Full text of DOI:10.1557/JMR.2000.0135

Reaction pathways at the iron–microspherical silica interface:
Mechanistic aspects of the formation of target iron
oxide phases
Sivarajan Ramesha)
Department of Chemistry, Bar-Ilan University, Ramat Gan-52900, Israel
Israel Felner
Racah Institute of Physics, Hebrew University, Jerusalem 91904, Israel
Yuri Koltypin and Aharon Gedankenb)
Department of Chemistry, Bar-Ilan University, Ramat Gan-52900, Israel
(Received 24 September 1999, accepted 29 December 1999)
Oxidative hydrolysis of elemental iron nanoclusters on hydroxylated surfaces such as
silica or alumina is known to be influenced by the degree of hydration of the surface.
The understanding and control of this process is crucial in the synthesis of iron oxide
coated silica microspheres with a desired magnetic property. The hydrolysis of iron
nanoparticles followed by heat treatment in the case of a hydrated microspherical silica
surface results in the formation of maghemite (␥–Fe2O3), whereas a dehydrated surface
yielded hematite (␣–Fe2O3) nanoparticles. The influence of adsorbed water on the
formation of intermediate iron oxides/oxidehydroxides and the mechanistic aspects of
their subsequent thermal dehydration iron oxide phases were investigated by
thermogravimetric analysis, Fourier transform infrared, and Mössbauer spectroscopies.
The reactions on both the hydrated and the dehydrated surfaces were found to proceed
through the formation of an x-ray amorphous lepidocrocite [␥–FeO(OH)] intermediate
and its subsequent dehydration to maghemite (␥–Fe2O3). Maghemite to hematite
transformation was readily facilitated only on a dry silica surface. The retardation of
the lepidocrocite → maghemite → hematite transformation in the case of a hydrated
silica surface is suggested to arise from strong hydrogen-bonded interactions between
the substrate silica and the adsorbed nanoparticles.
I. INTRODUCTION
in the presence of ammonia or an organic base as a
morphological catalyst is known to yield microspherical
silica with a narrow size distribution over a wider range.
The surface of the microspheres consists of single and
geminal silanols, silanols connected by an extensive
hydrogen-bonded network, and adsorbed water.
Hydroxylated oxide surfaces tend to react with tran-
sition metal clusters in their elemental state, with an
“oxidizing power” proportionate to the degree of hy-
droxylation.9–16 In general, zerovalent metals formed by
the thermal decomposition of adsorbed metal carbonyls
react with the surface hydroxyls, liberating hydrogen as
per the equation:14–16
The chemistry of the silica surface is probably the
most studied interfacial phenomenon over the past few
decades and still continues to exhibit rich variations in
terms of the surface chemical properties.1–6 The degree
of hydroxylation and hence the reactivity of the silica
surface are highly influenced by the origin of the silica.
Sol-gel-derived silica is normally obtained by the hy-
drolysis of an alkoxide, catalyzed by an acid or a base.
Acid-catalyzed hydrolysis proceeds through an electro-
philic reaction to an alcogel, whereas base-catalyzed re-
actions follow a nucleophilic pathway, leading to the
formation of hydrated silica powders.7 Stober et al.’s
method8 of hydrolysis of tetraethyl orthosilicate (TEOS)
M(CO)j + n(␴–OH) →
(␴–O−)nMn+ + (n/2)H2 + jCO
.
(1)
We have previously investigated the adsorption and re-
activity of elemental metal and metal oxide clusters such
as nickel,17,18 cobalt,19 iron,20 and rare-earth oxides21 on
the surface of microspherical silica. While nickel and
a)Present address: Department of Materials Science and Engineer-
ing, Bard Hall, Cornell University, Ithaca, New York 14853.
b)Address all correspondence to this author.
