Development of a micelle-enhanced high-throughput fluorometric method for determination of niclosamide using a microplate reader

Article abstract of DOI:10.1002/bio.3575

The present paper describes the development and validation of a simple and sensitive micelle-enhanced high-throughput fluorometric method for the determination of niclosamide (NIC) in 96-microwell plates. The proposed method is based on the reduction of the nitro group of niclosamide to an amino group using Zn/HCl to give a highly fluorescent derivative that was developed simultaneously and measured at λ_em 444?nm after excitation at λ_ex 275?nm. Tween-80 and carboxymethylcellulose (CMC) have been used as fluorescence enhancers and greatly enhanced the fluorescence by factors of 100–150%. The different experimental conditions affecting the fluorescence reaction were carefully investigated and optimized. The proposed method showed good linearity (r²≥?0.9997) over the concentration ranges of 1–5 and 0.5–5 μg/ml with lower detection limits of 0.01 and 0.008?μg/ml and lower quantification limits of 0.04 and 0.03?μg/ml on using Tween-80 and or CMC, respectively. The developed high-throughput method was successfully applied for the determination of niclosamide in both tablets and spiked plasma. The capability of the method for measuring microvolume samples made it convenient for handling a very large number of samples simultaneously. In addition, it is considered an environmentally friendly method with lower consumption of chemicals and solvents.

Full text of DOI:10.1002/bio.3575

Received: 9 July 2018
DOI: 10.1002/bio.3575
Revised: 4 October 2018
Accepted: 4 October 2018
R E S E A R C H A R T I C L E
Development of a micelle‐enhanced high‐throughput
fluorometric method for determination of niclosamide using a
microplate reader
1,2
3
|
Soad S. Abd El‐Hay
Fathalla F. Belal
1
Analytical Chemistry Department, Faculty of
Pharmacy, Zagazig University, Egypt
Abstract
2
Pharmaceutical Chemistry Department,
The present paper describes the development and validation of a simple and sensi-
tive micelle‐enhanced high‐throughput fluorometric method for the determination
of niclosamide (NIC) in 96‐microwell plates. The proposed method is based on
the reduction of the nitro group of niclosamide to an amino group using Zn/HCl
to give a highly fluorescent derivative that was developed simultaneously and
measured at λem 444 nm after excitation at λex 275 nm. Tween‐80 and carboxy-
methylcellulose (CMC) have been used as fluorescence enhancers and greatly
enhanced the fluorescence by factors of 100–150%. The different experimental
conditions affecting the fluorescence reaction were carefully investigated and
Faculty of Pharmacy, King Abdulaziz
University, Jeddah P.O. Box 80260, Saudi
Arabia
3
Analytical Chemistry Department, Faculty of
Pharmacy, University of Mansoura, Mansoura,
Egypt
Correspondence
Soad S. Abd El‐Hay, King Abdulaziz University,
Faculty of Pharmacy, Jeddah, Saudi Arabia.
Email: sabdulhai@kau.edu.sa
2
optimized. The proposed method showed good linearity (r ≥ 0.9997) over the
concentration ranges of 1–5 and 0.5–5 μg/ml with lower detection limits of 0.01
and 0.008 μg/ml and lower quantification limits of 0.04 and 0.03 μg/ml on using
Tween‐80 and or CMC, respectively. The developed high‐throughput method was
successfully applied for the determination of niclosamide in both tablets and spiked
plasma. The capability of the method for measuring microvolume samples made it
convenient for handling a very large number of samples simultaneously. In addition,
it is considered an environmentally friendly method with lower consumption of
chemicals and solvents.
KEYWORDS
niclosamide determination, high‐throughput method, fluorescence, microplate reader
1
|
INTRODUCTION
determination of niclosamide by derivatization reaction, which is
based on a 3 h reduction reaction of the nitro group in niclosamide
Niclosamide, (2,5‐dichloro‐40‐nitrosalicylanilide) (Figure 1) belongs to
the family of medicines called anthelmintic drugs that are effective
against most tapeworms, including the beef tapeworm, the dwarf
to an amino group by hydrogenation and catalyzed by Pd/C in
methanol:ethyl acetate (1:3), followed by reaction of the reduced
product with methyl‐β‐cyclodextrin and its application in pure and
[13]
[1]
tapeworm and tapeworm in dogs. Niclosamide has been determined
dosage forms.
