Analyst, June, 1979, Vol. 104, pp. 525-530 525 Spectrophotometric Determination of Trace Amounts of Free Cyanide in Prussian Blue G. J. Willekens and A. Van Den Bulcke Instituut VOOY Hygiene en Epidemiologie, Departement Farmatoxicologie, A fdeling Farmampee- en Standaardenonderzoek, Juliette Wytsmanstraat 14, 1050 Brussels, Belgium Trace amounts of free cyanide in Prussian blue are hydrolysed into hydro- cyanic acid. The latter is captured by a lithium picrate solution contained in a test-tube, which is placed in the reaction vessel. The colour change due to the resulting lithium isopurpurate is measured spectrophotometrically at 500 nm. This method can detect cyanide down to a level of 2.5 pg in 100 mg of Prussian blue and is accurate and reproducible. Keywords ; Cyanide determination ; insoluble and colloidal Prussian blue ; spectrop hotowetry Prussian blue is used as an antidote for metal intoxications such as those due to thalliumJ1s2 radioactive strontium3 and caesium.475 As the Prussian blue is given in daily doses of up to 20g6-* and generally over a 14-d period, the purity of Prussian blue is most important.Because of the extreme toxicity of cyanides their detection and determination in small amounts are of great importance. A great variety of methods are available for the identifica- tion of cyanides. Of these, spectrophotometry is the most convenient for determining small amounts. Prussian blue exists in either an insoluble or a colloidal form, the colloidal form being the so-called soluble formJg which is the most active antidote for thallium poisoning6910p11 owing to its small particle size (down to 0.03 pm).Because of the small particle size extraction of adulterants by solvents followed by filtration or centrifugation is impracticable. The free cyanides can, however, be converted into hydrocyanic acid, which diffuses as a gas to react with a suitable colorimetric reagent, either in solution or absorbed on a filter-paper. Distillation by heating an acidic mixture of the sample to give the hydrocyanic acid is not suitable for routine quantitative determinations as it is time consuming and requires complex apparatus. It may also give inaccurate results when small amounts of alkali metal hexacyanoferrate(I1) and hexacyanoferrate(II1) are present in the sample, as they break down into hydrocyanic acid.Hence, we focused our research on the detection and quantita- tive determination of the hydrocyanic acid evolved at 20 "C. From preliminary tests it appeared that a compromise between pH and time had to be found, which would result in a slow evolution of hydrocyanic acid by hydrolysis, with no addition of proton donors other than water. The choice of the reagent for detection was limited. Although procedures based on the Konig synthesis12-16 are generally considered satisfactory for determining small amounts of cyanide, the colour developed is not sufficiently stable. Alkaline picrates combine with hydrogen cyanide to give isopurpurates,l7 which are specific and stable colour reactions and the possibilities of using these reactions were examined.NO2 Picric acid Lithiu rn isopurpurate Attempts to use squares of filter-paper soaked with the reagent to react with the hydro- cyanic acid liberated from the test solution, by heating with sodium hydrogen carbonate526 WILLEKENS AND VAN DEN BULCI'E SPECTROPHOTOMETRIC AnaZyst, "ol. 104 and a slight excess of acid,l8 were unsuccessful as there was no significant change in the colour intensity with increase in the amount of hydrocyanic acid. Other tests such as the copper acetate - benzicline test,lS the test depending on the fonna- tion of Prussian blue,20 the copper sulphide test21 and the guaya resin test,22 although sensitive, proved to be unsatisfactory for the satme reason. The method finally developed used a simple gas-testing device (Fig.1). A