Amine-directed hydroboration: scope and limitations.

Iodine activation induces intramolecular hydroboration of homoallylic and bis-homoallylic amine boranes with good to excellent control of regiochemistry compared to control experiments using excess THF*BH 3. Deuterium labeling and other evidence confirm that the iodine-induced hydroboration reaction of homoallylic amine boranes occurs via an intramolecular mechanism equivalent to the classical 4-center process and without competing retro-hydroboration. Longer carbon chain tethers result in lower regioselectivity, whereas the shorter tether in allylic amines results in a switch to dominant intermolecular hydroboration. Regioselectivity in THF*BH 3 control experiments is higher for the allylic amine boranes compared to the iodine activation experiments, whereas the reverse is true for homoallylic amine borane activation.

Introduction

Unsaturated amine borane complexes have been known for many years to be isolable molecules that easily survive exposure to the air at room temperature.(1) This stability can be attributed to the strength of the bond between nitrogen and boron in the Lewis acid−Lewis base complex and is consistent with prior reports that demonstrate conversion of alkenylamine boranes into isomeric hydroboration products upon heating (eq ).(1b) Like other simple hydroborations,(2) this process likely involves reversible dissociation to borane and the free Lewis base ligand (i.e., the amine), followed by a conventional intermolecular hydroboration and subsequent internal N−B bond formation to afford the cyclic borane complex 1. In related studies, thermally induced borane transfer has been demonstrated from a saturated amine borane to an unsaturated amine, as shown in eq using trimethylamine borane as the source of borane.(1c) Formation of 2 requires cleavage of the N−B bond at some stage prior to hydroboration and is most easily understood if this occurs by a simple dissociative event. Also consistent is the well-known dissociation of amine boranes at similar temperatures in the absence of alkenes.(3)

We could find no cases in the prior literature to suggest that unsaturated amine boranes are capable of hydroboration by an intramolecular mechanism that might occur below the temperature for N−B dissociation. Similar conclusions have been reached for the comparably stable unsaturated phosphines boranes.(4) These issues are relevant to the prospects for heteroatom-directed intramolecular hydroboration,(5) a problem in hydroboration chemistry that was recognized long ago.(6) Prior to our work, the only well-defined examples of heteroatom direction via bonding interactions involving Lewis basic electron pairs are in the transition metal catalyzed hydroborations.(7) These reactions are believed to take place via an O−M bonding interaction, and not an O−B interaction.

As part of a program designed to develop new methods for heteroatom-directed hydroboration, we have initiated a study of amine borane activation. A preliminary account of this work has appeared,(8) as well as a report describing the analogous activation of phosphine boranes.(9) We now describe the activation of unsaturated amine boranes in detail.

According to the classical dissociative mechanism, hydroboration (HB) requires a trivalent borane to access the 4-center transition state. However, initial formation of an olefin borane π-complex via SN2-like alkene bonding at the backside of the amine borane N−B bond might also be considered, by analogy to the mechanism suggested for borane etherate hydroborations based on computational studies.(10) If such a pathway were to operate for unsaturated amine boranes 3, then a slow reaction would be expected for two reasons. First, the inherent stability of the N−B bond works against any mechanism that requires N−B bond cleavage, including an SN2-like pathway. Second, and more important, the orbital requirements for internal nucleophilic attack are not satisfied in the case of 3 or related homoallylic and bis-homoallylic amine boranes because the potential nitrogen leaving group bond would be endocyclic with respect to the hypothetical cyclic transition state.(11) This means that there should be no rate advantage for HB via an internal SN2-like displacement process involving the C=C subunit of an unsaturated amine borane acting in the role of the nucleophile, as long as the amine nitrogen acts as the leaving group. The above analysis implies that 3 would have to react via a dissociative pathway, presumably via π-complexes such as 4 or 5. These intermediates may well undergo further reaction with a preference for one of the two regioisomeric 4-center transition states due to electronic factors associated with the amine or amine borane subunits, but such effects are poorly understood and their role is difficult to anticipate. The goal of the present study was to determine whether a more predictable version of regiocontrolled hydroboration is feasible via amine borane intermediates where the N−B bond remains intact.

