I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
March, 1925
317
Rapid Method for Determination of Carbon Monoxide in Hydrogen' By A. T. Larson a n d C. W. W h i t t a k e r FIXED NITROGEN RESEARCH LABORATORY, WASHINGTON, D. C
T
HE methods commonly employed for the estimation of carbon monoxide in hydrogen involve either complete combustion of the gas mixture or the preferential catalytic combustion of the carbon monoxide,* followed in either case by the absorption and estimation of carbon dioxide. The method described in this paper is based on the following reaction: CO
+ 3Hz
CHI
+ Hz0
in which the water formed is a measure of the carbon monoxide present in the original gas mixture. When the carbon monoxide concentration falls within a certain range the water formed may be rapidly estimated by means of a modified dewpoint tester of the Regnault type. Experimental Procedure
A diagram of the apparatus is given in Figure 1. The gas mixture to be analyzed is first thoroughly dried and then passed through the catalyst contained in A , the temperature of which is maintained a t 290' to 310" C. A vapor bath of boiling beta-naphthol will give the right temperature. The effluent gases now pass into the dewpoint tester, R. This consists of an outer shell of glass, fitted with a rubber stopper, in which is set a polished nickel container. (A nickel box of the type furnished with certain types of stick shaving soap may be used for this purpose.) The container is pro-
barometric pressure and the pressure difference within and without the apparatus measured on the manometer a t C, the percentage composition may be calculated. As a temperature-regulating liquid, water is very satisfactory for temperatures above 0" C. Salt solutions cooled with ice may be used for temperatures as low as -40" C. Most types of dew-testing apparatus control the temperature by allowing air to bubble through ether which partially fills the inner cylinder, This method is quite satisfactory, particularly for temperatures above 0" C., but requires the addition of a dropping funnel for the replacement of ether lost by evaporation. A suitable condenser might serve equally well. In practice it has been found convenient to employ both devices. The accuracy of this method for estimating carbon monoxide depends primarily on the completeness with which carbon monoxide may be converted into methane. This has been accomplished by employing a catalyst prepared as follows: A 6 per cent solution of ammonium hydroxide (temperature about 40' C.) is slowly added to a 6 per cent solution of nickel nitrate (temperature about 40' C.) until precipitation is complete. After a thorough washing by decantation (use only distilled water), the precipitate is transferred to a Biichner funnel. When the water has been drained, the paste is placed in a porcelain mortar and ground to a thin paste. After a second drying on a Biichner funnel, the filter cake is broken into small lumps, for example 1-inch cubes, and dried in a ventilating oven. During the first 12 hours the temperature of the oven is held a t 80' to 90' C . The temperature is then raised from 125' to 150' C. and the drying continued for an additional 24 hours. The hard, compact granules which now remain are broken up into particles of 8 to 14 mesh, which is a convenient size for most purposes. About 15 cc. of the oxide thus formed are placed in catalyst tube and reduced in a stream of hydrogen a t a temperature of 290' to 310' C. Discussion of Method
Table I shows the results obtained with this apparatus in the examination of a nitrogen-hydrogen (ratio 1:3) gas mixture containing 2.26 per cent carbon monoxide. Expt. 1 2 3 4 5
vided with a 3-hole rubber stopper which suppwts a thermometer, calibrated accurately to 0.1' C., and an inlet and outlet tube for the passage of the temperature-regulating liquid. As the outside surface of the nickel cylinder is gradually cooled, a temperature is reached at which the water vapor content of the gases in contact with the cylinder reaches the saturation point, and a visible film of moisture forms on the bright nickel. By referring to a table giving the relation of vapor pressure of water to temperature, the partial pressure of water vapor in the gas may be ascertained. From the 1 3
Received October 17, 1924. Rideal and Taylor A n a l y s l . 44, 89 (1919).
Table I Deypoint Pressure C. Mm. 19.8 734.7 19.8 753.1 20.3 714.7 20.0 762.4 19.9 754.4
CO
Per cent 2.21 2.27 2.29 2.28 2.28
This method is probably as accurat'e as any of the analytical procedures commonly employed and has the added advantage of being rapid. Its usefulness is limited, however, to conditions under which the partial pressure of carbon monoxide lies below the vapor pressure of water at the temperature of the apparatus. Otherwise, condensation will take place in the connecting tubes. At a room temperature of 21" C. the maximum carbon monoxide content should not exceed a partial pressure of 19.83 mm. of mercury or 2.46 per cent of gaseous mixture. The lower limit of practicability is determined by the low temperatures available and by the skill of the operator. With a little practice 0.1 per cent of water which is detectable at about -20" C. may be easily estimated.
