Surface spectroscopy of room-temperature ionic liquids on a platinum electrode: A sum frequency generation study

Article abstract of DOI:10.1021/jp048260g

Sum frequency generation vibrational spectroscopy (SFG) and cyclic voltammetry (CV) measurements have been conducted on the ionic liquid/platinum electrode system. The room-temperature ionic liquid investigated is 1-butyl-3-methylimidazolium [BMIM]⁺ with [PF₆]^- or [BF₄]^- anions. SFG spectra are taken in situ under potential control from 2750 to 3300 cm^-1 (C-H stretching region). Polarization-dependent SFG spectra are used to extract orientation information for the cation at the electrode surface. The SFG spectra indicate that the cation changes orientation as the electrode potential is changed within the double-layer region. The plane of the imidazolium ring tips from 35?° at positive surface charge to a??60?° from the surface normal at negative surface charge. Further, the orientation change is different for [BMIM][PF ₆] than for [BMIM][BF₄]. A model for ions at the surface is presented based on these spectroscopic and electrochemical measurements.

Full text of DOI:10.1021/jp048260g

J. Phys. Chem. B 2004, 108, 15133-15140
15133
Surface Spectroscopy of Room-temperature Ionic Liquids on a Platinum Electrode: A Sum
Frequency Generation Study
Selimar Rivera-Rubero and Steven Baldelli*
Department of Chemistry, UniVersity of Houston, Houston, Texas 77204
ReceiVed: April 21, 2004; In Final Form: June 1, 2004
Sum frequency generation vibrational spectroscopy (SFG) and cyclic voltammetry (CV) measurements have
been conducted on the ionic liquid/platinum electrode system. The room-temperature ionic liquid investigated
is 1-butyl-3-methylimidazolium [BMIM]+ with [PF6]- or [BF4]- anions. SFG spectra are taken in situ under
potential control from 2750 to 3300 cm-1 (C-H stretching region). Polarization-dependent SFG spectra are
used to extract orientation information for the cation at the electrode surface. The SFG spectra indicate that
the cation changes orientation as the electrode potential is changed within the double-layer region. The plane
of the imidazolium ring tips from 35° at positive surface charge to ∼60° from the surface normal at negative
surface charge. Further, the orientation change is different for [BMIM][PF6] than for [BMIM][BF4]. A model
for ions at the surface is presented based on these spectroscopic and electrochemical measurements.
Introduction
Room-temperature ionic liquids are liquids composed only
of ions and are similar to molten salts such as NaCl.1 However,
a weak interaction between the ions allows them to be liquid at
room temperature.1 Due to the high ionic conductivity, non-
volatility, low vapor pressure, thermal stability, hydrophobicity,
and wide electrochemical window that ionic liquids possess,
these compounds have become a novel solution to problems
encountered with organic solvents and these molecules are a
prospective solution to the limitations encountered in electro-
chemical systems.2,3 These physical properties can be varied
by selecting different combinations of ions.4-6 Since the
electrochemical window of the pure ionic liquids depends on
the electrochemical stability of the cation and/or anion, under-
standing the ion behavior at the electrode surface leads to
improvement and implementation of the ionic liquid to the
desired system.7
Figure 1. Structure of 1-butyl-3-methylimidazolium cation, [BMIM]+,
with numbering scheme and pseudo C2 symmetry axis of the imida-
zolium ring.
Background. The most extensive work on ionic liquid
structure near the metal electrode surface is from results in
molten salt systems;21 specifically, those of liquid mercury,
magnesium,22 and liquid lead electrodes in contact with metal
halide salts.23,24 Experimentally, the interfacial structure is
inferred from electrocapillary and capacitance measurements.
The results must be interpreted in terms of a model (an
equivalent circuit) that can be quite complicated.22 The ions in
a molten salt near an electrode surface generally exhibit a
layered (alternating ion) structure. Surprisingly, the potential
of zero charge (PZC) of the metal/molten salt system is similar
to the corresponding ions in an aqueous solution.23 To our
knowledge, there are as yet no detailed electrochemical imped-
ance spectra (EIS) on room-temperature ionic liquids based on
imidazolium cations from which the interfacial structure is
deduced.
