A facility for the analysis of the electronic structures of solids and their surfaces by synchrotron radiation photoelectron spectroscopy

TLDR: The beamline design and its performance allow for a highly productive and precise use of the ARPES technique at an energy resolution of 10-15 meV for fast k-space mapping studies with a photon flux up to 2 ⋅ 1013 ph/s and well below 3 mev for high resolution spectra.

Abstract: A synchrotron radiation beamline in the photon energy range of 18-240 eV and an electron spectroscopy end station have been constructed at the 3 GeV Diamond Light Source storage ring. The instrument features a variable polarisation undulator, a high resolution monochromator, a re-focussing system to form a beam spot of 50 × 50 μm2, and an end station for angle-resolved photoelectron spectroscopy (ARPES) including a 6-degrees-of-freedom cryogenic sample manipulator. The beamline design and its performance allow for a highly productive and precise use of the ARPES technique at an energy resolution of 10-15 meV for fast k-space mapping studies with a photon flux up to 2 ⋅ 1013 ph/s and well below 3 meV for high resolution spectra.

Figures

FIG. 5. Estimated flux at the sample on the HR-branch, calculated for an exit slit opening of 100 µm and using the source flux from Fig. 2.

FIG. 6. Measured photoionization yield spectra of (a) Ne absorption and (b) He absorption in a gas cell at an exit slit height h = 0.03 mm using the grating of 800 lines/mm and a monochromator c-value of 2.25. In (b) the peaks are maked following notations in Ref. 16. (c) He gas absorption spectra at various h. The fits in (a), the inset of (b), and (c) are used to determine the energy resolution (fits shown as blue lines). (d) Extracted energy resolution, fit by ∆E = (1.3+h ·86) meV, with h in mm. The inset shows the flux. All data are measured using LH polarisation.

FIG. 7. Measured photon flux after the last mirror of the HR-branch. The undulator and cPGM with the grating of 800 lines/mm were operated at standard settings with a 6×6 mm beamline aperture and a c-value of 2.25. The exit slit was set to 100 µm.

FIG. 11. Rendering of the manipulator head. The three rotation axes with their respective angular ranges are indicated. The main support structure is machined in phosphor-bronze and shown in light blue. All thermal links (brown) are made in OFHC copper. Rotating parts are supported in polyimide (grade 2391 Tecasint) bushings. During ARPES measurements, the manipulator head is fully shielded with gold coated copper foil (not shown) to minimize the radiative heat load on the sample and distortions of the electrostatic field between the sample and analyzer entrance lens.

FIG. 8. Measured photoemission spectrum from polycrystalline gold using a photon energy hν = 20 eV and exit slit h = 10 µm, analysed as described in the text. A Gaussian function representing the combined experimental energy resolution of 2.6 meV is shown in grey.

FIG. 9. Measured spectrum of photoemission from polycrystalline gold for photon energy 25 eV (a) and 50 eV ((b) and (c)). The inset on the right side of each panel shows the same data multiplied by a factor 20. In (c) the beam passes through a 150 nm thick Al filter. The intensity scales in (b) and (c) are normalised to the same absolute values. The labelling of peaks is described in the text.

Full text

Article
Reference
A facility for the analysis of the electronic structures of solids and their
surfaces by synchrotron radiation photoelectron spectroscopy
HOESCH, M., et al.
Abstract
A synchrotron radiation beamline in the photon energy range of 18-240 eV and an electron
spectroscopy end station have been constructed at the 3 GeV Diamond Light Source storage
ring. The instrument features a variable polarisation undulator, a high resolution
monochromator, a re-focussing system to form a beam spot of 50 × 50 μm2, and an end
station for angle-resolved photoelectron spectroscopy (ARPES) including a
6-degrees-of-freedom cryogenic sample manipulator. The beamline design and its
performance allow for a highly productive and precise use of the ARPES technique at an
energy resolution of 10-15 meV for fast k-space mapping studies with a photon flux up to 2 ⋅
1013 ph/s and well below 3 meV for high resolution spectra.
HOESCH, M., et al. A facility for the analysis of the electronic structures of solids and their
surfaces by synchrotron radiation photoelectron spectroscopy. Review of Scientific
Instruments, 2017, vol. 88, no. 1, p. 013106
DOI : 10.1063/1.4973562
Available at:
http://archive-ouverte.unige.ch/unige:101914
Disclaimer: layout of this document may differ from the published version.
1 / 1

