Journal of Colloid and Interface Science 426 (2014) 209–212
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
The pH dependent surface charging and points of zero charge. VI. Update Marek Kosmulski Lublin University of Technology, Nadbystrzycka 38, PL-20618 Lublin, Poland
a r t i c l e
i n f o
Article history: Received 17 January 2014 Accepted 25 February 2014 Available online 12 March 2014 Keywords: Point of zero charge Isoelectric point Surface charge density Zeta potential Electrokinetic potential
a b s t r a c t The pristine points of zero charge (PZC) and isoelectric points (IEP) of metal oxides from the recent literature are summarized. This study is an update of the previous compilation (Kosmulski, 2009) [5] and of its previous updates (Kosmulski, 2009, 2011) [6,7]. Only the IEP of materials other than metal oxides are reported, and the PZC of such materials obtained by potentiometric titration and related methods are ignored. IEP of (nominally) CoO, Fe(OH)2, Gd2O3, Ni2O3, and Sb2O3 have been reported in the recent literature. Those materials have not been studied before. Ó 2014 Elsevier Inc. All rights reserved.
Introduction The pH-dependent surface-charging of solid particles in aqueous solutions of 1–1 electrolytes has been extensively studied. The points of zero charge (PZC) and isoelectric points (IEP) of metal oxides in 0.0001–0.1 M solutions of alkali halides, nitrates(V), or chlorates(VII) are termed pristine PZC/IEP, and they were compiled in the classical paper by Parks [1]. The pristine PZC are independent of the nature of the salt, its concentration and other experimental conditions, thus they characterize the surface-charging behavior over a broad range of experimental conditions. Ions other than alkali metal cations or halide, nitrate(V), or chlorate(VII) anions often adsorb specifically, and induce a shift of PZC and IEP away from the pristine value. The PZC/IEP observed in the presence of specifically adsorbing ions are of limited significance, because they depend on the nature of the salt, its concentration, and other experimental conditions, and they are ignored in the present review. For example, Mehdilo et al. [2] used H2SO4 to adjust the pH of dispersions of ilmenite. Their IEPs are not reported in this review, because sulfate anions are known to adsorb specifically on iron compounds, and to induce a shift in their IEP to low pH values. Several recent publications compile pristine PZC of a specific chemical compound(s) or of certain class of chemical compounds. For example Chorover compiled PZC of soil constituents [3]. In contrast, relatively few publications report PZC of broad ranges of materials. Extensive compilations of pristine PZC of metal oxides and other materials were published by Kosmulski [4,5]. The E-mail address: [email protected] http://dx.doi.org/10.1016/j.jcis.2014.02.036 0021-9797/Ó 2014 Elsevier Inc. All rights reserved.
classical Parks’ review is still more frequently cited as a reference for ‘‘recommended’’ PZC of metal (hydr)oxides than up-to-date reviews as illustrated in Fig. 1. Several recent publications, which cite [1,4,5] are discussed in this compilation. The recent review [5] was updated twice [6,7] and the recent results (2011–2013) and a few older results (overlooked in [5–7]) are compiled in the Supplementary material in Table 1 (not available in the printed copy of the journal). The significance of the pristine PZC/IEP and their applications were discussed in detail in Ref. [1–7], in many other compilations of PZC/IEP cited therein, and in handbooks of surface and colloid chemistry. The PZC reported in review papers are especially important for the scientists who do not determine the PZC of their specimens themselves, but who rely on the PZC of similar specimens taken from the literature. In principle the PZC/IEP ‘‘recommended’’ in various reviews are consistent, and the choice of this or another compilation is not crucial. Yet erroneous pristine PZC allegedly based on the review papers are commonplace in the recent literature. For example Sojka et al. [8] used an erroneous value of PZC of niobia (2.8); Usui et al. [9] used erroneous values of PZC of hematite (4.2–6.9) and of goethite (5.9–6.7); and Liu et al. [10] used an erroneous value of PZC of hematite (6.7), taken from old compilations. The PZC reported in review papers are also useful for the scientists who determined the PZC of their specimens experimentally. The PZC of similar specimens taken from the literature are used to verify the purity of the samples and the correctness of the procedures. Comparison of own results with the literature (original publications or compilations) is commonplace in papers reporting
210
M. Kosmulski / Journal of Colloid and Interface Science 426 (2014) 209–212
geographical locations (mine, region, country). The formula and concentration (concentration range) of the electrolyte(s), temperature, method, and instrument used in the surface charging measurements are reported in separate columns of Supplementary Table 1 (not available in the printed copy of the journal). Empty boxes denote that the information was not available or not applicable.
Methods Experimental
Fig. 1. Citations of selected compilations of PZC/IEP.
