Nano-CoF3 prepared by direct fluorination with F2 gas: Application as electrode material in Li-ion battery

Article abstract of DOI:10.1016/j.jfluchem.2016.10.003

Nano-CoF₃ powders were synthesized by direct fluorination at various temperatures (up to 300?°C) of cobalt nanoparticles and their electrochemical features in lithium battery were investigated. The structure and composition of the raw and fluorinated materials are explored by SEM/TEM, HRTEM, XRD and XPS experiments. The influence of the fluorination temperature on the crystallite size has been evidenced by XRD. It is shown that such fluorination process allows the direct synthesis of nano-CoF₃, at T_F2?=?100?°C. The XPS investigation of mixtures of CoF₃ and acetylene black has shown that, when handled in very dry atmospheres, CoF₃-based samples can be used to prepare electrodes for Li-ion batteries in safe conditions. The electrochemical reaction of nano-CoF₃ powders vs. lithium metal studied in organic medium involves a conversion process as described in the case of cobalt oxide. Successive impedance measurements were carried out at different discharge states to follow the modifications occurring at the electrode/electrolyte interface during the discharge step. The evolution of the reversible capacity retention of Li⁺ vs. time capacity of the powdered materials as well as the cycle life of the battery were investigated as a function of the fluorination temperature used for the preparation of the cobalt fluoride nanopowders. The best electrochemical performances were obtained with nano-CoF₃ powders prepared at fluorination temperature of 100?°C, for which a reversible capacity of about 390?mAh/g is obtained after subsequent cycles.

Full text of DOI:10.1016/j.jfluchem.2016.10.003

G Model
FLUOR 8868 No. of Pages 11
Journal of Fluorine Chemistry xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Journal of Fluorine Chemistry
journal homepage: www.elsevier.com/locate/fluor
Nano-CoF3 prepared by direct fluorination with F2 gas: Application as
electrode material in Li-ion battery
a
a
a
a
b
H. Groulta, , S. Neveu , S. Leclerc , A.-G. Porras-Gutierrez , C.M. Julien , A. Tressaud ,
*
E. Durandb, N. Peninb, C. Labrugerec
Sorbonne Universités, Univ. UPMC-Paris 6, Laboratoire PHENIX-UMR CNRS 8234, 4 place Jussieu, 75252 Paris, Cedex 05, France
ICMCB-CNRS, Université Bordeaux, 87, Av. Dr. A. Schweitzer, 33608 Pessac Cedex, France
Plateforme Aquitaine de Caractérisation des Matériaux, UMS 3626, CNRS-Université Bordeaux, 87 Av. Dr. A. Schweitzer, 33608 Pessac Cedex, France
a
b
c
A R T I C L E I N F O
A B S T R A C T
Article history:
Received 15 June 2016
Received in revised form 28 September 2016
Accepted 1 October 2016
Available online xxx
Nano-CoF3 powders were synthesized by direct fluorination at various temperatures (up to 300 ꢀC) of
cobalt nanoparticles and their electrochemical features in lithium battery were investigated. The
structure and composition of the raw and fluorinated materials are explored by SEM/TEM, HRTEM, XRD
and XPS experiments. The influence of the fluorination temperature on the crystallite size has been
evidenced by XRD. It is shown that such fluorination process allows the direct synthesis of nano-CoF3, at
TF2 = 100 ꢀC. The XPS investigation of mixtures of CoF3 and acetylene black has shown that, when handled
in very dry atmospheres, CoF3-based samples can be used to prepare electrodes for Li-ion batteries in safe
conditions. The electrochemical reaction of nano-CoF3 powders vs. lithium metal studied in organic
medium involves a conversion process as described in the case of cobalt oxide. Successive impedance
measurements were carried out at different discharge states to follow the modifications occurring at the
electrode/electrolyte interface during the discharge step. The evolution of the reversible capacity
retention of Li+ vs. time capacity of the powdered materials as well as the cycle life of the battery were
investigated as a function of the fluorination temperature used for the preparation of the cobalt fluoride
nanopowders. The best electrochemical performances were obtained with nano-CoF3 powders prepared
at fluorination temperature of 100 ꢀC, for which a reversible capacity of about 390 mAh/g is obtained after
subsequent cycles.
Dedicated to Prof. Steve H. Strauss, winner
of the 2016-ACS Award for Creative Work
in Fluorine Chemistry.
Keywords:
Cobalt trifluoride
Fluorine
XPS
Reactivity
Electrode
Lithium battery
Conversion reaction
Nanoparticles
ã 2016 Published by Elsevier B.V.
1. Introduction
For a long time, minor attention has been devoted to metal
fluorides due to their poor electronic conductivity. However, iron
The demand of clean and performing energy storage devices is
growing up and requires much attention for the development of
electrochemical systems such as Li-ion batteries. Among available
electrode materials to be used in such devices, main interest has
been focused on transition metal-based compounds. Indeed,
owing to the presence of empty channels or layers in the lattice,
these materials are able to insert/extract reversibly Li+ ions
without any detrimental structural changes. Therefore, the
improvement of the electrochemical battery performances could
imply the development of new classes of electrode materials able
to incorporate three electrons per 3d-metal, thus delivering higher
energy density, and exhibiting longer cycle life in addition to
moderate cost.
fluoro-compounds are of interest for batteries because of their low
cost and low toxicity. In addition, conversion reactions are involved
during the reduction of metal fluorides, enabling full utilization of
all the redox potentials of the host metal [1–8]: indeed, the metal
fluoride can be reduced fully to the metallic state according to the
reversible following reaction: xLi+ + xeꢁ + MFx $ xLiF + M. This
conversion reaction is characterized by full reduction of the metal
leading to the stabilization of nano-sized M/LiF domains, i.e. the
stabilization of metallic nano-metal embedded in a lithiated
matrix. This reaction is generally reversible due to the nanoscale
nature of the resulting composite. Moreover, MF3 such as FeF3
support also intercalation, forming LixMF3 compounds, as also
demonstrated for TiF3, FeF3 and VF3 [2,5,9]. It results a two-step
mechanism involving at first an one-electron intercalation reaction
(Reaction (1)) followed by a two-electrons conversion reaction
(Reaction (2)) according to:
* Corresponding author.
E-mail address: henri.groult@upmc.fr (H. Groult).
http://dx.doi.org/10.1016/j.jfluchem.2016.10.003
0022-1139/ã 2016 Published by Elsevier B.V.
Please cite this article in press as: H. Groult, et al., Nano-CoF3 prepared by direct fluorinationwith F2 gas: Application as electrode material in Li-
ion battery, J. Fluorine Chem. (2016), http://dx.doi.org/10.1016/j.jfluchem.2016.10.003

G Model
FLUOR 8868 No. of Pages 11
2
H. Groult et al. / Journal of Fluorine Chemistry xxx (2016) xxx–xxx
M3+F3 + Li+ + eꢁ ! LiM2+F3
and
(1)
(surfactant in cyclohexane) were mixed by mechanical stirring
for four days. The surfactant used was trioctylphosphineoxide. A
black magnetic precipitate was obtained in the organic solution,
which was washed several times with acetone and alcohol and
then dried. Finally, the powders were heat-treated at 250 ꢀC under
vacuum to remove traces of mercury.
