Analyst, September, 1979, VoL. 104, pp. 837-845 837 Determination of Nitrate in Raw, Potable and Waste Waters by Ultraviolet Spectrophotometry P. J. Rennie and A. M. Sumner* North West Water A uthority, Southern Division, Allport RoadlBridle Road, Bromborough, Wirral, Mersey- side, L62 6AB and F. B. Basketter North West Water Authority, Directorate of Scienti=fic Services, Dawson House, Great Sankey, Warrington, WA5 3 L W A method is proposed for the determination of nitrate in raw, potable and waste waters using ultraviolet spectrophotometry. The use of an activated carbon filter a t an elevated pH eliminates interference from organic matter, i.s., substances commonly assumed to be responsible for the related absorb- ances a t 275 and 210 nm. The procedure also removes the interferences of several cations that are precipitated out of solution. The development work leading up to the proposed method is discussed with reference to the relevant behaviour towards organic matter and the nitrate ion of activated carbon materials.The method has a limit of detection of 0.006 mg 1-1 N and a total standard deviation of 0.016 mg 1-l N at a nitrate concentration of 1.05 mg 1-1 N in potable water. No statistically significant difference was detected between the proposed method and an established automated method for a wide range of samples. Keywords : Nitrate determination ; water ; waste water ; ultraviolet spectro- Photometry ; activated carbon The various methods at present employed for the determination of nitrate in water fall into one of five main categories: the reduction of nitrate to ammonia; manual methods using chromogenic reagents; direct spectrophotometry ; ion-selective electrode methods; and reduction to nitrite.Of these techniques, one of the most reliable, simple and rapid is direct spectrophotometry in the ultraviolet (UV) region, but its range is limited by inter- ferences and the work described in this paper was intended to extend this range. The use of direct UV methods is widespread and especially suitable for waters of low organic content, although waters containing organic matter can be analysed by use of established correction factor technique~l-~ In this and previous ~ o r k , l - ~ “organic matter” is commonly assumed to be those materials responsible for related absorbances in the region of 200-300 nm.In this work, absorbances at 210 and 275 nm have been selected to indicate the presence of nitrate and “organic matter,” respectively. It should be pointed out however, that strictly the reference to “organic matter’’ means “materials absorbing at 275 nm.” We have used the absorbance at 275 nm as an indicator of the presence or absence of “organic matter’’ that could absorb also at 210 nm. The term “organic matter” is therefore used throughout the paper for brevity and because, as will be shown in the method validation section, it is believed that “organic matter” content is represented fairly accurately in the water samples of interest by “absorbance at 275 nm.” The inter-relationship of absorbances at 275 and 210 nm due to organic matter is based on observations of the similarity in absorption spectra of water samples in the region of 300 nm and below.For example,6 both synthetic solutions of humic and fulvic acid extracts and real samples illustrate well this inter-relationship, and we have also obtained UV spectra that support these findings. The correction factor techniques already mentioned involve the use of an arbitrary factor that corrects for the UV absorbance of organic matter at the wavelength selected for nitrate-ion determination. This approach cannot be used when the nitrate concentration is low and the organic content is high. The reason for such failures of the correction factor approach arises from the variability of the different organic compounds * Present address : North West Water Authority, Western Division, Merton House, Bootle, Merseyside, L20 3NH.838 Analyst, VoL.104 likelv to be present in different types of water. The variability of the correction factors and-the possible lowering of precision by the dilution required to bring the absorbances within the linear range of UV spectrophotometric analysis thus necessitate a non-arbitrary method with