NOVEMBER 15, 1936
ANALYTICAL EDITION
Summary Conditions have been established for the preparation in pure form of hexanitrato ammonium cerate, (NH4)&e(NO&, starting with a low-grade thorium-free mixture of 40 to 50 per cent ceric and cerous oxides containing 50 to 60 per cent of mixed oxides of praseodymium, neodymium, and lanthanum. The properties of this proposed new standard of reference in ceric oxidimetry are discussed and the indications pointing to its complex nature as distinguished from the double salt type of ceric salt, such as ceric ammonium sulfate, are pointed out.
45 1
The stability of solutions of the nitrato cerate in 0.5 N to 2.0 N sulfuric acid a t 100" C. has been studied and perfect stability shown. The many desirable properties possessed by the proposed new standard, which make its use in the new role desirable, are outlined.
Literature Cited (1)
Cuttioa and Toochi, Gazz. chim. ibal., 54, 628 (1924).
RECEIWDJuly 17, 1936
Construction of Glass Helices for Packing Fractionating Columns A Rapid Mechanical Method W. W. STEWART, Ontario Research Foundation, Toronto, Canada
S
INGbE-turn glass helices have been used by several investigators (I,$, S,6)as packing material in laboratory fractionating columns. Up to the present time their use, particularly in large columns, has been limited to some extent by rather slow and tedious methods of construction. A method for making these helices by hand was first described by Wilson, Parker, and Laughlin (7). A more detailed account of the construction, breaking, and sorting of glass helices has been reported by Roper, Wright, Ruhoff, and Smith (4), who wound glass spirals from Zmm. Pyrex rod by hand. The fiber diameter was about 0.6 mm., and the outside diameter of the coil 4.4 mm. About 5 cc. of finished product were produced in 1.5 hours, 1 hour of which was required for breaking and sorting. Recently, a partially mechanical method of constructing glass spirals has been described by Young and Jasaitis (8). The distilling column which has been used in this laboratory has a packed section of 300 cc. It would take about 80 hours to prepare a sufficient number of helices to pack this column by using the technic of Roper and his co-workers, whereas it was accomplished in 15 hours using the method here outlined.
Winding the Spirals A mechanical device for winding glass spirals similar to those described by Roper was constructed from Meccano parts. The design of this machine was based on that of a device developed by Tapp (6) for winding spirals from quartz fibers. The machine was built to wind directly a glass spiral, with a fiber diameter of about 0.6 mm. and 11turns per centbe meter, from a Pyrex rod 2 mm. in diameter. The fiber for the spiral was drawn from this rod as the spiral was wound on a winding form. The mechanical device was constructed in two parts: (1) a unit which rotated the winding form a t a uniform rate and, at the same time, moved this form in a direction a t right angles to the plane of rotation at a uniform speed; (2) a similar unit which fed the rotated glass rod onto the winding form a t a uniform rate. The details of construction of these units are clearly shown in Figure 1. The power plant was a Bodine electric motor, ty e CR2, equipped with a 595 to 1 reduction gear. The gear (ko. 27a, Meccano part number) on the motor shaft revolved at 16 r. p. m.
