Pyro-Borates, Spiro-Borates, and Boroxinates of BINOL - Assembly, Structures, and Reactivity

Article abstract of DOI:10.1021/jacs.7b02317

VANOL and VAPOL ligands are known to react with three equivalents of B(OPh)₃ to form a catalytic species that contains a boroxinate core with three boron atoms, and these have proven to be effective catalysts for a number of reactions. However, it was not known whether the closely related BINOL ligand will likewise form a boroxinate species. It had simply been observed that mixtures of BINOL and B(OPh)₃ were very poor catalysts compared to the same mixtures with VANOL or VAPOL. Borate esters of BINOL have been investigated as chiral catalysts, and these include meso-borates, spiro-borates, and diborabicyclo-borate esters. Borate esters are often in equilibrium, and their structures can be determined by stoichiometry and/or thermodynamics, especially in the presence of a base. The present study examines the structures of borate esters of BINOL that are produced with different stoichiometric combinations of BINOL with B(OPh)₃ in the presence and absence of a base. Depending on conditions, pyro-borates, spiro-borates, and boroxinate species can be generated and their effectiveness in a catalytic asymmetric aziridination was evaluated. The finding is that BINOL borate species are not necessarily inferior catalysts to those of VANOL and VAPOL but that, under the conditions, BINOL forms two different catalytic species (a boroxinate and a spiro-borate) that give opposite asymmetric inductions. However, many BINOL derivatives with substitutents in the 3- and 3′-positions gave only the boroxinate species and the 3,3′-Ph₂BINOL ligand gave a boroxinate catalyst that gives excellent inductions in the aziridination reaction. BINOL derivatives with larger groups in the 3,3′-position will not form either spiro-borates or boroxinate species and thus are not effective catalysts at all.

Full text of DOI:10.1021/jacs.7b02317

Subscriber access provided by McMaster University Library
Article
Pyro-borates, Spiro-borates and Boroxinates of
BINOL – Assembly, Structures and Reactivity
Gang Hu, Anil K. Gupta, Li Huang, Wenjun Zhao, Xiaopeng Yin, Wynter E. G.
Osminski, Rui H Huang, William D. Wulff, Joseph A Izzo, and Mathew J. Vetticatt
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b02317 • Publication Date (Web): 28 Jun 2017
Downloaded from http://pubs.acs.org on June 28, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 37
Journal of the American Chemical Society
1
2
3
4
5
6
Pyroꢀborates, Spiroꢀborates and Boroxinates of BINOL
7
8
9
– Assembly, Structures and Reactivity
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Gang Hu,†Anil K. Gupta,† Li Huang,† Wenjun Zhao,† Xiaopeng Yin,† Wynter E. G. Osminski,† Rui H.
Huang,† William D. Wulff,*† Izzo, J. A.;§ Vetticatt, M. J.*§
† Department of Chemistry, Michigan State University, East Lansing, MI 48824
§ Department of Chemistry, Binghamton University, Binghamton, NY 13902
wulff@chemistry.msu.edu
vetticatt@binghamton.edu
ABSTRACT. VANOL and VAPOL ligands are known to react with three equivalents of B(OPh)3 to
form a catalytic species that contains a boroxinate core with three boron atoms and these have proven to
be effective catalysts for a number of reactions. However, it was not known whether the closely related
BINOL ligand will likewise form a boroxinate species. It had simply been observed that mixtures of
BINOL and B(OPh)3 were very poor catalysts compared to the same mixtures with VANOL or VAPOL.
Borate esters of BINOL have been investigated as chiral catalysts and these include mesoꢀborates, spiroꢀ
borates and diborabicycloꢀborate esters. Borate esters are often in equilibrium and their structures can
be determined by stoichiometry and/or thermodynamics, especially in the presence of a base. The
present study examines the structures of borate esters of BINOL that are produced with different
stoichiometric combinations of BINOL with B(OPh)3 in the presence and absence of a base.
Depending on conditions, pyroꢀborates, spiroꢀborates and boroxinate species can be generated and their
effectiveness in a catalytic asymmetric aziridination was evaluated. The finding is that BINOL borate
species are not necessarily inferior catalysts to those of VANOL and VAPOL, but that under the
conditions, BINOL forms two different catalytic species (a boroxinate and a spiroꢀborate) that give
opposite asymmetric inductions. However, many BINOL derivatives with substitutents in the 3ꢀ and
3’ꢀpositions gave only the boroxinate species and the 3,3’ꢀPh2BINOL ligand gave a boroxinate catalyst
that gives excellent inductions in the aziridination reaction. BINOL derivatives with larger groups in
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 37
the 3,3’ꢀposition will not form either spiroꢀborates or boroxinate species and thus are not effective
1
2
3
catalysts at all.
4
5
6
1. INTRODUCTION
7
8
9
BINOL 1 is a C2ꢀsymmetrical biaryl ligand that has been widely used in a variety of asymmetric
reactions.1 The lack of ability of BINOL 1a to render high asymmetric inductions in many systems has
been related to the topological fact that its largest chiral pocket (naphthalene rings) and its active site
(hydroxy groups) are on opposite sides of the axis of chirality. The traditional solution to this limitation
is to employ BINOL derivatives which have substituents in the 3ꢀ and 3’ꢀpositions thus enhancing the
chiral pocket around the active site (Figure 1). This approach has led to the generation of a large family
of BINOL ligands which, as a whole, have made BINOL and its derivatives one of the most important
ligand types in asymmetric catalysis.2 Our approach to this problem has been to reꢀorient the direction
of the naphthalene rings such that they are projected in a vaulted fashion around the active site and
prototypical examples of this family of vaulted biaryls are isoꢀVAPOL 2, VANOL 3 and VAPOL 4.
These vaulted biaryl ligands have been demonstrated to be useful in a variety of useful asymmetric
reactions including DielsꢀAlder reactions,3,4 Mannich reactions,5 BaeyerꢀVilliger reactions,6 heteroatom
DielsꢀAlder reactions,7 the amidation8 and imidation9 of imines, the asymmetric reduction of imines,10
desymmetrization of aziridines,11aꢀe desymmetrization of diesters,11f the Petasis reaction,12 the
hydroarylation of alkenes,13 cisꢀaziridination of imines14,15 and transꢀaziridination of imines,16
benzoyloxylation of aryloxindoles,17 the 2ꢀazaꢀCope rearrangement,18 the Ugi reaction,21 azaꢀDarzens
reaction,19 hydrogenation of alkenes,20 propargylation of ketones,22 chorination and Michael reactions of
oxindoles,23 hydroacylation of alkenes,24 pinacol rearrangement,25a αꢀimino rearrangement,25b reduction
of aminals,26 Michael addition of alkynes,27a chromene functionalization,27b carbonylative
spirolactonization of enallenols,27c and controlled switching in a foldamer.27d More recently we have
reported the preparation of a number of derivatives of these vaulted ligands and these new ligand types
have been evaluated in the catalytic asymmetric aziridination reaction,15 DielsꢀAlder reactions,4 Ugi
reactions,21 and enantioselective protonation27e
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
ACS Paragon Plus Environment

