GeolCarp_Vol62_No2_181

www.geologicacarpathica.sk

GEOLOGICA CARPATHICA

GEOLOGICA CARPATHICA, APRIL 2011, 62, 2, 181—186 doi: 10.2478/v10096-011-0015-x

Introduction

Clays containing vanadium (V) have been described in a wide
range of geological environments (Muller et al. 1995). In these
materials, V is present in the + 3, + 4, and + 5 oxidation states,
which are closely related to the physicochemical conditions of
the clay formation (Premović 1984; Premović et al. 1986).
Chemical conversion of V to vanadyl (VO

) in natural waters

is strictly determined by both the acidity (pH) and the redox
potential (Eh). Therefore, VO

is an excellent indicator of the

geochemical conditions of clay formation and may provide
clues to the origin of the clay deposits of the past. In an inves-
tigation of these conditions, an accurate assessment of the por-
tion of V that occurs as vanadyl (VO

) is necessary.

Therefore, it is highly desirable to establish a relatively simple
and practical method for VO

determination.

Electron spin resonance (ESR) has become a popular and

useful tool for geochemists and geologically oriented scien-
tists. A large number of investigators working on paramag-
netic  ions  in  geological  materials  have  utilized  this  method
to probe the structural and dynamic aspects of ions. ESR is
especially  sensitive  to  paramagnetic  ions  and  has  excellent
resolution and signal reproducibility. Trace quantities of sev-
eral  transition  metals  may  be  detected  and  quantitatively
measured by ESR. Quantitative analysis of clays containing
these  trace  metals  has  not  been  as  extensive  by  ESR  as  by
other spectroscopic methods.

The use of ESR in the characterization of VO

com-

pounds is well established. A part of the V in clays exists as
VO

and ESR can be utilized to quantitatively determine

concentrations (Premović 1984; Premović et al.

1993a). In these studies, we compare the ESR signals from a
sample containing an unknown amount of VO

to that from

a reference sample with a known concentration of Cu

ions.

This determination is based on the assumption that the VO

content of the analysed sample and the Cu

content in the

A new method for determining the concentration of vanadyl

ions in clays

PAVLE I. PREMOVIĆ, BUDIMIR S. ILIĆ and DRAGAN M. ĐORĐEVIĆ

Laboratory for Geochemistry, Cosmochemistry and Astrochemistry, University of Niš, P.O. Box 224, 18 000 Niš, Serbia;

pavle.premovic@yahoo.com

(Manuscript received March 8, 2010; accepted in revised form October 11, 2010)

Abstract: A novel and simple method for quantitatively determining the concentration of vanadyl ions in clays using
electron spin resonance data has been developed. Several vanadyl standards with concentrations between 200—1000 ppm
were prepared in a mixture of glycerol and kaolinite (KGa-2). The anisotropic electron spin resonance (ESR) spectra were
recorded at room temperature, and the specific intensity of the line (attributed to nuclear spin m = —5/2||) was determined.
For vanadyl concentrations between 50 ppm and 200 ppm, the standards must be prepared by mixing kaolinite with known
vanadyl content (FBT2A-03) and kaolinite (GB1) containing no vanadyl. The method is applicable without modification to
other clays and clay-rich sediments containing vanadyl ions. The whole procedure is very suitable for routine work.

Key words: clay, kaolinite, vanadyl, determination, electron spin resonance.

reference sample (the standard) are related to each other by a
simple ratio:

]= ( SI

/SI

) [C

]

where SI is the specific signal intensity (the area under the cor-
responding  ESR  absorption  per  g  of  sample)  of  particular
paramagnetic ions. This assumption may not be valid because
the  transition  probabilities  (i.e.  the  coefficients  relating  the
specific  signal  intensity  to  the  number  of  absorbing  species)
for VO

and Cu

ions might not be equal (Aasa & V

nngard

1975). This method is also long because it requires an integra-
tion of the whole spectrum and is somewhat complex; the ab-
solute error is estimated to be at least  > 50 %, which is quite
high  for  the  ESR  technique.  Dielectric  differences  between
some  clays  and  standard  samples  may  create  additional  dis-
crepancies in analyses of VO

ions by the ESR method. The

present report describes a new method for determining the
concentration of VO

ions in clays which eliminates these

obstacles. The method has been tested with samples of natural
kaolinite containing V and VO