A molecular-level description on the arrangement of ions at
the metal/ionic liquid interface is needed. The problem arises
of how to envision double-layer structure and the potential of
zero charge (PZC) in a pure liquid electrolyte. By far, most
experimental and theoretical work on the double layer structure
is from aqueous systems using models such as Gouy-Chapman-
Sterns.25-27 This model assumes ions are surrounded by
solvation shells (of water) and have a nonuniform concentration
distribution at the electrode surface that forms the diffuse double
layer. This model is barely applicable to concentrated aqueous
Ionic liquids have been used in different systems such as
electrochemical cells,2,3,8,9 fuel cells,10,11 solar cells,12,13 liquid-
liquid extraction systems,14,15 and biphasic catalysis pro-
cesses.16,17 These systems depend on the interactions at the
interface for their performance; unfortunately, the interfacial
properties are not well characterized. To enhance the properties
of the ionic liquids and expand their applications, it is necessary
to understand their behavior at the interface. The electrochemical
behavior of haloaluminate ionic liquids has been widely studied,
but since their applications are limited due to their moisture
sensitivity the attention has been recently focused on more-
stable ionic liquids such as those containing imidazolium-based
cations (Figure 1) and anions such as hexafluorophosphate,
tetrafluoroborate, and others.7,18-20
This paper reports the results from sum frequency generation
vibrational spectroscopy of [BMIM][BF4] and [BMIM][PF6] on
a platinum electrode under potential control. These experiments
suggest that the imidazolium ring (Figure 1) reorients at the
surface in response to the surface charge with the ring orienting
more parallel to the surface at negative surface charge.
10.1021/jp048260g CCC: $27.50 © 2004 American Chemical Society
Published on Web 09/04/2004

15134 J. Phys. Chem. B, Vol. 108, No. 39, 2004
Rivera-Rubero and Baldelli
electrolytic solutions, and is not valid in the ionic liquids where
there is no solvent. Further, the results from the molten salts
(as stated above) are likely to be only an approximation to the
room-temperature ionic liquids discussed in this paper, since
these are molecular ions and their chemical structure probably
is important to the bonding and orientation at the surface. The
first assumption is that ions are directly adsorbed to the surface,
as in a Helmholtz layer. Also, several layers of ions may extend
into the liquid to compensate for the excess surface charge. This
model makes intuitive sense since the ions in the ionic liquid
are relatively ordered due to high Coulombic forces. According
to the Gouy-Chapmann-Stern model, the PZC is considered
to be the state where the surface concentration of ions is the
same as the bulk concentration.21,27,28 However, in a pure ionic
liquid with no solvent the ion concentration is essentially the
same throughout the system; only in the inner Helmholtz layer
is there a possible difference. There are theoretical treatments
of capacitance data that show it is possible to have surface excess
of ions at an electrode interface in a molten salt.27
with the Raman polarizability and the IR dipole transition, which
are averaged over the molecule orientation indicated by the
brackets, . N indicates the number of molecules contributing
to the SFG signal, ωIR and ωq refer to the frequency of the
incoming IR and the normal mode, respectively, and Γq is the
damping constant for the qth vibrational mode.
Experimental Section
Sample Synthesis and Preparation. The synthesis and char-
acterization of 1-butyl-3-methylimidazolium tetrafluoroborate,
[BMIM][BF4], and 1-butyl-3-methylimidazolium hexafluoro-
phosphate, [BMIM][PF6], is similar to that published by Gra¨tzel
et al.12 1-methylimidazole is mixed with 1-chlorobutane in a
1:1.1 molar ratio and refluxed under nitrogen at 65 °C for 52
h. The 1-butyl-3-methylimidazolium chloride, [BMIM][Cl], is
washed three times with ethyl acetate and dried under vacuum
at a temperature of ∼70 °C. [BMIM][Cl] is then mixed with
the acid of the desired anion in a 1:1.1 molar ratio and stirred
for 24 h. The final [BMIM][PF6] compound is simply washed
with water until no chloride is detected in the water. However,
for the [BMIM][BF4], it is first cooled in an ice bath and an
equal volume of cold dichloromethane is added. Then, the
compound is washed with ice-cold water until no chloride is
detected in the water.
After [BMIM][PF6] and [BMIM][BF4] are dried under
vacuum at a temperature of ∼70 °C, the chloride concentration
is detected directly in the ionic liquid using an ion-selective
electrode; both compounds have a chloride concentration e 18
ppm.
All chemicals are from Aldrich and used without further
purification. The water is deionized in a Millipore A10 system
with a resistivity of >18 MΩ‚cm and an organic concentration
of < 3 ppb. H NMR, LC-MS, and FT-IR are used for
characterization, and an ion selective electrode is used for the
chloride determination. Cyclic voltammetry is used to estimate
the water concentration (Figure 2).
The working electrode is made of polycrystalline platinum
10 mm in thickness by 0.25 in. in diameter and is prepared by
flame annealing in a hydrogen-air flame for ∼10 min. The
electrode is cooled in a glass chamber flowing with argon gas
for 10 min. Immediately upon removal from the cooling
chamber, a drop of ionic liquid is placed on the surface of the
Pt electrode. The electrode is then introduced into the SFG
electrochemistry cell. The procedure is analogous to the
preparation of well-characterized electrodes used in aqueous
electrochemistry by Clavilier,45 where it produces a clean surface
as determined by CV and UHV techniques.
