V , O L U M E 2 2 , NO. 5, M A Y 1 9 5 0 (2) rlm. (3) (4) (5)
(6)
SOC.Testing Materials, A.S.T.M. Designation D 8 7 1 4 6 T
for Testing Cellulose Acetate. Bacon, L. R., J . Franklin Inst., 221, 251-73 (1936). Cragg. L. H., J . Collozd Sci.. 1 , 261-9 (1946). Faxen, H., dissertation, Uppsala, 1921; Ann. Physik, 68, 89 (1922); Arkiu M a t . Astron. F g s i k , 17, KO.27 (1922). Malm. C . J., U. S.Patent 2,126,489 (Aug. 9, 1938).
661 (7) Malm, C. J.. Salo, M., and Vivian, H. F., Ind. Eng. Chem., 39, 168-9 (1947).
(8) Malm, C. J., and Smith, H. L., Jr., Ibid., 38, 941 (1946). (9) Stokes, G . , Trans. Cambridge PhiE. SOC.,8 (1845); 9 (1851);
Collected Papers, Vol.
1.
RECEIVED October 4, 1949. Presented before the Division of Cellulose CHEUICAL SOCIETY, Chemistry at the 116th Meeting of the AMERICAN Atlantic City, N. J.
Determination of Small Amounts of Triethylene Glycol in Air SAVL KAYE WITH THE TECHNICAL ASSISTANCE OF ANNE C. ADAMS Uniaersity of Chicago, Chicago 37, 111. A method is described for the quantitative determination of triethylene glycol vapor in air. By sampling 30 liters of air, concentrations of the order of 2 micrograms per liter (0.002 p.p.m.) may be determined simply and rapidly. This concentrktion is in the range usually required in processes of air sterilization. The air is passed through concentrated sulfuric acid, which absorbs all the triethylene gl>col present; the acid solution is heated on a steani bath for 30 minutes and cooled to 30" C., and then a freshly prepared solution of 1-naphthol in sulfuric acid is added. The intensity of the
T
HE vapors of triethylene glycol and of propylene glycol have been found to possess marked bactericidal activity when dispersed in small quantities in the air (9, 6, 7). The formar compound, having much the lower vapor pressure (0.0013 mm. of mercury a t 25" C., 4 ) , is the more efficient, and is currently receiving widespread application as an aerial disinfectant for occupied premises. The bactericidal effectiveness of triethylene glycol vapor depends upon the extent to which the atmosphere is saturated with this vapor. The saturation concentration of triethylene glycol vapor in the air in the absence of water vapor has been reported by Wise and Puck (11) to be of the order of 11 micrograms of glycol per liter of air at 25' C., and the variation of this saturation concentration with relative humidity and temperature of the air has been determined (IO). Losses of unknown magnitude, caused by condensation and by ventilation of the room, make it desirable to have an analytical method of determining the concentration of triethylene glycol in the air. This method must be sufficiently sensitive to determine quantities of glycol in the saturation range, between 1 and 10 micrograms per liter of air, a t normal temperatures and relative humidities, Wise, Puck, and Stral (f8)bubbled the air through water and determined colorimetrically the extent to which an aliquot of the aqueous glycol solution reduced a standard dichromate-sulfuric acid mixture. Because of the small quantities of triethylene glycol which must be analyzed, large volumes of air (ea. 300 liters) and long sampling periods must be used. Another method of determining triethylene glycol (3, 6) depends upon the condensation of the vapor upon a cooled surface, and subsequent measurement of the amount of light scattering caused by this condensed film. Instruments embodying this principle (glycometers) have been built and successfully used in these laboratories for several years. They are, however, complicated and eupensive, and are not yet commercially available. This paper describes a simple, rapid, and reasonably precise
yellow color formed is measured spectrophotometrically, and is proportional to the quantity of triethylene glycol present. The presence of water affects the development of color; different calibration curves must therefore be used for different relative humidities. A formula relates the optical density of the treated sample to both the water and the glycol content, so triethylene glycol concentrations may be calculated directly from spectrophotometer readings. When conditions of analytical treatment are carefully controlled, the method gives deviations of about *5% of the glycol concentrations. method for the analysis of triethylene glycol in air. Concentrated sblfuric acid is used as the collecting medium for the glycol, and the analysis is performed spectrophotometrically by determining the intensity of the color produced upon the addition of 1-naphthol t o the heated acid solution of glycol. APPARATUS AND MATERIALS
i\ c
c
c r
Figure Absorption Tube
