Borohydride-polyethylene glycol monomethyl ether derivatives: chemo- and stereoselectivity studies under phase transfer conditions

Borohydride-polyethylene glycol monomethyl ether derivatives: chemo- and stereoselectivity studies under phase transfer conditions

Reactive & Functional Polymers REACTIVE & FUNCTIONAL POLYMERS 33 (1997) 61-69 Borohydride-polyethylene glycol monomethyl ether derivatives: chemo-...

940KB Sizes 0 Downloads 1 Views

Reactive & Functional

Polymers

REACTIVE & FUNCTIONAL POLYMERS

33 (1997) 61-69

Borohydride-polyethylene glycol monomethyl ether derivatives: chemo- and stereoselectivity studies under phase transfer conditions James R. Blanton * Department Received

16 November

ofChemistry, The Citadel, Charlesron, SC 29409, USA

1995; revised version received 25 October

1996; accepted

18 February

1997

Abstract Borohydride-polyethylene glycol monomethyl ether (borohydride-PEGMME) derivatives were prepared by allowing low molecular weight polyethylene glycol (MW = 350) to react with either LiBH4, NaBH4, or KBH4. After comparing the reactivity of these borohydrides toward aldehydes, ketones, halides and esters, it was established that the LIBH4PEGMME derivative was the most active and the least selective, the NaBI&-PEGMME derivative exhibited moderate reactivity and selectivity, and the KBI-L-PEGMME derivative was the most selective and the least reactive. Additionally, the NaBb-PEGMME derivative was further functionalized by the addition of a chiral alcohol or chiral amino alcohol to the borohydride ion. In all cases, the activity of the phase transfer moiety increased and the enantiomeric induction that resulted from the reduction of acetophenone was low to moderate (23-40% ee). An attempted elucidation of the active reagent indicated that a significant amount of unreacted sodium borohydride was present in the reaction mixture. The presence of this achiral reagent probably accounts for the more modest levels of chiral induction that were observed. Keywords: Borohydride-polyethylene glycol monomethyl ether derivative; Chemoselectivity; transfer catalysis; Chiral catalyst; PEGMME-borohydride

1. Introduction In general, phase transfer catalysis has been widely accepted in industry and academia as a viable synthesis tool as depicted by the patent literature and the number of journal articles devoted to the topic [ 1,2]. In the past, a compound that has seen a great deal of use in this capacity is polyethylene glycol (PEG). More recently, there has been renewed interests in using PEG derivatives as a means of solubilizing reagents and catalysts because of the advantages this moiety imparts on the process. Namely, the reactions may be conducted in solution using the PEG *Tel.:

+I (803) 953 7789; e-mail: [email protected]

1381-5148/97/$17.00 0 1997 Elsevier Science B.V. All rights reserved. PII Sl381-5148(97)00017-5

Stereoselectivity;

Phase

derivative as a soluble species and recovered at the end of the process by precipitating the polymer. Specifically, PEG derivatives have been used as supports for the synthesis of tetrapeptides in yields that exceed the more traditional solid-phase synthesis approach [3]. Additionally, polyethylene glycol monomethylated ethers were used as solubilizing agents in the preparation of chiral catalysts used in the asymmetric dihydroxylations of olefins [4]. It should also be noted that polymeric reagents using PEG supports have been developed to exploit these same advantages. To illustrate this point, it has been shown that sodium borohydride-polyethylene glycol derivatives (PEG-NaBH4) are convenient and useful for reducing aldehydes, ketones, alkyl halides,

