|
1. |
Oxymetallation. Part 11. Synthesis of cyclic secondary alkyl peroxidesviathe peroxymercuration of α,ω-dienes |
|
Journal of the Chemical Society, Perkin Transactions 1,
Volume 1,
Issue 6,
1978,
Page 522-530
A. J. Bloodworth,
Preview
|
PDF (958KB)
|
|
摘要:
J.C.S. Perkin I Oxymetallation. Part I1 Synthesis of Cyclic Secondary Alkyl Per- .192 oxides via the Peroxymercuration of a,o-Dienes By A. J. Bloodworth and M. E. Loveitt, Christopher lngold Laboratories, Chemistry Department, University College London, 20 Gordon Street, London WC1 H 0A.J ,~ _______.-_ ___ Six cyclic peroxides of the form 0-(XCH,)CH[CH,],CH(CH,X)-b [X = HgCl (1). H (2). or Br (3) ; n = 1 (a) or 2 (b)] have been prepared in high yield via the reaction of penta-1.4-diene or hexa-1.5-diene with hydrogen per- oxide and mercury(l1) nitrate. The new compounds are identified by lH and proton-decoupled I3C n.m.r. spectro- scopy and additionally by mass spectrometry where X = H or Br; compound (1 a) isomerises to a (3-oxo-alcohol when dissolved in pyridine.The 1.2-dioxacyclopentanes are obtained as approximately equimolar mixtures of cis-and trans-isomers whereas the 1.2-dioxacyclohexanes contain about three times as much trans- as cis-isomer. The configurations are assigned on the basis of 13C and IH n.m.r. chemical shifts and variable-temperature n.m.r. spectra. PRINCIPALmethods for synthesising cyclic peroxides involve autoxidation, photo-oxygenation, or nucleo-philic displacement with hydrogen peroxide. Of these, only nucleophilic displacement can be used to make simple 1,2-dioxacyclo-pentanes and -hexanes, and con- ventional variations of the method are intrinsically unsuitable for preparing cyclic secondary alkyl com-pounds. The acid-catalysed reaction of 1,3-diols with concentrated hydrogen peroxide [equation (l)] has provided a series of 1,2-dioxacyclopentanes with tertiary carbon atoms next to the peroxide linkage3-5 and Part 10, A.J. Bloodworth and M. E. Loveitt, J.C.S. Perkin I, 1977, 1031. Preliminary account, A. J. Bloodworth and M. E. Loveitt, J.C.S. Chem. Comm., 1976, 94; corrigendum: cis and frans should be interchanged throughout. R. Criegee and G. Paulig, Ber., 1955, 88, 712. W. Adam and N. DurAn, J.C.S. Chew Com.m., 1972, 279. 3,3,6,6-tetramethyl-1,2-dioxacyclohexanehas been pre- pared by an analogous route.3 The displacement of methylsulphonyl groups by basic hydrogen peroxide has afforded the unsubstituted 1,2-dioxacyclohexane [equation (2) 697]. Even in these structurally favourable cases the yields are often no better than 30y0,3*637but when the bis- (methanesulphonate) method was used to synthesise the secondary compounds 3-methyl-5-(2-phenylethyl)-l,2-dioxacyclopentane and 3,5-bis-(2-phenylet h yl) -1,2-dioxa- cyclopentane the yields fell to 7 and 0.5%, respectively.s W.Adam and N. Duran, J. Org. Chem., 1973, 38, 1434. R. Criegee and G. Miiller, Chem. Bey., 1956, 89, 238. 'I H. D. Holtz, P. W. Solomon, and J. E. Mahan, J. Org. Chem., 1973, 38, 3175. P. M. Jacobs and A. H. Soloway, .I. Org. Chem., 1974, 39, 3427. A five-fold improvement in the yield of 3-methyl-5-(2-phenylethyl)-l,2-dioxacyclopentanewas recently achieved by carrying out the nucleophilic displacement with potassium superoxide and a crown ether in dimethyl sulpho~ide.~~* In recent years we have developed a new route to acyclic dialkyl peroxides which is particularly suitable for preparing secondary alkyl compounds.The method involves the combination of peroxymercuration and hydrogenodemercuration and has been used to provide t-butyl secondary alkyl peroxides [equation (3)lo] in much higher yields than are afforded by conventional HgX, NaBH,RCHXH, -XHgCH2CHROOBut CH3CHROOBut (3) routes. We have also shown that if hydrogen peroxide is used in the peroxymercuration stage then the p-mercurioalkyl hydroperoxide that is generated can itself function as the nucleophile in a subsequent addition [equation (4)]. The assumption underlying this work RCH:CH RCHXH,H202&XHgCH,CHROOH ___+ HgX* (XHgCH,CHRO), (4) was that by linking the two alkene functions together in an appropriate aw-diene, this system could be modified to provide a synthesis of cyclic peroxides in which both carbon atoms adjacent to oxygen are secondary [equation (5)1.of intramolecular alkoxymercuration and aminomercur- ation are both well established.” Furthermore they appear to