We began with the premise that an internal hydroboration may be possible starting from a stable amine borane of general structure 3 if a good exocyclic leaving group is introduced at boron, as in the iodoborane complex 6 (Scheme 1). We could imagine a dissociative SN1-like mechanism via an ion pair 7, or an SN2-like internal nucleophilic substitution process via the bonding interactions shown for 8. In either event, the initial result would be the formation of a tethered olefin π-complex ion pair 9. Bond reorganization via the fused bicyclic 4-center transition state 10 with subsequent ion pair recombination would then afford 11. Competition by a bridged bicyclic transition state 12 to give the regioisomeric product 13 is also conceivable, but this was expected to be the minor pathway.

Scheme 1

Results and Discussion

To test the above premise, a series of homoallylic amines were prepared, usually from the n class="Chemical">alcohols 14 by aminolysis of the tosylates 15 (Scheme 2). Amines 16, 18, and 20 were then easily converted into the isolable amine boranes 17, 19, and 21 by treatment with THF•BH3. Similar conventional methods were used to prepare homologous alkenylamine boranes 25, 27, and 31, 33.

Scheme 2

After a preliminary survey of common electrophiles, we quickly settled on iodine as the activating agent. Iodination of amine boranes is fast, easy to do, and has a long history.(12) According to the stoichiometry, molecular iodine is fully utilized in conversion of 3 to 6. Apparently, the initially formed byproduct HI reacts with 3 to generate a second equivalent of the iodoborane complex 6, together with hydrogen gas. When the iodine activation method was applied to homoallylic amine boranes, HB occurred at room temperature on a time scale of hours according to product assay after standard oxidative workup with alkaline hydrogen peroxide. However, initial attempts to characterize the presumed intermediates 6 or the initial products 11 and 13 using NMR methods were not informative. Numerous signals were observed in the crude 1H NMR spectrum prior to oxidative workup, as would be expected from the presence of diastereomers involving stereogenic boron as well as regioisomers, but the sheer complexity of the spectrum suggested the presence of additional unknown species. Thus, structures 6, 11, and 13 were tentatively drawn based on stoichiometry and product assay after oxidative workup.

As shown in Table 1, activation with 50 mol% I2 in CH2Cl2 converts a broad range of unsaturated amines into amino alcohols. Good to excellent regiocontrol was observed in the representative case of E-19b (entry 5; 13:1 36b:37b) as well as a number of other examples (entries 2, 6, 8, 12, 15, 18), leaving little doubt that an internally directed HB had been achieved. Somewhat improved 18:1 selectivity for 36b:37b was obtained in the iodine experiments when E-19b was recrystallized prior to activation (entry 6), but the 13:1 ratio using E-19b partially purified by simple plug filtration over silica gel (entry 5) has been retained in Table 1 to allow better comparisons with other entries where the amine borane substrate could not be crystallized.

Table 1

entrya	amine borane	yield from amine	R1, R2	R	n	products	ratiob	yield
1	17a	77%	H, H	H	1	34a:35a	1:1	69%c
2	E−17b	79%	H, Me	Me	1	34b:35b	43:1	85%c
3	E−17bd	79%	H, Me	Me	1	34b:35b	38:1d	83%c
4	19a	>90%	H, Bn	H	1	36a:37b	1:3	90%
5	E−19b	>90%	H, Bn	Me	1	36b:37b	13:1	83%
6	E−19b	76%	H, Bn	Me	1	36b:37b	18:1e	79%
7	E−19b	>90%	H. Bn	Me	1	36b:37b	10:1f	79%
8	Z−19b	>90%	H. Bn	Me	1	36b:37b	>20:1	87%
9	E−19c	>90%	H, Bn	Et	1	36c:37c	10:1	82%
10	E−19d	>90%	H, Bn	Ph	1	36d:37d	2:1	82%
11	E-19d(D3)	73%	H, Bn	Ph	1	36d(D1):37d(D1)	3:1	NA
12	Z-21c	77%	(CH2)4	Et	1	38c:39c	>30:1	92%
13	25a	41%	H,H	H	2	40a:41a	1:1.5	35%
14	E-25b	55%	H,H	Me	2	40b:41b	4:1	84%
15	Z-25b	71%	H,H	Me	2	40b:41b	14:1	92%
16	27a	63%	H,Bn	H	2	42a:43a	1:3	71%
17	E−27b	92%	H,Bn	Me	2	42b:43b	4:1	76%
18	Z−27b	83%	H,Bn	Me	2	42b:43b	39:1	80%
19	E−31b	63%	H,H	Me	3	44:45	2:1	95%
20	Z-31b	54%	H,H	Me	3	44:45	2:1	91%
21	Z−33b	83%	H,Bn	Me	3	46:47	1.5:1	63%

Conditions, 50 mol% I2/DCM, rt; yields refer to aminoalcohols isolated by chromatogaphy as a mixture of regioisomers after workup with NaOOH.