e-mail: gedanken@ashur.cc.biu.ac.il
944
J. Mater. Res., Vol. 15, No. 4, Apr 2000
© 2000 Materials Research Society

S. Ramesh et al.: Reaction pathways at the iron–microspherical silica interface
cobalt could be retained in their elemental form on hy-
Thermogravimetric analyses (TGA) in the temperature
range 25–900 °C under flowing nitrogen were carried out
employing a Mettler TG-50 analyzer. FTIR spectra were
recorded on an Impact 410 FTIR spectrometer (Nicolet,
U.S.A.) by a KBr disc method (2% by weight of sample
in 200 mg of KBr). Mössbauer spectroscopic investiga-
tions were carried out at 90 and 300 K, using a conven-
tional constant-acceleration spectrometer. The 57Fe
Mössbauer spectra were measured with a 50 mCi
57Co:Rh source, and the spectra were least-squares fitted
with two subspectra. The isomer shift values are relative
to Fe metal at 300 K.
drated silica even at elevated temperatures, iron nanopar-
ticles were found to be reactive toward the silica surface.
Instead of the expected elemental iron, we have ob-
served the formation of various iron oxides.20 In particu-
lar the formation of maghemite (␥–Fe2O3) and hematite
(␣–Fe2O3) nanoparticles were observed.
Two distinct pathways are likely to lead to such re-
sults: (i) different intermediate oxidehydroxides are
formed on differentially hydroxylated surfaces and they
undergo thermal dehydration in a subsequent step lead-
ing to distinct iron oxide phases. (ii) An identical inter-
mediate oxidehydroxide is formed on both surfaces
followed by a different thermal dehydration pathway to
lead to distinct iron oxide phases.
III. RESULTS
Extensive details on the synthesis, structure, and mor-
phological properties of the FeOx/SiO2 samples are
available elsewhere.20 Only the results relevant to the
investigation of mechanistic aspects will be presented
here. After sonication, as-prepared products did not show
any characteristic XRD peaks and were x-ray amor-
phous. Only after heat treatment did the products result in
distinct nanocrystalline products. The iron oxide phase in
the case of silica 1 was identified as ␥–Fe2O3/Fe3O4,
whereas the XRD indicated the formation of ␣–Fe2O3 in
the case of silica 2. Figure 1 shows the TEM micrograph
of the as-deposited iron product adhering to the silica 1.
In fact the synthesis and properties of iron oxide–silica
nanocomposites have been investigated earlier.22–27
Quite a few investigations28–30 have even focused on the
properties of the interfaces and phase transformations in
the iron oxide phases. In particular, an earlier report by
Lund and Deumesic30 described the formation of mag-
hemite and hematite phases in a silica gel matrix depen-
dent on the processing conditions. In this paper we report
the results of our experimental investigations on this as-
pect. Fourier transform infrared (FTIR), Mössbauer, and
thermogravimetric analysis techniques were employed to
address the specific problems outlined above.
II. EXPERIMENTAL
Details on the synthesis of the substrate microspheres
and the sonochemical deposition procedure have been
extensively discussed elsewhere,20 and only the neces-
sary outline is given here. Amorphous submicrospheres
of silica in the size range of 200–250 nm were synthe-
sized by a base-catalyzed hydrolysis of TEOS. The
spherical silica particles obtained as such, without any
heat treatment, will be referred to henceforth as silica 1.
A portion of the as-prepared silica was heated in vacuum
(10−5 torr) at 400 °C for 3 h to remove chemisorbed wa-
ter and other labile species. This sample will be referred
to as silica 2. Deposition of amorphous iron on the silica
spheres was carried out by sonicating a slurry of silica
and Fe(CO)5 in decalin with a direct immersion titanium
horn (Vibracell, 20 kHz, 100 W/cm2) under flowing ar-
gon. The solid products obtained after sonication were
washed with pentane, dried under vacuum, and heat
treated at a temperature of 400 °C under flowing argon
for 3 h. As-prepared and heat-treated samples were char-
acterized for structure and morphology by x-ray diffrac-
tion (XRD) and transmission electron microscopy
(TEM), respectively, as described in detail elsewhere.
Elemental composition was determined by energy dis-
persive x-ray analysis (EDX), employing a JEOL scan-
ning electron microscope.
