- Email: [email protected]
Kinetic fluorimetric determination of cyanide by means of its catalytic effect on the aerial oxidation of pyridoxal-5-phosphate oxalyldihydrazone
Talanra, Vol. 31, No. lOA, pp. 783-788, 1984 Printed in Great Britain. All rights reserved
Copyright
0
0039-9140/84 $3.00 + 0.00 1984 Pergamon Press Ltd
KINETIC FLUORIMETRIC DETERMINATION OF CYANIDE BY MEANS OF ITS CATALYTIC EFFECT ON THE AERIAL OXIDATION OF PYRIDOXAL-5-PHOSPHATE OXALYLDIHYDRAZONE S. RUBIO, A. GOMEZ-HENSand M. VALCARCEL Department of Analytical Chemistry, Faculty of Sciences, University of Cbrdoba, Cerdoba, Spain (Received
9 January
1984. Revised 26 April 1984. Accepted
7 May
1984)
Summary-A highly specific determination of traces of cyanide is described, based on the catalytic effect of cyanide on the oxidation of pyridoxal-5-phosphate oxalyldihydrazone by dissolved oxygen. The reaction is monitored by measuring the fluorescence of the oxidation product (I._ 350, I_,,, 420 nm), and allows determination of 3-180 ng/ml cyanide concentrations with relative standard deviations of 0.4-0.8%. The effect of 56 foreign ions has been investigated: 45 of them do not interfere, and only one, Hg(II), interferes when present at the same level as the cyanide. Its application to the determination of total cyanide in industrial and synthetic samples, without preliminary distillation, is reported.
It is well known that cyanide catalyses the oxidation of some aromatic aldehydes to acids, through formation of cyanohydrins,’ and this has been utilized for spectrophotometric and fluorimetric determination of cyanide.*-5 The reaction between the cyanohydrin of p-nitrobenzaldehyde and o-dinitrobenzene to give a highly-coloured blue compound, the doublycharged anion of o-nitrophenylhydroxylamine, has been used by Guilbault and Kramer.’ The rate of reduction of the o-dinitrobenzene was followed spectrophotometrically at 560 nm. Cyanide could be determined in the range 45450 ng/ml, and the method was highly specific. The fluorimetric methods for determination of cyanide have been reviewed and critically studied.6 The reaction of pyridoxal with oxygen to give 4-pyridoxolactone, a highly fluorescent compound, is catalysed by cyanide. 3-5Factors affecting the reaction rate have been studied by Takanashi and Tamura.’ The reagents are mixed in phosphate medium at pH 7.4 and heated at 50” for 60 min. The fluorescence is measured at pH 10 (achieved by adding sodium carbonate). The calibration graph is linear from 4 to 180 ng/ml. This reaction is regarded as the most sensitive fluorimetric method for cyanide.* Here, the analytical properties of a new compound, pyridoxal-5-phosphate oxalyldihydrazone (PPOH), are described for the first time, and its aerial oxidation is used for the kinetic fluorimetric determination of traces of cyanide by means of their catalytic effect on the oxidation reaction. Under the conditions proposed, the reaction is rapid, measurements being taken less than 7 min after the start. The method is very selective, no preliminary distillation is needed and cyanide concentrations as low as 3 ng/ml can be determined. 783
EXPERIMENTAL
Synthesis
of the reagent
Pyridoxal-5-phosphate (0.6 g) dissolved in the minimum of hot water-ethanol (3:2 v/v) solution was mixed with 0.13 g of oxalylhydrazide dissolved in hot water-ethanol mixture (1: 1v/v). The mixture was refluxed for 30 min, then let cool to room temperature. The yellow product was filtered off and washed with ethanol. Calculated for C,sH,,O,,N,P,, C 37.50x, H 3.82%; found C 37.5x, H 4.1%. Reagents Pyridoxal-5-phosphate oxalyldihydrazone (PPOH) solution 7 x IO-W, in methanol. Several drops of O.OlM
sodium hydroxide (enough to give an apparent pH of _ 7) were added to dissolve the reagent. This solution is stable for at least one month. Standard cyanide solution. Potassium cyanide (2.5 g, equivalent to I g of cyanide) was dissolved in 1 htre of 0.05M sodium hydroxide and standardized argentometrically.9 Ammonia-ammonium chloride 6M buffer solution, pH 9.8. Chloramine-T solution and pyridine-barbituric acid reugent.‘O
Reagent-grade chemicals and pure solvents were used. All anions tested were added in the form of their sodium or potassium salts. Apparatus Spectrofluorimeter. A Perkin-Elmer fluorescence spectrophotometer, model MPF-43A, fitted with a device for kinetic measurement and with l-cm quartz cells was used. The cell compartment of the spectrofluorimeter was kept at constant temperature by circulating water. The sensitivity was set at 3, and the excitation and emission slits were set to give 6-nm spectral band-pass. A set of fluorescent polymer samples was used daily to adjust the spectrofluorimeter to compensate for changes in source intensity. Cyanide distillation apparatus.‘0 Air-sampling equipment. A constant-flow
pump, DuPont model S2500; impingers, volume 30 ml; cellulose ester filters, 37mm in diameter and 0.8pm thick.
S. RUBIO ef al.
784
Procedure To lo-ml standard flasks add 0.5 ml of 7 x 10m4M PPOH solution, 5 ml of 6M ammonia-ammonium chloride buffer (pH 9.8), 2.5 ml of 4M sodium chloride and appropriate volumes of cyanide sample to give a final concentration of cyanide between 3 and 180 ng/ml. Make up to volume with distilled water, mix and after 1 min transfer a portion to a l-cm quartz cell kept at 62 t_ 0.1’. Wait 2 min before starting to record the fluorescence intensity (A,, 350, I,, 420 nm) as a function of time. Run a blank with no cyanide present. Calculate the reaction rate from the difference in the
slopes of the fluorescence-time plots for the sample and the blank.
