Design and construction of instruments with unique

specifications : applications to surface science and

crystal analysis
Citation for published version (APA):
Brinkgreve, P. (2001). Design and construction of instruments with unique specifications : applications to surface
science and crystal analysis Eindhoven: Technische Universiteit Eindhoven DOI: 10.6100/IR545757

DOI:
10.6100/IR545757

Document status and date:

Published: 01/01/2001

Document Version:
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be
important differences between the submitted version and the official published version of record. People
interested in the research are advised to contact the author for the final version of the publication, or visit the
DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page
numbers.
Link to publication

General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners
and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
• You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please
follow below link for the End User Agreement:
www.tue.nl/taverne

Take down policy

If you believe that this document breaches copyright please contact us at:
[email protected]
providing details and we will investigate your claim.

Download date: 09. mars. 2019

Design and construction of instruments
with unique specifications

Applications to surface science and crystal analysis

CIP-DATA LIBRARY TECHNISCHE UNIVERSITEIT EINDHOVEN

Brinkgreve, Peter

Design and construction of instruments with unique specifications :

applications to surface science and crystal analysis / by Peter
Brinkgreve. - Eindhoven : Technische Universiteit Eindhoven, 2001.
Proefschrift. - ISBN 90-386-2882-X
NUGI 841
Subject headings : precision engineering / experimental instruments; design and
engineering / scientific instruments; design and engineering / Ion Scattering Spectroscopy /
Double Toroidal Analyser / Channeling Facility / Vacuum Transport System

Druk: Universiteitsdrukkerij TU Eindhoven

© 2001, P. Brinkgreve
Design and construction of instruments
with unique specifications

Applications to surface science and crystal analysis

PROEFSCHRIFT

ter verkrijging van de graad van doctor aan

Technische Universiteit Eindhoven, op gezag van de
Rector Magnificus, prof.dr. M. Rem, voor een
commissie aangewezen door het College voor
Promoties in het openbaar te verdedigen
op donderdag 28 juni 2001 om 16.00 uur

door

Peter Brinkgreve

geboren te 's-Gravenhage
Dit proefschrift is goedgekeurd door de promotoren:
prof.dr.ir. M.J.W. Schouten
en
prof.dr. H.H. Brongersma

Aan Juul, Boet en Joosje

 i

Table of contents i
Introduction 1
Outline 6
References to introduction 6
1. Design principles 7
References to design principles 15
2. Low-Energy Ion Scattering Spectroscopy: LEIS 17
2.1 Introduction low-energy ion scattering spectroscopy 17
2.1.1 Low-energy ion scattering spectroscopy: Double Toroidal
electrostatic Analysers (EARISS, ERISS, Calipso)
Surface science and low-energy ion scattering [LEIS} 18
2.1.2 General design considerations 21
2.1.3 General description of LEIS with a double toroidal analyser [DTA] 23
2.2 General demands on the materials used 26
2.2.1 Vacuum demands on materials 26
2.2.2 Electric demands on materials 30
2.2.3 Magnetic demands on materials 30
2.3 Geometrical accuracy's and requirements 35
2.3.1 Geometrical accuracy's of the analyser 35
2.3.2 Geometrical accuracy's of the zoom lens 35
2.3.3. Geometrical demands and accuracy's of the diaphragm apertures 36
2.4 The actual size of the analyser 37
2.4.1 The detector 37
2.4.2 The collector plate 38
2.4.3 The specific geometry of the toroidal lenses 39
2.4.4 The distance between the target and the analyser 39
2.5 Connections between inner and outer elements 41
2.6 Discussion of solutions 43
2.6.1 The central hub 43
2.6.2 The hollow dowel pins and spacers 44
2.7 The diaphragm with variable apertures 47
2.8 The azimuth selector 52
2.9 Design goals and results 56
2.10 Summary LEIS 57
2.11 References to LEIS 58
ii

3. High-Energy Ion Scattering: HEIS 60

3.1 Introduction high-energy ion beam applications: channeling facility 60
3.1.1 High-energy ion beam applications: channeling facility 61
3.1.2 General design considerations 62
3.1.3 General description of the channeling facility 65
3.2 Details and components of the beam line vacuum system 70
3.2.1 Vacuum demands 70
3.3 Geometrical requirements 73
3.3.1 Accuracy's of the beam line 74
3.3.2 Accuracy's of the x-y-z translation stage 75
3.3.3 Accuracy of the rotation stage 79
3.4 Discussion of design goals and results 90
3.5 Conclusions HEIS 91
3.6 References to HEIS 92

4. The vacuum suitcase transport system 94

4.1 Introduction of the vacuum suitcase transport system 94
4.1.1 General design considerations 95
4.1.2 Vacuum demands 96
4.2 Requirements of the vacuum transport system 98
4.2.1 Requirements of the vacuum suitcase 98
4.2.2 Requirements of the universal sample holder 98
4.3 Discussion of the vacuum suitcase, load-lock and TU/e sample holder 100
4.4 References to the vacuum suitcase transport system 104
Conclusions 105
Personal remarks 106
Summarised vocabulary according to the "International Vocabulary of
Basic and General Terms in Metrology". 112
List of used acronyms 114
Samenvatting 116
Publications 119
Dankwoord 121
Curriculum vitae 122
 1

Introduction

The Central Design and Engineering Facilities [GTD] is a central unit of the organisation
of the Eindhoven University of Technology [TU/e]. It has the assignment to provide
technical support to the research-groups of the faculties.
The core-business of this GTD is:
• design, development and engineering of experimental equipment and apparatus in
commission of the research-groups (as far as those instruments are not commercially
available)
• specific operational technological support for the benefit of the research-groups.
For this business the GTD consists of a multidisciplinary design group and specialised
workshops. The professional disciplines are limited to mechanical and electric/electronic
engineering and, on a limited scale, physical engineering.
Over the years an incredible variety of subjects has been tackled: from electric motors for
possible biomechanical purposes, with an outside diameter of 0.65 mm up to precision
satellite communication ground stations, for satellite tracking, with reflectors of 5 m
diameter.
The GTD has even a history in support of the development of Continuously Variable
Transmissions [CVT]. The prototypes were produced some 35 years ago and still today
special gear-cases and flywheels are produced for a hybrid transmission system with an
integrated CVT. Recently, ultra high vacuum systems with special analysers, plasma
systems and several dedicated measuring instruments, like a "mice dynamometer", to be
placed inside a MRI-scanner, are under development.

In order to be capable of designing as well as actually building the experimental

instruments, the GTD utilises a wide variety of machines and technologies. One reason for
this is that a way of experimental instrument building is employed that is a mix of "already
building while designing" and "still designing while building". The classical fine
machining and optical diamond turning, several types of electric discharge machining,
plasma welding, vacuum furnace brazing and laser spot-welding, the necessary
electric/electronic and physical technologies do have a high quality level of operational
performance.

The GTD itself is responsible for the quality and quantity of these technologies. Often
special application of existing technology is necessary in order to remain capable of
2

offering appropriate and practical design solutions. These special applications have often
to be realised "in house".
The specialists involved, have been trained "on the job": their professional skills
progressed over the years whereby a strong personal engagement, talent and hard work is
needed to reach their level of performance. They were and still are willing to take on the
technical challenges, which were offered through the mostly difficult tasks that had to be
fulfilled, supporting the experimental scientists. Depending on the complexity, those
challenges are mostly encountered as team efforts; seldom on a personal basis. The more
successful the projects were, the more intense the interaction between scientists and
technicians was. This indicates that for the satisfying functioning of (technical) support-
groups and their (scientific) clients not only the professional skills, but also building and
maintaining good working relations are necessary conditions [1].

The question: "how to continue and further develop the professional skills of the GTD?”
arises when finding qualified technicians is getting more and more difficult. Recruiting
young talented technicians is necessary to replace the elderly specialists, which are going
to retire. The labour market has a shortage of qualified technicians, of any kind, and this
will be so for many years to come. For the last 5 - 10 years the GTD is trying in vain to
recruit a sufficient number of, preferably young and mechanically oriented, technicians
that have the "proper qualifications".
The Netherlands institutes for the practical education of technicians have a shortage of
pupils. What is more: the few pupils that do attend to these technical schools often do not
have the right talents and interests. Even the national "Leidse Instrumentmakersschool
[LIS], with the in my opinion too much reduced admittance levels, turns out but a handful
of often only little-talented instrument-makers. Certainly not enough for the needs of the
Dutch research and development institutes.
Specific education and in-house training of skills of young people with the appropriate
attitude seems the only logical choice to fill in those places in order to continue the
necessary technical support of this University of Technology [3]. This leads to an attempt
to make an inventory of what enables experimental scientists, designing engineers and
crafts-men of this university to co-operate in such a way that the experimental scientist
ends up with better experimental instruments, the designing engineer becomes a better
engineer and the craftsman has the satisfaction of making better instruments. This attempt
is worthwhile, because this co-operation exists already more than 25 years and it seems to
be a significant element of the stimulating and productive working situation.
 3

In the Netherlands national politics prescribe the size/capacity of experimental technical

research and the size of the related and necessary means. The distribution of those means
between institutes was done, until recently, on an "input driven" basis. More and more an
“output driven” model is used that should stimulate quality and relevance of research. For
the distribution of funds for research a competition mechanism is used today, that should
stimulate the quality of government research institutes and research-groups and the
relevance of research subjects.
In the institutes the internal politics determines the interpretation and spending of the
acquired means between the research-groups. Decisions concerning the distribution of
these means are difficult to influence by researchers and technicians themselves. All in all
the concerning amount of money is rather small compared to that in other western
countries. One reason among others is that in the Netherlands professional skill and
technical research are not items with a high social status and priority; both are too much
seen as costs and not seen as necessary conditions to continue and develop a high level of
a functioning of society. The reserved means for (technical) education and the part of the
national income reserved for scientific research are proof of this. As a result the financial
means for research will remain limited in the future. Those limited financial means have to
be spent with care and also have to be aimed at the future.

Of relevance is now the question what the true professional skills are of the group
instrument designers and builders, that functions in a unique organisational situation that
is not comparable to any other technical research organisation. Even though the present
GTD has a very high amount of unique and highly qualified instrument-making specialists,
with a strongly developed understanding of design principles, I do suppose that personal
qualifications alone are not the cause of the technical successfully "couleur locale".
What are the personal values of the daily-work situation? Which of those values have to do
with the characters of the persons involved and which of those are possibly to be
influenced by selection, training and specific education? Questions that are of interest
when it is realised that the quality and quantity of the technical support for the next
generation of experimental physicists is determined by the qualities of the new crop of
technicians. In order to design and fabricate successfully experimental instruments, the
engineers, instrument-makers and their customers (i.c. experimental scientists) have to find
each other in the daily exercised design philosophy and realisation strategy, that is aimed
to functionality. The philosophy here is that too much time is wasted in workshops (and not
in workshops alone) to fabricate useless details that have no relevance to functionality.
4

Most of the drawings and other workshop-instructions stipulate needless details and have
exaggerated tolerance demands that cost a lot of time to fabricate. Any over-specified
tolerance is an unsolved design-problem in the functional design philosophy. A part of this
time can better be used for functional design and functional fabrication in order to get the
most relevant performance for the least effort. These must be optimally tuned to and
integrated into one another [2].
On this University still exists a very much positive but also a critical point of view
concerning experimental technical research and the, for that purpose necessary, technical
support. Without high level sophisticated technical support there will be no high level
experimental scientific output. The question about the continuity of that support is of
relevance here. That is certainly true in a situation where several local scientists have the
personal need to do something about this continuity. They feel that something worthwhile
is going on which needs to be continued. For about 25 - 30 years there is, of service to the
scientists, a GTD that designs, develops and fabricates, in close collaboration with those
scientists, useful experimental instruments and equipment. This group of scientists may
realise that industrial/scientific activities that are successful are so for a period of 5 to 6
years. At this University the technical support of the GTD is already successful for 25
years, so the regular quality of it has to be more than average.
It still seems to be possible to exercise fore front research in physics, astronomy, chemistry
and many other research fields from the library. To an increasing extent, however, is
scientific research dependent on highly sophisticated instruments. To be able to evoke
phenomenon, to measure and to analyse, instruments (mostly extremely complex) are
indispensable. Partly such instruments, or essential parts of them, are possibly
commercially available or already designed and constructed at other institutes. Research
that really breaks new ground is strongly dependent on highly sophisticated instruments
that not yet exist. Very expensive set-ups are being built by interested parties together.
Costs and measuring-time are shared then, e.g. CERN and satellites for scientific
purposes.
A research-group that really is in front is on it's own whenever, for further progress,
hardly partners can be found. In this respect the research-groups of the TU/e are in a
unique position: the GTD with its specialised design and engineering capacities (with their
technical friends and relations in the different departments of industry) is trained and
triggered in the development of instruments, in close co-operation and collaboration with
the customer.
 5

Other Dutch Universities have chosen for other ways to organise their technical support;
recently most of those institutes have even closed down their technical support-groups. The
TU/e believes that it has the only real technology campus in the Netherlands.
The successful experimental scientists of the TU/e, but also their colleagues of other
research institutes, consider the present organisation of the GTD and the way it operates
of added value to the TU/e. This seems to indicate that analysis of the local working
process and the individual capacities are advisable. The purpose of the technical support,
the quality needed and the size of it have to be examined. Giving the "maximum possible"
technical support, depending on the available recourses, can be the purpose. “Maximum
possible" means that the support has to be of adequate quality and anticipating on future
developments: it has the obligation to keep up with time. "Maximum possible" does not
stipulate an absolute measure; it depends on the local balance between demand and
supply, the working-relations with the principals/scientists and their specific needs.
The specific need for technical support differs from scientist to scientist. Normally
speaking, the specific technical interpretation as such cannot be given by the individual
scientist. Satisfying results arise when the detailed filling-in is done together with the
support technicians with their specific technical knowledge: what the subjects are, is the
prime responsibility of the scientist, how they are implemented is entrusted to the
technicians. This "open" working-relation between scientist and technician, where the
scientific and professional views are brought together, can lead to amazing and frontier
breaking solutions.

This thesis considers the how and why of this rather unique combined action to achieve the
described results. In the background lives moreover the question how this situation can be
continued. The TU/e is a breeding ground for excellent experimental scientists. Continuing
the necessary technical support by recruiting young designing engineers and instrument-
makers and developing their operational skills demands a lot of worries.
The technical support supplied by the GTD is mainly of mechanical character. Mechanical
engineering and instrument design and development in particular are not possible
anymore without the support of electrical and physical engineering.
In this thesis, however, the focus will be on mechanical (precision) engineering.
Mechanical engineering is engineering mechanics that works; it certainly is neither
fundamental science nor numerical optimisation of details. The skills and insight that made
it possible for me to participate, for 25 years, in the design and development (engineering)
of many interesting and unique instruments that really worked, will be discussed.
6

Outline

This thesis deals with the design-principles that are the basis of the design strategy for the
fabrication of experimental instruments and how those principles are implemented in the
construction. The primary goal of this strategy is to lay down functional design specifica-
tions. This is achieved by a team effort of scientists, design engineers and makers, that is
focussed on functionality. The strategy to reach that goal is communication. Therefore,
makers and engineers do speak the same technical language. Especially this language has
to be based on insight and recognition of mutually understood and accepted design
principles. Mainly this comes down to defining design qualities in terms of kinematic and
static constraints. In the chapter on design principles these constraints will be considered
in detail.

References to introduction

[1] [Jones, 1967]

R.V. Jones: Instrument design: a pursuit and a profession.
Presidential Address to the Scottish Branch of the A.S.E. at Aberdeen.
[1a] [Jones, 1987]
R.V. Jones: Instruments and experiences.
Wiley & Sons, London Compilation of papers.
[2] [Moore, 1989]
W.R. Moore: Foundations of Mechanical Accuracy.
Library of Congress Catalog Card Number 73-127307.
[3] [Saris, 1993]
Frans W. Saris: Alles is Natuurkunde, maar Natuurkunde is niet alles,
hoofdstuk Freudiaans. ISBN 90-801704-1-0.
 7

1. Design principles

Kinematic and static design is aimed at constraining degrees of freedom [dof]:

rotations and translations. Rigid bodies are interconnected by means of constraints.
The relative position of one solid body with respect to another is fully and
sufficiently fixed when six degrees of freedom are constrained [1].
When a rigid body is restricted more than once in any of its dof the design is
statically indeterminate and, therefore, assembly stresses or thermal stresses are not
known.
Over-constrained designs of machines and instruments are expensive to fabricate.
Their parts have to be machined to fine tolerances. Only in this way it is possible to
limit the distortion of those parts due to internal assembly stress arising when they
are put together.
Correct kinematically and statically designed machines and instruments, however,
are less costly to fabricate: the least mechanical machining effort is needed for the
fabrication of components.

The mechanical accuracy of instruments is affected by various factors.

Under stationary temperature-conditions and in the absence of environmental
vibrations, the positions of components should be unambiguous and not affecting
negatively the geometry of the points of interest of instruments.
Correct kinematically designed instruments have the inherent capability to be very
accurate. Unnecessary distortion of components does not occur. Forces between
parts can be calculated and parts have to be dimensioned accordingly. Even the
stiffness of those parts and of a complete design can and should be determined
beforehand [1], [2] and [3]. Only by the use of kinematical and statical design-
principles, the dynamic behaviour of a design and its elements can be predicted.
Static and dynamic tolerances of position can be predicted beforehand instead of
becoming apparent afterwards when the instrument has been completely assembled.
In addition to inaccuracy from dynamics in the instrument, thermal behaviour and
the resulting influence on the geometry have to be considered also at the design
stage.
8

Small relative movement of "fixed" components under friction mainly causes

hysteresis. This movement can be prevented when the stresses in the construction
are known (i.e. statically determined design required) and limited.
Unknown stresses must not be built-in during assembly. Inevitably some part of the
unknown stresses is released by thermo/mechanical action during temperature-
changes (the result is a “new” unknown geometry, a new instrument is created at
every temperature excursion). This thermal snapping has to be considered and
fought with appropriate counter-measures, this again demands for kinematically
and statically determined constructions.
Temperature-variations, often non-homogenous, should only temporarily influence
the accuracy of instruments; i.e. instruments should not have temperature-
hysteresis. Good results can be achieved by constructing a thermal centre [tc] at a
point of interest. The instrument then remains accurate within a temperature-range.
This is more practical than being accurate at a single stationary temperature only.

Another factor that influences the accuracy of instruments is virtual play [vp]
(virtual backlash). This static position-error of components arises from stiffness
combined with friction. Hysteresis under load, be it intentional loading or
unintentional handling forces, has to be avoided as this results also in the creation
of “new” unknown geometries. Changes in the direction of load between
components will reveal this positional-error (fig. 1.1).
F

c s s tanα = c
α

model 1 F x

F Fu
sv
2 Fu
c Fu
α

model 2 F x

sv

model of mechanism hysteresis loop

Fig. 1.1
In the first model stiffness C is combined with real play s .
In the second model this stiffness is combined with a friction force Fu this
2 Fu
then results in the virtual play sv = as shown in the hysteresis loop.
C
 9

A design specifically aimed at low hysteresis is often possible to fabricate without

additional cost. The upgrading of designs not specifically aimed at low hysteresis is
costly and often not possible at all.
There are numerous forms of constraints of degrees of freedom possible for the
design and engineering of kinematically and statically determined constructions [3].
One frequently used form of constraint is the "rod" that is fixed between bodies. In
a correct kinematic and static application a rod is orientated in the direction of the
specific constraint. Fixtures between those rods and bodies can either be elastically
deformable (and for the purpose of "one time" adjustments even plastically
deformable) or be manufactured from parts that allow for sliding or rolling
movements. The elastically deforming fixture (elastic hinge or notch hinge) has no
friction and therefore has no hysteresis. Elastically deformed material in itself has
hysteresis of a very low order that can be neglected in most cases. The elastic hinge
does have, however, mechanical stiffness causing an elastic force as function of its
displacement.
The sliding fixture or joint (f.i. a ball joint) and to a lesser degree the rolling joint
(f.i. a self-aligning ball bearing) do have friction and therefore do have hysteresis.
Relative motion of components is best obtained with pre-loaded bearings and slides
for large movements and with elastic hinges for small movements. Both are free of
undesirable play.
Elastic hinges of 2D and 3D form are preferred for the design and engineering of
precision instruments of correct kinematic and static design. Elastic hinges are,
generally speaking, cheap to fabricate. Between the limits of elasticity their
behaviour is unambiguous [5]. With respect to temperature behaviour better results
can be achieved when performance-determining parts of such instruments, some
times even complete instruments, can be machined from one and the same piece of
material. This "Monolytic-Design" has no mounting surfaces between different
construction parts [7]. Heat barriers between different parts do not exist.
Temperature gradients between parts are minimised and temperature-hysteresis
does not exist (fig. 1.2). The application of this type of design of instruments is not
to be limited to the fabrication from customary metals only. Frontier shifting
solutions for scientific instrumentation can be reached when instruments, where
really necessary, can be executed from more exotic materials.
10

Body 1 Body 1

fixture=
elastic joint
mounting
surface
rod

Body 2 Body 2

Normal 2D or 3D demountable Monolytic 2D or 3D elastic joints

elastic deformable joints

Fig. 1.2
The place of hysteresis is the mounting surface between the
thin spring steel strip and the thick clamping block.