The fluorination of pristine Co nanoparticles was performed at
temperatures ranging from 100 to 300 ꢀC under 1 atm of elemental
fluorine (purity: 99.4–99.7%) in a nickel reactor. A decomposition
of nano-CoF3 occurs at higher temperatures giving rise to the
presence of CoF2. The composition of the fluorinated samples was
determined by coupling elemental analysis and X-ray diffraction
(XRD) measurements. The latter were performed with a Rigaku
LiM2+F3 + 2Li+ + 2eꢁ ! 3LiF + M0
(2)
The nature of the LiF/metal nano-composite may induce a high
reversibility of the conversion reaction, giving rise to large
reversible capacities. Indeed, the theoretical specific capacity of
positive electrode is significantly enhanced from ꢂ280 mAh/g for
layered intercalation compounds to more than 700 mAh/g for
metal di- or tri-fluorides. Thus, MF2 and MF3 compounds (with
M = Bi, Ti, V, Fe, Mn, Ni . . . ) have been already studied as positive
electrode materials [2–5,9,10]. For instance, Amatucci et al. [1]
reported that metal fluorides such as iron and bismuth fluorides
can be considered as suitable candidates for electrode materials for
rechargeable lithium batteries [1,8,10]. Recently, oxyfluoride and
hydroxyfluoride were considered as new candidates in such
rechargeable batteries [11–13].
Here, we focus our attention on the properties of nano-sized
cobalt trifluoride. Few articles have been devoted to the use of CoF3
as negative electrode [14]. The lack of reports may be due to the
instability of this metal fluoride in the presence of air that limits its
further development. CoF3 decomposes upon contact with water
to give HF, CoF2 and O2. In addition, as other metal fluorides, an
enhancement of its electrical conductivity, se, is requested for
applications. Among the possible routes to increase the electric
contact between particles, ball milling technique using various
carbonaceous products (graphite, black carbon, etc.) is one of the
most popular. This would lead to the formation of CoF3/C
composites [15] such as for FeF3. However, CoF3 is a strong
fluorinating agent and high energy arising from ball milling
performed in presence of CoF3 and carbon could lead to the
decomposition of CoF3 into CoF2 + F2, with formation of insulating
CFx phases. Therefore, severe precautions should be taken to
perform such surface modification of CoF3 powder by this
technique.
In spite of these limitations and because we perfectly know how
to handle all kinds of fluorinated species, we focused our attention
to the study the electrochemical behaviour in Li-ion battery of
nano- CoF3 prepared from direct fluorination of cobalt nano-
particles at different temperatures using F2 gas. It is well
established that the use of nano-sized particles greatly improves
the electrochemical activity, such as for other metal fluorides.
Optimizing the synthesis conditions in order to achieve nano-sized
crystallites of CoF3 will thus constitute one important issue of this
study. In the second part, the reactivity of CoF3 in different
atmospheres will be checked in the presence of acetylene black, a
material used during the fabrication process of lithium battery
electrodes.
Ultima III X-Ray diffractometer with
a
CuKa radiation
(l1 = 1.540598 Å). Global Rietveld profile refinement was made
to perform a fine analysis of the structure and to determine the
average particle size of CoF3 from the broadening of the Bragg
peaks (Scherrer formula) with respect to the instrumental one. The
experimental contribution of the diffractometer to the line
broadening was established using XRD patterns of standard LaB6
powder. The peak profiles were fitted using the pseudo-Voigt
function of Thompson-Cox-Hastings which allows refining sepa-
rately the angular dependence of both Lorentzian (HL) and
Gaussian (HG) components of FWHM (full width at half maxi-
mum). High resolution Transmission electron micrographs were
obtained with
microscope equipped with
a
Jeol JEM 100 CX II transmission electron
Jeol high resolution scanning
a
attachment (STEM-SEM ASID 4D). Scanning electron microscopy
images were obtained with a Jeol JEM 100 CX II scanning electron
microscope equipped with
a Jeol high resolution scanning
attachment (SEM-FEG). Finally, HRTEM images were collected
with JEOL JEM 2011. Surface characterizations were done by XPS
measurements using a ThermoFisher Scientific K-ALPHA spec-
trometer under ultrahigh vacuum conditions. The radiation was an
monochromatized Al source (1486.6 eV). All spectra were refer-
enced to C1 s located at binding energy (BE) of 284.8 eV. Surveys
were recorded at 200 eV pass energy and high resolution spectra
C1s, O1s, F1 s and Co2p at 40 eV pass energy, then fitted with an
AVANTAGE processing program provided by ThermoFisher. The
fluorinated samples were stored in a dry box until they were very
rapidly transferred into the XPS chamber.
The stability of the active CoF3 obtained after fluorination was
tested in order to ensure the process of the preparation of an
electrode, in same conditions as used in Li-ion batteries. The
purpose was to check the stability of CoF3 vs. a humid atmosphere
and also vs. divided carbon which was mixed to the active
component in order to enhance the electrode electrical conduc-
tivity. The fluoride powders were mixed thoroughly with various
amounts of acetylene black (AB) in a dry box containing only a few
ppm O2, H2O. The composite material was as follows: 80% CoF3,
20% AB. No binder such as PVDF was added, in order to have a clear
XPS answer on both C1 s and F1 s spectra, and also because this
material is reactive vs. CoF3. Afterwards, the mixture was
investigated by XPS, either in dry conditions, or after being left
for two days in the laboratory’s atmosphere.
2. Experimental section
Raw cobalt nanoparticles were obtained from an original
electrochemical route described elsewhere [16–18]. Briefly, the
electrochemical process consisted in the anodic dissolution of a Co
metal bar giving rise to metal ions (Co2+) in aqueous solution at
pH > 6 in order to avoid the reduction of protons; the homogeneous
medium was composed of sodium citrate or ammonium citrate,
Co2+ ions being dissolved as complexes. Due to their close time
constants, Co2+ ions were reduced on a liquid mercury cathode
producing metal nanoparticles directly dispersed in mercury. The
obtained material is a magnetic and conducting fluid that can be
considered as a source of metal nanoparticles since the metal can
be recovered by using a liquid–liquid extraction. The magnetic
nanoparticles were extracted using a mixture of surfactant in an
organic solvent. To succeed, magnetic mercury and solvent
The electrochemical lithium insertion/deinsertion reaction was
studied in 1 mol/L LiPF6-EC:DMC (1:1) solution (LP30, Merck) at
room temperature in a glove box (water content < 10 ppm) under
argon atmosphere. The working electrode was composed of 90% of
the active material, 5w% of acetylene black and 5w% of PVFD as
binder. The counter and reference electrodes were metallic lithium
foils. All potential values were referred to this Li0/Li+ reference
henceforth. The galvanostatic charge-discharge curves were
performed using
a potentiostat/galvanostat impedancemeter
(VMP3 Bio-Logic) in the potential range 0.02–4.0 V at various
current densities.
Please cite this article in press as: H. Groult, et al., Nano-CoF3 prepared by direct fluorination with F2 gas: Application as electrode material in Li-
ion battery, J. Fluorine Chem. (2016), http://dx.doi.org/10.1016/j.jfluchem.2016.10.003

G Model
FLUOR 8868 No. of Pages 11
H. Groult et al. / Journal of Fluorine Chemistry xxx (2016) xxx–xxx
3
3. Results
3.1. Analysis of the starting raw material: Co-based nanoparticles
The morphology of Co nanoparticles prepared by electrochem-
ical route was characterized by SEM and TEM. The SEM image
(Fig. 1a) reveals the formation of quasi-spherical aggregated
particles. TEM image presented in Fig. 1b shows the polydispersity
of the metal with a mean diameter of the order of 10 nm for cobalt
nanoparticles. XRD pattern of non-fluorinated cobalt nanoparticles
(Fig. 2) evidences the presence of strong diffraction lines of Co3O4
(ICDD 00-009-0418) as the main component, while CoO(OH)
(JCPDS 01-073-1213); Co(OH)2 (ICDD 00-030-0443), and Co (ICDD
01-089-7373) can also be detected as constituents. It is assumed
that the core of the metallic Co particles is thus covered by a
surface layer constituted by spinel Co3O4 and oxy-hydroxide CoO
(OH), whereas hydroxide Co(OH)2 may also be present. Thus, it is
suggested that after liquid–liquid extraction the cobalt nano-
particles were oxidized, due to their high reactivity vs. oxidation in
air.
