FOGG, CHAMS1 AND ABDALLA 465 - Delay coil (a) : Waste 4 m x Glassy Sample 0.58 mm carbon Pump injection detector Cadmium sponge column : Waste 10cm x detector 1 mm cadmium wire (c) -t- : Waste 0.58 mm carbon Pump 'ample 11.5 cm x detector injection 1-14 mm tube Fig. 1. Flow injection systctns uscd: (a) batchwise reduction niethod ; (b) continuous reduction method ; ancl (c) direct reduction mcthotl. Reagents Standard sodium nitrate solution, 1 x 10-1 M. Dissolve 2.13 g of sodium nitrate in water and dilute to 250 ml in a calibrated flask. Staizdard sodizrm ititritc solzitioit, 1 x 10-1 M. Dissolve 1.73 g of sodium nitrite in water and dilute to 250 ml in a calibrated flask. Acidic bromide eluent. Dissolve 100 g of potassium bromide in 350 ml of water, add 138 ml of concentrated hydrochloric acid, cool the solution, dilute to 500 ml and mix.Hydrochloric acid solzition, 0.3 M. Dilute 13.8 ml of concentrated hydrochloric acid to 500 ml with water. Spongy cadmiztm. Place one or two zinc rods (analytical-reagent grade, approximately 10 g total) in 250 ml of 20% m/V cadmium sulphate solution and allow to stand for 3 h at room temperature, moving the rods around periodically. Remove the rods, decant off the super- natant liquid from the deposited cadmium and wash the cadmium twice with distilled water. Cover the granules with water and macerate them for 10-15 s to produce a uniform fine size. Store under water. Prepare more dilute standard solutions by dilution. Prepare more dilute standard solutions by dilution. Preparation of the Reductor Column An Econo-Column (Bio-Rad Laboratories) was used in this work.Make a small mark on the side of the column about 0.5 cm above the spongy cadmium. Allow distilled water to flow through the column and store with the cadmium sponge under water. Before use, activate the cadmium sponge column by passing 0.3 M hydrochloric acid solution through it, and then wash with water again. Bring the water level down to the 0.5-cm mark. A suitable column for packing is about 10 cm long with an inner diameter of 7 mm. Fill the reductor column to about 7 cm with the spongy cadmium. Procedure Place a 10-ml calibrated flask below the column to collect emerging solution. Transfer slowly by pipette 1 ml of standard nitrate solution (1 x l O - S l x 10-1 M) to the top of the cadmium column, placing the tip of the pipette near the top of the cadmium bed in a manner such that the bed is not disturbed.Allow the solution to run through the column and into the466 FOGG et al. : FLOW INJECTION VOLTAMMETRY Analyst, Vol. 108 flask until the level of solution reaches the mark above the bed, then add about 2 ml of water at a time until the solution in the calibrated flask reaches the 10-ml mark. Mix, inject 25 p1 of the solution into acidic bromide eluent and note the reduction signal at the glassy carbon electrode. A 4-m delay coil and a flow-rate of 5 ml min-1 were used in this work. Determination of Nitrate using Continuous Reduction to Nitrite The reagents and column used in this study were the same as those used above for the batch- wise reduction of nitrate to nitrite.The only difference in this method was that the reductor column was connected to the sample injection valve such that solution passed through the sample loop of the flow injection valve after passing through the reductor. Sample solution was passed continuously through this system and the nitrite concentration in the sample solution was monitored at several stages by injecting the contents of the loop into the acidic bromide eluent in the flow injection voltammetric system. The system is shown in Fig. l(b). Determination of Nitrate by Direct Injection with a Flow Injection System Incorporating a Cadmium Wire Reductor Analytical-reagent grade cadmium wire of 1 mm diameter was used. In the procedure recommended, a 10-cm length of cadmium wire is inserted into an 11.5-cm length of PTFE tube (i.d.1.14 mm). This tube is then connected between the sample injection valve and the delay coil. The cadmium wire has to be reactivated as required and the frequency at which this has to be done depends on the amount of reducible material that has passed over it. The wire is taken out of the PTFE tube and washed vigorously with dilute hydrochloric acid before being replaced in the system. Aliquots (25 pl) of nitrate solution are injected directly into the system. A delay coil length of 4 m and a flow-rate of 5 ml min-l were used in this work. The flow injection system used is shown in Fig. l(c). Other reagents were prepared as described above. TABLE I BATCHWISE REDUCTION OF NITRATE : EFFECT OF LENGTH OF CADMIUM SPONGE COLUMN Nitrate and nitrite concentrations before dilution in the column = 1 x M.Length of reductor column/cm . . .. . . 