Hard-metals, temperature stable metals, but also brittle materials like quartz,
zerodur ® and ceramics can be precision machined to amazingly complex, but also
functional, work-pieces [4]. Machining those types of material is very much time
consuming and therefore expensive. It should be noticed that solving a design
problem by the application of exotic material is advisable only for a good design.
Such an expensive engineering solution can be cost effective, however, when
compared to a solution of more classical material that will not meet specifications.
Regular instrument-making workshop-routines and -technologies of excellent
quality together with more (modern) specialised technologies are needed for the
fabrication of such instruments. Different types of "Electric Discharge Machining"
[EDM] (spark eroding), Electro Chemical Machining [ECM], metal etching and
metal deposition techniques (micro galvanising), abrasive machining, together with
optical surface machining and even vacuum brazing have to be operated on a high
level of understanding. Only then it is possible to fabricate this type and the more
regular precision instruments of kinematic design. During the final assembly-stage
of any precision instrument and certainly for the fabrication of “one-off” precision
parts, manual routines, like lapping and honing, still remain quality determining
factors.
Apart from stationary factors there are also dynamic factors influencing the
accuracy of instruments:
 11

• Dynamic position-errors caused by the drive, arising during and after

movement. The position-error is dependent on the motion profile of the drive.
• Step-response of the initial stationary system, rather than the subsequent
mechanical resonance with the periodical repeated actuation, is the determining
factor.
Stiffness of the drive and backlash affect the step-response and thus also the
accuracy. When there is no backlash, high stiffness is best. If there is backlash,
however, high stiffness is not desirable. Hammering and eventually forging of the
construction-elements that bring about the clearance will occur with eventually
disastrous results. When backlash is traversed, a velocity difference between two
bodies concerned will be the result.
During impact, at the end of the backlash stroke,
kinematic energy 1 m v 2 is transformed into
2

the potential energy 1 cu 2 of the mass spring system concerned.

2
( c : stiffness, u : vibration amplitudes, v : impact velocity and m : mass).
m
This results in u = v . Thus u is small when the mass m is small and the
c
stiffness c is high.
The resultant maximum force is f = c. u = v m. c . Thus here a small mass is
advantageous but a high stiffness is disadvantageous.

No backlash in combination with high stiffness is preferable. Virtual play, the

positional error that is revealed when the direction of motion, between elements
under friction changes, has to be avoided. A high stiffness, no backlash and low
friction design are best.
E.g. for the rotation of bodies the statically determined solution: bearings with three
balls, is usually neither feasible nor practical. Pre-loading of ball bearings, with
numerous balls, is then unavoidable and running in tracks and adjusting pre-load
will be necessary to obtain high accuracy to at least the level where, for all the balls
concerned, the elastic deformation compensates for any deviation of diameter, etc.
At the cost of a fair amount of accurate machining and assembly time increased
stiffness with (relatively) low friction and no backlash will be the result.
Nonetheless this conforms to the foregoing statements about high stiffness and no
backlash design.
12

Dynamic position errors as a result of ground vibrations should also be avoided or

at least be small. Ground vibrations represent velocity jumps, which also have a
step response (when no active compensation is used). This response may affect the
positional accuracy, even long after the velocity jump occurred (trains and other
traffic passing by, slammed doors, etc.).
This positional accuracy will be minimal when the system is designed for minimal
energy storage: light and stiff. To provide maximum passive isolation against
ground vibrations a large mass should be supported on low stiffness springs.
The instrument to be isolated is then rigidly attached to this mass in a statically
determined way. Attenuation of ground vibrations by a factor 100 to 1000 can be
achieved this way [6].

At the TU/e, the above mentioned considerations are met by a multidisciplinary

team of instrument-makers, technological experts and designing engineers. A team
that has a history in experimental instrument building based on kinematic and static
design principles. This GTD-team in particular has been, from the beginning (1960-
1965), at the forefront of the practical applications of those design principles.
Experimental instrument building is in a way design and engineering "one-off"
instruments. Discrepancies between design-goals and engineering-possibilities are
solved here during the actual process of building those instruments.
As a customary strategy the technological and other realisation problems of new (to
be realised) commissioned instruments are pin-pointed first. This is the
responsibility of the multidisciplinary team. The crucial design problems are solved
on a theoretical basis as well as with a practical execution before the design and
engineering of the actual instrument is taken on.

As long as the foregoing design principles are honoured, a good design will be the
result. This applies to large constructions as well as fine mechanics, for general
machine building as well as precision engineering.
Perfect solutions can than be found for in fact any design request: for fatigue testing
equipment, measuring instruments, sample conditioning set-ups, etc.
The GTD has been doing this for the last 30 years and still is doing this today.
 13

Below, a selection is given of projects that were carried out by the GTD, in which
the author has been involved personally. The selection illustrates the wide scope of
the applications carried out.
1. Surface Roughness Measuring Machine: friction wheel driven surface
roughness (< 0.01 µm) measuring of large areas.
2. 3D Quadrupole Magnetic Field Measuring: a coil mounted on an accurate,
rotating aramid composite tube.
3. High Resolution (1:10-7) Numerically Controlled X-Y Table: photo-
lithography pattern generating.
4. Blood Coagulation Meter: the automation of a blood coagulation centre.
5. Monochromator for Daresbury Synchrotron Radiation Facility:
continuously variable radius monochromator fabricated with a bended
silicon-wafer.
6. Tribo Tester for large pre-loads (100 kN) by means of a hydrostatic
spherical oil bearing.
7. Winding Machine: laboratory scale prepreg fabrication.
8. High Frequency (16 GHz) Areals: fabricated from temperature stable
carbon prepregs.
9. Infrared Oven: fast heating radiation oven consisting of six optical,
elliptical cylinders with one common axis.
10. Pin on Disc and Pin on Plate: determining wear resistance of ceramics at
elevated temperatures.
11. Flywheel Hybrid Drive: studies on improving the efficiency of motorcars.
12. Inside Diameter Cutting Machine: tensioned blade with a bonded diamond
cutting edge, making fine cuts in very hard materials.
13. Foot Size Measuring Apparatus: which uses photo-sensors to determine
length and width of feet without touching them.
14. Invalid Chairs: two different prototypes of arm muscle driven wheelchairs.
15. Magnetisation Measuring Machine: 3D mapping of small magnetised
components.
16. Shock Tube: with a mechanically de-coupled measuring section to improve
stability/accuracy.
17. Fast Vacuum Switch: fast switching of electric mains with high power
(600 A by 15 kV).
14

18. Variable Capacitor: measuring frequency dependance of dielectric

permittivity.
19. Instrument for Defining Surfaces of Dental Filling: determination of wear
by means of height mapping.
20. Precision Fibre Drawing Set-up: fabrication of precision glass capillaries
such as used in dot-matrix printers for textiles.
21. Precision Mounting Frames: six degrees of freedom manipulator for the
positioning of dipoles for the Euterpe electron storage ring.
22. UHV SPA-LEED manipulator: x-y displacement of an electron beam over a
sample surface and also rotating this sample around two axes.
23. Single Particle Crusher: measurement of fracture forces of small (5 - 250
µm) abrasive particles.
24. Lateral Force Apparatus: quantitative measurements of forces (µN - mN)
and displacements (1 - 10 µm) in dynamic sliding.
25. Accelerating Cavity of the Race-track Microtron Eindhoven: coupling 9
accelerating precision diamond machined and vacuum-brazed OFHC cells
into one precision structure.
26. RAIRS-LEED set-up: measuring adsorbate, present on catalytic surfaces by
means of infrared spectroscopy.
27. Tensiometer for Electron Microscope: in situ tensioning of polymers
observed during microscopy.
28. Nano Indenter: determining hardness and youngs modulus of thin coatings.
Indentations 0.1 - 10 µm (resolution 1 nm) and forces, 1 N (resolution
10 µN).
29. Optical Measuring Machine: infrared spectroscopy on drums of copying-
machines.
30. Photo-Film Tester: measuring internal friction of automatic photo-camera
reels.
31. Micro Electro Discharge Machine: friction wheel actuated EDM machine
for fabricating up to 10 µm holes in < 0.1 mm metal foils.
32. Precision Linear Spectrometer: apparatus for performing x-ray absorption
fine structure measurements using a rotating anode as photon source.
33. Precision Stands for 3 and 5 meter diameter Arials: geostationary satellite
tracking with an accuracy in both azimuth and elevation < 0.001°.
 15

Three instruments (not so experimental anymore) that are in daily use, will be dealt
with in chapters 2, 3 and 4 of this thesis. These instruments have been designed by
the author in close co-operation with a number of experienced designing
instrument-makers. The fabrication was done on an experimental instrument-
making basis.
Those developments are chosen to illustrate the way that unique specifications of
physical instruments can be met with the design principles.

References to design principles

[1] [Hoek van der, 1989]

W. van der Hoek: Het voorspellen van het dynamisch gedrag en positioneernauw-
keurigheid van constructies en mechanismen. Lecture book TU/e nr. 4007.
[2] [Koster, 1996]
M.P. Koster.: Constructie principes. Uitgeverij Universiteit Twente, Enschede,
ISBN 9036508320.
[3] [Rosielle, 1995-1998]
P.C.J.N. Rosielle, E.A.G. Reker: Constructie principes, bedoeld voor het nauwkeu-
rig bewegen en positioneren. Lecture book TU/e.
[4] [With de, 1996]
G. de With, L.J.M.G. Dortmans: Constructief gebruik van brosse materialen.
De Constucteur, Okt. 1996, no. 10.
[5] [Jones, 1950]
R.V. Jones: Parallel and rectilinear spring movements. Journal of Scientific Instru-
ments (1951) 38-41.
[6] [De Bra, 1991]
D.B. de Bra: Vibration Isolation, text book tutorial notes ASPE Meeting Santa Fe,
1991.
[7] [Smith, 2000]
Stuart T. Smith: Flexures: elements of elastic mechanisms. Gordon and Breach
Science Publishers, ISBN 90-5699-261-9.
16

Fig. 2.1.1
Exploded view of the double toroidal LEIS analyser.
The dotted lines represent mumetal shields placed around the analyser.
 17

2. Low-Energy Ion Scattering Spectroscopy: LEIS

2.1 Introduction low-energy ion scattering spectroscopy

One example of the successful co-operation between a scientific research group and
the GTD is the development of several Low-Energy Ion Scattering [LEIS]
analysers of the group Physics of Surfaces and Interfaces. This development spans
a period of 15 years and is still continuing.
During this period 3 different LEIS analysers, together with all the supporting
systems, were designed and fabricated.
Calipso B.V. (one of the several commercial, technical and scientific application
units of the TU/e Holding) has one of the LEIS analysers in daily operation for
solving scientific problems for industry.
Proof of the unique possibilities of this type of LEIS is the growing demand for
specific surface analyses with it.
Recently the research school NIOK has decided that it is very important for
catalytic research in the Netherlands that a National LEIS Centre [NLC], based on
these LEIS analysers, is formed in Eindhoven.
The growing demand, from industry in particular, has stimulated the design and
development of still another analyser together with the mumetal vessel, specific
precision manipulator, transfer-system, sputtering units etc.
For TU/e standards this has been (and still is) a very big project: because of the
amount of undergraduate students (117), Ph.D. students (29) and very many
visiting scientists that participated in projects together with the scientific staff of the
group Physics of Surfaces and Interfaces and the technical specialists of the GTD.
An interesting feature of this multidisciplinary project is that almost all professio-
nal technical disciplines of the GTD contribute.
Designing engineers (mechanical, electrical as well as physical, together with their
supporting specialists) several designing instrument-makers with the regular
machine shop assistance and the special technological support participate. This
encompasses vacuum-furnace heat-treatment and brazing, laser welding, all types
of electric discharge machining, thin film deposition techniques, grinding and
polishing of optical parts and ceramics and sheet metal work. Welding specialists,
even the glass blowers, are still making their contributions today.
In order to achieve the accuracy of the parts and of the complete analyser the
technical performance and discipline of the participants involved, have to be of a
18

very high standard, but first of all the multi-disciplinary design of those analysers
has to be of superior quality.
Over the years the professional qualities and the quality of communication between
all involved individuals have proven here to be key-elements.
Many international scientific publications on the subject of these surface analysers,
the specific analysis and related subjects were a result of this project.
On a national level there were several technical publications and many lectures on
the mechanical design and engineering aspects of this "family" of analysers.

2.1.1 Low-energy ion scattering spectroscopy: Double Toroidal Electrostatic Analysers

(EARISS, ERISS, Calipso)
Surface science and low-energy ion scattering [LEIS]

The atomic composition and structure of the surface determine many of the
properties of a material. Processes such as adhesion, electron emission, corrosion
and catalysis are intimately connected to surface properties. During the last century
many techniques have been developed to analyse the surface and to try to connect
the material properties with physical and chemical properties. Since about 85 % of
the industrially important processes in both the chemical and the electronic industry
rely fully on surface and interface properties, it is clear that much energy and
money is invested in a better understanding of them. If, for instance, the efficiency
of an important catalyst could be improved with 1%, the gain for that specific
process could be of the order of US $ 107 - 1010 per year.
Most of the analysis techniques in surface science use the interaction of beams, of
particles (atoms, molecules, ions, electrons) or photons, with the surface [1]. The
number of combinations for incident and scattered particles is very large, especially
if one realises that parameters such as energy and direction of the incident of the
scattered particles have an important influence on the process one can study. For
instance, at high energies and large angles with the surface, the information will
generally come from deeper layers than in a case when low-energy particles under
grazing angles are being used. At present more than 100 different experimental
techniques are being used in the world to analyse surfaces [2] and [3] (Surface
Analytical Techniques).
A characteristic feature of a surface analysis technique is its information depth
(how deep the technique probes the surface). The information depth depends on the
 19

penetration of the incident particles and of the scattered particles/photons. In most

surface techniques the information depth is 10 nm (about 50 atomic layers) or less.

One of the techniques is Low-energy Ion Scattering [LEIS], sometimes called Ion
Scattering Spectroscopy (ISS). In this method ions of 0.1 - 10 keV impinge on the
surface to be studied. The ion impact will produce secondary electrons, sputtered
atoms, ions and molecules as well as photons. There also is a small probability that
the incident ions themselves are being scattered back by a hard collision with target
atoms. In LEIS one selectively studies these back scattered ions. By using noble
gas ions, such as He+, Ne+ and Ar+, it is possible to reduce the information depth to
one atomic layer. This is due to the very high electron affinity of the ions (it takes
very much energy to remove an electron from a noble gas atom and thus much
energy is gained in the reverse process). Since this affinity is higher than that of any
other atom in the periodic system, the ions will be very efficiently neutralised in
their interaction with the surface. The probability of being scattered as an ion is,
therefore, very low [4]. The fraction of the back scattered ionised noble gas
particles compared to the total number of all back scattered noble gas particles
(irrespective of their charge states) is only 10-1 - 10-5. Ions that penetrate and scatter
beyond the first atomic layer have much longer interaction times with the surface
atoms than those back scattered from the top layer. The chance of being back
scattered from deeper layers and still be in an ionised state can, therefore, be
neglected.
This means that the LEIS signals are very weak, but originate exclusively from the
outer atomic layer.

Thanks to the effective neutralisation of the noble gas ions, the back scattered ions
have suffered only one collision with an atom in the outer atomic layer of a surface.
In a LEIS experiment ions having a given energy impinge on the surface. The ions
will be scattered in all directions. Since the ions are heavy particles having high
energies, the Broglie wavelength and interference effects can be neglected in LEIS
(in contrast to low-energy techniques based on electrons). Therefore, the scattering
of ions in a LEIS experiment can be described by classical mechanics. It is in this
sense equivalent to a billiard ball game with balls having unequal masses.
20

By means of an electrostatic field the energy of the scattered ionised particles is

analysed. Conservation of energy and momentum in a collision dictate that, under
such conditions, there is a one to one relationship between the energy of the
scattered ion and the mass of the surface atom from which it was scattered. An
energy spectrum of the scattered ions thus is equivalent to a mass spectrum of the
surface atoms.
It is precisely the outer atomic layer that is analysed in LEIS that plays a dominant
role in the industrial processes mentioned above. Since the process that determines
the information depth (neutralisation of the ions) is universal, one can use LEIS for
any surface (as long as it can be put in a vacuum to allow for the transport of the
ions). The quantification of the analysis has long been a matter of dispute. If
neutralisation is so important in limiting the information depth, one would expect
that the ion neutralisation mechanisms depend heavily on the precise electronic
states of the surface atoms. It has been shown, however, that the dominant
neutralisation mechanism involves the core electrons of the atoms and therefore is
generally independent of the electronic state of the atoms and that of their
neighbours [4]. This makes the LEIS technique a unique tool in surface analysis.
For single crystal targets shadowing and masking effects in the scattering process
lead to characteristic features in the angular distribution of the scattered ions. From
this distribution one can even derive the atomic positions of the atoms in the
surface. The main principles of LEIS are known since 1967 [5], [6], [7].
A serious drawback of LEIS is the fact that the signals are very weak and that the
incident ions have so much energy that they will eventually destroy the surface to
be analysed. This has hampered widespread application of LEIS so far.

By taking full advantage of the new insights in modern technology a new

generation of Low-energy Ion Scattering instruments has been developed at the
Eindhoven University of Technology. These instruments are so sensitive that the
ion dose, that is required for an analysis, can be reduced by several orders of
magnitude. At such low values the damage is negligible. This development is the
result of a close collaboration between the Group Physics of Surfaces and
Interfaces (leader: Prof. dr. H.H. Brongersma), the Central Design and Engineering
Facilities (technical leader ir. P. Brinkgreve) and various other groups within the
Eindhoven University of Technology. Although the final design could only be
reached as a compromise among the requirements posed by surface science, ion
 21

optics, mechanics, data acquisition, etc., here the focus will be on the features of
the mechanical details.
Aim of the new analysers:
• Very efficient detection of the back scattered ions to lower the required ion
dose (dose that is necessary to obtain a certain accuracy in the analysis).
• Realise a scattering angle that permits a good mass resolution of the surface
atoms and a high yield for the scattered ions.
• Compatibility with surface science, in particular with applications as encoun-
tered in catalysis, adhesion, electron emission, etc.

2.1.2 General design considerations

The requirements formulated for the new analyser are partially conflicting. They
also put very heavy constraints on the mechanical design and the selection of
materials for the various components. For instance, the best mass resolution is
obtained for the strongest interaction of the ions and target atoms. This would
favour 180º scattering (for the definition of the scattering angle ! see fig. 2.3.3 page
36). Most industrial applications relate to materials with rough surfaces (certainly
on an atomic scale). Steep angles of the incident of scattered ions with the surface
are preferred to prevent that the signals are strongly dependent on the particular
surface roughness. On the other hand the probability for scattering over a very large
angle is very low and will lead to weak signals. There are also severe mechanical
constraints to the scattering angle. Since the incident ion beam should be well
defined, the lens elements and deflection plates for the ion optics have significant
dimensions. For large angle scattering this would physically interfere with the
analyser for the back scattered ions.
22

Fig. 2.1.2
Schematic view of the double toroidal analyser. Trajectories of scattered ions
with several energies are drawn.
 23

2.1.3 General description of LEIS with a double toroidal analyser [DTA]

Primary ions are incident along the symmetry axis and hit the target surface
perpendicularly (fig. 2.1.2).
Back scattered ions with the correct energy (the "pass-energy") are deflected by the
electrostatic fields between the lenses D1 - D2 and between D3 - D4 (fig. 2.1.2) in
such a way that their trajectory is halfway between D1 - D2 and halfway between
D3 - D4 and eventually these ions hit the first channel plate of the detector. Ions
with a somewhat higher energy will hit the first channel plate closer to the axis,
while ions with a somewhat lower energy will hit the channel plate at larger radii.
In this analyser one can simultaneously image a 12% bite of the energy distribution
of the scattered ions. Ions having an energy that deviates more than 6% from the
pass energy will strike the walls of the analyser and have a negligible probability of
reaching the detector.
Before the scattered ions enter the DTA they pass the "zoom lens", where they can
be accelerated or decelerated. By tuning the energy of the scattered ions by means
of the zoom lens, a different part of their energy distribution can be measured. By
using a number of acceleration and/or deceleration voltages, a full energy spectrum
is obtained.
The electron optical design of the analyser has been described by Hellings et al. [8].

