A New Method of Determining Chlorine Kinetic Isotope Effects

Article abstract of DOI:10.1021/ac9712355

Two methods have been used to measure the chlorine leaving group kinetic isotope effect for the S_N2 reduction of benzyl chloride to toluene by sodium borohydride in DMSO at 30.000°C. The reaction was monitored by titrating the unreacted borohydride ion. One method involved determining the chlorine isotope effect using the classical IRMS method, which requires the conversion of the chloride ions into gaseous methyl chloride that is analyzed in an isotope ratio mass spectrometric analyses (Hill, J. W.; Fry, A. J. Am. Chem. Soc. 1962, 84, 2763. Taylor, J. W.; Grimsrud, E. P. Anal. Chem. 1969, 41, 805.). Two different measurements using this method yielded isotope effects of k³⁵/k³⁷ = 1.007 19 ± 0.000 19 and 1.007 64 ± 0.000 19. The second method was a new technique where the ratio of the chlorine isotopes was obtained by fast atom bombardment mass spectrometry on the silver chloride recovered from the reaction, i.e., from the first step in the classical procedure. Therefore, the new method is much simpler and avoids the time-consuming preparation, purification, and recovery of the gaseous methyl chloride. Although the experimental error is larger (k³⁵/k³⁷ = 1.008 03 ± 0.00 10 and 1.008 02 ± 0.000 65) when the new technique is used to analyze the silver chloride samples from the same set of experiments that were used to measure the isotope effect by the classical method, the chlorine isotope effect found by the two methods is identical within experimental error. This large chlorine kinetic isotope effect indicates there is considerable C_α-Cl bond rupture in the S_N2 transition state.

Full text of DOI:10.1021/ac9712355

Anal. Chem. 1998, 70, 3548-3552
A Ne w Me t h o d o f De t e rm in in g Ch lo rin e Kin e t ic
Is o t o p e Effe c t s
K. C. We s ta w a y,*,† T. Koe rne r, Y.-R. Fa ng, J . Rudzi n˜ s ki, a nd P. Pa ne th*
†
†
‡
,‡
Department of Chemistry and Biochemistry, Laurentian University, Sudbury, Ontario, Canada, P3E 2C6,
and Institute of Applied Radiation Chemistry, Technical University of Ł o´ d z´ , Wr o´ blewskiego 15, 90-924 Ł o´ d z´ , Poland
suitable for direct measurement of their isotopic ratio,4 so
determining a chlorine isotope effect is both feasible and attrac-
tive.5,6 However, determining chlorine isotope effects is impeded
by the tedious procedure required for the conversion of chlorine
atoms into the gaseous methyl chloride that is required for
Two methods have been used to measure the chlorine
leaving group kinetic isotope effect for the S 2 reduction
N
of benzyl chloride to toluene by sodium borohydride in
DMSO at 3 0 .0 0 0 °C. The reaction was monitored by
titrating the unreacted borohydride ion. One method
involved determining the chlorine isotope effect using the
classical IRMS method, which requires the conversion of
the chloride ions into gaseous methyl chloride that is
analyzed in an isotope ratio mass spectrometric analyses
determining the 35Cl/ 37Cl isotopic ratio by isotope ratio mass
spectrometry (IRMS).1
,5-7
This obstacle could be avoided if the
isotopic composition of a reactant could be determined by direct
analysis. In fact, this can be accomplished by determining the
35Cl/ 37
Cl isotopic ratio of silver chloride samples by fast atom
(
Hill, J. W.; Fry, A. J . Am . Ch em . Soc. 1 9 6 2 , 84, 2 7 6 3 .
bombardment-isotope ratio mass spectrometry (FAB-IRMS). A
comparison of the classical conversion (IRMS) and a newly
developed FAB-IRMS method for measuring chlorine isotope
effects is presented.
