Deciphering a 20-Year-Old Conundrum: The Mechanisms of Reduction by the Water/Amine/SmI2 Mixture

Article abstract of DOI:10.1002/chem.201503104

The reaction of SmI₂ with the substrates 3-methyl-2-butanone, benzyl chloride, p-cyanobenzyl chloride, and anthracene were studied in the presence of water and an amine. In all cases, the water content versus rate profile shows a maximum at around 0.2 M H₂O. The rate versus amine content profile shows in all cases, except for benzyl chloride, saturation behavior, which is typical of a change in the identity of the rate-determining step. The mechanism that is in agreement with the observed data is that electron transfer occurs in the first step. With substrates that are not very electrophilic, the intermediate radical anions lose the added electron back to samarium(III) relatively quickly and the reaction cannot progress efficiently. However, in a mixture of water/amine, the amine deprotonates a molecule of water coordinated to samarium(III). The negatively charged hydroxide, which is coordinated to samarium(III), reduces its electrophilicity, and therefore, lowers the rate of back electron transfer, which allows the reaction to progress. In the case of benzyl chloride, in which electron transfer is rate determining, deprotonation by the amine is coupled to the electron-transfer step.

Full text of DOI:10.1002/chem.201503104

DOI: 10.1002/chem.201503104
Full Paper
&
Reaction Mechanisms |Hot Paper|
Deciphering a 20-Year-Old Conundrum: The Mechanisms of
Reduction by the Water/Amine/SmI Mixture
2
[
a]
Sandeepan Maity and Shmaryahu Hoz*
Abstract: The reaction of SmI with the substrates 3-methyl-
added electron back to samarium(III) relatively quickly and
the reaction cannot progress efficiently. However, in a mix-
ture of water/amine, the amine deprotonates a molecule of
water coordinated to samarium(III). The negatively charged
hydroxide, which is coordinated to samarium(III), reduces its
electrophilicity, and therefore, lowers the rate of back elec-
tron transfer, which allows the reaction to progress. In the
case of benzyl chloride, in which electron transfer is rate de-
termining, deprotonation by the amine is coupled to the
electron-transfer step.
2
2
-butanone, benzyl chloride, p-cyanobenzyl chloride, and an-
thracene were studied in the presence of water and an
amine. In all cases, the water content versus rate profile
shows a maximum at around 0.2m H O. The rate versus
2
amine content profile shows in all cases, except for benzyl
chloride, saturation behavior, which is typical of a change in
the identity of the rate-determining step. The mechanism
that is in agreement with the observed data is that electron
transfer occurs in the first step. With substrates that are not
very electrophilic, the intermediate radical anions lose the
Introduction
the development of this area, which was further developed by
[
12]
Procter et al.
Despite its synthetic importance,
[
11h]
Following the introduction of SmI as a single-electron-transfer
the mechanism of the
2
[
1]
reducing agent to organic chemistry by Kagan et al., many
researchers have striven to develop means to control its che-
reaction in question remains unclear, although several at-
tempts have been made to explain it. Suggestions include co-
[2]
[10]
moselectivity and enhance its reactivity. This was usually ach-
ieved by using additives, which either complexed to the sub-
ordination of the amine to SmI2, precipitation of Sm(OH) to
3
provide the driving force for pushing the reaction to the
[
11a]
strate or, more frequently, to SmI . Examples of the first group
right,
nucleophilic assistance by the amine in the reduction
2
[3]
[11e]
are iron, copper, and nickel salts. However, most efforts fo-
of benzyl halides,
and deprotonation of the water molecule
[
12f]
cused on additives that coordinated to the SmI itself. Ligands
to generate a better reducing agent
(hydroxide ion coordi-
2
for the samarium cation can affect the course of the reaction
nated to SmI ). Moreover, contradictory information on the ki-
2
[
4]
[12e]
in several ways. For example, hexamethylphosphoramide
netic rate order of water and amine has been reported.
A
[
5]
(
HMPA), which is one of the most commonly used ligands,
statement made by Procter et al. in a recent paper describes
faithfully the state of the art of this interesting reaction: “the
substantially increases the reduction potential of SmI . This re-
2
sults from greater energy gain upon coordination of HMPA to
mechanistic details of this process, including the critical role of
[6]
[12f]
samarium(III) than that to samarium(II). Proton donors, such
amine and H O additives, remained unclear”.
Herein, we
2
as MeOH and water, also coordinate to SmI , but without
report a mechanistic study of this reaction with the substrates
3-methyl-2-butanone, benzyl chloride, p-cyanobenzyl chloride,
and anthracene and show the role of the amine and water in
this often verbally referred to as “magic mixture”.
2
[
7,8]
having a significant effect on its reduction potential.
Their
added value stems from the ability to trap efficiently short-
lived radical anions generated by electron transfer from SmI2
[
7a,9]
to the substrate before back electron transfer takes place.
