ISSN:0003-2654    

DOI:10.1039/AN98409FX001

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST JANUARY 1984 VOL. 109 3 Results of an Inter-laboratory Analytical Quality Control Programme for Non-saline Waters The Severn Estuary Chemists’ Sub-Committee* The Severn-Trent South West Welsh and Wessex Water Authorities have a statutory responsibility for the quality of water in the Severn Estuary. A Joint Committee was formed to coordinate the activities of the Authorities in respect of the management of the water quality with special reference to the monitoring of polluting discharges entering the estuary. In order to carry out this duty effectively comparability of analytical results between the laboratories of the four Authorities was considered to be essential and a Chemists‘ Sub-committee was established to undertake this task. The Chemists‘ Sub-committee was required to establish within a period of 3 years and then to maintain an inter-laboratory analytical quality control programme for a total of 17 determinands.It was found that even for relatively simple determinands such as ammonia total oxidised nitrogen and chloride inter-laboratory comparability was not easy to achieve and that once achieved constant attention was necessary in order to maintain it. Hence the work required was much greater than at first envisaged and even so a perfect set of results was not possible for all determinands. This paper summarises the approach adopted illustrates some of the problems experienced and presents results of the inter-laboratory comparability finally achieved. Keywords Analytical quality control; inter-laboratory comparability; collaborative studies; river water analysis; sewage and trade effluent analysis The four Water Authorities that form the boundary of the Severn Estuary (Severn-Trent South West Welsh and Wessex) have statutory responsibilities regarding the quality of water in the estuary.There has been particular concern for many years about the pollution load entering the estuary via direct industrial discharges sewage effluents and rivers (these are termed “inputs”). In order to quantify these inputs with confidence it was necessary to establish comparability of analytical results between the various laboratories of the four authorities. A Chemists’ Sub-committee (CSC) (see Appen-dix) was established with the specific objective of achieving comparability of analytical results between all laboratories involved in the analysis of non-saline inputs to the estuary (the “Inputs Programme”).The CSC agreed that simply circulating samples to the participating laboratories for analysis by their current analyti-cal methods was unlikely to lead to comparability between the laboratories. Such exercises which have been widely used in other collaborative programmes1J and are commonly referred to as “round-robin” exercises have the disadvantage that they will provide information on the comparability achieved in a particular exercise only as measured by the over-all standard deviation. The round-robin exercise does not set objectives for precision and bias required nor will it readily enable the continuous performance of an individual laboratory to be assessed objectively nor enable the many possible sources of error in analytical methods to be identified and corrected.Therefore it was decided that a more systematic approach to comparability than round-robin exercises was required, ideally an approach that would enable the performance of each laboratory to be monitored objectively by comparison of results obtained with pre-set targets of maximum desirable total error (or accuracy). About the time of inception of the CSC such an approach had been proposed by the Water Research Centre (WRC) for achieving comparability within the group of laboratories participating in the Department of the Environment’s Harmonised Monitoring (HM) Scheme .374 The CSC felt that although this approach met its basic requirements a major disadvantage when applied to the * Correspondence regarding this paper should be addressed to: Dr.K. C. Wheatstone Severn-Trent Water Authority Lower Severn Division 141 Church Street Malvern Worcestershire WR14 2AN, UK. Severn Estuary Inputs Programme was that an exercise was held for each determinand in turn and consisted of several separate stages i.e. precision testing stage comparison of standards stage and bias testing stage each stage in an exercise being carried out sequentially on the successful completion of the previous stage. Consequently progress in achieving comparability between a group of laboratories although thorough was slow. Such a rate of progress was unacceptable in the context of the Inputs Programme for which inter-laboratory comparability was required for 17 determinands within 3 years.A compromise approach was adopted that combined the speed and simplicity of a round-robin exercise with many of the advantages of the HM Scheme approach. This paper provides a summary of the approach to comparability testing employed the problems experienced by individual laboratories in meeting the objectives and the level of inter-laboratory comparability finally achieved. A total of almost 40 exercises was necessary to achieve the comparability required; this paper provides only the results obtained in the final stages for each determinand. Full details of all exercises and associated work may be obtained on request. The Approach to Comparability Suite of Determinands The suite of determinands for which comparability between the participating laboratories was required for the Inputs Programme was as follows ammoniacal nitrogen total oxidised nitrogen silicate orthophosphate total phosphorus, chloride cadmium chromium copper iron lead mangan-ese mercury nickel zinc biochemical oxygen demand and organochlorine pesticides.Tolerable Error of Analytical Results After due consideration the tolerable errors of analytical results for the Inputs Programme were set as follows and this also enabled the results of exercises to be assessed objec-tively. In deciding the targets it was assumed that the total error of an analytical result is divided equally between systematic and random errors. This is an arbitrary but convenient sub-division of the total error that has been found to be acceptable in practice.4-4 ANALYST.JANUARY 1984. VOL. 109 ( a ) For metals and total phosphorus a maximum systematic error (or bias) for an individual analytical result of +20% of the concentration or half of the target limit of detection, whichever is the greater and a precision (as measured by the standard deviation) of +lo% of the concentration or one quarter of the limit of detection whichever is the greater. Target limits of detection are as follows: Determinand 1 Cadmium . . . . . . Copper . . . . . . . . Chromium . . . . . . Lead . . . . . . . . Nickel . . . . . . . . Zinc . . . . . . . . Iron . . . . . . . . Manganese . . . . . . Mercury . . . . . . Totalphosphorus . . .. Sewage and trade effluents 1.0pg1-1 10.0 pg 1-1 10.0 pg 1-1 10.0 pg I-’ 10.0 pg I-’ 10.0 pg 1- I 10.0 pg 1-1 10.0ug1-’ NA 0.1 mg 1-1 River water 0.2 pg I-’ 2.0 pg I-’ 2.0pgl-’ 2.0pgl-’ 2.0 pg I - ’ 0.1 1.181-1 NA NA NA 0.1 mg I-’ (b) For biochemical oxygen demand a maximum systematic error of k1.0mgl-1 and a precision of k0.5mgl-1 both relating to the sample whether diluted or not actually measured in the bottle. ( c ) For all other determinands the targets were k 10% of the concentration for maximum systematic error and +5’/0 of the concentration for precision. These tolerable errors may seem unduly large but experience has shown that for routine analysis they are realistic. Ideas that the total error of an analytical result is small (for example 5-10%) have not been confirmed in practice either in this or in other studies.4-7 Indeed McFarren et al.,7 when using results of inter-laboratory studies in order to categorise analytical methods objectively suggested that methods be judged excellent when the total error was f25% or less acceptable when the total error was +50% or less and unacceptable only when that total error was greater than The CSC had been given the requirement that the absolute minimum of analytical results should be reported as “less than” values for the purposes of the Inputs Programme. Therefore it was necessary to develop analytical methods and procedures that met the required percentage targets for all concentrations of determinands likely to be found in the “inputs” to the estuary.Consequently the methods employed by the CSC have limits of detection sufficiently small to allow the accuracy of an individual result to meet the targets regardless of the concentration. A good example to illustrate the reasoning behind this approach is provided by the results for lead in the River Severn at Hawbridge by Severn-Trent Water Authority’s routine water quality monitoring programme.8 For 1977-78, the mean concentration of lead was reported at “less than 0.04 mg 1-1” based on 25 samples (0.04 mg-1 being the limit of detection of the laboratory’s normal analytical method). At the same sampling point the mean flow over the same period was 9370 M1 d-1. From this it can be calculated that the mean mass load of lead in the River Severn at that point was “less than 374.8 kg d-1.” This is a statement that is unacceptably vague as the actual amount could be any value between 0 and 374.8 kg d-1.The analytical methods adopted for lead for the Inputs Programme consequently had limits of detection much smaller than the laboratory’s normal method which enabled real values to be reported for this determinand (and has in fact shown that the mean mass load of lead in the river at this point is about 40 kg d-1). +5O%. Analytical Quality Control Charts The first step in the CSC’s approach to comparability was to set up where not already in routine use analytical quality control charts in all laboratories for the determinands re-quired. By the use of such charts all laboratories were able to check routinely the precision of their analytical results.Comparability Exercise Format The comparability exercise format adopted by the CSC enabled all the stages i.e. precision testing comparison of standard solutions bias testing and where appropriate, spiking recovery to be carried out simultaneously and for several determinands at once rather than sequentially and individually as in the HM scheme. By taking the number of replicate batches of analyses as five it was possible for an inter-laboratory comparability exercise to be completed for a group of determinands within one working week. This convenience was offset by a lower degree of confidence in the results than if a larger number of replicates had been carried out. This was a calculated decision bearing in mind the conditions imposed by the time-scale required for the pro-gramme.The principles of the calculations employed were those used in the HM scheme.4.5 The “standard format” for comparability exercises was agreed to be the analysis by each laboratory of the following samples and solutions in duplicate and in random order on each of five days: a blank solution; the laboratory’s analytical quality control standard solu-tion; a standard synthetic solution of similar concentration to a standard synthetic solution of low concentration; a river sample; and an input other than a river-this varied and included sewage final effluent crude sewage and industrial efflu-ent. (b) ; Solutions (c)-(f) were prepared and circulated by the organis-ing laboratory.This format was followed for the “nutrient” and C1 determinands required for the Inputs Programme although for metals the format was varied slightly (see later). Check Exercises Once comparability had been achieved for a particular determinand or group of determinands check samples were circulated from time to time to all laboratories to ensure that this situation continued for the duration of the Inputs Programme. A river and another “input” sample were supplied for each such exercise and as these exercises were only checks on continuing comparability the format was reduced to duplicate analyses only for each determinand. These check exercises are still ongoing at approximately 3-monthly intervals. A summary of the application of the above approach to comparability testing is given below for all the determinands of interest in the Inputs Programme.Results and Discussion Determination of Metals in Sewage and Trade Effluents The first inter-laboratory comparability exercise for the metals of interest (Cd Cr Cu Fe Pb Mn Ni and Zn) involved the analysis of standard solutions samples and “spiked” samples by direct aspiration of the acidified samples into the atomic-absorption spectrophotometer. The results clearly demonstrated that such a method was unsatisfactory for most types of sample encountered being applicable only to concentrations of individual metals greater than about 500 pg 1-1 ANALYST. JANUARY 1984 VOL. 109 A more appropriate method was sought by laboratories testing various methods individually and which eventually culminated in an inter-laboratory comparability exercise objectively to compare and assess the candidate methods.The methods included concentration by evaporation solvent extraction and ion exchange prior to aspiration into the AAS instrument. From this exercise it was concluded that concen-tration by evaporation was the simplest and least tedious of the three methods and this was selected for further study. Early problems were experienced by some laboratories with the concentration by evaporation method owing to the lack of availability of a background absorption correction facility. Some laboratories also experienced problems with contamina-tion of the acid used for digestion while others found that their AAS instruments were in obvious need of servicing or even replacement.Each of these problems was identified and overcome in turn sometimes by the laboratory working alone and sometimes in collaboration with other laboratories. At every stage small exercises were held to monitor the progress of individual laboratories. The final inter-laboratory comparability exercise involved the analysis of all eight metals of interest in a circulated standard solution and a sewage effluent sample once each day for 4 days with all six laboratories using the standardised evaporation method. This method consisted of a 20-fold evaporation of the sample in the presence of nitric acid taking care to evaporate to fuming and not to allow the sample to boil dry. Correction for non-atomic (background) absorption was found to be essential for cadmium lead nickel and zinc.5 The results obtained from this exercise are given in Tables 1-4 from which it can be seen that most of the results were within the required targets. Where the targets were occasion-ally exceeded it was either only marginally or else it was due to an unacceptably large bias which in some instances was so large that the result was statistically rejected. These large outlying results were thought to be due to contamination of the sample during analysis and served to emphasise the care that needs to be taken in these determinations. For example, the high results obtained by laboratory 4 for cadmium and zinc for the sewage effluent sample were considered to be due to contamination.The results obtained for lead require special comment. Although all but one of the laboratories achieved acceptable results for the circulated standard the results for the sewage effluent fell into two distinct groups of concentration a factor of about two apart. The reasons for this are thought to be (a) contamination for laboratory 4 and possibly laboratory 1 and (b) omitting to use background absorption correction for laboratory 2. The lead concentration in the sewage effluent sample was checked by two independent laboratories using the same method who reported mean concentrations of 12.0 and 12.9 pg 1-1. This tends to confirm that the high results reported by laboratories 1 2 and 4 were erroneous. The low results for laboratory 3 for both standard and sample were caused by an incorrect laboratory standard.The results reported for chromium and iron for both the standard solution and the sewage effluent gave cause for concern. The range of results was large in all instances and no Table 1. Results for cadmium and chromium; Cadmium Chromium Circulated standard Circulated standard (true concn. 6.0 pg 1-1) Sewage effluent (true concn. 9.2 pg 1 - I ) Sewage effluent Mean Std. Rel. Max. Mean Std. Rel. Max. Mean Std. Rel. Max. Mean Std. Rel. Max. concn./ dev./ std. possible concn./ dev./ std. possible concn./ dev./ std. possible concn./ dev./ std. possible pg p! dev. bias dev. bias, 1 6.2 0.08 1.3 +4.9 0.9 0.05 5.6 ++ 8.3 0.29 3.5 -13.5 1.9 0.48 + ++ 2 5.7 0.04 0.7 -5.8 0.9 0.08 8.9 ++ 6.3 1.44 + ++ 2.1 0.63 + ++ 4.8 0.26 5.4 ++ 3 6.0 0.64 10.7* k12.5 0.7 0.15 + ++ 8.1 0.94 + ++ 4 5.4 0.56 10.4* -21.0** 4.8R 0.86 17.9** 9.8 0.96 9.8 +18.8 0.8 0.00 0.0 ++ 5 5.3 - - - 0.9 0.14 + ++ 4.5 - - - 2.0 0.50 + ++ 9.2 3.03* + ++ 0.5 0.33 + ++ 6 5.6 0.31 5.5 -12.7 1.0 0.08 8.0 ++ p! :! yo O/O tory 1 - 1 I-' % '% l5 1- y" Yo 1-1 I - Yo Y" 1-Labora- pg pg dev.bias p! dev bias, Mean 5.7 0.9 7.7 2.0 t Symbols as follows * result not significantly different from target (at 95% confidence level); * * result outside target; - single result only reported; + standard deviation within concentration target; + + maximum possible bias within concentration target; R statistically rejected value; ND not detected. Table 2. Results for copper and iron? Copper Iron Circulated standard (true concn.35.0 pgl-1) Sewage effluent Circulated standard (true concn. 466 pg 1- 1) Mean Std. Rel. Max. concn.1 dev.1 std. possible Labora- pg pg dev. bias, tory 1-1 1-1 ?Lo O/O 1 32.9 0.91 2.8 -9.1 2 35.8 2.7 7.5 +11.4 3 33.8 2.7 8.0 -12.5 4 32.1 0.65 2.0 -10.5 5 35.0 - -6 34.1 1.9 5.6 -9.0 -Mean 34.0 t For symbols see footnote to Table 1. Mean concn./ 1-1 29.2 31.9 30.2 26.1 25.0 28.9 Std. Rel. dev.1 std. I-' Yo 0.48 1.6 8.3 26.0** 0.46 1.5 0.82 3.1 1.6 6.4 1.5 5.2 28.6 pg dev., Max. possible bias, O/O +4.1 +45.7** +7.5 - 12.1 -19.2 +7.2 Mean concn.1 1-1 469 514 364 514 428 449 Std. Rel. dev.1 std. pg dev., I-' O/O 3.4 0.7 36.2 7.0 31.4 8.6 67.9 13.2* 16.1 3.6 456 - -Max.possible bias, O/O +1.5 +19.4 +27.4** -7.7 -29.8** -Sewage effluent Mean Std. Rel. Max. concn.1 dev./ std. possible I.18 dev. bias, 1-1 lY yo Yo 247 10.3 4.2 -14.3 244 3.3 1.4 -12.4 206 10.3 5.0 -29.2** 250 5.4 2.2 -11.1 357 37.4 10.5" +46.4** 342 34.7 10.2* +39.7** 27 6 ANALYST JANUARY 1984 VOL. 109 Table 3. Results for lead and manganese? Lead Manganese Circulated standard (true concn. 40 pg 1-1) Mean Std. Rel. Max. concn.1 dev.1 std. possible Labora- pg pg dev. bias, tory 1-1 1-1 Yo YO 1 40.4 2.78 6.9 +9.2 2 37.6 2.29 6.1 -12.7 3 35.0 3.1 8.9 -21.6** 4 38.9 4.53 11.6* -16.1 6 36.4 1.67 4.6 -13.9 - 5 39.6 - -Mean 38.0 t For symbols see footnote to Table 1.Sewage effluent Circulated standard (true concn. 42.6 pg 1-1) Sewage effluent Mean Std. Rel. concn.1 dev.1 std. 1-1 1-1 Yo 23.9R 1.11 4.6 38.3R 2.40 6.3 11.1 0.59 5.3 25.OR 3.2 12.8* 13.9 3.0* + 13.7 0.50 3.6 P8 Pg dev.9 12.9 Max. possible bias, Y O R R - 19.3 ++ +10.8 R Mean concn.1 1- p5 42.4 40.6 37.1 39.9 44.8 42.6 Std. Rel. dev.1 std. 1-1 Yo 1.75 4.1 1.89 4.6 3.28 8.8 1.14 2.9 0.71 1.7 41.2 Pg dev., - -Max. possible bias, YO -5.3 -9.9 -22.0** -9.5 k2.0 -Mean Std. Rel. concn.1 dev.1 std. pg dev. r 1-1 yo 39.0 0.41 1.0 30.1 4.50 15.0* 31.8 0.49 1.5 36.5 1.11 3.0 28.9 1.83 6.3 31.7 0.34 1.1 33.0 Max. possible bias, YO +19.6 -24.8** -5.4 +14.6 - 19.0 -5.2 ~ ~~ Table 4.Results for nickel and zinct Nickel Zinc Circulated standard (true concn. 8.8 pg 1-1) Sewage effluent Circulated standard (true concn. 92.6 pg 1-1) Sewage effluent Mean Std. Rel. Max. concn.1 dev. std. possible Labora- pg p dev. bias, tory 1-1 I- Yo O/O 2 6.8 0.29 4.3 ++ 3 6.4 0.42 6.6 ++ 4 9.7 0.75 7.7 +20.2** 1 8.1 0.75 9.3 -18.0 - 5 8.2 - -6 8.2 0.51 6.2 -13.6 Mean 7.9 t For symbols see footnote to Table 1. Mean Std. Rel. concn.1 dev.1 std. 1-1 1- Yo 5.3 0.29 5.5 5.3 0.65 + 3.8 0.13 3.4 5.5 1.27 + 5.0 0.96 + 4.4 0.85 + pg p dev.9 4.9 Max. possible bias, Y O +15.1 ++ ++ ++ ++ ++ Mean Std. Rel. concn.1 dev.1 std. P! dev. : 1- o/o 94.9 1.32 1.4 97.9 1.70 1.7 87.2 7.36 8.4 102.3 7.24 7.1 94.7 2.22 2.3 87.2 - -94.0 Max.possible bias, YO +4.2 +7.9 -15.2 +19.7 +5.1 -Mean concn.1 vg I-' 67.3 51.5 54.3 144.3R 49.4 58.0 Std. Rel. dev.1 std. I-' Yo 0.58 0.9 4.71 9.1 1.00 1.9 11.5 8.0 2.40 4.9 1.32 2.3 56.1 pg dev.7 Max. possible bias, Y O +21.2** -18.1 -5.3 - 17.0 +6.2 R common factor or reason could be found to which these differences could be attributed. This situation was clearly far from satisfactory and consequently one laboratory undertook a detailed investigation of the cause of the differences. The investigation showed that both chromium and iron could be subject to anomalous calibration graphs under certain flame conditions in the air-acetylene flame and that this was the probable cause of the differences.Details of the investigations and recommendations for overcoming the problem have been published.9Jo Determination of Metals in River Water For the determination of metals in sewage and trade effluents the evaporation and flame AAS method described above was considered satisfactory. However for the accurate determina-tion of cadmium lead copper nickel and zinc in river samples the target limits of detection obtainable by the evaporation method needed to be improved by a factor of about five. It was felt that only two analytical techniques were capable of meeting the limits of detection required namely concentration of the metals by means of a chelating resin followed by AAS and AAS with electrothermal atomisation.The chelating resin Chelex 100 had been widely used for the determination of low level metals in sea water,11-13 but for river waters the technique could not be directly applied because of the greater proportion of complexed and/or organically bound metals that would not be adsorbed into the resin. Consequently low results would be expected from samples of river waters. A method was eventually developed that overcame this problem involving treating the sample with potassium persulphate and nitric acid in order to break down the organic matter in the samples and to release the organically bound metals for analysis. The pH of the solution was adjusted to 6.3-6.7 and passed through a small column of Chelex 100. After rinsing to remove residual salts the chelated metals were removed from the column with a small volume of dilute nitric acid prior to analysis by flame AAS.Determination of low concentrations of metals by AAS with electrothermal atomisation presented difficulties for routine use owing to dissolved salts in natural samples which caused large suppressive interference effects (up to 80% suppression of the signal) particularly for lead. A method was developed whereby this interference was overcome by means of the addition of small amounts of lanthanum salts to the samples. Details of this method have been published.14-16 During the joint development of these two analytical techniques it was mutually beneficial to hold a series of small comparability exercises in order to monitor the progress and to check their accuracy.When the development work was completed it was decided that owing to the special precautions needed in analysing river samples the techniques were by no means suitable for routine use in all laboratories. Conse-quently only selected laboratories were equipped to carry out this analysis. The CSC also made available to another group involved in similar work the Humber Estuary ad hoc Group for Analy-tical methods and Inter-laboratory Testing details of the techniques developed. In-house precision testing was carried out by each of the laboratories together with coordinated tests on circulated river samples and standard solutions. A joint comparability exercise was held that involved fiv ANALYST JANUARY 1984 VOL.109 7 replicate determinations on the same day of a standard solution river water and river water spiked with the metals of interest (Lee Cd Cu Pb Ni and Zn). The river water was filtered through a 0.45-pm membrane filter into a large polythene container and the metals were preserved by the addition of 1 ml of concentrated nitric acid (low-in-metals grade) per litre of filtered sample. A bulk standard synthetic solution was also prepared and similarly preserved. The preserved river sample was divided into two, and to one aliquot was added a spike of a concentrated solution containing known amounts of the metals of interest. The river and spiked river samples and the synthetic solution were sub-sampled into small polyethylene bottles for distribu-tion to the participating laboratories.Two laboratories (numbers 3 and 5 ) used the Chelex 100 method the others used the electrothermal atomisation method. The results obtained are given in Tables 5-9 from which it can be seen that except for one precision result (laboratory 5 for zinc) and one bias result (laboratory 2 for lead) all laboratories met the targets for precision and bias on the circulated standard sample and spiked sample for all five metals. As these two laboratories were only marginally outside the target for only one sample each and all their other results were within the targets no further action was considered necessary. For the type of sample and levels of concentration of metals involved the results for the inter-laboratory exercise were considered to be very satisfactory overall.Mercury An initial survey of methods in use for mercury determina-tions indicated that all laboratories used the cold vapour electrothermal AAS measurement technique after reduction of mercury to the elemental form and the stripping of it from solution. A variety of reductants and measurement cell systems were in use and the first exercise was therefore designed to test the sensitivity of the measurement stage and the comparability achievable with inorganic mercury stan-dards. The results of this preliminary exercise demonstrated that comparability could be achieved for inorganic standards and indicated that the target detection limit was achieved by all laboratories. Various published methods for sample preservation and for the conversion of organically bound mercury to the inorganic form were tried and tested and a suitable procedure was adopted.Briefly this consists of taking about 100 ml of sample into a glass-stoppered borosilicate-glass bottle and adding Table 5. Results for cadmium Standard solution (true concn. 5.5 pg 1-1) River sample Mean Std. Rel. concn.1 dev.1 std. 1 5.4 0.14 2.6 2 5.5 0.11 2.0 3 5.4 0.29 5.4 4 5.6 0.44 7.9 5 6.2 0.00 0.0 Laboratory pg 1-1 pg 1-1 dev. ,% Mean 5.6 Max. possible bias YO -4.2 k1.9 -6.8 +9.4 + 12.7 Mean concn.1 1-1.2 1.2 1.2 1.4 1.5 Std. Rel. dev./ std. pgl-1 dev. YO 0.09 7.5 0.11 9.2 0.17 14.2* 0.12 8.6 0.05 3.3 1.3 * Result not significantly different from target (at 95% confidence level).Max. possible bias YO -14.3 - 15.7 -20.0 + 16.5 + 19.0 River sample + 5.5 pg 1-Mean concn.1 8.1 6.7 6.6 7.6 7.4 vg 1-Std. Rel. dev.1 std. pg 1-1 dev. YO 0.32 4.0 0.15 2.2 0.20 3.0 0.57 7.4 0.16 2.2 7.3 Max. possible bias YO +15.1 - 10.2 - 12.2 +11.5 +3.5 Table 6. Results for copper Standard solution (true concn. 24.0 pg 1-l) River sample River sample + 24.0 pg 1-1 Laboratory 1 2 3 4 5 Mean Mean Std. Rel. concn.1 dev.1 std. pg 1-1 pg 1-1 dev. '/O 21.2 1.41 6.7 21.5 0.36 1.7 24.2 1.21 5 .O 26.1 0.55 2.1 26.5 0.72 2.7 23.9 Max. possible bias YO -17.5 -12.1 +5.8 +11.3 +13.3 Mean Std. Rel. concn.1 dev.1 std. pgl-1 pgl-1 dev. YO 9.9 0.56 5.7 11.8 0.51 4.3 9.3 0.70 7.5 11.2 1.11 9.9 10.0 0.40 4.0 10.4 Max.possible bias YO -9.9 +18.1 - 17.0 + 17.8 -7.5 Mean concn .I 31.5 37.2 33.1 37.2 35.9 Yg 1-Std. Rel. dev.1 std. pg 1-1 dev. YO 1.35 4.3 0.40 1.1 1.09 3.3 0.60 1.6 1.35 3.8 35.0 Max. possible bias YO -13.1 +7.4 -8.4 +7.9 +6.2 Table 7. Results for leadt Standard solution (true concn. 18.0 pg 1-1) River sample River sample +18.0 pg 1-l Mean Std. Rel. concn.1 dev.1 std. pg dev., Laboratory 1-1 1-1 Y O 1 18.3 0.65 3.6 2 18.4 0.38 2.1 3 17.7 0.49 2.8 4 18.6 1.12 6.0 5 17.5 1.06 6.1 Mean 18.1 t For symbols see footnote to Table 1. Max. possible bias, Y O +5.6 +4.4 -4.3 +9.4 -8.9 Mean Std. Rel. concn.1 dev.1 std.1-1 I-' YO 1.6 0.38 + 1.8 0.36 + 1.9 0.42 + 1.0 0.00 0.0 0.4 0.00 0.0 pg dev.7 1.3 Max. possible bias, Yo ++ ++ ++ ++ ++ Mean concn./ 1-1 17.4 22.3 19.0 19.1 16.7 Std. Rel. dev.1 std. 1-1 Y O 0.27 1.6 0.49 2.2 0.54 2.8 0.74 3.9 1.26 7.5 18.9 pg dev., Max. possible bias, Y O -9.3 +20.4** +3.2 +4.8 - 18. 8 ANALYST JANUARY 1984 VOL. 109 Table 8. Results for nickel Standard solution (true concn. 