The toroidal lenses (D1, D2, D3 and D4 of fig. 2.1.2) do not have a simple geometry.
The geometry has been designed to give an unambiguous and linear image of the
energy of scattered ions on the first channel plate. A series of channel plates are
used to amplify the weak signals, generated by the incoming scattered ions.
The analyser is rotationally symmetric and the signal of scattered ions is imaged on
a circularly shaped detector. If the ions are scattered by the sample at the place
where the axis of rotational symmetry of the analyser intersects the sample-surface,
the scattered ions have no angular momentum with respect to this axis. These ions
therefore preserve their angular (azimuthal) distribution. Ions that are scattered at a
different place of the sample-surface will have some angular momentum. These
ions will move along a spiralling trajectory (fig. 2.1.3).
Since angular momentum is conserved during their motion, the angular velocity can
become quite large where the trajectory is close to the axis (near the detector).
24

Their spiralling motion can, therefore, nullify most of the angular information from
the scattered ions.

Fig. 2.1.3
Schematic figure of a scattered ion trajectory through the
diaphragm slits at an angle with the meridian plane AOB.

To be able to accept ions that originate from any part of the sample, an azimuth
selector (see section 2.8 page 52) has been designed that suppresses ions having
spirally curved paths. It is then possible to use a defocused or a rastered ion-beam,
which reduces further the surface-damage inflicted by the ion-beam. Another
advantage of the presence of an azimuth selector is that the primary ion beam
doesn't have to be aligned exactly with the axis of the analyser. This simplifies the
tuning of the beam considerably.

Since it is impossible to design an azimuth selector that only intercepts ions with
angular momentum, the loss of the ions without significant angular momentum
should at least be balanced by the larger area of the sample that can be used for the
analysis (the ion dose per surface area should be the same or preferably less than
that without the selector).
 25

The detector-assembly has two functions:

• the amplification of the signal
• the registration of the position of the ion impact in two dimensions.
Analysed ions fall on the first of the three channel plates. The channel plates are
used to amplify the signal given by the incoming ions. For every single ion there
are about 107 electrons produced. This cloud of electrons hits a detector plate (the
collector in fig. 2.1.2). Although the diameter of this cloud is significant, its centre
still has the same co-ordinates (radius and azimuth) as that of the ion impact on the
first channel plate [9].
The detector plate consists of an etched pattern of conducting rings and sickles in a
layer of gold. The gold layer has been sputtered on a glass support (see fig. 2.4.1).
The position of the centre of the cloud is determined from the distribution of the
electrical charge over the strips: which enables "Energy and Angle Resolved Ion
Scattering Spectroscopy" (EARISS). This type of LEIS is very important for the
more fundamental studies in surface science where one uses single crystal samples
[10].

For many applications of LEIS one deals with polycrystalline or amorphous

materials. This holds for most of the industrial applications. In these cases there is
no azimuthal information and the presence of an azimuthal selector would only
reduce the signals. In the "Energy Resolved Ion Scattering Spectroscopy" (ERISS)
and "Composition Analysis by Low-energy Ion scattering for Probing the Surface
on the Outside" (Calipso) set-ups the azimuthal selector is omitted.
Because the electron clouds lead to very short charge pulses (≈ 1 ns), high count-
rates are feasible and therefore, the detector has to meet high frequency demands.
Count rates up to 4.105 cts/s have been reached with an energy resolution of about
1 %.
26

2.2 General demands on the materials used

2.2.1 Vacuum demands on materials

For studies of reactive samples the analyser has to be used in UHV. For samples
having reactive surfaces, pressures of 10-10 mbar and lower are a necessity.
Outgassing of the materials used for the vacuum housing and the analyser is the
deciding factor for pump capacity here. Outgassing of surfaces is the result of
desorption of contaminants (e.g. from cooling-oil and tapping-grease as used in
normal workshop practice during the manufacturing of the various parts) and of ad-
and absorbed gasses from the atmosphere. Elevated temperatures, as used in “bake
out”-procedures, accelerate outgassing. Temperatures up to 180º - 250º C are used
for this purpose. After bake-out, the out-gassing of surfaces at room temperature is
greatly reduced. For stainless steel, for instance, the outgassing rate is about
10-3 mbar l /s m2 after one hour in vacuum and is reduced to 10-4 after 10 hours.
A serious bake-out procedure can reduce this to only 10-7 mbar l /s m2. Similar
results can be obtained for quartz and well-sintered alumina.

Surfaces and their respective material outgassing properties:

2
Item Surface [m ] Material

Vessel 0.8400 AISI 316

Analyser parts 1.2 AISI 316
(inner and outer lenses etc.
Assembly parts 0.2 AISI 316
(nuts, bolts, washers,
mounting flanges)
Manipulator:
main body 0.3 AISI 316
bellows x 1.4 (ø 140x110, 120 pair) AISI 316
bellows y,z 1.1(ø 80x50, 180 pair) AISI 316
≈ 5.1
Outgassing rate after 10 hours
-4 2
10 mbar l/s m
-4
Outgassing 5.1x10 mbar l/s
Magnetic shields (2x) 0.9 Mumetal
Outgassing rate after 10 hours
-4 2
10 mbar l/s m
-4
Outgassing 0.9x10 mbar l/s
Ceramic parts 0.05 Al2O3
Outgassing rate
Outgassing
Channel plates (3) 0.2 Metalised glas
Outgassing rate
Outgassing
Polyimide contact rings (5x) 0.0014 (ø 46x42.013 thick) Polyimide film, type NN

Table 2.1
 27

Apart from the manipulator (4 degrees of freedom, bellows type) substantial contri-
butions to outgassing come from the vessel itself and the mumetal shields, they also
cause pump restrictions.
It can be advantageous to design a special non-bellows type manipulator and to use
only one magnetic shield in combination with a mumetal vessel. The best is no
mumetal shield at all (feasibility studies are ever going on to determine the best
compromise between costs of vessel and shields versus pump-capacity).
Measures that can be taken to reduce the outgassing of chosen metal surfaces:
• Only constructions without enclosures (the air from enclosures will slowly leak
into the vessel and will show up in a restgas analyser as a leak: they are,
however almost impossible to locate with conventional leak testing equipment).
• Rolled materials used in the right direction (very high diffusion rates may arise
in the direction perpendicular to the applied pressure of the rolling process).
• Weldable materials, TIG exclusively, specialised welder and welding proce-
dures.
• Smooth (polished) surfaces with thin uniform defect-free oxide layer obtained
from a special etching procedure.
• Surfaces can be polished by a combination of mechanical polishing and
electrochemical etching: this can significantly reduce the roughness and thus
the physical surface, by more than an order of magnitude [11].
• Temperature-resistant design up to 250º C: statically determined design
required.
• Chemical oxidation forms a uniform oxide layer that strongly reduces the
diffusion of carbonaceous compounds to the surface.
For the construction of UHV vacuum systems (tubes, vessels, flanges etc.)
austenitic stainless steel (AISI 316) is the material of choice, world-wide. With this
material not only the above mentioned measures can be taken, but also it is:
• Easy to machine and easy to obtain a smooth surface.
• Easy to weld because of the low C content (18% Cr, 8% Ni, C ≤ 0.08%,
2% Mo, rest Fe).
• Leak tight in itself but also after welding or brazing.
• Corrosion resistant because, under normal atmospheric conditions, it forms a
thin but solid surface oxide layer.
28

• Bakable up to 400° C without changes in structure.

• Non-magnetic because of the austenitic structure.
• Readily available of high quality in all forms: sheet, bar and tubes.
Putting together UHV vacuum systems successfully requires painstaking attention
and disciple during assembly. A single fingerprint is enough to keep you pumping!
Example:
Assume that a fingerprint has a surface area of 2 cm2, a thickness of 0.1 mm, a
density of 1 g/cm3 and a molar weight of 100. When the material of the fingerprint
evaporates, this will produce 4.8 mbar l thus 4.8 1010 litre of gas at a pressure of
1.10-10 mbar! Even at a pressure of 10-8 mbar a vacuumpump of 400 l/s need 300
hours to pump it away!
Although degassing can greatly accelerate the desorption, part of the material will
decompose and react with the steel. This produces brown carbonaceous residues
that will increase the degassing of the steel for long periods.

Although the surface area of the polyimide is very small in comparison to that of
the vessel, such polymers may give rise to a significant "memory effect". When
exposed to air, polymers will absorb significant amounts of water and gasses,
which will be released again in vacuum. Although polyimide is compatible with
ultra high vacuum, it will slow down the pumping down procedure. The polyimide
rings have a total surface of 13.82 cm2 and a thickness of 0.13 mm. The volume
then is 0.18 cm3.
With 4% moisture-content, this is 72.10-4 g which is 72 10-4
. = 4.10-4 mol water.
18

This produces ≈ 9 mbar l.

The already mentioned vacuum pump will need ≈ 600 hours to pump that away at
10-8 mbar. Bake-out is the solution since the pressure increases then to ≈ 10-6 mbar:
only 6 hours are needed with the same pump capacity

The material of the sample holder depends on the sample to be studied. For specific
samples cleaning and annealing at temperatures of more than 1200º C is necessary.
In those cases the target holders are made from very pure single crystalline
molybdenum. (The single crystalline form facilitates the accurate machining). The
higher purity of the single crystals also reduces the outgassing, which is important
at these temperatures. Flash temperatures of 2500° C are employed incidentally
 29

(see also section 4.3 "Discussion of the vacuum suitcase transport system" page
100).
In fig. 2.2.1 the analyser is seen placed inside one of the two mumetal shields (these
shields are also shown as the dotted lines of the exploded view of fig. 2.1.1 page
16).

Fig. 2.2.1
Complete assembly of the analyser, mounted inside mumetal shields.
The assembly is fixed to the upper flange with "A-brackets" (in this
view the analyser is placed upside down)
30

2.2.2 Electric demands on materials

To deflect back scattered ions having energies of several keV the toroidal electric
fields have to be strong. Strong electrical fields, between the toroidal lenses, ask for
electric potentials of more than 10 kV between these lenses.

The necessary electrical isolation between lenses and other parts is one of the main
problems of the design of the optical layout and of the specific components of the
analyser.
Distances between components have to be dimensioned according to the electrical
potential differences between them.

For mechanical as well as electrical reasons the material alumina (Al2O3) has been
chosen for the fabrication of the isolating parts. Alumina is mechanically and
thermally stable. It can be precision ground to very fine tolerances.
In order to prevent field emission the surfaces of the high-voltage lenses, especially
the ones at a negative voltage (where the electron emission starts) require special
attention. For well-polished surfaces having not too small radius of curvature,
electrical potential differences of 1 kV for 0.5 mm are acceptable (thus 5 mm for
10 kV).
For fabricating delicate and small parts of the detector other materials with good
electrical isolating as well as temperature resistant properties are used. Kapton
(polyimide), Macor (machinable glass) and soft glass are a few of them.
Since the vacuum properties of these materials are not as good as that of alumina,
they have to be used by exception. Also, the electrical conductivity of Macor and
especially soft glass is higher than for alumina, which limits its use in some of the
applications.

2.2.3 Magnetic demands on materials

Magnetic fields will deflect charged particles and thus interfere with proper
operation of the energy analyser. The heavier ions are less affected than electrons.
For low-energy ion scattering a critical part is, therefore, the imaging of electrons
in the detector. Since the double toroidal analyser is also used for the energy
analysis of electrons (e.g. in a photoelectron spectroscopy experiment), severe
 31

constraints are put on the maximum magnetic field in the analyser. The required
high spatial resolution of the energy analyser and detector make this even more
severe.

The allowable magnetic field can be estimated from the deflection of a charged
particle in the magnetic field. When a homogeneous magnetic field of strength B
perpendicular to the motion of a particle of mass m, velocity v, energy E and charge
e, the particle will be describe an orbit, as a result of the so called "Lorentz-force".
The radius R of this orbit (fig. 2.2.3 page 32) is given by:
R = m v / (e B) = (2 m E ) / (e B) (2a)
(Units are here B: [tesla], m: [kg], v: [m/s], E : [Joule], e: [Coulomb]).

In LEIS it is more practical to use for the energy, mass and magnetic fields, the
units eV , mass units and Gauss, respectively. Here the conversion factors are:
1 eV = 1.6 × 10 −19 Joule,
. × 10−27 kg.
the mass of an ion being M x (mass H atom) = M × 167
M being the relative atom mass of the ion.
. × 10−19 Coulomb.
1 Gauss = 10 −4 Tesla and the charge being e = 16
In these units equation 1a becomes:
R( M ) = 1.5 ( M × E ) / B meter. (2b)
Example:
For a 4He+ ion of 1 keV in a magnetic field of 1 Gauss this gives a radius of
curvature of 95 m.

The radius of an electron is given by:

R(e − ) = 0.034 E / B meter (2c)
Example:
For an electron of 100 eV and a magnetic field of 1 Gauss, R = 0.34 m.
32

L
mv

R>>d: sin(α) = L/R ≅ α:

2
d ≅ L/2R

Fig. 2.2.3
Deflection of a moving charged particle in a magnetic field
(perpendicular to the plane of the drawing).
R is the radius of curvature.

When the double toroidal analyser is used for the detection of ions, a magnetic field
will first deflect the ions in the analyser and then the electrons that are generated in
the detector. Since the spatial resolution of the analyser should be 0.1 mm, the
maximum permissible displacement by the magnetic field thus also is 0.1 mm.

When the length L of the path of the charged particle is much smaller than the
radius of curvature R , the deflection d is given by:
d = 0.5 L 2 / R. (2d)

Example 1:
In the DTA's the path-length of the scattered ions is about 20 cm. For a spatial
accuracy of 0.1 mm, the radius of curvature should be (eq. 2d): 200 m. This radius
is obtained (eq. 2b) for 4He+ ions of 1 keV in a magnetic field of 0.5 Gauss.
The electrons that are produced by the ion impact on the first channel plate, and
that are multiplied in these plates, have to travel only a distance of about 1 cm from
the last channel plate to the detector plate. The centre of the electron cloud should
not be displaced by more than 0.1 mm by a stray magnetic field. When the
electrons leave the last channel plate, they have an average kinetic energy of the
order of 10 eV. They are accelerated towards the collector by a few 100 eV.
 33

Assuming an average kinetic energy of 100 eV, the radius of curvature (eq. 2d)
should be less than 50 cm, giving a maximum magnetic field of (eq. 2c): 0.7 Gauss.
Since the magnetic field in concrete steel reinforced buildings is between ¼ and
1 Gauss and welding of stainless steel vessels can give magnetic fields of
0.2–2 Gauss, only moderate precautions would be necessary for the transport of the
scattered ions in the analyser and for the electron cloud to the detector plate.

Example 2:
If the analyser is also used for the energy analysis of the electrons in X-ray
Photoelectron Spectroscopy [XPS], the constraints are much more severe. For
electrons of 100 eV, a path-length of 20 cm and an accuracy of 0.1 mm, the
permitted radius of curvature would be (eq. 2d): 200 m, but the maximum magnetic
field is now (eq. 2c) only 2 milliGauss. This is a very serious constraint on the
magnetic field in the region of the analyser. To avoid the existence of any (stray)-
magnetic fields, the analyser is therefore placed inside two mumetal shields. In
general, alloys with 50 - 80% nickel are used for their particular magnetic
properties. Mumetal in particular (80% nickel) has a very high magnetic
permeability (20.000). After any cold forming and after welding it has to be
annealed. The saturation-value is then restored to approximately 8500 Gauss. The
thickness of those shields is 1 mm; this is quite sufficient to achieve the necessary
low strength of the magnetic field inside the analyser. Two magnetic shields are
used under the presumption that only the outer shield will be magnetically saturated
(see fig. 2.2.1 page 29 and dotted lines of fig. 2.1.1 page 16).
For mounting purposes of various electrical and mechanical appurtenances, holes
have to be cut into the shields. Magnetic shielding is thereby locally interrupted.
Tubes of mumetal have been placed inside those holes to restore the shielding.
To be effective, these tubes have to protrude at least two times their diameter
outside the shields. Also, they have to be in solid contact with the shields to allow
magnetic flux passage without hamper.

Any remanent magnetism of the other metals used, even that of bolts, nuts and
washers etc, cannot be tolerated. During assembly all the metal parts are tested, the
classical way with a magnet, to be sure that no remanent magnetism is contained.
The magnetic field inside the mounted shields has been measured with a Hall-probe
and has a value of 1 - 2 milliGauss.
34

Fig. 2.3.1
A stack of inner elements: superior quality workmanship
 35

2.3 Geometrical accuracy's and requirements

2.3.1 Geometrical accuracy's of the analyser

Deviations of the electrical field between the lenses, caused by small deviations of
the mechanical geometry of the lenses, influence the linearity of image of the
scattered ions.
Small variations of the height of lenses D4 and D2 in relation to D3 and D1 do not
severely affect this image (see fig. 2.1.2).
Eccentricity, however, of lenses D4 and D2, in relation to D3 and D1, respectively,
will lead to an elliptical distortion of the image. An eccentricity of 0.01 mm causes
a distortion of 0.06 mm in the image on the detector [8].

For an energy resolving accuracy of better than 1%, the detector (R = 8 mm inside

and R = 22 mm outside) has to be accurate within 14 ≈ 0.15 mm. This leads to the
100
necessity that accuracy of form and position of the lenses, in the horizontal plane,
has to be better than 015
. ≈ 0.025 mm (see fig. 2.6.3a).
6

All this under the presumption that the detector is ideally concentric with respect to
either the inner or the outer lenses.

Form and roundness tolerances (± 2 µm) of the lenses in stainless steel (AISI 316)
can be achieved. This can be done without going to extreme labour intensive
workshop routines such as lapping and selective surface polishing (fig. 2.3.1).
Excellent workmanship, good quality precision machines and a 3D Co-ordinate
Measuring Machine [C.M.M.] are, however, necessities.
Still the sub-assembly of the inner and outer lenses has to be accurate within
± 20 µm.

2.3.2 Geometrical accuracy's of the zoom lens

The optical properties and the energy of scattered ions are adjusted by the zoom
lens. Accuracy and position here are somewhat less important, compared to the
analyser lenses.
36

Form and position tolerances of ± 0.05 mm are here aimed for (fig. 2.1.2).

2.3.3 Geometrical demands and accuracy's of the diaphragm apertures

The intensity of the scattered ion beam, that eventually hits the first channel- plate,
is determined by the apertures of the diaphragm (fig. 2.1.2). The sizes of the
diaphragm apertures are a compromise between high mass resolution (small
apertures) and sensitivity (large apertures). The optimal adjustment depends on the
application. The apertures determine also the accepted solid angle and the size of
the scattered beam; i.e. how much of the target can be “seen” by the analyser.
Analyser configurations with fixed as well as variable diaphragms are constructed.
The fixed diaphragm configuration has a tolerance in the scattering angle θ of
+ 0°
− 0.3°
145° .
The fixed diaphragm slits then have gaps of ≈ 0.6 mm since the area where the
primary beam (Ø ≈ 1 mm) hits the target, is completely “seen” by the analyser (fig.
2.3.3).

A deviation of max. 1% is the aim for the circular variation of the slits. That means
a variation of ± 6 µm is acceptable for the fixed diaphragms.

Ø 0.75
Primary Ion beam

Diaphragm slits
35

0.6
60

Target

Fig. 2.3.3
θ = scattering angle
 37

2.4 The actual size of the analyser

The dimensions of the analyser are, more or less, dictated by a number of physical
requirements (resolution detector, electrical isolation, sample size, noise, etc) and
by the size of the detector.

2.4.1 The detector

When a scattered ion hits the detector, secondary electrons are produced that are
subsequently amplified in a series of 3 Multi Channel Plates [MCP's]. These MCP's
find widespread application as image amplifiers (e.g. in night vision equipment).
The price of these MCP's is high. Especially for the larger diameters this is a
serious limitation. Physical restrictions of the size of the MCP's relate to their
maximum signal and their dark current. Spatial resolution is not a limiting factor in
the present application.
Since the number of ions that can be converted to electron pulses is fixed for a
given surface area (2.104 cts s-1 cm-2) one favours large areas for coping with high
count-rates. For low count-rates the dark current, which is also proportional to the
surface area, becomes important. For detection of trace elements with LEIS one
thus favours small surface areas. As a compromise MCP's with an effective outer
diameter of Ø 44 mm have been chosen.

This allows one to detect concentrates of about 3.105 cts/s (without serious
distortion of the pulses), while the dark-current remains low enough to detect for
instance a few ppm of gold in a rock sample.

