- Email: [email protected]
The absorptiometric determination of traces of cobalt with 1-nitroso-2-naphthol
VOL. 1 (1956)
THE
I35
CLINICA CHIMICA ACTA
ABSORPTIOMETRIC COBALT
WITH
DETERMINATION
OF
TRACES
OF
I-NITROSO-%NAPHTHOL by
G. MIDDLETON AND R. E. STUCKEY The British Drug Houses Ltd., City Road, London, N. I (England)
The significance of traces of cobalt in biological material has been the subject of a number of recent papers, although in most cases little attention has been paid to the details of the methods used for the determination. In attempting to employ such published methods for the determination of traces of cobalt in large quantities of, animal tissue (e.g. in liver) and in pharmaceutical preparations, it was found to be impossible to recover and determine the cobalt satisfactorily, and the procedures involved were therefore investigated more closely. An investigation of the preliminary destruction of biological tissues has already been reported’, 2 and the present paper, deals with the use of I-nitroso-a-naphthol for the absorptiometric determination of traces of cobalt. After the introduction by ILINSKI ANDVON KNORRE~O~ I-nitroso-z-naphthol as a reagent for cobalt, numerous other compounds of the same type were suggested. Fdr calorimetric determination of traces of cobalt, nitroso-R-salt has generally been preferred, as the reagent itself is stable and easily obtained in a relatively pure state; further, the reaction may be made specific for cobalt. The utility of the reaction is however, limited, since the colour cannot be concentrated and the reagent is not very stable in solution; in addition the unavoidable excess of reagent gives a yellow colour. (Recent techniques, involving the use of an alumina column for removal of excess of reagent and separation of cobalt from other metals, have considerably increased its possibilities43 “.) Unsulphonated compounds of the nitrosophenol group are all sparingly soluble in water, and offer the potential advantage that their metallic compounds may be extracted from aqueous solutions by an immiscible solvent. It would appear that one of the main reasons for their comparative obscurity lies in the difficulty of obtaining or keeping them in a sufficient state of purity. Thus SYLVESTER AND LAMPITT~ used r-nitroso-z-naphthol merely for the purpose of precipitating cobalt and other metals from the mineral constituents of a digest, followed by a dithizone procedure for the separation of cobalt and a final calorimetric determination with nitroso-R-salt. z-Nitroso-I-naphthol has also been recommended as a more sensitive reagent for cobalt than the I : 2 isomer. The use of r-nitroso-z-naphthol as a reagent both for the isolation and for the calorimetric determination of cobalt in animal products was proposed by PAULAIS*, who devised a method for the calorimetric determination of cobalt. In this method cobalt-nitrosonaphthol is extracted with chloroform from a solution of the ash, excess of reagent being eliminated by extraction with sodium hydroxide, and other metals by removal on a column of alumina. Attempts by the authors to apply this References p. 142
CC.MIDDLETON, R. E. STUCKE1-
136 method
to the determination
of cobalt
in liver and similar
VOL. 1 (1956)
materials
did not give
consistent results, but the method appeared to be promising and a more detailed study was made. The chief difficulties encountered in the application of the PAULAIS method were ir: ;.he complete removal of the iron, and also in that a low and constant blankvalue u ,js not obtained. It was ultimately found that the latter difficulty was due to the p: l:sence in the reagent of colonred impurities which were weakly adsorbed on alumina and passed into the final solution to a varying extent. The present paper describes a method, based on that of PAUL.IW, which has proved satisfactory for the determination of relatively minute amounts of cobalt in an’rnal tissues. A procedure for the removal of interfering impurities in r-nitrosoa-naphthol is given, together with a description of the behaviour of its metallic compounds
on alumina.
Interference
due to iron is eliminated
by reduction
of the
iron nitrosonaphthol compound with hydroxylamine. The method is applied to the determination of cobalt in the residue of mineral constituents obtained from animal matter
by the nitric
acid process*. EXPERIMENTAL
Comparison
of wktroso-a-naphthol
and a-nitroso-r-nafihthol
The recommendation of z-nitroso-r-naphthol is derived from the work of HEI.LUCCI~, and is only valid under certain special conditions, i.e. for the colour given b! a trace of cobalt in a large A volume of water. It is not necessarily applicable to other conditions, and observations showed that although z-nitroso-rnaphthol reacts with cobalt salts in the same manner as r-nitroso-z-naphthol, the resulting compound is less suited to absorptiometric measurcment than the I : 2 isomer, since the extinction curve for cobalt z-nitroso-r-naphthol (see Fig. I) shows a maximum of
ov-
Fig. I. hbsorption curves for metallic nitroso-naphthol derivatives. A, Co r-nitroso-r-naphthol ; R, Fe(ic) r-nitrosor-naphthol; C, Cr(ic) r-nitroso-z-naphthol; D, Co z-nitrosor-naphthol.
is about
double
Pa.wification
that
of the 2 : I compound
low intensity at 530 mp. It is therefore less suited to absorptiometric measurement than the I : z compound which ha: a well defined peak at 415 rnp at which point the absorptioii
at 530 m/l.
