AFL Aston Catalysis Surface Chemistry Slides

Prof Adam Lee [email protected]

8 lectures

Air

CH3010: Catalysis

1

During this course you should become familiar with:

• Experimental reaction kinetics – liquid and gas phase reactions • The importance of surface processes. • How surface structure affects the reactivity of materials. • Adsorption processes, physisorption and chemisorption • Adsorption isotherms - Langmuir and BET models. • Kinetics of surface reactions - Eley Rideal and Langmuir Hinshelwood mechanisms. • Surface analytical techniques, including XPS, AES.

Recommended Reading

• G.A. Somorjai Introduction to Surface Chemistry (Wiley)

• G.Attard, C.Barnes Surfaces (Oxford Chemistry Primer)

• E.McCash Surface Chemistry (OUP)

• P.W Atkins & J de Paula Physical Chemistry, 9th Edition Chapter 21 • “ “ Elements of Phys. Chem., 5th Edition Chapter 1

• Tutorial notes on surface chemistry, useful www.chem.qmw.ac.uk/surfaces

2

Learning objectives

3

Why do we care about kinetics?How fast a process occurs? - can predict practical/economic feasibility - can predict risks (heat release, pressure changes) - alter reactants - improve reactor engineering

What route is followed? - what steps are involved (e.g. dissociation, coupling) - what is the mechanism (how do atoms/molecules interact)

Typical experiments involve: 1. mixing reactants 2. initiating reaction (heat , light )

3. measuring T or P 4. measuring reactant/product concentrations

Crucial that Steps 1-3 >> Step 4

4

Experimental kineticsKinetic theory – reaction rates:

nAkdtAd ][][

−=

Experimental kinetics – concentrations of reactants

Aim of experiment to determine:

• Reaction order n - steps involved • Rate constant(s) k - intrinsic ‘speed’ of reaction, bigger = faster • Activation energy Ea - mechanism

Theory and experiment linked via the integrated rate equations

For reactions that occur over minutes or hours, k < 1x10-2 s-1

This measurement timescale is fine for standard undergraduate techniques:

e.g. pH meter , conductivity , photometry

Key considerations in choosing a method are the half-life t1/2 and mixing time.

5

Case study: 1

6

Case study: 1Consider thermal decomposition of cyclopentene:

Since pV = nRT (i.e. p ∝ n), and 1 mol of any gas occupies same V at same T

Measuring p versus time t allows us to track reaction progress

T = 500oC – 540 oC t1/4 ~1000 sec

Reaction vessel in furnace

Manometer

Cyclopentene

Vacuum pump

Transducer

7

T = 500oC – 540 oC t1/4 ~1000 sec

8

Case study: 1Consider thermal decomposition of cyclopentene:

Since pV = nRT (i.e. p ∝ n), and 1 mol of any gas occupies same V at same T

Measuring p versus time t allows us to track reaction progress

If all cyclopentene decomposes, pressure ↑ x2.

Partial pressure of CP at any time PCP = 2 PCP(t=0) – Pt.

Analysis

• Determination of order, n

1. Integrated plots

Plot ln(PCP) vs t and 1/PCP vs t and compare trend lines via R2 values.

Best fit (R2 closest to 1) predicts order.

BUT

Works best for synthetic data; “real” data difficult to distinguish.

9

Ln(CP)

ln  (C

P)

3.9

4.125

4.35

4.575

4.8

Time  /  min

0 5 10 15 20

R²  =  0.9752

1/CP

1  /  CP

0

0.0045

0.009

0.0135

0.018

Time  /min

0 5 10 15 20

R²  =  0.9536

10

2. Quarter lives method

1st order reactions t1/4 independent of PCP(t=0)

2nd order reactions t1/4 ∝ 1/PCP(t=0)

∴examine t1/4 at 3 different values of PCP(t=0) (at fixed T) → PCP(t=0) dependence

Can generalise this method to any value of n.

3. Initial rates method

ln(-Rt=0) = ln(k) + n ln(PCP(t=0))

∴plotting ln(-Ro) vs ln(PCP(t=0)) should be linear with slope = n.

Initial rates Ro can be obtained from the slope of the tangent at t=0 or more accurately by fitting a cubic polynomial to the data.

Need to repeat reax. at different T

• Determination of rate constant

Once we have determined n use appropriate integrated plot to determine k

• Determination of EActivation

Arrhenius plot of ln(k) vs 1/T slope = -EA/R

Ln(CP)

ln  (C

P)

4

4.2

4.4

4.6

4.8

Time  /  min

0 5 10 15 20

Need to repeat reax. at different T11

For reactions with t1/2 ~ 1 msec – 1 sec we usually use a flow method

• Continuous (discharge) flow

Very wasteful of reactants and needs moveable detector

12

A

B Reaction regionMixing region

Product (waste)

12

Case study: 2

Moveable detector

• Stopped flow method

Reactant syringes

Mixing chamber

Light sourcePhotomultiplier

detector

Stopping syringe

13

14

Uses much smaller amounts of reactants and employs a fixed detector

e.g. reduction of 2,2 dichlorophenolindophenol (DCIP) by ascorbic acid

DCIP → products

Can use optical absorption detection method: [DCIP ] ∝ absorbance

Faster reactions (t1/2 < 1 msec) – relaxation methods

For example systems at equilibrium

Monitor time to re-equilibrium system (e.g. [B] or [A]) after perturbation

t* = 1/(k1+k-1) and K = k1/k-1

The perturbation should occur in 10-6 – 10-7 seconds and may b e:

- pressure jump - temperature jump - high electric field pulse - ultrasonic vibrations - light pulse (flash photolysis) ultrafast spectroscopy (fs-ps!)

k1

A ⇔ B k-1

15

• Flash Photolysis – very fast reactions (k~50 000 – 500 000 s-1)

Pre-mixed volume of reactants in a photolysis cell subjected to light pulse producing atoms, radicals & excited states whose concentration is followed with time.

