3202
J. Org. Chem. 2002, 67, 3202-3212
Hydration with Mercuric Acetate and the Reduction with 9-BBN-H of 2-(1-Alkenyl)-4,6-dimethyl-s-triazines H. LeRoy Nyquist,* Edward A. Beeloo, and Lydia S. Hurlbut Department of Chemistry, California State University, Northridge, California 91330-8262
Rachel Watson-Clark† and David E. Harwell‡ Department of Chemistry and Biochemistry, University of California, Los Angeles, California 90095-1569
[email protected] Received August 1, 2001
Oxymercuration-demercuration of a double bond in conjugation with the 4,6-dimethyl-s-triazin2-yl substituent as in alkenes 1a,b gave anti-Markovnikov regioselectivity, which is explained by the electron-withdrawing nature of the triazinyl substituent. However, hydroboration of the conjugated alkenes with 9-BBN-H gave the corresponding alkanes 5a-c under normal workup conditions with or without oxidation. With time and without workup the hydroboration of 1b gave spectral evidence for the formation of intermediates 9-13 resulting from the migration of the 9-BBN moiety from the R-carbon to a ring nitrogen with concurrent formation of an exocyclic double bond to an R-carbon of the ring. Hydrolysis of the intermediates gave 5a-c. A possible mechanism involving successive allylic rearrangements is presented. Introduction In a previous paper we reported that the 4,6-dimethyls-triazin-2-yl substituent was electron withdrawing by both the field effect and the resonance or conjugative effect.1 In view of this influence, the substituent should destabilize any intermediate or transition state which may develop a positive charge on the adjacent R-carbon.2 Hence, it was of interest to investigate the extent to which this behavior, as well as that of possible steric factors, would influence the regioselectivity of the hydration of a conjugated carbon-carbon double bond by possibly forcing any developing positive charge to form on the less, or equally, substituted β-carbon. Thus, the substituent in alkenes 1a,b, upon oxymercuration-demercuration, would enhance the formation of alcohols 2a,b, but in hydroboration-oxidation, it would favor alcohols 3a,b.
The above triazinylalkenes, 1a-c, and 4 were synthesized for the above study, and we now wish to report the results of their reactions with mercuric acetate and 9-BBN-H. The former gave abnormal regioselectivity with 1a,b, and the latter reacted with 1a-c to give the unexpected corresponding alkanes 5a-c with either oxidative or hydrolytic workup. In addition, further investigations via 1H, 13C, and 11B NMR of the latter reaction with no workup gave conclusive evidence of the migration of the 9-BBN moiety from the R-carbon to a ring nitrogen with concurrent formation of exocyclic double bonds to the R-carbons of the ring. Similar behavior was shown to occur in the hydroboration of 2-(1propenyl)pyridine.3
Most previously reported general preparations of alkanes from alkenes via hydroboration have required acidic conditions and prolonged heating,4 and trialkylboranes are also reported to be stable under neutral conditions.5 Noteworthy exceptions particularly relevant to the present investigation are the hydrolyses under relatively mild conditions (pH 8) of the hydroborationoxidation of vinyl- and propenylpyridines3 and of substituted styrenes.2a
†
Current address: Clorox Corp., Pleasanton, CA 94612. Current address: Department of Chemistry, University of Hawaii, Honolulu, HI 96822-2275. ‡
(1) Nyquist, H. L.; Wolfe, B. J. Org. Chem. 1974, 39, 2591-2595. (2) (a) Brown, H. C.; Sharp, R. L. J. Am. Chem. Soc. 1966, 88, 58515854. (b) Brown, H. C.; Zweifel, G. J. Am. Chem. Soc. 1960, 82, 47084712. (c) Traylor, T. G. Acc. Chem. Res. 1969, 2, 152-160. (3) Brown, H. C.; Vara Prasad, J. V. N.; Zee, S.-H. J. Org. Chem. 1986, 51, 439-445. (4) (a) Brown, H. C.; Murray, K. J. Am. Chem. Soc. 1959, 81, 41084109. (b) Kabalka, G. W.; Newton, R. J., Jr.; Jacobus, J. J. Org. Chem. 1979, 44, 4185-4187. (5) Johnson, J. R.; Snyder, H. R.; Van Campen, M. G., Jr. J. Am. Chem. Soc. 1938, 60, 115-121.
