Upload
dangdung
View
226
Download
0
Embed Size (px)
Citation preview
Computational methods in catalysis
Horia Metiu
Department of Chemistry and Biochemistry
University of California at Santa Barbara
Santa Barbara, California
University of California at Santa Barbara
There are two important things about a chemical reaction
Thermodynamic equilibrium (what is the maximum conversion?)
Kinetics (how fast the reaction takes place?)
A catalyst changes the rate of reaction
It is not consumed or created during reaction
It does not change the maximum yield.
Enzyme catalysis
Homogeneous catalysis
Electrocatalysis
Photocatalysis
Heterogeneous catalysis
I’ll focus on ordinary heterogeneous catalysis
It is “easier” than the other areas of catalysis
It is the most important for economy
There are many opportunities for scientific discovery
One can contribute to economy and make some serious money
How important is ordinary heterogeneous catalysis ?
►80 % of chemical industry, 20% of GDPfuels (gasoline, diesel, aviation, heating)commodity chemicals (sulfuric acid, methanol, ethylene, propene, vinyl chloride, ammonia)monomers for all plastics including rubber polymers (polyethylene)cosmeticsdetergentsfuel cells (polymeric and oxide)
Prior to the discovery of oil most chemical industry was based on coal
Germany
Catalyst development: a case study
How catalytic research is done
What one person working in this field can do
The initial use of oil: distillation. heavy fraction: lubricants and lighting. lighter fraction: useless
Cars were invented. gasoline became useful the low quality of gasoline was the limiting factor in engine development (natural oil does not have the right kind of hydrocarbons)
Enter Eugene Houdry
rich, French mechanical engineer, interested in car racing.
Lignite to gasoline: 1922-1929. Pilot plant produced gasoline but not economical (not cheaper, not better). French government withdraws funding.
1930-1933: Goes to US. Switches from lignite to oil. Sponsored by Vaccum Oil Company. After testing hudreds of compounds he finds an aluminosilicate catalyst that produced gasoline of good quality (need branched alkanes and aromatics)
1933 Depression: the company told Houdry to find another sponsor
1933-1937: Sun Oil Company funds his research. Houdry finds a method to regenerate the catalyst. In 1937 a commercial unit started working and producing high octane gasoline (winning WW2)
Houdry also discovered a catalyst for making butadiene from oil, whichis copolymerized to make rubber. Also credited with winning the World War II.
Catalytic converter. Not used because it was poisoned by tetraethyl lead. After tetraethyl lead was banned converters came into use.
Better life through chemistry for us and for Houdry
Great opportunities: Most of the products chemical industry makes now are derived from oil by catalysis:
We need to convert almost all chemical industry from
oil to natural gas, or coal or CO2. We need new catalysts
and new processes for doing this
Is this kind of work possible today?
YES
The methodology is still empirical but computation will play a major role
Catalysis community
The three communities in catalysis
Surface science research
Academic “practical” catalysis research
Industrial catalysis research
The “sociology” of catalysis research community
Computational work:
binding sites and binding energy
spectroscopy
activation energies for reactions
rate constant
kinetic simulation
The main goal is working with systems which provide detailed, firm knowledge.
Surface science:
well defined surface (faces of single crystal)
perfect surface cleanliness (pressure 10-10-10-11 torr )
large number of tools (electron spectroscopy can be used)
Practical catalysis research in academia and industry
Use powder catalysts often amorphous
Use atmospheric pressure or higher
don’t know surface structure (electron microscopy)
surface purity is an issue
Main work:
Making catalysts (pay attention: MgO)
Testing chemical performance
Physical characterization (IR, Raman, EXAFS, NEXAFS, XRD, X-ray fluorescence, atomic analysis, NMR, ESR, Mossbauer, UV-Vis)
Flow gases or liquids through a bed of solid catalyst
Molecules stick to the solid, react, desorb and exit the bed
What are the goals of academic “practical” catalysis research?
