Telescoped approach to aryl hydroxymethylation in the synthesis of a key pharmaceutical intermediate

Article abstract of DOI:10.1021/jo400647t

An efficient synthetic approach leading to introduction of the hydroxymethyl group to an aryl moiety via combination of the Bouveault formylation and hydride reduction has been optimized using a rational, mechanistic-based approach. This approach enabled

Full text of DOI:10.1021/jo400647t

Article
pubs.acs.org/joc
Telescoped Approach to Aryl Hydroxymethylation in the Synthesis
of a Key Pharmaceutical Intermediate
Catherine N. Slattery, Rebecca E. Deasy, and Anita R. Maguire*
Department of Chemistry and School of Pharmacy, Analytical and Biological Chemistry Research Facility, University College Cork,
Cork, Ireland
Michael E. Kopach, Utpal K. Singh, Mark D. Argentine, William G. Trankle, and Roger Brian Scherer
Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, Indiana 46285, United States
Humphrey Moynihan
Eli Lilly SA, Dunderrow, Kinsale, Cork, Ireland
S
* Supporting Information
ABSTRACT: An efficient synthetic approach leading to introduction of the hydroxymethyl group to an aryl moiety via
combination of the Bouveault formylation and hydride reduction has been optimized using a rational, mechanistic-based
approach. This approach enabled telescoping of the two steps into a single efficient process, readily amenable to scaleup.
be accomplished by “cracking” paraformaldehyde or adding
solid paraformaldehyde to a preformed Grignard reagent.
However, the generation of anhydrous formaldehyde is often
extremely hazardous, as thermal runaway reactions can occur;
paraformaldehyde “unzipping” is very difficult to control and is
highly energetic. In addition, yields for formaldehyde-based
approaches are frequently very low with high impurity levels,
and quenching of excess paraformaldehyde is often problem-
atic. Hydroxymethylation can also be achieved by hydrolysis of
benzyl halides produced from a halomethylation reaction.3 A
significant downside to the halomethylation approach is that
benzyl halides are frequently strong lachrymators and the
halomethylation step produces bis(chloromethyl) ether, which
is a highly dangerous byproduct on all scales of operation. A
potentially attractive approach is Bouveault formylation, where
a dialkyl formamide such as DMF is reacted with a Grignard
reagent followed by an immediate reduction to produce the
target hydroxymethyl alcohol.4 However, due to side reactions,
the Bouveault formylation has not been generally useful on
preparative scales. To avoid common formylation side reactions
such as secondary nucleophilic addition, Meyers and Comins
developed 2-(N-methylformylamino)pyridine5 as an efficient
formylating reagent, which has been hypothesized to form a
INTRODUCTION
■
Hydroxymethylation is an important transformation in organic
synthesis and is often used to produce building blocks for more
complex targets. Edivoxetine·HCl is a highly selective
norepinephrine uptake inhibitor under development at Eli
Lilly and Co. for the treatment of depression (Scheme 1).1 An
Scheme 1. Edivoxetine·HCl Retrosynthesis
important intermediate for the synthesis of edivoxetine·HCl is
(5-fluoro-2-methoxyphenyl)methanol (1). A key requirement
in the edivoxetine·HCl synthesis is high regiochemical purity,
and as such, a route to produce 1 with high isomeric purity was
required.
The most common hydroxylation approach is reaction of a
Grignard reagent directly with formaldehyde or paraformalde-
hyde to produce the desired hydroxymethyl alcohol.2 This can
Received: March 27, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jo400647t | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
tight chelate, preventing release of the aldehyde under the
reaction conditions (Scheme 2).
preparation of 1, investigation of Grignard exchange chemistry
for the production of 3, and the development of a highly
efficient tandem acylation/hydride reduction amenable to use
on any scale for safe production of 1.
Scheme 2. Meyers and Bouveault Aldehyde Synthesis
RESULTS AND DISCUSSION
■
Two-Step Synthesis of Alcohol 1. The two-step synthesis
of (5-fluoro-2-methoxyphenyl)methanol (1) from 2-bromo-4-
fluoroanisole (2) was initially investigated on a 5 g scale using
2-methyltetrahydrofuran (2-MeTHF), tetrahydrofuran (THF),
and cyclopentyl methyl ether (CPME) as reaction solvents
(Table 1, entries 1−3). THF and 2-MeTHF were of particular
interest, because these solvents can be manufactured directly
from the renewable resource furfural.8 In this study, the
Grignard reagent was prepared in the selected solvent system
using catalytic DIBAL-H or iodine as an activator with a 12%
initiation charge of substrate.9 The acylation and reduction
steps were then carried out in the same solvent systems. For
each of the solvents examined, the aldehyde intermediate 4 and
alcohol product 1 were produced with good purity and in
generally high yields, albeit with a lower yield recorded for the
reaction conducted in THF due to soluble loss of product to
the aqueous layer during workup (Table 1, entry 2). While the
decreased yield recorded for the reaction in THF was a
concern, the purity in this case was very high and importantly,
in contrast to the syntheses employing 2-MeTHF and CPME,
the THF reaction proceeded without any significant solid
formation during the formylation and reduction steps, which
could be particularly advantageous for larger-scale reactions.
The preparation of 5-fluoro-2-methoxybenzaldehyde (4)
using N-formylmorpholine and N-formylpiperidine in place of
DMF was next examined. Reactions with N-formylmorpholine
were conducted in 2-MeTHF, THF, and CPME (Table 1,
entries 4−6). For each solvent system investigated, a reduction
in yield and product purity was recorded in comparison to
previous reactions employing DMF (Table 1, entries 4 vs 1, 5
vs 2, and 6 vs 3). In particular, a dramatic decrease in efficiency
was recorded for the synthesis of 4 in CPME, and therefore,
this solvent was not evaluated further in this study. Reduced
We envisioned that it may be possible to directly formylate
with DMF under Bouveault conditions using the Grignard
reagent 3 produced from commercially available 2-bromo-4-
fluoroanisole (2) to provide the aldehyde intermediate 4
(Scheme 3). We surmised that chelation of magnesium to the
Scheme 3. Proposed Synthesis of Aldehyde 4
α-methoxy group under Bouveault conditions would afford
benefits similar to those reported by Meyers for the N-pyridyl
system. Other formylating reagents such as N-formylmorpho-
line and N-formylpiperidine were also of interest, which may
afford similar results for the production of 1.6
In addition, Meyers further extended the acylation method-
ology by demonstrating that Grignard aldehydic intermediates
could be trapped in a controlled fashion with a second
nucleophile to produce tertiary alcohols.7 Meyer’s methodology
was of interest, as we sought to produce the greenest possible
synthesis of 1, which may require production of several
multiton quantities in the future. Herein we report the results
of our studies toward the synthesis of the pharmaceutical
intermediate 1 employing a combination of the Bouveault
formylation and hydride reduction reactions. This study
includes an examination of a two-step approach for the
Table 1. Two-Step Synthesis of 1
4
1
a
b
c
b
entry
solvent
formylating agent
DMF
formylation temp (°C)
crude yield (%)
purity (%)
crude yield (%)
purity (%)
1
2
2-MeTHF
THF
45
100.6
92.4
95.1
91.4
87.3
73.7
92.3
75.7
105.4
91.6
89.8
96.0
94.4
86.8
82.5
62.6
93.8
83.2
85.8
90.1
97.8
78.8
96.4
92.2
98.8
91.2
DMF
45
3
CPME
2-MeTHF
THF
DMF
45
4
N-formylmorpholine
N-formylmorpholine
N-formylmorpholine
N-formylpiperidine
N-formylpiperidine
DMF
45
5
45
81.7
89.3
6
CPME
2-MeTHF
THF
45
7
45
8
45
73.4
78.2
90.3
98.9
9
2-MeTHF
THF
room temp
room temp
10
DMF
a
b
c
Mass yield recovered for synthesis of aldehyde 4. As determined by GC-MS analysis. Mass yield recovered for synthesis of alcohol 1.
