Document G6bKYJg7REb6nqYp04n6v9EnV

OCTOBER 1947 Both blotting and print papers are removed from the soaking oath and, together, squeezed between two dry pieces of blotting paper which must overlap the soaked papers oy 0.76 inch on all sides. The pressure used should be approximately 550 pounds per square inch. The wider border of the drv papers will take jp the excess solution squeezed out of the soaked papers. Tilt! coined metal panel is then laid on the anode platen with -he metal of the pane} in contact with the mtital of the anode and the imbibition papers placed emulsion side in contnrt with the area of winch the desired porosity study is to be made. The soaked and squeezed blotting paper is laid congruent!;' on Lhu imbibition paper. .The cathode platen is then placed over the blotting paper. This entire s`sandwich" is centered in the press and uyee^ary pressure applied. The current is snitched on as soon n< ilie pressure required is obtained. After 20 seconds urrent is twitched off and pressure immediately released. The imbibition paper is placed in a developing bath of 2.5 grams of potassium ferroayanide and 2.5 grams of potassium Vrrieyariide per 100 tut. nfribriUed water. The prints-an* left in 779 this bath for 0.5 hour, washed in warm water until the yellow color of the ferricyanide is removed, and then dried bat wees blotting papers or on ferrotype tins. BIBLIOGRAPHY (I; Glazunov, A.. Chetn. Cotir., 1P3G. II, lUu. (2) llui., 1932, I, 130S. C3) Glazunov. A., and Jcnizek. L.. Korrosion u. MnnllscKms, 16. 3-J1-4 (October 1940). (4) Hormance, H. IV., Bui Lab. Record. 16. p. 370 (103?-). (5) Jbid.. IS. 209 (iO'lO), (0; Hinder. M. S,, Churchill. J. H., and Meats. R- B.. Metal Pray rws, 42, No. 6. 1070 (HrilTi. f7) international Critical Tallies, Vol. \ 1-. New 1 erk. McGraw' HiU Book Co.. 1929. (S) Slrzelba. II,, J\omt$%nu *>. .\f`'tnllscholz, 20, o, 6 (1944). Ksctivra April 10, jy-m Infrared Spectroscopic Analysis of Five isomer of 1,2,3,4,5,6-Hexachlorocyclohexane 1-. W. DA ASCI l, IVavat Jte'scarch Laboratory, Washington, D. C. The methods of infrared spectroscopy have been tipplied to the annlysis of the insecticide hcxnchlarocydohexane. The spectra of live pure isomer* of 1,2,3,4,5,6-hexaeliiorocycIohexxme nnd of n mixture of heptachlorocyclohcxnue impurities hove been determined in the 2- to 25-inicron range. A rtila- lively simple infrared method for the quiuitUoli>e determination tif each of the five isomers has been developed, which is satisfactory for the analysis of nearly jmre samples -.find relatively free of interfer ence from the usual type of impurities. The pos sible structures of the tsmuers are discussed. KCEXT advances m the use of this insecticide hexachloro- desired to obtain any information relative to the structure of the R cyclobexanc and tire inauguration of its commercial produc isomers audits effect upon insecriesdu! or toxicological properties. tion. ftirecasi the widespread use of this material. It is, of course, important ihaT both die insectiiddal and the toxicological proper PREVIOUS WORK ries be thoroughly investigated stud tins has been dampen'd by the fact that the cumrr.cr.vial Insecticide, frequently dcsknored a-4 HCH (}>oxrtihlorocycloheNanc) or BHC (benzene hexadilorkkri, *s not. a pure compound but consists of a mixture- of compounds, the.concentration and identity uf which arc nut easily determined. Thus, if is important that a.method of assay btrdevelopeu 'for use both in the evaluation of the insecticide and in the formulation of specifications for its purchase and application. Thu Industrial preparation of this compound by the chiurma- liori of ben2ano to form 1,2.3,4,5,i>hexnchlorocyclohDxa.riu results in a mixture? oi stereoisomers (alpha, beta, gamma, delta, and epsilon; as well as small quantities-of heptu- and octachloroevclo- iiejame. These isomers.differ only In the orientation of the chlo rine atoms about the cyclohexane ring. This stereoisomerism inducts a difference in bisectiridsi effeutiveni!.*;. the gamma iso mer, svhalaver its form, being the one especially endowed with insect Jidda! value. The- toxicity of the isomers toward warm blooded animals also varies, the gamma isomer being reported more toxic than DDT (5). In many cases--for example, where only the iiiseeinndal effectiveness is concerned--it juny bn suffi cient to know only the gamma isomer content of the product bur In other cases, especially In the complete evaluation of the toxi cological properties, it is import anr to know the- complete com position. * Skide (JG) and Jenkins ($) have prepared comprehensive revicu*5 of the literature oh hoxauhlorocyclohexan?, and lor the biological and toxioologlyrd aspects of tliia compound one is re ferred to these renews. Although the compound whs first pre pared by Michael Paraday in 1-S25, it wtus not until 19X2 that Van dor Linden established the fact- T.ixst the product was'a mixurn* of at least four isomer* (i>3}. In 1043 the insecticidal value of one of ;livsef the gamrun bonier, was discovered by F. 3. D. Thomas (18). Fur tin-- rcvison concentrated efforts to investi gate this compound have, been recent and the results are probably unpublished. A new isomer has been reported by Kaj ter, DuVall, and Alqubt- (9), bringing the number of known isomers to five. From Siudtfs article it appears- that up to 1945 the methods of determining the conjjxvshion of n crude preparation consisted of fractional solution and recrystalludtiem of lbe crude with appro priate solvents, A more recent article by llamsey and Patterson {13} oh partition chromatography offers an alternate means of analysis, Kauer et cl. mention that infrared spectroscopic tech niques were used- to determine the five isomers in hcxnohlorocvck'hexar.e crudes. The spectra of Kauer d ou, which extend from 2 in 15 mu, are in substantial agreement with the present work. SPECTROSCOPIC EOtni'MENT Ii wat the purpose of the present investigation to apply the All spectral measurements were made with a research.