2propenyl radicals

1
From combustion to astrochemistry How elementary
physics principles give predictive ability
Laurie J. Butler The University of Chicago
2-propenyl radicals Julie Mueller (Santa Clara
Univ. ? Oregon) Vinoxy radicals Johanna
Miller (Physics Today) Doran Bennett (Goldwater,
NSF)
Radicals Molecules with unpaired
electrons
with no
National Science Foundation Department of Energy
2
From combustion to astrochemistry How elementary
physics principles give predictive ability
Laurie J. Butler The University of Chicago
2-propenyl radicals Julie Mueller (Santa Clara
Univ. ? Oregon) Vinoxy radicals Johanna
Miller (Physics Today) Doran Bennett (Goldwater,
NSF)
3
From combustion to astrochemistry How elementary
physics principles give predictive ability
Laurie J. Butler The University of Chicago
2-propenyl radicals Julie Mueller (Santa Clara
Univ. ? Oregon) Vinoxy radicals Johanna
Miller (Physics Today) Doran Bennett (Goldwater,
NSF)
Helec?elec(ri) E?elec(ri)
4
Combustion Processes O2 reacts with hydrocarbon
fuels to abstract a hydrogen atom and fuel
molecules break at a C-H bond, creating radical
intermediates. The radicals subsequent chain
reactions release energy and form CO2. Unwanted
reactions produce pollutants, including NOx, PAHs
and soot.
5
We use laser light to generate specific molecular
radicals dispersed by their velocity and internal
energy, then study their subsequent reactions.
h?
(193 nm)
Cl
Cl
6
We use laser light to generate specific molecular
radicals dispersed by their velocity and internal
energy, then study their subsequent reactions.
h?
(193 nm)
Cl
Cl
ET
Eint radical hn??-??Do(C-Cl) - ET
148 -??????? - ET (kcal/mol)
7
We use laser light to generate specific molecular
radicals dispersed by their velocity and internal
energy, then study their subsequent reactions.
h?
(193 nm)
Cl
Cl
ET
Eint radical hn??-??Do(C-Cl) - ET
148 -??????? - ET (kcal/mol)
Maxwell
Einstein
Light quanta or photons Energyphoton h??hc/?
Energy ?(Amplitude)2 (nothing to do with
wavelength)
8
(No Transcript)
9
(No Transcript)
10
We use laser light to generate specific molecular
radicals dispersed by their velocity and internal
energy, then study their subsequent reactions.
h?
(193 nm)
Cl
Cl
Eint radical hn??-??Do(C-Cl) - ET
148 -??????? - ET (kcal/mol)
H
H
propyne
allene
Helec?elec(ri) E?elec(ri)
11
We use laser light to generate specific molecular
radicals dispersed by their velocity and internal
energy, then study their subsequent reactions.
h?
(193 nm)
Cl
Cl
Eint radical hn??-??Do(C-Cl) - ET
148 -??????? - ET (kcal/mol)
36.1
37.1
H2CCCH2 H allene H
Eint radical
HC? C- CH3 H propyne H
H2C C?- CH3 2-propenyl
12
(No Transcript)
13
Measure the Kinetic Energy ET imparted to the Cl
atom radical to disperse the radicals by
internal energy. Then determine product
branching as a function of internal energy in the
radical.
193 nm
nozzle
ionization source (electron impact at UofC,
tunable VUV at ALS)
-30 kV Al doorknob
skimmers
quadrupole mass spec.
