Applications of photoelectron velocity map imaging at high resolution

or

Photoionization dynamics of NH3 (B1E)

Katharine L. Reid

(Paul Hockett, Mick Staniforth)

University of Nottingham, United Kingdom

The importance of having control over the probe wavelength in pump-probe experiments

In our study of the photoionization dynamics of excited state ammonia, both of these features have been crucial.

(i) Control over resolution (see Alistair Green poster – intramolecular dynamics in toluene)

(ii) Awareness of continuum resonances (see Mick Staniforth poster – observation of shape resonances in small aromatic molecules)

In the absence of a molecular frame measurement a determination of the radial dipole matrix elements controlling an ionization process requires photoelectron angular distributions (PADs) to be measured for rotationally resolved states of the ion.

This is a challenge for photoelectron spectroscopy!

The highest resolution method of ion spectroscopy (ZEKE) does not allow the measurement of PADs

Can we achieve rotational resolution in small polyatomic ions using imaging?

Slow electron velocity map imaging*

*Osterwalder et al., JCP 121, 6317 (2004)

NH3

Approaching the molecular frame

X1A1 (v2 = 0; J = 1, K = 1) B1E (v2 = 4; J = 3, K = 2)

REMPI spectrum

X1A1 (v2 = 0) B1E (v2 = 4)

One-colour vs two-colour photoelectron images

X1A1 (v2 = 0; J = 1, K = 1) B1E (v2 = 4; J = 3, K = 2)

One-colour photoelectron spectrum

Choosing the probe wavelength

A closer look ...

The 11 ion rotational state vs probe energy

In what follows, a probe wavelength of 431.3 nm (Eprobe = 23186 cm-1) was chosen.

Rotationally resolved photoelectron images

X1A1 (v2 = 0) B1E (v2 = 4)

Extracted photoelectron spectra

Images for different B state (v2 = 4) rotational levels

X1A1 (v2 = 0; J = 1, K = 1) B1E (v2 = 4; J = 3, K = 2)

All extracted photoelectron angular distributions for v2 = 4

solid line is a fit to )()()()()( 6060404020200000 YYYYI

Expectations

The B state is predominantly p with some d character. The proportions are not known.

This leads to the expectation that the photoionization dynamics will be dominated by d and s waves, with a small contribution from p and f (atomic picture).

Previous work (our group, Softley and coworkers) has made assumptions about the dynamics and concluded there is no contribution from higher l partial waves.

For the LM terms

|l - l| L l + l

Any L = 6 contribution to a PAD must come from l ≥ 3 (f wave or higher). However, in order to make ion states with K = 1 or 5 selection rules require that l must be even.

Several strong ion peaks assigned to K = 1 and associated with PADs having |0| >> 0 are observed.

Therefore the ionization dynamics must involve g waves.

This information is missing in a ZEKE experiment …

By inspection:

There is sufficient data to fit the LM parameters to a model to determine the radial dipole matrix elements

1/2

1/2

( ) ( 1)

2 1)(2 1)(2 1)(2 1)(2 1)

(4 )

0 0 0

( , , , , ) ( ,

( , , , ) t t

t t

m K N MLM

ll mm N N K

t t t

t t t

t

l lii i

N N l L Ki

N N Kl l L l l L N N NiM M Mm m M N N Ki

C l m N C l

B N K N K

, , , ) exp[ ( )]t KM l l l lm N T r r i

00

( , , , )( , , , )

( , , , )LM

LM LMi i

i ii i

B N K N KN K N K

B N K N K

00 00( , , , )i iB N K N K

where

Based on Dixit and McKoy, 1985

Results of the fit: PADs

11 32

11 22

Results of the fit: prediction of photoelectron spectra

s p d f g

12.7 - 13.0 1.9 15.0 36.1 - 0.7 2.1 8.0 2.9 7.6 0.0

s p d f g

0 - 0 31 16 149 - 162 153 153 92 64 -

s p d f g

12.7 13.0 53.0 10.7 10.5

Results of the fit: parameters

Phases / degrees

Hockett et al. PRL, 102, 253002 (2009)

Squared radial dipole matrix elements / %

Simulating PADs following excitation of v2 = 3 using the parameters deduced from v2 = 4

We can also predict what we would expect to see if the excitation and ionization beams were

perpendicularly polarized

X1A1 (v2 = 0; J = 1, K = 1) B1E (v2 = 4; J = 3, K = 2)

Perpendicular polarizations: PADs at different detection angles

predicted measured

Tomographic reconstruction of angular distributions taken with perpendicular polarization geometries

(Thanks to Thomas Baumert!)

31 ion state

Raw reconstruction

Fit to spherical harmonics

00 20 20 22 22

44 4440 40 42 42

60 60 62 62 64 64

( , ) {1 ( , ) 2 ( ,0)cos2

( , ) 2 ( ,0)cos2 2 ( ,0)cos4

( , ) 2 ( ,0)cos2 2 ( ,0)cos4 }

I Y Y

Y Y Y

Y Y Y

Dependence of PADs in the perpendicular polarization geometry on ion rotational state

Reconstructed and fitted

Predicted

Conclusions

1. It is possible to achieve rotational resolution of NH3+ using VMI

2. We have measured the first rotationally resolved PADs for a polyatomic molecule

3. Control over probe wavelength is essential to optimize resolution and to avoid (or study) resonances

4. Our fit to radial dipole matrix elements appears to be robust (changing vibrational level and polarization geometry)

5. PADs reveal dynamics that are missed by ZEKE

6. We have approached the molecular frame through use of alignment by excitation. However, we are missing the even-odd l phase differences which can only be obtained when parity is broken

Prospects

This is probably the end of the line for rotational resolution in polyatomic molecules. However, we can make use of:

(i) Rotational coherence spectroscopy

(ii) Strong field alignment

In both cases the ability to treat different polarization geometries will be essential.

Acknowledgements

PAUL HOCKETT

Mick Staniforth

Dave Townsend

EPSRC