This study is primarily intended to provide a better understanding of the adsorption of ions on hematite (α-Fe 2
). In addition, due attention is given to the relation between the ionic adsorption and the colloidal stability of hematite sols.
Chapter 1. is concerned with the motivation and outline of the present study. The role of iron oxides in soils is discussed as an example of the practical importance of these materials. Several important plant nutrients (e.g. N, S, P, Mo and B) may be adsorbed on iron oxides in anionic form. Furthermore, the adsorption of cations by the soil and also the soil structure may be directly or indirectly affected by the iron oxides. When this study was started only a few fundamental colloid-chemical studies had been made on the iron oxide system. In particular, studies were lacking in which both the adsorption of ions and the colloidal stability had been discussed. Such investigations had been made, for example with silver iodide, and these studies served as a model for the present one. Hematite was chosen to represent the group of iron oxides because it plays an important role in (lateritic) soils and because it can be readily identified and also prepared and handled as a sol.
Chapter 2. deals with the preparation of the samples and their characterization, with special emphasis on the surface properties. A reasonably fast method was developed for the preparation of hematite. The samples contained little or none of the other forms of iron oxides but did contain traces of (strongly adsorbed) Ca 2+
and Mg 2+
ions. The electron micrographs indicated more or less globular particles with diameters between 400 and 700 Å. The specific surface area was determined with the BET method using a continuous flow technique and N 2
as adsorbate. Efforts to measure the surface area also by the negative adsorption method (at pH 4 with Ca 2+
as co-ion) encountered serious practical problems. In particular the release of Ca` ions by the samples hampered the measurements. By using Ca 45
this difficulty was largely overcome. The negative adsorption area was lower than the BET area, suggesting some degree of porosity.
Evidence for surface porosity was also obtained from N 2
adsorption data measured by the static method. The adsorption isotherm exhibits a hysteresis loop. It is a type IV isotherm implying that the pores have a width between 20 and 200 Å. Porosity is also evident, although less clearly, from the t
-plot method. A quantitative estimate of the degree of porosity from the t
-plot method is not possible (as yet) because of the lack of standard data for non-porous hematite.
Thermogravimetric data indirectly revealed that the surface layers of the hematite particles are hydrated or, more precisely, hydroxylated.
Measurements and calculations of the surface charge (due to adsorption of potential-determining ions) and of the counter charge (due to the distribution of indifferent ions), as a function of pH and salt concentration are described in Chapter 3. The surface charge, σ 0
was determined by potentiometric acidbase titrations. Preliminary measurements had indicated that the system could only be kept free of CO 2
by means of a gas-tight cell with a controlled N 2
atmosphere. The pH was measured with a glass electrode. A reference calomel electrode was connected with the solution (or the sol) in the titration vessel by means of a salt bridge. This construction led to fewer practical problems than the use of a 'VAN LAAR bridge' and the direct insertion of the calomel electrode in the solution. The fact that, using this device, the suspension effect does not interfere with the measurement of pH was of minor importance in this study.
The standard titration procedure used was a fast acid titration from pH 10 to 4. The 'wet' samples (always kept in contact with solution) were equilibrated at pH 10 before each titration. The amount of acid adsorbed during the slow adsorption step following the initial fast process did not amount to more than 20 % of the total adsorption. The reproducibility was about 3 %.
The point of zero charge (p.z.c.) was determined from the intersection point of the (σ 0
-pH) curves at different salt concentrations. To check the p.z.c.'s obtained in this way, the 'addition method' was also employed, i.e. the change in pH caused by the addition of (dried) hematite was recorded. Good agreement was obtained for the (2-1) electrolytes. For the other electrolytes studied the addition method gave lower values, possibly due to the drying of the samples.
At the end of Chapter 3. it is described how for some electrolytes both the surface charge and the counter charge have been measured under the same conditions.
The results and discussion of the surface charge and counter charge measurements are given in Chapter 4. The electrolytes studied were the chlorides of Li +
, K +
and Cs +
, the nitrates of Mg 2+
, Ca 2+
, Sr 2+
and Ba 2+
, and K 2
. The (σ 0
-pH) curves on hematite resemble those on other (iron) oxides. The curves are generally convex with respect to the pH axis and the surface charge attains much higher values (several tens of μC
) than on, for example, silver iodide and mercury. The very high surface charge values for precipitated silica and for glass reported by other workers have been attributed to penetration into the solid by the potential-determining ions as well as by the counterions. It was therefore checked whether this was also the case for hematite.
The results for Li +
and Mg 2+
at high salt concentration do indeed point in this direction. Under these conditions both ions are strongly preferentially adsorbed compared to the other alkali or alkaline earth ions. This is presumably related to the fact that of the cations studied only Li +
and Mg 2+
fit in the octahedral holes of the hematite lattice. Finally, it appears that the (σ 0
-pH) curves for Li +
and Mg 2+
are fairly well described by the porous double layer theory of LYKLEMA. For the other ions the indications for porosity are less clear. However, on the basis of the current double layer models for oxides it is equally impossible to demonstrate that the (σ 0
-pH) curves should be explained without assuming penetration.
