Highly efficient photochemically induced thiyl radical-mediated racemization of aliphatic amines at 30 °C

Article abstract of DOI:10.1021/jo702241y

(Chemical Equation Presented) UV irradiation in the presence of thiol enables the performance of highly efficient aliphatic amines racemization, under mild conditions at 30°C. The reaction proceeds via the reversible generation of prochiral α-aminoalkyl r

Full text of DOI:10.1021/jo702241y

Highly Efficient Photochemically Induced Thiyl Radical-Mediated
Racemization of Aliphatic Amines at 30 °C
Lucie Routaboul,† Nicolas Vanthuyne,‡ Ste´phane Gastaldi,*,† Ge´rard Gil,*,‡ and
Miche`le Bertrand*,†
Laboratoire de Chimie Mole´culaire Organique, UMR 6517, Chimie, Biologie, et Radicaux Libres,
Boite 562, UniVersite´ Paul Ce´zanne, Aix-Marseille III, Faculte´ des Sciences St Je´roˆme,
AVenue Escadrille Normandie-Niemen, 13397 Marseille Cedex 20, France, and Laboratoire de
Ste´re´ochimie Dynamique et Chiralite´, UMR 6180, Chirotechnologies: Catalyse et Biocatalyse,
UniVersite´ Paul Ce´zanne, Aix-Marseille III, Faculte´ des Sciences St Je´roˆme, AVenue Escadrille
Normandie-Niemen, 13397 Marseille Cedex 20, France
michele.bertrand@uniV-cezanne.fr
ReceiVed October 17, 2007
UV irradiation in the presence of thiol enables the performance of highly efficient aliphatic amines
racemization, under mild conditions at 30 °C. The reaction proceeds via the reversible generation of
prochiral R-aminoalkyl radicals. The latter may result either from a redox process between the thiyl
radical and the amine or from direct hydrogen atom abstraction by thiyl radical. As hydrogen atom donor,
the thiol plays a crucial role. While the racemization of both primary and secondary amines were fast
processes, the racemization of tertiary amines was sluggish. A tentative rationale is based on the
photostimulated amine-catalyzed oxidation of the thiol into the corresponding disulfide, which makes
the hydrogen atom donor concentration in the reaction medium drop up to trace amount at a rate that
depends on the nature of the amine.
Introduction
of a thermostable lipase.4 With the view to extend the scope of
this DKR methodology by using other enzymes, we have
investigated the feasibility of thiyl radical-mediated racemiza-
tion at lower temperatures. To the best of our knowledge,
aliphatic amine racemization proceeding at temperatures lower
than 70 °C under mild conditions, in particular in the absence
of strong bases, is unprecedented in the literature.5
Racemization of amines is a field of major importance to
industrial chemistry.1 The study of thiyl radical-mediated amine
racemization is an ongoing research topic in our group.2 We
have recently reported that the association of this radical process
with enzymatic resolution by CAL-B enabled the synthesis of
optically pure amides in good yields and high enantiomeric
excesses.3 Since the racemization was carried out at 80 °C using
AIBN as radical initiator, the success of the dynamic kinetic
resolution (DKR) process was intimately connected to the use
Results and Discussion
Without going into the details of how the R-aminoalkyl
intermediate radical is formed, the racemization process can be
summarized according to Scheme 1.
† Laboratoire de Chimie Mole´culaire Organique.
‡ Laboratoire de Ste´re´ochimie Dynamique et Chiralite´.
(4) For DKR, see: (a) Reetz, M. T.; Schimossek, K. I. Chimia 1996,
50, 668-669. (b) Choi, Y. K.; Kim, M. J.; Ahn, Y.; Kim, M.-J. Org. Lett.
2001, 3, 4099-4101. (c) Kim, M.-J.; Kim, W.-H.; Han, K.; Choi, Y. K.;
Park, J. Org. Lett. 2007, 9, 1157-1159. (d) Paetzold, J.; Ba¨ckvall, J.-E. J.
