Application of Central Composite Design
for the Determination of Exfoliating
Agents in Cosmetics by Capillary
Electrophoresis
with Electroosmotic Flow
Modulation
Yen Ling Chen*, Shiuh Jen Jiang2, Chia Hsien Feng3, Shih Wei Wang 4, Yi
Hui Lin1, Pei Yu Liu 3
1School of Pharmacy, College of Pharmacy, China Medical University,
Taichung, 40402, Taiwan
2Department of Chemistry, National Sun Yat-sen University, Kaohsiung
80424, Taiwan
3Department of Fragrance and Cosmetic Science, College of Pharmacy,
Kaohsiung
Medical University, Kaohsiung 807, Taiwan
4Department of Medicine, Mackay Medical College, Taipei 252, Taiwan
Correspondence:
Yen-Ling Chen, Ph.D., Assistant Professor.
Department of Fragrance and Cosmetic Science,
College of Pharmacy,
Kaohsiung Medical University,
Kaohsiung, Taiwan
E-mail:yelichen@kmu.edu.tw
Fax: 886-7-3210683
Tel: 886-7-3121101 ext. 2584
Page 1 of 33
URL: http://mc.manuscriptcentral.com/lanl E-mail: analyticalletters@email.wcu.edu
Application of Central Composite Design
for the
Determination of Exfoliating Agents in
Cosmetics by
Capillary Electrophoresis
with
Electroosmotic Flow
Modulation
Yen Ling Chen*, Shiuh Jen Jiang2, Chia Hsien Feng3, Shih Wei Wang 4, Yi
Hui Lin1, Pei Yu Liu 3
ABSTRACT
Multivariate analysis within central composite design is applied to simplify the optimization procedure and explore the interactions among all experimental parameters in the field of analytical chemistry. In this study, central composite design is used to identify the optimal capillary electrophoresis conditions with electroosmotic flow modulation to assess the contents of seven exfoliating agents in cosmetics. The influences of the phosphate concentration, cetyltrimethylammonium bromide concentration and methanol percentage on the response were evaluated by the use of the chromatographic exponential function. This strategy is based on the use of the chromatographic exponential
function as a response to simultaneously investigate the resolutions and separation times of 16 different CE conditions. By the use of statistical software, the optimized conditions were found to be as follows: 150mM phosphate solution (pH = 7) containing 0.5 mM
cetyltrimethylammonium bromide, 3 mM γ-CD and 25% methanol was used as the running buffer. To shorten the analysis time, an
electroosmotic flow modulating agent (cetyltrimethylammonium
bromide) was added to the separation buffer to reverse the EOF. During the method validation, the calibration plots were linear (r ≧ 0.998) with
high precision and accuracy in the homemade cosmetic matrix. The exfoliating agent contents of two commercial cosmetic products were determined and quantified by the optimized CE conditions, and the results coincided well with the LC-MS results. This CE method is feasible for application to the detection and analysis of exfoliating agents in cosmetics.
Key words: Capillary electrophoresis; Central composite design; Chromatographic
exponential function; Exfoliating agents; Cosmetics.
