XPS studies on the electronic structure of bonding between solid and solutes: adsorption of arsenate, chromate, phosphate, Pb2+, and Zn2+ ions on amorphous black ferric oxyhydroxide

Geochimica et Cosmochimica Acta, Vol. 64, No. 7, pp. 1209 –1219, 2000 Copyright © 2000 Elsevier Science Ltd Printed in the USA. All rights reserved 0016-7037/⫺1900 $20.00 ⫹ .00

Pergamon

PII S00167037(99)00386-5

XPS studies on the electronic structure of bonding between solid and solutes: Adsorption of arsenate, chromate, phosphate, Pb2ⴙ, and Zn2ⴙ ions on amorphous black ferric oxyhydroxide M. DING,1,† B. H. W. S.

1,

DE JONG,

* S. J. ROOSENDAAL,2 and A. VREDENBERG2

1

Institute for Earth Sciences, Utrecht University, 3584 CD Utrecht, the Netherlands Debye Institute, Section Interface Physics, Utrecht University, 3508 TA Utrecht, the Netherlands

2

(Received December 14, 1998; accepted in revised form October 5, 1999) 3⫺ 2⫺ 2⫹ Abstract—The adsorption of three anions (AsO3⫺ and Pb2⫹) has 4 , PO4 , and CrO4 ) and two cations (Zn been determined on a new form of amorphous black ferric oxyhydroxide that resembles in local structure ␤-FeOOH, akaganeite. The nature of the interaction between FeOOH substrate and the above mentioned adsorbates is characterised with X-ray photoelectron spectroscopy on the core and valence band levels of trivalent iron and oxygen using frontier molecular orbital theory as theoretical framework. Our findings indicate that substantial and variable charge transfer occurs in which the FeOOH substrate can function either as a Lewis acid or base in its interaction with different adsorbates. Our findings suggest that the valence band spectra can be used to estimate qualitatively the energy level separation between the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of the surface complexes and to assess the chemical affinity between substrate and adsorbate. Copyright © 2000 Elsevier Science Ltd

bates (Becker et al., 1997) and provide more direct and quantitative information of the sorbed complex, as for instance summarised by Brown (1990) in his spectroscopic studies of chemisorption reaction mechanism at oxide–water interfaces. Among these techniques, extended x-ray adsorption fine structure (EXAFS) has been widely adopted to probe the surface structure (Waychunas et al., 1995, 1996) primarily in an attempt to delineate the various geometries of adsorption complexes. From a quantum mechanical perspective the state of describing the interaction between adsorbate, the adions or admolecules, and substrate remains by and large wanting. To get to this next stage requires exploration of the nature of the chemical reaction between solid and solute. To accomplish this, the direction of charge transfer between substrate and adsorbate needs to be known, as well as the relative position of their empty and occupied energy states. XPS, as a surface analytical technique, provides the proper level of information to answer these questions or at least provide internally consistent rationalisations between spectroscopic and solution chemical observations (e.g., Martin and Smart, 1987; Rosso et al., 1999a; Rosso et al., 1999b). XPS is a surface analytical technique devised at the end of 1960s to provide chemical analyses of surfaces (Siegbahn et al., 1967). Since the early 1970s studies on FeOOH have been carried out focussing primarily on testing core levels spectra of pure materials. These studies revealed that chemical shielding affects the position of these core levels. To quantify the nature of these shifts Bagus and Bauschlicher (1980) carried out Hartree Fock calculations on an oxygen atom with varying charge showing that variations in valence state occupancy can cause a chemical shift of about 20 eV for the oxygen (1s) [O(1s)] core level. A simple formula based on these calculations relating O(1s)-binding energy with oxygen valence charge was subsequently devised to get an internally consistent oxygen charge for oxide materials (de Jong, 1989; de Jong et al., 1994; de Jong and Ding, 2000). Here, we shall test if the

1. INTRODUCTION

Hydrous ferric oxide has been the subject of a large number of studies reviewed thoroughly in recent chemical, geochemical, and hydrometallurgical books such as Cornell and Schwertman (1996), Dzombak and Morel (1990), Voigt et al. (1997), de Boer (1999), and Dutrizac and Harris (1996). In a study on self-sealing/healing iron oxide layer formation in porous media (Ding et al., 1996, 1997, 1998a,b; Ding, 1998), we have reported previously (Ding et al., 1998b) the synthesis of a new form of amorphous black ferric oxyhydroxide. This colloidal material, which upon heating transforms to hematite, resembles in local structure akaganeite. It has been characterised with X-ray photoelectron spectroscopy (XPS), conversion electron Mo¨ssbauer spectroscopy, scanning tunnelling microscopy, and classic aquatic adsorption experiments (Ding et al., 1998b; Ding, 1998). The material is ideally suited to study the nature of the interaction between solid substrate and adsorbates, because it is amorphous without preferred crystal planes, possesses a large surface area of about 300 m2/g, and adsorbs sufficient concentrations of adatoms or molecules to be detectable by surface analytical techniques. Extensive theoretical and experimental work on cation and anion absorption on ferric oxyhydroxides has been carried out in the past 30 years, focussing primarily on solutes in aquatic systems. However, much remains to be done in terms of detailed chemical reaction mechanisms involving characterisable elementary chemical steps between substrate and adsorbate. Recent developments in analytical technique have started to provide detailed structural and chemical observations of adsor-