Page 3 of 37
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The aziridination of imines with diazo compounds is arguably the most successful reaction to
date that is catalyzed by a VANOL or VAPOL derived catalyst.14ꢀ16 These aziridination catalysts were
originally prepared by heating the ligand with three equivalents of B(OPh)3 and then removing all
volatiles under vacuum as indicated by the procedure in Table 1.14a Data that compare the relative
effectiveness of BINOL, VANOL and VAPOL with imines with different protecting groups on the
nitrogen have been collected from the literature and presented in Table 1. In each case, it can be seen
that the BINOL derived catalyst gives far inferior asymmetric inductions when compared to the catalyst
generated from either VANOL or VAPOL. This observation is consistent with the hypothesis that
motivated the design and synthesis of the vaulted biaryl ligands: BINOL has a smaller chiral pocket
around the active site than VANOL or VAPOL. Also consistent with this hypothesis is the increase of
asymmetric induction for the aziridine 7a from 13% ee to 54% ee when the BINOL catalyst has 3ꢀ and
3’ꢀsubstituents (Table 1, entries 1 vs 2).28 Not consistent with this hypothesis is the fact that for all three
imines shown in Table 1 the catalysts derived from VANOL and VAPOL give nearly identical
asymmetric inductions. It would have been expected
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 37
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
that the VAPOL catalyst would have given higher asymmetric inductions than VANOL since the active
site is contained within a deeper chiral pocket in VAPOL than in VANOL.29 Any attempt to correlate
the effectiveness of these ligands and the size of the chiral pocket around the active site of course must
presuppose that the structure of the catalyst is the same for all of the ligands. Thus, while we have
recently gained some information concerning the structure of the boron based catalysts14d generated
from VANOL and VAPOL and B(OPh)3, not nearly as much is known about BINOL borates. The
purpose of the present work is to take a closer look at the structure and reactivity of various BINOL
borate esters that are generated upon interaction of BINOL and B(OPh)3. The findings are that the
supposition that BINOL, VANOL and VAPOL give the same type of catalyst is not correct and,
furthermore, that BINOL gives a mixture of two species under conditions where VANOL and VAPOL
give a single borate species. Computational analysis of the aziridination reactions catalyzed with the
different BINOL borate species supports the experimental observations that the low asymmetric
inductions with BINOL derived borate catalysts (Table 1) is due to the presence of two catalyst species
that give opposite enantiomers.
4
ACS Paragon Plus Environment

Page 5 of 37
Journal of the American Chemical Society
2. BACKGROUND
1
2
3
4
5
6
7
VANOL and VAPOL Catalysts and Preꢀcatalysts. The procedure involving heating the
VAPOL ligand with three equivalents of B(OPh)3 as indicated in Table 1 (Indirect Catalyst Assembly)
produces two species, and the 1H NMR, 11B NMR and mass spectroscopic analysis of this mixture
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
support the assignment of the mesoꢀborate 10 and the pyroꢀborate 11 as the components of this
mixture.14f The ratio of these two species varies with the method of preparation but typically the pyroꢀ
borate 11 is favored by a factor of ~4. The VAPOL pyroꢀborate was originally tentatively assigned as
the linear isomer,14a but more recent studies have lead to its reassignment as the cyclic isomer 11.14f
Neither the mesoꢀborate 10 nor the pyroꢀborate 11 are catalysts in the aziridination reaction, and the
mixture of these two species is in fact a preꢀcatalyst. Treatment of the mixture of 10 and 11 with an
imine leads to the generation of the boroxinate 12 which is the active catalyst and which consists of an
ion pair comprised of a chiral boroxinate anion and a protonated iminium.14d The complex 12 has been
fully characterized including Xꢀray analysis and crystals of this complex will catalyze the aziridination
reaction.14d A complex array of nonꢀcovalent interactions between the catalyst and the substrate in 12
were revealed by the Xꢀray diffraction analysis.14d We have subsequently found that it is not necessary
to prepare the preꢀcatalyst in order to generate the VAPOL boroxinate 12 (BOROX catalyst). The
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 37
covalent assembly of complex 12 can be induced in a few minutes at room temperature upon mixing
VAPOL, B(OPh)3 and an imine (Direct Catalyst Assembly), conditions under which there is no reaction
between VAPOL and B(OPh)3.14d It is also interesting to note that the formation of the three BꢀOꢀB
bonds in the boroxinate core of complex 12 from B(OPh)3 requires three equivalents of H2O. This is
consistent with the fact that the use of purified B(OPh)3 will not lead to the generation of the boroxinate
catalyst 12 unless 3 equivalents of H2O are added. Nearly all of the commercial samples of B(OPh)3
that we have examined over the years will generate active catalysts suggesting that these samples
contain at least three OH equivalents resulting from partial hydrolysis.14d In addition to aziridination
reactions,14ꢀ16 BOROX catalysts of the type 12 have been shown to be involved in 2ꢀazaꢀCope
rearrangements,18 a catalytic asymmetric Ugi reaction,21 and presumably in heteroatom DielsꢀAlder
reactions7 as well.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Known Borate Esters of BINOL. The most common structural types reported for borate esters
of BINOL are shown in Scheme 2.30 The reaction of BINOL with 1 equiv of B(OPh)3 has been reported
to give a compound for which the most probable structure was suggested to be the mesoꢀborate 13 but
other possible structures could not be excluded.31a,32 This material has been examined as a catalyst for
Mannich and heteroatom DielsꢀAlder reactions of imines,31 the heteroatom DielsꢀAlder reactions of
aldehydes,33 DielsꢀAlder reactions34 and aldol reactions.35 The reaction of BINOL with 0.5 equiv of
B(OPh)3 or B(OMe)3 has been reported to give the spiroꢀborate 14. Structure 14 has been suggested to
function as a chiral Lewis acid catalyst for a number of reactions including Mannich reactions,31h,36
heteroatom DielsꢀAlder reactions of imines,31h,37 and PictetꢀSpengler reaction.38 The catalytic function
of 14 was proposed to be as a chiral Lewis acid where interaction with a Lewis base disrupts one of the
boronꢀoxygen bonds to give the Lewis acidꢀLewis base complex 15.31h,38a There has been considerable
interest in the chiral anion unit in 14 and the sources of interest include use as a vehicle for the
resolution of BINOL,39 for examining chiral recognition in contact ionꢀpairs,40a as a structural unit in
solution and in the solid state,40b,c as a chiral shift reagent,41 and also as a chiral counterion catalyst in
aziridination reactions,42 in FriedelꢀCrafts reactions,43a in hydrogenations,43b in ringꢀopening of mesoꢀ
aziridiniums,43a and in the ringꢀopening of mesoꢀepoxides.44 The intriguing propellerꢀlike structure 17
was discovered by Kaufmann and Boese when they reacted BINOL with 1 equiv of BH2Br•SMe2
complex.34 This bisꢀtriaryl borate was isolated in nearly quantitative yield and contains a [6.6.6]ꢀ
propellane unit composed of three units of BINOL and two borons and was characterized by Xꢀray
diffraction. This complex has been shown to give high
6
ACS Paragon Plus Environment

