ANALYST, MAY 1992. VOL. 117 863 Monitoring and Assay of Water Treatment Additives by Alternating Current Tensammetry and Voltammetry: Scope and Limitations Pierre M. Bersier" Central Analytical laboratory, Ciba-Geigy Ltd., Basle, Switzerland William Neagle and David Clark Ciba- Geig y In d us tria I Ch em ica Is, Tra ffo rd Park, Man ch ester, U K The alternating current (a.c.) tensammetric behaviour of different commercially available water treatment additives is described. Possibilities and limitations of their routine determination by a.c. tensammetry at low levels (0.510 ppm) in different aqueous media are discussed. Indirect differential-pulse voltammetry via the 12-molybdophosphate derivative allows a classification between phosphino-containing and phosphorus-free water treatment compounds.Practical examples are given. Keywords: Water additives; routine determination; alternating current tensammetry; voltammetry The whole spectrum of industry, manufacturing and engineer- ing, textiles and chemicals, food and drinks, even leisure and service industries, depend on pure water. Pure water is, however, not widely or readily available. It often has to be chemically treated or obtained from sea-water in limited amounts. World-wide efforts to develop chemicals for water treat- ment are being made. These chemicals are essential for modern industry and desalination technology for controlling problems such as: (i) scaling: scaling is a build-up of solid material formed on the inner surface of boilers, for instance, when the concentration of the impurities in the water used exceeds their solubility limit and precipitation occurs;' (ii) microbiological fouling: fouling is the deposition of materials, normally in suspension, onto heat-transfer or other surfaces such as boilers; and (iii) corrosion: corrosion is the destruction of a metal by electrochemical reaction with its environment.' These problems must be dealt with in a safe and environmen- tally sound manner.* The need to control the concentration of water treatment chemicals and to control operating costs make it desirable to use cost-effective water treatment products that can be applied with minimum operator involvement.Chemicals used in non-precipitation programmes are either: (i) chelants, forming complexes with calcium and magnesium;' or (ii) sequestrants (solubilizing agents), which, in the same way as chelants, keep calcium and iron in solution, but are less corrosive.' These formulations typically contain phosphonates, poly- (acrylates), poly(methacrylates), poly(ma1eates) and poly- meric dispersants.Operating at threshold levels as opposed to the stoichiometric reaction of chelant reaction programmes, the polymers and phosphonates function primarily by altering o r distorting the crystalline structure of hardness precipitates. The new technology eliminates corrosion problems associated with chelants and the excessive precipitation common with phosphate treatments. Typical sequestering compounds include aminotri(methy1- enephosphonic acid) or NTP (Wayplex NTP) and hydroxy- ethylenediphosphonic acid (HEDPA) (see ref.3). Poly- (acrylates) and poly(methacrylates) play leading roles in todays most advanced treatment programmes.' Belasol, for instance, was developed to meet the needs of the oil industry. It ensures that sea-water pumped through permeable rock to push the oil up will not leave impurities behind which would clog the rock and block oil recovery.* Belgarde EV, on the * Present address: Gstaltenrainweg 61, 4125-Riehen. Switzerland. other hand, is a liquid polymer scale control additive based on an entirely new branch of poly(ma1eic acid) chemistry.* A factor that is common to all the compounds used is that their monitoring and the control of their concentration is difficult owing to the low levels encountered. In some instances this has led to their being blended with low levels of heavy metal ions as markers.' Analytical Methods Commonly used methods for determining phosphonate-con- taining additives are reported to be difficult, time consuming and plagued by interferences.3 A rapid method based on the ultraviolet (UV)-catalysed oxidation of the phosphonate moiety to orthophosphate has been reported.The phosphate compound is determined as (3-12-molybdophosphate by UV absorption at 700 nm.3-4 This method is, however, not specific for the intact phosphonate additive and additives as such. For the assay of some additives without a phosphonate moiety, fluorescence has been pro- posed.4 This method, although very sensitive, is not specific for the intact additive. Poly(acry1ates) and poly(methacry1ates) are highly surface- active and hence have a strong influence on the differential capacity of the electrical double layer at the mercury-water interface.5--8 They exhibit strong alternating current (a.c.) tensammetric signals.The a.c. tensammogram of a poly- (acrylic acid) such as PAA-1 is shown in Fig. 1 (cf. Table 1). Pospisil and Kuta studied the behaviour of maleic acid at a mercury electrode by a.c. polarographyg and its influence on the electrocapillary curve.") The present study shows that the direct a.c. tensammetric assay of poly( maleic compounds) is feasible. This paper discusses the possibilities and limitations of a.c. tensammetry for the direct assay and monitoring of selected modified phosphinocarboxylic acids (PCA- 1 , PCA-2, PCA-3 and HPA, cf.Table 1). A.c. tensammetry is also applicable to poly(acry1ic acids) (PAA-1-PAA-5, cf. Table 1) and poly- (maleic acids) (PMA-1, PMA-2, cf. Table 1), additives which contain no phosphinate groups, cf. Table 1. Experimental Apparatus A.c. tensammetric measurements were carried out with a Metrohm Polarecord 506 in conjunction with a Metrohm VA 633 multielectrode stand, using a hanging mercury drop as the working electrode.864 ANALYST, MAY 1992, VOL. 117 For the indirect differential-pulse voltammetric measure- ments of the p-12-molybdophosphate, a Polarecord 506 or 626 or an Amel Model 471 Multipolarograph and a dropping mercury electrode as the working electrode were used. A platinum wire was used as the counter electrode and a saturated calomel electrode as the reference electrode.The latter was connected to the cell by means of a double salt-agar bridge. All tensammetric and voltammetric measurements were made at room temperature (23 k OS'C), in de-aerated solutions. Reagents and Equipment The additives studied, which are summarized in Table 1, were all of the purest grade available. All other chemicals were of analytical-reagent grade and were used as received. Generic structures of the three groups of water treatment additive (I, I1 and 111) examined in this work are given in Table 2. The chemical composition of the model waters, viz., Ca-50, Ca-300, an artificial sea-water formulated according to DIN I I I I 0 - 500 - 1000 - 1500 NmV versus SCE Fig. 1 A.c. tensammograms of poly(acry1ic acid), PAA-1, recorded in this laboratory.Supporting electrolyte: model water Ca-300- 0.12 mol dm-3 sodium perchlorate, pH 4.4. Working electrode, hanging mercury drop. Applied alternating voltage, 15 mV (root mean square) at 75 Hz. Curve 1, supporting electrolyte; 2 , l ; 3,2; 4,4; 5.10; 6,20; 7,50; 8,100; 9,200; and 10,500 mg I-' of PAA-1. A, Rest current depression; and B, desorption peaks 5090011 and an artificial oil-loaded formation water, is given in Tables 3-5. Sep-Pak CI8 cartridges (Waters), when used for the assay of water treatment additives in polluted sea-water, were acti- vated with 5 ml of methanol and washed with 10 ml of doubly distilled water. Direct a.c. tensammetric assays were performed by adding the stock solution or sample solution to the appropriate supporting electrolyte.Procedure For the a.c. tensammetric assay, the following experimental procedure was applied. Stock solutions of each additive were prepared by dissolving approximately 150 mg of the additive of known concentration in 10 ml of doubly distilled water. For the indirect voltammetric determination of phosphonate-con- taining additives via fi-12-molybdophosphate, the UV-cata- lysed oxidation of the phosphino group to orthophosphate was performed in a glass cell with a thermostatically controlled heating mantle connected to a Lauda water-bath, using a high-pressure 125 W Hg tube. To 2 ml aliquots of the sample were added 1 ml of 1 mol dm-3 NaOH, 1 ml of concentrated H2S04 and 1 g of ammonium peroxydisulfate, and the volume was made up to 20 ml with doubly distilled water.The pH of this solution is about 7. The solution was irradiated for 20 min at 70 "C. The irradiated sample was then transferred quanti- Table 2 Generic structures of the three groups of water treatment additive (I, I1 and 111) examined (cf. Table 1) Group I : Phosphinocarboxylic acids- Group I1 : Poly(acry1ic acids)- Group 111: Poly(ma1eic acids)- +:&-::& Table 1 General and a.c. tensammetric data for the water treatment additives examined Compound PCA-1 (1) PCA-2 PCA-3 PAA-1 (11) PAA-2 PAA-3 PAA-4 PAA-5 PMA-1 (111) PMA-2 PMA-2 PMA-2 TCA ( W HPA P(MA/ME) P( MNSSA) Group Phosphinocarboxylic acid Phosphinocarboxylic acid Phosphinocarboxylic acid Poly(acry1ic acid) Poly(acry1ic acid) Poly(acry1ic acid) Poly(acry1ic acid) Poly(acry1ic acid) Poly(ma1eic acid) Poly(maleic acid) Poly(ma1eic acid) Poly(maleic acid) Triazinecarboxylic acid Hydroxyphosphonocarboxylic acid Maleic acid-ethylacrylate copolymer Maleic acid-st yrene-sulfonic acid copolymer Relative molecular mass -3000 -2700 -700 -2000 -5000 -4500 -2000 -3000 =700 500-550 -500 - 1250 -470 156 800-850 -1500 Lower detection limit (ppm) 1 1 1 1 1 1 1 1 1 1 1 1 0.1 10 1 1 Linear range 1-10 1-10 1-600 ( P P 4 - - - - - 1-(50) 1-600 1-60 0.1-20 10-50 1-10 - -ANALYST, MAY 1992, VOL.117 865 Table 3 Composition of Ca-50 and Ca-300 model waters Concentratiodmmol dm-3 Water S042- HC03- Ca*+ Mg2+ c1- Ca-50 0.4 0.3 0.5 0.5 1.1 Ca-300 0.4 6 3.0 3.0 6.1 Table 4 Composition of artificial sea-water11-12 Constituent Amount present*/g NaCl 28 MgSO4.7H20 7 MgC12.6H20 5 NaHC03 0.2 CaCI2.6H20 2.4 * In 985 ml of distilled water.Table 5 Composition of artificial oil-loaded formation water (density, 1.029 g cm-3; pH, 5.5; and ionic strength, 0.77 mol dm-3). The water is prepared by shaking artificial Gullfaks formation water with Gullfaks crude oil (SO + 50, v/v) for 24 h at ambient temperature Constituent Concentratiodmg I-L Na+ K+ Ca’+ Mg*+ Sr*+ Ba2+ c1- HC03- co32-- sop 14 570 330 1 040 305 260 50 25 600 400 0 0 Total dissolved solids: 42 555 tatively into a 50 ml calibrated flask and made up to the mark with doubly distilled water. To a 4 ml aliquot of this solution were added 4 ml of acetone and 2 ml of a solution containing 30 g of Na2MoO4-2H20, 24 g of tartaric acid and 90 ml of HCI (32%) per litre, and the solution was transferred into the polarographic cell.After de-aeration with pure nitrogen (99.998%) for 10 min, the voltammograms were recorded in the differential-pulse mode. The exact experimental conditions are given in the figure legends. Results and Discussion Quantitative Determinations Typical a.c. tensammograms of the different classes of additive (cf. Table 1) are illustrated in Fig. 2. The four additives, as shown in Fig. 2, were measured under identical experimental conditions. For a practical assay, however, the optimum conditions for each class and com- pound must be established. Variation in the concentration of the surface-active com- pound affects the depth of the depression of the current of the pure supporting electrolyte (curve 1 in Fig. 1) in addition to the change in the height and the position of the desorption peaks (cf.Fig. 1). Both effects can be exploited for the quantification of water treatment additives. Tensammetric waves frequently behave differently to fara- daic processes (cf. refs. 5-8, and references cited therein). A characteristic of polarographic and voltammetric techniques is the broad linear dependence over six or more decades. It is therefore much broader than that of most other instrumental methods. In tensammetry the dependence of the value of the measured capacity current on the concentration of the surfactant generally has a non-linear character. Hence a linear t c ‘ I I I I I I I I I I I 0 -500 -1000 -1500 -2000 0 -500 -1000 -1500 -2000 NmV versus SCE Fig. 2 I , PCA-I; (b) Group 11, PAA-1; (c) Group 111, PMA-1; and ( d ! Group IV, TCA.Curve 1: supporting electrolyte, 0.9 N lithium sulfate, pH 4; curve 2: additive concentration, 1.5 x mol dm-3 (cf. Table 4) A.c. tensammograms of water treatment additives: (a) Grou E I5O E a r 0, a z .