The investigation was undertaken with three specific objec-

tives: (a) to determine the VO

content in kaolinite using

ESR; (b) to evaluate ESR as a possible convenient, rapid, sen-
sitive and accurate quantitative method for the determination
of VO

in kaolinite and other clay minerals without chemical

pretreatment, and (c) to better understand the application of
ESR for analytical purposes.

Materials and methods

Inductively Coupled Plasma—Optical Emission Spectros-

copy (ICP—OES) analysis. V of the whole rock samples were
analysed by ICP—OES in the Laboratory for Physical Chem-
istry in the Institute of Nuclear Sciences Vinča. A Spec-

182

PREMOVIĆ, ILIĆ and

EVIĆ

GEOLOGICA CARPATHICA

GEOLOGICA CARPATHICA, 2011, 62, 2, 181—186

troflame ICP—OES instrument was employed, and Ar was
used as the plasma gas.

ESR measurements. The ESR measurements were per-

formed on finely-ground powders of the kaolinite samples in an
ESR quartz tube. Spectra were recorded on a Bruker ER-200 se-
ries ESR spectrometer with Bruker X-band bridges using stan-
dard 100 kHz field modulation. The measurements were made
at 9.3 GHz utilizing a rectangular TE cavity.

To maximize accuracy and precision, a sample tube was al-

ways  kept  inside  the  ESR  cavity  with  an  approximately  uni-
form  magnetic  field,  and  the  reproducibility  of  the  sample
positioning was achieved by using the same sample tube with
a fixed holder. Tests verified that all spectra for the measure-
ments  were  obtained  with  instrumental  parameters  that  gave
no  instrumental  effects  on  the  peak  height,  shape,  or  width.
The  measurements  were  run  at  a  modulation  amplitude  of
2 mT, a time constant of 100 ms, and a scan time of 16 min.
The  field  was  scanned  at  200 mT  when  the  entire  spectrum
was desired and at 20 mT when sensitive tracing of the —5/2||
resonance line (see below) was required. The instrument was
carefully tuned according to the manufacturer’s directions.

Samples and standards. Twenty one reference kaolinite

samples  were  selected  including  the  Nowa  Ruda  dickite.
These  samples,  which  originated  from  sediments  and  hydro-
thermal alteration, are listed in Table 1a together with their lo-
cation,  type,  origin,  total  V  content  and  references.  These
materials  were  selected  because  their  total  V  content  ranges
from 15 ppm to 2475 ppm. Table 1a also lists the VO

con-

centrations of the kaolinite samples studied and the fraction of
their V that occurs as VO

. Combined chemical, Fourier

transform infrared analyses, X-ray diffraction, and microprobe
analyses show that the samples studied contain 63—92 % ka-
olinite. Major impurities are illite and quartz.

A glycerol solution was prepared first by dissolving

known amounts of VOSO

O (Merck) in a solution con-

taining 1.5 ml concentrated H

and 0.5 ml deionized

O. This solution was then diluted to the desired VO

con-

centration (8000 ppm) with thorough agitation. Although the
VOSO

O reagent had impurity content below 10

—5

the real concentration of VO

ions in a weighted portion

could be different from the theoretical concentration because
of the loss of crystallization H

O and partial oxidation of

Sample name

Location

Type

Origin

[ppm]

V as VO

(wt. %)

Reference

KGa-1

Georgia (USA)

