Analyst, December, 1968, Vol. 93, $$. 821-826 821 A Simple Field Test for the Determination of Hydrogen Fluoride in Air BY B. S. MARSHALL AND R. WOOD (Ministry of Technology, Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, S.E. 1) A field method is described for determining hydrogen fluoride vapour in air at concentrations up to 20pg of hydrogen fluoride per litre. The gas is collected in an acidic solution of zirconium - Solochrome cyanine R reagent and the observed bleaching of the colour compared with standards. The apparatus used is simple to operate and the time required for a determination is less than 5 minutes. h dynamic method for the generation of standard atmospheres of hydrogen fluoride is also described. HYDROGEN fluoride is widely used in the chemical industry and in the petroleum industry in the synthesis of high octane spirit.It is also used in the manufacture of aluminium fluoride and synthetic cryolite for the production of aluminium. Considerable amounts of hydrogen fluoride are also used in the manufacture of refrigerants and fluorine-containing plastics. As hydrogen fluoride gas is highly toxic, with a threshold limit value of 3 p.p.m. v/v (2.5 pg per litre) in air,l there is a need for a simple, rapid field test for its detection in industrial atmospheres at this level. Any such test should preferably be capable of being used by non-scientific staff. Several methods for the determination of hydrogen fluoride in air have been reviewed by Farrah.2 Most of them involve the preliminary trapping and concentration of the gas in water or sodium hydroxide solution, or on filter-papers impregnated with calcium hydroxide, followed by a fluoride distillation and a final colorimetric determination of the fluoride ion.This general method, although normally precise and specific, is time consuming, requires considerable expertise with the necessary distillation apparatus and thus cannot be used to give a rapid check on the hydrogen fluoride concentration present in a factory atmosphere. The main requirements of a field test are ease of manipulation, rapid colour development and a good visual colour discrimination between neighbouring standards. When possible, a positive increase in colour, rather than a bleaching effect, with increasing concentration of air-borne contamination is desirable.With these points in mind we selected for examination two of the methods in general use in fluoride determinations for possible development of a field test for hydrogen fluoride. These were the lanthanum - alizarin - fluorine blue method3 and the zirconium - Solochrome cyanine R method: as modified by Dixoa6 A preliminary assessment of these two methods was carried out with aqueous solutions containing a range of total fluoride from 0 to 2.5 pg. This would be equivalent to the range of hydrogen fluoride concentrations collected from 500-ml samples of industrial atmospheres containing up to twice the present threshold limit value of the gas. Although the zirconium - Solochrome cyanine R method involves bleaching of the colour, it was found to give better visual colour differentiation for this range of fluoride concentrations than the positive colours of the lanthanum - alizarin - fluorine blue method and, consequently, was chosen for further examination as a possible basis for a field test.0 SAC; Crown Copyright Reserved822 MARSHALL AND WOOD: A SIMPLE FIELD TEST FOR THE EXPERIMENTAL [ktna@st, vol. 93 PREPARATION AND CALIBRATION OF STANDARD ATMOSPHERES OF HYDROGEN FLUORIDE- As hydrogen fluoride gas attacks glass, the apparatus for the generation of standard atmospheres of it was constructed so that all of the surfaces in contact with the gas were made of non-reactive materials, e.g., polythene. Prefiaration-Hydrogen fluoride gas from a low-pressure cylinder (supplied by the Matheson Company Inc.) maintained at 22°C in a thermostatically controlled bath was passed through a polythene restrictor and then into a polythene mixing chamber, into which 10 litres of diluting air per minute were flowing.By using a second restrictor, a short length of PTFE tubing with an adjustable screw-clip, 10ml of the dilute atmosphere per minute were passed into a second polythene mixing chamber, in which