The Newman-Kwart rearrangement: A microwave kinetic study

Article abstract of DOI:10.1021/jo702633g

(Chemical Equation Presented) The kinetic profile of the Newman-Kwart rearrangement has been evaluated using microwave heating. After first demonstrating equivalence between conventional convective heating and microwave heating, data was gathered and anal

Full text of DOI:10.1021/jo702633g

The Newman-Kwart Rearrangement: A Microwave Kinetic Study
John P. Gilday,† Philip Lenden,† Jonathan D. Moseley,*,† and Brian G. Cox‡
AstraZeneca, Process Research and DeVelopment, AVlon Works, SeVern Road, Hallen, Bristol, BS10 7ZE,
U.K., and AstraZeneca, Process Research and DeVelopment, Silk Road Industrial Park, Charter Way,
Macclesfield, SK10 2NA, U.K.
jonathan.moseley@astrazeneca.com
ReceiVed December 11, 2007
The kinetic profile of the Newman-Kwart rearrangement has been evaluated using microwave heating.
After first demonstrating equivalence between conventional convective heating and microwave heating,
data was gathered and analyzed to determine the effects of substituent, solvent, and concentration on the
reaction order. Reaction rate constants, Arrhenius constants, and activation energies have been determined.
The reaction rate shows strong sensitivity to the substituent and modest sensitivity to the solvent. At
high concentrations, the reaction order increases from the previously reported first-order to a mixed first/
second-order reaction. Overall, this re-evaluation of the Newman-Kwart rearrangement has shown the
reaction rate order to be more complex than previously thought. In addition, microwave heating has
proven ideal for the rapid collection of data to facilitate this type of kinetic study.
Introduction
instantaneous volumetric heating, lack of wall heating effects,
improved purity profile, reduced catalyst loading requirements,
and possible energy savings have been presented previously
elsewhere.1,2
Microwave-assisted organic synthesis (MAOS) was first
reported in 1986 and has grown exponentially since then, as
demonstrated by the number of recent reviews1 and books2 on
the subject. MAOS is now widely used within pharmaceutical
research departments since the features of small scale, fast
reaction time, and ready automation marry well with combi-
natorial and library synthesis techniques, which has provided a
step change in drug discovery programmes.2a Furthermore, in
addition to its now well-established use in post-graduate research
laboratories, microwave heating is a potential boon in under-
graduate teaching laboratories where automation and fast
reaction times can allow students to progress reactions to useful
outcomes within the time constraints of laboratory teaching
sessions.3 The other advantages of microwave heating such as
Early in our own investigations on microwave heating, we
realized that rapid heating and the consequent increased reaction
rates made microwave heating potentially ideal for kinetic
studies. As Strauss has pointed out, multiple data points are
needed for reliable kinetic studies, whether they be derived from
conventional or microwave heating.4 Microwave heating can
provide this data very quickly, and, in addition, can readily
access temperatures above the solvent bp under autoclave
conditions to provide a wider operating envelope for kinetic
analysis. However, in order to validate the microwave kinetic
studies, it was important to establish that conventional and
microwave heating were equivalent.
* Author to whom correspondence should be addressed.
† Avlon Works.
Results and Discussion
‡ Silk Road Industrial Park.
Because of the ongoing debate over nonthermal microwave
effects,5 we decided to concentrate our studies on a completely
reliable reaction for which there were no claims of any
(1) (a) Kappe, C. O. Angew. Chem., Int. Ed. 2004, 43, 6250-6284. (b)
Hayes, B. L. Aldrichim. Acta 2004, 37, 66-76. (c) Nu¨chter, M.; Ondruschka,
B.; Bonrath, W.; Gum, A. Green Chem. 2004, 6, 128-141. (d) Lidstro¨m,
P.; Tierney, J.; Wathey, B.; Westman, J. Tetrahedron 2001, 57, 9225-
9283. (e) Perreux, L.; Loupy, A. Tetrahedron 2001, 57, 9199-9223.
