Microwave-specific acceleration of a friedel-crafts reaction: Evidence for selective heating in homogeneous solution

Article abstract of DOI:10.1021/jo501153r

Thermally promoted Friedel-Crafts benzylation of arene solvents has been examined under both conventional convective heating with an oil bath and heating using microwave (MW) energy. Bulk solution temperatures-as measured by internal and external temperature probes and as defined by solvent reflux-were comparable in both sets of experiments. MW-specific rate enhancements were documented under certain conditions and not others. The observed rate enhancements at a given temperature are proposed to arise from selective MW heating of polar solutes, perturbing thermal equilibrium between the solute and bulk solution. Central to MW-specific thermal phenomena is the difference between heat and temperature. Temperature is a measure of the ensemble average kinetic molecular energy of all solution components, but temperature does not provide information about solute-specific energy differences that may arise as a consequence of selective MW heating. Enhanced chemical reactivity of the MW-absorbing solute can be described as a MW-specific extra-temperature thermal effect , because the measurable solution temperature only captures a portion of the solute kinetic molecular energy. Experimental factors that favor MW-specific rate enhancements are discussed with an eye toward future development of MW-actuated organic reactions, in which the observed thermal reactivity exceeds what is predicted from temperature-based Arrhenius calculations.

Full text of DOI:10.1021/jo501153r

Article
pubs.acs.org/joc
Microwave-Specific Acceleration of a Friedel−Crafts Reaction:
Evidence for Selective Heating in Homogeneous Solution
Michael R. Rosana, Jacob Hunt, Anthony Ferrari, Taylor A. Southworth, Yuchuan Tao, Albert E. Stiegman,
and Gregory B. Dudley*
Department of Chemistry and Biochemistry, Florida State University, Tallahassee, Florida 32306-4390, United States
S
* Supporting Information
ABSTRACT: Thermally promoted Friedel−Crafts benzylation of arene
solvents has been examined under both conventional convective heating
with an oil bath and heating using microwave (MW) energy. Bulk solution
temperaturesas measured by internal and external temperature probes
and as defined by solvent refluxwere comparable in both sets of
experiments. MW-specific rate enhancements were documented under
certain conditions and not others. The observed rate enhancements at a
given temperature are proposed to arise from selective MW heating of polar
solutes, perturbing thermal equilibrium between the solute and bulk
solution. Central to MW-specific thermal phenomena is the difference between
heat and temperature. Temperature is a measure of the ensemble average
kinetic molecular energy of all solution components, but temperature does
not provide information about solute-specific energy differences that may
arise as a consequence of selective MW heating. Enhanced chemical
reactivity of the MW-absorbing solute can be described as a MW-specific “extra-temperature thermal effect”, because the
measurable solution temperature only captures a portion of the solute kinetic molecular energy. Experimental factors that favor
MW-specific rate enhancements are discussed with an eye toward future development of MW-actuated organic reactions, in
which the observed thermal reactivity exceeds what is predicted from temperature-based Arrhenius calculations.
INTRODUCTION
■
From the seminal observations of Gedye1 and Majetich2 that
chemical reactions can be conducted in kitchen microwave
ovens, synthetic chemists have been intrigued by the possibility
that microwave heating can complement conventional heating
and provide specific advantages for certain applications.
Inspiration comes in part from food chemistry: certain foods
and beverages such as popcorn, coffee, and bacon are often
conveniently heated in a microwave oven, whereas people may
prefer that other items such as toast, pancakes, and steak be
prepared by conventional heating methods. Microwave ovens
heat by fundamentally different mechanisms than conventional
heat sources.3 The different heating processes (cf. Figure 1)
Figure 1. Simulated thermal image adapted from the product literature
produce different outcomes in complex food chemistry systems,
of an early MW manufacturer4 illustrating differences between MW and
convective heating after 1 min.
for example, prompting some to wonder if differences could
likewise be observed in the comparatively simple systems
involved in stepwise chemical synthesis.
mechanism by which MW radiation can couple resonantly with
a molecule in a way that could accelerate bond-changing events,
as Stuerga thoroughly addressed in a series of early papers.5
Many of the reported observations turned out to be
compromised by flawed temperature measurements.6 As
synthetic chemists have come to appreciate the experimental
MW Effects and Controversy. With perhaps varying
degrees of appreciation for the underlying dielectric loss
processes, chemists began seeking experimental evidence of
novel MW effects in organic synthesis. Certain early reports
described rate accelerations of several orders of magnitude over
conventional heating, leading to (probably unfounded)
speculation that nonthermal coupling of MW radiation with
molecules in solution can alter or reduce the activation energy of
the reaction. However, there is no obvious quantum mechanical
Received: May 23, 2014
Published: July 22, 2014
© 2014 American Chemical Society
7437
dx.doi.org/10.1021/jo501153r | J. Org. Chem. 2014, 79, 7437−7450

The Journal of Organic Chemistry
Article
challenges associated with measuring temperature under MW
heating, claims of dramatic and nonthermal rate accelerations
have largely dissipated.7 It is most reasonable to assume that
MW energy must first be converted into thermal energy, and it is
thermal energy (molecular motion) that drives chemical
reactivity.
Dramatic MW-specific effects in heterogeneous systems (cf.
food chemistry) are common, including heterogeneous syn-
thesis and catalysis. Two significant examples are presented here
(Figure 2). The first is a study by Hulshof et al.,16 in which MW
While claims of incredible MW effects recede from the
organic synthesis literature, there remains an anecdotal lore
associated with MW heating, as well as specific examples that are
difficult to dismiss as artifacts of experimental error. The
Yamada laboratory has reported several examples of enantiose-
lective reactions in which MW heating has resulted in enhanced
reactivity in comparison to conventional heating at the same
temperature (eq 1).8 The accuracy of their temperature
Figure 2. (left) Arcing of Mg turnings exposed to MW irradiation,
reproduced from ref 16b. This energy-transfer event is not caused by
convective heating. (right) MW-specific alteration of observed
thermodynamic equilibration between CO and CO2 in the Boudouard
reaction; reproduced from ref 18.
electromagnetic radiation was shown to accelerate formation of
Grignard reagents by magnesium surface energy-transfer events
that cannot be duplicated by conventional heating. The second
was a recent examination of the industrially relevant Boudouard
reaction:17 C + CO2 ⇆ 2CO. The highly endothermic (ΔH =
41 kcal/mol) production of CO becomes exergonic at high
temperatures due to favorable entropy; the equilibrium
concentration of CO exceeds that of CO2 only at temperatures
above 643 °C (916 K). Stiegman and co-workers reported that
MW heating of the solid carbon surface selectively accelerates
the forward reaction and lowers the temperature at which CO
becomes the major component to 213 °C (486 K).18 MW-
specific carbon surface activation events were postulated. These
examples illustrate the potential of MW energy to improve
chemical reaction processes in ways other than by bulk heating
of the system, but the macroscopic heterogeneity of these
systems was critical to the effects in both cases. Our objective is to
observe and quantify similar effects in macroscopically homogeneous
systems.
measurements is underscored by the fact that they report
similar enantioselectivities in both systems. Enantioselectivity
and rate usually correlate inversely, but the rate/selectivity
correlation is apparently somehow decoupled in the Yamada
system.
Pulsed MW heating (with intermittent cooling) proved
advantageous in a key Claisen rearrangement in Ley’s synthesis
of azadirachtin.9 Intermittent MW heating was more effective
than continuous MW heating (eq 2), and MW heating was
What many of these processes have in common is a MW-
absorbing catalyst or substrate in a less absorbing system, a
combination that was recognized early on as being of some
interest.19 This common profile of several systems for which
MW heating apparently provides an advantage is consistent with
(and in some cases helped guide) the design of the focal reaction
system of our study. It is also consistent with current theory on
the mechanisms of MW heating:
MW Theory. The conversion of microwave energy into
thermal energy requires a medium that interacts with and
absorbs MW electromagnetic radiation in such a way that heat is
generated.3
explored only after conventional heating options were
exhausted. We have investigated the impact of pulsed MW
heating on aryl Claisen rearrangements and found quantifiable
traces of MW-specific thermal effects.10
Dieters reported a MW effect on metal-catalyzed alkyne
trimerizations (eq 3).11 MW-specific advantages for this and
For a dielectric (i.e., nonconductive, nonmagnetic) material,
the MW → thermal energy conversion is macroscopically
related to its frequency-dependent permittivity. Bulk permittiv-
ityε, the ability of a substance to transmit an electric fieldis
expressed as the sum of its real and complex components: ε = ε′
− iε″. If no absorption is occurring (ε″ = 0), then the real
component e′ is equal to the dielectric constant, e. The
imaginary component ε″ characterizes the dielectric loss, which
is the source of heat. The dielectric loss tangenttan δ, the
portion of the electric field that is lost to heatis quantified as
the ratio of the imaginary and real components of the
permittivity function: tan δ = ε″/ε′. The value of tan δ varies
with temperature and the frequency of the applied field. It is
typically reported at 2.45 GHz, the standard frequency used in
other metal-catalyzed processes, especially alkene metathesis
reactions, have been described12 and questioned.13 MW effects
have also been documented in the synthesis of metallic
nanoparticles14 and on solvent molecules in proximity to a
metal surface.15
7438
dx.doi.org/10.1021/jo501153r | J. Org. Chem. 2014, 79, 7437−7450