proceed only if 5-or 6-membered ring formation results, but if this can occur, it is strongly preferred to any competing intermolecular reactions. This augured well for the proposed synthesis of 3,5-bis(mercurio-methyl)-l,2-dioxacyclopentane[equation (5),n = 13 and 3,6-bis(mercuriomethyl)-1,2-dioxacyclohexane [equation (5),n = 21. We now report the peroxymercuration of penta-1,4-diene and hexa-I ,5-diene and the isolation of the rgs’ulting cyclic secondary alkyl peroxides in high yield as the organomercury chlorides (1).We also describe how subsequent demercurations [equation (S)] provide excellent yields of the corresponding dimet hyl-l,2- dioxacycloalkanes (2) and bis(bromomethp1)- 1,2-dioxa- cycloalkanes (3). RESULTS AND DISCUSSION Peroxymercuration.-(a) Choice of mercury(11) salt. Although we employed mercury(I1) trifluoroacetate to prepare acyclic bis-p-mercurioalkyl peroxides [equation (4); X = CF,CO,], attempts to extend its use to the diene systems met with only moderate success. With hexa-l$-diene the procedure adopted in the earlier work with alkenes afforded 56% of crude 3,6-bis- (trifluoroacetoxymercuriomethyl)-l,2-dioxacyclohexane [equation (5); n = 2, X = CF,CO,], but n.m.r.spectro- scopy revealed the presence of impurities and the yield was halved in the process of isolating the pure organo- mercury chloride. No cyclic peroxide could be isolated r cHZHgXi CH,HgX CHz Br I CH 2HgCI1 Me I CH2Br cH~H~C[ Me The scheme of equation (5) requires that the un-saturated hydroperoxide initially formed should par- ticipate in the previously unknown process of intra-molecular peroxymercuration. The analogous reactions * Note added in proof. It has now been reported (M. F Salo-mon and R. G. Salomon, J. Amer. Chem. SOC.,1977, 99, 3500) that 1,2-dioxacyclopentane, 3,5-dimethyl- 1,Z-dioxacyclopentane, and 1,2-dioxacyclohexane can be prepared by the reaction of appropriate alkyl bis(trifluoromethanesu1phonates) with bis-(tributyltin) peroxide ; the method also affords 1,2-dioxacyclo- heptane and -octane.E. J. Corey, K. C. Nicolaou, M. Shibasaki, Y.Machida, and C. S. Shiner, TPtvahrdron Leftws, 1975, 3183. from a similar reaction with penta-1,4-diene, and the i.r. spectrum of the product suggested that extensive trifluoroacetoxymercuration had taken place. Originally rnercury(I1) trifluoroacetate was introduced lo (a)D. H. Rallard, A. J. Bloodworth, and R. J. Bunce, Chem. Comm., 1969, 815; (b) D. H. Ballard and A. J. Bloodworth, J. Chem. Soc. (C), 1971, 945; (c) A. J. Bloodworth and G. S. Bylina, J.C.S. Perkin I, 1972, 2433; (d) A.J. Bloodworth and I. M. Griffin, J. Organometallic Chem., 1974, 66, C1; (e) A. J. Bloodworth and I. M. Griffin, J.C.S. Perkin I, 1975, 195. l1 H. Straub, K. P. Zeller and H. Leditschke, ‘ Houben-Weyl.Methoden der Organischen Chemie,’ Band XIII/Zb Metallor- ganische Verbindungen : Hg, Thieme Verlag, Stuttgart, 1974, pp. 153, 160, 188, 308. in place of mercury(I1) acetate in peroxymercuration because the fluorines attenuate the nucleophilicity of the carboxylic acid and anion and eliminate or reduce competing acyloxymercuration.lod~e The initial results with dienes suggested that we needed to use a mercury(r1) salt with an anion of still lower nucleophilicity. We elected to try Hg(NO,),*H,O because it is commercially available and because there is evidence l1 that the products of intramolecular alkoxymercuration are far more stable towards strong acids than are acyclic ox ymercurials.The change to mercury(rr) nitrate proved highly successful and enabled us to isolate the organomercury chlorides (1) from both hexa-l$-diene and penta-l,4- diene in yields of over SOT, on a 5 mmol scale [equation (~1.Dichloromethane was the solvent used for peroxy- mercurations with mercury( 11) trifluoroacetate and has a; n=l b; n =2 been retained for the reactions with Hg(N03),*H20even though this salt is at best only sparingly soluble. The high strength (ca. 80%) hydrogen peroxide is added to a vigorously stirred suspension of Hg(NO,),*H,O in dichloromethane followed immediately by the diene.J.C.S. Perkin I alkane (1)as a white precipitate. Details of compounds (1) are provided in Table Z together with data for the cyclic peroxides prepared from them by hydrogeno- and bromo-demercuration (see later). The fact that pure 3,5-bis(chloromercuriomethyl)-1,2-dioxacyclopentane (la) is obtained in