A facility for the analysis of the electronic structures of solids and their surfaces by
synchrotron radiation photoelectron spectroscopy
M. Hoesch, T. K. Kim, P. Dudin, H. Wang, S. Scott, P. Harris, S. Patel, M. Matthews, D. Hawkins, S. G.
Alcock, T. Richter, J. J. Mudd, M. Basham, L. Pratt, P. Leicester, E. C. Longhi, A. Tamai, and F. Baumberger
Citation: Review of Scientific Instruments 88, 013106 (2017);
View online: https://doi.org/10.1063/1.4973562
View Table of Contents: http://aip.scitation.org/toc/rsi/88/1
Published by the American Institute of Physics
Articles you may be interested in
Invited Article: High resolution angle resolved photoemission with tabletop 11 eV laser
Review of Scientific Instruments 87, 011301 (2016); 10.1063/1.4939759
Laboratory-based high pressure X-ray photoelectron spectroscopy: A novel and flexible reaction cell
approach
Review of Scientific Instruments 88, 033102 (2017); 10.1063/1.4975096
High repetition pump-and-probe photoemission spectroscopy based on a compact fiber laser system
Review of Scientific Instruments 87, 123902 (2016); 10.1063/1.4969053
Electronic structure of buried LaNiO
3
layers in (111)-oriented LaNiO
3
/LaMnO
3
superlattices probed by soft x-
ray ARPES
APL Materials 5, 016101 (2017); 10.1063/1.4973558
Modular soft x-ray spectrometer for applications in energy sciences and quantum materials
Review of Scientific Instruments 88, 013110 (2017); 10.1063/1.4974356
Time-resolved photoemission apparatus achieving sub-20-meV energy resolution and high stability
Review of Scientific Instruments 85, 123904 (2014); 10.1063/1.4903788

REVIEW OF SCIENTIFIC INSTRUMENTS 88, 013106 (2017)
A facility for the analysis of the electronic structures of solids and their
surfaces by synchrotron radiation photoelectron spectroscopy
M. Hoesch,
1,a)
T. K. Kim,
1
P. Dudin,
1
H. Wang,
1
S. Scott,
1
P. Harris,
1
S. Patel,
1
M. Matthews,
1
D. Hawkins,
1
S. G. Alcock,
1
T. Richter,
1,2
J. J. Mudd,
1
M. Basham,
1
L. Pratt,
1
P. Leicester,
1
E. C. Longhi,
1
A. Tamai,
3
and F. Baumberger
3,4
1
Diamond Light Source, Harwell Science and Innovation Campus, Didcot OX11 ODE, United Kingdom
2
Data Management and Software Centre, European Spallation Source ERIC, Ole Maaløes Vej 3,
2200 Copenhagen, Denmark
3
Department of Quantum Matter Physics, University of Geneva, 24 Quai Ernest-Ansermet,
1211 Geneva 4, Switzerland
4
Swiss Light Source, Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland
(Received 24 October 2016; accepted 19 December 2016; published online 18 January 2017)
A synchrotron radiation beamline in the photon energy range of 18-240 eV and an electron spec-
troscopy end station have been constructed at the 3 GeV Diamond Light Source storage ring. The
instrument features a variable polarisation undulator, a high resolution monochromator, a re-focussing
system to form a beam spot of 50 × 50 µm
2
, and an end station for angle-resolved photoelectron spec-
troscopy (ARPES) including a 6-degrees-of-freedom cryogenic sample manipulator. The beamline
design and its performance allow for a highly productive and precise use of the ARPES technique
at an energy resolution of 10-15 meV for fast k-space mapping studies with a photon flux up to
2 · 10
13
ph/s and well below 3 meV for high resolution spectra.
C
2017 Author(s). All article content,
except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license
(http://creativecommons.org/licenses/by/4.0/). [http://dx.doi.org/10.1063/1.4973562]
I. INTRODUCTION
Photoelectron spectroscopy is a highly versatile tool for
the investigation of the electronic structure of solids and
their surfaces. The specific regime of very high energy reso-
lution angle-resolved photoelectron spectroscopy (ARPES)
has proven to be particularly successful for the measurement
of electrons at very low binding energies in crystalline
correlated electron systems
1–3
and is thus complementary
to transport measurements on one hand and to (scanning)
tunnelling spectroscopy on the other. Diamond Light Source
has built a vacuum ultraviolet (VUV) to soft x-ray beamline,
which, combined with an end station for high energy and
angular resolution photoelectron spectroscopy forms a highly
productive facility. Measurements are performed on both,
cleaved single crystals with high through-put, as well as
samples grown and prepared by complex surface science and
molecular beam epitaxy methods in the beamline vacuum
system. This instrument is unique in the coherent approach
taken to optimise all components into a well-balanced
ensemble. The design choices and design considerations for
the beamline have been described in Ref. 4.
The technique of ARPES at high energy resolution
requires a highly monochromatic intense photon beam of
photon energy E and energy spread ∆E, combined with
an electron spectrometer, which measures photoelectrons
with an energy resolution ∆E
ana
. The total energy resolution
may further be influenced by other effects such as the
sample grounding noise (∆E
etc
) to give a combined energy
a)
Moritz.Hoesch@diamond.ac.uk
resolution of
∆E
comb
=

