Chapter 3. Designing a PEI spectrometer for He(I) light source
3.3 Conclusion
From experimental and computational examination of various designs of the acceleration electrodes, the following conclusions were drawn.
1. The major source of background electrons is the photoemission from the repeller plate.
Therefore, the repeller plate should be placed away from the ionization region and a small retardation field may be applied to prevent photoelectrons emitted from the repeller to be transmitted to the acceleration region.
2. All insulators should be shielded to avoid charging.
3. Energy resolution (E/E) of higher than 1 % is obtainable even for a large ionization volume (i.e. of the order of millimeters). The acceleration electric field strength should be weak.
53
4. The resolution of an imaging system should be sufficiently high. In our case, a super-resolution (4096 4096 pixels) imaging system was needed to fully exploit the performance of our charged particle optics. The best energy resolution was 0.735 % at 5.461 eV (FWHM: 40 meV).
54
Figure 3.1 Photoelectron images observed by He(I) photoionization of a supersonic beam of Ar: (a) 6 mm aperture of He(I) light source was used. (b) 2 mm aperture of He(I) light source was used.
55
Figure 3.2 (a)-(c) Intensity-integrated raw photoelectron images of Kr at a stagnation pressure of 2.2 MPa at different electrode gap distance of 6 mm, 9 mm and 12 mm. (d denotes the electrode gap distance, apt denotes the aperture diameter of He(I) light source). denotes the directions of the polarization vectors of the VUV beam.
56
Figure 3.3 Difference between the anisotropy parameter calculated for Kr (2P3/2) by assuming different polarization degrees of light and the literature value 1.24.[ref. 1] The polarization degree of 74 % provides the best agreement.
57
Figure 3.4 Photoelectron kinetic energy distributions in He(I) photoionization of supersonic beams of Ar generated at different stagnation pressures of 0.4 (solid), 1.1 (dash dot), and 2.2 (dash) MPa.
58
Figure 3.5 (a)–(c) Intensity-integrated raw photoelectron images observed by He(I) photoionization of a supersonic beam of pure Ar, Kr and N2, respectively. (d)–(f) Photoelectron images of Ar, Kr and N2 after background subtraction. (g)–(i) 2D slice images of Ar, Kr and N2 obtained with the modified p-BASEX reconstruction. denotes the directions of the polarization vectors of the VUV beam.
59
Figure 3.6 (a) Photoelectron kinetic energy distribution in He(I) photoionization of supersonic beams of Ar. (b) Photoelectron kinetic energy distribution in He(I) photoionization of supersonic beams of Kr. The inset shows an expanded view in the PKE region between 6.0 and 8.0 eV. The solid red line indicates the best-fit Gaussian to the observed data; it has a FWHM of 246 meV at 7.22 eV.
60
Figure 3.7 (a) Photoelectron kinetic energy distribution in He(I) photoionization of supersonic beams of N2. The spectral feature agrees reasonable well with the literature [Ref.
3]. (b) Expanded view of He(I) photoelectron spectrum of N2 in the A band region. The solid red and dotted green lines show the overall distribution and individual Gaussian components obtained by least-squares fitting.
61
Figure 3.8 (a) 2D representation of the cylindrically symmetric potential in the VMI spectrometer. (b) Experimental geometry for photoelectron imaging experiment. The molecular beam and time-of-flight axis are located along the x axes; VUV light beam is located along y axis.
62
Figure 3.9 The simulated velocity resolution (/)y (black solid square) and (/)z (red solid circle) as a function of electrode spacing (upper panel) and the resolution difference [(/)y - (/)z] (lower panel).
1.5 2.0 2.5 3.0 3.5 4.0
36 38 40 42 44 46
0.8 1.0 1.2 1.4
V el oci ty resol uti on ( ) / % (
y(
z(
y- (
zSpacing / mm
Resolution diff
63
Figure 3.10 The simulated velocity resolution (/)y (black solid square) and (/)z (red solid circle) as a function of hole diameter (upper panel) and the resolution difference [(/)y - (/)z] (lower panel).
