Peng Tang1, Hyeon Jeong Lee1, Kevin Hurlbutt1, Po-Yuan Huang1, Sudarshan Narayanan1, Chenbo Wang2, Diego Gianolio3, Rosa Arrigo4, Jun Chen1, Jamie H Warner5,6, Mauro Pasta1,2. 1. Department of Materials, University of Oxford, Parks Road, Oxford OX1 3PH, United Kingdom. 2. Oxford Suzhou Centre for Advanced Research, 388 Ruoshui Road, Suzhou 215123, Jiangsu Province, P. R. China. 3. Diamond Light Source Limited, Harwell Science and Innovation Campus, Didcot, Oxfordshire OX11 0DE, United Kingdom. 4. School of Science, Engineering and Environment, University of Salford, Manchester M5 4WT, United Kingdom. 5. Materials Graduate Program, Texas Materials Institute, The University of Texas at Austin, 204 East Dean Keeton Street, Austin, Texas 78712, United States. 6. Walker Department of Mechanical Engineering, The University of Texas at Austin, 204 East Dean Keeton Street, Austin, Texas 78712, United States.
Abstract
Platinum single-site catalysts (SSCs) are a promising technology for the production of hydrogen from clean energy sources. They have high activity and maximal platinum-atom utilization. However, the bonding environment of platinum during operation is poorly understood. In this work, we present a mechanistic study of platinum SSCs using operando, synchrotron-X-ray absorption spectroscopy. We synthesize an atomically dispersed platinum complex with aniline and chloride ligands onto graphene and characterize it with ex-situ electron microscopy, X-ray diffractometry, X-ray photoelectron spectroscopy, X-ray absorption near-edge structure spectroscopy (XANES), and extended X-ray absorption fine structure spectroscopy (EXAFS). Then, by operando EXAFS and XANES, we show that as a negatively biased potential is applied, the Pt-N bonds break first followed by the Pt-Cl bonds. The platinum is reduced from platinum(II) to metallic platinum(0) by the onset of the hydrogen-evolution reaction at 0 V. Furthermore, we observe an increase in Pt-Pt bonding, indicating the formation of platinum agglomerates. Together, these results indicate that while aniline is used to prepare platinum SSCs, the single-site complexes are decomposed and platinum agglomerates at operating potentials. This work is an important contribution to the understanding of the evolution of bonding environment in SSCs and provides some molecular insights into how platinum agglomeration causes the deactivation of SSCs over time.
Platinum single-site catalysts (SSCs) are a promising technology for the production of hydrogen from clean energy sources. They have high activity and maximal platinum-atom utilization. However, the bonding environment of platinum during operation is poorly understood. In this work, we present a mechanistic study of platinum SSCs using operando, synchrotron-X-ray absorption spectroscopy. We synthesize an atomically dispersed platinum complex with aniline and chloride ligands onto graphene and characterize it with ex-situ electron microscopy, X-ray diffractometry, X-ray photoelectron spectroscopy, X-ray absorption near-edge structure spectroscopy (XANES), and extended X-ray absorption fine structure spectroscopy (EXAFS). Then, by operando EXAFS and XANES, we show that as a negatively biased potential is applied, the Pt-N bonds break first followed by the Pt-Cl bonds. The platinum is reduced from platinum(II) to metallic platinum(0) by the onset of the hydrogen-evolution reaction at 0 V. Furthermore, we observe an increase in Pt-Pt bonding, indicating the formation of platinum agglomerates. Together, these results indicate that while aniline is used to prepare platinum SSCs, the single-site complexes are decomposed and platinum agglomerates at operating potentials. This work is an important contribution to the understanding of the evolution of bonding environment in SSCs and provides some molecular insights into how platinum agglomeration causes the deactivation of SSCs over time.