* Author to whom correspondence should be addressed (bernard@ geo.uu.nl). † Present address: Los Alamos National Laboratory; Chemical Science & Technology Division; Environmental Science & Waste Technology. CST-7, MS J 514, Los Alamos, NM 87545, USA. 1209

1210

M. Ding et al.

oxygen and iron(III) charge will vary when an FeOOH substrate reacts with adsorbates enabling an assessment of the magnitude and direction of electron transfer between them. In addition, we shall rationalise how the valence band structure of FeOOH changes as a function of different adsorbates using a frontier orbital perspective as laid down by Fukui et al. (1954), Klopman (1968, 1974), and Hoffman (1988). The purpose of this research is twofold: to determine the charge transfer between the FeOOH surface and its adsorbates, and to study the nature of the interaction between FeOOH substrate and adsorbates from a frontier orbital perspective. To reach these goals we have carried out XPS measurements on the pristine FeOOH substrate and its surface complexes. By tracking O(1s) and Fe core levels (e.g., Fe(2p), Fe(3p), and Fe(3s)) binding energy shifts relative to those of FeOOH itself, we determined the direction of charge transfer between substrate and adduct. The shifts in valence band XPS spectra of the surface complexes provide a qualitative measure of the resulting charge transfer between highest occupied and lowest unoccupied molecular orbital (HOMO–LUMO) and the character of bonding on surface complexes. 2. EXPERIMENTAL METHOD

2.1. Synthesis of Amorphous Iron Oxyhydroxide Amorphous iron oxyhydroxide was prepared in a flask by hydrolysis of a stirred 0.1 mol/L Fe(NO3)3 solution with 0.1 mol/L NH3 䡠 H2O. The temperature was kept constant (at 50°C) with a recirculating water bath. Base was introduced by automatic titration at a constant rate of about 2 drops/s, up to a solution pH of 7.5 after which the solution was stirred for another 30 min. The resulting suspension was filtered, washed first with distilled water, then three times with a pH 7 washing solution, and diluted with distilled water.

2.2. XPS Specimen Preparation Sample preparation for our XPS measurements involved running individual adsorption experiments at a constant pH of 4.5 (for arsenate and phosphate), pH 5.0 (chromate), pH 5.5 (for zinc), and pH 4.5(for lead). The selected pH values for oxyanions reflect their maximum adsorption capacity and to prevent zinc and lead surface precipitation during adsorption. Ten milliliters of 0.1 mol/L arsenate (in the form of H3AsO4), phosphate (in the form of NaH2PO4), chromate (in the form of K2CrO4), lead (in the form of Pb(NO3)2), and zinc (in the form of Zn(NO3)2) were mixed with a 200-mL Fe(OH)3 suspension containing 0.056 g of ferric iron. The ionic strength of the solutions was controlled with 0.1 N NaNO3. The experiment was conducted in a sealed reaction flask with continuous N2 sparge, keeping the temperature constant at 25°C. The suspension was stirred for 24 h. After filtration, the residue was dried overnight in an oven at 105°C. The resulting black specimen consisted of X-ray amorphous FeOOH only as determined with conversion electron Mo¨ssbauer spectroscopy. The dried material was powdered and pelletized under vacuum.

2.3. XPS Measurements XPS measurements were carried out on OCTOPUS, a multichamber ultra high vacuum (UHV) preparation and analysis system connected to a 3-MV single-ended and a 6-MV tandem van der Graaff generator, using a Clam-2 hemispherical sector analyzer and a VG XR2F2 twin anode X-ray source with standard AL/Mg anodes. Spectra were recorded using the AlK␣ source operated at a power of 120 W and a constant pass energy of 20 eV for the analyzer, using the C(1s) spectral line at 284.7 eV as reference (see Seah, 1990). Measurement of the chemical shift in duplicate showed good reproducibility of the binding energies for the peak positions. No random shifts due to charging for this semiconducting material (direct gap about 2 eV) were observed.

3. RESULTS

3.1. Cations and Oxyanions Identification Using XPS Spectra 3⫺ The cations Zn2⫹ and Pb2⫹, oxyanions AsO3⫺ 4 , PO4 , and CrO2⫺ can be detected as adsorbates on a FeO(OH) substrate 4 with XPS, as shown in Figure 1. The measured binding energies of each element [48.7 eV As(3d); 132.3 eV P(2p); 577.5 eV Cr(2p); 1023.3 eV Zn(2p); 138.2 eV Pb(4f)] agree well with reference XPS data (Moulder et al., 1992). The full width at half maximum (FWHM) of about 5 eV for the Cr(2p) line suggests different site occupancies of this element on the substrate. Chemical analyses show the presence of about 8 wt.%, 7 wt.%, 6 wt.%, 10 wt.%, and 9 wt.% of As, P, Cr, Zn, and Pb, respectively, adsorbed on the FeOOH substrate.