Taylor, J. W.; Grimsrud, E. P . An a l. Ch em . 1 9 6 9 , 41,
8
0 5 .). Two different measurements using this method
yielded isotope effects of k3 5 / k ) 1 .0 0 7 1 9 ( 0 .0 0 0 1 9
3 7
and 1 .0 0 7 6 4 ( 0 .0 0 0 1 9 . The second method was a
new technique where the ratio of the chlorine isotopes was
obtained by fast atom bombardment mass spectrometry
on the silver chloride recovered from the reaction, i.e.,
from the first step in the classical procedure. Therefore,
the new method is much simpler and avoids the time-
consuming preparation, purification, and recovery of the
The dehalogenation of benzyl chloride by borohydride ion
-
4
-
C H CH -Cl + BH f C H CH -H + Cl + BH
3
(1)
6
5
2
6
5
2
was used as a model reaction for comparing the old and new
methods of measuring a chlorine isotope effect. The results also
provide information about the amount of C
the transition state of this S
2 reaction.1,8
R
-Cl bond rupture in
gaseous methyl chloride. Although the experimental error
is larger (k3 5 / k3 7 ) 1 .0 0 8 0 3 ( 0 .0 0 1 0 and 1 .0 0 8 0 2
N
(
0 .0 0 0 6 5 ) when the new technique is used to analyze
EXPERIMENTAL SECTION
the silver chloride samples from the same set of experi-
ments that were used to measure the isotope effect by the
classical method, the chlorine isotope effect found by the
two methods is identical within experimental error. This
large chlorine kinetic isotope effect indicates there is
Reagents. The isotope effect was measured with the same
batch of a distilled in glass grade dry DMSO [67-68-5] from
Caledon Laboratories Ltd. The sodium borohydride [16940-66-
2
] (Aldrich), which was stored in a vacuum desiccator once it
had been opened, was used as purchased. The benzyl chloride
considerable C
state.
r
N
-Cl bond rupture in the S 2 transition
[
100-44-7] (Aldrich) was purified by vacuum distillation. The
9
bp was 49-50 °C at 3 mmHg; lit. bp 51-52 °C at 4 mmHg.
Kinetic Measurements. All the glassware, the solvent, a
Mettler AE 160 semimicrobalance, and the reagents were placed
in an Instruments for Research and Industry model X-37-37
glovebag which was sealed, evacuated, and partially filled with
extra-dry nitrogen (Praxair) five times. All the solutions used to
monitor the reaction were standardized the day they were used.
Sodium borohydride and benzyl chloride stock solutions were
Chlorine kinetic isotope effects are a sensitive tool for distin-
guishing between alternative mechanisms.1 Although the mag-
nitude of these isotope effects is small (large chlorine kinetic
2
isotope effects are close to 1%), they can be used to indicate
whether a bond to a chlorine atom is broken or weakened in the
transition state.1,3 Chlorine has two stable isotopes in a proportion
prepared by diluting approximately 0.16 g (0.0042 mol) of NaBH
4
†
Laurentian University.
Technical University Ł o´ d z´ .
‡
(4) Weast, R. C., Ed. CRC handbook chemistry and physics, 61st ed.; CRC Press:
Boca Raton, FL, 1980; p B-263.
(5) Hill, J. W.; Fry, A. J. Am. Chem. Soc. 1 9 6 2 , 84, 2763.
(
1) Shiner, V. J., Jr.; Wilgis, F. P. In Isotopes in organic chemistry; Buncel, E.,
Saunders: W. H., Jr., Eds.; Elsevier: Amsterdam, 1992; Vol. 8, pp 272-
2
88.
(6) Taylor, J. W.; Grimsrud, E. P. Anal. Chem. 1 9 6 9 , 41, 805.
(7) McLennan, D. J.; Stein, A. R.; Dobson, B. Can. J. Chem. 1 9 8 6 , 64, 1201.
(8) Koerner, T. MSc. Dissertation, Laurentian University, Sudbury, ON, 1996.