Catalysts
for the electrochemical hydrogen evolution reaction (HER)
are desirable to produce hydrogen from clean sources like solar energy.[1,2] Platinum is among the most active HER catalysts known in acidic
media.[3] Platinum is a heterogeneous catalyst
with the HER occurring only on the metal surface.[4] This can be prohibitively inefficient because platinum
is so rare and expensive.[5]Single-site
catalysts (SSCs) are an emerging type of heterogeneous
catalysts in which isolated metal sites are supported on suitable
substrates.[6,7] Each metal site is well separated from other
metal atoms by a distance greater than the length of the metallic
chemical bond.[8,9] Platinum SSCs can be prepared
by synthesizing a platinum complex and adsorbing the whole complex
to a suitable substrate.[10−12] They have extremely high atom
utilization because every catalytic atom is exposed to the reactants
and available for reaction.[13,14] We choose to distinguish
these structures from single-atom catalysts (SACs), which have also
been shown to promote a broad range of catalytic reactions,[15−17] since the Pt SSCs we investigate are complexes of several atoms
surrounding the catalytically active metal atom.[18,19]Despite the promise of platinum SSCs for the HER, the bonding
environment
of platinum in these devices before and during operation is poorly
understood.[20−23] Establishing the structure–property relationship between
the platinum bonding and the electrochemical performance of the catalyst
is critically important to enabling design of tailored, high-performing
platinum SSCs.[24,25] Furthermore, platinum SSCs for
the HER have poor long-term stability, but the cause of their degradation
is poorly understood.[26] Operando spectroscopy
methods are a power tool to determine at the molecular level why these
catalysts deactivate over time.[27−30]In this work, we examine the model SSC system
of a platinum complex
with aniline and chloride ligands (trans-dianilinedichloroplatinum(II))
adsorbed onto graphene. Graphene is an excellent substrate because
of its high electronic conductivity, chemical stability[31] and the good contrast it provides with platinum
in electron microscopy.[32] At the same time,
aniline is stabilized via adsorption onto graphene through π–π
stacking interactions with the molecule’s aromatic ring.[10] Furthermore, aniline readily complexes with
platinum, and the molecule’s large size provides a steric hindrance
that can prevent platinum agglomeration.[33,34]We characterize the physical structure of the as-synthesized
SSC
with ex-situ scanning transmission electron microscopy (STEM) and
X-ray diffractometry (XRD). We examine the SSC’s oxidation
state and bonding environment with X-ray photoelectron spectroscopy
(XPS), X-ray absorption near-edge structure spectroscopy (XANES),
and extended X-ray absorption fine structure spectroscopy (EXAFS).
Finally, we use chronoamperometry with operando EXAFS and XANES, fitted
with structures calculated using density-functional theory (DFT),
to study how platinum’s bonding environment and coordination
number change as a function of applied potential during the HER. This
work is a good demonstration of using operando X-ray spectroscopy
techniques to probe the atomic-level structural evolution of platinum
SSCs for the HER during operation. This is an important contribution
to the understanding of the synthesis and stability of platinum SSCs.
Results
and Discussion
Synthesis and Physicochemical Characterization
The
platinum SSC is prepared according to a modified impregnation method
described in detail in the Experimental Procedures section in the Supporting Information and outlined in Figure a.
Figure 1
(a) Schematic illustration
of the synthesis process (5 steps) for
Pt SSC (gray, Pt; green, Cl; blue, N; black, carbon; white, H). Step
1: graphene nanosheets are grown by chemical vapor deposition (CVD)
on copper. Step 2: copper foil is etched after spin coating with poly(methyl
methacrylate) (PMMA) on top. Step 3: Thin-layer graphene coated with
PMMA is then transferred onto a polished glassy carbon (GC) electrode.
Step 4: after washing the PMMA with acetone and heating to 250 °C
to remove trace organic contaminants, aniline is drop cast onto the
graphene-coated GC electrode and dried. Step 5: chloroplatinic acid
in ethanol is then drop cast and annealed under an argon atmosphere.
This sample is referred to as the (Pt/Ani/Ar) sample. Control sample
(Pt/Ar) is prepared by the same method but without aniline. (b) X-ray
diffraction patterns for pristine GC and a (Pt/Ani/Ar) sample on GC;
low-magnification to high-magnification atomic resolution HAADF-STEM
images of (Pt/Ani/Ar) (c, d) and (Pt/Ar) (e, f).
(a) Schematic illustration
of the synthesis process (5 steps) for
Pt SSC (gray, Pt; green, Cl; blue, N; black, carbon; white, H). Step
1: graphene nanosheets are grown by chemical vapor deposition (CVD)
on copper. Step 2: copper foil is etched after spin coating with poly(methyl
methacrylate) (PMMA) on top. Step 3: Thin-layer graphene coated with
PMMA is then transferred onto a polished glassy carbon (GC) electrode.