3.2. O(1s) Spectra of FeOOH Surface Complexes After exposing the FeOOH substrate to solutions containing arsenate, phosphate, and chromate oxyanions, changes in peak position occur in the spectrum of O(1s) relative to that of pure FeOOH as indicated in Figure 2a. Our results show that due to the chemical adsorption between substrate and adducts, the peak of O(1s) spectra of surface complexes has shifted to a less negative binding energy for arsenate and phosphate relative to that of FeOOH, and a more negative binding energy for chromate. In contrast the peak shift is negligible for substrates on which the cations lead and zinc are adsorbed as illustrated in Figure 2b. 3.3. Fe(2p) and Fe(3s, 3p) Spectra of FeOOH Surface Complexes Similarly, core level Fe(2p) spectra of adsorbates show different degrees of shift relative to the pure substrate, as shown in Figure 3. In comparison with FeOOH, the Fe(2p) XPS spectra of adsorbates of arsenate and chromate show the largest shift to more negative binding energy, whereas those of phosphate, zinc, and lead show only a slight shift to less negative binding energy. Similar trends to those in Fe(2p) XPS spectra, are detected in Fe(3s, 3p) spectra of FeOOH adsorbates as shown in Figure 4. 3.4. Valence Band Spectra of FeO(OH) Surface Complexes Figure 5 presents the valence band spectra of surface complexes and substrate. The valence band spectra of FeOOH–As and FeOOH–P complexes are almost identical to that of FeOOH. Those of FeOOH–Cr, FeOOH–Zn, and FeOOH–Pb distinctively differ from that of FeOOH. 4. DISCUSSION

The experimental conditions used in our adsorbtion experiments precluded the occurrence of redox reactions. Thus, the charge transfer considered here is solely of the acid– base type. Two issues require discussion. First, we shall argue that charge transfer can be deduced from core level XPS shifts of iron and oxygen. Second, we shall discuss the nature of this charge transfer, the relation between HOMO and LUMO of substrate

XPS characterized electronic bonds between solid FeOOH and solutes

1211

Fig. 1. (a,b) Photoelectron spectra of anions and cations on black iron oxyhydroxide.

and adsorbate, as deduced from XPS variations of the valence band. 4.1. Effect of Charge Transfer on the Binding Energy of the Core Levels As observed in our O(1s), Fe(2p), and Fe(3s,3p) spectra of FeOOH surface complexes, core level binding energy shifts occur relative to those of the FeOOH substrate. These core level shifts reflect electron transfer in the valence band, where the actual charge transfer due to chemical reactions has to take place. Therefore, by tracking the change in core level peak position of the substrate with and without formation of surface complexes, it is possible to deduce the direction of electron transfer during adsorption and assess the relative acidity and basicity of substrate and adsorbate. 4.1.1. The relation between O(1s) binding energy and the direction of charge transfer in the valence states The O(1s) binding energy and intensity shift depend on the concentration of different oxygen atoms in particular on the surface of a material. These core level shifts can be correlated with variations in charge density of oxygen in the valence band as shown by Bagus and Bauschlicher (1980) and illustrated in Figure 6. A simple formula connecting O(1s) chemical shift

and charge can be constructed having as principal advantage that all charges derived in this manner are internally consistent (de Jong, 1989; de Jong et al., 1994). Qo ⫽⫺4.372 ⫹ [385.023 ⫺ 8.976 䡠 (545.509 ⫺ O1sBE)]1/2/4.488

(1)

Here Qo is the actual oxygen charge in a material and O1sBE is the O(1s) binding energy as determined from XPS. Although the calculated binding energies of Bagus and Bauschlicher (1980) vary between 545.4 eV(O0.0) and 515.2 eV(O⫺2.0), the range observed to date for metal oxides varies between 533.6 and 528.3 eV, which according to Eqn. 1 corresponds to an oxygen charge of ⫺0.66 and ⫺0.99 esu, respectively (Moulder et al., 1992). Of course, it should be noted that charges as such never represent absolute values but depend on the way electron density is partitioned over the constituent atoms in a molecule. However, the values calculated according to the above formula tend to be close to those deduced from ab initio Hartree Fock calculations on oxygen-containing molecules using a Mulliken population analysis, for example, ⫺0.73 for O in SiO2 vs. ⫺0.81 calculated for the bridging oxygen atom in H6Si2O7 (de Jong, 1989; Burkhardt et al., 1991), so fair consistency prevails. The oxygen charge of FeOOH and its variation on surface

