ANALYTICAL CHEMISTRY
1174
L’arious “fields” are delineated and marked specifically. Columns 1 to 4 contain the value of t (the temperature): columns 5 to 10 the values of t 2 ; columns 11 to 18 the values of ( t - 100jt3; columns 19 to 24 the value of Ro,and so on. For automatic calculation, Formula 3 has been replaced by the equivalent formula:
Rt = Ro
’
+ Kt + Lt* + J1 ( t - 100)
t3
(5)
wliere K = Ro.4, L = ROB,and .1f = RcC. In this equation, the four constants K O ,K , L , and .If define the resistance of a given thermometer. Values of 1, t 2 , and ( t - l O O ) t 3 from -200” to +500” C. weri: computed and punched into 700 cards in the first 18 columns. Each card represents one temperature value. Kest, the value of K O is “gang punched” (in each cnrd) by running the whole deck through th,e high speed reproducer. I n the same operation the information in columns 25 to 32 (number of the platinum resistance thermometer and the cards with negative temperatures) is alqo gang punched. After locking the value of li into a “multiplier,” and passing the deck through, the multiplier will punch product K t in rolumns 38 to 45, a t the rate of about 700 cards per hour. I n like manner, products Lt2 and ‘$1 ( t - 100)t3 are multiplied and punched in columns 46 to 62. lThe Droduct ,kl(t - 100)t3is reouired onlv on negative values of t. j The deck is now run again through the multiplier, which is now wired to add these products algebraically t o the IZO term and punch the Rt sum in columns 74 to 80. Figure 2 shows one of the
cards with the complete computation. The deck is now run through an “accounting machine,” which reads the information in the holes and makes a printcd record of all the information. Figure 3 shows a section of the final table giving the values of Rt for every degree, and all the computations in the event that checking is necessary. Thus in approsiniat,ely 4 Iiours’ machine time the entire computation of 700 temperatures is complet,e, and a printed record (triplicate if desired) of all the romputations is rendered. LITERATURE CITED
Callendar, H. L., Phil. Tmns. (London), 178, 160 (1887). Eckert, W. J., J . Chem. Edz~cntion,24, 54-7 (1947). Eckert, W. J., “Punched Card Methods in Scientific Computation,” Columbia University, New York. Thomas J. Watson Astronomical Computing Buieau, 1940. (4) Neyers, C. H., J . Rewarch .Tat[. Bur. Standards, 9, 807 (1932). (5) Schwab, F. R., and Smith, E. R., Ibtd., 34, 360 (1945). (6) Stimaon, H. F., Ibid., 42, 209 (1949). (7) Stull, D. R., Rev. Sci. Instruments, 16, 318 (1945). (8) Van Dusen, M. S...I. A m . Chem. Soc., 47, 326 (1925). (1) (2) 131
RECEIVED April 13, 1950. Presented before the Division of Chemical Eduoation a t the 112th Meeting of the A v E R I c h s CHEMICAL SOCIETY, New York, N. Y.
Sampling Certain Atmospheric Contaminants by a Small Scale Venturi Scrubber PAUL L. MAGILL, MYRA ROLSTON, J. A . MAcLEOD, AND R. D.-CADLE Stanford Research Institute, Stanford, Calif.
A portable device for sampling air contaminants permits the scrubbing of large volumes of air by a small volume of liquid. This facilitates chemical analysis of the collected materials by making it possible to obtain relatively high concentrations of contaminants in the scrubbing liquid. The action of this portable field unit is based on the principle used in the Venturi scrubbers employed to remove fumes from industrial stack gases. Its efficiency for the collection of several gaseous and particulate air pollutants is discussed. These pollutants include ammonia, sulfur dioxide, sulfuric acid, and sodium chloride.
