Mechanistic insights into boron-catalysed direct amidation reactions

Article abstract of DOI:10.1039/c7sc03595k

The generally accepted monoacyloxyboron mechanism of boron-catalysed direct amidation is brought into question in this study, and new alternatives are proposed. We have carried out a detailed investigation of boron-catalysed amidation reactions, through study of the interaction between amines/carboxylic acids and borinic acids, boronic acids and boric acid, and have isolated and characterised by NMR/X-ray crystallography many of the likely intermediates present in catalytic amidation reactions. Rapid reaction between amines and boron compounds was observed in all cases, and it is proposed that such boron-nitrogen interactions are highly likely to take place in catalytic amidation reactions. These studies also clearly show that borinic acids are not competent catalysts for amidation, as they either form unreactive amino-carboxylate complexes, or undergo protodeboronation to give boronic acids. It therefore seems that at least three free coordination sites on the boron atom are necessary for amidation catalysis to occur. However, these observations are not consistent with the currently accepted 'mechanism' for boron-mediated amidation reactions involving nucleophilic attack of an amine onto a monomeric acyloxyboron intermediate, and as a result of our observations and theoretical modelling, alternative proposed mechanisms are presented for boron-mediated amidation reactions. These are likely to proceed via the formation of a dimeric B-X-B motif (X = O, NR), which is uniquely able to provide activation of the carboxylic acid, whilst orchestrating the delivery of the amine nucleophile to the carbonyl group. Quantum mechanical calculations of catalytic cycles at the B3LYP+D3/Def2-TZVPP level (solvent = CH₂Cl₂) support the proposal of several closely related potential pathways for amidation, all of which are likely to be lower in energy than the currently accepted mechanism.

Full text of DOI:10.1039/c7sc03595k

View Article Online
View Journal
Chemical
Science
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: S. Arkhipenko, A.
Batsanov, M. Sabatini, V. Karaluka, T. Sheppard, H. Rzepa and A. Whiting, Chem. Sci., 2018, DOI:
10.1039/C7SC03595K.
This is an Accepted Manuscript, which has been through the
Volume
7 Number 1 January 2016 Pages 1–812
Royal Society of Chemistry peer review process and has been
accepted for publication.
Chemical
Science
Accepted Manuscripts are published online shortly after
www.rsc.org/chemicalscience
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 2041-6539
EDGE ARTICLE
Francesco Ricci et al.
Electronic control of DNA-based nanoswitches and nanodevices
rsc.li/chemical-science

Please do not adjust margins
Chemical Science
Page 1 of 17
View Article Online
DOI: 10.1039/C7SC03595K
Chemical Science
ARTICLE
Received 00th January
20xx,
Mechanistic insights into boron-catalysed direct amidation
reactions
Sergey Arkhipenko,a Marco T. Sabatini,b Andrei S. Batsanov,c Valerija Karaluka,b Tom D.
Sheppard,b Henry S. Rzepad and Andrew Whitinga*
Accepted 00th January 20xx
DOI: 10.1039/x0xx00000x
www.rsc.org/
The generally accepted monoacyloxyboron mechanism of boron-catalysed direct amidation is brought into question in this
study, and new alternatives are proposed. We have carried out a detailed investigation of boron-catalysed amidation
reactions, through study of the interaction between amines/carboxylic acids and borinic acids, boronic acids and boric
acid, and have isolated and characterised by NMR/X-ray crystallography many of the likely intermediates present in
catalytic amidation reactions. Rapid reaction between amines and boron compounds was observed in all cases, and it is
proposed that such boron-nitrogen interactions are highly likely to take place in catalytic amidation reactions. These
studies also clearly show that borinic acids are not competent catalysts for amidation, as they either form unreactive
amino-carboxylate complexes, or undergo protodeboronation to give boronic acids. It therefore seems that at least three
free coordination sites on the boron atom are necessary for amidation catalysis to occur. However, these observations are
not consistent with the currently accepted 'mechanism' for boron-mediated amidation reactions involving nucleophilic
attack of an amine onto a monomeric acyloxyboron intermediate, and as a result of our observations and theroetical
modelling, alternative proposed mechanisms are presented for boron-mediated amidation reactions. These are likely to
proceed via the formation of a dimeric B-X-B motif (X=O, NR), which is uniquely able to provide activation of the carboxylic
acid, whilst orchestrating the delivery of the amine nucleophile to the carbonyl group. Quantum mechanical calculations of
catalytic cycles at the B3LYP+D3/Def2-TZVPP level (solvent=CH2Cl2) suppport the proposal of several closely related
potential pathways for amidation, all of which are likely to be lower in energy than the currently accepted mechanism.
Introduction
Controlling boron in catalysis
Boron compounds have a rich, fascinating and complex
structural chemistry, especially related to clusters, cages,
complexes and anions associated with BꢀO1 and BꢀN2
containing systems. Indeed, such systems are well known to
“have a distinct tendency to disproportionate, apparently more
so than those of any other nonmetallic elements.”1a This
observation means that identifying and controlling the role
played by boron in reactions and catalytic processes is
particularly challenging and has proved so on many occasions.
Elegant single and cooperative solutions, exemplified in
Scheme 1, have enabled the precise reactivity control of boron
in different reactions. For example, in the catalytic asymmetric
Single centre B-catalyst
Cooperative B-X-B catalyst
vs.
B
B
B
B. 'Tetra-acetyl diborate'
A. 'Triacetoxyborate'
AcO
OAc
O
B
O
O
B
O
B
AcO
O
OAc
OAc
originally
proposed
actual
structure
Scheme 1. General structural role of boron in different catalytic processes
CBS reduction system3 both nitrogen and oxygen ligands
control boron coordination to tune an external boron to behave
cooperatively to control ketone Lewisꢀacid binding, activation
and hence, reduction. Similarly, precise boron centre control
can
be
employed
to
achieve
highly
efficient
hydroxynaphthaquinone chiral Lewisꢀacid binding, activation
and subsequent asymmetric DielsꢀAlder cycloaddition;4 the
boron control and prevention of diborateꢀdisproportionation
deriving from use of hindered binol ligands, assembled by
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 1
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 2 of 17
View Article Online
DOI: 10.1039/C7SC03595K
ARTICLE
Journal Name
careful manipulation of unstable acyloxyboron species. Indeed,
the model for these boron complexes was derived from
‘triacetoxyborate’ complexes of chiral juglone derivatives.
However, triacetoxyborate is not stable at ambient temperatures
and it rearranges to give tetraꢀacetyl diborate, as confirmed by
Xꢀray crystallography.5 These seminal works led us to conclude
that reaction mechanisms involving boron are nonꢀtrivial,
because the structural chemistry of boron is itself nonꢀtrivial.
It is generally the case that in the absence of a catalyst,
amines and carboxylic acids do not usually react to form
amides at ambient temperatures, and forcing conditions are
often required to promote this condensation reaction to access
amides.6 Although a direct thermally accessible reaction is
possible, direct amidation is better facilitated at lower
temperatures by using catalysts,7 with boronꢀbased systems
being particularly effective.8 Useful catalysts for direct
amidation include boric acid,9 borate esters10 and boronic
acids.11ꢀ13 Borinic acids have also been claimed to promote
direct amidation reactions under relatively mild conditions.14
Clearly, an understanding of the mechanism of these boronꢀ
mediated amidations is crucial for designing catalysts with
improved activity, but more importantly,
a fundamental
understanding of the role of the boron in such systems is
required. To this end, we have been interested in the structural
and mechanistic chemistry associated with boron compounds
employed in direct amidation for several years.12a The diversity
of the catalytic systems reported to date, from simple boric
acid9 or boronic acids11ꢀ13 to carefully designed cooperative
catalysts,15 including polyꢀboronꢀbased systems,16 poses a
challenge in terms of identifying key structural motifs which
are necessary for effective catalysis. Even considering boronic
acid catalysts alone, bifunctional systems containing nitrogen
and boron,12 or iodine/bromine and boron13 and electron
deficient aryl boronic acids11 have all been demonstrated to be
highly effective. Currently, a comprehensive understanding of
how the catalyst structure affects the activity is clearly lacking.
