Principal Component Analysis Applied to Compare the Quality of Trace Elements in Moroccan and Indian Henna

ISSN: 2665-8488 JOURNAL OF ANALYTICAL SCIENCES AND APPLIED BIOTECHNOLOGY 2020, Vol. 2, Issue 1 An International Open Access, Peer Reviewed Research Journal Pages: 32-37

Analytical Chemistry DOI: 10.48402/IMIST.PRSM/jasab-v2i1.19558

Principal component analysis applied to compare the quality of trace elements in Moroccan and Indian henna

Abdelaziz AIT SIDI MOU 1*, Mohammed DAOUDI 2, Mohammed El Amine GHANJAOUI 1, Bouchaib ELGAMANY 3, Rachid SLIMANI 4

1 Laboratory of Physical Chemistry and Bioorganic Chemistry, Faculty of Sciences and Techniques of Mohammedia (FSTM) 28830 Mohammedia, Morocco 2 Departement of physics, Faculty of Sciences and Techniques, 26000 settat, Morocco 3Laboratory of Materials, Electrochemistry and Environment, Faculty of Sciences Ibn Tofail University of Kenitra, 14060 Kenitra, Morocco 4 Laboratory of Spectroscopy, Molecular Modeling, Materials, Nanomaterials, Water & Environment- CERNE2D, Faculty of Sciences, Mohammed V University in Rabat, B.P. 1014 RP, Rabat, Morocco

ARTICLE INFO ABSTRACT Received April 7th, 2020 This work examines the distributions and comparison of trace elements in Received in revised form May 29th, 2020 Accepted June 1st, 2020 Moroccan and Indian henna samples. Multivariate statistical methods involving principal component analysis and hierarchical cluster analysis are applied. The measurements were performed using 22 elements (Al, As, Ba, Ca, Cd, Ce, Cr, Cu, Keywords: Er, Fe, K, La, Li, Mg, Mn, Na, Ni, Sm, Sr, Ti, Y, Zn) at µg.g-1 level determined Moroccan henna, Indian henna, by ICP-OES. The coefficients calculated for samples indicated strong positive Multivariate statistical, correlations among firstly K, Mg, and Li, on the hand La and Ba, on the other hand Trace elements, Al and Fe and finally Ti and Er. These strong correlations between these variables Qualitative study. show the similarity of the phenomena at the origin in henna samples. Also, negative correlations between K, Mg and Li with Ba then La, which indicates that these elements have different origins, the same result showed by Fe and Ce. The multivariable analysis indicates that three principal components (PC) were significant with eigenvalues higher than 1, contributing 100% of the observed variance. Thus, principal component 1 and principal component 2 show major contributions, whereas principal component 3 has a relatively lower effect. The same results obtained by hierarchical cluster analysis.

1. Introduction: Moroccan and Indian henna available from herbalists is a natural product used traditionally by women for dyeing and coloration of hairs and hands [1]. Lawsone (2-hydroxy-1,4 naphthoquinone) is a chemical present in henna and it can cause a red-brown coloration of the skin. The some works reported that this plant was applied as corrosion inhibitor [2], coloration of textiles [3] and in adsorption [4]. Women mixed various herbs or other substances like p-phenylenediamine with henna in order to give it a stronger color, which can present a health risk due to the presence of toxic products such as trace elements [5- 7]. Therefore, it is essential to develop sufficiently analytical methods for precise determination of trace metals in cosmetic products. Recently, several techniques, including Atomic Absorption Spectrometry (AAS) [8, 9], Energy dispersive X-ray Microanalysis [10], Inductively Coupled Plasma Atomic Emission Spectrometry (ICP-AES) [11], and Inductively Coupled

(* )Corresponding author. Tel.: +212 6 75 25 14 21 E-mail address: [email protected]

Journal of Analytical Sciences and Applied Biotechnology Ait Sidi Mou et al.