R
ECENT investigations of air pollution in metropolitan areas
have emphasized the need for a sampling device which makes it possible t o scrub large volumes of air with a small amount of scrubbing liquid. A device is also needed Yhich will sample these large volumes in a relatively short time. Most previous sampling instruments designed for the collection of air pollutants in a scrubbing liquid, such as impingers or bubbler trains (1, 8 ) , have the disadvantage of comparatively slow sampling rates or unduly large liquid volumes. Filtration techniques have been developed which are useful, and many advances have been made in this direction (10). However, filtering rates are usually slow and the neLassity for removing the filtered material from the filtering agent is often objectionable. The apparatus here described is a portable laboratory scale model of the Venturi scrubbers used for fume recovery in plant stack gases, and is based in principle on the industrial models described by Anthony (92). Essentially, it affords a means of injecting a scrubbing liquid into a rapidly moving stream of air a t a Venturi throat; the liquid stream is there reduced t o a fine spray with droplet acceleration, and subsequently the spray-gas mi\-ture is decelerated and separated. The efficient collection of particulate impurities probably is due chiefly to collision with water droplets and diffusion into the water ( 8 ) . Johnstone and Roberts
( 4 )have advanced a “diffusion theory” in which it is shown that aerosol particles of 0.1-micron diameter or less have sufficient Brownian movement to be considered aa acting like large gas molecules; thus the collection of such particles might be considered t o be analogous to gas absorption in liquid droplets. This type of scrubber has proved especially useful to this laboratory in sampling polluted city atmospheres, inasmuch as chemical analysis of trace substances is facilibted by their relatively high concentration in the scrubber liquid, as compared to other methods of sampling. It is valuable also because of its ability t o sample large volumes of air quickly during periods of peak contamination, which may be of short duration. I n addition to this application, the small Venturi-scrubber should also be usefu1,in industrial plants and other locations where it is desired to know the concentration of toxic substances in working areas. The use of this scrubber has been mentioned in several reports from this laboratory (6, 7,11). APPARATUS
Functional details of the mall scale Venturi scrubber used in this laboratory are diagrammed in Figure 1.
V O L U M E 22, NO. 9, S E P T E M B E R 1 9 5 0
DIFFUSING
1175
t.hruugh il constant-ievd device, is pravided to a reservoir in order t o make up for evaporation losses. The total liquidin the scrubber is about 150 ml. An areatype flowmeter is located in the water circulation line. It was found that the maximum possible Row gives optimum result^. This is achieved hy positioning the water jet in the Venturi throat by a trial and error met,had until the maximum Row rate is attained. ThiH rate was equivalent t,o 8 to 9 gallons of wator per 1000 cubic feet of ail. scruhbcd (about 0.25gallon p$minute) in an instrument of the design and dimensions shown in Figure 1. Connected to t.hr air line. leading from the separator is B vacuum gage which may he calibrated to indicate the rat,e of air Row through the instrument. The- width of the diffusing section a t the point oi attachment t,o t,hc ~eprtnltoris shout 0.75 inch.
SECTION
I ; IAL
\'AIR
EXHAUST
BAFFLE TO BREAK, LIQUID VORTEX
I 'WATER
narable to those of '6iphly efficicnt industrial.Venturi Reruhhek (4,5). Figure 2 shows the scrubber con__^.._^__I,__ L ^_I :-.. .:..-,- .---L., ^.._
RECIRCULATION LINE
:&
0.25 G.?M. Figure 1.
Fiinetional Details of Laboratory Scale Venturi
I l l 8 t.LIICIeIIDy "I
bulb Ulllall
Y I . I I I U I I 51:I~u,>,,tII
n n a "'"'""g"'e"
for particulate and gaseous air contaminants. The particulate contsminsnts were sodium chloride &wosoI, which is reldveiy easy to collect, and sulfuric acid mist, which is natorioudy difficult to rollect by wet mruhbing. The'gases used were ammonia and sulfur dioxide. These contaminants were introduced st a known rate into a &foot mixing tube through which air was passing'at 1000 cuhic ioot per minute. The atmosphere leaving the tube was sampled by means of the Venturi scrubber, and the resulting scrubber solutions were subjected to appropriate chemical analyses. Sampling time in each ease was 15 minutes.
r"
The results show that t,he effirienq of the scrubher may he markedly different for,various rontaminxnts or for eontarninants in differentphysical conditions. Sodium Chloride. Sodium rhluridr anrasol was dispersing a 15% aqueou. sdlulion of tho salt. in1 t,ube. Evaporatibn of t h r droplets left behind CI tielm Figurr 3 s h o w H ~gold-shadowedelectton'i
Figure 2.
