ISSN:0003-2654    

DOI:10.1039/AN98207FX025

出版商:RSC    

年代:1982

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98207BX027

出版商:RSC    

年代:1982

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98207FP065

出版商:RSC    

年代:1982

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98207BP071

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

July 1982 Vol. 107 No. 1276 The Analyst The Analysis of Cemented Carbides A Review R. S. Young Consulting Chemical Engineer, 1178 Beach Drive, Victoria, B.C., Canada, VSS 2M9 Summary of Contents Introduction Sample Preparation Sample Decomposition Determinations Aluminium Boron Carbon Chromium Cobalt Copper Iron Manganese Molybdenum Nickel Niobium Nitrogen Silicon Tantalum Titanium Tungsten Vanadium Zirconium Keywords : Review ; cemented carbides Introduction The chemical analysis of cemented carbides or “hard metals” is not an easy task. Difficulties commence with the preparation of the sample, because this involves the comminution of a material possessing a hardness exceeded only by that of diamond. Tungsten carbide, with cobalt as the binder, was the first cemented carbide to be developed, and is still of prime importance.For certain purposes, however, tungsten is often partly replaced with chromium, molybdenum, niobium, tantalum, titanium, vanadium or zirconium; with some carbides, nickel is used as the binder instead of cobalt. The cemented carbides may also, invariably or frequently, contain the following elements : aluminium, carbon, copper, iron, manganese, nitrogen and silicon. The complexity of the carbides, their resis- tance to dissolution by acid attack or alkaline fusion and the tendency of many constituents to hydrolyse or to divide when subjected to conventional separation procedures increase the difficulties of analysis. The sensitivity of niobium, tantalum, titanium, tungsten and zirconium towards atomic- absorption spectrophotometry is poor and they are prone to complex interferences.Aluminium, boron, tantalum and tungsten are unsuitable for determination by routine X-ray fluorescence spectroscopy, and the sensitivity of tantalum for optical spectrography is low. Volumetric methods have limited application for aluminium, niobium, silicon, tantalum, tungsten and zirconium in cemented carbides and are rarely used in commercial practice. The complexity of many carbides usually makes imperative several separations before a 72 1722 YOUNG: ANALYSIS OF CEMENTED CARBIDES Analyst, VoL. I07 colorimetric procedure can be carried out. Very few elements in carbides are amenable to determination by electrodeposition, polarographic, potentiometric or ion-selective electrode procedures. Sample Preparation The exceptional hardness of cemented carbides requires that the preparation of a sample be carried out with great care.A tungsten carbide mortar and a carbide-tipped pestle are generally employed, the latter usually driven by a small riveting hammer. All samples are crushed to pass a 100-mesh sieve, and many laboratories pulverise to -200 mesh. From some carbide materials it may be convenient initially to obtain representative sections by means of a diamond saw. Sample Decomposition Decomposition of cemented carbides can be effected by several methods. When the sample in a platinum dish is treated with hydrofluoric acid on a steam-bath, and nitric acid is added dropwise, complete dissolution will finally occur. Addition of perchloric acid, and heating to strong fumes of the latter, will eliminate hydrofluoric and nitric acids. Careful fusion with dry sodium peroxide in a nickel or zirconium crucible, followed by treatment of the cooled melt in hydrochloric acid, is another excellent decomposition pro- cedure.To avoid the vigorous reaction between sodium peroxide and some carbides, sintering, which can be carried out at a lower temperature, is often employed. Water is frequently used instead of hydrochloric acid to give a precipitate of cobalt, copper, iron, manganese, nickel, titanium and zirconium, leaving aluminium, chromium, molybdenum, tungsten and vanadium in solution; niobium and tantalum will be found in both the precipi- tate and solution. Carbides can be decomposed by fusion with sodium or potassium pyrosulphate in a silica flask, followed by dissolution of the cooled melt in ammonium citrate solution or in sulphuric acid.1 Boiling with a mixture of perchloric, phosphoric and sulphuric acids will also dissolve carbides.Boron carbide can be decomposed by sodium carbonate fusion in a platinum crucible, or by prolonged fusion with potassium persulphate in a silica flask. Decomposition of silicon carbide can be effected by sodium carbonate fusion in a platinum crucible; it is preferable to coat the walls of the crucible with the flux prior to the decomposition. Silicon carbide is not attacked by a mixture of hydrofluoric and nitric acids or by fusion with potassium pyrosulphate. This behaviour can provide a method of separating it from other components.Determinations If the composition of a carbide is unknown to the analyst, an initial qualitative examina- tion of the sample by optical emission spectrography or X-ray fluorescence is, if not manda- tory, at least most desirable. Aluminium Aluminium is not a major constitutent of cemented carbides, but a few types contain a small amount and a determination may be required. Separations are important in this analysis, and can be summarised as follows.2 Fusion with sodium peroxide followed by boiling with water will leave aluminium in solution, with chromium, molybdenum, tungsten, vanadium and part of the niobium and tantalum. A cupferron - chloroform separation in cold 10% sulphuric or hydrochloric acid will leave aluminium in the aqueous layer with chromium, cobalt, manganese, nickel and part of the copper.Separation with a mercury cathode in 0.3 N sulphuric acid will leave aluminium in solution, together with niobium, tantalum, tungsten, vanadium, zirconium and part of the manganese. Precipitation of aluminium by ammonia solution and ammonium chloride separates it from manganese ; copper and molybdenum are removed from aluminium with hydrogen sulphide in 0.2-0.5 N hydrochloric or sulphuric acid. When aluminium has been isolated from interfering elements it is usually precipitated as the hydroxide in ammoniacal solution and ignited to the oxide3; precipitation may also beJdy, 1982 A REVIEW 723 carried out in acetic acid - acetate solution with quinolin-8-01, weighing at 135 "C as the aluminium quinolate.4 Small amounts of aluminium, after isolation, can be determined spectrophotometrically with aluminon; the final aliquot should contain less than 0.1 mg of aluminium in 50 ml of solution.No interference from trace amounts of other elements accompanying aluminium in the usual separations of the latter in carbides is en~ountered.~ Atomic-absorption spectrophotometry can be used to determine aluminium, using the 309.3-nm line and a dinitrogen oxide - acetylene flame; about 1000 pg ml-l of sodium or potassium should be added to the test and standard solutions.&8 Optical spectrography, using the sensitive lines at 308.22, 309.27, 394.40 and 396.15 nm, can be used for small amounts of alurnini~m.~,~,~ Boron After fusion of boron carbide with sodium carbonate, the sample is acidified, treated with methanol and boiled to volatilise all of the boron as the methyl borate ester.The latter is caught and saponified in alkaline solution, the alcohol expelled and the boric acid titrated with standard sodium hydroxide solu t ion4 9 lo or determined spec t ropho t omet ricall y with curcumin, carminic acid or q~inalizarin.~,~ Boron can be determined by atomic-absorption spectrophotometry, using the mean sensitivity of 249.68- and 249.77-nm lines, with a fuel-rich dinitrogen oxide - acetylene flame. Small amounts of boron (10-500 p.p.m.) can be determined satisfactorily by optical spe~trography.~ The sensitivity of boron, however, is low.6 Carbon Because small changes in the carbon content of cemented carbides may significantly alter the properties of these materials, the determination of carbon is very important.The gravimetric procedure, involving the combustion of the sample in a stream of oxygen and collection and weighing of the carbon dioxide evolved, is almost invariably used. Most 200-mesh samples burn readily at about 985 "C without accelerators, but for 100-mesh material it is customary to cover the carbide with 2 g of iron drillings and sprinkle a layer of well ignited copper(I1) oxide over the top of the charge before heating at 1100-1 150 "C. The theoretical carbon contents of the carbides of iron, manganese, molybdenum and tantalum range from 6.2 to 12.5y0, those of chromium, niobium, and zirconium vary from 11.4 to 13.3y0, whereas the carbides of boron, silicon, titanium and vanadium have a carbon content of approxi- mately 19-3OyO.The determination of free carbon is sometimes required. To 1-5 g of carbide in a platinum dish add 15 ml of nitric acid, warm and add a few drops of hydrofluoric acid until solution is complete. Add 5 ml of orthophosphoric acid and heat until most of the hydrofluoric acid has been removed. Add 15 ml of a saturated boric acid solution and warm for several minutes. Wash thoroughly with hot water and dry in an oven at 105 "C for 1 h. Determine the carbon by combustion in the usual manner, correcting for any carbon in the asbestos by a blank run. Pure tungsten carbide has a carbon content of approximately 6.1%. Complex carbides, therefore, may have a wide range of carbon contents. Filter on a thin layer of ignited asbestos in a perforated platinum boat.Chromium A sodium peroxide fusion of a carbide, followed by a water leach of the cooled melt, will leave chromium in solution with aluminium, molybdenum, tungsten, vanadium and part of the niobium and tantalum. A cupferron - chloroform separation in cold 10% sulphuric or hydrochloric acid will leave chromium in the aqueous layer with aluminium, cobalt, man- ganese, nickel and part of the copper. Copper and molybdenum can be removed from chromium by hydrogen sulphide in 0.2-0.5 N sulphuric or hydrochloric acid. Aluminium can be separated from chromium with ammonia and ammonium chloride if the chromium is fully oxidised to the chromate state.2 The standard volumetric procedure using ammonium iron( 11) sulphate and potassium permanganate, or the spectrophotometric diphenylcarbazide method, can be employed for carbides after the separations mentioned above.4J1724 YOUNG: ANALYSIS OF CEMENTED CARBIDES Analyst, Vol.I07 Using the 357.9-nm line and either an air - acetylene or dinitrogen oxide - acetylene flame, chromium can be determined by atomic-absorption spectrophotometry. The presence of iron, nickel or vanadium causes chemical interferences ; 1000 pg ml-l of potassium should be added to both test and standard solutions to suppress ionisation.s-* Small amounts of chromium (1-2000 p.p.m.) can be determined by optical spectrography; the element has a high sensitivity and a number of useful emission lines, such as those at 425.4, 427.5 and 428.9 nm.9 X-ray fluorescence, using a lithium - sodium borate disc, can be employed for chromium in carbides.l*l2 Cobalt Cobalt, in concentrations ranging from approximately 3 to 25%, is the binding agent for all tungsten carbides and for all hard metals in which tungsten is the major carbide.Its determination is accordingly a frequent and important one in this industry. Cobalt can be separated from copper and molybdenum using hydrogen sulphide in 0.2- 0.5 N hydrochloric or sulphuric acid. After a sodium peroxide fusion of the carbide and boiling with water, cobalt remains as a precipitate accompanied by copper, iron, manganese, nickel, titanium, zirconium and part of the niobium and tantalum. A cupferron - chloro- form separation yields cobalt in the aqueous layer with aluminium, chromium, manganese, nickel and part of the copper.After prior removal of copper and molybdenum, a zinc oxide separation leaves cobalt in the filtrate, accompanied by only manganese and nickel. In an ammoniacal solution containing ammonium citrate, phenylthiohydantoic acid will precipitate cobalt; the latter will be accompanied by copper and part of the iron and nickel. Cobalt is separated from nickel by the familiar dimethylglyoxime or 1-nitrosonaphth-2-01 pro- cedures for nickel and cobalt, respectively. A favourite procedure for the determination of cobalt in carbides is a potentiometric method, based on the oxidation of cobalt in ammoniacal solution with potassium hexa- cyanoferrate(II1) . l v 4 Manganese is the only interfering element, and it can be removed initially by treatment with sodium chlorate and nitric acid, or cobalt can be separated with phen ylthiohydan toic acid.The standard 1-nitrosonaphth-2-01 precipitation of cobalt, following hydrogen sulphide and zinc oxide separations, is a very reliable procedure for all types of carbides. Cobalt is sometimes determined by electrodeposition of cobalt plus nickel in ammoniacal solution, with subsequent separation and determination of either element. A low cobalt content can often be advantageously determined by a spectrophotometric m e t h ~ d . ~ It is seldom that the amount of chromium, manganese, nickel or vanadium in a carbide sample will exceed the wide tolerance of the nitroso-R-salt procedure for these elements4 ; if so, cobalt can be readily isolated with 1-nitrosonaphth-2-01.No element normally found in cemented carbides will interfere in the excellent ammonium thiocyanate photometric met hod.4 Cobalt can be determined readily by atomic-absorption spectrophotometry, using the 240.7-nm line and an air - acetylene flame. Even high concentrations of chromium, copper, iron, manganese, molybdenum, nickel, silicon, titanium, tungsten and vanadium do not interfere.+8 For low contents of cobalt (10-1 000 p.p.m.) the spectrum of this element has a number of useful emission lines, and optical spectrography is a valuable analytical techniq~e.~ X-ray fluorescence can be used to determine cobalt, but chromium, iron, molybdenum and nickel give difficulties with this technique when it is applied to complex carbides. Copper Copper is not a major component of cemented carbides, but it is found in a few types and its determination may be required.Among the elements normally present in carbides, copper and molybdenum are the only ones precipitated by hydrogen sulphide in 0.2--0.5 N hydrochloric or sulphuric acid; molyb- denum sulphide is soluble in ammonium or sodium sulphide solution. The addition of tartaric or citric acid to the sample before the introduction of hydrogen sulphide will prevent the coprecipitation of certain elements such as niobium, tantalum, tungsten and vanadium.July, 1982 A REVIEW 725 After isolation of copper and molybdenum, and separation of the latter, the familiar volumetric iodide - thiosulphate method for copper4J1 is simple and reliable for this element in carbides.For a very low copper content, the spectrophotometric 2,2’-biquinoline procedure can be conveniently carried out on the copper sulphide precipitate after removal of m~lybdenum.~~~ Copper can be determined by atomic-absorption spectrophotometry, using the 324.8-nm line and an air - acetylene flame. Even a large excess of aluminium, chromium, cobalt, iron, manganese, nickel and vanadium in carbides has little effect on its absorption. Copper is an element with a high sensitivity to spectrochemical tests, and small amounts (10-2000 p.p.m.) can be readily determined by optical spe~trography.~ X-ray fluorescence can be used to determine copper over a wide range.l2 Iron Because iron is usually determined, after precipitation with ammonia solution, by the volumetric dichromate method, in which vanadium is the only interfering element, separa- tions are normally few and simple for an iron analysis. If vanadium is a component of a carbide sample, fusion with sodium peroxide, or subsequent treatment with sodium hydroxide if an acid dissolution is employed, will remove vanadium from iron.The volumetric dichromate procedure for iron is rapid and reliable, and is applicable to a wide range of iron ~ontents.~Jl For a low content of iron, the usual separations mentioned above will yield a solution ready for the 1 ,lo-phenanthroline spectrophotometric m e t h ~ d . ~ ~ ~ Atomic-absorption spectrophotometry can be used to determine iron ; the 248.3-nm line and an air - acetylene flame are employed.6-* Iron has a large number of spectral lines, and 5-5000 p.p.m.of this element in carbides can be determined by optical spe~trography.