-typt methods of infrared spectroscopy to the'analysis oi the insecticide prism spectrometer of high resolving power {10). Experimental be-XAchlorOcycIohexane. Since impurities such ft$ heptn- and records were mode by recording, as a function of wave length, the octacldoroeyolohexfines exist in only very small quantities in the energy traosmt tt?>d by a blank and that tranamiued by the sample commercial insecticide, and the manufacturers believe that these ao corresponding ink traces on the recorder chart. Slit widths impurities cau be reduced to a negligible quantity (7), only tho were changed periodically by ru.tminl control. five stereoisomers are considered in tho analysts. In addition to The wave-length moftsunuitents of the. trunsmisslon curves are the development and evaluation n; an analytical method it was aenurate to =0.01 minrr>;i nr hiurorand the percent- Irausmittance PLTEXP009275 WATER PCB-SD0000054573 780 i . |u1\ 1 ljr r rw ALPHA ISOMER 'V i/Hfi 1* i_irL_ -j-i/iir Y BETA ISOMER s V i Xi_ VOLUME 19, NO. 10 "W ki i 1i ti HU ri 1 -- 7-T^ j GAMMA ISOMER \|/< r~^ mrri/ir Mi 1i :H -L.Li DELTA ISOMER I wv/-v' ha sT V VI W1 / --t i_ i -r~ --i_____ EPSILON ISOMER --V l~ -------- rr-- / 11 Figure 1, Spectra of Hexachlorocyclohexane Jjpomcrs and Heptachlorocyclohexane measurements are accurate to about to 27c over the range 20 to- 80% transmittance. Quantitative spot measurements were made by setting on a particular wave length instead of by seam rung through a hand. By taking the average uf several deSec- vions'tho.spot measurements can be made accurate to 0.5% transmittance, Samples studied in solution were placed In amalgam-sealed absorption cells constructed and cidibraibd by the method of Smith and Miller {IT). Samples studied in the solid state were prepared by grinding with mortar and pestle and mixing wi'dt petrolatum, to form a paste which was placed between potassium bromide plates with an appropriate spacer. Tho blank for the solid samples was made in a similar manner using a potassium bromide and petrolatum paste. ' BASIS OF Ar-fALVnCA.L METHOD Infrared spectroscopy can bo used both for analytical purposes and for investigation of molecular structure. The bonds in the infrared spectrum are produced by preferential absorption of various wave lengths of infrared energy, the energies absorbed corresponding to the energies of certain molecular vibrations. Thus, each compound possesses a unique spectrum characterized by its atomic composition and arrangement. The spectra of geo metrical isomers, although they mnv have several bands which coincide, will, in general, by sufficiently different &l other wave lengths to be distinguishable. The spectra of the five known isomers of hcxachlorocyelohexane nre shown in Figure 1. Discontinuities are due to different con centrations. Each isomer has at least one absorption band in b spectral region where the others are relatively transparent. These differences in spectral absorption constitute & basis for spectroscopic identification of an isomer, either alone or in a mix ture with other isomers, The presence of the various isomers in a mixture can be recognized by absorption maxima at the follow ing wave lengths: Iso-nt-f A'pha Bs. Delta Ga^ma Kp{ito Wave tenjth. oi*a 12, $S 13. -16 13.22 11. SI *A U. A3 13.95 If the region beyond 15 mu is overlooked, for reasons discussed later, a consideration of >hc intensities of the absorptions shows that the magnitudes of the spectral differences at these wave lengths are greater than at any other set of wave lengths which might bo chosen. This basis for qualitative analysis of s mix lure is sufficient for quantitative analysts if the intensity of absorption ?.* each wave length is measured under controlled conditions. PLTEXP009276 WATER PCB-SD0000054574 OCTOBER } 9 4 7 78) in the sample's or by riiaertinee, using the measured values of the other four isomers. A superior approach, however, is to deter mine the bets, isomer directly by a separate spectroscopic analysis, using an acetone solution for measurement of the extinction at the beta wave length. t X) ! ..o-' . ......................... -->t `0 SO JO 40 50 c6o0n tKTanie i-./i.nr Figure 2. Working Curves in the- absence of molecular interaction, the. Jaw of Beer and Lambert states that for a solution of an absorbing materia) in a transparent solvent log J?