Stereochemistry of the Menthyl Grignard Reagent: Generation, Composition, Dynamics, and Reactions with Electrophiles

Article abstract of DOI:10.1021/acs.joc.8b02278

Menthyl Grignard reagent 1 from either menthyl chloride (2) or neomenthyl chloride (3) consists of menthylmagnesium chloride (1a), neomenthylmagnesium chloride (1b), trans-p-menthane (4), 2-menthene (8), 3-menthene (9), and Wurtz coupling products including symmetrical bimenthyl 13. The diastereomeric ratio 1a/1b was determined in situ by ¹³C NMR or after D₂O quenching by ²H NMR analysis. Hydrolysis of the C-Mg bond proceeds with retention of configuration at C-1. The kinetic ratio 1a/1b from Grignard reagent generation (dr 59:41 at 50 °C in THF) is close to the thermodynamic ratio (56:44 at 50 °C in THF). Carboxylation of 1 at -78 °C separates diastereomers 1a/b to give the anion of menthanecarboxylic acid (19) from 1a, which combines with unreactive 1b to give neomenthylmagnesium menthanecarboxylate (1b^I). The kinetics of epimerization for the menthyl/neomenthylmagnesium system was analyzed (ΔH^? = 98.5 kJ/mol, ΔS^? = -113 J/mol·K for 1b^I → 1a^I). Reactions of 1 with phosphorus electrophiles proceed stereoconvergently at C-1 of 1a/b to give predominantly menthyl-configured substitution products: PCl₃ and 2 equiv of 1 give Men₂PCl (6), which hydrolyzes to dimenthylphosphine P-oxide (7), whereas Ph₂PCl with 1 equiv of 1 gave P-menthyldiphenylphosphine oxide (27) after workup in air.

Full text of DOI:10.1021/acs.joc.8b02278

Subscriber access provided by UNIV OF NEWCASTLE
Article
Stereochemistry of the Menthyl Grignard Reagent: Generation,
Composition, Dynamics, and Reactions with Electrophiles
Sebastian Koller, Julia Gatzka, Kit Ming Wong, Philipp J. Altmann, Alexander Pöthig, and Lukas Hintermann
J. Org. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.joc.8b02278 • Publication Date (Web): 22 Oct 2018
Downloaded from http://pubs.acs.org on October 24, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 67
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Stereochemistry of the Menthyl Grignard Reagent: Generation, Composition,
Dynamics, and Reactions with Electrophiles
†
†
†
‡
‡
Sebastian Koller, Julia Gatzka, Kit Ming Wong, Philipp J. Altmann, Alexander Pöthig, and Lukas
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†
Hintermann *
†
Technische Universität München, Department of Chemistry, and Catalysis Research Center,
Lichtenbergstr. 4, 85748 Garching bei München
‡
Technische Universität München, Chair of Inorganic Chemistry, and Catalysis Research Center, Ernst-
Otto-Fischer Strasse 1, 85747 Garching bei München
lukas.hintermann@tum.de
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 2 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT
Menthyl Grignard reagent 1 from either menthyl chloride (2) or neomenthyl chloride (3) consists of
menthylmagnesium chloride (1a), neomenthylmagnesium chloride (1b), trans-p-menthane (4), 2-men-
thene (8), 3-menthene (9) and Wurtz coupling products including symmetrical bimenthyl 13. The
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
13
2
diastereomeric ratio 1a:1b was determined in situ by C NMR-, or after D O-quenching by H NMR
2
analysis. Hydrolysis of the C–Mg bond proceeds with retention of configuration at C-1. The kinetic ratio
1a:1b from Grignard reagent generation (dr 59:41 at 50 °C in THF) is close to the thermodynamic ratio
(
56:44 at 50 °C in THF). Carboxylation of 1 at –78 °C separates diastereomers 1a/b to give the anion of
menthanecarboxylic acid (19) from 1a, which combines with unreactive 1b to give
I
neomenthylmagnesium menthanecarboxylate (1b ). The kinetics of epimerization for the
‡
‡
menthyl/neomenthyl-magnesium system was analyzed (DH = 98.5 kJ/mol, DS = –113 J/mol·K for
1bI®1a ). Reactions of 1 with phosphorus electrophiles proceed stereoconvergently at C-1 of 1a/b to
I
give predominantly menthyl-configured substitution products: PCl
3
and two equivalents of 1 give
PCl with a single
Men PCl (6), which hydrolyzes to dimenthylphosphine P-oxide (7), whereas Ph
2
2
equivalent of 1 gave P-menthyl-diphenylphosphine oxide (27), after workup in air.
INTRODUCTION
Menthyl Grignard reagent 1 is prepared in the usual way from menthyl chloride (2) and magnesium in
etheric solution.1,2 This chiral organometallic reagent is widely used to attach menthyl groups to elec-
trophilic centers. Reactions have been described e.g. with carbon dioxide,3,4,5,6,7,8,9 formaldehyde,9
4
,1,10,11 PhPCl
,12, Ph
PCl,2,13,14 GeCl
,15 SnCl
or organotin chlorides.11a,d,16 The
dichloromethane, PCl
3
2
2
4
4
reagent has also been used in cross-coupling17,18 or transmetallation reactions. The products tend to
19
retain menthyl configuration, and the sequence from 2 via 1 to various substitution products (Men-E)
ACS Paragon Plus Environment

Page 3 of 67
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
has usually been assumed to proceed with retention of configuration in all steps (Scheme 1).8,9,14,20,21 The
occasional observation of epimeric products (Nmn-E) as minor components revealed that retention is
not strict. It is not immediately obvious if inversion occurs at the stage of metal insertion (Scheme 1, I),
by epimerization of menthyl- (1a) to neomenthylmagnesium chloride (1b) upon storage (II), during
electrophilic substitution (III), by epimerization of the final product (Scheme 1, IV), or in several of
those steps.22
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 1. Stereochemical aspects of reactions via menthyl Grignard reagent to electrophilic
substitution products.a
Me
Me
Me
E+
Mg
Cl retention
MgCl
Me
E
Me
retention
Me
Me
Me
Me
I
III
2
1a
Men-E
II
IV
inversion
inversion
Me
Me
Me
Mg
E+
Cl
Me
MgCl
E
retention
retention
Me
Me
Me
1b
Me
Me
3
Nmn-E
+
a) E = electrophile. Men-E and Nmn-E represent menthyl or neomenthyl configured substitution
products. Men (menthyl) is (1R,2S,5R)-2-isopropyl-5-methylcyclohex-1-yl, Nmn (neomenthyl) is
(
1S,2S,5R)-2-isopropyl-5-methylcyclohex-1-yl.
2
Duthie et al. have developed a refined picture of 1 and its reactions: they found equal amounts of dia-
stereomeric (3-D )-p-menthanes ((3-D )-4) by quenching 1 with D O and concluded that the reagent is a
1
1
2
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 4 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1:1 mixture of configurationally stable epimers 1a/b, which forms stereoconvergently from either 2 or
2
neomenthyl chloride (3). They also obtained the same product spectrum in reactions of Ph
2
SnCl with
2
Grignard reagents from either 2 or 3 and proposed that the higher nucleophilicity of 1a over 1b – and
thus kinetic resolution – was responsible for the usual preference of menthyl-configured substitution
products. Their model implies that 1b remains unchanged in the reaction, that a maximal yield of 50 %
of Men-E will be formed in reactions with a single equivalent of 1, and that use of two or more equiva-
lents of 1 might raise yields to quantitative. This analysis (convergence in I, resolution in III) assumes
that III proceeds with retention of configuration at C-1, and that 1a/b, as well as products Men-E/Nmn-
E, are configurationally stable under reaction conditions.22,23 Both assumptions are unproved; in terms
of deuteration, Knochel and coworkers observed that menthylzinc iodide is an epimeric mixture with dr
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
65:35, but may react with MeOD to give (3-D
of 1a/b, no experimental data is available. A complementary, direct analysis of reagent 1 and the
elucidation of its equilibration kinetics are required to benchmark the validity of the D O-quenching
1
)-4 with dr >99:1 (Men/Nmn).18,24,25 As for epimerization
2
method and permit secure assignments of the stereochemistry of steps I and III. Statements on the
composition of 1 and the stereoselectivity of III that are solely based on the Men/Nmn-ratio of a sub-
stitution product are insecure. Particularly so, if the isolated product quantities are substantially lower
than the molar amount of 1 applied, meaning that kinetic resolution may have occurred.
The present work emerged from an analysis of synthetic routes to P,P-dimenthylphosphane-based chiral
phosphorus ligands.26,27 We have previously reported on the nucleophilic chlorination of menthol (5)28
2
9
to menthyl chloride (2) and accompanying rearrangements. In continuation of those studies, we wished
1
10
1
to investigate the synthesis of Men
2
PCl (6) from 1 and PCl , which is low-yielding (13 %, 25–35 % )
3
and suffers from the difficult separation of the water- and oxygen-sensitive 6 from by-products. We
anticipated that hydrolytic workup would convert 6 to dimenthylphosphine P-oxide (7),30 whose
expected stability towards water and air might facilitate its isolation (Scheme 2).
ACS Paragon Plus Environment

