THE SYSTEM AMMONIA-WATER AT TEMPERATURES U P TO 150°C. AND AT PRESSURES U P T O TWENTY ATMOSPHERES I. L. CLIFFORD
AND
E. HUNTER
Research Laboratory of I . C. I . ( A l k a l i ) Ltd., Winnington, Northwich, Cheshire, England Received October 18, 1938 INTRODUCTION
The total vapor pressures of ammonia-water solutions a t temperatures between 0°C. and 120°C. and a t pressures up to 9 atmospheres have been measured by Mollier (1) who, however, determined no vapor compositions. Smits and Postma (2) examined the system in the region of the eutectic points and three-phase lines. They determined the pressures, temperatures, and compositions of certain solut,ions, including those existing in equilibrium with the various solid phases. Perman (3) has determined both total vapor pressures and vapor compositions between OoC. and 60°C. a t pressures below atmospheric. Perman's method consists of passing a known volume of air through the ammonia solution, and determining the weights of ammonia and water the air takes up. The partial vapor pressures of ammonia and of water can then be calculated from these weights, and the total pressure of the airammonia-water vapor phase. The work now reported was carried out in order to extend the knowledge of this system up to 150°C. and 20 atmospheres pressure. A dynamic method was used to determine vapor compositions between 60°C. and 100°C. and for pressures up to 1.5 atmospheres. The method was inconvenient for higher pressures. The total vapor pressures in the system up to 9 atmospheres were already known (4). For the higher range of temperatures and pressures it was found necessary to use a static method. Both total vapor pressures and vapor compositions were measured. EXPERIMENTAL
1. From 60°C. to 100°C.; pressures u p to 1.6 atmospheres
At temperatures as high as 60"C., Perman had difficulty in getting the air he passed through his saturators containing the ammonia solutions to 101
102
I. L. CLIFFORD AND E. HUNTER
take up its proper amount of ammonia and water. As saturators he used vessels of the wash-bottle type, and his difficulties were most probably due to the inefficiency of these when called upon to saturate air with the large quantities of ammonia and water provided by the high partial pressures at 60°C. A very efficient saturator designed by Bichowsky and Storch (5) has been adapted for vapor pressure work by Pierce and Snow (6). Its design is shown in figure la. The solution is held in the tube A. Air enters a t B and breaks into bubbles a t the jet C. These bubbles travel up the inclined tube D, becoming saturated with vapor from the solution. At the surface of the solution they break and the vapor travels back to the outlet a t E. Behind and below the jet C, a vent F is blown in the tube D. Here solution is drawn in and carried up the tube D between the bubbles. This ensures thorough mixing of the solution. The neck G, provided for filling purposes,
FIG.l a is closed by a rubber stopper which can be wired on. A sample of the solution can be withdrawn by way of the capillary tube H. Four vessels of this pattern, each 10 in. x 14 in., made of Pyrex glass, were joined together in series in the way shown in figure l b . It was not considered necessary to provide the two inner vessels with sampling tubes. The inlet and outlet tubes A and B were vertical. The whole apparatus was mounted in a frame which could be held in a thermostat capable of regulation to f0.01"C. The water level in the thermostat was maintained well above the stopper C. To prevent condensation from the gas on the walls of the tube B a heating coil D formed on a brass tube surrounding B was used. The tube was heated to 12O-14O0C. The air supply was kept a t a suitable pressure, indicated by the manometer E, by adjusting the screw clip F on a tube opening to the atmosphere. The clip G regulated the supply to the apparatus. Rubber pressure tubing was used, wired on a t all connections. The method of using the apparatus depended upon whether the solution
THE SYSTEM AMMONIA-WATER
103
in the saturators had a total vapor pressure below or above approximately 650 mm. a t the working temperature. With the former type of solution it is possible to pass through a sufficient amount of air without taking the pressure above atmospheric; this method will be described first. The saturators were filled to the correct level (the top of the inclined tubes) with an ammonia solution of known composition. After wiring on the stoppers the apparatus was placed in the thermostat and the heating coil was fitted over the outlet tube B. The pressure tubing from the air supply already described was wired on to the tube A with the clip G closed. A length of about 6 inches of Pyrex tubing (3-mm. bore) was sealed to the tpp of the tube 13, and bent over at right angles, as shown a t J in figure 2. The current to the heating coil and thermostat was then
