NOTES - The Journal of Physical Chemistry (ACS Publications)

Chem. , 1961, 65 (4), pp 690–702. DOI: 10.1021/j100822a022. Publication Date: April 1961. ACS Legacy Archive. Cite this:J. Phys. Chem. 65, 4, 690-70...
0 downloads 0 Views 2MB Size
NOTES

690

solution depletion by the solute-mercury reaction, since the current, in contrast with the current function, i/Ch‘/z, actually increases a t each nominally greater concentration (nominal concentration equals the “as-prepared” concentration), ie., while the true solute concentration is less than the nominal one due to depletion, the former, as indicated by the current, is in all cases greater than the last preceding nominal concentration. ;Ilthough the solute-mercury reaction may so interfere with the normal chemical processes a t the reference electrode surface as to produce the potential anomaly, the potential changes observed are more likely part of a more general pattern of behavior for the calomel electrode in sulfur dioxide, first noted 11y Cruse,*mho studied the cell Ag/AgCl/SOz, EtzKHzCl/Hg,Clz/Hg

( 11)

in which the calomel electrode acts as a cathode. For the first few hours, the potential remained fairly constant at 50 mv., which is close to the value (8) K. Cruse, 2. Elektrochem., 46, 571 (1940).

Vol. G5

calculated from thermal data; subsequently, it increased over a period of 12 hr. to a new and fairly constant value of 280 mv. This aging effect was observed in other cells, using either the calomel or Hg-Hg2Br2 electrode, but not in cells in which the Ag-AgC1 electrode was used with, e.g., a hydrogen electrode. Cruse concluded that the aging is characteristic of mercury-mercurous halide electrodes in sulfur dioxide, and that such electrodes are not suitable for measurements of more than a few hours duration. The potential change found in the present study, in which the calomel electrode functions as an anode, is logically in the opposite sense to Cruse’s observation. Experimental The material, apparatus and procedures used have been described.’

Acknowledgment.-The authors wish to thank the U. S. Atomic Energy Commission which helped support the work described.

NOTES DISSOCIATIOK OF MOLYBDENJM(V) CHLORIDE I N CARBON TETRACHLORIDE SOLUTION’ BY IRVING 31.PEAR SON^

AND

CLIFFORDS. GARNER

Department of Chemistry, Cninersify of California, Los Angeles 8.4, California Received A p r i l 11 1960 ~

Xolybdenum(V) chloride is a blue-black dimeric3&solid having normal melting and boiling points3b of 194 and 2G8”, respectively. The dark red-brown vapor apparently consists of MoC16 trigonal b i p y r a m i d ~ ,easily ~ decomposed to lower chlorides and Clz under appropriate conditions. Many studies of MoCI5 require specialized handling techniques because of the extreme sensitivity to moist,ure and oxygen, which form oxychlorides and, respectively, HC1 and C12. We give here observations on the hit,herto unreported solubility and dissociation of MoC15 in CCh solution, as well as on the behavior of MoC15 in vacuum distillation and the removal of oxychloride contaminants from MoC16. Experimental Argon, Hydrogen, Nitrogen.-Tank Ar ( I h d e ) , oilpumped tank Nz (Liquid Carbonic Co.) and electrolytic grade tank H, (Liquid Carbonic Co.) were each passed through Drierite and Ascarite, then through finely-divided Lr metal a t cu. 800” (Ar, AT,) or ca. 600” (H2). (1) Supported by U. S . Atomic Energy Commission under Contract AT(ll-1)-34,Project 12. (2) Electronics Division. The National Cash Register Company. Ilawthorne, California. (3) f a ) D. E . Sands and A. Zalkin, Acta Cryst., 12, 723 (1959); (b) H. Dehray, Compt. rend.. 66, 732 (1868). (4) 11. T’. G . E w e n i and 11.W.Lister, T r a n s . Faraday S o c . , 34, 1358 (1938).

Chlorine, Hydrogen Chloride.-Matheson tank C1, was purified by the method of Downs and Johnson.5 IIatheson tank HC1 was dried by passage through Mg(C104)Z. Carbon Tetrachloride.-J. T. Baker “Analyzed” CC1, was ourified (76.8” normal b.o.i essentiallv “ bv “ the method of W‘allace and Willard.6 Molvbdenum(II1) Chloride.-MoCL (Climax Molvbd e n u k Co.) was boiled with 12 f HCi, then n-ashed i i t h absolute ethanol and dried in a vacuum desiccator. Molybdenum(V) Chloride.-For most of these studies MoC15 was synthesized7,*from purified C12 and %Io metal (obtained by reduction of J. T. Baker “Analyzed” Moo3 with purified Hz at 1000°,followed by heating in anhydrous HC1 to remove possible oxide impurities). Climax Molybdenum Co. MoC15 (special lots selected for large lump size) was used in spectrophotometric studies of the dissociation. MoClj prepared from MooBand CC1, a t 4OOo9 was supplied by Professor S. Y. Tyree, Jr., and used in oxychlorideremoval studies with the same general results as our own MOC15. Removal of Oxychloride Contaminants from Molybdenum(V) Chloride.-MoC16 from the above sources generally had brown and green oxychloride coatings. \+’hen coatings were removed mechanically in the best dry boxes available to us new coatings formed rapidly. Distillation in a C1, atmosphere appeared to distil oxychloride with the MoC~S; attempts to fractionate were unsuccessful. Heating in z’ucuo or in Nz gave essentially no distillation except on approaching the melting point whereupon extensive release of C12 resulted, in accord with findings of Honigschmid and Wittmanns and thermodynamic estimatcls of Brewer and co-workers,1° but in apparent conflict n i t h a clainl” that I

,

(5) J. J. Downs and R . E. Johnson, J . A m . Chem. Soc.. 7 7 , 2098 (1955); J. J. Downs, Ph.D. Thesis, Florida State University, August. 1954. (6) C. H. Wallace and J. E. Willard, J . Am. Chem. SOC.,72, 5275 (1950). (7) A. Voigt and W. Biltz, 2 . anorg. Chem., 133,277 (1924). ( 8 ) 0. Hdnigschmid and G. Wittmann, ibid.,2119, 65 (1936). (9) K. Knox, S. Y . Tyree, J r . , R. U. Srivastara. V. Norman, J. Y. Bassett. Jr., and J. H. Holloway, J . A m . Chem. Soc.. 79, 3358 (1957). (10) L. Brewer, I,. A. Bromley. P. V’.Cilles and N. L. Lofgren, in L. L. Qiiill (ed.), “The Chemistry and hIetallurgy of Miscellaneouv

