Environ. Sci. Technol. 1989, 23, 1164-1 167
Ozone Fading of Triphenylmethane Colorants: Reaction Products and Mechanisms Danlel Grosjean,t Paul M. Whitmore,* and Glen R. Cass'
Environmental Engineering Science Department and Environmental Quality Laboratory, California Institute of Technology, Pasadena, California 9 1 125 James R. Druzlk The Getty Conservation Institute, 4503 Glencoe Avenue, Marina del Rey, California 90292-6537
Triphenylmethane compounds were exposed in the dark to ozone in air (10 ppm for 4 days), and the exposed samples were analyzed by mass spectrometry. There was no evidence for reaction between ozone and triphenylmethane and between ozone and the triphenylcarbinol pararosaniline base. In contrast, the triphenylmethane cationic dye Basic Violet 14 yielded substituted benzophenones and other aromatic compounds. These products are consistent with a mechanism involving ozone addition on the unsaturated carbon-carbon bond. The results are briefly discussed in terms of the ozone fastness of triphenylmethane dyes used as artists' pigments and industrial colorants.
Table I. Elemental Composition Data for 030 Mauve Sample element concn, pg/sample
AI
Si P S
c1
element concn, pglsample
K
32.4 1.7 2.5 f 0.2 0.6 0.la 13.2 f 0.7 1.0 f 0.1
1.7 f 0.1 19.4 f 1.0 2.8 f O.l* 0.7 i 0.4O 1.9 1.2
Ca cu Sb Ba
*
ODiagnostic element for lakes of triphenylmethane dyes (6, 7). *This corresponds to a copper phthalocyanine content of 25.4 ue / samole. Table 11. Compounds Studied
Introduction Recent studies carried out in this laboratory have focused on the reaction mechanisms involved in the ozone fading of organic colorants (1, 2). These studies were prompted by observations that a number of colorants, natural and synthetic, used in the formulation of artists' pigments, faded substantially when exposed, in the dark, to air containing ozone levels comparable to those actually recorded in urban air, including indoor museum settings (3-5). Since ozone damage to works of art is of obvious concern to the art conservation community, it is important, with regard to mitigation measures, to develop a better understanding of the relationship between potential damage and colorant structure. The colorants previously shown to be among the most ozone-fugitive (3-5) included a commercial watercolor, Winsor and Newton 030 Mauve, which reportedly contained copper phthalocyanine (CP), and a lake of the triphenylmethane colorant Basic Violet 14. For this watercolor sample, fading was accompanied by a shift in color toward that of CP alone. In addition, a number of green CP watercolors and a blend of CP with a substituted alizarin lake did not fade under the same conditions (4). This suggested that triphenylmethane colorants may be particularly susceptible to ozone attack and prompted us to explore the mechanisms of their reaction with ozone. Elemental analysis of the 030 Mauve colorant (Table I), together with solubility tests, indicated that the sample contained, in addition to the expected mixture of CP and Basic Violet 14, substantial amounts of aluminum, calcium, sulfur, and other impurities. For this reason, we elected to focus our study directly on the triphenylmethane component of the Mauve colorant, i.e., Basic Violet 14. Basic Violet 14 is a typical member of an important group of colorants, the triphenylmethane cationic dyes. Other members of this group include malachite green, Dept. of Chemical Engineering;also with DGA, Inc., 4526 Telephone Road, Suite 205, Ventura, CA 93003. *Present address: Research Center on the Materials of the Artists and Conservator, Mellon Institute, 4400 Fifth Avenue, Pittsburgh, PA 15213. 1164
Environ. Sci. Technol., Vol. 23, No. 9, 1989
compds
MW
Basic Violet 14O pararosaniline base" triphenylmethanea benzoic acid phthalic acid 4-aminobenzoic acid 4-amino-3-methylbenzoic acid" benzophenone 4-aminobenzophen~ne~
337.9 305.4 244.3 122.1 166.1 137.1 151.2 182.2 197.2
a See
mp,
OC
250 (dec) 205 (dec) 92-94 122-123 210 (dec) 188 169-171 49-51 121-124
purity, , ,A % nm -95 97 99 99 99 98 99 98
544
structures in Fieure 2.
crystal violet, fuchsin, and rosaniline. These dyes are structurally derived from the corresponding triphenylmethane and triphenylcarbinols (Figure 11, with the chromophore associated to para substitution with amino groups (6,r). We therefore included one triphenylmethane and one carbinol in our study as an aid to the interpretation of the reactivity of Basic Violet 14 toward ozone. The poor light fastness of triphenylmethane-based inks (8) and dyes (9-11) has been known for some time. With the exception of one study of malachite green at high reactant concentrations in methanol (12),the reactivity of triphenylmethane colorants toward ozone has not been previously investigated.
