The Comparison between Single Atom Catalysis and Surface
Organometallic Catalysis
Manoja K. Samantaray,
†
Valerio D’Elia,
‡
Eva Pump,
†
Laura Falivene,
†
Moussab Harb,
†
Samy Ould Chikh,
†
Luigi Cavallo,
†
and Jean-Marie Basset*
,†
†
King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900, Saudi Arabia
‡
School of Molecular Science and Engineering (MSE), Vidyasirimedhi Institute of Science and Technology (VISTEC), Wang Chan,
Payupnai, 21210 Rayong, Thailand
ABSTRACT: Single atom catalysis (SAC) is a recent discipline of heterogeneous
catalysis for which a single atom on a surface is able to carry out various catalytic
reactions. A kind of revolution in heterogeneous catalysis by metals for which it was
assumed that specific sites or defects of a nanoparticle were necessary to activate
substrates in catalytic reactions. In anothe r extreme of the spectrum, surface
organometallic chemistry (SOMC), and, by extension, surface organometallic catalysis
(SOMCat), have demonstrated that single atoms on a surface, but this time with
specific ligands, could lead to a more predictive approach in heterogeneous catalysis.
The predictive character of SOMCat was just the result of intuitive mechanisms derived
from the elementary steps of molecular chemistry. This review article will compare the
aspects of single atom catalysis and surface organometallic catalysis by considering
several specific catalytic reactions, some of which exist for both fields, whereas others
might see mutual overlap in the future. After a definition of both domains, a detailed
approach of the methods, mostly modeling and spectroscopy, will be followed by a detailed analysis of catalytic reactions:
hydrogenation, dehydrogenation, hydrogenolysis, oxidative dehydrogenation, alkane and cycloalkane metathesis, methane
activation, metathetic oxidation, CO
2
activation to cyclic carbonates, imine metathesis, and selective catalytic reduction (SCR)
reactions. A prospective resulting from present knowledge is showing the emergence of a new discipline from the overlap
between the two areas.
CONTENTS
1. Introduction B
1.1. The Comparison between Single Atom
Catalysis and Surface Organometallic Catal-
ysis B
1.2. Remarks on Heterogeneous Catalysis by
Single Atoms C
1.3. Definitions of SOMC, SOMF, and SCF D
1.4. Definition of Single Atom Catalysis (SAC)
and Single Alloy Atom Catalysis (SAAC) D
1.5. Analogies and Differences between SOMCat
and SAC E
2. Physicochemical Tools for Structure Determina-
tion E
2.1. Scanning Transmission Electron Microscopy
(STEM) E
2.2. X-ray Spectroscopy H
2.2.1. EXAFS Spectroscopy J
2.2.2. XANES and HERFD-XAS Spectroscopies M
2.2.3. Valence-to-Core X-ray Emission Spec-
troscopy (XES) N
2.3. Solid-State NMR Spectroscopy O
2.3.1. Standard Characterization of SOMC
Complexes O
2.3.2. Characterization of SOMCa t through
Heteroatoms Q
2.3.3. Characterization of SOMC Complexes
through Quadrupolar Nuclei R
3. Computational Tools S
3.1. Models S
3.2. Methods U
4. Classification by Types of Reaction and/or
Activation U
4.1. Hydrogenation U
4.1.1. Via SOMC V
4.1.2. Via SAAC W
4.1.3. Via SAC Y
4.2. Hydrogenolysis, Alkane Metathesis via C−H
and C−C Bond Activation Z
4.2.1. Via SOMC Z
4.2.2. Via SAAC AG
4.2.3. Via SAC AH
4.3. Light Alkanes Dehydrogenation AH
4.3.1. Via SOMC AH
Special Issue: Nanoparticles in Catalysis
Received: April 16, 2019
Review
pubs.acs.org/CR
Cite This: Chem. Rev. XXXX, XXX, XXX−XXX
© XXXX American Chemical Society A DOI: 10.1021/acs.chemrev.9b00238
Chem. Rev. XXXX, XXX, XXX−XXX
