1072 Analyst September 1983 Vol. 108 pp. 1072-1081 Studies of Calcium Ion-selective Electrodes in the Presence of Anionic Surfactants Anthony J. Frend Gwilym J. Moody and J. D. R. Thomas Afifilied Chemistry De#artment UWIST Cardifl CFl 3XA and Brian J. Birch Unilever Research Laboratory Bebington Wirral Merseyside L62 4XN Eighteen membrane systems based on a calcium bis{di[4-( 1,1,3,3-tetra-methylbutyl)phenyl]phosphate} ion sensor have been compared in potentio-metric studies with a membrane based on the sensor with a dioctyl phenyl-phosphonate solvent mediator with respect to interferences of calcium ion-selective electrodes by anionic surfactants especially by sodium dodecyl-sulphate (SDS) and sodium tetradecylbenzenesulphonates (ABS) . Electrodes made from poly(viny1 chloride) matrix membranes of the sensor with trioctyl phosphate solvent mediator are far superior to the other membrane systems in resisting interference by anionic surfactants and are shown to yield calcium ion levels in the presence of a wash liquor to match the expected values.Some improvement over dioctyl phenylphosphonate is also offered by decan-1-01 dodecan-1-01 and tetradecan-1-01 but the use of such solvent mediators impairs calcium ion selectivity and poly (vinyl chloride) plasticising qualities. The use of alternative polymer matrices based on poly(viny1idene chloride) and VAGH copolymer (hydrolysed vinyl chloride - vinyl acetate) offer no advantages. A nineteenth membrane obtained from a commercial supplier exhibited interference by SDS. An interesting effect is the increase (rather than the normal decrease) in e.m.f.observed for those electrodes from membranes in which the amount of free active sensor was reduced to low levels. By using optimum levels of sensor it was possible to fabricate electrodes exhibiting a zero e.m.f. change when changing the background 1 0 - 2 ~ calcium chloride to one that is also 1 0 - 3 ~ in SDS. Such membranes are based on decyl phosphate grafted to VAGH copolymer. X-ray fluorescence and chromatographic studies on membranes of the sensor with dioctyl phenylphosphonate in poly(viny1 chloride) show SDS to be a significant agent in leaching membrane components especially of dioctyl phenylphosphonate. Such observations are indicative that the interference of calcium ion-selective electrodes by anionic surfactants may be the result of the different solubilities of calcium - surfactant complexes in the solvent mediator of the membrane.Keywords Calcium ion-selective electrodes ; anionic surfactant interference In the absence of calcium ions the Orion 92-20 liquid membrane calcium ion-selective electrode (ISE) responds to anionic surfactants over limited concentration ranges with an anionic slope.1$2 Llenado observed that when this electrode was immersed in a solution containing calcium ions and a small aliquot of Sulframine (a linear-chain alkylbenzene sulphonate mixture with Cll 42.5%; C,, 34.6%; C,, 13.8%; and <C9 and >CIS 9.1% in the alkyl chain) was added there was a rapid negative shift in e.m.f. This same e.m.f. shift has also been observed for poly(viny1 chloride) matrix membrane calcium ISES.~ This work is concerned with further studies of this effect and of possible means of reducing the interference.Experimental Electrodes Poly(viny1 chloride) and other polymer matrix membrane electrodes with inner solution FREND MOODY THOMAS AND BIRCH 1073 of 10-1 M calcium chloride were assembled by the general procedure described previou~ly4~~ and the membrane compositions and electrode characteristics are summarised in Table I. Membrane No. I I1 I11 IV 17 VI VI I VIII IX X XI XI1 XI11 XIV xv XVI XVII$$ XVIII$$ XIX TABLE I COMPOSITION OF MASTER MEMBRANES AND ELECTRODE CHARACTERISTICS Composition* Slope at Limit of 25 "C/mV detectiont/ decade-1 M x 0.36 g DOPP$ + 0.17 g PVCY .. . . 0.36 g decan-1-01 + 0.17 g PVC . . . . 0.36 g dodecan-1-01 + 0.17 g PVC. . . . 0.36 g tetradecan-1-01 + 0.17 g PVC . . . . 0.36 g hexadecan-1-01 + 0.17 g PVC . . 0.36 g trioctyl phosphate + 0.17 g PVC . 