Goethite (α-FeOOH) was synthesized following Böhm's method as described in Schwertmann and Cornell 32. Briefly, 100 mL of freshly prepared 1 M Fe(NO₃)₃·9H₂O were mixed rapidly with 180 mL of 5 M KOH in a polypropylene beaker. All solutions used in this study were prepared with high-purity water (Milli-Q, Millipore, 18 MΩcm). The precipitated hydrous ferric iron was immediately diluted to 2 L with deionized water and heated in closed polypropylene flasks at 70°C for 60 hours. Afterwards, the precipitate was resuspended in high-purity water, centrifuged, and decanted. This washing procedure was repeated several times until the supernatant was free of chloride and then freeze-dried. X-ray powder diffraction confirmed the formation of α-FeOOH. The specific surface area was determined as 38 m²g^-1 by a multipoint N₂-BET adsorption method (surface area analyzer Gemini 2360, Micromeritics, Belgium).

Prior to the synthesis of DFOD, N-acetyl-O,O,O-triacetyl-desferrioxamin was synthesized as follows. DFOB (6.4 g, 10 mmol) and ground anhydrous K₂CO₃ (3 g, 20 mmol) were mixed with H₂O-free N,N-dimethylformamide (25 mL). This suspension was heated to 70°C under nitrogen atmosphere for 30 min, then cooled to room temperature and diluted with CH₂Cl₂ (100 mL). To this suspension, a solution of acetic anhydride (4.1 g, 40 mmol) in 20 mL CH₂Cl₂ was slowly added over a period of 30 min. The reaction mixture was stirred for 4 h and filtered. The filtrate was successively extracted with water (2 × 50 mL), saturated aqueous NaHCO₃ solution (50 mL) and aqueous sodium chloride solution (50 mL) and then dried over anhydrous Na₂SO₄. The solvents were removed in vacuo, and the N-acetyl-O,O,O-triacetyl-desferrioxamine was obtained as colorless thick oil (yield 6.7 g, 9.2 mmol, which correspond to 92% of the theoretical obtainable amount). Then, the N-acetyl-O,O,O-triacetyl-desferrioxamine was dissolved in 100 mL of methanol, cooled to 0°C with an ice bath, and the solution was saturated with gaseous ammonia. The solution was kept at room temperature for 4 h, and then refrigerated overnight. Crude DFOD was collected by filtration and re-crystallized twice from methanol/water for purification. Pure DFOD was obtained as colorless crystalline material. 4.8 g of DFOD were obtained corresponding to 84.7% of the theoretical obtainable amount. The melting point was 180 ± 1°C. The relative composition of elements for C₂₇H₅₀N₆O₉ was found to be: C, 53.93 (53.84); H, 8.44 (8.32); N, 13.76 (13.95)% (theoretical values are given in parenthesis). Detailed information on the characterization of DFOD is provided in additional file 1.

The adsorption of EDTA (Fluka, Switzerland) and the siderophore DFOD to goethite as a function of increasing SDS (Fluka, Switzerland) concentrations was studied at pH 6 in 40 mL batch reactors. SDS concentrations were kept well below the critical micelle concentration (CMC = 3.1 mM at 25°C in 0.01 M NaCl; 33). 30 mL batches of goethite (α-FeOOH) suspensions (solids concentrations of 7 g/L; pH 6.0 ± 0.05; 0.01 M NaClO₄ (as ionic strength buffer)) containing SDS at various concentrations were prepared and pre-equilibrated for 68 hours on an end-over-end shaker at 25 ± 1°C. Then, 10 mL of stock solutions containing 320 μM EDTA, or DFOD in 0.01 M NaClO₄ adjusted to pH 6.0 ± 0.05 were added to the suspensions. The total concentrations of EDTA, or DFOD were 80 μM in all batches and the total SDS concentrations varied from 0 to 600 μM. The suspension pH was measured using a pH electrode (Metrohm, 6.0234.100) and readjusted, if necessary, to pH 6.0 ± 0.05 by small additions of NaOH or HCl. All reactors were wrapped in aluminum foil to avoid photochemical reactions. After 24 hours equilibration time on an end-over-end shaker at 25 ± 1°C, 5 mL aliquots of each suspension were filtered through 0.025 μm cellulose nitrate membranes using polycarbonate syringe filter holders (25 mm diameter, Schleicher & Schuell, Germany). The concentration of adsorbed ligands was determined indirectly by measuring the ligand concentration left in solution after filtration. DFOD was immediately analyzed after filtration with a UV-visible spectrophotometer (Cary 50 BIO, Varian, Australia) as an iron complex in the presence of an excess of Fe(III) at 432 nm as described in 34. EDTA was also analyzed spectrophotometrically as iron complex following a procedure described by Bhattach et al. 35. Blank experiments without goethite were performed to monitor any losses due to adsorption of the ligands to the vessel or filter, but no significant losses were observed. During the 24 h equilibration time after the siderophore addition, some ligand-promoted dissolution of goethite may have occurred. However, based on the low dissolution rates of goethite at pH 6 in the presence of EDTA 36 and siderophores 37, the expected dissolved iron concentrations are in the low micromolar range and are not expected to affect the results.