Plasma Mass Spectrometry (ICP-MS) [12] have been used for the determination of elements like Pb and others in henna samples. At the same time, several analytical methods especially combined with chemometrics methods have been developed to study the characteristics and distribution of trace elements in different areas such as Food [13, 14], soil[15, 16], groundwater [17] agricultural soils [18], and the classification of essential oils and tea samples [19, 20]. In this work, we investigated for the first time the applied of different multivariate statistical for data interpretation of quality and comparison of trace elements in Moroccan and Indian henna.

2. Material and methods: 2.1. Sampling and analytical methods: Four samples of henna, two originally from Morocco and two from India were purchased from different Moroccan herbalists. Samples were stored in plastics flasks. The certified reference materials NIST 1573a Tomato Leaves (Gaithersburg, MD, USA) was analyzed in the same experimental conditions used for samples in order to evaluate the accuracy of the method. The blanks for the elements determined by ICP-AES were prepared by following all the analytical steps of the method in the absence of any sample or standard. As recommended in the respective certificate, the trace elemental concentration of the reference material was evaluated on a dry weight basis. All teflon digestion vessels were previously cleaned in a bath of 10% (v/v) nitric acid solution for 48 hours to avoid cross- contamination. Samples were grounded with a mixer grinder and the obtained powdered samples were sieved through a 250 μm mesh before their digestion. For microwave digestion, 0.1 g of each powder sample was weighed inside the digestion reactor. After that 2 mL of concentrated nitric acid, 2 mL concentrated hydrochloric acid, 2 mL concentrated hydrogen peroxide and 4 mL of ultra-pure water were added to the samples, waiting for the elimination of foam for 5 min and the reactors were closed and installed inside the oven. All samples were digested according to the following program: step 1, 5 min to reach 180ºC; step 2, 15 min at 180ºC; step 3, cooling down. After cooling to ambient temperature, the reactors were opened and sonicated to eliminate the nitrous oxide vapors, nine samples and one blank were treated simultaneously. The resultant solutions were transferred to a plastic flask and made up to 25 mL with ultra-pure water. Finally, the solutions were centrifuged to eliminate silica and the supernatant was directly feed to the ICP-AES [21].

2.2. Chemometric procedure for data analysis: Multivariate statistical methods are used to study the behavior, distribution and interrelationship of trace elements in henna samples. Correlations between trace elements are used almost in all related researches as it is simple and clear to optimize the information provided by the correlation’s matrix. Principal component analyses (PCAs) can reduce the complexity of large-scale data sets and are broadly used to elucidate relations among variables. The number of significant principal components is selected on the basis of the Kaiser criterion of eigenvalues higher than 1 (Kaiser, 1960) and a total explained variance is equal to or higher than 85%. Similarities among different elements concentrations in the samples were explored with Factor Analysis (FA). The extraction of factors was done through principal component analysis (PCA) and rotation using unrotated method. Cluster analysis (CA) is a group of multivariate technique, which allows the assembly of objects based on the characteristics. CA classifies objects, so that each object is similar to the others in the cluster with respect to a predetermined selection criterion. Hierarchical agglomerative clustering is the most common approach, which provides intuitive similarity relationships between any one variable and the entire data set and is typically illustrated by a dendrogram (tree diagram). The dendrogram provides a visual summary of the clustering processes, presenting a picture of the groups and their proximity with a dramatic reduction in dimensionality of the original data [22]. Correlation, principal component analysis (PCA) and Hierarchical cluster analysis (HCA) were carried out using STATISTICA 10 Software.

3. Results and discussion: 3.1. Characteristics of trace elements in henna samples: From the 42 studied elements, 22 elements (Al, As, Ba, Ca, Cd, Ce, Cr, Cu, Er, Fe, K, La, Li, Mg, Mn, Na, Ni, Sm, Sr, Ti, Y, Zn) were found in the Henna samples as reported by Ghanjaoui et al [21]. Table 1 summarises the concentration samples value of the trace element composition found in the Moroccan and Indian henna. From the data it can be seen that Moroccan samples are richer in Ca, K, Li, Mg, and Zn than Indian ones, while, the Indian samples are richer in Ba, Cd, Cr, Cu and Na than Moroccan samples. For the rest of elements, there is no significant variability between both types.