The Journal of Organic Chemistry
Article
kitchen microwave ovens. (It is important to bear in mind that
the energy associated with this frequency is insufficient to
promote bond-cleavage events.5 As noted above, MW energy
generally must first be converted into thermal energy to
promote reactivity.)
the transient domain is defined by the absorbing solute within its
solvent cage. Multicomponent solutions are macroscopically
homogeneous, but solvent cages produce transient heterogeneity at
the molecular level. According to Richert, “local heating can occur
in cases where absorptivity is a spatially varying quantity”.22c As
the solute absorbs MW radiation and converts it into thermal
energy, thermal equilibrium between the transient solvent cage
domains and the bulk solution is perturbed. The solvent cage
domain temperature will be higher than what can be
thermometrically recorded as the bulk temperature. The
temperature inside the solvent cage will be effectively higher than
the measurable temperature of the bulk solution.
This local heating model is conceptually similar to “temper-
ature sculpting in yoctoliter volumes” using laser photo-
excitation of gold nanoparticles, as described last year by Reiner
et al.,23 and it rationalizes the previously cited concept of
molecular radiators,24 in which MW-specific heating of a polar
solute results in bulk heating of the otherwise nonabsorbing
solution. We define a molecular radiator in the context of MW
chemistry as a strongly absorbing solute that converts applied MW
energy into thermal energy, generating heat in an otherwise poorly
MW-absorbing system. As heat from the molecular radiator flows
to the bulk solution and eventually out of the system to the
surroundings, there will necessarily be a thermal gradient in
which the molecular radiator (e.g., within its solvent cage) is
effectively hotter than the bulk solution, which in turn is hotter
than the surroundings. The simple analogy is using a radiative
space heater to heat a room: the radiator must be hotter than the
room to serve its purpose. This MW-specific heating effect cannot
be duplicated using convective heating.
The “transient solvent cage domain” model provides an
opportunity to design and exploit MW-specific thermal effects in
solution. If the MW-absorbing solute within its solvent cage is
also a reactant in a chemical process, and if thermally promoted
reactivity can occur more quickly than thermal equilibration of
the solution,25 then selective MW heating of the absorbing
solute and MW-specific reaction rate enhancements should be
observable. The observed reactivity would occur as a function of
the effective temperature within the domains, which would be
greater than the measurable temperature of the bulk solution.
The term “effective temperature” (Teff) is used here to describe
the average kinetic molecular energy of the MW-absorbing
solute molecules. Teff cannot be measured directly; it must be
inferred by studying chemical behavior (e.g., reaction kinetics).
Central Hypothesis. Our central hypothesis is that chemical
reactivity of a MW-absorbing solute in an otherwise-non-
absorbing system can be MW-actuated: accelerated using
applied MW energy to rates that cannot be duplicated by
conventional heating to similar reaction temperatures.26
For molecules in solution, classic Debye theory20 addresses
the MW energy loss processes that result in volumetric heating.
Molecular dipoles in solution rotate in response to the
oscillating electric field of the incident radiation. However,
due to collisional processes in solution, the frequency of the
molecular oscillations is out of phase with that of the electric
field. This results in loss processesoften thought of as
frictionalthat generate heat. The quantity of heat generated is
related to the frequency of the applied field, the size and dipole
of the absorbing molecule, the viscosity of the solution, and
other factors.
For a homogeneous (and efficiently stirred) solution, such
loss processes occur uniformly throughout the system, enabling
the rapid and efficient volumetric heating that is so intimately
associated with MW chemistry.21 For pure liquids (single-
component solutions), all molecules are identically suited to
interact with the applied field, and selective MW heating is not
observable.22 The same is effectively true for multicomponent
solutions in which the solvent (major solution component)
strongly absorbs MW energy, such as reactions conducted in
high tan δ solvents like DMSO and NMP. Given proper stirring,
thermal homogeneity should be maintained throughout the
system.
Model for Selective MW Heating in Solution. Richert and
co-workers offer a model for selective heating in slow-moving
liquids using a low-frequency electric field.22 Their model
includes transient domains in which solute configurational
temperature can be increased faster than the measurable bulk
solution temperature (Figure 3). The energy of the electric field
does work on the polar solute within the transient domain,
creating kinetic molecular energy that eventually propagates
through the liquid as heat.
We apply their model to MW heating of multicomponent
solutions in which the solvent is nonabsorbing and a specific
solute (minor solution component) is strongly MW-absorbing;
Central to this hypothesis is the distinction between thermal
energy (heat) and temperature. Temperature, an intensive
property, describes the ability of a substance to transfer heat;
heat is an extensive property.27 For solution-phase chemical
reaction systems, temperature reflects the ensemble average
kinetic molecular energy of the solution components. The
relationship between solution temperature and solute kinetic
energy applies uniformly only if the solute and solution are at
thermal equilibrium. The experiments described herein are
designed to perturb thermal equilibrium such that the average
kinetic molecular energy of the MW-absorbing solute is higher
than what would be expected on the basis of the macroscopically
observable bulk solution temperature.
Figure 3. Graphical illustration of the “transient solvent cage domain”
model for selective MW heating of strongly MW absorbing solutes in
an otherwise MW-transparent system. Interactions between the solute
and MW radiation generate heat within transient solvent cages,
perturbing thermal equilibrium between solute and solvent. Heat flows
from the domain to the bulk solution and then out of the system.
7439
dx.doi.org/10.1021/jo501153r | J. Org. Chem. 2014, 79, 7437−7450

The Journal of Organic Chemistry
Article
Scheme 1. Thermal Friedel−Crafts Benzylation, the Focal Reaction of This Study, with Key Experimental Design Parameters
Scheme 2. Previously Reported Thermal Friedel−Crafts Reaction29
We envisioned testing our central hypothesis using a reaction
system in which the reactant of interest is highly and uniquely
capable of absorbing MW energy, and any heat generated within
the system would emanate from this focal reactant. We designed
a system (cf. Scheme 1) in which an ionic benzyl-transfer
reagent28 undergoes thermolysis, nominally to a phenyl-
carbenium species, which then reacts with the nonpolar
aromatic solvent in a Friedel−Crafts substitution process.
Ionic species are typically strongly MW absorbing, and our
system was carefully designed to exclude any other potential
MW-absorbing materials. We ultimately selected nonpolar p-
xylene as the solvent (and Friedel−Crafts coupling partner),
employed custom-made reaction vessels constructed out of
MW-transparent quartz, and otherwise took efforts to moderate
bulk solution temperature while maximizing the selective
heating of the MW-absorbing solute. All of the reaction
mixtures were macroscopically homogeneous and vigorously stirred
throughout all of the experiments described in this report unless
otherwise noted.
Making a Soluble Salt. A series of analogues was prepared
and crudely screened for enhanced solubility.26 The salt was
stirred in toluene at either room temperature or at 35 °C and
then allowed to settle to the bottom of the flask. An aliquot was
then removed and concentrated under reduced pressure, and
the mass of residue was then correlated crudely with solubility.
The commercially available triflate salt (BnOPT) was sparingly
soluble in toluene at 35 °C. Increasing the size (and
hydrophobicity) of the heterocycle (e.g., quinoline deriva-
tives)30 correlated with enhanced solubility in toluene, but the
counteranion seemed to have a larger effect. Whereas the
tetrafluoroborate salt of 2-benzyloxy-1-methylpyridinium was
practically insoluble, the larger and more hydrophobic tetrakis-
[3,5-bis(trifluoromethyl)phenyl]borate counterion (BArF)31
enhanced solubility to the point at which homogeneous
solutions in toluene and other aromatic solvents could be
achieved by stirring in a warm water bath.
The experiments described below involved heating identical
solutions of 2-benzyloxy-1-methylpyridinium tetrakis[3,5-bis-
(trifluoromethyl)phenyl]borate (BnOPB, Figure 4) in either a
RESULTS AND DISCUSSION
■
First Attempts. We had previously reported thermal
Friedel−Crafts benzylation of aromatic solvents.29 A heteroge-
neous mixture of 2-benzyloxy-1-methylpyridinium triflate
(BnOPT) in solvent was heated with stirring at 80 °C in an
oil bath (e.g., Scheme 2). The salt dissolved upon heating, and
degradation of the salt coupled with benzylation of the solvent
reached full conversion typically within 24 h. Thermal ionization
of BnOPT to produce a phenylcarbenium-type species, 2-
methylpyridone, and triflate, followed by Friedel−Crafts
alkylation of the solvent, was postulated as the mechanistic
pathway.
Figure 4. 2-Benzyloxy-1-methylpyridinium tetrakis[3,5-bis-
(trifluoromethyl)phenyl]borate (BnOPB), the reactive and MW-
absorbing solute employed in these experiments.
Considering the ionic reagent and nonpolar solvents
employed in this study, we imagined possibly accelerating this
reaction process by substrate-selective MW heating in
homogeneous solution. However, the initial mixture is
heterogeneous. Selective MW heating of heterogeneous systems
is common, and in this case the ionic reagent selectively heated
and rapidly decomposed without ever going into solution. We
needed a more soluble version of our benzyl-transfer reagent.
preheated oil bath or a MW reactor. In each case, the mixture of
salt and solvent was warmed with stirring until a homogeneous
solution was obtained before starting the comparative heating
experiments. In this manner we were able to circumvent MW-
specific but unproductive decomposition of the ionic substrate
in the initial heterogeneous mixture and focus on MW-specific
effects on the homogeneous solutions.
7440
dx.doi.org/10.1021/jo501153r | J. Org. Chem. 2014, 79, 7437−7450

