CROATICA CHEMICA ACTA CCACAA 79 (1) 95¿106 (2006) ISSN-0011-1643 CCA-3069 Original Scientific Paper
Point of Zero Charge and Surface Charge Density of TiO2 in Aqueous Electrolyte Solution as Obtained by Potentiometric Mass Titration*
Tajana Preo~anin** and Nikola Kallay
Laboratory of Physical Chemistry, Department of Chemistry, Faculty of Science, University of Zagreb, Horvatovac 102 a, 10000 Zagreb, Croatia
RECEIVED OCTOBER 7, 2005; REVISED NOVEMBER 22, 2005; ACCEPTED NOVEMBER 23, 2005
Mass titration method was developed as a suitable tool for determination of the point of zero s charge (p.z.c.) and surface charge density ( 0) of metal oxide colloid particles at different ionic strengths. In the course of mass titration, subsequent portions of a metal oxide powder are added to an aqueous electrolyte solution, and pH of the equilibrated dispersion is measured. The pH of the system changes gradually and approaches a constant value of pH¥, which is in the case of a metal oxide free of impurities equal to the point of zero charge pHpzc. Counterion association shifts the pHpzc either to the acidic region (preferential adsorption of cations) or to the basic region (preferential adsorption of anions). Mass titration therefore enables detection of the difference between association affinities of counterions (cations and anions), which is important information about the equilibrium within the electrical interfacial layer. Such an analysis is not possible by the conventional acid-base potentiometric titration of the dispersion, since the location of the p.z.c. is based on the common intersection point (c.i.p.) of the data ob- tained at different ionic strengths. It is shown that c.i.p. may be used for locating the p.z.c. only Keywords in a »symmetric case«, when affinities of the anions and cations to associate with oppositely mass titration charged surface groups are equal. Analysis of the mass titration was performed on the basis of point of zero charge experimental data obtained with colloidal titania dispersed in an aqueous sodium chloride solu- surface charge density tion, and also by numerical simulation based on the Surface Complexation Model and the 2-pK counterion association mechanism of surface reactions. Increase in the NaCl concentration shifted the point of zero titan dioxide charge to the basic region, while the isoelectric point was shifted to the acidic region, indicat- sodium chloride ing higher association affinity of chloride ions compared to sodium ions.
INTRODUCTION measure the pH of the dispersion. The pH of the system changes gradually and approaches a constant value pH¥. Potentiometric mass titration was originally developed In the case of a pure metal oxide powder (absence of 1 by Noh and Schwartz as a method for determination of acidic or basic contamination), pH¥ is equal to the point the point of zero charge (p.z.c.) of pure metal oxides (in of zero charge pHpzc. Mass titration was successfully ap- the absence of acidic or basic impurities). The method is plied for determination of the point of zero charge of simple. One should add subsequent portions of a metal colloidal particles2 with low3 and high specific surface oxide powder to the electrolyte solution (or water) and areas (activated carbon).4 Mass titration was also propos-
* Dedicated to the memory of the late Professor Marko Branica. ** Author to whom correspondence should be addressed. (E-mail: [email protected]) 96 T. PREO^ANIN AND N. KALLAY ed for the point of zero charge determination of metal of active surface sites with the potential determining ions. oxide mixtures. The point of zero charge of a metal ox- For metal oxides, the potential determining ions are H+ ide mixture corresponds to the pH where the net surface and OH– ions. Within the Surface Complexation Model charge of the heterogeneous mixture is zero, while one (SCM), the 2-pK mechanism14,15 may be used for numeri- oxide bears a positive and the other a negative charge. cal simulation and interpretation of experimental data. Ac- For calculation, one needs to know the specific surface cordingly, surface charge is a result of the two-step pro- areas, mass fractions of components, and points of zero tonation of the surface groups produced by hydration of 5,6 charge of both metal oxides. The mass titration method the metal oxide surface was extended by Kallay and @alac7 for determination of ºMO– + ® ºMOH; o the point of zero charge of contaminated samples. The +H K1 (1) pH¥ value of contaminated dispersion is higher (basic º + ®º + o impurities) or lower (acidic impurities) compared to the MOH+H MOH2 ; K2 (2) point of zero charge. Interpretation of the mass titration o o provides information on the fraction of impurities in the where K1 and K2 are thermodynamic equilibrium con- powder and also on the point of zero charge.8 Mass titra- stants of the corresponding surface reactions. Generally, tion was found extremely suitable for determination of a thermodynamic equilibrium constant Ko is defined16 in the temperature dependency of the point of zero charge. terms of equilibrium activities (a) of species J involved One simply measures the temperature dependency of pH in the chemical reaction of a sufficiently concentrated metal oxide dispersion in the absence of impurities. The mass titration method was o n J K = Õ aJ (3) further developed for determination of surface charge den- J sity.9 The advantage of this method is that experiments For interfacial species S activity may be defined17 in can be performed at extremely low ionic strengths. terms of surface concentration G (amount of surface spe- The