Foam Fractionation under Total Reflux Recovery and Concentration of Ionic Species from Aqueous Solution BARRY L. KARGER, RUSTOM P. PONCHA, and MICHAEL M. MILLER Department of Chemistry, Northeastern University, Boston, Mass.
A new and improved apparatus, employing total reflux, has been developed for the recovery and concentration of ionic species from aqueous solution by foam fractionation. The foam is thermally broken at the top of the column by passing steam through a Friedrichs condenser. Methyl orange, 5 X lO-'M (0.17 p.p.m.1, can be completely recovered. A second apparatus, employing total reflux, is used with heat-sensitive solutes. In this case the foam is mechanically broken external to the column, and the collapsed foam liquid is recycled into the column.
ns
Steam in
IO
IlllSl
L-1-
J
F
fractionation is rapidly becoming effective for separating both surface-active (14) and non-surface-active species (6). A great deal of work has been reported on the removal of surfactants from waste streams (4) and radioactive metallic ions from nuclear wastes (13). Considerably less attention has been paid to conditions for recovery and concentration of ionic species, especially at the trace impurity level. Foam fractionation, however, should be useful for recovery and concentration, for the system operates more efficiently as the bulk concentration decreases (11). In other words, quantitative removal and accumulation of solutes in the foam should be possible. In a previous paper (9) a technique was described for increasing recovery and concentration in a foam column of ionic species which do not themselves form stable foams. A surfactant, oppositely charged to the solute, was added to the solution to produce the foam and carry the solute into the foam column. Enrichment occurred by means of total reflux, in which the foam at the top of the column was broken and the liquid from the collapsed foam was returned over lower layered bubbles. This process increased drainage of bulk liquid entrapped in the foam and increased reflux, interchange a t gas-liquid interfaces of more surface active with less surface active species. [The mechanism of this process has been described (9).] During the total reflux period, the foam was thermally broken with a heating tape wrapped around the OAM
764
ANALYTICAL CHEMISTRY
Figure 1. Apparatus for foam fractionation with total reflux
ing, in certain cases, in the formation of a hard-to-remove solid on the column walls. This effect, while negligible for 1-naphthylamine, was significant when solutes which could easily adsorb on the glass walls were used. Such solutes gave poor recoveries, even though completely removed from the bulk solution. A second problem arose from the fact that a finite time was necessary for the heating zone to cool to room temperature after the total reflux period. As a consequence, partial reflux occurred while the heating zone cooled, and the reflux time period could not be closely controlled. This factor was unimportant for long reflux periods, but contributed to nonreproducibility in the short total reflux periods. To minimize the problems, a new apparatus for foam fractionation under total reflux has been developed. In this paper we illustrate the usefulness of this apparatus and report the development of a second apparatus for foam fractionation under total reflux applicable to heat-sensitive molecules. EXPERIMENTAL
1. 2.
Two-way stopcock Porous glass frit, coarse, 15 mm. Kimax 28280 3. Funnel Foam column section, 2 0 cm. X 3 6 mm. 4. 4 0 / 5 0 standard-taper joints 5. Foam column proper, 2 3 cm. X 34 cm. 6. 2 4 / 4 0 standard-taper joints 7. Friedrichs condenser, Corning 2 6 4 0 8. Tube adapter, 1 OSo, Corning 8 8 4 0 9. Cone drive stirrer, Lo Pine 383-31 10. Glass collector funnel, 2 0 X 13 cm., 1 30°, 8-mm. stem 1 1. Wire basket, 7 X 9 cm. 12. Graduated cylinder, 1 0-ml.
top of the column. After this period, the heat was removed, and an appropriate volume of collapsed foam was collected. By this procedure, 1 X 10-4M l-naphthylamine could be completely recovered and concentrated 100-fold. Other workers have also shown the benefits of reflux in enrichment (2, 8, 1.2). The use of a heating tape to break the foam, however, presented several problems. In the first place the temperature of the column walls tended to be higher than a t the center of the column. Excessive drainage and evaporation occurred along these walls, result-
Apparatus. The improved apparatus for controlled recovery by foam fractionation is shown in Figure 1. The funnel foam column and 15-mm. porous coarse glass frit were similar to those used in the previous study (9). However, the length of the column proper was reduced from 90 cm. to 23 cm. The foam was thermally broken by passing steam through a Friedrichs condenser attached to the column proper. When the reflux period was concluded, the Friedrichs condenser was rapidly cooled with tap water and the gas flow was increased. The foam passed from the condenser into an adapter and thence into the foam breaker section. The foam breaker consisted of a spinning wire basket similar to that described by Rubin (10). The liquid from the collapsed foam was collected by a glass funnel and allowed t o fall into a 10-ml. graduated cylinder. Since solvent evaporation was minimal during total reflux, a cold condenser was not used during the reflux period. Procedure. I n a given run, a 150ml. solution of solute was pipetted into the funnel foam column section. The required amount of surfactant was then dissolved in this solution.
