applications of xanes and exafs to nanoparticle studiesable to provide structural information on nps...
Post on 12-Jun-2020
2 Views
Preview:
TRANSCRIPT
Office of Institutional Research, Planning & Effectiveness
Applications of XANES and EXAFS to nanoparticle studies
Janis Timoshenko
https://www.bnl.gov/chemistry/SDAN/ Department of Materials Science and Chemical Engineering, Stony Brook University, Stony Brook, NY 11794
XAFS Short Course 2016
Office of Institutional Research, Planning & Effectiveness
NPs properties can be very different from those of bulk materials: • Large surface-to-volume
ration • Structure relaxation • Different electronic
structure
Properties of NPs can be tailored by changing • NPs size • NPs shape • NPs composition • Intra-particle distribution
of different atoms • NPs support material • Ligands & adsorbates • ...
It is important to correlate atomic structure of NPs with their properties!
Why nanoparticles?
11/9/2016 XAFS short course 2016 2
J. Kleis, et al, Cat. Letters 141, 1067 (2011)
Alaqad and Saleh, J Environ Anal Toxicol 2016, 6:4
Office of Institutional Research, Planning & Effectiveness
Why XAS?
11/9/2016 XAFS short course 2016 3
Bulk
Nano
Bulk
Nano
TEM pictures of Thiol-protected Pd NPs:
XRD spectra of nanosized tungstates:
Sun et al, Langmuir, Vol. 22, No. 2, 2006
Anspoks et al, Solid State Commun. 183 (2014) 22.
XAS – one of a very few experimental techniques that are able to provide structural information on NPs structure : • Diffraction methods are not
applicable in case of small NPs ( < 3 - 4 nm)
• Electron-microscopy does not provide sufficient sensitivity to resolve structure in details
• Total scattering/PDF analysis provides information, complimentary to XAS, but lacks chemical sensitivity.
Office of Institutional Research, Planning & Effectiveness
Why XAS?
11/9/2016 XAFS short course 2016 4
Structure of NPs can change in situ: it is important therefore to study NPs in reaction conditions! Due to high-penetration depth of X-rays, complex sample environments can be used in synchrotron XAS experiments. XAS thus is and ideal technique for in situ and operando investigations Time-dependent XAS measurements are now enabled by high intensities of modern synchrotron sources.
Office of Institutional Research, Planning & Effectiveness
XANES and EXAFS
11/9/2016 XAFS short course 2016 5
EXAFS analysis: • Nearest neighbors:
types and numbers
• Bond length distributions
• Disorder effects
• Atomic dynamics
XANES analysis: • Oxidation state of
absorbing atom
• Local electronic structure
• Local symmetry
6900 7200 7500 7800 8100 8400
-1.0
-0.5
0.0
µ(E)
Energy E (eV)
EXAFS XANES hν
Office of Institutional Research, Planning & Effectiveness XAS studies of nanoparticles
Office of Institutional Research, Planning & Effectiveness
EXAFS
11/9/2016 XAFS short course 2016 7
Office of Institutional Research, Planning & Effectiveness
EXAFS in NPs
11/9/2016 XAFS short course 2016 8
W Co O
Bulk
Nano • NPs large surface-to-volume ratio →
reduced coordination numbers • Large disorder → large MSRD factors,
reduction of EXAFS amplitude • Limited contribution from distant
shells • Structure relaxation → changes in
interatomic distances • Asymmetric bond lengths
distributions
J. Timoshenko, Phys. Status Solidi A 212 (2015) 265-273.
Office of Institutional Research, Planning & Effectiveness
Size & shape
A.I. Frenkel et al J. Phys. Chem. B 105, 12689 (2001). J. Matos et al, Phys. Chem. Chem. Phys., 14 11457 (2012)
S2
S3
S4
S1
B. R. Cuenya et al, J. Am. Chem. Soc. 132, 8747 (2010) B. Roldan Cuenya, AIF, et al, Phys. Rev. B. 82, 155450 (2010)
One of the quantities, commonly accessible from EXAFS spectra, is average coordination number of absorbing atoms. Due to size effect, coordination numbers for nanoparticles are typically smaller than for bulk materials, and are directly linked to nanoparticle size and shape. NPs of different shapes will have different surface–to-volume ratios, hence different average coordination numbers and EXAFS spectra.