Scintillator
PMT
Eint radical hn-Do(C-Cl)-ET
14
C-Cl fission gives 2-propenyl radicals dispersed
by internal energy
193 nm
Cl
Cl
36.1
37.1
H2CCCH2 H allene H
HC? C- CH3 H propyne H
H2C C?- CH3 2-propenyl
E (kcal/mol)
15
C-Cl fission gives 2-propenyl radicals dispersed
by internal energy
193 nm
Cl
Cl
36.1
37.1
H2CCCH2 H allene H
HC? C- CH3 H propyne H
H2C C?- CH3 2-propenyl
Eint radical hn-Do(C-Cl)-ET
E (kcal/mol)
16
C-Cl fission gives 2-propenyl radicals dispersed
by internal energy
193 nm
Cl
Cl
ET
57.5
56.5
54.4
53.3
H2CCCH2 H allene H
HC? C- CH3 H propyne H
19.4
H2C C?- CH3 2-propenyl
Fast Cl atoms should have momentum-matched stable
mass 41 radicals
E (kcal/mol)
17
C-Cl fission gives 2-propenyl radicals dispersed
by internal energy
193 nm
Cl
Cl
ET
57.5
56.5
54.4
53.3
H2CCCH2 H allene H
HC? C- CH3 H propyne H
19.4
H2C C?- CH3 2-propenyl
Fast Cl atoms have momentum-matched stable mass
41 radicals
E (kcal/mol)
18
The radical co-fragments of the slower Cl atoms
have higher internal energy so dissociate to C3H4
(mass 40) H
Cl
ET
57.5
56.5
54.4
53.3
H2CCCH2 H allene H
HC? C- CH3 H propyne H
19.4
H2C C?- CH3 2-propenyl
E (kcal/mol)
19
The radical co-fragments of the slower Cl atoms
have higher internal energy so dissociate to C3H4
(mass 40) H
Cl
ET
57.5
56.5
54.4
53.3
H2CCCH2 H allene H
HC? C- CH3 H propyne H
19.4
H2C C?- CH3 2-propenyl
Allene/propyne products from radicals with a
given internal energy can be identified by their
velocity.
E (kcal/mol)
20
Exptl vs. RRKM branching ratio as a function of
internal energy in the radical
Data show allene yield is 2.0 (.05/-.15) times
higher for the higher internal energy radicals
than the lower internal energy ones. RRKM
prediction is 2.2 times higher. (ltEintgt15 kcal
ltEintgt3 kcal)
57.5
54.4
56.5
53.3
H2CCCH2 H
HC? C- CH3 H
Detect allene only w. 10.0 eV
allene
propyne
19.4
21
Exptl vs. RRKM branching ratio as a function of
internal energy in the radical
Data show allene yield is 2.0 (.05/-.15) times
higher for the higher internal energy radicals
than the lower internal energy ones. RRKM
prediction is 2.2 times higher. (ltEintgt15 kcal
ltEintgt3 kcal)
57.5
54.4
56.5
53.3
H2CCCH2 H
HC? C- CH3 H
Detect allene only w. 10.0 eV
propyne
19.4
22
Exptl vs. RRKM branching ratio
Data shows higher internal energy onset for
allene H and is fit well using RRKM prediction
ka(E)/(ka(E)kp(E))
57.5
54.4
56.5
53.3
H2CCCH2 H allene H
HC? C- CH3 H propyne H
Detect allene only w. 10.0 eV
19.4
23
The reactions thus far have been electronically
facile
e.g.
H
a'
a'
a'
One gets good predictive ability with quantum
scattering calculations or transition state
theory on a single ABO potential energy surface.
24
The reactions thus far have been electronically
facile
e.g.
H
a'
a'
a'
One gets good predictive ability with quantum
scattering calculations or transition state
theory on a single potential energy surface.
What classes of reactions are electronically
difficult or prohibitive, such that the
electronic wavefunction cant change as required
along the adiabatic reaction coordinate, and what
happens then?
25
Electronically Non-Adiabatic Processes
What are conical intersections and avoided
curve crossings? Why cant I just solve
HX(R)EX(R) for the nuclear dynamics? When do I
have to include nonadiabatic effects? How can
we probe this experimentally?
NH3 ?NH2 H
Biesner et al. JCP 91, 2901 (1989)
26
Why cant I just solve HX(R)EX(R) for the
nuclear dynamics?
Solve the electronic part of the SE to get a PES
(e.g U11(R)), then solve HX(R)EX(R), e.g. (TR
U11(R)) X(R) EX(R) for the nuclear dynamics of
the chemical reaction.
Below Energy along the IRC for vinoxy?ketene
H. Bennett et al. Barrier is calculated
at the MRCIQ level of theory (aug-cc-pvtz)
(3 core, 3 closed and 11 active
orbitals) is 43.3 kcal/mol)
H
ketene X(1A1)
vinoxy X(2A?)
27
How does the geometry change along the IRC for
vinoxy?ketene H?
Youd guess
H
H
ketene X(1A1)
vinoxy X(2A?)