common method for determination of the point cies per surface area), so that of zero charge is the potentiometric acid-base titration of 10,11 G S { } a dispersion. Comparison with blank titration in the aS = yS = yS S (4) absence of the dispersed solid oxide phase yields rela- G o { } tive values of the surface charge densities. Absolute val- where yS is the activity coefficient, and S denotes the ues of the surface charge densities are then obtained by relative value of surface concentration of S with respect setting the zero value at the common intersection point to the chosen standard value. There is no recommended (c.i.p.) for different ionic strengths. However, this meth- value for the standard surface concentration, but G o = od neglects the influence of counterion association on the 1 mol m–2 seems to be a suitable choice. Consequently, point of zero charge. It assumes the same association af- {S} becomes the numerical value of surface concentra- finity for both counterions (anions and cations), which is tion of species S expressed in moles per meter square: practically never the case. One still obtains the c.i.p. but { } G –2 S = S / mol m . The activity coefficient of surface this point may be far from the p.z.c.12 For these reasons, species S (yS) is defined through the difference in chemi- potentiometric acid-base titration cannot be used for ex- cal potentials of real (mreal) and ideal (mid) states at the amination of the p.z.c. dependency on the electrolyte surface concentration. m real m id j The aim of this article is to examine the possibility RT ln yS = S – S = zSF (5) of applying mass titration for determination of the effect of electrolyte concentration on the point of zero charge where F denotes the Faraday constant, zS is the charge j in order to deduce the difference in surface association number of surface species S and is the electrostatic affinities of cations and anions. The effect of electrolyte potential affecting their state at the interface. Ideal state concentrations on the mass titration of colloidal disper- corresponds to the zero value of the electrostatic poten- tial j being equal to that in the bulk of the solution. Equa- sions of TiO2 particles was examined by numerical simu- lation and experimentally. From mass titration data, the tion (5) considers only electrostatic effects as responsi- surface charge densities and the point of zero charge at ble for the nonideal behavior. Such an approximation is different ionic strengths were obtained. The results were correct since electrostatics plays a major role. For exam- compared with »electrolyte titrations« (increase of the ple, if the potential affecting state of a positively charged ionic strength in concentrated dispersions)13 and with surface group (z = +1) changes from 180 mV to –180 mV, electrokinetic measurements. the activity coefficient at room temperature will be chang- ed due to the electrostatic effect by 6 orders of magni- 3 –3 Surface Complexation Model tudes, i.e., approximately from 10 to 10 . Equation (5) can be rewritten as: The basic process at a metal oxide surface in aqueous j environment is the surface charging due to interactions yS = exp(zSF / RT) (6)
Croat. Chem. Acta 79 (1) 95¿106 (2006) POINT OF ZERO CHARGE AND SURFACE CHARGE DENSITY OF TiO2 97
In the case of the two-step protonation (2-pK mecha- Surface Charge Densities, Zero Charge Conditions nism) described by reactions (1) and (2), the activity and Surface Potential coefficients of charged interfacial species ºMO– and º + Total surface concentration of active surface sites G , MOH2 involved in surface reactions are: tot within the 2-pK model, as described by reaction equa- º – j y( MO ) = exp(–F 0 / RT) (7) tions (1), (2), (11) and (12), is equal to: y(ºMOH +) = exp(–Fj / RT) (8) 2 0 G G º G º + G º – tot = ( MOH) + ( MOH2 )+ ( MO )+ j G º + × – G º – × + where 0 is the (inner) surface potential affecting their ( MOH2 A )+ ( MO C ) (17) state at the interface. The activity coefficient of neutral ºMOH species is equal to 1. Accordingly, thermody- The corresponding surface charge densities in s b s namic equilibrium constants for the first (1) and second 0-plane ( 0) and in -plane ( b) are: (2) steps of protonation are defined on the basis of equa- s G º + G º – tions (3–8) as: 0 = F( ( MOH2 )– ( MO )+ G(ºMOH + × A–)–G(ºMO– × C+)) (18) {}º 2 o j MOH K1 = exp(F 0 / RT) × (9) {}º×– s G º – × + G º + × – MO aH+ b = F( ( MO C )– ( MOH2 A )) (19)
{}ºMOH + while the net surface charge density s , i.e., the effective K o = exp(Fj / RT) × 2 (10) s 2 0 {}º× charge directly bound to the surface per surface area be- MOH aH+ ing equal in magnitude but opposite in sign with respect Effective surface charge is reduced by association of to the so-called surface charge density of diffuse layer anions A– and cations C+ from the bulk of solutions (re- s d, is equal to: actions A and C, respectively): s s s s G º + G º – º + – ®º + × – o s =– d = 0 + b = F ( ( MOH2 )– ( MO )) (20) MOH2 +A MOH2 A ; KA (11) º – + ®º – × + o MO +C MO C ; KC (12) At low ionic strength (Ic) and also at low surface po- tentials (pH close to pHpzc), the association of counterions By introducing the activity coefficients into expres- is negligible, so the surface charge density in 0-plane could sions for thermodynamic equilibrium constants of be approximated by: counterion surface association reactions (11) and (12), and taking into account that interfacial ion pairs behave s » F(G(ºMOH +)–G(ºMO–)); as oriented dipoles with charged ends being exposed to 0 2 j j I ® 0orj ® 0 (21) different potentials, b and 0, one obtains: c 0
