J. Phys. Chem.

C 2009, 113, 3909–3916 3909

Monitoring Atomic Rearrangement in PtxPd1-x (x ) 1, 0.7, or 0.5) Nanoparticles Driven by

Reduction and Sulfidation Processes

Fabiano Bernardi,† Maria C. M. Alves,‡ Agnès Traverse,§ Dagoberto O. Silva,‡

Carla W. Scheeren,‡ Jairton Dupont,‡ and Jonder Morais*,†
Instituto de Fı́sica, UniVersidade Federal do Rio Grande do Sul (UFRGS), AVenida Bento GonçalVes,
9500, Bairro Agronomia, CP 15051, CEP 91501-970, Porto Alegre, RS, Brazil, Instituto de Quı́mica,
UniVersidade Federal do Rio Grande do Sul (UFRGS), AVenida Bento GonçalVes, 9500, Bairro Agronomia,
CP 15003, CEP 91501-970, Porto Alegre, RS, Brazil, Laboratoire de Chimie-Physique, UMR8000,
UniVersité de Paris-Sud, Bât. 349, 91405 Orsay Cedex, France
ReceiVed: June 20, 2008; ReVised Manuscript ReceiVed: NoVember 12, 2008

PtxPd1-x (x ) 1, 0.7, or 0.5) nanoparticles submitted to hydrogen reduction and posterior H2S sulfidation at
150 or 300 °C were characterized by in situ X-ray absorption spectroscopy (XAS) and X-ray photoelectron
spectroscopy (XPS). The in situ XAS measurements allowed monitoring of short-range order changes around
the Pt atoms induced by the thermal processes. The surface sensitivity and atom specific characteristics of
XPS provided additional information about the chemical state of the atoms present in the outermost layers of
the nanoparticles. Our experiments also indicate a Pd migration toward the surface of the nanoparticles driven
by the thermal processes. We observed that the reduction process is necessary prior to the occurrence of any
sulfur reaction and that the number of chemisorbed sulfur atoms is directly proportional to the quantity of Pd
atoms.

1. Introduction physicochemical characteristics of the metal atoms are modified,

or when both occur.
Increasingly intense research in the area of nanotechnology
In comparison with other bimetallic systems, the combination
is driven by the numerous potential applications of nanostruc-
of platinum and palladium is particularly advantageous in
tures in several fields, such as catalysis,1 optoelectronics,2 and catalysts used for hydroisomerization, hydrocracking, hydro-
information storage.3 Since catalytic processes are highly genation, and hydrotreatment. For example, a bimetallic catalyst
dependent on surface properties,4 the high surface-to-volume containing platinum and palladium supported on alumina has
ratio of nanoparticles makes them especially attractive for this been demonstrated to improve sulfur resistance in a diesel
use. The catalytic properties of metallic nanoparticles have been hydrotreating process more than a platinum monometallic
used in a variety of industrial processes. Moreover, forming catalyst.12 This improved sulfur resistance is generally proposed
bimetallic systems through the addition of a second metal to result from decreased electron density of the platinum in the
provides a method for tailoring catalytic activity and selectivity presence of palladium.13 Studies on the sulfidation of Pt and
through cooperative effects.5,6 Pt-Pd catalysts have mainly been done on supported catalysts,
Concerning environmental and clean-fuel legislation,7,8 re- such as Pt/Al2O3 or SiO2 and Pt-Pd/Al2O3 or SiO2.14-23 Few
cently, considerable attention has been paid to developing studies have addressed nonsupported nanoparticles.24-26 The
catalysts with high hydrogenation of aromatics and hydrodes- majority of studies on the structural properties of Pt-Pd
ulfurization activities.9,10 These studies indicate that particulate nanoparticles address systems in as-prepared and reduced
emissions in diesel exhaust gases can be reduced by decreasing conditions.12,23,26-30 To our knowledge, detailed studies concern-
the fuel’s sulfur content. A system highly active in the reduction ing the atomic structures and distribution induced by sulfidation
of aromatics has already been obtained, though it was very of nonsupported PtPd nanoparticles are nonexistent.
susceptible to sulfur poisoning.11 Thus, the use of these catalysts The goal of this paper is to characterize PtxPd1-x nanoparticles
is still limited by severe pretreatment requirements until sulfur with x ) 0.5, 0.7, and 1, thereby achieving a detailed description
tolerance can be greatly improved. of the Pt neighboring changes and the atomic arrangement inside
Since the mechanism of metal poisoning by sulfur compounds the nanoparticles as they undergo thermal treatments. The
involves strong chemisorption of the S-containing molecule on nanoparticles were submitted to two types of thermal processes:
the metal sites followed by its hydrogenolysis, as represented hydrogen reduction (named reduction); or hydrogen reduction
by Me0 + H2S S Me-S + H2, the resulting H2S may lead to followed by H2S sulfidation (named sulfidation). The results
the formation of a stable and inactive Me-S species on the were achieved with in situ X-ray absorption spectroscopy (XAS)
catalyst’s surface.9 However, this equilibrium may shift to the and X-ray photoelectron spectroscopy (XPS) with synchrotron
left-hand side when hydrogen pressure is high, when the radiation excitation.
XAS is widely employed to characterize the structural and
* To whom correspondence should be addressed. Phone: +55 51 electronic properties of catalysts and nanoparticles.31 One
33086525. Fax: +55 51 33086510. E-mail: [email protected]. advantage of XAS over other techniques is the possibility of
†
Instituto de Fı́sica, UFRGS.
‡
Instituto de Quı́mica, UFRGS. carrying out in situ measurements. In this work, in situ XAS
§
Université de Paris-Sud. analysis allowed us to obtain the short-range order and structure
10.1021/jp805465x CCC: $40.75  2009 American Chemical Society
Published on Web 02/18/2009
3910 J. Phys. Chem. C, Vol. 113, No. 10, 2009 Bernardi et al.

