March, 1925
I.VDUSTRIAL -4ND ENGINEERING CHEMISTRY
optimum time to add water is as soon as the rubber is smooth around the roll so that the water will be held on the roll and not fall into the mill pan. As the heat is actually generated at the bite of the rolls, nothing is gained by using a spray, and the water may more conveniently be poured on slowly, with care not to overcool the stock so that it breaks from the roll. Although many rubber men have a t times added water a t the elid of a batch to stop incipient scorching, the writers believe that the use of water as an aid in plasticizing is new, and whereas its use requires some discretion, large-scale operations have demonstrated that batches can be handled by this method under summer conditions that would otherwise be impossible. Type of Rubber
Another item to be considered is the rubber. The most uniform rubber on the market from the plasticity standpoint is pale crepe, with smoked sheets a fair second, and the brown crepes a poor third. The possibilities of grading raw rubbers by their plasticity figures are interesting. A tough lot of rubber will cause hotter and tougher batches and may cause scorched stock. Breaking down or massing the rubber on
297
separate mills puts the peak of the power consumption in a place where there is no scorching danger and leaves a somewhat more uniform rubber for further milling. Oil Softeners The use of the various types of oil softeners and greases, which has become common in recent years, introduces another means of varying the plasticity. The later these are added to the batch the softer the stock, the difference between adding the softener to the rubber a t the first of the batch instead of with the compound a t the end of the batch being often as much as fifty points on the plasticity. Differences of twenty points or more are also found between stocks using the same amounts of different softeners. Summary The plasticity of any given mill batch can be altered or controlled by changing one or more of the following factors: (1) temperature, (a) by mill water control, (h) by adding water on the batch, (c) by controlling gage of stock; (2) time of milling; (3) size of batch; (4) type of rubber; and (5) type of softeners, amount, and point of addition.
A Comparison of Various Tests upon Glue, Particularly Its Tensile Strength's' By Augustus H. Gill ~IASSACHUSETTS INSTITUTE OF T ~ C H N O L O CAMBRIDGE, CY, MASS.
X 1913 the results of a series of tests upon glue were published, covering a period of work of about fifteen years.3 The methods followed in the present paper are substantially those there given, but in addition the actual tensile strength of the glue itself has been determined, a t the suggestion of the late Friman Kahrs. The results were in manuscript several years before the article by Hopp4 appeared, and it is interesting to note that they are in substantial agreement with his results. I t seems likely that, inasmuch as little or no elongation was observed by the writer, the specimens were drier than those in the article cited. In addition to this actual tensile strength, the absolute viscosity in centipoises has been determined with the MacMichael viscometer. Finally, the grading by the Cooper Grades, the viscosity by a glass pipet, and the prices as of 1915 are included. Preparation of Test Strips
I
Inasmuch as the process of making the testing strip of glue is not altogether easy, it will be described here. One hundred and twenty grams of glue were allowed to soak overnight in 150 grams of water and then heated t o 60" C. until a completely homogeneous solution was obtained. It was poured into a slightly greased pan, 12 cm. (48/4 inches) X 21.5 cm. (8'/2 inches), making a layer 2 cm. ("4 inch) deep. Here it was allowed to remain until a skin formed over the surface, when it was turned to present the other side to the air until a skin formed upon this side, which required about 6 hours. The cake was then ready for drying. This was effected by tying or bolting it between pieces of galvanized iron wire gauze of 1.3-cm. ('/t-inch) mesh and hanging it up in front of a ventilator giving a draft of air a t a temperature of
* Received January * Contribution No
14, 1925. 86 from the Laboratory of Technical Analysis, Massachusetts Institute of Technology. THIS JOURNAL, 1, 102 (1916). 4 I b z d , 12, 1 (1920).
about 30" C. Three weeks are required to complete the drying. At the end of this time, the cake was sawed with a fine hack saw into three strips about 2.5 cm. (1 inch) wide and 18 cm. (7 inches) long. A miter box and a fine saw such as is used for sawing hardwood can be employed instead; the glue is very hard and dulls the saw easily. The strips were filed to remove the outside skin and give a piece 4.5 cm. (13/4 inch) X 2 cm. (3/4 inch) X 0.5 cm. (3//16 inch) in the center of the strip, and then put into the clamps of the testing machine, medium fine emery cloth with the emery side towards the specimen being used as a buffer. Without this precaution the specimen will usually break in the clamp. I n preparing and mounting this strip, several other precautions have to be observed. The portions in the jaws must be smoothly filed to present an even surface; the reduction of size mentioned in the middle, 4.5 cm. (13/4 inches), etc., must be in the form of a smooth curve, as sharp corners influence the break a t this place, like filing a scratch in a glass tube. The reduction in size of the specimen also serves the purpose of eliminating skin tension, removes air bubbles, and brings the break a t a convenient figure-about 70 kg. per sq. cm. (lo00pounds per square inch). Tensile Strength Determination The load must be applied a t a uniform rate and, except a t the start, preferably by hand, as the elongation is very slight. The fracture was cone-shaped a t each fractured end. In case air bubbles were found in the fracture, their size was calculated and deducted from the area of the cross section. The fractured pieces were collected, ground in a meat grinder, and the moisture determined by drying a t 105" C., giving from 12 to 15 per cent-about the normal amount. Viscosity Determination The absolute viscosity was determined by allowing 20 grams of glue to soak in 120 cc. of cold water for 12 hours,
2 98
INDUSTRIAL A N D ENGINEERING CHEMISTRY
Vol. 17, No. 3
Comparison of Various Glues
A
Hide and fleshings
236
191
AEx
17.5
38.5
739to800
10,50Oto11,350
217
3085
5.8
83.0
B
Bonecalf
170
130
1 Ex
16.5
36.3
817 to894
11,600 to 12,700
...