NMR assay.

Reaction quenched after 30 min.

Catalytic activation with 10 mol% I2.

Recrystallized substrate.

Activation using TfOH.

Conditions, 50 mol% I2/DCM, rt; yields refer to n class="Chemical">aminoalcohols isolated by chromatogaphy as a mixture of regioisomers after workup with NaOOH.

To test for competition by an intermolecular hydroboration pathway, the iodine activation of crystallized E-19b was conducted at two different concentrations. However, the ratio of regioisomeric aminoalcohols obtained with 0.1 or 1.0 M solutions of the purified substrate was identical (18:1 36b:37b), suggesting an intramolecular pathway. Additional evidence was provided by a control experiment where E-19b was treated with excess borane in THF (2 equiv). This gave a much lower ratio of 36b:37b (2.4:1) as expected for conventional intermolecular HB. Finally, E-19b was activated in the usual way, but in the presence of added β-methylstyrene (2 equiv) as a potential trap for dissociated borane species or other species capable of intermolecular HB. After oxidative workup, this experiment gave conversion of E-19b to 36b:37b as usual, along with recovered β-methylstyrene. Only traces of the alcohol products derived from β-methylstyrene were detected (<10%). Thus, the internal hydroboration was not suppressed by the external alkene.

Activation with triflic acid as the electrophile(9) was briefly compared with the iodine method using E-19b as the substrate (entry 7). Internal HB occurred on a similar time scale, but regiocontrol was lower (10:1 36b:37b). In view of the inconvenience of handling triflic acid compared to iodine, the procedure was not pursued further.

Higher regioselectivity was observed for the iodine-induced internal HB of Z-alkenes compared to the E-isomers (compare entries 8 vs 5, 15 vs 14, or 18 vs 17). Thus, with Z-19b as the substrate, the standard activation; oxidation procedure gave a single product 36b by NMR assay (entry 8). To establish detection limits, authentic1,4-amino alcohol 37b was prepared independently.(13) According to a calibration study, 37b would have been detected at the 5% level, indicating >20:1 36b:37b selectivity for the iodine-induced hydroboration of Z-19b. The role of olefin geometry was more pronounced for the bis-homoallylic amine boranes, especially for the N-benzyl derivatives (entries 18 vs 17). However, the relatively modest selectivities for some, but not all of the E-alkenylamine boranes indicate that the directing effect arises from a combination of factors. In this light, it was no surprise to find that the directing effect is not sufficient to overwhelm other factors that contribute to HB regioselectivity. For example, terminal alkenes reacted with low selectivity (entries 1, 4, 13, 16), as did a phenyl-substituted substrate (entry 10). In the latter example, the result can be understood by comparing the product ratio with that expected from the standard (intermolecular) HB selectivity for β-methyl-styrene, 86:14 in favor of the benzylic alcohol after oxidative workup.(14) For the amine-directed HB reaction (entry 11), the benzylic alcohol is the minor product (1:3 ratio).

Attempts to extend the amine-directed hydroboration to tertiary amine derivatives encountered complications. The alkenyl pyrrolidine borane Z-21c was relatively well behaved, and conversion of the alkene occurred within 5 h at rt, although the amino alcohol 38c was not recovered cleanly due to competing formation of the N-oxide during oxidative workup. In contrast, the dibenzylated analog of amine borane E-19b (R1, R2 = Bn) gave low conversion under the standard conditions, and unreacted amine was recovered together with the regioisomeric amino alcohols (1.0:3.5:1.5 ratio).