Page 7 of 37
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
asymmetric inductions in DielsꢀAlder reactions as a chiral Lewis acid, however, as might be expected,
the rates of reaction are quite slow.34,45 An isomer of 17 has been reported by Deꢀjun and coworkers
(Scheme 2).46 This compound isoꢀ17 was also found to be composed of three units of BINOL and two
borons and was characterized by NMR spectroscopy. It was reported to be formed from the
decomposition of the allylꢀboronate (not shown) that was generated from the reaction of BINOL 1a with
triallylborane upon standing for several days. Another interesting structure that has been reported is the
pyroꢀborate 16, which was reported to form upon refluxing BINOL in toluene with 1.1 equiv of boric
acid.47 The pyroꢀborate 16 was characterized by combustion analysis and by IR spectroscopy and was
used in the resolution of BINOL. The veracity of the structural assignment of 16 has been called into
question as it has been reported that heating BINOL and boric acid in refluxing benzene gives rise to the
Kaufmann’s propeller 17.39a Perhaps the earliest and most widely used class of triaryl borates are those
that have internal chelation of the boron to a quinone moiety. Kelly first prepared borates of the type 18
in 1986 for the purpose of effecting asymmetric DielsꢀAlder reactions on naphthoquinones.4,48,49
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 37
3. RESULTS AND DISCUSSION
1
2
3
4
5
6
7
8
9
Borate Esters from a 1:1 ratio of BINOL and B(OPh)3. While the reaction of VAPOL 4 with
three equivalents of B(OPh)3 gives a 4:1 mixture of pyroꢀborate 11 to mesoꢀborate 10 (Scheme 1), a
simple change in stoichiometry to a 1:1 mixture of VAPOL and B(OPh)3 gives a 17:1 mixture in favor
of the mesoꢀborate 10.14f Similarly, a 1:1 mixture of VANOL 3 and B(OPh)3 gives a 10:1 mixture in
favor of the mesoꢀborate.14f In contrast, in the present work we find that a 1:1 mixture of BINOL and
B(OPh)3 gives predominately the pyroꢀborate 19 (Figure 2, entry 2). Essentially all of the BINOL is
accounted for in 19 and unreacted BINOL. The assignment of the structure of 19 will be discussed later
(Scheme 6) and the complete assignment of all of the protons of 19 can be found in the Supporting
Information. The spectra in entry 2 of Figure 2 were taken on a sample generated by combining
BINOL and 1 equivalent of B(OPh)3 in CDCl3 for 10 min. Essentially the same spectrum was observed
when BINOL and 1 equivalent of B(OPh)3 were allowed to react in CD2Cl2 for 1 hour in the presence of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
11
4 Å molecular sieves according to the original report31a,32 (see supporting information for H and B
NMR spectra). If the B(OPh)3 is purified and stored in a dry box, essentially no reaction will occur with
BINOL (this was done with a 2:1 mixture of BINOL and B(OPh)3, see Figure 7a, entry 2). This lack of
reactivity with purified B(OPh)3 has been previously observed with VANOL or VAPOL.14d This is in
contrast to the reaction of BINOL with commercial B(OPh)3 shown in entry 2 of Figure 2. An
equivalent of H2O is needed to create the BꢀOꢀB linkage and it is our experience that most commercial
B(OPh)3 samples contain 20ꢀ30 percent phenol as a result of partial hydrolysis.14d Cleaner spectra for
the pyroꢀborate 19 could be obtained by heating BINOL with two or three equivalents of a boron source
(B(OPh)3 or BH3•SMe2 + phenol) as indicated in entries 3 & 4 in Figure 2.
8
ACS Paragon Plus Environment

Page 9 of 37
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. 1H NMR of BINOLꢀborate complexes. Unless otherwise specified, all NMR spectra were performed in
CDCl3. (1) Pure BINOL in CDCl3; (2) BINOL and B(OPh)3 (1:1) were combined in CDCl3 for 10 min. (3)
BINOL pyroborate 19 prepared by heating a mixture of BINOL, BH3•Me2S, phenol and H2O (1:2:3:1) at 55 °C.
(4) BINOL pyroborate 19 prepared by heating a mixture of BINOL and B(OPh)3 (1:3) at 100 °C. (5) BINOL
pyroborate 19 prepared by heating a mixture of BINOL, BH3•Me2S, phenol and H2O (1:3:2:3) at 80 °C. (6) A
mixture of BINOL, BH3•Me2S and phenol (1:1:1) was heated at 50 °C. (7) Kaufmann’s propeller prepared by
heating a mixture of BINOL and BH3•Me2S (1:1) at 80 °C.
Given the failure to prepare the mesoꢀborate ester 13 from BINOL and B(OPh)3 other
approaches were attempted. The first was to heat a 1:1:1 mixture of BINOL, BH3•SMe2 and phenol at
50 °C and the 1H NMR spectrum of the resulting mixture is shown in entry 6 of Figure 2. The BINOL
is completely consumed, but the major product is Kaufmann’s propeller 17 along with the pyroꢀborate
19 (doublet, δ = 6.17 ppm) as the second major product (the mole ratio of 17:19 = 53:47). Other minor
products were observed but their identity could not be determined. A clean spectrum of Kaufmann’s
propeller 17 was generated by heating a 1:1 mixture of BINOL and BH3•SMe2 (Figure 2, entry 7).
A second strategy for access to the mesoꢀborate 13 is shown in Scheme 4 and involves the
method of Kelly.48,49 Kelly had found that acetic acid will accelerate the reaction of BINOL derivatives
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 37
with boraneꢀTHF complex and thus would generate the intermediate borate mixed anhydrides of the
type 21 to activate the boron for reaction with 8ꢀhydroxynaphthoquinone species to give borate esters of
the type 18 shown in Scheme 2. Following Kelly’s procedure, BINOL 1a was allowed to react with 1
equivalent of BH3•THF complex and 1 equivalent of acetic acid in THF at 20 °C for 10 min and then all
1
2
3
4
5
6
7
8
1
volatiles were removed and the residue taken up in CDCl3. However, the H NMR spectrum revealed
9
that Kaufmann’s propeller had formed in 94% yield (Scheme 4). Addition of one equivalent of phenol
to the NMR tube revealed that after 10 min at room temperature, much of Kaufmann’s propeller had
exchanged into a number of species with the liberation of 23% of the BINOL. A complex mixture of
minor species were observed that included a 2% yield of the pyroꢀborate 19. Kaufmann’s propeller
remained the major borate ester species that contains a BINOL ligand. In light of the failure to inꢀsitu
generate the acetoxy borate ester derivative 21a (R=H) it should be pointed out that the successful
preparation of 18 by Kelly was with a 3,3’ꢀdisubstituted BINOL derivative (1d, R = Ph), a class of
compounds that are known not to form Kaufmann’s trimer.34
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Borate Esters from a 1:0.5 ratio of BINOL and B(OPh)3. The reaction of B(OPh)3 with two
equivalents of BINOL was reported by Yamamoto to give the spiroꢀborate 14 simply by stirring the two
in CD2Cl2 over 4 Å molecular sieves for 1 h at room temperature (Scheme 2).31h,52 We have repeated
this reaction and examined the 1H and 11B NMR spectra which are presented in Figure 3. The 1H NMR
spectrum (Figure 3a, entry 3) reveals that unreacted BINOL is present in substantial amounts (47%)
along with the pyroꢀborate 19 (6%).50 A small amount of a number of other species are present as
1
1
indicated by the H NMR spectrum but these have not been identified. The H NMR spectra of a 1:1
and 2:1 mixture of BINOL and B(OPh)3 are quite similar with unreacted BINOL as the major species
10
ACS Paragon Plus Environment

