depicted in Fig. 11c,d. Besides the major sextet that f itted
perfectly with hematite, an ill-defined doublet super-
imposed to the hematite sextet is present in all spectra of
the studied samples (Fig. 11e). This doublet can be
ascribed to the amorphous limonitic brownish material
which is the residue of the dissolution of dolomite. The
isomer shift (0.34–0.35 mm/s) and the quadrupole split-
ting (0.68 mm/s) at 300 K of these samples match well
with those of ferrihydrite (Cornell and Schwertmann 1996,
p 149). In the MS at 50 K, the doublet is still present, with
an isomer shift of 0.42 mm/s and a quadrupole splitting of
0.76 mm/s. These values are compatible with those found
by Childs and Baker-Sherman (1984) for synthetic
ferrihydrite at 77 K. The absence of a sextet at 50 K
indicates the poor cryst allinity of ferrihydrite, given th at
magnetic ordering is strongly dependent on crystallinity
(Cornell and Schwertmann 1996). The low crystallinit y of
ferrihydrite, on the other hand, explains its absence from
the XRD patterns. Mössbauer spectroscopy confirms,
therefore, that the brownish mate ri al obser v e d by op ti c al
petrography in the residue of the dolomite dissolution is
ferrihydrite. Based on the values of the relative subspect ral
areas, t he abundances of hematite and ferrihydrite were
estimated. It is seen from Table 2 that ferrihydrite is a
subordinate component in the soft ore.
Geochemistry
Analytical results for the 48 samples analyzed for major,
trace, and REE of the soft and hard ores are presented in
Electronic supplementary material Appendices B and C.
Average compositions and standard deviations of popula-
tions for the soft and hard iron ores are given in Tables 4
and 5. Statistics for variables below or very close to the
detection limit were not calculated. When the results of just
a few samples were below this limit, statistics were
calculated using half the detection limit. This procedure
was chosen to establish a distinction for those values, which
were slightly above the detection limit. The following
variables (with their respective detection limit) presented
analytical results below or very close to the detection limit:
Na
2
O (0.01%), K
2
O (0.01%), Sc (3 ppm), Be (3 ppm), Zn
(30 ppm), Ga (1 ppm), Rb (1 ppm), Nb (0.2 ppm), Mo
(2 ppm), In (0.1 ppm), Sn (1 ppm), Cs (0.1 ppm), Hf
(0.1 ppm), Tl (0.1 ppm), Pb (5 ppm), Bi (0.1 ppm), Au
(2 ppb), Br (0.5 ppm), Ir (5 ppm), and Se (3 ppm).
The soft iron ore consists almost entirely of Fe
2
O
3
(average 95.6 wt%), with FeO representing less than 0.5%
(average 0.1 wt%). The average values of major elements in
the massive and porous band of the soft ore were analyzed
separately to confirm the results of petrological analyses
and show that impurities are concentrated within the porous
bands (Table 4). These impurities consist almost entirely of
SiO
2
(average 1.7 wt%), MnO (average 1.7 wt%), and
Al
2
O
3
(average 0.8 wt%) and are mineralogically expressed
by chlorite, talc, and Mn oxides.
Trace element concentrations encountered in the soft ore
are very low (Table 4). Most elements show concentrations
of less than 10 ppm, except for Ba (33 ppm), V (60 ppm),
Cr (79 ppm), Y (19 ppm), Ni (24 ppm), Cu (79 ppm), and
∑REE (23 ppm). When compared to the average compo-
sition of dolomitic itabirite (Table 4), the soft ore is
Fig. 9 Photomicrographs of the
soft ore. a Microbanding of
massive and porous bands,
transmitted light. b, c Detail of a
porous band, transmitted light
(b), reflected light, PPL (c). d
Detail of the previous picture (b)
showing ferrihydrite (f)
surrounding aggregates and
individual crystals
of hematite
Miner Deposita (2008) 43:229–254 241