INTRODUCTION
The rough outer layer of the epidermis, called the stratum corneum, consists of several layers of flat, keratinized, non-nucleated cells and intercellular lipids. The stratum corneum serves as a barrier against damage by foreign substances and prevents transdermal water loss (Egelrud 2000). New stratum corneum cells are continuously produced by the basal layer. However, some diseases, such as xerosis,
ichthyosis, and acne, are caused by hyperkeratosis. To maintain a constant and normal stratum thickness, exfoliating agents are used to remove excess keratinocytes. The exfoliation process involves the detachment of corneocytes from one another by breaking the
desmosomes (corneosomes), which are the intercellular “rivets” that bind corneocytes together (Brysk 1992). Hydroxy acids, including alpha hydroxy acids and beta hydroxy acids, are the most common
exfoliating agents. Alpha hydroxy acids, such as glycolic acid, malic acid, tartaric acid, lactic acid, citric acid, and mandelic acid, are the most popular agents, but salicylic acid, often considered a beta hydroxyl acid, is also widely used (Saint-Léger 2007). In addition to their use in clinical treatment, alpha hydroxy acids and beta hydroxy acids are widely applied in the cosmetics field to enhance stratum corneum desquamation for the purpose of improving the appearance of skin (Berardesca 1997). The pH of the cosmetic formulation varies with both the nature of the hydroxy acids and their concentrations (Elsner 2000). The concentration of hydroxy acids in cosmetic products is limited to less than 10%, and their pH values must be more than 3.5 to avoid skin irritation.According to the regulations that have been issued by the Taiwan Food and Drug Administration, the concentration limit of salicylic acid in cosmetic products is below2% to avoid salicylism (Chin 2007). To ensure the proper quality control of commercial cosmetic
products and to minimize skin irritation due to the overuse of exfoliating agents, an analytical method that can qualitatively and quantitatively determine the exfoliating agents present in cosmetics is needed.
Capillary electrophoresis (CE) is a powerful separation technique that has enhanced the separation capabilities of a wide range of analytes in such fields as biology (Liu 2013, Zamfir 2013), environmental studies (Wang 2013), food science (Önal 2013), pharmaceuticals (Felix 2012), and cosmetic analysis (Liu 2012, Cheng 2012). The advantages of CE separation are the use of an automated system for precise quantitative analysis and the consumption of only a small amount of the reagent. Before the development of a successful analytical CE method, several factors should be considered, such as the electrolyte composition, concentration, and pH value, as well as the use of additives. The selectivity of CE can be manipulated by altering these parameters; however, traditional “one variable at a time” schemes are time consuming and fail to consider the interactions between factors (Dejaegher 2011). In recent years, experimental designs such as central composite designs and orthogonal methods have been widely applied for the optimization of CE (Bourdon 2013, Capella-Peiró 2006, Lamalle 2012, Li 2011, Liu 2006, Liu 2011) and HPLC (Acevska 2012, Kumar 2012) separation and extraction conditions (Ghambarian 2012, Ramandi 2011). A central composite design contains a two-level full factorial design, a star design, and twice as many center points to examine the factors. The points of the full factorial design are situated at the factor levels -1 and +1. The star points represent new extreme values (-α and +α) for each factor in the design (Hanrahan 2008). A central composite design is a powerful statistical tool for reducing the number of experiments required, evaluating the interactions between factors, and examining the global experimental domain. In this study, a central composite design was used to determine whether there was a significant change in the response at different factor levels.
The purpose of this study is to establish a simple and rapid CE method that can simultaneously analyze seven exfoliating agents and to determine their contents in cosmetic products using a central
composite design to acquire the optimized CE conditions. A homemade cosmetic matrix and a standard addition method were used to
exfoliating agents in the cosmetic products.
EXPERIMENTAL
Chemicals and reagents
All chemicals were of analytical grade. The malic acid and mandelic acid were
purchased from Alfa Aesar (Ward Hill, MA, USA). The glycolic acid and lactic acid
were purchased from Sigma-Aldrich (St. Louis, MO, USA). The tartaric acid, citric
acid, and salicylic acid were purchased from J. T. Baker (Phillipsburg, USA). The
sodium dihydrogen phosphate (NaH2PO4), hydrochloric acid (HCl),
sodium
hydroxide (NaOH), and cetyltrimethylammonium bromide were purchased from
Merck (Darmstadt, Germany). The fumaric acid, which was used as an internal
standard (IS), and γ-cyclodextrin were purchased from the Tokyo Chemical Industry
(Tokyo, Japan). Methanol (Mallinckrodt Baker, Phillipsburg, USA) was prepared as a
buffer additive. Double-distilled water prepared by a Milli-Q system (Millipore,
Bedford, MA, USA), which had a resistivity of 18.2 M_ cm at 25°C, was used to
prepare the buffer and related aqueous solutions. Sample I was a gel-type cosmetic
containing glycolic acid, citric acid, and lactic acid (NeoStrata Co., Princeton, USA),
and sample II was a liquid-type raw cosmetic material consisting of a complex of
alpha hydroxy acids containing glycolic acid, citric acid, lactic acid, malic acid, and
tartaric acid (Pro More Trading Co., Taipei, Taiwan). None of the samples were
labeled with the amounts of the hydroxy acids.