Am. Chem. Soc. 2005, 127, 17620-17621. (e) Parvulescu, A.; De Vos, D.;
Jacobs, P. Chem. Commun. 2005, 5307-5309. (f) Parvulescu, A. N.; Jacobs,
P. A.; De Vos, D. E. Chem.sEur. J. 2007, 3, 2034-2043. (g) Stirling, M.;
Blacker, J.; Page, M. I. Tetrahedron Lett. 2007, 48, 1247-1250. (h)
Crawford, J. B.; Skerlj. R. T.; Bridger, G. J. J. Org. Chem. 2007, 72, 669-
671.
(1) (a) Lawrence, S. A. In Amines: Synthesis, Properties and Applica-
tions; Cambridge University Press: Cambridge, 2004. (b) Breuer, M.;
Ditrich, K.; Habicher, T.; Hauer, B.; Kesseler, M.; Stu¨rmer, R.; Zelinski,
T. Angew. Chem., Int. Ed. 2004, 43, 788-824. (c) Hieber, G.; Ditrich, K.
Chim. Oggi 2001, 19, 16-20.
(2) (a) Escoubet, S.; Gastaldi, S.; Vanthuyne, N.; Gil, G.; Siri, D.;
Bertrand, M. P. J. Org. Chem. 2006, 71, 7288-7292. (b) Escoubet, S.;
Gastaldi, S.; Vanthuyne, N.; Gil, G.; Siri, D.; Bertrand, M. P. Eur. J. Org.
Chem. 2006, 3242-3250.
(3) Gastaldi, S.; Escoubet, S.; Vanthuyne, N.; Gil, G.; Bertrand, M. P.
Org. Lett. 2007, 9, 837-839.
(5) Ebbers, E.; Ariaans, G. J. A.; Houbiers, J. P. M.; Brugginks, A.;
Zwanenburg, B. Tetrahedron 1997, 53, 9417-9476.
10.1021/jo702241y CCC: $40.75 © 2008 American Chemical Society
Published on Web 12/13/2007
364
J. Org. Chem. 2008, 73, 364-368

Thiyl Radical-Mediated Racemization of Aliphatic Amines
SCHEME 1. Racemization Mechanism
2-Amino-4-phenylbutane (1) and its secondary and tertiary
derivatives (2 and 3) were selected as models to test different
experimental conditions. The procedure was then applied to a
series of related aliphatic amines bearing different functionalities
(Figure 1).
We first examined the possibility to use diazenes known to
decompose at low temperatures. 2,2′-Azobis(2,4-dimethyl-4-
methoxyvaleronitrile (V70), whose half-life is 10 h at 37 °C,6
was selected and used at 50 °C in degassed benzene. The results
were quite disappointing since it was necessary to add up to 1
equiv of V70 in five portions (0.2 equiv each 2 h from t0) over
8 h to only partially racemize amine 1 (28% ee and 17% ee
with methyl thioglycolate and cyclohexanethiol, respectively,
after 10 h at 50 °C). At a concentration 10 times higher than
the concentration used in the previous experiments (0.67 M
instead 0.067 M), amine 1 could be racemized in 6 h (in the
presence of 0.8 equiv of V70), but purification was made
difficult due to the formation of side products resulting from
the initiator, and no further attempt was made.7 In fact, at 50
°C the rate of hydrogen atom abstraction by thiyl radical was
too slow for the radical chain mechanism to proceed efficiently.
The slow rate and the necessity to use a large amount of initiator
made the procedure prohibitive for any application at preparative
scale.
FIGURE 1. Amine structures.
Therefore, we turned our attention to photochemical initiation,
which should enable direct homolytic cleavage of the S-H bond
and continuous production of thiyl radical.8 The reactions were
carried out in a benzene solution containing the amine (1, 2, or
3; 0.067 M) and 1.2 equiv of thiol (MeOCOCH2SH, C6H11SH,
or n-OctSH) in a quartz tube surrounded by 16 RPR-3000 UV
lamps in a Rayonet apparatus.9 Blank experiments demonstrated
without ambiguity that no racemization occurred when the
amines were irradiated alone in the absence of thiol.