Journal of Analytical Sciences and Applied Biotechnology Ait Sidi Mou et al.

Table 1. Contents of trace elements (μg g-1) values in henna samples. Element MS 1 MS 2 IS1 IS2 K 16294 ± 1068 14802 ± 82 6176 ± 27 5173 ± 149 Na 1922 ± 23 1982 ± 34 8402 ± 421 13803 ± 264 Ca 17818 ± 1167 12461 ± 52 9537 ± 930 8218 ± 159 Mg 5125 ± 321 4468 ± 21 2235 ± 126 1950 ± 38 Fe 1092 ± 51 2399 ± 122 1728 ± 123 749.3 ± 0.5 Al 1504 ± 163 3290 ± 117 2026 ± 250 610 ± 32 Mn 75.2 ± 0.4 75 ± 2 120 ± 5 65 ± 2 Sr 218 ± 16 147 ± 3 295 ± 6 353 ± 8 Cu 8.29 ± 0.06 6.7 ± 0.4 21 ± 6 11 ± 2 Li 10.3 ± 1.1 8.9 ± 0.2 1.5 ± 0.1 0.87 ± 0.08 Ti 21 ± 3 70 ± 3 85 ± 6 22.0 ± 1.6 La 6.3 ± 0.3 6.4 ± 0.5 9.0 ± 1.3 8.6 ± 1.4 Ba 27 ± 2 35.4 ± 1.3 164311 ± 1755 167786 ± 5732 Ce 46 ± 2 20.0 ± 1.2 24.4 ± 0.1 58 ± 8 Zn 12.3 ± 1.6 15.3 ± 0.6 6.6 ± 0.8 7.6 ± 0.5 Er 21 ± 3 70 ± 2 89 ± 3 24 ± 1 Y 7.2 ± 0.8

3.2. Multivariate statistical analysis: 3.2.1. Correlation analysis: The results of the correlations between trace elements (Table 2) illustrate the origins and behaviors of trace elements in the samples. The coefficients calculated for samples indicated strong positive correlations among firstly K, Mg, and Li, on the hand La and Ba, on the other hand Al and Fe and finally Ti and Er. These strong correlations between these variables show the similarity of the phenomena at the origin in henna samples. Also, negative correlations between K, Mg and Li with Ba then La, which indicates that these elements have different origins, the same result showed by Fe and Ce. Therefore, the main elements present in Moroccan henna show a fort correlation were K, Mg, Ca, Li, Zn and a low correlation were Fe, Al and the mains elements presents in Indian henna show a fort correlation were Ba, La, Na, Sr, Cu and a low correlation were Mn, Ti. The difference observed of quantities of the trace elements can be explained by the varieties of henna like Lawsonia inermis family due to soil origin. Also, that the powder mixture added in Indian henna as p-phenylenediamine can affect the variety of henna.

Table 2 . Correlation matrix for trace elements in the samples. K Na Ca Mg Fe Al Mn Sr Cu Li Ti La Ba K 1 Na -0.943 1 Ca 0.905 -0.834 1 Mg 0.997 -0.936 0.930 1 Fe 0.360 -0.579 0.034 0.311 1 Al 0.519 -0.709 0.203 0.474 0.984 1 Mn -0.339 0.025 -0.229 -0.333 0.334 0.241 1 Sr -0.886 0.939 -0.639 -0.8584 -0.746 -0.851 0.131 1 Cu -0.702 0.460 -0.516 -0.688 -0.032 -0.163 0.893 0.560 1 Li 0.999 -0.935 0.908 0.998 0.341 0.502 -0.356 -0.877 -0.713 1 Ti -0.147 -0.168 -0.332 -0.183 0.818 0.721 0.768 -0.274 0.542 -0.169 1 La -0.979 0.872 -0.843 -0.972 -0.315 -0.474 0.510 0.863 0.829 -0.982 0.246 1 Ba -0.992 0.928 -0.850 -0.983 -0.410 -0.563 0.394 0.912 0.755 -0.992 0.128 0.991 1 Ce -0.252 0.534 -0.016 -0.215 -0.957 -0.924 -0.586 0.630 -0.224 -0.232 -0.918 0.157 0.274 Zn 0.901 -0.835 0.638 0.874 0.546 0.671 -0.479 -0.932 -0.819 0.900 -0.028 -0.944 -0.946 Er -0.192 -0.126 -0.363 -0.226 0.788 0.685 0.790 -0.226 0.582 -0.213 0.998 0.293 0.176