such high yield indicates that the reaction with penta-1,4-diene is very clean. The corresponding 1,2-dioxacyclohexane (1 b) was isolated in lower yield and was not analytically pure, but examination of the crude organomercury nitrate l3C n.m.r. spectroscopy (see next section) revealed no additional products. (b) Characterisation of the bis(mercuriomethyZ)-1,2-dioxacydoalkanes. As in the identification of earlier peroxymercurials l?lowe have relied heavily on n.m.r. spectroscopy to establish the structures of the new compounds.The lH n.m.r. spectra (Table 2) are complex for two I CH 2HgCL reasons. First, each product is obtained as a mixture of diastereoisomers in which the mercuriomethyl groups are either cis or trans with respect to the ring (see later). Thus each peroxymercurial gives rise to a lH n.m.r. spectrum that is the sum of the non-identical spectra of TAI~LE1 1 ,Z-T)ioxacycloalkanes prepared r~inreaction of orto-dienes with hydrogen peroxide and mercury(rr) nitrate Found (yo) Calc. (yo)"/o trails-r-------h-----\ r------_, Yield Compound n X Isomer fi R1.p. ("C) C H C H (YO) (14 1 HgCl 50 119-121 10.3 1.4 10.5 1.4 92 (tlecomp.) (24 1 H 50 Liquid 58.4 9.85 58.8 9.85 68 (34 1 Br 60 Liquid 23.0 3.2 23.1 3.1 51 (1b) 2 HgCl 95 183 (dccomp.) 12.35 1.85 12.3 1.7 26 f (2b) 2 H 80 Liquid 62.85 10.65 62.05 10.4 72 (3b) 2 Br 75 63--70 Y 25.8 3.65 26.3 3.7 75 fl Calculated (to nearest 50/) from proton-decoupled 13C n.m.r. spectrum assuming equal responses for corresponding carbons of the isomers; checked from 'H n.m.r. spectrum where possible. Based on diene. Purified by g.1.c. Of crude product indicated by analytical g.1.c. to contain cu. 90% of required peroxide. For crude product, Found: C, 59.7; H, 10.O~o. f From Hg(0,CCF3)2; Hg(NOQ),.H,O afforded 82% of crude organoniercury chloride (75yo trans), m.p. 158-162 "C (decomp.) (Found: C, 11.2; H, 1.550/,, suggesting the presence of about 10% of inorganic material).g Rccrystallisation from CH,Cl,-pentane at -10 "C afforded pure truns-isomer, m.p. 91-92 "C. It is important to get the diene into the system quickly the individual diastereoisomers. Secondly, both exo-otherwise mercury(1r) nitrate and hydrogen peroxide cyclic and endocyclic methylene protons are magnetically react to evolve oxygen. An insoluble oil is formed which non-equivalent and with the methine protons attached is presumably a very concentrated solution of the to the asymmetric carbons constitute ABX systems. organomercury nitrate (4) in nitric acid. Treatment of Owing to this complexity it was often difficult to identify an aqueous solution of this oil with potassium chloride unambiguously the satellites arising from coupling with affords the bis(chloromercuriomethyl)-l,2-dioxacyclo-naturally occuring I99Hg nuclei.1978 525 The complication arising from nun-equivalence of in part by the base-induced cleavage that takes place in methylene protons is removed in the proton-decoupled pyridine solution [e.g. equation (8), X = CF,C02].1 13C n.m.r. spectra (Table 3), and these readily permit the A similar reaction [equation (9)] occurs with compound ratios of isomers to be determined (Table 1). The crude (la). This was monitored by lH n.m.r. spectroscopy organomercury nitrates were examined by 1% n.m.r, and was essentially complete after 4 days at room spectroscopy and these data are included in Table 3. temperature. The product had a broad singlet (lH TABLE2 1H N.m.r.spectra of 1,2-dioxacycloalkanes 7 -A-Cornpou ntl 'YL X Solvent Isomer CH,X ( la) 1 HgCl C,H,N * cis 7.7(d) 5.1 (m) 7.8 (m)trans 7.85 (d) 1 CH [CH,< 8.29 (dt)H CCl, cis 8.74(d) 5.75 (ddq) 7.23 (dt) trans 8.78 (d) 5.70 (tq) 7.81 (t)J Br CDCI, cis and trans 6.44(m) 5.35 (m) 7.37 (m) cis 7.7 (d) 1HgCl C5H5N trans 7.95(d) c*g 5.5 (br m) 8.1 (br m) (211) 2 H CH,C1, cis 8.84 (d) 5.94(m) 8.44(In)trans 9.02(d) 1 (3b) 2 Br CDCI, cis 6.40(m) 5.69 (m) 8.02(m)trans 6.61(d) 1 n and X as in Table 1 a See text for basis of assignment. Spcctruni recorded ininiediately after preparing solution. 3J 6.0Hz. ,J 11.5 Hz, J 6.5Hz. 7.2Hz. f3J 6.7Hz. v oJ(lssHg-lH) 241 Hz. *3J 7.0 Hz. AB part of ABX spectrum.TABLE3 Proton-decoupled 13Cn.m.r. spectra of 1,%dioxacycloalkanes 8 * (p.p.m.) Compound )L X Solvent lsomer '' CH,X CH [CH,In (4a) 1 HgNO, DZO-HNO, cis 26.69 81.65 52.49 trans 26.46 82.41 51.40 (la) 1 HgCl C5H5N'-CDCI, 1 31.60d 1 81.96 53.65 trans 53.42cis (?a) 1 H CDCI, cis 3'3.25 77.30 49.34 trans 18.40 77.04 48.61 (:la) I Hr CDCI, cis 32.51 79.68 43.59 trans 31.34 80.35 43.49 (4.h) 2 HgNO, D20-HN03 cis 28.09(?) 