Boundary Layers Between Silver Iodide and Aqueous Solutions at Low Temperatures BY B. VINCENT * AND J. LYKLEMA Laboratory for Physical and Colloid Chemistry Agricultural University de Dreijen 6 Wageningen The Netherlands Received 14th July 1970 Potentiometric titration studies have been made on AgI suspensions over the temperature range 0-20°C in the presence of LiN03 KN03 and RbN03. Points-of-zero-charge (P.z.c.) double-layer capacities and surface entropy data derived from the titration results indicate an increase in water " structure " as the temperature tends to O'C around the P.Z.C. region and particularly on the positive side. For strongly negativeIy charged surfaces the charge ordering effect on the dipoles tends to break down this structuring. The results of flocculation studies on AgI sols have been combined with the titration data to yield information about changes in double layer parameters with temperature.The results fit the suggestions made regarding changes in water structuring at the interface. During the last decades considerable progress has been made in the field of inter- facial electrochemistry especially with mercury electrodes. One of the important results has been that structural features of the liquid adjacent to the electrode e.g. the direction of orientation of water dipoles are reflected in measurable electro- chemical quantities such as the differential double-layer capacitance and the point of zero charge (P.z.c.). Hence inference on the structural properties of the interphase can be derived from these and other electrochemical measurements.Studies with non-metallic charge carriers are more scanty and as a rule less accurate but they have the advantage that for certain systems stable dispersions can be made. In those cases the double-layer measurements may be amplified by stability studies. The AgT-system has received special attention because stable AgT sols can be made as well as reproducibly operating AgI electrodes. Moreover its interfacial structural properties are of interest because AgI is a powerful cloud seeder. Some preliminary information on the boundary layer structure of this system has now been obtained by combining double-layer and stability studies as a function of temperature.' In view of the expectation that structure formation in the interface would be promoted by a decrease in temperature these measurements have now been extended to cover the temperature range down to 0°C.EXPERIMENTAL MATERIALS AgI suspensions were prepared by addition of 0.1 M AgNO to an equal volume of Surface areas were determined for each suspension used by * present address Dept. of Physical Chemistry University of Bristol Cantock's Close Bristol 0.1 M KI in the usual way.* BS8 1TS ; or I.C.I. Paints Division Ltd. Wexham Road Slough Bucks. 148 B . VINCENT AND J . LYKLEMA 149 comparison of surface charge values in C g-l from the experimental titration curve in lo-' M KNO, with surface charge values in pC cm-2 from standard curves for this system. Most of the experiments reported here were carried out on a suspension of surface area 1.34 m2 g-l. Sols were prepared by the similar addition of 2 .0 ~ M KI. Both suspensions and sols were aged at 80°C for 3 days. All salts used were of A.R. quality AgN03 KI LiNO, Union Chimique Belge ; KN03 Baker ; RbNO, Merck. The RbN0 was recrystallized twice from water. All the other salts were used as supplied. Water was doubly-distilled and passed down a AgI column before use. M AgNO to 2 . 2 ~ POTENTIOMETRIC TITRATIONS The potentiometric titration technique by which (surface charge pAg) or (surface charge PI) curves may be obtained for aqueous AgI suspensions has been extensively disc~ssed.~9 The apparatus used here was essentially similar except that the titration cell was constructed so that it could be completely immersed in a cryostat bath (Colora Ultra Cryostat KT 40 S). In this way the temperature of the cell was controlled to fO.l"C over a range 0-20°C.Potential measurements (f0.2 mV) were made using a potentiometer (Knick Praisiziono pH Metre type pH 34). The average reading of 4 or 5 Ag/AgI electrodes was taken. Calibration was carried out using standard solutions at PI 4.0 5.0 and pAg 4.0 5.0. The cell resistance was also checked periodically during a run (Phillips Bridge Gh 4249) to ensure that there was no blocking of the salt bridge capillary by AgI particles. Titration runs were performed in the following manner. A sample of concentrated suspension of known composition was weighed in a graduated flask. This gave the weight of AgI without resort to drying. The suspension was washed into the titration vessel with lo- M salt solution. A set of titrations was then carried out