11/9/2016 XAFS short course 2016 9
Office of Institutional Research, Planning & Effectiveness
Size & shape
Relation between NPs size and average coordination numbers can be obtained either analytically or numerically by modelling 3D geometries of clusters of different sizes and shapes
11/9/2016 XAFS short course 2016 10
Cuboctahedron
Icosahedron
13, 55, 147, 309, 561…
13, 55, 147, 309, 561…
13, 38, 79, 140, 225, 338, 483, 664…
Truncated octahedron
HCP
13, 26, 57, 89, 153, 214, 323, 421, 587
D. Glasner, A. Frenkel, AIP Conf. Proc. 882, 746-748 (2007).
Truncated (111) Cuboctahedron
10, 37, 92, 185, 326, 525,…
Truncated (001) Cuboctahedron
9, 34, 86, 175, 311, 504…
( )3111510
1352423
2co1 +++
++=
LLLLLLn
bulk1
3
1 161
431 n
Rr
Rrn
+
−≈
S. Calvin, et al, JAP 778, 94 (2003)
Montejano-Carrizales, et al, Nanostruct. Mater. 8, 269 (1997)
Office of Institutional Research, Planning & Effectiveness
Size & shape
Information on coordination number for the 1st coordination shell only is not sufficient to find the size and shape of NPs unambiguosly. Average coordination numbers for more distant coordination shells are more challenging to extract from EXAFS analysis, but they usually are more sensitive to size and shape effects.
11/9/2016 XAFS short course 2016 11
0 100 200 300 400 500 600 7005
6
7
8
9
10
11
12
CO309 or TO225 or IH147
EXAFSN1
CO55 or TO38
IH13
CO13 = TO13
Bulk (fcc) N1 = 12
Coordination numbers in CO, IH and TO clusters
Number of Atoms
CO IH TO
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 150
2
4
6
8
10
12
14
16
18
20
22
24Bulk 3NN and 5NN
Bulk 2NN
Bulk 1NN and 4NN
Coor
dina
tion
num
ber
L
1NN 2NN 3NN 4NN 5NN
0 10 20 30 40 50 60 70 80
Cluster diameter, Å
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 150
2
4
6
8
10
12
14
16
18
20
22
24
No overlap between 1NN and 2NN
1NN 2NN 3NN 4NN 5NN
Coor
dina
tion
num
ber
L
0 10 20 30 40 50 60 70 80
Bulk 2NN
Bulk 1NN and 4NN
Bulk 3NN and 5NN
Cluster diameter, Å
Spherical Hemi-spherical
Office of Institutional Research, Planning & Effectiveness
Crystallographic structure
J.A Varnell et al, Nature Commun. 7 (2016).
L. Menard et al, J. Phys. Chem. B. 110, 14564(2006)
Knowledge of coordination numbers, especially of those for more distant coordination shells and of multiple-scattering effects, allows one also to distinguish between different crystallographical structures, e.g., between close-packed and icosohedral structures, between fcc and bcc structures.
11/9/2016 XAFS short course 2016 12
Office of Institutional Research, Planning & Effectiveness
Order in heterometallic NPs
Chemical sensitivity of XAS method allows one to determine the ordering and distribution of atoms of different types with the nanoparticles of bimetallic (or more complex) materials. Again, coordination numbers obtained from EXAFS analysis can be used for this purpose. For such kind of analysis typically XAS data, obtained at absorption edges of both metals, are necessary.
