ANALYTICAL SCIENCES OCTOBER 2004, VOL. 20
2004 © The Japan Society for Analytical Chemistry
1443
Cyanide Reaction with Ninhydrin: Elucidation of Reaction and
Interference Mechanisms
Gabi DROCHIOIU,*† Ionel MANGALAGIU,* Ecaterina AVRAM,** Karin POPA,*
Alin Constantin DIRTU,*** and Ioan DRUTA*
*Organic Chemistry and Biochemistry Department, Al. I. Cuza University of Iasi,
11 Carol I, Iasi-700506, Romania
**Petru Poni Institute of Macromolecular Chemistry, 41a Ghica Voda, Iasi-700487, Romania
***Analytical Chemistry Department, Al. I. Cuza University of Iasi, 11 Carol I, Iasi-700506, Romania
A new sensitive spectrophotometric method has recently been developed for the trace determination of cyanide with
ninhydrin. Cyanide ion was supposed to act as a specific base catalyst. Nevertheless, this paper demonstrates that the
reported assay is based on a novel reaction of cyanide with 2,2-dihydroxy-1,3-indanedione, which affords purple or blue
colored salts of 2-cyano-1,2,3-trihydroxy-2H indene. Hydrindantin is merely an intermediary of the reaction. The
formation of a stable and isolable ninhydrin-cyanide compound has been confirmed by its preparation in crystalline form.
Also, it is thoroughly characterized by elemental as well as MS, IR, UV/VIS and 1H NMR analyses. The Ruhemann’s
sequence of reactions of cyanide with ninhydrin has been reconsidered and an adequate mechanism of the reaction is
proposed. As a consequence, the interference of oxidizers as well as copper, silver and mercury ions with the cyanide
determination has been elucidated.
(Received March 2, 2004; Accepted July 26, 2004)
Introduction
Although cyanides and hydrocyanic acid are most toxic, they
are nevertheless extensively used in many industries as well as
in agriculture.1 Hydrocyanic acid was reported to be present in
cigarette smoke as well.2 The origin of blood cyanide
concentrations may also be from hydrocyanic acid inhalation
from the combustion fumes of plastics.3,4 Due to its acute
toxicity, it is very important to monitor the cyanide
concentration using specific and sensitive analytical methods.
At present, the well-known Aldridge’s method based on the
formation of cyanogen bromide and its subsequent reaction with
pyridine and benzidine to form a highly colored polymethine
dye is still in use for the spectrophotometric determination of
low concentrations of cyanide.5 Its variants involve the use of
reagents, such as pyrazalone6 or barbituric acid,7 instead of
benzidine.
Cyanide may be qualitatively8,9 and/or
quantitatively2,10,11 determined by measuring the absorbance of
chromophores formed by the interaction of cyanide ion with
suitable reagents. A specific method for cyanide is based on
cyanide reactions with p-nitrobenzaldehyde and oVarious indirect methods have been
dinitrobenzene.10
developed based on the discharge of the color of metal
complexes by cyanide as a ligand-exchange reaction.1 A direct
and sensitive method for the determination of cyanide is
reported based on the conversion to ammonia, followed by UV
absorption spectrometry.12 Another spectrophotometric method
is based on the conversion of cyanide to cyanate using sodium
†
To whom correspondence should be addressed.
E-mail: gabidr@uaic.ro
hypochlorite, followed by acid hydrolysis to form ammonium
sulfate. The formed ammonium sulfate is determined based on
the Berthelot reaction using salicylic acid, sodium hypochlorite,
and sodium nitroprusside to form indophenol.1 Also, a sensitive
gas-chromatographic method for the determination of cyanide
in biological specimens, based on its conversion to cyanogen
chloride,13 as well as a fluorometric procedure14 were described.
One of the most recent spectrophotometric methods reported
for the trace determination of cyanide is the reaction of cyanide
with ninhydrin in an alkaline medium.15 This method is very
sensitive, highly specific, and relatively free from interference
by various species, and does not require heating or extraction.
Nevertheless, the authors have claimed that cyanide ion can act
only as a specific base catalyst. Consequently, the analytical
behaviors of the reaction, especially the interference and the
opportunity to be used in biological media, cannot be explained.