Find new and better catalysts
Improve existing catalysts (the 1 % crowd) (refineries)
Explain how existing catalysts work (it is sometimes useful: coking; Houdry)
Find a general organizing principle (explain)
Industrial research: money
conversion
rate
selectivity
long term stability
safety and environmental effects (Cd, Be)
cost (Ru for NH3)
heat management (the thresholds (methane,CO2), ethylene epoxidation)
Performance criteria for academic research in practical catalysis
Selectivity:
make only the compound you want
Partial oxidation of CH4:
methanol, formaldehyde, dimethyl ether, CO, CO2, H2O
CH2= CH2 and CH CH
Hydrogenate acetylene but not ethylene
H2 and CO
oxidize CO to CO2 but do not oxidize hydrogen to water
An exception from these rules:
catalysts that involve new principles or new classes of compounds
If you want to do computations to explain how a catalyst works look for:
important reaction for a product for which demand is growing
inexpensive and non-toxic catalyst
conversion over 40%
excellent selectivity
long lasting
How to model practical catalysts?
Ru catalyst supported on an oxide particle of oxide catalyst
Customary model
Highest area in a crystallite
Must be the most important face
Lowest surface energy
LaOCl(101)&
LaOCl(011)
LaOCl(100)&
LaOCl(010)
LaOCl(111)
LaOCl(110)
Es = 0.519 Es = 0.599 Es = 0.622 Es = 0.820
LaOCl(001)
Es = 0.106
LaOClLa (BSL)O (BSL)Cl (BSL)
CH4 + HCl + 1/2O2 = CH3Cl + H2O oxichlorination
MoS2 important catalyst for oil industry
Platelets
The face shown in the picture is the largest and it is inactive
Catalysis takes place at the edges.
Lesson 1: If the catalyst is crystalline you must consider all faces of the crystallite.
Lesson 2: Look at the role of defects: steps, missing atoms, small clusters on top of the surface (Taylor and active centers).
Oxide catalysts, oxidation reactions and Mars van Krevelen
Lesson 3: The catalyst is what becomes of the initial material in contact with the reactants and the products when it reaches the steady state.
CH4 + HCl + 1/2O2 = CH3Cl + H2O oxichlorination
People have used as catalyst
La2O3
or LaCl3
or LaOCl
Lercher & Dow Chemicals
Which of these is the catalyst?
None of the above
Do three experiments of methane oxichlorination simultaneously:
one with La2O3
one with LaOCl
one with LaCl3
After 10 hours of running the reaction you reach steady state: all three experiments have the same performance; the initial “catalysts” evolve to the same state which is the real catalyst
Lesson 3: The catalyst is what becomes of the initial material in contact with the reactants and the products when it reaches the steady state.
We expose the initial compound simultaneously to HCl and O2
Fact #1: LaCl3 and LaOCl react with O2 to make La2O3
and
Fact #2: La2O3 reacts with HCl to make LaOCl and then LaCl3
HCl puts Cl atoms on the surface and removes O atoms
O2 puts O atoms on the surface and removes Cl atoms
The surface is a mixture of O and Cl
The O to Cl ratio on the surface depends on the ratio of HCl and O2 in the gas and on temperature!!!!
This surface is the real catalyst; this is the surface to be modeled!!!
What to keep in mind when doing computational work:
we are working with imperfect models
we are working with an imperfect theory (DFT)
There are two extreme responses. Response 1:
work on well defined models for systems for which DFT is accurate (a small field; not many practical problems)
binding energies
activation energies and rate constants
spectroscopic quantities
Response 2: Learn how to do useful work in spite of inaccuracies.
Don’t try to get the exact activation energies: order the catalysts according to their activity and pass this information to you experimentalist collaborator.
Chemistry depends on differences in total energy ! binding energies, activation energies, vibrational frequencies
Find simple descriptor of the catalytic processmethane activation: activation energy of breaking the C-H bond
Trends in chemistry depend on differences of energy differencesCompare same reaction (e.g. H2 dissociation) on different catalysts (e.g. Pt, Pd, Ni, Cu). Compare activation energies forH2 dissociation on each metal.
The goal of such calculations is to increase the probability that the experimentalists find a good catalyst
It is very useful to be imbedded in an experimental group
A history lesson: ammonia synthesis
N2+ 3 H2 = 2NH3
Haber’s experiments in Karlsruhe: Osmium
Mittasch (BASF) 200 compounds; iron oxide with alkali; promoter
After Mittasch: 25000 catalysts have been tried; lower limit; We still use the Mittasch catalyst
Can we use theory to help us do better than this?
Professor: “remember now you got the brains”
The discovery of a new catalyst for ammonia synthesis
N2 + 3H2 = 2 NH3
(Norskov group + Topsoe)
N2 + surface = 2 N-surface
H2 + surface = 2 H surface
N-surface + H-surface = NH-surface
NH-surface + H-surface = NH2-surface
NH2-surface + H-surface = NH3-surface
NH3-surface = surface + NH3(gas)
Fact #1: Dissociation of N2 is the rate limiting step
Approximation #1: Bronsted-Evans-Polanyi
The activation energy for N2 dissociation is a linear function of the binding energy of N to the surface (importance; historical note)
Approximation #2: The stronger N binds to the surface the less reactive it is.
Consequences:
if N binds very strongly Evans Polanyi says that it is easy to dissociate N2 but N won’t react (poor catalyst)
if N does not bind strongly enough then you cannot dissociate N2 effectively (poor catalyst)
The best catalyst will bind N with moderate strength.
Note that because of these approximation all you need to do is calculate the binding energy of N !!!!
Calculate binding energy of N on all metals and alloys
Volcano curve: rate of ammonia production versus the binding energy of N to the metal catalyst.
Os, Ru best but Fe is used
Mo dissociates N2 easily but binds N too strongly: poissoning
Co binds N weakly but does not dissociate it efficiently
How about a CoMo alloy?
Topsoe: did experiments and found that
Co3Mo3N is as good as the best activated
Ru catalyst (which is too expensive to use)
Maybe this was a lucky accident!!
The same principle was used to find a new catalyst for the methanationreaction (conversion of CO and H2 to methane)
Best catalysts are Co and Ru but they are too expensive and are not used. The recommended catalyst was Ni
Ni is on one side of the volcano (below Co and Ru) and Fe is on the other side. A NiFe alloy is more efficient than Ni and cheaper.
The art of using DFT to contribute to catalysis science and practice
Catalysis offers computational chemists tremendous opportunitiesfor the next fifty years.
Make a reasonable model for the system
Find a simplifying descriptor of catalytic activity
Compare many catalysts for the same reaction
Give experimentalists systems to try and also systems not to try
Tell them that you are using intelligent, DFT-assisted guessing
This is better than trying 25,000 catalysts at random.
Problems and systems
Methane activation
stranded methane (flared or vented) turn it to liquids
methane as a raw material
now: methane to syn gas (CO + H2)
ammonia
methanol
gasoline
Can we convert methane directly (without syngas)?
methanol, formaldehyde, ethylene
yields are low, temperatures high, selectivity
CO2 utilization:
CO2 as oxidant (power plants):
CH4 +CO2 methanol, formaldehyde, CO + H2
Systems
cation doping
anion doping
oxide clusters on oxides
complex “oxides”: carbonates, phosphates, vanadates, molibdates, perovskites, etc
heterogenize homogeneous catalysts
make artificial enzymes
carbonic anhydrase with Zn cofactor
CO2 bicarbonate + protons
Protein
Active center
Pocket (channel)
Great selectivity
Do difficult chemistry (NH3)
Rate or structure simulations:
Difficulties:
flexibility, water matters, charge transfer is frequent; even remote proteins could matter
DFT for the active center
Classical treatment of protein and water (Monte Carlo or Molecular Dynamics)
Classical mechanics for protein and a continuum electrostatics for water
Blocking the pocket (Viagra)
Polymeric membrane electrolyte
Nafion
Carbon cloth anode Carbon cloth cathode
Carbon black particles
Pt catalyst at the cathode
Pt/Ru catalyst at the anodeH+
H+
H+
H+
H+
H2 or
CH3OH or
HCOOH
O2
What a fuel cell looks like: Electrocatalysis
Problems:
Pt and Pt/Ru too expensive and too rare
CO poisoning
Oxygen reaction too slow
Homogeneous catalysis: one phase
Wacker process:
H2C=CH2 + ½ O2 CH3CHO (acetaldehyde)
H2C=CH2 + ½ O2 CH3CHO (acetaldehyde)
Pd[Cl4]2- is considered to be the catalyst, but CuCl2 and H2O could also be considered catalysts.
Pd and goes back and forth between Pd2+ and Pd0: Cu is shifts from Cu2+ to Cu+; protons and Cl- are produced. Solvation is essential !!!
The reaction takes place efficiently at sites where the water molecules are in the right position. The water is part of the reaction coordinate
The ions interact with long range forces and these cannot be neglected