In 1995, Cabri et al. discovered that the addition of water
and amine to the reaction mixture significantly enhanced the
Results and Discussion
[
10]
reactions of SmI2. This discovery lay dormant for seven years
until it was brought to the public’s attention by Hilmersson
It seems most likely that a mechanism that necessitates the
presence of both water and amine must involve proton trans-
fer between these two parties, as suggested by Procter
[
11]
et al. The group of Hilmersson contributed significantly to
[
12f]
et al.
This is supported by the fact that a correlation was
[
a] Dr. S. Maity, Prof. S. Hoz
Department of Chemistry, Bar-Ilan University
Ramat Gan, 52900 (Israel)
found between the basicity of the amines and their activi-
[11h,12f]
ty.
In addition, Kudo and Kamochi showed that the
[
13]
K(Na)OH/water system was also an efficient combination;
E-mail: shoz@biu.ac.il
this demonstrated the importance of the presence of a hydrox-
ide ion. It is clear that of the two components, water and
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201503104.
Chem. Eur. J. 2015, 21, 18394 – 18400
18394
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 3. Rate constants as a function of water concentration in the reaction
of SmI with benzyl chloride in the presence of Et N.
2 3
cases: [SmI ]=1 mm, [substrate]=10 mm, and [amine]=
2
2
Figure 1. UV/Vis spectra of SmI in the presence of water.
2
0 mm. In Figure 3, the effect of varying the water concentra-
tion over the range of 0.001–1m in the reaction of SmI with
2
amine, water will usually preferentially complex to samarium.
benzyl chloride in the presence of Et N is shown.
3
This is evidenced from spectral changes of SmI caused by the
The reaction rate increases rapidly at the beginning of the
reaction and reaches a maximum at around 0.2m. Further ad-
dition of water results in a gradual decrease of the rate con-
2
incremental addition of water (Figure 1; similar behavior was
[11h]
observed also by Hilmersson and DahlØn ) and the lack of
[
12f]
any spectral changes in the presence of Et N. The absence of
stants. Similar results were obtained by Procter
and Hilmers-
3
[
11l]
isosbestic points in Figure 1 indicates the involvement of an in-
termediate structure(s) between noncomplexed and fully
son and Anker. This rate decrease may be the result of two
possible causes. The first is that, as the water concentration is
increased, second and even third solvation shells are built. This
may hamper inner-sphere electron transfer, and thus, lower
the rate. However, this should affect only substrates such as
water-complexed SmI . However, a close examination of
2
Figure 1 shows that isosbestic points do exist within the low
(
0–100 mm) and higher water concentration range (200–
00 mm), as shown in Figure 2. This implies that at a low con-
5
carbonyl derivatives, which are known to react with SmI by an
2
[
14]
inner-sphere mechanism.
However, it was suggested that
polyaromatic compounds, such as anthracene, react by an
[
14]
outer-sphere mechanism, and yet, anthracene exhibits the
same phenomenon (see the Supporting Information). More-
over, for all substrates examined, 3-methyl-2-butanone (with
Et N and tetramethylethylenediamine (TMEDA)), benzyl chlo-
3
ride (with Et N), p-cyanobenzyl chloride (with ethyldiisopropyl-
3
amine (EDIPA), since it was too fast with Et N), and anthracene
3
(
with Et N; see the Supporting Information), the maximum
3
effect is achieved at the same water concentration (0.2m). This
implies that the rate drop is not substrate dependent. A more
likely explanation is based on the rate of the deprotonation re-
action. A water molecule or any other proton donor, when at-
tached directly to the samarium cation, will enjoy a significant
acidity increase. In MeOH, this acidity increase amounts to
2
Figure 2. UV/Vis spectra of SmI in the presence of water at a concentration
of 0–100 mm. The inset shows the spectrum in the presence of 200–500 mm
of water.
[
15]
about 11 pK units. In THF, a solvent with a lesser ability to
a
stabilize anions, this effect may even be doubled. As a result,
deprotonation by the amine will be significantly enhanced.
However, the acidification of water molecules in the secondary
and tertiary solvation shells will be much diminished. Because
the amine cannot access the more acidic bound water mole-
cules in the primary solvation shell, the rate of deprotonation
will be drastically decreased and because, as we show below,
deprotonation rates play a crucial role in the reduction pro-
cess, the overall rate will be diminished.
centration of water SmI is converted into one type of absorb-
2
ing species (which may represent several different complexes
with the same absorption) and upon increasing the concentra-
tion of water this intermediate(s) is converted into the final
complex without going through an additional intermediate
with a different spectral absorption.
We first examined the effect of varying the water concentra-
tion on the reaction kinetics in the presence of a constant con-
centration of amine. All reactions were carried out under
pseudo-first-order conditions and the reaction rate was moni-
We now turn to the effect of the amine on the reaction rate.