The channel plates are the only parts of the EARISS, ERISS and Calipso analysers
that are not made and/or developed at the Eindhoven University of Technology.
Depending on the central space that is needed for the housing of the primary beam,
a central hole ≈ Ø 16 mm is therefore cut into them (fig. 2.1.2).
The channel plates are manufactured by Philips France.
38

2.4.2 The collector plate

The collector plate has to produce an image of the scattered ions. The image
originates from the first MCP. MCP's produce for each incoming ion a cloud of
electrons. Because of Coulomb repulsion among the electrons and because of the
scattering processes in the channels, an expanding electron cloud is produced at the
exit of the last channel plate. These effects will produce a rotationally symmetric
cloud, having a diameter of a few mm. The centre-line of the cloud intersects
perpendicular the first channel plate at the impact point of the scattered ion. The
structure of the collector plate is such that for each cloud this centre can be
determined. To be effective and to make sure that the proper centre of charge is
determined, the collector plate has to receive the charge of all the electrons in a
cloud. For that reason the outside diameter of the collector plate has to be a few mm
bigger than the last channel plate. The collector plate is a sputtered thin film pattern
on a glass-substrate. The collector plate is shown in fig. 2.4.1. The effective outside
diameter is Ø 48 mm.

Fig. 2.4.1
The collector plate
The sector where the electric wire connections are to be made is seen in the front.
 39

For signal transmitting purposes electric wire connections have to be made on the
surface of the detector plate. This is done in a sector of about 20°. In view of the
expanding charge clouds, this sector is increased to 40º. Effectively, however, a
sector of 40° of the detector plate is supposed not to be useful.

Because this sector is not suitable for imaging purposes it is called the “shaded
sector”. As a result the analyser is not completely rotationally symmetric and the
image of scattered ions covers only 320° of the full azimuth.

2.4.3 The specific geometry of the toroidal lenses

The rough form and dimensions of the toroidal lenses is based on experience with
and inside knowledge of other existing low-energy scattering instruments.
Accurate electron optics calculation have provided the specific geometry of the
lenses (Brongersma, Hellings et al. [8]).

2.4.4 The distance between the target and the analyser

The space between the target and the analyser is especially important for the
combinations of LEIS with other techniques. Optical pyrometry, which is used to
determine the temperature of the target, requires free sight on the target. When the
analyser is used for the energy analysis of electrons in an X-ray photoelectron
spectroscopy experiment, an X-ray source close to the target is necessary.
To enhance the yield of such an experiment even further, it is helpful if the target is
rotated towards the X-ray source. For such applications it is important to have
much space in front of the target.

The electro-optical dimensions of the analyser are thus determined by the size of
the channel plates, the characteristic distance between the channel plates and the
collector plate, the effective collector plate outside diameter and by the specific
dimensions of the lenses. The mechanical translation, the mechanical design, can
now be made.
40

Fig. 2.5.2b
Outer elements in assembled state

Fig. 2.5.2c
Outer elements separated
 41

2.5 Connections between inner and outer elements

2.5.1 As shown in fig. 2.1.1 (page 16) and fig. 2.1.2 (page 22) there are essentially two
"stacks" of elements. The inner elements (fig. 2.5.1 and fig. 2.3.1 (page 34)) and the
outer elements (fig. 2.5.2a, b and c), all do have the same centre-line.
Fig. 2.5.1 shows how the "stack" of inner elements can be assembled to a rigid unit.
Equally fig. 2.5.2a shows how the outer elements can be assembled.
As discussed in section 2.3.1 (page 35) the accuracy of the composition of the inner
and outer elements of the analyser lenses, the inner and outer zoom lens elements,
the diaphragm slits and the position of the detector are deciding factors for the
overall quality of the instrument.

Fig. 2.5.1
Inner elements

Fig. 2.5.2a
Outer elements
42

However, there cannot be any mechanical contact between the inner and outer
elements.
Mechanical contacts (insulating or conducting) would disturb the electrical fields
and thus the ion trajectories in the analyser.
Somehow the inner and outer parts have to be joined together because “you cannot
beat nature”. However, there are two field free zones where the unavoidable joints
between inner and outer elements can be made:
a). The zone just above the detector unit.
b). The zone between the diaphragm slits.
Easy and accurate dismantling of the detector unit rules out option a).

In section 2.4.2 the "shaded sector" of 40° is mentioned. The ions that should be
imaged on this 40° sector of the collector will not be detected and are thus of no
relevance. This reduces the transmission of the analyser by 11%. It follows that,
when spirally trajectories are neglected throughout the analyser, there is a
corresponding 40 ° sector where transmission of back scattered ions is of no
relevance.
The part of this 40° sector, which lies between the diaphragm slits, can thus be used
for joining inner and outer elements without additional loss in the transmission.
The demanded accuracy’s (chapter 2.3 page 35) together with temperature demands
ask for short and, if possible, symmetric joints.
Two small additional attachment zones are, therefore, positioned between the
diaphragm slits. During “bake-out” dimensional changes and accuracy determining
deformations of the analyser will be limited by those joints. These zones are
confined to 3° and, therefore reduce the transmission of the analyser by another
1.5%. The zones are spaced at 120° intervals.
In order to reduce blocking the paths of the scattered ions the least, these zones
have to be small. The remaining transmission of the analyser is 100-(11+2x1.5) =
86%.
 43

2.6 Discussion of solutions

2.6.1 The central hub

A central hub should ensure the concentricity of the inner and outer assembly of the
analyser. This hub consists of an inner and an outer part. These parts are joined
together as rigid as possible. The inner part of the hub is used for precision
mounting the inner elements of the lenses etc. (fig. 2.6.1, part nr. 25).

Fig. 2.6.1
Inner elements mounted on the central hub

The outer part of the hub is not only used for precision mounting of the outer
elements of the lenses, but indirectly also for mounting the analyser in the vacuum
vessel (fig. 2.6.2, part nr. 20).

Fig. 2.6.2
Outer elements mounted on the central hub
44

For joining those two parts of the hub, two useful mechanical solutions were found:
a). Machining the hub from one solid piece of stainless steel and in the meantime
leaving over the attachment zones.
b). Rough machining of the inner and the outer part separately. Vacuum brazing
the parts together with special attachment pieces.
Solution a) is time consuming and the final result is not very accurate because of
stress release of the metal. Furthermore the small dimensions of the “extra
attachment zones” are difficult to machine.
Solution b) is preferred: good stiffness and small signal blocking surfaces of the
attachments can easily be achieved. Good stiffness between the parts permits good
and accurate machining possibilities. Eventually the brazed hub can be fine
machined to the needed close tolerances.

2.6.2 The hollow dowel pins and spacers

The concentricity of the inner elements of the lenses in relation to the outer ones is
difficult to achieve (< 0.020 mm, see section 2.3.1 page 35).
The assembly of the lenses consists of many single elements. These single elements
form a kind of mechanical “chain”. Tolerances of those single elements are then
added-up to give the overall tolerance of the analyser (fig. 2.6.3a).

Fig. 2.6.3a
Analyser assembly without diaphragm slits.
 45

Fig. 2.6.3b

Fig. 2.6.4
Hollow dowel pin and pre-loaded spacers
46

Selective and corrective assembly will always be needed, regardless how well the
general design of the analyser and how accurate the single elements may be
fabricated.
To be successful in that respect, correct kinematical design is a must. Therefore, the
inner parts as well as the outer parts of the lenses are in each case joined with three
supports. The supports consist of hollow dowel pins and spacers of alumina.
The central hub (fig. 2.6.1 page 43) forms the common mounting platform for the
inner and outer elements: all the supports are attached to it.
The carefully ground alumina pins are put into cuttered holes of the metal parts
(fig. 2.6.4). The holes and pins do have a maximum clearance of a few microns.
The position of the holes in the single elements and the position of the holes in the
hub, determine the concentricity of the elements of the lenses.
The distances between parts and their parallelism are determined by means of
alumina spacers fitted around the pins.
The rough heights of the spacers are given by electric demands (chapter 2.2.2 page
30). The exact heights of those spacers are adjusted and tuned depending on the
geometrical needs. This is done by means of grinding and lapping.
Paired elements, pins and spacers, are pre-loaded to one another by means of
specially made stainless steel spring washers (see fig. 2.6.4 page 45).
The built in and thus restricted pre-load prevents the development of uncontrolled
loads during “baking out” periods (see chapter 2.2.1 page 26).

Stainless steel has a higher coefficient of thermal expansion than alumina

(α stainless = 16 µm/m °C, α alumina = 8,5 µm/m °C). A safe temperature
resistance is obtained and alumina parts are not damaged at elevated temperatures.
The accuracy of the construction is determined at 20º C. At elevated temperatures
bigger clearances are developed and the geometry is less accurate.
 47

2.7 The diaphragm with variable apertures

The necessary accuracy's of the diaphragms with fixed or variable apertures have
already been described in section 2.3.3 (page 36).
Openings up to 4 mm are required for the diaphragm with variable apertures. The
variable diaphragm apertures have to be located in exactly the same place where the
fixed diaphragm apertures should be. The problem is the imperative circular form
of the apertures
Existing, accurate, vacuum compatible solutions for variable apertures were not
found. Good diaphragms for optics, like the ones used in photo-cameras, are not
usable. Those diaphragms do not have a uniform cross section for narrow apertures.
Also, these diaphragms are certainly not easy to fabricate in a “bakable” version.
Exchanging fixed diaphragms is not possible because of the already rather crowded
and complicated surroundings of the designated location (fig. 2.6.3a and 2.6.3b).
One of the few remaining feasible solutions can be realised with paired fixed and
moving rings.

Chosen are fixed inner diaphragm rings that are a part of the inner hub (chapter
2.6.1). The outer rings move exactly vertically (parallel to the inner ring). In this
way they effectively change the opening of the diaphragm, seen from the target
position (fig. 2.7.1).

Ø 0.75
Primary Ion beam

Diaphragm slits
35
60

Target

Fig. 2.7.1
48

In principle, one would like to move both the inner and the outer ring of the
apertures symmetrically. In this way the average scattering angle of the scattered
ions would remain the same for all aperture widths. Since at these scattering angles,
the energy of the scattered ions is not very sensitive to the precise angle, this
symmetric movement is of less importance. The construction with moving inner
and outer rings would complicate matters without much gain. Therefore, the
solution with only moving outer rings is preferred. The execution of that solution
had to be of kinematical design (for reasons described in Design principles). It is
then correct to position each single outer ring on three supports and to manipulate
the position of those supports.
In order to obtain good stability and accuracy of position of the rings the supports
must be spaced as widely apart as possible.

It seems a bad solution to attempt simultaneously separate, accurate, in vacuum

measurement and positioning of the three supports for each ring (closed loop
control), over a range of 4 mm (six times for the two outer rings). The positioning
without measurement (open loop control) asks for an accurate synchronisation of
the vertical movement of the supports. The execution of this movement, by means
of rods and rockers, is described below.

In the 2D schematic view (fig. 2.7.2a) the rockers are turned when the incoming
rod B is moved up and down.

fixed inner diaphragm ring

moving outer ring

support
pivot point

pick-up point

transfer rod
rockers
incoming
rod

Fig. 2.7.2a
 49

The effective radii of the various rockers are set to the same value. The "transfer
rods" (the black ones in fig. 2.7.2a) have the same length as the distance between
the turning axes of the rockers. By small rotations all rockers can be seen turning
over the same angle. At the points A, the pick-up points of the supports, movement
in vertical direction is then equal.

In the real 3D version the pivot points of the "transfer rods" and rockers are placed
in a circle (fig. 2.7.2b).

Incoming rod (moving up/down)

Transfer rod

Rocker

Pick-up point

Pivot point
Pivot point

Expanding transfer rod

Fig. 2.7.2b
Horizontal view of one single moving diaphragm ring mechanism.

In order to pre-load the other five rods, one rod can expand by means of a spring.
The common pivot points of the rods and rockers are then clearance-free.
In reality the pivot points are displaced tangentially to the circle of departure. When
the rockers rotate, even slightly, the sum of the distances between the pivot points
changes, because they individually displace to a larger or smaller distance from the
centre of the circle. The expanding rod can also compensate for this variation. The
variation of the sum of the distances between pivot points is a result of fabrication
tolerances of the rods and rockers and their rotation and of the small changes in the
effective working diameter of the pivot points.
50

Fig. 2.7.3b

To move the two separate diaphragm-rings there are two sets of rods and rockers.
The sets are worked into each other and are mounted on the same flange of the
central hub (fig. 2.7.3a).
Small geometrical details determine the final accuracy of the explained solution.
 51

Fig. 2.7.3a
The black rings are the moving diaphragm rings.
52

2.8 The azimuth selector

The azimuth selector should prevent ions having too large an angular momentum
from entering the analyser. As explained before (section 2.1.1 page 24), this is quite
critical because of conservation of angular momentum in the rotationally symmetric
analyser. A small angular momentum will, therefore, lead to a large displacement at
the end of the analyser. For instance, for an azimuthal resolution of 4° on the
detector, the selector should block ions having a velocity ratio (azimuthal
velocity/total velocity) of only 0.004 or more (accepted spread at the selector in the
azimuthal angle: 0.20°).

There is space for the installation of a selector in the field-free zone between the
upper and lower apertures of the diaphragm (fig. 2.7.3a). Several options for the
azimuth selector were considered [10]. The most practical solution, till now, is to
use several identical discs with slits to form a mechanical selector mechanism. The
azimuth selector then consists of five discs, positioned at different, but well-defined
distances. These discs each have 1593 radially oriented slits of 60 µm width. The
discs are concentrically positioned with the axis of the analyser and they have the
same azimuthal position. The distances between those discs are chosen in such a
way as to prevent ions to pass that move at an angle of more than 0,2° with planes
through the axis. By using 5 discs, this is effective up to the 9th order. The discs are
concentrically positioned with the axis of the analyser and they have the same
azimuthal position.

The precise choice of the widths of the slots, the number of slots and the distances
between the discs is a compromise between high transmission (for the non-
spiralling ions), high azimuthal resolution, available space (between the apertures
of the diaphragm) and manufacturing constraints (dimensional stability, accuracy of
etching, etc.). Finally, discs with slots having a width of 62 µm were realised. Since
the first and last disc are 29 mm apart, the slot-width leads already for ideal slots to
an azimuthal spread of 0.12° degree in the selector (more than 2° at the detector).
A good compromise between dimensional stability and acceptable transmission is
obtained by placing the slots at 0.2° intervals (1583 slots in the sector of the disc
 53

that is used, see fig. 2.8.1a and 2.8.1b). The discs are made from 30 µm non-
magnetic stainless steel foil.

Fig. 2.8.1a

Fig. 2.8.1b
One disc of the azimuth selector. The enlargement shows the narrow slits;
springs on the inside of the foil are used to centre the foil in relation to the
axis of the analyser as given by the bore of the hub (fig. 2.7.3a page 51).
54

The technology hereby used is "chemical milling": a combination of photo-

lithography and etching. During the photo-lithographic process an, etchant
resistant, pattern is formed on the metal foil. During the following process, only the
bare metal of the foil is dissolved (the bare metal is not covered by the resistant
pattern). What remains is the designed metal product.
This technology is used mostly for the mass-production of duplicate metal-foil
products.
The five discs of the azimuth selector are duplicates of one another, with the
exception of the radial position of the slits (see fig. 2.7.3a).
Small products with accuracy’s of ± 3 µm can be produced by chemical milling
from 30"µm foil. Large products, like those azimuth selector discs, cannot be manu-
factured with the same accuracy. Deviations of form and measurement of etched
products are related to:
• Differences in etching velocities due to metallurgical differences across the foil
material.
• Thickness variations crosswise and along the rolling direction of the foil.
• The accuracy of the photo-lithographic pattern generating process.
High quality foil, produced in large quantities, will have limited metallurgical
differences. Selected pieces of foil material, taken from the same batch (reel), with
short distances between them, will have even smaller differences.
This foil material is rolled to a thickness of 30 µm with a tolerance of ± 1.5 µm.
This thickness variation is in reality an "etch-depth" variation that will result in a
product tolerance of at least ± 5 µm.
Thickness variations in the rolling direction of the foil, however, will be very small,
certainly when thickness differences are measured over short distances. Therefore
the patterns of the discs to be etched, have to be orientated in the same direction
(the rolling direction of the foil). Etched this way, the different discs will have the
same deviations of form between them.
Compared to the etching process, the photo-lithographic process is very accurate.
In order to limit the differences between the five discs even more, one "mother-
mask" is used. This mother-mask is locally covered to form the specific mask of
one of the five different discs (only the positions of the slits are different).
With the described precautions and measures, series of 10 of the different discs
were produced.
 55

Sets of five different discs were selected. Those sets had a conformity of ± 10 µm
and an absolute accuracy of ± 15 µm.
Because these discs do not have an "out of plane" structural stiffness on their own,
they are laid on thick supporting rings. Only the slots are left open by those rings.

Concentric positioning of the discs, in relation to the central axis of the analyser, is
achieved by means of "nests of springs".
Springs are etched on the inside of a disc; together they form a nest of springs (fig.
2.8.1a). These nests are slipped over the accurately machined bore of the central
hub (fig. 2.7.3a page 51).
Deviations of roundness of the etched discs are "cushioned" this way by the
springs. The concentric positioning of the discs was better than 2 µm.
Azimuthal positioning of the discs is achieved by pre-loading them against an
adjustable bar. This bar is mounted on the central hub and runs through the
rectangular holes of the discs.
Transmission of the azimuth selector was measured with laser-light and a photo-
erray. This was done by mounting a dummy azimuth selector sub-assembly on the
rotating main axis of a Sipp 3D CMM. Measured this way, the azimuthal
transmission varied between 10 and 12%.
The transmission of the genuine analyser, during real ion scattering experiments, is
slightly more: between 12 and 14%.
56

2.9 Design goals and results

Parameter Design goals Results

-9 -10
Vacuum base pressure 10 mbar 10 mbar
-10 -12
Pressure with baking out < 10 mbar 10 mbar
Magnetic field < 2 m Gauss 1 - 2 m Gauss
Electric potential difference 1 kV/0.5 mm > 1 kV/0.5 mm

Geometrical accuracy's
toroidal lenses Concentricity < 0.02 mm < 0.025 mm
zoom lenses Concentricity < 0.05 mm Within design specifications
diaphragm apertures Variation of slits < ± 6 µm Within design specifications
Azimuth selector discs
Temperature resistance No influence of bake-out No noticeable deviations
between calibrations.

Table 2.2

The geometrical accuracy of the analyser cannot be determined when it is

completely assembled. The results given in table 2.2 are obtained from sub-
assemblies, purposely mounted on a rotary table of a CMM.
Calibration of the complete analyser indicates a geometrical accuracy and
temperature stability within specifications.
 57

2.10 Summary LEIS

The development of a new type of analysers, the Double Toroidal Analysers,

required high precision mechanical constructions. Since these analysers are used in
surface science, many other constraints (compatibility with ultra high vacuum, high
voltage, very low magnetic field etc.) are imposed.
Thanks to the success of the design and construction of these analysers, a
spectacular (about a factor of 1000) improvement in the detection efficiency of
electrostatic analysers has opened the way to completely new ideas and possibilities
in the research of catalysis, polymers and ceramics. It is a good example how the
close interaction between designers and scientists enabled breakthroughs in science.
58

2.11 References to LEIS

[1] [Niemantsverdriet, 2000]

J.W. Niemantsverdriet: Spectroscopy in Catalysis. Second revised edition
2000, Wiley-VCN Verlag GmbH, ISBN 3-527-30200-x.
[2] [Grasserbauer, 1991]
M. Grasserbauer, H.W. Werner: Analysis of Microelectronic Materials and
Devices. John Wiley and Sons, ISBN 0471917133.
[3] [Rivière, 1990]
J.C. Rivière: Surface Analytical Technigues; chapter 16, Ion Scattering
Spectroscopy. Clarendon Press, Oxford, 1990.
[4] Goldberg, 1999]
E.C. Goldberg, R. Monreal, F. Flores, H.H. Brongersma, P. Bauer: New
model for ion neutralization at surfaces. Surface Science, 440 (1999), 875-
880.
[5] [Brongersma, 1972]
H.H. Brongersma, P.M. Mul: Ion Scattering a Spectroscopic Tool for Study
of the Outermost Atomic layer of a Solid Surface. Chem. Phys. Lett. 14
(1972), 380.
[6] [Bergmans, 1992]
R.H. Bergmans, P. Brinkgreve, H.H. Brongersma, J.P. Jacobs, C.A.
Severijns en R.G. van Welzenis: Oppervlakte-onderzoek met EARISS.
Tijdschrift voor Natuurkunde 58/14 (1992), 225-228.
[7] [Brongersma, 1990]
H.H. Brongersma, R.A. van Santen: Fundamental Aspects of
Heterogeneous Catalysis Studied by Particle Beams, chapter LowEnergy
Ion Scattering Investigations of Catalysis, NATO ASI Series, ISBN 0-306-
44002-4.
[8] [Hellings, 1989]
G.J.A. Hellings, H. Ottevanger, C.L.C.M. Knibbeler, J. van Engelshoven
and H.H. Brongersma: Potential Distribution and Focussing Properties of
Toroidal Deflection Plates. Journal of Electron Spectroscopy and Related
Phenomena, 49 (1989) 359-382.
 59

[9] [Ackermans, 1990]

P.A.J. Ackermans: Surface Analysis by Low-Energy Ion Scattering: Two
Dimensional Detecting and Qualification. Ph.D. thesis, TU/e, 1990.
[10] [Hellings, 1985]
G.J.A. Hellings, H. Ottevanger, S.W. Boelens, C.L.C.M. Knibbeler and
H.H. Brongersma: Surface Science, 162 (1985) 913.
[11] [Suurmeyer, 1989]
E.P.Th.M. Suurmeyer, J. Verhoeven: Vacuumtechniek. Nederlandse
Vacuüm Vereniging , ISBN 90-9002806-4.
[12] [Weisman, 1972]
C. Weisman: Welding Handbook. Amercian Welding Society. Library of
Congress Card Number: 74-346.
[13] [AWS, 1975]
A.W.S. Committee on Brazing and Soldering: Brazing Manual.
Library of Congress Card Number: 76-15811.
[14] [Maas, 1975-2001]
T.M. Maas: Private Communications.
60

3. High-Energy Ion Scattering: HEIS

3.1 Introduction high-energy ion beam applications: channeling facility

Another example of the fruitful collaboration of a scientific group and the GTD is
the design and engineering of the “Channeling”-facility of the group TIB
(Applications of Ion Beams) of the Department of Applied Physics.
The mechanical parts and their assembly were executed over a period of 2 years.
The installation in the vacuum vessel, the alignment of the ion-beam and the
calibration of the complete instrument took another 1.5 years.
Strict mechanical design-specifications were met. This resulted in a great deal of
professional satisfaction, in particular for the mechanical engineers involved, as
well as for the physicists.
Some remarkable measurements were possible because of the accuracy of this
channeling-facility. The mechanical design and the first measuring data were
successfully presented during the Ion-Beam-Analysis Conference at Balatonfüred,
Hungary 1993 [1], [2].
On a regular base several Ph.D. students and scientists are utilising the facility for
studies on lattice deformations and impurities of grown crystal layers (relevant for
semi-conductor research and industry).

The instrument consists of several sub-assemblies that were exclusively constructed

in the GTD workshops. Almost all-regular mechanical workshop disciplines
contributed, however, the most critical sub-assembly, the “gimbal-head” and drive-
system, was designed and fabricated by just one designing instrument-maker and
the author.
One of the main reasons for the quality of this channeling facility is the exceptional
standard of execution and fine-tuning of this sub-assembly.
Design solutions that were used for this goniometer were used in national
mechanical precision engineering publications [3] and lectures.
 61

3.1.1 High-energy ion beam applications: channeling facility

As stated in chapter 2.1.1 there are numerous surface-analysis techniques that use
ion beams. Among those are the low-energy (0.1 – 10 keV) ion scattering techni-
ques as well as the high-energy (3 – 30 MeV) ion scattering techniques ([4]:
Introduction, [5]: Summary).
Several analytical techniques can be distinguished in the high-energy region alone.
One group of those is called Rutherford Backscattering Spectroscopy [RBS]. In
RBS a high-energy, mono-energetic beam of He+ ions is incident upon a sample.
These incident ions collide elastically with the atoms in the upper layers of the
sample and may scatter back.
When RBS is applied on a monocrystalline sample, channeling can occur.
Channeling is a steering process arising when ions move through a crystal in a
direction almost parallel to a main axis or a plane. This phenomenon will take place
only when the angle between the direction of the incident ion and the crystal row is
small.
During channeling experiments the backscatter yield of an incident ion beam is
measured as a function of the angle of the ion beam, with respect to the main axial
or planar directions of a crystal. To determine those angles and accompanying
yields it is necessary to rotate the crystalline sample over three axes. Therefore, the
crystal has to be mounted on a three axes goniometer. Performing “angular scans”
is possible by rotating the crystal. The yield of the incident ion beam can now be
detected as a function of the goniometer angles (fig. 3.1.1).

2Ψ
Ψ
A B

Fig. 3.1.1
62

Here is Ψ the maximum angle between the direction of the incoming ion beam and
a crystal axis at which channeling still occurs. This Ψ can be derived analytically
([4]: J.A. Davies and [6]: J.A. Davies). The width of the angular scan (2 Ψ ) is
proportional to E −1/ 2 ( E is the incident beam energy). This means that the width of
the scan decreases with the energy of the incident beam. Therefore, a high-energy
ion beam is preferred, when a high angular accuracy is required.
In fig. 3.1.1 angular scans over the main rotation and the vertical tilt axes are shown
for 25 MeV particles on <100>Si.
2 Fu
Scans in opposite direction (plus, minus) exhibit the hysteresis, sv = as
C
discussed in fig. 1.1 (page 8). This is over 0.03º (distance A-B of fig. 3.1.1). The
width of the channeling dip 2 Ψ ≈ 0.2º.
Important parameters are the angular position of the minimum yield in the angular
scan, corresponding to the crystal axes or plane, the width 2 Ψ and the slope of the
dip at both sides of the minimum. The accuracy of the goniometer reflects in the
precision with which the minimum can be determined. The repeatability is the
closeness of the results of measurement when scanning from the same direction,
see: "Vocabulary according to the international Vocabulary of Basic and General
Terms in Metrology" (see page 112).

3.1.2 General design considerations

Not only pure crystals but also multi-layered crystals with lattice deformation can
be studied.
The band gap of semi-conductors is, among other things, very much dependent on
the lattice strain. During the fabrication-process of semi-conductors multi-layered
structures are grown on purpose. These (epitaxially) grown layers have lattice
parameters that differ slightly from those of the substrates on which they are
grown: directions will be affected accordingly.
The main point of concern of the design of the channeling facility is the possibility
of detecting the small deviations in channeling directions. Thus the accuracy, as
defined in 3.1.1, must be better then 0.01º and most likely ≈ 0.001º, particularly at
high beam energy when Ψ < 0.1º.
Of interest are also differences in the minimum yield of the angular scans caused by
defects in crystal structures.
 63

A parallel, mono-energetic incident beam of high-energy in combination with an

accurate three-axis goniometer and an accurate detector system is required for
characterisation of the strain and defects in crystal layers.
The required angular ranges of the three rotations are: 180º main rotation, 360º spin
rotation and ± 5º tilt rotation. The main rotation axis is (located) in the sample
surface, vertically. The spin axis is perpendicular to the sample surface and the
main axis (as wel). The tilt axis intersects perpendicular the main and spin axes in
their intersection (fig. 3.1.2).

Fig. 3.1.2

A particle accelerator produces a high-energy ion beam (MeV ions). There are
linear accelerators, e.q. the Van de Graaf accelerator and circular accelerators such
as cyclotrons. At the Eindhoven University a cyclotron is used.
In a cyclotron ions make very many cycles. Per cycle the voltage (= energy) of
those ions is raised a little bit until it eventually reaches the MeV range. A beam of
MeV ions is extracted from the cyclotron and guided into the scattering chamber
via a beam guidance system.
A precision goniometer, a set of movable detectors and an ion beam are put
together in the scattering chamber. Together they form the channeling instrument or
channeling facility (fig. 3.1.3 and 3.1.4). Ion scattering experiments in combination
with ion channeling can be accomplished with this set-up.
“Energy-loss in transmission” experiments can also be accomplished because of the
unique design of the goniometer head. For this purpose the incident ion beam is to
be transmitted through thin crystal samples. The energy distribution of transmitted
64

particles is then strongly dependent on the lattice orientation (≡ channeling direc-

tions) of the crystal with respect to the incident beam. The channeling effect can be
exploited in several ways and, therefore, it is an excellent tool for the investigation
of (multi-layered) crystalline samples. This project involved building the beam line,
the vacuum vessel with mounting frame, the precision positioning system for the
detector units and the goniometer itself (fig. 3.1.3).
Some commercially available goniometers at that time claimed accuracy's of
≈ 0.01º. However, we never could confirm such specification, nor find sufficient
reproducibility or low enough hysteresis. To corroborate on this we even carried
out elaborate measurements at one of the known companies in this field. Moreover,
no transmission goniometer was available at all.

Fig. 3.1.3 Channeling facility.

The beam line is seen entering through the elliptical hole in the
far sidewall. The scattering chamber is left from the centre.
The introduction of samples takes place through the VAT valve
seen in the foreground.
 65

3.1.3 General description of the channeling facility

The channeling facility consists of three main elements:

• The 6 degrees of freedom goniometer
• The detector stages
• The ion beam vacuum system
These main elements are put together into one vacuum vessel or scattering
chamber. Independence and ease of exchange for maintenance or upgrading the
different sub-assemblies was the main reason for the choice of their position in the
vacuum chamber.
Chosen was a cylindrical chamber with flat top and bottom flanges (fig. 3.1.4).

top flange
-10
2 x 10 m.bar
scattering chamber or channeling chamber

target

detector disk
ion beam

goniometer head: second and third axes

bottom flange

bellows
rod
rotary feed through: primary axis

folded leaf-springs (for z displacement)

X-Y platform

Z-displacement actuator
Z-displacement lever

Fig. 3.1.4
Schematic view of the channeling facility

The bottom flange of the channeling chamber is used as mounting base for the
goniometer. Two detector stages with detector disks are mounted, inside the
chamber, from the top flange. To facilitate accessibility of the detector set-up, is
here the matter of choice.
The ion beam intersects the axial middle of the chamber where also the target
position of the goniometer is situated. The goniometer consists of a translation
66

stage and a rotation stage. The translation stage is outside of the vacuum; the
rotation stage is partly within the vacuum.
A bellows and a differentially pumped rotary feed-through (V.G. DPRF 90) form
the mechanical interface between goniometer and channeling chamber.
The translation stage is hanging from the bottom flange by means of rods (fig. 3.1.4
x-y platform). X-y movement is allowed by elastic lateral displacement of those
rods. Folded leaf springs are mounted on the x-y platform; elastic deformation of
those leaf-springs allows pure Z-displacement of the rotary feed-through platform.
The application of the leaf-springs is so that they do not disturb on the chosen x-, y-
positions while moving in the z-direction is still possible.
The folded leaf spring mechanism is referred to as “compound compliance
mechanism”.
The rotating shaft of the feed-through is used as the primary axis of rotation of the
goniometer. Second and third axes are accommodated in the goniometer-head (fig.
3.1.4. and fig. 3.1.5).
Goniometer-head and attachments are mounted on the end of a large diameter thin
wall tube. Inside the bellows this tube is mounted on the rotating shaft of the feed-
through.

The purpose of a detector is to measure the scattered ion yield of the incident ion
beam during a scan. In order to measure the scattering yield at two different angles
simultaneously, two detectors have been incorporated in the design. Those
detectors are mounted on two-stepper motor driven stages that have the same
rotating axis, located at the middle of the top flange of the chamber. This rotating
axis is aligned, approximately, to the same orientation as the primary axis of the
goniometer.
Because of this geometry the centres of the detector disks can move in circles
around the middle of the goniometer-head (fig. 3.1.4).
The ion beam, the centres of the detector disks and the middle of the target are
situated in the same plane. The ion beam should intersect the target surface, for
obvious reasons.
 67

Fig. 3.1.5
Goniometer-head rear side
Goniometer-head mounted on large diameter (Ø 90 mm) thin wall (1.5 mm) tube.
The centre of the goniometer-head is in line with the ion beam. Mounting tube
protrudes through the bottom flange.
68

Surfaces of the samples mounted in the goniometer have to be aligned x-y-z with
respect to the ion beam. Different positions of incidence, where the beam hits the
sample surface, can then be chosen. Different sample geometries can then also be
accommodated within certain limits. Normally the sample surface is situated close
to two of the rotating axes of the goniometer. The third axis or spin axis intersects
the middle of the sample surface and is perpendicular to it (fig. 3.1.2 page 63).
Swaying of the point of incident over the sample surface, while rotating the sample,
is this way limited by the fixation of the incident beam on the same position of the
sample surface.

Fig. 3.1.6
TU/e sample holder based on kinematic mount principles.
The central hole is Ø 9 mm.

A sample holder permits easy changing of samples (fig. 3.1.6 and 3.1.7). TU/e
sample holders, more or less a TU/e standard for surface analyses, require a
specially adapted “load-lock”. This load-lock, specifically designed for the use of a
TU/e sample holder, can be used in combination with an existing vacuum transport
case. This vacuum transport case is used for sample transport of reactive samples
between a preparation facility and an analysing technique (see Chapter 4: The
vacuum suitcase transport system, page 94). Because of these transport and
mounting tools, powerful combinations with other “in house” analysing techniques
such as EARISS, ERISS, Calipso, Auger, various LEED and target modification
facilities are possible. Those proven tools provide the surface analysing community
of the TU/e with unique possibilities: the channeling facility is one of them.
 69

Fig. 3.1.7
Goniometer-head front side with mounted sample holder.
The mounting tube protrudes through the bottom flange of the channeling
chamber.
70

3.2 Details and components of the beam line vacuum system

differential turbo differential turbo differential

valve pump stage pump pump stage pump pump stage

scattering chamber

ion beam
-10
10 m.bar
-5
3 x 10 m.bar

rotating second beam first

vane slits stop slits

bellows

rotary feed
through

Fig. 3.1.8
Ion beam vacuum system

The purpose of the beam line system is to deliver on the target a stable, accurate
and mono-energetic ion beam of high-energy. A rotating vane and two sets of slits
are necessary tools for this purpose. The rotating vane, placed in front of the
channeling chamber, in the vacuum of the beam line system, is a diagnostic
instrument for determining the intensity of the ion beam. With the help of the sets
of slits, also placed in the vacuum of the beam line, the ion beam can be aligned
with respect to the goniometer target position and these slits also determine the
maximum divergence of the beam. The slit system should be able to limit the
angular spread to ≈ 0.03º - 0.1º. This divergence contributes to the angular accuracy
of the experiment.

3.2.1 Vacuum demands

Samples with reactive surfaces have to be studied. In the channeling chamber

pressures of 10-10 mbar and lower are, therefore, imperative. This is necessary for
measurement of e.g. surface reconstruction effects. However, so far only multi-
layer and transmission experiments have been carried out, which are less deman-
ding on base pressure. As can be concluded from fig. 3.1.8 there are many items
that influence the pressure in the chamber. Gas-flow into the vessel arises from the
rotary feed-through of the goniometer as well as from the beam guidance system.
 71

Outgassing of the surface of the chamber and its accessories is, therefore, not the
only source of vapour that degrades the vacuum.
The leak of the goniometer feed-through into the chamber is strongly reduced by
differentially pumping the feed-through housing. For this purpose a rotary vane-
pump is used (Leybold Trivac D8B).
The beam guidance-system operates at 3 x 10-5 mbar. The pressure drop to the
necessary 10-10 mbar of the chamber is accomplished by three differentially
pumped stages. Those UHV-HV gas-flow restrictions between the pumping stages
consist of tubes having an inside diameter of Ø 20 mm and a length of 400 mm (fig.
3.1.7). In between those restrictions are two large diameter vacuum tubes that are
evacuated by turbo-molecular pumps (Leybold Turbovac CSV 150 l/s). The base
pressure in those vacuum tubes is mainly dependent on the outgassing of their
surfaces and the inleaking gasflow of the beam guidance system.

Surfaces and material outgassing properties.

2
Item Surface [m ] Material

Vessel and tubes 2.3 AISI 316

Goniometer 0.5 (main tube Ø 90, length 300 AISI 316
Assembly parts 0.3 (nuts, bolts etc.) AISI 316
Bellows 2.1 (Ø 140 x 100, 140 pair) AISI 316
Detector discs 0.3 AISI 316
2
5.5 (5.5 m )
-4 2
Outgassing rate after 10 hours 10 mbar l/s m
-4
Outgassing 5.5 x 10 mbar l/s
Vacuum sections 2.0 (length 6000 mm) AISI 316
Beam line 0.4 (length 1200 mm) AISI 316
-4
Outgassing (6000 mm section) 2 x 10 mbar l/s
-4
Outgassing (1200 mm section) 0.4 x 10 mbar l/s

Table 3.1

Of interest is the leak flow rate Q of the beam line into the scattering chamber [11].

The leak flow rate of the first differentially pumped stage into the 6000 mm
section is given by: Q = C × ∆P where

Conductance: C (l/s) and Pressure drop: ∆P (mbar).

72

R3
Here is C = 0.9 l/s (approximation) where
L
10 3
R : diam (mm) and L : length of beam pipe (mm) ⇒ C = 0.9 x = 2.3 l/s.
400
The pressure drop over this section is set to 3 x 10-5 mbar.
This results in Q = 2.3 x 3 x 10-5 ≈ 7 x 10-5 mbar l/s.
The outgassing of the surface of this section is 2 x 10-4 mbar l/s (see table 3.1).
Pumped with 150 l/s, this section will have a base pressure

2.7 × 10−4
of ≈ 2 x 10-6 mbar.
150
With the same reasoning the inflowing leak flow rate into the second
(1200 mm long) section will be 2.3 x 2 x 10-6 mbar l/s.
The surface outgassing of this section is 0.4 x 10-4 mbar l/s (see table 3.1).
−5
The base pressure will be here 4.5 × 10 = 3 x 10-7 mbar.
150
This results in a leak flow rate of the beam line into the scattering chamber
2.3 x 3 x 10-7 ≈ 7 x 10-7 mbar l/s.

A base pressure of 2 x 10-7 mbar in the 1200 mm section has been obtained which
corresponds sufficiently well with the calculated 3 x 10-7 mbar.
It thus can be concluded that not the leak flow rate into the chamber is dominant, but
the outgassing (see table 3.1 page 71). This is mainly caused by the large surface of
the bellows and the goniometer. With the relatively small pump-capacity of 340 l/s a
base pressure of 8 x 10-9 mbar is measured in the chamber.
The design pressure of 10-10 mbar should be achievable if the outgassing is taken
care of by baking out and eventually the pump-capacity is increased.
 73

3.3 Geometrical requirements

The required accuracy's of the channeling facility are amongst others a conse-
quence of the chosen energy of the ion beam.
Meaningful channeling experiments are possible only when a well-defined parallel
ion beam of high-energy can be used (see section 3.2 page 70).
The highest energy of the ion beam (25 MeV) determines the necessary angular
accuracy of about 0.001º. This will be explained below in its consequences on
geometric accuracy, motion accuracy and position accuracy.

Apart from the properties of the ion beam itself, the accuracy of the channeling
facility is also determined by the geometrical stability between the ion beam slits,
the detectors and the goniometer-head.
Geometrical stability between the goniometer-head and the detectors is dependent
on the stiffness of the goniometer, the detector stages and the vacuum vessel.
A cylindrical vessel constructed with thick top and bottom flanges is very stiff by
nature. However, several apertures have to be welded into the wall of the vessel.
Local heat input during this welding brings about local distortions of the cylindrical
wall. Distortion influences negatively the natural stiffness of the vessel. As a
countermeasure a wall thickness of 3 mm instead of the theoretical necessary 2 mm
is chosen, thereby increasing the out of plane bending stiffness with a factor 3.4.

At the Eindhoven University of Technology ions with a maximum energy of

30 MeV can be extracted from the cyclotron. For 25 MeV He ions in <100> Si the
maximum angle for channeling Ψ = 0.09°. Therefore, the “well-defined” and
parallel ion beam must have an angular spread smaller than 0.09°.
For accurate measuring of channeling phenomena the accuracy's of the goniometer
have to be at least an order of magnitude smaller than 0.09°, but to establish the
shape of the channeling dip precisely we arrive at a design goal of 0.001º.
The detectors have to be capable of dissolving small differences in the yield of the
incident ion beam, which are angle dependent. For this purpose angle accuracy is
aimed at 0.1º. This requirement is less strict than for channeling directions. This is
due to the kinematic reaction parameter, which determines the energy of the
scattered ion beam.
 74

3.3.1 Accuracy's of the beam line

The supports of the beam line vacuum system and the channeling chamber are
placed on the very solid concrete floor of the TU/e Cyclotron Building. This
concrete floor is therefore the mounting platform of the channeling vessel and the
beam line vacuum system.
Temperature changes of the floor will be small, due to the specific characteristics of
the Cyclotron Building. Those temperature changes pass over long periods
(compared to the actual measuring time of channeling experiments). Therefore,
geometrical changes of the floor due to temperature changes need not be
considered.

In the beam line vacuum system two sets of slits are placed 7.68 meter apart. Each
set of slits can be adjusted in two perpendicular directions. Per direction the
adjustments have a range of 0 - 20 mm. With those slits the alignment of the ion
beam, with respect to the channeling vessel, and the angular spread of the beam can
be adjusted. A 2 mm gap between slits limits theoretically the angular spread of the
2×2
= 5.2 x 10 rad ≈ 0.03° (see fig. 3.1.8 and also section 3.2).
-4
beam to
7680
This compares favourable here with the maximum angle for channeling Ψ = 0.09°.
This angle and the spread together determine the measured width of the channeling
dip.
Measurements indicate that the influence of “mechanical noise” on the positional
stability of the concrete floor is small. Jumps in the velocity Vb are in the order of
0.1 mm/sec. For the scattering chamber these can result in a vibration amplitude of

U0 = Vb m (see page 11).