Page 11 of 37
Journal of the American Chemical Society
present (37% vs 47%, respectively, Figure 3a, entries 2 and 3). Comparing Figure 3a, entry 2 and
Figure 2, entry 2, the amount of pyroꢀborate 19 that is produced varies quite a bit and is dependent on
the batch of B(OPh)3 that is used.
1
2
3
4
5
A dramatic simplification of the spectra is observed when the mixtures of BINOL and B(OPh)3
are treated with the imine 5a. When the mixture of two equivalents of BINOL and one equivalent of
B(OPh)3 (Figure 3a, entry 3) is treated with one equivalent of the imine 5a (Figure 3a, entry 5), the
complex mixture of compounds is converted cleanly into a single compound which is identified as the
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
Figure 3a. H NMR of BINOLꢀspiroꢀborate complex 22aꢀ5a. Unless otherwise specified, all NMR
spectra were performed in CD2Cl2. (1) Pure BINOL; (2) A mixture of BINOL and B(OPh)3 (1:1) was
allowed to react in CD2Cl2 for 1 h at 25 °C over 4 Å molecular sieves. (3) A mixture of BINOL and
B(OPh)3 (2:1) was allowed to react in CD2Cl2 for 1 h at 25 °C over 4 Å molecular sieves (4) One
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 37
equivalent of imine 5a added to entry 2. (5) One equivalent of imine 5a added to entry 3
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3b. 11B NMR of BINOLꢀspiroꢀborate complex 22aꢀ5a. Unless otherwise specified, all NMR spectra were
performed in CD2Cl2. (1) Pure B(OPh)3 (2) A mixture of BINOL and B(OPh)3 (1:1) was allowed to react in
CD2Cl2 for 1 h at 25 °C over 4 Å molecular sieves. (3) A mixture of BINOL and B(OPh)3 (2:1) was
allowed to react in CD2Cl2 for 1 h at 25 °C over 4 Å molecular sieves (4) One equivalent of imine 5a
added to entry 2. (5) One equivalent of imine 5a added to entry 3
spiroꢀborate 22aꢀ5a (Scheme 5). The same clean formation of 22aꢀ5a is observed whether BINOL and
B(OPh)3 are first mixed with 4Å MS for 1 h followed by imine 5a, or whether BINOL, B(OPh)3 and
imine 5a are immediately mixed and allowed to react only within the time that it takes to load an NMR
1
tube and take the H NMR and 11B NMR spectra (Figure 3a, entry 5 and Figure 3b, entry 5). This
compound is assigned as the ionꢀpair 22aꢀ5a comprised of a spiroꢀborate anion and the protonated form
of the imine 5a and is formed in 94% yield.50 The 1H NMR spectrum in entry 5 of Figure 3a is of a 3:1
mixture of phenol to 22aꢀ5a as would be expected from the reaction of a 2:1:1 mixture of BINOL, imine
5a and B(OPh)3. The four coordinate anionic borate unit in 22aꢀ5a displays a characteristic sharp
absorption at δ = 8.9 ppm in the 11B NMR spectrum (Figure 3b, entries 5) and reveals the complete
disappearance of the broad three coordinate borate absorption at δ = 16 ppm, consistent with the
formation of 22aꢀ5a in very high yield. The spiroꢀborate 22aꢀ5a is also formed (75% yield)50 when one
12
ACS Paragon Plus Environment

Page 13 of 37
Journal of the American Chemical Society
11
equivalent of imine 5a is added to a 1:1 mixture of BINOL and B(OPh)3 but in this case the B NMR
spectrum indicates the presence of both the three and four coordinate borons in a 1.0:1.2 ratio (Figure
3b, entry 4). Taken together with the fact that the 1H NMR reveals (Figure 3a, entry 4) that BINOL is
completely consumed when the imine is added, this is consistent with conversion of all of the BINOL
and half of the B(OPh)3 to the complex 22aꢀ5a. Thus, the treatment of B(OPh)3 with either one or two
equivalents of BINOL leads only to partial conversion to a complex mixtures of compounds. However,
treatment of either of these mixtures with an imine gives clean conversion to a chiral spiroꢀborate.
Support for the structure 22aꢀ5a containing a spiroꢀborate anion and an iminium cation comes
from the work of Yamamoto and coworkers who have previously studied heteroatom DielsꢀAlder and
Mannich reactions of imines mediated by a stoichiometric amount of a catalyst prepared from two
equivalents of BINOL and one equivalent of B(OPh)3.31h They were able to isolate and determine the
Xꢀray structure of the complex 22aꢀ5d although no other spectral data for 22aꢀ5d was presented (Figure
4).31h This complex consists of the spiroꢀborate anion, a molecule of phenol and the protonated form of
imine (S)ꢀ5d. There is no direct Hꢀbonding between the protonated iminium and the spiroꢀborate
anion, but rather the protonated iminium is Hꢀbonded to the phenol, which in turn is Hꢀbonded to the
spiroꢀborate anion. This crystal was of the misꢀmatched pair of (S)ꢀBINOL and (S)ꢀ5d and perhaps in
the matched pair nonꢀcovalent contacts between the (S)ꢀBINOL core and protonated (R)ꢀ5d would be
more favorable. While studying the NMR spectra of the complex 22aꢀ5a containing an achiral imine
(Scheme 5), yellow crystals were serendipitously formed in the NMR tube and thus the opportunity to
determine the crystal structure of the ionꢀpair 22aꢀ5a (Scheme 3) presented itself. It was hoped that
nonꢀcovalent interactions between the spiroꢀborate core and the iminium could be found in 22aꢀ5a since
the imine 5a from which it is derived is not chiral.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
It was found that complex 22aꢀ5a also crystallized with a molecule of phenol but an Xꢀray
diffraction analysis of these crystals revealed that the nonꢀcovalent interactions present in complex 22aꢀ
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 37
5a (Figure 5) are much different than in Yamamoto’s complex 22aꢀ5d (Figure 4). There are several
direct interactions between the iminium ion and the spiroꢀborate core. The iminium proton is directly
Hꢀbonded to the spiroꢀborate anion with a biased bifurcated Hꢀbond that spans two of the oxygens of
the spiroꢀborate (Figure 5, d2 and d3). The phenol unit is also Hꢀbonded to the spiroꢀborate core but to
the opposite side of the anion to which the iminium is Hꢀbonded. Additional secondary interactions
between the iminium substrate and the spiroꢀborate catalyst include two CHꢀπ interactions between the
edges of the phenyl rings of the iminium and the faces of the naphthalene rings of the spiroꢀborate anion
(Figure 5, d1 and d4). As a result of these interactions, the benzylidene iminium plane (with the coplanar
iminium bond) forms an angle of 106.4° with a naphthalene ring of one of the BINOL’s. The nature of
the docking of imine 5a in the spiroꢀborate catalyst derived from (S)ꢀBINOL reveals that the Reꢀface of
the imine is exposed to nucleophilic attack. The correlation of this observation with the actual
stereoselectivity observed with spiroꢀborate catalysts of BINOL will be discussed below.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
14
ACS Paragon Plus Environment

