On-line monitoring of the photolysis of benzyl acetate and 3,5-dimethoxybenzyl acetate by membrane introduction mass spectrometry

Article abstract of DOI:10.1021/jo9608376

Membrane introduction mass spectrometry (MIMS) allows on-line monitoring of the products of photolysis (254 nm) of benzyl acetate in aqueous methanol and 3,5-dimethoxybenzyl acetate in water. The reaction mixture is continuously exposed to a silicone membrane through which analyte molecules permeate into a triple quadrupole mass spectrometer for qualitative and quantitative analysis. Ionization is achieved by either isobutane or ammonia chemical ionization, and ions characteristic of the reactant ester and its products are monitored simultaneously and continuously. Three products, benzyl methyl ether, ethylbenzene, and bibenzyl are observed in the benzyl acetate photolysis. Two products, 3,5-dimethoxybenzyl alcohol and 3,5-dimethoxyethylbenzene, are formed in the photolysis of 3,5-dimethoxybenzyl acetate. Quantitation is achieved through calibration using external standard solutions and, in the case of benzyl methyl ether, tandem mass spectrometry is used to verify product identification. During the photolysis of benzyl acetate, benzyl methyl ether and ethylbenzene are produced at onset with similar efficiencies. For the 3,5-dimethoxy ester photolysis, performed in aqueous solution, the efficiency of formation of the polar product 3,5-dimethoxybenzyl alcohol is about 300 times greater than that of the nonpolar product 3,5-dimethoxyethylbenzene. The results show that the relative reaction rates are dependent on the solvent and on the photon intensity and are consistent with earlier off-line experiments by Pincock et al. which showed that the photolysis proceeds through both ion and radical pair intermediates. To the best of our knowledge, the work reported here describes the first analysis of the photochemistry of an aralkyl ester in water and the first use of on-line mass spectrometry in a mechanistic study.

Full text of DOI:10.1021/jo9608376

J . Org. Chem. 1996, 61, 6627-6632
6627
On -Lin e Mon itor in g of th e P h otolysis of Ben zyl Aceta te a n d
,5-Dim eth oxyben zyl Aceta te by Mem br a n e In tr od u ction Ma ss
Sp ectr om etr y
3
Philip S. H. Wong, N. Srinivasan, N. Kasthurikrishnan, and R. G. Cooks*
Department of Chemistry, Purdue University, West Lafayette, Indiana 47907-1393
J . A. Pincock* and J . S. Grossert*
Department of Chemistry, Dalhousie University, Halifax, Nova Scotia, Canada, B3H 4J 3
Received May 7, 1996X
Membrane introduction mass spectrometry (MIMS) allows on-line monitoring of the products of
photolysis (254 nm) of benzyl acetate in aqueous methanol and 3,5-dimethoxybenzyl acetate in
water. The reaction mixture is continuously exposed to a silicone membrane through which analyte
molecules permeate into a triple quadrupole mass spectrometer for qualitative and quantitative
analysis. Ionization is achieved by either isobutane or ammonia chemical ionization, and ions
characteristic of the reactant ester and its products are monitored simultaneously and continuously.
Three products, benzyl methyl ether, ethylbenzene, and bibenzyl are observed in the benzyl acetate
photolysis. Two products, 3,5-dimethoxybenzyl alcohol and 3,5-dimethoxyethylbenzene, are formed
in the photolysis of 3,5-dimethoxybenzyl acetate. Quantitation is achieved through calibration using
external standard solutions and, in the case of benzyl methyl ether, tandem mass spectrometry is
used to verify product identification. During the photolysis of benzyl acetate, benzyl methyl ether
and ethylbenzene are produced at onset with similar efficiencies. For the 3,5-dimethoxy ester
photolysis, performed in aqueous solution, the efficiency of formation of the polar product 3,5-
dimethoxybenzyl alcohol is about 300 times greater than that of the nonpolar product 3,5-
dimethoxyethylbenzene. The results show that the relative reaction rates are dependent on the
solvent and on the photon intensity and are consistent with earlier off-line experiments by Pincock
et al. which showed that the photolysis proceeds through both ion and radical pair intermediates.
