MINIMUM? 'IHNHHWE ‘ > < Hi ‘ Will —lo_, 324M» H (ID—x0) THE FREPARATBGN AND 155.3%..3‘1TSQN OF THE GAMMA {SG‘fi'XER OF 1, 2;, 3, 4, 5, 6, HliXACi-QLC‘RCCJ‘!{LUDHEXANE Thesis far fin magma of [331. D. MACHKSAH STATE {liikifhEGE Ranger Lager; Sewr 47, 1-1; ‘:.:-1 .' .« L: M434 F. ,t '1' '7 A 3f " ‘ 7 9 .Y. PLACE IN RETURN BOX to remove this checkout from your record. To AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. . DATE DUE DATE DUE DATE DUE 001 c-Jcmcmmouo.pes-p.1s THE PREARAJI‘ION AND ISOLATION OF THE (mm 130m 01" 1,2,3,4,5,6 HEXACHLOROCYCLOHEXANE By ROGER LR) BAUR A 111313 Submitted to the School of Graduate Studies of nehigcn State College of Agricultnro'end.appliod Science in pox-6111 fulfillnent of the requirements for the degree of DOCTOROFPHIWOPHY Departmt of Chemistry 1949 ACKNGYLEDGMEM‘ The ”thor '1.h” to thuk me De To M“, for hie eid end guidance, without which thie work would not hove been poeeible. The euthor expreeeee hie eppreoietion to Dr. R. C. Hueton end Mr. J. F. Lee Veeux for their eld in obteinlng the Niegere Chenioel Divieion Fellowship for the yeer 1948, end e11 other eteff numbere end friends for the may kindneeeee they have ehovm. tttttttt ##tllt fittfi it . r‘. o q» T, ”14;“. q » " ' ‘.,-l w MW“ .4 5’ .. .1“): Ha TABLE OF CONTENTS Page Intr°m¢t1MOeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee 1 Theorwicueeeeeeeeeeeeeeeeeeeeeeeeeeeoeeeeeeeeeeeee 2 Chemical...onnun....o....o.u.............u... 8 Apmr‘t‘uleeeeeeeeeeeeeoeeeeeeeeeeeeeeeeeeeeeeeeeeeee 8 mly'heeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee 9 m3w..10neeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee 16 RM“ Of ada'eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee 16 Pfepu‘utlon Of heanhlorOOyOIOthOeeeeeeeeeee 18 saw‘tim 0f the 18011011.“...NHU"no..." 22 mltilhtion Mhfis 0f .mm‘ttoneeeeeeeeeeeee 22 Fnctim distillation"..."nun”..u 22 at“ dlltilhtion...u“an"....nu... 23 Chrmtogrtphyoo..n.on"...u................ 27 WWOOOOOOOOOOOOOOOOOOOOOOOOCOOOO00.000.00.000... 42 Bibliogr‘Plvooeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeo 43 ‘ C C D 0 O V . . . . INTRODUCTION 1, z, 3. 4, 6, 6 hexechlorocyclohenne, eleo known es benzene hexechlcride end 666, wee first nede by Feredey (l), in 1852. It wee not until 1887 thet Meunier (2), showed the presence of two isomers, the elphe isomer, melting et 157°. end bete isomer, herring e Inch higher melting point. Then in 1912 Ven der Linden (8) proved thet there were et leeet four iema's of this compound: ape. mp. 153° bets. mp. over 200° gem n.p. lOB°-lll° delte mp. 129°-132° In 194? the fifth isomer, epsilon, wee sepereted by Ieuer, DnVell, end Alquist (9). The melting point 1. zoo-227°C (eublined). Pure elphe end bete iecner eemples were prepared end their toxicity to insects wee investigeted in 1942 (3). When the gum isomer wee iso- leted in 1943 end its insecticidel properties eetiueted (3), it wee found to be more toxic to weevils then eny subetence heretofore tried. It wee subsequently verified thet the insecticidal property of bensene hemechlcride wee due elnost entirely to the gem isomer. This isomer hes since become e mjor insecticide. The purpose of this investigetion wee to study the effect of weve- lengthe of light on the fornetion of the isomers of hensene heeechlo- ride, end to investigate different methods of seperetion of the isomers. The odor rencvel frm the comerciel product wee else studied. 91- THEORH‘ICAL Addition of chlorine to benzene W. A. Noyes, Jr. end coworkers in studying the photochenicel eddi- tion of chlorine to benzene (4), heve found thet the edditicn rete et the beginning of the reection was neerly proportionel to the square root of the light intgisity, end proportionel to the pressures of chlo- rine end benzene. Vsry sss11 enounts of chlorine substitution products were found et reection tenperetures between 25 end 35°. with e deficiency of chlorine. With en excess of chlorine the reection geve dodecechloro- cyclohexene (c3c112). The eddition reection nechenisn.poetuleted by Noyes end coworkers wee: (e) 012+h9 a c1.c1* (b) oi‘+c12+c¢H6 s c636c12+c1 (c) CSH‘BCIz-pza2 : 0353015 Activeted c1“ furnished the ectivetion energy for the eddition re- ection (b). The compound 6536612 wee considered to be the first eddi- tion compound, end there wee sue evidence thet the chlorine edded'to this without further ectivetion to give hexechlorocyclohexene. Reection (b) es it stends norms that e three-body collision met teke plece. However, it is to be remembered that en internediete met heve e defi- nite life-span, so the reection could eleo be written: (b-l) 012+c1* g Ci3 0134-6686 = Csfieclz-ICI or (b-2) 01*tcsns : censci c636c1+c12 . 66H6012+01, . h. .“l—m . , e .i... . mu. _.__. ._ __.__ either reection being the equinlent of reection (b). Chronnt cgrefl In 1906, Micheel Tswett introduced e method with which he studied the green pigments of Issue. Although his experinaits did not result in the isoletion of the pure substencee. Tswott insisted on the greet inportence of his discovery. Very few peid this inportent new discov- ery the ettention it werrented, pertly beceuse his comprehensive book, 'Chronophylls in Plent end Aninel World“, published in 1910, bed only eppeered in Russien. The period free 1906 to 1931 cen be regarded es the letent period of chrontogrephy. In 1931 Kuhn end Ledsrer succeeded in eepereting elphe end bete cerotene. nking known tremendous possibilities of chrome- togrephy. In the period from 1936 to 1946 over one thousend pepere heve been published on chrontogrephic experinente. The theoreticel foundetions 'of chrontogrephy were elreedy recog- nised by Tewett. Though it cen be e function of mny eeperete proces- ses. chromtogrephy is prinerily due to selective edscrption of the con- ponents on the edeorbent, es centrested with their difference in solu- bility in the solvent used. . . For simplicity, essune e solution in which there ere two components, eech soluble to the sens degree. If such e solution in en inert solvent is poured into en edeorbent column, the solute will be edeorbed end re- noved from solution. if the edeorptive ettrection betweai the column end the couponents is of sufficient ssgnituds. Should tn. ettrection b. weeker then thet required to hold one conponsnt. thet component will 4"“? we- -.——._ " “- «‘. . .'_I". *tu.-_H_F___:Hfia __.._ _ peso through the column with the solvent, leeving the lore strongly edeorbed component on the column. It sometime heppens thet both can-- pcnents ere strongly edeorbed. In this one the difference in ettrec- tion of the two solutes towerd the column ceuses then to 'bend'uthet is, the two components ere sepereted. one being ebove‘the other on the colunn. What the nixture.firet cones in contect with the edsorbent, both solutes ere rnoved free solution. As we solution cones in, the nore strongly edeorbed component "descrbs' or displeces the less strong- ly edeorbed conponent. It moves doen the column to be edeorbed et e plece free from solutes. This process continues until ell of the stronger component hes been edeorbed, thereby displecing the weeker which moves to e plece lower on the colum.