Nitrogen gas was passed through a water saturator (to prevent spurious evaporation effects) and entered the solution through the coarse porous frit. The nitrogen flow rate was established a t 42 ml. per minute for the initial foam rise and reflux period and increased to 60 ml. per minute for collection purposes. The flow rate was measured upstream from the foam column by means of a soap bubble flowmeter. Unless otherwise specified, the volume of collapsed foam collected was 4 ml. Chemicals. Methyl orange was purified by crystallization from an acetone-water solvent mixture. The cationic surfactant, hexadecyltrimethylammonium bromide (Matheson, Coleman and Bell, technical grade), and the anionic surfactant, sodium lauryl sulfate (Fisher Scientific Co., U.S.P.) were also recrystallized from an acetone-water mixture, and both were dried in a vacuum desiccator. The phospholipid, vegetable lecithin (Nutritional Biochemicals Corp., 12-803), was used without pretreatment or purification. Analysis Procedures. The concentration of collected fractions of methyl orange was determined colorimetrically with a Beckman DU2 spectrophotometer. The solution was first adjusted to p H 1 to prevent interference from the surfactant. The wavelength maximum a t p H 1 was 504 mp, with an extinction coefficient of 4.53 X lo4. The phospholipid was analyzed for total phosphorus according to the procedure outlined by Bartlett (!). The hexadecyltrimethylammonium bromide was determined by the two-phase titration procedure of Cross (?), using sodium tetraphenyl boron as titrant. RESULTS AND DISCUSSION
The use of the Friedrichs condenser was a definite improvement over the heating tape apparatus. In the first place, during reflux, heat is applied internally over a wide portion of the foam column rather than a t the outer walls alone. Thus it is possible to apply less heat to break the foam, decreasing the previous problem of zones of excessive drainage and evaporation. Solid adsorption on the glass walls is lessened, resulting in a larger number of solutes which can be recovered. Secondly, after the total reflux period, the reflux zone is rapidly cooled by passing cold tap water through the condenser. Thus the reflux period is more closely defined, and this should produce more reproducible data for this apparatus than for the heating tape apparatus, especially for short reflux periods. For long reflux periods both systems should be reproducible. Recovery of 10-6M Methyl Orange. Methyl orange (MO) could not be successfully recovered with the heating tape method because of excessive adsorption on the glass walls; how-
OY
I
20
Figure 2. system
40
60
,
I
80
180
I60
Recovery of methyl orange as a function of time of nitrogen flow through MO = 10-'M HDT = 7 x 1 0 - 4 ~ pH = 10
ever, if lO-5M MO and 7 X 10-3M hexadecyltrimethylammonium bromide (HDT) are used at pH 10, Figure 2 shows that methyl orange can be recovered with the Friedrichs condenser apparatus and that total reflux is a definite aid to enrichment and recovery. The recovery measure is:
R=
I4 0
120
100 Time in m i n v w s
wt. collected wt. original
x 100
(1)
The total time from the start of nitrogen flow is plotted in Figure 2. The foam took 20 minutes to rise to the reflux zone, a t which point the total reflux period began. Approximately 80% of the methyl orange is recovered within a 1-hour total reflux period and after this period there is a slow rise in enrichment with time up to %yorecovery in a 3-hour total reflux period. It is not convenient to operate beyond 3 hours, because the foam column cannot be maintained. The slow increase in recovery for long reflux times and the fact that less than lOOyo is recovered are believed to be due to the slow removal of adsorbed methyl orange from the glass surfaces and not to its slow removal from the solution. In support of this belief, 100 ml. of bulk solution were evaporated down to 5 ml. after a 90-minute reflux period; subsequent analysis revealed no detectable methyl orange. A 10-4M 1-naphthylamine solution with lO-3M sodium lauryl sulfate a t pH 2 was next foam-fractionated with this apparatus. As with the heating tape method, essentially complete recovery occurred after a 1-hour reflux period. This further indicates that the incomplete recovery of methyl orange is due to the solute itself and that with appropriate compounds complete recoveries are possible with the apparatus of Figure 1.