62

J.R. BlantonIReactive

& Functional Polymers 33 (1997) 61-69

esters and epoxides under both homogeneous and phase transfer conditions [5-71. In addition to these uses, studies conducted under homogeneous conditions indicate that PEG derivatives prepared from sodium borohydride yield a moderate degree of chemo- and stereoselectivity when used to reduce aldehydes and/or prochiral ketones [8]. On a similar note, other groups have reported that borohydrides supported on polymers containing chiral functionalities or borohydrides derivatized by the addition of a chiral compound have been successfully used in asymmetric reductions of prochiral ketones [9-161. By using such an approach reductions involving moderate to good (50-75% ee) levels of enantiomeric induction were obtained [ 17,181. Recently, we reported that PEGMME-borohydride derivatives could be successfully employed under phase transfer conditions to quantitatively reduce ketones and alkyl halides at room temperature and at 85°C [7]. While there was an expected increase in activity at the elevated temperatures, the rates of reaction were also acceptable when the reactions were conducted at room temperature. Based on the possibility that a more selective species could be developed for use under such mild conditions, it was felt that a more comprehensive investigation utilizing a series of polyethylene glyco1 monomethyl ether-borohydride (PEGMMEborohydride) compounds under phase transfer conditions would be of interest. The following report describes the results of a more comprehensive investigation utilizing a series of polyethylene glycol monomethyletber-borohydride (PEGMME-borohydride) compounds derived from LiBH4, NaBI& and KBH4 under phase transfer conditions. 2. Experimental 2. I. Reagents and instruments All starting materials were of reagent grade and purchased from the Aldrich Chemical Company. All solvents were purified prior to use by

distillation. The gas chromatograph data was obtained using a Varian model 3400 gas chromatograph equipped with a 6-ft Carbowax column (lo%), TCD detector, and variable temperature capabilities. NMR data was obtained using a Varian model 360A nuclear magnetic resonance spectrometer. The infrared analyses were conducted through the use of a Perkin-Elmer model 1430 instrument. 2.2. Sodium borohydride-PEGMME preparation The sodium borohydride-polyethylene glyco1 methyl ether 350 (PEGMME 350), 1, was prepared by placing 0.13 g (3.3 mmol) of sodium borohydride in a 50-ml round-bottomed flask equipped with a side-arm, reflux condenser, magnetic stirrer bar, and septa. The flask was thoroughly purged with nitrogen and 1.3 g of polyethylene glycol monomethyl ether (M, = 350, 3.2 mmol) in 10 ml of toluene was added. The amounts of reagents were kept close to a 1 : 1 molar ratio as a means of insuring there would be sufficient hydride sites available for further functionalization or for use in reduction reactions. While vigorously stirring at 80”, the reaction was allowed to proceed until all hydrogen evolution had ceased (ca. 2 h). The yield of hydrogen was determined by a variation of a published procedure to ascertain the volume of hydrogen produced [ 191. From this information the average ratio of PEG chains to borohydride unit was found to be 1 : 1. Signals in the infrared spectrum that were consistent with boron to hydrogen stretches and boron to oxygen stretches were observed. These signals were also consistent to those previously reported for similar compounds [ 161. However, the ’H-NMR spectra gave markedly different results for the material that was soluble in toluene than for the suspended material. IR (mull): 2250, 1350, 1070 cm-‘; ‘H-NMR (soluble, ds-toluene): 6 3.8 (PEG), 1.75 (NaB&), 0.45 (NaBHd), -0.90 (NaBI&), -2.25 (NaBH4); ‘H-NMR (soluble, D20): 6 3.8 (PEG), 1.75

J.R. Blanton/Reactive

(NaBHd), 0.45 (NaBHb), -0.90 (NaBH4).