‘ It met be kept in nind thet such system ere not stetic. but dyne- nic. The ebove systems wee chosen with e neutrel solventu-cne which is not effectively edeorbed by the colunn. Actually, this is never the one. for the solvent list "wot“ the calm in order for the solute to cells in contect with it. It is nors e utter of degree of edecrption then mother it is edeorbed or’ not. Ln equilibriun is set up betwem the enount of the solvent end solute edeorbed on the column, end the quentity cf eech reneining in solution. The edeorbent contests with the solvent, end the reletive snout of the solute edeorbed depends on its ettrection to the solute conpered to its solubility, or its ettrec- tion to the solvent. That the solvent tends to be edeorbed lessens the ettrection of the solute towerd the calm. It cen be et once sssn thet even e sinple edsorption en en inert cclunn becones e phenomenon -4- _,‘... 4:*— v of new seperete processes, eech with its on cherecteristios; e11 of these edd up to give the totel effect cell chrontcgrephy. Just es s solvent will dissolve e limited emount of solute. so en edeorbent will edsorb e given enount of edsorbete. However, edsorption end solution ere opposite phenomem. Solution may be regerded es the eveporetion of e solute into e solvent until equilibrium is reeched be- tween the solute end the undiseclved meteriel. Adsorption is the con- densetion of e solute onto en edeorbent until equilibriun is reeched between the dissolved nteriel end condused solute. The processes in eech probebly ere nore or less the same. It is e utter of definition whether the edsorption, or solution for thet fitter, is physical or chem- icel. Just es s complex bet-semi solvent end solute mist fore in order thet the solute dissolve, so met e couple: form betwou the edeorbent end edsorbete in order thet the edsorbete be edeorbed. There ere two fectcrs which prinrily bring ebcut edsorption end solution; chemicel sinilerity end spetiel effects. Eech hes its seperete contribution. Spetiel effects mostly concern ionic end moleculer sise. end dipole mo- ment of molecules. sise is not 'importent in solutions except perhepe for very lerge differences between solute end solvent molecules. Dipole mounts ere inpcrtent. however. Bthencl dissolves in heavens due to chenicel sinilerity, end in weter probebly beceuse of both sinilerity end dipole moments. However, henne end weter ere neerly immiscible beceuse of diesinilerity of both. For the seme reesons. chercoel is en excellent edeorbent of cerbcn compounds, but not for nter, while silice gel is e good edeorbent of nter but not very setisfectory for cerbcn -5- compounds of low dipole moment. But silice gel edsorbs othenol, prob- ebly due to both dipole moment, end chemicel similerity through the hydroxyl group. The reletive forces of edsorption end solution lergely determine which wey end how fer the equilibrium for the solute will go. The reletive emount edeorbed depends upon the neture’of the complex formed end upon how strongly the substence is edeorbed. The type of com- plex depends both upon the structure of the edsorbete end upon the force ’with which it is held. Bob is somewhet interdependent. The structure is importent in two differmit weye; if it is e syimeetrioel structure it will ellow the molecules to epproech eech other more closely, thereby. increesing the emount of substence edeorbed. But if the structure of the edsorbete is such thet it eetiefies the ettrestive needs of the ed- sorbut, then the emount edeorbed is less. However, if these needs ere not completely setisfied by the «me enount of substence. then more us- teriel is edeorbed. Probebly both prinry end eecondery forces pley e pert, the nin difference being in the distence through which they ect. Thus , en edeorbed substence my be ney molecules deep, the inner mole- cules held by prinry velence forces end the ones towerd the outside of the edeorbed eree held by the eecondery velence forces of both the ed- eorbent end edsorbete. The reletive structurel symmetries betwem the edeorbent end edsorbete eleo essume en importent pert. The more sym- metricelly similer they ere the closer do their molecules epprcech eech other. thus bringing the prinry velence forces into the picture. In turn, this better orders the configuretion of the outside of the leyer, mining it eesier for the edsorbete molecules to nice use of prinry forces. Thus the emount edeorbed is greeter. -6- moleculer dipole forces probebly operete through e greeter dis- tence then.primery velence effects. Their purpose mey be to bring ebout close enough proximity of stone of different molecules so thet primery end eecondery velence forces cen pley en importent pert. The closeness of epprcech is still e function of reletive symetry, the dipole force merely increeeing the chence of the two molecules coming together. It must still be remembered thet edsorption is en equili- brium reection. Even if both molecules ere very dissimiler, the dipole force elonetney be of sufficient strength to sense epprecieble edsorp- tion, if the solute is not selectively edeorbed, end if the solvent- edsorbete ettrection is not too greet. -7. CHEMICAIB Acetone from J. '1‘. Beker or Merck end 00., wee used without purifice- tion, conforming to A.C.S. stenderds. Alumina. from Aluminum Compeny of Americe, 8-20 grede, 80-200 mesh, we used without ectiveting. Benzene, e Merck end Co. product, Reeth grede, wee used without purificetion. f Benzene Hezechloride, e comerciel semple, wee furnished by Niegere Chemicel Division, Food Lechinery end Chenicel Corp., Middleport, N. Y. Cerbon Disulfide, Baker's Anelys ed, conteined 0.0006% non-voletile neteriel, 0.00% sulfides end sulfur, less then 0.002% sul- fites end sulfetes. B.P. 46-47 C. _ Alpln, bete, delte, end gem isomers of benzene henchleride wee fur- nished by Niegere Chemicel Division, Food lbchinery end Chemicel Corp., Middleport, N. Y. Petroleum end Nepthe Skelleysolve B, from Shelley Oil 00., wee used without purificetion. APPARATUS Becknen Infrered Spectrophotometer, Model 13-2, edjusted to zsto.4°c. by weter. It wee used with sodium chloride optics end cell. Cell thick- ness wee epprorxinetely 0.28 mm -3- ANALYBIS Ileny nethods of enelysis here been edvenoed for the isomers, sons specific for the genme isomer elene, es the gem isomer ection on noe- quito leme (11), e deohlorieetion procedure (12), e polerogrephic nethod (13), end e cryoscopio nethod (14). The only nethods of estim- tion of ell the isoners ere methods of frectionel crystellisetion such es set forth by Sleds (3), pertition chrontogreplw (10) (17), end infre- red enelysis (15) (16). The method of Deesch (15) us used by the euthor. The nethod wes scourete to 0.001 g. iscner in 10 co. cerbcn disulfide solution. A single cell wes used throughout, the thickness being 0.28 .1. es oelculeted from cell thickness end extinction velnes of the iso- ners 511.1 in the work by Deesch. - As there wens nny solutions to be enelysed, the euthor set up nono- grephs for the elphe, delte, end gen isaers in cerbcn disultide solu- tion. Bete isoner wes ignored homes or its insolubility, end epsilon isomer wes ignored beceuse its concentretion wes low, end ‘beoeuse no pure isoeer wee eveileble. As eech issuer hes sons ebsorption et the elphe (12.64), delte (13.22 1.1) end gen (14.53 m) enelyticel wevelengths, the equetions follow! Mt) s(t)+D(t) s(d)+6(t) 8(8) : Kt) :1: 12-64 m A (d) g(e)4D(d) g(d)+G(d) 3(3) 3 E(d) et 13.22 m A (s) s(t)+D(s) 3(d)+G(s) 3(5) 8 3(3) tt 14-53 In where L(elphe) g A(e) : slope at elphe isoleer et 12.64 m, -9- A(delte) 'g A(d) . slope of elphe isomer et 13.22 m, . A(geml) : .