and meshed with the gear wheel (No. 31) on the horizontal drive shaft, causing the latter to revolve at 28 r.p.m. The pinion (No. 26) on the drive shaft meshed with the centrate wheel (No. 28) which drove the worm gear (No. 32). The worm gear meshed with the inion (No. 25) on the shaft driving the chain sprocket (No. 96af The chain, which was fastened at one end, passed under an idler sprocket and then over the driving sprocket. The other end of the chain was fastened to a spring which applied a tension, thus reventing a slipping motion as the drive sprocket meshed with tge chain links and carried the carriage forward. The forward motion was at the rate of 1 mm. per revolution of the winding form which rotated at 28 r. p. m. The glass spirals were wound on a 3.18-mm. (0.125-inch) diameter steel drill rod, which was coupled to the main drive shaft, and passed through a bearing fixed on the end perforated flange plate of the carriage. The free end of the winding form assed through a bearing which was fastened to anothw pergrated flange plate (right of Figure 2), the latter being fixed to the base board. These two bearings served to steady and guide the form during the winding operation. The end of the winding rod was slotted. The glass spirals were wound from a 2-mm. Pyrex glass rod, fed through a gas-air flame onto the winding form at right angles to it and in the same horizontal plane. The feeding mechanism is shown in Figure 1, the glass rod being clamped in the second carriage in exactly the same position as that occupied by the winding form in the first. The relative positions of the glassrod feeding device and the carriage bearing the winding form are shown in Figure 2. The inner cone of the gas-air flame was about 4 cm. high, and the outer tip was placed under the glass rod approximately 0.5 cm. from the winding form. The tip of the inner cone of the
FIGURE1. GLASSHELIX-WINDING APPARATUS
452
INDUSTRIAL AND ENGINEERING CHEMISTRY
flame was kept about 0.5 cm. below the Pyrex rod. Variations in the gas or air pressure will cause slight irregularities in the diameter of the spiral fiber. With a little practice, the operator can adjust the flame so that a uniform spiral is produced. Figure 2 shows a spiral being wound. The original Pyrex rod, the finished spiral, and single-turn helices are shown in Figure 3.
FIGURE2. GLASSSPIRAL BEINGW'OUND FORM
ON
WINDING
To begin winding a spiral, the two carriages were drawn back so that the end of the winding rod and the end of the glass rod almost entered the flame zone. Both motors were started simultaneously with a single switch. As the glass rod passed into the flame and softened, the tip of the rod was drawn out with a short bit of glass tubing, and this fiber was hooked into the slotted end of the winding rod. The finished spiral can be slipped off the free end of the steel winding form when the winding operation is completed.
Preparing the Helices The method used for cutting single-turn helices from the glass spirals is an improvement on that of any described method that has been noticed by the author. A long spiral was slipped over a 3.18-mm. (0.125-inch) diameter steel drill rod, then held firmly between the thumb and forefinger and each turn nicked by drawing a wedge-shaped pointed Carborundum glass-marking pencil along the top of the spiral. When the etched spiral was squeezed between the thumb and forefinger, each turn broke a t the nick. If any turns break less than three-fourths of a turn, they drop off the rod, thereby greatly reducing the period required for sorting. The number of helices breaking between three-fourths of a turn and one turn was very small, and these may be easily and quickly
VOL. 8, NO. 6