Fig. 2. XRD pattern of raw Co-based nanoparticles.
Table 1
Quantification of XPS spectra (atomic%, Scofield library).
To confirm this assumption, surface XPS measurements were
performed. Table 1 summarizes the elemental analysis of all
investigated samples. Note that 2.5% Hg was found at the surface in
the starting nano-Co particles due to the presence of the mercury
electrode in the electrochemical cell (see “Experimentals” above).
High-resolution Co 2p spectra show a spin-orbit splitting into 2p3/2
and 2p1/2 components at ꢂ789.0 eV range and 795.0–810.0 eV
range, respectively. As expected, the intensity ratio of the Co 2p3/2
and Co 2p1/2 peaks is close to 2. These components contain
qualitatively the same chemical information [19–21]. Therefore,
only the envelope relative to Co 2p3/2 peak was fitted (including
satellites) and discussed hereafter. Individual fitted parameters, i.e.
C
O
Co
F
Nano-Co
Nano-Co F100
CoF3-F300
CoF3-F300-AB-dry
CoF3-F300-AB-air
28.5
36.3
42.2
61.3
58.2
53.0
24.5
9.8
6.9
10.7
16.0
13.0
11.5
7.6
–
26.2
36.5
24.2
22.8
8.3
binding energy (BE), as well as the relative contribution of each
component (Ii%) to the total area of the Co 2p3/2 peak are calculated
for the different groups in Tables 2a and 2b.
Concerning the starting nano-Co sample, the XPS Co 2p3/2
spectrum (Fig. 3a) can be fitted into several asymmetrical
Table 2a
XPS data of pristine Co nanoparticles ꢁ Binding energies BE (eV), and relative
amounts of the different components Ii (%). in Co2p3/2 and O1 s spectra.
BE (eV)
Ii (%)
Assignment
Co 2p3/2
780.0
781.3
788.3
790.2
791.4
529.9
531.2
532.4
533.9
74.5
25.5
Co3O4
CoO(OH)
satellites
O 1s
38.3
42.8
15.6
3.3
Co3O4
CoO(OH) + O = C
O-C
H2O
Table 2b
XPS data of cobalt nanoparticles fluorinated at 100 and 300 ꢀC under F2 atmosphere.
[BE (eV), and relative amounts of the different components Ii (%) of Co 2p3/2 F1 s and
O1 s spectra].
Nano-Co F100
CoF3-F300
BE (eV) Ii (%) Assignment
BE (eV) Ii (%) Assignment
Co 2p3/2 780.8
15.9 Co3O4, CoO(OH)
55.1 Co-OxFy+ Co-OyFx
29.0 Co-Fz
–
–
–
781.8
784.6
788.3
790.9
684.5
686.6
530.0
531.5
533.1
782.6
784.6
788.8
793.2
684.9
687.2
531.2
532.3
533.6
51.1 Co-OyFx
48.9 Co-Fz
satellites
Satellites
F 1s
92.8 Fz-Co
93.6 Fz ꢁCo
7.2
F-C
6.4
9.8
F-C
CoO(OH),
O 1s
11.7 Co3O4
57.9
30.4 CoO(OH), O=C
Co-OxFy, O-C, Co-
OyFx
36.9 O=C
53.3 Co-OxFy, O-C
Co-OyFx
Fig. 1. (a) SEM and (b) TEM images of raw cobalt nanoparticles.
Please cite this article in press as: H. Groult, et al., Nano-CoF3 prepared by direct fluorinationwith F2 gas: Application as electrode material in Li-
ion battery, J. Fluorine Chem. (2016), http://dx.doi.org/10.1016/j.jfluchem.2016.10.003

G Model
FLUOR 8868 No. of Pages 11
4
H. Groult et al. / Journal of Fluorine Chemistry xxx (2016) xxx–xxx
Fig. 3. Fitted Co2p3/2, O1s, F1 s and C1 s XPS spectra of (a) nano-Co and (b) nano-Co F100 samples.
Fig. 4. Experimental X-ray diffraction powder patterns (red curve); Rietveld-refined profile (blue curve), and difference curve (green curve) of Co-nanoparticles fluorinated at
(a) 100 and (b) 300 ꢀC under pure F2 (p = 1 bar), during 3 days. Vertical bars are the positions of the Bragg reflections. (For interpretation of the references to colour in this figure
legend, the reader is referred to the web version of this article.)
components located at 780.0, 781.3, 788.3, 790.2 and 791.4 eV. The
first two components are assigned to the elastic peaks, whereas the
three last are satellites peaks. The main contribution at 780.0 eV
corresponds to Co3O4, which is consistent with the presence of a
satellite on the higher BE side of the main 2p component (higher
than 10.2 eV) [20–24]. The second component at 781.3 eV could be
due to either Co hydroxide or oxide-hydroxide, i.e. Co(OH)2 or CoO
(OH). As no satellite is found at 786.3 eV corresponding to the
fingerprint signature of the hydroxide, the component at 781.3 eV
can be therefore assigned to Co oxide-hydroxide, as reported in
[20]. It can be noted that the area of the two main components
represents 68.6 and 23.5% of the total area of the Co2p3/2 envelope
(Table 2a), respectively. The O1 s XPS spectrum is decomposed into
four peaks (Fig. 3a). The two main peaks observed at 529.9 and
531.2 eV are due to the oxygen from Co3O4, oxy-hydroxyl groups
shown previously on Co-based nano-particles exposed to oxidative
atmospheres [22].
To summarize, XPS measurements performed on pristine Co-
based nanoparticles evidence that these particles are fully covered
by layers of cobalt oxide and oxide-hydroxide. Let us consider now
the effect of the fluorination on these nano-particles.
3.2. Fluorination procedure of Co-based nanoparticles
The fluorination of cobalt nanoparticles was performed at two
temperatures (noted TF henceforth): 100–300 ꢀC. The correspond-
ing XRD patterns for TF = 100 and 300 ꢀC presented in Fig. 4a and b
are in total agreement with those of cobalt trifluoride CoF3 (JCPDS-
ICDD file 04-005-4245) [28]. It should be noted that the reaction
starts at even moderate temperatures (i.e. TF = 100 ꢀC), because of
the high reactivity of F2 gas vs. cobalt oxide and (oxy)hydroxide.
However, the crystallinity of the powder is improved when
increasing the temperature toTF = 300 ꢀC (Fig. 4b). CoF3 possesses a
trigonal structure with R-3c space group symmetry. The fluorina-
tion temperature has a great influence on the particles size as
evidenced by a narrowing of the XRD peaks with increasing
temperatures. Whatever, good refinements were obtained as a
and O
at 532.4 and 533.9 eV could be attributed to the pollution of OꢁꢁC
and water, respectively [25]. The presence of O
¼C pollution, respectively [20,23,26,27], whereas the peaks
¼C and OꢁꢁC groups
is also revealed by the 286–289 eV shoulder in the C1 s spectrum
(Fig. 3a). Such results account for the coverage of the raw cobalt
nanoparticles with a layer of Co oxide and oxide-hydroxide, as
Please cite this article in press as: H. Groult, et al., Nano-CoF3 prepared by direct fluorination with F2 gas: Application as electrode material in Li-
ion battery, J. Fluorine Chem. (2016), http://dx.doi.org/10.1016/j.jfluchem.2016.10.003

G Model
FLUOR 8868 No. of Pages 11
H. Groult et al. / Journal of Fluorine Chemistry xxx (2016) xxx–xxx
5
Table 3
morphology of the particles is kept even after fluorination at
TF = 100 ꢀC under F2 gas. Note that the material is partially
decomposed under the electron beam during the collection of
the TEM images.
