Janl-arv 1, 1920.] 



THE INDIA RUBBER WORLD 



217 



for several niiiiutes with boiling water. The solution is fil- 

 tered tlirough a folded filter paper and evaporated to a vol- 

 ume not to exceed 150 cc. After it has cooled it is poured 

 very slowly into a 5 per cent solution of pure tannic acid. 

 In the presence of glue a cloudiness will appear at lirst and 

 finally a precipitate as more of the glue solution is added. 

 Large percentages of glue give heavy curdy precipitates 

 while small quantities give decided cloudiness. In the ab- 

 sence of glue no permanent precipitate or cloudiness will 

 appear in the tannic acid solution as the liquid is added to it. 

 In order to test the reliability of this qualitative proce- 

 dure in detecting varying quantities of glue, mixings were 

 made up which contained from 0.9 per cent to 10.0 per cent 

 of glue, the balance of the compound being rubber (40 per 

 cent"), zinc oxide, magnesium oxide, gas, black, mineral rub- 

 ber and sulphur. They were run through the above proce- 

 dure with the following results: 



PEKCENTACE .\pPE.\RA.NCE OF THE SOLUTION ON .\dDITION TO 



OF Glce. Tannic Acid. 



0.00 Solution remains perfectly clear. No cloudiness. 



0.9 Distinctly cloudy. 



1.7 Cloudiness much denser than the above. 



2.6 Cloudiness almost opaque. Precipitate. 



3.5 .Absolutely opaque. _ Precipitate. 



4.3 Fairly heavy precipitate.' 



6..' Heavy precipitate. 



so \'trv heavy precipitate. 



10.0 Very heavy precipitate, which showed its curdy 



character very plainly. 



The ol)scrvations given above demonstrate conclusively 

 that even the smallest practicable amount of glue can readily 

 be detected by means of the procedure recommended here. 



Many experiments were carried out in order to ascertain 

 whether ingredients other than glue would dissolve in water 

 and precipitate the tannic acid solution and in this way be 

 mistaken for glue. It was found that magnesium oxide and 

 calcium oxide dissolve to a considerable extent in hot water 

 after they have been treated with cresol and the water solu- 

 tion that is obtained precipitates tannic acid when the re- 

 agent is not in excess. However, these precipitates are 

 readily soluble in excess of tannic aid, while the glue-tannate 

 precipitate is insoluble. Therefore it was recommended in 

 the method given above that the solution to be tested for 

 glue be added to the S per cent tannic acid solution. In this 

 way the magnesium and calcium precipitates will be dissolved 

 immediately as they appear and will become permanent only 

 when a considerable quantity of their solutions has been 

 added. In this way there will be no chance of mistaking 

 them for glue since the glue precipitate will appear imme- 

 diately and will be permanent, even when only a small quan- 

 tity of its solution has been added. We have been unable to 

 find anything else which would be present in this solution 

 and would give a permanent precipitate when added to tannic 

 acid. There is therefore no danger of misinterpreting the 

 tannic acid precipitate; when it is obtained the presence of 

 glue in the rubber mixing cannot be questioned. 

 QUANTITATIVE METHOD. 



However, since the method is not quantitative, after the 

 presence of glue has been proved, its amount must be de- 

 termined. The only means we have found by which this 

 can be done is to determine the nitrogen by the Kjeldahl 

 procedure and calculate the amount of glue from it. This 

 was found to be satisfactory and results were obtained which 

 were usually 0.1 per cent to 0.2 per cent higher than the true 

 value. The method is practically the same as that followed 

 out in most laboratories. It is as follows: 



Extract a two-gram sample with acetone for six hours, 

 and then with chloroform for three hours and dry. Transfer 

 the sample to a 7S0-cc. Kjeldahl flask. Add 25 to 30 cc. of 

 concentrated sulphuric acid, 10 to 12 grams of sodium sul- 

 phate, and about one gram of copper sulphate. Place the 

 flask on the Kjeldahl digesting apparatus, and heat gently 



until the first vigorous frothing ceases, then raise the heat 

 gradually until the liquid boils. Continue the boiling until 

 the solution becomes clear. Allow the flask to cool to 40 

 degrees to 60 degrees (If allowed to become thoroughly 

 cold the solution solidifies); dilute carefully with ISO cc. of 

 water; allow to cool. Add 100 cc. of 50 per cent sodium 

 hydrate solution, pouring it carefully down the side of the 

 flask, so that it does not mix immediately with the acid solu- 

 tion. Add about one gram of granulated zinc to prevent 

 bumping, and a piece of parafifine the size of a pea to dimin- 

 ish frothing. Connect the flask quickly with a condenser, 

 the delivery tube of which dips into a 500-cc. Erlenmeyer 

 flask, containing 50 cc. of tenth normal sulphuric acid, diluted 

 to about 100 cc. Shake up the Kjeldahl flask and heat gently 

 at first, increasing the flame as the danger of foaming over 

 diminishes, finally boiling briskly until about one-half of the 

 liquid has passed over into the receiver. Add one cc. of 

 methyl red solution, and titrate the excess acid by means of 

 standard sodium hydrate solution. 

 Calculation: 



100 rcc. H.SO, X normality — cc. NaOH 

 X normality). (0.014 X (5.56) 



= Percentage glue. 



Weight of sample 



When light-gravity stocks are analyzed as above, inexper- 

 ienced manipulators may not find glue when it is present. 

 Such stocks float on the surface of the acid and allow the 

 glue to be driven oflf by the heat before the nitrogen in it 

 has been converted into ammonium sulphate and therefore no 

 ammonia will be obtained. This difiiculty can be overcome 

 by constant agitation of the flask so that the rubber is kept 

 moistened with acid until it is attacked. This laboratory has 

 found it very convenient to add a small quantity of bromine 

 to the rubber, which makes a brominated mass that sticks to 

 the bottom when the acid is poured over it. There is there- 

 fore no danger of driving off any of the glue and the constant 

 agitation and attention otherwise necessary can safely be 

 dispensed with. 



CONCLUSIONS. 



1. The method presented will detect as little as 0.9 per cent 

 of glue in a rubber mixing. It has been found, by experi- 

 ment, that no substance other than glue will answer to this 

 test when it is carried out as directed, and the method is 

 reliable and convenient. 



2. When glue is present in a rubber sample its quantity 

 is arrived at by determining the nitrogen content by means 

 of the Kjeldahl procedure and calculating this to glue. We 

 have been able to find no other satisfactory quantitative 

 method which would give the correct percentage of glue. 



THE ADVANTAGES OF RECORDING THERMOMETERS. 

 .■\s an illustration of what can be accomplished by watching 

 results, the story is told of a large factory that had two vulcan- 

 izing departments in one of which an installation of six ther- 

 mometers on the vulcanizers gave greatly improved results. In the 

 other mill there were twelve vulcanizers, but because of some 

 difflculty with the vulcanizer man, the superintendent hesitated 

 to put in thermometers. Finally the man was induced to try 

 one out, and after a short trial he equipped all his vulcanizers, 

 and said that he wondered how he had gotten along without 

 them. In explanation he stated that before he had recording 

 thermometer? installed, if a heat came out badly he would 

 have to spend two or three days and sometimes nights person- 

 ally investigating before the cause was discovered. But, with 

 the recording thermometers, he first looked at the chart rec- 

 ords and if these agreed with the predetermined temperature 

 he knew that something was wrong with the mixture. Thus 

 he was enabled to put his finger on the trouble at once and 

 correct it without waste of time and without overmuch damaged 

 material. 