Portable Venturi Scrixbber
It is built, n i Lueitt,, which pennits olmrvatiorI of the spray sntl separation action. I.ueite was preferred to glass because oi ease of construction. Air is drawn through the devici at the rate of 32 cleaner. At the 0 to 3M) feet per 3arated from the acruhhinp liquid in t h i nycldnic s&xuator seeti;I". .e bottom of the turi throat. The ce8 the "driving ibbina liouid. A satisfactory water jet t,ube was found to be a lerigth Of grass tuhing, 6 mm. in inside diameter, with ~1st,ritight-cuitend; thifi end is bevolcd t.oward the center to produce a sharp edge. A connection, 1
8 Figure 3.
Sodium Chloride Particles fmm
Aemml
ANALYTICAL CHEMISTRY
1176 Scrubber Recovery of Sodium Chloride Aerosol
Table I.
NaCl Concentration in 4 i r , Mg. per Cu. Meter Calcd. from wt. of Calrd. from amount KaCl lost from genof NaCl in scrubber erator solution
Recovery, %
particles collected on a screen which had been in a sedimentation chamber containing the aerosol for 3 hours. Although the picture may overemphasize the percentage of large particles because of the short time allowed for sedimentation, it shows that all the particles were about 5 microns or less in diameter. The concentration of the salt in the air was determined by difference from ( a )the known weight of sodium chloride originally introduced into the generator, and ( b ) the amount of salt remaining in the generator after the dispersion period. This latter was determined by thoroughly washing the remaining solution from the generator, evaporating it to dryness, and weighing the residue. The scrubber solutions were analyzed for sodium content by the flame spectrophotometer. These recoveries, which averaged 96%, are presented in Table I. Sulfuric Acid. The sulfuric acid aerosol was formed by bubbling dried nitrogen through a small tube of fuming sulfuric acid. This was weighed immediately before and after the generation period to determine, through the weight of sulfur trioxide lost, the concentration of sulfuric acid produced in the test atmosphere. (Sulfur trioxide reacts rapidly with any moisture in the air to form a sulfuric hcid aerosol.) One set of experiments was performed to determine the scrubber’s efficiency in picking up the sulfur trioxide (sulfuric acid aerosol) immediately after it was generated from the bubbler. The sulfate content of the scrubber solution was measured turbidimetrically (9). Results of these experiments are shown in Table 11. The scrubber removed about 44% of this mist from the air. Table 11.
Scrubber Recovery of Sulfuric Acid Aerosol, as Sampled Directly from Generator €301
a
Generated, Mg.
SO8 Recovered byscrubber, 1Mg.
The efficiency of this device for sulfuric acid in aerosol form was lower than might be desired and was somewhat less than that reported by Jones ( 5 ) for an industrial scrubber. However, the concentrations of sulfuric acid in the gases a t the inlet of the scrubber described by Jones were a t least 1300 times the concentration used in the present work, and the concentrations of sulfuric acid in the exhaust gases from the scrubber he described were at lea& ten times greater than the concentrations in the atmospheres that were sampled by the laboratory scale scrubber. Moreover, the sulfuric acid aerosol Jones sampled may have had a larger mean particle size than those used in the present investigation. Because an extensive study of the design of the Venturi, solution jet, and flow rates was not undertaken, it is possible that the recovery of such aerosols could be improved through some modification of the apparatus.
Table 111. Scrubber Recovery of Sulfuric Acid Aerosol from Mixing Tube 50s in Air, Mg. per Cu. Meter Detd. froA loss in wt. of bubbler 0.278 0.268 0.156 0.096 0.148 0,073 0,053 1.75 1.95 1.90
Detd. by crucible train sampling
..