April, 1961 ?VIoClScan be vacuum sublimed away from MoCId a t 120" and the claim that MoC15 can be vacuum distilled a t 150°12a or even a t 60-70°.129 The latter paper also stated that solid MoC16 and gaseous MoC16 in equilibrium with it decompose to Cln and solid hfOC14 a t 98-162". Apparently MoC15 cannot be purified by vacuum distillation a t these temperatures unless the distillation is substantially faster than the dissociation. Repeated extraction of solid MoC16 with cc14 in vacuo iii a Pyrex apparatus allon-ing filtration of the solid from the extract and evaporation and re-use of the CCI, gave purified i\.loClSwith a Cl/hIo atom ratio of 5.00. The red-brown final CCI, extracts (including C1, in gas phase) had Cl/Mo atom ratios close to 5, whereas the first cc14 extracts were bright red with CI/?Ao atom ratios of 3.4-4.2. Measurement of Chlorine Produced by Dissociation of Molybdenum(V j Chloride in Carbon Tetrachloride Solution. -In the course of working with R'IoC16 in CCI, LUCUO Clz was found presmt in amounts grossly in excess of that which could be formed from any O? possibly present. Two methods were used (and a third attempted) to measure the Clz formed. Concentrations of Clz in the MoC4-CC1, solutions and in thi. vapor phase above each solution were deduced froni either the C1, in the vapor phase or the total Clz with the aid of the perfect-gas and Henry's laws.13 1. Distillation Method.-Aliquots (10- or 20-ml.) of ? V ~ O (puriiied C~~ by extraction method) in CCI, nere delivered from an evacuated buret equipped with a bellowstype brass vacuum valve with Cu-P) rex seals into an evacuated Pyrex evapor:ttor5 and the CCla, CIz and any HCI distilled in 2 min. ("fast") or in 18 min. ("slow") into traps immersed in liquid IYZ. The trap contents mere distilled into OZ-free:tqueous KI, and the 1,- formed titrated with standard n'aZSzO3,after which any HC1 a-as determined iodometrically . 2. Aqueous KI Method.-Aliquots of the same MoCl6CCI4 solutions n-ere delivered into the evaporator, then 0 2 free aqueous K I nxs rapidly added vith vigorous stirring and the 11-formed titrated with standard Na2SzO3. Clz so determined i: a measure of the "equilibrium" concentration if the assumptions he made that the l\IoCl6 is instantaneously hydrolyzed before further dissociation occurs and that oxidation-reduction is negligible here except between Clz and Ia-. 3 . Spectrophotometric Method .-Spectrophot ometric measurement of Clz in MoCl&CI, solutions was not considered feasiole because of complications. At the suggestion of Professor J. D. McCullough, we attempted to measure the intensity of the Clz absorption band a t 328 mp in known aliquots of the gas phase in apparent equilibrium with ?V1oCI5-CCl4 solutions. Because of considerable a t tack of MoClj by rnoieture during the extensive handling the results w r e useful only in showing independently of the other two methods that C1, is released when MoCI, dissolves in CC1,. Molybdenum and Chlorine Analyses .-lloC15 and its CC1, solutions and residue? were hydrolyzed in sealed apparatus with excess 1f KaOH, the ?\.Iooxidized to hlo(V1) with 30% H202, then m x s s H202and any CCI, pSesent removed by boiling. Aliquots of the colorless solutions mere used for determining M o by the a-henzoinoxime gravimetric methodI4 and C1 by Clarke's methodlo (satisfactory at pH 3.0-3.5 in the p~esenceof Mo(V1)). Dry Box.---Among several drv boxes used, the best was a Lucite box cbquippcd with an air lock, a Sa-arc purification system (No. 106, Caemco Iuc., Florida), and two Seoprene gloves Treated with Fluorolube 11 to reduce diffusion of 0, and moisture into the box. A second pair of ?;eoprene gloves was worn by the operator. A purified Ar flow was used inside the box. Although the box contained Materials: Thermodynamics," McGraw-Hill Book Co., Inc., N. Y . , 1st ed., 1950, Paper 8, p p . 276-311. (11) D. E. Couch and -4.Brenner, NBS Report No. 532G, J u n e 14, 1957, p. 11. (12) (a) S. .A. Schukarev, I. V. Vasil'kova and B. N. Sharupin, Zhur. Obshchei R h i m . , 26, 2093 (1956); (h) Vestnik Leningrad Unir., 14, No. 10, Ser. Fiz.i Khim., No. 2, 72 (1959). (13) Henry's law ooniitants were calculated from the data of IT. J. Jones, J . Chem. Soc., 99, 392 (19111, and N. W. Taylor and J. €I. Hildebrand, J . Am. Chem. Soc., 46, F82 (1923). (14) H. B. Knowles. Bur. Standards J . Research, 9 , 1 (1932). (15) F. E. Clarke, Anal. Chem., 22, 553 (1950).

KOTES

691

fresh PZOLand gave negative tests'6 for Oz with liquid Ka-K eutectic alloy and with a mineral-oil suspension of sodium benzophenone ketyl for 15-30 min. periods, hIoCl5 exposed to the dry-box atmosphere formed coatings within a minute or less.