The SFG-electrochemistry cell is an important development
to accomplishing these experiments. The cell is vacuum-tight
to >2 × 10-5 torr, which is possible because the Pt working
electrode is held in the cell with a 1/4-in Teflon Swage fitting
attached to the Kel-F shaft, (Figure 3). The entire cell is
fabricated only of inert polymers (Kalrez O-rings, Teflon, or
Kel-F parts) and glass, permitting aggressive cleaning in 50/50
nitric acid/sulfuric acid mixtures. The cell is thoroughly rinsed
in water from a Millipore A10 water system. The counter
electrode is a Pt wire and the reference is Ag/AgBF4 or Ag/
AgPF6 that has a potential of ∼ -200 mV vs NHE (Normal
Hydrogen Electrode).
Bulk Ionic Liquid Structure. Measurements on the bulk
structure of an ionic liquid are useful to understand the ion
interactions at the electrode surface. From the crystallographic
data of 1-alkyl-3-methyl imidazolium salts containing hexafluo-
rophosphate29 and tetrafluoroborate,30 it is known that the anion
is centered on the cation ring. The peak shift of the aromatic
hydrogen in NMR and IR spectroscopy has been used to study
the interaction of ions in the ionic liquid. The results indicate
that there is a slight preference for the anion toward the C(2)-H
bond, however, the anion interacts with the other ring protons
and thus is mostly centered on the imidazolium ring.31-34
Computer modeling of the bulk of the ionic liquid also indicate
this relative positioning of the ions.35,36 The ion attraction is
dominated by an electrostatic interaction but is also influenced
by hydrogen bonding and van der Waals forces, which are
-
determined by the anion identity. For PF6 and BF4-, these
interactions are similar though the IR and NMR results indicate
that the BF4- interaction is slightly stronger than for PF6-. This
difference is enough to dramatically affect the physical proper-
ties such as viscosity, water miscibility, and melting point.3,9,12,37
To probe the molecular-level information of the ions at the
electrode/ionic liquid interface, a surface-specific spectroscopic
technique is used. SFG is a second order nonlinear spectroscopy
technique sensitive to molecules in a noncentrosymmetric
environment. Most molecules in the liquid state are considered
to be in an isotropic environment; as a consequence SFG will
not be generated in the bulk, but only from the liquid/electrode
interface where the centrosymmetry is broken. The theory of
SFG will not be covered in detail; instead, the reader is referred
to different reviews.38-44 The intensity of the SFG signal is
proportional to the square of the induced polarization:
ESF
P
(2) ) |ø(2):EVisEIR|
(1)
(2)
ø(2) ) ø(2) + ø(2)
res
nr
N â(2)
(ωIR - ωq + iΓq)
(2)
res
ø
)
(3)
∑
where E refers to the electric field of the visible and infrared
Once the electrode with the ionic liquid drop on it is
introduced into the cell, the cell is evacuated to its base pressure
for ∼4 h at 60-70 °C. Next, a vessel containing the ionic liquid
is pressurized to just over 1 atm with Ar gas. A stopcock
connecting the vessel to the SFG cell is opened transferring
(2)
incoming beams and ø
is the second-order susceptibility
tensor that relates the interface response to the input light fields.
The ønr(2) arises from the nonresonant background of the surface
(2)
and the øres contains the molecular hyperpolarizability, â(2),

Sum Frequency Generation Study
J. Phys. Chem. B, Vol. 108, No. 39, 2004 15135
Figure 2. Cyclic voltammetry of neat (A) [BMIM][BF4] and (B) [BMIM][PF6] on Pt electrode. Scan rate ) 100 mV/sec. Vertical red lines
indicate the potential limit of the SFG experiment.
TABLE 1: [BMIM]BF4]. Amplitude Values for Peaks in the
SFG Spectrum of 1-butyl-3-methylimidazolium Cation on a
Pt Electrode at -800 mV and +1000 mV vs Ag/AgBF4
Reference. Number in Parentheses is the Percent Error of
the Amplitude
-800 mV
ppp
+1000 mV
ppp ssp
ssp
CH2(sym)
CH3(sym)
CH2(asym)
CH3(asym)
CH3(sym)(N3)
C(2)-H
2.2 (177)
1.0 (286)
-
3.8 (15)
1.6 (31)
-
-
7.8 (3)
3.8 (12)
8.6 (8)
-
0.9 (35)
3.7 (12)
5.7 (15)
5.1 (8)
-
-
3.7 (36)
4.2 (36)
0.9 (7)
1.3 (42)
-
-
-
1.3 (75)
2.8 (30)
H-C(4)C(5)-H 11 (5)
NR -1.1
3.7 (10)
-0.25
-0.07
-1.2
SFG signal exhibits a potential dependence, which only occurs
at the Pt/ionic liquid interface.
