Device and Technique for Rapid Determination of Effluent Fluorides GORDON C. HARROLD AND RALPH V. HURLBURT Industrial Health, Hygiene, and Safety Seruiee, Detroit, Mieh. 4 rapid field test for the determination of fluoride ion is described. Semiquantitative results have heen obtained in tests w i t h fluorine, hydrofluoric acid, and soluhle inorganicfluorides. Preliminary worlt shows that a n y fluorine compound that can be broken down on paper wetted w i t h various reagents and placed ahead of the test paper will give semiquantitative determinations of fluoride ion.
N
UMEROIJS tests have been run on various types of methods
far the pest 5 years in an effort to find srapid field method for the determination of fluorides as an atmospheric contaminant, particularly as fluorine might be connected with welding. The standard analytical method of Willard and Winter (6) is used for laboratory analytical procedures, but is unsatisfactory as a field test, in part because of the time involved in the distillation to remove impurities. A modification OF the alizarin sodium sulfonate and zirconium nitrate method was evolved which eliminates acetate interference, but this also was timoeonsuming. The range of greatest significance from the health standpoint involves concentrations From 1 to 10 p,p.m. and requires a more rapid analytical determination in the area where operations are being performed than can he obtained by m y method that involves carrying solutions to the laboratory for analysis after spending from 15 minutes to 2 hours in collecting the air sample. Some inorganic compounds of fluorine are gaseous and require an estimation of the concentration within a few minut,es, so that safety and health requirements may he met. Because methods involving the collection OF a sample and the subsequent transport of the sample to a lahoboratory do not meet this time requirement, methods for collecting the sample in a liquid medium in the presence of reagents that would immediately indicate the relative amounts of fluoride were tried. While working on this approach, i t was decided that the fast& det,ermination would be through the use of an ihdieating paper. Semiquantitative to quantitative results may often he obtained by pumping known amounts of contaminated air through t,he chemically treated paper (3). Experiments were conducted with papers treated with sodium alizarin sulfonate and airconiim chloride, hut these were not a6 satisfactory even under ultraviolet light as papers treated with zirconium salts and wxmonie acids. The paper finally selected was treated with p-dimethylaminoasoben.enearsonic acid and eireonium oxychloride (4). In dilute acid solution, an insoluble salt of zirconium aaohenzonoarsenate, brown in color, is deposited in the pores of the paper. The excess of pdimethylaminoazobenzenearsonio acid is washed out with dilute acid. When fluorides are drawn through the prepared paper, a colorless (ZrFa) ion and the freo rtaobenaenearsanic acid, which is red, arc formed. Thus, there is a distinct color change from the brown zirconium compound to the red of the free azo acid, and this hzs been made the basis for a semiquantitative field method for the determination of the fluorine ion. In very small quantities, such are dealt with in determining concentrations injurious to health, fluorine itself is hydrolyzed to hydrofluoric acid and this is assisted by moistening the paper with hydrochloric acid. Soluble fluorides, such as sodium fluoride, react on the hydrochloric acid-moistened paper to give an immediate fluoride ion reaction. The insoluble inorganic fluorides,
such as calcium fluoride, are retained by the moist filter paper and react to give the typical fluoride reaction, but not until a longer period of time has elapsed, The method is roughly quantitative in that values in the range from 6 to 10 p.p.m. can he determined within *I p.p.m. and from 10 to 40 p.p.m. with decreasing sensitivity as the concentration of fluoride ion increases. In concentrations below 6 p,p.m., the method becomes increasingly less sensitive to each stroke of the pump, and a1 2 p,p,m. may he estimated within -0.2 p.p.m. The accuracy at the given concentrations is largely determined by the number of strokes of the calibrated pump, These concentrations were chosen because they were the most important to health and safety. PREPARATION OF CHEMICALLY TREATED PAPER
The treated Dmer is made as directed hv Feiel and Rai-
exposed paper to turn br INTERFERENCES
Theoreticdly, those anions interfere which give insoluhle or stable soluble complex compounds with zirconium. Orsanic hydroxy acids, arsenates, phosphatos, sulfates, and thiosuliates
Figure 1. Preliminary P u m p and Paper Holder are included in this group. However, under the conditions of the test in the ranges set up, sulfates and arsrnates do not interfere in amounts below 50 p.p.m. Using the enlibration procedure described below, and testing for fluoride ion in cancentrations from 2 to 10 p.p.m., no deviation was noted where sulfate ion of 10 times the concentration of the fluoride ion was introduced into the test atmosphere. If sulfates are present in very high concentrations and no fluoride ion is present, there is a slight 1504
V O L U M E 21, NO. 12, D E C E M B E R 1 9 4 9
1505
this method of calibration would not give accurate results. Part of the fluoride was rauidlv removed on the wall surfaces and Some formed a complex ion. . The following system (Figures 2 and 3) was then set up for the calibration of the treated paper: To establish an atmosphere that would contain a definite value in parts per million of hydrofluoric acid, a dilute hydrofluoric acid solution was atomized into a 46.2-liter bottle at a constant rate. The air-acid mixture was drawn from this bottle into a 630-liter chamber with an eleetricslly driven air pump celibrated et 0.85 cubic foot Der minute. Two combination impinger and fritted-glass mist oollectors m r e placed between the pump and large ohamher with B set of T-tubes and pinchcocks. After pumping a t a constant rate on one bubbler until the rate of flow and the parts per million of hydrofluoric acid became constant. the flow was switched to thesecond bubbler and timed with
"""Yllil.