æ F ()jjb - ö – + ç 0 ÷ The zero charge condition at the surface is expres- g (ºMO × C ) = exp ç ÷ (13) è RT ø sed by three quantities, i.e., by the point of zero charge s (p.z.c.) corresponding to 0 = 0, by the isoelectric point z æ --F ()jjb ö (i.e.p.) corresponding to the electrokinetic potential = g º + × – ç 0 ÷ s ( MOH2 A ) = exp ç ÷ (14) 0 (and s = 0), and by the point of zero potential (p.z.p.) è RT ø j corresponding to 0 = 0. In the absence of specific ad- Accordingly, the corresponding thermodynamic sorption of ions and in the case of negligible or symmet- equilibrium constants are: ric association of counterions, all three points coincide and correspond to the state in which all electrical proper- exp(FRT (jj-×º× ) / ) {}MO–+ C ties diminish (s = s =0,j = z = 0). This electroneu- K o = b 0 = 0 s 0 C - trality condition is characterized by the so called pristine exp(-׺×FRTj / ) {}MO a 0 C+ 18,19,20,23 point of zero charge (pHppzc or pHeln) being re- {}º×-+ lated to the protonation equilibrium constants by: j × MO C exp(F b/RT) - (15) {}º×MO a + C 1 o o pHeln = log(K1 K2 ) (22) 2 -- ׺×{}+ – exp(FRT (jjb 0 ) / ) MOH2 A K o = = Electroneutrality condition can be experimentally A + exp(FRTj0 / )׺{}MOH ×a - 2 A achieved at low electrolyte concentration, and in such a {}º×+ – case: MOH2 A exp(–Fjb/RT) × (16) + {}º×MOH a – ® 2 A pHeln =pHpzc =pHiep =pHpzp; Ic 0 (23)
Croat. Chem. Acta 79 (1) 95¿106 (2006) 98 T. PREO^ANIN AND N. KALLAY
Since affinities of binding the cations and anions are where m is the mass of the metal oxide powder. These not necessarily the same, the shift in pHpzc,pHiep and pHpzp contributions should be also taken into account when con- with respect to pHeln might be expected at higher elec- sidering the changes in the amounts of free hydronium trolyte concentrations.21 For example, in the case of pre- Dn(H+) and hydroxyl Dn(OH–) ions: o o ferential adsorption of anions (KA > KC ), p.z.c. is shift- D + Dx Dx Dx ed to higher pH values and i.e.p. is shifted to lower pH n(H )=na – 2 – 1 – n (28) values compared to pHeln so that: pHpzc >pHeln >pHiep. D – Dx It may be shown that the point of zero potential pHpzp lies n(OH )=nb – n (29) between pHiep and pHpzc values (pHpzc >pHpzp >pHiep), and is close to pHeln, which depends on the characteris- Note that extents of reactions may have positive but tics of the interface.22,23 also negative values.16 In the latter case, the reaction pro- In the case of the 2-pK mechanism, the surface po- ceeds in the opposite direction, with respect to the direc- tential at 0-plane is, according to Eqs. (1) and (2), equal tion denoted by the reaction equation. According to Eqs. to: (27) – (29):
+ RT ln10 RT æ {}ºMOH ö Dn(H+)–Dn(OH–)=n – n – Dx – Dx = j = ()lnpH-- pH ç 2 ÷ (24) a b 2 1 0 pzp ç - ÷ u·m– Dx – Dx (30) F 2F è {}ºMO ø 2 1 The first term in Eq. (24) is Nernstian while the sec- Changes in the amounts of surface groups are: ond one is responsible for the lowering of the magnitude j j D º – Dx Dx D º + of the slope of the 0(pH) function. Potential 0 can be n( MO )= 1 – c; n( MOH2 )= 24 Dx Dx also approximated by the modified Nernst equation as: 2 – a (31)
D º – × + Dx D º + × – Dx j RT ln10 a n( MO C )= c; n( MOH2 A )= a (32) 0 = (pHpzp – pH) × (25) F In the following derivation, the solution in the ab- a where coefficient describes deviation from the ideal sence of solid metal oxide will be considered as the ini- a behavior characterized by = 1. For real metal oxide tial state (pH ), while the equilibrated dispersion after surfaces, the value of coefficient a is below 1 and de- in addition of the solid will be the final state (pHg). There- 25,26,27 pends on the kind of metal oxide, but also on the fore, the mass concentration of the solid is zero in the 28 composition of the electrolyte solution. initial, and g in the final state. Equations (28)–(32) yield the relationship between the changes in the hydronium Potentiometric Mass Titration Method and hydroxyl ion concentrations in the bulk of the solu- Surface charge of an oxide in aqueous environment is the tion and the mass concentration of the dispersion: result of surface reactions (1), (2), (11) and (12). In the D + D – g g G º + course of mass titration, when a solid metal oxide powder c(H )– c(OH )=u – s ( ( MOH2 )– G º – G º + – G º – + is added to an aqueous electrolyte solution, the amounts ( MO )+ ( MOH2 A )– ( MO C )) (33) (and concentrations) of H+ and OH– ions in the bulk of the solution are changing due to surface protonations (1) where g is the mass concentration of the metal oxide of and (2), but also due to possible neutralization of H+ and the specific surface area s. According to Eq. (18), defin- OH– ions in the bulk of the solutions: ing surface charge density in the inner plane of the inner layer of the electrical interfacial layer, and Eq. (33), the + – ® Dx o o H (aq) + OH (aq) H2O(l); n; Kn =1/Kw (26) changes in bulk concentrations of H+ and OH– ions are:
o o sgs where Kn and Kw are thermodynamic equilibrium con- Dc(H+)–Dc(OH–)=ug – 0 (34) stants of neutralization and water dissociation, respecti- F vely. The change of the extent of the reaction is gener- Changes in the concentrations of H+ and OH– ions 16 ally defined through the change of the amount of any may be obtained from pH measurements according to: D substance B participating in the chemical reaction, nB, n Dx D n o and its stoichiometric number B as = nB/ B. c Dc(H+)= (10–pHg –10–pHin) (35) The solid phase may contain a certain amount of acid y (na) or base (nb) as impurities. The fractions of acid and o c o o base impurities (a and b, respectively) in solid metal ox- Dc(OH–)= (10pHg–pKw –10pHin–pKw ) (36) ide powder and their difference u will be defined as: y
