INORGANIC ANALYSIS Analytical chemistry, upon which all other branches of the sci ence are ultimately dependent, has its source in the writings of the chemist-pharmacists of the iatrochemical period. Tachenius rec ognized many metals in solution together by their reactions in the wet way, but it was not until the phlogistic period that any sys tematic progress was made, more particularly by Boyle, to whom we owe the term analysis. Many of the reactions still utilized in qualitative work were known to Boyle, who also introduced cer tain plant extracts, notably litmus, for the recognition of acids and bases. Other workers of this period were Hofmann, Marggraf (who introduced the microscope into chemistry), Scheele, and especially Bergman, who devised methods by which the metals might be separated into groups according to their behaviour with certain reagents. The blowpipe (q.v.) was introduced into dry qualitative analysis by the mineralogist, Cronstedt, who applied it to the examination of ores, although it was left to Berzelius and Hausmann to bring about its general application. The colorations which sodium and potassium salts impart to the flame were known to Marggraf, but it is to Bunsen, who, with Kirchhoff, devised spectrum analysis, that the full value of the flame test is due.
While the work of the phlogistic period in this field was mainly of a qualitative character, some attempts of a quantitative nature were made by Marggraf and Black, while Bergman realized that elements need not be separated as such but might be isolated in the form of some suitable insoluble compounds. Klaproth, who with Vauquelin developed the quantitative analysis of minerals, proposed the ignition of precipitates before weighing them, if they were not decomposed thereby. Lavoisier, with his grasp of the importance of the composition by weight of chemical compounds, and because of his systematic use of the balance, must be con sidered as the first great exponent of quantitative analysis. The substantiation by Richter, and, independently by Proust, of the law of constant proportions and the formulation of the atomic theory by Dalton combined to give a fresh impetus to the devel opment of quantitative analysis, subsequently placed on a firm basis by Berzelius. The later researches of Rose, Wohler and Fresenius standardized various methods of analytical chemistry. The quantitative precipitation of metals by the electric current, although known to Faraday, was not applied to analytical chem istry until Gibbs, in 1865, worked out the electrolytic separation of copper. Since then the subject has been extensively studied, more particularly by Classen.
Volumetric analysis, possessing as it does many advantages over the gravimetric method, especially in technical work, was founded by Gay-Lussac, although rough application had been previously made ; the advantages of his carefully worked out methods of chlorimetry (1824), of alkalimetry (1828), and of the chloride– silver titration (183 2) were but slowly recognized. It was not until the application of potassium permanganate to the estimation of iron by Margueritte (1846) and of iodine and sulphurous acid to the estimation of copper and many other substances by Bunsen, that volumetric methods took their proper place alongside the gravimetric. Since that time, these methods have been rapidly developed, particularly by Mohr and Volhard.
The technique of quantitative micro-analysis has been devel oped mainly by Pregl, although Emich had previously indicated the possibility of working with small amounts of inorganic mate rials. The same basic principles and methods hold as for macro analysis, but apparatus and manipulation are necessarily modified when quantities of the order of a few milligrams only are being handled. The same period has seen the advance of the method of potentiometric titration, although this was first put forward in the case of acids and alkalis by Bottger as early as Qualitative Inorganic Analysis.—Any inference drawn from the results of the qualitative examination of a substance, whether by dry or wet methods, must always be confirmed by further tests; moreover, it should not be overlooked that the normal course of a particular reaction may be completely pre vented by the presence of interfering substances.
Dry Methods.—(I) Effect of Heat.—The substance when heated in a hard glass tube may evolve water, condensing on the cooler parts of the tube, derived from substances containing water of crystallization or of constitution and from certain ammonium salts. Its reaction to litmus paper should be observed; alkalinity is probably due to ammonia, acidity to readily decomposable salts.