In literature data, photolytic cleavage of the S-H bond is
generally performed with a low-pressure mercury lamp by
irradiating at 254 nm.8 We observed that racemization could
be performed in a quartz reactor with our apparatus. The
incidence of filtrating wavelengths below 300 nm by using Pyrex
tubes instead of quartz ones was also investigated in the absence
of any additional radical initiator.
FIGURE 2. Plots of amine 1 enantiomeric excess (ee) versus time,
during irradiation in the presence of octanethiol, cyclohexanethiol, and
methyl thioglycolate at 47 °C in a quartz reactor.
1.2 equiv of thiol was used. Conversely, the enantiomeric excess
never was less than 19% with n-OctSH and 25% with C6H11SH
after irradiation for 6 h. Owing to the less efficient radiative
intensity when wavelengths below 300 nm were filtered, the
racemization process was slowed down when irradiations were
carried out in a pyrex tube.10 In all cases, whenever racemization
was not completed, the ee seemed to reach a limit that could
not be overpassed upon prolonged irradiation without introduc-
ing and additional amount of thiol in the reaction medium.
If one admits that the racemization proceeds through revers-
ible hydrogen atom abstraction from the amine by thiyl radical,
then the initial rate should be related to the strength of the S-H
bond that is formed. The slowing of the racemization rate should
therefore be expected when going from methyl thioglycolate
to alkanethiols since the S-H bond dissociation enthalpy (BDE)
is stronger in methyl thioglycolate than in the other aliphatic
thiols.11 This implicitly admits that the forward hydrogen
abstraction is rate limiting for the process, at least at the initial
stage of the reaction. This is in accordance with experimental
observations. The slight difference observed between n-OctSH
and C6H11SH is likely to reside more in their compared steric
bulk than in the difference between their S-H BDEs. In regard
Figure 2 gives a survey of the influence of the nature of the
thiol on the racemization rate. The graphs show that methyl
thioglycolate is far more efficient than n-OctSH, itself slightly
more efficient than C6H11SH. With the former, amine 1
racemization was nearly completed in 3 h in a quartz tube when
(6) Feray, L.; Bertrand, M. In e-Encyclopedia of Reagents for Organic
Synthesis; Paquette, L. A., Crich, D., Fuchs, P., Molander, G., Eds.;
Wiley: Chichester, 2005; DOI:10.1002/047084289X.m00633.
(7) The 1H NMR spectrum of the crude mixture showed that amine was
degraded. The residual amine was difficult to quantify with respect to the
internal standard owing to the complexity of the mixture.
(8) For the photoinitiated cyclization of thiyl radicals, see: (a) Surzur,
J.-M. In ReactiVe Intermediates; Abramovitch, R. A., Ed.; Plenum: New
York, 1981; Vol. 2, Chapter 3, pp 177-186 and references cited therein.
(b) Kirpichanko, S. V.; Tolstikova, L. L.; Suslova, E. N.; Voronskov, M.
G. Tetrahedron Lett. 1993, 34, 3889-3892.
(10) In the presence of methyl thioglycolate, amine 1 racemization
reached completion in 5 h in a Pyrex vessel. The final ee was only 45% in
the presence of cyclohexane thiol under the same experimental conditions.
(11) Calculated S-H bond dissociation energies at 298 K according to
the G3B3(MP2) level of theory are 359.9 and 364.4 kJ/mol for n-BuSH
and HSCH2CO2Me, respectively; see ref 2.
(9) Spectral distribution ranges from 250 to 350 nm with maximum
intensity of irradiance at 300 nm.
J. Org. Chem, Vol. 73, No. 2, 2008 365

Routaboul et al.
to the use of a catalytic amount of thiol, the reaction rate is
likely to be limited in this case by the rate of the backward
hydrogen transfer from the thiol to the R-amino radical.