Page 15 of 37
Journal of the American Chemical Society
In an attempt to obtain an Xꢀray structure of the iminium catalyst complex 22aꢀ5a free of
1
2
phenol, solutions of this spiroꢀborate were generated from the reaction of a 2:1 mixture of BINOL and
BH3•THF complex with imine 5a but all attempts to grow crystals of this complex were unsuccessful.
We had previously found for the VAPOL boroxinate complex 12 (Scheme 1) that crystals of 12 were
much easier to come by for the MEDAM imine 5b than for the benzhydryl imine 5a (Table 1).14d This
also proved to be true in the present case as crystals of the catalystꢀsubstrate complex 22aꢀ5b could be
grown from a complex generated from BINOL 1a, BH3•THF and imine 5b (2:1:1). The binding of the
iminium substrate to the spiroꢀborate catalyst is essentially unchanged from that of complex 22aꢀ5a in
spite of the fact that it does not have an Hꢀbonded phenol. The bifurcated Hꢀbond (Figure 6, d2 and d3)
could also be located that link the two ions in the ion pair. The two CHꢀπ interactions between the
edges of the phenyl rings of the iminium
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 37
and the faces of the naphthalene rings of the catalyst are also present (Figure 6, d1 and d4). The angle
between the benzylidene unit of the iminium and the naphthalene of one of the BINOL’s is a little closer
to perpendicular in the present case (97.3°). An additional interaction is present in the structure of 22aꢀ
5b that is not in 22aꢀ5a, a CHꢀπ interaction between the one of the methyl groups of the imine 5b and
one of the naphthalene rings of the catalyst (Figure 6, d5). This same type of interaction exists in the
substrateꢀcatalyst complex 12 (Scheme 1) derived from the same imine and the VAPOL ligand which
gives a boroxinate catalyst. However, the binding of the rest of the imine in complex 12 is very
different than in the complex 22aꢀ5b. As for complex 22aꢀ5a, the binding of the iminium in complex
22aꢀ5b indicates that the reꢀface of the iminium is the one that is most open to nucleophilic attack and
how this correlates with stereoselectivity in the aziridination reactions will be considered below.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Borate Esters from a 1:3 ratio of BINOL and B(OPh)3. Our venture into the arena of borate
complexes with BINOL began with trying to understand why BINOL does not produce an effective
catalyst for the aziridination reaction when catalyst preparation was performed with the same protocol
used in the aziridination reaction with VANOL and VAPOL ligands shown in Scheme 1. Preꢀcatalyst
formation was performed in the same way by heating (S)ꢀBINOL with 3 equivalents of commercial
B(OPh)3, however, unlike VAPOL and VANOL which give mixtures of mesoꢀ and pyroꢀborates, the
1
reaction with BINOL gives clean conversion to the pyroꢀborate 19a shown in Scheme 6. The H NMR
spectra of this reaction mixture is shown in Figure 2 (entry 4) and reveals that BINOL reacts to give
essentially a single species with at most trace amounts of Kaufmann’s propeller 17 or any other species
containing a BINOL ligand. The main contaminants in this mixture are unreacted B(OPh)3 and a trace
1
amount of triphenoxyboroxine 20 (Scheme 3) as evidenced by the H NMR spectrum with the aid of
independently prepared and purified samples of each (See Supporting Information).
The pyroꢀborate 19a can also be generated by heating BINOL with 2 equivalents of BH3•SMe2,
3 equivalents of phenol and 1 equivalent of H2O (Figure 2, entry 3). The H2O is necessary to make the
BꢀOꢀB linkage. Even heating BINOL with 3 equivalents of BH3•SMe2, 2 equivalents of phenol and 3
equivalent of H2O afforded pyroꢀborate 19a as the major compound (Figure 2, entry 5). The pyroꢀ
borate 19a was characterized as the mixture obtained by the method shown in Scheme 6 which
produced 19a as essentially the only BINOL containing species in the mixture (Figure 2, entry 4). While
the NMR and mass spectral data were consistent with the structure of the cyclic pyroꢀborate 19a, an
alternative linear pyroꢀborate structure would be 23a. The assignment of all of the protons of the cyclic
1
pyroꢀborate 19a was made on the basis of H NMR experiments that were employed in the assignment
of the structure of the pyroꢀborate 11 of VAPOL14f that involved preparing the pyroꢀborate from perꢀ
16
ACS Paragon Plus Environment

Page 17 of 37
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
deuterated triphenylborate and these experiments are detailed in the supporting information. In
addition, the cyclic pyroꢀborate 19a was found to be 5.22 kcal/mole more stable than the linear pyroꢀ
borate 23a as determined at the B3LYP/6ꢀ311+G(d,p)//B3LYP/6ꢀ31G(d) level of theory and the details
can be found in the supporting information. The energy minimized structure for the cyclic pyroꢀborate
19a has one of the orthoꢀhydrogens of one of the phenoxy groups within a distance (CꢀC, 4.8 Å)
compatible with a CHꢀπ interaction with a naphthalene ring (see supporting information). In the energy
17
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 37
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
11
Figure 7a. B NMR spectra of boron complex derived from BINOL 1a and 1d in CDCl3. B(OPh)3 was used as
purchased unless otherwise specified. All NMR spectra were acquired within 10 min after the sample was
prepared. (1) Mixture of B(OPh)3 and imine 5e (1:1); (2) Mixture of (S)ꢀBINOL 1a and purified B(OPh)3 (2:1);
(3) Mixture of (S)ꢀBINOL 1a, purified B(OPh)3 and imine 5a (2:1:1); (4) Mixture of (S)ꢀBINOL, purified
B(OPh)3 and imine 5e (2:1:1); (5) Mixture of (S)ꢀBINOL 1a, BH3•THF and imine 5b (2:1:1); (6) Mixture of (S)ꢀ
BINOL 1a, purified B(OPh)3 with aziridine (R,R)ꢀ7a (2:1:1); a similar spectrum was observed with (S,S)ꢀ7a; (7)
Preꢀcatalyst made by heating a 1:4:1 mixture of (S)ꢀBINOL 1a, B(OPh)3 and H2O in toluene at 80 °C for 1 h and
then removal of volatiles; (8) Mixture of imine 5e and preꢀcatalyst (1:1) made in entry 7. (9) Preꢀcatalyst made by
heating a 1:4:1 mixture of substituted (S)ꢀBINOL 1d, B(OPh)3 and H2O in toluene at 80 °C for 1 h and then
removal of volatiles; (10) Mixture of imine 5a and preꢀcatalyst (1:1) made in entry 9; (11) A 1:1 mixture of imine
5a and preꢀcatalyst prepared by reaction of (S)ꢀBINOL 1d, BH3•SMe2, PhOH, and H2O (1:3:2:3) in toluene at 100
°C for 1 h and then removal of the volatiles. (12) A 10:1 mixture of imine 5a and preꢀcatalyst prepared by
reaction of (S)ꢀBINOL 1d, BH3•SMe2, PhOH, and H2O (1:3:2:3) in toluene at 100 °C for 1 h and then removal of
the volatiles.
minimized structure for the linear pyroꢀborate 23a, the closest such interaction is 5.4 Å. Since the
doublet at δ = 6.17 ppm (Figure 2) has been assigned as the ortho hydrogens on the phenoxy groups,
this chemical shift is most consistent with the cyclic pyroꢀborate 19a. The overall finding is that
BINOL has a much greater propensity for the pyroꢀborate over the mesoꢀborate than either VANOL or
VAPOL (Schemes 1 vs 6).
While the reaction of VAPOL with three equivalents of B(OPh)3 gives a mixture of mesoꢀand
pyroꢀborates 10 and 11, subsequent reaction with an imine gives the active catalyst as a single species,
18
ACS Paragon Plus Environment