switched on, and the latter was brought to the working temperature and there regulated. The absorption train consisted of two U-tubes, the first containing glass beads covered with 25 per cent sulfuric acid, the second, glass beads with concentrated sulfuric acid. A third tube filled with glass wool-phosphorus pentoxide absorbent was eventually discarded, for its weight never increased by more than one-fifth of a milligram during a run. The absorption train, after preparation and weighing, was fitted to the tube J with pressure tubing. The clip G was carefully adjusted to allow a slow, steady stream of bubbles to form and to pass through the apparatus and absorption system. During the run the tube J was gently warmed with a Bunsen flame to prevent condensation before the vapor reached the U-tubes. After a run, the clip G was closed and the absorption train was disconnected and weighed. The contents of the U-tubes were washed out, and the ammonia
104
I. L. CLIFFORD AND E, HUNTER
present was determined by distillation. The increase in weight of the U-tubes was normally between 0.5 and 1.0 gram. Finally a sample of the solution in the saturators was removed for analysis. A change in composition of one part per hundred of the NH, percentage was sometimes observed in the first saturator, where the air was admitted; in the last saturator the composition of the solution was always within '0.2 part per hundred of the original NH3 percentage. The composition of the solution in the last saturator was taken as that of the solution in equilibrium with the vapor collected. For solutions of the second type, those with total pressures between 650 mm. and 1100 mm. at the working temperature, it was necessary to devise some way of keeping the pressure in the saturators and absorption
FIQ.2
train at least 100 mm. above the total vapor pressure of the solution, in order to provide sufficient air to take up the ammonia and water. The apparatus shown in figure 2 was used. The U-tube A was made of Pyrex tubing (3-mm. bore) widening to 11-mm. bore tubing a t the U. To the base of the U, a third tube, B, was sealed and bent upwards. This was connected to the 50-cc. pipette bulb C, fitted to pressure tubing provided with the screw clip D. The apparatu's was held in an oven, E, fitted with mica windows. The oven was heated to keep the vapor passing through the U above its dew point. The side arm F was connected to the absorption system, the far end of which was connected to the tube H, dipping into a column of mercury, and to the t a p K. The U-tube A was filled with mercury to the level L.
THE SYSTEM AMMONIA-WATER
105
In preparation for a run, the apparatus was assembled in the way shown in figure 3. The taps of the absorption tubes and all rubber connections were wired on, and the warming up of the oven E was begun. The saturators were prepared as before, but when the temperature of the thermostat had risen to a point where the total vapor pressure of the complex in the saturators had reached approximately 650 mm., the tube J (from the saturators) was connected to the limb G of the U-tube with pressure tubing which was wired on. Subsequently, as the temperature of the thermostat rose towards the working temperature, the pressure of air, ammonia vapor, and water vapor in the saturators rose steadily. This
FIU.3
increase of pressure was balanced by applying a pressure of air through the tap K a t intervals, and watching the level of mercury in the limbs of A. After the thermostat had reached the working temperature and had been regulated, the mercury in A was drawn up into the bulb C, until there was a free passage for the vapor from one limb of A to the other, and the clip D was closed. Air was then admitted to the saturators and the run proceeded. The air finally escaped through H, bubbling up the tube M. Again, it was necessary to keep the upper tubes J, G, and F gently warmed. The run was brought to a n end by shutting off the air supply to the saturators and releasing the mercury in the bulb C, which closed the passage between G and F. The length of the limb F was usually sufficient to take
106
I. L. CLIFFORD AND E. HUNTER
the column of mercury which rose (owing to the pressure difference) when the absorption train was disconnected for weighing. At the highest pressures it was occasionally necessary to allow the thermostat to cool for some time before detaching the absorption tubes.
