Magnetite crystallization through oxidation and bioreduction: Processes for iron recovery from groundwater

Yvonne M. Mos

Research output: Thesisinternal PhD, WU


Most drinking water worldwide is produced from groundwater. Groundwater contains up to 1.8 mM aqueous ferrous iron. Since the recommended maximum iron concentration in drinking water is 5.10-3 mM (WHO), iron is removed in the purification process. The most common method for this is uncontrolled chemical iron oxidation, leading to precipitation of ferric iron oxides, followed by rapid sand filtration. This is an effective method, but produces large volumes of poorly dewaterable sludge with a low value. By recovering this iron in the form of magnetite, the value and application potential could be increased. Magnetite (Fe2+Fe3+2O4­)  is a crystalline and compact mixed valence iron oxide with strong magnetic properties and therefore a multitude of high-end applications including magnetic data storage and site-specific drug delivery. This thesis describes two processes for the crystallization of magnetite from aqueous ferrous iron and hydrous ferric oxide.

Research on crystalline materials often uses X-Ray Diffraction for phase identification. In this analytical technique, the sample is hit with X-rays that are scattered when they hit an electron. There are two types of scattering; elastic scattering yields a phase specific diffraction pattern, while inelastic scattering leads to a background noise. The use of copper radiation for the analysis of samples containing iron generally leads to a high background signal. This can be prevented by using cobalt radiation or take physical measures to decrease the detected level of background noise. Many examples exist in literature where diffractograms with a high background noise lead to ambiguous phase identification.

Crystallization of magnetite was achieved through the partial chemical oxidation of aqueous Fe2+ in a continuous stirred tank reactor. The reactor was fed with a medium of 50% tap water and 50% Milli-Q water. The reactor was operated at pH 5.5 or 6.0 for 24 hours, after which the pH was increased to 6.8 – 7.5, initiating magnetite crystallization. Alternatively, seeding the reactor with previously formed magnetite and operating it at a constant pH of 7.0 also yielded magnetite. Magnetite formation did not take place at a pH below 6.8 or when the reactor was operated at a constant pH without seeding. Color changes of the medium in the reactor indicate that magnetite formation took place via a green rust (GR) precursor phase. During the aforementioned experiments, nitrate was added as electron acceptor. However, mass balance calculations revealed that O­2, diffused into the reactor setup, also served as electron acceptor. This was confirmed in an experiment in which nitrate was omitted from the medium, which yielded magnetite.

The reactor was placed in a glovebox filled with N2 to eliminate the O2 diffusion. A controlled O2 flow was added to the reactor headspace in the form of compressed air. Magnetite crystallization was achieved in 100% Milli-Q water at an iron feed concentration equal to or above 1 mM. An iron feed concentration of 2.25 mM and a hydraulic retention time (HRT) of 34 hours yielded 69% magnetite. Reducing the HRT to 8 hours yielded 18% magnetite, while the fraction of side products, mainly lepidocrocite (γ-FeOOH), increased. This indicates an incomplete transformation of chloride-GR into magnetite at an HRT of 8 h. The main factor determining the iron phase formation however, was the Fe2+ concentration in the reactor. Magnetite crystallization took place at a concentration equal to or above 0.4 mM Fe2+ in the reactor.

Increasing medium complexity by using diluted tap water or growth medium, hindered magnetite crystallization. An experiment with medium consisting of 50% tap water and 50% Milli-Q water with the controlled addition of O2 to the headspace yielded lepidocrocite instead of magnetite, indicating the inhibition of the transformation of chloride-GR into magnetite. Increasing the percentage of tap water to 90% (and 10% Milli-Q water) led to the formation of goethite (α-FeOOH). This indicates the formation of carbonate-GR, that was oxidized by ambient air during sample preparation for XRD analysis. The formation of carbonate-GR was induced by the presence of carbonates in the tap water. The use of growth medium in an experiment without inoculation of microbial biomass led to the formation of lepidocrocite and goethite, while the experiment that was inoculated with activated sludge yielded vivianite (Fe3(PO4)­2). The latter was likely induced by the presence of phosphate and possibly vivianite in the added sludge.

Magnetite crystallization was also achieved by partial biological reduction of hydrous ferric oxide (HFO). Batch experiments were conducted in growth medium with acetate, lactate, glucose or ethanol as electron donor. While experiments with acetate did not yield magnetite, the crystallization of magnetite in the remainder of the batch experiments was independent of the electron donor. The main factors determining the iron phase formation in these experiments were the final pH and redox potential (ORP). Magnetite was formed when the final pH and ORP were in the range of 7.5 to 7.6 and -254 to -283 mV. Like in the partial oxidation process, observation of medium color changes indicated that magnetite crystallization took place through a GR precursor phase. At an ORP below -283 mV, the increased Fe2+ concentration stabilized the formed GR, thereby inhibiting the transformation into magnetite. The GR was oxidized to lepidocrocite by ambient air during sample preparation for XRD analysis.

An attempt to reproduce magnetite crystallization through partial bioreduction of HFO in a continuous reactor setup failed. The ethanol (electron donor) flow to the reactor was varied to induce various ORP conditions in the reactor, mimicking the final conditions in the batch experiments. However, all settings that induced reduction of HFO led to the formation of GR. Spiking the reactor containing the formed GR with a 0.5 M aqueous Fe3+ solution did lead to the formation of magnetite.

Magnetite was produced continuously through the partial oxidation process. The partial bioreduction process shows that magnetite crystallization is possible in the presence of ‘foreign’ ions. Both processes take place at circumneutral pH and require only small volumes of added chemicals. However, additional research is required to develop a process that is both continuous and applicable in a complex matrix and that is suitable for low iron concentrations. Therefore, a combination of both processes is proposed. Here, drinking water treatment sludge is partially bioreduced to form GR in a continuous reactor. This is then moved to a second reactor, where the controlled addition of O2 leads to the transformation of GR to magnetite.

Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • Wageningen University
  • Buisman, Cees, Promotor
  • Weijma, Jan, Co-promotor
Award date19 Oct 2018
Place of PublicationWageningen
Print ISBNs9789463434652
Publication statusPublished - 2018


Dive into the research topics of 'Magnetite crystallization through oxidation and bioreduction: Processes for iron recovery from groundwater'. Together they form a unique fingerprint.

Cite this