To the best of our knowledge, the work reported here describes the first analysis of the
photochemistry of an aralkyl ester in water and the first use of on-line mass spectrometry in a
mechanistic study.
In tr od u ction
is that the method of detection, mass spectrometry, is a
universal method, and hence a suit of compounds can be
examined simultaneously over the entire reaction period.
Membrane introduction mass spectrometry is a rapid,
rugged, analytical method which has been applied to the
on-line monitoring of volatile organic compounds (VOC's)
Little is known about the photochemistry of organic
compounds in water because their low solubility makes
characterization of products and determination of their
yields difficult. Also, the usual method used to follow
the kinetics of such reactions involves collection of
aliquots at chosen time intervals during the reaction and
their off-line analysis by the use of gas chromatography
4
5
in water and air, to in vivo analysis of VOC’s in blood,
6
to characterization of fermentation products, and to
7
monitoring such processes as wastewater treatment. A
1
or other spectroscopic techniques. These methods,
semipermeable membrane, usually a silicone polymer,
acts as an interface between the sample solution and the
vacuum of the mass spectrometer. The hydrophobic
membrane is permeable to relatively nonpolar and low
molecular weight compounds while the aqueous matrix
however, are time-consuming and do not allow real time
on-line monitoring of the photolysis process. Because
there appear to be no studies in the literature in which
on-line monitoring is used in order to perform mecha-
nistic investigations by mass spectrometry (although
reports on the monitoring of intermediates in bioreactors
is excluded. Analyte molecules permeate the membrane
by the process of pervaporation8 and are then ionized.
2
have appeared ), we were intrigued by the possibility
that MIMS (membrane introduction mass spectrometry)
The high resistance to fouling, and the mechanical and
chemical resistance of silicone membranes, adds to their
usefulness.
Most organic photochemical reactions involve only
radical intermediates, and, to a first approximation, the
polarity of the solvent is not expected to be a critical
3
could be used for continuously monitoring such reactions.
Recent advances in MIMS, especially the use of the direct
insertion membrane probe, used in conjunction with flow
injection analysis procedures of sampling, provide the
short response times which are desirable for on-line
reaction monitoring. The importance of the experiment
(4) (a) Kotiaho, T.; Lauritsen, F. R.; Choudhury, T. K.; Cooks, R.
G.; Tsao, G. T. Anal. Chem. 1991, 63, 875A. (b) Cooks, R. G.; Kotiaho,
T. ACS Symp. Ser. 1992, 508, 126.
*
To whom correspondence should be addressed.
Abstract published in Advance ACS Abstracts, September 1, 1996.
(5) Brodbelt, J . S.; Cooks, R. G.; Tou, C. J .; Kallos, G. J .; Dryzga,
M. D. Anal. Chem. 1987, 59, 454.
X
(
1) Hilborn, J . W.; MacKnight, E.; Pincock, J . A.; Wedge, P. J . J .
Am. Chem. Soc. 1994, 116, 3337.
2) (a) Lauritsen, F. R.; Gylling, S. Anal. Chem. 1995, 67, 1418; (b)
Heinzle, E. J . Biotechnol. 1992, 25, 81.
3) (a) Bauer, S. J .; Cooks, R. G. Am. Lab. 1993, October, 36. (b)
Bauer, S. Trends Anal. Chem. 1995, 14, 202.
(6) (a) Lauritsen, F. R.; Kotiaho, T.; Lloyd, D. Biol. Mass Spectrom.
1993, 22, 585. (b) Hayward, M. J .; Riederer, D. E.; Kotiaho, T.; Cooks,
R. G.; Austin, G. D.; Syu, M.-J ; Tsao, G. T. Proc. Contr. Qual. 1991, 1,
105.
(7) LaPack, M. A.; Tou, J . C.; Enke, C. G. Anal. Chem. 1991, 63,
1631.