80.34 32.62 trans 25.88 81.02 33.25 ( 1 2 HgCl C5H5N'-C,D, cis 28.80 79.5380.02 } 34.27trans 30.75f (2b) 2 H CDCI, cis 18.18 76.30 27.04 trans 18.79 77.08 31.58 (3b) 2 Br CDCI, cis 23.48 i::!: } 30.61trans 27.30 )L and X as in 'l'able 1. See text for basis of assignment. * bowntield from interrial Me,Si; spectra of conipounds (ad)and (4b)were referenced to internal 1,4-dioxacyclohexane[S(Me,Si) 67.391. C 70% v/v.d 1J(1SsHg-*3C) 1 783 Hz. e zJ(199Hg-13C) 62 Hz. J 1J(lssHg-13C)1 757 Hz. 0 zJ('ssHg-'3C) 72 Hz. 'J('9sHg-1%) 112, 4J(ls'Hg-'3C) 19 Hz. Comparison with model compounds, notably the n.m.r.) at T 6.8, assigned to ClHgCH,CO by comparison acyclic bis-(3-mercurioalkyl peroxides,l indicate that the with ClHgCH,COCH, (CH, at T 6.9),I and a signal chemical shifts and lg9Hg coupling constants (lH and (13C n.m.r.) at 6 68.8, assigned to CH(0H) by comparison 13C) and multiplicities ('H) are entirely consistent with with RCH(OH)CH2HgC1 [CH at 6 61.7 (R = H) or 77.0 (XHgCH2CHMe0l2 pyridlne XHgCH2COMe 4-MeCH(OH)CH,HgX (8) pyridine CLHgCHzCOCH,CH(OH)CH2HgCl (9) the proposed structures.The i.r. spectra of the organo- (R = Ph)].12 Compound (2a) was much more stable in mercury chlorides showed no absorptions in the OH- pyridine and did not change appreciably during 5 days stretching region, thus confirming the absence of hydro-at room temperature. peroxides. In addition to their synthetic importance, the reactions Acyclic bis- p-mercurioalkyl peroxides were identified 12 A. J. Rloodworth and R.Peters, unpublished results. with sodium borohydride and bromine discussed below provide additional evidence for the structures of the organomercury compounds. Hydrogenodemercuration.-Analy tical g.l. c. (silicone oil) indicated that the products obtained in almost (1) a; n= 1 bin= 2 quantitative yield by reducing each peroxymercurial ( 1) with sodium borohydride are mixtures containing a major and a minor component.Each major component (ca. 90% of the mixture) was isolated by preparative g.1.c. (silicone oil) and identified by lH and proton- decoupled 13C n.m.r. spectroscopy (Tables 2 and 3) as (a) vnle 84 I I I ?I-ii )5 m /e 43 J.C.S. Yerkin I Presumably the unsaturated alcohols (5)arise through partial deoxymercuration followed by reduction of the resultant unsaturated hydroperoxide [equation (1l)]. This contrasts with the behaviour of t-but yl peroxy- mercurials which afford epoxides as the byproducts,lh Me Me but the occurrence of borohydride-induced deoxy-mercuration has been reported previously for other 0xymercuria1s.l~ The observed fragmentation patterns are summarised in the Scheme.Speculation on the detailed mechanism of ion fragmentations is not warranted but the formation , ,/ m/e116 -CH,CH 0 ',/ -CdHVOiumle 43 inle 72 SCHEME Mass-spectral fragmentation pattern for (a) 3,5-dimethyl-1,d-dioxacyclopentane (2a) and (b) 3,6-diniethyl-1,2-dioxacyclo-hexane (2b) * Neutral loss supported by observation of the appropriate metastable ion. the dimethyl-l,2-dioxacycloalkane(2) (Table 1). The minor component of the product from compound (la) was shown to be pent-4-en-2-01 by comparing its g.1.c. retention time, lH n.m.r. spectrum, and 13C n.m.r. spectrum with those of an authentic sample; by analogy it seems probable that the minor component of the product from compound (lb) was hex-5-en-2-01.Thus the outcome of the reduction is summarised in equation (10)-l3 (a)F. G. Bordwell and M. L. Douglass, J. Amer. Chem. Soc., 1966,=, 993; (b)B. Giese, S. Gantert, and A. Schulz, Tetrahedron Letters, 1974, 3683. of (M -15)+and (M -44)+ ions can be envisaged as taking place via p-scission in the species produced by cleavage of the peroxide linkage in the molecular ion. On the other hand the formation of (M -18)+and (CH,CO) + ions appears to require rearrangements involving hydrogen migrations. Bromodemerc2Lration.-Bromodemercuration of acyclic peroxymercurials takes place under mild conditions and is generally a very clean reaction.l?l* The organo-mercury chlorides (1) behaved similarly [equation-( 12)] A. J.Bloodworth and I. M. Griffin, J.C.S. Perkin I, 1975, 695. 