The Journal of Organic Chemistry
Article
Preliminary Experiments at Constant Temperature:
Observations and Complications. Our initial experiments
involved heating homogeneous solutions of BnOPB in p-xylene-
d10 at 80 °C in either a MW reactor or an oil bath. We
monitored the reaction progression over time by analyzing small
ient for MW-assisted organic synthesis, particularly when
adapting procedures developed under conventional heating
conditions. However, overreliance on constant-temperature
mode has likely contributed to prior confusion over possible
MW effects.6 It also turns out that the constant-temperature
mode is not ideal for observing MW-specific thermal effects.10
Power and temperature profiles for this experiment are given
in Figure 5 (bottom). The system is subjected to high MW
power at the outset to generate heat and raise the temperature as
rapidly as possible, and then the power is reduced once the
desired temperature is achieved. If reactivity correlates with MW
power in any way, then any MW-specific reactivity would most
likely be observed in the early stages of the reaction, while the
power is high and the temperature is increasing. However, the
power and temperature profiles reveal fluctuations in recorded
temperature, including spikes as high as 85 °C. In response, the
MW reactor automatically adjusts the power; this adjustment
takes the form of a series of progressively smaller pulses of
applied MW power. The exact temperature and power profile is
nearly impossible to replicate across multiple experiments.
Moreover, because of lag time in the thermometric recordings, it
is likely that temperature swings are greater than what is
recorded: i.e., the temperature spike was probably higher than
the recorded temperature of 85 °C.
1
aliquots by H NMR spectroscopy. We chose p-xylene as the
solvent because of its symmetrical structure and convenient
chemical properties (reactivity, boiling point, etc.). Due to its
symmetrical structure, p-xylene is nonpolar and less MW-
absorbing that nonsymmetrically substituted arenes, and it
reacts with phenylcarbenium to produce a single product, which
simplified the analysis.
Our initial observations were intriguing but problematic.
Figure 5 (top) illustrates reaction conversion over time for an
These observations suggested to us that there might be
observable reactivity differences between oil bath and MW
heating in this system, but better control over the MW heating
profile is needed to support any firm conclusions. One can
achieve such control by setting a prescribed MW power profile.
Here, we prescribed a fixed and constant MW power, which
eliminates the potential for rapid and MW-specific temperature
fluctuations. (We have also been examining pulsed MW power
sequences, which perhaps better mimic the power profile in the
early stages of the experiment illustrated in Figure 5, as
described in an accompanying report.10)
MW-Specific Thermal Effects at High MW Power. We
examined the effects of MW heating at high constant power
(300 W) on the thermal reactivity of our system. As described
previously, identical homogeneous solutions of BnOPB in p-
xylene-d10 were vigorously stirred and heated either in the MW
or in a preheated oil bath. The MW reaction system was
designed to minimize heat formation other than by direct
interaction of MW energy with the ionic solute. p-Xylene was
deemed an appropriate solvent, because it is nearly MW
transparent yet suitable for Friedel−Crafts trapping of the
thermally generated phenylcarbenium species. Nothing else was
added to the reaction vessel except for the Teflon-coated stir
bar, which is not a significant source of heat in these
experiments. Importantly, the MW reaction vials were made
of quartz, because borosilicate glass (Pyrex) absorbs MW energy
and does contribute significantly to heat buildup in this system
(Figure 6).32
Conversion over time for the reaction conducted in our MW
reactor at full power (up to 300 W33) is shown in Figure 7. The
reaction reached 86% conversion after 10 min and was
essentially complete within 30 min. We monitored the
temperature using a calibrated external IR pyrometric sensor,
which correlates the surface temperature of the reaction vial with
the solution temperature, as measured by an internal fiber-optic
probe (vide infra). The recorded temperature was 35 °C at the
start of the experiment, because of the prior warming needed to
achieve a homogeneous solution of BnOPB in p-xylene-d10. The
maximum recorded temperature was 97 °C, and the average
Figure 5. (top) Reaction conversion over time for experiments heated
at 80 °C using the MW in constant-temperature mode (pink circles)
and a preheated oil bath (navy triangles); reproduced from ref 26.
(bottom) MW power and recorded temperature as a function of time in
the MW experiment.
experiment in a preheated oil bath (80 °C) and one in a CEM
Discover MW reactor programmed to heat at 80 °C (constant
temperature mode). The MW reactor automatically employs a
feedback loop that varies MW power to establish and maintain
the prescribed temperature. In the MW experiment we saw
evidence of dramatically enhanced reactivity at the early stages
of the reaction, but attempts to repeat this reactivity profile
produced erratic and variable results. We could not draw
concrete conclusions from these experiments.
The reproducibility problems with this MW experiment stem
from using variable MW power to try to achieve a prescribed
reaction temperature. Constant-temperature mode is conven-
7441
dx.doi.org/10.1021/jo501153r | J. Org. Chem. 2014, 79, 7437−7450

The Journal of Organic Chemistry
Article
measured temperature in the MW experiment, and that MW
energy was the only other variable, we concluded that MW
energy played a role in promoting thermal reactivity beyond
what would be expected on the basis of the system temperature;
this reaction system is MW-actuated.26,35 The distinction
between thermal energy (heat) of individual solute molecules
and recorded temperature of the bulk solution is a critical aspect
of this conclusion.
Control Experiment 1. Indirect MW heating through a
MW-absorbing jacket suppressed the observed effects. We
suspended our quartz reaction vial in a larger quartz flask and
immersed the vial in a bath of propylene glycol (ca. 80 mL). We
then heated this “shielded” system at 100 °C using either a
preheated oil bath or our CEM Discover MW reactor (constant-
temperature mode), again using a calibrated external IR sensor
to monitor internal solution temperature. The use of the highly
MW absorbing propylene glycol bath resulted in efficient bulk
heating using low and variable MW power,36 and this control
experiment also produced reaction conversions that tracked well
with oil bath heating (Figure 8). The MW-specific rate
enhancement is thus effectively suppressed by the propylene
glycol bath.
Figure 6. Heating curves for neat p-xylene in borosilicate (Pyrex, garnet
line) and quartz (gold line), showing that Pyrex can contribute
significantly to MW heating. Vials were kept open to the atmosphere
(open vessel) while the sample was stirred and heated using 300 W of
applied MW power.
Figure 7. Reaction conversion over time for experiments heated in the
MW at full power (300 W, gold squares) and in a 100 °C preheated oil
bath (red diamonds). The recorded MW temperature profile is also
plotted (black line).
Figure 8. Reaction conversion over time for experiments heated at 100
°C in the MW in constant-temperature mode (pink squares) and a
preheated oil bath (navy blue diamonds) through a bath of propylene
glycol (cf. inset cartoon), reproduced from ref 26.
over the course of the 30 min reaction was 86 °C. Temperature
changes occurred gradually and predictably,34 without the
fluctuations or spikes that were observed using variable MW
power.
It is practically impossible to replicate this temperature profile
using convective heating. Therefore, for comparative purposes,
we elected to conduct an analogous experiment at a constant but
higher temperature under convective heating. We reasoned that if
we observe lower conversion at higher temperature under
convective heating, then we could qualitatively attribute the
higher conversion in the MW experiment to MW-specific
effects.
The propylene glycol bath absorbs MW radiation and in turn
heats the solution convectively. Similar control experiments
have been reported using silicon carbide (SiC) vials,37 although
technically SiC vials neither exclude MW radiation nor prevent
MW dielectric heating (cf. Figure 9), particularly in systems
involving ionic MW-absorbing species: “the SiC tubeat
comparatively low temperaturesdoes not completely shield
the contained solvents from the electric field.”38 Our propylene
glycol design may be superior to SiC for this type of control
experiment.39
For this qualitative comparison, an otherwise-identical
reaction system was immersed in an oil bath that had been
preheated to 102 °C. (We separately examined the heating
profile of this reaction system, including the same vial and oil
bath, and found that the internal solution temperature reached
100 °C within 3 min.) Conversion over time for this reaction is
also plotted in Figure 7. In contrast to the MW experiment, we
observed a reaction conversion in the oil bath of only 12% after
10 min and 25% after 30 min. Considering that the measured
temperature of the oil bath experiment was higher than the
This control experiment supported our conclusion that MW
power is an important variable in promoting Friedel−Crafts
benzylation of xylene using BnOPB, because shielding the
reaction medium from MW radiation with the propylene glycol
bath resulted in regression of reactivity back to what is seen with
conventional convective heating in an oil bath. As noted
previously, heating these reaction mixtures in constant-temper-
ature mode using relatively low applied MW power produced no
discernible MW-specific effects, in contrast to experiments
designed and conducted in accord with our central hypothesis.
7442
dx.doi.org/10.1021/jo501153r | J. Org. Chem. 2014, 79, 7437−7450

