ANALYST, SEPTEMBER 1984, VOL. 109 1159 Spectrophotometric Determination of Cyanide with Pyridine and 4,4'-Dia m i nost il bene-2,2'-disu I p ho n ic Acid Enric Casassas, Roser Rubio and Gemma Rauret Analytical Chemistry Department, Barcelona University, Barcelona 28, Spain The use of 4,4'-diaminostil bene-2,2'-disulphonic acid (amsonic acid) as an alternative coupling agent to barbituric acid in the determination of cyanide is proposed. Amsonic acid possesses cost and availability advantages, and the stability of the pyridine - amsonic acid reagent is greater than that of pyridine - barbituric acid (1 7 d versus 24 h). The method can be applied at a wide range of cyanide concentrations, from 0.01 to 0.10 pg ml-1 a t 490 nm and from 0.10 to 0.60 pg ml-' at 545 nm. The detection limit is about the same order of magnitude as the standard method.Keywords: Cyanide determination; 4,4'-diaminostilbene-2,2'-disulphonic acid; utraviolet - visible spectro- photometry; p yridine The high toxicity of the cyanide ion at low concentrations necessitates its analysis in a variety of environmental samples (drinking waters, superficial waters, etc.) with a very low cyanide content. Among the many methods described for this purpose several spectrophotometric methods based on the Konig reaction'.' stand out because of their sensitivity. This reaction is based on the oxidation of cyanide to cyanogen halide, and on the subsequent reaction of this compound with pyridine to form glutaconic aldehyde, which hydrolyses and couples with an aromatic amine to give a polymethine dye.A great number of amines have been proposed as coupling agents for the analysis of cyanide, among them benzidine3.j and p-phenylenediamines; other proposed coupling agents are pyrazoloneh and barbituric acid.7.8 Bark and HigsonY.10 gave an extensive list of heterocyclic amines used instead of pyridine or as coupling agents. For cyanide determination in water and in waste water the methods using pyridine - pyrazolonel l-'3 or pyridine - barbituric acid as reagents are widely used; the latter was adopted as the APHA - AWWA - WPCF standard method. 14 These two reagents provide a very sensitive tnethod but they have low stability in solution and must be prepared just before use. It is interesting. therefore, to study new reagents that form stable solutions in pyridine - water and give a sensitivity at least of the same order of magnitude as the standard method.Unfortunately, barbituric acid is a con- trolled, expensive drug and can be obtained only with difficulty, particularly when a large number of determinations must be carried out." The Konig reaction was proposed initially for the determination of pyridine. At present, the coupling agent recommendedlh for dye formation in pyridine analysis is 4,4'-diaminostilbene-2.2'-disulphonic acid (amsonic acid), which was introduced by Fuentes-Duchemin and Casassas. 1' This coupling agent stands out in comparison with other amines and barbituric acid because of its sensitivity and the stability of the colour formed. These characteristics make the study of amsonic acid as a reagent for spectrophotometric cyanide determination interesting.The main variables affecting the application of the Konig reaction to the determination of cyanide are the type of oxidising agent used to obtain the cyanogen halide, the working pH and the nature of the buffer solution used. Bromine water,',j.'s chloramine-Th and N-chloro- succinimide,lg among others, have been used as oxidising agents. Phosphate solutionlj." and boric acid - borate solu- tion" have been used as buffers. In this work the influence of the above parameters on the polymethine dye formation were studied when amsonic acid was used as the coupling agent. and a new spectrophotometric method for the determination of small concentrations of cyanide is proposed. The proposed method is compared with the standard method14 for water and waste water analysis.Experiment a1 Apparatus Beckman spectrophotometer Acta M VII (UV - visible - IR). This is used in conjunction with a 1-cm optical path quartz cell. Radiometer PH64 pH meter. This can be fitted with a combined glass - calomel electrode GK2401B. Reagents Chloramine- T. Analytical-reagent grade material obtained from Probus; 0.1 and 1.0% aqueous solutions are prepared daily. Arsenic( IZI) solzction. A 2.6-g amount of arsenic(II1) oxide and 2.3 g of sodium hydroxide are dissolved in about 70 ml of