(9) Willi, A. V.; Ho, C.-K.; Ghanbarpour, A. J. Org. Chem. 1 9 7 2 , 37, 1185.
(
(
2) Maccoll, A. Ann. Rep. A, Chem. Soc. 1 9 7 4 , 71, 77.
3) Sims, L. B.; Fry, A.; Netherton, L. T.; Wilson, J. C.; Reppond, K. D.; Crook,
S. W. J. Am. Chem. Soc. 1 9 7 2 , 94, 1364.
3548 Analytical Chemistry, Vol. 70, No. 17, September 1, 1998
S0003-2700(97)01235-3 CCC: $15.00 © 1998 American Chemical Society
Published on Web 07/23/1998

and 0.95 g (0.0075 mol) of benzyl chloride (accurately weighed)
to 100 and 250 mL, respectively, with DMSO in volumetric flasks.
The substrate stock solution was transferred into a 250-mL round-
bottom flask fitted with a serum cap greased on the ground glass
joint, sealed with Parafilm and a wire. A 25-mL aliquot of the
and extracted with 100 mL of heptane. Then, the 100-mL heptane
layer was extracted with 25 mL of double-distilled water. All the
aqueous layers were made slightly basic with NaOH12 and reduced
to ∼100 mL on the rotary evaporator. The ionic strength of this
13
solution was increased by adding 10.0 g of KNO
3
[7757-79-1]
1
5
NaBH
4
stock solution was pipetted into three different 100-mL
3
(BDH) and the pH was reduced to below 2.0 with 1 M HNO .
Erlenmeyer flasks fitted with greased serum caps and sealed with
Parafilm. Then, the flasks were removed from the glovebag and
temperature equilibrated at 30.000 ( 0.002 °C for at least 1 h.
The reaction was started by transferring 25 mL of the benzyl
chloride stock solution with a 25-mL Hamilton gastight syringe
into the reaction vessel containing the borohydride ion stock
solution. At predetermined times, between 10 and 90% of
completion, a 1-mL aliquot of the reaction mixture was withdrawn
with a 1-mL Hamilton gastight syringe and diluted with 25.00 mL
of double-distilled water; the borohydride ion was titrated im-
mediately to a methyl red end point with 0.004 821 M HCl.10 The
second-order rate constant was determined using the integrated
form of a second-order rate expression that was first order in both
the nucleophile and the substrate.11 All the second-order kinetic
plots were linear up to 90% of completion and more than 90% of
the plots had correlation coefficients greater than 0.999.
The reaction mixture was then titrated in the dark under
photographic safety lamps to a potentiometric end point with a
standard AgNO
saturated calomel electrode connected to the solution through an
agar- KNO
salt bridge18 and the silver indicating electrode were
3
solution16 and a silver electrode.17 A reference
3
attached to a Fisher Scientific, model 50, pH meter set in the mV
mode. A Teflon stirring bar hit the silver electrode during the
titration.19 Then, the mixture was warmed and left in a dark
cupboard overnight so the colloidal AgCl precipitate would form
larger crystals.13 The AgCl precipitate was collected on a pre-
3
weighed sintered glass filter, washed with dilute HNO , dried in
the dark at 110 °C overnight, and cooled in a desiccator in a dark
cupboard. Finally, the filter and AgCl precipitate were weighed
and the weight of AgCl calculated.
The nitric acid titration, the chloride ion titration, and the
weight of silver chloride gave three independent estimates of the
f value that is used to calculate the isotope effect (Table 1).
Because all of the errors in f are e1.0%, the value of f is known
very accurately.