~ X-ray fluorescence can be used to determine iron, but for carbides it does not appear to be a popular technique. Manganese For carbides, manganese is generally determined by the sodium bismuthate volumetric method or the potassium periodate spectrophotometric p r ~ c e d u r e . ~ , ~ ~ In the former, the interfering chromium and vanadium are removed by sodium peroxide fusion, and cobalt and nickel by the isolation of manganese with sodium chlorate and nitric acid. In the spectro- photometric method, the coloured ions chromium, cobalt, copper and nickel interfere; chromium can be volatilised by boiling with perchloric and hydrochloric acids, copper can be removed with hydrogen sulphide in acidic solution and cobalt and nickel are separated from manganese by precipitation of the latter with sodium chlorate and nitric acid.The volumetric sodium bismuthate method is applicable to a wide range of concentrations of manganese, and the spectrophotometric potassium periodate procedure is thoroughly reliable for a low content of manganese in carbides. Using the 279.5-nm line and an air - acetylene flame, manganese in carbides can be deter- mined by atomic-absorption spectrophotometry. Even large excesses of chromium, cobalt, copper, iron, molybdenum, nickel and tungsten have little effect. Small amounts of manganese (44000 p.p.m.) can be determined by optical spectro- g r a p h ~ . ~ X-ray fluorescence can be employed for manganese, but its use for carbides does not seem to be widespread.Molybdenum In cemented carbides, molybdenum and copper are normally the only elements that are precipitated by hydrogen sulphide in 0.2-0.5 N hydrochloric or sulphuric acid. The presence of a small amount of tartaric acid in the hydrogen sulphide treatment will prevent the coprecipitation of elements such as niobium, tantalum, tungsten and vanadium. Molyb- denum is separated from copper by the solubility of the former sulphide in alkaline or ammonium sulphide solution. When molybdenum is thus isolated, it can be determined gravimetrically by the cc-benzoin726 YOUNG: ANALYSIS OF CEMENTED CARBIDES Analyst, Vol. 107 oxime method, or for a low content by the spectrophotometrk thiocyanate - tin(I1) chloride procedure .4 ,14 Atomic-absorption spectrophotometry can be used in molybdenum analysis, employing the 313.3-nm line and a dinitrogen oxide - acetylene flame.Interferences from some con- stituents of carbides can be overcome by adding 2% m/V ammonium chloride solution to both the standard and test solutions.s Molybdenum has a high sensitivity to optical spectrography, and it can be determined in carbides in the parts per million range by this te~hnique.9,~~ A wide range of molybdenum contents in carbides can be determined by X-ray fluores- cence.14 Nickel Nickel, like cobalt, serves as a binder for carbide particles, and its determination is a frequent and important one in the analysis of many cemented carbides. Nickel can be separated from copper and molybdenum by hydrogen sulphide in 0.2-0.5 N hydrochloric or sulphuric acid.If the carbide has been decomposed by sodium peroxide fusion, boiling with water leaves nickel as a precipitate with cobalt, copper, iron, manganese, titanium, zirconium and part of the niobium and tantalum. A cupferron - chloroform separation in cold 10% sulphuric or hydrochloric acid places nickel in the aqueous layer with aluminium, chromium, cobalt, manganese and part of the copper. Nickel is separated from cobalt by the conventional dimethylglyoxime or 1-nitrosonaphth-2-01 methods for nickel and cobalt, respectively. The gravimetric determination of nickel over a wide concentration range by dimethyl- glyoxime is one of the most reliable procedures in analytical chemistry, except in the presence of a high content of both cobalt and iron.4 The latter element can be quickly removed by diethyl ether extraction and cobalt separated with 1-nitrosonaphth-2-01. Occasionally nickel is found by the electrodeposition of cobalt and nickel in ammoniacal solution, with subsequent separation and determination of one of the elements.The colour produced by dimethylglyoxinie and the nickel(I1) ion furnishes the most convenient and reliable spectrophotometric procedure for a low content of nickel in carbides4p5 Atomic-absorption spectrophotometry can be used for the determination of nickel in carbides, employing the 232.0-nm line and an air - acetylene flame. The absorption of nickel is hardly affected by even a large excess of most elements. Optical spectrography can be applied to the determination of 10-1000 p.p.m.of nickel in carbidesg Nickel can be determined over a wide concentration range in cemented carbides by X-ray fluorescence.12 Niobium Copper and molybdenum are separated from niobium and tantalum by saturating with hydrogen sulphide a solution of the fused hydrogen sulphate sample in 20 ml of 20% tartaric acid containing 5 ml of 10% sulphuric acid. In a cupferron - chloroform separation in cold 10% sulphuric or hydrochloric acid, niobium is extracted into the chloroform layer together with iron, molybdenum, tantalum, titanium, tungsten, vanadium, zirconium and part of the copper. A mercury cathode separation in 0.3 N sulphuric acid will leave niobium in solution, accompanied by aluminium, tantalum, titanium, tungsten, vanadium and zir- conium.When a sample solution containing hydrochloric acid - hydrofluoric acid - water (5 + 4 + 11) is passed through a column of Dowex 1 or similar resin, niobium and tantalum are retained whereas virtually all other elements pass through the column. Niobium is eluted with 14% ammonium chloride - 4% hydrofluoric acid solution. After separation of niobium by ion exchange as outlined in the preceding paragraph, it can be precipitated with cupferron in cold 10% sulphuric or hydrochloric acid, carefully ignited in a platinum crucible at 1100 "C, and weighed as Nb205.4915 Niobium may also be determined volumetrically by passing the sample through a Jones reductor and titrating with standard potassium permanganat e solution.July, 1982 A REVIEW 727 Small amounts of niobium can be determined spectrophotometrically, using hydro- quinone or hydrogen peroxide or by extraction of its thiocyanate complex in hydrochloric acid with diethyl ether.lv4p5,l5 Neither niobium nor tantalum is really suitable for determination by atomic-absorption spectrophotometry ; both suffer from complex interference effects and low sensitivity.Optical spectrography and X-ray fluorescence have occasionally been employed for the determination of niobium in a few carbides, but their use is not general. Nitrogen The determination of nitrogen in cemented carbides is carried out by the familiar Kjeldahl method. The sample is dissolved in a mixture of sulphuric acid and potassium sulphate, with the aid of hydrochloric or perchloric acid if necessary; about 0.5 g of mercury(I1) oxide is added as a catalyst.1 Digestion is followed by neutralisation with sodium hydroxide, distillation of ammonia into a measured amount of standard sulphuric acid, and titration of the excess of acid with standard sodium hydroxide solution.Nitrogen is usually very low in carbides, and unless the latter may contain a small amount of a refractory nitride the determination of nitrogen is often omitted. Silicon The traditional separation of silicon by dehydration with hydrochloric, perchloric or sulphuric acid remains the most effective procedure. In carbides, most of the niobium, tantalum and tungsten will accompany the precipitate of silica. The determination of silicon is usually carried out gravimetrically by filtering and weighing in a platinum crucible the impure precipitate ignited at 1200 "C, adding 10 ml of hydro- fluoric acid and a few drops of sulphuric acid, evaporating to dryness, igniting again at 800 "C and re-weighing.The difference in masses represents s i l i ~ a . ~ ~ ~ ~ , ~ ~ The reason for the initial weighing of the crude silica after heating t o 1200 "C and the final weighing after ignition at 800 "C is that tungsten(V1) oxide becomes volatile at 850 "C. The initial temperature is essential for the complete dehydration of silica, and the final ignition is below the temperature at which tungsten commences to volatilise. Very small amounts of silica can be determined spectrophotometrically by the molybdo- disilicic acid method or the molybdenum blue p r o c e d ~ r e .~ ~ ~ ~ Using the 251.6-nm line and the dinitrogen oxide - acetylene flame, silicon can be deter- mined by atomic-absorption spectrophotometry.6 Optical spectrography and X-ray fluorescence are sometimes employed to measure silicon in carbides.9,12,16 Tantalum After elution of niobium from a Dowex 1 resin column with 14% ammonium chloride - 4% hydro- fluoric acid solution, tantalum is finally eluted with 14% ammonium chloride - 4% hydro- fluoric acid solution adjusted to pH 5.5 with ammonia solution. After isolation of tantalum as described above, it can be precipitated by cupferron in cold 10% sulphuric or hydrochloric acid, carefully ignited in a platinum crucible at 1100 "C and weighed as Ta,05.4115 Low contents of tantalum, after isolation, can be measured spectrophotometrically using pyrogallol in ammoniacal citrate - ammonium oxalate s o l ~ t i o n .~ ~ ~ ~ ~ ~ Like niobium, tantalum is subject t o difficulties when its determination in carbides is attempted by atomic-absorption spectrophotometry, optical spectrography or X-ray fluorescence. Most of the separations of tantalum have been outlined previously under Niobium. Titanium Because many cemented carbides contain 50% or more of titanium carbide, the deter- mination of titanium is important in this industry. Copper and molybdenum can be removed from titanium by hydrogen sulphide in 0.2-0.5 N hydrochloric or sulphuric acid. A sodium peroxide fusion and subsequent boiling with water leaves titanium as a precipitate, accompanied by cobalt, copper, iron, manganese, nickel and728 YOUNG: ANALYSIS OF CEMENTED CARBIDES Analyst, Vol.107 part of the niobium and tantalum. A mercury cathode separation in approximately 0.3 N sulphuric acid leaves in solution titanium, together with aluminium, niobium, tantalum, tungsten, vanadium, zirconium and part of the manganese. Extraction with cupferron - chloroform in cold 10% sulphuric or hydrochloric acid will yield titanium in the chloroform layer, with iron, molybdenum, niobium, tantalum, tungsten, vanadium, zirconium and part of the copper. Treatment with ammonia solution and ammonium chloride will precipitate titanium, accompanied by aluminium, chromium, iron, niobium, tantalum, part of the vanadium and trace amounts of cobalt, copper, molybdenum, nickel and tungsten.Titanium is conveniently separated from niobium and tantalum in hydrochloric acid - hydrofluoric acid - water (5 + 4 + 11) by passage through an ion-exchange column of Dowex 1 or a similar resin. Iron can often be advantageously separated from titanium by diethyl ether extraction. Titanium in carbides, even in samples containing 50% of titanium carbide, can be deter- mined by utilising the yellow to orange colour of titanium sulphate in the presence of hydrogen peroxide. Interferences are chromium, cobalt, copper, iron, molybdenum, nickel, niobium and vanadium, because their solutions are coloured or they form coloured com- pounds with hydrogen per0xide.l 9495911 Titanium can also be determined volumetrically, involving the reduction from the quadri- valent state by amalgamated zinc, lead, cadmium or aluminium ; interferences in carbides are chromium, copper, molybdenum, niobium, tantalum, tungsten and vanadium.4911 Although titanium can be determined by atomic-absorption spectrophotometry, using the 364.3-nm line and a dinitrogen oxide - acetylene flame, the element is rather insensitive and is prone to complex interferences.Tungsten Tungsten carbide, either singly or in combination with one or more other carbides, is by far the most important hard metal; a tungsten determination is therefore mandatory in virtually every cemented carbide. The various grades of tungsten carbide usually contain approximately 70-90y0 of tungsten. Copper and molybdenum can be separated from tungsten with hydrogen sulphide in 0.2- 0.5 N sulphuric or hydrochloric acid; addition of tartaric acid is advisable to prevent co- precipitation of tungsten.A sodium peroxide fusion and subsequent boiling in water leaves tungsten in solution with aluminium, chromium, molybdenum, vanadium and part of the niobium and tantalum. A mercury cathode separation in 0.3 N sulphuric acid yields tungsten in solution, accompanied by aluminium, niobium, tantalum, titanium, vanadium and zirconium. In the cupferron - chloroform separation in cold 10% sulphuric or hydro- chloric acid, tungsten is extracted in the chloroform layer with iron, molybdenum, niobium, tantalum, titanium, vanadium, zirconium and part of the copper. Tungsten can be determined gravimetrically by the conventional acid digestion - cinchonine procedure, which is reliable but lengthy.4911v12,17 Amounts of tungsten up to 0.15g can be measured spectrophotometricaly by a method that depends on the fact that when potassium thiocyanate and tin(I1) chloride in sulphuric - hydrochloric acid are added to a tungstate solution, a yellow colour derived from a tungsten(V) complex slowly forms.194s5917918 By the use of suitable aliquots, a wide range of tungsten contents can be determined by this procedure.In carbides this method may be subject to interferences from molybdenum, niobium, vanadium, appreciable amounts of iron and also from chromium, cobalt, copper and nickel, which give coloured ions ; appropriate separations are accordingly necessary. For most carbides, tungsten is separated by sodium peroxide fusion followed by a water leach and filtration.In solution with the tungsten, the only elements that interfere in the spectrophotometric method are chromium, molybdenum, vanadium and part of the niobium. If chromium is present it can be initially volatilised as chromyl chloride. In the presence of a fluoride, cupferron will not precipitate tungsten, thereby affording a separation from niobium and vanadium. Large amounts of molybdenum should be removed with hydrogen sulphide, but small amounts can be tolerated because molybdenum thiocyanate fades rapidly in strong acid. A tomic-absorption spect rop ho t ome t ry can be u t ilised to determine tungsten, employingJuly, 1982 A REVIEW 729 the 255.1-nm line and a very fuel-rich dinitrogen oxide - acetylene flame, but the sensitivity is not high.Tungsten is occasionally determined by optical spectrography and X-ray flu~rescence.~~J~ Vanadium Copper and molybdenum can be separated from vanadium with hydrogen sulphide in 0.2-0.5 N hydrochloric or sulphuric acid; the addition of tartaric acid will prevent the co- precipitation of vanadium. Sodium peroxide fusion, followed by boiling with water, will leave vanadium in solution, accompanied by aluminium, chromium, molybdenum, tungsten and part of the niobium and tantalum. A cupferron- chloroform separation will put vanadium in the chloroform layer with iron, molybdenum, niobium, tantalum, titanium, tungsten, zirconium and part of the copper; in the presence of a fluoride, tungsten is not precipitated by cupferron. Separation by a mercury cathode in 0.3 N sulphuric acid will leave vanadium in solution, with aluminium, niobium, tantalum, titanium, tungsten and zirconium.Vanadium can be removed from niobium and tantalum by the retention of the latter two on Dowex 1 or a similar resin from hydrochloric acid - hydrofluoric acid - water Vanadium can be determined by the persulphate - permanganate volumetric procedure, which depends on the fact that, in cool acidic solution containing no silver nitrate, vanadium can be reduced by adding an excess of iron(I1) sulphate, the excess destroyed with ammonium persulphate and the reduced vanadium finally titrated with standard potassium permanganate solution 919 Small amounts of vanadium (0.1-0.5y0) can be readily determined by the hydrogen peroxide spec tropho t omet ric procedure.*l1 9 1 9 The elements that interfere, either because they also give a yellow colour with hydrogen peroxide or because their ions are coloured, are chromium, cobalt, copper, iron, molybdenum, nickel, titanium and tungsten. Using the 318.4-nm line and a dinitrogen oxide - acetylene flame, vanadium can be deter- mined by atomic-absorption spectrophotometry. An addition of 1000 pg ml-l of potassium should be made to both the test and standard solutions. Large excesses of most elements found in carbides do not affect the absorption of vanadium.6 Optical spectrography and X-ray fluorescence can be used to determine various concentra- tions of vanadium in carbide^.^^^^^^ (5 + 4 + 11). Zirconium Copper and molybdenum are removed from zirconium by precipitation with hydrogen sulphide in acidic solution.Sodium peroxide fusion followed by boiling with water will leave zirconium as a precipitate, accompanied by cobalt, copper, iron, manganese, nickel, titanium and part of the niobium and tantalum. Cupferron - chloroform extraction will yield zirconium in the chloroform layer with iron, molybdenum, niobium, tantalum, titanium, tungsten and vanadium and part of the copper. A mercury cathode separation will leave zirconium in solution, accompanied by aluminium, niobium, tantalum, titanium, tungsten and vanadium. An excellent separation of zirconium from the other constituents of carbides is effected by precipitation with phosphoric acid in strongly acidic solution containing hydrogen peroxide; niobium and tantalum, which are carried down by zirconium, are the only interfering elements.Zirconium is separated from niobium and tantalum by the retention of the latter two in a column of Dowex 1 or a similar resin from hydrochloric acid - hydrofluoric acid - water (5 + 4 + 11). Zirconium is usually determined gravimetrically by either the mandelic acid method or the phosphate procedure.4J1JoJ1 For smaller amounts, a spectrophotometric determination, using Alizarin Red S, also known as sodium alizarin sulphonate, has wide application~.~~~0~~1 Although zirconium can be determined by atomic-absorption spectrophotometry, using the 360.1-nm line and a very fuel-rich dinitrogen oxide - acetylene flame, the sensitivity is low and the element is subject to complex interferences.Optical spectrography for low zirconium contents, and X-ray fluorescence for a wider range, are other techniques employed for the determination of this element in carbide^.^^^^^^^730 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. i8. 19. 20. 21. YOZJNG References Kaarik, K., and Thorgren, K.-E., Paper presented a t 27th Pittsburgh Conference on Analytical Young, R. S., “Separation Procedures in Inorganic Analysis,” Charles Griffin, High Wycombe, 1980. Tikhonov, V. N., “Analytical Chemistry of Aluminium,” (“Analiticheskaya Khimiya Alyuminiya”) , Young, R. S., “Chemical Analysis in Extractive Metallurgy,” Charles Griffin, London, 1971. Sandell, E. B., and Onishi, H., “Photometric Determination of Traces of Metals; General Aspects,” Thompson, K. C., and Reynolds, R. J., “Atomic Absorption, Fluorescence and Flame Emission Slavin, M., “Atomic Absorption Spectroscopy,” Second Edition, John Wiley, New York, 1978. Van Loon, J. C., “Analytical Atomic Absorption Spectroscopy,” Academic Press, New York, 1980. Barnes, R. M., Editor, “Emission Spectroscopy,” John Wiley, Chichester, 1976. Nemodruk, A. A., and Karalova, 2. K., “Analytical Chemistry of Boron,” Halsted Press, New York, Furman, N. H., Editor, “Scott’s Standard Methods of Chemical Analysis,” Sixth Edition, Volume 1, Bertin, E. P., “Introduction to X-ray Spectrometric Analysis,” Plenum, New York, 1978. Young, R. S., “The Chemical Analysis of Manganese,” Manganese Centre, Neuilly sur Seine, 1978. Busev, A. I., “Analytical Chemistry of Molybdenum,” Halsted Press, New York, 1971. Gibalo, I. M., “Analytical Chemistry of Niobium and Tantalum,” Halsted Press, New York, 1971. Myshlyaeva, L. V., and Krasnoshchekov, V. V., “Analytical Chemistry of Silicon, ’’ Halsted Press, Topping, J. J., Tulanta, 1978, 25, 61. Busev, A. I., Ivanov, V. M., and Sokolova, T. A., “Analytical Chemistry of Tungsten,’’ (“Analiti- Svehla, G., and Tolz, G., Talanta, 1976, 23, 755. Mukherji, A. K., “Analytical Chemistry of Zirconium and Hafnium,” Pergamon Press, Oxford, 1970. Elinson, S. V., and Petrov, K. I., “Analytical Chemistry of Zirconium and Hafnium,” Halsted Press, Received December 8 t h 1981 Accepted February 8th, 1982 Chemistry and Applied Spectroscopy, Cleveland, OH, March 2nd, 1976. Nauka, Moscow, 1971. Fourth Edition, John Wiley, New York, 1978. Spectroscopy,” Second Edition, Charles Griffin, London, 1978. 1971. Van Nostrand, Princeton, 1962. New York, 1974. cheskaya Khimiya Vol’frama”), Nauka, Moscow, 1976. New York, 1971.