& Functional Polymers 33 (1997) 61-69

(NaBHd), -2.25

2.3. General reduction procedure After phase transfer catalyst, la-c or 2a-c, was prepared, the mixture was allowed to equilibrate at the temperature at which the reaction was to be conducted. To the flask was added (via syringe) ca. 1 mmol of the substrate in 10 ml of toluene, ca. 0.5 mmol of a hydrocarbon reference was added as an internal standard, and the mixture was vigorously stirred during the course of the reaction. The reaction was monitored by GC until no starting material remained. The products were identified by comparing the GC retention times to those of the known compounds. Additionally, the yields of the products were based on the relative areas of the products to that of the internal standard. Because it was previously shown that the disappearance of the substrate in similar phase transfer reactions exhibits first-order kinetics, the rate constants were determined using a linear-least-squares regression program. 2.4. Preparation of chiral derivatives After the initial 2 h reaction period in which 3.2 mmol of the chiral substrate (endo-bomeol, (S)-2-methyl-1-butanol, or (S)-2-amino- 1-butanol) was added in 5 ml of toluene (via syringe) to the reaction mixture to form 2a-c, respectively. The mixture was stirred vigorously until all hydrogen evolution had ceased (ca. 2 h). In all cases the ratios of the moles of hydrogen evolved to the moles of borohydride were 1 : 1. At this point the reductions were carried out as described in the previous section using la-c as the reducing agent. IR (mull) 2a: 2250, 1350, 1070 cm-‘. IR (mull) 2b: 2250, 1350, 1050 cm-‘. IR (mull) 2c: 3350,2250, 1350, 1050 cm-‘. ‘H-NMR (soluble, D20) 2a: 6 3.45 (PEG), 3.2 (overlapped with PEG, no integral), 1.8 (NaBl&), 1.0 (s, 9H), -2.1 (NaB&). 0.5 (NaBH4), -0.85 (NaBb), ‘H-NMR (soluble, D20) 2b: S 3.55 (PEG), 3.25

la was formed,

63

(overlapped with PEG, no integral), 2.5 (m, 2H), 1.75 (NaBHb), 0.9 (d, 6H), 0.45 (NaBHd), -0.9 (NaB&), -2.25 (NaBH4). ‘H-NMR (soluble, D20) 2c: S 3.5 (PEG), 3.2 (overlapped with PEG, no integral), 2.3 (m, lH), 1.8 (NaBH& 1.4 (m, 2H), 1.O (t, 3H), 0.5 (NaBH4), -0.9 (NaBI&), -2.2 (NaBH4). 2.5. Preparative scale reaction The same procedure described previously was employed on a 10 mmol scale. The initial suspension of chiral reducing agent was prepared by allowing 0.38 g (10 mmol) of sodium borohydride to react with 3.50 g (10 mmol) of polyethylene glycol monomethyl ether in 40 ml of toluene. After the hydrogen evolution was complete, 0.89 g ( 10 mmol) of (S)-2-amino- 1-butanol in 10 ml of toluene was added. After the reaction was complete, 1.20 g (10 mmol) of acetophenone was added to the suspension and the reaction was allowed to proceed for 16 h at 23°C. The reaction was quenched by adding 25 ml of water to the mixture with vigorous stirring. The layers were separated and the organic layer was washed with 2 x 25 ml portions of water. The toluene layers was then separated and dried over MgS04. The organic mixture was concentrated to ca. 10 ml and transferred to a microware apparatus to isolate the product. The remainder of the toluene was removed via fractional distillation. The crude product was purified by vacuum distillation to yield 0.89 g (73%, 50-55°C 1 mmHg) of product. IR (neat): 3500, 3100, 1500, 1450, 1100 cm-‘. ‘H-NMR (CDC13): 6 7.6 (s, 5H); 5.2 (q, 1H); 2.2 (s, 1H); 1.8 (d, 3H). Asymmetric induction: 33% ee. 2.6. Determination of enantiomeric excess The enantiomeric excess was determined by the method described in the literature utilizing (-)-menthyl chloroformate to form diastereomerit carbonates [ 181. A 10% carbowax column was substituted for the columns described in the

J.R. BlantonIReactive & Functional Polymers 33 (1997) 61-69

64

Table 1 Derivatives containg chiral appendages a Chiral substrate Blank’ endo-(-)-borne01 (S)-2-methyl-1-butanol (S)-2-amino- 1-butanol (S)-2-amino- 1-butanol

Rate k x lo5 (s-l)

Enantiomeric excess (f6% ee) b

5.1 9.5 11.8 40.3 N/A

0.4 22 26 40 33d

1 la; M= Na lb; M= Li lc; M = K

la

-

R*OH [A”dG$3H4_x_y

Na

IR*oly

a The temperature for these reactions was maintained at 25 f 2°C. b The uncertainty for the technique was determined by taking the standard deviation of ten runs using a racemic mixture of the product alcohol. c The blank run was conducted using the PEGMME-borohydride derivative containing no chiral side group. d This reaction was conducted on a preparative scale and this is the % ee of the purified product.

