An Introduction To Surface Chemistry

An Introduction to Surface
Chemistry
written by
Dr. Roger Nix
Department of Chemistry
1. Structure of Metallic Surfaces
1.1 Surface Structure of Metals
In most technological applications, metals are used either in a finely divided form (e.g.
supported metal catalysts or in a massive, polycrystalline form (e.g. electrodes,
mechanical fabrications.
!t the microscopic level, most materials, with the notable exception of a few truly
amorphous specimens, can be considered as a collection or aggregate of single crystal
crystallites. "he surface chemistry of the material as a whole is therefore crucially
dependent upon the nature and type of surfaces exposed on these crystallites. In principle,
therefore, we can understand the surface properties of any material if we
#. $now the amount of each type of surface exposed , and
%. have detailed $nowledge of the properties of each and every type of surface plane.
("his approach assumes that we can neglect the possible influence of crystal defects and
solid state interfaces on the surface chemistry
It is therefore vitally important that we can independently study different, well&defined
surfaces. "he most commonly employed techni'ue, is to prepare macroscopic (i.e. si(e )
cm single crystals of metals and then to deliberately cut&them in a way which exposes a
large area of the specific surface of interest.
*ost metals only exist in one bul$ structural form & the most common metallic crystal
structures being +
bcc body-centred cubic
fcc face-centred cubic
hcp hexagonal close packed
,or each of these crystal systems, there are in principle an infinite number of possible
surfaces which can be exposed. In practice, however, only a limited number of planes
(predominantly the so&called -low&index- surfaces are found to exist in any significant
amount and we can concentrate our attention on these surfaces. ,urthermore, it is possible
to predict the ideal atomic arrangement at a given surface of a particular metal by
considering how the bul$ structure is intersected by the surface. ,irstly, however, we need
to loo$ in detail at the bul$ crystal structures.
I. The hcp and fcc structures
"he hcp and fcc structures are closely related + they are both based upon stac$ing layers of
atoms, where the atoms are arranged in a close&pac$ed hexagonal manner within the
individual layer.
"he atoms of the next layer of the structure will preferentially sit in some of the hollows in
the first layer & this gives the closest approach of atoms in the two layers and thereby
maximi(es the cohesive interaction.
.hen it comes to deciding where the next layer of atoms should be positioned there are
two choices & these differ only in the relative positions of atoms in the #st and /rd layers.
In the structure on the left the atoms of the /rd layer sit directly above those in the #st layer
& this gives rise to the characteristic ..!0!0!.. pac$ing se'uence of the hcp structure.
In the structure on the right the atoms of the /rd layer are laterally offset from those in both
the #st and %nd layers, and it is not until the 1th layer that the se'uence begins to repeat.
"his is the ..!0C!0C.. pac$ing se'uence of the fcc structure. 0ecause of their common
origin, both of these structures share common features +
#. "he atoms are close pac$ed
%. 2ach atom has #% nearest neighbours ( i.e. CN 3 #%
a! fcc structure
!lthough it is not immediately obvious, the ..!0C!0C.. pac$ing se'uence of the fcc
structure gives rise to a three&dimensional structure with cubic symmetry ( hence the
name 4 .
fcc structure
It is the cubic unit cell that is commonly used to illustrate this structure & but the fact that
the origin of the structure lies in the pac$ing of layers of hexagonal symmetry should not
be forgotten.
"he above diagram shows the atoms of one of the hexagonal close&pac$ed layers
highlighted in shades of red, and the atoms of another highlighted in shades of green.
b! hcp structure
"he ..!0!0!.. pac$ing se'uence of the hcp structure gives rise to a three&dimensional unit
cell structure whose symmetry is more immediately related to that of the hexagonally&close
pac$ed layers from which it is built, as illustrated in the diagram below.
II. The bcc structure
"he bcc structure has very little in common with the fcc structure & except the cubic nature
of the unit cell. *ost importantly, it differs from the hcp and fcc structures in that it is not a
close&pac$ed structure.
bcc structure
.hat is the co&ordination number of atoms in this structure 5

"here do #e go from here $
!n ordered surface may be obtained by cutting the three&dimensional bul$ structure of a
solid along a particular plane to expose the underlying array of atoms. "he way in which
this plane intersects the three&dimensional structure is very important and is defined by
using Miller Indices & this notation is commonly used by both surface scientists and
crystallographers since an ideal surface of a particular orientation is nothing more than a
lattice plane running through the /D crystal with all the atoms removed from one side of
the plane.
In order to see what surface atomic structures are formed on the various *iller index
surfaces for each of the different crystal systems we need to consider how the lattice planes
bisect the three&dimensional atomic structure of the solid. "o pursue this 'uestion for the
fcc , hcp and bcc systems you should select the appropriate option from the menu after
exiting from this section. !s you might expect, however, the various surfaces exhibit a
wide range of+
#. 6urface symmetry
%. 6urface atom coordination

and most importantly this results in substantial differences in 77.

/. 8hysical properties ( electronic characteristics etc. , and
1. 6urface chemical reactivity (catalytic activity, oxidation resistance etc.
1.% Surface Structure of fcc Metals
*any of the technologically most important metals possess the fcc structure + for example
the catalytically important precious metals ( 8t, Rh, 8d all exhibit an fcc structure.
"he low index faces of this system are the most commonly studied of surfaces + as we shall
see they exhibit a range of
1. Surface symmetry
%. Surface atom coordination
&. Surface reacti'ity
I. The fcc 1((! surface
"he (#99 surface is that obtained by cutting the fcc metal parallel to the front surface of the
fcc cubic unit cell & this exposes a surface (the atoms in blue with an atomic arrangement
of 1&fold symmetry
fcc unit cell
1((! face
"he diagram below shows the conventional birds&eye view of the (#99 surface & this is
obtained by rotating the preceding diagram through 1:; to give a view which emphasises
the 1&fold (rotational symmetry of the surface layer atoms.
"he tops of the second layer atoms are <ust visible through the holes in the first layer, but
would not be accessible to molecules arriving from the gas phase.
.hat is the coordination number of the surface layer atoms 5
"here are several other points worthy of note +
#. !ll the surface atoms are e'uivalent
%. "he surface is relatively smooth at the atomic scale
/. "he surface offers various adsorption sites for molecules which have different local
symmetries and lead to different coordination geometries & specifically there are +
o =n&top sites (above a single metal atom
o 0ridging sites, between two atoms
o >ollow sites, between four atoms
Depending upon the site occupied, an adsorbate species (with a single point of attachments
to the surface is therefore li$ely to be bonded to either one, two or four metal atoms.
II. The fcc11(! surface
"he (##9 surface is obtained by cutting the fcc unit cell in a manner that intersects the x
and y axes but not the z&axis & this exposes a surface with an atomic arrangement of %&fold
symmetry.
fcc unit cell
11(! face
"he diagram below shows the conventional birds&eye view of the (##9 surface &
emphasising the rectangular symmetry of the surface layer atoms. "he diagram has been
rotated such that the rows of atoms in the first atomic layer now run vertically, rather than
hori(ontally as in the previous diagram.
It is clear from this view that the atoms of the topmost layer are much less closely pac$ed
than on the (#99 surface & in one direction (along the rows the atoms are in contact i.e. the
distance between atoms is e'ual to twice the metallic(atomic radius, but in the orthogonal
direction there is a substantial gap between the rows.
"his means that the atoms in the underlying second layer are also, to some extent, exposed
at the surface
11(! surface plane
e.g. Cu11(!
"he preceding diagram illustrates some of those second layer atoms, exposed at the bottom
of the troughs.
In this case, the determination of atomic coordination numbers re'uires a little more careful
thought + one way to double&chec$ your answer is to remember that the CN of atoms in the
bul$ of the fcc structure is #%, and then to subtract those which have been removed from
above in forming the surface plane.
.hat is the coordination number of the topmost layer atoms 5
If we compare this coordination number with that obtained for the (#99 surface, it is worth
noting that the surface atoms on a more open (-rougher- surface have a lower CN & this
has important implications when it comes to the chemical reactivity of surfaces.
Do the atoms in the second layer have the bul$ coordination 5
)o & the fact that they are clearly exposed (visible at the surface implies that they have a
lower CN than they would in the bul$.
.hat is the coordination number of these second layer atoms 5
In summary, we can note that
#. !ll first layer surface atoms are e'uivalent, but second layer atoms are also exposed
%. "he surface is atomically rough, and highly anisotropic
/. "he surface offers a wide variety of possible adsorption sites, including +
o =n&top sites
o 6hort bridging sites between two atoms in a single row
o ?ong bridging sites between two atoms in ad<acent rows
o >igher coordination sites ( in the troughs
III. The fcc 111! surface
"he (### surface is obtained by cutting the fcc metal in such a way that the surface plane
intersects the x&, y& and z& axes at the same value & this exposes a surface with an atomic
arrangement of /&fold ( apparently @&fold, hexagonal symmetry. "his layer of surface
atoms actually corresponds to one of the close&pac$ed layers on which the fcc structure is
based.
fcc unit cell
111! face
"he diagram below shows the conventional birds&eye view of the (### surface &
emphasising the hexagonal pac$ing of the surface layer atoms. 6ince this is the most
efficient way of pac$ing atoms within a single layer, they are said to be -close&pac$ed-.
111! surface plane
e.g. *t111!
"he following features are worth noting A
#. !ll surface atoms are e'uivalent and have a relatively high CN
%. "he surface is almost smooth at the atomic scale
/. "he surface offers the following adsorption sites +
o =n&top sites
o >ollow sites, between three atoms
I+. ,o# do these surfaces intersect in irregular-shaped samples.
,lat surfaces of single crystal samples correspond to a single *iller Index plane and, as we
have seen, each individual surface has a well&defined atomic structure. It is these flat
surfaces that are used in most surface science investigations, but it is worth a brief aside to
consider what type of surfaces exist for an irregular shaped sample (but one that is still
based on a single crystal. 6uch samples can exhibit facets corresponding to a range of
different *iller Index planes. "his is best illustrated by loo$ing at the (clic$able diagrams
below.
(i an angled corner (ii a spherical tip
,rom the 0!?6!C picture gallery of 8rof. B. >ermann, ,rit(&>aber&Institut, 0erlin
S-MMA./
Depending upon how an fcc single crystal is cleaved or cut, flat surfaces of macroscopic
dimensions which exhibit a wide range of structural characteristics may be produced.
"he single crystal surfaces discussed here ( (#99, (##9 C (### represent only the most
fre'uently studied surface planes of the fcc system & however, they are also the most
commonly occurring surfaces on such metals and the $nowledge gained from studies on
this limited selection of surfaces goes a long way in propagating the development of our
understanding of the surface chemistry of these metals.
,or further information on other fcc metal surfaces you should ta$e a loo$ at +
6ection #.D + =ther 6ingle Crystal 6urfaces
Includes a brief description of high index fcc surfaces
with illustrative examples.
!tlas of fcc Crystal 6urfaces
! -clic$able map- of a wide range of fcc surfaces
(one part of a useful tutorial on metal crystallography
from 8er 6tol(e, !alborg Eniversity..
1.& Surface Structure of hcp Metals
"his important class of metallic structures includes metals such as Co, Fn, "i C Ru.
"he *iller Index notation used to describe the orientation of surface planes for all
crystallographic systems is slightly more complex in this case since the crystal structure
does not lend itself to description using a standard cartesian set of axes& instead the
notation is based upon three axes at #%9 degrees in the close&pac$ed plane, and one axis
(the c&axis perpendicular to these planes. "his leads to a four&digit index structure A
however, since the third of these is redundant it is sometimes left out 4
I. The hcp (((1! surface
"his is the most straightforward of the hcp surfaces since it corresponds to a surface plane
which intersects only the c&axis, being coplanar with the other / axes i.e. it corresponds to
the close pac$ed planes of hexagonally arranged atoms that form the basis of the structure.
It is also sometimes referred to as the (99# surface.
(((1! surface plane
e.g. .u(((1!
"his conventional plan view of the (999# surface shows the hexagonal pac$ing of the
surface layer atoms.
.hich fcc surface does it resemble 5
.e can summarise the characteristics of this surface by noting that +
#. !ll the surface atoms are e'uivalent and have CN3G
%. "he surface is almost smooth at the atomic scale
/. "he surface offers the following adsorption sites +
o =n&top sites
o >ollow sites, between three atoms
1.0 Surface Structure of bcc Metals
! number of important metals ( e.g. ,e, ., *o have the bcc structure. !s a result of the
low pac$ing density of the bul$ structure, the surfaces also tend to be of a rather open
nature with surface atoms often exhibiting rather low coordination numbers.
I. The bcc 1((! surface
"he (#99 surface is obtained by cutting the metal parallel to the front surface of the bcc
cubic unit cell & this exposes a relatively open surface with an atomic arrangement of 1&fold
symmetry.
bcc unit cell
1((! face
"he diagram below shows a plan view of this (#99 surface & the atoms of the second layer
(shown on left are clearly visible, although probably inaccessible to any gas phase
molecules.
bcc1((! surface plane
e.g. 1e1((!
Is the coordination of the second layer atoms the same as that of bul$ atoms 5
II. The bcc 11(! surface
"he (##9 surface is obtained by cutting the metal in a manner that intersects the x and y
axes but creates a surface parallel to the (&axis & this exposes a surface which has a higher
atom density than the (#99 surface.
bcc unit cell
11(! face
"he following diagram shows a plan view of the (##9 surface & the atoms in the surface
layer strictly form an array of rectangular symmetry, but the surface layer coordination of
an individual atom is 'uite close to hexagonal.
bcc11(! surface plane
e.g. 1e11(!
III. The bcc 111! surface
"he (### surface of bcc metals is similar to the (### face of fcc metals only in that it
exhibits a surface atomic arrangement exhibiting /&fold symmetry & in other respects it is
very different.
Top +ie# 2
bcc111! surface plane
e.g. 1e111!
In particular it is a very much more open surface with atoms in both the second and third
layers clearly visible when the surface is viewed from above. "his open structure is also
clearly evident when the surface is viewed in cross&section as shown in the diagram below
in which atoms of the various layers have been annotated.
Side +ie# 2
bcc111! surface plane
e.g. 1e111!
1.3 4nergetics of Solid Surfaces
!ll surfaces are energetically unfavourable in that they have a positive free energy of
formation. ! simple rationalisation for why this must be the case comes from considering
the formation of new surfaces by cleavage of a solid and recogni(ing that bonds have to be
bro$en between atoms on either side of the cleavage plane in order to split the solid and
create the surfaces. 0rea$ing bonds re'uires wor$ to be done on the system, so the surface
free energy (surface tension contribution to the total free energy of a system must
therefore be positive.
"he unfavourable contribution to the total free energy may, however, be minimised in
several ways +
#. 5y reducing the amount of surface area exposed
%. 5y predominantly exposing surface planes #hich ha'e a lo# surface free
energy
/. 5y altering the local surface atomic geometry in a #ay #hich reduces the
surface free energy
"he first and last points are considered elsewhere ( #.H 8articulate *etals, C #.@ Relaxation
and Reconstruction , respectively & only the second point will be considered further here.
=f course, systems already possessing a high surface energy (as a result of the preparation
method will not always readily interconvert to a lower energy state at low temperatures
due to the $inetic barriers associated with the restructuring & such systems (e.g. highly
dispersed materials such as those in colloidal suspensions or supported metal catalysts are
thus -metastable-.
It should also be noted that there is a direct correspondence between the concepts of
-surface stability- and -surface free energy- i.e. surfaces of low surface free energy will be
more stable and vice versa.
=ne rule of thumb, is that the most stable solid surfaces are those with +
#. a high surface atom density
%. surface atoms of high coordination number
(Note & the two factors are obviously not independent, but are inevitably strongly
correlated.
Conse'uently, for example, if we consider the individual surface planes of an fcc metal,
then we would expect the stability to decrease in the order
fcc 111! 6 fcc 1((! 6 fcc 11(!
"arning & the above comments strictly only apply when the surfaces are in vacuum. "he
presence of a fluid above the surface ( gas or li'uid can drastically affect the surface free
energies as a result of the possibility of molecular adsorption onto the surface. 8referential
adsorption onto one or more of the surface planes can significantly alter the relative
stabilities of different planes & the influence of such effects under reactive conditions (e.g.
the high pressureIhigh temperature conditions pertaining in heterogeneous catalysis is
poorly understood.
1.7 .elaxation 8 .econstruction of Surfaces
"he phenomena of relaxation and reconstruction involve rearrangements of surface ( and
near surface atoms, this process being driven by the energetics of the system i.e. the
desire to reduce the surface free energy ( see #.: 2nergetics of 6urfaces . !s with all
processes, there may be $inetic limitations which prevent or hinder these rearrangements at
low temperatures.
0oth processes may occur with clean surfaces in ultrahigh vacuum, but it must be
remembered that adsorption of species onto the surface may enhance, alter or even reverse
the process 4
I. .elaxation
Relaxation is a small and subtle rearrangement of the surface layers which may
nevertheless be significant energetically, and seems to be commonplace for metal surfaces.
It involves ad<ustments in the layer spacings perpendicular to the surface , there is no
change either in the periodicity parallel to the surface or to the symmetry of the surface.
-nrelaxed Surface
.elaxed Surface
d
1-%
9 d
bulk
!
"he lower picture shows the relaxed surface + the first layer of atoms is typically drawn in
slightly towards the second layer (i.e. d
#&%
J d
bul$

.e can consider what might be the driving force for this process at the atomic level ....
If we use a localised model for the bonding in the solid then it is clear that an atom in the
bul$ is acted upon by a balanced, symmetrical set of forces.
=n the other hand, an atom at the unrelaxed surface suffers from an imbalance of forces
and the surface layer of atoms may therefore be pulled in towards the second layer.
(.hether this is a reasonable model for bonding in a metal is open to 'uestion 4
"he magnitude of the contraction in the first layer spacing is generally small ( J #9 K &
compensating ad<ustments to other layer spacings may extend several layers into the solid.
II. .econstruction
"he reconstruction of surfaces is a much more readily observable effect, involving larger
(yet still atomic scale displacements of the surface atoms. It occurs with many of the less
stable metal surfaces (e.g. it is fre'uently observed on fcc(##9 surfaces, but is much more
prevalent on semiconductor surfaces.
Enli$e relaxation, the phenomenon of reconstruction involves a change in the periodicity
of the surface structure & the diagram below shows a surface, viewed from the side, which
corresponds to an unreconstructed termination of the bul$ structure.
"his may be contrasted with the following picture which shows a schematic of a
reconstructed surface & this particular example is similar to the -missing row model-
proposed for the structure of a number of reconstructed (##9 fcc metal surfaces.
6ince reconstruction involves a change in the periodicity of the surface and in some cases
also a change in surface symmetry, it is readily detected using surface diffraction
techni'ues (e.g. ?22D C R>22D .
"he overall driving force for reconstruction is once again the minimi(ation of the surface
free energy & at the atomic level, however, it is not always clear why the reconstruction
should reduce the surface free energy. ,or some metallic surfaces, it may be that the change
in periodicity of the surface induces a splitting in surface&locali(ed bands of energy levels
and that this can lead to a lowering of the total electronic energy when the band is initially
only partly full.
In the case of many semiconductors, the simple reconstructions can often be explained in
terms of a -surface healing- process in which the co&ordinative unsaturation of the surface
atoms is reduced by bond formation between ad<acent atoms. ,or example, the formation
of a 6i(#99 surface re'uires that the bonds between the 6i atoms that form the new surface
layer and those that were in the layer immediately above in the solid are bro$en & this
leaves two -dangling bonds- per surface 6i atom.
! relatively small co&ordinated movement of the atoms in the topmost layer can reduce this
unsatisfied co&ordination & pairs of 6i atoms come together to form surface -6i dimers-,
leaving only one dangling bond per 6i atom. "his process leads to a change in the surface
periodicity + the period of the surface structure is doubled in one direction giving rise to the
so&called (%x# reconstruction observed on all clean 6i(#99 surfaces L 6i(#99&(%x# M.
*ore examples +