Whatever the intimate mechanism is, the thiol concentration
is critical to ensure a complete racemization. The evolution of
the amounts of thiol and of the corresponding disulfide in the
reaction medium will be discussed later on.
An alternative pathway might involve electron transfer from
the amine to thiyl radical immediately followed by proton
abstraction by thiolate anion according eqs 5 and 6 (see below).
If a photostimulated (or a spontaneous) charge transfer from
the amine toward thiyl radical was responsible for the efficacy
of the process, then the rate of the reaction should be affected
both by the reduction potential of the thiyl radical intermediate12
and by the oxidation potential of the amine.13 According to these
values, methyl thioglycolate should be the most efficient thiol
in the series, but one would expect the tertiary amine to be more
reactive, which is not the case.
The influence of the nature of the amine on racemization was
investigated. The comparative data are shown in Figure 3 for
the reactions carried out on amines 1-3, 12, and 7 in a quartz
vessel. Plots already given for amine 1 have been reproduced
to facilitate the comparison. For this reason, all experiments
were carried out at 47 °C.
Amine 2 behaved rather similarly to amine 1 as regard to
racemization in the presence of methyl thioglycolate. But the
racemization of secondary amine 2 was faster than the racem-
ization of primary amine 1 in the presence of cyclohexane thiol.
The racemization of amine 3 was much slower, and what is
more, amine 3 was partially degraded when the reaction was
performed in the presence of methyl thioglycolate. No degrada-
tion, but at the same time no racemization at all, was registered
when amine 3 was irradiated in a pyrex tube whatever thiol
was used. Thereby the class of the amine has a very significant
impact on the rate of the photochemically induced racemization
process.
Owing to the presence of the electron-withdrawing carbethoxy
groups in γ-position relative to the reactive center, the behavior
of amines 2 and 3 might be considered as anomalous; therefore,
amines 12 and 7 racemizations were also monitored. Racem-
ization in the presence of methylthioglycolate was slightly faster
for amine 12 than for amine 2. Although faster than with amine
3 the racemization of tertiary amine 7 did not reach completion.
The final ee was 28% after 3 h.
FIGURE 3. Comparative plots of amine ee versus time for amines
1-3, 12, and 7 (quartz vessel, 47 °C) in the presence of 1.2 equiv of
C6H11SH (Figure 3a) or methyl thioglycolate (Figure 3b).
two electron withdrawing groups in position â relative to the
nitrogen atom, became significant at lower temperature.
Several tracks could be followed to propose a tentative
rationale to these experimental observations. Recent experi-
mental data and theoretical calculations came to the conclusion
that the class of the amine had little influence of the R-CH
BDE.14,2b This was in contradiction with previous reports that
stated the R-C-H BDE decreased in the order primary amine
> secondary amine > tertiary amine.15 Interpreting our data
according to R-C-H bond strength would contradict both since
racemization was faster for the secondary amine and the primary
amine than for the tertiary one.16
Our initial choice of secondary amine 2 and tertiary amine 3
as models was essentially governed by their easy access from
amine 1 via Michael addition to ethyl acrylate. The racemization
process, initiated thermally at 80 °C in the presence of AIBN
and a stoichiometric amount of thiol, was as efficient for amine
1 as for amines 2 and 3.2b However, the influence of one or
(12) The electron affinities of thiyl radicals were shown to give linear
correlation with S-H BDEs and with proton affinities; see: (a) Fattahi, A.;