/J = Kcl, where E =* extinction, F'o -- light transmitted'by a blank (pure solvent), I ** light trans mitted by the solution, K *=* extinction coefficient of absorbing material, c = concentration in grams per liter, and X = length or solution path in centimeters, If the absorbing materia] is a pure compound its extinction coefficient, .K, for a. given wave length can oe determined by measuring the extinction for a solution of concentration c in a cell of length J. For a material containing several components, this equation may be written: E = KjCil d* Ej Oil -f- ifjCiI -f* ................... (1) -EGe*.! = E -- KjCji -- KiCtl -- ,... ...........(2) where the.subscripts refer to component^ 1, 2,.3, etc.* The K*s cun be .determined from measurements on the pure components and if the above equation for a mixture of n components, is ap plied t-o a sample at 77 wave lengths then simultaneous equations obtained determine tiie concentrations of the n components in the mixture. These.equations can be solved exacilv by.any of several methods--Tor example, by the use of special computing machines, or by a-graphical method of successive approximations described below. In the application of this method to mixtures of. the five* isomers of hexachlorocyclohexane Equation 1 is applied at.each of the selected analytical wave lengths. At each wave length only one component absorbs strongly and therefore all but oho of the right- hand terms of equations will be small. By neglecting these small correction terms it is possible io obtain an approximate value of ihe concentration of the main absorbing components from the ex tinctions of the sample measured at ihe various wave lengths. These approximate eoneemr&ti.ons can then he used to calculate the magnitude of the correction terms and thus to determine more accurate values of the concentrations. To obtain the* experimental measurements required.in the ap plication of the method to hexachlorocydohexane a solvent which is transparent at each of the analytical wave lengths is required. Carbon disulfide appeared to be the only sufficiently transparent Solvent and an investigation of solubilities has shown that satis factory concentrations of nil the isomers except beta can be ob tained in tiiis solvent (Table 1). Since the beta isomer Is nearly Insoluble, ihe use of carbon disulfide effects a simplification fn the analysis by firmting the number of varhible components toiour. Thebera isomer, then, may be dt-termiued by total insolubles (assuming no foreign matter A.NALYT1CAI, PROCEDURE Calibration of the method is based upon absorption measure ments on solution* of the pure isomers at the various analytical wave lengtits. The absorption measurements (calculated as per cent transmittance) .are converted to extinctions by Beer's law and plotted against concentration to give the. working curves of Figure 2. The curves appear to be linear and straight lines have been drawn. The exact analytical wave-length setting nj the spectrograph is best determined by slowly scanning the band and plotting the record obtained- This is an especially neces sary procedure at the 11.31 mu gumma band which is on the side of a carbon disulfide (solvent.) band, Since it can be shown that the optimum accuracy for a measurement, is obtained for a sample that transmits 37.5% of the light-, a procedure of analysis speci fying corudfi coll thicknesses and sample concentrations will yield most accurate results far only a certain small range of sample composition. The procedure presented here la devised for sam ples con raining approximately 70% alpha, *5% beta, 13% gamma, 10% deitn, and less than 5% epsilon. This corresponds closely with the composition of the commercial products at the present time; for sample? batting appreciably different composition the results will be somewhat less accurate unless cell thicknesses anti sample concentrations are properly adjusted. However, reason able accuracy is obtained for transmit tain: measurements be tween 20 and 80%, so that even high gamma preparations such as that reported by Gunther (.5) could be analyst! with only small changes of rise analytical procedures. Table L Solubility of Hcxacblorocyclohexaue in Carbon Disulfide0 C./tf. Gxni n\& Alpha Delta Eetn Epsilon 0.1905 0.0500 0.0930 0.0007 0.0125 c Solubilities dfttfsrminfid by -.visaing residue upon er&poraiion of sol veer from 1 co. of saturated solution at roam ts-mperature (approximately 25* G.b There are several additional considerations in the choice of cell thickness, concentrations, and slit widths. U the solvent- ab sorption is high, as at 11.81 mu, a larger slit and/or a thinner cell is necessary to obtain full-scale deflections, and if the absorption band is very Strong* as in the 13.46 mu beta band,.the thinner ceil enables an appreciable concentration range to be covered. An unknown mixture should be sampled by some adequate procedure and about 0.3 gram weighed into each of two 10-ml. volumetric flasks and made up to volume with carbon disulfide and acetone, respectively. The appropriate solution is used at the various analytical wave lengths lmdor the same conditions used to determine the working curves. The transmittance measurements are converted to ex'tbictiona and the approximate concentrations are then determined fcv the working curves. The approximate concentrations for alpha, delta, gamma, and ep silon (all too