0.36 g tripentyl phosphate + 0.17 g PVC . . 0.36 g dioctylphosphite + 0.17 g PVC . . 0.18 g decan-1-01 + 0.18 g DOPP + 0.17 g PVC . . 0.18 g dodecan-1-01 + 0.18 g DOPP + 0.17 g PVC . . 0.18 g tetradecan-1-01 + 0.18 g DOPP + 0.17 g PVC . . 0.18 g hexadecan-1-01 + 0.18 g DOPP + 0.17 g PVC . . 0.36 g DOPP + 0.085 g poly(viny1idene chloride) + 0.085g PVC . . . . . . (vinylidene chloride) + 0.085 g PVC . . 0.085gPVC phosphoric acid (VAGHPl) + 0.17 g PVC . . 0.36 g DOPP + 0.17 g PVC .. Philips IS 561/SP Ca2+ membrane. . . . . . 0.36 g DOPP 4- 0.085 g VAGH + 0.085 g PVC . . . . 0.18g DOPP + 0.18g decan-1-01 + 0.085g poly-0.18g DOPP + 0.18g decan-1-01 + 0.85g VAGH + 0.36 g DOPP + 0.01 g VAGH grafted with monodecyl-30.9 27.5 26.5 28.1 29.7 30.1 31.0tt 33.4 22.0 31.8 28.8 27.3 35.6 32.1 30.2 26.2 30.4 (A) 27.8 -1.8 3.0 5.0 5.2 9.5 7.5 6.0t t 8.5 8.0 4.0 4.5 7.2 3.1 4.2 5.1 3.2 6.0 (A) 2.0 -kE:Ns: 2.411 0.25 0.61 0.53 0.55 0.043** 0.021tt $1 1.4 0.46 0.42 0.43 0.21 0.31 0.48 0.52 1.2 (A) 0.31 -* Including 0.04 g of calcium bis(di [4-( 1,1,3,3-tetramethylbutyl)phenyl]phosphate}. t Determined using calcium ion buffers based on STP (see Table 11).$ "a+] = 5 x 10-3M. DOPP = dioctyl phenylphosphonate. 7 PVC = poly(viny1 chloride). 1) Data from reference 6. ** "a+] = 0 . 0 5 ~ . t t Data from reference 7. $ $ too low for measurement. $5 Details of VAGHPl in reference 8. The 6 membranes (A-F) prepared for each of the compositions XVII and XVIII each contained 0 (A) 0.001 (B) 0.005 (C) 0.01 (D) 0.015 (E) and 0.02 (F) g of calcium bis {di [4-( 1,1,3,3-tetrarnethylbutyl)phenyl]phosphate} respectively. The master membranes were cast in the normal way4s5 except that those of the highest alcohols (membranes III-V) were redissolved in tetrahydrofuran and recast in order to obtain membranes of homogeneous composition. Poly(viny1idene chloride) or VAGH co-polymer alone that is without the additional support of poly(viny1 chloride) as in membranes XIII-XVI gave membranes of insufficient mechanical strength when used with normal proportions of dioctyl phenylphosphonate solvent mediator while smaller amounts of the phosphonate gave membranes of gel-like consistency Such matrix membranes of poly-(vinylidene chloride) or VAGH copolymer were therefore not evaluated in this study.Electrodes were conditioned overnight in 10-1 M calcium chloride solution and calibrated with serially diluted calcium chloride solutions in the 10-1-10-5 M range and at below M (0.1 M ionic strength) with sodium tripolyphosphate (STP) buffers (Table 11) set at pH 9 1074 FREND et al. STUDIES OF CALCIUM ION-SELECTIVE Analyst Vol. 108 TABLE I1 COMPOSITION OF STP-BASED (1.472 g 1-1) CALCIUM ION BUFFER SOLUTIONS OF 0.1 M IONIC STRENGTH The pH was adjusted to 9.0 with ammonia solution.[NaCl]/g 1-1 Total [Ca2+]/~ Free [Ca2+]/~ Ca2+ activity/M 3.57 5 x 10-4 1.02 x 10-6 4.08 x 10-7 3.62 1.0 x 10-3 2.41 x 9.64 x 3.68 1.5 x 10-3 4.37 x 10-6 1.70 x 3.74 2.0 x 10-3 7.31 x 2.92 x 3.85 3.0 x 10-3 2.09 x 8.36 x 3.80 2.5 x 10-3 1.21 x 10-5 4.84 x 10-6 Reagents and Materials All materials except the following were of the best analytical grade available. Calcium bis{ di [4- (1,1,3,3-tetramethylb~tyl)phenyl]phosphate) ,9 which has the formula Ca{ [CH3.C( CH,) ,.C,H,.O] 2.P( 0) 0) , and dioctyl phenylphosphonate,1° which has the formula C,H,.P(O) [O(CH,),CH,], were prepared as described previously. Dioctyl phenylphosphonate is also listed by Lancaster Synthesis and in the Alfred Bader Rare Chemicals List of the Aldrich Chemical Company.VAGH copolymer (a partially hydrolysed vinyl chloride - vinyl acetate copolymer of relative molecular mass ca. 23000) from Union Carbide (UK) Ltd. and poly(viny1idene chloride) homopolymer from ICI Ltd. were purified by first dissolving in tetrahydrof uran. The polymer was then precipitated by the dropwise addition of the solution into methanol. The separated precipitate was homogenised in a household homogeniser filled with de-ionised water. The powdered polymer was then filtered and dried