Clay Minerals (1990) 25, 65-71
T H E E F F E C T OF CO2 A N D O X I D A T I O N R A T E O N T H E
F O R M A T I O N OF G O E T H I T E V E R S U S
L E P I D O C R O C I T E F R O M A N F/~(II) S Y S T E M AT
pH 6 A N D 7
L. C A R L S O N *
AND U . S C H W E R T M A N N
Lehrstuhl far Bodenkunde, Technische Universitiit Miinchen, D-8050 Freising-Weihenstephan, Federal Republic
of Germany
(Received 2 March 1989; revised 4 August 1989)
ABSTRACT : To investigate the influence of carbonate on the formation of goethite and
lepidocrocite, ~200 samples were synthesized by oxidizing FeC12 solutions with air/CO2 gas
mixtures at ambient temperature and pH 6 and 7. The proportion of lepidocrocite in the
lepidocrocite/goethite mixtures (Lp/(Lp+ GO) decreased from 100 to 0% with increasing
[HCO~] in solution and with decreasing average oxidation rate (AOR). These two parameters
explained 81% of the variation of Lp/(Lp + Gt). At a given [HCO~], more goethite was formed
at pH 6 than at pH 7. The Lp + Gt mixtures contained 0-8 mg g-i carbon (CO which could not
be removed by washing. Ct reached apparent saturation at a [HCO~] equilibrium concentration
of ~ 6-8 and 60-80 mmol 1-~ at pH 6 and pH 7, respectively. In a plot of Ct vs. Lp/(Lp + GO all
data fell on the same line irrespective of oxidation parameters (pH, AOR). IR spectra showed
two broad bands at ~ 1300 and 1500 cm-I which can be assigned to distorted carbonate
adsorbed at the goethite surface. Identical bands were also found in a young, poorly crystalline
goethite formed from coal mine drainage in Ohio. It is suggested that carbonate anions direct the
polymerization of the double bands of FeO3(OH) 3 octahedra common to both minerals toward a
comer sharing arrangement, and thereby to goethite, whereas chloride permits edge-sharing as
in lepidocrocite.
The close association of goethite and its metastable polymorph lepidocrocite in m a n y soils
has not been satisfactorily explained. A detailed study of Fe oxide accumulations around root
channels (so-called pipestems) in South African soils showed, however, that goethite
dominated in the inner part of the pipestem, i.e. close to the root, whereas in the outer part
lepidocrocite was the main oxide (Schwertmann & Fitzpatrick, 1977). This distribution could
be due to small-scale differences in chemical environment within the rhizosphere. One of the
factors which may vary in the rhizosphere is Pco:, which, because of root respiration, should
decrease with increasing distance from the root. Schwertmann (1959) oxidized FeCI2
solutions with O2-CO2 mixtures and found that increasing amounts of goethite were formed
as the proportion of CO2 in the purge gas increased. Later, this result was substantiated by
Fey & Dixon (1981). G o o d m a n & Lewis (1981) found that the addition of N a H C O 3 during
oxidation of a FeC12 solution induced the formation of goethite rather than lepidocrocite. A
systematic study of the effect of carbonate is, however, still lacking and experiments were
therefore conducted in an attempt to quantify and possibly explain this effect on F e O O H
polymorph formation.
*Permanent address: Department of Geology, University of Helsinki, PO Box 115, SF-00171 Helsinki,
Finland
9 1990 The Mineralogical Society
66
L. Carlson and U. Schwertmann
METHODS
A total of ~ 200 samples were prepared by oxidizing 30 ml of a 0.1 M FeC12 (or FeSO4)
solution at room temperature with a gas mixture of air and CO2 in a 50 ml polyethylene
beaker at ambient temperature with stirring. The gas flow was accurately monitored with a
calibrated gas mixing apparatus (manufactured by H. Walz, Co., D-8521 Effeltrich, FRG).
Gas flows between 8 and 200 ml rain -1 air and 0 and 95 ml min -1 CO2 in various
combinations were used. A stock solution was prepared by dissolving fresh, analytical grade
FeC12.4H20 (or FeSO,.7H20) in distilled water, and then acidified with a known amount of
HC1 (or H2SO4) to a pH of 2.0-2.3 in order to prevent pre-oxidation. During oxidation the
solution pH was held constant (at pH 7 or 6) by addition of 1 M NaOH using a Radiometer
automatic titrator TTT 80.