high) now enable the extinctions at each wave length to be corrected for the absorption of the- interfering com* ponervts at that wave length. These corrections are also deter mined graphically from the correction curves and are evaluations of the small negative terms of Equation 2. The new values are overoorreeted, but a second evaluation using the now values will usually give results that ate within the exparimehtsl error, The PLTEXP009277 WATER PCB-SD0000054575 782 VOLUME 19, NO. 10 beta isomer U then determined in tin* acetone solution and cor have substantially more epsilon isuir.cr than the other rections are applied using the now known purcentuges oi alpha, gamma, delta, and epsilon. EXPEIU M EXTAX RESULTS products. DISCUSSION Since this method, like- any method employing working curves, is omotrical, its inherent errors are negligible, the ultimate ac The necessary calibration curves, shown in Figure 2. were con curacy being limited chiefly by the accuracy of the calibration structed from measurements on samples o? the ptue isomers. data (working curves,. If the analyses ore performed under the Absence of impurity absorption in the spectra (Figure 1) and same conditions used for the calibration, the latter need not be narrow melting point ranges, as determined by the usual capillary sbsblure but only experimentally reproducible. Thus, the &c* method, indicated that tjio samples were of suitable purity for curarv of the method, is determined almost entirely by the experi this work: mental reproducibility realized in the laboratory. Isomer Gamma Delts Alpha E&filop Beta Melting Point, a C. 1U.8-112.2 136.0-136.2 155.3-155.8 202-227 (eub)imsd 195-210 (sublimed) If care is used in sampling, making standard solutions, handling sample, etc., the most important errors affecting reproducibility arc probably those in the percentage tran$;iuUnnee measurements. However, for any given reproducibility a high degree of accuracy may be obtained by averaging a sufficiently large number oi The calibration curves were then used in the analysis of several measurements. Hence ii may be assumed that the calibration -ynriiclie solutions of pure isomers (Table I(), For one solution data (working curve?; can be determined ns .accurately as desired detailed results which indicate the method of successive approximations are presented. Separate solutions , were used for the beta isomer determination*, since this required a different solvent. Most of the data in Table II wore ob tained before the new epsilon isomer was reported, but a quantity of this isomer sufficient for. making one, fivec-omponcni mixture was later obtained and the results arc included as the last analysis. Table III contains the results of sev eral analyses of two crude samples from the commercial production of the In secticide. The presence of epsilon isomer was neglected in these analyses, and consequently the results for all other isomers are uncorrectcd for the epsilon isomer content;. This does not affect the usefulness of the data, since the cor rections introduced hv the small amounl of opsflon in these crude snmples are of negligible size in ail but the most ac curate analytical attempts. At a later date the epsilon content of these -sam ples was determined by a separate analysis and found to be 2 add 1FJ for products I and 2, respectively. It Table II. Analysis of Known Solutions (Alpha, ramrrsa. delta determloatfrris. Known C'jneerttra liens'. Biirnma " 7.16 grpmr p'tr liter, dcln * .11 grams pe? Bier. alpba -- 5.09 gtsiRt per liter. TctsJ B rr.ple - Ci.i'b grata? per p:*T) Esnerisiental Wave Length Substance c* Transmtnanrr Extinction Total Sampie Concent.-snao. . U..-' Liter 14.53 13.22 12.55 11 .SI Gamma CSi Delta CSj ci?bl Ganvma C?S- 21.50 70.50 3S;S2 S9.90 Si. 50 0.0611 o ospn 0.4109 O.Oi60 0 0SS2 0.0241 0.6&21 0.3550 0.0541 0.2900 7 03 7.90 W*v Length 13.22 12.5S ll.Sl Sabstjiticc Deha Alpha Gamma Co.-rertlons (First Approtimatinnl lion 0.5621 0.3850 0.0641 0.2900 Alpha 0.023 0.023 ... 0.014 Carree tt'-rr* Gamma Delta .... 0.0075 0.0025 .*... o.rm 0.0040 0.004 Totid O-Oii 0 0395 0.0055 0.018 Cur* reered Fxrinc. tion o.son 0 3343 0.057-4 0.2720 Cun reeled C&neec.rated 6.BO S 3D l?.22 1J.51 Delui Gamma WjLViLengils 14.53 IS.22 O1l2-.B5Sl) 0.5621 0.3050 0.0541 0.2000 Second Appr'nlttitttfrn 0 020 0.020 .. 0.01$ ___ 0.0075 0.0325 .... 0.027 0.D04 0. 004 t> 047 00276 0 0064 O.OlT Oilier Solutions Kiwn _ Jromer Cc-fteentratior- CD>eml**ica Alpbft (Oftrarna) 6.07 3.S7 32,41 <5.07) 42.35 0.6141 0.3376 0.057* 0.2730 6.P4S .40 21 .35 Eir*ena>eaial C-nn ranuatfra 5.00 3.54 32.40 rq IQ; 42.14 has been determined that the pres ence of 2% of epsilon found for prod uct 1 would cause errors of only 0.30. 0.02, and O.OD^, respectively, in the U,3 13.22 o3i2..s53n Gamma Della Alpha fGanunat 6.4S 35. 59 (3. 48) 45.61 6.11 4.70 35.SI in OQ) 45. D2 original experimentally determined eon* ceutrations of alpha, beta, delta, and gamma. These combined error? are well within the experimental errot 13.46 13.40 13.46 13.46 Beta {tou.fr Datenuiiations Beta 4.41 BBeettaa Bela 6.25 6.71 7.89 4.40 6.20 51..5505 of the analysis. The totaled percentages, includirig epsilon, Are 104.2 and 101.251- for prod ucts 1 and 2, respectively. The high values obtained for the total concentra tions in these analyses are discussed 14.53 13. &0 13.22 J2.5S (11.81) Flve-Ccr.