2.2 2.4 2.6 2.8
54 56 58 60 62 64 66
0.5 0.6 0.7
V el oci ty resol uti on ( ) / %
(
y
(
z(
y- (
zInner diameter / mm
Resolution diff
64
Figure 3.11 Cross-sectional view of the electrostatic lens system using three-electrode configuration (all units in millimeters). A molecular beam is introduced from the bottom and is irradiated by He(I) radiation at the position indicated by the cross ().
65
Figure 3.12 Velocity resolution simulated by performing trajectory calculation with 3D Simion software as a function of Vext/Vrep: (a) He(I) experiment condition: light beam diameter
= 4.2 mm. (b) FEL experiment condition: light beam diameter = 0.1 mm.
0.745 0.750 0.755 0.760 0.765 0.770
66
Figure 3.13 (a)-(c) Photoelectron images of Ar using the three-electrode configuration with solid repeller electrode. (d)-(f) Photoelectron images of Ar after replaced the solid repeller electrode to mesh repeller electrode.
67
Figure3.14 Cross-sectional view of the electrostatic lens system (all units in millimeters). A molecular beam is introduced from the bottom and is irradiated by He(I) radiation at the position indicated by the cross (). An example of the voltage setting in the measurement of photoelectron image of Kr is also shown in the figure.
68
Figure 3.15 (a) Simulated trajectory (red, blue and orange) of electrons emitted from the mesh electrode and photoelectrons from the photoionization point (black). The equipotential lines are also shown (light brown). (b) 2D representation of the cylindrically symmetric potential Φ(x,y) of (a).
69
Figure 3.16 Velocity resolution simulated by performing trajectory calculation with 3D Simion software as a function of (molecular beam dia./VUV beam dia.) ratio and VUV beam diameter is fixed at 4.3 mm. Noted that solid symbol indicated the ionization volume aligned along the VUV beam axis (Y axis), open symbol indicated the ionization volume aligned along the molecular beam axis (X axis), as indicated at the top of the figure.
70
Figure 3.17 Velocity resolution simulated by performing trajectory calculation with 3D Simion software as a function of Vext/Vrep: He(I) experiment with light beam diameter of 4.2 mm are indicated as solid symbol. FEL experiment with light beam diameter of 0.1 mm are indicated as open symbol.
0.855 0.856 0.857 0.858
0.0 0.1 0.2 0.3 0.4 0.5 0.6
()y - FEL ()z - FEL
Ve lo ci ty re so lu tio n (
) / %
V
ext/ V
rep@V
rep= -4000 V
()y - He(I) ()z - He(I)
71
Figure 3.18 Photoelectron images of Ar using a new design electrode. (a) before background subtraction. (b) background image observed without the sample gas. (c) after background subtraction. The distortion of the image was observed.
72
Figure 3.19 (b) and (d) are photoelectron images of Ar measuring with electrode (a) and (c), respectively.
73
Figure 3.20 (a) The configuration of the spectrometer including a charged insulator in 3D drawing. (b) The images simulated in three conditions: (i) no charged insulator (black), (ii) charged insulator with -800V (red) and (iii) charged insulator with -50 V (blue).
74
Figure 3.21 (a) The configuration of the spectrometer including a charged insulator in 3D drawing. (b) The images simulated in three conditions: (i) no charged insulator (black), (ii) charged insulator with -800V (red) and (iii) charged insulator with -50 V (blue).
75
Figure 3.22 Cross-sectional view of the final design of the electrostatic lens system (all units in millimeters). A molecular beam is introduced from the bottom and is irradiated by He(I) radiation at the position indicated by the cross ().An example of the voltage setting in the measurement of photoelectron image of Kr is also shown in the figure.
76
Figure 3.23 Velocity resolution simulated by performing trajectory calculation with 3D Simion software as a function of Vport/Vrep and Vext/Vrep.
77
Figure 3.24 (a) The geometries of the simplified electrodes. (b) Calculated velocity resolution as a function of the number of lens electrodes behind the extractor. The ionization volume was assumed to be 3.4 ϕ 4.2 mm.
78
Figure 3.25 Photoelectron images of Ar using a new design electrode with shielding all insulators. (a) before background subtraction. (b) background image observed without the sample gas. (c) after background subtraction. No distortion of the image was observed.
79
Figure 3.26 Simulated energy resolution (E) as a function of photoelectron kinetic energy using our newly designed electrode with different number of lens electrodes. The result of conventional three-electrode design is also shown (open green triangle) in the figure.