Page 5 of 67
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Scheme 2. In situ synthesis of P-chloro-dimenthylphosphane (6) and hydrolysis to
dimenthylphosphine P-oxide (7).
•
kinetic ratio or
•
stereocon-
epimerization?
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
vergence?
Me
Me
Me
Mg
D O
2
Cl
Me
THF
MgCl
Me
1 (α/β)
• retention ?
D
Me
Me
Me
Me
2
(β)
Cl
Cl
(3-D )-4
1
3 (α)
air- and water-
stable P-oxide?
• retention and PCl3
THF
resolution?
Me
Me
Me
Me
Me
H O
2
H
P
O
P
Me
Me
Cl
Me
Me
Me
Me
Me
6
7
If so accessible, 7 might find application as chiral secondary phosphine oxide (SPO) ligand,31,32 or as
33
platform chemical for chiral phosphine and phosphine oxide synthesis. Objectives of the present work
were thus: a) to study the synthesis and composition of Grignard reagent 1; b) to determine the diaste-
reomeric ratio 1a:1b by NMR spectroscopy in solution, and to validate D
analytical method for analyzing dr; c) to elucidate the temperature-dependent kinetics of equilibration of
a/b; d) to determine the stereochemistry of substitution of 1 with selected electrophiles; e) and to access
and isolate Men PHO (7).
2
O-quenching as quantitative
1
2
RESULTS AND DISCUSSION
Synthesis and Qualitative NMR Analysis of 1
3
4
Menthylmagnesium bromide had been prepared within a year of Grignard's discovery of his magnesium
5
O6,7 or THF solution.1,9,13
organometallics. The analogous chloride 1 was later obtained from 2 in Et
2
2
According to Duthie et al. reagent 1 is composed of non-epimerizing diastereomers 1a/b in a 1:1 ratio
in THF, irrespective of its generation from 2 or 3, which would imply kinetic control over the
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 6 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
diastereomeric ratio. Increasing the proportion of 1a in the mixture would then be desirable in syntheses
of menthyl configured targets, since the latter were assumed to emerge from the kinetically preferred
2
reaction of 1a with electrophiles. NMR spectroscopy should offer direct evidence of the composition of
Grignard reagent 1, and provide structural proof of all components.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
29
We prepared 1 from 2 (≥96 % purity) and excess magnesium, activated by iodine and 1,2-dibromo-
3
5
1
ethane, in (D
8
)-THF. The majority of signals in the H NMR spectrum (Figure 1) overlap in the aliphatic
region (d 0.5–2.5).
H
1
Figure 1. Excerpts from the H NMR spectrum (500 MHz) of 1 in (D
8
)-THF. The olefinic (d 5.5) and
high field (d 0) regions are scaled by a vertical factor of 16.
Olefinic signals at lower field are due to 2-menthene (8; d 5.51) and 3-menthene (9; d 5.36), sometimes
accompanied by y-menthene 10.36,37 A group of signals below d
0, centered by a major signal (d –0.08)
H
with multiplicity ddd (3JH,H = 13.3, 11.9, 2.9 Hz) show the typical shape of an axial H-1 in menthyl-
derivatives and are due to menthylmagnesium species that collectively belong to 1a. The major signal
likely represents MenMgCl(THF)
n
(1a'). A slightly broadened signal of similar shape at lower frequency
(
d –0.15) is ascribed to Men Mg as component of the Schlenk equilibrium (vide infra). Minor signals at
2
higher frequency tend to be less structured (see Figure S3 for examples) and may belong to MenMgX
compounds with other co-ligands or counter-ions.38,39 Integration of the CHMg-region against internal
ACS Paragon Plus Environment

Page 7 of 67
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
standard naphthalene in many experiments accounted for only 30–50 mol % of the RMgX-concentration
determined by titration.40 Therefore, H NMR spectroscopy gives an incomplete picture of the
1
1
3
composition of 1. The C NMR spectrum of 1 features a prominent signal at d 52.8, later assigned to C-
of 1b', which was set to 100 area-% as reference. Approximately 80 peaks at ≥5 area-% were detected
in total, besides those of the solvent and of the internal standard naphthalene. Signal sets for 2-menthene
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
7
37
(
8; 10 peaks, 19–24 area-%) and 3-menthene (9; 10 peaks, 10–12 area-%) were assigned by reference
4
1
42
data. Out of 26 major peaks (40–120 area-%), 6 were assigned to trans-p-menthane (4), and 10 each
43
to the predominant organomagnesium species 1a'/b'. Connectivity and peak assignments for 1a'/b' and
4
were elucidated by 2D NMR methods (HSQC, HMBC). The 13C NMR signals for the major
organometallic species are mapped in Figure 2.
MenMgCl (1a')
NmnMgCl (1b')
23.14
Me-C-5
37.41
23.27
Me-C-5
35.30
δ(13C):
5
5
37.66
44.93
37.70
33.43
44.49
30.92
4
3
6
1
4
3
6
1
H
30.57
H
2
9.40
2
49.50
MgCl
2
52.81
MgCl
3
5.92 1'
1' 35.45
H
Me-1'-b
Me-1'-ψ-eq
2.51
Me-1'-a
21.14*
Me-1'-ψ-ax
6.59
2
22.05*
1
1
δ( H):
−0.08 (H-1 ax)
+0.64 (H-1 eq)
Figure 2. NMR data in (D
8
)-THF for the major components in 1, MenMgCl (1a') and NmnMgCl (1b').
Solvent coordination or aggregation is not taken into consideration for simplicity. See the Supporting
1
Information for complete H NMR assignments.
The CHMg-unit of 1b' gives rise to a poorly structured multiplet at d
H
0.64 in (D
8
)-THF (cf. Figure 1).
This signal was often obscured in the more commonly analyzed THF–C
6
D
6
mixtures. Assignment of
signal sets to structures 1a'/b' relies on the characteristic shape of the H-1-signals. Furthermore, menthyl
13
and neomenthyl fragments are discernible by C NMR spectroscopy, where peaks for the diastereotopic
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 8 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
methyls of C-2-isopropyl are either clearly separated with one signal at d ≈ 15–17, the other at d ≈ 22–
4
4
2
3 (in Men-Y), or both found in close vicinity at d ≈ 21–22 (Nmn-Y). These chemical shifts reflect the
preferred conformation of the C-2-isopropyl group, which stands equatorial, with both methyl groups in
y-equatorial positions relative to the cyclohexane chair in Nmn-Y derivatives (Y = variable substituent;
Figure 3, b), unless Y becomes sterically very demanding (Figure 3, c). In most menthyl derivatives
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
including 1a', repulsion of groups Y at C-1 and iPr at C-2 causes the isopropyl group to rotate such that
one methyl arranges y-axial and the other y-equatorial relative to the cyclohexane chair (Figure 3, a).44,45
a)
b)
c)
Me
Me
Me
Y
H
H
Me
H
H
Me
Me
Y
Me
H
Y
Me
Me
H
Men-Y
Nmn-Y
Nmn-Y
small and
medium Y
small or large Y
large Y
Y = -AuCl
-SR,
-PPh {RuCl (arene)}
2 2
-
GePh3
PtBr(dppe)
-PPh (PdL ),
2
3
-
-Sn(Men)Ph2
Figure 3. Preferred conformations of Men-Y and Nmn-Y derived from NMR and X-ray crystal
structure data. a) Conformation of menthyl derivatives. b) Conformation of neomenthyl derivatives. c)
Ring-flipped conformation in neomenthyl derivatives with large Y-groups.45
1
3
At this point, 34 signals in the region d 15–48 of the C NMR spectrum of 1 remained unassigned. GC-
MS-analysis (Figures S15–S17) of a D O-quenched sample of 1 indicated that the signals belong to three
stereoisomeric hydrocarbons (m/z 278.3), namely bimenthyl (13), menthyl-neomenthyl (14) and
bineomenthyl (15). A batch of 1 was quenched with aqueous acid, washed with concd H SO and vacuum
distilled. The more volatile fraction consisted of 4, with a minor amount of y-menthane (1-isobutyl-3-
2
2
4
4
6
methylcyclopentane; 16, 96:4 ratio). Distillation of the residue in high vacuum gave a bimenthyl isomer
1
fraction (13:14:15 = 39:45:16) as colorless oil. Its H NMR spectrum covered the narrow shift range of
ACS Paragon Plus Environment