B
dx.doi.org/10.1021/jo400647t | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Table 2. Tandem Acylation/Reduction Strategy
a
b
c
entry
solvent
formylating agent
DMF
quench
crude yield (%)
purity of 1 (%)
impurity (amt (%))
1
2
3
4
5
6
2-MeTHF
THF
after 1 h
after 1 h
after 1 h
after 1 h
immediate
immediate
98.5
93.1
92.8
95.1
85.3
90.6
96.8
90.1
89.3
98.9
95.0
5a (2.9)
5a (0.3)
5b (4.9)
5c (8.9)
5a (0.3)
DMF
THF
N-formylmorpholine
N-formylpiperidine
DMF
THF
THF
2-MeTHF
DMF
100.1
5a (0.2)
a
b
c
Acetic acid quench performed either after 1 h or immediately after addition of LiBH4. Mass yield recovered after workup procedures. As
determined by GC-MS analysis.
purity levels and product recovery were also recorded for the
reaction examining N-formylpiperidine in THF (Table 1, entry
8 vs 2). Interestingly, a slight increase in the purity of 4 was
observed for the experiment employing N-formylpiperidine in
2-MeTHF (Table 1, entry 7 vs 1); however, this reaction was
not further explored due to the significant amount of solid
formation observed during formylation and difficulties in
removing excess N-formylpiperidine from the reaction mixture.
In a previous report by Olah and co-workers investigating the
preparation of aldehydes and ketones from N,N-dialkylamides
and Grignard reagents,10 it had been found that reactions must
be conducted at low temperatures (0−20 °C) to avoid the
occurrence of competing secondary reactions. In order to
investigate if such a limit was also critical for product purity in
our system under study, two experiments were conducted
examining a reduced temperature during formylation (Table 1,
entries 9 and 10). For both reactions examined a decrease in
the purity of aldehyde 4 was recorded (Table 1, entries 9 vs 1
and 10 vs 2), and as a result no change to the standard 45 °C
temperature for formylation was made for later scaleup studies.
Tandem Acylation/Reduction Strategy. It was sub-
sequently surmised that we may be able to telescope the
process under investigation by a tandem formylation/hydride
reduction strategy. There are limited reports on such an
approach, but the potential upside was large, including
elimination of the problematic aldehyde workup. For these
reasons we investigated the tandem addition strategy in both
THF and 2-MeTHF using the standard conditions previously
described for the two-step synthesis of 1, but without isolation
of the aldehyde intermediate 4, where lithium borohydride was
added to the unquenched tetrahedral intermediate.
The telescoped process was subsequently tested with N-
formylmorpholine and N-formylpiperidine as formylating
agents in THF (Table 2, entries 3 and 4).11 Once again,
successful formation of 1 was recorded in good yield; however,
a decrease in product purity was observed with enrichment of
reductive amination byproducts 5b,c. The impurities 5a−c
appear to form mainly during the quench by reaction of the
liberated secondary amine byproduct with the forming
aldehyde. For the DMF system, the dimethylamine byproduct
is volatile and can escape through the headspace at higher
temperatures (45 °C). However, for the N-formylmorpholine
and piperidine systems the corresponding morpholine and
piperidine byproducts are not volatile; hence, formation of the
reductive amination byproducts is not prevented and 5b,c are
thus observed at increased levels in the crude product mixture
(Table 2, entries 3 and 4 vs entry 2).
In an effort to further probe the reaction mechanism for the
one-pot process, two experiments were conducted examining
an immediate quench of the reaction mixture following
dropwise addition of the reducing agent (Table 2, entries 5
and 6). Interestingly, full consumption of the aldehyde
intermediate 4 was again recorded under the modified reaction
conditions. These observations combined with the lack of
exotherm when lithium borohydride is added to the tetrahedral
intermediate and no indication of hydrogen offgassing all
provide compelling evidence that reduction is occurring entirely
during the quench step.20
Grignard Exchange Chemistry. A potentially attractive
Grignard processing option which has emerged over recent
years is halogen−magnesium exchange chemistry, where
commercially available simple Grignard reagents undergo an
exchange reaction with an aryl halide to form a Grignard
complex.12 This approach has the benefit of directly avoiding
handling magnesium metal and minimizes the propensity for
runaway reactions. Knochel and co-worker discovered that the
use of the salt additive LiCl accelerated both the rate and the
efficiency of the reaction;13 thus, in this study both
commercially available i-PrMgCl and i-PrMgCl·LiCl exchanges
were explored.
We were pleased to discover that the proposed one-pot
formylation/reduction could be successfully conducted, provid-
ing 1 in 98.5% (90.6% pure) and 93.1% (96.8% pure) yields for
reactions in 2-MeTHF and THF, respectively (Table 2, entries
1 and 2). Importantly, purity levels for the isolated alcohol
product 1 was in the range of those previously observed for the
original two-step syntheses of 1 (Table 1, entries 1 and 2),
although the formation of the reductive amination byproduct
(5a) was observed.