2 2a; R* = endo-(-)-bornoxide 2b; R* = (S)-2-methyl-I-butoxide 2c; R* = (S)-2-amino-1-butoxide Fig. 1. Catalyst preparation. OH

report. In some cases, baseline separations were not obtained and this is reflected by the high uncertainties reported in Table 1. 3. Results and discussion

A 1a-c

Y

W

-Br Y = -H, -R, or -OR

3. I. General chemoselectivity

Fig. 2. General reaction.

During the course of this study, the relative abilities of the PEGMME-borohydride derivatives prepared from NaBH4, LiBI& and KBH4, (la-c), in reducing aldehydes, methyl ketones, esters and alkyl halides were investigated. The PEGMME-borohydride derivatives were prepared by allowing PEGMME (Mw = 350) to react with the appropriate borohydride for 2 h at SK, see Fig. 1. After allowing the mixture to cool to 25”C, the reactions were conducted under solid-liquid phase transfer conditions using toluene as the organic solvent, see Fig. 2. The progress of the reactions were monitored by GC analysis and the data obtained from these reactions have been summarized in Tables 2-4. As can be seen from the information in Tables 2-4, the general activity trend for these The basis for this trend derivatives is lb>la>lc. is the pseudo-first-order rate constant based on the disappearance of the substrate which shows the LiBHd-PEGMME derivative is substantially more effective in reducing all the substrates. For

Table 2 PEGMME-NaBb

reductions a,b

Substrate

k x IO5 (SC’)

Methyl benzoate Methyl benzoate I-Bromododecane 1-Iodooctane 1-Bromododecane Butanal 2-Pentanone Hexanal 2-Heptanone Heptanal 2-Octanone Octanal 2-Nonanone Benzaladehyde Acetophenone

2.1 f 0.5c 1.5 f0.4 5.3 f 0.6 18 I!Z2.0c 212 f 14 1.46 f 0.37 203 rt 14 1.96 f 0.29 248 f 25 8.10 ??0.55 209f20 6.72 ??0.25 183 * 15 6.71 & 0.81

0

a The reactions were conducted under the conditions described in the experimental section. b The rate constants were obtained as the averages of four separate runs with the standard deviations for these runs expressed as the uncertainties. ’Because of the slower reactions noted at 25°C. additional runs were carried out at 85 f 2°C to see if the rate of reaction increased to a more useable level.

J.R. Blanton/Reactive

Table 3 PEGMME-LiB&

reductions

& Functional Polymers 33 (1997) 61-69

a,b

Substrate

k x lo5 (s-l)

Methyl benzoate Methyl benzoate 1-Bromododecane 1-1odooctane 1-Bromododecane Butanal 2-Pentanone Hexanal 2-Heptanone Heptanal 2-Octanone Octanal 2-Nonanone Benzaladehyde Acetophenone

0.65 & 0.4 18 Zt 2.0c 2.1 Yko.5 7.8 & 1.0 22 + 2.0c 1905 * 40 181OYk 120 1870*40 1745 & 360 1770 + 50 1770 f 100 18905 50 1570f200 1970f25 1770 * 370

a The reactions were conducted under the conditions described in the experimental section. b The rate constants were obtained as the averages of four separate runs with the standard deviations for these runs expressed as the uncertainties. c Because of the slower reactions noted at 25°C. additional runs were carried out at 85 f 2°C to see if the rate of reaction increased to a more useable level.

Table 4 PEGMME-KBH4

reductions

a,b

Substrate

k x lo5 (s-l)

Methyl benzoate Methyl Benzoate 1-Bromododecane 1-Bromododecane

0 OC 0 OC 0 OC 27.3 * 1.0 0.397 f 0.011 33.9 * 2.0 0.474 k 0.037 40.5 f 2.5 0.646 f 0.050 33.5 k 2.5 0.544 f 0.025 33.4 ??2.0 0.484 f 0.070

I-Iodooctane 1-Iodooctane Butanal 2.Pentanone Hexanal 2-Heptanone Heptanal 2-Octanone Octanal 2-Nonanone Benzaladehyde Acetophenone

a The reactions were conducted under the conditions described in the Experimental section. b The rate constants were obtained as the averages of four separate runs with the standard deviations for these runs expressed as the uncertainties. c Because of the slower reactions noted at 25°C additional runs were carried out at 85 f 2°C to see if the rate of reaction increased to a more useable level.