In this study, a simple, sensitive micellar high‐
[
2–6]
in earlier studies using spectrophotometry,
high pressure liquid
[12]
throughput fluorometric method was developed for the determination
of niclosamide not only in pure form and tablets but also in plasma.
The automation and the high‐throughput properties of the proposed
method allowed the analysis of up to 96 samples simultaneously using
a 96‐microwell plate as a reaction vessel. Its efficiency enabled it to be
[7–11]
chromatography (HPLC)
and capillary electrophoresis.
Spectrofluorometry provides a sensitive and less time‐consuming
analysis for pharmaceutical drugs either in pure or dosage forms. Only
one fluorometric method has been developed previously for
Luminescence. 2018;1–7.
wileyonlinelibrary.com/journal/bio
© 2018 John Wiley & Sons, Ltd.
1

2
ABD EL‐HAY AND BELAL
2
2
.3
|
General procedure
The reduction procedure of NIC using Zn/HCl
.3.1
|
In total, 500 μl of the stock solution of NIC,0.01 g of zinc powder and
500 μl of hydrochloric acid (1 N) were added, then the flask was
allowed to stand for 15 min in a boiling water‐bath, cooled and the
solution filtered into a 10 ml volumetric flask, and then made up to
FIGURE 1 Structure of niclosamide
[
18]
the mark with MilliQ water.
used in quality control laboratories. Reduction of the nitro group to an
amino group has been utilized previously to retain the fluorescence of
2.3.2
|
Construction of the calibration graphs
[14–16]
some pharmaceutical drugs.
The use of microwell plates can provide a safer and more
economical tool for the determination of drug content not only in
pharmaceutical dosage forms but also in plasma using microvolumes
of samples, therefore decreasing the analysis cost and the risk of
human exposure to toxic solvents.
In total, 30 μl of the standard solution of the reduced form of NIC
50 μg/ml) was transferred into the wells of a 96‐microwell plate;
20 μl of Tween solution (0.1%, v/v) or 60 μl 0.2% w/v CMC solution
(
1
was added; 100 μl phosphate buffer (pH = 4) was added with Tween,
then the volumes were completed to 300 μl with methanol. The reac-
tion was completed simultaneously at room temperature (25 ± 5°C).
The fluorescence intensities of the resulting solutions were measured
at 444 nm after excitation at 275 nm by the monochromator micro-
plate reader. Blank wells were treated similarly using 30 μl of etha-
nol:acetone (1:1) mixture instead of the drug. The fluorescence
intensities of blank wells were subtracted from those of standard wells.
2
2
|
EXPERIMENTAL
Apparatus
.1
|
The fluorescence signal was measured in Costar® black 96‐well plates
Fisher Scientific, Pittsburgh, PA, USA) using a SpectraMax® M3
(
Multi‐mode microplate reader (Molecular Devices, Sunnyvale, CA,
|
2
.3.3
Analysis of NIC in tablets
USA). The excitation and emission wavelengths were 275 nm and
4
44 nm, respectively. A Jenway® 3510 Bench pH meter (Fisher Scien-
Ten Yomesan® tablets were accurately weighed, finely powdered and
mixed. An accurately weighed amount of powdered tablets equivalent
to 10.0 mg of NIC were transferred into a 10 ml conical flask and about
5 ml of ethanol: acetone (1:1) were added. After sonication for 15 min,
the mixture was filtered into a 10 ml volumetric flask and then
completed to the mark with the same solvent. Next, 500 μl of the stock
solution of NIC, 0.01 g of zinc powder and 500 μl of hydrochloric acid
tific Ltd, Leicestershire, UK) was used. Single and multichannel
pipettes were purchased from Pipetman Concept Multi® (Gilson Inc,
Middleton, WI, USA). Filtration using 0.8 μm Versapor Membrane
syringe filters was obtained from Gelman Laboratory (USA).
2
.2
|
Materials and reagents
(
1 N) were then added and then the procedure was completed as
•
Niclosamide, was obtained from Sigma‐Aldrich (Germany), purity
described under ‘Construction of the calibration graph’.
[17]
(
≥ 98.0%), as listed in the British Pharmacopoeia.
•
•
•
•
Stock solution: 1.0 mg/ml in ethanol: acetone (1:1) solvent
mixture.
