Formation of 2-trifluoromethylphenyl grignard reagent via magnesium-halogen exchange: Process safety evaluation and concentration effect

Article abstract of DOI:10.1021/op900040y

The thermal stability profile for a solution of 2-trifluoromethylphenyl magnesium chloride at 1.5Mconcentration in THF was determined using an Advanced Reactive System Screening Tool (ARSST). The solution generated by employing Knochel's magnesium-halogen

Full text of DOI:10.1021/op900040y

Organic Process Research & Development 2009, 13, 1426–1430
Technical Notes
Formation of 2-Trifluoromethylphenyl Grignard Reagent via Magnesium-Halogen
Exchange: Process Safety Evaluation and Concentration Effect
Wenjun Tang,* Max Sarvestani,* Xudong Wei, Laurence J. Nummy, Nitinchandra Patel, Bikshandarkoil Narayanan,
Denis Byrne, Heewon Lee, Nathan K. Yee, and Chris H. Senanayake
Department of Chemical DeVelopment, Boehringer Ingelheim Pharmaceuticals Inc., Ridgefield, Connecticut 06877, U.S.A
Abstract:
been reported.2 For example, an accident during the preparation
of 4-trifluoromethylphenyl magnesium chloride on a com-
mercial scale resulted in loss of life and destruction of a
chemical plant.3 Pfizer reported a severe decomposition during
the formation of 3-trifluoromethylphenyl magnesium bromide
causing extensive damage to a laboratory.4 It was believed that
the severe decomposition was initiated by highly active
magnesium particles generated in the reaction.2 This important
process safety issue has hampered the utility of trifluoromethyl-
substituted phenyl Grignard reagents in large-scale organic
synthesis.
In the last 10 years, Knochel and co-workers have developed
an attractive method for generating aryl Grignard reagents by
employing an alkyl Grignard reagent such as iPrMgCl to
promote metal-halogen exchange.5 This method enables the
transformation of many aromatic iodides and bromides into
aromatic Grignard reagents at or below room temperature. By
contrast, the traditional method mediated by magnesium metal
often requires heating and initiation to avoid a dangerous
induction period. From the process safety point of view, the
Knochel methodology would seem to possess attractive char-
acteristics for large-scale preparation. Further encouragement
for the use of Knochel’s method in an industrial context was
offered by Leazer and co-workers at Merck who described a
safer process for 3,5-bis(trifluoromethyl)phenyl Grignard reagent
from 3,5-bis(trifluoromethyl)bromobenzene and iPrMgBr.6 The
Merck scientists examined the stability of various trifluoro-
methyl-substituted phenyl Grignard reagents using Differential
Thermal Analysis (DTA) and a Reactive System Screening Tool
(RSST). They concluded Knochel’s method was safer for the
preparation of these reagents than the traditional method with
magnesium metal. Indeed the testing showed no exothermic
decomposition from Grignard solutions prepared with Knochel’s
The thermal stability profile for a solution of 2-trifluorometh-
ylphenyl magnesium chloride at 1.5 M concentration in THF was
determined using an Advanced Reactive System Screening Tool
(ARSST). The solution generated by employing Knochel’s
magnesium-halogen exchange protocol showed highly exothermic
decomposition. The decomposition begins at a low-onset temper-
ature accompanied by a rapid temperature and pressure rise.
Analysis of the decomposition mixture revealed the destruction
of trifluoromethyl group and formation of fluoride ion. This
decomposition profile was substantially attenuated by reducing
the concentration of the solution to 0.5-0.6 M. Thus, it is strongly
recommended that selecting an appropriate concentration for the
reagent based on calorimetric evaluation should be included with
procedural and engineering controls when considering any strategy
for safe scale-up of trifluoromethyl-substituted phenyl Grignard
solutions.
Introduction
The trifluoromethyl-substituted arene moiety is being ex-
ploited with increasing frequency for the design of pharmaceuti-
cal and agrochemical agents.1 The desire to evaluate the unique
properties of these compounds in biological systems has
stimulated the search for practical synthetic methods to prepare
them. Trifluoromethyl-substituted phenyl Grignard reagents are
versatile intermediates that could potentially satisfy the require-
ments. However, the preparation of these intermediates from a
suitable aryl halide by the traditional method using magnesium
metal is extremely dangerous. Several severe explosions have
* Authors for correspondence. E-mail: wenjun.tang@boehringer-ingelheim.com;
max.sarvestani@boehringer-ingelheim.com.
(1) (a) Thayer, A. M. Chem. Eng. News 2006, 84, 27. (b) Brands, K. M. J.;
Payack, J. F.; Rosen, J. D.; Nelson, T. D.; Candelario, A.; Huffman,
M. A.; Zhao, M. M.; Li, J.; Craig, B.; Song, Z. J.; Tschaen, D. M.;
Hansen, K.; Devine, P. N.; Pye, P. J.; Rossen, K.; Dormer, P. G.;
Reamer, R. A.; Welch, C. J.; Mathre, D. J.; Tsou, N. N.; McNamara,
J. M.; Reider, P. J. J. Am. Chem. Soc. 2003, 125, 2129. (c) Houlihan,
W. J.; Gogerty, J. H.; Ryan, E. A.; Schmitt, G. J. Med. Chem. 1985,
28, 28. (d) Welch, J. T.; Eswarakrishnan, S. Fluorine in Bioorganic
Chemistry; Wiley: New York, 1991. (e) Desai, R. C.; Cicala, P.; Meurer,
L. C.; Finke, P. E. Tetrahedron Lett. 2002, 43, 4569. (f) Kuethe, J. T.;
Wong, A.; Wu, J.; Davies, I. W.; Dormer, P. G.; Welch, C. J.; Hillier,
M. C.; Hughes, D. L.; Reider, P. J. J. Org. Chem. 2002, 67, 5993. (g)
Riachi, N. J.; Arora, P. K.; Sayre, L. M.; Harik, S. I. J. Neurochem.
1988, 1319.
(2) Waymouth, R.; Moore, E. J. Chem. Eng. News 1997, 75, 6.
(3) Ashby, E. C.; Al-Fekri, D. M. J. Organomet. Chem. 1990, 390, 275.
(4) Appleby, I. C. Chem. Ind. 1971, 120.
(5) (a) Abarbri, M.; Dehmel, F.; Knochel, P. Tetrahedron Lett. 1999, 40,
7449. (b) Boymond, L.; Rottla¨nder, M.; Cahiez, G.; Knochel, P. Angew.
Chem., Int. Ed. 1998, 37, 1701. (c) Jensen, A. E.; Dohle, W.; Sapountzis,
I.; Lindsay, D. M.; Vu, V. A.; Knochel, P. Synthesis 2002, 4, 565. (d)
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.
(6) (a) Leazer, J. L., Jr.; Cvetovich, R.; Tsay, F.-R.; Dolling, U.; Vickery,
T.; Bachert, D. J. Org. Chem. 2003, 68, 3695. (b) Leazer, J. L., Jr.;
Cvetovich, R. Org. Synth. 2005, 82, 115.
1426
•
Vol. 13, No. 6, 2009 / Organic Process Research & Development
10.1021/op900040y CCC: $40.75  2009 American Chemical Society
Published on Web 06/25/2009