Figure 3. Thin-layer electrochemical cell for SFG studies. The reflected
Another important feature of this optical setup is the reference
channel in the SFG spectrometer. The SFG signal is proportional
to the infrared intensity and since the infrared intensity is not
constant as a function of wavelength (due to the OPG/OPA
source and absorbance in the beam path) the SFG spectra must
be corrected for this. To ensure that the features in the SFG
spectrum are resonances of the surface molecules and not
modulations in the infrared intensity due to absorbance in the
ionic liquid, the following reference procedure is used. Infrared
light that is reflected from the Pt electrode is mixed in a z-cut
quartz with 532-nm light and produces a reference SF signal
that is only affected by the infrared fluctuations. The SFG signal
is adjusted by this reference signal to compensate for the changes
in infrared intensity. Since the detection system for this reference
setup is identical to the one for the SFG signal, it is highly
sensitive. In practice, it has been experimentally determined that
the spectra can be compensated by this method if the infrared
absorbance/fluctuations are less than 45%. In typical experi-
ments, the infrared intensity change (due to absorbance) is not
more than 7-8%. A reference is collected simultaneously with
each SFG spectrum.
IR is sent to the reference channel for normalization of the SFG signal.
the ionic liquid into the SFG cell without contact to air. The
cell can maintain the ionic liquid in an anhydrous state for
approximately 10 h, as determined by cyclic voltammetry. The
cell allows for control of potential during the SFG experiment
as well as for the measurement of the CV.
Optical Setup. An Ekspla 20 Hz picosecond Nd:YAG laser
is used to pump the optical parametric generation/amplification
system, OPG/OPA (LaserVision). The OPG/OPA is composed
of a set of KTP/KTA crystals that are used to generate the
second harmonic (532 nm) and the tunable infrared beams
(2000-4000 cm-1). The visible beam is at an angle from the
surface normal of 33° and the infrared was at an angle of 38°
at the ionic liquid/Pt interface with energy densities of 30 mJ/
cm2 and 20 mJ/cm2, respectively. Polarization control of the
visible beams is accomplished by a λ/2 waveplate antireflection
coated for 532 nm (CVI). The tunable infrared polarization was
changed using a zero-order phase retardation plate (λ/2) supplied
by ALPHALAS GmbH. Glan-Laser polarizers (CVI) are used
for polarization selection in these experiments.
Optical access to the electrode surface is possible in the thin-
layer cell design. The laser beams pass through an infrared
quartz window, through a thin layer of ionic liquid (∼10 µm
thick), then travel to the Pt electrode. The generated SF signal
is reflected out of the cell to the detection system described
below. To verify that the SF signal is from the Pt/ionic liquid
interface and not from the infrared quartz/ionic liquid interface,
two tests are performed. First, the Pt electrode is backed away
from the window, by ∼0.5 mm, so the infrared is completely
adsorbed by the ionic liquid. Under these conditions there is
no SFG detected using ssp or ppp polarizations. Second, the
Data Collection and Analysis. The data are acquired while
the infrared energy is continuously scanned from 2750 to 3300
cm-1 at 1 cm-1/sec. Each data point in a spectrum is an average
of 20 shots. The spectra presented are an average of five spectra
with error bars representing the standard deviation between the
spectra. The spectra are curve fitted according to eqs 1-3 using
nonlinear fitting functions in Origin with instrumental setting
for the error bars. As a result, the error in the peak amplitude
(Table 1 and Table 2) is not only the error in the fit; it also
accounts for the scatter in the data. The baselines in the spectra
are at zero and are determined by blocking the infrared to collect

15136 J. Phys. Chem. B, Vol. 108, No. 39, 2004
Rivera-Rubero and Baldelli
imidazolium ring are at ∼3050 cm-1, 3150 cm-1, and ∼3190
cm-1. They are due to the interaction peak, the C(2)-H or
H-C(4)C(5)-H(asym) peak, and the H-C(4)C(5)-H(sym),
respectively.32,46,47 Two other peaks at 2845 and 2870 cm-1 are
assigned to the CH2(sym) and the CH3(sym) of the butyl
chain.48-50 The CH3(sym) of the methyl group at N(3) is at
TABLE 2: [BMIM][PF6]. Amplitude Values for Peaks in the
SFG Spectrum of 1-butyl-3-methylimidazolium Cation on a
Pt Electrode at -500 mV and +1000 mV vs Ag/AgPF6
Reference. Number in Parentheses is the Percent Error of
the Amplitude
-500 mV
ppp
+1000 mV
ppp
∼2935 cm-1 46,47
.
The peak at 3050 cm-1 is called the interaction
ssp
ssp
peak. It is due to hydrogen-bonding interaction of the aromatic
C-H bonds with the anion32 or possibly the electrode.
In the ssp spectrum, there are four peaks observed. At 2875
cm-1 and ∼2955 cm-1 are the CH3(sym) and CH3(asym) of
the butyl chain. The peaks at 2920 and 3195 cm-1 are from the
CH2(asym) and the H-C(4)C(5)-H(sym), respectively.
The peak intensities show a potential dependence over the
range from -800 mV to +1000 mV. The CH3(sym) disappears
at low potential and reappears at high potential; the H-C(4)C-
(5)-H(sym) has the same behavior. In the ppp spectrum, peaks
change intensity in a similar way.
