Studies of Recovery by Foam Fractionation on 1-Naphthylamine SIR: The technique of foam fractionation for the separation of substances has been known for a long time (4). Foam separation takes advantage of the surface activity of certain solutes a t a gas liquid interface. When a solution is foamed, the active components collect a t the surface or foam layer, which can then be removed and broken to give a rich liquid product. An excellent review of the whole field can be found in an article by Rubin and Gaden (7). Initially, only surface active substances which formed stable foams were separated by foam fractionation, and the applications were therefore somewhat limited. More recently it has been found that even nonsurfaceactive compounds can be separated by this technique. For example, metal ions have been separated by adding to the solution a foaming agent which can complex the metal ions and carry them into the foam (8). Karger and Rogers (3) have separated organic compounds which do not form stable foams by themselves. They showed that in the case of an ionic solute, the component could be concentrated most effectively by using an ionic foaming agent of opposite charge. In this paper, a study has been made of the recovery of an ionic organic solute, protonated 1-naphthylamine, which does not form a stable foam. An oppositely charged surfactant, sodium lauryl sulfate, was used to produce the foam and carry the organic solute into the foam column. A new and simplified method has been developed to increase recovery. Foam fractionation is a particularly good method for recovery problems of ionic, nonsurface-active solutes due to the fact that the technique becomes more effective a t low concentrations as seen from the Gibbs adsorption equation ( 7 ) . I n other words, quantitative removal and accumulation of solutes in the foam should be easily achieved. Lemlich and Lavi (5) have shown that the enrichment of surfactant increases considerably with the increase of external reflux. Total reflux, in which the foam is broken and returned as liquid to the top of the foam column, has been tested by Schnepf and Kevorkian (9) and also by Brunner and Lemlich ( 1 ) ; in both cases considerably higher enrichments were obtained. Rogers and Olver (6) have pointed out that the increased enrichment of a solute in the foam may be due not only to reflux but also to the drainage of bulk liquid from the foam. I n these experiments the conditions have been 422
ANALYTICAL CHEMISTRY
A.
B. 1. 2.
3.
4. 5. 6.
7. 8. 9.
Figure 1. Apparatus for foam fractionation Apparatus for foam collection Two-way stopcock Porous glass frit, coarse, 15 mm., Kimax, 2 8 2 8 0 Funnel foam column section, 3 6 mm. X 2 0 mm. 40/50 standard tapered ioints. Foam column, proper, 36 mm. X 90 mm. Heating tape, Briskeat %” X 4’, 1 9 2 watts 19/22 standard tapered iolnts Reflux condenser, Metroware, ME4 1 7 - 8 3 Recovery tube, 3 6 mm. X 2 7 mm., 7 5 ’ bend
designed to increase both drainage and reflux and thus recovery. EXPERIMENTAL
Figure 1 shows the apparatus used in this work. In each run, 300 ml. of solution were pipetted into the column. Nitrogen gas was passed through a saturator (to prevent spurious evaporation effects) and entered the solution through a coarse porous glass frit. The flow was adjusted to 30 ml. per minute by means of a calibrated rotameter. A heating tape (Briskeat x 4’, 192 watts) controlled by a variac was wound near the top of the column, and the temperature in the
column was maintained a t about 105’ C. The foam rose to the top and broke in the heated section, the liquid flowing down as reflux. A small condenser was fitted a t the top of the column to prevent any vapors from escaping. Reflux time began a t the moment foam reached the heated section. After the reflux period, the heater was switched off, and the condenser was removed from the top of the column and replaced by the collector. The collector was also heated by a similar heating tape and fitted with a condenser to prevent any vapors from escaping. The gas flow was increased to 60-70 ml. per minute, and 3 ml. of the collapsed foam were collected in a graduated cylinder. The solutions used in the experiments were 2 x 10-3,?4 sodium lauryl sulfate and lO-‘M 1naphthylamine adjusted to pH 2 with HC1. Sodium lauryl