water, and concentrated hydrochloric acid is added to pH 7. The mixture is then diluted to 100 ml. Amsot?ic acid - pyridine reagent. This reagent is made up at two concentrations, 2 or 5% amsonic acid.To 2 or 5 g, respectively, of amsonic acid (Merck. zur Synthesis. purified by recrystallisation) 25 ml of pyridine (Fluka. puriss grade) and 1 ml of concentrated HCl are added and the mixture is diluted with water to 100 ml. The final pH of the yellow solution is 6.1-6.2, the solution being stored at 4 "C in a dark bottle. Cyanidesolutions. A stock solution containing 332 mg 1-1 of cyanide is prepared from potassium cyanide (Probus, analytical-reagent grade). The solution is standardised vol- umetrically by titration with 0.01 mol dm-3 silver nitrate solution using 5-(4-dimethylaminobenzylidene)rhodanine as the indicator. From this stock, two standard solutions with 33.2 and 1.66 mg 1 - l of cyanide. respectively, are prepared. Buffer solutions. These solutions are prepared as described by Bower and Bates.Borate - boric acid solutions are made from 1 mol dm-3 H3B03, 1 rnol dm-3 KCI and 1 mol dm-3 NaOH; Tris solution from 1.0 rnol dm-' of tris- (hydroxymethy1)aminomethane and 1 mol dm-3 of HC1: and phosphate solution from 1 rnol dm-' of KH2P04 and 1.0 mol dm-3 of NaOH. Procedure For low cyanide Concentrations (0.1 pg ml-1) To an aliquot of the cyanide-containing solution (maximum volume 5 ml), 15 ml of borate - boric acid buffer (pH 7.3) are1160 ANALYST, SEPTEMBER 1984, VOL. 109 added, followed by 1 ml of 1% chloramine-T solution; the mixture is shaken for 1 min and then 1 ml of arsenic(II1) solution is added. The reaction mixture is shaken again and 2 ml of pyridine - amsonic acid reagent are added.The mixture is next diluted to 25 ml with water (the final cyanide concentration must lie between 0.006 and 0.1 pg ml-1). The absorbance is measured against a blank at 490 nm no later than 2 min after addition of the reagent. For cyanide concentrations higher than 0.1 pg ml- * For cyanide concentrations between 0.1 and 0.6 pg ml-1 the above procedure is followed, but the absorbance is measured at 545 nm after 1 h. Results and Discussion Absorption Spectrum of the Coloured Compound The recommended procedure for low concentrations initially produces a strong orange solution (absorbance maximum at 490 nm). which turns red in about 1 h. During this time. the position of the spectral maximum shifts to 545 nm and the absorbance at the maximum decreases slightly. The red colour is stable for more than 30 min.At pH 6.5 the colour change is initially faster and the stability is also reached in 1 h. Fig. 1 shows the aborption spectra obtained from a solution contain- ing 0.185 pg ml-1 of cyanide at different times after the solution is prepared. A 0.6 al 6 0.4 +? :: 2 0.2 4 Wavelengthinm Fig. 1. Absorption spectra of a 0.185 pg ml- I cyanide solution after different reaction times. A. After 2 min: B. after 30 min: C. after 45 min: and D, after 60 min Effect of pH The course of the reaction is greatly influenced by the pH; the absorbance measured at 490 nm passes through a plateau at pH 7-7.5. In Fig. 2 the absorbance at 490 nm of a 0.185 pg ml-1 solution of cyanide within the pH range from 6.5 to 10.0 is shown. For the colorimetric determination pH 7.3 was adopted.The buffer composition affects the sensitivity of the method. Three types of buffer solution have been tested: borate - boric acid, KH2P04 - NaOH and Tris - HCI. The best results were obtained with the borate - boric acid buffer, as much lower absorbances resulted with the other buffers. The strong depressant effect of the phosphate ion agrees with that observed by Casassas and Duchemin2" in the nicotinic acid determination with amsonic acid, and by Bark and Higsonll in the colorimetric determination of cyanide with pyridine and p-phenylenediamine. Effect of the Nature and the Amount of Oxidising Agent The amount of oxidising agent has a noticeable influence on the colour intensity of the final solutions. and therefore on the detection limit of the procedure.It also affects the absorbance changes with time. No advantage has been observed when using other halogen- ating reagents (bromine and N-chlorosuccinimide - succinimide) rather than chloramine-T, so this last reagent has been adopted in the recommended procedure. In order to avoid losses of HCN upon acidification of the cyanide solution, the oxidation is carried out at the same pH in which coupling with amsonic acid occurs (pH 7.0-7.5). Halogenation requires a reaction time of 1 min, a pH of 7.3 and agitation. It has been observed that arsenic(II1) solution must be added in order to reduce the unused chloramine-T, otherwise the orange colour is quickly destroyed. 