Stock solutions of 1:1 Fe-EDTA, Fe-DFOB and Fe-DFOD complexes (120 μM in 0.01 M NaClO₄) were prepared by dissolving Fe(NO₃)₃·9H₂O (Fluka, Switzerland) in a 0.01 M NaClO₄ solution and adding EDTA, DFOB, or DFOD solutions at equimolar concentrations. The stock solutions were added to goethite suspensions pre-equilibrated with SDS at pH 6.0 ± 0.05 at 25 ± 1°C. The suspension pH was measured and readjusted if necessary to 6.0 ± 0.05 by small additions of NaOH or HCl. The experimental set up was similar to the adsorption experiments with free ligands as described above, except that the solids concentration was 2.5 g/L. The total concentration of Fe complexes in all reactors was 30 μM and the total SDS concentration was 0–600 μM. The equilibration time after the addition of the Fe complexes was 24 h. After filtration, the dissolved concentrations of iron complexes left in solution were immediately analyzed by inductively coupled plasma optical emission spectrometry (ICP-OES, Vista-MPX, CCD Simultaneous, Varian, Australia) at the wavelength 238.2 nm. In some experiments, the concentrations of 1:1 Fe complexes were also analyzed by UV-visible spectrophotometry. The lack of difference between these two methods indicated the formation of 1:1 Fe-ligand complexes (see additional file 2). In addition, Fe-DFOB adsorption isotherms were recorded using a similar set-up in the presence of 0, 15, and 800 μM total SDS, respectively. Samples were taken 24 hours after the addition of the iron complex, filtered, and iron concentrations were analyzed by ICP-OES.

The influence of increasing SDS concentrations on the adsorption of EDTA and the siderophores DFOB (data from 21) and DFOD is shown in Figure 2. SDS had almost no effect on the adsorption of EDTA, except for a slight decrease of EDTA adsorption at the highest SDS concentration (600 μM). Considering the hydrophilic nature of EDTA, hydrophobic interactions with co-adsorbed SDS are not expected. The doubly deprotonated H₂EDTA^2- (pK_a3 = 6.13) is the dominant species in solution in the absence of soluble iron at pH 6. Based on sorption and titration experiments combined with surface complexation modeling, Nowack and Sigg 38 proposed that the dominant surface species of EDTA at pH 6 is a triply deprotonated, singly charged binuclear inner-sphere surface complex (≡ Fe₂EDTA^-)

ATR-FTIR studies have conclusively shown that sulfate forms inner-sphere and outer-sphere surface complexes on goethite, depending on solution pH 39 At pH 6 and above, the dominant surface species was an outer-sphere surface complex, while at low pH sulfate was sorbed mainly as inner-sphere complex. Therefore, it seems likely that the sulfate head groups of SDS will also mainly form outer-sphere complexes at pH 6. Hence, little adsorption competition between SDS and EDTA is expected at low surfactant concentrations. Indeed, SDS had little effect on EDTA sorption at concentrations < 600 μM total SDS (Figure 2). Nevertheless, it is possible that adsorbed SDS forms a minor inner-sphere surface species via the sulfate head groups that may compete with EDTA for surface sites at the highest SDS concentration leading to the observed decrease of EDTA adsorption (Figure 2).

One could also assume that SDS may influence EDTA adsorption by modifying the charge at the mineral-water interface. Electrophoretic mobility measurements showed a reversal from positive to negative values upon formation of a surfactant bilayer on goethite 21. However, electrophoretic mobility measurements probe the electrical potential at the plane of shear, which is beyond the surfactant (bi)layer, i.e., at greater distance from the mineral surface. Under the assumption that EDTA forms only inner-sphere complexes and that the charge of adsorbed EDTA is located at or close to the surface its adsorption does not depend on variations of the potential above the hemimicelles and admicelles.