2.8 7.0 12.8 Signal for nitrate/pA . . 1 . .. . . 0.91 1.31 1.13 Signal for nitrite/pA . . .. . . . . 1.87 1.68 1.43 Signal for nitrite without reductor/pA . . . . 1.87 1.87 1.87 Results In the preliminary study of the batchwise reduction of nitrate solutions the effect of the length of the spongy cadmium column was studied first (see Table I ) ; control tests were also carried out using nitrite solutions to see the effect of the reductor on nitrite. Optimum reduc- tion of nitrate at the 1 x Clearly nitrite itself is being extensively reduced on longer columns. The effect of flow-rate through the reductor column was not critical at this level of nitrate, the current signals obtained being 1.17 and 1.15 pA at 2 and 6 ml min-l, respectively. The extents of reduction at various levels of nitrate and nitrite from 1 x to 1 x 10-1 M are compared in Table 11.The current signal M level is effected with a 7-cm column. TABLE I1 BATCHWISE REDUCTION OF NITRATE : SIGNAL SIZE AT VARIOUS PRE-DILUTION CONCENTRATION LEVELS Pre-dilution concentration of nitrite or nitrate/M f 1 o 1 x 10-5 1 x 10-4 1 x 10-3 1 x 10-2 1 x 10-1 Signal for nitrate/pA . . .. . . 0,026 0.040 0.200 1.63 10.8 28.0 Signal for nitrite/pA . . . I . . 0.026 0.041 0.212 1.80 16.0 93.3 Signal for nitrite without reductor/pA. . 0.026 0.041 0.217 2.13 16.2 93.3April, 1983 OF NITRATE AFTER REDUCTION TO NITRITE 467 values given for nitrite that had not been passed through the column (but had been similarly diluted) show the near rectilinearity of the nitrite signal up to about the 1 x loA3 M nitrite level.A reduction yield of SO-90~o can be achieved for nitrate at the 1 x loA3 M level; the yield falls rapidly at higher concentrations. 1 x M and 0.05 x 10-3-1 x 10-3 M were rectilinear. Signals obtained at the latter level are shown in Fig. 2; coefficients of variation (3-10 determinations) are typically <2.5% for sequential injections in this procedure and in the other procedures. Calibration graphs in the ranges 0.05 x C i Time d Fig. 2. Batchwise reduction of nitrate: calibration signals obtained a t the pre- dilution 0-1 x 10-3 M level of nitrate. Pre-dilution nitrate concentration : A, 0; B, 0.2; C, 0.4; D, 0.6; E, 0.8; and F, 1.0 x 10-3 M.The effect of the volume of sample solution that has passed through the reductor (and the sample injection loop) in the continuous reduction method at the 1 x 10-5 M level is shown in Table 111; the values given are for single runs. A reasonably steady value is attained after about 15 ml of sample solution has passed; there is a further sliglit increase up to about 50-100 ml and a slight decrease is obtained after about 130 ml. When nitrite at tlie same level is passed tlirough the system a similar signal to that of nitrate is observed. The signals obtained when 25 ml of solution have passed at levels of nitrate from 1 x M are shown in Fig. 3. An approach to rectilinearity is apparent up to about 1 x In a preliminary study tlie use of an eluent 3.2 >I in hydrochloric acid and 20% m/V in potassium bromide, tvhich is optimum for the determination of nitrite by direct injection,l was confirmed as being optimum also when nitrate was injected directly and reduced on-line.The effect of the length of cadmium wire used is shown in Table I V ; the optimum length is 10 cm. to 1 x M. TABLE 111 CONTINUOTJS REDUCTION METHOD : SIGNAL OBTAINED AT VARIOUS SAMPLE VOLUMES PASSED THROUGH COLUMN Nitrate and nitrite concentrations = 1 x M. Volume passed/ml . . . . .. 5 15 25 35 50 100 225 Signal for nitrate/pA . . . . 0.156 0.184 0.186 0.190 0.202 0.212 0.187 Signal for nitrite/pA . . . . 