improved accuracy. Such a method is proposed, which removes not only the organic matter prior to absorbance measurement, but also some cationic interferences that normally affect the spectrophotometric determination of the nitrate ion. RENNIE, sUMNER AND BASKETTER: DETERMINATION OF NITRATE Methods of Dealing with Organic Interference In addition to the factor methods already briefly mentioned, three other methods have been proposed to deal with interferences by organic matter.7 These methods involve the chemical reduction of the nitrate ion, the coagulation of organic matter with aluminium sulphate and the adsorption of organic matter by activated carbon.Two methods were attempted for the elimination of nitrate by its reduction in a sample to ammonia. These result in a solution, the UV absorbance of which when subtracted from the absorbance of the unreduced sample gives the absorbance directly attributable to the nitrate ion. Both methods were attempted on samples of raw water and for the zinc - copper couple method on nitrate standard solutions also. In the reduction by a zinc - copper couple,2?* it was inferred that organic matter was also reduced as the absorbance at 275 nm was decreased, The other reduction method involved heating the raw water with hydrazine hydrate in a boiling water-bath and then boiling, after the addition of sodium hydroxide, to expel ammonia.On acidification, the resultant solution exhibited an interfering absorbance at 210 nm, which was shown to be due to undestroyed hydrazine. These procedures cannot therefore be used to estimate the absorbance at 210 nm due to species other than nitrate and hence to deduce the nitrate concentration. So far as the aluminium sulphate coagulation methodS~l0 is concerned, there is evidence7 that removal of organic matter is only approximately 85% and indeed the APHA publicationlo recommends coagulation for removal of colour but still retains a correction factor for the remaining dissolved organic matter that absorbs at 275 nm.Recently, Brown and Bellingerll proposed a method for improving the correction factor procedure in the presence of humic acids and also suggested the use of an anion-exchange resin for the removal of most of the UV-absorbing organic materials. The difficulties encountered in dealing with organic matter and the potential effectiveness of activated carbon, particularly in view of the recent paper by Slanina et aZ.,12 prompted us to concentrate on the investigation of activated carbon methods. The adsorption of organic matter by activated carbon is well known but a less appreciated fact is that there is some e ~ i d e n c e ~ ~ ~ ~ ~ ~ ~ that nitrate ions are removed from aqueous solutions by activated carbon and that the effect may be pH dependent.In fact, activated carbon has been recommended13 for removing dissolved colour at alkaline pH prior to nitrate determination with chromogenic reagents. Although Slanina et aZ.12 investigated a number of carbons and recommended a particular one, they did not indicate the behaviour of the other carbons. It was therefore necessary for our purpose to investigate the organic matter content, nitrate content as received and the extent of nitrate retention of about 20 types of activated carbon. It was found that most of the carbons tested contained nitrate and organic matter that was easily leached out, and all the carbons retained nitrate to variable extents (19444% retention) under acidic conditions.Nitrate retention tests were made after washing the carbons free of nitrate and/or organic matter. One of the materials, ADC 33 (Sutcliffe, Speakman & Co.), was in regular use at a treat- ment plant and earlier work had indicated that under alkaline conditions contamination could be removed and nitrate was not retained. The effect of altering the pH was investi- gated in more detail with this carbon. The tests consisted of adding aliquots of carbon to a water of known nitrate concentration and containing organic matter, at several pH values between 1.5 and 12.6, mixing for 5 min, removing the carbon by filtration through Whatman GF/C paper and reducing the pH to 1.5-1.9 by adding the same volume of mixed acid reagent to each test solution before measuring the absorbance at 