2.3.4
|
Analysis of NIC in spiked plasma samples
Into screw capped tubes, 100 μl of drug‐free plasma was spiked with
various aliquots of NIC standard solution equivalent to 200–
Working solutions: the stock solution was diluted with ultrapure
deionized water (MilliQ) to obtain the required concentrations.
−1
1
000 μg ml . Next, 500 μl of mixture of ethanol:acetone (1:1) was
added, vortexed for 2 min and then centrifuged for 20 min at
500 rpm. Using syringe filters, the combined supernatant was with-
Ultrapure water was generated using a MilliQ water purification
system.
3
Methanol, Tween‐80 (0.1% v/v), didecyl dimethyl ammonium bro-
mide (DDAB) (0.2% w/v) aqueous solutions, were obtained from
Sigma‐Aldrich (Germany).
drawn, then collected in an evaporating dish and vaporized at 25°C
in range hood. The residue was reconstituted in 100 μl ethanol:
acetone (1:1) and the procedure was then applied as described before.
•
Zn powder, carboxymethylcellulose (CMC) and sodium dodecyl
sulphate (SDS; 94%) were purchased from Winlab Laboratory
chemicals, Leicestershire, UK and used as 0.2% w/v aqueous
solutions.
3
|
RESULTS AND DISCUSSION
The reduction of the nitro group of NIC using Zn/HCl into the
corresponding amino group forming the reduced niclosamide (RNIC)
has been utilized to retain the NIC fluorescence. This is due to the
formation of stable resonance products of NIC which exhibits intense
fluorescence.
•
•
•
Ethanol 95% (Hayman, UK), acetone (Tedia, USA).
Hydrochloric acid (BDH Chemicals, Poole, UK).
Yomesan® Tablets (Alexandria Company for Pharmaceuticals,
Egypt) labeled to contain 500 mg niclosamide per tablet.

ABD EL‐HAY AND BELAL
3
3
.1.2
|
Effect of volume of CMC or Tween‐80
The effect of volume of 0.1% v/v Tween‐80 and/or 0.2% w/v CMC
solutions on the fluorescence intensity was investigated by recording
FIGURE 3 Effect of different volumes of 0.1% v/v Tween‐80 or
0.2% w/v CMC solutions on RFI of RNIC (5 μg/ml)
SCHEME 1 Interaction of RNIC with Tween‐80 and CMC
As known, the addition of surfactants to a fluorophore solution at
concentrations over their critical micelles concentration increases its
[19–24]
quantum yield and/or its molar absorptivity (Scheme 1).
3
3
.1
|
Optimization of the assay parameters
Study of organized media
.1.1
|
The effect of using various organized media on the fluorescence of
RNIC was studied. These surfactants such as anionic (sodium dodecyl
sulfate, SDS), non‐ionic (Tween‐80), cationic surfactants didecyl
dimethyl ammonium bromide (DDAB) and macromolecules e.g. car-
boxymethylcellulose (CMC); 30 μl of every surfactant was added.
Tween and CMC greatly enhanced the relative fluorescence intensity
FIGURE 4 Effect of using different diluting solvents on RFI of RNIC
5 μg/ml) for Tween
(
RFI) of RNIC while both SDS and DDAB had no increasing effect on
(
the RFI of RNIC (Figure 2).
FIGURE 2 Effect of different organized media on RFI of RNIC
FIGURE 5 Effect of using different diluting solvents on RFI of RNIC
(5 μg/ml) for CMC
(
5 μg/ml)

4
ABD EL‐HAY AND BELAL
the fluorescence intensity over the range of 10–180 μl. It was found
that the fluorescence intensity increased as the reagent volume
increased and reached a maximum value at 120 and 60 μl for Tween
and CMC respectively, after which the blank reading was high.
Therefore, 120 and 60 μl of 0.1% v/v Tween‐80 or 2% w/v CMC
solutions were found to be the optimum volumes (Figure 3).
3.1.4
|
Effect of pH
Various types of buffers such as phosphate buffer and Tris buffers
over the pH range of 3–9, were used to study the effect of pH and
buffer volume on the RFI of micelle‐enhanced RNIC.
Using 100 μl of the buffer and 120 or 60 μl of Tween or CMC
respectively, it was found that the presence of buffer had no increas-
ing effect on the interaction between the fluorescent probe and CMC,
while the RFI of the fluorescent product increased in the presence of
buffer in the case of Tween (Figure 6).