Oxygen results from the decomposition of peroxides, chlorates, nitrates, iodates and similar oxygenated compounds, and also from oxides of the noble metals ; carbon dioxide from carbonates and organic substances, usually accompanied in the latter case by charring ; sulphur dioxide from many sulphides and thiosulphates. Chlorine, bromine and iodine are evolved from certain halide salts, particularly in the presence of oxidizing agents; oxides of nitrogen from nitrates. Some cyanides evolve cyanogen or hydro cyanic acid and possibly ammonia in the presence of water. All fluosilicates (silicofluorides) are decomposed on heating, evolving silicon fluoride.
A white sublimate is given by ammonium halide salts, mer curous chloride and bromide, oxides of arsenic ; a yellow sub limate may consist of sulphur (reddish brown when molten), arsenic sulphide and mercuric iodide (turning red on rubbing). A blackish deposit results from the condensation of violet iodine vapour or of mercuric sulphide, while most mercury compounds other than those mentioned give a grey deposit. Metallic mercury and arsenic appear as minute globules and a grey mirror.
Colour changes on heating and cooling afford useful informa tion; zinc oxide, titanium dioxide and stannic oxide are of yellow tone when hot, becoming white or whitish on cooling. Lead and bismuth oxides are, short of fusion temperature, brownish red when hot and yellow when cold. The oxides of manganese, cad mium, copper and iron also supply information by this test. All salts of organic acids are decomposed on ignition, usually with charring; the base will be left as carbonate, as oxide, or even reduced to the metallic state, according to the particular metal present.
(2) . Blowpipe Tests.—The substance is heated in a hollow on a charcoal block in the reducing flame of a blowpipe. Oxygenated salts such as chlorates and nitrates cause deflagration ; salts of the alkali metals tend to fuse' into the charcoal. The oxides or stable salts of the alkaline-earth metals and of magnesium, zinc, and aluminium give an infusible residue, white when cold, which in the case of strontia, lime, magnesia or zinc oxide glows brightly in the blowpipe flame. The infusible mass is then moistened with cobalt nitrate solution and again heated; a pink coloration indicates mag nesia (with a tendency to violet if present as phosphate or ar senate), a green colour is afforded by zinc oxide, while a blue sug gests aluminium, possibly silica, or alkaline-earth phosphates. If any other reaction occurs on this direct heating, it is better to carry out a further more decisive test on a mixture of the sub stance with sodium carbonate. Metallic globules are obtained with copper, gold, silver, lead, tin, bismuth and antimony, together with formation of incrustations in the case of silver, lead, tin, bismuth and antimony ; arsenic and cadmium also give analogous incrusta tions.
(3) . Bead Tests.—The colourless fluxes obtained by fusing borax (for coloured salts) or microcosmic salt (for silica, or if the substance is white) on a loop of platinum wire are capable of dissolving many metallic salts, often with production of charac teristic coloured glasses or enamels; the colour may differ accord ing as the bead is heated in the oxidizing or the reducing flame. Chromium gives a green, cobalt a blue bead ; copper in the oxidiz ing flame a bead, green when hot, blue on cooling, and with a tendency to red in the reducing flame. Iron gives yellow and green ish; nickel, brownish and yellow-grey; uranium and vanadium, yellow and green; manganese, violet and colourless, in the oxidiz ing and reducing flames respectively. Silica with microcosmic salt usually, but not invariably, gives a semi-transparent bead.
(4.) Flame Colorations.—When volatilized in a non-luminous flame, the salts of certain metals display characteristic colorations. This test is carried out by introducing into the flame by means of a platinum wire a portion of the substance, preferably moistened with hydrochloric acid, since the chlorides are comparatively vola tile salts. Sodium gives an intense yellow coloration, potassium violet, rubidium and caesium bluish-violet (the latter blue when really pure), calcium red, strontium and lithium crimson, barium yellowish-green and copper bright green ; while lead, arsenic and antimony (which should not be tested on platinum) give a greyish-blue.
The utility of this test is not confined to direct observation, since the coloration due to one metal may completely mask that of others, and for such cases the use of light filters is of assistance. Cobalt glass completely absorbs the sodium yellow and thus per mits the detection of potassium. The lithium and strontium col ours are not cut out by this blue glass ; a saturated solution of chrome alum effectively cuts out colorations due to sodium, cal cium, barium, lithium and strontium and transmits only those due to potassium, caesium and rubidium; solutions of permanganate and indigo-carmine also cut out the sodium colour.