210nm (for mixed acid reagent, see Reagents and Interference efects sections).The absorbances were compared with those of samples that had not been treated by the addition of carbon. The results (Table I) confirm the pH dependence of nitrate retention and organic matter adsorption; the nitrate retention appears to show a complex relationship with variation in pH but this was notSeptember, 1979 I N RAW, POTABLE AND WASTE WATERS BY UV SPECTROPHOTOMETRY 839 TABLE I EFFECT OF pH ON THE ADSORPTION PROPERTIES OF AN ACTIVATED CARBON (ADC 33) PH 1.5 3.0 5.8 7.3 10.4 12.0 12.6 Organic removal, yo 90 90 90 99 97 98 100 Nitrate-ion retention, % 41 59 24 63 57 0 0 investigated further. In the pH range 1.5-1.9, the variation of adsorption5 of nitrate solutions with pH is less than 7% of the absorbance reading and is therefore not significant in the present context.Removal of Organic Matter with Activated Carbon Methods investigated were (a) batchwise addition of powdered or granular activated carbons followed by filtration to remove the carbon, (b) use of a carbon column and (c) construction of a filter stack using materials impregnated with carbon. Firstly, a method was developed using ADC 33 activated carbon that consisted in membrane filtration of the sample and addition of powdered carbon (0.5 g) to 100 ml of filtrate adjusted to pH >12 by the addition of sodium hydroxide solution. This mixture was stirred for 5 min before removing the carbon by filtration. The pH of the filtrate was reduced to below 2 and the absorbance was measured at 210 nm against distilled water. There was no absorbance due to organic matter, i.e., at 275nm.A calibration graph was constructed using standards treated in the same manner. The drawbacks to this procedure were the need to weigh out the carbon, contamination of the glassware and the need for two filtrations. In an attempt to produce a less tedious method and to optimise the conditions used, a carbon column was prepared from granular carbon supported on a glass-wool plug. It was not possible to achieve a sufficiently long contact time for total removal of organic matter and so this approach was abandoned. Several materials impregnated with carbon, viz., filter-paper, fibreboard, woven fabric and polymer beads, were examined for organic matter and nitrate content and for the extent of nitrate retention under acidic conditions.It was found that these materials behaved in a similar manner to the granular activated carbons in that varying degrees of contamination with nitrate and/or organic matter were observed. Of the carbon filter media tested, an analytical-grade filter-paper (Schleicher & Schull No. 508 active carbon paper; Schleicher & Schiill, Dassel, West Germany; Anderman & Co. Ltd., East Molesey, Surrey) proved to be the most satisfactory with regard to the ease of construction of a filter stack (see Fig. l), the least cleansing required prior to use and the efficient removal of organic matter. I t was therefore decided to pursue a development of the method using the Schleicher & Schull paper under the alkaline conditions previously investigated with ADC 33 carbon.Preliminary tests with four, eight and sixteen layers of the carbon paper showed that sixteen layers were required in order to allow a sufficient contact time for the removal of organic matter from samples (the initial and residual absorbances at 275 nm were 0.152 and 0.004, respectively, using 10-mm cells). The extents of removal of organic matter with four and eight layers were 79 and goyo, respectively. At the same time tests with standard nitrate solutions showed that nitrate ion was not absorbed under alkaline conditions by the increased number of layers. Using this information we have developed the following proposed method using the filter assembly shown in Fig. 1.The extent of nitrate retention varied between 19 and 66%. Proposed Analytical Method Apparatus An ultraviolet spectrophotometer capable of measuring absorbances at 210 nm and a pair of matched ultraviolet-grade silica cells (path length 10 mm) were used. The filtration apparatus shown in Fig. 1, consisting of sixteen layers of 60 mm diameter Schleicher & Schiill No. 508 active carbon papers sandwiched between two 60mm diameter