Scheme 1. Formation of 2-trifluoromethylphenyl Grignard
reagent by magnesium-bromide exchange
like to report our studies concerning the exothermic decomposi-
tion of this solution at elevated temperature by using adiabatic
calorimetry in an Advanced Reactive System Screening Tool
(ARSST) and the influence of concentration on this event.
Results and Discussion
A sample (10 mL) of solution prepared as shown in Scheme
1 was transferred under nitrogen atmosphere to the reaction cell
in a closed 350-mL ARSST containment vessel. The purged
containment was filled with nitrogen (∼300 psig) and subjected
to a temperature ramp rate of ∼2 °C/min (see Experimental
Section for more details). To our surprise, a highly exothermic
event was observed with the onset temperature at about 80 °C.
The peak rate of temperature rise (dT/dt) reached 3200 °C/min,
and the peak rate of pressure rise (dP/dt) reached 6000 psi/min
(Figure 1). This severe decomposition was in sharp contrast to
Leazer’s report,6 where no severe exothermic events were
method, although no information about concentration was
mentioned in the paper.
Considering the inherently safer features of Knochel’s
method and the results of the Leazer investigation, we were
optimistic that employing 2-trifluoromethylphenyl Grignard
reagent would be an ideal strategy to accomplish one of our
recent process development challenges. Thus, by using Knoch-
el’s method, 2-bromobenzotrifluoride was converted to 2-tri-
fluoromethylphenyl magnesium chloride solution (∼1.1 M) with
>95% conversion as illustrated in Scheme 1. Here, we would
Figure 1. Formation of 2-trifluoromethylphenyl magnesium chloride (1.1 M). Data from ARSST quasi-adiabatic calorimetry: (a)
temperature vs time plot; (b) pressure vs time plot; (c) time derivative of temperature (temperature rise rate) vs temperature plot
(log scale); (d) time derivative of pressure (pressure rise rate) vs temperature plot (log scale). Both the heat-up and cool-down
portions are shown in (a) and (b). Only the heat-up portion is shown in (c) and (d).
Vol. 13, No. 6, 2009 / Organic Process Research & Development
•
1427