Instrument
array
detector was used. The study was performed using a fused-silica capillary (50 μm i.d.,
Polymicro Technologies, Phoenix, AZ, USA). The effective length to the detector was
40 cm, and the total length was 50.2 cm. The running buffer was set as a
phosphate/HCl solution (200 mM, pH = 7.0) containing 3 mM γ-CD, 0.5 mM
cetyltrimethylammonium bromide, and 25% methanol. The sample loading was
achieved by the application of pressure (0.5 psi, 20 s). The separation voltage was -15
kV (anode at the detector end), and the separation temperature was maintained at
25ºC. All operations, including the electropherogram acquisition, were computer-controlled, using the Beckman Coulter 32 Karat software system (Fullerton,
CA, USA). A Hitachi U-5100 Ratio Beam spectrophotometer (Hitachi, Tokyo, Japan)
was used for detection. The analyses were confirmed by LC-MS using a Waters 2695
liquid chromatography and a Waters Micromass ZQmass spectrometer (Waters
Corporation, MA, USA). All of the analyses were conducted using an electrospray
with a -15 kV probe. A Phenomenex Security Guard C-18 guard column and a
ZORBAX Eclipse XBD-C18 analytical column (2.1 × 100 mm, 3.5 m) were used.
The solvent gradient was composed of an acidified (0.1% formic acid) H2O/methanol
solution at a ratio of 100:0 for the first five minutes, after which it was ramped to
0:100 over 30 min. The flow rate was set at 0.2 mL/min.
Method validation
Stock solutions of the analytes were prepared in double-distilled water
diluted by double-distilled water as reference standards. To simulate the commercial
cosmetic matrix and decrease the interference of the matrix, a homemade emulsion
was prepared by adding phase A, including 1,4-butylene glycol (2.5 g) as a humectant,
xanthan gum (0.05 g) as a thickening agent, and EDTA-2Na (0.05 g) as a preservative
agent, and phase B, including OLIVEM 800® (2 g) and OLIVEM 1000®
(0.5 g) as
emulsifiers, phenoxyethanol (0.25 g) as a preservative agent and other ingredients,
such as caprylic/capric triglyceride (2.5 g), cyclomethicone (1.5 g), and squalene (1.5
g), as emollient agents. The blank emulsion was prepared by premixing phase A and
phase B separately and heating each to 80°C. Phase A was added to phase B under
stirring until the solution was completely emulsified; then, it was cooled to room
temperature. The emulsification was stored at room temperature until it was used. The
emulsions with different concentrations of exfoliating agents were weighed (200 mg),
and 1 mL of methanol was added as an extraction solvent. The samples were
ultrasonicated for 30 min and centrifuged at 12,000 rpm for 10 min; then, 900 μL of
the supernatant was removed from each sample. After drying, the residue was
reconstituted in 1000 μL of double-distilled water. The reconstituted solution was
filtered through a 0.45 μm poly(vinylidene fluoride) (PVDF) filter (Millipore,
Bedford, MA, USA) for CE analysis. The quantitation ranges of the calibration curves
for the analytes in the cosmetic matrix were 0.18-1.8 mM for malic acid, 0.09-0.9 mM for tartaric acid, 0.36-3.6 mM for citric acid,
0.045-0.45 mM for glycolic acid, 0.9
mM-9 mM for lactic acid, and 4.5-45 μM for both salicylic acid and mandelic acid.