-.ycr.e::f He' erudmuiou Osmmfi Ep=ilcir. DAefrlittas (Beta) ;'Gsr.u:!-v 4.37 2.05 7. 77 t'2o..5o1sJ (4.37? 25. ir 4 20 1 Si W .50 below after the inherent accuracy of the method itself has been considered. The spectra of products 1 and 2, as well as a third product, No. 3. are presented in Figure 3. Although the complete analysis for the latter prod uct is not included. It was found t<i Table III. Analyst of Crude Ht.vaeldorocyelohcrnne Industrial Products Product ic Product 2 14.29 13.14 0.S3 15.15 13. OS At phi. 12.53 mi 70.29 75.55 Delta, 13,22 on 8.53 6.70 0 39 Eeu, 13.46 mu lO.-SS * 0-3'' Avers?* o.r 10 determinations. PLTEXP009278 WATER PCB-SD0000054576 OCTOBER 19-47 The accuracy of an analysis, then, will be limited chiefly by the rrprodueibilUy of liie iranHtniiunco measurements on the unknown sample and can bo calculaied for any given reproducibility of transmittance measurement. Assuming, for example. that the measurements are made at the optimum value of 37.6% transmission, an experimental un certainty of 0.69c transmittance corresponds to =0.0056 in op tical density or to =1.34% in the extinction determination. Neglecting, fur the moment, the absorption of Interfering com ponents. this uncertainty t:i the optical density determination at each Hiniiytiual wave* length leads to the following uncertainties in the concentrations of the various Corners; Uncertainly* re Length Isomer Concern r&pen. Grams/Liter (on Total Sample) * 14 3 1J SI 12.5S 13 22 13.90 13 40 Gamma nMDH Aloha DiUn Epsilon Beta *0.074 *0. lol *0.583 *1.35 *0.114 *0.120 *0.25 *0.50 *1.95 *0.45 -=0.3S *0.40 ^Analytical uncertainty resulting from an experimental uncertainty of CUV* in TranamjliAQce measurement. Total aarspio concentration = 20 gram? pe liter. The figures in the right-hand cohutm represent tlie analytical precision which can be obtained under the conditions given. Greater precision in tnuismiltMiec measurement and the use of average values would lead 10 correspondingly less uncertainty in the analytical results. It was nasiur.ejl in the previous paragraph that the corrcctfoa for the absorption of interfering isomers ui each analytical wave length could "be determined precisely and that introduction of this correction, therefore, would not contribute to the uncertainty of the analytical results. The validity' of this assumption may be shown in the following manner. By computing (from Figure 2) the quantities of the various isomer* which have equal ab sorption at ouch of the analytical wave lengths as given in Table IV, it is seen that the absorption at each analytical wave length is relatively insensitive to the presence of other components, and therefore errors resulting from inaccuracies in rite corrections for interfering absorptions are negligible. PRODUCT NCX l 783 it is concluded, therefore, that if the per cent transmittance is measured to triihiu --0.5, it is possible by a careful application of the method to obtain analytical results which arc correct to within =0.5% lor the beta, gamma, delta, and epsilon isomers and to within --2% for the alpha isomer, the values based on total sample. This theoretics! accuracy (or reproducibility) agrees well with the experimental reproducibility obtained in Tabic HI. Table IV, Parts of Interfering isomer Equivalent lo One Part of Absorbing Isomer Microns ll.Sl 14.53 12. oS 13.22 13.90' 13.40 Gamma 1 1 42 35 23 Alpha 17 23 l 12.4 73 91 Delta 578 18 25 l 443 U Epsilon 35 400 0.7 125 1 135 In addition, to these factors the vvcr-nll accuracy ias distin guished from precision or reprortucibiiho') of the analyses for the, crude samples in Table Hi will also be determined by tho ac curacy of tho working curves and by the possible presence of iuterfering impurities, such oshepta- and oetachlurocyelohexancs. Tho high values of total concentration arc very probably due to ciTora. in tho working curves, siiiue It has b;-en observed that In general all analyses made in this investigation are consistently high, and in the particular case of the gauirun dotermuuuion tho values obtained in tho'^chotfk" du'-erntinutioa at ll.Sl mu areconsistently higher titan those determined hi- 14,53 mu. It is clear that these small calibration errors could be- 'eliminated in routine work and should not be* considered in evaluating the method. The presence of isomeric hepta- and. octachloroeycloh&xane impurities may cause error? in the results. However, comparison of the spectra of the crudes (Figure 3) with the spec tra sliown in Figure l substantiates the manufacturer's claim that very little impurity it piesoat. Hence, it is believed that the results have been unimpaired and that future products will likewise have negligible corrections from such source* or-error. In ftddk ion, the error caused by loss of solvent in filling t he absorption cells -would tend to give high results. This error could be avoided by using .special absorption cell and procedures, tin' devising of which would be feasible in routine applications. It his bent indicated that both 14,53 and 11.81 mu svvm !o bo suitable for gamma isomer analysis and .tiuce n two-band chuck for the active component would bo