(2) (a) Kappe, O.; Stadler, A. MicrowaVes in Organic and Medicinal
Chemistry; Wiley-VCH: Weinheim, Germany, 2005. (b) MicrowaVe
Assisted Organic Synthesis; Tierney, J. P., Lidstro¨m, P., Eds.; Blackwell:
Oxford, 2005. (c) MicrowaVes in Organic Synthesis, 2nd ed.; Loupy, A.,
Ed.; Wiley-VCH: Weinheim, Germany, 2006. (d) Hayes, B. L. MicrowaVe
Synthesis: Chemistry at the Speed of Light; CEM: Matthews, NC, 2002.
(3) (a) Katritzky, A. R.; Cai, C.; Collins, M. D.; Scriven, E. F. V.; Singh,
S. K.; Barnhardt, E. K. J. Chem. Educ. 2006, 83, 634-636. (b) Leadbeater,
N.; McGowan, C. Clean, Fast Organic Chemistry; CEM: Matthews, NC,
2006.
(4) Strauss, C. R. Angew. Chem., Int. Ed. 2002, 41, 3589-3590 and
relevant references therein.
10.1021/jo702633g CCC: $40.75 © 2008 American Chemical Society
Published on Web 03/22/2008
3130
J. Org. Chem. 2008, 73, 3130-3134

The Newman-Kwart Rearrangement
FIGURE 1. Comparison of extent of reaction with time for conventional (red) and microwave (blue) heating in the rearrangement of 1a (0.25 M)
in NMP at 140 °C.
TABLE 1. Rate Constants (140 °C) and Activation Parameters for
1a in DCB at 0.25 M
SCHEME 1
heating method
k/s-1 at 140 °C
A/s-1
Ea/(kJ mol-1
)
microwave
thermal
1.35 × 10-4
5.1 × 1010
116.2
118.7
1.13 × 10-4
7.8 × 1010
Comparison Between Conventional Thermal and Micro-
wave Temperature Control/Heating. Although we were
confident that the NKR gave equivalent reaction rates under
both conventional and microwave heating in general synthetic
applications,7 we decided to establish a higher level of accuracy
before beginning this study. Therefore, experiments were
performed to confirm that reactions carried out under microwave
heating showed similar kinetic behavior to those carried out
under conventional convective heating (i.e., in a thermostatically
controlled silicon oil heating bath). The initial comparison was
performed on identical scale in microwave reaction tubes with
stirring bars. The oil bath reaction tube was sampled at multiple
time points, while each microwave data point was taken from
a single reaction tube. This avoids a possible cumulative error
in the microwave case for heating up the microwave tubes, since
they must be cooled each time before sampling (although control
studies show these effects to be negligible for reaction times of
>10 min, even at low conversions). Figure 1 shows the
comparison for the reaction of 2-nitrophenyl-O-thiocarbamate
(1a) at 0.25 M in NMP at 140 °C under conventional and
microwave heating. The solid lines are best-fit lines and show
that the rate constants for the reactions agree within 4%.
Reaction of the 2-nitro substrate 1a in dichlorobenzene (DCB)
over the temperature range of 120-180 °C at the same
concentration showed very similar rates and activation param-
eters using the two methods of heating (Table 1).
The difference in rate constants at 140 °C corresponds to an
effective temperature for microwave control of 3.5 °C above
that for thermal control. In other cases, apparent small increases
in effective temperature of between 2 °C and 4 °C at around
150 °C were observed for reactions in which microwave heating
was used to achieve and control reaction temperature; the effects
were smaller at higher temperatures. This difference, if real,
suggests a small extent of superheating associated with the
microwaves, depending upon the efficiency of the reaction
stirring, or possibly a slight systematic error in the temperature
measurement.