...
4.7
67.8
C D
Hide
176
146
1 Ex
13.5
29.7
711 to810
10,100to 11,500
258
3665
5.4
78.0
First-run hide 160
138
li/r
13.0
28.6
697 to746
9900to 10,600
261
3700
5.3
75.3
140 118
118 116
11/2 11/4to IT/:
11.5 9.0
25.3 19.8
711to725 317to460
10,10Oto10,300 4500to 6540
277 185
3840 2635
4.0 3.3
57.0 47.1
E F
Bone Junkbone
heating gradually in thezviscometer cup and stirring until 60” C. (140’ F.) is reached, when the viscosity is read using the MacMichael instrument. This was done by different observers two years apart, which accounts for the difference in the results. Results In the accompanying table the first four columns of figures are those of the manufacturer. They are his arbitrary ratings and valuable only for comparison with each other. The jelly test and viscosity (done with a glass pipet) agree remarkably well, being, as Bogue says, “parallel functions.” T h e addition of hide stock to bone glue seems to increase both jelly test and viscosity, and fleshings seem to raise them still further. Whether this is generally true remains to be seen. I n the last nine columns it is to be noticed that the actual tensile strengths of all the glues except junk bone are approximately the same. The diminution in strength is probably due to the hydrolysis of the gelatin of the glue while being made. This uniformity of strength excludes this test as a means of discrimination.
24.0 25.2 13.5 13.1 14.0 15.3 26.0 21.0
Where actual strength is of importance, there is little advantage in using a high-grade glue. This is also true in the sand-paper industry, where a medium quality of glue is employed. Since the tensile strength of glue is about 704 to 845 kg. per sq. cm. (10,000 to 12,000 pounds per square inch) and wood 1408 to 2112 kg. per sq. cm. (20,000 to 30,000 pounds per square inch) and that of a glued joint about 352 kg. per sq. cm. (5000 pounds per square inch), the care taken in making this joint-the perfection of the union of glue and woodis the important factor. The strength determined by the bibulous paper test is in accord with the jelly and viscosity tests. It is to be noticed that the absolute viscosity does not agree with the pipet viscosity in the case of Sample D, first-run hide. Acknowledgment Acknowledgments are due Messrs. Codwise, Hale, Davis, Hanson, and Chirgwin, by whom the experimental work has been performed.
A Laboratory Circulating Pump for Corrosive Vapors’ By H. C. Kremers UNIVERSITY OF ILLINOIS, URBANA, ILL.
I
N ONE phase of researches on the rare earths in this
laboratory, large quantities of anhydrous chlorides of the rare earths are continually being prepared. It has been found by most workers that these chlorides are best and most conveniently dehydrated in a stream of dry hydrogen chloride gas. Since the rate of dehydration is then a function of the rate of flow of hydrogen chloride vapor, it is quite evident that large quantities of this gas must be generated in order to dehydrate any quantities of material. Quantities as large as several kilograms of rare earth anhydrous chlorides thus require many days in preparation, not to mention the inconvenience of generating large volumes of hydrogen chloride. As a result of these conditions and in attempting to speed up this operation by greatly increasing the flow of hydrogen chloride vapor, the circulating pump shown in the accompanying diagram was designed. Hydrogen chloride is generated by the action of concentrated sulfuric acid on sodium chloride. The latter is first well soaked with concentrated hydrochloric acid solution. The gas is dried in the usual manner by passing through the tower, 2. This tower is, of course, filled with glass beads and kept well moistened with concentrated sulfuric acid. The furnace, 9, of the ordinary chromel-wound type, contains a glazed porcelain core 75 mm. x 450 mm. with 1 Received
January 1, 1925.
a flange a t 8. This flange and the cap to fit have been made into an ordinary flat-ground joint similar to a desiccator and cover. By making a gasket of asbestos paper moistened with dilute water glass, and clamping on the cap, a joint perfectly tight to hydrogen chloride is made, which is capable of withstanding considerable heat. The incoming hydrogen chloride is preheated by the coil 7. The most satisfactory cement for attaching the preheater tube 7 to cap 8 is composed of very fine alundum cement moistened with dilute water glass. Upon drying and a t a few hundred degrees heat i t forms a hard semivitrified mass. The action of the circulating pump is clear from the diagram. The outer jackets, 5, are made of 50-mm. glass tubing and are about 700 mm. long. The inner plungers are, of course, sufficiently smaller in diameter to allow free action. Connecting rods are attached to the plungers by means of short pieces of brass tubing set over the projections on the plungers and then filled in with sealing wax. The remaining driving mechanism is self-explanatory and can be constructed to suit individual tastes and conditions. The liquid seal used in the jackets, 5, is a saturated paraffin oil, although concentrated sulfuric acid, or even mercury, might be used to advantage, depending upon conditions or gas used. The intake, 3, and outlet, 4, valves are made of glass and the valve seating is flat enough to prevent sticking of the valve during operation.