small flat-bottomed test-tube of approximately 2-ml capacity containing lithium picrate solution was placed in a glass-stoppered flask, which contained the sample of Prussian blue to which a definite amount of water had been added. The flask was stoppered immediately, protected from the light and allowed to stand over- night at 20 O C , after which the absorbance of the solution within the test-tube was measured against a blank prepared under the same conditions. The procedure, described below, easily detected 2.5 pg of cyanide added to 100 mg of Prussian blue (Table I). Experimental Reagents Lithium picrate solution. Dissolve 0.25 g of lithium carbonate and 0.5 g of picric acid in 80 ml of boiling distilled water. Cyanide standard solution. Dissolve 100 mg of potassium cyanide in 1000 ml of distilled water and dilute 5 ml of this solution to 100 ml.Prussian blue test sample (colloidal and insolu:ble). Obtained from Hopkin & Williams, Chadwell Heath, Essex. Allow to cool aind dilute to 100 ml. This solution is to be used within 12 h. Standard Additions Procedure Weigh 100-mg samples of the Prussian blue under test into each of five glass-stoppered, wide-necked, 200-ml flasks (Fig. 1). Add to four of them, respectively, 0.5, 1, 2 and 3 ml of the cyanide standard solution and sufficient distilled water to produce 5ml. To the fifth flask, used as a blank, add 5 ml of water. S'hake the flasks gently. Accurately pipette 1 ml of the lithium picrate solution into five small, flat-bottomed test-tubes and introduce each of them into one of the conical flasks and stopper the flasks immediately.Allow the flasks to stand overnight, protected from light, at 20 "C. Lithium picra , Prussian blue and cyanide ,te Fig. 1. Gas-testing device. Add to the contents of the flat-bottomed test-tubes 2 ml of the lithium picrate solution, To construct the calibration graph, plot the absorbances veysus the corresponding amounts mix and measure the absorbance in 1-cm cells against the blank at 500 nm. of cyanide added to the Prussian blue solutions and draw the regression line. Results and Discussion Neither the insoluble nor the colloidal (soluble) form of Prussian blue can be totally removed by filtration. Filtration of the colloidal solution, even through a micro-crystalline filter-paper (Whatman No.42) yielded a pale blue filtrate. Therefore, distillation of theJune, 1979 DETERMINATION OF TRACE AMOUNTS OF FREE CYANIDE IN PRUSSIAN BLUE 527 hydrocyanic acid was needed and although this distillation required a 24-h period the effective work for one determination was only 30min. The reaction was carried out at 20 "C with protection from the light. The recovery of different amounts of cyanide added to 100 mg of insoluble and colloidal Prussian blue was evaluated spectrophotometrically at 500 nm (Fig. 2). 240 3 20 400 4 80 5 60 Wave le ngt h/nm Fig. 2. Ultraviolet absorption spectra in water : solid line, lithium picrate; and broken line, lithium isopurpurate. Plots of the absorption values veysus the corresponding amounts of cyanide gave straight lines with a significant correlation coefficient of 0.9999 for both the insoluble and colloidal forms.Only the graph for insoluble Prussian blue passed through the origin, which means that this Prussian blue released no cyanide at 20 "C. The colloidal form lost approximately 0.7 pg per 100 mg of Prussian blue. This cyanide release was calculated from the total absorbance measured for each mixture and the deviation of the calibration graph on the y-axis. Subtraction of the related values yielded the cyanide release (Table I). The confidence limits (95% confidence interval) of the mean cyanide recoveries are given by x (o/dT x t*), where o/d< is the standard error, o the standard deviation, rz the number of measurements and t* is 1.96. The recovery errors were between -3.4% and +0.9%.These