PZC/IEP measurements. However, several recent studies reported rather unusual pristine IEP of common metal oxides, e.g., 6 for alumina [11] and 4.2 for Fe2O3 [12], without a reference to the existing literature on similar specimens. This is not clear whether the authors realized that their IEP were very different from the IEP of similar specimens reported by the others. The present review reports selected results from the literature. The pristine PZC and IEP are summarized in Supplementary Table 1 (not available in the printed copy of the journal). Erroneous or less credible results were deliberately neglected. Also the PZC/IEP reported in the original papers were verified when the original data (f potential or surface charge density r0 as a function of pH) were available. Several publications report f potential or r0 measured in multicomponent dispersions (dispersions containing various types of colloidal particles). The r0 in multi-component dispersions is additive, that is, the information about the composition of dispersion and the charging behavior of individual components makes it possible to predict the charging behavior of multi-component dispersions. In contrast there is no simple relationship between the apparent f potential observed in a mixture of various minerals and the composition of that mixture. Erroneous allegations in this respect are published now and then even in top-quality scientific journals, e.g., Eq. (2) in [13]. Kosmulski [14] studied the f potentials in dispersions containing two types of colloidal particles. The background subtraction procedure made it possible to calculate the f potential of individual types of colloidal particles in two-component dispersions. However, such a background subtraction procedure is only valid in the absence of heterocoagulation. The present review is focused on the PZC/IEP obtained in one-component dispersions, and the PZC/IEP allegedly obtained in multi-component dispersions (e.g., soils and rocks) were deliberately ignored.
Structure of Table 1 The structure of Supplementary Table 1 (not available in the printed copy of the journal) follows the same principles as Table 1 in Refs. [6,7]. The materials are organized into 13 categories according to their chemical composition. Within each category the materials are sorted by their chemical formula. Multiple specimens corresponding to the same (idealized) formula are sorted into the following categories: commercial, synthetic, and natural materials. The commercial specimens are characterized by their manufacturer and/or trade name. The synthetic materials are characterized by the recipe. The natural materials are characterized by
The terminology of experimental methods used in the literature is not consistent. Very often the same method is described by various names or the same name is used for different methods. The following terminology, adapted from previous compilations [4–7] is used in Supplementary Table 1 (not available in the printed copy of the journal): cip (common intersection point of potentiometric titration curves obtained at three or more ionic strengths) intersection (intersection point of potentiometric titration curves obtained at two ionic strengths) iep (isoelectric point obtained by electrokinetic or electroacoustic measurements) salt titration (salt addition) : addition of inert electrolyte (powder or concentrated solution) induces a shift in the pH of dispersion toward the pristine PZC. The pH value, at which salt addition does not induce any shift in the pH is equivalent to CIP. pH (natural pH of dispersions, e.g., obtained by mass titration). Such results were deliberately ignored in most materials. Silica (for which cip is not observed) is an exception. Moreover in a few papers reporting PZC of less common materials, the description of the method was not precise enough, and the ‘‘method’’ in Supplementary Table 1 (not available in the printed copy of the journal) is followed by ‘‘?’’. On top of measurements carried out by standard methods, a few publications present less common solutions. Corbett et al. [15] proposed a technique to measure the f potential of macroscopic specimens. They observed the velocity of tracer particles between two flat parallel electrodes as the function of the distance from the studied surface (in the range 0.1–1 mm), which was perpendicular to those electrodes. The new method has an advantage of using small specimens of solid and small volumes of liquid as compared to standard electroosmotic or streaming potential measurements, and it inherits most disadvantages of those methods (difficulties in the pH measurement, sensitivity to traces of impurities). Jordan et al. [16] obtained a hematite layer by atomic layer deposition and report a PZC of that layer at pH 5.5, derived from second harmonic generation (SHG) studies. Their PZC is substantially lower than the PZC/IEP usually reported for hematite powders. Sung et al. [17] studied an a-alumina single crystal, and report a PZC of (1–102) plane, derived from sum frequency vibrational spectroscopy (3230 cm 1 band) at pH 6.7. Their PZC is substantially lower than the PZC/IEP usually reported for alumina powders. Similar discrepancies (e.g., SHG vs. electrophoresis) are well-documented in the older literature. SHG is not considered a standard method in the present study, but a few PZC obtained by SHG are reported in Supplementary Table 1 (not available in the printed copy of the journal). There is a substantial difference between the present approach and that in Refs. [4–7], namely the PZC obtained by means of titration at one ionic strength (which in fact are natural pH values of dispersions) for materials other than metal oxides were ignored
M. Kosmulski / Journal of Colloid and Interface Science 426 (2014) 209–212
in the present update. It was shown recently [18] that in materials other than metal oxides, the IEP obtained in different laboratories and with various specimens of nominally the same chemical compound are relatively consistent, and the PZC obtained by titration are not. The charging curves of materials other than metal oxides usually do not show a sharp CIP, and the apparent PZC is sensitive to the solid-to-liquid ratio, ionic strength, and other experimental conditions. Accordingly the PZC obtained by titration at one ionic strength, mass titration, and equivalent methods (natural pH of dispersions) are ignored in the present update, also for metal oxides. Supplementary Table 1 (not available in the printed copy of the journal) compares the newly reported data with the median values of PZC/IEP from [5–7]. With 1 or 2 references in [5–7], the PZC/IEP from the older literature are explicitly reported. The number of entries (the second last column of Supplementary Table 1 (not available in the printed copy of the journal)) recalled in the present review is often lower than the numbers of entries in previews updates [6,7]. This is because numerous PZC from [5–7] (natural pH of dispersions) was not used as references in the present study. Section number in [5] is also indicated in Supplementary Table 1 (not available in the printed copy of the journal). In a few materials, ions other than H+ determine the sign of the f potential and r0. The PZC/IEP of those materials are reported in terms of the concentrations of the potential determining ions (e.g., pAg for silver halides) rather than pH. A recent study of Kallay et al. [19] revealed that the f potential of silver halides is not pHindependent (as it was originally believed). A shift in the pH from 6 to 3 induced a shift in the IEP of silver halides to higher pAg by about 1 pAg unit. Presentation of data Several publications cited in Supplementary Table 