Deuterium labeling experiments were utilized to probe the amine-directed hydroborations in greater detail. Labeled E-19d(D3) was prepared by the addition of the parent amine to THF•BD3 (generated in situ from NaBD4 + iodine).(15) An equivalent of E-19b was then added to E-19d(D3) and the mixture was activated with iodine as usual. After oxidative workup, the resulting amino alcohol products were analyzed for the presence of deuterium. Within limits of detection by ESMS(+) and 1H NMR, neither 36b nor 37b contained deuterium that might be traced to intermolecular crossover or disproportionation events involving E-19d(D3). Furthermore, 1H NMR integration revealed the presence of 36d(D1) and 37d(D1), the expected products of intramolecular deuteroboration. The control experiment using E-19d(D3) as the only substrate in the iodine activation gave identical spectral data for the signals assigned to 36d(D1) and 37d(D1). These results rule out a significant (>5%) role for retro-hydroboration and deuteroboration pathways that would be expected to produce 36d(D2) and 37d(D2) by H/D scrambling. Therefore, the reaction of E-19d(D3) with iodine involves a kinetically controlled, intramolecular hydroboration.

An experiment was carried out with substrate E-17b to learn whether a catalytic amount of iodine would be sufficient for conversion of the amine borane (Table 1, entry 3) to hydroboration products (Scheme 3). Treatment with 10 mol% iodine resulted in a somewhat slower conversion of the substrate compared to the experiment using 50 mol% iodine (2 h vs 0.3 h at rt), but oxidative workup gave the amino alcohol 34b as the dominant product with nearly the same yield and excellent selectivity (Table 1, entry 4 vs entry 3). This observation indicates that the initially formed hydroboration product 49 can react with the substrate E-17b by reversible hydride-iodide exchange, a process that would regenerate the activated 48 as well as the cyclic amine borane 50. The experiment was therefore repeated without oxidative workup, and formation of 50 was confirmed by isolation after chromatography (41% yield).

Scheme 3

The catalytic iodine procedure was also performed with E-19b. Initial attempts to oxidize the reaction mixture from hydroboration led to variable product recovery and a 7:1 ratio of amino alcohols 36b and 37b. The problem was traced to lower reactivity in the oxidative cleavage step for the presumed cyclic amine borane intermediates 53 and 54, structures that are relatively stable compared to the precursor iodoboranes 51 and 52 that accumulate under conditions of stoichiometric activation. Improved reactivity in the oxidative cleavage was achieved by the addition of 40 mol% of iodine to the product mixture obtained from catalytic hydroboration to force the conversion of 53 and 54 back to 51 and 52. This gave a 10:1 ratio of amino alcohol products E-36b:E-37b (79% yield). Of course, reactivation by adding iodine negates any practical advantage for catalytic activation in the hydroboration step for this case. No further attempt was made to characterize the mixture of regioisomers and epimers 53 and 54 or to identify the factors responsible for the diminished 10:1 regioselectivity compared to the stoichiometric activation method.

Iodine activation was explored in substrates 55 and 56, structures where complications due to regioisomer or epimer formation were not expected (Scheme 4). Stoichiometric activation of 55 gave no hydroboration products at rt, consistent with the pattern of slower reactions for the tertiary amine boranes as noted earlier. However, the analogous primary amine borane 56 behaved normally using either 50 mol% or 10 mol% iodine at rt. In the stoichiometric reaction, oxidative workup afforded the amino alcohol 60, but purification was easier after benzoylation to give the bis-benzoyl derivative 61. The latter structure provided an opportunity to determine whether the stereochemistry corresponds to the typical syn delivery of boron and hydride to the olefin via a 4-center hydroboration pathway. This was confirmed by the diaxial trans relationship for Ha and Hb in 61, assigned from the vicinal coupling constant (J = 9.5 Hz).

Scheme 4

The catalytic activation method using 10 mol% iodine at rt also worked well with 56. Complete consumption of olefin required less than two hours and cyclic amine borane 59 was isolated from the mixture in 46% yield after chromatography. Under these reaction conditions, propagation of the catalytic cycle likely occurs by reversible transfer of iodine from 58 to 56 to regenerate the activated iodoborane complex 57. The catalytic hydroboration could also be initiated by the addition of 10 mol% of Et3N· BH2I to 56, as expected according to the proposed catalytic cycle via reversible H/I exchange. The cyclic amine borane 59 could be oxidized with NaOOH to the amino alcohol 60 in 87% yield without further activation, provided that the oxidation was allowed to proceed for >8 h at rt. The stereochemistry of 60 obtained using the catalytic iodine method was confirmed by conversion to 61.

