- Email: [email protected]
Determination of sulphuric acid in process effluent streams using sequential injection titration
Talanta 52 (2000) 83 – 90 www.elsevier.com/locate/talanta
Determination of sulphuric acid in process effluent streams using sequential injection titration H. du Plessis, J.F. van Staden * Department of Chemistry, Uni6ersity of Pretoria, Pretoria 0002, South Africa Received 19 October 1999; accepted 18 November 1999
Abstract Sulphuric acid in process effluent streams from an electrorefining copper plant was analysed with a sequential injection (SI) titration system using sodium hydroxide as titrant. In the proposed SI titration system a base titrant, acid analyte and base titrant zone were injected sequentially into a distilled water carrier stream in a holding coil and swept by flow reversal through a reaction coil to the detector. The base zones contained bromothymol blue as indicator and the endpoint was monitored spectrophotometrically at 620 nm. The influence of carrier stream flow rate, acid and base zone volumes and titrant concentration on the linear range of the method was studied to obtain an optimum. A linear relationship between peak width and logarithm of the acid concentration was obtained in the range 0.006–0.178 mol l − 1 of H2SO4 for a NaOH concentration of 0.002 mol l − 1. The results obtained for the SI titration of process samples were in good agreement with a standard potentiometric method with an RSD B0.75% and a sample frequency of 23 samples h − 1. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Sequential injection titration; Sulphuric acid; Sodium hydroxide; Process effluent streams
1. Introduction Almost all copper is treated electrolytically during its production from ore. It is electrorefined from impure copper anodes or electrowon from leach/solvent-extraction solutions. A considerable amount of copper scrap is also electrorefined. In Presented at the 10th International Conference on Flow Injection Analysis (ICFIA ‘99), held in Prague, Czech Republic, 20 – 25 June, 1999. * Corresponding author. Fax: + 27-12-3625297. E-mail address: [email protected] (J.F. van Staden)
both electrorefining and electrowinning the electrolyte is an aqueous solution of H2SO4 (150–200 g l − 1, i.e. 1.5–2.0 mol l − 1) and CuSO4 (40–50 g l − 1 of Cu) [1]. Industrial waste water is often discharged into the sewage system. The waste water from the electroplating industry can be quite harmful to the sewer pipes and the municipal treatment plant, because free acids corrode concrete and all non-acid proof elements [2]. The sulphuric acid concentration in the effluent streams must therefore, be continuously monitored to adjust the pH of the waste water before discharge.
0039-9140/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 0 ) 0 0 3 2 2 - 2
84
H. du Plessis, J.F. 6an Staden / Talanta 52 (2000) 83–90
Conventional titrimetric procedures are timeconsuming, labour-intensive and huge amounts of reagents are used. In automatic titrators many of the repetitive operations of conventional titrations are mechanized and microcomputerized, but it is still based on batch operations and is therefore, unsuitable as process analyser. Many different flow injection (FI) titration methods have been developed since the late 1970s [3 – 16] and some have been implemented in process monitoring. One of the big disadvantages of these methods is the high reagent (titrant) consumption. The introduction of sequential injection (SI) titration in 1997 [17] overcame this drawback by replacing the continuous flowing titrant stream of the FI titration system with two titrant zones on either side of the sample zone in a distilled water carrier stream. The system was applied to a monoprotic strong acid, viz. hydrochloric acid. In the current study, it will be expanded to the determination of diprotic sulphuric acid in process effluent streams from a copper plating industry.
2. Experimental
2.1. Apparatus The sequential injection system depicted in Fig. 1 was constructed from the following components: a Gilson Minipuls 3 peristaltic pump (Villiers-le-Bel, France), a 10-port electrically actuated selection valve (Model ECSD10P; Valco Instruments, Housten, TX), and a Unicam 5625
UV–visible spectrophotometer (Cambridge, UK) equipped with a 10 mm Hellma (Mu¨llheim, Baden, Germany) type flow-through cell (volume 80 ml) for absorbance measurements. The absorbance of the carrier stream was monitored at 620 nm due to the blue colour of the bromothymol blue indicator when it is in the basic form. Data acquisition and device control were achieved using a PC30-B interface board (Eagle Electric, Cape Town, South Africa) and an assembled distribution board (MINTEK, Randburg, South Africa). The FlowTEK [18] software package (obtainable from MINTEK) for computer-aided flow analysis was used throughout for device control and data acquisition. Tygon tubing was used in the construction of the sequential injection system. The length and diameter of the tubing used are indicated in Fig. 1. The results obtained with the proposed sequential injection system were compared with the results from a manually performed potentiometric titration. The following apparatus were used in the potentiometric titration: a magnetic stirrer, a digital pH meter (Model 420A, Orion, USA) and a Triode™ pH electrode (Model 91-57, Orion) with Ag/AgCl internal reference system and a built in thermistor for automatic temperature compensation.
2.2. Reagents and solutions All reagents used were of analytical-reagent grade, unless specified otherwise. Deionised water from a Modulab system (Continental Water Sys-
Fig. 1. A schematic diagram of the SI titration system (BTB, bromothymol blue).
H. du Plessis, J.F. 6an Staden / Talanta 52 (2000) 83–90
85
Table 1 Device sequence for one cycle of the sequential injection titration system Time (s)
Pump
Valve
Description
0 5.0 17.0 17.6 21.5 24.5 25.1 29.0 35.0 36.0 37.0 150.0
Off Reverse Off
NaOH
Pump off, select NaOH solution (valve position one) Draw up NaOH solution Pump stop Select H2SO4 sample stream (valve position two to eight) Draw up H2SO4 sample Pump stop Select NaOH stream (valve position nine) Draw up NaOH solution Pump stop Select detector line (valve position ten) Pump stack of zones to detector Pump off, return valve to starting position (valve position one)
Sample Reverse Off NaOH Reverse Off Detector Forward Off
Home
tems, San Antonio, TX) was used to prepare all the aqueous solutions. The carrier stream was deionised degassed water.