1212

M. Ding et al.

Fig. 2. (a,b) O(1s) core level photoelectron spectra of anions and cations adsorbed on black iron oxyhydroxide and on pristine FeOOH.

complexation can be calculated with Eqn. 1, the results of which are summarised in Table 1. Inspection of Table 1 indicates that oxygen can either become more or less negative relative to the oxygen charge on FeOOH, which we took as point of reference. Thus, the oxygen charge of the FeOOH–Cr surface complex is less negative than that of FeOOH itself, indicating that during the chemisorption of Cr on FeOOH electron transfer occurred from substrate to adsorbate. In contrast, the oxygen charge in surface complexes of FeOOH–As and FeOOH–P is more negative than that of FeOOH, indicating that during chemisorption of As and P on FeOOH electron transfer from adsorbate to substrate occurs. In other words, the surface oxygens of the FeOOH substrate act as Lewis base, an electron donor, to chromate admolecules, and as Lewis acid, an electron acceptor to arsenate and phosphate admolecules. The magnitude of the O(1s) chemical shift indicates the strength of the surface–

oxyanion interaction without giving explicit information of the nature of this interaction. The oxygen charges in both zinc and lead FeOOH surface complexes are similar to those of the pristine substrate, indicating that adsorption of lead and zinc on FeOOH does not involve noticeable electron transfer between surface and adsorbate and suggesting that these cations form weak bonds with the FeOOH substrate. 4.1.2. The relation between Fe core level binding energy and the direction of electron transfer in valence states The relation between O(1s) binding energy and its actual charge in a compound should of course be consistent with shifts and charges for Fe core level binding energies. Experimentally these shifts are on the order of 5 eV for iron metal, FeO, and Fe2O3. The iron charge can be calculated after the oxygen

XPS characterized electronic bonds between solid FeOOH and solutes

Fig. 3. (a,b) Fe(2p) core level photoelectron spectra of anions and cations adsorbed on black iron oxyhydroxide and on pristine FeOOH.

1213

1214

M. Ding et al.

Fig. 4. (a,b) Fe(3s) and Fe(3p) core level photoelectron spectra of anions and cations adsorbed on black iron oxyhydroxide and on pristine FeOOH.

charge of these compounds has been determined from the O(1s) XPS shift. These iron charges can be correlated with observed iron core level XPS data, resulting in linear relations between iron core level binding energies and iron charge. The results of this procedure are summarised in Table 2, where the iron core level binding energies for Fe(2p1/2, 2p3/2), Fe(3s), and Fe(3p) are shown as well as the formulas connecting charge Q with binding energy X. All the binding energy data are from McIntyre and Zeatruk (1977) (see also Brundle et al., 1977; Graat, 1998; Roosendaal et al., 1998). With the formulas in Table 2 we can relate the observed Fe core level binding energies for FeOOH and its surface complexes to their iron charge as summarised in Table 3. Our results in Table 3 suggest that the adsorption of arsenate and chromate on FeOOH affects the charge on the iron atom much

more than phosphate, zinc, or lead. In particular, the adsorption of zinc and lead on FeOOH has virtually no effect on the charge density on Fe. Combining the results from Table 1 and Table 3, as illustrated in Figure 7, reveals that the surface oxygen atoms as well as the iron atoms act as Lewis acids on adsorption of phosphate: iron becomes less positive and oxygen more negative relative to iron and oxygen in FeOOH. It also reveals that the surface oxygen atoms and iron atoms act as Lewis bases on adsorption of chromate: iron becomes more positive and oxygen less negative relative to iron and oxygen in FeOOH, and that the adsorption of zinc and lead on FeOOH involves virtually no charge transfer. However, our results also reveal mixed behavior as indicated by the adsorption of arsenate on FeOOH in which the surface oxygen atoms become more negative and

XPS characterized electronic bonds between solid FeOOH and solutes

1215

Fig. 5. (a,b) Valence band spectra of FeOOH and FeOOH with adsorbed cations and anions.

the iron atoms more positive vis-a´-vis those for FeOOH (the oxygen atom acts as a Lewis acid and the Fe atom acts as a Lewis base). Our results suggest that the surface oxygen layer functions as a membrane for electrons. Atoms attached to both sites of this membrane donate or accept electrons in all likelihood predetermined by the difference in electron negativity between adsorbates and Fe(III). For the elements of relevance here the Sanderson (1983) electronegativity scale gives a value for oxygen of 3.65, Fe(III) 2.20, phosphorus 2.52, Cr(VI) 3.37, Pb(II) 1.92, and Zn 2.23. The relative electronegativity difference between atoms on both sides of the oxygen surface varies between negligible as in the case of Fe(III)–Pb(II) and Fe(III)– Zn(II) to very large in the case of Fe(III)–Cr(VI). Thus, adsorbtion of zinc and lead does not affect the charge distribution significantly, whereas chromate affects the charge distribution in such a way that not only electron density is pulled away from the Fe(III) ions in the substrate but also from the oxygen