Scheme 2. Magnesium-bromide exchange with iPrMgBr
iPrMgBr and not iPrMgCl as the reagent for their RSST and
DTA studies, we wondered whether this difference could
explain the contrast between our respective observations. We
thus employed a commercial solution of iPrMgBr (0.85-1.0
M) to prepare a solution of 2-trifluoromethylphenyl magnesium
bromide solution at ∼0.6 M concentration (Scheme 2). The
ARSST profile of this solution showed a much milder exo-
thermic decomposition with the onset temperature at about 100
°C, a peak rate of temperature rise at 50 °C/min, and a peak
rate of pressure rise at 10 psi/min (Figure 2). Although the
concentration of this Grignard reagent (0.6 M) is essentially
half that of the aforementioned 2-trifluoromethylphenyl mag-
nesium chloride solution (1.1 M), we found it striking that the
differences in both the peak rate of temperature and pressure
rise for the decomposition disproportionately large. To distin-
guish whether the results are due to a counterion effect (Cl vs
observed during both DTA and RSST studies. In order to
eliminate the possible involvement of magnesium particles
introduced from commercial Grignard reagent, the iPrMgCl
solution purchased from Sigma-Aldrich was filtered through a
10 µm filter prior to performing the magnesium-halogen
exchange. The experimental data once again showed a dramatic
decomposition event with similar magnitude. The low decom-
position onset temperature (∼80 °C) together with very rapid
temperature and pressure rise led to a process safety concern
for scale-up to pilot-plant batches. Since Leazer et al. employed
Figure 2. Formation of 2-trifluoromethylphenyl magnesium chloride (0.6 M). Data from ARSST quasi-adiabatic calorimetry: (a)
temperature vs time plot; (b) pressure vs time plot; (c) time derivative of temperature (temperature rise rate) vs temperature plot
(log scale); (d) time derivative of pressure (pressure rise rate) vs temperature plot (log scale). Both the heat-up and cool-down
portions are shown in (a) and (b). Only the heat-up portion is shown in (c) and (d).
1428
•
Vol. 13, No. 6, 2009 / Organic Process Research & Development

Table 1. Data summary from ARSST quasi-adiabatic calorimetry of various trifluoromethyl-substituted phenyl Grignard
solutions made with Knochel’s method at different concentrations
a The reactions were run under nitrogen at 15-25 °C in THF for 2-12 h. Mole ratio of aryl bromide: iPrMgBr/iPrMgCl ) 1:1.05. A 10-mL sample was taken for
adiabatic calorimetry experiment. See Experimental Section for details. b The mixture is highly exothermic during decomposition as two consecutive decomposition events
were observed.
Br) or a concentration effect, a solution of 2-trifluorometh-
ylphenyl magnesium chloride with the same concentration
(∼0.6 M) was prepared and evaluated by ARSST (Table 1,
experiments 3 and 6). The results were similar to those obtained
from the 2-trifluoromethylphenyl magnesium bromide solution.
This indicates that the decomposition rate is not dependent on
the counterions present but on the concentration of the Grignard
solution. Table 1 provides additional data that establish the direct
relationship between concentration and the severity of decom-
position. An extremely exothermic event was seen when the
concentration increased to 1.5 M while a Grignard solution at
0.5 M concentration exhibits a greatly subdued decomposition
profile (experiment 1 vs 4). A 3-fold reduction in concentration
provides a dramatic difference in the onset temperature (100
°C vs 70 °C) and the decomposition rates (peak dT/dt: 25 °C/
min vs 15,000 °C/min, peak dP/dt: 5 psi/min vs 70,000 psi/
min). This trend was also observed with 3-trifluoromethylphenyl
Grignard solution, 4-trifluoromethylphenyl Grignard solution,
and 3,5-ditrifluoromethylphenyl Grignard solution, although the
relative change in magnitude and onset temperature for the
decomposition events varied with substrates. Severe decomposi-
tions were observed of all the solutions at a high concentreation
(∼1.5 M). A significant exothermic activity with two consecu-
tive decomposition events was observed in the thermal profile
of 1.5 M 4-trifluoromethylphenyl Grignard solution (see Sup-
porting Information). On the other hand, little or no exothermic
activity was seen at 0.5 M concentration.
HPLC analysis of the decomposition products of 2-trifluo-
romethylphenyl Grignard solution revealed a complex mixture.
The 19FNMR spectrum obtained from the mixture displays a
wide range of 19F signals (-40 to -110 ppm), suggesting the
involvement of the trifluoromethyl group in one or more
decomposition pathways. The presence of fluoride ion in the
decomposition mixture is substantiated by ion chromatographic
analysis. Calculations based on the fluoride ion content and the
total fluorine content of the decomposition mixture indicate that
nearly one-fourth to one third of organic fluorine has been
transformed into ionic fluoride (see Supporting Information).
Vol. 13, No. 6, 2009 / Organic Process Research & Development
•
1429

