proceedings 2004 - RECENT TRENDS 2016



proceedings 2004 - RECENT TRENDS 2016
Recent Trends
in Charged Particle Optics and
Surface Physics Instrumentation
o f th e 9 th In te rn a tio n a l S e m in a r,
held in S k a lsk y d v u r n e a r B rn o , C z e c h R ep u b lic,
fro m Ju ly 12 to 16, 2 0 0 4 ,
o rg a n ise d by
th e In s titu te o f S c ie n tific In s tru m e n ts A S C R a n d
th e C z ec h o slo v a k M ic ro sco p y S o cie ty
held in honour o f Professor Arm in Delong
on the occasion o f his 80 th birthday
Edited by Ilona Miillerova
Published and distributed by the Institute of Scientific Instruments AS CR.
Technical editing by Marek Frank, printed by CCB Brno, Czech Republic.
Copyright © 2004 by the Institute of Scientific Instruments AS CR and individual contributors. All rights
reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any
form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the written
permission of the publisher, except as stated below. Single photocopies or single articles may be made for
private study or research. Illustrations and short extracts from the text of individual contributions may be
copied provided that the source is acknowledged, the permission of the authors is obtained and ISI AS CR
is notified.
The publishers have made every effort to reproduce the papers at their optimum contrast and with optimum
resolution from materials provided by the authors. The publishers accept no liability for the loss of image
quality resulting from supply of poor original material or improperly prepared digital records.
ISBN 80-239-3246-2
§ u o p Q u iu ijy jo s s a jo jj
Two years have elapsed, and once again we have the pleasure o f w elcom ing participants
to the R ecent Trends sem inar - the ninth in the series. From one point o f view this sem inar
is to differ from those that have gone before, since we have taken the liberty o f holding the
m eeting in honour o f Professor A rm in D elong on the occasion o f his 80th birthday.
Professor A rm in Delong was born on 29 January 1925 and began his career as one o f
a group o f three students who built the first C zech electron m icroscope in 1950. Soon after­
w ards he becam e a leading light in the electron m icroscopy com m unity in this country and
enjoyed w ide-ranging connections abroad. For nearly th irty years, betw een 1961 and 1990,
he was director o f our institute. D uring this tim e an unbelievably extensive range o f original
projects were undertaken in the fields of electron optics and m icroscopy w ith great success,
either w ith his personal participation or at least at his initiative. From my perspective as his
pupil (som ething o f w hich I w ill always be extrem ely proud) another aspect seem s equally,
if not more, im portant - under his leadership the Institute o f Scientific Instrum ents was for
decades a shining island in a surrounding sea o f repression o f the intellectual life o f a society
dispossessed of its political freedom .
Eight is surely the roundest of num bers, so an eightieth birthday may also seem the “round­
est”, and the one most worthy o f celebration. T his is surely particularly true of the birthday of
a man who rem ains creative to this day and shows no respect for his own oft-repeated tenet that
"innovations can only be produced by those still young enough not to have learned that what
they are aim ing for is im possible”. He learnt this long ago, but it doesn’t seem to bother him.
In the first session o f the sem inar, explicitly dedicated to Professor Delong, we w ill have the
pleasure of hearing about his recent developments and plans for the future. A num ber o f us will
also be adding com m ents and m em ories o f our work with him and our shared experiences.
In other respects we are, at first glance at least, m eeting up in an atm osphere not so very
different from that at our previous gatherings. C an this really be true, w hen ju st a short tim e
ago the C zech R epublic finally jo in ed the E uropean Union? Perhaps greater changes w ill be
seen in the everyday life of C zech scientists over the long term . So far all that has changed
is that new adm inistrative com plications have been added to the huge num ber that already
existed to burden the m anagers of scientific institutions. It seem s that, for scientists at least,
the crucial event was the fall o f the Iron C urtain, after w hich things began to develop sponta­
neously o f their ow n accord, as a result o f w hich we have already felt ourselves to be p art of
the E uropean scientific com m unity for years. H opefully the act of ctQssing the Czech border
this tim e was sim pler and a bit m ore pleasant for o ur foreign participants than it was before.
O ne other aspect should, however, be m entioned here. In the past, in a divided world,
we C zechs often profited from various form s of help from o ur colleagues in the European
Union, so we should ourselves now be ready to help those from less fortunate countries, or
at least from countries that are experiencing a fate sim ilar to our own, i.e. couritries in which
restrictions and a lack of freedom have been im ported and supported from the outside. H ope­
fully next tim e we w ill be able to offer som e financial support for the participation o f young
researchers from the third-w orld.
T his brochure is being distributed to sem inar participants on arrival, and this gives me
the opportunity to thank them for including this event in their diaries, already so full o f com ­
m itm ents and higher profile m eetings, and to w ish them a pleasant and fruitful week in the
H ighlands of B ohem ia and M oravia.
Luděk Frank
oi a n s
UUEUIJ31S3M r put: JafBips 0 ‘II0Ü 'il ‘zïl33 V
sarcmaNH iMvaa a a a v ra v A cinv N o u m o s a a
adODSOHDiiM n a o n v o n in n v d s w v s o n v n
h iim
î[TJB|0^ A PUB 3uOp(] V
uuBuisuunj Q puB J35JUIOJX g ‘asuaiiuoips 0 ‘pJBinuag d ‘oaipaiz D
■ysuio» j uojsjoj y ‘ipiuiipg S ‘J3<-laM N ‘P^IM IM‘-raipsg -jrç
Sdx ONioviAii d oa ir a n ia a o í h n s av3 n v :v d s 3 o n v n
sadODSOíOliM N o d io a a a a o 3 sa a o H iv D
noiss iw a
a a a ia
ÄDdaNa NOd.i.jaaa aaNNVHDixanw a m a o i w d a v o i i d o a h i
ojiqsBiun^ § puejinj^i y ‘u3ubj>i uba § ‘ouqrQ
NOdiDaaa s s in ih d im h
a o aDNViMdoaaad s h i no
HOiH v a o a a a ia ia
ä !)>j3 nh n v
um» d PUB U9J9UIOS lira g 'uo33ruy ut;a f j,aj
s a v îiw
SNaao>J3iiM aNV sasN aao> oii\i N oaiD aaa
s a u r a v a n o N is a o NOiiVNiiMiaa hiiay
A0>nupj3a S V
a o v a m s 3 h i h v sn
a a a ia w a a a£
íh a io s
siwia T
sw aiSA S avD U d o
N o d jja a a a o sa u n a d o a d a v ix v a v d a H i a iv a n o a v D o i a o o i v
[O.ipuü/W d puti jansips y
imss n i onioviaii
NOLLíTiosayoNVN >ioa m ív a a iiN D S ii- o v A a o a d A i
a sa
i:[opo¡s] \ pur; Ejminy y
iquidg f pin: i’|opD\; \
f ‘BjBJjny »
S 1 N 3 IN 0 3
R H am m ond, F. H. R ead, T. J. R eddish and R J. H arvey
R H orák and R Schauer
P. H rn čiřík and I. M üllerovà
J. Jin hua and A. K hursheed
J. Jirák. P. Č ernoch and R. A utrata
A. K hursheed
I. K onvalina and I. M ullerová
B. L encová
B. L encová, P. Jánský and J. Z lám al
G. L ilienkam p
F. M ika and L. F rank
M. O ral
A N A D D -O N S P E C T R O M E T E R /L E N S -S E C T O R P L A T E A TT A C H M E N T
M. O sterberg and A. K hursheed
a v iiN a a a a a ia
nt[Z X PUB ojunj^ 'g ‘ssnoy f ‘n;g g[ ‘Su^M 1
a o H ia w D iv a a a o a v
sw aiSA S a a o d in n w a o N O iiv a n w is
BjEjjny y pui; papíre^ y
a s a a o v n o A -M o a
S N oaiD aaa
iv m o is
a o iD a x a a
a o s a ia o iD a r v a i
>[3oçjoj-| j/\[ pire pAoouag g ‘>[30|A 1
w a s a o v n o A -M o a 3 hx a o N o isa a
ojiqseuin),] ç pire AO^Bpns ]/\[
SN ouvm m s
d v a i Noi n i s a o y y a c n a ia 0 ¿ -N 0 im s do s i s a i v n v 3 hjl
ejBJjny y fiur: jarreqos g
japrauqos ]/^ j pue Aos^ng y ‘sjauqg f h
‘oyfidajsj s Uauspo V ‘qssN a
y ‘asuaquoips ¡3
w aad-x a a A io sa a -a iM ii oN isn a a A a a sa o
s a s s a o o a a N o u v z u a N o v w a o sdiwvnaci aN ooasoD id
N o u v z M ii d o y o i D a i a a
N o u v n iiN iD S y o a s a o H ia w aA iivaniM is ö n v iv iN a iA ia a d x a
T!AO.i3[[!g\ 'i pin; >[in;jg g ‘ijBiuoQ-jg g\¡ ‘XíjsAoireuroy
adODSoaDHAi N o a iD a a a
o n in n v d s
. s is a iv n v y a o n v hoj
A oaaN a ayoi o u v i s o a i D a i a
SuuMog -f N pue peag
saoan os
noi ö n v
N o a iD a a a a o N o u v m w i s
S D iiv is o a iD a a a a o a w aa a m a o
h a
a iv r a n n 3H i
S3TJJ g pue J s q o n c y jaj
A oaaN a
w aiSA S iMvaa noi a a ssn D o a
a a iv D ia a a v a o aD N V w aoaaaa w aiSAS
R udolf Autrata a. Josef Jirák a b, Vilém Neděla a, Jiří Špinka a b
a Institute o f Scientific Instruments AS CR. Královopolská 147, 612 64 Brno, Czech Republic
h Departm ent o f Electrotechnology, Faculty o f Electrical Engineering and Communication,
BUT Údolní 53, 602 00 Brno, Czech Republic, fax. +420 541 146 147,
e-mail: [email protected]
W hen we are detecting signal electrons with an ionization detector in ESEM , the most
advantageous is to use, with respect to the effectiveness o f ionization, environment o f water
vapours in the specim en chamber. The m ost suitable are saturated w ater vapours (relative
humidity o f 100%). This environm ent is also advantageous for observation o f specimens
containing water. A t pumping o f the specimen chamber to the working pressure (typically
around 1000 Pa), in the case o f wet specimens, evaporation o f specimen can occur, or on the
contrary, condensation o f w ater on them. It means, that in some cases the specimen will not
remain in the original state. That is w hy we try to adjust and optimize the pumping procedure
and pum ping o f water vapors to minim ize deform ation o f the observed specimens.
For optim ization o f the pum ping procedure w e need to obtain information about amount
o f vapour in the vicinity o f the specimen. To measure humidity o f the environm ent the most
comm on m ethod used is the mechanism o f absorption o f water into polym er materials [1].
Impedance sensors, capacity sensors or sensors based on the measurem ent o f changes o f
mechanical properties are used, as well as other types.
Im pedance sensors use a polym er structure to w hich hydrophilic ions are bound, whose
charge is com pensated by opposite ions. W hen humidity is increased hydrophilic particles
absorb water. A typical representative o f such a material is N afion as a cation exchanger with
firmly bound sulpho groups, while hydrogen ions are the contraions. The impedance is then
measured at an alternating current. The changes o f the impedance o f the detector can be up
several orders. The dependence o f the impedance on the humidity o f the environment is
usually non-linear, and such sensors show certain hysteresis.
Capacity sensors use a hydrophobic polym er as dielectric (a material w ith small
perm ittivity is chosen) o f the capacitor. Even a small amount o f water absorbed on the surface
o f the detector, with regards to the high perm ittivity o f water (er = 80), significantly changes
the perm ittivity o f the configuration and, consequently, the capacity o f the sensor. This
dependence is linear. Polyam ides or derivates o f cellulose are used the most frequently.
The sensors based on changes o f mechanical properties '•contain a polymer, e.g.
polyvinylalcohol and finely dispersed carbon particles. Changes in volume o f the polym er due
to moistening or drying up, result in a change o f ohmic resistance between conductive grains
o f carbon (swelling-type sensors). This resistance is then measured. M echanical changes
caused by change o f hum idity are also used in piezo-resistor sensors.
For our purposes we used the capacity sensor [2] equipped with a sensor for temperature
measurement. All device is electronically controlled via a microprocessor. The output signal is
For optim ization o f the pum ping and filling-in processes w e utilized studies [3,4].
Tem perature o f w ater in the vapour developer, num ber o f repetitions o f cycles o f pumping and
filling-in water vapor, range o f pressures used for cyclical pumping and filling-in water vapors
as well as the level o f initial humidity in the specimen chamber, all have influence on the
pumping process. The lower is the tem perature o f the specimen, the better is the final result o f
the pum ping process.
However, it cannot be lower than the tem perature at w hich water
j o j | B ju 3 t u u o jiA u g s q j j o
' L Z Z ~ L Z Z 'd > 6 6 1
0 1 0 0 0 Z Z 9 Z lAISPM } h b j8 s q j Aq p s j j o d d n s sbm. >[jom s i q x
' 6 Z I - P c l ' á '£ 0 0 6 ‘ 6 ‘s is Á [ E i r e o i o ! j A ¡
JJ3A V '
J3pU tl
S T s X jB u y p u B 3 u u o jiu o ja j : ] j S -is q u a q o ij ' - f j u i u q x ‘ 'M " z u b j j
‘ Z L I Ï ° A X d o o s o jD ij/ \ [ j o
[t? ]
[ B i u n o f ‘ a d o D S O io ij A j u o j p g j g S u i u u b d s
[ f]
'0 0 0 t
? 5l3 !lU 3lt 3
‘jtr a r a X jo d
iiuXqnj s
á jo z u s s
(S) »
‘j B ^ s d o
000 l
♦ ♦
♦ ♦
'Z Ï Z - 8 Z Z d
‘Ç A O îjig iu ij
oos 5
♦ OS‘9
♦ ♦ ♦ 00‘9
00 z
*♦ ♦
0|oAo ßu;duind l'ßjd
• s S B U IB p p u s S s S u B q O J U B D IJ IU S IS Ä U B J U 3 A 3 jd O J j s p j o U I S U 3 U J I3 3 d S S U IU IB J U O O J3 JB A V
qjiM >]joa\ 3jqBU3 puB sjodBA J3JBA\ pajBjnjBS jo jusuiuojiAua üb 3JB3J3 p|noqs 3Uii§3J siqx
ju s p ijjn s 3A3iq3B oj sjEnbspB sq p p o q s qoiqM ‘ssuiij jn o j jsbsj jb £ puB z sdajs jB3ds>j >
■p3A3iqOB SI EJ 001 l
jo s m s s s jd jijun jnodBA jsjbm ui h ;j -sssbsjoui a in sssad s q x tóquiBqo sqj oj JsdopAgp
sq j uiojj jnodBA jsjbm ui ¡jij puB sajba uoijb S u ji sq j uado b <j 00/. J ° SuiASiqoB J3JJV '£
B J 0 0 ¿ O J S 3 S B 9 I3 3 P 3 jn S S 3 jd 3 q j U S q M ‘J U 3 U IO U I s q j O J
s p a a o o a d S u i d i u n d s q x '3 S B 3 J 3 s p o j s u i S a q s m s s a a d s q j j s j d o j p j s j e a v s q j j o
s q j J3 JJV
'p 3 A 3 iq o B
s n u iju o o []t m
s i ja q iu B q o
usm psds
p u B s jjb js J3 JB M j o
•(J3JBAV j o
jo j
¡ u i ç ’o j o
sqj ui
u o ijB J o d E A g
s m n jO A B p s s n s i
3 SB 3 j n o u i ) j s q u r e q o u s u i p s d s s q j j o
K jip iu in q
jb u tS u o
u o ijB J o d B A S
pA3 j sqj
(iju n
- S u i d u m d j j B j s u s q j ‘j a q m e q o a q j s s o j q
0 I '8 ‘Z » q u iB q a u s u ip s d s j o
3 iu n | O A b
[jBUis b j o s f u j g
:sdsjs 3 utmo [joj sqj jo sjsisuoo
3 u id u in d s jo j s q j s j b a v j o ju n o u iE
JI p sjssS S n s s q u b s sjn p so o jd S u idu in d u in u iijd o ub sS u ip u ij'p au oiju su i-aA oq B sq j u i o j j
3SB3 I3UUOJ sq j UT SB ]3A3] KjipiUItlq B q3HS SASiqOB jou píp 3A\ Q011 — 0 0 6 PUB Bd 0 0 1 1 “ 00 8
j o sS u b j aqj ui j o j ’b j 0 0 I I — 0 0 /. J ° sS u b j aqj ui u oijB (n ojp jb psAsiqoB si snjBA
jssq sq j jBqj ‘sjjn ssj jusu ism sE S iu sq j u io jj juspiAa si ji 'B j 0 0 6 ‘Bd 00 8 ‘Ed 0 0 ¿ J ° ssjn sssa d
[Bixiiuiui j o ssnjBA sq j oj S u id u in d u asoq o 3ABq say u osu B d u ioo j o j ' j 0q jb ajBjs p m b q ui jsjbm
3u|d335] jo j s in s s s j d jbu iiu iu i b si qoiqM ‘b j (,09 J ° snjBA sq j j s a o s jn s s s jd b jb S u idiu n d q siu ij
oj juBjjoduii si ji 'JsquiBqo u su iio sd s sq j ui Ajipiiunq Su iJinssj aqj u o so u s n y u i ju eoijiu S is b
SBq o s p ( i - 3 i j sq j 33s) 3 u ip X o usqM ssa n sssjd jb u iiu ioi puB ibuiixbui j o s3 u b j a q x
'D o Z Z -t3Ao
3jnjEi3dui3j B oj psjBaq si J3JBM jo J3UIEJU03 sqj jBqj 3jnsu3 oj jUBjJodiui si ji snqx 'b j 0 £ 6
J3AO ssjnsssjd jb )jíoa\ jo j ¡BoidXj si qoiqA\ ‘ssm sssjd JsqSiq SASiqoE oj sjqissod jou si ji puB
sjnooo Suqooo saisusjui uoijEJodBAS J3JEAV 3AISU3JUI j y 'sssoojd 3uiduind sqj jo jsjauiBJBd
juEjjodun jsqjouE si som os jodEA sqj ui j9jea\ jo 3injEJ3dui3j sq x 'soBjjns sqj uo ssjBSuspuoo
R. Autrata, V. N eděla
Institute o f Scientific Instrum ents A SCR, K rálovopolská 147 Bm o, Czech Republic,
vilem @ isibm
It is known, that study o f sam ples containing big am ount o f water, e.g. soft biological
tissues, can be realised by the technique o f environmental scanning electron m icroscopy. This
technique enables one to use high pressure o f w ater vapour in the specimen chamber (103 Pa)
and it thus creates conditions for m inim ization o f w ater evaporation from the sample, the
detection o f signal electrons w ith sufficient efficiency and neutralization o f a charge
originating on the surface o f insulation samples. Study o f biological specimens o f soft tissues
w ith fine and very fine hydrated structure still rem ains a big problem because this structure
can be easily dam aged by dehydration during the initiating phase o f vacuum pumping o f
a m icroscope. The attention is thus focused on setting up the conditions for complete
dehydration o f sam ples, ensuring the nature state o f a sam ple during the critical phase o f
pum ping o f a m icroscope. Such conditions can be created by respecting the physical
dependence o f pressure o f saturated w ater vapours on tem perature, w hich is depicted by
a curve presented in Fig. 1, and using special m ethods (water-containing gels, a chamber with
a mem brane, etc.), w hich m inim ize deviations o f wetness in the vicinity o f a specimen.
Violation o f the conditions presented in Fig. 1 leads to dehydration o f a sample or to coating
the surface topography o f a specim en by water. These states are presented in Fig. 2., w hich
presents a partially dried specimen o f small intestine tissue, on w hich forming w ater drops are
visible as a consequence o f a condensation process and a beginning o f coating the specimen
w ith water.
A m ethod o f cyclical irrigation o f the specimen chamber is one o f the methods which
preserve prolongation o f a relatively natural state o f a soft tissue sample. Experim entally set
up amount o f distilled w ater (approx. 0.5 ml for cham ber o f AQU ASEM m icroscope) is
inserted into the specim en cham ber before the beginning o f the pum ping process. The
pum ping o f the specim en chamber goes through the pressure limiting aperture, which
separates the cham ber o f differential pum ping from the specimen chamber. This pumping is
accom panied by a decrease o f the pressure down to the value o f a pressure o f saturated water
vapours (2062 Pa) corresponding with a tem perature o f distilled w ater in the specimen
cham ber (18°C). A nother decrease o f the pressure in the speciftren cham ber occurs after
evaporation o f distilled water. In that mom ent filling-in o f water vapour m ust start from the
external reservoir o f water, regulated by a valve. In order to achieve stabile high relative
w etness in the specim en chamber, it is necessary to repeat the irrigation process several times.
Another m ethod is based on observation o f w et tissues placed on a special support from
agar, w hose area is a sufficient supplier o f w ater, w hich gradually evaporates and thus
hydrates the specimen. [1]. The agar successfully m inim izes the influence o f dehydration
deviations on the observed specimen and, in connection with a m ethod o f cyclical irrigation o f
specim en cham ber, creates a very good m eans o f observation o f biological tissues in ESEM.
It is dem onstrated in Fig. 3, in w hich sam ples o f small intestine tissue observed at equal
tem peratures and pressures in the ESEM cham ber are presented. The upper parts o f figure
Fig. 3 (a, b, c) presents sam ples placed on a classical cooled specimen holder and the samples
placed on a special cooled agar base are shown in the bottom parts o f Fig. 3 (d. e, f). It is
'o iiq n d s y q o a z j aqj j o
s s o ir e p g j o
'¿01Z.90Č ÍIS^
P ub S 0IS 90Č S
se m :s[jom s iq x
'Í6V'V6V ‘sSuipssDOjj
‘£ 0 0 č J sq tu sjd ss ‘Xu b u u s q ‘u sp s a a a ‘£ 0 0 č s s u s jsju o q XdoDSOJOtj^ ‘(]A jaS3)
píděmi ju s z s jd (aoujBjAj) jtoajsss ^j jsib ^
: jj ‘S isq u a q o jj “ j^ ‘J3})ry ‘ J ‘js S ubç
A'0 0 8 'S v jq 'W 3 S 3 u! (s a u n S if u io u o q ) jo S d uíojJ a svq j v p a d s v
u o p u o (s a u n S if ja d d n ) uap¡oi[ u siu to a d s ¡ v o is s v p v u o p a D n jd a a iisa tu i ¡ p u is J o 3¡duiv¡¡ f a m S i j
U|LU g ¿ / Ed 0 0 8 / 0 o 2 (i
u |w 0 9 / e<d 0001- / Oo£ (a
¡ ' S i j m d n ¡as s u o itip u o o
J o aouanbasuoj o
U!W S ť / Bd 0SČI- / OoS (P
3A tllD J 3 d l U 3l UO S J t l o d v A J 3) V M
u o iin ¡ o iA
p a jn jn /n s J o a m s s d u d J o aouapuadap Jo a ¡ u n j j s u n S t j
s v s d o jp A3WM 'Z 3 -in S tj
U|W 06 / Bd 0 0 8 / OoO
uoijB ipjíqap Xq pasriBO s jn p r u is sdbjjtis j o siBUiBp
u m u n u tu i (isp [o q [boissb[o b u o paoBjd ssjdu res j o aseo sq j u¡ sb a m ss sjd puB sjniBJsdura}
j o ssrqBA su res sq j ib) M oqs u o d d n s jbSb sq j u o psoB |d s3|dures jBqj £ 'Sxj u io jj ajqB soqou
P 0 0 Z SP U!>-<1 ÍU 909H
R. Autrata. P. Schauer, P. W androl
Institute o f Scientific Instrum ents, AS CR. Brno. Czech Republic, wandrol(5>
D etection o f backscattered electrons (BSE) in scanning electron m icroscopy (SEM)
serves as an auxiliary m ethod in the study o f surfaces and com position o f materials.
Backscattered electrons have properties that are different from those o f usually used
secondary electrons. The achievem ent o f the theoretical lim it o f resolution (0.6 - 0.8 nm for
SE and 0.9 nm for BSE) depends not only on the properties o f the electron source, properties
o f electron optics, specim en preparation technique, type o f electrons, but also on the detection
system efficiency.
Several types o f detection systems have been designed for detecting signal electrons in
SEM. Even though in the last few years the noise and time characteristics o f sem iconductor
detectors and the properties o f channel plate detectors have been considerably im proved and
even though a great num ber o f SEM s are equipped with sem iconductor detectors to detect
BSEs. the scintillator-photom ultiplier still possesses the best signal-to-noise ratio and
bandw idth characteristics [1] The m ost im portant part o f such system is the scintillator.
Powder P47 and yttrium alum inium garnet (YAG) single crystal are the most efficient
scintillators even if pow der material o f P47 is suitable only for SE detectors.
It w as shown [2] that BSE detectors based on YAG single crystal enable to achieve
very high detection quantum efficiency (app 0.75 at 10 keV). The efficiency o f the YAG BSE detector depends not only on the optical properties o f the w hole detector system but also
(and especially) on the efficiency o f electron - photon transfer in a scintillator. This efficiency
is the lim iting factor for the achievem ent o f a high BSE resolution.
W orking out technology o f the preparation o f the YAG scintillator with the highest
light output after an im pact o f BSEs w as our aim.
For crystal grow th a small volum e o f w ater vapour in the growth atm osphere was
used. Due to the additional treatm ent o f the YAG discs in oxygen and hydrogen atm osphere at
very high tem peratures, the colour centres in the YAG crystal lattice were suppressed.