3
.1.3
|
Effect of different diluting solvents
Using different diluting solvents such as methanol, ethanol, water and
acetonitrile, it was found that methanol showed the highest RFI for
both of Tween and CMC as organized media (Figures 4, 5).
3
.2
|
Validation of the proposed method
The method was validated for linearity, sensitivity, precision, accuracy
[
25]
and robustness according to ICH Guidelines.
3
.2.1
|
Linearity, LOD and LOQ
Linearity was investigated by least‐squares linear regression of the
fluorescence intensities against the corresponding concentrations
(μg/ml). They were found to be linear over the concentration ranges
1–5, and 0.5–5 μg/ml for Tween 80 and CMC, respectively.
The method showed good linearity with high coefficient of
determination and small % RSD. The limits of detection (LOD) and
quantification (LOQ) were determined by establishing the lowest con-
centrations that could be detected and measured respectively accord-
25
ing to ICH Q2 (R1) recommendations as shown (Figure 7, Table 1):
LOD ¼ 3:3 Sa=b and LOQ ¼ 10 Sa=b
Where Sa ¼ Standard deviation of the intercept of the calibration curve
b ¼ Slope of the calibration curve
FIGURE 6 Effect of pH on RFI of RNIC (5 μg/ml) for Tween and
CMC
FIGURE 7 Fluorescence spectra of the reaction of (0.5‐5) μg/ml RNIC with both Tween and CMC

ABD EL‐HAY AND BELAL
5
TABLE 1 Analytical performance data for the determination of NIC
TABLE 3 Precision data for the determination of NIC by the pro-
by the proposed method
posed method
Parameter
Tween
CMC
0.5–5
Amount taken (μg/ml)
% found
% RSD
% error
Linearity range (μg/ml)
1–5
Tween‐80
Intercept (a)
56.68
469.82
256.3
Intraday
Slope (b)
762.35
1
101.01 ± 0.49
100.4 ± 0.2
0.48
0.2
1.01
0.4
Correlation coefficient (r)
SD of intercept (Sa)
0.9998
0.9997
2
1.8
1.96
1.414
0.01
0.03
5
99.88 ± 0.98
0.982
0.12
Percentage relative standard deviation, % RSD
Limit of detection, LOD (μg/ml)
Limit of quantitation, LOQ (μg/ml)
0.986
0.01
0.04
Interday
1
101.13 ± 1.41
100.99 ± 0.76
100.45 ± 1.26
1.39
0.75
1.25
1.13
0.99
0.45
2
5
CMC
3
.2.2
|
Accuracy and precision
Intraday
1
98.51 ± 0.62
100.9 ± 1.02
101.31 ± 0.37
0.63
1.01
0.36
1.49
0.9
To prove the accuracy of the proposed method, the results of the
assay of studied drugs were compared with those obtained using a
2
5
1.31
[
3]
reported method.
It was found that there was no significant
Interday
difference between them regarding accuracy and precision, as the
1
2
5
98.66 ± 0.57
101.2 ± 1.19
101.27 ± 1.27
0.58
1.17
1.26
1.34
1.20
1.27
calculated t‐ and F‐values were lower than the corresponding
[
26]
tabulated
(Table 2).
Intraday precision was assessed by analyzing three different con-
centrations three times during a day or every day for 3 days in case of
intraday and interday precision, respectively. The relative standard
deviations in all cases were found to be <2 indicating good precision
of the proposed method (Table 3).
3
.2.3
|
Specificity
The proposed method was found to be specific as there was no inter-
ference that may be arise from tablets excipients or encountered from
the matrix in biological fluids. (Table 4).
TABLE 2 Assay results for the determination of NIC in pure form by
the proposed and official methods
3
.2.4
|
Robustness
[
3]
Proposed method
Reported method
Amount Amount
taken found
μg/ml) (μg/ml) recovery (μg/ml) (μg/ml) recovery
Amount Amount
Robustness of the proposed analytical method expresses its ability to
remain unaffected by small but deliberate variations in its experimen-
tal conditions such as volume of surfactant (± 5 μl). It was found that
the minor changes in the experimental conditions did not significantly
influence the RFI of RNIC.
taken
found
%
%
(
Tween 80
1
2
3
4
5
1.01
2.01
2.96
3.99
5.02
101.38
100.50
98.72
4
4.06
6.08
101.54
101.4
6
12
18
11.87
18.21
98.92
99.79
101.21
3
3
.3
|
Applications
100.46
100.17
± 0.99
1.26
Mean
101.04
± 1.07
.3.1
|
Pharmaceutical applications
±
t
S.D.