Initial studies in this area focused on 2-MeTHF as reaction
solvent. At the outset low temperature (−10 °C and room
C
dx.doi.org/10.1021/jo400647t | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Table 3. Grignard Exchange Synthesis of 4
a
b
c
d
d
entry
Grignard reagent
solvent
temp (°C)
Grignard time (h)
formylating agent
crude yield (%)
4 (%)
6 (%)
e
1
2
3
4
5
6
7
8
i-PrMgCl
i-PrMgCl
i-PrMgCl·LiCl
i-PrMgCl
i-PrMgCl
i-PrMgCl
i-PrMgCl
i-PrMgCl
2-MeTHF
THF
reflux
reflux
reflux
reflux
reflux
30
2
1
1
1
1
4
4
4
DMF
DMF
DMF
101.4
103.5
102.0
99.8
71.6
73.7
78.3
90.2
76.5
98.8
96.7
98.5
0
20.8
18.5
8.2
THF
THF
N-formylmorpholine
N-formylpiperidine
DMF
THF
84.2
19.0
0.4
THF
93.2
f
THF
30
DMF
96.0
0.7
g
THF
30
DMF
96.7
0.2
a
b
Reaction time required for >99% Grignard exchange as determined by GC-MS analysis. Formylation conducted at 45 °C with 3 equiv of
c
d
e
formylating agent unless otherwise stated. Mass yield recovered after workup procedures. As determined by GC-MS analysis. Isolated product
contains 13.1% 2-bromo-4-fluoroanisole (2). Experiments were conducted examining increased amounts of i-PrMgCl in 2-MeTHF and longer reflux
f
g
periods,; however, for all reactions tested conversion to Grignard 3 was found to stall at ∼90%. Formylation was conducted at 30 °C. Two
equivalents of DMF was used.
temperature) Grignard exchange was investigated as described
by Leazer and co-workers;14 however, no Grignard exchange
was recorded, with 100% starting material 2 detected by GC-
MS analysis after 1 h. Consequently, reflux conditions were
implemented with successful Grignard exchange achieved after
2 h (Table 3, entry 1). As previously discussed in both the two-
step and one-pot syntheses in 2-MeTHF, significant solid
formation was observed upon charging DMF to the organo-
magnesium reagent 3, resulting in complicated workup, and
therefore this solvent was not evaluated further.
The next series of experiments employed THF as reaction
solvent. Complete Grignard exchange was successfully attained
with both i-PrMgCl and i-PrMgCl·LiCl after 1 h reflux (Table
3, entries 2 and 3). Negligible rate enhancement was observed
with i-PrMgCl·LiCl, and therefore this reagent was no longer
explored. Favorably, no solid formation was observed upon
addition of DMF to 3 in THF; however, purity levels for the
isolated aldehyde 4 were poor (73.7 and 78.2%, respectively)
with high levels of impurity 6 (20.8 and 18.5%, respectively)
detected by GC-MS analysis. Significantly, impurity 6 was also
observed with N-formylmorpholine and N-formylpiperidine
(Table 3, entries 4 and 5), albeit in lower levels for the latter
formylating agent, and thus the formation of 6 was independent
of the formylating agent employed. Furthermore, impurity 6
was not observed when 2-MeTHF was utilized (Table 3, entry
1), and thus a Grignard-based reaction incorporating ring-
opened THF with addition of the isopropyl moiety has been
postulated for its formation. This side reaction is likely radical
based and is precedented on the basis of the work of Hock and
co-workers.15
extended to the large-scale preparation of the Grignard reagent
3. The effect of temperature at the formylation stage was also
investigated and, as was observed for the two-step synthesis of
1, was found to have minimal effect on product purity (Table 3,
entry 7 vs 6). In an attempt to further optimize this procedure,
one experiment was conducted examining a reduced loading of
DMF during the formylation step. In initial Grignard exchange
studies (Table 3, entries 1−7), 3 equiv of DMF was employed
in line with literature procedures; however, the use of just 2
equiv of this reagent was found to be sufficient to permit full
conversion to aldehyde 4 with miminal impact on purity and
yield recorded (Table 3, entry 8).
Process Impurities. The principal process impurity
observed was 4-fluoroanisole (7) (Figure 1), which was
Figure 1. Impurities identified in the preparation of 1.
Temperature effects for the formation of impurity 6 were
subsequently explored, as the higher than anticipated impurity
level needed to be addressed prior to large-scale synthesis of 1
by Grignard exchange. When the halogen−magnesium
exchange reaction was performed at 30 °C, dramatically
reduced levels of impurity 6 (0.4%) were observed by GC-
MS analysis, with excellent purity (98.8%) of the isolated
aldehyde 4 (Table 3, entry 4). While a longer reaction time was
required for complete Grignard exchange at 30 °C relative to
reflux (4 h vs 1 h), the mild reaction conditions can be readily
found in varying amounts in all isolated crude samples of 1,
due presumably to the presence of adventitious water in the
reaction mixture or incomplete formylation. Additional
common impurities also observed by GC-MS analysis included
the Wurtz byproduct 8, the phenol byproduct 9, and the bis-
addition byproduct 10. The phenolic byproduct 9 had the
highest variability and was likely produced by hydrolysis of a
peroxide impurity generated from the reaction of the Grignard
reagent 3 with oxygen. As already discussed, impurity 6 was
observed exclusively during investigation of the Grignard
D
dx.doi.org/10.1021/jo400647t | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
exchange reactions. Byproduct 11 was also recorded at low
levels during the Grignard exchange study (≤0.3%) for
Grignard reactions conducted at 30 °C. 4-Fluoroanisole (7)
and 5-fluoro-2-methoxyphenol (9) were commercially available
and were analyzed by our GC-MS method to confirm the
retention times of these impurities. Compounds 5a−c, 6, 8, 10,
and 11 were isolated from the crude product mixtures by flash
chromatography and characterized to assign their structures.
Scaleup of the One-Pot Procedure. The synthesis of 1
was scaled to a 15 g scale of 2-bromo-4-fluoroanisole (2) using
the optimum conditions identified for the two-step, one-pot,
and Grignard exchange reactions. For this purpose, THF was
employed as the reaction solvent, representing the best choice
in terms of product purity and minimal solid formation;
however, loss of product to the aqueous layer during workup
procedures was a concern. The two-step synthesis of 1 was first
examined using DMF as formylating agent and with
formylation conducted at 45 °C (method a). Under these
conditions, the aldehyde intermediate and alcohol product were
obtained in 87.1% and 82.6% yields, respectively, giving an
overall yield of 72.3% for the two-step process (Table 4, entry
previously identified in the small-scale studies. The crude
yield recorded in this instance was found to be significantly
higher than that obtained for the alternative two-step
preparation of 1 (Table 4, entry 3 vs entry 1), due presumably
to the requirement for only one workup step in method b, thus
minimizing product loss to the aqueous layer.16
The scaled-up two-step synthesis of alcohol 1 employing
Grignard exchange chemistry for the formation of 3 was
conducted using 2 equiv of i-PrMgCl and DMF in THF
(method c), conditions deemed optimal from our previous
small-scale studies. Critically, the reaction temperature was set
at 30 °C during the Grignard formation step to avoid the
formation of significant quantities of byproducts 6 and 11.
Under these conditions, the desired product 1 was obtained in
high yield and excellent purity, with an increase in product
purity again recorded following crystallization (Table 4, entry 4,
see Figure 2 for impurities in crude product mixture).
