One-shot versus stepwise gas-solid synthesis of iron trifluoride: Investigation of pure molecular F2 fluorination of chloride precursors

Article abstract of DOI:10.1039/c3ce27033e

Iron trifluoride has been known for a long time and here we provide new insights into its detailed synthesis. Anhydrous iron trifluorides with different structures have been synthesized via one-shot and stepwise gas-solid fluorination under a gaseous flow of molecular fluorine from 150°C. The degree of hydration of the precursors has a direct influence on the crystalline phase obtained: one-shot fluorination of FeCl₂·4H₂O leads to the hexagonal tungsten bronze-type FeF₃ while fluorination of dehydrated iron chlorides gives the rhombohedral FeF₃ with traces of oxy(hydroxy)fluorides. A stepwise procedure at 150, 250 and 350°C tends to lead to pure phases and avoid the crystallization of oxy(hydroxy)fluorides in the final powder.

Full text of DOI:10.1039/c3ce27033e

CrystEngComm
View Article Online
View Journal | View Issue
PAPER
One-shot versus stepwise gas–solid synthesis of iron
trifluoride: investigation of pure molecular F2
fluorination of chloride precursors3
Cite this: CrystEngComm, 2013, 15,
664
3
a
a
a
a
a
Nicolas Louvain, Ahmed Fakhry, Pierre Bonnet,* Malika El-Ghozzi, Katia Gu e´ rin,
b
b
c
Moulay-Tahar Sougrati, Jean-Claude Jumas and Patrick Willmann
Iron trifluoride has been known for a long time and here we provide new insights into its detailed
synthesis. Anhydrous iron trifluorides with different structures have been synthesized via one-shot and
stepwise gas–solid fluorination under a gaseous flow of molecular fluorine from 150 uC. The degree of
hydration of the precursors has a direct influence on the crystalline phase obtained: one-shot fluorination
Received 15th December 2012,
Accepted 4th March 2013
of FeCl
chlorides gives the rhombohedral FeF
50 and 350 uC tends to lead to pure phases and avoid the crystallization of oxy(hydroxy)fluorides in the
final powder.
2 2 3
?4H O leads to the hexagonal tungsten bronze-type FeF while fluorination of dehydrated iron
3
with traces of oxy(hydroxy)fluorides. A stepwise procedure at 150,
DOI: 10.1039/c3ce27033e
2
www.rsc.org/crystengcomm
candidates and is, by far, the phase that attracted the most
attention from the early studies that reported capacities as high
Introduction
Metal fluorides are low cost inorganic compounds that have
been studied for their industrial applications such as smelting
in metallurgy, uranium enrichment process or optical material
21
15
as 140 mA h g at 3.4 V. It is a well-known inorganic fluoride
16
that is traditionally obtained as bulk from either iron chloride
17
or nitrate under anhydrous or aqueous HF; however, three
1
production. This applied development of the synthesis of
main reasons can be brought to mind in order to explain the
absence of industrialization as cathode materials:
inorganic fluorides allowed intense fundamental research in
the 1970s and 1980s as regards their electronic, and more
particularly, magnetic properties. In the last decades, research
subjects have been shifted toward the nanostructuration of
inorganic fluorides, a recently reviewed field, due to potential
(
3
i) like most inorganic fluorides, FeF behaves as an
2
18
insulator with a wide band-gap.
19
(
(
ii) FeF
iii) FeF
3
is sensitive to air humidity.
is mostly obtained as hydrated (please refer to the
3
3
4
–6
7
8–10
applications in catalysis, gas storage, photonics,
nanome-
next section).
11
12
13
decine,
tribology
and electrical energy conversion.
While the first and, to some extent, second points have
been addressed in the past by coating iron fluoride particles
with carbon, the synthesis of anhydrous iron trifluoride
remains mostly unexplored. To the best of our knowledge,
no account of its synthesis under molecular fluorine has been
reported to date even though it appears to be the most
Concerning the energy conversion, they are potential candidate
materials as cathodes for lithium-ion batteries through both
insertion and conversion reactions and they could improve the
theoretical specific capacity of the positive electrode (usually
made from lithiated metal oxides) from 274 mA h g (1 e ) to
more than 700 mA h g (3 e ).
21
2
2
1
2 13,14
Among the MF
3
fluorides (M
2
practical way of obtaining anhydrous ferric fluoride species.
=
metal), iron trifluoride, FeF , is one of the most promising
3
Generally, direct gas–solid fluorination uses anhydrous
2
,20
chlorides as precursors to obtain anhydrous fluorides.
Most
a
Institut de Chimie de Clermont-Ferrand CNRS UMR6296, Clermont Universit ´e ,
Universit ´e Blaise Pascal, 24 avenue des Landais BP80026, 63171 Aubi `e re, France.
E-mail: pierre.m.bonnet@univ-bpclermont.fr; Fax: +33 4 73 40 79 33;
Tel: +33 4 73 40 71 25
of the anhydrous transition metal chlorides, however, are very
21
hygroscopic and cannot be easily purified. On the other
hand, their hydrates appear to be a good alternative as
precursors but little is known about their usually complex
thermolysis and reactivity under fluorine gas. In this work, an
attempt is made to find the optimum conditions to synthesize
b
Institut Charles Gerhardt Montpellier, AIME, CNRS UMR5253, Universit´e
Montpellier 2, Place Eug `e ne Bataillon – CC 004, 34095 Montpellier cedex 05, France
c
Centre National d’Etudes Spatiales, Centre Spatial de Toulouse, 18 avenue Edouard
Belin, 31401 Toulouse cedex 9, France
anhydrous iron fluoride by direct F -gas fluorination through
2
3
Electronic supplementary information (ESI) available: Structural parameters of
the use of iron chlorides with different hydration degrees. A
systematic investigation of the kinetic processes of the
decomposition of the precursors and their thermal stability
iron fluorides and iron chlorides phases, thermogravimetric analysis,
M ¨o ssbauer spectroscopy and detailed fitting parameters. See DOI: 10.1039/
c3ce27033e
3
664 | CrystEngComm, 2013, 15, 3664–3671
This journal is ß The Royal Society of Chemistry 2013

View Article Online
CrystEngComm
Paper
has been undertaken in order to assess their reactivity, to
determine adequate reaction temperatures and to follow the
most efficient fluorination protocol, namely either a one-shot
a
Table 1 Crystallographic parameters of FeF
3
phases
b
r-FeF
3
HTB–FeF
3
Pyr–FeF
3
(
at a constant temperature) or stepwise (in a series of distinct
c
Space group
a, Å
b, Å
c, Å
a, u
b, u
R ¯3 c
Cmcm
Fd3m
stages of increasingly constant temperature) procedure.