6i(###&(HxH "he web&page of the 68* group at the ?aboratory of !pplied 8hysics at
?in$Nping Eniversity.
In this section, attention has been concentrated on the reconstruction of clean surfaces. It is,
however, worth noting that reconstruction of the substrate surface is fre'uently induced by
the adsorption of molecular or atomic species onto the surface & this phenomenon is $nown
as adsorbate-induced reconstruction (see 6ection %.: for some examples.
Summary
"he minimisation of surface energy means that even single crystal surfaces will not exhibit
the ideal geometry of atoms to be expected by truncating the bul$ structure of the solid
parallel to a particular plane. "he differences between the real structure of the clean surface
and the ideal structure may be imperceptibly small (e.g. a very slight surface relaxation or
much more mar$ed and involving a change in the surface periodicity in one or more of the
main symmetry directions ( surface reconstruction .
1.: *articulate Metals
!s mentioned in the Introduction, macroscopic single crystals of metals are not generally
employed in technological applications.
*assive metallic structures (electrodes etc. are polycrystalline in nature & the si(e of
individual crystallites being determined by the mechanical treatment and thermal history of
the metal concerned. Nevertheless, the nature and properties of the exposed polycrystalline,
metal surface is still principally determined by the characteristics of the individual crystal
surfaces present. ,urthermore, the proportions in which the different crystal surfaces occur
is controlled by their relative thermodynamic stabilities. "hus, a macroscopic piece of an
fcc metal will generally expose predominantly (###&type surface planes.
! more interesting case for consideration is that of metals in a highly dispersed system &
the classic example of which is a supported metal catalyst (such as those employed in the
petrochemical industries and automotive catalytic converters. In such catalysts the average
metal particle si(e is invariably sub&micron and may be as small as # nm . "hese metal
particles are often tiny single crystals or simple twinned crystals.
"he shape of these small crystals is principally determined by the surface free energy
contribution to the total energy. "here are two ways in which the surface energy can be
reduced for a crystal of fixed mass I volume +
#. 0y minimi(ing the surface area of the crystallite
%. 0y ensuring that only surfaces of low surface free energy are exposed.
If matter is regarded as continuous then the optimum shape for minimi(ing the surface free
energy is a sphere (since this has the lowest surface areaIvolume ratio of any /D ob<ect &
this is why li'uid droplets in free space are basically spherical.
Enfortunately, we cannot ignore the discrete, atomic nature of matter and the detailed
atomic structure of surfaces when considering particles of the si(e found in catalysts. If, for
example, we consider an fcc metal (eg. 8t and ensure that only the most stable (###&type
surfaces are exposed, then we end up with a crystal which is an octahedron.
=ctahedron exposing
D symmetry&related
fcc(###&type faces
( Note + there are D different, but crystallographically&e'uivalent, surface planes which have
the (### surface structure & the O###P faces. "hey are related by the symmetry elements of
the cubic fcc system.
! compromise between exposing only the lowest energy surface planes and minimi(ing the
surface area is obtained by truncating the vertices of the octahedron & this generates a cubo&
octahedral particle as shown below, with D (###&type surfaces and @ smaller, (#99&type
surfaces and gives a lower (surface area I volume ratio.
Crystals of this general form are often used as conceptual models for the metal particles in
supported catalysts.
"he atoms in the middle of the O###P faces show the expected CN3G characteristic of the
(### surface. 6imilarly, those atoms in the centre of the O#99P surfaces have the
characteristic CN3D of the (#99 surface.
>owever, there are also many atoms at the corners and intersection of surface planes on the
particle which show lower coordination numbers.
.hat is the lowest coordination number exhibited by a surface atom on this
crystallite 5
"his model for the structure of catalytic metal crystallites is not always appropriate + it is
only reasonable to use it when there is a relatively wea$ interaction between the metal and
the support phase (e.g. many silica supported catalysts.
! stronger metal&support interaction is li$ely to lead to increased -wetting- of the support
by the metal, giving rise to +
a greater metal&support contact area
a significantly different metal particle morphology
,or example
In the case of a strong metal&support interaction the metalIoxide interfacial free energy is
low and it is inappropriate to consider the surface free energy of the metal crystallite in
isolation from that of the support.
=ur $nowledge of the structure of very small particles comes largely from high resolution
electron microscopy (>R2* studies & with the best modern microscopes it is possible to
directly observe such particles and resolve atomic structure.
2xamples Dr. D. Qefferson, Eniversity of Cambridge (EB
>R2* ,acility, North .estern Eniversity (E6
1.; <ther Single Crystal Surfaces
!lthough some of the more common metallic surface structures have been discussed in
previous sections (#.%&#.1, there are many other types of single crystal surface which may
be prepared and studied. "hese include
high index surfaces of metals
single crystal surfaces of compounds
"hese will not be covered in any depth, but a few illustrative examples are given below to
give you a flavour of the additional complexity involved when considering such surfaces.
,igh Index Surfaces of Metals
>igh index surfaces are those for which one or more of the *iller Indices are relatively
large numbers. "he most commonly studied surfaces of this type are vicinal surfaces which
are cut at a relatively small angle to one of the low index surfaces. "he ideal surfaces can
then be considered to consist of terraces which have an atomic arrangement identical with
the corresponding low index surface, separated by monatomic steps (steps which are a
single atom high.
8erspective view of the
fcc(HH: surface
!s seen above, the ideal fcc(HH: surface has a regular array of such steps and these steps
are both straight and parallel to one another.
.hat is the coordination number of a step atom on this surface 5
0y contrast a surface for which all the *iller indices differ must not only exhibit steps but
must also contain $in$s in the steps. !n example of such a surface is the fcc(#9.D.H surface
& the ideal geometry of which is shown below.
8erspective view of the fcc(#9.D.H
surface
.hat is the lowest coordination number exhibited by any of the atoms on this surface
5
Real vicinal surfaces do not, of course, exhibit the completely regular array of steps and
$in$s illustrated for the ideal surface structures, but they do exhibit this type of step and
terrace morphology. "he special adsorption sites available ad<acent to the steps are widely
recognised to have mar$edly different characteristics to those available on the terraces and
may thus have an important role in certain types of catalytic reaction.
,or further information on the structure of metal surfaces you should ta$e a loo$ at +
*etal 6urfaces.

An Atlas of Models of Crystal Surfaces , by Q.,. Nicholas
(Rordon C 0reach, #G@:
! compilation of fcc surface structures (selectable
from a stereographic map.
,CC Crystal 6urfaces + from 8er 6tolt(e, !alborg Eniversity,
Denmar$.
6urfaces of fcc, bcc, and hcp metals and the
diamond and graphite structures.
"he 6urface 2xplorer (based at the ,rit( >aber Institute & an
on&line interactive surface structure utility.
Single Crystal Surfaces of Compounds
"he ideal surface structures of the low index planes of compound materials can be easily
deduced from the bul$ structures in exactly the same way as can be done for the basic
metal structures. ,or example, the NaCl(#99 surface that would be expected from the bul$
structure is shown below +
8erspective view of the
NaCl(#99 surface
In addition to the relaxation and reconstruction exhibited by elemental surfaces, the
surfaces of compounds may also show deviations from the bul$ stoichiometry due to
surface localised reactions (e.g. surface reduction andIor surface segregation of one or
more components of the material.
,or further information on the surface structure of compound materials you should ta$e a
loo$ at +
=xide 6urfaces The Surface Science of Metal Oxides , by S.2. >enrich and
8.!. Cox (CE8, #GG1
6urfaces of compounds with the sodium chloride
(NaCl, caesium chloride (CsCl and cubic
(incblende (Fn6 structure.
"he 6urface 2xplorer (based at the ,rit( >aber Institute & an
on&line interactive surface structure utility.
6urfaces of compound semiconductors, alloys and
oxides.
Rallery of 8ictures from the NI6" 6urface 6tructure Database
(courtesy of 8rof. B. >ermann.
%. Adsorption of Molecules on Surfaces
%.1 Introduction to Molecular Adsorption
"he adsorption of molecules on to a surface is a necessary prere'uisite to any surface
mediated chemical process.
,or example, in the case of a surface catalysed reaction it is possible to brea$ down the
whole continuously&cycling process into the following five basic steps +
#. Diffusion of reactants to the active surface
%. !dsorption of one or more reactants onto the surface
/. 6urface reaction
1. Desorption of products from the surface
:. Diffusion of products away from the surface
"he above scheme not only emphasises the importance of the adsorption process but also
its reverse & namely desorption. It is these two processes which are considered in this
6ection.
)otes on Terminology 2
Substrate & fre'uently used to describe the solid surface onto which adsorption can occurA
the substrate is also occasionally (although not here referred to as the adsorbent.
Adsorbate & the general term for the atomic or molecular species which are adsorbed (or are
capable of being adsorbed onto the substrate.
Adsorption & the process in which a molecule becomes adsorbed onto a surface of another
phase (note & to be distinguished from absorption which is used when describing upta$e
into the bul$ of a solid or li'uid phase
Coverage & a measure of the extent of adsorption of a species onto a surface (unfortunately
this is defined in more than one way 4 . Esually denoted by the lower case Rree$ -theta-,
Exposure & a measure of the amount of gas which a surface has seenA more specifically, it is
the product of the pressure and time of exposure (normal unit is the ?angmuir, where # ? 3
#9
&@
"orr s .
%.% ,o# do Molecules 5ond to Surfaces $
"here are two principal modes of adsorption of molecules on surfaces +
*hysical Adsorption *hysisorption !
Chemical Adsorption Chemisorption !
"he basis of distinction is the nature of the bonding between the molecule and the surface.
.ith ...
hysical Adsorption + the only bonding is by wea$ San der .aals & type forces. "here is no
significant redistribution of electron density in either the molecule or at the substrate
surface.
Che!isorption + a chemical bond, involving substantial rearrangement of electron density,
is formed between the adsorbate and substrate. "he nature of this bond may lie anywhere
between the extremes of virtually complete ionic or complete covalent character.
Typical Characteristics of Adsorption *rocesses
Chemisorption *hysisorption
"emperature Range
(over which adsorption
occurs
Sirtually unlimited
(but a given molecule may
effectively adsorb only over a
small range
Near or below the
condensation point of the gas
(e.g. Te J #99 B, C=
%
J %99
B
!dsorption 2nthalpy
.ide range (related to the
chemical bond strength &
typically 19 & D99 $Q mol
&#

Related to factors li$e
molecular mass and polarity
but typically :&19 $Q mol
&#
(i.e.
) heat of li'uefaction
Crystallographic 6pecificity
(variation between different
surface planes of the same
crystal
*ar$ed variation between
crystal planes
Sirtually independent of
surface atomic geometry
Nature of !dsorption
=ften dissociative
*ay be irreversible
Non&dissociative
Reversible
6aturation Epta$e ?imited to one monolayer *ultilayer upta$e possible
Binetics of !dsorption
Sery variable & often an
activated process
,ast & since it is a non&
activated process
"he most definitive method for establishing the formation of a chemical bond between the
adsorbing molecule and the substrate ( i.e. chemisorption is to use an appropriate
spectroscopic techni'ue, for example
IR ( see 6ection :.1 & to observe the vibrational fre'uency of the substrateIadsorbate bond
E86 ( see 6ection :./ & to monitor intensity C energy shifts in the valence orbitals of the
adsorbate and substrate
%.& Adsorption =inetics & "he Rate of !dsorption
"he rate of adsorption, "
ads
, of a molecule onto a surface can be expressed in the same
manner as any $inetic process. ,or example, when it is expressed in terms of the partial
pressure of the molecule in the gas phase above the surface+

R
ads
> k' P
x
where+ x & $inetic order
#$ & rate constant
& partial pressure
If the rate constant is then expressed in an !rrhenius form, then we obtain a $inetic
e'uation of the form +
R
ads
> A exp -E
a
? RT !. P
x

where E
a
is the activation energy for adsorption, and A the pre-exponential %fre&uency'
factor.
It is much more informative, however, to consider the factors controlling this process at the
molecular level....
"he rate of adsorption is governed by
#. the rate of arrival of molecules at the surface
%. the proportion of incident molecules which undergo adsorption
i.e. we can express the rate of adsorption (per unit area of surface as a product of the
incident molecular flux, ( , and the stic$ing probability , S .
R
ads
> S . F Lmolecules m
&%
s
&#
M
"he flux of incident molecules is given by the >ert(&Bnudsen e'uation
,lux , F > P ? %mkT!
1?%
L molecules m
&%
s
&#
M
where
& gas pressure L N m
&%
M
! & mass of one molecule L $g M
T & temperature L B M
"he stic$ing probability is clearly a property of the adsorbate I substrate system under
consideration but must lie in the range 9 J S J #A it may depend upon various factors &
foremost amongst these being the existing coverage of adsorbed species ( and the
presence of any activation barrier to adsorption. In general ,therefore ,
S > f ! . exp -E
a
? RT !
where, once again, E
a
is the activation energy for adsorption and f( is some, as yet
undetermined, function of the existing surface coverage of adsorbed species.
Combining the e'uations for S and ( yields the following expression for the rate of
adsorption +
Notes +
#. "he above e'uation indicates that the rate of adsorption is expected to be first order
with regard to the partial pressure of the molecule in the gas phase above the
surface.
%. It should be recognised that the activation energy for adsorption may itself be
dependent upon the surface coverage , i.e. E
a
3 E(.
/. If it is further assumed that the stic$ing probability is directly proportional to the
concentration of vacant surface sites (which would be a reasonable first
approximation for non&dissociative adsorption then f( is proportional to (#& A
where, in this instance, is the fraction of sites which are occupied (i.e. the
?angmuir definition of surface coverage.
,or a discussion of some of the factors which determine the magnitude of the activation
energy of adsorption you should see 6ection %.1 which loo$s at the typical 82 curve
associated with various types of adsorption process.
4stimating Surface Co'erages arising as a result of @as 4xposure
If a surface is initially clean and it is then exposed to a gas pressure under conditions where
the rate of desorption is very slow, then the coverage of adsorbed molecules may initially
be estimated simply by consideration of the $inetics of adsorption.
!s noted above, the rate of adsorption is given by + R
ads
> S . F
i.e.
where + )
ads
is the number of adsorbed species per unit area of surface.
In general, this e'uation must be integrated to obtain an expression for )
ads
, since the
stic$ing probability is coverage (and hence also time dependent.
>owever, if it is assumed that the stic$ing probability is essentially constant (which may be
a reasonable approximation for relatively low coverages, then this integration simply
yields+
%.0 *4 Cur'es 8 4nergetics of Adsorption
In this section we will consider both the energetics of adsorption and factors which
influence the $inetics of adsorption by loo$ing at the -potential energy diagramIcurve- for
the adsorption process.
"he potential energy curve for the adsorption process is a representation of the variation of
the energy (82 or E of the system as a function of the distance (d of an adsorbate from a
surface.
.ithin this simple one&dimensional (#D model, the only variable is the distance (d of the
adsorbing molecule from the substrate surface. "hus, the energy of the system is a function
only of this variable i.e. E 3 E(d
It should be remembered that this is a very simplistic model which neglects many other
parameters which influence the energy of the system (a single molecule approaching a
clean surface, including for example
the angular orientation of the !olecule
changes in the internal bond angles and bond lengths of the !olecule
the position of the !olecule parallel to the surface plane
"he interaction of a molecule with a given surface will also clearly be dependent upon the
presence of any existing adsorbed species, whether these be surface impurities or simply
pre&adsorbed molecules of the same type (in the latter case we are starting to consider the
effect of surface coverage on the adsorption characteristics.
Nevertheless it is useful to first consider the interaction of an isolated molecule with a
clean surface using the simple #D model. ,or the purposes of this tutorial we will also not
be overly concerned whether the -energy- being referred to should strictly be the internal
energy, the enthalpy or free energy of the system.
CAS4 I - *hysisorption
In the case of pure physisorption ( e.g. !r I metals , the only attraction between the
adsorbing species and the surface arises from wea$, van der .aals forces.
!s illustrated below, these forces give rise to a shallow minimum in the 82 curve at a
relatively large distance from the surface (typically d U 9./ nm before the strong repulsive
forces arising from electron density overlap cause a rapid increase in the total energy.
"here is no barrier to prevent the atom or molecule which is approaching the surface from
entering this physisorption well, i.e. the process is not activated and the $inetics of
physisorption are invariably fast.
CAS4 II - *hysisorption A Molecular Chemisorption
"he wea$ physical adsorption forces and associated long&range attraction will be present to
varying degrees in all adsorbate I substrate systems. >owever, in cases where chemical
bond formation between the adsorbate and substrate can also occur, the 82 curve is
dominated by a much deeper chemisorption minimum at shorter values of d .
"he graph above shows the 82 curves due to physisorption and chemisorption separately &
in practice, the 82 curve for any real molecule capable of undergoing chemisorption is best
described by a combination of the two curves, with a curve crossing at the point at which
chemisorption forces begin to dominate over those arising from physisorption alone.
"he minimum energy pathway obtained by combining the two 82 curves is now
highlighted in red. !ny perturbation of the combined 82 curve from the original, separate
curves is most li$ely to be evident close to the highlighted crossing point.
,or clarity, we will now consider only the overall 82 curve +
"he depth of the chemisorption well is a measure of the strength of binding to the surface &
in fact it is a direct representation of the energy of adsorption, whilst the location of the
global minimum on the hori(ontal axis corresponds to the e'uilibrium bond distance (r
e

for the adsorbed molecule on this surface.
"he energy of adsorption is negative, and since it corresponds to the energy change upon
adsorption it is better represented as E(ads or E
ads
. >owever, you will also often find
the depth of this well associated with the enthalpy of adsorption, *(ads.
(Note & the -heat of adsorption- , + , is ta$en to be a positive 'uantity e'ual in magnitude
to the enthalpy of adsorption A i.e. + 3 &*(ads
In this particular case, there is clearly no barrier to be overcome in the adsorption process
and there is no activation energy of adsorption (i.e. E
a
ads
3 9 , but do remember the
previously mentioned limitations of this simple #D model.
"here is of course a significant barrier to the reverse, desorption process & the red arrow in
the diagram below represents the activation energy for desorption.
Clearly in this particular case, the magnitudes of the energy of adsorption and the
activation energy for desorption can also be e'uated i.e.
E
a
des
3 E (ads or E
a
des
) &* (ads
CAS4 III - *hysisorption A Bissociati'e Chemisorption
In this case the main differences arise from the substantial changes in the 82 curve for the
chemisorption process
!gain, we start off with the basic 82 curve for the physisorption process which represents
how the molecule can wea$ly interact with the surface +
If we now consider a specific molecule such as >
%
and initially treat it as being completely
isolated from the surface ( i.e. when the distance, d , is very large then a substantial
amount of energy has to be put into the system in order to cause dissociation of the
molecule.
>
%
> V >
& this is the bond dissociation energy L ,(>&> M, some 1/: $Q mol
&#
or 1.: eS.
"he red dot in the diagram above thus represents two hydrogen atoms, e'uidistant (and a
long distance from the surface and also now well separated from each other. If these atoms
are then allowed to approach the surface they may ultimately both form strong chemical
bonds to the substrate .... this corresponds to the minimum in the red curve which
represents the chemisorption 82 curve for the two > atoms.
In reality, of course, such a mechanism for dissociative hydrogen chemisorption is not
practical & the energy downpayment associated with brea$ing the >&> bond is far too
severe.
Instead, a hydrogen molecule will initially approach the surface along the physisorption
curve. If it has sufficient energy it may pass straight into the chemisorption well ( -direct
chemisorption- ....
or, alternatively, it may first undergo transient physisorption & a state from which it can
then either desorb bac$ as a molecule into the gas phase or cross over the barrier into the
dissociated, chemisorptive state (as illustrated sche!atically below.
In this latter case, the molecule can be said to have undergone -precursor&mediated-
chemisorption.
"he characteristics of this type of dissociative adsorption process are clearly going to be
strongly influenced by the position of the crossing point of the two curves (molecular
physisorption vWs dissociative chemisorption & relatively small shifts in the position of
either of the two curves can significantly alter the si(e of any barrier to chemisorption.
In the example immediately below there is no direct activation barrier to dissociative
adsorption & the curve crossing is below the initial -(ero energy- of the system.
whilst, in this next case 7.
there is a substantial barrier to chemisorption. 6uch a barrier has a ma<or influence on the
kinetics of adsorption.
)ote - the depth of the physisorption -ell for the hydrogen !olecule is actually very s!all
%in so!e cases negligible'. but this is not the case for other !olecules and does not alter
the basic conclusions regarding dissociative adsorption that result fro! this !odel/
na!ely that the process !ay be either activated or non-activated depending on the exact
location of the curve crossing0
!t this point it is useful to return to consider the effect of such a barrier on the relationship
between the activation energies for adsorption and desorption, and the energy (or enthalpy
of adsorption.
Clearly, from the diagram
E
a
des
& E
a
ads
3 & E
ads
but, since the activation energy for adsorption is nearly always very much smaller than that
for desorption, and the difference between the energy and enthalpy of adsorption is also
very small, it is still 'uite common to see the relationship
E
a
des
) &*
ads
,or a slightly more detailed treatment of the adsorption process, you are referred to the
following examples of *ore Complex 82 Curves C *ulti&Dimensional 82 6urfaces.
%.3 Adsorbate @eometries 8 Structures on Metals
.e can address the 'uestion of what happens when a molecule becomes adsorbed onto a
surface at two levelsA specifically we can aim to identify
#. the nature of the adsorbed species and its local adsorption geometry (i.e. its
chemical structure and co&ordination to ad<acent substrate atoms
%. the overall structure of the extended adsorbateIsubstrate interface (i.e. the long
range ordering of the surface
"he latter topic is covered in detail in 6ection @.# + Classification of =verlayer 6tructures
whist this section will consider only the local adsorption geometry and adsorbate structure.
Chemisorption, by definition, involves the formation of new chemical bonds between the
adsorbed species and the surface atoms of the substrate & basically the same type of bonds
that are present in any molecular complex. In considering what type of species may be
formed on a metal surface, therefore, it is important not to abandon chemical common
sense and, if in doubt, to loo$ for inspiration at the structures of $nown metal&organic
complexes.
Chemisorption of ,ydrogen and ,alogens
*ydrogen %*1 '
In the >% molecule, the valence electrons are all involved in the >&> &bond and there are
no additional electrons which may interact with the substrate atoms. Conse'uently,
chemisorption of hydrogen on metals is almost invariably a dissociative process in which
the >&> bond is bro$en, thereby permitting the hydrogen atoms to independently interact
with the substrate (see 6ection %.1 for a description of the energetics of this process. "he
adsorbed species in this instance therefore are hydrogen atoms.
"he exact nature of the adsorbed hydrogen atom complex is generally difficult to determine
experimentally, and the very small si(e of the hydrogen atom does mean that migration of
hydrogen from the interface into sub&surface layers of the substrate can occur with relative
ease on some metals (e.g. 8d, rare earth metals.
"he possibility of molecular >
%
chemisorption at low temperatures cannot be entirely
excluded, however, as demonstrated by the discovery of molecular hydrogen transition
metal compounds, such as .(
%
&>% (C=/(8i&8r/ / , in which both atoms of the hydrogen
molecule are coordinated to a single metal centre.
*alogens %(
1
. Cl
1
. 2r
1
etc0'
>alogens also chemisorb in a dissociative fashion to give adsorbed halogen atoms. "he
reasons for this are fairly clear & in principle a halogen molecule could act as a ?ewis base
and bind to the surface without brea$age of the T&T bond, in practice the lone pairs are
strongly held by the highly electronegative halogen atom so any such interaction would be
very wea$ and the thermodynamics lie very heavily in favour of dissociative adsorption
L i.e. D(T&T V D(*&T% JJ % D(*&T M. Clearly the $inetic barrier to dissociation must
also be low or non&existent for the dissociative adsorption to occur readily.
!nother way of loo$ing at the interaction of a halogen molecule with a metal surface is as
follows + the significant difference in electronegativity between a typical metal and halogen
is such that substantial electron transfer from the metal to halogen is favoured. If a halogen
molecule is interacting with a metal surface then this transferred electron density will enter
the X antibonding orbital of the molecule, thereby wea$ening the T&T bond. !t the same
time the build&up of negative charge on the halogen atoms enhances the strength of the
metal&halogen interaction. "he net result of these two effects when ta$en to their limit is
that the halogen molecule dissociates and the halogen atoms interact with the metal with a
strong ionic contribution to the bonding.
>alogen atoms tend to occupy high co&ordination sites on the surface & for example, the /&
fold hollow site on fcc(### surfaces (A and the 1&fold hollow site on fcc(#99 surfaces
(5.
(A lan 3ie- (5
"his behaviour is typical of atomic adsorbates which almost invariably endeavour to
maximise their co&ordination and hence prefer to occupy the highest&available co&
ordination site on the surface.
!s a result of the electron transfer from the metal substrate to the halogen atoms, each
adsorbed atom is associated with a significant surface dipole.
Cross-section
=ne conse'uence of this is that there are repulsive (dipole&dipole interactions between the
adsorbed atoms, which are especially evident at higher surface coverages and which can
lead to a substantial reduction in the enthalpy of adsorption at specific coverages (if these
coverages mar$ a watershed, above which the atoms are forced to occupy sites which are
much closer together.
!nother feature of the halogen adsorption chemistry of some metals is the transition from
an adsorbed surface layer to surface compound formation at high gas exposures.
Chemisorption of )itrogen and <xygen
Oxygen
=xygen is an example of a molecule which usually adsorb dissociatively, but are also
found to adsorb molecularly on some metals (e.g. !g, 8t. In those cases where both types
of adsorption are observed it is the dissociative process that corresponds to the higher
adsorption enthalpy.
!s noted above, in the molecular adsorption state the interaction between the molecule and
the surface is relatively wea$. *olecules aligned such that the internuclear axis is parallel
to the surface plane may bond to a single metal atom of the surface via both
#. &donor interaction, in which the charge transfer is from the occupied molecular &
bonding molecular orbital of the molecule into vacant orbitals of &symmetry on
the metal (i.e. * =% , and
%. &acceptor interaction, in which an occupied metal d&orbital of the correct
symmetry overlaps with empty X orbitals of the molecule and the charge transfer
is from the surface to the molecule (i.e. * =% .
!lthough the interaction of the molecule with the surface is generally wea$, one might
expect that there might be a substantial barrier to dissociation due to the high strength (and
high dissociation enthalpy of the =3= bond. Nevertheless on most metal surfaces,
dissociation of oxygen is observed to be facile which is related to the manner in which the
interaction with the surface can mitigate the high intrinsic bond energy (see 6ection %.1
and thereby facilitate dissociation.
=nce formed, oxygen atoms are strongly bound to the surface and, as noted previously,
will tend to occupy the highest available co&ordination site. "he strength of the interaction
between adsorbate and substrate is such that the ad<acent metal atoms are often seen to
undergo significant displacements from the e'uilibrium positions that they occupy on the
clean metal surface. "his displacement may simply lead to a distortion of the substrate
surface in the immediate vicinity of the adsorbed atom (so that, for example, the ad<acent
metal atoms are drawn in towards the oxygen and the metal&oxygen bond distance is
reduced or to a more extended surface reconstruction (see 6ection #.@
Dissociative oxygen adsorption is fre'uently irreversible & rather than simply leading to
desorption, heating of an adsorbed oxygen overlayer often results in either the gradual
removal of oxygen from the surface by diffusion into the bul$ of the substrate (e.g. 6i(###
or Cu(### or to the formation of a surface oxide compound. 2ven at ambient
temperatures, extended oxygen exposure often leads to the nucleation of a surface oxide.
Depending on the reactivity of the metal concerned, further exposure at low temperatures
may result either in a progressive conversion of the bul$ material to oxide or the oxidation
process may effectively stop after the formation of a passivating surface oxide film of a
specific thic$ness (e.g. !l.
)itrogen
"he interaction of nitrogen with metal surfaces shows many of the same characteristics as
those described above for oxygen. >owever, in general N% is less susceptible to dissociation
as a result of the lower *&N bond strength and the substantial $inetic barrier associated
with brea$ing the N N triple bond.
Chemisorption of Carbon Monoxide
Depending upon the metal surface, carbon monoxide may adsorb either in a molecular
form or in a dissociative fashion & in some cases both states coexist on particular surface
planes and over specific ranges of temperature.
#. =n the reactive surfaces of metals from the left&hand side of the periodic table (e.g.
Na, Ca, "i, rare earth metals the adsorption is almost invariably dissociative,
leading to the formation of adsorbed carbon and oxygen atoms (and thereafter to
the formation of surface oxide and oxy&carbide compounds.
%. 0y contrast, on surfaces of the metals from the right hand side of the d&bloc$ (e.g.
Cu, !g the interaction is predominantly molecularA the strength of interaction
between the C= molecule and the metal is also much wea$er, so the *&C= bond
may be readily bro$en and the C= desorbed from the surface by raising the surface
temperature without inducing any dissociation of the molecule.
/. ,or the ma<ority of the transition metals, however, the nature of the adsorption
(dissociative v.Ws molecular is very sensitive to the surface temperature and surface
structure (e.g. the *iller index plane, and the presence of any lower co&ordination
sites such as step sites and defects.
*olecularly chemisorbed C= has been found to bond in various ways to single crystal
metal surfaces & analogous to its behaviour in isolated metal carbonyl complexes.
"erminal (-?inear-
(all surfaces
0ridging ( %f site
(all surfaces
0ridging I /f hollow
( fcc(###
0ridging I 1f hollow
(rare & fcc(#99 5