Whatman840 A n a l y s t , vol. 104 GF/C papers in a Carlson-Ford filter holder (Gallenkamp & Co. Ltd., London), was used in conjunction with a suitable vacuum source and a Biichner flask. To prepare a new carbon filter stack for use, wash with 400ml of 3.5% m/V sodium hydroxide solution followed by 200 ml of distilled water.This treatment gives reagent blank values of the order of 0.025 absorbance unit. Between batches of analyses, contact with a laboratory atmosphere may result in contamination of the filter stack with UV- absorbing substances and it may be necessary to clean the filter stack by the above alkaline procedure in order to maintain satisfactory reagent blank values. The top GF/C paper may be changed between samples if the filter exhibits significant headloss. The operational life of the prepared filter stack is dependent on the organic content of the samples; a typical capacity is 100 samples with an absorbance of 0.045 at 275 nm before organic breakthrough occurs. Glass bottles should be used for samples; these bottles and other glassware used for the method should be cleaned by treatment with concentrated sulphuric acid followed by thorough rinsing with distilled water.RENNIE, SUMNER AND BASKETTER: DETERMINATION OF NITRATE Section I elevation 80 ml capacity stainless- m- steel funnel -60 mm diameter GF/C ~ filter-papers 16 layers, 60 mm diameter, d+p Schleicher & Schull No. 508 active carbon ~ papers Filter base __ - Stainless-steel gauze Alkali-resistant bung (e.g., silicone rubber) for connection to 250-ml Buchner flask (vacuum required >/ 98 kPa) Fig. 1. Activated carbon filter assembly. Reagents Reagents of analytical-reagent grade are preferable unless otherwise stated. Distilled water should be used rather than de-ionised water, which could contain UV-absorbing materials derived from the ion exchangers and from organic matter present in the feedstock.I t is advisable to check that the reference distilled water does not contain materials that absorb at 210 nm. A 500-ml volume of the distilled water should be re-distilled from an all-glass apparatus and 200-, 50- and 150-ml sequential portions of distillate collected. Using the 50-ml portion in the reference cell, the other portions and the residue should have an absorbance reading of less than 0.01 for the original distilled water to be considered accept- ably free from UV-absorbing materials. Store in a polyethylene bottle (stable for about 3 months). 0.1 g of sulphamic acid in 500 ml of 5% V/V sulphuric acid. Prepare freshly by 50-fold dilution of a 100 mg 1-1 N stock standard solution (stable for 3 months) of potassium nitrate (0.721 8 g of dried potassium nitrate per litre).Dilute the standard solution as appropriate to produce 0.25, 0.5, 0.75, 1.0, 1.5 and 2.0 mg 1-1 N standard solutions. Borosilicate glass containers should be used. Sodium hydroxide solution, 3.5% m/V. Mixed acid reagent. Dissolve 5 Standard nitrate solution, 2 mg 1-1 N. This solution should be stored in borosilicate glass (stable for at least 2 months). Procedure than 2 mg 1-1 N of nitrate in a calibrated flask and mix. Add 5 & 0.1 ml of sodium hydroxide solution to 100 ml of sample containing not more Pass 30 ml of this solution throughSeptember, 1979 841 the filter and discard it, then filter 50ml of the solution and keep the filtrate. Pipette 40 ml of this filtrate into a 50-ml calibrated flask containing 5 ml of the mixed acid reagent, dilute to 50 ml with distilled water and measure the absorbance at 210 nm against water in the reference cell.Absorbances corrected for the reagent blank are then converted into milligrams of nitrogen per litre using a regressed equation from standards and a reagent blank treated in the same manner. A blank determination must be carried out with each batch of analyses by taking 100 ml of distilled water through the full procedure in place of the sample. Linear regression of the calibration graph gave a mean correlation coefficient of 0.9996 for a y = mx + c equation and the mean standard error of estimate was 0.016mg1-I N. An absorbance reading of about 0.4 is given by a 1 mg 1-1 N standard solution treated as described above.I N RAW, POTABLE AND WASTE WATERS BY UV SPECTROPHOTOMETRY The calibration graph