GEOLOGICA CARPATHICA, 53, 2, BRATISLAVA, APRIL 2002
71 — 77
ANALYSIS OF LOW CONCENTRATION OF FREE FERRIC OXIDES IN
CLAYS BY VIS DIFFUSE REFLECTANCE SPECTROSCOPY
AND VOLTAMMETRY
TOMÁŠ GRYGAR
1,*
, JIŘÍ DĚDEČEK
2
and DAVID HRADIL
1
1
Institute of Inorganic Chemistry AS CR, 250 68 Řež, Czech Republic; *grygar@iic.cas.cz
2
J. Heyrovský Institute of Physical Chemistry AS CR, Dolejškova 3, 183 23 Prague 8, Czech Republic
(Manuscript received October 4, 2001; accepted in revised form December 13, 2001)
Abstract: Ferric oxide admixtures in the concentration range 0.1—10 % in soils, sediments, and clay mineral samples can
be conveniently characterized by Vis diffuse reflectance spectroscopy (DRS) in the region of d-d electron transitions
close to 500 nm, voltammetry of microparticles, and voltammetry with carbon paste electroactive electrode (CPEE).
DRS also detects Fe(III) in the clay mineral matrix equally sensitively. Voltammetry of microparticles is suitable for
direct detection of free crystalline and amorphous ferric oxides. The determination limit of total free ferric oxides with
CPEE is ~0.01 %. The techniques were tested with six clay mineral samples. XRD was able to detect free FeOOH in only
one of them (1.6 %), DRS and CPEE also detected 0.19 % FeOOH in Be-3 (bentonite of Rokle deposit, Bohemia) and
0.04 % in reference montmorillonite SWy-2.
Key words: voltammetry, Vis spectroscopy, analysis, FeOOH, Fe
2
O
3
.
Introduction
Fe in oxic soils, sediments, and clay minerals is present in sev-
eral forms. In the total amount of Fe, free ferric oxides are an
important class of phases, and their analysis is a typical task of
solid-state speciation. Speciation by XRD is complicated by
poor crystallinity and low concentration of free ferric oxides,
commonly below the detection limit of about 1 % for well
crystalline and > 10 % for poorly crystalline species. Chemical
extraction by dithionite-citrate-bicarbonate (DCB) or ammoni-
um oxalate solutions is conventionally used for quantification
of the free ferric oxides, but the phase specificity of extrac-
tions is commonly overestimated. Mössbauer spectroscopy
and selective chemical extraction can be used if the content of
free ferric oxides is > 0.1 % (Komadel et al. 1998). Recently
another two convenient experimental techniques were reported
to be sufficiently sensitive and specific for this analytical task:
Vis diffuse reflectance spectroscopy (DRS) and voltammetry.
Although both these techniques are almost traditionally used
in analysis of inorganic solids, they have been applied to ferric
oxides in low concentrations for geochemical analysis as late
as in the last decade. Vis spectroscopy was proposed to deter-
mine soil goethite and hematite in concentrations > 0.1 %
(Malengreau 1996; Scheinost et al. 1998). DRS was also used
for phase identification in poorly crystalline palagonitic soil
(Morris et al. 1993). Voltammetry of microparticles (VMP)
was found to be similarly sensitive to detect and semi-quanti-
tatively analyze free ferric oxides in paleosoils (Grygar & van
Oorschot 2002) and in lacustrine sediments (van Oorschot et
al. 2001). Another electrochemical technique, voltammetry
with a carbon paste electroactive electrode (CPEE) was used
for quantitative analyses of pure synthetic ferric oxides (Le-
cuire 1975) and pure natural ilmenites (Andriamanana et al.