Determining the Chlorine (Leaving Group) Kinetic Iso-
tope Effect. The chlorine isotope effect was determined for the
N
S 2 reaction between sodium borohydride and benzyl chloride
in DMSO at 30.000 ( 0.002 °C. Six reactions were done at the
same time. One reaction was taken to 100% of completion. Three
different analyses (vide infra) indicated the complete reactions in
trials 1 and 2 had gone to 100.2 ( 0.65 and 99.73 ( 0.11% of
completion, respectively. The other five reactions were used to
obtain the 35Cl/ 37Cl ratio in the product after different fractions
of reaction.
A sodium borohydride stock solution was prepared under a
nitrogen atmosphere in a glovebag by diluting ∼1.65 g (0.044 mol)
of sodium borohydride to 1 L in a volumetric flask with DMSO.
Then, 150 mL of this solution was pipetted into five different 500-
mL round-bottom flasks. These solutions were used for the partial
reactions. Another sodium borohydride solution was prepared
Determining the Chlorine Isotope Effect Using the IRMS
Method. The silver chloride recovered from the reaction was
converted to methyl chloride using the procedure developed partly
by Hill and Fry5 and partly by Taylor and Grimsrud.6 Ap-
proximately 10 mg of AgCl was put into the reaction vessel (Figure
1), which was wrapped in foil to reduce photodecomposition, and
the vessel was connected to the vacuum line at cup seal C. The
vacuum line, with valves 12, 1, 11, 10, 9, and 7 open and valves 2,
-
2
8, and 5 closed, was evacuated to <5 × 10 mbar. The foil was
removed, and the bottom inch of the reaction vessel was immersed
3
in liquid nitrogen. Valve 1 was closed and 30 µL of CH I [74-88-
4] (Fisher Scientific) was injected into the reaction vessel line
through the rubber septum. The reaction vessel was immersed
in liquid nitrogen to the sidearm, valve 1 was opened, and the
reaction vessel was evacuated for 2 min. Then, valves 1 and 12
were closed, and the reaction vessel was removed from the line,
covered with foil, and heated in the dark at 110 °C for 48 h.
The methyl chloride sample for the mass spectrometric
analysis was collected on the vacuum line shown in Figure 1. The
by dissolving ∼0.49 g (0.013 mol) of NaBH
4
in 150.0 mL of DMSO.
This solution was used for the complete reaction. A benzyl
chloride stock solution was prepared by pipeting 75 mL of DMSO
onto ∼9.0 g (0.071 mol) of benzyl chloride in a 100-mL Erlenmeyer
flask. All of the flasks were fitted with greased serum caps, sealed
with Parafilm and copper wire, and temperature equilibrated for
at least 1 h. The reactions were started by injecting 5 mL of the
benzyl chloride stock solution with a Hamilton gastight syringe
into the reaction flask containing 150 mL of the sodium borohy-
dride stock solution. At predetermined times between 8 and 25%
of completion (see Kinetic Measurements), a reaction vessel was
removed from the bath, 100 mL of cold double-distilled water was
added, and the reaction mixture was titrated to a methyl red end
point immediately with standard 0.3086 M nitric acid.10 The
solution from the titration was transferred to a 500-mL separatory
funnel and extracted once with 100 mL of heptane. The aqueous
layer was put aside and the heptane layer was extracted with 25
mL of double-distilled water. The aqueous layers were combined
(
12) This prevented the loss of HCl in the rotary-evaporator.
(13) The crystal size of the AgCl precipitate is larger when the ionic strength is
high.14
(
14) Harris, D. C. Quantitative chemical analysis, 2nd ed.; W. H. Freeman and
Co.: New York, 1987; p 123.
(15) This prevented the precipitation of Ag
2
CO
was dried at 90 °C for 12 h and then cooled and stored in a
desiccator in the dark. The standard AgNO solution was prepared by
[7761-88-8] (BDH) to 1 L in a
3
.
(
16) The AgNO
3
3
diluting ∼9.5 g (0.056 mol) of dry AgNO
3
volumetric flask. The solution was stored in a dark bottle in the dark to
reduce photodegradation.
