Incremental solvation precedes ion-pair separation in enantiomerization of a cyano-stabilized Grignard reagent

Article abstract of DOI:10.1002/chem.201102253

Concerted or stepwise? Ion-pair separation is often proposed as a mechanism for enantiomerization of organolithium reagents in ethereal solvents. But what is the timing of the solvation and bond-cleavage events? Are they concerted, or does one precede the

Full text of DOI:10.1002/chem.201102253

DOI: 10.1002/chem.201102253
Incremental Solvation Precedes Ion-Pair Separation in Enantiomerization of
a Cyano-Stabilized Grignard Reagent
Neeraj N. Patwardhan, Ming Gao, and Paul R. Carlier*[a]
Dedicated to Professor K. Barry Sharpless on the occasion of his 70th birthday
Configurationally stable enantioenriched organolithium
compounds are important in asymmetric synthesis.[1] Studies
of the enantiomerization pathways of these compounds are
useful for improving the stereochemical outcome of these
reactions, and provide insight into the structure and dynam-
ics of these compounds.[1a,f,2] The composition of the organic
fragment,[1a,f,2a,d] the presence of additives,[2d,3] and the iden-
tity of the solvent[2d,4] are all known to affect configurational
stability. In contrast, few such studies have been reported
for enantioenriched organomagnesium compounds, perhaps
reflecting the relative paucity of these reagents.[5] Herein we
report Eyring analyses and reaction order studies of the rac-
emization of enantioenriched Grignard reagent 2 in ethereal
Scheme 1. Mg/Br exchange/deuteration reactions of (S)-(+)-1; structures
of 2 are not meant to imply a particular aggregation or solvation state.
solvents. Enantiomerization rates increase in the order Et2O
< 2-methylTHF < THF. A highly negative activation entro-
py for enantiomerization ((ꢀ49ꢁ4) eu) is seen under condi-
tions where reaction is zero-order in [Et2O] and first-order
in [Mg]. These results are consistent with an ion-pair separa-
tion enantiomerization mechanism wherein association of
tures. A key feature of our reinvestigation was the use of a
precooled CH3OD quench in place of room-temperature
D2O.[7] Delay times t varied from 60 to 3600 s and >95% re-
covery of [D1]-3 (>98% deuterium incorporation) was ob-
tained after workup even at the shortest time points (60 s)
at 175 K. First-order rate constants for enantiomerization of
2 (kenant) derive from slopes (ꢀ2kenant) of the plots of ln-
ꢀ
solvent precedes rupture of the Mg C bond. To our knowl-
edge this is the first experimental determination of the
timing of incremental solvation and ion-pair separation for a
Group I or II organometallic reagent.
G
We previously reported[5h] that enantioenriched cyano-sta-
bilized Grignard reagent 2 could be prepared by Mg/Br ex-
change[6] of enantiopure a-bromonitrile (S)-1 with iPrMgCl
(Scheme 1). The overall transformation of (S)-1 to [D1]-3
was shown to favor retention, and measurements of the e.r.
of [D1]-3 with increasing delay time t demonstrated that 2
possessed a racemization half life (t1/2ACTHNUGTRNE(NUG rac)) of 11.4 h in Et2O
at 173 K. The extrapolated e.r. at t=0 for 2 from these ex-
periments was 91:9, suggesting that Mg/Br exchange was not
perfectly retentive.
To explore the role of solvent in the enantiomerization
process, we reinvestigated this reaction in Et2O, 2-methylte-
trahydrofuran (2-MeTHF), and THF, at several tempera-
[a] N. N. Patwardhan, Dr. M. Gao, Prof. P. R. Carlier
Department of Chemistry, Virginia Tech
Blacksburg, VA 24061 (USA)
Figure 1. Determination of enantiomerization rate constants kenant and ex-
trapolated t=0 e.r. for 2 at 195 K and [Mg]total =0.0342m in Et2O, 2-
MeTHF, THF. CR is the mole fraction of the R enantiomer; note that
Fax : (+1)540-231-3255
E-mail: pcarlier@vt.edu
e.r.=100*XR : 100*
cept of each plot is ln
can be determined.