Results and Discussion Our experiments show clearly that Clz is released when M0C15 dissolves in CCl, a t 2-26", unlike PC15 and in contradiction with the inipression gained from the literature that nIoC15 dissolves in CCl, without reaction. Because of the fantastic sensitivity of hIoC15 to moisture and the low solubility of MoCl5in CC14, our experiments do not establish the nature of the &!to species formed in the dissociation. A!toC13 appears improbable since CC14 equilibrated for days with 1\IoC13powder had a solubility of f at 25" and our hIoC16CC1, solutions gave no solid phase on centrifugation. Residues from the total distillation of Clz and CC1, from MoC16-CC14 solutioiis were brown-black solids with green discolorations and C1/ N o atom ratios of 3.35-3.37, suggesting the presence of oxychlorides and possibly i\IoC14 (attack of hIoC15 by moisture in a t least one distillation was shown by finding 0.106 mmole of HC1 in addition to 0.066 mmole of Clz). Thermodynamic estimates of Brewer, et U Z . , ~ and ~ our own data are compatible with dissociation by the path


l Richmond, Virginia, Kovember 5, 1959 ( 2 ) See, f o r ~ u n n i i > l e . \. N Campbell and E. M. Kartzmark, Can. J C/LPVL57, 1 IO0 (1959). €1 5 Iireinemshers, Z physik. Chem., 1 1 , 80 (1893). 1'1'1 tnrl 1 C Ricri J .4m Chem Sac , 63,4305 (1931). IIf r < ,I C h m Lduc , 35, 510 (195s).

(

Vol. 65

perature and pressure constant) of a ternary system composed of mater and two salts, the components identified, respectively, as 1, 2, and 3. Assume a tie line joining a solid phase D with the solution A in equilibrium with the solid. This tie line is the locus of points of mktures from which the phases A and D will form at equilibrium. Consider two such mixtures on this tie line, B and C. If the points B and C are fixed, the direction of the tie line is known, although this would not fix the location of either A or L). If, however, any one of the coordinates A1, dzor ilgof A were also known16 that point would be fixed. This follows from the linearity of tie lines, which requires that

In like manner knoving any one of the coordiiiat,es 01, Dzor D 3sufices to fix point D, because

Thus in order to apply the method described here, the following must be known: (1) the location of point B, ( 2 ) the 1ocat.ion of point C, (3) oiie coordinate of point A, (4)one coordinate of point D. The first of these is available upon synthesis of niixture B. While in principle C could also be established by synt,hesis, this mould be highly impractical. Instead t,hat point is found indirectly, as described below, after determination of the third quantity. By the methods used here the third quantity is AI. It is determined by evaporation to dryness of a weighed fraction of the solution phase of mixture B. The fourth quant'ity is det,ermined as in the method of algebraic extrapolation, after the location of point C is fixed. A knowledge of -41found for the solution phase of mixture B would fix point C if A1 were known as some systematic function of the compositions of mixtures. Such a relat'ionship can be established by the detjerminat8ionof A1 for a series of mixt,ures having an arbitrary but fixed weight per cent. mater and variable ratios of the two s a h , i.e., variable C2/C2 C3 at const>aiitGI. Because t,he location of point C on the tie line is, in principle, arbitrary, C may be considered a member of the series of mixtures a t consttant C1. Thus data are available for a reference plot of Al vs. Cz/Cfz Ca a t known C1. For a value of Al determined for the solution phase of mixture B it is thus possible t o fix the compositjion of the corresponding point C from this plot; point's B and C must have the same A1 value if they lie on t,he same tie line. ,4 dctermination of A1 for a number of such mixtures us B thus h e s corresponding points C on the same tie lines, and defines both t'he solubility curve and the solid phase. As there are no t,ie lines in an invariant region, an invariant point cannot be established by the above method. This point must be fixed instead by extrapolation of Olie solubility curves t o t'heir

+

+

(6) The codrdinstes (in weight per cent.) of a point such as .Iare given by AI, Az and A3, where the subscripts refer to the components. (7) Equation 2 will be recognized as the basis of solid phase identification b y the method of algebraic extrapolation,* except t h a t t h e co111Dosition of the solution phase has been replaced by t h a t of a second mixture lying on the same tie line.

April, 1961

SOTES

603

Wt. % XaCl. point of intersection with one another. Determination of points on the solubilit,y 0 6 12 18 24 30 curves in the vicinity of the iiivariaiit 2 -t3.Oo point reduces the uncertainty in Lying E2 +2.50 the latter. The extrapolation is also 2 rendered less uncertain in that the two 3 +2.00 curves should int.ersect a t the AI found & experimentally by evaporation of frac- 0.; +1.50 t,ions of the sol-ution phase of mixtures +l.oo in the invariant region. bThe applicability of the method will 2 $0.50 depend upon the nature of the system. O.O0 As applied here, it obviously would not 5 be applicable to a system where a solu- fi -0.50 bility cume had. an essent,ially unvary0 10 20 30 40 50 ing value of AI. It would also be inWt. yo NaN08. valid for systems in the are Fig. 1.-solubility curves expressed as difference plots--(I) the either vol:hle o'r decomposed by heat. system KCl-NaC1-H20 at 25": c ,lit. values8; #, present T3-ork. I n addition for some systems wit'h such (2) The system NaBr-XaNOrHtO a t 25': 0, lit. v a I u e ~ 1 ~e, ; prescharacteristics as metastability or a rela- ent work. tively large num.ber of solid phases, it is ing evaporation to dryness. Evaporation R-as continued possible that t,ht: method would not be practical.

