Degradation of Curcumin Dye in Aqueous Solution ... - OSA Publishing

Degradation of Curcumin Dye in Aqueous Solution and on Ag Nanoparticles Studied by Ultraviolet–Visible Absorption and Surface-Enhanced Raman Spectroscopy ˜ AMARES, J. V. GARCIA-RAMOS, and S. SANCHEZ-CORTES* M. V. CAN Instituto de Estructura de la Materia. CSIC., Serrano, 121. 28006-Madrid, Spain

The chemical degradation of curcumin (CU) in aqueous solution and on silver nanoparticles was studied by means of ultraviolet (UV)-visible absorption and surface-enhanced Raman (SERS) spectroscopy at different pH levels and upon light irradiation. CU undergoes a chemical degradation in aqueous solution mainly when the pH is increased. The CU degradation is catalytically enhanced in the presence of Ag nanoparticles. In general, CU degradation implies two steps: (1) the breakdown of the interring chain connecting the two CU aromatic side rings, producing smaller phenolic compounds rich in carboxylate groups, and (2) polymerization of the resulting phenolic products, giving rise to phenolic polymeric products. The degradation–polymerization mechanism can be modulated depending on the experimental conditions. The chemical photoproducts resulting from the visible light irradiation are similar to the polycatechol products obtained when catechol is adsorbed on Ag nanoparticles. Index Headings: Curcumin; Catechol; Surface-enhanced Raman spectroscopy; SERS; Degradation; Photopolymerization.

INTRODUCTION Curcumin (CU) (Fig. 1) (1,7-bis(4-hydroxy-3-methoxyphenyl)-1,6-heptadiene-3,5-dione) is a natural yellow-orange dye that gives its name to the curcuminoid family. It is the main component of Curcuma longa L. rizomes, native to India and Southeast Asia. CU was known in Mesopotamia and was used by ancient Greeks and Romans. It was imported in 1612 to Europe, becoming a popular dye for silk scarves. This phenolic compound has also been employed for a long period of time as dye, medicine, and food additive1 in the Asian countries. Today it is also used as spice, in curry, and as food dye (E-100) and preservative. The medicinal activity of CU has been known since ancient times. In addition to its powerful antioxidant activity,2–5 this dye has anti-inflammatory properties,6 HIV antiproteases activity,7 and cancer preventive properties.8 It is well known that CU decomposes under alkaline conditions, by means of a rapid hydrolytic degradation,9 and light exposure,10 leading to other phenolic compounds such as ferulic acid, feruroyl methane, vanillin, and vanillic acid11 (Figs. 1b and 1c). The high reactivity of CU has been attributed to the hydrogen atoms in the b-diketone moiety. In recent works, we have successfully used surface-enhanced Raman spectroscopy (SERS) to study the chemical modification and polymerization of a large list of phenolic compounds.12–16 In these studies, we have reported the importance of pH and light irradiation on the degradation of these compounds. The SERS technique was demonstrated to be a very useful technique in the study of the stability of these polyphenolic compounds, and

this is why we have applied SERS to the study of the chemical degradation of CU. The SERS technique is based on the giant intensification of Raman emission for adsorbates in the presence of metal nanostructures.17,18 This technique can be applied in the study of poorly soluble compounds in water, as very low concentrations are required, with the additional advantage of the fluorescence quenching occurring on the metal surface,19 mainly in the case of dyes. The study of the chemical degradation of CU is interesting due to the enormous importance of these molecules from several points of view. CU chemical degradation could lead to a color change of importance for the artistic objects where this pigment could be present. For the medicinal and food industries, the study of CU degradation is also crucial to understanding the effects of the degradation products regarding their consequences in human health and the possible change in the color and organoleptic properties of food products in which CU is used as an additive. Curcumin degradation has been previously studied by ultraviolet (UV)-visible absorption spectroscopy and chromatographic techniques,9–11 which are the techniques usually employed in the study of the degradation of organic compounds. However, there have been no studies so far dealing with the application of SERS, a very sensitive technique that has been applied even in single-molecule detection,20 to the study of CU degradation. The application of SERS to the vibrational study of CU is interesting as metal nanoparticles induce a fluorescence quenching of the dye and allow its detection at trace concentrations due to the high sensitivity. However, this application is limited because of the chemical instability of the dye. Thus, a preliminary study of CU stability both in solution and on nanostructured metal surfaces is needed. The application of the SERS technique to the vibrational and degradation study of CU is advantageous with respect to other analytical techniques due to the fact that the same surface that induces the chemical change is the support that allows the detection of the CU degradation products. In addition, intermediate compounds produced during the chemical degradation can be stabilized due to their adsorption on the metal surface and can also be detected by SERS even at very low concentrations. In this work we have conducted a UV-visible and SERS spectroscopic study of CU both in solution and on Ag nanoparticles in order to better understand the reactivity of this dye. We have focused our attention on the effect of pH spectra and light irradiation due to the demonstrated importance of these factors on the dye stability.