separated as the helices are slipped off the end of the steel drill rod if a uniform packing of exactly one turn is desired. Helices of any required number of turns may be readily made by nicking the spirals a t the proper points. The loss due to breakage is low. About 10 cc. were lost in preparing 300 cc. of finished product, as compared with a loss of about 20 per cent with the method of Roper, Wright, Ruhoff, and Smith (4). The spiral was wound on this machine a t the rate of 350 turns per 12.5 minutes. However, with a few simple mechanical changes it is possible to increase the rate of winding considerably, if desired. The finished spiral was about 32 cm. long and had 11 turns per cm. In order to save time, the single-turn helices can be cut from one of these lengths of spiral while another spiral is being wound. Enough singleturn helices to occupy a packed volume of 20 cc. (about 1260), can be made by this method of construction in 1 hour. Some measurements upon average pieces of the finished product are as follows: fiber diameter, 0.65 * 0.02 mm.; outside diameter of the coil, 4.47 * 0.04 mm.; weight of 300 single turns, 2.64 * 0.04 gram. The 300 cc. of finished product required to pack the laboratory distilling column were obtained from 32 Pyrex rods 2 mm. in diameter and 92 cm. long. Table I shows the Meccano parts required for constructing a single carriage for winding the spirals. The total cost of these parts is about 15.00. TABLE I. PARTS REQUIRED No. of Parts Required 1 1 2 4 3 4 1 1 4 1
Mecanno Part No. 6 0a
7 8 9 12 13a 15 14 25
1
26
1 1 1 2 2
275 2s 32 37f 48b
7
52
1
96
1
96a
6
1265 63 69 20 94
1 10 4 2 feet
Description Perforated strip 38.1 mm. (1.5 inches) Perforated strip' 50.8 mm. (2 inches) Angle girder 62; mm. (24.5 inches) Angle girder: 318 mm. (12.5 inches) Angle irder 63.5 mm. (2.5 inches) Angle %rack& 12.7 mm. (0.5 inch) Axle rod 203 h m . (8 inches) Axle rod' 127 mm. (5 inches) Axle rod' 165 mm. (0.5 inches) Pinion wheel diameter 19.1 mm. (0.75 inch), 6.4 mm. (0.25 'inch) face Pinion wheel! diameter 12.7 mm. (0.5 inch:I, 6.4 mm. (0.25 inch) face Gear wheel 57 teeth Centrate wheel, 25.4 mm. (1 inch) 38 teeth Worm Nuts and bolts, 5.6 mm. (0.22 inch) Angle strip, 88.9 mm. (3.5 inches) X 12.7 mm. (0.5 inch) Perforated flange plate, 140 mm. (5.5 inches) X 63.5 mm. (2.5 inches) Sprocket wheel, 25.4-mm. (1-inch) diameter, 18 teeth Sprocket wheel, 19.1-mm. (0.75-inch) diameter, 14 teeth Flat trunnions Coupling Collar with set screw Flanged wheel, 25.6-mm. (1.125 inches) diameter Chain
Literature Cited (1) Fenske, M. R . , Tongberg, C. O., and Quiggle, D., IND.Exo. CHEM.,26, 1169 (1934). (2) Kistiakowsky, G . B., Ruhoff, J. R . , Smith, H. A,, and Vaughan, W. E., J. Am. Chem. SOC.,57, 876 (1935). (3) Laughlin, K. C., Nash, C. W., and Whitmore, F. C., Ibid., 56, 1396 (1934). (4) Roper, E. E., Wright, G . F., Ruhoff, J. R . , and Smith, W . R., Ibid., 57, 954 (1935). (5) Tapp, J. S., Can. J. Research, 6 , 584 (1932). (6) Tongberg, C. O., Quiggle, D., and Fenske, M. R . , IND. ENG. CHEM.,26, 1213 (1934). (7) Wilson, C. D., Parker, G . T., and Laughlin, K . C . , J. Am. C h m . SOC.,55, 2795 (1933). (8) Young, W . G . , and Jasaitis, Z., Ibid.,58, 377 (1936).
RECEIVED July 8, 1936.
FIGURE3. ORIGINALPYREXROD, FINISHEDSPIRAL,AND SINGLE-TURS HELICES
Direct Determination of Oxygen in Organic Compounds Containing Sulfur ter Meulen Method W. WALKER RUSSELL
AND
MAURICE E. MARKS, Metcalf Laboratory, Brown University, Providence, R. I.
I
The Catalyst
N PREVIOUS studies (2, 6, 7 ) the authors found that, when certain modifications were made, the ter Meulen method (3, 4, 5 ) for the direct determination of oxygen in organic compounds by catalytic hydrogenation gave satisfactory results. Thus when a very active thoria-promoted nickel catalyst was used, all oxygen was quantitatively converted to water in the analysis of compounds containing only carbon, hydrogen, and oxygen (6), and this was equally true when nitrogen (7) or small amounts of sulfur (9) were also present. The last work (W),which dealt primarily with the direct determination of total oxygen in oils whose sulfur and nitrogen content was below 0.1 per cent, appeared to justify further work with organic sulfur compounds. In an endeavor to make the method more generally applicable, the behavior of compounds containing considerable amounts of sulfur and various types of sulfur linkage has been studied. Even though sulfur is recognized as a serious poison for nickel catalysts, a method has been developed which has given satisfactory successive analyses with the several types of organic sulfur compounds studied.
