The mechanism of the first step of the Mitsunobu reaction

Article abstract of DOI:10.1016/j.tet.2015.05.099

Previous DFT calculations employing phosphine (PH₃) and dimethyl azodicarboxylate showed that a cyclic O,N-phosphorane was energetically favored relative to betaine formation. In this study strong experimental support for the formation of the phosphorane is described. The question of whether betaine formation occurs via nucleophilic attack of the phosphine on the azodicarboxylate as has been previously assumed, or via a [4+2] cycloaddition (cheletropic) reaction to give the O,N-phosphorane, followed by ring-opening to give the betaine has not been resolved. An answer to this intriguing question is provided.

Full text of DOI:10.1016/j.tet.2015.05.099

Accepted Manuscript
The mechanism of the first step of the Mitsunobu reaction
David Camp, Mark von Itzstein, Ian D. Jenkins
PII:
S0040-4020(15)00808-X
10.1016/j.tet.2015.05.099
TET 26817
DOI:
Reference:
To appear in: Tetrahedron
Received Date: 31 March 2015
Revised Date: 13 May 2015
Accepted Date: 26 May 2015
Please cite this article as: Camp D, von Itzstein M, Jenkins ID, The mechanism of the first step of the
Mitsunobu reaction, Tetrahedron (2015), doi: 10.1016/j.tet.2015.05.099.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical Abstract
To create your abstract, type over the instructions in the template box below.
Fonts or abstract dimensions should not be changed or altered.
The mechanism of the first step of the
Mitsunobu reaction. David Camp, Mark von
Itzstein and Ian D. Jenkins
Leave this area blank for abstract info.
Eskitis Institute, Griffith University, Brisbane, QLD, 4111, Australia.

1
ACCEPTED MANUSCRIPT
Tetrahedron
journal homepage: www.elsevier.com
The mechanism of the first step of the Mitsunobu reaction
David Camp a, Mark von Itzstein b and Ian D. Jenkins c, *
a
Griffith School of Environment, Griffith University, Nathan Campus QLD 4111 Australia.
b Institute for Glycomics, Griffith University, Gold Coast Campus QLD 4222 Australia.
c Eskitis Institute for Drug Discovery, Griffith University, Nathan Campus QLD 4111 Australia.
ARTICLE INFO
ABSTRACT
Article history:
Received
Received in revised form
Accepted
Previous DFT calculations employing phosphine (PH3) and dimethyl azodicarboxylate showed
that a cyclic O,N-phosphorane was energetically favored relative to betaine formation. In this
study strong experimental support for the formation of the phosphorane is described. The
question of whether betaine formation occurs via nucleophilic attack of the phosphine on the
azodicarboxylate as has been previously assumed, or via a [4+2] cycloaddition (cheletropic)
reaction to give the O,N-phosphorane, followed by ring-opening to give the betaine has not been
resolved. An answer to this intriguing question is provided.
Available online
Keywords:
electrocyclic
betaine
2009 Elsevier Ltd. All rights reserved
.
Michael addition
O,N-phosphorane
Mitsunobu
reaction.4,8 All three mechanisms appear a priori to be equally
1. Introduction
likely.
The Mitsunobu reaction is one of the most useful reactions in
organic synthesis, particularly for the inversion of configuration
of secondary alcohols.1 The first step in the Mitsunobu reaction
involves reaction of triphenylphosphine (Ph3P) with diethyl
azodicarboxylate (DEAD) or diisopropyl azodicarboxylate
(DIAD) to form a betaine 1 as originally proposed by Morrison2
and substantiated by Brunn and Huisgen over 40 years ago.3
Three possible mechanisms have been proposed for the first
step of the Mitsunobu reaction: (1) Michael-type nucleophilic
attack by the phosphine on the azodicarboxylate (Scheme 1),2-4
Scheme 3. Cheletropic Mechanism
The second possible mechanism appeared the most likely for a
while following the discovery that radicals are formed in the
Mitsunobu reaction5,9 (see later). However, in work that is not
widely known, the SET mechanism has been excluded by
Eberson and co-workers using cyclic voltammetry. The SET
reaction between Ph3P and DEAD was found to be far too
endergonic to be a realistic step under any conditions,10 leaving
mechanisms (1) and (3) as contenders. As far as we are aware,
there has been no definitive proof to exclude an initial cheletropic
reaction followed by ring-opening as being the mechanism of
formation of the Morrison-Brunn-Huisgen betaine 1.
R3P
R3P
CO2R'
N N
N N
R'O2C
R'O2C
CO2R'
1
Scheme 1. Nucleophilic Mechanism
(2) a single electron transfer (SET) process (Scheme 2),5 and (3)
an initial [4+2] cycloaddition (cheletropic) reaction6 followed by
ring-opening (Scheme 3).
Recently, Anders et al11 have performed an extensive density
functional [BP86/6-311++G(3df,3pd) level] investigation of the
hypersurface of the Mitsunobu reaction and shown that the first
step can lead to either a five-membered oxadiazaphosphole ring
(O,N-phosphorane) 2 or the betaine 1. Interestingly, the O,N-
phosphorane (2, R = H, R' = Me) was 12 kcal/mol more stable
Scheme 2. SET Mechanism
Each of these mechanisms is consistent with the observed
second-order reaction kinetics,7 and the irreversibility of the
* Corresponding author. Tel +61 7 3735 6025, Fax: + (07) 3735 6001, E-mail address: i.jenkins@griffith.edu.au

