The properties of chromogenic reagents used for the absorptiometric determination of aluminium in water are compared, and an experimental comparison of catechol violet, Eriochrome cyanine R and stilbazo has been made. Catechol violet is considered to be the most suitable and a method involving the use of this reagent has been developed. When using this method, the standard deviation of the analytical results varied from 0.004 to 0.008 mg 1-1 of aluminium for concentrations of 0.05 and 0-3 mg 1-1 of aluminium, respectively. With the exception of fluoride, substances normally present in treated waters did not cause important interference. The effect of fluoride (1 mg 1-1 or less) is tolerable for most purposes, but appreciably greater concentrations require preliminary removal of fluoride or correction for its effect. Satisfactory results have been obtained with the method in six other laboratories. The method is simple and rapid ; ten samples can be analysed in approximately l & hours, The method has advantages over other commonly used methods, and is recommended for use in water analysis laboratories.THE concentration of aluminium in potable water is often an important factor in assessing both the quality of the water and also the performance of the associated treatment plant. Concentrations of less than 0.1 mg 1-1 of aluminium are the common aim, and we required an analytical method capable of the following performance : Concentration range . . . . 0 to 0.3 nig 1-1 of aluminium Standard deviation >0.005 mg 1-1 of aluminium or 5 per cent. of the aluminium Bias .. .. .. . . > O -O l mg 1-1 of aluminium or 10 per cent. of the aluminium . . . . concentration (whichever is the greater) concentration (whichever is the greater)Several analytical techniques have been used for determining such aluminium concentrations, but we restricted our attention to absorptiometric methods (measuring in the visible region) because the necessary instrumentation is usually available in laboratories concerned with the analysis of potable water.When this work was started, aluminon was the most commonly used chromogenic reagent in this country. A number of standard methods using this reagent1-3 are based largely on the work of Packham,* who showed that aluminon has the following disadvantages:(i) the coloured product is a lake, and thus requires very careful control of all experimental conditions as well as the addition of a colloid stabiliser [gum arabic or poly( vinyl alcohol)] ;(ii) in order to achieve proper colour development, the sample must be heated, which is inconvenient and increases the analytical time;(iii) the calibration graph is not linear;(iv) close control of the temperature of the processed sample is desirable, the optical (v) fluoride causes very large negative errors; for example, 1.0 mg 1-1 of fluoride causes (vi) different batches of aluminon may give markedly different calibration graphs. density changing by approximately 0.6 per cent. "C-l at 20 "C; an error of -0.07 mg 1-1 of aluminium when 0-1 mg 1-1 of aluminium...

No abstract

A method, together with a modilication for obtaining high sensitivity, is described for determining the "reactive" silicon content of water; it is based on the absorptiometric measurement of solutions of reduced 8-molybdosilicic acid. The within-batch coefficient of variation of the optical-density difference, optical density of sample less optical density of reagent blank solution, varied from 7 per cent. to 0-2 per cent. for concentrations of 0.01 and 0.5 p.p.m. of silica, respectively. The limit of detection was about 0.001 p.p.m. of silica. The effects caused by several other substances have been determined. The analysis time is about 14 hours for a batch of 10 samples in duplicate.* "Reactive" silicon is defined in this paper as those forms of silicon-mainly monomeric and dimeric silicic acid-that react with ammonium molybclate in 10 minutes under the conditions of the method given in this paper.