.hilst the above structural diagrams amply demonstrate the inade'uacies of a simple
valence bond description of the bonding of molecules to surface, they do to an extent also
illustrate one of its features and strengths & namely that a given element, in this case carbon,
tends to have a specific valence. Conse'uently, as the number of metal atoms to which the
carbon is co&ordinated increases, so there is a corresponding reduction in the C&= bond
order.
>owever, it must be emphasised that a molecule such as C= does not necessarily prefer to
bind at the highest available co&ordination site. 6o, for example, the fact that there are /&
fold hollow sites on an fcc(### surface does not mean that C= will necessarily adopt this
site & the preferred site may still be a terminal or %&fold bridging site, and the site or site(s
which is(are occupied may change with either surface coverage or temperature. "he
energy difference between the various adsorption sites available for molecular C=
chemisorption appears therefore to be very small.
L ! description of the nature of the bonding in a terminal C=&metal complex, in terms of a
simple molecular orbital model, is given in 6ection :.1 M
Chemisorption of Ammonia and other @roup +?+I ,ydrides
!mmonia has lone pairs available for bonding on the central nitrogen atom and may bond
without dissociation to a single metal atom of a surface, acting as a ?ewis base, to give a
pseudo&tetrahedral co&ordination for the nitrogen atom.
!lternatively, progressive dehydrogenation may occur to give surface N>
x
(x 3 %,#,9
species and adsorbed hydrogen atoms, i.e.
N>/ N>% (ads
V >
(ads
N>
(ads
V % >
(ads
N
(ads
V / >
(ads
!s the number of hydrogens bonded to the nitrogen atom is reduced, the adsorbed species
will tend to move into a higher co&ordination site on the surface (thereby tending to
maintain the valence of nitrogen.
,eco!position frag!ents of a!!onia on an fcc%444' surface
(8icture adapted from the 0!?6!C 8icture Rallery by B. >ermann, ,rit(&>aber&Institut, 0erlin
=ther Rroup S and Rroup SI hydrides (e.g. 8>/ , >%=, >%6 exhibit similar adsorption
characteristics to ammonia.
Chemisorption of -nsaturated ,ydrocarbons
Ensaturated hydrocarbons (al$enes, al$ynes, aromatic molecules etc. all tend to interact
fairly strongly with metal atom surfaces. !t low temperatures (and on less reactive metal
surfaces the adsorption may be molecular, albeit perhaps with some distortion of bond
angles around the carbon atom.
2thene, for example, may bond to give both a &complex (A or a di& adsorption complex
(5+
(A Che!isorbed (5
Ethene
L ,urther examples + models of chemisorbed ethene and ethyne on Cu(### M
!s the temperature is raised, or even at low temperatures on more reactive surfaces (in
particular those that bind hydrogen strongly, a stepwise dehydrogenation may occur. =ne
particularly stable surface intermediate found in the dehydrogenation of ethene is the
ethylidyne complex, whose formation also involves >&atom transfer between the carbon
atoms.
Ethylidyne 5
this adsorbate preferentially occupies a
6-fold hollo- site to give pseudo-
tetrahedral co-ordination for the carbon
ato!0
"he ultimate product of complete dehydrogenation, and the loss of molecular hydrogen by
desorption, is usually either carbidic or graphitic surface carbon.
%.7 The Besorption *rocess
!n adsorbed species present on a surface at low temperatures may remain almost
indefinitely in that state. !s the temperature of the substrate is increased, however, there
will come a point at which the thermal energy of the adsorbed species is such that one of
several things may occur +
#. a molecular species may decompose to yield either gas phase products or other
surface species.
%. an atomic adsorbate may react with the substrate to yield a specific surface
compound, or diffuse into the bul$ of the underlying solid.
/. the species may desorb from the surface and return into the gas phase.
"he last of these options is the desorption process. In the absence of decomposition the
desorbing species will generally be the same as that originally adsorbed but this is not
necessarily always the case.
(!n example where it is not is found in the adsorption of some al$ali metals on metallic
substrates exhibiting a high wor$ function where, at low coverages, the desorbing species
is the al$ali metal ion as opposed to the neutral atom. =ther examples would include
certain isomerisation reactions.
Besorption =inetics
"he rate of desorption, "
des
, of an adsorbate from a surface can be expressed in the general
form +
R
des
3 k N
x
where x & $inetic order of desorption
# & rate constant for the desorption process
) & surface concentration of adsorbed species
"he order of desorption can usually be predicted because we are concerned with an
elementary step of a -reaction- + specifically,
I. Ato!ic or Si!ple Molecular ,esorption
!
(ads
!
(g

*
(ads
*
(g

& will usually be a first order process ( i.e. x 3 # . 2xamples include 7
. I Cu
(ads
.
(s
V Cu
(g
A desorption of Cu atoms from a . surface
Cu I C=
(ads
Cu
(s
V C=
(g
A desorption of C= molecules from a Cu surface
II. "eco!binative Molecular ,esorption
% !
(ads
!
% (g

& will usually be a second order process ( i.e. x 3 % . 2xamples include 7
8t I =
(ads
8t
(s
V =
% (g
A desorption of = atoms as =
%
from a 8t surface
Ni I >
(ads
Ni
(s
V >
% (g
A desorption of > atoms as >
%
from a Ni surface
"he rate constant for the desorption process may be expressed in an !rrhenius form,
k
des
3 A exp -E
a
des
? RT !
where E
a
des
is the activation energy for desorption , and

A is the pre&exponential factorA this can also be considered to be the -attempt
fre'uency-, , at overcoming the barrier to desorption.
"his then gives the following general expression for the rate of desorption
In the particular case of simple molecular adsorption, the pre&exponentialIfre'uency factor
( may also be e'uated with the fre'uency of vibration of the bond between the molecule
and substrateA this is because every time this bond is stretched during the course of a
vibrational cycle can be considered an attempt to brea$ the bond and hence an attempt at
desorption.
Surface .esidence Times
=ne property of an adsorbed molecule that is intimately related to the desorption $inetics is
the surface residence ti!e & this is the average time that a molecule will spend on the
surface under a given set of conditions (in particular, for a specified surface temperature
before it desorbs into the gas phase.
,or a first order process such as the desorption step of a molecularly adsorbed species +
M
(ads
M
(g
the average time ( prior to the process occurring is given by +
> 1?k
#
where #
#
is the first order rate constant (no proof given.
,rom the previously derived desorption formulae we $now that
k
#
> exp -E
a
des
? RT !
and if we also substitute for E
a
des
using the approximate relation E
a
des
) &*
ads
discussed in
6ection %.1 , then we get the following expression for the surface residence time
>
o
exp -H
ads
? RT !
where +

o
( 3 #I corresponds to the period of vibration of the bond between the adsorbed
molecule and substrate and is fre'uently ta$en to be about #9
&#/
s .
&. The Cangmuir Isotherm
&.1 The Cangmuir Isotherm
.henever a gas is in contact with a solid there will be an e'uilibrium established between
the molecules in the gas phase and the corresponding adsorbed species (molecules or
atoms which are bound to the surface of the solid.
!s with all chemical e'uilibria, the position of e'uilibrium will depend upon a number of
factors +
#. "he relative stabilities of the adsorbed and gas phase species involved
%. "he temperature of the system (both the gas and surface, although these are
normally the same
/. "he pressure of the gas above the surface
In general, factors (% and (/ exert opposite effects on the concentration of adsorbed
species & that is to say that the surface coverage may be increased by raising the gas
pressure but will be reduced if the surface temperature is raised.
"he ?angmuir isotherm was developed by Irving ?angmuir in #G#@ to describe the
dependence of the surface coverage of an adsorbed gas on the pressure of the gas above the
surface at a fixed temperature. "here are many other types of isotherm ("em$in, ,reundlich
... which differ in one or more of the assumptions made in deriving the expression for the
surface coverageA in particular, on how they treat the surface coverage dependence of the
enthalpy of adsorption. .hilst the ?angmuir isotherm is one of the simplest, it still
provides a useful insight into the pressure dependence of the extent of surface adsorption.
Important )ote & 6urface Coverage C the ?angmuir Isotherm
.hen considering adsorption isotherms it is conventional to adopt a definition of surface
coverage ( which defines the maximum (saturation surface coverage of a particular
adsorbate on a given surface always to be unity, i.e.
max
3 # .
This -ay of defining the surface coverage differs fro! that usually adopted in surface
science where the more common practice is to e'uate with the ratio of adsorbate species
to surface substrate atoms (which leads to saturation coverages which are almost invariably
less than unity.
&.% Cangmuir Isotherm - deri'ation from eDuilibrium
considerations
.e may derive the ?angmuir isotherm by treating the adsorption process as we would any
other e'uilibrium process & except in this case the e'uilibrium is between the gas phase
molecules (*, together with vacant surface sites, and the species adsorbed on the surface.
"hus, for a non&dissociative (molecular adsorption process we consider the adsorption to
be represented by the following chemical e'uation +
S - E A M
(g
3 S - M
where + S - E , represents a vacant surface site
Note - in -riting this e&uation -e are !a#ing an inherent assu!ption that there are a fixed
nu!ber of localised surface sites present on the surface0 This is the first !a7or assu!ption
of the 8ang!uir isother!0
.e may now define an e'uilibrium constant ( 9 in terms of the concentrations of
-reactants- and -products-
.e may also note that +
L 6 & * M is proportional to the surface coverage of adsorbed molecules, i.e.
proportional to
L 6 & X M is proportional to the number of vacant sites, i.e. proportional to (#&)
L * M is proportional to the pressure of gas ,
>ence, it is also possible to define another e'uilibrium constant, b , as given below +
Rearrangement then gives the following expression for the surface coverage
which is the usual form of expressing the ?angmuir Isotherm.
!s with all chemical reactions, the e'uilibrium constant, b , is both temperature&dependent
and related to the Ribbs free energy and hence to the enthalpy change for the process.
Note - b is only a constant %independent of ' if the enthalpy of adsorption is independent
of coverage - this is the second !a7or assu!ption of the 8ang!uir Isother!0
&.& Cangmuir Isotherm - deri'ation from kinetic
considerations
"he e'uilibrium that may exist between gas adsorbed on a surface and molecules in the gas
phase is a dynamic state, i.e. the e'uilibrium represents a state in which the rate of
adsorption of molecules onto the surface is exactly counterbalanced by the rate of
desorption of molecules bac$ into the gas phase. It should therefore be possible to derive an
isotherm for the adsorption process simply by considering and e'uating the rates for these
two processes.
2xpressions for the rate of adsorption and rate of desorption have been derived in 6ections
%./ C %.@ respectively + specifically ,

2'uating these two rates yields an e'uation of the form +
L#M
where the terms f( C f W( contain the pre&exponential surface coverage dependence of
the rates of adsorption and desorption respectively and all other factors have been ta$en
over to the right hand side to give a temperature&dependent -constant- characteristic of this
particular adsorption process, C(T .
.e now need to ma$e certain simplifying assumptions ... the first is one of the $ey
assumptions of the ?angmuir isotherm
!dsorption ta$es place only at specific locali(ed sites on the surface and the saturation
coverage corresponds to complete occupancy of these sites.
?et us initially further restrict our consideration to a simple case of re'ersible molecular
adsorption. i.e.
S - A M
(g
3 S - M
where S - represents a vacant surface site and S - M the adsorption complex.
Ender these circumstances it is reasonable to assume coverage dependencies for rates of
the two processes of the form +
!dsorption, f ! > c 1- ! i.e. proportional to the fraction of sites that are unoccupied.
Desorption, f F ! > cF
i.e. proportional to the fraction of sites which are occupied by
adsorbed molecules.
where is the fraction of sites occupied at e'uilibrium.
Note 5 these coverage dependencies are exactly -hat -ould be predicted by noting that the
for-ard and reverse processes are ele!entary reaction steps . in -hich case it follo-s
fro! standard che!ical #inetic theory that
#. the for-ard adsorption process -ill exhibit #inetics having a first order
dependence on the concentration of vacant surface sites0
%. the reverse desorption process -ill exhibit #inetics having a first order dependence
on the concentration of adsorbed !olecules0
6ubstitution into e'uation L#M then yields
where 2(T 3 (cWIc.C(T . !fter rearrangement this gives the ?angmuir Isotherm
expression for the surface coverage
where b ( 3 #I2(T is a function of temperature and contains an exponential term of the
form
b > ...... exp G E
a
des
& E
a
ads
! ? R T H > ...... exp G - H
ads
? R T H
Conse'uently, b can only be regarded as a constant with respect to coverage if the enthalpy
of adsorption is itself independent of coverage & this is the second ma<or assumption of the
?angmuir Isotherm.
,urther study + .hat is the ?angmuir Isotherm for dissociative adsorption 5
&.0 +ariation of Surface Co'erage #ith Temperature 8
*ressure
!pplication of the assumptions of the ?angmuir Isotherm leads to readily derivable
expressions for the pressure dependence of the surface coverage (see 6ections /.% and /./
& in the case of a simple, reversible molecular adsorption process the expression is
where b 3 b("
"his is illustrated in the graph below which shows the characteristic ?angmuir variation of
coverage with pressure for molecular adsorption.
Note that +
#. b at low pressures
%. # at high pressures
!t a given pressure the extent of adsorption is determined by the value of b + this in turn is
dependent upon both the temperature (" and the enthalpy (heat of adsorption. Remember
that the magnitude of the adsorption enthalpy (a negative 'uantity itself reflects the
strength of binding of the adsorbate to the substrate.
"he value of b is increased by
#. a reduction in the system temperature
%. an increase in the strength of adsorption
"herefore the set of curves shown below illustrates the effect of either (i increasing the
magnitude of the adsorption enthalpy at a fixed temperature, or (ii decreasing the
temperature for a given adsorption system.
! given e'uilibrium surface coverage may be attainable at various combinations of
pressure and temperature as highlighted below 7 note that as the temperature is lowered
the pressure re'uired to achieve a particular e'uilibrium surface coverage decreases.
& this is often used as <ustification for one of the main ideologies of surface chemistry A
specifically, that it is possible to study technologically&important (high pressure I high
temperature surface processes within the low pressure environment of typical surface
analysis systems by wor$ing at low temperatures. It must be recognised however that, at
such low temperatures, $inetic restrictions that are not present at higher temperatures may
become important.
If you wish to see how the various factors relating to the adsorption and desorption of
molecules influence the surface coverage then try out the Interactive Demonstration of the
?angmuir Isotherm (note & this is based on the derivation given in 6ection /./ .
Betermination of 4nthalpies of Adsorption
It has been shown in previous sections how the value of b is dependent upon the enthalpy
of adsorption. It has also <ust been demonstrated how the value of b influences the
pressureItemperature (8&" dependence of the surface coverage.
"he implication of this is that it must be possible to determine the enthalpy of adsorption
for a particular adsorbateIsubstrate system by studying the 8&" dependence of the surface
coverage.
Sarious methods based upon this idea have been developed for the determination of
adsorption enthalpies & one method is outlined below +
Step 1 + Involves
determination of a number of
adsorption isotherms (where a
single isotherm is a coverage I
pressure curve at a fixed
temperature.
Step % + It is then possible to
read off a number of pairs of
values of pressure and
temperature which yield the
same surface coverage
Step & + "he Clausius&
Clapeyron e'uation
may then be applied to this set
of (8&" data and a plot of ( ln
8 vWs (#I" should give a
straight line, the slope of
which yields the adsorption
enthalpy.
Note +
the value obtained for the adsorption enthalpy is that pertaining at the surface
coverage for which the 8&" data was obtained, but steps % C / may be repeated for
different surface coverages enabling the adsorption enthalpy to be determined over
the whole range of coverages.
this method is applicable only when the adsorption process is thermodynamically
reversible.
&.3 Applications 2 =inetics of Catalytic .eactions
It is possible to predict how the $inetics of certain heterogeneously&catalysed reactions
might vary with the partial pressures of the reactant gases above the catalyst surface by
using the ?angmuir isotherm expression for e'uilibrium surface coverages.
4xample 1 2 -nimolecular Becomposition
Consider the surface decomposition of a molecule ! , i.e. the process
A
(g
A
(ads
*roducts
?et us assume that +
#. "he decomposition reaction occurs uniformly across the surface sites at which
molecule ! may be adsorbed and is not restricted to a limited number of special
sites.
%. "he products are very wea$ly bound to the surface and, once formed, are rapidly
desorbed.
/. "he rate determining step (rds is the surface decomposition step.
Ender these circumstances, the molecules of ! adsorbed on the surface are in e'uilibrium
with those in the gas phase and we may predict the surface concentration of ! from the
?angmuir isotherm i.e.
> b.P ? 1 A b.P !
"he rate of the surface decomposition (and hence of the reaction is given by an expression
of the form
.ate > k
()ote - -e are assu!ing that the deco!position of A%ads' occurs in a si!ple uni!olecular
ele!entary reaction step and that the #inetics are first order -ith respect to the surface
concentration of this adsorbed inter!ediate
6ubstituting for the coverage, , gives us the re'uired expression for the rate in terms of
the pressure of gas above the surface
.ate > k b P ? 1 A b P !
It is useful to consider two extreme limits +
Cimit I1 + b.P 99 1 J then
1 A b.P ! K 1 and .ate K k.b.P
i.e. a first order reaction (with respect to the partial pressure of ! with an apparent
first order rate constant , #$ 3 #0b .
"his is the low pressure (or wea$ binding i.e. small b limit + under these
conditions the steady state surface coverage , , of the reactant molecule is very
small.