(
(
S0022-3263(96)00837-7 CCC: $12.00 © 1996 American Chemical Society

6
628 J . Org. Chem., Vol. 61, No. 19, 1996
Wong et al.
Sch em e 1. Gen er a l Mech a n ism for th e P h otolysis of Ben zyl Aceta te in Meth a n ol-Wa ter
factor in the reaction. However, some photochemical
reactions do involve ionic intermediates and, most inter-
estingly, some involve reaction pathways where both
radical and ionic species are important. The unique
properties (dielectric constant, viscosity) of water as a
solvent could therefore have significant mechanistic
effects. Recently, extensive results on the photochemis-
quartz tube was irradiated in a Rayonet photoreactor at a
wavelength of 254 nm. The solution was recirculated via
Teflon transfer lines to a direct-insertion membrane (silicone)
probe12 by a peristaltic pump at 1 mL/min. The probe was
inserted directly into the ion source of a triple quadrupole mass
spectrometer and was heated to 80 °C. Analytes permeating
the membrane were chemically ionized by either isobutane or
ammonia at an estimated pressure of 0.5 torr. External
calibration was achieved by flow injection of 500 mL aliquots
of either sample or standard solution through the membrane
probe. Data were acquired by scanning from m/ z 70 to 330
in 2 s. Benzyl acetate, benzyl alcohol, ethylbenzene, and
bibenzyl were obtained commercially while benzyl methyl
ether, 3,5-dimethoxybenzyl acetate, 3,5-dimethoxybenzyl al-
cohol, and 3,5-dimethoxyethylbenzene were synthesized by
published methods.1
try of arylmethyl esters in methanol have been re-
ported.1
,9,10
The general conclusion from these studies,
as outlined in Scheme 1, is that the excited state of the
ester reacts to give a radical pair by homolytic cleavage.
Direct heterolytic cleavage to an ion pair is not a major
pathway. Instead, the ion pair is formed by an electron
transfer process in the radical pair. The rate of this
,13
1
1
process, according to Marcus’s theory of electron trans-
fer, is dependent on the reorganization energy. The
solvent component of this energy will undergo large
changes as the dielectric constant of the solvent changes,
hence the expectation that the distribution of reaction
products should be dependent on the choice of solvent.
The objective of this work is, therefore, to explore the
application of the MIMS method for the on-line monitor-
ing of the photolysis of benzyl acetate in methanol-water
and 3,5-dimethoxybenzyl acetate in water. For the
unsubstituted ester, the solubility in water alone was low
and 1:1 (v/v) methanol-water was used. The expected
reaction intermediates (Scheme 1) are a radical pair and
an ion pair. For the radical pair, decarboxylation and
in-cage coupling gives ethylbenzene and carbon dioxide.
Out-of-cage dimerization of the benzyl radical gives
bibenzyl. For the ionic intermediate, trapping by metha-
nol when it is the solvent gives benzyl methyl ether, but
when water is the solvent, benzyl alcohol is produced.
Resu lts a n d Discu ssion
P h otolysis of Ben zyl Aceta te in Meth a n ol-Wa ter .
Mass spectra due to individual compounds must be
known in order to interpret data on reaction mixtures.
Table 1 summarizes the abundances of the major ions
in the mass spectra recorded by MIMS using isobutane
CI for aqueous methanol (1:1) solutions of benzyl acetate,
benzyl methyl ether, benzyl alcohol, ethylbenzene, and
bibenzyl. The concentrations were 100 mg/L each, except
that for benzyl alcohol a concentration of 1000 mg/L was
used. All the mass spectra show a common fragment ion
at m/ z 91, due to the benzyl group, and also the
+
protonated molecule [M + H] , except that benzyl alcohol
+
shows the [M-H] ion. On the basis of this information,
+
the [M + H] ions were chosen to monitor the compounds,
except that the concentrations of benzyl alcohol and
ethylbenzene could not be measured directly with this
reagent gas, because both give the same characteristic
ion, m/ z 107. Differentiation between these two com-
pounds was achieved by changing the CI reagent gas to
ammonia (all subsequent experiments used this reagent,
see below). Although tandem mass spectrometry (MS/
MS) could have been used to differentiate these two
compounds, the choice of another CI reagent gas is
Exp er im en ta l Section
A solution of 400 mg/L benzyl acetate in 1:1 methanol-
water or 3,5-dimethoxybenzyl acetate in water in an 8 mL
(8) (a) N e´ el, J . In Pervaporation Membrane Separation Processes;
Huang, R. Y. M., Ed.; Elsevier: New York, 1991. (b) Hwang, S. T.;
Kammermeyer, K. Membranes in Separation (Techniques of Chemistry,
Vol VII); Wiley: New York, 1975; Chapter VII.