1 (not available in the printed copy of the journal) report only the value of PZC/IEP, but they do not report the data points upon which those values were based. Those PZC/IEP are marked by an asterisk in Supplementary Table 1 (not available in the printed copy of the journal). Most publications cited in Supplementary Table 1 (not available in the printed copy of the journal) report the data points (f potential and/or r0 as a function of pH) and the PZC/IEP based upon those data points. In most studies the PZC/IEP were determined correctly from the data points. A few examples of controversial interpretation of data points were found. For example Missana et al. [20] report electrokinetic curves of titania for two ionic strengths. For 1.25 10 4 M NaClO4 they report data points in the vicinity of the IEP on its both sides, and an IEP at pH 6.5 which is close to the results found in other publications. However, for 10 2 M NaClO4, no data points in the pH range 4–7 were available, and an arbitrary line was drawn suggesting an IEP at pH 4.5. The apparent IEP at pH 4.5 were not reported in this review, because it is not supported by the data points. An interesting method of presentation of electrokinetic data was chosen by Ottofuelling et al. [21]. With pH as the abscissa and ionic strength as the ordinate (log scale) the iso-f potential lines were plotted every 10 mV, and the ranges (every 10 mV) were marked by different colors. Their IEP are not shown in Supplementary Table 1 (not available in the printed copy of the journal) because they used NaHCO3 to adjust the pH. The presence of carbonate anions is likely to induce a shift in the IEP of metal oxides. A non-standard method of presentation of electrokinetic data was chosen by Colombo et al. [22]. They studied various samples of hematite. With particle diameter as the abscissa and pH or zeta potential as ordinate they plotted two curves: natural pH(d) and f(d), in one graph. Taking each couple of data points corresponding to certain d, a f(pH) graph can be constructed from their data. A random mixture of positive and negative f potentials was obtained
211
at pH about 6, suggesting an IEP of hematite at that pH. This low IEP is not surprising since most studied specimens were obtained in the presence of specifically adsorbing anions (e.g., oxalate). Discussion of Table 1 Choice of specimens Many studies of PZC//IEP and many entries in Supplementary Table 1 (not available in the printed copy of the journal) refer to chemical compounds whose PZC/IEP are already well-documented. Usually the values reported in older publications for similar specimens were confirmed in the recent literature. There are more PZC/IEP reported for metal oxides than for all other chemical compounds together, and more PZC/IEP for aluminum, iron, and titanium oxides than for all other metal oxides together. Several commercial products that were very popular in the past (e.g. P-25 titania from Evonik/Degussa) received also much attention in the recent literature. A few PZC/IEP of less-well-documented materials were published in the recent literature. The IEP of CoO [45], Fe(OH)2 [79,80], Gd2O3 [45], Ni2O3 [45], and Sb2O3 [144] were recently published for the first time. The surface-charging measurements with those materials are challenging, because of their chemical properties. For example Fe(OH)2 is unstable against oxidation to Fe(III), and in spite of all efforts reported in [79,80], some degree of surface oxidation might have occurred. The surface oxidation might have affected the chemical composition of the surface (iron III and II rather than pure iron II), and thus the measured IEP. Gd2O3 absorbs CO2 from the atmosphere and turns into carbonate, which probably has lower IEP than oxide. Similar CO2 absorption and its effect on the IEP is well-documented for other lanthanides and for the 3-group metals’ sesquioxides. Apparently a commercial reagent was used in [45], and no efforts to remove CO2 from the material are reported. The IEP of Gd2O3 reported in [45] is lower than IEP of sesquioxides reported in the literature, probably due to CO2absorption. The point of zero charge of CaO [66,67] was recently reported, but in fact those experimental results are representative for hydroxide or carbonate rather than for oxide, which is very reactive. Most studies were carried out with powders, and the surfacecharging studies with single crystals are rare. IEP of ZnO single crystal [291] was lower than IEP reported for powders. Discrepancies between IEP of single crystals on the one hand and of powders on the other have been observed and discussed for alumina and for other metal oxides, but no generally accepted explanation of those discrepancies is available. Most electrokinetic studies reported in the recent literature were performed for one type of 1–1 electrolyte and for one ionic strength (cf. Supplementary Table 1 (not available in the printed copy of the journal)), and sometimes the nature and concentration of the electrolyte were not reported at all. Such studies are based on the assumption that certain electrolyte is inert, and no specific adsorption occurs. With metal oxides the inert character of alkali halides and nitrates is well-documented, and generally accepted, but with most materials other than metal oxides, the inert character of those electrolytes has yet to be proved. Electrokinetic curves of mixed Ti–Si oxide in two different electrolytes, at two different concentrations each [340] confirmed the inert character of those electrolytes, namely the IEP were independent of the nature of electrolyte and ionic strength. This is one of a few examples of systematic studies in this direction found in the recent publications. Temperature, pressure, and solvent effects on the PZC Most results reported in Supplementary Table 1 (not available in the printed copy of the journal) refer to room temperature and
212
M. Kosmulski / Journal of Colloid and Interface Science 426 (2014) 209–212
atmospheric pressure, and in a few papers, systematic studies of the temperature effects can be found. Temperature effect on the PZC and IEP of NiO [141], CIP of zirconia [295], CIP of alumina [58], and IEP of anatase [244] is reported in the recent literature. The f potentials of silica at various temperatures and pressures [173] and the f potentials of silica at 100 and 150 °C [167] have been reported. The recent studies confirm the previously found trends, namely the PZC shifts to low pH when the temperature increases. Temperature effect on the IEP of magnetite was studied in [399]. The IEP at 25 °C from that study are not reported in Supplementary Table 1 (not available in the printed copy of the journal), because it was obtained by arbitrary interpolation (no data points near the IEP). Lvov et al. [400] report the pH effect on the electrophoretic mobility of commercial silica at 20 bar. The mobility at 25 °C was negative and rather insensitive to pH over the range 4–8. The measurements were carried out at pressures up to 40 bar and temperatures up to 150 °C. Studies at such high temperatures and pressures cannot be carried out by means of commercial zetameters, and special equipment had to be designed. Very few studies of surface charging in nonaqueous and mixed solvents were published over the recent 3 years, and they were ignored in Supplementary Table 1 (not available in the printed copy of the journal). Turci et al. [401] report IEP of 3 commercial titanias in 10 % (v/v) aqueous DMSO at pH 5–5.5. At low ionic strengths, admixtures of non-aqueous solvents have been shown to have an insignificant effect on the IEP of titania. Therefore the above IEP are probably representative also for aqueous dispersions.