10.1021/jo010784e CCC: $22.00 © 2002 American Chemical Society Published on Web 04/24/2002
2-(1-Alkenyl)-4,6-dimethyl-s-triazines
J. Org. Chem., Vol. 67, No. 10, 2002 3203
Table 1. Results of Oxymercuration-Demercuration of Triazinylalkenes alkene
mole ratioa
% recovered 1
% 2b
1a 1b 1b 1b
1:1 1:1 1:2 1:3
c 55 29 12
57 39d 68 78
Table 2. Results of Hydroboration of Triazinylalkenes with 9-BBN-H in Hexane yields (mol %)a,b alkene
workup
5
2
1c
1a 1b 1b 1b 1b 1c 1ch 4
H2O2 H2O2 EtOH, H2O CD3OD, D2O noneg H2O2 H2O2 H2O2
62a 81a 72b 46b 23b,g 75 75 0
10a,e 0 0 0 0 0 0 60e
10f 10 8 10g 21f 26f
a
Alkene:Hg(OAc)2. b No detectable alcohol 3 by NMR. c Spectral data indicated the possibility of a few percent. d A higher temperature (53 °C) did not alter the yield.
Results and Discussion The results of the oxymercuration-demercuration of 1a,b are given in Table 1. The above results would support the conclusion that the triazine ring, as a substituent, is sufficiently electronwithdrawing to destabilize any developing charge on the more or equally substituted R-carbon atom within the intermediate bridged acetoxymercurinium ion.2c,6 Thus, the normal regioselectivity is not realized. In addition, two other factors need to be considered. First, in considering the steric influence of the triazine ring, it could direct the attacking mercuric acetate away from the R-carbon and toward the β-carbon and hence increase the yield of 3a,b, or second, it could inhibit the attack by water on the acetoxymercurinium ion at the R-carbon and thus favor the formation of 2a,b. The absence of any detectable alcohols 3a,b within the products would indicate that the first influence is negligible, and the second influence would not appear to be substantial in view of the fact that 3,3-dimethyl-2-butanol is formed in 94% yield from the oxymercuration-demercuration of 3,3-dimethyl-1-butene in which water attacks the carbon adjacent to the tert-butyl group.7 Thus, the regioselectivity appears to be primarily due to the electronic influence of the triazinyl substituent. A third factor to be considered is that the mercuric acetate, as a Lewis acid, may initially coordinate with the lone pair of electrons8 on a nitrogen of the triazine ring and thereby enhance the electron-withdrawing influence of the ring above and beyond that which the neutral ring possesses. The increased yield of alcohol 2b with an increase in the mole ratio of mercuric acetate to alkene may support this involvement of the mercuric acetate with the ring independent of the mercuration of the double bond. The hydroboration-oxidation with 9-BBN-H was carried out on the conjugated alkenes 1a-c and on the unconjugated alkene 4. The results are given in Table 2. In the normal hydroboration-oxidation, alkene 1a was the only conjugated alkene which gave evidence of the formation of any alcohol (∼10% 2a). On the basis of the experimental data from the hydroboration-oxidation of 1b and the large withdrawing resonance effect of the s-trianzinyl substituent,1 one would conclude that, in contrast to the hydroboration-oxidation of 2-vinylpyridine (R:β ≈ 1:2),3 the 9-BBN-H in this case added predominantly to the R-carbon (R:β ≈ 6:1) to form a (6) (a) Kitching, W. Organomet. Chem. Rev. 1968, 3, 61-133. (b) Chatt, J. Chem. Rev. 1951, 48, 7-43. (c) Olah, G. A.; Clifford, P. R. J. Am. Chem. Soc. 1971, 93, 2320-2321. (7) Brown, H. C.; Geoghegan, P. J., Jr. J. Org. Chem. 1970, 35, 1844-1850. (8) (a) Ramachandran, L. K.; Witkop, B. Biochemistry 1964, 3, 1603-1611. (b) Dean, P. A. W. In Progress in Inorganic Chemistry; Lippard, S. J., Ed.; Wiley & Sons: New York, 1978; Vol. 24, pp 109178. (c) Wirth, T. H.; Davidson, N. J. Am. Chem. Soc. 1964, 86, 43144318.