Step 4: after washing the PMMA with acetone and heating to 250 °C
to remove trace organic contaminants, aniline is drop cast onto the
graphene-coated GC electrode and dried. Step 5: chloroplatinic acid
in ethanol is then drop cast and annealed under an argon atmosphere.
This sample is referred to as the (Pt/Ani/Ar) sample. Control sample
(Pt/Ar) is prepared by the same method but without aniline. (b) X-ray
diffraction patterns for pristine GC and a (Pt/Ani/Ar) sample on GC;
low-magnification to high-magnification atomic resolution HAADF-STEM
images of (Pt/Ani/Ar) (c, d) and (Pt/Ar) (e, f).High-angle annular dark-field scanning transmission electron microscopy
(HAADF-STEM) of (Pt/Ani/Ar) and (Pt/Ar) are shown in Figure c and 1d and 1e and 1f,
respectively. (Pt/Ani/Ar) contains atomically dispersed platinum with
an average areal number density of about 0.5 nm–2 (Supplementary Figures S1 and S2). Some
aggregates of platinum are present;[34] these
are generally of size 1 nm or smaller (Supplementary Figure S2a). (Pt/Ar) mainly contains larger aggregates and
nanoparticles that are up to about 4 nm across (Supplementary Figure S2b).The XRD patterns for the
glassy carbon (GC) electrode and for (Pt/Ani/Ar)
are shown in Figure b.[35] The two patterns match well. In particular,
the pattern for (Pt/Ani/Ar) lacks the intense peaks at about 39.8°,
46.2°, and 68.5° expected for the (111), (200), and (220)
reflections in crystalline platinum (PDF# 04-0802). This implies that
the HAADF-STEM images, which show no platinum nanoparticles in (Pt/Ani/Ar),
are representative of the bulk sample. The STEM images show that treatment
with aniline promotes the formation of more atomically dispersed platinum.
Ex-Situ XPS and XAS
The platinum oxidation state was
investigated by XPS, and the spectra for (Pt/Ani/Ar) and (Pt/Ar) are
shown in Figure a.
Two standards are also shown: ammonium hexachloroplatinate(IV) ((NH4)2PtCl6) and platinum(II) chloride (PtCl2). The platinum 4f7/2 peak in PtCl2 has
a binding energy of 72.9 eV, consistent with the +2 oxidation state.
The (NH4)2PtCl6 has a higher binding
energy of 75.2 eV, which is also expected for its high oxidation state
of +4. The spectrum for (Pt/Ani/Ar) is well fitted by a 74% contribution
from a peak centered at 72.8 ± 0.1 eV and a 26% contribution
from a peak centered at 74.6 ± 0.1 eV (Figure a and Supplementary Figure S5b). These contributions likely arise from platinum(II) and
platinum(IV),[36] respectively, indicating
that the sample is a mixture of the two oxidation states.[24,25,37] (Pt/Ar) is well fitted by a single
peak centered at 73.0 ± 0.1 eV arising from the +2 oxidation
state (Supplementary Figure S6).[38,39] These oxidation-state assignments are consistent with the XANES
spectra shown in Figure b. Here, reference spectra for PtCl2 (+2) and (NH4)2PtCl6 (+4) have white line peak positions
at 11 565.8 and 11 567.7 eV, respectively. The spectrum
for (Pt/Ani/Ar) has a single peak centered at 11 566.7 eV.
This is expected of a mixture of platinum(II) and platinum(IV) and
is consistent with the XPS data.[37,38]
Figure 2
XPS spectra
of Pt 4f (a) and XANES spectra (b) recorded at the
Pt L3 edge for the synthesized Pt SSC and control sample
(Pt/Ar) compared with reference standards in oxidation states of platinum(II)
and platinum(IV). Vertical dashed red and blue lines in a indicate
the peak positions of Pt7/2 at the oxidation state of platinum(II)
and platinum(IV). (c) Corresponding k3-weighted Fourier-transformed Pt L3-edge EXAFS spectra
of samples in b and standard Pt foil. Vertical dashed lines or area
show the positions of Pt–N (red), Pt–Cl (blue), and
Pt–Pt (green) peaks.