Again, all reactions were carried out under pseudo-first-order
conditions and the rate was monitored at l=619 nm. The fol-
tored by following the disappearance of the SmI absorption
lowing concentrations were used in all cases: [SmI ]=1 mm,
2
2
at l=570 nm. The following concentrations were used in all
[substrate]=10 mm, and [water]=50 mm. This concentration
Chem. Eur. J. 2015, 21, 18394 – 18400
www.chemeurj.org
18395
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
of water was chosen because at higher concentrations the re-
actions with high amine concentrations were too fast to
follow. The effect of varying the amine concentration over the
The aforementioned results suggest that the second option
[Eq. (2)] is the actual scenario. The “magic mixture” is used in
cases in which, under normal conditions, that is, in the absence
of water and amine, the reaction does not (or hardly) take
place. This usually means that the electron-transfer step is
highly endothermic, and therefore, is highly reversible. In the
absence of a rapid successive step that prevents back electron
transfer from the radical anion to samarium(III) and pushes the
reaction to completion, the reaction will not take place. We
have embarked upon this problem previously and for cases in
which the “locking” step was protonation of the radical anion,
the simple solution we devised was to enhance the protona-
range of 0.001–0.4m in the reaction of SmI with 3-methyl-2-
2
butanone is shown in Figure 4. The rate profile exhibits a satu-
[
7a,9]
tion step.
This was achieved by replacing inefficient bimo-
lecular protonation by a proton donor from the bulk, through
efficient unimolecular protonation within the ion pair by
a proton donor already complexed to the samarium cation. We
realize that the magic mixture provides an alternative mecha-
nism. Instead of locking the substrate radical anion by proto-
nation, it prevents it from dispensing the added electron by re-
Figure 4. Rate constants as a function of Et
3
N concentration in the reaction
of SmI with 3-methyl-2-butanone in the presence of water.
2
III
ration effect. The kinetic order in amine before the onset of
ducing the electrophilicity of the acceptor: Sm . Reduced elec-
2
levelling off is one (slope=0.99, r =1.00). The corresponding
trophilicity is the outcome of the partial neutralization of the
effective charge by the negatively charged hydroxide. As
a result, the lifetime of the radical anion is increased, which en-
ables it to proceed further towards the products. Saturation
observed in the rate profile of the amine is in agreement with
this mechanism. At relatively low amine concentrations, the re-
action is first order in the amine. As the amine concentration
increases, the deprotonation rate is enhanced until the rate of
the back electron transfer step in Equation (2) and the rate of
deprotonation become similar. At this point, the bend in the
rate profile is achieved and electron transfer starts to become
the rate-determining step. Further increasing the amine con-
centration will therefore not affect the rate because deproto-
nation becomes a post-rate-determining step (Figure 5).
figures for other substrates and amines appear in the Support-
ing Information.
The rate profiles for water and amine provide a clear indica-
tion of the reaction mechanism and of the role these two com-
ponents have in the reaction. Assuming that the amine deprot-
onates a water molecule complexed to samarium, there are
three possible scenarios for this deprotonation of water by the
amine. The first is that deprotonation takes place before elec-
tron transfer [Eq. (1)]. In other words, the amine will deproto-
nate either partly or fully complexed SmI . The coordination of
2
À
II
OH to Sm will render the latter a better donor, and hence,
enhance the reaction. The second scenario is that deprotona-
tion takes place after electron transfer. This will generate a hy-
III
droxide anion bound to Sm [Eq. (2)] and will enhance the re-
III
action by virtue of diminishing the electrophilicity of Sm I ,
2
and hence, reduce the rate of back electron transfer. The third
possibility [Eq. (3)] is that deprotonation occurs simultaneously
with electron transfer (proton-coupled electron transfer
[
16]
(
PCET)). For the sake of simplicity, in Equations (1)–(3), A is
the substrate, R is a general substituent on nitrogen, and only
one water molecule is shown to be complexed to the samari-
um cation.
Figure 5. A rate profile showing electron transfer followed by deprotonation.
Amine concentration increases from a to c. The bend occurs at b.
Because the bend in the rate profile indicates a point at
which the rate of deprotonation by the amine is equal to the
rate of back electron transfer, one could expect that in less
electrophilic substrates, in which back electron transfer is
faster, the rate will be matched only at a higher amine concen-
Chem. Eur. J. 2015, 21, 18394 – 18400
www.chemeurj.org
18396
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tration. Indeed, whereas the reaction with 3-methyl-2-buta-
It should be pointed out that there are three bases present
in the reaction mixture: water, amine, and the minute amounts
of free hydroxide ion present in the solution due to the equi-
librium between water and amine. Their order of reactivity is
hydroxide@amine>water. However, because the water and
hydroxide concentrations are constant in the buffer experi-
ment, their effect will be reflected in the intercept of the plot
shown in Figure 6, whereas the slope is governed by increas-
ing amine concentration. We have performed an experiment at
the same water concentration in the absence of buffer with
none has its bend at [Et N]ꢀ0.05m (Figure 4), with a much
3
less electrophilic anthracene the bend is achieved at only
around 0.6m (see the Supporting Information).