Page 9 of 67
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
d 0.6–2.2 with signal overlap even at 950 MHz (Figure S18). Based on HMBC and HSQC and the d
C
criterion for discerning menthyl from neomenthyl fragments, all 40 carbon and 52 proton signals in 13–
15 could be assigned (Figures 4, S19, S21, S22).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
3.10
23.17
33.93
2
3.40
34.86
33.45
2
3.06* 22.31*
9.89
3
3.19
3
4.70
35.65
H
35.22 40.80
2
2
4.76 37.92
3.68
2
5.26
47.48
26.58
45.20
49.24
37.25
3
3.86
24.75 28.34
4
2
5.78
H
H
48.73
4
4.21
25.31
35.23
2
8.20
2
1.92
1
5.41
22.16
15.94
27.24
2.58
2
1.59*
Nmn
15
2
2.13*
2
Men
Men
Men
Nmn
Nmn
1
3
14
peak line
broadening (LB):
LB ≤1.5 Hz
LB 2 Hz
LB 3–4 Hz
LB 5–6 Hz
13
Figure 4. C NMR assignments for the Wurtz hydrocarbon byproducts Men-Men (13), Men-Nmn
(
14), and Nmn-Nmn (15). An asteriks (*) indicates interchangeable assignments.
Unlike symmetrical bimenthyl 13, hydrocarbons 14 and 15 show d line broadening in specific positions,
C
indicative of a dynamic chair-flip of the neomenthyl units. The oily hydrocarbon mixture separated
crystals of bimenthyl (13) upon standing, as revealed by NMR spectroscopy and X-ray crystal structure
analysis (Figure 5). Compound 13 melts at 105–106 °C, and is thus identical with a crystalline bimenthyl
isomer from the Wurtz-coupling of 2 described in the older literature.47
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 10 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 5. X-ray solid-state molecular structure of bimenthyl 13. Ellipsoids are shown at 50 %
probability.
Quantitative Analysis of Menthyl Grignard Reagent
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
40
Grignard reagent 1 was analyzed volumetrically for total alkylmagnesium, and for total magnesium by
4
8
1
EDTA-titration of acid-quenched samples. Quantitative H NMR spectroscopy (qNMR) with naphtha-
lene as internal standard permitted quantification of menthenes 8/9 by integration of the well-separated
olefinic signals. Since the d 0.5–4 region did not contain well resolved signals for integration, we usually
prepared 1 in non-deuterated THF and admixed C
D
6 6
for locking prior to NMR analysis. As in a previous
2
9
13
study of the composition of crude menthyl chloride, the high resolution and information density of C
NMR recommended use of this technique for quantitative analysis.49,50,51 Naphthalene (d 128.6) served
as internal standard under routine analysis conditions, and empirical correction factors for peaks of
5
2
individual components, positions and carbon types (CH vs. CH
2
vs. CH ) were applied. Correction
3
factors were derived from analysis of reference materials (4, 13–15) against internal standard, or by
1
referencing to quantitative H NMR data (8, 9). Since reference material is not available for 1a/b, their
5
2
correction factors were estimated from structurally similar model compounds. Deconvolutive peak
fitting was applied to reduce subjective errors and correct for peak overlap. Quantification of 1a/b
presented additional challenges due to the Schlenk and other ligand exchange equilibria. The population
of Men-Mg and Nmn-Mg units is not correlated with the signal intensity of a single species (1a'/b'), but
must be summed over various chemical forms, some of which display chemical exchange-broadened
signals (Figure 6).
a)
b)
c)
ACS Paragon Plus Environment

Page 11 of 67
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
13
Figure 6. Excerpts from H (top; 500 MHz) and C NMR (bottom; 101 MHz) spectra of Grignard
reagent 1. a) Reagent prepared from 2 and Mg in (D
8
)-THF (prepared at 70 °C, nominal concentration
0
.9 M); [RMg]/[Mg] = 0.49. b) Higher concentrated 1 in THF (50–70 °C, 1.55 M) with C added for
6
D
6
locking, displaying a high content of Men
2
Mg; [RMg]/[Mg] ratio = 0.9. c) Reagent prepared from 3
1
and Mg in (D
8
)-THF (rt, 0.9 M); [RMg]/[Mg] ratio = 0.4; H NMR at 400 MHz.
1
13
1
Figure 6a shows excerpts of the H and C NMR spectra for 1. The characteristic signal of 1a' in the H
NMR spectrum (H-1, d 0.08) correlates (by HMBC) with C-2 at d 49.5, whereas H-1 of NmnMgCl
1b'; d 0.64, cf. Figure 1) correlates with C-2 at d 52.8. The intensity ratio of the major C-2-peaks does
H
C
(
H
C
not reflect the ratio 1a/b: the major C-2 peak for 1a' is accompanied by broad signals, and the total
53
integral d
C
48–51 (1a-region) surpasses that of d 51–54 (1b-region). Figure 6b shows spectra of 1 at
C
5
4
high [R-Mg]/[Mg] ratio. As expected, the signal at d
H
–0.19 for Men
2
Mg is more intense than for a low
Mg and
either MenMgBr or more complex species. Satisfactory analytical results for the population and dr of
5
5
[RMg]/[Mg] ratio. The 1a-region contains at least three broadened signals for C-2 of 1a', Men
2
38
56
1
resulted from area integration of d 51–54 (for 1b) and d 48–51 (for 1a), after subtraction of overlapping
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 12 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
5
7
peak signals. The peak area correction factor for C-2 in 1a/b relative to internal standard signal was set
to the value for menthol (5) as model compound. This introduces an unknown systematic error, which
5
8
we estimate to be below 5–10 % of the absolute values, but will be even less important for the ratio
1a:1b, where such systematic errors cancel out.50a The data thus obtained for all key components in
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Grignard reagent 1 will be subject to additional systematic errors by variation of T and the NOE effect
1
with sample composition, but even if accuracy for individual component concentrations is limited,
parameter variation effects on the composition of 1 can now be systematically studied.
Table 1. Analyses of menthyl Grignard reagent 1, prepared from 2 or 3 under variable reaction
conditions.a
Me
Me
Me
Me
Me
H
H
H
H
MgCl
Me
MgCl
Me
H
H
H
Me
Me
Me
Me
Me
Me
Me
Me
1
a
1b
4
8
9
10
13 (Men2)
14 (Men-Nmn)
15 (Nmn )
2
Entry Solvent
t
Conc. [RMg] [Mg] Yield 1 [%] 1a:1b
4
8
9
13– 13:14:15 Recov.
°C] [M]b
[M]c
[M]d
[%]e
[%]g
[
[%] [%] [%] 15
%]f
[
Reactions starting from menthyl chloride (2)
1
2
2
3
4
THF
THF
rt
rt
1.8
0.9
0.97
0.64
1.59
0.90
53
72
54.1
74.0
55:45
62:38
12.2 5.5 2.9 25.9 36:47:17 100.5
9.7
3.5 1.9 20.1 32:50:18 109.4
a
2
(D O: 62:38)
THF
THF
50
50
1.8
0.9
1.44
0.77
1.88
0.95
80
86
72.5
74.3
59:41
59:41
6.4
8.3
2.4 1.2 11.8 32:50:18 94.5h
2.9 1.5 11.8 36:49:15 98.8
4a
5
(D
76.9
(D O:
2
O: 59:41)
THF
50i
0.9
0.76
0.85
85
58:42
7.2
2.3 1.2 11.7 31:57:12 99.4
55:45)j
5
6
7
8
9
a
2
THF
70
0.9
0.9
1.1
1.8
0.74
0.65
0.72
1.24
0.92
1.18
1.04
1.51
81
72k
67
70
87.5
45.1
68.6
71.4
55:45
56:44
53:47
55:45
6.5
1.6 0.9 8.8
38:45:17 105.3
Et
2
O
RT
15.3 6.0 3.1 29.2 37:47:16 99.1
106.6l
THF–PhMe RT
THF–PhMe 50
10.6 4.4 2.3 20.5 36:48:16
8.4
3.5 1.7 16.2 40:44:16 101.3
ACS Paragon Plus Environment