The NaOH consumption during oxidation was continuously recorded. The amount of
NaOH consumed after complete oxidation equals the sum of protons from three sources:
complete hydrolysis of iron, neutralization of the acid added initially, and neutralization of
carbonic acid to form HCO~- which is essentially the sole anionic form of carbonate at pH 6
and 7. Therefore, the concentration of HCO~ can be calculated from the difference between
the total amount of NaOH consumed minus the sum of acid added initially (see above) and
the amount of Fe used according to the reaction 4Fe 2+ + 02 + 8OH- --, 4FeOOH + 2H20.
The concentration of HCO~ was also determined in 18 samples by titrating the supernatant
with M HC1 down to the inflection point. The relation between both methods was HC1....
= 0.0759 + 0.943NaOH ..... r = 0.999. This relation was then used to estimate [HCO~] for all
samples. The amount of Fe oxidized divided by the time needed for total oxidation was taken
as the average oxidation rate (AOR). The fully oxidized product was washed three times with
distilled water and dried at 40~
Lepidocrocite (Lp) and goethite (Gt) were identified by X-ray powder diffraction (XRD)
using Co-Kc~radiation and a Philips PW 1050 vertical goniometer equipped with a graphite
reflected-beam monochromator. Gently pressed self-supporting specimens in an A1 holder
were used. The proportion of lepidocrocite in the mixture (Lp/(Lp + Gt)) was taken from the
areas of the lepidocrocite 020 peak (6.27 A) and the goethite 110 peak (4.18 A) measured
using a planimeter. Infrared (IR) spectra were recorded with a Beckman IR 4250
spectrometer using 1:300 Fe oxide :KBr discs at a recording rate of 150 cm -1 min -1. Total
carbon was determined with a Leybold-Heraeus CSA 302 instrument. Surface area was
estimated using ethylene glycol monoethyl ether (EGME) (Carter et al., 1965).
RESULTS
Effect o f carbonate, average oxidation rate and p H
In the following discussion [HCO~] is used even though CO2 and HzCO 3 also exist in the
system, because we believe that a charged species is more likely to react with the Fe oxide
surface than are neutral species. This is in line with IR studies by Russell et al. (1975) and
Rochester & Topham (1979) (see below). Nevertheless, because simple relations exist
between Pco2, pH and [HCO~-], all the following relationships also hold if Pco2 is used instead
of [HCO~-].
At pH 7 lepidocrocite was the sole product of complete oxidation if only air was used as the
purge gas. As Pco~ in the purge gas and thus as [HCO~-] in solution increased, increasing
Formation of goethite vs. lepidocrocite
LP021
Lp020
Gt110
mmol I:1 I.~G~
0
44
3
4
' 3b '
~
67
6s~,
2'0' ' 1 ' 0 "
CoKOl
mma L-1
1.0
52
0.42
63
O.22
89
0.09
0.65
?
3'0
4,
~
78,~
2b lb
CoK~
FIG. 1. X-raydiffractogramsof lepidocrocite-goethitemixtures produced at various [HCO~] in
solution.
proportions of goethite were formed until goethite was the only product above [HCO~-] of
100-120 mmol 1-1 (Fig. 1). The negative correlation between [HCO~] and Lp/(Lp + Gt)
(Fig. 2) was highly significant (r 2 = 0.69, n = 120). Eight goethite samples which formed at a
[HCO~-] higher ( > 120 mmol 1-1) than necessary to produce goethite as the sole phase were
excluded from this calculation.
Although a negative effect of [HCO~-] on Lp/(Lp § Gt) is obvious, the scattering of the
data in Fig. 2 is great. A subdivision of these data into groups of limited ranges of air flow (239; 40-80; 90-140 and > 150 ml min -1, respectively) yielded much higher r2-values of ~0.8
indicating that AOR is an important second factor in determining Lp/(Lp § Gt).
