Analyst, October, 1966, Vol. 91, @. 625-629 625 An Improved Method for Determining Residues of Diquat BY A. CALDERBANK AND S. H. YUEN (Imperial Chemical Industries Ltd., Agricultural Division, Jealott’s Hill Research Station, Bracknell, Berkshire) The ion-exchange method for determining diquat residues in potato tubers1 has been modified to give a simpler operating procedure and increased accuracy. Experiments on untreated samples with 0-08 to 0.2 p.p.m. of diquat added showed an average recovery of 76 per cent. with a standard deviation of 11 per cent. The method, with minor modifications, has been found to be generally applicable to other food crops and to water. For a 250-g sample, the limit of determination is 0-01 p.p.m. DIQUAT, or 1 ,l’-ethylene-2,2’-bipyridylium cation, which is manufactured in the form of its dibromide salt, is the active ingredient of Reglone and Preeglone herbicides.The method previously describedl relies on extracting diquat residues from potato tubers by boiling them with N sulphuric acid. The filtered hydrolysate is neutralised and passed through a cation-exchange resin column which retains the diquat together with some of the natural plant constituents. The diquat is then eluted with saturated sodium chloride solution and, after reduction, determined spectrophotometrically in the region of 379 mp. The main disadvantage of this method lies in the low average recovery (59 per cent.), and the rather tedious procedure involved in neutralisation and filtration. Since this method was developed, diquat has found increasing usage on a variety of crops, and the operating variables in the method have been subject to more critical examination.By modifying the conditions at the neutralisation, filtration and elution stages, increased recoveries have been obtained, and the revised method is suitable for routine analysis of large numbers of samples. Further, the method, with minor modifications, is now applicable to other crops, to animal tissue and excreta,2 milk2 and water. The general principles involved in the use of ion-exchange resins in residue analysis have already been discussed in some detail.3 The specific problems that are involved at the various stages of the method for determining diquat, and the modifications made, are described below. EXPERIMENTAL NEUTRALISATION OF ACID HYDROLYSATE- Traces of diquat are held quite firmly to starch, protein and leaf surfaces, and it has been found necessary to boil the macerated plant material with dilute sulphuric acid to release the diquat into solution.If the filtered acid hydrolysate is passed directly without neutralisa- tion through the cation-exchange resin, as described for the complementary bipyridylium herbicide, p a r a q ~ a t , ~ the recovery of diquat in the final effluent is low and the background absorption is high. This is because many of the hydrolysed plant constituents are basic and compete with diquat for adsorption sites on the resin under acid conditions. Further, the retained plant constituents are eluted along with diquat and increase the background absorption in the final effluent.Neutralisation of the hydrolysate has the advantage of giving better recoveries of diquat and considerably lower background absorption. This is because the ionisation of weakly basic materials in the hydrolysate is suppressed under neutral or alkaline conditions, and consequently these materials are then not readily retained by the resin column, leaving more sites available for diquat, the ionisation of which is not appreciably affected by change of pH. Neutralisation of the acid hydrolysis was previously carried out with calcium and sodium carbonates1 This is a tedious operation and has the disadvantage of causing losses of diquat by adsorption on to the precipitated calcium sulphate. The procedure is simplified con- siderably, and losses of diquat a t this stage are eliminated by neutralising with 10 N sodium hydroxide solution.The excess sodium ions in solution (about 0.5 N) do not interfere with the retention of diquat by the ion-exchange resin. Further, the end-point of alkali addition is easier to detect because the solution darkens appreciably and a fine precipitate usually forms.626 CALDERBANK AND YUEN: AN IMPROVED METHOD [Analyst, Vol. 91 The use of a filter aid is desirable to speed up filtration of the hydrolysate. It is now known that Hyflo Super-cel, which was previously used, removed an appreciable amount of