FIG. 1. TEM micrograph of the as-deposited iron product adhering to
microspherical silica surface in the case of silica 1.
J. Mater. Res., Vol. 15, No. 4, Apr 2000
945

S. Ramesh et al.: Reaction pathways at the iron–microspherical silica interface
The iron product consisted of spongy aggregates with
particle sizes in the range 5–10 nm. The EDAX profile of
the (␥–Fe2O3/Fe3O4)–SiO2 sample was measured and
the iron content estimated as 9.1%. The TGA profiles
for the substrate silica spheres and as-prepared amor-
phous iron products deposited on silica 1 and silica 2
showed a total weight loss of 16.6%, 17.7%, and 15.1%,
respectively. Bare substrate silica spheres and iron prod-
ucts on silica 1 showed a rapid weight loss below 200 °C,
whereas weight loss was slow in the case of silica 2.
Another significant feature was that the silica micro-
spheres coated with iron products showed weight loss up
to 800 °C, whereas no weight loss was observed in the
bare substrate silica beyond 600 °C. The FTIR spectra of
the bare substrate silica and the as-prepared iron products
on silica 1 and silica 2 are shown in Figs. 2(a)–2(c),
respectively. In the case of spectra (b) and (c) there is a
sharp and strong absorption around 1100 cm−1, which is
absent in the case of bare substrate silica spheres. The
absorption region in the range 2500–4000 cm−1 provides
information on the nature of free surface silanols,
hydrogen-bonded silanols, and adsorbed water. Hence
the absorption spectra in this range were resolved into the
component peaks by fitting to multiple Gaussian profiles.
The corresponding results for the substrate silica and the
iron products on silica 1 and on silica 2 are shown in
Figs. 3(a)–3(c), respectively.
The Mössbauer spectra at 90 K of the as-prepared and
crystallized iron products deposited on silica 1 are shown
in Figs. 4(a) and 4(b), respectively. The Mössbauer spec-
tra and parameters of the as-prepared iron product ob-
tained in the case of silica 2 were identical to those in the
case of silica 1 [Fig. 4(a)]. The as-deposited iron products
in the case of both silica 1 and silica 2 exhibited a broad
doublet in the central part of the spectrum, indicating the
absence of any long-range magnetic ordering. The hy-
perfine parameters of the major doublet at 90 K were
IS ס 0.46(1), with a quadrupole splitting of ⌬ ס eqQ/2 ס
0.78(1) and for the minor one IS ס 0.44(1) mm/s and
⌬ ס 1.30(1) mm/s, respectively, with a common line
width of 0.49(1) mm/s.
FIG. 2. FTIR spectra in the range 400–1500 cm−1 for (a) as-prepared
substrate silica microspheres, (b) as-prepared amorphous iron products
deposited on silica 1, and (c) as-prepared amorphous iron products
deposited on silica 2.
IV. DISCUSSION
ethyl silicate derived silica. This hydrogen-bonded net-
work obstructs the probing of the surface by an inert gas
and results in a low apparent surface area. The density of
hydroxyls on a hydroxylated surface can be calculated
from the weight loss and Brunauer–Emmett–Teller
(BET) surface area, using the following expression:31
The substrate silica microspheres showed a cumulative
weight loss of 16.6% over two distinct steps of 40–175
and 200–600 °C. The surface of the microspherical silica
consists of a small proportion of free silanols and a larger
proportion of silanols heavily hydrogen-bonded among
them and adsorbed water molecules. The thermal desorp-
tion of the adsorbed water can be connected to the weight
loss below 200 °C. The linear weight loss over a tem-
perature range 200–600 °C can be associated with the
expulsion of water within the pores. The heavy hydrogen-
bonding among the silanols is an important property of
nOH ס 66911[⌬m/(m0As)]
,
(2)
where ⌬m is the weight loss, m0 is the initial weight, and
As is the surface area expressed in m2/g. The conversion
factor of 66,991 yields the value of nOH in nm−2. A value
946
J. Mater. Res., Vol. 15, No. 4, Apr 2000

S. Ramesh et al.: Reaction pathways at the iron–microspherical silica interface
of 989/nm2 was obtained for the bare silica microspheres.
the probing gas used in the BET experiment resulting
in smaller surface area values. For this reason, the sur-
face area measured by BET for the microspherical silica
almost conforms to the geometrical outer surface area
calculated for uniform silica spheres (11.6 m2/g for
spheres of with a diameter of 250 nm and a density of
2.08 g/cm3).
