Novel Synthesis of Anhydrous and Hydroxylated CuF2 Nanoparticles and Their Potential for Lithium Ion Batteries

Article abstract of DOI:10.1002/chem.201800207

Anhydrous nanoscopic CuF₂ is synthesized from alkoxides Cu(OR)₂ (R=Me, tBu) by their reaction either in pure liquid HF at ?70 °C, or under solvothermal conditions at 150 °C using excess HF and THF as solvent. Depending on the synthes

Full text of DOI:10.1002/chem.201800207

A Journal of
Accepted Article
Title: Novel Synthesis of Anhydrous and Hydroxylated CuF2
Nanoparticles and Their Potential for Lithium Ion Batteries
Authors: Erhard Kemnitz, Thoralf Krahl, Friederike Wnkelmann, Andrea
Martin, and Nicola Pinna
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201800207
Link to VoR: http://dx.doi.org/10.1002/chem.201800207
Supported by

10.1002/chem.201800207
Chemistry - A European Journal
- 1 -
Novel Synthesis of Anhydrous and Hydroxylated CuF2
Nanoparticles and Their Potential for Lithium Ion Batteries
Thoralf Krahl1,2, Friederike Marroquin Winkelmann1, Andréa Martin1, Nicola Pinna1,
Erhard Kemnitz* 1,2
1 Humboldt-Universitä
̈t zu Berlin, Institut fü r Chemie, Brook-Taylor-Str. 2, D-12489 Berlin, GERMANY
2 Nanofluor GmbH, Rudower Chaussee 29, D-12489 Berlin, GERMANY
* tel: +49 30 2093 7555, fax: +49 30 2093 7277, mail: erhard.kemnitz@chemie.hu-berlin.de
Abstract
t
Anhydrous nanoscopic CuF2 is synthesized from alkoxides Cu(OR)2 (R = Me, Bu) by their reaction either in pure
liquid HF at -70°C, or under solvothermal conditions at 150°C using excess HF and THF as solvent. Depending on
the synthesis method, nanoparticles of sizes between 10 and 100 nm are obtained. The compound is highly
hygroscopic and forms different hydrolysis products under moist air, namely CuF2·2H2O, Cu2(OH)F3 and Cu(OH)F,
of which only the latter is stable at room temperature. CuF2 exhibits an electrochemical plateau at a potential of
~2.7 V when cycled versus Li in half cell Li-ion batteries, which is attributed to a non-reversible conversion
mechanism. The cell capacity in the first cycle depends on the particle size, being 468 mAh g-1 for ≈8 nm crystallite
diameter, and 353 mAh g-1 for ≈12 nm crystallite diameter, referred to CuF2. However, such a high capacity cannot
be sustained for several cycles and the capacity rapidly fades out. The cell voltage decreases to ~2.0 V for
CuF2·2H2O, Cu2(OH)F3 and Cu(OH)F. Because all the compounds studied in this work show irreversible conversion
reactions we conclude that copper-based fluorides are unsuitable for Li-ion battery applications.
1 Introduction
The search for new materials for Li-ion batteries has become essential during the last years,
because the compounds used nowadays as cathodes for commercial applications have
reached their limit of performance. The chemistry involved in these electrochemical
processes is based on the intercalation reaction of a structure able to intercalate Li+ ions in a
reversible way. The commonly used compounds such as LiFePO4 or LiCoO2, with practical
capacities comprised in the range of 120-180 mAh g-1, can theoretically host no more than 1
Li per unit formula, hence limiting the theoretical capacity of these materials. One of the
strategies that the scientific community has pursued is based on materials storing lithium
ions via conversion reactions enabling more than one lithium to react per formula unit. This
reaction allows reaching a much higher capacity than those obtained with intercalation
materials. This class of compounds comprises mostly metal oxides and fluorides.[1] The
conversion reaction follows the scheme below:
MaXb + bn Li ⇄ a M + b LinX
(1)
M = transition metal, n = stoichiometry of Li against X, X = anion (O, F)
Among these compounds, only the metal fluorides have redox potentials in the region of 2.6-
3.5 V being high enough to be used as cathode materials in Li-ion batteries. These high
potentials are due to the high electronegative character of the metal fluorine bond, and hence,
due to the high free energy of formation of LiF. The most studied new materials for
application in Li-ion batteries are the fluorides of Fe, Co, Ni and Cu (Table 1). In praxis, all
these transition metal fluorides display operating potentials lower than theoretically
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 2 -
predicted. This is caused by the very low conductivity of this class of compounds arising from
their high ionic character, and hence, the slow kinetics of the conversion reaction.
Table 1. Theoretical potential and capacity of selected transition metal fluorides.[1b]
Theoretical
potential (V) [a]
Theoretical capacity
(mAh g–1) [b]
Electronic
conductivity
Compound
FeF2
FeF3
CoF2
NiF2
CuF2
2.66
2.74
2.85
2.96
3.55
571
712
553
554
528
none
none
poor
poor
none
[a] MFn/M vs. Li/LiF, [b] Cth = nF/Mw
Among these metal fluorides, the copper(II) fluoride CuF2 presents the highest theoretical
potential of 3.55 V and a theoretical capacity of 528 mAh g-1. These properties make it a very
interesting electrode material for application in Li-ion batteries.[2]
However, in the literature, very poor cyclabilities have been reported, which may be due to
volume changes during the conversion reaction or dissolution of the active material along the
cycles.[3] Typically adopted strategies to overcome the poor performance of transition metal
fluorides, which is mainly attributed to their insulating character, are the reduction of the
particle size and their combination with conductive agents.[4] It can be stated that in general
the application of such materials in nanoparticular form is favored over materials consisting
of large crystallites.
It has also been shown recently that protecting the metal fluorides from the dissolution
through coating or the use of suitable additive in the electrolyte could improve the cycling
stability.[5]
The synthesis of anhydrous CuF2 is a challenging task. In aqueous chemistry, either the
dihydrate CuF2·2H2O or various hydroxide fluorides are obtained. Dehydration of CuF2·2H2O
does not lead to CuF2, but rather to different hydrolyzed products like Cu(OH)F, Cu2(OH)F3,
Cu3(OH)2F4 or even CuO depending on the reaction conditions.[6]
Anhydrous CuF2 is a highly hygroscopic white solid crystallizing in a Jahn-Teller distorted
rutile structure.[7] There are only few synthesis methods known. CuF2 can be synthesized as
bulk material by the reaction of Cu metal with F2 or a mixture Cl2/F2; or by reaction of copper
halides with F2.[8] CuO or CuF2·2H2O can be reacted with a stream of dry gaseous HF at high
temperatures (>300°C) to form CuF2.[9] In a different approach, first the complex fluorides
NH4CuF3 and (NH4)2CuF4 are formed by reaction of appropriate precursors (CuCl2, CuBr2 or
CuF2·2H2O) with NH4F or NH4HF2 either in a solvent or as solid-solid reaction, followed by
thermal decomposition.[10]
The well-established thermal decomposition of metal trifluoroacetates into fluorides,
especially for Ae(CF3COO)2 (Ae = Ca, Sr, Ba),[11] fails in the case of Cu(CF3COO)2, ending up
with a mixture of CuO and Cu2O.[12]
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 3 -
Most of these publications are quite old and are lacking some of the modern analysis
methods. Moreover, these methods are based on high temperature syntheses, resulting in
highly crystalline bulk material rather than nanoparticles.