The Journal of Organic Chemistry
Article
Figure 9. tan δ as a function of temperature for acetonitrile (purple),
silicon carbide (red), ethanol (blue), and an ionic liquid (bmimPF6,
green); reproduced from ref 38. The dielectric loss tangent (tan δ)
quantifies the ability of a substance to convert MW energy into heat. In
the temperature range relevant to our studies (≤100 °C), silicon
carbide reportedly absorbs MW energy and generates heat more
efficiently than most common organic solvents but less efficiently than
ethanol or a representative ionic liquid. p-Xylene (not shown) has a tan
δ value of close to 0 at room temperature, and this should not change at
higher temperatures. Reproduced with permission from the Royal
Society of Chemistry.
Kinetics and Magnitude of the Effect. The next phase of
our study was to characterize MW-specific thermal effects in
terms of deviation from the Arrhenius-based link between
measured temperature and kinetic reactivity. The Arrhenius
Figure 10. (top) First-order kinetic plots for the thermal benzylation of
p-xylene at 103.4, 94.9, and 85.7 °C. (bottom) Arrhenius plot of the
rate and temperature data. See the Supporting Information for larger
images and tabular data.
equationk = Ae−E /RTuses solution temperature as a
a
measure of kinetic molecular energy of the reactive species;
selective MW heating that perturbs thermal equilibrium would
result in reactivity that is inconsistent with temperature-based
Arrhenius predictions. It is important to note that any such
deviations would be attributable to molecular-level thermal
effects that cannot be captured by bulk temperature measure-
ments (vide infra for additional discussion and alternative
treatments of our experimental data).
We measured first-order rate constants for reactions in p-
xylene at three different temperaturesk = 0.369 × 10−4 s−1 at
85.7 °C, 1.09 × 10−4 s−1 at 94.9 °C, and 2.70 × 10−4 s−1 at 103.4
°C (Figure 10, left)and plotted the natural log of these
valuesln kagainst 1/RT to determine the Arrhenius
parameters (Figure 10, right). Reaction mixtures were heated
in an insulated oil bath, and reaction temperatures were
monitored using an internal thermometer. The activation
energy Ea = 30.2 kcal/mol and the pre-exponential term A =
9.64 × 1013 s−1 for first-order thermal benzylation of p-xylene
(solvent) using BnOPB were thus determined.
roughly 17x higher than in the oil bath experiment. This MW
reaction rate constant corresponds to a temperature of 128 °C
(401 K), 48 °C higher than the average recorded temperature
and 31 °C higher than the maximum recorded temperature.
The calculated temperature value of 128 °C in this crude
analysis represents the “effective temperature” (Teff) of the MW-
absorbing solute as it plays the role of both molecular radiator
and reactant. That is, the average thermal molecular energy of
the solute corresponds to it being heated at 128 °C. The
discrepancy between the bulk solution temperature and the
observed chemical reactivity can be explained by selective MW
heating of the polar solute, with subsequent chemical trans-
formations that occur faster than thermal equilibration of the
solute with the bulk solution (e.g., selective heating and thermal
reactivity occurring according to the “transient solvent cage
domain model” as described in the Introduction).41
The external IR temperature sensor does a reasonable job of
capturing the internal solution temperature in this system
(Figure 11). Subsequent to the experiments just described, we
acquired a Neoptix Reflex internal FO probe, calibrated it
between 150 °C and room temperature with a NIST-traceable
thermocouple, and used it to monitor solution temperature
directly in conjunction with the IR sensor. We heated our
reaction mixture at 300W while recording the solution
temperature simultaneously with the external IR sensor and
the internal FO probe. As can be seen in Figure 11, the IR sensor
data tracked closely with the FO probe data while the MW
power was on and before the solution started to boil. Close
examination of the tabular data (see Supporting Information)
reveals a 5−10 s delay between the internal FO probe and the
external IR sensor as the solution is heating up to around 100
°C. Once the solvent starts to evaporate and boil, the differences
Returning to the data presented in Figure 7,40 we can crudely
quantify the observed MW-specific effects. These calculations
roughly illustrate the magnitude of the impact of MW heating
beyond raising the bulk solution temperature. The MW-heated
reaction reached 86% conversion within 10 min, during which
time the recorded temperature had increased from 35 to 97 °C
(80 °C average). In contrast, the reaction in a preheated oil bath
(102 °C) had only reached 12% conversion.
On the basis of the Arrhenius parameters measured for p-
xylene, heating this reaction mixture at 100 °C for 10 min should
produce 11% conversion with a rate constant k = 1.89 × 10−4
s−1, which is in line with the 12% conversion observed in the oil
bath experiment. The 86% conversion observed in the MW
experiment corresponds to a rate constant k = 33 × 10−4 s−1,
7443
dx.doi.org/10.1021/jo501153r | J. Org. Chem. 2014, 79, 7437−7450

The Journal of Organic Chemistry
Article
Figure 12. Temperature profile for heating 7.5 mg of BnOPB in the
MW at 300W power with external cooling, as recorded by the internal
FO probe. A 35% conversion was observed at the end of this 5 min
experiment, whereas 5.2% conversion is expected based on integrated
Arrhenius calculations.
Figure 11. Temperature profile for a reaction mixture heated in the
MW at 300W as recorded by an external IR sensor (gold trace) and an
internal FO probe (garnet trace). The IR sensor tracks reasonably well
with the FO probe while the solution is heating, indicating that our IR
temperature data is reliable. Interestingly, the IR sensor then
underestimates the solution temperature as it boils and markedly
overestimates the temperature as it cools. These discrepancies likely
reflect dynamic heat-transfer events for which the IR sensor was not
calibrated, but they are not relevant to the studies here.
the expected reaction conversion after 5 min was 5.2%. This
expected level of conversion would be also achieved by heating
at a temperature of 99.8 °C to achieve the rate constant k = 1.8
× 10−4 s−1. We refer to the constant temperature (99.8 °C)
needed to produce the same expected conversion as what we get
by manually integrating the Arrhenius equation the variable-
temperature profile as the “integrated average temperature”. It is
generally slightly higher than the time-averaged temperature (in
this case, 99.8 vs 94.9 °C).
become more pronounced. However, these deviations are not
relevant to the studies at hand. They probably reflect heat-
transfer events that occur during solvent reflux and cooling for
which the IR sensor was not calibrated. Also evident in Figure 11
is that our p-xylene system heated faster than we had observed
previously in p-xylene-d10.26,40 Therefore, we reoptimized the
system for documenting MW-specific thermal effects on BnOPB
in using an internal probe.
We heated solutions of BnOPB in p-xylene with constant MW
power (300W) and recorded solution temperature continuously
using the FO probe. Bulk heating here was partially offset by
cooling the outside of the reaction vessel using a stream of cold
air.42 Thus, the internal solution temperature could be
moderated while applying high constant MW power of 300W.
To reduce the amount of bulk heat being produced, we also
reduced the amount of MW-absorbing substrate in each
experiment. A reaction mixture comprising 7.5 mg of BnOPB
in 2 mL of p-xylene in an open quartz vial was heated using
300W of applied MW power for 5 min while the outside of the
vial was actively cooled by a stream of cold compressed air. The
internal bulk solution temperature was recorded using the
internal FO probe and plotted as a function of time in Figure 12.
The highest recorded solution temperature was 105 °C, and the
average was 94.9 °C. As before, the maximum temperature was
recorded early in the experiment, and temperature decreased as
the reaction progressed. Reaction conversion as determined by
1H NMR spectroscopy was 35% at t = 5 min.
The 6.7-fold difference in observed (35%) vs calculated
(5.2%) reaction conversion is again attributed to selective MW
heating of the strongly absorbing reactant in an otherwise
nonabsorbing system. We assume that the magnitude of
selective MW heating is not constant during this 5 min window
as the bulk temperature increases, but we can quantify the time-
averaged MW effect over the 5 min course of the reaction. In this
case, to achieve a 35% conversion over 5 min requires substrate
thermal molecular energy that corresponds to an effective
temperature of 120 °C. This effective temperature is 15 °C
higher than the maximum recorded internal temperature and 20
°C higher than the integrated average solution temperature.
To put this in terms of rates, the hypothetical rate constant
that corresponds with this reaction reaching 35% conversion in
5 min is k = 14 × 10−4 s−1, whereas only 5.2% conversion would
be predicted on the basis of the measured temperature profile
for k = 1.8 × 10−4 s−1. The observed increase in conversion over
the Arrhenius temperature based prediction for the first 5 min
thus corresponds to a 7.8-fold increase in reaction rate.
Our discussion of expected conversion, hypothetical rate
constants, and effective temperature is meant as a guide for
understanding the relative magnitude of selective MW heating
in this system. This reaction design is unusual and extreme: an
ionic reagent in a nonpolar solvent at high dilution undergoing
unimolecular rate-determining thermolysis. Moreover, our
analysis focuses on the earliest stages of the reaction, during
which time the net temperature change is sharply positive. It is
unclear how easily these findings will translate to more
synthetically relevant systems, and how advantageous such
effects can be in organic synthesis. However, identifying these
effects is the critical first step toward addressing those questions,
and in this study we positively identify MW-specific thermal
effects.
The expected reaction conversion based on the bulk solution
temperature profile can be predicted using our Arrhenius
parameters. We manually integrated the Arrhenius parameters
over the course of the reaction. We calculated the instantaneous
rate for each time the temperature was recorded (at 1.0 s
intervals), calculated the expected reaction conversion during
the time that elapsed before the next temperature reading, and
then summed these conversions over the 5 min reaction
window. On the basis of the Arrhenius parameters and the
temperature profile, this time captured by an internal FO probe,
7444
dx.doi.org/10.1021/jo501153r | J. Org. Chem. 2014, 79, 7437−7450