65

instance, lb reduces aldehydes and ketones quantitatively to the corresponding alcohol in 5-10 min. Additionally, lb reduces both alkyl halides and esters at 25°C to the extent that both reactions are marginally viable. Furthermore, if the reaction temperature is increased to WC, the reactions are synthetically practical processes. Changing the catalyst to the PEGMME-NaB& derivative, la, increases the chemoselectivity to exclude esters (though a small amount of reduction is observed at elevated temperatures) with a concurrent increase in the amount of time required for quantitative reduction of the aldehydes and ketones. Additionally, only aldehydes are effectively reduced by the PEGMMEKBH4 derivative, lc, while ketones exhibit only a marginal degree of reactivity under the conditions of the reaction. As expected, the selectivities of the borohydride derivatives increased as the reactivities decreased. 3.2. Aldehydes vs. methyl ketones The ready availability of compounds that are aldehyde/methyl ketone pairs made this an attractive system for use as a competitive reaction to study any selectivity the borohydride derivatives, la-c, exhibited between aldehydes and ketones. In addition to the selectivity between the functional groups, the effect of increasing the size or changing the nature of the carbon group attached to the other site on the carbonyl group was also studied. Previously, it was found that derivatives prepared from sodium borohydride and polyethylene glycol (PEG) exhibited a 6 : 1 preference for benzaldehyde during a competitive reduction of a mixture of benzaldehyde and acetophenone. This was a substantial increase from the 1.2 : 1 preference for benzaldehyde that sodium borohydride exhibited [8]. In contrast to these findings, the data in Table 5 indicates that PEGMME-borohydride derivatives la, lc yielded much higher selectivities in favor of the aldehyde while lb produced very little selectivity. For instance, lcwas more selective with reactivity ratios between 62 and 72 to 1, la exhibited

J.R. Blanton/Reactive

66

& Functional Polymers 33 (1997) 61-69

Table 5 Methyl ketone vs. aldehyde selectivity a

Table 6 Relative amounts of dissolved hydride

Aldehyde : ketone

PEGMME derivative

Reactivity ratio b

Butanal : 2pentanone

la lb lc

1.057 f 0.086

Hexanal : 2-heptanone

la lb lc

25.4 f 0.8 1.087 zt 0.209 71.9f5.3

Heptanal : 2-octanone

la lc

30.6 & 2.5 1.000~0.156 62.6 f 6.2

la lb lc

1.222kO.159 62.0 f 6.9

lb Octanal : 2-nonanone

Benzaldehyde : acetophenone

la

lb lc

28.5 f 2.8 68.6 f 1.7

31.0f2.1

27.6 zt 3.6 1.160f0.225 70.7 f 12.8

a The reactions were carried out under the conditons described in the experimental section. b The ratio was determined using the pseudo-first-order rate constants illustrated in Tables 2-4 with the uncertainites represented by the standard deviation of the reactions.

aldehyde to methyl ketone reactivity ratios between 25 and 3 1 to 1, and lb gave reactivity ratios near unity. However, when the results from the homologous increase in the size of this chain were compared, it was apparent that any effect due to the size of this chain was negligible at best. In fact, even when the nature of the side group was changed from an aliphatic group to an aromatic moiety no significant change in the selectivity was noted. Overall, these findings indicated that only features that dramatically affect the environment near the carbonyl group has an effect on the selectivity of the catalyst (e.g., the hydrogen of an aldehyde relative to the carbons of a ketone). The nature of the solvent undoubtedly made a difference in the selectivities displayed by la-c as compared to the selectivities reported in the literature. In the prior studies, the reactions using both the PEG-borohydride derivative and sodium borohydride were conducted under homogeneous conditions using polar solvents. In contrast, the

