LIB R.A R.Y

OF THE UNIVER.5 ITY Of ILLI NOIS

Q-547 U6s \963 Suwiwet-

t

Return this book on or before the Latest Date stamped below.

Theft, mutilation, and underlining of books are reasons for disciplinary action and may result in dismissal from the University.

University of Illinois Library

W

•*=*s==5?;/

L161— O-1096

Digitized by the Internet Archive

in 2012 with funding from

University of Illinois Urbana-Champaign

http://archive.org/details/organicseminarab1963univ

ORGANIC SEMINAR ABSTRACTS

Summer, 19&5

Department of Chemistry and Chemical Engineering University of Illinois

SEMINAR TOPICS

Summer Session, 1963

A New Family of Sesquiterpene Lactones

R. Feiertag 1

The Cope Rearrangement

• J. C. Gaal 10

Recent Aryne Chemistry

R. Lambert l6

The Problem of Some Therraochromic Ethylenes

R. Puckett 25

Fragmentation in Solvolysis Reactions

W. F. Pickens # 33

Dimethyl Sulfoxide as Solvent and Reactant

P. Rivers k2

- 1

A NEW FAMILY OF SESQUITERPENE LACTONES

Reported by R. Feiertag

July 1, 1963

Introduction. --Recently several new sesquiterpene lactones have been isolated from plants of the genera Helenium, Parthenium, Ambrosia, and Balduina. It is the purpose of this seminar to review in detail the elucidation of the structures of the two most important members of this new family, tenulin and helenalin, to relate other members of the family to these two, and to review the known stereochemistry.

Before proceeding further, it will be helpful to mention briefly the methods used to determine the carbon skeletons of the compounds reviewed. In general, the compounds were reduced, and then dehydrogenated over a suitable catalyst. Reduction was accomplished by hydrogenation over Raney nickel under moderately \rigorous con- ditions (100° and 1500 p.s.i.)(l), and by the use of lithium aluminum hydride {2) or potassium borohydride ( 3) • Dehydrogenation was carried out at elevated tempera- tures ( 3OO-3600) over selenium (1,4), palladium charcoal (1,3), or palladium (2). In one case (4) the compound was dehydrogenated directly without prior reduction. In all cases, the products were azulenes, whose structures are given,. Several num- bering systems are used for azulenic systems (5). The system shown will be used in this seminar..

I. Guaiazulene II. Chamazulene (6) III Linderazulene (7,8) IV. Artemazulene

Tenulin and Helenalin. — Tenulin, Ci7H2205 (9) , and helenalin, C15H1804 (10), are found in Hel enium tenuf olium L. , (the bitterweed.j and H. autumnale l..> (the sneezeweed) respectively. On the basis of chemical and spectral evidence several formulas have been written for both tenulin and helenalin.

VI

VII

VIII

On dehydrogenation tenulin gives chama: e (II) and linderazulene (III) (3), and hele.ne.lin, chamazulene and guaiazulene (l)(2,4). It is seen that for, structures VI and VIII to be correct a methyl migration must have occurred during dehydrogenation. Analogous migrations are known (11). Moreover, aromatization would provide a strong driving force for migration.

The infrared spectra of both tenulin and helenalin show absorptions near 1585 and 1700 cm"1 (12,3). Coupled with the ultraviolet spectra (tenulin: \aax 226 mu, e 7000; helenalin: ^3, 220 mu, 6 12200; X^x 320 mu, 6 32) , and chemical tests, these indicate the presence of an a, ^-unsaturated ketone. Comparison with the infrared spectra of several conjugated 17-keto steroids (13) indicates the presence of cyclopentenone units.

The infrared spectra of both tenulin and helenalin show bands characteristic of the lactone functions drawn above (1772 cm"1 in tenulin (3); 1750 cm-1 in helena- lin (12)). Helenalin reacts slowly with strong base to form a water-soluble compound and is reprecipitated on acidification (12). These observations indicate the pre- sence of a lactone. Ozonolysis of helenalin yields formaldehyde, which shows the presence of a terminal methylene group (12). If hydrogenation of helenalin is

2 -

stopped after one mole of hydrogen has been taken up, the terminal double bond is preferentially saturated (l) . Subtraction of the ultraviolet spectrum of dihydro helenalin (Xnax 229 mu, € 65OO) from that of helenalin (A^ax 220 m\i, £ 12200) yields a spectrum identical with that of alantclactone ( IX) (^max 210 ^j e 10000) (2,l4) . Therefore, it is probable that helena- lin contains a chromophore like that of alantolactone. IX

Tenulin rearranges to isotenulin under mildly uasic conditions (9A5) unless the hydroxyl group is converted to an ether (l6). Both dihydrotenulin and its oxime rearrange smoothly to the corresponding compounds of the isotenulin series. Therefore, the hemiketal is not linked vinylogously to the ketone group, and the ketone is not involved in the rearrangement, which probably occurs by the mechani shown (3) •

OH®

sm

HO

0

6

l>

0

(J Co

/

0^ XCH^ 0

Although from the data presented thus far tenulin and helenalin might have either of the structures shown, n.m.r. spectroscopy (17) shows that structures VT and VTII respectively are the correct ones for these compounds. Because of the similarity of tenulin and helenalin, only the spectrum of tenulin will be treated in detail. Two vinyl protons are shown by the appearance of two pairs of doublets (intensity one proton each) at t 2.^5 and 3-90- 'This is clearly incompatible with structure V, but is easily explained in terms of structure VTj in which the vinyl protons split each other and are split by the C-l proton. There are four methyl groups in the spectrum of tenulin-- -a singlet at t 8-68 (intensity six protons), a doublet at x 8.73 } and a singlet at 1 8069. One of the methyls at t 8.68 is the masked acetyl, since it moves to 1 7*83 in the spectrum of isotenulin (Xa) in which there are also superimposed methyl doublets and a singlet at t 8.80. The presence of a methyl singlet in the spectrum of isotenulin and three in that of tenulin is

again incompatible with structure V but easily explained in terms of structure 'VT. Additional evidence for the correctness of this structure is provided by the perace- tic acid oxidation of dihydroisotenulin (XIa) and desacetyldihydroisotentilin (Xlb) to the corresponding lactones (XII), in whose spectra are found no new signals in the lactone range (t 4.5-6.5). Therefore, the new lactone adjoins a quaternary carbon atom.

The orientation of the lactone ring may also be determined using n.m.r. A doublet at t k.^-0 (intensity one proton) in Isotenulin moves to t 5-8 on deacety- lation and disappears from the spectrum of dehydrodesacetylisotenulin (Xc) in which the hydroxyl group has been oxidized to a carbonyl. Since the original peak is a doublet, the proton must be coupled to only one other, and the acetate is attached to C-6.. A triplet (J=9~H c.p.s.) at t 4.67 in tenulin and t 5-30 in iso- tenulin (intensity one proton), is further split (J=2-4 c.p.s.) by another proton. This signal is assigned to the C-8 proton* It changes only slightly during the transformations described above. Therefore, the lactone is attached to C-8.

Monobromoisotenulin(XIIl) is prepared by bromination of isotenulin and dehydro- bromination with potassium acetate (9)- An X-ray study of this compound (l8) confirms the assign- ments made with the use of n.m.r. In addition, the complete stereochemistry of the molecule is given. Both five-membered rings are trans -fused to the seven-member ed ring. Non-bonded steric Interactions between the angular methyl group and the hydrogens on C-8 and C-10 cause appreciable folding of the molecule.

Basic hydrolysis of dihydroisotenulin leads to a mixture of desacetyldihydroiso- tenulin and a new compound, desacetyldehydroalloisotenulin (XlVb) , an isomerization which involves reorientation of the lactone ring (19) . The n.m.r. spectra (17) show this very clearly. A doublet at Tj.jk in the spectrum of dihydroalloisotenulin (XlVa)is

XIII

0 CII3X JCAc

VI

XV

i

xvr

Base

(c)

Na2C03

1 . HSCH2CH2SH, BF3- Et20 £. Raney Ki, A

1. 2.

XII a, b<

R=Ae

R=H

HSCH2CH2SH }

BF3«Et20 Raney N1,A

CHpOH

XIX assigned to the C-6 hydrogen, and a triplet at t 4.88 which is further split by another proton, to the C-8 hydrogen. This second peak moves to t 6.06 in desacetyl- dihydroalloisotenulin (XlVb) and disappears when the hydroxyl group is oxidized to a carbonyl in dehydrodesacetyldihydroalloisotenulin (XIVc). Further evidence for this rearrangement is the fact that dihydrotenulin and desacetyldihydroisotenulin are easily isomerized to the corresponding alio- isomers in contrast to the pyrano- side ether of desacetyldihydroisotenulin (17). Both dehydrodesacetyldihydroisotenulin

- 4 -

(XIc) and its analog in the helenaiin series are converted to dibasic a, £_ unsaturated ketoacids (XV) when treated with sodium, carbonate,, while dehydrodesacetyldihydro- alloisotenulin (XIVc) is converted to a monobasic a ^-unsaturated ketoacid (XVl)(l9). These transf ormations } difficult to explain in terms of structures V and VTJ, are seen to be caused by opening of the lactone and cleavage of a (3~diketone in the case of the isotenulin and helenaiin derivatives., In the case of the alio derivative, no 8-diketone is present , and the second reaction cannot take place. A second dif- ference between the two series may be observed- The carbonyl group in desacetyldi- hydroisotenulin (XIb)(l7) and desacetyldihydroalloisotenulin (XIVb)(l9) may be reduced to give compounds XVTI and XVTII. When these compounds are reduced with lithium aluminum hydride > two triols (XIX) which differ in optical rotation and melting point are obtained (17,19) > indicating that epimerization has taken place at C-ll.

Treatment of tenulin with sodium bicarbonate gives a mixture of deacetyl- isotenulin and a new compound , desacetylneotenulin (3)° Neohelenalin, which is found in PL flexuosum Rafo (20) and IL_ mexicanum H.B.K. (21), has also been obtained from the isomerization of helenaiin during chromatography over basic alumina (22). The ultraviolet spectra (neotenulin: ^max 2^0 mu., £ 16000; neohelenalin: A^ax 235 mu, € 1780O; \jjax 208 mu., £ 13800) show that no chromophores have been destroyed, but ozonolysis of these compounds yields acetic acid (ozonolysis of the parent com- pounds does not). The n.m.r. spectra of the two compounds (17,22) show no vinyl protons present in neotenulin and only two in neohelenalin instead of the four pre- sent in the spectrum of helenaiin. In addition, no methyl singlet is observed. Instead> a new doublet (neotenulin: t 8.25;, Jc1.5 c.p.s.; neohelenalin: t 8.32, J=2 c.p.So), intensity three protons, is found- This is most easily explained as a vinyl methyl group split by long range coupling. The conversion of tenulin to desacetylneotenulin Involves isomerization to desacetylisotenulin which is then con- verted to desacetylneotenulin by the following reaction sequence in which the spirane compound may be an intermediate ( 17) ,

XX a. R-Ac (Neotenulin) b. R=H

Hydrogen chloride -chloroform converts helenaiin to a mixture of mexicanin A (21) and neohelenalin (22), Mexicanin A occurs naturally in IL_ mexicanum H.B.K. (21). The ultraviolet spectrum (A^x 212 mu., e 8400) indicates the possibility of an exocyclic double bond conjugated with a lactone as in alantolactone (lX)(l4)„ The infrared spectrum shows a weak absorption at 1630 cm"1 which is assigned to an unconjugated double bond and a strong band at 1750 cm"1 which is assigned to a lactone and an unconjugated cyclopentanone. The n.m.r. spectrum of mexicanin A acetate shows only one vinyl proton, a triplet at t 4.10= The only other difference between the spec- t tra of mexicanin A acetate and helenaiin acetate is the appearance of a triplet (intensity two protons) appearing at x 7.14 in the spectrum of the former. Both dihydrohelenalin and dihydromexicanin A are isomerized to dihydroneohelenalin by methanolic potassium bicarbonate. These observations are best explained by saying

that mexicanin A differs from helenalin only in the position of the eyclopentenone double bond. Mexicanin A may be written as XXIa and neohelenalin as XXII. The conversion of helenalin to neohelenalin and mexicanin A occurs by the following reaction sequence (23). Additional support for this sequence is gained from the observation that helenalin acetate, when treated with hydrogen chloride-chloroform,

XXII

R=H R=OAc

OAc XXIII

XXI a, b, does not isomerize, indicating the necessity for a free hydroxyl group.

Other Compounds of the Helenalin Series. — Balduilin. obtained from Balduina uniflora Nutt. (24), Linifolin A, obtained from IL linifolium Rydb. (25), and bige- lovin, obtained from JL_ bigelovii Gray (26) all have the formula C15H2o05. The infrared, ultraviolet, and n0m.r. spectra of these compounds are nearly identical to those of helenalin acetate, and they have been assigned structure XXIII. Che- mical tests are in accord with this assignment.

Hydrogenation of bigelovin to the dihydro compound reveals that, unlike helenalin, in which the exocyclic double bond is preferentially saturated (l), the eyclo- pentenone double bond is reduced first (26) . In addition, both bigelovin and its derivatives decompose in both aci- dic and basic solutions. This instability may be explained by the ability of bigelovin to undergo retroaldol ring- cleavage, and subsequent further decomposition. However, the reversal of dehydrogenation order as well as the relative stability of other isomers does not appear to have any simple explanation.

Linifolin B, C15H2o05, is found in H. linifolium Rydb. (25). The infrared and ultraviolet spectra reveal that although it has the features of other members of the helenalin series, the conjugated eyclopentenone is no longer present. The n.m.r. spectrum is similar to that of the acetate of mexicanin A (XXIb) , and on this basis it has been assigned the same gross structure.

Rarthenin (XXIV) C15H1804, is found in Parthenium hysterophorus L. (27). Reduction and dehydrogenation yields artemazulene (IV), and the infrared and ultraviolet spec- tra show a lactone and a eyclopentenone, both conjugated with double bonds (28). In addition, a tertiary hydroxyl group is found. It is resistant to chromic acid oxi- dation and acetylation, and parthenin is easily dehydrated. The n.m,rc spectrum of parthenin shows a methyl singlet at % 8.72 (C-5 methyl) and a methyl doublet at t 8.89 (C-10 methyl) (27,29) o When parthenin is dehydrated with anhydrous formic acid to anhydroparthenin (XXV), the methyl doublet disappears, and a singlet t 7.97 appears. This is in the vinyl methyl region, and indicates that the hydroxyl group is a to C-10, at C-l.

Formation of artemazulene suggests that the lactone ring is closed to C-6. This is confirmed by the n.m.r. spectrum which shows a doublet ("intensity one proton) at t 4.92, This is in the region of protons on carbon attached to a lactone or ester.

c

- o

COOH I

CH3-C-H I

CH2 I CH2

COOH

XXVII

.KMn04 H^

,03,CH3OH 3/ ^ —

XXIV H2, Pd-C

XXV

XXIX 0

Further evidence for this structure is provided by the ozonolysis of parthenin in methanol at -780 (27). Permanganate oxidation of norparthenone (XXVI) , the com- pound formed, yields S-(+) -O-methylglutaric acid (XXVTl)(30). This is obtained from the fragment of norparthenone between C-7 and C-l, and demonstrates that the C-10 methyl group is p.

Coronopilin ( XXVIII) is obtained from Ambrosia artemisiifolia L. , the common ragweed (31) . Both parthenin and coronopilin give dihydroisoparthenin (XXIX) on hydrogenation. On the basis of this reaction and spectral evidence it has been iden- tified as 1,2-dihydroparthenin.

Ambrosin (XXX), CisHxsOs, and damsin, C15H2003, are found in A. maritima L. (32).

Reduction and dehydrogenaTTTon or ambrosin

gives artemazulene (IV)(33)° The infrared spectrum shows the presence of an a, 3- unsaturated r-lactone and an CC,3-unsaturated cyclopenteneone (32). However, there is no absorption in the hydroxyl region. Hydro - genation of anhydroparthenin (XXV) yields two hexahydro derivatives, one of which is identical with tetrahydroambrosin (27). The n.m.ro spectrum of ambrosin shows the same features as that of parthenin with the exception of the vinyl peaks. These are each split into pairs of doublets. It is apparent that the vinyl protons are splitting each other and are being further split by the C-l proton.

Hydrogenation of ambrosin yields a dihydro and a tetrahydro derivative (32). In the infrared spectra of these compounds the carbonyl band moves to 1760 cm-1, indicating that this group is no longer con- jugated with a double bond. Damsin, although it contains an unconjugated cyclo- pentanone (infrared: 175^+ cm"1), is not identical with the dihydro compounds However, it yields a mixture of the dihydro and tetrahydro derivatives on hydroge- nation (3^)» ^Q dihydro compound is therefore dihydroisoambrosin (XXXl), and damsin has the structure XXXIII. This is confirmed by the n.m.r. spectrum (35) which has two vinyl doublets unsplit by any other protonc

Stereochemistry., --With the exception of bromoisotenulin (XIIl) and isotenulin (Xa), the complete stereochemistry of the compounds in the helenalin series is not known. However, a number of optical rotatory dispersion studies have been made which allow partial determination of the absolute configurations.

- 7 -

The isomerization of tenulin ( VI ) to isotenulin (Xa) destroys only the asymmetric centers at C-ll and C-12. Therefore the stereochemistry at C-l, C-5, C-6, C-7, C-8, and C-10 is the same as that of bromoisotenulin (XIIl).

The optical rotatory dispersion curves of balduilin (XXIIl)(.23), helenalin (VIII), tenulin (Vl), and isotenulin (Xa)(36) all show strong (-) -Cotton effects. Hydrogenation of the cyclopentenone double "bond causes this effect to he reversed; the curve of dihydrohelenalin has a (-) -Cotton effect, but those of tetrahydro- helenalin, dihydrotenulin, and dihydroisotenulin all have (+) -Cotton effects (36). The curves of tetrahydroparthenin, tetrahydroambrosin (XXXIl), and coronopilin ( XXXIII) also have (+) -Cotton effects (27). Therefore, all these compounds have the same configuration at C-l and C-5°

Tetrahydromexicanin A obviously differs from the others at either C-l or C-5, for its curve shows a weak (-) -Cotton effect, v/hich almost doubles on acetyla- tion (23) . It has been argued (23) that this is evidence for opening of the seven- member ed ring in a reverse aldol reaction and reorientation of the C-5 methyl group rather than a simple migration of the double bond., However, it is possible that hydrogenation of the C-l, C-2 double bond has caused the configuration at C-l to be reversed.

Isolation of S-(+) -a-methylglutaric acid (XXVTl) from the oxidation of nor- parthenone (XXVI.) demonstrates that the C-10 methyl group in parthenin (XXIV) is P (27) 'The equivalence of one of the hexahydro derivatives of anhydroparthenin (XXV) with tetrahydroambrosin (XXXII) shows that ambrosin (XXX) and parthenin (XXIV) have the same stereochemistry at C~5, C-6, C-7, and C-8, since only the asymmetric centers at C-l and C-10 are destroyed during the conversion of par- thenin to anhydroparthenin (27).