Advantages: - no mixing time

- timescales of reax. limited only by

pulse duration (<10-12 s with lasers)

- species formed in the centre of the cell

so can ignore wall reactions.

16

17

Reaction dynamicsAll previous methods consider systems where reactants have a thermal distribution of energy.

To study elementary reactions ideally want molecules possessing a specific energy, and only undergoing a single collision.

Solution: Molecular beams

Collimation+rapid expansion make beam: 1. intense 2. linear 3. minimal vibrational excitation

Energy can be tuned: 1. varying T 2. seeding beam with inert elements (e.g. He)

NH3 synthesis CH3OH synthesis

C2H4 oxidation

Surface Chemistry& Analysis

Heterogeneous Catalysis

Heterogeneous Catalysis

Chemical Surface

Modification

Chemical Surface

Modification

Corrosion Science

Medical Implants

Polymer ScienceTribology

OpticsCorrosion Science

Medical Implants

Polymer ScienceTribology

Optics

35% of global GNP

Electronic Devices

Gas sensors (CO)

18

Surface chemical processes$10 trillion!!

Gas phase: Controlling parameters: [A], collision rate zAA, collision cross-section σA ..

e.g. O2 P = 1 atm; T = 298 K → zAA = 2.8 x 109 molecule-1 sec-1

Surfaces: Rate of gas molecules colliding with surface (flux) = n x c

;

∴ Rate = P/(2πmkT)1/2 Hertz-Knudsen equation

e.g. O2 P = 1 atm; T = 298K → rate = 5 x 1023 cm-2 sec-1

Surface chemistry phenomenally fast due to higher collision probability → vast collisional cross-section 19

Surface Kinetics

n = N / V = P / (kT)

P,V,T

20

• Substrate (adsorbent) - the solid surface where adsorption occurs

• Adsorbate - the atomic/molecular species adsorbed on the substrate

Surface Terminology

What happens when a molecule collides with a surface?

Nothing - elastic collision, no energy transfer

Atoms/molecules feel attractive potential and “stick” to surface - a process termed ADSORPTION

The sticking probability s is defined as:

s = no.of molecules that stick 0 ≤ s ≤ 1 no. of molecular impacts

21

22

Distance from surface

adsorbate

δ- δ +δ+

Arises from van der Waals forces between a molecule/atom and surface

Consider changes in thermodynamic parameters upon adsorption:

S2 < S1 , ∴ ΔS < 0

Recall, ΔG = ΔH - TΔS Gibbs-Helmholtz equation

Since ΔG <0 for a process to occur spontaneously,

ΔH <0 , i.e adsorption must always be exothermic

3 degrees of freedom S1

2 degrees of freedom S2

AdsorptionSurface

23

Thermodynamics of adsorption

+ve-ve

24

For dissociative adsorption of molecules, e.g. X2, we need to consider energy: 1. breaking the X-X bond - dissociation energy DXX

2. making 2 new X-M (M=metal atom) bonds - bond strength EX-M

For the adsorption to be exothermic:

EX-M > DXX/2.

For N2, EN-M >420 kJ mol-1, H2, EH-M >200 kJ mol-1

∴ N2 adsorption is more specific.

EX-M high enough

Molecular adsorption Dissociation

Metal Metal Metal

Chemisorption • ΔH > 50 kJ mol-1 • chemical adsorption → chemical bond formation • Confined to single monolayer • Adsorption may be activated • Dissociation possible

Physisorption • ΔH < 50 kJ mol-1 • physical adsorption → van der Waals (perm./induced dipoles) • Multilayer adsorption • Unactivated adsorption • Enhanced at low T (Le Chatelier) • Non-dissociative

diffusion “2-D gas”

desorption

A2(g) ⇔ A2(a)

simple equilibrium between gas & adsorbed

Surface lifetime, τ = τo e ΔH/RT Frenkel eqn.

EX-M high enough  

A2(g) ⇒ 2 A(a)

adsorbed atoms can still diffuse across surface 25

Adsorption types

Chemisorption

• A true chemical bond forms between the adsorbate and the surface - involves electron transfer.

• Bonds within the adsorbate molecule are weakened. ⇒ fundamental to catalysis!

• Chemisorption process may be associative or dissociative.

e- transfer

26

ΔHads(kJ/mol) Dissociative ads: O2/W -600 N2/W -340 H2/W -160 H2/Ni - 80

Non-Dissociative CO/Ni -130 N2/Ni - 50

Bond dissociation energy of N2 = -950 kJmol-1 (Gas Phase) H2 = -430 kJmol-1

• ΔHads for dissociative adsorption > non dissociative

27

Consider a linear combination of 1s atomic orbitals when going from a diatomic molecule to an infinite solid

+

• Infinite Solid

[ ]n

[ ]n

[ ]n

Electronic Properties of Solids

28

In the solid state bands exist rather than discreet molecular orbitals. → Each band is a continuum of MO’s

Consider band structure for Sodium

1s2

2s2

2p6

3s1

1s Band

2s Band

2p Band

3s BandFermi Energy (EF)

Filled

Filled

Filled

½ full

Atom → extended structure

Band formation

29

Fermi energy

Recall from MO theory

• N atomic orbitals ⇒ N molecular orbitals levels in each band

• Each level has space for two electrons.

• The Fermi level is the highest occupied energy level in band structure (equivalent to the HOMO in a molecule).

• N.B. For a material to be a conductor, the Fermi level must lie in a region with unoccupied states above it.

EF

Insulator Conductor30

Density of States

1. N atomic orbitals ⇒ N levels in each band

2. Band width increases with orbital overlap

Recall:

• Distribn of energy levels within a band depends on band width

Strong overlap → Broad band → Low Density of States

• Density of States = No. energy levels per unit energy in a band

Weak overlap → Narrow band → High Density of States

31

d-orbitals ⇒ poor overlap ⇒ narrow band ⇒ high Density of states

s-orbitals ⇒ good overlap ⇒ wide band ⇒ low Density of states

Filled d states

sp band

sp band

Filled d states

Empty d states

EF

Partially filled d-band metalN(E)

N(E)

EF

Den

sity

of

Sta

tes

Den

sity

of

Sta

tes

Filled d-band metal

EF

32

What is the significance of the density of states?

• Chemisorption involves electron transfer to/from the surface

• Electrons can only be transferred to/from states near the fermi level (Ef).

• Direction of e- transfer can be:

Filled orbital on adsorbate → empty state on metal Filled states on metal → empty orbital on adsorbate

• High DOS at Ef → metal has lots of states available for electron transfer

• Transition metals (d block) are good catalysts

How do we determine the direction of e- transfer?

Recap:

• Combination of atomic orbitals in a extended solid result ⇒ ‘band’ formation

FERMI ENERGY (Ef) = Highest occupied level

DENSITY OF STATES (DoS) = Number of available energy levels at a particular energy.

• DoS at Ef tells us how good/bad solid will be at electron transfer (e.g.important for electronic interaction with adsorbates -see later)

Definition The difference in energy between an electron at rest just outside the metal surface and an electron at the Fermi Energy.

• Electrons are held in a potential well

• φ depends on the depth (W) of the potential well in which the electrons are held, and the number of electrons.

E > 0

E < 0W

EF

φEmpty levels Fermi Level

Bottom of valence band

E(φ)=W-EF

Zero Energy Level

Occupied States

The Workfunction (φ)

Inte

nsity

0.5

1/λ

Threshold = φ

• The work function (φ) is measured as the threshold photon energy required for photoemission

• φ is a measure of surface electronic properties and changes in the presence of adsorbates.

• φ is typically 1-5 eV

36

Work function changes induced by Chemisorption

• Recall chemisorption involves electron transfer

a. Electronegative adsorbates

e-EF

φ

If the LUMO of the adsorbate < EF then charge is transferred to the adsorbate.

Metal Adsorbate

LUMO

1. Ef ↓ hence φ ↑ i.e Δφ is +ve

ΔφEF

37

e-

EF

φ

b. Electropositive adsorbates

ΔφEF

• If the HOMO of the adsorbate >EF then charge is transferred to the metal.

HOMO

1. Ef ↑ hence φ ↓ i.e Δφ is -ve

38

Examples of work function changes

O/Cu(111) - Positive H/Mo(111) - Negative

Cs/W(100) - Negative

Cl/W(100) - Positive

39

Recap

WORK FUNCTION (Φ) = minimum energy for an electron to be emitted from the Fermi Level (Ef)

• Φ changes in the presence of adsorbates

- Chemisorption

- electronegative adsorbate → ΔΦ ALWAYS +ve (e.g. Cl)

- electropositive adsorbate → ΔΦ ALWAYS –ve (e.g. K)

40

Case Study CO/Pt(111)

Δφ varies with fractional coverage (θ)

Need to consider bonding of CO to surface

41

• Molecular Orbital Diagram for CO (with sp mixing):

• Electron configuration: (1σ)2 (2σ)2 (3σ)2 (4σ)2 (1π)4 (5σ)2 (2π)0

3σ

4σ

5σ

6σ

42

The 4σ orbital is localised on O

The 5σ orbital on the C atom

The 2π is anti-bonding

5σ

1πO

C 2π

4σ

So what happens at 0.33 ML?

43

Low coverages

Initial adsorption occurs into preferred on-top site

σ donation to metal dominates, hence φ decreases.

OC

• But CO-CO repulsion prevents occupation of every on-top site.

Vacant Metal d orbital

5σ orbital of CO

O C

Strong σ donor

Full metal d orbital

2π orbital of CO

π acceptor

44

Higher coverages

Occupy less favourable bridge sites.

Charge transfer now dominated by π* back-bonding

⇒ φ increases and returns to that of the clean surface

OCOC

O C

Weak σ donor π acceptor

45

ΔH

ads(

kJ/

Mol

)

- 145

- 96.4

- 48.2

Coverage (monolayers)

Infra-red measurements show:

< 0.33 ML → single band at ~2090 cm-1 → linear Pt-C=O species

> 0.33 ML → second band ~1850 cm-1 → bridging species

• The enthalpy of adsorption (ΔHads) is also coverage dependent

>0.33<0.33

• Is there evidence for such adsorption site changes?

0.33

46

Strong

Weak

Beginning of 20th century, use of nitrogenous fertilisers well established

Principal source NaNO3 (Chile)

Massive demand due to: • population growth • use as explosive

BUT

Sources predicted not to last more than 50 years

Other sources of ‘fixed’ nitrogen required! → ammonia

Process feeds roughly 33 % of world

Dissociative adsorption – NH3 synthesis

47

Haber-Bosch process (1909) commercialised by BASF

Alvin Mittasch tested >4000 catalytic materials in 1000 experiments to achieve direct combination of H2 and N2

1. Adsorption and dissociation of the reactants

N2(g) ⇒ N(a) + N(a) H2(g) ⇒ H(a) + H(a)

2. Diffusion and reaction of N and H atoms, and desorption of ammonia.

N(a) ⇒ NH(a) ⇒ NH2(a) ⇒ NH3(a) ⇒ NH3(g)

The catalyst must ‘get the balance right’:

• Adsorb N2 and H2 dissociatively • Allow N and H diffusion and reaction • Allow ammonia desorption

48

Fritz Haber Carl Bosch (Nobel Prize, Chemistry

1918/1932)

Potential energy diagram for dissociative adsorption

-ΔHchem

Transition State

H

-ΔHphysisorption

H + H

H2

Molecular state

Echem

Distance from Surface (nm)Ener

gy /

eV

Dissociated stateFe

Fe

HH

Gas phase enthalpy of dissociation

49

• Non-activated dissociative chemisorption

• Activated dissociative chemisorption

Crossover > E = 0

Crossover E < 0

- spontaneous dissociation

- need to put energy in to system (heat) to dissociate

e.g. H2 on Fe

e.g. CO on Ta

50

Molecular vs dissociative adsorption of CO

The catalytic activity of different metal surfaces reflects their interaction with

adsorbates.

Of particular importance is the role of the Fermi level.

Dissociated CO Molecular CO

51

The position of EF relative to LUMO of the adsorbate affects electron

transfer process (i.e. degree of back-bonding and hence dissociation)

Carbon Monoxide

Ti V Cr Mn Fe Co Ni

• Same principle applies to dissociation of other diatomics e.g. O2, N2, H2

• Back-bonding (and hence dissociation) can be tuned by adding dopants

e- transfer

• If Ef > LUMO → strong back-bonding → weakens C-O bond → Dissociation

52

P1 V1 = P2 (V1 + V2) - Kinetic theory of gases

P2’,V1,T

P1,V1,T

0,V2,T

P2,V1,TP2,V2,T

P2’,V2,T

P1,V1,T

0,V2,T

X

X

Open valve

Open valve

53

How do we measure adsorption?

P2’ < P2 , due to adsorption of gas molecules onto the surface

By measuring P2 and P2’ can calculate number of molecules adsorbed ∝ [P2 - P2

’] - Volumetric method

Plot of number of molecules adsorbed against equilibrium pressure at constant temperature

A useful quantity is the monolayer uptake - the number of molecules which just completely cover the surface.

This can be used to calculate the surface area of a solid

x

x

x

x

x

xx

Pressure (=P2’)

Num

ber

m

olec

ules

ad

sor

bed,

n

A2(g) ⇔ A2(a)

[A2(a)] = f (p)

Often use V (volume equivalent at stp) or θ (surface coverage) instead of n

54

Adsorption isotherms

Surface

Monolayer

55

By making certain assumptions Langmuir derived a simple equation to describe the adsorption isotherm of any atom/molecule.

Enables us to predict adsorbate coverage (θ) ➔ calculate reaction rates ➔ optimise reaction conditions (T, pressure)

Based on idea chemical equilibria exist during all reactions:

- stabilities of adsorbate vs. gas/liquid - temperature (surface and reaction media) - pressure (liquid conc.) above catalyst0

GAS/LIQUID reactants, products

solvents

CATALYST absorbate

Langmuir adsorption isotherm

56

Equilibrium between gas molecules M, empty surface sites S & adsorbates

e.g. for non-dissociative adsorption

S* + M S----M

Assumption 1: Fixed number of identical, localised surface sites

[S----M] ∝ θ adsorbate coverage

[S*] ∝ vacancies ∝ (1- θ) [M] ∝ gas pressure

∝ PReactants Products

Assumption 2: Adsorption is immobile – no surface diffusion

Assumption 3: Each adsorption site is occupied by only 1 adsorbate (only monolayer adsorption)

57

Equilibrium constant, b is

P)1(]tstanac[Re]oducts[Prb

θ−θ

==

Rearrange in terms of θ,

)bP1(bP+

=θ Langmuir Adsorption Isotherm

b is Langmuir equilibrium constant (∝ sticking-probability s ) - depends on ΔHads

Assumption 4: ΔHads and thus b is temperature & pressure independent

b

58

Rate of adsorption ↑ as b ↑

b=0.1

b=1

b=10

Surface saturated/ poisoned

Tune chemistry

Self-Assembled Monolayers

MoS2Au/(111) hydrodesulfurisation

catalyst

θ is the fractional surface coverage = n/nmonolayer

b is a constant for specific gas-solid equilibrium.

Rerrange Langmuir isotherm to relate experimental measureables, P and n

P/n = 1/bnm + (1/nmonolayer) P

Plot of P/n vs P should be linear with slope 1/nmonolayer

From nmonolayer, and the cross-sectional area of the adsorbate → calculate surface area of solid.

(often use volume of gas adsorbed at STP instead of number molecules)59

Applying Langmuir adsorption isotherm

)bP1(bP+

=θ

[C2H5Cl] Mass [C2H5Cl]/mass(mol/dm13) adsorbed

(g)0 0

1.17E103 3 3.90E1042.94E103 3.8 7.74E1045.87E103 4.3 1.37E1031.17E102 4.7 2.49E1031.76E102 4.8 3.67E103

Langmuir analysis of chloroethane adsorption on charcoal

[C2H

5Cl]

0

1.3

2.5

3.8

5

Mass  adsorbed  m  (g)0 0.0045 0.009 0.0135 0.018

[C2H

5Cl/]m

0.00E+00

1.00E-­‐03

2.00E-­‐03

3.00E-­‐03

4.00E-­‐03

[C2H5Cl]0.00E+00 4.50E-­‐03 9.00E-­‐03 1.35E-­‐02 1.80E-­‐02

y  =  0.1982x  +  0.0002

≡P for gases≡V for gases

60

Mass of monolayer = 1/slope = 1/0.198 = 5.05 g

Convert → moles → molecules

Surface area = molecules x area molecule

Langmuir analysis involving two measurementsA carbon sample adsorbs 25 cm3 of nitrogen at a pressure of 10 mbar and 41 cm3 at a pressure of 18 mbar. Making use of the Langmuir adsorption isotherm, determine the monolayer capacity of the sample. [R=8.314 J K-1 mol-1]

θ = 25/Vm at P=10 mbar , where Vm is saturation (monolayer) uptake. θ = 41/Vm at P=18 mbar

and θ = KP/(1+KP)

Thus 25/Vm = 10K/(1+10K) and 41/Vm = 18K/(1+18K)

Rearranging:

1/K = (10Vm - 250)/25 1/K = (18Vm - 738)/41.

Solve simultaneous equations for 1/K → Vm = 205 cm3 (make sure you can do this)

61

62

Langmuir Isotherm – adsorption stops at the monolayer

During physisorption multilayers form – need a better model to account for this

Monolayer volume

Mono- layer

Multilayer

Complex adsorption isotherms

         initial                        irreversible                maturation  I        maturation  II                dispersion attachment              attachment

Medical  devices -­‐hip  implants -­‐  prosthetics -­‐  scalpels

Fouling -­‐Bioreactors -­‐  boats -­‐  gum  disease

SLIME (matrix  protein)

Brunauer, Emmett and Teller developed more realistic model that: 1. Allows multilayer adsorption

2. Different enthalpy of adsorption of multilayers and monolayers

V = volume of N2 adsorbed Vm = monolayer volume P0 = saturation N2 vapour pressure at 77K P = applied N2 pressure

c accounts for enthalpies of adsorption

c = exp(ΔHDO - ΔHVAP

O)/RT

ΔHDO = enthalpy of desorption (strength of adsorbate-surface interaction)

ΔHVAPO = enthalpy of vaporisation (adsorbate-adsorbate interaction in multilayer)64

BET adsorption isotherm

ommo PP

cVc

cVPPVP .)1(1

)(−

+=−

BET Adsorption Isotherm

P = 1 + (c-1) Pn(P - Po) nmc nmc Po

A plot of P/V(P-Po) vs P/Po should be a straight line

Gradient = [(c-1)/Vmc], Intercept = 1/Vmc solve for c = Gradient / Intercept (substitute above to find Vm)

Vm → number of N2 molecules adsorbed,

Area occupied by single N2 molecule is 16.5 Å2

Total surface area = (no. N2 molecules) x (16.5 x10-20) m2

Isotherm is only valid for P/P0 = 0.05-0.3, outside it is not linear.65

x

x

xx

xP/V(P-Po)

P/Po

66

Example BET calculation 1. Calculate monolayer volume Vm

Intercept = 1/Vmc = 2x10-4

Gradient = (c-1)/ Vmc = 0.014 Gradient = (c-1) x 1/Vmc

= (c-1) x Intercept

Gradient = (c-1) x (2x10-4) = 0.014 ∴ (2x10-4)c - (2x10-4) = 0.014

c = (0.014 + (2x10-4))/(2x10-4) → c = 71

Intercept = 1/Vmc = 2x10-4 ∴ 1/(2x10-4)c = Vm

1/((2x10-4)x 71) = Vm → Vm = 70.4 cm3

2. Convert Vm to number of N2 molecules

PV = nRT P = 1.01x105 Pa (Nm-2); V = Vm in m3 (1cm3 = 1x10-6 m3)

R = 8.314 Jmol-1K-1; T = 298 K (SATP)

1.10x105 x 70.4 x10-6 = n x 8.314 x 298 ➔ n = 0.0029 moles

Number N2 molecules = n x NA = 0.0029 x 6.022x1023 = 1.75x1021 N2 molecules

3. Multiply by area of a single N2 molecule

Area of 1 N2 molecule = 16.2 Å2 = 16.2x10-20 m2

Surface area = 1.75x1021 x 16.2x10-20 = 283 m2

Determining the heat of adsorption For a gas at pressure P in equilibrium with the condensed phase at temperature T.

Clausius-Clapeyron equation

rearranges to:

where ΔHads = isosteric heat of adsorption (measured at constant θ)

Measure several adsorption isotherms at different temperatures:

Pressure (=P2

’)

Fractional surface

coverage θ

T3

T2

T1

T3 < T2 <T1 (θ increases with decreasing T)

For particular θ read off values for P1, P2, P3

Plot ln P vs 1/T and slope = -ΔHads /R

P3 P2 P1

θ = constant

67

2

lnRTH

dTPd adsΔ

="#

$%&

'

θ

!"

#$%

&−

Δ−=))

*

+,,-

.

122

1 11lnTTR

HPP ads

θ

• All gases Physisorb on any surface when T < condensation temp.

• Reactive gas Chemisorb on reactive surface when T>condensation temp.

• Langmuir isotherm - assumes all adsorption sites are identical - only useful for describing monolayers

• BET isotherm - better model for physisorption - takes account of varying ΔHads of mono- and multilayers - widely applied to surface area analysis

Adsorption summary

Highly organized macro-mesoporous Al2O3

68

Face Centred Cubic (fcc)

Body Centred Cubic (bcc)

Crystal structures

Most common structures for transition metals

69

Miller Indices Quick way to describe surface: 1. Find the intercepts of the plane with the 3 crystal directions or axes in

terms of primitive vectors (a, b, c)

= (2, 1, 3)

2. Take reciprocals

= (1/2, 1, 1/3)

3. Multiply resulting numbers by the smallest number that yields 3 integers ⇒

(h, k, l) notation

i.e. multiply by 6 ⇒ (h, k, l) = (3, 6, 2)

b

a

c

3

21

70

Miller Indices for a simple cubic lattice

z

yx

(100) (010)(100)(100) (010)(010)

(111)(111)

(110)(110)

71

(100) (110)

a

a√2

(111)

a√3/2

a√2

fcc

bcc

(100) (110) (111)

Atom packing in exposed crystal faces

72

In the bulk each atom is surrounded by 12 nearest neighbours → STABLE

Consider cleaving a metal crystal to generate a (111) plane

Atoms in the surface have lower coordination number

Bulk fcc(111)9 nearest neighbours12 nearest neighbours

Surface energy

73

fcc(100)

fcc(110)

Lower coordination number → more unstable/reactive → more bonds broken → higher surface energy

8 nearest neighbours

6 nearest neighbours

74

More open faces have higher surface energy i.e. for fcc γ111< γ100< γ110

6 nearest neighbours

9 nearest neighbours

8 nearest neighbours

75

76

Consider the surface decomposition of a molecule A , i.e.

A (g) ↔ A (ads) → Products

Let us assume that : • decomposition occurs uniformly across surface sites (not restricted to a few special sites)

• products are weakly bound to surface and, once formed, rapidly desorb

• the rate determining step (rds) is the surface decomposition step

Under these circumstances, the molecules of A on the surface are in equilibrium with those in the gas phase

➔ predict surface conc. of A from Langmuir isotherm

θ = b.P / ( 1 + b.P )

Assumption 4: ΔHads is coverage independent

Assumption 3: Only 1 adsorbate per site

Unimolecular decomposition

77

Rate of surface decomposition (∴reaction) is given by an equation:

Rate = k θ (assuming that the decomposition of Aads occurs in unimolecular elementary reaction step and that kinetics are 1st order in surface concentration of intermediate Aads)

Substituting for θ gives us eqn. for rate in terms of gas pressure above surface

Two extreme cases: • Limit 1 : b.P << 1 ; i.e. 1st order reaction (with respect to A) with an 1st order rate constant , k' = k.b

This is low pressure (weak binding) limit:

Rate = k b P / ( 1 + b P )

then ( 1 + b.P ) ~ 1 and Rate ~ k.b.P

➔ steady state θ of reactant v. small

78

• Limit 2 : b.P >> 1 ; then ( 1 + b.P ) ~ b.P and Rate ~ k

• i.e. zero order reaction (with respect to A)

This is the high pressure (strong binding) limit : steady state surface θ of reactant ~100%

Rate shows the same pressure variation as θ (not surprising since rate ∝ θ!)

Rate = k b P / ( 1 + b P )

79

Langmuir-Hinshelwood type reaction:

Assume that surface reaction between two adsorbed species is the rds. If both molecules are mobile on the surface and intermix then reaction rate given by following equation for bimolecular surface combination step:

Rate = k θΑ θΒ

Since θ = b.P / ( 1 + b.P ), when A& B are competing for same adsorption sites the relevant equations are:

A (g) ↔ A (ads) B (g) ↔ B (ads)

A (ads) + B (ads) AB (ads) AB (g)rds fast

Bimolecular reactions: 1

80

Look at several extreme limits: •Limit 1 : bA PA << 1 & bB PB << 1

In this limit θA & θB are both very low , and

Rate → k . APA . bBPB = k' . PA. PB 1st order in both reactants

•Limit 2 : bA PA << 1 << bB PB

In this limit θA → 0 , θB → 1 , and

Rate → k . bA PA / (bB PB ) = k' . PA / PB

Substituting these into the rate expression gives :

1st order in A negative 1st order in B

θ = b.P / ( 1 + b.P )

Rat

e

Pure A Pure B[A]/[B]

Competitive Adsorption

81

Eley-Rideal type reaction : Consider same chemistry

A (g) ↔ A (ads) A (ads) + B (gas) AB (ads) AB (gas)

last step is direct reax between adsorbed A* and gas-phase B.

82

A + B ➔ AB

rds fast

Rate = k θΑ [Β]

where [B] is pressure/conc in gas or liquid phase

[A ]/ [B]

Rat

e

A varied

Bimolecular reactions: 2

83

However Without extra evidence cannot conclude above reaction is Eley-Rideal mechanism… last step may be composite and consist of the following stages

B (g) ↔ B (ads) A (ads) + B (ads) AB (ads) AB (g)

with extremely small steady-state coverage of adsorbed B ➔

Test by monitoring rate • vary θΑ

• vary ratio of or over wide range

fast fast

slow

Langmuir-Hinshelwood not Eley-Rideal

B

A

pp

]B[]A[ need free sites

84

Calculated energy diagram

Langmuir-Hinshelwood: CO oxidation over Pt

Highest rate of CO2 production under slightly oxidising conditions: - a high concentration (~0.75 monolayer) of surface O - significant no. of Oa vacancies (empty sites) - CO adsorbs in vacancy with only small energy barrier

Reaction pathway

COO

CO(g)+O(a)

Example 1

CO(a)+O(a)

CO2 (g)

CO(g)+½O2(g)

85

Ru catalyst

O atoms

Eley-Rideal: CO oxidation over Ru

Highest rate of CO2 production under oxidizing conditions: - a high concentration (1 monolayer) of surface O - no surface CO detectable

Calculated energy diagram

Transition state

GAS

SURFACE

CO(g)+O(a)

Example 2

CO2 (g)

86

Oscillating reactions of carbon monoxide oxidation on platinum.

