Download PDF
Synthesis 2012; 44(16): 2531-2536
DOI: 10.1055/s-0031-1290810
special topic
© Georg Thieme Verlag Stuttgart · New York

Near-Ambient Temperature Halogen–Lithium Exchange of p-Bromoanisole and Related Substrates: Flow/Batch Studies

D. W. Slocum*
a   Department of Chemistry, Western Kentucky University, Bowling Green, KY 42101, USA
b   KinetiChem Inc., 19100 Von Karman Ave., Suite 400, Irvine, CA 926612, USA, Fax: +1(270)7455361   Email: donald.slocum@wku.edu   Email: paul.whitley@wku.edu
,
Thomas K. Reinscheld
a   Department of Chemistry, Western Kentucky University, Bowling Green, KY 42101, USA
,
Nick D. Austin
a   Department of Chemistry, Western Kentucky University, Bowling Green, KY 42101, USA
,
Damir Kusmic
a   Department of Chemistry, Western Kentucky University, Bowling Green, KY 42101, USA
,
Paul E. Whitley*
b   KinetiChem Inc., 19100 Von Karman Ave., Suite 400, Irvine, CA 926612, USA, Fax: +1(270)7455361   Email: donald.slocum@wku.edu   Email: paul.whitley@wku.edu
› Author Affiliations
Further Information

Publication History

Received: 01 March 2012

Accepted: 07 March 2012

Publication Date:
27 April 2012 (online)

 
 PDF Download Permissions and Reprints All articles of this category


Abstract

With the advent of flow chemistry, the norm has been reactions executed on the laboratory scale with flow rates of only a few mL/min. We bring to the community’s attention our investigation of the halogen–lithium (X–Li) exchange in a continuous flow reactor, the Synthetron™. This novel reactor is capable of orders of magnitude greater rates of flow than current microreactors. This paper details a problematic X–Li exchange using our promoted hydrocarbon media formulated batch studies as well as the comparative derived flow studies. All of these studies have the additional feature of being performed at ambient or near-ambient temperatures. From the initial discoveries of Wittig and Gilman in the late 1930s, it has been known that X–Li exchange of p-bromoanisole (p-BrA) is plagued by a secondary ortho-lithiation. Fine-tuning of promoted hydrocarbon media batch studies can increase the ratio of p-LiA/o-Li-p-BrA; results from the Synthetron™ studies afford a much superior ratio of >100:1. Gram quantities of derivatives from this exchange (employing two reactors) can be prepared in a few seconds. Rationales for these observations will be presented as well as initial studies and discussion for bromobenzene (PhBr), m-bromoanisole (m-BrA), and p-iodoanisole (p-IA).


#

Key words

continuous flow chemistry - meso/macroflow reactor - halogen–lithium exchange - aryl bromide exchange rates - hydrocarbon media

Flow chemistry is evolving into a desired alternative means to accomplish numerous synthetic undertakings. The norm has been reactions executed on the microscale. Flow rates of μL/min to a few mL/min of dilute reactant feed solutions are common. We bring to the community’s attention the macroscale potential of the Synthetron™, a patented device,[ 1 ] which provides rapid, intimate mixing by a novel shearing mechanism. The Synthetron is a most versatile reactor, which is capable of rapidly synthesizing both a few hundred milligrams as well as a few kilograms of a substance. To accomplish large-scale syntheses with this reactor, flow rates of well over 100 mL/min of concentrated solutions are utilized.

Our previous work with the Synthetron has featured batch versus flow studies of the halogen–metal (X–Li) exchange.[ 2 ] These studies were performed at ambient temperature using promoted hydrocarbon media. The impetus behind these studies was to make the X–Li exchange safer, greener, more sustainable and atom-economical while maintaining high product output. By use of hydrocarbon media instead of the conventional ether media normally used for the exchange, several problems associated with ether media are avoided. Among these are the volatility and flashpoint of the common ethers, their susceptibility to attack by alkyllithium reagents and their miscibility with water. Moreover, the use of only one equivalent of n-butyllithium instead of the conventional two equivalents of tert-butyllithium further increases the safety, greenness and atom-economy of the reaction. Performing a reaction effectively at ambient or slightly below ambient temperature significantly lowers the cost, should a large-scale synthetic process be desired.