C

The support of the chamber has a horizontal stiffness C ≈ 5 x 106 N/m.

The mass of the chamber (complete with all fittings) m ≈ 150 kg.

150
This results in U0 ≈ 0.1 x 10-3 ≈ 0.5 x 10-6 ≈ 0.5 µm.
5 × 10 6

Because of the limited stiffness of the A-frame beam line supports the mechanical
amplitudes of the beam line will be substantially larger. These larger amplitudes are
still of no importance to the accuracy of the alignment of the beam line in relation
to the scattering chamber and the attached goniometer.
 75

3.3.2 Accuracy's of the x-y-z translation stage

The goniometer consists of two sub-assemblies: the translation stage and the
rotation stage. Both are of statically determined design. High mechanical stiffness
and low weight of the goniometer are necessary conditions to obtain the desired
accuracy. Vibration amplitudes will be limited because of this high stiffness and
low mass in combination with the low “noise” level of the concrete floor.

Displacements of the x-y-z translation stage are relatively small. These are achieved
by elastic deformation of specific elements like rods and compound compliance.
Clearance free displacements are imperative for accurate constructions.
Displacements by means of elastic deformation are clearance-free and therefore
preferred. The turning angles of the rotation stage, however, cannot be realised by
means of elastic elements. The necessary large angles of rotation (see section 3.1.2
page 62) do not allow for this.
Ball bearings and ball-joints are used for the construction of this sub-assembly.
These are pre-loaded to exclude all clearance. However, virtual clearance (e.g.
2 times the friction divided by the stiffness, page 8) cannot completely be avoided
this way. By the use of low friction bearings and hysteresis free joints in
combination with construction elements of high stiffness this position-error can
only be limited, but should not exceed 0.03º (see section 3.1.1 page 62).

X-y translation:
Because of the statically determined design there will be virtually no thermal
memory and no thermal snapping in the elastic displacement x-y-z stage. The
influence of thermal effects on the accuracy of the construction is therefore limited
only to geometrical variations as a result of the thermal expansion of construction
elements. X- and y-displacements are used for a single adjustment to compensate
for geometrical deviations of the channeling facility and for different target sizes.
X- and y-displacements of ± 8 mm are possible. The deviations are a result of the
fabrication tolerances of components as well as deviations of the alignment of the
ion beam. Apart from alignment corrections, the z-displacement is also used
actively during experiments. Z-displacement (± 12 mm) is used to change the
incident position of the ion beam on the target in the goniometer-head.
76

Fig. 3.3.2
X-y-z translation stage
In the front is seen the x-y platform suspended from the bottom flange.
Two of three vertical rods can be seen. The rods can be seen to pass
through holes in the upper leaf-springs.
The sub-frame (the black tubes) attached to the x-y platform is connected to
the bottom flange by adjustable rods (see fig. 3.3.3). In the middle of the
x-y platform is the z-translation stage, with the rotary feed-through.
Two pair of folded leaf springs of the compound compliance can be seen,
mounted between the x-y platform and the z-stage.
 77

The x-y platform is supported from the bottom flange of the vessel by means of
three rods of fixed length (x-y platform fig. 3.1.4 page 65 and fig. 3.3.3). This
platform is a lightweight and stiff welded sheet metal construction. A sub-frame,
fabricated from tubes, is welded onto this platform. Between the sub-frame and the
bottom flange of the vacuum vessel three adjustable rods are fitted (fig. 3.3.3).

vacuum vessel = scattering chamber

bottom flange

bellows upper constriction

rod adjustable rod (one out of three)
subframe

X-Y platform

lower constriction

Fig. 3.3.3
Schematic view of the x-y translation

The adjustable rods together with the three rods of fixed length form a highly rigid
and statically determined connection between the bottom flange and the x-y
platform.
The x- and y-displacements are enabled by the elastic deformation of the upper and
lower constrictions (elastic hinges) in the rods.
Accurate translation of the x-y platform in relation to the bottom flange is possible
this way. Over the maximum stroke of 14 mm this results in a vertical (z)
displacement of less than ± 0.03 mm.
X- and y-displacements are adjusted by varying the length of the three adjustable
rods (fig. 3.3.2 and fig. 3.3.3).
The horizontal and vertical stiffness between the bottom flange and the x-y stage
varies between 1 x 106 and 3 x 106 N/m.
78

Fig. 3.3.4
X-y-z translation stage
Three vertical rods between the bottom flange of the vacuum vessel and the
x-y platform can be seen. One (horizontal) adjustable rod between bottom flange
and subframe is seen. One pair of folded leaf springs is seen in the front.

Fig. 3.3.5
The bottom flange of the z-translation stage.
One bearing of the worm wheel can be seen. Three mounting places,
radially extended, for leaf springs are seen on the outside of this flange.
Motors and encoders for the second and third axes drives are in the middle.
The motor and encoder for the primary axis is seen on the top left together
with one end of the Hooke’s joint pre-loading shaft.
 79

Z-translation:
By means of a compound compliance translation mechanism the z-stage is attached
to the x-y platform. This type of mechanism is ideally suited here because it leaves
open a lot of space in the centre of the x-y stage, where the rotary feed-through and
worm wheel drive are situated. The z-displacement is achieved by the elastic
deformation of the radially extended leaf-springs of the compound compliance.
The wide axial and radial spacing between the sets of leaf-springs results in high
lateral stiffness of the location of the z-stage (fig. 3.1.4 page 65).
Six folded leaf springs, in a symmetrical layout, instead of the necessary five, are
used for this mechanism.
Symmetry of motion, together with an unambiguous vertical drive force, and lateral
stiffness in two planes justify this choice.
Fabrication of more accurate parts and extra care, to obtain near stress free
assembly of this “over determined” construction, make this solution feasible.
Compound compliances of monolithic design is nowadays preferred in Eindhoven
for small strokes in a relatively small volume.
For the same reasons mentioned above, these are of over determined design also,
but of a much smaller size. However, being fabricated out of one piece of material,
there is no extra effort needed for the realisation and matching of accurate parts.
Problems during assembly of those over determined constructions do not occur
either.

Summary:
To achieve the necessary accuracy of the channeling facility the mechanical
stiffness of the x-y-z translation stage is of primary concern. Adjustments x, y and z
must be unambiguous and free of hysteresis but not necessarily very accurate.
Furthermore, during channeling experiments, displacements are of static nature.
The positional accuracy of the x-y-z translation stage will be limited by mechanical
amplitudes as a result of mechanical noise of the surroundings.

3.3.3. Accuracy of the rotation stage

The rotation stage has to perform three rotations independent of each other with a
high degree of accuracy that is imperative for channeling experiments (< 0.01°).
80

The primary rotation is performed by means of the rotation of a rotary feed-

through; the other two are performed by the goniometer-head.
The main features that contribute to the accuracy of the rotation stage are:
• clearance free statically determined design,
• low mass of the goniometer-head,
• high stiffness of the foundation of the goniometer-head,
• low friction design,
• high gear-ratio (i) drives integrated in the design of the goniometer-head, for
drive stiffness increases with i2,
• pre-loaded temperature resistant bearings,
• high quality classical craftsmanship for the construction of this delicate
instrument for maximum research competitiveness,
• application of full modern workshop technologies.
Those features are integrated in the design and execution of the rotation stage.
Therefore, they will not be dealt with individually, but will be discussed in
connection to one another. It follows that the construction of this design has to be
carried out with a lot of “finesse” and almost all construction activities have to be
geared to one another.
For example: the out of roundness of a worm wheels of the drives determines not
only the accuracy on its own, but also the amount of pre-load that is necessary to
make the concerning bearings clearance free and stiff. Also the friction torque of
those bearings is largely determined by this pre-load and therefore is of major
influence on the virtual backlash of the drives.

The primary rotation (range 360º):

One end of the shaft of the rotary feed-through (≡ primary axis) serves as the
foundation for a large diameter thin walled tube (Ø 90 x 1.5: being 20 times as stiff
and still of lighter weight than a Ø 25 mm solid bar, the commercially available
competitor uses a Ø 22 mm thin walled tube). The goniometer-head is mounted on
top of this tube. Solidly attached to the other end of the shaft of the feed-through is
a worm wheel. This worm wheel rotates between two large diameter pre-loaded
angular contact bearings that are spaced wide apart because they are mounted in the
top and bottom flange of the z-translation stage (fig. 3.3.5 page 78).
 81

Fig. 3.3.7a

Fig. 3.3.7b
Pre-loaded worm worm wheel drive for HR-LEED. Pre-loading by means
of elastic hinges. Hinges are EDMed in the front and rear chassis plates.
In the middle of fig. 3.3.7a is the TU/e sample holder. A: worm bearing
assembly.
82

A very stiff suspension of the goniometer-head is created this way.

The feed through housing is “floating” in its own seals around the shaft. The seals
of a differentially pumped feed-through can develop a lot of friction.
As a result the friction torque, between shaft and housing, can be quite high (in this
case up to 40 Nm initially). Low friction seals consequently had to be specially
made.
An A-bracket and elastic strut (fig. 3.3.6) blocks rotation of the feed-through
housing.
In general an A-bracket and strut is an excellent solution for blocking a rotation
torque in a statically determined way: it is simple, clearance free and stiff.

Fig. 3.3.6
A-frame and rod, still to be lowered to the sub-frame.

The drive of the feed-through is made backlash free by pre-loading the driving
worm in its bearings onto the driven worm wheel. For a long life and high stiffness
the axes of the worm and worm wheel have to be perpendicular within close
tolerances (< 0.01 mm, measured from the shaft of the worm).
Pre-loaded worm wheel drives of the rotation stage, for the primary as well as the
secondary rotation, are developments of the drive of a 5 degrees of freedom HR-
LEED manipulator designed by the author in another project (fig. 3.3.7a and fig.
3.3.7b page 81).
 83

Here the pre-load is applied by means of a rotation around the axis formed by
elastic hinges. Although the HR-LEED manipulator drive worked well, it has been
improved for the channeling rotation stage. As can be seen, the worm bearing
assembly cannot translate and rotate out of the plane (bearing assembly A of fig.
3.3.7a and b). The “best fit” between worm and worm wheel is therefore not
obtainable.
By allowing two further degrees of freedom an improved construction for the drives
of the rotation stages is realised. For the primary drive this is done by means of a
Hooke’s joint shaft for the support of the worm bearings assembly. The Hooke’s
joint shaft allows for two perpendicular crossing rotations. Put into two V-blocks
the shaft and the attached worm bearing assembly can move sideways as well as
rotate around two axes. Because of the extra rotation of the shaft, a best fit between
worm and worm wheel is possible. Stiffness of the drive is improved noticeably
and durability is improves at least theoretically.
Running in the drive was used to verify a correct tooth contact. This resulted in a
constant drive torque. The way to check the contact areas of worm and worm wheel
is by means of a thin coat of engineers blue applied to the worm wheel teeth.
One end of the Hooke’s joint pre-load shaft can be seen in fig. 3.3.5 (page 78).
Still a better solution would be possible when the sideways and angular movements
of the worm bearing housing were realised by elastic elements. This kind of design
is not applied because the necessary displacements for pre-loading the worm and
worm wheel were not known at the design stage.

The second and third rotation (range 180º and ± 5° respectively):

The goniometer-head that performs the second and third rotation is mounted on top
of the thin wall tube. Targets are located in the goniometer-head with their surfaces
on the intersection of the three rotation axes (fig. 3.1.7 page 69).
Another worm wheel drive, of the described design, performs the second rotation
(≡ spin rotation). By means of a “gimbal mount”, between the worm wheel and a
funnel-shaped gimbal body, the third rotation (≡ tilt rotation) is made possible
independent of the spin rotation (fig. 3.3.8 page 85). A gimbal mount gives
rotational freedom about two perpendicular axes as used for the suspension of
nautical compasses. The gimbal mount does not influence the spin rotation of the
gimbal body that is determined by the position of the worm wheel.
84

Fig. 3.3.9
Goniometer rear side.
The “two part garden gate”, a truss type structure, that determines only one
degree of freedom, is mounted between the gimbal body and the thin wall tube.
 85

Fig. 3.3.8
The spin rotation is brought about by the worm wheel.
Spin rotation of the gimbal body is still possible: the
coupling between gimbal body and A-type lever is
by means of a large diameter (3 balls) bearing.
Sample holders are mounted on one side of the gimbal body.
Surfaces of the samples will be located on the intersection
of the three goniometer rotating axes.

The funnel-shaped body originates as a result of the design requirement that

energy-loss and angular spread of scattered particles in transmission experiments
also have to be taken into account. Within a limited angular domain the ion beam
can be transmitted through the targets that are mounted on top of the funnel.
Without countermeasures the gimbal body rotates with the worm wheel but it can
rotate anywhere (around both the axes α and β of the gimbal mount) in the limited
space of the worm wheel rear side. The funnel of the gimbal body ends in a large
diameter ball bearing (fig. 3.3.8 and fig. 3.3.9). The outer ring of this four-point
contact bearing is held in position by a “two part garden gate” (a variation of the
folded leaf spring to be used in situations where large displacements are demanded)
86

and an A-type bracket. Gimbal body, together with the mounted sample holder, can
now be tilted by a lead-screw that displaces the A-bracket.
The assemblies of the goniometer within the vacuum need to be of temperature
resistant design (imperative for baking-out routines up to 200° C).
For the construction of precision instruments pre-loaded contact bearings are
preferred. There are no quality contact bearings of stainless steel on the market.
Separate bearing assemblies mounted into bodies in vacuum always create
undesirable virtual leaks, between the bearing cup and the body. For these reasons
special bearings have to be constructed for the drives of the goniometer. All
bearings for worms as well as worm wheels (and for the detector disk drives as
well) are integrally made in house. All are pre-loaded by membrane. The exact
position and mechanical characteristics of the membranes ask for special attention
(e.g. fig. 3.3.8 Pre-loading membrane, page 85).
The out of roundness and the out of plane of the fixed tracks versus the rotating
tracks determine, for one part, the variation of the pre-load force. Dimensional
changes due temperature differences e.g., as a result of baking out procedures,
determine this variation as well.
The geometrical accuracy of balls (of metal as well as ceramics) is incredibly high
as a result of mass production and selection techniques.
Design considerations of ball bearings in instruments often come down to: a large
number of balls between tracks or, the statically determining, three balls between
tracks.
Under sufficient pre-load the stiffness of bearings with numerous balls will be
much higher than those with three balls. Elastic deformation of the numerous
Hertzian contacts between ball and track has to be large compared to the accuracy
of balls and tracks. Under this condition all balls will be in contact and stiff
bearings will be the result.
In this construction there are no separate ball bearing assemblies within the
vacuum. All bearing tracks are machined directly into the rotating bodies and
frames.
Trailing the spin rotation of the worm wheel is the only function of the bearing on
the big end of the gimbal body, therefore, it is extremely low loaded. However, the
friction in the bearing itself is the point of concern. Therefore only three balls are
employed in this four-point contact bearing.
 87

The bearing diameter is larger than the distance between the gimbal mount pivot
points.
An unfavourable transfer ratio between the original friction forces of the bearing
and the resulting loads on the gimbal pivots is the result (fig. 3.3.7).
Position and accuracy of the gimbal mount is of major influence on the accuracy of
the spin as well as of the tilt-rotation. Therefore, a low and constant load on the
pivot points is required.
The friction of the bearing has also to be low to minimise the virtual backlash in the
drives of both the spin and the tilt rotations,
Stiffness of this bearing is of secondary concern. For obvious reasons the out of
roundness of this bearing has to be very small.
A special four-point contact bearing with only three balls is constructed. It is
carefully run in and it is fitted with a special ball cage to hold the balls, with low
friction, at 120° intervals. The outer ring is carefully machined to obtain the
required pre-load for a low, constant and long lasting friction torque.

All the vacuum bearings and pivot points are constructed with ruby balls. These
balls are readily available in different sizes and they are very accurate (∆R < 0.0001
mm per ball) and relatively cheap to buy.
In combination with stainless steel and UHV compatible Aluminium bronze
(CuAl10Ni) these balls have excellent engineering qualities.
There is no cold welding between mentioned materials and ruby even with
relatively high specific contact pressures for stainless steel 316 (> 500 N/mm2) and
baking out at 200° C.
In vacuum the self-lubricating qualities of ruby are reasonable. Friction between
parts is low and the quality of contact areas will not detoriate with time.
Instruments can be constructed by the appropriate use of ruby balls, with care of
both the manufacturer as well as the designer.
Stiffness and durability of ball bearings also are dependent of the geometrical fit
between balls and tracks.
Precision-machined bearing tracks are run in with steel balls. During run in the pre-
load is carefully raised in small steps to get a controlled amount of cold work into
the surface of the tracks. Eventually the tracks show a correct marking of the areas
88

of contact between balls and tracks. For final assembly ruby balls replace the steel
balls.
Rotations of the goniometer-head are driven by stepper-motors. Those motors are
mounted outside the vacuum on the bottom of the differentially pumped rotary feed
through shaft (fig. 3.3.5 page 78).
Bellows type rotary feed-throughs are used for the transfer of the drives to the
vacuum side. Special clearance free drive shafts are mounted between those feed-
throughs and the reduction gears of the goniometer-head. Clearance free Hooke’s
joints, comprising two forks at right angles, are used at both ends of these drive
shafts (fig. 3.3.10). By means of those joints the drive can be transmitted through an
angle. In this case the cross-pieces of the Hooke’s joints are made clearance free by
pre-loading them against the V- planes of the forks.

Fig. 3.3.10
Clearance free Hooke’s joint

The essential lengthening of the drive shafts is made possible by special diaphragm
shaft assembly’s (one of those shaft-assemblies is seen under the goniometer-head
in figs. 3.1.5 page 67 and 3.1.7 page 69). Between the centre and the outside of a
thin diaphragm a limited elastic out-of-the-plane displacement is possible.
 89

Fig. 3.3.11
Diaphragm shaft

The centre of such a diaphragm (0.05 mm) is micro plasma welded on a stainless
steel precision tube. A shaft with a dish and Hooke’s joint on one end and a three-
facet piston on the other end is slipped inside this tube (fig. 3.3.11).
When the outside of the diaphragm is welded onto the dish a drive shaft emerges,
that is rotation stiff and clearance free. This shaft is still capable of small
displacements between the centres of the Hooke’s joints.
90

3.4 Discussion of design goals and results

Lattices of pure crystals are an absolute reference standard for angle measurement.
The hysteresis and angular accuracy of the goniometer have been measured using
25 MeV He+ ions impinging on silicon <100> oriented crystals (see ref. [5]).
Calibration results obtained with this reference are given below in table 3.2.
Hysteresis is avoided when related angular scans are measured in the incidental
direction.

Parameter Design goal Experimental result

Goniometer Range Accuracy Range Accuracy

Main rotation 180° 0.001° -85° ↔ 257° < 0.001°
Spin rotation 360° 0.001° -136° ↔ 226° < 0.001°
Tilt rotation ± 5° 0.001° ± 9° < 0.001°
Hysteresis rotations 0.03° 0.03°
Repeatability < 0,001° < 0.001°
X translation ± 5 mm ± 8 mm
Y translation ± 5 mm ± 8 mm
Z translation ± 5 mm ± 12 mm
Vacuum
-8 -9
Base pressure 10 mbar 8 x 10
-10
with baking out 10 ?

Table 3.2
 91

3.5 Conclusions HEIS

The accuracy of the rotation stage of the channeling facility is obtained by careful
attention to the features as mentioned on page 80.
Design and engineering are integrated successfully as can be concluded from the
measured angular accuracy < 0.001° and a hysteresis < 0.03° for three axes. This
has been successfully applied and demonstrated in literature on solid state problems
(refs. [7], [8, [9] and [10]).