The proposed procedure was successfully applied for the quantification
of NIC inYomesan tablets. When the results of the method were statis-
(2.365)*
F
1.17
[
3]
tically compared with those obtained from the reported method. It
was found that there was no significant difference between them
regarding accuracy and precision, as the calculated t‐test and F‐values
using Student's t‐test and the variance ratio F‐test respectively were
(9.11)*
CMC
0.5
2
0.51
1.95
2.97
4.01
5.02
101.02
97.48
99.11
100.26
100.49
99.67
1.41
4
4.06
6.08
101.54
101.4
6
3
12
18
11.87
18.21
98.92
[
26]
lower than the corresponding tabulated ones
(Table 4).
4
101.21
5
3
.3.2
|
Application to biological fluids
Mean
101.04
± 1.07
±
t
S.D.
Niclosamide is one of the halogenated salicylanilides, which are highly
hydrophobic having a very low aqueous solubility. It was found that
after an oral dose of 5‐mg/kg, plasma concentration of niclosamide
was measurable up to 12 h at 3.8 ± 2.0 ng/ml, it was rapidly absorbed
with a Tmax less than 30 min after oral administration. Also, it showed
1.6
(
(
2.365)*
F
1.73
9.11)*
*Tabulated t‐test and f‐values P = 0.05.

6
ABD EL‐HAY AND BELAL
TABLE 4 Assay results for the determination of NIC using Tween‐80, CMC in Yomesan® tablets and plasma
Proposed method
[
3]
Tween‐80
CMC
Reported method
Amount taken Amount found
Amount taken Amount found
Amount taken Amount found
%
Parameter (μg/ml)
(μg/ml)
% found
(μg/ml)
(μg/ml)
% found
(μg/ml)
(μg/ml)
found
Yomesan®
1
2
4
5
0.99
2.02
4.05
4.96
98.94
100.92
101.16
99.13
1
3
4
5
0.97
2.99
3.95
4.99
96.77
4
4.06
6.08
101.54
101.4
99.51
6
98.78
12
18
11.87
18.21
98.92
99.89
101.21
101.04
± 1.07
Mean
100.04
± 1.16
99.77
±
t
S.D.
± 0.89
1.26 (2.447)*
0.85 (9.27)*
102.89
1.83 (2.447)*
1.44 (9.27)*
98.53
F
Plasma
1
2
3
4
5
1.03
2.02
3.07
4.15
5.09
2
3
4
5
1.97
2.97
4.09
5.06
4
4.06
6.08
101.54
101.4
100.87
102.47
103.67
101.71
102.32
± 1.08
98.90
6
102.29
101.10
12
18
11.87
18.21
98.92
101.21
Mean
100.2
101.04
± 1.07
±
t
S.D.
± 1.55
1.77 (2.365)*
1.02 (9.17)*
0.89 (2.447)*
2.1 (9.27)*
F
*Tabulated t‐test and F‐values P = 0.05.
[27]
the highest Cmax at 354 ± 152 ng/ml.
The proposed method was
[2] S. A. Abdel Fattah, Spectrosc. Lett. 1997, 30, 795.
successfully applied for determination of NIC in spiked plasma.
Statistical comparison of the obtained results with those obtained by
[3] K. Abdel‐Salam Attia, M. W. I. Nassar, A. W. Madkour, Eur. J. Biomed.
Pharm. Sci. 2015, 2, 1295.
[
3]
[4] H. G. Daabees, Anal. Lett. 2000, 33, 639.
the reported method using Student's t‐test and the variance ratio
[
26]
[5] A. A. Khier, M. M. Elhenawee, M. S. Elmasry, E‐J. Chem. 2008, 5, 1087.
F‐test
indicated no significant difference between them regarding
[
6] S. W. Shantier, E. Ali Elobaid, E. A. Gadkariem, World J. Pharm. Res.
015, 4, 122.
accuracy and precision (Table 4).
2
[
7] V. Amelin, A. Korotkov, A. Andoralov, J. AOAC Int. 2016, 99, 1600.
4
|
CONCLUSION
[8] G. Doran, M. M. Stevens, Anal. Methods 2014, 6, 6871.
[
9] L. Geis‐Asteggiante, S. J. Lehotay, A. R. Lightfield, T. Dutko, C. Ng, L.
