F OOT N OTE STO TABLE I

(a) Calculated by the method of Otozai, r-<ume, and Fukushima ( 1); values are for standard molecular

entropies in the gaseous state at 25°C. (b) Observed standard molecular entropies for the gaseous state at

zs • c .

Unless otherwise indicated. these values are taken from the tables of the American Petroleum Institute Re;;earch Project-44 (2). which appear to be the most reliable data to data.

(c) 5°.1 -s•b . (d) From measurements made at 25°C. unless otherwise indicated. (e) Cf. ref. (3). ( s:a c . o s ·~ ( ) f) The method ot McCullough~· 4 may be used to obtain calculated values of atandard molecular

entro1:-ics of 1-alkenes in the liquid state, which differ from e:q)erirnental quantities by only 0. 040/o. (g) Cf. ref. (5). (h) Mixture of ci3-and trans-isomers

( 6 ).

(i) At 18°C. (6). (j) Probable mixture

of isomers (7). (k) At l6°C. C/):-[rn) Ma;;lov (8) presents an alternative method of calculating

s •

for either the ~aseous or the liquid state of alkacliene3. (n) At saturation pressure. (p) Ap;;roximate

values for a mixture of cis-and trans-forms. (q) Cis-and trans-mixture; also undoubtedly contamin-ated with 1, 3-isomer {l:r;-cf. 15).

(r )

:Probably lo-;:-see footnote q. (s) Probably quite high (see footnotes q and r); the value calculated £rom group refractivities is 28. 9. (t) Pllrity uncertain { 16). (u) Cis-and b:·ans-mixture

(9 ).

{v) 1H ll°C. (9). {w) At l5°C. (9). (x) At 21°C. (9). {y) See ref.(TI). (z) ::iee ref. (11). (aa) At 20°C. {11). (bb) ')ea ref. (12). (cc) At

z o • c .

{12). (dd) Possibly contaminated with saturatetl material (13). {ee) At 21.4°C. (13). (ff) !;t 24.4°C.(l3).

(gg) The identity of this compound is still i!1 doubt.

Note: References to Tables I and II appear on page 34. (\.; ^• ..

28

comparison is made between the calculated and the observed values.

"'or the acyclic alkanes and al!<.enes, this met:-l.od _::>redicts entropies which are quite close to the obse.cveci. values, but consi stcntly low by 0. 2· 1. ·1 e. u.; the majority of ca3es show disagreement::~ of substar~ti.-

ally lesa than 1. 0 e. L\. The 0<1e exception falls in the case of trans-2• hexene, in which instance the calculated entr(}·py is high by 0. 5 e. u.

The method of Otozai, Ku1ne, an.Ci Fukushima, however, does not attempt to evaluate the effect oi cycle formali on or conjugation of multiple bonds on the overall entropy. Consequently, the differ-ences between the calculated and the observed values for the entropies of the cyclic systems and of the 1, 3-aU;adienes may be taken as a':>pro:d.matio•1s tc the corre G?Oncling entropies oi cyclization and of conjugation in a

given system . .According to the values in Table I, then, the loss b . entropy due to t_.t,.e formation o[ a formal six membered ring i3 about

17-21 e.u. , while restriction of rotation around a co;1jusated "sin;;:;!e"

bond of acyclic 1, 3-alkaclienea lead'; to an cntrony los:; of -t-5 e. u.

Unfortunately, entropy data for U.!lconju:?;atecl allca<iiene!5 are almost non-existent at present. In fact, the only diene in thi~3 category, whose entropy of formation has been determined, is 1, -1-pentadiene. It is exceedingly interesting that the observed entropy nf this compound

is lower than the Otozai, .I<:ume, and Fukushima prediction by '!:. 8 e. u. This value is 36-33% of tho maxirnun1 locs of entropy expected for the