diquat from solution by adsorption. Under comparable conditions, a coarser grade of a diatomaceous silicate, Celite 545, was found to retain less diquat, while allowing a quicker filtration.The coarsest grade of filter aid, Celite 560, gave a fast filtration rate, but did not completely remove the fine materials suspended in solution. ELUTION OF DIQUAT- Saturated sodium chloride solution was previously chosen for displacing diquat retained by the cation-exchange resin, Zeo-Karb 225 (containing 8 per cent. of divinylbenzene). Reduc- tion of diquat with sodium dithionite solution could be carried out in this eluant as effectively as in pure solution. Saturated sodium chloride solution, however, suffers from the dis- advantage that it produces tailing, so that recovery of diquat from the resin in 25ml of effluent is only 80 per cent. Hydrochloric acid (1 + 1) and saturated ammonium chloride solution (5 M) produce sharper elution peaks, the saturated ammonium chloride solution giving 95 per cent.recovery in 25 ml and quantitative recovery in 50 ml of effluent; both solutions introduce complications in the final assay. By modifying the conditions of the reduction it has been found possible to determine diquat accurately in saturated ammonium chloride solution, and consequently this solution is adopted as the preferred eluant. A 2.5 per cent. ammonium chloride solution (0.5 M) displaces diquat slowly from the selected cation-exchange resin, and it has been found that a 3.5-g bed of resin can be washed with up to 200 ml of 2.5 per cent. ammonium chloride solution before the adsorbed diquat starts to be eluted. In practice, the resin column is washed with 150 ml of this solution before the diquat is eluted.This procedure reduces the background absorption in the final effluent to a low level. REDUCTION AND DETERMINATION OF DIQUAT- When diquat is reduced in saturated ammonium chloride solution with 0.1 per cent. sodium dithionite in 3 N sodium hydroxide solution,l the colour fades rapidly. The stability of the colour is improved by decreasing the sodium hydroxide concentration, or by increasing the concentration of sodium dithionite. There is a limit to the concentration of sodium dithionite that can be used, because sodium dithionite absorbs strongly in the region of 379 mp, in which region reduced diquat has an absorption maximum. A reagent consisting of 0.2 per cent. sodium dithionite in 0-3 N sodium hydroxide gave a reduced diquat colour that was stable for about 3 minutes.A mixed reducing reagent containing 0.2 per cent. each of sodium dithionite and sodium metabisulphite in 0.3 N sodium hydroxide produced a reduced colour in saturated ammonium chloride solution that was stable for 30 minutes and this was stan- dardised. The conversion of diquat to its free radical is reversible, and if the dithionite is largely oxidised, the free radical will revert to diquat by losing an electron. The mixed reducing reagent may generate a buffered oxidation - reduction potential in which the free radical is more stable. In practice, the colour developed by adding the mixed reagent to ammonium chloride effluents from ion-exchange columns is rather less stable than the colour developed in pure ammonium chloride solution, presumably owing to the lower stability of the reduced back- ground absorption.However, the colour developed in the ion-exchange effluents is stable for 5 to 6 minutes, and thereafter fades slowly. For this reason, measurements of optical density should be completed within 5 minutes of adding the reagent. METHOD APPARATUS- Potato chipping machine, Macerator-A top-drive macerator, obtainable from Townson and Mercer, was used. BoiZing $asks---Round-bottomed flasks of 2-litre capacity fitted, by means of standard Heating mantles-300 watt and 1-litre capacity, with a maximum temperature of 450" C. Spectrophotometer-A Unicam SP600. ground-glass joints, with reflux condensers. An electrothermal heating unit containing six mantles is suitable.October, 19661 FOR DETERMINING RESIDUES OF DIQUAT 627 REAGENTS- Sulphuric acid, 18 N-Add cautiously with stirring 1 litre of sulphuric acid (sp.gr.1.84) to 1 litre of water and dilute the cooled solution to 2 litres with water. Octan-2-01. Celite 545-Obtainable from Johns-Manville Co. Ltd. Sodium hydroxide solution, 10 N. Indicator paper-eg., Universal test paper (Johnsons of Hendon Ltd.). Ethylenediaminetetra-acetic acid, disodium salt (EDTA) solution, 5 per cent. Cation-exchange resin-Permutit Zeo-Karb 225 (52 to 100 mesh), containing 8 per cent. of divinylbenzene in the sodium form. To prepare the column, wash 3-5 g of the resin with water into a 25-ml burette (i.d., 9 to 10 mm, and 50 cm long). Pass successively through the column at about 5ml per minute, 20ml of 2 N hydrochloric acid and 50ml of water.The column is then ready for use. Use a fresh column for each test. Ammonium chloride solution, saturated and 2.5 per cent. STANDARD SOLUTIONS OF DIQUAT- Stock solution, 250 9.p.m.