is linear up to at least 2.0 mg 1-1 N. Validation of the Proposed Method Statistical evaluation of the method was carried out according to the procedures given by Cheeseman and Wilson1* (viz., precision of standards and samples and spiking recovery). The method was compared with an established method and several specific interference effects of substances known to absorb in the UV region were investigated. Precision and recovery tests The precision target for the determination of nitrate in potable water had been set at a total standard deviation of an individual determination of 5% of concentration or 0.01 mg 1-1 N, whichever was the greater. To test the performance of the proposed method against this target, duplicate analyses were carried out, on each of five days, of distilled water, 0.5 and 1.0 mg 1-1 N standards, treated water from Lamaload reservoir (an upland impounded reser- voir in the Macclesfield area fed by the River Dean) and treated water from Lamaload reservoir spiked with a 0.474 mg 1-1 N addition.The results obtained from the last two samples were used to calculate the spiking recovery. The results were analysed to derive the corresponding within-batch (sw), between-batch (sb) and total (st) standard deviations. The results are summarised in Table 11; the number of degrees of freedom were derived14 from duplicate analyses on each of five days and are given in parentheses. TABLE I1 PRECISION OF ANALYTICAL RESULTS Standard deviation*/mg 1-1 N A f , Mean concentration Solution SW s b S t found/mg 1-l N 0.00 mg 1-l N .. .. . . . . . . 0.001 (5) - - 0.01 1.00 mgl-l N . . . . .. . . . . 0.011 (5) N.S. (4) 0.017 (6) 1.00 0.50 mg 1-1 N . . .. . . . . . . 0.012 (5) N.S. (4) 0.019 (6) 0.50 Treated Lamaload watert . . . . . . 0.005 (5) 0.015 (4) 0.016 (4) 1.053 Treated Lamaload water + 0.474 mgl-l N 0.012 (5) 0 (4) 0.012 (7) 1.551 * Figures in parentheses are the degrees of freedom. N.S. indicates that the result is not statistically f The absorbance of “Treated Lamaload water” plus reagents but without carbon filtration is of the significant. order of 0.065 at 275 nm. Analysis of variance14 showed that in only one instance (the sample) was there a signifi- cant component of between-batch variability, possibly reflecting unusually consistent deter- minations within each of the batches.A limit of detection of 0.006mg 1-1 N was calculated by using the formula 2t2/2(sw), where t is Student’s single-sided t for 0.10 probability and sw is the within-batch standard deviation of the blanks. The spiking recovery was 105 ~f 1.1%, where 1.1% is the 95% confidence interval (equal to s t / 4 5 , where s is the total standard deviation of the spiking addition, t is Student’s double- sided t for 0.10 probability and there were five batches). It must be remembered that This easily satisfies a target for potable waters of 0.02 mg 1-1 N.842 Analyst, vol. 104 because the method involves what is effectively a non-specific UV absorbance characteristic, even if interferences are present they will probably not be detected by a simple spiking recovery test based on additive absorbances.RENNIE, SUMNER AND BASKETTER: DETERMINATION OF NITRATE Comparison with established method Samples of raw, potable and waste waters from throughout the North West Water Authority area were taken. These samples were carefully split between two bottles; one sample was analysed in the Authority’s Rivers Division, Warrington Laboratory, using an automated analyser for nitrate determination, and the other was analysed by the proposed method. Both samples were stored in a similar manner at 4 “C and the nitrate determinations carried out at the same time using the respective methods. The automated method consisted in the reduction of nitrate by means of a cadmium column and determination of the total nitrite using 1-naphthylethylenediammonium chloride - acid sulphanilamide reagent, the original nitrite concentration being subtracted to obtain the nitrite derived from the nitrate content.The performance characteristics of the auto- mated method were as follows: range, 0-10 mg 1-1 N ; total standard deviations, 0.038 and 0.080 mg 1-1 N at concentrations of 1.0 and 7.5 mg 1-1 N, respectively (with 19 degrees of freedom) ; spiking