The surface potential depends, among other factors, on the protonation state of the surface and it has been observed that the adsorption of surfactants, especially at low concentrations, increases the protonation of oxide surfaces 4041. However, Nowack and Sigg 38 have reported that the adsorption of EDTA as a binuclear complex on goethite is constant between pH 5 and 7, which is consistent with our observation that possible SDS-induced changes of the surface protonation state at pH 6 were small enough not to influence the adsorption of EDTA. Considering Equation 1, the observation of constant EDTA adsorption at variable SDS concentrations suggests that SDS should have no effect on EDTA promoted dissolution rates. However, SDS may affect the formation of inner-sphere surface complexes, and hence, the surface speciation of sorbed EDTA, which may potentially affect dissolution rates (vide supra).

While adsorbed SDS had little influence on EDTA adsorption, a strong effect on DFOB and DFOD adsorption was observed (Figure 2). The adsorption of DFOB and DFOD increased strongly with increasing SDS concentrations. Despite their difference in molecular charge, the effects of the anionic surfactant on the adsorption of DFOB⁺ and DFOD⁰ were similar.

Cocozza et al. 16 found that DFOB and DFOD form mononuclear inner-sphere surface complexes with a single hydroxamate group coordinating with a surface site. The adsorption of DFOD involves no net change of surface charge in analogy to the adsorption of the monohydroxamate ligand acetohydroxamic acid 17. The adsorption of DFOB involves the increase of the surface charge due to the positive amine group, probably located relatively far from the goethite surface due to repulsive forces. The difference in charge of these two surface complexes should result in differences in the electrostatic interactions, which is consistent with the observation of somewhat higher adsorption densities of DFOD in the absence of surfactants under otherwise identical conditions 34.

The structures of the siderophores (Figure 1) include a pendant alkyl chain with five carbons giving rise to a local hydrophobic character of the molecule. Bonilha et al. 42 observed that the exchange constant between Na⁺ and alkylammonium ions (R-NH₃⁺, with R = C_nH_2n+1) in SDS micelles increases with increasing chain size, indicating that hydrophobic interactions contribute to the enhanced exchange constants. Therefore, increased siderophore adsorption in the presence of co-adsorbed SDS suggests that hydrophobic interactions between ligands and surfactants are responsible for the observed behavior. The observed effect of SDS on the adsorption of the two siderophores was similar despite their difference in molecular charge. These observations imply that co-adsorbed surfactants have little influence on electrostatic free energy contributions to siderophore adsorption. This is consistent with our interpretation of the absence of an effect of surfactants on EDTA adsorption as discussed above.

Adsorption of 1:1 Fe(III)-complexes was quantified by measuring the loss of Fe-complexes from solution by UV spectrophotometry and the loss of total Fe by ICP-OES. Both methods gave identical results (see additional file 2), confirming the adsorption of 1:1 Fe-ligand complex. Adsorption equilibrium of Fe-DFOB on goethite was reached after 24 h (see additional file 2). SDS had strong effects on the adsorption of the complexes of DFOB, DFOD, and EDTA (Figure 3a).

The adsorption of the Fe-EDTA complex decreased by almost 75% at the highest SDS concentration relative to adsorption in the absence of SDS. The adsorption behavior of free ligands compared to their iron complexes in the presence of surfactants can be interpreted in terms of their different adsorption modes. In solution, the dominant Fe-EDTA complex is a quinquedentate seven-coordinated bisaquo Fe(III)EDTA^-·(H₂O)₂ with a free carboxylic group 4344. Nowack and Sigg 38 suggested that the dominant surface complex at pH values below 7 is an outer-sphere complex (≡FeOH₂⁺·LFe^-), while an inner-sphere complex was postulated to dominate at higher pH values (≡FeOFeL^2-). Therefore, the contrasting effect of SDS on the adsorption of EDTA and its iron complex at pH 6 can be interpreted in terms of their adsorption as inner-sphere and outer-sphere complex, respectively. While the adsorption of the Fe-EDTA complex decreased with increasing SDS concentration (Figure 3a), co-adsorbed surfactants had little effect on free EDTA adsorption (Figure 2). As discussed above, the charge of surfactant aggregates on the goethite surface appeared to have little effect on electrostatic contributions to EDTA adsorption. On the contrary, the Fe-EDTA outer-sphere complex is adsorbed mainly by electrostatic interactions and its plane of adsorption can be expected at greater distance from the mineral surface. Therefore, decreasing adsorption of the negatively charged Fe-EDTA outer-sphere complex with increasing SDS adsorption suggests that electrostatic repulsion between admicelles and the iron complex decrease its adsorption. Nevertheless, some adsorption still occurred at SDS concentrations where electrophoretic mobility measurements indicated charge reversal 21. This behavior is consistent with the formation of inner-sphere surface complexes (Equation 4) as a minor surface species 38.

