of multivariate calibration methods are eliminated and All experiments were performed with analytical grade chemicals
the determination is carried out in a single step for each analyte. and solvents. Stock solutions (1000 mg L1) of Carmoisine and
In the present study a generalized net analyte signal stan- Sunset yellow (Aldrich) were prepared by weighing suitable
dard addition method (GNASSAM) is performed. It is similar to amounts of their powders and dissolving them in a volumetric
NASSAM, except that it is carried out using simultaneous ask with deionized water. The solutions were stored in a
standard addition for all analytes. Consumption of less sample refrigerator at 4  C; at this temperature CA and SY were stable
for several analytes compared to other routine standard addi- for at least 1 month. Working solutions with the required
tion methods is the main advantage of the offered procedure. In concentrations were prepared daily in phosphate buffer
other words, only one standard addition procedure was per- (0.04 mol L1, pH ¼ 7).
formed for analysis of multianalyte samples, therefore time and Britton–Robinson buffer solutions were prepared by dis-
sample size were saved. In some of the multivariate calibration solving appropriate amounts of boric acid, ortho-phosphoric
methods, such as partial least squares (PLS) and principal acid and acetic acid (glacial) in deionized water and adjusting
component regression (PCR), the concentration of the analytes the pH with sodium hydroxide (1 mol L1) to obtain a pH range
in the calibration step should be independent, but GNASSAM of 2–12. For pH 1–2 and 12–14, HCl and NaOH solutions were
can be applied for linearly dependent concentrations. used, respectively. Sodium phosphate buffer solution of pH ¼
In recent years there has been increasing interest in using 7 was prepared by dissolving a suitable amount of sodium
synthetic additives and food colours, especially azo dyes, in phosphate salt in water and adjusting its pH with sodium
commercial food products. Moreover, it is necessary to deter- hydroxide.
mine food dyes quantitatively and qualitatively in order to
comply with regulations in many countries. Carmoisine (diso- 2.3. Sample preparation for fruit candies
dium 4-hydroxy-2-[(E)-(4-sulfonato-1-naphthyl)diazenyl]naphtha- Thirty fruit candies (88 g) (Carnival, Shirin Asal, Ind. Co. Tabriz,
lene-1-sulfonate) and Sunset yellow (disodium 6-hydroxy-5-[(4- Iran) were ground in a mortar and dissolved in 250 ml of
sulfophenyl)azo]-2-naphthalenesulfonate) are two azo synthetic deionized water in a volumetric ask, then centrifuged and
dyes that are widely used in foodstuffs. Several studies have been ltered.
performed investigating new methods for analysis of colorants.
These methods utilize high performance liquid chromatog-
raphy,17–20 ion-pair liquid chromatography with photodiode array 2.4. Procedure
and electrospray mass spectroscopy detection,21 capillary elec- Sunset yellow and Carmoisine demonstrate strong overlap in
trophoresis22 and spectrophotometric methods.23–26 their ultraviolet-visible (UV-VIS) spectra (Fig. 1). pH changes can
In the present research, a novel method called the general- affect the spectra of dyes, therefore Carmoisine and Sunset
ized net analyte signal standard addition method (GNASSAM) is yellow solutions (20 mg L1) of different pH were prepared and
used. Utility of the method is illustrated by application in the effect of pH changes on their spectra were investigated. The
analytical determination of two common azo dyes, Sunset absorbances of CA and SY were stable at pH ¼ 4–8. Solutions
yellow (SY) and Carmoisine (CA), in fruit candies. The theory of were prepared in pH ¼ 7 for all the experiments described in
the NAS based methods used here has been described exten- this research. The net analyte signal standard addition method
sively in the literature,27–31 therefore in this paper we do not (NASSAM) was performed on synthetic Carmoisine and Sunset
describe the theory of them. yellow by preparing binary mixtures. In this method, the

This journal is © The Royal Society of Chemistry 2014 Anal. Methods, 2014, 6, 6110–6115 | 6111
 View Article Online

Analytical Methods Technical Note

each analyte in the samples were obtained by nding the

abscissa of the crossing point of the plotted line and the hori-
zontal axis (at ordinate equal to zero). For all samples, standard
addition procedures were replicated at least 3 times and the
average and standard deviation of the results were reported.