The use of ether-promoted cyclohexane media described in this article has some precedent in previous studies of ortho-lithiations carried out in our laboratory. Use of hydrocarbon media alone[ 3 ] or promoted hydrocarbon media containing measured equivalent of THF, MTBE, or tetramethylethylenediamine (TMEDA)[ 4 ] has demonstrated the ability of bis-chelating directing metalation groups as well as small amounts of an ether or bis-chelating amines to greatly accelerate ortho-lithiation. These considerations were derived from a seminal publication of Bauer and Schleyer,[ 5 ] which implicated a deficiently solvated n-butyllithium dimer as an intermediate in the ortho-lithiation reaction. No such clear demarcation exists for the X–Li exchange, thought to proceed through an ‘ate’ complex,[ 6 ] which is either an intermediate or the transition state for the exchange (Scheme [1]).

Zoom Image
Scheme 1 ‘Ate’ intermediate formation during the X–Li exchange

In the reactions described in this article, R2–X will be either an aryl bromide or iodide and R1 = n-butyl. At present, we can afford no insight into ethers accelerating the X–Li exchange reaction. Indeed, in our studies several exchanges have been efficiently carried out in neat cyclohexane media, but these were not as rapid. Flow investigation of the X–Li exchange in microreactors has been ongoing for a number of years. Yoshida et al. in particular have published several papers on the exchange of polybrominated aryls[ 7 ] and in several cases, substituted aryl bromides.[ 8 ] Other researchers have also contributed specific studies of flow aspects of the X–Li exchange.[ 9 ]

To further study the potential of the Synthetron, several exchanges were examined, which for one reason or another are known to be problematic. One of the initial findings of the co-discovers of the exchange, Wittig[ 10 ] and Gilman­,[ 11 ] was that p-bromoanisole (p-BrA) could not be efficiently exchanged under conditions used at that time (ambient temperature, Et2O, 1 equiv of n-BuLi) due to intervention of a secondary ortho-lithiation reaction (Scheme [2]). It was separately discovered that phenyllithium could be used to ortho-lithiate p-BrA,[ 10 ] thereby supporting the contention that p-BrA could be ortho-metalated by an aryllithium. In the ensuing years, p-BrA exchange has been run at –78 °C, a temperature that suppresses the secondary ortho-metalation.[ 12 ]

Zoom Image
Scheme 2 Exchange of p-BrA illustrating the secondary product formation

Examination of the secondary metalation problem suggested to us that, if the X–Li exchange could be made to take place rapidly and completely, little p-BrA would remain to undergo the secondary metalation. In particular, under the extremely rapid conditions extant in the Synthetron, where residence time is low, there would be no opportunity for the p-LiA to ortho-lithiate any remaining p-BrA. Using our established ambient temperature, ether-doped cyclohexane media (containing measured equiv of ether) protocol, the exchange characteristics of bromobenzene (PhBr) and ultimately of p-BrA were studied. To accomplish this, both batch and flow investigations were completed using these neat hydrocarbon or ether-doped solvent systems. Final trapping and measurement of the lithio-intermediate(s) generated was performed by quenching with ClSiMe3 (ClTMS) followed by GC analysis (Scheme [3]). To maintain atom-economy, only a measured one equivalent of n-butyllithium was utilized. Ultimately, in both batch and flow processes, the determined ratio of p-TMSA/o-TMS-p-BrA would indicate the selectivity of the exchange.

Zoom Image
Scheme 3 Analytics for the X–Li exchange of p-BrA

Table 1 X–Li Exchange of p-BrA in Ether-Doped Cyclohexane Media; Batch Studiesa

Entry

Promoter (equiv)

Time (min)

M

Relative % (GC)a

Anisole

p-BrA

p-TMSAb,c

p-n-BuA

4-Br-2-TMSA

 1

THF (0.25)

30

0.5

 4.5

16.5

66

1

 9

 2

THF (0.75)

60

0.5

 9.5

12

68

1

 6

 3

THF (1.0)

15

0.5

 9

 2

84

0

 9

 4

THF (1.0)d

15

0.5

 9

 2

78

0

 9

 5

MTBE (1.0)