1978 527 to provide the bis(bromomethyl)-l,2-dioxacycloalkanes unequal amounts of the cis-and trans-isomers can be (3) (Table 1). Partial separation of the diastereoisomers expected in general. occurred in the isolation of compound (3a) and this Using mercury(I1) nitrate, the &membered ring enabled us to group its 13C n.m.r. signals (see later). peroxide was obtained as an approximately equimolar (1) a; n =l b; n=2 The major isomer of compound (3b) was isolated by fractional crystallisation of the 3 : 1 mixture initially obtained. The bromodemercuration products were identified by lH and proton-decoupled l3C n.m.r.spectroscopy (Tables 2 and 3) and by mass spectrometry. Each mass spectrum showed three molecular-ion peaks with relative intensities 1:2 : 1 as expected for a compound containing two bromine atoms, two base peaks with m/e corresponding to (M -CH,Br)+, and strong peaks at m/e 93 and 95 [(BrCH,)+] and m/e 121 and 123 +[(BrCH,CO) i . Peaks corresponding to (AT -BrCH,-CHO)’ and to (M -H,O)+, (M -H,O -CH,Br)+, and (M -H,O -CH,Br -CO)+ (see Scheme) were either absent or very weak. mixture of isomers, while the 1,2-dioxacyclohexane contained about three times as much trans-as cis-isomer. By way of comparison the oxymercuration- reduction of hexa-l,6diene and hepta-l,6-diene [equation (15), n = 2 or 31 is reported to afford dimethyloxacyclo- alkanes with isomer ratios (cis :trans) of 19 : 72 and 61 : 19, re~pective1y.l~ We determined isomer ratios from signal strengths in the proton-decoupled 13C n.m.r.spectra (Table 3). Up to six resonances were observed for each cyclic peroxide and three correlations had to be made. First the lines were assigned to the appropriate carbon atom (to establish the vertical correlations in Table 3). For the 1,2-dioxacyclopentanes this could be done un- ambiguously from chemical shifts and relative intensities R and S Comparison with the mass spectra of compounds (2a and b) shows that the fragmentation pattern is greatly influenced by the presence of the bromine atoms. Loss of BrtH, is the strongly preferred mode of homo- lytic p-scission in molecular ions and formation of +BrCH, can be envisaged by heterolytic p-scission in either M’ or (M -CH,Br) +;[BrCH,CO]+ is presumably formed from the rearranged molecular ion, BrCH,%- [CH,],CH (OH) CH,Br. Diastereoisomerism.-Hydroperoxymercuration at the first double bond of the diene will generate the un-saturated hydroperoxide as a racemic mixture [equation (13)].The subsequent intramolecular peroxymercur-ation can proceed via electrophilic attack on either face [left or right in equation (14)] of the second double bond thereby affording diastereoisomeric products from each enantiomer of the hydroperoxide. Formation of cis trans (meso) (RR and SS) (the ring CH, gives a weaker signal than the other two carbons).For 1,2-dioxacyclohexanes any assignments that were not obvious from chemical shifts could be based on the values of 199Hg-13C coupling constants Me Me [compound (lb)] or on the a-,p-, and y-shielding effects of the substituents HgNO, [compound (4b)l and Br [compound (3b)l as determined from the 1,2-dioxacyclo- pentanes. Secondly the lines were divided into two sets, each 15 H, C. Brown, P. J. Geoghegan, J. T. Kurek, and G. J. Lynch, Organometallic Chem. Synth., 1970/71, 1,7. belonging to a single isomer (to establish the horizontal correlations in Table 3). This was a trivial problem where considerably different amounts of the isomers were present, as with all the 1,2-dioxacyclohexanes and with 3,5- bis( bromomet h yl) -1,2-dioxacyclopen tane (see Table 1).For compound (2a) an unequal mixture of isomers was obtained by reducing compound (3a) with tributylstannane [equation (IS)]. The sets for com-pounds (la) and (4a) were grouped by analogy with compound (3a) and may not be correct. Finally the sets were correlated with configuration and this is discussed below. Conjgurational Assignments.