3. Results and discussion

The absorbance spectra of CA and SY at different pH were
veried and pH ¼ 4–8 was found to be appropriate for spectra
Published on 14 April 2014. Downloaded by University of Sydney on 04/09/2014 06:46:20.

Fig. 1 Absorption spectra of Carmoisine and Sunset yellow. The

concentrations were 20 mg L1. Investigation of the bilinear behavior
for a mixture of Carmoisine and Sunset yellow.

concentration of the analyte of interest (CA or SY) was varied,

while the concentration of the interferent (CA or SY) was xed
(the plots weren't reported here). For the generalized net analyte
signal standard addition method (GNASSAM) binary synthetic
mixtures of Carmoisine and Sunset yellow were prepared as
samples (concentration of CA and SY were 8 and 6 mg L1,
respectively). First the spectrophotometric absorption of each
sample was recorded between wavelengths 400 to 650 nm with 1
nm intervals. Second, standard addition procedures were per-
formed for each sample. Finally, the spectra of the samples were
recorded aer each addition. The concentration of the added
analytes for CA and SY were 0, 2, 4, 6 and 0, 3, 6, 9 mg L1,
respectively.
A sensor selection approach based on a moving spectral
window strategy, in which a search for the minimum Error
Indicator (EI) was conducted at all possible spectral ranges and
the best wavelength ranges were studied.30
All observations were recorded and aer calculation of the
NAS for each sample, a diagram of the NAS versus the concen-
tration of added standards was plotted. The concentrations of

Fig. 3 (A) Absorption spectra of the samples from Table 1: (a) before
and (b–d) after addition of 2, 4 and 6 mg L1 of Carmoisine and 3, 6 and
9 mg L1 of Sunset yellow. (B) Net analyte signal of Carmoisine (a)
before and (b–d) after standard addition. (C) Net analyte signal of
Sunset yellow (a) before and (b–d) after standard addition. (D)
Fig. 2Error indicator function for investigation of the best wavelength Diagrams of the NAS of Carmoisine and Sunset yellow versus the
range of CA and SY in GNASSAM. concentration of the added standards.

6112 | Anal. Methods, 2014, 6, 6110–6115 This journal is © The Royal Society of Chemistry 2014
 View Article Online

Technical Note Analytical Methods

recording. pH ¼ 7 was selected as the best pH for the solutions. the analytes were obtained from the data matrix. The NAS
Buffer solutions were made using phosphate salt with the pH curves for Carmoisine and Sunset yellow are shown in Fig. 3B
was adjusted to pH ¼ 7 by addition of sodium hydroxide. All of and C, respectively. The norm of all NAS vectors is proportional
the solutions were made utilizing this buffer. The areas of high to the analyte concentration. As a matter of fact, NAS vectors are
absorbance in the UV-VIS spectra of Carmoisine and Sunset univariate data. Furthermore, the norms of the NAS versus the
yellow overlapped. The spectra of Carmoisine and Sunset yellow concentrations of the analytes were plotted and linear graphs
are shown in Fig. 1. Determination of these colorants by were obtained (Fig. 3D). Not only can the concentration of the
univariate methods required chemical separation, due to their analytes be simultaneously obtained with a single step proce-
spectral overlap. The initial necessity for applying any bilinear dure by GNASSAM, but also the calibration and prediction
method was examined and the bilinear behavior of the spec- steps, which are usual in multivariate calibration methods, were
Published on 14 April 2014. Downloaded by University of Sydney on 04/09/2014 06:46:20.