Further Optimization of the One-Pot Procedure. The
long-term aim of this project was the preparation of (5-fluoro-
2-methoxyphenyl)methanol (1) on a manufacturing scale. For
this purpose, the one-pot synthesis of 1 was identified as the
most suitable procedure, representing the best choice in terms
of minimum reaction steps, solvent usage, and overall product
yield, although notably product purity (94.2%) was slightly
lower for this method. As a result, a number of changes to the
original one-pot procedure (method b) were implemented
prior to manufacture in an attempt to decrease impurity levels
(the revised procedure is method d). First, the solvent for the
Grignard formation step was changed from THF to the more
environmentally benign solvent 2-MeTHF, which facilitated
downstream improved aqueous phase separations. THF was
still used for the formylation step, as significant solid formation
was observed at this stage in previous small-scale experiments
with other solvents. An altered workup procedure involving a
methanol/water quench prior to addition of the acetic acid and
an extended base wash period was also adopted to try to reduce
the byproduct 5a. As shown in Table 5, the optimized one-pot
Table 4. Large-Scale Experiments
crude
after crystallization
yield
a
purity of
purity of
b
yield
(%)
b
entry procedure
(%)
1 (%)
1 (%)
1
2
3
4
method a
method a
72.3
84.1
95.7
94.4
97.0
98.8
94.2
97.3
98.9
99.4
d
56.7
72.4
d
c
method b
method c
99.4
80.5
a
Mass yield recovered following reduction of 4 and after workup
b
c
procedures. As determined by GC-MS analysis. 2-MeTHF was
added to the reaction mixture during workup procedures to minimize
product loss to the aqueous layer. Successful crystallization was not
d
achieved due to the presence of suspected borohydride-derived
byproducts.16
Table 5. Optimization of the One-Pot Procedure
1). This yield could be increased to 84.1% by addition of 2-
MeTHF to the reaction mixture during the two workup steps,
thus minimizing loss of both 4 and 1 to the aqueous layer
(Table 4, entry 2). Significantly, the crude alcohol product was
isolated in excellent purity for both large-scale experiments
conducted using method a, with a purity of ≥98.9% achieved
following crystallization (see Figure 2 for impurities in the
crude product mixture).
b
amt (%)
Scaleup of the one-pot synthesis of 1 was next examined,
again using the optimal reaction conditions (method b)
a
entry procedure 2 (g) yield (%)
1
5a
9
10
1
2
3
method b
method d
method e
15
15
95.7
96.0
95.0
94.2
94.9
97.1
0.16
0.21
0.1
0.27
0.18
2.48
1.77
0.2
c
d
110
<0.1
a
Mass yield recovered following reduction of 4 and after workup
b
c
procedures. As determined by GC-MS analysis. For process safety
data for method e, see the Supporting Information. Average yield.
d
procedure (method d) provided the desired alcohol product 1
with a slightly increased yield and purity relative to the previous
procedure (method b). There was however no decrease in the
amount of 5a present in the crude reaction mixture.
The Commercial Procedure (Method e). With the
optimal procedure (method d) now in hand, the commercial
production of (5-fluoro-2-methoxyphenyl)methanol (1) was
next envisioned. Prior to commencement of the commercial
Figure 2. Comparison of GC-MS impurities in the crude product
mixture for reactions conducted using methods a−c.
E
dx.doi.org/10.1021/jo400647t | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
conditions: oven temperature program from 45 to 250 °C at 10 °C/
min, and the final temperature kept for 8.5 min; injector temperature
200 °C, split injection technique (25:1 split ratio); carrier gas
hydrogen, flow rate 1.0 mL/min; diluent used was toluene; ionization
energy 69.9 eV, in the electronic ionization (EI) mode; ion source
temperature 200 °C and scan mass range of m/z 50−500. HPLC
analysis was carried out using a Zorbax SB-C8 (25 cm × 4.6 mm × 5
mm) column under the following conditions: mobile phase A, 0.1%
H3PO4 in H2O; mobile phase B, acetonitrile; flow rate 1.0 mL/min;
gradient 0 min 10% B, 10 min 90% B, 16 min 90% B; wavelength 250
nm; ambient temperature.
campaign two additional changes were implemented to address
issues of process safety at the increased scale and impurity
levels, respectively. The changes were as follows. (1) The
Grignard initiation was performed at 10 °C, maintaining less
than 30 °C throughout the Grignard reagent formation. (2) In
contrast to previous experiments (one-pot procedure), the
Grignard reagent was transferred with magnesium sequestration
to a one volume THF solution containing 1.2 equiv of DMF.
This approach was envisioned to minimize the formation of
dimer 10 and allows for the use of a magnesium heel in the
Grignard formation step on a manufacturing scale. More
importantly, it was anticipated that this modified process could
be safely operated at all scales. As shown in Table 5, entry 3, the
modified commercial procedure (method e) was successfully
conducted on a 110 g scale of 2-bromo-4-fluoroanisole (2),
providing the desired alcohol product 1 in high yield (95.0%)
and importantly with high levels of purity (97.1%).
Significantly, alteration of the order of addition during the
formylation step resulted in a large decrease in the amount of
dimer 10 present in the crude product mixture (<0.2%). Other
standard process impurities were controlled well by method e;
however, four additional minor process impurities were
observed due to a less pure source of the 2-bromo-4-
fluoroanisole raw material which was used for the scaleup.17
As previously described (Table 4), the purity of the isolated
product 1 was increased to >99% following crystallization with
toluene/heptane (overall yield 88.0%).