It should be pointed out that the amine rate profile not only
supports deprotonation as a post-electron-transfer step, it also
rules out deprotonation as the first step [Eq. (1)]. Such a mecha-
nism necessitates kinetics that are first order in amine at all
concentrations. A leveling off may be achieved only in the un-
likely event that complete deprotonation of all samarium–
water complexes is achieved. Needless to say, in this kind of
scenario, the magic mixture should have exhibited a significant
a varied concentration of Et N. In this case, the concentration
3
of hydroxide ions is expected to increase in parallel to the
amine concentration and, indeed, the slope obtained was
much higher, which showed that the high reactivity of the hy-
droxide ion overcompensated for its low concentration (see
the Supporting Information). It should be pointed out that the
hydroxide, in this case, may function both as a general base
increase in the reduction potential of SmI , which was not ob-
2
[
11e]
served.
The suggested mechanism can be further supported thanks
to the fact that, for several decades, the mechanism of proton-
transfer reactions was intensively studied. Thus, the arsenal of
physical organic chemistry contains a very useful tool to inves-
tigate the case in hand. In terms of proton-transfer chemistry,
the topic is referred to as “specific versus general base cataly-
sis” and the diagnostic tool developed is aimed to distinguish
between cases in which proton transfer takes place at the pre-
equilibrium stage (specific catalysis) or at the rate-determining
and/or as a ligand for SmI . In the latter case, it may produce
2
a very strong reducing agent, which is in agreement with the
[
12f]
suggestion by Procter et al.
However, due to its minute con-
centration, it cannot be electrochemically detected.
We now turn to the feasibility of the third scenario, in which
electron transfer and deprotonation are concerted [Eq. (3)].
The highest probability for such a mechanism is with a sub-
strate for which electron transfer is rate determining with no
possibility of back electron transfer. Benzyl chloride fits this re-
quirement because electron transfer is concomitant with the
departure of the leaving group. Indeed, in this case, no satura-
tion kinetics were observed and a first-order reaction in amine
was observed up to 1.2m (Figure 7).
[17]
step (general catalysis). In the first case, the concentration of
À
the deprotonated species (OH complexed to SmI ) is pH de-
2
pendent. In the second case, the rate depends on the concen-
tration of the individual bases present in solution and their ki-
netic basicity. Thus, working in a buffered solution and chang-
ing the concentration of the buffer without changing the
buffer ratio will keep the pH constant. Nevertheless, the con-
centration of individual bases present in solution will increase
with increasing buffer concentration. Hence, if the reaction is
a specific base-catalyzed reaction, a plot of the rate versus
buffer concentration will be linear with a slope of zero because
the pH remains constant. Whereas if the reaction is general
base catalysis, that is, deprotonation is the rate-determining
step, since the free-base concentration increases with concen-
tration of buffer, while that of the hydroxide ion remains con-
stant, the plot will exhibit a positive slope. By using the couple
Et N/Et N·HI as the buffer (see the Supporting Information for
3
3
details) and varying its concentration, while the pH was main-
tained constant, a straight line with a positive slope was ob-
tained (Figure 6). This indicates that proton transfer takes place
at the rate-determining step.
Figure 7. Amine kinetic order in the reaction of benzyl chloride (10 mm)
with SmI (1 mm) and water (50 mm).
2
To exclude other mechanistic options, we performed the
buffer experiment on this system as well. The positive slope
2
obtained (1.6, r =0.9947; see the Supporting Information)
clearly shows a general base catalysis mechanism in which the
proton is transferred in the rate-determining step. Because the
aforementioned buffer diagnostic tool rules out the possibility
of pre-deprotonation as the first step [Eq. (1)] and since elec-
tron transfer is irreversible, we conclude that proton transfer is
coupled to the electron-transfer step (PCET) in the rate-deter-
mining step.
SavØant et al. convincingly showed that the introduction of
an electron-withdrawing substituent on the benzyl halides
Figure 6. General base-catalyzed experiment with 3-methyl-2-butanone
(
20 mm) and SmI
2
(2 mm) in THF/water (1/1 v/v).
Chem. Eur. J. 2015, 21, 18394 – 18400
www.chemeurj.org
18397
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[
18]
caused the mesolytic cleavage to be a stepwise reaction. We
therefore investigated the reaction of p-cyanobenzyl chloride.
As expected, the rate constant versus [amine] plot did not dis-
play a straight line up to high amine concentration, as found
for benzyl chloride, but rather showed leveling-off behavior
(
see the Supporting Information), which was typical of reversi-
ble electron-transfer reduction.