2
Tetrahedron
ACCEPTED MANUSCRIPT
than the betaine 1 (R = H, R' = Me) in acetonitrile. This
The formation of an O,N-phosphorane 4 rather than a betaine
suggests that the betaine 1 could indeed be formed via an initial
[4+2] cycloaddition (cheletropic) reaction followed by ring-
opening (Scheme 3). There are precedents for tervalent
phosphorus compounds participating in cheletropic reactions.
Thus, triphenylphosphite12 and various cyclic phosphites13 react
with azodicarboxylates to afford O,N-phosphoranes while
halogenophosphines undergo [4+2] cycloaddition with cis-1,3-
dienes, α,β-unsaturated ketones and α-dicarbonyl compounds to
form five-membered ring phosphorus heterocycles.14
5 with 9-phenyl-9-phosphafluorene (Scheme 4) provides strong
experimental support for the results of the DFT calculations of
Anders et al.11
In order to distinguish between Michael-type addition
(Scheme 1) or a pericyclic reaction (Scheme 3), a competition
experiment involving reaction of DIAD with a mixture of Ph3P
and 3 was undertaken. The relative reactivity of cyclic versus
acyclic tervalent phosphorus compounds has been reviewed by
Hudson and Brown.19 Hence, when phosphorus acts as a
nucleophile, as in a Michael-type addition, a small deviation
from the C–P–C angle of Ph3P ( 100°) to near that of the
tetrahedral angle of 109° would occur. However, the C–P–C
angle of the phosphafluorene ring of 3 is only 89°,20 so that an
increase in ring strain would be expected when 3 reacts as a
nucleophile and it would be expected to be less reactive than
Ph3P. On the other hand, in a cheletropic mechanism, the cyclic
compound, 3, should display enhanced reactivity relative to Ph3P
as formation of a pentacoordinate phosphorane would result in a
decrease in ring strain.19 It is well known that the presence of a
five-membered ring stabilizes pentacoordinate phosphoranes.15-17
Only betaine 1 (R = Ph, R' = Et, i-Pr or t-Bu) has been
observed by 31P NMR under typical conditions and not the
corresponding O,N-phosphorane 2. Moreover, the chemical shift
of 1 (R = Ph, R' = Et, i-Pr or t-Bu) varies little with solvent
polarity (e.g. δP +43.9/benzene to +45.4/DMF for R' = i-Pr)
indicating that any equilibrium between an O,N-phosphorane 2
and betaine 1 lies very much towards the betaine.4
Anders et al11 point out that three phenyl ligands on
phosphorus (rather than three hydrogens used in the calculations)
can be expected to stabilize the betaine 1 relative to the O,N-
phosphorane 2, so we thought it would be of interest to
investigate the reaction of 9-phenyl-9-phosphafluorene (3) with
an azodicarboxylate. Phosphine 3 is a close analogue of
triphenylphosphine but it contains a five-membered ring that
should facilitate formation of a phosphorane.15-17
Treatment of a 1:1 mixture of 3 (0.33 mmol) and Ph3P (0.33
mmol) with a limiting amount of DIAD (0.33 mmol) under
conditions similar to those employed above, resulted in the
formation of two peaks at δP +43.8 and –33.6 ppm in a 2:1 ratio,
respectively. The downfield signal corresponds to 1 (R = Ph, R' =
i-Pr) while the upfield peak at –33.6 ppm is due to the
spirophosphorane
4
(R'
=
i-Pr). As expected, signals
corresponding to the unreacted phosphines, Ph3P (δP –5.2 ppm)
and 3 (δP –9.3 ppm), were present in the complementary ratio of
1:2, respectively. Betaine formation in this system was
irreversible. Thus, addition of 3 to 1 (R = Ph, R' = i-Pr) resulted
in no change to the 31P NMR spectrum. Similarly, addition of
Ph3P to 4 did not afford 1 (R = Ph, R' = i-Pr) after one month at 4
°C. The irreversibility of betaine formation is consistent with the
results of Crich et al8 and, coupled with the mixed phosphine
experiment, shows that Ph3P reacts faster with DIAD than does 3.
This result is consistent with Michael-type nucleophilic addition
and inconsistent with a pericyclic reaction. In the case of the
phosphine 3, the first step of the reaction involves nucleophilic
addition to the azodicarboxylate to give the betaine 5 which
subsequently undergoes ring-closure to give the more stable O,N-
phosphorane 4. It is interesting to note in this regard that Anders
et al11 did not succeed in cleanly locating transition structures for
the attack of PH3 on the azodicarboxylate in either the gas phase
or solution.
Scheme 4. Phosphorane formation with 9-phenyl-9-phosphafluorene
2. Results and Discussion
Treatment of 9-phenyl-9-phosphafluorene (3, 0.33 mmol) in
tetrahydrofuran (THF, 2 mL) at 0 °C with DIAD (0.33 mmol)
under nitrogen gave a yellow-orange solution, the 31P NMR
spectrum of which showed a major peak at δ –33.6 ppm,
consistent with the formation of the spirophosphorane 4 (R' = i-
Pr). There was also a minor peak at δ +30.1 ppm corresponding
to the oxide of 3 [Lit18 δ +29.1 (CDCl3)]. Evidence for a rapid
equilibrium between the phosphorane 4 (R' = i-Pr) and the
betaine 5 (R' = i-Pr) was obtained from the marked solvent effect
on the 31P NMR chemical shift (Table 1).
Analogous results were obtained when a 1:1 mixture of Ph3P
and 3 was treated with methyl iodide instead of DIAD, i.e. the
Ph3P reacted faster than the cyclic analogue 3. Two peaks were
observed by 31P NMR spectroscopy (δP +23.2 and +26.7 ppm,
THF) in a 4:1 ratio corresponding to the methiodide salts of Ph3P
and 3, respectively. This confirms the nucleophilic nature of both
the alkylation reaction and the Michael-type addition to the
azodicarboxylate.
Table 1. 31P NMR data for Phosphorane 4
Solvent
Benzene
THF
δP (ppm)
–34.1
–33.6
CH2Cl2
CH3CN
–24.5
In addition to the P–N betaine 1 formed initially in the
Mitsunobu reaction employing Ph3P and either DEAD or DIAD,
a minor reaction pathway results in the formation very small
amounts of the radical species 7 and 8.5,9,10 These radicals are
formed9,10 by oxidation of the P–O betaine 6 by excess DIAD or
DEAD (or if the Ph3P is added to the azodicarboxylate rather
–21.9