AOR varied between 0.3 and 4.2/Jmol Fe s -1. It depended on stirring speed, shape of
stirrer, total gas flow (air and CO2) and size of bubbles; therefore, it was impossible to keep
AOR constant at the various gas flow rates. Generally, as AOR increased, Lp/(Lp -I- Gt) also
increased. In other words, under the experimental conditions of this study, slow oxidation
favours the thermodynamically more stable goethite over lepidocrocite (Schwertmann,
1959). The correlation between AOR and Lp/(Lp § Gt) was significant with r 2 = 0.49
(n = 120).
A multiple regression analysis considering the combined effects of [HCO~-] and AOR on
Lp/(Lp § Gt) at pH 7 resulted in the following equation:
Lp/(Lp + Gt) = 0.50(9) - 0.0067(10) [HCO~] + 0.167(70) AOR
with an r 2 of 0.81 (n = 120). Figures in parentheses give 95~ confidence limits of the last
digit(s). The partial regression coefficients indicate that within the experimental range of
[HCO~-] (0-120 mmol 1-1) and AOR (0.3--4.2 #mol s-l), both had about the same effect on
Lp/(Lp + Gt) but in an opposite direction.
L. Carlson and U. Sehwertmann
68
1.0'~s9
T".
^
0.8"1
/
1.0
9
9149
~ 9 9
0.8
9
]
0.6/
9
.
_4 0.4 I
""
qp 9
5
"..
9
9 9 ~
~
/
0.2]
.
"
0.6.
.
-4
"'"
9
.#
9
,., 0.4.
~
9
0.2
o9
9
0I
0
9
20
9
40
P
60
: - - - - ~
--
80
[HC03] r ~
:
:
.:
5....
P
9
1(30 120
8
'-1
12
[HOOt] mmol L-1
FIG. 2. Ratio of lepidocrocite to goethite
(Lp/(Lp + Gt)) produced at pH 7 as a function of
[HCO~] in solution. Number indicates number of
points.
FIG. 3. Ratio of lepidocrocite to goethite
(Lp/(Lp + Gt)) produced at pH 6 as a function of
[HCO~] in solution. Number indicates number of
points.
ct %
0.8,
opH6
e pl'l 7
o.6 ooo. :'o
o
0.4
o@
~
9
9
o%
9
9o
%
9
o
9
o
9
0.9.
0
012
o14
oZ6
Lp / ( Lp§
018 - 1~0
)
FIG.4. Relation between total carbon (Ct) and the lepidocrocite to goethite ratio (Lp/(Lp + Gt))
in the mixtures produced at pH 6 and pH 7.
Thirty-six samples were also prepared at p H 6 where [HCO~] at the same Pc9 is only one
tenth of that at p H 7. A t [HCO~-] = 0, lepidocrocite was again the only phase formed.
However, at any given [HCO~-], considerably more goethite was formed than at p H 7 (Fig. 3).
This indicates that the goethite-promoting effect of carbonate is much stronger at p H 6 than
at p H 7.
Carbonate retention of samples
The mechanism by which carbonate favoured formation of goethite over lepidocrocite was
further studied by determination of the total carbon content (Ct) of the samples. Ct ranged
between 0 and 8 mg g-1 and the carbon could not be removed by thorough water washing. In
a plot of Ct vs. Lp/(Lp + GO, samples synthesized at both p H 6 and 7 fall on the same line
(Fig. 4). The amount of carbonate taken up by the oxide was a function of [HCO~] in the
solution. The relationship has the shape of a saturation curve which is more clearly expressed
at p H 6 (Fig. 5a) than at p H 7 (Fig. 5c). A Langmuir plot of the p H 6 d a t a (Fig. 5b) yields a
~
Formation o f goethite vs. lepidocrocite
pH6
Ct (%)
0.6.
eo 9
~1
0.2.
0
o~
o 9
0.4
|
10
0
(el
1'5
pH 6
20"
J"
69
Pco2(kPa)
10
Ct
10-
~L~9
b = 0.78
r =0.98
n=15
b
I*o
~'5
pH7
Ct (%)
96
1 ,.i
ee
!~e
::t"
O!
o
oID 9
9
i~o
[HCO3] mm~ L-~
"goethlte
1800 1600 14~30 12iX) 1000 8130cm-1
c
~
'~J
~o
FIG. 5. Relation between total carbon (Ct) in the
products as a function of[HCO~] in solution at (a) pH
6; (c) pH 7; and (b) a Langmuir plot of the data for
pH 6.