We have shown herein that, in the Cabri–Hilmersson reac-
tions, electron transfer could occur either concomitantly with
proton transfer (PCET) or could precede it. The only substrate
in this study in which the two processes are merged into
a single transition state is benzyl chloride. Other substrates re-
acted in the sequential mode. The advantage of coupling in
PCET is that the system avoids high-energy species on the re-
action path. On the other hand, there are two probability
obstacles: one that is embedded in the entropy factor,
whereas the other is the encounter probability, which is con-
centration dependent. The activation entropy reflects the re-
duction in the degrees of freedom upon arriving at the
transition state. It is probably similar for both process, al-
though probably slightly more demanding for PCET. However,
the major difference is in the concentration-dependent en-
counter probability. In enzymes, when all reacting parties are
at the active site, the probability is, of course, very high. For
Figure 8. A log–log plot showing the amine kinetic order in the reaction of
SmI with 3-methyl-2-butanone.
2
Figure 9. A log–log plot showing the amine kinetic order in the reaction of
SmI with anthracene.
2
[19]
example, in cytochrome c oxidase,
electron transfer is ac-
We now turn to a quantitative analysis of the kinetic results,
which is enabled mainly because of the fortunate existence of
the saturation phenomenon in the rate versus amine concen-
tration profile. The overall rate equation is rate=k[A][NR3]
companied by proton transfer from tyrosine 244 cross-linked
to histidine 240. Needless to say, if protonation were external
(
bimolecular), the probability for a coupled process would
be drastically diminished. Clearly, benzyl chloride reacts
through PCET, although proton transfer is bimolecular because
it offers a lower energy path. Why then would the other sub-
strates not also benefit from this? We believe that they do and
there is also a PCET component in the other cases. The fact
that it is not so dominant results again from probability con-
siderations.
[SmI ]. By assuming a steady-state concentration for the radical
2
anionic intermediate in Equation (2), Equation (4) is obtained:
k k ½NR Š
1
2
3
rate ¼ k
½AŠ½SmI2Š
3
ð4Þ
þ k ½NR Š
À1
2
When Equation (4) is applied to the uprising limb of Figure 4,
@k [NR ], the equation will acquire the form given in Equa-
To clarify the above assertion, let us assume, for the sake of
k
À1
simplicity, that the substrate pairs first with SmI , which trans-
2
3
2
tion (5):
fers to it an electron in an equilibrium reaction. A longer life-
time of the ion pair produced in this reaction results in a great-
er probability for its encounter with an amine and deprotona-
k k ½NR Š
1
2
3
rate ¼
½AŠ½SmI2Š
ð5Þ
k
À1
tion of a water molecule bound to SmI . As a result, the elec-
2
tron affinity of the substrate, and hence, the lifetime of its radi-
cal anion increases, along with the contribution of the
sequential component to the overall reaction at the expense
of the PCET component. This conclusion is evidenced by the
shape of the plateau section for the various substrates. We
have two substrates that differ markedly in their electron affini-
ty: anthracene and 3-methyl-2-butanone. One can clearly see
in log–log plots that the slope at the plateau region is not
zero (Figures 8 and 9). With anthracene it is about 0.42
Therefore, the first-order rate constant (k) in Figure 4 is
equal to Equation (6):
k k ½NR Š
1
2
3
k ¼
½AŠ
ð6Þ
k
À1
The slope of the uprising limb in Figure 4 is given by Equa-
tion (7):
(
(
Figure 8), whereas for 3-methyl-2-butanone it is much smaller
0.18; Figure 9). This is, of course, in agreement with the
Dk
D½NR3Š
k1k2
¼ k ½AŠ
À1
ð7Þ
higher stability of the radical anion of the ketone, which favors
the stepwise reaction. The latter occurs through increased
probability of an encounter between the ion pair and the
amine compared with that of anthracene, for which this proba-
bility is smaller and leaves more room for PCET.
Because the slope can be measured in each case and the
concentration of the substrate is known, the value of k k /k
1
2
À1
could be determined. In the plateau region, k !k [NR ], the
À1
2
3
rate is equal to that given by Equation (8):
Chem. Eur. J. 2015, 21, 18394 – 18400
www.chemeurj.org
18398 ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
rate ¼ k ½AŠ½SmI Š
ð8Þ
water molecule may transfer a proton either to the carbonyl
radical anion or to the amine. In both cases, proton transfer
takes place between two heteroatoms. Why then is proton
transfer to the amine of the magic mixture more facile than
proton transfer from water to the oxygen atom in the radical
anion? The answer lies, most probably, in the realm of electro-
static interactions. We have shown that electrostatic interac-
tions between the radical anions
1
2
Namely, electron transfer is the rate-determining step and
the rate constant (k ) can be determined by back extrapolation
1
of the data from the plateau region to an amine concentration
of zero. Once k is known, the ratio k /k [Eq. (7)] can be de-
1
2
À1
termined. Relevant data is presented in Table 1.
Table 1. Kinetic analysis for the reactions with SmI
2
under Cabri–Hilmersson conditions.
of p-dimethylaminobenzophe-
III
none and [Sm ][2I], amounts to
[
a]
À1 À1
À1
Entry
Substrate (amine)
Dk/D (NR
3
)
k
1
[m
s
]
2
k /kÀ1 [m ]
À1 [20]
about 30 kcalmol .