1:82
1.98 1.90
Detd. by Venturi-scrubber 0.149 0.178 0.104 0.070 0 097 0.05.5 0.031 0.118 0.110 0.129
Recovery, % 53 7 66.5 66.7 72.9 65.8
75.0 58.6 67.4 56.7 67.8
A few runs were made to determine the effect on the efficiency for the collection of sulfuric acid of substituting a 0.5% aqueous solution of sodium hydroxide for water as the scrubbing agent. There was apparently no advantage in using a basic scrubbing agent. These results indicate that the apparatus can be used to determine in a semiquantitative manner very low concentrations of sulfuric acid in air. For precise results, it would probably be necessary to apply a predetermined correction factor which would depend on the efficiency of the scrubber for the acid in the partirular atmosphere being investigated.
Recovery, % 36 39 56 41 50
The sulfuric acid particles prepared in the manner described above are probably much smaller, and therefore more difficult to collect, than those found in normal atmospheres, because little time was allowed for them to grow by accumulating water. Another set of experiments was performed in which the sulfur trioxide was generated into the previously described &foot mixing tube, where several seconds elapsed between the time the mist was generated and the time of its collection by the scrubber. Again the scrubber solutions were analyzed turbidimetrically. The results of the tests carried out in this manner are presented in Table 111. The scrubber averaged 65% efficiency in the ten trials. A further check of the concentration of sulfuric acid in the aerosols was carried out by drawing the test atmosphere through a train of four Gooch crucibles, each three fourths filled with tightly packed asbestos ( 1 2 ) . After sampling, the asbestos was washed until free of sulfuric acid, and an aliquot of the filtered wash water was taken for turbidimetric analysis. This has been shown to be a very accurate method for determining the concentration of sulfuric acid in aerosol form. Sulfuric acid concentrations &s determined in this manner also appear in Table 111.
m9 N H 3 / m 3 A I R
Figure 4.
Recovery of Ammonia by Small Scale Venturi Scrubber
Ammonia. The long mixing tube was also used to determine the efficiency of the Venturi scrubber for removing ammonia from air by metering this gas from a cylinder into the intake end. Both distilled water and 0.1 N sulfuric acid were used m the solution in the scrubber during 15-minute runs. The ammonia content of the collecting solution was determined colorimetrically using Nessler’s reagent ( 3 ) . The efficiency using distilled water as the scrubbing solution increased with decreasing air concentrations of ammonia, approaching a maximum recovery of about
V O L U M E 22, NO. 9, S E P T E M B E R 1 9 5 0
1177 minimized. T o determine the possible extent of such interference, sulfur dioxide was metered into the mixing tube, and the scrubber sampled the air-sulfur dioxide mixture for 15-minute periods using distilled water as the solvent. The sulfur dioxide content of the resulting solution was determined by oxidizing the sulfite to sulfate, which was then determined turbidimetrically. The results are presented in Figure 5. The over-all efficiency varied from 1 to 9% as the sulfur dioxide concentradon was decreased from about 7 to 0.6 mg. per cubic meter. Probably high efficiencies for sulfur dioxide, if desired, could be obtained by using sodium hydroxide solutions in the scrubber. ACKNOWLEDGMENT
0
I
2070 at a concentration of 0.2 mg. per cubic meter of air. However, the recoveries averaged about 100% when 0.1 %, ’ sulfuric mid was used. The results are plotted in Figure 4. The Venturi scrubber would be expected to be inefficient for volatile substances, such as ammonia, unless some substance is present in the scrubber solution to convert these materials to a nonvolatile form. For a very small increment of time at the beginning of the sampling period the scrubber is efficient. During this time the scrubber solution consists essentially of pure water. -4s the scrubber operates, however, the solution rapidly