0.6 al c (0 n i 0.4 2 0.2 I I 9 0 ' 7 8 PH D Fig. 2. solution after a 2-min reaction time Absorbance at 490 nm Venus pH for a 0.185 pg ml- I cyanide Table 1.Absorbance change with time (for a 0.136 pg ml-1 cyanide solution) Time/min Oxidising agent A3y0 ,,,,, Oxidising agent 2 I .00 ml of 0.480 1 .OO ml of 10 0.10% chloramine-T 0.451 1 .0% chloramine-T 20 0.410 30 0.374 40 0.331 50 0.312 60 0.287 70 0.264 80 0.242 A ~ Y O nm 0.510 0.460 0.412 0.372 0.337 0.302 0.274 0.250 0.222 n m 0.263 0.279 0.295 0.308 0.316 0.323 0.323 0.323 0.323ANALYST, SEPTEMBER 1984, VOL. 109 1161 Table 2. Absorbance values at two different amsonic acid concentrations (readings at 490 nm taken after 1 min reaction time and at 545 nm after 1 h reaction time) Amsonic acid concentration, YO Cyanide 490 nm 545 nm concentration/ pg ml-1 2 5 2 5 0.012 0.036 0.061 0.016 0.028 0.037 0.140 0.157 0.066 0.087 0.062 0.240 0.261 0.119 0.135 0.093 0.356 0.385 0.178 0.197 0.136 0.481 0.510 0.502 0.260 Table 3.Data related to the quantification limit for the proposed method using amsonic acid and for the standard method using barbituric acid Amsonic acid method Barbituric 490nm 545nm at578nm acid method Log (molar absorptivity) Sandell sensitivity/ (on a cyanide molar basis) . . 5.00 4.77 5.06 pg Cm - 2 pg ml-1 0.039 0.066 0.036 centration factor = 2.5)/ pgml-1 . . . . . . . . 0.016 0.026 0.014 2 . 5 9 ~ 10-4 4.40X 10-4 2.27X . . . . . . . . Detection limit/yg ml- . . 0.013 0.022 0.012 Quantification limit/ Quantification limit (con- . . . . . . . . Y I I 0.2 t n 0.1 ! 0 5 10 15 Time/min Fig. 3. Absorbance (at 490 nm) versus reaction time for a 0.093 pg ml-1 cyanide solution using the proposed amsonic acid method using different oxidising agents.B . 1 .0 ml of 1 .O%, chlorarnine-T solution; C, 1 .O ml of o.lo/o chloramine-T solution; D, 0.1 ml of 0.1% chloramine-T solution; A, obtained from the standard barbituric acid method, at 578 nm Stability with Time When 0.1 ml of a 0.1 O/O chloramine-T aqueous solution is used for oxidation, the coloured compound (absorbance maximum at 490 nm) formed from cyanide solutions in the concentration range between 0.012 and 0.093 pg ml-1 is stable for about 7 min. For more concentrated test solutions the colour stability at 490 nm decreases more and more rapidly with time. On the other hand, when 1 ml of 1% chloramine-T solution is used the absorbance value is inititially increased, but the colour stability, particularly at concentrations higher than 0.09 pg ml-1, noticeably decreases.This is the reason why absorbance readings at 545 nm are recommended for the more concentrated solutions. In Table 1 absorbance values obtained at different times after reaction (up to 80 rnin) are listed. Values in the table are obtained from a solution containing 0.136 pg ml-1 of cyanide and using different amounts of chloramine-T. Fig. 3 shows the absorbance change occurring during 15 rnin when a 0.093 Table 4. Effect of diverse ions in cyanide determination Foreign ion CI- . . . . NO3- . . NO,- . . so32- . . S042- . . Fe( CN)64- CH3COO- Fe( CN)63 - Fe3+ . . . . s2- . . , . SCN- Foreign ion Cyanide concentration/ found/ Relative Added as yg ml-l pg ml-1 error, % NaCl 1000 0.19 - KN0, 1000 0.19 - NaNOz 1000 0.19 - Na2S03 1000 0.19 - Na2S04 1000 0.19 - CH3COONa.3Hz0 1000 0.19 - K,Fe( CN)6.3H20 2 0.19 - 10 0.20 5.3 4 0.21 10.5 11 0.21 10.5 Na2S.9Hz0 0.1 0.19 * 0.9 0.19 * 9 0.19 * 9 Turbidity KSCN 0.02 0.19 * 0.23 0.25 * 0.90 0.27 * K,Fe(CN), 2 0.20 5.3 FeCl, 2 0.19 - * These compounds decrease the stability with time of the coloured compound. pg ml-1 cyanide solution is treated according to the proposed procedure (readings at 490 nm) at three different oxidising concentrations, together with the change observed when the standard procedure using pyridine - barbituric acid reagent (readings at 578 nm) is applied to the same cyanide solution. Effect of Temperature Because of the volatility of cyanogen chloride, an attempt was made to maintain the solution in an ice-bath during the colour formation.3.4.No difference in absorbance was obtained whether the reaction was carried out at 0 "C or at room temperature. No increase in