Because
1
2
3
4
3-methyl-2-butanone (Et
3-methyl-2-butanone (TMEDA)
p-cyanobenzyl chloride (EDIPA)
3
anthracene (Et N)
3
N)
4130
2920
14390
250
19900
17400
26000
9400
21
18
55
3
the radical anion of the aliphatic
ketone of the present study is
much “harder” than the highly
delocalized benzophenone radi-
cal anion, its interaction with
III
À1
[21]
The electron-transfer rate constant (k ) for 3-methyl-2-buta-
“hard” Sm may be 5–10 kcalmol stronger. Thus, protona-
tion of the carbonyl radical anion will result in a loss of stabili-
1
none was determined in two experiments, each with a different
amine (Table 1, entries 1 and 2): (18700Æ1250)mÀ1 sÀ1. This
value is slightly lower than that for p-cyanobenzyl chloride
and, as expected, significantly higher than that for anthracene.
À1
zation energy of at least 35 kcalmol relative to proton trans-
fer to the amine. This rough estimate provides a good explana-
tion for the superiority of the magic mixture over direct proton
transfer to the carbonyl radical anion.
The k /k value in Table 1 is slightly higher for Et N than that
2
À1
3
for TMEDA. Because k is independent of the amine, this re-
À1
flects the relative kinetic basicity of the two amines. A compar-
ison of the results given in Table 1, entries 1 and 4, keeping in
mind that the deprotonation rate constants are identical in the
two cases, shows that the rate constant for back electron
Conclusion
Based on the evidence gathered, the most probable main reac-
tion mechanism for reduction by a mixture of water/amine/
transfer, k , is seven times larger for anthracene than that for
À1
2
SmI , is deprotonation of a water molecule complexed to SmI2
3
-methyl-2-butanone.
by an amine molecule after the electron-transfer step. This de-
Finally, we would like to discuss the differences between the
protonation, which binds a negatively charged hydroxide ion
two methods that enhance the reaction by competing with or
III
to Sm reduces its electrophilicity, and therefore, reduces the
effectively reducing the back electron transfer step [k ;
À1
rate of back electron transfer within the ion pair. By conduct-
ing a buffer experiment, it was shown that all reactions, at rela-
tively low amine concentration, were general base catalyzed.
In other words, deprotonation can be effected by any base in
the system, including water and hydroxide. The case of benzyl
chloride, for which electron transfer and cleavage of the CÀCl
bond take place in a single step [Eq. (3)], is unique among the
substrates, since no saturation is observed in the plot of rate
versus [amine]. Because electron transfer is irreversible and de-
protonation is rate determining, the reaction must be PCET.
Eq. (4)]. The first one is unimolecular protonation by proton
donors such as water or a molecule of methanol complexed to
SmI . The complexed proton donor protonates efficiently the
2
radical anion within the ion pair [increasing k in Eq. (9)], push-
2
ing the reaction forward.
The fact that, despite the ability of water molecules com-
plexed to SmI to unimolecularly protonate the radical anion,
2
there is a strong dependence of the rate on the amine concen-
tration. This indicates that the magic mixture is more effective
at preventing back electron transfer than protonation of the
radical anion, despite the fact that deprotonation is bimolecu-
lar. When protonation is supposed to take place on a carbon
atom, as is the case with anthracene, the explanation is rather
simple. Protonation on carbon, even if thermodynamically fa-
vored, is known to be slow and usually cannot compete with
proton transfer between two heteroatoms, such as between
Experimental Section
General
All reagents were purified prior to use by following standard pro-
[22]
cedures. Liquid reagents (substrates, water, and amines) were
degassed with argon prior to use. THF was dried and freshly dis-
tilled from sodium/benzophenone under an argon atmosphere.
SmI was freshly prepared prior to use by stirring samarium metal
2
[1b]
and 1,2-diiodoethane at room temperature. The concentration
of SmI2 was determined by UV/Vis spectroscopic measurements
[
17]
amine and water. Therefore, in these cases, the first method
internal protonation) is not expected to successfully compete
(l=619 nm; e=635). Kinetic and preparative reactions were car-
(
ried out in clean and dry glassware under a nitrogen atmosphere.
NMR spectra were recorded by using 300/400 MHz Bruker instru-
ments.
with the magic mixture. However, the case of carbonyl com-
pounds presents a more difficult problem. In this case, the
Chem. Eur. J. 2015, 21, 18394 – 18400
www.chemeurj.org
18399
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Kinetics
Chem. Soc. Rev. 2013, 42, 9155–9183; c) D. J. Edmonds, D. Johnston,
D. J. Procter, Chem. Rev. 2004, 104, 3371–3403; d) P. G. Steel, J. Chem.
Soc. Perkin Trans. 1 2001, 2727–2751; e) A. Krief, A.-M. Laval, Chem. Rev.