sulfate was recrystallized from a water acetone mixture, 1-naphthylamine from toluene, and both were dried in a vacuum desiccator. The concentration of l-naphthylamine in the foamate was determined spectrophotometrically after the 3 ml. of the collected liquid were diluted to the required volume with HC1 a t pH 2. The spectrophotometer used was a Hitachi Perkin-Elmer Model 139. The wavelength for maximum absorbance was 277.5 mp and the extinction coefficient 5.8 X lo3. It was determined that sodium lauryl sulfate did not interfere with the analysis a t this wavelength. The amount of sodium lauryl sulfate collected in the 3 ml. was determined as sulfate by the wet oxidation method of Du Bose and Holland (2). The surface tension determinations were made with a Fisher Model 19 du Noiiy Tensiometer. RESULTS AND DISCUSSION
When gas is passed through a solution of a surface active agent, a foam is produced which can then travel up the column. The foam consists of films of bulk liquid and surface active species adsorbed a t the gas liquid interface ( 7 ) . Bulk liquid drains countercurrent to the rising foam, thus thinning the film a t the top of the foam. This drainage is a factor in the concentration of the surface active agent a t the top. The thinning of the film can eventually lead to the breakage of the bubbles a t the top. When this breakage occurs, the surface active agent can then be adsorbed a t the gas liquid interface of the rising layers of bubbles. Interchange can occur between the more surface active and the less surface active species to produce a concentration of the most surface active agent
a t the top of the foam column. This phenomenon is called internal reflux. It should be noted that drainage can also cause interchange of surface active species, and thus reflux, due to the fact that the bulk liquid does contain some of the surface active species. To produce high recovery in foam fractionation, conditions should be devised to increase drainage and reflux. This increase can best be made by breaking the foam artificially a t the top of the column either thermally or mechanically ; however, thermal breakage is often more convenient. Essentially the principle involved is to prevent foam from rising out of the column by breakage at the top during the reflux period. After this total reflux period the breaker must be removed so that foam can go out of the column and be collected. I n thermal breakage, this removal can easily be achieved by switching off the heater whereas in mechanical breakage removal may not be so easy. A41so,in many types of mechanical breakage there is a possibility of trappage of foam or liquid in the breaker. However, when components are thermally unstable a t the foam breakage temperature, the mechanical breaker is preferable. The concentration of the surface active agent, sodium lauryl sulfate ( 2 x 10-3M), was a t the point Just prior to the critical micelle concentration as measured on the du Nouy tensiometer. This concentration results in the greatest foam stability. If higher concentrations are used, micelles would form which would adversely affect recovery and foam stability. If lower concentrations of surfactant are used, the foam column is less stable and less reproducible and it may become necessary to add surfactant during the run in order to maintain the foam column. In these studies the concern is not removal of foaming agent but only using it to remove solute and produce an enlarged gas-liquid interface. For the recoverv measure of wt. collected x 100, the results with wt. original thermal breakage are shown in Figure 2. I t can be immediately seen that after a reflux period of one hour, recovery is complete in the 3 ml. of the collected liquid. This result shows the high recovery possibilities of ionic solutes a t low concentrations by foam fractionation when drainage and reflux are emphazized. I t can be seen that the 1-naphthylamine represents only 14 p.p:m. in the solution. Recent experiments with dyes have shown that high recovery can be accomplished with 10-f-10-6M concentrations (1.40.14 p.p.m.) of solute. Thus the 10-4M concentration certainly does not repre-
IO0
80
* K
Y
60
I-MAPHTHY L AM I N € IO-*M
Y U
SODIUM LAURYL S U L F A T E 2X10-3M
a I-
z
40
L W
a.
2c
C
I
r
I
20
40
60
Figure 2.