it was diluted with a further 10 litres of air per minute to give a working atmosphere of about 2.5 pg of hydrogen fluoride per litre. The 10-ml per minute flow-rate was achieved beforehand by connecting the output of the second restrictor to a flow meter and adjusting to the required flow by using hydrogen fluoride free air. Hydrogen fluoride atmospheres of other concentrations were prepared by suitably adjusting the flow-rate of the second diluting stream of air.The polythene restrictor was made by removing the central wire, outer sheath and screening from a piece of co-axial cable, 70 mm long, closing down the bore of the remaining 5-mm 0.d. polythene tubing, by heating, on to a piece of 42 s.w.g. wire and then removing this wire when cold. Adiabatic cooling and consequent liquefaction of the hydrogen fluoride vapour issuing from the polythene restrictor were overcome by inserting, between the cylinder and the restrictor, a 65-mm length of stainless-steel tubing of 6-mm i d . wound with a low-output wire heater. Polythene screw-top bottles of 250 and 1000-ml capacity were easily adapted to serve as the first and second mixing chambers, respectively, and all of the pipework in contact with the hydrogen fluoride was made of polythene.No conditioning problems were experienced with this apparatus when used for the generation of hydrogen fiuoride atmospheres. Calibration-The output of the cylinder was checked by passing the gas issuing from the first restrictor into 0.1 N sodium hydroxide solution for a known length of time, and back- titrating the excess of alkali with standard acid. Under the above conditions, 31 ml of hydrogen fluoride gas were delivered per minute. The working atmospheres were calibrated by passing samples at the rate of 1 litre per minute for 5 minutes through 15 ml of water contained in an absorber fitted with a polythene inlet tube. It had previously been found that virtually 100 per cent. of the hydrogen fluoride at the atmosphere concentrations used was collected in one absorber.Other workers6 had shown that water was as efficient as an alkaline solution as a trapping agent for hydrogen fluoride from air samples. The fluoride collected was determined spectrophotometrically3 with the lanthanum - alizarin - fluorine blue reagent at 625 nm and the concentration of the atmosphere was calculated by reference to a previously prepared standard graph. DEVELOPMENT OF FIELD TEST- With the zirconium - Solochrome cyanine R reagent, Dixon6 has devised an improved spectrophotometric method for the determination of fluoride in water in the range 0 to 2.5 pg. This procedure, and the range of fluoride it covered, appeared to suit our requirements.The development of a test involving the use of visual colour standards equivalent to 0, 1-25, 2.5 and 5 pg of hydrogen fluoride per litre of air was envisaged. A 500-ml sample would normally be taken in the field by using a hand aspirator of 125-ml capacity. However, by reducing the number of aspirations and using the same set of colour standards, the range of hydrogen fluoride concentrations in air capable of being determined could be extended. Standard hydrogen fluoride atmospheres (500ml), containing 1.25, 2.5 and 5.0p.g of the gas per litre, were sampled at the rate of 125 ml per minute through 5 ml of water; 1 ml of each of the Solochrome cyanine R and zirconium solutions6 was added to the absorbing solutions after sampling.The respective colours obtained were compared with those of a similar set of solutions prepared from an aqueous fluoride solution and containing 0.6, 1.2 and 2.4 pg of fluoride. The visual colour matches obtained showed that the hydrogen fluoride was being quantitatively trapped and determined. Later, to minimise the number of opera- tions required, it was found that 5 ml of diluted, mixed reagent containing 1 ml of eachDecember, 19681 DETERMINATION OF HYDROGEN FLUORIDE IN AIR 823 of the chromogenic reagents could be used directly as trapping agent, and also gave colour matches with similarly prepared fluoride solutions over the range 0 to 2-4 pg of fluoride. The visual colour differentiation between the four fluoride levels, 0, 0.6, 1.2 and 2.4 pg, was considered to be entirely satisfactory for field test