(
(
(
9) DeCosta, D. P.; Pincock, J . A. J . Am. Chem. Soc. 1989, 111, 8948.
10) Hilborn, J . W.; Pincock, J . A. J . Am. Chem. Soc. 1991, 113, 2683.
11) Eberson, L. Electron Transfer Reactions in Organic Chemistry;
(12) (a) Bier, M. E.; Cooks, R. G. Anal. Chem. 1987, 59, 597. (b) Bier,
M. E.; Kotiaho, T.; Cooks, R. G. Anal. Chim. Acta 1990, 231, 175.
(13) Pincock, J . A.; Wedge, P. J . J . Org. Chem. 1994, 59, 5587.
Springer-Verlag: New York, 1987; p 32.

On-Line Monitoring of Photolysis by MIMS
J . Org. Chem., Vol. 61, No. 19, 1996 6629
Ta ble 1. Rela tive P ea k In ten sities in Isobu ta n e a n d Am m on ia CI Ma ss Sp ectr a of Meth a n ol-Wa ter (1:1) Solu tion s of
Ben zyl Aceta te, Ben zyl Meth yl Eth er , Ben zyl Alcoh ol, Eth ylben zen e, a n d Biben zyl. Ch a r a cter istic Ion s Used for
Mon itor in g Ar e Ma r k ed w ith a n Aster isk
mass spectral peaks with relative intensity in parenthesis
sample (MW)
isobutane CI
ammonia CI
benzyl acetate (150)
82 (5%), 85 (11%), 91 (100%), 93 (7%), 151* (40%)
90 (1%), 91 (2%), 108 (13%), 128 (1%), 168* (100%), 169 (8%)
91 (7%), 108 (14%), 121 (8%), 122 (13%), 140* (100%), 141 (8%)
benzyl methyl ether (122) 81 (3%), 85 (1%), 91 (100%), 92 (4%), 93 (2%),
23* (39%)
91 (100%), 92 (8%), 105 (3%), 107* (18%), 108 (7%),
09 (1%)
1
benzyl alcohol (108)
ethylbenzene (106)
bibenzyl (182)
91 (22%), 94 (8%), 108 (75%), 109 (6%), 126* (100%), 127 (8%)
1
65 (100%), 75 (2%), 89 (1%), 105 (2%), 106 (10%), 107* 82 (3%), 91 (100%), 105 (5%), 106* (20%), 107 (1%)
75%), 108 (2%)
91 (6%), 99 (6%), 105 (90%), 112 (9%), 181 (21%),
82 (3%), 183* (100%), 184 (4%)
(
87 (24%), 91 (38%), 108 (66%), 128 (5%), 200* (100%), 201 (2%)
1
preferred because it allows data to be taken at a greater
rate than would the MS/MS experiment.
Photolysis, using a low UV light intensity (two lamps
in a Rayonet photoreactor), of a water-methanol (1:1)
solution of benzyl acetate with continuous monitoring of
the reaction mixture gave the results that the ion
+
abundance at m/ z 151 (benzyl acetate, [M + H] )
decreases with time while the abundance of m/ z 123
+
(
benzyl methyl ether, [M + H] ) shows a corresponding
+
rise with time. The ion at m/ z 183 (bibenzyl, [M + H] )
has a very low abundance indicating that this is a very
minor product. Possible isomers of benzyl methyl ether
are 2-phenylethanol and phenyl ethyl ether. The proto-
nated 2-phenylethanol will undergo H
ion at m/ z 105 while the protonated phenyl ethyl ether
will lose C to give an ion at m/ z 95 upon collision-
induced dissociation.