-( a> 1,2-Dioxacyclo-hexanes. The more abundant isomer of the 1,2-dioxa- cyclohexane series was assigned the trans-configuration on the basis of comparisons with related cyclohexanes. A chair conformation (6) is envisaged, like that of the 1,4-dirnethylcyclohexane (7), with the substituents spending nearly all the time in equatorial positions. (i) Evidence from 13C n.m.7.chemical shifts. Analysis of many chemical shift data for compounds of known stereochemistry has provided reliable methyl substituent parameters for methylcyclohexanes.16 Since these para- meters vary considerably according to whether the substituent is axially or equatorially disposed, chemical shifts of the ring carbons can be used to distinguish configurations. Furthermore it has been shown that the effects of methyl substituents on the shieldings of ring carbons in 1,3-dioxacyclohexanes are analogous to, though sometimes quantitatively different from, those for the carb0cyc1e.l~ If we assume that the same J.C.S. Perkin I hydrogens l8 [see structure (S)]. These interactions similarly give rise to a shielding of the axial methyl carbon that is not experienced by an equatorial methyl carbon.Since the methyl groups of cis-1,4-dimethyl- cyclohexane spend 50% of the time in axial positions while those of the trans-isomer are essentially always equatorial, the methyl carbons of the trans-isomer resonate at lower field; the difference is found to be 2.6 p. p. ni .16 In our 3,6-dimet hyl- 1,2-dioxacyclohexane, where there is only one y-methylene in the ring, the methyl carbon resonance of the major isomer appeared 0.61 p.p.m. downfield of that of the minor isomer. A similar effect has been reported for the exocyclic CH,OR groups of cyclohexylmethanols and their derivatives.lS In our cyclic peroxides the downfield shift in the CH,X resonances of the major isomer relative to those of the minor isomer were 1.95 p.p.m.when X = HgCl and 3.82 p.p.m. when X = Br, again indicating a trans-configurat ion. (ii) Evidence from lH n.m.y. chemical shifts. The lH n.m.r. data for the exocyclic groups are also consistent with the major isomer being trans. The sterically induced charge polarisation that leads to shielding of axial carbons simultaneously results in deshielding of the attached hydrogens.20 Thus axial CH,X protons resonate at lower field than do equatorial ones, and trans-l,4-disubstituted cyclohexanes have the up$eZd absorption^.^^^^^ In our 1,2-dioxacyclohexanes the CH2X resonances of the major isomer were 0.25 (X = situation applies in 3,6-dimethyl-l,2-dioxacyclohexane, then the differences in chemical shifts of the ring carbons for the two isomers indicate that the major isomer has the trans-configuration [(6), X = HI.Thus the reson- ances of C, and Cp in the major isomer are respectively 0.78 and 4.54 p.p.m. downfield of those in the minor isomer; in the 1,4-dimethylcyclohexanesthe C, and CB resonances of the trans-isomer are respectively 2.5 and 4.7 p.p.m. downfield of those of the cis-isomer. The large shielding effect (ca. 5.4 p.p.m.) of an axial methyl group on the y-carbon of the cyclohexane ring has been ascribed to charge polarisation in the C-H bond induced by repulsive interactions between the attached l8 D. K. Dalling and D. M. Grant, J. Amer. Chem. SOC.,1967, 89, 6612; 1972, 94, 5318. G.M. Kellie and F. G. Riddell, J. Chem. SOC.(B),1971, 1030. D. M. Grant and B. V. Cheney, J. Amer. Chem. SOC.,1967,89, 5315. 1' HgCl), 0.21 (X = Rr), and 0.18 p.p.m. (X = H) upfield of those of the minor isomer. (iii) Evidence from variable-temperature n.m.r. spectra. Temperature-independent n .m. r. spectra are expected for trans (diequatorial) -3,6 -&met hyl- 1,2-dioxacyclo- hexane (6; X = H), but it should be possible to freeze out ring inversion in the cis-isomer [equation (17)] and so obtain line broadening and then richer spectra at sufficiently low temperatures. lH N.m.r. spectra of a mixture of isomers (ca. 7 : 1) were recorded in the range +30 to -125 "C. At -90 "C the methyl doublet of the minor isomer was broadened much more than that of the major isomer which re-mained resolved at -110 "C; the methyl doublet of cis-l,4-dimethylcyclohexanecoalesces at about -70 OC.,l Changes in the lH n.m.r.signal for the CH2Br protons l8 G. W. Buchanan, J. B. Stothers, and S-t. Wu, Canad. J. Chem., 1969, 47, 3113. 2o B. V. Cheney, J. Amer. Chem. SOC.,1968, 90, 5386. 21 N. Muller and W. C. Tosch, J. Chem. Phys., 1962, 37, 1167. 