Page 13 of 67
1
The Journal of Organic Chemistry
10
THF–PhMe 50
THF–PhMe 70
1.1
1.8
0.80
1.33
1.00
1.58
75
75
73.0
78.4
54:46
55:45
8.2
6.7
2.4 1.4 10.8 37:45:18 96.0m
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
1
1
1
1.6 0.9 8.5
32:54:14 96.2
1a
2
2
(D O: 56:44)
THF–PhMe 70
1.1
0.95
1.10
85
85.3
58:42
4.9
1.1 0.6 5.3
44:47:9
97.2
Reactions starting from neomenthyl chloride (3)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
105.0n
1
1
3
4
THF
THF
rt
0.9
0.9
n.d.
n.d.
n.d.
60
45.7
44.2
45:55
48:52
20.5 5.5. 2.5 20.3 36:45:19
9.9
5.1 2.7 20.7 35:46:19 83.7o
30
0.55
1.39
14a
(D
2
O: 42:58)p
100.2q
1
5
THF
50
0.9
0.66
1.37
73
63.6
54:46
13.6 3.0 2.0 16.6 35:48:17
a
b
Sample composition in mol % of initial 2 (or 3); concentrations in M = mol/L. Nominal concentration calculated from initial 2/3 and
c
40 d
2+
volumes of starting material and solvent, rounded to 0.1 M. Total concentration of RMg-entities by titration. Total Mg -concentra-
e
f
tion by EDTA titration. Yield of 1 calculated from [RMg]tot and nominal concentration. Value expressed in mol % of Men/Nmn
g
groups; the true molarity of dimers is half this value. Recovery is the sum of analytically detected Men and Nmn groups in mol %;
h
i
deviations from 100 % are due to undetected components and systematic errors. Including 0.2 mol % of 10. Reagent prepared at 50
j
2
°C over 1 h, then heated to 70 °C for 2 h. D
2
O-quenching result 1a/1b/16: 43.3:35.5:3.0 mol % by H NMR against C
6 6
D as internal
k
standard. Large deviation due to solvent loss and heterogeneity; NMR analysis performed after evaporation of Et
2
O and redissolution
l
m
n
o
in THF. Including 0.2 % of 10. Including 0.3 % of 10. Including 10.4 % of 3. NMR analysis after ≥ 4 h, with 1.0 % of remaining 3.
pD
O quench performed after 2 h, at 10 % of remaining 3. Including 1.4 % of 3.
q
2
Effects of Reaction Conditions on the Composition of 1
Effects of varying the solvent, reaction temperature and total concentration in the reaction of 2 and
magnesium were studied by analyzing the composition of the resulting Grignard reagent (Table 1, entries
1–12). The yield of 1 was higher at 50 or 70 °C than at ambient temperature or 30 °C, and also higher in
dilute (0.9–1.1 M) over concentrated (1.8 M) solution. Synthesizing the reagent in THF–toluene at 70 °C
or in THF at 50–70 °C gave similar, satisfactory results with a yield of 75–85 mol %, whereas diethyl
ether (entry 7) was less satisfactory. Reagent yields in THF and Et O agree with earlier reported titration
2
results.1,6 Spectroscopic yields of 1 do not correlate exactly with [RMg] from titration, since yield
calculations depend on estimated reaction volumes, which are inaccurate in small scale experiments.
In a first approximation, the diastereomeric ratio 1a:1b deviates not much from an average value of
57:43, irrespective of the reaction conditions. Complementary data for dr was in some cases obtained by
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 14 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
quenching a sample with D
2
O and analyzing the ratio of (3-D )-4 stereoisomers by H NMR spectroscopy
1
Figure 7).2,59
(
a)
b)
c)
d)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
1
Figure 7. H{ H} NMR spectra (61 MHz) of quenched samples of 1. a) D
2
O-quenched sample of
O-quenched sample of the reaction 3 + Mg in THF, before completion (Table 1,
entry 14a). c) Carboxylated and D O-quenched 1 composed mainly of a-(3-D )-4. d) H O-quenched
sample of 1 generated in (D )-THF. Most (D )-THF was removed by washing with water prior to
analysis; only residual (D )-THF is present. "y" marks peaks for (D )-16.
equilibrated 1. b) D
2
2
1
2
8
8
8
1
1
The signals d
D
1.65 and 0.94 correspond to the centers of multiplets in the H NMR spectrum of 4 at d
1.16 corresponds to D-1' of
1.18) from deuterolysis of sec-y-MenMgCl (12). A Grignard reagent from NmnCl
3; Table 1, entries 13–15) was largely identical to that from 2, as also evident from the similar NMR
H
1
.68 and 0.97 for equatorial and axial H-3, respectively. A small peak at d
D
3
9
y-menthane (16; d
H
(
spectra (Figure 6, a vs. 6, c)2,60 – at least if prepared at 50 °C (Table 1, entry 4 vs. 15). The room
temperature Grignard reaction of 3 was sluggish to initiate and required several hours for reaching
completion (entries 13, 14). A D
2
O-quenched sample after 2 hours at 30 °C gave (3-D )-4 with dr 42:58
1
2
at 90 % conversion (entry 14a; H NMR in Figure 7, b), which rose to dr 48:52 after 4 h (at 99 %
conversion; entry 14). Conversely, if 1 was generated from 2 at ambient temperature, a sample at 40 %
conversion displayed a higher dr 62:38 (Table 1, entry 2). Thus, Grignard reagent 1 is formed with a
ACS Paragon Plus Environment