Results The results which were obtained are given in table 1: The total vapor pressures were obtained by interpolation of the values given in the International Critical Tables, Vol. VIII, p. 362. The figures for the ammonia content of the vapors are probably accurate to 1.5 units for vapors of composition 30-70 per cent of ammonia. Beyond these limits these determinations are believed to be more accurate. 2. From 100°C. to 160°C.; pressures u p to $0 atmospheres
The apparatus which was used had been designed for the determination of the equilibrium relations of pressure, temperature, and compositions of solutions and vapors in uncondensed systems, a t temperatures above 100°C. and pressures below 20 atmospheres. A 4-liter bomb is used to hold the complex, the vapor volume being about two liters. There are arrangements for stirring, for measuring the temperature and pressure, and for withdrawing for analysis samples of solution and vapor. The apparatus is shown diagrammatically in figure 3. The bomb A, of stainless iron, is closed by bolting down the cover, C, over a copper washer. I n the centre of this cover is a pressure gland through which the stirrer D passes. The gland is packed with 8. E. A. rings, and the stirrer rotates a t 80 r.p.m. and has paddles in the liquor and vapor. Screwed on to the cover are two valves, E and F. Valve E has attached a tube which dips into the bomb to within an inch and a half of the bottom. The opening of F is flush with the lower side of the cover, and this valve connects the apparatus to the closed hydrogen manometer G, which is capable of measuring pressures up to 20 atmospheres. A connection from the cover of A to the valve H leads to a smaller bomb J (300 cc.) which is held on to the cover by a bracket. This smaller bomb is fitted with two valves K and L, K for withdrawing the vapor sample to an adsorption train, and L for the inlet of dry air to the farther end of J for sweeping out the contents. A copper-constantan thermocouple in a sheath is also fitted through the cover of A, a lens-ring joint being used. T o avoid the use of compensating leads, the thermocouple wires run directly to a cold junction and potentiometer. The apparatus is held in position by fixing the bomb A in a tripod in the base of an oil thermostat, the oil level being above the top of the smaller bomb J. The thermocouple was calibrated a t 10°C. intervals up to 200"C., and can be read accurately to within 0.1"C. The closed manometer is filled with pure dry hydrogen, and the column of gas is water-jacketed. The
107
THE SYSTEM AMMONIA-WATER
volume of the capillary had been determined for each 2 em. of its 70 cm. length. In order to calculate the pressure of the hydrogen from the height
PRESSURE
1
TABLE 1 AMMONIA I N LIQUID
T
=
I
AMMONIA I N VAPOR
60.0%.
per cent
per cent
0.240 0.286 0.355
0.96 1.97 3.40
17.2 28.8
0.366 0.416
3.66 4.69
0.430 0.487 0.584 0.630 0.842
4.94 6.04 7.90 8.66 11.99
1.104 1.451
15.38 18.91
atmosphere8
T
39.1 43.1 55.3 55.6 60.2 69.6 72.3
,
79 2 84.2 90.0
= 80.0"C.