100
90
5 5
Oxidation of PPOH and catalytic action of cyanide
The excitation and emission spectra of PPOH show a hypsochromic change in basic (&,, 350 nm, &,,, 420 nm) and acid (a,, 320 nm, A,,,, 395 nm) media in presence of such oxidizing agents as KIOj, H202, K,S,Os and NaIO,. Solutions of PPOH in ammonia-ammonium chloride buffer show identical
n2'
70-
E = 608 t _ bW SW_ ">
RESULTS AND DISCUSSION
At room temperature, PPOH is highly soluble in water or in sodium hydroxide solution, moderately soluble in dimethylformamide or methanol and slightly soluble in ethanol. Aqueous solutions are unstable whatever their pH. A 3.5 x 10d5M solution in methanol-water (I :24 v/v) has an absorption maximum at 305 nm and a very small absorption band at 360 nm in acidic and neutral media. In alkaline medium it shows a broad band with an absorption maximum at 305 nm. PPOH solutions are fluorescent (&, 365 nm, i,, 510 nm in acid medium and i,, 400 nm, i,, 510 nm in alkaline medium). The stability of the PPOH solutions is critically dependent on the medium. The reagent is stable in organic solvents such as methanol. Aqueous PPOH solutions are unstable owing to the autoxidation of the reagent by dissolved oxygen. This difference in behaviour occurs because there is a hydrolysis step prior to the oxidation, and this is favoured in an aqueous medium. For this reason, neutral aqueous solutions are more stable than acid or basic ones, because H+ and OH-- ions catalyse the hydrolysis. This autoxidation reaction occurs at pH values very close to that used for the determination of cyanide, which is why blanks are run. for the dissociation in The pK values methanol-water (1:24 v/v) were found”,‘2 to be 3.9, 8.2 and 10.4. The colour reactions of the reagent with 62 ions at various pH values were investigated; it reacts with Mg(II), Co(U), Ni(II), Mn(II), Fe(I1) and Fe(II1). Only the red Mg(IItPPOH complex formed at pH 4.2 and 50” (acetic acid-sodium acetate buffer) is of any photometric analytical interest. The fluorescence reactions are not important. Only V(V) and Al(II1) show fluorescence which is different from that of the reagent, but the sensitivity is low.
n2
,. : eo-
z a,
Analytical properties of the reagent
c c
3020-
rr
lo-A
I
1
1'
\
1
I
xx)
350
I 400
_
I \ 450
X(nm) Fig. 1. Catalytic effect of cyanide on the oxidation of PPOH by dissolved oxygen. Excitation and emission spectra. 1,l’: PPOH; 2,2’: PPOH + CN-. [PPOH] = 8 x lo-‘M; [CN-] = 0.4 pg/ml; [NH,Cl-NH,] = 0.4M. Graphs were recorded after a reaction time of I hr. Temperature 50°C.
fluorescence maxima (J.,, 350 nm, A,,,, 420 nm) to those obtained in the presence of oxidants but the formation of the fluorescent product is so slow that it takes several days for a detectable amount to be produced. In the presence of traces of cyanide, the oxidation rate of PPOH by dissolved oxygen is accelerated. This effect is shown in Fig. 1. The catalytic nature of the cyanide reaction is revealed by the fact that it does not take place in an inert atmosphere. When oxidants are utilized, the reagent is oxidized immediately, and cyanide then does not have a catalytic effect. Several tests have been made to characterize the oxidation product of PPOH and establish the mechanism of the reaction, with the following results (a) The characteristic absorption peak at 305 nm of the Schiffs base formed between pyridoxal-5phosphate and oxalylhydrazide disappears in the course of the reaction. This seems to be due to hydrolytic breaking of the azomethine group. (6) The fluorescence maxima corresponding to the reagent (L,, 400 nm, A,,,, 510 nm) disappear like the absorption maximum, but an intermediate product (J.,, 320 nm, &,,, 370 nm) can be observed before the formation of the final product (A,, 350nm, L,, 420 nm). This intermediate is in equilibrium with PPOH and subsequently with the final product, and its fluorescence characteristics are the same as those of the pyridoxal-5-phosphate.‘3 This seems to corroborate the supposition of hydrolytic breaking. (c) The fluorescence maxima of the oxidation product of PPOH are similar to those reported for a solution of pyridoxal-5-phosphate treated with hydrogen peroxide, which forms 4-pyndoxic acid.
From these observations it can be inferred that the likely oxidation mechanism consists of hydrolytic
Fluorimetric determination of cyanide
T
(‘Cl
785
PH
CPPOHI, IO’ M
Fig. 2. Effect of the temperature (A), pH (B) and PPOH concentration (C) on the reaction rate of the PPOHQCNsystem. breaking of the azomethine group and oxidation of the resulting compounds. Several attempts have been made to isolate the highly water-soluble fluorescent species from mixtures of the parent reagent and cyanide, but without success. Evaporation yielded a reddish brown liquid which would not crystallize. Addition of ethanol produced a yellow solid, the excitation and emission spectra of which were identical to those obtained in the analytical test, but this solid was water-soluble and non-extractable by water-immiscible organic solvents, and was contaminated by the reagent, so meaningful ultraviolet and infrared spectra could not be obtained. Attempts to obtain a pure product by thin-layer chromatography were unsuccessful. Injuence
of reaction variables on the oxidation
The temperature of the cell-holder was varied over the range 35-70” (Fig. 2A). The reaction rate increased linearly withtemperature increase between 35 and 60” but less steeply at higher temperatures. A temperature of 62” was selected for use. From Arrhenius plots of In k vs. reciprocal of the absolute temperature, the activation energy for the catalysed reaction was calculated to be 5.1 +_0.1 kcal/mole. The effect of pH on the reaction rate is given in Fig. 2B. The optimum pH is 9.7-10.0. The effect of the type of buffer (phosphate, carbonate, ammonia) was examined. The reaction develops only in the presence of ammonia-ammonium chloride buffer, and is linearly related to the buffer concentration up to at least 4M. A 3M concentration was chosen. The influence of the PPOH concentration was tested in the range 5 x 10-6-10-4M and is shown in Fig. 2C. A purely aqueous solution could not be used because of its instability, so a methanol solution was used. However, increasing the methanol content of the final aqueous methanol medium decreased the reaction rate, so its concentration in the solution measured was minimized. The rate of reaction rises with increasing ionic strength up to IM and then remains constant. The fluorescence intensity vs. time curves for different cyanide concentrations were recorded. The initial slopes indicate a first-order reaction with respect to cyanide. The various kinetic dependences on pH (Fig. 2B), reagent concentration (Fig. 2C) and buffer concen-
tration are summarized in Table 1, and the following equation is suggested for the oxidation of PPOH (3.5 x 10m5M) by dissolved oxygen at pH 9.85 in the presence of 3M ammonia buffer and cyanide as catalyst: d[PPOH],,/dt
= k[NH,Cl-NHJ3”[CN-1
in which [PPOH],, is the concentration of oxidized reagent and k is the conditional rate constant. This equation does not include the effect of the uncatalysed reaction. Calibration graphs
Three kinetic methods have been tested for the determination of cyanide: tangent ([CN-] 10-180 ng/ml), fixed-time ([CN-] 3-180 ng/ml) and variabletime ([CN-] 3-60 ng/ml).14 For the fixed-time method, measurements were made after 7 min. For the variable-time method, the inverse of the time necessary to obtain a relative fluorescence intensity of 30% was plotted against the cyanide concentration. The relative standard deviation for 50 ng/ml cyanide (P = 0.05, n = 11) is smaller for the initial-rate method (+0.4x) than the fixed-time (f0.8%) and the variable-time (+ 0.7%) methods. The initial-rate method is recommended when precision is the prime consideration and the fixed-time method when a low detection limit is the important factor. Interferences
The effect of 35 anions and 21 cations on the determination of cyanide was tested. The tangent method was used because a change in slope could be more clearly detected than a change in a single data point. The tolerance limits for interfering ions are
Table 1. Summary of kinetic data Dependence of V, on
Concentration range, M
W+l W+l” 1/W+]
4 5 X lo-“-7.9 X lo-” 7:9 x IO-“-2.5 x IO-“’ 2.5 x lo-‘“5.0 x IO-‘0 10-5-2.2 x 10-5 2.2 x 10~5-5.0x 10-S 5.0 X 10-5-10~4 2.0-4.0