This unusually high value, when compared to typical
values of 12–14 nm−2 observed in the case of aerosil and
pyrogenic silica,31 confirms the presence of extensive
hydrogen bonding among silanols and between silanols
and adsorbed water. A similar, high silanol density value
(700 nm−2) for ethyl silicate derived silica was also re-
ported by Burneau et al.,31 who attributed the discrep-
ancy to inner pore silanols, heavily hydrogen bonded
among themselves and with adsorbed water. Such a net-
work prevents the measurement of the inner pore area by
The iron product deposited on silica 1 also showed a
weight loss pattern similar to that of the bare silica sub-
strates, with two distinct stages of weight loss, except
that the sample continued to lose weight up to a tempera-
ture of 800 °C as opposed to a temperature of 600 °C in
the case of bare silica substrates. An excess weight loss
of 1.1% over the bare silica substrates is negligible if
the adsorbed decalin and Fe(CO)5 are taken into account.
This is very important, since it confirms that the iron
clusters are deposited on the silica surface in the elemen-
tal form and are oxidized/hydrolyzed with a sole avail-
able source of surface silanols/adsorbed water. The total
loss of water in this material however is not very differ-
ent from the bare substrate silica spheres as the interme-
diate iron oxidehydroxide also loses water in the
measured temperature range to form iron oxide. The iron
products deposited on silica 2 showed an almost linear
weight loss from room temperature up to a temperature
of 800 °C. A lower weight loss (even after considering
FIG. 3. FTIR spectra in the range 2500–4000 cm−1 and the compo-
nent peaks obtained by fitting to multiple Gaussian profiles for (a)
as-prepared substrate silica microspheres, (b) as-prepared amorphous
iron products deposited on silica 1, and (c) as-prepared amorphous iron
products deposited on silica 2. The dotted lines represent the experi-
mental data.
FIG. 4. Mössbauer spectra recorded at 90 K of (a) the as-obtained iron
product in the case of silica 1 and (b) the target iron product obtained
after heat treatment in the case of silica 1.
J. Mater. Res., Vol. 15, No. 4, Apr 2000
947

S. Ramesh et al.: Reaction pathways at the iron–microspherical silica interface
the adsorbed impurities of decalin and iron carbonyl)
present investigation as well. The component peak at
3214 cm−1 in the bare silica substrate sphere shifted to
3241 and 3257 cm−1, respectively, in the case of the
Fe/silica 1 and Fe/silica 2 samples.
compared to bare silica substrates is expected, since heat-
ing the silica sample under vacuum at 400 °C is known to
remove all the adsorbed water from the surface, thus
accounting for the absence of a rapid weight loss below
200 °C. However, heating under these conditions has not
resulted in the complete removal of hydroxyls from the
surface as can be determined from the total weight loss.
Although the reaction of iron clusters with surface
silanols proceed through an oxidehydroxide intermedi-
ate, an accurate estimation of the amount of water used in
the case of silica 1 and silica 2 could not be made from
TGA measurements, as the contribution of weight loss
due to the adsorbed solvent hydrocarbon could not be
separated.