recovery, 1oo.4y0 at a concentration level of 3 mg 1-1 N; maximum possible bias, 2.7% in a UK inter-laboratory exercise involving 19 laboratories. The results obtained by the two methods and the absorbances at 275 nm before and after carbon treatment as an indication of the efficiency of removal of organic matter are sum- marised in Table 111.The results indicate that there is good agreement between the two methods for the wide range of sample types considered. The differences (UV method minus automated method) between the pairs of results in Table I11 were examined using Student’s t-test.14 The results were dealt with in two groups, chosen such that the variance within each group was constant. This subdivision corresponded to the degree of dilution, which conveniently divides the samples into “water supply samples and Crewe effluent” and “sewage works effluents excluding Crewe,” i.e., “low” and “high” dilutions, respectively. For each group the t-value was calculated using the equation where 3 is the mean of the differences, n is the number of pairs and s is the standard deviation of the differences between pairs. The results of the examination are summarised in Table IV, which shows that the calculated t-values for each group are less than the tabulated, double-sided Student’s t for 0.05 probability and the respective number of degrees of freedom (n - 1).This indicates that there is no evidence of a statistically significant difference between the two methods over the wide range of sample types considered. The 275-nm absorbance readings (Table I I I) without carbon filtration are relatively so high that few samples could be analysed with confidence using the direct UV method1°,15 with a factor correcting for the presence of organic matter. This conclusion is based on the assumption of a correction factor of 3, i.e., the absorbance due to organic matter at 210 nm is three times that at 275 nm, and using the absorbance at 210 nm of nitrate ion as 1 mg 1-1 N = 0.4 unit.As an example, from Table 111, the Lamaload raw water sample absorbance due to organic matter at 210nm (without carbon) would be 3 x 0.104 = 0.312; the total absorbance measured (without carbon) at 210nm was in fact 0.626. Now, by using the correction factor method, the net absorbance at 210nm would be 0.214, which gives a nitrate concentration of 0.79 mg 1-1 N. This is approximately 75% of the actual concentra- tion, thus showing how the use of an assumed correction factor is unreliable. The results (Table 111) therefore illustrate how the proposed method extends the applicability of the UV method for nitrate determination to samples that contain significant amounts of organic matter.It is pertinent here to discuss how the results in Table I11 may be used in support of the assumption made in the introduction concerning the inter-relationship of absorbance readings at 210 and 275 nm, defined for the present context as “organic matter” but understood to mean “UV-absorbing substances.” The results given in Table 111, which cover a wide range of sample types (and therefore it may be expected that the samples will contain different organic species), show that for samples for which the absorbance at 275nm is significantSeptember, 1979 843 before carbon filtration, the absorbance falls to zero after carbon filtration.This indicates (according to the above assumption) that the organic matter has been totally removed by the carbon. Now, if any species absorbing at 210 nm (other than nitrate) were present, the UV method would give a higher “apparent” nitrate concentration than the automated method, which is specific for nitrate and is free from bias. There is no such bias between the two methods (see Table IV) and so this validates the assumption that absorbances at 210 and 275 nm are inter-related such that absence of absorbance due to organic matter at 275 nm indicates absence of absorbance at 210 nm. Further evidence is given by the sample from Hooton Borehole (Table 111), which has a negligible organic matter content as indicated by the absorbance at 275 nm without carbon filtration.The agreement between the results of the UV and automated methods confirms the absence of organic matter (and any other UV-absorbing material) that