The reproducibility was considerably better with the Friedrichs condenser than with the heating tape technique. While reproducibility for the heating tape method was 20y0 a t low reflux times with 1-naphthylamine, the Friedrichs condenser experiments showed a reproducibility within 5y0. Total Reflux as an Aid in Concentration of Solutes. The total reflux foam fractionation technique can be used for concentrating as well as recovering ionic solutes. A solution of 1 O - b M MO and 7 x 10-4A4 HDT at pH 10 was foamed with total reflux for 1 '/z hours. The collapsed foam after reflux was collected in 0.5-ml. samples up to 4 ml. rather than as one 4-ml. sample (Table I). The enrichment factor, E , is equal to
E= concentration of MO in collected foam concentration in original (2) A 200-fold enrichment of methyl orange has been effected in the first 0.5 ml. This enrichment factor should be compared with an E value of 10 for methyl orange previously obtained on 0.5-ml. samples (6). Thus the total reflux technique can be very effective for concentration as well as recovery. Table 1.
10-6M
Enrichment and Recovery of Methyl Orange in 0.5-MI. Samples
Sample No. 1 2
E 210 31 __
5.0 3.6 2.1 1.8 1.5 0.89
%R 70 10 .. 1.7 1.2 0.69 0.60
0.50 0.29
VOL 30, NO. 6, MAY 1966 e
765
Table II. Effect of Concentration of Methyl Orange on Recovery
Reflux time, min.
5 x 10-7~ 1 x 10-6M 5 x 10-7~ 1 x 10-5~ 5 x 10-7~ 1 x 10-6M 5 x 10-7~ 1 x 10-6M
30 60 90
Table 111.
yo R
Concentration
0
30 26 74 65 92 76 98 81
Comparison of Medium and Coarse Frits
Frit, mm. Medium, 35 Coarse, 15 Medium, 35 Coarse, 15 Medium, 35 Coarse, 15
Reflux time, min . 0 0 5 5
10 10
%R 42 (in 3 ml.) 28 (in 4 ml.) 54 (in 3 ml.) 38 (in 4 ml.) 70 (in 3 ml.) 47 (in 4 ml.)
There is a rapid decrease in solute concentration with each increment of the 4 ml. collected, with 94y0 of the total material in the first milliliter (Table I). In the reflux zone methyl orange is highly concentrated and the collection process causes a dilution. It was necessary to collect 4 ml., so that some of the collapsed foam could wash methyl orange from the collector funnel into the graduated cylinder. With suitable modification of the collection apparatus, it may be possible to effect the recovery in volumes smaller than 4 ml. Effect of Concentration of Solute on Recovery. Since removal becomes more efficient as the concentration of solute decreases, for any given reflux period before total removal, the more dilute the initial concentration of solute the higher should be the recovery, other things being equal. To test the effect of concentration on recovery, a solution of 5 x ~ o - ~MO M and 7 X 10-4M H D T a t pH 10 was foamed for various reflux periods. These results are compared with the foam fractionation of a lO-SM MO solution in Table 11. The yoR is larger a t all reflux times for the less concentrated material. In this particular case, the results reflect not only the more efficient removal of solute a t the lower concentrations, but also the decreased adsorption of the solute on the glass walls for the 5 X lO-7M solution. Both effects will result in a higher recovery of the 5 X lO-7M MO solution over the lO-5M MO solution. Essentially all of the 5 x lO-7M MO is recovered in 1 '/2 hour's reflux. The 5 x lO-7M solution represents an impurity level of only 0.17 p.p.m. 766 *
ANALYTICAL CHEMISTRY
Modification of Apparatus for More Rapid Removal and Recovery. I n all the experiments reported so far considerable time was expended to produce high recoveries-for example, it took 1 hour of total reflux to recover 80% of 10-5M MO. In an effort to decrease reflux time, the 15-mm. porous glass frit of the apparatus in Figure 1 was replaced by a 35-mm. porous medium glass frit. For a given gas volume, it was visually noted that more and smaller gas bubbles were produced for the medium frit. A 35-mm. fine porous glass frit was also tested but later abandoned because of the excessive pressures necessary to force gas through the frit. A comparison of % R for the coarse and medium frits as a function of total reflux time is shown in Table I11 for a flow rate of 30 ml. per minute. To assist in foam breakage of the wet foam produced by this frit,asbestos insulation was placed around the Friedrichs condenser. The test mixture was 1 X 10-5M MO and 1 X 10-3M HDT a t pH 10. A more rapid recovery is possible with the medium frit. The result is even more striking because the yo R is greater for the medium frit than the coarse frit in s/4 the collected volume. The improved speed of recovery for the medium frit is due to the increased rate of removal of methyl orange from the solution as