CH3(sym)
CH2 (asym)
N(3)CH3(sym)
CH3(asym)
C(2)-H
H-C(4)C(5)-H 13 (10)
NR -2.1
3.1 (32)
0.004 (5000)
3.5 (45)
0.018 (722)
15 (13)
5.5 (7)
-
3.8 (9)
2.5 (31)
2.0 (70)
1.9 (75)
0.3 (33)
-
1.6 (11)
1.3 (18) -0.86 (100)
0.29 (48)
-
27 (10)
23 (3)
-1.5
-
2.8 (9)
-0.71
6.2 (4)
-0.44
a background signal level. The detection system contains a
photomultiplier tube (Hamamatsu R3788) attached to a Jarrell-
Ash monochromator. Detection electronics consist of an SRS
250 boxcar average and an SR245 computer interface board.
These components are duplicated for the reference channel. The
tuning of the infrared OPG/OPA (LaserVision), the monochro-
mator, and the acquisition of SFG and reference signals are
achieved by a LabVIEW program.
For [BMIM][PF6] in the ppp spectrum (Figure 6), six
resonances are observed. Peaks at 2870 cm-1 and ∼2935 cm-1
are assigned to the CH3(sym) of the butyl chain48-50 and the
N(3)CH3(sym), while the peak for the CH2(asym) is at ∼2910
cm-1 The peak position, as well as the intensity, for the
.
Results
imidazolium ring showed a potential dependence. For the
H-C(4)C(5)-H(sym), the peak shifted from 3200 cm-1 to
∼3170 cm-1 when the potential changed from +1000 mV to
-500 mV. The C(2)-H vibration is clearly observed at 3130
cm-1 for an applied potential of -500 mV but seems to
drastically decrease its intensity and shifts to ∼3100 cm-1 when
the potential changes to +1000 mV. The interaction peak is
The SFG spectra of [BMIM][BF4] and [BMIM][PF6] for ppp
and ssp polarization are shown below in Figures 4, 5, and 6.
Qualitatively, the spectra are relatively complicated. There are
several resonances observed, as well as an intense nonresonant
background. Therefore, only the major peaks in the spectrum
are assigned and used in the analysis. The results exhibit several
vibrations in the C-H stretching region that are polarization
and potential dependent for both ionic liquids.
also present in the spectra around 3030 cm-1
.
The ssp spectrum has four peaks. At 2870 cm-1 is the CH3-
(sym), while the peaks at 2915 and 3190 cm-1 are from the
CH2(asym) and the H-C(4)C(5)-H(sym), respectively (Figure
For [BMIM][BF4] in the ppp spectrum, there are six
resonances observed at +1000 mV. The three from the
Figure 4. SFG spectra of [BMIM][BF4] at the Pt electrode at different applied potentials. (A) -800 mV. (B) 0 mV. (C) +500 mV. (D) +1000 mV.
Reference potential is Ag/AgBF4. Polarization is ppp. Line is a fit to the line shape equation.

Sum Frequency Generation Study
J. Phys. Chem. B, Vol. 108, No. 39, 2004 15137
Figure 5. SFG spectra of [BMIM][BF4] at the Pt electrode at different applied potentials. (A) -800 mV. (B) 0 mV. (C) +500 mV. (D) +1000 mV.
Reference potential is Ag/AgBF4. Polarization is ssp. Line is a fit to the line shape equation.
Figure 6. SFG spectra of [BMIM][PF6] at the Pt electrode at different applied potentials. (A) -500 mV. (B) 1000 mV. A and B are for ssp
polarization. (C) -500 mV. (D) +1000 mV. C and D are for ppp polarization. Reference potential is Ag/AgPF6. Line is a fit to the line shape
equation.
6). The resonance at 2965 cm-1 is due to the CH3(asym) on
the butyl chain.48-50
The cyclic voltammetry of ionic liquids [BMIM][BF4] and
[BMIM][PF6] are shown in Figure 2. These are similar to the

15138 J. Phys. Chem. B, Vol. 108, No. 39, 2004
Rivera-Rubero and Baldelli
Figure 7. Simulation of SFG signal intensity ratio ppp/ssp for the H-C(4)C(5)-H symmetric stretch for (A) [BMIM][BF4], (B) [BMIM][PF6].
The filled squares (9) and triangles (2) denote the experimentally determined ratio.
Figure 8. Orientation analysis of the terminal CH3 group on the butyl chain of [BMIM]+. (A) [BMIM][BF4] b) [BMIM][PF6]. Solid line is the
theoretical curve. Data points (9, 2) are experimentally determined ratio. Note: error bars for -800 mV in A extend from 0 to +2.0.
CV reported by Xiao and Johnson.8 and Conboy et al.51 The
potential is kept well below the oxidation and reduction limits
of +2.9 to -2.0 V vs Ag/AgBF4, respectively. Cyclic voltam-
metry of the ionic liquids indicate there is no redox processes
at these potentials and the ionic liquids are relatively free of
water and Cl- (see Experimental section and Figure 2).