1984), but it has not been used for analysis of sub-percent con-
centration in mixtures, although the detection limit of CPEE
can be as low as 0.02 % (Brainina & Vydrevich 1981). As it
was shown by using rock magnetism methods, free ferric ox-
ides in paleosoils and sediments bear information that can be
related to the paleoenvironment in the Quaternary (Dekkers
1997). Such analyses are enabled by particular sensitivity of
rock magnetic measurements to ferrimagnetic phases Fe
3
O
4
and
γ
-Fe
2
O
3
. However, there is a clear lack of methods that
would similarly sensitively detect antiferromagnetic and para-
magnetic pigment ferric oxides, mainly hematite and goethite,
whose ratio also has a well-established environmental diag-
nostic value (Cornell & Schwertmann 1996). The need for
such methods is particularly relevant with respect to wide-
spread occurrence of non-ferrimagnetic pigment ferric oxides
in the environment.
The aim of this study was to apply Vis spectroscopy and
voltammetry in analysis of free ferric oxides in selected well-
characterized clay mineral samples to evaluate the sensitivity
and specificity of these techniques, which are rather novel in
geochemical analysis. DRS seems to be a very promising tool
for detection of Fe(III) in free oxides as well as in clay miner-
al’s skeleton. Our aim is hence to promote application of these
techniques because their full utilization is conditioned by col-
lecting more data for comparative purposes taking into ac-
count the large structural variability of clay minerals.
Experimental
Samples
The samples of clay minerals are described in Tables 1 and
2. X-ray powder diffraction (XRD) was used for mineralogical
MECC ‘01
72 GRYGAR, DĚDEČEK and HRADIL
analyses with SIEMENS D-5005 (CuK
α
radiation, secondary
monochromator, 40 kV, 30 mA).
Samples KGa-2, PF-1, STx-1 and SWy-2 and their mineral-
ogical and chemical composition were provided by Source
Clay Repository of the Clay Minerals Society in Missouri,
U.S.A. The content of clay minerals was also confirmed by
additional XRD measurements under conditions described be-
low.
Samples KIC-8 and Be-3 were separated from the raw mate-
rial so that they would be representative and prepared for
measurement as non-orientated mixtures (dried and powdered
in an agate mill, measured in the 2
θ
range 3—80
°, step 0.02°
per 10 seconds). Then they were pre-treated by sedimentation
in distilled water and the clay fraction (< 4
µ
m) was separated
and sedimented on a glass slide to make clay aggregates orien-
tated along their basal crystal planes. These orientated speci-
mens were analyzed conventionally under air-dried conditions
(in the 2
θ
range 3—70
°, step 0.02° per 10 seconds) and under
ethylene glycol-solvated conditions (in the 2
θ
range 3—40
°,
step 0.02
° per 2 seconds). The raw data were processed by the
ZDS program for Windows (Ondruš 1997) employing the dif-
fraction pattern database (JCPDS 2000). Clay minerals were
interpreted in detail according to Moore & Reynolds (1997).
Chemical analyses of bulk clay samples were obtained in the
analytical laboratories of Gematest Ltd. and Laboratories of
the Geological Institutes of Charles University in Prague.
Voltammetry
The cell used for the electrochemical measurements is
shown in Fig. 1. Working electrode, either with immobilized
sample particles (VMP) or with carbon paste, saturated calom-
el reference and Pt-plate counter electrodes were used. In
VMP samples were deposited mechanically on the surface of a
paraffin-impregnated graphite rod (Grygar 1996; Grygar &
van Oorschot 2002). The working electrode was then touched
to the supporting electrolyte and fixed in the cell to minimize
the wetting of the sides of the carbon rod. This is necessary to
decrease the background current, which arises from oxygen
and water reduction on the graphite surface. Because the actu-
al amount of sample on the electrode surface in VMP is not
known, the technique is suitable for the phase identification
and relative comparisons.
