V O L U M E 2 1 , NO. 11, NOVEMBER 1 9 4 9 tivity of the device; this rate is controlled by manual setting of stopcock E.
1417 gave control within + 2 mm., and the resulting fractions proved pure.
This type of regulator has the advantage of easy operation with LITERATURE CITED
a positive force opposing the pressure of the still head. Elimination of the automatic stopcock of systems such as those described by Booth and Bozarth ( 1 ) greatly simplifies the adjustments necessary during distillation. .4lthough the greatest accuracy of pressure control with this type of regulator is in the region around atmospheric pressures, distillations a t as low as 250 mm. are accomplished easily. The control a t atmospheric pressure for distillations lasting as long as 8 hours, for 50 to 75 ml. of sample,
(1)
Booth, H. S., and Bozarth, A. R., I n d . Eng. Chem., 29, 470
(2)
Booth, H. S., and McNabney, R., IND.ENG.CHEM.,-4x.4~.ED.,
(1937); J . A m . Chem. SOC.,61, 2927 (1939). 16,131 (1944). (3) NcNabney, R., private communication. RECEIVED November 2 4 , 1948. Based upon work done i n connection with a research project sponsored b y the Office of Naval Research.
An Improved Tensimeter-Still WILLIAM P. RATCHFORD AND C. E. REHBEKG Eastern Regional Research Laboratory, Philadelphia 18, Pa. STUDIES of high-boiling derivatives of lactic acid, need INarose for a device which would serve as a simple low-pressure
still and also make possible accurate vapor pressure-i.e., boiling point-measurements in the low-pressure region-via., 0.02 to 10 mm. of mercury,
At pressures below 0.01 to 0.03 mm., noncondensahle gases diffuse against the distillate vapor stream and vitiate boiling point measurements ( 7 ) . Hence it is futile to attempt such measurements a t extremely low pressures, even though some substances can be distilled a t such pressures in the tensimeter-still. If a t approximately 0.03 mm. the boiling point of the distilland is too high to permit distillation without decomposition, it is advisable to use a molecular still, where permanent gases are pumped away efficiently and distillate molecules have a short path to the condenser. The tensimeter-still is similar in appearance to commercial equipment that has been described recently (2); the latter, however, is recommended for use in the range 0.001 to 0.1 mm., and is stated to have a separation efficiency of less than one theoretical plate. The present still gains its usefulness from vigorous agitation of the distilland, obtained by rocking the entire still to and fro. The agitation minimizes or eliminates superheating and inhomogeneities in the distilland, and thus facilitates accurate boiling point measurements and equilibrium distillation. The separation efficiency is at least one theoretical plate. APPARATUS AND METHOD
Figure 1. Tensimeter-Still A . Boiler, 500- or 1000-ml. flask with thermometer side arm 45/50 ground-glass joint Thermometers 12-mm. connection to vacuum gage 12-mm. connection to trap and pump F. Alembic, formed from 40-mm. tubing 6. Connection to pump
B. C. D. E.
Several types of stills m r e tried before the design described in this paper was evolved. The improved still has been in use for more than 2 years and has proved reliable and simple to operate over the desired range. It has been used to obtain vapor pressure information on a wide variety of lactic acid derivatives (10). Various means have been used to overcome difficulties in obtaining accurate boiling point measurements a t low pressures. The “automanometer still” of Hickman and Weyerts (8) exposes the distilland to a relatively high pressure drop between still pot and condenser, for it is precisely by measuring the pressure drop across one or more orifices that this still gains its usefulness. It is operable doivn to a condenser pressure of perhaps 0.5 mm. of mercury. On the other hand, the tensimeter-hypsometer of Hickman, Hecker, and Embree ( 7 ) offers negligible pressure drop but operates on total reflux, so that distillation is impossible. This device, moreover, in the present authors’ hands failed to operate satisfactorily above 2 mm., because of excessive bumping. It is stated to be operable down to 0.03 mm. of mercury.