Correlations The surface-charging behavior is correlated with several other physical properties. These correlations are indicated in Supplementary Table 1 (not available in the printed copy of the journal), in footnotes. The correlation between the IEP on the one hand and the minimum in colloid stability and the maximum in the particle size is the most well-known one, and numerous examples confirming such a correlation are presented in Supplementary Table 1 (not available in the printed copy of the journal). Most modern zetameters offer a possibility of a measurement of f potential and particle size using the same sample of dispersion. However, there are a few exceptions, e.g., the particle size of P-25 titania reported in [238] was rather insensitive to pH, also in the vicinity of the IEP. The role of the f potential as a parameter defining the colloid stability is sometimes overrated. For example Ref. [132] presents a table, in which intervals of absolute values of f potential are assigned to certain degrees of stability (instable dispersion, good stability, excellent stability, etc.). Such a precise assignment is incorrect, because the stability (at a constant f) depends on the ionic strength and Hamaker constant. In silica unlike in metal oxides, the stability minimum is not correlated with the IEP. This exception has been well-known for decades, and the recent studies, e.g., [189] confirm the exceptional behavior of silica. The IEP are correlated with a maximum in the yield stress and in the viscosity of concentrated dispersions. This well-known correlation is confirmed by a few examples reported in Supplementary Table 1 (not available in the printed copy of the journal).
Several studies report series of surface-charging data obtained for particles of various sizes having (nominally) the same chemical composition. For example effects of particle size on CIP of ferrihydrite [129] and on IEP of titania [219] were claimed. A graph PZC/ IEP (particle size) of hematite is presented in [100] (supporting information), but no clear correlation was found. Allegations about the particle size effect on PZC/IEP can also be found in the older literature, but the present author is skeptical about them. Namely, it is practically impossible to prepare dispersions, which vary only in particle size, and to eliminate the other factors (e.g., minor impurities), which are responsible for the scatter in the PZC/IEP of allegedly identical specimens (e.g., P-25 from Degussa) reported in different publications. Most likely the apparent effect of particle size on PZC/IEP is actually due to factors, which were beyond the control in course of experiments. The effect of crystallographic structure (rutile vs. anatase) on the IEP was studied [270], and the following acidity sequence: brookite > anatase > rutile was claimed. Such results are of limited significance, because the effect of surface chemistry on the IEP was neglected. For example one of the specimens was obtained in the presence of MgSO4. Both Mg(II) cations and sulfate anions adsorb specifically on the surfaces of metal oxides and shift their IEP. An increase in the IEP with the solid-to liquid ratio (50– 500 ppm) was reported for alumina and titania [53]. An increase in the IEP with the solid-to liquid ratio (1–100 ppm) was reported for four different powders [86]. These results were affected by specific adsorption of anions, namely, two studied specimens were iron (hydr)oxides obtained in the presence of sulfates. Another specimen was a commercial powder, which probably contained surface-active agent(s). Generally the IEP in the absence of specific adsorption do not depend on the solid-to liquid ratio, and experiments carried out at various solid-to-liquid ratios can be used to confirm the purity of specimens, and correctness of procedures. Fisher et al. [402,403] report the pH effect on the contact angles of plasma-treated silica substrates, and interpret the maxima in the contact angle(pH) curves as the isoelectric points. The correlation between the IEP and the maximum in the contact angle is well-know, but the contact-angle-measurement is not considered as a standalone method of IEP measurement in the present study, and such results are not reported in Supplementary Table 1 (not available in the printed copy of the journal). Unusual results Negative f potentials of MgCO3 at low pH and positive f potentials at high pH were reported in [132]. A mistake is very unlikely because for several other materials the same paper reports a usual behavior (negative f potentials at high pH and positive f potentials at low pH). Negative f potentials of travertine at low pH and positive f potentials at high pH were reported in [404]. A sign reversal from negative to positive on increase in pH is not an IEP in the sense discussed in the present paper. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcis.2014.02.036.