(g) .-'- ' ' . .1: 14.53 m, D(e1phe) g D(e) g delte et 12.64 m, D(de1te) :- D(d) u " " " et 13.22 nu, Mom-L) a 0(3) e " " " .1: 14.53 m, G(.1phn) n G(.) . " gem " .1: 12.61 m, G(de1te) . G(d) . * " " .1: 13.22 m. G(geuee) . G(g) : " n " .1: 14.55 m, g(e1phe) : g(e) g weight of elphe isomer in senple, 30101151) : g(d) : " " delte " " ' 5(M)=s(s). “ " su- " " " E(elphe) : E(e) g extinction et 12.64 m, E(de1te) . 3(a) . " " 13.22 .1, MW) : 3(8) : " “ 14.53 m. ‘ By inspection of the working curves (Figures 1, 2, end 3), A(e1phe), A(de1te), A(genn), D(elphe), etc. cen be found. So there ere three equetions which cen be used for celculeting the weights of the isours in solution, if the extinctions et the three nvelengths ere known. These working curves ere useeb1e with only one ebsorption cell thick- ness. Then, solving for g(e1phe), g(de1te), end g(gem), Eu) Du) GM E(d) D(d) G(d) 3mm) . 3(5) 0g) cm A ~10- 0.8 0.7 O e U' Extinction O a. 0 e 03 0.2 0.1 0.0 / / B ,. . / c / 6 1 2 18 24 30 36 42 Concentration (g./1iter) FIGURE 1 Working curve of alpha isomer 12.64 mu. 08 0.28 mm. cell 2.08 mm. alit A- Alpha I sonar 8- Gamma Iaomer C- Delta Iaomer Extinction 0.8 0.7 0.6 0.5 0.4 C e 03 0 e N 0.1 0.0 r X / / / M”? l 12 18 24 30 36 Cone entreti on (g./litar) FIGURE 2 Working curve of delta. iaomer 13.22 um. 08 0.28 mm. cel 2.55 mm. slit A- Delta Isomer 3- Alpha laomer c- Gemme Isomer Extinction 0.8 /)5 307 _ ___J -_ —1 A 0.6 :~—-——1 0.5 Oe4 _<,__ ‘M i i ! 0.3 i 0.2 0e]. 3 //V / r—rjf"””d 0.0 . ‘ 0 6 .12 18 24 30 36 42 Concentration (g./liter) FIGURE 3 Working curve of gamma isomer 14.53 mu. 082 0.28 mm. cell 1.55 mm. slit A- Gamma Iaomer B— Delte Iaomer 0- Alpha Iaomer Extinction 0.7 0 / 0.6 A/ / 0.4 / 0.3 / / / /% / / /. 0 6 12 18 24 30 36 42 Concentration (g./1iter) o o [1. a FIGURE 4 working curve of beta iaomer 13.46 mu. Acetone 0.28 mm. cell 1.01 mm. slit A- Beta Iaomer 3- Delta Isomer 0- Alpha Iacmer D- Gamma Iaomer Me) E(e) G(e) 1(a) 3(a) 0(a)) g(delta) : A9) 8:5) G(g) Ma) D(a) 3(a) 1(a) 0(a) 3(a) g(ge.mm.) .-_ filmy 3(6) A. Mt) 13(5) G“) A = Md) D00 000 11(3) N5) 6(5) Simplifying further, 301) = BOA G(g)D(d)-D(g)G(dl)-2£IL(E(d)G(s)-G(d)E(s))+G(t)(§£d)D£s)-D(d)E(g)) A 3(a) = AQXELQGBFGMLEQ))-Kt)%@Gia)-G$d)§(fl)+G(C)LA(d)E£e)-E(d)A(g)l ' 8(3) : A(a)(Q(_d)E(g)-§£d22(j))-DL6)LA(£)E(s)-E(dlALQHERKMdNKQ-MflM5)) A These three equation my be used to solve for the isomer weights in solu- tion. They apply only where Bew's law holds. In the case of the euthor's grephs, the velues for the slopes were: A.(e1pln) . 1.514 D(e1pha) . 0.275 G(e1phe) . 0.205 101.15.) . 0.0525 D(delte) . 1.575 G(delte) a 0.225 A(gemnn) I 0.345 D(gamnn) g 0.030 G(gennn) a 3.255 from whence A g 6.619 -11- Then, g(.) g 0.675 EKe)- 0.0290 E(d)- 0.0712 2K3) 1. 5(5) .-0.1515 e(.). 0.7559 2(4)» 0.0075 E(g) 3(3) 3-0.0320 21.)- 0.0557 E(d)+ 0.3112 E(g) hch equation my be set up es e greph so thet, knowing E(e1phe), 3(delte), end E(ge.mm), the home weights may be oelouleted. The pro- cedure for doing this can be hand in elmost any book on noeogrephy (18). As an example for preparing e nomogreph, the equation - 3(giio.0320 2(.)-0.0557 Exd)10.5112 2(3) my be set up in the determinant form. mq -0.0557 n 5(a) 1 -nq -0.0320 n E(e) l 3 0 0 zen h 1 an for the equation -- (0.0557 2(5) 3 - h - 0.0520 11.).) The quantities 2 end 3 ere scelers, 3 is e sealer used to adjust the proportions of the nomogreph, end E(elphe) end E(delte) ere unit vectors of the veriables in they direction. The variable 3 is elso e vector in the y direction, but it Ins e value dependent on the other veriebles. Letting 0.0557 n g 1, n g 17.95 0.0320 n - 0.5, n t 15063 fin So the ebove determimnt becomes, 11. 17.95 - 3(a) 1 -15.63 - 0.551.); 1 = 0 0 + 8.355h 1 Then, for the equation -- g(g) = 0.3112 3(3) + h m Ids) ' 1 ~nq ‘Oe3112 n 3(3) 1 g 0 0 E h 1 mm 2 : 8e355 men +0.3112n g +5.443, n g 17.49 The quotient mn of the‘rirst determinant must equal 1:5. quotient 55 of the seconifieoeuse the rector of 2 east be the sens in both def:- ninents. 80 the above determixent now becomes: III. 16 8g(g) 1 -17.49 -5.443 2(3) 1 g 0 0 48.356 h 1 Using determinants II end III e nomogreph cen be set up. Comidering the two determinants by the row: the third roe in eech is the same, so the h - line my be need as e base, with its z intercept equal to zero. Then, g(g) lies 16 units in the 5 direction, and extende 8 units in the z direction; it is to be mrked off in units up to 0.5 g. in the 8 units 01' length. E(g) is 17.49 units in -3_ direction, end extends 5.443 units in -z direction; within the 5.443 units of length it is calibrated up to 1.0. The E(e) line lies 17.95 in the -§_ direction, is one unit in length in the -z direction, end is marked up to one unit. The E(d) line is 15.63 units in the -_x_ direction, is 0.5 unit long in -z direction, end is marked up to one unit. The )1 line is not marked, as the value is not required. The values of 3 and n nay be varied at will so that the accuracy of the nonograph my be at an opti- ma. 0 The nonograph is operated through 31- line; a straight-edge is placed on the E(d) and E(a) values, and the intersection on the 3- line is noted. Then the straight-edge is placed across this 3- lino inter- section end E(g) measurement, and the g( g) value is read at the inter- section of the straight-edge end g(g) line. Values in full agreement with the working curve «16111.51... my be obtained .1: ease with . properly constructed nomograph. 5(2) 0.50 ‘ It h 0.25 . Bunits 17.49 units 16 units 1 5.1143 0.5 units .0 h E(s) -14- g(g)* .15“ h . .10~ .05" - E13) LJ I'e2 E(d) 1. 11 "e‘ 3(3) FIGURE S-NomOgraph of gamma isomer g(d) .1 .2 ‘ h d .1 ~ E1 3(3) 01" .0 . 1 E(a) e2“ h .3- 3(6)‘ FIGURE 6-Nom0graph of delta isomer In a similar mnner, working from the equations 1, from the equations for g(d), the determinants were: 7.446 -Ka) I 45.55 -0.1 13(5) 1 g 0 for 0.1513 3(0) .