Downloaded via KING ABDULLAH UNIV SCI TECHLGY on October 21, 2019 at 10:30:54 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
4.3.2. Via SAC AO
4.4. Methane Activation AQ
4.4.1. Via SOMC AR
4.4.2. Via SAC AT
4.4.3. Via SAAC AZ
4.5. Emerging Reactions in SOMCat and SAC
Chemistry AZ
4.5.1. Imine Metathesis AZ
4.5.2. Metathetic Oxidation of Olefins BB
4.5.3. Catalytic Hydrogenation of N
2
into NH
3
BC
4.5.4. Selective Catalytic Reduction of NOx by
Ammonia (NH
3
−SCR) BE
4.5.5. CO
2
Conversion for the Synthesis of
Organic Compounds BF
5. Conclusions and Perspectives BI
5.1. Hydrogenation BI
5.2. Hydrogenolysis BJ
5.3. Dehydrogenation BJ
5.4. Methane Activation BJ
5.5. Metathetic Oxidation BJ
5.6. Imine Metathesis BJ
5.7. Emission Control for SCR Catalysts for
Automotive Diesel NOx BJ
5.8. CO
2
Conversion to Organic Compounds BK
5.9. Affinities BK
5.10. Differences BK
5.11. Concluding Remarks and Future Develop-
ment BL
Author Information BL
Corresponding Author BL
ORCID BL
Notes BL
Biographies BL
Acknowledgments BM
Abbreviations Used BM
References BM
1. INTRODUCTION
1.1. The Comparison between Single Atom Catalysis and
Surface Organometallic Catalysis
Since the last century, there has been a considerable amount of
work dedicated to catalysis by supported metals, a classical
field in heterogeneous catalysis. Among the emerging concepts,
several parameters were identified as crucial for activity,
selectivity, and lifetime of heterogeneous catalysts based on
supported nanoparticles. Among these parameters, (i) metal
particle size, (ii) nature of exposed faces, (iii) defects sites,
corners, step sites, (iv) (strong) metal support interactions,
and (v) bimetallic effects (electronic or steric) can be
mentioned. Recently, a huge effort has been devoted to
catalysis on smaller and smaller supported nanoparticles.
1−5
Besides the crucial factors responsible for catalytic activity/
selectivity/lifetime already mentioned in (i−v), there is a limit
situation where the nanoparticle size is restricted to a single
metal atom on a surface. After several experimental
observations, it became evident to the catalytic community
that a single atom is able to achieve what, in the last, century
was believed to be the “privileged property” of an ensemble of
atoms.
6−14
Even now, the discovery of single atom catalysis
(SAC) is limited to a rather restricted, but growing, range of
catalytic reactions.
3,5,15−19
This new area describes single metal
atoms “supported “ or “embedded” on a solid support (oxide,
nitride, sulfide, etc.) or even at the periphery of a metal
nanoparticle.
20,21
The interaction between the single atom and
the surface can be of a different nature involving covalent,
coordination, or ionic bonds. Con sidering heterogeneous
catalysts i nvolving single atoms, the ru les of molecular
chemistry (whether it i s coordination or organometallic
chemistry) apply, and help to rationalize the structure and
the reactivity of this grafted, “deposited” atom.
In parallel to SAC, but mostly earlier than its recent
explosion, the discipline of surface organometallic chemistry
(SOMC), which progressively developed into surface organo-
metallic catalysis (SOMCat), emerged, involving single metal
atoms covalently or ionically bound to a solid support.
22−26
In
SOMCat, the catalytically active sites are formed by reacting
organometallic complexes or coordination compounds with
well-defined surfaces (oxides, metal nanoparticles, carbon, or
graphene) to achieve the target functionalities on the surface.