29

formation of a tr ue ring with five member s. The paucity of data requires some restraint in attempting; to interpret this value. Upon conGideration of the magnitude and dh·0ction of error in the calculated entropies, however, the di screpancy of 4. 8 e. u. in this case rnust

surely represent a signific~nt restriction in the molecular motions of 1. 4-pentadiene. The extent of thi::; restriction is quantitatively similar to the degree of restriction to rotation, which e::d.sta about t-l-te "single"

'bond of conjugated dienes. It is imposaible, however, on the basis of the· entropy data alone, to determine qualitatively the nature of this restricted motion. Since there is no evidence for intermolecular associ- ation of the alkenes (i.e. all the deviations fron~ the predicted entropy values are negative), there i:::; no reason to assume that l, <1-pentadiene should possess to any great extcmt a polymeric gas structure. It is likely, t..lJ.erefore, the restriction to motion is an internal one.

Two models which might account for the observec low entropy of 1,4-pentacliene can be suggested. The first oi the:;e is based on the familiar homoallylic system (64). In this model, the C-2 and C-4 atoms of the linear diene (IXa) are considered to be .;;lightly closer to one another than is the ca -.;;e in the corre ::>pending saturated or partially

saturated chain. It is presmr'Jed that the loss iLl entrory in this model would be balanced by a gain in resonance energy afforded by overlap of the p-orbitals on C -2 and C -4:. Due to the ~.~resence of the vinyl hydrogena at C-2 and C -4, conformation IXa,b would be expected to

30

be twi~tcd somewhat from planarity in order to OJ,-timize the resonance ar!d repulsive interactioru:;.

H H

~ ^{+~- H H}

IX a IXb

The second model assumeg that the cyclic conformation is itselZ somehow stabilized relative to the extended conformer by an interaction between the double bonds. Fixation of the fo!Qed conforrner might be attributed to either a charge transfer mechanisr1J, resultin~

in a coulombic attraction between the double bonds {X), or a delocaliza- tion mechanism (XIa, b c) not unlike the one deCJcribed for the first model. Cf these, the latter mechanism appears to be the more reasonable. The terminal hydrogens of XI would require that the double bonds be in approximately parallel planes rather than in the

same ::-lane in order that their mutual overlap be appreciable. Resonance of t..lle type Xla should be more important than that of Xlc on the basis of overlap alone. Both of theBe mechanism::; "vill be co:>sidered more quantitatively for the particular case of a 1, 5-diene in a later section.

X

31

( ) ( )

XI a XIb XIc

It should be emphasized t},.at the entropy data of Table I are for the ga secus state,* and consequently, do not r;ive any information on the condition of molecules in the liquid ;?hase. The higher dielectric constant usually a ssociated with the condensed phase woP .. ld, however, be expected to be rno::.·e conduci· . .,.e to formation of rnore polar s;:>ecie.:;

such as IX, X, and XI.

Molecular Volurr,es. The extended conforr!lation of a hydro- carbon chain will possess a greater molecular volume than any folded

species (e. g. cornpare the ism-rJers cyclopentane and 1-pentene in Table I). This simple fact suggests a second means of determining the presence of folded conformers in solutions of alkadienes. The average molecular volumes of a aumber of hydrocaroons are given in

Table I.· The first section of Table II draw3 a comparison between the rnolecular volume£> oi sonle normal alh:anes and the corresponding 1-alkenes. It is seen that the introduction of a terminal double bond decreases the molecular ·.-olurne by about 5. 7 :.:nl. / mole, but that thi:i

*

^{Data for the liquid state under corn}pa:-able conditions was not available in ma.ny cases of i:1tercst.

Table II Olefin a ~ db Re:::erence C ;mpoun A (Mol. Vol.) c A(Mol. Ref.) c 1 U-1 I-1 5.720

0.418

2 II-Z I-2

5.70 6

0.424 3 U-7 I-3 5. 712

0 . 42 1

4

ll-10

I-4 5.68

0 . 42

t• :; 11-13 I-5

5 . 6 2 9 0 . 42 6

6 II-16

I-6 5 . 575 0 . 425

w N mean

5 . 670

rnean 0. ·122 7 UI-7

U - 1 6 . 506 0. 5 01

8 lll-9d 11-5 5. l 0.4

9

UI-10 II-6

7 . 97 0 . 5 1 10 m - s

d 11-3 6. 04

0. 1 6

11

lll - 8d

II-'~

7.60

0 ?'" • ..,!) 12 III-lld ll-3

---- 0 . 46

l3 III-lld II-9

--- - o .