appropriate, analyses at both lnnds have been ineltided in the present scheme. Nevertheless, Hie recommendation oi the longer wave length nutsi be streK=ed< The set of working curves at 1-4.63 mu is more sensitive to tha presence of gamma isomer and provides a moru. accurate deter mination. In addition, tho M.3l mu band has the disadvantage of btdpg cm.the side of a -strong carbon disulfide absorption band, making it difficult to locate or check the position of the center of this band. Greater accuracy in the epsilon determination may be had by using.an absorption cell of greater thickness than was used in-this investigation. In the analytical work described here no correction was made for the small amount of beta isomer which will dissolve in the carbon disulfide solutions. Some finite interference in thy analy sis for the other isomers will exist', although under the con ditions of the analysis this interference was ton small to be meas ured. However, if the method is to be applied to sample solutions containing at least 2% of beta (the quantity required to saturate the sample solution), then any interference, however small, arising from the presence of beta In tht? sample may be completely eliminated by using in place of the "solvent" blank one which is saturated with pure beta isomer. In the application of tho method to routine analysis it might be desirable to measure directly the transmittance of the sample- PLTEXP009279 WATER PCB-SD0000054577 7&4 filled cell relative to lilt- cell insivad of ro^uring separably the transmittance of 6&h cel! relative u> a blank salt or potassium bromide plate. This cun be done by employing cells oT known and approximately equal libcLm-ssc-s, one Sled with solvent for beta-saturated solvent; and the other with sample solution. . There are other combinations of 1urnd? tli&t could be used in this analysis but iLose chosen appear to the host availably from 2 to Id mu. Beyond lo mu when* tits vibrations become more cb&raeu-riitic of the molecule a= a whole one would expect the bands for different. jsorn&rs to be more isolated. Such was found to bo the case. They as? strong and there is no inter ference between isomers If beta is first eliminated by using cnrbovi disulfide solutions. This Is necessary because beta has one band at H),65 mu which overlaps the nipha band at 13.&4 mu {sec- Figure 1). The equations then would be simplified to otic term and any isomer could be determined without knowing the eon* cenirntion of the otbor isomen-, experimentally or oihonvife, Hovvcver, since many laboratories arc not equipped to investigate the region beyond 15 mu, advantage hna not been taken of those fereat spectral differences. Instead, analysis has been confined to the region hvlow 2d mu in order bast the snalyUea! method might be used in any mfrarod laboratory. It appear* passible with the set of-choaw: Wave lengths to aim* only one solution iu acetone to determine all five isomers. One would then have to ob'.uin correction eurv* for sll isomers (iii'duding beta), for they arc ah soluble in acetone. Tbi? solvent hss strong absorption at about K33 ;tnd 12.75 mu but is rela tively transparent at fdl the analytics! wove lengths. Since U is a polar solvent there might be some nonfinesrUy in working curves caused by molectdar interaction. However. ia using this solvent for beta determination no such interaction was noticed. Finally, since the analytical bands are relatively broad* the quality of the spcctn>graphio equipment used fur this analysis is well above the minimum requirement- This instrument has h resoltuioa of approximately 0.7 cm."J at 10 mu when an esti mated 3 or 4 cm.**1 would be satisfactory. The resolutions claimed for most commercial instruments by their manufacturers are usually within this range. STRUCTURE OT THE ISOMERS The alicyclic compounds have not been too thoroughly inves tigated and the! r stereochemistry is not well understood, Sacbsc (io) io his theory of strainless rings postulated that a ring such as cyclohexane would be able to take two configurations now commonly designated the 0 or "boat" form and the Z or "chair'1 form. Such aa idea (for cyclohexane; is supported by Uie work of Haase! U) who investigated /nonochkiriimed cyclohexane and came to the conclusion that within one of these forms (the Z form) the chlorine in a monoriiloro derivative was able io assumetwo positions In the mojuente, mid-ring two forms that were ir*tercaavertrble. The two positions of the ehlsrine are on vafcncc bonds parallel and perpendicular to the axis of the molecule, there being six equivalent. bonds r,i each type in the Z form of cyclohexane. Ia an earlier article fey Kassel and Ottir (&) it is stated that the possibility of other less symr&sirieal forms is not excluded. Those other less -symmetrical form? would be the 0 forru and the planar form. The word 5,V.nnar ' is used here in its strictest sense--that is, all the cariboo atoms- are contained in one plane, the plane of the- ring. It ia apparent from the literature that tills term has been loosely used. For example, Parodi ill) siates that eyi.dokvxfuift ts a planar molecule. L-ike-vt-H.. the hesahydroxy derivative, inoaital, is discussed by K. T_ Shriser and ft. Adams (2) as a. planar ring. The eonuoutriou of "planitr" meaning a moloetilo whose atoms determine several planes which ar** parallel to one plane simplifies the discussion, but- oversimplifies whon one I? discussing the numb'sr of likely isomers in a derivative, Thu.