Solubility of Carbon Dioxide in Benzene at Elevated Pressure

simple laws of ideal solutions and ideal gases which are no longer applicable even under moderate pressures. At present the data in this field are con...
8 downloads 8 Views 416KB Size
Solubility of Carbon Dioxide in Benzene at Elevated Pressure SHEN-WU WAN AND BARNETT F. DODGE Yale University, New Haven, Conn.

HE increasing use of elevated pressures in chemical industry has made i t desirable to have more data on the effect of pressure on the liquid-vapor-phase equilibria in various systems. Such data are fundamental to all operations in which these two phases are brought into contact-for example, in distillation, condensation, absorption, and drying. T h e data are needed not only for direct application in the development of industrial processes, but also to aid in t h e formulation and testing of generalizations to replace the simple laws of ideal solutions and ideal gases which are no longer applicable even under moderate pressures. A t present the data in this field are confined to a relatively few systems under rather limited conditions of pressure, temperature, and concentration. The general plan for our investigation in this field covered various combinations of gases and liquids over the temperature range 0" to 300" C. and at pressures u p to 1000 atmospheres. The present paper reports some results on the composition of the coexisting liquid and vapor phases of the system carbon dioxide-benzene at 30", 40", 50", and 60" C. and a maximum pressure of 95 atmospheres with a few scattered tests at lo", 15", and 20" C. The main reason for the choice of this particular system was that carbon dioxide is a gas of great industrial importance, and it is of interest t o compare its solubility in various organic liquids with t h a t in water. Benzene was chosen as a typical, nonpolar organic liquid. The only previously published investigation of this system was t h a t of Sander (3) in 1912, who determined the liquidphase composition only, at 20", 35", 60°, and 100" C. u p to a

T

maximum pressure of 116 atmospheres. J. L. Marsh started an investigation on this system a t this laboratory in 1930. He developed a n apparatus based on a static method and made some measurements a t 25" and 50" C. which have not been published. Difficulties v-ith the operation of his a p paratus led to its abandonment and the design, by the present authors, of a new one based on a dynamic method. Although a static method appears t o be simpler and fundamentally sounder, the dynamic method seemed to offer practical advantages which made it more suitable for our purpose.

Apparatus and Procedure GESERAL OUTLIXEOF METHOD.I n choosing a method and developing an apparatus, the main desideratum was a relatively simple apparatus t h a t would not involve special complication in construction and operation, and yet would yield results of a fair order of accuracy-about 1per cent or better. Carbon dioxide from a storage cylinder is passed continuously a t substantially constant pressure through a series of vessels containing benzene maintained a t constant temperature. T h e first vessel, or presaturator, is maintained a t a temperature above or below the point a t which equilibrium is to be established in order to permit a n approach to equilibrium from both sides. The equilibrium vessel is preceded by a second presaturator maintained at the equilibrium temperature. After a sufficient time, the pressure and temperature are noted, and both phases in the equilibrium vessel are sampled for analysis. APPARATUS. A flow sheet of the apparatus is shown in Figure 1. Carbon dioxide enters a t A , flows through presaturators B and C into equilibrium vessel D,and is finally expanded into sampling line I. Liquid samples are withdrawn through line E. Bourdon gage G is for pressure control only, all final pressure measurements being made with the dead-weight piston gage, H. Contents of the equilibrium vessel are agitated by a magnetic stirrer operated by a solenoid, the connections to which are indicated a t F . Details of the saturators and equilibrium vessel are shown in Figure 2. Being of substantially orthodox design, they require no special comment. The first presaturator is placed in one oil bath controlled to about 0.1" C., and the second presaturator and the equilibrium vessel are in u second oil bath, controlled to 0.01 C. Heating is obtained by an electric immersion heater. For operationbelow room temperature, the first bath is cooled by a standard household refrigerating unit. The liquid-phase sampler is a small steel cylinder, approximately 3 cc. in volume, with valves a t each end. The gas phase is expanded directly into the analysis train, consisting of two U-tubes and two glass absorption vessels. The U-tubes, packed with copper wool and immersed in solid carbon dioxide, serve to condense the benzene from the vapor. The absorbers contain sodium hydroxide solution for the removal of the carbon dioxide. In a eneral way, the method and apparatus are somewfat similar to that employed by Saddington