38.0 pg I - ' ) River sample River sample + 38.0 pg 1 - I Mean Std. Rel. concn.1 dev.1 std. ELg pg dev.7 Laboratory 1-1 1-1 YO 1 35.7 1.52 4.3 2 34.4 1.08 3.1 3 33.7 1.98 5.9 4 39.4 1.52 3.9 5 42.4 1.23 2.9 Mean 37.1 Max. possible bias, O/O - 10.0 - 12.6 - 16.6 +7.6 + 14.7 Mean concn .I pg I-' 20.6 20.3 21.1 20.1 23.6 Std.Rel. dev.1 std. pg dev., I - ' Y O 1.52 7.4 0.76 3.7 0.67 3.2 1.78 8.9 1.48 6.3 21.1 Max. possible bias, YO -9.2 -7.2 k3.2 - 12.8 + 18.5 Mean Std. Rel. concn.1 dev.1 std. pg dev., I-' I-' Y O 53.3 0.40 0.8 54.8 0.68 1.2 52.6 0.49 0.9 61.2 1.79 2.9 62.5 2.28 3.6 56.9 Max. possible bias, YO -7.0 -4.8 -8.4 + 10.5 +13.6 Table 9. Results for zinc Standard solution (true concn. 47.0 pg 1-1) Mean concn.1 Laboratory 1- p9 1 45.4 2 53.4 3 48.1 4 52.3 5 49.3 ** Result outside target. Mean Std. Rel. dev.1 std. dev., 2.00 4.4 0.89 1.7 0.83 1.7 1.77 3.4 0.61 1.2 49.7 I- p9 Y O Max. possible bias, Y O -7.4 + 15.5 +4.0 +15.1 +6.2 River sample Mean concn.1 pg 1-1 32.4 32.4 26.1 31.2 28.8 Std.Rel. dev.1 std. dev., 1- p9 "/o 0.89 2.7 0.55 1.7 0.22 0.8 0.43 1.4 4.51 15.6** 30.2 Max. possible bias, Y O + 10.1 +9.0 - 14.3 +4.7 - 18.8 River sample +47.0 pg 1-1 Mean Std. Rel. Max. concn.1 dev.1 std. possible pg pg dev. bias, I-' 1-1 YO YO 81.5 4.82 5.9 +6.5 82.8 0.45 0.5 +3.0 84.9 2.69 3.2 1-8.2 80.8 75.8 0.76 1.0 -7.1 78.9 4.28 5.4 -7.4 Table 10. Results for mercury? Labora-tory 1 2 3 4 5 Mean Organic standard Inorganic standard (true concn. 0.936 pg 1-1) (true concn. 1 .000 pg I-') Mean concn.1 CLg I-' 0.953 0.958 1.064 0.930 1.018 Std. Rel.dev.1 std. I-' O/O 0.020 2.1 0.055 5.7 0.043 4.0 0.022 2.4 0.046 4.5 0.985 pg dev.3 Max. possible bias, YO +3.9 +8.0 + 18.0 -2.9 + 13.4 t For symbols see footnote to Table 1. Mean Std. Rel. concn.1 dev.1 std. dev., I- p rg Yo 0.975 0.044 4.5 1.064 0.037 3.5 1.294R 0'.063 4.9 0.959 0.030 3.1 1.040 0.071 6.8 1.010 Max. possible bias, YO -6.7 +9.9 -7.0 + 10.8 R River sample Mean Std. Rel. concn.1 dev.1 std. pg dev.7 :9 1-1 Yo 0.115 0.012 10.4 0.068 0.005 7.4 0.065 0.017 + 0.056 0.006 10.7 0.078 0.016 + 0.076 Max. possible bias, YO ++ ++ ++ ++ ++ River sample +0.936 pg 1-1 of organic Hg Mean Std. Rel. concn.1 dev.1 std. I-' I-' Yo 1.124 0.023 2.0 0.938 0.015 1.6 1.128 0.046 4.1 1.098 0.034 3.1 1.018 0.041 4.0 pg pg dev-7 1.061 Max.possible bias, YO +8.0 - 12.9 + 10.4 +6.5 -7.7 between 0.2 and 0.5 g of potassium persulphate (the amount depending on the expected organic content of the sample), followed by 1 ml of concentrated nitric acid (low-in-metals grade). The flask is stoppered and despatched to the labora-tory for analysis great care being taken to avoid contamina-tion. The sample is processed ready for analysis in the sample bottle by either bringing to the boil or if a concentration step is required evaporating five-fold. This pre-treatment converts any organic forms of mercury into the inorganic form suitable for analysis. The inter-laboratory comparability exercise was carried out on circulated standards and samples consisting of an inor-ganic mercury standard solution an organic mercury standard solution a river sample and the river sample spiked with organic mercury.The results obtained from five randomised replicate analyses in one batch are given in Table 10. With the exception of one bias result for the inorganic mercury standard solution (laboratory 3) all laboratories achieved the targets for both precision and bias for all the circulated solutions sample and spiked sample. Nutrients and Chloride The results obtained for the first inter-laboratory exercise for the nutrients ( i . e . ammonia total oxidised nitrogen silicate, orthophosphate and total phosphorus) and chloride were poor. Most laboratories failed to meet the targets for either the standard solutions or the samples or both for all determinands.However the advantages of the approach to comparability adopted were immediately obvious as in many instances it was possible to identify why individual labora-tories had failed to meet the targets and remedial action could then be taken. For example several laboratories failed to meet the targets for comparison of standard solutions and so clearly one major reason for the targets not being met was that the laboratory standards were inaccurate. Another factor was that most laboratories analysed the solutions using one range, typically a high range which was suitable for the sewage sample and the high concentration standard solution but completely inappropriate for the low-concentration river samples and solutions.Finally problems were identified wit ANALYST JANUARY 1984. VOL. 109 9 the precision achievable with certain analytical methods eg., distillation/Nesslerisation for ammonia and ion-selective elec-trodes for total oxidised nitrogen and the laboratories concerned were asked to investigate further and if necessary, change to an alternative method. After remedial action had been taken by the laboratories concerned the exercise was repeated with a fresh set of samples with similar concentrations. The results obtained were a considerable improvement over those obtained pre-viously although several laboratories still failed to meet the targets for some samples particularly the low-concentration standard solution and the river water sample.Further exercises were held to improve the quality of results so that the targets were met by the participants. Details of the results obtained from the final inter-laboratory exercise are con-sidered below. Ammonia For ammonia it can be seen that with the exception of laboratory 3 all laboratories achieved results that were either within or only just outside the required targets (Table 11). Laboratory 3 produced results well within the targets for precision but all the results for the river and low-concentration standards were outside the targets for bias. Clearly this laboratory had a fault in its low-concentration calibration standards during the period of the exercise; this was checked and found to be so and inter-laboratory comparability was confirmed in the next check exercise.Total Oxidised Nitrogen The results obtained for total oxidised nitrogen (Table 12) for precision and bias for the final exercise were all within the targets with the exception of the bias on the river sample for laboratory 6. Investigation revealed the cause of this large negative bias to be due to the problems with the automated continuous flow system used which occurred with real samples but not with synthetic solutions. It was overcome by the incorporation of a modified reducing reagent using hydrazine - copper - zinc instead of hydrazine - copper and changing the hydraulics of the analytical cartridge. This demonstrates the value of inter-laboratory comparability exercises involving real samples as this fault would not have been identified otherwise.In subsequent check exercises this fault has not reappeared. Silicate The results for the final exercise (Table 13) show that with the exception of one precision result and three bias results the targets have been achieved by all laboratories for all solutions. The one poor precision result and two of the bias results were only just outside the targets and taken overall the labora-tories’ results were considered acceptable. The laboratory that returned the bias result considerably outside the target checked its calculation but confirmed that the original result was correct. As their results for silicate on real samples in check exercises have been satisfactory it is considered that a simple arithmetic error was the probable cause of this spurious result.Orthophosphate The results for the final exercise (Table 14) were considered to be satisfactory although one laboratory just exceeded the bias target for its own low-level standard solution. As this laboratory achieved bias results well within the required targets for the circulated standards and samples its results were considered to be acceptable overall. For the river sample no bias result has been quoted for laboratory 2 as there was clear evidence from exchange of samples with another laboratory that their sample had been contaminated. Total Phosphorus The exercises involved the circulation of samples together with a standard solution of an organic phosphorus compound (adenosine-5-monophosphoric acid disodiuin salt; AMP).An organic phosphorus compound was used as it was considered to provide the best test of the laboratories’ digestion pro-cedures. If the results were satisfactory for this compound the procedures could be expected to be able to cope with the organic phosphorus compounds occurring in rivers and sewage effluents.17 The results obtained for the first exercise, however were far from satisfactory; all laboratories except one produced results considerably outside the precision and bias targets for the organic phosphorus standard. Most were also outside the targets for both the river and sewage effluent sample. Clearly the methods employed by the laboratories were unacceptable for total phosphorus determination. A closer inspection of results for the organic phosphorus standard revealed that most of the results were about only 50% of the actual concentration.The exception to this was one laboratory that had used a method developed by the Water Pollution Research Laboratory (WPRL). 18 This labor-atory’s results for the river and sewage effluent samples were also above the average. In view of this it was decided that all laboratories would standardise on the WPRL method. After a suitable interval of time to allow the laboratories to gain some experience with the method the exercise was repeated. The results obtained in the final exercise are given in Table 15 from which it can be seen that with only one small exception the results from all laboratories met the targets. The exception was laboratory 3 which had a result outside the bias target for its own standard solution (see also orthophos-phate above).This laboratory however obtained satisfac-tory results for all the other standards and samples and consequently no further action was considered necessary. Chloride The results obtained for chloride are summarised in Table 16 and were as expected generally acceptable. However the results were perhaps not as good as anticipated as two laboratories produced results outside both the precision and bias targets for the river sample. One of these laboratories also had the result for the sewage effluent outside the bias target. The reason for the outlying results was the same in both laboratories-the range of the continuous flow system em-ployed was unsuitable for the low concentrations found in this exercise.The normal working range in these laboratories could have been reduced to accommodate the low concentra-tions but then many routine samples would need dilution before analysis. After consideration of the implications of this for the laboratories’ normal workload together with the use made of chloride results in the Inputs Programme it was decided that a change of range was not appropriate. These results can be taken as confirming the choice of targets set for precision and bias which at first might seem unduly generous. Even for a relatively simple determinand such as chloride which is normally considered to be capable of precise and unbiased measurement problems have been experienced in achieving the targets.Biochemical Oxygen Demand Although biochemical oxygen demand (BOD) was not of primary importance for the Inputs Programme as oxygen deficiency is not a problem in the Severn Estuary it was considered that some knowledge of the mass input of BOD to the Estuary would be useful. Consequently one exercise wa Table 11. Results for ammonia (as N)t Laboratory's own standard Circulated standard (true concn. 0.68 mgl-1) Circulated standard (true concn. 4.0 mg 1-1) River sample Sewage Labora-tory 1 2 3 4 5 6 Mean Mean Std. concn.1 dev./ p p: 0.39 0.020 0.25 0.004 0.18 0.01 4.01 0.100 0.50 0.006 0.90 0.011 Rel. std. dev., YO 5.1* 1.6 5.6* 2.5 1.2 1.2 Max. poss . bias, YO -5.4 f0.9 - 12.9** +1.7 f0.7 k0.7 Mean Std.Rel. concn.1 dev.1 std. dev. p p yo 0.67 0.018 2.7 0.66 0.024 3.6 0.78 0.014 1.8 0.65 0.018 2.8 0.75 0.022 2.9 0.74 0.023 3.1 0.71 Max. poss. bias, YO -3.0 -5.0 +15.9** -6.0 + 12.2** 10.8** Mean concn.1 mg I - ' 4.05 4.13 4.03 3.96 4.16 3.90 Std. Rel. dev.1 std. mg dev., I-' Yo 0.25 6.2* 0.13 3.1 0.09 2.2 0.12 3.0 0.15 3.6 0.31 7.9** 4.04 Max. poss. bias, YO +4.9 +5.1 +2.1 -2.7 +6.2 -7.0 Mean Std. Rel. concn./ dev.1 std. dev., 0.36 0.02 5.6* 0.34 0.016 4.7 0.42 0.008 1.9 0.37 0.018 4.9 0.35 0.019 5.4* 0.34 0.019 5.6* ; yo 0.36 Max. poss. bias, Y O k3.2 -8.1 +18.0** +5.7 -5.8 -8.6 Mean Std. concn./ dev. ; p: 7.7 0.55 8.6 0.16 8.6 0.32 7.9 0.25 9.0 1.01 8.8 0.65 t For symbols see footnote to Table 1.Table 12. Results for total oxidised nitrogen (as N ) t Circulated standard (true concn. 3.4 mg 1 - 1 ) Laboratory's own standard Mean Std. Rel. Max. Mean Std. Rel. Max. concn.1 dev.1 std. poss. concn.1 dev.1 std. poss. Labora- mg mg dev. bias mg mg dev. bias, tory 1 - 1 1 - 1 Yo Yo I - I 1 - ' "/o O/" 1 5.10 0.167 3.3 -3.8 3.30 0.047 1.4 -3.7 2 1.97 0.049 2.5 -2.9 3.50 0.138 3.9 +5.3 3 0.99 0.031 3.1 -2.8 3.40 0.071 2.1 k1.2 4 15.9 0.22 1.4 -1.4 3.25 0.169 5.2* -7.3 5 0.96 0.012 1.3 -4.7 3.46 0.061 1.8 +2.8 6 4.46 0.037 0.8 -0.4 3.29 0.100 3.0 -4.9 Mean - 3.37 t For symbols see footnote to Table 1. Circulated standard (true concn. 18.3 mg 1-1) River sample Sewage Mean Std.Rel. concn.1 dev.1 std. mg dev. f3 1 - 1 yo 18.87 0.527 2.8 17.98 0.420 2.3 18.01 0.550 3.1 18.26 0.190 1.0 18.40 0.561 3.1 18.54 0.173 0.9 18.34 Max. poss. bias, YO +4.8 -3.1 -3.3 -0.8 +2.3 +1.9 Mean Std. Rel. concn.1 dev.1 std. mg mg dev., 1-1 I-' Yo 4.45 0.072 1.6 4.81 0.121 2.5 4.46 0.060 2.4 4.50 0.143 3.2 4.81 0.102 2.1 3.29R 0.073 2.2 4.61 Max. poss. bias, YO -4.4 +5.9 -4.6 -4.2 +5.6 R Mean concn.1 mg I-' 18.49 16.89 17.03 17.75 17.79 17.23 Std. dev. p: 0.612 0.339 0.574 0.228 0.665 0.31 Table 13. Results for silicate (as Si)? Labora-tory 1 2 3 4 5 6 Mean Circulated standard Circulated standard Laboratory's own standard (true concn.0.60 mg 1 - 1 ) (true concn. 2.4 mg 1-1) Mean Std. concn.1 dev.1 I- " 7: 0.60 0.014 2.28 0.084 2.03 0.022 7.82 0.147 1.00 0.005 4.71 0.077 Rel. std. dev., YO 2.3 3.7 1.1 1.9 0.5 1.6 Max. poss. bias, YO -9.8 -3.0 +2.1 -3.3 k0.3 +1.2 t For symbols see footnote to Table 1. Mean concn.1 mg I - ' 0.56 0.59 0.57 0.58 0.57 0.62 Std. Rel. dev.1 std. mg dev., I- I O/O 0.041 7.3* 0.031 5.3* 0.022 3.9 0.045 7.8** 0.004 0.7 0.018 2.9 0.58 Max. poss. bias, YO - 10.6** -4.7 -7.1 -7.7 -5.4 +5.1 Mean Std. Rel. concn.1 dev.1 std. mg mg dev., I - ' I - ' Yo 2.41 0.126 5.2* 2.43 0.146 6.0* 2.51 0.048 1.9 2.61 0.089 3.4 2.19 0.041 1.9 2.38 0.075 3.2 2.42 Max.poss . bias, YO +3.5 +4.8 +5.7 -t 10.9** -9.7 -2.6 River sample Sewage Mean Std. Rel. concn.1 dev.1 std. dev., 2.02 0.104 5.2* 2.08 0.059 2.8 1.96 0.022 1.1 2.17 0.122 5.6* 0.97R 0.007 0.7 2.05 0.025 1.2 : :5 yo 2.06 Max. poss. bias, Y O -4.9 +2.6 -5.5 +8.8 -1.2 R Mean Std. concn.1 dev. 17.59 0.305 18.64 0.067 16.93 0.143 17.05 0.452 18.48 0.465 18.75 0.227 pf 7: Table 14. Results for orthophosphate (as P)? Laboratory's own standard Circulated standard (true concn. 0.78 mg 1-1) Circulated standard (true concn. 2.0 mg 1-1) River sample Mean Std. Rel. Max. concn.1 dev.1 std. poss. Labora- mg mg dev. bias, 1 1.00 0.017 1.7 k1.0 2 0.04 0.001 2.5 -4.0 3 0.09 0.004 4.4 -12.3** 4 1.55 0.021 1.4 -3.9 5 0.50 0.011 2.2 k1.3 6 2.25 0.027 1.2 k0.7 tory 1- I I- 1 o/o YO - Mean t For symbols see footnote to Table 1.Mean concn . I mg 1 - 1 0.80 0.81 0.78 0.75 0.77 0.73 Std. Rel. dev.1 std. mg dev., I-' Yo 0.030 3.8 0.039 4.8 0.024 3.1 0.025 3.3 0.030 3.9 0.017 2.3 0.77 Max. poss. bias, O/O k4.8 +6.7 k1.8 -5.7 -3.5 -7.7 Mean concn.1 mg 1-1 1.97 1.99 2.00 1.93 1.98 1.94 ~~~ ~ Std. Rel. dev.1 std. dev. 7 yo 0.03 1.5 0.04 2.0 0.06 3.0 0.14 7.3* 0.07 3.5 0.03 1.5 1.97 Max. poss. bias, YO -2.4 -1.7 k1.7 -7.6 -3.0 -3.9 Mean concn.1 ;: 0.35 0.55R 0.34 0.36 0.36 0.33 Std. Rel. dev.1 std. dev., 0.023 6.6* 0.022 4.0 0.017 5.0 0.010 2.8 0.008 2.2 0.015 4.6 0.35 p Y Table 15.Results for total phosphorus (as P)t Laboratory's own standard Circulated standard (true concn. 2.0 mg 1-1) Circulated standard (trueconcn. 9.5mg1-1) River sample Labora-tory 1 2 3 4 5 Mean Mean concn .I 1- m: 1.02 0.98 0.08 0.84 1.03 Std. dev.1 1- m: 0.051 0.013 0.008 0.025 0.018 Re1 . std. dev., Yo 5.0 1.3 10.0 3.0 1.7 For symbols see footnote to Table 1. Max. poss . bias , O/O - 10.0 -2.8 -24.6** - 17.4 +4.0 Mean concn.1 mg I-' 1.71 2.17 2.03 2.15 2.18 Std. Rel. dev.1 std. mg dev., 1-1 Yo 0.165 9.6 0.136 6.3 0.057 2.8 0.048 2.2 0.027 1.2 2.05 Max. poss. bias, Yo -19.3 + 12.4 +3.2 +8.9 +9.8 Mean concn.1 mg 1-1 8.93 9.78 9.66 9.96 9.87 Std.Rel. dev.1 std. dev. :? Yo 0.21 2.4 0.45 4.6 0.29 3.0 0.66 6.6 0.19 1.9 9.64 Max. poss . bias, O/O -7.3 +5.7 +3.5 +8.9 +5.1 Mean Std. Rel. concn.1 dev.1 std. dev. p7 p yo 1.24 0.10 8.1 1.13 0.14 12.4* 1.32 0.08 6.1 1.24 0.03 2.4 1.29 0.05 3.9 1.24 Table 16. Results for chloride? Laboratory's own standard Circulated standard (true concn. 83 mg 1 - 1 ) Circulated standard (true concn. 186 mg 1-l) River sample Labora-tory 1 2 3 4 5 6 Mean Mean Std. concn.1 dev.1 ; ;: 122.5 4.06 98.4 0.52 25.3 0.44 25.1 0.32 90.1 0.39 157.7 1.60 -Re1 . std. dev., YO 3.3 0.5 1.7 1.3 0.4 1.0 Max.poss . bias, YO -3.1 -1.9 +2.2 +1.1 +0.4 -2.0 Mean concn.1 mg I-' 82.8 81.2 81.8 83.3 81.4 83.8 Std. Rel. dev.1 std. dev., 3.85 4.6 0.88 1.1 1.00 1.2 0.66 0.8 0.52 0.6 1.19 1.4 82.4 ; yo Max. poss. bias, Y O -2.9 -2.8 -2.1 +0.8 -2.3 + 1.8 Mean concn.1 ;: 182.3 178.3 181.7 185.7 183.1 182.8 Std. Rel. dev.1 std. dev. y yo 4.26 2.3 1.30 0.7 0.88 0.5 0.86 0.5 0.55 0.3 2.39 1.3 182.3 Max. poss. bias, YO -3.3 -4.5 -2.6 -0.4 -1.7 -2.5 Mean concn.1 mg I-' 18.8 21.7 20.9 21.4 20.7 18.8 Std. Rel. dev.1 std. dev. 1.65 8.8** 0.93 4.3 0.24 1.1 0.42 2.0 0.35 1.7 1.98 10.5** 20.4 ;3 yo t For symbols see footnote to Table 1 ANALYST JANUARY 1984 VOL.109 13 carried out to assess inter-laboratory comparability. The exercise involved the analysis of a blank (unseeded) a blank (seeded) the laboratories’ own standard glucose - glutamic acid solution a river sample and a sewage effluent sample using the “standard format.” As BOD is an empirical determinand in which the analytical method used could seriously influence the result obtained it was agreed that all laboratories would use the same method.19 The unseeded blank was used to calculate the results for the sewage sample and the seeded blank for the synthetic standard solutions. To ensure that different results were not produced simply by inappropriate dilution the dilutions to be used by all laboratories for the samples and standards were agreed beforehand.Clearly for a test such as BOD it is not possible to prevent biological activity in the river and sewage samples by sterilisation and still obtain a BOD result. Consequently the samples deteriorated over the 5-day period of the exercise, which made it impossible to assess bias for these real samples. Estimates of inter-laboratory bias were made however using the synthetic standard solutions. Similarly because of sample instability it was not possible to assess correctly the total standard deviation for the river and sewage samples and this was estimated by using the actual within-batch standard deviation (S,) obtained together with a between-batch stan-dard deviation (S,) calculation as a relative proportion of the between-batch standard deviation of the standard solutions.Such a calculated total standard deviation is thought to be the nearest to a “real-life” situation as can reasonably be obtained. The results obtained for the exercise are given in Table 17, from which it can be seen that with the exception of standard deviation values for the two standard solutions from labora-tory 3 all precision and bias results for samples and standards are within the targets. The reasons for laboratory 3’s standard deviations being outside the targets were investigated and appeared to be associated with the blank as their seeded blank values were very high and variable ( i . e . ranging from 1.1 to 3.2 mg 1-1 compared with 0.2-0.5 mg 1-1 obtained by other laboratories).Causes of the high blank values were investigated by the laboratory concerned and corrected. Organochlorine Pesticides A comparability exercise for organochlorine pesticides was undertaken on a round-robin format with the objective of simply assessing the capability of the participating laboratories for identifying correctly a range of pesticides. All laboratories employed their normal method of analysis, which involved essentially solvent extraction and concentra-tion of the pesticides followed by gas-chromatographic separation with electron-capture detection. To aid the positive identification of eluted peaks each laboratory analysed the solvent extract on two different chromatographic colymns. Three samples were circulated for the exercise consisting of (a) a synthetic solution containing a mixture of y-HCH aldrin and endrin in light petroleum (boiling range 40-60 “C) (b) a sewage works final effluent (from a sewage works known to receive effluent from a pesticide manufacturer) and (c) river water downstream of (b).The results obtained are detailed in Table 18 from which it was concluded that generally the agreement between labora-tories both qualitative and quantitative was fairly good. The exercise had however identified a number of points in which improvements could be made the chief of these being that the stocks of pure pesticide standards at several laboratories should be increased and that fresh pesticide calibration standards should be regularly prepared in order to improve quantitative results.Conclusions The programme of work to establish and maintain compar-ability of results for 17 determinands between laboratories Table 17. Results for biochemical oxygen demand as measured in the bottle? Laboratory 1 2 3 4 5 6 Mean Laboratory’s own standard Circulated standard (true concn. 4.40 mg 1-1) (dilution 1 in 50) (trueconcn. 1.80 mgl-l) (dilution 1 in 50) Mean concn.1 mg 1-1 4.10 4.18 4.38 4.04 4.20 4.32 Std. dev . I mg 1-1 0.19 0.18 0.87** 0.24 0.16 0.23 4.20 Max. possible biaslmg 1-1 -0.41 -0.32 -0.52 -0.50 -0.29 -0.21 t For symbols see footnote to Table 1. Mean concn.1 mg I-’ 1.71 1.69 1.88 1.88 1.86 1.79 Std. dev.1 mg 1-1 0.21 0.23 0.96** 0.19 0.16 0.17 1.80 Max.possible biaslmg 1-1 -0.21 -0.24 +0.64 +0.19 +O. 15 -0.11 River sample (no dilution) Sewage effluent (dilution 1 in 5) Mean Std. concn.1 dev.1 mg 1-1 mg 1-1 3.74 0.35 4.07 0.30 2.80 0.74* 3.28 0.50 3.73 0.16 4.10 0.18 3.62 Mean Std. concn.1 dev.1 mg 1-1 mg 1-1 5.54 0.34 3.80 0.20 4.28 0.42 4.80 0.32 5.54 0.66* 3.56 0.20 4.59 Table 18. Results for organochlorine pesticides? Synthetic mixture Sewage effluent River sample y-HCHI Laboratory pg 1-1 1 30 2 25 3 31 4 15R 5 29 6 29 Aldrinl 28 25 30 25 29 27 EL&’ Eldrinl 50 40 41 23R 40 44 pgl-’ (u-HCHI ELg1-I NSO NS 0.21 0.04R 0.17 0.14 y-HCHI 0.20 0.18 0.23 0.02R 0.32 0.21 pg I-’ Dieldrin1 0.16 0.12 0.21 ND 0.20 0.16 ELg 1-Aldrinl ND ND ND 0.01 ND ND I.18 I-’ (u-HCHI I%-’ NS NS 0.02 ND 0.10 0.03 y-HCHI ccgl-’ 0.05 0.04 0.03 0.04 0.14R 0.04 Dieldrin1 0.05 0.03 0.06 ND ND 0.04 pg I-’ Aldrinl ND ND 0.007 0.02 ND ND I.Lgl-‘ t For symbols see footnote to Table 1.0 NS no standard available at laboratory 14 ANALYST JANUARY 1984 VOL. 109 participating in the Severn Estuary Inputs Programme was arduous and for almost all determinands was much more difficult to achieve than at first expected. In the course of the collaborative work many laboratories found their existing analytical methods to be unsatisfactory and discarded them in favour of alternative procedures.The performance of some laboratory instrumentation was found to be unsatisfatory and had to be replaced and some new standardised methods and analytical techniques had to be developed to meet the demands of the programme. In general the exercises have shown that comparability of results can be achieved only by careful and constant attention to detail even for relatively simple determinands such as ammonia and total oxidised nitrogen. Once achieved com-parability cannot be taken for granted and it is essential that check exercises be regularly held with real samples to ensure that the required standard is maintained. The approach to achieving comparability employed in this work has subsequently been adopted by several other organ-isations for their own inter-laboratory comparability pro-grammes.This paper is published by kind permission of the Technical Working Party of the Severn Estuary Joint Committee (Chairman W. F. Lester Severn-Trent Water Authority), whose encouragement throughout the duration of the pro-gramme is gratefully acknowledged. Our appreciation is also due to Morlais Owens (Welsh Water Authority) for his advice, guidance and encouragement in this work. Appendix The period covered by the work described in this paper coincided with a consolidation of laboratory facilities within each Water Authority and a concomitant increase in the level of instrumentation available within most laboratories. This has inevitably meant that during this period changes in membership of the Chemists’ Sub-committee and indeed several changes in the number of laboratories participating in the programme have taken place.The complete list of Members and laboratories is given below together with an indication of the period of time over which they participated. This list is not the same order of laboratories as the coded listings given in the tables of results in the paper. Dr. K. C. Wheatstone Chairman (Severn-Trent Water Authority Birmingham); Mrs. E. M. Hewson Secretary, 1976-79 (Severn-Trent Water Authority Birmingham); Mr. M. M. Day Secretary 1979 onwards (Severn-Trent Water Authority Birmingham); Mr. M. R. Barker (Wessex Water Authority Divisional Laboratory Poole) (1975-76); Mr. J. G. Jones (Wessex Water Authority Bath); Dr. R. F. Mantoura (Institute for Marine Environmental Research) (1975-76); Mr. B. Milford (South West Water Authority, Exeter Laboratory) (1975); Dr. A. W. Morris (Institute for Marine Environmental Research) (1975-76); Dr. C. Pattinson (Welsh Water Authority Marine Laboratory Ponthir) (1978 onwards); Mr. A. Poole (Wessex Water Authority Divisonal Laboratory Saltford); Dr. E. Salt (Wessex Water Authority, Divisional Laboratory Bridgwater); Dr. W. Simpson (Welsh Water Authority Marine Laboratory Ponthir) (1977-78), Dr. J. Stoner (Welsh Water Authority Brecon); Mr. F. Sweeting (Wessex Water Authority Divisional Laboratory, Saltford); Mr. R. Toft (South West Water Authority Exeter Laboratory) (1976 onwards); Mr. J. E. Tomkin (Welsh Water Authority River Division Laboratory Caerleon; also rep-resenting Bridgend Hereford and Llanelli River Divisional Laboratories); Mr. C. Triner (Wessex Water Authority, Divisional Laboratory Poole); and Mr. K. Wagstaff (Severn-Trent Water Authority Regional Laboratory Malvern). 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. References Water Research Centre “Standard Analytical Samples,” (a) “Results of First Distribution,” WRA TM91 1974; ( b ) “Results of Second Distribution,” WRC TM94 1974; (c) “Results of Third Distribution,” WRC TM97 1974; ( d ) “Results of Fourth Distribution,” WRC TM98 1974; (e) “Results of Distributions 5 6 7 8 and 9,” WRC TR 65 1977; Water Research Centre Medmenham. United States Public Health Service Analytical Reference Service “A Series of Reports on the Results of Sample Distributions Between 1956 and 1971,” Robert A. Taft Sanitary Engineering Centre Cincinnati OH. “Analytical Quality Control for a Group of Collaborating Laboratories Inter-laboratory test,” TM96 Water Research Centre Medmenham 1974. Wilson A. L. Analyst 1979 104,273. Analytical Quality Control (Harmonised Monitoring) Com-mittee Analyst 1979 104 290. “Accuracy Required of Analytical Results for Water Quality Data Banks,” TR34 Water Research Centre Medmenham, 1976. McFarren E. F. Lishka R. J. and Parker J. H . Anal. Chem. 1970,42 358. “Water Quality 1977/78,” Severn-Trent Water Authority, Birmingham 1978 p. 145. Thompson K. C. Analyst 1978 103 1258. Thompson K. C. and Wagstaff K. Analyst 1980 105 641. Riley J . P. and Taylor D. Anal. Chirn. Acta 1968 40 479. Florence T. M. and Batley G. E . Talanta 1976 23 179. Abdullah M. I. El-Rayis D. A. and Riley J. P. Anal. Chirn. Acta 1976 84 363. Thompson K. C . Wagstaff K. and Wheatstone K. C., Analyst 1977 102 310. ‘ Bertenshaw M. P. Gelsthorpe D. and Wheatstone K. C., Analyst 1981 106 23. Bertenshaw M. P. Gelsthorpe D. and Wheatstone K. C., Analyst 1982 107 163. Morris A. W. personal communication. “Laboratory Procedure No. 15,” Water Pollution Research Laboratory Stevenage 1972. Department of the Environment “Analysis of Raw Potable and Waste Waters,” HM Stationery Office London 1972 p. 85. Paper A2155 Received March 8th 1982 Accepted September 15th 198

