Iournal of

Agricultural
MAY 1993
andFood
VOLUME 41, NUMBER 5
Chemistrv
0 Copyright 1993 by the American Chemical Society

SHORT COMMUNICATIONS

Benzene Production from Decarboxylation of Benzoic Acid in the

Presence of Ascorbic Acid and a Transition-Metal Catalyst
INTRODUCTION specific indicator of hydroxylradical (Winston et al., 1983).
Ascorbic acid (vitamin C) is a natural component of Although the three isomeric hydroxybenzoates have also
been measured as products of hydroxyl radical attack on
many foods and is often added to foods and beverages as benzoate (Loeble et al., 1951; Armstrong et al., 1960), we
a vitamin supplement or fortifier and promoted as an have not found any report of benzene production by this
antioxidant. Transition metals, e.g., Cu(I1) and Fe(III), reaction.
can catalyze the one-electron reduction of 02 by ascorbic
acid to produce the superoxide anion radical, which The present study shows that hydroxyl radical, gener-
undergoes spontaneous disproportionation to produce ated by the metal-catalyzed reduction of 02 and H202 by
hydrogen peroxide. Subsequent metal-catalyzed reduction ascorbic acid, can attack benzoic acid to produce benzene
of HzOz by ascorbic acid can generate the hydroxyl radical under conditions prevalent in many foods and beverages.
(Halliwell and Gutteridge, 1981): Since benzene has been shown to be carcinogenic(Maltoni

-
and Scarnato, 1979),ita potential formation in foods during
processing and storage should be of some concern.
Cu2++ H,Asc Cu+ + HAsc' (1)

+
c u + 0, - cu2++ 0,- (2)
MATERIALS AND METHODS
Desferal mesylate (desferioxamine) was a gift from Ciba
Pharmaceutical Co. Ascorbic acid stock solutions were prepared
20,- + 2H+ - 0, + H,O, (3)
fresh daily in 0.1 M HCl to stifle autoxidation. Hydrogen peroxide
stock solutionswere standardized colorimetricallywith saturated

Cu+ + H,O, - Cu2++ OH- + OH' (4)

TiOSO4 solution in 2 M &So4 (c = 717 M-l cm-l at 410 nm) (Ellis
and Sykes, 1973).
High-PerformanceLiquid Chromatography. The HPLC
system consistedof a Knauer 64 pump, a Rheodyne 7125injection
Sodium benzoate has been widely used as a food valve with 20-pL sample loop, and a Knauer variable-wavelength
UV detector operating at 204 nm with the response recorded on
preservative for many years and is generally recognized as a Kipp and Zonen BD41 strip chart recorder. The analytical
safe (GRAS). The optimum antimicrobial activity occurs column was a 100 X 4.6 mm Spheri-5reversed-phase Cs cartridge
in the pH range 2.5-4.0, indicating benzoic acid (pK, = with a 30 X 4.6 mm guard cartridge of the same packing material
4.2) is the active form of this antimicrobial agent, making (Applied Bioscience, Inc.). The mobile phase was 25% aqueous
it most suitable for foods and beverageswhich are naturally acetonitrile (spectral quality, Baker Chemical Co.) filtered
acidic (Chichester and Tanner, 1968). The sodium salt is through 0.2-pm membrane filters prior to use each day and was
preferred because of ita much higher solubility in aqueous pumped at a flow rate of 3.0 mL/min.
media. It is generally permitted in beverages a t a level of A 2.0 p M aqueous benzene standard solution was stored
less than 0.1% (ca. 7 mM), although higher levels are refrigerated in a capped container. Dilutions of this standard in
permitted for some foods. water were used as the working standards for HPLC analysis.
Benzene was well separated from other componentsin the reaction
Mathew and Sangster (1965) have shown that sodium mixtures under these conditions and was quantitated by com-
benzoate is decarboxylated by hydroxyl radical attack. parison of peak height with the standard solutions.
Sagone et al. (1980) used 14C-labeledbenzoate decarbox- Reaction Conditions. Reaction mixtures were placed in a
ylation to show hydroxyl radical generation by phagocytic constant-temperature bath at 25 OC for 15 min, at which time
leukocytes. Alkoxy1 radicals do not decarboxylate ben- an aliquot was removed and injected into the HPLC system for
zoate, suggesting benzoate decarboxylation is a relatively analysis. All experimentalconditionswere prepared and analyzed
0021-0581I931I44l-O693%04.00/0 0 1993 American Chemlcal Soclety
694 J. Agrlc. Food Chem., Vol. 41, No. 5, 1993 Short Communlcatlons