The calibration curves were established, with the corrected peak area ratio of each
analyte to the IS as the ordinate (Y) and the concentration of each respective analyte
in mM as the abscissa (X). Three concentrations of each analyte were chosen for the
analysis of the precision and accuracy of the proposed method. The limits of detection
(LOD) and the limits of quantification (LOQ) were determined by decreasing the
concentration of each analyte until the ratios of signal to noise equaled 3 and 10,
respectively (sampling for 20 s at 0.5 psi).
Sample pretreatment
Two hundred milligrams of Sample I was dissolved in 1 mL of methanol under
ultrasonication for 30 min and was then centrifuged at 12,000 rpm for 10 min. The
900 μL of supernatant was evaporated in a centrifugal vaporizer (CVE 2000, EYELA,
Japan) and reconstituted with 1 mL of double-distilled water. The reconstituted
solution of Sample I and 200 μL of each untreated Sample I were filtered through a
0.45 μm PVDF filter and diluted to suitable concentrations for CE analysis. Further
extraction was conducted using solid phase extraction with a Strata C18-E instrument
(Phenomenex, CA, USA) and was necessary for LC-MS analysis due to the complex
matrices of cosmetic products that causes interference in MS detection. The SPE
cartridges were preconditioned with 3 mL each of MeOH and water. After loading the
column. The
eluted solution containing the analytes was collected and diluted to a suitable
concentration for LC-MS analysis.
RESULTS AND DISCUSSION
Preliminary experiments
The UV spectra of the seven exfoliating agents under optimized separation
conditions (200 mM phosphate buffer (pH = 7.0) with 3 mM γ-CD, 0.5 mM
cetyltrimethylammonium bromide and 25% methanol) were obtained using a
spectrophotometer. The wavelength of the maximum absorbance for all of the
analytes in 400 μM was 200 nm. Among them, salicylic acid and mandelic acid have
the highest absorbance values because of their benzene structures. To identify the
appropriate separation conditions, preliminary studies were developed with a
phosphate buffer and several buffer additives to shorten the total analysis time and
improve the separation resolution. Because the pKa values of the seven exfoliating
agents were all less than 5.13, the phosphate buffer (pH = 7.0) was chosen to achieve
a high buffer capacity and to afford all of the analytes as anionic compounds. The
native charges on the silanol groups in the capillary form a double layer counterion,
and mobile cations from a diffuse layer generate electroosmotic flow (EOF) towards
the cathode. Electroosmotic flow is the driving force for both ionic and non-ionic
analytes that can push the analytes migrating in the capillary under an applied voltage
organic solvents
were used to manipulate the EOF by modifying the bonding of silica surface of the
capillaries. The cationic surfactant cetyltrimethylammonium bromide, used as an EOF
modulator, can be absorbed onto the capillary wall by the electrostatic attraction
between the positively charged ammonium moieties and the negatively charged
silanol groups. Above a 0.35 mM cetyltrimethylammonium bromide concentration, a
bilayer is formed, and the EOF is reversed (Terabe 1993). Methanol, as an organic
modifier, alters the electroosmotic flow and effective electrophoretic mobility of the
analytes. Therefore, the selectivity of the seven exfoliating agents could also be
improved by adding different amounts of methanol to the separation buffer. At the
same time, methanol may affect the interaction between the exfoliating agents and
γ-CD by altering the polarity of the separation buffer. Gamma-CD is usually used as a
chiral selector for an enantiomeric separation. However, γ-CD (1-5 mM) can
noticeably improve the separation efficiency in this study by altering the
electromobility of salicylic acid. Ultimately, 3 mM γ-CD was selected for use in
further studies.