microwave effect, and which was reliable under a wide range
of known parameters (temperature, pressure, solvent, etc.). We
have already reported our findings on re-evaluating the New-
man-Kwart rearrangement (NKR)6 under both microwave and
conventional thermal heating, and have shown that there is no
difference between the two under easily achieved and well-
controlled conditions.7 The NKR has previously been shown
to be a first-order, unimolecular rearrangement converting an
O- to S-thiocarbamate (Scheme 1; 1 to 2).8 The rate of reaction
is also well-known to be strongly dependent on the aromatic
substituent,6a,8,9 and we have more recently shown a modest
solvent effect also.7,10
It is important to note that the reaction mixture must be a
well-stirred homogeneous solution to avoid the potential problem
of localized superheating due to inefficient agitation.10 It is
equally important to measure the temperature that accurately
reflects the contents of the reaction vessel. Although others have
rightly pointed out the potential concerns with relying on the
use of external IR pyrometers,11 compared to the more reliable
fiber optic probes or shielded thermocouples inserted in the
reaction mixture for example, we have found external IR
pyrometers to be reliable for this reaction when regularly
calibrated and in the hands of a competent user.7
(5) de la Hoz, A.; Diaz-Ortez, A.; Moreno, A. Chem. Soc. ReV. 2005,
34, 164-178.
(6) (a) Newman, M. S.; Karnes, H. A. J. Org. Chem. 1966, 31, 3980-
3984. (b) Kwart, H.; Evans, E. R. J. Org. Chem. 1966, 31, 410-412.
(c) Lloyd-Jones, G. C.; Moseley, J. D.; Renny, J. S. Synthesis 2008, 661-
689.
(7) (a) Moseley, J. D.; Sankey, R. F.; Tang, O. N.; Gilday, J. P.
Tetrahedron 2006, 62, 4685-4689. (b) Moseley, J. D.; Lenden, P.
Tetrahedron 2007, 63, 4120-4125.
(8) (a) Miyazaki, K. Tetrahedron Lett. 1968, 2793-2798. (b) Araki, Y.;
Kaji, A. Bull. Chem. Soc. Jpn. 1970, 43, 3214-3219. (c) Kaji, A,; Araki,
Y.; Miyazaki, K. Bull. Chem. Soc. Jpn. 1971, 44, 1393-1399.
(9) Relles, H. M.; Pizzolato, G. J. Org. Chem. 1968, 33, 2249-2253.
(10) Moseley, J. D.; Lenden, P.; Thomson, A. D.; Gilday, J. P.
Tetrahedron Lett. 2007, 48, 6084-6087.
Finally, we further tested this point by studying the rear-
rangement of 1a at a fixed bulk temperature (140 °C) using
(11) Nu¨chter, M.; Ondruschka, B.; Weiâ, D.; Beckert, R.; Bonrath, W.;
Gum, A. Chem. Eng. Technol. 2005, 28, 871-881.
J. Org. Chem, Vol. 73, No. 8, 2008 3131

Gilday et al.
FIGURE 2. Rearrangement of 2-nitrophenyl-O-thiocarbonate (1a) in DCB at 150 °C. Experimental data are represented by the points, and the
solid lines represent values calculated using the rate constants (eq 1): k1 ) 2.02 × 10-4 s-1; k2 ) 1.42 × 10-4 M-1 s-1
.
TABLE 2. Comparison of Thermal and Microwave Reactions at
Different Microwave Powers for 1a in Xylene at Solvent Refluxa
under Open Vessel Conditions
under temperature control. Once superheating has been achieved
by high microwave power, the increase from 100 to 300 W
makes no significant difference; more power cannot be forced
into a system that is already superheated, even though the
temperature probe gradually indicates a higher temperature.12
For all systems, the same trends were observed, independent
of the method of heating. For the remainder of this work,
quantitative analysis refers to microwave reactions run under
temperature control for which more comprehensive results were
obtained.
Reaction Kinetics. Reactions were investigated at substrate
concentrations varying from 0.0625 to 4 M. Initial data treatment
was based on purely first-order kinetics, but, in all cases, there
was a regular increase in apparent first-order rate constant as
the concentration increased over this range. The effects were
particularly significant at concentrations of 1 M and above, with
the apparent rate constants being typically a factor of 2 higher
at a reaction concentration of 2 M compared with low
concentrations (0.25 M and below). It was also observed that,
within individual reactions at higher concentrations, there were
small but systematic deviations from exact first-order kinetics,
with the apparent rate constant decreasing with increasing extent
of reaction.