Polishing process o f the Y AG surface w as im proved. It w as found that this process caused
penetration o f the polishing m icroparticles into the surfaced microcracs. These particles can
be rem oved by the w ashing process in a special m ixture o f acids only at a suitable
Thanks to these steps, the light output from the m o d ifiei'scin tilla to r w as increased
about double times, in com parison with the older YAG (type I). The decay tim e 1/e was
decreased and m oves in the range 60-70 ns. Long tim e afterglow is also shorter.
The m odified single crystal, nam ed as Y AG II scintillator, was used to design the
planar Y AG -BSE detector that is tightly placed under the pole piece o f the SEM .A smaller
hole in the centre o f the YAG has been used for the reason o f higher collection o f BSEs
trajectories. DQE coefficient o f this detector w as increased from the original value 0.75 at
lOkeV to 0.85 at lOkeV.
Thanks to the im proved properties o f the Y AG -BSE detector, it was possible to
operate w ith low er PM T amplification, lower beam current and low er accelerating voltage.
All these conditions are suitable for the receiving o f higher BSE im age resolution.
R. Autrata, R. Hermann, M. M uller, A n efficient single crystal detector in SEM,
Scanning 14 (1992), 127-135
s u o jp s p
ju sp p u i jo ÁŽJ3U9 snsjsA sjejsájo
sjSuis jo jndjno ;q§i| dA pepu \ a jn S y
s jn S ij
•uoqjB3 uo ss p iy ß d pjoQ •£ s jn S tj
• sdB i d i j s u S b u i p a jB O O p | 0 { ) - p g j n S i j
■jopopp a s a - o v A
J B U B ] d q i l M p s p j o o s j S y x| Y x_l E Q p U E
SVB0 i ° 3-inprujs 33i}jE|J3dn§
(A9^) Aßj 0ua u ojp aia
> .
» SI 0
o °ro
js ro
(1 U U )
ON JuejS ‘3 |[q n d 3 y i p a z 3 aqj j o s s o u a p s j o X aisp B oy Xq p s y o d d n s
>[JOm siqj^
P 0 0 Z Spil3UX IU¿>39}¡
[3 ]
8£fr-/.£i7 ( 8 6 6 1 ) unDUE3 -jAjg s s s j g u o j [buoijbum jui
,U17 [ -OOJJ ‘IM 3S A 1 JOJ SJOJ33J3P jo iB ijp u p s (bjsXjo 3 I§ u is 'J sn B ip s d ‘B lB fln y -y
Jan Bärtle
Institute o f A pplied Physics, University o f Tübingen
A uf der M orgenstelle 10. D-72076 Tübingen, Germany; eMail: [email protected]
In the first design steps towards a novel electron optical system or component, initial paraxial
calculations are usually performed under assumptions, e.g. thin lens, deflectors with sharp cut
o ff fringing fields (SCOFF approximation). If the requested paraxial properties (such as
stigmatic imaging, m agnification, dispersion etc.) are obtained, the system or component may
be improved further on using more elaborate techniques [1-9].
W hen dealing with electron monochrom ators or energy filters, not only the four geometric
fundam ental rays xa, yp, xY, and ya are required, but also the two dispersive fundamental rays
xK and yK are o f interest. These two rays emerge in dispersive electron optical components
such as electrostatic toroidal condensers, magnetic sector fields or W ien filters. However,
these com ponents usually introduce second order aberrations not described in the paraxial
domain. The elim ination o f at least the m ost disturbing one o f these aberrations is mandatory
for the overall system performance. This influences the fundamental rays: to cancel some o f
the aberrations special symmetry conditions concerning the course o f the fundamental rays
have to be m et - despite o f using stigm ator elem ents which increase the system complexity.
To obtain a system fulfilling all desired requirem ents, especially the symmetry o f the
fundamental rays, one has to calculate the properties o f the electron optical system over a
large param eter range.
The program Paraxial was developed to provide such calculations.
Program principle
It is advantageous to describe the paraxial properties o f the electron opticakcom ponents o f
interest in term s o f transfer m atrices allowing rapid calculations o f the fundamental rays [10,
11]. The com ponents have certain user defined properties e.g. their lengths or their
excitations. These properties affect the matrix elements o f the transfer matrices and hence
influence the fundamental rays.
Besides the properties o f the components, the geometric fundamental rays or their first
derivative may be subject to conditions the user is supposed to enter. Typical conditions are
the generation o f interm ediate stigmatic or astigm atic images. A simplex algorithm [12] is
em ployed to vary the system under consideration to fulfil these conditions. A suitable system
can be found semi-automatically.
Program features
Paraxial handles the dispersive electron optical com ponents m entioned above, additional
quadrupole lenses, and thin (rotation free) round lenses as well. The operation o f Paraxial and
the input o f the data is performed interactively using a m odem graphical user interface.
Sim ultaneous calculation o f the rays is performed as well as the graphical display o f the
results, so the user may directly control the consequences o f his input. It is o f interest to know
exactly the position w here the rays intersect the optical axis (the ray value is zero in this case).
Paraxial detects these positions, not only zero but every user defined ray value.
The user can influence the program execution via options. Furthermore, dedicated output
options in m odem file formats are available to control the calculated fundamental rays and
[ £ j]
'( 0 Z.61 ) u o p u o g 's s s j j o iu io p E o y 'u S is s p s u s ¡ u o u p a p u i s a p d iu p o n l5 :so)[a\b¡-[ j
[^ []
'iq d p p /u io o 'p u B p o q 'M M M : o u ¡ p u B jjo g
[ j]
(9 6 6 1 ) U P O *x3o]o u q o o x j o X jisjOAraft ‘s is o q j ( j qd
:iu b j§ jo p u b a '[\!
S H ‘( ť 6 6 l ) I I ° A ‘JAI3 3 I ql£ l a o J d :ui ‘BAOOuog g
’ Z ‘( £ 6 6 1 ) frlO c Ï ° A ‘3 Id S j o ' ^ d :u¡ ‘o j u n ^ -g
'( 6 6 6 l) JPBJSUUEQ JO
X jis jo a iu q ‘s is o q j 'Q 'q d ‘ u a ip n b u a u o a iy a iq u n f su oiD W O A ip ou oj/y s a u p u S is 3( j qqE '» g
0 6 '( 9 6 6 1 ) 8 S 8 c 'Io A ‘3 Id S j o a o j j :ui ‘» n r » c] ‘q jjB g 3 T ‘uieJS Jop u b a jai
-osoDJijM u o jjo o jg - jd o 'A p y ‘s o q d g
3 s o j 3 ¡^\[ u o jjo o jg j d o A p y ‘osno>[ f
I ‘( ¿ 6 6 1) u ojb^ j B o o g ‘s s o jd 3 ^ 3 ‘s o ijd o o p i j j c d poSjB qo j o >(ooqpuBi[ :ui ‘ojunp\¡ 'g
[1 1 ]
[0 1 ]
•( 6 8 6 l ) SS3 J J X jisjo A iu ß o8 p u q u iE 3 ‘s a d p a y ¡D D u a iu n ^ : j b -jo s s o jd
•(¿ 6 6 1 ) U3 3 u ;q t)x j o X jisjOAiufj o q j jb p jo q o r n p o g 7 / y u d o u a q j j p j :s o q ([ -g
•(¿ 8 6 1) u o p u o g " s s o jj o iu io p B o y ‘s a p ij u v d p t â j t n p J o s o i)d Q :>[[U[[0 /w •]-[
i l l ‘( ť 6 6 l ) H
’I ‘( 17661) H
apijuoj paS jn yj Jo uSisap-sud aift uof souDisissy uaindiuoj
anz zSüAjpg
'( L í b l ) S i z d p g 'q y e g s n i s o j q u i y u u e q o f § e jjo ¿ \ ‘y iid o u a iio u ji/ a ig
lo q o n jg p u n q o s n g ui ‘yijdouauoutygig 9i/3Sij3M3tft atp Mqf¡ :jo z jo q o § O
S33U 3JSJ3>J
u iB jS oJd B|Bp ¡b u iS ijo o q j qjiAV pojB d u io o OJB s jjn s o j oqi puB p o jB jn o p o uooq OABq suiojsX s
[Bopdo u o jj o o p p o q s q q n d XpEOJ|B o u io s '¡DixnuDj qjiA\ pouiB jqo s q n s o j o qj ojB Jjsuouiop o x
S||I1S.1.I UHUrj(U,|
o u iij jjo q s B u iq jiM u iB jS o jd o qj osn o j o |q e oq p jn o q s oSpojAvoiDj [Boijdo u o j p o p qjiM jo sn
SM opuiM jB [nS oj y -uiojsXs SuijB Jodo S M opu[,\\ jjosojoifM o qj o iu i ¡ m x i u n j j o u o iib jS o iu i
ss3 [u ib 3 s gpiA oad o j p s s n sb a\ ([t? l] ‘ iq d jS Q ^ p u B jjo g ) jo o j ju s u id o p A s p s ib m ^ o s u js p o r n y
u u ! ) e ) u 3 u i 3 |d a i ! u i i u n o . i , |
sjs p u iB J B d jjB js oiBnbopBui S u isn uoq,-A ÁjiBioodso ‘s u o p n jo s p r u j o o o aiS o j [ibj Xbui
u iq ju o S jB x o [d a iis a q j ‘su o ijB n b s ,\b j [eiXB.icd s q j j o s u o q iq o s o ip o u o d a q i jo o jn iB U a q i 01 o iiq
•(su3j p u n o j u iq j b puB su a] a[o d m p B n b b S u isn p i x D M j u i p sjB jau aS s q ubo jn q ) 3[qE[iBAB
SI JU3U0dlU03 SU3| JSpUlJ/ÍO OU pUE SU O ipSygp 3|SuE JJEUIS SJEjnOIBD O) u o jjd o OU SEq JDIXDMJ
i 0 i& - = '•fh PUB SK)>1 =
PUB OKj) = U«t>)
ÁJl.iLuuiÁs a jis o d d o |ij [ n j 01 SABq puB p p i j o u p n p o q ; o j p o d s o j qjiM six b [Bopdo o qi jnoqE
o0 6 Áq Á[|BqinuiiZB pgjB joj o je sp[0[[ o ip u S e u i o q i SEOJOqw Ä jp u iu iÄ s p p o j o u o ao jo q jio jijjn j
o j OABq s jo jjij uoiyy^ o q j j o s p jo y o u p o p o q x
[ £ i ] u i u 0 i¡v}U3i 40 pjopuois o qj ‘( s z ^ i = 2,j,)
.C ipujujX s p p o [ ijjn j o j OABq so u o oipuS B U i ‘(oscj) - z§) X jp u iu iX s u o a o |i j |n j o j OABq sosuo]
o[odnapB nb o ijE js o jp o jo ‘u o is jo j SuiqsiuB A OABq js n u i six b (Bopdo poA jno b qjiAV suiojsX g
s u o i ) e ) i u i ! | u i k j S (u , |
■sonjEA [B u u jd o S u iq o B O j ¡ iju n s p u o o o s m o j o u io s u i s u o i jb ju S iju o o u io jsX s
lu o jo jjip
OOOl’OOS jn o q B X jp o id X j S0jB [n 0 |B 0 u iq ) u o S |B x o |d u iis o q i ‘u i o jq o jd onjBA |b ijiu |
o q j j o j ( s ju o u io jo x iJ iB u i = ) s jo p u iE J B d jjB js j o jo s o iq B jin s B qjjAV '( s s j n o o j o ‘s ju o u o d u io o
[B o ijd o u o j j o o p j o
ju n o u iB
oqj uo
S u ip u o d o p ) S 3 d
u jo p o u i u o
u o ijE [n 0 [B0 u io js X s o jo [d u jo o b s u u o j j o d j v i x d m j u o jjo q puB
spuooos j o
s u o ijo b jj u i
0|.Ol J ° ■"’^ u b j o q j u i s i XoBjnooB
o a ijb [ o j p o u iE jq o o q x 'o u i ij j j o q s b u i s o S u e j jo jo u r e jB d jo 3 jb [ js o j o j u iq ju o S jB x e jd u iis
o q j S0|qBU 0 s i q x
u o ijE in o jE O X e j o q j j o
X o u o io g jo o q j u o ju o d s u o o q OABq s j j o j j o [B lo o d s
■ s u o ijb 3 ijs o a u i j o q j j n j o j p jB 3 o j q jiA \ josB jB p 0 [ q B ip j b S u ip iA O jd s n q j OAijEAUop j s j i j j i o q j
A. S. Berdnikov
Institute o f Analytical Instrum entation, Rizhskij pr. 26, 198103 St.Petersburg, Russia. Email:
[email protected]
The report describes new BEM algorithm for 3D Laplace equation. The main goal o f our
investigation is to create the software w hich enables to calculate electric field with high
accuracy up to the boundaries o f the electrodes. BEM is unbeatable as soon as we consider
the accuracy as the main point. But to get the highest possible accuracy, especially combined
with the requirem ent to calculate the field accurately up to the boundaries, a lot o f
m athematical and algorithmical w ork should be done.
The list o f possible sources o f BEM errors include:
• approximate representation o f true boundaries by the discretized surfaces,
• approximate representation o f continuum set o f charge densities by the discrete subset
o f charge density functions,
• deviations o f num erically calculated integrals from analytically strict expressions,
• increased inaccuracies o f numerical integration procedure near the boundaries and at
the boundaries due to singular behavior o f BEM kernel functions at these points,
• approxim ate representation o f continuum boundary condition by discrete subset o f
collocation points or by some other fitting criterion,
• violation o f true boundary condition if collocation or fitting points are shifted with
respect to their nominal positions after approximate representation o f the boundaries,
• inaccuracy o f charge density values calculated by 3D solver which accumulate all
these approxim ations and inaccuracies,
• increased errors in charge density values near the edges o f the boundaries due to
singular behavior o f true charge density at these points,
• inconsistency o f BEM representation with particular problem if its mathematical
peculiarities are skipped by the program m er and selected BEM model is too narrow
(for example, it is impossible to represent by charges the potential which is constant
W hile some o f errors are inevitable, others may be eliminated by intelligent BEM model. Let
us consider it can be done.
To m inim ize the inaccuracy the boundaries are considered “as they are’-, assumi
that they are represented by general param etric expressions like
x = X ( p ,q )
n ( p , q ) = - y = Y (p ,q )
z = Z ( p ,q )
- for each param etric point ( p, q) from 2D rectangle p h < p < p h, qh < q < q h we can
calculate the surface point (X , F, Z) (and tangential and normal vectors o f the surface as
well, if necessary).
'u o ijeS ijsoau i siip S u u n p su o issn o sip juEjjodiui p u e [njosn Xjoa jo j AO>[Bpns p E q oij^ j q
puB 'y o d d n s oqj j o j
p jq ‘(o d o jn g ) XjojEJoqBg qojE osoy n zp B iu iq s ^uBqj oj o>[i[ pjnoA\ \
S J H 3 U ia 8 p 3 | M O U } | 3 y
•sooE jjns X jB punoq j o odBqs XjBjjiqjE js o iu je qjiM s p p i j o ijB jso jjo o p
Q £ j o suoijB]nD|BD u o i s p o j d q S iq j o j [ooj S u is iu io jd X joa e oq o j su io o s lu q ju o S je m o u oqj
‘s o ;jjo d o jd sjí j o u o ije S ijs o a u i s n o jo S u o jo iu s o jm b o j jo a [o s o qj p u e joX p o q s iu ij jo u X p jiu ijo p
si >(jom siq j q S n o q j iv 'o o B jjn s o qj iu o j j j e j si ju io d o q o jd o qj uoqA\ ( ^ . 0 1 - ) JOmo[ q o n u i si
i oxid s iq j jB iju ojod oSjBqo ju io d j o o qj s o ijjo d o jd o iu o u ire q o j o n Q s_ o i~ j o se m sosbo jso j j o j
(soS jE qo o o u o jo jo j qjiM s ju io d o d e jju s S u u o q q g p u uooM joq oouBjsip 3 qj UBqj jo |[ e u is q o n u i
si ooB jjns o qj puB ju io d o q o jd o qj uooM joq oouB jsip o qj uoqM ‘-o-i) ooB jjns X jB punoq o q j jb o u
[B ijuojod j o Xobjiidob S u ij|n s o j o q x uo ijB in 0[B0 p i ju o j o d puB X jisu sp oSjBqo j o j X jqiqBjs puB
( O O U O S jO A U O O j o p j o
X o E J t lO O E p 0 0 S p O J E J J S U O U lO p S3SB3 JS 3 J [ E O Ij X jB U E q jlA V S J U O U I I J O d x g
• s u o ijo u n j [ B u o i j i p p B o p n p u i o j p o p u o j x o o q p [ n o q s s s i j u B p S u i s j o X j B j q q o q j q S n o q j p
jb jiu iis
B u i p a jB jn o jB D
X o B Jn o o B u i o q j q jiM
oqj jo
X | [ B 3 ¡ jX ] B U B p o jB jn o jB O
" 3 'j -
o je
' '¡j
' ’f¡
o q UBD u o u d
s ju o u o d iu o o p p i j
D ju o u o d u io o
’s j o j j o
[E U O ijip p E o u q jiM
o q j o jiq M
X i| E o u o i u n u p s jB in o jB C ) s i p u B j E j n S o j s i ß
‘s p u
u o ijb jS o ju i
ju o u o d u io o j s j i j o q x
i ^ + /V + zXvhP,
+ b p d p ( z ‘ , l ' x ' b ' d ) '‘ó Y j j [ - — ..... ( h ‘d )
■h p d p ( z ' x ' x ' b ' d y è ^
: ( z ', \ ' x ) f l j o j u o is s o jd x o p o y ip o u i ui y
jo d o jd Xq p i u o ^ o qj iu o jj so iju E jn S u is o q j joB jjqns ubo 3 ^
s j u o p ijjo o o j o u o ijo o p s
-joiA Bqsq jB jnS uis p s j i s s p b
( " b ' " d ) ju io d a q o jd s q j jb 3 u 3ABq puB s d e jjo s s q j j s a o X ubsijX ibub p sjB jS sju i s q ubo X sqj jBqj
( : M ' ' x ' h ' d y é s u o is s o jd x a ¡boijX jbub q o n s 3ABq s a \ js q j a s o d d n § -pajBSJO si saiiuDjnSuisfo
lÜDuqil [[bo 3A\ qoiqM qoBOjddB jB p s d s s u u s j jB jn S s jji puB JB jnSuis s s s q j s jB u iu iq s o ^
Z3 + Zb + Z d t ’
z3 +
- j + z d h
z3 + ; b + z d h
z3 +
z3 + z h + zd ^
zb + zd A
:p s jB u iiu ip s q p jn o q s pire * o < - 3 jb u o ijb jS s ju i |B 0U 3iunu S u u n p X jjB jnS sjji
aABqaq o j s n o p i d s n s s jb s u u a j S u im o jjo j s q x
ju io d a q o jd o j j o q q S p u js s jb s u
iu o jj suoijB iA op p o z ip iiu o u ojb b puB d ojoqA\
a q j si qoiqM ‘ju io d ooB jjns o q j -
_ 3 + b - d - 3 - z + - b + _ d l^
-- -------- b p d p ( b ‘d ) 0
a ( b ‘d ) z
3>(i| soABqoq [3UJ3>[ [bj S ojui j o jojB uiuiouop ‘ooBjjns oqj JB3U si juiod oqojd
oqj uoq/w -ooBjjns oqj jbou si ju iod oq ojd oqj uoqA\ piuo>[ jbjS ojui oqj j o soijuB inSuis p u s
saijuB [nSojji j o u o ijB u iu iip oqj s i uiqjuoS|B
a\ ou jo juouiOAOjdiui u ibui o q x •£
( 'ojnpooojd uopB jgojui jBououinu
oqj se ‘XoBjnooE puB jo p jo oou oS joau oo su ib s oqj j o sXbmje oq oj ouioqos uoijBuiixojddB
oqj so>[Eiu ji ‘jo a o o jo j^ ) 'sjqSpA \ puB sop ou u oijejS oju i op isu i uoppiq ojb uojjEuiixojddB
|BiuiouX[od jo p jo q S iq j o spBjop oqj jBqj o s ‘^ su M om p n ,, oqj sb uo>[bj oje sop ou uoijbjSojui
jb son [ba ogjBqo oqj ‘ou ijn oj u oijbjSojui |B ouoiunu oqj qjiA\ juojsisu oo oq oj uoijB uiixojddB
oqj o>jeiu
u oijbjSojui j o
ojBjnooB ojoui
ooiA\j uoijbjSojui oqj o>)buj oj so[qBuo qoiqM posn oje ssun/njpub u n issrw f) jo jjo u oijbjSojui
oqj oziixiiuiiu o x oiqEjiAou; si uopBjSojui [Bouoiunu puB ‘Xi[boijX[bub s[ejS oju i ]/\[g g
ojb|iio|bo oj X jq iq issod ou s i ojoqj ‘o jq issod se [ejouoS se ojb soiJEpunoq jn o o o u is 'Z
M .J. van Bruggen, B. van Someren, P. Kruit
Delft U niversity o f Technology, Faculty o f Applied Sciences. Particle Optics Group
Lorentzweg 1, 2628 CJ Delft, The Netherlands
Present day M icro Electrical M echanical Systems (M EM S-) Technology allows the
fabrication o f m icrosources, -lenses and -m ultipoles, with dim ensions in the micrometer
range. M iniaturization o f these electron optical elem ents is favorable, because the spherical
and chromatic aberration coefficients scale correspondingly, leading to a superior
performance over conventional electron optical system s[l]. In addition, miniaturized columns
have a strongly reduced length and thus are less prone to external stray fields. Those m icro­
elem ents are perfectly suited to be fabricated in an array, achieving throughput enhancement
in lithography applications.
There are several groups in the world working on the developm ent o f microsources and
m icrocolumns[2-4]. Using such colum ns in an arrayed form is also under investigation for
application in lithography[5] and for microscopy purposes[6].
Electron optics
Scaling laws can be derived which show that the electron trajectory scales with the lens
dim ensions if all relative potentials stay the same[7]. If we want to keep the same electron
energy, it m eans that all lens potentials also stay the same. In that case, the electric field will
scale inversely proportional with the lens dimensions. This constant potential scaling mode is
favorable, since both the coefficients and aberration disks o f the spherical and chromatic
aberration scale by the sam e factor.
Practical challenges
W hile the constant potential scaling mode gives the best results in terms o f spherical and
chrom atic aberration improvement, the electric field between neighboring'electrodes will
scale up. resulting in a situation in which that field might become larger than the breakdown
field. Chang and coworkers used a selective scaling approach to circumvent this problem[8].
They mention a guideline o f 104 V/mm for the maximum field between electrodes. However,
this breakdown field scales better than linear w ith the inter-electrode spacing[9].
To prevent breakdown along the surface o f the insulating spacers between the electrodes, they
preferably have to be m ounted at the peripherals o f the lens assembly, where they can be
scaled up. Not only will this be more and more difficult at lens downscaling, also the effect o f
electrode bending under the relatively high fields will become more pronounced: there is a
constant trade o ff betw een prevention o f electrical breakdown and this bending.
In the fabrication o f m icrolenses, severe requirem ents are posed on the roundness and
roughness o f apertures and the alignment between them. Using an electron beam for
patterning, in com bination w ith self-aligned Reactive Ion Etching (RIE), some claim that a
roundness and alignment with an accuracy o f about 1 nm is possible, while edge roughness
could be in the order o f 10 n m [l]. Taking into account tolerances o f 0.1-1 %, this limits the
downscaling o f lens apertures to the m icrom eter regime.
There are more problems that arise when operating m icrom eter-sized lenses, such as heating,
charging o f the insulating material and contam ination o f the electrodes by the electron beam.
The influence o f these effects on the perform ance o f the system depends very much on the
specific configuration at hand and research will be necessary to say something about it.
8691 's s a /^ O N ‘(9)8 a '[ o u q s a i - p s -3e A j- îXejnpv -j i í u ja ^ -j -q î j h i ‘S u B q j
8861 <:sP 0 A M3N ‘SS3JJ u in u sid ‘soijdo uoi pire u o j p a j g î-jrç ‘i XSb jiz s
(8661) Ç\?L ‘Jdy/repM ‘(c)9 l 0 ïo u q o sx P S 's b a -f tuBqsjEpns S 'I ; X ‘m
(fr6 6 l)S 8 ie
■03Q/AON ‘( 9 ) j i g 'l o i n p a x
p s -3 e a T îs a w jq
a 'V
- ( T V ‘u R i i u s u p j
i'vLZ '33Q/A0N '(9)01 a lo u q o si -p § '^ A T íXEjnw < d i iu js ^ j - q î 'd 'H I ‘Sireq3
(9 6 6 1 )808£
•D 3Q /A 0N ‘(9)1/1 a ’l o u q o s i - p s 'OBA ' f î u B u i i s u p j Q ' v ■ u y jn 'a 'IATIAI - V G
(Z.66I) 66^1 'unf/ÂBjAj ‘(£)SI V ’lo u q D S i
'P S 3BA ï
'A ^ d 'AVS 'u n q o ^ î3 u b ^ ■§ P s q \ :ioq3 f - H i'A 'T ‘> R d
(6861) IS81 ‘330/ aon
• ( 9 ) l g l o u q o a i ps ' s e a T -II 3 P ï3 d S T f
> r a î S u m o J H I - V W ‘P-i°33]Ai
d 'Q - d ' H I ‘S u b i q
SS8 I L: » a / A°N ‘<9)L
'lo uq ssj,
-o b a
f ^
^ -y'W
[9 ]
[3 ]
■suresqqns j o XjiiBjrqd aqj sn s o j
puB dn ]i{ds oj sa s u a j- puE ssiniradE -oraïui
su©i ©A!P0 [qo
j o uraisXs E qjiA\ ‘jdaouoa XqdsjSoqiqouEU
uresq u c u p 3 p -ii[n u i p sso d o jd sq j
j o AV3IAJ3AO [BotiButsqos T s - in S y
sojnos A^uoqos
■ÀqdBjSoqjq utu 0 1-qns u u o jjsd
oj 9[qE aq oj juea\ s a \ l( i 'S ij sa s ) sd o o so jo iiu u o j p a p piBpirejs e ut ajnporn siq j SuiibjocLiodui
X g -s3 su 3 |0 j3 ia i Xq p ssn o o j XjjEnpiAipui ‘suresq q n s 3]dij[nui ojui s s j n o s /ty u o q o s b j o
ureaq pEOjq sq j dn siq d s jEqj 3|n p o u i s o m o s uiEsqijjnm e S u id o p A sp sjb 3a\ ‘iu su io u i siq j j y
D. Cubric“ *, S. van K ra n e n \ P. K ruitband S. Kum ashiroa
a Shim adzu research Laboratory (Europe), W harfside, Trafford W harf Road, M anchester
M17 1GP, UK
b Delft U niversity o f Technology, Faculty o f A pplied Sciences, Lorentzweg 1, 2628 CJ
Delft, The Netherlands
* E-mail: [email protected]
A high brightness electron source energy filter has been built and tested. The filter is o f the
W ien fringe filter type described previously [1], The optics o f the filter is designed to
particularly suit the Schottky electron source. A particular new feature incorporated into the
filter is an electrically isolated exit slit. Together w ith the independent control o f several
electric m ultipole fields at a single octupole unit, this enables detailed inspection o f the beam
profile and beam stabilization w ithin the filter that in turn makes the filter usable in
com m ercial instrum ents.