+1000 mV. However, the twist angle changes dramatically from
∼0-35° at -500 mV to about 65-90° at +1000 mV (Figure
7B).
Although the most reliable orientation information of imi-
dazolium is from the H-C(4)C(5)-H (sym) mode, some
qualitative assessment is made by the other modes. Quantitative
analysis of the alkyl CH vibrations is difficult due to the
presence of nonresonant background, the many overlapping CH
bands, and the interference between these two.
Discussion
Qualitatively, the peaks in the SFG spectra indicate that the
imidazolium ring is oriented with the C2 axis (Figure 1) along
the surface normal. These results seem consistent with recent
FTIR experiments.52
At +1000 mV the CH3(sym) (2875 cm-1) is strong but is
absent at the other potentials while the CH2(asym) (2915 cm-1
)
and CH3(asym) (2955 cm-1) are intense at positive potential
but weaker at negative potential. Overall, these observations
indicate a higher degree of ordering along the surface normal
at more positive potentials for the alkyl chains. Formal orienta-
tion analysis supports this interpretation as presented in (Figure
8). The methyl group for [BMIM][BF4] indicates that the chain
has a more perpendicular orientation at positive potentials. Thus,
at + 1000 mV the tilt angle is < 40° from normal while at
-800 mV it extends its range from 0 to 65°. Although this
orientation preference is only slight, the increased intensity
suggests an increase in polar orientation on the methyl group.
In this analysis the methyl group is assigned C3V symmetry and
free rotation around the C-C axis.
The orientation analysis is performed by the same method
as that used by Hirose et al.41-44 Simulations of peak intensity
vs tilt angle (of the C2 axis) are plotted for several twist angles
for [BMIM][BF4] in Figure 7A. Also plotted are the experi-
mentally determined intensity ratio for the H-C(4)C(5)-H
(3190 cm-1) of the imidazolium ring in ppp/ssp spectra. The
intersection of the experimental ratio with the calculated ratio
indicates the possible tilt and twist angles. At +1000 mV, the
C2 axis is tilted ∼45° from the surface normal and increases to
∼60° at -800 mV. The twist angle is relatively restricted to
near 0°; this is reasonable since the side chains, butyl and
methyl, hinder the twist. Also, there does not appear to be a
H-C(4)C(5)-H antisymmetric stretch so the H-C(4)C(5)-H
plane orientation is mostly a function of tilt angle θ.
For [BMIM][PF6], the analysis of the methyl group suggests
the butyl chain is more parallel to the surface > 60° at +1000
mV, while it is less than 40° at -500 mV. This is interpreted
as more degrees of freedom for the alkyl chain as the
imidazolium ring twists from parallel to perpendicular at the
surface.
The same analysis for [BMIM][PF6], however, indicates a
different potential-dependent orientation. For this ionic liquid
the tilt orientation is relatively constant from -500 mV to
+1000 mV and is about 12-26° at -500 mV and 23-34° at

Sum Frequency Generation Study
J. Phys. Chem. B, Vol. 108, No. 39, 2004 15139
This interpretation assumes a delta function distribution and
that the oriented cations do not form more than one monolayer,
that is, a Helmholtz layer at the surface. From previous
electrochemical studies using capacitance and electrocapillarity,
this model is reasonable.53
Acknowledgment. We gratefully appreciate funding for this
project provided by The Welch Foundation (E1531) and The
Petroleum Research Fund (37767.01-G5).
References and Notes
To estimate the coverage of imidazolium on the electrode,
cyclic voltammetry is used. The PZC of 1-ethyl-3-methyl-
imidazolium-BF4 was determined by Koch et al.,53 and found
to be -0.5 V vs Ag/AgCl (∼Ag/AgBF4). By integrating the
double-layer current from PZC to a given potential in the Pt
electrode, the charge accumulation is estimated. For [BMIM]-
[BF4], integration from -0.5 V to 1 V is approximately 2.8 ×
10-4 C/cm2, assuming an atomically flat surface (otherwise a
roughness factor of 2-3 should be applied).25 This charge value
is in agreement with those calculated from differential capaci-
tance. The capacitance at the Hg/[EMIM][BF4] is on the order
of 1 × 10-5 F/cm,2,53 which is a surface charge of about 1 ×
10-4 C, or 0.5 monolayer of electric charge (∼2 × 10-4
C/monolayer). Using this estimate on the amount of surface
charge it seems reasonable to envision that the ions adsorb to
the surface as a single layer and multilayers are not present at
the surface. This is consistent with a Helmholtz model of the
interface.
(1) Seddon, K. R. J. Chem. Technol. Biotech. 1997, 68, 351.
(2) Fuller, J.; Carlin, R. T.; Osteryoung, R. A. J. Electrochem. Soc.
1997, 144, 3881.
(3) Quinn, B. M.; Ding, Z.; Moulton, R.; Bard, A. J. Langmuir 2002,
18, 1734.
(4) Welton, T. Chem. ReV. 1999, 99, 2071.