15

0.5

34

 4.5

41

0.5

14.5

 6

none

15

0.5

20

26

49

0.5

 2

 7

THF (0.25)

30

1

 3

12

62

2

17.5

 8

THF (1.0)

30

1

10

 0

67

0

13.5

 9

MTBE (1.0)

15

1

 6.5

 2

68.5

1.5

15

10

none

15

1

1.5

27.5

69

0.5

0

a p-BrA, n-BuLi (1:1), 20 °C.

b ClTMS/GC analysis.

c Maximum EoE for the reaction.

d n-BuLi (2.0).

Study of batch process exchanges revealed that a reasonably high yield could be achieved while maintaining a satisfactory selectivity ratio (exchange product/secondary metalation product) using short reaction times (usually <30 min), ambient temperature, and THF–cyclohexane media (Table [1]). Adjustment of THF loadings and time periods [only the times of the highest extent-of-exchange (EoE) values are recorded in Table [1]] brought higher EoE values, but the selectivity ratios remained relatively low (Table [1, entries] 7 and 8). Interestingly, 1:1 exchange in neat hydrocarbon solvent (entries 6 and 10) afforded impressive selectivity, but the yields were low. The highest yield (84%, entry 3) had a selectivity ratio of 9.3:1, which was still deemed unsatisfactory.

In direct contrast to the results of the batch runs, a few test runs employing similar media ratios and conditions using the Synthetron afforded the TMS product from the exchange in a GC yield of 91% with ca. 1% of the product from the secondary reaction being detected (Table [2]).

Table 2 Synthetron-Facilitated X–Li Exchange of p-BrA

Entry

THF (equiv)

Reactor temp (°C)

Flow rate through reactor (mL/min)a

Relative % (GC)b

Anisole

p-BrA

p-TMSA

p-n-BuA

4-Br-2-TMSA

1

1.0

22

{100:100}

0.6

18.7

74.5

0.3

2.9

2

1.0

22

{75:75}

1.0

15.8

71.4

1.4

3.2

3

1.0

22

{125:125}

0.5

21.4

74.8

0.3

2.8

4

neat

14

{175:175}

0.8

3.9

91.5

0.9

0.9

5

2.0

12

{175:175}

0.6

4.6

91.1

0.2

1.1

a The first bracketed number represents the flow rate for the substrate solution, and the second represents the flow rate for the n-BuLi solution.

b GC analysis of functionalized (TMS) derivative.

A high flow rate was utilized in accordance with our initial ‘quick reaction’ hypothesis. We likely would not have achieved this level of success with the Synthetron so quickly had we not had the preliminary results from the batch process studies to guide us. It can be seen in Table [2] that both neat THF and about a 50:50 equivalent mixture of THF–cyclohexane with a very high flow rate were equally supportive of producing 91% yield of the exchange product. The 50:50 mixed media also served to avoid quantities of n-BuA formation, the product from the SN2 reaction of the n-butyl bromide with the p-LiA formed in the initial exchange.

Zoom Image
Scheme 4 Formation of a triarylcarbinol using two Synthetrons in sequence

To further our demonstration of the capabilities of the Synthetron, a meso-scale synthesis was undertaken. The p-BrA/o-Li-p-BrA was again generated in the reactor and fed directly into a second reactor into which was simultaneously fed a solution of benzophenone (Scheme [4]). Initially it was decided to use the best exchange (Table [2, entry] 4) results, but we knew we would need more THF in the benzophenone derivatization step due to the carbinol product precipitating in cyclohexane. The triaryl carbinol product was isolated in 80% yield (23 g) in 30 seconds. This would translate to 2.77 kg of carbinol product being produced in an hour.