-0.0075 3(5) - h 0 4.77711 1 10 +5 g(d) 1 59.148 -6.712 E(d) l = 0 for g(d) : h 4 0.7339 E(d) 0 4.777 b 1 The delta isomer nomograph was not very accurate, possibly to an in- correct value for the slope of a line on the workig curve. The alpha isomer nomograph that the author calculated was poorly designed and therefore not satisfactory. However, not mach research was put into the delta and alpha isomer nomogreplm, es the game isomer nomograph was the only one needed. With proper selection of ‘2 and n and careful determintion of slopes of lines, a nomograph could be constructed for all three isomers. Such a naaograph would give isomer weights in exact agreeeent with those obtained from the working curves by the custonnry lengthy procedure. DISCUSSION Removal of odor Attempts'were made to remove the pungent mold-like odore of the commercial isomer mixtures. Neither the fractional sublimation.proced- ure, the steam.distillaticn, nor the chromatographic separation of the isomers on alumina operating on the prepared benzene hexachlcride ef- fected this. Chromatographing through activated silica gel, litharge, Superfiltrol, Bentonite,’Florisil, magnesium1silicate and Permutit was also unsuccessfully tried, using Shelleysolve B as solvent. The removing of the odor chemically was also unsuccessful. The chemicals tried were: concentrated nitric acid, concentrated mixture of nitric and sulfuric acids, concentrated mixture of nitric and hydro- chloric acids, sodium chromte in concentrated sodium hydroxide, zinc with hydrochloric acid, and zinc with sodium.hydrcxide. These were mixed with the benzene hexachloride and let stand at room temperature over a five day period. Chlorine in.sodium.hydroride, chlorine in weter solution, and firm concmtrated nitric acid were also used over a period of six hours, with no appreciable remults. Chromatographic attempts to purify 0.P. grade benzene'before irra- diation.with silica gel, litharge, Superfiltrol, aluminum.oxide, Bento— nite, Florisil, magnesium silicate and Permtit also met with failure. .mr. Ernest Crocker's opinion of the odors of the pure isomer samples purified by a crystallisation methods were: alpha isomer stung the eyes, but the odor was not strong-dmore of a smart than an odor; beta isomer had a slight moldy odor with slight eyesting; gamma isaaer had a sweet fragrant odor; delta isomer had very little odor. The commercial benzene hexachloride odor compared to the odor of none of the puri- fied isomers. . -15; The benzene was partially chlorinated, distilled in a two-foot packed column, then the chlorination completed, with no success. The crude insecticide was also placed in a dessicator under vac- uum alongside a dish of activated charcoal. Adsorption methods were also used. The adsorbent was mixed dry in a pestle with the benzene hemachloride and let stand in the air over a long period afterwards. The unsuccessful adsorbents used were Superfiltrol, alumimm oxide, silica gel,‘Perm1tit, Florisil, end Bentcnite. The only two partially successful deodcrizations were by adsorp- tion with charcoal-«Ins with Norits 1', and the other with 20 mesh meafihfidchnmL mmnmuhtmournmmu.na.tm week period of exposure to air. An attempt to remove the benzene hexe- ‘ chloride with n-hexene and leave the odor with the Norite was unsuccess- ful. With Cenco activated charcoal, best results were obtained byshak- ing the powdered hexachlorids with the charcoal end .11th to stand covered over a two day period. The charcoal and the hexachloride were then separated by screening. No appreciable odor increase as noticed in the screened insecticide over a three month period in air, as judged by the author's associates. Quantitative analysis of the sepa- rated benzene hexachloride showed either that the delta and game isomers were adsorbed to a greater extent than the alpha isomer, or that part of both isomers (delta and gem) were converted to alpha isomer. No fur- ther tests were ads; the results of the amlyses have been shown in Table (I). TABLE I Analysis of deodorizsd samples of benzene hexachl orocyclohexans gample No. 1 Sample No. 2 Isomer Original Deodorizsd Original Deodorizsd Alpha 72% 77% 72% 66% Delta 12% 9% 13% 9% Game. 16% 14% 15% 5% hch sample the mixed with two separate samples of Cenco activated charcoal (20 mesh), end separated from each by screening. nch charcoal sample was about one third the weight of the benzene hexachlcride. Percentages were taken on sum of alpha, delta, and game weights. Praparation of hexachlorocyclohexene A quartz receptacle and quartz mercury vapor lamp were used below 4000 X incident light. Above 4000 X, . tungsten lamp with pyrex ooh- tainers were used. wherever filters were used, except in one case (the irridiation with 2557 X light (7) ), the pyrex filters were and. by Corning. The reaction as not kept mtirely fr” of stray light. Immediately prior to irradiation the benzene was boiled to remove dis- solved oxygen. (5). During the reaction chlorine was bubbled through the liquid benzene while being irradiated with light. The benzene hexachloride formed as a white oryetallims precipitate. After allowing the reaction to go nearly to canpleticn, the remining benzene was re- moved. The samples were then analyzed for alpha, delta, and gem isomer content by infrared analysis. The ismeers are generally formed in the approximate percentages: alpha, 60-70%; beta, 5%; gamma, 12-15%; delta 10%; epsilon, less than 5%. The differences in structure arise at the time of preparation of the hsneohloride, as the isomers are fairly stable. Considering the photoactivation of chlorine (4), 012 .10 .-. cl . 01‘ the incident radiant enu'gy met be of such a frequency that it is ab- sorbed. It also met possess adequate energ to activate the chlorine atom sufficiutly to add. The addition of tho first chlorine starts the reaction. The resulting molecule is unstable, and chlorine adds with- out further actigeticn. By automatically ruling out all strained forms of benzene hexachloride, the differences in the renaining five isomers my arise from either statistical or energy considerations, or both. Obviously, the spatial arrangemuts of the hydrogen on the benzene ring, when the chlorine atoms add, give rise to the different isomers. Data has so far shown that each ismeer is formed in a definite percentage of tho tot.1. Assuming that .11 irrediations were performed with light of adequate nergy to form all isomers, this, definite proportion is the result of the molecule, before chlorine addition, possessing a definite probability of any one structure. That this definite proportion my be effected by the energy of the incident light, as shown by the trend of Table II, seem to indicate that the activation energies of the various 10 11 12 13 14 15 16 17 18 19 20 TABLE II Alpha Delta 6.56. ‘lnveleggth.x I'°'°’ 222225. .122525 % % % 64.5 9.5 12.5 Mercury Arc 70.6 6.0 9.0 3660 62.0 5.6 7.6 3660 79.5 5.5 3.6 2557‘ 56.5 10.5 15.6 4356 75.6 6.5 9.5 4900 69.1 9.4 14.5 5250» 60.2 11.4 11.9 4356 62.4 10.0 15.2 5250 71.0 7.0 10.4 tungsten 65.6 6.5 6.7 diffuse daylight 56.7 10.5 15.6 5250 60.7 12.0 16.0 5900 56.6 10.6 14.9 6000 55.9 10.4 15.5 6400 52.7 12.7 15.2 6000 52.6 12.4 14.4 5250 56.7 10.5 16.7 4900 60.2 11.5 14.6 4200 *Bee reference (7). (Sample No. 5 did not pertain.tc this experiment) -20- benzene configurations are different. This means that, if the light activates the chlorine atom below a certain minim energy, it will not bring about addition to a certain configuration of the benzene molecule. However, it may add to one of a different configuration of lower addition energy. If so, the fact that the wavelength of light has any effect on the isomer proportions indicates that the further addition of chlorine is very rapid. Otherwise, the statistical rear- rmgellsnt of the activated "di-chloro benzene” atom would nullify any favored addition of a certain structure. As seen in Table II, with increase in wavelength, the alpha isomer tends to decrease, and the delta and gem isomers formed tend to increase. This shows that the alpha isassr configuration