(5) Suarez, P. A.; Einloft, S.; Dullius, J. E.; Souza, R. F.; Dupont, J.
J. Chim. Phys. 1998, 95, 1626.
(6) Holbrey, J. D.; Seddon, K. R. Clean Products and Processes 1999,
1, 223.
(7) Trulove, P. C.; Mantz, R. A. Electrochemical Properties of Ionic
Liquids. In Ionic Liquids in Synthesis; Welton, T., Wasserscheld, P., Eds.;
Wiley-VCH: Morlenbach, 2003; p 368.
(8) Xiao, L.; Johnson, K. E. J. Electrochem. Soc. 2003, 150, E307.
(9) Suarez, P. A.; Selbach, V. M.; Dullius, J. E.; Einloft, S.; Pianicki,
C. M.; Azambuja, D. S.; deSouza, R. F.; Dupont, J. Electrochim. Acta 1997,
42, 2533.
(10) Doyle, M.; Choi, S. K.; Proulx, G. J. Electrochem. Soc. 2000, 147,
34.
(11) Hagiwara, R.; Hirashige, T.; Tsuda, T.; Ito, Y. J. Electrochem. Soc.
2002, 149, D1.
(12) Bonhote, P.; Dias, A.; Papageorgiou, N.; Kalyanasundaram, K.;
Gra¨tzel, M. Inorg. Chem. 1996, 35, 1168.
The SFG data suggest that at high positive potential the cation,
[BMIM]+, is tipped with the imidazolium ring along the surface
normal and the ring becomes more parallel to the surface at
more negative potentials. This is reasonable since the PZC of
this system is near -500 mV;53 thus, at -800 mV the surface
has a negative charge. Previous results have shown that the
cation charge is mostly centered on the aromatic ring;54,55 thus,
this would be the lowest energy situation considering only the
electrostatic energy. To screen the charge, the ring orients
parallel to the surface. At positive surface charge the anion will
move into the inner Helmholtz layer for charge compensation
and the cation will then occupy the outer Helmholtz layer. This
model suggests that the cation is tipped along the surface normal
to make room for the anion to approach the surface at potentials
positive of PZC. Further, the effect is different for [BF4]-
compared to [PF6]- and is likely due to its size and/or charge
density.
(13) Papageorgiou, N.; Anthanassov, Y.; Armand, M.; Bonhte, P.;
Pettersson, H.; Azam, A.; Gratzel, M. J. Electrochem. Soc. 1996, 143, 3009.
(14) Rogers, R. D.; Huddleston, J. G.; Williauer, H. D.; Swatloski, R.
P.; Visser, A. E. Chem. Commun. 1998, 1765.
(15) Abraham, M. H.; Zissimos, A. M.; Huddleson, J. G.; Willauer, H.
D.; Rogers, R. D.; Acree, W. E. Ind. Eng. Chem. Res. 2003, 42, 413.
(16) Dupont, J.; deSouza, R. F.; Suarez, P. A. Chem. ReV. 2002.
(17) Suarez, P. A.; Dullius, J. E.; Einloft, S.; DeSouza, R. F.; Dupont,
J. Polyhedron 1996, 15, 1217.
(18) Wilkes, J. S.; Levisky, J. A.; Wilson, R. A.; Hussey, C. L. Inorg.
Chem. 1982, 21, 1263.
(19) Wilkes, J. S.; Zaworotko, M. J. J. Chem. Soc., Chem. Commun.
1992, 965.
(20) Ionic Liquids in Synthesis; Welton, T.; Wasserscheid, P., Eds.;
Wiley-VCH: Morlenbach, 2003, p 364.
(21) Bockris, J. O.; Khan, S. U. M. Surface Electrochemistry, 1st ed.;
Plenum Press: New York, 1993.
(22) Kisza, A. J. Electroanal. Chem. 2002, 534, 99.
(23) Graves, A. D. Electroanal. Chem. 1970, 25, 349.
(24) Graves, A. D.; Inman, D. Electroanal. Chem. 1970, 25, 357.
(25) Bard, A. J.; Faulkner, L. R. Electrochemical Methods, 2nd ed.; John
Wiley and Sons: New York, 2001.
There appears to be two slightly differing mechanisms for
cation reorientation at the electrode surface. In the case of
negative surface charge, the imidazolium ring is relatively
parallel to the electrode surface, presumably to screen the surface
charge. However, when the surface charge becomes positive
the ring tips from the surface to allow space for the anion to
enter the Helmholtz layer and shield the electrode charge. For
the [BMIM][BF4] and [BMIM][PF6] ionic liquids this occurs
in two different ways. In the [BMIM][BF4] case, the twist angle
of the ring is not greatly different at the positive and negative
potentials and only the tilt angle changes. However, for [BMIM]-
[PF6] the tilt angle is not changing as a function of potential,
but instead is changing with respect to the twist angle. These
results suggest that [BMIM]+ changes its polar tilt angle (θ) to
allow the anion access to the surface for [BF4]- while it changes
its twist angle to allow [PF6]- access to the surface.