trophotometric data of these dyes was conrmed. For this eliminated in GNASSAM, which can be observed from Fig. 3D.
purpose, the linearity and additivity of the spectrophotometric The position of the standard addition plot just depends on the
responses were investigated. The spectra of individual and concentration of the analyte. Further, the position is indepen-
mixture solutions and the sum of the spectra of Carmoisine and dent from the concentration of the interferent.
Sunset yellow were plotted. According to the obtained plots, The HAL/XS method uses all the factors for prediction in
there are no interactions between the analytes, and the signals order to build a method free from optimum factor estimation
indicate very good additive properties. and the HLA/GO method extracts interferent subspace with
removal of the analyte portion. These methods were investi-
3.1. Analysis of synthetic binary mixtures gated on synthetic mixtures.
The prediction error of a single component in the mixtures
In NASSAM one of the dyes (CA or SY) was chosen as the analyte, was calculated as the relative standard error of the prediction
while the other dye was considered as an interferent. Then, concentration:
known amounts of analyte standard solutions were added to the 0 N  2 11=2
X
samples, whereas the concentration of the interferent was xed. C^ j  Cj
B C
In GNASSAM both contributions are considered as analytes. B j¼1 C
R:S:E: ð%Þ ¼ BB X
C  100
For determination of the concentration of the analytes, known @
N
 2 C A
amounts of both analyte standard solutions were added to the Cj
j¼1
samples and the spectra of the mixtures were recorded. A
wavelength selection step was used to choose the optimum where N is the number of samples, Cj is the concentration of the
range of wavelengths. For this purpose, the correlation coeffi- component in the jth mixture and Ĉj is the estimated
cient of the tted straight line in the standard addition diagram concentration.
and also the value of the EI function were considered. Usage of The equations of the calibration graphs, correlation coeffi-
the EI function for wavelength selection showed that the nal cients, standard deviations, limit of detections (LODs) and
results were strictly affected by the selected range of wave- calculated selectivity & sensitivity for the synthetic binary
lengths (Fig. 2). mixtures of Carmoisine and Sunset yellow using the NASSAM,
The recorded spectra for binary mixtures of Carmoisine and GNASSAM, HLA-XS and HLA-GO methods are summarized in
Sunset yellow are illustrated in Fig. 3A. The solution spectra Table 1.27–31
were recorded in the range of 400–550 nm. Spectrum (a) in The RSE, selected wavelength ranges and calculated
Fig. 3A was considered as the unknown binary mixture and concentrations for Carmoisine and Sunset yellow using the
spectra (b–d) are those obtained aer adding known amounts of NASSAM, GNASSAM, HLA-XS and HLA-GO methods are
Carmoisine and Sunset yellow standard solutions. The NAS of summarized in Table 2.

Table 1 Comparison of NAS based methods for the determination of Carmoisine and Sunset Yellow (the analytical data were calculated from the
calibration graphs)

Parameter Analyte NASSAM HLA-XS HLA-GO GNASSAM

Equation of the CA Y ¼ 0.3691X  0.0119 — — Y ¼ 0.2876X  0.0136

calibration curve SY Y ¼ 0.3762X + 0.0084 — — Y ¼ 0.3762X + 0.0085
R2 CA 0.9999 0.999 0.999 0.9999
SY 0.9999 0.998 0.998 0.9999
SD CA — 0.022 0.016 0.0232
SY — 0.021 0.026 0.0480
LOD (mg L1) CA 0.18 0.07 0.05 0.18
SY 0.14 0.07 0.09 0.15
Selectivity CA 0.21 0.42 0.43 0.07
SY 0.25 0.71 0.61 0.11
Sensitivity CA 0.37 0.12 0.22 0.27
SY 0.38 0.23 0.23 0.37

This journal is © The Royal Society of Chemistry 2014 Anal. Methods, 2014, 6, 6110–6115 | 6113
 View Article Online

Analytical Methods Technical Note

Table 2 Comparison of the results of NAS based methods for analysis Table 3 Results of NASSAM and GNASSAM for real samplesa
of synthetic binary mixtures of Carmoisine and Sunset yellow
Concentration
Parameter Analyte NASSAM HLA-XS HLA-GO GNASSAM Method Analyte (mg L1) Recovery

Concentration CA — — — 8.01 (0.02) GNASSAM Carmoisine 9.96 100

(mg L1)  SY — — — 5.99 (0.05) Sunset yellow 8.99 99.9
(standard deviation) HPLC Carmoisine 9.96 —
%RSE CA 1.48 0.92 1.53 0.28 Sunset yellow 9.00
SY 1.86 2.77 2.92 0.66 NASSAM Carmoisine 9.95 99.9
l (nm) CA 400–550 — — 400–547 Sunset yellow 13.06 100.5
SY 400–550 — — 400–547 HPLC Carmoisine 9.96 —
Sunset yellow 13.00
Published on 14 April 2014. Downloaded by University of Sydney on 04/09/2014 06:46:20.

a
The concentrations of the added CA and SY standards in each step of the
standard addition procedure for the real samples were 0, 2, 4 and 6 mg L1.