surface layer. Unfortunately the electronegativity value for As(V) has not been assessed (see Huheey et al., 1993), but its magnitude should be slightly larger than that of phosphorus accounting for the more positive charge on Fe(III) and the more negative charge on surface oxygens. 4.2. Interaction of Valence Orbitals on FeOOH Surface We have argued in the preceding section that core level shifts in XPS spectra can be interpreted in terms of charge transfer between substrate and adsorbate. The remaining issue concerns the nature of the bonds between these surface clusters: Are the bonds “ionic” or “covalent”? How can they be discriminated?, and Are there kinetic consequences associated with one type versus the other? For instance, it remains puzzling why the chemical interaction between FeOOH and phosphate is stronger than between FeOOH and arsenate, as we observed in our competitive

1216

M. Ding et al.

Fig. 6. O(1s) core level shift as a function of charge variation in the valence band (de Jong, 1989). The observed range of values for metal oxides varies between 533.6 eV {O(nonbridging) in P2O5; Ca(NO3)2} and 528.3 eV(CaO) corresponding to an oxygen charge of ⫺0.66 and ⫺0.99 esu, respectively (Moulder et al., 1992).

aquatic chemical experiments discussed elsewhere (Ding et al., 1998b). To rationalise these differences in chemical affinity from valence band spectroscopy is difficult, not in the least because spectral assignments lack consensus as to which atomic orbitals participate in the molecular orbital that consitutes the top of the valence band. Despite this uncertainty we shall suggest some rationalisations by considering the frontier orbitals between substrate and adsorbate. Frontier orbitals are important in the study of interactions between two molecules because they determine the character of charge transfer and thus, the nature of the bonding. Two factors characterise this interaction: the difference in energy between

Table 1. Oxygen (1s) binding energies and oxygen charges for FeOOH and its surface complexes. Binding energy (eV)

Oxygen charge

Compounds

O

OH

O

OH

Charge transfer relative to O in FeOOH

FeOOH FeOOH-As FeOOH-P FeOOH-Cr FeOOH-Pb FeOOH-Zn

530.3 528.4 528.9 532.1 530.4 529.9

531.7 529.3 529.3 532.7 — 531.3

⫺0.859 ⫺0.982 ⫺0.949 ⫺0.747 ⫺0.853 ⫺0.885

⫺0.772 ⫺0.924 ⫺0.924 ⫺0.710 — ⫺0.797

Accepts electrons Accepts electrons Donates electrons No transfer No transfer

XPS characterized electronic bonds between solid FeOOH and solutes

1217

Table 2. The relation between Fe core level binding energy and its charge. Binding energy (eV)

Calculated charge

Compounds

Fe(2p)

Fe(3p)

Fe(3s)

O(1s)

Iron

Oxygen

Equationsa

Fe FeO Fe2O3

706.9 709.5 711.0

53.0 54.9 55.7

90.9 92.5 93.6

530 530

0 ⫹0.8786 ⫹1.3179

⫺0.8786 ⫺0.8786

QFe(2p) ⫽ 0.3233X ⫺ 228.51, r2 ⫽ 0.999 QFe(3p) ⫽ 0.4835X ⫺ 25.632, r2 ⫽ 0.998 QFe(3s) ⫽ 0.4925X ⫺ 44.744, r2 ⫽ 0.993

a

QFe(i) is the charge of Fe with respect to (i) core level, X is the binding energy of Fe(i) core level, r is the correlation coefficient.

the highest occupied and lowest unoccupied state between substrate and adduct and the overlap between these molecular orbitals on substrate and adduct as shown in Figure 8, according to Hoffmann (1988). Qualitative analysis of the orbitals and their interaction can be done by simple perturbation molecular orbital theory. This theory predicts that for two interacting orbitals of suitable symmetry and overlap, the smaller the difference in orbital energies, the greater the mutual perturbation or interaction. The interaction between two systems involved in MO interaction is described by the expression (Hoffmann, 1988): ⌬E ⫽

兩H eg兩 2 Ee ⫺ Eg

where, ⌬E is potential energy of the combined system, Ee and Eg are the excited (empty) acceptor state and filled ground donor state energy, respectively, and Heg is the strength of the perturbation that has a principal term the overlap integral between HOMO and LUMO. This overlap connects frontier orbital theory with Pearson’s hard and soft acid– base concepts (e.g., Fleming, 1976). Clearly, for a specific perturbation strength, the greater the energy separation of the two systems, the smaller the potential energy of the compound. Conversely, the smaller the energy level separation, the larger the potential energy of the compound. As XPS does not measure the LUMO states, the difference between HOMO and LUMO has to be inferred from the shift in energy of the top of the valence band after adduct formation. As operational notion we shall use that similarity in orbital energy combined with substantial orbital overlap, that is, the soft–soft interaction in Pearson’s acid– base theory will cause a shift in the top of the valence band toward more negative energies. When, on the other hand, orbital energies differences are large and in the absence of substantial overlap, that is, the hard– hard interaction, no, or a shift to less negative energy in valence band energy will occur. The highest occupied MOs in FeOOH are primarily non-