0.546 0.634 0.761 0.786 1.030 1.224 1.705
0.96 1.97 3.40 3.66 6.04 7.90 11.9
0.742 0.797 0.855 0.916 0.957 1.124 1.364
0.50 0.96 1.47 1.97 2.26 3.7 5.54
6.90 13.12 19.2 23.4 29.4 38.7 49.8
1.064 1.205
0.50 1.47
6.14 18.0
'
14.3 28.2 38.1 41.1 53.1 63.1 73.6) 73.41
of mercury in the capillary, it was necessary to know the quantity of hydrogen enclosed in the capillary. This was found by calibration
108
I. L. CLIFFORD AND E. HUNTER
against an open mercury column manometer a t 2 atmospheres pressure. I n addition, the higher pressure readings of the manometer (5 to 20 atmospheres) were checked against a calibrated Bourdon gauge. A cathetometer was used to read the position of the mercury in the capillary and t o measure the difference in level of the mercury in the capillary and in the reservoir. The pressure due to this head of mercury must be added to the calculated hydrogen pressure to give the pressure applied to the manometer. By closing a valve which lies between the reservoir and the body of the manometer, the gas space above the mercury in the reservoir could be evacuated without removing hydrogen from the capillary. To put a complex in the bomb the apparatus was evacuated while standing a t room temperature and the valve H closed. Two liters of complex was then drawn into the bomb A through the liquor sampling tube, care being taken not to introduce air with the complex. The thermostat was heated to the working temperature and regulated, and the contents of the bomb were stirred a t this steady temperature overnight. This was found to be a sufficient length of time to establish equilibrium. The manometer and thermocouple readings were taken. After evacuating the vapor sampling bomb J, the valve H was quickly opened and closed, isolating in J a sample of the vapor. The sample was drawn out by way of the valve K, through a sulfuric acid absorption train for analysis. When the pressure had been reduced to that of the service vacuum, air drawn through concentrated sulfuric acid was admitted through the valve L, and was swept through the bomb and train to remove the last traces of water and ammonia from J. Samples of solution were taken by opening the valve E and absorbing a few grams of the complex in a known weight of sulfuric acid in a system of traps which excluded the possibility of any loss of sample or acid. Usually a given complex was examined at, a whole series of temperatures between 100°C. and 150°C. New complexes were obtained either by emptying and re-filling the bomb, or by boiling off vapor from the previous complex. After some of the results had been obtained, it was found that the temperature of the complex in the bomb measured by the thermocouple, was always 3°C. below that of the outside bath, owing to heat losses through the connections from the bomb. This temperature difference was confirmed by experiments on the vapor pressure of water measured a t a number of temperatures in this apparatus. In order to make use of these first results the work was continued a t 97"C., 107"C., 117"C., 127"C., 137°C. and 147°C.
Results The experimental results are given in table 2. The error in the compositions in most of these, is about one part per hundred of the ammonia
TABLE 2 Isotherms of the system, ammonia-water PRESSURE
1
AMMONIA I N L I Q U I D
T
=
0.899 2.4 3.8
0 9.3 16.0
6.76
25.5
7.04
25.9
T
=
0 10.0
5.12 8.55
16.0 25.0
1.780 3.35 4.35 5.8 10.63
0 6.5 9.5 14.7
=
2.435 4.37
9.5
-
14.90
-
4.332 7.3
0
60.0 75.3 86.4
0
61.5 73.2 83.1
0
43.7 59.0 68.5 79.1
25.0
T 3.274 5.7 9.06
0 63.3 78.31 76.01 89.761 89,581 89.6
127°C. 0 6.43
5.62 7.31 12.44 13.6
per cent
107°C.
1.277 3.27
T
A M M O N I A I N VAPOR
97°C.
per cent
atmospheres
1
=
137"C, 0 6.5
25.0
0 42.0
65.9 78.5 80.2
0 6.5
0
9.3 11.84
9.5 13.6
50.2
16.3
19.5
75.0
110
I. L. CLIFFORD A N D E. HUNTER
percentage. In some cases, particularly a t the highest temperatures, the error in the vapor compositions is two parts per hundred. The error in the total vapor pressure determinations is less than two per cent.
FIQ.4 THE ISOTHERMS A N D ISOBARS IN THE SYSTEM
The experimental results are plotted as isotherms in figure 4. Data from tables of total vapor pressures given by Mollier (1) obtained by
11
112
I. L. CLIFFORD AND E. HUNTER
TEMPERATURE
degrees C .