absorbs the gas, until in a short time equilibrium conditions exist and the vapor pressure of the volatile substance in solution equals the partial pressure of this substance in the air. Obviously, gas collection by the scrubber from that time on would be zero. Thus the over-all efficiency of the scrubber will decrease with increasing lengths of runs. The decreasing efficiency with increasing con(bentration must result from deviations from Henry’s law. Sulfur Dioxide. Finally, because sulfur dioxide is often found along with sulfur trioxide in the atmosphere, the efficiency of the w u b b e r for collecting this gas was studied. It was expected that when pure water was used in the scrubber the efficiency for sulfur dioxide collection would be low, as was the cp,se in ammonia sampling. This would be an advantage when using the scrubber to collect sulfuric acid, inasmuch as interference from the dioxide, which in dilute solution oxidizes rapidly to sulfuric acid, would be
Acknowledgment is due the Smoke and Fumes Committee of the Western Oil and Gas Association, whose financial support and encouragement for the study of smog in Los Angeles made this work possible. LITERATURE CITED
(1) Am. Pub. Health .4ssoc., “Ventilation and atmospheric poliu-. tion. 11. Report of Subcommittee on Chemical Methods in
.4ir Analysis. Sampling and Sampling Devices,” Am. Pub. Health Assoc. Year Book, pp. 92-7, 1939-40. (2) Anthony, A. W., Jr., “Two Methods of Wet Scrubbing of Gasea for Reduction of Atmospheric Pollution,” Proc. Smoke Prevention Assoc. .4m., 26 pp., 1948. (3) Jacobs, M. B., “Analytical Chemistry of Industrial Poisons, Hazards, and Solvents, pp. 289-90, New York, Interscience Pub., Inc., 1941. (4) Johnstone, H. F., and Roberts, H. M., Ind. Eng. C h m . , 41. 241723 (1949). (5) Jones, W. P., Ibid., 41, 2424-7 (1949). (6) Magill, P. L., Am. Znd. Hug. Assoc. Q w r t . , 11, 55-64 (March 1950). (7) Magill, P. L., Proceedings of First National Air Pollution Sym(8) (9) (10)
(11) (12)
posium, Stanford Research Institute, Pasadena, Calif., Nov. 10 and 11,1949, p. 61. Moskowite, Samuel, Siegel, Jac, and Burke, 11’. J., N. Y . State Dept. Labor, I d . Bull., 19, 33-5 (1940). Sheen, R. T., Kahler, H. L., and Ross, E. M., IND.ENG.CHEW, ANAL.ED.,7, 262-5 (1935). Silverman, Leslie, Proceedings of First National Air’ Pollution Symposium, Stanford Research Institute, Pasadena, Calif., Nov. 10, and 11, 1949, p. 55. “Stanford Research Institute, “Smog Problem in Los -4ngeles County,” Second Interim Report, 1949. Western Precipitation Co., Los .4ngelc.s. Tech. Bull. 3-C: pp. 3-4.
R E C E I V E DMarch 21; 1950.
Spectrophotometric Determination of Phosphorus in Organic Phosphates W. R. SIMMONS’ A N D J. H. ROBERTSON, University of Tennessee, Knoxuille, Tenn. The molybdivanadophosphoric acid procedure for the determination of orthophosphate is adapted to the analysis of aliphatic phosphates after hydriodic acid conversion and to both aliphatic and aromatic phosphates after conversion by catalytic oxidation.
I
N A previous paper (6) the authors reported that hydriodic acid conversion of aliphatic phosphates yields colorless solutions, whereas similar treatment of aromatic phosphates gives solutions of varying color from pale green to pale orange; and that conversion of both aliphatic and aromatic phosphates by sulfuric-nitric-perchloric acid oxidation in the presence of a molybdenum catalyst produces slightly greenish colored solu1 Present address, Experiment Station, Hercules Powder Company, Wilmington,’Del.
tions. .4lthough the solutions obtained by hydriodic acid conversion of aliphatics appear to be colorless, when viewed through a long column they show a faint yellowish cast. Becuse of the importance of the molybdivanadophosphoric acid colorimetric method of determining orthophosphate, reported by M i e son (6) and modified by Kitson and Mellon ( 4 ) , it appeared desirable to establish the degree of applicability of the method to organic phosphates decomposed by hydriodic acid or by catalytic oxidation. The present paper describes this study.