3
than the other way around). The P–O betaine 6 could be formed
ACCEPTED MANUSCRIPT
directly by Michael-type addition to the carbonyl oxygen
(Scheme 5) or by “normal” betaine formation (Scheme 1)
followed by ring-closure to give the O,N-phosphorane 2 (R = Ph,
R' = Et or i-Pr) and then ring-opening at the P–N bond rather than
the P–O bond. Both radical species 7 and 8 are decomposed
immediately by carboxylic acids and do not appear to be
involved in normal Mitsunobu reactions.
References and notes
1. For a comprehensive review, see Swamy, K. C. K.; Kumar, N. N. B.;
Balaraman, E.; Kumar, K. V. P., Chem. Rev. 2009, 109, 2551-2651.
2. Morrison, D. C. J. Org. Chem. 1958, 23, 1072-1074.
3. Brunn, E.; Huisgen, R. Angew. Chem., Int. Ed. 1969, 8, 513-515.
4. Guthrie, R. D. G.; Jenkins, I. D. Aust. J. Chem. 1982, 35, 767-774.
5. Camp, D.; Hanson, G. R.; Jenkins, I. D. J. Org. Chem. 1995, 60, 2977-
2980.
6. Kumar, N. S.; Komana, P.; Vittal, J. J.; Swamy, K. C. K. J. Org.
Chem. 2002, 67, 6653-6658.
7. Kanzanian, T.; Mayr, H. Chem. Eur. J. 2010, 16, 11670-11677.
8. Crich, D.; Dyker, H.; Harris, R. J. J. Org. Chem. 1989, 54, 257-259.
9. Camp, D.; Campitelli, M.; Hanson, G. R.; Jenkins, I. D. J. Am. Chem.
Soc. 2012, 134, 16188-16196.
10. Eberson, L.; Persson, O.; Svensson, J. O. Acta. Chem. Scan. 1998, 52,
1293-1300.
11. Schenk, S.; Weston, J.; Anders, E. J. Am. Chem. Soc. 2005, 127,
12566-12576.
12. Arbuzov, B. A.; Polezhaeva, N. A.; Vinogradova, V. S. Izv. Akad.
Nauk SSSR, Ser. Khim. 1968, 2525-2529.
13. For examples, see: (a) Gonclaves, H.; Domroy, J. R.; Chapleur, Y.;
Castro, B.; Faudet, H.; Burgada, R. Phosphorus Sulfur 1980, 8, 147-
152. (b) Majoral, J. P.; Kraemer, R.; N'Gando M'Pondo, T.; Navech, J.
Tetrahedron Lett. 1980, 21, 1307-1310. (c) Navech, J.; Kraemer, R.;
Majoral, J. P. Tetrahedron Lett. 1980, 21, 1449-1452. (d) Hulst, R.;
van Basten, A.; Fitzpatrick, K.; Kellogg, R. M. J. Chem. Soc., Perkin
Trans. 1 1995, 2961-2963. (e) Li, Z.; Zhou, Z.; Wang, L.; Zhou, Q.;
Tang, C. Tetrahedron: Asymmetry 2002, 13, 145-148. (f) Kumar, K. V.
P. P.; Kumar, N. S.; Swamy, K. C. K. New J. Chem. 2006, 30, 717-
728.
Scheme 5. Formation of radicals in the Mitsunobu reaction
3. Conclusions
14. Smith, D. J. H. In Comprehensive Organic Chemistry; Barton, D. H.
R., Ollis, V. D., Eds.; Pergamon: Exeter, 1979; Vol. 2, pp 1233-1256.
15. Emsley, J. and Hall, D. “The Chemistry of Phosphorus”, Harper and
Row, London, 1976.
16. Ramirez, F. Bull. Soc. Chim. Fr. 1970, 3491-3519.