(
17) Harris, D. C. Quantitative chemical analysis, 2nd ed.; W. H. Freeman and
Co.: New York, 1987; p 694.
(18) Harris, D. C. Quantitative chemical analysis, 2nd ed.; W. H. Freeman and
Co.: New York, 1987; p 305.
19) This prevented a buildup of AgCl on the wire. Grimsrud reported that
AgCl on the silver electrode during the titration led to incorrect results.
2
0
(
(
(
10) Davis, W. D.; Mason, L. S.; Stegeman, G. J. Am. Chem. Soc. 1 9 4 9 , 71, 2775.
11) Laidler, K. Chemical kinetics, 2nd ed.; McGraw-Hill: New York, 1965; p 8.
(20) Grimsrud, E. P. Ph.D. Dissertation, University of Wisconsin, Madison, WI,
1971.
Analytical Chemistry, Vol. 70, No. 17, September 1, 1998 3549

Ta b le 1 . Th e Ave ra g e Fra c t io n o f Re a c t io n Ba s e d o n t h e HNO3 Tit ra t io n , t h e Ag NO3 Tit ra t io n , a n d t h e We ig h t o f
Ag Cl, t h e δ Va lu e s , a n d t h e Ch lo rin e KIEs fo r Tria ls I a n d II fo r t h e S N2 Re a c t io n s b e t w e e n S o d iu m Bo ro h yd rid e a n d
Be n zyl Ch lo rid e in DMS O a t 3 0 .0 0 0 ( 0 .0 0 2 °C Us in g t h e Co n ve n t io n a l IRMS Me t h o d
f
sample
HNO3
AgNO3
AgCl
av f
δ (35Cl/ 37Cl)a
Trial 1
k / k
35
37 b
δ (35Cl/ 37Cl)c
k / k
35 37 b
HA-1
HA-2
HA-3
HA-4
HA-5
HA-1C
HA-2C
HA-3C
HA-4C
0.0818
0.0915
0.0980
0.1110
0.1642
0.0813
0.0914
0.1006
0.1117
0.1626
0.0817
0.0931
0.1005
0.1131
0.1626
0.0816 ( 0.0002d
0.0920 ( 0.0008
0.100 ( 0.001
0.1119 ( 0.0009
0.1631 ( 0.0007
1.0000
1.0000
1.0000
1.0000
-13.88
-14.29
-13.96
-13.90
-13.87
-7.10
1.006 97
1.007 44
1.007 13
1.007 11
1.007 30
-11.65
-12.06
-11.73
-11.67
-11.64
-4.87
1.006 99
1.007 46
1.007 14
1.007 13
1.007 32
e
-7.17
-4.94
-7.14
-7.20
-4.91
-4.97
d
-4.92 ( 0.04d
(Ro ) 3.142)
av
-7.15 ( 0.04
(
Ro ) 3.149)
av
1.007 19
1.007 21
(0.000 19
0.000 19d
d
(
Trial 2
-9.86
HB-1
HB-2
HB-3
HB-4
HB-5
HB-6
HB-7
HB-1C
HB-2C
HB-3C
HB-4C
HB-5C
av
0.1074
0.1221
0.1342
0.1561
0.1564
0.1767
0.2479
0.1068
0.1207
0.1329
0.1540
0.1549
0.1750
0.2468
0.1070
0.1225
0.1332
0.1557
0.1558
0.1745
0.2469
0.1071 ( 0.0002
0.1218 ( 0.0008
0.1334 ( 0.0006
0.1553 ( 0.0009
0.1557 ( 0.0006
0.1754 ( 0.0009
0.2472 ( 0.0005
1.0000
1.0000
1.0000
1.0000
1.0000
1.007 53
1.007 84
1.007 90
-10.09
-10.09
-9.68
1.007 56
1.007 47
1.007 51
-9.52
-9.24
e
-2.82
-2.67
-2.64
-2.72
-2.78
d
-2.73 ( 0.07
(
Ro ) 3.136)
av
1.007 64
(
0.000 19d
a
These δ values were measured relative to a standard methyl chloride sample. b The chlorine isotope effect was calculated using eq 4. The Ro
c d
value is the average from the samples where f ) 1.000. These δ values were measured relative to a standard mean ocean chloride sample. The
standard deviation of the mean. Complete samples from the reaction taken to 100% of completion.