In addition to these highly efficient production abilities, insight into the occurrence of the secondary ortho-lithiation of p-BrA afforded by p-LiA has been gained. The ortho-H acidity of p-BrA is greater than that of anisole and therefore p-BrA can be conveniently ortho-lithiated with o-lithiodimethylaniline[ 13 ] whereas anisole itself cannot. Wittig originally observed that p-BrA could be ortho-lithiated at ambient temperature using phenyllithium.[ 10 ] Moreover, when 2,4-dibromoanisole is treated with one equivalent of n-butyllithium, only the ortho-bromine is exchanged.[ 14 ] We conclude that the ortho-position is far more reactive, whether the ortho-substituent is Br or H, than the para-position in p-BrA. A recent paper has shown that the Grignard exchange of 2,3,5-tribromoanisole exchanges the 2- and 3- positions equally, whereas the 5-position remains untouched.[ 15 ]

A second insight gained is that exchanges of PhBr and p-BrA are relatively slow in comparison to that of m-bromoanisole (m-BrA). Analysis was performed by derivatization of the lithio-intermediate(s) with ClTMS and subsequent comparison of relative GC spectral intensities. All samples were analyzed after short reaction times because the main goal of this study was to compare relative rates of reaction, not to maximize EoE. Furthermore, these short reaction times served the purpose of minimizing problematic secondary reactions, which would skew the GC and NMR results. When a 1:1:1 mixture of PhBr/ m-BrA/n-BuLi was exchanged in neat cyclohexane in a batch process, a more than 3:1 ratio of exchange of m-BrA was observed after 20 minutes. When a 1:1:1 mixture of PhBr/p-BrA/n-BuLi was similarly studied in neat cyclohexane, a slight excess of the TMS product (ca. 1.7:1) via exchange of p-BrA was found. When this same set of exchanges was performed in cyclohexane containing one equivalent of THF, exchange of PhBr was slightly faster (ca. 1.5:1) than p-BrA, with m-BrA remaining the fastest (>2:1 relative to both PhBr and p-BrA). In cyclohexane, as expected, the EoE’s were lower over the time periods examined. When these same mixtures were fed through the reactor, for both neat cyclohexane media and THF–cyclohexane­, PhBr exchanged marginally faster than p-BrA (Scheme [5]). The two reactants were fed into the Synthetron S3T1 reactor at 150 mL/min each when THF was employed. In neat cyclohexane, a slower 100 mL/min feed rate was used to allow a sufficient EoE for both substrates for viable analytical study.

Zoom Image
Scheme 5 An outline of the analytics for competitive exchange

It is our conclusion that the relative rates of exchange of these substrates are m-BrA, (o-BrA) > PhBr ≥ p-BrA. Therefore, the reason that p-BrA undergoes the observed secondary ortho-lithiation lies not only in the enhanced reactivity of the ortho-position of p-BrA, but also in the relatively slow exchange of the p-Br substituent. We should state here that by lowering reaction temperature, the secondary ortho-lithiation rate is suppressed relative to that for the exchange.

Further inquiry into the secondary metalation phenomenon led us to investigate the exchange of p-iodoanisole (p-IA). This exchange in batch studies provided an average of 95% maximum EoE in 5 minutes using three different media (Table [3]). Virtually no secondary reactions, neither the ortho-lithiation nor the n-butylation products were detected in the first 5 minutes, but the purity profiles deteriorated significantly within the first hour. These results suggest that an iodo substituent undergoes exchange much more rapidly than a bromine substitutent. Likely, this is due to the much greater size of iodine and to the weaker aryl C–I bond. This greater exchange reactivity for iodine was demonstrated again by a competition experiment. A 1:1:1 mixture of p-IA/p-BrA/n-BuLi was examined in a batch exchange. After 5 minutes, no p-IA remained, only p-BrA. Exchange of p-IA in the Synthetron (Table [4]) in neat THF (Table [4, entry] 1) afforded a 77% EoE with some p-IA (3.7%) remaining and 13.5% of the butylated product, n-BuA, being generated, as well as 2.5% of the secondary ortho-product. When a reactor p-IA exchange was performed with less THF at a faster pumping rate as well as 10 °C warmer (entry 2), no secondary product was found and butylated product was suppressed, yet 8% of starting material remained. The butyl iodide is certainly more prone to aryllithium attack than the butyl bromide, thereby making the unwanted butylated anisole more prevalent in the Synthetron due to superior mixing, especially when pumped slower and in neat THF. These p-IA results serve to illustrate the variability of reactor vs. batch comparisons where greater reactor parameter fine-tuning is necessary. Once fine-tuning has served to attain complete exchange, this process could be scaled. The batch process could not be scaled efficiently.