Crystallographic data of CoF3 nanoparticles prepared by treating nano-Co particles
at TF = 100 and 300 ꢀC.
TF (ꢀC)
a (Å)
c (Å)
Average particle size (nm)
XPS investigations were carried out on Co-based nanoparticles
fluorinated at 100 ꢀC due to the best conditions yielding fluorinated
powders with smallest particle size. The data were compared with
those of raw pristine material (Fig. 3). Although CoF3 is the main
phase observed in the XRD pattern, the presence of ꢂ25 at.% O is
detected at the surface by XPS spectroscopy (Table 1). The Co 2p3/2
spectrum is presented in Fig. 3b, for which individual binding
energy as well as the relative contribution of each component (Ii/%)
are given in Table 2b. This spectrum can be decomposed by five
asymmetrical peaks situated at 780.8, 781.8, 784.6, 788.3, and
790.9 eV. The first three components are attributed to the elastic
peaks, whereas the two last are due to satellite peaks. The peak
located at 780.8 eV can be assigned to remaining oxide/oxide-
hydroxide present at the surface, that is also confirmed by the
occurrence of the low BE component at 530 eV in the O1 s
spectrum (compare O1 s spectra in Figs. 3a and 3b). Both
contributions around 782 and 784.5 eV are the fingerprints of
cobalt in a fluorinated environment. If we refer to the XRD pattern,
it is possible to assess that Co is mostly present in its trivalent
oxidation state. From quantitative results, the oxygen rate is
similar to the fluorine one, which could induce various mixed (F, O)
anionic environments around the cobalt at the surface: Co-OxFy
and Co-OyFx, in which x is higher than y. This means two Co
environments, one oxygen-rich and the other one fluorine-rich.
Such assumption is reinforced by the presence in O1 s spectrum of
high BE component (Table 2b). The high BE peak at 533.1 eV could
be linked also to oxidized carbon species, which is also observed at
high BE in the C1 s spectrum (Fig. 3b). The main peak in the F1 s
region (Fig. 3b) observed at 684.5 eV (93% of the total area of the
F1 s region) is due to mostly fluorinated environment of Co,
whereas the minor contribution at 686.6 eV (7% of the total area of
100
300
5.174
5.028
13.263
13.218
11
77
Fig. 5. HRTEM image of CoF3 prepared at TF = 100 ꢀC.
function of the fluorination temperature, at TF = 100 and 300 ꢀC, as
shown in Fig. 4. The average CoF3 particle size increases with the
fluorination temperature from 11 to 77 nm (Table 3) for TF = 100
and 300 ꢀC, respectively.
HRTEM images obtained on Co nanoparticles fluorinated at
100 ꢀC are shown in Fig. 5. As expected, nanometric particles are
still observed after the fluorination procedure. It means that the
Fig. 6. Fitted Co2p3/2, O1s, F1 s and C1 s XPS spectra of (a) CoF3-F300, (b) CoF3-F300-AB-dry and (c) CoF3-F300-AB-air samples.
Please cite this article in press as: H. Groult, et al., Nano-CoF3 prepared by direct fluorinationwith F2 gas: Application as electrode material in Li-
ion battery, J. Fluorine Chem. (2016), http://dx.doi.org/10.1016/j.jfluchem.2016.10.003

G Model
FLUOR 8868 No. of Pages 11
6
H. Groult et al. / Journal of Fluorine Chemistry xxx (2016) xxx–xxx
the F1 s region) might be due to some CꢁꢁF bonding, as previously
In conclusion, the above results indicate that during the
fabrication of electrodes, a thin film of oxide/oxide-fluoride is
likely to be formed. This layer, which acts as an external layer, is
assumed to increase the electric conductivity of nano-particles.
Finally, one could also state that the interface layer is not
homogeneous and composed of nano-domains with CoꢁꢁO,
CoꢁꢁOꢁꢁF, etc. bonds. Such surface layer could protect the bulk
fluoride particles, avoiding the exfoliation of the lattice as it occurs
for example with graphite anode during the discharge process. It
induces enhanced electrochemical performances due to the grain
boundary contact. When pristine CoF3 is handled in very dry
atmosphere, the CoF3-based electrodes are assumed to be stable.
observed [29]
3.3. Reactivity of CoF3-based electrode
Before electrochemical investigation, samples were prepared to
test the stability of CoF3 in different atmospheres during the
elaboration process of the electrode. CoF3 and acetylene black (AB),
i.e. 80:20 ratio, were mixed thoroughly in a dry box. A part of the
samples was left for several days in these ultra-dry conditions (O2,
H2O < 2 ppm) (CoF3-AB-dry), whereas another part was brought
for a couple of days in a vial under laboratory atmosphere (CoF3-
AB-air). Results of the XPS investigations of the CoF3-AB mixtures
are shown in Fig. 6a–c. CoF3 powder fluorinated under a stream of
elemental fluorine for 12 h at 300 ꢀC is used as a reference (CoF3-
F300).
3.4. Electrochemical behaviour in Li-ion battery
As observed from XPS experiments, the decomposition of CoF3
occurs in air followed by the release of F2-gas. Consequently, the
use of CoF3 as cathode materials in Li-ion battery requests a severe
control of the working atmosphere. Indeed, as reported above,
conventional battery electrolytes react immediately in contact
with CoF3, leading to partial growth of solid interface film or a
second Co-F phase as demonstrated below. In this study, the
changes in the structural powders were analyzed by XRD
diffraction carried out before and after lithiation. Fig. 7 displays
the first discharge profile of a CoF3//Li cell that displays the states
of charge (dots) at which XRD diffractograms were collected
(Fig. 8): (a) raw CoF3 nano-powder prepared by direct fluorination
of nano-Co (b) electrode assembled in dry box before testing, (c)
electrode discharged at 1.2 V vs. Li and (d) electrode discharged at
0.02 V vs. Li. The XRD diagram of the grid supporting the active
material is also shown in (e). The diffractograms (a) and (b) are
similar and do not reveal any significant modification of the
structure and composition of the electroactive material. XRD
patterns exhibit one set of diffraction lines belonging to CoF3 and
A comparison of this reference sample with the one obtained
after fluorination at TF = 100 ꢀC (nano-Co F100 in Fig. 3) shows that
ꢂ10 at.% oxygen are only present at the surface, with respect to
25 at.% for Nano-Co F100 and 36.5 at.% F (Table 1). [Note that in
Fig. 6a, the fitted O1 s spectrum is represented full scale, despite
the small amount of surface oxygen]. The F:Co ratio (at.%) deduced
from these XPS analyses is about 3, in good agreement with the
chemical composition of CoF3. The Co2p3/2 spectrum is free from
the oxide/oxide-hydroxide component at low BE. Deconvolution of
this spectrum shows four components appearing at 782.6, 784.6,
788.8 and 793.2 eV. The first two components are the fingerprints
of Co in (oxy)fluorinated environment as follows: Co-OyFx (x > y)
and a more fluorinated environment, Co-Fz; the latter two
components being satellite peaks. From a comparison of these
XPS data with those of nano-Co fluorinated at 100 ꢀC (Table 2b), we
note a shift to higher BE of 0.8 eV for the main peak position,
together with an increase of the 784.6 eV area for CoF3-F300. These
changes correspond to a same trend. The shift from 781.8 to
782.6 eV could be due to a diminution of the number of oxygen
ligands around Co in Co(O;F) particles, whereas the increase of the
component at 784.6 eV could be assigned to an increase of the
number of Co-F bonds after the treatment at TF = 300 ꢀC treatment.
We should remark that for both fluorination treatments, the latter
(higher) BE is in agreement with that of Co3+-F bonds, since Co2+-F
bonds occur at lower BE of 783.2 eV for CoF2 [30]. The F1 s
spectrum of CoF3-F300 (Fig. 6a) is similar to that of nano-Co F100
(Fig. 3b). The main peak, observed at 684.9 eV (94% of the total area
of the F1 s region) is mostly due to fluorinated environments of
cobalt, FzꢁꢁCo. Concerning the O1 s spectrum, it could be noted an
increase of the component at higher BE: 53.3 vs. 30.4% Ii (%), which
corresponds to Co-OyFx with higher amount of fluorine.
peaks at 2u
= 26.80, 34.05 and 52.19ꢀ corresponding to a small
amount of CoF2. Note that diffraction lines relative to the presence
of the additive carbon cannot be observed due to its amorphous
nature.
From the XRD patterns of lithiated electrodes at 1.2 and 0.02 V,
we clearly observed the disappearance of the Bragg lines of pristine
CoF3, while a broad bump occurring at about 2
attributed to the appearance of metallic Co in good agreement with
the 2 -Co. Further, the electrochemical
line at 45.92ꢀ of
u
= 45ꢀ can be
u
a
discharge/charge tests were performed at low current density of
ꢃ5 mA/g for both CoF3 electrodes prepared at TF = 100 and 300 ꢀC.
When the sample is prepared and kept in very dry conditions
CoF3-F300-AB-dry (Fig. 6b), the Co 2p3/2, O1s, and F1 s XPS spectra
are similar with those of pristine CoF3-F300 sample (surface O
content = 7 at.%). However, various modifications are observed for
the sample left in ambient atmosphere CoF3-F300-AB-air (Fig. 6c),
in which the oxygen rate increases to 11 at.%. All Co2p, O1 s and F1 s
binding energies are modified, with the re-occurrence of oxidized
forms of Co at 780.9 eV and O1 s at 529.9 eV. In addition a shoulder
appears in the C1 s envelope for higher BE, above 288 eV. These
features and the appearance of a strong contribution at 686.7 eV in
the F1 s spectrum can be related to the decomposition of CoF3 in air
followed by release of F2-gas, which immediately reacts with
acetylene black to give CꢁꢁF bonds, similarly to those found in
carbon fluorides. It can be noted that in the XPS spectra of electrode
samples, the amount of surface oxygen for the sample left in air is
higher than the one of sample prepared in dry atmosphere and in
both cases the amount of carbon is increased because of the
presence of acetylene black.
Fig. 7. First discharge curve of a CoF3//Li cell. Dots are the states of discharge at
which XRD diffractograms were collected.
Please cite this article in press as: H. Groult, et al., Nano-CoF3 prepared by direct fluorination with F2 gas: Application as electrode material in Li-
ion battery, J. Fluorine Chem. (2016), http://dx.doi.org/10.1016/j.jfluchem.2016.10.003

G Model
FLUOR 8868 No. of Pages 11
H. Groult et al. / Journal of Fluorine Chemistry xxx (2016) xxx–xxx
7
the conversion reaction of cobalt fluoride into metal Co nano-
particles embedded into a LiF matrix according to the well-known
relation [3,5,33]:
CoFx + x Li+ + x eꢁ ! Co + x LiF.
with x = 2 and/or 3.
(3)
The structural evolution of CoF3 electrode studied by XRD
(Fig. 8) recorded at potential of ꢂ1.2 V (plateau) and at 0.02 V (end
of the discharge) ensures the conversion reaction. Note that the
observed extra capacity given the length of the plateau should arise
from an interfacial interaction of lithium within the Co/LiF matrix
[10,26,33] in addition to the formation of a polymer-like solid
electrolyte interphase (SEI) film [8]. After the complete reduction
(i.e. at 0.02 V), the galvanostatic cycles exhibit a continuously
decreased/increased profiles without voltage plateau, which is a
characteristic of conversion reaction. Also, the detection of the low
intensity XRD peaks of metallic cobalt with a bump shape confirms
the reversibility of the conversion mechanism.
As mentioned in the introduction, insertion/extraction of Li+ in
layered materials generally occurs in charge/discharge cycles
without any significant structural modifications. Thus, even if the
occurring volume expansion is due to the insertion of Li+, the
original host structure is restored after Li+ extraction. However,
their ability to insert a large amount of lithium is often limited to
about one lithium per mole of electroactive material. In contrast,
when the electrode material is composed of transition-based metal
fluoride, multiple-electron transfer can occur, and thus the metal
fluoride can be fully reduced to metallic state according to the
conversion reaction (Eq. (3)). Whatever the metal fluoride, the cell
voltage is strongly influenced by an overvoltage as described in
Refs. [8,10,34], which results from kinetic barriers. If the
overvoltage is large, the voltage variation is significantly influ-
enced by the kinetic terms and the voltage during the experiment
reaches the operating oxidation or reduction potential limit.
Consequently, during extraction, Li cannot usually be fully
extracted from the host lattice as it is claimed in the case of
CoF3. In other words, after the first reduction step, metallic cobalt
cannot be converted back to Co3+ but to Co2+, thus leading to a
progressive suppression of CoF3 domains as reported before
[35,36]. Therefore, after the first cycle, the reaction involved in the
electrochemical process is the following:
Fig. 8. XRD patterns of CoF3-based electrodes: (a) raw CoF3 nano-powder prepared
by direct fluorination of nano-Co with F2 (nano-Co F100), (b) Electrode assembled in
dry box before testing, (c) XRD pattern recorded after discharge to 1.2 V vs. Li, (d)
XRD pattern recorded after discharge to 0.02 V vs. Li and (e) stainless steel grid. Dots
indicate the diffraction lines of the CoF2 phase.
The discharge (Qdisch.) and charge (Qch.) capacity calculated in the
potential range 0.02–4 V vs. Li0/Li+, as well as the coulombic
efficiency (denoted CE) are reported in Tables 4 and 5.
The electrochemical potentials of the LiF/Co composite have
discussed by Zhou et al. [31] and those of cobalt fluoride films by Fu
et al. [32]. The theoretical voltage of the reduction of CoF3 by Li
linked by the reaction Li + CoF3 ! LiF + CoF2 is 4.85 V, while an
open-circuit voltage (OCV) of about 3.5 V was obtained in
experimental cell [32]. The electrochemical reaction of CoF2 with
Li was found partially reversible with a potential plateau at 1.1 V for
the first discharge that has been attributed to the formation of
metallic cobalt. However, it was reported that usual electrolytes
such as LiPF6 in EC-DMC have poor stability with the Co fluoride
electrode. High OCV values can be only observed during the first
discharge. The potential drops rapidly (Fig. 7) from an OCV of about
3.5 V to ꢂ1.2 V vs. Li00/Li+ depending on the current density and
then reaches a potential plateau due to the two-phase character of
the discharge. These results demonstrate very similar behaviour
reported by Zhou et al. [31].The presence of such low voltage
plateau below the theoretical operating potential is attributed to
CoF2 + 2 eꢁ + 2 Li+ $ Co + 2 LiF.
(4)
Finally, Maier et al. [10] reported a high overvoltage value of
1.58 V for Li-insertion reaction into CoF2.
Table 4
Influence of the fluorination temperature (TF) on the charge and discharge
capacities of CoF3 prepared under F2 atmosphere. [Values deduced from the 1st
cycle in 1 mol/L LiPF6-EC:DMC (1:1). CE: coulombic efficiency. Current density: ꢃ
5 mA/g].
TF (ꢀC)
Qdisch. (mAh/g)
Qch. (mAh/g)
CE (%)
100
300
1011
992
418
257
41.3
25.9
Table 5
Influence of the current densities on the charge and discharge capacities of CoF3
prepared at TF = 100 ꢀC under F2 atmosphere. [Values deduced from the 1st cycle in
1 mol/L LiPF6-EC: DMC (1:1). I = ꢃ5 mA/g. CE: coulombic efficiency (in%)].
I (mA/g)
Qdisch. (mAh/g)
Qch. (mAh/g)
CE (%)
5
20
140
694
41011
1046
824
418
385
257
101
41.3
36.8
31.1
29.3
Fig. 9. Galvanostatic charge-discharge curves of nano-CoF3 (TF = 100 ꢀC) in 1 mol/L
LiPF6-EC:DMC (1:1). I = ꢃ 5 mA/g.
345
Please cite this article in press as: H. Groult, et al., Nano-CoF3 prepared by direct fluorinationwith F2 gas: Application as electrode material in Li-
ion battery, J. Fluorine Chem. (2016), http://dx.doi.org/10.1016/j.jfluchem.2016.10.003

G Model
FLUOR 8868 No. of Pages 11
8
H. Groult et al. / Journal of Fluorine Chemistry xxx (2016) xxx–xxx
Fig. 11. Evolution of the charge and discharge capacities vs. cycle number obtained
in 1 mol/L LiPF6-EC:DMC (1:1) for CoF3 prepared at 100 and 300 ꢀC (current: ꢃ 5 mA/
g).
Fig.10. First galvanostatic charge-discharge curves of nano-CoF3 (TF = 100 ꢀC) under
various current densities.
No significant difference in the average plateau voltage and Q
values has been observed for low current density (ꢄ20 mA/g)
indicating that the diffusion of lithium ions controls the reaction.
This difference is more pronounced for higher rates. Thus, the
potential plateau decreases from around 1.5 to 1.0 V vs. Li00/Li+ at 5
and 140 mA/g due to internal polarization, respectively. The
capacities decrease also significantly with increasing current
densities (Table 5). Thus, Qch. decreases from 418 to 101 mAh/g
and the coulombic efficiency from 41.3 to 29.3%, for increasing
current densities from 5 to 694 mA/g, respectively.
The evolution of Qch. and Qdisch. vs. cycle number of in 1 mol/L
LiPF6-EC:DMC (1:1) electrolyte (rate: ꢃ5 mA/g) for CoF3 prepared
with two different fluorination temperatures is presented in Fig.11.
Reversible and stable capacities are observed after subsequent
cycles for the sample prepared at TF = 100 ꢀC (390 mAh/g), whereas
a drastic decay of the discharge capacity is observed for powders
prepared at TF = 300 ꢀC. This behaviour is explained by the particle
size, which is the largest for the sample prepared at TF = 300 ꢀC.
Thus, the capacity retention is higher for TF = 100 ꢀC than for
TF = 300 ꢀC.
Fig. 9 shows the galvanostatic charge-discharge curves of nano-
CoF3 (TF = 100 ꢀC) in 1 mol/L LiPF6-EC:DMC (1:1) at current density
5 mA/g. In these experimental conditions, the first charge/
discharge profile does not clearly evidence the presence of a
potential plateau but only a shoulder at around 2.4 V vs. Li00/Li+
(see inset of Fig. 9), which could be related to the insertion/
extraction of Li+ into/from the host lattice according to CoF3 +
xLi+ + xeꢁ ! LixCoF3 (x < 1). The capacity associated to this reaction
is very small, i.e. there is no evidence about the formation of a
ternary LixCoF3 compound. However, the origin of the source of
cobalt should have a decisive role in the present fluorinated
process. An optimization of the electrode preparation should be
necessary to improve the knowledge about such a reaction. After
the plateau, a continuous voltage drop down to the potential limit
(0.02 V vs. Li00/Li+) has been observed [10,34]. The first charge
capacity value is much higher than the theoretical one (694 mAh/g)
corresponding to the conversion reaction of CoF3. For instance, the
capacity Qdisch = 1011 mAh/g is obtained in the case of a fluorina-
tion temperature of 100 ꢀC, corresponding to the reaction with
about 4.4 Li ions per unit formula. As mentioned above, Li cannot
usually be fully extracted during the oxidation step, i.e. after the
first reduction step and Co metal cannot be converted back to Co3+
but to Co2+ [26]. Consequently, the CoF3 domains disappear after
the first charge step and the electrochemical reaction implies CoF2
according to the reaction Eq. (4). The theoretical capacity related to
this reaction is 553 mAh/g.
Finally, the lithiation conversion mechanism of CoF3 at the
electrode/electrolyte interface was investigated using a galvano-
static intermittent titration method coupled with electrochemical
XRD measurements have evidenced that an increase of TF
induces a strong modification of the particles size and has a direct
impact on the electrochemical performance of cobalt fluoride in Li-
ion battery, as often pointed out. Here, the highest the particle size
(Table 3), the lowest the irreversible capacity deduced at the end of
the first cycle, i.e. Qdisch ꢁ Qch., and the lowest the reversible
capacity (Qch.). Thus, for TF = 100 and 300 ꢀC, the reversible capacity
is 418 and 257 mAh/g, respectively (Table 4), for crystallite size
deduced 11 and 77 nm, respectively (see Table 3).
The first discharge-charge profiles obtained with CoF3 prepared
at TF = 100 ꢀC are given in Fig. 10 for various current densities. The
difference between the first and the subsequent cycle profiles
should arise from the occurrence of irreversible structural
modifications of the electrode materials during the insertion of
Li+ ions. An increase in the discharge current causes a decrease in
the charge voltages, which means that the values of the potential
plateau are strongly dependent on the discharge rates.
Fig. 12. OCV and CCV curves recorded after fourteen discharge/charge cycles
obtained in 1 mol/L LiPF6-EC:DMC (1:1) (rate: 5 mA/g) for nano-CoF3 (nano-Co
F100). Impedance measurements were carried out at each point indicated on the
curve. Duration of each galvanostatic pulse: 3 h. Rest time before impedance
measurement: 15 h. Inset: electric equivalent circuit for the electrode/electrolyte
interface.
Please cite this article in press as: H. Groult, et al., Nano-CoF3 prepared by direct fluorination with F2 gas: Application as electrode material in Li-
ion battery, J. Fluorine Chem. (2016), http://dx.doi.org/10.1016/j.jfluchem.2016.10.003

G Model
FLUOR 8868 No. of Pages 11
H. Groult et al. / Journal of Fluorine Chemistry xxx (2016) xxx–xxx
9
Fig.13. Experimental and fitted impedance spectra obtained at different points of the discharge curve given in Fig.11: (a) point 1, (b) point 3, (c) point 6, (d) point 12, (e) point
18 and (f) point 22.
impedance spectroscopy (EIS) measurements in the case of sample
prepared at TF = 100 ꢀC, which gives rise to the best electrochemical
performance. It is the reason why we decided to perform
impedance measurements after performing fourteen charge/
discharge cycles, i.e. after reaching a stabilization of the charge/
discharge cycling. EIS data at different points of the discharge curve
are illustrated in Fig. 12. To succeed, a constant discharge current is
applied to the electrode during a time t to reach a certain discharge
state. Then, the current is interrupted and a rest time of several
hours is imposed to the system. During this relaxation period, the
cell voltage tends to a constant value indicating that the system
tends to equilibrium, and after t = 4 h EIS measurements are
performed. Then, a further galvanostatic pulse is applied to the cell
and so on. In the present case, a current of 5 mA/g is applied to the
electrode for 3 h and the duration of the rest time step is 15 h. Point
referenced #6 on the curve presented in Fig. 12 is reached after six
successive discharge/rest time sequences. The experiments were
done until the potential reached 0.02 V vs. Li00/Li+ and gives rise to
the variation of the OCV (open circuit voltage) and CCV (closed
circuit voltage) curves. EIS spectra recorded at steps of discharge
referenced 1, 3, 6, 12, 18 and 22 to follow the influence of the depth
of discharge on the spectra are given in Fig. 13a–f, respectively.
Before studying the shape and the evolution of the impedance
spectra vs. discharge states, let us consider briefly the evolution of
the CCV and OCV curves (Fig. 12). The difference between OCV and
CCV was defined as the cell polarization. The polarization is about
0.75 V at the beginning of the discharge. Then, the polarization
gradually decreases with the increase in discharge to reach a rather
constant value of about 0.25 V after few discharge/rest time
sequences. This trend is consistent with the progressive generation
of conducting Co nano-clusters during the discharge.
As shown in Fig. 13 for the first discharge states, the EIS
diagrams exhibit a semi-circle at high-frequency range followed by
a very large second one. At the beginning of the discharge process,
Please cite this article in press as: H. Groult, et al., Nano-CoF3 prepared by direct fluorinationwith F2 gas: Application as electrode material in Li-
ion battery, J. Fluorine Chem. (2016), http://dx.doi.org/10.1016/j.jfluchem.2016.10.003

G Model
FLUOR 8868 No. of Pages 11
10
H. Groult et al. / Journal of Fluorine Chemistry xxx (2016) xxx–xxx
Fig. 14. Evolution of Re, Rct, RSEI and QSEI (inset) deduced from the fitting of the impedance spectra for different depth of discharge of nano-Co F100 sample (i.e. at different
points of the discharge curve given in Fig. 11).
the charge transfer is difficult to observe since the two semi-circles
obtained in this frequency domain overlap and the fitting of the
experimental curve is therefore more difficult. The contribution of
this loop hinders also the Warburg phenomenon at low frequen-
cies. Then, for deeper discharge states, the influence of the charge
transfer is reduced and a second semi-circle appears (for example
see Fig. 13d); in the same time, an inclined line with a slope of 45ꢀ
characteristics to Warburg contribution appears in the low
frequency region. Due to their close time constants, data cannot
be well extracted directly from the experimental curves and a
fitting of the experimental curves using the equivalent circuit given
in the inset of Fig. 12 is required. The electric equivalent circuit
used for the fitting of the experimental diagrams is composed of Re
relative to the ionic resistance of the electrolyte resistance in series
with an inductance L and with two (R, Q) loops. Note that constant
phase elements (CPE), i.e. pseudo-capacitance Q, are used instead
of pure capacitance to take into account the non-homogeneous
nature of the composite electrode and the electrode/electrolyte
interface. The first loop (high frequency region) is composed of the
resistance of the SEI layer generated during the discharge process
in parallel with its corresponding pseudo-capacitance QSEI. The
second loop (middle frequency range) contains the contribution
double layer pseudo-capacitance, Qdl, in parallel with two terms in
series: the charge transfer resistance (Rct) and the contribution of
Warburg diffusion (W). The latter must be considered for a perfect
fitting of the experimental curves even if it is not discussed
hereafter. The main data extracted from the fitting of the
impedance diagrams are given in Fig. 13.
obtained for example with FeF3 [12], the nano-Co formation
facilitates the cell reaction by enhancing electronic contact
between particles.
Finally, the variation of the pseudo-capacitance of the SEI
(denoted QSEI) is presented in the inset of Fig. 14. QSEI does not vary
significantly in the explored region vs. DOD, indicating a good
stability of the SEI layer upon discharge. The variation of QSEI vs.
DOD is in good accordance with the evolution of the associated
resistance The RSEI slightly decrease until point 9, before tending to
a rather constant value, is interpreted by a small increase of the SEI
thickness until this state and after its stabilization.
It should be noted that cobalt oxide and fluoride are conversion
materials exhibiting during the discharge process a long plateau
which appears in both cases. A simple interpretation of these
plateaus could be hazardous since several reactions can be
involved as the reactivity vs. CoF3, the formation of a passivation
layer, etc., but in our case electrochemical analysis, coupled with
XRD analysis of samples before discharge, partially discharged and
after discharge did not reveal the presence of Co oxide, but only Co
fluorides.
To summarize, the study of the electrochemical conversion
reaction after subsequent cycles by EIS measurements does not
highlight any significant modifications of the electrode materials
such as volume change, growth of the SEI layer, mechanical stress,
microstructural failures, and so on. The SEI layer is formed during
the first steps of the discharge and reaches rapidly a stable state,
allowing good electrochemical features and explaining the good
cycle life observed with the CoF3 nano-powders. The decrease of
the charge transfer resistance is an indication that the conversion
reaction does not give rise to detrimental modifications of the
electrode such as volume expansion or formation of isolated areas
from the bulk. Indeed, the two last phenomena should lead to the
decrease of the electrochemical performances of the battery and
generate local heating capable to strongly affect the security of the
battery, but this is not evidenced here.
Avery good agreement between the experimental and the fitted
curves are obtained (Fig. 13) in the entire frequency range and
whatever the discharged states. The evolution of Re, Rtc and RSEI vs.
the depth of discharge (DOD) is given in Fig. 14 The electrolyte
resistance Re does not vary with DOD. The resistance of the SEI
slightly decreases from point 1 (284
V) to point 9 (195 V) and then
tends to a rather constant value. This can be interpreted by the
stabilization of the SEI layer in terms of thickness and composition
during the first steps of the discharge process. Thus, the SEI layer
formed during the first discharge step is restored rapidly during
the first cycles. The charge transfer resistance drastically decreases
indicating that the reaction between lithium ions and electrons at
the electrode/electrolyte interface occurs more and more easily
with deeper discharge. The formation of Co/LiF nano-composites
during the discharge process does not constitute a limitation for
the electronic transfer at the interface and in contrast with results
4. Conclusions
Co nanoparticles synthesized in aqueous medium by electro-
chemical route were fluorinated between 100 and 300 ꢀC by F2 gas.
XPS measurements performed on the raw materials have pointed
out the presence of a coverage of the starting Co metal by cobalt
oxide and oxide-hydroxide. The nano-size of CoF3 particles is
maintained after fluorination treatments, at least at TF2 = 100 ꢀC.
Please cite this article in press as: H. Groult, et al., Nano-CoF3 prepared by direct fluorination with F2 gas: Application as electrode material in Li-
ion battery, J. Fluorine Chem. (2016), http://dx.doi.org/10.1016/j.jfluchem.2016.10.003

G Model
FLUOR 8868 No. of Pages 11
H. Groult et al. / Journal of Fluorine Chemistry xxx (2016) xxx–xxx
11
XRD measurements coupled with Rietveld refinements reveal that
the average size of the resulting CoF3 particles depends on the
fluorination temperature, the smallest size (11 nm) being obtained
for TF2 = 100 ꢀC. HRTEM investigations have evidenced well
crystallized cobalt fluoride nanoparticles. Mixtures of active
CoF3 material and acetylene black were tested in order to control
the preparation process of the electrode, in similar conditions as
used in the elaboration of Li-ion batteries. The XPS investigation
has shown that, when handled in very dry atmospheres, CoF3-
based mixtures are stable vs. traces of humidity and can be used to
prepare in good conditions electrodes for Li-ion batteries in safe
conditions.
The fluorinated cobalt nanoparticles prepared by fluorination at
100 ꢀC have been tested in 1 mol/L LiPF6-EC:DMC (1:1) electrolyte
as electrode materials in lithium battery. A reversible capacity of
390 mAh/g is obtained after subsequent cycles. Due to poor
electronic conductivity properties, CoF3 powder cannot accept
high discharge rate and the maximum one is 1C-rate. Efforts
focused on the modification of the surface are in progress to
enhance the electrochemical performances.
[8] P. Poizot, S. Laruelle, S. Grugeon, L. Dupont, J.M. Tarascon, Nature 407 (2000)
496–499.
[9] P. Liu, J.J. Vajo, J.S. Wang, W. Li, J. Liu, J. Phys. Chem. C 116 (2012) 6467–6473.
[10] H. Li, G. Richter, J. Maier, Adv. Mater. 15 (2003) 736.
[11] M. Duttine, D. Dambournet, N. Penin, D. Carlier, L. Bourgeois, A. Wattiaux, K.W.
Chapman, P.J. Chupas, H. Groult, E. Durand, A. Demourgues, Chem. Mater. 26
(2014) 4190–4199.
[12] D. Dambournet, K.W. Chapman, P.J. Chupas, R.E. Gerald, N. Penin, C. Labrugère,
A. Demourgues, A. Tressaud, K. Amine, J. Am. Chem. Soc. 133 (34) (2011)
13240–13243.
[13] M. Bervas, A.N. Mansour, W.-S. Yoon, J.F. Al-Sharab, F. Badway, F. Cosandey, L.C.
Klein, G.G. Amatucci, J. Electrochem. Soc. 151 (2004) A1878–A1885.
[14] S. Grugeon, S. Laruelle, L. Dupont, J.-M. Tarascon, Solid State Sci. 5 (2003) 895–
904.
[15] W. K. Behl, J.A. Read, US patent 2011/0091370 A1 (April 2011).
[16] R. Massart, B. Rasolonjatovo, S. Neveu, V. Cabuil, J. Magn. Magn. Mater. 308
(2007) 10.
[17] J.-C. Bacri, R. Perzynski, D. Salin, V. Cabuil, R. Massart, J. Magn. Magn. Mater. 62
(1986) 36.
[18] S. Neveu, R. Massart, V. Rocher, V. Cabuil, J. Phys. Condens. Matter 20 (2008)
204104.
[19] K. Artyushkova, S. Levendosky, P. Atanassov, J. Fulghum, Top. Catal. 46 (2007)
263.
[20] J. Yang, H. Liu, W.N. Martens, R.L. Frost, J. Phys. Chem. C 114 (2010) 111–119.
[21] K. Yasumoto, N. Koshizaki, J. Mater. Sci. Lett. 16 (1997) 215–217.
[22] P.Y. Keng, B.Y. Kim, I.-B. Shim, R. Sahoo, P.E. Veneman, N.R. Armstrong, H. Yoo, J.
E. Pemberton, M.M. Bull, J.J. Griebel, E.L. Ratcliff, K.G. Nebesny, J. Pyun. ACS-
Nano 3 (2009) 3143–3157.
Finally, one must underline that the mechanism from which the
lithium uptake occurs is not straightforward and requires more
investigations. The use of CoF3 is a challenge because of its high
reactivity. For example, the presence of shoulders during the
charge or the discharge process could be interpreted by the
presence of cobalt fluorides with various oxidation degrees.
[23] M.C. Biesinger, B.P. Payne, A.P. Grosvenor, L.W.M. Lau, A.R. Gerson, R.S.C. Smart,
Appl. Surf. Sci. 257 (2011) 2717–2730.
[24] V. Papaefthimiou, T. Dintzer, V. Dupuis, A. Tamion, F. Tournus, A. Hillion, D.
Teschner, M. Haevecker, A. Knop-Gericke, R. Schloegl, S. Zafeiratos, ACS Nano 5
(2011) 2182–2190.
[25] N. Motohira, T. Takai, N. Kamiya, K.-I. Ota, Bull. Chem. Soc. Jpn. 72 (1999) 1403.
[26] K.S. Kim, Phys. Rev. B 11 (1975) 2177–2185.
[27] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-ray
Photoelectron Spectroscopy, Physical Electronics, MN, 1995.
[28] A. Hepworth, K.H. Jack, R.D. Peacock, G.J. Westland, Acta Crystallogr. 10 (1967)
63–69.
[29] A. Tressaud, E. Durand, C. Labrugère, J. Fluorine Chem. 125 (2004) 1639–1648.
[30] J. Tan, L. Liu, S. Guo, H. Hu, Z. Yan, Q. Zhou, Z. Huang, H. Shu, X. Yang, X. Wang,
Electrochim. Acta 168 (2015) 225–233.
Acknowledgments
The authors express their gratitude to José Gomez (UPMC,
France) for his technical assistance, Sandra Casale (UPMC, France)
for TEM measurements and Stefan Borenjztein (UPMC, France) for
SEM experiments.
[31] Y. Zhou, W. Liu, M. Xue, L. Yu, C. Wu, X. Wu, Z. Fu, Electrochem. Solid-State Lett.
9 (2006) A147–A150.
[32] Z.-W. Fu, C.-L. Li, W.-Y. Li, J. Ma, Y. Wang, Q.-Z. Qin, J. Electrochem. Soc. 152 (2)
(2005) E50–E55.
References
[33] R.E. Doe, K.A. Persson, Y.S. Meng, G. Ceder, Chem. Mater. 20 (2008) 5274–5283.
[34] F. Leroux, G.R. Coward, W.P. Power, L.F. Nazar, Electrochem. Solid-State Lett. 1
(1998) 255–258.
[35] S. Laruelle, S. Grugeon, P. Poizot, M. Dolle, L. Dupont, J.M. Tarascon, J.
Electrochem. Soc. 149 (2002) A627.
[1] G.G. Amatucci, N. Pereira, J. Fluorine Chem. 128 (2007) 243.
[2] H. Arai, S. Okada, J.-I. Yamaki, J. Power Sources 68 (1997) 716.
[3] F. Badway, N. Pereira, F. Cosandey, G.G. Amatucci, J. Electrochem. Soc. 150
(2003) A1318.
[4] Y. Makimura, A. Rougier, J.M. Tarascon, Appl. Surf. Sci. 252 (2006) 4587.
[5] H. Li, P. Balaya, J. Maier, J. Electrochem. Soc. 151 (2004) A1878.
[6] Z. Johnson, S. Cordova, G.G. Amatucci, SAE Intern. J. Aerospace 1 (2009) 1018.
[7] N. Pereira, F. Badway, M. Wartelsky, S. Gunn, G.G. Amatucci, J. Electrochem. Soc.
156 (2009) A407.
[36] F. Badway, A. Mansour, N. Pereira, J.F. Al-Sharab, F. Cosandey, I. Plitz, G.G.
Amatucci, Chem. Mater. 19 (2007) 4129–4141.
Please cite this article in press as: H. Groult, et al., Nano-CoF3 prepared by direct fluorinationwith F2 gas: Application as electrode material in Li-
ion battery, J. Fluorine Chem. (2016), http://dx.doi.org/10.1016/j.jfluchem.2016.10.003