ISSN:0003-2654    

DOI:10.1039/AN9840900003

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST JANUARY 1984 VOL. 109 15 Studies with lmmobilised Chemical Reagents Using a Flow-cell for the Development of Chemically Sensitive Fibre-optic Devices Gordon F. Kirkbright Ramaies Narayanaswamy and Neal A. Welti Department of Instrumentation and Analytical Science UMlST Manchester M60 7QD UK Results of studies using a flow-cell and immobilised indicator dye reagents for the development of a chemically sensitive fibre-optic system for measurement of pH are presented. The indicator dye reagents are immobilised on a cross-linked styrene - divinyl benzene polymer matrix and reflectance measurements are made as a function of pH using a bifurcated fibre-optic system. Keywords Flow-cell; immobilised chemical reagents; chemically sensitive pH fibre-optic device Recently the development of chemically sensitive fibre-optic devices has been the subject of considerable interest.The techniques employed have normally been based on the immobilisation of chemical reagents that allow a colorimetric or fluorimetric indication of the chemical change in a solution. Fibre-optic pH sensors based on the use of classical indicator dye optical absorption have been developed for physiological and other uses. 1-3 Optical waveguides coated with chemical reagents which utilise changes in absorption of the reagent material have been employed to sense ammonia in the vapour phase."5 Fibre-optic sensors using changes in fluorescence have been reported for the measurement of pH,6 aluminium7 and glucose.* Chemically sensitive fibre-optic devices have several advan-tages over conventional potentiometric or voltammetric elec-trode sensors.One of the most important advantages is their inherent safety in operation due to their non-electrical nature. Other advantages of such sensors include their small size and flexibility capability of remote operation in hostile environ-ments rugged construction and reliability. Our present interest is in the development of fibre-optic transduction systems based on immobilised chemical reagents for the measurements of pH gases and metals for both industrial on-line applications and environmental and clinical purposes. In this paper we report our preliminary studies using a flow-cell to assist the development of chemically sensitive fibre-optic probes for measurement of pH.A reliable small sensor based on immobilisation of bromothymol blue on XAD-2 polymer has been developed as a result of these studies; the performance characteristics of this sensor and the development of a simple dedicated pH meter into which this is incorporated are to be described elsewhere. Experimental Instrumentation The arrangement of the flow-cell used in the study of the immobilisation of indicator dyes and its associated instrumen-tation is shown in Fig. 1. Radiation from a tungsten halogen lamp (12 V 55 W) was modulated using a rotating sector and is passed through a bifurcated fibre-optic. The fibre-optic employed was a bundle containing 16 polymer fibres (Quan-tum Jump Ltd. Liverpool) with a total diameter of ca. 1 mm. The sensing end of the fibre-optic was located in the flow-cell (Fig.2). Solutions of known pH were passed through the flow-cell at a constant rate using a peristaltic pump (Cole-Palmer Instrument Co. Model 7567-10). The reflectance signal through the fibre-optic was measured at a given wavelength using a grating monochromator (Varian Techtron AA4 Monochromator Australia) and a photomultiplier tube (Hamamatsu R213). The reflectance spectra were recorded on a chart recorder (JJ X - Y plotter Model PL51). The pH of Source Rotating sector t t Reservoir chromator t 1 I I Reservoir parameter recorder (P.M.T.) meter con t ro I Fig. 1. flow-cell Schematic diagram of instrumentation employed with the t lmmobilised ii disc 7dicator t Direction of flow Fig.2. Flow-cell assembly with immobilised indicator dye and fibre-opti 16 ANALYST. JANUARY 1984. VOL. 109 the test solutions was measured with a Model ECM201 pH meter (EDT Research London). The pH of the medium was altered by the addition of suitable aliquots of 1 M hydrochloric acid and sodium hydroxide solutions. Measurements were made by attaching the input and the reflectance output arms of the bifurcated fibre-optic to the source lens housing and monochromator respectively by means of light-tight fittings. Reagents Indicator dyes and chemical reagents used for the preparation of buffer solutions were purchased from BDH Chemicals Ltd. and XAD-2 polymer beads (a cross-linked polymer of styrene and divinylbenzene) also obtained from BDH Chem-icals Ltd. were used as the polymer support.Analytical-reagent grade methanol was employed. Immobilisation Procedure XAD-2 polymer beads were washed thoroughly with distilled water and then with acetone. The polymer was dried and stored. About 1 g of the polymer beads was placed in 10 ml of 0.1% indicator dye solution in methanol and left to stand for 4 h. The polymer with adsorbed indicator dye was washed with distilled water and then placed in the flow-cell using the pump. The polymer beads surrounded the fibre-optic in the flow-cell and were packed tightly so that they did not move. A constant flow-rate of liquid through the flow-cell containing immobi-lised indicator dye was maintained throughout the experi-ments; the flow-rate employed was 0.10 1 min-1. Reflectance Measurements The optimum wavelengths for measurements of the reflec-tance signal from the immobilised indicator dye may be determined from the reflectance spectra obtained.Fig. 3 shows the reflectance spectra recorded from the immobilised bromothymol blue indicator in the flow-cell at different pH values. The wavelength at which reflectance measurements were made was chosen as that wavelength at which a large change in the reflectance signal with pH was observed. Accordingly a wavelength of 593 nm was used for measuring reflectance signal from the immobilised bromothymol blue indicator. It was also found that this wavelength (593 nm) was also suitable for measurement of the reflectance signal for the other immobilised indicator dyes studied. The reflectance signals reported here are expressed as a percentage of reflectance and no reference wavelength to provide a ratio-metric measurement was employed.Results and Discussion Response Versus pH The reflectance signal measured from the six indicator dyes examined after immobilisation on the cross-lin ked XAD-2 polymer beads varies with pH as shown in Fig. 4. The approximate linear pH ranges and estimated indicator con-stants (pKI,) of the immobilised indicator dyes are presented in Table 1. It is noted that the indicator dyes studied all gave a linear response within a region of ca. 2 pH units. The response was observed to be reproducible rapid and reversible with changes in pH. It was observed that all the dyes except phenolphthalein were firmly retained in the polymer matrix and were not extracted by the aqueous solutions employed.The indicator phenolphthalein was extracted from the poly-mer matrix by alkaline solutions (methyl orange indicator dye was similarly extracted from the polymer matrix by acid solutions) and would not be suitable for use as a durable pH sensor in the immobilised state. Effect of Temperature and Ionic Strength The temperature coefficient of the pH response of immobili-sed bromothymol blue and thymolphthalein in the flow-cell was measured by recording the response of the indicator dyes while varying the temperature of the buffer solution between 25 and 45 "C. The temperature coefficient of the immobilised bromothymol blue and thymolphthalein expressed as change Table 1. Approximate linear pH ranges and estimated indicator constants of immobilised indicator dyes Apparent indicator constant Bromophenol blue .. 3.0-5.0 (3.0-4.6) 3.7 (4.1) Bromocresol purple . . 5.0-7.0 (5.2-6.8) 5.8 (6.1) Bromothymol blue . . 7.0-9.0 (6.0-7.6) 7.7 (7.1) Thymol blue . . . . 9.0-1 1 .O (8.0-9.6) 9.6 (8.9) Phenolphthalein . . . . 9.6-11.0 (8.3-10.0) 10.2 (9.6) Thymolphthalein . . . . 10.5-12.0 (8.310.5) ll.0(9.3) Indicator pH range* (PK,") * 9 t * Values in parentheses are from reference 9 in aqueous solution. 1. Approximate pKrn from this work. 100 350 500 650 Wavelengthlnm 800 0 -I-.- I I 1 I 2 4 6 8 10 12 PH Fig. 3 Reflectance spectra obtained from immobilised bromothymol blue in the flow-cell at a pH of (A) 2.77 (B) 7.07 (C) 8.19 (D) 9.18 and (E) 11.17 Fig.4. Reflectance signal as a function of pH for the immobilised (1) bromophenol blue (2) bromocresol purple (3) bromothymol blue, (4) thymol blue (5) phenolphthalein and (6) thymolphthalei ANALYST JANUARY 1984 VOL. 109 17 x A O B o c I I 1-I 7.0 8.0 9.0 10.0 PH Fig. 5. Effect of (A) 0.1 M (B) 0.5 M and ( C ) 1 .O M sodium chloride in the buffer solution on the response (reflectance vs. pH) for immobilised bromothymol blue in the flow-cell of pH per “C temperature change was found to be 0.013 k 0.003 and 0.015 k 0.003 respectively over this temperature range. The effect of variation of ionic strength on the response of the indicator dye bromothymol blue was investigated using a series of buffer solutions containing 0.1,0.5 and 1.0 M sodium chloride to vary the ionic strength; it was observed that pH response of the bromothymol blue indicator was independent of variation in ionic strength over this range as shown in Fig.5. Conclusions This study demonstrates the potential feasibility of chemical pH sensors based on immobilised indicator dye reagents. The indicator dyes studied present some useful pH ranges for measurements. The response characteristics yield rapid, reversible sensors that are independent of variations in ionic strength of sample solutions. Using the results of this study, we have developed a miniature self-contained fibre-optic pH probe for use in static or flowing solution sample streams. The fibre-optic probe has important advantages over an electrode in safety reliability applicability and cost. The performance characteristics of the fibre-optic probe are at present under investigation and will be reported in a later paper. 1. 2. 3. 4. 5. 6. 7. 8. 9. References Peterson J. I. Goldstein S. R. Fitzgerald R. V. and Buckhold D. K . Anal. Chem. 1980 52 864. Goldstein S. R. Peterson J. I. and Fitzgerald R. V. 1. Biomech. Eng. 1980 102 141. Peterson J. I . and Goldstein S. R. Diabetes Care 1982 5 , 272. Smock P. L. Orofino T. A. Wooten G. W. and Spencer, W. S . Anal. Chem. 1979 51 505. Giuliani J. F. Wohltjen H. and Jarvis N. L. Opt. Lett., 1983 8 54. Saari L. A. and Seitz W. R. Anal. Chem. 1982,54 821. Saari L. A. and Seitz W. R. Anal. Chem. 1983 55 667. Schultz J. S. Mansouri S. and Goldstein I. J. Diabetes Care 1982,5,245. Vogel A. I . “A Textbook of Quantitative Analysis,” Long-mans London 1979 p. 241. Paper A31248 Received August 8th 1983 Accepted September 23rd 198