[PPOH]“2 [PPOH]’ [PPOH]-3’2 [Buffer]3/*
S. RUBIO et al.
786 Table 2. Effect of various Tolerance
ions on the determination tangent method
of 50 ng/ml of cyanide
by the
ratio,
Ion added
ion to CN_____100
SCN-, F-, Cl-, Br-. S&,
SO:-,
ClOb,
S,O;-,
PO:-,
P,O;-,
Se%,
VO;, MOO:-,
I-, NO,, ClO;,
CO:-,
BrO;,
P,O;,,
SOS-,
IO;,
AsO;,
IO;,
AsO:-,
WO:-, B,O:-, acetate,
tartrate, citrate, K+, Ca2+>Sr2+, Ba2+> Zn*+, Cd*+, A13+, Cr3+, B?+,
Mg’+,
In3+,
Sb’+, Tl+, Pb*+, Ti4+, Zr4+ 75
C,O:-,
NO,
40
CrO:-,
S2-
20
Cu*+, Fe’+
10
Ni’+, Fe(CN)i-,
1
Ag+
less than
1
Hg2+
summarized in Table 2; 45 of the ions tested did not affect the determination of cyanide when present in lOO-fold ratio to it. Silver interferes when present at the same level as cyanide, but mercury(I1) interferes at even lower concentrations. Hence, the method is highly specific. Applications
The determination of microamounts of cyanide in industrial effluents is of interest owing to its toxicity. The pyridine-barbituric acid calorimetric method recommended” has a detection limit of 4 ng/ml for cyanide. To determine cyanide in water samples by this method, it is necessary first to separate it from interfering substances by distillation of hydrocyanic acid from the acidified sample. It is well known,’ and has been verified by us, that when the equipment and conditions described” for the distillation are used, the cyanide is not quantitatively recovered at levels below 1 pg/ml in the original sample. The reasons for this Table
3. Determination (Without
Sample
of cyanide
Cyanide preliminary
have been investigatedI and it was found that the hydrocyanic acid is completely distilled but not totally absorbed in the collection solution. The distillation also takes about 90 min. Because of its high selectivity, we have used the new method for determination of cyanide in several samples without preliminary distillation. Three types of sample were investigated. Water
in water
k + + f
0.4 pg/ml 0.3 pgg/ml 0.3 ng/ml 0.01 mg/m’
Standard method* 8.5 5.3 1.3 0.10
k + + *
air from
and air from an electroolatine
12.5 + 0.4 ng/ml 7.2 4.1 2.5 0.10
and
an
electroplating
factory.
Several samples of water and air were analysed by both the fluorimetric and the standard calorimetric method, with and without distillation. As the true cyanide concentration in the samples was unknown, the distillation was included, on the assumption that the determination could be taken as free of interferences if the results obtained with and without distillation, agreed. The air sample (90 litres) was aspirated at 1.5 l./min through three impingers connected in series, each containing 10 ml of 0.1 M sodium hydroxide; the solutions were mixed for
0.3 pg/ml 0.5 pg/ml 0.6 ng/ml 0.02 mg/m3
factorv
Cyanide, pglml (With preliminary distillation)
distillation)
Fluorimetric method*
1 2 3 4 5
Fe(CN)i-
co2+
5
-
Fluorimetric method* ‘7.1 + 0.2 4.2 + 0.2
*Average and standard deviation of five separate determinations. Sample: 1. Water intake to the baths. 2. Water from the first washing of the silver-plated pieces. 3. Water from the second washing of the silver-plated pieces. 4. Waste-water from a plating factory with zinc-plating, silver-plating baths. 5. Air from the workroom.
Standard method* 7.2 + 0.1 4.0 * 0.3 -
and chromium-plating
Fluorimetric determination Table 4. Determination
of cyanide in synthetic platingbaths Cyanide content, pgglml
Plating-bath 1 2 3 4 5 6
I
Added*
Found?