?$ ~

Experimental Salts, ACS or reagent grade, were recrystallized before use and dried until a sample showed no weight loss when heated overnight in a 100" vacuum oven. Salts were stored a t 150'. Il'ater was singly distilled and protected from CO, during stomge. Mixtures attained equilibrium a t 25 & 0.05' while constantly agitated with a magnetic stirring device. Two types of mixtures are required, corresponding to either B or (= as identified above. To distinguish betmeen the two, the former will be referred to as scanning mixtures and the latter as reference mixtures. While data for the systems were available from the literature, this fact \%-asessentially ignored to simulate conditions which might be encountered in unknown systems. Before preparing the mixtures a crude preliminary study of the solubility curves was made by adding water to known mixtures of the two salts, with stirring, until the solid phase had been reduced to a few small crystals. This is a convenient method of establishing optimum compositions of t,he scanning mixtures to be prepared, as well as indicating the general form of the solubility curve. Scanning mixtures were selected 60 that their compositions would lie near the solubility curves in order to minimize extrapolation errors inherent in the use of equation 1. The desired values of C1 t o be maintained constant in the reference mixtures 'were estimated visually during the preliminary crude study. For best results C1 should approach D1,but not so closely that insufficient solution phase is available for removal. ]?or the liquid-solid two-phase regions of both systems, a C, of 30 to 40 weight % was suitable. Mixtures were prepared by weighing in situ each component after it had been added to the solubility bottle. Total weights of scanning mixtures were a minimum of 15 g., while those of reference mixtures were 25 g.; weighings were made to the nearest whole mg. The latter mixtures were in larger amount to minimize the difficulty in obtaining a precisely preselected C1 value. Compositions of mixtures used in the determination of the weight per cent. mater of an invariant point were not precisely known. AS the approximate location of the point was known from the preliminary study, the only restriction in these invariant mixtures was that the composition should fall within t.ie limits estimated for the invariant region. Like values of dl for the solution phases of several supposedly invariant m.iutures were accepted as evidence that this value had been established. For both systems 12 hours were sufficient toattain equilibrium. Dupl.icate determinations from under- and oversaturation indicated no supersaturation tendcncies. Fractions removed from solution phases weighed a minimum of 1 g., and were withdrawn through filters. k3olution fractions werc weighed and dried in 100-ml. glass-stoppered volumetric flasks. These were used because the extended necks prevented loss of solids by spattering dur-

a t 150' until no further weight loss occurred over a 12 hour period. The minimum weight of residue obtained T V ~ S0.6 g. Weighings concerned wit~h solution fractions were made to the nearest 0.2 nig. Duplicate detcrmin:ttions of the nTeight per cent. water of solution phases agreed t + x d l y within =t0.05 weight yowater.

Results System 1.-Results are compared with accept'ed using the Of Rei11ders>8 in Fig' The comparison is made lvit'h a difference Plot as departure from the straight lines formed by joining the invariant point to those of t~heindividual salt,s in n'nter. The plot is based the 'lest curves, by eye, drawn throuf;h t,he data obtained in the present work. The solubilities of KCl and XaC1 in mater reported here are, respecti\Tely, 26.43 and 2G'44weight %. Accepted are 2R*30-26'73 for the formerg and 26-40-26.52 for the' htter." For the invariant point the values 11.13, 20.46 and 68.41 I ? . (11) (a) W. C. Rlasdde, I n d . Eng. Chem., 10, 34-1 i i ~ 1 8 ) : (1,) I (16) Jt EI RIGCI, J * Am. Cham! SOC,,67s 885 (1886)e

mn.

( 5 ) ”Condensed Chemical Dictionary.” 6th Edition; Reinhold Pub14 C O ~ DNew ., Y Q A N. IT., 10861

NOTES

April, 1961

695

TABLE I1 HEATOF COMBUSTION OF ETHYLENE CARBONATE Mol. wt. = 88.062 Sample wt. vac., g. (1)

1.60425 1.78164 1.51918 1.56722 1.47437

Theoretical

co2, g. (2)

2.40522 2.67091 2,27768 2.34970 2.21049

2.40316 2.67106 2.27775 2.34799 2.21037

30.6 34.0 29.0 29.9 28.2

0.119406 ,132315 .112862 ,116512 ,109697

21,513.7 23,840.0 20,334.5 20,992.2 19,764.2

69.0 69.0 69.0 69.0 69.0

5.8 5.8 6.4 6.9 5.8

8921.1 8897.3 8894.3 8908.2 8907.7

Mean 8905.7 Standard dev. of mean h 4 . 7 purity. The freezing points listed in the more recent literature, references (4) and (5), although not substantiated, checks more closelv the values obtained in this report. Apparatus and Procedure.-The heat of combustion was determined in a National Bureau of Standards calorimeter, manufactured by the Precision Scientific Company. Resistance measurements were made with Leeds and Northrop type, G-2 Mueller Bridge and a platinum resistance thermometer calibrated by the National Bureau of Standards. All combustions were made in a 340 cc. “Parr” Combustion bomb using purified oxygen. The bath temperature of the calorimeter was maintained a t 29.99 1 0 2 ” and calibrated with “Parr” Benzoic acid (calorific value 6318 cal./g.). The sample was dried in a vacuum desiccator a t room temperature to constant weight. A moisture analysis by the “Karl Fischer” method showed an insignificant quantity of water.

Results The results of the calibration are listed in Table I. I n Table I1 is listed the results on the heat of combustion of ethylene carbonate. After correcting to 25”, constant pressure, and including a modified “Washburn” correction, calculated according to Prosen,G a value of 3179.0 f 1.7 cal./g. is obtained. Acknowledgments.-The authors wish to acknowledge the assistance of Mr. R. Trask of Picatinny Arsenal, Dover, N. J. ( 6 ) F. D. Rosaini, ”Experimental Thermochemistry,” Editor, Ch. 6, Interscience Publishers, Inc., New York. N. Y.,1956.

FERROCENE AS A RADICAL ‘cSCA\7ENGER” IN THE RADIOLYSIS OF CARBON TETRACHLORIDE BY E. COLLIXSON, E”. S. DAINTON AND HUGHGILLIS Department of Phvszcal Chemistry, The University, Leeds, b, Enoland Recetaed September 19, 1960

Values of the radical yield, GR, for the radiolysis of deaerated, liquid carbon tetrachloride, using radiations of comparable linear energy transfer, differ widely. Gsing diphenylpicrylhydrazyl (DPPH) as a scavenger, Chapiro’ obtained 19 f 1; using the polymerization method, Seitzer and Tobolsky2 found a d u e of 10.2; with methyl methacrylate to scavenge the primary radicals and ferric chloride to oxidize the polymethyl methacrylate radicals so formed, the highest value obtained

was 8.73; and the radiation-induced exchange of C136-labelled chlorine with carbon tetrachloride suggests that GCc13 = 3.5 f 0.35 and that GR is 7.4 The solubility of ferrocene (Fn) perhaps in this solvent, its small positive oxidation potential and ability to react with free radical^,^ together with the recent evidence of Brand and SneddenG that it can be oxidized by iodine atoms generated photochemically, all point to the probability that ferrocene would efficiently scavenge chlorine atoms produced by irradiation of CCI, according to the reaction 1