59 14 III-l2d II -11

--- -

0.52 15

lll-l.i~d

II -12

-- -- 0 . 6 5

16 III-13d II-Hd

- - - - 0 . 6 4

mean 6.6~

Olefin a 17 18 III-16 19 III-17 20 III-18e 21 III-198 22 ill -20d

23

III -21 d 2~ III -22d

Table U (cont.) b Reference Compound ll-2 II -7 Il-5 r

n

-14: .. 1I-1·1i II -10

ll - 13

II-16

fl(Mol. Vol.)c 6. 472 5.61!

~(Mol. Ref. )c 0. 494· 0. '184 0.480

(o .

85)g (0. 50)~ 0.44 0 ..,~­ • c. :J 0.36 Footnotes to Table II: (a) Numbers refer to the compounds listed in Table I; {b) The reference compound is chosen, except where noted. to differ from the olefin in the first column only by the absence of the tel'minal double bond; (c) t.

=

Ref. C pd. -Olefin (i.e. decrements due to the precience of the new double bond are given as positiv~ value:::); (d) See footnote ::1 appropriate to the compound or its constants given in Table I; {e) No terminal double bond; (f) Data for .~-nonene (5) suggest it is too impure to use for comparison (Mol. Ref. calc. 43. 30, found 44. 78); (g) This value is included for completeness, although it is not possible to assess its sic;nificance with ·.fmch certainty. A decrement of more than 0. 69 in the molar refractivity is probably rDeaningful in the case of a cis-configuration, while the value o( 0. 58 seer.."ls to be a reasonable upper limit in

~ - sys te~s

( se-;-;;ef. 2).

y) (JJ

REFERENCES FOR TABLES I AND II 1. K. Otoza.i, S. Kume, and S. Fukushima, Bull. Chern. Soc. Jap., ~· 302 (1952). 2. F. D. Rossini et al., "Selected Values of Physical and Thermodynamic Properties of Hydrocarbons a~elated Compounds," American Petroleum Institute Research Project- 44, Carnegie P:;.-ess, Pittsburgh, Penn., 1953. 3. J. G. M. Brenner and G. D. ThomaEJ, Trans. Faraday Soc.,~· 779 (1947). 4. J.P. McCul1oughetal., J. Phys. Chem.,

g,

28')(1957). 5. :tv1. Bourguel, Bull. soc. chh:n. !-:'"'r., {4:),

!!.•

1475 {1927).

6.

:M. V. Cox, Bull. soc. chim. Fr., (4),

!?_ .

1551 (1925). 7. A .• Kirrmann, Cornpt. rend.,

.!!!•

1463 {1927); ~· Ann. dum. {10),

.!..!.. •

277 (192.9). 8. P. G. Maslov, Zhur. Fiz. Khim.,

.!.!. •

1242 (1957). 9. 0. Riebe, Arm. clrlm., (12), 11, 593 {1949); Chern . .P1b3tr.,

.±!:_ ,

2984·c (1950). 10. D. D. Coffznan and E. L. Jenner, J. Am. Chem. 3oc., ~· 2872 (1958). 11. ·A. L. Henne and K. W. Greenlee, J. Am. Chern. Soc., 65, 2020 (1943). 12. H. Pines and N. E. Hoffman, J, Am. Chern. 3oc.,

7.!:_.

4417 (1954). 13. J. L. Everett and G. A. R. Kon, J. Chern. Soc., 3131 (1950). 14·. B. H. Shoemaker and C. E. Boord, J. A.m. Chem.,

.?l •

1505 (1931). 15. C. Prevost, Bull. soc. chim. Fr., (5),

.!!•

89 (1941). 16. G. Vfiegt, Ph. D. Thesis, Ohio State University, Columbus, Ohio (1940).

w ,;.:..