180.5 170 160 150 140 130 120 110 100 90 80 70
60 50 40 30 25.3
I
TABLE 3 AMMONIA I N LIQUID
per cent
1
A M M O N I A I N VAPOR
per cent
0
0
3.4 6.6 10.1 14.0 18.0 22.0 26.3 30.8 35.6 40.6 46.0 52.2 60.0 70.2 87.0 100 '
20.5 36.3 51.5 64.8 74.2 81.1 87.0 91.7 95.1 97.4 98.8
100
Pressure = 8 atmospheres
30 20 18.5
0 0.3 3.1 6.4 10.2 14.1 18.3 22.4 26.7 31.4 36.4 41.9 47.5 54.1 62.0 73.5 94.6 100
159.3 150 140 130 120
0 2.6 5.9 9.6 13.5
171 170 160 150 140 130 120 110 100 90 80 70 60 50
40
0 2.3 20.7 38.9 55.5 68.0 77.3 84.0 89.5 93.3 96.2 98.2
100
0 19.5 40.0 57.1 69.9
113
THE SYSTEM AMMONIA-WATER
~
~~~~~
TEMPERATURE
I
TABLE 3-Continued AMMONIA IN LIQUID
I
AMMONIA IN VAPOR
Pressure = 6 atmospheres-Concluded degrees C .
110 100 90 80 70 60 50 40 30 20 9.7
per cent
per cent
17.6 22.0 26.6 31.4 36.6 41.9 47.5 54.0 62.0 74.0 100
78.4 84.8 90.0 94.2 97.0
100
Pressure = 4 atmospheres 144.1 140 130 120 110 100 90 80 70 60 50
40 30 20 10 0 -1.5
0
1.o 4.2 7.7 11.5 15.5 19.8 24.4 29.2 34.2 39.7 45.5 52.2 60.2 71.5 94.7 100
0 9.8 33.0 52.2 66.3 76.0 83.6 90.0 94.8 97.3
100
Pressure = 2 atmospheres 120.6 120 110 100 90 80 70 60 50
40 30
0 0.15 2.9 6.2 10.0 14.0 18.5 23.4 28.5 33.7 39.3
0 1.5 29.0 50.6 67.5 80.3 89.2 94.8 97.5 98.4
114
I. L. CLIFFORD AND E. HUNTER
TABLE 3-Continued TEMPERATURE
1
AMMONIA I N LIQUID
I
AMMONIA IN V A P O R
Pressure = 2 atmospheres-Concluded degrees C.
per cent
per cent
20
45.6 52.6 61.4 75.0 100
100
10 0 - 10
-18.5
100 90 80 70 60 50
40 30 20 10 0 - 10
-20
0 2.9 6.1 10.0 14.4 19.0 24.0 29.1 34.6 40.6 47.3 55.3 65.4 86.0
-30 -33.2
100
81.7 80.0 60 40 20 0 - 10 -20 -30 -40 -46.3
0 0.3 6.2 15.0 26.0 37.2 43.1 51.2 60.0 75.3 100
0 31.8 52.6 70.0 82.5 90.2 94.5 97.0 98.5 99.2
100
0 4.5 63.6 87.5 97.5
100
Pressure = 0 . 2 atmosphere 60.4 50 40 30 20
0 2.7 5.5 10.0 15.0
0 38.0 65.5 80.0 90.5
115
THE SYSTEM AMMONIA-WATER
TABLE 3-Concluded TEMPERATURE
I
AMMONIA I N LIQUID
I
AMMONIA I N VAPOR
degrees C .
per cent
per cent
10
20.0 25.3 31.8 100
95.5 98.0 99.0 100
0
- 10 -61.0
17.7 10 0 - 10 -20 -30 -40 -50 - 60 -70 -79.8 -82.9
0 1.7 5.0 9.5 15.0 21.0 27.5 34.3 41.5 51 .O 62.8 75.5
extrapolation of results determined up to 120°C. and 9 atmospheres, are also plotted on this graph and the agreement is good. Points from these isotherms, and from those of Perman, together with total vapor pressure values given in International Critical Tables (4) have been used to construct the isobars from 0.02 to 10 atmospheres in figure 5. The vapor branch of the 0.02 atmosphere isobar is unknown. A projection of the liquid and vapor and solid lines, determined by Smits and Postma, on the t - z base is shown on this figure and defines the lower temperature limits of existence of the solution phase. The course of the vapor branches of the isobars from 2 atmospheres to 10 atmospheres is somewhat uncertain in the temperature range 70" to 9O"C., where no vapor compositions have been determined and where there is a rapid change of curvature, but from analogy with the isobars of lower pressures, it is believed that the curves given in figure 5 are substantially accurate to within 1 per cent of the ammonia concentration. There are no points above 147°C. except those for pure water, but there is little doubt of the positions of the isobars since the curvature is so small in this region. Tables 3 and 4 have been constructed from figure 5. The values above 10 atmospheres in table 4 were obtained by int,erpolation of the straight
116
I. L. CLIFFORD AND E. HUNTER
TABLE 4 AMMONIA I N LIQUOR
I
PREBBURE
T per cent
=
0.197 0.439 0.717 1.084 1.559
T
=
A M M O N I A IN VAPOR
60°C.