Optimized results based on central composite design
Central composite design, which was the final optimization procedure, contained
16 runs with 2 center points and 3 factors (phosphate concentration, cetyltrimethylammonium bromide concentration, and methanol concentration) with
significant influences on the preliminary experiments. Each factor was examined at
five levels (-α, -1, 0, +1, +α), with α equal to (2f)1/4, where f is the
number of factors
in a rotatable circumscribed central composite design. If three factors are examined, α
is equal to 1.68. The central composite design established using five levels of the
three factors is shown in Table 1. The response factors were calculated using the
chromatographic exponential function, which was created by Morris et al. in 1996 to
simultaneously investigate the resolution and duration of electrophoretic separation
(eq. 1) (Morris 1996).
In eq. 1, Ri is the resolution between two contiguous peaks, Ropt is the
optimized
resolution and was set at 1.5, and n is the expected number of peaks. Moreover, tf is
the migration time of the final peak, tmax is the expected maximum
analysis time,
which was set at 20 min, and a is a factor for the slope adjustment. When a was higher,
the results depended more strongly on the resolution. Because the resolution was
more important than the analysis time in this study, a was set to 2. A lower
chromatographic exponential function value indicates that the results are closer to the
desired conditions, and the final objective of the optimization procedure is to
minimize the chromatographic exponential function. From the electropherograms of
runs 8, 9 and 14 (data are shown in Figure 1), we can observe that a lower value of the
chromatographic exponential function corresponds to higher resolution between each
peak. Figure 2 presents the three-dimensional response surface of the chromatographic exponential function obtained from JMP software as a function of
factors. The chromatographic exponential function plotted against the phosphate
concentration and the cetyltrimethylammonium bromide concentration is shown in
Figure 2 (A). When the cetyltrimethylammonium bromide concentration decreased
from 1.5 to 0.5 mM, the chromatographic exponential function decreased for all
phosphate concentrations. An increase in the phosphate concentration from 150 to 250
mM resulted in an increase of the chromatographic exponential function when the
cetyltrimethylammonium bromide concentration was fixed at 0.5 mM. In Figure 2 (B),
the effect of altering the methanol concentration and the cetyltrimethylammonium
bromide concentration on the chromatographic exponential function response can be
observed. When the percentage of methanol was maintained at a constant 22.6%, the
lowest value of the chromatographic exponential function at various cetyltrimethylammonium bromide concentrations can be observed. Figure 2 (C)
shows the chromatographic exponential function response plotted against the
methanol and phosphate concentrations. Samples with a 22.6% methanol
concentration with various phosphate concentrations have the lowest values of
chromatographic exponential function on the response surface. Meanwhile, it can be
observed that an increase in the phosphate concentration from 150 to 250 mM
resulted in an increase of the chromatographic exponential function value. To
minimize the response, it was necessary to use a low cetyltrimethylammonium
moderate methanol
concentration. From the central composite design results, 150 mM phosphate, 0.5 mM
cetyltrimethylammonium bromide, and 22.6% methanol were predicted as the optimal
conditions. However, the peaks of citric acid and glycolic acid achieved baseline
separation at a methanol concentration of up to 25% (results are shown in Figure 3).
Therefore, the optimized conditions with 150 mM phosphate, 0.5 mM cetyltrimethylammonium bromide, and 25% methanol were obtained after a slight
revision. To minimize the migration time and improve the peak shape for salicylic
acid and mandelic acid, the separation voltage was adjusted to -15 kV, and the total
analysis time was less than 12 min. In this study, we analyzed seven exfoliating agents
in a homemade cosmetic emulsion after sample pretreatment to simulate a real
cosmetic sample and observe whether any interference peaks appeared. The
electropherogram of the seven exfoliating agents separated by the optimized CE
conditions is shown in Figure 4. Separation efficiencies are usually represented by the
number of theoretical plates. We calculated the number of theoretical plates for all the
peaks in the Figure 4. The number of theoretical plates reached 1.3×105 plates for
malic acid, 1.5×105 plates for tartaric acid, 1.4×105 plates for glycolic
acid, 1.8×105
plates for citric acid, 9.7×104 plates for lactic acid, 1.1×105 plates for
salicylic acid
and 1.2×105 plates for mandelic acid, and all showed good separation
efficiencies.
