Catalytic Racemization of Activated Organic Azides

Article abstract of DOI:10.1021/acs.orglett.8b03168

The first detailed description of the catalytic racemization of activated benzylic and allylic azides under mild conditions is reported. A kinetic analysis of the observed racemization indicates a first-order dependence on azide, a first-order dependence

Full text of DOI:10.1021/acs.orglett.8b03168

Letter
pubs.acs.org/OrgLett
Cite This: Org. Lett. XXXX, XXX, XXX−XXX
Catalytic Racemization of Activated Organic Azides
Amy A. Ott and Joseph J. Topczewski*
Department of Chemistry, University of Minnesota Twin Cities, Minneapolis, Minnesota 55455, United States
S
* Supporting Information
ABSTRACT: The first detailed description of the catalytic racemization of
activated benzylic and allylic azides under mild conditions is reported. A
kinetic analysis of the observed racemization indicates a first-order
dependence on azide, a first-order dependence on catalyst, and that the
rate of racemization correlates to σ+. A variety of azides with varying substitution patterns undergo facile racemization, and
catalyst selection can tune this process.
Scheme 1. Mechanisms of Azide Rearrangement and
Racemization
ynthetic methods that provide direct access to enantioen-
Sriched products are highly desired. A powerful platform for
enantioselective synthesis is dynamic kinetic resolution
(DKR).1−6 DKR is especially powerful because racemic
starting materials can be converted to enantioenriched
products in both high yield and high selectivity. This
surmounts the 50% maximum theoretical yield barrier inherent
to kinetic resolution.7 A mechanism for rapid racemization is
crucial for DKR, which has limited this process to substrates
containing easily racemizeable stereocenters. Common DKR
substrates include carbonyl containing molecules with an α-
stereocenter.8−10 Alternatively, chiral secondary alcohols can
be racemized via an oxidation/reduction sequence.3 While
other mechanisms of racemization are known,1,2,11 the
applications of DKR remain limited. Identifying new
mechanisms of facile racemization for new substrate classes
could potentially open unexplored avenues for the develop-
ment of methods predicated on DKR. Azides are particularly
interesting substrates because they are chiral amine equiv-
alents, and chiral amines are an essential motif in many drugs
and drug candidates.
racemized. This report provides the first detailed account
documenting the catalytic racemization of organic azides.
Several catalysts for racemization are described that proceed
with an exceptional rate enhancement relative to the solvolysis
pathway (Scheme 1b vs 1c). This process is broad in scope and
is tunable based on catalyst selection.
We began this study with allylic azide 1a (Table 1). Prior
work indicated that this azide is configurationally stable in
nonpolar solvents at 100 °C for >1 week.18 Catalysts were
examined in dichloromethane at 50 °C so that any observed
racemization could be attributed to the catalyst. Azide 1a
(>99:1 er) was exposed to 10 mol % of various possible
catalysts for 24 h, and the reaction was monitored by chiral
HPLC. Initially, transition metal catalysts were investigated
(entries 1−4). As part of a report on an allylic substitution
reactions, Trost and co-workers suggested that the presence of
palladium may accelerate the allylic azide rearrangement.19
Exposure of azide 1a to Pd(0) sources led to minimal
racemization (entries 1−2). Exposure to a more electrophilic
Pd(II) salt led to considerable racemization along with
significant azide decomposition (entry 3). Based on Liu’s
observation,20 Lewis acidic salts were investigated (entries 5
and 6). Several of these catalysts resulted in racemization along
Recently, our group has become interested in using organic
azides as amine surrogates for DKR.12−14 We hypothesized
that the Winstein rearrangement15−17 could serve as a suitable
racemization pathway. An initial example afforded tertiary
azides in high er.12 These promising results led us to ask
fundamental questions regarding the mechanism of the allylic
azide rearrangement. Under most conditions, the rearrange-
ment occurred via a [3,3] sigmatropic pathway and no
racemization was observed (Scheme 1a).18 A second solvolysis
pathway led to azide racemization, which proceeded slowly in
methanol at 100 °C (Scheme 1b).18 While this was exciting,
racemization was slow, required elevated temperatures, and
resulted in significant decomposition of the azide through
methanolysis. Therefore, these conditions could not be readily
adapted to DKR. Nevertheless, the observed racemization
prompted a search for a suitable racemization catalyst.