The iron oxides and hydroxides have been charac-
terized by a wide variety of techniques including infra-
red (IR) spectroscopy.32,33 Of the IR spectra of various
iron oxidehydroxides and oxides listed by Cornell and
Schwertmann,34 only ferrihydrite and lepidocrocite
[␥–FeO(OH)] show sharp absorption above 1000 cm−1,
in the range of 400–3000 cm−1. In fact, the B2u and B3u
vibrational modes corresponding to the in-plane bending
of –OH in lepidocrocite, absorb at 1150 and 1018 cm−1,
respectively. The sharp absorption at 1100 cm−1 in the
case of iron products on silica 1 and silica 2 [Figs. 2(b)
and 2(c)] suggest that lepidocrocite is the most likely
intermediate in both cases.
The broad IR absorption in the range of 2800–
4000 cm−1 can be resolved to obtain important informa-
tion about the changes in the silanol structure on the
surface of silica. Of the various iron oxidehydroxides,
goethite [␣–FeO(OH)], akagenite, ␤–FeO(OH)], lepido-
crocite [␥–FeO(OH)], and feroxhyte [␦–FeO(OH)] ab-
sorb in the above region.32,33 This, when coupled with
the fact that hydrogen-bonded silanols themselves absorb
in this region, further complicates the analysis. Kondo
et al.35 have investigated the thermal behavior of the sila-
nol absorption range of ethyl silicate derived silicas. They
resolved the silanol absorption range (3000–3900 cm−1)
into various component peaks and observed that their
intensities varied as a function of temperature. Two of
the prominent component peaks observed in the present
study at 3214 and 3478 cm−1, respectively, matched
closely with the peaks assigned to hydrogen-bonded si-
lanols by Kondo et al.35 They assigned the component
peaks at 3260 and 3470 cm−1 to silanol networks linked
with hydrogen bonds of varying enthalpies of dehydra-
tion at 8.9 and 8.1 kcal/mol, respectively. The band at
3260 cm−1 showed a higher sensitivity to dehydra-
tion and showed a large blue shift, i.e., from 3260 to
3390 cm−1, on dehydration up to 300 °C, whereas the
peak at 3478 cm−1 was less sensitive to dehydration tem-
perature and remained virtually unchanged for treatments
up to 400 °C. A similar observation was made in the
The silica sample described by Kondo et al.35 showed
a shift of about 130 cm−1 on thermally induced breaking
of the hydrogen-bonded network followed by dehydra-
tion. Whereas a blue shift of 27 cm−1 of this sensitive
mode that could be brought about by mere adsorption of
iron atoms at room temperature stands to confirm the
extreme reactivity of iron clusters and the oxidative hy-
drolysis at the interface. Besides, 27 cm−1 is a significant
blue shift when compared to a shift of only 46 cm−1 in
the case of thermally treated silica microspheres. As ex-
pected, the component at 3470 cm−1 did not show any
significant shift. To make a rough estimate of the de-
hydration process the relative intensity of the sensitive
absorption 3214 cm−1 in comparison to the one at
3472 cm−1 was calculated. Though the dehydration due
to vacuum treatment and reaction with iron clusters could
not be separated in the case of Fe/silica 2, a lowest rela-
tive intensity of 0.77 confirmed that the amount of mo-
lecularly adsorbed water available for iron hydrolysis in
the case of silica 2 was definitely lower than in the case
of Fe/silica 1.
The possibility of the presence of ␣, ␤, ␥, and ␦
iron oxide hydroxides was examined using Ishikawa
et al.’s32,33 assignment of the infrared absorption of iron
oxidehydroxides in the 2800–3800 cm−1 range, where
silica and iron oxidehydroxides have strongly overlap-
ping absorptions. The ␣ and ␥ phases could not be as-
certained, as their strongest absorptions at 3150 and
3160 cm−1, respectively, could not be resolved. How-
ever, the ␤ phase has the strongest absorption at
3480 cm−1, and hence, a stronger relative intensity of the
3472 cm−1 component with respect to the 3257 cm−1
component in the case of Fe/silica 2 system suggests, at
first glance, that the intermediate oxidehydroxide in this
case may be ␤–FeO(OH). This assignment is ambiguous,
as the relative intensity is also affected by dehydration as
outlined earlier and is in clear contradiction to the IR
absorption at 1100 cm−1 [since ␤–FeO(OH) does not
show an absorption at this frequency]. In spite of this, the
presence of ␤–FeO(OH) cannot be ruled out at this stage.