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
The pH dependent surface charging and points of zero charge. VI. Update Marek Kosmulski Lublin University of Technology, Nadbystrzycka 38, PL-20618 Lublin, Poland
a r t i c l e
i n f o
Article history: Received 17 January 2014 Accepted 25 February 2014 Available online 12 March 2014 Keywords: Point of zero charge Isoelectric point Surface charge density Zeta potential Electrokinetic potential
a b s t r a c t The pristine points of zero charge (PZC) and isoelectric points (IEP) of metal oxides from the recent literature are summarized. This study is an update of the previous compilation (Kosmulski, 2009) [5] and of its previous updates (Kosmulski, 2009, 2011) [6,7]. Only the IEP of materials other than metal oxides are reported, and the PZC of such materials obtained by potentiometric titration and related methods are ignored. IEP of (nominally) CoO, Fe(OH)2, Gd2O3, Ni2O3, and Sb2O3 have been reported in the recent literature. Those materials have not been studied before. Ó 2014 Elsevier Inc. All rights reserved.
Introduction The pH-dependent surface-charging of solid particles in aqueous solutions of 1–1 electrolytes has been extensively studied. The points of zero charge (PZC) and isoelectric points (IEP) of metal oxides in 0.0001–0.1 M solutions of alkali halides, nitrates(V), or chlorates(VII) are termed pristine PZC/IEP, and they were compiled in the classical paper by Parks [1]. The pristine PZC are independent of the nature of the salt, its concentration and other experimental conditions, thus they characterize the surface-charging behavior over a broad range of experimental conditions. Ions other than alkali metal cations or halide, nitrate(V), or chlorate(VII) anions often adsorb specifically, and induce a shift of PZC and IEP away from the pristine value. The PZC/IEP observed in the presence of specifically adsorbing ions are of limited significance, because they depend on the nature of the salt, its concentration, and other experimental conditions, and they are ignored in the present review. For example, Mehdilo et al. [2] used H2SO4 to adjust the pH of dispersions of ilmenite. Their IEPs are not reported in this review, because sulfate anions are known to adsorb specifically on iron compounds, and to induce a shift in their IEP to low pH values. Several recent publications compile pristine PZC of a specific chemical compound(s) or of certain class of chemical compounds. For example Chorover compiled PZC of soil constituents [3]. In contrast, relatively few publications report PZC of broad ranges of materials. Extensive compilations of pristine PZC of metal oxides and other materials were published by Kosmulski [4,5]. The E-mail address: [email protected] http://dx.doi.org/10.1016/j.jcis.2014.02.036 0021-9797/Ó 2014 Elsevier Inc. All rights reserved.
classical Parks’ review is still more frequently cited as a reference for ‘‘recommended’’ PZC of metal (hydr)oxides than up-to-date reviews as illustrated in Fig. 1. Several recent publications, which cite [1,4,5] are discussed in this compilation. The recent review [5] was updated twice [6,7] and the recent results (2011–2013) and a few older results (overlooked in [5–7]) are compiled in the Supplementary material in Table 1 (not available in the printed copy of the journal). The significance of the pristine PZC/IEP and their applications were discussed in detail in Ref. [1–7], in many other compilations of PZC/IEP cited therein, and in handbooks of surface and colloid chemistry. The PZC reported in review papers are especially important for the scientists who do not determine the PZC of their specimens themselves, but who rely on the PZC of similar specimens taken from the literature. In principle the PZC/IEP ‘‘recommended’’ in various reviews are consistent, and the choice of this or another compilation is not crucial. Yet erroneous pristine PZC allegedly based on the review papers are commonplace in the recent literature. For example Sojka et al. [8] used an erroneous value of PZC of niobia (2.8); Usui et al. [9] used erroneous values of PZC of hematite (4.2–6.9) and of goethite (5.9–6.7); and Liu et al. [10] used an erroneous value of PZC of hematite (6.7), taken from old compilations. The PZC reported in review papers are also useful for the scientists who determined the PZC of their specimens experimentally. The PZC of similar specimens taken from the literature are used to verify the purity of the samples and the correctness of the procedures. Comparison of own results with the literature (original publications or compilations) is commonplace in papers reporting
210
M. Kosmulski / Journal of Colloid and Interface Science 426 (2014) 209–212
geographical locations (mine, region, country). The formula and concentration (concentration range) of the electrolyte(s), temperature, method, and instrument used in the surface charging measurements are reported in separate columns of Supplementary Table 1 (not available in the printed copy of the journal). Empty boxes denote that the information was not available or not applicable.
Methods Experimental
Fig. 1. Citations of selected compilations of PZC/IEP.