4.1 5.8 4.4 4.1 3.3 4.8 3.3
1.6+0.1 5.8 + 0.2 4.3 f 0.1 3.9 * 0.2 3.2kO.l 5.0 + 0.1 3.4 kO.1
*Initial concentration diluted lO,OOO-fold with potable water. tAverage and standard deviation of four separate determinations. Sample: 1. Brass-plating bath: CuCN (3%), Zn(CN), (0.9%), NaCN (4.9%), Na,CO, (1.5%), NH, (0.015%). 2. Brass-plating bath: CuCN (2.025%), Zn(CN), (5.0%), NaCN (3%), Na,CO, (2.25%). 3. Brass-plating bath: CuCN (2.6%), Zn(CN), (1. I%), NaCN (4.5%), Na,CO, (1.5%). 4. Brass-plating bath: CuCN (3%) Zn(CN), (2.6%), NaCN (3.8%), Na,CO, (1.5%). 5. Copper-plating bath: CuCN (2.6%) NaCN (3.4%) Na,CO, (1.5%), NaKC,H,0,.4H,O (3.6%). bath: Cd0 (2.6%), NaCN (9%), 6. Cadmium-plating Na,CO, (1.5%). 7. Silver-plating bath: AgCN (0.4%), NaCN (6%), Na,CO, (0.8%). analysis. sampling
A cellulose ester filter was inserted line to retain particulate matter.
in the
The results obtained are shown in Table 3. No cyanide was detectable in the water entering the plating bath (sample 1). The value found by the standard method without distillation was probably due to positive interferences, which were also found in samples 2 and 3. The agreement of the values obtained for these samples by the fluorimetric method with and without distillation shows that this preliminary step is not necessary. Owing to the low Table 5. Determination
1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN-
75 NO, 40 S2100 SCN100 BrlOOI20 Fe(II1) 10 Ni(I1) 5 Co(I1) 100 Br- 100 I20Sz- 6OSCN60 NO; 80 SCN5 Ni(I1) 6 Fe(II1) 20 Cu(I1) 200 Zn(I1) 20 Cu(I1) 100 Cd(I1) 2OCu(II) 1 Ag(1) 60Br 60SCN- 60160SCN- 40NO; 20Sz-
787
concentration of cyanide in sample 4, no results were obtained when distillation was used. The sample of air (sample 5) also contained a low amount of cyanide and gave the same results by the two methods. Water from synthetic plating solutions. The results obtained with the fluorimetric method for the real samples described above agree with those obtained by the standard method with prior distillation, but, owing to the low concentration of cyanide in some of these samples, it was not possible to obtain reliable results for all of them by use of distillation. For this reason, several synthetic plating solutions were prepared. I6 To obtain a final cyanide concentration similar to that in the waste-water from washing of plated pieces, these solutions were diluted lOOOO-fold with water. Table 4 shows the values obtained by the fluorimetric method without distillation, and the initial compositions of the solutions. The accuracy of the results proves that preliminary distillation is not necessary, at least for this type of sample. Other synthetic samples. A series of synthetic samples containing several mixtures of ions which usually interfere in other methods for cyanide determination were prepared, in order to explore the application of the fluorimetric method. Cyanide was determined in several binary, ternary and quaternary mixtures, and the results are summarized in Table 5 and compared with those obtained by the calorimetric method. The standard calorimetric method gave higher errors than the fluorimetric method for all the samples. Comparison with other methods
Comparison of the fluorimetric determination reported here with the fluorimetric procedures described in the literature ;eads to the following considerations. Only three other methods have determination limits similar to those in our method. They are based on (a) the catalytic action of cyanide on the
of 50 &ml
cyanide in synthetic samples Relative error, %t
Sample composition*
of cyanide
Fluorimetric method f0.5 -1.2 -0.4 +1.0 -0.7 +0.3 -0.5 -0.8 +I.2 -0.9 +1.1 -0.8 -1.2 +0.6 -0.3 -1.2 -0.7
*Each number means the ion/cyanide concentration ‘l-Average of five separate determinations.
ratio.
Standard method -69 -55 & + 100 -74 - 100 -7 -66 -55 -98 ti+100 B + 100 -20 -13 -12 -8 +79 $ + 100
S. RUBIO et al.
788
oxidation of pyridoxal to 4-pyridoxolactone,6 (b) the ligand-exchange reaction between cyanide and the non-fluorescent complex Cu(II)-leucofluorescein,” and (c) quenching of the fluorescence of 2-(o-hydroxyphenol)benzoxazole by copper.‘8 The first of these involves two steps in the sample preparation and a period of 50 min for the final fluorescence development, whereas only 3-7 min reaction monitoring is needed in our method reported here. Furthermore, the oxidation of pyridoxal is not specific for cyanide, there being numerous interferences.’ The second method has a relative standard deviation of 10% for 5 ng/ml cyanide, and sulphide and strongly oxidizing and reducing anions must be absent. At this cyanide level our method has a relative standard deviation of only 4.4%. No data have been reported on the selectivity and precision of the third method. This method has also been proposed for the determination of sulphide, and therefore this ion must be absent. The pyridine-barbituric acid method is also not specific, and a preliminary distillation is necessary to avoid interference. REFERENCES
R. T. Morrison and R. N. Boyd, Organic Chemistry, p. 656. Allyn and Bacon, Boston, 1973. 2. G. G. Guilbault and D. N. Kramer, Anal. Chem., 1966, 1.
38, 834.
Arch. Biochem. Biophys., 1960, 88, 366. 3. V. Bonavita, 4. S. Takanashi, Z. Tamura, A. Yoshino and Y. Lidaka. Chem. Pharm. Bull., 1968, 16, 758. 5. N. Oishi and S. Fukui, Arch. Biochem. Biophys., 1968, 128, 606. 6. A. Gbmez-Hens and M. Valcarcel, Analyst, 1982, 107, 465. 7. S. Takanashi and Z. Tamura, Chem. Pharm. Bull., 1970. 18, 1633. B. Santoni and P. Sanat. 8. M. Brebec, G. Delarue, Analusis, 1975, 4, 127. 9. I. M. Kolthoff and E. B. Sandell, Textbook of Quanlitatioe Inorganic Analysis, 2nd Ed., Q. 574. Macmillan, New York, 1947. 10. Standard Methods for the Examination of Water and Wastewater, 15th Ed., APHA-AWWA-WPCF, Washington, 1980. II. W. Stenstrom and N. Goldsmith, .I. PhJw. Chem., 1926, 30, 1683. 12. L. Sommer, Folia Fat. Sci. Natn. Univ. Purkynianae Brno, 1964, 5, I, 13. J. W. Bridges, D. S. Davies and R. T. Williams, Biochem. J., 1966, 98, 451. 14. K. B. Yatsimirskii, Kinetic Methodc of Analvsis. Chav_ . ter 3, Pergamon Press, Oxford, 1966: 15. P. D. Goulden. B. K. Afehan and P. Brooksbank. Anal. Chem., 1972, 44, 1845. Encyclopedia of Chemical Technology, 16. Kirk-Othmer, Uthea, Mexico. 17. D. E. Ryan and J. Holzbecher, Int. J. Enairon. Anal. Chem., 1971, 1, 159. 18. F. Vernon and P. Whitham, Anal. Chim. Acta, 1972,59, 155.