- Y 100 a 2 a c 3 50 c .- 0 cz 9 t 0 1 2 3 4 5 [PCA-I]/pg ml-’ Fig. 3 Calibration graph for the determination of PCA-1 (‘as is’), measured in 0.1 mol dm-3 sodium fluoride, pH 11. Different symbols indicate replicate measurements dependence between the measured signal height (depression of the current of the supporting electrolyte) or the height of the desorption peaks, respectively, is observed only at comparably low concentrations and over a narrow concentra- tion range (Fig. 3). The calibration graphs obtained under the experimental conditions exhibit linear ranges that depend on the compound under study and are therefore a characteristic of each additive.The linear ranges observed are summarized in Table 1. At low concentrations the standard additions method can be applied for the quantitative determination of the water treatment additives, provided that the sum of the analyte present after addition of the reference substance still falls within the linear part of the graph. For quantitative determination in low concentration ranges, the depression of the rest current (cf. Fig. 1) was exploited in most instances. The lowest detection limits are in the range 0.1-1 ppm (cf. Table 1). At higher additive concentrations, a dependence of the peak potential (Ep) on the logarithm of the additive concen- tration (log c) is observed in most instances (Fig.4).866 w 0 a 0 - s $ -20 u* -60 -40 . ANALYST, MAY 1992, VOL. 117 - - - +60 I 1 + +40 20 I -80 1 ,/ , , I -100 10 100 1000 10000 [PCA-Il/pg rnl-1 Fig. 4 Peak potential ( E J versus log of the PCA-1 concentration. The first desorption peak is exploited for quantitative assay. Different symbols indicate replicate measurements +215 rnV 1 NmV versus SCE --t Fig. 5 Differential-pulse voltammograms of 6-12-molybdophos- phate (procedure according to Fogg and co-workers.13.*4 Curve 1, sample after UV-catalysed oxidation; and curves 2 4 , sample after standard additions of phosphate (equivalent phosphate concentra- tions = 3.07, 6.08 and 9.03 pg ml-1 of polarographic solution) Classification of the Different Additives A serious shortcoming of tensammetry and of all electrochem- ical methods in general is their limited specificity and selectivity .The difficulties in determining surfactants in mixtures and interference problems are certainly among the drawbacks hindering the use of tensammetry as a viable analytical technique. The potential of the tensammetric peaks must differ by at least 0.2-0.3 V. The concentration of the most strongly adsorbed component must be such that the coverage of the electrode surface is less than 100%. Otherwise, only the peak of the most strongly adsorbed compound will be detected on the tensammograms.5-8 A direct differentiation between the four classes of water treatment additive (I-IV, cf. Fig. 2 and Table l), viz., (I): additives with phosphino groups; (11): without phosphino groups; (111): poly(maleic acids) (cf.Table 2); and (IV): miscellaneous, in the concentration range of interest [O. 1( 1)- 10 ppm] is not possible with the a.c. tensammetric procedure described here. The UV and voltammetric determination of the phosphino group, however, allows a preliminary classifica- tion between group I and groups I1 and 111. The voltammetric wave of (3-12-molybdophosphate ob- served at a glassy carbon electrode (cf. Fig. 5 ) has been used as Table 6 Comparison of differential-pulse voltammetric and spectro- photometric results for the determination of the phosphorus content of selected commercial water treatment additives after UV-catalysed oxidation of the phosphinate or phosphonate moiety to orthophos- phate Phosphorus (%) Polarographic Sample assay PCA- 1 0.92 0.97 0.98 PCA-2 0.98 0.76 PCA-3 2.54 2.67 HPA 9.4 9.3 9.1 Spectrophotometric assay 1.02 0.98 0.93 0.77 2.41 2.21 11.16 11.08 - - 80 I I E E .4 d J= P) r Y m a m 60 .- Y 40 2 3 >- m a "u 20 0" 0 2 4 6 8 10 [AdditiveVpg rnl-' Fig. 6 Influence of the relative molecular mass of phosphinocar- boxylic acid additives on the slope of the peak height. A, HPA; B, PAC-3; C, PCA-2; and D, PCA-1 the basis of a method for determining orthophosphate.13.14 Hence, voltammetry can serve as an alternative method to spectrophotometry for the determination of phosphate. Good agreement was found between phosphate concentrations determined by voltammetry and data obtained by a spectro- photometric method,*4 as shown in Table 6.The advantage of the electroanalytical methods used here is that both the voltammetric and the a.c. tensammetric assays can be carried out with the same instrument (a polarograph). Qualitative Determination Based on Tensammetric Measurements The shape and peak potentials of the tensammetric waves depend on the nature of the compound studied and, therefore, can, in some instances, give qualitative information. Jehrings showed that with increasing relative molecular mass (M,) (1000,3000, 5000,20 000) the peak potential of poly(ethy1ene glycols) moves progressively towards more negative values. The shift is a linear function of the reciprocal average M , . Resolved peaks are obtained. The peak width decreases with increasing M,.Bagdasarov et ~ 1 . 1 5 observed that the slope of the plot of capacity current versus concentration increased with increas- ing length of the hydrocarbon chain, for instance from C8 to CI2 (cf. Fig. 2 in ref. 15). Distinct differences between the slope of the measured capacity current (ic) (depression of theANALYST, MAY 1992, VOL. 117 867 1000 1 1 v v C 10 20 Additive/mg dm-3 Fig. 7 Adsorption of a water treatment additive onto the surface of iron powder as a function of the amount of additive. Medium: model water Ca-SO containin 4 g l - l of Fe (20h at 40°C). 0, pH 5.5 de-aerated solution; 8, pH 7.5 de-aerated solution; 0, pH 7.5 aerated solution; V, pH 9.5 de-aerated solution; and A, pure water, pH 7.5. A, 0.79; B, 1.2U1.26; and C, 1.53 mg g-l of Fe rest current or the peak height, respectively) and the M , of different phosphonate-containing additives of group I are observed (Fig.6). Hence, for the phosphinocarboxylic acids HPA (average M , =: 160), PAC-3 (average M , = 700) PCA-2 (average M , = 2700) and PAC-1 (average M , = 3000), provided that the amount of phosphonate present is known (the determination is carried out with the indirect procedure via (3-12-molybdo- phosphate described above), the slope of a graph of i, versus c furnishes qualitative information on the additive present, as is revealed by an inspection of Fig. 6. A distinction between the phosphinocarboxylic acids containing PCA-1 and PCA-2 with an average M , of 3000 and 2700, respectively, is, however, not possible, as shown in Fig.6. In practical applications, such as monitoring, not all of these compounds are present in a mixture. Hence, tensammetric measurements should permit both their determination down to fairly low levels and, for phosphinocarboxylic acids, with greatly different M , values, their qualitative determination/ identification. Applications In addition to the examples mentioned above, the following examples serve to illustrate the possible applications of the a.c. tensammetric techniques to the routine practical analysis of water treatment additives . Tensammetric determination of additives in the presence of iron (powder) performed on model waters Experiments run in the presence of iron powder (4 g per litre of model water Ca-50, for instance) showed no direct influence of iron or iron ions on the tensammetric assay of the additive examined.Changes in the concentration of the additive at the ppm level, owing to adsorption onto the surface of the iron powder, could be followed in de-aerated and aerated media. Hence the determination of the amount of an additive adsorbed on a surface and the thickness of the adsorbed layer can be monitored as a function of the given experimental conditions, such as temperature, composition, pH of the bath and time. The adsorption of an additive onto the surface of iron powder as a function of the amount of the additive is shown in Fig. 7. Determination of scale inhibitors in oil-loaded artificial sea- water and oil-loaded artificial formation water Concentrations of 0.5-400 ppm (range tested) of the scale inhibitor PCA-1 could be determined in oil-loaded artificial I I 1 1 10 100 1000 PCA-l/pg ml-1 Fig.8 Recovery of PCA-1 added to artificial sea-water (0, 0) and artificial formation water (A). Measured by the proposed a.c. tensammetric procedure sea-water and formation water (for details of the composition of these waters see Tables 4 and 5 ) in the presence of a de-emulsifier (100 and 200 ppm) by a.c. tensammetry. The assay was performed in 0.1 mol dm-3 NaF, pH 11, supporting electrolyte, after separation and accumulation on a Sep-Pak CI8 cartridge with or without prior extraction of the oil with CCI4, using the procedure described by Gruenfeld.16 A hanging mercury drop electrode was used as the working electrode. In pure 0.1 mol dm-3 NaF, pH 11, a linear relationship between the height of the desorption peak and the additivehhibitor concentration was found in the range 0.5-5 ppm.In the range 100-9000 ppm, a linear (EP versus log c ) dependence was observed. Recoveries of PCA-1 added to oil-loaded artificial forma- tion water and artificial sea-water are illustrated in Fig. 8. Conclusions A.c. tensammetry, combining high sensitivity with good precision, appears to constitute a very convenient electro- chemical procedure for the assay of water treatment additives, which represent a particularly important group of compounds in modern technology. It is therefore a valuable alternative to the methods commonly used. However, the practical analytical chemist, who is interested in the potential applications of tensammetric techniques, should be familiar with the variability of these techniques towards medium effects.The type and concentration of electrolyte may influence the wave shape and position far more than faradaic processes (cf. refs. 5-8). As stressed by Bond,” extreme care and commonsense must, therefore, be employed in using tensammetric techniques in routine analysis. The skillful technical assistance of H. G. Wenzel and J.-P. Worch is gratefully acknowledged. References Sendelbach, M. G., Chem. Eng., 1988, August, 127. Sykes, S . , Ciba-Geigy J.. 1988, 3, 10. Hach’s Water Analysis Handbook. Hach, Loveland, CO, Phosphonates. Range: 0-20 mg 1- 1 . Persulfate/UV Oxidation Method for Water, pp. 2-234-2-236. Clark, D., unpublished work. Bersier, P. M., and Bersier, J . , Analyst, 1988, 113, 3. Kalvoda, R., Pure Appl. Chem., 1987. 59. 715. Kalvoda, R.. and Parsons, R., Electrochem. Res. Dev., (Proc. UNESCO Forum. 1984, Publ. 1985; Chem. Abstr.. 105, 690 15n).868 ANALYST, MAY 1992, VOL. 117 8 Jehring, H . , Elektrosorptionsanalyse mit der Wechselstrompol- arographie, Akademie Verlag, Berlin, 1974. 9 Pospisil, L., and Kuta, J., Collect. Czech. Chem. Commun., 1968,33, 3040. 10 Pospisil, L., and Kuta, J., Collect. Czech. Chem. Commun., 1969,s. 3047. 11 Deutsche Industrie Norm (DIN) 50900, November, 1960. 12 Rompp Chemie Lexikon, eds. Falbe, J., and Regitz, M., Georg Thieme, Stuttgart, 1991, vol. M-PK, p. 2669. 13 Fogg, A. G., and Bsebsu, N. K., Analysr, 1981, 106, 369. 14 Fogg, A. G., Bsebsu, N. K., and Birch, B. J., Talanta, 1981,28, 473, and references cited therein. 15 16 17 Bagdasarov, K. N., Lokshina, G. A., Sadimenko, L. P., and Sokolov, V. P., Zh. Anal. Khim., 1986, 41, 171. Gruenfeld, M., Environ. Sci. Technol.. 1973, 7. 636. Bond, A. M., Modern Polarographic Methods in Analytical Chemistry, Marcel Dekker, New York, 1980. Paper 1 I01 516H Received April 2, 1991 Accepted December 13, 1991