Reactions of a Polyhalide Ionic Liquid with Copper, Silver, and Gold

Article abstract of DOI:10.1002/open.201800149

The reactions of copper, silver, and gold with the imidazolium-based polyhalide ionic liquid (IL) [C₆C₁Im][Br₂I] were investigated by using X-ray photoelectron spectroscopy (XPS), weight-loss measurements, and gas-phase mass spectrometry. All three Group 11 metals are strongly corroded by the IL at moderate temperatures to give a very high content of dissolved Cu^I, Ag^I, and Au^I species. The IL–metal solutions are stable against contact with water and air. The replacement of imidazolium with inorganic sodium cations decreased metal corrosion rates by orders of magnitude. Our results clearly indicate metal oxidation by iodide from dibromoiodide anions to form molecular iodine and anionic [Br-M^I-Br]^? (M=Cu, Ag, Au) complexes stabilized by imidazolium counterions. From experiments with a trihalide IL with imidazolium methylated at the 2-position, we ruled out the formation of imidazole–carbene as a cause of the observed corrosion. In contrast to Group 11 metals, molybdenum is inert against the trihalide IL, which is attributed to surface passivation.

Full text of DOI:10.1002/open.201800149

DOI: 10.1002/open.201800149
Reactions of a Polyhalide Ionic Liquid with Copper, Silver,
and Gold
Benjamin May,[a] Matthias Lexow,[a] Nicola Taccardi,[b] Hans-Peter Steinrꢀck,[a] and
Florian Maier*[a]
The reactions of copper, silver, and gold with the imidazolium-
based polyhalide ionic liquid (IL) [C6C1Im][Br2I] were investigat-
ed by using X-ray photoelectron spectroscopy (XPS), weight-
loss measurements, and gas-phase mass spectrometry. All
three Group 11 metals are strongly corroded by the IL at mod-
erate temperatures to give a very high content of dissolved
CuI, AgI, and AuI species. The IL–metal solutions are stable
against contact with water and air. The replacement of imida-
zolium with inorganic sodium cations decreased metal corro-
sion rates by orders of magnitude. Our results clearly indicate
metal oxidation by iodide from dibromoiodide anions to form
molecular iodine and anionic [Br-MI-Br]ꢀ (M=Cu, Ag, Au) com-
plexes stabilized by imidazolium counterions. From experi-
ments with a trihalide IL with imidazolium methylated at the
2-position, we ruled out the formation of imidazole–carbene as
a cause of the observed corrosion. In contrast to Group 11
metals, molybdenum is inert against the trihalide IL, which is
attributed to surface passivation.
1. Introduction
Ionic liquids (ILs) are a quite new class of salts that are often
liquid well below 1008C. They exhibit unique physicochemical
properties, such as extremely low vapor pressures, wide elec-
trochemical windows, and unusual solvation characteristics, to
name only a few, which makes them candidates for many ap-
plications.[1] ILs are mostly comprised of organic cations and
organic or inorganic anions that are covalently bound molecu-
lar entities. More unusual are ILs that are based on polyhalide
anions, such as triiodide [I3]ꢀ, due to the related Iꢀ +I2Ð[I3]ꢀ
equilibrium.[2]
hanced reactivity and higher yields in halogenation reactions.[5]
Moreover, they also exhibit stereo-[6] and regioselectivity,[7] and
the reusability of some of these agents has been demonstrat-
ed.[6–8] In addition, IL-based polyhalides have great potential as
“green” halogenation reagents due to their low volatility,
which makes them easier to handle than the base halogens.
Another application for trihalide ionic liquids is related to the
electrolyte in dye-sensitized solar cells (DSSCs). Classic DSSCs
based on IL ions make use of the [I3]ꢀ/Iꢀ redox couple.[9] Re-
cently, [Br3]ꢀ/Brꢀ-based redox couples have also come into
focus in this context.[10] Polyhalide ionic liquids are of interest
as components in both liquid and quasi-solid electrolytes.[11]
Understanding the behavior of polyhalide species in ionic liq-
uids is, therefore, an important step in optimizing redox
couple/electrolyte chemistry. Moreover, an understanding of
the corrosive nature of polyhalide-containing electrolytes
when in contact with metal electrodes—often made from pre-
cious metals and/or carbonaceous electrodes decorated with
metal electrocatalysts—is crucial for the long-term stability
issues of DSSC devices: it is known that iodine is corrosive to
metallic grid contacts, such as silver, when water and air are
present.[12] Furthermore, wet-chemical metal dissolution plays
an important role when etching gold microstructures in micro-
electronic devices by using aqueous iodide/iodine trihalide sys-
tems, or when employing cyanide-based chemistry for gold
leaching, to name only two examples. The latter has the disad-
vantage of high toxicity and disposal issues, the former gener-
ates volatile iodine formed during the etching process (a
recent review is provided by Green[13]).
Polyhalide anions have a long history in aqueous systems,
but have only recently been an active area of interest in ILs.
New synthesis techniques have allowed for an expansion of
the number of known polyhalide anions (particularly those of
the lighter halides), and several interesting potential applica-
tions for polyhalide ILs have emerged:[3] Tribromide salts have
been used in organic synthesis as convenient bromination
agents.[4] Tribromide- and trichloride-containing ILs exhibit en-
[a] B. May, M. Lexow, Prof. Dr. H.-P. Steinrꢀck, Dr. F. Maier
Lehrstuhl fꢀr Physikalische Chemie II
Friedrich-Alexander-Universitꢁt Erlangen-Nꢀrnberg
Egerlandstr. 3, 91058 Erlangen (Germany)
E-mail: florian.maier@fau.de
[b] Dr. N. Taccardi
Lehrstuhl fꢀr Chemische Reaktionstechnik
Friedrich-Alexander-Universitꢁt Erlangen-Nꢀrnberg
Egerlandstr. 3, 91058 Erlangen (Germany)
Supporting Information and the ORCID identification number(s) for the
author(s) of this article can be found under:
https://doi.org/10.1002/open.201800149.
One of the key properties of ILs is their low volatility, which
allows for investigations under ultrahigh vacuum (UHV) condi-
tions, such as X-ray photoelectron spectroscopy (XPS). Over
the last decade, XPS studies on a large number of nontrihalide
ꢁ 2018 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
This is an open access article under the terms of the Creative Commons
Attribution License, which permits use, distribution and reproduction in
any medium, provided the original work is properly cited.