Table 3 Five-Minute Batch Data for p-IAa

Entry

Promoter (equiv)

Relative % (GC)b

Anisole

p-IA

p-TMSA

p-n-BuA

4-I-2-TMSA

1

none

0.4

0.0

96.0

0.0

< 0.5

2

THF (1.0)

1.5

0.0

94.0

0.0

< 0.5

3

MTBE (1.0)

0.5

0.0

95.0

0.0

< 0.5

a p-IA, n-BuLi (1:1); 2.0 M n-BuLi added to 0.5 M p-IA, 22 °C.

b GC analysis of ClTMS-quenched solutions.

Table 4 Reactor Data for p-IA

Entry

THF (equiv)

Temp of reactor (°C)

Flow rate through reactor (mL/min)a

Relative % (GC)b

Anisole

p-IA

p-TMSA

p-n-BuA

4-I-2-TMSA

1

neat

 2

{125:125}

1.4

3.7

76.6

13.5

2.5

2

2.0

12

{175:175}

1.4

7.7

84.2

 5.3

0

a The first bracketed number represents the flow rate for the substrate solution, and the second represents the flow rate for the n-BuLi solution.

b GC analysis of ClTMS-quenched solutions.

These studies reflect the value of the Synthetron over batch exchanges when fast, competing reactions are involved. For p-BrA, a high ratio of exchange to secondary reaction product(s) was achieved, an exchange ratio far superior to that achievable through batch processes at these concentrations and temperature. That these results were attained at slightly below ambient temperature indicates that quantities, even large quantities, of desired derivatives of p-LiA could be prepared economically in rapid fashion. Additionally, some insights into relative rates of exchange of PhBr, p-BrA, m-BrA, and p-IA have been realized.

All research chemicals, including anhyd solvents used for reactions, were supplied by Aldrich Chemical Co. and used as received. GC analysis was performed on an Agilent 6850 instrument equipped with a FID detector and GC-MS analysis was performed on an Agilent­ 5973 MSD instrument containing a 6890 N Network system equipped with an FID detector. NMR spectra were performed on a JEOL Eclipse ECA 500 MHz instrument. Melting points were determined using a Mel-Temp® apparatus.

Precautions to exclude oxygen and/or moisture were implemented throughout all processes, save water quenches. All glassware was oven-dried, N2-purged and fitted with septa under a slight positive pressure of N2 before and during any reactant charges; needles and syringes were accordingly handled, including N2 purge prior to any reactant charge or aliquot sampling. No special handling other than the above was necessitated during reactor studies.


#

Batch Studies; General Procedure

Batch studies were performed on an 8 mmol scale. Flasks were charged with substrate and diluted to 2 M (unless otherwise noted; see Tables 1, 3) with cyclohexane and doping solvent (if applicable). For competitive runs, defined as those with more than one substrate, 8 mmol of each substrate was added to the flask and cyclohexane was added to bring the volume to 4 mL (2 M). The substrate(s) solution was placed in a water bath and n-BuLi (4 mL, 8 mmol) was added dropwise over a period of 45 s. Aliquots were taken by withdrawing 0.3 mL from the reaction mixtures; these were added to vials containing an excess of 2 M ClTMS in cyclohexane (0.5 mL). After 30 min, the vial solutions were quenched with sat. aq Na2CO3 (ca. 2 mL) and MTBE (ca. 2 mL to dilute properly for GC) was added. After vigorous shaking, the organic layer was removed and analyzed by GC and/or GCMS and/or 1H NMR (for TMS integration comparison only). Identities of GC and NMR peaks were confirmed by spiking with authentic samples. Competitive reactions were spiked with products from single-substrate reactions to confirm peak identity. Single-substrate results are contained in Tables 1 and 3, whereas competitive results are described in the body of this paper.