(1-XR). The slope of each plot is ꢀ2*kenant, and inter-
ꢀ
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201102253.
ACHUTGTNERN(NUG 2XRACHTUNGTERN(NUGN t=0) 1), from which the extrapolated t=0 e.r.
12250
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12250 – 12253

COMMUNICATION
As can be seen, at 195 K enantiomerization in Et2O is
seven- to eightfold slower than in 2-MeTHF and THF; the
same ordering is seen at the other temperatures as well. Be-
cause Mg/Br exchange is complete within 1 min even at
175 K, the y intercepts of these plots reflect the approximate
e.r. values of 2 following Mg/Br exchange. Whereas Mg/Br
exchange in Et2O at 195 K is highly retentive (extrapolated
t=0 e.r. is 94:6), reactions are significantly less selective in
2-MeTHF and THF (extrapolated t=0 er values of 79:21
and 59:41, respectively).
These divergent results suggest significant involvement of
the solvent in the both the Mg/Br exchange and enantiomer-
ization. To determine the activation parameters for enantio-
merization in these solvents, Eyring[9] analyses of the enan-
tiomerization rate constants were performed, using the
linear form DG° =DH° ꢀTDS° (Figure 2).[10]
appealing, merit additional scrutiny. One approach that
could discriminate between them, and eliminate other possi-
bilities, would be to measure reaction orders.[13]
For these studies we restricted ourselves to the use of
Et2O as solvent. In addition to providing the most retentive
Mg/Br exchange and smallest enantiomerization rate con-
stants, the structure of the Grignard reagent in solution is
considerably simpler in Et2O than in THF. Unlike the wide
array of structures proposed in THF,[14] in Et2O, alkyl mag-
nesium chlorides are known by ebullioscopy to be dimer-
ic.[14b,15] X-ray crystallography of Et2O solvates of RMgCl
show a preference for bisACHTNUTGRNEUNG
(m2-Cl) dimers,[16] suggesting this
structural motif in solution. Furthermore, our studies of the
stoichiometry of Mg/Br exchange of 1 with iPrMgCl in Et2O
are consistent with the formation of a mixed aggregate of 2
and iPrMgCl. Reaction of 1 with 1.1 equiv iPrMgCl in Et2O
(titrated, 10 min, 212 K, CH3OD quench), gave only (59ꢁ
5)% conversion of 1 (1H NMR), and (65ꢁ4)% of the ex-
pected amount of iPrBr (GC/MS, internal standard). In-
creasing the amount of iPrMgCl to 2.2 equiv gave >98%
conversion of 1, and (92ꢁ4)% of the expected amount of
iPrBr. We account for these observations by proposing that
iPrMgCl, which is known to be a dimer in Et2O,[15b] reacts
quickly with one equivalent of 1, to form a 2·iPrMgCl hete-
rodimer and iPrBr. Apparently reaction of the heterodimer
with another equivalent of 1 is slow under these conditions,
as is disproportionation of the mixed dimer to regenerate
the reactive iPrMgCl homodimer. In contrast, otherwise
identical studies in THF demonstrate that 1.1 equiv of
iPrMgCl (titrated) provides >95% consumption of 1, and
generation of (90ꢁ4)% of the expected amount of iPrBr,
consistent with the expected monomeric state of iPrMgCl in
THF.[15b]
Figure 2. Plot of calculated DG° values versus temperature T and activa-
tion parameters for the enantiomerization of 2 in Et2O. DG° values were
calculated from kenant values (175 K, 195 K, 212 K, 231 K) using the
Eyring equation. In most cases the error bars are smaller than the
symbol. Note that measurement of kenant at 231 K was practical only in
Et2O; in 2-MeTHF and THF at this temperature, Mg/Br exchange reac-
tions gave nearly racemic products.