cyclooctanoned
25 19 13g
a Based upon GLC data. b Based upon NMR data. c Recovered. Based upon moles of 9-BBN-H. e 1° alcohol. fBased upon epoxide yield. g Products distilled from reaction. h With 9-BBN-D in hexane.
d
transition state species, 6a, which places a partial positive charge on the primary β-carbon rather than the destabilized secondary R-carbon to ultimately form the adduct 7a. It is also reasonable that the alkane would arise only from adduct 7a in which any carbanion character on the R-carbon3 would be stabilized by the s-triazinyl ring.
When the hydroboration-oxidation was initially carried out on 1b,c with an excess of the alkene, and the normal period of time,9 a minor amount of the corresponding epoxy compound was also isolated.10 Consequently, subsequent reactions were carried out with 5-10% mole excess of 9-BBN-H. In view of the fact that the triazinylalkane 5b, and not the alcohol, was the only product isolated from the hydroboration-oxidation of 1b, it became apparent that only a source of hydrogen, and not an oxidizing agent, was involved in the workup. Thus, the workup of the hydroboration of 1b with ethanol and water in the absence of hydrogen peroxide gave isolated 5b (72 mol %). A comparable experiment with methanold4 and deuterium oxide gave 5b with increased multiplicity in the 1H NMR for the R-methylene proton due to the presence of deuterium which had replaced the 9-BBN moiety on the R-carbon. This mode of addition to give the adduct 7 was also consistent with the results of the reaction of 9-BBN-D with 1c. There was no spectral evidence in CDCl3 (1H and 11B NMR) for the migration of the 9-BBN moiety along the propyl group,11 but there was evidence in the 11B NMR spectra for the initial association of the 9-BBN-H dimer or monomer12 with the nitrogens of the ring (Kb for 3,5,6-trimethyl-1,2,4triazine: ∼7 × 10-12)13 as reported with amines,14 as well (9) Brown, H. C.; Knights, E. F.; Scouten, C. G. J. Am. Chem. Soc. 1974, 96, 7765-7770. (10) (a) House, H. O. Modern Synthetic Reactions, 2nd ed.; Benjamin: Menlo Park, CA, 1972; pp 306-310 and references therein. (b) Bunton, C. A.; Minkoff, G. J. J. Chem. Soc. 1949, 665-670. (11) Taniguchi, H.; Brener, L.; Brown, H. C. J. Am. Chem. Soc. 1976, 98, 7107-7108. (12) Wang, K. K.; Brown, H. C. J. Org. Chem. 1980, 45, 5303-5306. (13) Mason, S. F. J. Chem. Soc. 1959, 1247-1253.
3204
J. Org. Chem., Vol. 67, No. 10, 2002
Nyquist et al.
Table 3. NMR Chemical Shifts from the Hydroboration of 1b with 9-BBN-H in CDCl3a R-CHB< compd adduct 7b intermediatesd,e alkane 5bf
1H
2.98 (b)
R-CH2-
13C
11B
30.6 (b)
+84b
β-CH2-
ring CH3
γ-CH3
1H
13C
1H
13C
1H
13C
1H
13H
2.72 (t) 2.71 (t)
41.01 40.64
c 1.77 (m) 1.75 (m)
20.85 21.61 21.26
0.92 (t) 0.94 (t) 0.93 (t)
14.79 14.01 13.69
2.50 (s) 2.55 (s) 2.54 (s)
25.70 25.71 25.33
a Under argon in sealed NMR tubes. b See ref 20. c Initially obscured by 9-BBN signal. d Intermediates, 9b-13b, and/or 5b. e With time the 1H NMR spectra began to show eight small signals (total area ∼ 2-3 H): two triplets (δ 0.86 and δ 4.87), a quintet (δ 2.03), a multipet (δ 2.21), and two pairs of singlets (δ 4.04, 4.06 and δ 4.39, 4.48). The 13C NMR spectra also showed several small signals between δ 60 and 150. Within the 11B NMR spectra the δ +84 signal decreased as the δ +59 signal increased. f Authentic sample.