I 1 80 100 T I M E IN MINUTES
I
1
1
le0
160
180
Recovery of 1 -naphthylamine as a function of reflux time
sent a lower limit. In these experiments the collected product contains both the surfactant and the solute. If the goal is to obtain the pure solute, it may be done by a relatively easy extraction or the use of an ion exchanger to remove the surfactant. At zero reflux time, about 40y0 of the solute is recovered. This 40y0 recovery represents initial adsorption, internal reflux, and drainage since the foam takes about 15 minutes to reach the top of the heated section of the column. Figure 2 shows that the results are more scattered a t low reflux times. The cause of this scattering may be, for the most part, the rapid change in solute concentration during early reflux. When recovery is complete, the system reaches a time invariant situation for the l-naphthylamine. This situation is not necessarily a steady state, for nearly all the solute is a t the top of the foam and thus no more solute is being adsorbed in the foam layer from the solution. To have a true steady state it would be necessary to have a balancing of addition and removal of solute in the bulk solution. Schnepf et al. (8) have broken the foam and recycled it continuously into the bulk solution until no change occurred in the system. From a theoretical standpoint it is helpful to produce such a steady state, but if one wishes to recover components completely it does not seem necessary to recycle. During the longer reflux periods a solid was observed a t the top of the foam column because of the high concentration. This solid may be either the salt 1-naphthylamine hydrochloride or the surfactant-solute complex; both species are stable a t the temperature of the heated section of the column (105' C). It is possible to skim off the solid; however, a more efficient and
reproducible method is described in the experiments. To test whether 3 ml. were necessary to collect all the 1-naphthylamine, runs were made with a reflux period of one hour in which 1-ml, fractions were collected in succession up to 5 ml. The results indicated very poor reproducibility in the lst, 2nd, and 3rd ml. because of the carry over of solid and the dissolved 1-naphthylamine could not be maintained a t the same rate each time. However, it was found that in a total volume of 3 ml., essentially all the material was collected. Analysis of the sodium lauryl sulfate in the collected volume was also made in a few cases. During a one-hour reflux period, roughly 10% recovery of the surfactant could be made in the 3 ml. There was not a significant difference between the cases in which 10-4M 1-naphthylamine was present and those in which it was not present. To produce a higher recovery of the surfactant, it would be necessary to collect a much larger volume of foam because of the higher concentration of the surfactant. Thus the technique of thermal breakage is an aid in recovery and concentration not only for solutes, but also for the foaming agents themselves. In their production of a steady state, Schnepf et al. (8) have defined an enrichment ratio as: concentration in collected foam/concentration in the remaining bulk solution. With this measure the enrichment ratio goes from 68 a t zero reflux to infinity at complete recovery in the experiments ot this paper. When there is complete recovery, the enrichment ratio is not a good measure of the concentrating effect, as has been previously noted by Brunner and Lemlich (1). However, if a concentration ratio defined as conVOL. 37, NO. 3, MARCH 1965
* 423
centration of collected fosm/concentration of the original solution is used for the 1-naphthylamine, some measure of the concentrating effect can be obtained. With this Foncentration ratio, the in this work go from 40 at Ier0 reflux to loo at recovery. If a volume smaller than 3 ml. is collected, the concentration ratio would be higher. LITERATURE CITED
(1) Brunner, C. A., Lemlich, R., Ind. Eng. Chem. Fundamentals 2, 297 (1963).
(2) DuBose, B., Holland, v. B., Am. DyestugRe?J.34, 321 (1945). (3) Karger, B. L., Rogers, L. B., ANAL. cHEY. 33, 1165 (1961). (4) Kenrick, F. B., J . Phys. Chem. 16, 513 (1912). ( 5 ) Lemlich, R., Lavi, E., b’cience 134, 191 (1961). (6) Rogers, L. B., Olver, J. W., Ibid., 135, 430 (1982). \ - - - - I
(7jRuben, E., Gaden, E. L., Jr., “Separation Techniques,” H. M. Schoen, ed., D. 319. Interscience. New York. 1962. (8) Schnepf, R. W., kt al., Chek. Eng. Prog. 5 5 , 42 (1959). ( 9 ) Schne f , R. W., Kevorkian, V., personal 8on~munication, Eldib, I. A,,
“Advances in Petroleum Chemical Refining,” Vol. 7, K. A. Kobe, McKetta, J. J., Jr., eds., p. 98, Interscience, New York, 1963. RUSTOM P. POHCHA B.4RRY L. KARQER Department of Chemistry Northeastern University Boston, Mass. 02115 This work was supported by the Basic Research Fund of h’ortheastern University and the Food Division of the U. S.Army Natick Laboratory under Contract No. DA 19-129-AMC-302(N ).