*, inositol is stated to have eight possible isomers and one would VOLUME 19. NO. 2 0 assume riuri J/2.3,',o.fi-Uxartlurdcycloh'?sar.e would feav? the same number. However, if one buikls Fischer models of i.2,3Af5.(>-hexschi-7rocyriDhexarie in the 2 and C forms the following fac*Ji become apparent. For the Z or Chair Form- l. There are sixteen pc&sfbfc configurations if one ivaumes only one -chair form i> possible. 2. Of rbesaslvtetm, five ere refarive]y5tram]c<5and one of the five is she mirror image of another. _ 3. There sre eight different possible configurations assuming ail chairs sre ictereonverribie and strained bonds such z* In the eleven strained configurations of statement 2 are slab!*. For the C or Boat Forrq. I. Ttwfi* are eight passible isomew if the two chlorines on the* bow and etirc of the boat arc di rected to the outside of the boat (it.is too strained in the other position), 2. Of there eight only two aru relatively eUamtew and they are mirror images. Patterson and White (/;?; discuss the Z form iu much the same manner. If the logical Assumption U made that those forms which are $trainless are l.lw ones which will probably form in the preparation of an isomeric mixture, i: is evident that thwe sliould bo seven isomers of l.^Spko.G-hexachlorocycloiicxftm;, of which two pairs art- mirror images (set/: Figure 4}. This leaves five isomers whose infrared spectra would be unique (the mirror Lmages baing spectroseoplefllly id&nlicah. It should br rchsktiv*ly ca*y to ideutiiy the optically avtivi? pairs by the:r activity but stone the |i?;n; tsomen* inxxj.stigrvted up* peared to be optically a^-i^ A saturated solution of the erode insecticide in a 20-cirt. s!i ahawc-d no aeiiviry, which would wm to bubsthe fact iris; ihe optically active isomers, if present:, arc racemates, Actual structural rjeterminsiions have been eonfiricsl to the t^omor fi, o'). X-ray ana-lyris fessiiy charstterirva it be in the cribic system. It tb^r^fore has s eerst-rosytnrnvirh*. struc ture and has been sssirned ri.-ft l^t5, configuration in the chair form. The other structures must- await further invesrigatioa of their x-ray dUgratus which is boing attempted as the* prsaot time. X-ray data for the asw cpritoa isomer sre said to fee Ioduden in the report of its separation (&). The extern*! crystal smictun; of those i-josasrs may be iiifc.*rinati\-e on their internal molecular arrarrgemont and may be kelpful in di^unguisbing b^itvvt*n isomeric cr?*sml3 (Table. V), 1,38 Cl BONOS ALMOST PERPCHOlCUUlR TO AXIS or molcculx l I Mtnuw iruaci ;/ (/ w VI1\ wnfioh JUA0E5 JHTERCt>MVEaTiB*-e Figure A, Sketches of Seven Strainlfe>s.Configuration^ of 1,2.3, l,o,f)-Heinchlorocyclohexaiie PLTEXP009280 WATER PCB-SD0000054578 OCTOBER 1947 785 XfrOnrtr set. E Table V Crystulioercphic Data Tjp. Aj>pro*i- Siff= Tv* BWzfrtugeDce ii I lii. l S' 1.650 1.60-1.026 T.80-1.635 !:SM The- tvU tsomit T>u:u] to bo isometric, which eIi-eek-5 the eubre ciaKsihrariuii tv '.-ray analysis. Ii3 index of re-irnciion, by tiie immerstuti method and Beokc line, tvs* determined to hs 1/330. The solubility the compound? in the usual organic immersion oils necessitated ih* useuf potassium mercuric Sodtde- w*ser solutions which were found to cii&nge indsxrapidly. As a universal sregu \vn& not used. r.kv- onentsthm of the crystals wes Uf&htfurB. The cam are therefore 'offered ss a preliminary set of ino-s-Weiiic-nta. However, the otlmr. reur isomers axe known to heloug to the orthorhombic or monoriinic crystal clxs.Sc>:. The v$ry high birsSrhigime^ of the- de-Ua ;*>mcf ad the low 2V angk'- for alpha dts&iguish hltesa crystals from the rest. To distinguish gamma from opsiten would require wore exact information on their three refractive iridic?^ alpha, beta, and puina, Thus* data n~t.*re obtained in corijuneixon wiih.O, S. Tuttle ot the Crystal SdeuQQ sd U Xava! Jx^scarch laboratory, to whrtzn Urn author is indebted. The -rimUarity of gamma- and delta goes further than crystal structure. Their iafeannj absorption speetra have seme peculiar correspondences. The doublet structure : these two isomers at the C--II vibration frequencies of 3.84 end 3.3& nut is contrasted with Use singlei structure or the other three isomers. la the region irons 7.5 to $.5 mu there is a definite similarity of shape and fmonsity benviam bands (nr gamma and delta, a? well as aloha. At larger wave lengths when? foe total molecular con figuration is more important. the sinuiuriiy disappears. Beta and epsilon Uomors are also spe<nr?Jly similar in the shorter w&vc-Ienglh regions (2 to' 10 nut). The pvxsc striking feature of those two spectra \* their simplicity when compared with those of the othor isoimits. Simplicity or symmetry of strutum-. is the usual cause for simplified absorption spectra and therefore one might look for a mom or less symmetrical molecule for the "epsilon isomer- Of the remaining four configurations (1,3,5 = beta) the 1,2,3, chair form and the 1,4 boat form arc the most symmetrical (see Figure 4). The former bus a plane