Previously, an acyloxyboronꢀbased intermediate was
proposed as the reactive catalytic species,11a with computational
studies suggesting that it was energetically plausible (Scheme
2A).17 As shown in Scheme 2, this mechanism involved the
formation of the acyloxyboron derivative which then had to
undergo nucleophilic attack by an amine to give the amide,
after loss of water. However, despite the fact that DFT
calculations17 suggest that acyloxyboron species are feasible
intermediates in direct amidation catalysis, actual definitive
physical evidence of the mechanistic role of boron in such
reactions is sparse and could be misleading due to the potential
for numerous possible boron species present in such
reactions.12a Indeed, it should be noted that acyloxyboron
species are particularly prone to dimerization/oligomerization,
as mentioned above with regard to B(OAc)3 (vide supra and
Scheme 1).
Scheme 2. A) Currently proposed mechanism via
a monomeric acyloxyboron
intermediate. B) Summary of mechanisms proposed in this paper involving dimeric B-X-
B motifs.
In this work, we report recent studies in which we have
probed the types of species potentially involved in boronꢀ
catalysed amidation reactions using
experimental, structural and theoretical approach. As a result,
we are postulate that acyloxyboron species like (Scheme 2A)
a
combination of
I
are unlikely to be sufficiently reactive as acylating agents, and
we propose new mechanistic pathways for boronꢀmediated
amidation which proceed through BꢀXꢀB systems, as shown in
Scheme 2B. We also demonstrate their feasibility
experimentally and supported by computational methods.
Indeed, several routes for catalytic amidation are suggested, all
of which were calculated to have lower energy barriers than the
previously postulated single boron atom mechanisms.17
Importantly, they involve intermediates with two boron centres,
and take account of the possible role of both boronꢀnitrogen and
boronꢀoxygen interactions, which can be readily observed
between amines and boronic or borinic acids, and between
these boron compounds and carboxylic acids. These not only do
these observations point to a generic structural motif which can
potentially explain the success of cooperative,15 bifunctional12a
and polyboron amidation catalysts,16 but they may even explain
catalysis in other, related reactions.18
Results and Discussion
Observations on the acyloxy mechanism
In recent computational studies,17 both tetrahedral17a and
trigonal17b acyloxyboron species were considered as possible
active acylating agents, with the latter leading to a lower energy
barrier for the reaction with an amine. There is little
2 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 3 of 17
View Article Online
DOI: 10.1039/C7SC03595K
Journal Name
ARTICLE
experimental evidence for these proposed intermediates,11a and
Our first finding was that amines (e.g. benzylamine and
only
a few classes of wellꢀcharacterised acyloxyboron ethylenediamine) readily coordinate to borinic acids 3aꢀc
compounds have been reported.4,5,18 It is also not clear whether (freshly obtained from their corresponding ethanolamine
acyloxyboron compounds are effective acylating agents for complexes, see ESI) with formation of the corresponding “ate”ꢀ
amines; certainly, trigonal acyloxyboron compounds are likely complexes
to be Lewis acidic, and therefore, reaction of an amine at the observed by 11B NMR via a signal shift from δ 40ꢀ45 ppm for
boron atom should have a much lower energy barrier than the the borinic acids to a peak in the “tetrahedral” region around
4 (see Scheme 4). Complexes of type 4 were
3
alternative nucleophilic attack at the carbonyl leading to amide 2ꢀ6 ppm. Importantly, this diagnostic shift was observed in all
formation. Indeed, readily accessible acyloxyboron compound cases, and interestingly, when bisꢀ3,4,5ꢀtrifluorophenylborinic
1a19a reacted cleanly with benzylamine to give the ‘ate’ꢀ acid 3a was employed, the adducts 4a and 4b crystallised
complex
2 (Scheme 3). Even in the presence of excess amine, directly from the reaction mixtures allowing unambiguous
no conversion to amide was observed. An investigation of structural verification by Xꢀray crystallography (see Fig. 1).
previously reported tetrahedral acyloxyboron species had
Ar
Ar
OH
R-NH2
suggested that they were also ineffective as acylating agents for
amines. For example, borinate 1b, generated by reaction of
triethylborane with glycine, was recovered unchanged after
heating at reflux in toluene with benzylamine,19b and
preliminary work in our laboratories found that MIDAꢀboronate
1c also showed no reactivity towards amines under mild
conditions (others have reported related observations20) which
clearly demonstrate that such species are not effective acylating
agents. These results contrast with the mechanism examined
B
B
OH
R
N
H2
CDCl3, r.t., 5 min
Ar
Ar
11B NMR 2-6 ppm]
[
11B NMR40-45 ppm]
[
3
4
a; Ar = 3,4,5-(C6F3H2),
a; Ar = 3,4,5-(C6F3H2)
b; Ar = C6H5
c; Ar = 2-(C6ClH4)
d; Ar = 2-(C6IH4)
R = CH2Ph
b; Ar = 3,4,5-(C6F3H2),
R = CH2CH2NH2
RCOOH
PhCH2NH2
CH2Cl2/CDCl3,
r.t., 5 min
RCOOH, CH2Cl2 or
CDCl3, r.t., 5 min
computationally by Marcelli in which
a
tetrahedral
acyloxyboronate was proposed to undergo attack by an amine
in a reaction that typically takes place at room temperature.17a
Ar
Ar
O
H
R
B
N
H
O
rt or heating
O
1-2 eq
Ph NH2
O
Ph
Ph
O
O
B
O
B
Ph
Ph
X
R
N
H
Ph
[
11B NMR 1-3 ppm]
O
Ar = Ph or
2-ClC6H4
Ar
O
5
H2N
Ar
a; Ar = 3,4,5-(C6F3H2), R = (CH2)3Ph
b; Ar = 3,4,5-(C6F3H2), R = Ph
1a
2
[
33 pmm]
[
9 pmm]
δB
δB
No Amide
O
Scheme 4. Reactions between borinic acids 3 and amines to give complexes 4, and
subsequent conversion into aminocarboxylate complexes 5, which are not readily
converted into amides.
Ph
NH2
O
NH2
No amide formation
[ref 19b]
B
Toluene,
reflux, 2 h
Et Et
1b
O
O
N
Ph
NH2
B
< 5% Amide
O
CD3CN, r.t., 48 h
tBu
O
1c
Scheme 3. Monomeric acyloxyboron species typically show low reactivity towards
amines.
Borinic acids in direct amide formation
Analysis of carboxylic acid/amine/boronic acid reaction
systems is a complicated task because multiple equilibria
coexist in such reaction mixtures,12a leading to the presence of
many different species. Hence, in order to better understand the
potential catalytic role of boronic acids in direct amidation, we
examined related diarylborinic acids as model systems.
Interestingly, in our hands,14 these compounds were found to be
ineffective, unreactive direct amidation catalysts, and therefore,
ideal for probing the structural effects of coordination versus
reaction with both carboxylic acids and amines.