in the following order (1) 20"C+ -+ -addition of KI ; (2) 10°C- -++addition of AgN03 ; (3) O"C+ -+ -addition of KI ; (4) 20°C-+addition of AgN0,.At this point further inert electrolyte was added and the sequence repeated. STABILITY MEASUREMENTS These were performed using the " kinetic " method of Reerink and Overbeek.6 The change in optical density (at A = 556 nm) of the flocculating system was followed using a Vitatron spectrophotometer (Vitatron N.V. Dieren The Netherlands). This instrument incorporates a magnetic stirrer device. The output was fed to a Kipp Micrograph pen- recorder (Kipp N.V. Delft The Netherlands). The spectrophotometer cell-housing was thermostatted by pumping cooling liquid from the cryostat around its jacket. Temperature control was f0.5"C but at ambient temperature differentials of more than about 5°C it was necessary to enclose the apparatus within a dry-box to prevent condensation on the optical cell.The flocculation runs were carried out in cylindrical optical cells (0.9 ern diam.). 0.5 nil sol (AgI concentration 10 mM) were added with the aid of a syringe to 2 ml of salt solution in the cell which also contained a small glass-covered magnetic stirrer. Both the sols and salt solutions were adjusted to the necessary PI and stored at the required tempera- ture prior to use. RESULTS AND DISCUSSION SOLUBILITY PRODUCT L AND POINT OF ZERO CHARGE (P.z.c.) OF AgI pL values for AgI over the temperature range studied and at salt concentrations 10-3-100 M were calculated from the standard electrode potentials. These are shown in fig. 1 together with the values obtained by other auth0rs.l.There is an apparent continuous increase in the solubility of AgI with increase in temperature and with increase in inert electrolyte concentration. 150 AgI AT LOW TEMPERATURES The P.Z.C. (pAg") values (see table 1) was taken for each system as the mutual intersection point of the experimental "charge " (k charge with respect to an arbitrarily chosen reference point) against pAg curves. This point was generally well defined to within 0.2 pAg units except that the 100 M curves tcnded to cross the oo = 0 line at slightly higher pAg value. 1; I I d I I I 1 I I \ 10 2 0 3 0 4 0 5 0 6 0 temperature "C FIG. 1.-Solubility product (expressed as pL) of AgI as a function of temperature. I Lyklema,' lo-' M (with M ; 111 this work M K biphthalate present) ; 11 Honig and Hengst,' (a) 10-3 M (b) 10-1 M (c) 100 M.TABLE 1.-pAg VALUES FOR AgI authors ref. method salt ooc 10°C 2O0C 1 this work titration LiN03 7.10 6.40 5.80 KNO 6.95 6.20 5.65 2 Honig and Hengst 7 suspension effect KN03 5.51 5.51 5.52 3 Fairhurst 8 electrophoresis KNO - 5.54 5.40 RbN03 6.75 6.10 5.50 The three sets of data show apparent discrepancies. This work indicates a significant rise in pAg" with decreasing temperature. Those of Honig and Fairhurst are virtually independent of temperature over this range. Titration results however yield true P.Z.C. values whereas the suspension effect and electrophoresis give iso- electric points (i.e.p.). At higher temperatures the P.Z.C. values of Lyklema and the i.e.p. of Honig again diverge (e.g.,at 60°C; P.Z.C. = 5.45; i.e.p. = 4.38).* The were made in the presence of K-biphthalate and biphthalate ions tend to absorb specifically at the P.Z.C.which would in turn tend to move the P.Z.C. towards a higher pAg" value (e.g. at 20°C P.Z.C. = 5.72). This specific adsorption would presumably however decrease with increasing temperature whereas the divergence between i.e.p. and P.Z.C. increases. * Lyklema's measurements B . VINCENT A N D J . LYKLEMA 151 difference between the P.Z.C. and the i.e.p. appears to be minimal at about 20°C. As the solubility product itself shows no break around 20°C slructural changes in the solid Agl can hardly be invoked to explain the observed behaviour. A tentative explanation for the increase in pAg" on going from 20 to O"C as observed by us is as follows. At 20°C the P.Z.C.is very asymmetrical pAg" = 5.6 PI" = 10.6 pL (= pAg"+pI") = 16.2. Much less iodide is needed than silver to make the surface uncharged. At the same time the adsorption of Ag+ at the positive side of the P.Z.C. is very strong whereas positive AgI sols are rather unstable. One way of explaining these facts is to assume that at 20"CAgf absorbs largely in an associated form i.e. as AgN03. In the first place this would support the result that only relatively small amounts of I- are required to make the surface negative ; in the second place it explains the high adsorption at the positive side without concomitant rise in potential. At 0°C the P.Z.C. is less asymmetrical pAg" = 6.9 PI" = 10.5 pL = 17.4. At the same time the adsorbability of Ag+ at the positive side