␦–FeO(OH), on the other hand, is known to absorb
strongly at 3160 cm−1. If this were the intermediate hy-
droxide, it should have resulted in a component peak at
this frequency in the case of Fe/silica 1 and Fe/silica 2.
The absence of any notable difference among the three
spectra in this range indicates that the intermediate oxide
hydroxide is unlikely to be ␦–FeO(OH).
The Mössbauer spectrum of the crystallized product in
the case of Fe/silica 1 [spectrum 4(b)] showed a hyper-
fine magnetic splitting as clear evidence of long-range
magnetic ordering at low temperatures. The spectrum,
948
J. Mater. Res., Vol. 15, No. 4, Apr 2000

S. Ramesh et al.: Reaction pathways at the iron–microspherical silica interface
however, consisted of an additional doublet (7%), aris-
likely to be lepidocrocite ␥–FeO(OH) since this con-
forms to the descriptions of Fe3+ in a high-spin state and
oxidehydroxide intermediate showing a sharp IR absorp-
tion at 1100 cm−1. Hence a difference in the thermal
dehydration of the intermediate over the wet and dry
silica accounts for different end products obtained in the
case of silica 1 and silica 2.
ing from a small fraction of the precursor hydroxide left
uncrystallized. The interpretation of the spectrum yields
a magnetic hyperfine field of Heff ס 489(2) kOe, an
isomer shift of 0.39(1) mm/s, and a line width of
0.63 mm/s. These values are closer to the measured hy-
perfine values of ␥–Fe2O3 (Heff ס 520 kOe; IS ס 0.24
mm with an effective quadrupolar splitting of −0.20).
This discrepancy arises mainly as a result of smaller
portion of the uncrystallized precursor. Nevertheless it
suffices to confirm the identity of the crystalline phase as
␥–Fe2O3 as opposed to Fe3O4.
The Mössbauer spectra of the amorphous materials in
the case of Fe/silica 1 and Fe/silica 2 were identical
within the range of experimental uncertainties. The main
information derived from their analysis revealed the
presence of two quadrupolar doublets with a relative ra-
tio of 2:1, corresponding to inequivalent sites in the pre-
cursor material. The corresponding values listed in the
Results section are typical of Fe3+ in the high-spin state.
This can be attributed to the presence of Fe3+ in an x-ray
amorphous form of FeO(OH) intermediate.
The thermal dehydration of pure iron oxidehydroxides
to various iron oxides has been extensively studied, and
exhaustive documentation has been carried out by Cor-
nell and Schwertmann.34 On the basis of this work, a
schematic diagram showing the possible thermal dehy-
dration products during the dehydration of iron oxidehy-
droxides immobilized on silica is shown in Fig. 5. The ␣,
␤, and ␦ forms of iron oxidehydroxides and ferrihydrite
on thermal dehydration normally transform into hema-
tite, without going through an intermediate iron oxide
phase. On the other hand, lepidocrocite transforms into
hematite through an intermediate ␥–Fe2O3 and is known
to be sensitive to moisture content. The transformation of
␣–Fe2O3 is favored under dry conditions.36,37
Lund and Dumesic30 investigated the strong oxide-
support interactions between iron oxide supported on
silica gel prepared by the calcination or reduction of
silica gel impregnated with iron nitrate. They observed
The results of the TGA, FTIR, and Mössbauer inves-
tigations suggest that the intermediate species in the case
of both Fe/silica 1 and Fe/silica 2 are identical and most
FIG. 5. Schematic diagram showing the possible products that would result from the thermal dehydration of various iron oxyhydroxide inter-
mediates. (Based on Cornell and Schwertmann.34) Note that the ␣, ␤, ␦, and ferrihydrite phases on dehydration normally transform into hematite,
without going through any intermediate iron oxide phase. Lepidocrocite transforms into hematite through an intermediate mag-
hemite (␥–Fe2O3) phase. This is a crucial process that distinguishes between lepidocrocite and ferrihydrite, which are otherwise indistinguishable
in the present study.