A B
or + or
BMAM nn <
Random Intra-particle segregation
Inter-particle segregation
B
A
AB
AA
xx
nn
= 0=ABn
A.Frenkel, Z. Kristallogr. 222, 605 (2007) A.Frenkel, Chem. Soc. Rev. 41, 8163 (2012)
11/9/2016 XAFS short course 2016 13
A
ABAB N
Nn =
BMBAMAMM nxnxn +=
Partial coordination number: average number of specific bonds per atom Total average coordination number (can be used for size analysis)
A
AAAA N
Nn 2=
Office of Institutional Research, Planning & Effectiveness
Ordering parameter
Quantitavely, order in bimetallic alloy can be characterized by ordering parameter α. Positive (α >0) indicates tendency to clustering (segregation), while α = 0 corresponds to random arrangements of A and B atoms.
A.Frenkel, Z. Kristallogr. 222, 605 (2007) A.Frenkel, Chem. Soc. Rev. 41, 8163 (2012)
𝛼𝛼 = 1 −𝑛𝑛𝐴𝐴𝐴𝐴/𝑛𝑛𝐴𝐴𝐴𝐴
𝑥𝑥𝐴𝐴
11/9/2016 XAFS short course 2016 14
A B
A A
A
A A A
A B
B
B B
B B B
B A
A A
A A A
A
B B
B B
B
B
B A
A A A
A A
B B B B
B
031
≤≤− ABα
For 3D fcc bulk material:
𝛼𝛼𝐴𝐴𝐴𝐴 = −1 𝛼𝛼𝐴𝐴𝐴𝐴 = −1
𝛼𝛼𝐴𝐴𝐴𝐴 = 1 𝛼𝛼𝐴𝐴𝐴𝐴 = 1
𝛼𝛼𝐴𝐴𝐴𝐴 = 0 𝛼𝛼𝐴𝐴𝐴𝐴 = 0
Office of Institutional Research, Planning & Effectiveness
XAFS is sensitive to archtiecture of bimetallic NPs, and can e.g., distinguish between core-shell and random alloy structure models (which exhibit quite different catalytic properties).
S. Alayoglu et al, ACS Nano 3 3127 (2009)
Ru@Pt: 4.0nm Ru@Pt: 4.0nm
EXAFS Modeling
Bimetallic NPs: core-shell vs. random alloy
11/9/2016 XAFS short course 2016 15
Office of Institutional Research, Planning & Effectiveness
Simultaneous analysis of Multiple-edge EXAFS data provides composition and structure information on a system with multiple elements.
W. Du et al, J. Mater. Chem. 21 8887 (2011)
Ternary alloy structure
11/9/2016 XAFS short course 2016 16
Office of Institutional Research, Planning & Effectiveness
“A” edge
“B” edge
>1000 eV
Energy (ev)
As K
Au L3
µ(E
)
Au-As Au-Au As
-In
As-In, Au-As, Au-Au …
A-X1, A-X2, …
B-Y1, B-Y2, …
L. Menard et al, Phys. Rev. B 80, 064111 (2009) J. Liu et al, Chem. Mater (in press, 2016)
Overlapping edges
If no overlap:
Otherwise:
Similar situation, e.g., with Pt L3 and Au L3 edges, Pt L3 and Ir L3 edges,
The analysis is still possible!
11/9/2016 XAFS short course 2016 17
Office of Institutional Research, Planning & Effectiveness
Bond lengths
Due to surface tension, the bond lengths for atoms at metallic NPs surface can be significantly shorter than in bulk. EXAFS probes configuration average interatomic distance, averaged over all absorbing atoms (both at the surface and in the core of NPs). Thus the bond length, as estimated from EXAFS, will also be sensitive to surface-to-volume ratio. Changes in interatomic distances with accuracy better than 0.01 Å can usually be detected.