absorbs at 210 nm. The 275-nm absorbance readings with carbon filtration (Table 111) show that there are no significant amounts of organic matter remaining in the filtrates, thus confirming the absence I N RAW, POTABLE AND WASTE WATERS BY l J V SPECTROPHOTOMETRY TABLE I11 COMPARISON OF RESULTS OBTAINED BY THE PROPOSED METHOD AND BY AN ESTABLISHED AUTOMATED METHOD FOR RAW AND POTABLE WATERS Sample* Raw surface waters- Lamaload . . Sutton Hall (1 T’ 1) . . Alwen . . . . . . Goyt . . . . . . Langthwaite (1 + 1) . . Cant Clough . . . . Worthington (1 + l j . . Llangollen Canal (1 + 1) Stocks . . . .. . Arnfield . . . . . . Treated surface waters- Sutton Hall (1 + 1) . . Lamaload . . * . Alwen . . . . . . Langthwaite . . . . Hooton (1 + 1) . . Slag Lane . . . . Borehole waters- Haydock (1 + 3) , . Sewage works final efluents f- Adderley (1 + 19) . . Bunbury (1 + 19) . . Haslington ( 1 + 9) . . Crewe (1 + 3) . . . . Elton (1 + 19) . . AND FINAL EFFLUENTS Nitrate content/mg 1-I N - Automated Proposed method? method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Little Budworth South (1 + 9) Doddington ( 1 + 19) . . . . Mickle Trafford (1 + 9) . . Thornton Hough ( 1 + 9) . . Capenhurst (1 + 9) . . . . Wrenbury (1 + 19) . . . . Audlem (1 + 19) . . . . 1.05 3.15 0.30 0.85 0.80 0.55 4.00 3.10 0.45 0.80 1.11 3.04 0.47 0.80 0.70 0.50 4.00 3.06 0.40 0.70 2.40 2.70 1.05 1.04 0.40 0.52 0.80 0.72 3.20 3.20 7.40 7.40 0.05 0.06 35.5 0.1 28.9 33.5 16.95 18.8 39.4 19.5 25.4 12.3 23.9 23.6 32.4 0.2 30.9 32.5 16.9 18.9 38.3 19.8 25.0 13.1 24.3 22.9 Absorbance a t 275 nm - Without carbon filtration 0.104 0.097 0.151 0.040 0.067 0.080 0.079 0.115 0.062 0.048 0.030 0.064 0.044 0.036 0.006 0.021 0.023 0.016 0.033 0.007 0.019 0.020 0.018 0.007 0.016 0.018 0.016 0.009 0.006 With carbon filtration 0.000 0.000 0.005 0.000 0.000 0.001 0.005 - 0.001 0.003 0.006 0.000 0.000 0.000 0.000 -0.001 0.001 0.004 0.000 0.000 0.000 0.000 0.000 0.002 0.003 0.003 0.000 - 0.002 0.000 0.000 * Figures in parentheses show the dilution used in the proposed method.Results by the automated method were supplied by the North West Water Authority, Rivers $ The results for sewage works final effluents obtained by the proposed method were supplied ’Division Laboratory, Dawson House, Warrington. by the North West Water Authority, Southern Division, Hurleston Laboratory.844 Analyst, vol.104 of interference from organic matter. Throughout the initial running of the proposed method, further measurements of absorbance at 275 nm with and without carbon were made, and showed that the high efficiency of removal of organic compounds on the carbon has been maintained. RENNIE, SUMNER AND BASKETTER: DETERMINATION OF NITRATE TABLE IV STATISTICAL ANALYSIS OF RESULTS IN TABLE I11 Sample group Water supply samples and Crewe effluent (low dilution) Number of pairs, n . . . . . . . . . . 18 Mean of differences, Z mgl-1 N .. .. . . 0.009 Tabulated t (0.05 probability) . . . . .. 2.11 Standard deviation of differences, s mg 1-1 N Calculated t* . . . . . . .. . . . . 0.37 . . 0.107 * Calculated t = Z&/s. Sewage works effluents excluding Crewe (high dilution) - 0.105 1.30 0.27 2.23 11 Interference efects To avoid interference from nitrite, sulphamic acid was employed as used in a draft sub- mitted to the Standing Committee of Analysts99l5; the pH was adjusted to 1.8 to eliminate interferences from hydroxyl and carbonate ions.1° To determine the effects of other inter- fering substances on the method, graphs of absorbance at 210 nm against concentration were obtained (using the experimental conditions of the proposed method) of potential water contaminants known to absorb in the region of 210 nm.The effects of certain concentrations of the adventitious contaminants are summarised in Table V and the concentration that gave an absorbance at 210 nm equivalent to 0.02 mg 1-1 N is defined as the interference limit for the proposed method (denoted by an asterisk) in Table V. The lack of an inter- ference effect, compared with that found by direct UV spectrophotometry, of the sodium salt of dodecylbenzenesulphonic acid is due to the effectiveness of removal of organic matter by the filter. Whereas Fez+ and Fe3+ interfere significantly in direct UV methods, such