Page 15 of 67
1
The Journal of Organic Chemistry
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
small degree of retention of configuration at ambient temperature, but approaches dr 56:44 if prepared at
0 or 70 °C from either 2 or 3, which implies that the latter is an equilibrium ratio (vide infra).
The abundance of byproduct p-menthane (4) is always higher than the sum of menthenes 8 and 9. Radical
abstraction from the solvent provides no source of 4, because when 1 was prepared in (D )-THF, no trace
)-4 or solvent coupling-product was detected by H{ H} NMR after protic workup (Figure 7,
5
8
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
1
of (3-D
1
6
1
d). Byproducts 4, 8 and 9 are most abundant at room-temperature (Table 1, entries 1, 2, 13) and decrease
at higher temperature (entries, 3, 6, 12). Alkenes 8/9 are formed in similar amounts and isomeric ratio
from both 2 and 3 (entry 15), which excludes E2-elimination (with 1 as base) as an important reaction
6
2
pathway. This points to a specific mechanistic route for products 4/8/9, besides the obvious source of
S
some 4 by reaction of 1 with adventitious water. We suggest that b-H-elimination of surface-bound MgR
S
S
(
R = Men or Nmn) to surface-bound Mg–H and alkenes 8/9 is followed by combination of MgH and
SMgR to give 4. Bimenthyl isomers 13–15 will similarly be formed by combining two surface-bound
SMgR units, which agrees with the observation that the extent of Wurtz-coupling and the generation of
4
+8/9 is correlated. The observation of higher amounts of neutral byproducts (R–H, R–R) at ambient
temperature, but higher selectivity for Grignard reagent 1 at elevated temperature might be a consequence
of rate-limiting detachment of RMgCl from the metal surface. This process, involving major solvent
reorganization, will have a higher activation barrier than the release of apolar hydrocarbons (RH, RR),
which is likely limited by diffusion-controlled encounter of two surface groups. The stereoconvergence
from 2 and 3 to give 1a/b, the generation of bimenthyls 13–15 and the absence of solvent-derived-
S
products all point to a key mechanistic role of MgR-units, in which R inverts at experimental timescale,
much slower than a free radical, but faster than the final RMgX.63,64
Conclusions on the Generation of 1
The analyses in Table 1 validate the hypothesis that menthyl Grignard reagent 1 is a diastereomeric
mixture 1a/b, formed in a kinetically controlled and stereoconvergent process from either 2 or 3 in similar
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 16 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
amounts. Irrespective of reaction conditions, the ratio 1a/b varies only little around an average dr of
5
7:43. Presumably, this ratio is not just kinetically controlled, but close to thermodynamic equilibrium.
1
3
In spite of the valuable data obtained through C NMR spectroscopy, the faster and more accurate option
2
1
for analyzing the dr of 1a/b is H{ H} NMR spectroscopy of D O-quenched samples. The method gives
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
13
accurate results that agree with the in situ C NMR data, but does not suffer from complications caused
n
by species partitioning in the Schlenk equilibrium. Total [RMg] or individual R MgX concentrations in
mixtures can be quantified using the D O-quenching technique, if a deuterated internal standard is added
2
2
1
65
for H{ H} NMR analysis. Given that the ratio 1a:1b cannot be satisfactorily controlled by the reaction
conditions used for generating the Grignard reagent, an understanding of the stereochemistry of the
reactions of 1a/b with electrophiles becomes even more vital.
Carboxylation of Menthyl Grignard Reagent
Carboxylation of 1 has been reported repeatedly, but with inconsistent results:8,3,6 By pouring a THF
solution of 1 over solid CO
2
, Bose et al. obtained 76 % of acids 19/20 (Scheme 3) with a dr of 68:32,3b
8
while Grayson et al. obtained pure 19 (dr >100:1) in 88 % yield. By bubbling CO
2
into etheric 1, Smith
and Wright got 29 % of crystalline 19, and distilled another 32 % of a liquid (believed to be 20) from the
mother liquors.6,66 The same procedure provided Ueda et al. with 36 % of 19/20 in a ratio of 92.6:7.4,20c
67
whereas a 1975 patent found almost pure acid 19 (dr 99:1) in 53 % yield. Among the highest yielding
carboxylations is a patent that finds 93 % of 19 by concurrent addition of CO and 2 to magnesium in 2-
2
MeTHF–cymene at 90 °C.68
Scheme 3. Carboxylation of menthyl Grignard reagent 1, and synthesis of physiological coolant
compound WS-5 (22).
ACS Paragon Plus Environment

Page 17 of 67
1
The Journal of Organic Chemistry
2
Me
Me
Me
a)
1
2
. CO2
3
4
5
6
7
8
9
1
+
+
. H+ aq
CO H
CO H
2
2
CO H
2
Me
b)
Me
19
Me
Me
Me Me
Me
20
21
Me
Me
Me
1. SOCl2
O
H
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N
X
2.
OEt
CO Et
2
O
1
7 (X = MgCl)
Me
Me
NH ·HCl
2
18 (X = Cl)
22 WS-5
To understand the various results, we approached the problem with own experiments. Dry ice was added
2
9
to a Grignard reagent from crude 2 that contained tert-y-MenCl (18) as major impurity. Acid/base-
extraction gave a carboxylic acid fraction with 19, 20 and 21 in a ratio 77:19:4 (Scheme 3, a). Coupling
with ethyl glycinate gave a mixture of peptide esters (ratio 78:18:4) from which the major component 22
crystallized (Scheme 3, b). Compound 22 is commercially available under the name WS-569 as
7
0
physiological coolant for inducing a cooling sensation on skin. The substance has so far only been
described in the patent literature, thus we characterized it by 2D NMR methods in solution (Figure S26,
Tables S21/22) and by X-ray crystal structure analysis in the solid state (Figure 8).
Figure 8. Solid state molecular structure of WS-5 (22). Ellipsoids are shown at 50 % probability.
Oxygen is red, nitrogen blue. The common arrangement of the C-2 isopropyl group with y-axial and -
equatorial methyl groups (d 16.16 and 21.49) is evident. The unit cell contains six independent
C
1
2'
1'
molecules, which differ in the conformation of the peptidic side-chain [Men-C (O)-NH-C H -C (O)O-
2
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 18 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
''
2''
1
2'
1'
1'
1''
2''
C H
The individual conformers and their angles are: a) –75/–89; b) –87/–179; c) +91/+96; d) +86/+180; e)
60/+178; f) +58/–173. The pictured conformer is (c). See Figure S28 for depictions of all conformers.
2
C H
3
], in which the torsional angles C (O)–N–C –C (O) and C (O)–O–C –C are variable.
+
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The complexity of crude menthyl chloride had translated into regio- and stereoisomeric carboxylic acids
19–21) via metallation and carboxylation, but the ratio 1a/b (57:43) was not retained in acids 19/20 (dr
80:20). In ensuing carboxylation experiments, Grignard reagent 1 from purified menthyl chloride (2+3
(
≥
97 %) was used and the yield of 19/20 was determined by accurate qNMR methods. In one expriment,
1
was added dropwise to a saturated solution of CO in THF at –20 °C and the reaction was quenched
2
with water at the same temperature (Scheme 4, a).
Scheme 4. Carboxylation of Grignard reagent 1 with quantification of carboxylic acids 19/20.
Me
Me
Me
Me
1. CO , THF
2
(2 h)
MgCl
Me
MgCl
Me
. H+ aq
CO H
CO H
2
2
2
Me
Me
Me
Me
Me
Me
1
a
1b
19
20
a) Addition of 1 to CO2 (THF) at −20 °C
1
a
1b
Yield
dr
95:5
19
20
5
7%
43%
58.7%
55.9%
2.8%
b) Addition of CO2 (s) to 1 at ambient temperature
1
a
1b
Yield
dr
19
20
5
7%
43%
67.8%
89:11
60.2%
7.6%
A second experiment involved the addition of small pieces of CO (s) into the solution of 1 at ambient
2
temperature, which corresponds to an inverted addition mode (Scheme 4, b). The outcome in (b) agrees
with several literature reports that find 19 as major and 20 as minor component (dr ≈ 8:1; Scheme 4, b).
Surprisingly, experiment (a) in which 1a/b are continuously exposes to excess electrophile showed
higher selectivity for 19 (dr 19:1) rather than approaching the ratio 1a/b (57:43). The data in Scheme 4
ACS Paragon Plus Environment