ISSN:0003-2654    

DOI:10.1039/AN9820700721

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

Analyst, July, 1982, Vol. 107, pp. 731-736 731 Extraction and Analysis of Interstitial Water from Sandstone K. C. Wheatstone Severn-Trent Water Authority, Directorate of Scientz$c Services, 2297 Coventry Road, Sheldon, Birmingham, B26 3PU and D. Gelsthorpe Severn-Trent Water Authority, Directorate of Scientific Services, Nottinghawz Regional Laboratory, Meadow Lane, Nottingham, NG2 3HN This paper describes the laboratory work undertaken to develop a technique for extracting the interstitial water from borehole core samples of sandstone usir g a high-speed centrifuge. The effects of operating conditions and aspects of sample handling and preparation on recovery and potential contaminants are considered. Keywords ; High-speed centrifuging ; interstitial water ; sandstone ; ground- water quality The analysis of samples of water taken by pumping or depth-sampling a borehole provides information on the current quality of groundwaters, but does not provide any information to enable a prediction of the possible future levels of contaminants in the groundwater.This is an important question on which a Water Authority needs reliable data in order to manage its resources effectively. In the past, laboratory techniques used to provide the necessary data have involved leaching out the soluble components from a sample of crushed rock with distilled water followed by chemical analysis of the leachate. This method is very inaccurate at the levels of concentration normally found in groundwaters and is inappropriate for investigations in which small changes in concentration could be important in predicting future trends.Recently, a much more reliable technique of extracting water from samples of chalk has been deve1oped.l This technique uses a high-speed centrifuge to extract interstitial water from the chalk to enable chemical analysis to be carried out to characterise the change in chemical composition with depth of an aquifer. It was decided to develop this technique further for groundwater investigations based on Triassic Sandstones. The type of rock is most important in connection with the amount of interstitial water obtained; whereas 30- 40 ml of interstitial water can be extracted from chalk samples by high-speed centrifuging, the same amount of sandstone would only produce about 10 ml. Tests were also carried out to confirm that the wide range of determinands required on the extracted interstitial water were unaffected by the sample handling techniques employed. Equipment The centrifuge used for this work was an MSE High Speed 25 (MSE Ltd., Crawley, Sussex), which was fitted with an aluminium angle rotor (Catalogue No.43115-110), which would accept six sample cups. Initially, sample cups were constructed from Delrin (a high-density plastic) as previously rec0mmended.l However, significant levels of contamination of ammonia, total oxidised nitrogen, potassium, sodium and total organic carbon were found that could not be eliminated even by soaking in hydrochloric acid for several days. Conse- quently the cups were constructed from titanium to a modified design (Fig.1) that over- came contamination problems and made them lighter than the original plastic cups. Thus more rock sample could be processed in the titanium cups. The titanium sample cups were manufactured by Patelmo Engineering Limited, Bassett Road, Park Lane Industrial Estate, H alesowen, West Midlands. For a centrifuge operating at high speed, it was necessary to balance the rotor carefully. This was achieved by loading all six cups with sample and adjusting their mass to 500 & 0.2 g.732 WHEATSTONE AND GELSTHORPE : EXTRACTION AND H 1 cm Removable top o', CUP Roty lid Crushed sandstone . rock sample Rator * Glass-fibre fi Iter-pa per (GF/C) Integral perforated. plate Removable base of cup Extracted interstitial water ANALYSIS Analyst, Vol.I07 Fig. 1. Titanium centrifuge cup fitted in the angle rotor of the MSE High Speed 25 centrifuge. Investigation of Optimum Centrifuge Operating Conditions Centrifuging Speed Six samples of sandstone were selected, three from each of two cored boreholes, reflecting varying moisture contents and physical characteristics ; the essential details of the samples are given in Table I. The samples were centrifuged at a speed of 3000 rev min-l for 2 h after which the amount of water recovered was measured by weighing. This was repeated at speeds of 6000, 100oO and 12000revmin-l. The amount of water recovered at each centrifuge speed was expressed as a cumulative percentage of the total available water in the sample (Fig. 2). The total water content was determined by heating a separate sample of core at 105 "C for 2 h.Examination of Fig. 2 shows that the water removed from Sample B1 at a speed of 3000 rev min-l was only about 2% of that present in the sample; this increased to 45% at 12000 rev min-l. The other five samples produced much smaller differences between the lowest and highest speeds used, the general trend being a gradual increase in water removed with increasing speed. Only with sample B3 was all the water removed, and this was at the maximum speed. It was therefore decided that a speed of 12000revmin-l would be employed for all future work. TABLE I CHARACTERISTICS OF SANDSTONE SAMPLES USED FOR DETERMINATION OF CENTRIFUGE OPERATING CONDITIONS Sample Depth/m Moisture, % Comments Borehole A- A1 .. . . 10.3 14.5 Heterogeneous (a mixture of sandstone and pebbles) A2 ... . 54 6.0 Intermediate A3 .. . . 95 12.1 Homogeneous (red sandstone) B1 .. . . 10.25 7.7 Heterogeneous (mainly a mixture of pebbles and clay) B2 . . . . 54 9.6 Intermediate B3 . . . . 95.5 14.0 Homogeneous (yellow sandstone) Borehole B- Centrifuging Time Following the choice of centrifuge speed, further aliquots of the same six samples were centrifuged sequentially for periods of 0.5, 1, 2 and 4 h at 12000 rev min-l and the amount of water removed at each time interval was measured. Again these results were expressed as a percentage of available water and plotted against centrifuging time as shown in Fig. 3.July, 1982 OF INTERSTITIAL WATER FROM SANDSTONE 733 1 8 L I-' a, m 3 - m .- I-' I-' .- 2 c C - 0 3000 6000 9000 12000 Centrifuging speedhev min-' L - d 1 2 3 4 0 Centrifuging ti me/ h Fig.2. Amount of interstitial water removed Fig. 3. Amount of interstitial water removed from sandstone samples after centrifuging a t 12 000 rev min-l for various times. from sandstone samples after centrifuging for 2 h a t various speeds. This demonstrates that for all samples there was only a marginal increase in the amount of water removed between 1 and 4 h centrifuging compared with 0.5 h, and consequently a centrifuge time of 0.5 h was selected. Reliability of the Centrifuge Method Repeatability of Sample Volume To assess the reliability of the chosen operating conditions, samples were processed with a view to determining the repeatability of the volume of water extracted.A bulk sample of a sandstone core was prepared and sub-divided into centrifuge cups. Interstitial water was then extracted using the operating conditions described above and the volume obtained was measured by weighing. This was repeated using sandstone core samples obtained from several depths down a borehole. The results obtained are given in Table 11, from which it is evident that the repeatability of extraction of interstitial water from a particular sample is good, in most instances all results being within &1Oyo of the average. TABLE I1 REPLICATION OF INTERSTITIAL WATER RECOVERY Results are expressed in terms of percentage recovered. Depth of core sample/m r 7 33.3 33.3 35.7 38.1 31.0 33.3 Average . . . . 34.1 Range . . . . 31.0-38.1 8 43.6 45.2 47.7 45.4 43.8 46.5 45.4 43.6-47.7 16 73.2 68.7 73.6 70.9 72.5 74.5 72.2 68.7-74.5 23 72.5 70.4 65.2 66.7 77.6 69.8 70.4 65.2-77.6 Repeatability of Determinand Concentration determinands.and B3 (see Table I). Further tests were then carried out to investigate the repeatability of analysis for certain Three samples of sandstone core were selected for this investigation, B1, B2734 WHEATSTONE AND GELSTHORPE : EXTRACTION AND ANALYSIS Analyst, VoZ. I07 Bulk samples of core from each depth were sub-divided and placed into a series of centri- fuge cups and the interstitial water was extracted and analysed for electrical conductivity, chloride and total oxidised nitrogen. These three determinands give indications of the typical results that could be expected from determinations on these types of sample.The results obtained are given in Table 111, from which it can be seen that the homo- geneity of the sample has a definite effect on the replication of all three determinands; in all instances the sample B1 produced the most scattered set of results and sample B3 the least scattered. The concentrations of chloride and total oxidised nitrogen found for sample B2 were towards the lower end of the range employed (0-300mg1-1 for chloride and 0- 20 mg 1-1 for total oxidised nitrogen), and were therefore subject to proportionately larger error. TABLE I11 REPLICATION OF ANALYSES Determinand Sample Mean Electrical conductivity/pS cm-l . . .. B1 642 B2 360 B3 467 Chloride/mgl-l . . . . . . .. B1 85 B2 47 B3 58 Total oxidised nitrogenlmg 1-’ .. .. B1 5.2 B2 2.1 B3 3.8 * Based on 7 replicate analyses. 95% Confidence limits* 40 23 14 10 2.4 1.4 1.4 0.7 0.4 Fractionation Effect Previous work1s2 had indicated that there could be changes in the chemical composition of interstitial water depending on the amount removed. This “fractionation” effect had been observed for salinity with sediment samples2 and for the cations sodium, potassium and calcium in interstitial water from cha1k.l It was thought prudent to examine any possible fractionation effects for anions with sandstone core samples. Total oxidised nitrogen and chloride, together with electrical conductivity, were selected for study. The fractionation effect was examined by carrying out analyses on the interstitial water removed from a core sample (B3) after centrifuging at 3000,6000, 10000 and 12000 rev min-l for a period of 0.5 h.The results are detailed in Table IV from which it can be seen that there is close agreement between the concentrations found for each determinand at all four centrifuge speeds. No evidence of fractionation was observed. TABLE IV INVESTIGATION OF FRACTIONATION EFFECTS The sample was B3, borehole B. Centrifuge speed/ interstitial water Total oxidised conductivity/ . rev min-‘ removed, % nitrogenlmg 1-’ Chloride/mg 1-’ pS cm-l 3 000 73 4.6 58 465 6 000 69 4.7 58 495 10000 80 4.7 56 470 12 000 88 4.3 58 490 Amount of Electrical Sample Preparation and Analysis Sample Collection For ease of handling, especially with unconsolidated samples taken near the surface, and in order to preserve determinands such as nitrate and phosphate, it was proposed to freezeJuly, 1982 OF INTERSTITIAL WATER FROM SANDSTONE 735 the samples on site, transfer them into the laboratory in a frozen state and store them in a freezer until required for analysis.A further advantage of freezing is that interstitial water movement (particularly draining) during storage would be prevented. Investigations were carried out to ascertain whether this proposal affected the concentrations of determinands. Three determinands, electrical conductivity, chloride and total oxidised nitrogen, were selected, which, together with percentage moisture content, were thought to be representa- tive of the full analysis suite. Three samples (shallow, medium and deep) were selected from a borehole and were taken to the laboratory unfrozen.At the laboratory each sample was divided into two portions, one of which was analysed immediately, the other being frozen (-18 "C) for 1 week before analysis. Each analysis was carried out in duplicate and the results obtained are given in Table V. It was concluded that percentage moisture content and all three determinands were unaffected by freezing and this procedure was adopted for routine use. TABLE V COMPARISON OF INTERSTITIAL WATER FROM FROZEN AND UNFROZEN SAMPLES Electrical conduc- Moisture, yo tivity/pS cm-l & -7 Depth/m Unfrozen Frozen Unfrozen Frozen 6.0* 7.6 7.7 445 360 435 380 36.0 8.2 8.1 440 435 465 460 160.0 11.2 11.6 445 430 450 430 Chloridelmg 1-I & Unfrozen Frozen 48 41 45 46 50 52 50 52 62 66 63 66 Total oxidised nitrogenlmg 1-1 1 Unfrozen Frozen 4.9 3.2 4.8 5.5 10.1 10.7 12.4 11.1 6.9 6.0 6.6 5.8 * This sample was heterogeneous.Laboratory Procedure A frozen sandstone core sample, which was typically 25 cm long and 10 cm in diameter, was removed from the laboratory freezer and, by means of sharp blows, the top and bottom 4-5 cm of core were removed and discarded. The remaining sample was left at room temperature for 20-30 min when the outer 1.0-1.5 cm of core had thawed out sufficiently for it to be removed and discarded. (This was simple to achieve because as the sample thawed out the outer layer tended to crumble away.) The remaining centre section of the core sample, which was free from possible contamina- tion from drilling operations, was crushed and placed in a centrifuge cup.The five remaining centrifuge cups were loaded with other core samples treated as above, although occasionally it was necessary to load the same sample into two or more centrifuge cups in order to obtain sufficient interstitial water for analysis. When centrifuging was completed, the extracted water was transferred into a 35-ml stoppered glass tube. Each tube was labelled and stored in a refrigerator at 4 "C until the sample was required for analysis. By careful conservation of the extracted water, it was possible to determine up to 15 determinands from 15-20 ml of sample. The results of the application of this technique to the detailed investigation of high nitrate levels in groundwater have been published elsewhere .3 Conclusions The operating parameters of the high-speed centrifuge for extracting interstitial water from cored borehole samples have been examined and a routine laboratory technique has been developed.The reliability of the technique, in terms of both volume and quality of interstitial water obtained, has been thoroughly assessed and quantified. Aspects of core sample handling at the laboratory have been investigated and procedures instituted whereby the interstitial water is preserved unchanged and the results obtained are free from contamination by the drilling technique employed. A comprehensive analytical736 WHEATSTONE AND GELSTHORPE scheme has been developed for routine use whereby most of the major constituents of inter- stitial waters are accurately measured with only a small volume of sample. This assists the prediction of the effects of pollution of groundwater and enables a balanced assessment to be made of the causes and mechanisms involved. The authors thank Mr. C. W. Tait, who carried out the laboratory work, and Mr. W. F. Lester, Director of Scientific Services of the Severn-Trent Water Authority, for permission to publish this paper. References 1. 2. 3. Edmunds, W. M., and Bath, A. M., Environ. Sci. Technol., 1976, 10, 467. Rieke, H. H., and Chilingarian, G. V., “Compaction of Argillaceous Sediments. Lucas, J. L., and Reeves, G. M., Prog. Water Technol., 1980, 13, 81. Developments in Sedimentology,” Elsevier. Amsterdam, 1974. Received December 9th, 1981 Accepted February 26th, 1982