Cimit I% + b.P 66 1 J then
1 A b.P ! K b.P and .ate K k
i.e. a (ero order reaction (with respect to the partial pressure of ! .
"his is the high pressure (or strong binding i.e. large b limit + under these
conditions the steady state surface coverage , , of the reactant molecule is almost
unity.
In fact, the rate shows the same pressure variation as does the surface coverage (hardly
surprising since it is directly proportional to
4xample % 2 5imolecular .eaction (between molecular adsorbates
Consider a ?angmuir&>inshelwood reaction of the following type +
A
(g
3 A
(ads
5
(g
3 5
(ads
A
(ads
A 5
(ads
A5
(ads
A5
(g
.e will further assume, as noted in the above scheme, that the surface reaction between the
two adsorbed species is the rate determining step.
If the two adsorbed molecules are mobile on the surface and freely intermix then the rate of
the reaction will be given by the following rate expression for the bimolecular surface
combination step
.ate > k
!

0
,or a single molecular adsorbate the surface coverage (as given by the ?angmuir isotherm
is +
> b.P ? 1 A b.P !
.here two molecules ( ! C 0 are competing for the same adsorption sites then the
relevant expressions are (see derivation +
6ubstituting these into the rate expression gives +
=nce again, it is interesting to loo$ at several extreme limits
Cimit I1 + b
!
P
!
99 1 C b
0
P
0
99 1
In this limit
!
C
0
are both very low , and
.ate k b
!
P
!
b
0
P
0
> k' P
!
P
0
i.e. first order in both reactants

Cimit I% + b
!
P
!
99 1 99 b
0
P
0
In this limit
!
9 ,
0
# , and
.ate k b
!
P
!
? b
0
P
0
! > k' P
!
? P
0
i.e. first order in ! , but negative first order in 0
,rom this example it is readily seen that, depending upon the partial pressure and binding
strength of the reactants, a given model for the reaction scheme can give rise to a variety of
apparent $inetics + this highlights the dangers inherent in the reverse process & namely
trying to use $inetic data to obtain information about the reaction mechanism.
4xample & 2 C< <xidation .eaction
=n precious metal surfaces (e.g. 8t the C= oxidation reaction is generally believed to by a
?angmuir&>inshelwood mechansim of the following type +
C<
(g
3 C<
(ads
<
% (g
3 % <
(ads
C<
(ads
A <
(ads
C<
% (ads
C<
% (g
!s C=
%
is comparatively wea$ly&bound to the surface, the desorption of this product
molecule is relatively fast and in many circumstances it is the surface reaction between the
two adsorbed species that is the rate determining step.
If the two adsorbed molecules are assumed to be mobile on the surface and freely intermix
then the rate of the reaction will be given by the following rate expression for the
bimolecular surface combination step
.ate > k
C=

=
.here two such species (one of which is molecularly adsorbed, and the other dissociatively
adsorbed are competing for the same adsorption sites then the relevant expressions are
(see derivation +
6ubstituting these into the rate expression gives +
=nce again, it is interesting to loo$ at certain limits. If the C= is much more strongly
bound to the surface such that b
C<
P
C<
66 1 A b
<
P
<%
!

! and hence that 1 A b
<
P
<%
! A
b
C<
P
C<
! b
C<
P
C<
then the rate e'uation simplifies to give
In this limit the $inetics are half&order with respect to the gas phase pressure of molecular
oxygen, but negative order with respect to the C= partial pressure, i.e. C= acts as a poison
(despite being a reactant and increasing its pressure slows down the reaction. "his is
because the C= is so strongly bound to the surface that it bloc$s oxygen adsorbing, and
without sufficient oxygen atoms on the surface the rate of reaction is reduced.
0. -,+ 8 4ffects of @as *ressure
0.1 "hat is -ltra ,igh +acuum -,+! $
Sacuum technology has advanced considerably over the last %: years and very low
pressures are now routinely obtainable.
,irstly, we need to remind ourselves of the units of pressure +
"he 6I unit of pressure is the 8ascal ( # 8a 3 # N m
&%

Normal atmospheric pressure ( # atm. is #9#/%: 8a or #9#/ mbar ( # bar 3 #9
:
8a .
!n older unit of pressure is the "orr ( # "orr 3 # mm>g . =ne atmosphere is ca.
H@9 "orr ( i.e. # "orr 3 #//./ 8a .
.hilst the mbar is often used as a unit of pressure for describing the level of vacuum, the
most commonly employed unit is still the "orr. ("he 6I unit, the 8a , is almost never
used 4
Classification of the degree of vacuum is hardly an exact science & it very much depends
upon who you are tal$ing to & but as a rough guideline +
Rough (low vacuum + # & #9
&/
"orr
*edium vacuum + #9
&/
& #9
&:
"orr
>igh vacuum (>S + #9
&@
& #9
&D
"orr
Eltrahigh vacuum (E>S + J #9
&G
"orr
Sirtually all surface studies are carried out under E>S conditions & the 'uestion is #hy 5
"his is the 'uestion that we will address in 6ection 1.% .
0.% "hy is -,+ reDuired for surface studies $
Eltra high vacuum is re'uired for most surface science experiments for two principal
reasons +
#. "o enable atomically clean surfaces to be prepared for study, and such surfaces to
be maintained in a contamination&free state for the duration of the experiment.
%. "o permit the use of low energy electron and ion&based experimental techni'ues
without undue interference from gas phase scattering.
"o put these points in context we shall now loo$ at the variation of various parameters with
pressure

1. @as Bensity
"he gas density is easily estimated from the ideal gas law +
n 3 ( ) I 3 3 I (#0T L molecules m
&/
M
where + & pressure L N m
&%
M
# & 0olt(mann constant ( 3 #./D x #9
&%/
Q B
&#


%. Mean 1ree *ath of *articles in the @as *hase
"he average distance that a particle (atom, electron, molecule .. travels in the gas phase
between collisions can be determined from a simple hard&sphere collision model (see, for
example, !t$insW 8hysical Chemistry & this 'uantity is $nown as the !ean free path of the
particle, here denoted by , and for neutral molecules is given by the e'uation +

LmM
where + & pressure L N m
&%
M
&%/
Q B
&#

& collision cross section L m
%
M

&. Incident Molecular 1lux on Surfaces
=ne of the crucial factors in determining how long a surface can be maintained clean (or,
alternatively, how long it ta$es to build&up a certain surface concentration of adsorbed
species is the number of gas molecules impacting on the surface from the gas phase.
"he incident flux is the number of incident molecules per unit time per unit area of surface.
(Note & the flux ta$es no account of the angle of incidence, it is merely a summation of all
the arriving molecules over all possible incident angles
,or a given set of conditions (.T etc. the flux is readily calculated using a combination of
the ideas of statistical physics, the ideal gas e'uation and the *axwell&0olt(mann gas
velocity distribution.
Step 4 + it can be readily shown that the incident flux, ( , is related to the gas density above
the surface by

L molecules m
&%
s
&#
M
where + n & molecular gas density L molecules m
&/
M
c & average molecular speed L m s
&#
M
Step 1 + the molecular gas density is given by the ideal gas e'uation, namely
n 3 ) I 3 3 I (#T L molecules m
&/
M
Step 6 + the mean molecular speed is obtained from the *axwell&0olt(mann distribution of
gas velocities by integration, yielding

L m s
&#
M
where + ! & molecular mass L $g M
&%/
Q B
&#

3 /.#1#@
6tep 1 + combining the e'uations from the first three steps gives the >ert(&Bnudsen
formula for the incident flux

L molecules m
&%
s
&#
M
Note
#. all 'uantities in the above e'uation need to be expressed in 6I units
%. the molecular flux is directly proportional to the pressure

0. @as 4xposure - the LCangmuirL
"he gas exposure is measure of the amount of gas which a surface has been sub<ected to. It
is numerically 'uantified by ta$ing the product of the pressure of the gas above the surface
and the time of exposure (if the pressure is constant, or more generally by calculating the
integral of pressure over the period of time of concern.
!lthough the exposure may be given in the 6I units of 8a s (8ascal seconds, the normal
and far more convenient unit for exposure is the 8ang!uir, where # ? 3 #9
&@
"orr s . i.e.
(2xposureI? 3 #9
@
(8ressureI"orr ("imeIs

3. Sticking Coefficient 8 Surface Co'erage
"he stic#ing coefficient, S , is a measure of the fraction of incident molecules which adsorb
upon the surface i.e. it is a probability and lies in the range 9 & # , where the limits
correspond to no adsorption and complete adsorption of all incident molecules respectively.
In general, S depends upon many variables i.e.
S 3 f ( surface coverage , temperature, crystal face ....
"he surface coverage of an adsorbed species may itself, however, be specified in a number
of ways +
#. as the number of adsorbed species per unit area of surface (e.g. in molecules cm
&%
.

%. as a fraction of the maximum attainable surface coverage i.e.

& in which case lies in the range 9 & # .

/. relative to the atom density in the topmost atomic layer of the substrate i.e.

& in which case
max
is usually less than one, but can for an adsorbate such as >
occasionally exceed one.
Note +
#. whichever definition is used, the surface coverage is normally denoted by the Rree$
.
%. the second means of specifying the surface coverage is only usually employed for
adsorption isotherms (e.g. the ?angmuir isotherm. "he third method is the most
generally accepted way of defining the surface coverage.
/. a !onolayer (# *? of adsorbate is ta$en to correspond to the maximum attainable
surface concentration of adsorbed species bound to the substrate.
.e can also as$,
>ow long will it ta$e for a clean surface to become covered with a complete monolayer of
adsorbate 5
"his is dependent upon the flux of gas phase molecules incident upon the surface, the
actual coverage corresponding to the monolayer and the coverage&dependent stic$ing
probability ... however , it is possible to get a minimum estimate of the time re'uired by
assuming a unit stic$ing probability (i.e. S 3 # and noting that monolayer coverages are
generally of the order of #9
#:
per cm
%
or #9
#G
per m
%
. "hen
Time ? MC K 1(
#G
? F ! L s M

Summary - +ariation of *arameters #ith *ressure
!ll values given below are approximate and are generally dependent on factors such as
temperature and molecular mass.
Degree of Sacuum
8ressure
("orr
Ras Density
(molecules m
&/
*ean ,ree 8ath

(m
"ime I *?
(s

!tmospheric H@9 % x #9
%:
H x #9
&D
#9
&G
?ow # / x #9
%%
: x #9
&:
#9
&@
*edium #9
&/
/ x #9
#G
: x #9
&%
#9
&/
>igh #9
&@
/ x #9
#@
:9 #
Eltra>igh #9
&#9
/ x #9
#%
: x #9
:
#9
1
.e can therefore conclude that the following re'uirements exist for +
Collision ,ree Conditions 3U 8 J #9
&1
"orr
*aintenance of a Clean 6urface 3U 8 J #9
&G
"orr
Summary
,or most surface science experiments there are a number of factors necessitating a high
vacuum environment +
#. ,or surface spectroscopy, the mean free path of probe and detected particles (ions,
atoms, electrons in the vacuum environment must be significantly greater than the
dimensions of the apparatus in order that these particles may travel to the surface
and from the surface to detector without undergoing any interaction with residual
gas phase molecules. "his re'uires pressures better than #9
&1
"orr. "here are,
however, some techni'ues, such as IR spectroscopy, which are -photon&inIphoton&
out- techni'ues and do not suffer from this re'uirement.
(=n a practical level, it is also the case that the lifetime of channeltron and
multiplier detectors used to detect charged particles is substantially reduced by
operation at pressures above #9
&@
"orr.

%. *ost spectroscopic techni'ues are also capable of detecting molecules in the gas
phaseA in these cases it is preferable that the number of species present on the
surface substantially exceeds those present in the gas phase immediately above the
surface & to achieve a surfaceIgas phase discrimination of better than #9+# when
analysing ca. #K of a monolayer on a flat surface this re'uires that the gas phase
concentration is less than ca. #9
#%
molecules cm
&/
( 3 #9
#D
molecules m
&/
, i.e. that
the (partial pressure is of the order of #9
&1
"orr or lower.

/. In order to begin experiments with a reproducibly clean surface, and to ensure that
significant contamination by bac$ground gases does not occur during an
experiment, the bac$ground pressure must be such that the time re'uired for
contaminant build&up is substantially greater than that re'uired to conduct the
experiment i.e. of the order of hours. "he implication with regard to the re'uired
pressure depends upon the nature of the surface, but for the more reactive surfaces
this necessitates the use of E>S (i.e. J # x #9
&G
"orr.
It is clear therefore that it is the last factor that usually determines the need for a very good
vacuum in order to carry out reliable surface science experiments.
0.& 4xercises - "he 8ressure Dependence of Ras 8hase C
6urface !dsorption Characteristics
"his section provides a limited number of examples of the application of the formulae
given in the previous section to determine the +
Density of *olecules in the Ras 8hase
*ean ,ree 8ath of *olecules in the Ras 8hase
,lux of *olecules incident upon a 6urface
Rate of !dsorption of *olecules and 6urface Coverages
If you have not already been through 6ection 1.% then I would suggest that you stop now
and return to this page only after you have done so M
:ithin any one of the follo-ing sub-sections. it -ill be assu!ed that you have already done
the previous &uestions and !ay !a#e use of the ans-ers fro! these &uestions - you are
therefore advised to -or# through the &uestions in the order they are presented0
A. Molecular @as Bensities
.hat is the molecular gas density for an ideal gas at /99 B when the pressure is #9
&@
"orr 5
(in molecules m
&/

.hat therefore would be the gas density at a pressure of #9
&G
"orr 5
5. Mean 1ree *ath of Molecules in the @as *hase
Calculate the mean free path of C= molecules in a vessel at a pressure of #9
&1
"orr at /99 B.
(Collision cross section 3 9.1% nm
%
.
Now calculate the mean free path of C= molecules, again at /99 B, at a pressure of #9
&G
"orr .
C. 1luxes of Molecules Incident upon a Surface
Calculate the flux of molecules incident upon a solid surface under the following
conditions +
LNote & # u 3 #.@@ x #9
&%H
$g + atomic masses A !(= 3#@ u, !(> 3 # uM
=xygen gas ( 8 3 # "orr at /99 B
=xygen gas ( 8 3 #9
&@
"orr at /99 B
>ydrogen gas ( 8 3 #9
&@
"orr at /99 B
>ydrogen gas ( 8 3 #9
&@
"orr at #999 B
B. The =inetically Cimited -ptake of Molecules onto a Surface
"he rate of adsorption of molecules onto a surface can be determined from the flux of
molecules incident on the surface and the stic$ing probability pertaining at that instant in
time (note that in general the stic$ing probability itself will be dependent upon a number of
factors including the existing coverage of adsorbed species.
In the following examples we will assume that the surface is initially clean (i.e. the initial
coverage is (ero, and that there is no desorption of the molecules once they have adsorbed.
You should determine coverages as the ratio of the adsorbate concentration to the density
of surface substrate atoms (which you may assume to be #9
#G
m
&%
. In the first two
'uestions we will assume that the stic$ing probability is constant over the coverage range
concerned.
Calculate the surface coverage obtained after exposure to a pressure of #9
&D
"orr of C= for %9 s at /99
B & you may ta$e the stic$ing probability of C= on this surface to have a constant value of 9.G up to the
coverage concerned.
Calculate the surface coverage of atomic nitrogen obtained by dissociative adsorption after exposure to
a pressure of #9
&D
"orr of nitrogen gas for %9 s at /99 B & you may ta$e the dissociative stic$ing
probability of molecular nitrogen on this surface to be constant and e'ual to 9.#
In general, the stic$ing probability varies with coverage & most obviously, the stic$ing
probability must tend to (ero as the coverage approaches its saturation value. "hese
calculations are not 'uite so easy M
Calculate the surface coverage obtained after exposure to a pressure of #9
&D
"orr of C= for %99 s at /99
B & the stic$ing probability of C= in this case should be ta$en to vary linearly with coverage between a
value of unity at (ero coverage and a value of (ero at saturation coverage (which you should ta$e to be
@.: x #9
#D
molecules m
&%
.
3. Surface Analytical TechniDues
3.1. Surface Sensiti'ity 8 Surface Specificity
1. @eneral Sensiti'ity *roblems
"he problems of sensitivity and detection limits are common to all forms of spectroscopy A
some techni'ues are simply better than others in this respect 4 In its simplest form, the
'uestion of sensitivity boils down to whether it is possible to detect the desired signal
above the noise level.
In virtually all surface studies (especially those on single crystal substrates sensitivity is a
ma<or problem. Consider the case of a sample with a surface of si(e # cm
%
& this will have
ca. #9
#:
atoms in the surface layer. In order to detect the presence of impurity atoms present
at the # K level, a techni'ue must be sensitive to ca. #9
#/
atoms.
Contrast this with a spectroscopic techni'ue used to analyse a # cm
/
bul$ li'uid sample i.e.
a sample of ca. #9
%%
molecules. "he detection of #9
#/
molecules in this sample would
re'uire # ppb (one part-per-billion sensitivity & very few techni'ues can provide anything
li$e this level of sensitivity 4
"his is one reason why common spectroscopic techni'ues such as N*R (detection limit ca.
#9
#G
molecules cannot be used for surface studies, except on samples with very high
surface areas.
%. Surface Sensiti'ity 8 Specificity
!ssuming that a techni'ue of sufficient sensitivity can be found, another ma<or problem
that needs to be addressed in surface spectroscopy is distinguishing between signals from
the surface and the bul$ of the sample. In principle, there are two ways around this problem
+
#. "o ensure that the surface signal is distinguishable (shifted from the comparable
bul$ signal, and that the detection system has sufficient dynamic range to detect
very small signals in the presence of neighbouring large signals.
%. "o ensure that the bul$ signal is small compared to the surface signal i.e. that the
vast ma<ority of detected signal comes from the surface region of the sample.
It is the latter approach which is used by the ma<ority of surface spectroscopic techni'ues &
such techni'ues can then be said to be surface sensitive. "o illustrate this, we shall loo$ at
one way in which surface sensitivity can be achieved & this ma$es use of the special
properties of
..... ?ow 2nergy 2lectrons
& it is an approach employed in common surface spectroscopic techni'ues such as
!uger 2lectron 6pectroscopy (!26 (see 6ection :.%
T&ray 8hotoelectron 6pectroscopy (T86 (see 6ection :./
.e shall illustrate this approach by considering the latter techni'ue (T86 but you should
remember that the ideas that we develop are more widely applicable.
N*S - A surface sensiti'e techniDue $
In answering this important 'uestion, we will actually address the topics listed below & you
are strongly advised to wor$ through them in the order in which they are given.
Contents + .hat do we really mean by surface sensitivity I specificity 5
>ow can we demonstrate that T86 is surface sensitive 5
.hy is the T86 techni'ue surface sensitive 5
>ow can we use $nowledge of the I*,8 to calculate the thic$ness of surface films 5
>ow can we change the degree of surface sensitivity 5
"hat do #e really mean by surface sensiti'ity ? specificity $
*ost analytical techni'ues used in chemistry are -bul$- techni'ues in the sense that they
measure all the atoms within a typical sample (be it a solid, li'uid, solution or gas phase
sample & so if, for example, we were to analyse a sample consisting of a #9 nm thic$ layer
of one material (! deposited on a # mm thic$ substrate of another material (0 ....
7 and to as$ the 'uestion 7.
:hat -ould be the concentration of A in parts per !illion %pp!' -hich you -ould expect
to be revealed by a bul# analysis of such a sa!ple ;
.... we would get the same result (#9 ppm as if we were loo$ing at a sample in which
component ! was uniformly distributed throughout the bul$ of 0 such as to give the same
overall concentration in the sample as a whole.
0y contrast, a surface sensitive techni'ue is more sensitive to those atoms which are
located near the surface than it is to atoms in the bul$ which are well away from the surface
(i.e. the main part of the signal comes from the surface region & in the case of the first
sample therefore a surface sensitive techni'ue would produce a spectrum with enhanced
signals due to component ! , i.e.
I
!
? I
0
66 1(
&:
(where I
!
is the signal due to component ! etc.
! truly surface specific techni'ue should only give signals due to atoms in the surface
region & but that, of course, rather depends how the -surface region- is defined 4
2lectron spectroscopic techni'ues such as T86 are not completely surface specific
(although you will occasionally find this expression being used in that whilst most of the
signal comes from within a few atomic layers of the surface, a small part of the signal
comes from much deeper into the solid & they are best described as being surface sensitive
techni'ues.
>owever, we should also note that there are a few techni'ues (e.g. low energy ion
scattering LI66I?2I6M which are indeed virtually specific to the topmost layer of atoms.
,o# can #e demonstrate that N*S is surface sensiti'e $
Consider again the first of the two samples considered in the previous section & such
samples are often readily prepared by depositing material ! by evaporation onto a substrate
of material 0.
It is experimentally observed that in such a situation the T86 signals due to the substrate
(material 0 are rapidly attenuated, whilst those due to the condensed evaporant (material
! simultaneously increase to a limiting value.
Note + it is possible to measure deposition rates using e'uipment such as 'uart( crystal
oscillators and thus obtain independent verification of the deposited film thic$ness. !
detailed analysis of the variation of the signals due to ! and 0 is also capable of yielding
information about the mechanism of film growth.
"hy is the N*S techniDue surface sensiti'e $
"he soft T&rays employed in T86 penetrate a substantial distance into the sample ()m.
"hus this method of excitation imparts no surface sensitivity at the re'uired atomic scale.
"he surface sensitivity must therefore arise from the emission and detection of the
photoemitted electrons.
Consider, therefore, a hypothetical experiment in which electrons of a given energy, E
9
, are
emitted from atoms in a solid at various depths, d , below a flat surface.
.e shall assume that only those electrons which reach the surface and, at the time of
leaving the solid, still have their initial energy (E
o
are detected in this experiment. "his is
not literally true in an T86 experiment, but in T86 experiments it is only those
photoelectrons possessing characteristic emission energies and contributing to the pea$s
and that we consider in the subse'uent analysisA we do not consider photoelectrons which
have lost some energy and contribute to the bac$ground of the spectrum rather than a
specific pea$.
"here are two things which would therefore prevent an emitted electron from being
detected +
#. If it were -captured- before reaching the surface, or emitted in the wrong direction
and never reached the surface 4
%. If it lost energy before reaching the surface, i.e. if it escaped from the solid and
reached the detector, but with E J E
o