Two-Step Synthesis (Method a). Magnesium (2.10 g, 86.0
mmol) was suspended in THF (36.0 mL), and diisobutylaluminum
hydride (1 M solution in THF, 1.83 mL, 1.8 mmol) was added at 30
°C. 2-Bromo-4-fluoroanisole (2; 1.14 mL, 8.8 mmol) was added
dropwise, and the resulting mixture was stirred for 0.5 h. The
temperature was adjusted to 20 °C, and then 2-bromo-4-fluoroanisole
(8.35 mL, 64.4 mmol) diluted with THF (36.0 mL) was added over
0.5 h, maintaining a reaction temperature of <40 °C. When addition
was complete, the reaction mixture was stirred for 1 h, after which time
full Grignard reagent formation was confirmed by GC-MS analysis. A
solution of anhydrous DMF (6.91 mL, 89.0 mmol) in THF (36.0 mL)
was added dropwise over 0.5 h to the Grignard mixture at 45 °C. The
resulting mixture was stirred at 45−55 °C for 1 h and then quenched
by slow addition via cannula transfer to aqueous acetic acid (20 wt %,
105 mL) at 0 °C. 2-MeTHF (50 mL) was added to the reaction
mixture, and the phases were separated. The organic phase was washed
with aqueous sodium carbonate (5 wt %, 183 mL). The aqueous phase
was discarded, and the organic layer was concentrated by rotary
evaporation to give the crude aldehyde intermediate 4 as an orange
solid. A 10 wt % solution of 4 in THF was adjusted to 23 °C, lithium
borohydride (4 M solution in THF, 9.15 mL, 36.6 mmol) was added
dropwise, and the mixture was stirred for 1 h at this temperature. The
reaction mixture was subsequently transferred via cannula to an
aqueous solution of acetic acid (20 wt %, 105 mL) and stirred for 0.5
h. 2-MeTHF (50 mL) was added to the reaction mixture, and the
phases were separated. The organic phase was washed with aqueous
potassium hydroxide (10 wt %, 100 mL), and then the aqueous layers
were separated and discarded. The solvent was removed by rotary
evaporation to obtain crude (5-fluoro-2-methoxyphenyl)methanol (1;
9.60 g, 84.1%) as a yellow oil. The crude product was purified by
crystallization in toluene/heptane as follows. The crude alcohol 1
(9.60 g) was dissolved in toluene (12.4 mL), and the solution was
cooled to 0 °C. Heptane (15.7 mL) was added dropwise to the stirred
solution, and after addition of 3 mL of heptane the product 1 started
to crystallize without seeding. The remaining heptane was added over
20 min. Additional heptane (56.0 mL) was added slowly over 45 min,
and the mixture was stirred for 1.5 h. The suspension was filtered, and
the solid was washed with heptane (40.0 mL). The wet cake was dried
under reduced pressure at room temperature to obtain pure 5-fluoro-
2-methoxyphenyl)methanol (1; 8.28 g, 86.3% recovery from crude) as
a white crystalline solid with >99% purity (overall yield 8.28 g, 72.4%).
One-Pot Synthesis (Method b). Magnesium (2.10 g, 86.0 mmol)
was suspended in THF (36.0 mL), and diisobutylaluminum hydride (1
M solution in THF, 1.83 mL, 1.8 mmol) was added at 30 °C. 2-
Bromo-4-fluoroanisole (2; 1.14 mL, 8.8 mmol) was added dropwise,
and the resulting mixture was stirred for 0.5 h. The temperature was
adjusted to 20 °C, and then 2-bromo-4-fluoroanisole (2; 8.35 mL, 64.4
mmol) diluted with THF (36.0 mL) was added over 0.5 h, maintaining
a reaction temperature of <40 °C. When addition was complete, the
reaction mixture was stirred for 1 h, after which time full Grignard
reagent formation was confirmed by GC-MS analysis. A solution of
anhydrous DMF (6.91 mL, 89.0 mmol) in THF (36.0 mL) was added
dropwise over 0.5 h to the Grignard mixture at 45 °C. The resulting
mixture was stirred at 45−55 °C for 1 h. The reaction mixture was
adjusted to 23 °C, lithium borohydride (4 M solution in THF, 9.15
mL, 36.6 mmol) was added dropwise, and the mixture was stirred for 1
h at this temperature. The reaction mixture was subsequently
transferred via cannula to an aqueous solution of acetic acid (20 wt
%, 105 mL) and stirred for 0.5 h, and then the phases were separated.
CONCLUSIONS
■
The synthetic route to benzyl alcohol 1 leading to introduction
of the hydroxymethyl group via a combination of Bouveault
formylation and hydride reduction has been optimized using a
rational, mechanistic-based approach. This approach enabled
telescoping of the two steps into a single process, producing the
target compound 1 on a commercial scale in excellent in situ
yield (95%) and purity (97%). In addition, elimination of the
aldehyde workup reduces the overall process mass intensity by
greater than 20%. Conditions were developed which used 2-
MeTHF as the primary process solvent, which may be derived
from renewable resources. This approach is amenable to large-
scale manufacture and affords significant process safety
advantages relative to the formaldehyde and halomethylation
approaches. It is anticipated that this methodology could be
readily extended to the synthesis of other useful pharmaceut-
ical, fine chemical, and agricultural product intermediates.
EXPERIMENTAL SECTION
■
General Considerations. Unless otherwise noted, tetrahydrofuran
(THF) and 2-methyltetrahydrofuran (2-MeTHF) were distilled prior
to use over sodium benzophenone ketyl. Cyclopentyl methyl ether
(anhydrous, ≥99.9%) was used as purchased from Sigma-Aldrich. All
reactions were carried out under an inert atmosphere. NMR spectra
were recorded on a 300, 400, or 500 MHz NMR spectrometer. All
spectra were recorded at room temperature (∼20 °C) in deuterated
chloroform (CDCl3), unless otherwise stated, using tetramethylsilane
(TMS) as an internal standard. COSY and HETCOR correlations
were used to confirm the NMR peak assignments of all novel
compounds. 19F NMR spectra were recorded on a 400 MHz
spectrometer with complete carbon decoupling and referenced using
hexafluorobenzene (C6F6 in CDCl3 δ −162.2). High-resolution mass
spectrometry (HRMS) was performed on a TOF instrument in
electrospray ionization (ESI) mode; samples were made up in
acetonitrile. GC-MS analysis was carried out using a DB-WAX (30
m × 0.25 mm i.d. × 0.25 μm film) column under the following
F
dx.doi.org/10.1021/jo400647t | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
The organic phase was washed with aqueous potassium hydroxide (10
wt %, 100 mL), and then the aqueous layers were separated and
discarded. The solvent was removed by rotary evaporation to obtain
crude (5-fluoro-2-methoxyphenyl)methanol (1; 10.94 g, 95.7%) as a
yellow oil.
temperature of <30 °C. When the addition was complete, the
temperature was adjusted to 20 °C and the mixture was stirred for an
additional 1 h. HPLC analysis indicated less than 1% 2-bromo-4-
fluoroanisole (2) starting material remaining. In a separate 1 L
jacketed vessel anhydrous DMF (47.1 g, 0.64 mol, 1.2 equiv) and
nondistilled commercial THF (110 mL) were charged and the
temperature was adjusted to 45 °C. The Grignard reagent 3 was
sequestered from the elemental magnesium and added to an addition
funnel attached to the vessel containing DMF. (Note 1: the remaining
magnesium heel can be reused directly at the same scale or additional
magnesium added prior to the next experiment to attain the same
magnesium loading.) The Grignard solution was fed over 0.5 h to the
DMF solution stirred at 45−50 °C. (Note 2: during the Grignard
reagent feed a light suspension formed, which redissolved after
addition of ∼75% of the Grignard reagent feed.) After the feed was
complete, the remaining magnesium heel was washed with 2-MeTHF
(50 mL) and the rinsing added to the DMF solution. The resulting
mixture was stirred at 45−55 °C for 1 h and then cooled to 20 °C.