of I-nitroso-2-naphthoI*
A major difficulty experienced with the in the values of blank tests in which cobalt .__ -_.--
method was the variation was absent. This was ultimately
PAULAIS
found found
VOL. 1 (1956)
DETERMINATION OF TRACES OF COBALT
to be due to the presence pletely
adsorbed
and, in addition
in the reagent
by the alumina to the possibility
or aminonaphthol formed: moreover
respectively), the coloured
of coloured
impurities
‘37 which were incom-
column used. r-Nitroso-z-naphthol of simple oxidation a number impurities
or reduction
is very reactive (to nitronaphthol
of other coloured compounds may be formed in the manufacture are very
difficult to remove by recrystallisation. In the method of purification given by PAULAISg, I-nitroso-z-naphthol is dissolved in acetic acid, diluted with water and filtered; the nitroso-naphthol is extracted from the filtrate with chloroform and then extracted from the chloroform solution by dilute sodium hydroxide. This final alkaline solution is used as the reagent. This procedure does not, in our experience, produce a reagent solution of satisfactory purity, especially with some commercially available samples of I-nitroso-2-naphthol. It was found that dilution of the acetic acid solution precipitated a proportion of the r-nitroso-z-naphthol varying according to the temperature and time of standing before filtration. In one experiment the weight of the rejected precipitate corresponded to 67 per cent of the original material, only 26 per cent of the original amount being present in the solution. The first (insoluble) fraction was heavily contaminated with a red coloured substance (absorption maximum, 485 rnp) which was not a metallic compound of the reagent, and which was also present in a smaller, but still significant, amount in the second fraction. The quantity of reagent (3 mg) used by PAULAIS for a determination is too small to produce considerable errors from this cause but if other metals (e.g. zinc) are present in appreciable quantity a corresponding excess must be used as these metals compete for the reagent and the amount of coloured impurity introduced is thereby increased. It is not possible to allow for the colour produced by means of the usual blank test, since the absorption of the red impurity on the alumina is weak and variable, and an impure reagent does not give reproducible results in blank experiments. The result of a number of experiments showed that r-nitroso-e-naphthol could most readily be purified by formation of the comparatively insoluble sodium salt. A blank determination using 0.2 g of the purified sodium I-nitroso-2-naphthol showed the product to be free from the coloured metric determination of cobalt.
impurities
which interfere
An aqueous solution of sodium r-nitroso-2-naphthol, dilute for convenient use as a stock reagent. A concentrated
with absorptio-
although stable, is too solution (e.g. IO per cent)
may be made in glacial acetic acid, but this is unstable and, after a few weeks, appreciable quantities of coloured impurities are present. The most convenient reagent was found to be a suspension of I-nitroso-2-naphthol in dilute acetic acid, which can be kept in a dark bottle for several months without deterioration. The method of purification is as follows: Dissolve IO g of crude I-nitroso-z-naphthol in 300 ml of alcohol (methylated 95 per cent) and add 25 ml of 20 per cent aqueous sodium hydroxide. After allowing to stand for a few hours in a cool place, filter off the precipitated green sodium salt under reduced pressure and wash with a small amount of alcohol. Suspend the sodium salt in 2jo ml of alcohol, add 7 ml of glacial acetic acid and warm until solution is effected. Add 4 g of decolorising charcoal, boil, and filter. After cooling, add 25 ml of 20 per cent aqueous sodium hydroxide and after standing, remove the sodium salt by filtration as before and dry at about 50” C. The resulting product is a bright green crystalline powder. lirflw!lc~'.s b, rp
G. MIDDLETON,
138
R.
E.
STUCKEY
VOL.
1 (1956)
Metallic compounds of I-nitroso-2-naphthol A number of di- and tri-valent
metals form coloured compounds
with r-nitroso-o-
naphthol. These are produced directly in aqueous-generally somewhat acidsolution and may usually be extracted, together with excess of reagent, with chloroform, giving coloured solutions from which the excess of reagent may be removed by shaking with sodium hydroxide, or by treatment with alumina. From a chloroform solution the compounds formed with divalent metals are generally adsorbed on alumina, while those of trivalent metals pass through, and the Paulais method of separation of cobalt is based on this fact. The behaviour of various metals with the reagent is described below. Cobalt. The cobalt compound, prepared from a cobaltous salt, is not adsorbed on alumina, this in itself being an indication of its trivalent characterlo. This implies that in the formation of the compound the cobalt ion has been oxidised by excess of the reagent. A chloroform
solution
of the cobalt
salt of r-nitroso-2-naphthol
has an orange
red colour with an absorption maximum at 415 rnp (see Fig. I). Such a solution is stable and remains unaltered after standing for a few days in diffused light or by shaking with caustic soda or hydrochloric acid or by evaporation on a water-bath and re-solution. The colour intensity obtained from a given weight of cobalt salt was found to be proportional to the amount of cobalt present (at least up to IO lug of cobalt). Since the final solution contains only one coloured substance, methods of visual calorimetry may be used if desired. The quantitative recovery of cobalt by chloroform extraction naphthol compound was demonstrated by evaporating a chloroform
as r-nitroso-zsolution of the
cobalt compound to dryness, treating the residue with nitric and sulphuric acids to destroy the nitroso-naphthol, and repeating on this residue the process of formation of cobalt nitrosonaphthol and extraction with chloroform to the original volume. The colour obtained Chromium.
was equal to that of the original solution. It is known that chromium salts give a precipitate
with x-nitroso-
2-naphthol. It is, therefore, somewhat surprising to note that P.4ULAIS1’ specifically includes chromium (“Cr + 1I-et CrV1+“) among the metals which do not react; possibly this is because in his method it is precipitated as phosphate. If, however, a solution containing a trace of chrome alum, with sodium acetate, acetic acid and r-nitroso2-naphthol, is brought to the boil, there results a red colour which may be extracted with chloroform and which then passes through an alumina column. The chloroform solution of chromium I-nitroso-z-naphthol has an absorption maximum at 416 mp-practically the same point as cobalt (see Fig. I). Interference from chromium may be prevented by the addition of citrate to the reaction mixture. Iron. Ferric r-nitroso-a-naphthol forms a brownish solution with an absorption maximum at 408 rnp, and it also passes through alumina. In the presence of large quantities of iron the addition of phosphate (as used by PAULAIS), citrate or fluoride was not effective in ensuring the complete absence of iron from the chloroform extract. Since however the ferric complex is easily reduced to ferrous, which is absorbed on alumina, interference due to iron can be eliminated by the addition of hydroxylamine to the alkaline solution used to remove excess of the reagent. Ferrous I-nitroso-2-naphthol gives a bright green colour in chloroform solution and is thrown out as a blue precipitate on shaking this solution with dilute sodium
VOL.