The aim of this publication is to explore new approaches for the direct synthesis of nanoscale
CuF2 particles based on soft chemistry, and also to verify some of the previously reported
methods.
A promising candidate for the synthesis of copper fluoride is the fluorolytic sol-gel synthesis,
which was successfully applied to many different binary metal fluorides in the last decade
(MIF, MIIF2, MIIIF3; MI = Li, MII = Mg, Ca, Sr, Ba, Zn; MIII = Al, Y, Yb).[13] However, the synthesis of
copper(II) compounds was yet not reported. An appropriate metal precursor is reacted with
anhydrous HF dissolved in a non-aqueous solvent (mostly alcohols or ethers), resulting in
sols of nanoscale metal fluorides:
solv
MX푛 + 2 aHFsolv
→
nano-MF푛 + 2 HX
(2)
(M = metal; X = halide, alkoxide, carboxylate, etc.; solv = alcohol, ether; aHF = anhydrous HF)
In general, the most appropriate precursors are metal alkoxides, carboxylates or halides. The
optimal precursor depends on the metal ion, solvent and reaction conditions.
In this work, the fluorolytic sol-gel synthesis of anhydrous CuF2 is reported, as well the use of
CuF2 as positive electrode in Li-ion batteries. Cyclic voltammetry and galavanostatic cycling
are applied to investigate the electrochemical behavior. For a deeper understanding of the
lithiation mechanisms, oxygen-containing fluorinated copper compounds CuF2·2H2O,
Cu(OH)F and Cu2(OH)F3 were synthesized and their electrochemical properties compared.
2 Results and Discussion
2.1 Synthesis of anhydrous CuF2 and related compounds
2.1.1 Thermodynamics of the formation of CuF2
According to the hard and soft (Lewis) acids and bases (HSAB) concept, the copper(II) ion is
comparatively soft. This fact is a significant difference to the ions of Mg, Ca, Sr, Ba and Zn, for
which the fluorolytic sol gel synthesis of nanoscale binary fluorides is well known.[13] The
affinity of Cu2+ towards F– is low, but it is much higher towards the heavier halogens Cl– and
Br–. To illustrate the different reactivity of the different divalent metal ions, reaction
enthalpies for the fluorination of anhydrous divalent metal chlorides and acetates are given
in Table 2.
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 4 -
Table 2. Standard reaction enthalpy of divalent metal precursors with hydrogen fluoride (in kJ mol-1).
OAc- = CH3COO- (acetate). Thermodynamic data is taken from ref. [14]
MCl2(s) + 2 HF(g)
MF2(s) + 2 HCl(g)
→
M(OAc)2(s) + 2 HF(g)
→ MF2(s) + 2 HOAc(g)
M(OH)2(s) + 2 HF(g)
→ MF2(s) + 2 H2O(g)
M
Mg
Ca
Sr
-119.6
-138.3
-181.8
-195.8
-209.4
-61.1 [b]
-27.7 [b]
-73.5
-69.3
-48.4
-50.5
-5.3
-27.0 [a]
+3.5 [a]
Ba
Zn
Cu
+11.1
+39.5
+34.9
[a] These chlorides form chlorofluorides upon reaction with HF: MCl2(s) + HF(g)
→
MClF(s) + HCl(g): ΔH = -36.9 (Sr); -25.6 (Ba) kJ mol-1. [b] These fluorides are
hygroscopic and form hydrates.
Thus, the low affinity of Cu2+ towards fluoride must be compensated by a very basic counter
anion X– having a high reactivity towards HF (cf. Table 2, right column). Cu(OH)2 is not
suitable for that purpose, because CuF2 is highly hygroscopic and reacts with the water
formed during the fluorination. Instead, most suitable for this purpose are copper(II)
alkoxides Cu(OR)2. The methoxide and the tert-butoxide were synthesized (cf. experimental
section) and were used in the fluorolytic sol-gel process. Cu(OMe)2 is the base compound,
which can be directly synthesized from CuCl2 and LiOMe. Cu2+ is a strong oxidizing agent, and
therefore, all copper(II) alkoxides of primary and secondary alcohols are decomposed easily
at elevated temperatures into Cu2O or Cu and aldehydes or ketones. Tertiary alcohols are
more stable towards oxidants, and hence, reactions with Cu(OtBu)2 in either tert-butanol or
tetrahydrofuran can also be performed at higher temperatures. Other copper alkoxides of
primary and secondary alcohols can be synthesized in situ by refluxing Cu(OMe)2 with an
excess of the corresponding alcohol and simultaneous evaporation of methanol.
2.1.2 Synthesis of anhydrous CuF2
The simple reaction of stoichiometric amounts of solvated HF with Cu(OR)2 in alcohols or
THF does not lead to the formation of CuF2. Using a large excess of HF (10-20 fold), indeed,
CuF2 may be formed in small amounts (cf. next section), but the product is not pure. For a
complete conversion of the alkoxide, more drastic reaction conditions are necessary: (i)
Higher temperature and pressure using solvated HF under solvothermal conditions, or (ii)
performing the reaction in liquid unsolvated anhydrous HF.
Cu(OR)2 + 2 HF → CuF2 + 2 ROH (R = Me, tBu)
(3)
Cu(OR)2 (R = Me, tBu) reacts with a 10-fold excess of HF in THF solution in a Teflon-lined
autoclave under argon atmosphere at 150°C and 8 bar within 2 hours. CuF2 is obtained as
white precipitate (CuF2-1 from Cu(OMe)2 and CuF2-2 from Cu(OtBu)2 in Table 3). The
supernatant THF has a brownish appearance. Obviously, a part of copper(II) was reduced
(most probably to Cu2O or Cu), but the amount is very small (far below 1%). The product can
be simply decanted or filtered and subsequently washed with THF at -20°C. Unfortunately,
CuF2 seems to have a certain solubility in THF even at -20°C, therefore the synthesis yield is
comparatively low (<50%).
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 5 -
The reaction of Cu(OR)2 (R = Me, tBu) with pure liquid anhydrous HF at -70°C also leads to
white CuF2 within a few seconds upon stirring of the reaction mixture. After allowing the
reaction vessel to heat up to room temperature, excess HF evaporates. Adsorbed HF and the
alcohol formed during the reaction can be pumped off in vacuum. A further cleaning process
is not necessary, pure white CuF2 is obtained directly (CuF2-3 from Cu(OMe)2 and CuF2-4
Cu(OtBu)2 in Table 3). The yields are generally higher than in the autoclave reaction (70-
80%).
Analysis data of the products is given in Table 3, for XRD cf. Figure 1. The fluorine-to-copper-
ratio is nearly 2 in all products. The samples contain about 6-12% of organic residue, which is
typical for nanoparticles synthesized form organic precursors. The organic residue further
stabilizes the surface of the particles.[13] X-ray powder diffractograms of 3 out of these 4
synthesized CuF2 samples show exclusively reflections of CuF2 (Figure 1), which are
broädened due to the smäll cryställite sizes. According to Scherrer’s equätion, the cryställite
sizes range from 14 to 23 nm. Only CuF2-3, i.e. from the reaction of Cu(OMe)2 with liquid HF
at -70°C, is not pure in the XRD, but shows some additional unknown reflections between 10
and 20°.
Table 3. Properties of different samples of CuF2. Calcd (%): Cu 62.58, F 37.42.