Page 19 of 37
Journal of the American Chemical Society
the ionꢀpair 12, which contains a chiral boroxinate anion with a boroxine core (Scheme 1). The
1
2
3
opposite situation pertains with BINOL. The reaction of three equivalents of B(OPh)3 with
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7b. 11B NMR spectra of boron complex derived from BINOL analogs in CDCl3. B(OPh)3 was used as
purchased unless otherwise specified. (1) Mixture of B(OPh)3 and imine 5e (1:1); (2) Mixture of imine 5e and
preꢀcatalyst (1:1) made by heating a 1:4:1 mixture of (S)ꢀBINOL 1d, B(OPh)3 and H2O in toluene at 80 °C for 1 h
and then removal of volatiles; (3) Mixture of imine 5e and preꢀformed catalyst (1:1) made from substituted (S)ꢀ
BINOL 1b, B(OPh)3 and H2O (1:4:1); (4) Mixture of imine 5e and preformed catalyst (1:1) made from
substituted (S)ꢀBINOL 1c, B(OPh)3 and H2O (1:4:1); (5) Mixture of imine 5e and preformed catalyst (1:1) made
from substituted (S)ꢀBINOL 1h, B(OPh)3 and H2O (1:4:1); (6) Mixture of imine 5e and preꢀformed catalyst (1:1)
made from substituted (S)ꢀBINOL 1i, B(OPh)3 and H2O (1:4:1); (7) Mixture of imine 5e and preformed catalyst
(1:1) made from substituted (S)ꢀBINOL 1j, B(OPh)3 and H2O (1:4:1); (8) Mixture of imine 5e and preformed
catalyst (1:1) made from substituted (S)ꢀBINOL 1k, B(OPh)3 and H2O (1:4:1)
BINOL 1a gives a single species (pyroꢀborate 19a), but the subsequent reaction with an imine 5e gives a
5:2 mixture of the spiroꢀborate 22aꢀ5e and the boroxinate 24aꢀ5e, respectively, along with the tetraꢀ
phenoxyborate 25aꢀ5e. (Scheme 6). A very similar ratio was noted with the imine 5a (Scheme 6 and
Table 1). The presence of each of these species is most easily determined from the 11B NMR spectrum.
The distribution of these three species as a function of the nature of the BINOL ligand, the type of imine
and the stoichiometry of B(OPh)3 has been investigated and the results are summarized in Figure 7.
First it is to be noted that when the stoichiometry of the ligand, B(OPh)3 and imine is
19
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 37
2:1:1, a clean formation of the spiroꢀborate complex 22a is observed and this is true with imines 5a and
5e as shown by the presence of a sharp peak around δ ~ 9 ppm (Figure 7a, entries 3ꢀ4) in the 11B NMR
spectrum. The spiroꢀborate 22a can also be generated as a phenol free solution from the reaction of
BH3•THF with BINOL and imine 5b (Figure 7a, entry 5). Interestingly, the aziridine 7a (Table 1) is not
a strong enough base to assemble the spiroꢀborate anion 22a (Figure 7a, entry 6). When the ratio of
BINOL to B(OPh)3 is 1:4 one no longer sees a clean formation of the spiroꢀborate 22a (Figure 7a,
entries 4 vs 8). Instead, a 5:2:4 mixture of the spiroꢀborate 22a, the boroxinate 24a and the nonꢀchiral
borate 25a is observed. The major BINOL containing species is still the spiroꢀborate 22a but the
appearance of a peak at δ ~ 6 ppm indicates that the boroxinate 24a is also formed under these
conditions. A number of different boroxinate complexes of VANOL and VAPOL of the type 12
(Scheme 1) all have absorptions in the range of 5ꢀ6 ppm.14d We assume that the anionic BINOL borates
species 22a and 24a are in equilibrium, and since the spiroꢀborate 22a is still the major species in the
presence of 4 equivalents of B(OPh)3, one can deduce that the equilibrium favors the spiroꢀborate 22a.
In support of this assumption, Bull, James and coworkers have observed dynamic ligand exchange of
BINOL between chiral and nonꢀchiral boronꢀaryloxy complexes in a heteroatom DielsꢀAlder reaction.37
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As expected the neutral pyroꢀborate 19a is the major species in the presence of excess B(OPh) prior to
3
the addition of imine 5e. The borate species 25 with the absorption at δ ~ 2 ppm in the 11B NMR
spectrum can also be formed from just B(OPh)3 and imine 5e (entry 1) and thus this species was
determined not to contain a molecule of BINOL. The assignment of the peak at δ ~ 2 ppm as the
1
tetraphenoxyborate 25 is based on literature values for related tetraaryloxy borate salts.54 For the H
NMR spectra corresponding to the entries in Figure 7a, see supporting information.
A significantly different outcome is observed when 3,3’ꢀdisubstituted BINOL derivatives are
treated with B(OPh)3 (4 equiv) and then with an imine. For the 3,3’ꢀdiphenyl BINOL 1d only the
boroxinate 24dꢀ5a is formed with imine 5a (along with tetraaryloxyborate 25ꢀ5a) and there is at most a
trace of the spiroꢀborate 22dꢀ5a (Figure 7a, entry 10). This is interpreted to be the consequence of steric
interactions between the Ph groups on one BINOL with those on another in the possible spiroꢀborate
species 22dꢀ5a. Also to be noted is that the spectrum does not change whether one or ten equivalents of
the imine 5a is added (Figure 7a, entries 11 and 12). Among other BINOL derivatives, boroxinate
species of the type 24 were observed with ligands 1b (R = Br) and 1c (R = I) (Figure 7b, entries 3 and
4). As was the case for BINOL 1d (Figure 7b, entry 2) no spiroꢀborate 22b/22c was formed with
BINOL 1b or 1c as noted by the absence of absorptions at ~9ꢀ10 ppm in entries 3 and 4 of Figure 7b.
Only small amounts of boroxinates 24h/24i were obtained in the cases of BINOL derivatives 1h (R = 9ꢀ
phenanthryl) and 1i (R = 9ꢀanthracenyl) (Figure 7b, entries 5 and 6). The introduction of the very large
20
ACS Paragon Plus Environment