For the quantitative analysis a carbon paste electroactive
electrode (CPEE) was prepared as in reports by Lecuire (1975)
and Andriamanana et al. (1984). 2—15 mg sample, about
100 mg graphite powder (Electrocarbon Topo čany, Slovakia),
and 0.1 ml acetate buffer (acetic acid to Na-acetate 1 : 1, total
acetate 1 M) were mixed in an agate mortar to obtain paste of
characteristic butter-like consistency. An important prerequi-
site for quantitative analyses with CPEE is that the loading of
the carbon paste with the ferric oxides must be a few percent
or less to be completely involved in the electrochemical reac-
tions. Immediately after mixing, 20—30 mg of the paste was
spread over the surface of epoxide-resin impregnated graphite
rod, weighed, covered by a piece of microtene foil secured by
a rubber O ring (Fig. 1). The foil was laid on the paste not to
leave any air bubbles between the paste and foil, and than the
foil was perforated by a needle to ensure the electric contact
between the electrode and supporting electrolyte. Acetate
buffer (1 : 1, total acetate 0.2 M) was used as supporting elec-
Fig. 1. The scheme of the measuring cell and electrodes for the vol-
tammetric measurement. WE – working electrode, RE – refer-
ence saturated calomel electrode, CE – counter-electrode. CPEE
electrode consists of the carbon paste with the sample, covered by a
membrane (perforated microtene foil) fixed to a graphite rod by a
rubber ring. In VMP the sample is mechanically attached to the
graphite rod.
N
2
I
E
WE CE
RE
CPEE
VMP
membrane
O-ring
carbon paste
with sample
graphite
rod
graphite
rod
sample
POTENTIOSTAT
CELL
WORKING
ELECTRODES
acetate
buffer
Sample
Origin
Grain size (
µm)
Mineralogical composition (XRD)
KGa-2
Warren County, Georgia
< 2
poorly ordered kaolinite
KIC-8
Skalná nr. Cheb – ULK clay
< 4
poorly ordered kaolinite, illite, (goethite)
STx-1
Gonzales County, Texas
< 2
Ca-montmorillonite, (cristoballite)
SWy-2
Crook County, Wyoming
< 2
Na-montmorillonite, quartz, illite, (calcite)
Be-3
Rokle, Czech Republic
< 4
Ca-montmorillonite, illite, calcite, quartz (kaolinite)
PF-1
Gadsden County, Florida
< 2
palygorskite, quartz
Table 1: General description of clay samples. Parentheses stand for less than about 5 % of the phase.
ANALYSIS OF LOW CONCENTRATION OF FREE OXIDES IN CLAYS 73
STx-1
ν (cm
–1
)
18,100
18,750
19,600
20,500
21,000
21,800
10
7
*A
8.2
23
18
35
7.6
12
SWy-2
ν (cm
–1
)
18,250
18,700
19,400
19,900
20,600
21,300
21,700
10
7
*A
31
65
57
200
100
200
300
Be-3
ν (cm
–1
)
18,400
19,500
20,200
21,200
21,800
10
7
*A
660
3900
1680
2500
800
KIC-8
ν (cm
–1
)
18,700
19,500
20,300
20,900
21,800
10
7
*A
860
1640
660
7090
950
PF1-1
ν (cm
–1
)
18,100
19,600
20,800
10
7
*A
1.7
300
78
KGa-2
ν (cm
–1
)
18,400
19,300
20,200
21,800
10
7
*A
3.2
40
153
18
Identification
h
h
cl-1
cl-2
cl-2 or g
g
cl-3
KGa-2
STx-1
SWy-2
KIC-8
PF1-1
Be-3
SiO
2
43.90
70.10
62.90
45.03
60.90
42.64
TiO
2
2.08
0.038
0.09
0.62
0.49
3.72
Al
2
O
3
38.50
16.00
19.60
33.93
10.40
12.82
Fe
2
O
3
0.98
0.65
3.35
5.61
2.98
11.11
FeO
0.15
0.15
0.32
0.25
0.40
0.35
MgO
0.03
3.69
3.05
0.00
10.20
2.42
CaO
0.00
1.59
1.68
0.17
1.98
5.77
Na
2
O
<0.005
0.27
1.53
0.59
0.058
0.21
K
2
O
0.065
0.078
0.53
3.36
0.80
0.96
P
2
O
5
0.045
0.026
0.049
0.13
0.80
0.68
CO
2
-
0.04
0.05
-
0.11
3.17
Table 2: Elemental composition of clay mineral samples.
trolyte. PC-controlled potentiostat
µ
Autolab (EcoChemie
Utrecht, the Netherlands) was used in linear-sweep (in VMP)
or normal cyclic voltammetric mode (with CPEE). The soft-
ware package GPES 4.4 supplied by EcoChemie was used for
voltammetric data processing.