U

A

FIGURE 1. FLOW SHEET 95

INDUSTRIAL AND ENGINEERING CHEMISTRY

96

VOL. 32, NO. 1

tilled over P20s; only that portion t was r e t a i n e d w h i c h boiled b e t w e e n 79.9' a n d 80.1" C. Before charging into the apparatus it was boiled to remove dissolved gases. The w o r k i n g out of a satisfactory p r o c e d u r e for liquid sampling required considerable experimen tation. Direct expansion from the equilibrium vessel to an analytical train M A T E R IA L REMARKS gave inconsistent D E T A I L S FOR P h 6 A R E results. Possibly TOP HEAD DITTO I N CIRCULAR NO 61 U S D E P T the gas evolved GLAND N U T DITTO O F AGRICULTURE(^> PRE-SATURATOR BODY 0 ITTO when the presS T W I G H T - WAY VALVE DITTO sure was released SATURATOR BODY DITTO failed t o c a r r y along all t h e FIGURFI 2. DETAILS OF SATURATORS AND EQUILIBRIUM VESSEL liquid t h a t b e longed with it. This led to the and Krase (e) in their study of the system water-nitrogen at development of a separate pressure sampler, which gave conelevated pressures and temperatures, but with important differsistent results only after many trials had evolved a special ences in detail. technique. The main features of this technique were very PROCEDURE. With the particular dynamic method used, slow sampling and flushing of the contents of the sampler. the liquid phase is continually exposed to more gas and is Some trouble was at first encountered in vapor-phase bound to come to equilibrium if sufficient time is allowed. sampling because of the presence of considerable liquid in the The situation is different in the case of the gas phase because line, until it mas realized that the solubility of carbon dioxide i t makes a single pass through a given column of liquid and in benzene mas so great a t the high pressures that the volume must become saturated with the less volatile component durof the liquid phase had expanded to several times the volume ing this relatively short time of contact. It was believed that of the original benzene. When allowance was made for this the best procedure was to over-saturate the gas relative to the increase in volume, no further trouble with gas-phase samples equilibrium temperature by maintaining the first presaturator was encountered. 10' higher than the desired equilibrium temperature, and to Results depcnd on the other saturator and the equilibrium vessel to The runs a t IO", 15", and 20" C. and 150 pounds per square reduce the concentration of benzene to the proper value for inch pressure gave erratic liquid-phase samples with no trend equilibrium. This scheme was followed in most runs but of any kind discernilk We were inclined to attribute this some at 30" C. were made with an approach from the lowt o the presence of two liquid phases, but no proof was obtemperature side. Both procedures gave substantially checktained and the experiments of Sander at 20" C. and at presing results. sures up to 700 pounds per square inch did not indicate any A series of runs was made on a given filling of liquid, first such phenomenon. (Sander worked in glass and hence could at increasing pressure steps and then with decreasing ones. have observed the formation of more than one liquid phase.) Pressures and flow rates were controlled manually. At least Since our apparatus was not adapted to the investigation of a 2 hours were allowed to elapse before any samples were taken. system containing two liquid phases. no further work was Samples of both phases were taken a t I-hour intervals and done a t these temperatures. The results for 30°, 40°, 50°, immcdiately analyzed. A given run was completed when two and 60" C. are shown in the form of pressure-composition consecutive liquid samples agreed within the limit of experiisotherms in Figure 3. mental error. A considerable number of check runs was made to test the The amount of benzene in the sample was obtained from reproducibility of the whole procedure as finally developed, the increase in weight of the U-tubes. The carbon dioxide and the maximum deviation in the composition of any two absorbed in the standard sodium hydroxide solution was dcliquid phases was 0.4 mole per cerit carhon dioxide. In the termined by the usual procedure of precipitating the carmajority of cases the difference was of the order of 0.1 per bonate with barium chloride and titrating the solution with cent. standard acid and phenolphthalein indicator. Figure 3 shows that the agreement with Marsh's data at The carbon dioxide was the usual commercial grade s u p 50" C. is quite good. We are loath t o attempt any complied as a liquid in cylinders and it was employed without parison with Marsh's data at 25" C. because of the uncertainty further purification. The benzene was Merck's c. P., thioof extrapolating data in this region so close to the critical phene-free which was partially frozen, filtered, and then dis-

6'

JANUARY, 1940

97

INDUSTRIAL AND ENGINEERING CHEMISTRY

FIGURE 3. EQuILInRrmPHASE COMPOSITIONS OF THE SYSTEX CARBON DIOXIDE-BENZENE

temperature of one of the components. No wholly satisfactory comparison between our data and those of Sander can be made because the latter reported his results per unit volume of solution, and there are no data on the densities of the solutions. Sander did record the amount of benzene charged into his apparatus, but from a statement elsewhere in the paper one would infer that his figures on volume of solvent are only approximate. Accepting his value for benzene volume, we calculated his results to our basis and found poor agreement as shown in Figure 4. Sander's results are reported for 20°, 3 5 O , 60°, and 100' C., and the results in Figure 4 were obtained by graphical interpolation. It is of interest to see how well the results follow a simple solubility law such as Henry's law which may be expressed in the form,

y P = kx

where y

= z = P = k =

mole fraction of COZin gas phase mole fraction in liquid phase total pressure Henry-law constant

Values of the Henry-law constant, k , were calculated from the experimental data and are as follom-s: Mole Fraction COS in Liquid

30° C.