as with the nitrogen of a nitrogen-nitrogen double bond.15 Boron has also been reported to associate with oxygen in the 1,4-addition of 9-BBN-H in CDCl3 to give reduction of the double bond in 1-phenyl-2-alken-1-ones16 and in 4-hexen-3-one,17 as well as in the 1,2-reduction of a number of conjugated aldehydes to the corresponding allylic alcohols.18 In view of the above unanticipated results and the uncertainty of the origin of some products, such as cyclooctanone,19 the progress of the reaction of 9-BBN-H with 1b under argon in standard reaction apparatuses was abandoned. Thus, all subsequent reactions, except those for 11B NMR spectra, were carried out in degassed solvents under argon in sealed ampules or NMR tubes. Although abstraction of hydrogen from the glass or the deuterium from the solvent was considered, experimental evidence did not support either possibility. The products arising from the reactions were examined during, and/ or after, many months (4-20) at room temperature or at elevated temperatures (∼55 °C) for many days (2560) by NMR, GC-MS, and/or by quenching with deuterated solvents followed by GC-MS. The following spectral changes were observed for the reaction of 9-BBN-H with 1b in CDCl3 at room temperature. Within a few hours, in addition to the signals from the alkene and the 9-BBN moiety, the 1H, 13C, and 11B NMR spectra began to show signals consistent with those of adduct 7b as shown in Table 3. The broad signal at δ 2.98 within the 1H NMR spectra due to the presence of boron on the R-carbon requires some comment. A three-coordinate, but not a four-coordinate, boron on carbon is reported not to undergo two-bond coupling with hydrogen.21 Therefore, this situation is either anomalous or the boron is partially four-coordinated with a nitrogen of the ring or is still partially associated as a dimer. The 11B NMR spectra at δ +5922 would support an early (2 h) and an increasing association of a boron with a nitrogen as the reaction progressed. As the reaction continued, the signals from 7b increased and those of the alkene decreased. However, (14) (a) Brown, H. C.; Kulkarni, S. U. Inorg. Chem. 1977, 16, 30903094. (b) Brown, H. C.; Chandrasekharan, J.; Wang, K. K. J. Org. Chem. 1983, 48, 3689-3692. (15) Brand, U.; Hu¨nig, S.; Peters, K.; von Schnering, H. G. Chem. Ber. 1991, 124, 1187-1190. (16) Matsumoto, Y.; Hayashi, T. Synlett 1991, 349-350. (17) Boldrini, G. P.; Bortolotti, M.; Tagliavini, E.; Trombini, C.; Umani-Ronchi, A. Tetrahedron Lett. 1991, 32, 1229-1232. (18) Krishnamurthy, S.; Brown, H. C. J. Org. Chem. 1977, 42, 11971201. (19) (a) Bulgakov, R. G.; Tishin, B. A.; Vlad, V. P.; Maistrenko, G. Ya.; Ostakhov, S. S.; Tolstikov, G. A.; Kazakov, V. P. Khim. Vys. Energ. 1989, 23, 250-256. (b) Kulkarni, S. U.; Rao, C. G.; Patil, V. D. Heterocycles 1982, 18, 321-326. (20) Onak, T. Organoborane Chemistry; Academic Press: New York, 1975; p 7. (21) (a) Onak, T. Organoborane Chemistry; Academic Press: New York, 1975; p 9. (b) De Moor, J. E.; Van der Kelen, G. P. J. Organomet. Chem. 1966, 6, 235-241. (22) No¨th, H.; Wrackmeyer, B. Chem. Ber. 1973, 106, 1145-1164.