Kass, S. R. J. Org. Chem. 2004, 69, 9176-9183. (b) Janousek, B. K.; Reed,
K. J.; Brauman, J. I. J. Am. Chem. Soc. 1980, 102, 3125-3129.
(13) The oxidation potentials of primary, secondary, and tertiary amines
have been calculated by Fu et al.; see: (a) Fu, Y.; Liu, L.; Yu, H.-Z.; Wang,
Y.-M.; Guo, Q.-X. J. Am. Chem. Soc. 2005, 127, 7227-7234. The calculated
IP determined in acetonitrile for n-propylamine, di-n-propylamine, and tri-
n-propylamine at the B3LYP/6311++G(2df,2p)//B3LYP/6-31G(d) level of
theory, using the PCM solvation model, are 8.81, 8.03, and 7.38 eV,
respectively. The corresponding E0 vs NHE values are 1.63, 1.43, and 1.02
eV. The values are very close to experimental ones. For a quantitative
approach to IE and gas-phase basicity, see: (b) Cherkasov, A. R.; Jonsson,
M.; Galkin, V. J. Mol. Graphics Mod. 1999, 17, 28-42.
(14) (a) Wayner, D. D. M.; Clark, K. B.; Rauk, A.; Yu, D.; Armstrong,
D. A. J. Am. Chem. Soc. 1997, 119, 8925-8932. (b) Laleve´e, J.; Allonas,
X.; Fouassier, J.-P. J. Am. Chem. Soc. 2002, 124, 9613-9621. (c) Feng,
Y.; Wang, J.-T.; Huang, H.; Guo, Q.-X. J. Chem. Comput. Sci. 2003, 43,
2005-2013.
(15) (a) Burkey, T. J.; Catelhano, A. L.; Griller, D.; Lossing, F. P. J.
Am. Chem. Soc. 1983, 105, 4701-4703. (b) Griller, D.; Lossing, F. P. J.
Am. Chem. Soc. 1981, 103, 1586-1587.
(16) For the influence of amine class on the catalysis by aliphatic thiols
of the photoreduction of triplets carbonyls, see: (a) Stone, P. G.; Cohen,
S. J. Am. Chem. Soc. 1980, 102, 5686-5688. (b) Stone, P. G.; Cohen, S.
J. Am. Chem. Soc. 1982, 104, 3435-3440. (c) Khan, J.; Cohen, S. G. J.
Org. Chem. 1991, 56, 938-943.
366 J. Org. Chem., Vol. 73, No. 2, 2008

Thiyl Radical-Mediated Racemization of Aliphatic Amines
Other parameters might be taken into account such as amine
basicity17,18 or amine ionization potential.13,19 Since all properties
are intimately correlated, it is difficult to define which parameter
best accounts for the results.
The rate of formation of disulfide in the reaction medium
seems to be the most critical factor. The higher the amount of
disulfide is, the lower the amount of free thiol, i.e., the lower
the amount of hydrogen atom donor in the medium is. Even
though the homolytic cleavage of either the thiol or the disulfide
may initiate the production of thiyl radical, the reaction cannot
proceed unless the concentration of hydrogen donor enables
hydrogen atom transfer to the intermediate R-aminoalkyl radical
to become more efficient than competitive side reactions.
As the reaction proceeds, an increasing amount of disulfide
is formed at the expense of the thiol. Even spontaneous
formation of a small amount of disulfide was observed upon
mixing the reagents together at t0 just before starting irradiation.
The formation of disulfide can occur through different path-
ways.20 Equations 1-8 illustrate a panel of elementary steps
the occurrence of which may be envisaged in the reaction
medium. Thiyl radical and disulfide radical anions are closely
related to each other through an equilibrium which comes into
play any time thiyl radicals are generated through a redox
process from either thiols or disulfides (eq 3). Due to the
involvement of thiolate anion the equilibrium depends on the
concentration of the anion and thus, in our case, on the pKa of
the amine and on its concentration. It depends also on the nature
of the R group and, thus, on the acidity of the thiol. Reductive
generation of thiyl radical through direct electron transfer to
disulfide can be initiated with one electron reducing species
such as R-aminoalkyl radicals (eq 7). Although less likely, direct
reduction of thiyl radical to thiolate anion can also be envisaged
with these alkyl radicals known as being powerful reducing
agents13,15,21 (eq 8).