ISSN:0003-2654    

DOI:10.1039/AN9820700731

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

Analyst, July, 1982, Vol. 107, pp. 737-743 737 Homogeneous Precipitation of Palladium Dimethylglyoximate by Interchange Reactions of C=N Groups A. Rios and M. Valcarcel Department of Analytical Chemistry, Faculty of Sciences, University of Cdrdoba, Cdrdoba, Spain The formation of palladium dimethylglyoximate by homogeneous precipita- tion between pH 1 and 2 is described using several iminic derivatives of diacetylmonoxime (hydrazone and semicarbazone) , by reaction with an excess of hydroxylamine. Kinetic studies of the interchange of >C=N- groups were also carried out. Quantitative precipitation of palladium(I1) is useful for the gravimetric determination of micro-amounts of the metal with excellent results. It is shown that the physico-chemical properties of the precipitate are more advantageous than those shown by the precipitate obtained by the conventional method using dimethylglyoxime.Keywords : Palladium dimethy lglyoximate ; homogeneous precipitation ; iminic compounds ; hydroxylamine ; > C = N- interchange reactions This paper is a continuation of a study of the possibilities offered by interchange reactions of the >C=N- groups in analytical chemistry. These reactions were first described in 1973 by Valcfircel and Pinol as the cause of instability of photometric ligands in solution, in the system copper( I) - hydrazine (hydroxylamine) - 6-methylpicolinaldehyde azine. The reactions of interchange of >C=N- groups2: )C=N-X + Y-NH, + )C=N-Y + X-NH, R' R' consist of a net change of the X for the Y radical by means of the action of an excess of amine on an jmine compound.These reactions occur at a reasonable speed in aqueous medium at an appropriate pH. In this work we have made use of the formation of a palladium dimethylglyoximate, Pd(DMG) 2, precipitate by homogeneous precipitation using an in situ synthesis of dimethyl- glyoxime by the action of an excess of hydroxylamine on iminic derivatives of diacetyl- monoxime, viz., the hydrazone (DMH) , semicarbazone (DMS) and thiosemicarbazone (DMT) : CH,C=N-OH CH 3-C =N-OH R = NH, DMH DMG I + NH,OH(excess)+ 1 + H2N-R R = NHCONH, DMS CHS-C=N-R CH,-C=N-OH R = NHCSNH, DMT The reactions are carried out in the presence of palladium(I1) at acidic pH so that the dimethyl- glyoxime is formed in situ in its own reaction medium. Competition for palladium(I1) shown by the individual nitrogenous compounds referred to above is established.The principal characteristic of homogeneous precipitation techniques is that induction times are relatively long. This feature allows the gradual and the simultaneous formation of the precipitating reactant throughout the s~lution.~-~ Crystalline growth is favoured in this way, producing particles, which, as shown below, are only slightly contaminated and easy to manipulate. This technique, introduced by has been used on several occasions for the determination of palladium by in situ formation of the organic precipating reagent. Thus, in 1961, the homogeneous precipitation of palladium dimethylglyoximate with diacetyl and hydroxylamine was proposed by Kanner et a1.* Other organic reagents, such as f~rfuraldehyde,~ cyclohexane-1 ,2-dione,1° o-hydroxybenzaldehydell and indan-1-one738 RiOS AND VALCARCEL : HOMOGENEOUS PRECIPITATION OF Analyst, VOl.I07 oxime,l2 which precipitate palladium oximes in situ by reaction with hydroxylamine, have also been employed. Experimental Apparatus A Beckman 3500 pH meter with a combined glass - calomel electrode was used for all pH measurements. Absorbances were measured at a fixed wavelength with a Pye Unicam 500 digital spectrophotometer. A Perkin-Elmer 575 spectrophotometer with automatic control and temperature programming was employed for recording absorbance - time curves. Microphotography was performed with a Kyowa Microlux-73 photomicroscope equipped with a Nikon M-35s camera.Reagents water. All reagents were of analytical-reagent grade and solutions were prepared with distilled Diacetylmonoxime hydrazone, 0.02 and 1 yo m/V solution in ethanol. Diacetylmonoxime semicarbazone, 0.02% mlV solution in ethanol and 0.5% m/V solution in Diacetylmonoxime thiosemicarbaxone, 0.02% ml V solution in ethanol. Hydroxylamine chlorhydrate solution, 1 M. Palladium(II) standard solution, 1.248 g 1-l. N N-dimethylformamide. The solution was standardised gravimetrically with dimeth ylglyoxime .13 Procedure Synthesis of diacetylmonoxime semicarbaxone (DMS) l4 Two grams of semicarbazide chlorhydrate neutralised with 1 M sodium hydroxide solution are added to 2.7 g of diacetylmonoxime in ethanol. A 1- or 2-ml volume of glacial acetic acid is added and the mixture is shaken until a white precipitate appears, which is recrystal- lised from ethanol - water (1 + 1). C,H,,N,O, requires C 37.97, H 6.32 and N 35.44%; found, C 37.6, H 6.4 and N 35.2%.Elemental analysis. Synthesis of diacetylmonoxime thiosemicarbazone (DMT) l5 A 3.44-g amount of diacetylmonoxime is dissolved in 68.6ml of ethanol and mixed with a solution prepared by dissolving 3.104 g of thiosemicarbazide in 68.8 ml of hot water. The solution is filtered, then heated under reflux for 2 h, during which period it becomes yellow, and is subsequently evaporated under vacuum. The white residue is recrystallised from ethanol. C5Hl,N,0S requires C 34.48, H 5.88, N 31.20 and S 18.28y0; found, C 34.7, H 5.9, N 30.1 and S 18.7%. Elemental analysis. Synthesis of diacetylmonoxime hydrazone (DMH) To a solution of 2 g of diacetylmonoxirne in 40 ml of ethanol a solution containing 1.1 ml of lOOyo hydrazine hydrate in 10ml of ethanol is added drop by drop under reflux.The mixture is maintained under reflux for a further 1 h and approximately half of the solvent is then evaporated under vacuum. This is cooled and white crystals of hydrazone appear, which are recrystallised from ethanol - water (1 + 1). Elemental analysis. C,H,N30 requires C 41.73, H 7.87 and N 36.50% ; found, C 41.6, H 7.8 and N 36.9%. Gravimetric determination of palladium with DMS and DMH To 500-ml flasks containing a solution of 10-35 mg of palladium(II), 1.5 ml of 1 M hydro- chloric acid are added in order to control the pH, then 1.2 ml of 1 M hydroxylammonium chloride solution and 1.5 ml of a 0.5% m/V solution of DMS in NN-dimethylformamide per milligram of palladium present are added.When DMH is employed the amounts of reagentsJuly, 1982 PALLADIUM DIMETHYLGLYOXIMATE BY INTERCHANGE REACTIONS 739 added are 3.0 ml of 1 M hydroxylammonium chloride solution and 6.0 ml of a 1% m/V solution of DMH in ethanol per milligram of palladium present. Distilled water is added to give a final volume of 350 ml (concentration range 35-100 mg 1-1 of palladium). These solutions are maintained at 50 "C for 45 min if DMS is used and for 2 h with DMH. The solutions are cooled and filtered through a G-3 sintered-glass filter and washed until chloride ions have been eliminated. The samples are dried at 110 "C and weighed as Pd(C,H,O,N,),.The gravimetric factor was 0.316 1. Results and Discussion Study of Interchange Reactions in the Absence of Palladium( 11) Ultraviolet spectroscopy was used for studying the reactions in the absence of palladium(I1). The absorption maxima were determined for the compounds involved in the reactions at different pH values and it was shown that an excess of hydroxylamine at between pH 1 and 6 over DMS (Amax. = 260 nm), DMH (Amax. = 255 nm) and DMT (Amax. = 293 nm) leads to the formation of dimethylglyoxime (Amax. = 228 nm) in all instances (see Fig. 1). The kinetic study of these reactions, made by recording absorbance - time curves at the wavelength of maximum absorbance of the iminic compound or dimethylglyoxime, gave the following results. (a) The partial orders of the reaction were calculated and the rate equation in excess of hydroxylamine (the concentration value includes the rate constant) is suggested to be V = k[imine] [H+] where imine represents DMS, DMH or DMT.Rate constants, k (s-l), were calculated by plotting log(A, - A,) against time and the influence of different experimental variables was observed : In each of the three instances the reaction is more favoured at more acidic pH values (Fig. 2), whereas at neutral or alkaline pH the interchange reaction does not take place. InJuence of hydroxyZamine concentration. An increase in hydroxylamine concentration results in greater k values, especially when the molar ratio of hydroxylamine to imine is greater than 103 (Fig. 3). InJuence of temperature. An increase in temperature results in an increase in k , in accordance with the Arrhenius equation (Fig.4). By plotting log k against T-l, activation energies for each reaction of 17.260, 11.882 and 22.457 kcal mol-1 for DMS, DMH and DMT, (b) InJuence of PH. 225 250 275 300 3 Wavelengthlnm Fig. 1. Formation of dimethylglyoxime with diacetylmonoxime semicarbazone and hvdroxvlamine in acidic medium c ly) 1 0 - ~ 9 10-4 b 0 1 2 3 4 5 6 PH (spectrbphotdmetric measurement). Con- Fig 2. Influence of pH on rate con- centration of DMS: 7.6 x M. stants of interchange reactions.740 c I 2 1 0 - ~ RfOS AND VALCARCEL : HOMOGENEOUS PRECIPITATION OF Analyst, DMH DMS DMT - 1 0 1 2 3 pNHpOH Fig. 3. Effect of an excess of hydroxylamine on the interchange reactions (the iminic concentration is lo-*M).2.00 r I rn \ - = 3.00 0 -I 4.oa DMH + NH20H DMT+NH,OH \ 1 3.00 3.10 3.20 3.30 3.40 7-l x 1Op3/K V O l . I07 Fig. 4. Influence of temperature on the interchange reactions. respectively, were obtained: producing the reaction rate sequence DMH > DMS > DMT, which shows the facility to produce dimethylglyoxime in solution. For this type of reaction, an analogous mechanism to the hydrolysis and condensation of Schiff bases can be established, which explains the above-mentioned sequence as a result of the different electronic effects produced by hydrazone, semicarbazone and thiosemicarbazone groups.ls Maximum stability occurs when the rr-electron system is most extensively de- localised (as with the semicarbazone and thiosemicarbazone) .The difference in reactivity observed between DMS and DMT may be due to the difference in the electronegativity of oxygen and sulphur. Such a mechanism assures destruction of the rr-electron system. Therefore, it is evident that the interchange reactions studied assume a net loss of stabilisa- tion by resonance for DMS and DMT and that these reactions will be slower (less favoured) than for DMH, which has an electron system similar to that of dimethylglyoxime. Formation and Study of Palladium Dimethylglyoximate Reaction of palladium(II) with amines and imiiaes at pH 1.5 Reactions between palladium(I1) and the components of the reactions at the optimum pH for the precipitation of palladium dimethylglyoximate were studied spectrophotometrically. The results are given in Table I.As for the amines, neither hydroxylamine nor semi- TABLE I BEHAVIOUR OF PALLADIUM(II) WITH THE COMPONENTS OF >C=N- INTERCHANGE REACTIONS STUDIED (PH = 1.5) Reaction A mines- Hydroxylamine . . .. . . N o reaction Hydrazine . . .. . . Pd(I1) is reduced to PdO Semicarbazide . . .. . . KO reaction Thiosemicarbazide . . . . Two complexes are formed: Metal to ligand ratio Stability constant 2: 1 2.3 x 10" 1:2 4.9 x 1011 Imine compounds : all form complexes- Xmax./nm Metal to ligand ratio Stability constant DMS .. .. 325 1: 1 2.0 x 105 DMH .. .. 335 1: 1 3.5 x 105 DMT .. .. 325 1: 1 7.2 x 105 1:2 4.8 x lo1'July, 1982 PALLADIUM DIMETHYLGLYOXIMATE BY INTERCHANGE REACTIONS 741 carbazide reacts with palladium(I1) at this pH. Hydrazine reduces palladium(I1) to palladium, which is precipitated.This reaction is not observed in the presence of dimethyl- glyoxime. Thiosemicarbazide forms two stable soluble complexes with palladium( 11) and no palladium dimethylglyoximate is formed. All of the iminic compounds form complexes with palladium(I1) in solution at this pH. This does not prevent the corresponding hydrazonates and semicarbazonates of palladium( 11) from progressing towards the dimethylglyoximate in an excess of hydroxylamine. For this reason, the formation of these complexes of palladium(I1) with DMS and DMH is more favourable in the induction periods of the >C=N- interchange reactions than in the quantitative precipitation of palladium. On the other hand, it has been shown that thio- semicarbazone inhibits the formation of the precipitate.Therefore, there are two reasons why DMT cannot be the initial imine in the homogeneous precipitation, as both the imine (DMT) and the amine (thiosemicarbazide) interfere in the formation of the chelate Pd(DMG),. Induction period This is defined as the period of time between the mixing of the reagents and the appearance of the precipitate. Long induction periods result in better physico-chemical properties of the precipitate and a reduction of the contamination. The influence of several variables on the induction periods was studied (see Fig. 5 ) . An increase in temperature and the excess of hydroxylamine reduce the induction period ( I ) whereas an increased concentration of organic solvent [ethanol, acetone, dioxan and dimethyl- formamide were tried) retards the precipitation.20 40 60 80 TPC 30 20 10 0 25 50 75 Solvent, o/o V/V Fig. 5. Induction ,periods in the homogeneous precipitation of Pd(DMG), with DMS. Influence of (a) temperature, (b) excess of hydroxylamine and (c) several solvents: (A), dioxan, (B) dimethylformamide, (C) ethanol and (D) acetone (ordinate in hours for the last solvent). A graph of log1 against T-1 is a straight line similar to the Arrhenius equation. We can express this equation as where A and B are characteristic constants of each system. To extend the Arrhenius equation, B can be considered to be analogous to the activation energy for the homogeneous precipitation of Pd(DMG),. This would be an apparent constant that would include all of the processes that occur: the formation of other complexes of palladium(II), the formation of dimethylglyoxime by >C=N- interchange reactions and the formation of the precipitate itself.Calculated values for B are 4 x 103 for DMS and 5 x lo3 for DMH, which are in agreement with the experimental results, as the precipitate appears earlier with DMS. In order to compare the effects of the different solvents in the induction period in an absolute way, eliminating the effect that produces the solution of the iminic compounds itself, we have introduced the concept of a “limited period of induction” (I,) for each solvent.742 RfOS AND VALCARCEL : HOMOGENEOUS PRECIPITATION OF Anahst, vd. I07 The hypothetical induction period is defined as corresponding to Om1 of the solvent used.This is obtained by extrapolating the curve in Fig. 5(c) to the ordinate and has a characteristic magnitude for each solvent and reaction. In this hypothetical state, we assume that the effects of each solvent in the induction periods are additive, making it possible to eliminate, by simple subtraction from the I , values, the effects of dimethylformamide and ethanol in which the DMS and DMH, respectively, are dissolved, in this way obtaining the corresponding values for I , for the remainder of the solvents (Table 11). Knowing these values of I,, it is possible to establish relationships among the solvents, referring to the ease of the appearance of the precipitate: Reaction with DMS: ethanol < dioxan < DMF <acetone. Reaction with DMH: ethanol < acetone < dioxan < DMF.TABLE I1 1 0 VALUES FOR EACH SOLVENT I,/min * Solvent DMS DMH Dimethylformamide . . .. 6 11.5 Ethanol .. . . . . . . 1 0.5 Acetone . . . . . . . . 9 9.5 Dioxan . . . . . . .. 3 11.0 Photomicrographs Photomicrographs taken to compare the form and size of the particles of Pd(DMG), obtained homogeneously with DMS and DMH with hydroxylamine and those obtained conventionally by direct addition of dimethylglyoxime showed that the homogeneous techniques are more advantageous than those involving direct addition of the precipitating agent. Thermal behaviour , In agreement with literature data,17 the mass of the precipitate is constant at 240 "C, i e . , it maintains its stoicheiometric composition. Heating above this temperature can cause explosions, but by working with great care, quantitative conversion into palladium metal is possible.l* Sublimation of the metal has been shown to occur when it is heated at reduced pressure.Gravimetric Determination of Palladium Table I11 presents the results obtained for the determination of different amounts of palladium(I1) (32.1, 19.2 and 12.8mg) by using the technique proposed for DMS and DMH under Experimental. Standard deviation (s), average standard deviation (sm) and average relative error (P = 0.05) were calculated for five. determinations (t = 2.78) in each instance. As expected, less error occurs at higher concentrations of palladium, although all values are within accepted limits. As can be seen, the determinations for DMS were more precise, especially for the lower concentrations of palladium(I1).TABLE I11 STATISTICAL PARAMETERS FOR THE DETERMINATION OF PALLADIUM DMS DMH A r 1f A \ Palladium Confidence Error, Confidence Error, present/rng s Sm interval % s Sm interval % 32.1 0.07 0.03 32.1 f 0.1 0.26 0.07 0.03 32.1 f 0.1 0.26 19.2 0.07 0.03 19.2 f 0.1 0.43 0.11 0.05 19.1 f 0.1 0.72 12.8 0.11 0.05 12.8 f 0.1 1.08 0.16 0.07 12.7 f 0.2 1.53July, 1982 PALLADIUM DIMETHYLGLYOXIMATE BY INTERCHANGE REACTIONS 743 Interferences Among the elements of the platinum group, gold(II1) shows the strongest interference (Table IV), owing to its reduction to the elemental state by the reducing action of excess of hydroxylamine, co-precipitating with Pd(DMG), and producing an excess of errors. On the other hand, EDTA does not interfere at high concentrations, and can be used to mask base metals although the pH is not optimum for the formation of such complexes.TABLE IV HOMOGENEOUS PRECIPITATION OF Pd(DMG), IN THE PRESENCE OF OTHER SPECIES Palladium present in all instances: 32.1 mg. Amount of foreign Amount found/ Difference, Foreign species Salt used specieslmg mg % Os(VII1) . . .. oso, 64.2 32.1 0.0 Pt(1V) .. . . K2PtC1, 15.0 32.2 +0.3 Rh(II1) .. . . RhC1,.3H20 10.0 32.3 f0.6 Au(II1) . . . . HAuC1,.2H20 64.2 50.7 + 56.0 EDTA .. . . Na2C,,H,4N20,.2H20 321.0 32.3 + 0.6 Conclusion Two gravimetric techniques involving homogeneous precipitation have been proposed for the determination of micro-amounts of palladium, emphasising once again the important analytical utility of >C=N- group interchange reactions.Both techniques show clear advantages over the conventional procedure using dimethylglyoxime. This is reflected in better physico-chemical properties of the precipitate, which can be manipulated and filtered more easily. A t the same time, as a consequence of the relatively long induction periods that allow for the slow formation of large particles, contamination by adsorption and occlusion of the precipitate is considerably reduced. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. References ValcArcel, M., and Pino, F., Talanta, 1973, 20, 224. Patai, S., “The Chemistry of the Carbon-Nitrogen Double Bond,” Interscience, London, 1970, p. 81. Gordon, L., Salustsky, M., and Willard, H., “Precipitation from Homogeneous Solution,” John Wiley, Firsching, F., Talanta, 1963, 10, 1169. Cartwright, P. F. S., Newman, E. J., and Wilson, D. W., Analyst, 1967, 92, 663. Willard, H. H., Anal. Chem., 1950, 22, 1372. Willard, H. H., Anal. Chem., 1953, 24, 459. Kanner, L. J., Salesin, E. D., and Gordon, L., Talanta, 1961, 7, 288. Pino, F., Burriel, F., and Conejero, L. M., An. Quim., 1959, 55, 331. Velazquez, J. A., and Hileman, 0. E., Talanta, 1968, 15, 269. De, A. K., and Sahu, C., Sep. Sci., 1967, 2, 11. Bark, L. S., and Brando, D., Talanta, 1963, 10, 1189. Erdey, H. E., “Gravimetric Analysis, Part 11,” Pergamon Press, Oxford, 1965, p. 386. Martinez, P., Bendito, D. P., and Pino, F., An. Quim., 1973, 69, 747. Bendito, D. P., and Pino, F., Quim. Anal., 1967, 21, 31. Rios, A., Master Thesis, University of Cbrdoba, 1980, pp. 134-137. Flaschka, H. A., and Barnard, A. J., “Chelates in Analytical Chemistry,” Volume 2, Marcel Dekker, Beamish, F. E., and Van Loon, J. C., “Recent Advances in the Analytical Chemistry of the Noble Received October 16th, 1981 Accepted December 7th, 1981 New York, 1959. New York, 1969, p. 200. Metals,” Pergamon Press, Oxford, 1972, p. 431.

ISSN:0003-2654    

DOI:10.1039/AN9820700737

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

744 Analyst, July, 1982, Vol. 107, pp. 744-748 Determination of Metal Ions by Liquid Chromatograpy Incorporating Dithiocarbamates in the Eluent Roger M. Smith and Lawrence E. Yankey Deflartment of Chemistry, University of Technology, Loughborough, Leicestershire, LE11 3T U Dilute solutions of transition metal ions, including copper(II), cobalt (11), nickel(II), lead(I1) and iron(III), can be determined using reversed-phase liquid chromatography on Hypersil ODS by direct injection into an eluent containing sodium diethyldithiocarbamate (0.05 yo m/ V ) . The complexes formed are detected using a variable wavelength detector. The effects of reagent concentration, flow-rate and pH have been examined. Keywords : Transition metal determination ; copper determination ; liquid chromatography ; dithiocarbarnate ions ; on-column derivatisation In recent years there has been considerable interest in the use of liquid chromatography to analyse mixtures of metal ions as their neutral complexes with organic ligands.lV2 These include the diethyl- and tetramethylenedithiocarbamates of nickel, cobalt, copper, zinc, lead, manganese, chromium, cadmium, iron, mercury and bismuth, where separations have been reported on silica ge13-11 and nitrile-bondedl2~l3 and alkyl-bonded silica14-20 columns.Samples of 1-100 ng were determined using absorption detectors. In each instance the complex was formed by mixing the analyte solution with dithio- carbamate solution and then extracting the neutral complex into an organic solution, usually chloroform, which was then used as the sample for chromatography.Following our studies of the determination of dithiocarbamates by liquid chromatography with the incorporation of nickel or cobalt salts into the eluent as an ion-pair type reagent,21s22 we report here an extension of the concept of on-column complex formation to the reverse system. By using an eluent containing dithiocarbamate salts, metal ions can be injected directly as aqueous solutions and then separated and determined as their complexes by using a variable wavelength detector. While this work was being prepared for publication the same method was reported for copper ions using an electrochemical detector.23 A related technique in which quinolin-8-01 was incorporated into the mobile phase has been used to determine copper ions and to separate cadmium, copper, manganese, zinc, nickel, mercury and copper ions using absorption and electrochemical detecti0n.2~ An alternative method avoiding the solvent extraction of a metal complex has recently been described, in which the metal - dithiocarbamate complex in aqueous solution was trapped and concentrated on a pre-column and was then eluted on to the analytical column by a change in the eluent.25 Experimental Apparatus Liquid chromatography was carried out using a Pye Unicam XPS pump connected to a Shandon Southern column (10 cm x 5 mm i.d.) packed with Hypersil ODS and fitted with a Rheodyne 7010 valve injector with a 10-pl loop.The eluates were detected using a Cecil 2012 variable wavelength detector. The solvent flow-rate was 1.0 cm3 min-l. Reagents and Standards Metal salts.Dithiocarbamates. Ammonium NN-tetramethylenedithiocarbamate and sodium NN- diethyldithiocarbamate were obtained from Sigma London Chemical Co. Ltd., Poole, and Fisons Scientific Apparatus, Loughborough, respectively. Analytical-reagent grade metal sulphate or nitrate salts were used. MethanoZ. HPLC grade from Fisons Scientific Apparatus, Loughborough.SMITH AND YANKEY 745 Procedure Solutions (0.1-100 p.p.m.) of metal ions in water (10 pl) were injected on to the column and were eluted with methanol - water (80 + 20) containing either 0.1% m/V ammonium NN-tetramethylenedithiocarbamate or 0.05% m/V sodium diethyldithiocarbamate. The detector was set to an appropriate wavelength for each complex, in the range 320440 nm.Results and Discussion In order to simplify the liquid chromatographic determination of metal ions as dithio- carbamate complexes, it was proposed to eliminate the need for separate complexation and extraction steps. The inclusion of the dithiocarbamate ions in the eluent should permit the direct injection of the metal ion sample solution, complexation taking place on the column. The earlier studies examined the separation of the separately prepared complexes of both diethyl- and tetramethylenedithiocarbamates with metal ions. As the latter complexes have been reported to be the more table,^^,^' the use of ammonium tetramethylenedithio- carbamate (0.1% m/V) in the mobile phase was first examined. The reagent absorbed strongly up to 310 nm and a variable wavelength detector was used between 320 and 450 nm at the optimum wavelength for each complex.On injection of 200 p.p.m. solutions of cadmium(II), copper(II), lead(II), zinc(II), nickel(I1) or cobalt(I1) ions peaks were obtained (Table I). TABLE I CAPACITY FACTORS OF COMPLEXES FORMED ON INJECTION OF METAL IONS Metal ion Capacity factor* Cadmium(I1) . . . . .. 1.56 Lead(I1) . . . . .. 3.10 Nickel(I1) . . . . .. 3.18 Iron(II1) . . . . .. 4.00 Copper(I1) . . .. . . 5.96 Cobalt (11) t . . .. . . 3.87 * Eluent 80 + 20 methanol - water containing f Eluted as the cobalt(II1) complex. 0.05% m/ V sodium diethyldithiocarbamate. However, even at their maximum absorption wavelength the response of many of the ions was weak and repeatability was poor. The most sensitive and stable ion was cobalt(II), which forms a cobalt(II1) complex.22 I t gave a linear response from 1 to 10 ng but a non- linear response at higher concentrations.The peak height from 20-ng samples of cobalt(I1) had a repeatability coefficient of variance of 1.03%. Nickel salts appeared to give a partly insoluble complex as they gave a sharp peak with a long tail. They also interfered with the response of cobalt ions. These problems suggested that ammonium tetramethylenedithiocarbamate was not a suitable reagent. The use of sodium diethyldithiocarbamate (o.05y0 m/V) in the mobile phase was therefore studied. Direct injection of a range of metal ions gave peaks in the 320440 nm region and the order of elution (Table I) corresponded to that reported for the pre-formed c0mp1exes.l~ Using a model mixture of copper(II), cobalt(I1) and nickel(I1) ions and detection at 380 nm, it was found that the capacity factors were independent of eluent flow-rate and of the concentration of complexing agent from 0.01 to 0.15% m/V although at the lower reagent concentrations the peaks were broadened and the resolution was reduced. The use of metal sulphates or nitrates had no effect on the results.The retentions of cobalt, copper and lead complexes were unchanged when phosphate buffer, pH 6.69, or borate buffer, pH 9.72, were used as the aqueous component of the mobile phase, but the cadmium complex was eluted more rapidly at the higher pH. For most of the study the aqueous solution was unbuffered. Mercury(1) and -(II) salts gave two peaks and zinc(I1) gave a single peak but in all three instances the retention times were not reproducible.As in our earlier study22 the metal - dithiocarbamate complex appeared to be formed rapidly on injection. Unlike true ion-pair chromatography, negligible concentrations of746 SMITH AND YANKEY: METAL IONS BY LIQUID CHROMATOGRAPHY Analyst, Vol. I07 undissociated ions are present in the mobile phase as demonstrated by the insensitivity to reagent concentrations. The process can probably more accurately be termed on-column derivatisation. In order to examine the potential of the technique for the determination of metal ions in solution, calibration graphs were measured for four test ions over a range of concentrations (Table 11). TABLE I1 CALIBRATION GRAPHS FOR PEAK HEIGHTS AGAINST CONCENTRATION FOR 0.05% m/V SODIUM DIETHYLDITHIOCARBAMATE SOLUTIONS OF METAL IONS INJECTED INTO ELUENT CONTAINING Wavelength of Sample size/ Number Correlation Standard Intercept on Metal ion detectionlnrn ng of points S1opela.u.ng-l coefficient deviation absorbance axis Nickel .. .. .. 350 5-1.25 3 1.02 x lo-' 0.9997 0.045 x 0.36 x Cobalt .. .. .. 350 5-0.31 6 5.92 x lo-' 0.9978 0.68 x 0.58 x Iron(III).. .. .. 350 10-2.0 5 0.316 x 0.9947 0.10 x 10-9 0.328 x lo-* Copper .. .. .. 440 12.5-1.5 5 3.59 x lo-' 0.9989 0.81 X lo-' -0.63 x lo-' Small sample sizes, between 1 and 10 ng, of the ions give a linear response but with larger samples there was a trend towards giving a reduced response. Whether this is due to incomplete reaction or overloading the reagent is uncertain as there was no significant peak broadening.The absorptions of the different complexes at their optimum wavelength were very different, cobalt and copper being the most sensitive. A similar effect has also been noted with the pre-formed complexes, cobalt and copper showing a greater response than the lead or nickel complexes.ll Because it has no distinct absorption band a t a wavelength greater than 310 nm, the cadmium complex was exanlined at 320 nm, which is on the shoulder of the reagent band. As a result the response was noisy and very insensitive, amounts lower than 100 ng being difficult to determine. The calibration graphs for a number of the metals were not rectilinear, suggesting that some decomposition of the complex was occurring on the column.A similar loss was observed in our earlier study when sodium diethyldithiocarbamate was injected on to a column containing cobalt ions; however, in this instance cobalt ions were not a problem. Although only a few of the studies of the pre-formed complexes were quantitative, non- rectilinear calibration graphs have previously been observed for cadmi~m,~J nicke14p11 and lead complexes5 on silica columns, but in each instance the response was linear over the con- centration range examined. A mixture of copper, cobalt, lead and cadmium ions gave well resolved peaks, using 350 nm as a common detection wavelength, although the responses were very different. No interferences appeared to be occurring (Fig. 1). In order to examine the applicability of the technique to trace analysis a more detailed study of copper(I1) ions was carried out at low concentrations using 440 nm for the detec- tion.The repeatability of peak heights was determined for 2- and 0.5-ng samples (0.2 and 0.05 p.p,m. solutions) at a sensitivity of 0.02 a.u.f.s (Table 111). Although the peaks were significantly larger than the short-term base-line noise (5 mm), the base line was unstable and wandered. Both the drift and noise were more noticeable than at shorter wavelengths on the same sensitivity and probably indicated the limit of the deuterium lamp output. A calibration graph was constructed for copper solutions containing 0.05-0.6 p.p.m. (0.5-6 ng), replicate measurements being taken (Table IV). The correlation is only fair, largely reflecting the poor reproducibility of individual measurements.If the mean value of each concentration is used the apparent correlation coefficient improves to 0.9780. The copper solutions were also examined by flame atomic-absorption spectroscopy, which gave a correlation of 0.9937 but had the advantage of an integrated signal. The lowest concentra- tion was close to the limit of detection, although higher sensitivities can be obtained by electrothermal met hods. In a recent study using an eluent containing dithiocarbamates and electrochemical detection, a detection limit of 1 ng and linear range of 1-600 ng were reported.23July, 1982 INCORPORATING DITHIOCARBAMATES I N THE ELUENT 747 1 I I 0 5 10 Time/min Fig. 1. Chromatogram of a sample containing copper (100 p.p.m.), cobalt (5 p.p.m.), lead (10 p.p.m.) and cadmium (100 p.p.m.) ions injected into 75 + 35 methanol - water containing 0.05% m/ V sodium diethyldithiocarbamate, with detection a t 350 nm, 0.05 a.u.f.s.TABLE I11 REPEATABILITY OF PEAK HEIGHTS FOR 10-pl SAMPLES OF COPPER IONS FOR SAMPLE CONCENTRATIONS OF 0.05 AND 0.20 p.p.m. Peak height/a.u. x 0.05 0.20 3.76 13.0 4.40 11.2 3.84 14.4 3.64 13.9 3.20 11.4 4.00 13.3 3.20 12.4 Mean . . .. . . . . 3.72 12.8 Standard deviation . . . . 0.43 1.2 r A > Coefficient of variance, yo 11.6 9.4 * Determination a t 440 nm, 0.02 a.u.f.s. Eluent 80 + 20 methanol - water containing 0.05% m/ V sodium diethyldithiocarbamate. TABLE IV CALIBRATION GRAPH FOR COPPER IONS DETERMINED BY LIQUID CHROMATOGRAPHY Copper concentration, p.p.m.* 0.05 0.1 0.2 0.4 0.6 Slope, a.u.p.p.m.-1 . . .. . . .. Correlation coefficient . . . . .. Standard deviation . . .. . . .. Intercept . . . . .. . . .. Peak absorbance x 103 2.88, 2.08 6.4, 7.2, 9.28 13.8, 14.8 20.0, 18.2 27.4 0.042 21 0.966 1 0.002 1 0.002 92 Mean peak absorbance x lo3 2.48 7.6 14.3 19.1 27.4 0.041 73 0.9780 0.002 0 0.002 91 * 10-pl samples injected into 80 + 20 methanol- water containing 0.05% m/ V sodium diethyldithiocarbamate and detection at 440 nm, 0.02 a.u.f.s.748 SMITH AND YANKEY Conclusions The proposed method using a reagent-containing eluent appears to have considerable potential for the examination of solutions of metal ions in the 1-10 p.p.m. range by direct injection, the different ions being separated as their complexes.There is also the possibility of some degree of selective detection by the choice of the detector wavelength. Further studies to determine the scope of the method and its applicability to other metals are in progress. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. 12. 13. 14. 16. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 2 7. References Schwedt, G., GIT Fucha. Lab., 1979, 23, 640. Schwedt, G., Chromatogvaphia, 1979, 12, 613. Uden, P. C., and Bigley, I. E., Anal. Chim. Actu, 1977, 94, 29. O’Laughlin, J. W., and O’Brien, T. P., Anal. Lett., 1978, A l l , 829. Moriyasu, M., and Hashimoto, Y., Anal. Lett., 1978, A l l , 593. Liska, O., Guiochon, G., and Colin, H., J. Chromatogr., 1979, 171, 145. Liska, 0.. Lehotay, J., Brandsteterova, E., and Guiochon, G., J. Chromatogr., 1979, 171, 153.Lehotay, J., Liska, O., Brandsteterova, E., and Guiochon, G., J. Chromatogr., 1979, 172, 379. Liska, O., Lehotay, J., Brandsteterova, E., Guiochon, G., and Colin, H., J. Chromatogr., 1979, 172, Moriyasu, M., and Hashimoto, Y., Chem. Lett., 1980, 117. Edward-Inatimi, E. B., and Dalziel, J. A. W., Anal. Proc., 1980, 17, 40. Bannister, S. J., Sternson, L. A., and Repta, A. J., J. Chromatogr., 1979, 173, 333. Gaetani, E., Laureri, C. F., and Mangia, A., Ann. Chim. (Rome), 1979, 69, 181. Schwedt, G., Fresenius 2. Anal. Chem., 1977, 288, 50. Schwedt, G., Chromatographia, 1978, 11, 146. Schwedt, G., Fresenius 2. Anal. Chem., 1979, 295, 382. Schwedt, G., Chromatographia, 1979, 12, 289. Borch, R. F., Markovitz, J. H., and Pleasants, M. E., Anal. Lett., 1979, B12, 917; Chem. Abstr., Tande, T., Petterson, J. E., and Torgrimsen, T., Chromatographia, 1980, 13, 607. Yamazaki, M., Ichinoki, S., and Igarashi, R., Bunseki Kagaku, 1981, 30, 40; Chem. Abstr., 1981, 94, Smith, R. M., Morarji, R. L., Salt, W. G., and Stretton, R. J., Analyst, 1980, 105, 184. Smith, R. M., Morarji, R. L., and Salt, W. G., Analyst, 1981, 106, 129. Bond, A. M., and Wallace, G. G., Anal. Chem., 1981, 53, 1209. Berthod, A., Kolosky, M., Rocca, J.-L., and Vittori, O., Analusis, 1979, 7, 395. Haring, N., and Ballschmiter, K., Talanta, 1980, 27, 873. Hulanicki, A., Talanta, 1967, 14, 1371. Subramanian, K. S., and Meranger, J. C., Int. J . Environ. Anal. Chem., 1979, 7, 25. 384. 1979, 91, 150966. 197360. Received December 16th, 1981 Accepted February lst, 1982