is reduced (see titrations).The implication is that at 0°C these AgN03 ion pairs are more dissociated than at 20°C. Perhaps therefore due to an increase in water " structuring " at the interface the NO; ions are less easily adsorbed in the Stern layer. This explanation is corroborated by the surface excess entropy calculations (see later). POTENTIOMETRIC TITRATION RESULTS The titration curves for LiN03 KN03 and RbN03 are given in fig. 2 3 and 4 respectively. At low inert electrolyte concentrations ( lov3 M,) (do,/dT),, is negative. This is expected since here the diffuse layer term dominates the total double-layer capacity. There is a tendency however noticeable at low negative surface charges and particularly so on the positive side for (do,/dT),, to reverse sign in 10-1 and 1 M salt. By comparison with the curve at higher temperatures,l (do/dT),A becomes zero around 20"C i.e.the same region where the anomalous increase in pAg" begins. The differential double layer capacity values at the P.Z.C. are presented in table 2. TABLE 2.-DIFFERENTIAL DOUBLE LAYER CAPACITIES AT P.Z.C. (pF/cm2) 20°C 10°C 0°C i.e.c. LiNO3 KN03 RbN03 LiNO3 KNO3 RbNO3 LiNO3 KNO3 RbN03 loo M 18.0 24.8 30.0 1 9 . 2 23.5 29.0 17.7 20.5 22.5 lo-' M 1 5 . 0 17.5 23.0 16.6 17.0 1 8 . 0 14.0 16.0 16.0 M 9.5 10.2 11.6 10.7 11.5 12.5 9.5 10.0 10.0 M 5.4 6.0 6.2 6.0 6.0 6.0 6.6 6.6 6.6 The apparent specificity with regard to the nature of the cation particularly at higher salt concentrations where the capacity is dominated by the Stern layer term increasing in the order Li+<K+<Rb+ indicates some absorption of cations at the P.Z.C.as well as anions. This is also reflected in the slight dependency of the P.Z.C. on the nature of the cation. The decrease in capacity with lowering of temperature particularly from 10 to O"C again points to desorption of NO;-ions. (Since the decrease is relatively smaller for LiN03 this might indicate some desorption of the strongly hydrated Lif ions also). Desorption of NO ions on lowering the temperature from 20 to 0°C was also indicated on the positive side of the zero point of charge from a preliminary components-of-charge analysis of the type suggested by Lyklema. 152 AgI AT LOW TEMPERATURES - 5 -4 n N - I FIG. 2.-Surface chargepotential curves for AgI in the presence of various concentrations of LiN03 presence of various concentrations of LiN03 A 10" M ; B 10-1 M ; C M.- 20°C ; -- 10°C ; . . . 0°C. M ; D (E- Ep.z.c.)/mV FIG. 3.-Surface charge-potential curves for AgI in the presence of various concentrations of KN03 A 100 M ; B 10-1 M ; c 10-2 M ; D 10-3 M. - 2 0 0 ~ ; -- iooc; ... ooc. The differential capacity is much less temperature dependent at high negative surface charge when any water " structuring " would be opposed by the orientating effect on the water dipoles (e.g. in 10-1 M RbN03 +o = -250 mV C = 12.8 (20°C) 13.0 (lO°C) 13.2 (OOC) ,uF/cm2). INTERFACIAL ENTROPIES The derivative of the excess entropy of the AgT/solution interface with respect to the surface potential $o can be calculated from the temperature dependence of the B . VINCENT AND J . LYKLEMA 153 surface charge using either one of two equations derived previously by Bijsterbosch and Lyklema.O (af/a$O)T',as = -s"(CI-/F) + ( a a ~ / a T ) s p A g a s - ( a ~ O / a ~ s ) p * g T ( - s ~ + In - (C/F)(-si,+ +R ln a&+) (1) ( a f / a $ 0 ) T a s = sa(CAg+IF) +(a~O/aT)pI,asf (aaO/a&)pI,T(-s,"+ In (C/F)(-s;- + R In a*-) (2) In these equations the interfacial excess entropy q" is defined by where s denotes molar entropies the superscipts a and p apply to the solid (AgI) and liquid (W) phase respectively and subscript s refers to the salt. CI- and CAg+ are the contributions of the iodide and silver ions respectively to the total differential capacitance C. The way in which one splits C up into its component parts is immaterial. O Eqn (1) and (2) are equivalent the only difference being that in (1) the Ag+ ion is taken to be the potential-determining species whereas in (2) it is the I- ion.The equivalence of the two equations was corroborated by the computations although individual terms differed sometimes by a hundredfold the final values for (af/a$o)T,as agreed within a few percent. The general behaviour of (af/i?$o)T,a is to a large extent determined by the (dao/dT)pAg,as term. Further details on the method of computation are given in ref. (10). '1" = s' - saTAgI - SBrw S" = ( S - S" - SB)/A (3) (4) - 5 - 4 n $ - 3 Y W 1 g -2 - I / .' / I + 2 A -100 - 2 0 0 - 3 0 0 FIG. 4.