To distinguish goethite and hematite, sub-samples were
heated at 320
°C to convert goethite to protohematite (Grygar
1996) and voltammograms of original and heated samples
were compared (Grygar & van Oorschot 2002). Protohematite
is more reactive than the original goethite, and hence the re-
ductive dissolution peak of goethite is shifted to more positive
values after heating (Fig. 2B). Hematite peak position is not af-
fected by the heating.
Vis diffuse reflectance spectroscopy (DRS)
Spectra of homogenized powdered samples were recorded
using Perkin-Elmer Lambda 19 UV-Vis-NIR spectrometer
equipped with a standard device for measuring powder sam-
ples called “praying mantis”. Spectra were recorded in the re-
gion 200—900 nm with 1 nm increment and BaSO
4
as a refer-
ence. The remission function F(R
∞
) was calculated from the
Schuster-Kubelka-Munk equation F(R
∞
) = (1 R
∞
)
2
/2 R
∞
, where
R
∞
is the diffuse reflectance from a semi-infinite layer. F(R
∞
)
of solids is proportional to the concentration of absorbing spe-
cies in solids as with absorbance in the case of dissolved com-
pounds in solutions.
Two procedures were used for the characterization of sam-
ples: decomposition of the spectra to the Gaussian bands and
analysis of the second derivative mode of the spectra. The sig-
nal noise was removed by application of Fourier filter (cross 5
points). The second derivative of the spectra was connected
with smoothing by adjacent averaging cross 9 points. Data
processing was carried out using the Microcal Origin 4.1 soft-
ware (Microcal Software, Inc. U.S.A.).
The region of electron pair transition (EPT), (
4
T
1
+
4
T
1
)
←
(
6
A
1
+
6
A
1
), around 500 nm (15,000—30,000 cm
—1
) was used for
the estimation and characterization of Fe(III) species in sam-
ples. The wavenumber of the Fe(III) EPT transition in hema-
tite significantly differs from corresponding transitions of oth-
er Fe oxides (Scheinost et al. 1998). However, it is necessary
to point out, that the d-d transition of Fe(III) does not allow us
to distinguish goethite from other Fe oxides (ferrihydrite, lepi-
docrocite, maghemite, etc.). Spectral characteristics of the
samples studied are collected in Table 3.
Results and discussion
The colour of the six studied samples varied from very pale
grayish (samples with < 1 % total Fe
2
O
3
) to yellow (KIC-8) or
yellowish brown (Be-3). As for the ferric oxides, XRD was
only able to detect goethite in poorly ordered kaolinite sample
KIC-8.
Table 3: Wavenumbers
ν
and integral intensities A of absorption bands of clay mineral samples in 2nd derivative diffuse reflectance
spectra. Abbreviations: h – hematite, g – goethite, cl-1, cl-2, cl-3 – clay minerals’ skeletal Fe.
74 GRYGAR, DĚDEČEK and HRADIL
Voltammetry of Fe oxides can in principle use one of four
electrochemical reactions: reductive dissolution to Fe
2+
, re-
duction to metallic Fe and its re-oxidation to dissolved Fe
2+
,
and/or oxidation of total Fe
2+
obtained by the previous reac-
tions (Lecuire 1975; White et al. 1994; Grygar 1996). Reduc-
tive dissolution is suitable for identification of ferric oxide
phases (Lecuire 1975; Grygar 1996), but the analysis must en-
counter problem of large background current of working elec-
trode in the corresponding potential range (see the increasing
absolute value of current with decreasing potential in curves 1
and 2 in Fig. 2A). Formation of metallic Fe is only typical as a
side reaction in reduction of Fe
3
O
4
(White et al. 1994) and
α
-
and
γ
-Fe
2
O
3
polymorphs, but we did not observe that reaction
in the presented case. Re-oxidation of Fe
2+
is most suitable for
quantitative analysis of reducible Fe oxides (Lecuire 1975;
Andriamanana et al. 1984), as in the corresponding potential
range there is no significant background current (peak A1 in
Fig. 2C).