Figure 1. X-ray structures of Lewis adducts 4a (H-bonded dimer) and 4b.
Borinic acidꢀamine complexes
carboxylic acids (e.g. benzoic acid and 4ꢀphenylbutyric acid)
resulting in amineꢀcarboxylic acid complexes (Scheme 4).
The same complexes were also formed if all three
4 were found to react with
5
5
components (borinic acid, amine and carboxylic acid) were
mixed together simultaneously (see ESI for related reactions
and structures from the reactions of 3b and 3c). In many cases,
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 3
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 4 of 17
View Article Online
DOI: 10.1039/C7SC03595K
ARTICLE
Journal Name
these closely related products were also unambiguously
Having observed that borinic acids do not mediate amide
characterised by Xꢀray crystallography (see Fig. 2). It should be bond formation, we were surprised by a recent report describing
noted that complexes
are structurally closely related to the borinic acidꢀcatalysed amide formation.14 However, when we
5
acyloxyboron intermediate proposed for direct amidation attempted amidations under exactly the same conditions to
reactions (Scheme 2A), with the NH2 unit potentially able to those recently reported,14 amides could indeed be obtained in
hydrogen bond to the carbonyl oxygen to provide activation. good yield; an explanation for which was required.
Most importantly, however, complexes
5 could not be
Ar
B
converted to amides either by heating, or by addition of further
aliquots of acid, amine or base; a result which strongly
corroborates our hypothesis that, in general, singleꢀboron
acyloxy compounds are not competent acylating agents for
amines.
R
RCOOH [R = Ph or
-(CH2)3COOH]
O
B
O
B
Ar
Ar
O
B
O
B
6 days
Ar
Ar
B
OH
Ar
O
Ar
CDCl3, r.t., 5 min
Ar
O
H
7a
Ar
3a
Ar = 3,4,5-(C6F3H2)
+
ArB(OH)2
6a
6b
R = Ph
8a
R = -(CH2)3Ph
[M-H]m/z 725.1340
[calcd. 725.1343]
Scheme 5. Reaction of 3,4,5-trifluorophenylborinic acid 3a with 4-phenylbutyric or
benzoic acid, resulting in boroxine formation.
Figure 2. X-ray structures of amino-carboxylate borinic complexes 5a (showing
a
hydrogen-bonded dimer of two independent molecules) and 5b (disorder of phenyl
ring is omitted for clarity). For 5c and 5d see ESI.
In contrast, when bisꢀ3,4,5ꢀtrifluorophenylborinic acid 3a
was reacted with carboxylic acids alone, immediate formation
of a boron “ate”ꢀcomplex was observed according to 11B NMR
(Fig. 3), as shown in Scheme 5, and a subsequent slow
protodeboronation reaction occurred which generated a boronic
species (see Fig. 3). This was followed, in this case, by slow
crystallisation of the corresponding boroxine 7a (Fig. 4).
Hence, we propose that the “ate”ꢀcomplex observed by 11B
Figure 3. 11B NMRs of a mixture of bis-3,4,5-trifluorophenylborinic acid 3a with benzoic
acid after 5 min and after 6 days upon mixing (Ar = 3,4,5-trifluorophenyl).
NMR was
6 (Scheme 5 and Fig. 3) based on systems of this
type which have been isolated.21 In addition, the mass spectrum
of the mixture showed an m/z 725.1340 (calcd. m/z for 6a [Mꢀ
H] 725.1343), which links with the observation of
protodeboronationꢀdehydration
to
give
the
notably
uncomplexed boroxine 7a, and its corresponding boronic acid
8a. The fact that such reactions occur readily at room
temperature shows that formation of BꢀOꢀB species is
extremely facile. Both borinic and boronic acids can thus
readily form bridged BꢀOꢀB systems such as
6 and 7 (and 11,
Figure 4. X-ray structure of 3,4,5-trifluorophenylboroxine 7a in orthorhombic β-form at
120 K.
vide infra) which demonstrates that such motifs play an
important role, both in boron chemistry in general, and likely
also in direct amidation catalysis. It also exemplifies the ease
with which a borinic acid undergoes protodeboronation in the
presence of a carboxylic acid.
4 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 5 of 17
View Article Online
DOI: 10.1039/C7SC03595K
Chemical Science
ARTICLE
A)
Ph
Simultaneous mixing
O
O
N
Ar
10 mol% 3c,
5 Å MS, CH2Cl2
PhCH2NH2
B
Ph
Ar
A)
H2
B)
5c; Ar = 2-(C6ClH4)
O
10 mol% catalyst
Ph
Ph
B)
N
Ph
CO2H
H
5 Å MS, CH2Cl2
15 min "prestir",
then PhCH2NH2
3c
90% conv. ( catalyst)
91% conv. (8c catalyst)
C)
Cl
C)
Cl
B OH
5 Å MS, CH2Cl2
15 min, r.t.
Cl
3c
B
HO
OH
Filter & Concentrate
8c
Dissolve in CDCl
3
Scheme 6. Examining bis-2-chlorophenylborinic acid 3c as a catalyst for direct amide formation and 11B NMR spectra of product mixtures obtained in 3 examined cases.
After further investigations, we discovered that the key to chlorophenylborinic acid 3c and phenylacetic acid were stirred
catalytic reactivity was the 15 minute “prestir” period, as for 15 minutes with 5 Å MS, leading to a 1:6 ratio of borinic to
explained in Scheme 6. When the claimed catalyst, bisꢀ2ꢀ boronic species (Scheme 6C). Interestingly, the process of
chlorophenylborinic acid 3c, was exposed to a carboxylic acid protodeboronation occurred much more slowly when activated
in the presence of activated 5 Å molecular sieves (MS), 4 Å MS were used compared to 5 Å MS, so the particular
catalytic activity was observed. Without this “prestir” period, sieves used plays an important role in assisting the
no amide formation occurred. The significant finding was that deboronation reaction (it is worth noting that generally either 3
the 11B and 1H NMR spectra of the crude reaction mixtures or 4 Å MS are used for achieving dehydration in most
were different depending on whether the “prestir” period was amidation reactions as an alternative to azeotropic water
applied or not (Scheme 6A and B). This data indicated that if removal. See ref. 13b for a relevant discussion).
borinic acid 3c, phenylacetic acid and benzylamine were mixed
To further support the fact that boronic acid 8c is actually
simultaneously without the “prestir” period, aminocarboxylate formed from borinic acid 3c in situ under the “prestir”
complex 5c was formed and remained stable over at least one conditions and that 8c is the active catalyst, we compared 3c
week in solution (vide supra, Scheme 4 and Scheme 6A). When and commercial 2ꢀchlorophenylboronic acid 8c sideꢀbyꢀside
analyzing the product mixtures obtained after the “prestir” under the same conditions (Scheme 6B). Similar amide
period, significant amounts of boronic species were observed conversions were observed in the two reactions with 3c and 8c
(Scheme 6B). We propose, therefore, that borinic acid 3c acts (91% and 90% respectively), which supports our proposal that
as a preꢀcatalyst, and its protodeboronation occurs rapidly the boronic acid is the actual catalyst. It can, therefore, be
during the “prestir” period to give 2ꢀchlorophenylboronic acid concluded that borinic acids do not catalyse amidations
8c, which is well known to be an effective catalyst for direct directly, and can only do so through protodeboronation to the
amidation, as reported by Hall et al..13a This finding was further boronic acid.
verified by
a
separate experiment in which bisꢀ2ꢀ
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 5
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 6 of 17
View Article Online
DOI: 10.1039/C7SC03595K
ARTICLE
The fact that aminocarboxylate complexes
Journal Name
5
could not be
converted into amides seems to be rather inconsistent with the
commonly accepted mechanism for boronꢀmediated amidation
(Scheme 2A). Since borinic acids are stronger Lewis acids than
boronic acids, it might be expected that acyloxyborinic species
5
should actually be stronger acylating agents than the
acyloxyboronic analogues, and these complexes also have the
ability to form internal hydrogen bonds between the protonated
amine and the carbonyl. Since both boronic acids and boric acid
catalyse direct amidation,8 but borinic acids do not, it implies
that at least three, free coordination sites are needed on the
boron atom in order for amidation catalysis to be possible. This
conclusion led us to consider, therefore whether BꢀXꢀB dimer
formation might be essential for the observed catalytic activity,
and that a bridging Xꢀgroup (i.e. O or N) could be essential to
orchestrate the catalytic process. We therefore looked more
closely at the speciation that might be occurring in boronic
acidꢀmediated direct amidation catalysis, using both
experimental and theoretical methods.