(∆E)
2
+ (∆E
ana
)
2
+ (∆E
etc
)
2
. (1)
Since the resolving power E/∆E of a VUV monochromator
is limited, typically to values of about 20 000 (the extreme
reported is 100 000),
5
the technique uses low photon energies
E in the VUV range for high combined energy resolution.
The beamline photon energy range was thus selected to start
from hν = 18 eV. The core operation range, where the energy
resolution ∆E
comb
can be kept well below 10 meV then reaches
up to approximately 100 eV, above which a slightly reduced
energy resolution is used for band mapping, up to 240 eV.
In addition the design allows for a beam of reduced flux at
a photon energy of 500 eV, which is applied to core level
studies. At all photon energies the beam can be delivered
in linear horizontal (LH) or linear vertical (LV) as well as
circular left (CL) and circular right (CR) polarisations. This
paper describes the design and performance of the beamline
and its high resolution end station HR-ARPES. The paper
is organised in sections describing the layout, the undulator
photon source, the beam delivery design and performance, and
the end station design and performance before concluding.
Not described in this paper are the second branch for spatially
resolved ARPES and its end station nano-ARPES.
II. LAYOUT AND BEAMLINE INFRASTRUCTURE
The beamline starts inside the concrete shielding wall
with the undulator source and the front end. Optical elements
are located in a lead shielded hutch and a temperature
stabilised optics cabin. The hutch, optics cabin, and two
0034-6748/2017/88(1)/013106/9 88, 013106-1 © Author(s) 2017.

013106-2 Hoesch et al. Rev. Sci. Instrum. 88, 013106 (2017)
FIG. 1. Schematic layout of the optics at beamline I05-ARPES consisting of the undulator source, the common branch up to the plane grating monochromator,
and the two branches HR-ARPES and nano-ARPES.
end station rooms, a control room and a sample preparation
room are rooms built into the Diamond experimental hall.
The temperature in the optics rooms is kept stable to within
0.2
◦
C over a typical week. Two control and instrumentation
areas (CIA) adjacent to the beamline rooms provide space for
electronics racks for vacuum and motion control, computers
and network routers, pre-vacuum pumps, and a system of
Helium compressors for closed cycle cryostat operation.
The schematic optics layout of the two branches is shown
in Fig. 1. The beam is admitted into the optics through primary
slits that control the illumination of the first toroidal mirror
M1. A further set of slits just after M1 acts as the beamline
aperture and removes unwanted radiation from the edges of
mirror M1. The beam then passes over the plane mirror and
grating of the plane grating monochromator (PGM) and over
the cylindrical focussing mirror M3 into the exit slit, ES.
Alternatively the beam could be directed into the nano-ARPES
branch by inserting mirror M6 instead of M3, which can be
performed by a motorised motion. The dispersive direction
of the monochromator is vertical. The beam is refocussed on
the sample contained in the end station vacuum vessel by a
elliptical toroid mirror M4. The total length of the beamline
from source to end station is 50 m, about half of which is
contained in the storage ring shielding tunnel.
III. UNDULATOR SOURCE
The beamline is served from the long straight section I05
of the Diamond storage ring. As photon source, a 5 m long
variable polarisation undulator of Apple-II type
6
was selected.
The period length is λ
u
= 140 mm and the total length is 5 m,
thus accommodating n
u
= 34 periods and two half-periods at
the entrance and exit. It can produce all four polarisations,
LH, LV, CL, and CR in the range from 18 eV upwards. Higher
harmonic suppression is optimised in a quasi-periodic scheme
7
similar to the design described in Ref. 8.
To estimate the photon flux entering the beamline, Fig. 2
shows a calculation for the fundamental of a simpler fully
periodic 5 m long undulator HU140. In this case, the flux for
LH and LV is identical. The flux is calculated at the highest
brilliance point, where almost all of the beam is admitted into
the beamline aperture. The intensity of the fundamental of the
real undulator is slightly lower, up to 20%, due to the quasi-
periodic perturbation, which is not included in the prediction
calculation. Note that the flux evolves very smoothly over the
full range, which extends beyond 500 eV, while the beamline
optics calculations below cover only the range of 18-240 eV.
IV. MONOCHROMATOR AND REFOCUSSING
OPTICS DESIGN
The monochromator is of collimated Plane Grating
Monochromator (cPGM) type.
5,9–11
This design allows a
free choice of the included angle 2ϑ = α − β of the plane
grating and the associated anamorphic demagnification factor
c = cos β/ cos α (see the inset of Fig. 3 for a definition of
α and β). The optical functions of the beamline mirrors and
FIG. 2. Calculated flux of the first harmonic of a 5 m long HU140 undulator
at its peak brilliance in the Diamond storage ring.