2.2.1. Bromothymol blue indicator A stock solution was prepared by dissolving 0.4011 g of indicator (Riedel-de Hae¨n, SeelzeHannover, Germany) in 25 ml of 96% ethanol (Merck) and diluting to 100 ml with deionised water. 2.2.2. NaOH solutions NaOH solutions (0.001, 0.002 and 0.004 mol l − 1), containing 1.6× 10 − 3% of bromothymol blue (i.e. 1 ml of the bromothymol blue stock solution per 250 ml of base), were prepared by appropriate dilution of a 1.005 mol l − 1 NaOH solution (Titrisol standard solution (Merck) standardized against potassium hydrogen phthalate (Merck)). 2.2.3. H2SO4 standards Standards in the range 0.0018 – 0.56 mol l − 1 were prepared by appropriate dilution of a standardized 1.002 mol l − 1 H2SO4 solution. The 1.002 mol l − 1 H2SO4 solution was prepared by appropriate dilution from a 95 – 97% H2SO4 solution (Merck) and it was standardized against the 1.005 mol l − 1 NaOH solution. 2.2.4. Interferent stock solutions A 10 000 mg l − 1 Cu(II) stock solution was
prepared by dissolving 9.8228 g of CuSO4 × 5H2O (Carlo Erba) in 250 ml distilled water. A 10 000 mg l − 1 Ni(II) stock solution was prepared by dissolving 11.9647 g NiSO4 × 7H2O (Merck) in 250 ml distilled water. Working solutions were prepared by suitable dilution of the stock solutions.
2.2.5. Samples Effluent samples from a copper plating industry within the linear range of the method were directly analysed without any sample pretreatment. 2.3. Procedure The device sequence for the SI system is given in Table 1. The sulphuric acid standards or samples were arranged around the valve at valve positions two to eight (Fig. 1) making it ideally suitable for automation and process control. It is possible to analyze eight standards or samples at a time without the need to change the standard or sample after each analysis.
3. Results and discussion As seen previously [17], the dispersion of the acid sample zone in the SI titration system conforms with the ‘one mixing tank’ model, so that
H. du Plessis, J.F. 6an Staden / Talanta 52 (2000) 83–90
86
the physical distance between the equivalent points, teq, is proportional to the logarithm of the sample concentration according to the following equation [3]: teq =
Vm V ln 10 log C 0S − m ln 10 log (CTn) Q Q +
Vm SV ln 10 log Q Vm
teq = k1 log C 0S + k2
(2)
where k1 and k2 are constants. Thus, a linear relationship, with slope k1 and intercept k2, is obtained between teq and the logarithm of the sample concentration.
3.1. Method optimisation (1)
where C 0S is the concentration of the injected sample, CT is the concentration of the titrant, Vm is the volume of the mixing chamber, Q is the flow rate of the carrier stream, SV is the volume of the injected sample and n is the stoichiometric factor of the reacting components. In a system with a constant titrant concentration, a fixed sample and mixing chamber volume and a constant carrier stream flow rate, Eq. (1) can be simplified to
Fig. 2. (a) Influence of carrier stream flow rate on peak width and %RSD for a 0.02 mol l − 1 H2SO4 sample ( , peak width; and , %RSD). (b) Calibration curves at different carrier stream flow rates. H2SO4 mol l − 1, peak width A (relative) ( , 1.70 ml min − 1; , 1.96 ml min − 1; , 2.13 ml min − 1; and , 2.74 ml min − 1).
As can be seen from Eq. (1), there are a lot of factors that will influence the peak width of the injected acid sample zone, namely carrier stream flow rate, acid and base zone volumes and titrant concentration. In the SI titration of hydrochloric acid [17], the optimum parameters were chosen by just considering sensitivity and precision. In this study, it was decided to also investigate how the variation of a parameter will influence the linear range of the SI titration system. The different parameters were varied univariately, starting with the previously determined optimum conditions for the hydrochloric acid titration [17], with the exception of the sodium hydroxide concentration that was doubled to 0.002 mol l − 1 for the diprotic sulphuric acid. After optimisation of a parameter the new optimised value for the parameter was used for the optimisation of the next parameter.
3.1.1. Carrier stream flow rate The draw up times for the different zones were decreased as the flow rate was increased in order to keep the zone volumes constant at the different flow rates. As seen previously [17], the peak width of the recorded peak decreased with an increasing carrier stream flow rate (Fig. 2a), due to less dispersion of the zones at higher flow rates. The best precision (RSD values below 0.2 %) was obtained at flow rates below 2.13 ml min − 1, whereafter it deteriorated at higher flow rates probably due to the back pressure problems that were experienced at high flow rates [19]. In accordance with Eq. (1), the slopes of the calibration curves decreased with increasing flow rate (Fig. 2b). The linear range also decreased at a flow rate of 2.74 ml min − 1. Although the sensitivity was better at lower flow rates with quite the same level of precision, 2.13 ml min − 1 was chosen as optimum because the sample frequency was higher.
H. du Plessis, J.F. 6an Staden / Talanta 52 (2000) 83–90
Fig. 3. (a) Influence of base zone volumes on peak width and %RSD for a 0.032 mol l − 1 H2SO4 sample. Peak width A ( , peak width; and , %RSD). (b) Calibration curves at different base zone volumes. H2SO4 mol l − 1, peak width A (relative) (V(B1), volume base zone 1; and V(B2), volume base zone 2; , V(B1) of 430 ml and V(B2) of 215 ml; , V(B1) of 530 ml and V(B2) of 265 ml; and , V(B1) of 640 ml and V(B2) of 320 ml).