ChemistryOpen 2018, 00, 0 – 0
1
ꢁ 2018 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

IL systems have been successfully carried out, proving the
strength of this surface-sensitive technique (for reviews, see
Steinrꢀck[14] and Maier et al.[15]). XPS has also recently been ap-
plied to trihalide ionic liquid systems by Men et al.[16] By using
a conventional laboratory XPS setup and imidazolium-based
[Br2I]ꢀ and [I3]ꢀ ILs, the authors showed that the vapor pressure
was low enough to perform XPS and that beam damage ef-
fects were negligible when X-ray exposure was restricted. By
analyzing the shifts in binding energy and shake-up loss fea-
tures, Men et al. clearly showed that [Br2I]ꢀ and [I3]ꢀ are weakly
coordinating anions with very low basicity. The negative excess
charge on the anions is delocalized over the trihalide com-
plexes to form a p system. For [Br2I]ꢀ, the authors found the
anionic central iodine atom sitting in a linear [Br-I-Br]ꢀ complex
to be less electronegative than its bromine neighbors.[16] In
their XPS studies, Men and co-workers placed small IL droplets
onto a stainless-steel sample holder. Initial XPS studies by our
group to investigate similar trihalide ILs in the form of thin
films spread on polycrystalline gold foils (our standard inert
support) were stymied by the fact that the brownish IL films
first quickly darkened and then lost color under UHV. More-
over, these IL samples on gold showed an increased back-
ground pressure compared with conventional ILs, which gave
a first indication of a corrosion reaction between the IL and
gold to give volatile byproducts.
is the first systematic investigation into the etching of
Group 11 metals by a trihalide ionic liquid.
2. Results
2.1. Bulk Metal Corrosion
When the copper, silver, and gold foils were immersed in the
[C6C1Im][Br2I] IL, it was noted that the initially dark red-brown
IL turned black within minutes and became more viscous over
time. This darkening process occurred much faster with copper
foil than with gold and silver foils. With molybdenum foil, no
color change was noted.
All the Group 11 metal foils (gold, silver, and copper)
showed considerable mass loss after exposure to [C6C1Im][Br2I]
for 6 h at 408C (Table 1). Converting the metal mass loss to
metalmmol, copper was dissolved to the largest extent
(3.22 mmol Cu per 4.1 mmol of initial IL, that is, nominally a
0.78:1 molar ratio) compared with gold (0.30:1) and silver
(0.43:1). For gold, extended exposure to the IL was tested (Au-
21h, Table 1), with the second gold foil initially exposed to
[C6C1Im][Br2I] at 408C for about 11 h, then left in the IL without
heating for a further 10 h. The slightly increased mass loss of
gold after extended exposure showed that maximum gold dis-
solution had not been reached after 6 h; however, the 6 h ex-
posure will be discussed here in most cases for consistency. In
contrast to the Group 11 metals, molybdenum showed virtually
no mass loss. Visual inspection of the metals before and after
corrosion (Figure 1) revealed clear changes in the copper,
silver, and gold foils: instead of a shiny smooth surface, tarnish-
ed patches were seen on copper and silver. The apparent
greenish patina on the copper is consistent with the formation
of copper halide salts. Gold did not show any color changes,
but was clearly etched; the elongated grain structure typical of
a rolled polycrystalline foil became apparent after immersion
in the trihalide IL. To obtain a better microscopic view, a
0.1 mm thick gold foil and a 1 mm thick gold wire were partly
immersed in [C6C1Im][Br2I] for about 17 h at room temperature,
then scanning electron microscopy (SEM) pictures were record-
ed after the samples were rinsed with acetone (see Figure S1
in the Supporting Information). Whereas the unexposed part
of the foil revealed a smooth and flat appearance, the im-
To elucidate the effect of trihalide ILs on noble metals, we
carried out a systematic investigation of Group 11 metals (Au,
Ag, and Cu). These three metals from the same group have
been chosen because they are relevant electrode-contact ma-
terials in DSSCs and in electronic devices in general. They were
exposed to two imidazolium-based ILs with the trihalide anion
dibromoiodide [Br2I]ꢀ, namely 1-hexyl-3-methylimidazolium di-
bromoiodide ([C6C1Im][Br2I]) and 1-butyl-2,3-dimethylimidazoli-
um dibromoiodide ([C4C1C1Im][Br2I]). The latter was chosen to
elucidate the role of potential metal carbene formation. We
demonstrate that our corrosion procedure is a simple way to
synthesize IL systems with a high content of gold(I), silver(I),
and copper(I) species, which are stable against exposure to
ambient conditions and water. In contrast to the extensive
studies of transition metals, aluminum, and steel compounds
corroded by ionic liquids,[17] to the best of our knowledge this
Table 1. Change in mass (and in amount of substance) of the metal foils after immersion in [C6C1Im][Br2I] at 408C for 6 h. In case of Au-21 h, the Au foil
was kept in the IL at 408C for 11 h, then left for a further 10 h at room temperature. Mass losses are also given in dissolution / etching rates commonly
used in literature. Moreover, metal contents in the remaining IL solutions as derived from mass loss after corrosion are given in molar ratio (mol metal loss
: mol initial IL) and in molar content.