interference is eliminated in the proposed method as a result of the formation of insoluble TABLE V INTERFERENCE EFFECTS Interfering substance r A Species Chloride (Cl-) .. . . .. . . Bromide (Br-) . . .. . . . . . . Iodide (I-) . . .. . . . . . . Iron (Fe2+) . . . . .. .. . . Iron (FeS+) . . .. . . .. . . Dichromate (Cr20,2-) . . . . . . . . Manganese (Mn2+) . . . . . . . . Dodecylbenzenesulphonic acid, sodium salt 1 Concentration/ mg 1-’ 2 500* 5 000 1.0* 2.0 0.65* 1.30 13.5 27 1ooot 1 ooot 0.2 0.4 0.15* 0.30 100 1ooot 1ooot 3.5 7.0 Interference effect/mg 1-’ N With Without carbon filtration carbon filtration 0.02 0.02 0.06 0.06 0.02 0.02 0.04 0.04 0.02 0.02 0.05 0.05 (0.02 (0.02 (0.02 0.08 <0.02 1.4 (0.02 <0.02 <0.02 0.08 (0.02 1.75 0.02 0.02 0.08 0.08 (0.02 <0.02 (0.02 0.4 <0.02 (0.02 (0.02 0.06 (0.02 1.8 A I \ * Indicates “interference limit.” f Indicates highest concentration investigated.September, 1979 IN RAW, POTABLE AND WASTE WATERS BY UV SPECTROPHOTOMETRY 845 hydroxides at the elevated pH, the precipitates being removed by the GF/C paper on the filter.I t is apparent that the interference effects of anions are not reduced by carbon treatment. This effect could indicate that an anion-exchange mechanism is responsible for the retention of nitrate (and other) anions by activated carbon under neutral or acidic conditions. In the presence of alkali, such anion-exchange sites may not be available and so nitrate, chloride, bromide, etc., pass through unhindered, a situation that is analogous to “anion slip” and “regeneration” of anion-exchange resins. Conclusions The precision of the proposed method is satisfactory and the results for water samples, when compared with those obtained by the automated method, show good agreement.The method is less tedious and subject to fewer interferences than methods using chromogenic reagents, and is therefore more suitable for a wide variety of samples. The method is more economical on capital and running costs than an automated method and is therefore particu- larly recommended for small laboratories and low-throughput analyses. The authors thank their colleagues in the North West Water Authority, particularly Messrs. P. Morries, J. Collins and M. D. Nicholson, for useful discussions and for comparative inter-laboratory evaluation work, and Mr. J. Heron of the Freshwater Biological Association, Windermere, and Mrs. L. Newman, of the Water Research Centre, Information Services, for literature work. They are also grateful to Mr. G. Ainsworth, Director of Scientific Services, North West Water Authority and Mr. N. H. Girnson, Manager, Southern Division, North West Water Authority, for permission to publish this paper. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. References Hoather, R. C., and Rackham, R. F., Analyst, 1959, 84, 548. Morries, P., Proc. SOC. Wat. Treat. Exam., 1971, 20, 132. Sekiguchi, K., and Takishima, T., Nippon Kagaku Kaishi, 1975, 4, 642. Sekiguchi, K., Nippon Kagaku Kaishi, 1976, 10, 1553. Soares, M. I. V., Pereira, P. G. S., and Antunes, A. M., Revta Port. Quim., 1971, 13, 151. Wilson, A. L., J . Appl. Chem., 1959, 9, 501. Waters, W. F., Proc. Soc. Wat. Treat. Exam., 1964, 13, 298. “Approved Methods for the Physical and Chemical Examination of Water,” Institution of Water Engineers, London, 1960, p. 21. Cawse, P. A., Analyst, 1967, 92, 311. American Public Health Association, American Water Works Association and Water Pollution Control Federation, “Standard Methods for the Examination of Water and Wastewater,” Fourteenth Edition, American Public Health Association, New York, 1976, p. 420. Brown, L., and Bellinger, E. G., Wat. Res., 1978, 12, 223. Slanina, J., Lingerak, W. A,, and Bergman, Id., 2. Aitalyt. Chem., 1976, 280, 365. “Methods of Chemical Analysis Applied to Sewage and Sewage Effluents,” HM Stationery Office, Cheeseman, I<. V., and Wilson, A . L., “Manual on Analytical Quality Control for the Water “Methods for Examination of Water and Associated Materials. Nitrate in Some Nonsaline Waters London, 1956, p. 22. Industry,” Water Research Centre, Medmenham, 1978, TR66. and Effluents by Direct UV Spectrophotometry,” HM Stationery Office, London, in preparation. Received October 4th, 1978 Accepted April 9th, 1979