n n + – a = a ; b = b ; u = a–b (27) where y denotes the activity coefficient of H and OH m m ions in the bulk of the solution given by the extended De-
Croat. Chem. Acta 79 (1) 95¿106 (2006) POINT OF ZERO CHARGE AND SURFACE CHARGE DENSITY OF TiO2 99 bye-Hückel equation (Davies equation), while co is the ionic strength and of the adsorption of counterions on standard value of concentration (co = 1 mol dm–3). the mass titration curve are examined in next sections by According to Eqs. (34) – (36) the surface charge den- numerical simulations, and also experimentally. sity in the 0-plane is equal to:
o Numerical Simulations s Fu F c –pHg –pH pHg–pK o 0 = - (10 –10 in –10 w + s sg y Effects of the ionic strength, counterion association and o 10pHin–pKw ) (37) choice of the initial pH on mass titration curves and on surface charge densities were examined by numerical s In order to evaluate the surface charge density 0 us- simulation of the mass titration. Calculations were based ing this equation, one should know the degree of con- on the Surface Complexation Model assuming the 2-pK tamination u. Degree of acid or base contamination may mechanism (Eqs. 1–25). Calculations were performed for be obtained from the potentiometric acid-base titration a pure sample (u = 0). Parameters used in calculation cor- of the concentrated dispersion.7 If the value of u is un- respond to a typical metal oxide aqueous dispersion: s known, one can still obtain relative values of 0 and lo- –2 e s =50gm ; relative permittivity of water r(H2O) = cate the zero on the basis of the known point of zero G 78.54; T = 298 K; total surface concentration tot = charge. Another, and much better, approach is purifica- –5 –2 2 o 7 1.5×10 mol m (= 9.9 sites/nm ); K1 = 9.09×10 tion of the sample, leading to u = 0. To cover both acidic o o 4 o (lgK1 = 7.96) K2 = 3.16×10 (lgK2 = 4.50); capacitance and basic pH ranges, at least two experiments should be –2 of the inner Helmholtz layer C1 =1.5Fm ; capacitance performed, one with low and the other with high initial ¥ of the outer Helmholtz layer C2 = . According to the pH. o o chosen values of K1 and K2 :pHeln = 6.23. Calculations Equation (37) could be rewritten as: –3 were performed for different ionic strengths: Ic / mol dm –4 –3 –2 –1 o o =10 ;10 ;10 ;10 . 10–pHg –10pHg–pKw = (10–pHin –10–pHin–pKw )+ The following SCM parameters were calculated as a æ ö º g y ç ss 0 - ÷ function of pH: surface concentrations of species MOH, u (38) º + º – º +× – º –× + co è F ø MOH2 , MO , MOH2 A , MO C , surface charge s s s densities at different planes ( 0, b and s) and surface It is obvious that the slope of the tangent of curve potentials (j , jb » j ). Also, the pH of the dispersion o 0 d (10–pHg –10pHg–pKw ) vs. mass concentration: as a function of the mass concentration was calculated. --pH pH pK o As shown in Figures 1–5, in the course of mass ti- d()10gg- 10 w y æ ss ö = ç 0 -u÷ (39) tration, the pH of the metal oxide dispersion changes o è ø dg c F gradually, approaching the constant value of pH¥. The slope of the pH(g) functions decreases in magnitude, in- is in relation with the surface charge density s and with 0 dicating that the absolute value of the surface charge the degree of contamination u. If addition of metal oxide density s approaches zero. According to Eqs. (38)–(40), in dispersion does not change the pH, the value of 0 o when a constant value of pH in mass titration of pure (10–pHg –10pHg–pKw ) also remains constant, and the de- metal oxide (u = 0) is achieved, the surface charge den- rivative (39) is equal to zero, so that: s sity 0 is equal to zero and pH¥ =pHpzc. In Figures 1–5, the maximum displayed mass concentration was 500 g ss0 -u = 0 (40) –3 F dm , which in most cases was not high enough to pro- duce a constant pH value of pH¥. For that reason, the pH In the case of a pure metal oxide (u = 0), the surface values of dispersions with the mass concentration of 500 s –3 charge density 0 is equal to zero and pH¥ is equal to the gdm (pH500) and infinite mass concentration (pH¥) point of zero charge, pH¥ =pHpzc. were denoted by numerical values. According to numeri- In some cases, addition of a metal oxide powder cal simulations, the mass concentration at which an ap- produces a very dense dispersion in which mixing and proximately constant pH¥ is reached depends on the measurements of pH are difficult so that pH cannot reach electrolyte concentration and equilibrium parameters of a constant value (pHg
Croat. Chem. Acta 79 (1) 95¿106 (2006) 100 T. PREO^ANIN AND N. KALLAY
Figure 1. Effect of ionic strength on the mass titration and on sur- Figure 2. Effect of ionic strength on the mass titration and on sur- face charge density as calculated for the model metal oxide in ab- face charge density as calculated for the model metal oxide with K o K o sence of association of counterions at the surface ( A = C =0). significant and symmetrical association of counterions at the surface I –3 –1 –2 –3 –4 K o K o I –3 –1 –2 –3 c / mol dm =10 (—); 10 (·–·); 10 (---); 10 (····). pH500 ( A = C = 100). c / mol dm =10 (—); 10 (·–·); 10 g –3 –4 g –3 is pH of dispersion with =500gdm . In the calculations 2-pK (---); 10 (····). pH500 is pH of dispersion with =500gdm .In mechanism of the Surface Complexation Model was assumed, the calculations 2-pK mechanism of the Surface Complexation Eqs. (9, 10, 15–22, 24, 25, 37). Model was assumed, Eqs. (9, 10, 15–22, 24, 25, 37) Result: pH¥ =pHpzc =pHeln = 6.23. Result: pH¥ =pHpzc =pHeln = 6.23.