Henry-Law Constant at: 40° C.

50' C.

60' C.

INDUSTRIAL AND ENGINEERING CHEMISTRY

98

VOL. 32, NO. 1

ing values of IC, useful for approximate calculations, are obtained : L' c. k 30 40

1217 1420 1630

50

60

1913

These data give nearly a straight line when plotted. If Henry's law in the above form is applied to Sander's data, quite different trends are observed. sander reported no vapor compositions and hence it %-asnecessary to assume y = 1 in the Henry-laiv equation.) At 20°,35", and 100" C., k decreases rapidly a t first as mole fraction (or pressure) increases and then approaches a constant value. On the other hand, at 60" C., k is substantially constant over the whole range of the experiments to a maximum pressure of about 100 atmospheres. This comparison of our data with those of Sander doepnot depend on the uncertain value of the volume of benzene used as was the case when we tried t o compare the absolute results. We are unable to account for the difference in the trends of the two sets of data.

Literature Cited FIGURE 4. COMPARISON WITH SANDER'S DATA

(1) Dilley, J. (1929).

R., and Edwards, W. L., U.

(2) Saddington,

It is evident that Henry's law holds fairly well up to solutions as concentrated as 40 mole per cent. Averaging the values of k u p to and including 2 = 0.40 for the first three temperatures, and u p to and including 2 = 0.50 a t 60" C., the follow-

s. Dept.

Agr., C ~ T C61, .

W. A., and Krase, N. W.,J. Am. Chem.

SOC.,56,

353 (1934). (3) Sander, TV., 2. physik. Chem., 78, 513-49 (1912).

THISpaper is based on a dissertation presented by Shen-wu Wan in June, 1935, t o the faculty of the Graduate School of Yale Univereity in candidacy for t h e degree of doctor of philosophy.

2-Aminophenylanthraquinone and 1,2-Phthaloylcarbazole P. H. GROGGINS

0

Bureau of Agricultural Chemistry and Engineering, United States Department of Agriculture, Washington, D. C.

FEW years ago the writer reported on the preparation of /3-phenylanthraquinone-andits derivatives ( 2 ) . H e . showed that 4'-chloro-/3-phenylanthraquinone could

A

readily be converted to the corresponding amino compound by reacting with aqueous ammonia, giving a product which by titration was about 98 per cent pure. I n the case of the similar ammonolysis of 2-chloro-/3-phenylanthraquinone,the author was constrained to report that "efforts to titrate the 2-amino derivatives were unsuccessful, the results being uniformly low". The nitrogen content, horr-ever, was close to the theoretical. When 2'-amino-/3-phenylanthraquinonewas obtained by the cyclization of 2-amino-p-phenyl-o-benzoylbenzoicacid (S),a brown crystalline product, melting a t 200-201" C., was obtained. The crude product resulting from the ammonolysis of 2'-chloro-/3-phenylanthraquinonewith aqueous ammonia containing nitrobenzene and a copper catalyst also gave evidence of the presence of some amine, either by titration with sodium nitrite or by diazotizing and coupling with R salt. When this crude product was recrystallized from toluene and o-dichlorobenzene, long, silky, orange-colored needles (melting at 255" C.) were obtained. This crystalline product is not 2-amino-P-phenylanthraquinone,but 1,2-phthaloylcarbazole:

Obviously a ring closure occurred between the amino group and the 1 position of the anthraquinone nucleus, probably as a result of the oxidizing influence of the nitrobenzene. Weinmayr of the du Pont Company pointed out to the writer that under the conditions employed, the formation of 1,2-phthaloylcarbazole is not entirely surprising. Weinmayr also noted that the recrystallized product (melting at 255" C.) obtained b y us, corresponds to the melting point given for 1,2-phthaloylcarbazole, derived by heating I-aanthraquinonylbenzotriaaole in diphenylamine (I). It dissolves in concentrated sulfuric acid with a deep blue color and shows an instantaneous color change when a drop of nitric acid is added to the sulfuric acid solution; this is a reaction characteristic of the carbazole structure.

Literature Cited (1) Beilstein's Handbuch der organischen Chemie, Band XXI. p. 428, system 3230, Berlin, J. Springer, 1935. ( 2 ) Groggins, P. H., IND.ENG.CHEY.,22, 620 (1930). 3 . Patent .1,814,149 (1931). (3) Groggins, P. H., IT