Figure 1. (i) EXAFS signals at the Pt L3 edge for the PtxPd1-x nanoparticles for different x values and (ii) the corresponding Fourier transform for
the (a) as prepared, (b) reduced at 300 °C, and (c) sulfided at 300 °C. The black points represent the experimental data, and the gray line, the fitting.

around the nanoparticles’ Pt atoms as they were submitted to TABLE 1: Results Obtained from the Quantitative Analysis
heating in a reactive gas environment (reduction or sulfidation). of the EXAFS Data for the Coordination Shell Yielding the
We used a reactor specifically designed for these experiments, Coordination Number (N), Distance (R), and the Debye
Waller Factor (σ2) for the As-Prepared, Reduced
ensuring that the induced structural changes could be monitored.
(T ) 300 °C), and Sulfided (T ) 300 °C) Samples
The in situ measurements were carried out both in serial mode
at steady state conditions (constant temperature and gas flow) x pair N R (Å) σ2 (10-2 Å2)
and in the energy dispersive mode at a constant temperature, As-Prepared
which collected the time-dependent spectra after the addition 0.5 Pt-Pt 5.8 ( 0.3 2.748 ( 0.005 0.46 ( 0.06
of a reactive gas. Extended X-ray absorption fine structure Pt-Pd 4.0 ( 0.2 2.748 ( 0.005 0.86 ( 0.03
(EXAFS) and dispersive X-ray absorption spectroscopy (DXAS) 0.7 Pt-Pt 8.6 ( 0.3 2.735 ( 0.002 0.50 ( 0.03
measurements established the structural behavior of the Pt-Pd 2.2 ( 0.2 2.735 ( 0.002 0.53 ( 0.07
1 Pt-Pt 10.8 ( 0.3 2.750 ( 0.002 0.44 ( 0.02
nanoparticles.
Once we were confident that the thermal processes had Reduced
changed the structure of the nanoparticles, as observed by XAS, 0.5 Pt-Pt 5.9 ( 0.3 2.714 ( 0.004 0.63 ( 0.05
Pt-Pd 4.3 ( 0.3 2.714 ( 0.004 0.83 ( 0.05
we obtained complementary information by performing XPS 0.7 Pt-Pt 7.3 ( 0.4 2.734 ( 0.002 0.57 ( 0.04
on the post-treated samples. XPS provides the electronic Pt-Pd 2.6 ( 0.3 2.734 ( 0.002 0.72 ( 0.08
structure and chemical states of all elements present in the 1 Pt-Pt 11.6 ( 0.3 2.751 ( 0.002 0.89 ( 0.03
nanoparticles. The surface sensitivity and element-specific Sulfided
characteristics of XPS were also exploited to provide informa- 0.5 Pt-S 2.5 ( 0.1 2.262 ( 0.004 0.18 ( 0.04
tion about the chemical state of the atoms present in the Pt-Pt 4.1 ( 0.6 2.756 ( 0.008 0.8 ( 0.1
nanoparticles’ outermost layers. We have also associated the Pt-Pd 1.1 ( 0.1 2.756 ( 0.008 0.23 ( 0.08
XPS with argon sputtering to investigate if the nanoparticles’ 0.7 Pt-S 0.63 ( 0.07 2.26 ( 0.02 0.3 ( 0.1
cores have a different composition from their shells. By Pt-Pt 8.4 ( 0.4 2.760 ( 0.003 0.92 ( 0.04
comparing the XPS results obtained for the as-prepared, reduced, Pt-Pd 2.0 ( 0.2 2.760 ( 0.003 0.43 ( 0.05
1 Pt-S 0.5 ( 0.2 2.21 ( 0.04 2.2 ( 0.9
and sulfided samples, we were able to explore the influence of Pt-Pt 11.3 ( 0.3 2.752 ( 0.002 0.91 ( 0.06
the thermal treatments on their composition, chemical state, and
atomic rearrangement.
zolium hexafluorophosphate (BMIPF6). The resulting solution
Our main results show (i) nanoparticle composition (i.e., the
was reacted with molecular hydrogen at 75 °C and 4 atm for 5
x value) is a key factor in determining the amount of sulfur
min.32 The nanoparticles were isolated by centrifugation of the
that will be adsorbed (or chemisorbed) by the catalyst, provided
black solution formed by this reaction. TEM and XRD studies32
that the nanoparticle has been reduced prior to the H2S reaction;
showed that this procedure leads to the formation of 3.8-, 4.5-,
(ii) PtS, PtS2, PdS, and PdS2 are all identifiable after the
and 4.8-nm-average-diameter PtxPd1-x nanoparticles, for which
sulfidation process; and (iii) there is an atomic rearrangement
x ) 1, 0.7, and 0.5, respectively.
of Pt and Pd atoms inside the nanoparticles, leading to a
2.2. In Situ XAS Measurements. For the XAS experiments,
core-shell structure induced by the thermal processes under
about 10 mg of the nanoparticle powder was compacted to
controlled gas atmosphere.
produce 5-mm-diameter pellets. The pellet was introduced to
the previously described reactor,24 which allows for controlled
2. Experimental Methods thermal treatment of the sample under controlled gas flow.
2.1. Preparation of the Nanoparticles. PtxPd1-x nanopar- Transmission mode measurements were performed at the LNLS
ticles were prepared by dissolution of the Pt and Pd precursors, (Brazilian Synchroton Light Laboratory) at the XAFS1 beam-
Pt2(dba)3 and Pd(acac)2, in 1 mL of 1-n-buthyl-3-methylimida- line.33 The spectra were collected at the Pt L3 edge using a
Monitoring Atomic Rearrangement in PtxPd1-x J. Phys. Chem. C, Vol. 113, No. 10, 2009 3911