-phenylethanol gives a characteristic ion at m/ z 122
2
O loss to give an
2
H
4
1
4
F igu r e 1. The ion abundance of benzyl acetate (m/ z 168, [M
In addition, isobutane CI of
+
+
+
NH
+
4
m/ z 200, [M + NH ]), and ethylbenzene (m/ z 106, M ) are
4
]), benzyl methyl ether (m/ z 140, [M + NH ]), bibenzyl
4
2
+•
(
+
•
(M ), hence the possibility of m/ z 123 being 2-phenyl-
plotted as a function of time under ammonia CI condition. Note
the ion abundance at m/ z 106 is expanded 5 times. The
photolysis started at scan number 50. Ion abundances have
not been corrected for losses through the membrane; this
procedure changes the ion counts but has an insignificant
effect on the shapes of the curves.
ethanol is excluded. The ion at m/ z 123 was allowed to
undergo collision-induced dissociation (CID) with argon
(17 eV). The product ion spectrum of m/ z 123 from the
photolyzed solution displays two peaks, m/ z 91 (100%)
and m/ z 123 (63%), consistent with the results for a
standard solution of benzyl methyl ether. Thus, the ion
at m/ z 123 was unambiguously identified as the proto-
nated benzyl methyl ether.
Quantitation was achieved by calibration using exter-
nal standard solutions based on peak area of the ion
chromatograms calculated using the data system of the
instrument. The concentrations of benzyl acetate and
benzyl methyl ether were found to be 35 and 28 mg/L,
respectively, in this photolysis after 50 min. This cor-
responds to a yield of 9.4% of benzyl methyl ether at a
under ammonia CI conditions and hence they can now
be differentiated. Under ammonia CI, benzyl alcohol
+
gives characteristic ions of m/ z 109 ([M + H] ) and 126
+
([M + NH
4
]) while ethylbenzene has a characteristic ion
+
•
of m/ z 106 (M ). Detection limits (S/N ) 3) of benzyl
acetate, benzyl methyl ether, bibenzyl, ethylbenzene and
benzyl alcohol in 1:1 methanol-water solution were
determined as 5.0, 2.0, 2.0, 10, and 50 mg/L, respectively.
Figure 1 shows the abundances of ions characteristic of
benzyl acetate (m/ z 168, [M + NH4 ]), benzyl methyl
ether (m/ z 140, [M + NH4 ]), bibenzyl (m/ z 200, [M +
+
9
1% conversion of benzyl acetate. (Note that these values
+
do not consider losses at the membrane and so the
calculated yield represents a lower limit, see below.) The
isomer, 2-phenylethanol, which yields a characteristic ion
+
+•
NH
of time. Note that the photolysis started at scan number
0 and that it takes approximately 1 min more for the
4
]), and ethylbenzene (m/ z 106, M ) as a function
+
•
at m/ z 122 (M ) under isobutane CI conditions, is a
5
1
significant product (26%) in methanol photolysis, but is
solution to flow from the photochemical reactor to the
mass spectrometer.
not observed in methanol-water. Subsequent experi-
ments were similar, but two changes were made. First,
ammonia was used exclusively as the CI reagent gas.