ISSN:0003-2654    

DOI:10.1039/AN9840900015

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, JANUARY 1984, VOL. 109 19 Sequential Flow Injection Voltammetric Determination of Phosphate and Nitrite by Injection of Reagents into a Sample Stream Arnold G. Fogg and Nuri K. Bsebsu Chemistry Department, Loughborough University of Technology, Loughborough, Leicestershire, LEI I 3TU, UK For the intermittent analysis of large-volume water samples, such as a hydroponic fluid, a modified use of flow injection analysis is proposed. In this modification the water sample is made the eluent in the flow injection system and reagents for the determination of individual components are injected into the sample stream sequentially. As an illustration of this principle a sample solution containing phosphate and nitrite has been analysed in this way using acidic molybdate and acidic bromide reagents to determine phosphate and nitrite voltammetrically at a glassy carbon electrode.Phosphate was determined at the 0.5-50 x 10-5 M level and nitrite at the 0.55 x 10-4 M level without mutual interference and with good precision. Keywords: Flow injection analysis; phosphate determination; nitrite determination; reagent injection In the usual application of flow injection analysis (FIA), aliquots of sample are injected into a stream of reagent and the derivative that is formed is determined at a suitable detector.’ Previously, flow injection voltammetric procedures were developed in this laboratory for determination of phosphate2.3 and nitrite4 using this normal procedure. Early in these studies the determination of phosphate and other nutrient ions in hydroponic fluids using this approach was being considered.3 In this and similar applications the requirement is not for a rapid throughput of samples but for intermittent analyses at infrequent intervals of a single sample of changing composi- tion.Further in these applications the sample solution is in plentiful supply and is inexpensive. At that time in unpub- lished work the possibility of making determinations by injection of reagent into a sample stream was demonstrated. Johnson and Petty5 have reported a visible spectrophoto- metric method for the determination of phosphate using FIA with injection of reagent. They point out that whereas with normal FIA, in which the sample is injected into a reagent stream, the sample is diluted by dispersion as it passes to the detector, in reverse FIA, where the reagent is injected into a sample stream, the concentration of “sample” in the injected reagent plug increases as the plug passes to the detector.For this reason, assuming that a sufficiently concentrated reagent solution is injected, the sensitivity of the reverse technique should be greater than that of the normal technique. Johnson and Petty5 obtained a 5-fold increase in sensitivity for the reverse technique over that reported by R6iiCka and Hansenl using the analogous normal technique. In this work the application of the reverse technique with voltammetric detection has been demonstrated. Further, the sequential injection of different reagents to determine several constituents of a sample stream has been illustrated by the determination of phosphate and nitrite.Experimental Flow-injection analysis was applied as before2-4 except that the sample solution was used as the eluent. The flow of sample stream was usually produced by means of a Metrohm pressure bottle system (EA 1101) working at 0.8 bar, although a peristaltic pump (Gilford Minipuls 2) was used also. Injections were made by means of a Rheodyne injection valve (5020). The sample stream was presented to a glassy carbon electrode (Metrohm EA 286) in the wall-jet configuration. A Metrohm detector cell (EA 1096) was used, but without inserting the counter and reference electrodes and with partial immersion of the cell in electrolyte (0.01 M sulphuric acid). Contact between the electrolyte and the counter and reference electrodes was made by means of salt bridges.The glassy carbon electrode was cleaned with 1 M sodium hydroxide solution daily, or as required. It was found to be unnecessary to de-gas the sample stream. The potential of the glassy carbon electrode was held at +0.22V (versus S.C.E.) in phosphate determination and at +0.30 V in nitrite determina- tion using a PAR-174 polarographic analyser (Princeton Applied Research). Current signals were monitored on a Tarkan 600 Y - t recorder. All the results reported here were obtained with the Metrohm pressure vessel and a Metrohm detector cell. The same glassy carbon electrode was used throughout. Repeat injections of a reagent into a sample solution gave reproducible signals; coefficients of variation (eight injec- tions) were typically <1 YO.Reagents Standard orthophosphate solution, 3 X 10-3 M (285 pg ml-1 of POq3-). Dissolve 0.408 g of analytical-reagent grade potas- sium dihydrogen orthophosphate in water and dilute to 11 in a calibrated flask. This solution is 3 x l o - 3 ~ in phosphate. Prepare less concentrated standard solutions by dilution. Acidic molybdate solution, 0.5%mlV. Add 0.6 ml of analytical-reagent grade concentrated sulphuric acid to 60 ml of water. Dissolve 0.5g of ammonium molybdate in the resulting solution and dilute to 100 ml with water. Standard sodium nitrite solution, approximately 1 X 10-2 M. Dissolve approximately 0.172 g of analytical-reagent grade sodium nitrite, accurately weighed, in water and dilute to 250ml in a calibrated flask.This solution is 1 X 10-*M in nitrite. Prepare more dilute standard solutions from this solution. Acidic bromide solution, 20% mlV in potassium bromide and 3 . 2 ~ in hydrochloric acid. Dissolve 20g of potassium bromide in 70ml of water, add 27.6ml of concentrated hydrochloric acid, cool the solution, dilute to 100 ml and mix. Procedures Direct injection of acidic molybdate reagent into a stream of phosphate solution Inject 100 pl of 0.6% mlV acidic molybdate reagent into a stream of phosphate solution (0.5 x 10-5-50 x 10-5 M). Use a 3-m delay coil (0.58 mm bore).20 ANALYST, JANUARY 1984, VOL. 109 Direct injection of acidic bromide reagent into a stream of nitrite solution Inject 100 1-11 of 20% mlV acidic bromide reagent into a stream of nitrite solution (0.5 x 10-4-5 x 1 0 - 4 ~ ) .Use a 3-m delay coil (0.58 mm bore). Results Injection of the 2% mlV acidic molybdate reagent (1001-11) used previously2.3 in the normal FIA method into a sample stream of 2.0 x l o - 4 ~ phosphate was studied initially. The phosphate sample stream was prepared as required by a 10-fold dilution of the stock solution. The effect of the length of the delay coil between the injection valve and the detector cell at this concentration of phosphate was studied. The results given in Table 1 indicate that the optimum length for the delay coil is 3 m. Injections into a sample stream with a lower phosphate concentration (2 x l o - 5 ~ ) were made next using the 3-m delay coil and the same reagent. A proportionately much lower peak current value of 0.27yA was obtained and a double peak was apparent. The formation of a double peak would be expected when dispersion occurred only at each end of the sample bolus.Experience suggests, however, that they may be arising here also owing to an electroanalytical phenomenon or artefact that we have been unable so far to identify. The formation of a double peak was clearly unsatis- factory and attention was directed to studying the effect of the reagent composition. Reagent solutions were prepared with different concentrations of ammonium molybdate while the sulphuric acid concentration was kept constant at 0.6% V/V. Table 2 shows the effect of the concentration of ammonium molybdate used on the peak current obtained for a sample solution 2 x l o - 4 ~ in phosphate.The use of an ammonium molybdate concentration of 0.5% mlV was adopted. The optimum concentration of sulphuric acid used was found to be the same as previously, i.e., 0.6% VlV. The effect of the length of the delay coil on the signal obtained at the new optimum concentration of ammonium molybdate (0.5% mlv) was studied (see Table 3). The optimum length was found to be 3m, as before. The signals obtained when the optimised reagent (0.5% mlV in ammonium molybdate and 0.6% V/V in concentrated sulphuric acid) was injected into a sample stream of phosphate at the 2 X M levels were 0.61 and 6.18 PA, respectively. A rectilinear calibration graph was obtained using signals obtained with eluents 0, 1,2,3,4,10,20,30 and 40 X 10-5 M in phosphate. Determination of phosphate at the 1 X 1 0 - 6 ~ level was not possible, however, as the signal obtained was similar to the blank obtained by injecting reagent into distilled water.Attention was directed at the determination of nitrite by injection of reagent into sample solution. Injection of acidic bromide reagent that is 3 . 2 ~ in hydrochloric acid and 20% mlV potassium bromide into the sample stream of 2 x 10-4 M nitrite was studied initially. The nitrite sample stream was prepared as required by a 10-fold dilution of the stock solution. The effect of the length of the delay coil between the injection valve and the detector cell at this concentration of nitrite was studied. The peak heights at different delay-coil lengths are given in Table 4. The 3-m delay coil gave the largest signal.The effect of the reagent composition was studied next. Reagent solutions were prepared with different concentra- tions of potassium bromide while the hydrochloric acid concentration was kept constant at 3.2 M. The results in Table 5 show the effect of the concentration of potassium bromide used on the peak current obtained for a sample solution 2 X l o - 4 ~ in nitrite. Clearly the highest concentration of and 2 x bromide gives the greatest signal. The effect of the concentra- tion of hydrochloric acid used was then studied (see Table 6). Again the highest concentration gives the largest signal. Precipitation occurs at potassium bromide concentrations of >20% mlV in 3.2 M hydrochloric acid, and at hydrochloric acid concentrations >3.2 M in 20% mlV potassium chloride solution.Determination of nitrite at the 1 x 1 0 - 5 ~ level was not possible as the signal obtained was similar to that obtained for the blank, i.e., injection of reagent into distilled water. Nitrite can be determined at the 2 x l o - 3 ~ level but the signal is disproportionately less than that obtained at 2 x 10-4 M. Determinations of phosphate and nitrite were made on a series of solutions 0.5 x 10-5-50 x 10-5 M in phosphate and 0.5 X 10-4-5 x 10-4 M in nitrite. No mutual interference was observed within these limits. Typical signals are shown in Fig. 1. Table 1. Direct injection of acidic molybdate reagent: effect of delay coil length on peak current. Phosphate concentration of eluent = 2 X M. Molybdate concentration = 2% m/V Delaycoillength/m . .. . 1.0 2.0 3.0 4.0 5.0 Peakcurrent/pA . . . . . . 2.50 4.70 5.80 5.00 5.00* * Reproducibility bad. Table 2. Direct injection of acidic molybdate: effect of amount of ammonium molybdate on the peak current. Delay coil length = 3 m. Phosphate concentration of eluent = 2 x l o - 4 ~ Concentration of ammonium molybdateinreagent, YO m/V . . 0.5 1.0 1.5 2.0 Peakcurrent/pA . . . . . . . . 6.20 6.00 5.20 4.65 Table 3. Direct injection of acidic molybdate reagent: effect of delay coil length at optimised ammonium molybdate concentration. Phosphate concentration of eluent = 2 x 10-4 M. Molybdate concen- tration = 0.5% m/V Delaycoillength/m . . . . 1 2 3 4 5 Peakcurrent/pA . . . . . . 2.0 3.8 6.15 6.25* 5.75 * Reproducibility bad. Table 4. Direct injection of acidic bromide reagent: effect of delay coil length on peak current.Nitrite concentration in eluent = 2 X 10-4 M. Bromide concentration = 20% m/V. Hydrochloric acid concentration = 3.2 M Delaycoillength/m . . . . . . 1.0 2.0 3.0 4.0 Peakcurrent/pA . . . . . . . . 2.80 3.05 3.20 2.80 Table 5. Direct injection of acidic bromide reagent: effect of potassium bromide concentration on peak current. Nitrite concentra- tion in eluent = 2 X M. Hydrochloric acid concentration = 3.2 M. Delay coil length = 3 m Potassium bromide concentration, %m/V . . . . . . . . . , 5.0 10.0 15.0 20.0 Peakcurrent/pA . . . . . . . . 0.6 1.30 2.30 3.30 ~ Table 6. Direct injection of acidic bromide reagent: effect of hydrochloric acid concentration on peak current. Nitrite concentration in eluent = 2 x 10-4 M.Potassium bromide concentration = 20% m/V. Delay coil length =; 3 m Hydrochloric acidconcentrationh . . 1.0 2.0 3.0 3.2 Peakcurrent/pA . . . . . . . , 1.2 1.98 2.60 3.05ANALYST, JANUARY 1984, VOL. 109 21 1 C i A i Time 4 Fig. 1. Typical signals obtained for the determination of phosphate in the presence of nitrite using the reverse FIA procedure. Phosphate concentration in eluent: A, 2 x 10-5; B, 4 x 10-5; C, 6 x 10-5 and D, 6 x 10-5 M. Nitrite concentration in eluent: A-C, 1 X M; and D, zero Discussion PhosphatezJ and nitrite4 were determined previously by injection of sample solution into eluents consisting of acidic molybdate and acidic bromide reagents and monitoring at a glassy carbon electrode the molybdophosphate and nitrosyl bromide formed.Johnson and Petty5 studied the inversion of the roles of reagent and sample solution [which they termed reverse FIA (rFIA)] for the visible spectrophotometric determination of phosphate, and showed that increased sensitivity was obtained. This work has been concerned with assessing the reverse procedure for the determination of phosphate and nitrite voltammetrically. The voltammetric signals obtained in the reverse procedure were found to be of a similar magnitude to those obtained with the normal proce- dure, but it was not possible to make determinations at as low a level with the reverse procedure. This appears to be due to the particular characteristics of the voltammetric detection method. Thus, for successful voltammetry a reasonably conducting solution is required but unless additional elec- trolyte is added to the eluent, the blank is essentially pure water. Further, there is a background current associated with the reagents,2 and the signal associated with changes in this on injection appear to be more effectively minimised in the normal FIA method.A brief study of the addition of inert electrolyte to sample solutions has indicated that this does not readily solve the problem, but further studies are being made. Thus, whereas usable signals were obtained at the 1 x 1 0 - 6 ~ level of phosphate and nitrite by injection of sample into reagent, the minimum amounts to produce usable signals by injection of reagent into sample solution are 5 x 10-6 and 5 X 10-5 M , respectively. The determination of phosphate by the reverse FIA method has a useful rectilinear range and might be applied with advantage to the determina- tion of phosphate in samples such as hydroponic fluids.An advantage of using the sample as eluent is that contamination of the glassy carbon electrode occurs more slowly than it does in the normal FIA method (particularly for phosphate) in which reagent is passed over the electrode continuously during the determination. The satisfactory determination of nitrite by this reverse FIA procedure, except within a very narrow range of concentrations, has not proved possible so far. No mutual interference was found in the determination of phosphate and nitrite in the same eluent and the determina- tion of these ions by sequential injection of acidic molybdate and acidic bromide into the sample stream has been demon- strated.The sample mixture is somewhat artificial but the principle of sequential determination of constituents in a sample stream is clearly illustrated. The ions which are required to be determined in hydroponic fluids include phosphate, nitrate, potassium and ammonium. Calcium, sodium and certain heavy metals are also determined fre- quently. Computer-controlled systems for monitoring conduc- tivity, pH and temperature are currently in use in research institutes and elsewhere, and control systems for the monitor- ing of the ions mentioned above are being sought actively. Some progress has been made in this laboratory for on-line reduction of nitrate to nitrite prior to voltammetric determina- tion.6 We envisage the construction of a reverse FIA system in which both voltammetric and potentiometric detectors are used. Valve switching could be used to direct the sample stream to the required detector, to inject the required reagent for voltammetric determination and to change the potential of the voltammetric electrode as required. The authors thank Mr. G. S. Weaving of the National Institute of Agricultural Engineering for information and advice and for his interest in this work. N. K. B. thanks the people of the Socialist People’s Libyan Arab Jamahiriya for financial support and .leave of absence. 1. 2. 3. 4. 5. 6. References RfiiiEka, J., and Hansen, E. H., “Flow Injection Analysis,” John Wiley, New York, 1981. Fogg, A. G., and Bsebsu, N. K., Analyst, 1981, 106, 1288. Fogg, A. G., and Bsebsu, N. K., Analysf, 1982, 107, 566. Fogg, A. G., Bsebsu, N. K., and Abdalla, M. A., Analyst, 1982, 107, 1040. Johnson, K. S., and Petty, R. L., Anal. Chem., 1982,54,1185. Fogg, A. G., Chamsi, A. Y., and Abdalla, M. A., Analyst, 1983, 108, 464. Paper A3t142 Received May 19th, 1983 Accepted August 23rd, 1983

ISSN:0003-2654    

DOI:10.1039/AN9840900019

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, JANUARY 1984, VOL. 109 23 Determination of Chromium in Steel by Flame Atomic-absorption Spectrometry Using a Flow Injection Standard Additions Method* Julian F. Tyson and Ahyar B. ldris Department of Chemistry, Lough borough University of Technology, Lough borough, Leicestershire, LE77 3TU, UK The determination of chromium in steel by atomic-absorption spectrometry is briefly reviewed and the basis of the flow injection standard additions method explained, in which the novel configuration of using the sample as the carrier stream is employed. The effects of iron, fuel to oxidant ratio and dissolution procedure were investigated and a procedure is described that allows a conventional instrument-optimising strategy to be used, requires no releasing agents and uses pure chromium standard solutions.The selection of appropriate flow injection conditions is discussed in the light of the single well stirred mixing chamber model for dispersion. The application of the method is demonstrated by the analysis of six British Chemical Standard steels. Keywords : Flow injection analysis; standard additions method; a tom ic-a bso rp tion spectrometry; ch ro mium determination; steel analysis The determination of chromium in steel by atomic-absorption spectrometry with an air - acetylene flame has been exten- sively studied’-9 and is reported to be subject to a large number of interference effects. The factors that affect the slope and shape of the calibration graph include the presence of iron, the nature of the acids used for dissolution, the oxidation state of the chromium and the flame stoicheiometry.A variety of methods for overcoming these interferences have been proposed, including the addition of releasing or suppressing agents,3~5~6~7~9 separation and solvent extrac- tion,2.4 matrix matching,lJJOJl the use of the dinitrogen oxide - acetylene flame8JoJ1 and the use of a plasma.12 The interference effects of iron and acids on the chromium signal have been studied under various flame conditions.7.’3-15 The depressive effect of iron and the releasing action of ammonium chloride7J3J4 and quinolin-8-01-6 have been dis- cussed. The mechanisms of the interference from acids15 and the effects of other cations16 have also been explained in great detail. The effects of the oxidation state of chromium on the calibration graph have also been investigated.17-18 Most reports3.5-9-12 have advocated the use of a mixture of hydrochloric and nitric acids for sample dissolution, although various other acid mixtures have been used.1J,4JO711 In this paper, the use of aflow injection-based analogue of the standard additions method is described. This approach avoids the need to use releasing agents and considerably simplifies the volumetric manipulations of the conventional standard additions procedure, as the dispersion produced between the point of injection and the nebuliser can be designed to mimic the addition of standards to the sample followed by dilution to volume. A simple model for the dispersion behaviour obser- ved in flow injection - atomic-absorption systems, based on considering all the dispersion to be produced by a single well stirred mixing chamber, has been proposed19.2o and the use of this model to calculate the concentration of interferent appropriate for a given calibration sequence and dispersion has been described for the flow injection analogues of matching standard21 and standard additions21.22 methods.In the flow injection standard additions method the reverse configuration to the normal methods of flow injection analysis is used, in that the sample is used as the carrier stream into which are injected discrete volumes of the pure standards. The * Presented at the Royal Society of Chemistry Analytical Division meeting on “Research and Development Topics in Analytical Chemistry,” held at Loughborough University of Technology on March 28th and 29th, 1983.dispersion is designed (by suitable selection of volume injected and carrier tube dimensions) so that at the peak maximum (the measurement point) the appropriate ratio of interferent to analyte is achieved. The calculation takes account of the dilution of (a) the injected standard, (b) the analyte in the carrier and (c) the interferent in the carrier. As with all standard additions methods, the interferent to analyte ratio above which the depressive effect becomes constant must be known for the successful application of the method. The relationship between the relevant parameters is where C: is the concentration of interferent in the sample carrier stre-am, C: is the concentration of the top standard injected in the calibration sequence, D is the dispersion, CX is the concentration of the analyte in the sample carrier stream and Ril, is the minimum ratio of interferent to analyte necessary to achieve the maximum interference.c: = [CS/(D - 1) + CX]Ri,a . . . . (1) Experimental Apparatus This was as described previously? An air - acetylene flame was used throughout the work. The flame conditions and flow injection parameters used were as follows: Air rotameter reading . . . . . . . . 8.2 1 min-1 Acetylene rotameter reading . . . . 4.91min-1 Lamp current . . . . . . . . . . . . . . 6 mA Burner height , . . . . . . . 4.2 (arbitrary units) Slit width . . . . , . . . . , 1 (0.18 nm band pass) Wavelength . . . . . . . . . . . . 357.9 nm Pumping rate . . . . . . . .. . 5.95 ml min-1 Tube length . . . . . . . . . . 2.3 or 200 cm Tube i.d. . . . . . . . . . . . . . . 0.58mm Volume injected . . . . . . . . . . 100 or 50 1-11 The combination of a tube length of 2.3cm and with an injection volume of 1OOyl gave a dispersion of 1.2 and was used to analyse samples BCS 251/1, 254/1 and 25511. The combination of a tube length of 200cm and an injection volume of 50 1-11 gave a dispersion of 4 and was used to analyse samples BCS 261/1, 241/2 and 220/2. Reagents Chromium(II1) standards were prepared by serial dilution of a 1000 p.p.m. stocksolutionofchromium(II1) nitratein 1 nitric24 ANALYST, JANUARY 1984, VOL. 109 acid (BDH Chemicals). Iron(II1) solution (10 000 p.p.m.) was prepared by dissolving the appropriate amount of high-purity iron granules (BCS 149/3) in hydrochloric (sp. gr.1.18) and nitric acids (sp. gr. 1.42).9 Procedure Preliminary experiments The optimum pumping rate19 and the variation of dispersion with tube length and volume injected for a given tube diameter were established.22 The effects of iron and acids were investigated and the effect of fuel to oxidant ratio was studied by varying the acetylene flow-rate from 4.0 to 5.5 1 min-1 in the presence and absence of iron. For the dissolution of the samples, four different acid mixtures were investigated, namely a mixture of hydrochloric and nitric acids,9 a mixture of hydrochloric, nitric and perchloric acids,10 a mixture of sulphuric, nitric and hydrofluoric acids2 and a mixture of phosphoric, sulphuric and nitric acids.’ Steel samples It is necessary to know the approximate ratio of iron to chromium in the final sample solution, which should contain about 10 p.p.m.of chromium. If the approximate chromium content is unknown, then a preliminary experiment is requi- red. This also applies to the iron to chromium ratio, which should be established from a preliminary experiment or sufficient iron may be added to the final solution to ensure that the appropriate ratio is achieved. Transfer up to 0.5 g into a 250-ml PTFE beaker, add 10 ml of hydrochloric acid (sp. gr. 1.18) and 5 ml of nitric acid (sp. gr. 1.42). Cover the beaker with a clock-glass and heat gently until the sample has dissolved. Evaporate the solution just to dryness, coolanddissolve theresiduein 10 mlofhydrochloricacid(sp. gr.1.18), warming to trbtain complete dissolution. Cool and transfer the solution, with filtration if necessary, into a 100-ml calibrated flask. Dilute to volume with distilled water. Dilute thesolution so that the final solution containsabout 10p.p.m. of chromium and either about 500 p.p.m. of iron (if dispersion 4 is used) or about 3 000 p.p.m. of iron (if dispersion 1.2 is used). Use this final solution as the carrier stream and inject standards covering the range 0-20 p.p.m. Measure peak absorbance changes from the chart recording as either positive or negative values and plot against the appropriate standard concentration. Draw a smooth line through the calibration points and read off the concentration of the unknown solution from the intercept on the concentration axis.Results and Discussion The interference effect of iron is shown in Fig. 1. The degree of depression of the chromium signal increased sharply as iron was added and then levelled off at a mass ratio of iron to g 0.40 f 0.30 400 800 1200 1600 2 000 [Fe(lll)l, p.p.m. Fig. 1. Effects of iron on the absorbance of a 10p.p.m. chromium solution. A, Fuel-lean flame; B, fuel-rich flame chromium of 30 : 1. The depressive effect was greater with a fuel-rich flame than with a fuel-lean flame. The effects of hydrochloric and nitric acids in the presence and absence of iron are shown in Fig. 2. In the absence of iron, neither acid had much effect on the chromium signal but, in the presence of iron, the extent of the depression was not the same for both acids. The releasing effect observed with hydrochloric acid was presumably due to the formation of the metal chlorides, which have relatively low melting- and boiling-points.The effects of the other acids were complex and varied. They also depended on the presence of other cations or anions in the solution. The dissolution method of Nall et al.9 (hydrochloric and nitric acids) was judged to be the most suitable for the steels investigated in this study owing to its relatively simple effect on the chromium signal. It was also the most convenient method to use. Although a small amount of undissolved silica remained, this did not affect the accuracy of the results. The effect of the fuel to oxidant ratio is shown in Fig. 3. The maximum signal was obtained at a fuel flow-rate of 4.9 1 min-1, which was a slightly luminous flame.In the presence of iron, the depression was more severe for a fuel-rich than for a fuel-lean flame. An example of the recorder trace obtained is shown in Fig. 4 (dispersion 1.2) and the resulting calibration graph in Fig. 5. The standard deviations of the peak heights ranged from 1.4 X 10-3 to 5.1 x 10-3 absorbance units for the 0 and 21 p.p.m. standards, respectively. These values correspond to 0.48% and 1.20% relative standard deviation based on AA values. Naturally, as the concentration of standard injected approaches the concentration of the carrier stream, the relative standard deviation based on decreasing AA values increases. The shapes of chromium calibration graphs have been discussed by Thompson.17 The departure from the “normal” smooth curve shape observed here is in agreement with Thompson’s findings, although regions of negative or 0.60 I 1 0.401L-----.0.30 n PI U 0 1 .o 2.0 3.0 4.0 5.0 Concentration of HCI or H N 0 3 i ~ Fig. 2. Effects of acids on the absorbance of a 10 .p.m. chromium solution in the presence and absence of iron. A, CRromium in nitric acid; B, chromium in hydrochloric acid in the absence of iron; C , chromium in hydrochloric acid; and D, chromium in nitric acid in the presence of 1 000 p.p.m. of iron I I I Fuel rich Fuel lean I ~ 3.7 4.0 4.3 4.6 4.9 5.2 5.5 Acetylene flow-ratell min-’ Fig. 3. Effects of fuel to oxidant ratio. A, 10 p.p.m. of chromium; B, 10p.p.m. of chromium and 160p.p.m. of iron; C , 10p.p.m. of chromium and 660p.p.m.of iron; and D, 10p.p.m. of chromium and 1160p.p.m. iron25 0.4 0.3 0.2 0.1 Q d 0 - -0.1 -0.2 -0.3 ANALYST, JANUARY 1984, VOL. 109 21 0.5 - - - - - - - 4 0 I t Time Fig, 4. Typical chart recording for the flow injection standard additions method (AA, change in absorbance). The steel sample contained about 10 p.p.m. of chromium and the injected standards covered the range 0-21 p.p.m. in 3 p.p.m. increments /x I I X I 0.30 - 0.20 - 0.10 - Q a 3.0 6.0 9.0 12.0 15.0 18.0 21.0 C: p.p.m. Fig. 5. Typical calibration graph for the flow injection standard addition method (CS, concentration of standard). The sample concentration, CX, is obtained from the intercept on the CS axis zero slope on the calibration graph were not observed. In view of Thompson’s findings and conclusions (“the determination of chromium in the luminous air - acetylene flame optimised for maximum chromium sensitivity is not recommended’’) and from the extent of the depression observed at this fuel to oxidant ratio in the presence of iron, it may appear that the choice of this fuel to oxidant ratio for the analyses described here is not soundly based.However, it was decided that this value should be used, as by far the easiest and commonest way of setting up an atomic-absorption instrument for the determi- nation of any element is to set the various operating parameters to give maximum sensitivity, while nebulising a single pure standard solution of concentration calculated from the table of sensitivities in the manufacturer’s handbook. The instrument would then be used for determinations in which interferences were operating without any readjustment of parameters.The results of this study show that the flow injection standard additions method can be used successfully with this setting-up strategy. The results obtained for a number of BCS steels containing from 0.19 to 17.4% of chromium are shown in Table 1. Additional iron was added to the first three samples to achieve the necessary 30 : 1 mass ratio of iron to chromium to give the maximum interference effect on all the standards in the Table 1. Results for BCS steels Certified Sample value, YO Chromium found, % BCS 261/1 . . . . . . 17.4 17.4,17.4,17.5,17.6 BCS241/2 . . . . . . 5.35 5.33,5.36 BCS22012 . . . . . . 5.12 5.12,5.13,5.13,5.13,5.12 BCS 251/1 0.51 0.52,0.51 BCS254/1 .. . . . . 0.27 0.27,0.27 BCS255/1 . . . . . . 0.19 0.20,0.20 . . . . . . calibration sequence. As the top standard contained 21 p.p.m. of chromium and the flow injection conditions were selected to give a dispersion of 4, then the concentration of interferent in the carrier stream is calculated from equation (1) to be 525 p.p.m. for a sample concentration of half the top standard. In the experiments reported here the sample solutions were diluted so that the chromium concentration was about 10p.p.m. and sufficient iron(II1) solution was added so that the final solution contained an additional 500p.p.m. of iron. This, together with the iron already present in the samples, was considered to provide an adequate “safety margin.” As can be seen from Fig.5 , satisfactory results could have been obtained if the 15 p.p.m. standard were considered the “top” standard and so, in fact, there was a considerable safety margin. The other three samples contained a much higher ratio of iron to chromium and thus the dispersion could be decreased while still achieving the necessary maximum depressive effect. The effect of a change in top standard or sample concentrations on the concentration of interferent necessary for the successful application of the standard additions method can thus readily be calculated. Similar calculations can be performed for other interfering com- ponents of the solutions. In this study, for example, it was necessary to ensure that the effects due to the hydrochloric and nitric acids used in the dissolution procedure (see Fig.2) were taken into account when the final acidity of the sample solutions was considered. In theory, the equation could be used to calculate the dispersion necessary for the method to work for given values of the other parameters. Rearrangement of equation (1) gives The value of D, by definition, cannot be less than 1, so it is immediately apparent and there is a lower limit for C z (equal to CXRiIa) for successful application of the method. However, as Cz approaches this limit the value of D required becomes very large and there are two practical difficulties associated with large values of D. Firstly, the sensitivity, i.e., the slope of the calibration graph, is inversely proportional to D and thus, as D increases, the sensitivity decreases and the uncertainty in the interpolated value at AA = 0 (see Fig.5) increases. Eventually, of course, at large values of D, AA becomes indistinguishable from the noise on the signal. The second problem concerns the way in which dispersion is increased. If D is increased by increasing the length of tubing between the injector and the nebuliser then the peak is broadened and thus the time between injections must be increased to avoid carryover and cross-contamination. If D is increased by decreasing the volume injected then the precision becomes a problem as small changes in the volume injected cause large changes in the value of D (see Fig. 3 in reference 22). There is also a minimum volume that can be injected owing to the mode of construction of the injection valve.There are thus a number of practical restrictions on an upper value of D and so it appears sensible to select D with due regard to sensitivity, peak width and precision and then to calculate the concentration of interferent required from equation (1). This may mean that interferent has to be added to samples, if the concentration is not high enough, as was done in three of the analyses reported here. D = [C:Z?i,a/(Cg - CXRila)] + 1 . . . . (2)26 Conclusion In searching for solutions to analytical problems in which matrix interference effects are encountered, three approaches are, in general, applied. Either the analyte species is (a) separated literally from the interfering species prior to the final measurement step of the overall procedure, or (b) a figurative separation is achieved by the addition of selective reagents with appropriate control of reaction conditions or by appropriate use of some instrumental correction procedure or (c) the calibration procedure is designed to compensate for the interferences by ensuring that the standards are subjected to the same interferences as the samples, either by matching the standards to the samples if the nature and concentration of the interferents are known, or by standard additions if they are not.All of these approaches are used for analyses in which flame atomic-absorption spectrometry is used as the measure- ment stage. All of the methods have attractive theoretical features but all suffer from a number of practical disadvan- tages, not the least of which is the time taken to follow the method through for an individual sample.In practice, the analytical chemist uses professional skill and judgement to select the most appropriate approach for the particular problem. Generally, it appears that for the determination of minor alloying components of steels by flame atomic- absorption spectrometry, real separation methods (such as solvent extraction) are not much favoured in comparison with a combination of figurative separation (addition of releasing or protecting agents) and matching standards to samples (see, for example, reference 9). In this paper it has been shown that flow injection techniques for sample introduction to the spectrophotometer in conjunction with the design of the dispersion produced in the flowing stream (based on calcula- tions from equations derived from the simple hypothetical well stirred mixing chamber model for dispersion) can be used to provide an alternative to the conventional standard additions method. The flow injection-based method has the advantages of requiring fewer volumetric manipulations, being less time consuming and being interpolative rather than extrapolative.The procedure developed here also allows a straightforward instrument-optimising strategy to be used and ANALYST, JANUARY 1984, VOL. 109 could be readily adapted to composite analytical procedures in which more than one element is determined in each sample. Financial support for A. B. Idris from the National University of Malaysia is gratefully acknowledged. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. References Kinson, K . , Hodges, R. J., and Belcher, C. B., Anal. Chim. Acta, 1963, 29, 134. Donaldson, E. M., Talanta, 1980, 27, 773. Pandey, L. P., Ghose, A, Dasgupta, P., and Rao, A. S . , Talanta, 1978, 25, 482. Fogg, A. G., Soleymanloo, S . , and Burns, D. T., Talanta, 1975, 22, 541. Purushottam, A., Naidu, P. P., and Lal, S . S . , Talanta, 1973, 20, 631. Ottaway, J. M., and Pradhan, N. K., Talanta, 1973,20, 927. Barnes, L., Anal. Chem., 1966, 38, 1083. Cobb, W. D., Foster, W. W., and Harrison, T. S . , Analyst, 1976, 101,255. Nall, W. R., Brumhead, D., and Whitham, R., Analyst, 1975, 100,555. Thomerson, D. R., and Price, W. J., Analyst, 1971, 96, 321. Thomerson, D. R., and Price, W. J., Analyst, 1971, 96, 825. Atsuya, I., Anal. Chim. Acta, 1975,74, 1. Roos, J. T. H., and Price, W. J., Spectrochim. Acta, Part B, 1971, 26, 441. Roos, J. T. H., Spectrochim. Acta, Part B, 1972,27,473. Nasuta, T., Suzuki, M., and Takeuchi, T., Anal. Chim. Acta, 1970, 51, 381. Yanagisawa, M., Suzuki, M., and Takeuchi, T., Anal. Chim. Acta, 1970, 52, 386. Thompson, K. C., Analyst, 1978, 103, 1258. Green, H. C., Analyst, 1975, 100, 640. Tyson, J. F., Anal. Proc., 1981, 18, 542. Tyson, J. F., and Idris, A. B., Analyst, 1981, 106, 1125. Tyson, J. F., Appleton, J. M. H., and Idris, A. B., Anal. Chim. Acta, 1983, 145, 159. Tyson, J. F., Appleton, J. M. H., and Idris, A. B., Analyst, 1983, 108, 153. Paper A3i171 Received June 15th, I983 Accepted August lst, 1983