To address the possibility that the large negative DS° ob-
served in Et2O was due to aggregation of 2·iPrMgCl on the
enantiomerization path,[17] kenant was measured at 195 K over
a 25-fold range in [Mg]total, from 0.063 to 0.0025m (see the
Supporting Information). If Grignard reagent aggregation
necessarily precedes enantiomerization, the measured first-
In all three solvents, large negative entropies of activation
DS° were calculated (Et2O: (ꢀ49ꢁ4) eu; 2-MeTHF: (ꢀ64ꢁ
3) eu; THF: (ꢀ59ꢁ2) eu), accompanied by relatively small
enthalpies of activation DH° (Et2O: (5.7ꢁ0.8) kcalmolꢀ1; 2-
MeTHF: (2.3ꢁ0.6) kcalmolꢀ1; THF: (2.9ꢁ0.4) kcalmolꢀ1).
Capriati et al. recently reported[1f] a similarly high negative
entropy of activation ((ꢀ44ꢁ1) eu) for the enantiomeri-
order rate constants kenant would increase with increasing
[18]
[Mg]total
.
Yet, within experimental error, no change in
kenant was seen over this range, thus firmly establishing first-
order kinetics,[19] and ruling out required aggregation (or
deaggregation) at this temperature on the enantiomerization
path.
To directly test the hypothesis that the solvation number
of 2 increases in the rate-determining transition state, the re-
action order in [Et2O] was studied by performing the Mg/Br
exchange in Et2O/toluene mixtures at a fixed Grignard con-
centration of 0.00625m. The concentration of Et2O was
varied over a 158-fold range, from 0.056m (9 equiv Et2O) to
9.5m (neat Et2O, 1520 equiv) at two temperatures (212 K
and 195 K, Table 1).
zation of 1-lithio-1-(3-trifluoromethylphenyl)oxirane.
A
highly negative entropy of activation (ꢀ33 eu) for enantio-
merization was also reported for the related structure 1-
lithio-1-arylcyclopropane,[11] and smaller negative values
have been noted by other workers for enantiomerization of
other organolithium compounds.[2a,12] Such results are gener-
ally interpreted as evidence for enantiomerization by an
ion-pair separation mechanism, rather than by a “conducted
tour” mechanism.[2a] Two distinct phenomenon have been
proposed as the origin of the negative entropy of activation:
either an increase in solvation number concerted with ion
pair separation (solvent capture), or rearrangement of the
At 195 K, kenant remained unchanged until [Et2O] fell to
0.06m; note that a small but steady increase in t=0 e.r. to
98:2 is also seen as [Et2O] decreases.[20] However, at 212 K,
from the lowest Et2O concentration (0.06m) to pure Et2O
ꢀ
secondary solvent shell during Li C cleavage of a fully sol-
vated contact ion pair, to stabilize developing charge separa-
tion (solvent electrostriction).[12] These hypotheses, though
Chem. Eur. J. 2011, 17, 12250 – 12253
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12251

P. R. Carlier et al.
Table 1. Dependence of kenant and extrapolated t=0 e.r. of 2 on [Et2O].[a]
Entry
Temp
[K]
[Et2O]
kenant ꢁ 106 [sꢀ1
]
e.r. (t=0)[c]
[m][b]
1
2
3
4
5
6
7
8
9
195
195
195
195
195
212
212
212
212
212
9.5
25.2ꢁ5.4
24.4ꢁ5.5
22.6ꢁ5.8
21.6ꢁ2.5
12.9ꢁ3.1
179.4ꢁ14.7
145.3ꢁ37.9
114.5ꢁ31.8
70.8ꢁ17.8
12.4ꢁ1.6
94:6
97:3
97:3
98:2
98:2
91:9
95:5
95:5
96:4
96:4
0.95
0.23
0.15
0.056
9.5
0.95
0.23
0.15
0.056
10
[a] Reactions performed at [Mg]total =0.00625m. [b] Concentration of
Et2O in toluene, assuming neat Et2O is 9.5m at both 195 and 212 K.
[c] Extrapolated.
Scheme 2. Proposed mechanism for the enantiomerization of the
2·iPrMgCl heterodimer in Et2O, and possible intermediates 4–7.
solvation number at the resting state in pure Et2O. There-
fore the large negative entropy of activation for enantiomer-
ization ((ꢀ49ꢁ4) eu) measured in pure Et2O cannot be at-
tributed to solvent capture in the transition structure, and
must be due to solvent electrostriction. Reordering of the
secondary solvent shell to stabilize the increasing dipole
moment during formation of a separated ion-pair such as 6
should carry a significant entropic cost. Note that the depict-
ed pyramidalization of the cyclopropyl anion in 6 is support-
ed by calculations of the free anion;[25] reduced angle strain
is no doubt responsible for this preference.[26] In addition we
would like to point out that a “conducted tour” pathway[25]
to directly generate an ion-pair like 7 would also be consis-
tent with observed DS° value and cannot be ruled out.