11/9/2016 XAFS short course 2016 18
W. Huang, et al, Nature Materials 7, 308 (2008)
S. I. Sanchez et al. J. Am. Chem. Soc. 131 7040 (2009)
Office of Institutional Research, Planning & Effectiveness
Bond lengths
Other factors that can influence the bond lengths in the NPs are • interactions with
ligands
• interactions with the support
• interactions with the adsorbates
• vacancies • ...
11/9/2016 XAFS short course 2016 19
A.I. Frenkel, et al, J. Chem. Phys. 123, 184701 (2005).
0 1 2 3 4 5 60.0
0.2
0.4
0.6
0.8
1.0 a)
Au-S
Au-Au
FT M
agni
tude
, Å-3
r, Å
Au foil 6:1 3:1 1:1 1:2 1:3 1:6
0.1 1 100.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
R bulk-
R nano
(Å)
Au/thiol ratio ξ
M.W. Small et al, ACS NANO 6 5583 (2012) J. Chem. Phys. 123, 184701 (2005).
Anspoks et al Phys. Rev. B 86, 174114 (2012)
Office of Institutional Research, Planning & Effectiveness
Dynamics
Damping of EXAFS oscillations is a result of static and thermal disorder. The latter is related to the density of vibrational states. By measuring temperature dependency of disorder factor σ2, one can separate effects of structural and thermal disorder and also estimate the effective bond strength constants.
S. I. Sanchez et al., J. Am. Chem. Soc. 131 7040 (2009)
A. Frenkel, J. Rehr, Phys. Rev. B, 48, 585, (1993).
Disorder:
Einstein model:
Bond Strength:
projected VDOS:
11/9/2016 XAFS short course 2016 20
Office of Institutional Research, Planning & Effectiveness
Dynamics
NPs with different sizes, in different atmospheres and on different substrates may exhibit quite different dynamic properties due to changes in static disorder and bond strengths.
S. I. Sanchez et al., J. Am. Chem. Soc. 131 7040 (2009)
11/9/2016 XAFS short course 2016 21
Under He
0
0.002
0.004
0.006
0.008
0.01
0.012
0 100 200 300 400 500 600 700 Temperature (K)
MS
RD
(Å2 )
Pt Foil Pt 2.9 nm Pt 1.1 nm Pt 0.9 nm
0
0.002
0.004
0.006
0.008
0.01
0.012
0 100 200 300 400 500 600 700 Temperature (K)
MS
RD
(Å2 )
Pt Foil Pt 2.9 nm Pt 1.1 nm Pt 0.9 nm
Under H2
Office of Institutional Research, Planning & Effectiveness
Static disorder and strain energy
Static disorder can be used to characterize the relaxation of NPs structure, influence of adsorbates and NPs support. Knowing contribution of static disorder and effective bond strength constants, one can estimate the residual strain energy of the nanoparticle.
A. I. Frenkel, et al, Solid State Commun. 99, 67 (1996).
M. Small, et al ACS Nano 6, 5583 (2012)
4 6 8 10 12
0.000
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
2.9 nm1.8 nm bulk
Pt/γ-Al2O3 in He Pt/C in H Pt/C in He Pt/γ-Al2O3 in H
Stat
ic di
sord
er in
Pt-P
t (Å2 )
Pt-Pt Coordination number
1.0 nm
222ds σσσ += 2
21 σNkU =
WTVU += )( 2
21
sNkW σ=
Lower disorder for C support and for large sizes
11/9/2016 XAFS short course 2016 22
Office of Institutional Research, Planning & Effectiveness
Surface sensitivity
Surface effects are very important for understanding of NPs properties. Differential analysis of EXAFS data allow detection of small changes in the coordination environment of absorbing atoms on the surface of NPs, that would otherwise be undetected with conventional XAS techniques. Such approach can be useful, e.g., to study the influence of adsorbates on the local structure of nanoparticle.