The tensimeter-still (Figure 1) consists of a 500- or 1000-ml. round-bottomed short-necked flask fitted with a thermometer side arm and a 45/50 ground-glass joint. The joint is attached to a wide upright condenser with an alembic take-off located so that the vapor path is as short as possible. The condenser bears a thermometer and connections for a vacuum gage and pump. T o maintain vapor-liquid equilibrium and avoid bumping, the still is suspended from the side tubes near the top of the condenser and rocked to and fro by a crank powered by a stirrer motor. The crank is attached a t the ground joint after the flask is connected and the joint is well wrapped with asbestos tape. The joint is lubricated with silicone-type grease to eliminate the sticking sometimes encountered with ordinary grease. I n operation as a tensimeter, the rocker is started, and heat is applied with a mantle heater, while vacuum is maintained. When boiling begins, the connection to the pump is throttled down, and the pressure is adjusted to the desired value by means of a slow leak. Heating is continued until the sample is distilling gently. The heat input must be adjusted so that the difference between liquid and vapor temperatures is a minimum, indicating minimum superheating and pressure drop between flask and condenser. At several millimeters this difference is usually 1’ C. or less, but a t low pressures may become several degrees. To assist in condensing the vapors, a strip of moist cloth is sometimes wrapped around the condenser. The stopcock makes it possible to operate on total reflux (usually necessary on small samples). The boiling point is taken as the vapor temperature a t the prevailing pressure. T o make certain that the system is in equilibrium, the boiling is continued until the vapor temperature remains constant. The pressure is then adjusted to another value, and the process repeated. It is possible to cover the range 0.03 to 10 mm. in a few hours. The data thus obtained are adequate for the preparation of the usual log P US. 1/T charts.
ANALYTICAL CHEMISTRY
1418 For the pressure measurements, a triple-range McLeod gage covering the ranges 0 to 0.2,O to 1.0, and 0 to 20 mm. of mercury is used. This gage is useful because of its high precision a t lower pressures, where the boiling point changes rapidly with pressurefor example, a pressure change of 0.1 mm. of mercury a t 4 mm. changes the boiling point of butyl phthalate about 1a C. At 0.04 mm. a pressure change of only 0.003 mm. changes the boiling point 1 C. Hence if the boiling points are to be correct to 1 C., the pressure gage must have high precision a t low pressures; this high precision is obtained with the triple-range gage. To keep condensable vapors out of the gage a trap cooled with solid carbon dioyide is used. IO
i 4
3 2 I
0.8 c5 0.6 I 0.5 0.4
w-
a
0.3 02
3
01
p g; 0.04 0.03 0.02
0.0I 0.00 8 0006 0.005 0.004 0.003 0.002
0.00I
150 170 190 210 250 110 130 TEMPERATURE,%. Vapor Pressure of Phthalates
50 60 70 8090100 120
Figure 2.
that a linear extrapolation of considerable usefulness may also be made to 0.001 mm. Several factors tend to introduce errors into low-pressure measurements of boiling points, especially when the distilland is not adequately agitated: Evolution of vapor begins before the distilland reaches the boiling point, resulting in too low vapor temperature (6) ; the distilland becomes superheated because of poor heat transfer, resulting in too high vapor temperature; and in the distillation of a multicomponent distilland the surface layer becomes leaner in the more volatile components, resulting in nonequilibrium distillation, superheated vapor, and poor fractionation. The first two factors are important in tensimeter measurements because they tend to cause erroneous vapor-temperature observations. Fortunately, they have opposite effects and their net effect is thus reduced. In the lowpressure distillation of a mixture, all three factors must be considered, though it is believed that the first is of minor importance. The effects of the second and third factors, however, are additive and may be large, thus causing serious errors in observed boiling points. Their effect is minimized or virtually eliminated, however, by vigorous agitation of the distilland throughout the distillation. This is conveniently accomplished in the tensimeter-still by the reciprocating motion of the still imparted by the motor-activated crank. It is believed that the satisfactory results obtained in the present work were possible with equipment of this general type only because efficient agitation of the distilland was achieved. The agreement of these tensimeter data with vapor-pressure information obtained by other means [pendulum tensimeter (7), direct weighing and molecular dew (9)J indicates that a recent suggestion that "low pressure tensimeters relying on residual gas to measure the vapor pressure must be highly inaccurate in this range (0.15 mm.)" (6) is hardly valid. This suggestion was made on the basis of an observation during a distillation under conditions much different from those in the tensimeter-still. As a still, the device is useful in the preparation of many highboiling compounds, provided a low fractionating efficiency is sufficient-that is, the boiling points of the fractions to be separated are fairly wide apart. The preparation of dicarboxylic acid esters of various lactates (IO) may be used t o illustrate the use of the tensimeter-still in distilling organic compounds. After the crude IO 8
The McLeod gage measures only the pressure of the permanent gases present. If the distilland has been well stripped of solvents and other light fractions, and if it is not undergoing decomposition during distillation, the McLeod measurements are reliable.