Well-ordered kaolinite

Sedimentary 95

100

Gaite et al. 1997

BCH5

Charentes (France)*

Poorly-ordered kaolinite

“

198

174

Delineau et al. 1994

BCH6

“

185

155

“

CHA2

“

LAP1

“

105

“

SGN2

“

131

155

“

SGN3

“

191

201

“

FBT2

“

280

120

“

FBT 4

“

210

“

FBT 2A-02

“

220

132

“

FBT 2A-03

“

250

200

“

FBT 3A-01

“

270

166

“

FBT 3A-02

“

250

175

“

FBT 3A-03

“

225

147

“

Provins

Paris Basin (France)

“

210

Muller et al. 1995

Aranđelovac

Belgrade Basin (Serbia)

“

170

This work

Kolubara

“

155

“

GB1

Cornwall (England)

Well-ordered kaolinite

“

n.d.

–

Cases et al. 1982

Cigar Lake

Saskatchewan (Canada)

“

Hydrothermal 2475

2700

Mosser et al. 1996

“ “

(3 diluted)

“

825

900

“ “

(4 diluted)

“

619

675

“ “

(5 diluted)

“

495

540

Nopal

Chihuahua (Mexico)

Well-ordered kaolinite

“

115

Muller et al. 1990

Nowa Ruda

Silesia (Poland)

Well-ordered dickite

“

175

160

Balan et al. 2002

The content determined by ICP–OES. n.d. — not detected.

Each of the Charentes samples is represented by three letters indicating the name of the open-pit and a

number indicating its position

Table 1a: Location, type, origin and V (ppm)/VO

(ppm) contents of the selected kaolinite samples.

Sample name

Location

Type

[ppm]

V as VO

(wt. %)

Reference

CNS

New Mexico ( USA )

shale

150

Pillmore et al. 1987

SVS

“ “

175

Premović et al. 1993b

Fish Clay

Stevns Klint (Denmark)

marl

120

Premović et al. 1993a

Zvonce black shale

Zvonce (Serbia)

shale

100

Premović 1984

The content determined by ICP–OES.

The short notations for the Cretaceous-Paleogene boundary shales are: CNS — Canadian North Site and SVS — Starkville South Site.

Table 1b: Location, type and V (ppm)/VO

(ppm) contents of the selected kaolinite-rich sedimentary samples.

183

CONCENTRATION OF VANADYL IONS IN CLAYS – A NEW METHOD FOR DETERMINING

GEOLOGICA CARPATHICA

GEOLOGICA CARPATHICA, 2011, 62, 2, 181—186

ions. Therefore, just before preparing the standards, the

composition of VOSO

O was checked using the most

precise gravimetric/titrimetric methods of chemical analysis.
The relative standard deviation (RSD) was less than 5 %.

Changes in the loading Q factor of the ESR cavity can re-

sult in samples that have different dielectric properties or
surfaces. The above glycerol solution has a dielectric con-
stant (56 D) and cannot be used as a reliable comparison for
the relative VO

concentrations in kaolinite. For this reason,

standards were prepared by mixing and diluting the glycerol
VO

solution with KGa-2, which contains a very low

amount of VO

( < 5 ppm), to the desired VO

concentra-

tion. This mixture has a similar dielectric medium and Q pa-
rameter as the kaolinite samples. For the first set of
measurements, the VO

standards were prepared from the

glycerol solution and KGa-2 using a vibrating mill (Perkin-
Elmer) that covered the concentration range from 200 ppm
to 1000 ppm. For the second set of measurements, standards
were prepared from the FBT 2A-03 kaolinite, which contains
200 ppm of VO

, and diluted with GB1 kaolinite, which

contains no VO

(a detection limit 1 ppm), covering the

concentration range from 50 ppm to 200 ppm.