A series of chemical reactions has been used to establish that helenalin (VTIl) and balduilin (XXIIl)(24) are C-8 epimers. Saponification of tetrahydrobalduilin ( XXXIV) is effected by refluxing with potassium hydroxide. Under these conditions two isomers, desacetyltetrahydrobalduilin A (XXXV) and desacetyltetrahydrobalduilin B are obtained. Reacetylation of these compounds yields acetates different from ■ tetrahydrobalduilin (XXXIV), indicating that epimerization at C-ll and/or reorienta- tion of the lactone ring has taken place.. Chromic acid oxidation of tetrahydro- balduilin A (XXXV) yields a product identical with that obtained from allotetrahydro- helenalin ( XXXVII I. ). Only the C-8 and C-ll asymmetric centers have been altered by these reactions. Therefore, balduilin (XXIIl) has the same stereochemistry as helenalin (VTIl) at C-l, C-5, C-6, C-7, and C-10. This leaves only the relative configuration at C-8 in doubt, since in the parent compounds there is no asymmetric center at C-ll. Balduilin is different from helenalin acetate, and the two com- pounds must be C-8 epimers.

Treatment of the thioketal of tetrahydrobalduilin (XXXIX) with Raney nickel

- 8 -

L. Raney Ni,&

XXXIX

followed by saponification with sodium hydroxide yields compound XL in which the lactone ring has been reoriented. Oxidation of this compound with chromic acid

gives dehydrodesoxodesacetyldihydroalloisotenulin (XLl)(l9)o Therefore, balduilin and helenalin have the same configuration as isotenulin at C-l, C-5, C-6, C-7, and C-10. Helenalin may now be repre- sented by XLIIa and balduilin by XLIIb.

BIBLIOGRAPHY

1. R. Adams and Wo Herz, J. Am. Chem. Soc„ , 71, 2554 (19^9).

2. G. Mchi and Do Rosenthal, Jo Am. Chem. SoCo, 78, 3860 (195©). 3» Do Ho R„ Barton and P. deMayo, Jo Chemo SoCo, l42 (1956)0

4. Vo Herout, M. Romanuk and F, Sorm, Coll. Czech. Chem. Commun. , 21, 1359 (1956)0

5o Mo Gordon, Chem. Revs., 50, 127 (1952).

60 A. Meisels and A. Weizmann, J. Am. Chemo Soc. , 75, 3865 (1953) .

7o K. Takeda and W„ Nagata, Pharm. Bull. (Japan), 1, 164 (1953); C.A. , 48, 77l6g

(1954).

K. Takeda, Ho Minato and M, Ishikawa, Tetrahedron Letters, 3, 121 (1963)0

Eo P. Clark, Jc Am. Chem. Soc, 6l, 1836 (1939).

Eo Po Clark, J. Am. Chem, Soc, 58, 1982 (1936).

(a) J. Simonsen and D. H. R. Barton, "The Terpenes", Vol. Ill, Cambridge Press,

Cambridge, 1952, p. 34; (b) J. Simonsen and D. Ho R. Barton, "The Terpenes",

Vol. Ill, Cambridge Press, Cambridge, 1952, p. 84.

R. Adams and W. Herz, J. Am. Chem. Soc, 71, 2546 (1949).

R. N. Jones, Vo Z. Williams, M. J. Whalen and Ko Dobriner, J. Am. Chem. SoCo, 70,

2024 (1948) .

C. Asselineau and S. Bory, Compt. rend.,, 246, 1874 (1958)0

E. Po Clark, J. Am. Chem. Soc', 62, 597 (l$Pk)) .

Eo P„ Clark, J. Am. Chem. Soc, (d2, 2154 (1940) „

Wo Herz, W. A, Rohde, K. Rabindran, P. Jayaraman and N. Viswanathan, J. Am. Chem.

Soc, 84, 3857 (1962)0

Do Rodgers and Mazhar-ul-Haque, Proc„ Chem. SoCo, 92, (1963)°

Bo Ho Braun, W. Herz and K. Rabindran, J. Am. Chemo Soc, 78, 4423 (1956).

Wo Herz, R» B. Mitra, K» Rabindran and W. A. Rohde, J. Am. Chem. Soc, 8l, l48l

(1959).

A, Romo de Vivar and J. Romo, Chem, and Ind» (London), 882 (1959). Wo Herz, P» Jayaraman and Ho Watanabe, J» Am0 Chem. Soc, 82, 2276 (i960). Wo Herz, A. Romo de Vivar, J. Romo and N„ Viswanathan, J. Am. Chem. Soc, 85, 19 (1963)<

W. Herz, R. B. Mitra and P. Jayaraman, J. Am. Chem. Soc, 8l, 6o6l (1959). Wo Herz, J. Org. Chem., 27, 4043 (1962).

B. A. Parker and T. A. Geissman, J. Org. Chem., 27, 4127 (1962)0

- 9 -

27- W. Herz, H. Watanabe, M. Miyazaki and Y. Kishida, J. Am. Chem. Soc, Qk, 2601 (1962).

28. W. Herz, H. Watanabe and M. Miyazaki, J. Am. Chem. Soc. , 8l, 6088 (1959).

29. W. Herz, M. Miyazaki and Y. Kishida, Tetrahedron Letters, 2, 82 (1961).

30. A. Fredga, Arkiv. Kemi Mineral Geol. , 24A, No. 32 (19V7).

31. W. Herz and G. Ho'genaur, J. Org. Chem. , 26, 5011 (1961).

32. H. Abu-Shady and T. D. Soine, J. Am. Fharm, Assoc, k2, 387 (1953).

33. F. Sorm, N. Suchy and V. Herout, Coll. Czech. Chem. Commun. , 2k, 15^8 (1959)

34. H. Abu-Shady and T. D. Soine, J. Am. Eharm. Assoc., k£, 365 (1954). 35* R« B. Bates and V. Herout, private communication.

36. C. Djerassi, J. Osiecki and W. Herz, J. Org. Chem., 22, I36I (1957).

0 - The Cope Rearrangement

Reported by J.C. Gaa.l July 22, 1963

Introduction* The thermal rearrangement of substituted 1,5-hexadienes was first reported by Cope and Hardy in 19^0 (l) for a series of substituted malono- nitriles and malonic esters, as in equation 1. The reaction was proposed by Cope to be similar to the Claisen rearrangement of phenyl and vinyl aiiyl. ethers but

rc=a ^x

- :;

'n /A A__, -C - C = C' X = COOCaHs or GN

/ \ I mT „TT 1 - COOCJIu-, or CN ^ x"

CHs=CH-CH2 N> CH2-CH=CHa

with the oxygen atom replaced by a carbon atom.

Scope of the Reaction; The rearrangement was farther examined by Cope and coworkers and was shown to occur for a v y of 1,5 -hexadien.es with and without electronegative substitute (see Table l) , although these substituents enhance the rate (2), Vogel and coworkers (3 A) report that the reaction also occurs with cleavage of a cycloprcpyl or cyclobutyl ring, so that cis-l,2-aivinylcyclobutane (I) yields c i s, cis -1 , 5 -cyclopc tad iene (II) at 120° C. Doering and Roth (6) have

CD \-Kj> ^J-20^ . j (ir,

pointed out that the cyclopropyl ring cleavage greatly increases the reaction rate, due mainly • 19,1 kcal/mi released hy opening this ring.

In tramole cularity and Allylic Requirements g Cope, Hofmann and Hardy showed that the rearrangement is intramolecular by following the reaction of mixtures of ethyl (1-methyl I ~2~butenylmaIonate (III) and ethyl (l»methyl-l<-hexenyl) - allylcyanoacetate (IV) which have similar rates (7). These reactions resulted in

(]; CHa-C^c^C00C^a (IV) i*tte-Ui-^c • W1

CK3CH=CH-CHS^ ' ~"C0QCaH5 CHgaCH-CHa' COOCgHg

no cross products within their experimental error (the products were separated by fractional distillation). It should be noted that the structure of every reactant and product reported by Cope was demonstrated by the chemical methods available at the time,

Cope and coworkers have shown that the reaction requires the reversal of the ally! group with accompanying double bond shift and that it can be sterically in- hibited by groups on the carton to which the allyl is becoming bonded (7,8) .

Kinetic Information; Cop;- (5) and Walling, and Naimen (9) report first order kinetics for the reaction, A negative AS? of about -11. 5 e.u. was evaluated by Cope which he states is to be expected for a mechanism involving a cyclic transi- tion state due to loss of degrees of freedom from restricted atomic motion. This value can be compared with those for the Claisen rearrangement (ca -12 e.u. from selected reactions run by White, et al.(10)) and the Diels-Adler reaction of cyclopentadiene (-29 e.u.) (II).

Another parameter which .can be used to distinguish various reaction types is the volume of activation (AVH , which is the change In volume when a compound goes to its transition state from its ground state. Hammann lists various reaction types and, the expected sign of the AVr (12). In a series of high pressure experi- ments, Walling and, Naimen report that pressure enhances the reaction rate of the Cope rearrangement so that the AV/* is negative. Unfortunately, the usual method of treating pressure -rate data has been severely criticized recently by Benson and Berson (13), so that the exact value of AVr from Mailing's data depends on the

CQ -P d

0

B

0

to cd

U

a

0

K

0 ft O

u

0

ft

CJ

o

H

00

ft 0

H ■§

1

0

a)

CD

'0'

P

CD 05

CD

d

4-3

"£ ^

c

P P

P

CD g

i

0

j

cd (\)

>.

cd

-H

P> CD H

. — ^

'T5

p a

s — ,,

p I

H

d -p B

0

ft

"0* d 0

ft

0)

P

d

0 ft

g

0 i-o

a>

Q) ^-v

■

S is S

0

d

H

o o

d

O CD

d

o d «

d

I

0

>i

a? a

CD

05 d

(t

H O ^->

0

* — V

'd

id

O o5

T3

O CD

cd

H

cd H CD

-H

0

ft

0

C ^

•H

d t3

a

H

add

■d

d

0

ft

0

ft

CD

p

a5

t > a

: i

05 o

H

(U

cd <H

CD

i

Ss

5 g CD

cd

0

p

>>

•p

d

P

>, a

>»-P

^«H

P

0

P

^-s 8 ro

p

T3

Ctj

C

cd

CD

Cd

o ^-^

d

o

O >i

cd

d

■H

CD .---. !-i

CD

o

0

ft

-p

0

P

ft -H

! a>

a>

t

p

CD

C

d CD H

d

o

0

d

H

o

-p

0)

1

<-■

.-x d

p

CD

>-«. cy

CD

Ki

o

a d >3

0

!

d

0

>5

o

d

o

ft"

CJ

0 cu

Ei

O

a) p

CJ

°H

0.

■Ci CD d

°H

V£>

0

•H

d

cd

0

'0

B

cd

cd o d

cd

c»

0 TJ

CD

id

d d

cd

H

o

Cl CD

Tl

o^.

•H

Td

0

o

ft

o

ft

c

CU ft

ft

o

CD (U

o

'=>,-

>— ! >H P'

cd

cvj

0

T3

cd

ft

d

!

c

>s

d

<d ft

9

d

>a ft

c

d

cd

P-s H d

X

3 ,

c

cd

P1

d

d ft

cd

ft

Cd

.ft

1

ft >5

~t

05

»H I

n

<D

a

d ?>:> CO

0

CO

0

•p

d

0

>>

I

>5

O

'-■--

ft fl

>• r-

•

4->

a> d ft

XI

0

°H 0

H

d

0

ft

o

H

o

JH

CJ

H

-P d Xt

>> 0

.-J

o

^ s

CJ

d

P CD

a

d

o id

T i

0

ft

i

>i

ft

X ft

>>

d h

cy

d

d p> ^*

fl

ir\

CD

*\ 0

cd

fio

Q

-=fr

XJ

o

CU d

rQ

cd >s

ft

a;

CD d '

CD'

•»

o.H

W H

p

o

H

0

-P

CQ

x a>

P>

•a »d

i

'■d

ft CU H

»H

H

^O

• TO

o

0

H

CJ

H

0

°H

O ft

CD

d -p

.^h

••H

a ft >> T3

i

cd

olcd

O

d

CJ

>3

I

"

H I

s

>H CD

8

H -d-

cd

H

p

0

>5

o

ft

-S

CVi

CM

o ^t

a

i B

H

>5

1 -=t -p

M

>?

ft

Tl O

H

»H

o

=H

o

-P

8

>. i

CJ

H »H

>5

X

— t S CD

CD

-Q

a;

d o

CJ

n3

°H

^d

d

0

H

>>

ft

>>

0

a

o H

i

8 U

,-d

CD

>»h a

3

CD

X!

p. \o

:--

cd

^

>=>

-:j

B

>?

H >».-

H -p

p

d

d >s «h

«

XS

i

V

-p

>i

X

O

H

X

X*

>^. d

i>3

>i f

0!

D

P Xi T3

LT>

ftu-

.2 i lev? Scs i

'.

ft

X

o

V

■P

P

H CD

X!

H K>

a

r— 1

CD -P 6

•>!

i

°>.

IA

0

o

fn

1

0

<y

H ,£l

:P

rH •>

8

O

a CD K\

H

*>

ill

?H

"P

CM

B

S=!

cd ft

<U

cd 0,1

CU

>i

i ^

6

s

H

d

'd

>5

«*.

!

s

g

s.

S

CJ

CVJ CVJ H

H

l

,H

?>s

X5

H

r- 1

H

CU H

r— j

OJ H

H

.--,

• <•■ — * • — ».

S>s

>?

>>

CQ

CJ

X!

' j

v *

s — •*

N-— ''

x ^ v ^

N^^'

% ^ v^^*»

!>s i>

>:^

X3

P.

H

>5

a

CO

ft

i — i

i— ! :>>

ft*

>

ft -P

a*

H H

H

—* H

,d

H

•H H .-1

CD

P

0

CJ

lo

CO

6

>^

i>>

>s >^

>5

'>> >s

P

H

?*1 PS ?v

-

CL;

&

-T d ">

(*

s

>i\

xj

Xl

X

i^ ,X^

X!

tEJ X!

CD

cd

.d a' i

ft

a

ft

, r-, CO fJj

HP>!o

CQ

F

■&

-P

ft 0

-P

0

P> P

cd a;

■P

QJ

p -P

3

CVJ

-P 4-

w w w

H

3

s— !

H O

>3

CO

cdi

ft 0

d

H

0

ON

03 XI

o

B O,

Eh! VO

o

H

cd

p

0

o

cd

G

d

CO

H

cd

i

H*

d 0 ft o

?H ft

H

>5

x!

„ P>

d a

d *

O H

O r-\

O >s

XI

-P

in.

CO

w

u XI

CO

0

p

cd P 0 o cd o d

CO CJ

H

>_

H H

cd

H

,d -p 0

CVI I

H

x| P>

0

— i X

4-'

!

H CO

w xj

ON

o

LO..

M E, Xi

ON

O

o	o
t-	o
H	o
	cvi
0
p
cd	0
P	+j
0	cd
o	■p
cd	0
o	o
d	cd
CO	o
>i	d
CJ	CO
H	>s
>-,	CJ
H	ft
ft	>s
cd	H
i	ri
s »	cd
H	s
>j	^-~.
d	ft
•H	>>
>	d
H	»h
>s	j>
ft	ft.
O	>>
Sh	p>
ft	d
O	Xl
CQ	»
•H	d\|
CVi I	CM
H	ft
>s	>s
xl	XJ
P	■P
0	0
a	a
H	.— 1
— '	—
ft	ft
>*	>s
X	XS
P>	-P
fa	w

Lf\ in, .'x>,

f- CO O On !>- tr-

ee CQ CQ

!h ^ 5h XI X X-

8 CO

o o o o o o "o o o

t— c— o

ft ft CVJ

cd •p

0 CJ

cd o

d

CJ

0

0 P ,

cd +;

4.1 cd

S 0 ft « ^ ft^

O 05 ft

rt O H

cd d cd

S, d s

h 3 d

>.. ft >>

H H °ri

^ d ^

8 cd ft ___ i >-,.

^-.-. xf R>d -^

C? >> CJ

II?

0 s> CVJ xn ( — - J O >j f-1

CJ 0 X! >»X! ft bJ ft o>

g G if

ft H ft

— * — r i — >.

>; >• r>S

d fl ^

O f-

CM -^-^

KN O NO CO

ca ra 5-i f-i X XJ

o o

o o in o Co co

ft ft

0

Ctj P

0 CJ

cd o

, r in in b- cm" cm cm ^J\o'no"VS on

CO'CO'

r- 0-

O. ON 00

ra xl

XS CM

.-T ft

8

ft

ft +

co in. cvi cm O ft 5 O- f- lo

O S? to" ?VJ Oi ^> ' ONVO On f— O- On.

Q

5 S O: ON. ON.

ft d

cr 0

0

i/j KJ CQ CQ CQ CQ CQ U 5h 5h U 5-( fH !m

xj xS X X x! X Xl

CO OCMV0 1AH4 ft CVJ NO KN CVi

o o . cd •

CQ CQ cq x] r/J

!h 5-i 5h ft d W «

X X ,X ft h ^

CQ a ft! ^ co ,-cf- m co

- cvj 0 o Jn to.

ft CM

o o o o o o o o o o o o o o o o o o o m o o m co in in m (n. hooco t>-co co o

ft ft CVJ CVI ft ft ft ft KN

no m o o

ft ft ft m,

o o o o o o o o O O O lo,

CO CO CM .ft

ft CVI ft 8

o o

o o o o

ft" ft- ft ft

0

p

cd

ao

4-' CJ

0 H

o S

•Ti P

o o

d in

cd o

>> 5

CJ ..— ..

ft ft

ft p

ft 0

cd ft

i O

ft ft !>»ft

d >5

0-' X

=d -P

d 0

m ft

H ft

X xi p'

ta

Hq

0

ft 0

ft ft

U ft

p 5h

ft ft

d "H

O d

d o O pj ft o

Cd rH

II

ft. .^ ft ft

cd ft

s cd

ft 8 >sft

d >s

0 d ft 0

o x

ftx

ft O hH

ft >3 0 C.)

0 0 0

~p 4-3 ft

n5 cd cd

d d d

o o o

ft ft ft

cd o5

0

d

CD

d

05

0I!

0 d ' a 0

c 0 m c

0 ft »\ <i>

ft 'd ft -h

fd d 8 T3

cd -^ ft cd 0 0 d 0

ft X. 0 X!

* ' XS «

lo in ft lo

ft XJ p< ft ft ^j- ft

ss^	M
>> >5	>s
ft ft	p
ft ft	O
cd cd	5h
i	O 3
ft ^-n	ft
>> ft	f>s
d >s	d
0 d	0
ft CJ	ft
O ft	o
^ d	u

H ft

ft ft ft

>:, >S >-

,d x x

E^i S U.

8 I

ft ft ft ft

Ps P3 ?*5 »*>

d d x x

0 0 P P XS X! fl> 0

p., ft a a

8 S 8 f.

m m kn m,

0 CD

d d

0 0

ft ft

'd jd

cd cd

Pi X

0 0

X XJ

" i m,

in *••

*. ft

ft s

i ft

ft >»

>sXj

,d ft

ft 0

a fl

•rl ^

T5 I * ft'

ft- <* <». KN

KA !'