The Journal of Organic Chemistry
Article
Figure 13. Mathematical treatments of the Arrhenius equation to put observed MW-specific thermal effects in terms of Arrhenius parameters. The
MW reaction rate constant (k = 14.4 × 10−4 s−1) and integrated average temperature (T = 99.8 °C) do not equate for the Arrhenius parameters of this
reaction (A = 9.64 × 1013 s−1 and Ea = 30.2 kcal/mol). Our interpretation is that MW-specific thermal effects perturb thermal equilibrium between
solute and solvent, such that the measured solution temperature does not accurately reflect the solute kinetic molecular energy (option 1). The
effective temperature of the reactant was thus calculated to be Teff = 393 K (120 °C). However, one can alternatively quantify the observed MW rate
enhancement in terms of impacts on the pre-exponential term (A, option 2) or the activation energy (Ea, option 3), as shown. They are all part of the
same equation.
Control Experiment 2. We tested the effect of adding
crushed borosilicate glass chips to the stirred reaction mixture.
Our hypothesis was that heterogeneous additives can suppress
MW-specific rate enhancements, and this turns out to be the
case. The MW-specific rate enhancement that was observed in
the homogeneous system was not reproduced in this
heterogeneous system.
“nonthermal effect”, although it can be described as an extra-
temperature effect: the measurable bulk solution temperature
only captures a portion of the thermal energy available to the
solute. Under the perturbation of MW energy, the average
kinetic molecular energy of the absorbing solute is higher than
that of the nonabsorbing solvent. Thus, the recorded temper-
ature underrepresents the thermal energy of the ionic solute. In
this example experiment, the effective temperature (average
kinetic molecular energy) of the solute was Teff = 120 °C and the
integrated averaged temperature for the solution was T = 99.8
°C (cf. Figure 3).
One of the important next steps will be to examine
systematically the role of variables, including solute concen-
tration, solvent viscosity and heat capacity,40 MW power, and
perhaps frequency of the applied field (in consideration of the
frequency dependence of tan δ); such efforts are underway.10
Another may be to help develop theoretical models to account
for the influence of MW energy on specific solutes. Any such
future model would likely describe a difference in average kinetic
molecular energy between the absorbing solute and non-
absorbing solvent. While the bulk temperature reflects the
overall weighted average of kinetic molecular energies, solute-
specific perturbations in thermal energy should correspond to
solute-specific enhancements in chemical reactivity.
Alternative Explanations. Our discussion of effective
temperature within the context of the Arrhenius equation is
consistent with the transient solvent cage domain model of
selective MW heating. However, the focus on effective
temperature in the Arrhenius calculations is mathematically
arbitrary, as one could just as easily use the bulk temperature
values as measured and solve the Arrhenius equation for MW-
specific impacts on other values such as activation energy. We
do not favor the idea of there being any actual impact on the
activation energy (Ea) of the reaction, although quantifying
MW-specific thermal effects in terms of an imaginary impact on
Ea may have some merit.
If one takes the position that the measured solution
temperature accurately reflects the thermal energy available to
the substrate, then one can use the measured temperature and
solve the Arrhenius equation for imaginary impacts of MW
energy on either the effective activation energy or the pre-
exponential term. Such treatments are exemplified here (Figure
13) using data from the MW heating experiment monitored
with an internal FO probe, in which we observed 35%
conversion after 5 min at an integrated average temperature of
99.8 °C.
Perhaps surprisingly, the reaction system with added glass
chips heated overall less efficiently than in the absence of this
additive: qualitatively less external cooling was needed to
moderate the internal temperature to an average of 100.3 °C
(maximum recorded temperature of 105.0 °C). Reaction
conversion after 10 min in the presence of glass chips was
16% in the MW experiment, which is not much above the
reaction conversion expected (13%) on the basis of Arrhenius
calculations integrated over the recorded temperature profile.43
For comparison, an identical sample was heated in an oil bath at
100 °C for 10 min, which produced 12% conversion. This 12%
conversion is also consistent with Arrhenius-based expectations;
the calculated conversion after 10 min at 100 °C is 11%.
Borosilicate glass absorbs MW energy and helps heat the system
convectively (cf. Figure 6). These observations suggest that glass
chips also scatter and/or reflect applied MW radiation. Additives
that absorb and/or reflect MW energy are likely to disfavor
selective MW heating of the solute, which explains why the glass
chips attenuate the rate enhancements observed in the
homogeneous system.
Our Interpretation. We have evolving ideas as to how to
rationalize selective MW heating in homogeneous solution,
while recognizing that ours is an empirical study aimed at
validating selective MW heating. Drawing on ideas from the
MW-assisted organic synthesis (MAOS) literature,19 we first
conceptualized our observations in terms of enhanced transla-
tional and rotational kinetic molecular energy of the solute
leading to changes in the frequency and intensity of molecular
collisions needed to achieve reactivity.26 We later refined our
working hypothesis to focus on collisions and other molecular
events that occur within transient solvent cage domains.35b
We currently characterize the observed rate increase as a
“MW-specific thermal effect” that arises from perturbing thermal
equilibrium within a macroscopically homogeneous system. As
such, the measured temperature (an Arrhenius parameter) does
not adequately capture the thermal energy of the substrate of
interest, again underscoring the nonequivalence of heat and
temperature. Importantly for the larger discussion on MW
effects, we do not interpret this to be an example of a
7445
dx.doi.org/10.1021/jo501153r | J. Org. Chem. 2014, 79, 7437−7450

The Journal of Organic Chemistry
Article
The actual Arrhenius parameters for this reaction as given
above are Ea = 30.2 kcal/mol and A = 9.64 × 1013 s−1. However,
one can take the observed reaction conversion of 35%
corresponding to a rate constant k = 14.4 × 10−4 s−1and the
integrated average (measured) temperature of 99.8 °C and solve
for “imaginary” or “effective” MW impacts on these parameters.
Assuming that Ea is unchanged, the new pre-exponential under
MW heating would be A = 1.70 × 1015 s−1. Alternatively
assuming an unaltered pre-exponential, the imaginary impact on
activation energy would be ΔEa = −1.5 kcal/mol; the observed
rate acceleration is as if the activation energy were lowered to Ea
= 28.7 kcal/mol. One could also postulate partial impacts on
both A and Ea simultaneously, or even spread across all three
factors (A, Ea, and T), to rationalize the observed rate
enhancement. Our current thinking is that the most reasonable
explanation is selective MW heating of the solute to a higher
effective temperature than the rest of the solution. However,
these effects cannot be rigorously quantified in Arrhenius terms
because the system is not thermally homogeneous, and
Arrhenius calculations do not account for thermal gradients in
the reaction mixture.
Temperature Measurements. One of the fundamental
challenges of studying MW effects is that of accurately
measuring the solution temperature. Internal fiber-optic (FO)
temperature probes and external temperature measurement
using IR sensors offer complementary approaches to monitoring
system temperature, each with advantages and limitations. Fast-
responding internal FO probes are critical for optimal
temperature readings when highly MW absorbing systems are
being heated using variable MW power. However, an external
sensor can be appropriate when heating a poorly MW absorbing
system at constant MW power, which produces only gradual and
predictable changes in temperature, and a noninvasive external
sensor may even be preferred in some cases.44
The experimental caveats of using external IR sensors have
been thoroughly documented elsewhere.6 Here we discuss two
potential confounding issues with internal FO probes, and we
discuss how they were addressed in this study. The first problem
is that careless use of the probe can lead to erroneous
temperature readings if the tip of the probe rests against the
walls of the flask, which are not necessarily thermally
equilibrated with the solution.45 We took care to suspend and
secure the tip of the fiber-optic probe close to the middle of the
stirred solution but outside of the stirring vortex. The second
problem is more of a fundamental disadvantage: it is impossible
to probe a system without perturbing it. For example, fiber optic
probe attachments can restrict or alter heat flow from the system
to the surroundings depending on how the probe is secured in
the solution. On the other hand, with the internal FO probe one
can also use external cooling to accelerate release of bulk heat to
the surroundings.42 The external cooling mechanism would
compromise external temperature readings, but external cooling
does not confound internal solution temperature readings any
more so than does external heating. We employed both external
and internal sensors to collect our thermometric data, as there is
value in validating observations using both types of temperature
measurements. Both have a place in MW chemistry, although
one can always argue that the limitations of one and/or the
other may make it difficult to interpret the observed results.
Reflux Experiments. As noted in the previous section,
underlying complications inherent to MW heating experiments
make it difficult to establish the validity of any thermometric
data. We have used both accepted forms of MW temperature
measurementexternal IR sensors and internal FO probesto
record the temperature profile of experiments under MW
heating, and we have benchmarked our observations in these
systems against comparable experiments heated conventionally.
Differences in reactivity have been attributed to MW-specific
thermal effects: namely, selective MW heating of the absorbing
solute. However, one cannot perfectly replicate the bulk
temperature profiles for MW experiments using conventional
heating, and we do not presume that thermometric data can
capture all thermal events under MW heating. This is especially
true in our experiments under constant and high MW power, in
which we postulate that the thermal equilibrium is perturbed in
a way that cannot be achieved with convective heating. Such
thermal perturbations cannot possibly be captured thermo-
metrically.
We therefore pursued a third avenue for comparing
experiments under MW and convective oil bath heating in
which solvent reflux at atmospheric pressure was used to
moderate the thermal energy of the system. Thermometers are
ultimately calibrated against physical properties of substances;
for example, the melting and boiling points of water define the
Celsius scale. Identical solutions in identical systems being
heated at reflux under identical conditions must produce
identical results. We conducted a series of parallel experiments
in which identical solutions were heated at reflux under identical
conditions, except that one was heated using MW energy and
the other using a preheated oil bath. Using the boiling point of
our specific solution obviates the need to record the
temperature. In the absence of MW-specific effects, the reaction
outcomes should be identical, and any differences in reactivity
between the experiments would be attributed to MW-specific
effects.
As described previously,26 reflux experiments were conducted
in toluene instead of xylene because the boiling point of toluene
is more appropriate for monitoring these reactions over time.
Toluene is not an ideal solvent for probing MW-specific thermal
effects because it absorbs MW energy and produces heat more
efficiently than xylene. However, toluene turned out to be good
enough to support selective MW heating under certain
circumstances. The MW-specific effects were attenuated in
toluene in comparison to xylene, but they remained observable.
We reported parallel reflux experiments in which MW heating
resulted in 67% conversion after 30 min and oil bath heating
produced only 36% conversion after 30 min (Scheme 3).26,46
The difference in conversion here is less than what we observed
in the more MW transparent xylene solution, but it is significant
nonetheless.
Scheme 3. MW-Specific Acceleration of Friedel−Crafts
Benzylation of Refluxing Toluene
Having established a MW effect on this experiment, we then
examined our reflux system for thermometric evidence of
solvent superheating47 and/or nucleation-limited boiling
(“superboiling”),48 in which MW heating produces a sustained
but erratic reflux above the conventional boiling point of the
solvent. Solvent superheating is well established under both
7446
dx.doi.org/10.1021/jo501153r | J. Org. Chem. 2014, 79, 7437−7450