1 1956)
DETERMINATION
hydroxide (though not with potassium gives a blue-green band.
OFTRACES
hydroxide).
OFCOBALT
On alumina,
I39 the ferrous compound
Iron salts can be b0t.h oxidised and reduced by an excess of the reagent so that a mixture of the ferrous and ferric complexes is always obtained. Shaking a chloroform solution with hydrochloric acid partially converts the ferrous compound into the ferric. Zinc. An insoluble brown zinc compound can easily be obtained from zinc chloride and the reagent, but it is practically insoluble in chloroform and is decomposed by sodium hydroxide, so that zinc is eliminated before reaching the column. Copper and Nickel. These give a yellow to brown solution and similar coloured bands on alumina. The solutions in chloroform are largely decolorised by shaking with caustic alkali. Other metals. sium, aluminium, of cobalt.
Coloured compounds were not obtained from manganese, magneor bismuth salts under the conditions used for the determination
METHOD
Reagent Dissolve
0.5 g of the purified
sodium
salt of I-nitroso+naphthol
in 5 ml of
glacial acetic acid with the aid of gentle heat, cool, and add water to 50 ml. The resulting suspension (kept in an amber bottle) should be shaken before use. This reagent (IO ml) should give a colourless solution when a blank determination is carried
out by the method described
below.
The column Alumina to be used for the present purpose must first be tested for suitability by passing through it a few ml of dilute solution of cobalt I-nitroso-a-naphthol in alcohol-free chloroform and washing with a few ml of alcohol-free chloroform. If the colour does not pass through freely then the stock of alumina should be spread out on a sheet of paper and allowed to stand exposed to the atmoSphere until it is found to be satisfactory (standing overnight is usually sufficient). For the column there is used a piece of tubing of S-IO mm internal diameter and 50 mm length, drawn out below into a capillary so that it will sit evenly in the neck of a IO ml graduated flask. It is filled dry with I g of alumina which is covered by a thin wad of cellulose and then with a depth of about 0.5 cm of exsiccated sodium sulphate. The presence of alcohol in ordinary chloroform has a great influence on the chromatographic adsorption of the metallic compounds. It is not, however, necessary to use alcohol-free chloroform in the determination, since the washing which forms part of the process is sufficient to remove the alcohol before the solution is passed through the column. The absorptiometric
determination
of cobalt
To about 20-50 ml of a solution, containing up to IO pg of cobalt, add 0.5 g of sodium citrate and one drop of phenolphthalein solution. After neutralisation with sodium hydroxide, add I ml of glacial acetic acid followed by 5 ml of the reagent. Boil the mixture for one minute, cool, and transfer to a separating funnel. Measure out 12 ml of chloroform for the extraction, i.e. use 4 ml for the first extraction, followReferences p. 142
G. MIDDLETON,
140
VOL. 1 (Iy&)
R. E. STUCKEY
ed by 3.3, and 2 ml. Shake each chloroform extract in turn with 10 ml of a solution containing 5 per cent of potassium hydroxide and 0.5 per cent of hydroxylamine hydrochloride and then transfer it through the column into a IO ml graduated flask. In order to get all of the cobalt into this volume care must be exercised in the sepa ration. Finally,measure
the absorption
in a I cm cell at 415 mp.The
amount
of cobalt
present can then be derived from a standardisation curve prepared using known amounts of cobalt. The relation found in the authors’ work was expressed by the equation pg Co/ml = 1.74 x Elcm Smaller Application
amounts
of cobalt can be determined
by using a 4 cm cell.
to animal tissues, etc.
In applying the I-nitroso-z-naphthol method to the mixture of inorganic compounds obtained by the destruction of the organic matter of liver and animal tissues,
interference
due to the precipitation
of compounds
such as phosphates
of
calcium and iron in weakly acid solution can be prevented by using sodium hexametaphosphate. In this case a large excess of nitrosonaphthol reagent must be used in order to convert all of the reacting metals to their nitrosonaphthol compounds. It may be difficult, especially where large quantities of animal tissue have been taken, to obtain complete extraction of the cobalt in TO ml of chloroform, and with a very large excess of reagent there is the possibility of a trace of non-metallic colour coming through the column. In such cases it is better to re-work the chloroform solution first obtained. Details are as follows: Treat the acid solution of the mineral constituents (e.g. from 50-100 g of liver) with a few drops of phenolphthalein solution and a freshly prepared solution of 5 g of sodium hexametaphosphate (flake) in IO ml of water. Add sodium hydroxide solution until the mixture is slightly alkaline, followed by acetic acid sufficient
to restore the acid reaction,
and 2 ml of glacial acetic acid in
excess. To the acid mixture (which should show not more than traces of insoluble matter) add 20 ml of reagent, bring to the boil for one minute, cool, and extract with successive portions of chloroform, washing the chloroformic extracts successively with alkaline hydroxylamine solution before passing them through a column as described above. Evaporate the combined chloroform extract to dryness, heat the residue with 2-3 drops of sulphuric acid until well charred, then heat again with the addition of a few drops of nitric acid. Cool, add a few ml of water and evaporate down to fuming. Use this residue for the final determination of cobalt as described above. Result. The recovery of cobalt from solutions of mixtures, containing calcium the phosphate (I g) and ferric sulphate (0.5 g), made up to represent approximately main residual inorganic components of liver was checked by adding known amounts of cobalt and carrying out determinations as above. The results are shown in Table I. In order to check the uniformity of the results obtained from biological material, a large quantity of commercial fresh ox liver was digested with diluted nitric acid until fully dissolved with the exception of the fat. The fatty fraction (which contains practically no cobalt), was removed and aliquot portions of the remaining nitric acid solution were mineralised by the treatment described by us previouslyz. The
VOL.