As an additional probe of the stereochemical consequences of internal hydroboration, the bis-homoallylic amine borane 62 (obtained as a mixture of inseparable N-epimers) was subjected to the iodine activation conditions at rt (Scheme 5). Hydroboration was sluggish compared to the substrates of Table 1 (24 h for complete alkene consumption), but warming to 45 °C gave conversion within 4−5 h. Oxidation with NaOOH then afforded 1,3- and 1,4-amino alcohols 63 and 65 as major products in a 10:3 ratio. The syn stereochemistry obtained for each of these compounds is consistent with an internal mechanism. None of the 1,3-trans isomer 64 was found and only a small amount of 66 was detected, presumably resulting from some form of intermolecular hydroboration due to the higher temperature. The major product 63 was identified by comparison with the mixture of isomers 63 + 64 obtained by LAH reduction from ketone 67.(16) In contrast to the iodine-promoted reaction, intermolecular hydroboration of 62 using excess THF•BH3 gave a nearly equal mixture of all four possible isomers 63−66 by NMR assay.

Scheme 5

Having shown that homoallylic and bis-homoallylic amine boranes undergo intramolecular hydroboration upon treatment with iodine, we turned to the allylic amine analogues to learn whether an amine-directed HB pathway would be possible with the shorter tether (Scheme 6). The expected amino alcohol products did form in some examples using iodine activation and oxidative workup, but reproducibility was poor. This was surprising because the iodine activation of homoallylic amine boranes had consistently afforded HB products even though minimal precautions were taken to exclude air or moisture. By comparison, the allylic amine borane activations required carefully optimized inert atmosphere conditions to achieve substantial conversion. Despite these precautions, iodine activation of crotylamine borane 68a rarely gave better than 50% yield of amino alcohols 69a and 70a (5−7:1 ratio), and some experiments gave no amino alcohol products at all. Furthermore, several tries with the secondary N-benzylcrotylamine borane 68b afforded no hydroboration products upon treatment with 50 mol% of iodine at rt. Heating the reaction mixture at 45 °C did give partial conversion on a time scale of hours (6:1 ratio of 69b:70b along with ca. 20% recovered N-benzylcrotylamine), but iodine activation of the tertiary substrate 68 (R1 = R2 = Bn) gave no hydroboration regardless of temperature. Routine control experiments using excess THF•BH3 also gave unexpected results. Thus, treatment of 68a or 68b with excess THF•BH3 followed by NaOOH gave improved ratios of 69:70 after 2 h at 0 °C (a, 9:1; b, 10−13:1) compared to the iodine activation experiments.

Scheme 6

Further tests were conducted using the cyclic allylic amine boranes 71a,b to determine whether the stereochemistry of HB products would correspond to an internal hydroboration pathway (Scheme 7). However, 71a gave 1,2-trans amino alcohol 72a as the major product (≥3:1 72:73; iodine activation at rt over 4−5 h, followed by oxidative workup) while 71b gave no amino alcohols at room temperature. Formation of the trans isomer 72 suggests competition by an intermolecular mechanism, so the experiment was repeated in the presence of 1 equiv of β-methylstyrene in an attempt to shut down intermolecular pathways. This experiment should favor 73, the presumed product of the nitrogen-directed intramolecular pathway, but the major amino alcohol was the same trans-isomer 72 (2:1 72:73) although the product ratio was lower. Alcohols 74 and 75 were also obtained from this experiment, providing clear evidence for competing intermolecular HB. When iodine activation of 71a was performed in the presence of 3 equiv of β-methylstyrene, alcohols 74 and 75 were the exclusive products (6.5:1 ratio). This ratio is similar to the result from reaction of β-methylstyrene with THF•BH3 (74:75 = 5.7:1). The above experiments show that intermolecular HB pathways are not only accessible, but potentially dominant in the reaction of 71a under the conditions of iodine activation, in contrast to the behavior of homoallylic amine boranes.

Scheme 7

Similar reactivity trends were found in a comparison of amine boranes 76a and 76b, although conversion to hydroboration products was slower (Scheme 8). HB did occur if the resulting mixture was heated at 45−50 °C (2 h), resulting in the amino alcohol 78a (74%) after oxidative workup along with ca. 10% of cyclohexenylmethylamine. On the other hand, the corresponding sequence starting with 76b gave at most 10% of 78b, and ca. 90% of the material was recovered as the unreacted N-benzylcyclohexenylmethylamine.

Scheme 8

In an attempt to understand the puzzling reactivity and reproducibility trends, the crude product after iodine activation of 76a was examined at rt using 1H NMR spectroscopy. The olefinic signal of the starting complex 76a (δ ca. 5.65 ppm) was replaced by two new signals at δ ca. 5.85 ppm and 6.0 ppm together with several minor maxima. The major signal at 6.0 ppm, along with a broad absorption at δ ca. 7.5−7.8 ppm and a broad singlet at 3.6 ppm, were identified by comparison with the corresponding signals of cyclohexenylmethylammonium iodide 79. Evidence regarding the new signal at δ 5.85 ppm is less conclusive, but this absorption was enhanced by addition of free cyclohexenylmethylamine to freshly activated 76a, presumably containing 77a. This suggests a boronium salt structure 80 as the species responsible for the δ 5.85 ppm signal, but 80 could not be isolated to confirm this assignment. Furthermore, the presence of 77a after iodine addition could not be established due to the complexity of the 11B NMR spectrum.

The most important observation from the above study is that formation of the ammonium salt 79 is a major pathway resulting from iodine activation of 76a. We do not have detailed evidence regarding how 79 is formed, although the HI generated in the first stage of iodine activation would have to be the principal suspect. Direct protonation of the N−B bond via a three-center two-electron interaction is one possibility, and another is dissociation of the amine-BH2I complex followed by protonation at nitrogen. In either case, the results would indicate a weaker N−B bond for the allylic amine boranes compared to the homoallylic or bis-homoallylic amine complexes. This is consistent with the electron-withdrawing inductive effect of the olefinic sp2 carbon, a factor that is responsible for the lower pKa (by ca. 1 pKa unit) of ammonium salts derived from allylic amines compared to their saturated amine analogues.(17) The same inductive effect should lower the stability of allylic amine boranes, and could be the reason why allylic amine nitrogen becomes partially protonated during iodine activation. In any case, N-protonation requires that one or more trivalent boron species (presumably, BH3 or BH2I) would have to be released as the N−B bond is cleaved. This would explain why the allylic amine borane reactions are highly sensitive to experimental variables, why the products are mostly those expected from an intermolecular hydroboration, and why some of our experiments with the allylic amines gave no hydroboration products even though the reactions were conducted under nitrogen. Dichloromethane is not an ideal solvent for hydroborations because the borane-solvent complex is labile, at least compared to THF•BH3, and is easily destroyed by contaminants.

Given the insights obtained from the study of 76a, we reexamined the reactions of the crotylamine borane 68a. Immediately upon addition of iodine (ca. 5 mol%) to 68a, a broad singlet appeared at δ 7.35 ppm that shifted to δ 7.55 ppm as more iodine (50 mol% total) was added. Over the same time scale, the broad signal due to the complexed NH2 protons shifted from ca. δ 3.65 to 4.45 ppm. New olefinic signals also appeared, and basic oxidative workup returned unreacted crotylamine, as would be expected assuming that crotylammonium iodide is formed during activation. If sufficient care was taken to maintain inert conditions, the amino alcohols 69a and 70a were also recovered after oxidative workup (69% yield). However, we could not extract meaningful structural evidence regarding the hydroboration products prior to oxidative workup from the complex 1H or 11B NMR spectra.

In response to reviewer comments, one more series of NMR experiments was initiated in an attempt to obtain more evidence regarding the activated borane intermediates using simple model amine boranes as well as the well-behaved, primary homoallylic amine-derived borane E-17b. The latter substrate affords the best overall combination of regioselectivity (43:1) and product yield (>80%) in amine-directed HB, and has the simplest structure. Nevertheless, the 1H NMR spectrum after activation using 50 mol% iodine (0 °C to rt, 40 min) could not be interpreted beyond noting a broad signal at δ 7.45 ppm (ca. 0.1−0.15 H) suggesting minor formation of ammonium salts, as well as six maxima between δ 3.1 to 3.6 ppm with extensive splitting and overlap in the expected CH2NH region. Minor broad signals were also present, along with complex, overlapping upfield signals. The 11B NMR spectrum showed several maxima between δ 6 and δ −25 ppm. In striking contrast, when the same substrate E-17b was treated with 5 mol% iodine (catalytic conditions), the 11B NMR spectrum was relatively clean and consisted of the signals for unreacted E-17b together with the cyclized product amine borane 50. Evidently, the NMR complexity using 50 mol% iodine is due at least in part to the formation of N−B−I species, but control experiments with simple models (see below) suggest that the presence of N−H bonds in the substrate E-17b is the main problem. As already reported in the early literature, amine boranes lacking N−H bonds react cleanly with iodine. For example, treatment of the tertiary amine borane Me3NBH3 with iodine affords the iodoborane complex Me3NBH2I as a single species.(18a) We also observe a single well-resolved NMR signal with the expected triplet coupling (δ 11B = −9.3 ppm, t, J = 131 Hz). On the other hand, reaction of the known primary amine borane nBuNH2BH3(18b) with 50 mol% iodine in chloroform produces multiple 11B maxima. Minor signals observed at δ 18.1 ppm (diborane by chemical shift comparisons)(19) and δ −23 ppm (N−BH2X, br t, JBH = ca. 135 Hz) were relatively sharp, but accounted for less than 4% of total signal intensity. Two major peaks (δ −7 and −17 ppm) were also observed, but these resonances were broad and featureless, and could not be assigned with certainty, although the chemical shifts are near the range reported for dimeric hydrogen bridged amine boranes (δ 4 to −11 ppm) that might be formed via elimination of HI from nBuNH2BH2I, followed by 2 + 2 combination of the transient monomer [nBuNHBH2].(20)