Initial
[mg]
Final
[mg]
D
D
Dissolution rate
Etch rate
[mmminꢀ1
IL
[mg]
IL
Metal/IL molar
ratio
Metal
[mol%]
[mg]
[mmol]
[gdmꢀ2 dꢀ1
]
]
[mmol]
Au-6h 480
245
198
196
63
235
330
188
204
1
1.19
1.68
1.74
3.22
ꢁ0
18.9
6.89
12.3
13.7
ꢁ0
0.068
0.025
0.082
0.106
ꢁ0
1783 3.93
1935 4.26
1854 4.08
1874 4.13
1998 4.40
0.30:1
0.39:1
0.43:1
0.78:1
0.00:1
23
28
30
44
0
Au-21h 528
Ag-6h 384
Cu-6h 267
Mo-6h 329
328
&
ChemistryOpen 2018, 00, 0 – 0
www.chemistryopen.org
2
ꢁ 2018 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

magnitude as typical wet-chemical etching rates of 0.1 to
1 mmminꢀ1 in aqueous I2/Iꢀ or alkaline cyanide etching pro-
cesses.[13]
Due to the high metal contents in the IL solutions obtained
after dissolution, metal core-level signals were expected to be
clearly detectable XPS. After the 6 h corrosion experiments, the
solutions were placed onto a molybdenum XPS sample holder
and transferred to the XPS system. The chemical nature of the
dissolved metal species, as deduced by using XPS, will be dis-
cussed in detail in Section 2.2. The concentration values de-
rived from XPS intensities were directly compared with the
mass-loss values shown in Table 1; the quantification of metal
content is discussed below. For quantification, the measured
XPS intensities were converted into number of atoms (Table 2),
taking the corresponding atomic sensitivity factors (ASF) into
account. The signals of all atoms were normalized to the nitro-
gen signal that originated solely from the imidazolium cation
(that is, two nitrogen atoms per cation). For the Group 11
metals, metal-to-imidazolium ratios of between 0.4 and 0.5 to
one were found in the XPS spectra, whereas no molybdenum
signals were found for the Mo sample, which confirmed the
negligible corrosion of this metal. In Table 2, the reduced con-
tent of trihalide iodide is notable, particularly in the copper
corrosion experiment. Apparently, the corrosion of Group 11
metals goes along with a loss of iodide. Note that apart from
the IL and metal signals, minor amounts of oxygen and an
excess of carbon were also detected; these findings are dis-
cussed in Section 2.2.
Figure 1. Photos of metal foils after immersion in [C6C1Im][Br2I] for 6 h and
rinsing with acetone: a) copper, b) silver, c) gold, d) molybdenum. For direct
comparison, photos of a gold foil (0.1 mm thick, ꢁ4ꢂ18 mm) were taken
e) before and f) after corrosion under grazing illumination conditions.
mersed part displayed a rough surface with etched grooves
and pits. Moreover, the exposed Au wire clearly revealed a 6%
decrease in diameter, and the formerly smooth surface became
rough and decorated with etching pits on the mm scale (Fig-
ure S1). In contrast to the Group 11 metals, molybdenum ex-
hibited no visual changes in the surface; the negligible mass
loss and the unchanged surface appearance both indicate that
no corrosion of the molybdenum foil took place.
The mass losses of the foils after immersion for 6 h in the
polyhalide IL, and the quantification of the XPS signals of the
remaining solutions, consistently revealed that, for gold, silver,
and copper, a considerable amount of metal was dissolved
into the IL, whereas molybdenum was virtually unaffected. The
XPS spectra of the neat trihalide IL [C6C1Im][Br2I] and the solu-
tions after the corrosion experiments will be discussed below
in detail.
From the mass losses summarized in Table 1, it is clear that
the three Group 11 metals are susceptible to considerable bulk
corrosion by [C6C1Im][Br2I], which leads to an IL solution with a
metal content well above 20 mol%. For comparison with
common dissolution rates, such as metal mass loss per ex-
posed surface area per unit time [mgdmꢀ2 dꢀ1] and changes in
foil height per unit time [mminꢀ1], the corresponding values
are also given in Table 1. Note in the case of the most noble
metal, gold, the etching rate for 6 h is of the same order of
2.2. XPS Spectra of Neat [C6C1Im][Br2I] Before and After Cu,
Ag, Au, and Mo Dissolution
As described above, [C6C1Im][Br2I] quickly changed color from
dark brown to black on contact with the Group 11 foils in the
corrosion experiments. The black solutions were spread onto a
flat XPS molybdenum sample holder and placed in the fast
entry load lock of a UHV system overnight (for details, see the
Table 2. The composition of [C6C1Im][Br2I] before and after the bulk corrosion experiments (6 h at 408C; see Table 1). The nominal composition of
[C6C1Im][Br2I] is also given. Metal/Im denotes the ratio of metal cations to imidazolium cations.
C1s
0.208
O1s
N1s
0.364
I3p5/2
5.740
1
0.5
<0.1
0.3
0.5
Br3d
Cu2p1/2,3/2
Ag3d3/2,5/2
Au4f5/2,7/2
metal/Im
ASF
0.599
0
0.551
2
8.69
0
–
0.5
–
–
4.460
0
–
–
0.4
–
–
3.756
0
–
–
–
0.4
–
nominal
neat [Br2I]ꢀ
Cu-6h
Ag-6h
Au-6h
10
2
ꢂ2
ꢂ2
ꢂ2
ꢂ2
ꢂ2
11.7
14.4
16.5
15.2
17.3
0.2
1.2
1.3
0.6
1.4
1.7
2.2
1.9
2.1
2.0
0.5
0.4
0.4
0.0
Mo-6h
0.7
–
ChemistryOpen 2018, 00, 0 – 0
www.chemistryopen.org
3
ꢁ 2018 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Experimental Section). During degassing under vacuum, the
films lost considerable color to give a more-or-less transparent
liquid layer on the surface of the sample holder the following
day. This change in color from dark black to a fully clear film
was most pronounced for the solution from the copper corro-
sion experiment, in which the most metal was dissolved due
to the high etching rate (see Section 2.1.). Survey XPS spectra
and IL- and metal-related detailed core-level scans are shown
in Figures 2 and 3 for the neat degassed [C6C1Im][Br2I] trihalide
IL and for the IL solutions after corrosion of the gold, silver,
copper, and molybdenum foils for 6 h. The Calkyl peak at
284.8 eV served as an internal reference for the binding ener-
gies, as previously described.[18]
Figure 2 shows the survey spectra and the detailed IL-related
core-level spectra. In the C1s region, two deconvoluted peaks
originate from the five terminal carbon atoms of the alkyl
chain at 284.8 eV (Calkyl) and from the five carbon atoms direct-
ly attached to a nitrogen atom at 286.5 eV (Chetero). The inte-
grated areas of the two carbon signals of the neat IL revealed
a Calkyl/Chetero ratio of 1.1:1.0, which matches the expected 5:5
ratio to within the accuracy of our method (ꢃ10%). The
anion-related signals of the neat IL showed a single iodide
Figure 3. Metal core level regions of the IL solutions (red spectra) after cor-
rosion of b) gold (Au/IL), e) silver (Ag/IL), and g) copper (Cu/IL) foils for 6 h
at 408C in [C6C1Im][Br2I]. For comparison, the metal signals of freshly sput-
tered a) Au, d) Ag, and f) Cu foils before corrosion are shown in black along
with the AuIII signal (c, blue trace) of the tetrabromoaurate [C2C1Im][AuIIIBr4]
compound. For details, see the main text.