013106-3 Hoesch et al. Rev. Sci. Instrum. 88, 013106 (2017)
FIG. 3. Scaled diagram of the optical path and beam envelope for the HR-branch of the beamline. The inset shows the mirror and grating of the PGM in 1:1
scaling.
slits are illustrated in Fig. 3. Collimation in the vertical plane
is performed by the first toroidal mirror (CM), which also
forms the horizontal intermediate focus. The focussing mirror
(FM) captures the collimated beam and focusses it onto the
ES, at 11 m focal length, thus allowing a distance of more
than 12 m for the dispersion from the grating to the exit slit.
The intermediate horizontal focus is formed by horizontal
focussing of CM at 5 m upstream of the ES, thus making
the intermediate focus astigmatic. The refocussing mirror
(RFM) is an elliptical torus 6 m downstream of the ES and
2 m before the final focus position it demagnifies vertically
by 3 and horizontally by 11/2 (geometrical demagnification
factors). By design the nominal horizontal beam spot size is
approximately 50 µm over the whole photon energy range and
with exit slit openings between 20 and 200 µm the vertical spot
size varies between 7 and 70 µm. The nano-ARPES branch
employs a separate FM with a shorter focal length of 6 m to
make a stigmatic intermediate focus with the horizontal beam
waist on the ES of this branch.
The designuses single crystal silicon as mirror block mate-
rial and the heat load is managed by water cooling, thus keeping
all elements at room temperature or slightly above. At full
opening of the primary slits, highest ring current, and lowest
gap, the undulator can admit up to 1 kW of power onto the first
mirror M1. The bulk of this heat load, rather hard x-rays up to a
fewkeVphotonenergy, is absorbedin M1. A deflectionangle of
6
◦
was chosen off this mirror, which admits less than 100 W of
power onto the plane mirror M2 in the PGM. Since M1 deflects
horizontally, the tangential profile of the heat bump and heat
deformation only affects the horizontal focussing properties.
Thisdeformation wasoptimisedbya cross-sectional design and
attachment of the cooling brackets that allows the back part of
the mirror to get slightly warmer than the cooling water, thus
reducing the bending of the mirror block due to temperature
differences of the front surface and the bulk. M2, which has
a varying incidence position along its length according to the
varying angle geometry,
9
can develop a heat bump, the tangen-
tial profile of which directly affects the vertical focussing plane
and thus the energy resolution. This mirror employs internal
water cooling by channels that run along the long length of the
mirror inside the silicon and thus the heat is removed very effec-
tively and the resulting heat bumps are minimised. The internal
cooling has the additional advantage that no external cooling
brackets deform this 450 mm long mirror. Fully clamped into
its opto-mechanical holder, using metrology feedback from the
Diamond-NOM slope profilometer,
12
the mirror has a radius
>240 km and deviations of 0.16 µrad (RMS) along the tangen-
tial direction. The last optical elements where the heat load is
of concern are the gratings; even with their cooling brackets
attached the slope errors of these are kept below 0.25 µrad.
13
Theheatloadhere is smallandcoolingisperformed byprecisely
aligned and lightly attached copper side-brackets with tubing
that is optimised for minimum vibrations due to turbulences
in the water flow. Finally the horizontally deflecting focussing
mirror M3 also features a water cooled pad on the side of the
mirror, which is used to reduce potential heat drift issues for the
17 m long opticalarmthatfollowsthismirrorup to the reflection
from M4.
Mirror deflection angles were chosen as 6
◦
for M1, M3,
and M4 following a combined optimisation of the available
floor space in conjunction with the choice of mirror coatings.
Fig. 4 shows the calculated reflectivities of various coatings.
While the highest reflectivity is found around 100 eV for a
Rh coating, the amorphous carbon, calculated at a density
of 2.1 g/cm
3
, has the highest overall reflectivity and is free
of energy dependent features apart from the strong drop of
reflectivity on approaching the K-edge absorption (284 eV).
All elements are coated with amorphous carbon, except for the
gratings, where a metallic Pt coating was chosen. M2 deflects
by a variable angle between 3
◦
, used at the highest photon
energies, and 36
◦
, used at the lowest photon energy where still
a reflectivity of more than 60% is expected for an s-polarised
reflection [Ref. 14, Sec. 4.2]. The length of the grating blocks
and the plane mirror has been optimised together with the
vertical beam offset in the PGM. The latter is 35 mm and
with a grating block length of 150 mm and a ruled surface
length of 140 mm, the PGM can achieve high transmission
FIG. 4. Calculated reflectivity for s-polarised radiation at an incidence angle
of 3
◦
for various mirror coating materials.