Previously,16 we also developed a highly sensitive and selective
analytical procedure for cyanide based on its reaction with
ninhydrin, and showed that a novel reaction may occur,17 not a
catalytic one.18 Being very simple, accurate, fast, selective and
sensitive, this method is also useful in the determination of as
little as 0.025 µg ml–1 CN– in biological samples.19
Nevertheless, the process whereby the purple or blue
compounds are formed from ninhydrin remained obscure.
In the present work, the formation of a stable and isolable
ninhydrin-cyanide compound was confirmed by its preparation
in crystalline form. Also, most of the reaction products were
isolated and characterized both spectroscopically and by
elemental analyses. Further on, the reaction sequence was
compared with that in the literature.20 Thus, we established the
mechanism of reaction, which is precious for a cyanide assay,
and explained scientifically the interference. In addition, the
1444
ANALYTICAL SCIENCES OCTOBER 2004, VOL. 20
novel cyanide assay in environmental and biological samples
based on its reaction with 2,2-dihydroxy-1,3-indanedione has
been greatly improved.
Experimental
Instrument
An UVIKON 933-KONTRON double-beam UV/VIS
spectrophotometer with 1-cm matched cells of quartz was used
for spectral measurements. In addition, some reaction steps
were performed on a CINTRA 10e UV-visible spectrometer in
order to establish the most adequate reaction mechanism. The
FT-IR spectra were taken on a “Jasco 660-Plus” Fouriertransform infrared spectrometer using KBr-diluted samples
against a KBr standard (1 – 2% of analyzing sample). The
measured wave number range was 350 – 4000 cm–1 with a
resolution of 4 cm–1 and a scanning speed of 2 mm/s. 1H-NMR
spectra were recorded with a LUCY (DPX300) 300 MHz
spectrometer using D2O, acetone-d6, and DMSO-d6,
respectively, as solvents. Mass spectra were carried out with a
VESTEC-201, 2000 amu, with an ionization source, EI-CI. The
pH values were measured with a CG 837-Schott pH meter.
Elemental analyses were obtained from Organic Chemistry
Laboratory of Al. I. Cuza University of Iasi. Column
chromatography was performed with silica-gel grade 230 – 400
mesh. Melting points were obtained in open capillary tubes,
and were uncorrected.
Reagents
All chemicals were of analytical reagent grade (Merck), and
all solutions were prepared with milliQ grade water with R =
18.2 Ω. A standard solution of cyanide was prepared according
to Nagaraja et al.15 A 0.5% aqueous solution of ninhydrin in
2% sodium carbonate was prepared. Also, a 2.5 mol dm–3
sodium hydroxide solution was used. Nitrogen was bubbled
into each solution to release the interfering oxygen.
Synthesis of a cyanide-ninhydrin adduct
Potassium cyanide and ninhydrin were mixed in a 2:1 molar
ratio. Thus, 1.78 g (10 mM) of ninhydrin, 1, was added to 1.3 g
of KCN (20 mM) solved in a 2% solution of sodium carbonate.
The white powder of ninhydrin dissolved immediately to form a
deep red-colored solution. Upon adding a hydrochloric acid
solution, white-pink crystals separated (compound 6). The
reaction, which was carried out under nitrogen, proved to be
quantitative. The precipitate was washed out several times with
milliQ grade water on the filter paper and dried at room
temperature. All of the operations were also carried out under
nitrogen. The thus-obtained crystals melted at 124 – 126˚C
(uncorrected). The structure of the product 6 as well as that of
the other compounds in the reaction sequence was established
by elemental and spectral (MS, IR, 1H NMR) analyses. When
dissolved in solutions of sodium carbonate and sodium
hydroxide, respectively, compound 6 changed its color to red (5,
λmax 485 nm) and blue (4, λmax 590 nm). Upon evaporating the
solvent under a vacuum, the colored compounds were obtained
in crystalline form.
Hydrindantin formation
Small amounts of cyanide are supposed to catalyze the
formation of hydrindantin.15,22 In order to prove the correct
mechanism of hydrindantin formation, we additionally treated 1
mM ninhydrin (178 mg) with 65 mg of KCN, an equivalent
amount of cyanide, in the presence of sodium carbonate.
Fig. 1 Effect of the cyanide concentration on the shape of the
absorption curve.
Separately, the amount of cyanide was doubled. Both redcolored solutions were diluted accordingly, and their
absorbances were read over the wavelength range from 400 nm
to 600 nm. In addition, the reaction of ascorbic acid with
ninhydrin to afford hydrindantin was also performed. The three
spectra were compared with each other. It was assumed that
cyanide reacted stoichiometrically with ninhydrin to form
hydrindantin as a major product.
Recommended procedure
Samples of 10 – 100 µL of solutions containing less than 0.02
µg cyanide were pipetted into an Eppendorf vial of 1.5 mL, and
0.5 – 1.0 mL of ninhydrin reagent was added. To increase the
sensitivity, 0.5 – 1.0 mL of sodium hydroxide was sometimes
added. The resulting solution was bubbled with nitrogen,
capped and let to develop color. The deep-red or deep-blue
color was measured at 485 nm and 590 nm, respectively. A
reagent blank with no cyanide and a standard with 0.01 µg mL–1
cyanide were used.
Results and Discussion
Novel reaction of cyanide with ninhydrin
Hydrindantin, 3, seemed to be the major product of the
reaction when ninhydrin was treated 1:1 with potassium
cyanide. Both the solution of hydrindantin produced in the
reaction with ascorbic acid and that of cyanide-ninhydrin 1:1
adduct absorbed at 485 nm. Nevertheless, the former was
unstable and its color vanished in the long run. Under these
circumstances, we confirmed hydrindantin formation, as
previously shown by Nagaraja et al.15 Nevertheless, once
formed, hydrindantin possibly reacted with another molecule of
cyanide to afford another dye with a characteristic peak at 460
nm, which was superposed with the other at 485 nm. Its size
was 94% of the intensity of the 485 nm peak. This peak was
previously not reported.
Upon doubling the amount of cyanide, the absorbance at 485
nm increased 2.12 times, and that at 460 nm 2.98 times (Fig. 1).
Therefore, hydrindantin was the main, but not the sole, product
of the reaction, and it reacted with the excess of cyanide to
ANALYTICAL SCIENCES OCTOBER 2004, VOL. 20
1445
Scheme 2 Degradation of 2-cyano-1,2,3-trihydroxy-2H indene
under aerobic condition and at a high pH value.
Scheme 1 Reaction sequence for the formation of 2-cyano-1,2,3trihydroxy-2H indene, 6.
complete the formation of the novel compound (Scheme 1).
Even if this final product of the reaction proved to be quite
stable, it could be decomposed to hydrocyanic acid or cyanide
ions and hydrindantin. Thus, upon heating a solution of this
adduct in 2% sodium carbonate with a reagent for cyanide,
different from the ninhydrin reagent described above, the
reaction for cyanide ion was positive.
Analyses
Elemental (C, H, and N) analyses were compatible with the
proposed structure for 6. Having shown that the red and blue
colors of 6 were obtained simply by a change in the pH, we
considered that the colors were due to the anions of 2-cyano1,2,3-trihydroxy-2H indene. A similar discussion was made
previously for the indene structure of hydrindantin.21 The
transformation of ninhydrin, 1, into hydrindantin, 3, which is a
2H indene form, a more stable structure, would allow for two
ionizable groups. Therefore, the red color is attributable to the
monovalent anion, and the blue color to the divalent anion. The
ease of oxidation upon exposure to air and oxidizers is
consistent with the indene structure of 3 – 6.
The fragmentation schema was also in best agreement with
the proposed structure for 6. The molecular weight of 6 of 189
Da was determined by mass spectrometry. The main fragments
were found to be at 104, 188, 76, 133, 134, and 106 units,
respectively.
The 1H-NMR spectrum also confirmed the structure of 2cyano-1,2,3-trihydroxy-2H indene. The high peak at 7.90 ppm
was assigned to the four aromatic protons. Its shape suggested
that they are almost equivalent due to the specific structure of 6,
which contains the two enolic groups. Contrary, a structure
with two ketone groups would have generated different signals
in the 1H-NMR spectrum for the aromatic protons. The three
OH protons generated a large band, which was situated from
4.50 ppm to 5.33 ppm in the spectrum, in accordance with the
proposed structure.
The high peak at 700 cm–1 in the IR spectrum was assigned to
1,2-disubstituted ethylene. The cyan group gave a characteristic
signal at 2220 – 2250 cm–1, while the hydroxyl group absorbed
at 1200 cm–1. The OH groups associated by H-bonds presented
an intense band at 3200 – 3500 cm–1. Moreover, all of the other
signals in the IR spectrum were in good agreement with the
proposed structure of 6.
Also, all of the properties of 6 supported its structure, which
was confirmed spectroscopically. Thus, the red color of 5,
which is an anion form of 6, vanished upon shaking air into its
solution, particularly upon heating. Compound 6 proved to be a
powerful reducer due to its HO groups in positions 1 and 3,
which remind us of the similar behavior of the two HO enolic
groups in ascorbic acid. Moreover, compound 6 was quite
stable in the solid state or at a low pH value. The solution
changed color from red to blue with increasing pH. The color
again turned to white if the pH was decreased. Bromine,
chlorine and other oxidizers destroyed the colored compound to
form HCl and HBr. The presence of some reducers, such as
ascorbic acid, enhanced the color stability. Under anaerobic
reaction conditions, the red color was stable indefinitely, even at
100˚C. Also, under more alkaline conditions, the reaction gave
a blue solution, which was stable for weeks in the absence of
air. Therefore, cyanide reduced ninhydrin, 1, to hydrindantin, 3,
which reacted with another cyanide molecule to form a
stabilized 2H indene, 6. This one is stable under anaerobic
conditions, but can be easily oxidized to Ruhemann’s
compound, 7, which affords 8 and 9 with increasing pH
(Scheme 2).
Spectral and elemental characteristics
2-Cyano-1,2,3-trihydroxy-2H indene (6). White-pink crystals,
m.p. 124 – 126˚C. IR (KBr): ν– = 1200 cm–1 (OH group); 3250
cm–1 (OH group with H bond); 2240 cm–1 (CN group); 700 cm–1
(1,2-disubstituted ethylene). MS 104 [the highest peak, which
was assigned to C7H4O·+ resulted from M–(H+HCN+CO+·CHO)];
188 (M–1); 76 (C6H4+); 133 (C6H4C+(OH)CO or
134
[M–(HCN+CO)];
106
C6H4(CHO)C≡O+);
[M–(HCN+2C=O)]; 78 (C6H6+); 189 (the molecular peak); 161
(M–H–HCN); 51 (C4H3+-aromatic). 1H-NMR (DMSO-d6, 50˚C,
δ): 7.90 ppm (aromatic 4H); 4.50 ppm to 5.33 ppm (3HOH).
C10H7NO3: calcd. C 63.49; H 3.73; N 7.40; found C 63.10; H
3.55; N 7.55.
2-(2-Cyano-2-hydroxy-acetyl)-benzoic acid (8). Pale-brown
crystals, m.p. 158˚C. IR (KBr): ν– = 1770 cm–1 (carboxylic C=O
group); 1620 and 1640 cm–1 (C=O group); 2700 – 2850 cm–1
(carboxylic OH group); 1075 cm–1 (OH group); 1300 cm–1
(carboxylic OH group); 3400 – 3450 cm–1 (OH group with H
bond); 2210 cm–1 (CN group); 745 cm–1 (1,2-disubstituted
benzene); 1540 and 1600 cm–1, respectively (characteristic
peaks for the aromatic rings); 3000 – 3100 cm–1 (aromatic);
1360, 1390 and 2860 cm–1 (C–H bond). MS 106 (the main
signal), which was assigned to M–(CO2+OHC–CN); 135 was
assigned to M–(CO2+CN); 77 (C6H5+); 105 (C6H5–C≡O+); 134
(C6H5C(OH)CO+); and 51 (C4H3+-aromatic, a low peak). 1HNMR (DMSO-d6, 50˚C, δ): 5.67 ppm (alcoholic OH); 2.15 ppm
(aliphatic proton); 7.50, 7.62, 7.70, and 7.95 ppm, respectively
(aromatic protons); 8.45 ppm (carboxylic proton). C10H7NO4:
calcd. C 58.54; H 3.44; N 6.83; found C 58.30; H 3.50; N 6.80.
The reaction mechanism
The experimental data demonstrated that cyanide ion attacks
nucleophylically the C=O group at position 1 of ninhydrin, 1.
Then, the negative charged oxygen at position 2 (due to alkaline
medium) attacks nucleophylically the carbon atom of the cyan
group to release cyanate ion –OCN, and to afford an isomer of
1446
ANALYTICAL SCIENCES OCTOBER 2004, VOL. 20
Fig. 3 Ninhydrin reaction with small amounts of cyanide.
Fig. 2 Spectra of cyanide-ninhydrin complex in 2% Na 2CO3 30
min (1) and 24 h (2) after ninhydrin was added.
hydrindantin (1,2-dihydroxy-3-keto-3H indene). This one turns
into the 2H indene isomer, which is more stable. The process
whereby ninhydrin is rapid dissolved in a solution of KCN with
the formation of a purple solution was thus explained by the 2H
indene structure with the two ionizable HO groups at positions 1
and 3. Then, cyanide ion attacks hydrindantin at the C=O group
at position 2 to form 2-cyano-1,2,3-trihydroxy-2H indene or its
anions according to the pH value. The motive force of the
reaction seems to be the high stability of the indene structure
over the other structures involved in this study. When we put
ninhydrin in a sodium carbonate solution for 48 h, hydrindantin
have also been formed. It was red in color and, upon bubbling
air, its color vanished and ninhydrin resulted. The colorless
solution was able to react again with cyanide. The cyan group
also stabilizes the indene structure. Upon blocking with urea or
alcohol of one of the HO groups at position 2 of ninhydrin, 1,
the reaction between cyanide and ninhydrin was hindered. In
addition, Ruhemann’s compound 7 cannot form 2-cyano-1,2,3trihydroxy-2H indene, even if it reacts with cyanide, because
one HO group at position 2 was changed by a cyan group.
The spectra of colored solutions of the cyanide-ninhydrin
adduct showed a gradual disappearance of the yellow color,
which is characteristic of ninhydrin in the presence of sodium
carbonate (Fig. 2). In addition, the absorbance at 485 nm,
slowly decreased, and was practically absent after 2 weeks from
the beginning of the experiments. Contrary, the absorbance at
460 nm was constant over this interval. Thus, Fig. 2 shows the
good stability of the 2H indene derivative, as compared to
hydrindantin.
To prove the fact that the cyanate ion is really released as a
by-product, we identified it in the solution as ammonia after
hydrolysis with sulfuric acid and distillation with 32% sodium
hydroxide, as recommended by Deepa and coworkers.1 Thus,
the possibility of other reaction mechanisms has been excluded.
Stability of the colored product
Constant absorbance values were obtained 15 – 30 min after
adding a ninhydrin reagent with sodium carbonate as a function
of the cyanide concentration in the samples. When sodium
hydroxide was added to a deep-red colored solution, the
solution became instantaneously blue in color. Due to nitrogen
addition, the colors remained stable indefinitely. An increase in
temperature of up to 100˚C resulted in no fading of the color.
Upon bubbling air in each solution, the colors gradually faded
within a few minutes.
Nevertheless, when no nitrogen was used, the stability of the
colored solutions was low, especially in the presence of small
amounts of cyanide (Fig. 3). The experimental data suggested
the tendency of hydrindantin formation, which is less stable.
Nevertheless, each time the 2H indene derivative was present.
Its formation as well as the presence of hydrindantin clarified
the properties of the ninhydrin reagent and possible
interference.
Interference studies
The interfering effects of common anions and cations, which
may co-exist with cyanide, were studied. Also, the effect of
amino acids, ascorbic acid, and oxidizers, which may be present
in biological fluids, was investigated. In the free cyanide
determination, varying concentrations of interfering species
were introduced into 0.2 µg of cyanide, and the recovery of
cyanide was established following the procedure described
under the determination of free cyanide.
Copper formed a water-soluble, greenish 1:2 molar complex
with ninhydrin in a 2% sodium carbonate solution, which was
unable to react with cyanide (Scheme 3). Nevertheless, when
the ninhydrin was present in a large excess, a reaction with
cyanide was again possible. An excess of Cu2+ resulted in the
formation of a bluish precipitate of copper carbonate. Upon
adding sodium hydroxide, pH increased dramatically and a 1:1
molar complex of copper with ninhydrin occurred.
Nevertheless, the deep-green solution of the copper-ninhydrin
complex turned immediately into a dark-green precipitate.
Once formed, the red or blue color of cyanide-ninhydrin adduct
was not disturbed by copper ions. Therefore, it was established
that copper ions interfere with cyanide determination due to its
reaction with ninhydrin, not with 2-cyano-1,2,3-trihydroxy-2H
indene.
Silver and mercuric ions also interfere with cyanide
determination, making precipitates with the ninhydrin reagent.
Nevertheless, their interference was abolished by adding
hydroxylamine. The presence of some reducers, such as
ascorbic acid, enhanced the color stability. Nevertheless, large
amounts of ascorbic acid reacted with ninhydrin to form
hydrindantin, which masked the formation of red-colored salts
of 6. In the present method, ascorbic acid could be tolerated up
ANALYTICAL SCIENCES OCTOBER 2004, VOL. 20
1447
support and encouragement.
References
Scheme 3
solutions.
Reaction of ninhydrin with copper ions in alkaline
to 50 mg mL–1.
Application of the recommended method
Previously, samples of industrial and river water as well as
blood and urine were analyzed by the recommended assay. The
results of the determinations were in good agreement with those
made with a well-known spectrophotometric method.11,16,17
Hydroxylamine was added in the ninhydrin reagent only in the
case of samples of water being analyzed. Small amounts of
copper in blood and serum did not interfere with the
determination of cyanide.
Conclusions
The reaction of ninhydrin with cyanide has been reconsidered,
and evidence supporting an indene structure of a novel
compounds, 2-cyano-1,2,3-trihydroxy-2H indene has been
obtained. The main reaction product has been isolated in
crystalline form and thoroughly characterized. Therefore, an
adequate reaction mechanism was advanced, in which
hydrindantin is merely an intermediary of the reaction. The
interference of oxidizers as well as copper, silver and mercury
ions with the cyanide determination has been elucidated.
These data could be important to clarify the analytical
properties of the novel compounds involved and to improve the
procedure for cyanide determination in the environment and in
body fluids.
Acknowledgements
We gratefully acknowledge the financial support of the CNCSIS
Bucharest (Grant 97/2003).
The authors also express
appreciation to Professor Michael Przybylski for technical
1. B. Deepa, N. Balasubramanian, and K. S. Nagaraja, Anal.
Lett., 2003, 36(13), 2861.
2. G. Drochioiu, I. Mangalagiu, and V. Tataru, Analyst, 2000,
125, 939.
3. H. R. Wetherell, J. Forensic Sci., 1966, 11, 167.
4. I. Sunshine and B. Finkle, Int. Arch. Arbeitsmed., 1964, 20,
588.
5. W. N. Aldridge, Analyst, 1944, 69, 262.
6. J. Epstein, Anal. Chem., 1947, 19, 272.
7. G. V. L. N. Murty and T. S. Viswanathan, Anal. Chim.
Acta, 1961, 25, 293.
8. O. Wawschinek, H. Paletta, and W. Beyer, Arch. Toxikol.,
1968, 23, 52.
9. A. Brands, “Asphyxiant Gases, in Handbook of
Toxicology”, ed. T. J. Haley and W. O. Berndt, 1987,
Hemisphere Publishing Corporation, Washington, 472 –
503.
10. G. G. Guilbauld and D. N. Kramer, Anal. Chem., 1966, 38,
834.
11. S. Amlathe and V. K. Gupta, Fresenius’ J. Anal. Chem.,
1990, 338, 615.
12. S. Grieve and A. Syty, Anal. Chem., 1981, 53, 1711.
13. J. C. Valentour, V. Aggarwal, and I. Sunshine, Anal.
Chem., 1974, 46, 924.
14. D. Felscher and M. Wulfmeyer, J. Anal. Toxicol., 1998, 22,
363.
15. P. Nagaraja, M. S. H. Kumar, H. S. Yathirajan, and J. S.
Prakash, Anal. Sci., 2002, 18(9), 1027.
16. G. Drochioiu, Talanta, 2002, 56(6), 1163.
17. G. Drochioiu, Anal. Bioanal. Chem., 2002, 372(5 – 6), 744.
18. G. Drochioiu, I. Ardeleanu, T. L. Timofte, R. Danac, and I.
Druta, Ann. Stin†. Univ., “Al. I. Cuza”, Iasi, 2003, 9, 155.
19. G. Drochioiu and I. Mangalagiu, Pakistan J. Appl. Sci.,
2002, 2(6), 658.
20. S. Ruhemann, J. Chem. Soc., 1910, 97, 2025.
21. D. A. MacFadyen and N. Fowler, J. Biol. Chem., 1950,
186, 13.
22. T. C. Bruice and F. M. Richards, J. Org. Chem., 1958, 23,
145.