peak in the I3d5/2 region at 620.7 eV, and in the Br3d3/2,5/2
region at around 68.5 eV (note that, due to the small spin-
orbit-splitting of the Br3d signals, the binding energy values
refer to the peak maxima). Notably, the I3d5/2 and the Br3d sig-
nals of [C6C1Im][Br2I] are considerably shifted to higher binding
energies (by +2.4 and +1.1 eV), compared with the analogous
monohalide ILs [C6C1Im]I and [C6C1Im]Br, respectively (Fig-
ure S2). These shifts to higher binding energy are mainly relat-
ed to the delocalization of negative charge over the trihalide
anion compared with the single halide anions, and have al-
ready been discussed by Licence and co-workers.[16] Based on
the XPS binding energy shifts and Kamlet–Taft solvent parame-
ter measurements, the authors showed that the trihalide [Br2I]ꢀ
anion is an extremely weakly coordinating anion with very low
basicity. In our measurements, this is reflected by the larger
binding energy separation of the Calkyl and Chetero peaks for
[C6C1Im][Br2I] compared with the [C6C1Im]Br and [C6C1Im]I mon-
ohalide ILs (see the C1s spectra in Figure 2 and Figure S2). A
detailed explanation of the significance of this separation is
given in Ref. [19]. In short, a more strongly coordinating anion,
such as bromide, allows for greater partial electron transfer to
the imidazolium ring and leads to a lower positive charge and,
therefore, a shift in the Chetero peak towards lower binding en-
ergies. Thus, the large Chetero–Calkyl peak separation found for
[C6C1Im][Br2I] confirms the low coordination strength of the
[Br2I]ꢀ anion, as already proposed by Men et al.[16]
Quantification of the XPS intensities of neat [C6C1Im][Br2I] led
to an overall good agreement with the nominal composition
(see also Table 2), apart from the iodine content, which was
found to be too low. For the neat IL, an atomic composition of
iodine/bromine/nitrogen of 0.5:1.7:2.0 was found instead of
the expected 1:2:2 (for Mo-foil corrosion, a slightly higher
Figure 2. Wide (top) and detailed XP spectra of the neat [C6C1Im][Br2I] IL
(black) and of the solutions obtained after exposure of molybdenum (ma-
genta), gold (green), silver (blue), and copper (red) foils to [C6C1Im][Br2I] for
6 h at 408C. All spectra are referenced to aliphatic carbon at 284.8 eV.
&
ChemistryOpen 2018, 00, 0 – 0
www.chemistryopen.org
4
ꢁ 2018 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

iodine content in an atomic ratio of 0.7:2.0:2.0 was found). We
attribute this, at least partly, to some uncertainty in the atomic
sensitivity factor used here for the I3d5/2 core-level quantifica-
tion (note that a I/N ratio of 0.85:2.0 was found by XPS quan-
tification for [C6C1Im]I). Note that we derived the iodine intensi-
ty only from the main I3d5/2 line, which includes the broad
shake-up shoulder around 626 eV that contains some intensity
from the trihalide anions due to their delocalized charge distri-
bution.[16] Because the focus of this work is metal corrosion,
more detailed investigations must be carried out in the future
to elucidate this aspect.
Ag3d3/2,5/2 between 388 and 364 eV, and Cu2p1/2,3/2 between
960 and 925 eV. For comparison, the XPS spectra of Au0, Ag0,
and Cu0 in the form of freshly sputtered metal foils are also de-
picted (Figure 3).
Both the Au0 reference and the dissolved Au species show
the typical spin-orbit split Au4f5/2,7/2 doublet. Whereas the
Au4f7/2 level of the Au foil, that is, Au0, is found at 84.1 eV, in
the IL solution after corrosion it is shifted to higher binding
energy by 0.7 eV to 84.8 eV. This binding energy position is
consistent with an AuI species.[21] To rule out the presence of
AuIII in the solution, we also measured the tetrabromoaurate IL
analogue [C2C1Im][AuIIIBr4], which showed a much higher shift
of 2.8 eV to higher binding energy, that is, to 86.9 eV (see
Figure 3). Notably, the widths of the Au4f signals of the gold
foil and of the IL solution are virtually identical, which is an in-
dication that the AuI metal ion in solution is present in a uni-
form chemical environment.
The XPS spectra after the 6 h corrosion experiment with Mo
foil showed no Mo signals (e.g. Mo3d at around 230 eV),
which corroborates the negligible mass loss found for the IL-
exposed Mo foil (see Table 1). The only changes were a moder-
ate decrease in the [C6C1Im][Br2I] signals, along with an in-
crease in the Calkyl signal and the appearance of an O1s peak
at 530 eV. These additional Calkyl and O signals are attributed to
the appearance of nonvolatile contaminants in the near-sur-
face region after the corrosion experiment, which possibly
originate from residual glassware contamination and/or ad-
sorption from lab air. Such contaminants have been reported
to be surface-active[20] and, therefore, show up in XPS spectra
even at extremely low bulk concentrations.
Clean silver foil exhibits a sharp Ag3d doublet, with the
Ag3d5/2 level at 368.6 eV. In the IL solution after corrosion of
the silver foil, a shift in the Ag3d5/2 level to lower binding
energy by 0.7 eV (to 367.9 eV) with no loss features was ob-
served, which is indicative of the formation of AgI.[22] Similarly,
the Cu2p doublet of the solution was also shifted to lower
binding energies by 0.6 eV (Cu2p3/2 at 932.3 eV) with respect
to clean copper foil (Cu2p3/2 at 932.9 eV). The magnitude of
the shift to lower binding energies and the absence of pro-
nounced shake-up structures—which should be present for a
CuII species—are fully in line with the formation of a CuI spe-
cies.[23]
Next, we discuss the IL-derived XPS spectra in Figure 2 after
the 6 h corrosion experiments of gold, silver, and copper foils
at 408C. The most striking difference with molybdenum is the
reduction in I3d intensity (see the I3d5/2 region in Figure 2 and
Table 2). For the Au and Ag solutions, the I3d5/2 intensity de-
creased by 40 and 50%, respectively, compared with the signal
in neat [C6C1Im][Br2I]. For copper, for which the amount of
metal dissolution was found to be the highest (Table 1), the
I3d intensity was almost entirely absent. It is thus evident that
during the reaction of the Group 11 metals with the IL, iodide
from [Br2I]ꢀ is consumed. Because no precipitate was found, a
neutral iodine-containing species is apparently formed that de-
sorbs under vacuum conditions. The Br3d intensity did not de-
crease (and, therefore, no neutral IBr had formed), so it is most
likely that molecular iodine I2 was pumped away. After testing
in a separate UHV system, gas-phase mass spectra indeed
showed the release of I2 from [C6C1Im][Br2I] when the IL was in
contact with copper (see Figure S3). It is important to note
that the Br3d peaks given in Figure 2 did not exhibit signifi-
cant changes in binding energy or peak shape before and
after exposure to Cu, Ag, or Mo; only after Au corrosion were
the Br3d signals slightly shifted by +0.3 eV to higher binding
energy. Thus, the formation of free bromide (Brꢀ) during corro-
sion was ruled out due to the absence of single-halide Brꢀ sig-
nals in the Br3d region at lower binding energies.
2.3. Additional Gold-Foil Corrosion Experiments
As shown above, a considerable amount of metal in the form
of AuI, AgI, and CuI species was dissolved in the trihalide IL at
the expense of iodine. Whereas AgI and CuI species are stable
oxidation states in aqueous systems (i.e. they form stable pre-
cipitates with dissolved halide anions), AuI is metastable and
prone to disproportionation to AuIII and Au0 in water without
stabilizing ligands.[24] Moreover, gold is the most noble—and,
therefore, most unreactive—metal in this series. Thus, further
gold-foil corrosion experiments were carried out to elucidate
1) the effect of additional water, 2) the role of the proton in
the 2-position of the imidazolium ring, and 3) the replacement
of the imidazolium cation by an inorganic cation. The results
of these experiments are summarized below.
When the gold-corrosion experiment was performed under
anhydrous and water-saturated conditions for 6 h at 408C, we
found no significant differences in the amount of dissolved Au,
and also no significant difference in the XPS spectra (see
Table S1 and Figure S4) compared with the results described
above. This indicates that neither water nor oxygen are neces-
sary for (or inhibitory of) the corrosion of gold in contact with
[C6C1Im][Br2I].
Stable AuI-imidazole-carbene complexes are known in the
literature, for example, in the context of AuI homogeneous cat-
alysis.[25] To rule out the possibility that the corrosion process
found here requires or involves the formation of a stable AuI-
Direct evidence for the dissolution of Au, Ag, and Cu was
obtained from the survey spectra in Figure 2 and the detailed
metal spectra in Figure 3. Compared with molybdenum, for
which no metal signals could be detected in the IL, the XPS
spectra of the IL solutions after corrosion experiments with Au,
Ag, and Cu clearly showed the corresponding metal signals.
The detailed spectra in Figure 3 cover the most prominent
metal core levels, that is, Au4f5/2,7/2 between 92 and 79 eV,
ChemistryOpen 2018, 00, 0 – 0
www.chemistryopen.org
5
ꢁ 2018 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

carbene (which most likely would occur by abstracting the
most acidic proton at the 2-position of the imidazolium ring),
ample, in AuI–halide systems.[25]
This reaction mechanism explains the XPS metal signals of
oxidation state +1 in solution, the selective decrease in the
iodine signals in the XPS spectra and the increase in iodine in
the gas phase over the course of the reaction. We attribute the
color change of the trihalide ILs (from deep brown to black)
upon immersion of the Cu, Ag, and Au foils to the formation
of I2 that remains dissolved under ambient pressure, but is lost
under vacuum. The latter process was revealed by gas-phase
analysis and by the decoloration of the IL film on the sample
holder after degassing. According to the proposed reaction
mechanism, the amount of Br remains essentially constant, as
confirmed by using XPS. The fact that the Br3d XPS signals
from the trihalide IL before and after corrosion do not exhibit
new signals at different binding energy—which should be visi-
ble through significant peak shape changes—clearly supports
the proposed idea that the Br atoms remain in an anionic [Br-
MI-Br]ꢀ complex with the negative charge delocalized. The for-
mation of free bromide Brꢀ and free iodide Iꢀ anions can be
ruled out by using XPS because of their very different binding
energy positions.
1
we measured the H and 13C NMR spectra of the IL before and
after Au corrosion. These NMR spectra did not reveal any sig-
nificant changes due to the gold corrosion. As an example, we
1
show the corresponding H and 13C NMR spectra in Figures S5
and S6, respectively. In particular, the signals of the most acidic
2-proton at d=7.8 ppm and of the 2-carbon at d=137 ppm
did not decrease, which would be the case in carbene forma-
tion. Additionally, we tested
a
trihalide homologue,
[C4C1C1Im][Br2I], that was methylated at the 2-position to pre-
vent proton abstraction. Due to the higher melting point of
this IL, we carried out the corresponding corrosion experiment
at 508C, and the foil mass loss results are given in Table S1.
The mass loss of gold after 6 h was about 30% less than with
the nonmethylated IL, which might be related to the higher
viscosity of [C4C1C1Im][Br2I] even at 508C; the molar amount of
gold dissolved per mole of [C4C1C1Im][Br2I], however, was inde-
pendent of the methylation. Overall, the efficiency of Au corro-
sion was found to be very similar for [C6C1Im][Br2I] and
[C4C1C1Im][Br2I].
Finally, we explored whether the trihalide anion was able to
dissolve gold without the presence of an imidazolium counter-
ion. For this purpose, a saturated solution of an aqueous
Na[Br2I] system was prepared by dissolving NaBr (25 mmol)
and IBr (13 mmol) in water (3.2 mL; note that excess Brꢀ was
provided to stabilize AuI against disproportionation into Au0
and AuIII). A gold foil (305 mg) was then immersed into this
highly concentrated solution for 24 h at room temperature.
After rinsing and drying the remaining gold foil, we found a
mass loss of (12ꢃ1) mg, that is, about 0.06 mmol Au dissolved,
which proved that gold is moderately soluble under these cir-
cumstances. Note that about 1 mL of [C6C1Im][Br2I] or
[C4C1C1Im][Br2I] is able to dissolve at least ten times more gold
within 6 h.
For the most noble metal, Au, experiments with both a care-
fully dried IL and in the presence of water, showed no signifi-
cant difference in the corrosion efficiency, which ruled out any
significant influence of water. Moreover, the similar corrosion
rate of gold by [C4C1C1Im][Br2I] compared to the nonmethylat-
ed [C6C1Im][Br2I] IL and the lack of significant spectral changes
in the cation-related XPS peaks revealed that the imidazolium
C2 proton is not involved. This conclusion was confirmed by
the NMR spectra of the [C6C1Im][Br2I] IL after gold corrosion,
which showed no indication of an AuI–carbene complex. Final-
ly, the poor ability of highly concentrated aqueous Na[Br2I] to
corrode metallic gold indicated that the imidazolium cation is
not an innocent counterion, but plays an important role in the
dissolution of the anionic [Br-MI-Br]ꢀ complex. In particular, the
fact that the trihalide ILs oxidized gold to a stable +1 state,
even in the presence of excess water, is worth noting. AuI is
known to be relatively unstable in aqueous solution unless
there is a sufficient concentration of halide ions and a low
enough pH to maintain a stable complex, and will otherwise
disproportionate to Au0 and AuIII.[26] It is tempting to speculate
that there is a stabilizing solvation shell provided by the imida-
zolium cations around the [AuBr2]ꢀ complex, which makes the
imidazolium trihalide IL suitable for the corrosion process.
Here, further experimental evidence, such as crystal-structure
investigations and tests with other IL cations, is needed, but is,
however, out of the scope of this study.
3. Discussion
We have consistently shown by using weight loss and XPS ex-
periments that the trihalide ILs [C6C1Im][Br2I] and
[C4C1C1Im][Br2I] are able to dissolve the Group 11 metals
copper, silver, and gold efficiently by forming +1 metal ions at
the expense of iodine; this is in contrast to the situation for
molybdenum, which is not attacked. Based on these experi-
mental results, the following reaction mechanism of
Group 11 M0 metal with an [Br2I]ꢀ anion is highly plausible:
a
The lack of a mass loss of molybdenum in our corrosion ex-
periments, the lack of Mo signals in XPS spectra, and the lack
of volatile I2 in the gas phase clearly demonstrate that the less-
noble metal molybdenum is virtually inert against attack by
the trihalide IL. Most likely, the surface of this refractory metal
becomes rapidly passivated on contact with the trihalide IL,
which prevents further dissolution. To elucidate the nature of
molybdenum passivation, more experiments are foreseen in
the future.
1
M0 þ ½Br-I-Brꢄꢀ ! ½Br-MI-Brꢄꢀ þ = I2
2
in which the iodide of the trihalide anion oxidizes the metal
atom to oxidation state +1. The reduced iodine radical even-
tually recombines with another iodine radical, which leads to
dissolved iodine I2 that is later removed under vacuum condi-
tions due to its high vapor pressure. We propose that the MI
ion is stabilized by two Brꢀ anions arranged in a linear fashion
to form a [Br-MI-Br]ꢀ anionic complex, which is known, for ex-
&
ChemistryOpen 2018, 00, 0 – 0
www.chemistryopen.org
6
ꢁ 2018 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

treated with fine silicon carbide paper), gold (0.1 mm, MaTecK,
>99.99%), and molybdenum (0.1 mm, MaTecK, 99.95%) were
rinsed with acetone, dried, weighed, and then immersed in
[C6C1Im][Br2I] (ꢁ1 mL) in a glass tube under air. The tubes were
closed and then heated to 408C for 6 h in a water bath. A second
gold foil (619.8 mg) was also corroded in [C4C1C1Im][Br2I] by using
the same procedure, except that the reaction was carried out at
508C due to the higher melting point of the C2-methylated IL. For
each corrosion experiment, the boiling tubes were gently shaken
occasionally. After 6 h, a few drops of the liquid were removed
from the tubes and placed on a molybdenum sample holder to
form a thin layer for XPS analysis. The metal foils were removed
from the tubes, rinsed with acetone, dried, and reweighed. In addi-
tion, a 0.1 mm thick gold foil and a 1 mm thick gold wire were
half-immersed into [C6C1Im][Br2I] at RT for about 17 h. These sam-
ples were also studied by using SEM.
4. Conclusions
We believe that our study could be quite relevant for future
applications. The most obvious ones would be to use the in-
vestigated trihalide ILs as an etching medium for microfabrica-
tion, and even as a leaching medium to dissolve precious
Group 11 metals from metal waste, for example, from electron-
ic devices. One could also envisage recovering the metals by
using appropriate electrochemical processes. Due to their low
vapor pressure, the handling of trihalide ILs promises to be
safer than oxidizing halogens or the use of CN-based water
chemistry. Finally, the formation of stable and water-insensitive
AuI complexes in an IL medium may also open new possibili-
ties for AuI catalysis.
Two further corrosion experiments were performed with “dry”
[C6C1Im][Br2I], and “wet” [C6C1Im][Br2I] to elucidate the role of
water in the gold corrosion process: In the dry IL experiment,
[C6C1Im][Br2I] (2.072 g, 4.56 mmol) was dried by heating to 508C in
a continuously evacuated Schlenk flask for 24 h. A gold foil
(609 mg) was added and the vessel was purged with nitrogen
(Linde, 5.0) before being sealed against air contact. For the wet IL
experiment, the IL (1.933 g, 4.26 mmol) was added to a Schlenk
flask, followed by deionized water (543 mg, 30.1 mmol, 0.5 mL).
The two liquids were not miscible even after several minutes of ag-
itation, and the water formed a layer above the IL. This water layer
initially had a slight red-brown color. A gold foil (526 mg) was then
placed into the bilayered system and fully immersed in the IL
phase. Both Schlenk flasks were held in a water bath at 408C for
6 h. Samples of the ILs were taken for XPS analysis after the reac-
tion, and the masses of the remaining gold foils were measured.
Experimental Section
IL Synthesis
The RT trihalide IL [C6C1Im][Br2I] was prepared by mixing 1-hexyl-3-
methylimidazolium bromide ([C6C1Im]Br, purity 99%, purchased
from IoLiTec and used without further processing, highly viscous
liquid at RT) with iodine monobromide (IBr, Sigma–Aldrich, 98%) in
a 1:1 molar ratio. After mixing both components, we obtained a
deep red-brown liquid with a considerably lower viscosity than
[C6C1Im]Br. We avoided exposure of the IL to daylight to prevent
any photochemical reactions that might occur.
To investigate whether carbene formation played a role in the reac-
tion, we studied
a noncarbene-forming trihalide counterpart,
[C4C1C1Im][Br2I], methylated at the 2-position of the imidazolium
ring. This IL was prepared by mixing dry 1-butyl-2,3-dimethylimida-
zolium bromide (1.314 g, 5.634 mmol; [C4C1C1Im]Br, 99%, IoLiTec,
used without further processing, m.p. ꢁ908C) with IBr (1.175 g,
5.681 mmol), that is, a 1:1 molar ratio. When the two ingredients
were mixed in a mortar, the IL initially liquefied and a deep red-
brown liquid was obtained. The mixture was then placed in an
evacuated bell jar to remove adsorbed water from the hydrophilic
compound, which led to solidification after several minutes under
vacuum. After regrinding, the obtained deep-orange powder ex-
hibited a melting point in the range of 40–508C.
XPS Analysis
Before and after the reactions, XPS analysis was performed for all
IL samples. They were placed on a precleaned molybdenum
sample holder by spreading a thin liquid film (ꢁ0.1 mm thick) at
the bottom of a reservoir (dimensions 14ꢂ20ꢂ0.5 mm) milled into
the sample holder. The sample holder was introduced into a fast-
entry load lock (base pressure 5ꢂ10ꢀ7 mbar) and pumped down
overnight before being transferred to the main UHV system for
XPS analysis (base pressure better than 1ꢂ10ꢀ10 mbar).
1-Ethyl-3-methylimidazolium tetrabromoaurate ([C2C1Im][AuBr4])
was synthesized as an IL-analogous AuIII compound, which served
as an XPS binding energy reference: AuBr3 (454.6 mg, 0.98 mmol;
Sigma–Aldrich, 99.9%) and [C2C1Im]Br (193.7 mg, 1.01 mmol;
Sigma–Aldrich, >97%) were placed in a 50 mL 2-neck flask fitted
with a septum and a reflux condenser. Dry CH3CN (20 mL) was
added through the septum using a syringe, and the initial suspen-
sion was heated at reflux overnight. After cooling, the deep-red so-
lution was syringe-filtered (0.45 mm, cellulose) and transferred to a
50 mL flask. The solvent was evaporated by using a rotary evapora-
tor to give a dark-red solid (yield: 616.8 mg, 95.1 wt% of original
mass). The material was analyzed using inductively coupled plasma
optical emission spectrometry (ICP-OES), which revealed an Au
content of 30.3 wt% (nominal: 31.4 wt%).
In contrast to most IL samples, which were measured at RT, the
spectra with the IL [C4C1C1Im][Br2I] were collected at 45 and 558C.
The higher temperatures were necessary to melt these IL samples
to avoid charging. A type K thermocouple was used to monitor
the temperature of the heated samples, with a thermocouple con-
troller to regulate the voltage of the heating filament. During
measurements, the relative temperature was maintained to within
ꢃ0.18C; absolute sample temperature values were within ꢃ58C.[27]
In the XPS system, the radiation source was a non-monochromated
AlKa anode of a Specs XR-50 dual-anode source, operated at 12 kV
and a 20 mA emission current. All spectra were recorded in normal
emission to integrate over a maximum probing depth (7–9 nm, de-
pending on the kinetic energy[20,28]) by using a Scienta R3000 con-
centric hemispherical analyzer in constant pass energy mode with
200 eV pass energy for survey spectra and 100 eV for detailed
spectra. For the latter, the overall instrumental energy resolution
was 0.9 eV. The collected data were analyzed using CasaXPS and
charge-corrected, such that the aliphatic carbon peak was at
284.8 eV (for more details on the XPS system, see Ref. [18]). Note
Metal Foil Corrosion
To investigate the effect of the trihalide ILs on Group 11 metals,
15ꢂ20 mm foils of polycrystalline copper (0.1 mm thickness,
MaTecK, purity >99.99%), silver (0.125 mm, Chempur, 99.9%, pre-
ChemistryOpen 2018, 00, 0 – 0
www.chemistryopen.org
7
ꢁ 2018 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

[6] C. Chiappe, D. Capraro, V. Conte, D. Pieraccini, Org. Lett. 2001, 3, 1061–
1063.
that for the systems studied here, only minor charge corrections,
of the order of ꢃ0.1 eV, had to be applied.
[7] S. P. Borikar, T. Daniel, V. Paul, Tetrahedron Lett. 2009, 50, 1007–1009.
[8] J. Salazar, R. Dorta, Synlett 2004, 2004, 1318–1320.
[9] a) B. O’Regan, M. Grꢃtzel, Nature 1991, 353, 737–740; b) M. Grꢃtzel, Acc.
Chem. Res. 2009, 42, 1788–1798; c) A. Hagfeldt, G. Boschloo, L. Sun, L.
Kloo, H. Pettersson, Chem. Rev. 2010, 110, 6595–6663.
[10] K. Kakiage, T. Tokutome, S. Iwamoto, T. Kyomen, M. Hanaya, Chem.
Commun. 2013, 49, 179–180.
[11] M. Gorlov, L. Kloo, Dalton Trans. 2008, 2655–2666.
[12] S. Yanagida, Y. Yu, K. Manseki, Acc. Chem. Res. 2009, 42, 1827–1838.
[13] T. A. Green, Gold Bull. 2014, 47, 205–216.
Peaks in the XPS spectra associated with nonmetal elements were
fitted with a pseudo-Voigt profile with a 30% Lorentzian contribu-
tion. Doublets due to spin-orbit splitting (such as I3d) were fitted
with two peaks constrained to have equal full width at half
maxima (FWHM) and the expected peak-area ratio (e.g., 2:3 for
3d3/2/3d5/2). Metal peaks were fitted similarly, except the pseudo-
Voigt profiles had an 80% Lorentzian contribution. Shirley back-
grounds were used. Carbon 1s spectra were fitted with a two-peak
model, Calkyl and Chetero. The Chetero peak was constrained to a FWHM
of 1.1 times the width of the Calkyl peak.
[14] H. P. Steinrꢀck, Phys. Chem. Chem. Phys. 2012, 14, 5010–5029.
[15] F. Maier, I. Niedermaier, H. P. Steinruck, J. Chem. Phys. 2017, 146, 170901.
[16] S. A. Men, K. R. J. Lovelock, P. Licence, Chem. Phys. Lett. 2017, 679, 207–
211.
Acknowledgements
[17] a) B. Dilasari, Y. Jung, J. Sohn, S. Kim, K. Kwon, Int. J. Electrochem. Sci.
2016, 11, 1482–1495; b) Q. Zhang, Y. Hua, Z. Zhou, Int. J. Electrochem.
Sci. 2013, 8, 10239–10249; c) M. Uerdingen, C. Treber, M. Balser, G.
Schmitt, C. Werner, Green Chem. 2005, 7, 321–325.
[18] K. R. J. Lovelock, C. Kolbeck, T. Cremer, N. Paape, P. S. Schulz, P. Was-
serscheid, F. Maier, H. P. Steinrꢀck, J. Phys. Chem. B 2009, 113, 2854–
2864.
[19] T. Cremer, C. Kolbeck, K. R. Lovelock, N. Paape, R. Wolfel, P. S. Schulz, P.
Wasserscheid, H. Weber, J. Thar, B. Kirchner, F. Maier, H. P. Steinrꢀck,
Chemistry 2010, 16, 9018–9033.
[20] C. Kolbeck, M. Killian, F. Maier, N. Paape, P. Wasserscheid, H.-P. Steinrꢀck,
Langmuir 2008, 24, 9500–9507.
B.M., M.L., and H.-P.S. thank the European Research Council (ERC)
under the European Union’s Horizon 2020 research and innova-
tion programme for financial support, in the context of the Ad-
vanced Investigator Grant “ILID” to H.-P.S. (grant agreement no.
693398-ILID). Furthermore, we gratefully acknowledge Prof.
Jꢀrgen Schatz and Dr. Harald Maid of the FAU Organic Chemistry
Institute for the NMR spectroscopy investigations.
[21] H. Kitagawa, N. Kojima, T. Nakajima, J. Chem. Soc. Dalton Trans. 1991,
3121–3125.
[22] a) V. K. Kaushik, J. Electron. Spectrosc. 1991, 56, 273–277; b) S. W. Gaar-
enstroom, N. Winograd, J. Chem. Phys. 1977, 67, 3500–3506.
[23] M. C. Biesinger, L. W. M. Lau, A. R. Gerson, R. S. C. Smart, Appl. Surf. Sci.
2010, 257, 887–898.
Conflict of Interest
The authors declare no conflict of interest.
[24] S. Eustis, M. A. El-Sayed, J. Phys. Chem. B 2006, 110, 14014–14019.
[25] a) A. S. Hashmi, M. Rudolph, Chem. Soc. Rev. 2008, 37, 1766–1775; b) M.
Rudolph, A. S. Hashmi, Chem. Soc. Rev. 2012, 41, 2448–2462.
[26] a) G. H. Kelsall, N. J. Welham, M. A. Diaz, J. Electroanal. Chem. 1993, 361,
13–24; b) C. H. Gammons, Y. Yu, A. E. Williams-Jones, Geochim. Cosmo-
chim. Acta 1997, 61, 1971–1983.
[27] C. Kolbeck, A. Deyko, T. Matsuda, F. T. U. Kohler, P. Wasserscheid, F.
Maier, H.-P. Steinrꢀck, ChemPhysChem 2013, 14, 3726–3730.
[28] R. F. Roberts, D. L. Allara, C. A. Pryde, D. N. E. Buchanan, N. D. Hobbins,
Surf. Interface Anal. 1980, 2, 5–10.
Keywords: corrosion · ionic liquids · metal dissolution ·
photoelectron spectroscopy · surface analysis
[1] N. V. Plechkova, K. R. Seddon, Chem. Soc. Rev. 2008, 37, 123–150.
[2] J. S. McIndoe, D. G. Tuck, Dalton Trans. 2003, 244–248.
[3] a) H. Haller, S. Riedel, Z. Anorg. Allg. Chem. 2014, 640, 1281–1291; b) M.
Wolff, A. Okrut, C. Feldmann, Inorg. Chem. 2011, 50, 11683–11694.
[4] M. Fieser, L. Fieser, Fieser and Fieser’s Reagents for Organic Synthesis
Volume 1, WILEY-VCH Verlag, 1967.
[5] a) V. Kavala, S. Naik, B. K. Patel, J. Org. Chem. 2005, 70, 4267–4271; b) T.
Schlama, K. Gabriel, V. Gouverneur, C. Mioskowski, Angew. Chem. Int. Ed.
1997, 36, 2342–2344; Angew. Chem. 1997, 109, 2440–2442; c) V. M. Ze-
likman, V. S. Tyurin, V. V. Smirnov, N. V. Zyk, Russ. Chem. Bull. Russ. Chem.
B 1998, 47, 1541–1546.
Received: July 18, 2018
Revised manuscript received: September 4, 2018
&
ChemistryOpen 2018, 00, 0 – 0
www.chemistryopen.org
8
ꢁ 2018 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

FULL PAPERS
Soluble gold: The coinage metals Au,
Ag, and Cu are strongly corroded by
the dibromoiodide anions of the ionic
liquid 1-hexyl-3-methylimidazolium di-
bromoiodide ([C6C1Im][Br2I]) to form
iodine and stable metal(I)–bromide
complexes (see figure), as shown by X-
ray photoelectron spectroscopy and
gas-phase mass spectrometry.
B. May, M. Lexow, N. Taccardi,
H.-P. Steinrꢀck, F. Maier*
&& – &&
Reactions of a Polyhalide Ionic Liquid
with Copper, Silver, and Gold
ChemistryOpen 2018, 00, 0 – 0
www.chemistryopen.org
9
ꢁ 2018 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