Cycloaddition reactions between dicyclohexylboron azide and alkynes

Article abstract of DOI:10.1039/c3dt00068k

The room temperature 1,3-dipolar cycloaddition reactions of the boron azide, Cy₂BN₃ with the electron-poor acetylenes RCO ₂CCCO₂R, EtCCCOMe and HCCP(O)Ph₂ afforded new 1,2,3-triazoles. In the case of RCO₂CCCO₂R, a new macrocyclic product was isolated with loss of the R group.

Full text of DOI:10.1039/c3dt00068k

Dalton
Transactions
View Article Online
View Journal | View Issue
COMMUNICATION
Cycloaddition reactions between dicyclohexylboron
azide and alkynes†
Cite this: Dalton Trans., 2013, 42, 4795
Received 8th January 2013,
Accepted 11th February 2013
Rebecca L. Melen and Douglas W. Stephan*
DOI: 10.1039/c3dt00068k
www.rsc.org/dalton
The room temperature 1,3-dipolar cycloaddition reactions of the triazole derivatives of boron have been reported10 the chemistry
boron azide, Cy2BN3 with the electron-poor acetylenes RCO2Cu of boron azides and, in particular, their application in cyclo-
CCO2R, EtCuCCOMe and HCuCP(vO)Ph2 afforded new 1,2,3-tria- addition chemistry is very rare; currently, the only example is
zoles. In the case of RCO2CuCCO2R, a new macrocyclic product limited to the recent work by Curran et al.11 on the cyclo-
was isolated with loss of the R group.
addition chemistry of the electron-rich NHC-boryl azide reac-
tions with electron-poor alkynes, alkenes and nitriles to give
NHC-boryl triazoles, triazolidines, and tetrazoles respectively.
Herein, we explore the ability of the boron azide, Cy2BN3 (1)
to undergo click reactions with acetylenes under ambient con-
ditions. The boron azide 1 was conveniently prepared from the
1 : 1 stoichiometric reaction of Cy2BCl with Me3SiN3. Removal of
the Me3SiCl by-product in vacuo afforded pure 1 in high yields
(>90%) as a colourless oil.12 The 11B NMR spectrum displayed a
resonance at δ = 61 ppm. The azide 1 was found to be thermally
unstable decomposing with release of N2 above 55 °C.
Click chemistry is a powerful synthetic reaction in organic
chemistry which allows the rapid assembly of complex organic
compounds.1 Click reactions generate products in high yields
with relatively few by-products and support a wide range of
substituents and have seen widespread applications in drug
synthesis, material science, nanotechnology and polymers
inter alia.2 One example of click chemistry is the azide–alkyne
cycloaddition which selectively gives 1,2,3-triazoles which are
present in many pharmaceuticals, particularly antifungal
drugs.3 However, the thermal Huisgen 1,3-dipolar cyclo-
addition of alkynes with azides requires elevated temperatures
in order to complete the reaction which often results in a
mixture of regio-isomers.4 For example, the reaction between
methyl azide and propyne has a high activation barrier of
25–26 kcal mol−1.5 To overcome this problem, a Cu or Ru cata-
lyst is often employed in a metal-catalyzed Azide–Alkyne Cyclo-
addition (CuAAC or RuAAC respectively) offering rate
enhancements of up to 107 times with respect to the uncata-
lysed process.5–7 In the case of CuAAC this rate enhancement
is thought to be brought about by the formation of a Cu acety-
lide intermediate and as a result only terminal alkynes can be
used, however in RuAAC both terminal and internal alkynes
can partake in the reaction which proceeds via a ruthenacycle
intermediate.7,8
The first series of reactions explored were those involving
the reactions of Cy2BN3 with acetylenes RCuCH (R = Ph, p-tol,
4-tBuPh, SiMe3). However, initial attempts proved unsuccessful
1
at room temperature with no reaction observed by H and 11B
NMR spectroscopy even after one week. Raising the tempera-
ture to 50 °C exhibited no evident effect on the reaction by
NMR spectroscopy, whereas more elevated temperatures pro-
moted N2 release from the mixture and decomposition of the
boron azide. The lack of reactivity of 1 prompted us to utilise
more reactive acetylenes bearing a lower energy LUMO. The
room temperature stoichiometric reactions of 1 with the elec-
tron poor acetylenes EtCuCCOMe and Ph2P(vO)CuCH
cleanly led to the products 2 and 3 in moderate recovered
yields (Scheme 1). The synthesis of 2 appears to give the 1,4-
regio-isomer selectively as the only product on the basis of
in situ NMR spectra with no peaks evident due to the 1,5-product.