of the benzene molecule possesses greater activation energy than either the delta or game isomer configuration. Nothing quantitative can be said, as the author's data of Table II are not sufficiently accurate. As the rearrangement of the "di-chloro ben- zens" molecule is apparently slow compared to the further addition of chlorine, the factors bringing about a greater percentage formation of delta and gel-m isomers should therefore be: (a) lowered temperature, thereby decreasing theml agitation of the molecule, and (b) increased chlorine pressure, increasing the number of ‘di-chlcro benzene"-chlorine collisiom. Both of these bring about increased yields (19). There is also, apparently, an equilibrium between at least three of the isomers, alpha, delta, and gallon, in the solid state. Infrared analyses nde of the game isomer show that what we originally pure games isomer rearranged into alpha (4%) and delta (1%), luving 95% -21- gen: isomer. The tests were ands three years apart, on the same sample, kept at room tmsperature in darkness over that period, so no photochemical or mechanical reactions due to light or excessive heat took place. One is led to believe that the decomposition was either statistically brought about, or was due to thermodynamic instability. It us show: not to be photochemically unstable. If the alpha con- figuration of the benzene molecule is more difficult to add chlorine to, it my be that this benzene configuration, and consequently the alpha isomer, is in . lower Clergy state than tho delta and tho gamma ismesrs. If this is so, then the statistical weight of the alpha iso- mer configuration operates in conjunction with the Clergy differences between the isomers. These energy differmices are probably very smll, however. Separation of the isomers In the following separations, only attempts at separating the alpha, delta, end gen isomers were made, as they were more diffi- cult to separate. The beta isomer may be almost quantitatively sepa- rated by differential solubility in carbon disulfide, the beta being nearly insoluble (16). The epsilon isomer was ignored, as the author had none pure with which to construct working curves for infrared analysis. ' Distillation methods of separation A fractional distillation procedure was attempted. A two liter plastics reactor containing about 20 g. of crude insecticide was placed in an oil math. The reactor ms exhausted and the oil bath heated. The distillate collected as a white filn on the cover of the reactor, and was washed into a flask with acetone. When the acetone was evapo- rated, carbon disulfide was added prior to analysis with an infrared spectrograph. The residue as also analysed by the same method. Another nethod of separation tried was steam distillation. Steam was passed intoa water suspension of the isoners. The steam energing from the flask containing the suspension us condensed in a water con- denser and collected in a flask. The ismr which remained in the con- denser was washed out with acetone. The acetone solution was collected, evaporated, and carbon disulfide added to the residue prior to infrared analysis. The isomer remaining in the filtrate was separated by fil- tration through a Gooch crucible with a sintered-glass bottom. The isoner nixture was then dissolved with acetone, which was collected and and evaporated. Carbon disulfide was added to the solid for infrared analysis. The residue was not analysed. The results have been shown in Tables III and IV. Though the re- sults were not accurate, it is evidmt that game-enriched distillates were brought about in each distillation. It can be seen that the vapor pressure of the alpha isomer is appreciably lower than that of either the delta or the gain. The games tends to be lower than the delta vapor pressure also. However, the greatest differences in vapor pres- sure given in Slade's article fell in the neighborhood of 40°. The fractional distillation results contradict this, the temperature being probably closer to 100°. The results of both distillation tables my TABLE III Fractional distillation of benzene hexachloride Distil- late N0. 1 Distil- late No. 2 Distil- late Noe 3 Residue after Original Oil bath temp. 65° 140° 175° Pres- sure 1m» 60 mm. Tap. inside reactor about “0 about 70° about 103° leaner alpha delta alpha delta alpha delta alpha delta ganna alpha delta gar-n * Isomer percents are accurate to 1%. m 40.0 30.0 30.0 4.3 46.1 49.6 46.2 27.4 26.5 69.8 17.7 11.5 70. 12. 15. TABLE IV Steam distillation of benzene hemachloride . Emple isomer isomer Portion of distillate alpha 44. remaining in condenser dOIt‘ 18e gen 32. Portion of distillate alpha 41. passing through con- denser delta 4. ganm 49. Total distillate alpha 45. delt‘ 200 gam 30. Original alpha 70. dOIt‘ 120 8‘” 15a t Isoner percents are accurate to 1%. be approximate, as each was the third and last attempted. The tempera- tures listed for the fractional distillation probably lie somewhere be- tween the oil bath temperature and the taperature inside the reactor. Therefore, it is assumed that No. l was collected in the neighborhood of 55°, distillate N0. 2 about 100°, etc. In the steam distillation, a portion of the residue collected in the condenser while the rest raaained suspended in the water and vas collected in the distillate. Analyses mde of each were noticeably dif- ferent, as seen in Table IV. The Iain difference betwom them is in the percentage of delta isomer, mich ans larger in the condenser resi- due than in the distillate. The percentage of gamnn was less in the condenser by about the same amoumt, the alpha remaining nearly constant. An accurate graph of the Clausius-Clapeyron equation of the isomers, in conjunction with their relative molecular polarity, would clarify this somewhat. The vapor pressures listed in Blade's article are probably not accurate enough for this purpose, as smll amounts of impurities, perhaps in the form of lower-boiling aseotopes, would give incorrect results. Capturing the results of the two distillates, the only point at which they can be compared is at the distillate No. 2 of the fractional distillation, which took place near 100°. Comparison of the data of the fractional distillate No. 2 with that of the steam distillation shows that the alpha mostly renained in the residue in the fractional distillation. Again, investigation of the Clausius-Clapeyron graph in conjunction with the relative isomer polarites would help to explain this. -26- Rough calculations made on the basis of the Clausius-Clapeyron equation using Slade's vapor pressures of the isomers indicate that the molar heat of vaporisation of the alpha, delta, and gamm iso- mrs is of the order of 12 kcal, uhile that of the beta isomer is about 19 koal. The beta structure is know by x-ray analysis to be the symetrical 1,3,5 form. Its higher heat of vaporization could be explained by postulating that the greater molecular symetry allows closer approach of the molecules, and therefore, greater attraction of the molecules, thereby requiring more Clergy to separate them. That the heat of vapm-isation of the other three isomers is nearly the same is an indication of similarity in