Lithium borohydride (2 M solution in THF, 268 mL, 0.536 mol) was
added, and the mixture was stirred for 0.5 h at 20 °C. GC-MS analysis
(MeOH quenched sample) revealed <1% aldehyde 4. In a 3 L jacketed
vessel were charged water (330 mL) and methanol (220 mL), and
then the temperature of the contents was adjusted to 20 °C. The
reductive reaction mixture was transferred to the quench solution over
at least 0.5 h, maintaining a temperature of less than 30 °C. The
reaction vessel was rinsed with 2-MeTHF (660 mL), and the rinsings
were transferred to the quench vessel. Acetic acid (112.8 g, 1.88 mol)
was dissolved in water (330 mL), and then the aqueous acetic acid
solution was added to the quench mixture. After the mixture was
stirred for 0.5 h, the phases were separated and the lower aqueous
phase was removed (750 mL; pH 5). The organic phase was washed
with 2 × 550 mL of NaOH (0.5 N), and the aqueous rinsings were
discarded. (Note 3: the second base wash was stirred for 1 h to ensure
full hydrolysis of any remaining boronate ester. The pH of the first
base was ∼6 and that of the second base wash was ∼13.) The organic
layer was then washed with water (550 mL), and the aqueous layer
(pH ∼10) was separated and discarded. The organic layer was
concentrated, taken up in toluene, and evaporated to provide an oil
which was dried under high vacuum for 24 h to give the final crude
compound 1 (84.8 g, 101.1%, HPLC purity 97%). Compound 1 was
purified by crystallization in toluene/heptane using the method
described for the two-step synthesis of 1 (method a) to produce
alcohol 1 (78.0 g, 92.0% recovery from crude) as a white crystalline
solid with >99% purity (overall yield 78.0 g, 93.1%).
Grignard Exchange Reaction (Method c). Isopropylmagnesium
chloride (2 M solution in THF, 73.2 mL, 146.0 mmol) was added
dropwise over 0.5 h at 30 °C to a solution of 2-bromo-4-fluoroanisole
(2; 9.49 mL, 73.2 mmol) in THF (36.0 mL). When addition was
complete, the reaction mixture was stirred for 4 h, after which time full
Grignard reagent formation was confirmed by GC-MS analysis. A
solution of anhydrous DMF (11.33 mL, 146.3 mmol) in THF (36.0
mL) was added dropwise over 0.5 h to the Grignard mixture at 45 °C.
The resulting mixture was stirred at 45−55 °C for 1 h and then
quenched by slow addition via cannula transfer to aqueous acetic acid
(20 wt %, 105 mL) at 0 °C. The phases were separated, and the
organic phase was washed with aqueous sodium carbonate (5 wt %,
183 mL). The aqueous phase was discarded, and the organic layer was
concentrated by rotary evaporation to give the crude aldehyde
intermediate 4 as an orange solid. A 10 wt % solution of 4 in THF was
adjusted to 23 °C, lithium borohydride (4 M solution in THF, 9.15
mL, 36.6 mmol) was added dropwise, and the mixture was stirred for 1
h at this temperature. The reaction mixture was subsequently
transferred via cannula to an aqueous solution of acetic acid (20 wt
%, 105 mL) and stirred for 0.5 h, and then the phases were separated.
The organic phase was washed with aqueous potassium hydroxide (10
wt %, 100 mL), and then the aqueous layers were separated and
discarded. The solvent was removed by rotary evaporation to obtain
crude (5-fluoro-2-methoxyphenyl)methanol (1; 10.78 g, 94.4%) as a
yellow oil. The crude product was purified by crystallization in
toluene/heptane using the method described for the two-step
synthesis of 1 (method a) to produce alcohol 1 (9.20 g, 85.5%
recovery from crude) as a white crystalline solid with purity >99%
(overall yield 9.20 g, 80.5%).
Revised One-Pot Synthesis (Method d). Magnesium (2.10 g,
86.0 mmol) was suspended in 2-MeTHF (50.0 mL), and iodine (0.19
g, 0.73 mmol) was added. The reaction mixture was heated to 30 °C,
2-bromo-4-fluoroanisole (2; 1.14 mL, 8.8 mmol) was added dropwise,
and the resulting mixture was stirred for 0.5 h. The temperature was
adjusted to 20 °C, and then 2-bromo-4-fluoroanisole (8.35 mL, 64.4
mmol) diluted with 2-MeTHF (25.0 mL) was added over 0.5 h,
maintaining a reaction temperature of <40 °C. When addition was
complete, the reaction mixture was stirred for 1 h, after which time full
Grignard reagent formation was confirmed by GC-MS analysis. A
solution of anhydrous DMF (6.91 mL, 89.0 mmol) in THF (36.0 mL)
was added dropwise over 0.5 h to the Grignard mixture at 45 °C. The
resulting mixture was stirred at 45−55 °C for 1 h. The reaction
mixture was adjusted to 23 °C, lithium borohydride (4 M solution in
THF, 9.15 mL, 36.6 mmol) was added dropwise, and the mixture was
stirred for 1 h at this temperature. The reaction mixture was
subsequently transferred via cannula over 0.5 h to a solution of water
(50.0 mL) and methanol (50.0 mL), maintaining a quench
temperature of <30 °C. A solution of aqueous acetic acid (20 wt %,
105 mL) was added, the mixture was stirred for 0.5 h, and then the
phases were separated. The organic phase was washed once with
aqueous potassium hydroxide (10 wt %, 100 mL), stirred for 1 h with
aqueous potassium hydroxide (10 wt %, 100 mL), and finally washed
with water (100 mL). The aqueous rinsings were discarded, and the
organic layer was concentrated by rotary evaporation to obtain crude
(5-fluoro-2-methoxyphenyl)methanol (1) as a yellow oil (10.97 g,
96.0%).