Scheme 8. Reactions between phenylboronic acid 8a and benzylamine at reflux.
These examples further underlined (also observed for borinic
acids, vide supra) the importance of BꢀN interactions, which
should readily occur between the amine and any trigonal boron
species present in catalytic amidation systems. This is also
consistent with our hypothesis that any trigonal acyloxyboron
species is likely to react at the boron atom with an amine and
indeed, such interactions may facilitate interconversion to and
from boroxine species.
Boronic acids in direct amide formation
Having investigated the various interactions between
amines and boronic acids, we next examined reactions between
carboxylic acids and boronic acids. It was observed that boronic
acid 8c did not react with phenylacetic acid unless 5 Å MS
were introduced into the reaction mixture. In the presence of 5
Due to the complexity of carboxylic acid/amine/boronic
acid systems, the reactivity of boronic acids with amines and
carboxylic acids was initially explored separately. When treated
with an amine, boronic acids appear to form complexes of
general structures
9 . 3,4,5ꢀtrifluorophenyl boroxineꢀ
and 10 20
Å
MS alone, boronic acid 8c was converted into the
benzylamine complex 10a (Fig. 5) and the adduct with
ethylenediamine 10b (see ESI for Xꢀray) were formed at room
temperature and crystallised (Scheme 7), while phenylboroxine
complexes 10c and 10d with two or four benzylamine
molecules (Scheme 8 and Fig. 5) formed upon reaction of 8a
with two equivalents of benzylamine at reflux in toluene (either
with reflux condenser or with DeanꢀStark water removal
apparatus). An important observation was that the amine
promotes dissolution of the boronic acid in organic solvent by
formation of these boroxine complexes (see ESI).
corresponding boroxine 7c (Fig. 7AꢀB). With carboxylic acid
present, inital formation of boroxine 7c occurred (Fig. 7C), and
as more MS were added as well as phenylacetic acid, a new
“ate”ꢀcomplex (
δ 5 ppm, Fig. 7DꢀE) was formed, and the
equilibrium between it and boroxine 7c was pushed further
towards the new species. Careful crystallisation from this
heterogeneous system (containing solid 5 Å MS), allowed
crystals to be separated which were suitable for Xꢀray
characterisation. The product was unambiguously identified as
the “ate”ꢀcomplex 11c (see Fig. 6) which is structurally related
to analogous complexes reported elsewhere.22 Interestingly, in a
separate NMR reaction, addition of an excess of phenylacetic
acid to boronic acid 8c, in the presence of 5 Å MS shifted the
equilibrium even further towards the “ate”ꢀcomplex as seen by
11B NMR (see further examples in the ESI).
Scheme 7. Reactions between boronic acids 8a-d and amines at r.t.
6 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 7 of 17
View Article Online
DOI: 10.1039/C7SC03595K
Journal Name
ARTICLE
Figure 5. X-ray structures of 3,4,5-trifluorophenylboroxine-benzylamine adduct 10a and
phenylboronic acid derived boroxine-amine adducts 10c and 10d.
Figure 6. X-ray structures of B-O-B “ate”-complexes 11c and d.
Figure 7. 11B NMR spectra showing interactions between 2-chlorophenylboronic acid 8c, phenylacetic acid and 5 Å MS. A: 2-chlorophenylboronic acid; B: addition of 5 Å MS to (A),
leading to boroxine formation; C: addition of a tiny amount of 5 Å MS to a mixture of boronic and carboxylic acid; D: addition of more 5 Å MS; E: addition of even more 5 Å MS.
The amount of MS used also had a profound effect, and acid derivatives rather than act as amidation catalysts. This
with a large excess of MS it was possible to generate >85% of again highlights the need for the correct boron oxidation state,
the “ate”ꢀcomplex in the mixture (calculated by 11B NMR, see so that there are sufficient free coordination sites on the boron
Fig. 7E and ESI for further examples). The resulting species centre: borinic acids being unreactive but boronic acids/boric
11c is analogous to “ate”ꢀcomplexes
6 in the borinic acid series, acid being catalytically active.
which undergo subsequent protodeboronation to give boronic
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 7
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 8 of 17
View Article Online
DOI: 10.1039/C7SC03595K
ARTICLE
Journal Name
It should be noted that without 5 Å MS the formation of the acid, the “ate”ꢀcomplex was formed, as shown by 11B NMR,
diacyloxy BꢀOꢀB bridged complexes 11 did not occur to a but did not crystallise and only boroxine 7a crystallised from
significant extent; even when 4 Å MS were used, less than 5% the reaction mixture. In the case of orthoꢀsubstituted boronic
of the diacyloxy BꢀOꢀB complex 11c was observed. acids, the boroxines are likely to be destabilized due to steric
Interestingly though,
a similar diacyloxy BꢀOꢀB bridged effects, and thus, the formation of dimeric “ate”ꢀcomplexes 11
complex 11d could be isolated by reaction of 2ꢀ may be more favourable, enabling the formation of catalytically
iodophenylboronic acid 8d and phenylacetic acid. In contrast, competent species.
when boronic acids without an orthoꢀsubstituent, such as 3,4,5ꢀ
trifluorophenylboronic acid 8a, were reacted with phenylacetic
Figure 8. 11B NMRs of A): mixture of 2-chlorophenylboronic acid 8c and phenylacetic acid in the presence of 5 Å MS resulting in equilibrium between B-O-B dicarboxylate 11c (δ 5
ppm) and boronic species 7c (δ 30 ppm); B): addition of 1 equivalent (per “ate”-complex 11c) of benzylamine, 2.5 h, showing formation of a new complex 12; C): Addition of 2nd
equivalent of benzylamine, 20 min; D): Same mixture over 48 h.
For the former, this is due to reduced conjugation between the
aryl rings and the boroxine ring as a result of twisting of ~11°.
At 298 K, this translates into a ~105ꢀfold decrease in the
Ar
O
Ar
B
B
-3H2O
Ar
OH
8c/7c ꢀG298 +7.50 kcal mol-1
8e/7e +4.95 kcal mol-1
B
O
O
G
ꢀ
298
+3H 2O
B
OH
equilibrium concentration of the boroxine 7. Indeed, the high
[Calculated Values]
Ar
catalytic activity of such orthoꢀsubstituted systems may be due
to these key reactivity effects which can destabilise the
boroxine, a species which is likely to be an offꢀcycle resting
state of the catalyst.