-

Fn

+ C1+

Fn+Cl-

(1)

where Fn+Cl- denotes ferricinium chloride. We here report our observations on this system. Experimental May and Baker carbon tetrachloride was saturated with chlorine and then illuminated for 90 min. with an unscreened medium pressure mercury lamp in a quartz envelope, after which it was washed successively with 5 M aqueous KaOH and water, dried over calcium chloride and carefully distilled. Ferrocene was recrystallized once from methanol, dried in DUCUO and sublimed in vacuo a t 100’ into the empty irradiation cell, the solvent being added subsequently. co-60 yrays were used to provide a dose-rate of 3 X 1019e.v. 1.-’ sec. in the solution. Solutions of 5-ml. volume were degassed thoroughly before irradiation. The concentration of ferrocene during an experiment was followed by decanting the sample into a quartz spectrophotometer cell attached to the irradiation cell. Two sintered glass filters of porosities 2 and 3 between the irradiation cell and the spectrophotometric cell prevented the precipitate (Fn+Cl-; see Discussion) which was formed from interfering with the spectrophotometric measurements. The molar decadic extinction coefficients, determined from samples of freshly sublimed ferrocene weighed in air, were: emax = €807 = 785; E ~ =~ 310; ~ , €360 = 185; emln = €390 = 67; emnx = €460 = 105. Theee values agree as nearly as can be determined with those given in curve (e) of Fig. 1 of reference 6 .

Results and Discussion The variation of [Fn] with dose mas measured for solutions containing five different initial concentrations within the range 1.2 to 8.3 mhf. I n four of these the change in optical density was measured a t three different wave lengths and in the fifth solution a t two. The graphs of optical density with dose were linear down to a concentration of -0.5 mdf and correspond to G(-Fn) = 2.34 =t 0.07.

(1) A. Chapiro, J. Phys. Chem., 63, 801 (1959),and earlier papers cited therein. (2) W. H. Seitner and A, V, Tobolskr, d . Am. Chem. Sac., 71, 2687

(3) N. Colebourne, E . Collinson and work. ( 4 ) J. W.Schulte, J. Am. Chem. Soc., (5) P. L. Pauson, Quart. Revs., 9, 391 (6) J. C . D. Brand and W, Sneddsn,

(1866),

(1867).

F. 5. Dainton, unpublished 79, 4643 (1957). (1955). Trans. Psradnv S n r . , 119, 894

NOTES

696

There was evidence for a small back reaction in that the initial slope a t 307 mp of a solution of initial concentration 1.2 m M was 0.052 whereas the slope for a solution in which the concentration had been reduced by irradiation from an initial value of 8.3 to 1.2 mM was 0.037. The possibility that the irradiation products of pure carbon tetrachloride in some way affect the ferrocene or other products was excluded by direct experiment. To test the stoichiometry of the reaction, dry methanol was added under vacuum to the residue after the solvent had been pumped off an irradiated solution. The resulting solution was blue, having an absorption peak a t 620 mp. If this peak were due entirely to the ferricinium ion for which €690 = 362 in methanol,6 then the concentration of Fn+Cl- present was within 5% of that of the ferrocene destroyed. When a sufficiently large dose had been given completely to destroy the ferrocene, a small new absorption band (Ama, 320 mp) was observed. This was not due to chlorine which in carbon tetrachloride solution has a spectrum with Amax 332 mp. Possibly a substituted ferrocene is responsible for this band. If so the yield is small. If reaction 1 were the only origin of Fn+Cl- then we could conclude G a = 2.34. However it is possible that ferrocene may be oxidized by CC&+ (an abundant ion in the mass spectrum of CC1,) and CCla (almost certainly the precursor of the CsCls found when pure CC1, is irradiated). It is also possible that a fraction of such reactive radicals as chlorine atoms may abstract a hydrogen atom, the resultant ferrocene radical being the immediate precursor of the substituted ferrocene thought to be responsible for the spectrum with A,,,, a t 332 mp. An attempt to measure G(HC1) for a ferrocene solution was unsuccessful due to obscuration of the end-point by Fn and Fn+Cl-. We conclude that although solutions of ferrocene in carbon tetrachloride respond to irradiation in a typical “indirect action” manner the value of G(-Fn) = G(Fn+Cl-) = 2.34 is merely a minimum value for GE. Acknowledgment.-MTe wish to thank the Rockefeller Foundation and the General Electric Research Laboratory for financial aid.

Vol. 65

has ever been reported giving the concentrations of TDI, intermediates, and final product throughout the course of the reaction. Such data, applied t o recently derived equations,6 would permit more detailed characterization than has heretofore been possible of an important class of condensation polymers, polyurethans derived from TDI. Consequently, the following experiment and analysis were undertaken. TDI was treated with n-butyl alcohol-xylene solution a t 80’. The concentration of -XCO as a function of time was determined by transferriiig periodically withdrawn reaction samples to excess n-butylamine solution and back titrating wil h HC1. A representative curve is shown in Fig. 1. Values of --h’CO, as read from the smoothed out curve of Fig. 1are compiled in Table I, col. 2. TABLE I COMPARISON OF EXPERIMENTAL CONCENTRATIONS OF IsoCYANATE WITH CALCULATED VALUES Time,

hr.

Exptl.

Calcd. 1

Calcd. 2

0 0.100

0.250 .224 .205 ,187 .173 .162 .146 .134 ,126

0.25000 ,22235 .20208 .18657 ,17433 ,16443 ,14635 ,13409 .12520 .11843 ,11307 .lo869 .lo189 .09673 ,08914 .OF3352 ,07900

0.25000 ,22244 .20217 .18664 ,17438 ,16444 ,14630 ,13400 .12511 ,11834 .11300 .lo865 .IO190 .09680 .08925 .0835S .07894

.zoo

.300 .400 .500 .750

1 .ooo I . 250 1.500 1.750 2.000 2.500 3.000 4.000 5.000 6.000 Constants (l./eq.-hr.)

