CHARACTERIZATION OF SEDIMENTS FROM THE FUNDÃO DAM FAILURE

The accident with the Fundão dam, Mariana, MG, launched tons of iron ore tailings into the Rio Doce river system. Despite the natural composition of the river sediments, characterized by relatively common mineral phases such as quartz, muscovite, kaolinite and illite and even certain hematite and goethite contents, it is important verify the abundance of hematite and magnetite after the accident. The Mariana city post-accident sediments were analyzed by Syncrotron X-ray diffraction, EDXRF and AC Susceptibility. It was verified a significant iron content, which is not founded in the natural sediments in the same region before the accident with the Fundão dam, Mariana, MG. Some crystalline phases detected are originated from processing processes in the mineral extraction industry, as their crystals are of synthetic origin (not found in nature).


INTRODUCTION
The accident with the Fundão dam, Mariana, MG, launched tons of iron ore tailings into the Rio Doce river system. After a few days, the ore contamination plume reached the mouth, reaching the adjacent marine region. In a recent study Almeida et al [1] performed the characterization of the iron ore from the accident and spreading in the Rio Doce basin. Despite the natural composition of the river sediments, characterized by relatively common mineral phases such as quartz, muscovite, kaolinite and illite and even certain hematite and goethite contents, it is important verify the abundance of hematite and magnetite after the accident. Taking this into account, our working group used three different techniques to verify the abundance of crystalline phases present in post-accident sediments.

MATERIALS AND METHODS
X-ray diffraction measurements of the samples were performed at the Sincrotron National Light Laboratory at the National Center for Energy and Materials (CNEM-LNLS) in Campinas -Brazil. The measurements were taken on line XRD1 (Proposals 20170383 and 20170108) using a wavelength of 0.103287 (2) nm. Figure 01 shows the beam line XRD-1. All samples in this work presented the same avarege particle size. The sediments were grounded and sieved down to particle sizes were smaller than 53 micrometers. Wavelength calibration was performed with the NIST silicon standard. Measurements were made by transmission using a transparent capillary 0.5mm in diameter.
Elemental analysis was carried out by an energy dispersive X-ray fluorescence spectrometer (EDXRF), model EDX-720 (Shimadzu, Japan). Metallic elements ranging from Na to U were scanned for detection and measurement. The scan was performed by two channels (Na-Sc and Ti-U). All samples (filters and bulk) were measured using a rhodium X-ray tube, 15-50 kV, 1000 (auto) µA, 500 seconds of integration by channel and a Si(Li) detector. However, samples onto cellulose filters were analysed under vacuum medium. Laboratory filter blanks were also analyzed to evaluate analytical bias. The quantification was performed by calibration curves made by analysis of 47 certified reference materials (CRM) ranging from 40.3 to 58.8 µg cm -2 deposited on thin mylar membranes (Micromatter, USA) ranging from low-Z to high-Z elements (Na to Rg).
Magnetic susceptibility measurements of the samples were made at the UFES High Pressure Laboratory (PRESLAB) using an AC system. A simplified scheme of this home assembly is shown in figure 02. The assembly presents the following components: one field coil to apply magnetic field, two secondary pick up coils and a lock-in amplifier. The ac generator creates a current in the field coil, which creates an alternating magnetic field in the two inner secondary coils. The secondary coils are identical, but have been installed with opposite winding directions so that the potential difference Vb in one of the coils is equal in modulus and the opposite direction of the potential difference Va of the other secondary coil. Thus, the resulting value read from the Lock-in must be null. However, if a magnetic material is placed inside one of the coils, the sum of the two potential differences will be nonzero. In this assembly the Lock-in is responsible for measuring this signal, the voltage difference Va -Vb. The setup accuracy is 1ppm. The outer coil (FC) generates an oscillating magnetic field given by: Thus, considering A the area of the sensor coils, we will have: (2) Thus, the voltages in the secondary coils are:

th WORKSHOP OF CRYSTALLOGRAPHY APPLIED TO MATERIALS SCIENCE AND ENGINEERING
Meaípe, Guarapari -ES, Brasil, May 2019 So, since the sensor coils are coiled in phase opposition, we have: We see that from a simple relationship between the tensions involved we obtain the susceptibility of the material. Dividing  by the density of the sample we have  m , magnetic susceptibility.

(7)
In this setup magnetic susceptibility is only calculated from measurements of variation of the coil voltage induced, however one knows that it depends on other parameters involved, such as frequency and temperature. Taken into account, the measuring setup also records the temperature and current values in the generator coil.
The readings of the measured parameters are archived on the computer by the same program that automates the susceptometer. This software was also developed by PRESLAB and allows you to generate a file to be read in any graphic program that you want to analyze or process data.
The mounting of this apparatus is ideal for temperature sensitive magnetic susceptibility measurements. Temperature variation cannot be controlled and is thus obtained by the natural evaporation of liquid nitrogen that cools the susceptometer (and the sample contained therein). The sample is cooled in a few minutes to reach liquid nitrogen temperature (around 77 K), but the heating phase is quite slow (due to the susceptometer engineering itself that stores nitrogen in an almost adiabatic compartment) and may take several hours Then, it is possible to measure the magnetic susceptibility of the sample from 77 K to room temperature. If necessary the sample can be heated to a temperature of 373 K.

th WORKSHOP OF CRYSTALLOGRAPHY APPLIED TO MATERIALS SCIENCE AND ENGINEERING
Meaípe, Guarapari -ES, Brasil, May 2019

RESULTS AND DISCUSSION
A representative measure of X-ray diffraction at LNLS -CNPEM station XRD1 of the sediment samples found in the Mariana city from the Fundão dam failure is shown in figure 03. Figure 03 -Synchrotron X-ray diffraction pattern of sediment from city Mariana.
It was used a PDF-2 program to identify the crystallographic phase in the representative Mariana sediment sample. The Table I indicate crystallographic phases candidates presents in this difratogram, taken into account the 90% as confidence threshold. The crystallography phase content is associated with the crystal volume present in the sample total volume analyzed, taken into account the X-ray penetration and the spot size. As considering the density of each crystalline phase one can verify that this result is in agreement with the elemental analysis, which as carried out by an energy dispersive X-ray fluorescence spectrometer (EDXRF).
The Table II shows the elements from EDXRF present in the same representative Mariana city sediment sample.  As one can see, there exists stronger iron content in the representative sample. It was analyzed 10 samples. The statistical fluctuation of iron content was 10% in mass.
The AC magnetic susceptibility at room temperature of representative Mariana sediment sample was compared with Rio Doce river sediment after/before accident and standard compound GdCl3. The values are shown in the Table III.  (2016) 14.6(2)x10 -6 Mariana sediment after accident (2016) 12.9(2)x10 -6 One can noted that the magnetic susceptibility of Rio Doce river sediment in 2016 presented an increment of approximately 10 times, as compared with Rio Doce river sediment before accident in 2015.

CONCLUSION
The Mariana city sediments were analyzed by Syncrotron X-ray diffraction, EDXRF and AC Susceptibility. It was found a significant iron content, which is different in the natural sediments in the same region before the accident with the Fundão dam, Mariana, MG. Some crystalline phases found are originated from processing processes in the mineral extraction industry, as their crystals are of synthetic origin (not found in nature).