bonding O(2p)-like states (Sherman, 1985), thus differing from the full 2t2g iron three-dimensional crystal field orbitals that most likely make up the valence band edge of hematite (Zhang et al., 1993), although spectral assignments still remain open to debate (e.g., Welsh and Sherwood, 1989; Ma et al., 1993; Catti et al., 1995; Schedel-Niedrig et al., 1995). Inspection of the valence band photoelectron spectra (Fig. 5) of FeOOH before and after adsorption shows that the patterns of FeOOH–As and FeOOH–P are similar to that of FeOOH. The top of the valence band of FeOOH–As is shifted by about 3 eV to more negative binding energies relative to that for FeOOH and FeOOH–P. This trend is clearly not an indication of attachment strength because phosphate is preferred over arsenate in FeOOH adsorbtion according to our competitive experiments. Therefore, it may be that this shift is attributable to a difference in hydrogen bonding to the FeOOH substrate as argued by Morrison (1984). Chromate adsorbtion affects the nature of the valence band spectrum substantially more than arsenate and phosphate. The top of the valence band is shifted by about 2 eV to less negative binding energy relative to FeOOH, whereas the resolution of the peaks is substantially higher. This spectroscopic variation suggests that, next to charge transfer as deduced from the core level variations, bonds are broken and made that might indicate an inner sphere reaction. Similarly to chromate, zinc and lead adsorbtion on FeOOH show either a shift by about 2 eV to less negative binding energies, suggesting again the formation of an inner sphere complex. We conclude that the energy level separations between HOMO/LUMO in FeOOH and phosphate and arsenate must have been rather small so that after orbital interaction no large spectral variation occurs. Consequently, we expect the potential energy of the FeOOH–P and FeOOH–As surface complexes to be high and therefore, chemically difficult to dissociate. In contrast to phosphate and arsenate, the energy level separations between HOMO and LUMO in FeOOH–Cr, FeOOH–Pb, and

Table 3. Fe core level binding energies and Fe charges for FeOOH and its surface complexes. Fe core level energy (eV)

Fe charge

Compounds

Fe(2p3)

Fe (3p)

Fe (3s)

Fe(2p)

Fe(3p)

Fe(3s)

Aver.

FeOOH FeOOH-As FeOOH-P FeOOH-Cr FeOOH-Zn FeOOH-Pb

711.7 716.9 710.3 713.6 710.6 709.6

56.3 59.7 55.2 58.8 55.8 55.5

94.3 97.9 93.2 96.6 93.0 93.9

1.58 3.26 1.13 2.20 1.23 0.90

1.59 3.23 1.06 2.80 1.35 1.20

1.70 3.47 1.16 2.83 1.06 1.50

1.62 3.32 1.12 2.61 1.21 1.20

Charge transfser relative to Fe in FeOOH Fe Fe Fe Fe Fe

donates electrons accepts electrons donates electrons accepts electrons accepts electrons

1218

M. Ding et al.

Fig. 7. Synopsis of charge transfer across the oxide surface in FeOOH for the cations and anions considered in this study.

FeOOH–Zn cases must have been large, as reflected in the large spectral variation. Following the same argument, we therefore predict that chemical dissociation of FeOOH–Cr, FeOOH–Pb, and FeOOH–Zn is easier than that of FeOOH–P and FeOOH– As. These results agree well with our aquatic chemical experiments, where FeOOH–As and FeOOH–P complexes remain stable at pH values far above the point of zero charge, whereas the release or adsorption of cation or anion on FeOOH–Cr, FeOOH–Zn, and FeOOH–Pb is directly related to the point of zero charge. 5. SUMMARY AND CONCLUSIONS

Our XPS results show that the FeOOH surface can act as a Lewis acid or base in its interaction with different adatoms and admolecules. Thus, phosphate and arsenate oxyanions donate electrons to the substrate that acts as a Lewis acid. The FeOOH surface functions as a Lewis base in its interaction with chromate. The adsorption between the FeOOH surface and zinc or lead cations does not involve noticeable electron transfer. Our valence band XPS spectral results enable us to rationalise the relative affinity of absorption on FeOOH substrate, which in decreasing order goes from P, As, Cr, Zn, to Pb. Our research indicates the feasibility to probe with XPS the solid–solute

Fig. 8. Principles of HOMO–LUMO interaction according to Hoffmann (1988).