8c Ar=2-ClC6H4
8e Ar=4-ClC6H4
7c
7e
Scheme 9. Calculated ꢀG298 values for the formation of boroxines from o-
chlorophenylboronic 8c acid and p-chlorophenylboronic acid 8e.
Interestingly, when we attempted to reproduce the
reported10 generation of an “acyloxy” boron derivative as
described by Ishihara et al. using 3,4,5ꢀtrifluorophenylboronic
acid 8a heated at reflux with phenylacetic acid in toluene
(DeanꢀStark water removal), there was no obvious reaction by
either 11B and 1H NMR or ReactIR.
Having isolated the diacyloxy BꢀOꢀB complexes 11, we
attempted to react these systems with benzylamine (Figure 8).
Surprisingly, when 1 equivalent of amine per dicarboxylate
Such stericallyꢀinduced destabilization of boroxines was
indeed confirmed by relative free energy calculations26
(B3LYP+D3/Def2ꢀTZVPP/solvent
which cyclisation
=
dichloromethane) in
of
o
ꢀchlorophenyl boronic acid 8c to give 7c + 3H2O was found to
be 2.55 kcal/mol more endoenergic than formation of the
boroxine 7e from ꢀchlorophenyl boronic acid 8e (Scheme 9).
p
8 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 9 of 17
View Article Online
DOI: 10.1039/C7SC03595K
Journal Name
ARTICLE
complex 11c was added, the 11B NMR showed the appearance
of a new species 12 with a signal at 3 ppm, but 1H NMR
showed no amide formation even after 2.5 h (see ESI).
However, when a second equivalent of amine was added, amide
formation was seen within 20 minutes. This reaction of the
diacyloxy BꢀOꢀB bridged complex 11c with two equivalents of
benzylamine did not reach completion, and the signals we
1
assign to the new species 12 remained in the H NMR even
after 48 h. It is possible that this is due to the fact that there is
an excess of boron compound (in total) with respect to amine in
the reaction mixture, which indicates that an excess of amine
and/or carboxylic acid per catalyst molecule is required for
catalytic turnover and efficient reactivity.
We then went on to investigate the species present in
catalytic reaction mixtures containing boronic acid, amine and
carboxylic acid (Scheme 10). Interestingly, the behavior of
these systems was dependent on the order of reagent addition.
When the carboxylic acid was premixed with boronic acid 8c in
the presence of 5 Å molecular sieves, rapid formation of the
“ate”ꢀcomplex 11c was observed (Scheme 10A); subsequent
addition of amine led to complete formation of amide within 2
h. During the amidation reaction, small quantities of boroxineꢀ
amine complex 10e were also observed, along with a signal
Scheme 10. Observed boron-containing species in catalytic amidation reactions
containing boronic acid 8c, carboxylic acid and amine.
consistent with the unknown species 12 (δB 3 ppm). The amine
Summary of Observations
complex 10e became the dominant species after the amidation
was complete. In contrast, when boronic acid 8c was mixed
with the amine in the presence of 5 Å molecular sieves
(Scheme 10B), rapid formation of a boroxine amine complex
(e.g. 10e) was observed, though a lower 11B NMR shift of 15
ppm suggests that 10e was likely in equilibrium with a
compound analogous to 10d, in which two amine molecules are
coordinated to the boroxine. Addition of carboxylic acid led to
amide formation, but after 2 h the reaction had only reached
50% conversion. Again, during the amidation reaction both 10d
and 12 were observed, with the latter being the major species.
Finally, a reaction was carried out in which amine and
carboxylic acid were preꢀmixed to give the ammonium
carboxylate salt and boronic acid 8c was then added (Figure
10C). This exhibited similar behavior to the reaction shown in
Figure 10B, giving 53% conversion to amide after 2 h.
A full summary of possible intermediates generated from
borinic/boronic acids by reaction with carboxylic acids and
amines is shown in Scheme 11, along with likely pathways for
their formation. Thus, borinic acids
3 form inert complexes 5
(
via or postulated intermediate ) in the presence of amine
4
A
and carboxylic acid, but in the presence of carboxylic acid
alone, they can undergo protodeboronation (probably via
observed intermediate
Boronic acids undergo rapid complexation with amines to
give intermediate of type , which are rapidly converted
through to boroxineꢀamine complexes 10 via postulated
intermediate , facilitating dissolution of the boronic acid in
organic solvents. Direct condensation of a boronic acid with a
carboxylic acid should lead initially to , the previously
proposed reactive acylating agent.11a However, as we have
shown above, is likely to undergo selfꢀcondensation under
the dehydrating conditions to give “ate”ꢀcomplexes of type 11
in a similar fashion to the dimerization of B(OAc)3 (vide supra
Scheme 1). This latter complex could also potentially be
formed from , through condensation with two molecules of
6), to enter the boronic acid cycle.
8
9
B
8
C
C
Finally, when either B2O3 or boric acid were mixed with
benzylamine and phenylacetic acid (both insoluble in CDCl3),
some dissolution was observed and the resulting 11B NMR
showed the appearance of an “ate”ꢀcomplex (see ESI),
suggesting that these boron compounds can be pulled into
solution via complexation.
,
B
carboxylic acid. We tentatively propose that intermediate 12
may be the amineꢀbridged dimeric “ate”ꢀcomplex shown, which
is closely related to 11. This is based upon the fact that 12 is
formed by reaction of 11 with amine under dehydrating
conditions, and the fact that the NMR signals for 12 are similar
in line shape to those for complex
5 (see ESI). Under catalytic
conditions, a similar boron NMR peak is observed that may be
12, or a related compound derived from it.
Overall, these results lead to the conclusion that formation
of complexes from reaction of boron centres with both
carboxylic acids and amines (e.g.
5 and 12) is not only
observed at all three investigated boron oxidation levels
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 9
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 10 of 17
View Article Online
DOI: 10.1039/C7SC03595K
ARTICLE
Journal Name
(borinic, boronic and boric) but is also indicative that these Preliminary exploration of transition states involving
types of species are likely to be generated in catalytic amidation monomeric BꢀN species resulted in high overall energy barriers
reactions. It is also clear that BꢀOꢀB linkages and BꢀN (>30.0 kcal molꢀ1). Hence, pathfinder explorations of the
interactions are vitally important, and likely to be present in all potential energy surfaces were then carried out for the different
catalytically active boronꢀbased systems, i.e. particularly putative BꢀXꢀBꢀmediated catalytic cycles, organised around
boronic acids, and likely boric acid by extrapolation. As it is five possible routes based on the number of amines involved
not possible to be certain what other types of BꢀXꢀB systems (Figs. 9ꢀ13), in order to explore the mechanistic diversity of
can potentially form, hence, to more deeply understand the role these systems. Key transition state structures in each cycle are
of cooperativity resulting in direct amidation catalytic activity shown together in Fig. 14. To enable a high quality basis set to
between two boron atoms via a BꢀXꢀB system, we probed a be employed (Def2ꢀTZVPP), substituents were all modelled at
number of putative mechanistic pathways using quantum this stage using methyl groups. Each cycle included the reaction
mechanical modelling, all of which involve dimeric species free energies of the preꢀstep involving species used to assemble
potentially derived from 11 and/or 12
.
the reactant used as the start point in the cycle, and the energies
of all subsequent steps in the cycle, ending in the final products.
A final reaction free energy is included to reꢀgenerate the
starting reactant ready for a second catalytic cycle.