Reaction
method
[b]
Name
Reactant
Cu (%) F (%) C (%) H (%) res (%) [a] nF/nCu
XRD [c]
CuF2 only
L ≈ 23 nm
CuF2-1 Cu(OMe)2 solvothermal
CuF2-2 Cu(OtBu)2 solvothermal
58.80 34.50
58.63 34.40
0.37
1.43
2.76
0.53
6.70
6.97
1.96
1.96
CuF2 only
L ≈ 20 nm
nearly
amorphous +
unk. refl.
CuF2-3 Cu(OMe)2
liquid HF
56.32 33.30
56.10 32.20
2.77
4.25
1.48
3.75
10.38
11.70
1.98
1.92
CuF2 only
L ≈ 14 nm
CuF2-4 Cu(OtBu)2
liquid HF
[a] residue = 100–Cu–F. [b] molar ratio of F and Cu in the product. [c] L = size of the coherent scattering region according
to Scherrer’s equätion
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 6 -
1
2
*
3
4
*
calculated for CuF2
10
20
30
40
50
60
2 / °
Figure 1. XRD of different samples of CuF2 (cf. Table 3) and expected diffractogram calculated from
single crystal data. Miller indices for the observed reflections are given above. Note that not all
expected reflections are observed due to their broadness. (#) Overlap of 0 1 2 and 0 2 1. (*) Unknown
phase.
TEM images of CuF2 synthesized under solvothermal conditions show large aggregates of a
size of several hundred nanometers (Figure 2A). The size and shape of the primary particles
cannot be clearly determined, at least some particles are smaller than 50 nm. On the other
hand, the particles of CuF2 synthesized in liquid HF are smaller and less aggregated. The
primary particles are spherical, their size is around 10 nm, and the size of the aggregates
around 30 nm (Figure 2B). Thus, the reaction at -70°C in liquid anhydrous HF leads to
significantly smaller CuF2 particles than the solvothermal synthesis.
A
B
Figure 2. TEM images of (A) CuF2-1 from solvothermal synthesis (scale 50 nm) and (B) CuF2-4 from
reaction in liquid HF (scale 10 nm).
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 7 -
2.1.3 Synthesis of Cu(OR)F, CuF2·2H2O, Cu2(OH)F3 and Cu(OH)F
Cu(OR)F. As already mentioned in the previous section, the reaction of copper(II) alkoxides
with solvated HF at ambient temperature and pressure under argon atmosphere does not
lead to the formation of CuF2. Instead, a new class of compounds Cu(OR)F is formed:
Cu(OR)2 + HFsolv → Cu(OR)F + ROH
(4)
These compounds are not subject of this investigation. Details and further information can be
found in the supporting information (FTIR: Figure S1, XRD: Figure S2).
CuF2·2H2O. Synthesis of phase pure CuF2·2H2O from aqueous solutions is a challenging task,
because it requires accurate control of reaction conditions; otherwise mixtures of different
compounds are obtained. CuF2·2H2O readily undergoes hydrolysis reactions, resulting in
different hydroxide fluorides. The main side product is always Cu(OH)F, and to a lesser
extend Cu2(OH)F3, but even CuO was observed.
Therefore, a new synthesis from non-aqueous solvents was developed in order to reduce the
number of side reactions. The reaction of CuCO3·Cu(OH)2 in ethanol with an excess of
ethanolic HF (molar ratio HF:Cu = 3:1) in an open PFA flask leads nearly quantitatively to the
desired product within one hour under generation of CO2:
CuCO3·Cu(OH)2(s) + 4 HF + H2O → 2 CuF2·2H2O↓ + CO2↑ (in ethänol)
(5)
One equivalent of water is obviously absorbed from the surrounding air. The light blue
precipitate can be filtered off and dried in vacuum. This reaction in ethanol delivers a
nanocrystalline äggregäted product. These äggregätes häve ä size of ≈100-300 nm and
consist of primary particles with a size far below 50 nm (Figure 3A). When the product is
stored under argon, it is stable for long time (at least 6 months). At ambient air, it
decomposes within a couple of days into Cu(OH)F (Figure 4).
A
B
Figure 3. TEM images of (A) CuF2·2H2O from reaction in methanol (scale 50 nm) and (B) Cu(OH)F from
reaction in boiling water (scale 100 nm).
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 8 -
Upon heating under inert gas, the proposed dehydration mechanism of CuF2·2H2O is
different:[6, 15]
≈420°C
130-180°C
→
→
(
)
2 CuF2 ∙ 2H2O
Cu2 OH F3
−HF
Δ푚 =7.27%
CuO + CuF2
(6)
−HF
−3H O
2
Δ푚 =26.92%
2
1
The published temperatures of the two mass loss steps differ from each other. Our own
thermal analysis of CuF2·2H2O in an argon stream revealed a temperature of 176°C for the
first step and 391°C for the second step applying a heating rate of 10 K min-1 (cf. SI Figure S3).
At a lower heating rate of 1 K min-1, the first decomposition step is shifted to 156°C (cf. SI
Figure S4).
*
*
*
*
*
*
**
**
*
*
*
*
*
*
*
*
*
*
*
○
*
under argon
○
CuF2·2H2O
* *
at air
*
○
○
○
*
*
○
*
*
⨯
⨯
⨯
⨯
#
⨯
⨯
⨯
⨯
⨯
⨯
⨯
⨯
⨯
⨯
○
under argon
Cu2(OH)F3
at air
○
○
○
○
○
○
○
under argon
at air
Cu(OH)F
60
10
20
30
40
50
2 / °
Figure 4. XRD of hydrated copper(II) fluorides and hydroxide fluorides as prepared under argon and
after exposure to air for 5 days. (#) One unknown reflection, probably CuO.
Cu2(OH)F3. Cu2(OH)F3 is formed as a side product during reactions in aqueous solutions
containing Cu2+ and HF, but a synthesis of the pure product is not possible in this way.
Thermal dehydration of CuF2·2H2O upon heating in inert atmosphere theoretically leads to
the desired compound (eq. 6), but carried out as a bulk reaction, only product mixtures were
obtained and not a phase pure substance. Thus, the conditions of the thermal decomposition
need to be finely adjusted.
Nearly phase pure Cu2(OH)F3 is obtained upon thermal decomposition of CuF2·2H2O under
self-generated atmosphere (instead of applying vacuum or an inert gas stream) at 160°C:
2 CuF2·2H2O → Cu2(OH)F3 + 3 H2O↑ + HF↑
(7)
The compound is of a very light blueish color. It is highly moisture sensitive and has to be
stored and handled under argon. Although the crystal structure of Cu2(OH)F3 was not solved,
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 9 -
its published XRD pattern is in good agreement with what we report here.[6] At moist air, it
hydrolyzes within a few minutes into Cu(OH)F (Figure 4).
Cu(OH)F. Cu(OH)F seems to be a comparatively stable compound, when water and fluoride
are present in the reaction system. There are many ways for its synthesis. At ambient air, the
compounds CuF2, CuF2·2H2O, Cu2(OH)F3 and Cu(OR)F undergo hydrolysis and are finally
transformed into Cu(OH)F. Under these conditions, Cu(OH)F seems to be a stable
thermodynamic minimum.