at 60 “C under vacuum. Specially pure sodium dodecylsulphate (SDS) was obtained from BDH Chemicals Ltd., while sodium tetradecylbenzenesulphonate (ABS) was obtained from Unilever Research Laboratory.A “model soap powder’’ with the following component composition (yo mass in parentheses) was prepared (using water to bring the total of components to 100%) ABS (7.0) ; tallow soap (2.0) ; sodium silicate (10.0) ; sodium sulphate (10.0) ; STP (35.0) ; and sodium perborate (25.0). Procedures Efect of anionic swfactants on calcium ISEs The various electrodes (Table I) were tested for interference to SDS and ABS as appro-priate in conjunction with an Orion Model 90-02-00 double-junction reference electrode. The e.m.f. measurements were made with a Corning EEL Model 112 digital millivoltmeter -pH meter used in conjunction with a Servoscribe Model RE 4541 potentiometric chart recorder. The various test solutions were maintained at 25 & 0.1 “C. Normally the effect of added surfactant was examined in four background solutions, namely lo- and lo- M calcium chloride solutions 10-1 M sodium chloride solution and a calcium ion buffer solution formulated to give a calcium(I1) concentration of 1.70 x lo- M (pCa = 5.77) (Table 11).M) were added to the particular background solution (25 cm3) in which the calcium reference electrode pair had been equilibrated to a steady response. E.m.f. readings were noted for each aliquot added and further aliquots added until the background solution had reached a 10-3~ con-centration in surfactant. Each full run was normally performed six times by using a fresh membrane from the parent master membrane for each run. Except for electrodes from membranes 111 IV VI and XI calibrations for calcium ions following exposure to calcium ions were not possible; electrodes from membrane VI were the most robust and re-calibration with calcium gave responses near to those existing prior to contact with the surfactant.Thus aliquots (0.05 cm3) of surfactant solution (5 x Efect of anionic surfactants on re ference-type electrodes In checks on the effects of anionic surfactants on reference-type electrodes the Orion, Model 94-07 fluoride ISE was used as reference in conjunction with the electrode under study namely the EIL Model 1070030 pH electrode Orion Model 90-02-00 double September 1983 ELECTRODES IN THE PRESENCE OF ANIONIC SURFACTANTS 1075 junction reference electrode and a silver - silver chloride reference electrode from an EDT Research ISE body. An electrode made from membrane I (Table I) was used as the control.The e.m.f.s were monitored with an Orion Model 901 millivoltmeter fitted with two high impedance inputs in conjunction with a potentiometric chart recorder. The background solutions to which aliquots (0.05 cm3) of anionic surfactant solutions (5 x 1 0 - 2 ~ ) were added were pH 7 buffered (B) (solutions prepared in Radiometer Type S1326 pH 7.00 buffer solution) and unbuffered (A) potassium fluoride solutions M) that were also 10-5 M in potassium chloride. Analysis of calcium ions in detergent solutions The previously calibrated calcium ISE/Orion 90-02-00 double junction electrode pair were placed in a solution (25 cm3) of model soap powder of selected concentration. The solution was spiked with aliquots (0.025 cm3) of 1 M calcium chloride until the concentration of calcium chloride added was 5 x 1 0 - 3 ~ .Equilibrated e.m.f.s were noted for the various spikes. The calcium ion calibrations were compared with the calculated values by allowing for the effect of the model soap powder components. Mechanistic Experiments X-ray fluorescence studies Master membranes of membrane I composition (Table I) were cut into quadrants and each quadrant recast to full size. The resulting thin membranes were suspended for 48 h in appropriate stirred liquids (25 cm3) namely doubly de-ionised water and SDS solutions (5 x After rinsing thoroughly with de-ionised water and air-drying the surfaces of untreated and treated membranes were examined with a Model TN 2000 