purposes.PREPARATION OF COLOUR STANDARDS- A set of colour standards for use with the proposed test was conveniently prepared by using a standard sodium fluoride solution and solutions of the chromogenic agents. Details of the preparation are given later. With these standards it was found possible to determine the fluoride content of a sample at least to the nearest f0.625 pg per litre ( i e . , a quarter of the present threshold limit value) between 0 and 2-5pg of hydrogen fluoride per litre and at least to the nearest 1.25 pg per litre between 2.5 and 5Opg of hydrogen fluoride per litre. INTERFERENCES- Aluminium, phosphate and sulphate are well known interferences in the zirconium - Solochrome cyanine R method for fluoride,4 the first being a negative interference and the others positive interferences.The effects of these interferences on the proposed test were individually assessed by adding increasing amounts of each species in solution to a series of fluoride solutions equivalent to 0, 1.25, 2.5 and 5.0 pg of hydrogen fluoride per litre of air, and finally adding the mixed chromogenic reagent. The level of interference of each of the species was taken as the lowest concentration that would just produce a significant visual colour difference between the solution and an equivalent fluoride standard. The concentra- tions for aluminium, phosphate and sulphate were 1, 1.6 and 100, respectively, expressed as micrograms per litre of air. Besides the above study, the effects on the proposed test of several substances, which might occur together with hydrogen fluoride in industrial atmospheres, were examined.When the threshold limit value of any gas known to interfere with the proposed field test is present in an industrial atmosphere, then such a concentration constitutes a hazard in its own right, irrespective of the hydrogen fluoride concentration. In view of this, experiments were conducted to ascertain the effects on the proposed test of possible interfering gases and vapours up to an arbitrarily chosen equivalent of twice their present threshold limit value. Atmospheres were prepared containing 1.25, 2.5 and 5*0pg of hydrogen fluoride per litre by using, in the second dilution stage, air containing amounts of the interfering substances up to a concentration of twice their threshold limit value.These atmospheres were then sampled by the proposed test method. It was found that nitrogen dioxide and sulphur dioxide did not interfere at concentrations up to twice their present threshold limit values. i.e., 18 and 26 pg per litre, respectively. Also, sulphur trioxide did not interfere at concen- trations up to 1.6 pg per litre of air, ie., equivalent to twice the present threshold limit value for sulphuric acid. Experiments were also performed to assess the effect on the test of various amounts of water vapour present in atmospheres containing hydrogen fluoride by setting up, as above, standard atmospheres of gas and using air containing known amounts of water vapour for the second dilution stage.Water vapour did not interfere at concentrations up to 18mg per litre of air at 22" C. In view of the interfering effects of aluminium and phosphate, this field test is not recommended for the testing of industrial atmospheres polluted with fluoride-containing dusts. Several such materials used in industry, e.g., cryolite, apatite and various other phos- phate rocks, contain sufficient aluminium or phosphate to interfere with the proposed test. Although water vapour does not appear to interfere with the test, provided that con- densation does not occur in the sampling apparatus, it is recommended in certain instances that the inlet tube be washed with the mixed reagent after sampling. This is especially important when samples are taken of a damp atmosphere or one in which a dilute aqueous hydrogen fluoride spray has been used.In the sampling of such atmospheres, hydrogen fluoride may be trapped in water droplets on the inside of the inlet tube. The washing of this tube can be carried out by blowing carefully down the side-arm of the glass test-tube and thereby raising the level of liquid in the inlet tube to within 10 mm of its top. (This is equivalent to a relative humidity of 92.5 per cent.)824 