values are listed in Table I and indicate the acceptable accuracy of the spectrophotometric method using lithium picrate as a reagent. The results are listed in Table I. The calibration graphs (Fig. 3) at different temperatures show good linearity except for the colloidal form at 50 "C. The recovery errors were between -4.4% and +4% at 30 "C and between -4.6% and +11.6% at 40 "C. Thus, change of temperature had no significant influence on the recovery of cyanide added to 100mg of Prussian blue. The increase in temperature , however, increased exponentially the cyanide release of Prussian blue itself. The insoluble form released no cyanide at 20 "C when no cyanide was added. At 30 "C the release was 1.36 pg, at 40 "C 3.35 pg and at 50 "C 9.63 pg.For the colloidal form this cyanide release increased to a considerable amount , which was three times higher than for the insoluble form. Values of 0.61, 2.89, 8.67 and 22.56 pg were found at the different temperatures, respectively. Addition of cyanide, on the other hand, also increased this cyanide release and the greater the cyanide addition the greater was the cyanide release of Prussian blue itself, especially The influence of temperature on the recovery of the added cyanide was determined.528 WILLEKENS AND VAN DEN BULCKE: SPECTROPHOTOMETRIC Analyst, VoZ. 104 TABLE. I RECOVERY OF CYANIDE ADDED TO PRUSSIAN BLUE AND INFLUENCE OF TEMPERATURE ON THAT RECOVERY AND THE RELEASE OF CYANIDE BY PRUSSIAN BLUE Form of TemDeraturel Prussian blue Cvanide "C 20* 30t 40t Colloidal Insoluble Colloidal Insoluble Colloidal Insoluble Colloidal aidedlpg 0.00 2.50 5.00 10.00 15.00 0.00 2.50 5.00 10.00 15.00 0.00 2.50 5.00 10.00 15.00 0.00 2.50 5.00 10.00 15.00 0.00 2.50 5.00 10.00 15.00 0.00 2.50 5.00 10.00 15.00 0.00 2.50 5.00 10.00 15.00 0.00 2.50 5.00 10.00 15.00 Cyanide recovered/ P*.g 0.00 2.49 4.83 10.02 15.05 0.00 2.43 4.97 10.09 14.96 0.05 2.60 4.86 9.93 15.08 0.09 2.39 4.88 10.22 14.91 0.12 2.59 4.94 9.54 15.31 -0.18: 2.79 4.77 10.24 14.86 - 0.093 2.54 5.63 8.89 15.52 2.92 8.48 10.58 13.38 - 2.863 Cyanide Confidence limits released by Recoverv.& Prussian blue/ % - 99.8 96.6 100.2 100.3 97.3 99.4 100.9 99.7 104.0 98.8 99.0 100.5 95.6 97.6 102.2 99.4 103.6 98.8 95.4 102.1 111.6 95.4 102.4 99.0 101.6 112.6 88.9 103.4 116.8 169.6 105.8 89.2 - - - - - - - .I, Maximum Minimum Pt4 0.00 0.09 0.06 0.00 0.61 0.63 0.69 0.78 0.77 1.36 1.29 1.29 1.32 1.34 2.89 2.97 3.05 3.17 3.36 3.35 3.58 3.79 4.23 4.76 8.67 8.69 8.75 8.74 9.77 9.63 9.75 9.89 10.04 10.35 22.56 23.38 24.12 24.40 24.73 - * Each result is the mean of ten determinations. t Each result is the mean of three determinations.No confidence limits were calculated. These results were determined by calculation of the regression lines. for colloidal Prussian blue. The cyanide release of the insoluble form increased to a lesser extent. The influence of temperature, light, reaction time, acid and mass of sample on the release of cyanide from Prussian blue in the absence of cyanide was also investigated. The results are listed in Tables I1 and 111.Under the influence of light, and especially sunlight, there was an uncontrollable cyanide release. After an exposure of 24 h of the reaction vessel to daylight we found a cyanide release of 35 pg pler 100 mg of colloidal Prussian blue. The insoluble form released 3.85 pg per 100 mg. Even in the dark at 20 "C, 100 mg of colloidal Prussian blue lost 0.72 pg of cyanide. Under the influence of 0.06 N hydrochloric acid. at 37 "C, 100 mg of inSoluble Prussian blue released no cyanide. The colloidal form, however, released 1.90 pg. This value is notJzwze, 1979 DETERMINATION OF TRACE AMOUNTS OF FREE CYANIDE IN PRUSSIAN BLUE 529 0.800 0.700 r- 0.600 a, 0.500 0 2 2 a & 0.400 0.300 0.200 0.100 0 5 10 15 20 Cyanide concentration/pg Fig.3. Calibration graphs for the detec- tion of cyanide