8 O

O CQ

id 0

u a

0

d cd

•P 0 _j d

•9 S

o ft ft o CJ 5h >s ft

cj O ft ft

>i CJ

d !>»

ft O

> d <d d

CM & ' H

ft ft 1,5,

co ra

ft "H

olejs

0	0
d	d
0	0
ft	ft
T3	>d
cd	cd
ft	■P
d	d
0	0
ft	ft
O	O
ft	ft
o	o
>5	>i
o	CJ
ft	ft
'R	%
K	X
O	c
5h	U
^d	Ti
>5	>>
X	X

CQ ft

ft

d

CO

B

ft

a

ft

CQ

X P

m >s

cd

is

X ft

d

5h

+

0 >oa.

CQ Jh xj

CM

>

s

ft •p

0

CD

d

CQ CQ

5h O

X

$

cd

0 X

d o

ft

O

ft

XS

CJ

a

CO

ft

CQ !h

X!

LO-. *

mathematical interpretation that is used. In spite of this interpretive problem, an important feature of the results is that the value is not solvent dependent. Benson and Berson also suggest that temperature effects due to adiabatic compres-

n of the liquids could partially account for the small changes in rate which are observed by Walling.

Mechanism and Stereochemistry: Cope postulated a mechanism involving a cyclic transition state similar to the Claisen rearrangement for this reaction and showed the reversal of the point of attachment of the ally! group during the reaction ( see

Dve) Walling and Naimen interpreted their negative AVr as suggesting an exten- sively bonded, cyclic transition state which is similar to the product.

Taking model compounds from Egloff (lk) , a difference of -16 to -18 cc/mole is calculated for linear and cyclic compounds which are somewhat similar to starting materials and cyclic transition states. Walling suggests that the values of AVr should be somewhat less negative than these due to stretched bonds in the transition

te (11,15). It seems that the field of pressure -rate studies as applied to organic mechanisms is on shakier ground than the mechanism of the Cope rearrangement,

that it can scarcely be used as a valid criterion for the nature of the transition state (

Stereochemical details for 3 ^-dimethyl -i ,5-hexadienes have recently been re- ported in a series of papers by Doering and Roth,, They suggest that for hexadienes which have a choice of transition state geometry a four-center, chair -like arrange- ment is conformationally favored. The rearrangement of the 3 ,^-dimethylhexadienes to octadienes by the four- and six-center cyclic conformations proposed by Doering and Roth are shown in Figure 1(6).

..go

iter

-fM

"A

■ nter

>«, ni

n

'-%.

*i

: I

ra

H

$k

«JL

<—

SI

~-/e%

1 i.§Lli§"

As is snown, if the meso- isomer follows a four -center mechanism the product be cis ,trans~2,6-octadiene and if the six-center path is followed, then cijs, cis- and trans, trans -2, 6-octadiene o.re the products. On the other hand racemic starting materials result in ,cis ,trans -2 ,6-octadiene if the six-center path followed and the four-center method leads to ciSjCis- and trans , trans -2 ,6-octadiene . Meso and racemic-3,ii--dimethyl-l,5-hexadiene were separated by gas chromatography structure proofs effected by ozonolysis and oxidation to the known 2,3-dimethyl- succinic acids (16). Structures of the rearrangement products, the octadienes, were

L3 -

inferred from infrared spectra following their separation by gas chromatography,, The octadienes do not inter convert when heated to 230 JC. for 2k hours.

The product from the rearrangement of the roes o -isomer after six hours at 225 C. was a mixture of 2,6-octadienes consisting of 99 °l1° cis^trans- and O.jfo trans, trans - isomers o The calculated difference in free energy of activation implied by these yields favors the four-center path by 5.7 kcal/mc.le, The racemic- 3, k- dimethyl -1,5- hexadiene rearranges after 18 hours at l80 Go to 90$> trans , trans-, 10$ cis, cis- and less than 1$ cis jtrans-2,6-octadiene. Comparing the two four-center paths shows that the trans , trans -isomer is favored by approximately 2,0 kcal/mcle, which is not un- reasonable if the "axially" oriented methyl groups necessary for the cis ,cJLs -isomer are considered (6,l6) «,

Qua mechanical considerations suggest that the interaction of allyl radicals in the six-center path is less favored than the four-center one because the second it

Ltal of trie radical which contains the lone electron has a node at the central carbon atom (17) • On the other hand, the electron density is equally distributed over all three carbons in the first it -orbital (filled) « Tiiere is, therefore, always repulsion between the two central carbons of the radicals but there is a possibility

bonding forces between the terminal carbons which can overcome the repulsive force So

Not all Cope rearrangements have a choice of transition state geometry, however,, Rearrangement of ei_s-di.vinylcyclopropa.ne to I,4--cycloheptadiene (see above,? and re- actions of a- and 6-l~hydroxydicyclopentadiene (V and VI) to syn- and ant i - 8 -hydr oxy- d: cyclopentadiene reported by Woodward and Katz (l8, must follow the six-center mechanism (l6) . Doering and Roth tried to isolate cis-divinylcyclopropane formed by the^addition of methylene to cis ,cis ,cis-l,3 9,5 -hexatriene , but report that even at 4; C. the rearrangement to eycloheptadiene is too rapid to allow isolation of the cyclopropyl compound (19) e

(v

j'F'luctional" Isomerism ; Upon investigation of bicyclo (.5.1.0) 2,5-octadiene (IXa and IXb) , Doering and Roth report that, by using n„m.r„ spectra, they can ietect a rearrangement at -50 Ct which occurs approximately once per second per molecule and a1 l80"C occurs approximately 2.50 times per second per molecule „ Since the bieycTooctadiene is symmetrical, its rearrangement proiuct is the same as starting material and it has therefore been called a "degenerate" rearrangement, Examination of the cis- (IXa.) and trans -conf ormers (IXb; shows seven different types of hydrogens which might be differentiated by n.m.r. (labeled on IXb), but because the cis- c onf orme r undergoes rapid Cope rearrangement, actually only four types ar= observed (at l8o~C„'i due tc averaging (labeled a,b,x,v on IXa) „ A 2,6-

I

(IXb)

i)

- Ill

trans annular interaction is proposed in the cis- conformation which makes the trans - conformer the major contributing structure at -50 C, , but as the temperature is increased this energy barrier is overcome to yield the cis -conformation which may undergo rearrangement so that at ca 20 C, the n.m.r, spectrum shows two smeared peaks

Because the compound undergoes such rapid rearrangement, Doering and Roth have called it a "fluctional" structure,, They prevented the possibility of conformational inter conversion due to transannular repulsion by preparing the ketone tricyclo(3,5„l,0) 3 -2,7-dien -6-one (X) which is locked in the favorable els_- conformation. As a

result of very rapid rearrangement, this compound shows only three types of hydrogens (instead of the expected five) ir the n.m.r. spectrum even at -60'C

The limit of a "fluctional" structure in this series as proposed by Doering and Roth is tricyclo (3«3«2„0) deca~2,7,9-triene (XI) for which the trivial name "bullvalene" has been suggested. If the Cope re- arrangements of this molecule were very rapid (as they should be since it is held

1

in the favorable cis -conformation) , then all ten carbons could exchange positions (see Fig, 2) so that the n„m„r„ spectrum cf !*bullvalene " should show only one sharp peak which is an average of the properties of the protons in the following distri- bution among the various positions'; l/lO bridgehead character ^ 3/10 cyclopropyl character, 3/l0 vinyl character (next to the bridgehead.' and 3/10 vinyl character (next ■ yclopropyl ring' (19) ce in Fig, 2 that any cf the forms can be inter converted by an appropriate number of Cope rearrangements, Thus, for example, letting the bridgehead carbon be C-l, and numbering as in Fig, 2, the next bridge- head carbon must be either C-4, C-5^ or C-10, It would take two rearrangements to get C-2, C-7* or C~8 ; bridgehead and three reactions for the remaining three carbons. This means that the more than 1,2 millj n isomers of "bullvalene" are 1 ■■' ■ . 1 r i 'ie„

2

•^r<?

- 1.5 - Bibliography

1. A. Co Cope and E.M. Hardy, J. Am. Chem, Soc, 62, Ukl (19^0).

2. AoCo Cope and H. Levy, ibid . , 66, 1684 (19^0 .

3. E. Vogel, 0. Rods and K.H. Disch, Angew. Chem., 73, 3^2 (196l) . k. E. Vogel, K.H. Ott and K. Gajek, Ann., Gih> 172 fl96l) .

5. A.C. Cope, K.E. Hoyle and Do Heyl, J. Am. diem. Soc, 63, 1843 (19^1).

6. W. von E. Doering and W.R. Roth, Angew. Chem., 75, 27 (1963) .

7. A. C. Cope, CM. Hofmann and E.M. Hardy, J. Am.. Chem. Soc, 63, 1852 (19^1).

8. E.G. Foster, A.C. Cope and F. Daniels, ibid., 69, 1893 (19^7).

9. C. Walling and M. Naimen, ibid., 84, 262HT1962).

10. W.N. White, D. Gwynn, R. Schlitt, C. Girard and W. Fife, ibid., 80, 3271 (1958).

11. C. Walling and H.J. Schugar, ibid. , 85, 607 (1963).

12. S„ Hammann, "Ehysico -Chemical Effects of Pressure", Acad. Press, Inc., New York, N.Y., 19?7, Chap. 9.

13. S. W. Benson and J. A. Berson, J.Am. Chem. Sec. , 8^„ 152 (1962) .

14. Go Egloff, "Physical Constants of Hydrocarbons", Reinhold Publish. Corp,, New York, N.Y., 1939* vols. I, II.

15. C. Walling and D. D. Tanner, J. Am.. Chem. Soc, 85, 6l2 (1963) .

16. W. von E. Doering and W.R. Roth, Tetrahedron, 18, 6l (1962).

17. R.S. Berry in Doering and Roth, ibid., 18, 67 (1962) . 18„ R.B, Woodward and T.J, Katz, ibid., "5, 70 (.1959) .

19„ W„ von E. Doering and W.R. Roth, ibid., lg, 715 (1963) . 20. A.C. Cope and L. Field, J. Am.. Chem. Soc, 71, 1589 (1949).

- 16 - RECENT ARYNE CHEMISTRY

Reported by R. Lambert July 24, 1963

I. Introduction. --The entire area of aryne chemistry has been reviewed rather extensively through the latter part of 196c (l,2,3A>5)> therefore duplication will be avoided except for x.he summary in this section of the underlying principles. Since the initial studies by Wittig, Roberts and Huisgen*, aryne chemistry has pro- gressed to a point where it is a useful synthetic tool as well as an area of con- tinuing mechanistic interest. The following criteria may be considered qualitatively descriptive of aryne intermediates and reactions involving them; although any serious attempt to generalize the individual results is risky since the properties of a transient intermediate are apt to depend on the method and media in which it is generated.

(i) The intermediate is a short lived, extremely reactive electrophile, pos- sessing the normal benzene ring structure with a special triple bond between two equivalent, hydrogenless sp2 hydridized carbon atoms. This third bond results from the sideways overlap of these orbitals orthogonal to the jt system of the ring and containing antiparallel electrons. The bond is necessarily a weak one since these

orbitals are not mutually parallel, but its introduction into the benzene ring does not disturb the it clouds above and below it or its aromatic character.

(ii) The appearance of cine-substituted products whose isomer ratios are generally independent of the nature of the leaving group( s) and dependent upon the inductive and conjugative effect of ring substituents, and the steric requirements of the adduct and aryne.

(iii) A general failure of reactivity of benzyne toward nucleophiles to cor- relate with their pK^ values,

(iv) Formation of Diels-Alder adducts with dienes such as furan, anthracene, and tetracyclone. Bi- and triphenylene as well as biphenyl are characteristic pro- ducts when the intermediate is generated in an environment free of nucleophiles. (Methods C, D, E, and F below)

II. Methods of Generation and Characterization. --Only brief mention will be given to those methods which have been reviewed previously ( 1,2, J>} h} 5) (parts A-D below) or which have not received thorough investigation due to an inherently inef- ficient generation procedure (parts G-K below) .

A) Reaction of KNH2/NH.3 with haloarenes containing an ortho hydrogen atom.

B) Reaction of alkyl and aryl lithium compounds in ether, and R2r]LJ;//R2I\!H with halo- and dihaloarenes.

C) Reaction of Li-Hg In ether with dihaloarenes.

D) Reaction of Mg in ether with dihaloarenes.

E) Diazotization of anthranilic acid in a so'r+ion of r one e/ntr-n ted ethanolic HC1 with isomyl nitrite at 0°C. This method of Stiles (11,9,10,6) affords a 50- 60/0 yield of benzenediazonium 2-carboxylate rj^r'N2®(l) . When I is warmed to k-0-6on

^C0§ in an inert solvent it slowly decomposes with loss of N2 and C02 to yield benzyne.

The yield of N2 is usually quantitative (90-100$) while C02 appears to be the limiting

factor (yields range from 32-8c$>) . Friedman (12) has reduced the application of I

to synthetic practice by diazotizing anthranilic acid with amyl nitrite in solvents

such as CH2C12, CH3CN, THF, or acetone.

1) General Properties (9^10,30). — The inner salt I does not decompose at room temperature but detonates when shocked. It is soluble in water, insoluble in non- polar solvents and the appearance of a band at 2283 cm"1 in its IR spectrum suggests zwitterionic structure I over alternative cyclic species.

2) Typical Reactions (9,13,30).

1 + hs° 36 hrs.^ (OTT (8856)

^ 0h OH

11+ [j j] reflux, l^YXi) i<-*t\- HC1, v^A

I+tQroT^-^U

Triptycene

*See R. Heany's review (5) for leading references,

- 17 - Equations (a) and (b) below illustrate how neutral generation can extend the realm of useful aryne chemistry. The addition of t-BuOH in 78$ yield stands in sharp contrast with previous attempts at alkoxide additions which were mostly in the range of 0-<$ yields. Wo explanation by the authors is offered for the trace of salicylic acid detected in the products of this reaction. 0 0

(a) I + 0COH

40-60°. dry 0H

0CO0 (25*)

+ COp + N2 (65f)

+ 0-0 4- polymers

COO0

tBuOH (dry) '

OpN

OtBu

GCffo

Oplf

\

C00H (IT/o) trace Since Stiles chose benzene as solvent in this work (equations (a) and (d)) a

subsequent study on the decomposition of I in pure benzene was conducted.

(n\ T rfH ^° > ^) (16^)

^c; x -t- jyn ^ hrg> CQ^ + polymers + hydrocarbon fraction

Most of the reaction mixture was unreacted benzene but neither the amount, iso- lation, nor characterization of the polymeric material is described. The appearance of ester bands in the IR spectrum of the polymeric fraction however, suggest that this might arise from an attack by benzyne on I. Further analysis of the reaction mixture showed the presence of 2, 8, and 6$ yields of benzobicyclo (2.2.2) -octatriene, benzocycloBctatetraene, and biphenyl respectively. No bi- or triphenylene was reported; but Stiles postulates the formation of benzocycloo'ctatetraene from a ring expansion of dihydrobiphenylene.

The best evidence for the existence of a symmetrical aryne intermediate in these reactions is illustrated in equation (d) below. The ratio of the para to meta

0

R

(d)

Rr

'XX

coofc

N

©

0COOH

R!

R:

R

0 / s +

isomers from che two reactions was the same, within experimental error (1.0 - 0.2;. Deviations from a 50/50 p/m isomer ratio for the 5-nitro, and fluorocompounds (equa- tion d, R2 = F or N02) could not be explained solely on the basis of substituent inductive effects (l4,32,2) since the p/m ratio for the fluoro compound was 3-5 and that of the nitro compound was only 3.8. As Roberts points out (l4), conjugative effects between the substituent and the entering group for ^-substituted benzynes become important when a one step addition mechanism via a cyclic transition state is possible (as opposed to one that yields another charged intermediate). Assuming that this mechanism is operative here, Stiles argues that the conju- gative effect of the nitro group acts in opposition to its field effect with regard to polarization of the triple bond in benzyne, whereas for the fluoro group these effects are complementary.

In a further attempt to characterize the intermediate produced from the starting material Berry, Spokes and Stiles (15,13) prepared a solid N02 N02©

film of I, flash photolyzed it and recorded the U. V. spectrum of a transient species leading to the product biphenylene. Since biphenylene ' s ground and excited state spec- tra were distinctly different from that of the transient species, the latter was assigned to benzyne. Another compound (l6,17) ^^1 believed to yield

benzyne under thermal decomposition, was used |T (ll) in a comparative run. It was estimated that \/\ H j

under conditions leading to 80-90$ reaction of I only S 10-20$ of II decomposed. Precise e values are unavailable since length x cone parameters are unknown. No

- 18 - meaningful comparisons can be made therefore, of curves of absorbance vs. wavelength for the two compounds. After repeated photolysis of I, the ether soluble portion of the yellow percipitate which resulted was chroma tographed on alumina. Ei- and triphe- nylene were isolated and their identity demonstrated by matching U. V. spectral data and melting points with literature values. The yields of biphenylene were higher than triphenylene. Biphenylene shows intense absorption bands in_the 2^00 A region and weaker long wave bands, the strongest of which occur at 35^3 A. At 200-1000 micro-sec. after photolysis the absorption in the 3500 A region is broader and shows a transient continuous band which blots out the characteristic 21+00 A biphenylene bands over the rest of the spectrum^ One millisecond after photolysis the strongest long wavelength band comes at 3587 A. The authors conclude from this that biphenylene is produced with excess vibrational energy. That the transient continuous absorption was not due to active biphenylene alone, was shown by irradiating biphenylene. The characteristic 2^00 A absorption was not blotted out. On the basis of similar short life time maxima from the photolysis of I and II Berry (15) et. al, chose to assign the transient spectrum to benzyne.

F) Wittig and Hoffmann (l8,19,lo,20) employing the diazonium salt of o- aminobenzenesulfinic acid as a benzyne source appear to have a more reactive, less stable system than Stiles, and have made a more detailed attempt to characterize the intermediate. When III is warmed in an inert solvent to 10° it decomposes with

aN02

SOpH

WPt > NaOH

NallO;

HoSO,

N

II

SOpIIa

0^ 0

III

loss of N2 and S02 to benzyne.

(i) General Properties. --Ill is insoluble in water, soluble in alcohol, ether and an ether-benzene mixture. Its solubility characteristics and the lack of a diazonium band at 2300 cm-1 suggest the cyclic structure in preference to the zwit- ter ionic one.

Ill

Zn

Ac OH

In addition the following conversions have been carried out.