Results and discussion

The concentration of VO

equation:

] = (SI

/SI

) [C

]

where k indicates the kaolinite sample, st indicates the stan-
dard, and SI is the specific signal intensity (the integrated area
under the corresponding ESR absorption per g of the sample).
In the first approximation the discrepancy between the specif-
ic (signal) intensity of the VO

ions in the standards and ka-

olinite are considered to be analytically insignificant.

The kaolinite samples studied show multiline spectra simi-

lar to the spectra of VO

ions incorporated into the lattices

of numerous kaolinite samples, including KGa-1 (Premović
1984; Muller & Calas 1993; Gehring et al. 1993). A repre-
sentative X-band spectrum from KGa-l is shown in Fig. 1a.
A preliminary Q-band measurement indicates that this spec-
trum is a superimposition of the spectra of VO

ions located

in two different positions in the kaolinite structure. The spec-
trum shows an anisotropic pattern typical of axially symmet-
rical hyperfine coupling. A sharp intense peak near g = 2.002
is ascribed to defects that are always present in kaolinite
(Muller et al. 1992).

Figure 1 also illustrates the anisotropic ESR spectrum of:

(b) an initial solution of VOSO

O compound dissolved

in H

O and diluted with glycerol and (c) a standard

containing 1000 ppm of VO

. These spectra are typical of

those previously reported for VO

in either powder (poly-

crystalline) solids or extremely highly-viscous liquids
(Goodman & Raynor 1970).

Because only one line of the VO

anisotropic hyperfine

pattern is necessary for obtaining the integrated area, only a
narrow part of the VO

spectrum needs to be recorded. We

Fig. 1. First derivative, room temperature, anisotropic ESR spec-
trum of: a – KGa-1; b – An initial solution of VOSO

O dis-

solved in H

O and diluted with glycerol; c – A standard

containing 1000 ppm of VO

selected the first derivative

V hyperfine line marked with

= —5/2|| in the spectra of the kaolinite samples (see, for ex-

ample, Fig. 1a) and the 50—1000 ppm VO

standards (see,

for example, Fig. 1c). This line was chosen to keep the line-
width and lineshape similar and to minimize the interferenc-
es from both neighbouring VO

lines and other ESR active

species present. The anisotropy of the ESR parameters of
VO

in various clays has little or no effect on the linewidth

and lineshape of the —5/2|| resonance line.

184

PREMOVIĆ, ILIĆ and

EVIĆ

GEOLOGICA CARPATHICA

GEOLOGICA CARPATHICA, 2011, 62, 2, 181—186

To obtain maximum sensitivity, the spectrum must be re-

corded at a high power level. However, at high power, satu-
ration is possible. Therefore, we plotted the normalized
integrated area (A) of the —5/2|| line against the square root
of the microwave power (P

1/2

) for both the 1000 ppm stan-

dard and the KGa-l sample (Fig. 2a). The linear relationship
shows that quantitative work can be safely performed at
100 mW. At this power, no saturation broadening of the sig-
nal was observed for either the standards or the kaolinite
samples. Consequently, a high power of 100 mW was select-
ed for measurement, ensuring a high absolute intensity of the
—5/2|| line.

Figure 2b shows a plot of the integrated area of the —5/2 line

(vs. days) for five 1000 ppm standards prepared on five dif-
ferent days. The scatter of obtained points averaged to a
straight line (parallel to the day axis in Fig. 2b) with devia-
tions of < 5 %. The repeatability of these results is < 30 ppm
of VO

at the 1000 ppm level.

The spectrometric procedure described enabled us to

achieve sufficient reproducibility of the specific signal from
the standard and kaolinite samples with a VO

anisotropic

ESR spectrum. The RSD was less than 10 %. Although
VO

concentrations as low as 5 ppm were detected, the best

results were achieved for concentrations of about 50 ppm.

Although the kaolinite samples were not collected from

freshly exposed mine faces, repeated ESR analyses over the
course of several months showed no change in VO

content.

This result is consistent with previous studies of VO

stabil-

ity in kaolinite (Muller & Calas 1993; Gehring et al. 1993).
Similar experiments on VO

standards show that after six

weeks no oxidation had occurred (Fig. 2c).