Copyright
0
0039-9140/84 $3.00 + 0.00 1984 Pergamon Press Ltd
KINETIC FLUORIMETRIC DETERMINATION OF CYANIDE BY MEANS OF ITS CATALYTIC EFFECT ON THE AERIAL OXIDATION OF PYRIDOXAL-5-PHOSPHATE OXALYLDIHYDRAZONE S. RUBIO, A. GOMEZ-HENSand M. VALCARCEL Department of Analytical Chemistry, Faculty of Sciences, University of Cbrdoba, Cerdoba, Spain (Received
9 January
1984. Revised 26 April 1984. Accepted
7 May
1984)
Summary-A highly specific determination of traces of cyanide is described, based on the catalytic effect of cyanide on the oxidation of pyridoxal-5-phosphate oxalyldihydrazone by dissolved oxygen. The reaction is monitored by measuring the fluorescence of the oxidation product (I._ 350, I_,,, 420 nm), and allows determination of 3-180 ng/ml cyanide concentrations with relative standard deviations of 0.4-0.8%. The effect of 56 foreign ions has been investigated: 45 of them do not interfere, and only one, Hg(II), interferes when present at the same level as the cyanide. Its application to the determination of total cyanide in industrial and synthetic samples, without preliminary distillation, is reported.
It is well known that cyanide catalyses the oxidation of some aromatic aldehydes to acids, through formation of cyanohydrins,’ and this has been utilized for spectrophotometric and fluorimetric determination of cyanide.*-5 The reaction between the cyanohydrin of p-nitrobenzaldehyde and o-dinitrobenzene to give a highly-coloured blue compound, the doublycharged anion of o-nitrophenylhydroxylamine, has been used by Guilbault and Kramer.’ The rate of reduction of the o-dinitrobenzene was followed spectrophotometrically at 560 nm. Cyanide could be determined in the range 45450 ng/ml, and the method was highly specific. The fluorimetric methods for determination of cyanide have been reviewed and critically studied.6 The reaction of pyridoxal with oxygen to give 4-pyridoxolactone, a highly fluorescent compound, is catalysed by cyanide. 3-5Factors affecting the reaction rate have been studied by Takanashi and Tamura.’ The reagents are mixed in phosphate medium at pH 7.4 and heated at 50” for 60 min. The fluorescence is measured at pH 10 (achieved by adding sodium carbonate). The calibration graph is linear from 4 to 180 ng/ml. This reaction is regarded as the most sensitive fluorimetric method for cyanide.* Here, the analytical properties of a new compound, pyridoxal-5-phosphate oxalyldihydrazone (PPOH), are described for the first time, and its aerial oxidation is used for the kinetic fluorimetric determination of traces of cyanide by means of their catalytic effect on the oxidation reaction. Under the conditions proposed, the reaction is rapid, measurements being taken less than 7 min after the start. The method is very selective, no preliminary distillation is needed and cyanide concentrations as low as 3 ng/ml can be determined. 783
EXPERIMENTAL
Synthesis
of the reagent
Pyridoxal-5-phosphate (0.6 g) dissolved in the minimum of hot water-ethanol (3:2 v/v) solution was mixed with 0.13 g of oxalylhydrazide dissolved in hot water-ethanol mixture (1: 1v/v). The mixture was refluxed for 30 min, then let cool to room temperature. The yellow product was filtered off and washed with ethanol. Calculated for C,sH,,O,,N,P,, C 37.50x, H 3.82%; found C 37.5x, H 4.1%. Reagents Pyridoxal-5-phosphate oxalyldihydrazone (PPOH) solution 7 x IO-W, in methanol. Several drops of O.OlM
sodium hydroxide (enough to give an apparent pH of _ 7) were added to dissolve the reagent. This solution is stable for at least one month. Standard cyanide solution. Potassium cyanide (2.5 g, equivalent to I g of cyanide) was dissolved in 1 htre of 0.05M sodium hydroxide and standardized argentometrically.9 Ammonia-ammonium chloride 6M buffer solution, pH 9.8. Chloramine-T solution and pyridine-barbituric acid reugent.‘O
Reagent-grade chemicals and pure solvents were used. All anions tested were added in the form of their sodium or potassium salts. Apparatus Spectrofluorimeter. A Perkin-Elmer fluorescence spectrophotometer, model MPF-43A, fitted with a device for kinetic measurement and with l-cm quartz cells was used. The cell compartment of the spectrofluorimeter was kept at constant temperature by circulating water. The sensitivity was set at 3, and the excitation and emission slits were set to give 6-nm spectral band-pass. A set of fluorescent polymer samples was used daily to adjust the spectrofluorimeter to compensate for changes in source intensity. Cyanide distillation apparatus.‘0 Air-sampling equipment. A constant-flow
pump, DuPont model S2500; impingers, volume 30 ml; cellulose ester filters, 37mm in diameter and 0.8pm thick.
S. RUBIO ef al.
784
Procedure To lo-ml standard flasks add 0.5 ml of 7 x 10m4M PPOH solution, 5 ml of 6M ammonia-ammonium chloride buffer (pH 9.8), 2.5 ml of 4M sodium chloride and appropriate volumes of cyanide sample to give a final concentration of cyanide between 3 and 180 ng/ml. Make up to volume with distilled water, mix and after 1 min transfer a portion to a l-cm quartz cell kept at 62 t_ 0.1’. Wait 2 min before starting to record the fluorescence intensity (A,, 350, I,, 420 nm) as a function of time. Run a blank with no cyanide present. Calculate the reaction rate from the difference in the
slopes of the fluorescence-time plots for the sample and the blank.