VMP was used to directly detect reductive dissolution of
free ferric oxides (Grygar 1996; Grygar & van Oorschot 2002;
van Oorschot et al. 2001). Reductive dissolution:
FeOOH + 3 H
+
+ e
—
= Fe
2+
+ 2 H
2
O
(1)
Fe
2
O
3
+ 6 H
+
+ 2e
—
= 2 Fe
2+
+ 3 H
2
O (2)
is responsible for linear-sweep voltammetric peaks C1 and C2
in Fig. 2. Due to a large background current, the net electro-
chemical signal can only be obtained by subtracting the cur-
rents of the 1st and the 2nd scans (Grygar & van Oorschot
2002), as it is shown in Fig. 2A. Discrimination between goet-
hite and hematite, which are of approximately the same elec-
trochemical reactivity if their particles have the same size
(Grygar 1996), can be done by heating a small fraction of sam-
ple at 320
°C to produce the highly reactive form of hematite,
and comparing their voltammograms (Grygar & van Oorschot
2002). Due to the thermal conversion, the voltammetric peak
C2 of the heated sample is moved toward more positive poten-
tials if goethite is present (Fig. 2B). In such a way VMP was
able to detect goethite in samples KIC-8 and Be-3. Because in
VMP soluble reaction products including Fe
2+
can freely dif-
fuse to the bulk of the surrounding solution, re-oxidation peak
A1 cannot be observed using this technique.
The voltammograms of reductive dissolution C1 and C2 are
worse developed using CPEE (Fig. 2C). However, contrarily
to VMP, the ferrous salt produced by the reductive dissolution
of ferric oxides according to equations (1) and (2) is retained
in the bulk of the carbon paste and can be re-oxidized in subse-
quent anodic scans (peak A1 in Fig. 2C). Because the total
amount of sample in the paste is known and the charge corre-
sponding to this re-oxidation can be recalculated to weight of
Fe using the Faraday law (Lecuire 1975), quantitative analysis
is possible without any calibration. The charge is equal to the
integral of the voltammetric peak, (the gray-highlighted area
in Fig. 2C). Because of negligible background current and no
side reactions in the potential range of A1, CPEE determina-
tion is more sensitive than the VMP detection of free ferric ox-
ides. Using this approach, free ferric oxides were determined
in samples KIC-8, Be-3, and SWy-2. The results were ex-
Fig. 2. The voltammetric curves of sample KIC-8. A – linear-
sweep VMP of untreated sample, curve 1 and 2 are 1st and 2nd
scans, curve 3 is the difference between the 1st and 2nd scans multi-
plied by 5. B – VMP of original and heated sample, curves were
offset for clarity. C – cyclic voltammetry with CPEE, 1st and 2nd
scans, grey area corresponds to the charge of Fe
2+
oxidation. Peak
denotation: C1 reductive dissolution of amorphous Fe(III) oxides,
C2 reductive dissolution of crystalline Fe(III) oxides, A1: re-oxida-
tion of Fe
2+
dissolved from Fe(III) oxides, C3: reduction of Fe
3+
in
solution.