absorbance or stability was observed when the solution was warmed to 40 "C during the 20 rnin after the reagents were added."' Effect of Light Absorbance measurements were carried out at various times on a 40 pg ml-* cyanide sample solution kept in the dark. The stability with time, as well as the absorbance values, were of the same order of magnitude of those observed when working in the light, even when the solution was kept in the cell under direct irradiation at 490 nm. Effect of Reagent Concentration The reaction was carried out using amsonic acid solutions of concentrations of 2 and 5%.The more concentrated solution produced higher absorbances, but the stability with time was not enhanced. Table 2 lists absorbance values measured after 1 rnin at 490 nm and after 1 h at 545 nm using the two concentrations of amsonic acid solution. Stability of the Reagent Solution The stability of pyridine - amsonic acid solution was studied for a period of 17 d. For this purpose, the proposed procedure was applied to four cyanide solutions at different concentra- tions, and the colour development at each concentration was obtained with freshly prepared reagent solution and old solutions prepared 7 and 17 d previously, The absorbance readings at 490 nm are shown in Fig. 4.1162 ANALYST, SEPTEhIRER 19x4.VOL. 100 1.6 1.4 1.2 Q, m 1.0 e a 0" 0.8 0.6 0.4 0.2 / L 1 I 0 0.2 0.4 CN- concentration pg ml- Fig. 4. Absorbance at 490 nm ~'C'I.SII.Y cyanide concentration. obtained with A. freshly prepared reagent solution: R. \vith 7-d old reagent solution: and C . with 17-d old reagent solution A similar study was carried out by using the standard method with pyridine - barbituric acid. but in spite of pub- lished assertions about reagent stability it has proved inipos- sible to prepare solutions that are stable over a period of more than 24 h. Effect of Pyridine Concentration Pyridine must be present in excess for the chromogenic reaction to take place. The experiments carried out show that the recommended amount of pyridine in the proposed procedure gives the more reproducible results.Absorbance Variation with Cyanide Concentration. Agree- ment with Beer's Law The results obey Beer's law over a concentration range of cya>ide of between 0.006 and 0.6 pg ml-1 when working with both of the proposed methods (readings were taken at 490 or 545 nm. as described above) or with the pvridine - barbituric acid standard method. The linear regression equations are as follows: Proposed method at 490 nm: A = 0.010 + 3.86~; r=0.9999 Proposed method at 545 nm: A = 0.006 + 2 . 2 5 ~ : r=0.9997 Pyridine - barbitaric acid standard method at 578 nm: A = 0.010 + 4 . 4 2 ~ ; r=0.9999 where c is expressed in mg 1- I . In Table 3, several values related to the quantification limit are collected for the proposed method and for the standard method, which is given for comparison.Values in the first row are the log (molar absorptivities), expressed per mole of cyanide; in the second row. the sensitivities, in pg cm-2, according to Sandell are recorded: in the third row, the values are the detection limits for each method calculated from the blank readings and their standard deviations21 ; and in the fourth row, the quantification limits are given as three times the detection limits. If the samples of cyanide-containing water have been previously distilled in the presence of strong acid in order to eliminate interferences.14 the cyanides will be concentrated; in the last row of Table 3 the quantification limits are given, evaluated for a concentration factor of 2.5. When working with the method using the pyridine - barbituric acid reagent, the absorbance of the violet com- pound (maximum initially at 583 nm) suffers a rapid variation with time.qualitatively similar to that observed with amsonic acid. i.e.. in the first minute the absorbance increases to a maximum and then decreases. This observation is mentioned in the last edition of the APHA Standard Methods14 and a rigid standardisation of thc measurement time is recomnien- ded. Precision of the Method The precision was calculated separately from readings at 490 nm after 1 min and at 535 nm after 1 h from preparation of the co 1 o u re d so 1 u t i o n . The re 1 at i ve stand a rd de v i a t i CJ n o b t a i t i ed from results on ten test solutions containing 0.185 pg m - I of cyanide is 0.35% at 390 nm and 1.81% at 545 nm.Recovery of Cyanide from Alkaline Solutions of