3.1.2. Base zone 6olumes In the optimisation of the base zone volumes, the volume of base zone 1 (the base zone drawn up first) was always two times the volume of base zone 2 (the base zone drawn up after the acid sample zone) so that equal peak heights for base zone 1 and 2 were obtained after dispersion. The residence time of base zone 1 in the holding coil, as well as the distance it had to travel to the detector, was much longer than that for base zone 2, resulting in larger dispersion. From Fig. 3a it is clear that the peak width was not greatly affected by increasing the base zone volumes. The precision was more or less the same for base zone 1 volumes in the region 350 – 640 ml, whereafter it increased slightly at higher volumes. Calibration curves for peak width against the logarithm of the
87
acid concentration were constructed for base zone 1 volumes of 430, 530 and 640 ml. The slopes and intercepts of the different straight lines were constant, but the linear ranges extended to higher concentrations with increasing base zone volumes, i.e. 0.006–0.178 mol l − 1 H2SO4 for 430 ml, 0.010– 0.316 mol l − 1 H2SO4 for 530 ml and 0.010–1.000 mol l − 1 H2SO4 for 640 ml (Fig. 3b). The lower end of the linear range depends on the base zone concentration. At smaller base zone volumes the dispersion is bigger and hence the base zone concentration of the zone lower resulting in lower acid concentrations that can be detected. The height of the surrounding base peaks decreases as the acid sample zone concentration is increased, because more of the base are used in the neutralization reaction. The upper limit of the linear range is reached when the surrounding base peaks are totally consumed in the neutralization reaction. For larger base zone volumes higher acid concentrations can therefore be detected. Although a wider linear range was obtained for a base zone 1 volume of 640 ml, a volume of 430 ml for base zone 1 (215 ml for base zone 2) was chosen as optimum, because the sample frequency decreased with increasing base zone volume.
3.1.3. Acid zone 6olume As can be expected the peak width increased as the acid sample zone volume was increased (Fig. 4a). Precision deteriorated at sample volumes smaller than 90 ml. Calibration curves for peak width against the logarithm of the acid concentration were constructed for acid zone volumes of 70, 105 and 140 ml. In accordance with Eq. (1), the slopes of the different straight lines were constant while the intercept changed (Fig. 4b). The linear range shifted to lower concentrations as the sample volume was increased, i.e. 0.010–0.316 mol l − 1 H2SO4 for a sample volume of 70 ml, 0.006– 0.178 mol l − 1 H2SO4 for a sample volume of 105 ml and 0.003–0.178 mol l − 1 H2SO4 for a sample volume of 140 ml (Fig. 4b). Larger sample volumes use more of the surrounding base zones in the neutralization reaction so that lower acid concentrations can be detected at the lower end of the linear range, but not so high acid concentrations can be detected at the upper end of the linear
88
H. du Plessis, J.F. 6an Staden / Talanta 52 (2000) 83–90
range shifting the linear range to lower concentrations. A sample volume of 105 ml was chosen as optimum considering precision and sample consumption.
3.1.4. NaOH concentration As seen previously [17], the linear range of the SI titration system can easily be altered by just changing the titrant concentration, i.e. 0.0018– 0.100 mol l − 1 H2SO4 for 0.001 mol l − 1 NaOH, 0.006–0.178 mol l − 1 H2SO4 for 0.002 mol l − 1 NaOH and 0.010–0.562 mol l − 1 H2SO4 for 0.004 mol l − 1 NaOH. In accordance with Eq. (1), the sensitivity for the three different titrant concentrations were the same. A NaOH concentration of 0.002 mol l − 1 was chosen to evaluate the method. 3.2. Method e6aluation
Fig. 4. (a) Influence of sample zone volume on peak width and %RSD for a 0.032 mol l − 1 H2SO4 sample. Peak width A ( , peak width; and , %RSD). (b) Calibration curves at different sample zone volumes. H2SO4 mol l − 1, peak width A (relative) ( , 70 ml; , 105 ml and , 140 ml).
Fig. 5. SI titration peaks for different H2SO4 standards and a NaOH concentration of 0.002 mol l − 1. Time axis 8 s per division (112 s full scale). Absorbance units relative, 0.5 per division (5.5 at x-axis to 10.0 at top). The peak width was measured at a relative peak height of 8.80 (A, 0.006; B, 0.010; C, 0.018; D, 0.032; E, 0.056; F, 0.100; and G, 0.178 mol l − 1 H2SO4).
The optimum conditions for the SI titration of diprotic sulphuric acid was identical to that obtained for monoprotic hydrochloric acid [17], except for the titrant concentration that was doubled. It can therefore be concluded that for strong acids a difference in stoichiometric factor does not change the optimum conditions. The optimised SI titration system was critically evaluated with regard to linearity, accuracy, precision, sample frequency and interferences.
3.2.1. Linearity The SI titration peaks for the sulphuric acid calibration standards are shown in Fig. 5. The height of the surrounding base peaks decreased as the acid concentration was increased, because more of the base was used in the neutralization reaction. At acid concentrations above 0.178 mol l − 1 the base zones surrounding the acid sample zone disappeared, because it was totally neutralized by the high acid concentration. The calibration curve for the SI titration system is linear in the range 0.006–0.178 mol l − 1. The relationship between the peak width, teq, and the logarithm of the sulphuric acid concentration is given by teq = 11.62 log C(H2SO4)+ 33.87;
r 2 = 0.9986
The sensitivity for the SI titration of sulphuric acid is slightly higher than that for the SI titration of hydrochloric acid [17].