Tracer Europa energy-dispersive X-ray fluorescence spectrophotometer (Unilever Research).Additional master membranes were first suspended in 10-1 M SDS for 48 h and similarly examined. The membrane compositions were as follows (a) calcium bis(di[4-(1,1,3,3-tetramethylbutyl)phenyl]phosphate} (0.04 g) 2-nitrophenyl phenyl ether (Alfred Bader Library of Rare Chemicals of Aldrich Chemical Company) (0.36 g) and poly(viny1 chloride) (0.17 g); and (b) dioctyl phenylphosphonate (0.36 g) and poly(viny1 chloride) (0.17 g). Confirmatory thin-layer chromatography (TLC) experiments for leached components were carried out on the above suspension solutions following evaporation to dryness and dissolution of the residue in tetrahydrofuran (1 cm3). Chromatograms of the tetrahydrofuran extracts were run on silica gel with benzene - acetic acid (9 + 1) and detected with iodine vapour.and 10-1 M). Gas chromatogyaphy A conditioned electrode of membrane I (Table I) with a double-junction reference electrode were equilibrated to constant e.m.f. in 10 cm3 calcium chloride solution M). The solution was spiked with 0.02 cm3 SDS (5 x M) and when the e.m.f. had again stabilised (after ca. 5 min) the electrode pair was removed. The aqueous solution was extracted with dichloromethane (10 cm3) and the dichloromethane extract reduced in volume (to 2 cm2) by vacuum evaporation. This was chromatographed in a Perkin-Elmer Model F11 gas chromatograph fitted with a Carbowax 20M column using a flow-rate of 20 cm3 min-l at a temperature of 130 "C and the instrument injection temperature setting at No.8. The output was recorded on a Servoscribe Model 4541 potentiometric recorder with a sensitivity of 100 mV f.s.d. and a chart speed of 300 mm h-l. Standards of dioctyl phenylphosphonate and SDS were prepared in dichloromethane and similarly chromatographed. Results and Discussion The responses of calcium ISEs based on a calcium bis [di(4-octylphenyl)phosphate] sensor with a dioctyl phenylphosphonate mediator in poly(viny1 chloride) to added surfactants have already been d~cumented.~ The general effect of added anionic surfactant even at about 2 x M is to lower the e.m.f. response of the calcium ISEs when in contact with the various test solutions regardless of whether or not calcium is present in the cell test solution.s Similar behaviour has been observed in the present studies for calcium ISEs with a sensor o 1076 FREND et al.STUDIES OF CALCIUM ION-SELECTIVE Analyst Vol. 108 an isomer of calcium bis [di(4-octylphenyl)phosphate] namely calcium bis(di [4-( 1,1,3,3-tetramethylbutyl)phenyl]phosphate) (membrane I in Table I). Calcium ISEs based on either of these sensors behave ~imilarly,~Jl and these anionic surfactant interferences are undesirable. In these studies aimed at designing an electrode to diminish the undesirable effects, particular attention has been given to SDS but with some reference also to ABS. It is convenient to express the results of the investigation and their discussion under three main headings namely the effect of anionic surfactants on calcium ISEs with membrane compo-sitions as in Table I the effects of anionic surfactants on reference type electrodes and mechanistic studies based on X-ray fluorescence and chromatographic experiments.Effect of Anionic Surfactants on Calcium ISEs The membrane compositions summarised in Table I have been designed to demonstrate the extent of anionic surfactant interference on calcium ISEs based on various solvent mediators and polymer matrix compositions. Experiments on membranes XVII and XVIII were designed to determine any role by the calcium ion sensor itself. In all instances it is convenient to express the interferences in terms of the changes in e.m.f. on bringing the various background solutions up to M in anionic surfactants (Tables I11 and IV) a level approximating to that normally used in practice.Calcium ISEs with diferent solvent mediators Table I11 summarises the changes in e.m.f. for the respective electrodes according to membrane