a result of the greater gas-liquid surface area and the greater amount of bulk liquid entrained in the foam. While the medium frit is advantageous in concentration and recovery problems, the coarse frit of narrow diameter is probably better suited for selective removal of solutes from the bulk solution. A small number of large bubbles would allow more selective adsorption because of decrease in gas-liquid interface and would 'carry less entrapped bulk liquid than a large number of small bubbles. Isolation of Solute from Surfactant in Foamate. The product collected in the collapsed foam contains solute and excess surfactant and a further experiment must be performed to recover pure solute. This separation is most conveniently carried out by an ion-exchange process in which the surfactant is retained on an ionexchange resin. The isolation of pure solute from surfactant was tested with a solution of 5 X 1 0 d 4 M M O a n d7 X 10-3MHDT a t p H 10. One gram of Amberlite I R 120 CP (Rohm and Haas) high porosity resin in the hydrogen form was first converted to the sodium form with 6M NaOH. The solute was separated from surfactant in a batch process by mixing 3 ml. of solution with 1 gram of resin. After equilibration, the solution was decanted and the resin was washed, to give a final volume of 10 ml. Analysis
I
Figure 3. Total reflux foam fractionation with mechanical breakage 1.
Foam column No. 9 two-hole rubber stopp'er 1 1-mm. borosilicate glass tubing Cone drive stirrer, La Pine 383-31 Plexiglas cover Wire basket, 7 X 9 cm. Glass collector funnel, 2 0 X 1 3 cm., 1 3 0 ° , 2-mm. stem 8. 3-way stopcock, 2 mrn. 9. Reflux dripper
2. 3. 4. 5. 6. 7.
of the h a 1 solution revealed all the methyl orange and no detectable HDT. Thus ion exchange is a simple and effective method for isolation of solute from surfactant in the foamate. Recovery of Heat-Sensitive Solutes. Thermally labile compounds cannot be recovered in the apparatus in Figure 1, for foam breakage must be accomplished by some means other than heat. While such procedures as bombardment with a-particles (7) or the use of sonic beams (6) have beem employed to break foams, it was felt that for analytical scale operation a mechanical method would be simplest and most convenient. The mechanical breaker shown in Figure 3 was thus devised for heatsensitive molecules. The foam traveled out of the column up the 11-mm. tubing into the spinning wire mesh basket, where it was broken. The collapsed foam then dripped back into the column a t a controlled rate by means of a three-
way stopcock. (The back-pressure exerted by the column would also play a role in the rate of returned flow; however, this aspect was not investigated in these studies, as the flow rate was the same in each run.) The stopcock also directed the liquid into a graduated cylinder when collection was desired. Since the liquid from the foam may be in the collector funnel for a substantial time, a Plexiglas cover was placed atop the glass funnel to prevent evaporation by air convection from the spinning wire basket. To increase rate of recovery, the medium frit was used with this apparatus. By suitable regulation of the threeway stopcock it was possible to maintain any reflux ratio desired. When the foam was broken inside the column, as in Figure 1, the reflux ratio must always be infinity; however, externally breaking the foam allowed adjustment of the reflux ratio. The experimental design of Figure 3 does not differ greatly from that of Lemlich and his coworkers. In one case Brunner and Lemlich (2) pump the externally collapsed foam into the top of the foam column. In a second case Lemlich and Lavi (8) gravity-feed a portion of the externally collapsed foam into the center of the column. In our apparatus the foam is broken externally and returned by gravity to a point close to the top of the foam column. Preliminary studies with the mechanical breaker total reflux apparatus were first made with 10-5M MO and l O - 3 M H D T a t pH 10. Results comparable to those with the Friedrichs condenser apparatus were obtained, but the apparatus in Figure 3 required a slightly larger collection volume for successful recovery of methyl orange. Runs were then made with a heatsensitive solute, which for the most part was the phospholipid, vegetable lecithin. This phospholipid molecule is zwitterionic a t neutral pH, so that for successful removal and recovery of the solute, a zwitterionic surfactant must be used or the p H must be altered to neutralize
Table IV. Recovery of Lecithin as a Function of Total Reflux Time ( 2 X 10-SM sodium lauryl sulfate, 10-6M lecithin, pH 1, collect 10 ml.)