EXPERIMENTAL Received 5 May 2006; accepted 5 September 2006. * Author to whom correspondence should be sent. E-mail: imts158@iem. cfmac.csic.es.

1386

Volume 60, Number 12, 2006

Materials. Curcumin was purchased from Acros (98%). Stock solutions of the dye at a concentration of 102 M were

0003-7028/06/6012-1386$2.00/0 Ó 2006 Society for Applied Spectroscopy

APPLIED SPECTROSCOPY

FIG. 1. (a) Structure of CU, and the degradation products reported in the literature: (b) trans-6-(4 0 -hydroxy-3 0 -methoxyphenyl)-2,4-dioxo-5-hexanal (enol form), and (c) vanillic acid.

prepared in DMSO. In this solvent no chemical change with time was observed. All the reagents employed were of analytical grade and purchased from Sigma and Merck The aqueous solutions were prepared by using triply distilled water. Silver colloids were prepared by reduction of silver nitrate with tri-sodium citrate dihydrate and hydroxylamine hydrochloride according to a previous work.21 The colloids were activated before adding CU. This activation consisted of a partial aggregation of the colloidal particles, and to accomplish this, 12 and 20 lL of 0.5 M potassium nitrate solution were added to 0.5 mL of the citrate and hydroxylamine Ag colloids, respectively. This activation is a prerequisite for SERS spectra to be observed at a higher intensity, as we have demonstrated in other works.22,23 Then, 1 lL of the CU solution in DMSO was added to 1 mL of the Ag colloid in order to reach a CU concentration of 105 M. Therefore, the final DMSO concentration in the solution was 0.1%. Nitric acid and sodium hydroxide were employed to vary the pH. In the last case, no nitrate was added, since the hydroxide also induced the colloid aggregation. Samples for the UV-visible spectroscopy in water were prepared by adding 1 lL of the CU solution in DMSO to 1 mL of water so that the CU and DMSO were at concentrations of 105 M and 0.1%, respectively. Quartz cells with an optical path of 1 cm were used. For the analysis of the influence of Ag nanoparticles in CU degradation, an aliquot of the same samples prepared for the SERS measurements was centrifuged for 20 minutes at 10 800 rpm, in order to remove the Ag nanoparticles. SERS spectra were recorded 1 min after the sample preparation. Furthermore, for the irradiation experiments, 20 lL of the SERS sample were placed in a capillary and irradiated at different times with the 457.9 nm line of a Spectra Physics Model 165 Arþ laser using a radiation power of 80 mW. Instrumentation. The SERS spectra with excitation in the visible region were recorded using a micro-Raman Renishaw RM2000 instrument, working under macro conditions, by using the 514.5 nm radiation line of a Spectra Physics Model 163-C4210 Arþ laser. The laser power at the sample was 1 mW. The resolution was set at 4 cm1 and the geometry of micro-Raman measurements was 1808. Fourier transform Raman and FT-SERS spectra were obtained by using an RFS 100/S Bruker spectrometer. The 1064 nm line, provided by a Nd:YAG laser, was used as the excitation line. The resolution was set to 4 cm1 and a 1808

FIG. 2. UV-vis spectra of 105 M CU in DMSO/H2O 0.1% (v/v) at different pH values.

geometry was employed. The output laser power was 150 mW in the case of SERS and solution measurements and 50 mW in the solid samples. Ultraviolet–visible absorption spectra were recorded with a Cintra 5 spectrometer. The samples were put in cuvettes with an optical path of 1 cm.