1978 529 of the major isomer of 3,6-bis(bromomethyl)-1,2-dioxa-exocyclic methylene group. This is based, by analogy cyclohexane did occur in a similar variable-temperature with the 1,2-dioxacyclohexanes, on the idea that the study. However these protons constitute the AB part carbon of a pseudoaxial CH,X group should experience a of an ABX system and the changes are believed to arise shielding due to sterically induced charge polarisation from changes in CjAB due to modification of the rotamer that is not felt by a pseudoequatorial group.However CHzXCH2x II --H T c "Hz X populations [equation (IS)]. That only 8 lines were observed for CHAHRBr at -85 "C is consistent with a trans(diequatoria1) configuration ; the non-equivalent CH,Br groups of an axial-equatorial co,nformation could give rise to 16 lines for the cis-isomer at low temperature. \ CH2 rSr He Again in keeping with a trans-configuration there was no splitting or appreciable broadening of the three resonances in the proton-decoupled 13C n.m.r. spectrum of the major isomer of compound (3b) as the temperature was lowered to -75 "C.Although no single piece of evidence is compelling, the chemical-shift data and the variable-temperature results when taken together present a strong case for assigning the trans-configuration to the major isomer of the 1,2-dioxacyclohexane series. (b) I ,2-Dioxacyclopentanes. The conformations of several 1,2-dioxacyclopentanes with differing degrees of methyl substitution at C(3) and C(5) have been studied and will be discussed in a separate publication.22 The ring has a half-chair or envelope shape with the methyl groups in pseudo-equatorial or -axial positions, as illustrated by the conformers (9) of trans-(Za). This (90) (9b) results in the cis-and trans-isomers of compound (2a) being easily identified by their methylene signals in the lH n.m.r.spectrum (Table 2). Thus it was easy to establish that the major product obtained by reducing compound (3a) [equation (l6)l was the &isomer and so this configuration was correlated with the stronger set of 13C n.m.r. lines (Table 3). The configurations of the remaining 1,2-dioxacyclo- pentanes are assigned on the assumption that each trans- isomer will have the upfield 13C n.m.r. signal for the 22 A. J. Bloodworth and J. A. Khan, in preparation. the same charge polarisation should deshield the attached protons, yet trans-(2a) has the upfield methyl doublet. Other factors must be influencing these chemical shifts and the assignments to compounds (la], (3a),and (4a) should be regarded as only tentative.HA EXPERIMENTAL Commercial penta- 1,4-diene and hexa- 1,5-diene (high grade), deuterium oxide (99.5 atorno/,), and Hg(l\'O,),.H,O and NaBH, (reagent grade) were used without further purification. Other reagents and solvents were as described in Part 10.1 Spectra were recorded as described previously.' When obtaining n.m.r. spectra at low temperatures the solvent for compound (2b) was (CD,),CO-CDCl, (to -90 "C) or CF,Cl,-CH,Cl, (below -90 "C), and for compound (3b) was (CD,) ,CO ('H) or PhMe-CH,Cl,-C,D, ( I3C). General procedures for peroxymercuration, hydrogeno- demercuration, and bromodemercuration are described below and details of individual compounds are presented in Tables 1-3. Peroxymercuration.-To a vigorously stirred suspension of mercury(r1) nitrate monohydrate (10 rnmol) in dichloro- methane (25cm3) was added 80-85% hydrogen peroxide 5-9 mmol) followed immediately by a solution of diene (5 mmol) in dichloromethane (5 cm3).The mixture was stirred for 10 min; a heavy oil separated. The super- natant liquid was decanted. The oil was washed with dichloromethane (15 cm3) and then treated as follows. (a) To prepare a solution for 13Cn.m.r. spectroscopy. The oil was evacuated (12 mmHg, 5 niin) and dissolved in D,O (5om3), and a few drops of 1,4-dioxacyclohexane were added as an internal standard. For the product from hexa-1,5-diene it was necessary to add four drops of ~M-HNO, to get all the oil into solution. (b) To isolate the organornercury chloride (1). The oil was dissolved in water (5 cm3, plus a few drops of ~M-HNO,for the product from hexa- 1,5-diene), an aqueous solution of potassium chloride (10 mmol in 15 cm3) was added, and the mixture was stirred vigorously for 45 min.The whitc precipitate of bis(chloromercuriomethyl)-1,2-dioxacyclo-alkane was filtered off and dried in vacuo. Hydrogenodemercuration.