ML Fb(.0Ac)4 TTT

NMH * 1J1

/ S02

-GO;

U=IJ-

in 5hi> yield from the same reaction run in a N2 atmosphere, complementary experiment in which the epoxide is converted

III + furan reacts to form 1-naphthol rather than the expected l,4-dihydro-l,4- epoxy- naphthalene. That this was an air induced isomerization was shown by the isolation of (f{ The

to naphthalene by air reflux was also reported. A solution of III in THF 4- anthra- cene gives 25-30$ yields of triptycene. When III was treated with tetramethyl- ethylene, a hydrocarbon Ci2Hls resulted which was then assigned structure IVa on the basis of its IR spectrum. In the light of fimmons (19) work, the sterically hindered 1,2-cycloaddition reaction would hardly be expected to occur instead of the observed a Hylic alkylation (29) • The decomposition of III in the presence of alpha -pyrone gives a 36$ yield of naphthalene in one step (equation e below). Ill reacts with phenylazide to give a kn° yield of 1-phenylbenztriazol (equation f below). These adducts were unobtainable using base-generated benzyne. That no products corresponding to attack at the ester carbonyl of alpha pyrone were observed in the diene addition of equation (e) suggests strongly that an uncharged symmetrical benzyne intermediate is present. Decomposition of III in water gave a

■::

(e) m+\^

(f) III + 0-N=N=N

2)

•S02

-CO;

-H2,-S02

II

N

(hli)

©UH: l

?H3 CH2

CH<

IVa

31/° yield of phenol, and in ethanol it gave a Gsfp yield of phenetole. Since

and 8c$ yields of W2 and S02 were observed in the phenetole synthesis an almost quantitative addition of ethanol to benzyne was achieved. In order to show that these reactions proceed via a benzyne intermediate and not by the alternate substitution- elimination mechanism shown below, Wittig performed the following experiment. 0E,

N^ Et

+EtOH,-N;

so^e (LiC1 added;

in

-r

Ilia

IVb

SOP0

When an excess of independently prepared IVb was introduced into the ethanol III

reaction mixture, IVb was recovered in over 80$ yield while a small fraction went

on to react with benzyne to yield V. The overall yield of phenetole was not specified.

When III was decomposed in weakly nucleophilic solvents such as ether, acetone, or chloroform, the benzyne formed attacked the undecomposed precursor ( III) . Chro- matographic separation of the mixture gave sulfur and nitrogen containing oils with characteristic IR sulfone bands, as well as biphenylene in yields ranging from 2-$$ (but no triphenylene) . Biphenylene (38-52$) and triphenylene (2$) were obtained when III was detonated in a vacuum. For the chemical characterization of the inter- mediate from the gas phase decomposition of III, a Diels-Alder addition with furan vapor was attempted, and a 0.6$ yield of the adduct 1-naphthol was isolated. In order to see if benzyne can appear as a diradical Wittig decomposed III in nitric oxide. The reaction products contained no /^^™'* (v^)> but 2-nitrobiphenylene (2$)

was isolated in addition to biphenylene (52$) and triphenylene (2/°). Therefore, the existence of radical benzyne cannot be demonstrated by radical trapping.

•Since the decomposition of III avoids the use of strong bases Wittig (19) has been able to extend the benzyne nucleophilicity scale to include halide ions and al- cohols. The mechanism for benzyne formation using Methods A and B above has been shown by earlier workers (1,2,3,^,5) "to proceed according to the scheme below with the ratio of k2 to k_! depending on the nature of the leaving group, with kx the rate determining step. Wittig has demonstrated the reversibility of the second step through the isolation of aryl halide s from the reaction of lithium halides with III in THF. The protons for reaction with the intermediate anion are believed to come from adven- titious water.

+ :Be-!±-, BH + X

k_i

k_;

1)1 +^

where X = I, Br, CI

From a series of competition reactions using lithium halides -first in THF and then in ethanol, the following nucleophilicity scale resultedi _,

Ie- 65>Br -8>C1Q-1>7 This scale indicates that benzyne is not the highly reactive, unselective interme- diate previously proposed (1,2,3,4,5). In order to exclude the possibility that the reaction was proceeding through an intermediate other than benzyne the following experiment was performed. As shown in the scheme below, compounds VII and VIII have been shown to undergo the indicated substitution, however they both do so very much more slowly than III decomposes, and therefore cannot take part in the main reaction path. This, however, says nothing as to whether the hypothetical Ilia participates in the course of the main reaction.

-20-

+

Xs

III

11^

\

>

X

bU2-

+ H

■+)

Ilia

•S02

VII VIII A

( -so2)\

TT0 /

-y

N2+ +H /

G) Wawzonek (21) reports the generation of benzyne by electrolytic reduction of halobenzenes. This method, if successful, would allow studies in both aprotic and protic media. Both dihalo and monohalobenzenes were used and attempts at trap- ping intermediate benzyne with furan showed isolated yields of alpha naphthol of 1$ , but only for the o-Br substituted benzenes.

H) LeGoff (22) reports that diphenyliodonium 2 -carboxylate ( IX) may serve as an aprotic precursor of^bcnzyne.

I If

n C00J

-r C02 + 0-1 +

hv

-I-

The intermediate has been trapped with tetracyclone to give^a 68$ yield of 1,2,3,4-tetraphenyl naphthalene and with anthracene to give a 2^$ yield of triptycene. At 325° IX gives biphenylene in low yield.

I) Kampmeier (23) reports the photolytic decomposition of o-diiodobenzene (X) to benzyne. The proposed scheme below shows that the initial liberation of I* is a highly reversible step which seriously hinders the reaction. Therefore, the total concentration of I2 was kept low by repeated extractions of the reaction mixture with NaHS03.

a:

x

In cyclohexane X gave iodobenzene as the chief product and in benzene 2-iodo-biphenyl (42$) plus triphenylene (4$) and some biphenyl were isolated. That 2-iodobiphenyl may serve as the precursor of triphenylene was shown by irradiating some in benzene. A 7$ yield of triphenylene resulted. Reactions involving the usual trapping agents were run - 2-iodobiphenyl was the chief product when tetracyclone was used, with 10$ of the Diels-Alder adduct also being found. In none of these reactions was bi- phenylene ever isolated, or detected.

J) Wittig (l6) has investigated the thermal decomposition of o-iodophenyl- mer curie iodide to benzyne but found it to be an inferior method of generation. Attack by I, Hg, and Hgl complicate the studies considerably L:nd seriously reduce the steady- state concentration of the intermediate. Phthaloyl peroxide has likewise been rejected by Wittig (17) as a suitable method.

K) Simmons (24) has investigated the possibility of using silver o-halo-ben- zoates to generate benzyne. Of the many possibilities tried only the ortho chloro compound gave any results which could be interpreted to show that the reaction was proceeding via a benzyne mechanism. The yield of o-chlorophenylbenzoate from a 200° decomposition was 50$.

III. Synthetic Applications. - -A) Aromatic Nucleophilic Cyclizations. The

principle workers in this field have been Bunnett (25,26,27,28,4) and Huisgen (3).

The intramolecular addition of a nucleophile situated on a side chain of an aryne is

really just a special case of older intermolecular alkylations on arynes where nucleo-

philes, either in competition with the generating base or the base itself, were used

as adducts. Successful ring closures using carbanions, nitranions, and sulfur anions

have been reported (25,26,27) (equations g_£) . Since approximately the same yields

from ortho and meta isomers resulted, the elimination-addition mechanism is suggested.

That direct displacement of bromine by RS~ does not occur was Illustrated by: (i)

refluxing XI or XII with KaOEt/EtOH for four hours which gave only 0.9$ product;

97$ starting material was recovered and (ii) ff^^MI-C-O' Slip tt „

UA || r 1^ IJ.R.

1 Br s CH3

(s)

IHC-0 II

ErS(or Cl)

6 NH2

- 21 -

II P

S

IJ=C-0

l&

WH.

/

NH-C-0

HH2

Br (or Cl)

H-C-0 I SH

N

percent yields-

o-Br o-Cl

90

o2

x

m-Br m-Cl

/

0

72 67

The difficulty in effecting alkoxide additions to benzyne has been noted by earlier workers (see section II) and Bunnett ascribes the singular success of this cycliza- tion reaction (equation h) to the excellent stereochemical orientation and the added stability of the anion due to the carboxamidc linkage. Equation (i) represents

U_0 __^_ „ vl ^

0'

(h)

NH<

\X0 a typical nitranion addition to benzyne

X=Er

X=C1

i2io

6cf?°

(i)

(kh1o)

In studying the addition of carbanions to benzyne Bunnett (25,26) has uncovered a serious limitation to the scope of the ring closure reaction. If a side chain anion is produced with the negative charge on an atom adjacent to the aromatic ring, and if the greater portion of this charge is not concentrated in the side chain, then the negative charge will be smeared out on the ring (mesomerism) causing deacidi- fication of the aromatic protons, and consequently no benzyne formation is achieved (equation j). ^^ e

■ f r6>'

o *"* ~ o

II « lie

9CHH2 e CH-CMH

QT m^>

@

Cl

?H-C00 Cl

©

(N.R.)

(J)

V

Cl

NH<

Cl

(N.R.)

In equation (k) below if R=H, the base abstracts protons from both methylene and nitrogen atoms forming a dianion in which the negative charge is not sufficiently localized in the side chain, and therefore no reaction occurs.

(k)

Cl

CH20

N-C=0

I

R

Sh-

MH-

R-CH3; 91$ yield

R=H : N.R.

When an additional keto group is introduced in place of the phenyl group, then both carbonyls stabilize the charge so that ring closure proceeds with a 78$ yield (equation JL) - — r -1 0

U)

\

Cl

M-C

0

II

■ CH2-CCH3

j

N-

1

•i?

CH^^C CH3

&[

6

y>

CCH-

f OH H

- 22

Hauser and Morris (31) have confirmed Bunnetts1 finding indirectly.

<5.CN 0CH2C1

;grfH^(0H3)^;gr<N(cH3)

Cl CN CI

ci

^

CN i

c-ch20

(m)

N( CH3) 2

(9$)

In reaction (m) below a l6# yield was taken as evidence for the activation of alpha hydrogen by an amide carbonyl group.

ci © ^ y H

.CH3 §T* L I) if R=CH3; 16* yield

t 0

When R=H cyclization takes place to form a benzoxazole ring (equation h) in 37> yield. If the amide carbonyl is replaced by a sulfonyl linkage the expected cycli- zation product formed initially reacts further with KNH2 to reopen the ring (equa- tion ( n) ) . n - sj ;• "■ ®

:CH2 I

so2

Equation (n)

KHHP Ma

^NX I CH.

LT>S02

XIII

I'JHCH-

That attack by NH2- takes place at the sulfonyl group instead of on the ring hydrogen of the initial product is deduced from the reaction products in which no primary aromatic amine could be found, and from a comparison of XIII' s n.m.r. spectra with a known sample of the initial product synthesized by an unequivocal route. Bunnett (26) has also found that aliphatic esters, nitriles, sulfones and ketones with (D-ortho chlorophenyl groups give homocyclic ring closure products (equation (o)).

Equation ( o)

(CH2)n-CH2Z

•S-

NH;

\

Cl

Mi.

(CH2)A

•J /where z=cyano, sulfonyl, ^ — CH — J carbethoxy

Z

By varying both chain length and the nature of the activating group (z), Bunnett hoped to gain mechanistic insight as well as to determine the optimum con- ditions for synthesis. The order: cyano, sulfonyl, carbethoxy, is roughly in accordance with the activation capacity for the group in terms of % yield. Ortho- chloro compounds were shown to give the same ring closure products as meta isomers but in higher yields. Reaction times were usually on the order of 15 minutes. Ques- tions which remain unanswered from this work are: (i) Why is proton removal ortho to the side chain preferred, i.e., ff1*^^"?^ r^T^f*-

(©pa r Q-%+% rather than Q =z +BH

X +B:

seems to be the case since yields greater than 50$ have been observed (see equations (g) and (h). (ii) Is this additional reaction a two step or concerted process, i.e. is it the conjugate acid of the side chain anion which adds to benzyne in one step or does the side chain anion add to the benzyne bond to give an aryl anion which is sub- sequently protonated.

B) Diels-Alder and Analogous Additions. --Simmons (7) has investigated the reac- tion of strained bicyclic hydrocarbons with benzyne (equation t below).

F

(t) i^Y - ™F

^Br

+ Mg +

XIV

- 23 - Only the exo isomer was uncovered in the products. XIV reacts with KMn04 and

Br2/CCl4, shows C=C in IR at 6-39 M- (1564 cm-1) and n.m.r. showed it to be a 1:1 adduct C13H12. It was known that bicycloheptene double bonds give 1:1 adducts with dienophiles and that predominant formation of the exo isomer results. Sim- mons thought that the bulky benzyne molecule would give the exo adduct. Structure proof for the adduct was made through comparisons of its n.m.r. spectra with a known sample of the endo isomer. Subsequent chemical transformation, on both compounds and comparison of those spectra showed Simmon's assumption to be correct.

H

(u)

+ Mg +

Br

(XV)

H

The hydrogenation product from XIV was shown to be identical with XV. The exclu- sive exo geometry has been attributed to the ease of steric approach to the less hindered side of the bicycloheptene double bond. Thus it would appear that ben- zyne reactions are more susceptible to steric controls than previously thought' (1,2,3,4,5) , and since reactions (t) and (u) both give exo adducts no unusual elec- tronic factors need be invoked. In the reaction of benzyne with cyclohexadiene (7) two adducts were obtained (combined k&jo yield). The Diels-Alder adduct was

+ Mg +

Br

+ C12H12

65$ 35$ characterized by its IR and n.m.r. spectra and by Analogy with

Wittig's results (17). The C12H12 product could not be obtained pure by VPC but the n.m.r. spectra of the crude product showed it to be a 1:1 adduct with bands expec- ted of a benzycyclobutene structure (a line of area two at t = 6.51). In the reaction of 1-octene a 17$ yield of trans-l-phenyl-2-octene resulted. This hydrogen abstrac- tion reaction is pictured as going through a cyclic quasi-six membered ring.

+ CH2=CHC6H13 ^jC

11 -,Lr

0CH2C=C-C5Hn

H

Although the molecular mechanism has not been established for these reactions Simmons feels that this is a likely guess since skeletal rearrangements in other cycloaddition reactions are absent. He argues that if the possibility of a zWitterion intermediate did exist then reaction (m) would undoubtedly have resulted in a rearranged product. Contingent to this argument however, is the assumption that such a rearrangement is faster than the four membered ring closure. Wo evidence is given to support this assumption.

Stiles (10) has achieved the following additions to his intermediate.

a

coo

e

+

XVII on

treatment with Pd/C absorbs on equivalent of Hp to. form XVTIa-

XVII

XVIIa

- 24 - ..lie proposed mechanism to explain the products is given below. XVII formation is shown as arising by a 1,2 cycloaddition reaction, and phenanthrene is pictured as arising by a 1,4 cycloaddition through an anion intermediate.

T jQ Mmcrizes [O T__jOj 0J 0 0

two >teps

CH-CH

Attempts to extend the 1,2 cycloaddition reaction led to this result,

H

1. 2.

3-

5- /•

o.

7- 8.

9-

10. 11.

12.

13- 14.

15. 16.

IT-

18.

19- 20. 21. 22. 23- 24.

25- 26. 27. 28.

29. 30.

31. 32. 33.

3^. 35-

f

c HI .

c I

OEt

CHCH

BIBLIOGRAPHY

G. Wittig, Angew. Chem. , 6g, 245 (1957).

J. D. Roberts, Chem. Soc. (London) Spec. Fubl. No. 12, 115 (1958).

R. Huisgen, Chapt. 2 in " Organometalic Chemistry", H. H. Zeiss, ed. , Reinhold

Publishing Corp. , New York, N. Y.

J. F. Bunnett, J. Chem. Ed., j>8, 278 (1961).

H. Heaney, Chem. Revs., 62, 8l (1962) .

M. Stiles, R. G. Miller and U. Burckhardt, J. Am. Chem. Soc, 85, 1792 (1963).

H. E. Simmons, J. Am. Chem. Soc, 8^, lo57 (1961).

E. Wolthuis, J. Org. Chem., 26, 2215 (1961).

M. Stiles and R. Miller, J. Am. Chem. Soc, 82, 3802 ( i960) .

M. Stiles, U. Burckhardt and A. Haag, J. Org. Chem., 27, 1+715 (1962).

R. G. Miller and M. Stiles, J. Am. Chem. Soc, 85, 179$ (19'$3)-

L. Friedman and F. M. Logullo, J. Am. Chem. Soc, 85, 15^9 (1963).

R. S. Berry, G. N. Spokes and M. Stiles, J. Am. Chem. Soc, 82, 5240 (i960).

J. D. Roberts, C. W. Vaughan, L. A. Carlsmith and D. A. Semenow, J. Am. Chem. Soc. ,

78, 611 (1956).

R. S. Berry, G. N. Spokes and M. Stiles, J. Am. Chem. Soc, 84, 3570 (1962).

G. Wittig and H. F. Ebel, Ann., 65O, 20 (1961).

G. Wittig and R. W. Hoffmann, Angew. Chem., 72, 5o4 (i960).

G. Wittig and R. W. Hoffmann, Chem. Ber. , £5, 2718 (1962).

G. Wittig and R. W. Hoffmann, Chem. Ber., £5, 2729 (1962).

G. Wittig and R. W. Hoffmann, Angew. Chem., 7^, 435 (1961).

S. Vfewzonek and J. H. Wagenknecht, J. Electro. Chem. Soc, 110, k20 (1963).

E. LeGoff, J. Am. Chem. Soc, 84, 3786 (1962).

J. A. Eampmeier and E. Hoffmeister, J. Am. Chem. Soc, 84, 3787 (1962) .

H. E. Simmons, J. Org. Chem., 25, 691 (i960).

J. F. Bunnett, T. Kato, R. R. Flynn and J. A. Skorcz, J. Org. Chem., 28, 1 (1963).

J. F. Bunnett and J. A. Skorcz, J. Org. Chem., 27, 3836 (1962).

J. F. Bunnett and A. J. Sisti, J. Org. Chem., 2bJ 3^39 (1961).

J. F. Bunnett and B. F. Hrutfiord, J. Am. Chem. Soc, 8£, 1691 (1961).

E. M. Arnett, J. Org. Chem., 2£, k^ko (I96l).

R. Huisgen and P. Knorr, Ifeturwissenschaften, 48, 716 (1961).

C. R. Hauser and G. F. Morris, J. Org. Chem., 2o, 4740 (1961).

G. E. Hall, E. M. Libby and E. L. James, J. Org. Chem., 28, 311 (1963).

F. N. Jones, M. F. Zinn and C. R. Hauser, J. Org. Chem., 2§, 663 (1963).

G. Koebrich, Angew. Chem., jk, 428 (1962).

G. Wittig, Angew. Chem. (internat. ^d.) , 415 (1962).

- 25 - THE PROBLEM OF SOME THERMOCHROMIC ETHYLENES

Reported by R. Puckett July 29, 1963

Introduction. - A number of organic compounds have been observed to change color upon heating both in the solid state and in solution. This phenomenon is known as thermochromism. Photochromism refers to the same type of behavior upon irradiation with light of a suitable wavelength. A number of these compounds have also been observed to undergo a color change when subjected to pressure. The latter has been called piezochromism.

A recent paper (1) on the crystal and molecular structures of the a and £ poly- morphic forms of 9,9'-bixanthenylidene (II) suggests that much of the evidence for the mechanism of thermochromism and related effects in the bixanthenylidenee and bian- thronylidenes needs to be reevaluated. This seminar will deal with studies performed in this area on both these classes of compounds and will attempt to summarize the evidence available at the present time. A recent review (2) on the general topic of thermochromism treats these compounds briefly. Also, Korttfm has summarized his own work, mainly on the bianthronylidenes, more thoroughly (3). This seminar will not deal with the thermochromism of the spiropyrans.