Experimental Methods Compounds Studied. Basic Violet 14, pararosaniline base, and triphenylmethane were obtained from commercial sources and were used without further purification. In order to facilitate the identification of reaction products and the interpretation of mass spectra fragmentation patterns, mass spectra were also recorded for a number of benzoic acids and benzophenones. Relevant information regarding these compounds is given in Table 11. The corresponding structures are shown in Figure 2. Ozone Exposure Protocols. Triphenylmethane, pararosaniline base, and Basic Violet 14 were deposited on Teflon membrane filters (-10 mg/filter) and exposed in the dark to purified air containing 10 ppm ozone. These exposures were carried out for 96 h at 24 "C and 120%
0013-936X/89/0923-1164$01.50/0
-
0 1989 American Chemical Society
TRIPHENY LCARBINOL
TRIPHENYLMETHANE
TRIPHENYLMETHANE CATIONIC DYE
r2
OXIDATION
*
k R'
R'
'R
DYE MALACHITE GREEN (BASIC GREEN 4) BRILLIANT GREEN (BASIC GREEN 1 ) CRYSTAL VIOLET (BASIC VIOLET 3) PARAROSANILINE CHLORIDE (BASIC RED 9 )
N 'R
N+, A-
R'
C I NUMBER
R
XI
42000
-CH3
-H
-N(CH3);!
42040
-CH*CH3
-H
-N(CH2CH3)2
42555
-CH3
42500
-H
'R
x2
AC204HHSO;
-N(CH&
-N(CH&
CI-
-NH2
-NH2
CI-
Figure 1. Structures of some triphenylmethane dyes.
relative humidity, at a flow rate of 1.0 L/min. Control samples (not exposed to ozone) were included for each colorant. This exposure protocol has been described in detail in ref 1 and 2. Mass Spectrometry Analysis. After exposure to ozone, the Teflon filter deposited colorant samples were analyzed directly by probe-insertion MS. The MS analyses were carried out on a Kratos Scientific Instruments Model MS-25 hexapole, double-focusing magnetic sector instrument. All samples were analyzed in the chemical ionization (CI) mode with methane as the reagent gas. Operating conditions have been described in detail in ref 1,which also specifies our convention for structural information on reactants and products to be deemed positive, probable, or tentative. Reaction products that could go undetected include (a) those too volatile to remain on the substrate at the completion of the exposure experiment, (b) those formed in yields of less than 0.1%, (c) those with a molecular weight higher than 517, the upper limit of our mass scans, and (d) those with very low vapor pressure, no detectable amount of which could reach the spectometer's ionization source.
Results and Discussion Mass Spectra of Reactants and Other Reference Compounds. Methane chemical ionization mass spectra were recorded for all compounds listed in Table 11. The spectra of the three triphenylmethanes of interest were consistent with the structures given in Figure 2. The CI spectrum of Basic Violet 14, which was successfully recorded in spite of the ionic character of the compound, exhibited clusters of four fragments (m/z = x , x + 1, x + 2, x 3) as expected from the isotopic distribution of 35Cl and 37Cl. Details of this spectrum are discussed elsewhere (23) along with those of the spectrum of pararosaniline base. No impurities could be detected from examination of the spectrum of Basic Violet 14. That of pararosaniline base suggested diaminobenzophenone as a minor (-5%) impurity. The CI spectrum of benzoic acid was recorded and matched well with earlier literature data (14). The E1 and CI spectra of phthalic acid have been discussed previously (1). Those of triphenylmethane and of the other aromatic compounds are summarized in Table 111. The spectra of the substituted benzoic acids included MH (base peak), MH - H 2 0 and MH - COz as major fragments. Those of the benzophenones exhibited major MH and XC6H4C0fragments (X = H, NHJ.