28
NO. The transition state is twisted. (Should we
force it to stay planar?)
H
H
ketene X(1A1)
vinoxy X(2A?)
29
H
ketene X(1A1)
vinoxy X(2A?)
30
H
ketene X(1A1)
vinoxy X(2A?)
31
H
ketene X(1A1)
vinoxy X(2A?)
32
H
ketene X(1A1)
vinoxy X(2A?)
33
H
ketene X(1A1)
vinoxy X(2A?)
34
A?
H
ketene X(1A1)
Ethan Alquire!
vinoxy X(2A?)
35
A?
a'
a?
H
ketene X(1A1)
vinoxy X(2A?)
36
H
3A?
A?
a'
a?
H
ketene X(1A1)
vinoxy X(2A?)
37
H
3A?
A?
a'
a?
A
H
ketene X(1A1)
vinoxy X(2A?)
38
H
3A?
A?
Conical Intersection
a'
a?
A
A
H
ketene X(1A1)
A?
vinoxy X(2A?)
39
H
3A?
A?
Avoided Crossing
a'
a?
A
A
H
(1a?)2(10a)1(2a?)2
ketene X(1A1)
(1a?)2(10a)2(2a?)1
A?
vinoxy X(2A?)
40
H
Now use HX(R)EX(R) for the dynamics? Maybe
3A?
A?
Avoided Crossing
a'
a?
A
H
ketene X(1A1)
vinoxy X(2A?)
41
Competing dissociation channels of the vinoxy
radical
OsbornNeumark excited vinoxy to the B(2A?) state
and assumed internal conversion to the ground
state. Their RRKM calc. agreed with their rough
product branching. (they could not find H
ketene TS, so used 3 kcal exit barrier per Benson
analogy)
RRKM 4 1 Expt 4?2 1
41.0 kcal
37.6 kcal
H ketene
CH3 CO
42
Competing dissociation channels of the vinoxy
radical
MatsikaYarkony showed the early dynamics is via
a B/A avoided crossing (Yamaguchi) and calculated
X/A conical intersections promoting internal
conversion.
RRKM 4 1 Expt 4?2 1
Our G3//B3LYP calculations found the H ketene
TS (slightly higher barrier) high energy RRKM
branching is similar to Osborn/Neumark estimate
of 80.
43
Try chloroacetaldehyde as a precursor for vinoxy
X(2A?).
a?
193 nm
Cl
Do the nascent radical co-products have enough
internal energy to dissociate?
Eint radical hn-Do(C-Cl) - ET 148 - 73.4 -
ET
(energies should be accurate to ?2 w.r.t. the H
ketene asymptote)
44
Nascent vinoxy radicals have internal energies
spanning the calculated H ketene and
isomerization (acetyl?CH3CO) transition states
45
Does vinoxy X(2A?) branch to the CH3 CO product
channel?
isomerize?
CH3 CO
CH3CO
46
YES
Does vinoxy X(2A?) branch to the CH3 CO product
channel?
isomerize?
CH3 CO
CH3CO
Fit assumes all nascent vinoxy radicals
dissociate to CH3CO with no KE release. The
poor fit on the fast side is telling us that
vinoxy radicals from the highest KE C-Cl
bond fissions, i.e. vinoxy radicals with the
lowest internal energies, do not dissociate to
CH3 CO.
47
Which vinoxy X(2A?) dissociate to the CH3 CO
product channel?
isomerize?
CH3 CO
CH3CO
Vinoxy radicals from these C-Cl fissions do.
Note The observed CH3 CO product channel
evidenced no significant recoil KE. This is
inconsistent with prior studies of the secondary
dissn of CH3CO from acetone photolysis. It
evidences distinct dissn dynamics of acetyl
approached from the vinoxy?acetyl isomerization
TS.
48
Which vinoxy X(2A?) dissociate to the CH3 CO
product channel?
isomerize?
CH3 CO
CH3CO
Vinoxy radicals from these C-Cl fissions
do. Lets look at their internal energies.
49
Only vinoxy radicals with internal energy greater
than 41?2 kcal/mol isomerize and dissociate to
acetyl?CH3CO .
This agrees with the calculated G3//B3LYP and
G2Q barrier.