In the case of 3, both the 1,4- and 1,5-products are formed in a
3 : 2 ratio as seen in the crude 1H NMR spectrum before re-
crystallisation. The smaller proportion of the 1,5-regio-isomer
is presumably due to the steric conflict between the cyclohexyl
groups on boron and the phenyl groups on phosphorus.
Only a handful of boron azide compounds of the type R2BN3
(R = alkyl or aryl) have been reported previously9 and, although
Department of Chemistry, 80 St. George StreetUniversity of Toronto, Toronto,
Ontario, Canada, M5S 3H6. E-mail: dstephan@chem.utoronto.ca
†Electronic supplementary information (ESI) available: Experimental details,
Crystals of 2 suitable for X-ray diffraction were grown from
a saturated solution of 2 in CH2Cl2 at −35 °C and crystals of 3
were formed by the slow evaporation of the solvent from a
NMR data, DFT calculations and details of the crystal structure determination of
2, 3 and 5. CCDC 918102–918104. For ESI and crystallographic data in CIF or
other electronic format see DOI: 10.1039/c3dt00068k
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 4795–4798 | 4795

View Article Online
Communication
Dalton Transactions
Table 1 Geometric parameters for the triazole rings in 2, 3 and 5 in relation to
previously reported triazole structures found on the CSD
Compound
Averagea
2
3
5
Bond length/Å
N(1)–N(2)
N(2)–N(3)
N(3)–C(4)
C(5)–N(1)
C(4)–C(5)
1.348
1.436
1.383
1.371
1.400
1.352(2)
1.302(2)
1.350(3)
1.349(3)
1.395(3)
1.346(3)
1.311(3)
1.358(3)
1.333(3)
1.380(3)
1.332(2)
1.333(3)
1.351(3)
1.343(3)
1.370(3)
a Search of the CSD.
In both cases the fused C2N4B2 bicyclic frameworks are essen-
tially planar (max. deviation from planarity 0.054 Å for 2 and
0.084 Å for 3) with the unreacted azide moiety lying slightly
out of the plane of the heterocyclic ring; the plane of the BN3B
moiety forming an angle to the heterocyclic B2N2 plane of
2.62° and 8.15° for 2 and 3 respectively. However there are
some marked variations in N–N, C–N and C–C bond lengths in
relation to previously reported 1,2,3-triazoles (Table 1); the
bonding in triazoles typically exhibits C–N bond lengths con-
sistent with a delocalised bonding pattern comparable to pyri-
dine (C⋯N at 1.34 Å) rather than imines or amines (CvN at
1.28 Å and C–N at 1.47 Å). With the exception of the N(1)–N(2)
bond in 2 the structures of 2 and 3 exhibit shorter bonds than
in conventional triazoles with a particularly marked shorten-
ing of N(2)–N(3). The nature of the bonding within the fused
heterocycle was further probed by DFT (B3LYP/6-311G*+) cal-
culations and NBO studies (see ESI†). The NBO analysis based
on the gas-phase geometry-optimised structure revealed sub-
stantial delocalisation and a strongly polar structure with con-
siderable bis-imine character within the triazole ring (Fig. 2).
NMR spectroscopic studies of 2 and 3 show that each
species is intact in solution with no indication of dissociation
of the coordinating Cy2BN3 moiety. Even the addition of coor-
dinating solvents such as THF or pyridine did not promote the
dissociation of the Cy2BN3 group or azide (vide infra),
suggesting that the bicyclic frameworks of 2 and 3 are robust
and that the remaining azide moiety is deactivated towards a
further cycloaddition reaction. The lack of reactivity of 1 with
acetylenes bearing mildly electron-withdrawing groups indi-
cates that the acetylene–azide reaction occurs through inter-
action of the HOMO of the azide with the LUMO of the alkyne.
However the DFT calculations indicate that the HOMO of both
1 and 2 are of similar energy and both of π-non-bonding char-
acter. This suggests that the lack of reactivity may be due to
greater steric shielding (by the cyclohexyl groups) of the azide
Scheme 1 Synthesis of compounds 2 and 3.