+
Table 111. Methane Chemical Ionization Mass Spectra of Triphenylmethane and Relevant Aromatic Compounds mlz
245 244 243 168 167 166 165 91 79 78 77 138 137 123 122
120 94 92 152 151 134 108 107 106
abund, %I of base peak (BP)
fragment structure
Triphenylmethane (MW = 244) 3 MH 15 M 3 (Ph),C, Ph=C6H6 14 13C contribtn to BP" 100 (BP) (Ph)zCH 7 (Ph)zC 9
PhCH, H+ PhH, H+ PhH C ~ H (aromatic) B 4-Aminobenzoic Acid (MW = 137) 100 (BPI MH 41 M MH -NHb 11 6 M - NHb 43 MH - HzO' 44 MH - COP' 8 C6H4NHZ 4-Amino-3-methylbenzoic Acid (MW = 151) 100 (BP) MH 57 M 42 MH - HzOb 55 MH - COzb 16 M - COZ, MH - COOH 15 CH&HBNHZ
Benzophenone (MW = 182) 56 MH 18 M 7 13Ccontribtn to BP" 100 (BP) C&CO 22 C6H6 4-Aminobenzophenone (MW = 197) 100 (BP) MH 198 35 M 197 121 5.4 13c of 1 2 P 120 70 NHzCsH4CO 63 C&CO 105 92 9 NHZC6H4 77 10 CBH6 These fragments have diagnostic value. bAnd/or benzoic acid (MW = 122) as a minor imDuritv. 'TvDical of all benzoic acids. 183 182 106 105 77
Ozone-Triphenylmethane Reaction: Products and Tentative Mechanism. The mass spectra of control and Environ. Sci. Technol., Vol. 23, No. 9, 1989
1165
Table IV. Products of the Reaction between Gas-Phase Ozone and Teflon-Deposited Basic Violet 14
NHp,CI N H 2 G C Z OH o
reaction product
L7CaCH3
4-AMINO-BENZOIC
NH2
ACID
benzophenones NH2C6H4COCeH3(COOH)NHzc NH2C6H4COC6H3(CH20H)NHzC NH&6H4COC6H3(OH)NH2c NH2C6H4C0C6H3(CH3)NHzd NH2C&14COC6H4NH2d substituted benzenes phthalimided aminobenzoic acidc benzoic acidd benzoquinoned benzaldehydec phenol
NH2
BASIC VIOLET 14 I C 1 42510)
qCH3 NH2
COOH
H , N a E -OH
4-AMINO-
h
3 -METHYL
BENZOIC ACID
NH2 /I
PARAROSANILINE BASE
0
QClQrNH' /I 0
TRIPHENYLMETHANE
ozone-exposed samples of triphenylmethane were identical, i.e., there was no evidence for reaction between ozone and triphenylmethane or for any other triphenylmethane loss process in our exposure protocol. In particular, there was no evidence for even minor oxidation of triphenylmethane to triphenylcarbinol. In the same way, pararosaniline base did not react measurably with ozone under the conditions of our study: the CI spectrum of the exposed sample was identical with that of the control sample, with the exception of minor fragments (corresponding to less than 2% yields) at mlz = 107 and 109, which possibly correspond to benzaldehyde (MW = 106) and benzoquinone (MW = 108), respectively. These products may also have formed by oxidation of the diaminobenzophenone impurity. There H
'N'
H
gy
257 243 229 227 213
tentative* tentative tentative probable positive
147 137 122 108 106 94
148 138 123 109 107 95
tentative positive positive positive probable positive
+
H
"/
00,
+ 0,NH2
256 242 228 226 212
was no discernable change in the deep violet color of pararosaniline base after exposure to ozone. In contrast, Basic Violet 14 yielded a number of products including substituted benzophenones and aromatic compounds (Table IV). Major products included diaminoand methyldiaminobenzophenone,benzoquinone, benzoic acid, and a product of molecular weight 147 (MH = 148) whose mass fragmentation pattern matches that of phthalimide. Comparison of the CI spectra of control and exposed samples also indicated that a substantial amount of Basic Violet 14 was still present at the completion of the 96-h exposure experiment. Ozone attack on triphenylmethane compounds may involve several reaction sites: (a) the carbon-carbon unsaturated bond, (b) the carbon-nitrogen unsaturated bond, (c) the aromatic rings, (d) the amino group nitrogen atoms, and (e) the saturated aliphatic carbon atoms. Of these, reaction centers a and b are specific to Basic Violet 14, and reaction center d is absent in triphenylmethane.