The cycle involving one amine is illustrated in Fig. 9. The
preꢀassembly condensation step (blue) indicates a modestly
endoenergic step (ꢀꢀG298 +9.4 kcal molꢀ1) producing water,
which agrees with the need for water removal using molecular
sieves for the reaction to proceed. With the reactant in the cycle
set to a relative energy of 0.0, the high point is reached with Cꢀ
O cleavage (TS3) corresponding to
corresponding to relatively slow ambient temperature
ꢀ
G298 +23.4 kcal molꢀ1,
a
reaction. The product in this cycle is +0.3 kcal molꢀ1 relative to
the reactant, but ꢀ4.2 kcal molꢀ1 is recovered in the next
regeneration step (red). Such a computational exploration can
only demonstrate the viability of a single pathway to product; it
does not preclude, of course, that lower energy pathways might
not be discovered upon a more complete exploration of the
potential energy surface. These variants might include lower
energy conformations of the various species explored. Rather
more certain is that the free energy barrier surmounted during
the cycle is probably an upper calculated bound to any true
value. It should also be noted that the model used for the
substituents (methyl) may introduce further uncertainty, but
since the purpose here was a broad exploration of various
mechanistic alternatives, we did not undertake the much more
computationally resourceꢀintensive recalculation using larger
substituents.
A cycle involving two amine species is illustrated in Fig.
10. The preꢀassembly step (blue) is rather more endoenergic
(ꢀ
G298 +25.4 kcal molꢀ1) than the previous cycle, but the
energetic maximum in the cycle (TS3,
relative to the assembled reactants is very similar. The product
of the first cycle is exoenergic by
G298 22.3 kcal molꢀ1, offset
in part by the regeneration step for the second cycle being
ꢀ
G298 +24.8 kcal molꢀ1)
Scheme 11. Boron species potentially generated from reaction of both borinic acids
and boronic acids with carboxylic acids and amines. Compounds which have been fully
characterized in this study by X-ray crystallography are shown in green, and those
which have been observed by NMR and/or mass spectrometry are shown in blue.
Proposed intermediates during the formation of the observed boron species are shown
in black.
ꢀ
ꢀG298 +16.5 endoenergic.
The threeꢀamine cycle (Fig. 11) again shows similar overall
behavior. The condensation reaction resulting in the initial
cycle reactant is endoenergic by
G298 +24.4 kcal molꢀ1,
matched by the exoenergic products (
G298 ꢀ24.8 kcal molꢀ1),
and with a regenerative step for the start of the second cycle
being endoenergic by
G298 +14.1 kcal molꢀ1. The rate limiting
TS3 has
G298 21.1 kcal molꢀ1, slightly more favourable than
the previous two cycles.
ꢀ
Mechanistic insights and theoretical calculations
ꢀ
We decided to formulate different diboron systems, linked
by oxygens or nitrogens and with different complexes,
including amine and carboxylate and their corresponding salts,
in order to have a better chance of locating the types of systems
that might be responsible for direct amidation catalysis.
ꢀ
ꢀ
10 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 11 of 17
View Article Online
DOI: 10.1039/C7SC03595K
Journal Name
ARTICLE
The fourꢀamine cycle (Fig. 12) has a highly endoenergic which forms exoenergically from the bicyclic reactant related to
G298 +37.9 kcal molꢀ1), again offset by the the previous cycles, but with a transition state of presumed low
condensation step (
ꢀ
products being exoenergic by 32.4 kcal molꢀ1. The regeneration barrier corresponding to reactant reorganization which we were
step for the second cycle is endoenergic by +24.9 kcal molꢀ1 unable to locate. This suggests that the barriers resulting from
and the overall barrier originating from TS1 is 25.4 kcal molꢀ1. species with BꢀNꢀB bridges are similar to those with BꢀOꢀB
This makes this model cycle less favourable.
bridges.
These five pathfinding explorations have thus identified a
Finally, a variation of the one amine cycle involving
utilization of a second amine to form a BꢀNꢀB catalytic model corresponding to a thermal reaction at 298K, with the
intermediate was examined (Figure 13), for which the catalyst caveat that further optimization of these very complex catalytic
assembly step is similarly endoenergic (
to the BꢀOꢀB bridged two amine route above. The highest free the future, we plan to exploit these core models to study
ꢀ
G298 +26.0 kcal molꢀ1) combinations may always reveal even lower energy routes. In
energy point in the catalytic cycle (TS2, ꢀ
G298 +23.2 kcal molꢀ substituent effects on the energies.
1) is now relative to the initial tetrahedral intermediate (TI1),
Figure 9. Catalytic cycle involving one amine (data26 sub-collection DOI: 14469/hpc/1854). The total computed free energy for each stationary point is shown in Hartree, with the
relative energies shown in parentheses in kcal mol-1. The DOI for data repository entries for individual species are shown in the form e.g. 10.14469/hpc/1885. Energies in blue are
for the pre-assembly stage, in black for the cycle proper and in red for the regeneration step to initiate a new cycle.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 11
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 12 of 17
View Article Online
DOI: 10.1039/C7SC03595K
ARTICLE
Journal Name
Figure 10. Catalytic cycle involving two amines (data26 sub-collection DOI: 14469/hpc/1622; the labels have the same meaning as in Fig. 9).
Figure 11 Catalytic cycle involving three amines (data26 sub-collection DOI: 14469/hpc/1621; the labels have the same meaning as in Fig. 9).
12 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 13 of 17
View Article Online
DOI: 10.1039/C7SC03595K
Journal Name
ARTICLE
Figure 12. Catalytic cycle involving four amines (data26 sub-collection DOI: 14469/hpc/1728;the labels have the same meaning as in Fig. 9).
Figure 13. Catalytic cycle involving B-N-N bridging group (data26 sub-collection DOI: 14469/hpc/3188; the labels have the same meaning as in Fig. 9).
All five of the catalytic cycles investigated have reasonable lower overall energetics than the monomeric acyloxyboron
thermal barriers for the catalytic mechanism (Fig. 14). It is pathway previously proposed11a and computationally
difficult, however, at least purely on the basis of the computed examined.17 Finally, on the basis of these calculations, the
energetics of such catalytic processes, to confidently resting state of the catalyst would be expected to be a bridged
discriminate between these. Importantly, all these systems have (BꢀOꢀB or BꢀNꢀB) system containing tetrahedral boron atoms
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 13
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 14 of 17
View Article Online
DOI: 10.1039/C7SC03595K
ARTICLE
Journal Name
(e.g. TI2 in Figure 13), consistent with the signals observed boronic acid catalyst provides some support for mechanisms
from NMR analysis of catalytic reaction mixtures (
δ
B 3 pmm).
involving an active species containing both of these
components, especially if its formation is dependent on BꢀN
interactions. Crucially, there appears to be a restriction on the
Kinetic analysis of a catalytic amidation reaction
Preliminary studies of the reaction kinetics for a boronic quantity of active catalyst that can be generated from free
acid catalyzed amidation reaction revealed some interesting boronic acid (somewhere between 5ꢀ10 mol%), which can be
findings (Scheme 12 and ESI). The selection of reaction lifted at higher amine concentrations.
conditions and amine/carboxylic acid reactants was guided by
the need for a homogenous reaction mixture in order to obtain
useful kinetic data. Reaction of 4ꢀphenylbutylamine with
benzoic acid12b was carried out in tertꢀamyl methyl ether
(TAME) under DeanꢀStark conditions10f using 8c as the
amidation catalyst. Analysis of reaction profiles was carried out
using the Burés method (see ESI).23 The reaction was first order
in catalyst up to a loading of 5 mol%, but a catalytic loading of
10 mol% did not provide a significant rate acceleration over 5
mol%. Interesting, the first order dependence on catalyst was
observed up to 10 mol% loading, when 1.4 equivalents of
amine were present. This may suggest that the amine is
important for generating the active catalytic species, perhaps
related to dimer 12. A positive order of reaction with respect to
the amine concentration was also observed at 3.5 mol% catalyst
loading. Unexpectedly, increasing the concentration of
carboxylic acid had a negative effect on the reaction rate, and
even a slight excess of 1.2 equivalents of carboxylic acid led to
almost complete inhibition of the reaction at 3.5 mol% catalyst
loading; the effect was less pronounced at higher catalyst
loadings.