1. Introduction
A motive for this work is to use the filter for high-resolution low voltage scanning electron
microscopy. The basis o f this application is in decreasing effects o f chrom atic aberration and
consequently in im provem ents o f the spatial resolution [2], A high resolution LVSEM
requires both high source beam brightness and low electron energy spread. Our filter limits
the brightness loss by m atching the dispersion to the im age size o f the source and by reducing
the Boersch effect through reducing the beam current entering the filter. The source image at
the exit slit is very small hence only little dispersion is needed and a short W ien filter
operating in low -excited mode is used.
2. Experim ental results and discussion
The electron optical transfer part o f the energy filter is built as a small one-lens SEM and it
can im age an exit slit (nanoslit) and surrounding parts by m easuring the electric current
received by the slit at each beam position. Figure 1 shows the sample current images
obtained around the nanoslit holder tube. The sample current increases from white to black.
Figure la shows various nanoslit structures that could be used for tuning and beam
monochrom atization. Figure lb shows the top edge o f the nanoslit holder tube imaged using
the same process as before. Cleanness o f the inlet aperture o f the m onochrom ator could
significantly influence the beam profile and consequently the beam brightness. We have used
a beam profile m easurem ent to measure and quantify this effect. The beam profile at the
nanoslit can well be described w ith a bell function S(x):
___________ I__________
S('X) ~ (1 + 4 (2 '"’ - \ )(x / F W H M fY
where p is a shape param eter; p= 1 gives a Lorentzian a n d p » \ gives a G aussian distribution,
the FW H M is a full width at h alf maxim um and I is maximum intensity. Figure 2 shows the
focused beam profiles at the nanoslit for two inlet aperture conditions. The difference in the
shape and in the peak intensity comes from the inlet aperture charging effect. The shape
param eters for the tw o m easurem ents are 5 and 0.7, respectively. The brightness could easily
decrease by 10 tim es due to the aperture effect. It is therefore im portant to have a means o f
testing the perform ance w ithin the filter, as we have, and to identify and rectify the effect.
601 (6661) LZP ‘Y M13IAI P W ’« su i - p n ^ ‘jiru ^ - j puB 5pop\[ f a H
a n ‘ooor ‘iix w anna ‘ipux d p«b ^oojaj m h "v^Sín 'c m ‘qvreg -37
•sjujod 831*831
surejuoo ireos 'uirl o i si m sta j o p p i j ' \ s / uiri iL'Z s! uoisradsip sq x 'suresq (jqSu)
pssjsdsip puB (ip ]) pasjadsip-uou jo ‘jq s m u QOS E JSao psmsBaui sp jo jd u re sg '£ o jn S ij
J3A0 p s jn s B s j^ -sa jn y a d B p [ u ; p a ir e a p Á jju s js jjip
ío j
jq so u B u s q j
■;tjs u iu o s i b
s j í j o j d u r e s g -j s i n S i j
•sqnj jsp jo q jqsouEU s q i j o s S p s d oj s q i
s q tq i s p jo q
(q ‘jqsouB]\[ ( b
•rnrl 001 s! aasia j o p p i j
s q j p u n o iB p s u r e jq o ss fte u n j u s u n o s jd u re s s q x ' I s in S ij;
'33U3J3JU03 3 q j JE p3JU3S3id 3q ][TM pB3idS
XSJ3U3 s q i puB u o is js d s ip s}u3ui3jnsB3U i s s s u jq S u q s q j u o sjíB jsp j s q y r y -jq s u iu 0 0 S E J3A0
p3insB 3U i 3 j e 3S 3qx 'IIISOUBU s q j je ssd B q s uiB sq p s s js d s ip puB p ssjra d sip -u o u SAVoqs £ 3 j n 8 i j
P e tr Č ižm ár
I n s titu te o f Scientific In s tru m e n ts, A SC R , K rálovopolská 147, 61264 B rn o , C zechia
cizm [email protected]
O u r version of th e m u ltich an n e l energy a n a ly z e r[1] [2] co n sists of tw o m ain p a rts. A t
first it is a n elec tro n o p tic a l p a rt, w hich se p a ra te s th e e n terin g electro n s by th e ir energy.
T h e n th e electrons lan d o n to a scin tilla tin g screen. S econdly th e re is th e o p tica l p a rt,
w hich should p ro je c t th e light signal on th e sc in tilla to r o n to a d e te c tin g C C D . T h ere
are several w ays how to do th is.
In o rd er to d eterm in e, w hich design is m o re useful for th e m u ltich an n el an aly z er, we
need to know th e efficiency a n d th e re so lu tio n of th e o p tic al system . B y efficiency
th e ra tio b etw een th e ligh t energy leaving th e sc in tilla tio n screen a n d th a t reach in g th e
C C D is m e a n t. T h e efficiency can b e e stim a te d from th e m ax im u m angle a betw een tw o
ray s e m itte d from one p o in t reach in g th e first o p tic a l elem ent. T h e efficiency p rim arily
d ep en d s on th e size of th e first o p tical elem ent. If th e re w ere no o th e r loses th e efficiency
w ould be:
F or th e reso lu tio n calc u latio n , w here we n eed to d e te rm in e th e size o f th e sp o t, to w hich
one p o in t is p ro je c te d , we need to tra c e th e rays. T h e re w ere tw o designs calculated:
th e tw o-lens design a n d th e elip so id al-m irro r design.
In th e tw o-lens design th e re so lu tio n d e p en d s on th e sh a p e of th e lenses, as seen in figs
1 a n d 2 as well as on th e d ista n c e from th e c e n tra l p o in t. In b o th designs th e sp o t size
exceeds 10m m . It is possible to decrease th e sp o t b y red u cin g th e size o f th e lenses,
w hich decreases th e efficiency.
T h e o th e r p o ssib ility is to use a n elipsoidal m irro r to reflect-th e b e a m a n d to focus it
o n to th e C C D . T h e sh ap e of th e m irro r follows from th e req u e st to p ro je c t th e p o in t
X o to one p o in t on th e C C D . T h e n th e sh a p e is elipsoidal. Sizes of sp o ts on th e C CD
w ere c a lc u lated w ith resu lt in figs.3, 4, 5 a n d 6. O bviously, th e re so lu tio n d ep en d s on
th e d istan ce from X o, w here th e re a re no a b e rra tio n s, a n d on th e size of th e m irror.
T h e n e a re r to th e Xo , th e b e tte r th e re so lu tio n is. T h e only a b e rra tio n s needed to be
co nsidered are th e m issalig n m en t ones, unlike w ith th e lens system s.
M .Jack a, M .K irk, M.M .E1 G o m ati, M .P ru tto n , R ev. of Sci. In stru m . 70(1999)2282
F.H . R ead , R ev. of Sci. In stru m . 73(2002)1129
T h e w ork is s u p p o rte d b y th e A S C R G ra n t A gency u n d e r g ra n t no. IA A1065304
u l o q 'i :o z is d o d d im - l u i o d jv d iu d o o y i
u u o f d o u v i s i p d y i u o o o u d p u o d o p d z i s i o d g :g -6i j
u l o ç ’i : d z i s d o d d i m - s n o o f p u v U0 i i 0d ÿ d } j :g •6%
m o j : o z i s d o x i i u L - í u i o d ¡ v d i u o o oí¡i
u u o f o o u v is ip o y i u o o o u o p u o d o p d z is i o d g : f ' f y j
m oj
: o z i s d O d d im - s n o o f p u v
u o i i o o if o y :g
'6i j
[luo]o x w o jj eouBisip
sdsudi pdiudido
ñ j v n b d m i m , u ß i s o p s u d j- o c n j^ :% -6i j
u ß is o p d d o o s d p i s u o \-o (n x - I
Armin D elong and V ladim ír Kolařík
DELONG INSTRUM ENTS a.s., Bulharská 48, 612 00 Brno, [email protected] dicom
E Introduction
Field emission emitters have significantly improved the coherence o f electron guns in
electron microscopes and allowed to reach atom ic-scale resolution in the case o f crystal
imaging. Even now, thirty years after the first application o f field emission emitters, one can
hardly say that the developm ent o f electron emitters has been finalized.
2. Cold field emission emitters
The principle o f cold field emission from the point o f view o f physics is relatively
simple and depends theoretically on two param eters only: the work function and the electric
field gradient close to the cathode surface. However, the real situation is much more
complicated. A more realistic description o f field emission o f electrons is as follows: being
emitted in a real technical dynamical vacuum, electrons are accelerated and hit the anode with
an energy o f several kVs. Their energy is sufficient for ionization and desorption o f adsorbed
gases and vapors on the surfaces being hit by the electron beam. Desorbed ions are
accelerated and not only destroy the cathode by sputtering, but change at the same time the
work function o f the cathode surface. Em ission current fluctuations are the results o f this
effect. This noise deteriorates different applications o f electron beam devices.
Two param eters can to a certain extent improve the unfavorable situation. It is first o f
all the im provem ent o f the vacuum by heating the electron gun to tem peratures over 100°C
for several hours. As an acceptable approach, vacuum systems pumped by ion getter pumps
may be considered. The attainable pressure should not be worse than 1O'8—10“9 mbar. The
second parameter, which can improve the emission properties, is the choice o f suitable
material o f w hich the emitter is to be made. It is to be noted that the proposed solution should
have one important property: emitters m ust be realized w ithout any special and expensive
equipment. M anufacturing costs, especially material costs, should be acceptable. Schottky
emission cathodes w hich are widely used are too expensive and far away o f the idea o f using
such cathodes with the aim o f improving the param eters o f simple instruments, which could
be cheaper at the same time.
3. M aterial for cold field emission emitters
The im provem ent o f field emission emitters, especially their emission current stability is
made possible by choosing a suitable material, which will better resist the ion bombardment.
Such m aterials are carbides o f transition metals [Zr, Hf, N b, Ta]. They feature significant
advantages against the formerly used ones: reduced w ork function, more stable emission and
resistance to ion bombardment. There are different ways to realize such emitters [1],[2],
It appears that the main problem connected with the realization o f cold field emitters
with sufficiently stable emission current is the ion bombardment. The idea o f an appreciable
pressure decrease has been rejected, being tim e consum ing and expensive. Fortunately, one
option still exists: to cover the point o f an electron emitter with a dielectric layer which will
protect the em itting tip from foreign material adsorption, ion bom bardment or even act as
a radiator thus protecting the tip against overheating [3], Electrons from the tip are emitted
into the conduction band o f the dielectric and under the influence o f the field, they are
transported through the conduction band to the surface from which they escape into the
vacuum [fig.l]. Electrons are heated during their transport with no change in their energy
opoqjEa u o is s iu is -p p ij ja q ij u o q jE j '£ '3 ï J
'J3ÁBJ a u jo s p íp
qjiM psjE03 jajjiuia
ue uiojj
U O IS S IU I3 U O J J O S p J O U I E j S E i p p U B Q ' [ ' S l J
£00c "Oř l"Sel LS6 XdoosojouiBfltn ‘Suijboo oujospíp b qjiM
sopoqjED uoissiura-ppij pjoo jsqij-uoqjBO jo uoijBsqiqBJS ‘\j x
puE‘'S'j/\¡ esnoj/\¡
¿661 unf/XBpv '£ ( g i ) y jo u ip a x -p s -dea f "ssoiAap uoissiuis p p ij jo j sjbijsjbui
dB3 puBq spi^v ‘T f lio-I[ I puE ‘ f f ou io n j ‘ g ^ V 1043 ‘ f D
‘'A 'A Aoím qz
6661 "Č8~ÍL "6L XdoDsojoiuiEJJif) 'suoijBonddE soiAsp puE sjajjiura p p ij psBoo-spiqjBO
uiojj U0ISSIUI3 p p ij psDUEqug ‘ ^ j siaeq puB'-g 3!X ‘'V/W sppejAj ‘W á JsiuuoqjBqj
£ 8 6 1 ‘6Se - 6 P Í ‘P ° N ‘ l £ 'Ioa‘'osoidt]/\[ uojpajg-f ‘saijjadojd
uoissiuia oiuoiuusqj jp q j puE sspiqjEO ájojobjjsj jo poqjsui uoijEJBdsjd A\au y ‘'}¡ BpE^
[ 3]
•X]uo mnnDEA sjBJspoui S u u in b aj
■jodBDqo qoruu sq []ia\ jsjjiu is u oissiura p p ij pjoa sq j p q j p a p a d x s aq u eo j¡ 'p p ij siq j ui
sjuouisA siqoE Mau S u u q K p jn s [JIM qojB3S3j jaqjjn j puE Sui3E jno3ua á j s a sq oj su isa s ja jjiu is
siq j j o o jd p u u d a q x JEqui
q í J ° uinnoBA e ui uoijBJsdo sjq E ip j jsS oj sjqE ai3A\ s a \ "5(J0m
3U |jnoj jo j p sp u aju i 3je qoiqM s s o q j X jjE padss ssd o o so jo iu i u o jjo a p j o spuií[ \\u SAOjdun
UBD a ojnos u o jjD sp j o pui>[ E qons jBqj SMoqs sjsjjiu ia ssaq j qjiM p s u p S a o u su s d x a j n o
'ssn jo j ÁpBai s | J3JJIU13 sq j ‘uinnoEA m j 0 00c 1® sjn o q je js a s s
jo j J3JJJUI3 aqj SuijEsq J 3 jjy 'uoijB zuauiX jod j s j je jsäb ] uiri £ 'o b 3aeS qoiqA\ ‘ÁjiS03SiA e qjiM
UIS3J X xods ojui pad d ip sem d p aqj S u iu B sp oiuosBjjjn J 3 jjy H O BN J ° u oijn jos jsjbm ¡oui i'o
UI X|[B3|JX|0JJD3[3 p sq o js SEM dij 3qj puE X xod s S uijonpuoo oj >[onjs sea \ ju su iejij UOqjED
a q x '[ č '3 y ] p ssn sem ju su iejij d o o | pjEpuBjs e ‘jajjiura sq j jo j asBq e s y -jsS oj Xsb3 á j s a si
puB iurl¿ jnoqB si j s q ij siq j j o jajsuiB ip s q g '[sj3m X [od 'S's] sjbijsjbuj ju s js jjip í?UTuaqjSu3Jjs
jo j p ssn X i[eu uou st qoiqA\ 'Jsqij uoqjBO u io jj psjEdajd si dij sq j 'ssbd siq j uj -[t7] 3JnjEJ3j|[
sq j uj p sq ijo sa p jsjjiu is oqj sq oj sjBsddB adXj jajjiu ia S u ijssjsju i X jE p sd ss u y
q.-)iqA\ ‘SJBU3JBU1 j o uoijBUiquiOD E qons p u y oj Xjuo suiij j o jsjjeui b st j¡ -jajjiu is j o sdXj siqj
qj]A\ poASiqoB u a sq 3AEq qoiqM sjjn saj aqj S u iq u o sa p sjsdE d
jejsass sjb
a jsq j, 'uoijnqjjjsip
M. E scher1*, M. M erkel1, N. W eber1, S. Schmidt2, F. Forster2, F. Reinert2, C. Ziethen3,
P. Bernhard3, G. Schönhense3, B. Krömker4, D. Funnemann4
1Focus GmbH, 65510 H ünstetten Germany
2U niversität des Saarlandes, FR 7.2. Experim ental Physik, 66041 Saarbrücken
’Johannes Gutenberg Universität Mainz, Institut fiir Physik, 55099 Mainz, Germany
4O micron NanoTechnology GmbH, 65232 Taunusstein, Germany
* m [email protected] focus-gm
A new instrum ent for Imaging XPS has been developed in close cooperation between the
University o f Mainz, the University o f the Saarland, Focus GmbH and Omicron GmbH - the
NanoESCA. The instrum ent consists o f a Photo Emission Electron M icroscope (PEEM ) optic
as entrance lens and a newly developed double pass energy filter (patent pending) that con­
sists essentially o f two hem i­
spherical analysers. The en­
ergy filter is designed to give
an achromatic image that is
corrected for the aberration
o f dispersion o f a single
hemispherical analyser.
The microscope allows im ag­
ing with chemical contrast
(Imaging ESCA) by energy
filtering o f photoelectron
images at kinetic energies up
to 1.6 keV, w hich are typical
for XPS. Additionally, non
energy-filtered PEEM im ag­
ing and a pulse counting
m ode for quantitative small
spot spectra are im plem ented
Fig. 1: Schematic layout o f the nanoESCA instrument. The
into the instrument. A sche­
three paths o f the electrons are indicated:
matic layout o f the instru­
1. PEEM -M ode, 2. Selected area spectroscopy,
ment showing the electron
3. Energy filtered ESCA imaging. The grey box en­
paths o f the three different
velops the elem ents o f the PEEM mode.
modes is shown in Fig. 1.
First characterization m easurem ents were carried out at the Bessy II synchrotron radiation
source (Berlin) and w ith laboratory sources. Fig. 2. shows measurem ents made on
a Cuo.9 sBio.o2 poly-crystal that p roof the segregation o f the Bi towards the grain boundaries o f
the crystal. The low background intensity o f the Bi 4f-spectra taken on the Cu-grains dem on­
strates the very good capability o f the instrum ent for spectroscopic imaging. We obtained an
analyser energy resolution o f AE < 190 meV (FW HM ). This was derived from the width o f
the Fermi edge at room tem perature assum ing a photon line width o f 80 meV. The measure-
J O S e 'd ‘( c 0 0 3 ) J J o d s y |B n u u y K s ss g ‘-[B p u u B iu a u u ry q
m i ' d ‘(ZOOZ) p o d a y p m u u y X s ss g ‘ IB P i p i u n p s 'S
fr9 8 Z .M £ I'O N -Z )iJ HIRJ3 ‘j g j A i g Xq p a p u n j XpjBd
sbm p a f o j d
s iq x
•y^sui 0 6 1 = lul|" '3 V j o u o ijtq o s a i issX jbub
'(S1°P) SldlUBS 3UIBS s q j JO 3Sp3-TlXIJ3X IjqSl'a
■aSBixii ; g s q j u i jq S ijq sjq isiA ‘(sjo p p s n u ) sa u B p u n o q
u ib jS a q i u o pire ( s jo p u s d o ) su ib jS s q j u o |SA3)-3J03 £ / / J t ^ g 3H)
S u iS b u ij :i j s q
■qipiM u iu i[ j o s jq s p u e X S jaro ssBd A 30S q iiM spB iu
3jom sju 9 u i3 jn sB 9 iu a q j '|BisAJO-X|()d jo 0j g 86 onQ b j o s u o ijb S ijs3 a u i o i d o a s o jp s d s
[/\0 ] A6 J0 U0
6 u|pu|q
[/\8] A6 j 8U8 6uipuiq
. ...
- - ,- 1 1 1 . , , __'
ooe I
\ f (!9 ) seuepunoq uo —•— ' 009 ‘
(no) u|bj6 uo —0 — •
:se0je p0 ;ej6aiu| . 009
A8 009 = Aq
A © e 9 ' 9 5 L = a3
•(UIU09-017 u o ijrq o s a j ]BJ3jB])
SuiSBUII IM g g d 3qj
“ f101
J3JJ1J X8j3U3 s q j Xq psjBJ
B qjIM SJU3ra3jnSB3JAJ
JBql 3JB3ipUI 3 3 jnO S -y\n AJOlEJOqB]
;u 3 u iu o jia u 3 u o jjo jq s u X s s q j u i suopB jqiA jB oiuB qosui Xq p a jiu iij m o u o j d n si s u o i p s p
'S d X S u iS b u ii
u o p r q o s a i |BJ3jB| o q j ■% ( g f g g ) j o jsb jju o o b qjiM pA O |-ajoo d z IV
IB ua^B j s a i n p r u j s a u q u iu 0 0 £ p s A jo s sj Xjjnjssaoons OABq 3M a p o u i - y Q s g S u iS b u ij aqi 8 u iS [j
001 M o p q s u o p n jo s a i X S jsu s uiB jqo o j p a d x a
[buiiuiui b joj
p s u S is s p si ju s u ir u js u i a q j s y 'sq ip iM jq s puB X S jsua ssBd
aasX|BUB jB D u aqdsiureq
jbuoijusauoo 3]3uis b jo
-u o o si s i q j -[z‘ \] \ s OS J° ^§-i3U3 ssBd
q z J°
a q j S uiuinssB
u o q r q o s a i XSjouo p o p o d x s s q j qjiM
q iiw XSj s u s oipui> [ \ s 0 0 9
inoqE jb u 35[bj
sb m ju s u i
A. Feltz, U. Roll, G. Schafer. J. W estermann
OMICRON N anoTechnology GM BH, Limburger Str. 75, 65232 Taunusstein, Germany
During the last few years, we have developed a new electron column as an excitation source
for UHV applications such as low voltage SEM, SAM, small spot Auger spectroscopy,
cathodolum inescense and others. The instrum ent was developed in a cooperation between
LEO (now ZEISS SM T-NTS), CEOS GmbH and OMICRON N anoTechnology GmbH.
Although the instrum ent is based on the LEO Gemini electron and uses a similar electron
optics, it is constructed alm ost entirely from different parts and materials to make it UHV
com patible and bakeable to 180°C.
FIG 1 SEM image and lineprofile
Spot Size:
dX = 2.3 nm
Distance [nm]
Gold islands on graphite, raw data.
15 keV beam energy, 400 pA beam
current, 8 mm working distance.
One main feature o f this electron column is the ability to achieve <3 nm ultimate spot size at
working distances as large as 8 mm, but it is also capable o f small spot sizes at low beam
energies, such as <10 nm at 1 keV and 1 nA.
In the present work we now report about the results obtained with Auger spectroscopy and
scanning Auger microscopy utilising this source. The ultimate spatial resolution in scanning
A uger mode was so far found to be below 10 nm (see Fig. 2.).
We explicitly distinguish between the spatial A uger resolution and the probe diameter. The
latter is related to the SEM resolution, which has been found to be 6 nm for this measurement
(10 keV. 1 nA). The spatial Auger resolution however is also influenced by the energy
dependent interaction volume, edge effects, image contrast and other Auger-intrinsic
param eters fl. 2. 3. 4], Although these early papers describe many o f the effects that we see,
no studies are so far available in our param eter space with sub 10 nm probe sizes, beam
currents in the range o f 1 nA, and beam energies from 30 keV down to 1 keV.
As an example we present a combined SEM, SAM, and Auger spectroscopy experiment (Fig.
3.). The aim o f this experiment was to determine why the growth o f Ag islands on this Si
substrate showed such irregular island shapes. Usually the islands are well orientated
alongside the com er hole directions ({2,-1,-1} and {1, 1,-2}) o f the underlying S i ( l l l )
surface (com pare with Fig. 2). The spectra and SAM images revealed a contam ination o f the
sample with Molybdenum and Sulfur.
K £ -£ £ £ (£861) Sel s o u s p s ^ B jjn s 'x n u z e j 7
(6861) 99£-t?S£ T i 'I ° A ‘sisXpmy s o b jís ju j puč s o B jjn s ‘xnBzt? 3 (
(8861) ? 9c-IS Č T I I°A ‘sisX jE u y s o b jjs ju i p u s SD Bjm s ‘\e ja p B u io o ¡ g p\[TA[
91 e -60e ( 6 ¿ 6 l) S8 33U3IDS soB jang '\e 19 p E u io g | g -yv'JAI
’( 1 9 3 ‘sjsK b jo u o jaj p s iq m s s s y j p s ) SJ3ÄB[ j s i s s j p s m ja ru jso u B u j o
uo p E S p ssA u i jajx S 3M1 3C1 II!M u o p e u iq u io o s iip j o j uopB apddB
a q j ‘j s a s m o h 'jo d s 3 |d u re s
su re s s q j u o u o p r q o s s j a id Ol oj UMop 3[bos m m u i o j j m ooz b ssjqB ua ji puB ‘S u ra o p is o d
d p ’S-3 j o j |n j d p q X p m sjjx s p u n o j u s a q seq j ^ g s pire j ^ x S i ° uoijBuiqm oo a q x IALLS qi!M
u o p e u iq iu o o u; p s s n u s s q o sjb SBq lu im a r j A H f l 3lP ‘suopB opddB j s 8 n y p u e jrç g g J0J s s p i s s g
(um) eouB(sig
mu 9
jo uoprqossj b
S JB 3 A 3 J
S b iu t jA jg s
•juauno uresq y u [ ’A sii 0 1 X Sjsus uresq
*3mBjj js d uim £[ ‘p x i d t?8£x ClS ‘(9 + d )/(9 -d)
LA3 06£ 1B punojS>pBq ' a s cS£ ‘N N W
'(% 0 8 -% 0 c) u o p n jo sa i sg p a m u 5
u o p n jo s sj jA jys ajBm pjfj :
I’ Hammond1*, F H Read2, T J Reddish’ and P J Harvey2
'Department o f Physics, University o f Western Australia, Perth WA 6009, Australia.
