THE VJSCOSlTY OF CERTAIN AMIDES. 1935CCV.-The Viscosity of Certain Amides.By ALBERT ERNEST DUNSTAN and ALBERT GEORGE MUSSELL.THE peculiarly constitutive nature of viscosity confers on it con-siderable value in discriminating between two possible types ofstructure. The amides may react as though they possessed theOH formulae R*CO-NH, or R*CqNH (Tafel and Enoch, Bet-., 1890,23, 1550; Lander, Trans., 1903, 83, 418), although by comparisonwith a genuinely hydroxyiminic compound, such as glycollimino-hydrin, Hantzsch and Voegelen (Ber., 1901, 34, 3142) regardR-CO-NH, as the correct formulation for the amides. From aphysical point of view the amides are undoubtedly amociate1936 DUNSTAN AND MUSSELL :(Auwers, Zeitsch. physikal. Chem., 1893, 12, 689; 1894, 15, 33;1897, 23, 449; 1899, 30, 521), and the fact that they form additivecompounds (Titherley, Trans., 1901, 79, 413) indicates that thegroup *CO*NH, possesses considerable residual affinity.I n furthersupport of the usual amidic structure may be cited the work ofHantzsch and Dollfus (Ber., 1902,35,226) and Schmidt (Ber., 1903,36, 2459). Fawsitt (Proc. Roy. Soc, Edin.., 1904, 25, 1, 51) foundthat the fatty amides in aqueous solution were non-electrolytic, andgave normal depression of the freezing point. The same chemist(Electrochemist and Metallurgist, 1904, 3, 664) determined theviscosities of a few amides in aqueous solution, and showed thatthe viscosity increased with increasing molecular weight.Meldrum and Turner (Trans., 1908, 93, 876), in an ebullioscopicexamination of a considerable number of amides, found that 90 percent.of those used were associated in benzene, 80 per cent. inether, 80 per cent. in chloroform, 10 per cent. in acetone, 45 percent. in water, and probably 20 per cent, in ethyl alcohol. Theyconnected these results with the dielectric constants of the solvents,and considered that the carbonyl group or the nitrogen atom wasthe centre of association.Turner and Merry (Proc., 1910, 26, lB), using Ramsay andShields’ method, state that all the amides investigated by them areassociated.The object of the present work was t o investigate the viscosity ofthe amides with the view of detecting any hydroxylic nature, forsuch structure has been found to affect very materially this propertyO H in aqueous solution.Again, if the amides possessed the -C‘qNHgroup, then, in pyridine solution, salt formation would probablyoccur, with enhanced viscosity. Further, the viscosity of the amidesin the free condition had not been pr$viously investigated, and itwas of interest to determine the equivalent viscosity as being likelyto throw light on their relative degrees of association.EXPERIMENTAL.The viscosity of a large number of amides, both fatty andaromatic, has been determined f o r the free substances and foraqueous and pyridine solutions. The materials used were a,s follows :Carhami&.-Kahlbaum’s purest, 111. p. 132O.Formamide.-Kahlbaum’s, redistilled under diminished pressure.2.916 grams, boiled with alcoholic potash (70 C.C.of 1*155N),used, after boiling, 16-25 C.C. of N-sulphuric acid, whenceMCO-NH, = 99.7 per cent. Boiled with decomposition at 204O1758 mm. ( 2 0 8 O corr.); b. p. 141°/65 mm., 136O/50 mm. BruhTHE VISCOSITY OF CERTAIN AMIDES. 1937(Zeitsck. physikal. Chem., 1894, 16, 214) gives b. p. 111-112°/14 mm.A cetamide.-Kahlbaum's, m. p. 8 2 O , redistilled b. p. 215*5O/749 mm.Propionamide.-Kahlbaum's, m. p. 80-81O.n-Butyramide.-Kahlbaum's, m. p. 116'.Thioca~b anzide.-Schuchar dt's, m. p. 149O, crystallised fromd cetaniZide.-Prepared from aniline, recrystallised several timesBenzarnide.-Kahlbaum's, recrystallised from hot water, m. p.BenzaniZide.-Schuchardt's, m. p. 160°, crystallised from alcohol.T~~~ocarbaniZide.-Schuchsrdt's, recrystallised twice from alcohol,m.p. 153O.Metlz ylacetaniZide.-Prepared from methylaniline, recrystallisedfrom alcohol, m. p. lolo, b. p. 2 3 7 O .DiphenyZca;l.bam,ide (carbanilide).-Prepared from ketobenz-oxazole and aniline, crystallised from much alcohol, m. p. 234' (insealed tube).water.from water, m. p. 1 1 2 O , b. p. 283O.128O.Uretlmne.-Kahlbaunl's, m. p. 51°, b. p. 179O/749 mm.Cyanuric d cid.-Kahlbaum's, recrystallised from hot water.FormaniZide.-Schuchardt's, recrystallised from hot aqueousPhthaZuniZ.-Schuchardt's, m. p. 205O.Plttha.limide.-Schuchardt's, recrystallised from alcohol, m. p.2 30°.Sol ZI en t 8.-Conductivi t y water, p yridine p art>ly from Kahlb aum,dried over potassium hydroxide, b. p. 114--116O/742 mm.