Fig. 1 POV-ray depiction of 2 (top) and 3 (bottom). Hydrogen atoms are
omitted for clarity.
saturated solution of 3 in toluene. Both 2 and 3 crystallise in
ˉ
the triclinic space group P1 with one molecule in the asym-
metric unit (Fig. 1).‡ Previously, boron azides bearing bulky
substituents have been shown to be dimeric in solution and
the solid state.13 Thus the structures of both 2 and 3 can been
viewed as arising from the 1,3-dipolar cycloaddition of the
Fig. 2 NBO partial charges (left) and bond orders (right) for 2 based on the
alkyne with one azide group of the [Cy2BN3]2 molecule. DFT-optimised (B3LYP/6-311G*+) geometry.
4796 | Dalton Trans., 2013, 42, 4795–4798
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Communication
Scheme 2 Structures of 4–7.
Fig. 3 POV-ray depiction of the dimeric structure of 5. Hydrogen atoms are
omitted for clarity.
functionality in 2 than in 1, consistent with the increase in
ring size from a 4 to a 5-membered ring.
Interestingly, the high-resolution DART (and ESI+) mass CH2Cl2) and even the more polar coordinating solvents (water,
spectrometry on 2, did not show the presence of 2 but instead alcohols, DMSO and DMF). Owing to the insoluble nature of 5,
displayed peaks at m/z = 316.3 and 631.5 corresponding to solution state characterisation by either NMR or mass spectro-
[4 + H]+ and the dimer [4′ + H]+ (Scheme 2) with the correct scopy proved impossible but the composition of 5 was
isotope distribution pattern, suggesting that heating 2 may result additionally confirmed by elemental analysis and solid state
in the release of Cy2BN3 to generate 4 which could dimerise to NMR spectroscopy.
form 4′. Formation of 4′ may also occur via cycloaddition of a
The loss of alkyl groups in these cycloaddition reactions is
further equivalent of alkyne with 2. In order to investigate this unprecedented. Indeed, in the equivalent reaction reported by
reactivity, a solution of 2 and excess EtCuCOMe in d8-toluene Curran,11 there was no evidence for loss of alkyl groups during
1
was heated for 6 h at 80 °C. The resulting in situ H spectrum the cycloaddition of the NHC-stabilised boron azide with
showed that a different product was formed in high yield RCO2CuCCO2R (R = Me or Et). However, the recurrent for-
which we tentatively assign to the click product 4 or 4′ mation of 5 indicates that 5 is not only a thermodynamic sink
(Scheme 2), though crystals have proved elusive to date. In this but that this appears a preferred outcome for these ester
context, the formation of the mono-cycloaddition product 2 derivatives and would appear to suggest that the O-donor is
can be viewed as a stable intermediate on the way to the for- capable of displacing azide from the boron centre. Notably the
mation of the double cycloaddition product 4′.
ketone EtCuCCOMe does not react in the same manner nor
Subsequent investigations to assess further the effects of does addition of THF to 2 or 3 reveal any evident reactivity.
the electron withdrawing substituents on the acetylene, led us Formation of the triazole ring would indicate that the boron
to examine the 1 : 1 stoichiometric reactions of
1 with azide initially undergoes the expected cycloaddition with the
t
RCO2CuCCO2R (R = Me, Et, Bu) in toluene. In all cases the alkyne but the ester carbonyl appears then to undergo intra-
product was unexpectedly found to be the macrocycle 5. molecular coordination to the boron centre with concomitant
Storage of a saturated toluene solution at room temperature or elimination of the alkyl group from the ester.
at −35 °C afforded colourless cubic crystals of 5 suitable for
X-ray diffraction (Fig. 3). Changing the reaction stoichiometry and 1 : 3 [RCO2CuCCO2R (R = Me, Et):azide] stoichiometric
(1 : alkyne = 3 : 2) also afforded 5 but in lower isolated yields.
reactions were followed by in situ 1H NMR spectroscopy.
In an attempt to determine the fate of the R group the 1 : 1
Compound 5 crystallises in the triclinic space group P1.‡ Although the 1H NMR spectra are complicated by the presence
The structure comprises a central C8O4B2 macrocycle located of the cyclohexyl groups, throughout these experiments three
about a crystallographic inversion centre with two sets of three signals corresponding to two compounds in the 1H NMR spec-
crystallographically related fused 5-membered rings (Fig. 3). trum were always observed due to the ethyl (or methyl) groups.