Page 19 of 67
1
The Journal of Organic Chemistry
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
suggest that dr 19/20 is not controlled by availability of (excess) electrophile, but by kinetic preference
of 1a over 1b in the reaction with electrophile. The lower yield and higher dr in (a) reflect improved
kinetic resolution at lower reaction temperature. To verify this assumption, not only the reaction products
19/20, but also the unreacted Grignard reagents 1a/b must be analyzed, which was achieved by D
2
O-
quenching in combination with use of a deuterated internal standard for H{ H} NMR. Such ex-
periments prove that the reaction of 1 with excess CO at –78 °C is limited to epimer 1a and proceeds
with retention of configuration at C-1 to give 19 (Scheme 5, a). Neomenthyl epimer 1b remains largely
unreacted and gives rise to enriched 3a-(D )-4 after quenching with D O (Scheme 5, a). In repeat experi-
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
1
71
2
1
2
ments, the kinetic resolution was consistently high, up to the extent where no more 3b-(D
1
)-4 could be
detected (≥99.5:0.5).
Scheme 5. Kinetic resolution of menthyl Grignard reagent 1 by low-temperature carboxylation,
and epimerization studies.
a)
Me
Me
Me
Me
Me
Me
1
. CO2 (11 equiv.)
THF, −78 °C, 1 h
+
+
+
MgCl
Me
MgCl
Me
CO H
CO H
D
D
2
2
2. D O
2
3
. H+
Me
Me
Me
Me
19
Me
Me
Me
Me
Me
Me
1
a
1b
20
3α-(D )-4
1
dr
(68:1)
3β-(D )-4
1
dr 57:43
mol% by 1/2H NMR: (57.6%)
(0.0%)
(39.4%)
(0.6%)
b)
Me
Me
Me
Me
Me
3
. D O
2
1
. CO , THF
2
+
+
−
78 °C, 1 h
4. H
+
MgX
Me
MgX
D
D
CO H
2
2. Vacuum
Me
Me
Me
Me
Me
Me
Me
Me
Me
(
- CO2)
1
1
b (X = Cl)
1a (X = Cl)
1a (X = O CMen)
2
3α-(D )-4
1
3β-(D )-4
19
1
I
I
b (X = O CMen)
2
a) D O-quenching for analysis of unreacted Grignard reagent. b) Experimental course for studying the
2
epimerization of resolved neomenthylmagnesium menthanecarboxylate 1bI.
Kinetics and Equilibrium in the Epimerization of 1
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 20 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The kinetic resolution of 1 with CO gives access to highly enriched solutions of neomenthyl-configured
2
I
Grignard reagent 1b , from which excess electrophile can be completely removed in vacuum. The
7
2
I
kinetics of epimerization of chiral Grignard reagent 1b could now be studied by warming to a target
temperature and taking samples for D O-quenching (Scheme 5, b). Analysis of the dr of 3-(D )-4 by
H{ H} NMR provided consistent results as shown in Figure 9. The Nmn-Grignard reagent is configu-
2
1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
1
rationally rather stable at 0 °C, but epimerizes notably upon warming to 30 °C (t1/2 = 14 h).73
100
80
60
Nmn 70 °C
Men 70 °C
Nmn 50 °C
Nmn 50 °C
Nmn 30 °C
Nmn 30 °C
Nmn 0 °C
Men 0 °C
40
20
0
0
60
120
180
240
300
t [min]
I
Figure 9. Temperature-dependent epimerization of NmnMg(O CMen) (1b ) in THF solution. The 300
2
min data-point at 0 °C was extrapolated for graphical reasons. Curves are interpolated and serve as
guidelines only.
Clean first-order kinetics with t1/2 of 62 min (50 °C) and 8 min (70 °C) operate at higher temperature.
I
I
‡
‡
The activation energy is ≈ 100 kJ/mol. The rate 1b ®1a is characterized by DH = 98.5 kJ/mol and DS
‡
‡
I
I
= –113 J/mol·K by Eyring analysis (DH = 96.1 kJ/mol and DS = –110 J/mol·K for 1a ®1b ), with
calculated epimerization half-lives of 39 h at 22 °C, or 65 days at 0 °C.73
Scheme 6. Configurational stability of secondary alkyl-Grignard reagents.
ACS Paragon Plus Environment

Page 21 of 67
1
The Journal of Organic Chemistry
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a)
endo/exo = 59:41
MgBr
(or 48:52)
endo
MgBr
• epim.: t1/2 ≈ 5 h @ r.t.
exo
b)
endo/exo = 96:4
•
•
stable @ r.t.
stable: ≥12 h @ 67 °C
MgCl
MgCl
• epim.: <12 h @ 111 °C
exo
endo
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
c)
Nmn/Men = 58:42
MgX
• epim.: t1/2 = 39 h @ 22 °C
MgX
Men
ΔH‡ = 98.5 kJ/mol
ΔS‡ = −113 J/mol·K
Nmn
1 I
1aI
b
X = menthylcarboxylate
d)
•
stable ≤ −30 °C
MgCl
MgCl
• racem.: t1/2 = 5 h @ −10 °C
krac = 3.46 10-5 s-1
Ph
Ph
(
)
a) Norbornylmagnesium halides are known to slowly epimerize at rt. Reported data are faintly
74
conflicting. b) Bornylmagnesiumchloride forms as kinetic endo/exo 67:33 mixture, but changes to
9
6:4 upon heating.75,76 c) Menthylmagnesium compounds slowly epimerize at rt (this work). d)
Hoffmann's open-chain chiral Grignard reagent racemizes quickly at rt.79
7
7
Primary alkyl-Grignard reagents invert fast already below room temperature. Properties of secondary
alkyl-magnesium species are collected in Scheme 6. Norbornylmagnesium bromide equilibrates over a
day at room temperature (Scheme 6, a)74,78 whereas bornylmagnesium chloride is stable under the reflux-
conditions of its synthesis, but epimerizes if refluxed in the higher-boiling solvent toluene (Scheme 6,
b). Both the solvent and counter-ion differ between systems, making it difficult to assess the importance
I
of additional methyls on configurational stability. In any case, cyclic 1b (Scheme 6, c) is more inert than
Hoffmann's open-chain secondary alkyl Grignard reagent, which racemizes below 0 °C (Scheme 6, d).79
I
I
The 1a /b equilibrium ratios dr 38.3:61.7 (50 °C; after 144 h) and 39.5:60.5 (70 °C, after 1.5 h) deviate
markedly from those seen in the thermal equilibration of 1a/b (22 °C; dr 57.3:42.7 after 1 d, increased to
5
8.2:41.8 after 3 months; 50 °C, dr 55.6:44.4; 70 °C, dr 54.8:45.2). This is not contradictory, because the
I
I
major chemical species in 1 differ from those in carboxylated reagent 1a /b by the presence of chloride
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 22 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
vs. menthyl carboxylate as counter-ion (Scheme 5, b). Yet, the carboxylated reagent shows a surprising
I
prevalence of the axially substituted Nmn-diastereomer 1b , which may be a consequence of secondary
dimerization and solvation equilibria,80 because MgX usually favors the equatorial position in
cyclohexanes.81,82 Independent confirmation that the change of conformational preference is caused by
the carboxylate counter-ion was obtained by combining 1 with lithium salts of hindered carboxylic acids
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
and determining the composition of the in situ generated (Men/Nmn)Mg(O
2
CR) species after thermal
equilibration. With 2,4,6-trimethylbenzoate (mesitoate), the shift was from dr 57:43 (1a/b, rt) to 47:53
II
II
2
71
(
1a /1b , 60 °C; after 24 h). The recovery of RMgX (by quantitative H NMR ) was 98 %, which proves
that the change of composition is caused by isomerization and not kinetic ablation (e.g., by hydride
7
1
III
III
elimination) of one isomer. In another experiment with lithium pivalate, the resulting 1a /1b was
allowed to equilibrate at 70 °C and the reaction progress was followed by sampling and D O-quenching
2
(
Figure 10).
60
Me
Men [%]
Nmn [%]
5
6
Me
Mg
O Me
Men
O
52
III
β-(3-D )-1a
1
4
4
8
4
Me
Me
Me
O
Mg
Nmn
O
40
III
α-(3-D )-1b
0
15
30
45
t [min]
60
75
90
1
III III
Figure 10. Equilibration of in situ prepared MenMg(OPiv)/NmnMg(OPiv) (1a /b ) in THF solution
III
III
at 70 °C. A solution of 1 was mixed with LiOPiv and heated to 70 °C. Equilibrium ratio 1a /1b =
4
2.9:57.1 (after 69 h).
III
The initial excess of menthyl diastereomer 1a epimerizes with a half-life of 26 min (cf. 8 minutes for
I
I
III
II
1
a /1b ) to a new equilibrium at 42.9 % 1a , which is less than in case of mesitoate (46.8 % 1a ), but
ACS Paragon Plus Environment