0.158 0.164 0.172 0.169 0.182 0.194 0.222468 FOGG et al.: FLOW INJECTION VOLTAMMETRY Annlyst, Vol.108 Time d Fig. 3. Continuous reduction method : signals obtained a t various nitrate con- centration levels: A, 0; B, 1 x C, 1 x 10-5; D, 1 x 10-4; E, 1 x and F, 1 x lo-' M. The decrease in signal with longer lengths is probably due to increased dispersion or possibly further reduction of nitrite. The signals obtained for various levels of nitrate and nitrite injected into the recommended system are given in Table V. These are compared with the nitrite signal attained in the absence of the cadmium wire. The yield of nitrate is seen to vary from 18% at the 1 x M level. The size of the signal is increased by increasing the residence time by using a slower flow-rate; the signal varied from 0.71 pA at 1.5 ml min-1 to 0.35 pA at 8 ml min-1. Signals obtained in the range 0.1 x lOw4-l x M level to 4% at the 1 x M are shown in Fig.4. TABLE IV DIRECT INJECTION METHOD : EFFECT OF LENGTH OF CADMIUM WIRE ON SIGNAL Nitrate concentration = 1 x M. Length of wirelcm . . .. 0 2 5 10 20 40 Signal/pA . . .. . . 0.015 0.087 0.238 0.381 0.365 0.341 Discussion Nitrite sample solutions can be injected into an acidic bromide eluent in a flow injection system and determined voltammetrically at a glassy carbon electrode by the reduction signal of the nitrosyl bromide pr0duced.l In this work the extension of this method to the determina- tion of nitrate has been studied. Commonly, nitrate is reduced to nitrite and determined by visible spectrophotometry using the diazotisation properties of the nitrite. The first procedure given here simply illustrates the use of the flow injection voltammetric method previously TABLE V DIRECT INJECTION METHOD : SIGNAL SIZE AT VARIOUS COT 1' CENTRATION LEVELS Length of cadmium wire = 10 cm.Equivalent concentration of nitrate or nitrite/M f A 7 0 1 x 10-6 1 x 10-5 1 x 10-4 1 x 10-3 i x 10-2 Signal for nitrate/pA . . .. . . 0.0198 0.0198 0.0234 0.115 0.474 2.038 Signal for nitrite/pA . . .. . . 0.0198 0.0300 0.122 0.734 4.70 25.8 Reduction yield of nitrite from Signal for nitrite without wire/pA . . 0.0198 0.0300 0.126 0.833 5.55 34.7 nitrate, % . . .. .. .. - - 18 15 10 4April, 1983 OF NITRATE AFTER REDUCTION TO NITRITE 469 I , I Time Fig. 4. Direct injection method: calibration signals obtained in the range 0.1-1 x 10-4 M nitrate: A, 0; B, 0.2; C, 0.4; D, 0.6; E, 0.8; and F, 1.0 x 1 0 - 4 ~ .published in the determination of nitrate after it has been reduced first to nitrite by means of a cadmium sponge column. Clearly cadmium ions introduced into the sample during reduction do not interfere and most other reagents commonly used to reduce nitrate to nitrite would not interfere with the determination. Flow injection analysis has so far been little used for the intermittent analysis of sample streams or of large solution samples in which the determined concentration may be varying. If part of the sample stream is allowed to flow through the sample loop of the flow injection valve, the solution can be sampled and the determinand determined as required. Alterna- tively, and possibly more satisfactorily, sample can be pumped from the stream, or from a large solution sample, as required through the sample loop at specified times to allow sampling and determination.Suitable laboratory applications might include solution kinetic studies and tablet dissolution or drug availability studies. This study has suggested an application in which nitrate sample may be pumped at pre-determined times through a reductor column and sample loop, and may then be sampled and the resulting nitrite determined in the flow injection system. The preliminary work carried out so far indicates that the only difficulty would be in maintaining the level of activity of the reductor column. The introduction of an occasional reactivation cycle may solve this problem. The ideal procedure, however, would involve direct injection of the nitrate sample solution into the flow injection system.Results obtained using a cadmium wire to reduce nitrate on-line have been reported here. Yields of nitrite with this system have been low (<20%) and depend on the condition of the surface of the cadmium wire. Nevertheless, the procedure as developed so far may have applications in areas where nitrite levels need to be monitored with only low precision. By means of valves before and after the cadmium wire it may prove to be possible to regenerate the cadmium wire in situ. The results presented here indicated three possible approaches to the determination of nitrate using a flow injection voltammetric method for nitrite. Further work is planned to develop these systems for automatic use, but