(corr.),partly from crude coal-tar pyridine fractionated with a long rod-and-disk column, b. p. 116O.Magnitude of Experimental Er-ror.-Densities were taken inSprengel pyknometers of 2 C.C. capacity. Two determinations weremade, and these usually agreed within a milligram. The densityerror is not greater than 0.05 per cent. Times of flow were takenuntil about five concordant results were obtained agreeing withinone second; as an example the following may be quoted :alcohol, m. p. 46--47O, b. p. 271O.Butyramide, 16.88 per cent. in water.Greatest mean error=O*17 sec.3' 27'4''; 3' 27'4'' ; 3' 27.6'' ; 3' 27'5" ; 3' 27.7" ; 3' 27.6''. Average=$ 27*53^.The viscosity error may be taken as k0.1 per cent.To obtain the equivalent viscosities, two or more determinationsA curve was then were made at different (low) concentrations1938 DUNYTAN AND MUSSELL :drawn practically linear near the origin.By arranging theequivalent concentration so low its mol. wt.j20, it was possible tointerpolate from this linear portion, and so obtain a more correctvalue than could be observed directly.The viscosities at high temperatures are not so accurate, since it isvery difficult t o secure effective thermo-regulation. An examplewill illustrate the method :Temperature Temperaturea t start. at end. Time of flow.104-8" 105.0" 5.28105'0 103.8 5-31104.8 104-6 5.27105'0 105.7 5.27L 2 v Avernge temperature 104 '8The viscosity error is probably k0.5 per cent.The following tables show the resu1t.s obtained; the figures in thefirst column representing percentages of the substance in the solvent.Average 5 -282TABLE I.Amides in Aqueous Solution at 25O.(1).Carbamide-Density.1 *02 1 -0008-13 1.01811.89 1 '02915.47 1'03323.12 1.05933'28 1.08738.13 1.10246.18 1.125(2). Formanzide-l * i O 0.999940'22 1-05452.2 1-06975.42 1.11579.15 1.118100.0 1.132(3). Acetamide-0.78 0.99775-82 1 -00117-69 1.00825.95 1,01537'21 1,0226926 1 -038Viscosity.0*008950.009390.009690 *010350.010880.01 2520 '013480 015610 -009210'011820.012830'021010 '0 22 T 5 ;:;;;;: }0.008900.009910.01 2320'014580.018750'04442(4). Propionavtide-Density.1.245 0'99745.65 0 998411 071 1 .on022-18 1.00641 -48 1.01170.65 1-011( 5 ) .mBt6tyramide-1 *02 0.99728-11 0'997416.88 0.998417'92 0.9987(6). Formic acid-3 05 1 0 0 427-38 1.06073-55 1.161100 0 1 '209Viscosity.0.009190*010250.01 1880.015360'030300.059800*009060'011220 *0145 80 '01 5 110*009070*010410 -0 1 3790.01 5 80* From two specimens from KahlbaumTHE VISCOSITY OF CERTAIN AMIDES.Amides in Pyridine Solution at 25O.Thiocarbamide-Density.5 -52 0.994912.57 1'019Aeetamide-3.87 0.979416'27 0.9959Benamide-5'16 0.983912-46 0,9971Benaccnil ide -5'40 0-98299.02 0'990212-75 0'9964Thwcarbanilide-7-40 0.991914.51 1 *005Met h y hcetamide-6.59 0.978611.2 0.981916-44 0.9844Carbanilide-5.69 0.98417-19 0.9862Urethane-9-09 0'983214'96 0 9899Viscosity.0.012510*020190.009870-013810*010550 *013 3 10 '0102'10 -01 1210 '012520*011590*014160.009670'020220 *O 108 20*010450.010790.010700 .o 1202Propioxamidc -Density.7'92 0.976923.75 0.981713'66 o m 9Cyanuric acid-3.20 0'9874Forwunilide-10.32 0-990319'14 1'004Phtha Zad-3-55 0.9821Phtha Zinaide-5.05 0987611.93 1.006Formanzide-7.77 0-987111'10 0993517.12 1*G0075.65 0.98148 $1 0.9826Acetamide-Carhamide-0'91 0.9802PyridilLe *-100-0 0.97461939Viscosity.0 *0104 70.012080*016150*010380.010890 '01 3 350.0096060.0098780.01 1820'010640.01 1600.013650.010050 010880-009390.00884* Hartley, Thomas, and Applebey (Trans., 1908, 93, 544) give 0.00885.The Viscosity of the Amides at High Temperatures.To obtain comparable information as to the relative state ofassociation of the amides, it was necessary to work at somewhatelevated temperatures, seeing that propionamide melts at 81° andacetamide at 82O.A description of the apparatus used may be ofinterest.A large beaker filled with cylinder oil of high flash-point wassupported on wire gauze and jacketed with asbestos paper, throughwhich two opposite longitudinal slits were cut for observations oftmhe viscometer. Through the metal lid passed a thermometergraduated in fifths of a degree, a stirrer connected with a Henricimotor, and the viscometer. The latter was designed with theview of increased accuracy in filling.The usual method (Ostwald,Physico-Chemicd Measurements, 1894, 163) is to run in a know1940 DUNSTAN AND MUSSELL :volume of liquid from a pipette. This is obviously impossible whenworking with a substance which is solid at the ordinary tem-perature, unless the pipette can be kept at a sufficiently high tem-perature, so that the instrument shown in Fig. 1 was used. It isof the Ostwald type, and merely possesses two etched lines at theUFIG. 