-Surface charge-potential curves for AgI in the presence of various concentrations of RbN03 A 100 M; B 10-1 M; c 10-2 M; D 10-3 M. - 2 0 0 ~ ; -- iooc; ... ooc. The values obtained for (~r,f'/a$o),,,s were plotted as a function of $o graphically integrated and finally plotted as a function of a.using the P.Z.C. as the reference point. The results are given in fig. 5 together with some results obtained previously at higher temperatures in the presence of lQ-3 M K biphthalate.'O The data shown 154 AgI AT LOW TEMPERATURES are for LiN03 since this would be expected to reveal structural changes in the inter- facial water more strongly than KN03 or RbN03 because the masking effect of specific adsorption on the negative side is least for the Li+ ion. (The curves for KN03 and RbN03 are in fact similar but the trends are less marked.) b F -. -_ - - - - -_____ '. . o ;5Oc 8 5 c - 0 . 5 00 pWm2 FIG. 5.-Tnterfacial excess entropy with respect to the point of zero charge as a function of surface charge for decimolar solutions of LiN03 - current work ; -- previous work lo in which K-biphthalate was also present.Interfacial entropies are composite and complex quantities. One should therefore not over-interpret the data. Nevertheless some interesting trends seem to emerge. First even around 20°C the variation of y" with o0 is more than tenfold the amount that could be accounted for by the adsorption entropies of counterions calculated from their specific adsorption energies. Its order of magnitude suggests changes in interfacial water structure. A remarkable feature is that at low temperatures y' increases with increasing negative surface charge whereas at higher temperatures the trend is the reverse. At high temperatures where hydrogen-bonded structuring will be weak this has been interpreted lo as being due to structure promotion resulting from the increased polarizing action of the field as the surface is made more negative.This is to be combined with structure breaking effects of the NO -counter-ions adsorbed strongly at the positive side. At low temperatures the situation is reversed. The decrease in entropy with increasing positive charge in combination with the marked decrease in capacitance suggest again some expulsion of NOT-ions with an increasing hydrogen-bonded structure formation. This is in agreement with the interpretation offered for the shift in the P.Z.C. Given the observed trends it follows that at a temperature around or slightly above 20°C a transition must take place where y" is largely independent of go. (The exact temperature cannot be deduced from the data given because the high temperature curves are obtained in the presence of biphthalate).At this temperature with increas- ing negative charge the two opposing trends i.e. the increase in dipole ordering and decrease in water structuring would tend to cancel each other. STABILITY The stability of a hydrophobic sol measured as the flocculation concentration (cf in mmol/l.) reflects the charge and potential distribution in the electrical double B . VINCENT AND J . LYKLEMA 155 layer. As a consequence stability data can serve as an additional source of informa- tion. However its applicability with respect to the detection of structure formation is limited because stable sols can only be made at sufficient high surface charge i.e. in the region where any effect of structure formation is relatively weak.There are two ways in which structure formation would be expected to affect cf (i) electrostatically because if there is structure formation the extent of counterion adsorption and hence the potential of the outer Helmholtz phase t+hd is changed; (ii) kinetically because the rate of counterion transfer from the diffuse part of the double layer to the Stern layer during the encounter of the particles as required by DLVO-theory is reduced. In fig. 6 and 7 flocculation concentrations under various conditions and the corresponding double layer parameters are shown. The flocculation concentration apparently decreases with decreasing temperature at PI 4.6 but the opposite is true at PI 6.6. The surface charge shows a slight maximum at around 10°C for all three cations at both PI.The trends in cf and o0 are not parallel. This is not expected a priori since cf reflects the charge distribution whereas go is the total charge. In order to analyze these data further Stern potentials specific adsorption inner layer capacities and dielectric constants have been calculated using an analysis to be discussed elsewhere,ll but which may be summarized briefly as follows. From the flocculation data and an ionic components of charge analysis on well-established potentiometric titration data for KN03 at 20°C a value of A = 2 . 4 0 ~ J was established for the Hamaker constant for AgI in aqueous systems. This value was then used to calculate the Stern potentials and hence the other double-layer parameters from the flocculation concentrations and titration data.2 i- I I I 0 10 2 0 temperature "C FIG. 6.