M. Small, et al Phys. Chem. Chem. Phys., 16 26528 (2014)
∆˗EXAFS ≈ 𝑆𝑆02𝑁𝑁𝑠𝑠𝑓𝑓 𝑘𝑘 𝑒𝑒−2𝑟𝑟1′𝜆𝜆 𝑘𝑘 𝑒𝑒
−2𝑘𝑘2σ1′2
𝑘𝑘𝑟𝑟1′2 2𝑘𝑘𝑘𝑘𝑘𝑘 sin 2𝑘𝑘𝑘𝑘1′ + 𝛿𝛿 𝑘𝑘 + 𝜋𝜋
2
11/9/2016 XAFS short course 2016 23
Office of Institutional Research, Planning & Effectiveness
Validation of theoretical models
EXAFS data can be used for validation of structure models, obtained in DFT, ab-initio MD or classical MD simulations. By comparing the calculated configuration- and time-averaged EXAFS spectra with experimental data, one can demonstrate the ability of structure models to reproduce the 3D local structure and dynamics of the material.
Structure models, obtained in ab-initio MD simulations for thiol-protected Au NPs, corresponding theoretical EXAFS spectra and their comparison with experimental data
EXAFS spectra for structure models, obtained in classical MD simulations with empirical potentials for Au NPs, and their comparison with experimental data
D. F. Yancey et al, Chem. Science 4, 2912-2921 (2013) J. Timoshenko et al, Cat. Today. (2016), http://dx.doi.org/10.1016/j.cattod.2016.05.049
11/9/2016 XAFS short course 2016 24
Office of Institutional Research, Planning & Effectiveness
3D NPs structure via RMC method
EXAFS data can also be used directly to construct a 3D structure model of the nanoparticle using reverse Monte Carlo approach. In the RMC-EXAFS approach a 3D structure model is optimized by random displacements of atoms with the aim to minimize the difference between experimental and calculated configuration-averaged EXAFS data. This approach allow to reconstruct the distribution of atoms within the NPs taking into account contributions of distant coordination shells, multiple-scattering effects and non-Gaussian shapes of bond lengths distributions.
J. Timoshenko et al, Cat. Today. (2016), http://dx.doi.org/10.1016/j.cattod.2016.05.049
11/9/2016 XAFS short course 2016 25
Office of Institutional Research, Planning & Effectiveness
Non-homogeneous samples
Often the investigated sample will contain different NPs with different sizes and compositions. Such non-homogeneity of sample may result in significant artifacts in EXAFS analysis. E.g., if sample of bimetallic random alloy NPs consists of NPs, where concentration of metal A differs significantly from particle to particle, the apparent A—A coordination number will be higher than expected for average A concentration in the sample.
11/9/2016 XAFS short course 2016 26
Theoretical coordination numbers for clusters of N = 100 atoms, calculated assuming Gaussian compositional distribution:
( )
−−∝ 2
2
2exp)(
c
xxxσ
ρ
Assume that particles are random alloys of the same size (nMM), but have inter-particle compositional disorder:
The apparent coordination number then is:
∫
∫= 1
0
1
0
)(
)()(~
dxxx
dxxnxx
nAA
AA
ρ
ρx – concentration of metal A
A. I. Frenkel, Chem. Soc. Reviews, 41, 8163 (2012)
Atomic % Ir
0 10 20 30 40 50 60 70 80 90 100N
umbe
r of P
artic
les
0
4
8
12
16
20
Office of Institutional Research, Planning & Effectiveness
Non-homogeneous samples
Similarly, if the broad distribution of NPs sizes is present in the sample, EXAFS analyis may yield an overestimated NPs size.
11/9/2016 XAFS short course 2016 27
Distribution of NPs cluster order:
The apparent coordination number then is:
A. I. Frenkel, Annu. Rev. Anal. Chem. 4 23 (2011)
Apparent coordination number as a function of the standard deviation of the size distribution.
σ= 0.15 will cause 10-20% error in CN
Office of Institutional Research, Planning & Effectiveness
Mixtures
11/9/2016 XAFS short course 2016 28
EXAFS for mixture: 𝜒𝜒 𝑘𝑘 = ∑ 𝜒𝜒𝑖𝑖 𝑘𝑘 𝑤𝑤𝑖𝑖𝑖𝑖 𝑤𝑤𝑖𝑖 - molar fraction of i-th component
A.Frenkel, Z. Kristallogr. 222, 605 (2007)
Common scenario: sample is a mixture of two species: 1) reduced monometallic nanoparticles (fraction w1) and 2)molecular precursors (fraction w2)
Obtained coordination numbers:
𝑛𝑛 𝑃𝑃𝑃𝑃 − 𝑃𝑃𝑃𝑃 = 4.0 ± 0.9 The interpretation depends critically on the independent knowledge of the
state of heterogeneity of the system! Assuming a homogeneous model: ca. four Pd––Pd and two Pd––S bonds per Pd atom → clusters should have fewer than 13 atoms! But TEM measurements indicated much larger clusters (26 Å in diameter)... Assuming a two-component system → 40% of Pd atoms are in clusters, remaining part - Pd-sulfide complexes.
Office of Institutional Research, Planning & Effectiveness
XANES
11/9/2016 XAFS short course 2016 29
Office of Institutional Research, Planning & Effectiveness
EXAFS - tool of choice for many structural studies But...
Why XANES?
Analysis gets very complex for heterogeneous systems
Disorder reduces significantly the amplitude of signal
Works well for systems with symmetric, Gaussian like bond-length distributions, but it is often not the case for nanoscale systems
XANES: • Less sensitive
to disorder
• Good for time-resolved studies
• Analysis of heterogeneous systems is well-established
• Good for electronic structure and site symmetry studies, but more complex information can also be extracted
6900 7200 7500 7800 8100 8400
-1.0
-0.5
0.0
µ(E)
Energy E (eV)
EXAFS XANES hν
Office of Institutional Research, Planning & Effectiveness
Valence state
Analysis of XANES data can be used straightforwardly to characterize changes in the electronic state and local symmetry of absorbing atoms. For instance, changes in valence state result in shift of absorption edge.
11/9/2016 XAFS short course 2016 31
Belli et al, Solid State Communications 35 355 (1980)
Office of Institutional Research, Planning & Effectiveness
Symmetry
In particular, the K-edge absorption is caused by the transition of a 1 s electron of absorbing atom from its core atomic state to a final unoccupied state. It follows dipole (Δl = ± 1) or quadrupole (Δl = 0, ± 2) selection rules, hence is sensitive to symmetry.
11/9/2016 XAFS short course 2016 32
Ti K-edge XANES in PbTiO3
Vedrinskii et al, J. Phys.: Condens. Matter 10 9561 (1998) Jaoen et al, Phys. Rev. B 75, 224115 (2007) Frenkel et al, Phys. Rev. B 71, 024116 (2005). Frenkel et al Phys. Rev. Lett. 99, 215502 (2007). Anspoks et al, Phys. Scr. 89 044002 (2014)
The 2nd pre-peak in Ti K-edge XANES in titanates originates from transitions of Ti 1s electron to the p–d hybrid orbitals (Ti 3d orbital is mixed with the O 2p orbitals due to the Ti displacement from the center of the octahedra. The intensity of the pre-peak is proportional to the squared displacement of the Ti.
Office of Institutional Research, Planning & Effectiveness
Coordination
11/9/2016 XAFS short course 2016 33
The XANES region of 1 mM [Cr3+(OH2)6]3+ (aq) and [Cr6+O4]2- (aq) are very different.
Office of Institutional Research, Planning & Effectiveness
Charge transfer
Changes in intensities of NPs XANES features upon temperature / pressure changes can be related to charge transfer processes between NPs, support and adsorbates.
11/9/2016 XAFS short course 2016 34
Energy (eV)11564 11566 11568 11570 11572 11574 11576
Nor
mal
ized
inte
nsity
0.4
0.6
0.8
1.0
1.2
1.4
673 K550 K488 K393 K294 K
Energy (eV)11564 11566 11568 11570 11572 11574 11576
Nor
mal
ized
inte
nsity
0.4
0.6
0.8
1.0
1.2
1.4
50,000 ppm 25,000 ppm10,000 ppm400 ppm
Energy (eV)11564 11566 11568 11570 11572 11574 11576
Nor
mal
ized
inte
nsity
0.4
0.6
0.8
1.0
1.2
1.4
50,000 ppm 25,000 ppm10,000 ppm400 ppm
Energy (eV)11564 11566 11568 11570 11572 11574 11576
Nor
mal
ized
inte
nsity
0.4
0.6
0.8
1.0
1.2
1.4
673 K550 K488 K393 K294 K
M.W. Small et al, ACS NANO 6 5583 (2012) J. Chem. Phys. 123, 184701 (2005).
50,000 ppm H2 400 ppm CO
H2 at 673 K
CO at 673 K
Office of Institutional Research, Planning & Effectiveness
Charge transfer
Changes in intensities of NPs XANES features upon temperature / pressure changes can be related to charge transfer processes between NPs, support and adsorbates. Integral differetial XANES signal may be used to highlight and quantify the differences.
11/9/2016 XAFS short course 2016 35
M.W. Small et al, ACS NANO 6 5583 (2012) J. Chem. Phys. 123, 184701 (2005).
Office of Institutional Research, Planning & Effectiveness
Charge transfer
Changes in intensities of NPs XANES features upon temperature / pressure changes can be related to charge transfer processes between NPs, support and adsorbates. Integral differetial XANES signal may be used to highlight and quantify the differences.
11/9/2016 XAFS short course 2016 36
M.W. Small et al, ACS NANO 6 5583 (2012) J. Chem. Phys. 123, 184701 (2005).
.)(1
),( / PCBTe
APTS Tn +++
= −α
)(),( ),( PCBTPTAPTS ++= θ
Particle- adsorbate
Particle- support
Gas- support
Office of Institutional Research, Planning & Effectiveness
Linear combinations
Commonly investigated sample is a mixture of different species (e.g., of reactants and reaction products). Common approach to XANES analysies relies on linear combination fitting, where the XANES spectrum of unknown material is expressed as linear combination of reference spectra. Relatively short time, required to acquire a single XANES spectrum, makes XANES studies especially attractive for time-resolved investigations of nanomaterial structure. J. A. Rodriguez et al J. Am. Chem Soc. 124 346 (2002)
11/9/2016 XAFS short course 2016 37
A. Kuzmin and J. Chaboy, IUCrJ 1 571 (2014)
𝜇𝜇 𝐸𝐸 = �𝜇𝜇𝑖𝑖 𝐸𝐸 𝑤𝑤𝑖𝑖𝑖𝑖
𝑤𝑤𝑖𝑖 - molar fraction of i-th component (fitting paremeter)
Office of Institutional Research, Planning & Effectiveness
Linear combinations
Commonly investigated sample is a mixture of different species (e.g., of reactants and reaction products). Common approach to XANES analysies relies on linear combination fitting, where the XANES spectrum of unknown material is expressed as linear combination of reference spectra. Relatively short time, required to acquire a single XANES spectrum, makes XANES studies especially attractive for time-resolved investigations of nanomaterial structure. J. A. Rodriguez et al J. Am. Chem Soc. 124 346 (2002)
11/9/2016 XAFS short course 2016 38
A. Kuzmin and J. Chaboy, IUCrJ 1 571 (2014)
𝜇𝜇 𝐸𝐸 = �𝜇𝜇𝑖𝑖 𝐸𝐸 𝑤𝑤𝑖𝑖𝑖𝑖
𝑤𝑤𝑖𝑖 - molar fraction of i-th component (fitting paremeter)
Isosbestic points
Office of Institutional Research, Planning & Effectiveness
PCA analysis
If the reference compounds and/or their total number are not known, Principial Component analyis can be employed. PCA allows one to obtain minimal linear-independent set of reference spectra, required to fit all experimental data.
11/9/2016 XAFS short course 2016 39
= . . A U λ2
λ1
… λN
VT
N data sets, M energy points in each measurement ⇒ data matrix A(N,M)
M×N
M×N
N×N
N×N
Data
Eigenvectors (components)
Eigenvalues (importance
of the components)
Eigenvectors (importance of the
components for individual spectrum)
λ1 > λ2 > … > λN
A. I. Frenkel et al, . Chem. Phys., 116, 9449 (2002). Q. Wang et al J. Chem. Phys. 129, 234502 (2008)
Office of Institutional Research, Planning & Effectiveness
Principial Component Analysis allows one to discriminate the most important contributions to experimental spectra, and, hence, allows one to find the number of different components in the mixture. Identification of components can be done by comparing the PC with spectra of reference materials, or by fitting of XAS spectra, constructud from PC.
Q. Wang et al J. Chem. Phys. 129, 234502 (2008)
Time-resolved XANES of H2 reduction of Ce0.8Cu0.2O2
First Principial Components
Residual for fit with 2 PC
Fit of experimental spectrum with 2 PC
Residual for fit with 3 PC
Fit of experimental spectrum with 3 PC
11/9/2016 XAFS short course 2016 40
PCA analysis of XANES data
Office of Institutional Research, Planning & Effectiveness
XANES modelling
Similar approaches can be used for ab-initio XANES modelling as in case of EXAFS modelling. XANES modelling, however, currently usually is less accurate then EXAFS modelling.
J. Timoshenko et al, Phys. Chem. Chem. Phys. 18 19621 (2016)
Comparison of ab-intio Pd K-edge XANES spectra with experimental data for some Pd-containing reference compounds.
11/9/2016 XAFS short course 2016 41
J. J. Rehr, et al, Phys. Chem. Chem. Phys. 12 5503 (2010).
Office of Institutional Research, Planning & Effectiveness
XANES modelling
Nevertheless, XANES modelling still can be used for interpretation of experimental spectra. In particular, calculated XANES spectra can be used for linear combination analysis of experimental data, when the experimental spectra for corresponding reference compounds are not available. This approach can help, e.g., to find the placement of dopant atoms in the metallic nanoparticle.
J. Timoshenko et al, Phys. Chem. Chem. Phys. 18 19621 (2016)
Ab-intio XANES spectra for different locations of dopant atom:
Fit of experiment XANES with linear combinations of calculated spectra:
11/9/2016 XAFS short course 2016 42
Office of Institutional Research, Planning & Effectiveness
RIXS & HERFD
Electronic structure of absorbing atoms can be probed in even more details using Resonant Inelastic X-ray Scattering and High Energy Resolution Fluoroscence Detection techniques. These methods provides improved sensitivity to changes in electronic structure than conventional XANES, and can be used, e.g., to analyze charge transfer processes due to interaction with adsorbates.
A. Elsen et al, J. Phys. Chem. C 119 25615 (2015)
11/9/2016 XAFS short course 2016 43
Office of Institutional Research, Planning & Effectiveness
Combination with other techniques
The complete characterization and understanding of complexity of nanoscale systems is possilble only by combining the power of XAS with the possibilities provided by other experimental techniques (TEM, X-ray scattering, optical spectroscopy techniques, etc.).
Y. Li, et al, Nature Communications, 6, 7583 (2015) S. Zhao, Chem. Cat. Chem 7 3683 (2015)
Combined XAFS/XRD/Raman
Combined XAFS/XRD/DAFS Combined XAFS/DRIFTS
11/9/2016 XAFS short course 2016 44
top related