Second, a high UV light intensity (13 lamps) was used
in an attempt to increase the yield of the product
bibenzyl. Table 1 summarizes the relative ion abun-
dances in the ammonia CI mass spectra of standard
aqueous methanol (1:1) solutions of benzyl acetate, benzyl
methyl ether, benzyl alcohol, ethylbenzene, and bibenzyl
when examined by MIMS. Note that benzyl alcohol and
benzyl methyl ether have different characteristic ions
There is no signal due to the ions which are charac-
teristic of benzyl alcohol, m/ z 109 and 126, throughout
the whole photolysis period, indicating that benzyl alcohol
is not formed in significant quantities under these
conditions. The ion abundance of benzyl methyl ether
started to increase at scan number 90 as did that of
ethylbenzene and bibenzyl although the low sensitivity
to ethylbenzene and the low rate of formation of bibenzyl,
a bimolecular reaction product, make their onsets more
difficult to locate. The low sensitivity for ethylbenzene
is due to the difficulty of protonation under ammonia CI
conditions since its proton affinity (191.6 kcal/mol) is
(
14) Sigsby, M. L.; Day, R. J .; Cooks, R. G. Org. Mass Spectrom.
1
979, 14, 273.

6
630 J . Org. Chem., Vol. 61, No. 19, 1996
Wong et al.
a
are due to contributions from several compounds (Table
1
). The concentrations of benzyl acetate, benzyl methyl
ether, bibenzyl, and ethylbenzene were found to be 32,
4, 7.6, and 32 mg/L, respectively, after about 20 min of
4
photolysis. These measurements were made by calibra-
tion using external standard solutions; they correspond
to 92% conversion of benzyl acetate and the correspond-
ing yields of benzyl methyl ether, ethylbenzene, and
bibenzyl are 15%, 12%, and 3.4%, respectively. Note that
the ethylbenzene data are the least precise because of
the low abundance of the characteristic ion, m/ z 106. It
is expected that benzyl alcohol should be formed in a yield
comparable to that of benzyl methyl ether (44 mg/L).
However, the detection limit for benzyl alcohol is only
5
0 mg/L, so it is not observed. A separate off-line
photolysis experiment was performed under static condi-
tions with product analysis being made only at the end
of the reaction period and hence no reagent loss being
possible. Note that under these more sensitive condi-
b
tions, a small signal was observed due to the ion [PhCH
2
-
+
OH + NH
4
], characteristic of benzyl alcohol, indicating
that benzyl alcohol is indeed formed during the photoly-
sis. Previous off-line experiments in methanol solvent
showed yields of 26%, 19%, and 23%, for benzyl methyl
ether, ethylbenzene, and bibenzyl, respectively, together
with other products in a total 94% yield.1 The difference
in relative yields between the on-line aqueous/methanol
and the off-line methanol experiments is simply ascribed
to the increased polarity of the solvent which significantly
reduces the formation of the nonpolar product bibenzyl.
F igu r e 2. (a) Calibration scans for benzyl methyl ether under
ammonia CI conditions. (b) Least squares fit of the calibration.
The poor mass balance in the MIMS experiments is
primarily due to analyte loss through the membrane as
the solution is recirculated while the limited membrane
permeability to the more polar compounds such as benzyl
alcohol and 2-phenylethanol, both of which have similar
detection thresholds, contributes to the errors in the mass
balance. The loss of benzyl acetate through the mem-
brane was investigated in a separate experiment in which
a 400 mg/L benzyl acetate solution was aspirated through
the membrane probe over the same period of time used
for the photolysis reactions but with the UV light off. It
was found that the signal due to benzyl acetate decreases
steadily and eventually reached 60% of its initial value.
Similar experiments were also performed to investigate
analyte loss through the membrane for benzyl methyl
ether, bibenzyl, and ethylbenzene. It was found that they
show similar percentage losses, in the range of 35-45%,
through the membrane. The fact that a linear relation-
ship exists between the concentrations of the unimolecu-
lar reaction products (see Figures 4 and 7) is a further
indication that the analyte losses do not differ signifi-
cantly. In order to account quantitatively for the mass
imbalance, we performed the following estimates. If we
assume that the percentage loss is independent of
concentration, the corrected yields of benzyl methyl ether,
ethylbenzene, and bibenzyl are 25%, 20%, and 5.7%
respectively, based on 40% analyte loss through the
membrane. If benzyl alcohol is formed in a yield com-
parable to that of benzyl methy ether, then a total
corrected yield of 76% is obtained. The remaining 24%
may be partly due to 2-phenylethanol, which is formed
in 26% yield in methanol photolysis,1 which is not
detected by MIMS due to the limited membrane perme-
ability to this compound. The errors in analyte loss
through the membrane can presumably be reduced by
doing larger scale photolysis such that the losses are
F igu r e 3. An ammonia CI mass spectrum of the benzyl
acetate solution after photolysis for 20 min.
lower than that of ammonia (204.0 kcal/mol).15 Figure
2
1
a shows calibration results for benzyl methyl ether, m/ z
40, and Figure 2b shows the least squares fit to a set of
calibration data. The precision of the experiment is
represented in these results.