5.362(1)
5.362(1)
5.362(1)
57.94(2)
57.94(2)
57.94(2)
103.86
2
7.423(3)
12.730(4)
7.526(3)
90
10.325(2)
10.325(2)
10.325(2)
90
90
90
Foreword on iron fluoride crystallographic phases
Traditionally, iron(III) fluoride is obtained by allowing anhy-
drous FeCl
apparatus until HCl evolution ceases, according to eqn (1).
90
90
to react with anhydrous HF in a gas–solid
c, u 3
3
V, Å
711.17
12
1100.7
16
Z
Pore Ø, Å
ICSD #
PDF #
Ref#
—
3.1
35359
01-076-1262
26
1–2.5
202047
00-38-1305
27
FeCl
3
+ 3HF = FeF
3
+ 3HCl
(1)
41120
00-033-0647
23
The reaction proceeds even at room temperature and yields
amorphous FeF , which needs to be heated up to 600 uC in
order to induce crystallization.
anhydrous FeF crystallizes in the rhombohedral space group
a
3
It should be noted that iron-based fluoride structures are
28 b
1
6,22
analogous to the aluminium and the titanium ones.
The cell is
Under these conditions,
c
given in its rhombohedral setting. In the ref. 25, although the
space group is indicated as P6/mmm, there is apparently no report of
the structural data and we indicated here the data of the hydrated
form from which it is derived.
3
R ¯3 c and its structure is isostructural with a distorted ReO -type
3
structure, which can also be described as an A-site deficient
perovskite, and consists of corner-sharing FeF6 octahedra
extended in all three directions of space (Fig. 1a and S1,
2
3
ESI
3
). Single crystals can be produced through hydrothermal
2
4
reaction of FeF ?3H O in concentrated HF at 390 uC.
3
2
Hydrated phases of iron fluorides have also been isolated,
structurally determined and characterized (please refer to the
The rhombohedral form of FeF , denoted r-FeF , is also
known as a-FeF . At least two other polymorphs and two
3
amorphous phases are known in the literature. Firstly, the
hydrothermal reaction of amorphous FeF and H O at 360 uC
leads to the hexagonal tungsten bronze (HTB) FeF ?0.33H O
3 2
that possesses relatively large channels (Ø y0.31 nm) along
the c axis where zeolitic water molecules are accommodated
3
3
ESI
Fig. S4, ESI
ferric chloride at room temperature (a) and 50 uC (b, ICSD
3
3 2
for a detailed summary). Most notably a- and b-FeF ?3H O
2
5
(
3
) are grown from a hydrofluoric acid solution of
3
2
29
3
0,31
#
14134, PDF #00-032-0464),
monohydrate FeF ?H O after vacuum treatment although clear
evidence have yet to be provided.
HTB–FeF is obtained as the hydrate FeF
ESI ) by solvothermal reactions, or by synthesis in fluori-
nated ionic liquid. Other hydrated forms of iron fluorides are
and are said to give the
3
2
2
4,32
The aforementioned
?0.33H O (Fig. S2,
2
6
(
Fig. 1b and S2, ESI
dehydrated without structural change to give anhydrous FeF3
with a HTB structure (HTB–FeF ). Secondly, a topotactic
oxidization of the mixed valence fluoride NH Fe F , which
3
). Above 120 uC, this compound can be
3
3
2
26
3
3
3
3
4
2
6
34
Fe
2
F
5
?7H
2 2 5 2
O (no ICSD, PDF(I) #00-045-0883), Fe F ?2H O
exhibits an ordered pyrochlore structure, in a boiling solution
of Br in acetonitrile at 81 uC leads to the pyrochlore-type FeF
3
5–37
(ICSD #1167 and 201737, PDF #00-049-1263, Fig. S5, ESI ),
3
2
3
3
8
Fe1.9F4.75?0.95H
2
O
(no ICSD, PDF(B) #00-028-0483)
and
Those
(
3
Pyr–FeF ), which structure consists of corner-sharing FeF6
39,40
3 8
Fe F ?2H O (ICSD #37140 and 38366, Fig. S6, ESI ).
2
3
octahedra creating 3D interconnected channels of hexagonal
compounds, obtained through solution-based methods, are
mixed-valence iron fluorides and are closely related to the open
pyrochlore or HTB structures.
From this short overview of the iron fluoride phases, it
should now be evident that direct F gas–solid fluorination has
2
7
section (Fig. 1c and S3, ESI
open with large channels. It is noteworthy that annealing of
the pyrochlore form of FeF led to structural transformations
to HTB–FeF3 and then r-FeF3 at 280 uC and 380 uC,
3
). The two former structures are
3
2
2
2
respectively. All the anhydrous FeF
3
phases have been
never been fully investigated for the synthesis of iron fluorides.
thoroughly investigated by means of M o¨ ssbauer spectroscopy
2
5
and magnetism. Their structural parameters are summar-
ized in Table 1.