-1.2
-0.9
-0.6
-0.3
0
-6
-4
-2
0
E (V vs. SCE)
I (
A)
µ
1
2
3
C1
C2
A
-1.6
-1.2
-0.8
-0.4
0
-1.2
-0.9
-0.6
-0.3
0
E (V vs. SCE)
I (
A
)
µ
original
heated
B
-0.3
-0.2
-0.1
0
0.1
0.2
-1
-0.5
0
0.5
1
E (V vs. SCE)
I (mA)
C3
C1,C2
A1
1
2
C
ANALYSIS OF LOW CONCENTRATION OF FREE OXIDES IN CLAYS 75
Sample
XRD
DRS
VMP
CPEE
KGa-2
goethite
KIC-8
goethite
goethite
goethite
1.6 % FeOOH
STx-1
SWy-2
(goethite)
0.04 % FeOOH
Be-3
goethite, hematite
goethite
0.19 % FeOOH
PF1-1
(goethite)
pressed as FeOOH because DRS identified goethite as the ma-
jor free ferric oxide in the samples (Table 3). The results of the
quantitative CPEE analysis are summarized in Table 4.
In contrast to the report by Xiang & Villemure (1995), we
did not observe any signs of redox cycling of the clay miner-
al’s skeletal Fe ions. Negligible electrochemical activity of
skeletal Fe in montmorillonite enables its application as modi-
fier in carbon paste electrodes for analysis of dissolved elec-
troactive species (Navrátilová & Kula 2000).
The UV-Vis absorption spectra of ferric oxides exhibit
strong charge-transfer absorption bands with maximum
around 40,000 cm
—1
, followed by medium-intense d-d electron
pair transition (EPT) bands around 25,000 and 20,000 cm
—1
and weak d-d bands around 14,000 and 11,000 cm
—1
. The
charge-transfer band is very sensitive but not phase specific.
Although the hematite band at around 11,500 cm
—1
is enough
specific to hematite (Morris et al. 1993; Scheinost et al. 1998),
it is too weak to be applied in trace phase analysis. The EPT
bands are hence most appropriate for the characterization of
Fe(III) oxides in soils (Scheinost et al. 1998). The EPT absorp-
tion bands were distinguished according to the minima in the
2nd derivative spectra as in the previous reports (Malengreau
et al. 1996; Scheinost et al. 1998). Scheinost et al. gave wave-
number ranges of the bands by goethite (20,300—20,900 cm
—1
)
and hematite (17,700—19,200 cm
—1
, median 18,800 cm
—1
). Fer-
rihydrite band occurs at 20,000—20,700 cm
—1
, but we have not
got any proof of its presence in the samples studied.
The noise removal based on the application of Fourier filter
followed by the second derivative of the spectra is significant-
ly better for the spectra evaluation than the cubic spline fitting
procedure applied by Scheinost et al. (1998) and Malengreau
et al. (1996), as the Fourier filter does not alter the shape of the
spectrum including weak bands (shoulders). The 2nd deriva-
tive spectra were decomposed into Gaussian curves corre-
sponding to individual absorption bands and the areas of these
Gaussian bands were used as a measure of the intensity of
Fe(III) absorption bands in the spectrum; the results are col-
lected in Table 3. Because the absorption coefficients of the in-
dividual species are not known, integral intensities can only be
used to evaluate a relative order of concentration of the free
ferric oxides in the samples. Their goethite content decreased
in the following order:
KIC-8 > Be-3 > SWy-2 >> PF-1 >> STx-1
Goethite is absent in KGa-2. Hematite is only present in a
small amount in Be-3 and in traces in SWy-2, STx-1. Ferric
oxides found in samples are given in Table 4.
Fig. 3. 2nd derivative DR spectra of three samples. The wavenum-
bers of minima are given, Ge stands for goethite and He for hema-
tite, further absorption bands are identified in Table 3.
wavenumber / cm
-1
0
0.4
20300
18700
He
19500
21800
20900
Ge
18000
22000
20000
0.2
-0.2
KIC-8
kaolinite, illite
10^6*
2nd derivati
of intensi
ty
ve
Table 4: The evaluation of free ferric oxides in the clay mineral samples. Brackets denote a trace amount of the ferric oxide phase.