More definite, however, is the information afforded by observa tion of the flame through the spectroscope, which is of particular value for ascertaining the purity of precipitates obtained in quan titative work. Of the rarer metals which give flame colorations, indium (blue) and thallium (green) were discovered by means of the spectroscope, as indeed were also rubidium and caesium. The instrument is applicable to the detection of many other metals by examination of their spark spectra.
l'l'et Methods.—In qualitative analysis of a substance by wet methods it is necessary to obtain it in solution. Portions should be successively tested with water, dilute acids, strong acids and aqua regia (q.v.), first in the cold and then with warming. For sub stances insoluble in all these reagents, fusion with sodium car bonate alone or in admixture with potassium nitrate, followed by acidification, must be employed. In certain cases sodium peroxide or potassium bisulphate (pyrosulphate) may be used, while for the investigation of some refractory silicates treatment with hydro fluoric acid in the presence of sulphuric acid is useful.
The procedure for the detection of metals in solution consists of separating them first into groups by means of volatile reagents ; the group precipitates are then examined separately. From the cold solution, not too concentrated, hydrochloric acid (if this has not already been used in effecting solution) precipitates lead (in part), silver and mercurous salts. The filtrate is rendered nearly neutral and sulphuretted hydrogen passed in to saturation ; this causes mercury (-ic), lead (incompletely removed in the fore going operation), copper, bismuth, cadmium, arsenic, antimony and tin to be precipitated as sulphides ; the filtrate from these is boiled until free from sulphuretted hydrogen, finally with addition of enough oxidizing agent such as nitric acid to convert iron from the ferrous to the ferric state. Ammonium chloride and ammonia are now added to precipitate aluminium, chromium, iron and titanium as hydroxides; if phosphoric acid is present, however, these metals are precipitated in part or wholly as phosphates, pos sibly together with some of the metals of later groups. In this case the precipitate is dissolved in a minimum of hydrochloric acid, a portion of the solution being specially tested for iron, and the remainder boiled with ammonium acetate and a slight excess of ferric chloride, whereby the phosphates are completely elimi nated as ferric phosphate, together with the excess of iron as basic acetate ; the filtrate is combined with that from the ammonia precipitation. It may be noted that some other acid radicals, e.g., fluorides, borates and silicates, also interfere by precipitation at this stage, while in the presence of certain organic compounds, such as tartaric acid, precipitation in this group may be completely inhibited. From the filtrate from the iron group, ammonium sul phide precipitates nickel, cobalt, zinc and manganese, to the fil trate from which ammonium carbonate is added. A precipitate is obtained in the presence of calcium, barium and strontium, leav ing sodium, potassium and magnesium in solution, the last of these metals not being precipitated by ammonium carbonate in the pres ence of ammonium salts. This completes the separation into groups.
Group I.—The white precipitate formed by cold hydrochloric acid is boiled with a considerable quantity of water, and the solu tion filtered immediately ; lead chloride dissolves and may be identified by the yellow precipitate, insoluble in acetic acid but soluble in caustic soda, formed with potassium chromate. The residue on the filter paper is digested with ammonia, whereby silver chloride passes into solution, leaving a black residue of in definite composition which retains the mercury. Silver chloride is reprecipitated from the ammoniacal solution by nitric acid.