H. du Plessis, J.F. 6an Staden / Talanta 52 (2000) 83–90
89
Table 2 Comparison of results for process samples determined by SI titration system and a manually performed potentiometric titration method Sample
1 2 3 3 4 4 5 5
Potentiometric titration
SI titration
C(H2SO4) (mol l−1)
C(H2SO4) (mol l−1)
RSD (%) (n = 6)
Recovery (%)
0.02510 0.03651 0.01966 0.01966 0.04034 0.04034 0.00862 0.00862
0.02424 0.03493 0.02134 0.01924 0.04298 0.03760 0.00897 0.00879
0.33 0.44 0.28 0.38 0.19 0.13 0.74 0.61
96.57 95.67 108.55 97.87 106.54 93.21 104.07 101.97
3.2.2. Accuracy and precision The accuracy of the proposed SI titration system was evaluated by comparing its results with that of manually performed potentiometric titrations for five samples. The results revealed a good agreement between the two methods (Table 2). Precision for the sulphuric acid standards and the five samples was better than 0.75% (Tables 2 and 3), which is remarkably better than in the flow injection titration systems where RSD values of between 0.5 and 3% were obtained [3 – 5,15].
titration. A positive interference of 1% on the peak width may look small, but because of the logarithmic relationship between peak width and concentration this translates into an unacceptably large error. The concentration of copper and nickel in the effluent streams from an electroplating plant is usually in the range of 0–20 mg l − 1 for copper and 5–40 mg l − 1 for nickel [2] and would therefore not interfere in the determination of sulphuric acid.
3.2.3. Sample frequency For the more concentrated acid samples it took 155 s to complete one analytical cycle, resulting in a sample frequency of 23 samples h − 1. Although this is much lower than in high-speed FIA titrations [4], the titrant consumption is much lower.
4. Conclusions
3.2.4. Interferences Copper and nickel are main constituents in electroplating solutions and it is therefore, important to determine if they interfere in the determination of the sulphuric acid concentration in the effluent streams. A series of solutions with a fixed sulphuric acid concentration (0.02 mol l − 1), but increasing copper or nickel concentrations was prepared and run on the SI titration system. A positive interference was observed for both copper and nickel at concentrations bigger than 100 mg l − 1 (Table 4), probably due to the formation of the respective metal hydroxides during the
The proposed SI titration system is ideally suited as process analyser for the analysis of sulphuric acid in process effluent streams. It is simple, fully computerized and the titrant consumption is much lower than in FI titrations, in the order of 700 ml of base and 100 ml of acid Table 3 Precision of the proposed SI titration system for the H2SO4 standards C(H2SO4) (mol l−1)
RSD (%) (n = 6)
0.006 0.010 0.018 0.032 0.056 0.100 0.178
0.55 0.13 0.25 0.31 0.21 0.46 0.32
H. du Plessis, J.F. 6an Staden / Talanta 52 (2000) 83–90
90
Table 4 Interference of copper and nickel on the peak width of a 0.02 mol l−1 sulphuric acid standard C(Ni) (mg l−1)
Peak width
% Interference
C(Cu) (mg l−1)
Peak width
% Interference
0 10 20 40 60 100 200 400
13.36 13.38 13.37 13.36 13.40 13.49 13.71 13.95
– 0.15 0.07 0 0.30 0.97 2.62 4.42
0 10 50 100 200 400 600
13.34 13.33 13.35 13.58 13.82 14.57 15.36
– −0.07 0.07 1.80 3.60 9.22 15.14
sample per analysis. The system is also very flexible. The linear range can easily be adjusted to fit specific needs by just changing the acid or base zone volumes or the titrant concentration. Accuracy is comparable with manually executed titrations.
References [1] A.K. Biswas, W.G. Davenport, Extractive Metallurgy of Copper, 3rd edn., Pergamon Press, Oxford, 1994. [2] S.E. Jorgensen, Studies in Environmental Science. Industrial Waste Water Management, vol. 5, Elsevier, Amsterdam, 1979. [3] J. Ru˚zˇicˇka, E.H. Hansen, H. Mosbaek, Anal. Chim. Acta 92 (1977) 235. [4] A.U. Ramsing, J. Ru˚zˇicˇka, E.H. Hansen, Anal. Chim. Acta 129 (1981) 1. [5] D.R. Turner, S. Knox, M. Whitfield, Anal. Proc. 24 (1987) 360.
.
[6] J.F. Tyson, Anal. Proc. 24 (1987) 362. [7] J. Toei, Fresenius Z. Anal. Chem. 330 (1988) 484. [8] G. Nagy, Z. Feher, K. Toth, E. Pungor, A. Ivaska, Talanta 26 (1979) 1143. [9] Z. Fehe´r, I. Kolbe, E. Pungor, Analyst 116 (1991) 483. [10] B. Fuhrmann, U. Spohn, Anal. Chim. Acta 282 (1993) 397. [11] K.G. Schick, Adv. Instrum. 39 (1984) 279. [12] N. Ishibashi, T. Imato, Fresenius Z. Anal. Chem. 323 (1986) 244. [13] J.F. van Staden, Water SA 12 (1986) 93. [14] T. Imato, A. Katafuchi, T. Hamazoe, J. Yagi, K. Takahara, N. Ishibashi, Anal. Sci. 5 (1989) 777. [15] R.H. Taylor, J. Ru˚zˇicˇka, G.D. Christian, Talanta 39 (1992) 285. [16] R.H. Taylor, C. Winbo, G.D. Christian, J. Ru˚zˇicˇka, Talanta 39 (1992) 789. [17] J.F. van Staden, H. du Plessis, Anal. Commun. 34 (1997) 147. [18] G.D. Marshall, J.F. van Staden, Anal. Instrum. 20 (1992) 79. [19] G.D. Marshall, J.F. van Staden, Process Control Qual. 3 (1992) 251.