Reflux time, min. n 15 30 90 120
%R 23 ~31 36 64 73
creases. In liquid-liquid extraction the ability to extract solutes from an aqueous phase and recover them in a nonaqueous phase is lower as the solute concentration is decreased. Likewise in chromatography, recovery becomes more difficult a t low concentration levels. However, in foam fractionation removal and recovery seem to improve as the concentration of the solute is decreased. LITERATURE CITED
one of the charges. The latter course was chosen in this work. Since hydrolysis of the phospholipid occurs readily in basic media, it was decided to lower the pH to neutralize the phosphate anion and use as surfactant sodium lauryl sulfate (anionic). pH 1 was selected for this study rather than more acidic pH’s which might decompose the phospholipid and affect foam stability. At pH 1 a substantial portion of the lecithin was recovered. A solution of 10-5M vegetable lecithin and 2 x l O + M sodium lauryl fate a t pH 1 was used to determine the usefulness of the total reflux apparatus in Figure 3. The effect of reflux time on recovery in 10 ml. of collapsed foam is shown in Table IV. Recovery improves with reflux time and thus the apparatus is effective for recovery of heat-sensitive molecules. The increased time for a given recovery relative to methyl orange in Figure 2 is probably a reflection of the reflux apparatus and the lecithin itself. CONCLUSIONS
In the new and simplified total reflux apparatus fractionation is able to remove, recover, and concentrate solutes a t the trace concentration level from aqueous solution. In many of the separation techniques currently available on an analytical scale, recovery of separated fractions decreases as the quantity of solute de-
(1) Bartlett, G. R., J . Biol. Chem. 234, 466 (1959). (2) Brunner, C. A., Lemlich, R., Znd. Eng. Chem. Fundamentals 2, 297 (1963). (3) Cross, J. T., Analyst 90, 315 (1965). (4) Grieves, R. B., Crandall, C. J., Woods, R. K., Intern. J. Air Water Pollution 8 , 501 (1964). (5) Haas, P. A., Johnson, H. F., A.Z.Ch.E. J. 11. 319 (1965). (6) Karger, B. L.; Rogers, L. B., ANAL. CHEM.33,1165 (1961). (7) Kato, R., Toshihiko, K., J . A p p l . Phys. 34, 708 (1963). (8) Lemlich, R., Lavi, E., Science 134, 191 (1961). (9) Poncha, R. P., Karger, B. L., ANAL. CHEM.37,422 (1965). (10) Rubin, E., Ph.D. thesis, Columbia University, 1963. (11) Rubin, E., Gaden, E. L., Jr., “Sepa-
ration Techniaues,” H. M. Schoen,. ed... p. 319, Interscience, New York, 1962. (12) Schnepf, R. W., Kevorkian, V., nersonal communication. in Eldib. r - - - I. A., “Advances in Petroleum Chemicd Refining,” Vol. 7, p. 98, K. A. Kobe and J. J. McKetta, Jr., eds., Interscience, New York, 1963. (13) Schonfeld, E., et al., U. S . Atomic Energy Comm., Rept. NYO-9577 ’
(1960). (14) Skomoroski, R. M., J . Chem. Educ. 40,470 (1963).
RECEIVED for review January 10, 1966. Accepted March 14, 1966. Division of Analytical Chemistry, 149th Meeting, ACS, Detroit, Mich., April 1965. Research undertaken in cooperation with the U. S. Army Natick (Mass.) Laboratories under contract No. DA 19-129-AMC302(N), assigned No. TP-25 in the series of papers approved for publication. The findings in this report are not to be construed as an official Department of the Army position.
VOL. 38, NO. 6, MAY 1966
767