5-Fluoro-2-methoxybenzaldehyde (4): white solid; GC-MS
retention time 11.5 min; δH (400 MHz) 3.93 (3H, s, OCH3), 6.96
(1H, dd, JHH 9.0, JHF 3.8, ArH), 7.24−7.30 (1H, m, ArH), 7.51 (1H,
dd, JHH 8.3, JHF 3.3, ArH), 10.42 (1H, d, J 3.2, CHO); δF (400 MHz)
−120.7. Spectral characteristics for 4 are consistent with previously
reported data.18
(5-Fluoro-2-methoxyphenyl)methanol (1): white crystalline solid;
GC-MS retention time 14.5 min; δH (400 MHz) 2.21 (1H, t, J 6.5,
OH), 3.83 (3H, s, OCH3), 4.66 (2H, d, J 6.5, CH2OH), 6.78 (1H, dd,
JHH 8.9, JHF 4.3, ArH), 6.94 (1H, td, JHH 8.5, JHF 3.1, ArH), 7.04 (1H,
dd, JHH 8.7, JHF 3.1, ArH); δF (400 MHz) −124.3. Spectral
characteristics for 1 are consistent with previously reported data.1
1-(5-Fluoro-2-methoxyphenyl)-N,N-dimethylmethanamine (5a):
white solid: mp 64−66 °C; GC-MS retention time 9.4 min; δH (300
MHz) 2.44 (6H, s, 2 × CH3), 3.73 (3H, s, OCH3), 3.97 (2H, s, CH2),
6.80 (1H, dd, JHH 9.0, JHF 4.5, ArH), 6.93−7.05 (2H, m, ArH); δC
(75.5 MHz) 49.9 (2 × CH3), 55.9 (OCH3), 60.6 (CH2), 111.9 (CH,
3
2
Commercial Procedure (Method e). In a 1 L jacketed vessel
magnesium (32.6 g, 1.34 mol) was suspended in nondistilled
commercial 2-MeTHF (220 mL), iodine (1.36 g, 0.0054 mol) was
then added, and the mixture was cooled to 10 °C under N2. 2-Bromo-
4-fluoroanisole (2; 10 g, 0.049 mol) was added, and a strong exotherm
(10 °C) was observed within a couple of minutes. The exotherm
quickly tailed, and once the mixture reached 10 °C, a solution
containing 2-bromo-4-fluoroanisole (2; 100 g, 0.488 mol) and 2-
MeTHF (110 mL) was added over 1.5 h, maintaining a reaction
d, JCF 8.1, aromatic CH), 117.0 (CH, d, JCF 22.8, aromatic CH),
2
120.8 (CH, d, JCF 22.8, aromatic CH), 154.6 (aromatic C), 155.0
(aromatic C), 157.8 (aromatic C); δF (400 MHz) −123.6; HRMS
(ES+) exact mass calcd for C10H15FNO (M + H)+ 184.1138, found
184.1134.
4-(5-Fluoro-2-methoxybenzyl)morpholine (5b): yellow oil; GC-
MS retention time 15.4 min; δH (300 MHz) 2.47−2.52 (4H, sym m, 2
× CH2), 3.52 (2H, s, CH2N), 3.70−3.75 (4H, sym m, 2 × CH2), 3.80
(3H, s, OCH3), 6.77 (1H, dd, JHH 8.9, JHF 4.4, ArH), 6.89 (1H, td, JHH
G
dx.doi.org/10.1021/jo400647t | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
8.9, JHF 3.2, ArH), 7.14 (1H, dd, JHH 9.2, JHF 3.2, ArH); δC (75.5 MHz)
53.7 (2 x CH2), 56.0 (OCH3), 56.1 (CH2N), 67.1 (2 × CH2), 111.3
AUTHOR INFORMATION
Corresponding Author
*E-mail for A.R.M.: a.maguire@ucc.ie.
Notes
■
3
2
(CH, d, JCF 8.2, aromatic CH), 113.7 (CH, d, JCF 23.0, aromatic
2
3
CH), 116.6 (CH, d, JCF 23.6, aromatic CH), 128.1 (C, d, JCF 6.9,
aromatic C), 153.8 (C, d, 4JCF 2.2, aromatic C), 157.1 (C, d, 1JCF 237.9,
aromatic C); δF (400 MHz) −124.1; HRMS (ES+) exact mass calcd
for C12H17FNO2 (M + H)+ 226.1243, found 226.1233.
The authors declare no competing financial interest.
1-(5-Fluoro-2-methoxybenzyl)piperidine (5c): orange oil; GC-MS
retention time 13.5 min; δH (400 MHz) 1.38−1.50 (2H, m, CH2),
1.52−1.68 (4H, m, 2 × CH2), 2.42 (4H, t, J 4.7, 2 × CH2), 3.48 (2H,
s, CH2N), 3.79 (3H, s, OCH3), 6.75 (1H, dd, JHH 8.9, JHF 4.4, ArH),
6.86 (1H, td, JHH 8.5, JHF 2.9, ArH), 7.15 (1H, dd, JHH 9.3, JHF 2.7,
ArH); δC (75.5 MHz) 24.3 (CH2), 26.1 (2 x CH2), 54.6 (2 × CH2),
ACKNOWLEDGMENTS
■
Enterprise Ireland is acknowledged for the financial support of
C.N.S. and R.E.D. This work is dedicated to the memory of
Prof. A. I. Meyers.
3
55.9 (OCH3), 56.4 (CH2N), 111.1 (CH, d, JCF 8.1, aromatic CH),
REFERENCES
■
2
2
113.2 (CH, d, JCF 23.0, aromatic CH), 116.6 (CH, d, JCF 23.0,
(1) Eli Lilly and Company. Treatment of Pervasive Developmental
Disorders with Norepinephrine Reuptake Inhibitors, WO2005/20975
A2, 2005.
(2) (a) Schwarz Pharma, Ltd. Shortened Synthesis Using
Paraformaldehyde or Trioxanes, US8,115,028, 2012. (b) Miteni, S.
P. A., Process for the Preparation of 3,5-bis(trifluoromethyl)
benzylalcohol, WO2005/035472A1, 2005. (c) Gillman, H.; Catlin,
W. E. Org. Synth. 1941, 1, 188.
(3) (a) Hoffman LaRoche, Inc. Chloromethylation of Benzene
Compounds, US3,465,051, 1969. (b) Knoll, A. G. Process for
Producing Veratyl compounds, US2,695,319, 1954. (c) Dozeman, G.
J.; Fiore, P. J.; Pols, J. P.; Walker, J. C. Org. Process Res. Dev. 1997, 1,
137−148.
(4) Bouveault, M. L. Bull. Soc. Chim. Fr. 1904, 31, 1306−1322.
(5) Comins, D.; Meyers, A. I. Synthesis 1978, 403−404.
(6) (a) Frechet, J. M.; Amaratunga, W. Tetrahedron Lett. 1983, 24,
1143−1146. (b) Zanka, A. Org. Process Res. Dev. 2000, 4, 46−48.
(7) Meyers, A. I.; Comins, D. Tetrahedron Lett. 1978, 19, 5179−5182.
(8) (a) Pace, V.; Hoyos, P.; Castoldi, L.; Dominguez, P.; Alcantara, A.
R. ChemSusChem 2012, 5, 1369−1379. (b) Sitthisa, S.; Resasco, D. E.
Catal. Lett. 2011, 141, 784−791.
(9) The Grignard reagent formation is highly exothermic with a
reaction enthalpy of −291 kJ/mol and TAd = 141 °C. For this reason it
is critical that the substrate feed is not commenced until initiation is
verified. It typically takes 2−4 min to achieve a productive initiation;
however, high levels of water in the system (>500 ppm) can
significantly delay the onset of initiation. However, once initiation
occurs, the Grignard formation is feed rate limited and the exotherm
can be controlled by the substrate dosage rate.
(10) Olah, G. A.; Surya Prakash, G. K.; Arvanaghi, M. Synthesis 1984,
1984, 228−230.