"he process by which an electron can lose energy as it travels through the solid is $nown as
-inelastic scattering-. 2ach inelastic scattering event leads to + (i a reduction in the electron
energy (ii a change in the direction of travel.
In the diagram below arrows are used to represent possible tra<ectories of emitted electrons
from a particular source atom. "he orange arrows represent electrons travelling with their
initial energy, E
o
, whilst the blue arrows represent electron tra<ectories subse'uent to an
inelastic scattering event. =nly the orange traces which extend out of the solid therefore
correspond to electrons which would be detected in our hypothetical experiment.
In this case, therefore, only a small fraction of the emitted electrons would have been
detected.
Clearly if the source atom had been closer to the surface, then a greater fraction of the
emitted electrons would have been detected since +
#. "he detector would subtend a greater (solid angle at the emitting atom (this is
actually not a significant factor in real experiments.
%. ,or an electron emitted towards the surface, there would be less chance of inelastic
scattering before it escaped from the solid.
So ho# does the probability of detection depend upon the distance of the emitting
atom belo# the surface $
?et us first simplify the problem by only considering those electrons emitted directly
towards (normal to the surface & in our hypothetical experiment we can represent this by a
smaller detector located directly above the emitting atoms.
"he probability of escape from a given depth, 8(d, is determined by the li$elihood of the
electron not being inelastically scattered during its <ourney to the surface.
Clearly, *d
%
! 9 *d
#
! but, in order to 'uantify this we need to $now more about the
inelastic scattering process & this brings us to the sub<ect of the ..
Inelastic Mean 1ree *ath IM1*! of electrons.
"he I*,8 is a measure of the average distance travelled by an electron through a solid
before it is inelastically scatteredA it is dependent upon
"he initial $inetic energy of the electron.
"he nature of the solid ( but most elements show very similar I*,8 vWs energy
relationships.
"he I*,8 is actually defined by the following e'uation which gives the probability of the
electron travelling a distance, d , through the solid without undergoing scattering
*d! > exp - d ? !
where is the I*,8 for the electrons of energy E .
(Note + 3 f(E, and this inelastic mean free path, which relates to the movement of
electrons in the solid, is completely unrelated to the mean free path in the gas phase once
they escape from the solid.
"he graph below illustrates this functional form of 8(d 7
7 and it is clear that the probability of escape decays very rapidly and is essentially (ero
for a distance d U : . It is therefore useful to re&plot the function for a range of (dI from
9&: .
.hat is the probablility of escape when 7 d 3 # 5
.hat is the probablility of escape when 7 d 3 % 5
.hat is the probablility of escape when 7 d 3 / 5
It is also interesting to loo$ at this same problem from a different point of view. ?et us
assume that there are many sources of electrons uniformly distributed at all distances from
the surface of the solid (which is in effect the situation pertaining in all T86 experiments
and again arrange to only detect those unscattered electrons which emerge normal to the
surface & then we can as$ ...
:hat is the distribution of the depths of origin of those electrons that -e do detect ;
"his new function, letWs call it 8W(d, actually has the same exponential form as 8(d & i.e. the
fre'uency of detection of electrons from different depths in the solid is directly
proportional to the probability of electron escape from each depth.
6o if we allow our detector to collect #99 electrons, we can as$ ...
>ow many electrons (on average will have come from within a distance of one I*,8 from the surface
5
(i.e. how many originate from a distance, d , such that 9 J (dI J # 5 .
"his is a relatively simple problem in integration & try it 4
In summary, we can conclude from the answer to this 'uestion that ...
... the maOority of electrons detected come from #ithin one LimfpL of the surface.
Sirtually all ( U G:K of the electrons detected come from within / I*,8s of the surface.
"arning & beware of terms such as - sa!pling depth - and - electron escape depth -. It is
not always clear whether these refer to #, / or : I*,8s (or something else completely.
So ho# big is the IM1* of typical electrons #ithin a solid such as a metal $
"he graph below shows the so&called -universal curve- for the variation of the I*,8 with
initial electron energy (in eS .
It is reasonably accurate for most metals, although experimental measurements on different
metals do show a substantial scatter about the general curve.
=ther classes of solids also exhibit 'ualitatively similar universal curves, although the
absolute values of the I*,8 of electrons in materials such as inorganic solids and polymers
do differ 'uite significantly from those in metals.
"his I*,8 data is more usually displayed on a log-log plot, since this more clearly shows
how the I*,8 varies at low $inetic energies & the graph below shows the general curve for
metals re&plotted in this new format.
"he I*,8 exhibits a minimum for electrons with a $inetic energy of around :9 &#99 eS A at
lower energies the probability of inelastic scattering decreases since the electron has
insufficient energy to cause plasmon excitation (the main scattering mechanism in metals,
and conse'uently the distance between inelastic collisions and the I*,8 increases.
In summary ,
"he Inelastic *ean ,ree 8ath (I*,8 in metals is typically less than +
#9 !ngstroms ( # nm for electron energies in the range #: J EIeS J /:9
%9 !ngstroms ( % nm for electron energies in the range #9 J EIeS J #199
i.e. the I*,8 of low energy electrons corresponds to only a few atomic layers.
"his means that any experimental techni'ue such as T86 which involves the generation
and detection of electrons of such energies will be surface sensitive.
,o# can #e use kno#ledge of the IM1* to calculate the thickness of surface films $
.e can now consider a situation where a substrate (or thic$ film of one material, 0 , is
covered by a thin film of a different material, ! . "he T86 signal from the underlying
substrate will be attenuated (i.e. reduced in intensity due to inelastic scattering of some of
the photoelectrons as they traverse through the layer of material ! .
"he probability of such a scattering event for any single photoelectron passing through this
layer is simply given by +
* > exp - t ? !
where + t is the thic$ness of the layer of material ! .
It follows that the overall intensity of a T86 signal arising from 0 is reduced by this same
factor, i.e. if the intensity of this signal in the absence of any covering layer is I
o
then the
intensity I in the presence of the overlayer is given by +
I > I
o
exp - t ? !
It also follows that it is possible to estimate the thic$ness of a deposited layer (using the
above e'uation, provided the reduction in the substrate signal is $nown (i.e. if spectra are
ac'uired before and after deposition of the covering film.
,o# can #e change the degree of surface sensiti'ity $
,or a given techni'ue such as T86 which relies on inelastic electron scattering to provide
the surface sensitivity, we can consider how it might be possible to increase the surface
sensitivity when analysing a particular substrate.
=ne approach would be to try and ensure that the electrons being detected have $inetic
energies corresponding to the minimum in the universal curve for the I*,8. "he problem
with this approach is that
#. it re'uires us to be able to vary the T&ray radiation so that we can ad<ust the photon
energy, h , to give us the desired $inetic energy for the photoemitted electrons A
this is not generally possible with conventional laboratory x&ray sources (although
tunable x&ray radiation is available at large synchrotron sources
%. whilst it may be possible to ad<ust the $inetic energy of some particular
photoelectrons to correspond to the minimum in the universal curve, this effect may
not hold for all the other photoelectron pea$s associated with the sample.
! much more generally useful approach is to collect photoelectrons at a more gra(ing
emission angle.
"he $ey point to recognise here is that it is actually the distance that the electrons have to
travel through the solid (rather than the distance of the emitting atom below the surface
that is crucial in determining their probability of escape without inelastic scattering & only
in the special case considered previously of normal emission are these two 'uantities the
same.
Consider, for example, emission at some angle to the surface normal.
=nce again @/K of the electrons will originate from atoms located such that the electrons
have to travel a distance, d , which is less than # through the solid (see above. Ender
these measurement conditions, however, all these atoms are located within a smaller
distance, x , from the surface where
x > d cos
6o as the emission angle ( is increased, cos decreases, the analysed region becomes
more surface localised and the surface sensitivity is increased.
=ne example of how this effect can be put to good use is in the study (and diagnosis of
surface segregation. "he graph below shows simulated data for the variation of the T86
signal intensities with emission angle for
a random alloy of #9K of element ! in element 0, with no segregation (lower blue
curve
the same random alloy, but with surface segregation of ! to give a surface
monolayer of pure element ! (upper red curve
3.%. Auger 4lectron Spectroscopy
!uger 2lectron 6pectroscopy (Auger spectroscopy or !26 was developed in the late
#G@9Ws , deriving its name from the effect first observed by 8ierre !uger, a ,rench
8hysicist, in the mid&#G%9Ws. It is a surface specific techni'ue utilising the emission of low
energy electrons in the Auger process and is one of the most commonly employed surface
analytical techni'ues for determining the composition of the surface layers of a sample.
*,/SICAC 5ASIS
!uger spectroscopy can be considered as involving three basic steps +
(# !tomic ioni(ation (by removal of a core electron
(% 2lectron emission (the !uger process
(/ !nalysis of the emitted !uger electrons
"he last stage is simply a technical problem of detecting charged particles with high
sensitivity, with the additional re'uirement that the $inetic energies of the emitted electrons
must be determined.
.e will therefore concern ourselves only with the first two processes & before starting,
however, it is useful to briefly review the electronic structure of atoms and solids, and
associated nomenclature.
4lectronic Structure - Isolated Atoms.
"he diagram below schematically illustrates the energies of the various electron energy
levels in an isolated, multi&electron atom, with the conventional chemical nomenclature for
these orbitals given on the right hand side.
>owever, scientists wor$ing with x&rays tend to use the alternative nomenclature on the left
and it is this nomenclature that is used in !uger spectroscopy. "he designation of levels to
the B,?,*,7 shells is based on their having principal 'uantum numbers of #,%,/,7
respectively.
It is convenient to expand the part of the energy scale close to the vacuum level in order to
more clearly distinguish between the higher levels ....
"he numerical component of the B?*.. style of nomenclature is usually written as a
subscript immediately following the main shell designation.
?evels with a non&(ero value of the orbital angular momentum 'uantum number ( l U 9,
i.e. p,d,f,.. levels, show spin&orbit splitting. "he magnitude of this splitting, however, is too
small to be evident on this diagram & hence, the double subscript for these levels (i.e. ?
%,/
represents both the ?
%
and ?
/
levels.
4lectronic Structure - Solid State
In the solid state the core levels of atoms are little perturbed and essentially remain as
discrete, localised (i.e. atomic&li$e levels. "he valence orbitals, however, overlap
significantly with those of neighbouring atoms generating bands of spatially&delocalised
energy levels.
"he energy level diagram for the solid is therefore closely resemblant of that of the
corresponding isolated atom, except for the levels closest to the vacuum level.
"he diagram below shows the electronic structure of Na metal +
The Auger *rocess 8 Auger Spectroscopy
Now let us return to the sub<ect of !uger spectroscopy & in the following discussion, the
!uger process is illustrated using the B, ?
#
C ?
%,/
levels. "hese could be the inner core
levels of an atom in either a molecular or solid&state environment.
I. IoniPation
"he !uger process is initiated by creation of a core hole & this is typically carried out by
exposing the sample to a beam of high energy electrons (typically having a primary energy
in the range % & #9 $eS. 6uch electrons have sufficient energy to ionise all levels of the
lighter elements, and higher core levels of the heavier elements.
In the diagram above, ionisation is shown to occur by removal of a B&shell electron, but in
practice such a crude method of ionisation will lead to ions with holes in a variety of inner
shell levels.
In some studies, the initial ionisation process is instead carried out using soft x&rays ( h 3
#999 & %999 eS . In this case, the acronym T!26 is sometimes used. !s we shall see,
however, this change in the method of ionisation has no significant effect on the final
!uger spectrum.
II. .elaxation 8 Auger 4mission
"he ioni(ed atom that remains after the removal of the core hole electron is, of course, in a
highly excited state and will rapidly relax bac$ to a lower energy state by one of two routes
+
#. N-ray fluorescence , or
%. Auger emission
.e will only consider the latter mechanism, an example of which is illustrated
schematically below ....
In this example, one electron falls from a higher level to fill an initial core hole in the B&
shell and the energy liberated in this process is simultaneously transferred to a second
electron A a fraction of this energy is re'uired to overcome the binding energy of this
second electron, the remainder is retained by this emitted Auger electron as $inetic energy.
In the !uger process illustrated, the final state is a doubly&ioni(ed atom with core holes in
the ?
#
and ?
%,/
shells.
.e can ma$e a rough estimate of the B2 of the !uger electron from the binding energies
of the various levels involved. In this particular example,
=4 > 4
B
- 4
?#
! - 4
?%/

L .hy is this answer not li$ely to be very accurate 5 M
)ote 5 the 9E of the Auger electron is independent of the !echanis! of initial core hole
for!ation0
"he expression for the energy can also be re&written in the form +
=4 > 4
B
- 4
?#
A 4
?%/
!
It should be clear from this expression that the latter two energy terms could be
interchanged without any effect & i.e. it is actually impossible to say which electron fills the
initial core hole and which is e<ected as an !uger electron A they are indistinguishable.
!n !uger transition is therefore characteri(ed primarily by +&
#. the location of the initial hole
%. the location of the final two holes
although the existence of different electronic states (terms of the final doubly&ioni(ed atom
may lead to fine structure in high resolution spectra.
.hen describing the transition, the initial hole location is given first, followed by the
locations of the final two holes in order of decreasing binding energy.
i.e. the transition illustrated is a B?
#
?
%,/
transition .
If we <ust consider these three electronic levels there are clearly several possible !uger
transitions + specifically,
B ?
#
?
#
B ?
#
?
%,/
B ?
%,/
?
%,/
In general, since the initial ionisation is non&selective and the initial hole may therefore be
in various shells, there will be many possible !uger transitions for a given element & some
wea$, some strong in intensity. !ER2R 682C"R=6C=8Y is based upon the
measurement of the $inetic energies of the emitted electrons. 2ach element in a sample
being studied will give rise to a characteristic spectrum of pea$s at various $inetic energies.
"his is an !uger spectrum of 8d metal & generated using a %.: $eS electron beam to
produce the initial core vacancies and hence to stimulate the !uger emission process. "he
main pea$s for palladium occur between %%9 C /19 eS. "he pea$s are situated on a high
bac$ground which arises from the vast number of so&called secondary electrons generated
by a multitude of inelastic scattering processes.
!uger spectra are also often shown in a differentiated form + the reasons for this are partly
historical, partly because it is possible to actually measure spectra directly in this form and
by doing so get a better sensitivity for detection. "he plot below shows the same spectrum
in such a differentiated form.
Summary
!uger 2lectron 6pectroscopy (!26 is a surface&sensitive spectroscopic techni'ue used for
elemental analysis of surfaces A it offers
high sensitivity (typically ca. #K monolayer for all elements except > and >e.
a means of monitoring surface cleanliness of samples
'uantitative compositional analysis of the surface region of specimens, by
comparison with standard samples of $nown composition.
In addition, the basic techni'ue has also been adapted for use in +
#. !uger Depth 8rofiling + providing 'uantitative compositional information as a
function of depth below the surface (see 6ection H.:
%. 6canning !uger *icroscopy (6!* + providing spatially&resolved compositional
information on heterogeneous samples (see 6ection H.%
3.&. *hotoelectron Spectroscopy
8hotoelectron spectroscopy utili(es photo&ioni(ation and energy&dispersive analysis of the
emitted photoelectrons to study the composition and electronic state of the surface region
of a sample.
"raditionally, when the techni'ue has been used for surface studies it has been subdivided
according to the source of exciting radiation into +
N-ray *hotoelectron Spectroscopy
N*S!
& using soft x&ray (%99&%999 eS radiation to examine
core&levels.
-ltra'iolet *hotoelectron Spectroscopy
-*S!
& using vacuum ES (#9&1: eS radiation to examine
valence levels.
"he development of synchrotron radiation sources has enabled high resolution studies to be
carried out with radiation spanning a much wider and more complete energy range ( : &
:999V eS but such wor$ is, and will remain, a very small minority of all photoelectron
studies due to the expense, complexity and limited availability of such sources.
*hysical 5asis
8hotoelectron spectroscopy is based upon a single photon inIelectron out process and from
many viewpoints this underlying process is a much simpler phenomenon than the !uger
process.
"he energy of a photon is given by the 2instein relation +
E > h
where h & 8lanc$ constant ( @.@% x #9
&/1
Q s
& fre'uency (>( of the radiation
8hotoelectron spectroscopy uses monochromatic sources of radiation (i.e. photons of fixed
energy.
In T86 the photon is absorbed by an atom in a molecule or solid, leading to ioni(ation and
the emission of a core (inner shell electron. 0y contrast, in E86 the photon interacts with
valence levels of the molecule or solid, leading to ionisation by removal of one of these
valence electrons.
"he $inetic energy distribution of the emitted photoelectrons (i.e. the number of emitted
photoelectrons as a function of their $inetic energy can be measured using any appropriate
electron energy analyser and a photoelectron spectrum can thus be recorded.
"he process of photoioni(ation can be considered in several ways + one way is to loo$ at
the overall process as follows +
A A h A
A
A e-
Conservation of energy then re'uires that +
4A! A h > 4A
A
! A 4e-!
6ince the electronWs energy is present solely as $inetic energy (B2 this can be rearranged
to give the following expression for the B2 of the photoelectron +
=4 > h - 4A
A
! - 4A! !
"he final term in brac$ets, representing the difference in energy between the ioni(ed and
neutral atoms, is generally called the binding energy (02 of the electron & this then leads
to the following commonly 'uoted e'uation +
=4 > h - 54
!n alternative approach is to consider a one&electron model along the lines of the following
pictorial representation A this model of the process has the benefit of simplicity but it can be
rather misleading.
"he 02 is now ta$en to be a direct measure of the energy re'uired to <ust remove the
electron concerned from its initial level to the vacuum level and the B2 of the
photoelectron is again given by +
=4 > h - 54
N="2 & the binding energies (02 of energy levels in solids are conventionally measured
with respect to the ,ermi&level of the solid, rather than the vacuum level. "his involves a
small correction to the e'uation given above in order to account for the -or# function (
of the solid, but for the purposes of the discussion below this correction will be neglected.
4xperimental Betails
"he basic re'uirements for a photoemission experiment (T86 or E86 are+
#. a source of fixed&energy radiation (an x&ray source for T86 or, typically, a >e
discharge lamp for E86

%. an electron energy analyser (which can disperse the emitted electrons according to
their $inetic energy, and thereby measure the flux of emitted electrons of a
particular energy

/. a high vacuum environment (to enable the emitted photoelectrons to be analysed
without interference from gas phase collisions
6uch a system is illustrated schematically below+
"here are many different designs of electron energy analyser but the preferred option for
photoemission experiments is a concentric hemispherical analyser (C>! which uses an
electric field between two hemispherical surfaces to disperse the electrons according to
their $inetic energy.
3.&.1 N-ray *hotoelectron Spectroscopy N*S!
,or each and every element, there will be a characteristic binding energy associated with
each core atomic orbital i.e. each element will give rise to a characteristic set of pea$s in
the photoelectron spectrum at $inetic energies determined by the photon energy and the
respective binding energies.
"he presence of pea$s at particular energies therefore indicates the presence of a specific
element in the sample under study & furthermore, the intensity of the pea$s is related to the
concentration of the element within the sampled region. "hus, the techni'ue provides a
&uantitative analysis of the surface co!position and is sometimes $nown by the alternative
acronym , 26C! (2lectron 6pectroscopy for Chemical !nalysis.
"he most commonly employed x&ray sources are those giving rise to +
*g B
radiation + h 3 #%:/.@ eS !l B
radiation + h 3 #1D@.@ eS
"he emitted photoelectrons will therefore have $inetic energies in the range of ca. 9 & #%:9
eS or 9 & #1D9 eS . 6ince such electrons have very short I*,8s in solids (see 6ection :.# ,
the techni'ue is necessarily surface sensitive.
4xample 1 - the N*S spectrum of *d metal
"he diagram below shows a real T86 spectrum obtained from a 8d metal sample using
*g B
radiation
& the main pea$s occur at $inetic energies of ca. //9, @G9, H%9, G#9 and G%9 eS.
6ince the energy of the radiation is $nown it is a trivial matter to transform the spectrum so
that it is plotted against 02 as opposed to B2.
"he most intense pea$ is now seen to occur at a binding energy of ca. //: eS
.or$ing downwards from the highest energy levels ......
#. the valence band (1d,:s emission occurs at a binding energy of ca. 9 & D eS
( measured with respect to the ,ermi level, or alternatively at ca. 1 & #% eS if
measured with respect to the vacuum level .
%. the emission from the 1p and 1s levels gives rise to very wea$ pea$s at :1 and DD
eS respectively
/. the most intense pea$ at ca. //: eS is due to emission from the /d levels of the 8d
atoms, whilst the /p and /s levels give rise to the pea$s at ca. :/1I:@# eS and @H/
eS respectively.
1. the remaining pea$ is not an T86 pea$ at all 4 & it is an !uger pea$ arising from x&
ray induced !uger emission. It occurs at a $inetic energy of ca. //9 eS (in this case
it is really meaningless to refer to an associated binding energy.
"hese assignments are summarised below ...
It may be further noted that
there are significant differences in the natural widths of the various photoemission
pea$s
the pea$ intensities are not simply related to the electron occupancy of the orbitals
4xercise 1 - the N*S spectrum of )aCl
"he diagram opposite shows an energy level diagram for
sodium with approximate binding energies for the core
levels.

If we are using *g B
( h 3 #%:/.@ eS radiation ...