The Journal of Organic Chemistry
Article
MW and conventional heating, whereas superboiling is MW-
specific. Neither was observed in our stirred systems to any
significant degree, although we can document both superheating
and MW-specific superboiling in an unstirred system (Figure
14).35b
reactivity. Temperature (intensive property) measures the
degree of thermal energy in a system at thermal equilibrium.
MW heating can perturb thermal equilibrium between
absorbing and nonabsorbing components of a system in such
a way that the measured temperature cannot accurately capture
the thermal energy of the absorbing components.
A graphical summary of representative experiments described
over the course of these studies is provided in Figure 15. We
Figure 15. Graphical summary of temperature data and calculations (in
°C) from representative experiments reported herein. The measured
average temperature T(ave) is represented in the front row of blue
columns, and the calculated effective temperature T(eff) is represented
in the back row of red columns. T(ave) and T(eff) roughly coincided in
the following experiments: oil bath, oil bath with boiling chips, MW
with propylene glycol (PG) bath, and MW with boiling chips.
Measured T(ave) and calculated T(eff) values were markedly different
for MW heating of homogeneous solutions at high power, whether the
temperature was monitored with an external IR sensor or an internal
FO probe. The maximum recorded temperature T(max) in specific
MW experiments was as follows: (a) T(max) = 97 °C (in p-xylene-d10);
(b) T(max) = 111 °C; (c) T(max) = 105 °C; (d) T(max) = 105 °C.
Figure 14. Plots of temperature over time for MW heating of our
reaction system under three sets of conditions; reproduced from ref
35b. The top (green) line shows initial superheating followed by
sustained superboiling, presumably due to the lack of proper nucleation
sites in the absence of stirring. (This is the only experiment in which we
did not vigorously stir the solution.) The lower two lines show the
temperature profiles of a properly stirred system. The black line shows
the temperature of the homogeneous system that is the focus of our
study, and the red line shows the temperature of the system with
boiling chips added.
conducted a series of thermal benzyl-transfer experiments in
which macroscopically homogeneous and well-stirred reaction
mixtures were subjected to either MW heating or convective (oil
bath) heating to promote thermal reactivity. In cases where the
ionic substrate (BnOPB) is uniquely capable of interacting with
the applied MW field, clear differences in reaction rate were
observed. The exact magnitude of the rate enhancement varied
for different experimental setups, but MW-specific rate enhance-
ments were predictably observed under our prescribed conditions. In
cases where MW-absorbing and/or -reflecting species were
introduced into the system, the MW-specific effects on BnOPB
reactivity were effectively and reproducibly blocked. These
observations support our central hypothesis that the chemical
reactivity of a MW-absorbing solute in an otherwise-non-
absorbing system can be MW-actuated: accelerated using MW
energy to rates that cannot be duplicated by conventional
heating to similar reaction temperatures.
We interpret and rationalize these observations in terms of
selective MW heating of the solute within transient solvent cage
domains. Thermal activation of the absorbing solute occurs in
solution within transient solvent cages. In contrast, the
nonabsorbing solvent can only be heated indirectly by collisions
with the solute. The solute must be hotter (i.e., at a higher
effective temperature) than the solvent in order to transfer heat
to the solvent.
In the absence of stirring, we measured solvent superheating
to temperatures as high as 140 °C before nucleation began,
followed by rapid and chaotic release of gas and heat.47 Once the
turbulence of the initial vaporization events subsided (within the
first 2−5 min), the recorded temperatures fluctuated around
126 °C for the duration of the analysis. This is consistent with
nucleation-limited superboiling, which is a consequence of the
lack of nucleation sites in the bulk solution (where heat is being
generated) limiting the rate at which phase change from liquid
to gas can occur.48 Superboiling does not occur under
conventional heating because heat is introduced through the
walls of the flask, where there are ample nucleation sites. It has
been reported that boiling stones and/or proper stirring reduce
or eliminate both types of superheating effects,48c and this is
what we see as well. The recorded reflux temperature of our
stirred solution was 111.8 °C, and a stirred solution with boiling
chips boiled at 110.8 °C under our laboratory atmospheric
conditions. In the latter case the heterogeneous boiling chips
provide additional nucleation sites and may perturb the system
in other ways as well (cf. Control Experiment 2). On the basis of
these experiments, we rule out any significant contributions of
bulk solvent superheating, in line with our conclusions of MW-
specific effects related to selective heating of the solute.
CONCLUSION
The focus of this work was to establish the possibility of
selective MW heating in macroscopically homogeneous
solutions. The current reaction system is unusual in that it
involves an ionic substrate dissolved at ultrahigh dilution in a
nonpolar solvent, which doubles as the cation scavenger to track
reactivity of the salt. The practical synthetic applications of this
particular reaction are limited. The value of this work stems
■
The thermal Friedel−Crafts benzylation of toluene and xylene
using BnOPB can be MW-actuated, in that MW energy can
promote reactivity beyond temperature-based expectations.
Central to understanding MW-specific thermal effects is the
realization that heat and temperature are different. Thermal
energy (heat, an extensive property) promotes chemical
7447
dx.doi.org/10.1021/jo501153r | J. Org. Chem. 2014, 79, 7437−7450

The Journal of Organic Chemistry
Article
with a condenser and submerged in a preheated oil bath (125 or 145
°C). Timing of the experiment began when the solvent began to reflux.
The round-bottomed flask was removed from the oil bath after the
prescribed reaction time, and a three-drop aliquot was removed (using
a 9 in. disposable borosilicate glass Pasteur pipet) as soon as the reflux
line fell below the reflux condenser joint.
from the fact that it establishes parameters under which selective
MW heating is possible in homogeneous solution. Practically
every conceivable condensed-phase chemical reaction system is
heterogeneous at the molecular level; thus, the potential
implications here cover a wide range of reaction systems.
Experiments described in this paper support the conclusion that
the conversion of MW energy into heat within transient solvent
cage domains can be faster than propagation of thermal energy
through the bulk solution by convective heat transfer
mechanisms. Thus, the thermal equilibrium is perturbed,
resulting in observably enhanced reactivity of the absorbing
solute. Our results can support a better understanding of
previous reports of unusual MW effects in organic synthesis, and
they provide a blueprint for future examination of other systems
in which selective MW heating may be possible.
General Procedure for the Benzylation of Toluene in
Refluxing Toluene in a Microwave Reactor. 2-Benzyloxy-1-
methylpyridinium tetrakis(3,5-bis(trifluoromethyl)phenyl)borate
(BnOPB; 150.0 mg, 0.1410 mmol) was placed in a clean, dry, quartz
round-bottom flask with an elongated neck equipped with a stir bar. To
the salt was added 30 mL of dry toluene via a graduated 60 mL syringe.
The flask was then gently heated with a heat gun with stirring until the
solution became homogeneous. Once the solid was dissolved, a three-
drop aliquot was removed (using a 9 in. disposable borosilicate glass
Pasteur pipet) to ensure no reactivity prior to subjection to the reaction
conditions. The quartz round-bottom flask was then placed in the
microwave reactor, followed by placement of the open vessel
microwave attenuator. The flask was fitted with a condenser, followed
by subjection to microwave irradiation (300 W). Timing of the
experiment began when the solvent began to reflux (visible from the
top of the instrument though the open-vessel attenuator). After
reaction for the prescribed interval, the microwave irradiation was
ceased. A three-drop aliquot was removed (using a 9 in. disposable
borosilicate glass Pasteur pipet) as soon as the reflux line fell below the
reflux condenser joint.
EXPERIMENTAL SECTION
■
General Experimental Methods. Commercially available starting
materials, reagents, and solvents were used without further purification
unless otherwise noted. NMR spectra were obtained using one of
several solvents including DMSO-d6, acetone-d6, benzene-d6, and
toluene-d8. All collected aliquots were stored in small vials with
phenolic or Teflon-lined caps until prepared for and analyzed by NMR
spectroscopy.
General Procedure for the Benzylation of p-Xylene in a
Propylene Glycol Bath within a Microwave Reactor. 2-
Benzyloxy-1-methylpyridinium tetrakis(3,5-bis(trifluoromethyl)-
phenyl)borate (BnOPB; 10.0 mg, 0.0094 mmol) was placed in a
clean, dry, quartz microwave reaction tube equipped with an oval-
shaped stir bar. To the salt was added 2.0 mL of p-xylene via syringe (2
× 1.0 mL aliquots). The vial was then gently heated with a heat gun
with stirring until the solution became homogeneous. Once the solid
was dissolved, a three-drop aliquot was removed (using a 9 in.
disposable borosilicate glass Pasteur pipet) to ensure no reactivity prior
to subjection to the reaction conditions. The tube was suspended in a
custom-made quartz round-bottom flask filled with propylene glycol
and equipped with a stir bar using copper wire (see the Supporting
Information for photographs of this experimental setup). The round-
bottom flask was then placed into a CEM Discover Benchmate
microwave reactor. The prescribed reaction conditions were pro-
grammed into the instrument using Synergy software as an open-vessel
reaction, allowing for timed aliquots to be removed without
interruption of the experiment. The reaction was started, and aliquots
were removed at desired intervals by inserting a glass pipet
momentarily to remove the correct approximate amount needed.
Excess solution was quickly transferred back into the reaction vessel.
General Procedure for the Benzylation of p-Xylene in a
Large Oil Bath for Collection of Data for Kinetic Plotting. 2-
Benzyloxy-1-methylpyridinium tetrakis(3,5-bis(trifluoromethyl)-
phenyl)borate (BnOPB; 13.5 mg, 0.0127 mmol) was placed in a
small, clean, dry borosilicate vial equipped with a small cylindrical stir
bar. To the salt was added 3.0 mL of p-xylene via syringe (3 × 1.0 mL
aliquots). A heat gun was used to heat the sample as quickly as possible
to the approximate oil bath temperature, as monitored by an internal
fiber optic probe, and then the vial was submerged in a large, insulated,
and preheated oil bath. Three-drop aliquots were removed at
prescribed intervals. The temperature was noted by a fiber optic
probe at each collection interval, with the average taken for use in
kinetic calculations.
General Procedure for Benzylation of p-Xylene in an Oil
Bath. 2-Benzyloxy-1-methylpyridinium tetrakis(3,5-bis(trifluoro-
methyl)phenyl)borate (BnOPB; 10.0 mg, 0.0094 mmol or 7.5 mg,
0.0071 mmol) was placed in a clean, dry ∼6 mL lime glass test tube
equipped with a triangular-shaped or small cylindrical stir bar. Any
additives (glass boiling chips etc.) were added at this time, prior to
addition of the solvent. To the salt was added 2.0 mL of p-xylene via
syringe (2 × 1.0 mL aliquots). The tube was then gently heated with a
heat gun with stirring until the solution became homogeneous. Once
the solid was dissolved, a three-drop aliquot was removed (using a 9 in.
disposable borosilicate glass Pasteur pipet) to ensure no reactivity prior
to subjection to the reaction conditions. The test tube was then
submerged in an oil bath preheated to 2 °C above the desired reaction
temperaturethe time of reaction began at this point. Three-drop
aliquots were removed at the desired intervals according to the
prescribed conditions, returning any excess solution to the reaction test
tube.49
General Procedure for the Benzylation of p-Xylene in a
Microwave Reactor. 2-Benzyloxy-1-methylpyridinium tetrakis(3,5-
bis(trifluoromethyl)phenyl)borate (BnOPB; 10.0 mg, 0.0094 mmol or
7.5 mg, 0.0071 mmol) was placed in a clean, dry, quartz 10 mL
microwave reaction tube equipped with a triangular-shaped or small
cylindrical stir bar. To the salt was added 2.0 mL of p-xylene via syringe
(2 × 1.0 mL aliquots). The tube was then gently heated with a heat gun
with stirring until the solution became homogeneous. Once the solid
was dissolved, a three-drop aliquot was removed (using a 9 in.
disposable borosilicate glass Pasteur pipet) to ensure no reactivity prior
to subjection to the reaction conditions. The reaction tube was then
placed into a CEM Discover Benchmate microwave reactor. The
prescribed reaction conditions were programmed into the instrument
using Synergy software as an open-vessel reaction, allowing for timed
aliquots to be removed without interruption of the experiment. If
external cooling was used, then the pressure of compressed air blown
on the outside of the reaction vial was controlled by hand. The reaction
was started, and aliquots were removed at desired intervals by inserting
a glass pipet momentarily to remove the correct approximate amount
needed. Excess solution was quickly transferred back into the reaction
vessel.
Procedure for the Synthesis of 2-Benzyloxy-1-methylpyr-
idinium Tetrakis[3,5-bis(trifluoromethyl)phenyl]borate
(BnOPB). Sodium tetrakis[3,5-bis(trifluoromethyl)phenyl]borate (Na-
BArF; 2.29 g, 2.58 mmol) was placed in a clean, dry 100 mL round-
bottom flask equipped with a stir bar and flushed with nitrogen.
Separately, 2-benzyloxy-1-methylpyridinium triflate (BnOPT; 0.900 g,
2.58 mmol) was placed in a clean, dry 250 mL round-bottom flask
equipped with a stir bar and flushed with nitrogen. Tetrahydrofuran was
added via syringe to each flask (∼50 mL per flask). Each flask was
General Procedure for the Benzylation of Toluene in
Refluxing Toluene in an Oil Bath. 2-Benzyloxy-1-methylpyridinium
tetrakis(3,5-bis(trifluoromethyl)phenyl)borate (BnOPB, ;150.0 mg,
0.1410 mmol) was placed in a clean, dry, quartz round-bottom flask
with an elongated neck equipped with a stir bar. To the salt was added
30 mL of dry toluene via a graduated 60 mL syringe. The flask was fitted
7448
dx.doi.org/10.1021/jo501153r | J. Org. Chem. 2014, 79, 7437−7450