XPS spectra
of Pt 4f (a) and XANES spectra (b) recorded at the
Pt L3 edge for the synthesized Pt SSC and control sample
(Pt/Ar) compared with reference standards in oxidation states of platinum(II)
and platinum(IV). Vertical dashed red and blue lines in a indicate
the peak positions of Pt7/2 at the oxidation state of platinum(II)
and platinum(IV). (c) Corresponding k3-weighted Fourier-transformed Pt L3-edge EXAFS spectra
of samples in b and standard Pt foil. Vertical dashed lines or area
show the positions of Pt–N (red), Pt–Cl (blue), and
Pt–Pt (green) peaks.The EXAFS spectra plotted in Figure c (along with those of the standards) give information
about the platinum bonding environment. The standard samples of PtCl2 and (NH4)2PtCl6 have one
dominant peak in the range from 1.5 to 2.3 Å with the center
at 1.96 Å, suggesting that this peak can be associated with Pt–Cl
bonding. The most intense peak in the spectrum for (Pt/Ani/Ar) is
at 1.96 Å with a shoulder at 1.66 Å. The peak at 1.96 Å
should arise from the presence of Pt–Cl bonds. This shoulder
peak at 1.66 Å is indicative of platinum–nitrogen bonding.[40,41] Finally, there is the absence of obvious Pt–Pt bonding at
2.73 Å (green curve in Figure c), which also suggests that the platinum atoms are
mainly isolated from each other. We therefore interpret these data
to mean that the aniline and chloride coordinate to the platinum to
form a complex molecule (Supplementary Figure S7).These XPS and XANES data indicate that (Pt/Ani/Ar)
and (Pt/Ar)
are reduced in situ from platinum(IV) (the oxidation state of platinum
in the precursor H2PtCl6) to platinum(II). The
cause of the reduction of Pt(IV) to Pt(II) is the ethanol used as
a solvent to prepare the H2PtCl6 solution (Supplementary Figure S8).[42] However, in (Pt/Ani/Ar), the platinum forms a complex with
aniline (as confirmed by the bonding data from the EXAFS spectra).
This is the reason that (Pt/Ani/Ar) is only partially reduced with
platinum(IV) still present. The STEM, XPS, XANES, and EXAFS data suggest
that the aniline-treated sample contains atomically dispersed platinum
in the form of molecular complexes (Pt single-site complexes) physisorbed
onto the graphene substrate.
Operando XAS Measurements
Operando
hard XAS at the
platinum L3 edge was performed at different potentials
to investigate the evolution of the platinum single-site complex’s
oxidation state and chemical bonding environment during the HER.[23,30,43] A customized three-electrode
cell setup (Supplementary Figure S10) was
mounted to investigate the aniline-treated sample (Pt/Ani/Ar). The
chronoamperometric performance of (Pt/Ani/Ar) at various potentials
is shown in Figure a (with an enlarged view shown in Figure b). The areal current densities are shown
for at least 1000 s of operation at each of 10 voltages: 0.4, 0.3,
0.2, 0.1, 0.06, 0.02, 0, −0.02, −0.04, and −0.06
V versus the reversible hydrogen electrode (RHE). The open-circuit
voltage (OCV) is 0.6 V versus RHE. None of the tests at a positive
potential have an areal current density greater than 0.1 mA cm–2. This is to be expected since the HER is only thermodynamically
favorable below 0 V versus RHE. Indeed, the data for the test at 0
V show a minimal areal current density until about 600 s (out of approximately
1200 s in total) when the current jumps suddenly and then gradually
increases to about −0.3 mA cm–2. The areal
current densities for the subsequent potentials of −0.02, −0.04,
and −0.06 V versus RHE are −0.5, −0.9, and −1.5
mA cm–2, respectively. The Tafel slope, calculated
using the 4 negative potentials, is 60 mV dec–1.