35

value decreases as the length of the chain increa.oes. ? resumably the introduction of a second ter::.J.1inal double bond :>hould bring about a similar reduction in molecular size, if no other changes occur. In 1,4-alh:.adienes, however, the inclusion of a >3econd double bon3.. in

the terrninal position results in an avera0e rnolecular volume decrease of 6. 6 ml . /mole, which is 17% greater than anticipated. The one exception is found in the case of 2-methyl-1,4-pentadiene. The purity of the compound upon which the measurer.nent was rnade is uncertain in this instance. The second double bond of 1, 5-hexadiene produceD a 14·o/e larger volume decrement:: than does the fir st. The decr.ease associated with each of: tb.e bo:1.d:; of 1, 9-decadiene is the :'lame, 3ug- gesting that no special conformation advantage is provided by the presence of the second multi-,:le bond in the lon-ser chain.*

The data available do not allow the calcul2.tion of. the r:noleculn.r volumes of all the dienes at the same temperature. However, molar refractivities, which are temperature independent, may be obtained for all the hydrocarbons of Table I. The introduccion of a double boi:lci produces a net decrease in the rnolar refracdvity, since the two

•

hydrogen atoms removed have a slightly larger refracdvily than the carbon-carbon double bond. The chnn;~e as :::;oda.ted ·.;..:it..h the

*

^{It should be noted that the volu::-.'1e} contractions observed may simply reflect increa.:Jed interrr:olecular att:;:-active forces which

accompany an increase in molecular polarities. If thig were the case, then longer chains should al3o sl:.ow slight deviations.

36

incorporation of one terminal double bond is about 0. 42.. In almost aU of the 1, 4· , 1, 5-, and 1, 6-alltadienes of Table I, the change in molar refractivity is :;ignificantly lar;.~er tha.n this value. This is indicative of an abnormal change in the polarizability of the diene s, relative to the 1-alkcnea, which r:night possibly be related to a shift in the conformational equilibrium in favor of the folded conformer.

Viscosities. Viscosity measurements may be used to obtain the ldnetic theory collision area of a molecule in the gas or liquid phase {65). Since the collision area depends on the conformation of the molecule, viscosity data for the hydrocarbon::~ of Tabl<:;! would be expected to be very useful

in

establishing the conformational nature of alkadienes. Although the viscositie::~ of a large number of hydrocarbons have been rnea ::;ured, the alkadiene'::> have been alnlO:Jt completely

ignored. Cummings, McCoubrey and Ubbelohde (('6), however, have

1.rnea3ured the viscositie3 of n-he~:ane, 1-hexene, an•.i !, ~-he:~adiene

in the ga.a ·::.hase, ancl·have concluded that the proportion oi folde0.

conformers increases in the order r.-1entioned at a given ternperature ..

More dat&. a:re r1eeded before it will be possible to rr1ake any useful generalizations about the conformation of alkadienes. At present, however, t..'here is a strong suggeotion that certain diolefins

r.t1ay show~ 3reater preference for the more cornpact, folded conforma-

tions than do their partially or totally saturated analogues. It is likely

3'7

that the phenomenon may be explained in part by the £act thc.t the introduction of a terminal double bond has the ~£feet of lowering the energy barrier to rotation about the adjacent C -C bond relative to the paraffin (67). The concomitant increase in rotational freeclom . how- ever, should lead to a relative increase in the molecular entropy, which is not observed. (see Table I). Also, it is difficult to see why the longer chain alkadienes (e.g. 1, 9-decadien~) ar~ not more folded relative to the parent paraffins. if more facile rotations around C -C bonds were the only important conformational effects as::>ociated with the intro- duction of double bonds. It was proposed, therefore, th.at another mechanism(s). involving an electronic interaction of the double bond:J, might be important in deter rP.i:-ring the conformation of diolefinic chains of. intermediate len~th.