Calcd. 1 Calcd. 2

.118 .113 .lo9 ,102 ,097 ,090 ,084 ,078

Std. deviation

B1

kz

kr

k4

8 . 5 4 4.00 1 . 9 2 0 . 5 3 9.21 6.00 1.20 0 . 6 3

THE REACTION OF WITH n-BUTYL ALCOHOL BY ARMAND Dr GIACOMO E. I . du Pont de Nemours R. Cornpanu, Eastern Laboratory, Gzbbstown A’. J . Received September 18. 1060

(1) I. C. Kogon, J. Ow. Chem., 24, 438 (1959). (2) M. E. Bailev, et al., I n d . Eng. Chem.. 48, 794 (1956). (3) 1cI. Morton and M. A. Deies, dbstracta of Papers, Div. of Paints and Plastics, A.C.S. Meeting, September, 1956. (4) J. J. Tazuma and H. K. Latourette, ref. 3. ( 5 ) J. Burkus and C. F. Eckert, J. Am. Chem. Soc.. 80,5948 (1958).

0.00115

Assuming that reaction of each -NCO group in eq. 1 with a primary alcohol obeys second-order kinetics17one obtains the following set of independ-

TOLUEP\’E-8,4-DI1SOCYANL4TE

Although numerous studies of isocyanate concentration vs. time for the reaction of toluene2,4-diisocysnate (TDI) with a primary alcohol have appeared in the l i t e r a t ~ r e , l -no ~ analysis of the data

meq./ml.

0.00118

NCO

NH-CO-OR

I

SH-CO-OR (6) A. Di Giaeomo, J . Poly. Sei.. in prem. (7) R. G. Arnold, et al., Chem. Revs., 67, 47 (1957).

April, 1961

697

;?;OTES

TABLE I1

--

9%-NCO Consumed

Calcd. 1

0 11.o 19.2 25.4 30.3 34.2 41.5 46.4 49.9 52.6 54.8 56.5 59.2 61.3 64.3 66.6 68.4

1.0000 0.7819 .6266 .5117 .4238 .3551 .2370 .1645 .1174 .0854 .0631 .0472 .0273 .0162 .0062 .0026 .0011

--

-

CALCULATED MOLESOF (T), (Q), ( R ) AND (S) us. OJ,-NCO CONSUMED (Q1-y -(It)c 0) Calcd. 2

Calod. 1

Calcd. 2

Calcd. 1

Calcd. 2

Calcd. 1

1.0000 0.7827 .6279 .5131 .4254 .3567 .2386 .1658 .1185 ,0864 .0639 .0479 .0278 .0166 .0064 .0026 .0011

0.0000

0.0000 ,0032 .0106 .0200 .0303 .0412 ,0682 ,0938 .1176 .1397 .1599 .1787 .2126 .2422 .2924 .3340 .3697

0.0000 ,0381 .0622 .0796 ,0880 ,0944 .lo08 ,1000 ,0959 .OR04 ,0842 ,0783 .0667 .0565 .0405 ,0292 .0214

0.0000 ,0233 .0372 .0456 .0504 ,0531 ,0541 ,0514 .1473 .0427 .0382 ,0341 ,0259 .0211 .0129 ,0080 ,0051

0.0000 .1770 .3011 .3914 .4591 .5109 .5960 .6438 .6711 .6863 .e940 .6968 .6938 .6848 .6601 .6338 .6084

Constants (l./eq.-hr.)

Cnlcd. 1 Calcd. 2

.0030 ,0101 .0190 .0291 .0397 ,0663 .0918 .1152 .1378 .1585 .1777 .2118 .2424 .2931 .3345 .3691 kl

8.54 0.21

kr

ks

4.00 6.00

1.92 1.20

‘k

0.53 0.63

(S)------

Calcd. 2

0.0000 .1908 .3243 .4214 .4938 .5491 .6302 .6889 .7166 .7313 ,7378 .7392 ,7328 .7202 .6883 .6554 .6242

Std. deviation, meq./ml.

0.00118 0.00115

Kotwithstanding the uncertainty in the values of the rate constants, the values, which were of primary concern, of (T) and of (Q) as a function of the % -XCO consumed were computed with satisfactory accuracy by this procedure. The nearly equal values obtained with each of two sets of rate constants are shown in Table 11. Values of thi: intermediates, (R) and (S), which are subject l o a greater uncertaintry, are also tabulated. The assistance of Dr. A. L. Squyres in running the computer is gratefully acknowledged. (IO) D. M. Simons and R. G . Arnold, J. Am. Chem. Soc.. 78, 1658

0

0

1

2

3 4 5 6 Time, hr. Fig. 1.-[NCO] us. time. Reaction of T D I with n-butyl alcohol in xylene solution a t 80”.

ent equations (where each rate “constant” depends slightly on hydroxyl8 and urethang concentration). -d(T)/dt = ( h Ics)(T)(OH)

+

-

-d(R)/dt = MR)(OH) kr(T)(OH) -d( S)/dt = kr( S)( OH) ki( T)( OH) 2(T) (R) (S) = total-NCO

+

+

(2)

Equations 2 were fitted to the -NCO vs. time curve with the aid of a Bendix G-15 computer. This automatic curve fitting procedure was found to be insensitive to the value of kz as may be seen from Table I, columns 3 and 4. Nonetheless, the approximate values of kl = 8.9, kz = 5, ks = 1.6, kq = 0.6 all in liters/eq.-hr., are an improvement over reported values derived from similar data by analyses based on only two rate constants. Assignment of the largest value to the para- rather than the ortho-isocyanate group was based on studiesl0 which demonstrated unequivocally its greater reactivity. (8) J. W. Baker, et al., J . Chem. Soe., 9, 19, 24, 27 (1949). (9) J. W. Baker and J. B. Holdsworth, J . Chcm. SOC.,713 (1947).

(1956).

SPECTROPHOTOR’IETRIC EVIDENCE FOR INTERACTION BETWEEN CHLOROFORM AND MONOETHYLAMINE BY LEONSEGAL Southern Reoional Research Laboratorv,l New Orleans, Louiszana Received October 6 , 1960

I n an earlier paper by Segal and Jonassen2 evidence for interaction between chloroform and ethylamine was reported, based on heat of mixing, distillation, and water-washing of chloroform-amine and hexane-amine mixtures. A change in refractive index was also reported which seemed to indicate that a 1:l complex was formed in the chloroform-amine system. From the data a t hand the authors concluded that a hydrogen-bonded complex was formed, where the bonding was through the C-HcN bond. More recent evidence for the hydrogen bonding power of chloroform is discussed by Pimentel and 1Ll~Clellan.~An infrared spec( I ) One of the laboratories of the Southern Utilization Research and Development Division, Agricultural Research Service, U. 9. Department of Agriculture. (2) L. Segal and € B. I.Jonaasen, J . Am. Chem. SOC., 74, 3697 (1952).