FIGURE 4. Plots for the disappearance of methyl thioglycolate and
for the formation of the corresponding disulfide upon irradiation in
the presence of amines at 30 °C.
anism in eqs 5 and 6. However, the ionization potential of the
amine does not correlate with the rate of formation of the
disulfide. The formation of disulfide is slower in the presence
of primary amine 1 than in the presence of a tertiary one (N-
methyl-N,N-di-n-octylamine), but the formation of the disulfide
and the concomitant disappearance of the thiol go slightly slower
in the presence of a secondary amine (i-Pr2NH taken as model)
than with amine 1.
(17) pKa values in water are 10.69 for n-propylamine, 11.09 for
diethylamine, and 11.09 for triethylamine; see: (a) Lu, G.; Grossman, J.
E.; Lambert, J. B. J. Org. Chem. 2006, 71, 1769-1776. The corresponding
pKa values in acetonitrile are 18.4 for ethylamine, 18.8 for diethylamine,
and 18.5 for triethylamine; see: (b) Li, J.-N.; Fu, Y.; Liu, L. Guo, Q.-X.
Tetrahedron 2006, 62, 11801-11813. Solvent effect on amine basicity is
well known. A subtle differentiation, according to the nature of the alkyl
substituent, might be induced by a solvent of low dielectric constant like
benzene. See: (c) Headley, A. D. J. Org. Chem. 1991, 56, 3688-3691. (d)
Safi, B.; Choho, K.; Geerlings, P. Chem. Phys. Lett. 1999, 300, 85-92. (e)
Caskey, D. C.; Damrauer, R.; McGoff, D. J. Org. Chem. 2002, 67, 5098-
5105.
(18) Proton transfer would induce the formation of an unreactive salt in
the reaction medium (the R C-H BDE is strengthened when the amine lone
pair is protonated). The formation of the ammonium thiolate is an
equilibrium. The more basic the amine is and the more acidic the thiol is,
the more the equilibrium should be displaced to the right and, as a
consequence, the least efficient the hydrogen abstraction should be. The
pKa value for methyl thioglycolate is 7.8, and the pKa value for n-BuSH is
10.9; see: McMillen, D. F.; Golden, D. M. Annu. ReV. Phys. Chem. 1982,
33, 493-532.
(19) For other photoinduced electron transfer processes involving amines,
see: (a) Pischel, U.; Nau, W. M. J. Am. Chem. Soc. 2001, 123, 9727-
9737. (b) Pishel, U.; Nau, W. M. J. Phys. Org. Chem. 2000, 13, 640-647.
(c) Dossot, M.; Allonas, X.; Jacques, P. Phys. Chem. Chem. Phys. 2002, 4,
2989-2993.
(20) For general reviews, see: (a) Asmus, K.-D.; Bonifacˇic´, M. In
S-Centered Radicals; Alfassi, Z. B., Ed.; Wiley: Chichester, 1999; Chapter
5. (b) Chatgilialoglu, C.; Guerra, M. Thiyl radicals. In Chemistry of Sulphur-
Containing Functional Groups; Patai, S.; Rappoport, Z., Eds.; Wiley:
Chichester, New York, 1993; pp 363-94. (c) For EPR studies of
photoreaction of aliphatic thiols leading to disulfide radical anions, see:
Cremonini, M. A.; Lunazzi, L. Placicci, G. J. Org. Chem. 1993, 58, 3805-
3810.
The evolution of the amounts of thiol and that of the corre-
sponding disulfide was investigated under different experimental
conditions monitoring the reaction by H NMR using penta-
1
methylbenzene as internal standard in C6D6.22 When methyl thio-
glycolate was irradiated alone for 3 h, less than 10% of thiol
was consumed to give the disulfide. When the thiol and MeN-
(n-Oct)2 (taken as a model for tertiary amine) were mixed in
the absence of irradiation, less than 6% of thiol was consumed
and converted into disulfide within 3 h. Figure 4 illustrates how
the irradiation of the thiol in the presence of the different types
of amines stimulates the formation of disulfide.