Resonance and Conformation in Non-Conjugated Alkadienes

It has been suggested (pp. 29-31) that the foldeci conformation of certain non-conjugated alkadienes might be stabilized by an electron delocalization mechanisrn arising from a non-z~ro overlap of the

classically isolated r. -system::J. The nature of the overlap woulc differ for the chair {Fig. 3) and boat (Fig. 5) folding of the diene.

Chair Conform.:;.:r. In the chair conformation of a 1, 5-alkacliene, only overlap between the p-orbitals on C- 1 and C-6 (i!""'ig. 3) is extensive. These orbitals are almost coa~dal, and can be made exactly so with only

38

a very small deformation (less than 5°) of the C-C-C bond an::.;le at C -3 or C -4. In this co;.1figuration of the diene, the approach of C -1 and C -6 is favored by the fact that there is no rnutual eclipsing of the hydrogen atoms at these centers. However, folding of the chain in this fashion produces two eclipsed interactions, which are avoided in a linear conformer; namely, the vinyl hydrogens at C -2 and C -5 are brought into opposition with their neighbors on C -3 and C -4, producing a relatively·

weak repulsive interaction due to the 120° angle at the double bond.

In order to assess the importance of 6e p-p overlap in stabiliz- ing the chair conformer, ::.;imple molecular orbital calculations based en the Hlickel approximation were carried out for a number of rotarneric forms leading to the lin1iting chair configuration. These forms are obtained from the conforrner of Itigure 3 by simple rotations of C -5 around the C -C bond in eithe:r a clockwise or a counterclockwise

3 4

manner. Rotation in either direction by 180° will give a particular linear conformation of the diene. Frecise measurernents of angles and distances were made from accurately scaled (Dreiciing) molecular rr1odels. These parameters were used to obtain values for the cr - and n-overlap integrals from standard tables (68). -....<;hich in turn were used to calculate the required resonance integrals (69). The resulting

resonance energy (R. E.) of these rotamers is given in Figure 4 a"' a function of the C

1

-c

₆ internuclear distance.

Q

^{. ' .}

^.

I• I

,

^~

39

0 .--:-.- ^.

_'

.

.

.

.

\ ·,I '

\

0.8 0.6 l

" , Rl(~)

o.4 0.2 o.o 1.5 2.0

' '

\ \ \ \

' ' ' '

....

' 2.5 3·0 r(i) ___. Fig. 4

... ____ ... _ 3·5 4.0 Resonance energies calculated for rotameric conformations of 1,5-hexadiene related to the chair form shown

in

Fig. 3 "(see text).

,j:. 0

The s;ecular equation associated with the conformation shown in Figure 3 and with t:hoae arising ~rom a ccmnterclockwi3e rot.."ttion of the residue at C -5 :c-esernble;.; that .::;f 1, 3-~u.tadicne, and consequently the small resonance energy is expected (solid line). The function is ma}dmized when the C

1

-c

6 ciL;tance cm:re-:;ronds roughly to that of a C -C single bond, since the overlar:.' is :rirnarily cr ir:. character.

If the rotation is performed in the clock•.viae sense, the p-orbitn.l at C -5 begins to overlap increasingly with that o.Z C -1, while the orbital on C-6 becomes leas involved in overlap. The relative stabilization (dotted line. Fig. 4) dec1·eases sharply due to the appearance of n.

nodal plane between C -1 a:::~d C -5 in the highe!:lt bonding molecular orbital; the resonance energy reaches a minir.num when the overlap of

C-5 and C-6 with C-1 are about equal (ca. 2.1

A.).

As r otation in this direction ia continued, bonding between C -1 and C -5 becomes more important and antibonding between C -1 and C -6 becomes less important, resulting in an increa~:;e in the del·.Jcalization energy. A

second maxin1un1 is reached at a distance of about 2. 4 /.,. , which co:;.·-

0

resnond3 to a C

1-C .. internuclear se·;aration r;i. about 1. 9 A. The

.