26
The situation is different from the previous case (SAC)
because the supported organometallic compounds keep at least
part of their ligands after grafting (Scheme 1 for metal oxide
support). The latter aspect is crucial for catalysis. In catalysis by
design,
27
the supported complex contains moieties (fragments
A and B of Scheme 1) that are selected based on the proposed
mechanism derived from well-established steps in molecular
chemistry. Other ligands (spectator ligand X of Scheme 1)
serve to control oxidation state, geometry, and d
n
configuration
of the metal within the catalytic cycle.
Besides the evident d ifferences discussed above, both
disciplines p resent similarities and chances for mutual
enrichment. SAC and SOMCat both deal with isolated metal
atoms on surfaces, some with preexisting ligands and some
without ligands. In both cases, applying isolated single atoms in
heterogeneous catalysis helps in understanding elementary
steps as they resemble elementary steps in molecular catalysis
or coordination chemistry, where single atoms are surrounded
by ligands. This suggests that both methodologies are not
mutually exclusive. For example, one can consider that SAC
adopts a coordination sphere already present in SOMC-
prepared catalysts in the presence of reactants A and B and
component X (Scheme 1).
Furthermore, the SOMC methodology can be employed to
prepare single atoms. By combining grafting of precursors on
surfaces via SOMC with a thermolytic, nonoxidative procedure
that removes all organic ligands, isolated atoms may be
produced on the surface.
28
Several alkane dehydrogenation
catalysts (vide infra) have been prepared by this strategy
having a proposed M-O
s
structure (O
s
: oxygen atom of the
oxide support) without coordinated ligands on the metal.
29
Scheme 1. Link between Single Atom Catalysts (SAC) on
Oxide and Catalysts Prepared by SOMC on Oxides
a
a
M = catalytically active metal; M
s
= metal atom of the oxide support.
Chemical Reviews Review
DOI: 10.1021/acs.chemrev.9b00238
Chem. Rev. XXXX, XXX, XXX−XXX
B
Many reviews have already appeared recently both in SAC
1
and SOMCat;
26
the target of this review article is, in the first
place, to provide a critical and comprehensive overview about
similarities and differences between SOMCat and SAC, and,
additionally, to identify emerging reactions, areas of possible
overlap, and how the two disciplines could mutually benefit
each other. The comparison will be in terms of tools, structure,
reactivity, and mechanistic understanding of elementary steps
when they are known or proposed. Several types of “classical”
heterogeneous chemistry reactions have been selected as a
result of their relative importance either in both fields or in
each field separately: hydrogenation, dehydrogenations,
reactions involving C−C bond formation and breakage like
hydrogeno lysi s or alka ne metathesis, methane activation,
transformation, and functionalization. Furthermore, to com-
plete our overview, we will discuss emerging reactions in both
fields such as imine metathesis, metathetic oxidation or
selective catalytic reduction, and CO
2
conversion for organic
synthesis. Preference will be given to publications where a
tentative mechanistic interpretation has been proposed.
Several aspects will be considered in this review article: (i)
single atom catalysis on support (SAC), (ii) single atoms on
mono or plurimetallic nanoparticles or single atom alloy
catalysis (SAAC) and (iii) single atoms in SOMCat. In the first
case (i), the single atoms are located either at the periphery or
on the surface of an oxide, nitride, carbon nitrides, sulfide,
graphene, and microporous material including zeolites, MOF.
In some particular cases (ii), the single metal atom has been
isolated at the periphery of a metal nanoparticle, but it is the
only species which exhibits catalytic activity (e.g., Pt/Sn). This
situation is at the limit when an “ensemble of metal atoms” in a
nanocluster decreases to such an extent that only one single
metal atom is involved in the catalysis. In the third case (iii),
the single metal atoms are covalently linked mostly to the
surface of the oxide support but display ligands (spectator or
functional) inherited by the organometallic or coordination
compound precursors.