This is comparable to the Tafel slope of platinum nanoparticles for
the HER.[44,45]
Figure 3
(a) Stable chronoamperometric curves of Pt single-site
complex
at various applied potentials. (b) Enlarged red dashed area in a shows
the chronoamperometric curves at a current density smaller than 0.1
mA cm–2. (c) Operando XANES spectra recorded at
the Pt L3 edge of the synthesized Pt single-site complex
at different working potentials for the HER in 0.5 M H2SO4. (Insets) Energy shifts of the Pt L3 absorption
edge (left) and white-line peak (right). (d) Corresponding k2-weighted Fourier-transformed EXAFS spectra
(magnitude component). Vertical dashed lines or area show the Pt–N
(red), Pt–Cl (orange), and Pt–Pt (gray) peaks.
(a) Stable chronoamperometric curves of Pt single-site
complex
at various applied potentials. (b) Enlarged red dashed area in a shows
the chronoamperometric curves at a current density smaller than 0.1
mA cm–2. (c) Operando XANES spectra recorded at
the Pt L3 edge of the synthesized Pt single-site complex
at different working potentials for the HER in 0.5 M H2SO4. (Insets) Energy shifts of the Pt L3 absorption
edge (left) and white-line peak (right). (d) Corresponding k2-weighted Fourier-transformed EXAFS spectra
(magnitude component). Vertical dashed lines or area show the Pt–N
(red), Pt–Cl (orange), and Pt–Pt (gray) peaks.The average of eight XANES spectra, collected during
operation
at each potential, is shown in Figure c. These data show that the oxidation state of platinum
decreases from a mixture of +4 and +2 to 0. When the potential is
decreased from 0.6 (OCV) to 0 V versus RHE, the positions of both
the white line (red arrow and right inset in Figure c) and the absorption edge (left inset in Figure c) in the XANES spectra
show a negative shift of 0.4 eV. Those shifts suggest a continuous
decrease of platinum valence from a mixture of platinum(II) and platinum(IV)
to metallic platinum.[23,46]The bonding environment
of the platinum also changes as the applied
potential varies, as shown in the EXAFS spectra in Figure d. As the potential is decreased
from OCV to 0, the peaks at 1.66 and 1.96 Å, arising from Pt–N
and Pt–Cl bonds, respectively, both decrease in intensity.
However, the peak arising from Pt–N (blue vertical line) decreases
faster than that arising from Pt–Cl (orange vertical line).
This indicates that as the potential is decreased from OCV and while
the oxidation state of the platinum falls to 0, the aniline molecules
are dissociating from the complex. The chloride ions are also dissociating
from the platinum atoms but to a lesser extent compared to the aniline
molecules. At 0 V, the SSC has two distinct EXAFS spectra (yellow
and olive curves in Figure d and isolated in Figure S12):
one at a low areal current density (−0.1 mA cm–2) before the onset of the HER and one at a higher areal current density
(−0.3 mA cm–2) after the onset of the HER.
Before the onset of the HER, the bonding environment still shows appreciable
Pt–Cl bonding and some Pt–N bonding. After 600 s and
the beginning of hydrogen evolution, almost all Pt–N bonding
character is lost and the Pt–Cl bonding feature decreases significantly.
The EXAFS spectrum is instead dominated by Pt–Pt bonding. The
rapid breaking of Pt–N and Pt–Cl bonds at 0 V may be
caused by hydrogen molecules aggregated on the SSC. It has been experimentally
shown that the diffusion and agglomeration of platinum atoms can be
induced by exposure to a hydrogen atmosphere.[47−50] Since more reduced platinum structures
after 600 s can boost the evolution of hydrogen by lowering the Gibbs
free energy of hydrogen adsorption in acidic electrolytes,[23,51] the sudden increase in current density here (the yellowish line
discussed in Figure a) can probably be justified accordingly.
Structure–Electrochemistry
Correlation
The previous
section elucidated the evolution of the Pt single-site complex into
metallic nanoclusters at potentials below 0 V; the details of this
transformation need to be quantitatively examined together with the
catalytic activity. The effect of applied potential on both the areal
current density and the platinum oxidation state in the SSC is shown
in Figure a. The height
of the XANES normalized white lines is summarized (black plot in Figure a) to quantify the
evolution of platinum’s oxidation state and the electron occupancy
of platinum’s 5d orbitals.[52,53] The white-line
height decreases precipitously from OCV (0.60 V) to about 0 V versus
RHE. This is consistent with the complete filling of the platinum
5d valence orbitals and the formation of metallic platinum.[23] At the same potential, the corresponding areal
current density increases precipitously as the HER becomes thermodynamically
favorable, although they are still relatively low (from −0.001
to −0.08 mA cm–2). There is an obvious drop
in white-line height at around 0 V, and the white-line height stabilizes
at around 1.30 after 0 V, indicating the oxidation state of platinum
remains as 0 at potentials below 0 V. Corresponding to the decrease
in oxidation state, a significant increase in current density is also
recorded at 0 V (red plot in Figure a).After the increase, a linear relationship between
the base-10 logarithm of the areal current density and the potentials
is presented.