The titration was made with standard sodium hydroxide and
showsacheekwithin0.5np.m Figure 2.
Calibrating Paper w i t h Known Q u a n t i t i e s of Hydrofluoric Acid
reaction, but the resultant color is a light salmon pink instead of FL red pink. Under usual conditions, sulfates do not constitute a serious interference, hut phosphates will interfere and will give approximately the same color effect as the fluoride ion when in higher concentrations than those fluoride concentrations of from I t o 40 p.p.m. with which we me largely concerned.
a t a constant rate, sufficient sample was drawn to give the firs; perceptible pink on the standard paper. The number of strokes required to do this is established before the titration is made; thus, there isa. definite relation herween this number and the parts per million of hydrofluoric acid determined lat.er.
PUMP
A pump of 50-ml.capacity was chosen, as this volume of air
from a. health and safetv viewnomt A head ro hold the test papers was built as sh&n in Figure 1. (Complete improved test set and calibrated test papers are manufactured by the Production Equipment Company, 6432 Cas8 Ave., Detroit, Mich.) A Plastacele insert holder 1.25 inohes (3.1 ern.) in diameter with 3r 0.25-inch strip mross the center provided an opening that permit,tcd the air to flow through the paper at a rate that permitted standardization of t.he reaction of the fluoride ion with the test paper. CALIBRATION
Some of the early .attempts a t calibration were unsuccessful hut are briefly mentioned here because they are a logical and rapid means of obtaining au answer and, as such, have heen used by a number of investigators in several universities in verifying the accur&~y of the fluoride field test. A 46.21iter battle was coated with a fine film of paraffin and volumetrically measured amounts of 48% hydrofluoric acid were introduced in a paraffin block plaoed on the bottom of the bottle. Air was circulated over the drop of hydrofluoric acid until i t was diffused throughout the jar. A number of preliminary tests indicated that about 50 strokes of the pump were needed for 1.1 p.p.m., 17 strokes for 3.5 p.p.m., and 10 strokes for 5.6 p.p.m. No quant,itative significance was attached t o these results because of the inequalities of distribution and doubts as to the concentration of the concentrated hydrofluoric acid in the authors' possession. In order t o determine the fluorine mare accurately, potassium fluoride dihydrate was weighed on a microbalance in amwnts .calculated t o give values in the range from 0.5 to 10 p.p.m. !n the 46.2-liter bottle employed. This fluoride salt was placed in the hollow of a paraffin block and a drop of concentrated sulfuric acid was dropped into the depression in the block after the bottle was closed. The evolved hydrogen fluoride was recirculated throughout the bottle by means of a rubber bulb. Two separate determinations a t 3 p.p.m. resulted in a very light pink coloration a t from 4.5 t o 5 strokes of the pump in both instances. Later tests showed that distribution was not uniform: and that
Figure 3.