Determination of sulphuric acid in process effluent streams using sequential injection titration H. du Plessis, J.F. van Staden * Department of Chemistry, Uni6ersity of Pretoria, Pretoria 0002, South Africa Received 19 October 1999; accepted 18 November 1999
Abstract Sulphuric acid in process effluent streams from an electrorefining copper plant was analysed with a sequential injection (SI) titration system using sodium hydroxide as titrant. In the proposed SI titration system a base titrant, acid analyte and base titrant zone were injected sequentially into a distilled water carrier stream in a holding coil and swept by flow reversal through a reaction coil to the detector. The base zones contained bromothymol blue as indicator and the endpoint was monitored spectrophotometrically at 620 nm. The influence of carrier stream flow rate, acid and base zone volumes and titrant concentration on the linear range of the method was studied to obtain an optimum. A linear relationship between peak width and logarithm of the acid concentration was obtained in the range 0.006–0.178 mol l − 1 of H2SO4 for a NaOH concentration of 0.002 mol l − 1. The results obtained for the SI titration of process samples were in good agreement with a standard potentiometric method with an RSD B0.75% and a sample frequency of 23 samples h − 1. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Sequential injection titration; Sulphuric acid; Sodium hydroxide; Process effluent streams
1. Introduction Almost all copper is treated electrolytically during its production from ore. It is electrorefined from impure copper anodes or electrowon from leach/solvent-extraction solutions. A considerable amount of copper scrap is also electrorefined. In Presented at the 10th International Conference on Flow Injection Analysis (ICFIA ‘99), held in Prague, Czech Republic, 20 – 25 June, 1999. * Corresponding author. Fax: + 27-12-3625297. E-mail address: [email protected] (J.F. van Staden)
both electrorefining and electrowinning the electrolyte is an aqueous solution of H2SO4 (150–200 g l − 1, i.e. 1.5–2.0 mol l − 1) and CuSO4 (40–50 g l − 1 of Cu) [1]. Industrial waste water is often discharged into the sewage system. The waste water from the electroplating industry can be quite harmful to the sewer pipes and the municipal treatment plant, because free acids corrode concrete and all non-acid proof elements [2]. The sulphuric acid concentration in the effluent streams must therefore, be continuously monitored to adjust the pH of the waste water before discharge.
0039-9140/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 0 ) 0 0 3 2 2 - 2
84
H. du Plessis, J.F. 6an Staden / Talanta 52 (2000) 83–90
Conventional titrimetric procedures are timeconsuming, labour-intensive and huge amounts of reagents are used. In automatic titrators many of the repetitive operations of conventional titrations are mechanized and microcomputerized, but it is still based on batch operations and is therefore, unsuitable as process analyser. Many different flow injection (FI) titration methods have been developed since the late 1970s [3 – 16] and some have been implemented in process monitoring. One of the big disadvantages of these methods is the high reagent (titrant) consumption. The introduction of sequential injection (SI) titration in 1997 [17] overcame this drawback by replacing the continuous flowing titrant stream of the FI titration system with two titrant zones on either side of the sample zone in a distilled water carrier stream. The system was applied to a monoprotic strong acid, viz. hydrochloric acid. In the current study, it will be expanded to the determination of diprotic sulphuric acid in process effluent streams from a copper plating industry.
2. Experimental
2.1. Apparatus The sequential injection system depicted in Fig. 1 was constructed from the following components: a Gilson Minipuls 3 peristaltic pump (Villiers-le-Bel, France), a 10-port electrically actuated selection valve (Model ECSD10P; Valco Instruments, Housten, TX), and a Unicam 5625
UV–visible spectrophotometer (Cambridge, UK) equipped with a 10 mm Hellma (Mu¨llheim, Baden, Germany) type flow-through cell (volume 80 ml) for absorbance measurements. The absorbance of the carrier stream was monitored at 620 nm due to the blue colour of the bromothymol blue indicator when it is in the basic form. Data acquisition and device control were achieved using a PC30-B interface board (Eagle Electric, Cape Town, South Africa) and an assembled distribution board (MINTEK, Randburg, South Africa). The FlowTEK [18] software package (obtainable from MINTEK) for computer-aided flow analysis was used throughout for device control and data acquisition. Tygon tubing was used in the construction of the sequential injection system. The length and diameter of the tubing used are indicated in Fig. 1. The results obtained with the proposed sequential injection system were compared with the results from a manually performed potentiometric titration. The following apparatus were used in the potentiometric titration: a magnetic stirrer, a digital pH meter (Model 420A, Orion, USA) and a Triode™ pH electrode (Model 91-57, Orion) with Ag/AgCl internal reference system and a built in thermistor for automatic temperature compensation.
2.2. Reagents and solutions All reagents used were of analytical-reagent grade, unless specified otherwise. Deionised water from a Modulab system (Continental Water Sys-
Fig. 1. A schematic diagram of the SI titration system (BTB, bromothymol blue).
H. du Plessis, J.F. 6an Staden / Talanta 52 (2000) 83–90
85
Table 1 Device sequence for one cycle of the sequential injection titration system Time (s)
Pump
Valve
Description
0 5.0 17.0 17.6 21.5 24.5 25.1 29.0 35.0 36.0 37.0 150.0
Off Reverse Off
NaOH
Pump off, select NaOH solution (valve position one) Draw up NaOH solution Pump stop Select H2SO4 sample stream (valve position two to eight) Draw up H2SO4 sample Pump stop Select NaOH stream (valve position nine) Draw up NaOH solution Pump stop Select detector line (valve position ten) Pump stack of zones to detector Pump off, return valve to starting position (valve position one)
Sample Reverse Off NaOH Reverse Off Detector Forward Off
Home
tems, San Antonio, TX) was used to prepare all the aqueous solutions. The carrier stream was deionised degassed water.