Bluhm, J. Chromatogr. A 2012, 1258, 43.
A simple and sensitive micelle‐enhanced high‐throughput fluorometric
method for determination of NIC in 96‐microwell plates was
developed and validated. The method is based on the simple reduction
of nitro groups in niclosamide with Zn/HCl for the amino group to
retain NIC fluorescence followed by enhancement of the fluorescence
through using Tween‐80 or CMC. In addition, using microwell plates
can offer a safer, environmentally friendly and economical method
for the determination of niclosamide. The high sensitivity of the
proposed method allowed the determination of the studied drug in
pharmaceutical dosage forms as well as in spiked plasma with high
recoveries of the analyte.
[10] S. J. Lehotay, A. R. Lightfield, Anal. Bioanal. Chem. 2018, 410, 1095.
[11] C. Paghadar, N. H. Vadia, Arabian J. Chem. 2015. https://doi.org/
1
0.1016/j.arabjc.2014.11.064, Ahead of Print
[
12] J. M. Cabot, E. Fuguet, M. Roses, Electrophoresis 2014, 35, 3564.
[
13] M. Algarra, B. B. Campos, J. E. Rodriguez‐Borges, F. G. Sanchez, J. M.
Lopez‐Romero, J. C. G. E. da Silva, J Incl Phenom Macro 2012, 72, 89.
[
[
[
[
14] F. Belal, A. A. Al‐Majed, S. Julkhuf, N. Y. Khalil, Pharmazie 2003, 58, 874.
15] F. Belal, F. Ibrahim, Z. Sheribah, H. Alaa, Luminescence 2017, 32, 713.
16] F. Belal, F. Ibrahim, Z. Sheribah, H. Alaa, Luminescence 2017, 32, 1299.
17] British Pharmacopoeia Commission, Her Majesty's Stationery Office,
London 2005.
[
18] B. M. H. Al‐Hadiya, in Profiles of Drug Substances, Excipients and
Related Methodology, (Ed: H. G. Brittain) Vol. 32, Academic Press
CONFLICT OF INTEREST
2
005 67.
The authors declare no conflict of interest.
[
[
[
19] Z. Chang, H.‐T. Yan, J. Lumin. 2012, 132, 811.
20] J. Du, H. Wu, Y. An, Y. Shi, X. Guo, L. Du, Anal. Methods 2016, 8, 2778.
ORCID
21] J. Du, H. Wu, X. Du, Z. Zhao, X. Zhao, Y. Shi, X. Guo, L. Du, Can. J.
Chem. 2017, 95, 512.
Soad S. Abd El‐Hay http://orcid.org/0000-0003-1880-3699
[
22] A. S. Fayed, M. A. Hegazy, E. E. Abbas, N. N. Salama, RSC Adv. 2016, 6,
REFERENCES
1
07246.
[1] D. C. Blood, V. P. Studdert, C. C. Gay, Saunders comprehensive veteri-
[23] T. D. Smirnova, S. N. Shtykov, E. A. Zhelobitskaya, M. I. Safarova,
nary dictionary, 3rd ed., Elsevier Saunders, Edinburgh, New York 2007.
J. Anal. Chem. 2017, 72, 562.

ABD EL‐HAY AND BELAL
7
[
[
24] H. Wu, G.‐Y. Zhao, L.‐M. Du, Spectrochim. Acta, Part a 2010, 75A,
624.
1
How to cite this article: Abd El‐Hay SS, Belal FF. Develop-
ment of a micelle‐enhanced high‐throughput fluorometric
method for determination of niclosamide using a microplate
reader. Luminescence. 2018;1–7. https://doi.org/10.1002/
bio.3575
25] Presented in part at the ICH harmonized tripartite guideline, valida-
tion of analytical procedures: Text and methodology, Q2 (R1),
current step 4 version, parent guidelines on methodology, 6 Novem-
ber 1996, incorporated in November 2005.
[26] J. C. Miller, J. N. Miller, Statistics and Chemometrics for Analytical
Chemistry, 5th ed., Harlow, England 2005.
[27] T.‐K. Y. Yi‐Wei, K.‐T. L. Chang, W.‐C. Chen, H.‐T. Yao, S.‐J. Lan, Y.‐S.
Wu, H.‐P. Hsieh, C.‐M. Chen, C.‐T. Chen, J. Food Drug Anal. 2006, 14,
329.