Journal of Analytical Sciences and Applied Biotechnology Ait Sidi Mou et al.

3.2.2. Principal component analysis: PCA was used to identify the sources of trace elements and their intrinsic relations, thus we analyzed the concentrations of trace elements from henna samples. The results are shown in below. Table 3 shows the results of the PCA applied to the trace element. This multivariable analysis indicates that three principal components (PC) were significant with eigenvalues higher than 1, contributing 100% of the observed variance. Thus, PC1 and PC2 show major contributions, whereas PC3 has a relatively lower effect.

Table 3. Total variance of trace elements in henna samples by PCA. Eigen value % Total - variance Cumulative % PC1 9.623076 60.14 60.14 PC2 5.295363 33.09 93.24 PC3 1.081561 6.75 100

Table 4 shows the factor loadings of the principal components. The first component (PC1) explained 60% of the total variance and displayed loadings on K, Na, Ca, Mg, Sr, Cu, Li, La, Ba and Zn but the factor loadings of Fe, Al, Mn, Ti, Ce and Er were negatively correlated elements than those of K, Ca, Mg, Li and Zn; therefore, the behavior of elements in this group may be independent. PC2 is composed mostly of Fe, Al, Mn, Ti, Er and Ce, contributes 33% of the variance.

Table 4. Factor Loadings of trace elements in henna samples. Factor 1 Factor 2 Factor 3 K 0.982368 -0.115461 0.147047 Na -0.947244 -0.204678 -0.246648 Ca 0.811421 -0.290728 0.507024 Mg 0.968553 -0.149252 0.199070 Fe 0.503025 0.822333 -0.265960 Al 0.647098 0.731281 -0.215622 Mn -0.333006 0.781142 0.528132 Sr -0.950570 -0.297225 0.089860 Cu -0.717399 0.541151 0.438742 Li 0.979965 -0.136784 0.144772 Ti -0.027340 0.998860 -0.039131 La -0.974988 0.220927 0.024282 Ba -0.994524 0.100609 -0.028276 Ce -0.370780 -0.928302 0.027898 Zn 0.958860 -0.013708 -0.283548 Er -0.075523 0.996832 -0.024962

In Figure 1, the graphical projection of the PCA results shows that the first component (PC1), accounts for 60% of the total variance and has loadings on K, Na, Ca, Mg, Sr, Cu, Li, La, Ba and Zn, reveals an association of strongly correlated elements. Where, Ca close to K, Mg, Li and Zn, suggested there are the main elements presented in Moroccan henna samples whereas Ba and La close to Na and Sr, are the main elements presented in Indian henna samples. The second component (PC2), which describes 33% of the total variance, has loadings on Fe, Al, Mn, Ti, Er and Ce. Er and Ti are similarly distributed in the two samples, while Ce, Mn and Cu are strongly distributed in Indian henna, although Fe and Al are distributed in Moroccan henna.

Journal of Analytical Sciences and Applied Biotechnology Ait Sidi Mou et al.

Factor Loadings, Factor 1 vs. Factor 2 Rotation: Unrotated 1.2 ErTi 1.0 Mn Fe 0.8 Al

0.6 Cu

0.4 La 0.2 Ba Zn 0.0 MgLiK Na Factor 2 Factor -0.2 Sr Ca

-0.4

-0.6

-0.8 Ce -1.0

-1.2 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Factor 1 Figure 1. Loading plots of trace elements in the space defined by two components.