-Dissolve 0.1229 g of pure diquat dibromide monohydrate, C,,H!,N,Br,.H,O (mol. wt., 362.1 ; 50.9 per cent. cation), in saturated ammonium chloride solution and make up to 250 ml with saturated ammonium chloride solution. Solution A, 10 p.p.m.-Dilute 10 ml of stock solution to 250 ml with saturated ammonium chloride solution. Also prepare standard solutions B, C, D and E (containing 1.5, 1.0, 0.5 and 0.25 p.p.m. of diquat) by diluting 15, 10,5 and 2.5 ml, respectively, of solution A to 100 ml with saturated ammonium chloride solution. These solutions are stable under normal laboratory conditions, but must not be exposed to sunlight. Reducing reagertt-Dissolve 0.2 g each of sodium dithionite and sodium metabisulphite in 100 ml of 0.3 N sodium hydroxide.This solution should be prepared immediately before use, and should on no account be used after keeping for more than 1 hour. It should be stored in 1-02 bottles (tightly covered with screwed-on lids) which are kept in a desiccator. Solid sodium dithionite is unstable in the presence of moisture. EXTRACTION AND CHROMATOGRAPHIC SEPARATION OF DIQUAT- Take about 1000 g of potato tubers at random from the sample provided, wash free from soil and remove surplus water with a dry cloth. Cut each tuber into four approximately equal segments and reject two that are opposite.Cut the remaining segments into small pieces and mix them thoroughly. Weigh out 250 g of tuber pieces into a macerator jar, add 200 ml of water and macerate for 3 minutes. Transfer the macerated material to a boiling flask. Rinse the jar with 50 ml of water, and add the rinsings to the contents of the flask, followed by 25 ml of 18 N sulphuric acid and 1 ml of octan-2-01. Mix well, attach a reflux condenser to the flask and heat to boiling on a heating mantle. Swirl the mixture occasionally, to prevent local overheating and charring, until the solution is boiling steadily. Boil under reflux for 5 hours and allow to cool (the solution can be left overnight at this stage). Prepare a Celite 545 filter as follows: moisten a Whatman No, 5 filter-paper with water under suction on a 16-cm Buchner funnel, supported on a 2-litre filter flask.Pour 150 ml of an aqueous suspension containing l o g of Celite 545 over the paper, suck dry, and discard the filtrate. Wash down the condenser with 50 ml of water, and filter the contents of the flask by moderate suction through the prepared filter. Suck the residue dry and wash with 150ml of water. Transfer the filtrate to a 2-litre beaker, and neutralise to pH 8 to 9 by adding, slowly with stirring, 10 N sodium hydroxide (about 45 ml) with a Universal test paper as external indicator (the colour deepens near the end-point and a fine precipitate usually forms). Add 50 ml of EDTA solution (the pH of the solution should now be 6 to 7) and re-filter the solution through a Celite 545 filter as before.Transfer the solution (900 to 1000 ml) to a 1-litre separating funnel suspended above the prepared resin column. Allow the solution to percolate through the column at a flow-rate of 5 to 10 ml per minute. Remove the funnel, and wash the column at 3 to 4 ml per minute628 CALDERBANK AKD YUEN: AN IMPROVED METHOD [Analyst, Vol. 91 with 150 ml of 2.5 per cent. ammonium chloride solution. Elute the diquat by passing saturated ammonium chloride solution through the resin column at about 1 ml per minute. Collect 50 ml of the effluent in a 50-ml calibrated flask, and mix. DETERMINATION OF DIQUAT- Transfer by pipette 10.0 ml of the effluent into a 15-ml glass-stoppered test-tube, add 2.0 ml of reducing reagent, and mix. Within 5 minutes, measure the optical density of the solution at 375, 379, 383 and 385 mp in a 4-cm optical cell, against a reference solution consisting of 10 ml of saturated ammonium chloride solution and 2 ml of reducing reagent.Call these E,,, and, respectively. Measure, concurrently with each series of analyses, the optical densities at 379mp of 10-0-ml aliquots of standards B, C, D and E, after adding 2-Om1 of reducing reagent. Construct a calibration graph relating optical densities of the standards to concentrations of diquat in p.p.m. CALCULATIOX- absorption by means of equations (1) and (2), and call the corrected optical densities and respectively. Correct the optical density at 379mp of the sample solution under test for irrelevant El379 == 3.79 E.379 - 2.28 E375 - 1.52 E385 . . .. * * (1) E”379 = 2.49 (2E37g - E375 - E.383) .. .. .. * (2) Use the mean of E’3,9 and Then the diquat content in parts per million to read off from the prepared calibration graph the concentration of diquat in the final effluent. Call this amount Y p.p.m. Y 100 5 percentage recovery - - - x RESULTS RECOVERY- To establish the accuracy of the improved method, 47 recovery experiments were carried out on 250-g portions of untreated potato tubers, with 0.08 to 0.2 p.p.m. of diquat added before hydrolysis. The recoveries (Table I) ranged from 51 to 97 per cent. with an over-all mean of 76 per cent. (standard deviation is 2 1 1 per cent.). TABLE I RECOVERY OF DIQUAT ADDED TO UNTREATED POTATO TUBERS Diquat added, Diquat recovery, per cent. Number of experiments p.p.m. mean standard deviation 22 0.08 7 5 f 8 19 0-1 76 & 13 6 0.2 79 & 14 By means of the two derived equations the background absorption contributed by natural plant constituents is eliminated, leaving only that contributed by diquat.The mean corrected optical densities at 379 mp, measured in a 4-cm optical cell and obtained on un- treated potato tubers, were usually found to be within 20.02, and this is equivalent to apparent diquat contents of +0-007 p.p.m. Taking into account the average apparent diquat content of untreated samples, the average recovery and the size of sample taken, the limit of determination for the improved method is considered to be 0.01 p.p.m. APPLICATION TO OTHER CROPS AND TO WATER- Because of its quick action and lack of residual activity in the soil, diquat has proved useful in a wide range of crops for pre-emergence and post-emergence weed control, pre- harvest desiccation and aquatic weed contr01.~ f6 With minor modifications, the improved method has been applied satisfactorily to a variety of other food crops and to water.For samples of high dry matter, such as seeds, and for those, such as grass and cereal straws, that have been in direct contact with the spray and are likely to contain higher residues,October, 19661 FOR DETERMINING RESIDUES OF DIQUAT 629 the amount taken for analysis can be conveniently reduced to 25 or 50 g. The diquat residues present in these samples can be quantitatively extracted into N sulphuric acid by boiling for 3 hours. Diquat is also an efficient aquatic herbicide at concentrations in the water of 1 p.p.m.or below. For analysing treated water, acid hydrolysis is not required, and traces of diquat that may be present are concentrated in the usual way by passing the sample solution through the resin column, followed by washing and elution. The results obtained by applying this ion-exchange method to various crops and to water are summarised in Table 11. Applica- tion of the method to the determination of residues of diquat in animal tissue, excreta and milk is described elsewhere.2 TABLE I1 SUMMARY OF THE RESULTS OBTAINED BY APPLYING THE IMPROVED METHOD TO VARIOUS CROPS AND TO WATER Period of boiling, Limit of determination, Recovery, Sample applied* Size of sample hours p.p.m. per cent. Potato tubers Fruits . . ::} 250g Cereal grains Peas and beans . . Cotton seeds . ’‘1 . Rape seeds . . Sunflower seeds , . J Grass . . Cereal straws : } Vegetables. . .. 50g Alfalfa and clover 25 g Silage . . .. 5 3 3 0 0.01 66 to 80 0.05 0.1 60 to 75 70 to 85 0.03 0.003 90 to 100 * These samples have been analysed with satisfactory results. We thank Mr. R. H. McKenna for assistance with the experimental work. REFERENCES 1. 2 . 3. 4. 5. 6. Calderbank, A., Morgan, C. B., and Yuen, S. H., Analyst, 1961, 86, 569. Black, W. G. M., Calderbank, A., Douglas, G., and McKenna, R. H., J . Sci. Fd Agvic., in the press. Calderbank, A., in Gunther, F. A., Editor, “Residue Reviews,” Springer-Verlag, Berlin, Gottingen and Heidelberg, 1966, Volume 12, p. 14. Calderbank, A., and Yuen, S. H., Analyst, 1965, 90, 99. Calderbank, A., and Crowdy, S. H., Ann. Rep. Appl. Chew., 1962, 47, 536. Springett, R. H., Outl. Agric., 1965, 4, 226. Received December 13th, 1965