... at what $inetic energy will the Na #s
photoelectron pea$ be observed 5
(the #s pea$ is that resulting from photoionisation of the
#s level
... at what $inetic energy will the Na %s and %p
photoelectron pea$s be observed 5
Spin-<rbit Splitting
Closer inspection of the spectrum shows that emission from some levels (most obviously
/p and /d does not give rise to a single photoemission pea$, but a closely spaced doublet.
.e can see this more clearly if, for example, we expand the spectrum in the region of the
/d emission ...
"he /d photoemission is in fact split between two pea$s, one at //1.G eS 02 and the other
at /19.% eS 02, with an intensity ratio of /+% . "his arises from spin&orbit coupling effects
in the final state. "he inner core electronic configuration of the initial state of the 8d is +
(#s
%
(%s
%
(%p
@
(/s
%
(/p
@
(/d
#9
....
with all sub&shells completely full.
"he removal of an electron from the /d sub&shell by photo&ioni(ation leads to a (/d
G
configuration for the final state & since the d&orbitals ( l 3 % have non&(ero orbital angular
momentum, there will be coupling between the unpaired spin and orbital angular momenta.
6pin&orbit coupling is generally treated using one of two models which correspond to the
two limiting ways in which the coupling can occur & these being the 8S (or Russell&
6aunders coupling approximation and the 7&7 coupling approximation.
If we consider the final ionised state of 8d within the Russell&6aunders coupling
approximation, the (/d
G
configuration gives rise to two states (ignoring any coupling with
valence levels which differ slightly in energy and in their degeneracy ...
%
D
:I%
g
<
3 %xO:I%PV# 3 @
%
D
/I%
g
<
3 %xO/I%PV# 3 1
"hese two states arise from the coupling of the 8 3 % and S 3 #I% vectors to give permitted
< values of /I% and :I%. "he lowest energy final state is the one with maximum < (since the
shell is more than half&full, i.e. < 3 :I%, hence this gives rise to the -lower binding energy-
pea$. "he relative intensities of the two pea$s reflects the degeneracies of the final states
(g
<
3 %< V #, which in turn determines the probability of transition to such a state during
photoioni(ation.
"he Russell&6aunders coupling approximation is best applied only to light atoms and this
splitting can alternatively be described using individual electron l&s coupling. In this case
the resultant angular momenta arise from the single hole in the d&shellA a d&shell electron
(or hole has l 3 % and s 3 #I%, which again gives permitted 7&values of /I% and :I% with the
latter being lower in energy.
"he pea$s themselves are conventionally annotated as indicated & note the use of lower
case lettering
"his spin&orbit splitting is of course not evident with s&levels (l 3 9, but is seen with p,d C
f core&levels which all show characteristic spin&orbit doublets.
Chemical Shifts
"he exact binding energy of an electron depends not only upon the level from which
photoemission is occurring, but also upon +
#. the formal oxidation state of the atom
%. the local chemical and physical environment
Changes in either (# or (% give rise to small shifts in the pea$ positions in the spectrum &
so&called che!ical shifts 0
6uch shifts are readily observable and interpretable in T8 spectra (unli$e in !uger spectra
because the techni'ue +
is of high intrinsic resolution (as core levels are discrete and generally of a well&
defined energy
is a one electron process (thus simplifying the interpretation
!toms of a higher positive oxidation state exhibit a higher binding energy due to the extra
coulombic interaction between the photo&emitted electron and the ion core. This ability to
discri!inate bet-een different oxidation states and che!ical environ!ents is one of the
!a7or strengths of the =S techni&ue0
In practice, the ability to resolve between atoms exhibiting slightly different chemical shifts
is limited by the pea$ widths which are governed by a combination of factors A especially
the intrinsic width of the initial level and the lifetime of the final state
the line&width of the incident radiation & which for traditional x&ray sources can
only be improved by using x&ray monochromators
the resolving power of the electron&energy analyser
In most cases, the second factor is the ma<or contribution to the overall line width.
Exa!ples of che!ical shifts
2g.# =xidation states of titanium
L"i %p spectraM.

Angle Bependent Studies
!s described in 6ection :.#, the degree of surface sensitivity of an electron&based techni'ue
such as T86 may be varied by collecting photoelectrons emitted at different emission
angles to the surface plane. "his approach may be used to perform non&destructive analysis
of the variation of surface composition with depth (with chemical state specificity.
Exa!ples
!ngle&dependent nnalysis of a silicon wafer with a native oxide layer
(courtesy of 8hysical 2lectronics, Inc. (8>I
3.&.% -ltra'iolet *hotoelectron Spectroscopy -*S!
In E86 the source of radiation is normally a noble gas discharge lamp A fre'uently a >e&
discharge lamp emitting >e I radiation of energy %#.% eS .
6uch radiation is only capable of ionising electrons from the outermost levels of atoms &
the valence levels. "he advantage of using such ES radiation over x&rays is the very
narrow line width of the radiation and the high flux of photons available from simple
discharge sources. "he main emphasis of wor$ using E86 has been in studying +
#. the electronic structure of solids & detailed angle resolved studies permit the
complete band structure to be mapped out in #&space.
%. the adsorption of relatively simple molecules on metals & by comparison of the
molecular orbitals of the adsorbed species with those of both the isolated molecule
and with calculations.
3.0. +ibrational Spectroscopy
Sibrational spectroscopy provides the most definitive means of identifying the surface
species generated upon molecular adsorption and the species generated by surface
reactions. In principle, any techni'ue that can be used to obtain vibrational data from solid
state or gas phase samples (IR, Raman etc. can be applied to the study of surfaces & in
addition there are a number of techni'ues which have been specifically developed to study
the vibrations of molecules at interfaces (22?6, 6,R etc..
"here are, however, only two techni'ues that are routinely used for vibrational studies of
molecules on surfaces & these are +
#. I. Spectroscopy of 'arious formsQ e.g. .AI.SQ MI.!
%. 4lectron 4nergy Coss Spectroscopy 44CS !
I. Spectroscopy
"here are a number of ways in which the IR techni'ue may be implemented for the study
of adsorbates on surfaces.
,or solid samples possessing a high surface area +
"ransmission IR 6pectroscopy + employing the same basic experimental geometry as that
used for li'uid samples and mulls. "his is often used for studies on supported metal
catalysts where the large metallic surface area permits a high concentration of adsorbed
species to be sampled. "he solid sample must, of course, be IR transparent over an
appreciable wavelength range.
Diffuse Reflectance IR 6pectroscopy ( DRI,"6 + in which the diffusely scattered IR
radiation from a sample is collected, refocused and analysed. "his modification of the IR
techni'ue can be employed with high surface area catalytic samples that are not sufficiently
transparent to be studied in transmission.
,or studies on low surface area samples (e.g. single crystals +
Reflection&!bsorption IR 6pectroscopy ( R!IR6 + where the IR beam is specularly
reflected from the front face of a highly&reflective sample, such as a metal single crystal
surface.
*ultiple Internal Reflection 6pectroscopy ( *IR + in which the IR beam is passed
through a thin, IR transmitting sample in a manner such that it alternately undergoes total
internal reflection from the front and rear faces of the sample. !t each reflection, some of
the IR radiation may be absorbed by species adsorbed on the solid surface & hence the
alternative name of !ttenuated "otal Reflection (!"R.
.AI.S - the Study of Adsorbates on Metallic Surfaces by .eflection I. Spectroscopy
It can be shown theoretically that the best sensitivity for IR measurements on metallic
surfaces is obtained using a gra(ing&incidence reflection of the IR light.
,urthermore, since it is an optical (photon inIphoton out techni'ue it is not necessary for
such studies to be carried out in vacuum. "he techni'ue is not inherently surface&specific,
but
there is no bul$ signal to worry about
the surface signal is readily distinguishable from gas&phase absorptions using
polari(ation effects.
=ne ma<or problem, is that of sensitivity (i.e. the signal is usually very wea$ owing to the
small number of adsorbing molecules. "ypically, the sampled area is ca. # cm
%
with less
than #9
#:
adsorbed molecules (i.e. about # nanomole. .ith modern ,"IR spectrometers,
however, such small signals (9.9#K & %K absorption can still be recorded at relatively high
resolution (ca. # cm
&#
.
,or a number of practical reasons, low fre'uency modes ( J @99 cm
&#
are not generally
observable & this means that it is not usually possible to see the vibration of the metal&
adsorbate bond and attention is instead concentrated on the intrinsic vibrations of the
adsorbate species in the range @99 & /@99 cm
&#
.
Selection "ules
"he observation of vibrational modes of adsorbates on metallic substrates is sub<ect to the
surface dipole selection rule.
"his states that only those vibrational modes which give rise to an oscillating dipole
perpendicular (normal to the surface are IR active and give rise to an observable
absorption band.
,urther information on the selection rules for surface IR spectroscopy can be found in the
review by 6heppard C 2r$elens L!ppl. 6pec. /D, 1H# (#GD1M. It also needs to be
remembered that even if a transition is allowed it may still be very wea$ if the transition
moment is small.
Exa!ples of "eflection I" Spectra
2g.# Nitric oxide (N=
adsorption on a 8t surface.
2g.% >CN adsorption on a 8t
surface
4lectron 4nergy Coss Spectroscopy 44CS!
"his is a techni'ue utilising the inelastic scattering of low energy electrons in order to
measure vibrational spectra of surface species + superficially, it can be considered as the
electron&analogue of Raman spectroscopy.
"o avoid confusion with other electron energy loss techni'ues it is sometimes referred to as
>R22?6 & high resolution 22?6 or S2?6 & vibrational 2?6
6ince the techni'ue employs low energy electrons, it is necessarily restricted to use in high
vacuum (>S and E>S environments & however, the use of such low energy electrons
ensures that it is a surface specific techni'ue and, arguably, it is the vibrational techni'ue
of choice for the study of most adsorbates on single crystal substrates.
"he basic experimental geometry is fairly simple as illustrated schematically below & it
involves using an electron monochromator to give a well&defined beam of electrons of a
fixed incident energy, and then analysing the scattered electrons using an appropriate
electron energy analyser.
! substantial number of electrons are elastically scattered ( E 3 E
o
& this gives rise to a
strong elastic pea# in the spectrum.
=n the low $inetic energy side of this main pea$ ( E J E
o
, additional wea$ pea$s are
superimposed on a mildly sloping bac$ground. "hese pea$s correspond to electrons which
have undergone discrete energy losses during the scattering from the surface.
"he magnitude of the energy loss, E 3 (E
o
& E, is e'ual to the vibrational 'uantum (i.e.
the energy of the vibrational mode of the adsorbate excited in the inelastic scattering
process. In practice, the incident energy ( E
o
is usually in the range :&#9 eS (although
occasionally up to %99 eS and the data is normally plotted against the energy loss
(fre'uently measured in meS.
Selection "ules
"he selection rules that determine whether a vibrational band may be observed depend
upon the nature of the substrate and also the experimental geometry+ specifically the angles
of the incident and (analysed scattered beams with respect to the surface.
,or metallic substrates and a specular geometry, scattering is principally by a long&range
dipole mechanism. In this case the loss features are relatively intense, but only those
vibrations giving rise to a dipole change normal to the surface can be observed.
0y contrast, in an off&specular geometry, electrons lose energy to surface species by a
short&range impact scattering mechanism. In this case the loss features are relatively wea$
but all vibrations are allowed and may be observed.
If spectra can be recorded in both specular and off&specular modes the selection rules for
metallic substrates can be put to good use & helping the investigator to obtain more
definitive identification of the nature and geometry of the adsorbate species.
"he resolution of the techni'ue (despite the >R22?6 acronym 4 is generally rather poor A
19&D9 cm
&#
is not untypical. ! measure of the instrumental resolution is given by loo$ing at
the ,.>* (full&width at half maximum of the elastic pea$.
"his poor resolution can cause problems in distinguishing between closely similar surface
species & however, recent improvements in instrumentation have opened up the possibility
of much better spectral resolution ( J #9 cm
&#
and will undoubtedly enhance the utility of
the techni'ue.
In summary, there are both advantages and disadvantages in utilising 22?6, as opposed to
IR techni'ues, for the study of surface species It offers the advantages of ...
high sensitivity
variable selection rules
spectral ac'uisition to below 199 cm
&#

but suffers from the limitations of ...
use of low energy electrons (re'uiring a >S environment and hence the need for
low temperatures to study wea$ly&bound species, and also the use of magnetic
shielding to reduce the magnetic field in the region of the sample
re'uirement for flat, preferably conducting, substrates
lower resolution
Applications of +ibrational Spectroscopy
=ne of the classic examples of an area in which vibrational spectroscopy has contributed
significantly to the understanding of the surface chemistry of an adsorbate is that of +
Molecular Adsorption of CO on Metallic Surfaces
!dsorbed carbon monoxide usually gives rise to strong absorptions in both the IR and
22?6 spectra at the (C&= stretching fre'uency. "he metal&carbon stretching mode (ca. 199
cm
&#
is usually also accessible to 22?6.
"he interpretation of spectra of C= as an adsorbed surface species draws heavily upon IR
spectra from related inorganic cluster and coordination complexes & the structure of such
complexes usually being available from x&ray single crystal diffraction measurements.
"his comparison suggests that the C= stretching fre'uency can provide a good indication
of the surface coordination of the molecule + as a rough guideline,
(C&=
C= ( gas phase %#1/ cm
&#
"erminal C= %#99 & #G%9 cm
&#
0ridging ( %f site #G%9 & #D99 cm
&#
0ridging ( /f I 1f site J #D99 cm
&#
Exa!ple +
R!IR spectrum for C= chemisorbed on a clean 8t surface.
"he reduction in the stretching fre'uency of terminally&bound C= from the value observed
for the gas phase molecule ( %#1/ cm
&#
can be explained in terms of the Dewar&Chatt or
2lyholder !odel for the bonding of C= to metals.
"his simple model considers the metal&C= bonding to consist of two
main components +
A + this is a bonding interaction due to overlap of a filled &-lone
pair- orbital on the carbon atom with empty metal orbitals of the
correct symmetry & this leads to electron density transfer from the
C= molecule to the metal centre.
5 + this is a bonding interaction due to overlap of filled metal d
(and p 5 orbitals with the X antibonding molecular orbital of the
C= molecule. 6ince this interaction leads to the introduction of
electron density into the C= antibonding orbital there is a
conse'uent reduction in the C= bond strength and its intrinsic
vibrational fre'uency (relative to the isolated molecule.
,or a more detailed discussions on the bonding of C= to metals, you are recommended to
refer to one of the following +
- !dvanced Inorganic Chemistry - by ,.!. Cotton C R. .il$inson (:th 2dn. pp. :D & @% .
- 6olids C 6urfaces + a chemistWs view of bonding in extended structures - by R. >offman
pp. H#&H1.
3.3. Secondary Ion Mass Spectrometry
"he techni'ue of 6econdary Ion *ass 6pectrometry (6I*6 is the most sensitive of all the
commonly&employed surface analytical techni'ues & this is because of the inherent
sensitivity associated with mass spectrometric&based techni'ues.
"here are a number of different variants of the techni'ue +
6tatics 6I*6 + used for sub&monolayer elemental analysis
Dynamic 6I*6 +
used for obtaining compositional information as a function of
depth below the surface
Imaging 6I*6 + used for spatially&resolved elemental analysis
!ll of these variations on the techni'ue are based on the same basic physical process and it
is this process which is discussed here, together with a brief introduction to the field of
static 6I*6. ,urther notes on dynamic and imaging 6I*6 can be obtained in 6ection H.1 .
In 6I*6 the surface of the sample is sub<ected to bombardment by high energy ions & this
leads to the e<ection (or sputtering of both neutral and charged (VI& species from the
surface. "he e<ected species may include atoms, clusters of atoms and molecular fragments.
In traditional 6I*6 it is only the positive ions that are mass analysed & this is primarily for
practical ease but it does lead to problems with 'uantifying the compositional data since the
positive ions are but a small, non&representative fraction of the total sputtered species. It
should be further noted that the displaced ions have to be energy filtered before they are
mass analysed (i.e. only ions with $inetic energies within a limited range are mass
analysed.
"he most commonly employed incident ions (denoted by I
V
in the above diagram used for
bombarding the sample are argon ions ( !r
V
but other ions (e.g. al$ali metal ions, Ra
V

are preferred for some applications.
"he mass analyser is typically a 'uadrupole *6 analyser with unit mass resolution, but
high specification time&of&flight ("=, analysers are also used and provide substantially
higher sensitivity and a much greater mass range (albeit at a higher cost.
Static SIMS
In static 6I*6 (66I*6 the aim is to obtain sufficient signal to provide compositional
analysis of the surface layer without actually removing a significant fraction of a
monolayer & i.e. to be able to analyse less than ca. #9
#1
atomsImolecules (about #9K of a
monolayer for a #cm
%
sample. "he techni'ue is then capable of providing information
about the topmost single atomic layer of the surface.
Exa!ple + static 6I*6 spectra from the surface of 8",2 (polytetrafluoroethylene
Note the substantial
differences between the
positive and negative ion
spectra.
In each case the most
intense pea$s correspond
to relatively stable positive
and negative ions
respectively.
3.7 Temperature-*rogrammed TechniDues
"here are a range of techni'ues for studying surface reactions and molecular adsorption on
surfaces which utilise temperature&programming to discriminate between processes with
different activation parameters. =f these, the most useful for single crystal studies is +

"emperature 8rogrammed Desorption ("8D
.hen the techni'ue is applied to a system in which the adsorption process is, at least in
part, irreversible and T&programming leads to surface reactions, then this techni'ue is often
$nown as +

"emperature 8rogrammed Reaction 6pectroscopy ("8R6
>owever, there is no substantive difference between "8R6 and "8D.
"he basic experiment is very simple, involving
#. !dsorption of one or more molecular species onto the sample surface at low
temperature (fre'uently /99 B, but sometimes sub&ambient.
%. >eating of the sample in a controlled manner (preferably so as to give a linear
temperature ramp whilst monitoring the evolution of species from the surface bac$
into the gas phase.
In modern implementations of the techni'ue the detector of choice is a small, 'uadrupole
mass spectrometer (Z*6 and the whole process is carried out under computer control
with 'uasi&simultaneous monitoring of a large number of possible products.
"he data obtained from such an experiment consists of the intensity variation of each
recorded mass fragment as a function of time I temperature. In the case of a simple
reversible adsorption process it may only be necessary to record one signal & that
attributable to the molecular ion of the adsorbate concerned.
"he graph below shows data from a "8D experiment following adsorption of C= onto a
8d(### crystal at /99 B.
6ince mass spectrometric detection is used the sensitivity of the techni'ue is good with
attainable detection limits below 9.#K of a monolayer of adsorbate.
"he following points are worth noting +
#. "he area under a pea$ is proportional to the amount originally adsorbed, i.e.
proportional to the surface coverage.
%. "he $inetics of desorption (obtained from the pea$ profile and the coverage
dependence of the desorption characteristics give information on the state of
aggregation of the adsorbed species e.g. molecular vWs dissociative.
/. "he position of the pea$ (the pea# te!perature is related to the enthalpy of
adsorption, i.e. to the strength of binding to the surface.
=ne implication of the last point, is that if there is more than one binding state for a
molecule on a surface (and these have significantly different adsorption enthalpies then
this will give rise to multiple pea$s in the "8D spectrum.
"he graph below shows data from a "8D experiment following adsorption of oxygen on
8t(### at D9 B.
.hich pea$ corresponds to the state which is more wea$ly adsorbed 5
Theory of T*B
!s discussed in 6ection %.@, the rate of desorption of a surface species will in general be
given by an expression of the form +
R
des
> . N
x
exp - E
a
des
? R T ! L#M
where "
des
& desorption rate ( 3 &d)Idt
x & $inetic order of desorption (typically 9,# or %
E
a
des
& activation energy for desorption
In a temperature programmed desorption experiment in which the temperature is increased
linearly with time from some initial temperature T
o
, then +
T > T
o
A .t and dT > .dt L%M
"he intensity of the desorption signal, I(T , is proportional to the rate at which the surface
concentration of adsorbed species is decreasing. "his is obtained by combining L#M and L%M
to give
L/M
"his problem may also be considered in a rather simplistic graphical way &the $ey to this is
to recognise that the expression for the desorption signal given in the above e'uation is
basically a product of a coverage term ( )
x
& where ) depends on T and an exponential
term (involving both E
a
and T .
Initially, at low temperatures E
a
UU "T and the exponential term is vanishingly small.
>owever, as the temperature is increased this term begins to increase very rapidly when the
value of "T approaches that of the activation energy, E
a
.
0y contrast, the pre&exponential term is dependent upon the coverage, )(T, at the
temperature concerned & this term will remain at the initial value until the desorption rate
becomes of significance as a result of the increasing exponential term. "hereafter, it will
decrease ever more rapidly until the coverage is reduced to (ero. "he shaded area is an
approximate representation of the product of these two functions, and hence also an
approximate representation for the desorption signal itself & whilst this illustration may be
overly simplistic it does clearly show why the desorption process gives rise to a well&
defined desorption pea$.
If you wish to see exactly how the various factors such as the $inetic order and initial
coverage influence the desorption profile then try out the Interactive Demonstration of
"emperature 8rogrammed Desorption (note & this is based on the formulae given in 6ection
%.@ A we will however continue to loo$ at one particular case in a little more detail.
CAS4 I Molecular adsorption! & in this case the desorption $inetics will usually be first
order (i.e. x 3 #
"he maximum desorption signal in the I(T trace will occur when ( dI I dT 3 9 ,
i.e. when
>ence , remembering that the surface coverage changes with temperature i.e. ) 3 )(T,
where we have substituted E
a
for E
a
des
purely for clarity of presentation.
6ubstituting for d)IdT from e'tn.L/M then gives
"he solution is given by setting the expression in s'uare brac$ets to be e'ual to (ero , i.e.
where we have now defined the temperature at which the desorption maximum occurs to be
T 3 T
p
(the pea# te!perature .
Enfortunately, this e'uation cannot be re&arranged to ma$e T
p
the sub<ect (i.e. to give a
simple expression of the form T
p
3 ....., but we can note that +
#. as E
a
des
(the activation energy for desorption increases, then so T
p
(the pea$
temperature increases.
%. the pea$ temperature is not dependent upon, and conse'uently does not change
with, the initial coverage, )
t39
.
/. the shape of the desorption pea$ will tend to be asymmetric, with the signal
decreasing rapidly after the desorption maximum.