interface and to reveal the nature of chemical reactions in terms of charge and energy transfer between frontier orbitals. Qualitatively the energy level separation can be estimated between HOMO and LUMO of the frontier orbitals for surface complexes. The energy difference between the HOMO and LUMO in the FeOOH–P complex and the FeOOH–As complex are small resulting in a substantial potential molecular energy. This implies strong chemical bonding between FeOOH and phosphate and arsenate. However, the energy differences between HOMO and LUMO in FeOOH–Cr, FeOOH–Pb, and FeOOH–Zn are large. Therefore, the potential molecular energies of the FeOOH–Cr, FeOOH–Zn, and FeOOH–Pb complexes are small, resulting in a weaker chemical bond formation in these complexes. Our aquatic chemistry experiments (Ding et al., 1998b) provide evidence consistent with these theoretical consideration. Acknowledgment—I thank the Faculty of Earth Sciences at Utrecht University for providing an “AIO” fellowship to carry out my Ph.D. study, in particular Professor Dr. R. D. Schuiling for his support of this research. REFERENCES Bagus P. S. and Bauschlicher C. W. (1980) Core binding energy for free negative ions of oxygen O0 to O2⫺. J. Electron Spectrosc. Relat. Phenom. 20, 183–190. Becker U., Hochella M. F. Jr., and David J. V. (1997) The adsorption of gold to galena surfaces: Calculation of adsorption/reduction energies, reaction mechanism, XPS spectra, and STM images. Geochim. Cosmochim. Acta 61, 3565–3585. Brown G. E., Jr. (1990) Spectroscopic studies of chemisorption reaction mechanisms at oxide–water interface. In Mineral–Water Interface Geochemistry (ed. M. F. Hochella Jr. and A. F. White), Vol. 23, pp. 309 –364. Reviews in Mineralogy. Brundle C. R., Chuang T. J., and Wandelt K. (1977) Core and valence level photoemission studies of iron oxide surfaces and the oxidation of iron. Surf. Sci. 68, 459 – 468. Burkhardt D. J. M., de Jong B. H. W. S., Meyer A. J. H. M., and van Lenthe J. H. (1991) H6Si2O7: ab initio molecular orbital calculations show two geometric conformations. Geochim. Cosmochim. Acta 55, 3453–3458. Catti M, Valerio G., and Dovesi R. (1995) Theoretical study of electronic, magnetic, and structural properties of ␣-Fe2O3 (hematite). Phys. Rev. B 51, 7441–7450. Cornell R. M. and Schwertmann U. (1996) The Iron Oxides, Structure, Properties, Reactions, Occurrence and Uses. VCH. de Boer C. B. (1999) Rock Magnetic Studies on Hematite, Maghemite and Combustion-Metamorphic Rocks. Geologica Ultratraiectina. Mededelingen van de Faculteit Aardwetenschappen Universiteit Utrecht, #177.

XPS characterized electronic bonds between solid FeOOH and solutes de Jong B. H. W. S. (1989) Glass: Ullmann’s Encycl. Ind. Chem. A 12, 365– 432. de Jong B. H. W. S., Ellerbroek D., and Spek A. L. (1994) Lowtemperature structure of lithium nesosilicate, Li4SiO4, and its Li1s and O1s X-ray photoelectron spectrum. Acta Cryst. B 50, 511–518. de Jong B. H. W. S. and Ding M. (2000) XPS derived electronegativities for oxide materials. Abstract ACS, San Francisco Meeting, March 26 –30, USA. Ding M. (1998) Self-sealing/Healing Isolation and Immobilization Caused by Chemical Discontinuities in Porous Media. Geologica Ultraiectina. Mededelingen van de Faculteit Aardwetenschappen Universiteit Utrecht, #164. Ding M., van der Sloot H. A., and Geusebroek M. (1996) Interface reaction affects Fe3⫹ mobility in jarosite and alkaline coal fly ash. Ceram. Trans. 72, 475– 484. Ding M., van der Sloot H. A., and Geusebroek M. (1997) Interaction of Fe3⫹, Ca2⫹, and SO2⫺ 4 at the jarosite/alkaline coal fly ash interface. Proc. 30th Intern. Geol. Congr. 19, 207–222. Ding M., Geusebroek M., and van der Sloot H. A. (1998a) Interaction precipitation affects the resistance to transport in layered jarosite/fly ash. J. Geochem. Expl. 62, 319 –323. Ding M., de Jong B. H. W. S., Niessen B., Roosendaal S. J., van der Meer J. P. M., and van Hinsberg, V. J. (1998b) Characterization of amorphous black ferric oxyhydroxide: X-ray photoelectron spectroscopy, conversion electron Mo¨ssbauer spectroscopy, and chemical adsorption. (submitted). Dutrizac J. E. and Harris G. B., eds. (1996) Iron Control and Disposal: Part 8. Disposal Practices for Iron Residues. Proc. 2nd Intern. Symp. Iron Control in Hydrometallurgy. Can. Inst. Mining, Metallurgy and Petroleum. Quebec. Dzombak D. A. and Morel F. M. M. (1990) Surface Complexation Modelling: Hydrous Ferric Oxide. John Wiley & Sons. Fleming I. (1976) Frontier Orbitals and Organic Chemical Reactions. Wiley. Fukui K., Yonezawa T., and Shingu H. (1954) Molecular orbital theory of orientation in aromatic, heteroaromatic, and other conjugated molecules. J. Chem. Phys. 22, 1433–1442. Graat P. (1998) The initial oxidation of iron and iron nitride. Ph.D. Thesis, Delft University. Hoffmann R. (1988) Solids and Surfaces: A Chemist’s View of Bonding in Extended Structures. VCH. Huheey J. E., Keiter E. A., and Keiter R. L. (1993) Inorganic Chemistry, 4th ed. HarperCollins. Klopman G. (1968) Chemical reactivity and the concept of charge and frontier controlled reactions. J. Am. Chem. Soc. 90, 223–234. Klopman G. (1974) The generalized perturbation theory of chemical reactivity and its applications. In Chemical Reactivity and Reaction Paths (ed. G. Klopman), pp. 369. Wiley. Ma Y., Johnson P. D., Wassdahl N., Guo J., Skytt P., Nordgren J., Kevan S. D., Rubensson J.-E., Boeske T., and Eberhardt W. (1993) Electronic structure of ␣-Fe2O3 and Fe3O4 from O K-edge absorption and emission spectroscopy. Phys. Rev. B 48, 2109 –2111. Martin R. R. and Smart R. St. C. (1987) X-ray photoelectron studies of anion adsorption on goethite. Soil Sci. Soc. Am. J. 51, 54 –56.