J. Mater. Res., Vol. 15, No. 4, Apr 2000
949

S. Ramesh et al.: Reaction pathways at the iron–microspherical silica interface
5. H.E. Bergna, L.E. Firment, and D.G. Swartzfager, Advances in
retardation of the maghemite to hematite transition in
certain samples. This was attributed to strong oxide–
silica interactions arising from a substitution of impurity
cations. However, our experimental observations indi-
cate that the maghemite to hematite transformation could
be retarded due to the presence of strong hydrogen-
bonded interactions between the maghemite phase and
silica surface in the case of silica 1.
Such a view is not without support as can be seen in
a recent report by del Monte et al.,38 wherein the
authors synthesized iron oxide–silica nanocomposites
by the hydrolysis of TEOS by FeCl3 и 6H2O and
Fe(NO3)3 и 9H2O salts in ethanol medium. The interest-
ing aspect of this work lies in the fact that only the water
of hydration in the iron salts was used for the hydrolysis
of TEOS. In the case of Fe(NO3)3 и 9H2O ␥–Fe2O3 was
the final iron oxide product whereas hydrolysis with
FeCl3 и 6H2O resulted in the formation of hematite.
Though the mechanism proposed by del Monte et al.
involved redox reactions with impurity hydrocarbons,
they recognized that the reaction with chloride salt,
where the water of hydration was retained, resulted in
hematite compared to Fe(NO3)3 и 9H2O, where water
was available for the hydrolysis resulting in a maghemite
end product.
Chemistry Series 234 (American Chemical Society, Washington,
DC, 1994).
6. I.S. Chuang and G.E. Maciel, J. Am. Chem. Soc. 118, 401 (1996).
7. D. Seagal, Chemical Synthesis of Advanced Ceramic Materials
(Cambridge University Press, Cambridge, United Kingdom, 1989).
8. W. Stober, A. Fink, and E. Bohn, J. Colloid Interface Sci. 26, 62
(1968).
9. A.K. Smith, F. Hugues, A. Theolier, J.M. Basset, R. Ugo,
G.M. Zadherigi, J.L. Bilhou, V. Bilhou-Bougnal, and W.F. Gray-
don, Inorg. Chem. 18, 3104 (1979).
10. F. Solymosi and M. Pasztor, J. Phys. Chem. 90, 5312 (1986).
11. G.S. McNulty, K. Cannon, and J. Schwartz, Inorg. Chem. 25,
2919 (1986).
12. P. Basu, D. Panyotov, and J.T. Yates, Jr., J. Am. Chem. Soc. 110,
2074 (1988).
13. S.K. Purnell, X. Xu, D.W. Goodman, and B.C. Gates, J. Phys.
Chem. 98, 4076 (1994).
14. A. Brenner, D.A. Hucul, and S.J. Hardwick, Inorg. Chem. 18,
1478 (1979).
15. A. Brenner and D.A. Hucul, Inorg. Chem. 18, 2836 (1979).
16. D.A. Hucul and A. Brenner, J. Phys. Chem. 85, 496 (1981).
17. S. Ramesh, Yu. Koltypin, R. Prozorov, and A. Gedanken, Chem.
Mater. 9, 546 (1997).
18. S. Ramesh, Y. Cohen, D. Aurbach, and A. Gedanken, Chem.
Phys. Lett. 287, 461 (1998).
19. S. Ramesh, Y. Cohen, R. Prozorov, K.V.P.M. Shafi, D. Aurbach,
and A. Gedanken, J. Phys. Chem. B 102, 10234 (1998).
20. S. Ramesh, R. Prozorov, and A. Gedanken, Chem. Mater. 9, 2996
(1997).
21. A. Patra, E. Sominska, S. Ramesh, Yu. Koltypin, Z. Zhong,
H. Minti, R. Reisfeld, and A. Gedanken, J. Phys. Chem. B 103,
3361 (1999).