ISSN:0003-2654    

DOI:10.1039/AN9820700744

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

Analyst, July, 1982, Vol. 107, PP. 749-754 749 Rapid Screening Technique Utilising Hig h-performance Liquid Chromatography for Assessing Acrylamide Contamination in Effluents Leslie Brown, Michael M. Rhead and Keith C. C. Bancroft John Graymore Chemical Laboratory, Department of Environmental Sciences, Plymouth Polytechnic, Drake Circus, Plymouth, Devon, PL4 8A A A reliable method for the determination of acrylamide, suitable for the rapid screening of sewage and industrial effluents, is presented. Using initial sample clean-upJ which involved the passage of the sample through a mixed- resin bed containing anionic (OH-) , cationic (H+) and hydrophobic resins, it was found possible to reduce the major interfering organic compounds and inorganic ions, present in such samples, to acceptable levels.Final resolution from interferences was achieved by reversed-phase high-performance liquid chromatography and quantification by ultraviolet detection. Precisions of 10% at 5 and 10 pg 1-1 and 4% at 100 pg 1-1 of acrylamide have been obtained for a variety of spiked river waters, sewage and china clay works effluents. Keywords : Rapid trace acrylamide determination ; aqueous solutions ; high- performance liquid chromatography ; environmental flollution Acrylamide polymers are used extensively to aid water clarification at water-treatment works1 and for conditioning sludges2 The manufacture of these polyelectrolytes results in contami- nation by acrylamide monomer,l and the high chronic toxicity of acrylamide makes it undesirable in potable water ~upplies.~-~ The recommended maximum acrylamide con- centration in potable waters is 0.25 pg 1-1.6 Commercial polyelectrolytes used as coagulants during the preparation of potable waters may not contain in excess of 0.05% of a~rylamide.~~~ However, unregulated polyelectrolytes may be used for effluent treatment and such poly- acrylamides may contain up to 5% of m~norner.~ Recent advances in the polymerisation procedures have meant that many non-potable water-grade polymers now have monomer contents of less than o.3y0.839 Reduction in the acrylamide content of polymer^^^^ seems to have reduced the impact of acrylamide pollution resulting from the use of certain poly- acrylamides.1° The use of high monomer content polymers, acrylamide based sewer grouts,ll the recycling of waste paperlo and the manufacture of polyacrylamides’ may still result in significant acrylamide pollution of water courses.Studies in the laboratory12 and in situ13 have inferred that little loss of acrylamide from water would occur as a result of adsorption by inanimate matter. Although biodegradation of acrylamide O C C U T S ~ ’ ~ ~ ~ ~ both the laboratory7314 and the in situ13 studies have suggested that the residence times would be in the order of days for natural waters. Discharges of acrylamide at concentrations of between 6 and 50 pg 1-1 have been shown to reduce greatly a river’s aquatic invertebrate diversity.13 As acrylamide will pass through a water-treatment works7 and is stable in tap water for in excess of 2 months14 it would be prudent to control both the environmental and human health impact of acrylamide pollution through the monitoring of effluent discharges.A rapid method for the detection of acrylamide at trace levels is therefore required. Previous methods of direct analysis16-19 have had detection limits of between 100 and 1000 pg 1-l. For grossly polluted samples we have found such techniques to be unsuitable at these levels unless a suitable clean-up stage was utilised. Concentration of acrylamide from an aqueous solution appears to be possible only by derivatisation.20 Techniques involving the production of a suitable acrylamide are of necessity somewhat time consuming. This paper reports a sensitive (5 pg 1-I), rapid, routine method of analysis using a resin clean-up treatment followed by high-performance liquid chromatographic detection.750 BROWN et d.: RAPID SCREENING TECHNIQUE UTILISING HPLC And‘YSt, ‘Vd. 107 Experimental Apparatus A Shandon Southern Products Ltd. column packer was used to prepare analytical columns packed with Hypersil octadecylsilane (ODS) phase. All other high-performance liquid chromatographic columns were purchased as packed columns from the appropriate manu- facturers. A Millpore Swinnex filtration apparatus (0.25 cm diameter) fitted with Whatman GF/F (0.7 pm) filters was used to filter the samples. A Waters 6000A solvent delivery system, fitted with a Hypersil ODS 5 pm column (100 x 5 mm i.d.) and either a Perkin-Elmer LC75 spectrophotometer with autocontrol unit or a Pye LC UV and a Perkin-Elmer 023 recorder were used.A Waters U6K injection valve was employed to inject the sample on to the analytical column. A Scientific Glass Engineering syringe (100 pl) was used to inject the sample. Pre - treatment Column Elga mixed anionic (OH-) and cationic (H+) resins, as used in their commercially available de-ioniser packs, were used without any prior treatment. The hydrophobic resin (Amberlite XAD-2) was washed successively with distilled water, propan-2-01, hexane, propan-2-01 and distilled water prior to use. An aliquot (100 g) was batch cleaned and then stored in distilled water in the dark prior to use. Resin columns were produced from glass tubing (240 x 10 mm) to which a funnel (approximately 50 ml), a glass sinter and a tap were fitted. The column was filled with distilled water prior to packing. Sufficient mixed Elga resins were added to give a bed height of approximately 140 mm.The differing resin densities invariably led to banding of these resins but this did not adversely affect the columns’ performance. The remaining 100 mm was filled with Amberlite XAD-2 and the column top plugged with glass-wool. New columns were eluted with approximately 100 ml of distilled water prior to use. Column life is dependent on the purity of samples being analysed, ranging from more than 100 samples for certain river waters to less than 10 samples for raw sewage. Noticeable discoloration of the resins, particularly the white Amberlite XAD-2, usually occurred prior to column failure. Columns should be examined at suitable periods by treatment with acrylamide standard solutions (10 pg 1-l) to test for acrylamide adsorption.Columns should be discarded if they either begin to adsorb acrylamide or they allow levels of inorganic ions or dissolved organic compounds, which seriously interfere with the high-performance liquid chromatographic analysis, to elute. Exhaustive tests have not been carried out on the complete range of resin products available, but it should be noted that certain products appear either to have a limited adsorptive capacity for acrylamide or to leach unacceptable levels of components from the column itself. Flow-rates through the column depended on the closeness of the packing. Flows of between 2 and 10 ml min-l have proved suitable for the analysis. Analytical Column A selection of columns were examined to find the most appropriate.The columns tested were as follows: self-packed Hypersil ODS (particle size 5-7 pm, 100 x 5 mm i.d.); Spheri- sorb ODS (particle size 10 pm, 250 x 4 mm i.d.) (Phase Separations Ltd.) ; Partisil PXS, ODs-2 (particle size 10-25 pm, 250 x 4 mm id.) (Whatman Inc.); and pBondapak ODS (particle size 10 pm, 250 x 4 mm i.d.) (Waters Associates Inc.). All columns were found to elute a contaminant peak (as shown for the Hypersil column in Fig. 1) shortly after the void volume. The nature of this peak is not known but its magnitude usually increased with column use. The use of the relatively short wavelength (202 nm) used for acrylamide determination, and the high ultraviolet spectrophotometer sensitivity setting necessary to detect low concentrations of acrylamide (less than 100 pg 1-l) greatly emphasised the importance of this interference, which would not be noticed in less critical analyses.The commercially packed columns examined achieved resolution of acrylamide from this interference by varying degrees, but were not capable of maintaining sufficient resolution when used regularly to allow acrylamide to be assayed at concentrations of less than 20 pg 1-l. The acrylamide elution pattern when using the Hypersil phase was radically different from that of the other packing materials. When using Hypersil, acrylamide eluted well before the interference, whilst with the other phases acrylamide eluted either on the Columns should be re-packed if allowed to run dry.JuZy, 1982 FOR ASSESSING ACRYLAMIDE CONTAMINATION I N EFFLUENTS 75 1 trailing edge or shortly after the contaminant peak.To date, several columns have been packed with different batches of Hypersil phase and all showed comparable performances. Reagents Solvents. HPLC solvent deaeration gas. Helium. Acrylamide monomer. Propan-2-01 and hexane (Fisons HPLC grade). Electrophoresis grade, purity greater than 99%. A solution of acrylamide (100mg1-l) in distilled water was prepared for each set of analyses. A dilute acrylamide solution (1 mg 1-1) was prepared from the former and aliquots of the latter solution were used to prepare spiked samples (Fig. 1). Determination of Acrylamide Monomer in Aqueous Solutions Samples were immediately filtered on collection, using a Millpore Swinnex fitted with Whatman GF/F filters, to reduce microheterotrophic activity, and stored in a refrigerator (4 "C) in the dark prior to analysis. For the samples tested no acrylamide loss was noted during 1 week's storage. Samples may be stored without filtration provided microbial activity is reduced either by boiling for 30s or by the addition of mercury(I1) chloride. Either technique should be accomplished as soon as possible.Whilst most samples may be stored untreated for 16 h without acrylamide loss, more rapid loss of acrylamide has been noted for samples with high microbial activity (e.g., the waters of an activated sludge tank). Samples stored unfiltered should be filtered (using Whatman GF/F filters) prior to resin treatment. Paper filters should not be used as they may leach acrylamide owing to the use of polyacrylamides in the manufacture of many paper products.The acrylamide solution (or aqueous sample) was passed through the mixed-resin bed and the first bed volume discarded. The second bed volume was collected. The column was next washed with at least three bed volumes of distilled water prior to the next analysis. An aliquot (100 pl) of the sample eluate was injected into the injection valve and chromato- graphed under the following conditions : column, Hypersil ODS 5 pm (100 x 5 mm) ; mobile phase, distilled water; flow-rate, 1 ml min-l; pressure, 1000 lb in-2; detector wavelength, 202 nm; chart speed, 0.5 cm min-l ; and absorbance scale, 0.01. According to the sensitivity required duplicate aliquots of sample (1-100 pl) were injected.For maximum sensitivity the absorbance scale was set on 0.004 (Perkin-Elmer LC75 0.04 a.u.f.s. into a 1-mV recorder) or 0.005 (Pye LC UV). Peak heights of less than three times the base-line noise were not considered significant. Preparation of Calibration Graphs Standard solutions of acrylamide (5, 7, 10, 50, 70 and 100 pg 1-1) were treated in the same manner as the samples described previously. The slope of the calibration graph [absorbance (peak height) versus concentration] did not change by what was considered an appreciable amount for an eluent flow-rate of 1 ml min-l over the period studied (Fig. 2). Calculations 100-pl injection on an absorbance scale of 0.004 a.u.f.s., i.e., The concentration of acrylamide in a sample was calculated from the absorbance of a mean absorbance of concentrate volume of concentrate injected absorbance scale loo 0.004 Total absorbance = Interferences The levels of inorganic and organic ultraviolet absorbing molecules present in the river water, sewage and china clay effluents tested prevented analysis for acrylamide if the resin clean-up technique was not utilised.Results and Discussion Selection of Liquid Chromatographic Column and Wavelength The Spherisorb, Bondapak and Whatman ODS columns tested were unable to resolve752 BROWN et al. : RAPID SCREENING TECHNIQUE UTILISING HPLC Analyst, VoZ. 107 acrylamide from an unknown contaminant. This contaminant, although also present when using a Hypersil ODS column (Fig. l), was adequately resolved and allowed acrylamide determinations at maximum stable sensitivity (0.004 a.u.f A).The detection wavelength of 202 nm was chosen instead of the wavelength of maximum absorbance (197 nm) as (i) the selection of 202 nm as opposed to 197 nm considerably reduced interference from the un- retained inorganic ions eluting immediately before the acrylamide peak and the unknown contaminant eluting soon afterwards (Fig. 1) ; (ii) absorbance ratioing of the acrylamide peak at 195, 197, 199, 202, 205 and 210 nm for a variety of samples occasionally gave uncharac- teristic ratios for wavelengths below 202 nm; and (Z) the sensitivity loss at wavelengths longer than 202 nm prohibited their use. - Analysis of Water Samples for Acrylamide Monomer Water types tested included peat bog waters with high humic acid content, the waters of various local rivers, china clay process waters, paper mill process waters, untreated, primary and secondary treated sewage from a sewage works receiving greater than 90% domestic sewage and untreated and primary treated sewage from a works that received a high level of industrial discharges.Acrylamide was only detected in river and sewage samples that had been artificially dosed with acrylamide during in situ acrylamide degradation studies13 and in paper mill process waters.1° Examples of all sample types were dosed with 5, 10 and A selection of water types were examined to assess the suitability of the method. C 70 60 50 a C $40 2 n a 30 20 10 1 L b) i 3 1 2 3 4 5 6 7 8 910 0 1 2 3 4 5 6 7 8910 ‘“’r? d)--lt # E J 01 2 3 4 5 67 8 910 0 1 2 3 4 5 6 7 8910 Ti me;m i n Fig.1. Resolution of acrylamide from interferences using a Hypersil ODS 5 pm column and eluting with distilled water. Chromatographic conditions : column, Hypersil ODS 5 pm (100 x 5 mm); eluent, distilled water; flow-rate, 1 ml min-l; pressure, 1000 lb in-2; chart speed, 0.5 cm min-l; wavelength, 202 nm; absorbance scale, 0.004 a.u.f.s. = 100 units. (a) 100-p1 injection of acrylamide spiked distilled water (5 pg 1-I), A = inorganic anions and unretained organic compounds, B = acrylamide (5 pg 1-l) and C = unknown (column artefact?) ; ( b ) , 100-pl injection of acrylamide spiked distilled water (10 pg l-l), D = acrylamide (10 pg 1-l) ; (G) 100-pl injection of sewage effluent, - - - untreated, - resin treated; (d) 100-p1 injec- tion of acrylamide spiked sewage effluent (10 pg l-l, - - - untreated, - resin treated), E = acrylamide (10 pg 1-l).As the acrylamide peak appeared on the trailing edge of the solvent peak, the acrylamide was quantified after drawing a tangent to exclude solvent interference.July, 1982 FOR ASSESSING ACRYLAMIDE CONTAMINATION I N EFFLUENTS 753 100 pg 1-1 of acrylamide and the data used to calculate the precision at these concentrations (Fig. 2). Confirmation of acrylamide concentration for certain acrylamide-containing environmental samples was obtained using the method of Brown and Rhead.20 Excessive interferences were noted even when using the resin technique for certain untreated/primary treated sewage and untreated paper mill process waters.Standard addition techniques were used to assess the detection limits for these samples. Concentrations that could be assayed with a precision of 10% ranged from 20 to 100 pg 1-1 of acrylamide. 180 5 160 E 2 140 t : 120 Lc : 100 0 80 2 60 - 40 f! a ([I c. r-" 20 0 20 40 60 80 100 120 Concentration of acrylamide/pg I-' Fig. 2. Acrylamide calibration graph. Chromatographic con- ditions as in Fig. 1. Absorbances were calculated for a 100-pl injec- tion. The samples were: 0, 5, 7, 10, 50, 70 and 100 pg 1-1 unspiked and acrylamide spiked distilled water, river water, china clay and sewage effluent samples. Conclusions The neutral hydrophilic nature of acrylamide prevents its adsorption by certain anionic, cationic and hydrophobic resins. This property has been utilised in the proposed method to develop a mixed resin clean-up technique that effectively results in the analysis of acrylamide in a solution approximating to organic free, de-ionised water, even for a polluted sample such as a sewage effluent.The presence of ultraviolet absorbing inorganic ions and organic compounds in most natural and polluted waters prevented the determination of trace concentrations (5 pg 1-1) of acrylamide monomer if the proposed clean-up technique was not utilised (Fig. 1). Certain untreated discharges contained such high levels of inter- ferences that the detection levels were raised to more than 100 mg 1-1 without resin treatment and from 20 to 100 pg 1-1 ( & l O ~ o ) with clean-up. The latter levels are, however, still suitable to screen certain discharges, in particular the effluent from polyelectrolyte manu- factures' and waters near sites used for the in situ preparation of acrylamide grouts.ll The instrumental analysis times for this high-performance liquid chromatographic method are directly comparable to those of previous techniques.1*-20 The previous direct methods,16-19 however, have lower sensitivities (100-1 000 pg 1-l) and we have not found them to be applicable, even at this level, for most polluted waters.We found, in practice, that the bromination derivatisation techniques previously used to detect acrylamide at concentrations of 0.2-100 pg 1-1 (either with gas-chromatographicz1*22 or high-performance liquid chromatographicz0 analysis) are both tedious and time consuming.The manipulative requirements etc. needed for the reproducible preparation of a,/3-dibromopropionamide resulted in an average sample throughput of 16 samples per 2-d cycle per person, with a maximum of 32 samples per week per person. Sample turnover has been improved in other754 BROWN, RHEAD AND BANCROFT laboratories (which use the Hashimoto technique22) by having a member of staff specialising solely in the derivatisation stage. Although the resin technique as described may have a short column life when using grossly contaminated samples, this was not too deleterious as preparation of replacement columns was rapid. It takes approximately 3 min to re-pack and 10 min to wash (using 100 ml of distilled water) a new column. Five columns were run concurrently (it takes 30min to repack five columns), thus enabling five samples to be resin treated per 20-min period.As the chromatographic analysis was completed whilst the next batch was being resin treated a sample throughput of 15 samples per hour per person was accomplished. This represents a very significant increase in sample capacity, thus enabling the technique to be used for screening discharges to water courses. The authors thank South West Water, Mr. R. Lisle of Allied Colloids and Dr. P. Sykes of Crosfield Polyelectrolytes. The work in this paper was carried out under contract from the Department of the Environment (DGR 480/306) and publication is with their permission. 1. 2 . 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.22. References Rothwell, E., Water Treat. Exam., 1974, 23, 373. Novak, J. T., and O’Brien, J. H., J . Water Pollut. Control Fed., 1975, 47, 2397. McCollister, D. D., Oyen, F., and Rowe, V. K., Toxicol. Afipl. Pharmacol., 1964, 6, 172. Davenport, J. G., Farrel, D. F., and Marksumi, S., Neurology, 1976, 26, 919. Keeson, C. M., Lawson, D. H., and Baird, A. W., Postgrad. Med. J., 1977, 53, 16. Ministry of Housing and Local Government, Water Treat. Exam., 1969, 18, 90. Croll, B. T., Arkell, G. M., and Hodge, R. P., Water Res., 1974, 8, 989. Lisle, R., Allied Colloids, Bradford, personal communication, 1980. Sykes, P., Crosfield Polyelectrolytes, Warrington, personal communication, 1980. Brown, L., Rhead, M. M., and Bancroft, K. C. C., Water Pollut. Control, 1980, 79, 507. Igisu, H., Goto, I., Kawamura, Y., Kato, M., Izumi, K., and Kuroiwa, Y., J . Neurol. Psychiatry, Brown, L., Bancroft, K. C. C., and Rhead, M. M., Water Res., 1980, 14, 779. Brown, L., Rhead, M. M., Hill, D., and Bancroft, K. C. C., Water Res., 1982, in the press. Brown, L., Rhead, M. M., Bancroft, K. C. C., and Allen, N., Water Res., 1980, 14, 775. Lande, S. S., Bosch, S. J., and Howard, P. H., J . Environ. Qual., 1979, 8, 133. Betso, S. R., and McLean, J. D., Anal. Chem., 1976, 48, 766. Husser, E. R., Stehl, R. H., Price, D. R., and DeLap, R. A., Anal. Chem., 1977, 49, 154. Ludwig, F. J., Sr., and Besand, M. F., Anal. Chem., 1978, 50, 185. Skelly, N. E., and Husser, E. R., Anal. Chem., 1978, 50, 1959. Brown, L., and Rhead, M. M., Analyst, 1979, 104, 391. Croll, B. T., and Simkins, G. M., Analyst, 1972, 97, 281. Hashimoto, A., Analyst, 1976, 101, 932. 1975, 38, 581. Received October 21st, 1981 Accepted February 22nd, 1982