Page 23 of 67
1
The Journal of Organic Chemistry
I
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
higher than for menthanecarboxylate (39.5 % of 1a ) as counter-ions. The influence of counter-ion X in
MenMgX on Men/Nmn-equilibrium composition and epimerization kinetics is substantial and
contributes to the complexity of reaction systems involving reagent 1 with other ionic components.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Stereochemistry and Relative Kinetics of Electrophilic Substitution at C-1 of Menthyl Grignard
Reagent
The result of the low-temperature carboxylation of 1 (Scheme 5, a) is expressed in tabular form in Table
2
, entry 1. Based on the known initial amount of 1a and 1b, analysis of remaining MenD (b-(3-D
and NmnD (a-(3-D )-4)) in the product mixture reveals how much of each diastereomer was converted
in a reaction. This quantity is expressed as 1a/D1b in mol % (–56/–3 mol %) and compared to the
quantity of epimeric substitution products formed (+58/+0 mol %; Table 2, last column). Since the
numbers are close, the conclusion can be reached that CO reacts with 1 by a retentive mechanism. This
1
)-4)
1
D
2
agrees with the observation of retention of configuration at the stereogenic carbon in the carboxylation
of Hoffmann's chiral Grignard reagent (cf. Scheme 6).79a Reactions of other H- and P-centered
electrophiles with Grignard reagent 1 were similarly investigated, and the results are collected in Table
2.
Table 2. Quantitative investigation of the reactions of Grignard reagent 1 with various
electrophiles.a,b
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 24 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Me
Me
Me
Me
a) E+
Me
Me
Me
b) D O
Me
2
Me
Me
Me
E
Me
+
D
Me
+
Me
E
retention
or inversion
ClMg
Me
D
retention
1
a/b
Men-E
Nmn-E
Men-D
Nmn-D
Entry RMgX (1a, 1b)b,c Td
E+e
[mol %]
Men-D
Nmn-D dr
R-Eg
(D1a)f
(D1b)f
[
mol %]
[°C]
(D
1
)-4
(mol %)
[
mol %]
[mol %]
39 (-3)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
100 (57, 42)
–78
CO
2
(xs)
1 (-56)
1:99
MenCO
2
H (19; 58 %);
NmnCO H (20; 0 %)
2
2
3
4
5
6
100 (57, 42)
100 (57, 42)
100 (57, 42)
100 (57, 42)
100 (57, 42)
–78 ® rt
–78 ® rt
–78 ® rt
–78 ® rt
H
2
2
O (25)
O (50)
36 (-21)
5 (-51)
41 (-1)
47:53
–
–
–
–
H
28 (-13) 16:84
32 (-9) 51:49
30 (-11) 52:48
n.d. n.d.
MeOH (50)
33 (-23)
MenOH (50) 32 (-25)
–78 ® rth PCl
(200)
n.d.
MenPO
H
(25a; 92 %);
(25b; 6 %)
3
2
2
NmnPO
2
H
2
7
8
8
300 (170, 127)
300 (170, 127)
300 (170, 127)
100 (57, 42)
0 ® 50
PCl
PCl
PCl
3
(100)
(100)
(100)
25 (-145) 34 (-93) 43:57
41 (-129) 49 (-78) 45:55
44 (-126) 48 (-79) 48:52
Men
Men
Men
2
POH (7; 61 %)
POH (7; 75 %)
POH (7; 73 %)
ai
bi
0
3
3
2
2
0
9
0 ® rt
Ph
2
PCl (100) 8 (-49)
21 (-21) 28:72
MenPOPh (27; 72 %)
2
a
+
General conditions: 1 was added to electrophile E in THF at –78 or 0 °C. After stirring for 1–3 h, reactions were
b
quenched with D
2
O and warmed to rt. Reactant quantities and analytical compositions are given in mol %, rounded to
c
the next nearest integer number. See the supporting information for more precise values expressed in µmol. The
composition of 1 varies from 1a/b = 57:43 to 58:42, with ca. 1 mol % of y-MenMgCl (12) present; the composition of
d
e
+
starting 1 (57:42:1 mol %) represents an average value. Room temperature (rt) is ca. 22 °C. E = electrophile, quantity
f
in mol % relative to 1 in brackets. Remaining quantity of 1a/b upon conclusion of the reaction, as determined by D
2
O-
2
1
quenching and H{ H} NMR quantification of (D
1
)-4. The number in brackets indicates change (loss) of organometallic
g
h
reagent over the course of the reaction. R-E = electrophilic substitution product. Reaction duration 3 h at –78 °C,
i
followed by warming to rt with stirring overnight. 8b is a repetition experiment of 8a.
1
3
2
Coinciding analytical results for the dr of Grignard reagent 1 from C NMR and H NMR analysis (after
O-quenching) prove that hydro-de-metalation of 1a/b proceeds with retention of configuration at C-1.
This is also demonstrated in the quenching of 1b with D O to give essentially pure a-(3-D )-4 (Figure
), which implies stereospecifity. Yet, protonation of 1a and 1b could still proceed at different rates.
D
2
2
1
9
Reaction of 1 with a limited amount of protons, followed by quenching with excess D O reveals the
2
ACS Paragon Plus Environment