Materials and methods
FeCl ?4H O (98%) and FeCl ?6H O (98+%) have been pur-
2
2
3
2
chased from Sigma-Aldrich and Prolabo, respectively, and
used without further purification. Pure molecular fluorine F
2
gas (98–99%, 27 bars) has been purchased from Solvay
Chemicals, Solvay Fluor GmbH, Germany. The dehydrated
precursors and fluorinated products have been stored and
handled in an argon-filled glovebox with a humidity level
below 0.05 ppm.
Fig. 1 Different structural phases of anhydrous FeF
11 1¯ ] direction (a), HTB–FeF
, viewed along the c axis (b), Pyr–FeF
the [101] direction (c).
3
:r-FeF
3
, viewed along the
[
3
3
, viewed along
This journal is ß The Royal Society of Chemistry 2013
CrystEngComm, 2013, 15, 3664–3671 | 3665

View Article Online
Paper
CrystEngComm
Dehydration of the precursors
X-ray powder diffraction analysis
X-ray powder diffraction (XRPD) experiments have been
carried out on PANalytical X-Pert Pro diffractometer
FeCl ?2H O and FeOCl–FeOOH have been obtained by
2
2
dehydration of FeCl
2
?4H
2
O at 120 uC and FeCl
3
?6H
2
O at 150
a
uC, respectively, under dynamic vacuum by using a B u¨ chi
drying glass oven for 15 h.
equipped with a diffracted beam monochromator Cu Ka
source and an X’celerator linear detector by using a home-
made measurement cell that allows the data to be collected
under a static inert argon atmosphere. The counting time was
200 s and the angle step size 0.0836u.
Fluorination experiments
FeCl
nated under pure molecular fluorine up to 350 uC. The direct
-gas fluorination process was performed in a dedicated
2 2 2 2
?4H O, FeCl ?2H O, and FeOCl–FeOOH have been fluori-
M o¨ ssbauer spectroscopy
F
2
4
1
fluorine equipment using special handling procedures.
A
M o¨ ssbauer spectroscopy has been measured on freshly
5
7
cylindrical nickel reactor of approx. 50 cm in length and
offering a volume of approx. 1 L was used for the reaction. The
heating was ensured by a horizontal tubular oven surrounding
the reactor, with a domain of homogeneous temperature of
approx. 10 cm. All samples were set in a passivated nickel boat.
A typical one-shot fluorination procedure can be described
as follows: approx. 250 mg of iron chlorides are placed on a
passivated nickel boat, and the powder surface exposed to the
prepared samples stored under argon. Fe M o¨ ssbauer spectra
were recorded in the constant acceleration mode and in
transmission geometry on a standard M o¨ ssbauer spectrometer
composed of electronic devices from Ortec and Wissel. A
5
7
Co(Rh) source with a nominal activity of 370 MBq was used.
The source and the absorber were always kept at room
temperature. The thickness of the absorbers is about 10 mg
2
2
cm . The isomer shift is given relative to an a-Fe standard at
2
reactive atmosphere is manually limited to 1–2 cm , which
room temperature.
induces a powder thickness of a few millimetres. The boat is
placed at the centre of the fluorination oven, which is flushed
for 2 h with nitrogen while the temperature is increased to 100
uC (heating ramp of approximately 10 uC min ). Then,
fluorine is introduced, and the temperature set to 350 uC.
After 15 h, fluorine flow is stopped and the reactor oven is
flushed with nitrogen until the system cools down to room
temperature.
Typical stepwise fluorination procedure: the pre-treatment
protocol is similar to the previous one. After flushing the oven
with nitrogen at 100 uC, fluorine is introduced and the
temperature set to 150 uC for 5 h. Then, the temperature is set
to 250 uC for 5 h and finally 350 uC for 5 h. Afterwards, the
reactor is allowed to cool down to room temperature under
nitrogen flow. Both one-shot and stepwise fluorination
procedures are summarized in Scheme 1.
Results and discussion
2
1
About iron chloride precursors and their thermal stability
Many transformations may occur when a sample is subjected
to heat under specific conditions: melting, sublimation, phase
transition or decomposition. In solid–gas reactions, it is
primordial to assess these solid-state transformations. Thus,
solid-state kinetic data about the iron chloride precursors are
of practical interest in order to rationalize their reactivity
under fluorine and anticipate the formation of one phase over
another.
3
Ferrous chloride tetrahydrate, FeCl ?4H O (Fig. S8, ESI ), is
2
2
reportedly thermally stable up to 103 uC. It yields the
dihydrate, FeCl ?2H O, when heated up to approximately 110
2
2
4
2
uC (Fig. S9, ESI
3
), and the remaining two water molecules are
removed at 158 uC and 212 uC, respectively, along with some
4
3
dehydrochlorination. In flowing nitrogen, the entire first
dehydration step of FeCl ?4H O is rate-controlled by a
2
2
nucleation and growth mechanism (A model), while subse-
quent steps obey a contracting phase boundary (R3) model.
The rate-limiting step of the A model is assumed to be the
formation and growth of nuclei, which are finite quantities of
product in the reactant lattice; after formation, a nucleus
grows and the nucleation rate is different from that of nuclei
growth. For a geometrical contraction mechanistic (R) model,
nucleation is assumed to be instantaneous throughout the
surface and the rate-limiting step is the progress of the
product layer from the surface of the crystal inward and is
4
4
different for various crystal morphologies. These complex
dehydration processes are most probably accompanied by
decomposition of either FeCl
and thus the absence of structural data about FeCl ?H O.
2 3
or FeCl in residual moisture,
4
5
2
2
After 15 h at 120 uC under dynamic vacuum, commercial
FeCl ?4H O yields the dihydrate form of ferrous chloride, as it
can be seen from the XRPD patterns (Fig. 2a–d).
Scheme 1 Description of the two different fluorination protocols used in our
2
2
2
study: one-shot versus stepwise direct gas–solid F fluorination.