heating method/
(microwave power supplied)
k/s-1
temperature/°C
oil bath (0 W)
4.02 × 10-5
3.95 × 10-5
7.9 × 10-5
8.1 × 10-5
140b
microwave (∼25 W)
microwave (100 W)
microwave (300 W)
138-142c
154-159c
160-165c
a Conversion of 1a to 2a at 140 °C and 0.44 M in o-xylene (bp 138-
139 °C). b Measured by calibrated thermocouple. c Measured by calibrated
IR pyrometer.
refluxing m-xylene as the solvent (bp 138-139 °C) at a
concentration of 0.44 M. The microwave reactor can be used
in open vessel mode with a conventional round-bottomed flask
heated to the solvent’s bp. The microwave supplies high power
initially to achieve the set point temperature (140 °C) but then
provides only enough power to maintain this temperature
(typically ∼25 W for a 10 mL sample in this case). Working
with a larger sample (10 mL vs 2 mL) and at the solvent bp,
where the latent heat of evaporation tends to stabilize the bulk
temperature at the desired set point, made this experiment more
robust. In addition, because it is in an open vessel, the
microwave reaction vessel could be sampled while being heated,
which required only one reaction mixture to be prepared and
heated to temperature. The conventional convective heating
reaction mixture was run under identical conditions of concen-
tration, scale, vessel, etc. to achieve a similar level of reflux.
Once data had been collected at normal reflux in the microwave
case (using ∼25 W), the power was increased to a fixed 100
W setting so that the reactor was now heated under power
control, not temperature control. Further data points were taken,
and then the power was ramped up to the maximum 300 W
available. However, it was noticeable that, under both these sets
of conditions, the IR pyrometer gradually recorded an increase
in temperature, eventually up to 165 °C, indicating significant
superheating of the solvent.11,12 The apparent first-order rate
constants under these conditions are shown in Table 2.
It was thought possible that these effects are due to a change
in reaction medium as the substrate concentration is increased,
but this was not consistent with the relative insensitivity of the
reaction to solvent (see below). Furthermore, the systematic de-
viations from first-order kinetics are difficult to reconcile with
a medium effect. The data were therefore analyzed as a mixture
of first and second-order kinetics, according to eq 1 for sub-
strate S:
-d[S]/dt ) k1[S] + k2[S]2
(1)
A typical result is shown in Figure 2 for the rearrangement of
1a to 2a in DCB at 150 °C. The results show an excellent quan-
titative fit to eq 1, with the calculated curves corresponding to
k1 ) 2.02 × 10-4 s-1; k2 ) 1.42 × 10-4 M-1 s-1
The results, while confirming the equivalence of the two
heating methods at low power, show clear evidence of super-
heating at high microwave power, but there is no evidence of
any fundamental difference between the reactions carried out
under conventional thermal and microwave heating (∼25 W)
Similar behavior was found for all substrates, solvents, and
temperatures, irrespective to the form of heating (microwave
or thermal), and it was typically observed that
k1/k2 ≈ 2 M
(2)
(12) Chemat, F.; Estveld, E. Chem. Eng. Technol. 2001, 24, 735-744.
3132 J. Org. Chem., Vol. 73, No. 8, 2008

The Newman-Kwart Rearrangement
TABLE 3. Rate Constants for the Rearrangement of
Aryl-O-thiocarbamates 1a-d in DCB at 140 °C
TABLE 5. Activation Parameters for the Rearrangement of
Aryl-O-thiocarbamates
∆Hq1/
∆Sq1/
∆Hq2/
∆Sq2/
a
aryl substituent (X)
k1 (s-1
)
k1(X)/k1(2-NO2)
system/solvent
J mol-1 a J K-1 mol-1 a kJ mol-1 b J K-1 mol-1 b
2-NO2
9.6×10-5
5.7×10-5
1.0×10-6
3.4×10-8
1
0.59
0.010
0.00035
4-NO2
2-NO2/DCB
2-NO2/DMA
4-NO2/DMA
115.9 ( 1.6 -44.7 ( 3.9 100.8 ( 3.0 -83.7 ( 8.0
111.8 ( 1.6 -51.4 ( 4.0 104.2 ( 3.0 -76.3 ( 7.0
120.9 ( 1.8 -41.2 ( 3.9 100.8 ( 3.0
4-CO2Meb
4-Clc
-96 ( 9
4-CO2Me/DCB 122.2 ( 1.8 -66.5 ( 5.0
4-Cl/DCB 134.6 ( 2.0 -64.7 ( 5.0 131.3 ( 4.0 -77.3 ( 8.0
a First-order process. b Second-order process.
a Equation 1. b Extrapolated from measurements at 210-250 °C, using
k1 ) 9.54 × 109 exp(-126300/RT) s-1
.
c Extrapolated from measurements
at 210-250 °C using k1 ) 1.19 × 1010 exp(-138800/RT) s-1
.