In contrast, the adsorption of Fe-DFOB and Fe-DFOD increased over the same range of surfactant concentrations (Figure 3a). The adsorption of Fe-DFOD was higher than that of Fe-DFOB at low surfactant concentrations. At high SDS concentrations, higher adsorption of Fe-DFOB than that of Fe-DFOD was observed. Figure 3b shows the influence of low and high SDS concentrations on the adsorption isotherms of Fe-DFOB on goethite at pH 6. Again, increasing surface concentrations of the Fe-DFOB complex were observed with increasing surfactant concentration. Over a wide pH range (pH 2–11), including the experimental pH of this study, the charge of the iron complexes of the siderophores Fe-DFOB⁺ and Fe-DFOD⁰ is the same as the charge of free siderophores 45. The formation of 6 five-membered rings completely satisfies the coordinative requirements of Fe(III) resulting in extremely stable complexes. Therefore, it is unlikely that the chelated iron adsorbs as inner-sphere surface complex and Fe-DFOB⁺ and Fe-DFOD⁰ are expected to interact with the surfactant-coated goethite surface by electrostatic or hydrophobic interactions as outer-sphere surface complexes. Increasing adsorption of both complexes with increasing surfactant concentrations again underlines the importance of the hydrophobic interactions between the hydrophobic backbone of siderophores and the surfactant tail. However, the effect of hemimicelle/admicelle formation on Fe-DFOB⁺ adsorption was stronger than the effect on Fe-DFOD⁰ adsorption (higher adsorption of Fe-DFOB⁺, Figure 3a). In contrast to the neutral Fe-DFOD⁰, the adsorption of Fe-DFOB⁺ benefits not only from hydrophobic interactions but also from electrostatic interactions with the negatively charged admicelle. The positively charged Fe-DFOB⁺ experiences electrostatic repulsion resulting in low surface concentrations compared to Fe-DFOD⁰ at small surfactant concentrations. At higher surfactant concentrations (< 100 μM total SDS), Fe-DFOB⁺ experiences attractive forces due to charge reversal (see electrophoretic mobility in 21) caused by the formation of hemimicelle and admicelle and its adsorption exceeded Fe-DFOD⁰ adsorption. This electrostatic effect is not seen in the adsorption of free siderophores forming mainly inner sphere complexes resulting in the location of the charge of the surface complex closer to the mineral surface as discussed above.

Table 1 provides the adsorbed concentrations of SDS and DFOB to goethite at pH 6 as influenced by total SDS concentration in suspension. The ratio Δ(DFOB_ads)/ΔSDS_ads) indicates how much additional DFOB was adsorbed with additional adsorption of SDS. These values are only rough estimates, because they are based on SDS adsorption data determined in the absence of the siderophore 21, which was found to have a minor effect on SDS adsorption in preliminary experiments. The strong decrease of the ratio Δ(DFOB_ads)/Δ(SDS_ads) with increasing SDS concentration suggests that the effect of adsorbed SDS on the adsorption of DFOB was largest at low SDS surface coverage. Similar differences in ratios were observed for the co-adsorption of fluorocarbon alcohols 46 and 1-pentanol 47 in the presence of adsorbed surfactants on mineral surfaces. These alcohols possess hydrophobic and charged structural moieties that contribute to electrostatic and hydrophobic interactions with adsorbed surfactants, in analogy to the mechanism of siderophore adsorption discussed above. Lai et al. 46 and Lee et al. 47 suggested that low amounts of adsorbed surfactant create two different regions to which amphiphilic compounds can co-adsorb. One region is located between the head groups of the surfactants, and the second region is the hydrophobic perimeter arising from patchwise adsorption of bilayered surfactant aggregates mainly present at low surfactant concentrations 47. They suggested that the fraction of alcohol adsorbed at the perimeter can be large only at low surfactant concentrations when the aggregates are small and sparse on the mineral surface. Similarly, Asvapathanagul et al. 48, suggested that the structure, nature, or arrangement of adsorbed surfactants at low concentrations had a much stronger effect on co-adsorption of solutes than the actual adsorbed amount. The high Δ(DFOB_ads)/Δ(SDS_ads) ratios observed at low SDS loadings suggest that DFOB adsorbs in the two different regions mentioned above.