Experimental Facts . Existence of Color Change . -- As early as 1909 (4) it

was noted that warm solutions of 9,9' -bianthronyl- idene (I) were green while at room temperature they were lemon-yellow. In 1928 9,9'-bixan- thenylidene was observed to behave similarly, changing from light yellow to blue -green (5). Crystals of bianthronylidene are pale yellow at room temperature (6). When heated to about 265°Co, they are reported to become light green and then dark green. The color change is not j jj reversible for the crystals once the dark green

state has been attained; however, solutions of bianthronylidene exhibit thermochromic behavior, the color change occurring at a much lower temperature (70 -180 C.) and being completely reversible (7).

The phenomenon of photochromism in these two classes of compounds seems to be closely related to that of thermochromism. Hirshberg first noted that at tempera- tures below -70 Co solutions of bianthronylidene changed color markedly when submit- ted to ultraviolet irradiation of frequency 27,^00 cm. x (8) . The color was observed to fade upon warming to room temperature. Furthermore the phenomenon did not mani- fest itself if the irradiating light was less energetic than approximately 22,000 cm."1.

Ultraviolet and Visible Spectra. - In 1950, Theilacker et al. (7) studied the visible and ultraviolet absorption spectra of bianthronylidene at various tempera- tures (Fig. 1) and observed that there was a band in the visible region at 15,000 - l6, 000 cm.—1 which could be correlated with the color change and whose molar absorp- tivity coefficient showed a strong temperature dependence, decreasing with decreasing temperature and obeying the Beer-Lambert law. The remainder of the spectrum showed no strong temperature dependence. Thus this band was assigned as being characteris- tic of the thermochromic property of bianthronylidene. On the basis of the rever- sibility of the color change and the fact that a plot of log €max for the 15,00 0 cm»T band versus l/T gave a straight line, it was proposed that the thermochromism of this compound was caused by a thermal equilibrium between two different forms, A and B, of the molecule, form A being "normal" and form B being colored.

Grubb and Kistiakowsky obtained similar spectral results and were able to show a slight solvent dependence by employing both acetophenone and decalin (9) - The spec- tral measurements were also duplicated by Hirshberg and Fischer, but they did not specify their solvent (10). The spectrum of the closely related compound 9,9' -bi- xanthenylidene (II) is substantially the same as that for 9,9 ' -bianthronylidene (I) except that molar absorptivity coefficients are somewhat smaller (7) » Hirshberg showed that the temperature dependent band at 15,000-16,000 cm. -1 of the colored thermochromic modification is practically identical with a corresponding band pro- duced by the photochromic form (10) .

26-

O H

Korti!im and Bayer in a recent commu- nication have stated that 1,3>6!,88- tetramethyl-9,9 ' -dianthronylidene ( III) exists in three forms in solution (11) „ Their evidence is mainly spectral, They report the photochromic form (form B) with the usual maximum at 15,000 cm., 1 and another form (form C) with smaller maxima at 21,000 and 22,500 cm,"1. The uncolored form is denoted as form A. The remaining ultraviolet spectrum of form C is completely different from that of form Ao Solvolysis of the sulfuric acid adduct of A in 10$ aqueous alcohol at -90 C„ is reported to give exclusively form Bo Both B and C revert to A at higher temperatures.

10000

20000

30000 40000

cm

Fig. 1. Visible and ultraviolet spectrum of 9*9' -bianthronylidene (I).

Thermodynamics and Kinetics of the Color Change . - Assuming that the colored and uncolored forms were in equilibrium, Theilacker has determined equilibrium con- stants from an evaluation of the molar absorptivity coefficients of the 15,000 - 16,000 cm. l band at various temperatures (7) ° A plot of the equilibrium constant versus the reciprocal of the absolute temperature enabled him to obtain the heat of transformation between forms A and B, For 9>9 ! -bianthronylidene (I) , a value of 3»4 + 0.2 kcal./mole was obtained „ The corresponding value for 9^9,-t)ixa^'fchenylidene (II) was 4.9+0.1 kcal./mole. Grubb and Kistiakowsky obtained similar results and were further able to show that the heat of transformation had a slight solvent dependence (9) o In a later study, 2,4,5 ' ,7' -tetramethyl-9,9 ' -dianthronylidene (IV) was shown to have a AH value of 6.7+0.3 kcal./mole (12) .

In 1953 Hirshberg and Fischer found that the frequency factors for the thermal reversion reaction were not smaller than those en- countered in common first order reactions (10). This indicates that the reversion process is not due to a transition involving a change in electron spin multiplicity . Wasserman and Davis (13) report that the bleaching of the thermochromic, photochromic and piezochromic forms of 9 >9 '-bianthronylidene is a first order pro- cess in which the rate constants for all three forms are identi- cal within "10$ uncertainties ". Their criterion of error is un- specified o They interpret this as evidence that all three forms are identical. However, as Mills and Nyburg have pointed out (14) , this could merely mean that all three return to the colorless form by the same rate determining step. It has been noted that the lifetime of the photochromic state of 2,4,5 ' ,7' -tetramethyl-9,91 -dianthronylidene (IV) in solution is at least 104 larger than that of known triplet states of organic materials (usually several seconds) (15) . This however does not discount the existence of the colored form of this compound in a triplet state,

Dipole Moment Studies. - In 1950, Bergmann and Fischer reported that 9,9'-bian-

- 27 -

thronylidene (l) in solution exhibited a dipole moment of 1 Debye (l6)o In contrast Korti!im and Buck have reported that both 9,9'- bianthronylidene and 9,9 ' -bixanthenylidene have no dipole moment in solution (17) . Through a study of the molar polar izability of solu- tions, Hirshberg has demonstrated that the colored form of xanthyli- deneanthrone (V) produced by irradiation is more polar than the non- colored modification (10) „

Magnetic Measurements . - As we shall see, thermochromism and photochromism have been interpreted in terms of triplet -triplet transitions o Therefore, various magnetic measurements have been made with somewhat conflicting results . Nilson and Fraenkel found that solutions of bianthronylidene in decalin and dimethylphthalate at various concentrations showed paramagnetic absorption in the temperature range l40 -200 C (l8) . In more concentrated samples the intensity of absorption was found to decrease with decreasing temperature to a point where it could no longer be detected. In more dilute solutions it was only possible to observe that no absorption could be detected below a certain temperature (generally 1000-1500C.)„ Kortilim and Theilacker observed that the photochromic form of bianthronylidene was para- magnetic by magnetic susceptibility measurements (15) . These workers maintain that the thermochromic form is diamagnetic however, mainly on the failure to observe paramagnet- ism of the colored form produced at high temperature. It should be pointed out that they used a magnetic balance method which is considerably less sensitive than that of paramagnetic resonance absorption.

Wasserman has shown that pyridine solutions of bianthronylidene at 25°C<J exhibit electron spin resonance which increases with temperature (19) » The signal is reported to consist of five main components having relative intensities of lik°.6°.k:l, indicating the interaction of an_unpaired electron with four equivalent protons. The absorption maximum at 15,950 cm.-1 and shoulder at 17,390 cm, _1 were described as being sensitive to both oxygen and excess bianthronylidene. Ether solutions of the green form of bianthronylidene (obtained by sublimation at reduced pressure) showed no absorption at -77 Co, but did exhibit a resonance at -k6°C„ which then vanished as the color of the solution faded (19). Measurements on methyl 9,9* -bianthronylidene -3-carboxy late by magnetic balance measurements at 210 C. showed no evidence of a biradical (20). No ESR signal was observed in a series of substituted bianthronylidenes and bixantnenylidenes, but the authors admit that this may have been due to unusually broad resonance lines (21).

o It has also been found that bianthronylidene shows a paramagnetic absorption at 265 C. in the solid state which is retained upon cooling (18) „ Matsunaga has reported that bianthronylidene shows a drastic decrease in diamagnetic susceptibility as the color change occurs in the solid state (22). The color change is also accompanied by complete vanishing of sharp x-ray powder diffraction patterns. Mills and Nyburg re- port no paramagnetic absorption in their solid a-modification of bixanthenylidene (Ik). Stereochemical Studies. - In an effort to obtain some knowledge of the actual con- formations of forms A and B, Korti!fm and co-workers prepared a series of substituted bianthronylidenes and determined the presence or absence of thermochromic properties. The results are shown in Table 1.

Table 1

Derivative of I Thermochromic Properties Reference

2,^,5 ST '-tetramethyl- Thermochromic band at 155°C 12

1,3, 6 ',8' -tetramethyl- No Thermochromism 12

1^*5',8I -tetramethyl- No Thermochromism 12

2,3,6!,7!-dibenzo- Weakly Thermochromic 23

3,J+,5',6!-dibenzo- Thermochromic 23

!^2,7' ,8- -dibenzo- No Thermochromism 23

These results indicate that substitution in the 1,1' or 8,8s positions prevents thermo- chromic behavior. Although 1,3,6' ,8' -tetramethyl-9,9 '-bianthronylidene is non-thermo- chromic, Kortilim reports that in the preparation of this compound a deep green solution is formed which immediately reverts to the yellow color of the "normal" compound (12).

- 28 -

Some substituted bianthronylidene s which do not exhibit thermochrornism have been shown to be photochromic„ Ehotochromism is exhibited by 2,4,5 ' .,7' -tetramethyl-9,9'- bianthronylidene (IV) (15) as well as 1,3,6' ,8' -tetramethyl-9, 9' -bianthronylidene (III) (11). In contrast, 9,9' -bixanthenylidene (II), which is thermo chromic, does not exhibit photochromism (10) .

An examination of models reveals that there is an exceptionally high amount of steric hindrance present at positions 1,1' and 8,8' even in unsubstituted bianthronyli- dene. Indeed a completely coplanar system can be completely eliminated (Fig. 2).

Theilacker et al. , prepared the series of compounds VI -VIII in an effort to study their optical activity (24). These workers were able to resolve the quinine salts of VI j but upon acidification with HC1, racemization took place immediately., They were unsuccessful in attempts to resolve either VII or VIII. Recently it has been reported indirectly that the optically active isomers of 9,9' -bithioxanthenylidene-4-carboxylic acid (IX) racemize spontaneously in dioxane (25) „ The amount of crowding in the 1,1' and 8,8' positions in this compound is comparable to that in 9*9 ' -bianthronylidene (i) „

0 0

C02H ^^^\/^^ COsH

Fig. 2, Drawing of ring system showing overlap of hydrogen atoms in con- flicting position.

VII

COaH

COpH

VIII

IX

However, this sulfur compound is not thermochromic| and at the time of this writing no thermochromic bianthronylidene or bixanthenylide has been resolved.

Structure of 9 »9 ! -Bianthronylidene . - In 195^ 9 Harnik and Schmidt (26) , as part of a study of the molecular and crystal structures of a series of "overcrowded" aromatic compounds, published the crystal structure of 9^9' -bianthronylidene (I). By the appli- cation of two-dimensional molecular Fourier transform methods, they were able to eluci- date the following structure. The "overcrowded" carbon atoms, 1,1' and 8,8', are found to be 2.90 A apart, this being made possible by a 40 rotation of the benzene rings out of the plane of the central ethylenic system. The exocyclic bonds to the carbonyl carbon are deflected out of the planes of their respective benzene rings by 8 , thus

giving a boat-like shape to the center rings. Furthermore r^C~C bond angle is com-

00 pressed from its normal value of 120 to 113 . All bond lengths are of approximately

normal length. The conformation, because of the "wing-like" structure of each of the two molecular halves, has been called the "folded conformation" of bianthronylidene. Crystal and Molecular Structure of 9 ?9' -Bixanthenylidene. - Recently some very significant x-ray work (l) was done by Mills and Nyburg on 9>9 '-bixanthenylidene (II). These workers were able to obtain good crystals of both the a(deep blue -green) form and the 3( yellow) form. The a crystals were obtained by subliming the 6 form in a 250 -130 C. gradient at about 10 3 mm. Hg. Complete structural analyses were per- formed on both polymorphic forms, employing the Fourier molecular transform method1. This study showed that both forms possessed the folded conformation previously assigned only to the yellow "normal" state. Bond lengths and distances are shown in Figures 3 and 4. Estimated errors are + 0.02 A in bond lengths and + 1.5° in bond angles.

1*39

l-38>

1- 40

1*37

1-39

Fig, 3o a-form of 9,9'-bixanthenylidene

Fig. k. p-form of 9,9 bixanthenylidene ,

As can be seen, there are no striking structural differences between the a and |3 polymorphs o In fact the only substantial differences between the two forms in the solid state are in the modes of molecular packing. These are shown, in Figures 5 and 6< In both these figures the molecules drawn in full lines are those which have their mid- points lying in the center layer of the unit cell. The molecules in broken lines have their midpoints lying in layers half a cell above or below. In both cases the outer rings of the molecules stack on top of each other in a columnar array. The hetero- cyclic rings also stack in this manner in the a-form with oxygen atoms alternating from left to rightj but in the (3-form the heterocyclic rings do not stack in columns. Thus it has been shown that in 9*9 '-bixanthenylidene both the "normal" and colored modifications have the same molecular conformation in the solid state, and thermo- chromism cannot be due to an equilibrium between two conformational forms of the molecule „

Fig, 5° a -bixanthenylidene in x-projection.

Fig, 6, P -bixanthenylidene in z -project! on.

- 30 -

Theories - As the facts about these molecules have developed, a number of expla- nations of their the rmo chromic behavior have appeared. In retrospect some may now be dismissed out of hand,, However, at present there is still no satisfactory accounting for the facts .

Betaine Theory . - Scho'nberg has proposed that form B has a betaine-like structure such as X ( 6) , This theory has been discounted due to the inde- pendence of the thermochromic absorption spectrum on the acidity of the solvent and the fact that the rate of disappearance of the green modification at -50 C is greater in ethanol than in isooctane ~oy a factor of 2 ( 27) .

Biradical Theories, - Several workers (9) , (28) have proposed that the colored modification produced by the photochromic tech- nique is a biradical in which the two halves of the bianthronyli - dene molecule are twisted with respect to each other, Matlow has

performed molecular orbital calculations on both bianthronylidene and bixanthenylidene to determine the relative magnitude of energy differences between the ground state and both coplanar and perpendicular triplet states (29) . He has assumed the ground state is a coplanar singlet and has used 20 kcal./mole as an approximation to the potential barrier to twisting in the ground state „ This figure was obtained by analogy with the racemization of optically active 2,2* -dibromo-^,^' -dicarboxydiphenyl. On the assump- tion that the steric hindrance of the singlet and triplet states is the same, the following results (Table 2) are obtained. Thus, these calculations indicate that the

X

		Table	2
Compound		AE	(kcalo/mole)
	Coplanar		Perpendicular	Obs, Thermal Exc,
bianthronylidene bixanthenylidene	15 0 996 26,1		3 0 3335 10.8	3^-3*5 h.9

perpendicular biradical is of considerably lower energy than the coplanar one, How- ever it is also obvious that the calculations are valid only for order of magnitude comparisons.

Woodward and Wasserman have suggested XI as one resonance form of a biradical for the green form of bianthronylidene (23.)= A twisted biradical should be stabilized by substituents in the 1,1' and 8,8' positions. How- ever Woodward noted that the opposite was true. This together with the observed ESR of this compound led to the proposal of the coplanar biradical. There has been some controversy as to whether the photochromic and thermochromic modifications are identical (30) , (31), (32), (33), The majority of evidence seems to indicate that they are.

Obviously if any of the various biradical theories is to be given any sound basis, then quantitative magnetic measure- XI

ments must be obtained which establish the presence of two

unpaired electrons, As was seen in part 1, some workers have reported that the colored forms of these compounds are paramagnetic, indicating the presence of a radi- cal species; but experimental difficulties seem to have prevented quantitative -measurements. At this point, the proposal of biradicals is unsupported by direct evidence.

Conformational Equilibrium Theory , - An examination of the stereochemical data suggests that the phenomenon of thermochromism of the bianthronylidenes is restricted by the amount of steric hindrance in the 1,1' and 8,8' positions, This would seem to indicate that planarity of the molecule is somehow connected with its thermochromic properties ,

On this basis Korttfm has identified molecular form A with that conformation which was found to be present in the yellow crystalline form of 9 >9? -bianthronylidene (I) by Harnik and Schmidt (26), the so-called "folded conformation" (Fig. 7). For form B he has proposed a model in which the two halves of the molecule are twisted with

31

Fig, 8, Twisted conformation proposed for form Bo

Fig, 7° Folded conformation,

respect to each other, thus straining the ethylenic bond (Fig. 8) „ As is readily seen from an examination of the models, in order for a dianthronylidene molecule to pass from form A to form B, it is necessary for either carbon atoms 1 and l1 or 8 and 8' to pass each other, This would be extremely unlikely, as the hydrogen nucleus of H-l lies only 0,3 A in projection from C-l1 (14) , [Studies on optical activity however seem to indicate indirectly that the conflicting carbon atoms can pass each other in solution (24) ,(25) » With regard to this topic, Mills and Nyburg (l4) have made the interesting if somewhat unprecedented suggestion that racemization may occur through the hydrogen atoms "changing their allegiance ". ]

In any event a thermal equilibrium between conformational forms cannot be valid for the thermochromism of bixanthenylidene in the solid state, and evidence seems to point against it in solution,

Spectroscopic Models,

H

S

100

<D H

o

S

o

50-

0

s

o

-40 9 000

s"

?0

= s

00

LO^OOp a

b^ =

T1

m m mil

Two models have been pro- posed to account for the thermo- chromism and photochromism of the bianthronylidenes and bixanthenylidenes , The first (Fig, 9) was proposed by Kortu'm (28) to explain the photochromic behavior of 2,4,5 ' ,1 ' -tetra- methyl-9,9 % -bianthronylidene (IV) 0 The second model (Fig„10) ,, rtvas originally proposed by Grub'o T (IIDand Kistiakowsky (8) to account for both the thermochromism and photochromism of these types of compounds. In both models T (II) photochromic bands are attributed to transitions from low lying (I) triplet states to upper excited triplet states. Kortu'm has in- Spectroscopic dicated that thermochromism Model 2 could be attributed to the pre- sence of a set of singlet states of his proposed twisted conformation (3^). In model 2, approximately >5 kcal, above the ground state I there is a thermally populated triplet state or set of triplet states II of the molecule. From state II there is a transition of 15,000-16,000 cm _1 to an upper triplet state or set of triplet states III, In this model both photo- chromic and thermochromic bands are represented by triplet -triplet transitions between states II and III, For a thermochromic compound such as bianthronylidene, the lowest lying triplet state can be reached by thermal excitation. If the compound is non- thermochromic, then the lowest lying triplet state is too high to be attained thermal] y. The temperature dependence of the thermochromic bands is explained as follows, As

Fig, 9

Spectroscopic Model 1

Fig, 10.

the temperature increases, state II b

D r» ""vmi<*\ t

•-< ,j.ii\-j

easmgly popula oed 9 and thus the malar

- 32 ~ absorptivity coefficient for the 15,000-16,000 cm,"1 transition increases . As the temperature is decreased, state II becomes depopulated, accounting for the fading and final disappearance of color,,

Mills and Nyburg support the hypothesis that the transition between states II and III is dependent on environment (ik) . There has not been a sufficient amount of work performed on the effect of solvents in solution thermochromism or intermolecular environment in solid state thermochromism for any valid conclusions to be drawn. Work in this area could possibly shed much light on the mechanism of thermochromism in the s e c omp ound s ,

It should be pointed out that neither the presence of a biradical nor a triplet state has been conclusively shown. The theory of thermochromism and related effects in the bianthronylidenes and bixanthenylidenes is at present very much open for concrete suggestions and valid working models,

BIBLIOGRAPHY

1, J.F.D. Mills and S.C. Nyburg, J, Chem, Soc, 308 (1963).