Figure 3 shows the ammonia CI mass spectrum of the
solution after photolysis for about 20 min. Characteristic
peaks show the presence of ethylbenzene (m/ z 106,
+
•
[
PhCH
2
CH
3 2
] ), benzyl methyl ether (m/ z 140, [PhCH -
+
OCH
CH
NH
3
+ NH
4
]), benzyl acetate (m/ z 168, [PhCH
2
O(CO)-
+
3
+ NH
4
]) and bibenzyl (m/ z 200, [PhCH
2
CH Ph +
2
+
4
]). Other prominent ions, such as m/ z 91 and 108,
(15) Lias, S. G.; Liebman, J . F.; Levin, R. D. J . Phys. Chem. Ref.
Data 1984, 13, 695.

On-Line Monitoring of Photolysis by MIMS
J . Org. Chem., Vol. 61, No. 19, 1996 6631
F igu r e 5. The ion abundance of 3,5-dimethoxybenzyl acetate
+
(
[
m/ z 211, [M + H ]), 3,5-dimethoxybenzyl alcohol (m/ z 169,
F igu r e 4. A plot of concentration of ethylbenzene versus the
concentration of benzyl methyl ether. The slope, after correct-
ing the concentration units to mol/L, gives the relative rate of
formation of ethylbenzene with respect to benzyl methyl ether.
+
M + H ]), and 3,5-dimethoxyethylbenzene (m/ z 167, [M +
H ]), plotted as a function of time under ammonia CI condition.
The photolysis started at scan number 30.
+
small compared to the volume of the reaction vessel. If
corrections to the observed ion abundances shown in
Figure 1 are made to account for losses of analyte through
the membrane, only insignificant changes in the data are
found; these results are therefore not reproduced.
Information on relative rates of product formation can
be obtained by comparing initial rates of formation
derived from the initial slopes in Figure 1, each of which
is linear for the first few minutes. Note that the errors
due to analyte loss are small at the beginning of the
photolysis. The initial rates of formation of benzyl
methyl ether, ethylbenzene, and bibenzyl were found to
be in the ratio of 1:0.93:0.15. The threshold rate of
formation of bibenzyl is about 1 order of magnitude
slower than the rate of formation of ethylbenzene and
benzyl methyl ether. The relative rate, after correcting
concentration units to mol/L, can also conveniently be
obtained by plotting the concentration of ethylbenzene
against benzyl methyl ether, as shown in Figure 4. The
ratio of the concentrations of the two products remains
essentially constant (0.90) over the full photolysis period.
This result is as expected for two products being formed
from two competing pathways and indicates that they
are both primary photoproducts.
F igu r e 6. An ammonia CI mass spectrum of the 3,5-
dimethoxybenzyl acetate solution after photolysis for about 12
min. The symbol Ar represents the 3,5-dimethoxyphenyl group.
the difference in volatility between the two compounds
is not large. Its negligible formation is expected because
this compound can only be formed by dimerization of
benzyl radicals which have escaped from the solvent cage,
which the high viscosity of water appears to prevent.