FIG. 6. IR spectra of samples with increasing (from
top to bottom) goethite proportion produced at pH 6
and various Pco~, and a natural, poorly crystalline
goethite formed from coal mine drainage at Belmont
Co., Ohio, USA.
saturation b-value for Ct of 7.75 mg g-1 (650 #mol g-l). Saturation was reached at a [HCO~]
value of ~ 6 - 8 mmol 1-1 at pH 6, and at a [HCO~] value of ~ 6 0 - 8 0 mmol 1-1 at pH 7. At
saturation, only goethite was formed.
Russell et al. (1975) found up to ~ 200/~mol CO2 g-1 adsorbed by goethite with a surface
area between 50 and 100 m 2 g-1. The surface area of two representative goethites was found
to be 230 and 234 m 2 g-1. Assuming a linear relation between carbonate adsorbed and surface
area, the amount of carbonate retained by our goethites could then well be explained by
surface adsorption.
From these results it is concluded that carbonate is essentially associated only with goethite
and not with lepidocrocite, and that it is most likely adsorbed at the goethite surface rather
than incorporated into its structure. This receives further support from I R spectra which
showed the typical features of goethite and lepidocrocite. In addition, two strong but broad
bands at ~ 1300 and 1500 cm -~ appeared, their intensities increasing with the increasing
proportion of goethite (Fig. 6). The intensities of these bands correlated significantly with Ct
(r = 0.79, n = 65) and thus with [HCO~-] and Pco~ in the system. Russell et al. (1975) and
Rochester & Topham (1979) have assigned these bands to the C--O stretch of adsorbed,
distorted CO32The mode of carbonate bonding by goethite is not yet clear9 Yapp (1987) found between
0.06 and 1.3 m o l e ~ entrapped carbonate in natural goethites and suggested a solid solution
70
L. Carlson and U. Schwertmann
between FeOOH and Fe(CO3)OH. Our synthetic goethites contain up to 6 mole~ C but there
is no indication of structural incorporation. A preliminary survey of some soil goethites has
not shown any IR bands which could be assigned to carbonate. However, a poorly crystalline
goethite from an ochreous precipitate in a mine drainage sample (pH 7-3) showed the
carbonate bands in its IR spectrum (Fig. 6). The Ct of this sample is 29-4 mg g-1. (The
authors are grateful to Dr. J. M. Bigham, the Ohio State University for supplying this IR
spectrum). Characteristically, the goethite is very young and it appears likely that with time
the goethite may recrystallize and repel the carbonate.
The effect of sulphate
A slight although much weaker favouring effect of SO42- on goethite was noticed.
Whereas only lepidocrocite was formed in a pure Cl-system at [HCO~-] = 0, 3-21~ goethite
occurred in a lepidocrocite-goethite oxidation product from a 0.1 M FeSO4 solution at pH 7.
The samples showed a strong IR band at 1125 cm -1 and weak bands at 970 and 1170 cm -1
which can be assigned to adsorbed SO4 (Harrison & Berkheiser, 1982).
Crystallinity of goethite and lepidocrocite
As judged by the width at half height (WHH) of its XRD lines, the lepidocrocite was
reasonably well crystallized when it was the sole phase and had a surface area of only 80 m 2
g-1. As the proportion of goethite in the mixture increased the lepidocrocite lines became
broader (Fig. 1), whereas the goethite line-width seemed to be independent of the
mineralogical purity of the samples. The surface areas of the two mixtures with Lp/(Lp + Gt)
~0.5 were 175 and 213 m 2 g-1.
DISCUSSION
The experiments simulate soil conditions with respect to pH, partial pressure of CO2, and
temperature, but not so well for [Fe:+]. Under these conditions, the nature of the anion, rate
of oxidation and pH turned out to be the major factors determining the lepidocrocite to
goethite ratio.
The mechanisms through which these factors act and interact in the formation of the two
FeOOH polymorphs are not clear. At pH 7, the oxides were formed via a green rust phase.