APPARATUS- Absorbers-Several are required, each comprised of a glass test-tube with a polythene inlet tube (see Fig. 1). For ease of dispensing reagents, make a mark at the 5-ml leveI. Each test-tube should be fitted with its own inlet tube. MARSHALL AND WOOD: A SIMPLE FIELD TEST FOR THE FIELD TEST FOR THE DETERMINATION OF HYDROGEN FLUORIDE IN AIR [Analyst, Vol.93 A= Rubber bung B=Glass tube, 13mm i.d., C= Polythene inlet tube, 5mm i.d., 18-5-cm over-all I5.5-cm over-all length length Fig. 1. Absorber Asj5irator-A rubber-bulb hand aspirator (obtainable from Siebe Gorman and Co. Ltd,, Davis Road, Chessington, Surrey) adjusted to deliver 125ml of sample per minute. REAGENTS- All reagents should be of analytical-reagent quality unless otherwise stated. Zircoaizlm soZzctiort-l>issolve 36 mg of zirconium oxychloride octahydrate, ZrOC1,.8H20, in water, add 1201111 of concentrated hydrochloric acid and dilute to 500ml. Solochrome cyanivze R (C.I. 43820) soZzctio.n--Dissolve 60 mg of the purified reagent in water, add 24ml of N hydrochloric acid and dilute to 250ml. This reagent, as normally supplied, contains variable amounts of sodium sulphate.The dye can be separated from it by extraction with methanol. The extract is evaporated to dryness under reduced pressure, and the resulting purified material used to prepare the reagent solution. Mixed reageat-Combine 10ml of each of the zirconium and Solochrome cyanine R solutions and dilute to 200ml with water. This solution appears to be stable for a con- siderable time, but to obviate any possible contamination problems, it is recommended that fresh mixed reagent be prepared for each series of tests. Stavzdard j h r i d e solutioa-Dissolve 2.121 g of sodium fluoride in water, add 1 ml of 0.1 N sodium hydroxide solution and dilute to 1 litre with water. Dilute 2-5 ml of this solution to 1 litre to give a solution containing 2.4pg of fluoride per ml, i.e., equivalent to 2-5pg of hydrogen fluoride per ml. PROCEDURE- In an uncontaminated atmosphere, well away from the suspected source of hydrogen fluoride, place 5 ml of the mixed reagent in the absorbing tube, insert the inlet tube and connect the aspirator to the side-arm.Transfer the assembled apparatus to the sampling site, and draw a 500-ml sample (i.e., four aspirations) of the atmosphere through the reagent. Remove the apparatus to an uncontaminated atmosphere. Disconnect the aspirator fromDecember, 19681 DETERMINATION OF HYDROGEN FLUORIDE IN AIR 825 the side-arm of the glass test-tube. If the sample has been taken in a humid atmosphere, wash the inlet tube by gently blowing down the side-arm, and raise the level of the mixed reagent to within about 1Omm of the top of the inlet tube.Place a finger on the end of the inlet tube to hold the mixed reagent in that position for a few seconds. Allow the liquid to flow back into the glass test-tube and remove the inlet tube. Compare the colour of the sample solution, preferably in daylight, in turn with 5ml of each of the fluoride colour standards contained in tubes of similar diameter to the sample test-tube. View through the depths of the respective liquids against a white (paper) background. Should the level of hydrogen fluoride with a 500-ml sample of atmosphere be above 5 p g per litre, a more accurate determination of the true concentration can be made by using a 125-ml sample. The standard that gives a colour match with the sample solution is then multiplied by four to give the amount of hydrogen fluoride in micrograms per litre present in the atmosphere.When a test is required on an atmosphere that is either not readily accessible or possibly contains high concentrations of hydrogen fluoride, the sampling should be carried out as follows. Connect the site to be tested to the sampling apparatus with a length of polythene tubing. Start aspirating and continue until a colour change is obtained in the mixed reagent. Dis- connect the absorber and rapidly replace it with another containing fresh mixed reagent. Carry out the test as described above. It is recommended that an operator sampling high concentrations of hydrogen fluoride in an atmosphere should have suitable respiratory pro- tection.NOTE- should be rinsed with a few millilitres of