1999, 99, 745–777; f) G. A. Molander, C. R. Harris, Tetrahedron 1998, 54,
3321–3354; g) G. A. Molander, C. R. Harris, Chem. Rev. 1996, 96, 307–
Stopped-flow kinetic measurements were carried out by using
a Hi-Tech stopped-flow spectrometer. All reactions were performed
under pseudo-first-order conditions. Rates of reactions were moni-
338.
tored by following the disappearance of the SmI absorbance at
2
[
[
[
3] H. B. Kagan, Tetrahedron 2003, 59, 10351–10372.
4] R. A. Flowers II, Synlett 2008, 10, 1427–1439.
5] a) J. Inanaga, M. Ishikawa, M. Yamaguchi, Chem. Lett. 1987, 1485–1486;
b) J. B. Shotwell, J. M. Sealy, R. A. Flowers II, J. Org. Chem. 1999, 64,
l=619 (for the amine profile) and 570 nm (for the water profile).
To ensure that the expected products were also obtained under ki-
netic conditions, preparative reactions were conducted with reac-
tant concentrations similar to those used in the kinetic measure-
ments.
5251–5255; c) R. J. Enemærke, T. Hertz, T. Skrydstrup, K. Daasbjerg,
Chem. Eur. J. 2000, 6, 3747–3754; d) D. V. Sadasivam, J. A. Teprovich, Jr.,
D. J. Procter, R. A. Flowers II, Org. Lett. 2010, 12, 4140–4143; e) E.
Prasad, R. A. Flowers II, J. Am. Chem. Soc. 2002, 124, 6895–6899.
6] C. N. Rao, S. Hoz, J. Am. Chem. Soc. 2011, 133, 14795–14803.
7] a) M. Amiel-Levy, S. Hoz, J. Am. Chem. Soc. 2009, 131, 8280–8284; b) A.
Yacovan, S. Hoz, I. Bilkis, J. Am. Chem. Soc. 1996, 118, 261–262.
8] See, however, at high water concentration: E. Prasad, R. A. Flowers II, J.
Am. Chem. Soc. 2005, 127, 18093–18099.
9] M. Amiel-Levy, S. Hoz, Chem. Eur. J. 2010, 16, 805–809.
10] W. Cabri, I. Candiani, M. Colombo, L. Franzoi, A. Bedeschi, Tetrahedron
Lett. 1995, 36, 949–952.
11] a) A. DahlØn, G. Hilmersson, Tetrahedron Lett. 2002, 43, 7197–7200; b) A.
DahlØn, G. Hilmersson, Tetrahedron Lett. 2003, 44, 2661–2664; c) A.
DahlØn, A. Petersson, G. Hilmersson, Org. Biomol. Chem. 2003, 1, 2423–
2426; d) A. DahlØn, G. Hilmersson, Chem. Eur. J. 2003, 9, 1123–1128;
e) A. DahlØn, G. Hilmersson, B. W. Knettle, R. A. Flowers II, J. Org. Chem.
General procedure for preparative reactions under kinetic
conditions
[
[
A solution (10 mL) of the substrate (benzyl chloride (0.63 g;
[
[
4
(
.98 mmol), 3-methyl-2-butanone (0.43 g; 4.98 mmol), anthracene
0.44 g; 2.47 mmol), or p-cyanobenzyl chloride (0.75 g; 4.98 mmol))
in THF containing triethylamine (0.20 g, 1.97 mmol) and H O
2
[
[
(
(
0.054 g, 3 mmol) was added to a solution of SmI (10 mL) in THF
0.1m) in a volumetric flask. The final concentrations of the differ-
2
ent reactants in the reaction mixture were as follows: SmI , 0.05m;
substrates, 0.25m (except for anthracene, which was 0.125m); Et N,
2
3
0
.1m; and H O, 0.15m). After the decolorization of SmI , the reac-
2 2
tion mixture was filtered and the filtrate was diluted with chloro-
form (25 mL). The organic solution was washed with a 0.1m solu-
tion of HCl (10 mL). The aqueous layer was extracted with chloro-
form (3”10 mL). The solutions in chloroform were combined, dried
over anhydrous sodium sulfate, concentrated under reduced pres-
sure, and analyzed by NMR spectroscopy and ESI-MS. The products
for 3-methyl-2-butanone, benzyl chloride, p-cyanobenzyl chloride,
and anthracene were 3-methyl-2-butanol, toluene, p-cyanotoluene,
and 9,10-dihydroanthracene, respectively.
2003, 68, 4870–4875; f) A. DahlØn, A. Sundgren, M. Lahmann, S. Oscar-
son, G. Hilmersson, Org. Lett. 2003, 5, 4085–4088; g) T. A. Davis, P. R.
Chopade, G. Hilmersson, R. A. Flowers II, Org. Lett. 2005, 7, 119–122;
h) A. DahlØn, G. Hilmersson, J. Am. Chem. Soc. 2005, 127, 8340–8347;
i) T. Ankner, G. Hilmersson, Tetrahedron Lett. 2007, 48, 5707–5710; j) A.