100
90
5 5
Oxidation of PPOH and catalytic action of cyanide
The excitation and emission spectra of PPOH show a hypsochromic change in basic (&,, 350 nm, &,,, 420 nm) and acid (a,, 320 nm, A,,,, 395 nm) media in presence of such oxidizing agents as KIOj, H202, K,S,Os and NaIO,. Solutions of PPOH in ammonia-ammonium chloride buffer show identical
n2'
70-
E = 608 t _ bW SW_ ">
RESULTS AND DISCUSSION
At room temperature, PPOH is highly soluble in water or in sodium hydroxide solution, moderately soluble in dimethylformamide or methanol and slightly soluble in ethanol. Aqueous solutions are unstable whatever their pH. A 3.5 x 10d5M solution in methanol-water (I :24 v/v) has an absorption maximum at 305 nm and a very small absorption band at 360 nm in acidic and neutral media. In alkaline medium it shows a broad band with an absorption maximum at 305 nm. PPOH solutions are fluorescent (&, 365 nm, i,, 510 nm in acid medium and i,, 400 nm, i,, 510 nm in alkaline medium). The stability of the PPOH solutions is critically dependent on the medium. The reagent is stable in organic solvents such as methanol. Aqueous PPOH solutions are unstable owing to the autoxidation of the reagent by dissolved oxygen. This difference in behaviour occurs because there is a hydrolysis step prior to the oxidation, and this is favoured in an aqueous medium. For this reason, neutral aqueous solutions are more stable than acid or basic ones, because H+ and OH-- ions catalyse the hydrolysis. This autoxidation reaction occurs at pH values very close to that used for the determination of cyanide, which is why blanks are run. for the dissociation in The pK values methanol-water (1:24 v/v) were found”,‘2 to be 3.9, 8.2 and 10.4. The colour reactions of the reagent with 62 ions at various pH values were investigated; it reacts with Mg(II), Co(U), Ni(II), Mn(II), Fe(I1) and Fe(II1). Only the red Mg(IItPPOH complex formed at pH 4.2 and 50” (acetic acid-sodium acetate buffer) is of any photometric analytical interest. The fluorescence reactions are not important. Only V(V) and Al(II1) show fluorescence which is different from that of the reagent, but the sensitivity is low.
n2
,. : eo-
z a,
Analytical properties of the reagent
c c
3020-
rr
lo-A
I
1
1'
\
1
I
xx)
350
I 400
_
I \ 450
X(nm) Fig. 1. Catalytic effect of cyanide on the oxidation of PPOH by dissolved oxygen. Excitation and emission spectra. 1,l’: PPOH; 2,2’: PPOH + CN-. [PPOH] = 8 x lo-‘M; [CN-] = 0.4 pg/ml; [NH,Cl-NH,] = 0.4M. Graphs were recorded after a reaction time of I hr. Temperature 50°C.
fluorescence maxima (J.,, 350 nm, A,,,, 420 nm) to those obtained in the presence of oxidants but the formation of the fluorescent product is so slow that it takes several days for a detectable amount to be produced. In the presence of traces of cyanide, the oxidation rate of PPOH by dissolved oxygen is accelerated. This effect is shown in Fig. 1. The catalytic nature of the cyanide reaction is revealed by the fact that it does not take place in an inert atmosphere. When oxidants are utilized, the reagent is oxidized immediately, and cyanide then does not have a catalytic effect. Several tests have been made to characterize the oxidation product of PPOH and establish the mechanism of the reaction, with the following results (a) The characteristic absorption peak at 305 nm of the Schiffs base formed between pyridoxal-5phosphate and oxalylhydrazide disappears in the course of the reaction. This seems to be due to hydrolytic breaking of the azomethine group. (6) The fluorescence maxima corresponding to the reagent (L,, 400 nm, A,,,, 510 nm) disappear like the absorption maximum, but an intermediate product (J.,, 320 nm, &,,, 370 nm) can be observed before the formation of the final product (A,, 350nm, L,, 420 nm). This intermediate is in equilibrium with PPOH and subsequently with the final product, and its fluorescence characteristics are the same as those of the pyridoxal-5-phosphate.‘3 This seems to corroborate the supposition of hydrolytic breaking. (c) The fluorescence maxima of the oxidation product of PPOH are similar to those reported for a solution of pyridoxal-5-phosphate treated with hydrogen peroxide, which forms 4-pyndoxic acid.
From these observations it can be inferred that the likely oxidation mechanism consists of hydrolytic
Fluorimetric determination of cyanide
T
(‘Cl
785
PH
CPPOHI, IO’ M
Fig. 2. Effect of the temperature (A), pH (B) and PPOH concentration (C) on the reaction rate of the PPOHQCNsystem. breaking of the azomethine group and oxidation of the resulting compounds. Several attempts have been made to isolate the highly water-soluble fluorescent species from mixtures of the parent reagent and cyanide, but without success. Evaporation yielded a reddish brown liquid which would not crystallize. Addition of ethanol produced a yellow solid, the excitation and emission spectra of which were identical to those obtained in the analytical test, but this solid was water-soluble and non-extractable by water-immiscible organic solvents, and was contaminated by the reagent, so meaningful ultraviolet and infrared spectra could not be obtained. Attempts to obtain a pure product by thin-layer chromatography were unsuccessful. Injuence
of reaction variables on the oxidation
The temperature of the cell-holder was varied over the range 35-70” (Fig. 2A). The reaction rate increased linearly withtemperature increase between 35 and 60” but less steeply at higher temperatures. A temperature of 62” was selected for use. From Arrhenius plots of In k vs. reciprocal of the absolute temperature, the activation energy for the catalysed reaction was calculated to be 5.1 +_0.1 kcal/mole. The effect of pH on the reaction rate is given in Fig. 2B. The optimum pH is 9.7-10.0. The effect of the type of buffer (phosphate, carbonate, ammonia) was examined. The reaction develops only in the presence of ammonia-ammonium chloride buffer, and is linearly related to the buffer concentration up to at least 4M. A 3M concentration was chosen. The influence of the PPOH concentration was tested in the range 5 x 10-6-10-4M and is shown in Fig. 2C. A purely aqueous solution could not be used because of its instability, so a methanol solution was used. However, increasing the methanol content of the final aqueous methanol medium decreased the reaction rate, so its concentration in the solution measured was minimized. The rate of reaction rises with increasing ionic strength up to IM and then remains constant. The fluorescence intensity vs. time curves for different cyanide concentrations were recorded. The initial slopes indicate a first-order reaction with respect to cyanide. The various kinetic dependences on pH (Fig. 2B), reagent concentration (Fig. 2C) and buffer concen-
tration are summarized in Table 1, and the following equation is suggested for the oxidation of PPOH (3.5 x 10m5M) by dissolved oxygen at pH 9.85 in the presence of 3M ammonia buffer and cyanide as catalyst: d[PPOH],,/dt
= k[NH,Cl-NHJ3”[CN-1
in which [PPOH],, is the concentration of oxidized reagent and k is the conditional rate constant. This equation does not include the effect of the uncatalysed reaction. Calibration graphs
Three kinetic methods have been tested for the determination of cyanide: tangent ([CN-] 10-180 ng/ml), fixed-time ([CN-] 3-180 ng/ml) and variabletime ([CN-] 3-60 ng/ml).14 For the fixed-time method, measurements were made after 7 min. For the variable-time method, the inverse of the time necessary to obtain a relative fluorescence intensity of 30% was plotted against the cyanide concentration. The relative standard deviation for 50 ng/ml cyanide (P = 0.05, n = 11) is smaller for the initial-rate method (+0.4x) than the fixed-time (f0.8%) and the variable-time (+ 0.7%) methods. The initial-rate method is recommended when precision is the prime consideration and the fixed-time method when a low detection limit is the important factor. Interferences
The effect of 35 anions and 21 cations on the determination of cyanide was tested. The tangent method was used because a change in slope could be more clearly detected than a change in a single data point. The tolerance limits for interfering ions are
Table 1. Summary of kinetic data Dependence of V, on
Concentration range, M
W+l W+l” 1/W+]
4 5 X lo-“-7.9 X lo-” 7:9 x IO-“-2.5 x IO-“’ 2.5 x lo-‘“5.0 x IO-‘0 10-5-2.2 x 10-5 2.2 x 10~5-5.0x 10-S 5.0 X 10-5-10~4 2.0-4.0