structure, the similarity aris- ing from the supposition that their molemales met all be less symet- rical, for there is only one molecule theoretically possible that gives an isometric crystal structure (16). Chrontggaphy About 35 g. of crude isomer mixture were shaken with 250 cc. of Skelleysolve B and allowed to stand overnight. The mixture was fil- tered and ushed once with a small amount of Skelleyeolve B. The mixture us retained until use. Then the sample was prqared for the column by taking a definite ammnt and evaporating the Skelleysolve B. The residue was taken up in the desired amount of the solvent used for the chromtograph. In every case but one, the solvent volume was the same as the volume of solvent evaporated. All chromtographs used an alumina column. They were packed by vacuum applied at the bottom. The tube was tapped while the vacuum -2 7- was applied to settle the alumina. The benzene henchlcride was added to the top of the column dissolved in a given solvent used for that chronatograph. The percolate was caught in fractions of a given volume, from ixich the solvent us evaporated by reduced pres are. The residue was taken up in carbon disulfide for infra- red analysis. For the study of the action of benzme hexachloride on alumina, the chrontogralmic method used by the author wmld be classed as partition chrcmmtography. The mobile solvent us Shelleysolve B, the immobile solvent carbon disulfide. It was not strictly true in this case tlmt carbon disulfide was imobile, because the two solvents were both adsorbed to some extent. However, the carbon disulfide was probably adsorbed more, as predicted by a slightly higher dielectric constant than Skelleyeolve B, and as shown by the fact that at the start of a chromtogram, the odor of Skelleysolve B always appeared in the percolate before that of carbon disulfide. Alumim is a very active adsorbent. There are nmny examples of oheaical reaction ixich take place on an alumixm column. The fact that carbon disulfide-Skellqrsolve B mixtures and even acetone wozld not remove all the adsorbed isomer ny be ascribed to this high activity. The amount retained was roughly [roportional to the quantity of alptm and gamma isomers added, as seen in Table v, amount- ing to around 50.% in most cases. According to the data in Table VII, the amount retained was indede out of column lergth. It my have been due to deterioration of a certain percentage of the isomer. The amount retained ‘I nearly constant between ratios of 131 ad 2:1 Skelleys olve B 5335337553EITFI; 1:1 3:2 2:1 TABLE'V Milligrams alpha isomer added 57 114 170 27 42 85 23 46 87 83 85 228 342 97 116 ' 101 180 Alpim isomer retained on six inch by one- m1: inch alumina column eluted with solvent of ratio shown. Skelle olve B When Efiuflide 1:1 3:2 2:1 Milligranm gamnm isomer added 27 53 80 11 7 37 10 31 26 29 24 106 159 62 56 37 74 Gamna isomer retained on six inch by one- half inch alumina column eluted with solvent of ratio shown. -29- Bkellysolve B to carbon disulfide, as sham by Table V. Most varia- tion in Table V can be accounted for as errors in analysis, varia- tions in column packing, etc. However, the two discrepancies for the gem isomer retained when 53 and 106 mg. were added could not have been the to expe'imental errors. Their cause was not knom, but they say have bee: (he to atmospheric humidity differmces. Alumina has a great affinity for water vapor, which would tend to prevent as much game. from being adsorbed. Cm‘tain structural predictions smy be nde from a study of the relative adsorbability of the isomers. The stmcture of the beta iso- mer has been found by x-ray analysis (20) to be the symmetrical 1,3,5 form. The structures of the other isomers have not been proved. How- ever, it has ben shown statistically (2) that the 1,2,4 form ims the greatest probability of fornti on, and this be bed: called the alptm isomer. The one chair form has hem: called the epsilon configuration. Of the renining two, the 1,3 and the 1,2,3 fern, not my predictions inve been undo. However, relative adsorption of the 1,3 and 1,2,3 forum, as well as the 1,2,4 alpha form, give indication as to their structure. Structures of the greatest symtry tend to have greater heats of vaporization, indicating greater intermolecular attraction. The structure of the isomer with the greatest synunetry would therefore tend to be adsorbed more, «he to this same symetry. On a silicic acid column (10) the order of elation is alpha first, then gamma. The beta traces and delta can be removed by acetone, which is both -m- a better solvent and is probablyadsorbed more on the column. The author's work with alumixm adsorbent substantially confirmed this; though beta was not investigated. This adsorption sequence indicates that the beta and the delta isomers lmve the greatest symmetry. X-ray armlysis of fie beta has confirmed this (20); the other configuration of highest symmetry is the 1,2,3 form, which is therefore probably the delta isomer. The renmining two, the 1,2,4 form (pobably alpha) and 1,3 form may be the alpha and the gem structures, as their symmetries are least. If the alpha form is the 1,2,4 configuration, as it may be on the basis of statistical calculations (3), this leaves the gamma form as the 1,3 configuration. They are eluted very close together from an adsorbent column, also evidmce of their similarity of mole- cular p‘operti es. It is known that there is one isometric form of hexachlorocyclo- hexane, the other four isomers belonging to either the orthorhombic or monoclinic crystal classes (16). The chromatographic adsorption and crystal symmetry can be smde into a table: m Ogdm' of glutign alpha 1 game 2 delta 3 or 4 beta 4 or 3 Probable Prob able crys tal system configration monoclinic l ,2 , 4 monoclinic l , 3 wthorhombic 1,2,3 is ometri c (by x-ray analysis) 1,3,5 No adsorption studies of the epsilon isomer have bem made, but on the symmetry basis, it should act on an adsorbent column similarly -31- to the alpha isomer as it belongs to the monoclinic class. Chrontographs 0f mixfires of the alpha, delta and game. isomers were nude on alumim to determine the effect of variations in percola- tion rate, column length, change in ratio of Skelleysolve B to carbon disulfide, and amount of isomers added. The results have been shown in Figures 7 through 17. The number of the 25 ml. sample containing benzmie hexachloride has been shown along the absissa. The alpha iso- ner first appeared between the 100-125 cc and 150-175 cc fraction in every case with a 6 inch column. The amoxmt of prerun was not measured accurately. Along the ordinate have been shown the weights of the isomers in each fraction. It can be seen that in every case the alpha isomer was eluted first, followed by the gamma home. In every case the delta isomer was retained on the column, which could be partially eluted with acetone. This has been discussed later. The effect of diffm'ences in percolation rate can be sem by'com- paring Figures 7 and 8. Both columns were 6" x é". The 15 cc. sample contained 170 mg. alpha, 52 mg. delta and 60 ng. gamma isomr. The solvent ratio was a 1:1 mixture of Skelleysolve B-carbon disulfide. Figure 7 showed