Vol. 65

698

WAVELENGTH (MICRONS)

Fig. 1.-Infrared

spectra of ethylamine in carbon tetrachloride and in chloroform.

t m m of ethylamine in chloroform has been published in connection with studies on gossyp01,~but the bands are poorly resolved and no cognizance was taken of any possible solvent-solute interaction and its effect on the infrared spectrum. This paper reports infrared spectrophotometric evidence for an interaction between chloroform and ethylamine which further confirms hydrogen bonding of the type C-HcN. Experimental Anhydrous ethylamine was drawn as a vapor from a cylinder, condensed in an all-glass apparatus, and stored in a refrigerator over sodium hydroxide pellets. Samples titrated with standard hydrochloric acid indicated a punty of 100%. Carbon tetrachloride and chloroform were of spectrophotometric grade. Infrared absorption spectra were obtained at a rate of 0.5 p per minute with a Perkin-Elmer Model 216-double-beam infrared spectrophotometer m t h sodium chloride prism, operated with the following settings: resolution, 927; gain, 6; response, 1; and suppression, 3. Cell thickness was 0.48 mm., and solution concentrations were 20.8 g./l. (0.46M ) .

Results and Discussion I n the infrared spectrum of ethylamine in carbon tetrachloride where association of the amine does not occur, only the absorptions of the NHz group are of interest here. These have been identified after extensive consultation of several works6 as (3) G. C. Pimentel and A. L. McClellan, “The Hydrogen Bond.” Reinhold Publ. Co., New York, N. Y., 1960, pp. 196-199. (4) R. T.O’Connor, P. Von der Haar, E. F. DuPre, L. E. Brown and C. H. Pominski, J . Am. Chem. Soc., 78, 2368 (1954). (5) Use of a aompany and/or product named by the Department does not imply approval or recommendation of the product t o the exclusion of others which may also be mitable. (6) R. N. Jones and C. Sandorfy in “Chemical Application of Suectroacopy,” ad. bv W. Weat. Interscience Pitbl., New York, N. Y., 1956, pp. 247-680: H.M. Randall, R. G. Fowler, N. Fuson and J . R. Dangl, “Infrared Determination of Organic Btrncturei.” D. Van Noatrand

being the unbonded N-H and NH2 stretching of a primary amine at 2.94 p, the NH2 scissoring mode at 6.15 p, the NH2 wagging mode at 8.92 p , the C-N stretchings of a primary amine at 9.24 and 9.51 p, an N-H bending at 10.39 p, and NH2 deformations a t 11.24 and 11.95 p (Fig. 1). The spectrum of ethylamine in chloroform shows shifts in the locations of several of these absorption bands and the presence of new bands. These shifts, and particularly the presence of the new bands, indicate the presence of hydrogen bonding in this system. The strong absorption now appearing a t 3.10 p indicates a strongly bonded NH2 group. The new mode at 4.00 p definitely fixes the bonding as that of C-H+N. An amine band in this region does arise from an amine hydrochloride where the group =KHz+, and >NH+, but no hydrochlois -”a+, ride is present here. Gordy,’ after observing the appearance of an absorption band at 4 p in systems involving chloroform, bromoform, pyridine, apicoline, and piperidine, concluded that it was an N H vibrational band resulting from a hydrogen bond formed by sharing of the proton of the C-H haloform group and a N of the amine. He also concluded that this band is the most definite evidence yet for the existence of the hydrogen bond. Thompson and PimenteP have more recently reported the appearance of a strong band in this region for the solid obtained by freezing a gaseous mixture of chloroform and triethylamine at 77°K. Co.. New York, N. Y., 1949; L. J. Bellamy, “The Infra-red Spectra of Complex Molecules,” John Wiley and Sons, New York, N. Y., 1954; N. R. Colthup, J . Opt. SOC.Am., 40, 397 (1850). (7) W. Gordy, Nature, 142, 831 (1938); J . Cham. Phys., 0 , 163 (1839). (8) W.E,Thompson and G. C. Pimental, 2. 18lsktruehsm., 64, 748 (1960).

NOTES

April, 1961 However, they attribute this band, which shifted to 5.32 p when CDC13 was substituted for CHC13, to the C-H bending mode of hydrogen-bonded chloroform. Other changes in the ethylamine spectrum arise from this association between chloroform and ethylamine. The NH2 scissoring mode has shifted from 6.15 to 6.30 p. I n the 8-p region now appears a strong, broad band whose origin is quite puzzling. The NH2wagging mode has shifted down to 8.85 p , the N-H deformation a t 10.39 has shifted down to 10.00 p , and the NH2 deformations a t 11.24 and 11.95 U, have shifted down to 11.12 and 11.64 p. Two new absorptions now appear in the C-N stretching region, located at 9.12 and 10.72 p. The 9.12-p band can be reasonably assigned to a new NH2 deformation as it seems to be a harmonic of the bands at 3.10 and 6.30 p , but assignment of the 10.72-p band cannot be made yet. Acknowledgment.- The author is indebted to 3lis E. R. NIcCall for the infrared spectra.

2

699

3

I

1

1

1

1

I

I

I

4

5

6

7

8

9

IO

I1

PH.