atmospheres
0 5 10 15 20
1
per cent
0 56.4 75.8 84.1 91.3
80°C.
0 5 10 15 20 25 30
0.467 0.90 1.48 2.19 3.05 4.14 5.55
0 47.8 69.9 81.2 87.0 90.8 ' 93.3'
0 5 10 15 20 25 30
0.692 1.30 2.00 2.91 4.12 5.55 7.12
0 48.5 67.5 77.1 $4.5 89.1 92.4
0 5 10 15 20 25 30
1 1.82 2.75 3.88 5.31 7.18 9.48
0 45.6 64.5 75.3 82.8 87.9 91.3
T 0 5 10 15 20 25 30
=
110°C.
1.414 2.41 3.58 5.03 6.91 9.20 11.98
0 41.8 61.5 73.7 81.4 86.2 89.0
'
117
T H E SYSTEM AMMONIA-WATER
TABLE 4-Concluded A M M O N I A IN LIQUOR
I
PRESSURE
T
1
A M M O N I A IN VAPOR
= 120°C.
per cent
atmospheres
per cent
0 5 10 15 20 25 30
1.96 3.22 4.70 6.55 8.83 11.72 15.46
0 39.3 60.0 72.7 79.2 83.2
0 5 10 15 20 25 30
2.67 4.27 6.15 8.42 11.33 15.04 19.8
0 37.7 58.2 69.6 76.8 82.0
0 5 10 15 20 25 30
3.57 5.55 7.83 10.70 14.37 19.20
0 34.7 53.8 67.2 75.7
-
T 0 5 10 15 20 25 30
-
= 150°C.
4.70 7.14 10.02 13.58 18.6
-
0 31.3 31.5 66.4 75.5
lines obtained when log P is plotted against the reciprocal of the absolute temperature. SUMMARY
1. The compositions of the vapors in equilibrium with ammonia-water solutions have been determined a t 60"C.,8O"C., 90°C., and 100°C. for solutions with total vapor pressures up to 1.5 atmospheres.
118
I. L. CLIFFORD AND E. HUNTER
2. The total vapor pressures and compositions of vapors in equilibrium with ammonia-water solutions have been determined within the limits 100-150°C. and 1 to 20 atmospheres. 3. The isotherms in the system have been constructed for temperature intervals between 60°C. and 150°C. 4. From the results of the work reported and those of Perman, Mollier, and Smits and Postma (3, 1, 2), the isobars for 10, 8, 6, 4, 2, 1, 0.5, 0.2 and 0.02 atmospheres have been obtained. The authors wish to express their thanks to the Directors of Imperial Chemical Industries Ltd., for their permission to publish this work, which was carried out in the Research Laboratory of their subsidiary company, I. C. I. (Alkali) Ltd. REFERENCES (1) MOLLIER:Z. Ver. deut. Ing. 62,2,1315 (1908). (2) SMITSAND POSTMA: Proc. Acad. Sci. Amsterdam 17 ( l ) , 182 (1914). (3) PERMAN: J. Chem. SOC.83, 1168 (1903). (4) International Critical Tables, Vol. 111, pp. 362, 234. (5) BICHOWSKY AND STORCH: J. Am. Chem. SOC.37,2696 (1915). (6) PIERCE AND SNOW: J. Phys. Chem. 31,231 (1927).