Good for oxididation

‘Inert’ towards O2

Can adsorb CO

Ultra High Vacuum Equipment

• Measurements require Ultra High Vacuum (< 10-7 torr)

- trajectories of electrons/ions used in analysis remain unperturbed

- surface kept free of contamination

Apparatus for Surface Analysis

87

Retarding Field Analyser

Electron analysers

88

Concentric hemispherical analyser (CHA)

+ ve

- ve

- ve

89

X-ray Photoelectron Spectroscopy (XPS)

Typical photon sources: Mg Kα = 1254 eV Al Kα = 1487 eV

Elemental analysis of surfaces

Binding Energy

Kinetic Energy

hνKE = hν - BE

90

KE of emitted electron varies depending on photon energy

Binding Energy eV

XPS spectrum of zirconium

91

Limiting resolution depends on X-ray line width.

- higher resolution achieved using monochromator

92

Monochromated X-rays

Can also use synchrotron source ⇒ continually variable source of monochromated X-Rays

Bremsstrahlung background

93

Auger Electron Spectroscopy (AES)

Can use X-ray or electron excitation source

Typical electron excited Auger spectrum

94

The Auger process

Initial excited state

Auger process

Final state

EAUGER

Ew

Ex

Ey

Ep

e-

Ground state

Photoelectron

95

Energy of emitted Auger electron involving levels w, x and y:

Ewxy ≈ Ew(Z) - Ex(Z) - Ey(Z)

Internal atomic rearrangement – independent of energy source used.

Assign elements in surface using tables of electron binding energy.

e.g. calculate the kinetic energy of the OKLL Auger transition: 493-512 eV (depending on oxide environment)

96

Why are low energy electron spectroscopies surface sensitive? - high KE strong scattering by neighbouring atoms - low KE excitation of

Surface sensitivity

Kinetic Energy eV

Esc

ape

Dep

th /n

m

Universal escape depth curve

97

Electrons excited by backscattering and emitted

Auger/photo electrons excited by incident beam and directly emitted

Excitation source

Auger/photoelectrons excited but not lose too much energy to be emitted

98

Can use XPS/AES to study thin films/coatings

Frank-Van der Merwe

Stranski-Krastanov

Applications

99

Simultaneous Multilayers

Volmer-Weber

100

Id/Io = exp (- d/lcosθ)

θ e-

d

Io = intensity of substrate peak for clean surface

Id = intensity of substrate peak after growth of film

l= escape depth of emitted electron θ = angle between surface normal and detector

d = film thickness

101

Angular Resolved XPS

I1/I2

Angle

I1I2

I1 I2

I1/I2

Angle

I2

I1

I2I1

I2

Fe Cr

102

Electron binding energy shifts depending on atom neighbours - increases if surrounded by electron withdrawing groups (e.g. O, F) - decreases “ “ “ “ donating groups (e.g. K, Ca, H)

Chemical environment

103

Accurate measurement of BE and ‘chemical shift’ tells us - surface functionalisation (e.g. polymer coating) - oxidation state of elements (e.g. rusting)

N 1s XP spectrum of NH3 oxidation catalyst

104

105

106

0.25

0.20

0.15

0.10

0.05

0

θ C2H

4 / M

L

0 1 2 3 4Exposure / L

Monolayer

285 284 283 282

2.81.230.160.06

Expo

sure

/ L

Binding Energy / eV

C 1s Fast XP spectra of C2H4 adsorption on Pt(111)

Multilayer

C2H4

C2H4

Pt(111)

C2H4

C2H4

Ethene adsorption

• Precursor-mediated adsorption

di σ/π-bound

• Single adsorbate

107

Ethylidyne

285 284 283 282500

400

300

200

100

Binding Energy / eV

285 284 283 282

126183

238293

353467

621

Tem

perat

ure / K

Binding Energy / eV

C2H4C 1s Fast XP spectra of C2H4 reaction on Pt(111)

Tem

pera

ture

/ K

CHx

Carbon• Surface coverage = 0.25 ML

H2

H2

Time-resolved XPS

108

• Stable ethylidyne intermediate ΔEact = 75 kJmol-1

3.5 3.7 3.9 4.1 4.3 4.5 T-1 / 10-3 K-1

ln (R

ate)

-7

-8

-9

-10

-11

-12

-13

Eact = 57 ± 3 kJmol-1

ν = 1x1010±.0.5 s-1

Temperature / K 100 200 300 400 500 650

H2 D

esor

ptio

n 3 L C2H4

1st order kinetics

0.1

0.2

0.3

0

0.4

0.5

C :

Pt ra

tio

Carbon C2H3

di-σ C2H4 Total C

C2H4

CHx Cgraphite H2 H2

C2H4,H2

Surface reaction kinetics

Permit quantitative analysis of surfaces • Sensitive to 0.1-1% monolayer (1012 - 1013 atoms/cm2) • Oxidation state information? - if oxidation state of element changes → electron binding energy changes ➔ chemical shift - complex for AES as 3 electrons involved, hard to interpret shifts

• Surface sensitivity depends on kinetic energy of emitted electron - 50-500 eV Auger electrons are highly surface sensitive

• Auger electron energy independent of energy of excitation source

XPS/AES summary

109

RAIRS

Reflection Absorption Infra Red Spectroscopy

Surface vibrational spectroscopy

110

2884

1339

1118

C-H symmetric stretch C-H deformation C-C stretch

RAIRS of ethylidyne Pt(111)

Heat

111

Selection Rules:

Angle of incidence (degrees)