resolution of about 0.13 eV. The base pressure used inside the
chamber was about 1.3 × 10-9 mbar. The monochromator
photon energy calibration was done at the Si K edge (1839 eV).
An additional calibration of the analyzer’s energy was performed
after every sample change using a standard Ag foil (Ag 3d5/2
peak at 368.3 eV). We also considered the C 1s peak value of
284.5 eV as a reference to verify possible charging effects. The
samples were placed on carbon tape, and the XPS measurements
were obtained at a 45° takeoff angle at room temperature. The
samples were studied before and after ion sputtering using an
Ar+ ion beam at 3 keV for 10 min at a pressure of 4.0 × 10-6
mbar and impinging at a grazing incidence of 10° with respect
to the sample surface.
2.5. Data Analysis. The EXAFS data were analyzed in
accordance with the standard procedure of data reduction,31
using IFEFFIT.36 FEFF was used to obtain the phase shift and
amplitudes.37 The EXAFS signal (k) was extracted then
Fourier-transformed using a Kaiser-Bessel window with ∆k
range of 8.1 Å-1. Single and multiple scattering events were
Figure 2. Comparison of EXAFS signals at the Pt L3 edge after considered in the fitting procedure. The S02 was fixed at 0.84
sulfidation process at 150 and 300 °C. The fitting is represented by the for all samples, and in some cases, it was necessary to apply
gray line.
cumulant expansion (c3 and c4 values around 10-4).
XPSPeak version 4.1 was used to fit the XPS results. All
channel-cut Si (111) crystal and three argon-filled ionization peaks were adjusted using a Shirley type background and an
chambers. A standard Pt foil was used to calibrate the mono- asymmetric Gaussian-Lorentzian sum function (25% Lorent-
chromator.Thespectrawereacquiredintherangeof11 440-12 200 zian contribution). The fwhm of the XPS components were
eV with 2 eV steps and 2s/point. Two to four scans were allowed to vary 0.6 eV around their typical values.
acquired to improve the signal-to-noise ratio. Experiments at
Pd K edge (24 350 eV) were not performed due to low photon
flux for this energy range at LNLS. 3. Results
The XAS spectra of the PtxPd1-x as-prepared samples were 3.1. In Situ XAS Measurements. The EXAFS spectra at
measured at room temperature. The samples were then submitted the Pt L3 edge for bimetallic nanoparticles (PtxPd1-x x ) 0.5
to three processes: (i) the temperature was increased at a rate and 0.7) as well as monometallic Pt nanoparticles were collected
of 15 °C/min to the selected temperature (150 or 300 °C) under for the as-prepared, reduced at 300 °C and sulfided at 300 °C
a He flux of 13 cm3/min. This temperature was maintained under samples. The EXAFS signals (k) and the Fourier transforms
He flow for 20 min. (ii) After this, the reduction process was (FT) as well as the best fit of the signal (gray line) are presented
begun under a gas mixture of 78% He (≈13 cm3/min) + 22% in Figure 1.
H2 (≈ 4 cm3/min), for 20 min. (iii) After reduction, sulfidation The EXAFS signals for the as-prepared PtxPd1-x nanoparticles
was accomplished by flowing 75% He (≈13 cm3/min) + 21% for x ) 0.5, 0.7 and 1 (Figure 1a) are generally similar to those
H2 (≈ 4 cm3/min) + 4% H2S (≈0.7 cm3/min) for 30 min. XAS of bulk Pt, where the characteristic signature of the fcc structure
spectra were acquired at the end of each step. The total gas can be found in the nanoparticles. For x ) 1 the overall shape
pressure at the sample was kept at approximately 35 psi. is identical to bulk Pt with a small dumping of the oscillations.
2.3. In Situ DXAS Measurements. Dispersive XAS mea- Pd replacement of part of the Pt atoms in the bimetallic systems
surements were performed in the same reactor used for the XAS induces stronger signal damping as the Pd content increases,
experiments at the LNLS DXAS beamline.34 The monochro- which is clearly seen for k > 5 Å-1 in Figure 1a. The damping
mator consists of a curved Si(111) crystal (dispersive polychro- could be due to particle structural disorder and (or) size
mator) that focuses the beam in the horizontal plane down to reduction. Since TEM indicates a slight diameter increase when
about 200 µm and in the vertical plane to about 500 µm. The x goes from 1 to 0.5, structural disorder is assumed to be the
detector was a position-sensitive CCD camera. The reactor was cause of the oscillation damping.
placed in the beamline, taking care to place the pellet at the The first peak in the FT of standard Pt corresponds to
X-ray focal point. The measurements were performed at the Pt platinum atoms in the position of nearest neighbors. The peak
L3 edge during sulfidation of the Pt0.5Pd0.5 nanoparticles at 300 is split due to variation of the backscattering amplitude with k
°C using the same gas fluxes as in the XAS measurements. The for heavy atoms.31 The FT of PtxPd1-x nanoparticles (x ) 0.5
time resolution was about 100 ms, and a spectrum was collected and x ) 0.7) has two peaks associated with Pt-Pt and Pt-Pd
every 18 s, with an accumulation time of 3 s. bonds in the coordination shell. The absence of contributions
2.4. XPS Measurements. After each process, upon comple- at greater distances indicates only short-range order. For x )
tion of the XAS measurements, the samples were introduced 1, only Pt-Pt bonds are observed, and there is a reduction in
into the SXS beamline35 endstation (also at LNLS) where XPS the contribution of higher shells compared to Pt metal.
measurements are taken. The spectra were collected using an After reduction at 300 °C (Figure 1b), the Pt0.5Pd0.5 presents
InSb (111) double crystal monochromator at a fixed photon a shift in the Pt-Pt and Pt-Pd distances for lower values. No
energy of 1840 eV. The hemispherical electron analyzer significant changes for the x ) 1 and 0.7 systems were observed.
(Physical Electronics model 10-360) was set at a pass energy It is interesting to note how the two bimetallic systems (x )
of 23.5 eV, and the energy step was 0.1 eV, with an acquisition 0.5 and 0.7) behaved differently during the sulfidation at 300
time of 500 ms/point. The use of synchrotron radiation excitation °C (Figure 1c). The system with the higher Pd content is strongly
provided spectra with intense signal and an excellent overall sulfided, as can be observed by the strong change in the EXAFS
3912 J. Phys. Chem. C, Vol. 113, No. 10, 2009 Bernardi et al.

Figure 3. (a) Evolution of the XANES measurements at the Pt L3 edge for the Pt0.5Pd0.5 nanoparticles submitted to the sulfidation process at
300 °C. The time axis corresponds to the elapsed minutes from the beginning of the sulfidation process. (b) Comparison of selected XANES curves
(solid lines), where the dotted line is used to guide the eyes. (c) Plot of the first (black line) and the last (red line) spectra for better comparison of
the intensities in the near edge structure.

pattern that yields a strong Pt-S contribution at about 1.8 Å in fewer S atoms were bonded to the nanoparticles. The fitting
the FT. For x ) 0.7, only a small asymmetry in the FT first results indicated that at 300 °C, the N value was (2.5 ( 0.1),
peak is observed that can be associated with the formation of a but at 150 °C, N ) (0.8 ( 0.2).
Pt-S bond. We observed that sulfidation occurs only when the 3.2. In Situ DXAS Measurements. The DXAS experiment
samples have been previously reduced under the conditions was used to get preliminary information on the kinetics of
described above. sulfidation of the Pt0.5Pd0.5 nanoparticles at 300 °C, and the
Typical R-factors in the quantitative analysis (Table 1), results are shown in Figure 3. The nanoparticles were previously
provided by the IFEFFIT code, were between 0.008 and 0.023, reduced using the same procedure described for the XAS
illustrating good fit quality. For the bimetallic systems, the measurements. The zero on the time scale corresponds to the
number of Pt-Pt or Pt-Pd neighbors does not follow the sample introduction of the H2S. There is an increase in the intensity of
composition; that is, the NPt-Pd/NPt-Pt ratio is different from the absorption edge that corresponds to a 2p f 5d transition.
(1 - x)/x. This is due to the oxidation of Pt, which yields more empty
For all the as-prepared samples, the metal-metal distances electronic states in the 5d band. A change in the position of the
(Pt-Pt ) 2.77 Å and Pd-Pd ) 2.73 Å) are close to those of first oscillation maximum beyond the main edge is also
a fcc lattice and match, within the acceptable range of noticeable. The position of this feature is stabilized after 7.5
uncertainty, that of pure fcc Pt. For the reduced samples, there min of sulfidation and remains unchanged (Figure 3b). Thus,
is significant bond contraction, from 2.748 to 2.714 Å, for the one may consider sulfidation a slow process, when compared
bimetallic system x ) 0.5. to the DXAS measurement but fast in comparison to the
The N values from the EXAFS results for the as-prepared acquisition time of a XAS spectrum, which is ∼40 min. For
samples are in agreement with the TEM findings.32 Indeed, since this reason, the XAS measurements presented in this paper were
the coordination number is related to the average nanoparticle performed on samples that had already reached a steady state.
diameter,38 it increases with the addition of Pd in the alloy. After 3.3. XPS Measurements. Figure 4 shows the Pt 4f and Pd
the reduction process, the coordination number of Pt increases 3d photoemission spectra for the as-prepared, reduced at
for all particles. For x ) 1, it goes to 11.6 ( 0.3 (almost 300 °C and sulfided at 300 °C PtxPd1-x nanoparticles (x ) 0.5,
corresponding to bulk Pt), whereas the Pt coordination numbers 0.7, and 1). For comparison purposes, all the spectra are
for the alloy samples add up to only 9.9 ( 0.3 (x ) 0.7) and presented with the same x and y scale, together with the fitting
10.2 ( 0.3 (x ) 0.5). In addition, the reduced samples present curves. Sulfur 2p photoemission spectra were also collected (not
a decrease in the coordination number with the addition of Pd. shown here) and were observed only in the sulfided samples.
These results indicate that a sintering has occurred for the pure The peak positions and the chemical components obtained from
Pt nanoparticle, and the presence of Pd inhibits the sintering of the XPS fitting results are presented in Table 2 for the sulfided
a dispersed Pt phase in the alloy samples. This role of Pd is in nanoparticles.
agreement with previous results found in the literature.12,39 For the as-prepared samples (Figure 4a), the Pt 4f region
Therefore, the atomic rearrangement with formation of a displays the Pt 4f7/2 and Pt 4f5/2 doublet at 70.8 and 74.1 eV,
core-shell structure in the Pt-Pd nanoparticles may also work respectively. These energies correspond to Pt-Pt bonds (Pt0).
as a mechanism to restrain the sintering process. The Pt 4f spectra obtained for all as-prepared nanoparticles are
We have also compared the sulfidation of Pt0.5Pd0.5 nanopar- quite similar, independent of their composition. The Pd 3d region
ticles at 150 and 300 °C (Figure 2). The sulfidation at a lower displays the Pd 3d5/2 (335.2 eV) and Pd 3d3/2 (340.5 eV) doublet
temperature shows a less pronounced Pt-S peak, indicating that as well as the Pt 4d3/2 (332.0 eV) peak.40 The binding energy
Monitoring Atomic Rearrangement in PtxPd1-x J. Phys. Chem. C, Vol. 113, No. 10, 2009 3913

Figure 4. XPS measurements at Pt 4f and Pd 3d regions for different x values for (a) as prepared, (b) reduced at 300 °C, and (c) sulfided at
300 °C. The vertical lines indicate the binding energies of the observed chemical components.

values for the Pd 3d peaks correspond to Pd-Pd bonds (Pd0).41 A closer analysis of the Pd 3d region allows us to obtain
The Pd0 component has a shakeup satellite, as described information about both the Pd and the Pt atoms, since it displays
previously.42 The satellite is located at 7.2 eV below the Pd photoemission lines at similar kinetic energies and, conse-
3d5/2 peak, and it was considered in the Pd0 component fitting quently, comparable electron mean free paths of about 2 nm.
procedure. The peak at 336.4 eV43 corresponds to the palladium Therefore, each Pd 3d region spectrum of the bimetallic samples
oxide (PdO) component. allows us to monitor the relative intensity of the Pt and Pd peaks
After reduction (Figure 4b), the observed Pt and Pd chemical within the same probed range. Hence, the evolution of the Pt
components remain the same. The PdO component is still 4d3/2/Pd 3d peak area ratio after the reduction and sulfidation
present, which is probably the result of brief air exposure prior processes can be compared, and its variations could be related
to the XPS measurements. to different Pt and Pd depth distribution.
After sulfidation (Figure 4c), the Pt0 component is no longer After the reduction process (Figure 4b), there is a decrease
observed in the Pt 4f region for all samples. A new component in the Pt 4d3/2/Pd 3d peak ratio as compared to the as-prepared
appears (Pt 4f7/2 at 72.0 eV), which was assigned to the Pt-S samples (Figure 4a). This implies that upon reduction, Pd content
bonds.20 For the pure Pt nanoparticles (x ) 1), the PtS2 increases at the probed depth of 2 nm. The Pt 4d 3/2 (in the Pd
component is observed (Pt 4f7/2 at 73.4 eV), in addition to the region) is no longer observed in the sulfided samples (Figure
PtS at 72.0 eV. The Pd0 components for the bimetallic samples 4c), indicating the enrichment of Pd atoms at the probed depth.
are also no longer observable and are replaced by PdS for x ) Since the Pt 4f photoelectrons have higher kinetic energy than
0.7 and by PdS2 for x ) 0.5. the Pt 4d3/2 ones, the Pt 4f must originate from deeper atoms;
3914 J. Phys. Chem. C, Vol. 113, No. 10, 2009 Bernardi et al.

TABLE 2: XPS Results Obtained from the Fitting of the TABLE 4: Comparison of the Results Obtained from XPS
Experimental Curves with Identification of the Different Measurements for the Sulfided Pt0.5Pd0.5 Submitted to 150
Chemical Components for the Sulfided Nanoparticles at and 300 °C
300 °C
before sputtering after sputtering
before sputtering after sputtering T (°C) peak Eb (eV) chem comp Eb (eV) chem comp
x peak Eb (eV) chem comp Eb (eV) chem comp
150 Pt 4f7/2 72.0 PtS 70.8 Pt0
0.5 Pt 4f7/2 72.0 PtS 70.8 Pt0 73.4 PtS2 72.0 PtS
72.0 PtS Pd 3d5/2 337.0 PdS2 335.2 Pd0
Pd 3d5/2 337.0 PdS2 335.2 Pd0 337.0 PdS2
336.3 PdS S 2p3/2 162.2 PtS 162.2 PtS
S 2p3/2 162.2 PtS 162.2 PtS 163.3 PdS2
163.3 PdS2 161.8 PdS 300 Pt 4f7/2 72.0 PtS 70.8 Pt0
0.7 Pt 4f7/2 72.0 PtS 70.8 Pt0 72.0 PtS
72.0 PtS Pd 3d5/2 337.0 PdS2 335.2 Pd0
Pd 3d5/2 336.3 PdS 335.2 Pd0 336.3 PdS
336.3 PdS S 2p3/2 162.2 PtS 161.8 PdS
S 2p3/2 161.8 PdS 161.8 PdS 163.3 PdS2 162.2 PtS
162.2 PtS 162.2 PtS
1 Pt 4f7/2 72.0 PtS 70.8 Pt0
73.4 PtS2 72.0 PtS Table 4 shows the XPS results for Pt0.5Pd0.5 nanoparticles
S 2p3/2 169.2 S+6 169.2 S+6 sulfided at two temperatures, 150 and 300 °C. The overall XPS
162.2 PtS 162.2 PtS results (Tables 2 and 4) indicate that sulfur is present in different
161.8 PtS2
chemical states: PtS, PdS, PtS2, and PdS2. In addition, the S 2p
region indicates the presence of S6+ and SO4-2 compounds only
TABLE 3: Pt 4d 3/2/Pd 3d Peak Area Ratio Obtained from for Pt nanoparticles sulfided at 300 and 150 °C, respectively.
the XPS Fitting for the As-Prepared, Reduced (300 °C), and
Sulfided (300 °C) PtxPd1-x Nanoparticles (x ) 0.5 and 0.7)
Discussion
Pt 4d3/2/ Pd 3d peak area ratio
x sample before sputtering after sputtering
We may highlight two main results of the study presented
here. The first one concerns the dependence of the sulfidation
0.5 as-prepared 0.28 ( 0.03 0.33 ( 0.02 process on the sample’s chemical state (either oxidized or
reduced 300 °C 0.20 ( 0.01 0.31 ( 0.02 reduced), sample composition (x value), temperature, and
sulfided 300 °C 0.00 ( 0.01 0.13 ( 0.01
0.7 as-prepared 0.76 ( 0.02 0.77 ( 0.03 duration of the thermal processes. The second one involves the
reduced 300 °C 0.52 ( 0.01 0.76 ( 0.02 observation of the atomic rearrangement inside the nanoparticles.
sulfided 300 °C 0.054 ( 0.02 0.47 ( 0.01 Regarding the sulfidation, it takes place only after PtxPd1-x
reduction under the conditions described here. For Pt0.5Pd0.5
its signal is clearly detectable after sulfidation. The significant nanoparticles, the sulfidation is achieved in 7.5 min, as seen by
decrease in the Pt 4f signal when comparing the as-prepared in situ DXAS measurements. The observed sulfidation process
sample to the reduced one and to the sulfided one also reflects is considerably faster than similar results found in the literature.
the rearrangement of the Pt atoms toward the nanoparticle core. For example, Bando et al.15 have obtained a sulfidation process
To confirm the observation of Pd surface enrichment, we at 280 °C that is accomplished after ∼100 min for the PtPd/
submitted the samples to Ar+ sputtering to eliminate their Al2O3 system. Once reduction is achieved, sulfur incorporation
outermost atoms. Figure 5 compares the XPS measurements increases proportionally with the amount of Pd. The presence
for the Pt0.5Pd0.5 nanoparticles before (i) and after (ii) sputtering. of platinum sulfide observed for all sulfided samples was also
The as-prepared sample shows no significant changes after obtained for sulfided-supported Pt catalysts.19-21 As established
sputtering, which proves that sputtering does not induce atomic previously,21 PtS is responsible for a bonding effect in some
rearrangement. The Pt 4f region results (Figure 5a) demonstrate reactions instead of the structural effect caused by sulfate
a clear increase in intensity after the sputtering for the reduced compounds. The XPS results from the S 2p region shows the
and sulfided nanoparticles with a simultaneous decrease in the presence of SO42- and S6+ compounds with energies of 168.2
Pd peak intensity (Figure 5b). Moreover, the Pt 4d3/2 peak, which and 169.2 eV, as reported previously.19-21 We believe that the
is not visible in the sulfided samples, reappears after sputtering. sulfate components are due to the samples’ exposure to air
Thus, one can deduce from these observations that there is a between the XAS and XPS measurements.
modification of the atomic arrangement inside the nanoparticles The sulfidation of Pt (x ) 1) nanoparticles at 300 °C produces
after either the reduction or sulfidation processes, resulting in a PtS and PtS2, and the latter has not been found in similar
Pd-enriched surface and a Pt-enriched core. Table 3 shows the investigated systems.19-21 In contrast, only PtS was observed
Pt 4d3/2/Pd 3d peak area ratio for the as-prepared, reduced at for Pt0.7Pd0.3 and Pt0.5Pd0.5 nanoparticles sulfided at the same
300 °C, and sulfided at 300 °C PtxPd1-x nanoparticles (x ) 0.5 temperature, in addition to PdS and PdS2, respectively. There-
and 0.7) before and after sputtering. This table provides a fore, PtS2 is not formed for alloy samples, which have created
quantitative argument for the observed Pd surface enrichment. a Pd protecting shell. We believe that the interaction between
This rearrangement is accompanied by a modification of the Pt and Pd atoms may promote an electron deficient character
metal-sulfur bonds. The Pt0 component is observed again after of the Pt atoms in the interface between the core and shell
sputtering (Figure 5a), indicating that PtS is present at the regions, as previously reported.9,12,13,15,22,39 In contrast, PtS2 and
surface. Simultaneously, the presence of PdS2 is no longer PdS2 are present in Pt0.5Pd0.5 nanoparticles sulfided at lower
observed after sputtering and is replaced by the Pd0 and PdS temperature (150 °C), implying a less pronounced electron
components. We observed analogous behavior for the Pt0.7Pd0.3 deficient character of Pt. It is noteworthy that an electron
nanoparticles in the XPS analysis. deficient character of Pd has also been observed.44 Hence, these
Monitoring Atomic Rearrangement in PtxPd1-x J. Phys. Chem. C, Vol. 113, No. 10, 2009 3915

Figure 5. (a) Pt 4f and (b) Pd 3d XPS regions before (i) and after (ii) sputtering with Ar+ for Pt0.5Pd0.5 nanoparticles. The vertical lines indicate
the binding energies of the observed chemical components.

Figure 6. Atomic rearrangement model suggested for PtxPd1-x (x ) 0.7 or 0.5) samples for the (i) as prepared, (ii) reduced, and (iii) sulfided
conditions.

results suggest that the temperature is also an important the inner to the outer region. A scheme of the atomic
parameter that influences the final sulfidation products. organization inside the nanoparticles is proposed in Figure 6.
Concerning the second main result, we clearly observed the The core-shell structure suggested here has been noticed by
atomic rearrangement within the PtxPd1-x nanoparticles, which other authors in previous papers.23,26,12,29,30 In ref 12, the authors
is induced by the thermal processes under hydrogen or sulfur found this same structure by XPS measurements performed
atmosphere. Using the surface sensitivity of the XPS technique before and after argon bombardment for a sample with a Pt/Pd
and monitoring photoemission peaks at a similar probe depth, ratio of 3/1 after reduction at 300 °C, starting from a random
we are able to monitor the chemical composition near the surface distribution for the as-prepared samples. Guillon et al.22 have
region of the nanoparticles. On the basis of our findings, we found that a PtPd/Al2O3 system under exposure to sulfur has a
propose that initially the Pt and Pd atoms have no preferential core-shell structure with a PdS shell, but without Pt-S bonds.
site inside the nanoparticles. The reduction process promotes a In contrast, we have found an important contribution of the Pt-S
migration of Pd atoms to the nanoparticles’ surface. A post- bond in our samples from both XPS and EXAFS analysis. We
sulfidation process increases the tendency of atomic rearrange- do not attribute the difference to the support of Al2O3 used22
ment and promotes the formation of metal-sulfur bonds at the because Pt-S bond has been found previously15 for a PtPd/
nanoparticle surface. Thus, both thermal processes induce the Al2O3 system. We believe that this difference may be due to
formation of a core-shell structure in the bimetallic nanopar- the higher temperatures employed here (300 °C), as opposed
ticles. In addition, XPS signals of metal-metal bonds reappear to 100 °C in ref 22. In fact, for the PtPd nanoparticles sulfided
on sulfided samples submitted to sputtering. For example, after at 150 °C (see Figure 2), we found a weak contribution of Pt-S
sputtering of the sulfided Pt0.5Pd0.5 sample, one observes a from XAS, and the core-shell structure was also observed by
change in chemical states from PtS to (PtS + Pt0) and from XPS.
PdS2 to (PdS + Pd0). It can be understood as a gradient of PtSz Hwang et al.,45 in a theoretical work, deduced an equation to
or PdSz compounds (z ) 0, 1, or 2) with an increase in z from predict the atomic distribution of nanoparticles on the basis of
3916 J. Phys. Chem. C, Vol. 113, No. 10, 2009 Bernardi et al.

EXAFS analysis. In this case, the equation takes into account (9) Navarro, R. M.; Pawelec, B.; Trejo, J. M.; Mariscal, R.; Fierro,
values obtained from measurements at both Pt and Pd edges to J. L. G. J. Catal. 2000, 189, 184.
(10) Yasuda, H.; Yoshimura, Y. Catal. Lett. 1997, 46, 43.
calculate the J factor, as named by the authors. Therefore, by (11) Barbier, J.; Lamy-Pitarra, E.; Marecot, P.; Boitiaux, J. P.; Cosyns,
calculating just the J factors for the Pt L3 edge, we have found J.; Verna, F. AdV. Catal. 1990, 37, 259.
values that agree with the core-shell structure proposed in our (12) Fiermans, L.; De Gryse, R.; De Doncker, G.; Jacobs, P. A.; Martens,
J. A. J. Catal. 2000, 193, 108.
work formed after the thermal processes. (13) Lee, J.-K.; Rheen, H.-K. J. Catal. 1998, 177, 208.
Coming back to the role of Pd in sulfidation, we have noticed (14) Bensaddik, A.; Caballero, A.; Bazin, D.; Dexpert, H.; Didillon, B.;
that the bonding enthalpies of a Pd-S and Pt-S formation are Lynch, J. Appl. Catal., A 1997, 162, 171.
(15) Bando, K. K.; Kawai, T.; Asakura, K.; Matsui, T.; Bihan, L. L.;
practically the same under thermodynamical equilibrium. Nev- Yasuda, H.; Yoshimura, Y.; Oyama, S. T. Catal. Today 2006, 111, 199.
ertheless, the sulfidation process for our nanoparticles is (16) Fujikawa, T.; Tsuji, K.; Mizuguchi, H.; Godo, H.; Idei, K.; Usui,
improved by the presence of Pd atoms. Taking into account K. Catal. Lett. 1999, 63, 27.
the atomic arrangement proposed in Figure 6, in which Pd atoms (17) Vaarkamp, M.; Miller, J. T.; Modica, F. S.; Lane, G. S.; Konings-
berger, D. C. J. Catal. 1992, 138, 675.
are localized preferentially at the nanoparticle surface after a (18) Kobayashi, J.; Shimizu, T.; Mizushima, T Bull. Chem. Soc. Jpn.
reduction process, we suggest that Pd atoms display an increased 2000, 73 (10), 2395.
probability to bond with sulfur in the case of low-dimensionality (19) Appay, M.-D.; Manoli, J.-M.; Potvin, C.; Muhler, M.; Wild, U.;
Pozdnyakova, O.; Paál, Z. J. Catal. 2004, 222, 419.
configuration. (20) Paál, Z.; Muhler, M.; Matusek, K. J. Catal. 1998, 175, 245.
(21) Paál, Z.; Matusek, K.; Muhler, M. Appl. Catal., A 1997, 149, 113.
Conclusions (22) Guillon, E.; Lynch, J.; Uzio, D.; Didilon, B. Catal. Today 2001,
65, 201.
In summary, in situ XAS measurements of the PtxPd1-x (x ) (23) Bazin, D.; Guillaume, D.; Pichon, Ch.; Uzio, D.; Lopez, S. Oil.
1, 0.7 or 0.5) nanoparticles were performed during reduction Gas Sci. Technol. 2005, 60 (5), 801.
(24) Bernardi, F.; Alves, M. C. M.; Scheeren, C. W.; Dupont, J.; Morais,
and sulfidation processes at 150 and 300 °C. The vicinity of Pt J. J. Electron Spectrosc. Relat. Phenom. 2007, 156-158, 186.
atoms was modified by the elimination of lighter atoms after (25) Wang, Y.; Toshima, N. J. Phys. Chem. B 1997, 101, 5301.
reduction and by the incorporation of S during sulfidation. The (26) Toshima, N.; Harada, M.; Yonezawa, T.; Kushihashi, K.; Asakura,
K. J. Phys. Chem. 1991, 95, 7448.
in situ DXAS measurements showed a fast process of sulfur (27) Cho, S. J.; Kang, S. K. Catal. Today 2004, 93-95, 561.
incorporation in the nanoparticles. The presence of S atoms (28) Chen, C.-H.; Hwang, B.-J.; Wang, G.-R.; Sarma, L. S.; Tang, M.-
increases with the amount of Pd, which improves the capability T.; Liu, D.-G.; Lee, J.-F. J. Phys. Chem. B 2005, 109, 21566.
of the nanoparticles to form metal-sulfur bonds. The XPS (29) Hansen, P. L.; Molenbroek, A. M.; Ruban, A. V. J. Phys. Chem.
B 1997, 101, 1861.
measurements supported the XAS observations and identified (30) van den Oretelaar, L. C. A.; Nooij, O. W.; Oerlemans, S.; Denier
the different chemical states of the elements present in the van der Gon, A. W.; Brongersma, H. H.; Lefferts, L.; Roosenbrand, A. G.;
nanoparticles. In particular, they showed that the reduction and van Veen, J. A. R. J. Phys. Chem. B 1998, 102, 3445.
(31) X-ray Absorption: Principles, applications and techniques of
sulfidation processes induce an atomic rearrangement with a EXAFS, SEXAFS and XANES in Chemical Analysis; Chemical Analysis;
formation of a core-shell structure, with a Pd-enriched shell Koningsberger, D. C., Prins, R. ,Eds.; John Wiley & Sons: New York, 1988;
and a Pt-rich core. We conclude that the migration of Pd atoms Vol. 92.
(32) Scheeren, C. W.; Machado, G.; Teixeira, S. R.; Morais, J.;
toward the surface and the number of Pd atoms in the Domingos, J. B.; Dupont, J. J. Phys. Chem. B 2006, 110, 13011.
nanoparticles are key factors in the sulfur trapping mechanism (33) Tolentino, H. C. N.; Ramos, A. Y.; Alves, M. C. M.; Barrea, R. A.;
from the gas phase. We plan to perform future experiments on Tamura, E.; Cezar, J. C.; Watanabe, N. J. Synchrotron Radiat. 2001, 8,
similar systems but containing higher Pd concentrations and to 1040.
(34) Tolentino, H. C. N.; Cezar, J. C.; Watanabe, N.; Piamonteze, C.;
monitor the atomic local order around the Pd atoms, as well. Souza-Neto, N. M.; Tamura, E.; Ramos, A. Y.; Neueschwander, R. Phys.
Scr. 2005, T115, 977.
Acknowledgment. We thank the support given by the LNLS (35) Tolentino, H.; Compagnon-Cailhol, V.; Vicentin, F. C.; Abbate,
staff. Work funded by CNPq, CT-PETRO, CT-ENERG, and M. J. Synchrotron Radiat. 1998, 5, 539.
(36) Newville, N. J. Synchrotron Radiat. 2001, 8, 322.
LNLS (DXAS 3384, DXAS 4556, DXAS 5305, XAFS1 5269, (37) Zabinsky, S. I.; Rehr, J. J.; Ankudinov, A.; Albers, A. C.; Eller,
XAFS1 5695, and SXS 5734 proposals). F.B. thanks CNPq for M. J. Phys. ReV. B 1995, 52, 2995.
his Ph.D. fellowship. (38) Gregor, R. B.; Lytle, F. W. J. Catal. 1980, 63, 476.
(39) Lin, T.-B.; Jan, C.-H.; Chang, J.-R. Ind. Eng. Chem. Res. 1995,
34, 4284.
References and Notes (40) Moulder, J. F.; Stickle, W. F.; Sobol, P. E.; Bomben, K. D. In
Handbook of X-ray Photoelectron Spectroscopy; Chastain, J. , Ed.; Perkin-
(1) He, J.; Ichinose, I.; Kunitake, T.; Nako, A.; Shirashi, Y.; Toshima, Elmer Corporation: Eden Prairie, MN, 1992.
N. J. Am. Chem. Soc. 2003, 125, 11034. (41) Militello, M. C.; Simko, S. J. Surf. Sci. Spectra 1997, 3 (4), 387.
(2) Kamat, P. V.; Huehn, R.; Nicolaescu, R. J. Phys. Chem. B 2002, (42) Martensson, N.; Nyholm, R.; Johansson, B. Phys. ReV. Lett. 1980,
106, 788. 45 (9), 754.
(3) Sun, S.; Murray, C. B.; Weller, D.; Folks, L.; Moser, A. Science (43) Kim, K. S.; Gossmann, A. F.; Winograd, N. Anal. Chem. 1974, 46
2000, 287, 1989. (2), 197.
(4) Frenkel, A. I.; Hills, C. W.; Nuzzo, R. G. J. Phys. Chem. B 2001, (44) Toshima, N.; Yonezawa, T.; Kushihashi, K. J. Chem. Soc. Faraday
105 (51), 12689. Trans. 1993, 89 (14), 2537.
(5) Roucoux, A.; Schulz, J.; Patin, H. Chem. ReV. 2002, 102, 3757. (45) Hwang, B.-J.; Sarma, L. S.; Chen, J.-M.; Chen, C.-H.; Shih, S.-C.;
(6) Sinfelt, J. H. J. Catal. 1973, 29, 308. Wang, G.-R.; Liu, D.-G.; Lee, J.-F.; Tang, M.-T. J. Am. Chem. Soc. 2005,
(7) Directive of the European Parliament and of the Council on the 127, 11140.
Quality of Petrol and Diesel Fuels, Brussels, COM 241 final, 2001.
(8) Cooper, B. H.; Donnis, B. B. L. Appl. Catal., A 1996, 137, 203. JP805465X