Acquiring the predicted optimized conditions by using the responses of only sixteen
runs in a central composite design was less time consuming and had a low
experimental cost. The interactions between the experimental parameters and the
responses were also modeled in this study.
Method validation using a homemade cosmetic
emulsion
To evaluate the quantitative applicability of the established CE method in cosmetic
samples, five different concentrations of seven exfoliating agents were analyzed using
fumaric acid (50 μM) acting as an IS to calculate the peak area ratios via dividing the
corrected peak areas of each analyte. The corrected peak area that was used with the
standard was calculated as follows: (peak area of an analyte/migration time of the
analyte) divided by the (peak area of the IS/migration time of the IS). Calibration
curves were obtained from linear peak area ratios with the X axis as concentration and
the Y axis as peak area ratios. Table 2 shows the regression equations with high
linearities, and the correlation coefficient (r) of each curve was greater than 0.998 (n =
3). The LODs were 625 nM for mandelic acid and salicylic acid, 12.5 μM for citric
acid and tartaric acid, 25 μM for malic acid, 100 μM for glycolic acid and 125 μM for
lactic acid (S/N = 3; 0.5 psi, 20 s injection). The LOQs were 2 μM for mandelic acid
and salicylic acid, 40 μM for citric acid and tartaric acid, 75 μM for malic acid, 400
μM for glycolic acid and 900 μM for lactic acid (S/N = 10; 0.5 psi, 20 s injection).
The relative standard deviations (RSDs) and relative errors (REs) were calculated on
acid (at levels of
0.1, 0.5 and 0.8 mM), glycolic acid (at levels of 0.4, 2.0 and 3.2 mM), citric acid (at
levels of 0.05, 0.25 and 0.40 mM), lactic acid (at levels of 1.0, 5.0 and 8.0 mM),
mandelic acid and salicylic acid (each at levels of 0.005, 0.025 and 0.040 mM). The
results indicated that the RSD and RE values were less than 9.23% (data shown in
Table 2). Lower RSDs and REs indicate that the developed method could estimate the
concentration values with good accuracy and precision in a homemade cosmetic
emulsion using this CE method.
Application
Two commercial products (Samples I and II) were subjected to quality control
analysis using the central composite design-optimized CE conditions. The
electropherograms obtained from the analysis of Samples I and II are shown in Fig. 5.
Standard addition was used to quantify the exfoliating agents in the commercial
products instead of a calibration curve of standards to avoid the matrix effect issue.
The recoveries of this method were briefly studied by spiking and mixing the
commercial products with known amounts of all analytes. The analytical results
indicate that the recoveries of the method are all above 95%. The exfoliating agent
contents were 0.34% glycolic acid, 0.21% citric acid, and 0.39% lactic acid in Sample
I; and 0.55% malic acid, 0.60% tartaric acid, 12.64% glycolic acid, 2.01% citric acid,
and 10.63% lactic acid in Sample II. Liquid chromatography-mass spectrometry
laboratories, especially in
the fields of pharmacology, drug discovery, and qualitative and quantitative analysis.
To identify these commercial cosmetic products containing several exfoliating agents,
LC-MS was used as tool for further verification. All analyses had similar retention
times with very poor resolution because of the high polarity of the samples. However,
the MS spectra of the seven exfoliating agent standards in the negative ionization
mode showed the formation of [M-H]
-.[M-H]- ions with m/z values of 133.1 for malic acid, 149.1 for tartaric
acid, 75.2 for glycolic acid, 191.1 for citric acid, 89.2 for lactic acid, 137.1 for salicylic acid, and 151.1 for mandelic acid. To avoid
interference in the LC-MS detection from the complicated sample matrix characteristic of cosmetics, solid phase extraction was applied to exclude some hydrophobic compounds by directly loading the sample without washing and eluting procedures. The characteristic peaks in the MS spectra in the negative ionization mode for Sample I, containing glycolic acid, citric acid, and lactic acid, and Sample II, containing malic acid, tartaric acid, glycolic acid, citric acid, and lactic acid, were clearly identified (data were not shown). Based on these results, the proposed CE method was established to be suitable for the determination of the exfoliating agents in cosmetic products.