The anion is an essential constituent of the green rust in that it balances the positive charge of
the FelII,II-hydroxy layers. Therefore, the properties of the green rust vary with the nature of
the interlayer anion. In fact, the stability of the green rust increases in the order C1 < SO4 <
CO3 and so may its stability against oxidation (Taylor & McKenzie, 1980; R. M. Taylor,
pers. comm.). It is possible, therefore, that the goethite-favouring effect of carbonate is partly
due to a slower rate of oxidation of carbonate green rust as compared to the other forms of
green rust.
It seems, however, that there must be a dominant, direct effect of the anion, because at
equal rates of oxidation the carbonate system produces more goethite than the sulphate and
chloride systems. Furthermore, at pH 6 at which hardly any green rust is formed, carbonate
has a very strong effect on promoting goethite.
The goethite, while forming, binds appreciable quantities of carbonate rather firmly, but
much less sulphate and no chloride. As the proportion of goethite in the mixture is better
Formation o f goethite vs. lepidocrocite
71
correlated with the amount of carbonate retained by the product than with [HCO~] in
solution, it appears that the carbonate needs to interact with the Fe to favour actively goethite
formation. However, the type of interaction during goethite formation is not clear.
Structurally, both lepidocrocite and goethite consist of identical double rows of edge-sharing
FeO3(On) 3 octahedra, but they differ in the way these double rows are linked together: via
edges in lepidocrocite to form zig-zag layers, and via corners in goethite. Possibly, the various
anions play a role during crystallization when small double-band units link together to
develop the third dimension of the crystal. It is suggested that carbonate and, to a lesser
extent, sulphate hinder edge-sharing of the double band because of their specific adsorption
on these units, whereas chloride does not.
ACKNOWLEDGMENTS
This study was performed during several stays by Liisa Carlson at the Lehrstuhl f/Jr Bodenkunde, Technische
Universit/it Mfinchenin Freising-Weihenstephan, kindly funded by the Deutsche Forschungsgemeinschaftin
cooperation with the Academy of Finland and the Emil Aaltonen Foundation. The skilful experimental work
of Mr. H. Fechter is gratefully acknowledged. We also thank Dr. J. M. Bigham for a helpful review of the
manuscript.
REFERENCES
CARVERD.L., HEILMANM.D. & GONZALESC.L. (1965) Ethylene glycol monoethyl ether for determining
surface areas of silicate minerals. Soil Sci. 100, 356-360.
FEY M.V. & DIXONJ.B. (1981) Synthesis and properties of poorly crystalline hydrated aluminous goethites.
Clays Clay Miner. 29, 91-100.
GOODMANB.A. & LBWISD.G. (1981) M6ssbauer spectra of aluminous goethites (et-FeOOH).J. Soil Sci. 32,
351-363.
HARRISONJ.B. & BERKHEISERV.E. (1982) Anion interaction with freshly precipitated hydrous iron oxides.
Clays Clay Miner. 30, 97-102.
ROCaESTERC.H. & TOPrIAMS.A. (1979)Infrared studies of the adsorption of probe moleculesonto the surface
of goethite. J. Chem. Soc., Faraday Trans. I 75, 872-882.
RUSSELLJ.D., PATERSONE., FRASERA.R. & FARMERV.C. (1975) Adsorption of carbon dioxide on goethite
(ct-FeOOH) surfaces, and its implications for anion adsorption. J. Chem. Soc., Faraday Trans. 1 71, 16231630.
SCnWERTMA~ U. (1959) Uber die Synthese definierter Eisenoxyde unter verschiedenen Bedingungen.
Z. anorg, allg. Chemic 298, 337-348.
SCHWgRTMANNU. & FITZPATRICKR.W. (1977)Occurrence of lepidocrocite and its association with goethite in
Natal soils. Soil Sci. Soc. Am. J. 41, 1013-1018.
TAYLORR.M. & MCKENZmR.M. (1980) The influence of aluminum on iron oxides. VI. The formation of
Fe(II)-AI(III) hydroxychlorides,-sulfates, and -carbonates as new members of the pyroaurite group and
their significance in soils. Clays Clay Miner. 28, 179-187.
YAPP C.J. (1987) A possible goethite-iron(III)carbonate solid solution and the determination of CO/partial
pressures in low-temperature geological systems. Chem. Geol. 64, 259-268.