0 1 2 3 4 5 6
0.0 0.1 0.2 0.3 0.4
Energy res olut ion E (eV)
Photoelectron kinetic energy (eV)
7 lens electrodes
6 lens electrodes
5 lens electrodes
4 lens electrodes
3 lens electrodes
2 lens electrodes
1 lens electrodes
conventional 3 elect.
80
Figure 3.27 PKE distributions determined by He(I) PEI of supersonic beams of Kr using a (512 512 pixels) CCD camera. The inset shows an expanded view in the PKE region between 6.0 and 8.0 eV. The solid red line indicates the best-fit Gaussian to the observed data;
it has a FWHM of 210 meV at 7.22 eV.
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
0.0 0.2 0.4 0.6 0.8 1.0
2
P
1/2In te n sit y ( a .u .)
Photoelectron kinetic energy (eV)
2
P
3/26.0 6.5 7.0 7.5 8.0
FWHM
210 meV
81
Figure 3.28 Velocity resolution evaluated using photoelectron imaging of Kr as a function of Vext/Vrep. A low-resolution (512 512 pixels) CCD camera was used without image processing. Due to insufficient resolution of the camera, the focusing curve is dull. The inset shows velocity resolution evaluated using photoelectron imaging of Ar with super-resolution (4096 4096 pixels) imaging system. The focusing curve becomes much sharper and the resolution varies as a V-shape for the Vext/Vrep ratio. The error bar corresponds to the fitting error (fwhm).
82
Figure 3.29 PKE distributions determined by He(I) PEI of supersonic beams of Ar using a super-resolution (4096 4096 pixels) imaging system. The inset shows an expanded view in the PKE region between 5.1 and 5.6 eV. The solid red line indicates the best-fit Gaussian to the obeerved data; it has a FWHM of 40 meV at 5.461 eV.
0 1 2 3 4 5 6
0.0 0.2 0.4 0.6 0.8 1.0
Int ens iy (a. u. )
Photoelectron kinetic energy (eV)
5.1 5.2 5.3 5.4 5.5 5.6
2
P
1/22
P
3/2FWHM
40 meV
83
Figure 3.30 The definition of the individual four quadrants of the image is shown in the inset.
The observed PAD of (a) 2P1/2 and (b) 2P3/2 of Kr, obtained from each quadrant and the corresponding least-squares fits are shown with open circles and solid lines. The determined anisotropy parameters, the average and the standard deviation obtained in each quadrant are also shown.
84
Table 3.1 Anisotropy parameter β for krypton as obtained in this work in comparison with the literature values.
β
Ionic state eBE(eV) This work Previous reports
2P3/2 14.00 1.23 1.24(2), 1 1.30(4),7 1.30(5),8 1.29(5),9 1.37(2),10 1.20(5)11
2P1/2 14.65 1.09 1.21(2),1 1.23(4),7 1.23(5),8 1.25(5),9 1.37(3),10 1.20(5)11
Errors (±, in the last digit unless indicated otherwise) given in parentheses.
Table 3.2 Anisotropy parameter β for argon as obtained in this work in comparison with the literature values.
β
Ionic state eBE(eV) This work Previous reports
2P3/2 15.76
Errors (±, in the last digit unless indicated otherwise) given in parentheses.
85
3.4 References
1 Kreile, J.; Kurland, H. D.; Seibel, W.; Schweig, A. Nucl. Instr. and Meth. 1991, 308 621.
7 Kreile, J.; Schweig, A. J. Electron Spectrosc. Relat. Phenom. 1980, 20, 191.
8 Karlsson, L.; Mattson, L.; Jadrny, R.; Siegbahn, K.; Thimm, K. Phys. Lett. A 1976, 58, 6.
9 Dehmer, J. L.; Chupka, W. A.; Berkowitz, J.; Jivery, W. T. Phys. Rev. A 1975, 12, 1966.
10 Niehaus, A.; Ruf, M. W. Z. Phys. 1972, 252, 84.
11 Carlson, T. A.; Jonas, A. E. J Chem. Phys. 1971, 55, 4913.
12 Kibel, M. H.; Leng, F. J.; Nyberg, G. L. J Electron Spectrosc. Relat. Phenom. 1979, 15, 281.
13 Mason, D. C.; Mintz, D. M.; Kuppermann, A. Rev. Sci. Instrum. 1977, 48, 926.
14 Hancock, W. H.; Samson, J. A. R. J Electron Spectrosc. Relat. Phenom. 1976, 9, 211.
86
Chapter 4
He(I) Photoelectron Imaging of Benzene and Pyridine
The preceding chapters have focused on the designing of a photoelectron imaging spectrometer. In this chapter, He(I) ultraviolet photoelectron imaging of benzene and pyridine in supersonic molecular beams are discussed.