“Department o f Physics and Astronomy, University o f Manchester, Manchester M l 3 9PL, UK.
^Department o f Physics, University o f Windsor, Windsor, Canada N9B 3P4.
*email address: [email protected]
Most storage rings and circular charge particle accelerators use magnetic fields but for studies
in atomic and molecular physics a purely electrostatic storage ring has the advantages o f small
size, low power consumption, no magnetic hysteresis and the ability to store ions o f any mass as
electrostatic deflection depends on energy and not on momentum. The ‘ELISA’ storage ring [1]
was the first o f this type and other similar electrostatic ion storage systems are being designed
and constructed [2,3].
Conventional 180° hemispherical deflectors were not used in ELISA because it seems that
their ‘strong focusing’ property [1] causes instabilities. Instead 160° sector hemispherical
deflectors were used initially because o f their equal focussing in the ‘vertical’ and ‘horizontal’
directions. Later the deflectors were changed to a cylindrical geometry [4] to avoid the reduction
o f the beam lifetimes at high beam intensities caused by the existence o f narrow beam waists.
We have designed (and are constructing) the electrostatic ring 'ERS' (Electron Recycling
Spectrometer) that will be used initially as a novel source o f mono-energetic electrons and later
as a storage device for positrons and polarised electrons. The ERS incorporates 180°
hemispherical deflectors because o f their good energy dispersion. In the present paper we
consider the stability o f the ERS and describe a new mode that has been found to be stable.
The ERS is intended primarily for use with low-energy electrons. A cut-away view o f it is
shown in Fig. I . It consists o f two hemispherical deflection analyzers (HDA) between which are
electrostatic lens systems. The C P03D program [5] has been used to model it.
View o f one half o f the Electron Recycling Spectrometer showing one o f the
hemispherical deflection analysers and the lens stacks at its entrance and exit.
'8 8 6 1 LPJOJ x O 'SS3J<[ uop u sjE O sotido apij.wj paS.wifj fo s'jisai/j a i/j ‘u o sm b j q f [/J
6861 ‘sjj3snipEsse|/\[ S u ip say ‘s>|oog
s n s s j s j "p; | j -snjrunddy otfijusps Suip¡mg ‘u e |d o 3 'y ' [Aj puB s ia b q j j ‘sjoo|^| » 'f [9]
/Luo3'S3ijdouojpa|3'A\AA\\//:djji] ajis qs^V sqi uo 3|qB|iBAB ‘suibjSo j j O d 3 [S]
( 1OOt) SO I 1 6 1 J3S s^Md u o s js p a j a ' q puB J3)[0W ¿ ■$ fr]
•(eooč) £ S t 9Pl J3JU| d¿H Su^oga-ipiuiqos 'H Pub dduunps 'y ‘ipspM ¿ 3 [£]
'(cOOc) S6S ‘Z 8f V 'HPIAI 'isui - p n ^ sqcuBjuBAV j ‘BpoN » ‘Bpiq^ » ‘sq B u si -j
' ( ¿ 6 6 1) 18Č 'P 6 £ V 'm 3lAl 'l s u l ' P n N ‘J 3 1I0 IAI ' d ' S
S 0 3 U .1 J0 J.1 »
•XSjsus sojn os sq i uiojj lu sjajjip si XSjsus
p 3 jb i sip u 3 i|M jnoiABqsq sq i (£) puE j s p u io j p s d s sq j j o uoprqosaj XSjsus aqi ( 3 ) ‘suoyEjjaqB
j o s p s j j s 3i)i (1) aprqoui |iB p p ui p sssn osip sq ||im apoui m su siq i j o ssiy a d o jd a q x
■XUIBUI iiun 3 i|i si S7Y qoiqA\ ui sp ou i |BU0IJU3au0D ajoui sq i jo j 9SB0 sq j jou si qoiqA\ ‘XjiqSqs
Xjba y a n 3i|l puB S3SUS| 3Lji j o sjspuiBJBd sq i uaqA\ 3|qBis su|B uisj uoijBjado j o sp ou i siq i
IBqi puij
su 0 !iri|0 Aaj om i jsy B p sjo isa i s; inq uojirqoASj auo jajjB paysA iu usq i si sSbuu s q x
W XZ = W
3 u [a \o ||o j sq i j] puB i_y = n j¡ uisisXs lu ssa id sq i joj paysjiBS sq ubd qo|qM puB n j o srqBA Xub joj
S3A|S qo;qM
n \
= "N
nv y
Jw m
§ uim oj | oj sq j Suisn Xq Xem
jsaou b ui
uoijipuoo Á iijiq ^ s siqj /Çjsijes oj uasoqo 3ABq
\l \ siiqjo aqi j o
Xji|iqBis sqj jo j uoi}ipuo3 |BJ9U3S aqi puB
’SP\[ nM = SSN
usqi si sojnos sqi 01 >|3Bq sojnos sqi uiojj xijibui jsjsu b ji ||b j3ao sqj_ 'sojnos sqi ib }Bqi
SB 3 UIBS sqi si isSjbi aqi ib XSjsus u o j p s |3 sqj iBqi os ouisuiuiXs si uiaisXs sqi iBqi sjsq psmnssB
s ! il V a H S11J ° l 33UBJ1U3 sqi puB 3Djnos aqi j o suom sod aqi 6} js ja j q puB ¿ 3J3qM puB
ZJ - Ö = ZN
‘ !i ~ J = l.Y
¡x ' x z - y
Zy \ z -
y / y
s; (X|3A|p3ds3j SLUJB J3M0) pue jsd d n sqj
j o 3[ppiui 3qi u| 3JB qoiqA\) p 8aei sq i 01 sojnos sqi ujojj x u ieu i jsjsu b jj sqj jo j xuibuj psujqujoo
sq i VQH 341 PUB S3SU9) 3ql JO S30UJBUJ J3JSUBJ1 3ql JOJ [9] SUOIlIUjjap |BU0IJU3AU0D 3ql §UISf)
■uoiirqoASj ajSuis b j o uoiiE|nui|s siBjnaoB ub Xq psqsqqBjss sjb p q i ssdijibui jsjsu b ji
sssn iBqi p o q p u i b pado|3A3p g jo jsja q i 3ABq
's jo jjs |BOuauinu 3A|iB|nuin3 j o aSBiUBApBSip
aqi SBq s¡qi inq ‘auop 3ABq 3M qoiqA\ ‘su0|in|0A 3j j o sjsquinu 38jb| qSnojqi ssuopsfB j} sqi
sibjSsiui 01 ujBjSojd uoiiB|nuiis b ssn oi s| saïAsp sq i j o XiqiqBjs aqi Xpnjs o} Xbm snoiAqo s q x
P. Horák, P. Schauer
Institute o f Scientific Instrum ents AS CR. Brno, Czech Republic. horak(®,
Polysilanes - a broad class o f organic m aterials whose basic building block is a chain
built up o f silicon substituted by alkyl or aryl groups. Properties o f these substituents
significantly affect properties o f polysilanes. This material is very interesting because o f its
chemical, electrical and optical properties [1]. In spite o f great research interest in recent years
cathodolum inescent (CL) properties o f polysilanes were not yet studied. Poly[methyl(phenyl)
silane] (PM PhSi) is a typical representative o f polysilanes which was prepared by the Wurtz
coupling polym erization. Thin layers o f PM PhSi were prepared from toluene solution by
a spin coating technique [2], The material was applied on the quartz glass substrate and
covered with the alum inium (Al) film. The A1 film protected the specimen from charging and
reflected photons emitted under the specimen surface towards the quartz glass substrate.
The m ethod o f PM PhSi study was based on the measurem ent o f CL intensity after
passing through the specimen (Figure 1). Electron beam, emitted from a wolfram cathode,
accelerated and focused in the excitation part, struck the Al deposited specimen. Emitted
photons which passed through the specimen and through the substrate were led by a light
guide to the detection part.
The excitation part was based on the rebuilt electron microscope TESLA BS 242.
Electron beam energy was variable from 1 to 60 keV. The deflecting system allowed
modulation o f the electron beam and enabled measurement, not only in the continual mode,
but also in the pulse mode. This was very
im portant for the
measuring in the
Electron beam
synchronous mode as well. The detection part
was based on a photom ultiplier tube [3].
Deflecting system
The first PMPhSi specimens o f the
A l layer
thickness o f 2
covered with 50 nm o f the
\ \ \ \ ,
photom ultiplier tube TESLA 65 PK 415. The
Quartz glass
excitation energy o f 10 keV was used for
cathodoluminescence intensity measurement
in the continual mode (i.e. the deflecting
( Light guide
system is switched-off). Current density o f the
incident electron beam was approximately
5 nA / m m 2. The CL emission from PMPhSi
Emited photons
was smaller by 2 orders o f magnitude in
comparison with the CL emission from the
YAG:Ce single crystal and measured under
the same conditions [4], Strong influence o f
the background, above all o f the cathode light,
was proved and also measured by switching
F ig u re 1 Layout o f the experim ental arrangem ent
o ff high voltage (only w olfram filam ent is
fo r the study o f cathodolum inescent properties o f
polysilanes including the specim en p a rt details.
c l - I '( 9 6 6 1 ) 6 ( íu s u i s j d d n s ) X d o a s o jo ij^ S u iu u b d s : j ‘jsnBqDS í y ‘ B j B j j n y
'0Z.-/.9 ‘(cOOc) -inAQ X^jsjb^s ‘u o ijB ju s u iru js u j s o is X q j soBjjng; puE s o i j d o s p i y B d
p a3 aB q 3 u i s p u s j x j u s a s y u o j e u u u s s 'j s j u i q j¿ j o m y :
‘B j E j j n y í j ‘ja n B q o s
' L Z Í - 6 l£ ‘(6661) 6-8 J '3 u g puB -p S 'Jsjb^m :'S >pm dssN
'0 ll7 l-6 S £ l '(6861) 6 8 A3>I ' ^ 3 : 'f B N " 0 '^ » I I M
s a r> u a jo j.i> |
•ssjbjs sjqEjSEjaui Suijesjo
a jE o ip u i X bui ( g jjB d ) s j n u i u i l(Jot^ l s ifl u i sb |p M sb ( y jjB d ) a jn u iiu ^ o t 7 P UB 4 l0 l 3 lll u ! s a j i j o
a q j j o s d o j s s q j j o s a S u E q o s q j ’snjBA X jis u a ju i [B ijiu i a q j j o % \ Á [uo p a q D E a i j n q ‘3|qB U BA ui
s n o iA S jd s q j j o
s j n u i i u ^06
'p a A J s s q o s b m ( %
JBqi o j [B O ijuapi X jjB au s b m s j n u i i u l)t0 c l
c ) ^ J is u s ju i X 3
aqj jo
snjBA X jis u s ju ;
á j s a o o s j j e i j j e j -(q d E jS s q j j o
jjB d )
s u o ijip u o o s u ib s s q j j s p u n u iB sq u o j j o s p a q j Xq p s j p x g u i e 3 e s b m i S H J H d 3 lIl ‘s -in jE ja d u is j
u i o o j je a d o o s o j o i u i a q j j o
js q u r e q o uin n o B A s q j u i j j s j u s i u p s d s g q j q jiM ‘u o i j E j p x s o u
j o s a n o q q z j a U V '( s s j n u i u i o í u ; % [ jn o q E ) u o i j E j p x s j o s s j n u i i u q ç j s j j e ‘á jm o js s j o u i q o n iu
j n q ‘u o S u io S [|¡ js s b m íSH dJA ld J ° u o ijB p B iS s p a q x '( X jis ira ju i s q j j o s s B a jo s p % o/. jn o q B )
s s j n u i u i o í j s j y s q j u i u s a s s b m (q d B jS s q j j o y jiE d ) u s i u p s d s m s u a q j j o u o ijE p B jS s p j s s j s b j
u o ijB p B iS a p p u s j E i u b o j p s u S is s B s b m [S q d lM d J ° A i s u a j u i
1 3
j o a sB S Jo a p s q x
•(ZH>[ [ ‘ -IUUI / y u Ç ‘ y\3>[ 0 1 ) SAOqB U3JJUM SE 3J3 M JU3UI3jnSE3UI S iq j JB
s u o ijip u o o ¡ B ju s u iu s d x s
3 - m S ij) q d B jS 3 iiu q ju E S o [ iu i 3 s u o p s jjo jd p u s jajsiu jjO A O U B u
1 3 is q d J A ld 3 4J J ° s3 u E q o s q x ' g Z Í Z N V d l N i l
ui-> po| s iq j 3 u is n p ajn sB S U j sbm X jis u s ju i
j3 jaiu j[O A O u B u ui->[30[ s q j j o
jnduT s o u a j s j s j s q j o j p s jo a u u o o sbm jo jE ja u a S as]n d aq j
j o jn d jn o j s S S i j j a q j u i o ij us^ bj s s jn d s d u s j s j s j s q x 'zh^I I sbm u o ijB jn p o u i j o X o u s n b s jj s q x
1 £ £ 0 ‘ >I1 J o je js u s S s s jn d s q j S u is n p s jB | n p o u i sbm u iB a q u o jjd s [3 s q x 's s io u puB punoj3>(DEq
sqj jo
3 3 u 3 ri|ju i s q j s j B u iu iip
o j p a sn sbm
s p o u i s n o u o jq o u Á s
s q j u i SuunsE3j/\[
(g u v d ) sdinuiiu o fijx d u u o f duniDAddiud} uioou jn adoosojoiut f o udquiDifo uinnoüx v ui SuiADdj/o sunoif
0 1 udifo úpudnbdsqns p u n ( y juvd) sdjnuiw 06 AOIp d J n s v d w Kjisudjw j j ¡S H dW dIo soudpudddp diui¿ £ d jr tS tj
[UjLU] UOj)6)jOX3
011 001
__ _
__ __ __
9 p o u j s n o u o ji^ o u Á s 6 u ¡s n Áq
P0JB JjLU!|0 - |0A 0| p u n DJD>|OBg
0 p o iu s n o u o jip u A s u| p a jn s e a iu iSMdlAId i ° u o ije p e jß a Q
P. Hrnčiřík and I. M ullerová
Department o f Electron Optics. Institute o f Scientific Instruments o f the A cademy o f Sciences
o f the Czech Republic. K rálovopolská 147, Brno 61264, Czech Republic
email: pecan'a'i
The inelastic and elastic mean free paths (IMFP and EMFP) determine the largest
sample thickness usable for transmission electron microscopy (TEM). At primary electron
energies norm ally used for TEM (>50 keV ). both mean free paths decrease as the primary
energy is lowered. If the primary energy is lowered to below about 100 eV. IMFP is predicted
to stop decreasing and to begin growing again [1, 2], This opens up the exciting possibility o f
very low voltage TEM o f sufficiently thin samples, with poorer resolution but greatly reduced
radiation damage compared to conventional TEM.
Previous investigations o f this effect employed electrons field-emitted from a nanotip
[3], The incident electron energy was determined primarily by the tip-to-sam ple distance, and
was not adjustable below about 40 eV. Similar experimental set-up as ours was proposed in
[4], but with no experimental results at that time.
In our arrangement, we have used a scanning low energy electron microscope (SLEEM)
designed for the operation down to zero landing energy [5], An addition detector o f
transm itted electrons m odifies this microscope. The resulting combination o f SLEEM /
SLETEM (scanning low energy transm ission electron microscope) uses the cathode lens to
decelerate the prim ary electron beam ju st in front o f the specimen surface, and is able to reach
a resolution o f a few nm even at landing energies in units o f eV. This provides with flexibility
to investigate the transm ission o f electrons through thin samples at electron energies as low as
1 eV.
A set up o f the detection part o f the instrum ent is shown in Figure 1. The primary
electrons at 10 keV are decelerated by finely adjustable negative bias to the specimen. Signal
electrons, both reflected and transmitted, are accelerated to detectors at ground potential. The
SLEEM detector is formed by an annular disk o f YAG single crystal with the diameter o f the
opening around the optical axis o f 0.3 mm; the SLETEM detector consists o f a PIN diode
collecting the total transm itted signal. The geometry o f SLETEM and the collection efficiency
o f the transm itted detector were optimised by using SIM ION 3D software [6].
The experimental instrum ent is based on the ultrahigh vacuum SEM Tesla BS350 that
was equipped by a Schottky field emission gun. The vacuum system and control electronics
were fully renovated, too.
We have shown first images at low energies o f a holey carbon foil approximately 20 nm
thick [7J. In this study we used a 5 nm carbon foil; the images recorded at different energies
are shown in Figure 2. Unfortunately, no indication o f increase in the foil transparency at
l - 2 e V compared to 20-100 eV was shown again. At certain energies we have noticed
a higher contribution o f secondary electrons to the total signal from the specimen. The signal
from the foil is then higher than that from places where is no foil. Landing energy, at which
this effect appears, depends on the foil thickness. This could be very easy method how to
measure the sample thickness.
More detailed study is necessary in the future in order to understand the inherent
•y 3 g y j o Xdu3§ V ju b j q J O 5 0 ^ 5 9 0 0 9 '° N p a f o i d ipjB SSSj Xq p s jjo d d n s si
s iq x
■sssjd u i 'u in iS p Q ‘d a s M ju y > 0 0 c J 1AI3 J ° 's o j j :ui ••[B js j ^ u id u jjx
' ¿ I L ‘B I§J03r)‘BJUBUV
'5 6 6 1 ‘s a id o x p siJIV PUB Â a p a io j;o o d g ssBpv u o s o u 3J3j u o 3 S W S V PJ £t? ‘T a m Ba
'60£ ‘£00č J3IA3SI3 ‘(S3>[A\BH
puB S u iS b u ij ui -A py ‘ q ^ u e j j puB [ BAoa3 [[njAj
'M 'd
P 3 ) 81 1
s o is A iy
•g - o u j g ‘tO O c SP U3J1 J u s a s 'H u 0 J B u i a i s s -jB U iaju i q ig j o
o o - y :u i j b js "IAIT"PM u s s t o ^ j
[g ]
Z6 (9661) 81 SuiuuBOS ‘"H n z iu iiq s V ™ T 'Z § u i a
[BoisXqj ‘ H ß l ^
101 ( ¿ 6 6 l ) 8 t u o ra M ‘'H '3 T 3 3 U 3 d s
' ZZ5 ( 0 L 6 \ ) I
'A ^ £ (P PUB ‘A300Z, O 'A 3 0t? (q
"A3 Z (BJ °
u o jp s p SuipuBj sqj jb U3>[BJ ‘p o usd uirl
jo pu3 jsddoo b uo uiu g
jo sssiqoiq} b j o [ioj uoqjBD b jo ssSbuii (M opq) H 3 I H 3 S P ul3 (SAoqB) JA J 3 3 3 S '-Z a j n g i j
•jusiuruisui JA 13133S / W 3 3 3 S 3lP J ° S1JBCI uo p o ap p jo dn-jas {B ju suiu sdxa a q x ¡I a -in S ij
Jiang Jinhua and Anjam Khursheed
National U niversity o f Singapore, 4 Engineering Drive 3, Singapore 117576.
e-mail: [email protected]
The finite elem ent method is a powerful numerical method widely used to solve field
distributions in charged particle optics [1, 2]. One o f the m ost challenging aspects o f using
the finite elem ent m ethod is to create a suitable mesh. This is a non-trivial task since there are
several requirem ents that m ust be simultaneously satisfied. Firstly, for reasons o f optimizing
program run-tim e and m inim izing mem ory usage, the mesh should be concentrated in areas
o f high field strength. Secondly, the mesh spacing should change smoothly, particularly in
and around regions o f high field strength. Thirdly, the mesh should fit a wide variety o f
electrode shapes, both polygon shapes as well as curved boundary shapes. Fourthly, the mesh
should be structured, this m akes it easier to use high-order interpolation techniques for
plotting accurate trajectory paths o f charged particles [3], Lastly, the mesh generator should
be interactive and easy to use. This paper describes how such a mesh generator was
developed w ithin the Electrical and Computer Engineering department o f the National
University o f Singapore. The mesh generator w as written on a visual C++ platform.
The mesh generator functions by highlighting region blocks on a background region block
m esh and locally m odifying or transform ing it to the desired shape. Figs la and lb illustrate
this process for polygon electrodes, while Figs 2a and 2b illustrate this for elliptical shapes.
Figure 1: Creating polygon shapes
I - js y J o f r T I ra id s ii3
[£ ]
■)7| - £ ¡ d ‘s S u ip s s o o .y jB u iu ia s ^8 ‘u o ijB iu s u iru jsu i
s o is K q j s o B jm s puB s o n d o a p o i B j psS jB qQ u i s p u s j x lu a o o y ‘BAOousq g
[3 ]
‘u o j s o g 'j 3A \n j^ ‘s o i jd o s p i y B j p s§ jB q 3 m p o q jsj/^ j u s u i s j g s j r a i j s q x ‘p a s q sm if)} -y
u o p r q o s j / ^ g j üb u io jj s s m j ¡ B p u s jo d m b g (d)
q s s j ^ s u i j (q )
s^pojg uoiSsy (b)
9[duiBX3 o |o d n jp B n f) y \<¿ s j n S i j
(3 )
(q )
•qssui siqj Suisn
uoprqos ju s u isp 3JIUIJ B jo saut] [Bpusjodmbs sqj SMoqs o g 'S ij 'q£ puB B£ ’sSi j ui uMoqs
SB ‘3ÍJU30 3ip JB UOlSsj 3JBnbs AjISUSp q§U] B UIOJJ SuipBjS puB JBjnOjp sq OJ S3pOJJ33]3
3JBnbs Sutuuojsubjj S3AJ0AUT qoiqM 3|duiBX3 3[0drupBnb B st SutA\0|[0j sq x 'Xjqioouis
ssSireqo Xjisusp ssoqM qssui b sjbsjo
uraiSoid sqj puB p s p p s d s X|sno3iiBj(nuiis sq ubo
suoi§3i Xjisusp qSiq ]BJ3A3g 'qssui sqi spBjS oj XjpiqB sqj ssjBiodiooui jojbjsusS qssui sq x
ssdBqs j B o p d p p 8 u ijb 3 J 3 :j s m S i j
J. Jirák. P. Černoch , R. Aulrata
Institute o f Scientific Instrum ents ASCR. Královopolská 147, 612 64 Brno. Czech Republic
* Department o f Electrotechnology FEEC BUT. Údolní 53, 602 00 Brno, Czech Republic
[email protected]
For signal electrons detection at environmental conditions, it is often used a detector that
utilizes impact ionization in the gaseous environment o f a specimen chamber for a signal
amplification. An electrostatic field created in a space between a grounded specimen and
a detector electrode provides the energy needed for efficient ionization o f secondary
electrons. A voltage o f several hundred volts is attached to the detector electrode. Danilatos
f 1] proposed an electrode system o f a ionization detector with several concentric electrodes,
where a higher diam eter o f the electrodes brings a higher contribution o f backscattered
electrons to the detected signal.
In our experiments, the properties o f the electrode system consisting o f four concentric
electrodes were studied. The electrode system showed in Fig. 3 was created by the printed
circuit board technology. The inner diam eter o f the sm allest electrode is 2.5 mm; the outer
diam eters o f the electrodes are 4. 9, 14 and 19 mm. The gap betw een the electrodes is
0.5 mm. For this electrode system, a dependence o f a signal level from a golden foil specimen
on pressure was measured at the varied interconnections o f the electrodes. Results o f the
measurem ent where the signal was detected from the electrode A which was on a potential o f
400 V and outer electrodes were either not contacted or grounded are plotted in Fig. 1. In
following measurem ents, the signal from the electrodes A and B on the potential o f 400 V,
respectively from the electrodes A, B and C on the same potential (Fig. 2), was detected and
the outer electrodes were not contacted or grounded. The measurem ents show the increase o f
the detected signal level caused by adding o f additional electrodes. The configuration with the
grounded outer electrodes provides always a higher signal level in comparison with not
contacted outer electrodes. Also the maxima o f the signal level occur at a higher pressure for
all m easurem ents with grounded outer electrodes.
A sim ilar electrode system with electrodes divided into two halves - right and left halve was prepared for other experiments. A specimen, consisting o f A1 metallization on a silicon
substrate, was observed by this electrode system. An image o f the specimen at the electrodes
interconnection, w here the signal was detected from the left halve o f the inner electrode with
a voltage o f 300 V and all other electrodes were grounded, is shown in Fig. 4a. The same
specimen and similar detection conditions but with 300 V potential also on the right halve o f
the inner electrode is pictured in Fig. 4b. A shadow effect visible in this figure shows that in
the case o f signal detection with a ionization detector, when the process o f impact ionization
lakes place in the space between the specimen and the detector, the effect o f the position o f
the detection electrode on the image contrast manifests itself.
This research has been supported by Grant Agency o f A cademy o f Sciences o f the Czech
Republic, grant No. S2065107.