wavenumber / cm
-1
-0.01
0
0.02
18100
He
19600
20900
Ge
18000
22000
20000
0.01
10^
6*
2n
d
d
e
ri
v
at
i
o
f
in
te
n
s
it
y
PF1-1
palygorskite
ve
0
0.004
21800
21100
20500
Ge
19600
18750
18100
He
18000
22000
20000
0.002
-0.002
He
wavenumber / cm
-1
10^
6*
2n
d
d
e
ri
v
a
ti
o
f i
n
te
n
s
it
y
STx-1
montmorillonite
ve
76 GRYGAR, DĚDEČEK and HRADIL
Table 5: The overall evaluation of DRS and voltammetry as tools for analysis of admixtures of free ferric oxides.
Method
Specificity
Detection/determination limit
DRS
~500 nm
Directly distinguishing goethite and hematite,
overlap of goethite with ferrihydrite and lepidocrocite
~0.1% (detection),
semi-quantitative comparisons
VMP
red. dissolution
Indirectly distinguishing goethite and hematite,
distinguishing amorphous and crystalline ferric oxides
0.1–0.5% (detection)
CPEE
Fe
2+
re-oxidation
Only for sum of reductively dissolved Fe oxides
~0.01% (determination)
Other Fe(III) octahedral species than free ferric oxides must
be responsible for the other three absorption bands observed in
the samples and denoted cl-1 to cl-3 in Table 3. As follows
from the report by Bishop et al. (1993), Fe(III) montmorillo-
nite also absorbs in this region. The Fe(II) and Fe(II)-Fe(III)
electronic absorption bands are present at longer wavelengths
(~800 nm, Komadel et al. 1990; Morris et al. 1993), and tetra-
hedral Fe(III) are also of different spectral properties than oc-
tahedral Fe(III) (Lever 1984).
Comparing the presence of bands cl-1 to cl-3 to the kind of
clay minerals, we propose the following identification of those
bands. Band cl-1 is mainly typical for 2 : 1 clay minerals, cl-2
for both 2 : 1 and 1 : 1 clay minerals, and cl-3 is dominant in
samples with a significant content of mica-illite. It is notewor-
thy that 2nd derivative spectra enable resolution of as many as
seven bands in SWy-2 in a relatively narrow bandwidth, how-
ever, the fact that all these bands are present in at least one of
the other samples indicates that certain well-defined species
are responsible for them. Further study would be necessary
with a larger set of clay minerals with skeletal Fe(III) to identi-
fy those species.
Conclusions
The comparison of the methods described above is summa-
rized in Table 5. Both Vis diffuse reflectance spectroscopy
(EPT bands in the range approx. 17,500—22,500 cm
—1
) and
voltammetry can compete with XRD and other analytical tech-
niques for speciation of Fe(III), especially in analysis of free
ferric oxides in natural solid samples such as soils, sediments,
and separated clay mineral fractions. Sensitivity of these tech-
niques is generally much better than that of XRD. Further-
more, the simultaneous speciation of free ferric oxides and oc-
tahedral Fe(III) in the clay mineral structure is a unique and
yet not exploited possibility of Vis spectroscopy.
Direct evidence of free ferric oxides by VMP can possibly
be affected by a simultaneous presence of reducible species,
such as Pb compounds and oxygen, which must be expelled
from electrochemical system. On the contrary we are not
aware of any interferences in the determination of total free
ferric oxides by re-oxidation of Fe
2+
in CPEE. The low deter-
mination limit of CPEE (~0.01 %) is enabled by the fact, that
charges of about 0.1 mC (~0.06
µ
g Fe
2
O
3
) can be convenient-
ly determined with a common potentiostat and with the carbon
paste containing only a few milligrams of sample.
Acknowledgment: The work on speciation of Fe oxides in a
mineral matrix was solved in the framework of the project
supported by Grant Agency of CR (Project 205/00/1349).
References
Andriamanana A., Lamache M. & Bauer D. 1984: Etude electro-
chemique de differentes ilmenites. Electrochim. Acta 29,
1051—1054.