Figure 4
(a) Amplitude of the normalized white-line peak of the
Pt L3-edge operando XANES spectra (black) and corresponding
catalytic
activity (red) at different potentials. Log(j) is
calculated using a relatively stable current density at the end of
each CA test. (b) Coordination number of Pt–N, Pt–Cl,
and Pt–Pt fitting paths for operando EXAFS data of the Pt single-site
complex at selected potentials by 0 V. Error bars in b represent the
uncertainties of the fitting results. (c) Contribution of different
paths (Pt–N, Pt–Cl, and Pt–Pt) to fit the corresponding
EXAFS spectra. Details of the fitting results shown in K-space and R-space are found in Figure S13. (d) Proposed schematic model for the evolution
of the (Pt/Ani/Ar) sample from a single-site complex to subnanoclusters
based on the operando EXAFS analysis (gray, Pt; green, Cl; blue, N;
black, carbon; white, H).
(a) Amplitude of the normalized white-line peak of the
Pt L3-edge operando XANES spectra (black) and corresponding
catalytic
activity (red) at different potentials. Log(j) is
calculated using a relatively stable current density at the end of
each CA test. (b) Coordination number of Pt–N, Pt–Cl,
and Pt–Pt fitting paths for operando EXAFS data of the Pt single-site
complex at selected potentials by 0 V. Error bars in b represent the
uncertainties of the fitting results. (c) Contribution of different
paths (Pt–N, Pt–Cl, and Pt–Pt) to fit the corresponding
EXAFS spectra. Details of the fitting results shown in K-space and R-space are found in Figure S13. (d) Proposed schematic model for the evolution
of the (Pt/Ani/Ar) sample from a single-site complex to subnanoclusters
based on the operando EXAFS analysis (gray, Pt; green, Cl; blue, N;
black, carbon; white, H).EXAFS curve-fitting analysis on the platinum’s EXAFS spectrum
at OCV allows for the determination of a plausible molecular structure
for the platinum complex adsorbed onto graphene. Considering eight
possible platinum complexes described by (C6H5NH2)ClPt for x = 1 or 2 and y =
1, 2, 3, or 4, the Pt–N and Pt–Cl bond lengths (about
2.05 and 2.30 Å, respectively) determined by density-functional
theory modeling of the isolated molecular complexes are the starting
point for the analysis. The operando EXAFS data are best described
by the structural model for trans-dianilinedichloroplatinum(II),
(C6H5NH2)2Cl2Pt(II), within the fitting range (k-range, 3–12.5
Å–1, R-range, 1.1–2.5
Å, misfit factor 0.8%). This structure is consistent with the
XPS and XANES spectra of the (Pt/Ani/Ar) sample (most of the Pt is
at an oxidation of +2) and is similar to the previously reported crystal
(C6H5NH2)2PtCl2 complex prepared by mixing aniline with H2PtCl6.[54,55]Performing EXAFS curve-fitting analysis
(platinum’s first
(1.5–2.1 Å) and second (2.3–3.1 Å) coordination
shells) on the spectra collected at different potentials results in
the curves in Figure b and Table S1, which show the average
number of platinum, chlorine, and nitrogen atoms coordinated to each
platinum atom (i.e., the coordination number) as a function of applied
potential. Specifically, Figure c highlights the evolution of the EXAFS spectra while
decreasing potential in the R space fitted with the
contributions of the Pt–N, Pt–Cl, and Pt–Pt paths.
At OCV, the coordination number for both chlorine and nitrogen is
about 2.5. As the potential is lowered, the coordination number for
nitrogen decreases to about 1.0 by a potential of 0.1 V versus RHE.