1 (1956)
DETERMINATION
cobalt was then estimated as above. ,ug of Co per IOO g of fresh liver.
OF
TRACES
The results
TABLE
OF
COBALT
141
were as follows:
4.7, 4.8, 4.4, 4.95
1
RBCOVERY OF COBALT .hDDED TO CAIXIUM PHOSPHATE (I g) AND FERRIC SULPHATE (0.5 g) ~~____. ~~ ~~ (‘0 added
(‘0 found
/i :
I’S
2.0
1.95
4.0
3.97
8.0
8.00
Results have shown that the method can be applied to the determination cobalt in all types of biological material after destruction of the organic matter Recoveries of known amounts of cobalt added to blood and urine were within limits of experimental error.
of 1’ 2. the
ACKNOWLEDGEMENT
The authors wish to thank the Directors permission to publish this work.
of The British
Drug Houses Ltd., for
SUMMARY A method is proposed for the determination of cobalt with r-nitroso-z-naphthol based on the separation of interfering metals by the use of an alumina column; iron is removed after reduction with alkaline hydroxylamine. The use of a pure reagent is essential and a method of purification is given. Absorption curves are given for the I-nitroso-z-naphthol compounds of cobalt, iron (ferric) and chromium (trivalent), and also for cobalt z-nitroso-I-naphthol. The latter compound is shown to be less suitable for absorptiometric determination of cobalt than the I : 2 isomer.
Les auteurs proposent une mCthode de dosage du cobalt par le I-nitroso-znaphtol, basCe sur la skparation des m&aux genant par l’emploi d’une colonne d’alumine; le fer est enlevC apr&s rkduction par l’hydroxylamine alcaline. L’emploi d’un rCactif pur est essentiel et une mCthode de purification est d&rite. Les auteurs donnent les courbes d’absorption des composiis du I-nitroso-2-naphtol avec le cobalt, le fer (ferrique) et le chrome (trivalent) de m&me que du composC du 2-nitrosoI-naphtol avec le cobalt. 11s montrent que ce demier compos? se p&e moins bien au dosage absorptiomCtrique du cobalt que l’isomkre I : 2. ZUSAMMENFASSUNG
Eine Methode zur Bestimmung von Kobalt mit I-Nitroso-a-naphtol wird beschrieben, welche sich auf die Abtrennung der stijrenden Metalle mit Hilfe einer Aluminiumoxyd-%ule griindet ; das Eisen wird nach Reduktion mit alkalischem Hydroxylamin entfernt. Die Anwendung eines reinen Reagenzes ist wesentlich und HefevPllces p. 142
G. MIDDLETON,
142
R. E. STUCKEY
VOL.
1 (1956)
eine Reinigungsmethode wird beschrieben. Es werden Absorptionskurven fiir die I-Nitroso-z-naphthol-Verbindungen von Kobalt, dreiwertigem Eisen und dreiwertigem Chrom, sowie fiir die z-Nitroso-I-naphtol-Verbindung von Kobalt beschrieben. Es wird gezeigt, daf.I die letztere Verbindung fur die absorptiometrische stimmung von Kobalt weniger geeignet ist als das I : 2 Isomere.
Re-
PE3IOME npcahal.aeTCa MeTO Ha
pa3,&CACHIIM.
,U,AXOIIpe~eAeHuB Ko6aAbTa
MeUIZUO~X
MeTahOB
IIpll
IIOMOl$JSl
C I-HUTpO30-2-Ha+TOAOM OKHCU
aAK,MHHHR;
)l(eAe30
OCHOBzWHbIfi )‘CTpaHSETCff
Heo6xoncmo nprnfenenne PHCTO~O pearewa, MeTod osu%eHus ~0~0p0r0 npuBonuTca. OnpeAeAeHbI ~pmbni norAo%eHuX .i+w I-HuTp030 2-Ha@oAoBbIx CoeduHeHufi K06aAbTa,XeAe3a (qeppureworo) u Tpex-BanemHoro xpoMa, a Tame w coeduHeHua Ko6aAbTac 2-HuTpo30-I-Ha~ToAoh4.IIocAe~eeco~uHeHue BBAXeTCII MeHee yao6-1M AAX a6Cop6guoHuo-MeTpusecKoro OnpeneAeHElR KoGaAbTa, PeM u30Mep I: 2. IIOCAC
BOCCTaHOBA‘2HIlSl
I&JJeAO¶HbIM
WlJpOKCIIABMEIHOM.
REFERENCES I G. MIDDLETON AND R. E. STUCKEY, Analyst, 78(195x) 532. 2 G.MIDDLETON AND R. E. STUCKEY, Analyst,79(Ig54) 138. 3 M. ILINSKI AND G. v. KNORRE, Ber., r8 (1885) 699. 4 J. A. DEAN, Anal. Chem., 23 (1951) 1096. 5 E. JENSEN, Anal. Chem. Acta, 7 (1952) 561. 6 N.D.SYLVESTER AND L.H.LAMPITT,J. SOL. Chem. Znd., Trans.,59 7 I. BELLUCCI,G~~~. chim. ital., 49 (1919) 294. 8 R. PAULAIS. Ann. pharm. Frank.. 4 (1~46) IOI. g R. PAULAIS, ibid., p. 108. IO R. PAULAIS, ibid., p. 106. II R. PAULAIS, ibid., p. 105.