1.same level on the same limb. The viscometeris cleaned and dried, and the compound underobservation is distilled into it. Where distillationis impossible, the melted substance must be filteredin. Unless filtration or distillation is resorted to,it is practically impossible to secure freedomfrom particlm of dust, which will inevitablychoke the capillary. This is the most fertilesource of error in viscometry.The instrumentfilled just above the marks is placed in the oil-bath, and levelled by means of a gravity bobhung from the supporting clamp. After remain-ing for ten minutes to attain the bath tem-perature (a thin flame obtained from a Bunsenburner after unscrewing the chimney, and con-trolled by a long lever on the tap, gives excellenttemperature regulation), the liquid is adjusted tothe marks by a capillary pipette, and times offlow are taken and averaged. Absolute densitieswere measured in a 10 C.C. bottle-shaped pykno-meter filled with fused compound. A correctionwas applied for the known coefficient of expansionof the glass, the volume being determined at2 5 O and 4 5 O by the water content.following way :The viscometer wits calibrated with ethylene dibromide in theLog viscorneter constant = log r] - log time - log density.At 105O,Time of flow at 105O= '77.4 secs., and at 120°= '71.9 secs.Density at 105" = 2.009, and at 120°= 1.979.whence log K,, = 5.6138, and log K,, = 5.5965.Hence the viscosity of a compound at 120° or 1 0 5 O is obtainedfrom the equation q = K x time x density.The ethylene dibromide used for this purpose boiled at 129'5O/749 mm.for ethylene dibromide = 0.00639 (Thorpe and Rodger).A t 120°, q 9 9 ,, =0.00562 ,, 9 TEE VISCOSlTY OF CERTAIN AMIDES. 1941TABLE 11.Amides at High Temperatures.105" :Formamide ............Acetamide ............Propionamide ......Urethane ............120" :Formamide ............Acetaiiiide ............Propionamide ......Urethane ............Formzlnilide .........Acetanilide............Methylacetanilide ...v.0.007680'01320.01270 -00 9 160.006590*01060.01030'007 150.01650'02220'00818d.1.0610.9800.9331 *0051.0500.9670-9250.9911.0761 '0310.977Mol. wt. q x 106/M.V.45 18159 21973 16289 103.545 15459 17473 13689 77'8121 147135 170149 53 -6Discussion of Results.The Amides in Aqueous Solution.Formamide, acetamide, propionamide, n-butyramide, and carb-amide were examined in aqueous solution. I n Fig. 2 are plottedthe viscosity-concentration curves of the above amides, togetherwith those of formic acid and methyl alcohol for comparison.In connexion with forma.mide, Walden (Zeitsch.Electrochem.,1908, 14, 718) found a high value for the latent heat of fusion(50.4, acetamide being 69.4). Noreover the expression for l ' eformamide is 8.31, and for acetamide 11.51, giving the coefficientsof association 1-62 formamide, and 1-17 acetamide.The rapid increase of viscosity with increasing amide concentrationis remarkable, and indicate very considerable aisociation on thepart of these compounds. Comparing formamide and formic acid,which yield curves of similar type up to a concentration of 30 percent., it is evident that beyond this limit the formamide curve risessteeply until the relatively high viscosity 0.0326 is reached for thepure substance.x 106 has a seriesconstancy; for example, for alkyl chlorides it is 37.4, and forketones 43-3 (Dunstan and Thole, J .Chim. Phys., 1909, 7, 204),whilst for associated compounds values in great excess of these areNow it has been shown that the quantity M. ??q x 106obtained, For formamide, - - - 682, and for formic acid this M. V.expression = 415. There is thu? little doubt that formamide in thefree state is extremely associated, and although these numbers d1942 DUNSTAN AND MUSSELL :not give an exact value for the degree of association, yet it ispossible, qualitatively, to obtain a very fair idea of the relativeextents of the molecular complexity. By measuring off the curves,the values of the viscosity coefficients at equiinolecular concen-trations, the following numbers are obtained :Mol.wt. Amide.87 Rutyramide ..................59 Acetamide., ..................60 Carbarnide ....................45 Formamide ..................46 Foriiiic acid ..................32 Methyl alcohol ...............73 Propionamide ...............FIG. 2.0*04000.0320 55-2**9) 22 0.02407hi,0*01600~00800Equir. d e n t viscositya t mol. wt./8 per cent.12211010294909299The above equivalent viscosities in aqueous solution illustrate thedissociating action of the solvent. Formamids is almost completelybroken down, since its equivalent viscosity is nearly identical withthat of formic acid, which does not exist associated with water inaqueous solution.Methyl alcohol, on the other hand, is mostprobably associated with the solvent. There is a steady increasTHE VISCOSITY OF CERTAIN AMIDES. 