Frequently asked questions

Q1. What are the contributions in "A facility for the analysis of the electronic structures of solids and their surfaces by synchrotron radiation photoelectron spectroscopy" ?

A synchrotron radiation beamline in the photon energy range of 18-240 eV and an electron spectroscopy end station have been constructed at the 3 GeV diamond light source storage ring this paper. 

Q2. What is the photon flux curve at low photon energies?

The photon flux curve first rises gently from low photon energies with slight humps at 30 and 50 eV and then reaches a maximum at 118 eV before it falls with a rapid drop off at 240 eV. 

Q3. What is the way to detect the 3rd harmonic in LV?

At 50 eV, the 2nd harmonic is very weak in the LV polarisation, while the 3rd harmonic is observed in LV but not in the other polarisations. 

Q4. How many samples are inserted into the load lock?

5-10 samples at a time are inserted into the load lock and transferred to the interface chamber (IC) after a load lock pumpdown of 3–5 h. 

Q5. Why is the fundamental of the real undulator slightly lower?

The intensity of the fundamental of the real undulator is slightly lower, up to 20%, due to the quasiperiodic perturbation, which is not included in the prediction calculation. 

Q6. How can the sample be rotated around the axis?

the sample can be rotated around three axes passing through the sample surface with a sphere of confusion of ∼200 µm. 

Q7. What are the rooms built into the Diamond experimental hall?

The hutch, optics cabin, and two0034-6748/2017/88(1)/013106/9 88, 013106-1 © Author(s) 2017.end station rooms, a control room and a sample preparation room are rooms built into the Diamond experimental hall. 

Q8. What is the advantage of the angle-sweeping lenses on the analyser?

Furthermore the recent development of angle-sweeping lenses on the analyser would enable the measurement of relevant sections of momentum space without rotation or any movement of the sample and thus a faithful measurement from a single small spot on the sample surface. 

Q9. How does the flux of the beamline evolve?

Note that the flux evolves very smoothly over the full range, which extends beyond 500 eV, while the beamline optics calculations below cover only the range of 18-240 eV. 

Q10. What is the resolving power of the beamline?

At these settings, the beamline delivers a resolving power E/∆E = 25 000 with a flux of up to 1012 ph/s (see Fig. 7).The linear regression of the energy resolution in Fig. 6(d) can be used to eliminate the contribution of the grating dispersion to the bandwidth, and estimate the limit of instrument resolution. 

Q11. How many periods is the beamline length?

The period length is λu = 140 mm and the total length is 5 m, thus accommodating nu = 34 periods and two half-periods at the entrance and exit. 

Q12. What is the main problem with the multichannel amplifier?

For the typical high count rate operation of their instrument, the bottle-neck here is, however, the use of a multichannel-plate amplifier, which ages over time and thus develops deviations from flat response across the detector. 

Q13. What are the key limiting factors of the resolution of the mirror?

The key limiting factors of the resolution are identified as follows: (a) Residual slope errors of the plane grating and plane mirror, including cooling water-flow induced vibrations. 

Q14. How much energy resolution is achieved at the smallest slit?

The smallest recommendable exit slit setting below which no substantial gain in energy resolution is achieved is approximately h = 0.015 mm. 

Q15. What is the resolving power of a VUV monochromator?

The total energy resolution may further be influenced by other effects such as the sample grounding noise (∆Eetc) to give a combined energya)Moritz.Hoesch@diamond.ac.ukresolution of ∆Ecomb = (∆E)2 + (∆Eana)2 + (∆Eetc)2. (1)Since the resolving power E/∆E of a VUV monochromator is limited, typically to values of about 20 000 (the extreme reported is 100 000),5 the technique uses low photon energies E in the VUV range for high combined energy resolution. 

Q16. What is the optimum mirror angle for the gratings?

Mirror deflection angles were chosen as 6◦ for M1, M3, and M4 following a combined optimisation of the available floor space in conjunction with the choice of mirror coatings.