e
Varian 90-P gas chromatograph (GC) was operated at an injection
temperature of 150 °C, a flow rate of 60 mL/ min, an oven
temperature of 70 °C, a filament current of 150 mA, and an
attenuation of 1. The attenuation of the Hewlett-Packard model
HP3396A integrator was -2. When the GC had stabilized, valves
2
and 5 were closed and valves 3, 4, and 6 were opened allowing
highpurity grade helium (Praxair) to enter the injector of the GC.
Valve 4 and then valve 3 were closed keeping the GC at the normal
pressure. A reaction vessel was clamped to the vacuum line at
C. With valves 1, 2, 11, 10, 9, 7, and 6 open and valves 3, 4, 5, 8,
-
2
and 12 closed, the line was evacuated to <5 × 10 mbar. The
bottom of the U tube was immersed in liquid nitrogen, valve 1
was closed, valve 12 was opened, the reaction vessel was warmed,
3 3
and the CH Cl and the unused CH I were distilled into the U tube
in 2 min. The liquid nitrogen was raised to the top of the U tube,
valve 1 was opened, and the line was evacuated for 2 min. Valves
1
and 2 were closed, the reaction vessel was removed from the
line, valves 9, 7, and 6 were closed, and valves 8 and 5 were
opened. The U tube was warmed to room temperature, and valve
Fig u re 1 . The vacuum line used in converting the silver chloride
into methyl chloride and for the purification of the methyl chloride
samples for the IRMS analyses.
3
and then valve 4 were opened injecting the contents of the U
tube onto the 6 ft by 1/
8
in. o.d. stainless steel Porapak Q, 100-
120-mesh column (Chromatographic Specialties). After 90 s, valve
4
and then valve 3 were closed and the bottom of spiral trap A
had eluted, trap A was immersed in liquid nitrogen, valves 5 and
8 were closed, valves 6, 7, and 9 were opened, and the oven
was immersed in liquid nitrogen. When all of the methyl chloride
550 Analytical Chemistry, Vol. 70, No. 17, September 1, 1998
3

temperature on the GC was increased to 150 °C to elute the methyl
Ta b le 2 . Th e Ave ra g e Fra c t io n o f Re a c t io n Ba s e d o n
t h e HNO3 Tit ra t io n , t h e Ag NO3 Tit ra t io n , a n d t h e
We ig h t o f Ag Cl, t h e Cl/ Cl Ra t io , a n d t h e Ch lo rin e
iodide. The section of the vacuum line from stopcock 1 to
-
2
3 5
3 7
stopcock 5 was evacuated to <5 × 10 mbar, a methanol-liquid
nitrogen bath was raised over spiral trap B (to capture any water),
valve 10 was closed, the liquid nitrogen was removed from spiral
trap A, and the bottom inch of the breakseal tube was immersed
in liquid nitrogen. Spiral trap A was warmed. This caused the
pressure in the vacuum line to rise to ∼5 mbar and then decrease
KIE fo r Tria ls 1 a n d 2 Us in g t h e FAB-IRMS Me t h o d
35Cl/ 37Cl ratio ((SD)
k35/ k37 ((SD)a
sample
av f
Trial 1
HA-1
HA-2
HA-3
HA-4
HA-5
complete
0.0816
0.0920
0.1000
0.1119
0.1631
1.0000
3.1343 ( 0.000 19
3.1425 ( 0.000 26
3.1400 ( 0.000 20
3.1417 ( 0.000 19
3.1376 ( 0.000 19
3.1156 ( 0.000 21b
1.0063
1.0090
1.0082
1.0089
1.0077
-2
to ∼6 × 10 mbar as the CH
3
Cl condensed in the breakseal tube.