2.2.1. Bromothymol blue indicator A stock solution was prepared by dissolving 0.4011 g of indicator (Riedel-de Hae¨n, SeelzeHannover, Germany) in 25 ml of 96% ethanol (Merck) and diluting to 100 ml with deionised water. 2.2.2. NaOH solutions NaOH solutions (0.001, 0.002 and 0.004 mol l − 1), containing 1.6× 10 − 3% of bromothymol blue (i.e. 1 ml of the bromothymol blue stock solution per 250 ml of base), were prepared by appropriate dilution of a 1.005 mol l − 1 NaOH solution (Titrisol standard solution (Merck) standardized against potassium hydrogen phthalate (Merck)). 2.2.3. H2SO4 standards Standards in the range 0.0018 – 0.56 mol l − 1 were prepared by appropriate dilution of a standardized 1.002 mol l − 1 H2SO4 solution. The 1.002 mol l − 1 H2SO4 solution was prepared by appropriate dilution from a 95 – 97% H2SO4 solution (Merck) and it was standardized against the 1.005 mol l − 1 NaOH solution. 2.2.4. Interferent stock solutions A 10 000 mg l − 1 Cu(II) stock solution was
prepared by dissolving 9.8228 g of CuSO4 × 5H2O (Carlo Erba) in 250 ml distilled water. A 10 000 mg l − 1 Ni(II) stock solution was prepared by dissolving 11.9647 g NiSO4 × 7H2O (Merck) in 250 ml distilled water. Working solutions were prepared by suitable dilution of the stock solutions.
2.2.5. Samples Effluent samples from a copper plating industry within the linear range of the method were directly analysed without any sample pretreatment. 2.3. Procedure The device sequence for the SI system is given in Table 1. The sulphuric acid standards or samples were arranged around the valve at valve positions two to eight (Fig. 1) making it ideally suitable for automation and process control. It is possible to analyze eight standards or samples at a time without the need to change the standard or sample after each analysis.
3. Results and discussion As seen previously [17], the dispersion of the acid sample zone in the SI titration system conforms with the ‘one mixing tank’ model, so that
H. du Plessis, J.F. 6an Staden / Talanta 52 (2000) 83–90
86
the physical distance between the equivalent points, teq, is proportional to the logarithm of the sample concentration according to the following equation [3]: teq =
Vm V ln 10 log C 0S − m ln 10 log (CTn) Q Q +
Vm SV ln 10 log Q Vm
teq = k1 log C 0S + k2
(2)
where k1 and k2 are constants. Thus, a linear relationship, with slope k1 and intercept k2, is obtained between teq and the logarithm of the sample concentration.
3.1. Method optimisation (1)
where C 0S is the concentration of the injected sample, CT is the concentration of the titrant, Vm is the volume of the mixing chamber, Q is the flow rate of the carrier stream, SV is the volume of the injected sample and n is the stoichiometric factor of the reacting components. In a system with a constant titrant concentration, a fixed sample and mixing chamber volume and a constant carrier stream flow rate, Eq. (1) can be simplified to
Fig. 2. (a) Influence of carrier stream flow rate on peak width and %RSD for a 0.02 mol l − 1 H2SO4 sample ( , peak width; and , %RSD). (b) Calibration curves at different carrier stream flow rates. H2SO4 mol l − 1, peak width A (relative) ( , 1.70 ml min − 1; , 1.96 ml min − 1; , 2.13 ml min − 1; and , 2.74 ml min − 1).
As can be seen from Eq. (1), there are a lot of factors that will influence the peak width of the injected acid sample zone, namely carrier stream flow rate, acid and base zone volumes and titrant concentration. In the SI titration of hydrochloric acid [17], the optimum parameters were chosen by just considering sensitivity and precision. In this study, it was decided to also investigate how the variation of a parameter will influence the linear range of the SI titration system. The different parameters were varied univariately, starting with the previously determined optimum conditions for the hydrochloric acid titration [17], with the exception of the sodium hydroxide concentration that was doubled to 0.002 mol l − 1 for the diprotic sulphuric acid. After optimisation of a parameter the new optimised value for the parameter was used for the optimisation of the next parameter.
3.1.1. Carrier stream flow rate The draw up times for the different zones were decreased as the flow rate was increased in order to keep the zone volumes constant at the different flow rates. As seen previously [17], the peak width of the recorded peak decreased with an increasing carrier stream flow rate (Fig. 2a), due to less dispersion of the zones at higher flow rates. The best precision (RSD values below 0.2 %) was obtained at flow rates below 2.13 ml min − 1, whereafter it deteriorated at higher flow rates probably due to the back pressure problems that were experienced at high flow rates [19]. In accordance with Eq. (1), the slopes of the calibration curves decreased with increasing flow rate (Fig. 2b). The linear range also decreased at a flow rate of 2.74 ml min − 1. Although the sensitivity was better at lower flow rates with quite the same level of precision, 2.13 ml min − 1 was chosen as optimum because the sample frequency was higher.
H. du Plessis, J.F. 6an Staden / Talanta 52 (2000) 83–90
Fig. 3. (a) Influence of base zone volumes on peak width and %RSD for a 0.032 mol l − 1 H2SO4 sample. Peak width A ( , peak width; and , %RSD). (b) Calibration curves at different base zone volumes. H2SO4 mol l − 1, peak width A (relative) (V(B1), volume base zone 1; and V(B2), volume base zone 2; , V(B1) of 430 ml and V(B2) of 215 ml; , V(B1) of 530 ml and V(B2) of 265 ml; and , V(B1) of 640 ml and V(B2) of 320 ml).