3.2.3. Cluster analysis: The trace element was used to examine the similarity of takaout plant by means of hierarchical cluster analysis. The dendrogram obtained applying the 1-Pearson r and Ward’s method defines four different groups (Fig. 2). By using the Ward method the production of small clusters is favored in our studies. Cu and Mn are in cluster 1. They are remarkably different from the other elements in terms of 1–Pearson r, which implies a poorly effect from the other elements. Er, Ti, Al and Fe form cluster 2, mainly under the influence of Ce, Ba, La, Sr and Na form cluster 3 and finally, Zn, Ca, Mg, Li, and K form cluster 4, which result is in accord with the PCA.

Ward`s method 1-Pearson r 7

Linkage Distance Linkage

0 Cu Mn Er Ti Al Fe Ce Ba La Sr Na Zn Ca Mg Li K Figure 2. Hierarchical dendrograms from the hierarchical clustering method.

4. Conclusion: In this study, trace elements were used as input data for identification of the differentiating characteristics of Moroccan and Indian henna samples. The strong positive correlations among firstly K, Mg, and Li, on the hand La and Ba, on the other hand Al and Fe and finally Ti and Er. These strong correlations between these variables show the similarity of the phenomena at the origin in henna samples. Also, negative correlations between K, Mg and Li with Ba then La, which indicates that these elements have different origins, the same result showed by Fe and Ce. A principal component analysis performed in these data allowed a reduction in the data matrix to three principal components. The first component (PC1) explained 60% of the total variance and displayed loadings on K, Na, Ca, Mg, Sr, Cu, Li, La, Ba and Zn but the factor loadings of Fe, Al, Mn, Ti, Ce and Er were negatively correlated elements than those of K, Ca, Mg, Li and Zn; therefore, the behavior of elements in this group may be independent. PC2 is composed mostly of Fe, Al, Mn, Ti, Er and Ce, contributes 33% of the variance. The same results found by hierarchical cluster. In resume, the results obtained by 36

Journal of Analytical Sciences and Applied Biotechnology Ait Sidi Mou et al. multivariate statistical analysis shows that the Moroccan and Indian henna samples have a different some groups of trace elements, and can explain the diverse of quality and chemistry proprieties such as color of these products.