of sym metry and a center of symmetry and the latter luts two planes of symmetry, Since the ypsHcm concentmlion Is iho lowest in the preparation of Om. insecticide, {he structure of higher energy content would probably Ik> epsilon. This corresponds to the 1,4 boat form (4). Validity of ihh designation rests cm rather * slender evidence and it is offered as s. suggested structure to B$s\miy when matching future data with possible structured Another spectral peculiarity, whidi -most also be dealt with in general' terms at Htu jir^ssm time, is the apparent doubling of iho bom bands at 7.6 and 3.12 mu in the alpha, cbtUs, and gamma Hornon. The single beta band a? 7.GO mu split? into two si 7.42 lo 7.82 mu for alpha, 7.52 to 7.G7 mu for deha, aiid 7.45 to 7.ST mu Air gnmnui. Similarly. with the beta band nt S.12 mu the two baud*- for rtpb.s are at 7.92 to $ IS mu at $.03 to 5.24 mu for delta, and a; $.00 to-S.21 mu for gawmii. According to Has- mum ( / 4\ who investigated cyelvhemne by infrared methods. ib* biuuds.u'j the range 700 to 1300 cm.*1 (7.5 io 14 mu) are due to CH; freeing ami twisting vibrations or C--C stretching vi- bmtlona. The apparent splitting of the beta bsnds could hi due to tbit two kinds of--CflCl--groups, depending upon -which bond of each carbon -he chlorine hud added. The two type.? of bonds would correspond to those: anzutiomsi by Bu?:s.el in hU mono- chlora derivatives. Tridorina atoms in tHe beta structure are attached by only one of tho types of bonds (pcrpamlivulAr to the jeon of the molecule). This might- explain the doublet C--hX vibrations mentioned previously but the singlet C--H vibration for i he alpha isomer contradicts the hypothesis. Further work along these liiiits will h-e repijnr-d ni a later date. SUMMARY AND CONCLUSIONS Tr.e jw-c;ra of five purv bomvra of hue insecticide 1,2.3.4,5.0hv.vudih.coe.yeloVio.'iime and of u mixture of hvj?tavhlcp.*eycl&hoxan-r impuririvs luivw besiu determined in the 2- to 25-microa range. The differences in the spectral absorption of the isonitsrs arc -sufficlsTitJy great to permit the qualitaUv* and quitrdfurive anslyv-X of mixture hy irdrared methods. Ar* Infrared method Ic? the quanUtelive deiermintton of each of the live, komers ;n the inseotietde has bean dav't-lopad and a r-wommendetl procedure is de=crihvd in detail. TL-e method U rulstividy simplv for a fivv-somn>nent- analysis and could be applied cither Its control or to research determinarious. It has been successfully applied to the analysis uf & few syathetie iiitxnm* #' jmrs isomers and to crude samples from the commercial proilucrion of the insvatkido. For all but one of the isomers the precision of the method is numerically eqimt to the precision t>: the per cent fnmsnBUanw mwisuremenis, and die cev*r-aU aacaracy of the s&othod is limited oniv by its precitioh. It is Ciuiduuivj tba.t. tlit* method ri entirely sa-UiTerury for tba azsnlysU of nearly pure samples of ihc ingectlrids .end U relatively iTec of Interferwicu from the usual iypo of impuririea encountered. It U t bat (h3% accuracy in Rtiultticai results can read ily be obtKi'iei; in must infrared iaboratories by a careful ftppilcxtri-m of th.e method. The possible* strui`turas of the isomers are fliseassed; seven are relatively stnimk'ss forms and of theta two pairs are optical ;soiptre. Preliminary crj'staliographic data for the five known isomers are given which distinguish ail but two by optical propectiey X'he nceepred cubic classification for the. structure of the belli isomer is i.-bccked by x-ray and crysialbgraphic data, ACKNOWLEDGMENT The author gratefully acknowlotlges the advice and guidance of D, 0. Smith, iivad i>l 0u? Spectroscopy Section at-ibis labora tory. He nUo wb.iios u> cypress his appreciation to the Hooker Eloetroehemicnl Company for the samples of.aiplia, beta, gamma, and dehu isomers; to Dow Chemical Company for the epsilon isomer; and to the1 Bureau of Entomology and Phutt Quaranttno far thb mixture of heptachk-rocyclohcxanca. LITER ATtTKE CITED U) Plckinswi. K. G.. and Biiicke, C., /, Am. Chttn, Sou.. 50t 764 as*28>. (2j 73,. ''Organic Cliembtry/' Vol. I, 2nd ed., p. 337, New York. John Wiley & Sons,-1944. (3) Gunther. F. Chemistry A* Industry, 1946. 309; <4i 0. Fids. JCicwi' &#&* >UU. 3,52 (1943). t7A Hnsseb 0.. and Oitar, B.. Chim.'Zcntr., 1942, XI, 2353. (Q> iieadrirkx, S. B. on.1 .BDidke. C., J. Am. Chm. Snc-, 48, 3007 0026;; <71 Hooker Elt'ctrochendcal Go., private rommucication. (fi) Jenkins. D, IV,, Chemical Warfare Service, Afrhcsi IKsUicn lUpnri 55 (Sept. 26. 19451. , iiii Knuor. K. B.. DuV&il, K_3.4 and Alquiat. F. W., Abstract* of Papers. \ ICnU Meeting. Ax. Guru. SoC.-p. 52lj Ixd- EJfU. Cubm.. 32, vm <UU7>. 00; Niclann. J. lt.< Crawford. ,F. \V., and Smith, D. C., J. OpM< Sbe. ArA., Ui be published. (li; ParreiL M.. Com-pt. 215, 13 (1242;. ;12) Patte.rean, A. L.x and 'Whic-c, T~ N., Z. Kmst^ 78, 76 (19TI). (13) finm-E^v, L. L,, and Patrereon, W, X-, /. J.tn>c. Dj-icicl Apr. Chtm... 29. 3S7 (1246). i!4- Rauus*su ib S., J. Chan. PhV3.. II. 249 (1943). 05) Sr^Uce, &tr.. 23. 1363 (IS00L 110) Slade, IG Ckrm&?*j # tndurin?, 2945.314. (17) Smith. D. C.. nnd MiHsr. E. C., J. Q#i*l Sot. Anu, 34, 130 0&44). Htcnrio. Ma?eb 42. ltJ-47- PLTEXP009281 WATER PCB-SD0000054579