It continues to decrease to about 0.5 by 0 V (but before the onset
of the HER). The chlorine coordination number stays higher than the
nitrogen coordination number. This indicates that the aniline molecules
dissociate from the complex before the chloride ions do, likely due
to the more ionic character and higher bond energy of the Pt–Cl
bonds compared to the Pt–N bonds.[56] The platinum coordination number increases as the potential is lowered,
ultimately reaching a value of about 7.0 by 0 V versus RHE. This indicates
the agglomeration of platinum as the potential is lowered (Supplementary Figure S14), which is also in line
with the reduction of the oxidation state of platinum described in Figure a.[57]Together these data elucidate how the platinum bonding
environment
in an SSC changes as a function of applied potential during operation,
as described schematically in Figure d. As the potential is lowered from OCV, the dianilinedichloroplatinum(II)
decomposes as the aniline dissociates from the platinum, followed
by the dissociation of the chloride ions. Simultaneously, the 5d valence
orbitals of platinum are filled. By the onset of the HER at 0 V versus
RHE, the dianilinedichloroplatinum(II) complexes have completely dissociated
and the metallic platinum(0) single atoms agglomerate into nanoclusters
which catalyze the HER.
Conclusions
In this work, we use
operando, synchrotron X-ray absorption spectroscopy
to characterize how the oxidation state and bonding environment of
platinum changes in an SSC for the HER at the molecular level. First,
we show successful synthesis of a platinum SSC, confirming the platinum
is atomically dispersed by STEM and XRD. We observe a mixture of platinum(II)
and platinum(IV) by XPS and XANES, and we confirm platinum coordination
to aniline by ex-situ EXAFS. We then probe the oxidation state and
molecular structure as a function of applied potential using operando
EXAFS and XANES with synchrontron radiation. We find that as the potential
is lowered from OCV the platinum oxidation state decreases, breaking
bonds between platinum atoms and aniline ligands first. The Pt–Cl
bonds also break but to a lesser extent than the Pt–N bonds.
Finally, once the onset of the HER at 0 V versus RHE is reached, the
platinum is completely reduced and metallic platinum atoms agglomerate
to form catalytic nanoclusters. Our work suggests that many single-site
catalysts that achieve atomically dispersed platinum by ligand coordination
do not remain atomically dispersed during device operation. This demonstrates
that researchers should be extremely careful when assuming that the
bonding environment of the as-synthesized SSC will remain the same
during device operation. Furthermore, this operando study establishes
two important structure–property relationships for aniline-complex
platinum SSCs. First, platinum exists as metallic platinum (that is,
in the oxidation state 0) while catalyzing the HER. Second, the molecular
structure of the platinum complexes is lost due to the reduction of
the platinum. These findings highlight the necessity for operando
techniques to follow the dynamic behavior of SSCs during operation
and provide critical new insight into why SSCs deactivate over time.
Such experiments can also provide an important contribution to an
atomic understanding of the structural evolution of SSC systems as
well as SACs for other critical electrochemical reactions, such as
the oxygen reduction reaction and carbon dioxide reduction reaction.
These results and the operando X-ray spectroscopy method used in this
study will accelerate the development of stable, high-performance
platinum SSCs for clean hydrogen production.
Authors: Grazia Malta; Simon A Kondrat; Simon J Freakley; Catherine J Davies; Li Lu; Simon Dawson; Adam Thetford; Emma K Gibson; David J Morgan; Wilm Jones; Peter P Wells; Peter Johnston; C Richard A Catlow; Christopher J Kiely; Graham J Hutchings Journal: Science Date: 2017-03-31 Impact factor: 47.728
Authors: Abdulsattar H Ghanim; Jonathan G Koonce; Bjorn Hasa; Alan M Rassoolkhani; Wei Cheng; David W Peate; Joun Lee; Syed Mubeen Journal: Front Chem Date: 2018-11-06 Impact factor: 5.221
Authors: Can Li; Zheng Chen; Hong Yi; Yi Cao; Lei Du; Yidong Hu; Fanpeng Kong; Richard Kramer Campen; Yunzhi Gao; Chunyu Du; Geping Yin; Igor Ying Zhang; Yujin Tong Journal: Angew Chem Int Ed Engl Date: 2020-05-20 Impact factor: 15.336