Sampler. The sampler used is a glass a b s o r p tion tube (Figure 1). Air is drawn through the tube by means of a vacuum pump, which is protected from corrosion by a soda-lime tube placed in the line. The flow of air is controlled by a suitable critical flow orifice installed between the sampler and the pump. It is desirable to have as little absorbing fluid as possible left in the tube after draining, and to prevent the liquid from splashing over into the vacuum system during operation. The Vigreux-type indentations have been found helpful in the second regard. When 10 ml. of sulfuric acid are used in the tube illustrated, whose over-all height is 26 em., the acid level is about 3 cm. above the outlet. When air is being drawn through the acid a t a rate of 15 liters per minute, the acid level rises to about 6 cm. above the base. The arrangement of the ground-glass joint (24/40) at the cap is convenient for refilling the bubbler. The sulfuric acid acts as stopcock lubricant; both the regular and the siliconebase types of stopcock grease were found to interfere with the test. The rubber retaining washer must be removed from the stopcock. Other Equipment. A spectrophotometer (Coleman Universal, Model 14, was used in this laboratory), a steam bath, and a water bath capable of maintaining a temperature of 30" C. are required. A vacuum pump and appropriate flowmeters should be available for taking the samples. Reagents. Triethylene glycol (air sterilization grade, Carbide and Carbon Chemicals
662
A N A L Y T I C A L CHEMISTRY
Corporation) distilled between 104" and 105" C. a t 0.4 mm. of mercury, was employed as the standard. Concentrated sulfuric acid, reagent grade, and 1-naphthol, melting point 95-96" (Eastman Kodak), are the other reagents.
limits of accuracy of the instrument (approximately 3% of the measured glycol concentration for transmittances between 20 and 7Oy0 of the control, or optical densities from 0.1 to 0.7).
SAMPLING PROCEDURE
EFFECT O F WATER
Ten-milliliter portions of concentrated sulfuric acid are dispensed, by means of an acid buret, into heavy-walled borosilicate glass ignition tubes, 25 X 200 mm. These tubes are placed in a rack and covered to protect them from dust. When a sample is required, the contents of one of the tubes are poured into the absorbing tube and the ignition tube is clamped below to receive the liquid after air has been drawn through. With the bubbler illustrated, 9.8 of the original 10.0 ml. drain from the apparatus in about 15 seconds. Thus, if successive samples are to be taken from atmospheres of changing glycol content, it is well to rinse the absorption tube with sulfuric acid between samples. For tests on more static atmospheres, however, the carry-over of glycol from one sample to the next introduces negligible error if only short intervals intervene.
Samples of triethylene glycol and acid to which more water ha5 been added develop IRSS color with the naphthol reagent than they would were' no water added. If this is not taken into consideration high humidities can introduce serious error into the determination of the triethylene glycol content of the air. In Figure 2, A represents the transmittance ua. glycol relationship in the absence of water, while C indicates the relation in the presence of 0.50 ml. of water in each sample. From a series of tests with different amounts of added water, one obtains a family of straight lines of differing slope. By relating the slopes of these lines to the water content, we obtain, for our test conditions, the relation
The method is most accurate when the sample contains between 0.05 and 0.25 mg. of triethylene glycol. A maximum of 18 liters of air per minute may be drawn through 10 ml. of acid in the sampler illustrated without observable loss by splashing, while air flow rates between 18 and 12 liters per minute have not produced observable differences in collection efficiency. In order t o simplify the procedure, the authors have installed standard orifices in vacuum lines, so that a constant rate of 15.0 liters of air per minute is maintained. All samples are taken for 2 minutes when normal "glycolized" atmospheres are being analyzed; the 30 liters of air thus sampled provide the amount of glycol required t o perform the spectrophotometric determination in the most accurate range. The time of sampling may be varied for particularly low or high glycol concentrations. ANALYTICAL PROCEDURE
After a series of samples has been taken, the ignition tubes are placed in a boiling water bath for 30 minutes, then removed and placed on a bath a t 30" C. (A rack that can be transferred from one bath to the other is a convenience here.) The contents of the tubes reach bath temperature in about 5 minutes, and this interval is used for the preparation of the 1-naphthol solution, which must be freshly made immediately before each test. The reagent is dissolved in concentrated sulfuric acid; 0.100 gram is added for each 21.0 ml. of sulfuric acid used. When the ignition tubes have cooled t o 30" C., 1 ml. of the 1-naphthol solution is added to each, and the mixtures are shaken. The tubes and contents are allowed to remain at 30" C. for an additional 15 minutes to allow full color develupment, and are then transferred to individually calibrated cuvettes. A tube containing only 10 ml. of unaerated sulfuric acid is boiled and cooled along with the samples, is treated with naphthol solution, and then serves as the blank for the entire series. The intensity of yellow color produced in the test samples is com ared with the blank by measuring the relative transmittance ofP light of wave length 410 mp, where maximum absorption occurs, as was determined by a wave length-transmittance experiment. CALIBRATION CURVE