s 4
3 RESULTS AND DISCUSS103
The charts give boiling point curves for three common plasticizers (Figure 2) and for two potential plasticizers made from lactic acid (Figure 3), plotted as log Pus. 1/T, 'K. Each curve is linear, and although the experimental points go down only to 0.02 to 0.03 mm. of mercury, the curves have been extrapolated to 0.001 mm. The information in the literature (1, 9,6, 7 , 11) on the vapor pressures of these common plasticizers is compared with these data in Table I. The observed values were taken from the boiling point curves, and the literature values from similar published curves. The experimental points define a straight line that is in good agreement with published values; the observed boiling points of 2-ethylhexyl phthalate may be somewhat low, as this was a commercial plasticizer sample. As a measure of the agreement to be expected, it may be pointed out that six publications in the period 1930 to 1946 give the boiling point of butyl phthalate a t 3 mm.as166°t0170"C. (1,9-6,7,11). Extrapolation of the lines in Figure 2 gives 52" and 92" C. a t 0.001 mm. for butyl and 2-ethylhexyl phthalates, respectively. The literature (direct determination, 7, 9) gives 56" and 96" C. It is thus evident that the tensimeter-still can provide reliable vapor-pressure data over the range 0.03 to 10 mm. of mercury and
2 (j
2
1.0
08
22 :: 0.6
0.3
v,
w
(I:
0.2
a
0.I 0.08 0.06 0.05 0.04 0.03 0.02
0-n-DECYLLACT-
0.0I TE MPERATURE,"C.
Figure 3.
Vapor Pressures of Lactic Acid Derivatives
V O L U M E 21, NO. 11, N O V E M B E R 1 9 4 9
1419 LITERATURE CITED
Table 1.
Comparison of Boiling Points with Published Values
Plasticizer Butyl phthalate 2-Ethylhexyl phthalate n-Octyl phthalate
Boiling Points at F’arious Pressures 4 mm. of H g 0 4 mm. of H g 0 . 0 4 mm. of H g Obsd. Lit. Obsd. Lit. Obsd. Lit. 174 172 132 133 96 97 220 223 177 179 141 141 237 238 191 192 153 , ,
reaction mixture has been obtained, the lower boiling reagents are distilled from the product in a Vigreux still, and then the residue is distilled a t low pressure in the tensimeter-still. The product thus obtained may be redistilled to obtain vapor-pressure data; if the reaction is clean-cut, however, the distillation and vapor-pressure determinations may be effected simultaneously. The reliability of the boiling point measurements indicates that the tensimeter-still has a separating efficiency of at least one theoretical plate.
Burrows. G.. J . SOC.Chem. Ind.. 65. 360-5 11946). Distillation Products, Inc., Rochester 13, N. Y . , “High Tacuum Distillation,” Bugust 1947. Gardner, G. S., and Brexer, J . E., Ind. Eng. Chem., 29, 178-81 (1937). Hickman, K. C. D., J . Franklin Inst., 221, 363-402 (1936). Hickman, K. C. D., J . Phps. Chem., 34,627-36 (1930). Hickman. K. C. D., Record Chem. Progress (Kresge-Hooker Sci. Lib.), 9, 15-25 (1948). Hickman, K. C. D., Hecker, J . C., a n d Embree, N. D., ISD. ENG. CHEM.,A s . 4 ~ ED., . 9, 264-7 (1937). Hickman, K. C. D., and Weyerts, W.,J . Am. Chem. SOC.,52, 4714-28 (1930). Kapff, S. F., and Jacobs, R. B., Ret. Sci. Instruments, 18, 581-4 (1947).