Although definitive structural evidence was not obtained despite considerable effort, the NMR studies do show that complexity is inherent in the iodine activation of amine boranes that contain N−H bonds, such as the primary amine borane E-17b. Despite this complexity, the amine-directed hydroborations proceed with good to excellent regioselectivity and conversion in a number of cases. Because multiple species are present in solutions containing the iodine-activated intermediates, the isolation of internal hydroboration products in good yields supports the notion that the unidentified activated intermediates are formed reversibly. According to this interpretation, the byproducts of iodine activation undergo interconversion with iodoborane complexes such as 6, and internal HB eventually drives the equilibrium to cyclic amine boranes.

Summary

Homoallylic or bis-homoallylic amine borane activation using molecular iodine occurs via an intramolecular mechanism to favor the products expected from a bicyclic hydroboration transition state. The process satisfies the requirements for a simple intramolecular mechanism, including tests based on regiocontrol, stereochemistry, and deuterium labeling. Added external alkene does not compete effectively for the reactive hydroborating agent, although there is minor release of active hydroborating species capable of intermolecular HB in some cases, as evidenced by the formation of the trans-amino alcohol 66 (ca. 7% relative yield) starting from 62. In general, the amine-directed reaction is not very sensitive to the presence of traces of air or moisture, and requires only minimal precautions as long as there is sufficient reactivity for internal HB at room temperature. Finally, formation of the expected cyclic amine borane products has been confirmed by isolation from experiments conducted under mild catalytic conditions using 5 mol% iodine.

In contrast, the allylic amine borane activation experiments do not satisfy any of the tests for an intramolecular reaction. Although we cannot exclude minor competition by an intramolecular mechanism in some of the allylic substrates, there is no compelling reason to propose that the relatively strained internal hydroboration transition states are feasible in the allylic series. To the contrary, there is evidence suggesting substantial cleavage of the N−B bond already at room temperature.

In another contrast between the homoallylic and allylic amine boranes, the latter undergo regioselective intermolecular hydroboration using the standard THF•BH3 procedure to form primarily the 1,2-amino alcohols(21) after oxidative workup, while the homoallylic substrates react nonselectively. Iodine activation dramatically improves regioselectivity with the homoallylic and bis-homoallylic amine boranes, but actually lowers selectivity for the allylic analogues compared to the excess THF•BH3 control experiments. Several early reports describing the intermolecular hydroborations of allylic amine substrates with THF•BH3 have encountered similar regioselectivity,[22,23] but we have found no prior reports of directed (i.e., intramolecular) hydroborations via intermediates having an intact N−B bond throughout the sequence. The possibility that certain amine boranes may be capable of intermolecular hydroboration with N−B bonding in the transition state has been raised, based on experiments using chiral amine boranes.(24) Mechanistically, this proposal constitutes a plausible extension of the idea that some hydroborations may involve an SN2-like mechanism,(10) but other explanations for low ee merit consideration. Our work has not addressed intermolecular reactions beyond attempts to repeat one of the chiral amine examples(24a) on small scale.(25) Our studies as described above establish that activation of the isolable homoallylic amine borane complexes follows an intramolecular mechanism, and that related reactions can be exploited for the regiocontrolled synthesis of 3-aminoalcohols.