#

Synthetron Studies; General Procedure

Substrate solutions were prepared identical to the batch studies and kept under a slight positive pressure of N2 during all syringe pump withdrawals. A 25 mL solution of substrate(s) (50 mmol of a single substrate or 50 mmol of each of the two competing substrates, 2.0 M) in cyclohexane or cyclohexane–THF, was prepared and reacted with commercial 2.0 M n-BuLi in cyclohexane (25 mL). Reagent solutions were delivered through Teflon tubing using independent SYR-2200 dual programmable syringe pumps obtained from J-KEM Scientific. The single reactant or mixture of two reactants was fed into the Synthetron S3T1 reactor[ 1 ] at 100 mL/min each when in neat cyclohexane while 150 mL/min used when THF was employed [(to provide more EoE (see text)] at 12 °C (chilled with a circulating chiller). The exit streams were directed into 250 mL round-bottomed flasks (immersed in ice baths) containing excess ClTMS (12 mL, 120 mmol) diluted to 60 mL (2 M) with cyclohexane. After 30 min, the solutions were quenched with sat. aq Na2CO3 (ca. 15 mL) and a sample of the organic layer was diluted and was analyzed by GC and/or GCMS and/or crude NMR.


#

High-Output Synthesis of (4-Methoxyphenyl)diphenylmeth­anol in the Synthetron

A 50 mL solution of p-BrA [12.7 mL, 100 mmol (2.0 M)] in THF was prepared and reacted with commercial 2.0 M n-BuLi in cyclohexane (50 mL). Reagent solutions were fed at 100 mL/min each into the Synthetron S3T1 reactor at 12 °C. A combined flow rate of 200 mL/min in the first reactor provided a residence time of 0.034 s for the Br–Li exchange step. The exit stream was directed into a second Synthetron S3T1reactor[ 1 ] at 12 °C that was simultaneously fed a 3 M solution of benzophenone (50 mL solution prepared with 27.6 g, 150 mmol) in cyclohexane (containing 15 mL of THF for solubility) at a flow rate of 66.7 mL/min (utilizing a Büchi pump, Model C-610). Only 33.3 mL of the prepared 50 mL benzophenone solution was added to ensure essentially a 1:1:1 ratio of n-BuLi/ p-BrA/benzophenone. A combined flow rate of 266.7 mL/min in the second reactor provided a residence time of 0.026 s for the nucleophilic addition step and an overall production time of 30 s. The exit stream from the second reactor was directed into a flask containing brine (50 mL). The biphasic-quenched reaction mixture was transferred to a separatory funnel where the aqueous layer was separated and back-extracted with MTBE (25 mL). The combined organic layers were washed with brine (20 mL), dried (Na2SO4), and concentrated in vacuo to provide a viscous oil. Overnight trituration at r.t. with cyclohexane (10 mL) and filtration afforded >98% (GC) pure (4-methoxyphenyl)diphenylmethanol (22.7 g, 80%) as a white solid; mp 80.2–81.4 °C (Lit.[ 16 ] mp 82 °C).

1H NMR (500 MHz, CDCl3): δ = 2.8 (s, OH), 3.8 (s, 3 H), 6.8 (d, J = 9.2 Hz, 2 H), 7.2 (d, J = 9.1 Hz, 2 H), 7.3 (complex m, 10 H).

13C NMR (125 MHz, CDCl3): δ = 55.4, 81.8 113.3, 127.3, 127.9, 128.0, 129.3, 139.3, 147.2, 158.8.


#
#

Acknowledgment

Support of this research was under the auspices of NSF, CHE 070021. Support of our preliminary studies was provided by the Petroleum Research Fund, PRF 42090-B1. Additional support was received from the Western Kentucky University Research Foundation. Continual consultation with Dr. Jeffrey C. Raber, President of KinetiChem, Inc., re Synthetron™ reactor parameters was greatly appreciated. Finally, thanks are due to Steven Bush, Brian Jones, and Maria DiLoreto for their laboratory assistance.



   Permissions and Reprints All articles of this category
Zoom Image
Scheme 1 ‘Ate’ intermediate formation during the X–Li exchange
Zoom Image
Scheme 2 Exchange of p-BrA illustrating the secondary product formation
Zoom Image
Scheme 3 Analytics for the X–Li exchange of p-BrA
Zoom Image
Scheme 4 Formation of a triarylcarbinol using two Synthetrons in sequence
Zoom Image
Scheme 5 An outline of the analytics for competitive exchange