Pig. 1.-Effect of a tyrosyl-carboxylate ion hydrogen bond on the sharpness of the transition of Fig. 12 of the previous paper.3

T H E SHARPNESS OF THE TRSNSITION I N ItEVERSIBLE PROTEIN D E N ATUR ATIOW in the intermediate pH-region by one tyrosyl€1. ,4.SCHERAO.4, R. A. SCOTT,' G. I. LOEB,A. ~ A K A J I & ~ A carboxylate ion side-chain hydrogen bond. Thus, AUD J. HERIIANE, JR. it can be seen that side-chain hydrogen bonding increases the sharpness of the transition. As shown Department of Chemistry, Cornell Unzwrszty, Ithaca, N . Y . previously, it also increases the transition temperaRrcewed Octobei 7, 1960 ture. This is an addendum to a recent paper3 on reIt is easiest to see the effect of cross-linking a t a versible protein denaturation, and makes use of the pH \There AHOH and AFOH are zero, i.e., where no model presented therein. The purpose of this side-chain hydrogen bonds exist. Under these note is to show how permanent cross-links and side- conditions, the transition temperature (from eq. 33) chain hydrogen bonds affect the sharpness of the i9 transition in reversible denaturation. (TZ - 4) AH'rea The observed standard free energy of denaturaTtr AS% tion is given by eq. 19, 33 and 34 for the model and used previously.

+APH

AFoobsa = (n - 4)AHores- TASOo 48'0 = (n - 1)AS'res ASOJv A F O= ~ - R T ~In (I - z,,)

+

(33) (34) (19)

Thus, since AXQ, is negative15Ttr is raised by crossThe fraction of the molecules denatured is given3s4 linking, as pointed out by Schellman4 and F10ry.~ (See also Table I of previous paper.3) Further, by eq. 36. (da 'dT)T,, decreases with increasing cross-linking, = [I $- eAF'ol,sd/RT]-l (36) i . e . , the transition becomes broader. For example, Defining the sharpness4 of the transition as d a l d T in the model used p r e ~ i o u s l y the , ~ value of (do!/ at 1he transition temperature, T t r , where A F n o b 3 d dT)Ttr is decreased from 0.0250 to 0.0178 by the = 0, we obtain cross-linking. (n - 4)AHor,,-I- AHOE The effect of increasing n was pointed out by rt, = 4RT2tr S ~ h e l l m a n . ~I n the absence of cross-links and sidewhere AHOH is given by eq. 20.3 The terms AFOH chain hydrogen bonds large n leads to and A H o arise ~ from the side-chain hydrogen bonds. Ttr AHOreJAS4es The pH-dependence of (da 'dT)Ttr arises from and to increased values of (da/dT)T,,. the pH-dependence of AHOH and of T t r . This is Proteiiis may, therefore, differ from synthetic illustrated in Fig. 1 for the model previously de- polypeptides, as far as the sharpness of the transi.;cribed,3 in which the crystalline unit is stabilized tion in reversible denaturation is concerned, if (1) This work was supported b y research grant No. E-1473 from cross-links and side-chain hydrogen bonds are t h e National Institute of 4llergy and Infectious Diseases of the present in the protein but absent from the polyYational Institutes of Health. Public Health SeiTice, and by grant peptide. Obviously, any other phenomena which C, 0401 from the National Science Foundation affect the enthalpy or entropy of the transition ( 2 ) U. S. Public Health Service Pre-doctoral fellow, National Heart Institute, 1959-1960. (such as breaks in the helix, hydrophobic bonds, (d) H. A. Scheraga, J Phys. Chem., 64, 1917 (1960). etc.) may also affect the sharpness. (4) J . A. Sohellmen Compl. rend. trau. lub. CarZaZ~eru,8tv. chtm,, 19,

(3

230 (1955).

(a)

P.

J. Flow, J . A m . Chtm. SPC., 78, 6222 (1856).

NOTES

700 PROTONATION I N N-METHYLACETAMIDE BY GIDEONFRAEXKEL~

Gates and Crellin Laboratoraes of Chemzstry, Calafornza Instztute of Technology, Pasadena, Calzfornza

AHARONLCJEWENSTEIN ARD N M R Laboratory, Reazmann Indtztute, Rehouoth, Israel

SAULMEIBOOII Bell Telephone Labs., Murray H z l l , S. J. Recewed Septembei 81,1960

I n a recent note by Spinner3 a discussion of the site of protonation in amides is given. In that note it is suggested that the n.m.r. evidence for 0protonation in K-methylacetamide and other amides is inconclusive. As we have shovn prev i o u ~ l y ,the ~ , ~N-methyl proton resonance of this molecule in very acidic solution consists of a symmetrical doublet presumably due to spin-spin interaction with a single N-hydrogen. Spinner, however, interprets this doublet as being due to non-equivalence of the methyl groups in different isomeric forms resulting from hindered rotation. We wish to point out that the symmetry of the doublet makes such an interpretation highly improbable as the equal intewity of the components would lead to the conclusion that the two isomeric forms have very nearly the same energy. Two experiments were performed t o shorn unambiguously that the splitting of the 3-methyl resonance in protonated N-methplacet~amide is due t o spin-spin interaction rather than a chemical shift. 1. The N-methyl resonance of a solution of Kmethylacetamide in 100% HzS04 consists of two equal lines while in 100% DzSOlthe same resonance consists only of a single line. 2. The separat#ion of the RI-methyl doublet in protonated N-methylacetamide is independent of the magnetic field (3.7 f 0.3 c.p.s. both at 33 and 60 Me.). It is therefore clear that K-methylacetamide protonates predominantly on oxygen. (1) Contribution No. 2580. (2) Department of Chemistry, The Ohio State University, Evans Chemical Laboratory, 88 W. 18th Avenue, Columbus 10,Ohio. (3) E. 6. Spinner, J . Phys. Chem., 64,275 (1960). (4) G. Fraenkel and C. Niemann, Proc. Nail. Acad. Sci. (U.S.), 1 4 , 688 (1958). ( 5 ) A. Berger, A. Loewenstein and 8. Meiboom, J. A m . Chem. SOC., 81, 62 (1959).

ANALYSIS OF T H E INTRINSIC VISCOSITY OF A POLMMER UNDERGOING SIMULTANEOUS CROSSLINKING A4ND DEGRADATION BY MALCOLM DOLE Department of Chemistry, Northwestern University, Euanston, Illinois Receiaed October 18, 1961)

The effect of simultaneous crosslinkine and derradation on the intrinsic viscosity of a” polym& has recently been considered theoretically in this country by Kilb,’b2 using the theory of Zimm and Kilb,3 and in Japan by S a i t ~I: , ~okutis and Inokuti

Vol. 65

and I