CONCLUSIONS
A simple, efficient and fast analytical method based on capillary electrophoresis has
been developed and optimized for the determination of seven exfoliating agents in
cosmetic products using central composite design. Statistical experimental design
techniques such as central composite design are useful for studying the effects of
factors and the interactions between factors. The chromatographic exponential
function was used as the response variable for the optimization of the CE conditions,
which can be optimized efficiently with a minimal amount of experimentation. This
method has been successfully established and applied to the quantification of
exfoliating agents in commercial cosmetic products.
We gratefully acknowledge the support of the National Science Council of Taiwan
and the NSUSU-KMU Joint Research Project (#NSYSUKMU 102-P015) for funding
this work.
The authors have declared no conflict of interest
REFERENCES
Acevska, J., G. Stefkov, R. Petkovska, S. Kulevanova and A. Dimitrovska. 2012.
Chemometric approach for development, optimization, and validation of different
chromatographic methods for separation of opium alkaloids. Anal. Bioanal. Chem.
403: 1117-1129.
Berardesca, E., E. Distante, G. Vignoli, C. Oresa and B. Green. 1997. Alpha
hydroxyacids modulate stratum corneum barrier function. Br. J. Dermatol. 137:
934-938
Bourdon, F., M. Lecoeur, M. Duhaut, P. Odou, C. Vaccher and C. Foulon. 2013. A
validated micellar electrokinetic chromatography method for the quantitation of
dexamethasone, ondansetron and aprepitant, antiemetic drugs, in organogel. J. Pharm.
Biomed. Anal. 86: 40-46.
Brysk, M.M. and S. Rajaraman. 1992. Cohesion and desquamation of epidermal
stratum corneum. Prog. Histochem. Cytochem.25:1-53.
Capella-Peiró, M.E., D. Bose, M.F. Rubert and J. Esteve-Romero. 2006. Optimization
of a capillary zone electrophoresis method by using a central composite factorial
design for the determination of codeine and paracetamol in pharmaceuticals. J.
Chromatogr. B Analyt. Technol. Biomed. Life Sci. 24: 95-101.
Cheng, Y.C., C.C. Wang, Y.L. Chen and S.M. Wu. 2012. Large volume sample
stacking with EOF and sweeping in CE for determination of common preservatives in
cosmetic products by chemometric experimental design. Electrophoresis. 33:
443-1448.
Chin, R.L., K.R. Olson and D. Dempse. 2007. Salicylate toxicity from ingestion and
continued dermal absorption. Cal. J. Emerg. Med. 8:23-5.
Dejaegher, B. and Y. Vander Heyden. 2011. Experimental designs and their recent
advances in set-up, data interpretation, and analytical applications. J. Pharm. Biomed.
Anal. 56:141-158.
Egelrud, T. 2000. Desquamation in the stratum corneum. Acta Derm. Venereol. Suppl
208:44-45.
Elsner, P. and H.I. Maibach. 2000. Cosmeceuticals Drugs vs. Cosmetics, New York,
USA: Marcel Dekker.
Felix, F.S., L.M. Ferreira, O. Rossini Pde, C.L. do Lago and L. Angnes. 2012.