Crystallographic pure Cu(OH)F in larger amounts can be easily synthesized according to
Berzelius by boiling CuF2·2H2O for 5 hours in water in an open plastic beaker:[16]
CuF2·2H2O → Cu(OH)F + H2O↑ + HF↑
(8)
The so-obtained product consists of aggregated particles, the size of the aggregates is several
hundred nanometers (Figure 3B). The size of the primary particles is around 100 nm. This is
in accordance with the narrow reflections in XRD (Figure 4). The product is stable at moist air
for at least several weeks, most probably even months (Figure 4).
2.1.4 Summary of synthesis
An overview about the preparative reactions leading to CuF2 and its hydrolysis products is
given in Figure 5.
The most suitable reactants towards CuF2, and especially nano-CuF2, are copper(II) alkoxides
Cu(OR)2. The reaction of Cu(OMe)2 with HF would be the easiest way towards CuF2, but this
reaction requires harsh conditions. At elevated temperature and pressure, only one
methoxide group is fluorinated resulting in the formation of Cu(OMe)F. This compound is
very stable, as long as water is excluded. Even the reaction of Cu(OMe)2 with liquid
anhydrous HF does not lead to pure CuF2. Contrary to this, Cu(OtBu)2 is much more reactive.
It reacts with liquid HF at –70°C to spherical CuF2 nanoparticles below 10 nm in diameter.
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 10 -
CuCO3·Cu(OH)2
+ HF
in MeOH
CuF2
CuF2·2H2O
+ 20 eq HF
in EtOH, r.t.
not quantitative
excess aHF
-70°C
not quantitative
100°C, 5h
in H2O
+ 1...2 eq HF
in EtOH, r.t.
+ 10 eq HF
in THF, 2.5 h
160°C, 8 bar
air
(days)
Cu(OtBu)2
Cu(OEt)2
Cu(OEt)F
Cu2(OH)F3
air
(minutes)
Cu(OH)F
Cu(OMe)2
Cu(OMe)F
+ 1...20 eq HF
in MeOH, r.t.
air
(minutes)
+ 2 LiOMe
− 2 LiCl
air
(several month)
H2O, r.t.
days
CuCl2
CuO
Figure 5. Overview about the synthesis reactions in the CuF2 system. Thick red arrows: most suitable
reactions towards CuF2. Dashed arrows: product mixtures. Green boxes: compounds used for
electrochemical investigations. aHF = anhydrous HF, Me = methyl, Et = ethyl, tBu = tert-butyl, AcOMe =
methyl acetate, AcOtBu = tert-butyl acetate.
In solution, complete fluorination of Cu(OMe)2 or Cu(OtBu)2 to CuF2 is achieved under
solvothermal conditions in THF at 150°C and 8 bar resulting in a nanocrystalline products
forming aggregated particles. In this case, the reaction with Cu(OMe)2 is more convenient,
because the synthesis of Cu(OtBu)2 requires an additional reaction step.
Thus, the two most suitable synthesis reactions for CuF2 are:
1. The reaction of Cu(OMe)2 with excess HF in THF at 150°C for 3 h, resulting in
nanocrystalline aggregated material.
2. The reaction of Cu(OtBu)2 with excess of pure unsolvated HF at –70°C for 30 min,
resulting in slightly agglomerated particles with diameters below 10 nm.
Nearly all compounds containing CuII and fluoride are moisture sensitive. Even the hydrate
CuF2·2H2O is not stable in air. The only notable exception is Cu(OH)F, which is formed upon
exposure to air from CuF2, CuF2·2H2O, Cu2(OH)F3 and Cu(OR)F (R = Me, Et).
2.2 Electrochemical properties
2.2.1 Electrochemistry of anhydrous CuF2
As already mentioned, metal fluorides intrinsically show very low electric conductivity and
their integration in electrode materials is difficult. However, their conductivity can be
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 11 -
increased by combining nanoscale materials with conductive additives. For example, ball
milling with carbon black increases the conductivity, enables homogeneous dispersion of the
two components, and therefore, lowers the insulating character of the active material.[4b]
Upon ball milling of CuF2 with carbon black, diffraction peaks coming from metallic copper
are present (cf. SI Figure S5A). This creation of metallic copper may be due to the reaction
between the CuF2 and the carbon black induced by the energy impact of the mill:
CuF2 + x C → (1–x/2) CuF2 + x/2 Cu + CFx
(9)
XRD measurements on the first peaks of CuF2, after the ball-milling process, point to a
decrease of the crystallite sizes for the both compounds, nevertheless the compound CuF2-4
still exhibits a smaller crystallite size than CuF2-1 (cf. SI Figure S5B).
Two different samples of CuF2, synthesized under solvothermal conditions or in liquid HF (cf.
Table 3), respectively, were evaluated by cyclic voltammetry and galvanostatic
charge/discharge cycling. Figure 6 shows cyclic voltammetry of the small CuF2-4 particles
within two different potential windows.
5
anodic
to 1 V
4
to 2 V
3
2
1
0
-1
cathodic
1
2
3
4
5
Voltage vs. Li/Li+ (V)
Figure 6. Cyclic voltammetry of CuF2-4/CB (cf. Table 3) between 1 and 4.5 V and 2 and 4.5 V. Sweeping
rate is 1 mV s-1. CB: carbon black.
On the cathodic sweep down to 2 V, a broad peak can be seen starting from ~2.7 V, having its
maximum at ~2.5 V. This broad peak can be assigned to the reduction of the fluoride material
to metallic copper and lithium fluoride. Extending the cathodic sweep to 1 V, an additional
broad peak can be seen at 1.5 V and another one after 1 V. As no reaction is expected at such
low potentials, we can assume that these two peaks can be assigned to side reactions
probably with the electrolyte.
On the anodic sweep, no obvious peak is present before ~3.2 V. At this region, a large
oxidation peak is observed, which is assigned to the dissolution of the metallic copper into
the electrolyte (instead of reconversion to CuF2) during the cathodic sweep. To validate this
assumption, a copper plate has been used as cathode for linear sweep potential, which should
simulate the metallic copper produced during the reduction of the copper(II) fluoride. As
expected, a similar signal can be seen in the potential range of 3.5-3.7 V (cf. SI Figure S6).
Furthermore, we can well assign the peak at 2.5 V during the cathodic sweep being the only
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 12 -
reduction of the CuF2, because for the cyclic voltammetry the current produced on the anodic
sweep during the dissolution of the metallic copper is similar between the potential window
2-4.5 V and 1-4.5 V.
2nd
1st cycle
1st cycle
3rd
3rd
2nd
4.5
4.0
3.5
3.0
2.5
2.0
4.5
4.0
3.5
3.0
2.5
2.0
(A)
(B)
2nd
3rd
3rd
2nd
100
1st cycle
400
1st cycle
0
100
200
300
400
500
0
200
300
500
Capacity (mAh g-1)
Capacity (mAh g-1)
Figure 7. Galvanostatic charge/discharge profiles at 0.1C rate of (A) CuF2-4/CB (small particles,
synthesis in liquid HF) and (B) CuF2-1/CB (big particles, solvothermal synthesis), Cf. Table 3 and
Figure 2 for details. CB: carbon black.
Figure 7 shows the three first discharge/charge profiles of two different samples of CuF2
cycled at 0.1C. The main difference between these two compounds is the size of the
crystallites, which after ball milling is roughly 50% larger larger for CuF2-1 synthesized under
solvothermal conditions (cf. SI Figure 5B). For conversion materials in Li-ion batteries a
smaller particle size is generally traduced by an easier diffusion of the electrons and ions
within the electroactive material. This would provide faster kinetics of reaction, better
reversibility and therefore better performances. It is thought that the reaction path length
within this kind of material is around 10 nm.[17] Sizing the particles around this diameter
should minimize, and hopefully overcome the conductivity problems. It will also promote the
dissolution of metallic copper produced during the conversion reaction, because a larger
surface will be exposed to the electrolyte. The samples of CuF2 show a flat plateau located
around 2.7 V, which can be ascribed to the reduction of the CuF2 to metallic copper and
lithium fluoride (cf. CV experiments). For further confirmation, ex-situ X-ray diffraction has
been carried out on the electrode material at the end of the discharge (cf. SI Figure S5a). Clear
broad signals of lithium fluoride can be observed.
The difference of morphology between the two CuF2 samples will bring slightly different
profiles of discharge. The smaller particles (Figure 7A) present a higher capacity release,
namely 413 mAh g-1, the larger particles (Figure 7B) a capacity of 329 mAh g-1, which
represents 78% and 62% of the theoretical capacity, respectively, including the organic
residue (cf. Table 3). This assumption is confirmed by cycling the composite materials at a
higher rate; when the first discharge is operated at 1C, the loss in capacity is smaller for the
compound with smaller crystallites (cf. SI Figure S7). It confirms the better performance for
smaller particles. The following charges show for both compounds a big plateau around 3.5 V,
which can be ascribed to the dissolution of the metallic copper according to the CV
experiment (Figure 6). A second plateau around 4.2 V is caused by another side reaction. This
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 13 -
plateau at high potential could be due to the decomposition of the electrolyte due to the
catalytic power of metal fluorides as it has been already observed in the literature.[18] Almost
no capacity is released during the second discharge of the CuF2-4 material, which could be
due to the total dissolution of the converted metallic copper. It shows that none of the
converted material has been reconverted back during the charge process. The compound
CuF2-1 shows a very small capacity during the second discharge process, which could be due
to the incomplete conversion of the nanoparticles during the first discharge. During the third
discharge almost no capacity is observed.
Table 4. Electrochemical properties of CuF2, Cu(OH)2 and related compounds. Thermodynamic data is
taken from ref.[14] Measured capacity is for the active material only including the organic residue, but
not the carbon black.
Measured
capacity at 0.1C
rate, 1st cycle
Measured
capacity at 0.1C
rate, 10th cycle
Theoretical
potential
(V) [a]
Theoretical
capacity
Measured
potential
(V)
Crystal
structure
Compound
[b]
(mAh g–1
)
(mAh g–1
)
(mAh g–1
)
distorted
rutile [7]
chain
1: 329 (353) [c]
4: 413 (468) [c]
CuF2
3.55
3.46
528
~2.5
~2.0
none
CuF2·2H2O
390
533
538
structure
[19]
565
504
585
77
72
37
Cu2(OH)F3
Cu(OH)F
≈3.2 [d]
≈3.0 [d]
unknown
~2.0
~2.0
layered
structure
[20]
layered
Cu(OH)2
2.62
549
structure
[21]
[a] vs. Li/(LiF,LiOH). [b] Cth = nF/Mw. [c] 1: 100 nm, 93.3% CuF2; 4: 10 nm particle size, 88.3% CuF2 (cf. Table 3).
Values in brackets are corrected for the organic residue and refer to pure CuF2. [d] estimated from CuF2 and
Cu(OH)2 using an arbitrary correction of -0.1 V for the unknown enthalpy of formation of Cu(OH)xF2-x
.
2.2.2 Electrochemistry of CuF2·2H2O, Cu2(OH)F3 and Cu(OH)F
When a certain amount of oxygen replaces fluorine ions in metal fluorides their conductivity
is improved. This should lower the average ionic character of the compound and decrease the
polarity of the metal fluorine bond. The exact role of hydroxide groups in metal fluorides has
not been yet fully discussed, but it is known to play an important role. This has been shown
especially for the iron fluoride system, where the total fluorination of the compound is quite
hard following a soft-chemistry approach. The introduction of water into the structure
facilitates the diffusion of the lithium ions within the framework by creating channels, in
which the ions can diffuse more easily, and hence, improving the performance of the
material.[22] The expected theoretical potentials and capacities of the relevant copper(II)
fluoride hydrate or hydroxide materials are summarized in Table 4. It is expected, that the
cell potential of Cu(OH)xF2-x is lower than that of anhydrous CuF2. This effect is mainly caused
by the lower enthalpy of formation of LiOH compared to LiF.
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 14 -
0.5
0.0
0.5
0.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0
-2.5
-3.0
-0.5
-1.0
-1.5
-2.0
-2.5
-3.0
-0.5
-1.0
-1.5
-2.0
-2.5
-3.0
1st cycle
2nd cycle
1st cycle
2nd cycle
1st cycle
2nd cycle
CuF2·2H2O/CB
Cu2(OH)F3/CB
Cu(OH)F/CB
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
Voltage vs. Li/Li+ (V)
Voltage vs. Li/Li+ (V)
Figure 8. Cyclic voltammetry of different oxygen containing copper(II) fluoride compounds on carbon
black (CB).
5
4
3
2
1
5
4
3
2
1
5
4
3
2
1
1st cycle
2nd
2nd
1st 2nd cycle
1st cycle
10th
10th
10th
CuF2·2H2O/CB
Cu2(OH)F3/CB
Cu(OH)F/CB
2nd
10th
1st cycle
2nd
10th
2nd
1st cycle
10th
1st cycle
0
100 200 300 400 500 600
0
100 200 300 400 500 600
Capacity (mAh g-1)
0
100 200 300 400 500 600
Figure 9. Galvanostatic discharge/discharge of oxygen containing copper(II) fluoride compounds on
carbon black (CB) at 0.1C rate.
Figure 8 shows cyclic voltammetry and Figure 9 the galvanostatic discharge/charge profiles
at different cycles of the oxygen-containing materials. Here, the potential window has been
chosen to be 1-4.5 V allowing the reduction of the compounds, as it is usually done for
bivalent metal fluorides.[23] The galvanostatic profiles of CuF2·H2O, Cu2(OH)F3 and Cu(OH)F
are quite similar, namely a big flat plateau of reduction located around 2 V showing higher
capacity delivery compared to their theoretical capacity. The conversion occurring should be
similar to that of CuF2. This extra capacity is quite common for bivalent metal fluorides,
because the operating potential of these materials is too close to the one of side reactions,
which could occur between the electrolyte and the active material or the current collector.
The following charge does not have specific plateau and the capacity is only partially
recovered.
The subsequent discharges occur at a slightly higher potential, which could be ascribed to the
decrease of ionicity of the compounds after the first discharge.[24] The three compounds show
poor capacity retention (Figure 10). Charge and discharge show the same profile along all the
cycles.
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 15 -
It is noteworthy to point out that for all of them no dissolution of metallic copper can be seen
on the cyclic voltammetry and galvanostatic discharge/charge. Thus, the reduced compounds
are not exactly the same as the ones formed during the discharge of pure CuF2. Indeed, the
produced insulating phase would not be pure lithium fluoride, so the metallic copper
particles will not be encapsulated into an insulating phase. Furthermore, the reactions
involved in the conversion occur at lower potential. Thus, it could be that the following
reconversion of the materials occur at lower potential as well, avoiding reaching the potential
at which the metallic copper starts dissolving into the electrolyte.
600
CuF2-4/CB (small)
CuF2-1/CB (big)
500
CuF2·2H2O/CB
400
300
200
100
0
Cu2(OH)F3/CB
Cu(OH)F/CB
0
5
10 15 20 25 30 35 40 45 50
Number of cycles
Figure 10. Capacity retention during charge/discharge cycles at 0.1C rate. CB: carbon black. CuF2-4
~10 nm particle size, CuF2-1 ~100 nm particle size (cf. Table 3 and Figure 2 for details).
2.2.3 Summary of electrochemical properties
Anhydrous CuF2 as composite with CB shows a desired high discharge capacity at a relatively
high potential for cathode conversion materials. Smaller particles exhibit a higher capacity
than larger particles due to the higher reactions kinetic. Unfortunately, the following charge
of the compound is concealed by the dissolution of the metallic copper nanoparticles formed
during the discharge. No subsequent capacity is shown arising from a reconversion of the
active material, presenting therefore an irreversible behavior for CuF2. Lithium ion half cells
using our CuF2 materials in electrode formulations show a working potential of ~2.7 V. The
cell capacity depends on the particle size being 413 mAh g-1 for ~8 nm crystallite size and 329
mAh g-1 for ~12 nm crystallite size, which is ~78% and ~62% of the theoretical capacity.
Correcting these values for the organic residue, and thus, referring them to pure CuF2, the
capacities are 468 mAh g-1 (88%) and 353 mAh g-1 (67%). Flat plateaus in the galvanostatic
charge/discharge experiments indicate a conversion mechanism. However, the charge
processes are not reversible and the capacity is nearly lost already after the 3rd
charge/discharge cycle. This non-reversibility is caused by the dissolution of the metallic
copper produced during the previous discharge.
If the process would be reversible the materials would be a strong competitor for replacing
state of the arts compounds since the energy density delivered by the pure CuF2 during the
first discharge is twice as high as the nowadays used compounds as cathode for Li-ion
batteries (~1100 Wh g-1 for the 10 nm sized CuF2-4 against ~500 Wh g-1 for the LiFePO4).
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 16 -
The compounds CuF2·2H2O, Cu2(OH)F3 and Cu(OH)F as composites with carbon black also
show a high capacity discharge, but at a lower working potential of ~2.0 V. This is attributed
to the formation of LiOH, decreasing the free enthalpy of the conversion reaction. Compared
to pure CuF2, they show higher capacities of 500-600 mAh g-1. Additionally, they show a
better reversibility of the electrochemical processes. A complete loss of capacity is reached
after 20-30 charge/discharge cycles. This behavior is attributed to the lithium compounds
formed during reduction. Here, not pure LiF is formed, but rather a mixture of LiF and LiOH.
This composite might not completely cover the formed particles of metallic Cu, ensuring still
a certain electric conductivity. Furthermore, the reconversion mechanism occurs for a couple
of cycles at lower potential, because the dissolution of metallic copper, which should occur at
~3.5 V, is hindered by the lower working potential of the electrode material. However, for the
CuF2·2H2O, Cu2(OH)F3 and Cu(OH)F materials the capacities along the cycles still decrease
quickly, although the cyclability is slightly better than that of anhydrous CuF2.
Although, the CuF2 samples studied in this work show rather poor reversibility, the capacities
exhibited during the first discharge, reaching up to 600 mAh g-1, make them interesting
cathode materials for primary battery applications.
3 Conclusion
Nanoscale CuF2 can be successfully prepared from copper(II) alkoxides and HF, using either
solvothermal synthesis in THF or liquid anhydrous HF under strict exclusion of moisture.
Depending on the reaction conditions, nanoparticles of a size between 10 and 100 nm are
formed. We show here the possibility to implement a new synthesis approach allowing the
nanosizing of CuF2. This is a far more suitable way for any applications in electrochemistry
than the ball milling of micron-sized particles, which is commonly performed in order to
study intrinsically low conductive compounds.
CuF2 is a very hygroscopic compound, readily forming hydrates and hydroxide fluorides upon
contact with moisture. Out of these, only Cu(OH)F is stable under air. Novel synthesis routes
were developed to obtain the phase pure compounds CuF2·2H2O and Cu2(OH)F3. Both of them
are unstable under air and undergo hydrolysis.
Based on the comprehensive synthesis routes developed in this work (cf. Figure 5), we have
investigated the electrochemical properties of copper(II) fluoride and derivatives in an
attempt to evaluate their applicability in Li-Ion batteries and to assess a clear structure-
property correlation. All in all, it can be stated that although the theoretical capacity and
potential of copper(II) fluoride make it a potential candidate, its high potential vs. Li/Li+ leads
to the dissolution of metallic copper during the charge and to irreversible operation. On the
other hand, hydroxyfluorides and hydrated copper fluorides show a partial reversibility, but
also, as expected, a lower working potential making them not as interesting as pure metal
fluorides.
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 17 -
4 Experimental Part
4.1 Reactants
Methanol was obtained from Sigma Aldrich Co. (99.6%) and refluxed over Mg for 3 days.
Other alcohols (ethanol, n-propanol, n-butanol, t-butanol) were refluxed over Na for 3 days.
Tert-butyl acetate was dried by stirring over freshly dehydrated molecular sieve 4 Å for 3
days. Tetrahydrofuran was refluxed over sodium wire for 3 days.
Copper(II) chloride CuCl2 was obtained from ABCR, lithium methoxide was obtained from
Strem. These compounds were handled and stored under argon in a glove box.
Solutions of anhydrous HF in methanol, ethanol and tetrahydrofuran were prepared by
bubbling a gaseous HF/argon mixture through the solvent in a FEP bottle under cooling. The
exact procedure is described elsewhere.[25] Caution: HF is a hazardous agent and has to be
used under restricted conditions only!
4.2 Synthesis of copper precursors
As long as not stated otherwise, all work was carried out under argon using standard Schlenk
technique and the products are stored under argon atmosphere. Caution: HF is a hazardous
agent and has to be used under restricted conditions only!
Abbreviations: PFA = perfluoroalkoxy polymer, THF = tetrahydrofuran, Me = CH3, tBu = tert-
C4H9.
Cu(OMe)2.[26] 10.86 g (80.8 mmol) brown CuCl2 was dissolved in 100 ml methanol resulting
in a clear green solution. 6.14 g LiOMe (161.6 mmol) was dissolved in 80 ml methanol. The Li
solution was transferred into the Cu solution under vigorous stirring; the vessel was purged
with 20 ml methanol. A mixture of blue and green precipitates is obtained, which are
Cu(OMe)2 and Cu(OMe)Cl, respectively. The reaction mixture is refluxed for 2-3 hours. The
green precipitate now vanished. After cooling and sedimentation overnight, the blue
precipitate is filtered and washed repeatedly with methanol (ca. 7-10 times with portions of
60 ml), until the filtrate contains no more chloride (tested with AgNO3). The blue residue was
dried in vacuum at 40°C. Yield: 9.70 g Cu(OMe)2 (77.2 mmol, =95.5%).
Cu(OtBu)2.[26] 5.05 g (40.2 mmol) blue Cu(OMe)2 was dispersed in 80 ml tert-butyl acetate
CH3COO-tBu and refluxed for 1 h. The solid color turned from deep blue to slightly greenish
1
blue. After this, the reaction mixture is evaporated in vacuum to /4 of its original volume to
remove formed methyl acetate. The procedure was repeated after adding another 50 ml of
tert-butyl acetate. After final cooling, remaining solvent was filtered off, and the residue was
dried in vacuum at 40°C. Yield: 4.22 g Cu(OtBu)2 (20.1 mmol, =50%).
CuF2·2H2O. 15,12 g CuCO3·Cu(OH)2 (68.4 mmol) was suspended in 30 ml ethanol in an open
PFA flask. 15.1 ml of 27.16 M ethanolic HF (410.4 mmol, i.e. Cu:HF = 1:3) was added and
stirred for 1 hour, until no CO2 evolves anymore. The light blue precipitate is filtered and
dried at 50°C in vacuum. Yield: 17.55 g (172.6 mmol, =93.3%). Elemental analysis calcd (%):
Cu 46.2, F 27.6; found batch 1/2: 46.3/46.8, 24.4/27.2. XRD CuF2·2H2O (PDF 6-143).
Cu2(OH)F3. 1.00 g CuF2·2H2O (7.27 mmol) was packed three times crosswise into aluminum
foil to ensure a self-generated atmosphere during decomposition. This pack was annealed in
a muffle type furnace for 3 h at 160°C at the open air. The still hot pack is transferred into an
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 18 -
argon filled Schlenk tube. After cooling under argon, the tube was transferred into an argon
filled glove box, and the foil pack was opened. A pale blue powder is obtained. Yield: 0.691 g
(3.44 mmol, =94.6%). Elemental analysis calcd (%): Cu 63.2, F 28.1; found 60.6, 28.3. XRD
Cu2(OH)F3 (PDF 6-170).
Cu(OH)F.[16] 5.00 g CuF2·2H2O (36.3 mmol) was suspended in 100 ml water in an open PFA
flask. The flask was heated in an oil bath at the open air to boiling for 5 h. Formed HF was
allowed to evaporate. Evaporated water was successively replaced. Meanwhile a light blue
precipitate was formed. After cooling, the precipitate was filtered off, followed by drying at
80°C. Yield: 2.33 g (23.4 mmol, =64.5%). The product is stable at air, but decomposes into
CuO in contact with cold water within one day. Analysis (theoretical values in brackets):
Elemental analysis calcd (%): Cu 63.8, 19.1; found: 63.8, 19.1. XRD Cu(OH)F (PDF 7-306).
4.3 Synthesis of nano-CuF2
All reactions were carried out under inert argon atmosphere. Products were stored under
argon. Caution: HF is a hazardous agent and has to be used under restricted conditions only!
CuF2 – solvothermal reaction. General procedure: In an autoclave with Teflon inset (total
volume 40 ml) 2.00 g Cu(OMe)2 or 3.34 g Cu(OtBu)2 (15.92 mmol) were filled in. 10.0 ml of
17.59 M HF solution in THF (176 mmol, i.e. Cu:F ≈ 1:11) were ädded. The vessel wäs closed,
and then annealed for 2.5 h at 150°C. The pressure in the autoclave raises to approximately 8
bar. After cooling, the autoclave was opened, and the content was transferred into a PFA flask.
The reaction mixture consisted of a white precipitate in a light brown supernatant liquid. The
brown color was less distinct for the Cu(OtBu)2 precursor. By cooling the flask to -20°C, more
precipitate is formed. The precipitate was then filtered off. The residue was washed with
small amounts of pre-cooled (-20°C) THF and dried in vacuum at room temperature.
Depending on the number of washing cycles, a white to slightly brownish product was
obtained. The yields vary from 0.5 to 0.7 g. The product still contains absorbed organic
residue (ca. 6-8%), i.e. the total yield of CuF2 is between 28 and 40%. Subsequent washing
with THF reduces the yield. Obviously, CuF2 is slightly soluble in THF.
CuF2 – liquid HF. 2.00 g Cu(OMe)2 or 3.34 g Cu(OtBu)2 (15.92 mmol) were placed in a 250 ml
three necked PFA flask, equipped with a magnetic stirrer. The flask was cooled to -70°C and
hold at this temperature. A gas stream of HF/argon (ratio ca. 1:1) was introduced, until ca. 10
ml of liquid HF were formed. Meanwhile the solid changed color from blue to white, but did
not dissolve. The reaction mixture was stirred for 10 min at -70°C. Then the cooling was
removed and the mixture was stirred under a light argon flow, until HF was evaporated.
Residual HF was removed in vacuum. The yield is usually around 1.4 g. The product still
contains absorbed organic residue (ca. 10-12%), i.e. the total yield of CuF2 is around 75-80%.
For detailed analysis data cf. Table 3.
4.4 Characterization
XRD. X-ray diffraction were measured on a Seiffert 3003TT diffractometer (Bragg-Brentano
geometry) using Cu Kα radiation (if not stated otherwise). Samples were prepared under
argon in the glove box. For the measurement they were protected with amorphous
polystyrene foil and sealed with Kel-F grease. The average crystallite size L was calculated
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 19 -
from Scherrer’s equätion 퐿 ≈ 휆/(훽푘ℎ푙 cos휃). The measured linewidth 훽푚,ℎ푘푙(2휃) was
corrected with the intrinsic linewidth 훽 =0.12° of the equipment: 훽 = 훽푚2,ℎ푘푙 −훽02.
√
0
ℎ푘푙
FTIR. Spectra were obtained using a Varian 3000 spectrometer in ATR mode (golden gate,
diamond crystal). The design of the ATR cell allowed handling and measuring of substances in
an argon atmosphere under exclusion of moisture.
TEM. Transmission electron microscopy images were recorded on a Philips CM 200
microscope equipped with a LaB6 cathode and operated at 200 kV.
Elemental analysis. Cu content was determined by dissolving the compound in diluted
hydrochloric acid and titration with 0.01 M Na2S2O3 using amylose as indicator. F content was
determined by soda-potash fusion followed by measuring the total fluoride concentration
with a fluoride sensitive electrode. CHNS analysis was done using a standard EuroVector
EuroEA 3000 analyzer.
4.5 Electrochemical testing
All the subsequent operations were realized in an Ar-filled glove box.
Materials preparation. The active materials were beforehand mixed together with
appropriate amounts (see below) of Carbon Black (Timcal, super C) by the mean of a
planetary mill PULVERISETTE 7) for 8 h at 750 rpm using zirconia beakers and beads under
argon atmosphere. 0.5 g of the mixed powder was used for each ball milling, the ratio of this
powder to the zirconia beads was 1:14.
Electrodes preparation. The collected milled powders were blended with poly(vinyliden-
difluoride) (PVdF, Alfa Aesar) in N-Methylpyrrolidone (NMP, Sigma Aldrich , anhydrous
99.5%). The final mass ratio between the active materials, carbon black and PVdF was
65:25:10. The homogeneous slurries were spread over aluminum foils with a razor blade for
a thickness of 20 µm and left more than two days to allow the NMP to be evaporated. Then
disks of 18 mm of diameter were cut to be used as electrodes. The electrochemical
performances were evaluated via half-cell tests using 1 M LiPF6 (ABCR, 99.9% battery grade)
in a non-aqueous mixture of ethylene carbonate (EC, ABCR, 99%), propylene carbonate (PC,
ABCR, 99%) and dimethyl carbonate (DMC, ABCR, 99%) with a volume ratio of 1:1:1 as an
electrolyte, and glass fiber separator (Whatman) against metallic lithium disks (Sigma) on a
Bio-Logic VMP3 potentiostat/galvanostat. The calculations of the specific capacities are done
taking in account only the active materials, including the organic residue, but excluding the
carbon black added during the ink formulation and ball milling process. The uncertainties
related to the calculations of the capacities given in the manuscript based on at least two
batteries run in parallel were always below 4%.
5 Acknowledgements
Dr. F. Emmerling (BAM) is gratefully acknowledged for measuring X-ray diffraction data of
Cu(OR)F. Dr. M. Feist (HU Berlin) is gratefully acknowledged for measuring thermal analysis
of CuF2·2H2O. Florian Schütz is gratefully acknowledged for assistance in working with liquid
HF. This work was partially funded by the Deutsche Forschungsgemeinschaft DFG KE
489/21-1.
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 20 -
6 References
[1]
[2]
[3]
a) J. Cabana, L. Monconduit, D. Larcher, M. R. Palacín, Adv. Mater. 2010, 22, E170-
E192; b) N. Nitta, F. Wu, J. T. Lee, G. Yushin, Mater. Today 2015, 18, 252-264.
J. K. Seo, H.-M. Cho, K. Takahara, K. W. Chapman, O. J. Borkiewicz, M. Sina, Y. S. Meng,
Nano Research 2017.
a) F. Wang, R. Robert, N. A. Chernova, N. Pereira, F. Omenya, F. Badway, X. Hua, M.
Ruotolo, R. Zhang, L. Wu, V. Volkov, D. Su, B. Key, M. S. Whittingham, C. P. Grey, G. G.
Amatucci, Y. Zhu, J. Graetz, J. Am. Chem. Soc. 2011, 133, 18828-18836; b) X. Hua, R.
Robert, L.-S. Du, K. M. Wiaderek, M. Leskes, K. W. Chapman, P. J. Chupas, C. P. Grey, J.
Phys. Chem. C 2014, 118, 15169-15184.
[4]
[5]
a) L. D. Carlo, D. E. Conte, E. Kemnitz, N. Pinna, Chem. Commun. 2014, 50, 460-462; b)
F. Badway, A. N. Mansour, N. Pereira, J. F. Al-Sharab, F. Cosandey, I. Plitz, G. G.
Amatucci, Chem. Mater. 2007, 19, 4129-4141.
a) J. Chun, C. Jo, S. Sahgong, M. G. Kim, E. Lim, D. H. Kim, J. Hwang, E. Kang, K. A. Ryu, Y.
S. Jung, Y. Kim, J. Lee, ACS App. Mater. Interfaces 2016, 8, 35180-35190; b) X. Wang, W.
Gu, J. T. Lee, N. Nitta, J. Benson, A. Magasinski, M. W. Schauer, G. Yushin, Small 2015,
11, 5164-5173.
[6]
[7]
C. M. Wheeler, H. M. Haendler, J. Am. Chem. Soc. 1954, 76, 263-264.
a) C. Billy, H. M. Haendler, J. Am. Chem. Soc. 1957, 79, 1049-1051; b) P. Fischer, W.
Hälg, D. Schwarzenbach, H. Gamsjäger, J. Phys. Chem. Solids 1974, 35, 1683-1689; c) J.
C. Taylor, P. W. Wilson, J. Less Common Met. 1974, 34, 257-259.
[8]
[9]
a) F. Ebert, H. Woitinek, Z. Anorg. Allg. Chem. 1933, 210, 269-272; b) O. Ruff, M. Giese,
Z. Anorg. Allg. Chem. 1934, 219, 143-148.
H. von Wartenberg, Z. Anorg. Allg. Chem. 1939, 241, 381-394.
[10] a) R. Juza, H. Hahn, Z. Anorg. Allg. Chem. 1939, 241, 172-178; b) W. J. de Haas, B. H.
Schultz, J. Koolhaas, Physica 1940, 7, 57-69; c) H. M. Haendler, F. A. Johnson, D. S.
Crocket, J. Am. Chem. Soc. 1958, 80, 2662-2664; d) K. C. Patil, E. A. Secco, Can. J. Chem.
1972, 50, 1529-1530; e) S. I. Troyanov, I. Morozov, Y. M. Korenev, Russ. J. Inorg. Chem.
1993, 38, 909-989.
[11] T. Y. Glazunova, A. I. Boltalin, P. P. Fedorov, Russ. J. Inorg. Chem. 2006, 51, 983-987.
[12] M. Mosiadz, K. L. Juda, S. C. Hopkins, J. Soloducho, B. A. Glowacki, J. Fluorine Chem.
2012, 135, 59-67.
[13] E. Kemnitz, J. Noack, Dalton Trans. 2015, 44, 19411-19431.
[14] a) D. D. Wagman, W. H. Evans, V. B. Parker, R. H. Schumm, I. Halow, S. M. Bailey, K. L.
Churney, R. L. Nuttall, J. Phys. Chem. Ref. Data 1982, 11, Suppl. 2; b) M. Binnewies, E.
Milke, Thermochemical Data of Elements and Compounds, Wiley-VCH, Germany,
Weinheim, 1999.
[15] P. Ramamurthy, E. A. Secco, Can. J. Chem. 1969, 47, 2185-2190.
[16] J. J. Berzelius, Annalen der Physik 1824, 77, 169-230.
[17] L. Li, R. Jacobs, P. Gao, L. Gan, F. Wang, D. Morgan, S. Jin, J. Am. Chem. Soc. 2016, 138,
2838-2848.
[18] Q. Guan, J. Cheng, X. Li, W. Ni, B. Wang, Chin. J. Chem . 2017, 35, 48-54.
[19] S. Geller, W. L. Bond, J. Chem. Phys. 1958, 29, 925-930.
[20] G. Giester, E. Libowitzky, Z. Kristallogr. 2003, 218, 351.
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 21 -
[21] H. R. Oswald, A. Reller, H. W. Schmalle, E. Dubler, Acta Crystallographica Section C
1990, 46, 2279-2284.
[22] D. E. Conte, L. Di Carlo, M. T. Sougrati, B. Fraisse, L. Stievano, N. Pinna, J. Phys. Chem. C
2016, 120, 23933-23943.
[23] Y. T. Teng, F. Wei, R. Yazami, J. Alloys Compd. 2015, 653, 434-443.
[24] Y. T. Teng, S. S. Pramana, J. Ding, T. Wu, R. Yazami, Electrochim. Acta 2013, 107, 301-
312.
[25] T. Krahl, D. Broßke, K. Scheurell, B. Lintner, E. Kemnitz, J. Mater. Chem. C 2016, 4,
1454-1466.
[26] J. V. Singh, B. P. Baranwal, R. C. Mehrotra, Z. Anorg. Allg. Chem. 1981, 477, 235-240.
This article is protected by copyright. All rights reserved.

10.1002/chem.201800207
Chemistry - A European Journal
- 22 -
TOC Graphics
HF
solvothermal
Cu(OR)2
nano-CuF2
4
3
2
1
0
-1
1
2
3
4
5
Voltage vs. Li/Li+ (V)
TOC Text
Nanoscale CuF2 is directly prepared by reaction of anhydrous HF with Cu(OR)2 without a
further milling process.
It has a high electrochemical potential of 2.7 V in lithium ion batteries. Smaller particles show
higher capacities than larger particles, but the capacity is lost after recharge. Hydroxylated
CuF2 has a lower potential but a higher capacity and a slightly better cyclability.
This article is protected by copyright. All rights reserved.