Rehberg, C. E., Dixon, M.B., and Meiss, P. E., preprint booklet of papers presented before Division of Paint, Varnish, and Plastics Chemistry at 113th Meeting, Bu. CHEM.SOC., Chicago, Ill., pp. 105-8, llpril 21, 1948. Schicktane, S. T., J . Research Satl. Bur. Standards, 14, 685-92 (1935). RECEIVED Soveinher 24, 1948. Presented before the Division of Paint,
ACKKOW LEDGV ENT
The authors wish to thank Harry John of this laboratory for his suggestions and for the construction of the McLeod gage.
Varnish, and Plastics Chemistry a t t h e 113th Meeting of t h e AMERICAN CHEMICAL SOCIETY, Chicago, Ill., a n d a t t h e Meeting-in-Miniature, Philadelphia Section, AMERICAX CHEMICALSOCIETY, January 22, 1948. Paper 3396, Biireau of Agricultural and Industrial Chemistry.
Rapid Colorimetric Estimation of Phenol R. W. AIAIHTIN, Plastics Laboratory, General Electric Company, Pittsjield, .llass. describing the use of 4-aminoantipyrine as a reagent A PAPER for the quantitative estimation of phenolic fungicides ( 6 ) prompts the report of a similar procedure for the estimation of phenol. The procedure, which has been in use in this laboratory for nearly 3 years, was developed when a method for the very rapid determination of small quantities of phenol in salt brines was required. I n the range 0.2 to 2 p.p.m., phenol can be estimated either in water or in concentrated salt brines within f 10%. A number of procedures have been developed for the determination of phenol in both lo^ and high concentrations. I n low concentrations the methods of Gibbs ( 4 ) , Folin and Dennis (S), Houghton and Pelly ( 6 ) , and Stoughton (8) have found the most use. Recently Lykken, Treseder, and Zahn ( 7 ) have modified Stoughton’s procedure and applied it to the determination of phenols in petroleum and allied products. The main advantages of the present method are the ease and rapidity with which a determination can be made and the fact that it can be applied without modification to a concentrated salt brine containing only a few tenths of a part per million of phenol. The color reaction is based on the condensation of 4aminoantipyrine with phenols in the presence of alkaline oxidizing agents. The formation of the antipyrine dyes which result has been represented by Emerson ( Z ) , n ho discovered the reaction, as taking place according to the follon ing euample:
REAGENTS
4-Aminoantipyrine. The compound, which is a stable material, was prepared from antipyrine according to the procedure outlined by Eisenstaedt ( 1 ) . A 2% stock solution was prepared by dissolving the purified compound in distilled water. Potassium ferricyanide, 2y0 aqueous solution. Ammonium hydroxide, 2 3‘ solution. Phenol. Freshly distilled, water-white phenol was used to prepare the standard solutions containing 10 and 100 p.p.m. PROCEDURE
Transfer 50 ml. of the sample to a 50-ml. tall-form Nessler tube and add 0.3 ml. of the 2y0 4-aminoantipyrine solution and 1 ml. of 2 11; ammonium hydroxide. Mix the contents of the Sessler tube thoroughly, add 1 ml. of the 2% potassium ferricyanide solution, and mix again. Compare the reddish color produced by the sample in a Fischer nesslerimeter or similar suitable apparatus with standards prepared from phenol solutions of knoum concentration. Simple visual comparison of the colors can also be employed, with some sacrifice in accuracy. Where the phenol content of a salt brine is to be determined, sodium chloride may be added to the standard phenol solutions, so that the known and unknown solutions contain approximately the same amount of sodium chloride. However, in most cases no difference in the intensity of the colors produced in salt or water solutions is noted and this added precaution is not necessary. Both in the presence and in the absence of salt in the solutions, the colors develop almost immediately, so that the samples need not qtand more than 2 m i n u t e before the readings are taken. DISCLSSIOA O F RESLLTS
With the apparatus used, the most sensitive color range for a 50-ml. sample was found between 0.2 and 2 p.p.m. of phenol. Samples of greater concentration may be used by properly diluting them before the analysis. Table I shows typical results. Assuming an excess of 4-aminoantipyrine, the color reaction with phenol Tvill occur over a rather \vide range of concentrations for the base and oxidizing agent. However, maximum sensitivity is dependent upon the amount of each reagent used. The color reaction requires that the solution be slightly basic; and, although sodium hydroxide or sodium carbonate can be used,