The effect of experimental conditions on the formation, stability (in the presence of reagents used for destroying molybdophosphoric acid) and reduction of a-and p-molybdosilicic acids -has been determined. The effectiveness of several reagents for preventing inierference from phosphate has also been investigated. Both cr-molybdosilicic: acid reduced by stannous tin and p-molybdosilicic acid reduced by 1 -amino-2-naphthol-4-sulphonic acid, should be suitable for precise methods of determining "reactive" silicon in water.AT the steam pressures and temperatures used in modern power stations, silicic acid is appreciably soluble in steam, The concentration of silicon in boiler and make-up water is therefore controlled to ensure that the steam contains not more than about 0.02 p.p.m. of silica; above about 0.02 to 0.04 p.p.m. of silica undesirable deposits may be formed on turbine blades.Methods are therefore needed for deteirmining both the "reactive"* and total silicon concentrations in a range of aqueous samples from steamwater circuits in power stations.Thus the methods must be suitable for conde:nsate, feed-water and steam (0-002 to 0-05 p.p.m. of silica), make-up water (0.005 to 0-2 p.p.rrt. of silica) and boiler water (0.2 to 75 p.p.m. of silica).The ratio of phosphate (p.p.m.) to silica (p.p.m.) is usually between 0 and 20, but occasionally ratios as high as 50 may occur.The methods must also be precise, so that they can be used for determining small amounts of "non-reactive" silicon as the difference between the total and "reactive" silicon contents.Absorptiometric methods in which the reduced aand P-molybdosilicic acids are used appear to be the only sufficiently sensitive means of determining extremely small concentrations of silicic acid. They can also be used for determining total silicon provided all forms of silicon are converted to silicate or monomeric silicic acid in an earlier step.This paper reports an investigation of the molybdosilicic acids to determine suitable conditions for each method; Parts 112 and 1113 give details of the methods and the results obtained.Two forms of molybdosilicic acid, a-and p-, are of analytical interest and were first distinguished by S t r i ~k l a n d . ~Each form can be reduced to several different blue compounds with various reducing agents, and Strickland4 has determined the properties of several forms and the conditions for their formation. The relative amounts of ci-and P-molybdosilicic acids produced and their stabilities appear to be affected by factors such as the concentrations of acids, molybdate, neutral salts and reagents such as oxalic or tartaric acid used for destroying molybdophosphoric acid. However, the literature does not give quantitative values for all these effects, and they have, therefore, been investigated.Boiler waters often also contain added phosphate. EXPERIMENTAL APPARATUS, REAGENTS AND TECHNIQUE-All optical-density measurements were made in 4-cm cuvettes with a Hilger Cvispek spectrophotometer against distilled water in the reference cu...

The method described in Part 11 for determining "reactive" silicon has been shown to be suitable for analysing boiler waters containing phosphate and alkali. This paper describes the tests performed to confirm the suitability of the method in the presence of up to 5.0 mg of phosphate. Although the magnitude of interference from phosphate is small for silicon concentrations exceeding 0.1 p.p.m. of silica, the nature of the effect was found to be complex, and has been studied in detail. The effects of many other substances have also been determined. Coefficients of variation of about 0.4 per cent. have been obtained in the analysis of boiler waters.PART I11 describes a method for determining "reactive" silicon in relatively pure waters such as de-ionised water, feed-water, make-up water and steam. The method was expected to be suitable for analysing boiler waters that may contain appreciable concentrations of phosphate and alkali. However, the effect of these two substances had not been tested in sufficient detail for a definite recommendation to be made. Such tests have now been performed, and the results presented in this paper confirm the suitability of the method for analysing boiler waters.Two variants of the method are described in Part 1I.l In the first, portions of samples are analysed in 100-ml calibrated flasks; in the second (the modification for higher sensitivity), 100-ml portions are analysed in polythene bottles. In view of the alkalinity of many boiler waters, the latter technique was adopted as described below under "Method." The method given in this paper is suitable for all types of power-station waters but the tolerances given for times and reagent concentrations are smaller than those in Part I1.l This is to ensure that the effects caused by phosphate are small. METHOD APPARATUS-Polythene bottles, 4-OY 8-oz capacity-The analyses are carried out in polythene bottles, and the magnitude of the interference caused by phosphate may depend on the particular bottles that are used (see under (iii), p. 637). Therefore, before a set of bottles is used for analysis, it is prudent first to check that the effect of phosphate is adequately small. When the bottles have been shown to be satisfactory they should be reserved for this analysis only. REAGENTS-The water, tartaric acid solution, reducing-agent solution and standard solutions of silicon should be prepared as described in Part 1I.l AcidiJied molybdate solution-Dissolve 89 g of ammonium molybdate, (NH,) ,Mo70,,.4H2O, in about 800ml of water at room temperature. Add 63ml of 98 per cent. sulphuric acid cautiously to 100 ml of water, with stirring, and allow the mixture to cool. Add the acid to the molybdate solution, cool it to room temperature, and dilute it to 1 litre with water.Both the acidified molybdate solution and the tartaric acid solution have been found to be adequately stable for at least 6 months.* For details of earlier parts of this series, see reference list, p. 641.f "Reactive" silicon is defined in this paper as those forms of silicon, ...