2, J,H. Day, Chem, Rev., 63, 65 (1963)0

3, Go Kortflm, Angew, Chem,, ?0, ik (1958) . k. H, Meyer, Chem, Ber,, k2, JK3 (1909)=

5, A, Scho'nberg and 0= Scheutz, Chem, Ber,, 6l, 4?8 (1928) .

6, A.- Scho'nberg, A, Ismail and W. Asker, J, Chem, Soc, kk2 (19^6).

1. Wo Theilacker, G, Korti&i and G, Friedheim, Chem, Ber., 83, 508 (1950) .

8, Y. Hirshberg, Compt, rend,, 231, 903 (1950).

9, W.T. Grubb and G.B. Kistiakowsky, J, Am, Chem, Soc, 72, k±9 (1950),

10, Y, Hirshberg and E, Fischer, J, Chem, Soc, 629 (1953).

11, G„ Kortflm and G, Bayer, Angew, Chem, (Int, Ed,), 2, kk (1963) ,

12, G, Korttfm, W, Theilacker, H, Zeininger and H. Elliehausen, Chem, Ber,, 86, 29k

(l953)o

13, E, Wasserman and R,E, Davis, J, Chem, Fhys , , 30, 1367 (1959). Ik. J,F,D, Mills and S,C, Nyburg, J, Chem, Soc, 927 (1963) ,

15, Go KortiUm, W. Theilacker and G„ Littmann, Z. Naturforsch, 12a, lj-01 (1957) »

16, E,Do Bergmann and E, Fischer, Bull, Soc Chim, France, lOSh (1950),

17, G, Kortiftn and M. Buck, Z, Elektrochem, , 60, 53 (1956),

18, W,G, Nilson and G,K„ Fraenkel, J, Chem, Fhys,, 21, 1619 (1953),

19, E„ Wasserman, J, Am, Chem, Soc, 8l, 5006 (1959.).

20, W, Theilacker, G„ Korttfm and H. Elliehausen, Z, Naturforsch, 9b, 167 (195J+) .

21, Y. Hirshberg and S.I. Weissman, J, Chem, Fhys,, 28, 739 (1958).

22, Y, Matsunaga, Bull, Chem, Soc Japan, 29, 582 (1956) .

23, W, Theilacker, G. Kortflm, H, Elliehausen and H. Wilski , Chem, Ber., 89, 1578 (1956) .

2k. W, Theilacker, G, Kortflm and H. Elliehausen, Chem, Ber., 89, 2306 (1956) ,

25. JoF5D, Mills and S.C, Nyburg, Ref, 3, J. Chem, Soc, 308 (1963) .

26. E, Harnik and G.M.J. Schmidt, J, Chem, Soc, 3295 (195k).

27. R.Bo Woodward and E, Wasserman, J. Am, Chem, Soc, 8l, 5007 (1959),

28. G. Korttfm, W, Theilacker and V. Braun, Z. Physik, Chem, (Frankfurt) , 2, 179 (195^).

29. So Matlow, J, Chem. Fhys., 23, 152 (1955),

30. Y, Hirshberg and E. Fischer, J, Chem. Fhys,, 23, 1723 (1955).

31. G, Kortflm, W. Theilacker and V, Braun, J, Chem, Fhys,, 23, 1723 (1955),

32. Y. Hirshberg and E, Fischer, Angew. Chem,, 70, 573 (.195"oT.

33. G. Kortiftn, Angew. Chem,, 70, 573 (1958) .

3k, J,F,D, Mills and S.C, Nyburg, Ref. 9a,, J. Chem, Soc, 92? (1963) .

- 33 - Fragmentation in Solvolysis Reactions

Reported by W. F. Pickens August 5, 1963

In the broadest sense fragmentation reactions comprise a large part of organic chemistry, including every example of the decomposition of a relatively large mole- cule into two or more smaller fragments. (Simple elimination reactions are not con- sidered as fragmentation reactions.) Thus, retrograde condensations, thermal decom- positions, degradations, and fragmentations observed in mass spectrometry studies are all formally part of this rather broad and diverse subject. It is the purpose of this seminar to consider only heterolytic fragmentation in solvolysis reactions, which may be formulated according to the general scheme 1>2 where X is a nucleophilic

c - d

n

+ x©

©a = b

donating electron density by either a conjugative or inductive effect (e.g. a= alkyl, aryl, -c-®, H0-, R0-, R2N-, etc.), and a,b,c, and d are atoms such as carbon, nitrogen, and oxygen which are capable of forming multiple bonds. The electron deficiency at d is relayed to b by heterolysis of the b-c bond, and is stabilized either transiently or permanently at b by electron release from a. In the third fragment, c=d, a multi- ple bond is formed, at least temporarily. The fate of the fragments a-b © and c=d vary with their structure and environment.

Fragmentation reactions, although frequently observed, have received little at- tention from a mechanistic point of view, and have not been as thoroughly investigated as nucleophilic substitutions, eliminations, and rearrangements with which they often compete. Until recently no attempt had been made to incorporate fragmentation into a general system of chemical reactivity.

As is the case with many other kinds of reactions, certain fragmentations are subject to acid and base catalysis as shown in the acid catalyzed cleavage of 1,3- diols,3

1 1 1

HO-C-C-C-OH 1 1 1

— — ► HO-C-C~C-OH2 iii

HO=C.

+ JC = (C + HpO

the base catalyzed cleavage of 1,3-diol derivatives (X = aryl-S020-) 111 :B 0 1 • 1 ' ^ ^

HO-C-C-C-X < > O-C-C-C-X > 0 = C + C =

ill 111

CL +

X0

and the base catalyzed cleavage of 8 -halo ketones.6

1 OH® © QHi 1

0 = C-C-C-X

R ' '

-> o-c-c-c-x

> RCOaH + c = or + X

0

The general subject of fragmentation in solvolysis

eactions has been discussed This review hall emphasize the work published after 1961, up to August 1, 1963. Grob's most

by C.A. Grob,7->8'9 mainly in terms of his own important contributions.

recent series of articles

10-16

contains many older references not included here. The

Y -amino halides will be treated in detail, and the principles elucidated from the study of this class of compounds will be applied to other systems.

Fragmentation in Degradation. - Fragmentation should be recognized as a useful degradation tool, as the following work by Franck and Johnson illustrates.17 These workers wanted to degrade the heterocyclic ring of the natural product veratramine (I) to give a product useful for comparison with totally synthetic material.

Clf

- y\ -

Treatment of I with N-chlorosuccinimide gave N-chloroveratramine (II) which presumably fragmented under solvolysis conditions to give the imino compound (III) which was not isolated but underwent hydrolysis to the aldehyde (IV) upon addition of hydrochloric acid. The overall yield (1 — -»IV) was 9*$.

Fragmentation of Y-amino halides. - The y -amino haiides are particularly suit' able models for a detailed study of mechanism and stereochemistry of fragmentation reactions.18'19 The stoichiometry and kinetics of their reactions are simple and re presentatives of different but fairly definite geometry are readily available. An early example of the fragmentation of Y-amino halides is the conversion of halogen- dihydroquinidine (V) to niquidine (Vl).2°~23

OH 1 R-CH

X CH-CH3

CH=CH-CH3

R-C

VI

R=

Also, Adamson reported that treating substituted Y-amino alcohols (VII) with boiling acetic anhydride gave diphenylethylene derivatives (VIII) instead of the expected amino olefins (IX).24

\

R 6

f T

> - CH2 - CH - C - OH VII fi

0

(CH3-C^O

R H-

\

C = C&

VIII

-#-

R

1

!N - CH2 - C = C02 IX

Grcb has chosen the Y-amino halides as a major class of compounds for the elucidation of the factors involved in fragmentation. For Y-amino halides, two extreme mechanisms of fragmentation can be postulated: (a) a one-step, concerted process, and (b) a two-step process, both of which will obey first order kinetics

N

N -

(a)

'G

/

c - c - c

I I I

X

#

N^C -- C-

' 1 1

- x

*

0

->®N=C + C = C + X

N

I I

c - c

1 4r '

I 0

CQ + X

1

T

N=C

— C

© 0

+ x

substitution , elimination, quaternization

In the one -step process (a) ionization is assisted by electron release from the amino group. The consequent acceleration of ionization, called the "frangomeric" effect, should make fragmentation the only observable reaction. In the two-step pro- cess (b) the amino group is not directly involved in the ionization step, and a car- bonium ion is formed at a normal rate. Fragmentation then becomes one of several competing reactions in a fast, product -determining step.

The one -step process (a) will occur most readily if all the electron pairs directly involved (e.g. the X-Cq, Cq-Cq, Cg-Oy, and Qy-N bonds, and the lone pair on nitrogen) lie in one plane, or in two planes which intersect in the Cg-Gy bond.25 Thus, stereoelectronically favorable conformations can be derived from the coplanar, staggered form (X) by rotation around the Cg-Cy bond, as indicated by the arrow. The

v V H3C ,9

USfB

\ V

XII

X

XIII

XIV

skev form (XI) is present in 3 £ -substituted tropanes (XII) , and the eclipsed form (XIII) is present in ^-substituted quinuclidines (XIV). Fragmentation of any of the rotational configurations of X corresponds to a trans elimination, whereas if the X-Cq, and Cg-Cy bonds are cis as in XV, the fragmentation corresponds to a less favor- able cis elimination.

The results of a kinetic and product study of some 3-tropanyl chlorides in QCffo ethanol at 62.0° are shown below.7

H3C

k,106(sec."1) rel

V

r

CH3

CH2~CH=CH2

300

2.5'IG4

s

-}

CH2-CH=CH2

16.7

1400

3a -tropanol (7*$) 3P -tropanol (6$) tropidine ( 20$)

7.72

645

CI

OR

(calc.) 0.012 1

The more basic tertiary amine (XVT) reacts considerably faster than the secondary amine (XVII) , In the a-isomer (XVIII) the CT-Ca bond is not parallel with the Cg-Cy bond, and no fragmentation is observed. The fact that the rate is increased relative to cyclohexyl chloride and that the major product ( 3<2-tropanol) has retention of configuration suggests that the basic nitrogen atom participates in the ioni- zation as shown at the right.

Archer and coworkers have reported the cleavage of 3^-tropanyl chloride with aqueous ethanol in the presence of potassium cyanide to give a mixture of 2-allyl-l- methylpyrrclidine-5-nitriles (XIX) ,26"2B The a-isomer (3a-tropanyl chloride) gave

CH3

CH^ - CH=CH2

CH2-CH=CH2 XIX

- 36 -

only substitution with nucleophilic reagents;27 the products showed retention of configuration*

The coplanar y -amino halide system 4-bromo quinuclidine (XX) also reacts by the synchronous fragmentation mechanism.7 Frangomeric assistance apparently is responsi- ble for a very much faster rate of solvolysis of the bridgehead bromide (XX) compared to the carbon homomorph (XXI). Indeed, the rate of solvolysis of XX in 80$ ethanol at 40 7 is faster than that of isopropyl bromide,29 even though the leaving group is at the bridgehead position of a bicyclic system.

Br CH2 Msec."1) krel

xx 1 y \&\ 4.6-10"5 5-yiO4

xxi \\ \ > Ml 8.7*io"10 l

(CH3)2 CH-Br > (CH3)2 CH-OR 1.17*10'6 at 50° 29

(CH3)3 C-Br > (CH3)3 C-OR 2.VlO~3 2.7°106

In an attempt to approach^ a rigid, coplanar staggered conformation of y -amino alcohol derivatives, Grob has studied the cis -isomer of 3-dimethylamino cyclohexyl tosylate (XXII),9 A high rate relative to its carbon homomorph (XXIIIa) was observed, and fragmentation was the only reaction. The trans -isomer (XXIV) in which the break-

CH3 /^-^^? f^ -^

7-^^-^/o^ * k^f(CH3)2 59

c;^a

CH3 XXII

(CH3)2CH-^^0Tos (CH3)2CH

XXIII a) cis b) trans

OR

28$ fragmentation

► .50$ elimination 0.64

8$ substitution OTos 8$ ring closure

ing bonds are nonparallel reacts at a slightly lower overall rate than its homomorph (XXIIIb) . This, along with the mixture of products formed, indicates an unaccelerated two-step process o9

The hydrolysis of 3-bromo-3,3-dimethyl-l-N,N-dimethyl butylamine (XXVa) at 60° in aqueous sodium hydroxide gave a 48$ yield of dimethylamine by fragmentation, along with formaldehyde and isobutylene.10 Also a mixture (~>l:l) of the amino alcohol (XXVb) and the amino olefin (XXVI) was found.

(CH3)2 N-CH2 - CH2 - C(CH3)2 - X (CH3)2 N-CH2 - CH = C(CH3)2

XXV XXVI

a) X = Br

b) X = OH

The bromide (XXV a) reacted with silver nitrate in acetonitrile in the presence of triethylamine to give 60$ fragmentation.10

- 57 -

In the hydrolysis of 1 -methyl -3 (1' -bromo-1' -methyl ethyl) piperidine (XXVII a) ■with aqueous sodium hydroxide, fragmentation occurred to the extent of 6cfp to give the secondary amine (XXVIII a) isolated as the tosylate (XXVIII t>) and formaldehyde. Substitution gave jk1° of the amino alcohol (XXVII b) , and elimination gave 6fo of a mixture of olefins in which the compound XXIX predominated.10

^N-C(CH3)2-X ^E m C(CH3)2 /\ C(CH3)=CH2

^ x™11 ta3 xxviii tH3 XXIX

a) X = Br a) R = H

b) X = OH b) R = OTos

The 3 -(lf -bromo-11 -methyl ethyl) quinuclidine (XXX a) gave 62$ of XXXI by frag- mentation, some of XXXII by elimination- and a little of XXX b by substitution when treated with aqueous sodium hydroxide.1

CH=C( CH3) 2

C(CH3)2-X r^ r^N= c(c%)2 /^\.CH2-X

n

xxx xxxi ^ik xxxii \n^ xxxiii

6h?

13

a) X = Br a) X = Br

b) X = OH b) X = OTos

The primary bromide (XXXIII a) and the corresponding p_-toluene sulfonate ( XXXIII b) did not fragment due to the difficulty of ionization. Only substitution was observed with these compounds in aqueous sodium hydroxide.10

The fragmentation of secondary y -amino halides of conformational ambiguity follows a more complex pattern. Several N-alkyl-4-chloro piperidine derivatives have been studied, but there is no simple relationship between the size of the sub- stituent on nitrogen and the overall rate or the extent of fragmentation.9 These reactions appear to follow a two step accelerated mechanism involving nitrogen par- ticipation. The results of solvolysis in QCffo ethanol at 62 are shown below.9

R	krel	p fragmentation
- H	1400	100
- CH2-CH3	330	100
- CH(CH3)2	295	100
- CH3	165	95
- C(CH3)3	105	92

OR

A further example of the failure of concerted fragmentation to occur when the bonds are not properly oriented is provided by a study of suitable cyclopentane derivatives.8 In the trans form of 2-(dimethylamino methyl) -cyclopentyl tosylate (XXXTV) the Ca-0Tos and Cg-Cy bonds are not coplanar . In the cis compound (XXXV) the bonds are coplanar, but concerted fragmentation in this case would correspond to an unfavorable cis elimination. Both compounds (XXXIV and XXXV) solvolyze at about the same rate as cyclopentyl tosylate. Ring closure is favorable in the case of the trans isomer, as shown by the appearance of 6cff> of the azetidinium salt and a slight increase in rate due to a small participation effect of the amino group.8

- j}8 - substitution

rel

Q

XXXIV

CH2~N( CH3) 2

OTos

/^t-* CH2-N(CH3)2

XXXV

OTos

■» elimination

ring closure (60$)

/I + H2C = N ( CH3) 2

substitution elimination

OR

The effects of methyl groups in the a-, £-, and y -positions of the y -aminopropyl chain have been studied.9 In the case of primary or secondary chlorides, ring closure to an azetidinium salt is the only observable reaction in Qcffo ethanol at 56°. 9 The rates of solvolysis of the y -amino chlorides relative to the rates of solvolysis of their carbon homomorphs are shown below.

y -amino chloride

k/kh

( CH3) 2 N-CH2-CH2-CH2-CI

( CH3) 2 N~CH2-C( CH3) 2-CH2-CI

( CH3) 2 W-CH2-CH2-CH( CH3) -CI

6'io3 5'106 225

% ring closure 100

100

homomorph

( CH3) 2 CH-CH2-CH2-CH2-CI

( CH3) 2 CH-CH2-C( CH3) 2-CH2-CI

(CH3)2 CH-CH2-CH2-CH(CH3) -CI

The chemistry of azetidines was reviewed in I96I.30

In the case of tertiary chlorides, the rates of solvolysis of Y -amino chlorides are practically the same as the rates of solvolysis of their carbon homomorphs, in- dicating that the amino group is not participating in the rate determining step. In the stepwise process the nature of the Y-substituent should affect the product com- position markedly, but only slightly the rate at which products are formed. In the examples below, solvolysis occurred at 56 in 80$ ethanol.9

R-CH2-CH2-C( CH3) 2-CI

slow.

