Talanta 52 (2000) 91 – 99 www.elsevier.com/locate/talanta Titration of strong and weak acids by sequential injection analysis technique Silla Maskula, Johan Nyman 1, Ari Ivaska * A, bo Akademi Uni6ersity Process Chemistry Group, c/o Laboratory of Analytical Chemistry, Biskopsgatan 8, FIN 20500 Turku-A, bo, Finland Received 25 October 1999; received in revised form 2 February 2000; accepted 8 February 2000 Abstract A sequential injection analysis (SIA) titration method has been developed for acid-base titrations. Strong and weak acids in different concentration ranges have been titrated with a strong base. The method is based on sequential aspiration of an acidic sample zone and only one zone of the base into a carrier stream of distilled water. On their way to the detector, the sample and the reagent zones are partially mixed due to the dispersion and thereby the base is partially neutralised by the acid. The base zone contains the indicator. An LED-spectrophotometer is used as detector. It senses the colour of the unneutralised base and the signal is recorded as a typical SIA peak. The peak area of the unreacted base was found to be proportional to the logarithm of the acid concentration. Calibration curves with good linearity were obtained for a strong acid in the concentration ranges of 10 − 4 –10 − 2 and 0.1– 3 M. Automatic sample dilution was implemented when sulphuric acid at concentration of 6 – 13 M was titrated. For a weak acid, i.e. acetic acid, a linear calibration curve was obtained in the range of 3 ×10 − 4 –8 ×10 − 2 M. By changing the volumes of the injected sample and the reagent, different acids as well as different concentration ranges of the acids can be titrated without any other adjustments in the SIA manifold or the titration protocol. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Titration; Strong acid; Weak acid; Sequential injection analysis 1. Introduction Interest for the automatic titrations arise from the fact that many of the acid – base titrations in the process industry are still performed manually, * Corresponding author. Tel.: + 358-2-2154420. E-mail address: ari.ivaska@abo.fi (A. Ivaska) 1 Present address: Raisio Chemicals, PO Box 101, FIN21201 Raisio, Finland. i.e. the control of the process solution concentrations takes place off-line in a laboratory. Automation of these analytical procedures would shorten the time between sampling and obtaining the results. Possibility for human errors in the analytical work will be diminished and reduction of reagents used and waste produced will also be obtained. Possible techniques, which can be used for automatisation of titration procedures, are the sequential injection analysis (SIA) and its parent 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 3 - 4 92 S. Maskula et al. / Talanta 52 (2000) 91–99 technique flow injection analysis (FIA). For industrial purposes SIA titration presents a more suitable alternative, because the FIA titration methods have usually too high reagent consumption for on-line applications. FIA technique does not either offer the robustness and reliability needed in industrial processes. For instance, the use of peristaltic pumps in FIA instruments causes elongation of the tubing, which requires more frequent maintenance for the apparatus. SIA instruments are also totally computer-controlled, which enables their connection to other process controlling devices and fast modification of measurement protocols if needed. The FIA titration technique itself is well established and has been used in several applications [1]. In classical acid– base FIA titration, a sample of acid is injected into a carrier stream of base, which contains an indicator. On its way to the detector the sample is dispersed into the carrier stream and thereby gradually neutralised in both the leading and tailing edges of the sample zone. In both of these edges, an element of fluid exists where the amount of acid is equivalent with the amount of base. Peak width between these two equivalence points is proportional to the logarithm of the sample concentration. The same procedure has also been adapted for SIA titration in order to reduce the reagent and sample consumption [2]. In that study the continuous base stream was replaced by two base zones on each side of acidic sample zone in a carrier stream of distilled water, and the obtained peak width was used for determination of the logarithm of sample concentration. In this paper we will describe a SIA titration method, where a sample zone of acid and only one reagent zone of base are sequentially injected into a carrier stream of distilled water. The injected zones are first stacked into the holding coil and when the flow is reversed, they are dispersed, partially into each other and partially into the carrier stream on their way to the detector. The base zone contains an indicator, which colour is detected with an LED-spectrophotometer. The concentration gradients, which are formed, can be thought to consist of small individual elements of fluid, each of them having a