Carbanion inversion of 7 and N ! C migration would ulti-
mately lead to enantiomerization of 2. Such an “ionogenic
conducted tour” is thus a possible alternative to direct for-
mation of 6 by a traditional ion-pair separation mechanism.
In closing, we note that highly negative entropies of acti-
vation have been seen in several cases for dissociative reac-
tions involving carbocationic intermediates.[27] In these ex-
amples, rate-limiting coordination of solvent is not likely,
and solvent electrostriction provides a reasonable explana-
tion for the observed large negative DS° values (up to
(ꢀ33ꢁ2) eu[27a]). It thus seems likely that solvent electro-
striction plays a significant role in the kinetics and thermo-
dynamics of many ionogenic reactions, including the enan-
tiomerization of stabilized organolithium compounds in
ethereal solvents.[1f,11,28]
Figure 3. Plot of kenant versus [Et2O] in toluene co-solvent at 212 K at
[Mg]total =0.006m. See Scheme 2 for the definition of k2, kꢀ1, and k1, and
the Supporting Information for derivation of the equation relating them
and [Et2O] to kenant
.
(9.5m), kenant increases 14.5-fold. As Figure 3 illustrates, satu-
rating, not simple first-order dependence of kenant on [Et2O]
is seen.
Saturation behavior in [donor solvent] has been reported
for the deprotonation of hydrazones by LDA in TMEDA/
hexanes,[21] and deprotonation of ketones by LiHMDS in
R3N/toluene.[22] The rate constant data in Table 1 and the
saturation behavior depicted in Figure 2 are consistent with
a increase in the resting solvation number of 2·iPrMgCl
from [Et2O]=0.06m (9 equiv Et2O) to neat Et2O (9.5m,
1520 equiv) (Scheme 2).
Based on the available X-ray crystal structures of Et2O–
ACTHNUTRGNEUNG
solvates of RMgCl[16] and other bis(m2-X) dimers,[23] at low
[Et2O], 2·iPrMgCl likely exists as a disolvate (e.g. 4). Al-
though fits of kenant to [Et2O] (see the Supporting Informa-
tion) do not allow us to discriminate between addition of
one or two solvent molecules, for simplicity we assume for-
mation of a trisolvate in pure Et2O. One possible structure
for the trisolvate would be the open dimer 5; such a struc-
ture has been located computationally for the Et2O trisol-
vate of the MeMgCl dimer.[24] However the zero-order de-
pendence of kenant on [Et2O] at high [Et2O] does allow us to
rule out an increased solvation number in the rate-determin-
ing transition structure for enantiomerization, relative to the
12252
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12250 – 12253

Enantiomerization of a Cyano-Stabilized Grignard Reagent
COMMUNICATION
[8] a) D. C. Hsu, P. C.-H. Lam, C. Slebodnick, P. R. Carlier, J. Am.
Chem. Soc. 2009, 131, 18168–18176; b) P. R. Carlier, D. C. Hsu,
S. A. Bryson, Top. Stereochem. 2010, 26, 53–92.
[9] H. Eyring, Chem. Rev. 1935, 17, 65–77.
[10] No significant change in calculated activation parameters or their
standard errors was seen when the more traditional plotting method
Acknowledgements
We gratefully acknowledge financial support from the NSF (CHE-
075006), and thank Professors David B. Collum and James M. Tanko for
helpful discussions.
( lnACHTUNGTRENU(NG kenant/T) vs. 1/T) was applied.
[11] D. Hoell, C. Schnieders, K. Mꢃllen, Angew. Chem. 1983, 95, 240–
241; Angew. Chem. Int. Ed. Engl 1983, 22, 243–244; Angew. Chem.
Int. Ed. Engl 1983, 22, 299–306.