type and background solutions for added SDS and ABS. In general these show the interferences for the M calcium chloride solutions to be similar for each type of membrane for added SDS and is greatest for an electrode made from a commercial calcium ISE membrane (membrane XIX). E.m.f. changes for 10-1 M sodium chloride solutions are generally less pronounced and slightly less so again in the STP buffer system. The striking feature for added ABS is the considerably more pronounced e.m.f. changes for the 1 0 - 4 ~ and TABLE I11 ADDED SURFACTANT INTERFERENCES FOR CELLS WITH CALCIUM ELECTRODES IN VARIOUS SOLUTIONS AE (s.d.for n = 6) caused by 10-3 M surfactant in solutions/mV Membrane /- A . No. lo- M CaC1, SDS surfactant data-I -70 (3.4) I1 -24 (2.1) I11 -0.4 (0.1) IV -1 (0.5) V -45 (5.0) VI -3 (0.7) VI I -61 (3.6) VIII -75 (2.9) I X -30 (1.4) X -24 (3.6) XI -3 (0.3) XI1 -58 (3.8) XI11 -76 (4.0) XIV -81 (3.1) xv -86 (6.1) XVI -34 (3.1) XIX -113 (5.2) ABS surfactant data-I -65 (3.0) I11 -1 (1.1) VI -2 (1.0) XI -3 (0.1) X -21 (6.1) XI1 -60 (2.1) M CaC1, - 68 (4.1) -25 (3.0) -3 (1.7) -50 (3.7) -2 (0.4) -65 (2.9) -65 (6.1) -40 (2.1) -27 (2.9) -52 (2.9) -79 (3.6) -82 (5.2) -44 (2.7) -70 (3.6) -2 (1.0) -2 (1.2) -95 (3.9) -91 (4.0) -72 (2.1) -61 (1.7) -124 (5.1) -99 (6.1) -94 (4.7) 10-1 M NaCl -40 (2.5) -10 (1.3) -3 (2.5) -32 (2.1) -34 (1.7) -36 (3.1) -12 (1.7) -8 (1.3) -2 (0.4) -59 (1.5) -50 (2.9) -60 (3.8) -38 (3.1) -2 (1.1) -3 (0.9) -35 (3.7) --43 (2.1) -10 (1.4) -3 (0.5) -18 (3.1) -15 (2.1) -42 (3.1) STP buffer system' -32 (6.9) -2 (1.5) -2 (0.6) -20 (3.0) -28 (4.1) -44 (5.1) -15 (2.1) -8 (2.0) -29 (3.1) - 11 (2.0) -3 (1.0) -1 (1.0) ------23 (2.1) -11 (2.3) -2 (0.3) -13 (2.9) -7 (1.5) -20 (2.9 September 1983 ELECTRODES IN THE PRESENCE OF ANIONIC SURFACTANTS 1077 calcium chloride solutions compared with those for the This is a reflection of the considerably increased interference occurring at ca.3.5 x 1 0 - 4 ~ ABS level and observed for all membranes studied in a Such increased interference led to S-shaped plots for e.m.f.ve~sus [ABS]. The replacement of dioctyl phenylphosphonate solvent mediator with C1,,-Cl4 alkan-1-01s (membranes 11-IV) significantly reduces SDS interferences with improvement as the alkyl chain is lengthened (Table 111). Also there are significant transient increases in e.m.f. for the c1&6 alkan-1-01s (membranes 11-V) with associated lengthening of response times for reaching the equilibrium e.m.f. The reduced interferences for the alkan-1-01 mediator electrodes and previously noted for decan-l-ol,S suggest that the interference effect may be due to the different solubilities of calcium - surfactant complexes in the solvent mediator of the membrane. It is also possible that the interference is related to the leaching of the solvent mediator from the membrane surface by the surfactant in solution.The supporting evidence for this view is the stiffening of membrane I on soaking in SDS solution thus indicating the leaching of plasticising dioctyl phenylphosphonate from the membrane and is discussed here with the mechanistic data. Of considerable significance is the relatively small interferences shown by membrane VI for the circumstances of Table 111 that is except for M calcium chloride solution with added ABS. This is a considerable gain over the previously reported3 reduction of inter-ference by the partial replacement of dioctyl phenylphosphonate with decan-1-01 especially as the replacement of dioctyl phenylphosphonate by decan-1-01 is at the expense of lost calcium ion selectivity.6 Erosion of selectivity is less likely for the single solvent mediator of membrane VI and previous studies’ have shown the trioctyl phosphate solvent mediator to be compatible with good calcium over inorganic ion selectivity.The relatively small surfactant interference of membrane XI does not have the assured supportive quality of normal calcium ion selectivity nor of a single solvent mediator. M solutions. M calcium chloride background. Such “transients” are not observed at 35 “C. Calcium ISEs with diferent polymer matrices The data for membranes I and XIII-XVI (Table 111) indicate that the partial replacement of the poly(viny1 chloride) polymer matrix with either VAGH or poly(viny1idene chloride) does not offer any improvement in electrode behaviour.Calcium ISEs with diferent levels of sensor This series of experiments was based on membranes XVII and XVIII for SDS added to The notable effect here is that the level of calcium ion sensor has an effect on the level of interference by SDS to the extent that for low levels of free sensor-grafted sensor ratios (membranes XVII) and of free sensor to dioctyl phenyl-phosphonate ratios (membranes XVIII) the interference is manifested by increases in e.m.f. on adding M calcium chloride (Table IV). M SDS to the background M calcium chloride solution. TABLE IV ADDED SDS ( M) INTERFERENCES (e.m.f CHANGES) FOR CELLS WITH CALCIUM ELECTRODES I N M CALCIUM CHLORIDE Membrane type (see footnote $5 in Table I) A B C D E F A r \ E.m.f. changes for membranes XVII/mV .. . . 45 14 -16 -37 -51 -63 E.m.f. changes for membranes XVIII/mV . . . . 14 10 -46 -64 -68 -68 Based on the data for membranes XVII it was possible to design an electrode with the optimum free sensor to grafted sensor ratio. For this the experimental change in e.m.f. on making the background 1 0 - 2 ~ calcium chloride solution equivalent to 1 0 - 3 ~ in SDS was +1 mV. The null point for the system of membranes XVIII occurs at a free sensor to dioctyl phenylphosphonate ratio of 6 x but such a membrane is more difficult to reproduce and the electrodes from the membranes so fabricated gave experimental change 1078 FREND et d. STUDIES OF CALCIUM ION-SELECTIVE AndJJSt VOZ. 108 in e.m.f. of +3 to -4mV on making the background 1 0 - 2 ~ calcium chloride up to 1 0 - 3 ~ in SDS.Membrane XVII(A) is based on a grafted decyl phosphate sensor and matches the calcium ISEs based on free calcium bis(didecy1phosphate) with dioctyl phenylphosphonate in poly-(vinyl chloride)8J2 in general ISE performance. Because the grafted and ungrafted sensors give electrodes of similar lifetimess9l2 there is no advantage obtained by the long synthetic procedure in producing the grafted sensor membrane. Unfortunately calcium (mono [4-(1,1,3,3-tetramethylbutyl)phenyl]phosphate ] grafted to VAGH copolymer gives electrodes of poor calcium over sodium selectivity.8 TABLE V ADDED SURFACTANT RESPONSES FOR CELLS WITH REFERENCE TYPE ELECTRODES veYsm AN ORION 94-07 FLUORIDE ELECTRODE E.m.f. changes caused by M surfactant in solutions A and B/mV Orion 90-02-00 r A 7 Orion pH double-junction Silver - silver electrode electrode chloride electrode 7-7 * * Surfactant added A* Bt A B A B Sodium butylsulphate (SBS) .. . . 0 0 1 -0.7 9 5 Sodium dodecylsulphonate (SDS) . . 3 2 12 1.0 14 -49 Sodium tetradecylbenzene sulphonate (ABS) . . 2 4 35 0 25 -21 Sodium tetradecylsulphate (STS) . . . 20 -0.5 0 0 6.0 0.5 * A 10-3 M potassium fluoride + 10-5 M potassium chloride (unbuffered). t B as for A but buffered to pH 7. Effects of Anionic Surfactants on Reference Type Electrodes It is not easy to draw conclusions from the data on e.m.f. changes of various reference type electrodes used in conjunction with the Orion Model 94-07 fluoride ISE brought about by 10-3 M surfactant in solutions A and B (Table V).However the Orion Model 90-02-00, double-junction electrode used as reference in this work is insensitive to added surfactant in the pH buffered solutions B that is assuming that the fluoride electrode is also insensitive and does not suffer from a compensating interference. The responses in the unbuffered solutions are not as reassuring. Analysis of Calcium Ions in Detergent Solutions Calcium ion determination was carried out by spiking calcium chloride into solutions of model soap powder at various levels of added soap powder. The calcium ion levels deter-mined using calcium ISEs with selected membranes (I VI VIII IX X and XI) for 1 2 3, 4 and 5 mM of added calcium chloride are compared with calculated values13 in Table VI. TABLE VI FREE Ca2+ ACTIVITIES IN SIMULATED WASH LIQUORS FOUND BY CALCIUM ELECTRODES OF VARIOUS MEMBRANES Experimental aca8+ (s.d.for n = 6) determined with electrodes of various membranes M I* VI+ VIII* IX* X* XI* M aCa'+ (calculated)/ [Caz+] added] r- -3 g 1-1 model soap powder added-1 1.6 (1.5) x 10-77 1.2 (0.6) x lo-' 1.9 (1.3) x lO-?t 2.3 (0.4) X 3.0 (0.6) X lo-?? 4.1 (1.1) x lo-?? 1.1 x lo-" 2 4.5 (2.1) x lo-?? 3.6 (1.3) x lo-' 5.9 (2.0) x 10-lt 6.9 (0.9) x lo-'? 2.0 (0.9) x 2.5 (1.5) x 3.6 x 3 6.0 (1.0) x l O - 7 t 3.6 (1.3) x 8.3 (2.7) x lO-?t 4.1 (0.7) x 9.0 (1.5) X 1.1 (0.3) x 4.3 x 4 1.2 (0.6) x 10-6 1.0 (0.4) x 2.1 (0.6) x 10-6 1.0 (0.6) x lo-'. 2.3 (0.8) x 3.1 (1.0) x 1.6 x lo-' 5 1.7 (0.8) x 2.1 (0.6) x lo-' 3.1 (0.9) x lo-" 1.9 (0.8) x lo-* 4.6 (1.6) x 5.9 (1.1) x 1.9 x 5 g 1-1 model soap powder added-1 1.0 (0.4) x 10-7t 2.7 (1.0) x 10-0 1.1 (0.6) x 10-7t 4.9 (1.9) x 10-7t 1.8 (0.5) x 10-7t 2.6 (0.8) x 10-7t 8.9 x 10-7 2 1.5 (0.7) x 10-7t 3.5 10.51 x 10-8 2.1 10.71 x 10-7t 7.2 12.51 x 10-7't 4.4 (0.9) x 10-7t 4.9 (1.01 x 10-7t 2.5 x 10-0 3 6.9 (i.6j x 10-7+ 4.9 (0.4j x 10-6 6.5 (i.7j x lo-'+ 1.0 (0.2j x 10-6' 1.0 (0.3j x 10-6' 1.6 (0.5j x 10-6' 5.4 x 10-6 4 8.7 (0.4) x lo-'? 8.7 (2.4) x 10-6 9.0 (2.7) x lo-?? 1.9 (0.1) x lo-' 2.5 (0.9) X 3.4 (0.9) x 1.4 x lo-& 5 1.1 (0.6) x 10-o 2.0 (0.8) x 1.5 (0.6) x lo-" 5.3 (0.9) x 6.1 (0.2) x 10-o 7.5 (2.0) X lo-" 7.7 x Figures in parentheses are s.d.(n = 6). t Determined by extrapolation of calibration line September 1983 ELECTRODES IN THE PRESENCE OF ANIONIC SURFACTANTS 1079 Electrodes based on membranes I and VIII perform poorly in the simulated wash liquors.Improved performance is observed when the membranes contain mixtures of dioctyl phenyl-phosphonate with decan-1-01 (IX) dodecan-1-01 (X) and tetradecan-1-01 (XI). However the outstanding membrane is the one based on the trioctyl phosphate solvent mediator thus confirming the superiority observed in the solvent mediator parameter studies as reported under Results and Discussion. The electrode gave measured calcium ion levels close to the calculated values especially at the 3 g 1-1 level of powder. Mechanistic Studies The various experiments concerning interferences of calcium ISEs by anionic surfactants indicate apart from any other factors that the solvent mediator is significantly involved.Further Oesch and Simon1* have indicated that the loss of solvent mediator and/or electro-active species from the polymeric membrane phase into solution determines the lifetimes of electrodes. In view of the important role of dioctyl phenylphosphonate solvent mediator in imparting calcium ion selectivity when used with organophosphate calcium ion sensors it is pertinent to determine whether the anionic surfactant interference may be aggravated by the loss of these membrane components. This has been attempted here by X-ray fluorescence and chromatographic studies. (A) Expansion I h (B) Expansion (C) Expansion I I P CI Ca Energy+ Fig. 1. XRF spectra of recast Membrane I quadrants leached by A de-ionised water; B, 5 x 10-3 M SDS; and C 10-1 M SDS 1080 FREND et at.STUDIES OF CALCIUM ION-SELECTIVE Analyst VOl. 108 X-ray fluorescence (XRF) The XRF spectra of recast membrane I quandrants subjected to leaching by de-ionised water 5 x Fig. 2 shows the XRF spectra of specially fabricated membranes of (a) calcium sensor + 2-nitro-phenyl phenyl ether + poly(viny1 chloride) and (b) dioctyl phenylphosphonate + poly(viny1 chloride) after exposure to de-ionised water and to 10-1 M SDS. Fig. 1 shows a progressive decrease in the phosphorus content of the membrane surface in passing from (A) to (C). It is only in (C) that the calcium peak is decreased significantly. Fig. 2 confirms these observations. Considering the spectra together in relation to the normal levels of surfactant in solution (10-3~) in the main experiments above it can be deduced that the leaching of the solvent mediator is likely to be significant.However it has to be noted that the treatments in the XRF studies have been severe and prolonged. Neverthe-less a change in the dioctyl phenylphosphonate concentration at the membrane surface can be significant in altering the boundary potential. The thin-layer chromatography experiments on leached residues from membrane I confirm that the surfactant promotes the leaching of dioctyl phenylphosphonate (RF = 0.83) and sensor (R = 0.21) from the membrane at 10-1 M SDS. Under the milder conditions pro-vided by 5 x M SDS and 10-1 M SDS are shown in Fig. 1 (A Band C respectively). M SDS only dioctyl phenylphosphate can be detected in the leachate. P CI Ca P CI Fig.2. XRF spectra of specially fabricated membranes of (a) calcium bis{di [-4- ( 1 1,3,3-tetramethylbutyl) phenyllphosphate} and 2-nitrophenyl phenyl ether in poly(viny1 chloride) and (b) dioctyl phenylphosphonate in poly(viny1 chloride) after exposure to (1) de-ionised water and (2) 10-1 M SDS. Energy ___+ Gas chromatography studies These were directed to determining leaching of dioctyl phenylphosphonate from membrane I under normal interference conditions rather than the prolonged exposures of the XRF and TLC studies. Standard dioctyl phenylphosphonate yielded two gas - liquid chromatographic peaks with retention times of 5.0 and 12.0 min respectively. Peaks of the same retention times were obtained for the calcium - SDS extract as described under Experimental but no such peaks were observed when SDS was absent.Such peaks confirm the synergistic effect of SDS in leaching dioctyl phenylphosphonate. Conclusion This work shows the composition of the poly(viny1 chloride) matrix membrane system organophosphate-based calcium ISEs to be critical when anionic surfactants are present. The studies directed to the normal calcium ISEs based on membrane I that is the calcium bis{ di [a-( 1,1,3,3-tetramethylbutyljphenyl]phosphate) sensor with dioctyl phenylphosphonate in poly(viny1 chloride) indicates that the presence of anionic surfactant at the M leve September 1983 ELECTRODES IN THE PRESENCE OF ANIONIC SURFACTANTS 1081 has a detrimental effect on membrane components not otherwise observed in the absence of surf act ant.An important outcome of the potentiometric studies is the considerably reduced anionic surfactant interferences by calcium ISEs based on calcium bis{ di[4-(l11,3,3-tetramethyl-butyl)phenyl]phosphate) with trioctyl phosphate as solvent mediator (membrane VI). This can mark a significant step foward in the determination of free calcium ions in the presence of anionic surfactants. The authors thank the Science and Engineering Research Council for a studentship (to A. J.F.) under the scheme for Co-operative Awards in Science and Engineering in conjunction with Unilever Research Laboratory Port Sunlight; Mr. R. Lee of Unilever Research Labora-tories for obtaining the XRF spectra; and Union Carbide (UK) and ICI Ltd. are thanked for their gifts of VAGH copolymer and poly(viny1idene chloride) homopolymer respectively. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Birch B. J. and Clarke D. E. Anal. Chim. Acta 1973 67 387. Llenado R. A. Anal. Chem. 1975 47 2243. Craggs A. Moody G. J. Thomas J. D. R. and Birch B. J. Analyst 1980 105 426. Moody G. J. Oke R. B. and Thomas J . D. R. Analyst 1970 95 910. Craggs A. Moody G. J. and Thomas J . D. R. J . Chem. Educ. 1974 51 541. Craggs A. Keil L. Moody G. J. and Thomas J . D. R. Talanta 1975 22 907. Moody G. J. Nassory N. S. and Thomas J. D. R. Analyst 1978 103 68. Hobby P. C. Moody G. J. and Thomas J . D. R. Analyst 1983 108 581. Craggs A. Delduca P. G. Keil L. Key B. J. Moody G. J. and Thomas J. D. R. J . Inorg. Nucl. Craggs A. Delduca P. G. Keil L. Moody G. J. and Thomas J. D. R. J . Inorg. Nucl. Chem., Moody G. J. and Thomas J. D. R. Ion-Sel. Electrode Rev. 1979 1 3. Keil L. Moody G. J. and Thomas J. D. R. AnaZyst 1977 102 274. Birch B. J . personal communication. Oesch U. and Simon W. Anal. Chem. 1980 52 692. Chem. 1978 40 1483. 1978 40 1943. Received January 26th 1983 Accepted April 15th 198