slightly different concentration as the neighbouring ones. These small volume elements of reagent (base) are mixed with small volume elements of the sample (acid). When the flow pattern is kept constant, the ‘number’ of reagent (base) fluid elements, which are neutralised, depends on the initial concentration of the sample. At the same time, as the base in a fluid element is neutralised, the colour of the acid-base indicator in that particular element is also changed. The detector senses the colour of the unneutralised base and the signal is recorded as a typical SIA peak. When the initial concentration of the acid increases more base fluid elements are neutralised and the width of the peak decreases. At the same time, the area of the peak also decreases. The observed peak area, which corresponds to the amount of unreacted base, was experimentally found to be proportional to the logarithm of acid concentration. As long as the manifold, reagent concentration, the injected volumes and the flow rates are maintained constant all the samples undergo a similar dispersion and the obtained peak areas can be compared with each other. 2. Experimental 2.1. Chemicals All chemicals used were of p.a. grade and obtained from Merck except the a-naphtholbenzein, which was obtained from Fluka. 96% ethanol was obtained from Primalco Oy, Finland. Distilled, deionized water (Millipore) without degassing was used throughout the experiments. It was also used as the carrier solution. The Bromthymol Blue (BTB) dye solution was prepared by dissolving 0.40 g of the indicator in 25 ml of 96% ethanol and making the final volume to 100 ml with 10 − 2 M sodium tetraborate solution. The a-naphtholbenzein dye solution was prepared by dissolving 0.10 g of the indicator in 100 ml 96% ethanol. 2.2. Standards NaOH solutions were prepared by appropriate dilutions of 4.89 M stock solution, standardised S. Maskula et al. / Talanta 52 (2000) 91–99 against potassium biftalate, C6H4(COOH)COOK. All NaOH solutions contained 3×10 − 5 M BTB. The HCl solutions in the range 0.1– 3 M were prepared by appropriate dilutions of 5.02 M stock solution, standardised against the 4.89 M NaOH solution. H2SO4 solutions were prepared by appropriate dilutions of concentrated acid. Acetic acid (HAc) solutions were prepared by appropriate dilutions of 5.0 M stock solution. 2.3. Apparatus A commercial SI flow system SIAmate™ analyser from Arctic Instruments Oy Ab (Turku, Finland) was used in the experiments [3]. The software package of the instrument (AnalySIA) was used in all experiments for device control, data acquisition and for integration of the peak areas. The instrument has a built-in zoomable LED-specII LED based photometer with a 0.76 mm light path. The photometer was trimmed against the darkest indicator solution before the experiments. The absorbance was monitored with a 635 nm LED. The blue colours of BTB and a-naphtholbenzein have their absorption maxima at 620 and at 650 nm, respectively. Absorbances of the used reagent solutions at these wavelengths were measured with a conventional spectrophotometer equipped with a 10 mm cuvette (Diode Array Spectrophotometer 8452A, Hewlett Packard). Absorbance values of 0.83 and 0.61 were 93 measured with the used reagent solutions of BTB and a-naphtholbenzein, respectively. Absorbances measured with the LED-specII during the experiments were lower than these values due to the shorter path length and dilution in the manifold. A schematic diagram of the sequential injection flow system is shown in Fig. 1. All tubing in the system was 0.76 mm i.d. polytetrafluoroethylene (PTFE). The length of the holding (HC) and reaction (RC) coils were 200 and 100 cm, respectively. 2.4. Procedure For the acid titration without sample dilution one measurement cycle consisted of the following steps: (1) washing of the sample line with a new sample solution; (2) aspiration of the desired sample volume; (3) aspiration of the desired reagent volume; (4) flow reversion and the detection of the blue colour, when the aspirated zones were pumped through the reaction coil to the detector and then to the waste. When the sample dilution was used, the measurement cycle was otherwise the same as described above, but after the step (2), part of the sample zone was dispensed to the auxiliary waste. In all experiments the measurement cycle was repeated three times for each concentration. The collected data was median filtered, which minimised the disturbances in the final signal caused by possible air bubbles. 3. Results and discussion 3.1. Zone dispersion Fig. 1. A schematic diagram of the SI system used. HC is the holding coil, RC is the reaction coil. Auxiliary waste was used as the dilution conduit. Zone dispersion, i.e. how well the injected zones are mixed with each other, is regarded as one of the most important parameters in SI analysis [4,5]. The flow profiles of the reagent and sample zones as they reached the LED-spectrometer were studied by injecting BTB dye solution into a colourless carrier stream. As the measurement protocol was the same as during the titration experiments, and because no chemical reactions took place, the obtained tracer curves show only 94 S. Maskula et al. / Talanta 52 (2000) 91–99 Fig. 2. (a) Detector response for increasing injection volumes of BTB; (b) peak height (A max) plotted as a function of the injected sample volume (), and the term − log (1− A max/A 0) also as a function of the injected sample volume ( ). A 0 is the maximum absorbance of the dye. The absorbance values are given in arbitrary units. the physical dispersion that the zones undergo on their way to the detector. A series of growing sample volumes of BTB was aspirated in the SI system in order to determine the sample volume, which has to be used to obtain half of the maximum absorbance, S1/2 (Fig. 2a). The term − S. Maskula et al. / Talanta 52 (2000) 91–99 log (1−A max/A 0) and the absorbance values are shown in Fig. 2b as the function of the sample volume. The term − log (1−A max/A 0) shows a linear relationship with the sample volume, which is accordance with the theory of controlled dispersion in FIA [1]. The S1/2 value for the used manifold was found to be 160 ml. The reagent and sample volumes used in this work giving the optimal titration results are shown in Table 1. Tracer curves for these injection volumes are shown in Fig. 3. It can been seen in Fig. 3 that the reagent and sample zones overlap each other almost completely for every volume combination thus guaranteeing the best dispersion conditions. 95 This is in good agreement with results obtained by Růžička et al. [4]. They found that the optimum conditions for SI single reagent based chemistry are obtained when the aspirated sample zone volume is 50.5S1/2 and at the same time the reagent zone volume is at least twice as large as the sample zone volume. 3.2. Titration of strong acid Hydrochloric acid in two different acid concentration ranges, 10 − 4 –10 − 2 and 0.1– 3 M, was titrated with sodium hydroxide. In order to optimise the linearity of the calibration curve different Table 1 Optimum volumes of reagents and samples in the titrations of strong and weak acids studied in this work Acid Concentration range (M) Reagent volume (ml) Sample volume (ml) HCl HCl H2SO4 HAc 10−4–10−2 0.1–3 6–13 2.6×10−4–8×10−2 175 175 175 175 65 50 50 80 (= 1.06S1/2) ( = 1.06S1/2) (= 1.06S1/2) (= 1.06S1/2) (= 0.41S1/2) (= 0.31S1/2) (= 0.31S1/2) (= 0.5S1/2) Dilution volume (ml) – – 175 – Fig. 3. The obtained tracer curves for the reagent, 175 ml, and different aspiration volumes of the sample, 80, 65 and 50 ml. The absorbance values are given in arbitrary units. 96 S. Maskula et al. / Talanta 52 (2000) 91–99 Fig. 4. Calibration curves for titration of 0.1 – 3 M HCl (line A), and for titration of 6 – 13 M H2SO4 with dilution (line B). concentrations of NaOH and different combinations of the reagent and sample volumes were studied. The best calibration curve with R 2 = 0.992 (not shown) for the titration of 10 − 4 –10 − 2 M HCl was obtained by using 0.5 ×10 − 3 M NaOH as reagent, 175 ml of reagent and 65 ml of sample. The best calibration curve for titration of HCl in the concentration range 0.1– 3 M was obtained by using 0.1 M NaOH together with reagent and sample volumes of 175 and 50 ml, respectively. The linearity of the obtained calibration curve was R 2 =0.995. The calibration curve for the range 0.1– 3 M is shown as line A in Fig. 4. The error bars (n =3) for the individual points are also included in the figure. Typical detector responses for the titration of 0.1– 3 M HCl are shown in Fig. 5. All the response curves show a presence of a ‘shoulder’ point, marked with ‘’ in Fig. 5. This point moves to the left when the acid concentration in the sample is increased. At the same time the ‘shoulder’ becomes less and less pronounced. According to our experience occurrence of the ‘shoulder’ is strongly dependent on the volume ratio between sample and reagent and on their concentrations as well. In our method, however, the area under the peak is measured. Therefore the ‘shoulder’ has not the same effect in the evaluation of the results as it would have in the traditional FIA titration where the evaluation is based on the width of the peak. 3.3. Titration of strong acid with NaOH with automatic dilution It was observed that when acid concentrations higher than 3 M was titrated, the change in the refractive index of the sample resulted in changes in the baseline, making the calculation of the peak areas difficult. Influence of the change in the refractive index can already be seen in the slightly falling baseline in Fig. 5, when acid concentrations higher than 0.5 M were titrated. For concentrations 5 3 M acid these changes were though so small that the effect on the determination of the peak areas was negligible. However, for titration of higher concentrations, an automatic sample and standard dilution procedure was used [6]. One S. Maskula et al. / Talanta 52 (2000) 91–99 of the ports of the selection valve was connected to auxiliary waste and the holding coil was used as the dilution conduit. The sample was first aspirated into the holding coil, where a concentration gradient of the sample was formed. Part of the sample gradient zone was then injected into the line of auxiliary waste. The remaining part of the sample gradient zone that was left in the holding coil, i.e. the tail of the sample peak, was then titrated with the base. Tracer curves for the titration experiments with dilution are shown in Fig. 6. The tracer curves are shown for the optimum experimental conditions, where 50 ml of sample was first aspirated into the holding coil. Then 175 ml was injected to the auxiliary waste and the sample tail remaining in the holding coil was then used as the sample. This highly diluted tail of the original sample can be seen in Fig. 6 (observe the different scales on the different axes in Fig. 6). When high concentrations of strong acids are used as the sample this tail has a concentration which is high enough to give repeatable titration results. It can be concluded from the curves in Fig. 6 that according to the proposed 97 procedure an approximate dilution factor of 100 was obtained. The dilution method was tested by titration of sulphuric acid in the range of 6 – 13 M. Linearity of the obtained calibration curve was good, R 2 = 0.993. The calibration curve for the range 6 – 13 M is shown as line B in Fig. 4. The error bars (n =3) for the individual points are also included in the figure. The SI peaks of the titration were similar to the peaks shown in Fig. 5. 3.4. Titration of weak acid with NaOH In the titration of acetic acid, HAc, a-naphtholbenzein was used as the indicator, because its colour transition range (pH 9.0– 11.0) better coincides with the pH at the equivalence point of the HAc titration. The obtained detector response was similar to the response in the strong acid titration. In these experiments, the acid concentration was varied from 3 ×10 − 4 –8×10 − 2 M. Different NaOH concentrations and different acid/base aspiration volumes were studied to optimise the linearity of the calibration curve. The best calibration curve was obtained by using Fig. 5. Detector response for titration of 0.1 – 3 M HCl with 0.1 M NaOH. The absorbance values are given in arbitrary units. 98 S. Maskula et al. / Talanta 52 (2000) 91–99 Fig. 6. Tracer curves for the reagent, 175 ml, and the diluted sample (observe different scales on different axes). The absorbance values are given in arbitrary units. 0.5×10 − 3 M NaOH and injection volumes of 175 ml of the reagent and 80 ml of the sample. The linearity of the calibration curve (not shown), was satisfactory, R 2 =0.989. 4. Conclusions The results obtained in this study indicate that the proposed method can be used in automatic SIA titration of different concentrations of strong and weak acids. By changing reagent concentration and injection volumes, acids in different concentration ranges can be titrated without any other adjustments of the SI manifold or the titration protocol. Even high concentrations of acid can be titrated by using automatic dilution procedure. Compared with the earlier SIA titration methods the reagent consumption can therefore be further reduced. One possible field where the proposed method can be used is the monitoring of acid concentra- tions in different industrial processes. In industrial processes analyte concentrations are often high, but their approximate concentrations are well known beforehand and their concentrations normally vary over a rather narrow range. The automatic dilution minimises even the disturbances caused by the changes in refractive index, which usually occur when spectroscopic methods are used to detect high concentrations. The method offers also a certain flexibility. By changing one or several of the following variables — the injected sample volume, injected reagent volume or reagent concentration — different concentration ranges of acid can easily be titrated. As the SI system is totally computercontrolled, change of these variables does not cause any physical reconfiguration of the flow manifold. The SI system can also be part of the control system in an industrial process, where the signal, i.e. the acid concentration, is used as one of the control parameters. S. Maskula et al. / Talanta 52 (2000) 91–99 99 Acknowledgements References This work has been supported by the Academy of Finland as a part of the A, bo Akademi Univer-sity Process Chemistry Group, a National Centre of Excellency. S.M. and J.N. thank for the support from the National Graduate School of Chemical Engineering. The authors thank Markus Kass for performing some of the experiments. [1] J. Růžička, E.H. Hansen, Flow Injection Analysis, second ed, Wiley, New York, 1988. [2] J.F. van Staden, H. du Plessis, Anal. Comm. 34 (1997) 147. [3] N. Kullberg, M. Vilén, P. Sund, et al., Talanta 49 (1999) 961. [4] T. Gübeli, G.D. Christian, J. Růžička, Anal. Chem. 63 (1991) 2407. [5] J.F. van Staden, H. du Plessis, S.M. Linsky, et al., Anal. Chim. Acta 345 (1997) 59. [6] A. Baron, M. Guzman, J. Růžička, et al., Analyst 117 (1992) 1839. .