Received
(1940) 57.
December
qth,
1955
THE
I35
CLINICA CHIMICA ACTA
ABSORPTIOMETRIC COBALT
WITH
DETERMINATION
OF
TRACES
OF
I-NITROSO-%NAPHTHOL by
G. MIDDLETON AND R. E. STUCKEY The British Drug Houses Ltd., City Road, London, N. I (England)
The significance of traces of cobalt in biological material has been the subject of a number of recent papers, although in most cases little attention has been paid to the details of the methods used for the determination. In attempting to employ such published methods for the determination of traces of cobalt in large quantities of, animal tissue (e.g. in liver) and in pharmaceutical preparations, it was found to be impossible to recover and determine the cobalt satisfactorily, and the procedures involved were therefore investigated more closely. An investigation of the preliminary destruction of biological tissues has already been reported’, 2 and the present paper, deals with the use of I-nitroso-a-naphthol for the absorptiometric determination of traces of cobalt. After the introduction by ILINSKI ANDVON KNORRE~O~ I-nitroso-z-naphthol as a reagent for cobalt, numerous other compounds of the same type were suggested. Fdr calorimetric determination of traces of cobalt, nitroso-R-salt has generally been preferred, as the reagent itself is stable and easily obtained in a relatively pure state; further, the reaction may be made specific for cobalt. The utility of the reaction is however, limited, since the colour cannot be concentrated and the reagent is not very stable in solution; in addition the unavoidable excess of reagent gives a yellow colour. (Recent techniques, involving the use of an alumina column for removal of excess of reagent and separation of cobalt from other metals, have considerably increased its possibilities43 “.) Unsulphonated compounds of the nitrosophenol group are all sparingly soluble in water, and offer the potential advantage that their metallic compounds may be extracted from aqueous solutions by an immiscible solvent. It would appear that one of the main reasons for their comparative obscurity lies in the difficulty of obtaining or keeping them in a sufficient state of purity. Thus SYLVESTER AND LAMPITT~ used r-nitroso-z-naphthol merely for the purpose of precipitating cobalt and other metals from the mineral constituents of a digest, followed by a dithizone procedure for the separation of cobalt and a final calorimetric determination with nitroso-R-salt. z-Nitroso-I-naphthol has also been recommended as a more sensitive reagent for cobalt than the I : 2 isomer. The use of r-nitroso-z-naphthol as a reagent both for the isolation and for the calorimetric determination of cobalt in animal products was proposed by PAULAIS*, who devised a method for the calorimetric determination of cobalt. In this method cobalt-nitrosonaphthol is extracted with chloroform from a solution of the ash, excess of reagent being eliminated by extraction with sodium hydroxide, and other metals by removal on a column of alumina. Attempts by the authors to apply this References p. 142
CC.MIDDLETON, R. E. STUCKE1-
136 method
to the determination
of cobalt
in liver and similar
VOL. 1 (1956)
materials
did not give
consistent results, but the method appeared to be promising and a more detailed study was made. The chief difficulties encountered in the application of the PAULAIS method were ir: ;.he complete removal of the iron, and also in that a low and constant blankvalue u ,js not obtained. It was ultimately found that the latter difficulty was due to the p: l:sence in the reagent of colonred impurities which were weakly adsorbed on alumina and passed into the final solution to a varying extent. The present paper describes a method, based on that of PAUL.IW, which has proved satisfactory for the determination of relatively minute amounts of cobalt in an’rnal tissues. A procedure for the removal of interfering impurities in r-nitrosoa-naphthol is given, together with a description of the behaviour of its metallic compounds
on alumina.
Interference
due to iron is eliminated
by reduction
of the
iron nitrosonaphthol compound with hydroxylamine. The method is applied to the determination of cobalt in the residue of mineral constituents obtained from animal matter
by the nitric
acid process*. EXPERIMENTAL
Comparison
of wktroso-a-naphthol
and a-nitroso-r-nafihthol
The recommendation of z-nitroso-r-naphthol is derived from the work of HEI.LUCCI~, and is only valid under certain special conditions, i.e. for the colour given b! a trace of cobalt in a large A volume of water. It is not necessarily applicable to other conditions, and observations showed that although z-nitroso-rnaphthol reacts with cobalt salts in the same manner as r-nitroso-z-naphthol, the resulting compound is less suited to absorptiometric measurcment than the I : 2 isomer, since the extinction curve for cobalt z-nitroso-r-naphthol (see Fig. I) shows a maximum of
ov-
Fig. I. hbsorption curves for metallic nitroso-naphthol derivatives. A, Co r-nitroso-r-naphthol ; R, Fe(ic) r-nitrosor-naphthol; C, Cr(ic) r-nitroso-z-naphthol; D, Co z-nitrosor-naphthol.
is about
double
Pa.wification
that
of the 2 : I compound
low intensity at 530 mp. It is therefore less suited to absorptiometric measurement than the I : z compound which ha: a well defined peak at 415 rnp at which point the absorptioii
at 530 m/l.