Figure 6 shows the ammonia CI mass spectrum of the
solution after photolysis for about 12 min. Characteristic
peaks show the presence of 3,5-dimethoxyethylbenzene
P h ot olysis of 3,5-Dim et h oxyb en zyl Acet a t e in
Wa ter . When the solvent is changed to pure water, the
methyl ether is no longer formed because the nucleophile
(m/ z 167 and 184), 3,5-dimethoxybenzyl alcohol (m/ z
3
CH OH is not present. Under ammonia CI conditions,
1
69) and 3,5-dimethoxybenzyl acetate (m/ z 211 and 228).
the substituted acetate, the alcohol, ethylbenzene, and
+
bibenzyl all give characteristic [M + H ] ions which occur
Their concentrations were found to be 0.5, 174, and 6.6
mg/L, respectively, by calibration using external standard
solutions. This corresponds to 98% conversion of the
acetate and the corresponding yields of the alcohol and
3,5-dimethoxyethylbenzene are 55% and 0.2%. The loss
of the alcohol through the membrane was also measured
to be about 40%. If the analyte loss is also considered,
as discussed above, the corrected yield of the alcohol is
92%. Since the alcohol is more polar than 3,5-dimethoxy-
ethylbenzene, its relatively high yield is expected in the
polar solvent. Information on relative rates of product
formation can be obtained by comparing initial rate of
formation derived from the initial slopes in Figure 5. The
ratio of initial rates of formation of the alcohol and
substituted ethylbenzene were found to be 319:1. The
at m/ z 211, 169, 167, and 303, respectively. The detec-
tion limits of 3,5-dimethoxybenzyl acetate, 3,5-dimethoxy-
benzyl alcohol, and 3,5-dimethoxyethylbenzene in water
were determined as 5.0, 50, and 0.2 mg/L respectively.
During photolysis at high UV light intensity (13 lamps),
the reactants and products were monitored continuously.
Figure 5 shows the intensities of ions characteristic of
3
,5-dimethoxybenzyl acetate, 3,5-dimethoxybenzyl alco-
hol, and 3,5-dimethoxyethylbenzene as a function of time.
Note that there is no signal due to the ion characteristic
+
of the substituted bibenzyl (m/ z 303, [M + H ]). Since
bibenzyl permeates readily through the membrane with
a detection limit of 2 mg/L, the substituted bibenzyl is
expected to permeate through the membrane, too, as

6
632 J . Org. Chem., Vol. 61, No. 19, 1996
Wong et al.
Con clu sion s
The capability of performing on-line monitoring of
photolysis reactions by mass spectrometry is shown for
the first time and is used to investigate the mechanism
of benzyl ester photolysis. Significantly, this capability
is demonstrated to extend to the study of reactions in
aqueous solution. The high selectivity of the experiment
makes it a promising technique in studying relative rates
of product formation, although improvements are needed
in precision and sensitivity. These are likely to come as
a wider range of membrane materials, more suitable than
are silicone polymers to the compounds studied here,
becomes available.
The results show that the relative photolysis rates are
dependent on the solvent, and the observed products are
consistent with earlier demonstrations by Pincock et
F igu r e 7. A plot of concentration of 3,5-dimethoxybenzyl
alcohol versus the concentration of 3,5-dimethoxyethylbenzene
1
,9,10
(total time 12 min). The slope gives the relative rate of
al.
based on off-line experiments that the photolysis
formation of 3,5-dimethoxybenzyl alcohol with respect to 3,5-
dimethoxyethylbenzene.
proceeds through both ion and radical pair intermediates.
Obvious advantages of MIMS over the methods used
previously for such studies are that MIMS is simple, it
provides real time data during photolysis, and sample
pretreatment, such as solvent extraction or other pre-
concentration procedures, is not needed.
relative rate can also conveniently be obtained by plotting
the concentration of the alcohol against the substituted
ethylbenzene, as shown in Figure 7. The ratio of the
concentrations of the two products remains essentially
constant (285:1) over the full photolysis period. The fact
that the alcohol is formed about 300 times faster than
the substituted ethylbenzene also indicates that solvent
has a significant influence on the mechanism of this
photochemical reaction.
Ack n ow led gm en t. We thank Dr. Harry Morrison
for the loan of a Rayonet photoreactor. The work at
Purdue was supported by the Office of Naval Research.
P.W. thanks Eli Lilly and Company for Fellowship
support.
J O9608376