■/Cdoosoj3ij/\[ uojjoajg SuiuuBog
Jb j u s u i u o j i a u j j
'd 05c ‘8661 ‘SS3Jd aiuispEoy ‘Asup^s
jo suotjBpunoj : q r> ‘S 0 1 V 3 IN V Q
qt/ Si J U O J03JJ3 M o p cq s ‘J O J 3 3 J 3 p U O I J B Z I U O I |BJU31U83S
p3A[Bq Xq p sjD sjsp sjB jjsq n s u o o q is u o u o i j b z i [[B J 3 u i [ y 'q'Et? '§ !J
[I ]
JO J3 3 J3 p U O IJB Z IU O I
[B ju su iS ss ■£ 'S ij
p s p u n o jS jo psjD B juoa jo u q s p o j j s s p
00t? 1B 3 ‘9 ‘V ss p o jj3 3 ]3
UOIJ33J3p ‘SJOdBA J3JBA\ JO 3jnSS3jd UO (3A3) [BUÍfrs JIOJ U3p|Og JO SSOUSpUSdSQ '£ 'S lJ
pspunojS jo psjoEjuoo jou q ‘j "g sspojjD sp
001? J® V spo jjo ap
U01J33J3P "SJOClBA J3JBM JO 3jnSS3jd UO [3A3| [BUSlS [IOJ U3p|0§ JO S30U3pU3dS(J \ 'S lJ
Anjam Khursheed
National University o f Singapore, 4 Engineering Drive 3, Singapore 117576. [email protected]
In principle, many analytical techniques can be combined with the normal operation o f the SEM,
but in practice, they are limited to a few routinely used add-on attachments [1], The main reason
for this lies in the way SEMs are currently designed. A t present, the SEM column typically
consists o f an electron gun, electromagnetic lenses, scan coils and apertures, see Fig. I. The
purpose o f this column is to produce a highly energetic (1- 30 keV) beam o f electrons that is
successively focused into a sub-micron probe and raster scanned over the sample’s surface. The
problem with this conventional SEM design is that most the electrons and photons scattered back
from the sample travel towards the column and are therefore difficult to detect and analyze.
Moreover, the distance between the lower pole-piece o f the objective lens and the surface o f the
sample, commonly called the working distance, is relatively small, typically restricted to be
between 3 to 25 mm. There is therefore little room in this design to mount detectors and
spectrometers that can efficiently collect the scattered electrons and photons generated by the
primary beam-specimen interaction or analyze their energy spectra with high resolution.
The SEIM proposal described here does not restrict detectors or spectrometers to be located in the
small space between a conventional SEM objective lens and the specimen. The new SEM design
layout is depicted in Fig. 2. The central concept behind the design is to lie a conventional SEM
column on its side, so that the primary beam initially travels in a horizontal direction and then to
turn it through 90 degrees by magnetic sector plates so that the primary beam strikes the
specimen in the normal vertical direction. The specimen in this case is located inside or just
above the objective lens, which is integrated with the specimen chamber. The scattered electrons
and photons are therefore not obstructed by the SEM column which lies in a horizontal direction
located well away from the sample. The scattered electrons and photons are directed on to a
hemispherical region above the specimen, unobstructed by the SEM column. An array of
electron and photon detectors/spectrometers can be mounted on this hemispherical collection
surface, essentially unrestricted by space constraints.
The new SEM design has several advantages over conventional designs. Firstly, it is predicted to
have high transport efficiency for all scattered electrons and photons, typically over 80%.
Secondly, the energies o f the scattered electrons can be analyzed with high precision, where the
energy resolution is expected to be typically better than 10'4. Thirdly, it provides for better
separation between different types o f scattered electrons. Fourthly, the new SEM design can
easily be extended to incorporate time multiplexed columns and multi-column arrays. The
redesigned SEM presented here is referred to as the "Spectroscopic SEM”, or simply “SPSEM”,
since it naturally combines SEM imaging information with energy spectroscopy.
'¿661 ‘V °À M3N ‘sssJd 3>I3
".ü°lJ0 O f Áq psjips ‘soijdo sp iy B J p sS je q j jo >|ooqpuBH ‘6 Jai<M 3 ‘JRSOj -j, iaj
jnoXB] iMHSdS -Z 3Jn3ij
jaqiueqo ipuipsds/sipi 9Aipafqo psiBJSsjuj
-----------------------------------• • •
■ •—H
___ *.
uuirqoo JAIHS
qgiq ajSuB-api^w
, qSiq 3|§UB-A\oajBfs[
I SU O Jpdp X§J3 U9
k''SUOjp3|3 X§J3U3
3DBjjns ..
puB iq§n
UOip3||OD [B0|J3qdsiUJ9H
sjsp u io jp sd s /sjo p sp Q Xbj- x pue iq § n ‘uojpajg jo Xauy
I/V3S I^uojiusAuo^ : [ dm2\ j
j3quiBq3 ipunoads
jo p s p p
SU9J 3AjP9fqO
ajnysdy |buij [J----S9SU9JJ9SU9pUO^
u n r) u o j p 3[3
POOZ sp u ^ l 1U3D9H
1. Konvalina and I. M ullerova
Institute o f Scientific Instruments, Academ y o f Sciences o f the Czech Republic,
Kralovopolska 147. 612 64 Brno, Czech Republic; e-mail: [email protected]
In order to collect the secondary electrons (SE), scanning electron microscopes (SEM)
are equipped w ith the Everhart-Thom ley (ET) type detector [1], The electrostatic field o f the
front grid, biased to a positive potential o f several hundred volts (see Fig. 1), is to attract all
SE o f kinetic energy below 50 eV or at least those from the SE spectrum peak at 1+3 eV.
However, the detection quantum efficiency (DQE) o f such detectors has been found to be
significantly lower than one [2], which is mainly given by their low collection efficiency. The
electrostatic field o f the grid cannot sufficiently penetrate toward the specimen and influence
the trajectories o f SE owing to grounded electrodes surrounding the specimen (specimen
alone and its holder, specim en stage, pole piece o f the objective lens, etc).
We calculated the collection efficiency (CE) o f the ET detector as a function o f working
distance (see Fig. 2), for the arrangem ent shown in Fig. 1. Software SIM ION ver. 7.0 [3] was
used for the simulations o f trajectories and equipotentials. An example o f the potential
distribution around the grid on a positive potential is shown in Fig. 3. The C E drastically
drops down to several per cent with small working distances necessary for SEM to operate in
a high-resolution mode. Results are in good agreem ent with the previous work [4] and also fit
well our measured data o f DQE [5], The w hole energy spectrum o f SE was taken into
account, too. but it does not influence the results too much in comparison with the case when
only one SE energy (E se = 5 eV) was used. M oreover, the influence o f the specimen diameter
on CE w as exam ined [6].
N ow we test the im pact o f param eters o f the grid [7] and scintillator on CE (see Fig. 4).
There is no grid in the example in Fig. 5 and the scintillator o f 10 mm in diam eter and biased
to +10 kV is simply im m ersed into the tube at a ground potential. Fig. 6 shows the field
penetration from the scintillator (again at +10 keV) through the grid at a grouqd potential.
Recently, new sophisticated configurations o f SE collection systems have appeared in
the commercial SEMs that transport particles by electrostatic and magnetic fields toward an
“in lens” detector w ith high collection efficiency. It is evident from the example shown in Fig.
7, that CE rem arkably increases for the upper “in-lens” detector (better signal to noise ratio o f
the image - SNR) but unfortunately the contrast o f subm icrom eter crystals disappears due to
influence o f a strong magnetic field at the specimen surface that extracts SE toward the upper
detector and suppresses shadowing phenomena. High topographical contrast rem ains for the
low er SE detector o f a configuration similar to that in Fig. 1 but SNR is, o f course, low.
Several other examples led us to study the collection efficiency o f the SE detectors in more
details. [ 10]
[8 ]
T.E. Everhart and R.F.M. Thom ley, J. Sci. Instrum. 37 (1960) 246.
D.C. Joy et al. Scanning 18 (1996) 533.
DA. Dahl, in: Proc. 43ri ASMS Conf. on Mass Spectrometry and Allied Topics. Atlanta (1996) 717.
M. Balasubramanyam et al, Nucl. Instrum, and Methods in Phys. Res. A 363 (1995) 270.
L. Frank et al, in: Proc. o f EMC 2004, Antwerp, Belgium, in print.
1. K onvalina and 1. M ullerova, Microsc. M icroanal. 9 (suppl. 3) (2003) 108.
F.H. Read et al, Nucl. Instrum, and Methods in Phys. Res. A 427 (1999) 363.
K. M atsuda et al, in: Proc. o f EMC 2004, Antwerp, Belgium, in press.
H. Kazumori, JEOL N ew s 37E (2002) 44.
The study is supported by FEI Czech Republic. Ltd.
IAI3S J 00Z.9 103f »Ml
u! [6] OH^u) 3SJ° Jop9pp jaddn Usu9j
UI„ UB Xq pUB (ÿ 3j) J0P 3J3P gS J9AV0I
pjBpuBjs 9qj Xq ua^B] ‘[8] j n n !S-§W
IV/e0 zlV 9M1u! spiyed DIUIBJ33 *¿ *§ij[
|BijU9jod punoj§ B
JB puS šiji qSnojqj A5! 01+ Ie JO^BJipuios
luui oi = dsp ‘urai Qz =PM ‘9^°! papunojS b ui (iqSiJ) uuu qzjo (yaj) oi
Xq p9SJ9uiuu ‘\y\ 01+ WJ01BI|iiutos 9qi joj sjBiiugjodmbg p9jB[noiB3 •§ '3i j
‘(A5! 01+ =3sn ‘A 00£+ =§n
= dsp ‘uiui oč = PA\ ‘A 9 Ç=3S3) P!j§ 3lPJ° S9Jim psdBqs arui
qiiM S9IJOP9ÍBJJ 3S p9jB[nuiis jo 9jpunq b jo m9IA q £ 'p -§ ij
A 00e+ =§n Pub uiui
01 = dsp ‘A 9Ç =^Sg JOJ PM SnSJ9A HD paiB ^D ^ 'Z SlJ
[uiui ] aotjwp 6uj)|JOM
(uiui e = PM ‘uuu 01 = dsP) A 00£+ spojpap ub Xq
paiBjnuns puS aqj joj sjBijuajodmba p9iB|no[B3 •£ Sij
•J9J9UIBip U9UlID9ds 91{J
UI p9Sn JuâtÜ9§UBJJB 9qi JO U0IP 9S-SS0J3 ’X 'S ij
0 o
s® o<tT
Pn) pus v
(0>n) J0(D||!*UPS \
jopepp lg
B. Lencová
Institute o f Scientific Instruments, AS CR. Královopolská 147, 612 64 Brno, Czech Republic,
e-mail: [email protected] isibm
1. Introductory remarks
In the tim e o f mainframe computers one o f the principal reasons for writing one’s own
programs w as the price o f the software and the fact that general-purpose software was not
often available. Homemade programs provided complete control over all input (data structure,
interface) and output and the im plem entation o f specific algorithms [1], Very often they were
ju st a product o f student4s work, which w as not completely debugged and in no way userfriendly; this w as not too serious, because the author was mostly the user o f the software.
Such programs were frequently short-lived, undocum ented and unsupported after the student
finished. This approach is certainly not optimal and it had to be changed. “M odem ”
com putations in electron optics were largely influenced by the first order Finite Element
Method (FEM ) software developed by Munro, for which the source code was published [2].
W ith the arrival o f PCs, much program m ing effort is being invested in interfaces, examples
being SIM ION based on the Finite D ifference M ethod (FDM) [3] and CPO programs based
on the Charge Density M ethod (CDM ) [4], The appearance o f interfaces changes as operating
systems develop, and backward compatibility is not completely guaranteed. The programming
languages also develop with time. The programming skills o f students are rather decreasing
because many o f the problem s that had to be solved before by programming can now be
analyzed with Maple, Matlab, or even Excel. In my opinion the popularity o f Linux among
physics students also channels their interest into a different direction. It is also difficult to
oblige students to contribute to the existing software: even if they spend most o f their time
sitting behind a PC, they are frequently unw illing or unable to do serious programming work.
For a new student the use o f Fortran can be an obstacle; in reality it is rather the lack o f
experience in any program m ing language and technology and often the lack o f “sitzfleisch”.
The size o f m ost existing packages is also too big to allow additions.
Electron optics (EO) as such is not o f great comm ercial interest and thus only limited funding
is available. A consequence o f this is that the existing software is not easy to combine with
programs like AutoCAD, Matlab, Maple, or Excel. Each EO program uses a slightly different
approach and terminology. It is also not simple to jum p from one"BO com putation package to
another. The EO program s are often not the primary activity o f their authors, who are mostly
busy with other tasks, prim arily teaching and research. Also not too many citations can be
made on software („program w riters“ tend not to cite anybody else’s work, commercial users
o f available software often do not cite in publications their tools). Writing software has a low
status because it is considered to be a routine task rather than scientific work.
Applications o f EO software, ranging from detector design via computations o f individual
lenses to system optimization, are ju st too numerous to be covered by a single software
package. A qualified and proper use o f programs for particle optics needs some previous
knowledge; in general this level o f knowledge and education in particle optics is on the
decrease. M ost specific applications that can be analyzed w ith specialized EO software are
not very transparent to students who prefer to use programs like SIMION, with which
reasonable-looking output can be produced in a short time [5, 6],
spoqjaui uouB[odjf)iui aqj jo Xub qjiM pasti aq ubo (so |odopo put; sajodnjxas ‘sajodtupEnb)
sju au iap ajodijinui jaqSiq jo s p p ij aqj ui uaAa ‘s p p ij uoijaayap puB Suisnaoj jo uoijBuiquioa
Xub ui uoijoui jo uoijBnba apijjBd aqj sa \|o s 3upB jj Xb'jj '3 0 d S J ° aSBspBd S Ä SV H 1/3H 1
ui pasn spoqjaui [-^l] d y z J0 [£ l] 33!IS 3lP
saqsaui iM gj ui uoijEjodiajui q j aqj
jpM puBjsjapun oj paau aqj si sapBjsqo jsaSSiq aifj j o au(j 'Xjojs jaqjouB jaX si SupBjj Xe»
'NOIJMIS SB M3ns ureiSojd E lajajd
oj uiaas oqA\ ‘sjuapnjs Xq paA paai XyBaijsEisnqjua Xjba jou bjb Xaqj os ‘s jp s s j jo Suisssaoid
auq-jjo amos puE suiBjgoid Q V 3 U0UIUI03 UBqj Supjuiqj jo sdXj ju sjsjjip XpqSijs e aiinbai
Xsqj M3 \ 3 \\o | i joojd-juspnjs XjjjBd are Xaqj os ‘o u jg puB jjp c j fix 1B sasm oa saijdo apijjBd
ui pasn ajaM surejSoid a q j sjjnsaj \B|dsip oj puB saijjadojd p a n d o jo uoijBjnduioa jo j
EjEp SuijBaja jo j aaBjiajui SOQ -taqiouB pire ‘(sjuiod 000003 W " az !s qsaui uo jiuiq ‘asnoui
jo asn ou ‘sajBuipjooa z puE j jo oijbj jaau o a ui jndjno ou ‘sauij qsaui uo pajiuii| si uiooz
‘sa|ijB pui SA\opui;Yv jo pBajsui sa|ij g o d H °lu ! sjjnsai puB XijauioaS jo sjndjno paiqdBiS
“X flM Il J0 SA\opu;/w jiq- 2£ ‘sjaAup saiqdBiS uiapoui qjiM XjijiqijEduioa moj ‘saiqdBiS mojs) jm ;ij jo j jbosbj puEjjog ui uajjiiM aoBjjajui SOO B s3sn saShapsd aqj jo qoBg 'ap j
qajBq SOQ B B!A pap au uo a X([Bnjnui ajB jBqj suoijEjnduioa jo j sajqEjnaaxa puB saaBjjajui jasn
JBJ3 A3S j o Suijsisuoa ‘ssSB^OBd aqi [je jo ainjam js piSu aqj ajB 3 0 d S J ° samjBaj a ja p s q o
juauiuioa siqj asn jou op sjasn bjbmjjos jsoui
tjndui BjBp aqj ui juauiuioa ajqBjins b SuiXjpads Xq aaBjjajui Jasn juajjna aqj uiqjiM pappB si
dooj Bjjxa aqx ’sjndjno j o Suissaaojd puB sjjnsaj aqj j o uoijBzijBnsiA XsBa sajqBua qaiqM ‘ajqBj
ajdujis e oj paanpaj si sjjnsaj jndjno j o junouiB aSjBj a q j sjjnsaj aqj jo jndjno g g 3 X g UB
puB ‘suaj jaqjo aqj uo bSejjoa aqj spug qaiqM ‘Suijooqs Xq saijjadojd jBaijdo jo uoijBjnduioa
aqj ui aSBjjOA aqj jaAO dooj BJjxa ub psppn 3ABq 3A\ siqj jo,[ 'S3SU3[ sqj jo suo jo sSbjjoa
aqj uo saijjadojd p aijd o aqj jo aouspusdsp aqj a\ou>( oj jubm ba\ ajaqM ‘uoijisod a§Euii
puB joafqo paxij qjiM rnajsXs aijBjsojjaaja sua[-OMj jo uoijBjnduioa b si ajduiExa u y ’saSE^aEd
ujEjSojd j o amjorujs pi3 u b Xq paxij XjSuiuiaas ajB Xaqj j i uaAa ‘suoijBjnduioa aqj ui sajnjBaj
M3U ppB OJ XsBa uaqj SI JI -JSJldUJOD J3JSBJ puE M3U B qjIM 3]ldui00 puB Xjipoui OJ XsB3 snqj
puB UEIJJOJ UI ]|B 3JB SUOIJBjnduiOO [[E JOJ S3]qBjn03X3 ‘psjusuinoop XjqEUOSBSJ SI 3JBMJJ0S aqj
‘paqaEaj sq usa s p p ij j o XaBjnaaB qSiq ‘XjuojaBjsijBS 5(joa\ sajnpooi aqj []B jBqj si 3 0 d S
jo ajnjBaj aAijisod ja q jo u y '[ 2 [] j|naijj¡p jou ajB japjo-qSiq b ajinb jo sa]odij]niu jo sjojaayap
apauSBUi puB aijBjsojjaap ajduns jo j sjuauoduioa p p ij ajodijjnj^ [ j j ] uinuiixBui pjaij
j o o/o ¡Q-Q uBqj jajjaq p]aij jbixb aqj jo XaBjnaaB sapiAOjd sjuiod qsauj 00008 tp!-" pajnduioa
sua] ounsx aqj ' 3 |diuEX3 j o j [oi ‘6] sju siu a p aujauiuiXs Xijbuoijbjoj jo uoijBjnduioa
p p ij jo j uoijBuiijsa XaBjnaaB aijBuiojnB sqj si 3 0 d S u î paanpojjui Xjjuaaaj ajnjBaj
snbiufi ' 3 J b uo puoDss i UBqj sssi ui auop aq a\ou ubo sjuiod 0009 J ° qssm b ui sua|
pajBjnjBS b jo [g] Ç86 I ui ounsx j o uoijBjnduioa jqSiiuaAO uř^jajsB j SuijjaS ajB sjajnduioa
asnBaaq Xiuibui si siqx 'ajqissod ajB sjuiod uoijiiui auo oj a so p sazis qsaui ‘spuoaas
0£ punoiB spaau sjuiod 000003
qsaui e ui uoijBjnduioa aqj XyBaidXj isaqsaui aSjB| Suisn
Xq paAOjduii si XatunaaB uoijBjnduioa aq x 'aiduiis puB jsbj si ji jEqj aSBjUBApE aqj SBq qaiqM
‘W g j Jsp-io jsjij aqj Xq paurejqo ajB sp p ig SupBjj Xbj puB saijjadojd p aijd o jo uoijBjnduioa
3M1 ‘0 3 ui uoijBjnduioa p p ij aqj jaAoa 3 0 d S HMM psAIos aq irea jBqj suiaiqojd aqx
•suiBjSojd aqj qjiM îjjom aqj
aAOjduii XpuBaijiuSis sjjnsaj j o sjndjno païqdBjS aqj puB ‘qsaui auy jo uoijBjauaS aijBuiojnB
‘qsaui jndui jo uoijBjauaS aqj jo j aoEjjajui jasn [BaiqdBi3 jo asn aqx '[/,] sjoqjnB jB dpuud
OMj aqj Xq Xpuanbasqns pus ‘y p Q n i 1SI JE padopAap XjjjBd ‘( 3 0 d S ) suoijBjnduio 3
saijd o apiJJB j jo j aiBMjjos payBa suiBjSojd jo jas jno Xyauq ssnasip mou sn jag
suoijnjndiiio 3 sa ijd o siaijjR j j o j 3 j b m } j o s •j
The next problem is to use efficiently the large amount o f output provided by ray tracing.
“Single ray tracing” o f up to 20 rays with many possible outputs o f trajectory and other data
after some given step has the disadvantage o f off-line graphics and large output files to
process; graphics output needs an external 2D program SGPlot for DOS or Windows, which
is no more supported. The “multi ray tracing” has as a result ju st the final positions available
only in one selected plane and some care is needed with the interpretation o f results. On the
other hand, the accuracy o f the ray tracing results can be very high, so that we can easily
derive from the final positions the geometrical aberration coefficients o f the 5th order and the
chromatic aberrations o f the 4lh rank [15, 16].
A drawback o f SPOC is that it is still incomplete. The programs for computations o f fields o f
misaligned lenses and deflectors, for com putations o f aberrations o f deflectors in electrostatic
lenses or for com putations o f combined electrostatic and magnetic lenses with or without
deflectors are still unfinished.
3. Electron Optical Design (EOD) program
A w indows version o f SPOC is badly needed, because “standard” user knows how to click
with a mouse and expects to find everything in Help file. Unfortunately even the W indows
them selves develop, e.g. a change from W indows 2000 to W indows XP for most programs
written for them w as not completely smooth. Compilers also change (we wanted to stay with
Visual Fortran, but our com piler changed the producer from M icrosoft via Digital and
Compaq to Intel); their graphical possibilities improve so that the same results as with more
popular C or C++ compilers can be obtained. Then there are common problems as with any
software - debugging and testing o f complex programs is tedious, proper documentation is
boring, sufficient num ber o f illustrative examples is a m ust if the programs are to be
What is EOD? A s announced in 2000 [17], it was then about to be released soon. The main
intention was to combine a new W indows interface with the FEM field computations as well
as with the com putation o f optical properties and ray tracing, i.e. virtually combining all what
is in the five SPOC packages in a single program written in Visual Fortran and using a lot o f
graphics. In reality, since then even more features were implemented, like the use o f dielectric
materials, N eum ann boundary condition, accuracy check, 3D display o f traced rays. The
critical part w as not the new interface; in fact it was written (by J. Zlámal) in quite a short
time at the beginning o f his PhD study in ISI (for the FEM programs in SPOC all four
packages have their separate interface programs although most o f the program features are the
same). With the new interface the number o f lines and points in the mesh for FEM
com putations is limited only by the allocated com puter memory.
Ray tracing has been programmed anew. The slice method used in accurate ray tracing by
TRASYS [13] was not implemented and it will probably even not be required. The ZRP
method w as newly program m ed and recently complemented with bi-cubic and bi-quintic 2D
interpolation. The variable-step 7th-8th order Runge-Kutta-Fehlberg method, in TRASYS
rigidly implem ented to comply with the required output step, can also be used in EOD as well
as 4lh-5th order variable-step Runge-Kutta procedures allowing easier interpolation o f results.
Unfortunately for the release o f EOD, attention w as then concentrated on successful
implementation o f space charge computations in ion and electron beams [18], Current effort is
directed back on finishing EOD with procedures to evaluate, from ray tracing results, paraxial
optical properties and aberrations as well as on im plem enting the “standard” procedure based
■sSuipaaoojd asaqj ‘p u re jz f p ro Á>[suEf ,| ‘BAoauaq g [jj []
1 0 1 ‘I '[°A 'o u jg ‘000c W 3 íin 3 30Jd ‘pui?IZ T puc EAoauaq g [¿ i ]
•sSuipaaaojd a s a q j' [bjq ‘[Al [91 ]
•55 ‘9661
‘..spuajj juaaay,, a o jj '|b iu r |7 j- pire BAoauaj g [g j]
£ £ l (£661) MOc 3Id S
3 T PUB >l!Pujq 3 T [H ]
(0661) 86 cV « y s ^ iy qjajAj jjsui p n N ‘BAoauai g pire ^ipssiyW o ‘q y e g ’3 T [£l]
'60l7f (Z.661) 89 'um jjsuj p s Aatf ‘?A oauai g p ro 31193 -jaj [ 3 1 ]
> 002 ‘s-mipojq > ¡3 §V ISI ‘?A oauai g f x 1]
6PI (VOOZ) 6 ISV sa>f sÁM<[ ui qpjAj jjsui 'P ^ N ‘?A oauai g |()(|
£9Z (ZOOS) £6 -JoiuiBajin ‘BAoauai g
[6 ]
\Z \ ‘5861 ‘8 ¿ °N ’» S J uo3 '^M d JSUI ‘q P S V 3 "S P™ «unsx
'9S (0 6 6 1) 86 cV s s y 's^Md u! q p ^ jjsui '[oiin " lu ip ss i,^ -9 pire BAOOiraq g
•sSuipaaaojd asoqi ‘BAOJSnn^ 'i puB emi[bauo-)| ] [ 9 ]
•s8uip3330jd 0S3l[l ‘BJBJjny y pUB [OjpUR/W ,|
•iuoa’saijdouojjaap-A\A\M//:djjq 33S ‘uiBjSoid s a s - 0 d 3 'PB3» H 3
L\L ‘S6 6 I eju e[jv ‘u jaad s ssbjaj ju o o SPMSV pj£t> 3°Jd 'j i p a V 'd
SZ.61 sS puqureo n ‘u o d a y í f U I 3 3 1 3 /0 3 0 3 ‘cu im ^ 3
¿68 ‘9861 ‘III u ed ‘98 IAI3S ‘3u a i ^ pus BAoauag g
•sjndjno j o Sm ssaaojd aqj j o amos j b a o 35(bj
° l 3 3 D X 3 3>|i[ s[ooj uM oiq-ipM Joqjo a j q i s s o d s b q a n u i s b a s n oj asuas saíjBiu os|B |[ qjBd
jqSrejjs aqj u i o j j auiij oj auiij u i o j j payaAip u i e S b ajB sm j i uaAa ‘SA\opu[\\ Suisn j o spuajj aqj
qjiA\ d n qajBa oj S u i A j j o s j e aje s m X ( s n o T A q o }i SuiAOjdun j o 3 j b m j j o s aqj ui sdBS aqj Suqjij
jb p a u i r e sjaafojd juajjna uopuBqB oj p a S i |q o b j b aA \ auiij oj auip u i o j j pus ‘uiaqj u i o j j sjp sa j
jajjaq ja8 oj ‘suiBjgojd gupsixa aA0jdui| 01 sn aajoj suoijBaqddE puE suiajqojd Maj^[ ’ssaaojd
í ? u i p u a - j a A a u e s i j § j u i s a i j d o u o j p a j a j o j a j B .w j jo s j o
j u a u i d o p A a p a q j ‘a A o q B u w o q s s y
su o isn p u o j
POOZ SPU'}JI iU333}¡
u o ijE jja q E a q j
iu o jj
s j u a p i j j a o a u o i j E j j a q B S u i j j a S p u B u o p E n b a /( j o j a a f B j j p i x E J B d S u i a j o s
B. Lencová, P. Jánský and J. Zlámal*
Institute o f Scientific Instruments, AS CR, Královopolská 147, 612 64 Brno, Czech Republic.