-An ice-cold solution of sodium borohydride (1640 mmol) in QM-so~Iu~hydroxide (20- 40 cm3) was added to a stirred suspension of the organo- mercury chloride (1) (7-9.5 nimol) in dichloromethane (40 cm3) at -10 "C at such a rate that the temperature remained below 0 "C. The mixture was kept at -5 "C for 30 min, and then allowed to warm slowly to room tem- perature; two colourless layers and a bead of mercury were obtained.The organic layer was separated and the aqueous layer extracted with more dichloromethane (40 cm3). The combined dichloromethane solution was dried (MgSO,) and concentrated at >, 12 mmHg. The volatile fraction was collected in a trap at -80 "C and each 10 cm3 portion was examined by lH n.m.r. spectroscopy to confirm that it contained no peroxide. Crude (2a) was distilled (b.p. 32 "C at 19 mniHg) to remove the last trace of dichloro- methane. (a) Chromatography. The dimethyl-l,2-dioxacyclo-alkanes were purified by g.1.c. using a Varian Aerograph 712 instrument fitted with a column (10 ft x 3/8 in 0.d.) of silicone oil on Supasorb (40-60 mesh); 100-200 pl were injected for each run and the carrier gas was nitrogen.For an oven temperature of 50 "C and carrier-gas pressure of 14 lb in-2, the product from penta-1,Cdiene had fractions with retention times 9.8 (pent-4-en-2-01) and 22.4 niin [compound (2a)l. For an oven temperature of 75 "C and carrier gas pressure of 11 lb in-,, the product from hexa- 1,5-diene had fractions with retention times 2.0 (unknown), 13.0 (perhaps hex-5-en-2-01), 20.5 (major) plus 23.0 (minor) [compound (eb)], and 33.0 min (unknown) (Found for fraction 2: C, 69.6; H, 11.35. Calc. for C,H,,O: C, 72.05; H, 12.1%). (b) Mass spectra. The following values of m/e (76 of base peak) were observed. (i) For compound (2a): 102 (22), 87 (9), 84 (8), 69 (22), 58 (18), and 43 (100).(ii) For compound (2b): 116 (6.5), 101 (7.5), 98 (ll), 83 (35), 72 (7), 55 (loo), and 43 (100-t ; this peak was off-scale for the lowest sensitivity galvanometer). Bromodevnercuration.-Bromine (9 mmol) in dichloro-methane (5 cm3) was added to a suspension of compound (la) (4.5 inniol) in dichloromethane (25 om3) and the mixture stirred for 25 h. The solution was then filtered and the dichloromethane removed from the filtrate (12 mmHg). The residue was extracted with a mixture of pentane (90 cm3) and dichloromethane (10 cm3) and the solvent removed from the extract (12 mmHg) to yield compound J.C.S. Perkin I (3a). Compound (3b) was similarly prepared on half the scale using a mixture of pentane (95 cm3) and dichloro- methane (5 cm3) for the extraction.Mass spectra. The following values of nz/e (yoof base peak) were observed. (i) For compound (3a): 262 (12), 260 (24), 258 (12), 179 (5), 177 (lo), 175 (7), 169 (24), 167 (loo), 165 (loo), 151 (lo), 149 (lo), 125 (20), 123 (52), 121 (36), 95 (60), and 93 (64). (ii) For compound (3b) : 276 (6), 274 (12), 272 (6), 181 (loo), 179 (loo), 125 (5), 123 (lo), 121 (9), 100 (25), 99 (18), 96 (13), 95 (32), 94 (13), and 93 (32). Reaction of Hexa- 1,5-diene with Hydrogen Peroxide and Mercury(I1) TriJuoroacetate.-To a stirred solution of niercury(r1) trifluoroacetate ( 10 nimol) in dichloromethane (25 cm3) was added 80-85% hydrogen peroxide (5-9 nimol) followed immediately by a solution of hexa-l,&diene (5 mmol) in dichloromethane (5 cm3).After 15 min the solution was washed with water (15 cm3), dried (MgSO,), and evaporated (12, then 0.05 nimHg) to afford crude organomercury trifluoroacetate. The proton-decoupled 13C n.m.r. spectrum (CDCl,) showed strong lines at 6 79.59 (CH), 32.95 (ring CH,), and 28.30 (CH,Hg) plus six lines 6 27.2-29.5 and 33.75) three or four times less intense, of which two may belong to the second isomer of the 1,e-dioxa- cyclohexane. The organomercury trifluoroacetate was redissolved in dichloromethane ( I5 cm3), aqueous potassium chloride (10 mmol in 10 cm3) was added, and the mixture was stirred for 35 min. The dichloromethane layer was separated, dried (MgSO,), and evaporated ( 12 niniHg). The resultant crude organoniercury chloride was washed with dichloro- methane (3 cm3) to leave a residue of analytically pure 3,6-bis(chlorontercuriornethyl)-1,2-dioxacyclohexane( lb). Reduction of 3,5-Bis(bromomethyl)- 1,2-dioxacyclopentane. -A mixture of compound (3a) (0.9 mniol) and tributyl- stannane (10.5mniol) was stirred at room temperature for 60 min. Trap-to-trap distillation at 0.05 mmHg then yielded compound (2a) (0.034 g, 37%). We thank the S.1I.C. for financial support (to M. E. L.), Laporte Industries Ltd. for the gift of hydrogen peroxide, and Mr. J. A. Khan for a pure sample of compound (2a). [7/1192 Received, 6th JuZy, 19773
ISSN:1472-7781
DOI:10.1039/P19780000522
出版商:RSC
年代:1978
数据来源: RSC
|
|