added to a 100mg per 5 ml of Prussian blue solution at different temperatures : 0, insoluble state ; and A, colloidal state. similar to that reported by Kamerbeek,' who incubated colloidal Prussian blue for 4 h with gastric juice (pH 2) and 0.1 N hydrochloric acid in the presence of oxygen and could not detect any release of cyanide. The mass of Prussian blue did not influence the cyanide release. Conclusion The spectrophotometric determination of cyanide by use of a lithium picrate solution is a simple method, which is based on the distillation of hydrocyanic acid and is used in all instances where interfering substances may hamper direct detection of cyanide. The proposed method is a standard additions procedure and permits the simple detection of cyanide down to a level of 2.5 pg.TABLE I1 CYANIDE RELEASED BY COLLOIDAL PRUSSIAN BLUE IN WATER UNDER DIFFERENT CONDITIONS The results given are the amounts of cyanide released, in micrograms; each value is the mean of three determinations. Mass of Prussian bluelmg 7 Temperaturel'C Acid Time/h Daylight 50 20 - 24 + 8.0 0.3 24 20 1.4 30 - 24 1.4 37 0.06 N HCI 2 4.4 40 - 24 50 - 24 - 19.1 - - - - - 100 150 200 250 16.2 18.2 29.1 35.4 0.7 1.4 2.2 2.8 2.9 4.6 5.7 7.8 1.9 1.9 2.0 3.0 8.4 12.1 17.1 30.0 23.6 30.8 36.3 47.2530 WILLEKENS AND VAN DEN BULCKE TABLE I11 CYANIDE RELEASED BY INSOLUB:LE PRUSSIAN BLUE IN WATER UNDER DIFFERENT CONDITIONS The results given are the amounts of cyanide released, in micrograms; each value is the mean of three determinations. Mass of Prussian blue/mg I A Temperaturel’C Acid Time/h Daylight 50 100 150 200 20 - 24 + 2.6 3.9 4.3 4.2 20 - 24 - 0.05 0.09 0.05 0.09 24 - 0.6 1.0 1.4 1.5 30 37 0.06 N HCl 2 40 - 24 - 1.8 3.1 3.9 5.4 50 - 24 - 5.9 10.8 14.4 17.4 - - - - - - 7 250 4.3 0.07 2.0 1.4 7.8 18.2 The authors thank Prof.Dr. Apr. van Peteghem (Rijksuniversiteit Gent) for literature communications and Mr. Legrand (I.H.E., Brussels) for statistical analysis. They are obliged to Mr. J. W. Lightbown of the National Institute for Biological Standards and Control, Hampstead, London, for assistance with the manuscript. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. References Dvorak, P., Naunyn-Schmiedebergs Arch. Ex*. Ptzth.Pharmak., 1971, 269, 48. Stevens, W., Van Peteghem, C., Heyndrickx, A, and Barbier, F., Int. J . Clin. Phavmac. Thev. Toxic., 1974, 10, 1. Borisov, V. P., I l k , L. A., Kendysh, M. M., Shomorokhova, T. N., Mikhailovich, S. M., and Seletskaya, L. I., Hlth Phys. Probl. Int. Contaiw. Proc. IRPA (Int. Rudiat. Pvot. Ass.), 2nd Eur. Congr. Radiat. Prot., 1972; Chem. Abstr., 1973, 79, 143311. Havlicek, F., Int. J. Appl. Radiat. Isot., 1968, 19, 487. Madshus, K., and Stromme, A., 2. Naturj., 1968, 236, 391. Kamerbeek, H. H., Rauws, A. G., Ten Ham, M., and Van Heyst, A. N. P., Acta Med. Scand., 1971, Kamerbeek, H. H., “Therapeutic Problems in Thallium Poisoning,” Proefschrift Utrecht, Gianotten, Rauws, A. G., Kamerbeek, H. H., and Ten Ham, Id., “Verslag 130e Wetenschappelijke Vergadering,” Pascal, P., “Trait6 de Chimie MinCrale,” Masson, Paris, 1933, p. 862. Dvorak, P., 2. Ges. Exp. Med., 1960, 151, 89. Dvorak, P., Arzneimittel-Forsch., 1970, 20, 1886. Konig, W., J. Prakt. Chem., 1904, 69, 105. Konig, W., 2. Angew. Chem., 1905, 70, 115. Aldridge, W. N., Analyst, 1944, 69, 262. Aldridge, W. N., Analyst, 1945, 70, 474. Epstein, J., Analyt. Chem., 1947, 19, 272. Auterhoff, H., and Boehme, K., Arch. Pharm., Bed., 1968, 301, 793. Schwapowalenko, A. M., Chem. ZentBl., 1930, 11, 588. Moir, J., Chem. News, Lond., 1910, 102, 17. Vichoever, A., and Johns, C. O., J . Am. Chem. SOG., 1915, 37, 601. Barnebey, 0. L., J . Am. Chem. Soc., 1914, 36, 10!)2. Prodanov, P., Izv. Khim. Inst. Bulg. Akad. Nauk, 1964, 10, 277. 189, 321. Tilburg, 1974. R.I.V., Utrecht, Bilthoven, 1973. Received September 12th, 1978 Accepted December 28th, 1978