[PPOH]“2 [PPOH]’ [PPOH]-3’2 [Buffer]3/*
S. RUBIO et al.
786 Table 2. Effect of various Tolerance
ions on the determination tangent method
of 50 ng/ml of cyanide
by the
ratio,
Ion added
ion to CN_____100
SCN-, F-, Cl-, Br-. S&,
SO:-,
ClOb,
S,O;-,
PO:-,
P,O;-,
Se%,
VO;, MOO:-,
I-, NO,, ClO;,
CO:-,
BrO;,
P,O;,,
SOS-,
IO;,
AsO;,
IO;,
AsO:-,
WO:-, B,O:-, acetate,
tartrate, citrate, K+, Ca2+>Sr2+, Ba2+> Zn*+, Cd*+, A13+, Cr3+, B?+,
Mg’+,
In3+,
Sb’+, Tl+, Pb*+, Ti4+, Zr4+ 75
C,O:-,
NO,
40
CrO:-,
S2-
20
Cu*+, Fe’+
10
Ni’+, Fe(CN)i-,
1
Ag+
less than
1
Hg2+
summarized in Table 2; 45 of the ions tested did not affect the determination of cyanide when present in lOO-fold ratio to it. Silver interferes when present at the same level as cyanide, but mercury(I1) interferes at even lower concentrations. Hence, the method is highly specific. Applications
The determination of microamounts of cyanide in industrial effluents is of interest owing to its toxicity. The pyridine-barbituric acid calorimetric method recommended” has a detection limit of 4 ng/ml for cyanide. To determine cyanide in water samples by this method, it is necessary first to separate it from interfering substances by distillation of hydrocyanic acid from the acidified sample. It is well known,’ and has been verified by us, that when the equipment and conditions described” for the distillation are used, the cyanide is not quantitatively recovered at levels below 1 pg/ml in the original sample. The reasons for this Table
3. Determination (Without
Sample
of cyanide
Cyanide preliminary
have been investigatedI and it was found that the hydrocyanic acid is completely distilled but not totally absorbed in the collection solution. The distillation also takes about 90 min. Because of its high selectivity, we have used the new method for determination of cyanide in several samples without preliminary distillation. Three types of sample were investigated. Water
in water
k + + f
0.4 pg/ml 0.3 pgg/ml 0.3 ng/ml 0.01 mg/m’
Standard method* 8.5 5.3 1.3 0.10
k + + *
air from
and air from an electroolatine
12.5 + 0.4 ng/ml 7.2 4.1 2.5 0.10
and
an
electroplating
factory.
Several samples of water and air were analysed by both the fluorimetric and the standard calorimetric method, with and without distillation. As the true cyanide concentration in the samples was unknown, the distillation was included, on the assumption that the determination could be taken as free of interferences if the results obtained with and without distillation, agreed. The air sample (90 litres) was aspirated at 1.5 l./min through three impingers connected in series, each containing 10 ml of 0.1 M sodium hydroxide; the solutions were mixed for
0.3 pg/ml 0.5 pg/ml 0.6 ng/ml 0.02 mg/m3
factorv
Cyanide, pglml (With preliminary distillation)
distillation)
Fluorimetric method*
1 2 3 4 5
Fe(CN)i-
co2+
5
-
Fluorimetric method* ‘7.1 + 0.2 4.2 + 0.2
*Average and standard deviation of five separate determinations. Sample: 1. Water intake to the baths. 2. Water from the first washing of the silver-plated pieces. 3. Water from the second washing of the silver-plated pieces. 4. Waste-water from a plating factory with zinc-plating, silver-plating baths. 5. Air from the workroom.
Standard method* 7.2 + 0.1 4.0 * 0.3 -
and chromium-plating
Fluorimetric determination Table 4. Determination
of cyanide in synthetic platingbaths Cyanide content, pgglml
Plating-bath 1 2 3 4 5 6
I
Added*
Found?
4.1 5.8 4.4 4.1 3.3 4.8 3.3
1.6+0.1 5.8 + 0.2 4.3 f 0.1 3.9 * 0.2 3.2kO.l 5.0 + 0.1 3.4 kO.1
*Initial concentration diluted lO,OOO-fold with potable water. tAverage and standard deviation of four separate determinations. Sample: 1. Brass-plating bath: CuCN (3%), Zn(CN), (0.9%), NaCN (4.9%), Na,CO, (1.5%), NH, (0.015%). 2. Brass-plating bath: CuCN (2.025%), Zn(CN), (5.0%), NaCN (3%), Na,CO, (2.25%). 3. Brass-plating bath: CuCN (2.6%), Zn(CN), (1. I%), NaCN (4.5%), Na,CO, (1.5%). 4. Brass-plating bath: CuCN (3%) Zn(CN), (2.6%), NaCN (3.8%), Na,CO, (1.5%). 5. Copper-plating bath: CuCN (2.6%) NaCN (3.4%) Na,CO, (1.5%), NaKC,H,0,.4H,O (3.6%). bath: Cd0 (2.6%), NaCN (9%), 6. Cadmium-plating Na,CO, (1.5%). 7. Silver-plating bath: AgCN (0.4%), NaCN (6%), Na,CO, (0.8%). analysis. sampling
A cellulose ester filter was inserted line to retain particulate matter.