Page 25 of 67
1
The Journal of Organic Chemistry
2
71
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
unreacted amount and dr of 1a/b by H NMR. In the reaction of 1 with 0.25 equiv of water, 1a is
selectively attacked while 1b initially remains (Table 2, entry 2). With twice the amount of water (0.5
equiv), 1b is also partially protonated, but at a lower rate than 1a, such that the remaining 1 is depleted
in the latter (Table 2, entry 3). Water is a more stereoselective acid in this reaction than alcohols are
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
entries 4, 5). Kinetic resolution by protonation (or: hydronation) is only relevant if the hydron source is
limiting, but will not play a role if excess D O is used to quench 1. Neither is a dynamic kinetic resolution
expected in view of the slow epimerization of 1a/b under quenching conditions (at rt or below).
Turning to phosphorus electrophiles, reagent 1 was allowed to react with excess PCl in the cold and the
resulting, sensitive menthyl dichlorophosphane (24) converted to phosphinic acid RPH(=O)(OH) (25)
2
3
8
3
by quenching with water (Table 2, entry 6). The high yield and diastereoselectivity (dr 15:1 ) prove that
the reaction proceeds with a high degree of stereoconvergence from both 1a and 1b to 24a. Thus, the
reported high yield of 90 % of 24a from 1,11a exceeding by far the amount of available epimer 1a in 1 is
confirmed by our experiment.
Combination of 1 with PCl in a 2:1 ratio, followed by heating to reflux is the standard procedure to
3
1
10
1
prepare Men
2
PCl (6). The reported yield is low at 13 % or 25–35 % . Analysis of the reaction mixture
is considerably simplified by quenching with D
7 (Scheme 7). The usual procedure (mixing at 0 °C, then heating) followed by protic, aqueous workup
gave a fair spectral yield of 7 (Table 2, entry 7). The remaining 1 (dr 43:57; analyzed as (3-D )-4)
2
O (1®[3-D ]-4) and hydrolysis of 6 to phosphine oxide
1
1
displayed less enrichment than expected for a kinetic resolution, although this ratio might have been
affected by epimerization of 1a/b upon heating. Performing the reaction at 0 °C throughout further
increased the yield of 7, without affecting the dr of unreacted Grignard reagent (Table 2, entries 8a/b).
The results show that both substitution events with nucleophile 1 along PCl
3
® MenPCl (24) ®
2
Men PCl (6) proceed stereoconvergently with respect to C-1 of 1, with both diastereomers 1a and 1b
2
contributing substantially to produce 6, with a slight kinetic preference for 1a in reactions with P–Cl
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 26 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
electrophiles. Even though the generation of 7 was limited to 75 mol-%, and unreacted 1 was still present,
no major organophosphorus species was detected in the organic phase. Failure to detect hydrolysis
product 25 (MenPO
2
H
2
) means that both 24 (and thus also the more reactive PCl ) must have been spent
3
at the conclusion of the reaction. The unaccounted loss of phosphorus in entries 7–9 is approximately
half the molar quantity of unaccounted loss of RMgX, which suggests that redox processes accom-
panying the main reaction lead to reduced, water-soluble phosphorus species and neutral hydrocarbons
as side-products.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 7. Reaction of 1 with phosphorus electrophiles.
a)
PCl3
THF
Me
Me
Me
1
a/b
1
a/b
PCl2
P
THF
Me
Me
Cl
b)
Me
Me
24a
Me
Me
Ph PCl
2
6
H O
2
Me
Me
Me
Me
Me
O2
air
H
P
Ph
Ph
PPh2
Me
P
O
O
Me
Me
Me
Me
Me
Me
2
6
27
7
The reaction of 1 with Ph
stated that quenching the reaction of equimolar 1 and Ph
original 1 (i.e., 1a) converted to MenPPh , the other half (i.e., 1b) returning NmnD (a-[3-D
2
PCl2,13,14 was also analyzed (Table 2, entry 9). A previous, qualitative report
PCl with D O gave a mixture with half the
]-4). In our
2
2
2
1
hands, the analytical yield of oxide 27, which emerged from phosphane 26 during workup in air, exceeds
the initially present quantity of 1a (Table 2, entry 9; Scheme 7). Even though the kinetic preference for
1
a is marked, roughly half of epimer 1b must have reacted with inversion of configuration at C-1 to
account for the observed yield (cf. D1a/b in Table 2, entry 9), which fits the pattern observed with PCl
®24), and which agrees with the stereoconvergent reactions of bornyl- and fenchylmagnesium reagents
3
(
ACS Paragon Plus Environment

Page 27 of 67
1
The Journal of Organic Chemistry
with P-Cl electrophiles.75a Phosphine oxide 27 was isolated and structurally characterized by X-ray
crystallography (Figure 11). The conformation of the menthyl group is consistent with earlier
observations (Figure 3).
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 11. Solid-state X-ray molecular structure of dimenthylphenylphosphine P-oxide (27). Ellipsoids
are shown at 50 % probability. Hydrogens are omitted for clarity. Oxygen is red, phosphorus purple.
Synthesis and Properties of Dimenthylphosphine Oxide
In the preparative study of the reaction of PCl
3
with of 1 to give Men PCl (6), we were initially guided
2
by the hypothesis that 4 equivalents of 1, corresponding to approximately two equivalents of 1a, might
provide a higher yield, if 1a is the exclusive nucleophile.2,75b Yet, variation of the ratio PCl
3
/1 over 1:1,
1
:3, 1:4 and 1:6 indicated that already 1:3 was optimal, and use of additional Grignard reagent had a
counter-productive effect. Hydrolysis of the 1:3 reaction mixture (containing 6) gave the expected, previ-
30
ously unknown dimenthylphosphine P-oxide (7; Table 2, entries 7, 8, and Scheme 7) that was readily
1
quantified by H NMR in crude reaction mixtures. Chromatography required acetic acid as additive, with
which 7 forms a hydrogen-bonded complex that must be split with base prior to solvent evaporation. The
product was obtained as slowly crystallizing oil in 50% yield. Spectral properties agree with a C -
1
1
13
symmetric structure, with two signal sets for diastereotopic menthyl groups ( H, C NMR), a single
3
1
1
P{ H} NMR resonance (d
P
41.8), and a characteristic doublet of doublets at d 6.77 with the large
H
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 28 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
31
1
J( P, H) of 435.5 Hz for the central P(O)H-unit. The identity of this material as 7 was further secured
by single crystal X-ray crystallographic analysis (Figure 12).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 12. Solid state molecular structure of dimenthylphosphine P-oxide (7). Oxygen is colored red,
phosphorus purple. Ellipsoids are shown at 50 % probability. The unit cell contains three independent
molecules of very similar geometry, of which one is displayed.
Based on the 2D NMR analysis (Figure S29, Table S25), the diastereotopic menthyl groups (MenRe,
Si
3
Re
3
Men ) differ by a JP,H coupling (5.6 Hz) from H-P to H-1 in Men , which is absent ( JP,H ≈ 0 Hz) for
Si
H-1 in Men . This observation is consistent with the solid-state conformation in the X-ray structure
Re
Si
having torsional angles of –169° (to H-1 in Men ) or –50° (to H-1 in Men ). The P=O oxygen is involved
Re
Si
in short intramolecular contacts with H-1' of the isopropyl group in Men (237 pm), and H-2 of Men
(283 pm), and those interactions are reflected by higher frequency d signals for the respective nuclei,
H
compared to their diastereotopic counterparts. The usual conformation of C-2-isopropyl in menthyl
derivatives is observed (Figure 3, a).
CONCLUSIONS
ACS Paragon Plus Environment

Page 29 of 67
1
The Journal of Organic Chemistry
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Menthyl Grignard reagent (1) has been known and used in asymmetric synthesis for many years. Yet, its
exact chemical composition and reaction stereochemistry have remained insecure in the absence of direct
analytical studies. In the reaction sequence starting from 2 via 1 and leading to various substitution
products (Scheme 1), the stereochemistry of the first and second steps were not known with certainty,
and neither was the configurative stability of 1a/b established. Methods to analyze the composition 1a/b
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
had relied on reactions with electrophiles (CO
kinetic resolution effects. Even for D O-quenching, the assumption that the dr of (3-D
of 1a/b (i.e., retention of configuration in proto-de-metalation) was unproved.
In the present study we have for the first time analyzed menthyl Grignard reagent in situ by combining
, D
2 2
O), the results of which could have been affected by
2
1
)-4 reflects the dr
1
13
H and C NMR methods. MenMgX (1a') and NmnMgX (1b') were confirmed as major organometallic
components of the reagent, besides hydrocarbons trans-p-menthane (4), 2-menthene (8), 3-menthene (9)
and three diastereomeric Wurz products 13–15. The Schlenk equilibrium is operative and produces
considerable amounts of Men Mg from 1a at low halide concentration, whereas 1b is less notably
2
involved in equilibria. The Schlenk equilibrium complicates quantitative NMR analysis, since minor and
13
dynamic species display exchange-broadened signals. The intensity of the characteristic C NMR signals
C
d 49.5 and 52.8 for the major components 1a' and 1b' do not reflect the dr 1a:1b. Use of dibromoethane
or other halogenated reagents for initiating the Grignard reaction has the beneficial effect of shifting the
Schlenk equilibrium towards 1a'/b' and giving easier to analyze NMR spectra. However, integration of
1
3
defined ranges of the C NMR spectrum (d 48–51 for C-2 of 1a, d 51–54 for C-2 of 1b) led to consistent
results for the 1a/b-dr that agreed with the D O quenching results and confirm retention of configuration
in deuterolysis of the C–Mg bond. Based on this benchmarking experiment, the technically simple and
2
2
1
fast H{ H} NMR analysis of 1a/b after D
2
O-quenching can be carried out with confidence. Besides
)-4 signals (d 0.94 and 1.65), analysis in presence of a
analyzing the peak area ratio of the a/b-(3-D
1
D
ACS Paragon Plus Environment