(26) Schmickler, W. Interfacial Electrochemistry, 1st ed.; Oxford
University Press: New York, 1996.
(27) Bockris, J. O.; Reddy, A. K. N. Modern Electrochemistry: Ionics,
2nd ed.; Plenum Press: New York, 1998; Vol. 1.
(28) Devanathan, M. A.; Tilak, B. V. Chem. ReV. 1965, 65, 635.
(29) Gordon, C. M.; Holbrey, J. D.; Kennedy, A. R.; Seddon, K. R. J.
Mater. Chem. 1998, 8, 2627.
(30) Holbrey, J. D.; Seddon, K. R. J. Chem. Soc., Dalton Trans. 1999,
13, 2133.
(31) Carmicheal, A. J.; Hardacre, C.; Holbrey, J. D.; Seddon, K. R.;
Nieuwenhuyzen, M. Proc. Electrochem. Soc. 2000, 99-41, 209.
(32) Dieter, K. M.; Dymek, C. J.; Heimer, N. E.; Rovang, J. W.; Wilkes,
J. S. J. Am. Chem. Soc. 1988, 110, 2722.
(33) Fannin, A. A.; King, L. A.; Levisky, J. A.; Wilkes, J. S. J. Phys.
Chem. 1984, 88, 2609.
(34) Haung, J. F.; Chen, P. Y.; Sun, I. W.; Wang, S. P. Inorg. Chim.
Acta 2001, 320, 7.
(35) Urahata, S. M.; Ribeiro, M. C. J. Chem. Phys. 2004, 120, 1855.
(36) Morrow, T. I.; Maginn, E. J. J. Phys. Chem. B. 2002, 106, 12807.
(37) Cammarata, L.; Kazarian, S. G.; Salter, P. A.; Welton, T. Phys.
Chem. Chem. Phys. 2001, 3, 5192.
(38) Shultz, M. J.; Baldelli, S.; Schnitzer, C.; Simonelli, D. J. Phys.
Chem. B 2002, 106, 5313.
Conclusion
Sum frequency generation has been used to study the
adsorption and orientation of ions in a room-temperature ionic
liquid at the platinum electrode surface. The results indicate
that the cation has a preferred orientation that changes as a
function of applied potential. At relatively negative surface
charge the imidazolium ring is nearly parallel to the surface,
while it tilts along the surface normal at more positive potential.
(39) Buck, M.; Himmelhaus, M. J. Vac. Sci. Technol., A 2001, 19, 2717.
(40) Huang, J. Y.; Shen, Y. R. Sum Frequency Generation as a Surface
Probe. In Laser Spectroscopy and Photochemistry on Metal Surfaces; Dai,
H. L., Ho, W., Eds.; World Scientific: Singapore, 1995.
(41) Hirose, C.; Yamamoto, H.; Akamatsu, N.; Domen, K. J. Phys.
Chem. 1993, 97, 10064.
(42) Hirose, C.; Akamatsu, N.; Domen, K. Appl. Spectrosc. 1992, 46,
1051.

15140 J. Phys. Chem. B, Vol. 108, No. 39, 2004
Rivera-Rubero and Baldelli
(43) Hirose, C.; Akamatsu, N.; Domen, K. J. Chem. Phys. 1992, 96,
997.
(44) Akamatsu, N.; Domen, K.; Hirose, C. J. Phys. Chem. 1993, 97,
10070.
(50) Snyder, R. G.; Strauss, H. L.; Elliger, C. A. J. Phys. Chem. 1982,
86, 5145.
(51) Fitchett, B. D.; Knepp, T. N.; Conboy, J. C. J. Electrochem. Soc.
2004, in press.
(45) Clavilier, J.; Faure, R.; Guinet, G.; Durand, R. J. Electroanal. Chem.
1980, 107, 205.
(46) Carter, D. A.; Pemberton, J. E. J. Raman Spectrosc. 1997, 28, 939.
(47) Carter, D. A.; Pemberton, J. E.; Woelfel, K. J. J. Phys. Chem. B
1998, 102, 9870.
(48) MacPhail, R. A.; Strauss, H. L.; Snyder, R. G.; Elliger, C. A. J.
Phys. Chem. 1984, 88, 334.
(49) Snyder, R. G. J. Chem. Phys. 1965, 42, 1744.
(52) Nanbu, N.; Sasaki, Y.; Kitamura, F. Electrochem. Commun. 2003,
5, 383.
(53) Nanjundiah, C.; McDevitt, S. F.; Koch, V. R. J. Electrochem. Soc.
1997, 144, 3392.
(54) Hardacre, C.; Holbrey, J. D.; McMath, S. E.; Bowron, D. T.; Soper,
A. K. J. Chem. Phys. 2003, 118, 273.
(55) Hanke, C. G.; Price, S. L.; Lyndon-Bell, R. M. Mol. Phys. 2001,
99, 801.