o o centration: pH¥ =pHeln = 6.23. As expected, for a given presented in Figures 1 and 2 for negligible (KA = KC = s o o pH, the surface charge density 0 was found to be higher 0) and symmetric counterion association (KA = KC >0), at a higher ionic strength. The common intersection the value of pHpzc coincides with the state in which all point (c.i.p) coincides with pHcip =pHpzc =pHeln, which electrical properties disappear, i.e., coincides with the s s is due to the absence of counterions association. surface electroneutrality point. In such a case, 0 = s =0 j z Figure 2 represents the effect of ionic strength in the and 0 = = 0, so that pH¥ =pHpzc =pHiep =pHpzp = o o pH =pH , so the common intersection point may be case of symmetric counterion association (KA = KC = eln cip 100). It is clear that the pH(g) function approaches the used for determination of the point of zero charge. constant pH¥ value sooner (i.e., at lower mass concen- Figures 3 and 4 correspond to the unsymmetrical o ¹ o trations of the solid) compared to the system in the ab- counterion association (0 < KA KC > 0). Due to the sence of counterion association. For relatively high ionic high counterion association equilibrium constants, the –2 –3 strengths (above 10 mol dm ), pH500 was found to be pH¥ is reached earlier, practically at a mass concentra- –3 practically equal to pH¥ =pHeln = 6.23. The surface tion of 200 g dm . These results show that pH¥ =pHpzc, charge density was significantly increased due to the as- i.e., that the point of zero charge could be obtained by sociation of counterions. The common intersection point the mass titration method also in the case of high and un- is again at pH¥ =pHeln = 6.23. According to the results symmetrical counterion associations. As shown in Fig-
Croat. Chem. Acta 79 (1) 95¿106 (2006) POINT OF ZERO CHARGE AND SURFACE CHARGE DENSITY OF TiO2 101
Figure 3. Effect of ionic strength on the mass titration and on sur- Figure 4. Effect of ionic strength on the mass titration and on sur- face charge density as calculated for the model metal oxide with face charge density as calculated for the model metal oxide with K o K o preferential association of anions at the surface ( A = 10000; preferential association of cations at the surface ( A = 100; K o I –3 –1 –2 –3 K o I –3 –1 –2 –3 C = 100). c / mol dm =10 (—); 10 (·–·); 10 (---); C = 10000). c / mol dm =10 (—); 10 (·–·); 10 (---); –4 g –3 –4 g –3 10 (····). pH500 is pH of dispersion with =500gdm .Inthe 10 (····). pH500 is pH of dispersion with =500gdm .Inthe calculations 2-pK mechanism of the Surface Complexation Model calculations 2-pK mechanism of the Surface Complexation Model was assumed, Eqs. (9, 10, 15–22, 24, 25, 37). was assumed, Eqs. (9, 10, 15–22, 24, 25, 37). Result: pH¥ =pHpzc >pHeln = 6.23. Result: pH¥ =pHpzc
Croat. Chem. Acta 79 (1) 95¿106 (2006) 102 T. PREO^ANIN AND N. KALLAY
Figure 5. Effect of initial pH (pHin) on the mass titration as calcu- Figure 6. Mass titrations of TiO2 at different ionic strengths ad- –3 –4 –3 –2 lated for the model metal oxide with significant and symmetrical justed with NaCl. Ic / mol dm =10 (l); 10 (s); 10 (n); K o K o –1 association of counterions at the surface ( A = C = 100). Sur- 10 (u). Temperature: 25 °C. Surface charge densities were cal- I –4 –3 face charge density was also displayed. c =10 mol dm ,pHin culated from mass titration data by means of Eq. (37). =7(¡,·–·); 6 (u,––); 5 (s,---); 4 (o,····). In the calculations 2-pK mechanism of the Surface Complexation Model was assum- ed, Eqs. (9, 10, 15–22, 24, 25, 37). Result: pH¥ =pHpzc =pHeln = 6.23. grams per liter. After mixing, the dispersion pH was about 3.5 due to acidic impurities in the commercial sample. Dis- persion was treated with an ultrasonic probe and pH was Then, an analogical experiment should be conducted, adjusted by adding a small amount of NaOH to pH » 7. Af- but starting from basic pH of the solution. If the same ter adjusting the pH, the dispersion was close to the isoelec- pH¥ is obtained, it proves that pH¥ =pHpzc. This proce- tric point, the particles were aggregated and sedimentation dure can cover a broad pH range, as shown in Figure 5. occurred. Supernatant was separated from the sediment by The results of numerical simulations were supported by decantation. Conductivity and pH of the supernatant were experimental findings, presented in the next section. measured. To the sediment, which was in fact a very con- centrated dispersion (mud, slurry), redistilled water was ad- ded and the dispersion was again treated with ultrasound. If EXPERIMENTAL the pH was far from the isoelectric point, no sedimentation –2 Materials: TiO2 (P-25, Degussa); s =50gm , NaCl (p.a., occurred. In such a case, the pH was again adjusted with Fluka), HCl (0.1 mol dm–3, titrival, Fluka), NaOH (0.1 mol NaOH to pH » 7. The procedure was repeated until the su- dm–3, titrival, Fluka), standard buffers (Merck), redistilled pernatant conductivity reached the constant and low value and decarbonated water. of about 10 mScm–1. (Conductivity of redistilled water was m –1 Purification: The TiO2 sample was purified by extended about 3 Scm .) Use of glassware was avoided so that the washing as follows. The titania powder was dispersed in possible dissolution of silica would not contaminate the TiO2 water. Mass concentration of the dispersion was about 50 surface due to adsorption.