Fig. 3.5.1
Detector stages (top) with mounted detector disks, goniometer-head (centre)
and the ion beam entry port (right) seen together through the window of the
channeling vessel.
92

3.6 References to HEIS

[1] [Dijk, 1997]

P.W.L. van Dijk: High-energy ion channeling facility and its applications,
Ph.D. thesis, TU/e 1997.
[2] [Dijk, 1993]
P.W.L. van Dijk, L.J. van IJzendoorn, P. Brinkgreve, M. Janszen and
M.J.A. de Voigt: Channeling with high-energy particles using a 3 - 30 MeV
cyclotron, Proceedings IBA Conference Balatonfüred, Hungary (1993).
[3] [Soemers, 1989]
H.M. Soemers, D.M. van Dijk, P. Brinkgreve: Hoge instelnauwkeurigheid
met conventionele technieken, Mikroniek 29 (1989) 140-149.
[4] [Thomas, 1976]
J.P. Thomas and A. Cachard: Material characterization using ion beams, see
J.A. Davies, NATO Advanced Study Institutes Series, Plenum Press, ISBN
0-306-357283.
[5] [Rivière, 1990]
J.C. Rivière: Surface Analytical Techniques, Clarendon Press, Oxford,
1990.
[6] [Feldman, 1986]
Fundamentals of surface & thin film analysis, Feldman & Mayer, North
Holland (1986).
[7] [Selen, 2000]
L.J.M. Selen, L.J. van IJzendoorn, P.M. Koenraad and M.J.A. de Voigt:
Evidence for strain in and around In As quantum dots in Ga As from ion
channeling experiments, Physics Review, 8270 (2000) 61.
[8] [Selen, 2000]
L.J.M. Selen, F.J.J. Jansen, L.J. van IJzendoorn, M.J.J. Theunissen, P.J.M.
Smulders and M.J.A. de Voigt: Strain measurements in ultra-thin buried
films (< 50 Aº) with RBS ion channeling. Nucl. Instr. and Meth., 492
(2000) 161-163.
 93

[9] [Selen, 2000]

L.J.M. Selen, F.J.J. Jansen, L.J. van IJzendoorn, M.J.J. Theunissen, P.J.M.
Smulders and M.J.A. de Voigt: RBS ion-channeling for strain analysis in
buried nano-films (< 6 mm). Submitted to Journal of Applied Physics.
[10] [Selen, 2000]
L.J. Selen, P.J.M. Smulders, L.J. van IJzendoorn and M.J.A. de Voigt:
Planar ion-channeling on buried nano-films discriminates between ion-atom
potentials. Submitted to Physics Review Letters.
[11] [Suurmeijer, 1989]
E.P.Th.M. Suurmeijer, J. Verhoeven: Vacuumtechniek, Nederlandse
Vacuüm Vereniging. ISBN 90-9002806-4.
94

4. The vacuum suitcase transport system

4.1 Introduction of the vacuum suitcase transport system

The range of techniques applied for the characterisation of materials is extremely

wide (see also chapter 3.1.1 page 61). This is on the one hand due to the wide scope
of the materials to be investigated (catalysts, ceramics, composites, polymers, semi-
conductors etc) and on the other hand a consequence of the great variety of
properties that scientists are interested in.
Since each technique imposes its own perspective on the material under study, a
careful choice of complementary techniques is generally required (see [1] "catalyst
characterisation" and [2] "spectroscopic techniques").

Each technique has its specific spectral domain, resolution, sensitivity and
conditions. It thus is common practice to characterise a given material sample with
many different analytic tools. For problems where the outermost atoms of the
surface play an important role, transfer of a sample in the open air, under normal
atmospheric conditions, is generally out of the question. It would lead to
fundamental changes of the composition and structure of the surface. This implies
that preparation (including techniques for cleaning and surface modification) as
well as all relevant analysis techniques should be included in the same UHV
system.
In a modern surface science laboratory, the number of necessary techniques is very
large. For instance, in the group Surface and Interface Analysis at the TU/e some 9
analysis and 10 preparation techniques are regularly used in its research.
A combination of these techniques in one system would create serious logistic
problems. Also, the possible cross-contamination of samples and the inevitable
maintenance of such a system would pose extreme difficulties and lead to major
downtime of the entire system.

Stand-alone systems, that contain only the most essential techniques for a certain
type of project, are therefore used at the TU/e. Transport of samples between the
various set-ups is realised here by using a transport vessel (the so-called “vacuum
suitcase”). The transport takes place under vacuum or in a special inert atmosphere.
Without exposing the samples to the outside atmosphere, the following analytical
 95

and preparation facilities can be combined: EARISS, TOF-LEIS, ERISS,

CALIPSO, SPA LEED, XPS, Auger, NODUS, Mini Mobis, High Resolution EELS
together with glove boxes for spin coating of polymers and UV treatment,
sputtering facilities, temperature and gas processing, ALE and differential pressure
set-up ([3] "differentially pumped pressure cell").

In the early days of surface science at the TU/e it still was possible to implement
the transfer systems and special load-locks (the sub-assembly of a vacuum system
where the samples are locked in or out) in a relatively easy way. In view of the
importance of a transfer capability in surface science, the group Surface and
Interface Analysis (leader: Prof.dr. H.H. Brongersma) received in the eighties a
major grant from the Netherlands Foundation for the Advancement of Science
(ZWO/NWO) to realise this in Eindhoven.
A vacuum suitcase, universal sample holder and transfer mechanisms were
developed. Since then, new equipment was practically always designed in such a
way that transfer was possible. It would be an immense task to modify all existing
set-ups now. At present, samples can be transferred among most of the set-ups of
the groups of Surface Interface Analysis (Prof.dr. H.H. Brongersma) and Physics
and Applications of Ion Beams and Accelerators (Prof.dr. M.J.A. de Voigt). Also
specially designed sample holders are exchangeable between set-ups of the two
groups.

4.1.1 General design considerations

Although the transfer principle is very simple, it is rarely used in other laboratories
and certainly not on such a large scale as in Eindhoven.
The reason for this is twofold: an integral transfer system between different
operational set-ups is difficult to implement afterwards, commercially available
manipulators and sample holders are not compatible to one another.
For the ease of operation of analytical and preparation techniques it is usually
necessary to mount the materials under study on sample holders. The great variety
of analytic and preparation techniques leads to many special requirements on the
mounting of those samples and thus on the versatility of the sample holders. The
mechanism of transferring the samples from the suitcase into the different set-ups
(and vice versa) also has to meet many special demands.
96

Vacuum transfer of samples is hampered also by problems encountered when the

suitcase is connected to the various set-ups. Transfer should be such that the
samples are absolutely not exposed to air. Thus after connection and before the
suitcase is opened, the space between suitcase and set-up has to be evacuated.

4.1.2 Vacuum demands

In view of the discussion in chapter 2.2.1 (page 26) one may think that it is essential
that the sample transfer takes place under UHV conditions. This is very difficult to
achieve and would make the transfer very cumbersome and time consuming.
Fortunately, there are many specimens that are not very reactive towards the restgas
in the vacuum suitcase. The absence of energetic radiation (ions, electrons, photons),
which would lead to reactive dissociation products, helps in this respect. The
situation is similar to that of the high purity gloveboxes used for sample preparation
(spincoating of polymers, catalysts etc.) and for mounting of the samples. In those
gloveboxes the atmosphere contains less than 1 ppm oxygen and water. This means
that the partial pressures of those constituents are less than 10-3 mbar. In a vacuum
suitcase having a base pressure of only 10-5 mbar, the partial pressures are thus
already 2 orders of magnitude lower than in the best gloveboxes!
Reactive specimens will get contaminated in this vacuum suitcase, but the large
reduction in the partial pressures of the reactive gases, in comparison to the
pressures of those gases in the ambient atmosphere, reduces the contamination
enormously. This is especially true, since the transfer can be fast (within a few
minutes). Since the diffusion of the contamination from the surface into the sub-
surface layers of the specimen generally requires the passage over an energy
barrier, the mild exposure to the contaminants in the glove box and the vacuum
suitcase will generally limit the contamination to the outer surface. This is even true
for some very reactive materials like steel and aluminium. To generate a clean
surface after transfer with the suitcase, a relatively simple treatment performed in
situ in a specific set-up, is therefore often sufficient.
 97

Fig. 4.2.1
The vacuum suitcase
The side port can be used for an extra vacuum pump and/or a pressure gauge.
The lower picture shows the opened suitcase with the sample holder stack
protruding through the mounting flange.
98

4.2 Requirements of the vacuum transport system

4.2.1 Requirements for the vacuum suitcase (fig. 4.2.1)

• Easy transfer to the manipulators of the various set-ups.

• Simultaneous storage of a number of samples in the suitcase, enabling the
transfer of several samples at the same time.
• Vacuum 10-5 mbar or better.
• Coupling and uncoupling of the suitcase without exposing the samples to air.
• Small and easy to carry for short transfer times.
• Ease of coupling and decoupling.

Fig. 4.2.2
TU/e sample holder

4.2.2 Requirements of the universal sample holder (fig. 4.2.2)

• Maximum sample size Ø 10 mm.

• Non magnetic materials must be used (see chapter 2.2.3 page 30).
• Easy mounting of samples.
• Low heat capacity.
• Short distance from the sample to the heating element that is to be mounted
underneath the sample holder.
• Design and materials must be resistant against flash temperatures of
2500º C.
• Ease of coupling and de-coupling to transfer mechanisms and manipulators.
 99

• Unambiguous positioning in the target positions.

100

4.3 Discussion of the vacuum suitcase, load-lock and TU/e sample holder

The basis of the design of the vacuum suitcase is the TU/e sample holder. Being of
statically determined design this sample holder is easily and accurately positioned
in a sample holder/adapter (fig. 4.3.1)

Fig. 4.3.1

At appropriate positions, the geometry of this adapter is machined into different

assemblies and set-ups. By means of a bayonet coupling, mounted on top of an
elastic nose-piece, the different sample holders can be caught (fig. 4.3.2).

Fig. 4.3.2
Elastic nose-piece with bayonet coupling.
 101

Loading and unloading the suitcase and transport of samples within the set-ups is
usually accomplished by means of a Magnetic Linear and Rotary Motion Drive
[MLRM] manipulator.
A MLRM uses a magnetic coupling to impart linear and rotary movements to a
shaft. The shaft is completely enclosed inside a thin walled stainless steel tube that
is part of the vacuum system of the particular set-up (fig. 4.3.3 page 103).
Sliding over the tube is an actuating (steering) magnet. The elastic nose-piece is
mounted on the far end of the shaft.

When the suitcase is coupled to the load-lock entry valve and before this valve is
opened, the space enclosed between suitcase and the outer side of the gate valve is
evacuated. For that purpose the housing of the commercially available gate-valve is
specially reworked.
After opening, the sample holder stack of the suitcase is advanced to the coupling
position. Here the sample holders are transferred to the manipulator of the
particular set-up.
Movements of the MLRM, necessary for catching the sample holders, can be seen
through a glass view-port and manually actuated by moving the actuating magnet.

The sample holder design with two pairs of mushroom shaped stubs provides also
another transfer possibility. Perpendicular to the transfer rod that couples and de-
couples the sample holders from the suitcase, a different transfer mechanism, that
uses the same method of coupling and de-coupling, can be implemented. For a
more elaborate vacuum system this creates an extra design possibility.

Sample holders are accurately positioned (0 - 0.05 mm) by two mushroom stubs
and the centre of a horizontal lip (fig. 4.3.1).
During transfer the stubs slide into a semicircular guide-way and the horizontal lip
slides over a plane, oriented through the middle of the guide-way. Together, guide-
way and plane, form the adapter.
Between sample holder and adapter 5 degrees of freedom are defined (four of those
restrictions between the guide-way and the stubs and the fifth between slip and the
plane).
The sliding motion of the transfer cycle is arrested by an adjustment screw or a
solid stop machined into the rear side of the adapter. Springs or elastic mechanisms
102

machined into the metal of the adapter impose relatively low forces on the contact
areas. For positioning accuracy the springs must be stiff in the sliding direction
(limited virtual clearance! page 8).
Contact forces result in friction that retains the sample holders in position after the
transfer cycle is completed and the MLRM is withdrawn.
Unavoidable misalignments between the orientation of sagging transfer rods (free
lengths of more than 1 meter are common practice) and adapters are compensated
by the elastic nose-piece (fig. 4.3.2 page 100) and the widened entry side of the
guide-way.
Sample holders and adapters are made of different materials for specific purposes.
The most common type sample holder is fabricated from stainless steel 316, the
mushroom stubs are laser spot welded in the body. This type can be used up to
400º C. Sample holders are coated with TiN to prevent binding to the adapter at
temperatures up to 600º C. Recently sample holders are fabricated from the exotic
mono crystalline Mo for 2500º C + flash temperatures. At those temperatures the
sample holder (≈ 20 mm outside measurement) will have a substantial elongation:
∆l ≈ 2500 x (20 x 10-3) x 10 = 500 µm (related to a stainless steel adapter). The
design of the adapter has to cope with this.
TU/e sample holders are not only used for normal or high temperature applications,
but also for extremely low temperatures ([4] "cryogenic ion beam applications").

Conclusion:
A continuous development of sample holders made of special materials takes place
and the number of set-ups in which they are applied is still growing.
This proves the versatility and user-friendliness of the TU/e sample holder. The
vacuum suitcase transfer-capability, among a wide and still growing variety of
analytical and preparation techniques, has established itself as being an important
tool for the surface science and surface analysing community of the TU/e.
 103

Fig. 4.3.3
104

4.4 References to the vacuum suitcase transport system

[1] [Brongersma, 1990]

R.W. Joyner in “Fundamental Aspects of Heterogeneous Catalysis
Studied by Particle Beams”. Eds H.H. Brongersma and R.A. van Santen,
Plenum Press, NATO ASI Series B 265 (1991), ISBN 0-306-44002-4.
[2] [Niemantsverdriet, 2000]
J.W. Niemantsverdriet: Spectroscopy in Catalysis.
Wiley-VCH Verlag GmbH; ISBN 3-527-30200-X.
[3] [Jansen, 2000]
W.P.A. Jansen, A.W. Denier van der Gon, G.M. Wijers, P.W. van den
Hoogen, J.A.M. de Laat, T.M. Maas, E.C.A. Dekkers, P. Brinkgreve, H.H.
Brongersma: Differentieel gepompte drukcel, NEVACBLAD, 38 (2000)
83-87.
[4] [de Jong, 1997]
M.P. de Jong, A.J.H. Maas, L.J. van IJzendoorn, S.S. Klein, M.J.A. de
Voigt, Journal of Applied Physiscs, 82 (1997) 1058.
 105

Conclusions

By the implementation of the design principles in the design of the LEIS analysers and the
HEIS channeling set-up unique specifications are met:
• The LEIS analysers do have a spectacular improvement in sensitivity (about a factor of
1000).
• The HEIS channeling set-up has an angular resolution below 1/1000 degree. This is
unique in the world for a small laboratory scale set-up.
Those instruments were the result of a successful team effort of experimental scientists and
mechanical engineers. Because of this collaboration it was possible to start the execution of
those instruments immediately after the agreement upon the design specifications. This
was done successfully on an experimental instrument building basis: designing and
building at the same time. This was possible because the design goals and the workshop
possibilities were already tuned to one another in the design specifications. This way
valuable time was gained because useless details, that have no relevance to functionality,
were not fabricated.
106

Personal remarks

How to obtain the professional skills and insight of a mechanical designing engineer, how
to expand them and how to keep up may be impossible to indicate, but looking at the way a
creative (obsessed) hobbyist developed into a designing engineer that is an active member
of a successful designing and engineering group may be of help here.
Apart from the inventive-creative qualities, the required knowledge and skills for the
designing mechanical engineer consist mainly of two categories: the theoretical-analytical
and the practical-analytical ones.

The theoretical-analytical knowledge was obtained during the studies mechanical

engineering at the TU/e.
Through the years my practical-analytical knowledge has vastly expanded as a result of a
continuous contact with graduate students, Ph.D. students and the scientific staff of this
University. Contacts with the group Constructions and Mechanisms gave continuous
impulses. From the beginning of my designing activities Prof. ir. W. van der Hoek and in
particular my late friend ing. N. van Dijk and later on dr.ir. N. Rosielle were always
willing to discuss, analyse and give response in order to reach better design solutions.

Long before the mechanical-engineering studies were attended, a lot of machinery had
already passed through my hands, most of it with disastrous results and now and then with
a simple success.
The necessity and willingness to understand specific design and realisation solutions
stimulated practical-analytical thinking. The need of keeping third-hand mopeds, motor
bikes and automobiles in running conditions, being the main cause.
Well known to me were the basics of machining: during the secondary school it was
possible to attend special classes of the Philips Industrial School on a voluntary basis.
Filing, drilling, turning, milling and tool sharpening were learned under the supervision of
classical professional teachers. As a result of the accumulation of practical knowledge, the
relevance of theoretical subjects of the TU/e studies was decided.
Aware and unaware, the attention (interest) given to different subjects of the studies was
rather different (the marks were varying accordingly).

For designing engineers the usefulness of "penser avec les mains" should not be under-
estimated. Practical knowledge of basic machining and of machines and their components
 107

are a necessity in order to be able to practise it on a useful level. Testing and evaluation of
my final graduation subject and of many professional subjects later on, would never have
been possible without practical insight and knowledge.
The fabrication of realistic models and critical parts of proposed designs still are and will
remain a useful "design tool". As a counterpart to computer modelling and computer
analyses it is of incalculable value.
Neither has to be confused with the fabrication of machine-parts by professionals. In
principal every reasonable talented hobbyist can fabricate most of the needed parts of
constructions, if need be in an adapted form. Really accurate dimensioning, reproduci-
bility, low production times and other qualities of machine-parts can only be achieved by
trained professional craftsmen.
The skilled instrument-makers of the GTD achieve that, but they also attended lectures on
specialist engineering subjects in order to understand better design principles and
solutions. Because of this they are more able to give relevant comments and contributions
during the process of optimising proposed design solutions. The GTD achieves optimising
of designs by a continuous dialogue between designers and fabricators, even during the
execution stage. This way these instrument-makers are uniquely trained at the cutting edge
of functional design.
The integration of functional design and professional execution regularly results here in
frontier shifting solutions.
On itself this is reason enough to bring the appreciation of craftsmanship to another level.

For the designing engineer there is more than understanding why and what there is to
design and fabricate. The saying: "As a punishment you should make it yourself" is also of
major importance for any design and engineering process. This is an important aspect of
functional design: getting the best result for the least effort.
The amount of labour that is going to be invested in any proposed design has to be decided
on this basis.
Competing in any mature technical sport makes this point of departure extremely clear.
Time schedules that have to be met, in combination with the need to be competitive, oblige
to focus on the essential elements of the sport: being there next time when "the flag drops",
is one of the most important ones. Whether it be an Olympic sailing campaign, car-racing,
outboard racing, or the very demanding free flight model air-plane competition, in the end
it all comes down to the same: functionality of technical design.
108

In any technical sport a continuous process of servicing, repairing and development of

material is going on. Depending on the complexity of the specific competition material this
can be a time consuming process.
The functionality and quality of this material (and the results of the last race) determine on
which part of the process time has to be spent. For a given team the deadline for departure
to the next event on the race-calendar, however, limits the amount of available time and by
that determines the options.
By building, modifying and continuously up-dating competition material a lot of practical
experience was obtained, but also the awareness of the necessity of quality: "without
quality it will not work at the right moment" (f.i. R.V. Jones: "Most Secret War").

Mechanical instruments and apparatus consist of an endless collection of construction

parts and details. In most cases it is not the quality of the design alone, but also the quality
of execution of the components that determine the merits of a complete instrument.
Attention to detail is the deciding factor here for the practical usefulness of an instrument.
But there is still more than the quality of a design and its execution. For instance the
difference between theoretical design objectives and the requirements that matter in actual
use can be enormous: a satisfactory result arises when there is no difference between those
two.
My first designed and built racing car and the inevitable end of it, against an uprooted
tree, made that very clear to me.
A few years later, when a friend died in a racing-car that was prepared by me, yet another
relevant element was added: be sure that design and execution have been done
professionally. This means you have to be sure that things have been done the best possible
way.
For designers and engineers there is always the compulsive obligation to be sure that
delivered equipment fulfils all the necessary requirements, which can be much more than
that what the principal asks for.

In a professional technical world there is no room for trifling work and make shift
solutions: attention has to be paid to the necessary details and requirements that possibly
have not been explicitly stated but are imperative for reliability and usefulness.
 109

Looking back at the studies mechanical engineering, I realise that systematic thinking,
analysing and calculating of dimensions were added then to my "blacksmith"-level of
practical thinking.
An already existing instinctive way of finding solutions for design problems remained, but
ways to systematise, analyse and determine dimensions were acquired. A more or less
"lettered" technician was formed, a technician that knows what types of supporting
knowledge there are and how to obtain them.
Intuitive originated solutions could be judged on their merits in an early stage of the
design process, usable elements were recognised and selected.
As a compensation for the theoretical accent of the studies, the years were filled with
"penser avec les mains". This was done on a rather professional level, because of the easy
access to all thinkable machining possibilities, and the response of the professional
technicians that worked on the research programmes of the University.
Racing-cars, circuit-racing powerboats, motorbikes, special trailers for Olympic sailing
campaigns and not to forget the continuous updating and bringing to perfection of my
private competition material (masts and spars, fittings, sails) were important subjects. The
continuous challenge of working with modern fibres for structural purposes started then.
This was an instructive period of active mechanical engineering that consumed much more
than 40 hours weekly, not only as a result of the demands from professional boat-racer and
racing car driver friends.
Most of this time was spend in a competitive atmosphere, resulting in an extra stimulus to
achieve the necessary functionality of solutions, because there was always the time
pressure of one race-calendar after another.