1219

Mclntyre N. S. and Zeatruk D. G. (1977) X-ray photoelectron spectroscopic studies of iron oxides. Anal. Chem. 49, 1521–1529. Moulder J. F., Stickle W. F., Sobol P. E., and Bomben K. D. (1992) Handbook of X-ray Photoelectron Spectroscopy. Perkin-Elmer Corporation Publ. Morrison W. H. (1984) Aqueous adsorption of anions onto oxides at pH above the point of zero charge. J. Coll. Interf. Sci. 100, 121–127. Roosendaal S. J., Giebels L. A. M. E., Vredenberg A. M., and Habraken F. H. P. M. (1998) Determination of photoelectron attenuation lengths in thin oxide films on iron surfaces using quantitative XPS and elastic recoil detection. Surf. Interf. Anal. 26, 758 –765. Rosso K. M., Becker U., and Hochella M. F. (1999a) Atomically resolved electronic structure of pyrite {100} surfaces: An experimental and theoretical investigation with implications for reactivity. Am. Min. 84, 1535–1548. Rosso K. M., Becker U., and Hochella M. F. (1999b) The interaction of pyrite {100} surfaces with O2 and H2O: Fundamental oxidation mechanisms. Am. Min. 84, 1549 –1561. Sanderson R. T. (1983) Polar Covalence. Academic Press. Schedel-Niedrig Th., Weiss W., and Schloegl R. (1995) Electronic structure of ultra thin ordered iron oxide films grown onto Pt(111). Phys. Rev. B 52, 17449 –17460. Seah M. P. (1990) Charge referencing techniques for insulators. In Practical Surface Analysis, 2nd ed. (ed. D. Briggs and M. P. Seah), pp. 79. Wiley. Sherman D. M. (1985) SCF-X␣-SW MO study of Fe3⫹ coordination sites in iron oxides: Applications to spectra, bonding and magnetism. Phys. Chem. Min. 12, 311–314. Siegbahn K., Nordling C. N., Fahlman A., Nordberg R., Hamrin K., Nedman J., Johnsson G., Bermark T., Karlsson S. E., and Lindberg B. (1967) ESCA: Atomic, Molecular and Solid State Structure Studied by Means of Electron Spectroscopy. Almqvist and Wiksells. Voigt J. A., Wood T. E., Bunker B. C., Casey W. H., and Crossey L. J. eds. (1997) Aqueous chemistry and geochemistry of oxides, oxyhydroxides, and related materials. Mater. Res. Symp. Proc. 432. Waychunas G. A., Davis J. A., and Fuller C. C. (1995) Geometry of sorbed arsenate on ferrihydrite and crystalline FeOOH: Re-evaluation of EXAFS results and topological factors in predicting sorbate geometry, and evidence for monodentate complexes. Geochim. Cosmochim. Acta 59, 3655–3661. Waychunas G. A., Fuller C. C., Rea B. A., and Davis J. A. (1996) Wide angle X-ray scattering (WAXS) study of “two-line” ferrihydrite structure: Effect of arsenate sorption and counterion variation and comparison with EXAFS results. Geochim. Cosmochim. Acta 60, 1765–1781. Welsh I. D. and Sherwood P. M. A. (1989) Photoemission and electronic structure of FeOOH: Distinguishing between oxide and oxyhydroxide. Phys. Rev. B 40, 6386 – 6392. Zhang Z., Boxall C., and Kelsall G. H. (1993) Photoelectrophoresis of colloidal iron oxides 1. Hematite (␣-Fe2O3). Colloids Surf. A 73, 145–163.