3.2. Analysis of real sample mixtures

Fruit candies were ground in a mortar and dissolved in deion-
ized water in a volumetric ask, then centrifuged and ltered.
The fruit candies solution was tested and only Carmoisine was
observed, therefore a suitable amount of Sunset yellow that was
not in the real sample was spiked (13 mg L1 in NASSAM and
9 mg L1 in GNASSAM). Standard additions in the NASSAM and
GNASSAM methods were successfully performed. The concen-
trations of the standards in GNASSAM were 0, 2, 4 and 6 mg L1.
A sensor selection approach for searching for the minimum
error indicator (EI) was investigated for the real samples (the EI
plots were not shown here).
The recorded spectra of the real samples are illustrated in
Fig. 4A. Spectrum (a) in Fig. 4A was considered as the unknown
solution of the real sample and spectra (b–d) are those obtained
aer adding known amounts of Carmoisine and Sunset yellow
standard solutions. The NAS for the analytes were obtained
from the data matrix. The NAS curves for Carmoisine and
Sunset yellow are shown in Fig. 4B and C, respectively. The
norm of the NAS versus the concentration of the analyte was
plotted and linear graphs were obtained (Fig. 4D and E).
Because of the matrix effect, the slopes of the standard addition
plots for the synthetic and real examples are different.
In our proposed method, high performance liquid chroma-
tography (HPLC) was applied for nding the subspace in the
binary mixtures of the real samples. The absorbance spectra for
the concentration of the analyte, which were measured by
HPLC, were subtracted from the absorbance of the mixtures.
The remaining absorbance is the absorbance of the real sample
without analyte, which was considered as the subspace. In the
applied HPLC method, the mobile phase consisting of phos-
phate buffer (10 mM, pH ¼ 7) as solvent A and a mixture of
methanol : acetonitrile (80 : 20, v/v%) as solvent B was prepared
and a gradient elution program was applied. The HPLC results
Fig. 4 (A) Absorption spectra of real sample: (a) before and (b–d) after were compared with the results of the proposed method. The
addition of 2, 4 and 6 mg L1 of Carmoisine and 2, 4 and 6 mg L1 of results of the proposed method were in good agreement with
Sunset yellow. (B) The net analyte signal of Carmoisine (a) before and those of the HPLC method, as reported in Table 3.
(b–d) after standard addition. (C) The net analyte signal of Sunset
yellow (a) before and (b–d) after standard addition. (D) Diagrams of the
NAS of Carmoisine versus the concentration of the added Carmoisine 4. Conclusions
and (E) diagrams of the NAS of Sunset yellow versus the concentration
of the added Sunset yellow. A novel method named the generalized net analyte signal
standard addition method (GNASSAM) was presented for

6114 | Anal. Methods, 2014, 6, 6110–6115 This journal is © The Royal Society of Chemistry 2014
 View Article Online

Technical Note Analytical Methods

simultaneous determination of analytes that have spectra 12 B. E. H. Saxberg and B. R. Kowalski, Anal. Chem., 1979, 51,
overlap. GNASSAM shows some advantages in comparison with 1031.
other methods. It is performed in only one standard addition 13 K. Asadpour-Zeynali and M. Bastami, Spectrochim. Acta, Part
procedure for analysis of multianalyte samples, therefore time A, 2010, 75, 589.
and sample consumption were saved. What's more, it is 14 K. Asadpour-Zeynali, M. R. Majidi and M. Tahmasebpour,
possible to analyze data even when the concentrations are J. Serb. Chem. Soc., 2009, 74, 789.
linearly dependent. The selection of an appropriate signal range 15 K. Asadpour-Zeynali and P. Naja-Marandi, Electroanalysis,
also improves the results and this point should be considered in 2011, 23, 2241.
the analysis of data. 16 K. Asadpour-Zeynali, M. R. Majidi, J. Vallipour and S. Jafari,
J. Chin. Chem. Soc., 2011, 58, 353.
Published on 14 April 2014. Downloaded by University of Sydney on 04/09/2014 06:46:20.