PEGMME-borohydride

% Soluble hydride a

Normalized values b

Li Na K

1ooc 2.11 1.04

47.4 1.00 0.493

a The amount of dissolved hydride in the organic layer was determined by separating the organic layer into a separate vessel after the catalyst was prepared and quenching the mixture with excess n-butyraldehyde. The moles of n-butanol formed after 24 h at room temperature was assumed to be equal to the equivalents of active hydride present. b The percentage of soluble hydride was normalized with respect to the amount of soluble PEGGMME-NaB& . ’Effectively all the solid had dissolved in the organic layer.

use of a less polar solvent seemed to increase the selectivity of the reaction. A possible explanation for this polarity/selectivity relationship could be the partitioning of the catalyst during the phase transfer process which limited the amount of active hydride present in the organic layer. If this were the case, the overall selectivity of the aldehyde relative to the methyl ketone could be attenuated due to the lesser amount of active hydride being available for reaction as compared to the soluble analogs. To test this idea, compounds la-c were prepared in the usual manner, but the toluene layer was removed from the solid suspension after the mixture was allowed to equilibrate at room temperature. To this was added an excess of n-butanal and the mixture was allowed to stir for 24 h at which time no additional reaction was noted. In this fashion the effective amount of active hydride in the organic layer was obtained. As illustrated by the data in Table 6, the amount of dissolved hydride varied with the PEGMME-borohydride catalyst. Using the NaB&-PEGMME, la, as the reference, the more reactive LiBHd-PEGMME, lb, contained 47.4 times more active hydride in the organic layer, while the less reactive KBI&-PEGMME, lc, contained 0.474 times as much active hydride in the organic layer. Certainly, solvent effects relative to the ion pairs must play a role in the

J.R. Blanton/Reactive

& Functional Polymers 33 (1997) 61-69

process, but this information also shows the ability of the phase transfer catalyst to cross between the phases plays an important part as well [20]. 3.3. Asymmetric syntheses Because of its moderate activity and chemoselectivity, la was chosen as the base phase transfer moiety from which to prepare chiral reducing agents. The chiral catalysts were prepared by allowing la to react with a chiral alcohol at 85°C for 2 h. At this time all hydrogen evolution had stopped and gas chromatographic analysis showed that no soluble, chiral substrate remained. After equilibrating to 25°C acetophenone was added via syringe and the mixture was allowed to stir for 16 h. The product mixture was treated as described in the literature to yield diastereomeric carbonates which could be analyzed by gas chromatography [21]. The asymmetric reduction (Fig. 3) of acetophenone using catalysts 2a-c yielded enantiomeric excesses in the range of 20-40%. As illustrated in Table 1, the results from the blank run gave the expected racemic product. Additionally, the uncertainty of the technique was determined by the standard deviation of the ten blank runs and was found to be f6% ee. When either endo-borne01 or (S)-2-methyl-1-butanol was used as the chiral appendage to form 2a and 2b, respectively, the degree of chiral induction were approximately the same when the uncertainty was considered. Additionally, the

67

rate constants were also similar. However, when (S)-2-amino-1-butanol was used to prepare the chiral catalyst, 2c, a significant increase in the rate constant occurred in addition to a substantial increase in the chiral induction relative to catalysts 2a and 2b. Initially, it was felt that the multidentate nature of the (S)-2-amino-1-butanol contributed to the increase in activity as a result of the increased number of electron donating groups on the borohydride moiety. The increase in chiral induction was further attributed to the greater influence of the stereogenic center of the amino alcohol as a result of it closer proximity to the borohydride ion. However, in all cases, the degree of chiral induction could only be described as moderate. While the 40% ee obtained from reducing acetophenone using 2c was consistent with previously reported values (and substantially better than the results obtained using 2a and 2b), it was only a marginally useful synthesis tool. As a means of determining the source of the problem, ‘H-NMR analysis was used to provide some possible answers for these rather modest results. Because the catalyst is solid, the actual identity of the previously reported PEG-and PEGMME-borohydrides was never completely elucidated in previous studies. For instance, the initial studies reported by Santaniello did not include any NMR data (carbon or proton) with the IR data that was used to support the identity of the reagent. Additionally, other groups have shown that B-H protons are sometimes difficult

h-c *

4b Fig. 3. Asymmetric

reductions.