Temperature *rogrammed .eaction Spectroscopy
In "8R6 a number of desorption products will normally be detected in any one experiment
& this is where mass spectrometric detection and multiple ion monitoring really becomes
essential.
! good example of "8R6 results are those obtained from dosing formic acid ( >C==>
onto a copper surface.
"he spectra contain two regions of interest & the first at around room temperature where
there is desorption of hydrogen (!Iz 3 % alone, and the second between 199 C :99 B
where there are coincident desorption pea$s attributable to hydrogen and carbon dioxide
(!Iz 3 11.
"he lower temperature hydrogen pea$ is at about the same temperature at which hydrogen
atoms recombine and desorb as molecular hydrogen from a hydrogen&dosed Cu(##9
surface i.e. the $inetics of its evolution into the gas phase are those of the normal
recombinative desorption process. 0y contrast, the higher temperature hydrogen pea$ is
observed well above the normal hydrogen desorption temperature & its appearance must be
governed by the $inetics of decomposition of another surface species. "he carbon dioxide
pea$ must also be deco!position li!ited since on clean Cu(##9 carbon dioxide itself is
only wea$ly physisorbed at very low temperatures.
*ore generally, you should note that coincident desorption of two species at temperatures
well above their normal desorption temperatures is characteristic of the decomposition of a
more complex surface species.
.hat do you thin$ is the surface species responsible for the high temperature pea$s in this particular
case 5
7. <'erlayer Structures 8 Surface Biffraction
7.1 Classification of <'erlayer Structures
!dsorbed species on single crystal surfaces are fre'uently found to exhibit long&range
ordering A that is to say that the adsorbed species form a well&defined overlayer structure.
2ach particular structure may only exist over a limited coverage range of the adsorbate, and
in some adsorbateIsubstrate systems a whole progression of adsorbate structures are formed
as the surface coverage is gradually increased.
"his section deals with the classification of such ordered structures & in most cases this
involves describing the overlayer structure in terms of the underlying structure of the
substrate.
"here are two principal methods for specifying the structure A these are
#. .oodWs notation
%. matrix notation.
0efore we start to discuss overlayer structures, however, we need to ma$e sure that we can
ade'uately describe the structure of the substrate 4
The Concept of the Surface -nit Cell
"he primitive unit cell is the simplest periodically repeating unit which can be identified in
an ordered array & the array in this instance being the ordered arrangement of surface
atoms. 0y repeated translation of a unit cell, the whole array can be constructed.
?et us consider the clean surface structures of the low index surface planes of fcc metals ...
The fcc1((! surface
"he fcc(#99 surface has 1&fold rotational symmetry (-s'uare symmetry- & perhaps it
should not surprise us therefore to find that the primitive unit cell for this surface is s'uare
in shape 4
fcc1((! surface
"wo possible choices of unit cell are highlighted & it is clear that a unit cell of this si(e is
indeed going to be the simplest possible repeating unit for this surface. "he two alternatives
drawn are in fact but two of an infinte number of possibilities A they have the same
shapeIsymmetry, si(e and orientation , differing only in their translational position or
-origin-.
.hichever we choose then it is clear that we can indeed generate the whole surface
structure by repeated translation of the unit cell A for example ....
In fact, I personally prefer the alternative choice of unit cell which has the corners of the
unit cell coincident with the atomic centres.
.e now need to thin$ how to define the unit cell shape, si(e and symmetry & this is best
done using two vectors which have a common origin and define two sides of the unit cell ...
,or this fcc(#99 surface the two vectors which define the unit cell, conventionally called a
1
C a
%
, are +
the same length i.e. [a
1
[ 3 [a
%
[
mutually perpendicular
0y convention , one also selects the vectors such that you go anticloc$wise from a
1
to get
to a
%
.
L Note + the length of the vectors a
1
C a
%
is related to the bul$ unit cell parameter, a , by [a
1
[
3 [a
%
[ 3 a I % M
The fcc11(! surface
In the case of the fcc(##9 surface, which has %&fold rotational symmetry, the unit cell is
rectangular
fcc11(! surface
0y convention, [a
%
[ U [a
1
[ & if we also recall the convention that one goes anticloc$wise to
get from a
1
to a
%
, then this leads to the choice of vectors shown.
The fcc111! surface
.ith the fcc(### surface we again have a situation where the length of the two vectors are
the same i.e. [ a
1
[ 3 [ a
%
[ . .e can either $eep the angle between the vectors less than G9
degrees or let it be greater than G9 degrees. "he normal convention is to choose the latter,
i.e. the right hand cell of the two illustrated with an angle of #%9 degrees between the two
vectors.
fcc111! surface
<'erlayer Structures
If we have an ordered overlayer of adsorbed species (atoms or molecules, then we can use
the same basic ideas as outlined in the previous section to define the structure.
"he adsorbate unit cell is usually defined by the two vectors b
1
and b
%
. "o avoid
ambiguities, it again helps if we stic$ to a set of conventions in choosing the unit cell
vectors. In this case +
#. b
%
is again selected to be anticloc$wise from b
1
.
%. if possible, b
1
is chosen to be parallel to a
1
and b
%
parallel to a
%
.
=nce the unit cell vectors for substrate and adsorbate have been selected then it is a
relatively simple matter to wor$ out how to denote the structure.
"oodFs )otation
.oodWs notation is the simplest and most fre'uently used method for describing a surface
structure & it only wor$s, however, if the two unit cells are of the same symmetry or
closely&related symmetries (more specifically, the angle between b
1
C b
%
must be the same
as that between a
1
C a
%
.
In essence, .oodWs notation first involves specifying the lengths of the two overlayer
vectors, b
1
C b
%
, in terms of a
1
C a
%
respectively & this is then written in the format +
( [b
1
[I[a
1
[ x [b
%
[I[a
%
[
i.e. a ( % x % structure has [b
1
[ 3 %[a
1
[ and [b
%
[ 3 %[a
%
[ .
"he following diagram shows a ( % x % adsorbate overlayer on an fcc(#99 surface in
which the adsorbate is bonded terminally on&top of individual atoms of the substrate.
Substrate 2 fcc1((!
Substrate unit cell
Adsorbate unit cell
"he unit cells of the (#99 substrate and the ( % x % overlayer are both highlighted.
"he next diagram shows another ( % x % structure , but in this case the adsorbate species is
bonded in the four&fold hollows of the substrate surface. =f course, only a very limited
section of the structure can be shown here & in practice, the unit cell shown would repeat to
give a complete overlayer structure extending across the substrate surface.
Substrate 2 fcc1((!
Substrate unit cell
Adsorbate unit cell
"he highlighted unit cells of the adsorbate and substrate are identical in si(e, shape and
orientation to those of the previously illustrated ( % x % structure.
0oth this and the previous structure are examples of pri!itive ( % x % , or p( % x % ,
structures. "hat is to say that they are indeed the simplest unit cells that may be used to
describe the overlayer structure, and contain only one -repeat unit-.
,or the purposes of this tutorial I shall follow common practice and omit the preceding -p-
& referring to such structures simply as ( % x % structures (in spo$en language, -t-o by
t-o- structures.
6uch ( % x % structures are also found on other surfaces, but they may differ mar$edly in
superficial appearance from the structure on the fcc(#99 surface. "he following diagram,
for example, shows a ( % x % structure on a fcc(##9 surface
Substrate 2 fcc11(!
Substrate unit cell
Adsorbate unit cell
"he adsorbate unit cell is again twice as large as that of the substrate in both dimensions & it
retains the same aspect ratio as the rectangular substrate unit cell (# + #.1#1 and does not
exhibit any rotation with respect to the substrate cell.
"he following diagram shows yet another ( % x % structure, in this case on the fcc(###
surface ...
Substrate 2 fcc111!
Substrate unit cell
Adsorbate unit cell
!gain, the adsorbate unit cell is of the same symmetry as the substrate cell but is scaled up
by a factor of two in its linear dimensions (and corresponds to a surface area four times as
large as that of the substrate unit cell.
"he next example is a surface structure which is closely related to the ( % x % structure + it
differs in that there is an additional atom in the middleIcentre of the ( % x % adsorbate unit
cell.
6ince the middle atom is -crystallographically e'uivalent- to those at the corners (i.e. it is
not distinguishable by means of different coordination to the underlying substrate or any
other structural feature, then this is no longer a primitive ( % x % structure.
Substrate 2 fcc1((!
c % x % !
% x %!.03
Instead it may be classified in one of two ways +
(i !s a centred ( % x % structure i.e. c( % x %
L where we are using a non&primitive unit cell containing % repeat units M
or ...
(ii !s a - ( % x %R1: - structure , where we are specifying the true primitive unit cell .
In using the latter .oodWs notation we are stating that the adsorbate unit cell is a factor of
% larger than the substrate unit cell in both directions and is also rotated by 1: degrees with
respect to the substrate unit cell.
)ote 5 if the >central> ato! is not co!pletely crystallographically e&uivalent. then the
structure for!ally re!ains a p%1x1' unit cell but no- has a basis of t-o adsorbate ato!s
per unit cell0
In some instances it is possible to use a centred unit cell description for a structure for
which the primitive unit cell cannot be described using .oodWs notation & for example, the
c( % x % structure on the fcc(##9 surface shown below.
Substrate 2 fcc11(!
c % x % !
!s a final example, the next diagram illustrates a commonly&observed structure on fcc(###
surfaces which can be readily described using .oodWs notation.
Substrate 2 fcc111!
& x &!.&(
(You should confirm for yourself that the adsorbate unit cell is indeed scaled up from the
substrate cell by the factor given and rotated by /9 degrees 4
Matrix )otation
"his is a much more general system of describing surface structures which can be applied
to all ordered overlayers + 'uite simply it relates the vectors b
1
C b
%
to the substrate vectors
a
1
C a
%
using a simple matrix i.e.
Matrix )otation
& remember a
1
, a
%
, b
1
and b
%
are vectors.
"o illustrate the use of matrix notation we shall now consider two surface structures with
which we are already familiar ...
Substrate 2 fcc 1((!
% x %! o'erlayer
,or the (% x % structure we have +
0y contrast, for the c(% x % structure +
Substrate 2 fcc 1((!
c% x %! o'erlayer
we have
Summary
=rdered surface structures may be described by defining the adsorbate unit cell in terms of
that of the underlying substrate using +
#. "oodFs )otation + in which the lengths of b
1
and b
%
are given as simple multiples
of a
1
and a
%
respectively, and this is followed by the angle of rotation of b
1
from a
1
(if this is non&(ero.
%. Matrix )otation + in which b
1
and b
%
are independently defined as linear
combinations of a
1
and a
%
and these relationships are expressed in a matrix format.
7.% Co# 4nergy 4lectron Biffraction C44B !
?22D is the principal techni'ue for the determination of surface structures + it may be used
in one of two ways &
#. Zualitatively + where the diffraction pattern is recorded and analysis of the spot
positions yields information on the si(e, symmetry and rotational alignment of the
adsorbate unit cell with respect to the substrate unit cell.
%. Zuantitatively + where the intensities of the various diffracted beams are recorded
as a function of the incident electron beam energy to generate so&called I&S curves
which, by comparison with theoretical curves, may provide accurate information on
atomic positions.
In this section, we will only consider the 'ualitative application of this experimental
techni'ue.
4xperimental Betails
"he ?22D experiment uses a beam of electrons of a well&defined low energy (typically in
the range %9 & %99 eS incident normally on the sample. "he sample itself must be a single
crystal with a well&ordered surface structure in order to generate a bac$&scattered electron
diffraction pattern. ! typical experimental set&up is shown below.
=nly the elastically&scattered electrons contribute to the diffraction pattern A the lower
energy (secondary electrons are removed by energy&filtering grids placed in front of the
fluorescent screen that is employed to display the pattern.
If you want to $now what the real thing loo$s li$e then clic$ on the photo 4
5asic Theory of C44B
0y the principles of wave&particle duality, the beam of electrons may be e'ually regarded
as a succession of electron waves incident normally on the sample. "hese waves will be
scattered by regions of high localised electron density, i.e. the surface atoms, which can
therefore be considered to act as point scatterers.
"he wavelength of the electrons is given be the de 0roglie relation +
.avelength, > h ? p ( where p & electron momentum
Now ,
p 3 !.v 3 (%!E
#
#I%
3 (%!.e.3
#I%

where ! & mass of electron L $g M
v & velocity L m s
&#
M
E
#
& $inetic energy
e & electronic charge
3 & acceleration voltage ( 3 energy in eS
3U .avelength, > h ? %m.e.V !
1?%

.hat is the wavelength of electrons of energy (i %9 eS 5
( Note & h 3 @.@% x #9
&/1
Q s , e 3 #.@9 x #9
&#G
C , !
e
3 G.## x #9
&/#
$g .
.hat is the wavelength of electrons of energy (ii %99 eS 5
,rom the above examples the range of wavelengths of electrons employed in ?22D
experiments is seen to be comparable with atomic spacings, which is the necessary
condition for diffraction effects associated with atomic structure to be observed.
Consider, first, a one dimensional (#&D chain of atoms (with atomic separation a with the
electron beam incident at right angles to the chain. "his is the simplest possible model for
the scattering of electrons by the atoms in the topmost layer of a solidA in which case the
diagram below would be representing the solid in cross&section with the electron beam
incident normal to the surface from the vacuum above.
If you consider the bac$scattering of a wavefront from two ad<acent atoms at a well&
defined angle, , to the surface normal then it is clear that there is a -path difference- (d
in the distance the radiation has to travel from the scattering centres to a distant detector
(which is effectively at infinity & this path difference is best illustrated by considering two
-ray paths- such as the right&hand pair of green traces in the above diagram.
"he si(e of this path difference is a sin and this must be e'ual to an integral number of
wavelengths for constructive interference to occur when the scattered beams eventually
meet and interfere at the detector i.e.
d > a sin > n.
where + & wavelength
n & integer (..&#, 9, #, %,..
,or two isolated scattering centres the diffracted intensity varies slowly between (ero
(complete destructive interference A d 3 (n V \ and its maximum value (complete
constructive interference A d 3 n. & with a large periodic array of scatterers, however, the
diffracted intensity is only significant when the -0ragg condition-
a sin > n.
is satisfied exactly. "he diagram below shows a typical intensity profile for this case.
"here are a number of points worth noting from this simple #&D model
#. the pattern is symmetric about 3 9 (or sin 3 9
%. sin is proportional to # I 3
#I%
(since is proportional to # I 3
#I%

/. sin is inversely proportional to the lattice parameter , a
"he aforementioned points are in fact much more general & all surface diffraction patterns
show a symmetry reflecting that of the surface structure, are centrally symmetric, and of a
scale showing an inverse relationship to both the s'uare root of the electron energy and the
si(e of the surface unit cell.
!s an example we can loo$ at the ?22D pattern from an fcc(##9 surface. In the diagram
below the surface atomic structure is shown on the left in plan view, as if you are viewing
it from the position of the electron gun in the ?22D experiment (albeit greatly magnified.
"he primary electron beam would then be incident normally on this surface as if fired from
your current viewpoint and the diffracted beams would be scattered from the surface bac$
towards yourself. "he diffraction pattern on the right illustrates how these diffracted beams
would impact upon the fluorescent screen.
"he pattern shows the same rectangular symmetry as the substrate surface but is -stretched-
in the opposite sense to the real space structure due to the reciprocal dependence upon the
lattice parameter. "he pattern is also centrosymmetric about the (99 beam & this is the
central spot in the diffraction pattern corresponding to the beam that is diffracted bac$
exactly normal to the surface (i.e. the n 3 9 case in our #&D model.
"he above illustration of the diffraction pattern shows only the -first&order- beams i.e. it is
representative of the diffraction pattern visible at low energies when only for n 3 # is the
angle of diffraction, , sufficiently small for the diffracted beam to be incident on the
display screen.
0y contrast, the diagram below shows the diffraction pattern that might be expected if the
energy of the incident electrons is doubled & some of the second order spots are now visible
and the pattern as a whole has apparently contracted in towards the central (99 spot.
"his is what the real diffraction patterns might loo$ li$e 7
In the case of such simple ?22D patterns, it is possible to explain the diffraction pattern in
terms of scattering from ro#s of atoms on the surface. ,or example, the rows of atoms
running vertically on the screen would give rise to a set of diffracted beams in the
hori(ontal plane, perpendicular to the rows, thus leading to the row of spots running in a
line hori(ontally across the diffraction pattern through the (99 spot. "he further the rows
are apart, then the closer in are the diffracted beams to the central (99 beam. "his is,
however, a far from satisfactory method of explaining ?22D patterns from surfaces.
! much better method of loo$ing at ?22D diffraction patterns involves using the concept
of reciprocal space + more specifically , it can be readily shown that &
- The obser'ed C44B pattern is a scaled! representation of the reciprocal net of the
pseudo-%B surface structure -
(No proof given 4
"he reciprocal net is determined by (defined by the reciprocal vectors +
a
1
E C a
%
E (for the substrate and b
1
E C b
%
E (for the adsorbate
Initially we will consider <ust the substrate. "he reciprocal vectors are related to the real
space unit cell vectors by the scalar product relations +
a
1
. a
%
E 3 a
1
E . a
%
3 9
and
a
1
. a
1
E 3 a
%
. a
%
E 3 #
,or those not too $een on vector algebra these mean that +
a
1
is perpendicular to a
%
E , and a
%
is perpendicular to a
1
E
there is an inverse relationship between the lengths of a
1
and a
1
E (and a
%
and a
%
E
of the form +
[a
1
[ 3 # I ( [a
1
E[ cos ! , where ! is the angle between the vectors a
1
and a
1
E.
Note + when ! 3 9 degrees (cos ! 3 # this simplifies to a simple reciprocal relationship
between the lengths a
1
and a
1
E.
2xactly analogous relations hold for the real space and reciprocal vectors of the adsorbate
overlayer structure + b
1
, b
1
E , b
%
and b
%
E .
"o a first approximation, the ?22D pattern for a given surface structure may be obtained
by superimposing the reciprocal net of the adsorbate overlayer (generated from b
1
E and b
%
E
on the reciprocal net of the substrate (generated from a
1
E and a
%
E
4xample 1
?et us now loo$ at an example & the diagram below shows an fcc(#99 surface (again in
plan view and its corresponding diffraction pattern (i.e. the reciprocal net .
.e can demonstrate how these reciprocal vectors can be determined by wor$ing through
the problem in a parallel fashion for the two vectors +
a
1
E must be perpendicular to a
%
a
%
E must be perpendicular to a
1

a
1
E is parallel to a
1
a
%
E is parallel to a
%

"he angle, ! , between a
1
C a
1
E is (ero "he angle, ! , between a
%
C a
%
E is (ero
>ence, [ a
1
E[ 3 # I [ a
1
[ >ence, [ a
%
E[ 3 # I [ a
%
[
If we let [ a
1
[ 3 # unit, then [ a
1
E[ 3 # unit. [ a
%
[ 3 [ a
1
[ 3 # unit, therefore [ a
%
E[ 3 # unit.
?et us now add in an adsorbate overlayer & a primitive ( % x % structure with the adsorbed
species shown bonded in on&top sites & and apply the same logic as <ust used above to
determine the reciprocal vectors, b
1
E and b
%
E, for this overlayer.
b
1
E must be perpendicular to b
%
b
%
E must be perpendicular to b
1