E-mail: [email protected] isibm
*Institute o f Physical Engineering, Faculty o f M echanical Engineering, TU Brno,
Technická 2, 616 96 Brno, Czech Republic. E-mail: zlam [email protected]
W hat is EOD? The program Electron Optical Design contains interface in W indows, all
program s for 2D com putations by the first order FEM, ray tracing in computed fields,
simulation o f electron and ion sources, and com putation o f space charge limited beams - for
further details see [1], M ost parts o f the software are in the beta developm ent phase, and they
are already used in a num ber o f projects. For example, in order to simulate properly the
behavior o f ion optical systems for ion beam deposition at TU Brno, the ion source
com putations were started by including the plasm a creation inside the ion gun [2]. We can
then start the com putations on the plasm a meniscus and take into account the emission
properties o f the gun as well as space charge effects in ion beams transported in the system.
Intensive boron beam transport was studied with EOD as a part o f a diplom a project in
cooperation w ith Salford University; the space-charge limitations were quite severe [3],
Ion sources and beam transport simulations are more critical on the proper evaluation o f the
space charge than electron guns and transport o f beams we w ant to analyze for the electron
beam welding applications [4], because the electrons are lighter particles and they are soon
accelerated to a sufficiently high energy. First, in order to test the performance o f EOD, we
have checked our results on the com putation o f a Pierce gun used as a test 20 o f CPO-2DS
program [5]. Figure 1 shows the geometry o f a simplified gun; proper voltages at selected
points are introduced with auxiliary electrodes and the N eum ann boundary condition is used
on the outer radial boundary. The left part shows the beam in the first step o f the iterative
procedure; the current in the beam and the emitting area are adjusted automatically to provide
vanishing field on the em itting surface. The final result provides the beam shown on the righthand side o f figure 1. The current transported in the gun is 9.1 »A at 10 V on anode, 98 % o f
the expected value [5]. In realistic electron guns we have to pay attention to the mesh used in
the immediate vicinity o f the emission surface in order to get accurate results. The geometry
o f the problem can be approxim ated very accurately as the num ber o f coarse mesh lines
defining the problem can be very large. There is more freedom now in the selection o f the fine
mesh. The mesh size is limited only by the amount o f mem ory that can be allocated to the
com putations and eventually by round-off errors. The graphical outputs o f EOD are also
improved: figure 2 shows a cut-away 3D display o f the gun geometry. The gun analyzed [4] is
intended for 60 keV beams with m aximum pow er up to 2 kW.
The EOD program is planned to replace soon the SPOC packages f 1]. Before that, still more
testing with all relevant exam ples studied before has to be done, Help must be further
im proved, and the evaluation o f the optical properties finished.
B. Lencová, these proceedings.
J. Zlámal. PhD Thesis, TU Brno 2003
P. Jánský, diplom a thesis, TU Brno 2002.
P. Jánský, B. Lencová. and J. Zlámal, M icroscopy and M icroanalysis 9 [Sup. 3], p. 2223 (Proc. MC2003, Dresden).
CPO-2DS program, see http://w ww .
'(/A 5! Z -isMod
09 ) Suip|3M u re sq -g jo j unS u o j p a p aqj j o ÁBjdsip q £ ÁEME-inj : j s jn S ij
•jqSu sq j je j|n s3 j iira jsisu o o -jp s ¡buij ‘jjsj
sq j ]E S3UOJ33ÍBJJ u o j p s p j o p s |Eiiiui - unS 30J3TJ [EoupuqÁo aqj j o suoi)B )ndiuo3 : [ S jnS lJ
G. Lilienkamp
Technical University o f Clausthal, Institute o f Physics and Physics Technologies,
Leibnizstrasse 4, D-38678 Clausthal-Zellerfeld. Gerhard.Lilienkam [email protected]
W ithin the last tw enty years Low Energy Electron M icroscopy (LEEM ) and Photo Electron
Emission M icroscopy (PEEM ) have been applied to numerous experiments in basic physics
and chemistry. Especially LEEM and PEEM have brought new insights in fields like metal
and sem iconductor epitaxy, catalytic reactions at solid surfaces and magnetic properties o f
thin films. Due to the parallel image acquisition LEEM and PEEM have an outstanding
performance in studying dynamic processes. The wide range o f accessible contrast
mechanisms allows the exam ination o f the surface topography on a mesoscopic scale as well
as the determ ination o f details on the atomic level such as m onoatomic steps and the local
periodicity (by n-LEED ). The spectroscopic version o f the LEEM has already shown its value
in NEXAFS studies or im aging o f characteristic photoelectrons excited by synchrotron
radiation. Parallel im aging A uger electron microscopy using the LEEM gun as primary
electron source is also possible. However, there is still a lack o f chemical information without
an access to synchrotron radiation.
B inding E n e rg y /e V
B inding E n e rg y /e V
Binding E n e rg y /e V
Fig. I : a) PEEM image showing the segregation o f SrO islands on SrTiO}. b) MIEEM image
o f a segregated island. Field o f view 50 pm. Binding energy 13.8 eV. c) Spectra obtained
from a MIEEM image series fo r the regions (a), (b) and (c)(see text).
In the present w ork we have used metastable (3S) He*-atom s to excite surface valence band
electrons and thereby achieve a spectroscopic contrast. We call this surface probe Metastable
Im pact Electron Em ission M icroscopy (M IEEM ). Fig. 1. shows a first example: images and
spectra taken from a La doped SrTiO? (001) single crystal after heat treatm ent in ambient air
at 1300°C for 120h. In the PEEM image islands o f different size and contrast are clearly
visible. An Auger analysis shows that these islands mainly consist o f Strontium Oxide.
A series o f M IEEM images (one o f these images is shown in F ig.l b)) gives additional
S d 3 3 d [qiqiui y ‘Rsjoqons A ‘«Miuno S ‘PAV ueH ‘d u n q u aim O ‘qoeq-im H
[ I]
8ZZ-1ZZ (čOOt) 8 ‘saiu rejso o jp aia jo jBuinof je p p/y\ u b h
9£I-Č£I (£00č) SESEES a o u sp g ao ejm s [B p duiB ^usqn 'O
3]qiss333B Á[snoauBj|nuiis are ja p jo oiuioje sq j puB 3|bos-uiu b uo s jre p p |E3iqdBj3odoj 'B pp
aidooscupods suBaui jBqx '3 ui3buh p p i j >|jBp puB 0 3 3 1 - ^ ^q js p jo oiuiojb aqi j o u o ijn q u jsip
[B30[ sq j inoqB u o ijb u u o ju i SASiqoE oj 3[qissod os[B si ji JAJ331 qi!M lBqi lno sjuiod £ 'S i j
£Xg 3lj¡ Suipsiß u SI 03.1D Jl/Suq 31/1
'iuD3q (z /l-'O ) 3lt> lIH‘n l n R (R u! SD suws (3
'3MPHUIS %lUZ(ZXl)d 3lf¡ (o U01SU3JX3
3i\i si33[f3.i adujs ¡ifSuq 3i¡¡ ‘A 36Z=3 l°ds
(Z/I~‘Z/I~) 3W %u!sn ‘sSvwi p p i f ¡¡.mp (q
A 36P=3 '^USU3p-y¡ ¡us.iaffip if/iM suMUvd
Q337-ri(D JUOUf U011DV3A 31/1 fo ¡U31U3X01U
31/1 fo uoipsuip 31/1 ui Sutjuiod SI MOJAD
(OlOHX/X uo uo!1»P!xO-:H
siinuit ¡a¡ 3 3 j (3 ■
A 9= AIS (P
‘A £= AIS (■? A t i fo (AIS) sSvnoA
-1X Y IS v >°
(R u o im p x s
(S(Z)*3H ^R psumjqo sjnf.ins psonpsA
-usSo.ipXi/ pun pszipixo sqi /o oufoads uoissuus
lio UOllOpiXQ-^H :Z'% !3
UO.IJ33/7J (17 ( Q ll) t f íírs ¡
:£'s U
p8onp0j (OUU)Hid/X »
paz!P!xo(on.)L|y/X ■
■[j] u o ip B ai sq j 3 u u n p u o p n q u jsip a i um issBiod sqj
j o 3 ui3euii os[b sa\o[|b p u s sjq isso d jusuiuSissb jE a p b ss>[bui J/V33I1AI snonS iqureun js3 iio |
ou si 33Bjans aqi j o sS b jsao o puB jsBjjuoD u s s M p q u o ijE p j aqj uinissBjod j o 33U3S3id aqj qjiM
jnq iA133d Áq psSBuii sq osjb ireo u o ip B ai siqj g jd p u u d u[ 'su o ip B aj opXjBjBO p sjo m o jd i|b>(ib
j o Xpnjs aqj si ju s u iu s d x a siqj j o uoijb aijo u i g q x 'soBjjns (O Il)q'H psJSAOosjd uimssBjod b
uo uoijBpixo usSoapXq 3 HÁ|BJB3 sq j ÄpiUBU IAI33IIA1 J ° uoijBoqddB puosas B SMOqs £ '§ i j
■uoubSijssaui ja S n y snoiAsad b Xq p3A3iq3B sq jou p p o o sjjnssj
nsoqj •[[] lo f!.I J0 “-O !! SB q 3ns sssBqd q o u -} l jo suoz jsp jo q sp¡A\ uirl £-¿ b Áq pspunouns
3JB Áaqj jBqi puB ío^S ^q paiBUiaüO) sjb spuB[si o j S 3qi lBMl sjb sasj sp sfq o umou^
uiojj B jpads qiiA\ uosuBduioo y isa ja ju i j o suoiSaj ju sjsjjip ui Ájisuaiui sqj Suiibj33jui Áq
sausnbss sSboii aqj uiojj pspB ijxs sísav B jpsds sq x '(o [ '3 y ui uass sq ubo se uopEouojui
POOZ SP“3.<X 1U333)1
F. M ika and L. Frank
Institute o f Scientific Instrum ents ASCR, Královopolská 147, 612 64 Brno, Czech Republic.
fum [email protected] isibm
Functional details o f sem iconductor structures keep decreasing in size, now to the low
order o f 10 2 nm. Among the structure elements the locally doped patterns play crucial role so
that tools are needed for their observation. For fast diagnosis and quality check o f the
sem iconductor structures the scanning electron microscope is useful because o f its wide range
o f magnification, availability o f different signal modes, speed o f data acquisition and non­
destructive nature o f the technique in general, especially at low voltage operations. The
dopant concentration in sem iconductor is quantitatively determined via acquisition o f signal
o f the secondary electron (SE) emission [1] in such a w ay that the image contrast is measured
between areas o f different type or rate o f doping.
The contrast mechanism was assigned to presence o f external patch fields above the
specimen surface [2], which balance the local differences in the inner potential connected
with the doping. The new theory relies upon subsurface fields formed between semiconductor
and the carbon contam ination layer, w hich behaves similarly to a metal and grows on the
specimen surface owing to electron impact under standard vacuum conditions [3,4],
Evidences have been collected about crucial influence o f surface overlayers on the contrast,
however suspicion has arisen that the contrast formed in this way is not inherent to all SE
emission but exhibits some angular sensitivity. A part o f the study reported here was oriented
to determ ining the relation betw een the p/n contrast and the polar angle o f signal emission.
By means o f the standard side-attached Everhart-Thomley detector, the range o f
acceptance in polar angles can be varied by changing the working distance - this was checked
by 3D simulation o f the signal electron trajectories by m eans o f the SIM ION 7.0 program
package (see Fig. 1A,B). Evidently, at larger working distances smaller polar angles (taken
from the surface norm al) are accepted. Experim ents were made on boron-diffused p+ type
patterns into phosphor doped n type Si (111) substrate with the surface passivated by etching
in HF. The device was CFE SEM JEOL 6700F at the primary energy 1 keV. Micrographs in
Fig 2 show a strong dependence o f the contrast on the working distance, namely its decrease
toward longer working distances, inversion and increase to high levels with the p type
brighter, as anticipated on the basis o f previous results [4].
Further confirmation was sought by repeating the experiment in other two SEMs with
different working distances and detector geometries (Tescan 5130 Vega and Tesla BS 343)
and the contrast w as measured in dependence o f the electron energy. Fig. 3 shows again much
higher contrast at a long working distance. H igh p/n contrast w ith p type brighter is obviously
concentrated to low polar angles o f the SE emission, w hich is in accordance to the model
relying upon subsurface fields streaming the slow electrons toward surface.
[1J Perovic, D., D. et al. Ultramicroscopy. 1995, vol. 58, p. 104-113.
[2] Sealy, C., P. et al. Journal o f Electron Microscopy 2000, vol. 49. p. 311-321.
[3] El-Gomati, M. M., et al. IEEE Transactions: Electron Devices. 2004, vol. 51, p.288-292.
[4] Mtillerová, 1., El-Gomati, M .,M ., Frank, L. Ultramicroscopy. 2002, vol. 93, p. 223-243.
[5] Support o f GA CR under the Programme for Support o f the Targeted Research and
Development, no. S2065301, is gratefully acknowledged.
SaO U EJSip 8 u i ) |I 0 M J U 3 J 3 J J ip OMJ J 0 J JSE JJU 03
• ( m u í 9 j p ir e g - x o i d d e )
3 l]J J O 9 0 U 3 p U S d 3 p Á S j3 U 3 3 1 (1
'■£ OLÍ
■UIUI 6 6 Č P U E / / 8 I
‘v n ‘£'8 (IP I sqj m oij) S3 DUBJSIP 3ui>¡ioa\ aqi io j \sy[ [ je
dOOZ.9 lO H f u! Jo p a p p
1 3 „J3M.OI,, sqj ipiM psSBUii ‘sjBijsqns ( i l l ) ÎS 3C^ J u °JU! uoisnjjip ucuoq pauiajjB j : j r j j j
• s a o u B is ip S u i í j j o m
ju s is jj ip OMi
io j
o ‘¿ N O I F M S q i ! M s a u o p a f B J j u o j p a p
j b u S is
3 ip j o u o ip j n u i i s Q £ : [
q u
£ O íd
[A3] A6 jd u a rneaq A jeuid
luiug l-
Z O ld
M artin Oral
Institute o f Scientific Instruments, AS CR Brno, Czech Republic
Today, when computers become faster, it is possible to simulate the behavior o f a great
number o f particles in a beam. Direct ray tracing through an optical system is very accurate,
but still slow for some simulations, even with up to 100 particles calculated in a second. The
calculation o f particle positions in a spot profile with the use o f aberrations gives information
about a few m illions o f particles per seconds. It is still not feasible to deal w ith every single
particle in a real beam, but the available computing power allows us to analyze an ensemble
adequately representing all particles in the beam. Then we can calculate a current density
profile in a beam with a high accuracy. For this purpose, a program was developed.
A simple algorithm o f the profile computation was developed and tested. The method
assumes a general particle source o f finite size and energy width and an optical system with
deflectors. The profile is recorded in a given plane perpendicular to the optical axis.
Real sources em it particles from a finite area with a given density in coordinates, slopes
and energy. The m ethod assumes the source to be defined in a plane at za with coordinates x0
and y„ and slopes x ’a and y ’0 with energy E (figure 1). In general, the density distribution
function o f the source is therefore a function o f all those coordinates, slopes and energy. The
program assum es that the density function is separable, i.e. a product o f three functions o f
only coordinates, only slopes and only energy. The considered source is rotationally
symmetrical. In addition, the Gaussian distribution g 0(t) = (1 1^¡27iG )e x p [ - / 2 / ( l a 1)] with
a = FW H M / V81n2 is used for coordinates and energy, the distribution in slopes is expected
to be uniform w ithin a given aperture angle. The overall density distribution function o f the
source is (w ritten w ithout the respective FW HMs)
g (x 0, y 0,x 'a , y \ ,E ) = g 0( J x J V y J ) - [ ( z a - z 0)l(nR a2) ] • # „ ( £
where Eo is the mean energy o f the beam and the term in the square brackets is the reciprocal
solid angle, to w hich particles are emitted, with the assumption Ra « R a. It is the
distribution in slopes, za is the position o f a (virtual) aperture, Ra its radius.
The optical system is described by transfer matrix and aberration coefficients
(calculated i.e. w ith [6]), geometrical aberrations o f the third order and chrom atic aberrations
o f the first order are assumed [1-5]. The positions in the desired target plane are calculated as
a paraxial position plus the aberrations. The aberration formulas used in the program can be
found in [2] and in manuals [3]. For systems w ith small starting positions x„ and y„ the
computation o f off-axial aberrations can be switched o ff to speed up the simulation slightly.
The profile computation begins with the division o f the planes in the source, in the
aperture and in the target into small regions (figure 1); the energy range is subdivided into
sufficiently small intervals. Then from a center o f a square on the source is started a test
particle, whose direction is chosen to go through centers o f all o f the squares on the aperture.
The position o f the particle at the target is calculated with the aberration expression and its
current is added to an appropriate square on the target. The initial energies o f the particle are
set to mean energies o f single energy subintervals. The energy division is made so that each
i ut+l g (E - E„) d E = const. The current
contribution A I o f that test particle corresponds to current contributions o f all the particles that
would travel from the same source square, through the same aperture square with energies
Z00Z o ujg ‘IBnuEui J9sn o¿ £ sXsejx 'BAOouoq g
Z00Z OUJ8 ‘spnireui josn
0¿ f uoisjoa ‘(iq u io j ‘d o ij "dojdg ‘c n g ) uoijEjnduioo soqdo opijjEd jo j suibjSojj :baoou3J g
¿-£l5č-£6^8-0 N aS I '¿661 '0 1 1 ssoJd O JO ‘sondo opiyBd p o áreq o jo 5)ooqpuBH :p P j j o p o T
•¿O I¿90c 'MS'M joafojd oqi Xq pojJoddnS
L - í l íc-e6t^8-0 MaSI '¿661 'ssoJd 3HD ‘sopdo a p ijie j p o S m p j o ^ooqpuBH ■W PJJOFO T
(£006) LZ-9Z 'S - [£ ’dns] 6 sisXibubojoijai
pire Xdoosojoij/V 'Suioejx Xb>[ Xq sju o p ijjo o j uoijEiioqy jo uopEinop;} iBAOouoq g J e j o 'W
(W>61) £ ° N '¿6 W dO 'sosuoj oijEjsojioop jo soiyodojd jo uoiiBjnduio^ :ouaq -jm -BAOouoq g
9£c-clc (c86l) £ °M '19 >l!ldo :AI W ¿ ‘w \ l \
(3861) c'O N 'IfW lO III WBd'91-1 (2861) I 'ON '19 WdO HUBd '06£-l¿£ (1861) P °N '09
>|UdQ 1 IJE j ' suiojsXs XqdBiSoijjfi uiBoq u o j p o p j o
sis Xjeue
p o iio u m ^ : n q j
) ’| j ‘ojunj/\| r|
M oqs sjBq [3ao[-Xbj3 o q x 'iurl ç ç ’o t Kq pojjiqs sjb sjods pssnoojop oqj j o sjojuoo ‘uirl
si ooirejsip u o ip o y a p s q x pBJUi i si sjSire ojnyadB aqx ’([¿] anjBA [BoidÁj e)
si (sojBuipjooo ui jajhA V í) oo-inos uoi oqj j o azis [BnjjiA oqj p u s 'A s m ç ’o si lAJHAVd
'A 3 1 °1 13S PUB pozqB uuou si uresq oqj j o X§jouo uboui o q x oDUBjsip Sui>[joa\ uiui
u o ip o y o p o[qnop q;iM urajsÁs suoj-oavj oijB jsojjosp ub ui ss jijo jd ureoq poiB|no[B 3
auB|daßeiu! UBjssneo
uiu q ç
o 1 puB
[mrl] x
uni o í - snooj.3Q
qjBd a p iy B d jsa; b qjiM poqjoui aqj j o ujbjS eip oijBuraqos : i a in S ij
[g] " i ojnS y; u i OJB p o q ;o u i o q i qjiAV p o jn d u io o s o jijo jd X jisuop juo-uno j o s q n s a ^ u o ijB jn d u io o
o q j u i poaopisuoo 3JB nx puB ° y upB j o qi uiqjiA \ ssjBnbs A p io 'siBAJOjuiqns X Siouo puB sojBnbs
j o su o ijB u iq u io o o jq is s o d o q ; []B j o j s u o p u o o q SBq s jn p o o o jd s iq i jo ]jb p o q s iu y si s p j o j d o q x
( I o jn S ij o qj u i piuiBjX d a q i) 3[§ub p q o s Ä JB juoiuap o q j si ,a'v
",xy p n p o j d s q x 'ssijijU B nb
§ u ip u o d s o jjo o j o u o isiA ip o q j u i sd o js o jb S B jpp o q x i o S j b i oqj jb uiBoq o qj j o ju o jjn o [Bjoj
oq j s¡ / ojoqA\ ‘ 3\7 " / ’V " ,w " AV"-W(3 ‘ "Â ‘ " ,x ‘
‘ " X )% I = 7V os ‘lEAJOjuiqns su re s o qi u io jj
M. Osterberg and A. Khursheed
National University o f Singapore. 4 Engineering Drive 3, Singapore 117576. sg
This paper presents a magnetic spectrometer/lens unit intended to acquire both image as well
as spectroscopy information inside a SEM. The unit is small enough so that it can be placed
inside the specimen chamber as an add-on attachment and comprises o f deflection plates that
function together with a low aberration add-on miniature perm anent magnet im mersion lens
Although various spectrometers have been previously proposed for the SEM in the context o f
techniques such as voltage contrast, auger analysis and depth profiling, they do not presently
exist as SEM add-on attachments, nor are they capable o f capturing the full energy spectrum,
from the low energy secondary electrons to the elastic backscattered electrons. N ot only does
the attachment proposed here aim to acquire high resolution spectral information over the
entire energy range, but it also aims to capture high resolution topographic images, typically
in the nanometer resolution range.
The attachm ent proposed in this paper works by providing a stigmatic-free 90 degrees
deflection o f the primary beam in the SEM. In this respect, it resembles the LEEM setup [2].
The deflector unit works much like a round lens and can give magnification or
dem agnification o f the probe. N orm ally one thinks o f magnetic deflection as only being able
to focus an electron beam in a plane parallel to the deflector plates, called in-plane focus.
However, by making use o f the fringe fields that are always present at the edges o f magnetic
sector plates, one can obtain out-of-plane focusing o f the beam if it i§ inclined nonperpendicularly to the plate edge. The idea is then to put the perm anent magnet objective lens
on its side so that the deflected primary beam enters it horizontally, striking the specimen
which is located inside it. The add-on lens can provide demagnification o f greater than 10.
If the specimen is negatively biased to a value close to that o f the beam energy (i.e. creating
a retarding electric field) electrons emitted from the specimen will be accelerated back
towards the deflector unit. The strong fields inside the lens will collimate the scattered
electrons so that most o f them will enter the deflector unit. On traversing the magnetic sector
fields, the scattered electron will be bent in a direction away from the incoming primary beam.
An auxiliary pair o f magnetic sector plates then turn the scattered electrons in a preferred
direction, typically towards the scintillator detector in the SEM. If now the strength o f this
auxiliary sector plates is ramped and a slit is put between the deflection unit and detector, it is
then possible to acquire the energy spectrum o f the scattered electrons.