50
Does vinoxy X(2A?) branch to the ketene H
product channel?
(mass 42)
?
H
Cl signal was .0095 cts/laser shot This signal
is .00044 cts/laser shot (integrated for 5.4
million laser shots)
m/e42 15o,10.6 eV
pulsed beam bkgd
51
Does vinoxy X(2A?) branch to the ketene H
product channel?
(mass 42)
?
H
NO! The upper bound of H Ketene is less than a
percent! (fit shown assumes a 1.6 quantum yield)
m/e42 15o,10.6 eV
pulsed beam bkgd
52
What, then, is the source of the m/e42 signal?
Another source of ketene?
(mass 42)
?
A 2nd dissn channel of the precursor?
HCl
m/e42 15o,10.6 eV
pulsed beam bkgd
53
What, then, is the source of the m/e42 signal?
Another source of ketene?
(mass 42)
?
A 2nd dissn channel of the precursor?
HCl
54
What, then, is the source of the m/e42 signal?
Another source of ketene?
(mass 42)
?
A 2nd dissn channel of the precusor?
HCl
YES.
55
Our expts produce vinoxy radicals in the ground
electronic state, so can test whether the
conventional assumption that the
electronic wavefunction readjusts rapidly as the
nuclear dynamics evolves on the ground state PES
gives a good prediction for the product branching.
TS1 41.8 kcal/mol
H
ketene X(1A1)
a'
a?
vinoxy X(2A?)
a?
56
The RRKM prediction for the H ketene CH3
CO product branching for the E distribution in
our work is 43
57
The 43 RRKM prediction for the H ketene CH3
CO product branching would give a signal to
slightly faster arrival times of the ketene
signal from HCl elimination - we know the
magnitude of the predicted ketene signal from
vinoxy (assumes ketene from each process has the
same photoionization cross section).
ketene signal from HCl, expand time-axis...
58
The 43 RRKM prediction for the H ketene CH3
CO product branching would give a signal to
slightly faster arrival times of the ketene
signal from HCl elimination - we know the
magnitude of the predicted ketene signal from
vinoxy (assumes ketene from each process has the
same photoionization cross section).
ketene signal from HCl, change y axis...
59
The 43 RRKM prediction for the H ketene CH3
CO product branching would give a signal to
slightly faster arrival times of the ketene
signal from HCl elimination - we know the
magnitude of the predicted ketene signal from
vinoxy (assumes ketene from each process has the
same photoionization cross section).
ketene signal from HCl
60
The 43 RRKM prediction for the H ketene CH3
CO product branching would give a signal to
slightly faster arrival times of the ketene
signal from HCl elimination - we know the
magnitude of the predicted ketene signal from
vinoxy (assumes ketene from each process has the
same photoionization cross section).
RRKM predicted ketene signal from vinoxy
The RRKM prediction for the branching to ketene
H from vinoxy is a factor of about 200 too
high! (if we scale it down x 200 it could be
buried in the background)
ketene signal from HCl
61
The 43 RRKM prediction for the H ketene CH3
CO product branching would give a signal to
slightly faster arrival times of the ketene
signal from HCl elimination - we know the
magnitude of the predicted ketene signal from
vinoxy (assumes ketene from each process has the
same photoionization cross section).
RRKM predicted ketene signal from vinoxy
The RRKM prediction for the branching to ketene
H from vinoxy is a factor of about 200 too
high! (if we scale it down x 200 it could be
buried in the background)
ketene signal from HCl
62
The RRKM prediction for the H ketene CH3
CO product branching was 43 . . .
63
But DATA SHOWED branching to H ketene is LOWER
BY x 200!
64
In regions where ?elec(R) changes (H-E)X(R)
H
3A?
A?
Coupling causes hop to other PES when
a'
a?
A
A
H
(1a?)2(10a)1(2a?)2
ketene X(1A1)
(1a?)2(10a)2(2a?)1
A?
vinoxy X(2A?)
65
In regions where ?elec(R) changes (H-E)X(R)
H
3A?
A?
Coupling causes hop to other PES when
a'
a?
A
A
H
(1a?)2(10a)1(2a?)2
ketene X(1A1)
(1a?)2(10a)2(2a?)1
A?
vinoxy X(2A?)
66
(No Transcript)