b
1
E is parallel to b
1
b
%
E is parallel to b
%

"he angle, 0 , between b
1
C b
1
E is (ero "he angle, 0 , between b
%
C b
%
E is (ero
>ence, [ b
1
E[ 3 # I [ b
1
[ >ence, [ b
%
E[ 3 # I [ b
%
[
[ b
1
[ 3 %[ a
1
[ 3 % unitsA [ b
1
E[ 3 \ unit. [ b
%
[ 3 %[ a
%
[ 3 % unitsA [ b
%
E[ 3 \ unit.
!ll we have to do now is to generate the reciprocal net for the adsorbate using b
1
E and b
%
E
(shown in red.
"hatWs all there is to it 4
4xample %
,or the second example, we will loo$ at the c( % x % structure on the same fcc(#99
surface. "he diagram below shows both a real space c( % x % structure and the
corresponding diffraction pattern +
In many respects the analysis is very similar to that for the p( % x % structure, except that+
#. [ b
1
[ 3 [ b
%
[ 3 % units A conse'uently [ b
1
E[ 3 [ b
%
E[ 3 #I % units.
%. the vectors for the adsorbate overlayer are rotated with respect to those of the
substrate by 1:;.
Note that the c( % x % diffraction pattern can also be obtained from the pattern for the
primitive structure by -missing out every alternate adsorbate&derived diffraction spot-. "his
is a common feature of diffraction patterns arising from centred structures.
7.& .eflection ,igh 4nergy 4lectron Biffraction
.,44B!
?ow 2nergy 2lectron Diffraction (?22D utili(es the inherent surface sensitivity
associated with low energy electrons in order to sample the surface structure. !s the
primary electron energy is increased not only does the surface specificity decrease but two
other effects are particularly noticeable
#. forward scattering becomes much more important (as opposed to the bac$ward
scattering observed in ?22D
%. the scattering angle (measured from the incident beam direction tends towards #D9
degrees for bac$&scattering and 9 degrees for forward scattering.
In order to extract surface structural information from the diffraction of high energy
electrons, therefore, the techni'ue has to be adapted and the easiest way of doing this is to
use a reflection geometry in which the electron beam is incident at a very gra(ing angle & it
is then $nown as Reflection >igh 2nergy 2lectron Diffraction (R>22D.
"he diagram above shows the basic set&up for a R>22D experiment, with the sample
viewed edge&on. In practice the display screen is usually a phosphor coating on the inside
of a vacuum window (viewport and the diffraction pattern can be viewed and recorded
from the atmospheric side of the window. "he small scattering angles involved are
compensated for by using relatively large sampleIscreen distances.
"he sample can be rotated about its normal axis so that the electron beam is incident along
specific crystallographic directions on the surface.
In order to understand the diffraction process we need to consider how the electron beam
can interact with the regular array of surface atoms in this experimental geometry. It is
worth noting, however, that the use of glancing incidence ensures that, despite the high
energy of the electrons, the component of the electron momentum perpendicular to the
surface is small. Ender these conditions an electron may travel a substantial distance
through the solid (in accord with the much longer mean free path of such high energy
electrons without penetrating far into the solid. "he techni'ue, conse'uently, remains
surface sensitive.
Now consider the plan 'ie# of a surface illustrated below in which we concentrate
attention on <ust one row of atoms (shown shaded in pale blue running in a direction
perpendicular to the incident electron beam (incident from the left
In addition to the change in momentum of the electron perpendicular to the surface, which
leads to the apparent reflection, the diffraction process may also lead to a change in
momentum parallel to the surface, which leads to the deflection by an angle when loo$ed
at in plan view. Constructive interference occurs when the path difference between ad<acent
scattered -rays- ( a sin is an integral number of wavelengths (i.e. the same basic
condition as for ?22D. "his gives rise to a set of diffracted beams at various angles on
either side of the straight through (specularly reflected beam.
"hatQ if anyQ ad'antages does .,44B offer o'er C44B $
In terms of the 'uality of the diffraction pattern absolutely none 4 & moreover, diffraction
patterns have to be observed for at least two sample alignments with respect to the incident
beam in order to determine the surface unit cell. >owever , ....
#. "he geometry of the experiment allows much better access to the sample during
observation of the diffraction pattern. "his is particularly important if it is desired to
ma$e observations of the surface structure during growth of a surface film by
evaporation from sources located normal to the sample surface or simultaneous
with other measurements (e.g. !26, T86.
%. 2xperiments have shown that it is possible to monitor the atomic layer&by&atomic
layer growth of epitaxial films by monitoring oscillations in the intensity of the
diffracted beams in the R>22D pattern.
0y using R>22D it is therefore possible to measure, and hence also to control, atomic
layer growth rates in *olecular 0eam 2pitaxy (*02 growth of electronic device
structures & this is by far and away the most important application of the techni'ue.
7.0 4xamples - Surface Structures
"his section provides wor$ed exercises in the classification of surface structures using both
.oodWs notation and matrix notation, and in the determination of surface coverages.
)ote - all surface coverages discussed in this section are defined in the conventional
surface science !anner. by reference to the nu!ber density of the underlying layer of
surface ato!s of the substrate
Structure I1
>ow would this centred structure observed on fcc(#99 surfaces be described using
.oodWs notation 5
.hat is the coverage of adsorbate for this surface structure 5
Structure I%
>ow would this structure on the fcc(#99 surface be described using .oodWs
notation 5
)ote & in this instance the structure illustrated is but one of two e'uivalent domains &
related by the symmetry of the substrate. "he second domain is shown below.
It differs from the first only in that the closely&pac$ed rows of the adsorbate now run at G9
degrees to their direction in the first domain. 8atches of this domain structure would exist
on the surface with statistically&e'ual probability to the other domain. It obviously
corresponds to the same adsorbate coverage and must possess identical properties
(electronic, thermodynamic C reactivity.
Structure I&
>ow would this adsorbate structure on the fcc(##9 surface be described using
.oodWs notation 5
Structure I0
>ow might this adsorbate structure on the fcc(### surface be described using .oodWs
notation 5
Structure I3
"his shows an alloy surface layer on an fcc(### substrate & such a layer might be formed
by two metals which alloy if one is evaporated onto a (### single crystal surface of the
other.
>ow might this superstructure be described using .oodWs notation 5
.hat are the relative concentrations of the two elements in the surface layer
( specified as the ratio of !+0 atoms 5
Structure I7
"he diagram below shows a -coadsorption- structure in which adsorption of two different
species has led to the formation of an ordered surface layer containing both species & such
structures ( albeit more complicated than the one shown are $nown for a number of pairs
of adsorbates on a variety of surfaces e.g. ben(ene I C= & Rh(###
"he driving force for the formation of such structures may be an attractive dipole&dipole
force between the two different adsorbed species (this might occur if, for example, one acts
as an electron donor whereas the other acts as an electron acceptor on the substrate
concerned.
>ow would this structure be classified using .oodWs notation 5
.hat are the relative concentrations of the two species in the coadsorption structure 5
Structure I:
"he diagram below shows a molecular adsorption structure in which a diatomic molecule is
bonded terminally to substrate atoms but is inclined to the surface normal. "his particular
structure has not been observed but similar structures involving a periodic tilting of the
molecular axis have been discovered, e.g. C= I 8d(##9 .
>ow would this structure be classified using .oodWs notation 5
Summary
"he structures of ordered adsorbate overlayers may be defined by specifying the adsorbate
unit cell in terms of the ideal substrate unit cell & in many cases, such as the examples
given, this can be done using the simple .oodWs notation and this is the common practice.
In more difficult cases it may be necessary to use matrix notation.
In all the examples studied +
#. the overlayers were simple structures in which the adsorbate unit cell dimensions
corresponded to one of the atom&atom spacings in the underlying substrate structure
A this is generally the case for adsorbed overlayers of gaseous molecules. *ore
complex -coincidence- and -incommensurate- structures are commonplace when
the overlayer consists of one or more atomic layers of a distinct chemical
compound e.g. oxide overlayers on metals.
%. it was assumed that the underlying substrate structure was undisturbed by the
adsorption of species on the surface. .hilst such distortions may generally be
small, this is far from always being the case and a number of examples involving
adsorbate&induced reconstruction of the topmost layer of substrate atoms have now
been well documented ( e.g. Ni(#99 c(%x%&C .
:. Surface Imaging 8 Bepth *rofiling
:.1 5asic concepts in surface imaging 8 localised
spectroscopy
*ost surface spectroscopic techni'ues involve probing the surface by exposing it to a flux
of -particles- (h , e
&
, !
V
.... and simultaneously monitoring the response to this
stimulation by, for example, measuring the energy distribution of emitted electrons. In their
most basic form, these techni'ues collect information from a relatively large area of surface
() mm
%
. In most cases, however, there are variations of these techni'ues which permit
either ,
#. information to be collected only from a specific region of the surface & that is ....
Cocalised Spectroscopy
or
%. the spatial distribution of a property (eg. simple topography or elemental
concentration to be mapped & that is ....
Surface Imaging Surface Microscopy!
"he re'uirement in both cases is for spatial localisation of the spectroscopic techni'ue.
"his may be achieved in one of two ways ,
A. Cocalisation of the *robe
6patial localisation of a spectroscopic techni'ue can be achieved by focusing the probe so
that it interacts with only a small region of the surface.
"he ease with which this can be accomplished and the spatial resolution (degree of
localisation attainable depends primarily upon the nature of the probe & charged particles
(e.g. electrons and ions are easily focused A neutral particles (e.g. x&ray photons and neutral
atoms are not readily focused.
.ith this localisation of the probe into a small area, it is clearly advantageous to collect as
many as possible of the outgoing particles whose detection completes the spectroscopic
process in order to maximi(e the signal&to&noise.
In the diagram above, the probe is stationary and incident upon a particular small region of
the surface in a manner suitable for localised surface spectroscopy. In order to carry out
surface imaging or mapping, either
the probe must be scanned (rastered across the surface , or
the surface must be moved under the probe beam.
5. Spatially-localiPed Betection
"he alternative approach is to allow a large area of the sample to be exposed to the probe
but only to collect signal (i.e. emitted particles from a small region of the sample. "hus, in
this instance, the spatial localisation is occurring on the detection side of the techni'ue.
In principle, this is easy & it simply re'uires an appropriate lens system to collect as much
emission as possible from the desired part of the surface and focus it onto a detector as
illustrated in a highly schematic fashion below.
In practice, this approach has only really been successfully employed in surface analysis
when the emitted particles are electrons.
In order to get an image of the surface using a single detector it is necessary to either
#. scan the sample position underneath the detection system.
%. ma$e use of additional scanning plates in the electon&optical focusing system.
!n alternative approach to obtaining images using this same basic arrangement involves
employing a linear array or two&dimensional array of detecting elements. In this type of
system, the emission from different parts of the surface is effectively focussed onto
different parts of the array detector as schematically illustrated below.
6ome additional scanning mechanism is still re'uired to build&up an image or map of
anything other than the directly imaged region of surface, but the use of multiple (array
detectors greatly enhances the speed of image ac'uisition and ultimately the image 'uality.
In summary, spatial localisation of a surface spectroscopic techni'ue may be obtained by
either focusing the probe or imaging only a small part of the sample surface onto a detector
using a lens system. Images and compositional maps are then usually obtained by scanning
the focused probe in the first case, and either translating the sample or employing
sophisticated opticalIdetector systems in the latter case.
:.% 4lectron Microscopy - S4M 8 SAM
"he two forms of electron microscopy which are commonly used to provide surface
information are
Secondary 4lectron Microscopy S4M !
& which provides a direct image of the topographical nature of the surface from all the
emitted secondary electrons
Scanning Auger Microscopy SAM !
& which provides compositional maps of a surface by forming an image from the !uger
electrons emitted by a particular element.
0oth techni'ues employ focusing of the probe beam (a beam of high energy electrons,
typically #9 & :9 $eS in energy to obtain spatial localisation.
A. Secondary 4lectron Microscopy S4M !
!s the primary electron beam is scanned across the surface, electrons of a wide range of
energies will be emitted from the surface in the region where the beam is incident. "hese
electrons will include bac$scattered primary electrons and !uger electrons, but the vast
ma<ority will be secondary electrons formed in multiple inelastic scattering processes
(these are the electrons that contribute to the bac$ground and are completely ignored in
!uger spectroscopy. "he secondary electron current reaching the detector is recorded and
the microscope image consists of a -plot- of this current, I , against probe position on the
surface. "he contrast in the micrograph arises from several mechanisms, but first and
foremost from variations in the surface topography. Conse'uently, the secondary electron
micrograph is virtually a direct image of the real surface structure.
"he attainable resolution of the techni'ue is limited by the minimum spot si(e that can be
obtained with the incident electron beam, and ultimately by the scattering of this beam as it
interacts with the substrate. .ith modern instruments, a resolution of better than : nm is
achievable. "his is more than ade'uate for imaging semiconductor device structures, for
example, but insufficient to enable many supported metal catalysts to be studied in any
detail.
5. Scanning Auger Microscopy SAM !
"he incident primary electrons cause ioni(ation of atoms within the region illuminated by
the focused beam. 6ubse'uent relaxation of the ioni(ed atoms leads to the emission of
!uger electrons characteristic of the elements present in this part of the sample surface (see
the description of !uger spectroscopy in 6ection :.% for more details.
!s with 62* , the attainable resolution is again ultimately limited by the incident beam
characteristics. *ore significantly, however, the resolution is also limited by the need to
ac'uire sufficient !uger signal to form a respectable image within a reasonable time
period, and for this reason the instrumental resolution achievable is rarely better than about
#:&%9 nm.
,or a description of a modern combined 62*I6!* instrument (a SR *ICR=?!0 /:9,
and examples of some real images obtained with such an instrument, clic$ on the picture
(courtesy of "hermo SR 6cientific

:.& Imaging N*S
"he combination of the features of T&ray photoelectron spectroscopy (in particular,
'uantitative surface elemental analysis and chemical state information & see :./ together
with spatial localisation is a particularly desirable option in surface analysis. >owever,
whilst much progress has been made in developing the techni'ue of imaging T86, there is
still a considerable research effort being devoted to improving the available spatial
resolution beyond that which is currently available.
Different manufacturers of imaging T86 systems have adopted different strategies for
obtaining spatial localisation & including all of those mentioned in 6ection H.# .
6pecificially, these include
#. ?ocalisation of the probe, by focusing the incident x&rays.
%. ?ocalisation by selected (limited area analysis.
/. Ese of array detectors, with associated imaging optics.
1. Imaging N*S by N-ray 1ocussing
Entil very recently this was not a popular option in commercial instruments, since x&rays
are rather difficult to focus (in fact x&ray lens design is a science in its own right and
alternative focusing techni'ue, usually employing specially&bent diffracting crystals, have
to be employed.
Nevertheless, with recent improvements in technology, an x&ray spot si(e of better than #9
m has been achieved using this approachA images may then be obtained either by scanning
the sample under the focused x&ray beam, or by scanning the microfocused x&ray beam.
,or a description of a modern scanning 26C! microprobe (the Zuantum %999 which
employs x&ray focusing clic$ on the picture
(courtesy of 8hysical 2lectronics, Inc. (8>I

%. Cimited Area Analysis
"he most common approach to localisation of analysis, is to restrict the area of the sample
surface from which photoelectrons are analysed using a combination of lenses and
apertures in the design of the electron energy analyser.
"he main problem with this approach is that as the sampled area is reduced, so is the
collected signal & conse'uently, there is a direct trade&off between spatial resolution and
data collection time.
"he practically achievable spatial resolution is rarely better than #99 m.
Imaging of the sample surface may then be achieved by either
#. translating the sample position under the electron energy analyser, so that the
analysed region is moved across the surface.
%. incorporating electrostatic deflection plates within the electron optics to move the
region from which electrons are collected across the sample surface.
=ne advantage of this techni'ue is its relative simplicity (and hence relatively low cost &
but this is reflected in the poor performance 4
&. Array Betectors and Imaging <ptics
"here are a number of innovative designs for imaging spectrometers which combine array
(i.e. multiple&segment detectors with sophisticated imaging optics to obtain electron&
energy resolved images at much faster ac'uisition rates. "he spatial resolution achievable
using this approach is also higher than that for either of the other techni'ues mentioned &
state&of&the&art instruments give better than : m resolution
,or a description of a state&of&the&art imaging T86 instrument (the 26C!?!0 %:9 based
on this type of technology clic$ on the picture
(courtesy of "hermo SR 6cientific

State of the Art
"his is one area in which an ultra&high intensity x&ray source offers ma<or advantages and
the highest performance imaging T86 systems (scanning photoemission microscopes are
therefore those based at synchrotron sources. Esing (one&plate technology to focus the x&
rays (combined with multi&segment detectors to enhance data ac'uisition rates it is
possible on such systems to obtain T86 images with a resolution of better than #99 nm
(see, for example, the 2?2""R! 26C! microscopy web&pages.
In summary, the spatial resolution currently achievable with commercial imaging T86
instruments limits their potential applications & nevertheless there are many areas of
materials science where the information obtainable is incredibly useful and the relatively
poor spatial resolution (compared, for example, with electron microscopic techni'ues such
as 6!* is more than offset by the benefit of concurrent chemical state definition.
:.0 SIMS - Imaging 8 Bepth *rofiling
"he basic ideas behind the 6I*6 (6econdary Ion *ass 6pectrometry techni'ue have
already been discussed in 6ection :.: .
6ince the techni'ue utilises a beam of atomic ions (i.e. charged particles as the probe, it is
a relatively easy matter to focus the incident beam and then to scan it across the surface to
give an imaging techni'ue.
A. Surface Imaging using SIMS
If the aim of the measurement is to obtain compositional images of the surface formed
from the secondary ion spectrum with minimum possible damage to the surface, then the
main problem is to ensure that sufficient signal is obtained at the desired spatial resolution
whilst minimi(ing the ion flux incident on any part of the surface.
"his is most easily achieved by switching from the traditional instrumental approach of
using continuous&flux ion guns and 'uadrupole mass spectrometer detectors, to using
pulsed ion sources and time&of&flight ("=, mass spectrometers. "he "=, mass
spectrometers are a much more efficient way of ac'uiring spectral data, and also provide
good resolution and sensitivity up to very high masses. Esing such instruments, 6I*6
images with a spatial resolution of better than :9 nm are obtainable.
5 SIMS Bepth *rofiling
"he aim of depth profiling is to obtain information on the variation of composition with
depth below the initial surface & such information is obviously particularly useful for the
analysis of layered structures such as those produced in the semiconductor industry.
6ince the 6I*6 techni'ue itself relies upon the removal of atoms from the surface, it is by
its very nature a destructive techni'ue, but also, ideally suited for depth profiling
applications. "hus a depth profile of a sample may be obtained simply by recording
se'uential 6I*6 spectra as the surface is gradually eroded away by the incident ion beam
probe. ! plot of the intensity of a given mass signal as a function of time, is a direct
reflection of the variation of its abundanceIconcentration with depth below the surface.
=ne of the main advantages that 6I*6 offers over other depth profiling techni'ues (e.g.
!uger depth profiling is its sensitivity to very low concentrations ( ) ppm of elements &
again this is particularly important in the semiconductor industry where dopants are often
present at very low concentrations.
"he depth resolution achievable (e.g. the ability to discriminate between atoms in ad<acent
thin layers is dependent upon a number of factors which include +
#. the uniformity of etching by the incident ion beam
%. the absolute depth below the original surface to which etching has already been
carried out
/. the nature of the ion beam utili(ed (i.e. the mass C energy of the ions
as well as effects related to the physics of the sputtering process itself (e.g. ion&impact
induced burial.
:.3 Auger Bepth *rofiling
!s described in 6ection :.% , !uger 6pectroscopy is a surface sensitive spectroscopic
techni'ue yielding compositional information.
In its basic form it provides compositional information on a relatively large area ( ) # mm
%
of surface using a broad&focused electron beam probe. In this manner, sufficient signal
can be readily obtained whilst $eeping the incident electron flux low, and thus avoiding
potential electron&induced modifications of the surface. !s a conse'uence the techni'ue is
non&destructive when used in this manner.
"o obtain information about the variation of composition with depth below the surface of a
sample, it is necessary to gradually remove material from the surface region being
analysed, whilst continuing to monitor and record the !uger spectra.
"his controlled surface etching of the analysed region can be accomplished by
simultaneously exposing the surface to an ion flux which leads to sputtering (i.e. removal
of the surface atoms.
,or example, suppose there is a buried layer of a different composition several nanometres
below the sample surface.
!s the ion beam etches away material from the surface, the !uger signals corresponding to
the elements present in this layer will rise and then decrease again.
"he diagram above shows the variation of the !uger signal intensity one might expect from
such a system for an element that is only present in the buried layer and not in the rest of
the solid.
In summary, by collecting !uger spectra as the sample is simultaneously sub<ected to
etching by ion bombardment, it is possible to obtain information on the variation of
composition with depth below the surface & this techni'ue is $nown by the name of Auger
,epth rofiling
:.7 Scanning *robe Microscopy STM ? A1M !
In the early #GD9Ws two I0* scientists, 0innig C Rohrer, developed a new techni'ue for
studying surface structure & Scanning Tunnelling Microscopy ( 6"* . "his invention
was 'uic$ly followed by the development of a whole family of related techni'ues which,
together with 6"*, may be classified in the general category of 6canning 8robe
*icroscopy ( 68* techni'ues. =f these later techni'ues, the most important is Atomic
1orce Microscopy ( !,* .
"he development of these techni'ues has without doubt been the most important event in
the surface science field in recent times, and opened up many new areas of science and
engineering at the atomic and molecular level.
5asic *rinciples of S*M TechniDues
!ll of the techni'ues are based upon scanning a probe (typically called the tip in 6"* ,
since it literally is a sharp metallic tip <ust above a surface whilst monitoring some
interaction between the probe and the surface.
"he interaction that is monitored in +
STM & is the tunnelling current between a metallic tip and a conducting substrate which
are in very close proximity but not actually in physical contact.
A1M & is the van der .aals force between the tip and the surfaceA this may be either the
short range repulsive force (in contact&mode or the longer range attractive force (in non&
contact mode.
,or the techni'ues to provide information on the surface structure at the atomic level
(which is what they are capable of doing 4 +
#. the position of the tip with respect to the surface must be very accurately controlled
(to within about 9.# ] by moving either the surface or the tip.
%. the tip must be very sharp & ideally terminating in <ust a single atom at its closest
point of approach to the surface.
"he attention paid to the first problem and the engineering solution to it is the difference
between a good microscope and a not so good microscope & it need not worry us here,
sufficient to say that it is possible to accurately control the relative positions of tip and
surface by ensuring good vibrational isolation of the microscope and using sensitive
pie(oelectric positioning devices.
"ip preparation is a science in itself & having said that, it is largely serendipity which
ensures that one atom on the tip is closer to the surface than all others.
?et us loo$ at the region where the tip approaches the surface in greater detail ....
... the end of the tip will almost invariably show a certain amount of structure, with a
variety of crystal facets exposed ...
7 and if we now go down to the atomic scale ....
... there is a reasonable probability of ending up with a truly atomic tip.
If the tip is biased with respect to the surface by the application of a voltage between them
then electrons can tunnel between the two, provided the separation of the tip and surface is
sufficiently small & this gives rise to a tunnelling current.
"he direction of current flow is determined by the polarity of the bias.
If the sample is biased &ve with respect to the tip, then electrons will flow from the surface
to the tip as shown above, whilst if the sample is biased Vve with respect to the tip, then
electrons will flow from the tip to the surface as shown below.
"he name of the techni'ue arises from the 'uantum mechanical tunnelling&type mechanism
by which the electrons can move between the tip and substrate. Zuantum mechanical
tunnelling permits particles to tunnel through a potential barrier which they could not
surmount according to the classical laws of physics & in this case electrons are able to
traverse the classically&forbidden region between the two solids as illustrated schematically
on the energy diagram below.
"his is an over&simplistic model of the tunnelling that occurs in 6"* but it is a useful
starting point for understanding how the techni'ue wor$s.
In this model, the probability of tunnelling is exponentially&dependent upon the distance of
separation between the tip and surface + the tunnelling current is therefore a very sensitive
probe of this separation.
Imaging of the surface topology may then be carried out in one of two ways+
#. in constant height mode (in which the tunnelling current is monitored as the tip is
scanned parallel to the surface
%. in constant current mode (in which the tunnelling current is maintained constant as
the tip is scanned across the surface
If the tip is scanned at what is nominally a constant
height above the surface, then there is actually a periodic
variation in the separation distance between the tip and
surface atoms. !t one point the tip will be directly above
a surface atom and the tunnelling current will be large
whilst at other points the tip will be above hollow sites
on the surface and the tunnelling current will be much
smaller.
! plot of the tunnelling current vWs tip position therefore shows a periodic variation which
matches that of the surface structure & hence it provides a direct -image- of the surface (and
by the time the data has been processed it may even loo$ li$e a real picture of the
surface 4 .
In practice, however, the normal way of imaging the
surface is to maintain the tunnelling current constant
whilst the tip is scanned across the surface. "his is
achieved by ad<usting the tipWs height above the surface
so that the tunnelling current does not vary with the
lateral tip position. In this mode the tip will move slightly
upwards as it passes over a surface atom, and conversely,
slightly in towards the surface as it passes over a hollow.
"he image is then formed by plotting the tip height (strictly, the voltage applied to the (&
pie(o vWs the lateral tip position.
4xamples of STM Images
,or a more comprehensive set of examples you should try one of the following sites +
"E Sienna 6"* RroupWs 6ite Images from 8eter SargaWs group at Sienna, including some
excellent atomic resolution images of metallic surfaces.

"he I0* 6"* Image Rallery *ore for fun than anything else 4 & some of the best images
ever recorded, including classic demonstrations of 'uantum
confinement and atom manipulation, but all disguised with a
bit of an artistic touch.

=micron Instruments &
!pplications Rallery
! collection of 68* images, including some exceptionally
well&resolved 6"* images.

Seeco Instruments &
Nano"heatre
! good collection of images illustrating the wide range of
applications of scanning probe microscopy techni'ues.

Summary
In summary, the development of the various scanning probe microscopy techni'ues has
revolutioni(ed the study of surface structure & atomic resolution images have been obtained
not only on single crystal substrates in E>S but also on samples at atmospheric pressure
and even under solution. *any problems still remain, however, and the interpretation of
68* data is not always as straightforward as it might at first seem. "here is still very much
a place for the more traditional surface structural techni'ues such as ?22D.
"his introduction to 6"* has concentrated on the non&invasive imaging applications of the
techni'ue, yet there is increasing interest in using such techni'ues as a tool for the actual
modification of surfaces. !t the moment this is still at the -gimmic$y- stage, but the longer
term implications of being able to manipulate surface structure and molecules at the atomic
level have yet to be fully appreciated + we can but await the future with interest 4
(urther "eading 5 for a !ore !athe!atical treat!ent of STM loo# at -6canning "unneling
*icroscopy + ! "utorial- fro! the ,epart!ent of Che!istry ? 2ioche!istry at the
@niversity of Auelph0

An Introduction To Surface Chemistry

Transféré par

Informations du document

Description originale:

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

An Introduction To Surface Chemistry

Transféré par

Droits d'auteur :

Formats disponibles

An Introduction to Surface

*ean ,ree 8ath

( h 3 #%:/.@ eS radiation ...

Vous aimerez peut-être aussi