1.2. Remarks on Heterogeneous Catalysis by Single Atoms
In heterogeneous catalysis, catalysts transform molecules into
new molecu les, ma cromolecule s, or soli ds. These tran s-
formations typically (except for electron transfer reactions)
involve creation of covalent bonds between one or several
surface atoms with one or several atoms of the substrate. The
thus formed reaction intermediate has been named surface
organometallic fragment (SOMF) or surface coordination
fragment (SCF) and can be prepared on purpose.
26,27
This
very simple concept was at the origin of the development of
SOMC and later SOMCat.
23−26
SOMFs can be characterized
with specific tools of molecular chemistry as well as surface
science. In several cases, the rules of molecular chemistry
(regarding structure or reactivity) apply to these surface
organometallic fragments. This enables rationalization of the
elementary steps of heterogeneous catalysis by understanding
the way bonds are broken or made. These elementary steps are
very different from the well-known and accepted classical
elementary steps of adsorption, desorption, diffusion, etc.
SOMFs or SCFs are unparalleled tools to design catalysts,
mimic and understand reaction intermediates and mechanisms
in heterogeneous catalysis, or even discover new reactions yet
unknown in homogeneous or heterogeneous catalysis (Scheme
2).
27
There are also more and more examples showing that
single metal atoms with appropriate ligands (SOMFs and
SCFs) are able to achieve new catalytic reactions.
27
In most
cases, the choice of appropriate ligands was a pure direct
prediction of a catalytic reaction. The most relevant examples
are Ziegler−Natta depolymerization, alkane metathesis, or
metathetic oxidation of olefins.
30,31
There are more and more examples showing that a single
metal atom on the metallic surface of a nanoparticle (SAAC)
can achieve multistep reactions which were previously believed
to occur only on a large number of atoms assumed to be the
“active sites” of classical heterogeneous catalysis. The term
“ensemble effect” was advanced to explain this phenomenon.
6
The main difference between catalysis by single atoms and
homogeneous catalysis is the presence of a surface which plays
the role of a solid, rigid ligand with redox, acid−base, as well as
physical properties (e.g., porosity, hydrophilicity, hydro-
phobicity, semiconducting properties, etc.). With oxide
supports, the electron count of SOMFs and SCFs (including
support and spectator ligands) are easily rationalized by the
classical electron counts, oxidation state, and d
n
configurations
of molecular chemistry.
32
The oxide surface can be regarded as
an infinite “pool” of oxygens, or oxygen containing species,
which plays the role of X or L ligands in the MLH Green
formalism.
33
The composition of the surface in terms of
adsorbed water molecules, surface hydroxyls, or metal oxides
(M-O-M; M: metal, e.g., Si−O−Si on silica) is mostly
governed by the temperature of dehydroxylation of the
support. These M-O-M surface species may also coordinate
the grafted metals via an oxygen lone pair, contributing to the
global metal electron count (L ligand).
Besides the contribution of the surface to the electron count
of the grafted metal, spectator ligands, which do not appear as
participating directly in the elementary steps of the mechanism,
are sometimes crucial ligands in the control of the electron
density of the metal and also of its oxidation state, d
n
configuration and geometry.
In this review article, we will focus mostly on the role of
SOMF, SCF, and spectator ligands without entering into the
complex situation of the interaction between a metal fragment
and its support. The role of metal/support interaction will be
considered in a separate article.
Scheme 2. Definition of Catalysis by Design: The Catalytic
Cycle Is Entered by a Presumed Reaction Intermediate
Called Surface Organometallic Fragment (SOMF) or
Surface Coordination Fragment (SCF) or One of Its Closest
Precursors
a
a
Reproduced with permission from ref 32. Copyright 2018 Oxford
University Press.