Then, most of breakseal tube was immersed in liquid nitrogen,
valve 9 was closed, valve 10 was opened, and the line was
evacuated for 2 min. Finally, the breakseal tube was sealed and
sent for mass spectrometric analysis in a Micromass Sira 9 dual-
inlet isotope ratio mass spectrometer at the University of Waterloo.
The recovery of the methyl chloride, based on the weight of silver
chloride put in the reaction vessel, was 93 ( 2%.
av
1.0080
(
0.0010
Trial 2
HB-1
HB-3
HB-4
HB-6
HB-7
complete
0.1071
0.1334
0.1557
0.1754
0.2472
1.0000
3.1355 ( 0.000 15
3.1400 ( 0.000 15
3.1373 ( 0.000 14
3.1360 ( 0.000 14
1.0074
1.0090
1.0082
1.0079
1.0076
3
5
37
The Cl/ Cl ratios in the starting material, R
o
, and in the
3.1343 ( 0.000 10b
chloride ion, R , formed after several known, but different, fractions
f
3.1138 ( 0.000 05
of reaction, f, were obtained by analyzing the methyl chloride from
av
1.0080
(0.0006
complete and partial reactions, respectively, in a Micromass Sira
9
dual-inlet isotope ratio mass spectrometer. The 35Cl/ 37Cl ratios
a
The chlorine isotope effect was calculated using eq 4. The Ro value
were reported as δ values,
b
is from the sample where f ) 1.000. Average from three different
AgCl samples from the reaction taken to 100% of completion.
δ ) 1000((R - R )/ R )
(2)
r
s
r
(
about 0.2 mm thick and 4 mm in diameter). The silver plate
where R
standard mean ocean chloride (SMOC) sample and R
the 35Cl/ 37Cl ratio in the benzyl chloride starting material, R
the 35Cl/ 37Cl ratio of the chloride ion formed during the reaction,
. Rearranging eq 2 gives
r
is the 35Cl/ 37Cl ratio of a reference methyl chloride or a
is either
, or
with the adhered solid sample was mounted on the tip of the direct
insertion probe. Xenon atoms (6 keV) hitting the surface of the
probe at an incidence angle of 45° were used for ionization. The
negative ions formed in this way were accelerated through a 5-kV
potential and detected in a Faraday cup collector system. The
spectra contained almost exclusively chlorine peaks at m/ z ) 35
and 37. The mean value of the isotopic ratio was obtained from
between 30 and 50 separate determinations, each of which was
an average of 10 individual measurements. The total ion current
under the above conditions was stable for at least 0.5 h. It is worth
noting that the very stable ion current allows one to measure the
isotope ratio with the same precision when a single-ion counter
is used.
s
o
R
f
R ) R (1 - (δ/ 1000))
(3)
s
r
The R
of AgCl recovered from a reaction taken 100% to completion, i.e.,
reacted for at least 10 half-lives (Table 1), and the R value was
o
value was determined from at least three different samples
f
obtained from at least four samples taken at different fractions of
reaction ranging from 8 to 25% of completion (Tables 1 and 2).
The chlorine KIEs were calculated using eq 4, where f is the
average fraction of reaction (Table 1).
Finally, the chlorine isotope effect was calculated using eq 4.
The fraction of reaction, the 35Cl/ 37Cl ratios in the benzyl chloride
starting material (the complete or R
ratios of the chloride ion formed during the reaction (the R
f
o
values), and the 35Cl/ 37Cl
3
5
37
k / k ) log(1 - f)/ [ log(1 - (R / R )f )]
(4)
o
f
values)
found using the FAB-IRMS method are in Table 2.