-Experimental parameters under various conditions (a) flocculation concentrations ; (b) surface charge. counter ion type LiN03 -; KN03 ... ; RnN03 .... PI 4.6( x) ; 6.6(0). The specific adsorption curves (expressed as a percentage of the surface charge) indicate little or no dependence on temperature at PI 4.6 but at PI 6.6 there appears to be a definite fall-off in specific adsorption of cations on lowering the temperature from 20 to 0°C. Thus at highly negative surface charges (as at PI 4.6) the specific adsorption is largely governed by electrostatic forces and is therefore largely tempera- ture independent over this range but at lesser negative surface charges (as at PI 6.6) there is at least some effect from water structuring leading to some desorption of counter ions on lowering the temperature.The Stern-layer capacity falls slightly with increasing temperature and is in the order Rb+>K+>Li+ as expected but is also in the order PI 6 . 6 ~ ~ 1 4.6. Since specific adsorption increases with decreasing PI and there are probably no significant changes in inner layer thickness this could reflect changes in the inner layer dielectric 156 AgI AT LOW TEMPERATURES constant. (At such strong specific adsorptions (-50 %) the occupancy of the Stern layer is such that the polarizabilities of the adsorbed ions themselves probably contribute significantly to the dielectric constant of the inner layer.) The mean value at 20°C 7.5& 1.5 corresponds well with the value of about 6 quoted for the inner layer at the Hg+water intex-face.12 The mean value rises to about 8.0+ 1.5 at O'C which again supports the idea of increased hydrogen-bonded structuring of the inter- facial water although only slight at these negative charges.22 r 7Ol- ................. x.. ....................... x 0 ..................... ...... .... __- - 40 19- '.. -. (0) I 10 0 2 0 10 It3 2 0 0 20- - 10 ........ .... ........ ....... .... ......... . . . . . . . 9 - .... ......... ..... .... .o ... 0 '0 I& " ' >----- 'o-----_o ' I ' ' ' . . ""0 --A- -_____ -x X - I I (j. (d) 3 0 12- 2 0 10 2 0 0 10 ternperaturel'c FIG. 7.-Derived double layer parameters at the flocculation concentration (a) Stern potential ; (b) specific adsorption (expressed as a percentage of surface charge) ; (c) Stern-layer capacity ; (d) Stern-layer dielectric constant.Counter-ion type LiN03 -; KN03 ... ; RbN03 .... PI 4.6 ( X) ; 6.6 (0). CONCLUSIONS It is generally observed that adsorption increases with decreasing temperature. For the adsorption of I-ions on AgI this trend was c0nfirmed.l The more detailed study of the adsorption of potential determining ions reported in this paper shows that this trend is not followed at temperatures between 0 and 20°C. At the negative side of the P.Z.C. the adsorbed amount (and hence the surface charge a,) is almost independent of temperature whereas on the positive side a. tends to decrease with decreasing temperature. The effect is that of increased structure formation in the aqueous layer adjacent to the AgI surface at less negative or more positive charges together with consequent desorption of NOT-ions.It is fully corroborated by surface excess entropy calculations. Flocculation experiments are less informative because only data at high negative surface charges are available. In this region the structure formation effect is less pronounced. However the stability trends observed at lower negative surface charge agree well with the theory suggested above. There remains the discrepancy between p.z.c.-values from titration and i.e.p.-values from the suspension effect and B . VINCENT AND J . LYKLEMA 157 electrophoresis. The experimental data seem to be well established. Unfortunately we cannot offer a satisfactory explanation for this. We thank the Royal Society for the provision of an Overseas Fellowship to one of us (B.V.) and also thank Miss Olga van Hiele for technical assistance. ' J. Lyklema Disc. Faraday SOC. 1966 42 81. see e.g. B. H. Bijsterbosch and J. Lyklema J . Colloid Sci. 1965 20 665. Int. Crit. Tables to be published. J. Lyklema and J. Th. Overbeek J . Colloid Sci. 1961 16 595. G. L. Mackor Rec. Trav. Chim. 1951 70,763. H. Reerink and J. Th. Overbeek Disc. Farduy SOC. 1954 IS 74. ' E. P. Honig and J. H. Th. Hengst J. Colloid Interface Sci. 1969 30 109. ' A. L. Smith private communication of unpublished work by D. Fairhurst. l o B. H. Bijsterbosch and J. Lyklema J. Colloid Interface Sci. 1968 28 506. l 2 J. O'M. Bockris M. A. V. Devanathan and K. Muller Proc. Roy. SOC. A 1963 274 55. J. Lyklema Trans. Faraday SOC. 1963 59,418. B. Vincent B. H. Bijsterbosch and J. Lyklema to be published.