Chemical Reviews Review
DOI: 10.1021/acs.chemrev.9b00238
Chem. Rev. XXXX, XXX, XXX−XXX
C
1.3. Definitions of SOMC, SOMF, and SCF
When a metal is grafted directly at the surface of an oxide
(silica, alumina, ceria, silica−alumina, mesoporous materials,
etc.) by classical methods of SOMC, the following bonds and
interactions are possible (Scheme 2):
• Metal−surface bond: σ-bond between a surface oxygen
(which behaves as an X-ligand, Scheme 1) and a metal
(M). It formally oxidizes the metal by one unit. In the
following, the metal grafted on a support will be written
as “[M]”.
M(O)M
x
x
[
]= − −
• Metal−surface interaction: π-bond or coordination
bond between a surface oxygen (L-ligand) and the
grafted metal [M]. This interaction formally does not
oxidize the metal. It is also represented by
(
O) M−−→[ ]
A surface oxygen which is covalently linked to a metal is
not only an X-ligand but can also act as an L-ligand
because its lone pair overlaps to a certain extent with the
π-orbitals of the metal. Besides direct reaction with the
precursor, there are many examples where a [M]−Ror
[M]−H group can open an adjacent Si−O−Si
moiety with formation of [M]− O−Si and, respec-
tively, Si-R or Si−H via classical σ-bond metathesis
mechanism (Scheme 3).
34,35
• Metal−functional fragment bond: σ-bond between the
grafted metal [M] and the functional fragment “R”
which formally oxidizes the metal by one unit. This
fragment “R” is supposed to be an intermediate among
all the elementary steps of a catalytic cycle (in contrast
to the SSP and the support). The fragment “ R” allows
focusing on reactivity or interconversions of surface
moieties. One SOMF can convert into another SOMF:
[M]−R → [M]−R′. This conversion can be linked to
the reactivity of “R” and might influence the elementary
steps of heterogeneous catalysis (e.g., a metal alkyl giving
a metal (carbene)(hydride)). The fragment “R” can have
several functions (mono, bis, tris, tetra). This means that
a single metal atom can possess simultaneously several
different/or identical fragments (e.g., R can be a hydride
or an alkyl).
• Metal-spectator bond: Typically, the spectator ligand
(X-ligand in Scheme 1) does not directly take part to
reactions but plays a crucial role in fine-tuning the
electronic or steric effects of [M] and in defining the
oxidation states. In rare cases, the same kind of ligand
can serve as spectator or as catalytically active moiety
according to the reaction (for instance metal oxo
functionalities in olefin metathesis reactions
36
or in
oxidative dehydrogenation of propane).
37
A model of chemical environment of a single atom prepared
via SOMC methodology is represented in Figure 1. To note,
SOMC complexes that are directly bound to the surface, and
where the surface acts as rigid ligand, are evidently different
from supported homogeneous catalysts where the metal atom
is tethered to the support surface via flexible linkers that
generally coordinate the metal via noncovalent interactions.
38
1.4. Definition of Single Atom Catalysis (SAC) and Single
Alloy Atom Catalysis (SAAC)
Many reviews have been written recently on SAC.
3,15−18
Behind this acronym, there is a proliferation of examples in
which single atoms are found or claimed to be involved in
several catalytic reactions.
18
This makes it difficult to give a
precise definition of single atom catalysis. The most frequently
accepted concept, evolving with the progress of synthesis and
characterization tools, is that “isolated” atoms “adsorbed,”
“chemisorb ed,”“embedded, ” and “immobilized” on a
“support”, exhibit catalytic properties in a still limited number
of reactions.
The fact that a single supported metal atom displays catalytic
properties has been considered as a kind of revolutionary
concept in the heterogeneous catalysis community. There was
a general belief that metal nanoparticles were responsible for
the catal ytic activity o f metal containing catalysts. The
contradiction came from the research of size effects or of the
crystallographic position necessary to achieve a given reaction.