Table I. Effect of Metal Ions and Ligands on Benzene

Production from Benzoic Acid
addition [benzene],nM
none 6.7 f 1.9
0.1 mM desferal NDb
0.05 mM CuSO4 + 0.1 mM desferal 23.3 f 1.7
0.25 mM CuSO4 + 0.1 mM desferal 27.0 f 1.1
1.0 mM CuSO4 + 0.1 MM desferal 37.3 f 3.1
Y J 2.5 mM CuSO4 + 0.1 mM desferal 29.6 f 3.4
OO 5 Io rb jo
[Ascorbate] mM
is 39
4.0mM CuSO4 + 0.1 mM desferal 29.3 f 0.6
0.05 mM FeS04 21.6 f 0.5
Figure 1. Dependence of benzene production on ascorbic acid 0.25 mM FeS04 13.2 f 1.9
concentration. Reaction conditions were as in Table I, with 0.25 1.0 mM FeS04 11.9 k 0.6
mM CuS04 and varying ascorbic acid concentrations. 0.25 mM FeS04 + 0.25 mM EDTA 17.6 f 0.3
Each reaction mixture contained 50 mM sodium phosphate
P
E39
"1 buffer,pH 3.0,6.25 mMsodium benzoate,8.0 mM ascorbic acid, 10.5
mM hydrogen peroxide,and the additionsindicated in atotal volume
of 2.0 mL. Reaction mixtures were incubated at 25 "C in sealed vials
and aliquota removed at 15 min for analysis of benzene. * ND, none
detected.

[Hydrogen Peroxldr] (mM)

Figure 2. Dependence of benzene production on hydrogen

peroxide concentration. Reaction conditions were as in Table I,
with 0.25 mM CuS04 and varying H202 concentrations.
0 2 4 2, 5 , J
in triplicate. Buffers for the pH dependence study contained 50 ,
3
,
35
,
1
,
45
,
5 55 5

mMphosphoric acid and 50 mM acetic acid (combined), and pH PH

was adjusted with NaOH. NaC104 was added to buffers to give Figure 3. Dependence of benzene production on pH of the
afinalionicstrengthofO.lOM,exceptatpH7.0the ionicstrength solution. Buffer solutions were prepared as described in the text.
was 0.14 M without added NaC104. Reaction conditions were as in Table I, with 0.25 mM CuSO4 and
GC/MS Analysis of Benzene in Headspace. The reaction varying buffer solutions.
mixture was heated to 45 "C, and 500 rL of headspace gas from
the reaction vial was analyzed on a Hewlett-Packard (5890/5971, duced in the absence of H202 when the reaction mixture
Series 11) gas chromatograph/mass spectrometer. A 40 nM was heated to 50 "C for 3 h (25.0 i 1.4 nM benzene
benzene standard in water was used as an external standard for produced, with 8 mM ascorbic acid, 6.2 mM sodium
comparison of retention time and mass spectrum of benzene in benzoate, and 0.25 mM CuSO4 in 50 mM phosphate buffer,
the reaction mixture. pH 3.0). Hydroxyl radical could be formed under these
RESULTS AND DISCUSSION conditions by the sequential reduction of atmospheric O2
present in solutions, as shown in eqs 1-4.
Quantitation of Benzene. The chromatographic Dependence of the Reaction on Transition-Metal
system gave a linear response for benzene standards over Catalyst. Benzene production was optimum at 1.0 mM
a concentration range of 0-100 nM (r2 > 0.981, with a CuSO4 concentration (Table I); higher concentrations of
retention time of approximately 8 min. The lower limit this metal ion resulted in a small decrease in benzene
for detection was approximately 1nM. A time-dependent production. When CuSO4 was omitted from the reaction
study of a typical reaction mixture showed an increase in mixture, there was a significant amount of benzene
benzene production up to ca. 10 min, with no further produced (6.7 f 1.9 nM), which could be completely
change in benzene concentration up to 40 min. Conse- suppressed by addition of 0.1 m M desferioxamine,an iron
quently, all subsequent reaction mixtures were analyzed chelator that suppresses iron-catalyzed production of
for benzene at 15 min after mixing (unless indicated hydroxyl radical by trace amounts of iron in distilled water
otherwise). and buffer salts (Buettner, 1986).
Dependence of the Reaction on Ascorbic Acid and Addition of 0.05 mM FeS04 to the reaction mixture, in
Hydrogen Peroxide. Benzene production increased with the absence of added CuSO4, gave optimum benzene
increasing ascorbic acid concentration at low ascorbic acid production; higher concentrations of FeS04 resulted in
concentrations but decreased with increasing amounts of suppressionof benzene production. These results suggest
ascorbic acid when its concentration exceeded that of these metal ions are necessary to catalyze hydroxyl radical
benzoic acid (Figure 1). This dependence on ascorbic acid production but at higher concentrationsmay interfere with
concentration indicates hydroxyl radical production is benzene production, possibly by catalyzing reactions in
dependent on the concentration of ascorbic acid, but the the free-radicalchain mechanism that do not yield benzene.
ascorbic acid competes with benzoic acid as a scavenger When an equimolar amount of EDTA was added with
of the hydroxyl radical at higher concentrations. 0.25 mM FeS04, there was a small increase in benzene
Benzene production increased linearly with H202 con- production relative to addition of iron without EDTA.
centration until the concentration of H202 exceeded the pH Demndence of the Reaction. The benzene yield
ascorbic acid concentration, reflecting a change in the was measured for this reaction over the pH range 2-7. The
limitingreagent (Figure 2). Under the standard conditions maximum amount of benzene was produced at pH 2, with
(25 OC and 15-min reaction time), there was no benzene a sharp drop in benzene production from pH 3 to pH 5
detected when either ascorbic acid or H202 was omitted (Figure 31, and benzene was not detectable in reaction
from the reaction mixture. However, benzene was pro- mixtures at pH 7. These results indicate that hydroxyl
Short Communications J. Agric. Food Chem., VOl. 41, No. 5, lQQ3 685