68

JR. BlantonIReactive

& Functional Polymers 33 (1997) 61-69

to see by ’H-NMR [ 161. In spite of these limitations, a few pieces of the borohydride puzzle were solved using ‘H-NMR to investigate both the solid suspension and the soluble materials present in the toluene layer. From the ‘H-NMR data, it was estimated that ca. 50% of the sodium borohydride in the reactions forming 2a and 2b was not derivatized, and ca. 35% of the sodium borohydride remained unreacted during the formation of 2c. In general, all the reaction mixtures were found to be complex mixtures of unidentified borohydrides. Additionally, the fact that the PEGMMEborohydrides will be used under phase transfer conditions further complicates the identification process. Since a variety of borohydrides were present, it was reasonable to assume the partitioning of these species between the solid and liquid phases would be affected. To study this possibility, the solvent was carefully removed in vacua to yield a small amount of white solid which was stored under nitrogen. The solid was dissolved in ds-toluene and the ‘H-NMR spectrum obtained. From the data, a very small amount of unreacted sodium borohydride was observed in addition to a signal at 6 3.8 which corresponds to only the polyethylene glycol side chain, Since it has been shown that not all B-H ‘H-NMR signals are readily observable, one cannot automatically assume that all sites on the borohydride unit have reacted with the polymer. However, it is relatively safe to assume that multiple polymer chains reacting with a single borohydride ion has occurred to some extent because the solid that remained suspended in the organic layer also shows the presence of the polymer side chain. It is currently thought that the increased solubility of the soluble fraction is the result of two or more polyethylene glycol chains being present. Being more soluble in the organic layer, a species with multiple PEGMME groups would be an effective phase transfer catalyst whether or not it still possessed active hydrides. The remaining, insoluble material yielded no appreciable ‘H-NMR signals after an attempt was made to dissolve it in ds-toluene. This solid was then

dissolved in D20 and subjected to ‘H-NMR analysis after the slight evolution of hydrogen had stopped. The ‘H-NMR spectrum showed the presence of the polyethylene glycol side chain at 6 3.8 in addition to the strong signals at 6 1.75, 0.45, -0.90, and -2.25 that are characteristic of sodium borohydride. This finding is consistent with those previously reported for lactic acid/alcohol/borohydride systems in which it was more fully ascertained that such systems were complex mixtures of borohydrides of differing stoichiometries using 13C-NMR techniques [ 161. Based on a review of the ’H-NMR data, the increased activity and enantiomeric induction that was observed when 2c was used as compared to 2a and 2b was explained as follows. The mixture of borohydrides that comprised 2c contained a lesser amount of unactivated, achiral sodium borohydride than did 2a and 2b. This increased functionalization was the result of the multidentate nature of the amino alcohol starting material which undoubtedly allowed for reaction with multiple equivalents of borohydride ion. 4. Conclusion In summary, the results of this study indicate that useful chemo- and stereoselective phase transfer moieties can be prepared using PEGMME-borohydride derivatives. As expected, the degree of chemoselectivity varied with the borohydride starting material. The least chemoselective and most active derivative was lb, which rapidly effected complete reduction of aldehydes, ketones, and alkyl halides at room temperature. The most chemoselective and least reactive catalyst was lc, which was only effective in reducing aldehydes at room temperature. Overall, the most versatile catalyst was la, which exhibited a modest chemoselectivity between the aldehydes and ketones at lower temperatures while maintaining a moderate catalytic activity. Because of its moderate reactivity, la was further derivatized to yield chiral phase transfer catalysts which were successfully used in reducing acetophenone to study the degree of