17. (a) Gorenstein, D. G.; Luxon, B. A.; Findlay, J. B.; Momii, R. J. Am.
Chem. Soc. 1977, 99, 4170-4172. (b) Swamy, K. C. K.; Pavan, M. P.;
Srinivas, V. Top. Heterocycl. Chem. 2009, 20, 99-145.
18. van Kalkeren, H. A.; Leenders, S. H. A. M.; Hommersom, C. R. A.;
Rutjes, F. P. J. T.; van Delft, F. L. Chem. Eur. J. 2011, 17, 11290 –
11295.
Evidence is provided that the first step in the Mitsunobu reaction,
the reaction of a phosphine with DIAD or DEAD, can produce
either an O,N-phosphorane or a betaine, depending on the
phosphine employed. This represents the first experimental
confirmation of previous DFT predictions. Evidence is also
provided for a rapid equilibrium between the O,N-phosphorane
and the corresponding betaine. Significantly, we now know that
the first step in the Mitsunobu reaction proceeds via a Michael-
type nucleophilic attack by the phosphine on the
azodicarboxylate, and not via a concerted pericyclic reaction or a
SET mechanism.
19. Hudson, R. F.; Brown, C. Acc. Chem. Res. 1972, 5, 204-211.
20. Alyea, E. C; Feruson, G.; Gallagher, J. F. Acta Cryst. (1992). C48,
959-961.
21. Nesmeyanov, N. A.; Rebrova, O. A.; Mikul'shina, V. V.; Petrovsky, P.
V.; Robas, V. I.; Reutov, O. A. J. Organometal. Chem. 1976, 110, 49-
57.
4. Experimental Section
Materials. Triphenylphosphine, diisopropyl azodicarboxylate,
and methyl iodide were commercial samples and used without
further purification. THF was dried and distilled prior to use. 9-
Phenyl-9-phosphafluorene (3) was synthesized according to the
procedure of Nesmeyanov et al.21
31P NMR competition experiments. Ph3P (94.4 mg, 0.36 mmol)
and 9-phenyl-9-phosphafluorene (93.6 mg, 0.36 mmol) were
dissolved in dry THF (3 mL) under nitrogen in a 10 mm NMR
tube. The solution was cooled to 0 °C and DIAD (18 ꢀL, 0.09
mmol) then added dropwise over 1 min to the swirled, cooled
solution. A cap was placed on the tube and sealed with Parafilm
before recording the 31P NMR spectrum at 10 °C on a 300 MHz
instrument within 5 min. The experiment was repeated using
methyl iodide in place of DIAD. All spectra were acquired at an
operating frequency of 121.47 MHz using a 45° flip angle, 3 s
recycle delay, and a 0.33 s acquisition time with gated
decoupling. Negative 31P chemical shifts are upfield of external
phosphoric acid (85%).
Acknowledgement.
We thank Griffith University for financial support.
* Corresponding author. Tel +61 7 3735 6025, Fax: + (07) 3735 6001, E-mail address: i.jenkins@griffith.edu.au