Res

pons

eEp E’p

Es

E’s

Surface

IR

112

Metal Surface Selection Rule

Vibrational mode of surface molecule needs dipole moment perpendicular to the surface to be detected

113

HREELS

High Resolution Electron Energy Loss Spectroscopy

Sample Slits

Electron Gun

Monochromator

Detector

114

Surface

Incident electron beam interacts with vibrating molecule - loses/gains energy

- energy change → vibrational frequency

‘Specular’ detection: θreflection = θincidence

115

Dipole Scattering: Electric field of e- interacts with dipole ┴ to surface due to molecular vibration. Strongest in SPECULAR direction

Impact Scattering: Impact of e- with molecule excites vibration. Not dependent on orientation of dipole. Best viewed OFF SPECULAR

116

CO/Ni(111)

CO/Pt(111)

117

HREELS of C2H4

Why do we see the C=C stretch?

118

HREELS vs RAIRS Resolution > 2meV 0.1meV Pressure Range < 10-8 torr UHV to >1 bar Spectral Range > 10 meV >100 meV

• RAIRS low frequency range is limited by detector, while for HREELS it is limited by elastic peak width.

• Vibrating molecule must have a dipole perpendicular to surface

• In HREELS can observe dipoles parallel to surface using ‘impact mode’ and working off specular

Summary

119

Group frequencies

120

121

Low Energy Electron Diffraction (LEED)

Structural characterisation of surfaces

122

dsinθ = nλ

d

Incident electron beam Diffracted electrons

Surface diffraction occurs due to regularly spaced unit cells

123

dsinθ = nλ

and

sin θ = opposite/hypotenuse = y/L

∴y/L = nλ/d or separation of spots ∝ 1/d (atom/molecule spacing)

θ

LL

y

Sample

124

Possible structures of p(2x2) oxygen unit cell on Cu(100)

LEED pattern for O2/Cu(100)

a

b

+ 0.25 ML Oadsorbed

Clean Cu(001)

125

126

Temperature-programmed desorption

127

Pt(111)

128

Quadrupole Mass Spectrometer

H2

Temperature / K 100 200 300 400 500 650

H2 D

esor

ptio

n 3 L C2H4

Stepwise decomposition

C2H3

CH3

CH2

For a simple 1st order desorption the activation energy (Eact) for desorption

can be calculated from the desorption peak temperature using the

Redhead equation:

β = heating rate (typically 1-1000 K sec-1)

Tp = peak temperature

ν = constant (frequency factor) = 1015 sec-1

Eact = RTp [ln(νTp/β) - 3.46] Redhead equation

129

CO/Pt(112)- stepped surface - get info about preferential adsorption site

Weakly-bound

Strongly-bound

130

TiO2

Can also follow surface reactions

Temperature / K

O

Acetophenone decomposition on TiO2(001)

131

Scanning tunnelling microscopy (STM)

Gerd Binnig & Heinrich Rohrer (Nobel Prize, Physics 1986)

132

TipSample

Electronic Wavefunction

Fermi Level

Vacuum Level

φS φT

At a certain applied potential electrons tunnel to or from tip.

Tunnelling current sensitive to tip sample separation

Maintain a constant tunnelling current while scanning tip across surface.

Motion of tip towards or away from surface ⇒ topography of surface.

133

134

Xenon on nickel (110) CO on platinum (111)

Fe on Cu(111)

Adsorbate-induced reconstruction

N2/Cr(110)

135

Chemical Contrast

Pt/Rh(100)

136

Physical interaction with the surface and tip

Vertical displacement is registered by deflection of laser

Compile a topographic image of surface

Can record images of - insulating samples

- solid liquid interface

- biological samples

Atomic force microscopy (ATM)

137

138

AFM image of contact lens recorded in saline solution

Hydrophobic lens Lens with hydrophilic coating

139

140

• Supported metal particle can expose different crystal faces.

• In addition there are steps & defects within each particle. - these are low coordination sites - region of high potential energy ➔ facilitate bond dissociation

Structure sensitivity

141

Structure Sensitivity occurs when reaction requires specific active sites: (any mix of step, terrace, kink atoms)

Density of steps and dominant crystal face reflects the metal particle size

∴changing particle size modifies rate

Stepped surfaces Stepped + kinked surface

(100)

square

(111)

hex

142

100xNN(%)Dispersion

T

S=

Consider total fraction of available surface sites:

Spherical particles

if Ns = total no. of surface atoms NT = total atoms in particle

For small particles (< 20Å) Dispersion → 1

∴if Activity ∝ SA, then ⇑ particle size will ⇓ rate (per mass of catalyst)

provided exposed surface atom arrangement unchanged

143

Structure sensitive test:

Consider CO + 3H2 → CH4 + H2O

Compare specific TON (per surface site)

Ni (100)

9% Ni/Al2O3

5% Ni/Al2O3

If reaction requires specific (4-coord) active site expect

• constant ΔEact observed

• higher rate over surfaces with most (100) sites larger particles

144

Structure sensitive vs insensitive reaction:

Cyclohexane hydrogenolysis • High step/kink densities → high rates • Reaction requires defect sites

contrast with (de)hydrogenation which proceeds over diverse surface arrangements

Reaction kinetics tell us about the active site

-H2

-CHx

• The importance of surface processes

• How surface structure affects the reactivity of materials

• Adsorption processes, physisorption and chemisorption: - activated vs non-activated adsorption

• Adsorption isotherms: Langmuir and BET models

• Kinetics of surface reactions: - Eley Rideal and Langmuir Hinshelwood mechanisms

• Surface analytical techniques: specifically XPS/AES

145

Summary