A conspicuous feature of the (σ 0
-pH) curves is the fact that the p.z.c. shifts due to the presence of certain electrolytes. This phenomenon has hitherto not been described for oxides. In the presence of KCI the p.z.c. occurs at pH 8.5. In the presence of LiCl and the alkaline earth nitrates a shift of the p.z.c. towards lower pH values was found. In contrast to this, the p.z.c. shifted towards a higher pH in the presence of K 2
. These shifts are the result of (specific) adsorption of cations and anions, respectively, at the p.z.c. The direction in which the p.z.c. shifts was found to be in accordance with the thermodynamically derived ESIN-MARKOV coefficient. When verifiable, namely at high surface charge, fair quantitative agreement was also found between the theoretical and experimental values for this coefficient.
The measurements of the surface charge and the counter charge under the same conditions gave the expected results regarding both the distribution of the counter charge over cations and anions and the surface charge-counter charge balance.
Chapter 5. deals with the method and results of the stability experiments. The classical method of visually determining the coagulation concentration in a series of coagulation tubes was applied. The diffuse double layer potential was calculated from the coagulation value using the DLVO theory for spherical particles. The diffuse layer charge was then calculated and, by substraction of this charge from the surface charge, the charge in the STERN layer (and/or the solid phase) was obtained.
It is shown that the monovalent cations and anions in the double layer are for the greater part (more than 85 % under coagulation conditions) located in the STERN layer and/or the solid phase. For the alkaline earth ions (and presumably also for the sulfate ion) this percentage rises to virtually 100 % as is apparent from the instability of the hematite sol when these ions are present as counterions. With the alkaline earth ions reversal of charge occurs between pH 8.5 and pH 6.5. The expected shift in the iso-electric point towards a higher pH was not observed, possibly because the stability measurements were not sensitive enough in this respect.
It appears that with the electrolytes discussed so far the ions act as normal counterions, even though they may be strongly specifically adsorbed. This is not the case with phosphate. The adsorption of this ion is therefore treated separately in Chapter 6. There the relevant literature is briefly discussed first. A workinghypothesis is then postulated which implies that adsorption of phosphate ions occurs by exchange reactions (as given in Chapter 6.) of these ions with surface groups, in particular -OH and -OH2+
. It is furthermore assumed that the phosphate dissociation equilibria do not alter upon adsorption. The experiments performed concern the direct measurement of (1) the adsorption of phosphate (2) the amount of acid required to keep the pH constant during this adsorption and (3) the stability of the hematite sols in the presence of phosphate.
Regarding the adsorption of phosphate the proposed exchange reactions explain the adsorption on the negative hematite surface and the increase in adsorption with decreasing pH and increasing concentration of indifferent electrolyte. It was not possible to explain the relatively strong increase in adsorption around pH = p K
. (It was suggested that this feature might be related to incomplete equilibration with the solution). The amount of acid consumed during the adsorption of phosphate is also in accordance with the exchange reactions. However, the influence of the concentration of indifferent electrolyte was not always clear. The adsorption of phosphate and the stoichiometry of the exchange reactions could thus partially be explained in terms of the proposed reactions. Further verification is needed. The stability measurements indicated, i.a., that after coagulation of the hematite sol at very low concentrations only polyvalent phosphate ions (i.c. the bivalent ions) could produce restabilization. With regard to the exchange reactions this means that these ions act as potential-determining ions. It is finally noted that there are no definite indications for the occurrence of more than one adsorption mechanism.
Summarizing it is concluded that the adsorption of ions on hematite appears to be characterized by a large variety of interactions with the surface. In order of increasing degree of interaction, the following sequence can be set up:
1. the chlorides of K +
and Cs +
. More than 85 % of the ions in the STERN layer and/or the solid phase. No shift in p.z.c. (pH 0
2. LiCl. Behaviour at low salt concentrations similar to KCl. At high salt concentrations penetration into the solid phase and shift in the p.z.c. (pH 0
= 7.8 in I M).
3. K 2
. The p.z.c. shifts to a higher pH (pH 0
= 9.6). No restabilization.
4. the nitrates of Ca 2+
, Sr 2+
and Ba 2+
. (Almost) completely in the Stern layer. The p.z.c. shifts to a lower pH (pH 0
= 6.5). Reversal of charge between pH 8.5 and 6.5.
5. Mg(NO 3
. Behaviour at low salt concentrations similar to the other alkaline earth ions. At high salt concentrations penetration as with Li +
6. Phosphate. Exchange against surface groups (mainly -OH and -OH2+
). Reversal of charge between pH 8.5 and 6.0. Polyvalent ions potential-determining.
A further conclusion is that a better understanding of the adsorption of ions on iron oxides, and on oxides in general, can be achieved by using a combination of surface-chemical and colloid-chemical techniques, in particular the measurements of the surface charge and the stability.