The hepta-cyclic framework is essentially planar (max. devi- We assign these resonances tentatively to the model intermedi-
ation 0.145 Å). Whilst the 14-membered macrocycle offers a ates 6 and 7 in the formation of the macrocycle 5. Since the
potential O4-donor set for metal coordination, the cyclohexyl reactions were performed under rigorously anhydrous con-
groups appear to sterically hinder coordination (see later). ditions the loss of RO− by ester hydrolysis can be excluded. It
Compound 5 was found to be insoluble in all common organic is plausible that RN3 could be formed in the reaction, although
solvents (pentane, hexane, diethyl ether, toluene, THF, this was not observed presumably due to its high volatility.
ˉ
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 4795–4798 | 4797

View Article Online
Communication
Dalton Transactions
The macrocycle 5 was found to be remarkably robust being
air and water stable. Indeed, attempted alkylation reactions
with MeI, BnCl and EtBr at N(2) in the triazole ring and
attempts to encapsulate a Li+ cation within the macrocyclic
core have, to date, proved unsuccessful.
1 H. C. Kolb, M. G. Finn and K. B. Sharpless, Angew. Chem.,
Int. Ed., 2001, 40, 2004.
2 J. E. Moses and A. D. Moorhouse, Chem. Soc. Rev., 2007, 36,
1249; H. C. Kolb and K. B. Sharpless, Drug Discovery Today,
2003, 8, 1128; W. H. Binder and R. Sachsenhofer, Macro-
mol. Rapid Commun., 2007, 28, 15; J. F. Lutz, Angew. Chem.,
Int. Ed., 2007, 46, 1018; M. V. Gil, M. J. Arévalo and
O. López, Synthesis, 2007, 1589.
3 R. S. Bohacek, C. McMartin and W. C. Guida, Med. Res.
Rev., 1996, 16, 3.
4 K. V. Gothelf and K. A. Jorgensen, Chem. Rev., 1998, 98, 863.
5 F. Himo, T. Lovell, R. Hilgraf, V. V. Rostovtsev,
L. Noodleman, K. B. Sharpless and V. V. Fokin, J. Am.
Chem. Soc., 2005, 127, 210.
6 For examples of CuAAC see: C. Spiteri and J. E. Moses,
Angew. Chem., Int. Ed., 2010, 49, 31; V. V. Rostovtsev,
L. G. Green, V. V. Fokin and K. B. Sharpless, Angew. Chem.,
Int. Ed., 2002, 41, 2596; V. D. Bock, H. Hiemstra and
J. H. Van Maarseveen, Eur. J. Org. Chem., 2006, 51;
M. Meldal and C. W. Tornøe, Chem. Rev., 2008, 108, 2952;
J. E. Hein and V. V. Fokin, Chem. Soc. Rev., 2010, 39, 1302;
L. Liang and D. Astruc, Coord. Chem. Rev., 2011, 255, 2933.
7 For examples of RuAAC see: B. C. Boren, S. Narayan,
L. K. Rasmussen, L. Zhang, H. Zhao, Z. Lin, G. Jia and
V. V. Fokin, J. Am. Chem. Soc., 2008, 130, 8923; L. Zhang,
X. Chen, P. Xue, H. H. Y. Sun, I. D. Williams,
K. B. Sharpless, V. V. Fokin and G. Jia, J. Am. Chem. Soc.,
2005, 127, 15998; L. K. Rasmussen, B. C. Boren and
V. V. Fokin, Org. Lett., 2007, 9, 5337.
In conclusion studies of the reactivity of the boron azide
Cy2BN3 with acetylenes RCuCR reveal that electron-donating
substituents show no reactivity suggesting the dominant inter-
action can be considered as a normal electron-demand cyclo-
addition, i.e. the dominant orbital interaction is between the
HOMO of the 4π 1,3-dipole (azide) and the LUMO of the 2π
dipolarophile (alkyne). The use of electron deficient acetylenes
results in a lowering of the LUMO leading to a cycloaddition
reaction forming the expected triazole. Unlike other ‘click’
reactions of this type, these reactions proceed rapidly at room
temperature in the absence of a catalyst. The boron azide (1) is
considered to be dimeric in solution and the second azide
appears deactivated and NMR evidence for cycloaddition at
this second site only appears to occur at elevated temperatures.
In the case of RCO2CuCCO2R an unprecedented rearrange-
ment occurs generating an incredibly stable macrocyclic
product contain 7 fused rings via elimination of the alkyl (R)
groups. Perhaps most importantly this work demonstrates
fine-tuning of the electronics of the acetylene is key to the
modulation of the reactivity and the nature of the products
formed.