J.R. Blanton/Reactive

& Functional Polymers 33 (1997) 61-69

enantiomeric induction. It was shown that enantiomeric excesses of 20-40% were possible depending on the identity of the chiral appendage. In order to achieve the 40% ee, a more multidentate-type group was required as illustrated by the data obtained using 2c as the chiral catalyst. In an attempt to understand why the moderate levels of asymmetric induction were not higher that what they were, ’H-NMR studies devoted to better elucidating the identity of the active hydride(s) were conducted. The ‘catalyst’ was observed to be an complex mixture of borohydrides with containing a variety of functionalized species and approximately 3550% of unreacted sodium borohydride. Additionally, such a finding might provide a viable explanation why similar phase transfer catalysts reported in the literature have induced enantiomeric excesses of the same magnitude. Overall, this study illustrates that PEGMMEborohydride systems provide reagents that can be used under phase transfer conditions to reduce a variety of functional groups. Furthermore, the successes outlined here more than justify further investigation into developing similar chiral catalysts that are more effective by reducing or eliminating the presence of the unreacted sodium borohydride. A possible solution to this problem, as illustrated by the results using 2c, would be to use chiral agents that possess more than one site that will bind to the borohydride ion. In this manner, more borohydride ion will be derivatized to contain a chiral appendage which should lead to greater enantiomeric inductions. This is an area that will receive a large degree of attention in future studies.

69

References [ll CM. Starks, C.L. Liotta and M. Halpem, Phase Transfer Catalysis. Chapman and Hall, New York, NY, 1994. PI M. Aglietto, E. Chiellini, S. D’Antone, G. Ruggeri and R. Solaro, Pure Appl. Chem., 60 (1988) 415-430; and references cited therein. [31 E. Santaniello, A. Fiecchi, A. Manzocchi and P Ferraboschi, J. Org. Chem., 48 (1983) 3074-3077. [41 E. Santaniello, P Ferraboschi, A. Fiecchi, P Grisenti and A. Manzocchi, J. Org. Chem., 52 (1987) 671-674. [51 J.R. Blanton and S.J. Kruzska, Microchem. J., 47 (1993) 120-126. 161 H. Han and K.D.Janda, J. Am. Chem. Sot., 118 (1996) 7632-7633. 171 J. Zhu and L.S. Hegedus, J. Org. Chem., 60 (1995) 58315837. 181 E. Santaniello, P. Ferraboschi, A. Fiecchi, P. Grisenti and A. Manzocchi, Gazz. Chim. Ital. 117 (1987) 701-703. [91 E. Chiellini, R. Solar0 and S. D'Antone, Polym. Sci. Tech., 24 (1984) 227-248; and references cited therein. 1101L. Deloux and M. Srebnik, Chem. Rev., 93 (1993) 763784; and refrences cited therein. Ull J.C. Briggs, P. Hodge and Z. Zhengpu, React. Polym., 19 (1993) 73-80. [121 M.J. O’Donnell, S. Wu and J.C. Huffman, Tetrahedron, 50 (1994) 4507-45 18. [131 M. M. Midland, Chem Rev., 89 (1989) 1553-1561. [141 H. Sakuraba, N. Inomata and Y. Tanaka, J. Org. Chem., 54 (1989) 3482-3484. 1151 A. Akelah and D.C. Sherrington, Polym. Commun., 24 (1983) 147-150. [161 G. Bianchi, F. Achilli and A. Gamba, J. Chem. Sot. Perkin Trans., (1988) 417-422. 1171 K. Adjidjonou and C. Caze, Eur. Polym. J., 30 (1994) 395398. [181 K. Adjidjonou and C. Caze, Eur. Polym. J., 31 (1995) 749754. 1191 Sodium Borohydride Digest, 3rd edition. Morton Intemational, Andover, Mass., 1995. [201 G.W. Gokel, K.A. Arnold, T. Cleary, R. Friese, V. Gatto, D. Goli, C. Hanlon, M. Kim, S. Miller, M. Ouchi, I. Posey, A. Sandler, A. Viscariello, B. White, J. Wolfe and H. Yoo, ACS Symp. Ser., 326 (1987) 24-37. [211 J.W. Westly and B. Halpem, J. Org. Chem., 33 (1968) 3978-3980.