The activities were usually divided by the number of surface
atoms (TOF) in the particles. Some reactions were claimed to
be “facile” and the TOF was just depending on the number of
surface atoms irrespective of their position at the surface of the
particle. Others were quali fied as “demanding” because many
parameters were playing a role on the measured TOF, in
particular, the particle size. For those “demanding” reactions,
the activity was found to depend on metal particle size, but
nobody was expecting the extreme situation where only a
Scheme 3. Example for [M]−Surface Bond and [M]−
Surface Interaction and Opening of Siloxane Bridge through
[M]−R Ligand
Figure 1. Representation of a typical SOMF or SCF containing oxide support, transition metal, spectator ligands, and functional ligands and
comparison with supported homogeneous catalyst.
Chemical Reviews Review
DOI: 10.1021/acs.chemrev.9b00238
Chem. Rev. XXXX, XXX, XXX−XXX
D
single metal atom could achieve catalytic reactions by itself and
be recovered intact at the end of a catalytic run.
In addition, it is known that single atoms existing in the gas
phase possess catalytic properties,
39,40
there are also dis-
ciplines, such as homogeneous catalysis, where single metal
atoms catalyze many reactions in the presence of ligands. Also,
in biocatalysis, a single metal atom surrounded by appropriate
ligands could achieve difficult reactions (such as for example
ammonia synthesis).
41, 42
For a single atom which is
“adsorbed,”“chemisorbed,”“embedded,”“immobilized,” or
”grafted” on (or in) a support, the type of bonding between
this atom and the surface, the bulk, the periphery of this
support is crucial for the reactivity of the catalyst. The basic
rules of chemistry apply to this so-called interfacial situation
defining the interaction between surface and single atom as
covalent, ionic, coordination, electron transfer, etc. The direct
consequence is the precise definition of the oxidation state by
rationalizing the electron count of this single atom “attached”
to the surface or embedded in the bulk of the support.
The following examples of SACs have been reported:
• Single metal atoms may be linked to “unusual” (with
respect to SOMC) or innovative materials such as
carbon,
43
carbon nanotubes,
44
carbon nitrides,
17
MOFs-
derived matrices,
45
silica layers grown by ultrahigh
vacuum techniques,
14,46
etc., which possess chemical or
physical functionalities. It was appealing to ascribe some
catalytic properties to the site isolation concept.
• Single atom catalysis was also claimed to occur when the
isolation of the single site occurs in a metallic
nanoparticle for which the “active metal atom” was
dispersed in a kind of matrix of inactive metal (SAAC).
In classical heterogeneous catalysis, a so-called “ dilution
effect,attheoppositeofensembleeffect,”
47
was
observed.
• Another example of single atom catalysis is related to
isolated metal atoms in semiconductors. It was
discovered that the electronic properties of such semi-
conductors change with the incorporation of single
metal atoms. Single atoms were proposed to participate
in the energy level of the electron transfer process (e.g.,
photocatalytic water splitting).
48
The consequence of this ambiguity is the di fficulty to
propose a unifying definition and theory behind SAC.
1.5. Analogies and Differences between SOMCat and SAC
SAC and SOMCat are both domains dealing with a single
atom “attached,”“chemisorbed,”“embedded,”“immobilized,”
and “grafted” to/in/on a support. In SOMCat, the support is
mostly an oxide or an oxygen containing material (e.g., silica,
MOF, zeolites). In SAC, a lot of supports have been tested
including carbon, carbon nanotubes, carbon nitrides, MOF,
semiconductors, and metal particles.