- BENZOPHENONE
Figure 2. Structures of compounds studied.
H +
MH" identificn
Reagent adduct MH peak in methane chemical ionization mass spectrum. *See Experimental Section and ref 1 for definitions of tentative, probable, and positive identification. Minor product. Major product.
BENZOPHENONE
4 -AMINO
MW
1
B,,
__c
0
-
m:::
I
c -0'
O C HNH2 ,
@ = BASiC VIOLET
NH2
14
0.
&fC&;;;
NH2
1,2,3- TRIOXOLANE
CARBON -CARBON BOND SCISSION
f 0' I JQrCQCH3
NH2
NH2
CRIEGEE BlRADlCALS
c
10, (SAME AS FOR
@)
i
NH2
M,MNH2 I1
0
Flgure 3. Tentative mechanism for the ozone-Basic Violet 14 reaction. 1166
Environ. Sci. Technol., Voi. 23, No. 9, 1989
6-4
0 II
JfJrc7Q;;3 NH2
0
2
0
CARBONYL PRODUCTS
In the liquid phase (15)ozone reacts rapidly with unsaturated carbon-arbon bonds by 1,3 dipolar addition (see details of mechanism below); the electrophilic addition of ozone to the nitrogen atom in aromatic amines yields nitroaromatic and aromatic ring oxidation products, e.g., nitrobenzene and benzoquinone from aniline; the addition 7 of ozone to C=N bonds yields epoxides, e.g., C6H5CHNCH,O from C6H5CH=NCH3;and the reaction of ozone with aromatic rings may involve atom and bond attack to form phenols and ring-opening carbonyls, respectively (15). The relative reactivity of these reaction centers toward ozone in the liquid phase decreases according to the sequence >C=C< > amino > >C=N- > aromatic ring > saturated aliphatic
I
A similar sequence of reactivity toward ozone is observed in the gas phase as well (16). Our results for the heterogeneous reaction between gaseous ozone and substratedeposited triphenylmethanes also appear to be consistent with this reactivity sequence. Thus, triphenylmethane and pararosaniline base, which lack the more reactive >C=C< reaction center, were unreactive toward ozone under the conditions of our study. Of the two reaction centers specific to Basic Violet 14, one, the C=N bond, is even less reactive toward ozone than the amino groups of pararosaniline; the corresponding epoxide products were not observed. This leaves only the C=C bond as the initial site of ozone attack. A tentative mechanism for the ozone-Basic Violet 14 reaction, consistent with the above discussion and with currently accepted mechanisms for ozone addition to the C = C bond in substituted olefins (ref 15-17, and references therein), is outlined in Figure 3. The mechanism involves the steps of 1,3 dipolar addition, formation of the Criegee biradicals from the initial 1,2,3-trioxolane adduct, and carbon-carbon bond scission to yield the carbonyl products, in this case methyldiaminobenzophenone. Examination of resonance hybrid forms suggests diaminobenzophenone as another product of the same reaction sequence. Benzophenone products have also been identified in the oxidation of other triphenylmethane dyes by ozone in methanol (12), by hydrogen peroxide in the presence of light, both with dye powders and in solution (lo), and by photooxidation of the dyes as powders, on cellulose and in aqueous solution (9-11). The one-ring oxidation products listed in Table IV are structurally consistent with further oxidation of the parent dye and of the benzophenone products. Ozone Fastness of Triphenylmethane Dyes. While substituent effects, lake composition, substrate, and other factors may have significant effects on the ozone fastness of triphenylmethane colorants, the results of our study suggest that members of this group that contain unsaturated carbon-carbon bonds may fade upon exposure to ozone. Triphenylmethane colorants that are expected to be ozone-fugitive include the amino-substituted cationic dyes Malachite Green, Brilliant Green, Crystal Violet, Pararosaniline Chloride (see structures in Figure l),and Methyl Green, among others. Acid triphenylmethane dyes