Cyanide Using the Proposed Method In the determination of "total cyanide'' (free and complexed cyanide) distillation or micro-diffusion techniques are com- monly used. in which hydrogen cyanide is absorbed i n an alkaline medium. As this is ;I well established step. common to all spectrophotometric methods for the determination of cyanides. the recovery of cyanide has only been studied in this work for solutions in which the sample has been directly dissolved in the same amount of sodium hvdroxide solution a s is used as an absorbant in the distillation and micro-diffusion techniques, as applied to cyanide determination in water samples. Because, in the proposed procedure, a large excess of borate buffer is added.the calibration graph obtained in this instance is nearly the same as was described for pure cyanide samples and the recovery is quantitative. Interferences Although cyanide determination in waste water samples implies a previous separation step in order to remove the interfering ions that are commonly present in such samples, the proposed method using amsonic acid has been studied with regard to the direct effect of sonfe ions on the absorbance. The results obtained for a 0.19 pg m-I cyanide solution in the presence of these ions are given in Table 4. Several ions (particularly sulphide and thiocyanate) interfere in the proposed method in ii manner analogous to that in which they interfere in all other colorimetric methods for cyanides based on the Kiinig reaction.Conclusions The method proposed for the determination of cyanide in water and waste water, using pyridine - amsonic acid as a reagent has some advantages over the pyridine - barbituric acid method. Firstly, it uses a reagent that is much more stable in solution. cheaper and more easily available. More- over, the method can be applied to a wide range of cyanide concentrations: within the range 0.1-0.010 pg ml-1 absor- bance readings are best taken at 490 nm 2 inin after the coloured solution is prepared, whereas within the range 0.10-0.60 pg mi-' it is better to perform the absorbance measurements at 545 nm after a 1-h delay. The detection limit is of about the same order of magnitude as that of the standard method using pyridine - barbituric acid reagent.Although the colour stability and the sensitivity of the proposed method are slightly lower than those of the standard method, the aforementioned advantages justify the use of the new reagent. The authors thank CIRIT (Comissici Interdepartamental de Recerca i Innovacio Tecnologica from the Generalitat de Catalunya) for the help received in support of this work.ANALYST. SEPTEMBER lW4, VOL. 109 1163 1 . 3 . 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 3 -. References Kiinig. W.. J . Prukt. C'hem.. 19114. 69. 105. Kiinig, W., Z . Angew. C'hem., 1905. 69. 115. Aldridge, W. N.. Analyst, 1944. 69, 262. Aldridge. W. N . , Anu!\'.sf, 1945. 70. 474. Bark. L. S . . and Higson. H. G.. Tulunta. 1964, 11. 471. Epstein, J . , Anal. C'hem.. 1947. 19, 272. Asmus. E.. and Garschagen. H . . Fr~senius 2. A n d . C'hem.. 1953, 138. 414. Murthy, G . V. L. N . . and Viswanathan, T. S . , Anal. Chim. Acta, 1961. 25. 293. Bark, L. S . , and Higson. H. G.. Tulanra, 1964. 11, 621. Bark. L. S . , and Higson, H. G.. A n u l ~ , 1963, 88, 751. Rodier, J . . "L'Analyse de I'Eau," Fifth Edition, Dunod, Paris. 1975, p. 251. Serfass, E. J . , and Freeman. R. B., Plating, 1952, 267. Ludzack. F. J . , Moore, W. A.. and Ruchhoft, C. C.. A n d . Chern.. 1954, 26, 1784. APHA. AWWA and WPCF. "Standard Methods for the Examination of Water and Wastewater," Fifteenth Edition, American Public Health Association. Washington. DC. 19x1. 15. 16. 17. 18. 19. 20. 21. Botto, R. I . , Karchrner, J. H., and Eastwood, M. W., Anal. Chem.. 1981, 53, 2375. Snell. F. D., and Ettre, L. S . , Edirors. "Encyclopedia of Industrial Chemical Analysih," Volume 17. Intersciencc. New York, 1973. p. 432. Fuentes-Duchemin, J., and Casassas. E., Anal. Chim. Acfu. 1969.44, 462. Russel, F. R.. and Wilkinson, N . T . , Anulyst, 1959, 84, 751. Lambert, J . L., Ramasarny, J . , and Paukstelis, J . V., Anal. Chem., 1975, 47, 916. Casassas, E . , and Fuentes-Duchernin, J . , Inf. Quim. Anal., 1971, 25, 159. Massart, D. L., Dijkstra, A . , and Kaufrnan. L., "Evaluation and Optimization of Laboratory Methods and Analytical Procedures," Elsevier, Amsterdam, 1978. Puper A4129 Received January 23rd, 1984 Accepted March 30th, 1984