DahlØn, A. Nilsson, G. Hilmersson, J. Org. Chem. 2006, 71, 1576–1580;
k) T. Ankner, G. Hilmersson, Org. Lett. 2009, 11, 503–506; l) T. Ankner, G.
Hilmersson, Tetrahedron 2009, 65, 10856–10862.
[12] a) M. Szostak, M. Spain, D. J. Procter, Org. Lett. 2012, 14, 840–843; b) M.
Szostak, M. Spain, D. J. Procter, Org. Lett. 2014, 16, 5052–5055; c) M.
Szostak, B. Sautier, M. Spain, D. J. Procter, Org. Lett. 2014, 16, 1092–
General procedure for buffer experiments
1095; d) M. Szostak, M. Spain, A. J. Eberhart, D. J. Procter, J. Am. Chem.
Soc. 2014, 136, 2268–2271; e) M. Szostak, M. Spain, A. J. Eberhart, D. J.
Procter, J. Org. Chem. 2014, 79, 11988–12003; f) M. Szostak, M. Spain,
D. J. Procter, Chem. Eur. J. 2014, 20, 4222–4226.
The rate of reduction of 3-methyl-2-butanone and benzyl chloride
was measured in the presence of buffer. Buffer solutions were pre-
pared in water with a 1:1 mixture of Et N and its corresponding
3
[
13] Y. Kamochi, T. Kudo, Chem. Lett. 1991, 20, 893–896.
[14] R. J. Enemærke, K. Daasbjerg, T. Skrydstrup, Chem. Commun. 1999, 343–
344.
iodide salt, [Et NH][I]. Kinetic measurements were performed in
3
a 1:1 (v/v) mixture of THF and H O. The solution of SmI and sub-
2
2
strate in THF was placed in one syringe and the buffer solution in
the other. The reactions rates were monitored by following the
[15] A. A. Neverov, G. Gibson, R. S. Brown, Inorg. Chem. 2003, 42, 228–234.
[16] a) D. R. Weinberg, C. J. Gagliardi, J. F. Hull, C. F. Murphy, C. A. Kent, B. C.
Westlake, A. Paul, D. H. Ess, D. G. McCafferty, T. J. Meyer, Chem. Rev.
decay of the SmI absorbance at l=560 nm.
2
2
012, 112, 4016–4093; b) R. Yella, S. Hoz, Org. Lett. 2013, 15, 5262–
265; c) R. Yella, S. Hoz, Org. Lett. 2014, 16, 3876–3879.
17] R. P. Bell, The Proton in Chemistry, 2nd ed., Chapman and Hall, London,
973.
18] C. P. Andrieux, A. L. Gorande, J.-M. SavØant, J. Am. Chem. Soc. 1992, 114,
892–6904.
5
Preparation of [Et NH][I]
3
[
[
1
HCl (0.154 g, 4.21 mmol; 0.48 mL of 32% HCl in H O) was added to
2
Et N (0.426 g, 4.21 mmol) in THF (5 mL). The product [Et NH][Cl]
3
3
6
was filtered and dried by applying high temperature and vacuum.
It was then dissolved in acetonitrile and NaI (0.63 g, 4.21 mmol)
was added to the solution. This resulted in the precipitation of
NaCl, which was separated through filtration. The filtrate was con-
centrated under reduced pressure and dried in vacuum.
[
[
[
19] S. Yoshikawa, A. Shimada, Chem. Rev. 2015, 115, 1936–1989.
20] H. Farran, S. Hoz, Org. Lett. 2008, 10, 4875–4877.
21] The difference in the interaction energy between p-dimethylaminoben-
À1
zophenone and p,p-dichlorobenzophenone is 3 kcalmol . Therefore,
for an aliphatic ketone, such as 3-methyl-2-butanone, in which the
charge cannot be delocalized onto an aromatic ring, the electrostatic
interaction ought to be very much higher.
Keywords: electron transfer · kinetics · proton transfer ·
reaction mechanisms · samarium iodide
[22] D. D. Perrin, W. L. F. Armarego, Purification of Laboratory Chemicals,
rd ed., Pergamon Press, New York, 1989.
3
[
[
1] a) J. L. Namy, P. Girad, H. B. Kagan, New J. Chem. 1977, 1, 5–7; b) P.
Girard, J. L. Namy, H. B. Kagan, J. Am. Chem. Soc. 1980, 102, 2693–2698.
2] For some recent reviews, see: a) X. Just-Baringo, D. J. Procter, Acc.
Chem. Res. 2015, 48, 1263–1275; b) M. Szostak, M. Spain, D. J. Procter,
Received: August 6, 2015
Published online on November 3, 2015
Chem. Eur. J. 2015, 21, 18394 – 18400
www.chemeurj.org
18400
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim