Influence of substituents and functional groups on the surface composition of ionic liquids

Article abstract of DOI:10.1002/chem.201304549

We have performed a systematic study addressing the surface behavior of a variety of functionalized and non-functionalized ionic liquids (ILs). From angle-resolved X-ray photoelectron spectroscopy, detailed conclusions on the surface enrichment of the functional groups and the molecular orientation of the cations and anions is derived. The systems include imidazolium-based ILs methylated at the C2 position, a phenyl-functionalized IL, an alkoxysilane-functionalized IL, halo-functionalized ILs, thioether- functionalized ILs, and amine-functionalized ILs. The results are compared with the results for corresponding non-functionalized ILs where available. Generally, enrichment of the functional group at the surface is only observed for systems that have very weak interaction between the functional group and the ionic head groups. Scratching the surface: The surface behavior of various functionalized ionic liquids (ILs) is studied by angle-resolved X-ray photoelectron spectroscopy and conclusions on surface-enrichment effects and molecular orientation of cations and anions are derived (see figure; Im=imidazolium, Tf=trifluoromethanesulfonyl). Overall, surface enrichment is only observed for functional groups that interact weakly with the ionic head groups.

Full text of DOI:10.1002/chem.201304549

DOI: 10.1002/chem.201304549
Full Paper
&
Ionic Liquids
Influence of Substituents and Functional Groups on the Surface
Composition of Ionic Liquids
Claudia Kolbeck,[a] Inga Niedermaier,[a] Alexey Deyko,[a] Kevin R. J. Lovelock,[a]
Nicola Taccardi,[b] Wei Wei,[b] Peter Wasserscheid,[b] Florian Maier,[a] and Hans-
Peter Steinrꢀck*[a]
Abstract: We have performed a systematic study addressing
the surface behavior of a variety of functionalized and non-
functionalized ionic liquids (ILs). From angle-resolved X-ray
photoelectron spectroscopy, detailed conclusions on the sur-
face enrichment of the functional groups and the molecular
orientation of the cations and anions is derived. The systems
include imidazolium-based ILs methylated at the C2 posi-
tion, a phenyl-functionalized IL, an alkoxysilane-functional-
ized IL, halo-functionalized ILs, thioether-functionalized ILs,
and amine-functionalized ILs. The results are compared with
the results for corresponding non-functionalized ILs where
available. Generally, enrichment of the functional group at
the surface is only observed for systems that have very weak
interaction between the functional group and the ionic head
groups.
Introduction
ment at the interface will be different from that in the bulk.
Various surface-sensitive methods under ambient pressure and
under ultra-high vacuum (UHV) conditions have been em-
ployed to study the gas/IL interface (see refs. [6,7] and referen-
ces therein). The majority of these studies mainly focused on
non-functionalized ILs. Consequently, the knowledge about the
influence of functional groups on the surface composition is
still scarce.[8–18] For non-functionalized ILs with long alkyl chains
(e.g., [CnC1Im]+ systems (nꢀ4, Im=imidazolium)), a common
consensus has been reached concerning the molecular ar-
rangement at the surface: the ILs are oriented with the alkyl
chains sticking out towards the gas phase, forming an aliphatic
overlayer above a more polar sublayer, which consists of the
cationic and anionic head groups.[6] Introducing a functional
group at the terminal position of the alkyl chain leads to in-
creased surface tension as was shown by Tariq et al. in a com-
prehensive review of the surface tensions of ILs. They suggest-
ed that the functional group interacts with the polar sublayer
and thereby disrupts the aliphatic overlayer, an interaction that
results in the observed higher surface tension.[19] However, one
should abstain from drawing general conclusions from this
study alone as the nature of the functional group and of the
cationic and anionic head groups strongly influence their intra-
and intermolecular interactions. For instance, a concomitant vi-
brational sum frequency spectroscopy (VSFS) and neutral
impact collision ion scattering spectroscopy (NICISS) study of
ethanolammonium nitrate (EtAN) showed that the hydroxy
group and the ammonium head group lie below the methyl-
ene groups at the outer surface.[10,11] However, when introduc-
ing the hydroxyethyl chain into an imidazolium-based IL,
namely 1-(2-hydroxyethyl)-3-methylimidazolium tetrafluorobo-
rate ([(HOC2H4)C1Im][BF4]), the hydroxy group is found at the
outermost position at the surface, as was shown by molecular
Ionic liquids (ILs) are molten salts with a melting point below
1008C. The enormous variety of cation–anion combinations
(some authors speak of >1018 possibilities[1]) enables tuning of
the physico-chemical properties of ILs over a wide range. Func-
tional groups can be introduced into the chemical structure to
adapt ILs for specific tasks, for example, the capture of CO2
from flue gas by amine-functionalized ILs,[2,3] or the modifica-
tion of silica supports by reaction with silyl-functionalized ILs.[4]
Such functionalized ILs are known as “task-specific ionic liq-
uids” and they are used in a variety of applications, ranging
from catalysis and organic synthesis to extraction and dissolu-
tion processes.[5]
In many applications, the interface of the IL with its environ-
ment (gas, liquid, solid) plays an important role. Therefore,
knowledge about interface properties, like surface composition
and surface tension, and their relation to the chemical struc-
ture is of pivotal importance for choosing the right IL for a spe-
cific application. Owing to the break in symmetry at the inter-
face, the chemical composition of and the molecular arrange-
[a] Dr. C. Kolbeck, I. Niedermaier, Dr. A. Deyko, Dr. K. R. J. Lovelock, Dr. F. Maier,
Prof. Dr. H.-P. Steinrꢀck
Lehrstuhl fꢀr Physikalische Chemie II
Friedrich-Alexander-Universitꢁt Erlangen-Nꢀrnberg
Egerlandstrasse 3, 91058 Erlangen (Germany)
Fax: (+49)9131-85-28867
E-mail: hans-peter.steinrueck@fau.de
[b] Dr. N. Taccardi, W. Wei, Prof. Dr. P. Wasserscheid
Lehrstuhl fꢀr Chemische Reaktionstechnik,
Friedrich-Alexander-Universitꢁt Erlangen-Nꢀrnberg
Egerlandstrasse 3, 91058 Erlangen (Germany)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201304549.
Chem. Eur. J. 2014, 20, 3954 – 3965
3954
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
dynamic simulations (MDS)[9] and angle-resolved X-ray photo-
electron spectroscopy (ARXPS).[8]
cation or anion in a specific ionic liquid. Therefore, we have
performed a systematic study addressing the surface behavior
of a variety of functionalized and also some non-functionalized
ILs. Table 1 provides an overview of the studied systems, which
include imidazolium-based ILs methylated at the C2 position,
From these few studies, it is evident that at the moment no
general conclusions can be drawn and no prediction can be
made concerning the molecular arrangement of a particular
Table 1. Summary of ILs investigated in this study.
Abbreviation
Structure
IUPAC-Name
Spectra
in
[C4C1Im][TfO]
[C8C1Im]Br
1-butyl-3-methylimidazolium trifluoromethylsulfonate
1-methyl-3-octylimidazolium bromide
Figure 1
Figure 1
Figure 1
[C8C1Im][TfO]
1-methyl-3-octylimidazolium trifluoromethylsulfonate
[C8C1Im][Tf2N]
[C4C1C1Im][TfO]
[C8C1C1Im]Br
1-methyl-3-octylimidazolium bis[(trifluoromethyl)sulfonyl]imide
1-butyl-2,3-dimethylimidazolium trifluoromethylsulfonate
1,2-dimethyl-3-octylimidazolium bromide
Figure 1
Figure 1
Figure 1
Figure 1
Figure 1
[C8C1C1Im][TfO]
[C8C1C1Im][Tf2N]
1,2-dimethyl-3-octylimidazolium trifluoromethylsulfonate
1,2-dimethyl-3-octylimidazolium bis[(trifluoromethyl)sulfonyl]imide
1-methyl-3-(3-phenyl-propyl)imidazolium bis[(trifluoromethyl)-
sulfonyl]imide
[(PhC3H6)C1Im][Tf2N]
[((EtO)3SiC3H6)C1Im]Cl
[(C4F9C2H4)C1Im]I
Figure 2
Figure 3
Figure 4
Figure 5
1-methyl-3-(3-triethoxysilanepropyl)-imidazolium chloride
1-methyl-3-(3,3,4,4,5,5,6,6,6-nonafluorohexyl)imidazolium iodide
1-dodecyl-3-methylimidazolium 4-fluorobutylsulfonate
[C12C1Im][FC4H8SO3]
[C2C1Im][ClC4H8SO3]
[C8C1Im][ClC4H8SO3]
1-ethyl-3-methylimidazolium 4-chlorobutylsulfonate
1-methyl-3-octylimidazolium 4-chlorobutylsulfonate
Figure 6
Figure 6
[(ClC2H4)C1C1Im][TfO]
[(BrC2H4)C1Im][Tf2N]
[(IC2H4)C1Im][Tf2N]
1,2-dimethyl-3-(2-chloroethyl)imidazolium trifluoromethylsulfonate
Figure S1
Figure S2
Figure S3
1-(2-bromoethyl)-3-methylimidazolium bis[(trifluoromethyl)sulfonyl]-
imide
1-(2-iodoethyl)-3-methylimidazolium bis[(trifluoromethyl)sulfonyl]-
imide
1-[2-(2-methoxy-ethoxy)ethyl]-3-methylimidazolium
bis[(trifluoromethyl)sulfonyl]imide
in text
only
[Me(EG)2C1Im][Tf2N]
[(MeSC2H4)C1Im][Tf2N]
1-methyl-3-(2-methylthio)-ethylimidazolium bis[(trifluoromethyl)-
sulfonyl]imide
Figure 7
1-ethyl-3-methyl-2-methylthioimidazolium bis[(trifluoromethyl)-
sulfonyl]imide
[C2C1(MeS)Im][Tf2N]
Figure S4
1-methyl-3-(3-dimethylaminopropyl)imidazolium trifluoromethyl-
sulfonate
[(Me2NC3H6)C1Im][TfO]
[(Me2NC3H6)PBu3][FAP]
Figure 8
Figure 8
Figure 9
(3-dimethylaminopropyl)tributylphosphonium tris(pentafluoroethyl)-
trifluorophosphate
[(HOC2H4)2Me2N]
[H2NC2H4SO3]
di(2-hydroxoethyl)-dimethyl-ammonium 2-aminoethylsulfonate
Chem. Eur. J. 2014, 20, 3954 – 3965
www.chemeurj.org
3955
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
a phenyl-functionalized IL, an alkoxysilane-functionalized IL,
halo-functionalized ILs, thioether-functionalized ILs, and amine-
functionalized ILs. In addition, results for corresponding non-
functionalized ILs are recapitulated for comparison, if necessa-
ry. In each case, ARXPS is used to study the chemical composi-
tion and the molecular arrangement at the IL/vacuum inter-
face. XP spectra of the relevant core levels are measured at
two emission angles. At 08 (with respect to the surface
normal), information about the first 7–9 nm of the surface is
obtained, whereas at 808 only the outermost IL layers with an
information depth of 1–1.5 nm are probed. Thus, a visual com-
parison of the 08 and 808 measurements provides direct infor-
mation on orientation and/or enrichment of molecules at the
surface. A preferential increase/decrease in core-level intensity
with increasing emission angle indicates a higher/lower con-
centration of this element in the outermost layers compared
with the bulk.
shows that the C2 methylation increases melting point and vis-
cosity of the IL, whereas density, conductivity, and polarity are
decreased.[21–24] The impact of C2 methylation on the surface
tension was studied by Freire et al.[25] and Yoshida et al.[23] Both
groups found lower surface tension for the non-methylated
system than for the methylated one. Based on the observed
change of the surface tension, we also expect a change in the
chemical composition of the outer surface. To systematically
study the effect of the C2 methylation on the surface composi-
tion, we used ARXPS to investigate ILs with anions of different
size and ILs with the same anion, but with different alkyl chain
lengths. We compare the C1s spectra of [C8C1C1Im]Br,
[C8C1C1Im][TfO], [C8C1C1Im][Tf2N], and [C4C1C1Im][TfO] in 08 and
808 emission, with the C1s spectra of the respective non-meth-
ylated systems. As [C8C1C1Im]Br and [C8C1C1Im][TfO] are solid at
room temperature, all spectra shown in Figure 1 were mea-
sured at 350 K to rule out possible temperature effects.[26] To
allow for a visual comparison of the different systems, all spec-
tra were normalized to the N1s signal of [C8C1Im]Br in 08 emis-
sion and the binding energy was referenced to the Calkyl signal
at 285.0 eV. The Calkyl signal stems from all carbon atoms of the
alkyl (octyl or butyl) chain that are solely bound to carbon or
hydrogen atoms. The carbon atoms of the imidazolium head
group, atoms that are bound to at least one nitrogen atom,
are combined in a signal at approximately 286.5 eV (denoted
as Chetero). For the [TfO]ꢁ- and [Tf2N]ꢁ-based ILs, an additional
Results and Discussion
Methylation of C2 position of imidazolium systems
([CnC1Im]+ vs. [CnC1C1Im]+)
One efficient way to modify the physical properties of imidazo-
lium-based ILs is the methylation of the C2 position of imid-
azolium cations. Comparison with the non-methylated systems
Figure 1. C1s spectra of [CnC1Im]X (black) and [CnC1C1Im]X (red) systems in 08 (top) and 808 (bottom) emission at 350 K. Difference spectra ([CnC1C1Im]Xꢁ
[CnC1Im]X) are depicted in blue. From left to right: n=8, X=Brꢁ; n=8, X=[TfO]ꢁ; n=8, X=[Tf2N]ꢁ; n=4, X=[TfO]ꢁ. To enable a visual comparison, the spec-
tra were normalized to the N1s signal area of the imidazolium nitrogen atoms and referenced to the Calkyl signal at 285.0 eV. At the top of the figure, the
chemical structure of the [C8C1C1Im]+ cation including the atom assignment is shown.
Chem. Eur. J. 2014, 20, 3954 – 3965
www.chemeurj.org
3956
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
peak (not shown) at approximately 292.5 eV is observed, which
is attributed to the CF3 group of the anion.
At 08 emission, all [CnC1C1Im]+ systems display an increased
C1s signal area compared with the respective [CnC1Im]+ sys-
tems owing to the additional carbon atom. By subtracting the
C1s spectrum of the [CnC1Im]+ system from the C1s spectrum
of the [CnC1C1Im]+ system, difference spectra are obtained,
which are also depicted in Figure 1. All difference spectra show
a peak at approximately 286 eV, which is attributed to the ad-
ditional methyl group of the [CnC1C1Im]+ cation. The areas of
these peaks fit well (ꢂ15%) with the intensity expected for
one carbon atom. The minimum at the higher binding energy
side of the difference spectra is attributed to the different elec-
tronic structure of the imidazolium ring of the [CnC1C1Im]+ sys-
tems compared with the [CnC1Im]+ systems. The signal of the
C2 carbon is shifted slightly towards higher binding energies,
and the C1s signals of C4 and C5 carbon atoms as well as the
N1s signal of the imidazolium nitrogen atom (not shown) are
shifted towards lower binding energies (see also ref. [27]).
As already discussed, the alkyl chains of imidazolium cations
at the surface in [CnC1Im]+ systems (nꢀ4) tend to orientate
themselves towards the vacuum, a trend that is reflected by
an increase in the Calkyl signal intensity in 808 emission. The
comparison of the 808 spectra of the [CnC1Im]+ and
[CnC1C1Im]+ systems in Figure 1 shows a very similar increase
in the Calkyl signal, suggesting a comparable alkyl enrichment
for both systems. To verify that the differences in the spectra
are indeed only due to the additional carbon atom in the
[CnC1C1Im]+ systems, we also show difference spectra for 808
emission in Figure 1. The analysis of these spectra shows that
the difference for each IL is a single peak at 286 eV that is due
to the additional carbon atom, with only negligible changes in
the position of the Calkyl peak at 285 eV. Interestingly, for all sys-
tems, the peak at 808 resulting from the additional carbon
atom at approximately 286 eV is somewhat larger than that at
08. This indicates that the methyl group at the C2 atom is pref-
erentially orientated towards the vacuum, an observation that
is in line with the ARXPS study by Lockett et al., who postulat-
ed the same surface orientation for [C4C1C1Im][BF4].[8] This ori-
entation effect is most distinct for [C8C1C1Im]Br. It is attributed
to the low molecular volume and the spherical shape of the
bromide counterion in comparison to the bigger and more
elongated [TfO]ꢁ and [Tf2N]ꢁ anions.
Figure 2. C1s, N1s, and F1s spectra of [C8C1Im][Tf2N] (black) and
[(PhC3H6)C1Im][Tf2N] (red), in 08 (left) and 808 (right) emission. Grey dashed
lines are added as guides to the eye. At the top of the figure, the chemical
structure of [(PhC3H6)C1Im][Tf2N] is shown.
287.1, and 285.2 eV are assigned to the two carbon atoms of
the [Tf2N]ꢁ anion (CCF ), the five carbon atoms with nitrogen
3
atoms as their neighbor (Chetero), and the seven carbon atoms
of the alkyl chain (Calkyl), respectively. For [(PhC3H6)C1Im][Tf2N],
the C1s peak at 285.1 eV is assigned to the phenyl-functional-
ized side chain (labeled Cphenyl). Within the margin of error, this
binding energy matches the value for the Calkyl signal
(285.2 eV) of [C8C1Im][Tf2N], a fact that indicates that the differ-
ences in C1s binding energy resulting from the different hy-
bridization of alkyl and phenyl carbon atoms, are below the
resolution of our apparatus. Interestingly, the intensity of the
signal at approximately 285.1 eV is identical for both ILs, al-
though the phenyl-functionalized side chain contains one addi-
tional carbon atom. This behavior is attributed to the fact that
for phenyl rings, intensity of the main C1s peak is transferred
to a p!p* shake-up satellite at approximately 291.7 eV (de-
noted as Csat in Figure 2).[28] Similar shake-up satellites are also
observed for the imidazolium ring, but are removed from the
spectra by the subtraction of a three-point linear background.
The binding energies and intensities of all other XPS signals of
[(PhC3H6)C1Im][Tf2N] match those of [C8C1Im][Tf2N], indicating
that the phenyl group has no influence on the core levels of
the imidazolium ring and the anion.
Phenyl-functionalized IL
The influence of a phenyl-functionalized side chain on the sur-
face orientation of imidazolium cations was investigated by
comparing ARXPS measurements of [(PhC3H6)C1Im][Tf2N] and
[C8C1Im][Tf2N]. In Figure 2 (left column), the corresponding
C1s, N1s, and F1s spectra at 08 emission are depicted in red
([(PhC3H6)C1Im][Tf2N]) and black ([C8C1Im][Tf2N]). In the N1s
region, two peaks are observed at 402.2 and 399.6 eV, which
originate from the two nitrogen atoms of the imidazolium ring
(NIm) and the nitrogen atom of the [Tf2N]ꢁ anion (NTf N), respec-
2
tively. For [C8C1Im][Tf2N], the peak assignment in the C1s
region has already been mentioned: the three peaks at 293.1,
Chem. Eur. J. 2014, 20, 3954 – 3965
www.chemeurj.org
3957
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Information on the surface orientation can be derived from
the 808 emission spectra of [C8C1Im][Tf2N] and [(PhC3H6)C1Im]-
[Tf2N] shown in Figure 2 (right column). To enable a visual
comparison with the 08 emission spectra, the 808 emission
spectra were multiplied by an empirical factor (see Experimen-
tal Section), and grey dashed horizontal lines were added as
guides for the eye. For [C8C1Im][Tf2N], the increase of the Calkyl
intensity at 808 (compared with 08), and the simultaneous de-
crease of all other XPS signals reflect the previously reported
surface enrichment of alkyl chains, which form a non-polar
layer above the imidazolium rings and the anions. In contrast,
for [(PhC3H6)C1Im][Tf2N], the Cphenyl signal shows only a minor
increase at 808. Despite this weak increase, the signals of the
imidazolium ring (Chetero and NIm) decrease in the 808 emission
spectra by ~10%, indicating that the phenyl groups are on
average closer to the outer surface than the imidazolium rings.
Analysis of the [Tf2N]ꢁ anion signals shows an increase in the
F1s peak and a decrease in the O1s peak (not shown) at 808
compared with 08, whereas the intensities of NTf N, CCF , and
2
3
S2p (not shown) stay constant. This behavior indicates that
the [Tf2N]ꢁ anion orientates itself at the surface with its CF3
groups pointing towards the vacuum and the SO2 groups
pointing towards the bulk, an observation in agreement with
conclusions based upon Rutherford backscattering experi-
ments.[29] As both the Cphenyl and the F1s signal show a similar
increase in intensity (by approximately 7%) on going from 08
to 808, we propose that the phenyl rings and the CF3 groups
lie next to each other at the outer surface; this contrasts with
the situation for [C8C1Im][Tf2N] where the alkyl chains lie above
the CF3 groups.
The different surface orientations of [(PhC3H6)C1Im]+ and
[C8C1Im]+ cations can be explained by the different interaction
of alkyl and phenyl groups with the imidazolium rings and the
anions. Owing to the non-polar nature of alkyl chains,
[CnC1Im]+-based ILs with nꢀ4 form a heterogeneous structure
in the bulk, with the alkyl chains segregating into non-polar re-
gions, while the imidazolium rings and the anions form more
polar regions.[30,31] In contrast, the quadrupole moment of
phenyl groups promotes their interaction with the anions and
cationic head groups. This is concluded from molecular dy-
namic simulations as well as X-ray crystallography and neutron
diffraction experiments on IL–benzene mixtures, the results of
which showed that the anions prefer an equatorial position
around benzene, whereas the imidazolium cations lie below or
above the benzene ring.[32–36]
Figure 3. Si2p, O1s, C1s, N1s, and Cl2p spectra of [((EtO)3SiC3H6)C1Im]Cl in
08 (black) and 808 emission (green), and the chemical structure of the IL.
the IL are superimposed. To simplify the analysis, the signal is
treated as the superposition of two peaks, similar to the non-
functionalized [CnC1Im]+ systems. The C1s peak at 286.5 eV
(Chetero) contains the carbon atoms bound to nitrogen or
oxygen atoms; all other carbon atoms, that is, those bound to
only carbon or to silicon atoms, contribute to the peak at
284.9 eV (Calk,Si). The quantitative analysis of the 08 emission
matches the nominal atom ratio within the margin of error of
ꢂ5%, confirming this assignment and also the purity of the IL.
Information on the orientation of the IL at the outer surface
is obtained from the 808 emission spectra. The imidazolium
N1s signals and the anion Cl2p signal show the most pro-
nounced changes: both decrease when going from 08 to 808
emission, an observation that indicates a loss of atom density
in these species at the outer surface compared with the bulk.
The larger decrease for the N1s signal (to ~65% of the intensi-
ty at 08) than for the Cl2p signal (to ~75%) indicates a larger
mean distance from the imidazolium ring to the outer surface
than from the chloride anion to the surface; in other words,
the anion lies slightly above the imidazolium ring. On the
other hand, the Si2p and O1s signals from the triethoxysilane
group gain intensity at 808, indicating an enrichment of the
functionalized chain at the outer surface. This preferential sur-
face orientation of the IL cation is also reflected in the C1s
spectra: the Chetero signal predominantly stems from the imida-
zolium ring (5 out of 8 atoms), whereas the Calk,Si signal origi-
nates solely from the functionalized chain. On going from 08
to 808 emission, the Calk,Si signal increases and the Chetero signal
Alkoxysilane-functionalized IL [((EtO)3SiC3H6)C1Im]Cl
In Figure 3, the XP spectra of the alkoxysilane-functionalized IL
[((EtO)3SiC3H6)C1Im]Cl are depicted for both emission angles, 08
(black) and 808 (green). In the Si2p, O1s, N1s, and Cl2p re-
gions, only one signal is observed, as expected from the struc-
ture of the IL. Note that the Si2p and the Cl2p levels are spin-
orbit split; for the Si2p level, the splitting is not resolved,
whereas it is clearly seen for the Cl2p level, with the expected
intensity ratio of 1:2 for the 2p1/2 and 2p3/2 components. In the
C1s region, the signals of the different carbon environments of
Chem. Eur. J. 2014, 20, 3954 – 3965
www.chemeurj.org
3958
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
decreases in intensity, an observation that confirms that the
functionalized chains point towards the vacuum, whereas the
imidazolium rings lie below.
These observations are in line with the reported high surface
activity of polysiloxanes. Marginal polysiloxane contamination
can be detected at the outer surface of an IL, even for bulk
concentrations below the detection limit of NMR spectrosco-
py.[14,37–39] A contamination with such surface-active substances
may impose huge problems for all interface-related processes
as they modify the surface properties, often in an uncontrolled
way. However, we can also take advantage of the surface activ-
ity of alkoxysilanes and siloxanes. For example, catalysts and
other substances could be placed preferentially at the surface/
interface of an ionic liquid medium by introducing these
groups into their chemical structure.
Halo-functionalized ILs
Perfluorinated alkyl compounds are known for their surface ac-
tivity in aqueous[40,41] and in IL systems.[42–45] The very rigid and
non-coordinating CꢁF bonds lead to very weak dispersive in-
termolecular interactions between the perfluorinated alkyl
chains and, therefore, perfluorinated systems are energetically
unfavored in the bulk.[46] In a molecular dynamics simulation
study of the perfluorinated system [(CnF2n+1C2H4)C1Im][Tf2N]
with n=2–6, Smith et al. showed that the perfluorinated ILs
form a heterogeneous structure in the bulk similar to that of
their non-perfluorinated analogues ([CnC1Im][Tf2N]).[47] Thereby,
the perfluoroalkyl chains tend to stick together in a non-polar
region, whereas the surrounding polar region consists of the
cationic head groups and the anions.[47] Therefore, we expect
that perfluorinated alkyl chains are also enriched at the outer
surface of the IL, similar to the [CnC1Im]+ systems. To verify this
assumption, we investigated the perfluorinated IL
[(C4F9C2H4)C1Im]I and the fluoro-functionalized IL [C12C1Im]-
[FC4H8SO3].
Figure 4. C1s, N1s, F1s, and I3d5/2 spectra of [(C4F9C2H4)C1Im]I in 08 (black)
and 808 emission (green) taken at approximately 355 K. At the top of the
figure, the chemical structure of the IL is shown.
[C12C1Im][FC4H8SO3] is not as obvious. In Figure 5, the C1s, F1s,
N1s, and O1s spectra are depicted. The quantitative analysis of
the 08 spectra (black) confirms the purity of the sample. Inter-
estingly, on going from 08 to 808 emission, a slight decrease in
In Figure 4, the ARXP spectra of [(C4F9C2H4)C1Im]I are depict-
ed. The spectra were taken at a temperature of approximately
355 K to ensure the sample was liquid and to evaporate resid-
ual 1-iodo-1H,1H,2H,2H-perfluorohexane, which was used as
a starting material in the IL synthesis. The quantitative analysis
of the 08 emission spectra (black spectra in Figure 4) matches
the nominal atom ratio to within the margin of error, indicat-
ing the high purity of the IL. In the C1s region, three signals
are observed: the peaks at 291.7 and 294.1 eV are attributed to
the CF2 and CF3 groups, respectively. All other carbon atoms
contribute to the peak at 286.7 eV (Chetero). By changing from
08 to 808, the F1s and the CCF signals increase in intensity,
x
whereas the Chetero, N1s, and I3d5/2 signals decrease. This be-
havior unequivocally demonstrates the enrichment of the per-
fluorinated chains at the outer surface, whereas the imidazoli-
um ring and the iodide anion lie below. The identical decrease
of the N1s and I3d5/2 peaks at 808 indicates that the imidazoli-
um rings and the anions have the same mean distance to the
outer surface.
Figure 5. C1s, F1s, N1s, and O1s spectra of [C12C1Im][FC4H8SO3] in 08 (black)
and 808 emission (green). At the top of the figure, the chemical structure of
the IL is shown.
In contrast to the perfluorinated IL, the surface enrichment
of the fluoro-terminated chain of the fluoro-functionalized IL
Chem. Eur. J. 2014, 20, 3954 – 3965
www.chemeurj.org
3959
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
the F1s intensity is observed instead of the increase expected
from surface enrichment. This can be explained by the surface
orientation of the [C12C1Im]+ cation. As already discussed, alkyl
chains tend to stick out towards the vacuum. This is also the
case here, as can be seen from the increase in the Calkyl signal
at 808. Because of this alkyl enrichment, all other signals are
damped. The degree of damping indicates the mean distance
of the species to the outer surface. At 808 emission, the imid-
azolium N1s signal and the O1s signal from the SO3 group of
the anion both decrease to approximately 59% of their 08
emission intensity. At the same time, the F1s signal only de-
creases to about 89%. This indicates that the cationic and
anionic head groups have the same mean distance to the
outer surface and that the fluorobutyl chain points towards
the vacuum. Despite this orientation, the F1s signal experien-
ces a slight decrease at 808 because the functionalized chain is
shorter than the alkyl chain and thus the F atoms at the end of
the chain are not at the outer surface. The results for
[(C4F9C2H4)C1Im]I and [C12C1Im][FC4H8SO3] show that the pre-
ferred surface orientation of the fluorinated chains towards the
vacuum is independent of the charge of the head group, that
is, it occurs for both functionalized cations and functionalized
anions.
As the next step, we addressed the effect of the counter ion
on the degree of surface ordering. We compared two ILs with
the same functionalized 4-chlorobutylsulfonate anion, but dif-
ferent cations, namely [C2C1Im][ClC4H8SO3] and [C8C1Im]
[ClC4H8SO3], where the cations differ from each other only by
the length of the alkyl chain at the imidazolium ring. In a previ-
ous study, it was shown that this anion preferentially orientates
itself at the outer surface with the 4-chlorobutyl chain pointing
towards the vacuum and the SO3 group towards the bulk.[17,18]
Evidence for this orientation can also be seen in Figure 6,
which shows the C1s, N1s, Cl2p, and O1s spectra of [C2C1Im]
[ClC4H8SO3] (black) and [C8C1Im][ClC4H8SO3] (red) for 08 (left
column) and 808 emission (right column). To remove the effect
of the different IL density and, thus, to enable a direct compar-
ison, the XP spectra of [C8C1Im][ClC4H8SO3] are normalized to
the N1s signal of [C2C1Im][ClC4H8SO3] in 08 emission. For both
ILs, the increase of the Cl2p signal and the decrease of the O
and S signals at 808 confirm the proposed preferred orienta-
tion of the functionalized chain with the Cl atoms towards the
vacuum and the SO3 groups towards the bulk. For each IL, the
ionic head groups have the same distance to the outer surface,
as is deduced from the identical decrease of the O1s and N1s
signals at 808. When comparing the two ILs, the more pro-
nounced decrease at 808 for [C8C1Im][ClC4H8SO3] stems from
the stronger attenuation of the head group signals by the
longer octyl chains compared with the ethyl chains of [C2C1Im]
[ClC4H8SO3].
Figure 6. C1s, Cl2p, O1s, and N1s spectra of [C2C1Im][ClC4H8SO3] (black) and
[C8C1Im][ClC4H8SO3] (red) in 08 (left) and 808 emission (right); data partly
from Ref. [18]. The XP spectra of [C8C1Im][ClC4H8SO3] are normalized to the
N1s signal intensity of [C2C1Im][ClC4H8SO3] measured at 08 emission to ac-
count for the different IL density.
0.66 for [C8C1Im][ClC4H8SO3]. The larger value for the latter IL
indicates that the surface orientation is more pronounced for
the IL with the longer alkyl chain. This effect is attributed to
the enhanced dispersive intermolecular interactions of the 4-
chlorobutyl chain with the octyl chain of [C8C1Im]+ compared
with the ethyl chain of [C2C1Im]+. The somewhat smaller abso-
lute increase in the Cl2p signal at 808 for [C8C1Im][ClC4H8SO3]
compared with [C2C1Im][ClC4H8SO3] is assigned to the fact that
the functionalized chain is shorter than the octyl chain; thus,
the Cl atom at the end of the functionalized chain is not at the
outer surface and is therefore damped by the longer octyl
chain. This effect does not occur for the IL with the shorter
ethyl chain.
The degree of the preferred surface orientation of the func-
tionalized anion chains can be evaluated by the ratio of Cl2p
and O1s signals. For both ILs, quantitative analysis of the bulk-
sensitive geometry (08 emission) matches the nominal value of
0.33 to within the margin of error, as is expected for clean ILs.
In the surface-sensitive geometry (808), the Cl2p/O1s signal
ratio rises significantly; to 0.56 for [C2C1Im][ClC4H8SO3] and to
Chem. Eur. J. 2014, 20, 3954 – 3965
www.chemeurj.org
3960
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Comparing the Cl2p/O1s signal ratio for [C2C1Im][ClC4H8SO3]
and [C8C1Im][ClC4H8SO3] (see above) to the value of 0.41 for
the F1s/O1s signal ratio of [C12C1Im][FC4H8SO3] in 808 emission
shows that the degree of surface orientation is smaller for
[C12C1Im][FC4H8SO3] in spite of the longer alkyl chain. It is sug-
gested that this is also due to weaker dispersive interactions of
alkyl chains with CꢁF bonds than with CꢁCl bonds, which have
a greater polarizability.[46,48]
Furthermore, we studied and compared three different ILs
with a haloethyl chain attached to the imidazolium ring,
namely the chloro-functionalized IL [(ClC2H4)C1C1Im][TfO], the
bromo-functionalized IL [(BrC2H4)C1Im][Tf2N], and the iodo-
functionalized IL [(IC2H4)C1Im][Tf2N] (Table 1). For these three
ILs, ARXPS reveals more or less isotropic behavior of the func-
tionalized cations (i.e., no significant surface enrichment) at
the outer surface (spectra are provided in the Supporting Infor-
mation as Figures S1–S3). From the lack of orientation of the
halogen-functionalized chain towards the vacuum (which was
observed, e.g., for [ClC4H8SO3]ꢁ systems—see above), we con-
clude that the haloethyl chains are too short to form a hetero-
geneous structure at the outer surface. This is in line with pre-
vious investigations on [CnC1Im][Tf2N] where alkyl surface en-
richment was only found for nꢀ4, but not for shorter
chains.[15,49]
Figure 7. C1s, N1s, S2p, F1s, and O1s spectra of [(MeSC2H4)C1Im][Tf2N] in 08
(black) and 808 emission (green), along with the chemical structure of the IL.
Thioether-functionalized ILs
In a previous study, we showed that ether-functionalized (i.e.,
ethylene glycol-functionalized) cations, like [Me(EG)nC1Im]+
(n=1,2,3), do not exhibit a preferred surface orientation owing
to the formation of intra- and intermolecular hydrogen bonds
between the oxygen atoms and the acidic hydrogen atoms of
the imidazolium ring.[15,49] Herein, the related thioether-func-
tionalized IL [(CH3SC2H4)C1Im][Tf2N] is studied. The respective
XP spectra in 08 and 808 emission are depicted in Figure 7. The
single peaks in the F1s and O1s spectra are assigned to the
anion, the C1s, N1s, and S2p regions each contain one cation
peak and one anion peak. In the C1s and S2p region, the peak
at higher binding energy is attributed to the IL anion and the
signal at lower binding energy to the IL cation. In the N1s
region, the peak assignment is reversed. The purity of the
sample is once again verified by the quantitative analysis of
the 08 measurement, which matches the nominal atom ratio to
within the margin of error.
We first address the surface orientation of the [Tf2N]ꢁ anion
by comparing the XP spectra at 08 and 808 emission in
Figure 7. Whereas the Canion, Nanion, and Sanion signals show no
distinct angular dependence, the F1s signal increases and the
O1s signal decreases at 808. This indicates a preferred surface
orientation of the [Tf2N]ꢁ anion for the cis conformation with
its CF3 groups pointing towards the vacuum and its SO2
groups pointing towards the bulk. This finding is in agreement
with the results of the fluoro-functionalized ILs previously dis-
cussed and observed in the literature for [CnC1Im][Tf2N] sys-
tems.[8,29,50–52]
signal. However, the Ncation signal, which particularly probes the
position of the imidazolium ring relative to the outer surface,
is lower in intensity (89%) at 808 emission (note that the Ccation
signal does not show pronounced angle dependence as the
carbon atoms in the ring and in the chain cannot be distin-
guished by the limited energy resolution). From this behavior,
we propose that the imidazolium ring lies slightly below the
[Tf2N]ꢁ anion with the thioether-functionalized chain pointing
towards the vacuum. However, as the changes in the ARXP
spectra are small, this preferred orientation is not very pro-
nounced.
In addition, we also studied the thioether-functionalized IL
[C2C1(MeS)Im][Tf2N] (Table 1). The methylthio group at the C2
position of the imidazolium ring has an effect on the electronic
structure of the cation, as was shown in a previous study.[53]
The impact on the surface composition is, however, not dis-
tinct. The angle-dependent changes in the XP spectra (see Fig-
ure S4 in the Supporting Information) are only minor and
show no clear tendency of the thioether chain pointing to-
wards the vacuum.
Amine-functionalized ILs
The effect of the cationic head group on the surface enrich-
ment of functionalized chains is investigated for a 3-dimethyl-
aminopropyl chain substituted on the two different cationic
head groups, 1-methylimidazolium and tributylphosphonium.
The corresponding ILs are [(Me2NC3H6)C1Im][TfO] and
[(Me2NC3H6)PBu3][FAP]. Note that the results were partly al-
For the cation, no angular dependence is observed for the
thioether-functionalized chain as shown by the constant Scation
Chem. Eur. J. 2014, 20, 3954 – 3965
www.chemeurj.org
3961
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ready presented in ref. [18] and are included in this work to
provide a comprehensive overview of the influence of func-
tional groups on the surface composition of ILs. The N1s spec-
tra of the imidazolium-based IL in Figure 8a contain peaks
from the cation head group at 403.0 eV (NIm) and the tertiary
Figure 9. N1s, O1s, C1s, and S2p spectra of [(HOC2H4)2Me2N][H2NC2H4SO3] in
08 (black) and 808 emission (green), along with the chemical structure of the
IL. Note the small carbon contamination, which does not affect the deduced
conclusions, however.
Figure 8. N1s spectra of [(Me2NC3H6)C1Im][TfO] and N1s and P2p spectra of
[(Me2NC3H6)PBu3][FAP] in 08 (black) and 808 emission (green), along with the
chemical structures of the ILs; data partly from Ref. [18].
to carbon and/or hydrogen atoms. As no additional signals are
observed in the XP spectra and the quantitative analysis of the
IL signals matches the nominal atom ratio to within the
margin of error, we attribute the C1s signal to a hydrocarbon
contamination (Ccontam). On increasing the emission angle from
08 to 808, a strong increase in the intensity of the Ccontam signal
is observed, indicating the high surface activity of the contami-
nation. In addition, the Namine signal increases in intensity,
amine group of the cation at 400.3 eV (Namine). The Namine signal
increases at 808, whereas the NIm signal slightly decreases. This
behavior indicates that the amine groups tend to stick out to-
wards the vacuum. This result is confirmed by a molecular dy-
namics simulation by Xing et al., who calculated the surface
orientation of the IL [(H2NC3H6)C1Im][BF4]. They found that the
primary amine group tethered to the imidazolium cation is
also preferentially orientated towards the vacuum.[12] For the
phosphonium-based IL, both the N1s and the P2p spectra
have to be considered (Figure 8b and c). The N1s signal at
399.6 eV is attributed to the amine group. The two P2p signals
at 133.2 and 134.9 eV are assigned to the phosphonium cation
(Pcation) and the [FAP]ꢁ anion (PFAP), respectively (note that the
spin-orbit splitting for the P2p level is not resolved). The ab-
sence of any intensity changes between 08 and 808 emission
indicates an isotropic arrangement of the IL at the outer sur-
face, contrary to the imidazolium-based IL.
whereas the S2p and OSO signals of the SO3 group decrease to
3
approximately 73% of their 08 emission intensity. This suggests
that the [H2NC2H4SO3]ꢁ anion preferentially orientates itself at
the surface with the amine group pointing towards the
vacuum and the negatively charged SO3 group towards the
bulk, similar to the [(Me2NC3H6)C1Im]+ cation. The surface ori-
entation of the [(HOC2H4)2Me2N]+ cation is deduced from the
behavior of the OOH and the Nammon signals. The Nammon signal
decreases at 808 emission compared with the 08 measurement,
whereas the OOH signal intensity stays constant within the
margin of error, indicating that the mean distance from Nammon
to the outer surface is larger than for OOH. A similar preferred
surface orientation of hydroxyethyl chains tethered to a cationic
head group was previously reported by Lockett et al. and Pen-
sado et al., who studied the surface orientation of the 1-(2-hy-
droxyethyl)-3-methylimidazolium cation by ARXPS and MDS,
respectively.[8,9] Our observations imply that for the investigat-
ed primary and tertiary amine-functionalized ILs, the neutral
functional groups form the outer surface, whereas the ionic
head groups lie below.
As an additional amine IL, one with a primary amine func-
tionality, we studied [(HOC2H4)2Me2N][H2NC2H4SO3], which con-
tains, next to the amine group in the anion, two hydroxy
groups in the cation. The corresponding core-level spectra are
depicted in Figure 9. The two N1s signals at 402.8 and
399.7 eV stem from the ammonium (Nammon) and the amine ni-
trogen atoms (Namine), respectively. The two peaks in the O1s
region are attributed to the SO3 group at 531.7 eV (OSO ) and
3
to the hydroxy groups at 533.1 eV (OOH). As all carbon atoms
of the IL are connected to hetero atoms, we expect only one
peak in the C1s region. However, next to the Chetero peak at
286.6 eV, a second signal is observed at 285.3 eV, which is
a binding energy typical for carbon atoms that are only bound
Conclusion
We have performed a systematic ARXPS study to investigate
the surface behavior of a variety of functionalized and some
non-functionalized ILs, including imidazolium-based ILs meth-
Chem. Eur. J. 2014, 20, 3954 – 3965
www.chemeurj.org
3962
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ylated at the C2 position, a phenyl-functionalized IL, an alkoxy-
silane-functionalized IL, halo-functionalized ILs, thioether-func-
tionalized ILs, and amine-functionalized ILs. The main results
and conclusions can be summarized as follows.
vacuum occurs for both functionalized cations and functional-
ized anions.
The effect of the counterion on the degree of surface order-
ing was studied for [C2C1Im][ClC4H8SO3] and [C8C1Im]
[ClC4H8SO3], that is, cations with different alkyl chain lengths.
For both ILs, the 4-chlorobutyl chain is enriched at the outer
surface, but the surface orientation is more pronounced for
[C8C1Im][ClC4H8SO3], owing to the enhanced dispersive inter-
molecular interactions of the 4-chlorobutyl chain with the
longer alkyl chain. A comparison with [C12C1Im][FC4H8SO3] re-
veals a lower degree of surface orientation despite the longer
alkyl chain, an observation that is attributed to weaker disper-
sive interactions of the alkyl chain with CꢁF bonds than with
CꢁCl bonds, owing to the higher polarizability of the
latter.[46,48]
Methylation at the C2 position of imidazolium systems
The influence of methylation of imidazolium cations at the C2
position was studied as a function of cation alkyl chain length
([CnC1C1Im]+ with n=4,8) and anion size (Brꢁ, [TfO]ꢁ, [Tf2N]ꢁ).
As for the non-methylated systems, the alkyl chains tend to
orientate themselves towards the vacuum. For all the methyl-
ated systems, a preferred orientation of the methyl group at
the C2 position towards the vacuum is observed, an observa-
tion that is in line with the study by Lockett et al. for
[C4C1C1Im][BF4].[8]
A comparative analysis of three halogen-functionalized ILs
with short chains, namely [(ClC2H4)C1C1Im][TfO], [(BrC2H4)C1Im]-
[Tf2N], and [(IC2H4)C1Im][Tf2N], revealed no specific surface en-
richment. This is attributed to the fact that haloethyl chains
are too short to form a heterogeneous structure at the outer
surface, in line with previous investigations on [CnC1Im][Tf2N]
for n<4.[15,49]
Phenyl-functionalized ILs
The surface arrangement was studied for [(PhC3H6)C1Im][Tf2N]
and compared with that of imidazolium cations with non-func-
tionalized alkyl chains of comparable length ([C8C1Im][Tf2N]). In
contrast to the alkyl chains, no preferential orientation towards
the vacuum is found for the phenyl-functionalized chains. Nev-
ertheless, the phenyl groups are on average closer to the outer
surface than the imidazolium rings. The [Tf2N]ꢁ anion orien-
tates itself at the surface with its CF3 groups pointing towards
the vacuum and the SO2 groups pointing towards the bulk.
The phenyl rings and the CF3 groups lie next to each other at
the outer surface, in contrast to [C8C1Im][Tf2N], where the alkyl
chains lie above the CF3 groups. The differences are attributed
to the quadrupole moment of the phenyl group, which pro-
motes its interaction with the anions and cationic head
groups, in line with molecular dynamic simulations as well as
diffraction experiments of IL–benzene mixtures.[32–36]
Thioether-functionalized ILs
These studies were motivated by previous work on the ether-
functionalized ILs [Me(EG)nC1Im]+ (n=1,2,3), which showed no
preferential surface orientation owing to the formation of
intra- and intermolecular hydrogen bonds between the
oxygen atoms and the acidic hydrogen atoms of the imidazoli-
um ring.[15,49] For the related thioether-functionalized ILs
[(CH3SC2H4)C1Im][Tf2N] and [C2C1(MeS)Im][Tf2N], the [Tf2N]ꢁ
anion shows a preferred surface orientation for the cis confor-
mation, with its CF3 groups pointing towards the vacuum and
its SO2 groups pointing towards the bulk, in agreement with
the results of the ether-functionalized ILs.[8,29,50] The imidazoli-
um ring is found to be slightly below the [Tf2N]ꢁ anion, with
the thioether-functionalized chain of [(CH3SC2H4)C1Im][Tf2N]
showing a weak preferential orientation towards the vacuum.
Alkoxysilane-functionalized ILs
For [((EtO)3SiC3H6)C1Im]Cl, we find that the functionalized
chains point towards the vacuum, whereas the imidazolium
rings lie below. The imidazolium ring is found at a larger dis-
tance from the outer surface than the chloride anion, indicat-
ing that the anion lies slightly above the cation head group.
The surface enrichment of the functionalized chain is in line
with the reported high surface activity of polysiloxanes.[14,37]
Amine-functionalized ILs
We studied two ILs with a 3-dimethylaminopropyl chain substi-
tuted on two different head groups, [(Me2NC3H6)C1Im][TfO] and
[(Me2NC3H6)PBu3][FAP]. Although for the imidazolium-based IL
the amine groups tend to stick out towards the vacuum, for
the phosphonium-based IL an isotropic surface arrangement is
found. An additional amine-functionalized IL was studied,
[(HOC2H4)2Me2N][H2NC2H4SO3], which contains two hydroxy
groups in the cation next to the amine group in the anion.
The [H2NC2H4SO3]ꢁ anion preferentially orientates itself at the
surface with the amine group pointing towards the vacuum
and the negatively charged SO3 group towards the bulk, simi-
lar to the [(Me2NC3H6)C1Im]+ cation. The [(HOC2H4)2Me2N]+
cation orients itself with the OH group closer to the surface, an
observation that implies that the neutral functional groups
Halo-functionalized ILs
Owing to their very weak dispersive intermolecular interac-
tions, perfluorinated alkyl chains are expected to be enriched
at the outer surface of the IL, similar to aqueous systems. This
behavior is indeed observed for the perfluorinated chains of
[(C4F9C2H4)C1Im]I, with the imidazolium rings and the anions
positioned underneath and at the same mean distance to the
outer surface. A corresponding study of the fluoro-functional-
ized IL [C12C1Im][FC4H8SO3] demonstrates that the preferred
surface orientation of the fluorinated chains towards the
Chem. Eur. J. 2014, 20, 3954 – 3965
www.chemeurj.org
3963
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
donated to us by Merck. All other ILs were synthesized in our labo-
ratories. Details on the synthesis are given in the Supporting Infor-
mation.
form the outer surface, whereas the ionic head groups lie
below.
General conclusions
ARXPS
From the described results, we can deduce some general
trends: The occurrence of surface enrichment of functional
groups/chains and of non-functional alkyl chains depends on
the interplay/competition between the interactions of the
ionic head groups, dispersive interactions between the ionic
head groups and the chains/functional units, and dispersive in-
teractions between the chains/functional units.
Thin IL films (~100 mm) were prepared by deposition of the corre-
sponding IL onto a planar Au foil (20 mmꢂ15 mm). These samples
were then introduced into the ultra-high vacuum (UHV) system
through a loadlock. The in situ heating of the ILs was done by
heating from the reverse side of the samples. A thermocouple at-
tached near the sample gave temperature readings with an error
of ꢂ58C.
The ARXPS measurements of the ILs [((EtO)3SiC3H6)C1Im]Cl,
[(C4F9C2H4)C1Im]I, and [C12C1Im][FC4H8SO3] were conducted with
a VG ESCALAB 200 system using non-monochromatized AlKa radia-
tion (hn=1486.6 eV) at a power of 150 W (U=15 kV, I=10 mA)
and a pass energy of 20 eV. All other ILs were measured with the
same chamber, but with a (newly installed) SPECS XR 50 X-ray
source and a VG Scienta R3000 electron analyzer, using non-mono-
chromatized AlKa radiation at a power of 250 W (U=12.5 kV, I=
20 mA) and a pass energy of 100 eV. For both series, the overall
energy resolution was approximately 0.9 eV. Unless stated other-
wise, the binding energies are denoted as measured, that is, no
referencing of the binding energy scale to a particular core level
was applied. In some cases, the 808 spectra were shifted by
~0.2 eV to correct for minor charging.
At the kinetic energies used (800–1300 eV), the inelastic mean free
path (IMFP) of photoelectrons is approximately 3 nm in organic
compounds.[20] Therefore, the information depth (=3x IMFP) at 08
emission is 7–9 nm, whereas it is only 1–1.5 nm at 808. To account
for the overall lower intensity in 808 emission and to enable
a direct visual comparison of the 08 and 808 spectra, the 808 spec-
tra were multiplied by an empirical factor.[14]
For the C1s spectra of [Tf2N]ꢁ- and [TfO]ꢁ-based ILs, a three-point
linear background subtraction was used. For all other core-level
spectra a two-point linear background subtraction was applied. All
peaks were fitted by using a Gaussian–Lorentz profile with 30%
Lorentz contribution. From the resulting peak areas and by taking
into account the sensitivity factors for the different elements,
quantitative information was obtained on the stoichiometry of the
near-surface region. The atomic sensitivity factors (ASFs) were cali-
brated for our specific experimental setup by the process reported
by Kolbeck et al.[14] This procedure yields ꢂ5% accuracy for the de-
noted atomic concentration values. The results of the quantitative
analyses of all studied ILs are given in the Tables in the Supporting
Information. Beam damage was observed for some ionic liquids
after 2–3 h of X-ray exposure; all results in this paper are from ex-
periments recorded before beam damage was detected.
1) For weak interactions between the functional units/chains
and the head groups, typically, surface enrichment of the func-
tional groups/chains is observed when the chain length is suffi-
cient (nꢀ4); examples are imidazolium-based ILs with non-
functionalized, halogen-functionalized, alkoxysilane-functional-
ized, thioether-functionalized and amine-functionalized alkyl
chains. For short chains (n<4), that is, chloro- or iodoethyl, no
such enrichment is found. These effects are found for function-
al groups attached to the cationic and also the anionic head
groups.
2) In the case of significant interactions of the functional
group with the ionic head groups, for example, through hydro-
gen bonding for ether groups or quadrupole interactions for
phenyl groups, surface enrichment of the functional chains is
suppressed, in contrast to non-functional chains of the same
length.
3) In the case of surface enrichment of the functional
groups/chains, the cationic and anionic head groups typically
form a polar layer underneath, at equal or very similar distance
to the outer surface.
4) For ILs with functionalized/non-functionalized chains on
both the anion and the cation (e.g., [C12C1Im][FC4H8SO3]), a sur-
face arrangement is observed with both chains pointing to-
wards the vacuum, but in the case of [C12C1Im][FC4H8SO3], the
F atoms are not at the outer surface due to the shorter length
of the functionalized chain.
5) Dispersive interactions between the functional groups/
chains on both anions and cations contribute to enhanced sur-
face enrichment. Thus, double-functionalized ILs with short
functionalized chains at the anion and the cation, for example,
[(HOC2H4)2Me2N][H2NC2H4SO3], show enrichment of both func-
tional groups.
6) Independent of the occurrence of surface enrichment of
the cation, for all studies with the [Tf2N]ꢁ anion, the anion
shows a preferred surface orientation for the cis conformation,
with its CF3 groups pointing towards the vacuum and its SO2
groups pointing towards the bulk.
Acknowledgements
H.-P.S., F.M., and C.K. thank the DFG for support through grant
STE 620/9–1. P.W. thanks the European Research Council (ERC)
for financial support for this research in the context of an Ad-
vanced Investigator Grant to P.W (H2-SMS-Cat, No 267376). In-
frastructural contributions by the Erlangen Cluster of Excel-
lence “Engineering of Advanced Materials” are also gratefully
acknowledged. The authors thank Prof. B. Etzold for the syn-
thesis of the IL [((EtO)3SiC3H6)C1Im]Cl.
Experimental Section
Materials
A summary of all ILs studied in this paper is given in Table 1.
[C4C1Im][TfO], [C4C1C1Im][TfO], [C8C1Im][TfO], and [C8C1Im]Br were
purchased from Merck, and [C8C1C1Im]Br and [C8C1C1Im][Tf2N] were
Chem. Eur. J. 2014, 20, 3954 – 3965
www.chemeurj.org
3964
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[27] T. Cremer, C. Kolbeck, K. R. J. Lovelock, N. Paape, R. Wçlfel, P. S. Schulz,
P. Wasserscheid, H. Weber, J. Thar, B. Kirchner, F. Maier, H.-P. Steinrꢀck,
Chem. Eur. J. 2010, 16, 9018.
[28] D. Briggs, M. P. Seah, Practical Surface Analysis Vol. 1, 2nd ed., Springer-
Verlag, Berlin, 1990.
Keywords: functionalized ionic liquids · surface analysis ·
surface enrichment · surface orientation · X-ray photoelectron
spectroscopy
[29] K. Nakajima, A. Ohno, H. Hashimoto, M. Suzuki, K. Kimura, J. Chem. Phys.
2010, 133, 044702.
[1] A. J. Carmichael, K. R. Seddon, J. Phys. Org. Chem. 2000, 13, 591.
[2] M. Hasib-ur-Rahman, M. Siaj, F. Larachi, Chem. Eng. Process. 2010, 49,
313.
[3] F. Karadas, M. Atilhan, S. Aparicio, Energy Fuels 2010, 24, 5817.
[4] L. Vidal, M. L. Riekkola, A. Canals, Anal. Chim. Acta 2012, 715, 19.
[5] S. G. Lee, Chem. Commun. 2006, 1049.
[6] K. R. J. Lovelock, Phys. Chem. Chem. Phys. 2012, 14, 5071.
[7] C. S. Santos, S. Baldelli, Chem. Soc. Rev. 2010, 39, 2136.
[8] V. Lockett, R. Sedev, S. Harmer, J. Ralston, M. Horne, T. Rodopoulos,
Phys. Chem. Chem. Phys. 2010, 12, 13816.
[9] A. S. Pensado, M. F. C. Gomes, J. N. C. Lopes, P. Malfreyt, A. A. H. Padua,
Phys. Chem. Chem. Phys. 2011, 13, 13518.
[10] D. Wakeham, P. Niga, C. Ridings, G. Andersson, A. Nelson, G. G. Warr, S.
Baldelli, M. W. Rutland, R. Atkin, Phys. Chem. Chem. Phys. 2012, 14, 5106.
[11] C. Ridings, G. G. Warr, G. G. Andersson, Phys. Chem. Chem. Phys. 2012,
14, 16088.
[12] H. B. Xing, Y. Yan, Q. W. Yang, Z. B. Bao, B. G. Su, Y. W. Yang, Q. L. Ren, J.
Phys. Chem. C 2013, 117, 16012.
[13] N. Paape, W. Wei, A. Bosmann, C. Kolbeck, F. Maier, H.-P. Steinrꢀck, P.
Wasserscheid, P. S. Schulz, Chem. Commun. 2008, 3867.
[14] C. Kolbeck, M. Killian, F. Maier, N. Paape, P. Wasserscheid, H.-P. Steinrꢀck,
Langmuir 2008, 24, 9500.
[15] 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.
[16] F. Maier, Angew. Chem. 2011, 123, 10315; Angew. Chem. Int. Ed. 2011, 50,
10133.
[17] C. Kolbeck, I. Niedermaier, N. Taccardi, P. S. Schulz, F. Maier, P. Wassersc-
heid, H.-P. Steinrꢀck, Angew. Chem. 2012, 124, 2664; Angew. Chem. Int.
Ed. 2012, 51, 2610.
[18] I. Niedermaier, C. Kolbeck, N. Taccardi, P. S. Schulz, J. Li, T. Drewello, P.
Wasserscheid, H.-P. Steinrꢀck, F. Maier, ChemPhysChem 2012, 13, 1725.
[19] M. Tariq, M. G. Freire, B. Saramago, J. A. P. Coutinho, J. N. C. Lopes,
L. P. N. Rebelo, Chem. Soc. Rev. 2012, 41, 829.
[20] R. F. Roberts, D. L. Allara, C. A. Pryde, D. N. E. Buchanan, N. D. Hobbins,
Surf. Interface Anal. 1980, 2, 5.
[21] P. Bonhote, A. P. Dias, N. Papageorgiou, K. Kalyanasundaram, M. Grꢃtzel,
Inorg. Chem. 1996, 35, 1168.
[22] F. F. C. Bazito, Y. Kawano, R. M. Torresi, Electrochim. Acta 2007, 52, 6427.
[23] Y. Yoshida, O. Baba, C. Larriba, G. Saito, J. Phys. Chem. B 2007, 111,
12204.
[24] P. W. Lohse, R. Bursing, T. Lenzer, K. Oum, J. Phys. Chem. B 2008, 112,
3048.
[30] Y. T. Wang, G. A. Voth, J. Am. Chem. Soc. 2005, 127, 12192.
[31] J. N. A. C. Lopes, A. A. H. Padua, J. Phys. Chem. B 2006, 110, 3330.
[32] R. M. Lynden-Bell, M. G. Del Popolo, T. G. A. Youngs, J. Kohanoff, C. G.
Hanke, J. B. Harper, C. C. Pinilla, Acc. Chem. Res. 2007, 40, 1138.
[33] K. Shimizu, M. F. C. Gomes, A. A. H. Padua, L. P. N. Rebelo, J. N. C. Lopes,
J. Phys. Chem. B 2009, 113, 9894.
[34] J. Lachwa, I. Bento, M. T. Duarte, J. N. C. Lopes, L. P. N. Rebelo, Chem.
Commun. 2006, 2445.
[35] J. D. Holbrey, W. M. Reichert, M. Nieuwenhuyzen, O. Sheppard, C. Harda-
cre, R. D. Rogers, Chem. Commun. 2003, 476.
[36] M. Deetlefs, C. Hardacre, M. Nieuwenhuyzen, O. Sheppard, A. K. Soper,
J. Phys. Chem. B 2005, 109, 1593.
[37] J. M. Gottfried, F. Maier, J. Rossa, D. Gerhard, P. S. Schulz, P. Wassersc-
heid, H.-P. Steinrꢀck, Z. Phys. Chem. (Muenchen Ger.) 2006, 220, 1439.
[38] C. Kolbeck, T. Cremer, K. R. J. Lovelock, N. Paape, P. S. Schulz, P. Was-
serscheid, F. Maier, H.-P. Steinrꢀck, J. Phys. Chem. B 2009, 113, 8682.
[39] K. R. J. Lovelock, E. F. Smith, A. Deyko, I. J. Villar-Garcia, P. Licence, R. G.
Jones, Chem. Commun. 2007, 4866.
[40] H. Hoffmann, J. Wurtz, J. Mol. Liq. 1997, 72, 191.
[41] M. Monduzzi, Curr. Opin. Colloid Interface Sci. 1998, 3, 467.
[42] N. Li, S. H. Zhang, L. Q. Zheng, J. P. Wu, X. W. Li, L. Yu, J. Phys. Chem. B
2008, 112, 12453.
[43] N. Li, S. H. Zhang, L. Q. Zheng, T. Inoue, Langmuir 2009, 25, 10473.
[44] P. F. Long, J. F. Chen, D. Wang, Z. Q. Hu, X. D. Gao, Z. R. Li, J. C. Hao, J.
Phys. Chem. B 2012, 116, 7669.
[45] T. L. Merrigan, E. D. Bates, S. C. Dorman, J. H. Davis, Chem. Commun.
2000, 2051.
[46] B. E. Smart, J. Fluorine Chem. 2001, 109, 3.
[47] G. D. Smith, O. Borodin, J. J. Magda, R. H. Boyd, Y. S. Wang, J. E. Bara, S.
Miller, D. L. Gin, R. D. Noble, Phys. Chem. Chem. Phys. 2010, 12, 7064.
[48] R. H. Boyd, L. Kesner, J. Chem. Phys. 1980, 72, 2179.
[49] F. Maier, T. Cremer, C. Kolbeck, K. R. J. Lovelock, N. Paape, P. S. Schulz, P.
Wasserscheid, H.-P. Steinrꢀck, Phys. Chem. Chem. Phys. 2010, 12, 1905.
[50] M. Lꢄsal, Z. Posel, P. Izak, Phys. Chem. Chem. Phys. 2012, 14, 5164.
[51] K. Nakajima, S. Oshima, M. Suzuki, K. Kimura, Surf. Sci. 2012, 606, 1693.
[52] O. Hçfft, S. Bahr, M. Himmerlich, S. Krischok, J. A. Schaefer, V. Kempter,
Langmuir 2006, 22, 7120.
[53] N. Taccardi, I. Niedermaier, F. Maier, H.-P. Steinrꢀck, P. Wasserscheid,
Chem. Eur. J. 2012, 18, 8288.
[25] M. G. Freire, P. J. Carvalho, A. M. Fernandes, I. M. Marrucho, A. J. Queima-
da, J. A. P. Coutinho, J. Colloid Interface Sci. 2007, 314, 621.
[26] C. Kolbeck, A. Deyko, T. Matsuda, F. T. U. Kohler, P. Wasserscheid, F.
Maier, H.-P. Steinrꢀck, ChemPhysChem 2013, 14, 3726.
Received: November 20, 2013
Published online on March 18, 2014
Chem. Eur. J. 2014, 20, 3954 – 3965
www.chemeurj.org
3965
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim