Molecular orientation of CN adsorbed on Pd(110)

Article abstract of DOI:10.1063/1.1574794

Information on azimuthal molecular orientation as well as on the tilt of the C-N molecular axis on the Pd(110) surface was derived by combining angle-dependent NEXAFS and full-solid-angle PED measurements. NEXAFS data and PED yield independent and consistent conclusions, indicating a nearly parallel CN adsorption on Pd(110), with preferential azimuthal orientation of the C-N molecular axis along the [001] direction. Furthermore, the presence of a 25° tilt of the C-N axis from the surface plane was found with the C atoms closer to the surface than N ones. While the near-parallel bonding geometry of CN found by previous ARUPS data was confirmed, the azimuthal orientation of the bond axis contradicted the conclusions based on the interpretation of photoemission data.

Full text of DOI:10.1063/1.1574794

Molecular orientation of CN adsorbed on Pd(110)
F. Bondino, E. Vesselli, A. Baraldi, G. Comelli, A. Verdini, A. Cossaro, L. Floreano, and A. Morgante
Citation: The Journal of Chemical Physics 118, 10735 (2003); doi: 10.1063/1.1574794
View online: http://dx.doi.org/10.1063/1.1574794
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/118/23?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Synthesis and characterization of oriented graphitelike B–C–N hybrid
J. Appl. Phys. 99, 084902 (2006); 10.1063/1.2189008
C 70 adsorbed on Cu(111): Metallic character and molecular orientation
J. Chem. Phys. 116, 7685 (2002); 10.1063/1.1467346
Functionalization of silicon step arrays II: Molecular orientation of alkanes and DNA
J. Appl. Phys. 90, 3291 (2001); 10.1063/1.1397297
Scanning tunneling microscopy study of the molecular arrangement of meta- and para-xylene on Pd(111)
J. Vac. Sci. Technol. A 19, 1993 (2001); 10.1116/1.1366698
Orientation and symmetry of ethylene on Pd(110): A combined HREELS and NEXAFS study
J. Chem. Phys. 112, 5948 (2000); 10.1063/1.481168
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sat, 20 Dec 2014 01:01:41

JOURNAL OF CHEMICAL PHYSICS
VOLUME 118, NUMBER 23
15 JUNE 2003
Molecular orientation of CN adsorbed on Pd„110…
F. Bondino,a) E. Vesselli,b) A. Baraldi,b) G. Comelli,b) A. Verdini, A. Cossaro,b) L. Floreano,
and A. Morganteb)
Laboratorio TASC–INFM, s.s. 14 Km 163.5 in Area Science Park, 34012 Trieste, Italy
͑Received 2 December 2002; accepted 25 March 2003͒
The bonding geometry of the C–N molecule in the saturated c(2ϫ2) layer on the Pd͑110͒ surface
has been determined by combining polarization-dependent near edge x-ray absorption fine structure
and full-solid-angle x-ray photoelectron diffraction ͑PED͒. The N K-edge spectra display a strong
dependence on the polar and azimuthal orientation of the light polarization with respect to the
sample surface. A strong forward scattering peak along the ͓001͔ direction is present in the
full-solid-angle photoelectron diffraction data of the C1s core level. Both the position of the C1s
PED forward scattering peak and the angular dependence of the N K-shell absorption spectra
provide direct evidence that the CN molecules is oriented with the molecular axis along the ͓001͔
surface direction, at variance with earlier conclusions based on angle-resolved valence level
photoemission data. The forward scattering peak in the C1s PED data further indicates that the N
atoms lie above the C atoms, with the C–N molecular axis tilted by 25°Ϯ4° with respect to the
surface plane. The close analogy of this geometry with the results of previous structure
determinations of CN adsorbed on Ni and Rh ͑110͒ surfaces is discussed. © 2003 American
Institute of Physics. ͓DOI: 10.1063/1.1574794͔
I. INTRODUCTION
the surface ͓001͔ direction, tilted by approximately 20° from
the surface plane. On Pd͑110͒, CN species, originated by
complete dissociation of C2N2 at room temperature, form an
ordered c(2ϫ2) layer at saturation. In this case, indirect
information on the CN adsorption geometry was previously
derived from ARUPS data.11 The lift of the ␲ orbital degen-
eracy in ARUPS data11 indicated that CN is bonded either
parallel to the surface or with a significant tilt from the sur-
face normal. Previously, ARUPS was also used to determine
the orientation of CN species on Ni͑110͒.12 The ARUPS data
of CN on Pd͑110͒ and on Ni͑110͒ show almost identical
angular dependence of the photoemission features. In both
cases, ARUPS data were interpreted in terms of CN axis
CN is an interesting molecule having a model character
for the surface interactions of organic and polymeric nitriles,
where the CN functional groups provide important centers
for surface attachment.1
Moreover, the formation of cyanide intermediates has
been detected in the catalytic reduction of NO by C2H4 ,2 in
the reaction between NO and carbon,3 in the reaction of CO
with coadsorbed nitrogen4 or NO, both on supported
catalysts5 and single crystals.6 The formation of these cya-
nide intermediates is an unwanted reaction step in the re-
moval of NO by the three-way catalyst, mostly composed of
supported transition metal particles such as Pt, Pd, and Rh on
Al2O3 or SiO2 , since part of the adsorbed CN hydrogenates
and desorbs as poisonous HCN, while the remainder decom-
poses, shifting the desorption of N2 to higher temperatures.
On metal surfaces in ultra-high vacuum environment the
CN bonding geometry is rather peculiar. Unlike most ad-
sorbed diatomic molecules ͑i.e., CO, NO, N2 , etc.͒7 and con-
trary to expectations drawn from coordination chemistry,8 on
most metal-vacuum interfaces CN does not bind in the typi-
cal end-on geometry, but with the molecular axis essentially
parallel to the surface.9–18
¯
aligned along the 110 direction. However, this conclusion
͓
͔
was questioned by a recent NEXAFS and PED measure-
ments of c(2ϫ2)–CN on Ni͑110͒, which indicated a CN
axis oriented along the surface ͓001͔ direction.13
Here we report on the measurement of the angular de-
pendence of NEXAFS N K-edge resonance intensities and
the angular dependent distribution of the photoelectron inten-
sity of c(2ϫ2)-CN on Pd͑110͒. The conclusions derived
both from PED and NEXAFS data are consistent and provide
the azimuthal orientation and the tilt angle of the CN mol-
ecules relative to the Pd͑110͒ substrate. The results of the
present work confirm the nearly parallel bonding configura-
tion of CN species on Pd͑110͒ deduced from ARUPS data,
but indicate a ͓001͔ azimuthal orientation of the C–N mo-
lecular axis, at variance with ARUPS results.
Full quantitative structural determinations of the CN ad-
sorption geometry are limited to two studies, on Ni͑110͒ us-
ing near edge x-ray absorption fine structure ͑NEXAFS͒ and
9
photoelectron diffraction ͑PED͒ and on Rh͑110͒ using dy-
namical low energy electron diffraction ͑LEED͒.18 In both
cases the C–N molecular axis was found to be aligned along
II. EXPERIMENT
a͒
Author to whom correspondence should be addressed. Electronic mail:
The experiments were performed at the ALOISA
beam-line19–21 of ELETTRA, the Trieste synchrotron
facility.
bondino@tasc.infm.it
Also at: Dipartimento di Fisica, Universita di Trieste, Via Valerio 2,
34127 Trieste, Italy.
b͒
`
0021-9606/2003/118(23)/10735/6/$20.00
10735
© 2003 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sat, 20 Dec 2014 01:01:41

10736
J. Chem. Phys., Vol. 118, No. 23, 15 June 2003
Bondino et al.
Sample position was set automatically using a modified
VG–CTPO motorized manipulator with six degrees of free-
dom, which allows to freely select, with 0.01° accuracy, the
orientation of both the photon beam grazing angle (0°–10°)
and of the sample surface with respect to the radiation polar-
ization vector.
At the same time, the emission direction can be chosen
with high flexibility, since a wide portion of the full-solid
angle above the sample can be probed for any surface orien-
tation, thanks to the combined rotation of the frame hosting
the analyzer and of the whole experimental chamber. The
photon beam is linearly polarized with the electric field vec-
tor in the horizontal plane.
Scanned-angle PED patterns were obtained by measur-
ing C1s, N1s, and Pd3d5/2 signal at the peak energy maxi-
mum and at an additional higher kinetic energy value aside
the peak to obtain reference data for linear background sub-
traction. The photon-beam grazing angle was set to 4° with
the electric field vector parallel to the surface plane, i.e., in
transverse electric ͑TE͒ polarization. Automatic angular
scans were performed by computer-controlled stepping mo-
tors rotating the detector in a plane perpendicular to the scat-
tering one by varying the emission angle ␪ from the surface
normal (␪ϭ0°) to the grazing emission (␪ϭ84°) with 1°
constant ␪ steps and rotating the sample around the surface
normal by varying the azimuthal angle ␾ with 2° constant ␾
steps from Ϫ10° to 100° with respect to the Pd surface ͓001͔
direction ͓Fig. 1͑a͔͒. The azimuthal angle ␾ was scanned
over a range of 110°, which includes all the features of the
full-solid-angle pattern, due to the ␲/2 symmetry of the sur-
face. The full-solid-angle PED pattern was obtained by a
symmetric extension of the measured ␲/2 portion.
FIG. 1. ͑a͒ Experimental set up for full-solid-angle PED data acquisition at
the ALOISA end-station. The azimuthal angle ␾ of the analyzer was varied
by rotating the sample around its normal. The polar angle ␪ of the analyzer
was varied by rotating the chamber hosting the analyzer around the beam
axis in the plane highlighted in the figure. The grazing incidence angle ͑␣͒
was fixed and the polarization vector was parallel to the surface plane. ͑b͒
Sketch of the experimental geometry used for the acquisition of the Auger
Yield NEXAFS spectra. ␪ is the angle of the polarization from the surface
␧
plane. ␾ is the azimuthal angle of the polarization with respect to the Pd
␧
surface ͓001͔ direction. At fixed ␾ ϭ0, the angle ␪ was varied by rotating
␧
␧
the sample by ⌬␪␧ around the beam axis. At the same time the chamber
hosting the analyzer ͑C͒ was rotated by the same amount ͑⌬␪␧͒ around the
beam axis, in order to have a constant emission angle for all the spectra with
different ␪ . At fixed ␪ , ␾ was changed by rotating the sample around
␧
␧
␧
the surface normal, while keeping the polarization ␧ vector parallel to the
During the PED scans the precession of the surface nor-
mal due to some residual misalignment of the crystal on the
sample holder was automatically corrected. The required cor-
rections were preliminarily determined by measuring the
beam specular reflectivity from the sample surface using a
diode detector.
surface (␪ ϭ0, TE polarization͒. In both cases, the grazing incidence angle
(␣in) was fixed.
␧
face normal, keeping the sample in TE polarization ͓Fig.
1͑b͔͒.
The PED plot presented in the following shows the in-
tensity in a linear gray scale of C1s core level. Each point in
the displayed patterns shows the intensity measured at a spe-
cific ͑␪, ␾͒ angular position, in a polar coordinate system,
where ␪ is the radial distance from the center of the plot. The
center of the plot corresponds to normal emission and the
circumference to grazing emission. For each polar angle, the
PED data have been normalized to the average intensity cal-
culated over the azimuthal scan.
The absorption at the nitrogen K-edge was measured by
recording the intensity of the nitrogen KVV Auger electron
transition as a function of the photon energy. Auger and pho-
toemitted electrons were collected with a 35 mm mean radius
hemispherical analyzer with ϳ1° acceptance angle.22 All
NEXAFS spectra were normalized to the photon flux, scaled
to the same intensity before the absorption and finally back-
ground subtracted by fitting the data in the pre-edge region.
Polarization-dependent NEXAFS experiments were per-
formed by rotating either the sample and the whole chamber
together around the beam axis or the sample around the sur-
The cleaning procedure of the Pd͑110͒ surface involved
repeated cycles of Ar ion bombardment at room temperature
and annealing to 1050 K, followed by oxygen treatment at
625 K and reduction in hydrogen, until the surface exhibited
a sharp 1ϫ1 reflection high energy electron diffraction
͑RHEED͒ pattern and no traces of contaminants were de-
tected by photoemission. The azimuthal orientation of the
surface was evaluated using RHEED.
Cyanogen (C2N2) was produced by thermal decomposi-
tion of AgCN in a tube attached to the gas line; the gas was
dosed by background exposure via a leak valve.
The saturation c(2ϫ2)-CN/Pd(110) adlayer was pro-
duced by dosing 10 L of C2N2 at 325 K sample temperature.
At this temperature C2N2 adsorbs dissociatively as a single
CN species.23 In agreement with previous LEED investiga-
tions of the saturated CN layer,11 we observed by RHEED an
ordered c(2ϫ2) periodicity.
Background pressure in the UHV chamber, during the
measurements, was always in the low 10Ϫ10 mbar range. All
measurements were collected at room temperature.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sat, 20 Dec 2014 01:01:41

J. Chem. Phys., Vol. 118, No. 23, 15 June 2003
Molecular orientation of CN adsorbed on Pd(110)
10737
FIG. 3. Top: set of Auger yield N K-shell NEXAFS spectra recorded as a
function of the light-polarization polar angle (␪ ), at fixed surface azimuthal
␧
orientation ͑␧ along the ͓001͔ direction͒, fixed photon incidence (4°) and
with the analyzer kept nearly normal to sample surface. Bottom: plot of the
FIG. 2. Top: set of Auger yield N K-shell NEXAFS spectra recorded as a
*
*
␴ /␲ intensity ratio as a function of the light-polarization polar angle.
z
function of the light-polarization azimuthal angle (␾ ), with the electric
␧
field in the surface plane ͑TE polarization͒.
care and should be verified by single hole cross-section
measurements,25 in agreement with previous studies, we as-
*
sign the broad feature at 410 eV to the transition to a ␴
shape resonance. Notice that, in the spectrum measured with
III. RESULTS AND DISCUSSION
A. NEXAFS
¯
the electric field along the 110 direction, the 410 eV broad
͓
͔
Figure 2 displays a set of N K-shell NEXAFS spectra,
measured with the electric field ␧ in the surface plane (␪
ϭ0°, TE polarization͒ at different azimuthal angles ␾ ,
*
*
peak is not present. The plot of the ␴ /␲ intensity ratio as
a function of the azimuthal angle of the polarization is dis-
played in the bottom panel of Fig. 2. The azimuthal depen-
dence of the NEXAFS spectra and, in particular, the in-
␧
␧
where ␾ is the angle between ␧ and the ͓001͔ direction ͓Fig.
␧
1͑b͔͒. The NEXAFS spectra were measured by recording the
Auger KVV intensity with the electron analyzer aligned
along the surface normal. The main peaks in the spectra are
very similar to the features present in the corresponding
¯
equivalence of the ͓001͔ and 110 directions, clearly
͓
͔
indicate that the C–N molecular axis is not perpendicular to
the surface.
Further evidence of this conclusion is derived from the
series of Auger yield N K-edge spectra measured as a func-
tion of the beam-polarization polar angle, at fixed surface
azimuthal orientation ͑␧ along the ͓001͔ direction͒, fixed
photon incidence (4°) and with the analyzer near to the
sample normal ͑Fig. 3͒.
9
N K-edge NEXAFS spectra of CN on Ni͑110͒ and CN on
Pd͑111͒.14 A single sharp peak close to 395 eV with a small
shoulder at 393 eV is observed when the electric vector is
¯
oriented along the 110 direction of the surface plane. In
͓
͔
accordance with previous studies, the sharp peak close to
*
395 eV is assigned to a 1s→␲ transition.
The orientation of the linear light polarization was
By changing the azimuthal orientation of the electric
changed from perpendicular to ␪ ϭ90°, i.e., transverse
͓
␧
¯
field vector from the 110 to the ͓001͔ direction, the inten-
magnetic ͑TM͒ polarization͔ polari nearly parallel to the sur-
͓
͔
*
sity of the ␲ resonance decreases and a broad feature grows
face plane (␪ ϭ5°, ϳTE polarization͒ by simultaneously
␧
up at 15 eV higher photon energy. The energy position of the
rotating the sample and the chamber around the beam axis
while keeping the incidence photon beam angle and the
emission direction fixed with respect to the sample surface
͓see Fig. 1͑b͔͒.
*
broad feature is consistent with the location of the ␴ shape
resonance observed in gas-phase NEXAFS spectra of mol-
ecules containing the CN group24 and with the feature ob-
served in the NEXAFS spectra of CN on Ni͑110͒ and
Pd͑111͒.9,14 Although the assignment of the features above
the photoionization threshold should be approached with
The spectra display a strong angular dependence. At ␪
␧
*
ϭ90° only the ␲ resonance at 393 eV is present. By chang-
ing the polar orientation of the electric field vector from 90°
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sat, 20 Dec 2014 01:01:41

10738
J. Chem. Phys., Vol. 118, No. 23, 15 June 2003
Bondino et al.
*
reduced and the ␲ resonance is clearly splitted in two com-
ponents. The component located at 2 eV higher photon en-
*
ergy can be assigned to residual x intensity. Correspondingly,
the component at 2 eV lower energy with respect to the main
peak in the spectrum measured at ␪ ϭ0°, ␾ ϭ90° is asso-
␧
␧
*
ciated with residual ␲ intensity. Thus, it is clear that the
z
C–N molecular bond lies in a plane parallel or nearly paral-
lel to the surface approximatively oriented along the ͓001͔
direction, even if residual ␲ and ␲ intensities are present
x
z
for geometries where they should be absent according the
selection rules for NEXAFS resonances. In particular, the
presence of the feature assigned to ␲ for ␧ oriented along y,
x
i.e., the ͓001͔ direction, could be due to a vibrational motion
of the nitrogen atom or to a twist of the CN bond from the
͓001͔ direction. The presence of the ␲ for ␧ parallel to sur-
z
face could be either due to a static tilt of the CN bond from
sample plane or to the residual unpolarized light or to vibra-
tional motions of the nitrogen atom or a combination of these
factors.27
From a simple visual inspection of the NEXAFS spectra,
the orientation of the CN molecular axis has been qualita-
tively determined. A quantitative analysis of the NEXAFS
data must be approached with care for the difficulty implied
in handling the normalization, background corrections and
deconvolution of the different effects ͑imperfect light polar-
ization, vibrations͒. A rough estimate of the orientation of
C–N bond axis has been derived from the evaluation of the
FIG. 4. Comparison of the NEXAFS spectra measured for ␧ along the ͑a͒
¯
110 , ͑b͒ ͓001͔, ͑c͒ ͓110͔ directions of the sample.
͓
͔
*
*
␴ /␲ intensity ratio with the electric vector orientation in
*
*
to 5° the ␲ component intensity decreases and the ␴
*
the angle-dependent near-edge spectra, by fitting the ␲
peaks with Gaussian functions and the ␴ resonance with
shape resonance grows up.
*
According to the dipole selection rules governing the
K-shell NEXAFS resonances,26 the intensity associated with
a specific molecular orbital is largest if the electric vector of
the light points in the direction of that molecular orbital and
it vanishes when the electric vector is perpendicular to that
Gaussian functions having a width linearly dependent on the
photon energy, as proposed by Outka and Stohr.28 Neglecting
¨
*
*
the degree of linear polarization, the ␴ /␲ intensity ratio
indicates that the C–N bond lies along the ͓001͔ direction,
tilted by approximately 15°Ϯ10° from the surface plane.
Another set of N K-edge data was collected using total
electron yield ͑TEY͒ detection mode. Except for the TEY
mode giving a much smaller signal to background ratio, the
*
orbital. In our case the ␲ states can only be excited when
the electric vector is perpendicular to the C–N molecular
axis, proving therefore that the molecular orientation is
highly tilted (Ϸ70°–80°) from the surface normal. Further-
more, from the azimuthal-dependent NEXAFS spectra
shown in Fig. 2, it can be concluded that the CN species are
azimuthally aligned along the ͓001͔ direction.
*
*
same behavior of the ␲ and ␴ resonances as a function of
the electric field orientation with respect to the surface nor-
mal was found.
In Fig. 4 the NEXAFS spectra measured for ␧ along the
¯
B. Full-solid-angle PED
sample normal and for ␧ along the sample 110 , ͓001͔ di-
͓
͔
rections are compared.
The experimental full-solid-angle PED pattern, mea-
sured for C1s with photon energy of 947 eV in TE polariza-
tion, i.e., with the photon polarization vector parallel to the
surface plane, is reported in Fig. 5.
The most prominent feature appearing in the experimen-
tal PED pattern is a strong peak along the surface ͓001͔
direction, at approximately 25° from the sample plane. Be-
cause of the high kinetic energy of the photoelectrons, this
peak can be safely attributed to forward scattering of C1s
electrons due to other atoms located at a higher vertical po-
sition.
¯
By labeling the 110 , ͓001͔ and surface normal direc-
͓
͔
tions as x, y, and z, respectively, the single sharp resonant
peak observed for ␧ perpendicular to the surface is assigned
*
to the 1s→␲ excitation, while the sharp resonance appear-
z
¯
ing in the spectrum recorded with ␧ along the 110 direc-
͓
͔
*
tion in the surface plane is assigned to a 1s→␲ excitation.
x
*
The energy position of the ␲ resonance ͑observed at
x
*
␪ ϭ0°, ␾ ϭ90°) and the ␲ excitation ͑visible at ␪
␧
␧
␧
z
ϭ90°) differs by ϳ2 eV. The lifting of the degeneracy of
*
the two ␲ orbitals is consistent with a configuration where
the C–N molecular bond is lying in a plane parallel or nearly
The possibility that the scatterer atom is a first layer Pd
atom can be excluded with the following considerations. A
forward scattering peak due to scattering between C and Pd
atoms would be present in the same azimuthal position if the
*
parallel to the surface because the ␲ molecular orbitals in-
z
teract directly with the substrate. We further observe that in
*
the spectrum at ␪ ϭ5°, ␾ ϭ0°, the intensity of ␲ is much
␧
␧
z
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sat, 20 Dec 2014 01:01:41

J. Chem. Phys., Vol. 118, No. 23, 15 June 2003
Molecular orientation of CN adsorbed on Pd(110)
10739
tions and in the symmetric mapping of the ␲/2 angular por-
tion of experimental data. The best agreement between ex-
perimental and theoretical data, in particular concerning the
forward scattering peak position, was found for ␣ϭ25°
Ϯ4°. The corresponding calculated PED pattern is reported
in Fig. 5͑b͒.
Since the measurements were collected at room tempera-
ture, we point out that anisotropic vibrations of the scatterer,
e.g., a hindered translation of CN along the ͓001͔ direction
and possible differences in the vibrational amplitudes in the
¯
͓001͔, 110 , and ͓110͔ directions, which have not been ac-
͓
͔
counted in our analysis, could affect the value of the tilt of
the CN molecular axis derived from PED and NEXAFS data
with the assumption of static molecules.
This determination of the C–N molecular bond orienta-
tion indicates that CN on Pd͑110͒ follows the general adsorp-
tion behavior of CN species, preferentially bonded with C–N
axis nearly parallel to the surface in metal-vacuum inter-
faces. In particular, the intramolecular orientation of CN on
Pd͑110͒ is consistent with that of CN on Ni͑110͒ determined
by photoelectron diffraction9 and CN on Rh͑110͒ determined
by LEED–IV.18 In both cases the molecule was found to be
located within the ͑110͒ grooves, on top of a second layer
substrate atom, oriented along the ͓001͔ direction, with the
nitrogen atom bridging two first layer metal atoms and the
carbon atom bound to both the first and second layer metal
atoms. Thus, CN adsorbed molecules display a general pref-
erence towards an alignment along the ͓001͔ direction on all
the investigated ͑110͒ transition metal surfaces.
The CN lying-down configuration has been tentatively
attributed to strong electrostatic and polarization contribu-
tions as CNϪ binds to metal surfaces.30 The adsorption con-
figuration found on Pd͑110͒, Ni͑110͒, and Rh͑110͒ is consis-
tent with a threefold coordination and a combined ␲ and ␴
bonding. This preferential adsorption geometry could be
used as a starting point for the calculation of the electronic
properties of the adsorbed CN species to provide a better
physical and chemical insight into the intriguing nature of
CN bond with metal surfaces, whose understanding is still
poor.
FIG. 5. Full-solid-angle C1s PED pattern measured at 947 eV photon en-
ergy ͑a͒ with the PED simulation obtained for a C–N tilt angle of 25° from
the surface plane ͑b͒. Each point in the displayed patterns shows the nor-
malized intensity in a linear gray scale of the C1s core level measured at a
specific ͑␪, ␾͒ angular position, in a polar coordinate system, where ␪ is the
¯
radial distance 110 directions, respectively. ͑c͒ Schematic representation
͓
͔
of the CN adsorption geometry as derived from PED and NEXAFS mea-
surements.
C atoms were located at bridge sites on the second Pd layer,
below the first Pd atomic layer. However, considering a
rough estimation based on the covalent radii of C and Pd
͑refraction effects can be neglected at these energies͒, the
polar position of this peak would be at much grazing emis-
sion (ϳ0°–5° from the sample plane͒ than that observed
(ϳ25°). Therefore, the scatterer atom responsible for the
forward scattering enhancement is likely to be the nitrogen
atom bonded to the emitting carbon. This implies that the
C–N bond is oriented along the surface ͓001͔ direction, with
the C–N axis inclined from the surface plane and that the C
atoms are closer to the surface than the N atoms.
An analogous N1s full-solid-angle PED pattern col-
lected using photon energy of 810 eV in TE polarization did
not show the presence of any forward scattering features, but
it appeared uniform over all the measured region. The lack of
forward scattering features further strengthens the conclusion
derived from C1s PED data, that the N atoms lie above the
emitting C atoms.
The azimuthal position of the peak with respect to the
substrate directions is determined accurately by recording si-
multaneously the C1s and Pd3d5/2 signal during PED data
acquisition.
To derive quantitatively the C–N bond orientation,
single scattering calculations have been performed using the
MSCD code29 at several tilt angles ͑␣͒ of the C–N bond ori-
ented along the ͓001͔ direction. The simulations take into
account the symmetry of the outgoing wave, refraction at the
surface potential step, inelastic effects and isotropic vibra-
tions of the atoms. The existence of two equally populated
CN domains has been considered both in the PED calcula-
IV. CONCLUSIONS
Information on azimuthal molecular orientation as well
as on the tilt of the C–N molecular axis on the Pd͑110͒
surface has been derived by combining angle-dependent
NEXAFS and full-solid-angle PED measuremens.
NEXAFS data and PED yield independent and consis-
tent conclusions, indicating a nearly parallel CN adsorption
on Pd͑110͒, with preferential azimuthal orientation of the
C–N molecular axis along the ͓001͔ direction. Furthermore,
the presence of a 25° tilt of the C–N axis from the surface
plane is found with the C atoms closer to the surface than N
ones. While the near-parallel bonding geometry of CN found
by previous ARUPS data11 is confirmed, the azimuthal ori-
entation of the bond axis contradicts the conclusions based
on the interpretation of photoemission data.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sat, 20 Dec 2014 01:01:41

10740
J. Chem. Phys., Vol. 118, No. 23, 15 June 2003
Bondino et al.
15 M.E. Kordesch, Th. Lindner, J. Somers, W. Stenzel, H. Conrad, A.M.
Bradshaw, and G.P. Williams, Spectrochim. Acta, Part A 43, 1561 ͑1987͒.
16 M.E. Kordesch, W. Stenzel, H. Conrad, and M. Weaver, J. Am. Chem.
Soc. 109, 1878 ͑1987͒.
ACKNOWLEDGMENTS
This work was supported by the Italian INFM, by
MURST under the program ‘‘COFIN01’’ and by Sincrotrone
Trieste S.C.p.A. We gratefully acknowledge R. Gotter for
technical assistance.
17 M.E. Kordesch, W. Feng, W. Stenzel, M. Weaver, and H. Conrad, J. Elec-
tron Spectrosc. Relat. Phenom. 44, 149 ͑1987͒.
18 F. Bondino, A. Baraldi, H. Over, G. Comelli, S. Lizzit, G. Paolucci, and R.
Rosei, Phys. Rev. B 64, 085422 ͑2001͒.
19 L. Floreano, G. Naletto, D. Cvetko et al., Rev. Sci. Instrum. 70, 3855
͑1999͒.
1 F.P. Netzer and M.G. Ramsey, Crit. Rev. Solid State Mater. Sci. 17, 397
͑1992͒.
2 R.M. van Hardeveld, A.J.G.W. Schmidt, R.A. van Santen, and J.W. Ni-
emantsverdriet, J. Vac. Sci. Technol. A 15, 1642 ͑1997͒.
3 L.A. DeLouise and N. Winograd, Surf. Sci. 154, 79 ͑1985͒.
4 P. Jakob, Chem. Phys. Lett. 263, 607 ͑1996͒.
20 F. Bruno, L. Floreano, A. Verdini, D. Cvetko, R. Gotter, A. Morgante, M.
Canepa, and S. Terreni, J. Electron Spectrosc. Relat. Phenom. 127, 85
͑2002͒.
21 Details about the experimental chamber can be found at http://
www.tasc.infm.it/tasc/lds/aloisa/aloisa.html
5 D.R. Mullins, L. Kundakovic, and S.H. Overbury, J. Catal. 195, 169
͑2000͒.
22 R. Gotter, A. Ruocco, A. Morgante, D. Cvetko, L. Floreano, F. Tommasini,
and G. Stefani, Nucl. Instrum. Methods Phys. Res. A 467, 1468 ͑2001͒.
23 A. Baraldi, S. Lizzit, M.G. Ramsey, and F.P. Netzer, Surf. Sci. 416, 214
͑1998͒.
6 J.H. Miners, A.M. Bradshaw, and P. Gardner, Phys. Chem. Chem. Phys. 1,
4909 ͑1999͒.
7 H. Over, Prog. Surf. Sci. 58, 249 ͑1998͒.
8 A.G. Sharpe, The Chemistry of Cyano Complexes of Transition Metals
͑Academic, New York, 1976͒.
24 P. Hitchock, M. Tronc, and A. Modelli, J. Phys. Chem. 93, 3068 ͑1989͒.
25
¨
B. Kempgens, H.M. Koppe, A. Kivimaki, M. Neeb, K. Maier, U. Hergen-
9 N.A. Booth, R. Davis, D.P. Woodruff et al., Surf. Sci. 416, 448 ͑1998͒.
10 M.E. Kordesch, W. Stenzel, and H. Conrad, Surf. Sci. 186, 601 ͑1987͒.
11 M.G. Ramsey, G. Rosina, F.P. Netzer, H.B. Saalfeld, and D.R. Lloyd, Surf.
hahn, and A.M. Bradshaw, Phys. Rev. Lett. 79, 35 ͑1997͒.
26
27
28
29
¨
J. Stohr, NEXAFS Spectroscopy ͑Springer-Verlag, Berlin, 1992͒.
´
M. Perez Jigato, W.K. Walter, and D.A. King, Surf. Sci. 301, 273 ͑1994͒.
¨
Sci. 218, 317 ͑1989͒.
M.G. Ramsey, D. Steinmuller, F.P. Netzer, S. Kostlmeier, J. Lauber, and
D.A. Outka and J. Stohr, J. Chem. Phys. 88, 3539 ͑1988͒.
12
¨
¨
MSCD code, Y. Chen and M.A Van Hove ͑private communication͒, http://
¨
`
`
N. Rosh, Surf. Sci. 309, 82 ͑1994͒.
electron.lbl.gov/mscdpack/; Y. Chen, F.J. Garcıa de Abajo, A. Chasse,
R.X. Ynzunza, A.P. Kaduwela, M.A. Van Hove, and C.S. Fadley, Phys.
Rev. B 58, 13121 ͑1998͒.
13 K. Besenthal, G. Chiarello, M.E. Kordesch, and H. Conrad, Surf. Sci. 178,
667 ͑1986͒.
30
14 J. Somers, M.E. Kordesch, Th. Lindner, H. Conrad, and A.M. Bradshaw,
Surf. Sci. 188, L693 ͑1987͒.
K. Hermann, W. Muller, and P.S. Bagus, J. Electron Spectrosc. Relat.
¨
Phenom. 39, 107 ͑1986͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sat, 20 Dec 2014 01:01:41