The calibration curves were prepared by determining the

area under the peak of the —5/2|| resonance line and multiply-
ing or dividing that line by factors required to put all of the
standard and kaolinite sample peak areas on the same set-
ting. The peak areas were evaluated using the built-in com-
puter of the spectrometer. This computer performed the
appropriate integrations with baseline corrections.

A linear calibration curve was obtained by plotting the

specific intensities of the —5/2|| lines of the standards against
the VO

concentrations. Fig. 3a shows this plot in the 200—

1000 ppm range. Using this plot as the calibration curve, the
concentration of VO

was determined for the kaolinite sam-

ples (Table 1a) by recording the VO

spectrum and the spe-

cific intensity of the —5/2|| line. It should be noted that the
calibration curve obtained by plotting the specific intensities
of the —7/2, + 5/2 and + 7/2 lines gave similar results but
with less accuracy and precision.

A rough ESR estimate indicates that the VO

content of

the Cigar Lake kaolinite sample is much higher than
1000 ppm. Therefore, this sample was diluted three, four and
five times with the KGa-2 kaolinite. The calibration curve
shown in Fig. 3a was then used to determine the VO

con-

tents of these diluted samples (Table 1a). The VO

content

in the Cigar Lake kaolinite sample is 2700 ± 100 ppm
(Table 1a).

The use of the KGa-2/glycerol mixture as a standard is not

suitable, especially for routine analysis when many samples
with relatively low concentrations ( < 2 00 ppm) of VO

Fig. 2. Saturation behaviour of the —5/2|| VO

resonance line in: a – A

standard containing 1000 ppm of VO

and KGa-l; b – The repeat-

ability of the integrated area using the 1000 ppm standard; c – Effect
of time on the —5/2|| VO

resonance line of the 1000 ppm standard

stored in air; d – The repeatability of the integrated area using a sec-
ondary standard at the 100 ppm level. N is the normalized intensity of
the —5/2|| line; A is the integrated area of this line.

185

CONCENTRATION OF VANADYL IONS IN CLAYS – A NEW METHOD FOR DETERMINING

GEOLOGICA CARPATHICA

GEOLOGICA CARPATHICA, 2011, 62, 2, 181—186

need to be handled. For this reason, new (hereinafter second-
ary) standards were prepared by mixing GB1, which has no
detectable VO

, and the FBT 2A-03 kaolinite, which is rela-

tively enriched with VO

(Table 1a).

Eight secondary standard samples were prepared with 50,

60, 75, 100, 125, 150, 175 and 200 ppm of VO

. A very

good linear relation was obtained as is shown in Fig. 3b. The
linearity shows that these secondary standards are a good
tool for the quantitative determination of VO

in the range

of 50 to 200 ppm. Table 1a also lists the number of kaolinite
samples with VO

content lower than 200 ppm. The calibra-

tion curve shown in Fig. 3b was used to obtain these results.

All secondary standard concentration results were con-

firmed by performing two or more experiments in five days
and through the use of more than one GB1 and FBT 2A-03
sample in each case. Fig. 2d indicates that the repeatability
of these results is < 10 ppm of VO

at the 100 ppm level.

The above method of VO

determination was extended to

the  kaolinite-rich  Cretaceous-Paleogene  boundary  shales
CNS/SVS, the smectite-rich marl Fish Clay and the kaolin-
ite-rich  Zvonce  black  shale,  (Table 1b).  The  calibration
curve  shown  in  Fig. 3b  was  also  used  to  obtain  the  results
listed in Table 1b. Preliminary results indicates that a similar
method can be used to determine the concentration of other
paramagnetic ions, such as Cu

and Mn

in clays.