ISSN:0003-2654    

DOI:10.1039/AN9840900023

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, JANUARY 1984, VOL. 109 27 Micro-determination and Separation of Titanium in Environmental Samples by Spectrophotometry Using a Liquid Ion Exchanger K. Sobhana Menon and Yadvendra K. Agrawal Analytical Laboratories, Pharmacy Department, Faculty of Technology and Engineering, M.S. University of Baroda, Baroda, India A sensitive and selective method for the micro-determination of titanium, based on the direct measurement of light absorption by a liquid ion exchanger phase that has absorbed the coloured complex species derived from the sample, has been 'developed. The sensitivity of the titanium(1V) - benzohydroxamic acid (BHA) complex extracted with a liquid ion exchanger, Aliquat 336, in the presence of thiocyanate is ten times greater than that of the titanium(1V) - BHA complex extracted with hexanol or isoamyl alcohol.The method can be applied to the concentration, determination and separation of titanium(1V) in industrial effluents and environmental samples containing very low levels of the metal. Keywords: Titanium determination; spectrophotometry; liquid ion exchanger; environmental samples Liquid ion exchangers have shown great potential for the separation of metal ions in mineral acids, organic acids and complexing agent media. 1-5 Although a few attempts have been made to develop colour in the liquid ion exchanger phase6 after extraction from the aqueous phase, no work has been reported so far on the extraction of coloured species into the liquid ion exchanger phase. The method described here is based on the direct measurement of light absorption by the ion exchanger phase that has absorbed the coloured complex species.The method can be applied to the direct detection, pre-concentration, determination and separation of metals present in industrial effluents and environmental samples. In most instances the sample species are converted into complex anions of higher valency within the ion exchanger phase, leading to considerable enhancement of the sensitivity. Hydroxamic acids have proved to be very sensitive and reasonably selective reagents for the trace determination of rnetals.7.8 In acidic media benzohydroxamic acid (BHA) forms a stable yellow complex with titanium(IV),'~'O which can be extracted with hexanol or isoamyl alcohol. This paper describes the extraction of the titanium(1V) - BHA complex with a liquid ion exchanger, Aliquat 336, which can be applied to the determination of microgram amounts of the metal in the chemical form and in the environment, and to its separation from closely associated metals such as vanadium and molyb- denum, by selective elution.Experimental Chemicals and Reagents All chemicals were of analytical-reagent or general-purpose reagent grade from BDH Chemicals and E. Merck, unless otherwise specified. Benzohydroxarnic acid. This was prepared as described elsewhere.11 The method was slightly modified by washing the acid with light petroleum after recrystallisation from ethyl acetate and drying under vacuum for the removal of the last traces of the solvents. Titanium standard solutions. A stock solution of titanium was prepared by heating 0.2 g of pure titanium(1V) oxide in a Pyrex flask with 8.0 g of ammonium sulphate and 25 ml of concentrated sulphuric acid.After cooling, the resulting solution was transferred into a 250-ml calibrated flask. The Pyrex flask was washed with 5% sulphuric acid and finally diluted to the mark with water. The solution was standardised spectrometrically with N-phenylbenzohydroxamic acid (PBHA)12 and the titanium content was found to be 0.400 mg ml-1. The solution was diluted as and when required. Ion Exchangers Amberlite LA-1 [(N-dodecyltrialkylmethyl)amine] (Rohm & Haas, Philadelphia, PA, USA), Aliquat 336 (tricaprylmethyl- ammonium chloride) (Fluka, Buchs, Switzerland) and tri- octylamine (Fluka) dissolved in suitable diluents, in varying proportions, were used.Apparatus A VSU-2P (Carl Zeiss, Jena) spectrophotometer with matched quartz cells was used for spectral measurements. A Varian Techtron AA-5 atomic-absorption spectrophoto- meter equipped with a Himamatsu photomultiplier tube and an HCN-Ti hollow-cathode lamp was used for titanium determination. The measurements were made at 364.3 nm with a 20mA current and a dinitrogen oxide - acetylene flame. The details of the procedure were as described by the EPA.13 Procedure A sample solution containing 2-20 pg of the metal was taken in a 60-ml separating funnel. Adjustment of the aqueous phase to a total volume of 15ml of 6 - 8 ~ hydrochloric acid, using concentrated hydrochloric acid and water, was carried out after adding 2 ml of a 0.1% solution of BHA in water followed by 1.5-2ml of 2% ammonium thiocyanate solution.The mixture was shaken gently with 15ml of a 4% solution of Aliquat 336 in benzene for about 1 min. The two layers were allowed to settle. After drying over anhydrous sodium sulphate, the organic layer was transferred carefully into a 25-ml calibrated flask. The anhydrous sodium sulphate was washed with the solvent and the washings were collected. Finally, the combined extract was diluted to the mark with the solvent and the absorbance was measured at 360 nm against a reagent blank. To calculate the distribution ratio, D, and the percentage extraction, E , the titanium concentration in the aqueous phase was determined14 from a calibration graph. The metal content in the organic phase was determined by the procedure given below.Titanium was eluted from the ion exchanger phase with 2 N sulphuric acid and determined spectrophotometrically . l5 For the elution of titanium, the layer obtained after extracting the complex from the aqueous phase was shaken with 5 ml of 2 N sulphuric acid for 5 min. The two phases were28 ANALYST, JANUARY 1984, VOL. 109 allowed to settle and the aqueous layer was withdrawn carefully. The organic layer was again shaken with 2 ml of 2 N sulphuric acid and the aqueous layer was collected. The combined extract was then mixed with 2ml of hydrogen peroxide (20 volume) and diluted to 10 ml with 2 N sulphuric acid. The absorbance was measured at 410 nm. The amount of titanium was calculated from a calibration graph.Determination of titanium in sample solutions A titanium sample solution was prepared by dissolving a high-purity steel sample (standard) in hydrofluoric acid and, after evaporation nearly to dryness, fuming the residue with 10ml of concentrated sulphuric acid and then diluting to 500ml with 4% sodium oxalate solution. An aliquot of the solution was taken and 10 ml of 0.1% tin(I1) chloride solution were added to reduce iron(III), vanadate, etc. The titanium content was determined according to the procedure. Determination of titanium in industrial effluents A 100-ml volume of the effluent was taken, acidified with hydrochloric acid and about 1 g of sodium sulphide was added to precipitate any molybdenum present. The contents were filtered and hydrogen sulphide was boiled off from the filtrate by adding concentrated nitric acid.A 20-ml volume of 0.1% tin(I1) chloride solution was added and the titanium was determined according to the procedure. Determination of titanium in plant and soil samples A 1-g amount of the soil or plant (ash) sample was digested with a mixture of perchloric and nitric acids and centrifuged to remove the siliceous material. The mother liquor was concen- trated by heating and diluted to 100 ml with 0.05 M hydro- chloric acid. An aliquot of the solution was taken and about 1 g of sodium sulphide was added. After centrifugation, the mother liquor was heated with a few drops of concentrated nitric acid to boil off hydrogen sulphide, diluted and then 10 ml of 0.1% tin(I1) chloride solution was added.Finally, the titanium content was determined according to the procedure. Results and Discussion The extractability of the titanium(1V) - BHA complex with a liquid anion exchanger shows the anionic nature of the complex. The yellow complex is extracted with a 4% solution of Aliquat 336 from a medium of 6-8 M hydrochloric acid. The extracted species has maximum absorbance at 360 nm. Addi- tion of thiocyanate favours the extraction through a synergic effect and increases the molar absorptivity, although the absorbance maximum remains unchanged. Effect of Variables on the Extraction Acids and acidity The titanium complex was extracted with different mineral acids, viz., hydrochloric, sulphuric and perchloric acids. Extraction was very poor in sulphuric acid and with higher concentrations of perchloric acid (>5 M) (Table 1).Extraction with 6-8 M hydrochloric acid was quantitative and complete. At lower hydrochloric acid concentrations (<4 M) the intensity was lower but increased steadily as the hydrochloric acid concentration increased, to a maximum at 6-8 M. The results are given in Table 1. Reagent Concentration The absorbance of titanium was constant with the use of excess of the reagent. Extraction with various concentrations of the reagent showed that 1-2 ml of 0.1 M BHA was adequate for quantitative extraction of the sample and the mixed ligand Table 1. Effect of acids and acidity Acid Concentration/ M HCI . . . . . . 2 4 6 8 HCI04.. . . . . 2 4 6 H 2 S 0 4 . . . . . . 4 8 Extraction, Y O 58.0 85.5 99.9 99.9 60.0 87.0 Turbidity 40.7 63.0 Table 2.Effect of various diluents on the extraction (YO) of titanium(1V) with Aliquat 336 Aliquat 336concentration, '/O Diluent Benzene . . . . . . Toluene . . . . . . , . Xylene . . . . . . . . Hexane . . . . . . , . Kerosene . . . . . . Chloroform . . . . . . Carbon tetrachloride . . 4 99.9 82.5 90.1 65.3 68.9 57.4 48.9 6 99.9 80.3 90.7 65.8 68.8 50.8 48.5 complex of titanium, whereas extraction was incomplete at lower concentrations. Electrolytes The extraction was carried out under the optimum conditions with various amounts of sodium chloride and potassium chloride. The electrolyte had no effect on the extraction of the titanium complex. Ammonium thiocyanate concentration The use of ammonium thiocyanate favours the extraction and increases the intensity of the colour.For maximum extraction and colour intensity, 1.5-2 ml of a 2% solution of ammonium thiocyanate was found to be sufficient. Aliquat 336 concentration The optimum concentration of Aliquat 336 was studied by varying the concentration from 0.5 to 15% in benzene. (Caution-Benzene is highly toxic and appropriate precau- tions should be taken.) The extraction was quantitative from 4% and remained constant up to 10%. A tendency to form an emulsion was observed at higher concentrations of the ion exchanger. A 4% solution of Aliquat 336 in benzene was adopted for extraction. Diluents Titanium was extracted with 4 and 6% Aliquat 336 solution in various diluents. Equilibration was effected by maintaining the ratio of organic to aqueous phase at 1 : 1.The extraction was very poor with chloroform and carbon tetrachloride and xylene showed incomplete extraction. The extraction was complete and quantitative with benzene and a clear separation was obtained. Hence benzene was used as the diluent in subsequent work (Table 2). Type of liquid anion exchanger Titanium was extracted with three extractants in various diluents (Table 3). Aliquat 336 in benzene was found to be the best for extraction.ANALYST, JANUARY 1984, VOL. 109 29 Table 3. Effect of different liquid anion exchangers on the extraction of titanium(1V) Liquid Extraction, anion exchanger Diluent YO Aliquat336(4%) . . . . Benzene Xylene Chloroform Xylene Chloroform Xylene Chloroform Amberlite LA-1 (4%) . . Benzene Trioctylamine (4%) .. Benzene 99.9 90.1 57.4 72.3 64.0 38.0 50.4 39.0 15.8 Equilibration time and stability The extraction was very rapid and required only a few seconds for quantitative and complete extraction. The time of shaking was varied from 15 s to 5 min. The extraction was quantitative within 30 s. The complex extracted under optimum conditions was stable for several days. Optical properties The colour system obeyed Beer's law from 0.2 to 3.5 p.p.m. of titanium at 360nm and the optimum range (Ringbom p1ot)lh was 0.1-2.5 p.p.m. The molar absorptivities were 6.82 X 1031 mol-1 cm-1 (without thiocyanate) and 1.8 X lo4 1 mol-1 cm-1 (with thiocyanate), and the molar absorptivities of titanium(1V) - BHA complexes extracted from isoamyl alcohol9 and hexanoll" were 1.48 X l o 3 and 2.32 X 103 1 mol-1 cm-1, respectively.Composition and Stability Constant The composition of the titanium mixed ligand complex was studied by the slope ratio method,l7 i.e., by plotting a graph of the logarithm of the distribution coefficient of metal [logD,,,] against the logarithm of the ligand concentration [log(lig- and)]. The extraction was carried out by taking a fixed amount of titanium in the presence of (a) a constant amount of thiocyanate and Aliquat 336 and varying the concentration of BHA, (b) a constant amount of BHA and Aliquat 336 and varying the concentration of thiocyanate and (c) a constant amount of BHA and thiocyanate and varying the concentra- tion of Aliquat 336. In all three instances the plot of logD,,, against -log(ligand) (Fig.1) gave a straight line with slope of 2, 1 and 1, respectively, which indicates that the compositions of the complexes are (i) titanium: BHA = 1 : 2, (ii) titanium : BHA : SCN- = 1 : 2 : 1 and (iii) titanium : BHA : SCN- : Aliquat 336 = 1 : 2 : 1 : 1. Possible reactions are as follows: Ti02+ + 2(C6H5CONHOH)2 + C1- + TiO(C6H5CONH0)2Cl- + 2H+ TiO(C6H5CONH0)2Cl- + SCN- -+ TiO(C6H5CONH0)2SCN- + c1- TiO(C6H&ONH0)2SCN- + R4N+ + R4N+ TiO(C6H5CONH0)2SCN- where R4N+ represents the cationic part of the liquid anion exchanger. Hence the extracted species is probably R4N+[TiO- (BHA)2SCN]-. The stability constant of the mixed ligand complex determined by the spectrophotometric method18 was 6.94 x 1013. Stripping After extraction of titanium into the organic phase, it was stripped with 15 ml of varying concentrations (0.05-5 M) of sulphuric acid, hydrochloric acid, nitric acid, sodium sulphate, sodium chloride, sodium carbonate and sodium hydroxide 100 Q 8 10 -J 1 0.01 0.1 LogIRl 1 Fig.1. Logarithmic raphs of distribution coefficient versus ligand concentration: A, log[&HA]; B, log[SCN-1; and C, log[Aliquat 336) Table 4. Effect of diverse ions. Ti taken: 20 pg per 25 ml Tolerance limit/ Foreign ion Added as Ag+ . . . . . . AgN03 Be2+ . . . . . . BeS0, Mg2+ . . . . . . MgSO, Ba2+ . . . . . . BaC1, Ca2+ . . . . * ' Ca(N03)2 Sn2+ . . . . . . Sn(N03)2 Pb2+ . . . . . . Pb(N03)2 co*+ . . . . . . COCI, cu2+ . . . . . . CUSO, Hg2+ . . . . . . HgC1, Ni2+ . . . . . . NiC1, Zn2+ . . . . . . . ZnSO, Mn2+ . . . . . . MnCl, Cr3+ . . . . .. CrCI, A13+ . . . . . . A1Cl3 M o ~ O ~ ~ ( + . . . . (NH4)hMo7024 Zr4+ . . . . . . Z T ( N O ~ ) ~ C1- . . . . . . NaCl Br- . . . . . . NaBr I- . . . . . . . . NaI CH3COO- . . . . CH3COONa Cit3- . . . . . . Citricacid Sod2- . . . . . . Na2S04 . . . . . . Na3P04 Cd2+ . . . . . . CdS0, As3+ . . . . . . As203 v5 + . . . . . . NHiVOi U6'- . . . . . . U02(CH3C00)2 * Stripped with 0.5 M hydrochloric acid. t Stripped with 0.2 M sodium acetate solution. $ Washed with 0.2 M sodium chloride solution. mg 25 30 30 30 30 40 40 40 30 30 40 30 25 35 20 25 * 40 40.1 8$ 10* 5* 50 20 30 50 50 50 30 solutions. The stripping was complete with 0.2 N sulphuric acid. The volume of sulphuric acid required was found to be low and elution was complete even with 5-8 ml of 0.2 N sulphuric acid.The metal content was determined. photometrically by the hydrogen peroxide method,14 as the stripping agent will not interfere in the method. Effect of Diverse Ions Titanium was extracted and separated in the presence of a large number of different ions (Table 4). Interference studies were made by measuring the absorbance of the extracted liquid ion exchanger phase and conditions were established for the removal of interfering ions from the organic phase by eluting with suitable solvents. The tolerance limit was set as the amount30 ANALYST, JANUARY 1984, VOL. 109 Table 5. Determination of titanium in the standards and environmental samples Sample Certified value, ‘h Cast iron . . . . . . . . . . . . . . . . 0.102 Silicon - aluminium alloy . . . . . .. . . . 0.210 Ferrotitanium . . . . . . . . . . . . . . 40.00 Soil samples - Nandesari soil sample - Nandesari plant sample (Dhatoora) . . . . . . - Pulp and paper unit I t Chrome sample - . . . . . . . . . . . . . . . . . . . . . . . . . . . . Plants - . . . . . . . . . . . . . . . . . . - . . . . . . . . . . . . Paint and pigment unit I$ . . . . . . . . . . - Chrome industrial waste - Monozite sand - . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . * The values are averages of 10 determinations. t Samples from Nepanagar paper mill. $ Sample from GIDC area of Vapi. Obtained by present method* 0.10% 0 * 22% 40.0”” 10.0 p.p.m. 6.5 mg 28.5 p.p.m. 29.8 mg 62.3 p.p.m. 168.0 p.p.m. 13.0 mg 10.3 mg 25.0 mg Obtained by AAS 0. 1 O/O 0.21% 40.1 Yo 9.5 p.p.m.6.3 mg 29.2p.p.m. 29.5 mg 61.5 p.p.m. 170.0 p.p.m. 13.1 mg 10.2 mg 24.9 mg of foreign ion causing a change of k0.02 absorbance unit or k2% error in the recovery of titanium. Moderate amounts of various metal ions commonly associated with titanium were tolerated and also most of the common anions. Fluoride interfered seriously. Lower concentrations of vanadium (<40 pg) did not interfere. The vanadium, extracted together with titanium, can be stripped first by shaking the organic phase with 10 ml of 0.2 M sodium acetate solution for 5 min. Titanium was subsequently stripped with 10 ml of 2 N sulphuric acid. The interference of molybdenum(1V) was removed either by precipitation with sulphide prior to the extraction or the extracted molybdenum(1V) was eluted from the organic phase by washing the organic phase with 5ml of 0 .2 ~ sodium chloride solution. Moderate amounts of chromium, iron and zirconium were tolerated in the presence of tin(I1) chloride and any portion of the metals extracted into the organic phase can be removed by shaking first with 5 ml of 0.5 M hydrochloric acid for 5 min before the separation of titanium. Uranium also can be separated from titanium by stripping with 5 ml of 0.5 M hydrochloric acid and titanium can be recovered with 10 ml of 2 N sulphuric acid. Results for the determination of titanium in steel and environmental samples are given in Table 5. Conclusion The proposed method is suitable for the pre-concentration, separation and simultaneous micro-determination of titanium in the chemical form, alloys and steels and in environmental samples. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. References Onishi, H., and Toite, Y., Bunseki Kagaku, 1965, 14, 1141. Cerrai, E., and Testa, C., Anal. Chim. Acta, 1962, 26, 204. Moore, F. L., Anal. Chem., 1957, 29, 1660. El Yamani, I. S., Farah, M. Y., El Aleim, F. A. Abd, Talanta 1978,25, 523. Pinacva, S. G., Pyatinitskii, I. V., and Demutskaya, L. N., Ukr. Khim. Zh., 1978,44, 978. Shukla, J. P., and Subramanian, M. S . , J. Radioanal. Chem., 1979, 50, 159. Agrawal, Y. K., Rev. Anal. Chem., 1980, 5, 3. Agrawal, Y. K., and Patel, S. A., Rev. Anal. Chem., 1980,4, 237. Bag, S. P., and Khastagir, A. K., J. Zndian Chem. SOC., 1978, 55, 74. Agrawal, Y. K., Chem. Era, 1975, 11, 21. Hauser, E. R., and Renfrow, W. B., Jr., “Organic Synthesis,” John Wiley, New York, 1944, p. 7. Afghan, B. K., Marryati, R. G., and Ryan, D. E., Anal. Chim. Acta, 1968, 41, 131. “Methods for Chemical Analysis of Water and Wastes,” EPA-600-4-79-020. US Environmental Protection Agency, Ohio, 1979. Sandell, E. B. , “Colorimetric Determination of Traces of Metals,” Interscience, New York, 1965, p. 876. Sandell, E. B . , “Colorimetric Determination of Traces of Metals,” Interscience, New York, 1965, p. 870. Ringbom, A., Z . Anal. Chem., 1949, 21,332. Branko Tomazic, B., and Jerome O’Laughlin, W., Anal. Chem., 1973,45, 1519. Harvey, A. E., and Manning, D. L., J. Am. Chem. SOC., 1950, 72, 4488. Paper A3122 Received January 21st, 1983 Accepted August 1 Oth, 1983