TABLE 4. Rearrangement of 2-Nitrophenyl-O-thiocarbamate (1a)
at 140 °C in Solvents
dichlorobenzene to the very highly polar solvents N,N′-dim-
ethylacetamide (DMA) and N-methyl 2-pyrrolidinone (NMP).
Earlier studies have also shown relatively little variation among
aprotic solvents of moderate to high polarity, but a significant
rate increase when the reactions are carried out in formic acid
as solvent.7
a
solvent
104 k1/s-1 (eq 1)
A/s-1a
Ea/(kJ mol-1
)
DCB
DMA
NMP
0.96
1.04
1.26
0.40
4.7 × 1010
4.4 × 1010
n/d
116.2
115.7
n/d
o-xylene
n/d
n/d
Activation Parameters. Enthalpies (∆Hq) and entropies
(∆Sq) of activation for several systems were determined using
eqs 3 and 4:
a Arrhenius parameters from k1 ) Ae-E /RT; n/d ) not determined; DCB
) 1,2-dichlorobenzene; DMA ) N,N-dimethylacetamide; NMP ) N-methyl-
2-pyrrolidinone.
a
k ) (kBT/h) exp(∆SqR) exp(-∆HqRT)
(3)
It follows that, at substrate concentrations of around 0.2 M and
below, the first-order process dominates, with the second-order
process contributing <10% to the initial rate and a decreasing
proportion as the reaction proceeds. At these concentrations,
the difference between k1, determined via eq 1, and via a simple
first-order rate law (neglecting any second-order contribution)
is within experimental error ((5%). At reaction concentrations
of 2 M, however, approximately 50% of the initial rate is due
to the second-order process.
or
ln k/T ) ln(kB/h) + ∆SqR - ∆HqRT
(4)
Figure 3 shows k1 values for 2-nitrophenyl-O-thiocarbamate
(1a) in DCB over the temperature range of 120-170 °C plotted
according to eq 4.
Substituent Effects on Reaction Rates. Rate constants were
measured in DCB at 140 °C for several O-aryl thiocarbamates
with differing electron-withdrawing aryl-substituents, and the
first-order rate constants, k1 (eq 1), are listed in Table 3. The
results agree with those from earlier studies, which show a strong
substituent effect on the reaction.6a,8,9
Solvent Effects on Reaction Rates. The reaction rates were
found to show little sensitivity to solvent in the (aprotic) solvents
investigated. Table 4 shows rate constants, k1, obtained for the
first-order rearrangement of 2-nitrophenyl-O-thiocarbamate (1a)
at 140 °C in various solvents. Only in the very low-polarity
solvent o-xylene was there a noticeable decrease in reaction rate
constants relative to those observed for solvents varying from
Results for the various systems are listed in Table 5.
The entropies of activation of the bimolecular process are
strongly negative and are comparable to typical values observed
for the coming together of two molecules to form the transition
state, namely, -50 to -160 J K-1 mol-1 in the gas phase and
frequently somewhat smaller in solution.13 The activation
enthalpies are also generally lower than those of the corre-
sponding first-order process.
The first-order reactions also show quite strongly negative
entropies of activation, reflecting the loss in conformational
freedom resulting from the formation of the proposed four-
membered transition state (Figure 4). Earlier work suggested
that entropy losses upon forming the transition state are lower
FIGURE 3. Activation plot (eq 4) 2-nitrophenyl-O-thiocarbamate (1a) in DCB. Points correspond to experimental values; solid line calculated
according to eq 4 with ∆Hq) 115.9 kJ mol-1, ∆Sq) -44.7 J K-1 mol-1
.
J. Org. Chem, Vol. 73, No. 8, 2008 3133

Gilday et al.
reactant concentration. Quantitative analysis of the data over a
range of reactions and concentrations strongly suggests that this
is due to a change in reaction order at higher (>1 M)
concentrations. Overall, the reaction order of the Newman-
Kwart rearrangement is more complex than previously thought,
especially at the high concentrations typically used in preparative
and synthetic procedures. Finally, microwave heating has proven
to be an ideal method for the rapid acquisition of data for a
kinetic study of this kind.
FIGURE 4. Proposed four-center transition state.
Experimental Section
For each series of experiments, homogeneous solutions of
the appropriate aryl-O-thiocarbamate 1a-d were made up at
the required concentrations (0.0625-4.0 M) in the appropriate
solvent (NMP, DMA, DCB, or o-xylene). Equal volumes
(typically 2.0 mL) were heated with stirring in thick-walled glass
sealed tubes in a CEM Discover monomode 300 W microwave
reactor with IR temperature monitoring and noninvasive pressure
transducer. The heating time to reach the set temperature was
typically 30-90 s, depending on the scale, the maximum
wattage supplied (100-300 W), and the temperature required
(100-250 °C). The heating time is not included in the quoted
hold time for any given procedure, but control studies show
that the heating time has a negligible effect on overall conver-
sion.
Conventional (oil bath) and microwave open vessel experi-
ments were performed at the same bulk temperature in o-xylene
at reflux (∼140 °C) and sampled at regular time points (typically
every 10 min). Open vessel reflux in the microwave reactor
required ∼25 W to maintain a 10 mL sample once at 140 °C.
Later microwave experiments were performed under fixed
wattage control (100 or 300 W), whereupon the IR pyrometer
gradually registered higher temperatures (as the microwave
reactor retained more heat), while the bulk sample was notion-
ally still at reflux.
FIGURE 5. Possible bimolecular transition state structure.
in systems with sterically bulky substituents in the 2-position,
because of hindered motion in the ground state,9 but this does
not appear to be a significant factor for the substrates investi-
gated here.
A recent computational study concluded that the unimolecular
rearrangement is reasonable (concentration would not be
considered in gas-phase calculations) and proceeds in synchro-
nous fashion through the four-membered cyclic transition state
proposed (Figure 4) not involving an intermediate.14 Given the
conclusions of this computational study, one might propose that
the bimolecular reaction could proceed through a synchronous
mechanism via an analogous eight-membered transition state
as postulated in Figure 5. This might be expected to exhibit a
lower strain energy, but the entropy losses associated with the
bringing together of two reactant molecules would mean that
this process would be relatively disfavored at lower concentra-
tions. As noted above, quite high concentrations (∼2 M) are
required before the second-order process makes a similar
contribution to the overall reaction rate to that of the first-order
process.
Conversions were determined by HPLC analysis of diluted
samples and are corrected for relative response factors (RRFs).
The aryl-S-thiocarbamate products (2a-d) had been isolated
previously to provide reference markers for the HPLC method
and RRF values.7
Conclusions
In summary, we have reported a microwave-facilitated kinetic
study of the Newman-Kwart rearrangement. This has confirmed
that the aromatic substituent has the greatest effect on the
reaction rate, with somewhat smaller variations arising from
change in solvent. The reaction rate (percentage conversion at
a given time) has been shown to increase steadily with increasing
Supporting Information Available: HPLC method; general
experimental conditions; preparation and data on aryl-O-thiocar-
bamates 1a-d; physical and spectroscopic data on aryl-S-thiocar-
bamates 2a-d; general notes on the use of microwave reactors in
obtaining kinetic data points; primary kinetic data for activation
parameters (Tables S1-S5); additional references for Supporting
Information. This material is available free of charge via the Internet
at http://pubs.acs.org.
(13) Hammett, L. P. Physical Organic Chemistry: Reaction Rates,
Equilibria, and Mechanisms; McGraw-Hill: New York, 1970.
(14) Jacobsen, H.; Donohue, J. P. Can. J. Chem. 2006, 84, 1567-1574.
JO702633G
3134 J. Org. Chem., Vol. 73, No. 8, 2008