the result of a percolation rate of 25 cc./6 min, Figure 8 the result of a percolation rate of 25 cc./12 min. 70 mg. alpha and 23 mg. games were elated pure; 18 mg. alpim and 20 g. game were in the mixed por- tion in Figure 7. In Figure B, 68 mg. alpha and 37 mg. gamma appeared pure, while 17 q. alpim and 17 mg. gamma were elated mixed. The slower elation rate caused the isomers to define a sharper band. This -32- 40 milligrams to U o o H O 40 Hilligrams (0 03 O O H O benzene hexachloride ”Kg; c " 4 6 0' 10 12 14 $0. of 25 m1. fraction containing FIiURS 7 (1) Alpha isomer (2) Gamma isomer \px / e \\\a “a I No. 4 6 of 25 m1. c. U 10 benzene hexachloride 12 fraction containing 191311713 8 (1) Alpha isomer (2) Gama isomer 14 was because the components as they flowed down the column were prob- ably in concentration closer to those of equilibrium. The further any from equilibrium, or the faster the rate, the less distinct the band became. When the two canponents were introduced into the column, the solute flow was so rapid for the faster column, that the alumina particles had no chance to remove from solution as nany as thq were able. Naturally all were removed later, but the efficiency of initial separation was impaired. As this took place during the entire run for the faster rate, the bands were spread out. Too slow a rate is possi- ble also, though there are no examples of it here. If the rate of flow is of the order of diffusion velocities, then the components tend to come to equilibrium throughout the whole column, and less efficient separations can again be apected. No run with Ice than the 2 co. per minute was performed. An example of too fast elution has been shown in Figure 7, as compared to Figure 8. It can also be seen by comparing these two graphs that the faster elution rate reqzired more solvent. This was also because the solvent flow was so rapid that the solute did not have time to come to equilibrium between the solvent and adsorbent. Consequently, the solvent did not ranove all the solute it was able; thus more solvmt was required. A difference could also be noticed between the first and last part of each band. The first part had a steeper slope in every case. and the differences were greater as the elution velocity increased, as sea also by the three previously mentioned graphs. Non-attainment of equilibrium explained this by the same reasoning. It follows that the sharper is the band, the higher is the maxim for the same amount eluted. -33- Also studied was the effect of varying the amounts of solute introduced into the column. The solute was nade up to a nearly saturated solution filich varied from 6 to 30 cc. in volume. The re- sults can be seen by comparing Figures 7 and 9, and Figures 10, 11 and 12. TABLE VI The effect of ing amount of sample on a 6 inch alumim column. ‘1 m 8 0 1y- mge Re we we Figure Rate solve B Isoma‘ Sample Fluted Eluted retained carbon di- pure mixed on column sulfide 7 25cc/6min. 1.1 alpha 170 70 13 82 delta 52 O 0 62 gas-1a 80 23 20 37 9 25cc/5min. 1:1 alpha 57 32 2 as delta 18 O O 18 gam , 27 15 2 10- lo 25cc/8nin. 2.1 alpha 170 73 12 as delta 52 O O 52 gamma 80 46 10 24 11 25cc/6min. 2.1 alpha 223 as 39 lol delta 70 O O 70 gamma 106 42 27 37 12 26cc/3nin. 2.1 alpha 342 160 20 162 delta 104 O O 104 gamma 159 56 19 85 It was found that the efficiency of separation on a six inch col- umn was not greatly altered with variations of 100 to 500 ng. solids in a carbon dieulfide-skslleysolve B ration of 1:1, or with variations of 300 to 600 milligram solids in a solvent of a 2:1 ratio. That is. the area of the sections where the isomers overlapped varied directly with the weight of the sample. The solvent volume separating the -34- 16 12 40 01 O Jilligrams m o H O Milligrams m Rex 2 ‘1 4 6 8 10 12 No. of 25 m1. fraction containing benzene hexachloride FIGURE 9 (1) Alpha isomer (2) Gama isomer /"\ / Nl-LGN KN. j 4 6 8 10 12 No. of 25 ml. fraction containing benzene hexachloride FIGURE 10 (1) Alpha isomer (2) Jarma isomer 80 Hilligrams ab 0 o o (J O Milligrams A O 4 I 8 No. of 25 m1. fraction containing benzene hexachloride 311L132 12 (l) -upha isomer (2) Gama isomer 1 l ’ ‘LflE-fl :v" 2 4 6 I 10 12 14 16 No. of 25 m1. fraction containing benzene hexachloride FIGURE 11 (1) Alpha isomer (2) Game. isomer /°\ )/ W l \ ix 10 12 14 16 elution heads of the isomers tended to become more separated with greater solute samples. No measurements were nude on the compara- tive volume of pro-run; this would tell whether the alpha isomer came through sooner or the gamma came through later. Probably, the alpha isomer came through sooner and also the gamma later, due to the fact the gamma tended to desorb the alpha, but also because there was an equilibrium set up. That is, alpha tended to desorb game also, but this reaction was not donimnt. The eluemt volume increase as sea in Figures 10, 11 and 12 was also partly due to increase in per-- colation rats. The volumes of the original solution prior to introducing them into the column were also imestigated, as seen in Figures 9 and 13. Rob figure is the plotted result of a sample of 57 lg. alpha, 18 lg. delta and 27 g. gem chromatographed thrmgh a 6" 1 %" alumina col- umn using a solvent consisting of a 1:1 ratio (1' 8kelleysolve B-carbon disulfide. Percolation rates were 25 cc./5 min. for Figure 9, and 25 cc./6 min. for Figure 13. In Figure 9. 32 mg. alpha and 15 mg. gamna was eluted pure, with 2 :3. alpha and 2 mg. gem mixed, with 23 mg. alpha, 18 lg. delta and 10 mg. gems retained. For Figure 13, 27 lg. alpha, and 16 mg. game. were eluted pure, 3 mg. alph. and 2 lg. gamna nixed, an! 27 lg. alpha, 18 mg. delta and 11 q. gain retained on the oolunn. The solution volume was varied ten-fold, from 5 co in Figure 9 to 50 co. in Figure 15, with little noticeable effect on the eluate. The difference in eluate volume required to completely elute each isomer was probably due to the difference in percolation rate. -35- Milligrams 40 3C 20 10 20 15 illigrame ‘ 7 “a (:1 /\ \k 0 v v 0 2 4 6 8 10 No. of 25 m1. fraction containing benzene heXachloride FIGURE 13 (1) Alpha isomer (2) Gamma isomer .22p2\1L\\1 2 0 ’{::ff’ k\\fk“1r-~ -—~— 4 >— ——~<)—-———-—)— —————‘ -~>——-———~—v 12 —— e»— C!) e d if: a v4 v—i 1. \ 4 \\‘ IF§J}4¥fi>\}\&N O ‘ v 0 8 12 16 20 2A 28 32 so. of 25 ml. fraction centaining benzene hexachloriie slowis 16 (1) Alpha isomer (P) 323:1 isomer The conditions for the chromatograph of Figure 15 were: column, 27” x %” alumina; Shelleysolve B-carbon.disulfide solvent ratio, 121; sample. 170 mg. alpha, 52 mg. delta, 80 lg. gamma, and percolation rate 25 co./% min. The conditions of Figure 16'were the same as for Figure 15, except that the percolation.rate was 25 cc./W’nin. .A comparison of these two graph also shows the effect of changing the percolation rate. These two chrmtographs were the only ones nde on a 27“ column with this sample of alumina. The isomers have been Just resolved. with 89 mg. alpha and 41 ng. gunma recovered in Figure 15, and 83 ng. alpha and 39 mg. game. recovered in Figure 16. 