A calibration curve must be prepared to relate the measured per cent transmittance t o the glycol content. A stock solution is repared by slowly adding cold sulfuric acid to LOO0 gram oPtriethylene glycol until the volume is nearly 100 ml. The h a 1 dilution to 100 ml. is made a t room temperature. Appropriate volumes of this concentrated solution are diluted with sulfuric acid, so that a series of solutions containing from 0.01 to 0.40 mg. of triethylene glycol per milliliter is obtained. Such standard solutions have been kept in a refrigerator for months without a significant change in response. One milliliter of each of these solutions is added to 9 ml. of concentrated sulfuric acid contained in an ignition tube, and these standards are then treated as described above. Line A on Figure 2 indicates the relation of optical density, D = (2 - loglo yotransmittance) t o the amount of triethylene glycol in each sample. The relationship is a straight line up to a t least 0.25 mg. of glycol in the sample, and is reproducible within the
Milligrams of TEG =
D
2.6
- 3(grams of
H20)
where the weights of triethylene glycol and of water are those added to the sample, and D is the optical density. Knowing the volume, the temperature, and the relative humidity of the air being sampled, we may convert optical densities to glycol concentrations by this expression. Calculation is considerably simplified when the air being sampled is of constant temperature, and when a constant volume is taken. Thus, for 30-liter samples, at about 23" C., substitution in the above formula gives 17800 .Micrograms of TEG/liter = 145 - RH where RH is the relative humidity. A psychrometric determination is required with each sample. .4lthough this dependance upon the humidity of the air introduces an additional variable, it does not in practice operate as a serious drawback t o the use of the method, because the relative humidity of the air is of primary importance in determining the effectiveness of glycol as an aerial disinfectant, and must be regularly determined in the course of any program utilizing glycol. The fact that the analytical procedure for glycol requires a knowledge of the humidity, then, introduces no new complications. COLLECTION EFFICIENCY
Wise, Puck, and Stral ( 1 2 ) have reported that the first of two absorption tubes containing water succeeded in removing only about 70% of the glycol from the air; when two such tubes were used in series, the total recovery was approximately 90%. With 10 ml. of sulfuric acid as the collection medium, a single absorber appears to be completely efficient, as indicated by the close agreement of concentrations determined analytically in saturated atmospheres with saturation concentrations determined from vapor pressure measurements. Furthermore, when two acid absorbers were used in series, there was a t no time any measurable amount of glycol recovered in the second sampler. ACCURACY AND PRECISION OF METHOD
The accuracy of the results obtained from air containing triethylene glycol is not readily determined, because the normal methods of volatilizing this material into the room, by the use of a heated surface, lead to losses of not inconsiderable size. Thus, in this laboratory (9) it has been estimated that only about 50% of the quantity of material lost from a vaporizing device appears in the air of even a sealed room. British workers ( 1 ) have confirmed and extended these findings. In order t o obtain atmospheres containing triethylene glycol vapor in concentrations which could be estimated by reasonably certain independent methods, a
V O L U M E 22, NO. 5, M A Y 1 9 5 0 0.9
0.8 0.7
0.6
I
I -
c
c
663
TRIETHYLENE GLYCOL IN SULFURICACID AT^ MC
@ No Added Water @ 0.2 MI.WaterAdded @ 0.5 MI.Wattr Added
1
/
tobacco smoke was sampled. Gross amounts of dust or traces of rubber or soap in the ignition tubes imparted a deep red color to the samples when naphthol was added, whereas traces of copper and iron caused a green coloration. The effect of various organic materials has thus far been investigated only in a preliminary way. The test is not specific for triethylene glycol, for propylene glycol and diethylene glycol give qualitatively similar results. An important source of error will be introduced into any of the methods used for determination of triethylene glycol if the temperature of the surface from which it is heat-volatilized is high enough to cause thermal decomposition. Tentatively, this temperature may be taken as 140’ C.
Table 11. Triethylene Glycol Concentrations Determined by Pairs of Samplers Pair No.