[12] H. Ahlbrecht, J. Harbach, R. W. Hoffmann, T. Ruhland, Liebigs
Ann. 1995, 211–216.
Keywords: electrostriction · enantiomerization · Grignard
reagents · ion-pair separation · reaction mechanisms ·
solvents
[13] D. B. Collum, A. J. McNeil, A. Ramirez, Angew. Chem. 2007, 119,
3060–3077; Angew. Chem. Int. Ed. 2007, 46, 3002–3017.
[1] a) A. Basu, S. Thayumanavan, Angew. Chem. 2002, 114, 740–763;
Angew. Chem. Int. Ed. 2002, 41, 716–738; b) J. Clayden, Tetrahe-
dron Organic Chemistry, Vol. 23, Organolithiums: Selectivity for Syn-
thesis, Pergamon, Amsterdam, 2002, pp. 169–240; c) V. Capriati, S.
Florio, A. Salomone, Top. Stereochem. 2010, 26, 135–164; d) R.
Scholz, G. Hellmann, S. Rohs, D. ꢂzdemir, G. Raabe, C. Vermeeren,
H.-J. Gais, Eur. J. Org. Chem. 2010, 4588–4616; e) S. Florio, R.
Luizi, Chem. Rev. 2010, 110, 5128–5157; f) V. Capriati, S. Florio,
F. M. Perna, A. Salomone, Chem-Eur J 2010, 16, 9778–9788.
[2] a) R. E. Gawley, Top. Stereochem. 2010, 26, 93–133; b) R. W. Hoff-
mann, Top. Stereochem. 2010, 26, 165–188; c) R. Scholz, G. Hell-
mann, S. Rohs, G. Raabe, J. Runsink, D. ꢂzdemir, O. Luche, T.
Heß, A. W. Giesen, J. Atodiresei, H. J. Lindner, H.-J. Gais, Eur. J.
Org. Chem. 2010, 4559–4587; d) F. M. Perna, A. Salomone, M.
Dammacco, S. Florio, V. Capriati, Chem. Eur. J. 2011, 17, 8216–
8225.
[3] a) T. K. Beng, T. I. Yousaf, I. Coldham, R. E. Gawley, J. Am. Chem.
Soc. 2009, 131, 6908–6909; b) I. Coldham, D. Leonori, T. K. Beng,
R. E. Gawley, Chem. Commun. 2009, 5239–5241.
[4] a) F. Marr, R. Frçhlich, D. Hoppe, Org. Lett. 1999, 1, 2081–2083;
b) H. Ikemoto, M. Sasaki, K. Takeda, Eur. J. Org. Chem. 2010,
6643–6650; c) D. C. Kapeller, F. Hammerschmidt, Chem. Eur. J.
2009, 15, 5729–5739; d) S. Wu, S. Lee, P. Beak, J. Am. Chem. Soc.
1996, 118, 715–721.
[5] a) H. M. Walborsky, A. E. Young, J. Am. Chem. Soc. 1964, 86,
3288–3296; b) C. Hamdouchi, M. Topolski, V. Goedken, H. M. Wal-
borsky, J. Org. Chem. 1993, 58, 3148–3155; c) V. Schulze, R. W.
Hoffmann, Chem. Eur. J. 1999, 5, 337–344; d) R. W. Hoffmann, B.
Hçlzer, O. Knopff, K. Harms, Angew. Chem. 2000, 112, 3206–3207;
Angew. Chem. Int. Ed. 2000, 39, 3072–3074; e) R. W. Hoffmann,
P. G. Nell, R. Leo, K. Harms, Chem. Eur. J. 2000, 6, 3359–3365;
f) R. W. Hoffmann, Chem. Soc. Rev. 2003, 32, 225–230; g) B.
Hçlzer, R. W. Hoffmann, Chem. Commun. 2003, 732–733; h) P. R.
Carlier, Y. Zhang, Org. Lett. 2007, 9, 1319–1322; i) J. Beckmann, A.
Schutrumpf, Org. Biomol. Chem. 2009, 7, 41–42.
[6] a) P. Knochel, W. Dohle, N. Gommermann, F. F. Kneisel, F. Kopp, T.