At the University a substantial part of the weekly "design evaluation sessions" of

Prof. W. van der Hoek, where the work of graduate students was to be evaluated, were
dedicated to the technical evaluation of the competition results and competition disasters
(blown engines, broken masts, uncontrollable road-holding, almost won championships
etc).
One thing and another were experienced in groups and not on an individual basis,
therefore this was an extremely fascinating and intensively instructive period.
In connection with the foregoing sessions more in depth knowledge was obtained: from
those design evaluation sessions originated the "design principles" that are dealt with in
110

the so-called "Devils Picture Book", the name of the lecture-book of

Prof. W. van der Hoek. Since 1985 this lecture-book became the responsibility of
Prof.dr.ir. R. Koster and since 1990, when Prof.dr.ir. R. Koster transferred his teaching
activities to the University of Twente, dr.ir. N. Rosielle continued their work at the TU/e.
Those design principles and examples of good engineering practice were often a direct
result of thorough analyses and evaluation of the final design subjects of the engineering
graduates of Prof. W. van der Hoek.

In the beginning those weekly evaluation sessions could be compared with a sort of
"technical market place", where a continuous process of bargaining about technical merits
of design elements and principles was going on.
However, familiar with σ = E. ε , I learned at that time to find a challenge in realising the
consequences in full depth.
Already in 1676 Robert Hooke formulated the basic statement of linear elasticity. About
100 years later Thomas Young proclaimed that the proportionality between stress ( σ ) and
strain ( ε ) is : σ = E. ε . Hence E : Youngs Modulus. For elastic deformation this is the
constitutive equation. This law is a very useful approximation of the behaviour of most
materials within the range of recoverable small strains and is commonly known to all
technicians.
As a rule, however, materials are intrinsically non-linear. Nowadays the interest of
scientists is directed to problems related to material that is non-linear.
For me the consideration of fundamental choices for any design originate here: can it be
executed with purposely built-in elastic deformation of elements or has it to be realised
with, preferable low friction, rolling components; what are the fundamental reasons for a
specific design solution etc. etc. During those sessions my engineering insight was formed.

Also during that time it became very clear to me that successful engineering is a team
effort. When the goal of that effort is to get the best results for the least effort: then design
engineers and makers together focus their efforts on functionality. The strategy to reach
that goal is communication. Therefore, they all have to speak the same technical language
and their management should be aware of the tremendous advantages and landslide
differences w.r.t. conventional design and fabrication.
The guiding principle for me and my colleagues over the last 30 years has been: defining
design qualities in terms of kinematic and static constraints, together with the actual
fabrication of instruments based on the same principles as well.
 111

Judged by the amount of fabricated useful instruments, of a large variety, it has proven to
be an extremely inspiring and workable principle.
112

Summarised vocabulary according to the "International Vocabulary of Basic and

General Terms in Metrology"

1. Accuracy (of measurement): Closeness of the agreement between the results of a

measurement and the true value of the measurand.
2. Accuracy (of a measuring instrument): Ability of a measuring instrument to give
responses close to a true value.
3. Repeatability (of results of measurements): Closeness of the agreement between the
results of successive measurements of the same measurand carried out under the same
conditions of measurement.
4. Repeatability (of a measuring instrument): Ability of a measuring instrument to
provide closely similar indications for repeated applications of the same measurand
under the same conditions of measurement.
5. Reproducibility (of results of measurements): closeness of the agreement between the
results of measurements of the same measurand carried out under changed conditions
of measurement.
6. Sensitivity: change in the response of a measuring instrument divided by the
corresponding change in the stimulus (input signal).
7. Resolution: smallest difference between indications of a displaying device that can be
meaningfully distinguished.
8. Stability: ability of a measuring instrument to maintain constant its metrological
characteristics with time.
9. Uncertainty of measurement: parameter, associated with the result of a measurement,
that characterises the dispersion of the values that could reasonably be attributed to the
measurand.
10. Measurement: set of operations having the object of determining a value of a quality.
11. Principle of measurement: scientific basis of a measurement.
Example: the thermoelectric effect applied to the measurement of temperature.
12. Method of measurement: logical sequence of operations, discribed generically, used in
the performance of measurements.
13. Results of a measurement: value attributed to a measurand, obtained by measurement.
Notes: When a result is given it should be made clear whether it refers to:
• the indication
• the uncorrected result
• the correct result
 113

and whether values are averaged.

14. Indication (of a measuring instrument): value of a quantity provided by a measuring
instrument.
15. Sensor: element of a measuring instrument or measuring chain that is directly affected
by the measurand.
• rotor or a turbine flow meter
16. Detector: device or substance that indicates the presence of a phenomenon without
necessarily providing a value of an associated quantity.
• halogen leak detector.
114

List of used acronyms

AES Auger Electron Spectroscopy

AISI American Iron and Steel Institute
ALE Atomic Layer Epitaxy

Calipso Composition Analysis by Low-Energy Ion Scattering for Probing

the Surface on the Outside
CERN Centre Européenne pour la Recherche Nucléaire
CVT Continuously Variable Transmission
CMM Co-ordinate Measuring Machine
cts counts
DDP Des Duivels Prentenboek: Lecture book on mechanical design
principles
dof Degree of freedom
DTA Double Toroidal Analyser
EARISS Energy and Angle Resolved Ion Scattering Spectroscopy
ECM Electro Chemical Machining
EDM Electro Discharge Machining
ERISS Energy and Resolved Ion Scattering Spectroscopy
GTD Central Design and Engineering Facilities
HEIS High-Energy Ion Scattering
HR- EELS High Resolution Electron Energy Loss Spectroscopy
HR-LEED High Resolution Low-energy Electron Diffraction
IBA Ion Beam Analysis
ISS Ion Scattering Spectroscopy
LEIS Low-Energy Ion Scattering
LIS Leiden School for Professional Instrument-makers
MCP Multi Channel Plate
MRI Magnetic Resonance Imaging
Mini Mobis Modular Back-scattering Ion Spectrometer
MLRM Magnetic Linear and Rotary motion drive Manipulator
NLC National LEIS Centre
NODUS Non Destructive Ultra-sensitive Spectrometer
RBS Rutherford Back Scattering
 115

SPA-LEED Spot Profile Analysis-LEED

TC Thermal Centre
TIB Applications of Ion Beams
TIG Tungsten Inert Gas welding
TOF-LEIS Time of Flight - LEIS
TU/e Eindhoven University of Technology
UHV Ultra High Vacuum
UV Ultra Violet
vp virtual play
ZWO/NOW Netherlands Foundation for the Advancement of Science
XPS X-ray Photo-electron Spectroscopy
116

Samenvatting
Ontwerp en vervaardiging van instrumenten met unieke specificaties
Toepassingen voor oppervlakte fysica en kristal analyse

De GTD ontwerpt en vervaardigt experimentele apparaten en instrumenten ten behoeve

van het wetenschappelijk onderzoek van de TU/e. De ontwerpen van deze apparatuur
komen tot stand door een gezamenlijke inspanning van wetenschappelijk medewerkers,
ontwerpers en uitvoerende technici, die gericht is op het vinden van functionele
oplossingen. De gevolgde strategie is die van overleg, waarbij betrokkenen een zelfde
technische taal spreken. Deze taal is gebaseerd op kennis en inzicht van gemeenschappelijk
begrepen en aanvaarde ontwerpprincipes. In belangrijke mate komt dit neer op het defi-
niëren van ontwerpcriteria op basis van belemmeringen en vrijheidsgraden: het kine-
matisch en statisch bepaald construeren.
Door deze principes toe te passen, kunnen perfecte oplossingen voor nagenoeg alle
ontwerpvraagstukken worden gevonden. Dit geldt zowel voor de algemene werktuigbouw
als voor de precisietechnologie.
Ten opzichte van de gebruikelijke overbepaalde constructies, zijn kinematisch en statisch
bepaalde constructies nauwkeuriger en beter te dimensioneren en desondanks goedkoper
doordat minder machinale bewerkingen nodig zijn voor de fabricage van onderdelen.
De GTD past deze ontwerpprincipes reeds 30 jaren toe en een selectie van daarmee
gerealiseerde instrumenten wordt genoemd. Een drietal experimentele instrumenten ten
behoeve van de analyse van oppervlakken en materialen, waarvoor ik verantwoordelijk
was voor de constructie, wordt uitvoerig behandeld om te illustreren dat aan unieke
specificaties kan worden voldaan door het toepassen van deze principes:
• De dubbel toroïdale (LEIS) analysatoren.
Hierbij worden edelgas ionen met een bekende en lage energie (0,1 - 10 keV) op een te
onderzoeken oppervlak geschoten. De energieniveaus van de teruggestrooide ionen
geven informatie over de samenstelling van het beschoten oppervlak; de richting van
de teruggestrooide ionen geeft informatie over de structuur van het oppervlak.
De energieniveaus worden bepaald met behulp van een uiterst nauwkeurig
elektrostatisch (dubbel toroïdaal) lenzenstelsel, waar hoge elektrische potentiaalver-
schillen tussen de verschillende systeemdelen worden aangebracht.
Ionenverstrooiing aan sterk reactieve materialen is alleen in ultra hoog vacuüm (UHV)
mogelijk, reden waarom strenge eisen worden gesteld aan de materiaalkeuze en de
temperatuurbestendigheid van de constructie.
 117

De constructieve eisen zijn deels tegenstrijdig en tevens complex door het grote aantal
geometrische, elektrische, magnetische en vacuümtechnische eisen.
De (succesvolle) analysatoren hebben een gevoeligheid die een factor 1000 beter is dan
die van vergelijkbare elektrostatische ionenverstrooiers. Hierdoor zijn nieuwe ideeën
op het gebied van onderzoek naar katalyse, polymeren en keramiek ontstaan.
• De hoge energie (HEIS) channeling opstelling.
Channeling is een proces dat kan plaatsvinden wanneer ionen door een kristal bewegen
in een richting die nagenoeg parallel is aan een hoofdrichting of een vlak van dit
kristal.
Naarmate de energie van de ionen(bundel) groter is, wordt de hoekafwijking, tussen de
bewegingsrichting van de ionen en de hoofdrichting van het kristal waaronder
channeling nog plaatsvindt, kleiner: naarmate de energie van de ionenbundel groter
wordt, kunnen de hoofdrichtingen en roostervlakken nauwkeuriger bepaald worden.
Op de TU/e wordt de hoog energetische ionenbundel (3 - 30 MeV) van het Cyclotron
in combinatie met een 6 graden van vrijheid channeling goniometer gebruikt om
metingen aan kristallen te verrichten.
Deze goniometer met de bijbehorende detectoren is geplaatst in een vacuümvat waarop
het ionenbundel geleidingssysteem van het Cyclotron is aangesloten.
Er worden hoge eisen gesteld aan deze goniometer. De belangrijkste is het grote
instelbereik over drie assen met een nauwkeurigheid < 0,001º.
De geometrie van de goniometer is zodanig dat ook metingen in transmissie (de bundel
gaat door het monster heen) mogelijk zijn.
Om aan reactieve monsters te meten, moet het vacuüm in het vat beter zijn dan 10-8
mbar. Uitstookbaarheid van het in vacuüm gedeelte van de goniometer en van het
detectormechanisme is dan een voorwaarde.
De channeling meetopstelling voldoet ruimschoots aan de gestelde eisen. Dit wordt
o.m. bewezen door de succesvolle metingen aan halfgeleider kristallen zoals beschre-
ven in de daaraan gerelateerde publicaties.
• Het vacuüm transportsysteem.
Er zijn veel technieken nodig om materialen te kunnen karakteriseren. Niet alleen is de
verscheidenheid van te onderzoeken materialen groot, maar ook zijn veel verschillende
materiaaleigenschappen van belang en daarom moeten veel verschillende metingen
worden verricht. Iedere meettechniek heeft zijn specifieke werkgebied. Eén materiaal
wordt daarom met verschillende meettechnieken onderzocht. Normaliter heeft iedere
meettechniek een specifieke materiaal- monsterhouder.
118

Het combineren van meerdere meettechnieken in één systeem heeft (te) veel nadelen.
Stand-alone meetsystemen hebben de voorkeur. Indien transport van monsters tussen
de verschillende meetsystemen nodig is, wordt het blootstellen aan de atmosfeer een
nadeel. Dit nadeel wordt verholpen met het vacuüm transportsysteem, waarmee
monsters onder vacuümcondities worden verplaatst. Dit systeem moet dan wel
technisch/fysisch verantwoord, efficiënt en gemakkelijk te bedienen zijn.
Er is een universele monsterhouder, een vacuümtransport koffer en een transfer
mechanisme ontwikkeld. Deze monsterhouder wordt in een twintigtal meettechnieken
en behandelkamers toegepast, waaronder de LEIS analysatoren en de HEIS channeling
opstelling. Deze installaties zijn ingericht voor het gebruik van die ene en dezelfde
statisch bepaalde monsterhouder.
Deze monsterhouders worden voor bijzondere toepassingen eventueel van speciaal
materiaal vervaardigd (flitstemperaturen tot 2500º C zijn toegepast).
Dit unieke vacuüm transportsysteem heeft bewezen een belangrijk gereedschap te zijn
voor het onderzoek van de oppervlakte fysica en voor de oppervlakteanalyse aan de
TU/e.
 119

Publications

1. P. Brinkgreve, C. Snijders, D.M. van Dijk: Foot size measuring apparatus has photo-
sensors moved longitudinally and transversely of foot to produce signals when illumi-
nated from opposite sides of foot.
Patent NL 7900596 (A 19800729), EP 14022 (A 19800806) EP 14022 (B 19841114),
DE 3069615 (G 19841219).
2. P. Brinkgreve, D.C. Koningsberger: Kinematic X-ray analysis apparatus - has
backlash-free movement mechanism interlinking source, crystal and centre positions.
Patent NL 8300927 (A 19841001), US 4637041 (A 19870113), CA 1223093
(A 19870616), EP 118965 (B 19880525), DE 3471525 (G 19880630), NL 184386
(19890201).
3. P. Brinkgreve, T.M.J. Maas, D.C. Koningsberger, J.B.A.D. van Zon, M.H.C. Janssen,
A.C.M.E. van Kalmthout, M.P.A. Viegers: A linear spectrometer designed for EXAFS
spectroscopy. Proceedings of an International Conference, Springer Proc. Phys.
(1984), 2 (EXAFS Near Edge Struct. 3) 517-19.
4. F.W.H. Kampers, F.B.M. Duivenvoorden, J.B.A.D. van Zon, P. Brinkgreve,
M.P.A. Vliegers, D.C. Koningsberge: Laboratory EXAFS spectrometer, principles and
applications. Solid State Ionics, vol. 16 (1985) 55-63.
5. P. Brinkgreve: Bouw en ontwikkeling van de EXAFS goniometer. De Constructeur,
26 (1987) 6-38.
6. P. Brinkgreve: Bouw en ontwikkeling van de EXAFS goniometer. Mikroniek, 27
(1987) 33-36.
7. P. Brinkgreve, D.C. Koningsberger: Kinematic X-ray analyses apparatus.
Patent US 4637041 (19870113 (cited in 005 later patents)), US 1984-588280
(19840312), NL 1983-927 (19830315), US 4637041 (19870113).
8. F.W.H. Kampers, T.M.J. Maas, J. van Grondelle, P. Brinkgreve, D.C. Koningsberger:
An in situ cell for transmission EXAFS measurements on catalytic samples. Review of
Scientific Instruments, vol. 60 (1989) 2635-8.
9. H.M. Soemers, D.M. van Dijk, P. Brinkgreve: Hoge instelnauwkeurigheid met
conventionele technieken. Mikroniek, 29 (1989) 140-149.
10. R.H. Bergmans, P. Brinkgreve, H.H. Brongersma, J.-P. Jacobs, C.,A. Severijns and
R.G. van Welzenis: Oppervlakte-onderzoek met EARISS. Nederlands Tijdschrift voor
Natuurkunde, 14 (1992) 225-228.
120

11. R.H. Bergmans, P. Brinkgreve, H.H. Brongersma, J.P. Jacobs, C.A. Severijns,
R.G. van Welzenis: Surface study with EARISS [energy and angle resolved ion
scattering spectrometer]. Nederlands Tijdschrift voor Natuurkunde (1992), 58 (14)
225-8.
12. J.I.M. Botman, C.J. Timmermans, B. Xi, H. Heller, H.L. Hagedoorn, P. Brinkgreve,
E. Dekkers, J. Moerel: Dipole design for the Euterpe storage ring. Proceedings of the
1993 Particle Accelerator Conference, Proceedings of the IEEE Particle Accelerator
Conference, New York, USA, vol. 4 (1993) 2892.
13. P.W.L. van Dijk, L.J. van IJzendoorn, P. Brinkgreve, M. Janszen and
M.J.A. de Voigt: Channeling with high-energy particles using a 3 - 30 MeV cyclotron,
Proceedings IBA Conference Balatonfüred, Hungary (1993).
14. R.W. de Leeuw, J.E. Copens, T. Drouen, C.J. Timmermans, J.I.M. Botman,
H.L. Hagedoorn, P. Brinkgreve: The accelerating cavity of the Racetrack Microtron
Eindhoven, Proc. Eur. Part. Acc. Conf. Barcelona (1996).
15. R.W. de Leeuw, J.E. Coppens, T. Drouen, C.J. Timmermans, J.I.M. Botman,
H.L. Hagedoorn, P. Brinkgreve: The accelerating cavity of the Racetrack Microtron
Eindhoven. EPAC96, Fifth European Particle Acelerator Conference, Institute of
Physics Publishing, Bristol, UK (1997) 1964-6.
16. P. Brinkgreve, E. Dekkers, P.C.J.N. Rosielle: Frit Jig Assembly.
Patent US 609371 (20000725), US 1999-400961 (19990922), EP 1998-203230
(19980925).
17. P. Brinkgreve, E. Dekkers, P.C.J.N. Rosielle: Frit Jig Assembly for cathode ray tube,
has bolt with a resilient metal pin which is flexible in direction transverse to the
longitudinal axis of the bolt.
Patent US 6093073 (A 20000725 (200038)), EP 1051723 (A1 2001115 (200059)),
18. W.P.A. Jansen, A.W. Denier van der Gon, G.M. Wijers, P.W. van den Hoogen,
J.A.M. de Laat, T.M. Maas, E.C.A. Dekkers, P. Brinkgreve, H.H. Brongersma:
Differentieel gepompte drukcel, ontwerp en bouw van een differentieel gepompte
drukcel voor oppervlakteonderzoek onder reactiecondities. NEVACBLAD, 38 (2000)
83-87.
 121

Dankwoord

Tot nu toe heb ik een zeer inspirerend en dankbaar werkgebied gekend. Zeer veel tech-
nische uitdagingen zijn mij geboden, zowel op de TU/e als daarbuiten. Als ik daarop
terugkijk, is het duidelijk dat er gedurende de jaren veel vertrouwen in mij gesteld is.
Zowel mijn werkomgeving, waaronder zeker ook het College van Bestuur, mijn vrienden
uit alle richtingen van technische wedstrijdsporten als vooral mijn familie hebben mij altijd
de noodzakelijke steun en vrijheid gegeven die voorwaarden zijn om deze uitdagingen aan
te kunnen gaan. Ik kan ze daarvoor niet genoeg bedanken.
Voor de vaktechnische, meestal prachtige en geraffineerde uitvoeringen, van mijn "pente-
keningen" kan ik slechts alle medewerkers van de GTD als team bedanken: hulde.
Het noemen van namen in een situatie als deze is gevaarlijk, maar zonder de bijzondere
inspiratie en inspanningen van Hidde Brongersma, Wim van der Hoek, Nick Rosielle,
Martien de Voigt en van de leden van de Promotie Commissie in het algemeen, was er van
dit boekje beduidend minder terechtgekomen.
122

Curriculum vitae

Peter Brinkgreve, geboren 13-2-1941 te 's-Gravenhage.

1959 Eindexamen HBS-b, Gemeentelijk Lyceum Eindhoven
1971 Afgestudeerd bij prof.ir. W. van der Hoek, vakgroep Constructies en
Mechanismen, faculteit Werktuigbouwkunde, Technische Universiteit
Eindhoven
1971 Werkzaam bij de vakgroep Constructies en Mechanismen, faculteit
Werktuigbouwkunde, Technische Universiteit Eindhoven.
Onderwerp: Ontwerp en realisatie van een testvoertuig met computer-
gestuurde wielophanging.
1972 Medewerker van de Gemeenschappelijke Technische Dienst van de
Technische Universiteit Eindhoven als constructeur, later in de functie van
hoofd stafgroep constructie en technologie.
1997 Hoofd Techniek en Innovatie van de Gemeenschappelijke Technische
Dienst
2001 Hoofd van de Gemeenschappelijke Technische Dienst