Page 21 of 37
Journal of the American Chemical Society
2,4,6ꢀ(iꢀPr)3C6H2 groups in the 3ꢀ and 3’ꢀpositions of BINOL (1j) causes the boron to forgo
incorporation into either the spiroꢀborate 22j or the boroxinate 24j (Figure 7b, entry 7). The only borate
species observed in this case is the tetraꢀphenoxyborate 25 which exists as a salt with the protonated
form of imine 5e. A similar observation was made in the case of the BINOL derivative 1k with large
triphenylsilyl groups at the 3ꢀ and 3’ꢀpositions (Figure 7b, entry 8).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Comparison of the BINOL spiroꢀBorate 22 and the BINOL Boroxinate 24 as Catalysts in the
Aziridination Reaction. We have not observed a spiroꢀborate species of the type 22 with either the
VAPOL or VANOL ligand although this will be the subject of additional study. The fact that BINOL
can be induced to cleanly generate the spiroꢀborate 22 presented the opportunity to investigate the
ability of this chiral borate to induce asymmetric induction in the aziridination of imines with ethyl
diazoacetate. The data in Table 2 include results with catalysts generated with both a 1:2 and a 3ꢀ4:1
ratio of B(OPh)3 to BINOL. For comparison, results of catalysts prepared from a 1:2 and a 3ꢀ4:1 ratio
of B(OPh)3 to VAPOL are also included in the table (entries 2 and 3). Our previous studies indicate that
VAPOL does not form a spiroꢀborate for steric reasons and thus the result in entry 3 is thought to most
likely involve a boroxinate catalyst. A control experiment was performed where only B(OPh)3 was
used as the catalyst (no ligand) and a 78% yield of the aziridine 7a was isolated (entry 1). Since it was
shown that the reaction of B(OPh)3 and an imine can generate the tetraphenoxyborate 25 (Figure 7a,
entry 1), this control experiment provides evidence that this species catalyzes the background reaction.
However, as will be discussed below, we have evidence that the background reaction is not the source
of the lower inductions for the BINOL catalyst. As has been determined to be the case for VAPOL,14b
BINOL catalysts give higher asymmetric inductions in the series, BUDAM imines > MEDAM imines >
benzhydryl imines under boroxinate forming conditions with a 3 : 1 ratio of B(OPh)3/BINOL (Table 2,
entries 10ꢀ12). As was shown by the 11B NMR spectra in entries 3ꢀ5 in Figure 7a, the use of a 1:2 ratio
of B(OPh)3 to BINOL gives the clean formation of the spiroꢀborate 22 and as can be seen from entry 13
in Table 2 this species does catalyze the aziridination and gives the aziridine 7a in 26% ee. The
asymmetric induction can be increased from 26% ee to 76% ee with 100 mol% catalyst, but for reasons
we don’t understand, the yield falls to 8% (entry 14). The % ee can also be increased by using the
MEDAM imine 5b giving 7b in 63% ee (entry 15). When the ratio of B(OPh)3 to BINOL is 3:1, a
11
mixture of spiroꢀborate 22 and boroxinate 24 is produced as indicated by the B NMR spectra and as
indicated in Scheme 6 the ratio of 22:24 is 100:40 with imines 5e and 100:41 with imine 5a. For imine
5a, this catalyst mixture results in the formation of the aziridine 7a in 24% ee and this rises to 67% ee
with the BUDAM imine 5c (entries 10 and 12). For the boroxinate catalysts of VANOL and VAPOL, it
has been found that the asymmetric inductions of a range of substrates is slightly higher for toluene as
solvent than methylene chloride.14a This solvent effect was found also to be true for the boroxinate
catalyst generated from the BINOL ligand 1d (entries 18 vs 19). The 11B NMR spectra reveal that
ligand 1d forms the boroxinate species to the exclusion of the spiroꢀborate (Figure 7a, entries 10ꢀ12).
21
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 37
However, the catalysts generated from a 1:2 ratio of B(OPh)3 to BINOL could not be studied in both
solvents. This catalyst mixture was not soluble in toluene but was completely soluble in methylene
chloride (entry 13). When the reaction in entry 13 is run in toluene there is a solid present in the
reaction mixture during the entire course of the reaction and the results were variable and not deemed
reliable (see supporting information). On the other hand, the catalyst generated from a 3:1 ratio of
B(OPh)3 and BINOL is soluble in toluene (entries 8 & 9).
1
2
3
4
5
6
7
8
9
The most interesting observation to be made from the data in Table 2 is that the BINOL catalyst
prepared with a 1:2 ratio of B(OPh)3 to BINOL gives the opposite configuration of aziridine 7a than the
catalyst generated from a 4:1 ratio (entries 10 vs. 13). Since the former gives a catalyst consisting of
just the spiroꢀborate 22a (Figure 7a, entry 3) and the latter gives a catalyst consisting of a mixture of the
spiroꢀborate 22a and the boroxinate 24a (Figure 7a, entry 3), this suggests that as catalysts, the spiroꢀ
borate 22a and the boroxinate 24a give opposite asymmetric inductions in the aziridination reaction.
The 11B NMR studies summarized in Figure 7 reveal that the 3,3’ꢀdiphenyl BINOL ligand 1d produces
a catalyst that contains only the boroxinate 24 and none of the spiroꢀborate 22 (Figure 7a, entries 10ꢀ12)
Thus, this presents an opportunity to determine how the BINOL boroxinate catalyst 24 with the
boroxinate core will function in the aziridination catalyst unencumbered by the presence of the
competing spiroꢀborate catalyst 22 and, furthermore, to determine how it compares to the boroxinate
VAPOL catalyst 12. As expected, the diphenyl BINOL derivative 1d gave much higher asymmetric
induction than the unꢀsubstituted BINOL giving 70% ee under conditions where BINOL gives only 13%
ee (Table 2, entry 16 vs. entry 8). This is still significantly lower selectivity than observed for the
boroxinate VAPOL catalyst 12 under the same conditions (91% ee, Table 2, entry 2). These results
reveal that while the (R)ꢀenantiomer of the pure BINOL boroxinate catalyst 24d gives the (2S,3S)ꢀ
aziridine 7a, the same enantiomer of this aziridine is obtained from the (S)ꢀenantiomer of the pure spiroꢀ
borate BINOL catalyst 22a. Interestingly, an asymmetric induction up to 90% ee was obtained with the
BUDAM phenyl imine 5c when the boron source was switched from B(OPh)3 to BH3•SMe2 (entry 23 vs
entry 24). However, no significant changes were observed for the imines 5a and 5b (entries 17 vs 21
and entries 20 vs 22). Additionally, comparable results were obtained when the catalyst was formed
from BINOL 1d by the direct method involving simply mixing 1d with B(OPh)3 and imine 5a as when
the catalyst was prepared by the indirect method involving the initial generation of a preꢀcatalyst as
indicated in Scheme 1 (75% ee, entry 16 vs. 76% ee, entry 17). It should be noted that Chen and
coworkers reported a huge drop in enantioinduction (34% ee) using BINOL derivative 1d and an imine
derived from 4ꢀSO2CH3ꢀC6H4CHO for the catalytic asymmetric aziridination reaction.14e
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
22
ACS Paragon Plus Environment

Page 23 of 37
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
A question arises in comparing the asymmetric induction observed for the VAPOL boroxinate
catalyst (94% ee, Table 2, entry 4), and the BINOL boroxinate catalyst (76% ee, Table 2, entry 17). Is
this difference the result of an inherent difference in the boroxinate complex 12 with VAPOL and the
boroxinate complex 24d for the BINOL ligand 1d? Alternatively, could the difference be due to a
greater background reaction by the nonꢀchiral borate species 25 for the BINOL boroxinate than for the
53
54
55
56
57
58
59
60
VAPOL boroxinate? We have obtained evidence that this is not the case. Depending on the
23
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 37
stoichiometry, varying amounts of the nonꢀchiral borate species 25 is formed. For example, the BINOL
boroxinate 24d generated from the reaction of the BINOL ligand 1d with 4 equivalents of a boron
precursor is formed with significant amounts of the nonꢀchiral borate 25 (38% of 24d) (Figure 7a, entry
10), whereas, with only 3 equivalents of boron the boroxinate 24d is formed with no discernable amount
of the nonꢀchiral borate 25 (Figure 7a, entry 11). Despite this, the formation of the BINOL boroxinate
24d under these two conditions gives the same asymmetric induction (Table 2, entries 17 and 20).
In order to examine the electronic and steric effects of other substituted BINOL ligands, a series
of catalysts were generated from several different 3,3’ꢀdisubstituted BINOLs with 3ꢀ4 equivalents of
B(OPh)3 (or BH3•SMe2 and PhOH) and the results from the screening of these catalysts in the
aziridination of imines 5a and 5f with ethyl diazoacetate are is presented in Table 3. Since the 11B NMR
spectra revealed that the dibromo and diiodo BINOL derivatives 1b and 1c gave relatively clean
catalysts that contained the boroxinate complex 24b/24c but at most traces of the spiroꢀborate 22b/22c
(Figure 7b, entries 3 and 4) these catalysts as expected gave higher asymmetric inductions (~50% ee)
than the BINOL itself (Table 3, entries 2 and 3 vs. entry 1). The finding that 3,3’ꢀdisubstituted BINOL
derivatives give higher asymmetric inductions in the aziridination reaction than the parent BINOL had
been previously observed by Wipf and Lyon (entries 8ꢀ11).28 Interestingly, while they tried ligands
bearing metaꢀ (entry 11) and paraꢀsubstituted aryl rings (entries 8ꢀ10) at the 3,3’ positions, no ligand
bearing an orthoꢀsubstituted aryl ring was examined nor was the unsubstituted phenyl analog 1d
examined. The observation that the BINOL derivative 1j (R = 2,4,6ꢀ(iꢀPr)3C6H2) containing isoꢀpropyl
groups in the orthoꢀpositions gives the aziridine 7a in –2% ee (entry 14) was in fact anticipated from the
11B NMR spectrum in entry 7 of Figure 7b since it reveals the absence of both the spiroꢀborate catalyst
22j and the boroxinate catalyst 24j. However, the reaction does occur in 78% yield (when R = 2,4,6ꢀ(iꢀ
Pr)3C6H2) which can be explained by the fact that the only borate species observed in 11B NMR
spectrum was the nonꢀchiral borate 25. Also as expected for the same reasons, a similar observation
(1% ee, entry 16) was made for the BINOL derivative 1k (R = SiPh3). While a 58% ee (entry 11) was
reported by Wipf and coworkers for aziridine 7f with a catalyst prepared from the 2ꢀnaphthyl BINOL
derivative 1g (R = 2ꢀnapthyl), nearly racemic aziridines 7a were obtained when the BINOL derivatives
1h (R = 9ꢀphenanthryl) and 1i (R = 9ꢀanthracenyl) were employed (entries 12ꢀ13). This suggests that
orthoꢀsubstitution in an aryl ring in the 3,3’ꢀpositions of BINOL hinders the formation of the boroxinate
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
11
catalyst thus lowering the asymmetric induction. This is supported by the B NMR spectrum which
indicate the formation of a much reduced amount of the boroxinate catalyst 24h and 24i (Figure 7b,
entries 5 and 6). It is interesting to note that the present
24
ACS Paragon Plus Environment