O
O
cat. 8c
Ph
OH
Ph
Ph
TAME (0.25 M),
reflux, Dean-Stark
Ph
N
H
H2N
Cl
O
8c
Ph(CH2)4NH2
Ph
OH
B(OH)2
•First order (2.5-5 mol%)
•First order up to 10 mol%
with 1.4 eq amine.
•~0.5 order (with 3.5 mol%
catalyst loading)
•Negative order
•Excess acid shuts down
amidation reaction
Figure 14. The computed structure for: a) TS3 shown in Fig. 9; b) TS3 shown in Fig. 10;
c) TS3 shown in Fig. 11; d) TS1 shown in Fig. 12 and e) TS2 shown in Fig. 13. Indicated
bond length are in Å. The activation free energies ꢀG298 for the five transition states are
respectively 23.4, 25.2, 21.1, 25.4 and 23.2 kcal mol-1
.
Scheme 12. Preliminary kinetics studies on a boronic acid catalyzed amidation reaction.
At present, this observation is difficult to explain, but it
appears that excess carboxylic acid prevents formation of the
active catalytic species, e.g. via formation of offꢀcycle species
such as the ammonium carboxylate salt or dimer 11. Similar
observations regarding catalyst inhibition by carboxylic acid
were noted in Zrꢀcatalysed amidation reactions.24
All the mechanisms outlined above, as well as those
previously proposed in the literature, involve an essentially
unimolecular turnover limiting step, so reaction kinetics do not
provide a useful tool for distinguishing between them. The fact
that the reaction kinetics show a first order dependence on the
catalyst concentration is consistent with either a monomeric or
Summary and conclusions
We have investigated carboxylic acid/amine/boron catalyst
systems experimentally and computationally and have shown
the potential importance of both the formation of BꢀXꢀB
species and of BꢀN interactions in these systems, leading to the
proposal of new catalytic cycles for boronꢀcatalysed amidation
reactions involving boronic acids, though by extrapolation,
boric acid is highly likely to behave similarly. We have
demonstrated that whereas boronic acids can catalyse amidation
reactions, borinic acids are ineffective as catalysts themselves,
and in fact, undergo protodeboronation instead to derive
boronic acids in situ. In turn, the boronic acid analogue is
catalytically competent and causes downstream amidation
catalysis. This work clearly indicates that at least three free
dimeric active catalyst, provided it remains as
a
monomer/dimer throughout the catalytic cycle.23c,24 The
positive interaction between the concentration of amine and
14 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 15 of 17
View Article Online
DOI: 10.1039/C7SC03595K
Journal Name
ARTICLE
coordination sites are needed on the boron atom to enable coordinates (IRCs) were used to define the start and end species
effective catalysis to take place, and hence, borinic acids are of all transition states. Research data management (RDM) is via
precluded from amidation reactivity. Indeed, our proposed deposition into a collectionꢀbased data repository28 where
mechanisms can account for this and other observations access is via the appropriate FAIR data DOI.29
regarding amidation catalysis. Firstly, the formation of an
General experimental methods
active dimeric boron catalyst appears to be impossible at the
borinic acid oxidation level, as rapid complexation of an acid
General experimental details and data files are available via
a
and amine molecule leads to the formation of a catalytically
inactive complex . The success of orthoꢀsubstituted boronic
acids,13 including bifunctional systems,12 as catalysts can also
data repository.29 Selected experimental procedures and data
are included in the ESI.
5
potentially be rationalized by their preference for the formation Synthetic methods
of dimeric species over trimeric boroxines. Furthermore, this
All synthetic methods and compound characterization data is
supports our previous observation that cooperative catalysis
using two different boronic acids can lead to enhanced yields of
amides in challenging cases.15 Our proposed mechanisms are
also consistent with a recent report that a complex B/N/O
heterocyclic system displays high activity in catalytic amidation
reactions.16
available via a data repository or in the ESI.26
X-ray crystallography
Crystallographic and NMR full data files (in Mpublish format)
are available via a data repository.29
The findings in this study indicate the clear importance of
bringing together multiple techniques to understand the various
entities involved in amidation catalysis, i.e. the combined use of
NMR (1H, 13C and especially 11B), IR, Xꢀray crystallography,
reactivity studies and the importance of examining multiple
potential reaction pathways by theoretical methods, as
exemplified by the five related key transition states summarised
in Fig. 14. In this case, this comprehensive approach has
allowed us to move away from a seemingly 'accepted'
Acknowledgements
We thank Durham University for Doctoral Fellowship funding
(SA), GlaxoSmithKline and UCL Chemistry for supporting a
PhD studentship (MTS), and Pfizer for providing an EPSRC
CASE award (VK). We also thank Dr Dmitry S. Yufit
(Chemistry Department, Durham University) for solving Xꢀray
crystal structures of compounds 4a, 5c, 11d and βꢀ7a.
mechanism17 which arguably has little supporting experimental Notes and references
evidence,11a and to propose instead closely related low energy
alternatives. We consider that the mechanism shown in Fig. 13
1
a) M. F. Lappert, Chem. Rev., 1956, 56, 959–1064; b) G.
Heller, Topics in Current Chemistry, 1986, 131, 42-78.
M. D. Bosdet and W. E. Piers, Can. J. Chem., 2009, 87, 8-29.
E. J. Corey, R. K. Bakshi and S. Shibata, J. Am. Chem. Soc.,
1987, 109, 5551-5553.
may be the most likely to be effective under catalytic conditions
using boronic acids, due to our experimental observations of the
reactivity of dimer 11 with amine, and the direct 11B NMR
analysis of catalytic reaction mixtures. The key point though is
that this is a guide, and individual systems may change between
different modes of action depending upon both catalyst and
substrate properties (electronics, substituents, etc). However,
we anticipate that these mechanistic insights will provide
valuable assistance for the design of more active boron
amidation catalysts in the future.
2
3
4
5
T. R. Kelly, A. Whiting and N. S. Chandrakumar, J. Am. Chem.
Soc., 1986, 108, 3510–3512.
a) B. M. Trost, D. O'Krongly and J. L. Belletire, J. Am. Chem.
Soc. 1980, 102, 7595-7596; b) A. Dal Negro, L. Ungaretti and
A. Perotti, J. Chem. Soc., Dalton Trans., 1972, 1639-1643.
a) H. Charville, D. Jackson, G. Hodges and A. Whiting, Chem.
Commun., 2010, 1813-1823; b) H. Charville, D. Jackson, G.
Hodges and A. Whiting, M. R. Wilson, Eur. J. Org. Chem.,
2011, 5981–5990.
a) H. Lundberg, F. Tinnis, N. Selander and H. Adolfsson,
Chem. Soc. Rev., 2014, 43, 2714-2742; b) R. M. de
Figueiredo, J.-S. Suppo and J.-M. Campagne, Chem. Rev.,
2016, 116, 12029–12122.
a) R. M. Lanigan and T. D. Sheppard, Eur. J. Org. Chem., 2013,
7453-7465; b) K. Ishihara, Tetrahedron, 2009, 65, 1085-1109;
c) B. M. Monks and A. Whiting, Chapter 5, "Sustainable
Catalysis: Challenges and Practices for the Pharmaceutical
and Fine Chemical Industries," eds: P. J. Dunn, K. K. Hii, M. J.