ISSN:0003-2654    

DOI:10.1039/AN9840900027

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST. JANUARY 1984. VOL. 100 31 Solid - Liquid Separation after Liquid - Liquid Extraction: Spectrophotometric Determination of Iron(l1) by Extraction of its Ternary Complex with 2,2’-Dipyridyl and Tetraphenylborate into Molten Naphthalene Masatada Satake and Toru Nagahiro Faculty of Engineering, Fukui University, Fukui 9 10, Japan Bal Krishan Puri* Department of Chemistry, Indian Institute of Technology, Hauz Khas, New Delhi-170016, India A method for the spectrophotometric determination of iron(l1) after extraction of its ternary complex with 2,2’-dipyridyl and tetraphenylborate into molten naphthalene has been developed. Iron(ll) reacts with 2,2’-dipyridyl to form a water-solu ble coloured complex. This complex cation forms a water-insoluble stable ternary complex in the presence of sodium tetraphenylborate, which is easily extracted into molten naphthalene in the pH range 2.8-7.6 by vigorous shaking for a few seconds. The solid naphthalene containing the iron 2,2’-dipyridyl tetraphenyl borate complex is separated by filtration and dissolved in acetonitrile.The absorbance is measured at 521 nm against a reagent blank. Beer‘s law is obeyed in the concentration range 2.2-65.5 pg of iron in 10 ml of acetonitrile solution. The molar absorptivity and Sandell’s sensitivity are 8.89 x lo3 I mol-1 cm-1 and 0.0063 pg cm-2 at 521 nm, respectively. The interference of various ions has been studied in detail. The method has been applied to the determination of iron in standard metallic samples and the results are compared with those obtained by the 1 ,lo-phenanthroline method.Keywords: Solid - liquid separation; spectrophotometry; iron determination; metal and alloy analysis 2,2‘-Dipyridyl has been widely employed in the detection and determination of certain metal ions. 1-3 It reacts with iron(I1) to form a water-soluble, deep red complex, Fe(di~y)~z+, which forms water-insoluble, stable ternary complexes with anions such as C104-, SCN-, Hg142-, HgBr42- and Cd142-. As a result, some metal ions and anions can be extracted into suitable organic solvents as ternary complexes.&6 This further increases the sensitivity and an improvement in selectivity can be achieved. Since 1969, we have developed a method7-11 for the spectrophotometric determination of trace metals using naph- thalene as an extractant. In this study, the technique was used for the extraction of iron(II), which forms a stable ternary complex with these complexing reagents at high temperatures. We have observed that the complex cation Fe(di~y)~2+ reacts with the tetraphenylborate anion to form a very stable, water-insoluble, coloured ternary complex.This ternary complex cannot be extracted into most organic solvents, such as benzene, toluene, xylene, dichloroethane, nitrobenzene, isoamyl alcohol, isobutyl methyl ketone and isoamyl acetate because of the low solubility of the complex in these solvents at room temperature. However, the complex can be quantit- atively extracted merely by contact with molten naphthalene or by vigorous shaking for a few seconds. The interference of various ions has been studied. The method has been applied to the determination of iron in alloys and metals with extraction of the ternary complex into naphthalene, and the results were compared with those obtained by the 1 ,lo-phenanthroline method.12 The main advantage of the present method is that equilib- rium distribution of the complex in the two phases is attained in a few seconds owing to the high temperature. As only a small amount of the organic solvent (1.2 g) is required for the complete extraction of the metal ion using about 50 ml of the aqueous phase, the sensitivity of the method is enhanced as the whole phase may be taken for the analysis.Ordinary * Author to whom correspondence should be addressed. non-aqueous organic solvents have more or less the same solubility in water, which causes errors in the determination of metals owing to the change in the volume of the organic to the aqueous phase, whereas naphthalene is a solid crystal, immiscible with water and completely separated from the aqueous phase at room temperature.Hence the proposed method is the most suitable for the extraction of the complex without encountering difficulties in phase separation. The iron( 11) dipyridyl complexes, which are produced with colour- less inorganic anions as described above, are unstable and dissociated in aqueous solution at high temperature, but in the presence of the organic tetraphenylborate anion and complex formed is very stable and hardly dissociated. Hence the procedure can be successfully applied to the extractive spectrophotometric determination of iron.Experimental Equipment A Hitachi, Model 200-20, double-beam spectrophotometer was used for the absorbance measurements. pH measure- ments were made with a Toa-Dempa HM-SA pH meter, equipped with a combined calomel and glass electrode assembly. Reagents Doubly distilled water and analytical-reagent grade acids and salts were used throughout unless stated otherwise. Standard iron(ZII) solution, 5 p.p.m. Dilute 5ml of 1000 p.p.m. standard iron(II1) solution to 1 000 ml with water. 2,2’-Dipyridyl solution, 0.4%. Dissolve 0.4 g of 2,2’- dipyridyl in a small volume of concentrated hydrochloric acid and dilute with distilled water to 100 ml in a calibrated flask. Acetic acid - ammonium acetate buffer. Mix 1 M acetic acid and 1 M ammonium acetate solution in suitable amounts to give buffer solutions in the pH range 3-6.32 ANALYST, JANUARY 1984, VOL.109 Ammonia - ammonium acetate buffer. Mix 1 M ammonia solution and 1~ ammonium acetate solution in suitable amounts to give buffer solutions in the pH range 8-11. Naphthalene. Check the purity spectrophotometrically before use in the range 200-700 nm. Acetonitrile. Check the purity spectrophotometrically before use in the range 200-700 nm. Sodium tetraphenylborate solution, 1 Yo. Hydroxylammonium chloride solution, 2%. General Procedure To an aliquot of iron solution in a beaker add 2.0 ml of 2% hydroxylammonium chloride, 2.5 ml of 0.4% 2,2'-dipyridyl solution and adjust the pH to 4.5 with 2.0 ml of acetate buffer solution and dilute aqueous ammonia.Transfer the solution into a 100-ml round-bottomed flask, add 2.0 ml of 1% sodium tetraphenylborate solution and allow the mixture to stand for 1-2 min. The flask was heated to about 60 "C on a water-bath. Add 1.2g of naphthalene, stopper the flask and continue to heat until the naphthalene melts. Remove the flask from the water-bath and shake it vigorously until naphthalene solidifies into very fine crystals. Separate the naphthalene from the aqueous phase by filtration through a small glass filter-funnel and dry it in an oven. Dissolve the naphthalene in acetonitrile and dilute it to lOml with acetonitrile in a calibrated flask. Place a portion of this solution in a 1-cm cell and measure the absorbance at 521 nm against a reagent blank. Results and Discussion Absorption Spectra The absorption spectra of the reagent and ternary iron(I1) complex in naphthalene - acetonitrile solution were measured against water (Fig.1). The iron(I1) - 2,2'-dipyridyl tetra- phenylborate complex shows maximum absorption in the range 518-523 nm, where the reagent does not absorb. All absorbance measurements were made at 521 nm in subsequent studies, Effect of pH The effect of pH on the extraction of the complex was studied at 521 nm with a sample containing 30 yg of iron, keeping the other conditions constant. A graph of absorbance against pH (Fig. 2) showed that the absorbance of the complex is dependent on pH, the maximum absorbance being obtained between pH 2.8 and 7.6. A pH of 4.5 was chosen for all subsequent measurements.Effect of Amount of Reducing Agent Various amounts of hydroxylammonium chloride were added to a solution containing 301.18 of iron(II1) and the general procedure was followed. Iron(II1) was reduced to iron(I1) quantitatively when 0.5-12 ml of 2% hydroxylammonium chloride solution were used. Therefore, 2.5 ml of 2% solution were used in all subsequent work. Effect of Reagent Concentration Extractions of the complex were carried out at fixed pH with 2.0 ml of 1% tetraphenylborate solution, and with the addition of various volumes of 0.4% 2,2'-dipyridyl solution (Fig. 3). The extractions were quantitative when more than 0.3 ml of 0.4% solution was used. Effect of Tetraphenylborate Concentration Various volumes of 1 % tetraphenylborate solution were added to a sample solution containing 30 pg of iron and 2.5 ml of 0.4% 2,2'-dipyridyl solution at pH 4.5, the general ' 0.: I 1 1 420 460 500 540 580 Wavelengt h/n rn Fig.1. Absorption spectra of 2,2'-dipyridyl - tetraphenylborate and Fe(I1) - 2,2'-dipyridyl- tetraphenylborate in naphthalene - acetonitrile solution. Fe(III), 30 pg; pH, 4.5; 2% NH20H.HCI, 2.0ml; 0.4% 2,2'-di yridyl, 2.5 ml; 1% tetraphenylborate, 2.0ml; naphthalene, 1.2 g; sgaking time, 30 s; reference, water o'6 ' I f I I I I 1 0 2 4 6 8 PH Fig. 2. conditions as in Fig. 1 Effect of pH on extraction. Wavelength. 521nm: other I I I 1 6 1 2 3 4 5 6 0.4% D i p y r i d y I /m I Fig. 3. in Fig. 2 Effect of reagent concentration on extraction. Conditions as 1 2 3 4 1 2 3 4 5 0.1% Kalibor/rnl 1% Kalibor/rnl Fig.4. Conditions as in Fig. 2 Effect of tetraphenylborate concentration on extraction.ANALYST, JANUARY 1984, VOL. 109 33 0.6 I 1 0.2 ' I I I I 20 40 60 80 100 Volume of aqueous phase/ml Fig. 5. Effect of aqueous phase volume on extraction. Conditions as in Fig. 2 procedure being followed. Extraction was complete when the volume of 1% tetraphenylborate solution was in the range 0.1-5.0ml (Fig. 4). Therefore, 2.0 ml of 1% tetraphenyl- borate solution were used in subsequent work. Effect of Buffer Solution It was found that the addition of 0.5-5.0ml of acetate buffer solution caused virtually no variation in the absorbance. Hence in all experiments 2.0 ml of buffer solution were used. Effect of Addition of Naphthalene The amounts of naphthalene were varied from 0.4 to 2.0 g and the extractions were carried out by the general procedure.The extraction was quantitative in this range, but above 2.0 g it was difficult to dissolve the naphthalene in the limited amount of acetonitrile (10ml). Hence 1.2g of naphthalene was used as the most suitable amount. Effect of Volume of Aqueous Phase As the volume of the organic phase is small compared with that of the aqueous phase, it was essential to study the effect of the volume of the aqueous phase on the extraction. When the latter was varied between 10 and 100m1, the absorbance remained constant up to a volume of 55ml. Above this volume, the extraction was not quantitative (Fig. 5). Effect of Shaking Time The extraction of the complex into molten naphthalene (81-85OC) was found to be very rapid and no change was observed when the shaking time was varied from 3 to 120s. Effect of Standing Time after Addition of Reagents To study the effect of standing time after addition of the reagents, the iron(I1) - 2,2'-dipyridyl tetraphenylborate com- plex in a solution containing 30 pg of iron was allowed to stand at room temperature for 1-30min and then extracted into molten naphthalene. The rate of formation of the complex was very fast even at room temperature and a standing time of 1-2 min before the extraction was sufficient.Effect of Standing Time A mixture of the ternary complex and naphthalene was dissolved in acetonitrile and the effect of standing time on the absorbance was investigated. The ternary complex in naphthalene - acetonitrile was stable for up to 10 d.Beer's Law and Sensitivity Using the optimum conditions described above, a calibration graph for iron was constructed at 521 nm and was found to be Table 1. Tolerance limits for diverse ions and salts. Iron, 30 pg; pH, 4.5; naphthalene, 1.2 g Tolerance Tolerance Ion Cr6+ . . . . . . . . co2+ . . . . . . Zn2+ . . . . . . Ni2+ . . . . . . . . cu+ . . . . . . . . Cd2+ . . . . . . Sn2+ . . . . . . . . Bi3+ . . . . . . . . Pb2+ . . . . . . . . Mg2+ . . . . . . Hg2+ . . . . . . limit/pg 5 000 2 000 8 000 100 300 800 400 100 8 000 150 3 000 Salt limit/mg NaI . . . . . . 7000 K C I O ~ . . . . . . 300 KN03 . . . . . . 3000 CH,COONa.3H20 4 000 KSCN . . . . . . 1000 Na2S04 . . . . 2000 NaCl . . . . . . 1000 Na2C204 . . . . 1 KH2P04 . . . . 400 KCN .. . . . . 1 Na2tartrate.2H20. . 100 Na3citrate.2H,0 . . 5 linear over the concentration range 2.2-65.5 pg of iron in 10 ml of acetonitrile. The molar absorptivity and Sandell's sensi- tivity were calculated to be 8.89 X 103 1 mol-1 cm-* and 0.0063 pg cm-2 of iron, respectively (for an absorbance of 0.001) at 521 nm. Ten replicate analyses of a sample solution containing 30 yg of iron gave a mean absorbance of 0.477 with a standard deviation of 0.0029 and a relative standard deviation of 0.6%. Choice of Solvent Various organic solvents for dissolving the mixture of the ternary complex and naphthalene were tried. The complex is soluble in dimethylformamide, dimethyl sulphoxide and acetonitrile, but insoluble in dioxane, chloroform, benzene, toluene, xylene , chlorobenzene, o-dichlorobenzene , isoamyl acetate, isobutyl methyl ketone, carbon tetrachloride, nitro- benzene, 1,2-dichloroethane, diethyl ether, hexane, amyl alcohol, isoamyl alcohol, butan-1-01 and ethyl acetate.It is unstable in dimethylformamide and the resulting solution is unsuitable for absorbance measurements. Hence acetonitrile was the most suitable solvent for dissolving the complex. Effect of Diverse Ions Sample solutions containing 30 yg of iron and various amounts of different alkali metal salts or metal ions were prepared and the determination of iron was studied. The pH of the solution was adjusted to 4.5 and the general procedure was applied. The tolerance limits for diverse ions are given in Table 1. Obviously the method is fairly selective. Determination of Iron in Standard Materials and Metal Alloy Samples The procedure has been applied successfully to the determina- tion of iron in standard materials and real samples.The results are in reasonably good agreement with those obtained by the conventional spectrophotometric procedure using 1 , 10- phenanthroline12 (Table 2). Silicon - bronze alloy (NBS, SRM-158a) A 100-mg amount of this alloy was dissolved in a mixture of 10 ml of hydrochloric acid (1 + 1), 1 ml of concentrated nitric acid, 3ml of concentrated sulphuric acid and l m l of 30% hydrogen peroxide. The mixture was then gently heated on a boiling water-bath. The excess of the acids was evaporated and the volume of the sample was adjusted to 200ml with water. A 3.0-ml aliquot of this sample was placed in a separating funnel and 40 ml of hydrochloric acid (1 + 1) were added.The iron in this sample was extracted by vigorous shaking for 5 min with 20 ml of isobutyl methyl ketone. It was34 ANALYST JANUARY 1984, VOL. 109 Conclusion Table 2. Determination of iron in standard materials and real samples Sample SRM-158a silicon - bronze alloy NBS NBS SRM-85b aluminium alloy . . . aluminium Metallic Iron can be determined spectrophotometrically using 2,2‘- dipyridyl but many metal ions interfere and the sensitivity is low. Liquid - liquid extraction can improve the selectivity but it cannot be applied in the present instance as the ternary complex has low solubility in almost all organic solvents or the interfacial phase separation is not clear. The proposed technique has not only solved this problem, but has also enhanced the sensitivity of the method because only 1.2 g of naphthalene is required for the complete extraction of the metal ion.Metal ions that form thermally unstable complexes can interfere in liquid - liquid extraction spectrophotometric methods, but have no effect in the proposed technique, Concentration ,lO-phenanth- Certified proposed method, rolirie method, Concentration obtained by Obtained by value, % % * /O o * 1.23 1.25 f. 0.02 1.24 k 0.02 0.24 0.233 f. 0.002 0.239 k0.002 resulting in better selectivity. (powder) . . - 0.133 kO.001 0.137 kO.001 CoS04.7H20 - 0.004 3 f 0.000 2 0.004 5 rt: 0.000 1 * Average for five individual samples. then back-extracted from the organic phase with 25ml of water and then the proposed method or the 1,lO- phenanthroline procedure was applied.Standard aluminium alloy (NBS, SRM-85b) or metallic aluminium powder A 1.0-g amount of the aluminium alloy or 2.0 g of aluminium powder were accurately weighed and then placed into a 100-ml beaker t o which were added 40-50ml of hydrochloric acid (1 + 1) and 3.0 ml of 30% hydrogen peroxide. On heating on a boiling water-bath, the mixture dissolved completely and the excess of hydrogen peroxide decomposed. After cooling, the sample was diluted to 200 ml with water in a calibrated flask. An aliquot of this sample solution was taken and the proposed procedure was followed. Analysis of Cobalt Salt A salt (CoS04.7H20) was dissolved in water and then analysed as described above for the standard material sample. References 1. Welcher, F. J . , “Organic Analytical Reagents,” Volume 111, Van Nostrand, London, 1964, p: 68. 2. Rao, A. L. J., and Puri, B. K., Zh. Anal. Khim., 1973,28,183. 3. Schilt, A. A., “Analytical Applications of 1 ,lO-Phenanthroline and Related Compounds,” Pergamon Press, Oxford, 1969. 4. Poluektov, N. S., and Nazarenko, V. A., J . Appl. Chem. U.S.S.R., 1937, 10, 2103. 5. Yamamoto, Y., and Kotsuji, K., Bull. Chem. Soc. Jpn., 1964, 37, 594. 6. Kotsuji, K., Yoshimura, Y., and Ueda, S., Anal. Chim. Acta, 1968, 42, 225. 7. Fujinaga, T., Satake, M., and Yonekubo, T., Bunseki Kagaku, 1970, 19, 216. 8. Fujinaga, T . , Satake, M., and Yonekubo, T., Bull. Chem. SOC. Jpn., 1973,46,2090. 9. Satake, M., and Yamauchi, T., Mem. Fac. Eng. Fukui Univ., 1977, 25, 107. 10. Satake, M., Anal. Chim. Acta, 1977, 92, 423. 11. Puri, B. K., and Gautam, M., Mikrochim. Acta, 1979, I, 515. 12. American Public Health Association, American Water Works Association and Water Pollution Control Federation, “Stan- dard Methods for the Examination of Water and Waste Water,” Thirteenth Edition, American Public Health Associa- tion, New York, 1971, p. 189. Paper A311 22 Received May 5th, 1983 Accepted August Ist, 1983