A test experim ent is performed in a conventional SEM to investigate the add-on spectrometer
concept outlined above. The setup is shown schematically in Figure 1. The deflector unit is
controlled by four electrom agnetic sector plate pairs. Focusing conditions and the direction o f
the outgoing beam is determ ined by individual currents in each o f the electromagnetic coil
¿611-681 l d ‘(8661) ‘S '» 31 A3>I 'J-mS ‘'F l 3 duiojg iaj y
0 1 c > 0 c d (100c) ÍT ‘Suiuubds ‘qo>| H •§ pire -qBiddruE>[ N ‘paaqsjnqx -y
[ Z]
S0 3 U 0 J 0 j a } |
■ a a u a ja ju o a a q j je p a ju a s a jd a q
jA jg s
JB U 0U U 3A U 03 B a p iS U l S u i J B j a d o l|U tl s u a j / j a j a u i o j j a a d s J ^ g S U O -p p B a q j UIOJJ S JJtlS aj JEIJIUJ
UJBSq [3|[EJBd B UI 3uiJ|nsaJ ‘SU31 jojaafojd B SE p3JBJ3do U33q SBq JI JBJ OS 'paiJBA aq UE3
jiun uoijaayap aq; qSno.iqj SuipABJj JajJB Cibui ud aqj jo uoijisod [Baoj aqj ‘sajBjd uotjaayap
aqj ui sqjSuajjs pjaij aijauSBui aqj SuijsnfpB auiij aures aqj je puB jiun jojaayap aqj aAoqB
uoijisod j 3ao-ssojo aqj Suia’jba Ay a§Bim aajj aijsiuSijs b sbaiS z jo p as oj pareduioa £ pus [
sjopns ui jaSuojjs sauiij aajqj AjqSnoj pjaij oipuSbui b jBqj punoj si j¡ ASjaua ureaq AjBiuud
y\a>l 9 qjiM pajEjado si j/\[gs un8 uajsSunj 009SNSÍ lO H f B ‘sjuauiuadxa jbijiui aqj u¡
•sajBjd ÁJEq|xnE (p puB sajBjd u o ija a y ap -jso d (c ‘sajB]d u o ip a |ja p
u ieui (z ‘sa)E[d u o ija a g a p -a jd ([ -dirías jiun su a j/ja ja u io jjaa d s aqj j o aijB uiaqas :i ajn3; j
uicaq ,v.iiíuii.i(j
uirujaads uojjaap
pajajjBas aqj ajinbaB oj pasn si (uA\oqs jou) j o p a p p jaq d ij[n u io jo q d /jo jB |i|ju p s JAI3S Ibuuou
aqj puB a u n j ui paduiE j si p sajBjd j o qjS uajjs p p i j a ip u S s iu a q g 'z J o p a s ui jBqj oj £ j o p a s
ui qjSuajjs p p i j aijauSBui Xjjbjiiuis puB z J o p a s u; jBqj oj [ j o p a s ui qj 8 u a jjs pjaij aijauSEiu
aqj j o oijbj aqj Áq p a u iu n a p p ájuibiu si puEq ja q jo aqj u o snooj a u B |d -jo -jn o '£ pire z
‘ [ sjo jaa s ui s p p i j aijauSBui j o u o ija a g a p pauiq u io a aqj u io jj sauioa Suisnaoj auBjd-uj sjiBd
Michael Rauseher and Erich Plies
Institute o f Applied Physics, University o f Tuebingen, A uf der Morgenstelle 10,
D-72076 Tuebingen, email: michael.rauseher!®,
Low energy (< 3 keV I focussed ion beam milling can provide a useful tool for the preparation
of cross-sectional samples for transmission electron microscopy, for example for in-situ
reworking of am orphisized specimens in the TEM [1 ]. Due to the increasing influence of the
chrom atic aberration spot size contribution in low energy operation some effort has to be
made in order to keep the total spot size sufficiently small and the corresponding spot current
density sufficiently large, respectively.
The design and favourable mode of operation o f a dedicated low energy focussed ion
beam system based on immersion optics is presented. The beam is accelerated using a gun
lens and the intermediate lens space is on high potential [2J. Beam retarding is accomplished
within the objective lens. W ith this design the target is field free and can be kept on ground
potential. This indeed is very sim ilar to the basic operation principle of many modern
scanning electron microscopes [3]. As all deployed voltage-levels are comparatively low,
both condenser as well as objective lens can easily be operated in the internal acceleration
mode, consequently improving their general optical performance [4], In order to keep the
working distance to a minimum and avoiding the technical difficulties of dynamic
electrostatic deflection on high potential, beam scanning is realised using a split final
objective lens electrode. Figure 1 shows an outline o f the system design together with a sketch
of the potential distribution.
Figure 2 shows the system performance in terms o f spot size d vs. beam current I for a
beam energy of 3 keV and an immersion ratio (ratio of booster potential to final beam
potential) of <t>booster/‘i >i = 8/3. Both, the results o f third order optical analysis (using the RPSalgorithm |5 |), as well as data from direct ray-tracing (obtained using the IMAGE software
package [6 ]) with and without Coulom b interactions are shown. As expected, the influence of
interactions is a major limiting factor for the achievable minimum spot size. Further analysis
of this effect will be presented.
‘uiu o ç n /is a o jn o s
iu o ijeju sie d
/ ( 8 j3 U 3
a ip
jo j
ç qjpiM A Sjsus ||n j 'j s / y r l 01 ^ J is u s ju i jtqnS uE
psuinssB i:ji;p aarn o ij u u u ç [ j o 33UEjsip 3up[j0M
sqj jo j ] ju sjjn o uresq ’sa p azts jo d s j o s u u a j ui aouEuuoj.iod uiojsÂg
o jn 3 y
[vu] uirat|i
[UJU] p
•<J) |EIJU3J0d
[EIXE pniE[3J-30jnOS pUE IU3JSÂS |E3Ijdo U0ISJ3UIUII JO SuiMEjp 3I)EUI3q0g :[ n jtlS y
sepojpe|06u!uubds J
Lueeq P8IBJ601UI i|i!Msu0| 8A!P0fqo |
sapojp0|0 iu0luu6üb I
0 60}U| i ^ m
LUE0q P ; b j
0 q n j J0U||
© o jn o s uo! |Ei LU pm bii
IOI |ej}U0ioa p0}B|0j-0Qjnos
p n s a a w ¿ q saSEîpEd g o V W I P™ S D IId O :suoijEinoiE 3 j o j pssn s u ib jS o jj [9 ]
(9661) l o n o i w do
d ‘m-œg a r [ç]
(£661) £9 ‘3S XdossojoîiuEjjin ‘p s s u Q M M
(£661 ) I£ ‘£9£ V S3d s^ d
M13W 'flsui p n ^ ‘p q s iq a s g pus ‘s a q j g ->p3g s [£]
■(17661) W Z ‘0171 qjMOJQ [ejsâjo 7 ‘nzsuoA H 'B iutqso N ‘MOI^S I >l'-d "X \Z\
00 0 Z ‘u a S u iq a n x j o ÂiisjaAïuf) ‘u o ije jjs s s iq ‘îpnEUQ >| [ [ |
Frank H. Read
Department o f Physics and Astronomy, University o f Manchester, Manchester M13 9PL, UK
e-mail address:
In the present study we consider the limitations to the ultimate accuracy o f the BEM
(Boundary Element Method) for 3D electrostatic systems, with and without space-charge. It
seems useful to do this because the BEM (also known as the Integral Equation, Surface Charge or
Charge Density Method) is known to be usually faster and more accurate for such systems than
other methods. Some evidence for this is provided by a comparison o f 8 different benchmark
tests carried out in 1997 and 1999 [1,2] when it was shown that for computing times of the order
o f 1000s on a 100MHz Pentium PC the BEM program CPO [3] was o f the order o f 10 to 100
times more accurate than the best FDM (Finite Difference Method) or FEM (Finite Element
Method) programs [4-6] available at that time.
The sources o f inaccuracy for systems that do not involve a cathode or space-charge are
( 1) the finiteness o f the number N o f segments into which the electrodes are subdivided,
(2) imperfections in the shapes of the segments, (3) imperfections in the distribution o f charge
density on the segments, (4) the existence o f field singularities, (5) the approximations used in
evaluating the potentials and fields due to the segments, ( 6 ) numerical inaccuracies in inverting
the relevant matrix and (7) machine inaccuracies.
Additional sources o f inaccuracy for systems that involve a cathode and/or space-charge are
(1) the finiteness o f the number M o f trajectories, (2) the inaccuracies in simulating the
space-charge cloud in front o f a flat thermionic cathode, (3) the approximations used in
generalizing to a cathode that is curved or not fully saturated, (4) the inaccuracies in assigning the
space-charge o f trajectories and (5) the inaccuracies in calculating the potentials and fields due to
the space-charges o f the cathode and trajectories.
These sources o f inaccuracy will be carefully considered for a variety o f electrostatic
problems, including ( 1) calculations o f capacitances and associated singularities, (2 ) comparisons
o f potentials and fields with known analytic solutions. (3) calculations o f lens parameters and (4)
calculation o f the parameters o f energy and mass spectrometers and ion traps.
[1] D. Cubric, B. Lencova and F. H. Read, Electron Microscopy and Analysis 1997, Institute of
Physics Conference Series vol. 153 (1997) 91.
[2] D. Cubric, B. Lencova, F. H. Read and J. Zlamal, Nucl. Instrum. Meth. A427 (1999) 357.
[3] CPO programs, available on the web site
[4] S1M10N Program, information available at
[5] SIM3D Program, information available at
[6 ] ELENS Program, information available at http://www.isibrno,cz/~bohunka
Frank H. Read 3 * and Nicholas J.Bow ring1’
“Departm ent o f Physics and Astronomy. University o f Manchester. M anchester M13 9PL. UK
hDepartm ent o f Engineering, Manchester M etropolitan University. M anchester M l 5 6BF1. UK
*e-mail address: [email protected]
Sim ulation of therm ionic cathode space charges when the Langinuir-Child law is valid
In the simulation o f cathodes by the Boundary Elem ent M ethod (BEM) the surface o f
the cathode is effectively replaced by a sheet o f charge that lies on the surface and that is
subdivided into segments. We now describe for the first time the technique used in the CPO
programs [ 1] since 1994 to simulate the behavior o f thermionic cathodes.
The simplest example is the space-charge limited current o f a planar diode when the
cathode tem perature T is zero and the current density is given by the Langmuir-Child Law
[2,3], The essence o f the present m ethod is to divide the total charge Q in a cathode region o f
depth d into N regions that each have the charge Q/N and then to replace these distributed
charges by thin sheets o f charge at the centres o f the regions. This is illustrated in Fig. 1, for
N = 4. The only description o f a sim ilar technique seems to be that o f Andretta et al [4] who
used a single sheet that carries the whole charge density O, placed at the cathode itself.
Figure 1. Schematic illustration o f the positions o f 4 charge sheets, indicated by bold
lines, that replace the space charges in the regions enclosed by the broken lines. In
practice the first sheet would lie much closer to the cathode than indicated here.
The field produced by the N sheets is piece-wise linear between the cathode and anode
and it can be shown that the values o f the field, and also o f the potential, are correct at the
boundaries o f the regions.
In the practical im plem entation o f this [1] the full simulation o f a system that contains
cathode and other space-charges is o f course iterative. The trajectories in any given iteration
use the space-charges, including those in the cathode region, produced in the preceding
ju a m S s s sq j oj psuSissE si aSjEqs oiEudojddE sq j usq i uoiS sj siqi ui sjusutífes sq j j o ano
sssso jo X jopafBjj e j ¡ 'soqnj jo s[[sd aSjEqo-SDEds aqj j o suoisirauiip aqj uEqj j3[|buis qonui
s¡ snipEj s p o q p o oqj usqM [Bijuasss si qoiqM ‘u o i3 sj jsjn o sq j o} u o iS sj apoqjBD jau u i sqj
qojEiu 01 si osodjnd sji 'sapoqiBO x sa u o o jo j Xjuo sisixa u o iS sj sp o q p o sjB ipsousjui s q x
•siusuiSss s s sq j uo ssSjEqo sq i uiBjqo oj S3poqjB3 jEUB|d j o j j u s i u j b s j j SAoqB sqj
a s q R j s u s S oj 3 [qB u a sq 3 ABq 3 A\ opoip [Eousqds JO [B3lJpill|Xo B JO J uoijE nbs S^U OSSIOJ 8 uiA|0 S
jira u iS ss juajBd sq j Xq p sp u ajq n s 3 [3 u e pijos sures sq j [[ij oj p sp u sd x a s j b jnq sjusuiS ss
opoqjBa jusjEd j p q i
ssdBqs sures sq j 3AEq uoiS aj spoqjEO
js u u i
sq j ui sjusuiS ss s q x
■ X p A i p s d s s j u o i S s j .s S j B q o - s o E d s
o jB ip a u jjs ju i,
u o i8 a j
p g jE q o -s o B d s
s p o q jE O ,
‘j p s j i
s q j j o s i u s i u S s s s q j j u a s s j d s j s a u q p s q s B p p u E p j o q ‘u i q j s s j s a s u e j } s q x
s p o q jE O
'Z a m S ig
ui psjB jjsnin S| s iq x •p3iE[nm]s S u p q uraisXs sq j j o js s j sqj
j o 3joqA\ oq] jo j souopolB Ji sq ; j o ssíÜEqo-soEds sqj pjoq jsq j ssqrq jo S[po aSjBqD-SDBds aqj
suiejuoo jEqj u o iS sj ja jn o , aqj si p jiq j aqj puE siu sraS ss j o p s j s q y n j e suiejuoo p q j ^uoiSaj
apoqjBO 3)Eip3UJJ3iui, sq j si puooos sqj 'SAoqB psssnosip sjusiuS ss aSjBqD-SDEds spoqjEO
aqi suiejuod p q j ^uoiSsj a p o q p o jsu u i, sq i si is jij sq j :su o i3 sj a sjq i qsm Suijsip 3 ^
aaejjns apoqjes jo ajnjBAjiw jo jusuijksjx
•qjiA\ ijEsp sq XjisBa a jo js js q j
ssijiuL iojiun-uoN Ju su iS a s ]u3JEd sq j
U3a i 3
b joj
sdEqs puE azis
aS jsqo j o s p s q s S u ip u o d ssjjo o s q x
a p o q p o aqj oj paijd d s p p i j sqj ui
aqj 3ABq uaq) p s u iS s s a p o q p o
Ju su iS as a p o q p o sq j j o qoE3
poA \o||oj si SAoqE p a u ip n o a jn p so o jd s q x 'SjuauiSas jB [nguB psj
JB[n8uBui oiui pspiA ipqns
si aaE jjns spoqjBo jEUE[d sq j iBqj si uoiiE jusuisiduii ¡BOipBjd sq) j o p sd sB js q jo u y
to o z
sp u ^ i i u m j x
Sim ulation of therm al energy effects
Langmuir [5] has given a careful treatm ent o f the effects o f the initial thermal energy
spread o f the emitted particles on the space-charge distribution and current from planar
cathodes. In general the space-charge cloud in front o f the cathode creates a potential
minim um Vm that reflects electrons o f lower energy back into the cathode surface and thus
reduces the current that reaches the external region. This minim um becomes the effective
source and is known as the virtual cathode.
We use an iterative procedure to obtain a self-consistent solution o f the Langmuir
equations to establish the current density and the depth and position o f the potential
minimum. Then the charge in each sub-region is integrated and assigned to the corresponding
cathode space-charge segment. The velocity distribution o f the trajectories that pass the
potential barrier
-ie that are emitted by the virtual cathode- is assumed to be Maxwellian.
An example o f the near-cathode potential distribution obtained by this procedure is
shown in Fig. 3. Here a planar diode is simulated for a cathode that has kT= 0.1 e V and for a
field that is 10 V/mm in the absence o f space-charge The results obtained with three different
values o f the num ber N o f cathode space-charge segments are shown. At the highest value o f
N used here the potential and position o f the virtual cathode have converged approximately to
the values obtained by applying L an g m u irs method analytically. Another test o f the
technique is o f course the value o f the current density j that reaches the anode. Here j is
approxim ately 1% larger than the accurate value obtained analytically.
Figure 3. Potential distribution as a function o f the distance z, in mm, in front o f a
planar cathode that has kT= OAeV, obtained with the C P 03D S program, showing
the existence o f the virtual cathode. Results are shown for three values o f the
number N o f cathode space-charge segments.
Sim ulation of liquid metal ion sources
These sources are capable o f producing currents o f a few uA from source diameters o f a
few nm | 6J but the effects o f stochastic ion-ion scattering tend to increase the final energy
spread by several eV. We have used the C P 03D S program to model the stochastic scattering,
/UIO3'S0IjdouOjp3|3'MMM//:dHq 9JXS q3M 3lfl UO 9|qB[lBAB ‘SUIBjSojd OdO
6861 ‘>|J°Á
a\ 3[s[ ‘s s a jj oiuiapsoy ,,‘so p d o u o jp a ig jo sa jd p u ijj,, ‘jadsB ^ ’g pus sa>[MBH '¿\V d
'M )c-96l OOOZ) 6 ISV H>W lunuisuj p n /y ‘g u iíM o e fN PUB PB3ÍI H 'í
'LSZZ-08ZZ ( e O O t ) PL umuisui p $ visy ‘g u u M o g f N PUE PM >I H'.l
'SOI l (£661) ř9 u tnum i P S ^>1 I P p o T
'S£l7-6It7 (£361) IZ MN '* % / MpuiguEg \
-ť801 (9861) e i - a a
1 3 SUBJL 3 3 3 1 ‘iuububz -o P™ iuijbjaj i \ ‘B jp jp u y -yi
8861 <5I-10Á M3N 's s s jj iu n u 3 |j „‘saijd o uo] puB u o jp s jg ,, ‘iASejizs '(/\[
[ 8]
[9 ]
•S3U3J3JU03 sq i )B p aju asaid oq |||m uoijB |nui|s siqj j o sjjnsai a q x
•spouB puB apoqjBO aq; u sa M p q sagBjs p sjío jju o o 6 ui pasBajoui uaaq 3ABq sqjSua]
snouBA asaqj 3|durexa js jij siqj uj
-ajBjnooB K p u a p ijjn s su ib iu sj 8ui8bj3ab agjBqo-aoBds
aqj p u s 3AISS30X3 jou si sdajs X jopafej} j o
jsq u in u jbjoj aqj jBqj os ‘XpjBijdojddB
asBajoui oj poA\o[[B oq jsnui sqjSua] snouBA osoqi aoBjjns apoqjBa aqj u io jj opoooj
suoj oqj s y
' 8 uijohbds oqsBqoojs j o uoijB]no[B3 oqj jo j pap ašu si jBqj Áiisuop ogjBqo-ooBds
aqi S uiS bjsab j o j ssouBjsip ip u is puB sagiBqo Á jo p a fejj aqj ihbjuod jBqj soqnj ogjBqo-ooBds
oqj jo j npBJ [jBUis ‘uopB jgajui \jo p o f e ji jo j sqjguaj dajs jjo q s osn oj ájcssooou a jo ja ja q j
si )| uoigoj siqi uj
iuu A\aj js jij aqj uiqjiM Kpsoui 00B|d o>[Bi o jo jajo q j sjuaAa Suuojjbos
'02 XpjBiuixojddB j o jo p B j b Áq sasBajosp Ajisuap agjBqa aqj puB
uoi-uoi juajoiA jso u i nq j
g XpjBuiixojddB j o jo p B j b Xq sosBajoui ÁjpojaA uoi aqj utuç js jij aqj u i q j i ^
’aauBjsip qjiM
X|5p m b X jsa aguBqa apoqjBa aqj j o ju o jj ui suoijjpuoD aqj jBqj si X j|naijjip 3 |d p u u d a q x
" lin a ijjip /(i|Buoijd33xa aq oj Suqapoui
siqj asnBO jcq j s p a j ja ja q jo auios 3jb a jaq j jnq sp^aij q8iq qans jb aauB jjoduii siqiS qS au
j o asjnoo j o si apoqiBa sqi j o ju o jj ui aSjBqa-SDBds a q x
'aoBjjns sjí jb iuu//i 01 J ° PRU
b SBq jBqj luuç snipBj j o sa in o s uiniUBg p in b q b pajapoui 3ABq sa \ ajduiBxa js jij aqj uj
j o 3uij31)bos ousBqDOjs jo j [^ ‘¿] X|snoiA3jd p a q u a sa p anbiuqoaj aqj j o uoijBzqBjauaS b 8uisn
V. Romanovsky, M. El-Gomati, L. Frank*, I. Müllerová*
Departm ent o f Electronics, U niversity o f York, York YOlO 5DD, [email protected]
* Institute o f Scientific Instrum ents AS CR, Brno, Czech Republic
A scanning low energy electron microscope
(SLEEM ) was realized w ith a cathode lens in which
the negatively biased specimen is used as the
cathode. In this arrangement, electrons pass through
the microscope at high energy and are decelerated to
low energy before landing on the sample. Signal
electrons are accelerated back to the detector, which
is part o f the column. This design brings plenty o f
signal even in the landing energy range o f tens or
units o f eV. This is comparable with the known
LEEM method. In many cases the contrast itself and
its variation, as a function o f electron energy, does
not bring enough data for reliable image
interpretation. M oreover, exploitation o f low-energy
electrons enables observation o f differently doped
sem iconductor regions [1], However, the use o f the
Fig. 1. N ew scanning low energy
primary electrons with low energy brings some
electron microscope accommodated
problems e.g. low source brightness, increased
in the CM A
aberrations and sensitivity to stray fields (i.e.
defocusing the probe by SE collecting field) [2]. The
scanning A uger microscopy (SAM) is well-established experimental method. Surface
elem ental mappings in SAM could be affected by artifacts due to surface topography and
subsurface atomic num ber variations. A com bination o f SAM and SLEEM in one device
would therefore provide a sufficient tool to solve the problems inherent in these individual
methods. A lthough m uch has already been achieved in the area o f detection o f slow electrons,
none o f the methods currently know n is suitable to be built into a scanning illumination
colum n for A uger microprobe analysis.
Figure 1 shows the high resolution and fully electrostatic mini-colum n accommodated
inside the cylindrical m irror analyser. The colum n has a diam eter o f only 45 mm and is
85 m m long. Operating beam voltage is up to 10 kV. A Schottky emitter is used as the source.
The cathode, suppressor and extractor are constructed as one pre-aligned module with
dim ensions 20x20 mm (provided by YPS [4]). This configuration insures very high precision
o f mechanical alignment. In the middle o f the colum n is a built-in detector for low energy
electrons, which consists o f a m irror lens, m icrochannel plate and detection electrodes. The
detector is shown in Fig. 3. It is divided into six equal segments covering 360° angle. In the
middle o f the m irror lens is a small aperture w ith two important functions: deflecting signal
electrons onto the detector and acting as the prim ary beam -lim iting aperture. Between the
cathode module and m irror lens is another lens with a system o f alignm ent deflectors. This
allows the m icroscope to be used in two different modes. In the high-resolution mode, the
• (^ n ) X jspog pXo^j sqj Xq psjjoddns sem josfojd
uioo'pj[-sdA'AWiM//:djjq 'pjg sssjn o s s q o ij ^ io ^
'OEc'd ‘8661 -tsâupds ‘XdoDsojraijAj u o ip s jg Sutuubd§ " g ‘jsu ip ^ j
'66 (000c) 18 XdoosojDiuiBJJin “W ‘pB iuorym ‘'I ‘b a o js jp w “q ‘^p B ij
>01 (S66I) 8S X dossoioiureum ‘P 13 CTQ ‘oiaoj3<i
■jsbjjuos oiqdBjgodoj
s p o u i-g ry y ui s S euij 'ç 'Si j
‘A 300S X Sjsus ‘u s u ip s d s
8 u ijs sj 3 / n y ‘spuuB qo
s p u S is p su iu in s
g ry n > 'S y
s Sbuii
-[i ]
'jo p a ja p puuB qo -9
ui-ijm q a\ su sq x ■£ -S ij
'sjdures sures sqj uiojj V F Í3
qiiM u s:p j Bjjssds js S n y sqj SMoqs z 3Jn8i j 'spoui g ry y ui spuuBqo xis jp uiojj sp u S is j o
uoijBuiuins Suisn psSBun sjduiBS sures sqj SMoqs qoiqM
ifliM uosiiBduios ui psAjssqo sq
UBD jsejjuos siqdBjSodoj ui ssBaioui sSjbj y sobj Asqj jojasjsp qoiqM uo 8uipusdsp sjn o p o
ju sjsjjip ui sjosjjs sS p s jqSi[q8iq Xem siqj ui psjBSJS sqdEjSojsijAj ’sSeuii jnojoo B uuoj
oj psuiquios s j p s s j sqj puB in o p o g r y g jusjsjjip
b 01 psjjSAuos sba\ j p s s i qsB3 ‘[f] psjsBijqns
spuuB qo sjísoddo u io u sp u S ig -spoui g{JH
ui sjB ijsqns uoqjBO b u o ssp ijjB d p jo 8 j o sSbuii
A 300I2=d3 ‘P S iJ ao sídures sqj jo
B jpsds j s 8 n y ' j 'Si j
_®k 4—
üb SMoqs ç ’S i j 'SuiSbuii X Sjsus m oj jo j X jp p s d s s
•puuB qa qoB3 u io jj ssSbuii ui ss o u s is jjip
s 8 jbj SAJSsqo oj sjq isso d sí jj 'su iij sures sq j jb
spuuB qo xis | p sÄBjdsip 3JBA\yos m su s q g jb u S is
jndui j o sp A sj 0 0 0 ‘S9 j 3a o S A p ssj oj sjqB sjb
X sqg -sjíq 91 j o u o p n jo s s j 3 ABq sjndui y y -sjuiod
960t’x 960t7 sí u o ijn jo ss j Suiuubos uiniuixBjAj 'piB3
jsjn d u io o p ssd s-q S iq b qjiA\ Xjj33Jip sjstjijd u re
U0 IJB]0SI-0 jd 0
jB ijusjod b u o jBoy sjsijijduiB s q x 'uoijBuiquioo s p is s o d Xub ui jsq jsS o j p su io f sq oj sp u u e q o
m o|jb puB j0 ss3 3 0 jd 0 j3 iu i b Xq usA ijp 3JB sX sp j s s s q x 'sXb|3j psjB |d p p S sjnjB iuiuiqns
Xq SJ3IJI|duiB3jd |BI33dS qjIA\ p3J33UU03 3IB JOJ33J3P 3qj JO SJU3UI§3S XIS XIV
M3IA j o p p ij SuiaiS ‘pssn sí uisjsXs uoijos|j3p sSejs sjqnop ‘sijB jsojjssp 3]od-8 ub ‘Suiuubss
sqj j o j "yu ç-)? punojE jusjjns síduiBS j3Aipp ubo sm apoui siqj uj ’aiduiBS sqj jiq s u o j p s p
sqj i p XjjBsu puB sjnjjsdB sqj ui P3JB3J3 si J3AOSSOJ3 sqj 3poui JU3JJ113 qSiq sqj uj 'SuiSbuii
JOJ pssn 3JB SIXB sqj UO SUOJJ33P X p o PUB sjnjjsdB Supiuill sqj 3J0J3q P3JB3J3 SI J3AOSSOJ3
P. Schauer and R. Autrata
Institute o f Scientific Instruments, AS CR, Brno, Czech Republic, [email protected] isibm
In S(T)EM an image is formed using a focused electron beam, which is scanning across
a very small part o f the specimen surface. A scintillation detection system consisting o f
a scintillator, light-guide and photom ultiplier (PM T) processes only one pixel o f the image at
any given moment. N ot only efficiency, but also kinetic properties o f such a system are o f
great im portance. Scintillation detectors can show a noticeable difference in detective
quantum efficiency (DQE) due to the bad electron-photon energy conversion and/or light
losses in the optical part o f the detector. Up to now, some studies were engaged in
measurem ent o f S(T)EM detectors performance ascertaining very low DQE for some
detectors, but no suggestion has been made to optimize the detector set-up. To find the neck
o f a detection system, one m ust exam ine the w hole detection path (Figure 1) step by step.