in the
The results obtained are shown in Table 3. No cyanide was detectable in the water entering the plating bath (sample 1). The value found by the standard method without distillation was probably due to positive interferences, which were also found in samples 2 and 3. The agreement of the values obtained for these samples by the fluorimetric method with and without distillation shows that this preliminary step is not necessary. Owing to the low Table 5. Determination
1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN1 CN-
75 NO, 40 S2100 SCN100 BrlOOI20 Fe(II1) 10 Ni(I1) 5 Co(I1) 100 Br- 100 I20Sz- 6OSCN60 NO; 80 SCN5 Ni(I1) 6 Fe(II1) 20 Cu(I1) 200 Zn(I1) 20 Cu(I1) 100 Cd(I1) 2OCu(II) 1 Ag(1) 60Br 60SCN- 60160SCN- 40NO; 20Sz-
787
concentration of cyanide in sample 4, no results were obtained when distillation was used. The sample of air (sample 5) also contained a low amount of cyanide and gave the same results by the two methods. Water from synthetic plating solutions. The results obtained with the fluorimetric method for the real samples described above agree with those obtained by the standard method with prior distillation, but, owing to the low concentration of cyanide in some of these samples, it was not possible to obtain reliable results for all of them by use of distillation. For this reason, several synthetic plating solutions were prepared. I6 To obtain a final cyanide concentration similar to that in the waste-water from washing of plated pieces, these solutions were diluted lOOOO-fold with water. Table 4 shows the values obtained by the fluorimetric method without distillation, and the initial compositions of the solutions. The accuracy of the results proves that preliminary distillation is not necessary, at least for this type of sample. Other synthetic samples. A series of synthetic samples containing several mixtures of ions which usually interfere in other methods for cyanide determination were prepared, in order to explore the application of the fluorimetric method. Cyanide was determined in several binary, ternary and quaternary mixtures, and the results are summarized in Table 5 and compared with those obtained by the calorimetric method. The standard calorimetric method gave higher errors than the fluorimetric method for all the samples. Comparison with other methods
Comparison of the fluorimetric determination reported here with the fluorimetric procedures described in the literature ;eads to the following considerations. Only three other methods have determination limits similar to those in our method. They are based on (a) the catalytic action of cyanide on the
of 50 &ml
cyanide in synthetic samples Relative error, %t
Sample composition*
of cyanide
Fluorimetric method f0.5 -1.2 -0.4 +1.0 -0.7 +0.3 -0.5 -0.8 +I.2 -0.9 +1.1 -0.8 -1.2 +0.6 -0.3 -1.2 -0.7
*Each number means the ion/cyanide concentration ‘l-Average of five separate determinations.
ratio.
Standard method -69 -55 & + 100 -74 - 100 -7 -66 -55 -98 ti+100 B + 100 -20 -13 -12 -8 +79 $ + 100
S. RUBIO et al.
788
oxidation of pyridoxal to 4-pyridoxolactone,6 (b) the ligand-exchange reaction between cyanide and the non-fluorescent complex Cu(II)-leucofluorescein,” and (c) quenching of the fluorescence of 2-(o-hydroxyphenol)benzoxazole by copper.‘8 The first of these involves two steps in the sample preparation and a period of 50 min for the final fluorescence development, whereas only 3-7 min reaction monitoring is needed in our method reported here. Furthermore, the oxidation of pyridoxal is not specific for cyanide, there being numerous interferences.’ The second method has a relative standard deviation of 10% for 5 ng/ml cyanide, and sulphide and strongly oxidizing and reducing anions must be absent. At this cyanide level our method has a relative standard deviation of only 4.4%. No data have been reported on the selectivity and precision of the third method. This method has also been proposed for the determination of sulphide, and therefore this ion must be absent. The pyridine-barbituric acid method is also not specific, and a preliminary distillation is necessary to avoid interference. REFERENCES
R. T. Morrison and R. N. Boyd, Organic Chemistry, p. 656. Allyn and Bacon, Boston, 1973. 2. G. G. Guilbault and D. N. Kramer, Anal. Chem., 1966, 1.
38, 834.
Arch. Biochem. Biophys., 1960, 88, 366. 3. V. Bonavita, 4. S. Takanashi, Z. Tamura, A. Yoshino and Y. Lidaka. Chem. Pharm. Bull., 1968, 16, 758. 5. N. Oishi and S. Fukui, Arch. Biochem. Biophys., 1968, 128, 606. 6. A. Gbmez-Hens and M. Valcarcel, Analyst, 1982, 107, 465. 7. S. Takanashi and Z. Tamura, Chem. Pharm. Bull., 1970. 18, 1633. B. Santoni and P. Sanat. 8. M. Brebec, G. Delarue, Analusis, 1975, 4, 127. 9. I. M. Kolthoff and E. B. Sandell, Textbook of Quanlitatioe Inorganic Analysis, 2nd Ed., Q. 574. Macmillan, New York, 1947. 10. Standard Methods for the Examination of Water and Wastewater, 15th Ed., APHA-AWWA-WPCF, Washington, 1980. II. W. Stenstrom and N. Goldsmith, .I. PhJw. Chem., 1926, 30, 1683. 12. L. Sommer, Folia Fat. Sci. Natn. Univ. Purkynianae Brno, 1964, 5, I, 13. J. W. Bridges, D. S. Davies and R. T. Williams, Biochem. J., 1966, 98, 451. 14. K. B. Yatsimirskii, Kinetic Methodc of Analvsis. Chav_ . ter 3, Pergamon Press, Oxford, 1966: 15. P. D. Goulden. B. K. Afehan and P. Brooksbank. Anal. Chem., 1972, 44, 1845. Encyclopedia of Chemical Technology, 16. Kirk-Othmer, Uthea, Mexico. 17. D. E. Ryan and J. Holzbecher, Int. J. Enairon. Anal. Chem., 1971, 1, 159. 18. F. Vernon and P. Whitham, Anal. Chim. Acta, 1972,59, 155.