fresh reagent and the inlet tube dried before re-use. To avoid cross-contamination between samples, each absorber, i.e., test-tube and inlet tube, PREPARATION OF FLUORIDE COLOUR STANDARDS- To a set of four 100-ml graduated flasks add 0, 5, 10 and 20 ml of the dilute standard fluoride solution. Add 5 ml each of the zirconium and Solochrome cyanine R solutions to each flask and dilute to volume with water. These standards represent 0, 1-25, 2.5 and 6 pg of hydrogen fluoride per litre of air, respectively. The standards, if stoppered, are stable for at least 1 week. APPLICATION OF METHOD- The proposed test was assessed and checked under field conditions at various sites where hydrogen fluoride was being used.The check testing was normally carried out by taking three concurrent samples for each atmosphere analysed and using a 3-way manifold and a common sampling inlet. The three samples taken were as follows. (i) A 500-ml sample by the proposed test over a period of 5 minutes. (G) A 5-litre sample taken at 1 litre per minute, with 50 ml of mixed reagent as absorbing solution. This was a ten times scaling-up of the field test, and the fluoride collected was determined visually as in (i) by using colour standards. (iii) A 5-litre sample taken at 1 litre per minute, with 15ml of water as absorbing solution to which lanthanum - alizarin - fluorine blue reagent was added after sampling. The fluoride was determined spectrophotometrically, with reference to a previously prepared calibration graph.Table I gives the results obtained in one check-testing series carried out in and around a fume cupboard in a laboratory in which hydrogen fluoride was being used. Some of the samples were deliberately taken inside the fume cupboard to enable an assessment of the field test to be made at high levels of hydrogen fluoride contamination. The results shown in Table I, and those obtained at other sites, appeared to confirm the validity of the proposed field test for the determination of hydrogen fluoride in air. It should, however, be mentioned that in some of the early check tests performed, high and erratic results were obtained when using the field method compared with the other two concurrently taken samples.This was traced to the contamination of the sampling absorber, the polythene inlet tube in particular, between the taking of successive samples. It appeared826 MARSHALL AND WOOD TABLE I COMPARISON OF THE RESULTS OF THE ANALYSIS OF HYDROGEN FLUORIDE CONTAMINATED ATMOSPHERES BY THE PROPOSED FIELD TEST AND TWO OTHER METHODS Hydrogen fluoride concentration found, pg per litre, by- r A -l Lanthanum - alizarin - fluorine blue reagent, Sample Field test Field test x 10 spectrophotometrically 1 0 0 0.26 2 0.625 0.625 1.0 3 0.625 0.625 0.625 4 0-625 0-625 0-86 5 1-25 1-25 1.76 6 1.25 1-25 1-12 7 2-5 2 4 2.6 8 3.75 3.76 2.5 9 1@0* 1o.ot 9.3 10 20.0* 20.0f. 24-52 * A 125-ml sample taken at the mid-point of time during which the sample for spectrophotometric determination was taken. t A 2.5-litre sample. $ A 1-25-litre sample. that as the volume of absorbing solution used in the field test was small compared with those used in the other two concurrent tests, it was relatively more susceptible to contamina- tion from the inlet tube used. This effect was obviated either by using fresh absorbers for each test or washing the absorber components with mixed reagent between tests and drying the inlet tube. The test apparatus should always be assembled in an uncontaminated atmosphere. This work was carried out on behalf of the Department of Employment and Productivity Committee on Tests for Toxic Substances in Air. We thank the Government Chemist for permission to publish this paper. REFERENCES 1. 2. 3. 4. 5. 6. “Ministry of Labour, Safety, Health and Welfare, New Series No. 8, Dusts and Fumes in Factory Farrah, G. H., J. Air Pollut. Control Ass., 1967, 17, 738. Greenhalgh, R., and Riley, J. P., Analytica Chim. Ada, 1961, 25, 179. Megregian, S., Analyt. Chenz., 1964, 26, 1161. Dixon, E. J., in preparation. Pack, M. R., Hill, A. C., Thomas, M. D., and Transtrum, L. G., A.S.T.M. Special Technical Received June 27th, 1968 Atmospheres, Third Edition, H.M. Stationery Office, London, 1966. Publication No. 281, 1958, p. 27.