radical attack on benzoic acid may yield benzene, but Ellis, J. D.; Sykes, A. G. Kinetic Studies on the Vanadium(I1)-
attack on the benzoate anion apparently yields other Titanium(1V) and Titanium(II1)-Vanadium(1V) Redox Re-
products, such as hydroxybenzoates, phenol, or biphenyl actions in Aqueous Solutions. J . Chem. SOC.,Dalton Tram.
(Loeble et al., 1951; Armstrong et al., 1960; Sugimore et 1973,537-543.
al., 1960; Sakumoto et al., 1961). Halliwell, B.; Gutteridge, J. M. C. The Importance of Free
Positive Identification of Benzene in Reaction Radicals and Catalytic Metal Ions in Human Disease. Mol.
Mixtures by W/MS Analysis. Complete reaction Aspects Med. 1985,8, 89-193.
mixtures were placed in sealed vials and warmed to 45 "C, Loebl, H.; Stein, G.; Weiss, J. Chemical Reactions of Ionizing
and the headspace gas was analyzed by GC/MS. There Radiations in Solution: VIII. Hydroxylation of Benzoic Acid
was a major peak corresponding to benzene, in both GC by Free Radicals Produced by X-rays. J . Chem. SOC.1951,
retention time and mass spectrum, in the headspace gas 405-407.
of the reaction vial. This peak was not observed in reaction Maltoni, C.; Scarnato, C. First Experimental Demonstration of
blanks that were lacking ascorbic acid, nor was it observed the Carcinogenic Effect of Benzene. Med. Lavoro 1979, 70,
in headspace gas of vials containing ascorbic acid but 352-357.
lacking sodium benzoate in the reaction buffer. Mathew, R. W.; Sangster, D. F. Measurement of Benzoate
To our knowledge, this is the first demonstration of Radiolytic Decarboxylation: Relative Rate Constant for
benzene production from benzoic acid in the presence of Hydroxyl Radical Reaction. J . Phys. Chem. 1965,69,193&
a hydroxyl radical generating system. The conditions for 1946.
this reaction suggest that benzene could be produced in Sagone, A. L., Jr.; Decker, M. A.; Wells, R. M.; Demoko, C. A
food products, especially acidic beverages, containing the New Method for the Detection of Hydroxyl Radical Production
combination of ascorbic acid and sodium benzoate. The by Phagocytic Cells. Biochim. Biophys. Acta 1980,628,90-
97.
metal ion catalysts are most likely present in water used
in the preparation of these foods. Although the yield of Sakumoto, A,; Tsuchihashi, G. Radiation-induced Reaction in
benzene is extremely small (<50 nM or <1 ppb) under an Aqueous Benzoic Acid Solution. 11. Determination of
Products by Isotope Dilution Method. Bull. Chem. SOC.Jpn.
conditions selected to approximate relative amounts of 1961,34,663-667.
these compounds in foods or beverages, the combination
of ascorbic acid and sodium benzoate in foods and Sugimore, A.; Tsuchihashi, G. Effect of Metal Ions on the
Radiation-induced Decarboxylation of Aqueous Benzoic and
beverages should be evaluated more carefully, in view of SalicylicAcid Solutions. Bull. Chem. SOC.Jpn. 1960,33,713-
these results. 714.
ACKNOWLEDGMENT Winston, G. W.; Harvey, W.; Berl, L.; Cederbaum, A. I. The
Generation of Hydroxyl and Alkoxy1 Radicals from the
G.D.L. thanks the Long Island University Faculty Interaction of Ferrous Bipyridyl with Peroxide. Biochem. J .
Federation for release time from teaching duties to pursue 1983,216,415-421.
this research. We are grateful to Hewlett-Packard Co. for
the generous donation of the GUMS instrument to the Lalita K. Gardner and Glen D. Lawrence.
Chemistry Department. Chemistry Department, Long Island University,
LITERATURE CITED Brooklyn, New York 11201
Armstrong, W. A.; Black, B. A.; Grant, D. W. The Radiolysis of
Aqueous Calcium Benzoate and Benzoic Acid Solutions. J. * Author to whom correspondence should be addressed.
Phys. Chem. 1960,64, 1415-1419. Temporary address (through July 1993): Department of
Buettner, G. R. Ascorbate Autoxidation in the Presence of Iron Veterinary Physiology, Sokoine University of Agriculture,
and Copper Chelates. Free Radical Res. Commun. 1986,1, P.O.Box 3017, Morogoro, Tanzania.
349-353.
Chichester,D. F.; Tanner, F. W., Jr. Antimicrobial Food Additives.
InHandbook ofFood Additiues;Furia, T. E., Ed.; The Chemical Received for review December 28,1992. Accepted January 28,
Rubber Co.: Cleveland, OH, 1968. 1993.