Croat. Chem. Acta 79 (1) 95¿106 (2006) POINT OF ZERO CHARGE AND SURFACE CHARGE DENSITY OF TiO2 103
Figure 8. Effect of NaCl concentration on pHpzc (dashed line) and l on pH values ( ) obtained by electrolyte (NaCl) titration of TiO2 dispersion (g =120gdm–3)at25°C.
Figure 7. Mass titrations of TiO2 for different initial pH: pHin = 9.8 (¡); 6.7 (r);4(l). Ionic strength was adjusted with NaCl: –4 –3 Ic =10 mol dm . Temperature: 25 °C. Surface charge densities were calculated from mass titration data by means of Eq. (37).
Figure 9. Effect of pH and ionic strength (controlled with NaCl) on z I –3 Experimental Procedures and Conditions: The pH was mea- electrokinetic -potential of TiO2 dispersion at 25 °C. c / mol dm sured by a combined glass-Ag/AgCl electrode (Metrohm =10–3 (u), 10–1 (¡). 6.0233.100). The electrode system (glass-Ag/AgCl electrode) was calibrated with three standard buffers. The systems were strengths. The maximum mass concentration was 120 g dm–3, kept under an argon atmosphere and thermostated at 25 °C. because the more concentrated dispersions are too dense, and The equilibrium was achieved by the application of ultra- it would be difficult to stir them and measure the pH. The sounds (ultrasonic probe) and by stirring the dispersion. results are in agreement with numerical simulations demon- strated in Figures 1–4. As shown in Figure 6, an approxi- Potentiometric Mass Titration mately constant pH was reached only for the dispersion at I –2 –3 Potentiometric mass titration was used to determine the ef- higher ionic strengths ( c >10 mol dm ). According to fect of ionic strength on the surface charge density and on Eq. (40), the point of zero charge for the ionic strengths of –2 –1 –3 » 10 and 10 mol dm was found to be pHpzc 6.3, or the point of zero charge of TiO2. Initial pH of water was ad- slightly higher. For lower ionic strengths, the pH value did justed by addition of HCl (pHin = 4) or NaOH (pHin = 9.8). –3 –4 –3 –2 not reach the plateau, so p.z.c. could not be evaluated. Sur- Constant ionic strength (Ic / mol dm =10 ;10 ;10 ; 10–1) was controlled by NaCl. Weighed amounts of titania face charge densities vs. pH were calculated from the mass powder were added to the aqueous solution of NaCl of pre- titration data by means of Eq. (37). As expected, the sur- face charge densities increased with the ionic strength. determined acidity (pHin) in subsequent portions. After each addition and equilibration (5 minutes under ultrasound According to numerical simulations (Figure 5), the and 10 minutes of stirring), the pH was measured. Figure 6 point of zero charge at low ionic strengths could be deter- represents the mass titrations of TiO2 at different ionic mined by choosing the initial pH of the dispersion close to
Croat. Chem. Acta 79 (1) 95¿106 (2006) 104 T. PREO^ANIN AND N. KALLAY pHpzc. In order to examine this finding experimentally, face charge densities, and point of zero charge of metal mass titrations of titania at low ionic strength (10–4 mol oxide colloidal dispersion was examined by numerical si- dm–3) were performed using three different initial pH val- mulation, and also experimentally using titania aqueous ues. The results (Figure 7) are in agreement with numerical dispersion. Experimental results agreed with the numeri- simulations (Figure 5). Initial pH values were adjusted to cal simulation, thus supporting the idea of applying the be in the acidic region far from and close to the point of mass titration method for examination of equilibrium at zero charge (pH = 4 and pH = 6.7), but also in the basic in in solid/liquid interface. Mass titration was found to be a region (pHin = 9.8). At a sodium chloride concentration of –4 –3 very simple and useful tool for determination of the point 10 mol dm , the point of zero charge of TiO2 was found s of zero charge. To avoid limitations of the method, one to be at pHpzc = 6.1. Surface charge densities 0 were also evaluated. By performing two runs, the first with the initial should first purify the sample so that the pH of highly acidic solution and the second with the initial basic solu- concentrated dispersion corresponds to the point of zero tion, the broad pH region was covered. charge. The second problem is that in some cases one is not able to prepare enough concentrated dispersion. This Electrolyte Titration problem can be avoided by adjusting the initial pH (of the liquid medium) close to the point of zero charge, prior Initial TiO dispersion was prepared by addition of a rela- 2 to addition of the solid phase. Such measurements can tively large amount of purified TiO powder (g =120gdm–3) 2 produce accurate results, which are important because to the aqueous HCl solution (c =10–4 mol dm–3). Weighed portions of solid sodium chloride were then added to the they indicate the possible preferential association of cat- initial dispersion so that ionic strength increased from 10–4 ions with respect to anions and vice