PZC/IEP measurements. However, several recent studies reported rather unusual pristine IEP of common metal oxides, e.g., 6 for alumina [11] and 4.2 for Fe2O3 [12], without a reference to the existing literature on similar specimens. This is not clear whether the authors realized that their IEP were very different from the IEP of similar specimens reported by the others. The present review reports selected results from the literature. The pristine PZC and IEP are summarized in Supplementary Table 1 (not available in the printed copy of the journal). Erroneous or less credible results were deliberately neglected. Also the PZC/IEP reported in the original papers were verified when the original data (f potential or surface charge density r0 as a function of pH) were available. Several publications report f potential or r0 measured in multicomponent dispersions (dispersions containing various types of colloidal particles). The r0 in multi-component dispersions is additive, that is, the information about the composition of dispersion and the charging behavior of individual components makes it possible to predict the charging behavior of multi-component dispersions. In contrast there is no simple relationship between the apparent f potential observed in a mixture of various minerals and the composition of that mixture. Erroneous allegations in this respect are published now and then even in top-quality scientific journals, e.g., Eq. (2) in [13]. Kosmulski [14] studied the f potentials in dispersions containing two types of colloidal particles. The background subtraction procedure made it possible to calculate the f potential of individual types of colloidal particles in two-component dispersions. However, such a background subtraction procedure is only valid in the absence of heterocoagulation. The present review is focused on the PZC/IEP obtained in one-component dispersions, and the PZC/IEP allegedly obtained in multi-component dispersions (e.g., soils and rocks) were deliberately ignored.
Structure of Table 1 The structure of Supplementary Table 1 (not available in the printed copy of the journal) follows the same principles as Table 1 in Refs. [6,7]. The materials are organized into 13 categories according to their chemical composition. Within each category the materials are sorted by their chemical formula. Multiple specimens corresponding to the same (idealized) formula are sorted into the following categories: commercial, synthetic, and natural materials. The commercial specimens are characterized by their manufacturer and/or trade name. The synthetic materials are characterized by the recipe. The natural materials are characterized by
The terminology of experimental methods used in the literature is not consistent. Very often the same method is described by various names or the same name is used for different methods. The following terminology, adapted from previous compilations [4–7] is used in Supplementary Table 1 (not available in the printed copy of the journal): cip (common intersection point of potentiometric titration curves obtained at three or more ionic strengths) intersection (intersection point of potentiometric titration curves obtained at two ionic strengths) iep (isoelectric point obtained by electrokinetic or electroacoustic measurements) salt titration (salt addition) : addition of inert electrolyte (powder or concentrated solution) induces a shift in the pH of dispersion toward the pristine PZC. The pH value, at which salt addition does not induce any shift in the pH is equivalent to CIP. pH (natural pH of dispersions, e.g., obtained by mass titration). Such results were deliberately ignored in most materials. Silica (for which cip is not observed) is an exception. Moreover in a few papers reporting PZC of less common materials, the description of the method was not precise enough, and the ‘‘method’’ in Supplementary Table 1 (not available in the printed copy of the journal) is followed by ‘‘?’’. On top of measurements carried out by standard methods, a few publications present less common solutions. Corbett et al. [15] proposed a technique to measure the f potential of macroscopic specimens. They observed the velocity of tracer particles between two flat parallel electrodes as the function of the distance from the studied surface (in the range 0.1–1 mm), which was perpendicular to those electrodes. The new method has an advantage of using small specimens of solid and small volumes of liquid as compared to standard electroosmotic or streaming potential measurements, and it inherits most disadvantages of those methods (difficulties in the pH measurement, sensitivity to traces of impurities). Jordan et al. [16] obtained a hematite layer by atomic layer deposition and report a PZC of that layer at pH 5.5, derived from second harmonic generation (SHG) studies. Their PZC is substantially lower than the PZC/IEP usually reported for hematite powders. Sung et al. [17] studied an a-alumina single crystal, and report a PZC of (1–102) plane, derived from sum frequency vibrational spectroscopy (3230 cm 1 band) at pH 6.7. Their PZC is substantially lower than the PZC/IEP usually reported for alumina powders. Similar discrepancies (e.g., SHG vs. electrophoresis) are well-documented in the older literature. SHG is not considered a standard method in the present study, but a few PZC obtained by SHG are reported in Supplementary Table 1 (not available in the printed copy of the journal). There is a substantial difference between the present approach and that in Refs. [4–7], namely the PZC obtained by means of titration at one ionic strength (which in fact are natural pH values of dispersions) for materials other than metal oxides were ignored
M. Kosmulski / Journal of Colloid and Interface Science 426 (2014) 209–212
in the present update. It was shown recently [18] that in materials other than metal oxides, the IEP obtained in different laboratories and with various specimens of nominally the same chemical compound are relatively consistent, and the PZC obtained by titration are not. The charging curves of materials other than metal oxides usually do not show a sharp CIP, and the apparent PZC is sensitive to the solid-to-liquid ratio, ionic strength, and other experimental conditions. Accordingly the PZC obtained by titration at one ionic strength, mass titration, and equivalent methods (natural pH of dispersions) are ignored in the present update, also for metal oxides. Supplementary Table 1 (not available in the printed copy of the journal) compares the newly reported data with the median values of PZC/IEP from [5–7]. With 1 or 2 references in [5–7], the PZC/IEP from the older literature are explicitly reported. The number of entries (the second last column of Supplementary Table 1 (not available in the printed copy of the journal)) recalled in the present review is often lower than the numbers of entries in previews updates [6,7]. This is because numerous PZC from [5–7] (natural pH of dispersions) was not used as references in the present study. Section number in [5] is also indicated in Supplementary Table 1 (not available in the printed copy of the journal). In a few materials, ions other than H+ determine the sign of the f potential and r0. The PZC/IEP of those materials are reported in terms of the concentrations of the potential determining ions (e.g., pAg for silver halides) rather than pH. A recent study of Kallay et al. [19] revealed that the f potential of silver halides is not pHindependent (as it was originally believed). A shift in the pH from 6 to 3 induced a shift in the IEP of silver halides to higher pAg by about 1 pAg unit. Presentation of data Several publications cited in Supplementary Table 