R

XXXVI

( CH3) 2CH-

H2N-

(CH3)2 N-

CH3 3 N-

k°104(sec."1)

4.73 h.66

3.55

R-CH2"'CH2

(CH3)

fast

1.0

0.99

0.75

(extrapolated) 1.5*10~2 (extrapolated) 9.6*10~3

•> products

dP fragmentation

0 20

50

0 0

A more careful study of the N,N,a,a-tetramethyl-Y -aminopropyl chloride case (XXXVI) showed products corresponding to Wf* fragmentation, 25$ substitution, 25$ elimination, and yfo ring closure.9

The mechanism changes again if two additional methyl groups are introduced into ^he 0-position. The compound XXXVII solvolyzes 125 times as fast as its homomorph, indicating participation of the Y-amino group.9

(CH3)2 N-CH2-C(CH3)2-C(CH3)2-C1 -^-> (CH3)2 N^- -C(CH3)2 — 8?1 l4> fragmentation

XXXVII

H2C - C(CH3)2 * 20$ elimination

This effect can be explained by an internal solvotation by the nitrogen assisting the ionization of the chloride ion without formation of a convalent bond, since the

- 39 -

azetidinium salt (which is stable under the reaction conditions) is not formed. This mechanism also applies when the geminal methyl groups occupy the y -position. The ratio of the rate of solvolysis for compound XXXVIII to the rate of solvolysis of its carbon homomorph is 27. 9

(CH3)2 N-C(CH3)2-CH2-C(CH3)2-C1 K CH3) aN^-- C( CH3) a— %¥ 8($ fragmentation

XXXVIII (CH3)2C - CH2 > 20$ elimination

Fragmentation Induced by Rearrangement. - In unassisted fragmentations the carbonium ion center may be provided by preliminary rearrangement instead of initial ionization at that point. As shown in the reaction scheme below, a 1° 5-amino carbonium ion may rearrange to a fragmentable Y -amino carbonium ion.12

1 ' "?" - i , • "?" ^Q / - -c-

N-C-C-C-C e > N-C-C-C-C > N = C + C = C - C -

1 ' -c- " ' ' ' -c- ' ' - ' -c- ' I I I

A suitable compound for this study was 4-tosyloxy-methy] quinuclidine (XXXIX) which was found to undergo the following reaction.12

CH2OTos Ciy© @

o

> >

XXXIX K n ^N

Solvolysis of the homomorph, 1-tosyloxymethyl-bicyclo [2.2.2] octane (XXXX), had been investigated previously and was found to produce the alcohol ( XXXXI). 31

CH2-0Tos CK£) ^ >

xxxx ^ ^^ XXXXI

The kinetic measurement of the rate of solvolysis for neopentyl tosylate (XXXXII) was also carried out for comparison.12

(CH3)3 C-CH2-OTos — » (CH3)3 C-CH2e — -* (CH3)2 C-CH2-CH3 — >(CH3)2 C-CH2-CH3 XXXXII

'The results are summarized below for 0.01 M solutions in 80$ ethanol with 0.015^ M trimethyl amine at 116.00 +0.03 . 12

Fragmentation of a-Amino Ketoximes

compound k°105(sec„ ~1) krei

XXXIX 11. 02+0. 05 1 (Second Order Beckmann Reactions)

(.1.00+0.02 Oximes of a-amino ketones represent

systems capable of fragmentation to

XXXX 4.^2+0.09 k.k

|2.0oK>.02 2.1 nitriles and ternary iminium salts.32™35

XXXXII 1 2. 09+0.01 (0.0173M BtaH) The ability to fragment is greatly

enhanced by esterification or etheri- fication of the hydroxyl group. Recent work by Grob14>15,16 describes the fragmenta- tion of various tosylates, 2.4-dinitrophenyI ethers, and benzoates of a-amino ketox- imes of known configuration. 3 The field of abnormal Beckmann reactions has been well reviewed in these papers 14>15 as well as in a recent general treatment of the Beckmann rearrangement.36 Also, Grob's experimental work in this field cannot re- ceive fair treatment here. Therefore, only a few general observations will be noted. The ant! forms are much more reactive than the syh forms (lc^kg^ 2*103) ,14 Fragmen- tation of the anti compounds corresponds to a coplanar trans elimination, and pro- cedes quantitatively in 80$ ethanol at 13°, whereas the syn compounds give both fragmentation (corresponding to a less favorable cis elimination) and the normal Beckmann rearrangement.14

- 40 -

^N R ^<s .R n ^ _ 0

anti

C - c'

^N R ~ ^C — CX^ -> <$ N = C+RC=N + XW

c,vn C - C. R > C - C - N

+xe

The fragmentation of p-keto ether-oximes37and of a Y-oximino ketone38 has also been reported.

Fragmentation of p-Bromo Cinnamic Acids. - The conversion of cinnamic acid di- bromide to 0-bromostyrene has been well investigated.39"41

Br Br Q

fi - CH - CH - C02 > fi HC = CHBr + C02 + Br

The decarboxylations of p-bromo cinnamic acids provide further evidence for the stereospecifieity of fragmentation.9 The sodium salt of the cis compound (XXXXIII) , having the ~Br and -C02 groups trans, reacts by an accelerated synchronous mechanism to give phenyl acetylene exclusively. The sodium salt of the trans compound (XXXXIV) reacts slower and gives both phenyl acetylene and acetophenone . The decarboxylation rates have been compared with the rate of solvolysis of a-bromostyrene in water at 100°. 9

a © rel

<b. _ C02 n

xc m"z' — *• 0-C = CH + C02 + Br 103

Brj NH

^XXXXIII

6 3. j?

NC = CX ^ > fi - C - CH3 + <b - C = CH + C02 10

Br' ^C02@

xxxxxv

Br'

0 II

C = CH2 — — - > p - C - CH3

Another very similar fragmentation caused by the treatment of the a, 0 -unsatu- rated chloroaldehyde (XXXXV) with base produces a phenyl acetylene as shown below.42 CI ' 0 CI 0 Q

H3C0 ^^^ C=CH~CH H3C0/r^C=CH-Cv HaCO^^C^CH

if PIT y if H 0H * T IJ + HC02H+C1

H3C0 ^y HaCO^^ HaCO^y

XXXXV

Bibliography

1. C.A. Grob and W. Baumann, Helv. Chim. Acta, ^8, 59^ (1955)*

2. C.A. Grob, Experientia, 13, 126 (1957).

3. H.E. Zimmermann and J. English, Jr., J. Am. Chem. Soc, j6, 22$k (195*0 , and references therein.

h. R.B. 'Clayton, H.B. Henbest, and M. Smith, J. Chem. Soc, 1957, 1982.

- in -

5. H.B. Henbest and B.B. Milliard, J. Chem. Soc, i960, 3575-

6. F. Nerdel, H. Goetz, and M. Wolff, Ann. Chem., 632, 65 (i960).

7. C.A. Grob, in "Theoretical Organic Chemistry", (Report on the Kekule Symposium, London, 195 8) , Butterworth, London, 1959, p. ll£.

8. C.A. Grob, Bull. Soc. Chim. France, i960, I36O.

9. C.A. Grob, Gazz. Chim. Ital., 92, 902 (I962).

10. C.A. Grob and F. Ostermayer, Helv. Chim. Acta, 45, 1119 (1962) .

11. C.A. Grob, F. Ostermayer, and W. Raudenbusch, Helv. Chim. Acta, 45, 1672 (1962) .

12. C.A. Grob, R.M. Hoegerle, and M. Ohta, Helv. Chim. Acta, 45, 1823 (1962) .

13. H.P. Fischer and C.A. Grob, Helv. Chim. Acta, 45, 2528 (1962) .

14. H.P. Fischer, C.A. Grob, and E. Renk, Helv. Chim. Acta, 45, 2539 (I962) .

15. H.P. Fischer and C.A. Grob, Helv. Chim. Acta, 46, 936 (1963) •

16. C.A. Grob, Helv. Chim. Acta, 46, 1190 (1963) , in press.

17. R.W. Franck and W.S. Johnson, Tet. Letters, 1963, No. 9, 545.

18. C.A. Grob, Angew. Chem., 69, 680 (1957).

19. C.A. Grob and F.A. Jenny, Tet. Letters, i960, No. 23, 25.

20. E.M. Gibbs and T.A. Henry, J. Chem. Soc, 1939, 240.

21. W. Solomon, J. Chem. Soc, 194l, 77.

22. H.S. Moscher, R. Forker, H.R. Williams, and T.S. Oakwood, J. Am. Chem. Soc, 74, 4627 (1952).

23. R.B. Turner and R.B. Woodward, in Manske and Holmes, "The Alkaloids", Academic Press, New York, N.Y. , 1953, p. 21.

24. D.W. Adamson, Nature, 164, 500 (1949).

25. A.T. Bottini, C.A. Grob, and E. Schumacher, Chem. Alnd. (London), 1958, 757.

26. S. Archer, T.R. Lewis, and B. Zenitz, J. Am. Chem. Soc, 79, 3603 (1957) .

27. S. Archer, M.R. Bell, T.R. Lewis, J.W. Schulenberg, and M.J. Unser, J. Am. Chem. Soc, 79, 6337 (1957).

28. S. Archer, T.R. Lewis, and B. Zenitz, J. Am. Chem. Soc, 80, 958 (1958) .

29. C.G. Swain, R.B. Mosely, and D.E. Bown, J. Am. Chem. Soc, 77, 3731 (1955).

30. P. Tschampel, U„ of 111. Organic Seminar Abstracts, Semester II, I96O-6I, p. 127.

31. C.A. Grob, M. Ohta, E. Renk, and A. Weiss, Helv. Chim. Acta, 4l, 1191 (1958) .

32. R.K. Hill and R.T. Conley, Chem. & Ind. (London), 1956, 1314.

33. R.K. Hill and R.T. Conley, J. Am. Chem. Soc, 82, "6*45 (i960) , and references therein.

34. H.P. Fischer, C.A. Grob, and E. Renk, Helv. Chim. Acta, 42, 872 (1959).

35. H.P. Fischer and C.A. Grob, Tet. Letters, i960, No. 26, 22.

36. L.G. Donaruma and W.Z. Heldt, in Adams, "Organic Reactions", Vol. 11, Wiley, New York, N.Y. , I96I, p. 1.

37. R.K. Hill, J. Org. Chem., 27, 29 (1962), and references therein.

38. W. Eisele, C.A. Grob, and E. Renk, Tet. Letters, 1963, No. 2, 75.

39. E. Grovenstein, Jr., and D.E. Lee, J. Am. Chem. Soc, 75, 2639 (1953), and references therein.

40. S.J. Cristol and W.P. Norris, J. Am. Chem. Soc, 7£, 2645 (1953), and references therein.

41. H. Sliwa and P. Maitte, Bull. Soc. Chim. France, 1962, 369.

42. K, Bcdendorf and P. KLoss, Angew. Chem., 75, 139 (I963) .

- k2 -

DIMETHYL SULFOXIDE AS SOLVENT AND REACTANT

Reported by P. Rivers August 7 , 1963

Dimethyl sulfoxide* is a colorless, nearly odorless liquid which boils at l8o° (760 mm Hg) and melts at I8.50. It has a specific gravity of 1.1014 at 20^ and a refractive index of 1.2+783 (at 20°). Although this compound has been known since I867 (l) it is only within the last five years that it has become commercially avai- lable at reasonable prices. The two main sources of supply in this country are the Crown Zellerbach Corporation and Stepan Chemical Company. Commercially it is prepared by the oxidation of dimethyl sulfide using Nr03, N02, N204 or NO (with air) . Other physical properties are given in Table I and a more comprehensive coverage can be found in a pamphlet distributed by Crown Zellerbach Corporation (2)

	Tabl	e I
Heat of Vaporization		(189°)	132 cal/g
Heat of Solution		(20°)	60 cal/g
Heat of Fusion		(189°)	20 cal/g
Heat of Combustion		(25°)	6050 cal/g
Coefficient of Expansion		0.00088 per
Dielectric Constant			h5
Molecular Weight			78.13

This discussion will be divided into two main sections } (i) DMS0 as a reac- tion solvent and (il) those systems in which DMS0 is both solvent and reactant.

I. DMSO as a Solvent . --This discussion will be restricted to reactions in which basic or neutral conditions are employed due to limitations of space, the recent interest in this particular area and its greater importance to synthetic and physical organic chemistry. In terms of its solvent properties, we are mainly concerned with two questions: how and why does DMSO differ from other solvents and what reactions can be carried out in DMSO most advantageously and why.

1. Base-Catalyzed Reactions. (a) Base catalyzed hydrogen-deuterium exchange. Recently Cram has conducted a study of the stereochemical factors which influence electrophilic substitution at saturated carbon atoms. His earlier work in this field has been. covered in a previous University of Illinois Organic Seminar (3) and will not be discussed except in terms of its relevance to DMSO. Cram, et al. (k) , have been able to show a wide variation in stereospecificity for electro- philic substitution at saturated carbon atoms, depending on the solvent employed. Suitable metal alkoxides were observed to undergo cleavage with retention of con- figuration in t-butyl alcohol. In sharp contrast, the same reaction occurred at much lower temperatures with almost complete racemization in DMSO. Significant

ovf

H,' ^ H ' B H,-" R5/

differences, in rates which were most pronounced in DMSO, were observed as the metal ion was changed; K>Na>Li.

Recently, Cram described the base catalyzed hydrogen-deuterium exchange reac- tion of optically active 2-phenylbutane, 1-phenyl-l-methylethane and 2-methyl-3- phenylpropionitrile. Cram, Rickborn and Knox (5) found that these compounds could be racemized at 25° in DMSO at a rate that could be compared to that attained at 173° in t-butyl alcohol. Extrapolated rates of proton abstraction were calculated to be 107 to 10s times greater in DMSO than in t-butyl alcohol. Compared to results reported by Cram, Rickborn, Kingsbury and Haberfield (6) this would suggest a spread of eleven powers of ten in going from methoxide ion in methanol to t-butoxide ion in DMSO. It is important to note that the dielectric constants of the two solvents differ only slightly; 3k for methanol and k$ for DMSO.

^Throughout this discussion dimethyl sulfoxide, (CH3)sS0, will be abbreviated as DMSO. In a few cases in the literature DMS is used but DMSO is the predominate form.

- k3 -

(b) Wolff -Kishner Reduction. --Evidence has been presented that the

Wolff -Kishner reaction can be conducted at room temperature using sublimed potas- sium t-butoxide in DMSO (7) as compared to the normal reaction temperature of 100- 200°. Yields ranging from 60-90? were obtained in all cases reported. Alkali metal t-butoxide, even when dried under the severest conditions, are known to exist as a complex of the salt and alcohol (9) •> The use of sublimed t-butoxide has been recommended by several authors (8, 9,31) when an alcohol-free system is desired. The sublimed potassium salt is now commercially available (lC).

(c) Acidity of Hydrocarbons. — Hofmann, Muller and Schriesheim (8) have stu- died the ionization rates of toluene and several polyalkyl benzenes using sublimed potassium t-butoxide in tritated DMSO. Since benzene and t-butylbenzene remained unchanged after extended reaction periods and elevated temperatures, it was con- cluded that only a-hydrogens can undergo exchange. It was predicted that this medium could be used to study hydrocarbons of pKa^AO. (NH3~35> CH4c^58) ( ^7) •

(d) Isomerization of Alkenes. --Schriesheim and Rowe (9), using 2-methyl-l- jentene found that anionic isomerization could be carried out in DMSO with sub- limed t-butoxide salts at temperatures of 40-70°.

CH3 CH3

CHg = C-CH2CH2CH3 CH3-C = CHCH2CH3

CH3

The formation of 4-methyl-l-pentene, CH3-CHCH2CH = CH2> occurs only after extended reaction periods (230 hrs) . These results are explained on the basis of the rela- tive stabilities of the following carbanions. The resonance forms which would lead

(1)

(2)

fc)

to the formation to 4-methyl-l-pentene would be less favorable since they are simi- lar to secondary and tertiary carbanions, while those resulting in 2-methyl-2-pen- tene involve primary and secondary carbanions.

Schriesheim, and co-workers (11,12,13), have investigated the effects of ring size on the rates of formation of anions from alkylidenecycloalkanes (i) and cyclo- alkanones ( II) . The base catalyzed exo- to endo -rearrangement using potassium t-

CH2	0
II	II
/c\	/CN
CH2 CH2	CH2 CH2
X(CH2)/	^(CHs)^
n = 1 to 5	II n = 1 to 5

butoxide in DMSO and the base-catalyzed bromination of the ketones were studied and found to correlate closely. Since these results run counter to Brown's I-strain

Exo-.	Endo rearrangement		Bromination
n	rel. rates ( Cq-	=1)	n	rel.	rates (Cs=l)
1	1070		1		15=5
2	k$k		2		9-9
3	1		3		1.0
k	5.8		h		1.7
5	17		5		3.fc

theory (lM, they were rationalized following the explanation of Corey (15) for the orientation of bromine in a-bromoketostcroids . According to this, the rate of pro- ton removal will be greatest when there is maximum opportunity for overlap between

kk -

the sp-->p orbital made available by the leaving hydrogen and the rt-orbitals or the exo-cyclic system. The authors state that Dreiding models show that this geometrical requirement follows the observed rate order,

(e) Nucleophilic Displacements. --When bromobenzene in DMSO was treated with a solution of sublimed potassium t -but oxide in DMSO at 25° for 15 hrs., the resulting product was t-butylphenylether in 86$ yield (5)=. Similar results were obtained in t-butyl alcohol at 175° for nine hours in 37> yield. By competition experiments, fluorobenzene was found to react at l/25 the rate of bromobenzene. o-Fiuorotoluene could be converted to o-cresol, contaminated by less than 3^ of the met a- isomer. o-Bromotoluene gave a mixture of k parts m-cresol and 1 part o-cresol. The parent compounds were suggested as being the corresponding t-butyl ethers which probably decomposed in the process of V.P.C. analysis. These results would definitely point to the possibility of an aryne intermediate,

Kornblum, Berrigan and LeWoble (l6,17.) have observed definite solvent dependence in the course of the alkylation of ambident anions. Using sodium phenoxide (16) and sodium naphthoxide (l8), it was concluded that the carbon to oxygen alkylation ratio reflected both the effects of dielectric constant and protonic nature of the solvent. Table II shows clearly that the nature of the substrate anion is also important in determining the c/o ratio.

Solvent

Dimethylformamide

DoM.S.O.

Ethylene glycol dimethyl ether

Tetrahydrofuran

Methanol

Ethanol

2,2,2-trifluoroethanol

Water

•able II*
	Phenoxide
Dielectric
Constant	C$ (#
37	0 91
*5	__
7	0 99
7'	0 96
3k	0 96
2k	0 99
27	26 62
80	2k 65

J10

0	97
0	y5
22	70
36	60
3k	57
28	52
85	7
8k	10

*Taken from data on benzyl bromide

(f) Radical Anions. --Russell and Janzen ( 19) have reported the facile formation of radical anions from a variety of nitroaromatics in the presence of strong base, both in t.-butyl alcohol and DMSO. Similar results were obtained from nitrobenzene (20) and from compounds capable of forming dianions in the

presence of their unsaturated analogues (21). The electron transfer was found

[0-N=N-0f + 0-N=H- 0 ^—> 2[0-N=N-0]~

to take place most effectively in DMSO, Other results (22) seem to imply that electron transfer between organic donors and acceptors is a general reaction observable for numerous carbanions, nitranions, phenoxide and mecaptide anions even at dilute concentrations in DMSO and t-butyl alcohol.

iL, Reactions without -Base-Catalysis. (a) Dehydration- --Traynelis , Hergen-

rother, Livingstone and Valicenti (23) observed that upon heating 2-hydroxy- 2,3^,5-tetrahydrobenzo [b] thiopin (III) in a large excess of DMSO at 160- l8o° for nine to sixteen hours, water was eliminated, A number of alcohols

OH

DMSO

- 45 -

were shown to undergo this reaction in DMSO. The reaction does, however, appear to be limited to secondary alcohols. Subjection of 1-pentanol and p-phenethyl alcohol gave only unchanged starting materials. Gillis and Beck (2^) had hoped that this would be a convenient route to the exomethylene norbornanes, but in- stead found that ohe reactions of l,k diols in DMSO led to high yields of tetra- hydrofuran derivatives, y -/

CH2OH

CHoOH

OH OH (CH3)2-C-(CH2)2-C(CH3)2

DMSO

TBo^ 52$

In the second reaction the expected dienes were also found and identified. The authors proposed a cyclic transition state However, only polymeric and starting

0) *'0$&

/

0'i'"i

3 1 - CH3 H <^CH3

materials were isolated when the reaction was attempted with cis-butene-I,4-diol.