Fig. 3. Specific signal intensities of the —5/2|| lines of the standards
against the VO

concentration. a – The standards of 200—

1000 ppm. Experimental error bars for the data points are ± 10 %
for VO

and ± 50 10

for specific intensity; b – The standards of

50—200 ppm. Error bars for the data points are ± 10 % for VO

and

± 20 10

for specific signal intensity.

Conclusions

A new, fast, simple and straightforward method for deter-

mining the concentration of VO

in clays and clay-rich sedi-

ments using ESR is described. This method eliminates
difficulties of the previously reported method employing the
Cu

standard. In addition, the proposed method does not re-

quire any pretreatment of the clay and tedious/complex inte-
gration of the whole spectrum. The newly developed method
is sensitive enough to enable quantization of VO

at low

( 50 ppm) and high ( > 2500 ppm) concentrations. These ad-
vantages encourage the application of the proposed method in
routine clay analysis of VO

Acknowledgments:  Reliable  samples  for  this  study  were
provided  through  a  variety  of  sources  and  the  authors  ex-
press their deep thanks to a number of scientists who collect-
ed  specimens  for  analysis.  We  thank  to  Michael  A.
Karakassides,  Peter  Komadel  and  an  anonymous  reviewer
for their constructive comments. Funding for P.I. Premović’s
ESR work at Université Pierre et Marie Curie (Paris) was ob-
tained  from  le  Ministere  Francais  de  l’Éducation  National,
de l’Enseignement Supérieur et de la Recherche.

References

Aasa R. & V

nngard T. 1975: EPR signal intensity and powder

shapes: Are examination. J. Magn. Reson. 19, 3, 308—315.

Balan E., Allard T., Morin G. & Calas G. 2002: Incorporation of

in dickite: A spectroscopic study. Phys. Chem. Miner. 29,

4, 273—279.

Cases M., Fripiat J., Francois J. & Letellier M. 1982: Thermody-

namic and microdynamic behavior of water in clay suspen-
sions and gels. J. Colloid Intef. Sci. 89, 2, 378—400.

Delineau T., Allard T., Muller J.P., Barres O., Yvon J. & Cases

J.M. 1994: FTIR reflectance vs. EPR studies of structural iron
in kaolinites. Clays and Clay Miner. 42, 3, 308—320.

Gaite J.M., Ermakoff P., Allard T. & Muller J.P. 1997: Paramagnet-

ic Fe

: A sensitive probe for disorder in kaolinite. Clays and

Clay Miner. 45, 4, 496—505.

Gehring A.U., Fry I.V., Luster J. & Sposito G. 1993: The chemical

form of vanadium (IV) in kaolinite. Clays and Clay Miner. 41,
6, 662—667.

Goodman B.A. & Raynor J.B. 1970: Electron spin resonance of transi-

tion metal complexes. Adv. Inor. Chem. Radiochem. 13, 135—362.

Mosser C., Boudeulle M., Weber F. & Pacquet A. 1996: Ferriferous

and vanadiferous kaolinites from the hydrothermal alteration
halo of the Cigar Lake uranium deposit (Canada). Clay Miner.
31, 3, 291—299.

Muller J.P. & Calas G. 1993: Genetic significance of paramagnetic

centers in kaolinites. In: Murray H.H., Bundy W. & Harvey C.
(Eds.): Kaolin genesis and utilization. Clay Miner. Soc., Boul-
der, 261—289.

Muller J.P., Ildefonse P. & Calas G. 1990: Paramagnetic defect cen-

ters in hydrothermal kaolinite from an altered tuff in the Nopal
Uranium deposit, Chihuahua, Mexico. Clays and Clay Miner.
38, 6, 600—608.

Muller J.P., Clozel B., Ildefonse P. & Calas G. 1992: Radiation-in-

duced defects in kaolinites: An indirect assessment of radionu-
clide migration in the geosphere. Appl. Geochem. 7, 1, 205—216.

Muller J.P., Manceau A., Calas G., Allard T., Ildefonse P. & Haze-