Group II.—The precipitate formed by sulphuretted hydro gen may contain the black mercuric, lead and copper sulphides, dark brown bismuth and stannous sulphides, yellow cadmium and arsenic sulphides, orange-red antimony sulphide, dirty yellow stan nic sulphide and whitish sulphur, this last resulting from the action on sulphuretted hydrogen of oxidizing agents such as ferric salts, chromates, nitrates, etc. Warming with yellow ammonium sul phide dissolves out the arsenic, antimony and tin sulphides, which are reprecipitated from the solution of sulpho-salts, after filtra tion, by addition of hydrochloric acid ; tin is always reprecipitated as yellowish stannic sulphide. If traces of copper are being looked for, sodium sulphide should be used for this extraction ; polysul phides are necessary to dissolve stannous sulphide. The mixture of the three sulphides is digested with warm ammonium carbonate solution, which dissolves the arsenic ; this is reprecipitated on acidifying, after filtration, and confirmed by due tests. The residual sulphides of antimony and tin are dissolved in strong hydrochloric acid and these two elements tested for individually. When arsenates are present it is preferable to add sufficient sul phurous acid solution to reduce them to arsenites before passing in sulphuretted hydrogen.
The residue from the ammonium or sodium sulphide digestion is warmed with moderately dilute nitric acid, which dissolves the remaining sulphides, other than that of mercury, with separation of sulphur. A little sulphuric acid is added to the solution, to gether with an equal volume of alcohol; any white precipitate is lead sulphate. As the alcohol must now be boiled off, it is advis able to test separately for lead; if this metal is absent alcohol need not be used. Ammonia in excess added to the filtrate from the lead precipitates white bismuth hydroxide ; in the event of the filtrate being colourless, direct passage of sulphuretted hydro gen precipitates yellow cadmium sulphide. A blue filtrate from bismuth indicates copper, and in this case sufficient potassium cyanide must be added to destroy the blue colour before testing for cadmium.
Group IIIa.—In the absence of phosphates (the presence of which is readily ascertained by warming with ammonium molyb date in nitric acid solution) this group consists of the gelatinous hydroxides of titanium and aluminium (white), iron (red-brown) and chromium (greenish), usually accompanied by a little zinc (especially when chromium is present) and manganese, and pos sibly traces of the group IV. metals precipitated by carbonate present in the ammonia. (To avoid this last contamination in quantitative work hexamethylenetetramine has been suggested in place of ammonia.) The precipitate is dissolved in a little hydro chloric acid ; titanium is readily detected in a portion of the solu tion by the yellow colour it affords with hydrogen peroxide in acid solution. Another portion of the solution is tested for iron with potassium ferrocyanide (blue precipitate) or thiocyanate (deep red coloration), while the third portion is boiled with sodium peroxide until the excess of reagent is decomposed; this dissolves aluminium and chromium hydroxides, the latter being converted to the yellow chromate. By boiling the filtered solution with an excess of ammonium chloride, aluminium hydroxide is again precipitated.
Group IIIb includes the sulphides of cobalt and nickel (black), of zinc (white) and of manganese (usually buff coloured). The precipitate is digested with cold dilute hydrochloric acid and the solution filtered; the filtrate, which contains zinc and manganese, is boiled to expel sulphuretted hydrogen, and then hydrogen per oxide is added, followed by ammonium chloride and ammonia to precipitate manganese as a dark brown hydroxide. Zinc is de tected in the filtrate by reprecipitation as sulphide, after decom position of the peroxide by boiling. The presence of cobalt is readily ascertained in the black residue from the hydrochloric acid digestion by the borax bead test; the remainder of the sulphides is then dissolved in a little aqua regia, the solution diluted and rendered ammoniacal and an alcoholic solution of dimethylgly oxime added. Nickel is indicated by a yellow coloration, followed almost immediately by a red flocculent precipitate ; in the pres ence of cobalt the filtrate is brown, turning deeper and redder on addition of ammonium sulphide.
Group IV.—The precipitated carbonates are dissolved in a little dilute hydrochloric acid and the flame test applied. To a portion of the solution, calcium sulphate solution is added; if precipita tion is immediate, barium is indicated, but if delayed, strontium. In the absence of any precipitation, the remainder of the original solution is tested for calcium by addition of ammonia and am monium oxalate. If, however, barium is present, ammonium acetate is added, followed by potassium chromate to precipitate barium chromate ; strontium and calcium are again precipitated as carbonates and then dissolved in the minimum quantity of hydrochloric acid. Ammonium sulphate now serves to precipitate strontium, the filtrate from which is tested for calcium by means of ammonium oxalate as noted above.