(11) One experiment was also conducted with N,N-dibutylforma-
mide as the formylating agent. This experiment provided alcohol 1
with a purity of 89.1%.
aromatic CH), 129.0 (C, d, 3JCF 6.9, aromatic C), 153.7 (C, d, 4JCF 2.0,
1
aromatic C), 157.2 (C, d, JCF 237.5, aromatic C); δF (400 MHz)
−124.3; HRMS (ES+) exact mass calcd for C13H19FNO (M + H)+
224.1451, found 224.1442.
4-(5-Fluoro-2-methoxyphenyl)-5-methylhexan-1-ol (6): yellow
oil; GC-MS retention time 16.9 min; δH (400 MHz) 0.72 (3H, d, J
6.7, CH3), 0.95 (3H, d, J 6.7, CH3), 1.17 (1H, bs, OH), 1.29−1.36
(2H, m, CH2), 1.48−1.58 (1H, m, one of CH2), 1.74−1.87 (2H, m,
CH and one of CH2), 2.82−2.89 (1H, bm, CH), 3.56 (2H, t, J 6.6,
CH2OH), 3.77 (3H, s, OCH3), 6.74−6.86 (3H, m, ArH); δC (75.5
MHz) 20.6 (CH3), 20.9 (CH3), 28.2 (CH2), 30.8 (CH2), 32.9 (CH),
3
44.0 (CH, bs), 56.1 (OCH3), 63.1 (CH2), 111.5 (CH, d, JCF 8.3,
2
aromatic CH), 112.3 (CH, d, JCF 22.8, aromatic CH), 114.6 (CH, d,
2JCF 22.8, aromatic CH), 135.2 (C, d, 3JCF 6.4, aromatic C), 154.1 (C,
d, 4JCF 2.0, aromatic C), 157.3 (C, d, 1JCF 237.5, aromatic C); δF (400
MHz) −124.0; HRMS (ES+) exact mass calcd for C14H22FO2 (M +
H)+ 241.1604, found 241.1594.
5,5′-Difluoro-2,2′-dimethoxybiphenyl (8): white solid: mp 114−
117 °C; GC-MS retention time 16.6 min; δH (500 MHz) 3.75 (6H, s,
OCH3), 6.86−6.91 (2H, m, ArH), 6.97−7.03 (4H, m, ArH); δF (400
MHz) −124.4; HRMS (ES+) exact mass calcd for C14H13F2O2 (M +
H)+ 251.0884, found 251.0879. Spectral characteristics for 8 are
consistent with previously reported data.19
Bis(5-fluoro-2-methoxyphenyl)methanol (10): white solid: mp
105−108 °C; GC-MS retention time 22.3 min; δH (400 MHz) 3.47
(1H, d, J 5.1, OH), 3.80 (6H, s, 2 × OCH3), 6.24 (1H, d, J 5.1, CH),
6.79−6.84 (2H, m, ArH), 6.91−6.99 (4H, m, ArH); δC (75.5 MHz)
3
56.0 (2 × OCH3), 66.8 (CH), 111.4 (CH, d, JCF 8.1, 2 × aromatic
2
2
CH), 114.4 (CH, d, JCF 23.1, 2 × aromatic CH), 114.8 (CH, d, JCF
24.5, 2 × aromatic CH), 132.2 (C, d, 3JCF 6.6, 2 × aromatic C), 152.8
(C, d, 4JCF 2.1, 2 × aromatic C), 157.2 (C, d, 1JCF 238.6, 2 × aromatic
C); δF (400 MHz) −123.3; HRMS (ES+) exact mass calcd for
C15H13O2F2 (M − OH)+ 263.0884, found 263.0881.
2-(2,3-Dimethylbutan-2-yl)-4-fluoro-1-methoxybenzene (11):
sticky white solid; GC-MS retention time 9.5 min; δH (400 MHz)
0.73 (6H, d, J 6.9, 2 × CH3), 1.24 (6H, s, 2 × CH3), 2.60 (1H, qu, J
6.9, CH), 3.78 (3H, s, OCH3), 6.74−6.86 (2H, m, ArH), 6.94 (1H, dd,
JHH 11.1, JHF 3.1, ArH); δC (75.5 MHz) 18.1 (2 x CH3), 23.5 (2 ×
(12) Knochel, P.; Dohle, W.; Gommermann, N.; Kneisel, F. F.;
Kopp, F.; Korn, T.; Sapountzis, I.; Vu, V. A. Angew. Chem., Int. Ed.
2003, 42, 4302−4320.
(13) Krasovskiy, A.; Knochel, P. Angew. Chem., Int. Ed. 2004, 43,
3333−3336.
2
CH3), 32.4 (CH), 41.4 (C), 55.7 (OCH3), 112.1 (CH, d, JCF 21.9,
(14) Leazer, J. L.; Cvetovich, R.; Tsay, F.-R.; Dolling, U.; Vickery, T.;
Bachert, D. J. Org. Chem. 2003, 68, 3695−3698.
3
aromatic CH), 112.3 (CH, d, JCF 8.4, aromatic CH), 115.0 (CH, d,
2JCF 24.0, aromatic CH), 140.4 (C, d, 3JCF 6.0, aromatic C), 154.5 (C,
d, 4JCF 2.1, aromatic C), 156.9 (C, d, 1JCF 236.5, aromatic C); δF (400
MHz) −124.3; HRMS (ES+) exact mass calcd for C13H20FO (M +
H)+ 211.1498, found 211.1495.
(15) Hock, H.; Kropf, H.; Ernst, F. Angew. Chem. 1959, 71, 541−545.
(16) During scaleup an additional impurity was observed by 1H
NMR analysis which was not identified in the previous small-scale
studies. This impurity appears to be due to the reaction of
dimethylamine with the borohydride or borohydride derived by-
products, although its full structure has not been confirmed.
(17) The impurities reported according to HPLC retention time are
as follows: (1) unknown, 0.68% (RRT 0.74); 5-fluoro-2-methoxy
benzoic acid, 0.45% (RRT 0.95); positional isomer 1, 0.59% (RRT
1.02); positional isomer 2, 0.31% (RRT 1.05).
ASSOCIATED CONTENT
■
S
* Supporting Information
Text, tables, and figures giving details of the quench study,
process safety data for the commercial procedure (method e)
(18) Laali, K. K.; Koser, G. F.; Subramanyam, S.; Forsyth, D. A. J.
Org. Chem. 1993, 58, 1385−1392.
(19) Kar, A.; Mangu, N.; Kaiser, H. M.; Tse, M. K. J. Organomet.
Chem. 2009, 694, 524−537.
1
and H and 13C NMR, COSY, and HETCOR spectra for the
novel compounds 5a−c, 6, 10, and 11. This material is available
free of charge via the Internet at http://pubs.acs.org.
H
dx.doi.org/10.1021/jo400647t | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
(20) Further quench studies, including exploration of the sequence
and relative rates of the various steps in the process, are outlined in the
Supporting Information.
I
dx.doi.org/10.1021/jo400647t | J. Org. Chem. XXXX, XXX, XXX−XXX