1943in viscosity as the molecular weights of the amides become greater.According to Meldrum and Turner (Zoc. c i t . ) , the amides areassociated in aqueous solution with the possible exceptxion ofcarbamide (formamide was not examined by them).The position of carbamide is interesting, seeing that the curvelies almost exactly midway between those for acetamide andformamide. Although of nearly identical molecular weight, theequivalent viscosity is considerably lower than that of acetamide.Here, however, another consideration should be advanced.Viscosity is not entirely a matter of molecular mass or molecularvolume.What may be termed the molecular shape or symmetrycannot be ignored, and in accordance with this view it is foundthat the viscosity of iso-compounds differs from that of the normalisomerides. It may happen that carbamide is a more symmetricalcompound than acetamide, in which case, although the degree ofassociation might be the same, the molecular viscosity would be less.The Amides in Py?.idine Solution.The choice of pyridine was made in the expectation that if anytendency existed towards the structure *CGNH on the part of theamides, it would be developed by the well-known basic propertiesof this solvent. With the possible exceptions of thiocarbamide,cyanuric acid, and thiocarbanilide, this hope was not realised, butat the same time it will be evident from the curve that anapproximate separation of the amides in the order of theirmolecular complexity is achieved.Pyridine is a dissociating solvent(von Laszczynski and von Gorski, Zeitsch. EZektroc7tem., 1897, 4,299), and, like water, tends to break up the aggregates presented toit. It is particularly noteworthy that the amide the molecularviscosity of which was smallest in the fused state, methylacetanilide,affords the curve with the least upward tendency, that is, the leasteffect on the viscosity of the solvent. For the sake of comparison,viscosities of the solutions have been measured at equivalent con-centrations, mol. wt./20 per cent.OHThe order then becomes:Mol.wt.4559149897312160123At mol. wt./20Amide. per cent.Formamide ......... 0 '0093Acetamide ......... 0 '0095Methylacetanilide . 0.0097Urethane ............ 0.0097Propionamide ...... 0 -0097Forrrianilide ......... 0 -0098Carbamide ......... 0'0102Ph thalimide ......... 0 -0 102Mol. wt.12113576.212223197129228At mol. wt.120Amide. per cent.Benzamide ......... 0 -0 108Acetanilide ......... 0 -0108Thiocarbamide.. .... 0*0109Carbanilide ......... 0 -01 10Phthalanil ......... 0 *0114Benzanilide ......... 0-0116Cyanuric acid ...... 0'0120Thiocarbanilicle ... 0.0131944 POWER ANb ROGERSON!The value for carbamide is not of the same order of accuracy asthe other amides, seeing that a concentration of 1 per cent. was thehighest obtained. If it be granted that the amides in certaininstances do react as acids, then the high molecular viscosities ofthiocarbanilide and benzanilide would be explained, and thisexplanation is at least a possible one. But at the same time thenon-acidic nature of the fatty amides is emphasised. The valuesin pyridine solution are complicated therefore by two causes: (1)the dissociation more or less complete suffered by the dissolvedsubstance; (2) the effect of m y acidic nature of the solute. Whenthe equivalent viscosities in pyridine are compared with those inaqueous solution, it is again noticed that formamide has the lowestvalue; this, of course, may be in each case due to the fact thatit is the most easily dissociated, but it must be emphasised oncemore that the question of molecular symmetry cannot be ignoredin drawing comparisons based on viscosity determinations. Weintend extending this research to the investigation of the generalquestion of solutions in the amides, and particularly to the questionof relative viscosities at corresponding temperatures,In conclusion, we desire to express our gratitude to Mr. W. E. S.Turner for many useful suggestions and criticism, and to theResearch Fund of the Chemical Society for a grant in aid of thework.PHYSICAL CHEMICAL LABORATORY,E A S T HAM TECHNICAL COLLI4CGR