4.1 He(I) Ultraviolet Photoelectron Imaging of Benzene
4.1.1 Introduction
Benzene (C6H6) is one of the most fundamental polyatomic molecules, and its photoelectron spectroscopy has been performed using UV1-14 and X-ray light sources.15-19 The configuration of valence electrons in benzene is
[2e2g()] [3a1g()] [2b1u()] [1b2u()] [3e1u()] [1a2u()] [3e2g()] [1e1g()]
in the ascending order of the orbital energy. Theoretical studies predict that eight ionic states can be observed by photoionization of benzene with He(I) radiation (21.22eV). There were some controversies in the assignment of observed He(I) photoelectron spectrum of benzene1,20. One controversy was the assignments of the D1 and D2 states between 11 and 13 eV. Lindholm et al.6 suggested the electron removals from the 3e2g() and 1a2u() orbitals as D2[A2E2g] and D1[B2A2u] states of C6H6+
, respectively. Although this assignment has been supported by a number of researchers,4,7,16,21,22 Turner et al.1 and Price et al.5,23 preferred the reversed assignment. Carlson et al.4 measured the angle-resolved He(I) photoelectron spectrum and found that the anisotropy parameter gradually increases between 11 and 13 eV.
They suggested D1[A2E2g] and D2[B2A2u]. This assignment has been further supported by
87
Mattsson et al.24 and Sell et al.25 Carlson et al.26 and Baltzer et al.27 measured the energy dependence of the anisotropy parameters using synchrotron radiation and compared with calculations based on the multiple scattering X method. These studies established D1[A2E2g] and D2[B2A2u]. Another controversy concerned the assignments of the band at 15.5 eV.
Jonsson et al.28 assigned this band as the electron removals from the 2b1u() orbital, whereas Price et al.5 assigned it to 3a1g(). This problem was settled by Gelius et al.16,23 who supported Jonsson based on the intensity analysis of the X-ray photoelectron spectrum.
4.1.2 Interference from the residual water vapor in the chamber
We performed PEI experiments using continuous supersonic beams of polyatomic molecules. However, the signal level was not much higher than the background photoionization signal of the residual water vapor in the chamber. The partial pressure of water in the ionization camber was measured to be ~ 2 10-8 Torr using a residual gas analyzer (Stanford Research Systems, RGA200). Figure 4.1 (a) shows a photoelectron image obtained with a continuous beam of benzene 8 % in He expanded with a stagnation pressure of 1.2 atm; the signal and background images were integrated for 100 min, and the background image obtained without the molecular beam has already been subtracted. This image exhibits a vertical band along the VUV light path, which is due to photoionization of water vapor. The photoelectrons signal of water remains in Fig. 4.1 (a) even after the background subtraction, indicating water vapor pressure changed with and without the molecular beam. Figures 4.1 (b) and (c) show ion images observed with and without the molecular beam, respectively; ions were not mass-selected. These images spatial distribution of ions along the light path of He(I) radiation. Bright elliptical spot in the center is due to benzene ions produced in the molecular beam, whereas the vertical band arises from water ions.
88
4.1.3 Results
To enhance the signal/background contrast ratio, we replaced a continuous beam with a pulsed beam and we added a turbo molecular pump (650 L/s) to the ionization chamber. This allowed us to increase the ratio of the partial pressure of benzene against water vapor during the gas pulse. We time-gated both CCD and MCP so as to detect photoelectrons only when a pulsed beam was introduced into the ionization region. In order to synchronize the readout of the CCD camera with a gas pulse, we used a low-resolution CCD camera (512 512 pixels) that accepts an external trigger. Figure 4.2 (a) shows a photoelectron image measured with a pulsed supersonic beam of benzene 18 % seeded in He expanded at a stagnation pressure of 0.55 atm; the signal and background images were integrated for 90 min at a data acquisition rate of 20 Hz, and the latter was subtracted from the former to obtain Fig. 4.2(a). It is seen that the background photoelectron signal is considerably smaller than the case using a continuous molecular beam. As the vacuum system can allow 100 Hz of gas pulses at a stagnation pressure of 0.55 atm, the duty cycle of our PEI system can be easily increased by a factor of five by not synchronizing the CCD camera.