Quantification of terbinafine in pharmaceutical tablets using capillary electrophoresis
with contactless conductivity detection and batch injection analysis with
amperometric detection. Talanta. 15:220-225
Ghambarian, M., Y. Yamini and A. Esrafili. 2012. Three-phase hollow fiber
microextraction based on two immiscible organic solvents for determination of
tricyclic antidepressant drugs: comparison with conventional three-phase hollow fiber
Hanrahan, G., R. Montes and F. Gomez. 2008. Chemometric experimental design
based optimization techniques in capillary electrophoresis: a critical review of modern
applications. Anal. Bioanal. Chem. 309:169-179.
Kumar, P., A. Rubies, R. Companyó and F. Centrich. 2012. Hydrophilic interaction
chromatography for the analysis of aminoglycosides, J. Sep. Sci. 35: 498-504.
Lamalle, C., R. Djang’Eing’A Marini, B. Debrus, P. Lebrun, J. Crommen, P. Hubert,
A.C. Servais and M. Fillet. 2012. Development of a generic micellar electrokinetic
chromatography method for the separation of 15 antimalarial drugs as a tool to detect
medicine counterfeiting. Electrophoresis. 33: 1669-1678.
Li, X., X. Xu, D.R. Albano and T. You. 2011. Optimization using central composite
design for antihistamines separation by nonaqueous capillary electrophoresis with
electrochemical and electrochemiluminescence detections. Analyst. 21:5294-5301.
Liu, J.J., S.P. Li and Y.T. Wang. 2006. Optimization for quantitative determination of
four flavonoids in Epimedium by capillary zone electrophoresis coupled with diode
array detection using central composite design. J. Chromatogr. A 27: 344-349.
Liu, M., Y. Zheng, Y. Ji and C. Zhang. 2011. Development and validation of a
capillary electrophoresis method for the enantiomeric purity determination of
RS86017 using experimental design. J. Pharm. Biomed. Anal.55:93-100. Liu, P.Y., Y.H. Lin, C.H. Feng and Y.L. Chen. 2012. Determination of
hydroxy acids
in cosmetics by chemometric experimental design and cyclodextrin-modified
Liu, X., Y. Yu, H. Gu, T. Zhou, L. Wang, B. Mei and G. Shi. 2013. Simultaneous
determination of monoamines in rat brain with Pt/MWCNTs@Pdop hybrid
nanocomposite using capillary electrophoresis--amperometric detection.
Electrophoresis. 34: 935-943.
Morris, V.M., J.G. Hughes and P.J. Marriott. 1996. Examination of a new chromatographic function, based on an exponential resolution term, for use in
optimization strategies: application to capillary gas chromatography separation of
phenols. J. Chromatogr. A. 755: 235-243.
Önal, A., S.E. Tekkeli and C. Önal .2013. A review of the liquid chromatographic
methods for the determination of biogenic amines in foods. Food Chem. 138: 509-515.
Ramandi, N.F., N.M. Najafi, F. Raofie and E. Ghasemi. 2011. Central composite
design for the optimization of supercritical carbon dioxide fluid extraction of fatty
acids from Borago officinalis L. flower. J. Food. Sci., 76: 1262-1266. Saint-Léger, D., J.L. Lévêque and M. Verschoore. 2007. The use of hydroxy acids on
the skin: characteristics of C8-lipohydroxy acid. J. Cosmet. Dermatol. 6: 59-65.
Shintani, H. and J. Polonský.1996. Handbook of Capillary Electrophoresis applications, Oklahoma: Springer.
Terabe, S.1993. Micellar electrokinetic chromatography. California, USA: Beckmann.
Wang, X. and H.Wang. 2013. Epigenotoxicity of environmental pollutants evaluated
by a combination of DNA methylation inhibition and capillary electrophoresis-laserinduced
fluorescence immunoassay.Anal. Bioanal. Chem. 405: 2435-42.
Zamfir, A.D., C. Flangea, A. Serb, A.M. Zagrean, A.M. Rizzi and E. Sisu. 2013.
with
electrospray ionization mass spectrometric detection. Methods Mol. Biol.