We sought to identify a catalyst for azide racemization that
would (i) significantly increase the rate of racemization, (ii)
proceed without significant decomposition of the azide, and
(iii) broaden the scope of azides that could be efficiently
Received: October 4, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.8b03168
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Other cationic gold(I) complexes were investigated (Table
1, entries 11−14). These results indicate that other gold(I)
complexes with noncoordinating counterions are suitable for
racemization. Significantly, minimal azide decomposition was
observed even with rapid racemization. Decreasing the catalyst
loading slowed racemization (entry 15 and SI); however,
activity could be restored by increasing the temperature and
concentration (entry 16). As a control experiment, AgSbF6 was
used as the catalyst (entry 17), which led to both racemization
and significant azide decomposition. We proceeded to study
catalyst 4 because it is readily available and fully soluble,
provided rapid racemization at room temperature, and caused
minimal azide decomposition.
Table 1. Initial Screen for Racemization Catalyst
a
b
entry
catalyst
temp
er
% azide remaining
With a catalyst identified, we were interested in illuminating
the mechanism of racemization. Using the method of initial
rates,22 the reaction was found to be first-order in both azide
and gold catalyst. To gain more insight, a series of substituted
azides were synthesized (Figure 1, 1a−1e). For each azide, the
1
2
3
4
5
6
7
8
9
10
11
12
13
Pd(PPh3)4
Pd2(dba)3
PdCl2(NCPh)2
(cod)2RhOTf
Cu(OTf)2
Zn(OTf)2
Tf2NH
catalyst 1
Zn(OTf)2
catalyst 1
Ph3PAuSbF6
catalyst 2
catalyst 3
catalyst 4
catalyst 4
catalyst 4
AgSbF6
50 °C
50 °C
50 °C
50 °C
50 °C
50 °C
50 °C
50 °C
rt
rt
rt
rt
rt
81:19
98:2
68
89
76
23
61
86
65
82
94
88
84
75
85
97
99
99
72
64:36
59:41
60:40
50:50
82:18
50:50
82:18
50:50
50:50
52:48
78:22
57:43
98:2
c
c
c
c
14
rt
rt
50 °C
rt
cd
,
15
16
17
e
60:40
50:50
a
b
Measured using chiral HPLC analysis. Determined by calibrated
HPLC analysis using 1,1-di-tert-butylbiphenyl as a standard. Reactions
Figure 1. Plot of log(k/k0) vs σ+ for substituent effects on the
racemization of azides 1a−1e using catalyst 4. Rates were measured in
duplicate, replicates shown.
were run in duplicate, and the average yield and er values are reported.
c
d
Room temperature = rt. Reaction time of 10 min. Reaction with 1
e
mol % catalyst. Reaction with 1 mol % catalyst at 0.1 M [N3] and at 3
h.
initial rate of racemization was measured using chiral HPLC
analysis. The rate of racemization correlated to the Hammett
parameter σ+ (Figure 1, R2 = 0.97, ρ = −5.9), indicating
buildup of positive charge in the transition state. The large
negative ρ value is consistent with formation of a carbocation
intermediate.23 Based on the known binding mode of a similar
gold complex to adamantyl azide24 and the kinetic analysis, a
mechanism is proposed (Scheme 2). The gold catalyst likely
binds the azide (step (i)), ionizes the azide (step (ii)),
recombines with the carbocation (step (iii)), and dissociates
(step (iv)).
with appreciable decomposition of azide 1a (entries 4−6 and
Supporting Information). Gratifyingly, Zn(OTf)2 led to full
racemization with minimal azide decomposition (entry 6).
Triflamide, as a Brønsted acid catalyst, was found to be
minimally effective (entry 7). A report by Toste and co-
workers suggested that cationic gold salts may cause partial
racemization of some azides.21 Catalyst 1 was chosen as a
representative gold(I) complex. Exposure of azide 1a to 10 mol
% of this catalyst led to complete racemization with minimal
decomposition of the azide. Based on these results, Zn(OTf)2
and gold catalyst 1 were selected for further study.
Other organic azides could be racemized under these gold
catalyzed conditions (Table 2). In order to tease apart subtle
A secondary screen was conducted to determine whether
Zn(OTf)2 or catalyst 1 was more efficient (Table 1, entries 9
and 10). Racemization was slow with Zn(OTf)2 at room
temperature (entry 9). However, full racemization occurred
with gold catalyst 1 at room temperature over 24 h (entry 10).
This indicated a greater rate of acceleration with the gold(I)
catalyst relative to Zn(OTf)2. A more detailed kinetic analysis
indicated that catalyst 1 is capable of full racemization within
10 min at room temperature. This is considerably faster than
the uncatalyzed solvolysis conditions, which proceeded at 100
°C in methanol over several hours (Scheme 1b). This clearly
indicates a remarkable acceleration of the racemization rate to
a level that could conceivably be compatible with DKR
applications.
Scheme 2. Proposed Mechanism for Azide Racemization
B
DOI: 10.1021/acs.orglett.8b03168
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
verified (entry 2). Other tertiary benzylic azides demonstrated
a range of activity based on the substituents (entries 3−5).
Diaryl secondary azides reached a 50:50 er within 6 h (entries
6 and 7). These results are particularly exciting because they
demonstrate that this process is not restricted to allylic azides
and is operable on a much broader class of compounds. Less
activated azides also reacted, albeit at a decreased rate under
these dilute conditions (entries 8 and 9). Other activated
secondary azides were stable under these conditions (entries
10−12).
Table 2. Scope of Azides That Can Be Racemized
With several suitable catalysts identified in the initial screen,
we hypothesized that the catalyst and substrate combination
could be fine-tuned to (i) increase the reactivity of less
activated azides or (ii) prevent decomposition for the most
activated azides. For model substrate 1a, the use of a decreased
AgPF6 loading resulted in racemization with less decom-
position (Table 3, entry 1). Furthermore, the reactivity of the
Table 3. Alternate Catalysts for Racemization
a
a
Determined by HPLC analysis using 1,1-di-tert-butylbiphenyl as a
Determined by HPLC analysis using 1,1-di-tert-butylbiphenyl as a
standard. Reactions were run in duplicate, and the average yield and
er values are reported. Measured using chiral HPLC analysis. Time
points are 10 min and 1 h at 6 mM [N3]. Reaction with 3 mol %
catalyst loading. Time points are 1 and 3 h. Room temperature = rt.
standard. Reactions were run in duplicate, and the average yield and
er values are reported. Measured using chiral HPLC analysis. Time
points are 1, 5, and 10 min. Time points are 1 and 3.5 h. Yield
determined by GC analysis. Time points are 1, 3.5, and 6 h. Room
b
c
b
c
d
d
e
e
f
temperature = rt, PMP = para-methoxyphenyl.
silver salt could be tuned by the addition of ligand (entry 2).
Secondary azides 11 and 13 were stable to catalyst 4, but
treatment with AgPF6 resulted in near-complete racemization
in under 5 h at room temperature (entries 3 and 5).
Racemization could be further accelerated by mildly heating
the reaction (entry 4). Racemization of azide 10 was sluggish
under the initially optimized conditions, but increasing the
differences in the rate of racemization, the following experi-
ments were conducted under dilute conditions (5 mM). As
expected based on the Hammett correlation, a tertiary allylic
azide underwent rapid racemization (entry 1). A doubly
benzylic and tertiary azide racemized and gradually decom-
posed to the corresponding styrene, which was independently
C
DOI: 10.1021/acs.orglett.8b03168
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(7) Keith, J. M.; Larrow, J. F.; Jacobsen, E. N. Practical
Considerations in Kinetic Resolution Reactions. Adv. Synth. Catal.
2001, 343, 5−26.
(8) Tai, A.; Watanabe, H.; Harada, T. Stereochemical Studies of the
Hydrogenation with Asymmetrically Modified Nickel Catalysts; The
Hydrogenation of Methyl 2-Alkyl-3-Oxobutyrate. Bull. Chem. Soc. Jpn.
1979, 52, 1468−1472.
(9) Noyori, R.; Ikeda, T.; Ohkuma, T.; Widhalm, M.; Kitamura, M.;
Takaya, H.; Akutagawa, S.; Sayo, N.; Saito, T.; Taketomi, T.; et al.
Stereoselective Hydrogenation via Dynamic Kinetic Resolution. J. Am.
Chem. Soc. 1999, 111, 9134−9135.
temperature and concentration resulted in moderate race-
mization in 5 h (entry 6). Highly active substrates can be
efficiently racemized under more mild conditions with
Zn(OTf)2 (entries 7−8). While azide 2 began to decompose
using catalyst 4, Zn(OTf)2 resulted in minimal decomposition
(entry 8).
Herein we disclose the catalytic racemization of activated
organic azides. The rate of the reaction is correlated to the
Hammett constant σ+, implicating the involvement of a
carbocation intermediate. The reactivity can be modulated
by the choice of catalyst. Investigations into utilizing this
racemization for a downstream functionalization are currently
underway.
(10) Jurkauskas, V.; Buchwald, S. L. Dynamic Kinetic Resolution via
Asymmetric Conjugate Reduction : Enantio- and Diastereoselective
Synthesis of 2, 4-Dialkyl Cyclopentanones. J. Am. Chem. Soc. 2002,
124, 2892−2893.
́
(11) Carmona, J. A.; Hornillos, V.; Ramírez-Lopez, P.; Ros, A.;
ASSOCIATED CONTENT
́ ́
enza, J.; Gomez-Bengoa, E.; Fernandez, R.; Lassaletta, J.
■
Iglesias-Sigu
̈
M. Dynamic Kinetic Asymmetric Heck Reaction for the Simultaneous
Generation of Central and Axial Chirality. J. Am. Chem. Soc. 2018,
140, 11067−11075.
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.8b03168.
(12) Ott, A. A.; Goshey, C. S.; Topczewski, J. J. Dynamic Kinetic
Resolution of Allylic Azides via Asymmetric Dihydroxylation. J. Am.
Chem. Soc. 2017, 139, 7737−7740.
Experimental procedures and data (PDF)
(13) Porter, M. R.; Shaker, R. M.; Calcanas, C.; Topczewski, J. J.
Stereoselective Dynamic Cyclization of Allylic Azides : Synthesis of
Tetralins, Chromanes, and Tetrahydroquinolines. J. Am. Chem. Soc.
2018, 140, 1211−1214.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: jtopczew@umn.edu.
ORCID
(14) Ding, P. G.; Hu, X. S.; Zhou, F.; Zhou, J. Catalytic
Enantioselective Synthesis of α-Chiral Azides. Org. Chem. Front.
2018, 5, 1542−1559.
(15) Gagneux, A.; Winstein, S.; Young, W. G. Rearrangement of
Allyl Azides. J. Am. Chem. Soc. 1960, 82, 5956−5957.
(16) Feldman, A. K.; Colasson, B.; Sharpless, K. B.; Fokin, V. V. The
Allylic Azide Rearrangement: Achieving Selectivity. J. Am. Chem. Soc.
2005, 127, 13444−13445.
Amy A. Ott: 0000-0002-2018-3276
Joseph J. Topczewski: 0000-0002-9921-5102
Notes
́
(17) Liu, R.; Gutierrez, O.; Tantillo, D. J.; Aube, J. Stereocontrol in a
The authors declare no competing financial interest.
Combined Allylic Azide Rearrangement and Intramolecular Schmidt
Reaction. J. Am. Chem. Soc. 2012, 134, 6528−6531.
ACKNOWLEDGMENTS
(18) Ott, A. A.; Packard, M. H.; Ortuno, M. A.; Johnson, A.; Suding,
̃
■
V. P.; Cramer, C. J.; Topczewski, J. J. Evidence for a Sigmatropic and
an Ionic Pathway in the Winstein Rearrangement. J. Org. Chem. 2018,
83, 8214−8224.
We thank Prof. Megan Fieser (University of Southern
California) for her many helpful discussions. Financial support
was provided by the University of Minnesota and The
American Chemical Society’s Petroleum Research Fund
(PRF No. 56505-DNI1). This research was supported by the
National Institute of General Medical Sciences of the National
Institutes of Health under Award Number R35GM124718. We
also acknowledge NIH Shared Instrumentation Grant No.
S10OD011952.
(19) Trost, B. M.; Pulley, S. R. On the Flexibility of Allylic Azides as
Synthetic Intermediates. Tetrahedron Lett. 1995, 36, 8737−8740.
(20) Liu, R. A Combined Allylic Azide Rearrangement and
Intramolecular Schmidt Reaction - Discovery, Development, And
Application; Ph.D Dissertation, University of Kansas, 2003.
(21) Lalonde, R. L.; Sherry, B. D.; Kang, E. J.; Toste, F. D. Gold (I)-
Catalyzed Enantioselective Intramolecular Hydroamination of Al-
lenes. J. Am. Chem. Soc. 2007, 129, 2452−2453.
(22) Espenson, J. H. Chemical Kinetics and Reaction Mechanisms;
McGraw-Hill, Inc.: New York, 1995.
(23) Jaffe, H. H. A Reexamination of the Hammett Equation. Chem.
REFERENCES
■
(1) Pellissier, H. Recent Developments in Dynamic Kinetic
Resolution. Tetrahedron 2011, 67, 3769−3802.
(2) Pellissier, H. Dynamic Kinetic Resolution. Tetrahedron 2003, 59,
8291−8327.
Rev. 1953, 53, 191−261.
(24) Dash, C.; Yousufuddin, M.; Cundari, T. R.; Dias, H. V. R. Gold-
Mediated Expulsion of Dinitrogen from Organic Azides. J. Am. Chem.
Soc. 2013, 135, 15479−15488.
̈
(3) Verho, O.; Backvall, J. E. Chemoenzymatic Dynamic Kinetic
Resolution: A Powerful Tool for the Preparation of Enantiomerically
Pure Alcohols and Amines. J. Am. Chem. Soc. 2015, 137, 3996−4009.
(4) Bhat, V.; Welin, E. R.; Guo, X.; Stoltz, B. M. Advances in
Stereoconvergent Catalysis from 2005 to 2015: Transition-Metal-
Mediated Stereoablative Reactions, Dynamic Kinetic Resolutions, and
Dynamic Kinetic Asymmetric Transformations. Chem. Rev. 2017, 117,
4528−4561.
(5) El Gihani, M. T.; Williams, J. M. J. Dynamic Kinetic Resolution.
Curr. Opin. Chem. Biol. 1999, 3, 11−15.
(6) Huerta, F.; Minidis, A.; Backvall, J.-E. Racemisation in
̈
Asymmetric Synthesis. Dynamic Kinetic Resolution and Related
Processes in Enzyme and Metal Catalysis. Chem. Soc. Rev. 2001, 30,
321−331.
D
DOI: 10.1021/acs.orglett.8b03168
Org. Lett. XXXX, XXX, XXX−XXX