of I-nitroso-2-naphthoI*
A major difficulty experienced with the in the values of blank tests in which cobalt .__ -_.--
method was the variation was absent. This was ultimately
PAULAIS
found found
VOL. 1 (1956)
DETERMINATION OF TRACES OF COBALT
to be due to the presence pletely
adsorbed
and, in addition
in the reagent
by the alumina to the possibility
or aminonaphthol formed: moreover
respectively), the coloured
of coloured
impurities
‘37 which were incom-
column used. r-Nitroso-z-naphthol of simple oxidation a number impurities
or reduction
is very reactive (to nitronaphthol
of other coloured compounds may be formed in the manufacture are very
difficult to remove by recrystallisation. In the method of purification given by PAULAISg, I-nitroso-z-naphthol is dissolved in acetic acid, diluted with water and filtered; the nitroso-naphthol is extracted from the filtrate with chloroform and then extracted from the chloroform solution by dilute sodium hydroxide. This final alkaline solution is used as the reagent. This procedure does not, in our experience, produce a reagent solution of satisfactory purity, especially with some commercially available samples of I-nitroso-2-naphthol. It was found that dilution of the acetic acid solution precipitated a proportion of the r-nitroso-z-naphthol varying according to the temperature and time of standing before filtration. In one experiment the weight of the rejected precipitate corresponded to 67 per cent of the original material, only 26 per cent of the original amount being present in the solution. The first (insoluble) fraction was heavily contaminated with a red coloured substance (absorption maximum, 485 rnp) which was not a metallic compound of the reagent, and which was also present in a smaller, but still significant, amount in the second fraction. The quantity of reagent (3 mg) used by PAULAIS for a determination is too small to produce considerable errors from this cause but if other metals (e.g. zinc) are present in appreciable quantity a corresponding excess must be used as these metals compete for the reagent and the amount of coloured impurity introduced is thereby increased. It is not possible to allow for the colour produced by means of the usual blank test, since the absorption of the red impurity on the alumina is weak and variable, and an impure reagent does not give reproducible results in blank experiments. The result of a number of experiments showed that r-nitroso-e-naphthol could most readily be purified by formation of the comparatively insoluble sodium salt. A blank determination using 0.2 g of the purified sodium I-nitroso-2-naphthol showed the product to be free from the coloured metric determination of cobalt.
impurities
which interfere
An aqueous solution of sodium r-nitroso-2-naphthol, dilute for convenient use as a stock reagent. A concentrated
with absorptio-
although stable, is too solution (e.g. IO per cent)
may be made in glacial acetic acid, but this is unstable and, after a few weeks, appreciable quantities of coloured impurities are present. The most convenient reagent was found to be a suspension of I-nitroso-2-naphthol in dilute acetic acid, which can be kept in a dark bottle for several months without deterioration. The method of purification is as follows: Dissolve IO g of crude I-nitroso-z-naphthol in 300 ml of alcohol (methylated 95 per cent) and add 25 ml of 20 per cent aqueous sodium hydroxide. After allowing to stand for a few hours in a cool place, filter off the precipitated green sodium salt under reduced pressure and wash with a small amount of alcohol. Suspend the sodium salt in 2jo ml of alcohol, add 7 ml of glacial acetic acid and warm until solution is effected. Add 4 g of decolorising charcoal, boil, and filter. After cooling, add 25 ml of 20 per cent aqueous sodium hydroxide and after standing, remove the sodium salt by filtration as before and dry at about 50” C. The resulting product is a bright green crystalline powder. lirflw!lc~'.s b, rp
G. MIDDLETON,
138
R.
E.
STUCKEY
VOL.
1 (1956)
Metallic compounds of I-nitroso-2-naphthol A number of di- and tri-valent
metals form coloured compounds
with r-nitroso-o-
naphthol. These are produced directly in aqueous-generally somewhat acidsolution and may usually be extracted, together with excess of reagent, with chloroform, giving coloured solutions from which the excess of reagent may be removed by shaking with sodium hydroxide, or by treatment with alumina. From a chloroform solution the compounds formed with divalent metals are generally adsorbed on alumina, while those of trivalent metals pass through, and the Paulais method of separation of cobalt is based on this fact. The behaviour of various metals with the reagent is described below. Cobalt. The cobalt compound, prepared from a cobaltous salt, is not adsorbed on alumina, this in itself being an indication of its trivalent characterlo. This implies that in the formation of the compound the cobalt ion has been oxidised by excess of the reagent. A chloroform
solution
of the cobalt
salt of r-nitroso-2-naphthol
has an orange
red colour with an absorption maximum at 415 rnp (see Fig. I). Such a solution is stable and remains unaltered after standing for a few days in diffused light or by shaking with caustic soda or hydrochloric acid or by evaporation on a water-bath and re-solution. The colour intensity obtained from a given weight of cobalt salt was found to be proportional to the amount of cobalt present (at least up to IO lug of cobalt). Since the final solution contains only one coloured substance, methods of visual calorimetry may be used if desired. The quantitative recovery of cobalt by chloroform extraction naphthol compound was demonstrated by evaporating a chloroform
as r-nitroso-zsolution of the
cobalt compound to dryness, treating the residue with nitric and sulphuric acids to destroy the nitroso-naphthol, and repeating on this residue the process of formation of cobalt nitrosonaphthol and extraction with chloroform to the original volume. The colour obtained Chromium.
was equal to that of the original solution. It is known that chromium salts give a precipitate
with x-nitroso-
2-naphthol. It is, therefore, somewhat surprising to note that P.4ULAIS1’ specifically includes chromium (“Cr + 1I-et CrV1+“) among the metals which do not react; possibly this is because in his method it is precipitated as phosphate. If, however, a solution containing a trace of chrome alum, with sodium acetate, acetic acid and r-nitroso2-naphthol, is brought to the boil, there results a red colour which may be extracted with chloroform and which then passes through an alumina column. The chloroform solution of chromium I-nitroso-z-naphthol has an absorption maximum at 416 mp-practically the same point as cobalt (see Fig. I). Interference from chromium may be prevented by the addition of citrate to the reaction mixture. Iron. Ferric r-nitroso-a-naphthol forms a brownish solution with an absorption maximum at 408 rnp, and it also passes through alumina. In the presence of large quantities of iron the addition of phosphate (as used by PAULAIS), citrate or fluoride was not effective in ensuring the complete absence of iron from the chloroform extract. Since however the ferric complex is easily reduced to ferrous, which is absorbed on alumina, interference due to iron can be eliminated by the addition of hydroxylamine to the alkaline solution used to remove excess of the reagent. Ferrous I-nitroso-2-naphthol gives a bright green colour in chloroform solution and is thrown out as a blue precipitate on shaking this solution with dilute sodium
VOL.