3.1.2. Base zone 6olumes In the optimisation of the base zone volumes, the volume of base zone 1 (the base zone drawn up first) was always two times the volume of base zone 2 (the base zone drawn up after the acid sample zone) so that equal peak heights for base zone 1 and 2 were obtained after dispersion. The residence time of base zone 1 in the holding coil, as well as the distance it had to travel to the detector, was much longer than that for base zone 2, resulting in larger dispersion. From Fig. 3a it is clear that the peak width was not greatly affected by increasing the base zone volumes. The precision was more or less the same for base zone 1 volumes in the region 350 – 640 ml, whereafter it increased slightly at higher volumes. Calibration curves for peak width against the logarithm of the
87
acid concentration were constructed for base zone 1 volumes of 430, 530 and 640 ml. The slopes and intercepts of the different straight lines were constant, but the linear ranges extended to higher concentrations with increasing base zone volumes, i.e. 0.006–0.178 mol l − 1 H2SO4 for 430 ml, 0.010– 0.316 mol l − 1 H2SO4 for 530 ml and 0.010–1.000 mol l − 1 H2SO4 for 640 ml (Fig. 3b). The lower end of the linear range depends on the base zone concentration. At smaller base zone volumes the dispersion is bigger and hence the base zone concentration of the zone lower resulting in lower acid concentrations that can be detected. The height of the surrounding base peaks decreases as the acid sample zone concentration is increased, because more of the base are used in the neutralization reaction. The upper limit of the linear range is reached when the surrounding base peaks are totally consumed in the neutralization reaction. For larger base zone volumes higher acid concentrations can therefore be detected. Although a wider linear range was obtained for a base zone 1 volume of 640 ml, a volume of 430 ml for base zone 1 (215 ml for base zone 2) was chosen as optimum, because the sample frequency decreased with increasing base zone volume.
3.1.3. Acid zone 6olume As can be expected the peak width increased as the acid sample zone volume was increased (Fig. 4a). Precision deteriorated at sample volumes smaller than 90 ml. Calibration curves for peak width against the logarithm of the acid concentration were constructed for acid zone volumes of 70, 105 and 140 ml. In accordance with Eq. (1), the slopes of the different straight lines were constant while the intercept changed (Fig. 4b). The linear range shifted to lower concentrations as the sample volume was increased, i.e. 0.010–0.316 mol l − 1 H2SO4 for a sample volume of 70 ml, 0.006– 0.178 mol l − 1 H2SO4 for a sample volume of 105 ml and 0.003–0.178 mol l − 1 H2SO4 for a sample volume of 140 ml (Fig. 4b). Larger sample volumes use more of the surrounding base zones in the neutralization reaction so that lower acid concentrations can be detected at the lower end of the linear range, but not so high acid concentrations can be detected at the upper end of the linear
88
H. du Plessis, J.F. 6an Staden / Talanta 52 (2000) 83–90
range shifting the linear range to lower concentrations. A sample volume of 105 ml was chosen as optimum considering precision and sample consumption.
3.1.4. NaOH concentration As seen previously [17], the linear range of the SI titration system can easily be altered by just changing the titrant concentration, i.e. 0.0018– 0.100 mol l − 1 H2SO4 for 0.001 mol l − 1 NaOH, 0.006–0.178 mol l − 1 H2SO4 for 0.002 mol l − 1 NaOH and 0.010–0.562 mol l − 1 H2SO4 for 0.004 mol l − 1 NaOH. In accordance with Eq. (1), the sensitivity for the three different titrant concentrations were the same. A NaOH concentration of 0.002 mol l − 1 was chosen to evaluate the method. 3.2. Method e6aluation
Fig. 4. (a) Influence of sample zone volume on peak width and %RSD for a 0.032 mol l − 1 H2SO4 sample. Peak width A ( , peak width; and , %RSD). (b) Calibration curves at different sample zone volumes. H2SO4 mol l − 1, peak width A (relative) ( , 70 ml; , 105 ml and , 140 ml).
Fig. 5. SI titration peaks for different H2SO4 standards and a NaOH concentration of 0.002 mol l − 1. Time axis 8 s per division (112 s full scale). Absorbance units relative, 0.5 per division (5.5 at x-axis to 10.0 at top). The peak width was measured at a relative peak height of 8.80 (A, 0.006; B, 0.010; C, 0.018; D, 0.032; E, 0.056; F, 0.100; and G, 0.178 mol l − 1 H2SO4).
The optimum conditions for the SI titration of diprotic sulphuric acid was identical to that obtained for monoprotic hydrochloric acid [17], except for the titrant concentration that was doubled. It can therefore be concluded that for strong acids a difference in stoichiometric factor does not change the optimum conditions. The optimised SI titration system was critically evaluated with regard to linearity, accuracy, precision, sample frequency and interferences.
3.2.1. Linearity The SI titration peaks for the sulphuric acid calibration standards are shown in Fig. 5. The height of the surrounding base peaks decreased as the acid concentration was increased, because more of the base was used in the neutralization reaction. At acid concentrations above 0.178 mol l − 1 the base zones surrounding the acid sample zone disappeared, because it was totally neutralized by the high acid concentration. The calibration curve for the SI titration system is linear in the range 0.006–0.178 mol l − 1. The relationship between the peak width, teq, and the logarithm of the sulphuric acid concentration is given by teq = 11.62 log C(H2SO4)+ 33.87;
r 2 = 0.9986
The sensitivity for the SI titration of sulphuric acid is slightly higher than that for the SI titration of hydrochloric acid [17].