(21) (a) Wayner, D. D. M.; Dannenberg, J. J.; Griller, D. Chem. Phys.
Lett. 1986, 131, 189-191. (b) For the theoretical prediction of the absolute
standard redox potentials of a series of R-amino radicals, see ref 13. The
values calculated in solution in acetonitrile, at the B3LYP/6-311++G-
(2df,2p)//B3LYP/6-31G(d) level of theory using the PCM solvation model,
There is a clear relationship between the nature of the amine
and the rate of formation of the disulfide which argues in favor
of the hypothesis of a photostimulated electron-transfer mech-
•
range from - 1.5 V to 0.36 V (E° ) -0. 66 V for H2NCH2 ; - 1.47 V for
•
(Me)2NCH2 ). (c) Armstrong, D. A.; Rauk, A.; Yu, D. J. Am. Chem. Soc.
1993, 115, 666-673.
J. Org. Chem, Vol. 73, No. 2, 2008 367

Routaboul et al.
TABLE 1. Quantitative Data for the Racemization of Amines
processes. The lowest isolated yield was registered for amine 8
(bearing an additional tertiary amine group, entry h). It can be
noted that the racemization of amphetamine 9 (that could not
be performed at 80 °C without degradation) was also very
efficient (entry i). The bifunctional amines 11, 13, and 14 were
degraded when irradiation was carried out in quartz vessel.
However, 11 and 13 could be racemized without too much
degradation in a Pyrex vessel at 47 °C and 30 °C, respectively
(entries l′ and n).
1-14 upon Irradiation in the Presence of Methyl Thioglycolatea
yield (%)
amine
entry
(ee0, %)
time (h)
NMRb
isolatedc
ee (%)
a
1 (94)
2.5
3.5
3
3
3
3.5
3
3.5
1
3
1
2.5
3
3
4
3
99
92
99
96
96
83
97
86
99
nd
96
97
94
83
63
95
79
82
62
81
76
84
85
63d
77
62
13
8
<0.5
33
4
a′
b
c
d
d′
e
e′
f
g
h
i
k
l
l′
m
n
o
1 (94)e
2 (94)
3 (94)
4 (99)
4 (99)e
5 (98)
5
6
7
The racemization of tertiary amines was slow and could not
be completed (final ee close to 30% after irradiating for 3h)
(entries c and g).
5 (98)e
6 (98)
15
30
21
10
4
7 (98)
8 (>99.5)
9 (93)
10 (90)
11 (93)
11 (93)e
12 (91)
13 (>99)f
14 (>99.5)
Conclusion
The thiyl radical mediated racemization of aliphatic amines
could be performed under mild experimental conditions at 30
°C under photochemical initiation. The racemization of primary
and secondary amines proceeded in high yield, and all experi-
ments reported herein were very clean and fast. Tertiary amine
racemization was slower. Thioglycolic derivatives are among
the most efficient aliphatic thiols. The reaction proceeds through
the reversible formation of a R-aminoalkyl radical. Therefore,
the concentration of thiol as the hydrogen atom donor must be
kept at such a level as to avoid competitive evolutions of the
prochiral intermediate radical. Any oxidative transformation of
the thiol into the corresponding disulfide is detrimental to the
racemization process. Therefore, the efficiency of the racem-
ization depends on the complex interplay of several factors such
as R C-H BDE, amine ionization potential, amine basicity, thiol
acidity. The low-temperature range opens perspectives regarding
the performance of dynamic kinetic resolutions associating the
photochemically induced reaction with enzymatic resolution by
thermolabile enzymes. Assays are currently under investigation,
results will be reported in due course.
degradation
87
97
88
53
84
74
17
2
16
3.5
degradation
a Conditions: amine (0.067M); thiol (1.2 equiv); solvent: benzene; 30
°C; hν (quartz vessel, unless otherwise stated). b Determined using pen-
tamethylbenzene as internal standard. c Yields were determined after
derivatization in trifluoroacetamide unless otherwise stated. d Isolated
without derivatization. e The reaction was carried out in a Pyrex vessel at
47 °C. f The reaction was carried out in a Pyrex vessel at 30 °C.
It must be noted that in the case of amine 3, the disappearance
of the thiol is by far much slower than in the other cases. The
presence of the two carboxylic esters in position â-relative to
the nitrogen atom in 3 modifies the properties in such a way
that the formation of disulfide has little incidence after irradiating
for 3 h (less than 5%). In this case, the incidence of the
substituents on either the R-CH BDE or the amine ionization
potential could provide a rationale.
The racemization efficiency is undoubtly connected with the
disappearance of the thiol from the reaction medium. Partial
racemization of amine 1 could be achieved when starting from
the thioglycolic disulfide, i.e., in the absence of thiol, but the
ee was only 36% after 7 h, and partial degradation occurred
Experimental Section
General Procedure for Racemization Experiments. In a typical
experiment, a 0.06 M benzene solution of amine (100 mg) and
methyl thioglycolate (1.2 equiv) was irradiated at 30 °C in a quartz
tube (or Pyrex tube depending on the stability of the substrate (see
Table 1)) in a Rayonet apparatus (RPR-200, 16 UV lamps RPR-
3000) for 1-4 h. The workup was adapted according to the nature
of the amines. For primary amines, after concentration, the residue
was diluted with CH2Cl2 (5 mL). Triethylamine (3.2 equiv) and
trifluoroacetic anhydride (3 equiv) were added at 0 °C. The solution
was stirred for 3 h at room temperature. The solvent was removed
in vacuo and the crude material was diluted with Et2O, washed
with HCl (0.1 N) and brine, and dried under Na2SO4. After
concentration, the crude material was subjected to flash column
chromatography on silica gel (AcOEt/pentane) to give the trifluo-
roacetamide. For secondary and tertiary amines, after concentration,
the crude material was purified by flash column chromatography
on silica gel (Et3N/pentane) to give the pure amine.
1
(47% H NMR yield). Conversely, starting with 1.2 equiv of
methyl thioglycolate, 1 was racemized in 2.5 h (13% ee, Table
1, entry a).
Finally, the influence of the temperature upon the racemiza-
tion of amine 1 was investigated. Lowering the temperature to
30 °C had no significant effect on the characteristic feature of
the reaction, as far as it was carried out in quartz vessel. The
optimized conditions at 30 °C were applied to the series of
aliphatic amines given in Figure 1. The results (final ee, and
isolated yields) are reported in Table 1. For sake of comparison,
some satisfying results obtained at 47 or 30 °C in a Pyrex vessel
are also given (entries a′, d′, e′, l′, and n).
The racemization of primary and secondary amines proceeded
in high yield, and all reactions were very clean and fast
(22) Only methyl thioglycolate enabled the monitoring of the reaction
by 1H NMR while irradiating in C6D6 in the presence of pentamethylbenzene
as internal standard. The signals of the methylene protons and of the
methoxy group in both compounds are distinct from each other. The δ values
for the methoxy group and the methylene protons are 3.20 and 2.66,
respectively, in methyl thioglycolate. The δ values for the methoxy group
and the methylene protons are 3.25 and 3.17, respectively, in the corre-
sponding disulfide. It can be noted that the methylene signal in methyl
thioglycolate remains a doublet (J ) 8.1 Hz) in the presence of amine 3,
but due to a fast exchange of the SH proton, no coupling is observed in the
presence of amine 1 and the other amines in the series. This confirms the
low basicity of amine 3.
Acknowledgment. A gift of Novozym 435 from de
Novo Nordisk (Denmark) is gratefully acknowledged. We
thank the National Agency for Research for funding (ANR, no.
Blan06-3_142460).
Supporting Information Available: Experimental procedures
and NMR spectra for compounds 1-14. This material is available
free of charge via the Internet at http://pubs.acs.org.
JO702241Y
368 J. Org. Chem., Vol. 73, No. 2, 2008