Acknowledgements
8 F. Himo, T. Lovell, R. Hilgraf, V. V. Rostovtsev,
L. Noodleman, K. B. Sharpless and V. V. Fokin, J. Am.
Chem. Soc., 2005, 127, 210.
We would like to thank Prof. M. Pilkington (University of
Brock for crystallographic data collection) and NSERC of
Canada is thanked for financial support. DWS is grateful for
the award of a Canada Research Chair.
9 P. I. Paetzold, Z. Anorg. Allg. Chem., 1966, 345, 79;
H. U. Meier, P. I. Paetzold and E. Schroeder, Chem. Ber.,
1984, 117, 1954; P. I. Paetzold, P. P. Habereder and
R. Muellbauer, J. Organomet. Chem., 1967, 7, 45;
J. Muenster, P. I. Paetzold, E. Schroeder, H. Schwan and
T. von Bennigsen-Mackiewicz, Z. Anorg. Allg. Chem., 2004,
630, 2641; P. I. Paetzold, Chem. Ber., 1983, 116, 1531;
S. Biswas, H. F. Bettinger and I. M. Oppel, Inorg. Chem.,
2010, 49, 449.
10 K. Niedenzu and K. R. Woodrum, Inorg. Chem., 1989, 28,
4022; C. P. Brock, A. L. Companion, L. D. Kock and
K. Niedenzu, Inorg. Chem., 1991, 30, 184.
11 E. Merling, V. Lamm, S. J. Geib, E. Lacôte and D. P. Curran,
Org. Lett., 2012, 14, 2690.
12 Covalent azides are potentially explosive and reactions were
performed on small scale behind blast shields. Cy2BN3·Py
has been synthesised previously from the reaction on (Cy)-
(Ph)BCl and NaN3 in pyridine. See: Gmelin Handbook: B:
B-Verb.4, 8, 260.
13 W. Fraenk, T. M. Klapötke, B. Krumm, P. Mayer, H. Nöth,
H. Piotrowski and M. Suter, J. Fluorine Chem., 2001, 112,
73; W. Fraenk, T. M. Klapötke, B. Krumm and P. Mayer,
Chem. Commun., 2000, 667.
Notes and references
‡Crystal data for 2, 3 and 5 were collected on a Bruker APEX-II diffractometer.
Structures were solved by direct methods and refined by full matrix least squares
based on F2 using the SHELXTL program package.14
ˉ
Crystal data for 2: C30H52B2N6O, M = 534.40, triclinic P1, a = 9.1297(4), b =
9.2734(4), c = 19.1974(8) Å, α = 79.937(2), β = 80.421(2), γ = 73.642(2)°, V =
1523.67(11) Å3, μ(Mo-Kα) = 0.71, T = 150(2) K, Z = 2, Dc = 1.165 Mg m−3, F(000) =
584, independent reflections 5345 (Rint = 0.026). Two of the four cyclohexyl
groups were found to be disordered over two sites and refined with 50 : 50 dis-
order and a constrained geometry R1 (I > 2σ(I)) = 0.055, wR2 (all data) = 0.136,
S = 1.192 (all data).
ˉ
Crystal data for 3: C38H55B2N6OP, M = 664.47, triclinic P1, a = 9.5954(13), b =
11.2785(15), c = 17.351(2) Å, α = 97.004(4), β = 97.151(4), γ = 101.004(5)°, V =
1808.3(4) Å3, μ(Mo-Kα) = 0.116, T = 150(2) K, Z = 2, Dc = 1.220 Mg m−3, F(000) =
716, independent reflections 6296 (Rint = 0.033). R1 (I > 2σ(I)) = 0.058, wR2 (all
data) = 0.123, S = 1.121 (all data).
ˉ
Crystal data for 5: C80H132B6N6O8, M = 1370.78, triclinic P1, a = 10.3184(10),
b = 14.6317(15), c = 15.0290(16) Å, α = 69.271(5), β = 88.761(5), γ = 75.300(5)°, V =
2046.8(4) Å3, μ(Mo-Kα) = 0.069, T = 150(2) K, Z = 1, Dc = 1.112 Mg m−3, F(000) =
748, independent reflections 7186 (Rint = 0.040). R1 (I > 2σ(I)) = 0.070, wR2 (all
data) = 0.161, S = 1.210 (all data).
14 SHELXTL, Bruker AXS, Madison, WI, USA.
4798 | Dalton Trans., 2013, 42, 4795–4798
This journal is © The Royal Society of Chemistry 2013