Page 25 of 37
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
work reveals that BINOL ligand 1d bearing a phenyl group in the 3 and 3’ꢀpositions is superior to the
BINOL ligand 1f with a paraꢀ2ꢀnaphthylphenyl group (entries 5 vs 10). Finally, in 2011, Chen and
coworkers reported similar observations with BINOL derivatives 1i and 1k when they performed
55
56
57
58
59
60
catalytic asymmetric aziridination reactions with an imine derived from 4ꢀSO2CH3ꢀC6H4CHO.14e
25
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 37
Computational Analysis of the Asymmetric Inductions for the SpiroꢀBorate 22 and the
Boroxinate 24. We investigated the origins of opposite asymmetric induction for the reaction of 5a and
6 catalyzed by (R)ꢀBINOLꢀspiroꢀborate 22a and (R)ꢀBINOLꢀboroxinate 24a using the B3LYP55 method
employing a 6ꢀ31G* basis set (for the entire system) as implemented in Gaussian ´09.56 The mechanism
of the aziridination reaction of MEDAM imine 5b and 6 catalyzed by VANOLꢀbased boroxinate has
been established using experimental kinetic isotope effects (KIEs) and theoretical calculations.14g The
reaction proceeds via initial carbonꢀcarbon bondꢀformation (TS1) between 5a and 6 to reversibly yield
the diazonium ion intermediate followed by rateꢀdetermining SN2ꢀlike attack of the imine nitrogen with
concomitant displacement of N2 (TS2) to furnish the final cisꢀaziridine product 7a (Scheme 8). The
origin of enantioselectivity is obtained from the relative energetics of TS2 in the pathways leading to the
major and minor enantiomer of 7a.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Rather than assume the same mechanism for the analogous reaction catalyzed by 22a and 24a,
we thought it appropriate to model both the CꢀC bond forming (TS1) and ringꢀclosing (TS2) transition
structures leading to both the major and minor enantiomers for both catalysts. Generation of the energy
profile of the reaction coordinate is expected to provide insight into the enantiodivergence observed in
reactions catalyzed by these two distinct species 22a and 24a. Such a comprehensive analysis of a large
computational system is challenging; but our extensive experience in modeling these reactions allowed
Figure 8. Reaction Coordinate ꢀ spiro-borate (22a) catalyzed aziridination reaction of 5a and 6
26
ACS Paragon Plus Environment

Page 27 of 37
Journal of the American Chemical Society
us to streamline the process. It must be noted here that the use of benzhydryl instead of MEDAM imine
and BINOL instead of VAPOL significantly reduced the size of the model system compared to our
previous studies. This enabled treatment of the full model system using the DFT method, which is
expected to provide more reliable energetics compared to our previous studies that utilized the hybrid
DFT:semiꢀemperical ONIOM method.14g
1
2
3
4
5
6
7
8
9
The reaction pathways leading to the major and minor enantiomers of the cis-aziridine product
7a catalyzed by (R)ꢀBINOLꢀspiroꢀborate 22a are shown in Figure 8. The highest energy barrier in the
reaction coordinates for the major (Red, si attack) and minor (Blue, re attack) enantiomer is the SN2ꢀlike
ringꢀclosing step. The enantioselectivity is therefore a result of the difference in energy, ∆∆G‡, of the
two ringꢀclosing transition structures spiroꢀRe-TS2 and spiroꢀSi-TS2. This difference is found to be 1.2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 9. Relevant transition structures along the reaction coordinates leading to the major and minor
enantiomer of product 7a in the reaction of 5a and 6 catalyzed by spiro-borate 22a.
27
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 37
kcal/mol, which corresponds to 82 %ee at room temperature – with the preferred product corresponding
to initial attack of 6 to the si-face of imine 5a. While this is an overꢀestimation of the experimental
enantioselectivity (26% ee; Table 2, entry 13), it is in qualitative agreement with experiment. All four
relevant transition structures along with key distances of bondꢀforming and stabilizing interactions (with
distances in Å) are shown in Figure 9. Rateꢀdetermining transition structures in both pathways (spiroꢀ
Re-TS2 and spiroꢀSiꢀTS2) benefit from similar stabilizing interactions, namely (a) a CH…O nonꢀ
covalent interaction between the αꢀCH of 6 and an anionic spiroꢀborate oxygen atom, and (b) an
intramolecular Hꢀbonding interaction between the NꢀH and the carbonyl moiety of the diazonium ion
intermediate.57 Transition structure spiroꢀSiꢀTS2 is lower in energy than spiroꢀReꢀTS2 since both these
interactions, (a) & (b), are enhanced in spiroꢀSiꢀTS2 (rCH..O=1.91 Å and rNH..O=2.09 Å) versus spiroꢀReꢀ
TS2 (rCH..O=1.93 Å and rNH..O=2.26 Å).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The reaction pathways leading to the major and minor enantiomers of the cis-aziridine product
7a catalyzed by the (R)ꢀBINOLꢀboroxinate 24a is shown in Figure 10. The highest energy barrier along
the reaction coordinate for the major enantiomer (Blue, re attack) is for the CꢀC bondꢀforming step
(boroxꢀReꢀTS1). Intriguingly, the highest energy barrier for the minor enantiomer (Red, si attack) is the
SN2ꢀlike ringꢀclosing step (boroxꢀSiꢀTS2). This presents a rather unique situation where the origin of
enantioselectivity in a catalytic asymmetric reaction is determined by the comparison of the relative
energy of different steps along the reaction coordinate of the two enantiomeric pathways. The ∆∆G‡,
the energy difference between these two highest energy steps, is 1.5 kcal/mol – with the preferred
product corresponding to initial attack of 6 to the re-face of imine 5a. This corresponds to a prediction
of
89%
ee
at
room
temperature
(of
the
opposite
enantiomer
of
7a
Figure 10. Reaction Coordinate ꢀ boroxinate (24a) catalyzed aziridination reaction of 5a and 6
28
ACS Paragon Plus Environment

Page 29 of 37
Journal of the American Chemical Society
with respect to the reaction catalyzed by 22a) if the boroxinate catalyst could be formed exclusively.
There is no experimental result corresponding to this theoretical finding since 24a is always formed as
the minor component in the catalyst mixture, regardless of the ratio of ligand to B(OPh)3 used. However,
this result suggests that, if formed exclusively, 24a would be a rather selective catalyst for this
transformation. Additionally, our computational study validates the hypothesis that the low
enantioselectivity, observed when a mixture of 22a and 24a catalyzed the aziridination reaction of 5a
and 6, results from the opposite absolute asymmetric induction observed with the two catalysts. All four
of the transition structures for the reaction catalyzed by 24a, along with distances for key stabilizing
interactions are shown (Figure 11).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 11. Relevant transition structures along the reaction coordinates leading to the major and minor
enantiomer of product 7a in the reaction of 5a and 6 catalyzed by borox-borate 24
Which is the better catalyst? We have shown that when a 3:1 ratio of B(OPh)3 to (S)ꢀBINOL is used,
(S)ꢀBINOLꢀspiroꢀborate 22a and (S)ꢀBINOLꢀboroxinate 24a are formed in the ratio of 5:2 with imine
5a (Scheme 6 and Supporting Information). The enantiomer of the aziridine 7a that is formed under
29
ACS Paragon Plus Environment