Analytical Concentration, ?/Liter Sampler 1 Sampler 2 Mean 1.98 1.75 1.87 2.06 2.08 2.09 2.22 2.33 2.44 2.33 2.44 2.21 2.40 2.45 2.43 2.91 2.92 2.89 3.05 3.06 3.06
DISCUSSION
I
0.1
0.05
I
0.15 0.20 Milligrams TEG in Sample
Figure 2.
Table I.
I
I 0.25
I
1
0.3
0.35
Calibration Curve
Triethylene Glycol concentration of Saturated Air .-
Temperature, C. Relative humidity, yo .lnalytical concn., microgramsfliter ?Jean analytical concn. Saturation concn. (Wise and Puck, I f )
Teat I 23.9 58 2.91, 2.91, 2.70, 2.61,3.06, 2.96, 2.67 2.83 0.15
Teat I1 20.0 54 1.95,1.72,1.72, 2.11. 1.85,1.95, 1.95,1.95 1.90 * 0.10
2.75
1.95
f
tightly sealed experimental room (8)wa.s used. Several large bed sheets were soaked with liquid glycol, and hung in the chamber for 2 days with a large fan playing directly upon them. I t is reasonably certain that a t the end of this time the air was saturated with triethylene glycol vapor. The results of two such tests are presented in Table I. The average deviation from the mean of the analytical values is jeen to be *5.3y0 in both tests, and the saturation concentpations, determined in a different fashion, fall within this range. In another test, two 30-liter air samples were taken simultaneously from an experimental chamber into which triethylene glycol was being heat-volatilized continuously. This method of evaporation usually produces mist as well as vapor, and thus the absolute concentration a t any time is not known. The results obtained from several such pairs of samples analyzed by the above method are presented in Table 11. The average of the deviations of each pair from its mean is 2.5%. INTERFERING SUBSTANCES
In a previous method for the analysis of triethylene glycol (le) ordinary room air often gave readings which formed a significant proportion of those obtained from air containing glycol. A series of tests in the authors’ laboratory and in several other “normal” atmospheres indicated that ordinary room air gives a negative or negligible response when tested by this method. A red-orange color developed, however, when air containing a large amount of
The 1-naphthol solution must be prepared immediately before use; the authors have not yet found a solvent in which the naphthol is stable, which does not interfere in some way with the test. If the naphthol solution is permitted to stand for more than 15 minutes, color formation is greatly diminished. This diminution in color also occurs if the 1-naphthol solution is added to the hot acid solutions of glycol. At least 30 minutes in the steam bath are necessary; increasing this time to 1 hour appears to produce no greater color, while heating for more than 1 hour appears to decrease color formation. The color develops to a maximum within 15 minutes after addition of the naphthol reagent, and fades but slowly thereafter. The authors are a t present attempting to determine the mechanism of this reaction, whose details have been determined in an empirical fashion. The method has, however, provided a useful tool for performing routine analyses of the air in experimental rooms. ACKNOWLEDGMENT
The authors are indebted to E. W. Dunklin for supplying redistilled triethylene glycol. LITERATURE CITED
Bourdillon, R. B., Lidwell, 0. M., and Lovelock, J. E., “Studies in Air Hygiene,” p. 122,London, Medical Research Council, 1948.
Puck, T. T., Reu. Sci. Instruments, 19,16 (1948). Puck, T. T.,Robertson, 0. H., Hamburger, M., and Hurst, V., J . Infectious Diseases, 76,216 (1945). Puck, T. T., and Wise, H., J. Phys. Chem., 50,329 (1946). Puck, T. T., Wise, H., and Robertson, 0. H., J . Ezptl. Med., 80, 377 (1944).
Robertson, 0. H., Bigg, E., Puck, T. T., and Miller, B. F., Ibid., 75, 593 (1942). Robertson, 0. H., Puck, T. T., Lemon, H. M., and Loosli, C. G.,Science, 97, 142 (1943). Robertson, 0. H., P u c k , T. T., and Wise, H., J. Ezptl. Med., 84,559 (1946).
Wise, H., Ind. Eng. Chem., 41,633 (1949). Wise, H., Ph.D. thesis, University of Chicago, 1947. Wise, H., and Puck, T. T., Science, 105,556 (1947). Wise, H., Puck, T. T., and Stral, H. M., J . Biol. Chem., 150. 61 (1943).
RECEIVED September 30, 1949. Preaented before the Division of Analytical and Micro Cbemiatry a t the 116th Meeting of the AMERICANCHEMICAL SOCIETY,Atlantic City, N. J. Work done under a grant from the Biological Department, Chemical Corps, Camp Detrick, Frederick, Md.