References: 1. El Hilaly, J., Hmammouchi, M., Lyoussi, B., (2003). Ethnobotanical studies and economic evaluation of medicinal plants in Taounate province (Northern Morocco). J Ethnopharmacol, 86, 149-158. 2. Jasim Al Hasan, N.H., Alaradi, H., Khadhim Al Mansor, Z.A., Jabbar Al Shadoo, A. H., (2019). The dual effect of stem extract of Brahmi (Bacopamonnieri) and Henna as a green corrosion inhibitor for low carbon steel in 0.5 M NaOH solution. Case Studies in Construction Materials 11 e00300. 3. Rahman Bhuiyan, M.A., Islam, A., Ali, A., Islam M.N., (2017). Color and chemical constitution of natural dye henna (Lawsonia inermis L) and its application in the coloration of textiles. Journal of Cleaner Production 167 (20), 14-22. 4. Davarnejad, R., Karimi Dastnayi, Z., Kennedy, J. F., (2018). Cr(VI) adsorption on the blends of Henna with chitosan microparticles: Experimental and statistical analysis. International Journal of Biological Macromolecules, 116, 281–288. 5. Onder, M., Atahan, C., Oztas, P., Oztas, O.M., (2001). Temporary henna tattoo reactions in children. American Journal of Contact Dermatitis, 12 (2), 126. 6. Brancaccio, R.R., Brown, Lance H., Chang, Y.T., Fogelman, J.P., Mafong, E.A., Cohen, D.E., (2002). Identification and Quantification of para-Phenylenediamine in a Temporary Black Henna Tattoo. American Journal of Contact Dermatitis.13 (1), 15-18. 7. Muhammad, H.S., Muhammad, S., (2005). The use of Lawsonia inermis linn. (henna) in the management of burn wound infections. African Journal of Biotechnology. 4 (9), 934-937. 8. Lekouch, N., Sedki, A., Nejmeddine, A., Gamon, S., (2001). Lead and traditional Moroccan pharmacopoeia. Sci Total Environ. 280, 39-43. 9. Kang, I.J., Lee, M.H., (2006). Quantification of para-phenylenediamine and heavy metals in henna dye. Contact Dermat. 55: 26–29. 10. Abdel-wahab, S.I., Ain N.M., Abdul A.B., Elhassan M.M., Ibrahim T.Z.T., (2009). Energy dispersive X-Ray Microanalysis of Elements content of Medicinal Plants used Traditionally as Anticancer cure. Res J Biol Sci, 4, 547-549. 11. Jallad, K.N., Espada-Jallad, C., (2008). Lead exposure from the use of Lawsonia inermis (Henna) in temporary paint-on- tattooing and hair dying. Sci Total Environ, 397, 244-250. 12. Bernth, N., Hansen, O.C., (2005). Survey of chemical substances in consumer products. Steen Færgemann Hansen og Eva Pedersen Teknologisk Institut. Danish Environmental Protection Agency, (Danish EPA), No 65. 13. Gonzálvez A., Armenta, S., de la Guardia, M., (2010). Adulteration detection of argan oil by inductively coupled plasma optical emission spectrometry. Food Chemistry, 121, 878–886. 14. Diniz, P.H.G.D., Pistonesi, M.F., Alvarez, M.B., B, de Araujo, M.C.U., (2015). Simplified tea classification based on a reduced chemical composition profile via successive projections algorithm linear discriminant analysis (SPA-LDA). Journal of Food Composition and Analysis, 39, 103–110. 15. Aslam, S.N., Huber, C., Asimakopoulos, A.G., Steinnes, E., Mikkelsen, Ø., (2019). Trace elements and polychlorinated biphenyls (PCBs) in terrestrial compartments of Svalbard, Norwegian Arctic. Science of the Total Environment, 685, 1127– 1138. 16. Fei, Q., Hongbing, J., Qian, L., Xinyue, G., Lei, T., Jinguo, F., (2014). Evaluation of trace elements and identification of pollution sources in particle size fractions of soil from iron ore areas along the Chao River. Journal of Geochemical Exploration 138, 33–49. 17. Moraa, A., Mahlknechta, J., Torres-Martíneza, A., (2018). Evolution and multivariable analysis of trace elements in groundwater of an agricultural area in a semi-arid region of Mexico. Energy Procedia, 153, 196–201. 18. Soriano-Disla, J.M., Janik, L., McLaughlin, M.J., Forrester, S., Kirby, J., Reimann, C., (2013). The EuroGeoSurveys GEMAS Project Team 1; The use of diffuse reflectance mid-infrared spectroscopy for the prediction of the concentration of chemical elements estimated by X-ray fluorescence in agricultural and grazing European soils. Applied Geochemistry, 29, 135–143. 19. Faraga, N.F., El-Ahmadyb, S.H., Abdelrahmana, E.H., Naumannc, A., Schulzc, H., Azzama, S.M., El-Kashourya, E.l.S.A., (2018). Characterization of essential oils from Myrtaceae species using ATR-IR vibrational spectroscopy coupled to chemometrics. Industrial Crops & Products, 124, 870–877. 20. Chen, Q., Zhao, J., Zhang, H., Wang, X., (2006). Feasibility study on qualitative and quantitative analysis in tea by near infrared spectroscopy with multivariate calibration. Analytica Chimica Acta, 572: 77–84. 21. Ghanjaoui, M.E., Cervera, M.L., El Rhazi, M., de la Guardia, M., (2014). Assessment of Trace Elements In Traditional Moroccan Cosmetics By Inductively Coupled Plasma Atomic Emission Spectroscopy. International journal of scientific & technology research, 3(10), 104-112. 22. Shrestha, S., Kazama, F., (2007). Assessment of surface water quality using multivariate statistical techniques: a case study of the Fuji river basin. Japan Environ Model Softw, 22, 464–475.