V. CONCLUSIONS
22. A.M. Homola, M.R. Lorenz, C.J. Mastrangelo, and T.L. Tilburg,
IEEE Trans. Magn. MAG-22, 716 (1986).
23. A.P. Philipse, M.P.B. van Bruggen, and C. Padhmamanoharan,
Langmuir 10, 92 (1994).
24. M.N. Asuha, J. Mater. Sci. Lett. 12, 1705 (1993).
25. C. Chaneac, E. Tronc, and J.P. Jolivet, J. Mater. Chem. 6, 1905
(1996).
26. L. Zhang, G.C. Papaefthymiou, and J.Y. Ying, J. Appl. Phys. 81,
6892 (1997).
27. Y. Liu, A. Wang, and R.O. Claus, Appl. Phys. Lett. 71, 2265 (1997).
28. D. Niznansky, J.L. Rehspringer, and M. Drillon, IEEE Trans.
Magn. 30, 821 (1994).
29. G. Ennas, A. Musinu, G. Piccaluga, D. Zedda, D. Gatteschi,
C. Sangregario, J.L. Stanger, G. Concas, and G. Spano, Chem.
Mater. 10, 495 (1998).
The oxidation and hydrolysis of elemental iron on the
surface of microspherical silica surface are influenced by
the extent of hydration of the surface. While a totally
hydrated surface results in the formation of maghemite
(␥–Fe2O3), partial dehydration of the silica surface leads
to the formation of hematite (␣–Fe2O3) as the final iron
oxide phase. Lepidocrocite ␥–FeO(OH) has been shown
to be the intermediate oxidehydroxide species in both
cases. While the transformation of ␥–Fe2O3 into hematite
on a previously dehydrated silica surface is readily ac-
complished, the presence of water retards this transfor-
mation under identical conditions of heat treatment.
30. C.R.F. Lund and J.A. Dumesic, J. Phys. Chem. 85, 3175 (1981).
31. A. Burneau, O. Barres, J.P. Gallas, and J.C. Lavalley, Langmuir 6,
1364 (1990).
32. T. Ishikawa, S. Nitta, and S. Kondo, J. Chem. Soc., Faraday
Trans. 1 82, 2401 (1986).
33. T. Ishikawa, W.Y. Cai, and K. Kandori, J. Chem. Soc., Faraday
Trans. 88, 1173 (1992).
34. R.M. Cornell and U. Schwertmann, The Iron Oxides: Structure,
Properties, Reactions, Occurrence and Uses (VCH Verlagsgesell-
schaft, Weinheim, Germany, 1996).
ACKNOWLEDGMENTS
A.G. thanks the Israel Ministry of Science for support-
ing this research through the grants for infrastructure.
Yu.K. thanks the Ministry of Absorption and the Center
for Absorption in Science for financial support. The au-
thors thank Dr. Shifra Hochberg for editorial assistance.
35. S. Kondo, M. Muroya, and K. Fujii, Bull. Chem. Soc. Jpn. 47, 553
(1974).
36. H. Naono and K. Nakai, J. Colloid Interface Sci. 128, 146 (1989).
37. E. Tronc, J.P. Jolivet, and J. Livage, Hyperfine Interact. 54, 737
(1990).
REFERENCES
1. R.K. Iler, The Chemistry of Silica, Solubility, Polymerization,
Colloid and Surface Properties and Biochemistry (Wiley-
Interscience, New York, 1979).
2. L.C. Klein, Annu. Rev. Mater. Sci. 15, 227 (1985).
3. L.L. Hench and J.K. West, Chem. Rev. 33, 90 (1990).
4. B.A. Morrow and A.J. McFarlan, Langmuir 7, 1695 (1991).
38. F. del Monte, M.P. Morales, D. Levy, A. Fernandez, M. Ocana,
A. Roig, E. Molins, K.O. Grady, and C.J. Serna, Langmuir 13,
3627 (1997).
950
J. Mater. Res., Vol. 15, No. 4, Apr 2000