Both domains have studied various catalytic reactions that
occur exclusively on single atoms. Some of these reactions are
identical, some are different. Some transformations exist in
SOMCat but not in SAC and vice versa. This difference is
because in SOMCat efforts were, initially, mainly concerned
with hyd rogenolysis and alkane metathesis u sing early
transition metals,
25,27
whereas SAC has only recently risen to
wide popularity
1
and has not yet covered many existing
processes. Additionally, SOMCat has been mainly focused on
early transition metals, whereas most works on SAC use late
transition or noble metals. Nevertheless, these gaps are due to
be filled in next few years because (a) the field of application of
both domains is progressively being expanded; for instance,
SOMCat has been recently applied to reactions such as
oxidative
37
and nonoxidative dehydrogenation of propane,
49
CO
2
conversion to carbonates,
50
and imine metathesis,
51
just
to cite some. Along with continued growth of SAC, it is
expected that the number of transformations available for both
techniques will strongly increase. (b) There is no specific
reason, if not because of tradition, why SAC should not involve
early transition metals. It has been shown that single atoms of
tantalum can be synthesized on silica support using a cluster
source and used in catalysis.
46
Similarly, SOMC syntheses of
late transition metals
52
and noble metal
53
complexes have been
reported although they were generally not used in catalysis as
such but for the synthesis of supported single-site metal−oxo
complexes or nanoparticles.
Therefore, whereas the gaps discussed above will eventually
be bridged, the fundamental difference between the method-
ologies is the approach behind each specific domain:
In simple terms, we should say that single atom catalysis is
the discovery that many catalytic reactions can occur on an
isolated metal atom embedded on an inert support. This was a
kind of revolution in heterogeneous catalysis. However, in the
absence of a well-defined coordination sphere on the metal
atom, prediction and tuning of catalytic activity are generally
very challenging with SAC.
In SOMCat, the concept at the origin of the field was a kind
of predictive concept of catalysis by design
30
for which the
elementary steps of molecular chemistry can be applied to
discover (or improve) reactions. This was made possible by
starting a catalytic reaction from a well-defined supported
reaction intermediate (SOMF or SCF), which means that the
SOMCat strategy is predictive, provided a presumed
mechanism at the origin of the catalyst synthesis.
2. PHYSICOCHEMICAL TOOLS FOR STRUCTURE
DETERMINATION
2.1. Scanning Transmission Electron Microscopy (STEM)
The most convincing evidence of a full atomic dispersion is
giv en simply by imaging the SACs or SOMC-prepared
complexes. During the two last decades, electron microscopy
science has benefited from a huge improvement in the lateral
resolution of scanning transmission electron microscopes.
54
This is because the spherical aberration induced by the used
electromagnetic round lenses are now compensated mostly by
the integration of aberration correction devices.
55
Those past
developments allowed new possibilities for the analysis at the
atomic scale, which coincided with the emergence of synthesis
protocols for the immobilization of single atoms.
Hence, the technique which is the most successful is the
aberration corrected scanning transmission electron micros-
copy (STEM) coupled to a high-angle annular dark field
(HAADF) detection. By using a combination of detector and
camera length allowing a large collection angles (β >70−200
mrad), electrons arising from Rutherford scattering are
collected while electrons deviated by coherent elastic scattering
are mostly excluded (removing the phase contrast). The
intensity I in the resulting images is then, in first
approximation, given by I α t · Z
α
(α = 1.5−2) with a
thickness t and an average atomic number Z.
56
As a direct
consequence, the contrast in HAADF-STEM i mages is
sharpened when the difference between atomic number of
Chemical Reviews Review
DOI: 10.1021/acs.chemrev.9b00238
Chem. Rev. XXXX, XXX, XXX−XXX
E
![Figure 2. Aberration-corrected HAADF-STEM images of the sample prepared by adsorption of Au(CH3)2(acac) in zeolite NaY. (a) Initially prepared sample. (b) Sample after treatment in flowing CO + O2 (during CO oxidation catalysis) for 30 min. (c) FFT of the experimental image shown in (a). (d) FFT of the experimental image shown in (b). (e) Simulated framework and theoretical diffraction pattern of zeolite Y in the [110] projection. (f−h) Zoomed-in views of the regions shown in rectangles in (a) and (b). Reproduced with permission from ref 57. Copyright 2012 John Wiley and Sons.](/figures/figure-2-aberration-corrected-haadf-stem-images-of-the-3bff8jei.webp)