Korn, I. Sapountzis, V. A. Vu, Angew. Chem. 2003, 115, 4438–4456;
Angew. Chem. Int. Ed. 2003, 42, 4302–4320; b) F. F. Fleming, Z.
Zhang, W. Liu, P. Knochel, J. Org. Chem. 2005, 70, 2200–2205.
[7] This modification significantly improved the e.r. of the product [D1]-
3, by avoiding unnecessary warming of the reaction during the
quench (see Figure 1 and Table 1). Thus Mg/Br exchange in Et2O is
slightly more retentive than reported in our earlier publication
(Ref. [5h]).
[14] a) S. Sakamoto, T. Imamoto, K. Yamaguchi, Org. Lett. 2001, 3,
1793–1795; b) D. Seyferth, Organometallics 2009, 28, 1598–1605.
[15] a) J. Tammiku-Taul, P. Burk, A. Tuulmets, J. Phys. Chem. A 2004,
108, 133–139; b) F. W. Walker, E. C. Ashby, J. Am. Chem. Soc.
1969, 91, 3845–3850.
[16] a) C. Dohmeier, D. Loos, C. Robl, H. Schnçckel, D. Fenske, J. Orga-
nomet. Chem. 1993, 448, 5–8; b) D. E. Hibbs, C. Jones, F. A. Ri-
chards, J. Chem. Soc. Dalton Trans. 1999, 3531–3532.
[17] P. Brandt, F. Haeffner, J. Am. Chem. Soc. 2003, 125, 48–49.
[18] As we show in the Supporting Information, the term 2XR-1 is equiv-
alent to {[(R)-2] ꢀ [(R)-2]e}/{[(R)-2]0 ꢀ [(R)-2]e}, where [(R)-2],
[(R)-2]e, and [(R)-2]0 are the time-dependent, equilibrium, and initial
concentrations of (R)-2. For our studies on the dependence of kenant
on Grignard concentration, we chose the easily measurable quantity
[Mg]total: Note that [(R)-2]e and [(R)-2]0 are both proportional to
[Mg]total, and [(R)-2] is related to [Mg]total by exponential decay from
[(R)-2]0. Although this first-order rate constant kenant does not mea-
sure the decay of [Mg], it measures the rate of decay of the excess
concentration of (R)-2 over its equilibrium value, which is related to
[Mg]total. It is therefore appropriate to study the effect of varying
[Mg]total on kenant
.
[19] J. H. Espenson, Chemical Kinetics and Reaction Mechanisms, 2nd
ed., McGraw-Hill, New York, 2002, pp. 16–17.
[20] Note that the high e.r. (98:2) seen at [Et2O]=0.06m at 195 K allows
us to set 49:1 as a lower limit on the retentive versus invertive deut-
eration stereoselectivity.
[21] M. P. Bernstein, F. E. Romesberg, D. J. Fuller, A. T. Harrison, D. B.
Collum, Q. Y. Liu, P. G. Williard, J. Am. Chem. Soc. 1992, 114,
5100–5110.
[22] P. J. Zhao, D. B. Collum, J. Am. Chem. Soc. 2003, 125, 4008–4009.
[23] F. Bickelhaut in Grignard Reagents: New Developments (Ed.: J.
Herman J. Richey), Wiley, New York, 2000, pp. 299–328.
[24] T. Mori, S. Kato, J. Phys. Chem. A 2009, 113, 6158–6165.
[25] P. R. Carlier, Chirality 2003, 15, 340–347.
[26] G. Boche, K. Harms, M. Marsch, J. Am. Chem. Soc. 1988, 110,
6925–6926.
[27] a) H. Kwart, P. A. Silver, J. Org. Chem. 1975, 40, 3019–3026; b) M.
Oki, Pure Appl. Chem. 1989, 61, 699–708; c) M. Oki, H. Ikeda, S.
Toyota, Tetrahedron Lett. 1998, 39, 7729–7732.
[28] N. Deora, P. R. Carlier, J. Org. Chem. 2010, 75, 1061–1069.
Received: July 21, 2011
Published online: September 21, 2011
Chem. Eur. J. 2011, 17, 12250 – 12253
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12253