3
666 | CrystEngComm, 2013, 15, 3664–3671
This journal is ß The Royal Society of Chemistry 2013

View Article Online
CrystEngComm
Paper
FeOCl + H
2
O = FeOOH + HCl
+ 0.5H
(5)
(6)
FeOOH = 0.5Fe
2
O
3
2
O
The M o¨ ssbauer spectrum, recorded after the thermal
treatment of FeCl ?6H O, consists of a broad doublet that
confirms the presence of iron oxychloride with an isomer shift
3
2
2
1
of d = 0.37 mm s and a quadrupole splitting of D = 0.90 mm
2
1
50–53
3+
s
(Fig. S18, ESI
3
).
In addition to the Fe doublet from
3
+
the oxychloride phase, a second Fe contribution appears to
be necessary to fit the experimental spectrum (d = 0.26 mm s
and D = 0.54 mm s ) and can be attributed to the
paramagnetic Fe species of lepidocrocite, c-FeOOH. Thus,
the thermal treatment of the ferric chloride hydrate leads to a
partial decomposition and the product is a mixture of poorly
2
1
2
1
3
+
54
Fig. 2 XRPD patterns of the different iron chloride precursors against their
crystallized FeOCl and amorphous FeOOH, according to eqn
simulations (* indicates peaks from FeCl
3
?6H
2
O).
3+
(4) and (5) and the XRPD data (Fig. 2g–h). These two Fe
signatures can also be evidenced in the M o¨ ssbauer spectrum
of FeCl ?4H O (Fig. S18, ESI ) and their contribution is about
14% of the signal while the main contribution is a Fe doublet
2
2
3
2
+
3 2
3
The ferric chloride hexahydrate, FeCl ?6H O (Fig. S11, ESI ),
undergoes simultaneous dehydration and dehydrochlorina-
2
1
21
with d = 1.22 mm s and D = 2.98 mm s , corresponding to
5
5
3+
tion from its molten phase in the temperature range 100–200
FeCl ?4H O. This underlines the presence of Fe impurities
2
2
4
6,47
uC,
melting of ferric chloride hydrate taking place around
in the ferrous chloride tetrahydrate and confirms the strong
tendency of hydrated ferrous chlorides to be oxidized in the
4
0 uC, and FeOCl and FeOOH are among the main decom-
4
8,46
45
position products that have been identified.
It is note-
presence of air.
worthy that the crystal structures of both compounds are
similar (Fig. S15, ESI ). According to the kinetics study, below
50 uC the dehydration is the main mechanism and follows
Hydrated ferrous and ferric chlorides dehydrate under inert
atmosphere following different kinetic models as regards the
temperature. Up to this point we may speculate that molecular
fluorine will react differently with the chloride precursors
depending on the heating process. While ferrous chloride
hydrates should not be so sensitive to different heating
temperatures, ferric chlorides follows two distinct kinetic
pathways when heated below or above 150 uC and thus should
give rise to different reactivity.
3
1
second order kinetics (F2); the rate law of a reaction-order-
based model is based on the reaction order, similar to the
same rate expressions in homogeneous kinetics (e.g., gas or
4
4
solution phases). Above 150 uC, the dehydration–dehydro-
chlorination follows a diffusion-based model, in which a
surface encrustation of a solid product tends to grow covering
the liquid surface of the molten ferric chloride; once such an
encrustation is grown to sufficient thickness, diffusion of
reactants into the reaction sites or products away from the
reaction sites through the moist surface becomes the slow
process and, therefore, rate-controlling, i.e. a hard shell of one
decomposition product limits the diffusion of by-products
One-shot fluorination of the chloride precursors
The XRD patterns of FeCl ?4H O, FeCl ?2H O and FeOCl–
2 2 2 2
FeOOH fluorinated at 350 uC for 15 h are shown in Fig. 3a. For
the ferrous chloride tetrahydrate, the product obtained is a
mixture of two crystallized phases: r-FeF
3
and HTB–FeF
and r-FeF
3
3
. From
(
H O and HCl) and thus prevents further decomposition,
the XRD pattern (Fig. 3a, ex 4H O), HTB–FeF
2
3
are
2
clearly distinguishable. When ferrous chloride dihydrate is
fluorinated in the same conditions, the resulting powder is
mainly composed of r-FeF (Fig. 3a, ex 2H O), as shown by the
3 2
intense peaks at 2h = 23.8u, 33.5u, 34.5u, 40.2u, 48.8u and 54.2u,
corresponding to (011), (211), ( ¯1 01), (102), (022) and (123) hkl
leading to a core–shell kind of situation. Heating hydrated
ferric chloride at 150 uC under dynamic vacuum for 15 h leads
to the formation of a poorly crystallized iron oxychloride,
FeOCl (Fig. 2g–h). The synthesis of FeOCl is usually done over
the course of several days by heating up a mixture of Fe
2
O
3
and
4
9
planes, respectively. The amount of crystallized HTB–FeF is
FeCl
3
at 350 uC according to eqn (2).
3
reduced in the present case, and its characteristic peaks can be
FeCl + Fe O = 3FeOCl
(2)
(3)
found at 2h = 13.8u, 23.6u and 27.8u (hkl planes (110), (002) and
3
2 3
(220), respectively). Finally, the fluorination of anhydrous
FeOCl–FeOOH led to the crystallization of a mixture of r-FeF
3
FeCl
3
?6H
2
O + 5FeCl
3
= 6FeOCl + 12HCl
and FeOF (or FeF ), without any trace of crystallized HTB–FeF3
2
It has also been reported that FeOCl could be obtained by
heating stoichiometric amounts of FeCl ?6H O and FeCl to
50 uC for 60–80 min, as written in eqn (3), and it has been
proposed that the decomposition of FeCl ?6H O takes place
according the following sequence (eqn (4)–(6)).