versa. The results to 1 mol dm–3. The pH values of the dispersion were mea- with titania dispersed in an aqueous sodium chloride so- sured after each addition of NaCl and after equilibration (5 lution, indicating the unsymmetrical association of the minutes under ultrasound and 10 minutes of stirring). counterions, are also supported by electrokinetic results; Increase in the electrolyte concentration caused an in- pHiep and pHpzc were shifted in opposite directions with crease in the dispersion pH to pH = 6.3 (Figure 8). Dashed respect to their values obtained at a low ionic strength. It line demonstrates the dependency of pHpzc on the ionic was found that chloride ions exhibit higher affinity to- strength determined from the mass titrations. At relatively wards association with oppositely charged surface groups high ionic strengths of 10–2 mol dm–3 and 10–1 mol dm–3, than sodium ions. the point of zero charge is pH » 6.3 (Figure 6). At the pzc Accuracy of the point of zero charge determination low ionic strength of 10–4 mol dm–3, the point of zero charge by the mass titration method enables a comparison of pH is lower, i.e., pH = 6.1 (Figure 7). Discrepancy of the data iep pzc and pH values. In the case of unsymmetrical associa- obtained from the electrolyte titration with p.z.c. data from pzc mass titration is due to the insufficient mass content of tita- tion, these two values are different and their shifts caused nia in the electrolyte titration experiments. At higher mass by electrolyte addition are in opposite directions. If these concentrations this discrepancy might disappear, which is two points coincide at high ionic strength, or if they do not hard to prove experimentally because the dispersion is too depend on the electrolyte concentration, this means that the dense at high mass concentrations. Increase of pHpzc with association equilibrium constants of the anions and cat- sodium chloride concentration indicates the preferential as- ions with oppositely charged surface groups are equal. In o o sociation of anions with respect to cations: KCl– > KNa+ . such a case, all zero charge conditions coincide, so that pHeln =pHpzc =pHiep =pHpzp. This finding is useful for lo- Electrokinetic Measurements cating the point of zero potential, necessary for evaluation of the surface potentials from the electrode potentials of The electrokinetic z-potential was measured using an the corresponding single crystalline electrodes.23,28,35 It is ELS-800 Otsuka instrument. Titania dispersion was pre- suitable for a given metal oxide to find an electrolyte with pared by adding a small amount of powder to the aqueous solution (g » 5mgdm–3). Ionic strength and pH were con- the same surface association affinities of cations and an- trolled by addition of NaOH, HCl and NaCl. The isoelectric ions. Such an electrolyte should possess equal pHpzc –3 –3 (mass titration) and pHiep (electrokinetics), and also inde- point of TiO2 particles at Ic =10 mol dm is at pHiep = –1 –3 pendence of these two pH values of the electrolyte concen- 6.2, while at Ic =10 mol dm pHiep = 5.8. The shift of tration. It is important to note that the common intersection the pHiep value to the acidic region indicates preferential o o point in acid-base titrations of a dispersion at different association of anions with respect to cations KCl– > KNa+ , as deduced from electrolyte and mass titrations. This find- ionic strengths cannot be used as a tool for location of the ing is in agreement with the literature.30–34 point of zero charge, since this approach implies the inde- pendency of pHpzc on the electrolyte concentration, i.e., it assumes the same association affinities of cations and an- DISCUSSION AND CONCLUSION ions at the surface, which is actually never the case. The effect of ionic strength, in a broad range (10–4 mol The mass titration method can be also used for de- dm–3 …10–1 mol dm–3), on the mass titration curve, sur- termination of the surface charge density in 0-plane. To
Croat. Chem. Acta 79 (1) 95¿106 (2006) POINT OF ZERO CHARGE AND SURFACE CHARGE DENSITY OF TiO2 105 cover the entire pH range of interest, one should perform 15. W. Rudzinski, R. Charmas, W. Piasecki, J. M. Cases, M. several experimental runs at different initial pH values Francois, F. Villieras, and L. J. Michot, Colloids Surf. A 137 of the solutions to which the samples of solid powder (1998) 57–74. are added. Four runs might be enough, two in the acidic 16. I. Mills, T. Cvitaš, K. Homann, N. Kallay, and K. Kuchitsu, Quantities, Units and Symbols in Physical Chemistry,2nd and two in the basic region. This method does not re- ed., Blackwell Scientific Publications, Oxford, 1998. quire titration in the absence of the solid phase (blank ti- 17. N. Kallay, T. Preo~anin, and S. @alac, Langmuir 20 (2004) tration) and the problem of locating the point of zero 2986–2988. charge disappears. It is especially useful for determina- 18. M. A. Pyman, J. W. Bowden, and A. M. Posner, Aust. J. tion of the surface charge density in the pH range close Soil Res. 17 (1979) 191–195. to the point of zero charge. 