H. du Plessis, J.F. 6an Staden / Talanta 52 (2000) 83–90
89
Table 2 Comparison of results for process samples determined by SI titration system and a manually performed potentiometric titration method Sample
1 2 3 3 4 4 5 5
Potentiometric titration
SI titration
C(H2SO4) (mol l−1)
C(H2SO4) (mol l−1)
RSD (%) (n = 6)
Recovery (%)
0.02510 0.03651 0.01966 0.01966 0.04034 0.04034 0.00862 0.00862
0.02424 0.03493 0.02134 0.01924 0.04298 0.03760 0.00897 0.00879
0.33 0.44 0.28 0.38 0.19 0.13 0.74 0.61
96.57 95.67 108.55 97.87 106.54 93.21 104.07 101.97
3.2.2. Accuracy and precision The accuracy of the proposed SI titration system was evaluated by comparing its results with that of manually performed potentiometric titrations for five samples. The results revealed a good agreement between the two methods (Table 2). Precision for the sulphuric acid standards and the five samples was better than 0.75% (Tables 2 and 3), which is remarkably better than in the flow injection titration systems where RSD values of between 0.5 and 3% were obtained [3 – 5,15].
titration. A positive interference of 1% on the peak width may look small, but because of the logarithmic relationship between peak width and concentration this translates into an unacceptably large error. The concentration of copper and nickel in the effluent streams from an electroplating plant is usually in the range of 0–20 mg l − 1 for copper and 5–40 mg l − 1 for nickel [2] and would therefore not interfere in the determination of sulphuric acid.
3.2.3. Sample frequency For the more concentrated acid samples it took 155 s to complete one analytical cycle, resulting in a sample frequency of 23 samples h − 1. Although this is much lower than in high-speed FIA titrations [4], the titrant consumption is much lower.
4. Conclusions
3.2.4. Interferences Copper and nickel are main constituents in electroplating solutions and it is therefore, important to determine if they interfere in the determination of the sulphuric acid concentration in the effluent streams. A series of solutions with a fixed sulphuric acid concentration (0.02 mol l − 1), but increasing copper or nickel concentrations was prepared and run on the SI titration system. A positive interference was observed for both copper and nickel at concentrations bigger than 100 mg l − 1 (Table 4), probably due to the formation of the respective metal hydroxides during the
The proposed SI titration system is ideally suited as process analyser for the analysis of sulphuric acid in process effluent streams. It is simple, fully computerized and the titrant consumption is much lower than in FI titrations, in the order of 700 ml of base and 100 ml of acid Table 3 Precision of the proposed SI titration system for the H2SO4 standards C(H2SO4) (mol l−1)
RSD (%) (n = 6)
0.006 0.010 0.018 0.032 0.056 0.100 0.178
0.55 0.13 0.25 0.31 0.21 0.46 0.32
H. du Plessis, J.F. 6an Staden / Talanta 52 (2000) 83–90
90
Table 4 Interference of copper and nickel on the peak width of a 0.02 mol l−1 sulphuric acid standard C(Ni) (mg l−1)
Peak width
% Interference
C(Cu) (mg l−1)
Peak width
% Interference
0 10 20 40 60 100 200 400
13.36 13.38 13.37 13.36 13.40 13.49 13.71 13.95
– 0.15 0.07 0 0.30 0.97 2.62 4.42
0 10 50 100 200 400 600
13.34 13.33 13.35 13.58 13.82 14.57 15.36
– −0.07 0.07 1.80 3.60 9.22 15.14
sample per analysis. The system is also very flexible. The linear range can easily be adjusted to fit specific needs by just changing the acid or base zone volumes or the titrant concentration. Accuracy is comparable with manually executed titrations.
References [1] A.K. Biswas, W.G. Davenport, Extractive Metallurgy of Copper, 3rd edn., Pergamon Press, Oxford, 1994. [2] S.E. Jorgensen, Studies in Environmental Science. Industrial Waste Water Management, vol. 5, Elsevier, Amsterdam, 1979. [3] J. Ru˚zˇicˇka, E.H. Hansen, H. Mosbaek, Anal. Chim. Acta 92 (1977) 235. [4] A.U. Ramsing, J. Ru˚zˇicˇka, E.H. Hansen, Anal. Chim. Acta 129 (1981) 1. [5] D.R. Turner, S. Knox, M. Whitfield, Anal. Proc. 24 (1987) 360.
.
[6] J.F. Tyson, Anal. Proc. 24 (1987) 362. [7] J. Toei, Fresenius Z. Anal. Chem. 330 (1988) 484. [8] G. Nagy, Z. Feher, K. Toth, E. Pungor, A. Ivaska, Talanta 26 (1979) 1143. [9] Z. Fehe´r, I. Kolbe, E. Pungor, Analyst 116 (1991) 483. [10] B. Fuhrmann, U. Spohn, Anal. Chim. Acta 282 (1993) 397. [11] K.G. Schick, Adv. Instrum. 39 (1984) 279. [12] N. Ishibashi, T. Imato, Fresenius Z. Anal. Chem. 323 (1986) 244. [13] J.F. van Staden, Water SA 12 (1986) 93. [14] T. Imato, A. Katafuchi, T. Hamazoe, J. Yagi, K. Takahara, N. Ishibashi, Anal. Sci. 5 (1989) 777. [15] R.H. Taylor, J. Ru˚zˇicˇka, G.D. Christian, Talanta 39 (1992) 285. [16] R.H. Taylor, C. Winbo, G.D. Christian, J. Ru˚zˇicˇka, Talanta 39 (1992) 789. [17] J.F. van Staden, H. du Plessis, Anal. Commun. 34 (1997) 147. [18] G.D. Marshall, J.F. van Staden, Anal. Instrum. 20 (1992) 79. [19] G.D. Marshall, J.F. van Staden, Process Control Qual. 3 (1992) 251.