Calibration Apparatus
A series of 39 points was determined, some of which show wide distribution (Figure 4). However, i t is known that a portion of this is due to failure to establish a constant flow before suitohing to the second bubbler. Care also must be taken to avoid Interfering gases in the laboratory, such as ammonium hydroxide or hydrochloric acid. The extreme deviations found, including vdues with known errors, are shown in Figure 4. For the calibration curve (Figure 5), all values in parts per million a t a given number of strokes on the pump were averaged. Any values that varied 1 p.p.m. from this average were discarded and the average of the remaining factors was used to estahhsh fhe find ouve. The calibration curve represents
ANALYTICAL CHEMISTRY
1506
61.57, of all points determined. All discarded values were in the less sensitive part of the curve above 6 p.p.m. PRECAUTIONS
In preliminary work, it was observed that uniformity of treatment of the paper employed is essential to good results. The authors found that a Whatman Xo. 41-H paper is most suitable for the test and the results specified were obtained when this paper was used. Uniformity of deposition of chemicals upon the paper is important, and it is essential to wash out completely the excess 40
’
36
number of strokes of the pump. It is recommended that a definite number of strokes of the pump be taken, so that quantities of fluoride that would be immediately hazardous may be determined quickly. For example, if after 3 strokes of the pump in which 150 cc. of contaminated air have been drawn through the paper a reaction of greater or less intensity has occurred, dangerous concentrations exist and the necessity for immediate evacuation is indicated. If, on removal from the apparatus, the paper shows two very faint lips of pink coloration against the brown background, it will indicate the concentration to be 11 p.p.m. The Erst appearance of pink on the brown background is taken as the end point of the reactions. Should no pink appear, the same paper may be replaced and another five strokes of the pump taken. By taking these strokes in groups of five, it may be roughly determined after 18 strokes that no great danger exists, as the indicated concentration will be less than the 3 p,p.m. stated t o be safe by the United States Public Health Service. iifter the first examination of three strokes, when it is determined that danger does not exist and that immediate evacuation of the area is not imperative, fresh papers
32
CI
28
i 0
7 24. 0 0
I Y
120
0 2
01
4 8 12 16 NUMBER OF STROKES ON 50 C C PUMP
20
Figure 4. All Points Obtained in Determination of Hydrofluoric Acid p-dimethylaminoazobenzenearsonic acid with great care before washing with alcohol and drying in vacuo. The paper was kept in a sealed vacuum chamber but there was, apparently, no great decomposition when it was kept in sealed brown bottles. Paper 3, 6, and 12 months old proved to be as sensitive as freshly made paper, Paper has been kept under carefully controlled conditions for 5 years iTith no detectable loss in sensitivity. Safety considerations, in actual field use, dictate replacement after 3 months to one year because treated paper exposed to light and oxidizing influences is not as sensitive as paper not so exposed. Exposed paper will give indications of concentrations of fluoride in the range of 6 p.p.m. but with a considerable loss in accuracy after one week’s exposure. Such exposed papers are not dependable. FIELD T E S T PROCEDURE
Prepared zirconium azoarsenate paper is placed in the special fixture attached t o a 50-cc. pump. The paper is dipped into 2 N hydrochloric acid immediately before being placed in the apparatus; in the field, the paper may be placed in the holder dry and then wetted with the acid added with a dropper just before use. Air is drawn slowly through the filter paper for a given
0
4 8 12 16 N W B E R OF STROKES ON 5 O C C PUMP
20
Figure 5. Average Calibration Curve
should be used und about l i strokes should be t,aken tu see whether the suggested 3 p.p.m. value has been approached. - b y method useful in a particular situation ma!- he used with the calibration chart. LITERATURE CITED
(I) Assoc. Offic. Agr. Chemists, “Official and Tentative Methods of Analysis,” 5th ed., p. 529, 1940. (2) Clifford, P. A., J . Assoc. O f i c . Agr. chemist.^. 24, 350 (1941). (3) Dept. of Scientific and Industrial Research. “Methods for t h e Detection of Toxic Gases i n Industrl-,” London, H.M. Stationery Office, 1938. (4) Feigl, F., and Rajmann, E . , Xikrochemie, 12,133 (1932). (5) Scott, W. W., “Scott’s Standard Methods of Chemical Analysis,” 5th ed., Vol. 2. p. 2209, New York, D . Van Nostrand Co., 1939. (6) Willard, H. H., and Winter, 0 . B.. ISD.ENG.CHEM..4 X A L . ED., 5 , 7 (1933).
RECEIVED January 8, 1949.