(e.g., Acid Fuchsin, CI 426851, hydroxy-substituted members of the rosolic acid series (e.g., Aurin, CI 43800), and members of the xanthene chromophore group such as the Rhodamines (e.g., Basic Red 1,CI 45160) are also expected to fade upon exposure to ozone. Acknowledgments
Mass spectrometry analyses were carried out by Dilip K. Sensharma and John Wells, Department of Chemistry, University of California, Los Angeles, CA. The XRF analysis of the Mauve sample was performed at N U , Inc., Portland, OR. Dixie Fiedler prepared the several versions of the manuscript and Nancy Tomer drafted the figures. Registry No. 03, 1002815-6;triphenylmethane, 519-73-3; basic violet 14,632-99-5;pararosaniline, 467-62-9; benzoic acid, 65-850; phthalic acid, 88-99-3; 4-aminobenzoic acid, 150-13-0;4-amino3-methylbenzoic acid, 2486-70-6; benzophenone, 119-61-9; 4aminobenzophenone, 1137-41-3; diaminobenzophenone, 611-98-3; methyldiaminobenzophenone,121125-68-6;benzoquinone, 10651-4; benzaldehyde, 100-52-7; phenol, 108-95-2; phthalimide, a5-41-6.
Literature Cited (1) Grosjean, D.; Whitmore, P. M.; De Moor, P. C.; Cass, G. R.; Druzik, J. R. Enuiron. Sci. Technol. 1987,21,635-643. (2) Grosjean, D.; Whitmore, P. M.; Cass, G. R.; Druzik, J. R. Enuiron. Sci. Technol. 1988, 22, 292-298. (3) Shaver, C. L.; Cass, G. R.; Druzik, J. R. Environ. Sci. Technol. 1983, 17, 748-752. (4) Drisko, K.; Cass, G. R.; Whitmore, P. M.; Druzik, J. R. In Wiener Berichte uber Naturwissenschaft in der Kunst; Vendl, A., Pichler, B., Weber, J., Eds.; Verlag ORAC: Vienna, 198.5186; Vol. 213. (5) Whitmore, P. M.; Cass, G. R.; Druzik, J. R. J . Am. Znst. Conserv. 1987,26,45-58. (6) The Chemistry of Synthetic Dyes; Venkatamaran, K., Ed.; Academic Press: New York, 1952; Vol. 2, pp. 705-739. (7) The Chemistry of Synthetic Dyes; Venkatamaran, K., Ed.; Academic Press: New York, 1971; Vol. 4, pp. 103-157. (8) Arney, J. S.; Jacobs, A. J.; Newman, R. J.; J. Am. Znst. Conserv. 1979, 18, 108-117. (9) Iwamoto, K. Bull. Chem. SOC.Jpn. 1935, 10, 420-425. (10) Desai, C. M.; Vaidya, B. K. J. Zndian Chem. SOC.1954,31, 261-264. (11) Porter, J. J.; Spears, S. B., Jr. Text. Chem. Color. 1970,2, 191-195. (12) Matsui, M.; Takase, Y. Senryo to Yakuhin (Dyestuffsand Chemicals) 1982, 27, 10-17. (13) Grosjean, D.; Sensharma, D. K.; Cass, G. R. Chemical Ionization Mass Spectra of Artists’ Pigments and Dyes: Indigos, Anthraquinones and Triphenylmethanes. Manuscript in preparation, California Institute of Technology, 1988. (14) Hichikawa, H.; Harrison, A. G. Org. Mass Spectrom. 1978, 13,389-396. (15) Bailey, P. S. Ozonation in Organic Chemistry; Academic Press: New York, 1982; Vol. 2. (16) Atkinson, R.; Carter, W. L. P. Chem.Rev. 1984,84,437-470. (17) Martinez, R. L.; Herron, J. T. J. Phys. Chem. 1987, 91, 946-953. Received for review October 10,1988. Accepted March 20, 1989. This work was supported by a contract with the Getty Conservation Institute, Marina del Rey, CA.
Environ. Sci. Technol.. Vol. 23, No. 9, 1989 1167