Figure 4.2 (b) shows the slice image obtained by taking the inverse Abel transform of Fig.
4.2 (a). As mentioned in chapter 1, photoionization with unpolarized light generates a PAD that is cylindrically symmetric about the light propagation direction (indicated by the arrow in the figure). Figure 4.2 (c) shows the energy-dependent anisotropy parameter determined from the image. Although the concentration of benzene (18 %) in the pulsed beam was high, no signature from benzene dimers has been identified at ionization energy (IE) of 8.65 eV29 in our photoelectron spectrum. No vibrational structure was resolved with our energy resolution of 0.2–0.3 eV. The vibrationally-resolved He(I) photoelectron spectrum of the first band has been reported by Baltzer et al.20; however, has not been determined at a vibrational resolution. The energy-dependent anisotropy parameter determined in the present study is in excellent agreement with the literature as presented in Fig. 4.2(c) and Table 4.1.11,12,19 In the
89
lowest photoelectron band, decreases with the electron binding energy. Carlson and Anderson4 have observed the same feature and suggested that it may be due to the Jahn–Teller effect in D0. However, the influence of Jahn–Teller splitting on has not been elucidated experimentally or theoretically for any molecular system.30 Another possible origin for the variation of is a Coulomb phase. Since for photoionization from the 1e1g() orbital increases with increasing PKE,19,20,31 is expected to diminish at higher binding energies (i.e., lower PKEs) within the first band.
4.2 He(I) Ultraviolet Photoelectron Imaging of Pyridine
4.2.1 Introduction
Pyridine (C5H5N) is an important model system in relation to biologically active nicotinic acid and the nucleotides of cytosine, uracil, and thymine. It is also an example of heterocyclic aromatic molecules. The replacement of a carbon atom with a nitrogen atom in a benzene ring creates a lone-pair orbital. The lowest photoelectron band of pyridine has
These early photoelectron experiments suggested all three possible assignments of
--n,1, 35 -n-,36 and n--37 - 40. The resolution of this confusion required additional information. Utsunomiya et al.41 performed the first angle-resolved PES of pyridine in 1978 and found that near the ionization threshold increases from 0.2 to ~ 0.6 in the middle of the first band. They referred that ionization to the n-1 state of 2,6-lutidine (dimethyl-pyridine)
90 in agreement with experimental studies. MRDCI predict that the first two ionic states have an energy difference of 0.7 eV, whereas the other methods predict differences less than 0.2 eV.
So far, angle-resolved PES of pyridine has been limited to the first three ionic states. In this study, the energy-dependent anisotropy parameter is determined for the entire energy region accessible with He(I) radiation.
4.2.2 Results
Figure 4.3 (a) shows a photoelectron image measured with a pulsed supersonic beam of 10 % pyridine seeded in He. The background image has been subtracted. The left half is the raw image and the right half is the slice image obtained by inverse Abel transform. Figure 4.3 (b) shows the photoelectron energy spectrum and extracted from the slice image. The numerical values of are also compared with the literature41,42 in Table 4.2. A microdischarge of MCP occurred when a high density molecular beam impinged on the detector, we needed to restrict the stagnation pressure for supersonic expansion to be 0.2 atm. (The observed microdischarge was a specific problem of this detector.) Because of the low stagnation pressure of 0.2 atm, a long (3 hours) integration time was needed to measure the image. In the future, microdischarge should be avoided by redesigning the PEI spectrometer so that a molecular beam travells parallel to the face of the imaging detector.