Group V.—The basic radicals not precipitated by group re agents comprise magnesium, sodium, potassium and ammonium. The solution is evaporated to dryness and the residue heated to dull redness until cessation of fumes indicates that all ammonium salts have been expelled. The residue is dissolved in water and a portion tested for magnesium by the addition of ammonium chlo ride, ammonia and sodium phosphate. The presence of sodium and potassium is indicated by a flame test, but the introduction of sodium through the medium of the reagents must not be over looked; potassium may be confirmed by precipitation as platini chloride or cobaltinitrite, while ammonia is tested for by boiling a portion of the original substance with caustic soda solution.
Alternative schemes of separation exist, and there are also nu merous confirmatory tests for individual metals. It should be emphasized that any inference drawn from the result of the appli cation of a particular group reagent presupposes the absence of all metals belonging to previous groups. The system outlined above includes only the commoner bases, but each of the rarer metals falls into one or other of the groups and for the sake of completion, these are now given without further comment than that the presence of these metals of necessity complicates the identification scheme described above.
Group I.—Thallium (-ous).
Group II.—Gold, the platinum metals, germanium, vanadium, tungsten, molybdenum, selenium, tellurium.
Group IIIa.—Uranium, beryllium, zirconium, hafnium, gallium, indium, thorium, columbium, tantalum, scandium, cerium and the other "rare-earth" metals.
Group IIIb.—Thallium (-ic).
Group IV.—Radium.
Group V.—Caesium, rubidium, lithium.
Acid Radicals.—The detection of acid radicals does not lend itself to such a systematic serial procedure as that available for the bases. Rose noted the nature and colour of the precipitates obtained in neutral and acid solutions with barium chloride, silver nitrate, mercurous nitrate and calcium chloride on separate por tions of the solution. The method mostly in use, however, is that due to Bunsen, who divided the acid radicals into groups accord ing to the solubility of the silver and barium salts. It is generally desirable to carry out these tests after removal of the heavy metals by suitable treatment with sodium carbonate; moreover, the solution should not be too strong.
Group I.—Silver salts insoluble in water and in nitric acid; barium salts soluble in water :—Chloride, bromide, iodide, f erro and f erri-cyanide, thiocyanate, hypochlorite.
Group II.—Silver salts insoluble in water but soluble in nitric acid; barium salts soluble in water :—Sulphide, selenide, telluride, nitrite, acetate, cyanate, hypophosphite.
Group III.—Silver salts (a) white, (b) coloured, and soluble in nitric acid; barium salts insoluble in water but soluble in nitric acid :—(a) Sulphite, selenite, tellurite, phosphite, carbonate, iodate, borate, molybdate, selenate, tellurate, meta- and pyro phosphate; (b) phosphate, arsenate, arsenate, vanadate, thiosul phate, chromate, periodate.
Group IV.—Silver salts soluble, barium salts insoluble in nitric acid:—Sulphate, fluoride, fluosilicate (silicofluoride).
Group V.—Both silver and barium salts soluble :—Nitrate, chlo rate, perchlorate, manganate, permanganate.
Group VI.—Non-volatile acids forming soluble alkali salts only :—Silicate, tungstate, titanate, columbate, tantalate, zircon ate.
The acid radicals having been placed in their appropriate groups, special tests for the individuals are applied which cannot be de tailed here. It is interesting to note that certain pairs of salts or free acids are incompatible in solution, thus iodic and hydriodic acids react to form iodine, hypochlorite and sulphite give chloride and sulphate, hydrosulphuric and sulphurous acids liberate sul phur.
Filtration is effected by means (I) of folded paper discs of various texture specially prepared for the purpose ; (2) of a pad of asbestos supported in a special form of crucible such as that designed by Gooch; or (3) of a porous "sintered" glass diaphragm fused into glass tubing. Whichever means is adopted, it is essen tial that the filtering medium should be subjected to the same heat treatment before the preliminary weighing as it will be sub sequently (except when paper is to be incinerated).