Table 4. Relative Yields of Products from Reaction Solutions of 9-BBN-H with Alkene 1a rel yields (%) alkene 1b 1b 1b 1b 1b 1b 1b-d6 1b 1b 1b 1b 1b 1c 1c
solvent hexaned,e hexaneh,i hexaneh,i hexaneh,i hexaned,l CDCl3i,n CDCl3i,n CDCl3i,n CD2Cl2i,n C6D6i,n THF-d8d,i,n THFd,l CDCl3i,n CDCl3i,n
tempb rtf 56 56 55 75 rt rt 60 rt rt rt 75 rt rt
time alkane alkene adduct (months)c (5) (1) (7) 4 9 days 1 2 2 days 2 20 7 days 4 20 1 52 h 20 days 4
75g 35j 50k 35 81m 71o 85 97o 100o,q 51 37o 61s 64o 85o
25g 22j 24k 10 19m 25o 0o 0o 40o 39s 17o 15o
43j 26k 55 4o 15p 3o 0o 49r 23o 19o 0o
a Based upon GC-MS data before workup unless otherwise designated. b Bath temperature, °C. c Unless otherwise designated. d Commercial 9-BBN-H solution. e Vial with septum. f Precipitate formed. g Absolute yields relative to nonane: 5b, 62%; 1b, 21%. h Sealed ampule. i Prepared 9-BBN-H solution. j Absolute yields relative to nonane: 5b, 38%; 1b, 23%; 7, 46%. k Absolute yields relative to nonane: 5b, 38%; 1b, 18%; 7, 20%. l Reaction flask (Ar). m Absolute yields relative to nonane: 5b, 61%; 1b, 15%. n Sealed NMR tube. o Based upon NMR data (alkane yield also includes intermediates 8-13). p m/z 399 (7-d6 + 9-BBN-H). q Absolute yields relative to 2,4,6-tripropyl-s-triazine: 5b, 52 mol %. r m/z 393 (7 + 9-BBN-H). s Absolute yields relative to bibenzyl: 5b, 33 mol %; 1b, 25 mol %.
after 3-4 days at room temperature, all of the signals attributed to 7b had maximized and a sharp triplet at δ 2.72 (R-CH2) in the 1H NMR spectra became increasingly evident. Thus, the signals from 7b began to decrease as new signals consistent with those of the alkane 5b began to appear and continue to increase over a period of months. However, in addition to the signals attributed to 5b, the final 1H NMR spectrum showed a total of eight small signals (Table 3) totaling approximately two to three protons. The final 13C NMR spectrum also gave an increasing number (∼10) of small signals between δ 60 and 150 which a DEPT spectrum showed to be sp2 carbons. Essentially the same sequence of NMR spectra was obtained when the reaction was carried out at 60 °C for 8 days. Since the additional signals could not be assigned to 7b or 5b, they were attributed to species resulting from the rearrangement of 7b to possible species such as 8a-c and 9a-c through 13a-c. Table 4 shows the relative yields of products from the GC-MS analyses of the reaction solutions of 1b with 9-BBN-H in a variety of solvents before hydrolysis. The transformation of 7b into 5b and intermediates 9b-13b appears to proceed more rapidly in more polar solvents. Quenching of the NMR solutions with water after completion of the reaction caused all of the 1H NMR signals unique to the intermediates 9b-13b to disappear and leave only
2-(1-Alkenyl)-4,6-dimethyl-s-triazines
those signals attributable to 5b and the 9-BBN moiety, whose broad envelope (δ 1.25-2.0) had sharpened. Within the 13C NMR spectra the signals between δ 60 and δ 150 disappeared and the broad signals due to the 9-BBN moiety (δ 23.3 and δ 33.2) narrowed.