1 (not available in the printed copy of the journal) report only the value of PZC/IEP, but they do not report the data points upon which those values were based. Those PZC/IEP are marked by an asterisk in Supplementary Table 1 (not available in the printed copy of the journal). Most publications cited in Supplementary Table 1 (not available in the printed copy of the journal) report the data points (f potential and/or r0 as a function of pH) and the PZC/IEP based upon those data points. In most studies the PZC/IEP were determined correctly from the data points. A few examples of controversial interpretation of data points were found. For example Missana et al. [20] report electrokinetic curves of titania for two ionic strengths. For 1.25 10 4 M NaClO4 they report data points in the vicinity of the IEP on its both sides, and an IEP at pH 6.5 which is close to the results found in other publications. However, for 10 2 M NaClO4, no data points in the pH range 4–7 were available, and an arbitrary line was drawn suggesting an IEP at pH 4.5. The apparent IEP at pH 4.5 were not reported in this review, because it is not supported by the data points. An interesting method of presentation of electrokinetic data was chosen by Ottofuelling et al. [21]. With pH as the abscissa and ionic strength as the ordinate (log scale) the iso-f potential lines were plotted every 10 mV, and the ranges (every 10 mV) were marked by different colors. Their IEP are not shown in Supplementary Table 1 (not available in the printed copy of the journal) because they used NaHCO3 to adjust the pH. The presence of carbonate anions is likely to induce a shift in the IEP of metal oxides. A non-standard method of presentation of electrokinetic data was chosen by Colombo et al. [22]. They studied various samples of hematite. With particle diameter as the abscissa and pH or zeta potential as ordinate they plotted two curves: natural pH(d) and f(d), in one graph. Taking each couple of data points corresponding to certain d, a f(pH) graph can be constructed from their data. A random mixture of positive and negative f potentials was obtained
211
at pH about 6, suggesting an IEP of hematite at that pH. This low IEP is not surprising since most studied specimens were obtained in the presence of specifically adsorbing anions (e.g., oxalate). Discussion of Table 1 Choice of specimens Many studies of PZC//IEP and many entries in Supplementary Table 1 (not available in the printed copy of the journal) refer to chemical compounds whose PZC/IEP are already well-documented. Usually the values reported in older publications for similar specimens were confirmed in the recent literature. There are more PZC/IEP reported for metal oxides than for all other chemical compounds together, and more PZC/IEP for aluminum, iron, and titanium oxides than for all other metal oxides together. Several commercial products that were very popular in the past (e.g. P-25 titania from Evonik/Degussa) received also much attention in the recent literature. A few PZC/IEP of less-well-documented materials were published in the recent literature. The IEP of CoO [45], Fe(OH)2 [79,80], Gd2O3 [45], Ni2O3 [45], and Sb2O3 [144] were recently published for the first time. The surface-charging measurements with those materials are challenging, because of their chemical properties. For example Fe(OH)2 is unstable against oxidation to Fe(III), and in spite of all efforts reported in [79,80], some degree of surface oxidation might have occurred. The surface oxidation might have affected the chemical composition of the surface (iron III and II rather than pure iron II), and thus the measured IEP. Gd2O3 absorbs CO2 from the atmosphere and turns into carbonate, which probably has lower IEP than oxide. Similar CO2 absorption and its effect on the IEP is well-documented for other lanthanides and for the 3-group metals’ sesquioxides. Apparently a commercial reagent was used in [45], and no efforts to remove CO2 from the material are reported. The IEP of Gd2O3 reported in [45] is lower than IEP of sesquioxides reported in the literature, probably due to CO2absorption. The point of zero charge of CaO [66,67] was recently reported, but in fact those experimental results are representative for hydroxide or carbonate rather than for oxide, which is very reactive. Most studies were carried out with powders, and the surfacecharging studies with single crystals are rare. IEP of ZnO single crystal [291] was lower than IEP reported for powders. Discrepancies between IEP of single crystals on the one hand and of powders on the other have been observed and discussed for alumina and for other metal oxides, but no generally accepted explanation of those discrepancies is available. Most electrokinetic studies reported in the recent literature were performed for one type of 1–1 electrolyte and for one ionic strength (cf. Supplementary Table 1 (not available in the printed copy of the journal)), and sometimes the nature and concentration of the electrolyte were not reported at all. Such studies are based on the assumption that certain electrolyte is inert, and no specific adsorption occurs. With metal oxides the inert character of alkali halides and nitrates is well-documented, and generally accepted, but with most materials other than metal oxides, the inert character of those electrolytes has yet to be proved. Electrokinetic curves of mixed Ti–Si oxide in two different electrolytes, at two different concentrations each [340] confirmed the inert character of those electrolytes, namely the IEP were independent of the nature of electrolyte and ionic strength. This is one of a few examples of systematic studies in this direction found in the recent publications. Temperature, pressure, and solvent effects on the PZC Most results reported in Supplementary Table 1 (not available in the printed copy of the journal) refer to room temperature and
212
M. Kosmulski / Journal of Colloid and Interface Science 426 (2014) 209–212
atmospheric pressure, and in a few papers, systematic studies of the temperature effects can be found. Temperature effect on the PZC and IEP of NiO [141], CIP of zirconia [295], CIP of alumina [58], and IEP of anatase [244] is reported in the recent literature. The f potentials of silica at various temperatures and pressures [173] and the f potentials of silica at 100 and 150 °C [167] have been reported. The recent studies confirm the previously found trends, namely the PZC shifts to low pH when the temperature increases. Temperature effect on the IEP of magnetite was studied in [399]. The IEP at 25 °C from that study are not reported in Supplementary Table 1 (not available in the printed copy of the journal), because it was obtained by arbitrary interpolation (no data points near the IEP). Lvov et al. [400] report the pH effect on the electrophoretic mobility of commercial silica at 20 bar. The mobility at 25 °C was negative and rather insensitive to pH over the range 4–8. The measurements were carried out at pressures up to 40 bar and temperatures up to 150 °C. Studies at such high temperatures and pressures cannot be carried out by means of commercial zetameters, and special equipment had to be designed. Very few studies of surface charging in nonaqueous and mixed solvents were published over the recent 3 years, and they were ignored in Supplementary Table 1 (not available in the printed copy of the journal). Turci et al. [401] report IEP of 3 commercial titanias in 10 % (v/v) aqueous DMSO at pH 5–5.5. At low ionic strengths, admixtures of non-aqueous solvents have been shown to have an insignificant effect on the IEP of titania. Therefore the above IEP are probably representative also for aqueous dispersions.