81 Q. alpha and 39 13. game. ‘were retained on the column for Figure 15, and 87 mg. alpha and 41 mg. gamna were retained for figure 16. The alpha isomr started to be eluted‘with the 450-475 cc. fraction in each case. Figure l? is file result of a chromatograph through a 27" x fi" alunim column also, but the alumim was from a newer sample. The sample was nude up of pure isomers, 153 ng. alpha, 63 ng. delta and 76 2;. game. isomer. 9l cg. alpha and 29 mg. game. were eluted pure. 'while 62 lg. alpha, 65 ng. delta and 47 mg. gamma‘were retained on‘the column. The pure gamma eluted was low because the elution.was stopped before all the gun: was eluted tint was possible to elute. The perco- lation.rate‘was 25 cc./5§“nin., and the Shelleywolve B-carbon disulfide ratio was 1:1. ‘It can be sea: that the alpha and gamma isomers wa'e again.just resolved. The afiissa has been labeled "co. eluted” instead of "no. of 25 cc. sample containing benssne hexachloride”. The greater volum of solvent containing benzene hmchloride was probably due to -37- Milligra13 01 k) 0 400 illigraus Q. ’2) o P A H 3 O 100 40 30 10 \ w \\ 2 // 800 1000 1200 1400 1600 ll. eluted EIJURJ l7 {1) Alpha isomer (2) Gamma isomer 10 12 14 16 18 20 22 24 (\3 (7b 28 3O 32 FIGURE lB-flo. of :5 :1. fraction the greater activation state of the alumna. That the activation state of the alumira caused no relatively greater separation demon- strated that the ratio of the strength of adsorption of the alpha and gamma isaners on alumina was the same in each case. The results of the four experiments sade eluting the column with acetone are shown in Table VII. The amount of delta recovered was in- dependent of column length , and approximately 70% of the original delta added. The amounts of alpha and ganma varied between zero and 12%. This variation nay lave been due to instability. as my reactions are 1 known to take place on alumina. It could also have been due to humid- ity changes, though this is doubtful. Ramsay and Patterson published their findings of a chromtographic separation of the alpha, delta, and game. isomers of benzue hexachlor- ide (10), the results of which were graphed in Figure 18. The beta traces and delta rained on the column, and were removed with acetone. Samples prepared from crude hexachlorocyclohexane did not give clear- - cut separation between the alpha and game. isomers. The column mater- ial used was silicic acid, with n-hanane the mobile solvent and nitro- methane the immobile solvent. No figures could be given concerning the per cent recovery from a chrontogram of a commercial sample, as the original amlysis was not available. Calculations of the author of the thesis nude by assuming that the eluent betwem two bands con- sisted of only those two constituents indicate that total gamma con- tained 6% impurity and alpha 33%. Tests run using pure isomers gave recovery of 93% alpha, 84% delta and 98% gamma which were separated wall on the column. TABLE'VII Weights of isomers eluted from column after chromtggraphing Column Figure Isaner Mg. into Mg. eluted % isomer eluted length No. column by acetone 27 17a alpha 153 19 12 delta 55 40 73 game. 76 18 24 2 7 16 alpha 1 70 6 4 delta 52 37 71 gem 80 5 6 6 an alpha 114 0 0 delta 35 2 6 74 gm 53 5 10 6 7 alpha 170 13 8 delta 52 57 69 gonna 80 5 4 t The concentration of the gamnn isomer in this sample was high because the carbon disulfide-'Skelley-solve B elution was not continued to the point there no gamma appeared in the eluent. as The graph for the eluent of this column has not been in- eluded. -59.. TABLE‘VIII Solubilities of alpha, gamma and delta isomers Solvent £22333. mntane“ carbon-disulf‘idgM acetone"I alpha 0.9 5.90 13.9 butt 0.1 0.07 10.3 gamma 2.2 19.95 43.5 delta 1.6 9.50 71.1 Sclubilities given in g./100 cc. *Solubilities taken from Blade's (5) work. “Solubilities taken Daasoh's (16) work. The author used an alumina column, with Skelleysolve B as the mobile solvent and carbon disulfide as the immobile solvent. Yields of 50% alph, 50% gamma and 70% delta were obtained. The alpha and gamna fractions were pure 1 each other and of delta. The delta con- tained traces up to arund 43% of combines alpha and gamma. impurities. -41.. SUMRY Investigation of the effect of wavelength of light between wave- lengths of 2500 and 6500 2 on the preparation of benzene hexachloride revealed that relatively more alpl'a tended to be formed at lower wave- lengths with corresponding decrease in gamma and delta formed. At higher wavelengths the gamma and delta isomers tended to be firmed in greater quantities with corresponding decrease in the alpha isomer. Applying this to the mechanism of the addition of chlorine to benzene, it was deduced that the second and third chlorine molecule additions were relatively fast. The disagreeable odor of crude benzene hazachloride was partially removed by mixing it dry with activated charcoal. The odor did not return over a three month period whether the charcoal and benzene hexa- chloride were separated or not. Infrared analysis indicates that rela- tively more game. and delta is omers were adsorbed by the charcoal. Chrontographio separation on aluaim using carbon disulfide- Skelleysolve B mixtures as solvent, resulted in pure alpha and gonna isomers being separated, each in 50% of the original isomers entered into the column. By eluting the column with acetone, 75% of the delta isomer was recovered, about m% pairs. The effects of elutim rate, amount of sample. column lngth, and solvuit ratio were dismissed. 0n the basis of the order of chrantographic adsorption, the delta isomer was assigned the 1,2,3 structure and the game. the 1,3 structure. The dissymetry of the 1,3 (gamma) and 1.2.4 (alpha) isomer configurations was also evidmt from chromatographic adsorption. Nomcgraphs were con- structed for the analysis of the game and delta isomers. -42- 19. 20. BIBLIOGRAPHY Faraday, Phil. Transaction, 2.225. Meunier, Ann. Chem. Phys., 12, 223-84 (1887). Slade, Chem. and Ind., £932, 314-9. Smith, Noyes, and Hart, J. Am. Chem. Soc., _5_§, 4444-59 (1933). Luther and Goldberg, Zeit. Physikol. Chem, 56, 43-56 (1906). Waters, Coverhill and Robertson, J. Chem. Soc., 1241, 1168-72. Peskov, J. Phys. Chem., 21, 382-401 (1917). VanderLindai, Ber., 45;, 236 (1912). Kauer, mVall and Alquist, Ind. Eng. Chem., 32, 1335 (1947). Ramsay and Patterson, J. Off. Agr. Chane, 337-46 (1946). Heal and Dum, forthcoming publication. LaClair, Anal. Chem., 32, 241-5 (1948). m‘agt, m1. Chem” 33, 737-40 (1948). Bowen and Pogorelskin, Aral. Chem., 32, 346-8 (1948). Tremor, Walker, Arison and Buhs, forthcoming publication. Daasch, Anal. Chem., _1_9, 779-85 (1947). Aepli, Water and Gall, Anal. Chem., 32, 611-3 (1948). Lipka, J., “Graphical and Mechanical Computation, John Wiley and Sons, New York, 1918. Noyes and Ldghton, "The Photochemistry of Gases”, Reinhold Publishing 00., 1941, p. 295. Dickenson and Belioke, J. Am. Chem. Soc., go, 764 (1928). .43- JUL 1 4 to mu 27 '58 MAY 25" 88921.3. ‘m-‘aunlnn.lm_ v . «(A"ll~('lfl_el T547 227941 758259 Saur T547 227941 8259 Saur‘ ; The preparation and isolation of the gamma isomer Of 1,293.4,5’6‘ hexachlorocyclohexanc. vdtv i Efi AAIKC a. K '1 r‘. 1‘ I.)- ' e " $1 xii-{.12 Np: . ’.. I» l .\ '/ (1. '7‘ IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII an»):((mewmllnwl