The following experimental evidence supports the formation of alkane 5, adduct 7, and intermediates 9-13 in the reaction of 9-BBN-H with 1a-c. Although intermediate 8 is a feasible species, no experimental evidence was found to support its existence. (1) When the reaction of 1b with 9-BBN-H in CDCl3 was permitted to proceed only until the adduct 7b was fully formed (70 h, room temperature) and then quenched with methanol-d4, the 1H NMR gave a multiplet at δ 2.65 (R-CHD) and showed essentially a quantitative conversion of 7b to 5b relative to bibenzyl as an internal reference. However, when the reactions were permitted to proceed for longer periods of time past the formation of 7b before quenching, the yields of alkane 5b relative to bibenzyl in the NMR or to nonane in GC-MS were within 40-60% depending upon the solvent (Table 4). These lower yields could indicate that unidentified products were formed before quenching as a result of disproportionation between the intermediates. Also, after deuterium quenching, unidentified products (5-6%) containing deuterium were frequently found in the GC-MS of the products. The MS also showed an enrichment of deuterium in the spectrum of 5b (m/z 150-155) relative to that of 5b from nondeuterated quenches (m/z 150152). The 13C NMR spectrum of the 5b from the deuterium quench showed 13C-D coupling23 both at the carbons of the ring methyl groups (1J ) 19.6 Hz) and the (23) Friebolin, H. Basic One- and Two-Dimensional NMR Spectroscopy, 2nd ed.; VCH Publishers: New York, 1993; pp 49-50 (translated by J. K. Becconsall).
J. Org. Chem., Vol. 67, No. 10, 2002 3205
R-carbon (1J ) 19.5 Hz) giving evidence consistent with the formation of 5b from the quenching of intermediates 9-13. The 9-BBN-O-CH3 (65-75% yield) from the quenching experiments did not show any incorporation of deuterium. Derivatization of the 9-BBN moiety at the end of the reaction with ethanolamine24 or oxidation with peroxide gave yields of approximately 90% of the ethanolamine adduct and 95% of the cis-1,5-cyclooctanediol. Slow evaporation (14 months) of the CDCl3 of the reaction solution from the hydroboration of 1b open to the air gave crystals which upon crystallographic analysis showed the presence of a 1:1 complex of boric acid and 1b together with some 5b and from which, with time, both of the latter evaporated.25 (2) Low-temperature 1H NMR (400 MHz) enabled the signal at δ 2.03 to be resolved into a quintet and the signals between δ 4 and δ 5 to be resolved into a triplet (δ 4.87) and two pairs of singlets (δ 4.04, 4.06 and δ 4.39, 4.48). The quintet and the triplet together with the small upfield triplet (δ 0.86) are supportive of the presence of the propylidene group of intermediates 9b and 10b. Likewise, the presence of the two exocyclic methylene groups and the n-propyl groups of intermediates 11b13b is supported respectively by the two pairs of singlets (2J ) 0) and two large triplets (δ 0.94 and δ 2.72) and a sextet (δ 1.77). The latter three multiple signals are also contributed to by the n-propyl group of any alkane 5b present. The possible presence of 5b before quenching was indicated by a semiquantitative comparison in the 1 H NMR in which the calculated areas from the ring methyl groups of 9b-13b (relative to the areas of the sp2 hydrogens) could account for only 55-60% of the total area from all of the ring methyl groups. Experimentally, the presence of 5b before quenching was also supported by the GC-MS analyses of the reaction solutions (Table 4). The most probable sources of 5b before quenching would arise from hydrogen abstraction from between the intermediates (disproportionation) or from the 9-BBN moiety. (3) To further ascertain whether the methyl groups of the ring or the allylic γ-methyl groups of 1b,c were sources for the hydrogen that replaced the 9-BBN moiety, the following deuterated compounds were synthesized. Experimentally, within 6 h, the hydroboration of 1b-d6
(24) (a) Rondestvedt, C. S., Jr.; Scribner, R. M.; Wulfman, C. E. J. Org. Chem. 1955, 20, 9-12. (b) Kramer, G. W.; Brown, H. C. J. Org. Chem. 1977, 42, 2292-2299. (25) The crystal structure will be reported elsewhere.
3206
J. Org. Chem., Vol. 67, No. 10, 2002 Scheme 1
(93 atom % D) in CDCl3 began to show a sharp triplet at δ 2.72 for the R-methylene group. After approximately 5 days, a broad multiplet (2J ) 2 Hz) began to appear superimposed upon the sharp triplet. This multiplicity at δ 2.72 continued to increase until the original broad signal at δ 2.98 (>CHB