Ultraviolet Spectrophotometric Determination of Trace Quantities of Phosgene in Gases SIR. Ai spectrophotometric method for the determination of phosgene in gases has previously been described ( 1 ) . This method depends upon the absorption of 1,3-diphenyIurea formed when phosgene is allowed to react with aniline in aqueous solution. It is possible, by incorporation of a n extraction step, to increase the sensitivity fold. Aside from a color reaction of phosgene with 7-(nitrobenzyl)pyridine, described by Lamouroux ( d ) , the modified method is the most sensitive spectrophotometric one yet available for traces of phosgene. APPARATUS AND REAGENTS
A Cary recording spectrophotometer, Model l l M S , with matched 1-em. silica cells, was used for absorbance measurements. A manually operated spectrophotometer may be used, if enough points are plotted. The 1,3-diphenylurea was a recrystallized product, melting point 235.5-7O C. Freshly distilled aniline was used to prepare solutions in water containing 2 mg. of aniline per ml. Phosgene (99.5 mole %) was obtained from the Matheson Co. The 1-pentanol was Eastman practical grade. The chlorine, nitric oxide, and carbon monoxide mere commercial grade gases. All other reagents were of analytical quality. CALIBRATION
Prepare a solution of 100 mg. of 1.3-diphenylurea in 100 ml. of methanol. Dilute a IO-ml. aliquot to 1000 ml. with water. Add a 1-ml, aliquot of the aqueous solution to a separatory funnel containing 25 ml. of the aniline solution. Add 1 ml. of concentrated sulfuric acid and approximately 9 ml. of a 1 : l mixture of 1-pentanol and nhexane. Shake the separatory funnel for 3 minutes, allow the layers to separate, and make the organic layer to volume in a 10-ml. volumetric flask by washing the separatory funnel with a small volume of the solvent mixture. 424
e
ANALYTICAL CHEMISTRY
Scan the spectrum of a portion of this solution in the ultraviolet from 290 to 215 mp, using the hexane-pentanol mixture as a reference solution. Draw a base line tangent to the curve at approximately 280 mp and 220 mp. Read the absorbance at 257 mp and divide the diphenylurea concentration by the absorbance to calculate a coefficient C. This should have a value of about 68.5 pg. per 10 ml. per absorbance unit when a‘ 1-cm. light path is used. Bubble the gas under investigation through 25 to 50 ml. of the aqueous aniline solution at such a rate that not more than 2 mg. of phosgene are passed per minute ( 1 ) . Transfer the solution to a 125-ml. separatory funnel; add 1 ml. of concentrated sulfuric acid and about 9 ml. of hexane-pentanol solvent mixture. Extract and make to a 10-ml. volume. If necessary, take an appropriate aliquot and dilute to volume with the solvent mixture. Scan the spectrum of a portion of this solution in a 1-cm. absorption cell from 290 to 215 mp. Determine the net absorbance at 257 mp as above. Call this quantity A . Milligrams of phosgene in 10 ml. of solution = A X C X 0.466. The factor 0.466 represents the molecular weight ratio of phosgene to diphenylurea. RESULTS AND DISCUSSION
To test the method, small quantities of phosgene were weighed in sealed capillary tubes. These tubes were broken under n-hexane in which phosgene is very soluble. The resulting solution contained 18.4 pg. of phosgene per ml. The phosgene content mas initially 17.7 p g , per ml. by the extraction method. This decreased to 15.6 pg. per ml. in 3 days, after which it remained constant for a period of several weeks. This initial change in concentration was probably caused by reaction of the phosgene with solvent impurities. The limit of detection of phosgene per
10 ml. of solution is approximately 1 pg. In air, 0.01 p.p.m. (1 p.p.m. = 4.05 pg. per liter of air) can be determined by blowing about 30 liters of air through the aniline solution and then using the extraction procedure. Two different techniques were used to allow the phosgene to react with aniline in the aqueous solution. First, portions of the standard solutions were vaporized and aspirated through the aniline solution. In a second approach portions of the standard solutions were shaken with the aniline solutions. The results are summarized in Table I. Although the results are quantitative in only a few instances, they show better than 90% recovery in most cases. This is sufficient for practical applications. Difficulties encountered in handling phosgene solutions quantitatively and in assuring complete contact of phosgene with aniline account for the low results. They are not caused by the extraction which is shown in Table I1 to remove 1,3-diphenylurea quantitatively. The pH of the aniline solution may change when combustion products or vent gases are bubbled through it. The effect of pH on the reaction of phosgene with aqueous aniline was investigated by shaking 1.00 ml. of a n-hexane solution containing 15.6 pg. of phosgene with 50 ml. of the aqueous
Table I.
Determination of Phosgene
Aspirated samples Shaken samples Phosgene Phosgene Phosgene Phosgene taken, found, taken, found, rg.
rg.
6 6 7 7
5 5 6 7
2 2 7 7
8 9 9 1
rg.
1 3 9 13 26
3 1 4 2 4
PG.
1 2 9 13 26
0 8 1 2 3