1 1956)
DETERMINATION
hydroxide (though not with potassium gives a blue-green band.
OFTRACES
hydroxide).
OFCOBALT
On alumina,
I39 the ferrous compound
Iron salts can be b0t.h oxidised and reduced by an excess of the reagent so that a mixture of the ferrous and ferric complexes is always obtained. Shaking a chloroform solution with hydrochloric acid partially converts the ferrous compound into the ferric. Zinc. An insoluble brown zinc compound can easily be obtained from zinc chloride and the reagent, but it is practically insoluble in chloroform and is decomposed by sodium hydroxide, so that zinc is eliminated before reaching the column. Copper and Nickel. These give a yellow to brown solution and similar coloured bands on alumina. The solutions in chloroform are largely decolorised by shaking with caustic alkali. Other metals. sium, aluminium, of cobalt.
Coloured compounds were not obtained from manganese, magneor bismuth salts under the conditions used for the determination
METHOD
Reagent Dissolve
0.5 g of the purified
sodium
salt of I-nitroso+naphthol
in 5 ml of
glacial acetic acid with the aid of gentle heat, cool, and add water to 50 ml. The resulting suspension (kept in an amber bottle) should be shaken before use. This reagent (IO ml) should give a colourless solution when a blank determination is carried
out by the method described
below.
The column Alumina to be used for the present purpose must first be tested for suitability by passing through it a few ml of dilute solution of cobalt I-nitroso-a-naphthol in alcohol-free chloroform and washing with a few ml of alcohol-free chloroform. If the colour does not pass through freely then the stock of alumina should be spread out on a sheet of paper and allowed to stand exposed to the atmoSphere until it is found to be satisfactory (standing overnight is usually sufficient). For the column there is used a piece of tubing of S-IO mm internal diameter and 50 mm length, drawn out below into a capillary so that it will sit evenly in the neck of a IO ml graduated flask. It is filled dry with I g of alumina which is covered by a thin wad of cellulose and then with a depth of about 0.5 cm of exsiccated sodium sulphate. The presence of alcohol in ordinary chloroform has a great influence on the chromatographic adsorption of the metallic compounds. It is not, however, necessary to use alcohol-free chloroform in the determination, since the washing which forms part of the process is sufficient to remove the alcohol before the solution is passed through the column. The absorptiometric
determination
of cobalt
To about 20-50 ml of a solution, containing up to IO pg of cobalt, add 0.5 g of sodium citrate and one drop of phenolphthalein solution. After neutralisation with sodium hydroxide, add I ml of glacial acetic acid followed by 5 ml of the reagent. Boil the mixture for one minute, cool, and transfer to a separating funnel. Measure out 12 ml of chloroform for the extraction, i.e. use 4 ml for the first extraction, followReferences p. 142
G. MIDDLETON,
140
VOL. 1 (Iy&)
R. E. STUCKEY
ed by 3.3, and 2 ml. Shake each chloroform extract in turn with 10 ml of a solution containing 5 per cent of potassium hydroxide and 0.5 per cent of hydroxylamine hydrochloride and then transfer it through the column into a IO ml graduated flask. In order to get all of the cobalt into this volume care must be exercised in the sepa ration. Finally,measure
the absorption
in a I cm cell at 415 mp.The
amount
of cobalt
present can then be derived from a standardisation curve prepared using known amounts of cobalt. The relation found in the authors’ work was expressed by the equation pg Co/ml = 1.74 x Elcm Smaller Application
amounts
of cobalt can be determined
by using a 4 cm cell.
to animal tissues, etc.
In applying the I-nitroso-z-naphthol method to the mixture of inorganic compounds obtained by the destruction of the organic matter of liver and animal tissues,
interference
due to the precipitation
of compounds
such as phosphates
of
calcium and iron in weakly acid solution can be prevented by using sodium hexametaphosphate. In this case a large excess of nitrosonaphthol reagent must be used in order to convert all of the reacting metals to their nitrosonaphthol compounds. It may be difficult, especially where large quantities of animal tissue have been taken, to obtain complete extraction of the cobalt in TO ml of chloroform, and with a very large excess of reagent there is the possibility of a trace of non-metallic colour coming through the column. In such cases it is better to re-work the chloroform solution first obtained. Details are as follows: Treat the acid solution of the mineral constituents (e.g. from 50-100 g of liver) with a few drops of phenolphthalein solution and a freshly prepared solution of 5 g of sodium hexametaphosphate (flake) in IO ml of water. Add sodium hydroxide solution until the mixture is slightly alkaline, followed by acetic acid sufficient
to restore the acid reaction,
and 2 ml of glacial acetic acid in
excess. To the acid mixture (which should show not more than traces of insoluble matter) add 20 ml of reagent, bring to the boil for one minute, cool, and extract with successive portions of chloroform, washing the chloroformic extracts successively with alkaline hydroxylamine solution before passing them through a column as described above. Evaporate the combined chloroform extract to dryness, heat the residue with 2-3 drops of sulphuric acid until well charred, then heat again with the addition of a few drops of nitric acid. Cool, add a few ml of water and evaporate down to fuming. Use this residue for the final determination of cobalt as described above. Result. The recovery of cobalt from solutions of mixtures, containing calcium the phosphate (I g) and ferric sulphate (0.5 g), made up to represent approximately main residual inorganic components of liver was checked by adding known amounts of cobalt and carrying out determinations as above. The results are shown in Table I. In order to check the uniformity of the results obtained from biological material, a large quantity of commercial fresh ox liver was digested with diluted nitric acid until fully dissolved with the exception of the fat. The fatty fraction (which contains practically no cobalt), was removed and aliquot portions of the remaining nitric acid solution were mineralised by the treatment described by us previouslyz. The
VOL.