Unsupported and quartz-supported nickel catalysts, both unpromoted and thoria-promoted, have been investigated. Ordinarily any of these catalysts entirely sorbed any sulfur compounds not held back by the platinized quartz cracking surface. I n fact, hydrogen sulfide was not obtained from any active catalyst under conditions of analysis, or even at 450" C., unless uncracked benzene passed through the system. When this occurred hydrogen sulfide was evolved and the catalyst became completely poisoned. Considerable amounts of sulfur could be present in the system before the catalysts failed in analysis. Thus with aliphatic compounds, in whose analysis the catalyst was the limiting factor, from eight to thirteen runs could be made, thereby introducing some 200 to 400 mg. of sulfur, before the 5 to 10 grams of unsupported promoted catalyst failed to give quantitative conversion of oxygen to water. The smaller nickel content of the supported promoted catalysts allowed about four successive runs on 20 grams of this material. Certain samples of unsupported catalyst, both unpromoted and thoria-promoted, exhibited a marked and persistent increase in blanks immediately following an analysis-for example, an initial blank of 0.5 mg. per half hour might increase to 2 t o 3 mg. directly after a run. However, after passing hydrogen for a few hours, the blanks decreased and finally became constant below their initial values. Observations of this phenomenon led to the belief that hydrogen sulfide was in some way acting upon the catalyst. This contention was strengthened by the results of experiments in which hydrogen sulfide was passed through the tube under the conditions of analysis. I n the cases of both promoted and unpromoted catalysts, the blanks were considerably increased thereby. It seems necessary to assume, therefore, that hydrogen sulfide, and perhaps other sulfur compounds which may exist a t this stage in the analysis, are capable of accelerating the reduction of nickel oxide which escaped reduction by hydrogen alone during the preparation of the catalyst. Small amounts of such oxide are known to exist within catalyst granules even after prolonged reduction. It is interesting to note that the increased blanks persisted for some hours after the passage of hydrogen sulfide had been discontinued. This may signify a slow reaction with adsorbed hydrogen sulfide or sulfur, or a specific catalytic effect. Although the phenomenon of increased blanks was not observed when reduction was prolonged for 2 weeks, it was shown by certain samples of unsupported catalyst, among a number which were believed to be identical in composition and mode of preparation, but which were reduced for 24 hours only. Some factor not under control, therefore, appears to have been responsible for a variation in the amounts of residual oxide present in the more rapidly reduced catalysts. A simple and very effective solution of the problem was found to lie in the use of granular quartz-supported nickel catalysts. The nickel was now present only in very thin layers which were readily freed from all but negligible amounts of oxide by reduction in hydrogen overnight a t 500" C. It is recommended, therefore, that the supported, promoted nickel be used in all analyses of sulfur compounds as catalyst, but not as cracking surf ace.
The Cracking Surface Cracking surfaces composed of platinized quartz granules, and also of nickelieed quartz granules, both with and without thoria, were investigated. While the nickel-coated surfaces appeared slightly more active, they proved unsatisfactory because of large blanks. The platinized quartz cracking surface (6) proved very efficient for most organic sulfur compounds studied. More or less sulfur from the compound was always retained by this cracking surface. Because the aliphatic compounds cracked easily to gaseous products between 600" and 800" C., depositing little carbon, the efficiency of the cracking surface was maintained, and the capacity of the catalyst to resist sulfur poisoning became the limiting factor in determining the number of successive runs possible with one tube-filling. Aromatic compounds which may require cracking temperatures up to 1100" C. deposited considerable carbon which diminished the activity of the cracking surface as successive runs were made. Thus the amount (about 20 cc.) of platinized quartz employed allowed about five successive analyses to be made of either p-xylenesulfonic acid or di-p-tolyl sulfoxide, about three of sulfobenzoic anhydride, or about two of diphenyl sulfone. Failure occurred when the cracking surface was no longer able to prevent easily condensable decomposition products from passing through. A larger amount of cracking surface should make possible an increase in the number of successive analyses of aromatic sulfur compounds. Of the aromatic compounds studied, diphenyl sulfone required the highest temperature for cracking-i. e., about 1100" C.-while for the other aromatics temperatures down to 800" C. sufficed. That the number of successive analyses possible with a given aromatic compound offered a criterion of its cracking characteristics f o l l o ~ from s observations that the rate of vaporization of the sample had to be decreased in the above sequence in order to obtain satisfactory cracking. Thus the full extension of the analysis time which has been recommended for organic sulfur compounds was necessary only in the cases of the compounds most difficult t o crack. 453