As seen in Fig. 4.3 (b), of the first band between 9.2 and 10.2 eV increases with the binding energy in agreement with Utsunomiya et al.41 and Piancastelli et al.,42. This band has
91
contribution of D0 and D1 associated with electron removal from 11a1(n) and 1a2() orbitals, respectively. In He(I) UPS of pyridine, the first band corresponds to PKE between 12.02 and 11.02 eV. In this PKE region, increases only gradually with PKE if it is due to the energy-dependent Coulomb phases.31 The observed increases rather rapidly with the electron binding energy (i.e. decreasing PKE), which indicates that the variation of in the first band is not due to Coulomb phases and rather due to D1 overlap with D0. The origin of D1 has not been determined yet.
The fourth and fifth bands between 12.5 and 13.5 eV have been assigned to ionization from b2 and b1, respectively.41 As far as we know, has not been determined at the electron binding energies over 14 eV. The variation of in this region is similar to that of benzene; has a deep minimum at about 15.8 eV for pyridine and 14.9 eV for benzene. All these cation states are due to the removal of an electron from the orbitals (see Fig. 4.3). beam by introduction of a pulsed beam. The major source of the background photoelectrons are from residual water vapor in the photoioniztion chamber. The photoelectron anisotropy parameters determined for benzene and pyridine were in good agreement with the literatures.
92
The anisotropy parameters were determined for the first time for pyridine at the electron binding energies over 14 eV. Higher energy resolution is obtainable in principle by centroiding calculations of the light spots on the phosphor screen of the detector. However, our super-resolution imaging system is currently able to handle up to 256 light spots in a single frame (30 frames/s), which is too low for He(I) PEI. This problem will be solved in the future when a higher frame rate of a camera and a faster digital image processing circuit become available.
93
Figure 4.1 (a) Photoelectron image of a continuous supersonic beam of 8% benzene seeded in He (the background image has been subtracted). (b) Ion image with the supersonic molecular beam. (c) Ion image without the molecular beam. The arrow indicates the propagation direction of the He(I) radiation.
94
Figure 4.2 (a) Symmetrized photoelectron image of a pulsed supersonic beam of benzene. (b) the slice image obtained from (a). (c) The energy-dependent anisotropy parameters (upper panel, solid circle) and photoelectron spectra (lower panel) obtained from (b). Data from [ref.
12], [ref. 19] and [ref. 11] are given as open squares (□), open triangles () and open inverted triangles (), respectively. The assignments of the ionized orbitals are indicated at the vertical ionization point [ref. 19]. The error bars are the fitting errors.
95
Figure 4.3 (a) Symmetrized photoelectron image of a pulsed supersonic beam of 10%
pyridine seeded in He (left half) and the slice image taking the inverse Abel transform (right half). (b) The energy-dependent anisotropy parameter (upper panel, solid circles) and photoelectron spectra (lower panel) obtained from (a). Data from [ref. 42] and [ref.41] are given as open squares (□) and open triangles (), respectively. The assignments of the ionized orbitals are indicated at the vertical ionization point [ref. 44]. The error bars are the fitting errors.
96
Table 4.1 Electron binding energy (in eV) and anisotropy parameters () for benzene
This study Sell et al.c Carlson et al.d Mattsson et al.f
97
98
99
Table 4.1 (Continued)
This study Sell et al. Carlson et al. Mattsson et al.
Orbital eBE eBE eBE eBE
19.13 0.09(5)
19.19 0.16(5) 19.20 0.06(12)e
19.25 0.15(5) 19.31 0.09(6) 19.37 0.03(6)
aThe orbital assignments for benzene are those in ref. [19]. bErrors (, in last digit unless indicated otherwise) given in parentheses. cThe energy-dependent anisotropy parameters are reproduced from ref. [25]. dThe energy-dependent anisotropy parameters are reproduced from ref. [19]. e was determined by integrating over the area of the 2e2g() band. fReference [24].
100
Table 4.2 Electron binding energy (in eV) and anisotropy parameters () for pyridine.
this study Utsunomiya et al.c Piancastelli et al.d
orbitala eBE eBE eBE
101
Table 4.2 (Continued)
this study Utsunomiya et al. Piancastelli et al.
orbital eBE eBE eBE
102
Table 4.2 (Continued)
this study Utsunomiya et al. Piancastelli et al.
orbital eBE eBE eBE
103
aThe orbital assignments for benzene are those made in ref. [44]. bErrors (, in last digit
aThe orbital assignments for benzene are those made in ref. [44]. bErrors (, in last digit