Determining the Chlorine Isotope Effect by the FAB-MS
Method. The ratio of chlorine isotopes was also measured21 using
a modified, hybrid model MI 1201E FAB-isotope ratio mass
spectrometer (PO Electron, Ukraine) equipped with three collec-
tors: two Faraday cups with fixed focus and an electron multiplier
with pulse analyzer and a single-ion counter. Both detector types
yield comparable precision in isotope ratio determinations.
The silver plate that is mounted on the tip of the probe was
washed with nitric acid, water, and acetone. Then, ∼5 mg of silver
chloride was deposited on the clean silver plate and the plate was
heated gently to melt the sample. Precautions were taken not to
exceed 480 °C and to keep the dimensions of the samples constant
RESULTS AND CONCLUSION
The error in a kinetic isotope effect calculated from the isotopic
composition of a product after small and complete conversion is
determined by the error in the fraction of reaction, f, and the error
o f
in determining the isotopic ratios, i.e., R and R . To minimize
the contribution of the error in the fraction of reaction to the
overall error in the isotope effect, the fraction of reaction was
determined by three independent analyses. Because the error
in determining the average f is less than 1.0% of the value of f
(
Table 1), the fraction of reaction was known precisely and
accurately. This ensured that the error from this source was
small.
In trial 1, the del values of the methyl chloride samples from
the reaction were measured relative to both a standard methyl
(
21) Rudzi n˜ ski, J. Book of Abstracts, 14th International Mass Spectrometry
Conference, Tampere, Finland, 1997; p 267.
Analytical Chemistry, Vol. 70, No. 17, September 1, 1998 3551

chloride sample and a SMOC sample (Table 1). This procedure
gave two different del values for each of the methyl chloride
samples in trial 1. However, the isotope effects calculated from
the two sets of del values were in excellent agreement, i.e.,
nutritional sciences for determining the isotopic ratio of calcium
samples, although in these cases, the sensitivity rather than the
isotopic precision is more important.22
Finally, the chlorine leaving group isotope effects found for
S
N
2 reactions range from approximately 1.0057 to 1.0098.23
1
.007 19 ( 0.000 19 and 1.007 21 ( 0.000 19, respectively (Table
1
). This clearly demonstrates that the magnitude of the isotope
Because the magnitude of the isotope effect increases with the
amount of C -Cl bond rupture in the transition state, the chlorine
leaving group isotope effect of 1.007 45 ( 0.000 19 found for the
2 reaction between benzyl chloride and borohydride ion
effect does not depend on the standard chosen for the IRMS
analysis.
The chlorine leaving group kinetic isotope effects obtained by
the IRMS method in two different experiments are given in Table
R
S
N
indicates there is substantial C
state.
R
-Cl bond rupture in the transition
1
. The agreement between the two isotope effects is similar to
7
that reported by Stein and McLennan, who found that repeated
measurements of a chlorine isotope effect differ by ∼0.0004.
The chlorine isotope effects obtained using the FAB-MS
method (Table 2) are slightly less precise than those found using
the IRMS method. It is important to note, however, that the FAB-
MS and IRMS isotope effects are the same within the experimental
error of the methods. An important factor in the comparison of
the isotope effects is that the lower precision of the FAB-MS
isotope effect is compensated by its simplicity. It is worth noting
that a similar protocol has been used in the biological and
Received for review November 12, 1997. Accepted June
, 1998.
2
AC9712355
(
22) O’Brien, K. O. Book of Abstracts, 6th International Symposium on the Synthesis
and Applications of Isotopically Labeled Compounds, Philadelphia, PA, Sept
1997; p 52.
(
23) Shiner, V. J., Jr.; Wilgis, F. P. In Isotopes in organic chemistry; Buncel, E.,
Saunders: W. H., Jr., Eds.; Elsevier: Amsterdam, 1992; Vol. 8, pp 290-
292.
3552 Analytical Chemistry, Vol. 70, No. 17, September 1, 1998