(Fig. 3a, ex FeOCl). Peaks from another minor phase can be
identified at 2h = 27.1u, 35.2u and 53.2u and indexed in the
parameters of the rutile-FeOF (ICSD #2875; hkl planes (110),
(101) and (211), respectively). Nevertheless an ambiguity still
3
2
6
3
1
2
3
2
4
6
exists regarding the nature of this phase as FeF is also a rutile-
2
type structure (ICSD #73729). The hydration of the precursor
has a direct influence on the crystallization of the iron fluoride
FeCl ?6H O = FeOCl + 5H O + 2HCl
(4)
3
2
2
This journal is ß The Royal Society of Chemistry 2013
CrystEngComm, 2013, 15, 3664–3671 | 3667

View Article Online
Paper
CrystEngComm
Fig. 3 XRPD patterns of iron chloride precursors FeCl
2
?4H
2
O, FeCl
2
?2H
2
O and FeOCl fluorinated at 350 uC for 15 h by the one-shot procedure (denoted ex 4H
and FeOF (a) and M o¨ ssbauer spectra of the same materials (b).
2
O, ex
2
H
2
O and ex FeOCl, respectively) against the simulation of r-FeF
3
, HTB–FeF
3
phases, as emphasized in the case of the soft template route,
and the overall attribution remains the same but the major
component is the well-crystallized r-FeF (67%), while the
contribution of HTB–FeF3 is only 22%. This result is in good
agreement with the previous XRD analysis.
5
6
and serves as a structural stabilizer of the open HTB–FeF
When the chloride precursor is thoroughly dehydrated, no
HTB–FeF can be detected (Fig. 3a, ex FeOCl). Moreover, there
3
.
3
3
is no evidence of the iron trifluoride trihydrate phase being
crystallized at any stage, a statement supported by the
thermogravimetric study of the products shown in Fig. S17,
Turning to the M o¨ ssbauer spectrum obtained for the
fluorinated FeOCl (Fig. 3b), it can be fitted using
a
paramagnetic doublet with an isomeric shift of d = 0.43 mm
2
1
21
ESI.
3
The results highlight the low water content in all the
s
(red line) and a quadrupole splitting of D = 0.51 mm s
,
products; at 240 uC, the total weight loss is less than 1% for the
three products.
and three sets of sextets with isomeric shifts of d = 0.49, 0.49
2
1
and 0.32 mm s and quadrupole splitting of D = 0.05 (dark
2
1
In addition to the XRD analysis, which is restricted to
crystallized solids, a spectroscopic M o¨ ssbauer study has been
carried out in order to investigate the amorphous, i.e. non-
crystallized, proportion of the fluorinated samples. Fig. 3b
shows the M o¨ ssbauer spectra of the fluorinated iron chlorides
powders. The obtained results are summarized in Table 2. A
quick glance at the spectra informs us about their similar
green line), 20.02 (light green line) and 20.09 mm s (orange
line), respectively. Among the sextets two of them are similar
to those found previously and correspond to the two types of
r-FeF , crystallized and poorly crystallized. The last minor
3
sextet indicates the presence of some hematite in this sample,
5
7
features that are characteristic of different Fe(III) species.
Crystalline r-FeF has a long-range magnetic order at room
temperature and thereby exhibits a six-line Zeeman-split
spectrum, while in its paramagnetic phase crystalline FeF
shows a single-line spectrum with a zero or negligibly small
Table 2 Room-temperature M o¨ ssbauer parameters for the three fluorinated
iron chloride precursors; reported values for known compounds are given for
comparison
3
3
2
1a
21b
d, mm s
D, mm s
H, Tesla
%
5
8
quadrupole splitting. The spectrum of ex FeCl
2 2
?4H O can be
c
FeOF
0.39, 0.44
0.483
0.45
0.45
0.46
0.85, 0.52
0
0.55
0.23
0.17
0.01
20.15
0.25
0, 0
38
0
0
0
40.4
33.8
0
46, 54
100
100
100
50
35
15
fitted by taking into account mainly a doublet (red line) with
59
r-FeF
am-FeF
HTB–FeF3
3
2
1
58
an isomeric shift of d = 0.46 mm s and a small quadrupole
3
6
0
2
1
splitting of D = 0.17 mm s , and two sets of sextets with
ex FeCl
2
?4H
2
2
O
O
2
1
isomeric shifts of d = 0.49 mm s and quadrupole splitting of
0.49
2
1
D = 0.01 (dark green line) and 20.15 mm s (light green line).
The doublet can be assigned to HTB–FeF whereas the sextet
having the largest hyperfine field (dark green line) is attributed
to r-FeF . The sextet with smaller hyperfine field is likely to
result from a badly crystallized r-FeF . As for ex FeCl ?4H O,
0.49
0.46
ex FeCl
2
?2H
22
3
0
0
0.43
0.49
0
0
.49
.49
0.02
20.16
0.51
0.05
20.02
20.09
40.6
35.2
0
40.3
35.1
51.5
67
11
42.6
29.4
25.1
2.8
3
ex FeOCl
3
2
2
.49
.32
the spectrum of ex FeCl
2
?2H
2
O also consists of a large doublet
2
1
21
(d = 0.46 mm s
and D = 0.25 mm s ), and two sets of
a
b
c
2
1
21
Isomeric shift. Quadrupole splitting. Non-published data.
sextuplets (d = 0.49 mm s and D = 0.02 and 20.16 mm s ),
3
668 | CrystEngComm, 2013, 15, 3664–3671
This journal is ß The Royal Society of Chemistry 2013

View Article Online
CrystEngComm
Paper
Table 3 Room-temperature M o¨ ssbauer parameters for FeOOH and
FeO(OH)12xF
x
21
21
d, mm s
D, mm s
FeOOH
FeO(OH)12xF
x
0.26
0.41
0.54
0.51
most probably coming from the decomposition of FeOOH (eqn
6)).