19. J. Lyklema, J. Colloid Interface Sci. 99 (1984) 109–117. 20. J. Lützenkirchen and P. Magnico, Colloids Surf. A 137 (1998) 345–354. 21. J. Sonnefeld, Colloids Surf. A 190 (2001) 179–183. REFERENCES 22. D. Kova~evi}, A. ^op, and N. Kallay, Evaluation and Us- age of Electrostatic Potentials in the Interfacial Layer, in: 1. J. S. Noh and J. A. Schwarz, J. Colloid Interface Sci. 130 A. Delgado (Ed.), Interfacial Electrokinetics and Electro- (1989) 157–164. phoresis, Marcel Dekker Inc., New York, 2002, pp. 99– 2. K. Bourikas, J. Vakros, C. Kordulis, and A. Lycourghiotis, 121. J. Phys. Chem. B 107 (2004) 9441–9451. 23. N. Kallay, A. ^op, T. Preo~anin, and D. Kova~evi}, Annales 3. M. Harju, E. Levänen, and T. Mäntilä, J. Colloid Interface Universitatis Mariae Curie-Sklodowska, Sectio AA Chemia Sci. 276 (2004) 346–353. 60 (2005) 47–64. 4. S. Wang, Z. H. Zhu, A. Coomes, F. Haghseresht, and G. Q. 24. T. W. Healy, D. E. Yates, L. R. White, and D. Chan, J. Elec- Lu, J. Colloid Interface Sci. 284 (2005) 440–446. troanal. Chem. 80 (1977) 57–66. 5. J. P. Reymond and F. Kolenda, Powder Technology 103 25. N. Kallay, R. Sprycha, M. Tomi}, S. @alac, and @. Torbi}, (1999) 30–36. Croat. Chem. Acta 63 (1990) 467–487. 6. T. Preo~anin, Z. Brzovi}, D. Dodlek, and N. Kallay, Progr. 26. M. J. Avena, O. R. Camara, and C. P. De Pauli, Colloids Colloid Polym. Sci. 112 (1999) 76–79. Surf. A 69 (1993) 217–228. 27. N. Kallay, A. ^op, E. Chibowski, and L. Holysz, J. Colloid 7. S. @alac and N. Kallay, J. Colloid Interface Sci. 149 (1992) Interface Sci. 259 (2003) 89–96. 233–240. 28. N. Kallay, Z. Dojnovi}, and A. ^op, J. Colloid Interface 8. N. Kallay and S. @alac, Croat. Chem. Acta 67 (1994) 467– Sci. 286 (2005) 610–614. 479. 29. P. Hesleitner, D. Babi}, N. Kallay, and E. Matijevi}, Lang- 9. T. Preo~anin and N. Kallay, Croat. Chem. Acta 71 (1998) muir 3 (1987) 815–820. 1117–1125. 30. W. Janusz, Polish J. Chem. 68 (1994) 1871–1880. 10. J. Lützenkirchen, J. F. Boily, L. Lovgren, and S. Sjoberg, 31. R. O. James and G. A. Parks, Characterization of Aqueous Geochim. Cosmochim. Acta 66 (2002) 3389–3396. Colloids by their Electrical Double-Layer and Intrinsic 11. J. Lützenkirchen, J. Colloid Interface Sci. 290 (2005) 489– Surface Chemical Properties, in: E. Matijevi} (Ed.), Surface 497. and Colloid Science, Vol. 12, Plenum Press, New York, 12. A. Breeuwsma and J. Lyklema, J. Colloid Interface Sci. 43 London, 1982, pp. 119–216. (1973) 437–448. 32. Y. G. Berubé and P. L. de Bruyn, J. Colloid Interface Sci. 28 13. P. H. Tewari and A. W. McLean, J. Colloid Interface Sci. 40 (1968) 92–105. (1972) 531–539. 33. R. Sprycha, J. Colloid Interface Sci. 102 (1984) 173–185. 14. D. E. Yates, S. Levine, and T. W. Healy, J. Chem. Soc., Fara- 34. W. Janusz, J. Radioanal. Nucl. Chem. 134 (1989) 193–198. day Trans. I 70 (1974) 1807–1818. 35. T. Preo~anin, W. Janusz, and N. Kallay, unpublished paper.
SA@ETAK
To~ka nul-naboja i povr{inska gusto}a naboja TiO2 u vodenim otopinama elektrolita dobivene potenciometrijskom masenom titracijom
Tajana Preo~anin i Nikola Kallay
Masena titracija pokazala se pogodnom metodom za odre|ivanje to~ke nul-naboja (p.z.c., point of zero s charge) i povr{inske gusto}e naboja ( 0) koloidnih ~estica kovinskih oksida. Postupak se sastoji od postupnog dodavanja kovinskog oksida u otopinu elektrolita i mjerenju pH suspenzije. Svakim dodatkom, pH suspenzije
Croat. Chem. Acta 79 (1) 95¿106 (2006) 106 T. PREO^ANIN AND N. KALLAY
se mijenja dok ne dosegne stalnu vrijednost, pH¥, koja je u slu~aju ~istog kovinskog oksida jednaka to~ki nul- naboja, pHpzc. Asocijacijom protuiona pomi~e se pHpzc u kiselo podru~je (ja~a asocijacija kationa) ili u lu`nato podru~je (ja~a asocijacija aniona). Rezultati masene titracije pru`aju informacije o razlici u afinitetu povr{inske asocijacije protuiona (aniona i kationa), va`nom parametru ravnote`e unutar elektri~nog me|upovr{inskog slo- ja. Odre|ivanje to~ke nul-naboja klasi~nom potenciometrijskom kiselinsko-baznom titracijom suspenzije teme- s ljeno je na zajedni~kom sjeci{tu 0(pH) krivulja (c.i.p., common intersection point) za razli~ite ionske jakosti. Pokazano je da se c.i.p. mo`e koristiti za odre|ivanje to~ke nul-naboja samo u slu~aju kada je asocijacija anio- na i kationa sa suprotno nabijenim povr{inskim vrstama jednaka. Analizirani su eksperimentalni rezultati masene titracije suspenzije koloidnih ~estica titanijevog oksida u vodenoj otopini natrijeva klorida, te rezultati nume- ri~kih simulacija temeljenih na 2-pK mehanizmu i modelu povr{inskog kompleksiranja. Pove}anje koncentraci- je natrijeva klorida pomi~e to~ku nul-naboja u lu`nato podru~je, dok se izoelektri~na to~ka pomi~e u kiselo podru~je, {to ukazuje na ja~u asocijaciju kloridnih iona sa suprotno nabijenim povr{inskim mjestima u odnosu na natrijeve ione.
Croat. Chem. Acta 79 (1) 95¿106 (2006)