^~

resonance energy a~sociated with thia 1.-naximurn may be increased by allowing for some angle deformation in the model, but the total

energy of the five memberec! ring conformation would still be le3:5 tha:.1 that of the unatrained six mernbered rL>g. It may also be rJointed out that when the overlap between C - 1 and C-5 i~; maximized, the

42

confor n1a tion of the chain is analozous to the "envelope" form of cyclo?entane, and the C -3 and C -4: rnethylene groups are eclipsed as also are the groups on C -1 and C ·5. These c::msidera.tions are in

agreement with the fact that rings with five member;; are usually minor products when 1, 5-dienes are cyclized {vide infra).

Boat ConfoL"mer. :ivfolecular model9 of t.he bon.t confvrn>ation o~

a 1, 5-diene indicate t:hat a-overlap between the F-orbitals on C -l and C -5 is almost as extensive as that between C -1 ancl C -6 (Fig. 5).

--

A" "' long as the overlap betv;een these pairs of centers is equal, the resonance energy predicted by the simple Hlicl.cel-type treatment is zero; that is, the bonding o1·bitals of the ethylene units are symmetrically split, such that there is no net increase in t..'h.e delocalization energy.* However, the boat folding of a 1, 5 -diene is fully eclipsed, anc'.. for t..hi s rea son alone would not be expected to be a stable conformation. Rotations in either sense around the C,, -C. axis should relieve the steric com-

.) 4

pressions due to the eclipsed groups, and at the same time w.::)Uld change the nature of the overlap in the system.

For very small counterclockwise rotation3 the a overlap

between the pairs of centers remains essentially equal, ancl there is no gain in delocalization ene:r~:w· For larger rotations (solid line, Fig. 6),

*

^{For the special case when all resonance integrals for nei:,;hbor-}

ing atoms are equal to unity, the secular determinant is the same ag that obtained for cyclobutadienc, and there is only one bonding orbital;

the otl:'ler orbital having become non -bonding an•:~ clege:nerate.

43

.... w

..

0.8 0.6 r

RE(~)

0.4 0.2 o.o 1.5

, , , _, ^2.0

I

,

I I I

--- ... , ' ' ' ' ' ' ' ' 2.5

~.o

r(i) --+ Fig. 6

' ' ^{' '}

....

...

~·5

4.0 Resonance energies calculated for rotameric c;onfontationa of 1,5-hexa.dime related to the boat fol'll shown

in

l'ig. 5 (see text).

*"' *"'

·15

however, the ;,:--orbitals at C- 1 and C-Z both overlap extensively with that of C -6 as also do those at C-5 and C -6 with the orbital on C-2. The resultinG secular determinant rcgembles that cbtained f:>r bicyclo- butadiene. Larger increases in re sonai1CC cn.er~y are predicted a s overlap between C -2 and C -6 predominates. Continued rotation

diminishes the overlc::.p and the opportunity for electron delocalization, and the resonance energy r eturns to ;;;;ero when C-1 and C-6 are

0

separated by about 3. 7 A. This is illustrated by the solid line in Figure 6.

l\iluch the s;:une behavior i ~J observed when clockwi se rotations are considered (dottecl li11e in Fig. 6). The broadness of the curveg in Figure

o

iG due to the fact that ,>mall rotations, which do not result in large changes in the overlai-, value·3, give rise to relatively large separation between C -1 and C -6.

The P ossibility of lntran1olecular Complex Forr.1at.ion. That other- wise stable specie3, ._..,hen suitably paired, give rL'>e to molecular

complexes is a widely studied phenomenon, which has been compre- hensively r·eviewed by a number of author s {70). A v-u iety of mechan- isms for complex formation have been proposed, with the Mullikin theory of donor -acceptor interaction {71) apoearing at present to be the most widely accepted. In vtew of this, =-~•ention should be made o£ the possibility .for confo:;:-mational fixation via a mutual (i.e. intra- molecular) complexing of the olefinic centers of the 1,5-dienc. If