(b) Elimination Reactions. --Recently, Cram nas begun a study of solvent effects on the N-amine oxide elimination reaction (?)• Threo and erythro-N,N- dimethyl-3-phenyl-2-butylamine oxide, which upon elimination yield cis and trans- 2 -phenyl -2 -but ene and 3-phenyl-l-butene, were studied in DMSO (25-50°T, tetrahy- arofuran (25-100°), and aqueous mixtures of these solvents.

0 N(CH3)2 0 0

1 I 20°-138° \ X

CH3-CH-CHCH3 J ., C=CHCH3 + CHCH=CH2

iv CH3 v CH3 VI

The production of" V was completely stereospecific in all media as measured by V. P. C. analysis. For the threo-isomer of IV the ratio of V to VI changed from k-9 in DMSO to 19 in tetrahydrofuran to 7 S-n water (132°) - The erythro isomer of IV for the same solvents gave ratios of l6, 19 and 11 respectively. All reactions were repor- ted to be first order in substrate. Rate comparisons for the threo compounds at 25° are as follows:

kDMS0/kH20'~ ]-°5 kTHF/kH20 " 1C)6 and for the erythro:

kDMS0/kH20~104 kTHF/'kH20~105 Rate depressions, relative to the pure solvent, of approximately 200 for yC$ THF and 5 to 10 for 90$ DMSO (l<# H20) were noted. This effect reflects the deacti- vation of the N-amine oxide oxygen by hydrogen bonding. It is less in DMSO because the solvent is able to compete with the substrate for hydrogen bonds.

Any attempt to correlate the varied results obtained in any dipolar aprotic solvent is difficult and especially so in this case. It is felt, however, that most of the solvent properties of DMSO can be explained on the basis of its inabi- lity to solvate small anions through hydrogen bonding while possessing the ability to solvate cations and to function well as either a nucleophile or electrophile. Cram, Kingsbury and Rickborn (25) explain the difference between protonic 'reten- tion" solvents and aprotic "racemization" solvents, of which DMSO is an outstanding example, on the basis of asymmetric vs. symmetric solvation.

,R

%

► C-H

D-0

/

I

R

- k6 - D-Q-R

►ce o —

/ 'H^ \R^~

-4^' \)-R

/ '""'M*^

h/0ne

a symmetrically solvated ion pair asymmetrically solvated

intimate ion pair

*C-D D-C" / \

R

/

ROH 0.

t

D CiimmiD^ XR

►C-D

Retention configuration

loss of configuration symmetrical solvated iori

0

►C-D /

)-R

DMSO ,_ \ i Vs T^ n ►> (CH3)2~S""""" C m.M.rD-0

\

R

ROH

^C-H H-C / \ loss of configuration

asymmetrically solvated anion

— f ch3) 2a11""" S9"""" s( ch3)

3^2

symmetrically solvated anions

Hence the difference in the stereospecificity of base-catalyzed hydrogen-deuterium exchange in DMSO vs. t -butyl alcohol can possible be explained on the basis of the inability of DMSO to contribute hydrogen bonding. The tremendous differences in reaction rates can also be related to this same feature. There is considerable evidence that anions are solvated in pro tonic solvents by hydrogen bonding. Cram has postulated that the alkoxide ion is present in DMSO as a "naked" anion, unsta- bilized by hydrogen bonding and hence is much more reactive than one well solvated. It has been shown in other cases that base strengths of anions are greatly enhanced in DMSO, It has even been claimed (26) that elimination of HBr from alkyl bromides can be catalyzed by fluoride ion in this solvent to yield the corresponding alkenes. This is probably a combination of base strength enhancement and the tendency of DMSO to solvate large^ highly polarizable anions .

The dehydration reactions can possibly be related to the electrophilicity of the solvent, which is probably heightened at elevated temperatures. At room tem- perature, DMSO is believed to exist in chains with hydrogen bonding between the methyl hydrogens and oxygens (27) or more likely with sulfur-oxygen linkages (28). This order is reflected in its high heat of vaporization and low heat of fusion.

CH3 OH3 CH3

I /CH3

S— >-0'

l/CHs

,/

OH.

0

As with water at 37° > DMSO undergoes a structural change between kO and 60°. Plots of physical properties against temperature show discontinuities at this point. The high polarizability of DMSO is certainly of assistance in the dehydrobromi nation reaction,,

It is difficult, at this time, to correlate the effects of dielectric constants in most of the reactions discussed. The cases reported by Cram, where it appears to be unimportant, are in sharp contrast to its inferred prominence in the naphthoxide enolate work of Kornblum. Schriesheim (9) also found no obvious relationship between the isomerization rate of 2-methyl-l-pentene and the dielectric constant of the medium.

II. DMSO as Reactant and SoIvento--Up to this point we have considered DMSO as an inert solvent. A number of reactions of this solvent have been observed to take place, often under identical conditions to those In which we assume inertness.

- k7 -

New reactions of this solvent are constantly being reported. In considering the reactions previously discussed it should he remembered that in no case was the fate of the solvent thoroughly explored.

The stability of an anion Of to the sulfoxide group has long been a highly debatable question, mainly because of the uncertainty in defining the nature of the sulfur-oxygen bond. Last year Corey and Chaykovsky (29) presented the first in a series of observations on the chemistry of the parent compound of a possible family of carbanions stabilized by the ° group; methylsulfinylcarbanion It is easily generated by dissolving (R-S-)

sodium hydride or sublimed potassium t-butoxide in excess DMSO at 65-70 under nitrogen. Sodium amide has also been found to work equally well. The methylsul- finylcarbanion is strongly basic , as was shown by its reaction with triphenyl methane producing a deep red solution which was assumed to be the triphenylcar- banion. An equilibrium constant of 21 - k (25°) was reported for the reaction, clearly showing that the methylsulfinylcarbanion is more basic than the triphenyl- carbanion. Chlorobenzene was found to undergo substitution reactions with excess sulfinylcarbanion to yield methyl benzyl sulfoxide (6.7 mole excess) at room temperature. Keaction of the anion with benzaldehyde and benzophenone gave com- pounds VII and VIII OH 0 OH 0

0-CHCH2SCH3 (0)2-C-CH2SCH3

VII VIII

Observing that methylsulfinylcarbanion would react with phosphonium salts to form Wittig type ylides Corey and Chaykovsky (30) attempted the use of the carbanion itself as a Wittig reagent. Walling and Bollyky (31) found themselves in the same field when they attempted to study the base -catalyzed homogeneous hydrogenation of benzophenone in DMSO. In simultaneous publications both groups reported that the products 01 the reactions were 1,1-diphenyl ethylene, diphenyl- methane and 1,1-diphenylcyclopropane. Both MaH and sublimed potassium t-butoxide were used as bases. Walling proposed the following reaction scheme:

1. CH3SCH3 + C4H90e .7=1 C4H9OH + CH3S0CHf

2. CH3S0cl2 + Ph2C0 ~ — w CH3SOCH2CPh20@

3. CH3SOCH2CHi20@ t CH3SC| + CH2=C.Ph2

k, CH3S0Ch| + CH2^CRi2^± CH3SOCH2CH2?Ph2 0

5. CH3SCH2CH2CEh2 ■ ->. CH3SOCHCH2CHFh2

6. CH3SOCHCH2CHPh2

7. CH3bOCH2CH2Sph2

-*- CH3S0CH=CH2

+ CHPh2 — ii— >- CH2Ph2

CH3S

•s+ zy

0

Step 3 is suggested as being analogous to the Witt iff reaction and probably proceeds via a cyclic transition state;

0

t ©

CH3-SCH2CPh20

£\ C1

f^c-cl

ch3 v r.cH2

\ S^

\ 0

CH3S02

+ CH2=CPh2

Since 1,1-diphenylethylene was postulated as the intermediate in the formation of diphenylmetnane and 1,1-diphenylcyclopropane, Walling and Bollyky subjected it to the reaction conditions and did obtain the expected diphenylmetnane (l]?j and 1,1-diphenyl cyclopropane (&>) . However, the unexpected, and so far unex- plained formation of 3,3-diphenylpropene was the major reaction (77/°).

- li8

Similar results for the benzophenone reaction were also reported by Chay- kovsky and Corey and interpreted in the same way. These workers, however, were able to identify two additional products; diphenylacetaldehyde and 1,1-diphenyl- 2-methylthioethylene. The aldehyde formation was explained as follows-

CH3fOCH2CRi20

Ph.oC

CH2 + CH3SO

Ri2CHCHO

It should be pointed out that Cope (32) was unable to observe ring opening in the same epoxide in strong base, while Cristol, Douglass and Meek (^-8) were able to isolate diphenyl acetaldehyde upon treatment of the epoxide, with phenyl lithium. No explanation was given for the production of l,l-diphenyl-2-methylthioethylene, Ri2C=GHSCH3.

Greenwald, Chaykovsky and Corey (33) have shown that the addition of the anion to phosphonium salts greatly facilitates the Wittig reaction and even unreactive ketones like camphor gave a 7^7° yield under these reaction conditions.

Kuhn (3^) and Kuhn and Trischman (35) observed that DMSO,, in contrast to amine oxides, does not produce the expected alkylsulfonium halides in reaction with alkylhalides, but instead gives the trimethylsulfoxonium halides (iodides being the only halide used). This was later confirmed by Major and Hess (36) • Smith and Winstein (37) found that while o-alkylation was kinetically favored S-alkylation was thermodynamically preferable. It should be pointed out that other halides will not undergo the reaction under the conditions used with the iodides.

According to Corey and Chaykovsky (38), treatment of th.: trimethylsulfoxonium halides (IX) in DMSO with strong base, produces the dimethyl sulfoxonium methylids (X). The methylide (X) was found to be highly reactive toward carbonyl

( CH3) 3S=<

0

(ix) (CH3)

>

Rl-C-R2 + (CH3)2S^

0

CH2 JO-

(X)

Ri-C

V

<>

Ep + DMSO

CH;

&

:0

groups. All reactions were carried out at 25 for one hour and at 50 for an

additional hour. The practicality of this reagent is demonstrated by the prepa- ration of cyclooctanone and cycloheptaldehyde.

0^ HO CHpMp

ll1o

II

Ha M;

MH3

Gcffo

y BF3«EtpO CHO

Compounds which are susceptible to Michael addition (e.g., benzalacetophenone) responded quite differently. The products in these cases were found to be cyclo- propane derivatives in high yield. Only minor amounts of the epoxide could be isolated. 0

0-CH=CHC-0 -f (X)

H

0

0 C-0

Because of the interesting and unusual results obtained with the sulfoxonium methylide, Corey and Chaykovsky (39) have begun a study of the dimethyl sulfonium

- U9 -

methylide; (CH3)2S=CH2 (XI). The sulfonium ylides had been reported earlier by Franzen, Schmidt and Mutz (ko) to be unstable. Corey was able to prepare the dime- thyl sulfonium methylide in DMSO-THF solution at -10°. It was found to have a half- life of a few minutes at 0° compared to the sulfoxonium ylide which can be stored for several days at 25°. The dimethyl sulfonium methylide was found to react with a variety of aldehydes and ketones. Similar to the sulfoxonium ylide , the sulfonium compond gave rise to the corresponding epoxides. Yields ranged from 75 to 97$ • In

/. contrast to the sulfoxonium

_CH fCH- ) S ylides, which tend to form

' 2 3 2 cyclopropane s with potential

Michael reactants, the sulfonium compound selectively affords the epoxides.

On the basis of relative stabilities of the sulfoxonium and sulfonium ylides, along with their relative rates of base-catalyzed hydrogen-deuterium exchange (h6) , Corey and Chaykovsky concluded that the sulfonium ylide is a much more effective methylene transfer agent than the sulfoxonium.

Replacement Reaction. --Ratz and Sweeting (40,4l) have demonstrated another un- usual reaction of DMSO. If 3,9-dichloro-2,4,8,10-tetraoxa-3,9-diphosphaspiro [5<5] undecane-3,9-dioxide is dissolved in "anhydrous" DMSO the following general reaction occurs. 0 0

C1.+ (_DMS^ §_£.

Oxygen Exchange. — Kornblum, et al. (12), in an earlier note reported transfor- mation of phenacyl halides to keto-aldehydes by merely dissolving the halide in DMSO at room temperature for nine hours. Benzyl bromide was later (hj) found to undergo 0 0 the reaction in the presence of sodium bicar-

II DMSO N j bonate or if the halide is first converted to

2 v the tosylate and then treated with a mixture

of sodium bicarbonate in DMSO. Nothing is said about whether the medium was anhydrous so that water may have taken part in the reaction at some stage.

Searles and Hays (kh) have oxidized sulfides to sulfoxides and these have been

oxidized to sulfones (45). / x >_

v J' R2S + (CH3)2S0Z^: R2S0 + CH3SCH3

R2S0 + R2S0 > R2S02 + R2S

Conclusion. --DMS0 is an excellent solvent for systems in which very strongly basic conditions are desired. This partly explains the great interest being shown in this compound. Other strongly basic systems usually lack either ease of manage- ment, stability, reproducibility or all three. DMSO surmounts all of these prob- lems. Aside from the numerous reactions already mentioned it should be interesting to observe the effects of DMSO on the formation of carbenes, pyrolysis of esters, the malonic ester synthesis, the Favorskii reaction, the Perkin ring closure, the Stevens rearrangement, the Smiles rearrangement, the Von Riehter reaction, the study of carbanion geometry and nucleophilic displacements.

There are, however, other problems involved in the use of DMSO. It is known that the solvent disproportionates at elevated temperatures (l60-l8o°) .

2( CH3) 2S0 r CH3SCH3 + ( CH3) PS02

As pointed out earlier, no one has reported a thorough investigation of the sol- vent to ascertain what changes might have occurred. The normal workup by addition of water makes this somewhat difficult. Its hygroscopicity requires that care be taken if anhydrous conditions are desired. It is also known to react violently with some acyl halides and very strong acids.

It is certain that as our knowledge of DMSO increases so will its use and vice versa.

- 50 -

BIBLIOGRAPHY

1. A. Saytzeff, Ann., L44, l48 (1867).

2. "Technical Information on Dimethylsulf oxide", Crown Zellerbach Corporation, Camus, Washington.

3. D. F. O'Brien, University of" Illinois Organic Seminar, 2nd Semester, 1959-1960.

4. D. J. Cram, J. L. Mateos, F. Hanck, A. Langemann, K. R. Kopecky, W. D. Nielsen, and J. Allinger, J. Am. Chem. Soc. , 8l, 5774 (1959) and preceeding papers in the series.

5. D. J. Cram, B. Rickborn and G. R. Knox, J. Am. Chem. Soc, 82, 64l2 (i960) .

6. D. J. Cram, B. Rickborn, C. A. Kingsbury and P. Haberfield, J. Am- Chem Soc, 83, 3678 (1961).

7. D. J. Cram, M. R. V. Sahyun and G. R. Knox, J. Am. Chem. foe, 84, 1734 (1962) .

8. J. E. Ilofmann, R. J. Muller and A. Schriesheim, in press.

9. A. Schriesheim and C. A. Rove, J. Am. Chem. Soc, 84, 3160 (1962).

10. Mine Safety Applicance Company.

11. A. Schriesheim, R. J. Muller, C. A. Rove, Jr., J. Am. Chem. Soc, 84, 3164 (1962)

12. A. Schriesheim, C. A. Rove, Jr. and L. Easlund, J. Am. Chem. Soc, 85, 2111

( 1963) •

13. S. Bank, C. A. Rowe, Jr. and A. Schriesheim, J. Am. Chem. Soc, 85, 2115 (1963). Ik. H. C. Brown, J. H. Brewster and H. Shechter, J. Am. Chem. Soc, Jb, 467 (1951!)-

15. E. J. Corey, J. Am. Chem. Soc, 76, 175 (1954).

16. N. Kornblum, P. H. Berrigan and W. J. Lelfoble, J. Am. Chem. Soc, 85, 111*1 (1963)

17. N. Kornblum, P. H. Berrigan and W. J. LeWoble, J. Am. Chem. Soc, |§, 1257 (i960)

18. N. Kornblum, R. Seltzer and P. Haberfield, J. Am. Chem. Soc, 85, 1148 (1963).

19. G. A. Russell and E. Janzen, J. Am. Chem Soc, 84, 4153 (1962).

20. G. A. Russell, A. J. Moye and K. IJagpal, J. Am. Chem. Soc, Ok, kl^k (1962).

21. G. A. Russell, E. G. Janzen and E. T. Strom, J. Am. Chem. Soc, Ok, 4l55 (1962) ■

22. Private communication

23- V. J. Traynelis, W. J. Hergenr other, J. R. Livingstone and J. H. Valicenti,

J. Org. Chem., 27, 2377 (1962) .

2k. B. T, Gillis and P. E. Beck, J. Org. Chem., 28, 1308 (1963).

25. D. J. Cram, C. A. Kingsbury and B. Rickborn, J.. Am. Chem. Soc, 8^, 3688 (1961).

2o. A. J. Parker and Banthorpe, unpublished work from Reference 28.

27. H. L. Scha'fer and W. Schaffernicht, Angew. Chem., 72, 6l8 (i960).

28. A. J. Parker, Quart. Rev., lb, 163 (19&2).

29. E. J. Corey and M. Chaykovsky, J. Am. Chem. Soc, 84, 866 (19S2).

30. M. Chaykovsky and E. J. Corey, J. Org. Chem., 28, 254 (1963).

31. C. Walling and L. Bollyky, J. Org. Chem., 28, 256 (1963).

32. A. Cope, P. A. Trumbull and E. R. Trumbull, J. Am. Chem. Soc, 80, 2844 (1958). 33- R. Greenwald, M. Chaykovsky and E. J Corey, J. Org. Chem., 28, 1128 (1963).

34. R. Kuhn, Angew. Chem., 6g, 570 (1957).

35. R. Kuhn and H. Trischmann, Ann., 6ll, 117 (1958).

36. R. T. Major and H. J. Hess, J. Org. Chem., 2J>, 1563 (1958).

37. S. G. Smith and S. Winstein, Tet., J, 317 (1958).

38. E. J. Corey and M. Chaykovsky, J. Am. Chem. Soc, 84, 867 (1962).

39. E. J. Corey and M. Chaykovsky, Tet. Letters, 169 (1963).

40. R. RStz and 0. J. Sweeting, Tet. Letters, 529 (1963).

41. R. R&tz and 0. J. Sweeting, J. Org. Chem., 28, lol2 (1963).

42. N. Kornblum, J. W. Powers, G. J. Anderson, W. J. Jones, H. 0. Larson, 0. Leonard and W. M. Weaver, J. Am. Chem. Soc, 7^, 6562 (1957).

43. IT. Kornblum, W. J. Jones and G. J. Anderson, J. Am. Chem. Soc, 8l, 4-113 (1959).

44. S. Searles and H. R. Hays, J. Org. Chem., 2J5, 2028 (1958).

45. F. Ostermeyer and D. S. Tarbell, J. Am. Chem. Soc, 02, 3572 (i960) .

46. W. von E. Doering and A. H. Hoffmann, J. Am. Chem. Soc, 77, 521 (1955).

47. R. P. Bell, Proton in Chemistry. Cornell University Press, Ithaca, Hew York (1959), P- 87.

48. S. J. Cristol, J. R. Douglass and J. S. Meek, J. Am. Chem. Soc, 7J5, 8l6 (1951).

y

i