Krische and M. T. Williams, Wiley, Hobokon, NJ, USA, 2013;
d) F. Ferdousi and A. Whiting, Chapter 13, "Green Catalytic
Direct Amide Bond Formation" in Green and Sustainable
Medicinal Chemistry: Methods, Tools and Strategies for the
21st Century Pharmaceutical Industry,” eds. L. Summerton,
H. F. Sneddon, L. C. Jones and J. H. Clark, RSC Green
Chemistry series, Cambridge, 2016.
6
7
8
Experimental
Computational methods
Pathfinder explorations of the potential energy surfaces for
putative catalytic cycles all utilised the Def2ꢀTZVPP basis set25
and the B3LYP density functional augmented with the Grimme
D3 dispersion energy correction,26 a combination selected after
earlier evaluation for catalytic cycles27 and as implemented in
the Gaussian 09, versions D01 and E.01 and Gaussian 16
version A.03 programs. A solvent correction using the CPCM
method and dichloromethane parameters was employed
throughout and all energies are corrected for thermal
contributions and entropy by calculation of normal mode
frequencies, with transition states exhibiting the required single
9
P. Tang, Org. Synth. 2005, 81, 262.
negative force constant. All free energies correspond to a 10 a) P. Starkov and T. D. Sheppard, Org. Biomol. Chem., 2011,
9
, 1320-1323; b) R. M. Lanigan, P. Starkov and T. D.
standard state at 298K of 0.0445 M (1 atm). Intrinsic reaction
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 15
Please do not adjust margins

Please do not adjust margins
Chemical Science
Page 16 of 17
View Article Online
DOI: 10.1039/C7SC03595K
ARTICLE
Journal Name
Sheppard, J. Org. Chem., 2013, 78, 4512-4523; c) V. Karaluka,
R. M. Lanigan, P. M. Murray, M. Badland and T. D. Sheppard,
Org. Biomol. Chem., 2015, 13 10888-10894; d) R. M.
Lanigan, V. Karaluka, M. Sabatini, P. Starkov, M. Badland, L.
Boulton and T. D. Sheppard, Chem. Commun., 2016, 52
8846-8849; e) T. Maki, K. Ishihara and H. Yamamoto, Org.
Lett., 2006, , 1431–1434; f) M. T. Sabatini, L. T. Boulton and
Research Computing Services data repository, 2016, DOI:
10.14469/hpc/1620 and sub-collections therein.
,
,
8
T. D. Sheppard, Sci. Adv., 2017, 3, e1701028.
11 a) K. Ishihara, S. Ohara and H. Yamamoto, J. Org. Chem.
1996, 61
Yamamoto, Macromolecules, 2000, 33, 3511–3513; c) K.
Ishihara, S. Ohara and H. Yamamoto, Org. Synth., 2002, 79
, 4196-4197; b) K. Ishihara, S. Ohara and H.
,
176–185; d) T. Maki, K. Ishihara and H. Yamamoto, Synlett
2004, 1355–1358; e) A. Sakakura, T. Ohkubo, R. Yamashita,
M. Akakura and K. Ishihara, Org. Lett., 2011, 13, 892–895; f)
A. Sakakura, R. Yamashita, M. Akakura and K. Ishihara, Austr.
J. Chem., 2011, 64, 1458–1465; g) R. Yamashita, A. Sakakura
and K. Ishihara, Org. Lett., 2013, 15, 3654–3657; h) K.
Ishihara and Y. Lu, Chem. Sci., 2016, 7, 1276–1280.
12 a) K. Arnold, B. Davies, R. L. Giles, C. Grosjean, G. E. Smith
and A. Whiting, Adv. Synth, Catal., 2006, 348, 813-820; b) K.
Arnold, A. S. Batsanov, B. Davies and A. Whiting, Green
Chem., 2008, 10, 124-134; c) K. Arnold, B. Davies, D. Hérault
and A. Whiting, Angew. Chem., 2008, 47, 2673-2676; d) S.
Liu, Y. Yang, X. Liu, A. S. Batsanov and A. Whiting, Eur. J. Org.
Chem., 2013, 5692-5700.
13 a) R. M. Al-Zoubi, O. Marion and D. G. Hall, Angew Chem. Int.
Ed., 2008, 47, 2876−2879; b) N. Gernigon; R. M. Al-Zoubi and
D. G. Hall, J. Org. Chem., 2012, 77, 8386-8400.
14 T. M. El Dine, J. Rouden and J. Blanchet, Chem. Commun.,
2015, 51, 16084–16087.
15 S. Liu, Y. Yang, X. Liu, A. S. Batsanov and A. Whiting, Eur. J.
Org. Chem., 2013, 5692-5700
16 H. Noda, M. Furutachi, Y. Asada, M. Shibasaki and N.
Kumagai Nature Chem., 2017, 9, 571-577.
17 a) T. Marcelli, Angew. Chem. Int. Ed. 2010, 49, 6840-6843; b)
C. Wang, H.-Z. Yu, Y. Fu and Q.-X. Guo, Org. Biomol. Chem.,
2013, 11, 2140-2146.
18 P. Zhang, R. M. Kriegel and J. W. Frost, ACS Sustainable
Chem. Eng. 2016, 4, 6991−6995.
19 a) P. Paetzold, P. Bohm, A. Richter and E. Scholl Z.
Naturforsch., 1976, 31b, 754–764; b) G. H. L. Nefkens and B.
Zwanenburg, Tetrahedron, 1983, 39, 2995-2998.
20 B. Garrigues and M. Mulliez, J. Organomet. Chem., 1986,
314, 19-24.
21 A. Schäfer, W. Saak, D. Haase and T. Müller, Angew. Chem.
Int. Ed., 2012, 51, 2981-2984.
22 a) B. Wrackmeyer, E. Khan and R. Kempe, Z. Naturforsch.,
2008, 63b, 275–279; b) R. Köster, A. Sporzynski, W.
SchüBler, D. Bläser and R. Boese, Chem. Ber., 1994, 127,
1191-1199.
23 a) J. Burés, Angew. Chem. Int. Ed., 2016, 55, 2028–2031; b) J.
Burés Angew. Chem. Int. Ed., 2016, 55, 16084–16087; c) J.
Burés, Top. Catal., 2017, 60, 631–633.
24 H. Lundberg, F. Tinnis, J. Zhang, A. G. Algarra, F. Himo, H.
Adolfsson, J. Am. Chem. Soc., 2017, 139, 2286–2295.
25 F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005,
7, 3297-3305
26 S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys.,
2010, 132, 154104-18.
27 A. Armstrong, R. A. Boto, P. Dingwall, J. Contreras-García, M.
J. Harvey, N. Mason and H. S. Rzepa, Chem. Sci., 2014, 5,
2057-2071.
28 M. J. Harvey, A. McLean and H. S. Rzepa, J.
Cheminform., 2017, 9:4.
29 S. Arkhipenko, A. S. Batsanov, M. T. Sabatini, V. Karaluka, T.
D. Sheppard, H. S. Rzepa and A. Whiting, Imperial College
16 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 17 of 17
Chemical Science
View Article Online
DOI: 10.1039/C7SC03595K
75x43mm (300 x 300 DPI)