17 D. Sumita, K. Subhash and D. Mukul, J. AOAC Int., 2010, 93,

Acknowledgements 1503.
18 Q. Chen, Sh. Mou, X. Hou, J. M. Riviello and Zh. Ni,
Financial support from the University of Tabriz is gratefully
J. Chromatogr. A, 1998, 827, 73.
acknowledged.
19 M. A. Prado and H. T. Godoy, J. Liq. Chromatogr. Relat.
Technol., 2002, 25, 2455.
References 20 N. Vachirapatama, J. Mahajaroensiri and W. Visessanguan,
J. Food Drug Anal., 2008, 16, 77.
1 L. Xu and I. Schechter, Anal. Chem., 1997, 69, 3722. 21 M. R. Fuh and K. J. Chia, Talanta, 2002, 56, 663.
2 N. M. Faber, Anal. Chem., 1998, 70, 5108. 22 R. A. Frazier, E. L. Inns, N. Dossi, J. M. Ames and
3 S. Wold, M. Sjostrom and L. Eriksson, Chemom. Intell. Lab. H. E. Nursten, J. Chromatogr. A, 2000, 876, 213.
Syst., 2001, 58, 109. 23 G. Karaoglan, G. Gumrukcii, M. Ozgur, A. Bozdogan and
4 K. Asadpour-Zeynali, R. Ghavami, R. Esfandiari and B. Asci, Anal. Lett., 2007, 40, 1893.
P. Sohyli-Azad, J. AOAC Int., 2010, 93, 1995. 24 E. Dinç, E. Baydan, M. Kanbur and F. Onur, Talanta, 2002,
5 A. J. Berger, T. W. Koo, I. Itzkan and M. S. Feld, Anal. Chem., 58, 579.
1998, 70, 623. 25 L. F. Capitan-Vallvey, M. D. Fernandez, I. D. Orbe,
6 H. C. Goicoechea and A. C. Olivieri, Anal. Chem., 1999, 71, J. L. Vilchez and R. Avidad, Talanta, 1998, 47, 861.
4361. 26 L. F. Capitan-Vallvey, M. D. Fernandez, I. D. Orbe,
7 H. C. Goicoechea and A. C. Olivieri, Chemom. Intell. Lab. J. L. Vilchez and R. Avidad, Analyst, 1997, 122, 351.
Syst., 2001, 56, 73. 27 A. Lorber, Anal. Chem., 1986, 58, 1167.
8 N. R. Marsili, M. S. Sobrero and H. C. Goicoechea, Anal. 28 A. Lorber, K. Faber and B. R. Kowalski, Anal. Chem., 1997, 69,
Bioanal. Chem., 2003, 376, 126. 1620.
9 A. E. Mansilla, I. D. Meras, M. J. R. Gomez, A. Munoz de la 29 M. M. Galera, D. P. Zamora, J. L. M. Vidal, A. G. Frenich,
Pena and F. Salinas, Talanta, 2002, 58, 255. A. Espinosa-Mansilla, A. M. Pena and F. S. Lopez, Talanta,
10 A. M. Pena, A. E. Mansilla, M. I. A. A. Valenzuela, 2003, 59, 1107.
H. C. Goicoechea and A. C. Olivieri, Anal. Chim. Acta, 2002, 30 H. C. Goicoechea and A. C. Olivieri, Analyst, 1999, 124, 725.
463, 75. 31 K. Asadpour-Zeynali, E. Saeb, J. Vallipour and
11 F. Bosch-Reig and P. Campins-Falco, Analyst, 1988, 113, M. Bamorowat, J. AOAC Int., 2014, 97, 252.
1011.

This journal is © The Royal Society of Chemistry 2014 Anal. Methods, 2014, 6, 6110–6115 | 6115