The Journal of Organic Chemistry
Article
heated in an oil bath set at 40 °C and stirred. THF was gradually added
via syringe until no solid particles remained in either flask. Once the
reagents were dissolved, both flasks were opened, and the NaBArF−
THF solution was quickly poured into the BnOPT−THF solution,
using a magnet to prevent transfer of the stir bar. The solution was
again capped and stirred under nitrogen overnight, after which time the
THF was removed under reduced pressure to yield a clear yellow oil.
α,α,α-Trifluorotoluene (PhCF3) was added to make a clear, yellow
solution with a precipitate. The precipitate (presumably NaOTf) was
removed by filtration. BnOPB was crystallized from the PhCF3 solution
by adding hexane. Once the solution became cloudy, it was placed on
ice. The resulting precipitate was filtered off as a yellow solid. A second
recrystallization from PhCF3/hexanes yielded a fine, shiny white solid
in 63% yield (∼1.7 g). 1H NMR (400 MHz, DMSO-d6): δ 4.04 (s, 3H),
5.66 (s, 2H), 7.47−7.53 (m, 3H), 7.60−7.62 (m, 4H), 7.65 (broad s,
8H), 7.76 (broad s, 4H), 7.85−7.87 (d, 1H), 8.53−8.58 (dt, 1H), 8.72−
8.74 (dd, 1H) ppm. 19F NMR (400 MHz, DMSO-d6): δ −64.18 (s,
24F) ppm.
(11) Teske, J. A.; Deiters, A. J. Org. Chem. 2008, 73, 342−345.
(12) (a) Michaut, A.; Boddaert, T.; Coquerel, Y.; Rodriguez, J.
Synthesis 2007, 2867−2871. (b) Coquerel, Y.; Rodriguez, J. Eur. J. Org.
Chem. 2008, 1125−1132. (c) Deck, J. A.; Martin, S. F. Org. Lett. 2010,
12, 2610−2613. (d) Samojłowicz, C.; Borre,
́
E.; Mauduit, M.; Grela, K.
Adv. Synth. Catal. 2011, 353, 1993−2002.
(13) Dallinger, D.; Irfan, M.; Suljanovic, A.; Kappe, C. O. J. Org. Chem.
2010, 75, 5278−5288.
(14) (a) Gerbec, J. A.; Magana, D.; Washington, A.; Strouse, G. F. J.
Am. Chem. Soc. 2005, 127, 15791−15800. (b) Washington, A. L., II;
Strouse, G. F. J. Am. Chem. Soc. 2008, 130, 8916−8922. (c) Gawande,
M. B.; Shelke, S. N.; Zboril, R.; Varma, R. S. Acc. Chem. Res. 2014, 47,
1338−1348.
(15) Tsukahara, Y.; Higashi, A.; Yamauchi, T.; Nakamura, T.; Yasuda,
M.; Baba, A.; Wada, Y. J. Phys. Chem. C 2010, 114, 8965−8970.
(16) (a) van de Kruijs, B. H. P.; Dressen, M. H. C. L.; Meuldijk, J.;
Vekemans, J. A. J. M.; Hulshof, L. A. Org. Biomol. Chem. 2010, 8, 1688−
1694. (b) Dressen, M. H.; Kruijs, B. H. V. D.; Meuldijk, J.; Vekemans, J.
A.; Hulshof, L. A. Org. Process Res. Dev. 2007, 11, 865−869.
(17) Weissermel, K.; Arpe, H.-J. Industrial Organic Chemistry, 2nd ed.,
VCH: New York, 1993; p 14.
ASSOCIATED CONTENT
* Supporting Information
■
S
(18) Hunt, J.; Ferrari, A.; Lita, A.; Crosswhite, M.; Ashley, B.;
Stiegman, A. E. J. Phys. Chem. C 2013, 117, 26871−26880.
(19) Perreux, L.; Loupy, A. Tetrahedron 2001, 57, 9199−9223.
(20) Debye, P. Polar Molecules; Chemical Catalog Co.: New York,
1929.
Figures giving NMR spectra of various crude reaction mixtures,
expanded kinetic plots and data, and plots of temperature vs
time in MW experiments. This material is available free of
charge via the Internet at http://pubs.acs.org.
(21) (a) Kappe, C. O.; Dallinger, D.; Murphree, S. S. Practical
Microwave Synthesis for Organic Chemists. Strategies, Instruments, and
Protocols; Wiley-VCH: Weinheim, Germany, 2009. (b) Microwave
heating as a tool for sustainable chemistry; Leadbeater, N. E., Ed.; CRC
Press: Boca Raton, FL, 2010.
(22) (a) Huang, W.; Richert, R. J. Phys. Chem. B 2008, 112, 9909−
9913. (b) Huang, W.; Richert, R. J. Chem. Phys. 2009, 130, 194509.
(c) Khalife, A.; Pathak, U.; Richert, R. Eur. Phys. J. B 2011, 83, 429−
435.
(23) Reiner, J. E.; Robertson, J. W.; Burden, D. L.; Burden, L. K.;
Balijepalli, A.; Kasianowicz, J. J. J. Am. Chem. Soc. 2013, 135, 3087−
3094.
(24) (a) Kaiser, N. F. K.; Bremberg, U.; Larhed, M.; Moberg, C.;
Hallberg, A. Angew. Chem., Int. Ed. 2000, 39, 3596−3598.
(b) Steinreiber, A.; Stadler, A.; Mayer, S. F.; Faber, K.; Kappe, C. O.
Tetrahedron Lett. 2001, 42, 6283−6286. (c) Rodríguez, A. M.; Prieto,
P.; de la Hoz, A.; Díaz-Ortiz, A.; García, J. I. Org. Biomol. Chem. 2014,
12, 2436−2445.
AUTHOR INFORMATION
Corresponding Author
*E-mail for G.B.D.: gdudley@chem.fsu.edu.
Notes
■
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This research was supported by a COFRS grant from the FSU
Research Foundation. A.E.S. (NSF-CHE 0911080) and G.B.D.
(NSF-CHE 1300722) benefit from grants from the National
Science Foundation. We thank Prof. Ranko Richert (Arizona
State) and Prof. Geoff Strouse (FSU) for helpful discussions.
We thank Tom Dusek for making quartz reaction vessels to our
custom specifications and Po-Kai Chen for periodically checking
and confirming the accuracy of MW power levels in the sample
cavity.
(25) For selected examples of chemical processes thought to occur
within transient solvent cages, see: (a) Liu, Q.; Wang, J. K.; Zewail, A.
H. Nature 1993, 364, 427−430. (b) Goldstein, S.; Czapski, G. J. Am.
Chem. Soc. 1999, 121, 2444−2447. (c) Zhang, Y.; Reynolds, N. T.;
Manju, K.; Rovis, T. J. Am. Chem. Soc. 2002, 124 (33), 9720−9721.
(26) Rosana, M. R.; Tao, Y.; Stiegman, A. E.; Dudley, G. B. Chem. Sci.
2012, 3, 1240−1244.
REFERENCES
■
(1) Gedye, R.; Smith, F.; Westaway, K.; Ali, H.; Baldisera, L.; Laberge,
L.; Rousell, J. Tetrahedron Lett. 1986, 27, 279−282.
(2) Giguere, R. J.; Bray, T. L.; Duncan, S. M.; Majetich, G. Tetrahedron
Lett. 1986, 27, 4945−4948.
(3) Metaxas, A. C.; Meredith, R. J. Industrial Microwave Heating; IET:
1983
(4) Schanche, J. S. Mol. Diver. 2003, 7, 291−298.
(5) (a) Stuerga, D. J. Microw. Power Electromagn. Energy 1996, 31, 87−
100. (b) Stuerga, D. J. Microw. Power Electromagn. Energy 1996, 31,
101−113.
(6) (a) Herrero, M. A.; Kremsner, J. M.; Kappe, C. O. J. Org. Chem.
2007, 73, 36−47. (b) Kappe, C. O. Chem. Soc. Rev. 2013, 42, 4977−
4990.
(7) Laurent, R.; Laporterie, A.; Dubac, J.; Berlan, J.; Lefeuvre, S.;
Audhuy, M. J. Org. Chem. 1992, 57, 7099−7102.
(8) (a) Nushiro, K.; Kikuchi, S.; Yamada, T. Chem. Lett. 2013, 42,
165−167. (b) Nushiro, K.; Kikuchi, S.; Yamada, T. Chem. Commun.
2013, 49, 8371−8373.
(9) Durand-Reville, T.; Gobbi, L. B.; Gray, B. L.; Ley, S. V.; Scott, J. S.
Org. Lett. 2002, 4, 3847−3850.
(10) Chen, P.-K.; Rosana, M. R.; Dudley, G. B.; Stiegman, A. E. J. Org.
Chem. 2014, DOI: 10.1021/jo5011526.
(27) For a discussion on the confusion between “heat” as a scientific
construct and its colloquial usage, see: Wiser, M.; Amin, T. Learning
and Instruction 2001, 11, 331−355.
(28) (a) Poon, K. W. C.; Dudley, G. B. J. Org. Chem. 2006, 71, 3923−
3927. (b) Poon, K. W. C.; Albiniak, P. A.; Dudley, G. B. Org. Synth.
2007, 84, 295−305.
(29) Albiniak, P. A.; Dudley, G. B. Tetrahedron Lett. 2007, 48, 8097−
8100.
(30) Nwoye, E. O.; Dudley, G. B. Chem. Commun. 2007, 1436−1437.
(31) Nishida, H.; Takada, N.; Yoshimura, M.; Sonoda, T.; Kobayashi,
H. Bull. Chem. Soc. Jpn. 1984, 57, 2600−2604.
(32) As a side note, it is clear to us from our experiments that quartz
vials (as opposed to Pyrex) should be employed in any studies aimed at
probing MW-specific thermal effects.
(33) The instrument ramped up to full power (nominally 300 W)
within the first 60 s and maintained full power for the duration of the
experiment. However, at that time (cf. ref 26) we were relying solely on
instrument power readings; thus, we cannot confirm that “full power”
7449
dx.doi.org/10.1021/jo501153r | J. Org. Chem. 2014, 79, 7437−7450