ISSN:0003-2654    

DOI:10.1039/AN9840900031

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, JANUARY 1984. VOL. 109 35 Study of the Formation of Vanadium(1V) Complexes with Some Triphenylmethane Reagents and Cationic Surfactants Maciej Jarosz and Zygmunt Marczenko Department of Analytical Chemistry, Technical University, ul. Noakowskiego 3, 00-664 Warsaw, Poland The optimum conditions for the formation of ternary complexes of vanadium(1V) with Eriochrome Cyanine R (ECR), Chrome Azurol S (CAS) and Pyrocatechol Violet (PV) in the presence of zephiramine, cetyltrimethylam- monium (CTA) or cetylpyridinium (CP) ions have been determined. The complexes are formed a t pH 3.5-5.5 and suitable molar absorptivities are 5-8 x 1041 mol-1 cm-l. The higher sensitivities of the methods based on ternary systems are due t o the higher R : V(IV) molar ratios in ternary complexes (with cationic surfactants) than in binary complexes.Of the methods studied, that based on the system with ECR and CTA is proposed for the determination of vanadium as it has the highest sensitivity ( E = 7.9 X lo4 I mol-1 cm-1 a t 575 nm). This method has been applied t o the determination of vanadium in steel. Keywords: Vanadium determination; ternary complexes; spectrophotometry; triphen ylmethane reagents; cation ic surfactan ts Spectrophotometric methods for the determination of vanad- ium generally exhibit low sensitivity. 1 The molar absorptivi- ties of the most sensitive methods with triphenylmethane reagents do not exceed 3.7 x 1041 mol-1 cm-1.2 A significant increase in the sensitivity of the methods based on binary complexes of metals with chelating reagents (mainly tri- phenylmethane reagents) can be obtained by introducing a third component, a cationic surfactant, into the system."-" The mechanism of the action of cationic surfactants has been the subject of many publications. The effects observed after introducing cationic surfactants to the binary system, viz., a bathochromic shift of the absorption maximum of the ternary complex relative to that of the binary complex and an increase in molar absorptivity, were attributed to the inter- action of the cationic surfactants with anionic forms of the reagent3.6.7 or to solubilisation of the binary complexes with micelles of the cationic surfactants.83' Savvin et al.9 introduced the term hydrophobic interactions.This paper describes a study of the conditions of the formation of ternary vanadium complexes with Eriochrome Cyanine R (ECR), Chrome Azurol S (CAS), Pyrocatechol CH3 CH3 I Pyrocatechol Violet Eriochrome Cyanine R CH3 CH3 I 1 Chrome Azurol S Violet (PV) and the cationic surfactants cetyltrimethyl- ammonium bromide (CTA), cetylpyridinium chloride (CP) and tetradecyldimethylbenzylammonium chloride (zephiramine) in order to find the best system for the spectrophotometric determination of vanadium.The systems vanadium(1V) - CAS - zephiramine/CTA have been described elsewhere. 10~1 Experimental Apparatus The absorbances were measured using a VSU2-P spectropho- tometer land the absorption spectra were recorded on a Specord ultraviolet - visible spectrophotometer. In all meas- urements 1-cm cells were used.The pH measurements were carried out using an ELPO N-517 pH meter. Reagents Vanadium(V) standard solution, 1 mg ml-1. A 2.295 0-g amount of ammonium vanadate (NH,V03) was dissolved in water and 5 ml of concentrated ammonia solution were added. The solution was acidified with 10ml of concentrated hydrochloric acid and diluted to 1 1 with water in a calibrated flask. Working solutions of vanadium(1V) were prepared by taking suitable aliquots and reducing the vanadium(V) by means of 5% ascorbic acid solution. Eriochrome Cyanine R (Loba), Chrome Azurol S (BDH) and Pyrocatechol Violet (POCh). The reagents were purified as described elsewhere. 12 Aqueous solutions of the reagents were used, of concentrations 1 x l o - 3 ~ ECR or CAS and 2 x 10-3 M PV. Cetyltrimethylammonium bromide (International Enzy- mes), cetylpyridinium chloride (Loba) and zephiramine (ICN).The commercial preparations do not require further purification. The results of elemental analyses (C, H, halogen) agreed with the theoretical values. Aqueous solutions of these reagents were used of concentration 1 X 10-2 M. Procedure for Determining Vanadium(1V) with ECR and CTA Pour a solution containing 2-16yg of vanadium(1V) into a 25-ml beaker. Add 3 ml of the CTA solution and 3 ml of the ECR solution. Adjust the pH to 4.9k0.3 using ammonia solution. Transfer the solution into a 25-ml calibrated flask and dilute to the mark with water. After lOmin, measure the absorbance at 575 nm against a reagent blank. Results Preliminary studies have indicated that vanadium(V) , unlike vanadium(IV), does not form ternary complexes with the reagents considered here.36 ANALYST, JANUARY 1984, VOL.109 Vanadium(1V) - Eriochrome Cyanine R - Cationic Sur- factant Systems Effects of acidity and excess of reagents The effect of acidity on the measured absorbances was examined in the pH range 3-6. The optimum pH ranges for the formation of the ternary complexes were 4.9k0.3, 4.5 kO.1 and 5.2 k 0 . 2 for CTA, CP and zephiramine, respectively. Maximum absorbances were obtained when the excess of ECR with respect to vanadium was higher than 10. In the system with CP the absorbance increased at a higher excess of ECR (Fig. 1). A hypsochromic shift (ca. 10nm) of the absorption maximum in the ternary system was observed as the excess of ECR increased.The optimum molar excesses of the cationic surfactants with respect to vanadium were 75-200,75-125 and 75-150 for CTA, zephiramine and CP, respectively. At a low excess of zephiramine (below the turbidity ranget3) a significant bathochromic shift (from 550 to 620nm) of the absorption maximum of the formed complex with respect to the position of A,,,, of the binary complex V-ECR, and also a hyperchromic effect, were observed (Fig. 2). At an excess of the cationic surfactants above the turbidity range the A,,,,, of the ternary complex exhibited a hypsochromic shift to about 580 nm and the measured absorbances attained maximum values. Similar changes in the absorption spectra were observed when CTA and CP were used. The complexes of vanadium(1V) with ECR and zephir- amine/CTA were stable for at least 30min after the pH had been adjusted to a suitable value.With CP the absorbance decreased after 5 min. Composition of the ternary complexes The molar ratio of ECR to vanadium(1V) in the ternary complexes, determined by the method of isomolar series (Fig. 3), was close to 1 : 1.5 (when zephiramine was used as the third component) or 1 : 2 (with CTA or CP). The complexes formed are weak, as indicated by the lack of sharp maxima on the curves. The molar ratio of vanadium to the cationic surfactant in the ternary complexes leading to the maximum absorbances cannot be determined because a large excess of the cationic surfactant is necessary for the formation of these complexes. Spectrophotometric determination of vanadium (IV) Ternary complexes of vanadium with ECR and cationic surfactants can form a basis for its spectrophotometric determination.Calibration graphs obtained under optimum 0.8 a, 0.6 0 S m -F a 0.4 0.2 [Vl : [ECRI Fig. 1. Effect of the excess of ECR with respect to V(1V) on the systems V - ECR - CTA/CP/zephiramine. cv = 1 X M; C c r A = ccp = c,, h = 1 X 1 0 - 3 ~ . Ternary systems (absorbances measured against byank): 1, with CTA (A = 575 nm); 2 , with CP ( h = 575 nm); and 3, with zephiramine (A = 580 nm) conditions obeyed Beer's law up to a vanadium(1V) concen- tration of 0.7 pg ml-1 when CTA and zephiramine were used and in the range 0.1-0.5 pg ml-1 when CP was used. The molar absorptivities were 7.9 x 1041 mol-1 cm-1 at 575 nm, 7.7 X lo4 1 mol-1 cm-1 at 580 nm and 7.1 x 1041 mol-1 cm-1 at 575 nm for CTA, zephiramine and CP, respectively.The precision of the most sensitive method, based on the system V(1V) - ECR - CTA, is indicated in Table 1. The method is not selective. Numerous metals that readily hydrolyse in slightly acidic media interfere. Hence, before the determination, vanadium should be separated from other metals, e.g., by using highly selective extraction with N-benzoyl-N-phenylhydroxylamine (BPHA).' The determination of vanadium in steel was carried out as follows. A small amount of steel (ca. 20 mg) was dissolved by heating with 10ml of concentrated hydrochloric acid plus a few drops of concentrated nitric acid. The solution obtained was evaporated almost to dryness and then diluted with 5h1 hydrochloric acid to 20ml and shaken with two successive portions of 0.1% BPHA solution in chloroform.The combined extracts were evaporated to dryness and the residue was mineralised with a mixture of concentrated sulphuric and nitric acids. After dilution of the solution, vanadium was reduced to vanadium(1V) using ascorbic acid and determined by means of ECR and CTA [see Procedure for Determining Vanadium( IV) with ECR and CTA]. The content of vanadium in the steel sample was (2.70 2 0.15) x lo-*% (seven determinations). The recovery of vanadium was 91% (the amount of vanadium added was 5 pg). Vanadium(1V) - Chrome Azurol S - Cationic Surfactant Systems Conditions of complex formation The optimum pH ranges for the formation of the ternary complexes of vanadium(1V) with CAS and the cationic surfactants were 4.2 k 0.2, 4.3 k 0.3 and 4.5 k 0.2 for CP, CTA and zephiramine, respectively.Maximum absorbances in the presence of CTA were obtained when the molar excess of CAS with respect to vanadium was higher than 10. With CP the absorbance increased continuously with increasing excess of CAS. When zephiramine was used a maximum was observed at a molar excess of CAS of 8-10. An increasing excess of CAS with respect to vanadium resulted in a hypsochromic shift of h,,,, of the ternary complex (cu. 5nm). 0.8 0.6 (r, c m a 0.4 2 0.2 5 500 560 580 600 Wavelengthhm 700 Fig. 2. Effect of the excess of zephiramine with respect to V(1V) on. absorption spectra in the system V - ECR - zeph. cv = 1 x 10-5 M ; cECR = 1.2 x 10-4 M . Molar excess of zephiramine: 1, without zephi- ramine; 2, 1 : 1 ; 3 , 1 : 3; 4, 1 : 50; 5 , 1 : 100; and 6, 1 : 200ANALYST, JANUARY 1984, VOL.109 37 0.6 cu 0.4 C m +2 z n a 0.2 I 1 I 1 1 : l 1 : 2 I I I I Molar ratio V : ECR Fig. 3. Determination of the molar ratio of V(IV) to ECR in the ternary complexes with CTA/CP!zephiramine by means of the method of continuous variations. cv + cECR = 2.5 x M; cnA = ccp = C,,ph = 1 x 1 0 - 3 ~ . 1, With CTA (h = 575 nm); 2, with CP ( h = 575 nm); and 3, with zephiramine (A = 580 nm) The optimum molar excesses of the cationic surfactants were 75-125 for CTA, 150-200 for CP and 100-175 for zephiramine. When CTA or zephiramine was used the absorbance increased up to a maximum value with increasing excess of the surfactant (above the turbidity range). An increase in the excess of CP up to 150-fold resulted in a linear decrease in absorbance.The absorption spectra of the V(1V) - CAS - cationic surfact- ant ternary complexes with various excesses of the cationic surfactants showed similar courses to those of the complexes with ECR. The V(1V) - CAS - CTA/zephiramine ternary complexes were stable for at least 30min. When CP was used the maximum absorbance decreased after 5 min. Composition of the ternary complexes The molar ratios of CAS to vanadium in the ternary complexes, determined by the method of isomolar series (Fig. 4), were approximately 1 : 1 , 2 : 1 and more than 2 : 1 for CTA, zephiramine and CP, respectively. Spectrophotometric determination of vanadium(IV) The molar absorptivities for the methods of vanadium determination based on the ternary complexes involving CAS and cationic surfactants were 7.1 x 104 1 mol-1 cm-1 at 595 nm for zephiramine, 6.0 x 1041 mol-1 cm-1 at 600 nm for CTA and 5.4 x 104 1 mol-* cm-1 at 600 nm for CP.The calibration graphs were straight lines for vanadium concentrations up to 0.7 pg ml-1 (zephiramine), up to 0.8 pg ml-1 (CTA) and in the range0.15-0.7 pgml-l (CP). Vanadium(1V) - Pyrocatechol Violet - Cationic Surfactant Systems Conditions of complex formation Ternary complexes of vanadium( IV) with PV and cationic surfactants were formed in the pH ranges 3.6 k 0.2,4.0 k 0.3 and 4.3 k 0.3 for CP, zephiramine and CTA, respectively. Table 1. Precision data for the determination of vanadium(1V) Vanadiumlpg Standard Confidence 7.6 x 10-2 1.97 zk 0.07 2.0 1.97 8.0 8.01 4.9 x 10-2 8.01 f.0.05 16.0 16.20 1.08 x 10-I 16.20 k 0.10 Added Found deviation*/pg limits? * For seven determinations. t Probability level = 0.95. Maximum absorbances of the ternary complexes were obtained at molar excesses of PV of 10-20 (zephiramine), 15-20 (CP) and 20-25 (CTA). When zephiramine was used a constant absorbance was obtained at molar excess of PV higher than 10. When CP was used at least a 15-fold excess of PV was necessary. In the presence of CTA the optimum PV excess was 20-25. The absorption maximum did not shift with increasing excess of PV. The optimum molar excesses with respect to vanadium of the cationic surfactants for the formation of ternary complexes involving PV were to 80-200 (CP), 200-275 (zephiramine) and 100-175 (CTA).As the excess of CP or zephiramine (above the turbidity range) increased the absorbance increased up to a constant value. With an increasing excess of CTA the absorbance initially decreased and attained a constant value only at a 100-fold molar excess. Ternary complexes of vanadium(1V) with PV and cationic surfactants were not formed below the turbidity range. The absorbance of ternary complexes of vanadium(1V) with PV and CTA/zephiramine did not change for 30min. When CP was used the absorbance decreased after 5 min. Composition of the ternary complexes The PV:V(IV) molar ratio in the ternary complexes was determined by the method of isomolar series (Fig. 5). It was slightly higher than 2 when CTA was used and slightly lower than 3 when zephiramine or CP was used. The number of PV molecules per vanadium atom in the complexes was, on average, higher than in the systems with ECR and CAS.Spectrophotometric determinution of vanadium (IV) The sensitivity of the spectrophotometric determination of vanadium using PV and cationic surfactants depends strongly on the nature of the surfactant. For zephiramine E = 7.5 X lo4 at 660nm and for CP E = 5.0 x 1041 mol-'cm-1 at 660 nm. The determination ranges are up to 0.5 pg ml-1 of vanadium for zephiramine and 0.15-0.6 pg ml-1 of vanadium for CP. The system V(1V) - PV - CTA cannot be used for the spectro- photometry of vanadium because the calibration graph is not straight for a wide range of vanadium concentrations. Discussion and Conclusions Of the nine systems of vanadium(1V) with selected triphenyl- methane reagents and cationic surfactants examined as a basis for the spectrophotometric determination of vanadium(IV), that with ECR and CTA, which exhibit the highest sensitivity (molar absorptivity 7.9 x l o 4 1 mol-1 cm-I), can be recom- mended.The methods based on the V(1V)-ECR- zephiramine and V( IV) - PV - zephiramine systems have similar sensitivities. In the presence of cationic surfactants, spectrophotometric methods for determining vanadium with the chromophoric reagents ECR, CAS and PV are far p o r e sensitive than those with binary systems. Compared with molar absorptivities of 1.8-3.7 x l o 4 for binary systems,2 values of 7.1-7.9 X lo4 can be attained for ternary systems. An advantageous increase in Ah ( L a x .complex - L a x . reagent) is also observed.38 ANALYST. JANUARY 1984. VOL. 109 Molar ratio V : CAS Fig. 4. Determination of the molar ratio of V(1V) to CAS in the ternary complexes with CTAICPlzephiramine by means of the method of continuous variations. cv + cCAS = 2.5 x lops M ; CCTA = ccp = C,,ph = 1.5 x l o - 3 ~ . 1, With CTA (h = 600nm); 2, with CP (h = 600 nm); and 3, with zephiramine (h = 595 nm) The increased sensitivity results from the fact that in the binary systems the V(1V) : R molar ratio is 2 : 1-1 : 1,2 whereas in ternary systems the ratios, as shown in this work, are in the range 1 : 1-1 : 3. Undoubtedly cationic surfactants (and espe- cially their micellar forms) permit higher R : V(1V) molar ratios in ternary complexes.This effect is observed with high concentrations of cationic surfactants, above the turbidity range.13 The shifts of the absorption maxima that occurred with ternary complexes with ECR or CAS following an increase in the concentration of added cationic surfactants or chromo- phoric reagents are closely connected with the character of the interactions between the triphenylmethane reagent and the cationic surfactant. The large bathochromic shift of I.,,,, below the turbidity range is a result of electrostatic interac- tions between a positive charge localised on the nitrogen atom of the cationic surfactant and anionic forms of the triphenyl- methane reagent. These interactions are stronger when the localisation of the positive charge on the nitrogen atom is more compact.It depends on the size of the inductive effect from alkyl groups bonded with the nitrogen atom and also on steric hindrances (screening of the nitrogen atom by large substituents). The hypsochromic shift of the absorption maximum of the ternary complexes above the turbidity range (Fig.2) is caused by hydrophobic interactions9 of micellar forms of the cationic surfactants. Micelles of cationic surfactants are formed more readily when their molecules contain larger hydrophobic substituents. Hydrophobic interactions are hindered by rigid fragments of the molecules, e.g., the pyridine ring in CP. These interactions lead to an increased energy state of the electrons in conjugated bonds and increased reactivity of chromophoric reagents, which results in the formation of complexes with higher R : V(1V) molar ratios than below the turbidity range.In ternary complexes with PV, hydrophobic interactions play an important role. These complexes are formed only above the turbidity range, and the positions of their absorp- tion maxima do not change with increasing concentration of cationic surfactants. Their formation depends on the dissoci- ation of protons of the phenol groups of PV, which is made easier by micellar forms of cationic surfactants. The number of molecules of chromophoric reagents in these complexes is greater than in complexes with ECR or CAS. 1 : l 1.2 1 . 3 Molar ratio V: PV Fig. 5. Determination of the molar ratio of V(IV) to PV in the ternary complexes with CTA/CP/zephiramine by means of the method of continuous variations.cv + cpv = 2.5 x 10-5 M (CP, zeph) or 5 X 10-5 M (CTA). cmA = 2 x l o - 3 ~ ; ccp = 1.5 x 10-3 M; C,,ph = 2.5 X 10-3 M. 1, With CTA (h = 660 nm); 2, with CP (h = 660nm); and 3, with zephiramine (A = 660 nm) In the optimum pH ranges for the formation of the ternary complexes, vanadium(1V) can occur in two forms, VO(OH)+ or VO(OH)2.14 In solutions containing ECR, CAS or PV and cationic surfactants at these acidities doubly and triply ionised forms of these reagents15316 are in equilibrium. It has been found previously13 that the ternary complexes of aluminium with ECR, CAS and PV are formed when aluminium occurs as Al(OH)2+ or A1(OH)3 and the triply ionised form of the reagent (ECR, CAS, PV) predominates. Ternary complexes of vanadium(1V) and aluminium(II1) with the examined reagents are formed in the pH ranges in which these metals occur as univalent hydroxy complexes or hydroxides and the chromophoric reagent in doubly or triply ionised forms.1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. References Marczenko, Z., “Spectrophotometric Determination of Ele- ments,” Ellis Horwood, Chichester, 1976. Janssen, A., and Umland, F., Fresenius 2. Anal. Chem., 1971, 254,286. Tikhonov, V. N., Zh. Anal. Khim., 1977, 32, 1435. Savvin, S. B., Crit. Rev. Anal. Chem., 1979, 8, 5 5 . Marczenko, Z.,Crit. Rev. Anal. Chem., 1981, 11, 195. Chernova, R. K., Zh. Anal. Khim., 1977,32, 1477. Savvin, S. B., Chernova, R. K., Belousova, V. V., Sukhova, L. K., and Shtykov, S . N., Zh. Anal. Khim., 1978,33, 1473. Martynov, A . P., Novak, V. P., and Reznik, B. E., Ukr. Khim. Zh., 1978,44, 203. Savvin, S. B., Marov, I. N., Chernova, R. K., Shtykov, S. N., and Sokolov, A. B., Zh. Anal. Khim., 1981, 36, 850. Horiuchi, Y., and Nishida, H., Bunseki Kagaku, 1969,18,850. Gordeeva, I . N., and Mescherakova, D. N., Vest. Leningr. Tekhn. Univ., 1981, 4, 95. Langmyhr, F. J., and Klausen, K. S . , Anal. Chim. Acta, 1963, 29, 149. Marczenko, Z., and Jarosz, M., Analyst, 1982, 107, 1431. Nazarenko, V. A., Antonovich, V. P., and Nevskaa, E. M., “Metal Ions Hydrolysis in Dilute Solutions” (in Russian), Atomizdat, Moscow, 1979. BureSova, I., Kubaii, V., and Sommer, L., Collect. Czech. Chem. Commun., 1981,46, 1090. Skarydova, V., and cermakova, L . , Collect. Czech. Chem. Commun., 1982,47, 1310. Paper A31104 Received April Ilth, 1983 Accepted August 25th, 1983

ISSN:0003-2654    

DOI:10.1039/AN9840900035

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, JANUARY 1984, VOL. 109 39 Direct Complexometric Determination of Aluminium and Moderate to Low Amounts of Titanium and Iron Using Tartaric Acid as a De-masking Agent Samarendra Dasgupta and Birendra Chandra Sinha Analytical Chemistry Division, Central Glass and Ceramic Research Institute, Calcutta-700032, India Narendra Singh Rawat Department of Chemistry, Fuel and Mineral Engineering, Indian School of Mines, Dhanbad-826004, India A critical study has been made on the masking of titanium and de-masking of the Ti0 - EDTA (ethylenedi- aminetetraacetic acid) complex with tartaric acid. It reveals that tartaric acid de-masks Ti0 - EDTA selectively and quantitatively with the release of EDTA at pH 5.3 from a solution at 60-70 "C containing Ti0 - EDTA, Al - EDTA and Fe - EDTA complexes.The release of EDTA from Al - EDTA is, however, effected by the usual practice of de-masking with sodium fluoride or ammonium fluoride. Based on these observations, procedures are given for the direct complexometric determination of aluminium, titanium and iron present in the same analyte by stepwise indirect titration with EDTA, which does not involve prior separation of the metals. The method is, therefore, very simple, rapid and accurate and has been applied successfully to determine aluminium and moderate to low amounts of titanium and iron present in common aluminosilicate materials. Results for the determinations compare favourably with both the certified values and values obtained by other standard methods. Keywords: Aluminium titanium and iron determination; complexometry; eth ylenediaminetetraacetic acid; tartaric acid; aluminosilicate In common aluminosilicate materials, A1 is invariably asso- ciated with moderate to low amounts of Ti and Fe.In such a system combined A1 and Ti(A1 + Ti) and Fe are complexome- trically determined by a back-titration with ethylenediamine- tetraacetic acid (EDTA)lJ at pH 5.3. The A1 value is obtained indirectly from the total of (A1 +Ti) after deducting the Ti (equivalent to Al), determined separately by colorimetric or other methods. The limitations of these methods are as follows: the indirect determination of A1 from the total (A1 + Ti); the adverse effect of the intense yellow Fe(II1) - EDTA complex on the end-point of the titration when Fe is present in significant amounts; and the sluggish end-point caused by large amounts of Ti, although some workers3 have applied this method to the determination of Ti and A1 present in large amounts.Under such conditions, the separation of Ti and Fe is imperative and is fulfilled by a cupferron - chloroform extrac- tion4.5 Again, the complexometric determination of a large amount of Ti by an indirect titration with EDTA in the presence of H2026 requires the removal of both A1 and Fe from the system. Fe, however, offers no difficulty as it can be determined in such a mixture by direct titration with EDTA at pH 2-3 in the presence of sulphosalicylic acid.7 Pribil and Veselyg devised a method for the complexometric determina- tion of Ti, Fe and A1 when present together. Ti is separated as the hydroxide in the presence of triethanolamine, which masks the Fe and A1.Ti(OH)4 is then dissolved and Ti determined as Ti - H202 - EDTA at pH 1-2.9 From the filtrate A1 and Fe are determined by the usual back-titration procedure.However, this method has the disadvantage that Fe is also coprecipitated when present in large amounts. The method of Voinovitch et al. 10 for the complexometric determination of A1 in the presence of Ti and Fe with an excess of EDTA, using ZnC12 as the back-titrant (in the presence of tartaric acid, diammonium phosphate and fluoride) at pH 5-6 and with a dithizone indicator appears attractive. However, the absorption of the dithizone indicator, which occurs because of the precipitation of titanium phosphate (formed after phosphate addition) interferes with the end-point determination.Interference is also caused by tartaric acid, which tends to complex zinc, the back-titrant, and disturbs the end-point. Recently Tikhonovll recommended a procedure using only tartaric acid and zinc chloride in an acetate buffer (pH 5 . 9 , which was also tried initially in this work but was found to be completely unsatisfactory owing to similar interference from tartaric acid. Therefore, owing to the lack of suitable methods presented in the literature, this work has been aimed at providing a direct determination, by a stepwise indirect titration with EDTA for Ti, Fe and A1 when present together. The method developed does not involve any lengthy separation techniques. Experimental All reagents were of analytical-reagent grade.Standard EDTA solution, 0.025 M. Dissolve 9.30 g of the disodium salt of EDTA in water and dilute to 11. Standardise the solution by direct titration against standard zinc chloride solution (0.025 M) in a hexamine buffer solution (pH 5.3) with xylenol orange indicator. Lead nitrate solution, 0.025 M. Dissolve 8.28 g of lead nitrate [Pb(NO&] in water, acidified with a few drops of 8 N HN03 and dilute to 11 in a calibrated flask. Check the molarity of the solution by titration with standard EDTA solution in a hexamine buffer solution (pH 5.3) using xylenol orange as indicator. Iron(III) ammonium sulphate solution, 0.025 M. Dissolve, with warming, 6.0 g of iron(II1) ammonium sulphate [NH4Fe- (S04)2.12H20] in water containing 10 ml of 18 N HzS04.Cool and dilute to 500 ml in a calibrated flask. Check the molarity by titrating a hot (40-50 "C) solution against the standard EDTA solution at pH 2-3 using sulphosalicylic acid7 as the indicator. Titanium solution, 0.025 M. Weigh 1.0 g of titanium dioxide (Ti02) in a platinum basin. Add 20ml of 40% HF, 10 ml of 18 N H2S04, heat and evaporate the contents of the basin until copious fumes of SO3 are evolved. Cool, dissolve the titanium sulphate in 400 ml of 2 N H2SO4 on a steam-bath. Dilute the solution to 500ml in a calibrated flask with 2~ H2S04 and standardise the solution at pH 5.3 and at 30 k 5 "C in the40 ANALYST, JANUARY 1984, VOL. 109 presence of H2026 by back-titrating an added excess of EDTA with standard lead nitrate solution, using xylenol orange as indicator.Aluminium sulphate solution, 0.025 M. Weigh 5.93 g of potash alum [KAl(SO&. 12H20], dissolve in water, acidify with a few drops of 1 2 ~ HCl and dilute to 500ml in a calibrated flask. Standardise the solution at pH 5.3 by back-titrating an added excess of EDTA with standard lead nitrate solution using xylenol orange as indicator. Hexamine buffer solution, pH 5.3. Dissolve 20 g of hexam- ine in 100 ml of water and adjust the pH to 5.3 by adding 12 N HCl and checking with a pH meter. Tartrate solution. Dissolve 10 g of tartaric acid in 100 ml of water. Adjust the pH of the solution to 5.3 by adding 2 N NaOH solution and checking with a pH meter. Hydrogen peroxide, 30 volume. Hydrofluoric acid, 40%. Sulphuric acid, 18 N. Hydrochloric acid, 6 and 12 N.Nitric acid, 8 N. Potassium hydrogen sulphate, solid. Xylenol orange solution (slightly acidified) 0.2%. Ammonium fluoride, solid. Potassium hydroxide solution, 2 and 5 N. Preparation of Sample Solution Weigh 0.5g of a well ground (80-90pm) and dried (105- 110°C) sample in a platinum basin. Moisten with water and add 10 ml of 40% HF and 2 ml of 18 N H2SO4. Evaporate the contents on a sand-bath until copious fumes of SO3 are evolved and ultimately to dryness. Fuse the contents with 5 g of KHS04. Dissolve the melt with heating in water that contains 5ml of 1 2 ~ HCl. Cool the solution and dilute to 250 ml in a calibrated flask. Determination Procedure Pipette an aliquot (25-30ml) of the solution into a 250-ml conical flask and pipette an excess of EDTA (in at least a 5-ml excess over the stoicheiometric requirement for quantitative formation of the A1 - EDTA, T i 0 - EDTA and Fe - EDTA complexes) into the flask.Neutralise the solution with 2~ KOH solution in the presence of xylenol orange indicator when the solution turns slightly red. Acidify with a few drops of 1 2 ~ HC1 until the solution turns yellow. Add 20ml of hexamine buffer solution (pH 5.3), dilute to 100ml and boil the solution for 5 min. Cool to room temperature (30 k 5 "C), add 5 ml of hexamine buffer solution and titrate the standard lead nitrate solution to a sharp, red end-point. Add 20ml of tartrate solution (the colour of the solution turns yellow), boil for 5 min and titrate again with standard lead nitrate solution at 60-70 "C, the end-point being indicated by a sharp change of colour from yellow to red.Again add 2g of NH4F, boil the solution for 5min, cool to room temperature and similarly titrate the solution with lead nitrate solution. Perform a blank titration for the excess of EDTA added at the start against lead nitrate at the same pH and room temperature in the presence of xylenol orange indicator. The second and third titres correspond to Ti and Al, respectively, while the difference between the last and the sum of the first, second and third titres is a measure of Fe. by step from Al-EDTA and TiO-EDTA complexes. If, however, a complexing agent, which can preferentially de-mask either A1 or Ti from their EDTA complexes before the addition of fluoride, the problem is solved. Preliminary experiments with citric acid, oxalic acid, lactic acid and tartaric acid as de-masking agents have indicated tartaric acid to be the most promising.12 Experiments have been carried out by adding a known excess of EDTA to different concentrations of Ti solutions containing tartaric acid, boiling and back-titrating the excess of EDTA with lead nitrate solution at pH 5.3 using xylenol orange indicator and hexamine.Lead nitrate solution was selected as a back-titrant instead of zinc because the tartrate present in the solution complexes with zinc but not lead. The results for the lead nitrate solution back-titration (Table 1) correspond to the actual amount of EDTA added, indicating the complete masking of Ti with tartaric acid. Under the same conditions, experiments have also been performed to determine whether tartaric acid can de-mask Ti from its EDTA complex.A known excess of EDTA has been added in solutions containing different amounts of titanium. The solution is brought to pH 5.3 with hexamine, boiled and titrated at room temperature (30 k 5 "C) against lead nitrate solution using xylenol orange as indicator. Tartrate is then added, the solution is boiled and titrated with lead nitrate solution at the same pH and temperature to measure the release of EDTA. The lead nitrate solution titres for the second titration (Table 2) indicate a quantitative release of EDTA from the Table 1. Masking of titanium by tartrate against EDTA titration Lead nitrate (0.025 M) Ti (0.025 M) taken/ EDTA (0.025 M) for back-titration/ ml added/ml ml - 10.10 10.10 1.20 10.10 10.10 2.40 10.10 10.10 3.60 10.10 10.10 4.80 10.10 10.10 Table 2.De-masking of titanium from Ti0 - EDTA by tartrate EDTA (0.025 M) Ti (0.025 M) added in taken/ml excess/ml 1.20 10.10 2.40 10.10 3.60 10.10 4.80 10.10 Lead nitrate (0.025 M) for back- titration/ ml 8.88 7.65 6.50 5.33 Ti (0.025 M) obtained by difference/ ml 1.22 2.45 3.60 4.77 Ti (0.025 M) obtained from release of EDTA by tartrate/ ml 1.25 2.38 3.62 4.80 Results and Discussion In developing the method for the complexometric determina- tion of Al, Ti and Fe in aluminosilicates using a stepwise indirect titration with EDTA, the main problem has been how to release EDTA step by step from a mixture of A1 - EDTA, Ti0 - EDTA and Fe - EDTA. The usual addition of fluoride to release EDTA from A1 - EDTA does not release EDTA step c .- C 3.00 I I I I I I I (0 20 30 40 50 60 70 80 90 -I Titration temperature/"(= Fig.1. Effect of temperature on the quantitative release of EDTA by tartrate from Ti0 - EDTA in the presence of A1 - EDTA indicated by the lead nitrate titration of the released EDTA. The broken line represents the theoretical value for titaniumANALYST, JANUARY 1984, VOL. 109 41 Table 3. Determination of Al, Ti and Fe in synthetic solutions A1 taken/ mg 16.75 13.40 6.70 3.85 0.67 A1 found/ mg 16.73 13.42 6.70 3.84 0.69 Ti Devia- taken/ -0.02 1.44 +0.02 5.76 - 2.88 -0.01 4.32 +0.02 5.76 tion/mg mg Ti found/ mg 1.48 5.71 2.85 4.29 5.80 Fe Devia- taken/ +0.04 1.35 -0.05 4.05 -0.03 1.35 -0.03 4.05 +0.04 4.05 tiodmg mg Fe found/ Devia- mg tionlmg 1.32 -0.03 4.00 -0.05 1.40 +0.05 4.03 -0.02 4.05 - Table 4.Determination of A1203, Ti02 and Fe203 in some aluminosilicate materials Found, YO Mean, ‘/O Certified value, “/o Sample A1203 Bauxite . . . . 58.05 58.17 58.20 Flint clay (NBS 97) 38.76 38.70 38.82 Firebrick . . . . 41.15 41.13 41.22 Clay . . . . . . 34.50 34.58 34.45 Ti02 9.95 9.86 9.90 2.45 2.37 2.35 3.90 3.95 3.85 3.19 3.26 3.15 Fez03 A1203 Ti02 Fe203 A1203 Ti02 Fe203 1.90 58.14 9.90 1.89 58.25* 10.03? 1.87$ 1.82 1.95 1.05 38.76 2.39 0.97 38.808 2.409 0.980 0.90 0.95 3.60 41.17 3.90 3.65 41.05* 4.001 3.66$ 3.70 3.65 0.40 34.51 3.20 0.39 34.44* 3.12t 0.40$ 0.35 0.42 * After deducting TiOz equivalent to AI2O3 from combined (AI2O3 + Ti02) determined with EDTA. t Spectrophotometric determination with H202. $ Spectrophotometric determination with orthophenanthroline.0 Certified values from NBS. Ti0 - EDTA complex establishing quantitative de-masking of Ti. However, when A1 is present along with Ti a very interesting observation is the drifting of the end-point of the second titration at room temperature (30 f 5 “C) and the titre is less than the theoretical value indicating partial release of EDTA by tartrate from the Ti0 - EDTA complex. However, this irregularity has been overcome with a solution temperat- ure of 60-70°C in the second titration after the addition of tartaric acid. Fig. 1 shows the effect of temperature during the second titration on the stability of the Ti -tartrate complex or in other words the release of EDTA from the TiO-EDTA complex with 5.76 and 13.4 mg of Ti and Al, respectively.It is evident from Fig. 1 that the optimum temperature of the solution for titration is 60-70°C at which the end-point is stable and the correct value for Ti taken (equivalent to EDTA released on addition of tartrate) is obtained; at temperatures between 25 and 50°C, results are lower and drifting of the end-point is considerable; at temperatures >70 “C, the results approximate to the correct values but the end-point of titration is not sharp and is unsatisfactory. The release of EDTA from the A1 - EDTA complex is, however, effected by the addition of NaF or NH4F. Based on the above observations, a stepwise complexome- tric method for the direct determination of A1 and moderate to low amounts of Ti and Fe in common aluminosilicates has been developed.However, the application of the method is limited by the presence of large amounts of Ti and Fe. Mn, Ni, Co, Zn and Pb interfere by increasing the Fe value but their presence is unexpected in common aluminosilicates. Trace amounts of phosphorus, which are likely to contaminate refractory bricks, do not interfere. Table 3 shows the results of the determination of Al, Ti and Fe in synthetic solutions and Table 4 those of clays, firebricks, bauxites, etc. The results are comparable to the actual amounts taken, to the certified values or to the values obtained by other standard methods. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. References Brady, G. W. E., and Gwilb, J . , J . Appl. Chem., 1962,12,75. Grechko, L. I . , and Fartushnaya, L. H., Zuvod. Lab., 1980, 46, 565; Anal. Abstr., 1981,40, 2B206. Chasar, A. G., and Flork, G. R., Analysr, 1969, 94, 695. Bennett, H., Hawley, W. G., and Eardley, R. P., Trans. Br. Ceram. SOC., 1962, 61, 201. Bhargava, 0. M. P., Tulantu, 1979, 26, 146. Sinha, B. C., and Roy, S. K., J . Inst. Chem., India, 1974, 46, 19. West, T. S., “Complexometry with EDTA and Related Reagents,” Third Edition, BDH Chemicals Ltd., Poole, 1969, p. 187. Pribil, R., and Vesley, V., Talanta, 1963, 10, 383. Bieber, B., and Vecera, Z . , Collect. Czech. Chem. Commun., 1961, 26, 2081. Voinovitch, I. A., Guedon, D., and Louvrier, J . , “The Analysis of Silicates,’’ Israel Programme for Scientific Trans- lation Ltd., Jerusalem, 1966, p. 250. Tikhonov, V. N., Zh. Anal. Khim., l982,37,435;Anal. Abstr., 1982,43,6B83. Chen, Tsun-Chai, and Wei, Chung-Mei, Fen Hsi Hua Hsueh, 1979, 7, 327; Anal. Abstr., 1980, 40, 2B55. Paper A31143 Received May 19th, 1983 Accepted August 2nd, 1983