the present results will allow analytical chemists experienced in particular application areas to assess whether these approaches hold any advantages for them.Full experimental details have been given so that other workers can reproduce our conditions and results exactly before making any modification to suit their particular requirements. A standardisation cycle would also be needed. A.Y.C. thanks the Lebanese University for leave of absence and the Lebanese Government M.A.A. thanks the University of Khartoum for leave of absence and The authors thank Mr. I. W. Burns and Mr. G. M. for financial support. The British Council for financial support. Telling of Unilever Research for helpful discussions. References 1. 2. Fogg, A. G., Bsebsu, N.K., and Abdalla, M. A., Analyst, 1982, 107, 1040. Fogg, A. G., Bsebsu, N. K., and Abdalla, M. A., Analyst, 1982, 107, 1462. Received November lst, 1982 Accepted December 21st, 1982April, 1983 OF NITRATE AFTER REDUCTION TO NITRITE 469 I , I Time Fig. 4. Direct injection method: calibration signals obtained in the range 0.1-1 x 10-4 M nitrate: A, 0; B, 0.2; C, 0.4; D, 0.6; E, 0.8; and F, 1.0 x 1 0 - 4 ~ . published in the determination of nitrate after it has been reduced first to nitrite by means of a cadmium sponge column. Clearly cadmium ions introduced into the sample during reduction do not interfere and most other reagents commonly used to reduce nitrate to nitrite would not interfere with the determination. Flow injection analysis has so far been little used for the intermittent analysis of sample streams or of large solution samples in which the determined concentration may be varying.If part of the sample stream is allowed to flow through the sample loop of the flow injection valve, the solution can be sampled and the determinand determined as required. Alterna- tively, and possibly more satisfactorily, sample can be pumped from the stream, or from a large solution sample, as required through the sample loop at specified times to allow sampling and determination. Suitable laboratory applications might include solution kinetic studies and tablet dissolution or drug availability studies. This study has suggested an application in which nitrate sample may be pumped at pre-determined times through a reductor column and sample loop, and may then be sampled and the resulting nitrite determined in the flow injection system.The preliminary work carried out so far indicates that the only difficulty would be in maintaining the level of activity of the reductor column. The introduction of an occasional reactivation cycle may solve this problem. The ideal procedure, however, would involve direct injection of the nitrate sample solution into the flow injection system. Results obtained using a cadmium wire to reduce nitrate on-line have been reported here. Yields of nitrite with this system have been low (<20%) and depend on the condition of the surface of the cadmium wire. Nevertheless, the procedure as developed so far may have applications in areas where nitrite levels need to be monitored with only low precision. By means of valves before and after the cadmium wire it may prove to be possible to regenerate the cadmium wire in situ. The results presented here indicated three possible approaches to the determination of nitrate using a flow injection voltammetric method for nitrite. Further work is planned to develop these systems for automatic use, but the present results will allow analytical chemists experienced in particular application areas to assess whether these approaches hold any advantages for them. Full experimental details have been given so that other workers can reproduce our conditions and results exactly before making any modification to suit their particular requirements. A standardisation cycle would also be needed. A.Y.C. thanks the Lebanese University for leave of absence and the Lebanese Government M.A.A. thanks the University of Khartoum for leave of absence and The authors thank Mr. I. W. Burns and Mr. G. M. for financial support. The British Council for financial support. Telling of Unilever Research for helpful discussions. References 1. 2. Fogg, A. G., Bsebsu, N. K., and Abdalla, M. A., Analyst, 1982, 107, 1040. Fogg, A. G., Bsebsu, N. K., and Abdalla, M. A., Analyst, 1982, 107, 1462. Received November lst, 1982 Accepted December 21st, 1982