The Journal of Organic Chemistry
Article
equated to 300 W. We now measure the power output periodically to
confirm that the magnetron delivers 300 W of applied MW power into
the sample cavity. Also, we erroneously reported the average recorded
temperature as 92 °C, which included extended heating time after
reaction was complete; the average over the 30 min reaction time was
actually only 86 °C. We apologize for these oversights, which may have
caused us to underestimate the impact of MW heating on our reaction
system.
(34) It was interesting to us that the maximum temperature was
recorded early in the process, prior to full consumption of the starting
material. The solute composition changes over time (from substrate to
product) as the reaction progresses, and from this early temperature
maximum we infer that the starting materials are more absorbing (and
consequently produce more heat) than the products of this reaction.
(35) Kappe and co-workers later reported that heating similar reaction
systems in constant-temperature mode using relatively low MW power
did not produce any rate enhancement. However, their use of low and
variable MW power was inconsistent with our design parameters, which
called for high and constant MW power (cf. our control experiments
using low and variable MW power). Prof. Kappe declined our offers to
work with his laboratory to resolve any experimental discrepancies. For
more information, see: (a) Kappe, C. O.; Pieber, B.; Dallinger, D.
Angew. Chem., Int. Ed. 2013, 52, 1088−1094. (b) Dudley, G. B.;
Stiegman, A. E.; Rosana, M. R. Angew. Chem., Int. Ed. 2013, 52, 7918−
7923.
just in those experiments designed to probe MW-specific effects. Wang,
T.; Intaranukulkit, T.; Rosana, M. R.; Slegeris, R.; Simon, J.; Dudley, G.
B. Org. Biomol. Chem. 2012, 10, 248−250.
(42) Hayes, B. L. Aldrichim. Acta 2004, 37, 66−77.
(43) The integrated average temperature was 102 °C, and the effective
temperature was 104 °C. As discussed above, the integrated average
temperature is the temperature that would produce the same
conversion as the measured variable temperature profile according to
Arrhenius predictions (i.e., 13% over 10 min), and the effective
temperature is the temperature that would produce the same
conversion as was observed in the experiment: 16% over 10 min.
(44) (a) Geuens, J.; Kremsner, J. M.; Nebel, B. A.; Schober, S.;
Dommisse, R. A.; Mittelbach, M.; Tavernier, S.; Kappe, C. O.; Maes, B.
U. Energy Fuels 2007, 22, 643−645. (b) Razzaq, T.; Kappe, C. O.
Tetrahedron Lett. 2007, 48, 2513−2517.
(45) The external IR sensor, in contrast, is calibrated to correlate vial
temperature with solution temperature.
(46) The Kappe laboratory repeated these experiments and observed
a similar MW-specific rate enhancement; they reported 75% conversion
under MW heating and 41% conversion using oil bath heating after 30
min (see the Supporting Information for ref 35a). They went on to
show that adding heterogeneous “boiling chips” (made by crushing
various materials, including borosilicate glass) negated the MW-specific
rate accelerations. As discussed above in the control experiment 2,
heterogeneous additives can suppress MW-specific thermal effects in
our homogeneous system.
(47) (a) Baidakov, V. G. Explosive Boiling of Superheated Cryogenic
Liquids; Wiley-VCH: Weinheim, Germany, 2009. (b) Debenedetti, P.
G. Metastable Liquids; Princeton University Press: Princeton, NJ, 1996.
(48) (a) Baghurst, D. R.; Mingos, D. M. P. J. Chem. Soc., Chem.
Commun. 1992, 674−677. (b) Saillard, R.; Poux, M.; Berlan, J.;
Audhuy-Peaudecerf, M. Tetrahedron 1995, 51, 4033−4042. (c) Chemat,
F.; Esveld, E. Chem. Eng. Technol. 2001, 24, 735−744.
(36) This type of experimental designvariable MW power, highly
absorbing system, external temperature readingsis prone to “false
positive” identification of MW effects, which perhaps accounts for the
slightly elevated conversions observed in the early stages of this
experiment. This minor discrepancy notwithstanding, our interpreta-
tion is that there is not a significant difference in conversion between
MW and oil bath heating through a propylene glycol bath.
(37) Gutmann, B.; Obermayer, D.; Reichart, B.; Prekodravac, B.;
Irfan, M.; Kremsner, J. M.; Kappe, C. O. Chem. Eur. J. 2010, 16, 12182−
12194.
(49) Reactions run in p-xylene show a solvent impurity peak present
at ∼4.65 ppm. Additionally, a triplet near 5.3−5.4 ppm appears as a
result of minor degradation of BnOPB in some NMR solvents. Efforts
were made to minimize the exposure of our samples to NMR solvents
prior to spectrum acquisition.
(38) Robinson, J.; Kingman, S.; Irvine, D.; Licence, P.; Smith, A.;
Dimitrakis, G.; Obermayer, D.; Kappe, C. O. Phys. Chem. Chem. Phys.
2010, 12, 10793−10800.
(39) Assuming that propylene glycol absorbs at a level comparable to
that ethylene glycol (i.e., better than ethanol), our 80 mL bath of
propylene glycol likely blocked MW radiation more effectively than
would a 2−3 mm thick wall of silicon carbide. Further experiments
aimed at validating this assumption are in progress and will be reported
separately.
(40) We switched from p-xylene-d10 to p-xylene for cost reasons, once
it became apparent that we could monitor the reactions by 1H NMR in
either case. No solvent kinetic isotope effect is expected (or observed)
for rate-determining first-order thermolysis of the substrate. However,
we have observed different heating profiles (perhaps due to different
heat capacities) for deuterated vs nondeuterated solvents in MW
experiments, which can affect the magnitude of observed MW effects.
Relationships among isotope effects, solvent heat capacity, and MW
effects are beyond the scope of the present study and will be
investigated in future work.
(41) An alternative explanationthat the solution temperature “must
have been significantly higher than the values recorded using external
IR sensor technology” in this experimenthas been advanced in the
literature.35 We rule this out because: (1) our sensor was calibrated
repeatedly against an internal thermocouple probe to correlate the
external surface temperature with the internal solution temperature for
a range of relevant temperatures and for each reaction vial (cf. Figure
11); (2) there is no physical reason to suspect erroneous temperature
readings in this particular experiment under constant MW power, in
contrast to strongly absorbing systems heated under variable MW
power (cf. the propylene glycol control experiment and reference 36);
and (3) a faulty IR sensor (i.e., that is off by an average of ca 48 °C
here) would produce inaccurate readings systematically, as opposed to
impacting specifically this experiment. Our IR sensor has proven
reliable for measuring temperature in diverse MW experiments and not
7450
dx.doi.org/10.1021/jo501153r | J. Org. Chem. 2014, 79, 7437−7450