Up to this point, the phase identification stemming from
(
XRD has been confirmed by the M o¨ ssbauer data and the
isomer shifts are in the expected range for high-spin
5
7
octahedrally coordinated iron(III). The last contribution to
examine is the one coming from FeOF, seen in the XRD
pattern of fluorinated FeOCl (Fig. 3a). The Fe(III) signal of
FeOF cannot be distinguished in the M ¨o ssbauer spectrum.
Nevertheless the doublet of the fluorinated FeOCl (Fig. 3b, red
line) can be understood if we consider it to originate from
Fig. 4 XRPD patterns of fluorinated FeCl
against the simulation of FeOF, r-FeF
2
?4H
2
O following the stepwise protocol
3
3
and HTB–FeF ; the pattern of the product
from the one-shot procedure is shown for comparison (Al contribution is
coming from the sample holder).
FeOOH, present as a decomposition product of FeCl
3
?6H
2
O.
tend to decompose by following a diffusion-based model, thus
forming composites. This situation, as shown in the previous
section, leads to a partial fluorination at 350 uC of the inner
material, supposedly FeOOH, and to the presence of
FeO(OH)12xFx in the final product. Starting the fluorination
experiments at 150 uC should then prevent the diffusion-based
kinetics of decomposition of the dehydrated iron(III) chloride
precursors, and consequently avoid the formation of oxy(hy-
droxy)fluorides.
FeOOH may only be partially fluorinated at 350 uC under
molecular fluorine because it might compose the inner core of
the FeOCl–FeOOH composite. On the other hand FeOCl, which
composes the outer part of the composite, is fully fluorinated
and gives rise to r-FeF . Consequently the doublet can be
3
assigned to FeO(OH)12xFx, a partially fluorinated FeOOH.
Table 3 shows the effects of the fluorination on the M o¨ ssbauer
parameters of FeOOH.
FeO(OH)12x
F
x
possesses an increased isomeric shift (d =
In order to verify if our assumptions are correct, FeCl ?4H O
2
2
2
1
21
0
.41 mm s ) compared to FeOOH (d = 0.26 mm s ) and this
can be explained by taking into account the fact that the
isomer shift increases with the electronegativity of ligands,
and fluorine is more electronegative than OH.
substitution of F for OH modifies the iron environments
enough to produce iron sites that have different isomer shifts
d) and quadrupole splittings (D) (see Table 3). Similar results
have already been reported for other iron-based hydroxides.
This suggests that the FeOF signature seen in the XRD pattern
Fig. 3a) corresponds in reality to FeO(OH)12x
In the one-shot fluorination process, the hydration of the
and FeOCl precursors have been submitted to a stepwise
fluorination at 150 uC, 250 uC and finally 350 uC. Between each
fluorination step, the sample temperature is decreased back to
6
1–63
The
room temperature under N
analysis. The results are depicted in Fig. 4–6.
For FeCl ?4H O, the first fluorination step (5 h, 150 uC)
2
and a portion is taken away for
2
2
2
2
(
unambiguously produces HTB–FeF as the main crystallized
3
6
4
phase. As already emphasized for other systems, the presence
of hydration water molecules stabilizes the open bronze
(
x
F .
precursors has a direct influence on the crystallization of
different iron trifluoride phases. When the hydration of the
starting material is high (FeCl ?4H O), HTB-type FeF is the
2 2 3
main product, and while the hydration of the precursors
decreases, the main crystallized phase changes from HTB–
FeF
3
to r-FeF
3
(Fig. 3a), along with the presence of
FeO(OH)12xF
. The M o¨ ssbauer spectroscopy allows for a fine
x
analysis of the composition of the fluorinated samples and
supports the XRD phase identification, most importantly
regarding the presence (or absence) of oxy(hydroxy)fluorides
and amorphous phases.
Stepwise fluorination and effect of the hydration
The one-shot procedure described previously led to mixtures
of phases. In the case of dehydrated ferric chlorides, it even
Fig. 5 XRPD patterns of fluorinated FeOCl following the stepwise protocol
produced oxy(hydroxy)fluorides,
a result that could be
3
against the simulation of FeOF and r-FeF ; the pattern of the product from the
explained by taking into account the solid-state kinetics of
the precursors. Above 150 uC, dehydrated iron(III) chlorides
one-shot procedure is shown for comparison (Al contribution is coming from
the sample holder).
This journal is ß The Royal Society of Chemistry 2013
CrystEngComm, 2013, 15, 3664–3671 | 3669

View Article Online
Paper
CrystEngComm
The majority of transition metal fluoride (or oxide–fluoride)
frameworks, such as those investigated for cathode materials
in Li-ion technology, have been obtained through high
temperature reaction chemistry. New and/or original precur-
sors could improve the reaction conditions and, as it can be
seen in the present case, reduce reaction temperature. The
stepwise fluorination process of FeOCl gives rise to a well-
crystallized r-FeF after 15 h.
3
Conclusions
It was possible to synthesize two polymorphs of anhydrous
iron trifluoride at relatively low temperature through the
careful choice of the precursors and their hydration degree by
following a stepwise fluorination protocol. It was possible to
2 2
Fig. 6 Comparison of XRPD patterns of fluorinated FeCl ?4H O and FeOCl–
FeOOH precursors following the stepwise and one-shot procedures (D = r-FeF
= HTB–FeF , * = FeO(OH)12x ).
3
,
#
3
F
x
obtain r-FeF at 150 uC from FeOCl–FeOOH composites, while
3
the fluorination of FeCl
the hexagonal tungsten bronze type FeF
2
?4H
2
O at the same temperature led to
, HTB-FeF (Fig. 6).
3
3
5
6
structure of FeF3. The following fluorination steps at 250 and
50 uC lead to an increased crystallinity of the HTB phase with
no other phase detected in the XRD pattern (Fig. 4).{
Compared with the one-shot fluorination at 350 uC that leads
Increasing the temperature did not produce any other phases.
Combined analytical and physical techniques have demon-
strated the impossibility of isolating single phased iron
trifluorides from one-shot fluorination. These results demon-
strate the necessity to investigate the thermal behaviour of
precursors prior to submitting them to fluorination.
3
to a 1 : 1 mixture of r-FeF
stepwise protocol provides a route towards almost single-
phased HTB–FeF
3 3
and HTB–FeF (Table 2), this
3
.
When the stepwise procedure is applied to the dehydrated
chloride precursor FeOCl–FeOOH, the first step leads to the
crystallization of anhydrous r-FeF
3
, which continues to grow as
Notes and references
a single crystallized phase after the subsequent 250 and 350 uC
steps (Fig. 5), with no detectable traces of crystallized
oxy(hydroxy)fluorides.
1 G. Villalba, R. U. Ayres and H. Schroder, J. Ind. Ecol., 2007,
11, 85–101.
2
3
4
5
6
7
8
9
Inorganic Solid Fluorides, ed. P. Hagenmuller, Academic
Press Inc., 1985.
P. P. Fedorov, A. A. Luginina, S. V. Kuznetsov and V.
V. Osiko, J. Fluorine Chem., 2011, 132, 1012–1039.
A. Sekiya, H. D. Quan, M. Tamura, R. X. Gao and J. Murata,
J. Fluorine Chem., 2001, 112, 145–148.
E. Kemnitz, S. Wuttke and S. M. Coman, Eur. J. Inorg.
Chem., 2011, 2011, 4773–4794.
E. Kemnitz and D.-H. Menz, Prog. Solid State Chem., 1998,
As expected from the study of the thermal stability and
kinetics of the precursors, the stepwise procedure has a major
influence on the reaction products. Hydrated ferrous chloride
precursor transformation below 150 uC is rate-controlled by a
nucleation and growth mechanism, and the same compound
tends to transform above 150 uC by following a contracting
phase boundary model, which rate-limiting step is the
progress of the product layer from the surface of the crystal
inward. By employing the stepwise fluorination procedure, the
relatively high water content of the precursor helps stabilizing
the HTB phase and this transformation is not rate-limited.
For the FeOCl–FeOOH precursor, the influence of the
stepwise protocol is emphasized by the absence of crystal-
lization of oxy(hydroxy)fluorides in the final product, as shown
by the XRD analysis (Fig. 5). This behaviour can be rationalized
by considering the kinetic data. Indeed, if heated above 150 uC,
ferric chloride decomposes by following a diffusion-based
model that induces the formation of an encrustation. Such a
hard layer slows down the diffusion of reactants into reaction
sites. By heating the FeOCl–FeOOH precursor at 150 uC, this
diffusion-based model is by-passed.
26, 97–153.
G. Scholz, R. St o¨ ßer, J. A. Momand, A. Zehl and J. Klein,
Angew. Chem., Int. Ed., 2000, 39, 2516–2519.
W. M. Patterson, P. C. Stark, T. M. Yoshida, M. Sheik-Bahae
and M. P. Hehlen, J. Am. Ceram. Soc., 2011, 94, 2896–2901.
S. Fujihara and K. Tokumo, J. Fluorine Chem., 2009, 130,
1
106–1110.
0 J. Lucas, F. Smektala and J. L. Adam, J. Fluorine Chem.,
002, 114, 113–118.
1 D. K. Chatterjee, L. S. Fong and Y. Zhang, Adv. Drug Delivery
Rev., 2008, 60, 1627–1637.
1
1
2
12 G. N. Nehme, Wear, 2012, 278–279, 9–17.
13 G. G. Amatucci and N. Pereira, J. Fluorine Chem., 2007, 128,
243–262.
1
1
4 J. Cabana, L. Monconduit, D. Larcher and M. R. Palac ´ı n,
Adv. Mater., 2010, 22, E170–E192.
5 H. Arai, S. Okada, Y. Sakurai and J.-I. Yamaki, J. Power
Sources, 1997, 68, 716–719.
{
The presence of r-FeF
fluorination step at 150 uC may stem from Fe(III) impurities in FeCl
please refer to ESI
3
as a minor phase in the product obtained from the first
2
2
?4H O
(
3
)
3
670 | CrystEngComm, 2013, 15, 3664–3671
This journal is ß The Royal Society of Chemistry 2013

View Article Online
CrystEngComm
Paper
1
1
1
1
2
6 Handbook of Preparative Inorganic Chemistry, ed. G. Brauer,
Academic Press Inc., 2nd edn, 1963, vol. 1.
7 L. Francke, E. Durand, A. Demourgues, A. Vimont, M. Daturi
and A. Tressaud, J. Mater. Chem., 2003, 13, 2330.
8 R. F. Li, S. Q. Wu, Y. Yang and Z. Z. Zhu, J. Phys. Chem. C,
41 J. Grannec and L. Lozano, in Inorganic Solid Fluorides, ed.
P. Hagenmuller, Academic Press Inc., 1985, pp. 17–76.
42 B. Morosin and E. J. Graeber, J. Chem. Phys., 1965, 42,
898–901.
43 S. C. Shu and C. M. Shiao, Acta Chim. Sin., 1960, 26, 124.
44 A. Khawam and D. R. Flanagan, J. Phys. Chem. B, 2006, 110,
17315–17328.
45 S. B. Kanungo, J. Indian Chem. Soc., 2004, 81, 750–756.
46 S. Kanungo and S. Mishra, J. Therm. Anal., 1996, 46,
1487–1500.
47 S. Kanungo and S. Mishra, J. Therm. Anal., 1997, 48,
385–401.
2010, 114, 16813–16817.
9 A. Demourgues, N. Penin, D. Dambournet, R. Clarenc,
A. Tressaud and E. Durand, J. Fluorine Chem., 2012, 134, 35–43.
0 K. Lutar, H. Borrmann and M. Leblanc, in Advanced Inorganic
Fluorides: Synthesis, Characterization and Applications, ed. T.
Nakajima, A. Tressaud and B. Zemva, Elsevier, Amsterdam, 1st
edn, 2000, pp. 5–49.
2
2
2
2
2
1 W. Armarego and C. Chai, Purification of Laboratory
Chemicals, 2009.
2 J.-M. Gren `e che, F. Varret, M. Leblanc and G. F ´e rey, Solid
State Commun., 1987, 61, 813–816.
3 M. Leblanc, J. Pannetier, G. F ´e rey and R. De Pape, Rev.
Chim. Miner., 1985, 22, 107–114.
4 G. F ´e rey, M. Leblanc, R. De Pape, M. Passaret and M.
P. Bothorel-Razazi, J. Cryst. Growth, 1975, 29, 209–211.
5 G. F ´e rey, M. Leblanc, R. De Pape and J. Pannetier, in Inorganic
Solid Fluorides, ed. P. Hagenmuller, Academic Press Inc., 1985,
48 J. Frydrych, L. Machala, M. Hermanek, I. Medrik,
M. Mashlan, J. Tucek, J. Pechousek and V. K. Sharma,
Thin Solid Films, 2010, 518, 5916–5919.
49 M. D. Lind, Acta Crystallogr., Sect. B: Struct. Crystallogr.
Cryst. Chem., 1970, 26, 1058–1062.
50 G. A. Fatseas, P. Palvadeau and J. P. Venien, J. Solid State
Chem., 1984, 51, 17–37.
51 R. H. Herber and Y. Maeda, Inorg. Chem., 1981, 20,
1409–1415.
52 R. H. Herber and Y. Maeda, Physica B+C, 1980, 99, 352–356.
53 E. Kostiner and J. Steger, J. Solid State Chem., 1971, 3,
273–275.
54 C. E. Johnson, J. Phys. C: Solid State Phys., 1969, 2,
1996–2002.
394–413.
2
6 M. Leblanc, G. F ´e rey, P. Chevallier, Y. Calage and R. De
Pape, J. Solid State Chem., 1983, 47, 53–58.
2
7 R. De Pape and G. F ´e rey, Mater. Res. Bull., 1986, 21,
9
71–978.
55 J. N. Bull, R. G. A. R. MacLagan, C. M. Fitchett and W.
C. Tennant, J. Phys. Chem. Solids, 2010, 71, 1746–1753.
56 C. Li, L. Gu, J. Tong, S. Tsukimoto and J. Maier, Adv. Funct.
Mater., 2011, 21, 1391–1397.
57 F. Menil, J. Phys. Chem. Solids, 1985, 46, 763–789.
58 M. Eibsch u¨ tz, M. E. Lines, L. G. Van Uitert, H.
J. Guggenheim and G. J. Zydzik, Phys. Rev. B: Condens.
Matter, 1981, 24, 2343.
2
2
3
3
3
8 A. Tressaud, J. Fluorine Chem., 2011, 132, 647–655.
9 A. H. Nielsen, Z. Anorg. Allg. Chem., 1940, 244, 85–88.
0 G. Teufer, Acta Crystallogr., 1964, 17, 1480–1480.
1 D. G. Karraker and P. K. Smith, Inorg. Chem., 1992, 31, 1118–1120.
2 Y. Macheteau and P. Charpin, C. R. Hebd. Seances Acad.
Sci., 1972, 275, 443.
3
3
3
3
3
3
3 C. Li, L. Gu, S. Tsukimoto, P. A. van Aken and J. Maier, Adv.
Mater., 2010, 22, 3650–3654.
4 K. J. Gallagher and M. R. Ottaway, J. Chem. Soc., Dalton
Trans., 1975, 978–982.
5 Y. Laligant, J. Pannetier, P. Labbe and G. F ´e rey, J. Solid
State Chem., 1986, 62, 274–277.
6 W. Hall, S. Kim, J. Zubieta, E. G. Walton and D. B. Brown,
Inorg. Chem., 1977, 16, 1884–1887.
7 D. B. Brown, E. G. Walton and J. A. Dilts, J. Chem. Soc.,
Dalton Trans., 1980, 845–850.
8 P. Charpin and Y. Macheteau, C. R. Hebd. Seances Acad.
Sci., 1975, 280, 61.
59 I. E. Malka, A. Błachowski, K. Ruebenbauer, J. Przewo ´z nik,
˙
J. Zukrowski, T. Czujko and J. Bystrzycki, J. Alloys Compd.,
2011, 509, 5368–5372.
60 Y. Calage, M. Leblanc, G. F ´e rey and F. Varret, J. Magn.
Magn. Mater., 1984, 43, 195–203.
61 Y. Hazony, R. C. Axtmann and J. W. Hurley Jr, Chem. Phys.
Lett., 1968, 2, 440–444.
62 R. C. Axtmann, Y. Hazony and J. W. Hurley Jr, Chem. Phys.
Lett., 1968, 2, 673–676.
63 Y. Hazony and R. C. Axtmann, Chem. Phys. Lett., 1971, 8,
571–573.
3
4
9 E. Herdtweck, Z. Anorg. Allg. Chem., 1983, 501, 131–136.
0 M. Leblanc, G. F ´e rey, Y. Calage and R. De Pape, J. Solid
State Chem., 1984, 53, 360–368.
64 M. Ati, M. T. Sougrati, G. Rousse, N. Recham, M.-L. Doublet,
J. C. Jumas and J.-M. Tarascon, Chem. Mater., 2012, 24,
1472–1485.
This journal is ß The Royal Society of Chemistry 2013
CrystEngComm, 2013, 15, 3664–3671 | 3671