ISSN:0003-2654    

DOI:10.1039/AN9820700749

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

Analyst, July, 1982, Vol. 107, pp. 755-760 755 Biogenic Amine Resolution in Tissue Extracts of Rat Brain Using lon-pair High-performance Liquid Chromatography with Electrochemical Detection * S. A. Pleece, P. H. Redfernt and C. M. Riley School of Pharmacy and Pharmacology, University of Bath, Claverton Down, Bath, BA2 7 A Y E. Tomlinson Sub-faculty of Pharmacy, University of Amsterdam, Plantage Muidergracht 24, Amsterdam, The Nether- lands A method for determining biogenic amine concentrations in extracts of brain tissue using high-performance liquid chromatography with electrochemical detection is described. Using a Hypersil stationary phase and a mobile phase based on methanol and containing 0.1 yo V / V sulphuric acid, the effects of varying methanol concentrations and of varying concentrations of the ion- pairing sodium octylsulphate have been investigated.The optimum separation conditions were achieved with a mobile phase containing 10% V/V methanol, 0.1% V / V sulphuric acid and 6 x M sodium octylsulphate. Under these conditions, and with a potential of 1.1 V (positive) at the glassy carbon working electrode, it is demonstrated that picogram amounts of noradrenaline, dopamine, 5-hydroxytryptamine, trypto- phan and 5-hydroxyindole acetic acid can be determined in tissue extracts of rat brain. Keywords ; High-performance liquid chromatography ; catecholamines ; indole- amines ; electrochemical detection Investigation of the mode of action of centrally acting drugs commonly requires measurement of neurotransmitter concentrations in small samples of brain tissue.Until recently, radio- enzymatic assays by, for example, Palkovits et aZ.1 have provided the most sensitive method for the determination of non-peptide transmitters such as noradrenaline (NA), Ei-hydroxy- tryptamine (5HT) and related compounds. However, with the development of high- performance liquid chromatography it has become widely recognised that this technique can form the basis of an assay system, which is equally sensitive and at the same time considerably simpler and quicker. Many groups have developed high-performance liquid chromatographic assays for either catecholamines, e.g., Moyer and Jiang2 and Wagner et aZ.,3 or indoleamines, e.g., Koch and Ki~singer.~ The procedure described here, developed during the course of an investigation of the effects of antidepressant drugs on neuronal mechanisms controlling appetite, allows the determination of both catecholamines and indoleamines in the same sample.The method, which uses an ion-pairing agent to optimise the separation of cationic and neutral species of interest, is widely applicable for the determination of monoamines in body fluids and tissue extracts. Experimental Apparatus The high-performance liquid chromatographic system was custom built from a Constametric I11 pump (Jones Chromatography, Llanbradach), an RE 541.20 Servoscribe recorder (Jones Chromatography), a 7125 Rheodyne injection valve fitted with a 20-4 loop (Jones Chromato- graphy) and a Bioanalytical Systems LC-4 amperometric detector (Anachem Ltd., Luton, Bedfordshire) connected to a flow cell fitted with a glassy carbon working electrode and a silver - silver chloride reference electrode.The column components were supplied by HETP (Macclesfield, Cheshire). The column temperature was maintained at 35.0 & 0.1 "C by immersion in a thermostatically controlled water-bath (Gallenkamp, Type 400-010, London). * A preliminary account of this work was presented as a poster communication at the Joint NL - UK Symposium on Quantitative Organic Analysis, Noordwij kerhout, The Netherlands, April 22-24, 1981. t To whom correspondence should be addressed. -756 PLEECE et at?. : BIOGENIC AMINE RESOLUTION IN TISSUE Analyst, Vd. I07 Materials Dopamine hydrochloride, 5-hydroxyindole-3-acetic acid, 5-hydroxytryptamine oxalate and tryptophan were supplied as Sigma-grade reagents by Sigma (Poole, Dorset); nor- adrenaline bitartrate (puriss grade) was supplied by Koch-Light (Colnbrook, Buckingham- shire) and sodium octylsulphate was supplied as 98% pure by Cambrian Chemicals and was used as received.All other chemicals were of AnalaR grade and were supplied by either BDH Chemicals (Poole , Dorset) or Fisons (Loughborough, Leicestershire) . The chromato- graphic stationary phase used was 5 pm ODS Hypersil (Shandon) supplied by Jones Chroma- tography. Alumina (Type A, 100 - 200 mesh) was supplied by Laport Industries (Widnes, Cheshire). Water was doubly distilled from an all-glass still. Procedures ODS Hypersil (2.7 g) was slurried in methanol - chloroform (1 + 9, 75ml) and packed upwards into a 200 x 5 mm i d .stainless-steel column as described by Bristow et The mobile phases were freshly prepared each day and de-gassed by purging with helium. At the end of each working day, the column was cleaned by the passage of 100 ml of methanol - water (1 + 1) followed by 100 ml of methanol. All other chromatographic procedures were as described previously.6 Female Wistar rats (University of Bath strain) were used throughout this study and were sacrificed at the same time each day by decapitation. The hypothalamus was removed, transferred immediately into a sealable polythene bag and stored in liquid nitrogen. At the time of assay each hypothalamus, while still frozen, was crushed to a fine powder in a metal anvil. The contents of the anvil were transferred into a centrifuge tube and extracted into 1 ml of water containing 10% V/V methanol, O.lyo sulphuric acid and 6 x 10-4 M sodium octylsulphate by vortex mixing for 5 min.The tube was centrifuged for 5 min at 5000 g, the supernatant liquid was removed by aspiration and the extract then passed through a 0.5-pm filter (Millipore, London). A 20-4 aliquot was then injected on to the high-perform- ance liquid chromatographic column. Results In high-performance liquid chromatography, the isocratic separation of compounds exhibiting a wide range of affinities for the stationary phase requires optimisation of the mobile phase composition, so that the compound retained longest is eluted within a reason- able analysis time and the compound with the shortest retention time is well separated from the solvent peak. Using the ion-pair high-performance liquid chromatographic mode, Riley et aZ.' investi- gated these criteria by examining the effect of methanol concentration and pairing-ion concentration on the retention of the five solutes of interest. Under the conditions employed (Hypersil stationary phase with a mobile phase containing 0.1% V/V sulphuric acid), 5- hydroxyindoleacetic acid (5HlAA) was un-ionised and noradrenaline (NA), dopamine (DA) , 5-hydroxytryptamine (5HT) and tryptophan (TRY) were in their cationic form.The effect of increasing the concentrations of the ion-pairing agent, sodium octylsulphate, is shown in Table I, from which it can be seen that the retention of the cationic species was directly related to the sodium octylsulphate concentrations, whereas, as was to be expected, the retention of the un-ionised 5HlA.A was unaffected.Table I1 shows that the effect of increasing the methanol concentration in the mobile phase was to decrease the retention of all five solutes. Potential of the Electrochemical Detector The effect of changing the potential difference between the electrodes of the electro- chemical detector is shown in Fig. 1. For NA and DA the critical potential is between +0.4 and +0.6 V; 5HlAA and 5HT also show a significant increase in peak height at this potential, presumably indicating conversion of the aromatic OH group into a quinone structure. A higher potential, between 0.8 and 1.2 V (positive) was required to detect ionisation of the heterocyclic nitrogen of TRY, 5HT and 5HlAA.Optimum separation conditions were achieved with a mobile phase containing 10% V/V methanol, 0.1% V/V sulphuric acid and 6 x loA4 M sodium octylsulphate.Jzcly, 1982 180 160 E E 2 120 2 8 0 - a" 4 0 - 0 - (D Y c 0, 1- EXTRACTS OF RAT BRAIN USING ION-PAIR HPLC 5HIAA - - NA DA - TABLE I EFFECT OF ION-PAIRING AGENT (SODIUM OCTYLSULPHATE) CONCENTRATION ON RETENTION Sodium octylsulphate concentration*/M x lo-* 0 h Solute k NA .. .. . . 0.12 DA . . .. . . 0.27 6-HT . . .. . . 0.64 TRY .. .. . . 2.59 6-HIAA . . . . 1.66 4 & e' k 0.00 0.58 0.01 1.07 0.02 2.60 0.00 10.00 t 1.77 6 & e' k 0.00 0.63 0.01 1.30 0.03 3.02 t 1.52 -0.01 13.1 8 & e' k 0.00 0.77 - 0.02 1.63 - 0.03 3.47 -0.04 14.7 t 1.51 * k = Capacity ratio and e' = mobile phase enhancement factor.t Not retained by ion-pair mechanism. TABLE I1 EFFECT OF METHANOL CONCENTRATION ON RETENTION Methanol concentration, yo V/ V 757 f I 20 10 * 1 Solute k e' k NA . . .. . . 0.63 0.00 0.67 DA . . .. . . 1.30 0.07 0.53 5HT .. . . 3.02 0.13 4.61 5HIAA. .. . . 1.62 0.13 2.78 TRY .. .. . . 13.1 0.14 18.7 Fig. 2(a) shows that chromatography of hypothalamic extracts under these conditions produced separation of the five solutes from each other, althoUgh NA was not adequately separated from unidentified material present in the mobile phase. Two strategems were adopted to overcome this problem without prolonging the appearance of TRY. The first [as illustrated in Fig. 2(b)] involved the use of two flow-rates. Initial separation of NA and DA is achieved with a flow-rate of 1 ml min-l; with the appearance of DA, the flow-rate is then increased to 2 ml min-l.The second method avoided the use of two flow-rates but included a separation stage for NA and DA. After the methanol extrac- tion of powdered brain, a 0.5-ml aliquot of the centrifuged supernatant liquid was adjusted to pH 8.6 with 3 M Tris buffer and shaken for 15 min with 90 mg of activated alumina. The supernatant liquid was removed and the alumina washed, once with 1 ml of 0.06 M Tris buffer at pH 8.6 and twice with water. Desorption from the alumina was achieved758 PLEECE et al. : BIOGENIC AMINE RESOLUTION IN TISSUE Analyst, Vol. I07 Inject alumina extract Fig. 2. (a) Elution of amines from hypothalamic extract; (b) use of two flow-rates, 1 ml min-l initially, increased to 2 ml min-l after the appearance of DA; and (c) prior alumina separation of NA and DA.using 0.5 ml of 0.05 M perchloric acid. It was convenient to inject the 20-4 aliquot of this sample containing NA and DA 10min after the 2 0 4 aliquot of the original extract con- taining the indoleamines, so that NA and DA were detected between the peaks of 5HT and TRY [Fig. 2(c)]. This method has the further attraction of producing two DA peaks, thus providing an added cross-reference of extraction efficiency and reproducibility. A comparison of aqueous standards and brain extracts, together with an indication of percentage recovery is given in Table 111. TABLE I11 COMPARISON OF AQUEOUS STANDARDS AND BRAIN EXTRACTS Aqueous standards of: NA, 1 pg ml-' and DA, 5HT, 5HIAA and TRY, 200 ng ml-l.Solute r A \ Parameter NA DA 5HT 5HIAA TRY Peak height of standard/cm . . .. . . 7.0 6.5 4.7 7.5 4.1 Peak height of samplelcm . . .. . . 6.2 2.4 3.1 2.8 3.1 Peak height of spiked samplelcm . . . . 12.8 8.1 7.0 7.8 6.9 Recovery, yo . . .. .. .. . . 94 88 83 67 93 Discussion Increasing the pairing-ions concentration enhanced the retention of the cationic species , and decreasing the methanol concentration enhanced the retention of all solutes. A mobile phase enhancement factor, e, may be defined as the effect of changing the mobile phase composition so as to produce an enhancement of retention and is given by: .. .. ' (1) e = 10g(k,k~-~) . . .. where k, and k, are the capacity ratios of a particular solute under different mobile phaseJuly, 1982 EXTRACTS OF RAT BRAIN USING ION-PAIR HPLC 759 conditions. The relative mobile phase enhancement factor, e', may be defined as the enhancement factor of a solute, X, relative to the enhancement of a reference compound. In this instance the least retained solute, NA, was chosen as the reference solute.The relative mobile phase enhancement factor is given by .. . . - ' (2) e' = ex - eNA .. .. The relative mobile phase enhancement factors for each subsequent change in mobile phase composition that produced an increase in retention are given in Tables I and I1 and it can be seen that e' is independent of the pairing-ion concentration but is dependent solely an the methanol concentration. This shows that sodium octylsulphate enhances the retention of the cationic species to the same extent; however, the methanol has the greatest enhancing effect on the most retained compound, because e' increases with increasing retention.These results are indicative of a non-concomitant relationship between pairing-ion concentration and organic modifier concentration and may be rationalised further by the application of linear free energy relationships. The relationship between the capacity ratio of a solute x, k , and the capacity ratio of the least retained, reference compound, kNA, is given by where T is an extra thermodynamic parameter describing the ph ysico-chemical difference between the two solutes responsible for their separation. Rearranging equation (3) gives Previously it has been shown8 that selectivity, r, in both ion-pair and non ion-pair reversed phase high-performance liquid chromatography, may be related to hydrophobic fragmental constants (eg., Hansch 7~ valuesg) by means of linear free-energy relationship, e.g., .. .. - - (5) r = a Z f + b .. . . Substitution of equation (5) into equation (4) gives It has been shown that when alcoholic mobile phases are employed in conjunction with hydrophobic stationary phases containing no residual silanol groups, the intercept term is not significantly different from zero and equation (6) reduces to Equation (7) is a linear free energy relationship between logk, and relative solute hydro- phobicity. The retention data given in Tables I and I1 have been analysed by least-squares linear regression (Table IV) using hydrophobic fragmental constants,1° calculated relative to TABLE IV RETENTION DATA ANALYSED BY LEAST-SQUARES REGRESSION Hydrophobic fragmental constants calculated relative to NA (ie., ~ N A = 0).Mobile phase bgki?A A I 7- Methanol, Sodium octylsulphate/ Standard % v/v M x 10-4 Coefficient deviation 20 0 -0.944 0.044 20 4 -0.347 0.041 20 6 -0.247 0.051 20 8 -0.161 0.075 10 6 -0.189 0.056 Regression analysis L .I a - Coefficient deviation coefficient of correlation 1.16 0.058 0.997 0.049 1.17 0.061 0.997 0.051 1.16 0.075 0.997 0.063 1.14 0.079 0.994 0.067 1.27 0.081 0.994 0.069 Standard Correlation Standard deviation760 PLEECE, REDFERN, RILEY AND TOMLINSON noradrenaline, that is XfNA = 0. It can be seen from Table IV that the slope coefficient, a, (which determines the magnitude of the selectivity of the chromatographic system), is independent of pairing-ion concentration and that it decreases with increasing methanol concentration.The intercept term, on the other hand, depends on both methanol and pairing-ion concentration. These conditions indicate that the desired retention of the reference compound, NA, may be achieved with different combinations of methanol and pairing-ion concentrations, and that ideally the over-all analysis time could be reduced owing to a reduction in the value of a by using higher concentrations of both methanol and sodium octylsulphate. Unfortunately this particular rationale is inapplicable in the present study as the decrease in retention for 5HIAA owing to increasing methanol concentration cannot be compensated for by an increase in sodium octylsulphate concentration.Therefore, for the compounds discussed, in order to produce separation within an acceptable time span, it is necessary either to change the flow-rate or to use alumina to separate out NA and DA. Both of these methods can be successfully used to assay catecholamines and indoleamines in picogram amounts in fragments of brain tissue. The peak heights (in centimetres -+ standard error of the mean, n = 4) obtained for a 20 pg ml-l standard solution were as follows: NA 9.5 & <0.1; DA 2.3 &- 0.1; 5HIAA 2.5 & 0.1; 5HT 1.5 & < O . l ; and TRY 1.1 & 0.1. With apparatus that is only slightly more sophisticated it should be possible to incorporate the alumina separation of DA and NA into a preparative pre-column; equally, if the number of assays to be performed warranted two high-performance liquid chromatographic systems in parallel, NA and DA could be assayed in a system incorporating ion-pairing sodium octylsulphate, while if indoleamines are separated using a mobile phase devoid of sodium octylsulphate, the TRY peak will appear within 8 min. S.A.P. is in receipt of a research studentship from the Pharmaceutical Society of Great Britain. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. References Palkovits, M., Browstein, M., Saavedra, J. M., and Axelrod, J., Brain Res., 1974, 77, 137. Moger, T. P., and Jiang, M., J . Chromatogr., 1978, 153, 365. Wagner, J., Palfreyman, M., and Zraika, M., J . Chromatogr., 1979, 164, 41. Koch, D. D., and Kissinger, P. T., J . Chromatogr., 1979, 164, 441. Bristow, P. A., Brittain, P. N., Riley, C. M., and Williamson, B. F., J . Chromatogr., 1977, 131, 57. Riley, C . M., Tomlinson, E., Jefferies, T. M., and Redfern, P. H., J . Chromatogr., 1979, 162, 153. Riley, C. M., Tomlinson, E., and Jefferies, T. M., J . Chromatogr., 1979, 185, 197. Tomlinson, E., Jefferies, T. M., and Riley, C . M., J . Chromatogr., 1978, 159, 315. Iwasa, J., Fujita, T., and Hansch, C., J . Med. Chem., 1965, 8, 150. Rekker, R. F., “The Hydrophobic Fragmental Constant,” Elsevier, Amsterdam, 1977. Received November 9th, 1981 Accepted January 20th, 1982

ISSN:0003-2654    

DOI:10.1039/AN9820700755

出版商:RSC    

年代:1982

数据来源: RSC