1 (1956)
DETERMINATION
cobalt was then estimated as above. ,ug of Co per IOO g of fresh liver.
OF
TRACES
The results
TABLE
OF
COBALT
141
were as follows:
4.7, 4.8, 4.4, 4.95
1
RBCOVERY OF COBALT .hDDED TO CAIXIUM PHOSPHATE (I g) AND FERRIC SULPHATE (0.5 g) ~~____. ~~ ~~ (‘0 added
(‘0 found
/i :
I’S
2.0
1.95
4.0
3.97
8.0
8.00
Results have shown that the method can be applied to the determination cobalt in all types of biological material after destruction of the organic matter Recoveries of known amounts of cobalt added to blood and urine were within limits of experimental error.
of 1’ 2. the
ACKNOWLEDGEMENT
The authors wish to thank the Directors permission to publish this work.
of The British
Drug Houses Ltd., for
SUMMARY A method is proposed for the determination of cobalt with r-nitroso-z-naphthol based on the separation of interfering metals by the use of an alumina column; iron is removed after reduction with alkaline hydroxylamine. The use of a pure reagent is essential and a method of purification is given. Absorption curves are given for the I-nitroso-z-naphthol compounds of cobalt, iron (ferric) and chromium (trivalent), and also for cobalt z-nitroso-I-naphthol. The latter compound is shown to be less suitable for absorptiometric determination of cobalt than the I : 2 isomer.
Les auteurs proposent une mCthode de dosage du cobalt par le I-nitroso-znaphtol, basCe sur la skparation des m&aux genant par l’emploi d’une colonne d’alumine; le fer est enlevC apr&s rkduction par l’hydroxylamine alcaline. L’emploi d’un rCactif pur est essentiel et une mCthode de purification est d&rite. Les auteurs donnent les courbes d’absorption des composiis du I-nitroso-2-naphtol avec le cobalt, le fer (ferrique) et le chrome (trivalent) de m&me que du composC du 2-nitrosoI-naphtol avec le cobalt. 11s montrent que ce demier compos? se p&e moins bien au dosage absorptiomCtrique du cobalt que l’isomkre I : 2. ZUSAMMENFASSUNG
Eine Methode zur Bestimmung von Kobalt mit I-Nitroso-a-naphtol wird beschrieben, welche sich auf die Abtrennung der stijrenden Metalle mit Hilfe einer Aluminiumoxyd-%ule griindet ; das Eisen wird nach Reduktion mit alkalischem Hydroxylamin entfernt. Die Anwendung eines reinen Reagenzes ist wesentlich und HefevPllces p. 142
G. MIDDLETON,
142
R. E. STUCKEY
VOL.
1 (1956)
eine Reinigungsmethode wird beschrieben. Es werden Absorptionskurven fiir die I-Nitroso-z-naphthol-Verbindungen von Kobalt, dreiwertigem Eisen und dreiwertigem Chrom, sowie fiir die z-Nitroso-I-naphtol-Verbindung von Kobalt beschrieben. Es wird gezeigt, daf.I die letztere Verbindung fur die absorptiometrische stimmung von Kobalt weniger geeignet ist als das I : 2 Isomere.
Re-
PE3IOME npcahal.aeTCa MeTO Ha
pa3,&CACHIIM.
,U,AXOIIpe~eAeHuB Ko6aAbTa
MeUIZUO~X
MeTahOB
IIpll
IIOMOl$JSl
C I-HUTpO30-2-Ha+TOAOM OKHCU
aAK,MHHHR;
)l(eAe30
OCHOBzWHbIfi )‘CTpaHSETCff
Heo6xoncmo nprnfenenne PHCTO~O pearewa, MeTod osu%eHus ~0~0p0r0 npuBonuTca. OnpeAeAeHbI ~pmbni norAo%eHuX .i+w I-HuTp030 2-Ha@oAoBbIx CoeduHeHufi K06aAbTa,XeAe3a (qeppureworo) u Tpex-BanemHoro xpoMa, a Tame w coeduHeHua Ko6aAbTac 2-HuTpo30-I-Ha~ToAoh4.IIocAe~eeco~uHeHue BBAXeTCII MeHee yao6-1M AAX a6Cop6guoHuo-MeTpusecKoro OnpeneAeHElR KoGaAbTa, PeM u30Mep I: 2. IIOCAC
BOCCTaHOBA‘2HIlSl
I&JJeAO¶HbIM
WlJpOKCIIABMEIHOM.
REFERENCES I G. MIDDLETON AND R. E. STUCKEY, Analyst, 78(195x) 532. 2 G.MIDDLETON AND R. E. STUCKEY, Analyst,79(Ig54) 138. 3 M. ILINSKI AND G. v. KNORRE, Ber., r8 (1885) 699. 4 J. A. DEAN, Anal. Chem., 23 (1951) 1096. 5 E. JENSEN, Anal. Chem. Acta, 7 (1952) 561. 6 N.D.SYLVESTER AND L.H.LAMPITT,J. SOL. Chem. Znd., Trans.,59 7 I. BELLUCCI,G~~~. chim. ital., 49 (1919) 294. 8 R. PAULAIS. Ann. pharm. Frank.. 4 (1~46) IOI. g R. PAULAIS, ibid., p. 108. IO R. PAULAIS, ibid., p. 106. II R. PAULAIS, ibid., p. 105.
Received
(1940) 57.
December
qth,
1955