RESULTS AND DISCUSSION Ultraviolet–Visible Spectroscopy of Curcumin. Effect of pH. Ultraviolet-visible spectra of 105 M CU in DMSO/ H2O 0.1% (v/v) solutions at different pH values are shown in Fig. 2. On increasing the pH value the absorption is enhanced while the maximum position shifts from 424 nm at pH 3 to 466 nm at pH 11. This fact is attributed to the deprotonation of CU hydroxyl groups. The pK values of an aqueous solution of 105 M CU have been calculated by Bernabe´-Pineda et al.:24 pK1 ¼ 8.38 6 0.04, pK2 ¼ 9.88 6 0.02, and pK3 ¼ 10.51 6 0.01. From these data and absorption spectra, we can deduce that successive CU ionizations induce the observed absorption maximum shift to higher wavelength. Thus, the maximum at 466 may correspond to the CU3 trianion. Besides this change, other smaller absorptions were seen at 272 and 355 nm at alkaline pH, revealing that the pH increase could also induce a chemical degradation of CU leading to smaller aromatic compounds. Effect of Time. Figure 3 shows UV-visible spectra of solutions of 105 M CU in DMSO/H2O 0.1% (v/v) at pH 5.0 and 12.0, as well as their time evolution. At pH 5 (Fig. 3a) a quick decrease of the 424 nm absorption intensity of the maximum is observed after the first 5 minutes, accompanied by the absorption increase at 355 nm. Then, the spectrum remains almost unchanged during the following three hours, with a slight increase of the 355 nm absorption. After 24 hours, a great decrease of the intensity of the maximum at 424 nm is observed. The absorption at 355 nm would correspond to a first intermediate degradation product bearing a shorter conjugated p bond system in comparison to CU. Most probably this first degradation product corresponds to the intermediate


1387

FIG. 3. Time evolution of UV-visible spectra of 105 M CU in DMSO/H2O 0.1% (v/v) at different pH values: (a) 5 and (b) 12.

product of Fig. 1b, which has been identified by many authors as the main degradation product of CU.9 Ultraviolet–visible spectra obtained at pH 12 (Fig. 3b) show a progressive decrease of the absorption intensity at 466 nm and a slight shift of this maximum to lower wavelengths. Also, an intensification of bands at 270 and 350 nm is observed. Three hours later a strong degradation of CU was produced, as deduced from the large weakening of the 466 nm absorption and the appearance of new absorption maxima at 247, 299, 350, and 567 nm, which are better seen in the difference spectrum (180 min 0 min) shown in Fig. 3b. The changes occurring at alkaline pH and longer degradation times suggest that the CU degradation takes place by the previous formation of smaller aromatic species (responsible for the band at 350 nm and then for those at 247 and 299 nm), which may correspond to the degradation products indicated in Fig. 1, and which progressively polymerize, giving rise to bigger polymeric products (responsible for the absorption maximum at 567 nm) only seen at longer times. Effect of Ag Nanoparticles. The CU degradation was also followed when the molecule is in the presence of Ag nanoparticles at pH 12 (Fig. 4a). In order to observe the CU absorption spectra it was necessary to remove the Ag nanoparticles by centrifugation, since the extinction due to the resonance of metal plasmon overlaps the CU absorption. The changes observed on Ag are similar to those observed in DMSO/H2O at the same pH (Fig. 4b): a strong absorption decrease of the band at 450 nm and the intensification of the 355 nm absorption band. However, the changes occurring in the presence of Ag nanoparticles are quicker in the first stages due to the catalytic effect of the metal. Figure 4b shows the UV-visible absorption spectra of CU in DMSO/H2O and in the presence of Ag nanoparticles 180 min after the sample preparation. The main differences seen between both spectra are the shift of the 450 nm maximum

1388


FIG. 4. (a) UV-visible spectra of 105 M CU on Ag nanoparticles at pH 12, after removing the nanoparticles; (b) comparison between the UV-visible spectra of CU on Ag (solid line) and in aqueous medium (DMSO/H2O, 0.1 %) at pH 12 and after 180 min (dashed line).

toward 417 nm and the remarkable absorption increase observed above 500 nm. The larger shift to shorter wavelengths is attributed to the progressive acidification of the sample due to the formation of carboxylic groups in the medium as a consequence of the CU degradation. Furthermore, the absorption increase at longer wavelengths is attributed to the subsequent polymerization of CU degradation products, catalyzed by the Ag nanoparticles. Surface-Enhanced Raman Spectra. Effect of pH. Figure 5 shows the SERS spectra of 105 M CU at different pH values, along with FT-Raman spectra of solid CU and in DMSO exciting at 1064 nm. The Raman spectrum of CU in DMSO solution (Fig. 5b) does not significantly change in relation to CU in the solid state (Fig. 5a), except for bands appearing in the 1320–1200 cm1 region. Since these bands can be attributed to m(C–O) and d(C¼C–H) vibrations of the inter-ring chain,25 the observed changes are probably due to conformational changes occurring in this chain in different media. The SERS spectrum of CU at pH 4.0 (Fig. 5c) is similar to the Raman spectrum of CU in DMSO solution, thus indicating that in general terms the molecule does not undergo a significant chemical change at these conditions. Nevertheless, the weakening of the 961 cm1 band and some changes occurring in the 1300–1100 cm1 region indicate that the

FIG. 6. SERS spectra on Ag citrate nanoparticles of 105 M CU at pH (a) 4 , (b) 6, and (c) 12. Excitation at 514.5 nm. FIG. 5. (a) FT-Raman spectra of CU in powder form and (b) in DMSO (102 M, * ¼ solvent bands). FT-SERS spectra of 105 M CU on Ag citrate nanoparticles at pH (c) 4, (d) 6, and (e) 12. Excitation at 1064 nm.

molecule is rearranged upon adsorption on the metal surface, inducing conformational changes that mainly involve the interring chain connecting both aromatic rings. However, at pH 6 (Fig. 5d) the SERS spectrum undergoes evident changes, among which the most outstanding one is the practical disappearance of the alkene m(C¼C) band at 1632 cm1 and the appearance of strong bands at 1487, 1264, 1158, and 1124 cm1, while the aromatic m(C¼C) band at 1597 cm1 shifts to 1576 cm1. These changes are even more pronounced at pH 12 (Fig. 5e), where broad features at 1581, 1486, 1387, 1260, 1162, and 1126 cm1 are seen. These spectral changes indicate that CU undergoes a clear degradation on increasing the pH, consisting of the elimination of the inter-ring aliphatic chain connecting the aromatic moieties with the appearance of other aromatic compounds, with new C–O and C–O–C bonds attributed to alcohol and ether groups, and carboxylate groups responsible for the bands at 1387 and 930 cm1. These products could correspond to vanillic acid (Fig. 1c) and/or other related hydroxybenzoic compounds.9,26 The broadness of the SERS features suggests a large heterogeneity of the resulting CU products, which may include phenol and carboxylate aromatic compounds together with polymeric aromatic compounds as also deduced from the UV-visible spectra. Figure 6 shows SERS spectra of CU at various pH values but with excitation at 514.5 nm. When comparing these spectra with those obtained at 1064 nm, different profiles can be seen for each pH value. However, the SERS spectra at pH 6 and 12 obtained at 514.5 nm are similar to those at pH 4 and 6, respectively, at 1064 nm, i.e., an apparent lower degradation of CU is seen at 514.5 nm. Since the CU products have more

intense absorptions toward the UV region, this effect is explained on the basis of the selective and resonant enhancement of still non-degraded CU when using the line at 514.5 nm as the excitation line. Effect of Light Irradiation. Figure 7 shows FT-SERS spectra of 105 M CU at pH 6.0 on Ag nanoparticles prepared by reduction with citrate and after irradiation for different times using the 457.1 nm line. The effect of light irradiation was twofold: a considerable SERS intensity increase and, on the other hand, a clear change in the spectral profile. The irradiation induces a general enhancement of the SERS bands, which, 150 min after the irradiation, becomes one order of magnitude more intense. This effect can be due to either a modification of the aggregation state of the Ag nanoparticles or an induced adsorption of CU products upon light irradiation. On the other hand, the spectral profile changes resulting from the irradiation can be better seen in the difference spectrum between the SERS obtained after 150 (Fig. 7c) and 30 min (Fig. 7b) of laser irradiation, as shown in Fig. 8a. In the last spectrum, narrow bands corresponding to the degradation products occurring after the CU adsorption and irradiation on the Ag nanoparticles are seen, which contrast with the broad features seen at alkaline pH (Fig. 8b). In particular, the intense bands below 700 cm1 reveal the presence of polymerized products.14,16 We have seen that the SERS difference spectrum shown in Fig. 8a is surprisingly very similar to the SERS spectrum of catechol (Fig. 8c). In previous work we have demonstrated that catechol is able to polymerize on Ag nanoparticles, giving rise to polycatechol,14 which in essence is a polymer composed of aromatic benzene rings linked through inter-ring C–O–C or C–C bonds, which corresponds to the spectrum of Fig. 8c. A similar result was observed for catechol adsorbed on TiO2 by LanaVillarreal et al.27 Thus, one of the consequences of light irradiation is the strong polymerization of the small CU hydroxybenzoic products resulting from the CU degradation


1389

FIG. 7. FT-SERS spectra of 105 M CU (a) on Ag citrate nanoparticles at pH 6 before irradiating, and after irradiation at 457.9 nm for (b) 30 min and (c) 150 min. Excitation at 1064 nm.

upon its adsorption on the metal surface, yielding polyaromatic compounds similar to polycatechol. The absence of lower intensity of bands below 700 cm1 at high pH indicates that the polymerization is much more limited in these conditions, under which CU seems to be degraded to smaller aromatic products rich in carboxylate groups. Furthermore, the polymerization of CU products, observed upon light irradiation in the presence of Ag nanoparticles, is related to the higher absorption observed in the region above 500 nm in Fig. 4a. The only differences observed between the SERS spectrum of irradiated CU and polycatechol are better seen in the difference spectrum of Fig. 8d and may correspond to the aliphatic groups resulting from the catechol ring opening, which does not occur in CU. It is interesting to note that no irradiation effect was observed when CU was adsorbed on hydroxylamine-prepared Ag nanoparticles. This is attributed to the existence on its surface of a relatively high number of chloride ions adsorbed on its surface. These ions are able to passivate the catalytic centers existing on the Ag surface, as has also been reported by other authors.28,29 Hence, the photopolymerization induced by the laser irradiation needs the existence on the Ag surface of active sites.

CONCLUSION Curcumin undergoes a chemical degradation in aqueous solution that is stronger at high pH and in the presence of Ag nanoparticles. In general, this degradation implies two steps: (1) the breakdown of the inter-ring chain connecting the two CU aromatic side rings, producing smaller phenolic compounds rich in carboxylate groups, and (2) polymerization of the resulting phenolic products, giving rise to phenolic polymeric products. The CU degradation is catalytically enhanced in the presence of Ag nanoparticles and mainly at extreme conditions: at alkaline pH and upon laser irradiation.

1390


FIG. 8. (a) Difference spectrum (Fig. 7c – Fig. 7b); (b) FT-SERS spectrum of 105 M CU on Ag citrate nanoparticles at pH 12, (c) FT-SERS spectrum of catechol (7 3 103 M) on Ag at pH 12; and (d) difference between spectra (c) and (a).

In particular, the laser irradiation of CU on Ag nanoparticles and at alkaline pH induces the formation of polymerization photoproducts similar to those formed by polymerization of catechol rendering polycatechol. ACKNOWLEDGMENTS This work was supported by Ministerio de Educacioń y Ciencia ( FIS200400108 grant) and Comunidad Autonoma de Madrid (GR/MAT/0439/2004 and S-0505/TIC/0191 (MICROSERES) grants). This work also received support from the Red Tema´ tica del Patrimonio Histo´ rico (C.S.I.C.). M.V.C. acknowledges C.S.I.C. for an I3P fellowship founded by the European Social Fund.

1. V. S. Govindarajan, Crit. Rev. Food Sci. Nutr. 12, 199 (1980). 2. T. Masuda, K. Hidaka, A. Shinohara, T. Maekawa, Y. Takeda, and H. Yamaguchi, J. Agric. Food Chem. 47, 71 (1999). 3. T. Masuda, K. Hidaka, T. Toi, H. Bando, T. Maekawa, Y. Takeda, and H. Yamaguchi, J. Agric. Food Chem. 50, 2524 (2002). 4. S. V. Jovanovic, S. Steenken, C. W. Boone, and G. Simic, J. Am. Chem. Soc. 121, 9677 (1999). 5. Y. M. Sun, H. Y. Zhang, D. Z. Chem, and C. B. Liu, Org. Lett. 4, 2909 (2002). 6. Y. J. Surh, Mutat. Res. 428, 305 (1999). 7. S. Zhihua, R. Salto, J. Li, C. Craik, and P. R. Ortiz de Montellano, Bioorg. Med. Chem. 1, 415 (1993). 8. B. B. Aggarwal, A. Kumar, and A. C. Bharti, Anticancer Res. 23, 363 (2003). 9. Y. J. Wang, M. H. Pan, A. L. Cheng, L. I. Lin, Y. S. Ho, C. Y. Hsieh, and J. K. Lin, J. Pharm. Biomed. Anal. 15, 1867 (1997). 10. A. Sundaryono, A. Nourmamode, C. Gardrat, S. Grelier, G. Bravic, D. Chasseau, and A. Castellan, Photochem. Photobiol. Sci. 2, 920 (2003). 11. H. H. Tonnesen and J. V. Greenhill, Int. J. Pharm. 87, 79 (1992). 12. S. Sanchez-Cortes and J. V. Garcıá-Ramos, Spectrochim. Acta, Part A 55, 2935 (1999). 13. S. Sanchez-Cortes and J. V. Garcıá-Ramos, J. Coll. Interface Sci. 231, 98 (2000). 14. S. Sanchez-Cortes and O. Francioso, J. V. Garcıá-Ramos, C. Ciavatta, and G. Gessa, Coll. Surf. A 176, 177 (2001).

15. M. Alvarez-Ros, S. Sanchez-Cortes, O. Francioso, and J. V. GarcıáRamos, J. Raman Spectrosc. 32, 143 (2001). 16. S. Sanchez-Cortes and J. V. Garcıá-Ramos, Appl. Spectrosc. 54, 230 (2000). 17. L. C. T. Shoute and G. R. Loppnow, J. Chem. Phys. 117, 842 (2002). 18. M. Moskovits, Rev. Mod. Phys. 57, 783 (1985). 19. P. C. Lee and D. Meisel, J. Phys. Chem. 86, 3391 (1982). 20. S. Nie and S. R. Emory, Nature (London) 275, 1102 (1997). 21. M. V. Canãmares, S. Sanchez-Cortes, D. Go´mez-Varga, C. Domingo, and J. V. Garcıá-Ramos, Langmuir 18, 8546 (2005). 22. S. Sanchez-Cortes, J. V. Garcia-Ramos, and G. Morcillo, J. Coll. Interface Sci. 167, 428 (1994). 23. S. Sanchez-Cortes, J. V. Garcia-Ramos, G. Morcillo, and A. Tinti, J. Coll. Interface Sci. 175, 358 (1995).

24. M. Bernabe´-Pineda, M. T. Ramı´rez-Silva, M. Romero-Romo, E. Gonza´lezVergara, and A. Rojas-Hernańdez, Spectrochim. Acta, Part A 60, 1091 (2004). 25. T. M. Kolev, E. A. Velcheva, B. A. Stamboliyska, and M. Spiteller, Int. J. Quant. Chem. 102, 1069 (2005). 26. A. Khurana and C.-T. Ho, J. Liq. Chromatogr. 11, 2295 (1988). 27. T. Lana-Villarreal, A. Rodes, J. M. Pe´rez, and R. Go´mez, J. Am. Chem. Soc. 127, 12601 (2005). 28. M. Alvarez-Ros, S. Sanchez-Cortes, O. Francioso, and J. V. GarcıáRamos, Can. J. Anal. Chem. Spectrosc. 48, 132 (2003). 29. C. Minero, G. Mariella, V. Maurino, D. Vione, and E. Pelizzetti, Langmuir 16, 8964 (2000).


1391