While these pathways are not easily understood, we speculate
that the main energy release could be largely due to the
formation of magnesium fluoride species owing to its high
lattice energy (the lattice energy of MgF2 is -2957 kJ/mol).
In summary, we have characterized the effect of concentra-
tion on the thermal decomposition of trifluoromethyl-substituted
phenyl Grignard solutions generated with Knochel’s method
using adiabatic calorimetry. The results demonstrate the im-
portance of this factor in gauging the risks involved with the
scale-up of these reagents despite the fact that they have been
prepared by Knochel’s method.7 It is our hope that the chemical
development community will benefit from this information
when an assessment is required to determine whether these
reagents can play a role at some stage in the development of a
biologically active substance containing the trifluoromethyl-
substituted arene moiety. A more effective strategy for safe
scale-up is possible when the concentration factor is combined
with procedural and engineering controls. In our scale-up
program, we successfully produced 83 L (50 mol) of 2-trifluo-
romethylphenyl magnesium bromide solution in THF at 0.6 M
concentration.
be applied and proved to be the same outcome. A ∼10 mL
sample was withdrawn during or after the completeness of
magnesium-bromide exchange and subjected to adiabatic
calorimetry study using an ARSST.
Data Collection of Adiabatic Calorimetry. Experimental
data were collected using ARSST manufactured by Fauske and
Associates.8 ARSST is a quasi-adiabatic instrument that works
on basis of heat loss compensation principle. The basic
component of the ARSST includes a spherical 10-mL glass test
cell, its surrounding “bottom heater” jacket and insulation,
thermocouple, pressure transducer, and a 350-mL containment
vessel that serves as both pressure simulator and safety vessel.
Tests were performed in the open test cell in closed containment.
Nitrogen pressure in the containment vessel is used to suppress
the boiling point of the sample. The sample temperature is
measured by a thermocouple inside the test cell. A magnetic
stir bar is placed inside the test cell and driven by an external
magnetic stirrer. A key feature of the apparatus is its low
effective heat capacity relative to that of the sample (low ꢀ
factor). Thus, the heat released by chemical reaction goes to
heat up the sample with negligible energy absorbed by the test
cell itself. A fill tube is used to add the mixture to the purged
test cell. ARSST containment including the reaction cell was
purged many times with nitrogen before introducing the reaction
mixture to the system by syringe through the fill tube. All
ARSST data were collected with the use of 2 °C/min temper-
ature ramp polynomial. All ARSST data points in the plots were
smoothed over five data points. In all experiments 10 mL of
the reaction mixture was used.
Experimental Section
General. All reagents were purchased from commercial
sources and used without further purifications unless otherwise
specified. All reactions and adiabatic calorimetry experiments
were performed under N2 atomosphere. HPLC analysis was
performed on an Agilent 1200 series with Halo C8 column;
flow rate of 1.3 mL/min; UV detection at 205 and 220 nm;
mobile phase of water (0.1% H3PO4) and MeCN/THF (95:5,
v/v) with gradient composition.
Supporting Information Available
Adiabatic calorimetry data of all the experiments including
temperature-time profile, temperature rise rate-temperature
profile, pressure-time profile, and pressure rise rate-temperature
profile as well as fluorine analysis of the decomposition
products. This material is available free of charge via the Internet
at http://pubs.acs.org.
General Procedure for Preparation of Trifluoromethyl-
Substituted Aryl Grignard Solution. To a 250-mL three-
necked flask equipped with magnetic stirrer and thermometer
was charged a solution of iPrMgCl (47 mL, 2.0 M, 94 mmol,
1.05 equiv) or iPrMgBr (104 mL, 0.9 M, 94 mmol, 1.05 equiv)
in THF. To the mixture at ∼15 °C was added neat trifluoro-
methyl-substituted bromide (89 mmol, 1.0 equiv) or its THF
solution (THF was added to adjust the concentration) over
10-15 min while controlling the temperature between 15-25
°C. The resulting solution was stirred at 15-25 °C for 2-12
h, after which HPLC analysis (aliquot into MeOH) showed
>95% conversion of bromide. For low concentrations (e1.1
M for aryl magnesium chloride, e0.75 M for aryl magnesium
bromide), a reversed addition (addition of iPrMgCl/ iPrMgBr
to a THF solution of trifluoromethyl-substituted bromide) can
Received for review February 25, 2009.
OP900040Y
(7) Song et al. reported another case of thermodecompositions of Grignard
solutions prepared by Knochel’s method: Reeves, J. T.; Sarvestani, M.;
Song, J. J.; Tan, Z.; Nummy, L. J.; Lee, H.; Yee, N. K.; Senanayake,
C. H. Org. Process Res. DeV. 2006, 10, 1258.
(8) For more information, see: Fauske, H. K. Process Saf. Prog. 2006, 25,
120.
1430
•
Vol. 13, No. 6, 2009 / Organic Process Research & Development