Washing is generally necessary for the purpose of removing substances which are to be determined subsequently or which will interfere with the treatment of the precipitate, but at the same time it must not be carried too far. No substance is completely insoluble, and many of the compounds utilized in analytical work are appreciably soluble, even in the wash-liquors specially devised for use with them; it may even be necessary to apply correction for this solubility. In the case of heavy crystalline precipitates which settle readily, most of the washing can be carried out by decantation; precipitates which are light or bulky and gelatinous are transferred to the filter, the last traces being removed by me chanical means if necessary, and there washed by repeated treat ment with water or other liquor. The filter should be allowed to drain as completely as possible between washings.
The precipitate is prepared for weighing by drying in an oven maintained at a definite temperature or by more drastic heat treat ment. In the case of a precipitate on paper, ignition in a crucible may sometimes be carried out without any preliminary drying, sometimes after drying in a hot air-oven, while occasionally it is necessary to burn the paper apart from the precipitate before completing the ignition. All vessels after heating must be allowed to cool thoroughly in a desiccator before weighing; platinum may be weighed after about io minutes' cooling, porcelain after 20 min utes'. In all cases, the appropriate heat treatment should be re peated until no change in successive weighings is recorded. From the amount of weighed product, which is of known definite chem ical composition, the quantity of the constituent sought may be calculated. Blank tests should be made on the reagents, as these may introduce small quantities of the particular constituent which it is desired to determine.
Where readily applicable, electrolytic methods have much in their favour, especially their simplicity and cleanliness. Perhaps the most generally applied determination is that of copper, espe cially in its various alloys with zinc and tin ; the solutions elec trolyzed should be acid, though not too strongly so, with sulphuric or, more commonly, with nitric acid, this latter being used for the preliminary removal of tin. A rotating cathode of platinum gauze is the most convenient; the presence of a little hydrogen peroxide promotes the formation of a bright deposit, particularly from sulphuric acid solution. It is necessary to allow for bismuth which accompanies the copper. In solutions of the acidity re quired for deposition of copper, lead is completely deposited on the anode. If the quantity of lead is small, no great error is made in considering the deposit as the dioxide ; but should more than 2cg. of lead be present, the dioxide obtained may contain as little as 84% of lead, instead of the theoretical 86.6%. In such cases, or in any event if manganese also be present, the deposit should be dissolved from the anode, and converted into lead sulphate for weighing. Deposition of lead, however, from solutions containing a higher proportion of nitric acid than is permissible for copper, affords practically the correct dioxide.
Another useful electrolytic determination is that of zinc, espe cially in aluminium alloys of high zinc (12-14%) content. This is effected, after removal of copper, by electrolysis of the zincate solution, the aluminium being retained in solution by means of tartaric acid when the alkali is added. A gold rotating cathode is most convenient for this process; allowance must be made for traces of iron and manganese simultaneously deposited. The pres ence of chlorides, nitrates and ammonium salts is not permissible. In most other cases where electrolytic methods can be applied, preliminary separations by chemical means are necessary.
Calibrated glassware for volumetric work should not be ac cepted as correct unless checked or supplied with a reliable cer tificate of calibration; otherwise an accumulation of errors may result in a difference of several units per cent. The usual vessels are the flask, pipette and burette; the first of these when filled at 15° C to the graduation on the neck contains a specified vol ume, from which aliquot parts may be withdrawn by means of pipettes. Finally, the standard solution is, in general, run out from the burette, usually graduated in millilitres and tenths, until by appropriate signs visible to the eye or by electrical means, to which reference is made below, the reaction is found to be com plete. This procedure is termed "titration." The usual volumetric processes may be broadly divided into three classes: (a) acidimetric and alkalimetric; (b) oxidation and reduction ; (c) precipitation methods. The strength of standard solutions is commonly expressed in terms of normality ; a normal solution is one which contains in one litre that weight of the re agent which corresponds to 1.008g. of hydrogen or its equivalent, having regard to the particular reaction involved. This weight is either the molecular weight, expressed in grams, or some simple fraction thereof.
According to Arrhenius's Ionic Theory, salts do not exist in aqueous solution entirely as such but are split up more or less completely into their constituents or "ions" ; thus a solution of hydrochloric acid is regarded as containing hydrogen and chlorine ions H. and C1', while sodium hydroxide exists in solution as Na. and OH'. The chemical reaction between these two bodies, HC1+NaOH -- NaCl+H:O, is expressed ionically at H•+Cl'+ Na•+OH'—Na•+Cl'-}-H20 or, by removing the terms common to both sides The hydrogen ion and the hydrovyl ion are characteristic of acids and alkalis respectively and, so long as one or other of these predominates, the solution is said to be acid or alkaline in reaction. Indicators are very weak acids or weak bases whose colours in the ionized and in the non-ionized state dif fer ; constitutional alterations may also come into account. These colour changes occur within limited ranges of hydrogen-ion con centration, varying with the particular indicator, and it is upon this principle that the use of indicators in acidimetry and alkalimetry depends. Of those in common use, it may be said in general that phenolphthalein should be used for the titration of weak acids with strong bases, and methyl-orange for weak bases with strong acids. Litmus and methyl-red are both sensitive to carbon dioxide ; if these indicators are used when carbonates are present, it is neces sary to expel the carbon dioxide by boiling. Bicarbonates react neutral to phenolphthalein ; no indicator is suitable for the titra tion of weak bases with weak acids (see INDICATORS).
The numerous volumetric methods based on oxidation and re duction are often of considerable accuracy and ready application. Thus many metals form two (or more) series of salts differing in their degree of oxidation and which may be converted one to the other by means of appropriate oxidizing or reducing solutions of known strength. The oxidizing agents in most general use are potassium permanganate, iodine and potassium dichromate ; others employed for more special purposes are potassium bromate, iodate and ferricyanide.
Solutions for oxidations with potassium permanganate and di chromate must be sufficiently acid ; with the former the presence of sulphuric acid is preferable to that of hydrochloric acid, al though the reducing effect of this on the permanganate is prac tically negligible in very dilute cold solutions, especially in the presence of manganous sulphate. With permanganate, the end point of the titration is readily observed by the lasting pink colour imparted to the solution ; in the case of dichromate the end point may be noted by the use of an external indicator solu tion or preferably by using diphenylamine as an internal indicator. Iodine forms with starch a deep blue compound, which serves to indicate the presence of a very small quantity of the oxidizing agent ; this reagent is used in conjunction with sodium thiosul phate, less frequently with sodium arsenite.
Solutions of reducing agents do not maintain their strength so well as oxidizing agents, but they may be preserved fairly well in a suitable apparatus in which air is replaced by an inert gas such as carbon dioxide. Stannous and chromous chlorides are used and more especially titanium trichloride or sulphate. These last are very powerful reagents of extensive application. In the precipita tion methods of volumetric analysis, the standard reagent may he added until no further precipitation occurs, until a precipitate just appears, or until the end point is shown by means of an indicator, internal or external, with which a small excess of the reagent pro duces a characteristic change. The majority of the determinations in this category belong to the third type ; thus chloride is titrated with silver using potassium chromate or fluorescein as indicator, silver with thiocyanate in nitric acid solution using a ferric salt as indicator.
The end point in volumetric work may also be found in many cases by potentiometric methods. These depend upon the fact that when a metal or hydrogen is brought into contact with solu tions containing these same substances in the ionized state, a dif ference of potential is set up, the magnitude of which varies with the nature of the substances in contact and with the concentration of the ions in solution. As titration proceeds it generally happens that the potential difference, measured by suitable electrical means, undergoes a marked change at the end point, which may thus be determined. The method can be applied to acidimetric, reducing and oxidizing reactions, but in general holds little advantage over the older processes, the basic principle being the same in each case. The potentiometric method is, however, of especial value for the determination of hydrogen-ion concentration (q.v.), the impor tance of which in analytical as well as other work is becoming in creasingly recognized.
The most suitable methods for the estimation of any element will be found in the article devoted to that element.