ISSN:0003-2654    

DOI:10.1039/AN9840900039

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, JANUARY 1984, VOL. 109 43 Spectrophotometric and Analogue Derivative Spectrophotometric Determination of Cobalt with 2,2’-Dipyridyl-2-benzothiazolyl hydrazone Raj Bhushan Singh,” Tsugikatsu Odashima and Hajime lshiit Chemical Research Institute of Non-Aqueous Solutions, Tohoku University, Sendai, 980, Japan 2,2’-Dipyridyl-2-benzothiazolylhydrazone (DPBH) reacts with cobalt(l1) t o form a 1 : 3 (metal : ligand) complex having an absorption maximum at 530 nm in weakly acidic to alkaline media. The complex, once formed; remains stable even when its solution is acidified by addition of hydrochloric acid. The complex is extractable into chloroform, benzene and 4-methylpentan-2-one. The chloroform extract has an absorption maximum a t 504nm and gives a constant absorbance in the pH range 2.0-9.3.Apparent molar absorptivities and sensitivities for an absorbance of 0.001 in the proposed procedures without and with extraction are 2.50 x lo4 I mol-1 cm-l, 2.35 ng cm-2 and 3.43 x lo4 I mol-1 cm-l,l.72 ng cm-*, respectively. Both proposed methods are selective for cobalt and especially in the method without extraction none of the other ions usually encountered interfere in the cobalt determination. Application of one of the proposed methods to the determination of cobalt in standard iron and steel samples and further sensitisation of the methods by employing analogue derivative spectrophotometry are also described. Keywords : Cobalt determination; 2,2’-dip yridyl-2- benzothiazolylh ydrazone; spectrophotometry; extraction spectrop h otome tr y; analogue derivative spectrop hotom etr y The use of hydrazones in analytical chemistry is increasing because of their easy synthesis, high sensitivity and selectivity towards various metal ions.They are widely used as colori- metric, fluorimetric and gravimetric reagents and acid - base indicators. Their analytical applications have been reviewed. 1 Systematic studies have been carried out in our laboratory on the analytical use of hydrazones with special reference to benzothiazolylhydrazones.2-5 One such hydrazone, 2,2’- dipyridyl-2-benzothiazolylhydrazone (DPBH), has been reported as a spectrophotometric reagent for the determina- tion of iron in aqueous solutions containing a surfactant.6 In the work presented here the complex formation of DPBH with cobalt(I1) was studied spectrophotometrically in order to utilise for the determination of micro-amounts of cobalt.It was found that this reagent is very selective and sensitive for cobalt(I1) in aqueous acidic media. The method developed is further sensitised by extraction of the cobalt- DPBH complex and also by the introduction of an analogue derivative spectrophotometric technique. Experimental Reagents All reagents were of analytical-reagent grade and all solutions were prepared with distilled, de-ionised water, unless stated otherwise. DPBH solution, 2.5 x 1 0 - 3 M. Prepared by dissolving the required mass of DPBH in ethanol by heating on a water-bath. DPBH was synthesised as described earlier.6 Standard cobalt(ZI) solution. Prepared by dissolving 0.5 g of metallic cobalt (99.99% pure) in 10 ml of nitric acid (1 + 1) and diluting to 500ml with water. Working solutions were prepared by dilution of this solution with water. Apparatus The apparatus described earlier6 was used for measurements of the absorbance, the absorption spectrum and the derivative spectrum.* On study leave from Bareilly College, Bareilly-243005, India. 1- To whom correspondence should be addressed. Procedure Ordinary spectrophotometry in aqueous acidic medium (procedure A ) To an aliquot containing less than 59 pg of cobalt(I1) in a 25-ml calibrated flask, add suitable masking agents if necessary, 2 ml of 2.5 x 10-3 M DPBH solution and 2 ml of 1 M acetate buffer (pH 4). After allowing the mixture to stand for a few minutes in order to complete the complexation, add 6ml of hydro- chloric acid (1 + 1) and dilute to the mark with water.Measure the absorbance of the resultant solution at 530 nm against a reagent blank prepared under the same conditions using 1-cm cells. Ordinary spectrophotometry with extraction (procedure B ) To an aliquot containing less than 17.2 pg of cobalt(I1) in a 50-ml separating funnel, add suitable masking agents if necessary, 2 ml of 2.5 x 10-3 M DPBH solution and 2 ml of 1 M monochloroacetate buffer (pH 2.5) and dilute to aproximately 20ml with water. Extract the cobalt complex into 10 ml of chloroform by shaking mechanically for a few minutes. Allow the phases to separate and transfer the organic layer into a flask containing about l g of anhydrous sodium sulphate in order to dehydrate it.Measure the absorbance of the extract at 504nm against a reagent blank prepared under the same conditions using 1-cm cells. Second-derivative spectrophotometry When the cobalt content of the coloured solution or the extract prepared by the procedures described above is too low to give a measurable absorbance, re-prepare them from the beginning using 2.5 x 10-4 instead of 2.5 x 1 0 - 3 ~ DPBH solution. Record the second-derivative spectrum from 650 to 400 nm against a reagent blank by using a combination of both first- and second-order differentiation circuits of No. 6 and a scan speed of 300nm min-1 and measure the second- derivative value (the vertical distance from a peak to a trough or that from the base line to a trough of the peak). Dissolution and pre-treatment of iron and steel samples To about 0.1 g of the sample add 10 ml of aqua regia and 10 ml of 60% perchloric acid, heat to decompose the sample and44 ANALYST, JANUARY 1984, VOL.109 evaporate the mixture to dryness. After cooling to room temperature dissolve the residue in 10 ml of hydrochloric acid (1 + l), filter to remove any undissolved portion (silica) and wash with a small amount of hydrochloric acid (1 + 1). Add 10 ml of 4-methylpentan-2-one to the filtrate and washings and shake for 5min to remove iron(II1). Separate the aqueous phase, heat it to remove most of the acid and dilute to 100 ml with water. Use 2ml or another appropriate aliquot of the resultant solution for the determination. Results and Discussion Properties of the Complex and Effect oi'pH Cobalt(I1) reacts with DPBH in weakly acidic to alkaline media to form a yellow - orange complex that is insoluble in water above pH 3, but soluble below pH 3 because of protonation at the pyridine nitrogen in the complex molecule.The complex, once formed, remains stable even when the solution is acidified and gives a maximum absorption at 530 nm against a reagent blank (curve A Fig. 1). Fig. 2 shows the effect of the amount of hydrochloric acid (1 + 1) added after the complexation was complete at pH 4. The absorbance increases gradually with increasing amount of the acid owing to the protonation of the complex and free DPBH exists; however, the variation in the absorbance is very sma!l and the absorbance can be considered to be almost constant when 6 k 0.5 ml of hydrochloric acid (1 + 1) are added per 25 ml of the final volume.Hence further studies were carried out by adding 6 ml of the acid in 25 ml of final solution after the complexation was complete, in order to avoid possible interferences from other ions. 0.4 al C m e g 0.2 n Q 0 1 2 0 0) > .- $ 1 4- .- 5 2 ? c 3 % -0 (7, 400 500 600 Wavelengthinm Fi . 1. (a) Absorption spectra of DPBH and its cobalt complex and (by second-derivative spectra of cobalt - DPBH complex. Cobalt(II), 720 p.p.b.; DPBH, 1 x M; A, C, E and F, cobalt - DPBH complex against reagent blank; B, reagent blank against water; and D, reagent blank against chloroform. Continuous lines, in 1.4 N hydrochloric acid solution; and broken lines, in chloroform - I L I I I n o 3 6 9 12 15 8 0.2 ' a HCI addedhl Fig.2. Effect of amount of hydrochloric aicd (1 + 1) added after complexation was complete. Cobalt(II), 823 p.p.b.; DPBH, 2 X M; reference, reagent blank The cobalt - DPBH complex is extractable into chloroform, benzene, 4-methylpentan-2-one and 1 ,Zdichloroethane and gave the largest absorbance in chloroform; hence chloroform was used as the extraction solvent in the procedure with extraction. The complex is quantitatively extracted into chloroform in the pH range 2.0-9.3 as shown in Fig. 3 and has a maximum absorption at 504nm against a reagent blank (curve C Fig. 1). Effect of DPBH Concentration A 4-fold molar excess (in procedure A) and a 5-fold molar excess (in procedure B) of DPBH were required in order to obtain a constant absorbance.The excess of DPBH was not critical in either instance and did not interfere. Effect of Shaking Time Variation of the shaking time from 1 to 15min revealed that 1 min was sufficient for complete extraction of the complex. The extraction of cobalt was reproducible; 99.7% of that present in the aqueous phase was extracted by a single extraction, according to the results obtained by a single extraction followed by the determination of cobalt in the organic and aqueous phases. One extraction, therefore, seems to be sufficient for the determination of cobalt. Effect of Other Variables in Extraction The absorbance of the extract remained constant even when the organic to aqueous phase ratio was varied from 1 : 1 to 1 : 5 and ethanol concentrations up to 15% in the aqueous phase did not affect the absorbance of the organic phase. Stability of the Complex The cobalt-DPBH complex in aqueous acidic medium is stable and gave a constant absorbance even after 10h.The extracted complex is also stable and gave a constant absorb- ance even after 13 h. Composition of the Complex The molar composition of the complex formed under the conditions for the determination of cobalt was ascertained by Job's method of continuous variations and the molar-ratio method. Both methods indicated the formation of a 1 : 3 (metal : ligand) complex under either of the conditions em- ployed for the determination of cobalt. Calibration Graph, Sensitivity and Precision Straight-line calibration graphs passing through the origin were obtained using the recommended procedures.The equations of the lines obtained by a least-squares treatment were Co(p.p.m.) = 2.3514 . . . . . . (1) Co (p.p.m.) = 1.72A . . , . . . (2) where A is the absorbance and equations (1) and (2) L g 0.2 I 2 I I I I I 0 2 4 6 8 10 PH Fig. 3. Effect of pH on the extraction of cobalt - DPBH complex. Cobalt(II), 720 p.p.b.; DPBH, 2.5 x l o - 4 ~ ; reference, reagent blankANALYST, JANUARY 1984, VOL. 109 45 correspond to the calibration graphs in procedures A and B, respectively. The optimum ranges for the determination of cobalt, the sensitivities for an absorbance of 0.001 and the molar absorptivities calculated from equations (1) and (2) were 2.7-58.9 pg, 2.35 ng cm-2 and 2.50 X 1041 mol-1 cm-1 in procedure A and 0.8-17.2 pg, 1.72 ng cm-2 and 3.43 x 104 1 mol-l cm-1 in procedure B, respectively.Two series of ten standard solutions each containing 20.6 and 7.2pg of cobalt were analysed by the recommended procedures. The results gave relative standard deviations of 0.71 and 0.46% in procedures A and B, respectively. Effect of Diverse Ions The effects of foreign ions on the cobalt determination by procedures A and B are summarised in Tables 1 and 2, respectively, from which it can be seen that as a rule both procedures are selective for cobalt, except that nickel(I1) interferes seriously in procedure B because of complex formation with DPBH, thiosulphate interferes in procedure A because of precipitation and the tolerance limit for palladium- (11) is low in both procedures because the palladium complex is equally stable in acidic medium.In addition, in procedure A Table 1. Tolerance limits for various ions in the determination of 20.6 pgof cobalt by procedure A. Tolerable error: +3% Ion Tolerance limit C1-, Br-, NO3-, C104-, S042-, P043-, tartrate, I- . . . . . . . . . . . . . . . . . . 130mg Thiocyanate . . . . . . . . . . . . . . 100 mg F-, oxalate . . . . . . . . . . . . . . 40 mg Thiourea . . . . . . . . . . . . . . 30 mg Thiosulphate . . . . . . . . . . . . . . Precipitated Mn(I1) . . . . . . . . . . . . . . . . 3 000 pg* Ca(II), Mg(I1) . . . . . . . . . . . . . . 2 000 pg* Al(III), Cd(II), Fe(III)t, Pb(II), Zn(I1) . . . . 2000 pg Cu(II)$,$ . . . . . . . . . . . . . . 1 500 pg Ag(I), Hg(II)§, V(IV, V)q . . . . . . . . 1000 pg citrate, thioglycollate, ascorbate . .. . . . 200 mg* . . . . . . . . . . . . . . Cr(II1, VI) 500 Pg Ni(II)§ 100 Pg Pd(I1) 2 Pg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . * Maximum tested. t 2 ml of 1 M sodium fluoride solution were added. $ 2 ml of 1% thiourea solution were added. § 4 ml of DPBH solution were added. n2 ml of 0.1 M sodium oxalate solution were added. Table 2. Tolerance limits for other ions in the determination of 7.2 pg of cobalt by procedure B. Tolerable error: k3Y0 Ion Tolerance limit C1-, Br-, I-, NO3-, C104-, S042-, PO4-, Tartrate, citrate . . . . . . . . . . . . 100 rng Thiocyanate . . . . . . . . . . . . . . 60 mg Oxalate . . . . . . . . . . . . . . . . 30mg F-, thiourea . . . . . . . . . . . . . . 20 mg Ca(II), Mg(I1) . . . . .. . . . . . . . . Al(III), Cd(II), Mn(II), Pb(II), Zn(1I) . . . . 1000 pg thioglycollate, thiosulphate, ascorbate . . . . 200 mg* 2 000 pg* Cr(II1,VI) 200 Ag(I), Fe(II1)t 100 I.I-8 50 Yg V(IV,V)$ . . . . . . . . . . . . . . Cu(II)§ 30 Pg Hg(I1) 15 Pg Pd(I1) 2 Pg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ni(I1) . . . . . . . . . . . . . . . . <1pg * Maximum tested. t 1 ml of 1 M sodium fluoride solution was added. $ 1 ml of 0.1 M sodium oxalate solution was added. § 1 ml of 1% thiourea solution was added. with acidification the amounts of foreign ions tolerated are larger than those in procedure B with extraction. In the absence of masking agent amounts of copper(1I) and iron(II1) tolerated are 100 and 30 pg, but these limits can be increased to 1500 and 2000pg by masking them with thiourea and fluoride, respectively.Nickel(II), which interferes seriously in procedure B, can be tolerated up to 1OOpg provided that sufficient DPBH is present in the solution to react completely with cobalt(I1) before the addition of hydrochloric acid (1 + 1). Application to Actual Samples In order to confirm the usefulness of the proposed methods, procedure A was successfully applied to the determination of cobalt in standard iron and steel samples. The results are summarised in Table 3. Sensitisation by Employing Analogue Derivative Spectrophoto- metry Ishii and co-workers have previously reported that derivative spectrophotometry using an analogue differentiation circuit is extremely effective for the sensitisation of ordinary spectro- photometry.7.8 As an example of sensitisation, the second- derivative spectrophotometric determination of cobalt is described here.Selection of conditions for measurement of the second- derivative value As both the time constant of the analogue differentiation circuit and the scan speed of the spectrophotometer affect the second-derivative value (the vertical distance from a peak to a trough or that from the base line to a trough of the spectrum) in second-derivative spectrophotometry, these need to be selected to give a well resolved large peak (to give good selectivity and higher sensitivity in the determination). This is done on the basis of the breadth of the band in the ordinary absorption.spectrum. Each differentiation circuit in our apparatus has six additional circuits with different time constants, which can be varied easily by turning a kn0b.7.~ They are represented by circuit numbers from 1 to 6 and an increase in the circuit number means an increase in the time constant. In Fig. 4 the second-derivative spectra of the cobalt - DPBH complex solution measured with varying cir- cuit number or scan speed are shown; circuit No. 6 and a scan speed of 300 nm min-1 are seen to be preferred for the cobalt determination. The second-derivative spectra of the complex solution measured under the recommended conditions have already been shown in Fig. 1 in comparison with the corresponding absorption (zero-derivative) spectra.Table 3. Determination of cobalt in standard iron and steel samples by procedure A Cobalt found, YO Sample Proposed method Certified value Iron and steel, JSS-611-5 . 0.364 0.36 0.362 0.366 0. 368 Average: 0.3& Iron and steel, JSS-606-5 . 0.47, 0.468 0.468 0.471 Average: 0.47() 0.4646 ANALYST, JANUARY 1984, VOL. 109 0.2 0 0) .- +4 .- 9 0.2 8 L Q) 5 * 0.4 0.6 0.8 4 240 300 600 U 1 500 600 400 500 600 Wavelengthhm Fig. 4. Influence of ( a ) scan s eed (with circuits all No. 6) and (b) circuit number (with scan spee8300 nm min- 1) on second-derivative s ectra of cobalt - DPBH complex in 1.4 N hydrochloric acid solution. &balt(II), 115 p.p.b.; DPBH, 1.5 x 1 0 - - 5 ~ ; reference, reagent blank; slit width, 1 nm. Numerical values indicate scan speeds in ( a ) and first- and second-differentiation circuits in ( b ) Calibration Graph The calibration graph, prepared by plotting the second- derivative value of coloured solutions prepared in aqueous acidic medium versus the cobalt concentration, was a straight line passing through the origin when either the peak-to-trough values or the base line-to-trough values were plotted. The equation was Co (p.p.b.) = 1370 .. . . . . (3) for the peak-to-trough measurements and Co (p.p.b.) = 159D . . . . . . (4) for the base line-to-trough measurements, where D is the second-derivative value represented by the conversion of the value into absorbance. Fig. 5 shows a calibration graph obtained under the recommended conditions, from which it can be seen that cobalt down to the 7 p.p.b.level can be easily determined by the proposed method. A similar calibration graph was obtained in the extraction method. Comparison with Other Methods Several hydrazones have been proposed for the determination of cobalt, e.g., pyridine-2-aldehyde-2-quinolylhydrazone,~ 2,2’-dipyridyl-2-pyrimidylhydrazone,1() 2-benzoylpyridine-2- pyridyl hydrazone, 1 I benzil mono( 2-pyridylhydrazone)I2 and 0.E 0.6 - m 0) .- 4- > f 0.4 .- L U 8 v) 0.2 0 Fig. 5. 20 40 60 80 100 Co, p.p.b. An examde of the calibration graDh for second-derivative spectrophotometri. A, peak-to-trough Galies and B, base line-to- trough values. Circuits, all No. 6; scan speed, 300 nm min-I; slit width, 1 nm; recorder sensitivity, x 1 ; reference, reagent blank; DPBH, 1.5 x 10-5 M ; hydrochloric acid concentration, 1.4 N ; cobalt(I1) concentra- tions (a) 7.2, (b) 14.4, (c) 28.8, (d) 57.6and (e) 115 p.p.b. most recently biacetylmonoxime-2-pyridylhydrazone. l 3 In comparison with methods using these hydrazones the pro- posed methods seem to have similar sensitivity but higher selectivity. In addition, extremely high sensitivity is obtained in the proposed methods by introducing analogue derivative spectrop ho tome try. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. References Singh, R. B., Jain, P., and Singh, R. P . , Talanta, 1982,29,77. Odashima, T., and Ishii, H., Anal. Chim. Acta, 1975,74,61. Odashima, T., Anzai, F., and Ishii, H., Anal. Chim. Acta, 1976, 86,231. Odashima, T., Satoh, S., and Ishii, H., Nippon Kagaku Kaishi, 1982,1322. Ishii, H., Odashima, T., and Imamura, T., Analyst, 1982,107, 885. Singh, R. B., Odashima, T., and Ishii, H., Analyst, 1983,108, 1120. Ishii, H., and Koh, H., Nippon Kagaku Kaishi, 1980,203. Ishii, H., and Satoh, K., Fresenius 2. Anal. Chem., 1982,312, 114. Singhal, S. P., and Ryan, D. E . , Anal. Chim. Acta, 1967,37,91. Singh, R. B., Jain, P., Garg, B. S . , and Singh, R. P . , Anal. Chim. Acta, 1979,104,191. Going, J. E., and Pflaum, R. T., Anal. Chem., 1970,42,1098. Pflaum, R. T., and Tucker, E. C., Anal. Chem., 1971,43,458. Asuero, A. G., Rosales, D., and Rodriguez, M. M., Analyst, 1982,107,1065. Paper A31245 Received August5th, 1983 Accepted August22nd, 1983

ISSN:0003-2654    

DOI:10.1039/AN9840900043

出版商:RSC    

年代:1984

数据来源: RSC