Bishop J.L., Pieters C.M. & Burns R.G. 1993: Reflectance and
Moesssbauer spectroscopy of ferrihydrite-montmorillonite as-
semblages as Mars soil analog materials. Geochim. Cosmo-
chim. Acta 57, 4583—4595.
Brainina Kh.Z. & Vydrevich M.B. 1981: Stripping analysis of sol-
ids. J. Electroanal. Chem. 121, 1—28.
Cornell R.M. & Schwertmann U. 1996: The iron oxides. VCH Wein-
heim, Germany, 375—432.
Dekkers M.J. 1997: Environmental magnetism: an introduction.
Geol. Mijnbouw 76, 163—182.
Grygar T. 1996: Electrochemical dissolution of iron(III) hydroxy-
oxides: More information about the particles. Coll. Czech.
Chem. Commun. 61, 93—106.
Grygar T. & van Oorschot I.H. M. 2002: Voltammetric identification
of pedogenic iron oxides in paleosol and loess. Electroanaly-
sis, in press.
JCPDS 2000: Powder Diffraction File, PDF-2, International Centre
for Diffraction Data, Newtown, PA, USA.
Komadel P., Lear P.R. & Stucki J.W. 1990: Reduction and reoxida-
tion of nontronite: extent of reduction and reaction rates. Clays
and Clay Miner. 38, 203—208.
Komadel P., Grygar T. & Mehner H. 1998: Reductive dissolution and
Mössbauer spectroscopic study of Fe forms in the fine fractions
of Slovak Fe-rich bentonites. Clay Miner. 33, 593—599.
Lecuire J.-M. 1975: Réduction Électrochimique des Oxydes de Fer.
Application a la Mesure de non Stoechiométrie. J. Electroanal.
Chem. 66, 195—205.
Lever A.B.P. 1984: Inorganic electronic spectroscopy. Elsevier,
Amsterdam, 1—452.
Malengreau N., Bedidi A., Muller J.P. & Herbillon A. J. 1996: Spec-
troscopic control of iron oxide dissolution in two ferralitic
soils. European J. Soil Sci. 47, 13—20.
Moore D.M. Reynolds R.C. 1997: X-ray diffraction and the identifi-
cation and analysis of clay minerals. Oxford University Press,
Oxford.
Morris R.V., Golden D.C., Bell J.F., Lauer H.V. & Adams J.B. 1993:
Pigmenting agents in Martian Soils: Inferences from spectral,
Mössbauer, and magnetic properties of nanophase and other
iron oxides in Hawaiian Palagonitic Soil PN-9. Geochim. Cos-
mochim. Acta 57, 4597—4609.
Navrátilová Z. & Kula P. 2000: Cation and anion exchange on clay
ANALYSIS OF LOW CONCENTRATION OF FREE OXIDES IN CLAYS 77
modified electrodes. J. Solid State Electrochem. 4, 342—347.
Ondruš P. 1997: ZDS – software for X-ray powder diffraction anal-
ysis. ZDS Systems Inc., Prague, Czech Republic.
Scheinost A.C., Chavernas A., Barrón V. & Torrent J. 1998: Use and
limitations of second-derivative diffuse reflectance spectrosco-
py in the visible to near-infrared range to identify and quantify
Fe oxide minerals in soils. Clays and Clay Miner. 46, 528—536
van Oorschot I.H. M., Grygar T. & Dekkers M.J. 2001: Detection of
small concentrations of fine-grained iron oxides in soils and
sediments by voltammetry of microparticles. Earth Planet. Sci.
Lett. 193, 631—642.
White A.F., Peterson M.L. & Hochella M.F. 1994: Electrochemistry
and dissolution kinetics of magnetite and ilmenite. Geochim.
Cosmochim. Acta 58, 1859—1875.
Xiang Y. & Villemure G. 1995: Electrodes modified with synthetic
clay minerals: evidence of direct electron transfer from struc-
tural iron sites in the clay lattice. J. Electroanal. Chem. 381,
21—27.