Correlations The surface-charging behavior is correlated with several other physical properties. These correlations are indicated in Supplementary Table 1 (not available in the printed copy of the journal), in footnotes. The correlation between the IEP on the one hand and the minimum in colloid stability and the maximum in the particle size is the most well-known one, and numerous examples confirming such a correlation are presented in Supplementary Table 1 (not available in the printed copy of the journal). Most modern zetameters offer a possibility of a measurement of f potential and particle size using the same sample of dispersion. However, there are a few exceptions, e.g., the particle size of P-25 titania reported in [238] was rather insensitive to pH, also in the vicinity of the IEP. The role of the f potential as a parameter defining the colloid stability is sometimes overrated. For example Ref. [132] presents a table, in which intervals of absolute values of f potential are assigned to certain degrees of stability (instable dispersion, good stability, excellent stability, etc.). Such a precise assignment is incorrect, because the stability (at a constant f) depends on the ionic strength and Hamaker constant. In silica unlike in metal oxides, the stability minimum is not correlated with the IEP. This exception has been well-known for decades, and the recent studies, e.g., [189] confirm the exceptional behavior of silica. The IEP are correlated with a maximum in the yield stress and in the viscosity of concentrated dispersions. This well-known correlation is confirmed by a few examples reported in Supplementary Table 1 (not available in the printed copy of the journal).
Several studies report series of surface-charging data obtained for particles of various sizes having (nominally) the same chemical composition. For example effects of particle size on CIP of ferrihydrite [129] and on IEP of titania [219] were claimed. A graph PZC/ IEP (particle size) of hematite is presented in [100] (supporting information), but no clear correlation was found. Allegations about the particle size effect on PZC/IEP can also be found in the older literature, but the present author is skeptical about them. Namely, it is practically impossible to prepare dispersions, which vary only in particle size, and to eliminate the other factors (e.g., minor impurities), which are responsible for the scatter in the PZC/IEP of allegedly identical specimens (e.g., P-25 from Degussa) reported in different publications. Most likely the apparent effect of particle size on PZC/IEP is actually due to factors, which were beyond the control in course of experiments. The effect of crystallographic structure (rutile vs. anatase) on the IEP was studied [270], and the following acidity sequence: brookite > anatase > rutile was claimed. Such results are of limited significance, because the effect of surface chemistry on the IEP was neglected. For example one of the specimens was obtained in the presence of MgSO4. Both Mg(II) cations and sulfate anions adsorb specifically on the surfaces of metal oxides and shift their IEP. An increase in the IEP with the solid-to liquid ratio (50– 500 ppm) was reported for alumina and titania [53]. An increase in the IEP with the solid-to liquid ratio (1–100 ppm) was reported for four different powders [86]. These results were affected by specific adsorption of anions, namely, two studied specimens were iron (hydr)oxides obtained in the presence of sulfates. Another specimen was a commercial powder, which probably contained surface-active agent(s). Generally the IEP in the absence of specific adsorption do not depend on the solid-to liquid ratio, and experiments carried out at various solid-to-liquid ratios can be used to confirm the purity of specimens, and correctness of procedures. Fisher et al. [402,403] report the pH effect on the contact angles of plasma-treated silica substrates, and interpret the maxima in the contact angle(pH) curves as the isoelectric points. The correlation between the IEP and the maximum in the contact angle is well-know, but the contact-angle-measurement is not considered as a standalone method of IEP measurement in the present study, and such results are not reported in Supplementary Table 1 (not available in the printed copy of the journal). Unusual results Negative f potentials of MgCO3 at low pH and positive f potentials at high pH were reported in [132]. A mistake is very unlikely because for several other materials the same paper reports a usual behavior (negative f potentials at high pH and positive f potentials at low pH). Negative f potentials of travertine at low pH and positive f potentials at high pH were reported in [404]. A sign reversal from negative to positive on increase in pH is not an IEP in the sense discussed in the present paper. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcis.2014.02.036.