Electron-photon conversion optimization
The main com ponent o f a scintillation detector is the scintillator. The scintillator provides
energy conversion from electrons to photons. It has to be very fast, possess high efficiency o f
electron-photon conversion, and it has to em it the light in the spectral region o f high PMT
sensitivity. To optim ize the scintillator one m ust use a direct excitation m ethod and measure
its cathodolum inescent (CL) properties, i.e. its energy conversion efficiency, decay tim e and
spectral emission characteristics. These CL properties can be measured using the equipment
built in our laboratory [1], The excitation unit o f this equipm ent is formed by a column o f an
electron m icroscope w ith an electrostatic deflection system and a blanking diaphragm, so the
continuous properties (CL efficiency) as well as kinetic properties (rise and decay times) can
be measured. The equipm ent is controlled and the data are processed by a personal computer
using IEEE-488 (GPIB) interface bus. Some tens o f different single crystal materials were
measured at our laboratory [2], O f these, single crystals o f YAG:Ce, YAP:Ce and P47 were
chosen as the most interesting ones for S(T)EM applications. Alternatively, the transparent
sintered YAG:Ce ceram ics with optically good isotropy and pore-free structure (as presented
by Ikesue [3]) seem to be interesting.
Optical properties optim ization and PMT choice
In the separate step o f optim ization the materials o f the light-guide, photocathode, conductive
and reflective film s and o f the optical cem ent have to be evaluated. Optical characteristics
(such as optical transmittance, position and width o f the emission band, spectral response o f
and matching to PM T photocathode sensitivity) have to be utilized at this step o f
optimization. In fact, the PM T choice, including its photocathode matching, is a relatively
simple task, because PM Ts are well developed comm ercial components w ith precise data
Photon collection and system geom etry optimization
Design o f the system including geom etry is the m ost dem anding step o f optimization. The
efficient geometry is dependent on a lot o f optical quantities o f all components used, and no
simple m ethod is available for this step. To determine the photon transport efficiency, the
Monte Carlo simulation method has been developed [4], In Figure 2, the significance o f this
step o f optim ization is demonstrated by presenting the efficiency o f the bad and good
geom etry o f the BSE scintillation detector for the S 4000 Hitachi SEM. The good system
possesses transport efficiency o f about 400 % com pared with the bad one.
'0LZ ‘(5661)
£9£ ‘F sd}L sfy d
"umjjsuj 'pnpj : f ‘jojXbx í'g ‘oiunj/\[
■Çgt71 d t 'I0A ‘0002 ‘anqndsy ipazo -ouig ‘z i m m Z 3Md
‘t^-UW i'd ‘s t r a p s
0Č0I ‘(0003) 801 udf ™S UWU3J f : -y ‘3ns3>[i
'I ‘(9 6 6 1) 6 'id d n s osouotpy Suiuuvog :<¡ ‘js n m p s pus >j 'eiru in y
■¿££-d t -I0A 'cOOS ‘B3HJV qjnos ‘UBqjnQ ‘g ¡ m D l f o ™ Jd ' "d ‘BjBjjny - d ManBips
S33U 3J3J3^[
qosz 3 sqj jo XousSy jubj£) sqj j o fcfcI£/W)/£0l ‘°N JubjS aqj Xq psyoddns sbm ^ jom siqx
[g] souopofE.ii u o j p a p XjBpuooos
j o uonEjnuiTS Suisn poqsijqnd uooq OAEq suiojsXs u o i p o p p uoijb ^ ijuios j o s o p u o p ijjo
uo)p 3 i[o o j o sjjn sa j 3upsojoiu[ ’s u o j p s p jeuS is j o uopoojjoo soouon|jui Á|[BijuBjsqns
ipjqM ‘S u ip u n o jjn s j o p o p p aqj oj p sjB p j jou oje uopE ziiupdo j o sdojs snoiAOjd oqx
u o p K /u u p d o U 0 ip 3 ||0 3 u o jj3 3 |3 p u n ‘u o p is o d ‘u o p e jn S ij u o a , i o p o p ( |
•joje |[ i}ui3s oi[i j o 33Bjjns aq) je iinod uoiiB ipxs
3ql JO SSJEUipjOOO 3l|l UO Á3U3I3IJJ3 liOdSUBJl
UOlOqd 0L[J JO 33U3pU3dsp 3l[l SlU3S3jd qdtiJTÎ Q£
■\Jiauioo3 poo§ sqi (q) pBq 3qi ( e ) Suisssssod suisisXs
uo|)33;3p u o i ; e | [ i j u |3 S qSnojqj iJOdsuEJ] uoioqd 3q]
sqj J O UO!lB|nUIIS 0]JE3 OlUO^Sj 'Z '§ ! i
J O S 3 I3 U 3 ]3 1 JJ3
sq i j o s3i[qui3SSE jo/puB sjjEd |EnpiAipui jo j spo q p u j
SAiiEinuns pus |B)U3UJU3dx3 Suisn dais \q dais
qied uoji33j3p 3|oqM oqi suiurexa jsniu auo ’luaisAs
UO[133)3p BJO ÎJ33U 3qi pUlJ OX JO]3313p U0I1B|[[1UI3S
1NH(1)S B u!
U0|i33i3p j o
3U ||jno -|
)U 0 jjn o A
IBüBis I___________
uoitBOiidmriLU uojioaisoioqd
(Aouaioif/a uonoanoo)
apouÄp i s i oj ijodsuej}
fp/srf iuniuenb apomeooioiid)
SU0Jl03|3010qd 01 UOjSJSAUOO
(■■■ ‘U 0I)03I)3J 'UOIlOBJ/aj)
0 1 uuoj) edBosa uoioqd
uiaisAs UOII33I0Q 3Sa paziuiudo
(" ‘Buuaueos ‘ ¡djosqe ‘ penaj)
spm ß }q6i| ui S9S0|
( - ‘U 0 l)0 3 lj9 J ‘UOIIOBJ/aj)
0 -| oi ßui|dnoo uoioqd
( - ‘uoiioaiiaj ‘uoijoBj/aj)
joiBjjiuios LU0J1 adBOsa uoioqd
(Áouaioiiia Luniuenb 70)
u o q e ja u aß uoioqd
(Áouaioiua uoiioanoo)
uoqoanoo u o jp a is
S U O J I O 0 |0
luaisA s u o q o a ia a g s s p a z ju jjid o u n
G. Schönhense. A. Krasyuk. D. Neeb, A. Oelsner, S. Nepijko, H. J. Elmers
Institut für Physik, Johannes-Gutenberg-Universität. D-55099 Mainz, Germany
([email protected] uni-m
A. Kuksov. C. M. Schneider
Institut für Festkörperforschung IFF-6, Forschungszentrum Jülich, D-52425 Jülich. Germany
1. Introduction
I'he developm ent o f a dynamic im aging technique with time-resolution in the sub-nanosecond
regime is driven by the need for methods to explore and control the fast dynamic response o f
magnetic thin film elements. This knowledge is param ount for industrial applications, e.g. for
new magnetic storage devices such as hard disks with a greater storage density or magnetic
RAM and spintronics elements. W hen the dynamic behaviour comes into play, high storage
densities are inevitably correlated to fast data transfer rates and therefore short switching
times o f the order o f a few nanoseconds down to below one nanosecond.
The majority o f methods for magnetic imaging with high lateral resolution work on the
basis o f scanning-probe techniques. The sequential scanning o f lines along the sample surface
contradicts a high tem poral resolution. There are optical methods to achieve the desired
temporal resolution like magneto-optical Kerr microscopy, but the lateral resolution is limited
by diffraction o f light and therefore not sufficient to investigate devices with critical
dim ensions o f the order o f 100 nm. A highly prom ising approach is the method o f X-ray
cxcitcd photoem ission electron microscopy (X-PEEM) because the required precise timing is
achieved by exploiting the excellent time-structure o f Synchrotron radiation. To our
knowledge this method so far provides the fastest kind o f electron microscopy with
a resolution o f presently 14 ps and potential o f further improvement.
2. Experim ental technique
As high-speed probe we em ployed short Synchrotron radiation pulses at BESSY (Berlin) and
the ESRF (Grenoble) in the soft X-ray range. Very short probe-pulses down to At = 1.6 ps
length were achieved in the novel "low-alpha“ operation mode at BESSY [1], for many
experim ents the typical lengths o f 50-100 ps o f the standard modes proved to be sufficient.
Fast magnetic field pulses have been generated by passing short current pulses through
coplanar waveguide devices with magnetic structures being lithographically prepared on the
stripline surface, see F ig .l. A stroboscopic pump-probe set up [2] with a variable delay (few
ps step width) between field pulse and Synchrotron (probe) pulse allowed to take snapshots o f
the dynamic response o f the magnetic domain structure as a function o f delay time. X-ray
magnetic circular dichroism XM CD yields magnetic contrast with the magnetization
com ponent along the photon beam (horizontal component) showing up as grey value [3].
Results presented in section 3 have been taken at moderate time resolution at the ESRF
in 16-bunch inode (photon pulses 105 ps, period 176 ns), those o f section 4 in the “low-alpha”
mode o f BESSY (1 .6 ps pulses at a period o f 2 ns). Perm alloy (FeNi) structures were prepared
on a Cu micro stripline. The current pulse I through the stripline gives rise to a magnetic field
pulse H (cf. Fig. 1). determined in situ via a transient change o f image size caused by the
electric pulse when travelling across the field o f view [4], The rise time o f the magnetic field
pulse (more precisely, the slope o f the temporal field profile) governs the dynamic response,
because it determ ines the strength o f the Fourier components in the frequency spectrum o f the
pulse. The slope was about 1 m T / 500 ps and twice that value in the section 3 and 4
experiments, respectively.
([p] wojJ'üiop) sSps Cj if\[
difl jD Q J W X °> 3nP s! isvxiuoo o ij3 u S o j\ 7/ s s jn d p p i j m iauSm u M /odiun n J o 3%ps Suipoaj
3\ji S u u n p 8ui.i Koj¡o w a 3 ¿ o j o u u d u v d u w u io p a¡///o s s S d w i p y g g j lo tfsd o u s fo S3U3$ :£ S i j
■[ç] apon . [ |\ 'N 0 0 »Ml Suisn uoijEnba uaq[if)-zi|i{sji']-nB puB 'j aqj uo pasEq uoijB[tia|BO
jEaiureuXp b Xq pauiEjqo uaaq scq u.iaiiBcl imbuí op pajB[nuus a q x X[]Ejuozuoq p a ju a u o si
ureaq uojo q d aqj ‘asjn d p ja ij aqj sajBOipui h ‘uoijoa.iip uoijEzijauSEui ]BOO[ aqj ajouap SM Ouy
a |y o jd asjnd aqj ui pa>[jBui sb as[nd p jaij aqj j o aSpa 8 u|pB 3 [ aqj oj p u o d sa jjo a sjoqsdßus
jaq jo a q x pai|ddE uaaq aAEq sas[nd p p i j jo sa u a s snoiAaad aayB ajEjs a|qBjsEjaui b sjo a y aj
a§Biui sd o a q x 'sapX o g0 l x Z. J3A0 SuiSbjoab oj S u ip u o d sajjo a s 021 XjiEaidXj aiaM sau n j
a jn so d x a aqj Xjisuajui j o su o sb sj jo .[ uoijBjnp su ç i puB apnjqduiB [ ui £ jnoqB j o asjnd p p i j
JB[odiun b j o uoijBOi[ddE S u u n p ua>(Ej uaaq aABq sjoqsdBUS j o sa u a s a q x '("uiu 0£ ssa iq o iq j
‘ p i uirl i i ‘ P '° m rl o s ) S u u X ouB uuaj p a jn jo tu jso ja iu i b jo j sjjn sa i A\oqs £ pire j s a m S ij
u o ijk jo j uoijBzijauSeui juo.ioqo.iui put; jua.iaqo ) ■£
(¡tfSu) sSn¡d
3M)lHOJOllU IfJlM. uspjoif 3]dlUDS 3]q03SuDlj3X3-/¡f{f¡ 31{JJ'O l/dVaSo/Ol/d pUD (ifol) lioipiuusj fo
s3jnpnjisoA3tw oiisuSmu i{ji¡n apinSatWM. uvuv¡doo 3i¡i J o aSrnuL sdoosouotw ¡ooitdQ : j S i j
The Landau-type feature (flux closure structure, see lower left image in Figs. 2 and 3) visible
on the left-hand side o f the ring is disturbed by the field pulse resulting in an increase o f the
"dow n” domain, as anticipated from the field direction H. The "up” domain reacts by forming
stripes perpendicular to H (see top images in Fig. 2). This behaviour was verified in the
simulation and is obviously caused by the tendency to minim ize the magnetic stray field
energy in the transient state. In the plateau region the average grey level suggests
a predom inantly downward magnetization. The fact that the magnetic flux runs from top to
bottom through the ring leads to the formation o f a so-called "onion state” as indicated in the
top row o f Fig. 3.
Diffuse domain boundaries give evidence o f deviations from a strictly periodic
switching behaviour (averaging over 7 x l0 8 cycles in the stroboscopic mode). However, the
structure always ends up in the same final domain configuration as is clear from the relatively
sharp pattern at the foot o f the trailing edge o f the field pulse (right image in Fig. 3). Even at
zero field the pattern is not yet fully relaxed, visible by comparison with the image at the foot
o f the leading edge. It shows characteristic stripes in the region o f the Landau structure.
Restoring the Landau pattern (driven by minim ization o f the stray field) obviously needs
more time. The simulation reveals that this slow relaxation is caused by vortices being pinned
close to the right hand side o f the pattern; see "Sim ." in Fig. 3. The simulated pattern is
considerably sm aller than the experimental structure so that the num ber o f vortices is different.
4. Observation o f magnetic modes
In the high tim e-resolution measurem ents at BESSY the field pulses were shorter and had
a higher slope. A t these conditions eigenoscillations o f the magnetic structure, so-called
magnetic modes are excited. The characteristic frequencies o f these modes lie in the range o f
several GHz, i.e. their observation requires a tim e-resolution in the region o f a few ps. Figure
4 shows first results revealing a magnetic mode in a Permalloy structure o f 15
x 30 (im.
■p3SpS[A\OU}[3E X[]nj3JEJ§
sl ..££11 ddS sssszojdsSunisisijsuSBW sípuqssB Jjjn uniiBj§ojdj>[undj3Áups„ qSnojqj
!):I(I SE ip.w SE (00S9NE0 P«b 6/VCI8lS>IS0 ‘S/IPMÍlIS^ISO sjuejS) j g w 8 ¿q yoddns
Ibioubui j uoijEJsdoos puE jjoddns ju sjp o x s jo ; j h s H 3lP PUB ÄSSH 8 J ° JJB1S 3xfl ’sluBlP 3M
•/jiuuioo/Ao 8 'jsiirqjBui//:djjq ‘¿pinQ s,su3sß JM/MOO '-isyod 'O 'Cl PUB snqBuoa f -ja
psjjiuiqns ‘je J3 q33|\[ ’a
6£81-S8Z. I 'H (cOOc) 59
'-ß o a j day -ssusquoqss 'O P«e Js p p u q ss 'H 3
juud ui (fOOč) 'u to u sijj p x -.ijjads . ip z ft f ‘j e js js u s p o 'V tqoJEiu £Z suijuo
iqnd (ťOOč) V • '% / l d dV Pub 898-£98 (£00c) 9L V s‘üld ld d V ‘ F I3 >pXsBJ5i -y
uoijE3iunuuuo3 sjEAud ‘^pEpuojj "M PTO 3P 'A S S a 8 'MAVM
}U 3 U I3 § p 3 [A V 0 U íp V / S33U3-I3J3JI
•sjnjnj JB3U sqj u¡ suiosjsao sq ](im jEqj sjuibjjsuos
|E.->iuqo3j oj snp jsqjEJ jnq sjnjBu |E dpuud jo jou 3je jsjjej sq x -poqjsui sqj jo suoijejiuiii
ju3S3jd sqj osjE jnq jBijusjod sqj sjo sy sj jjnssj siqx 'Zjjr) S 'č l jnoqE jo X ousnbsjjusSp
UE JB SpOUI 3[J3u8euJ B JO S3U3piA3 SSAl3 (sd \,\ jnoqß) UOIjr\[OSSJ SIUIJ pSAOjduil JB U35(EJ
3|duiBX3 |Eup sq x 'sm iS sj (sijEjs-isEnb -3-1) 3s|nd-p[3ij-M0(s sqj ui jussqB sjb jBqj ssjn jssj
3ISBq UIB]dx3 P|n03 UOlJBjnmiS OIJ3USeUJOJ31UI y '3JE3S su-qns 3qj UO SJ33JJ3 |B3iuiBuXp oj
JO SuiSeIUI 3ld03S0q0JJS JOJ JU3UJ!J3dx3 UE JO SJJtlSSJ pAOU p3JU3S3jd 3AEq 3M ‘U0IS11[3U03 UI
■3|qiSlA 3JB (UOIJE||I3SO 3qj JO S3SBqd UIEJJ33 '3'l) SSUJIJ UIBJJ33
je sjoqsdBus X[u q X jsnonuijuoo XjEpunoq uiBmop sq j j o jusiusaouj XjojB[jpso sq j mo | joj oj
3Sje | ooj sbav X jjEsp sd QZ j o X E pp-siuij sqj j o JUSUISJ3UI s q x uoijisod lu n u q q m b s sjí jnoqB
u o ijE |jp so (psdujEp) e sjíq iq x s ssSeuii sq j j o sjjuss sq j ui uiEiuop íjjBp psdEqs-puouiBip sq x
I sns.t3A Btuno juBuisDojdsip ,11/1 MO-tfpaiDiuijsa sd z t f f ) ç 7 /
inoqD [o X jusnbs.ifuaSia un id u o iw u p s o u w u io p d vrz ‘spoiu yi/suSnui n J o sjo ysd o u g :p S i j
Mikhail Sudakov* and Sumio Kumashiro
Shim adzu Research Laboratory Ltd.,
W harfside, Trafford W harf Road. Manchester. M l 7 1GP. UK
* corresponding author: [email protected]
In a typical ion trap operation cycle ions are trapped within fast oscillating quadrupole
field for more than 10ms in presence o f neutral buffer gas. Comparatively low voltage
periodic excitation fields are used to excite motion o f trapped ions mass selectively with the
use o f resonance phenomena. Ion trap performance in terms o f resolving power and scan
speed is sensitive to small distortions o f trapping potential. Correct ion trap simulation should
be able to reproduce accurately the effects o f ion collisions with buffer gas and the influence
o f small deviations o f the trapping potential from a pure quadrupole field.
The influence o f several simulation parameters on the accuracy o f ion trap simulation
with the use o f SIM ION-7.0 was investigated. Prediction for the secular frequency o f ion in
a RF trapping field was used as a criterion o f simulation accuracy. It was found that the
secular frequency o f ion oscillations w ith the use o f SIMION potential arrays could differ
from theoretical prediction by few percent. Three possible reasons for that were investigated:
1 ) insufficiently small integration step; 2 ) errors in electrical field interpolation from potential
grid (numerical differentiation); 3) errors in field solution by finite difference (FD) method. It
was found that simulation with the use o f standard 4-th order Runge-Kutta procedure with
integration step 0.01 o f the RF period provides prediction for secular frequency with accuracy
better than 1 ppm (in a pure quadrupole field). Errors due to numerical differentiation were
estim ated to be below lOppm for geometries o f practical interest. Finally investigation was
concentrated on the quality o f SIM ION field solution.
Errors in field solution from SIMION appear from tree major factors: 1) field faults
due to different boundary conditions; 2 ) errors due to finite difference approximation o f
Laplass operator; 3) errors due to stepwise approximation o f the electrode shape. According
to SIM ION-7.0 manual [1] potential array is refined with "insulator” boundary conditions.
Due to this SIMION cannot reproduce field solution accurately for problems with other
boundary conditions such as. for example, "free space”. Fortunately for ion trap simulations
the area enclosed by electrodes is o f importance. It was found that field faults due to incorrect
boundary conditions for axially symmetric ion traps with hyperbolic electrodes (theoretical
geometry) are below 5 ppm. provided that the size o f the array covers at least 2ro, where ro is
an inscribed radius o f the quadrupole field. Errors due to finite difference approximation o f
the Laplass operator were estim ated using a model o f 2D quadrupole w ith circular electrodes.
A 10-digit accurate field solution for the same problem [2] was used for comparison. Potential
array was created with a step O.Olro and grid potentials were assigned to correct values taken
from accurate solution. In order to avoid boundary condition problems all boundary points
were replaced by electrodes with potentials from accurate solution. Although the grid
potentials in this case were correct to at least 10 digits the refine procedure is triggered by the
difference between continuous Laplass operator and it’s FD analogue. Potential array was
refined with convergence objective o f le-7. The absolute difference between accurate field
solution and SIMION field was found to be below 5ppm within entire simulation volume.
Finally, the errors due to the stepwise shape o f electrodes were analysed. The same potential
arrays as previously were used, but with electrode points assigned to constant voltages. By
ï6 Z Z -0 6 Z Z dd
‘ £006
‘L V *
‘ u io j j o o d s s se jm u i i u o o
pidB'y ‘p p i f a/odopo pdppv LjiiM sajodnjpvnb jo a u ij ‘SBjSnoQ'Q pire A ojppng'j^
S 3 -I3 'd ‘0006 ‘3 3 3 N ollBPI 7 nnun^[s,j3sß ()¿ uoisudy[ q ç u o m S 'W ( j Q
[ []
•juouirujsui jeiijob oqj ui ooBjd oíjej jou Xbui qoiqM
‘sopnjqdure uoijBjqiA qSiq qjiM suoi j o jnoiA Eipq jb su iju o u ojqissod j o ojbme oq p(noqs
jo sn ouiij ouïes oqj j y 'odBqs joojjod j o so p o jjo o p u o posEq uoijBjnuiis UBqj jojjoq uoijBnjis
[Bjuouiijodxo oonpo.idoj UEO SÁBJJR jBijuojod NOIJAÍIS u0 pssBq uoijB jnuiis 'u id d o o i -»ao
Aobjiioob qjiM uo ijoipoid OAOiqoB oj jo p jo ui junoooB ojui uo'fRj puB p 0)Bn|BA0 oq ueo p o q p u i
Q j jo sjo jjo p p i j '[[Buis X p u o p y jn s si dojs p u S oqj popiAOjd ‘soujouiooS dEij uoi jBOidXj
jo j ojEinooE ojinb oq oj p u n o j sem uoijrqos p p i j N O M IS 3 M1 sisXjeue jo jjn so j e s y
•suoijBjnuiis dBJJ uoi ui Buouiouoqd jeoui[uou o|qBjo¡pojd XjpjBq jo j ojqisuodsoi
oq UEO qoiqM ‘sjuouoduioo ojodijinui jopjo qSiq jo uoijBuiqiuoo b sb poquosop oq ueo
sjojjo p p i j '(°J [0'0 M opq dojs p u § ) SÁBJJB qgnoi jo j juooiod woj sb Siq sb oq ueo pus dojs
p u 8 e oj [Euoijjodojd si oouajojjip oqj jEqj punoj sem ji su 0ijn | 0s ojbjiioob qji.w uosuBduioo
I. Vlček. B. Lencová and M. Horáček
Institute o f Scientific Instruments. AS CR, Královopolská 147, 612 64 Brno. Czech Republic.
E-mail: [email protected] isibm
Our aim is to redesign the low-energy SEM (a scanning LEEM) in an UHV apparatus
designed in our institute [ 1] to allow the detection o f the angular distribution o f signal
electrons. For this purpose we have to separate the signal electrons from the primary beam
with a Wien filter [2] and project the image o f the back-focal plane o f the objective lens on an
area-sensitive detector (aback-illum inated CCD [3]). Therefore we have to design a new
electrostatic optics working in UHV.
The arrangem ent o f the key parts is
electron beam from a Schottky gun
lens on the specimen. For the beam
the lens. The beam aperture must be
signal beam.
shown in figure 1. The crossover o f the primary 5 keV
is imaged with the electrostatic objective lens/cathode
scanning a two-stage deflection system is positioned in
placed above the W ien filter so that it does not limit the
The Wien filter is producing magnetic dipole fields with eight poles o f equal size, w hich are
electrically insulated and serve also as electrodes for electrostatic dipole field. This
arrangem ent guarantees a perfect overlap o f the two mutually perpendicular dipole fields. The
Wien filter is stigmatic for the primary beam if we superimpose a suitable electrostatic
quadrupole field on its electrodes. For our application the deflection o f the signal beam can
be as low as 15 degrees. Its minim um value is given by the requirem ent to position transport
optics close to the W ien filter. By changing the orientation o f the fields in the filter, it is
possible to direct the signal on several detectors.
Another advantage o f the Wien filter is that it has straight axis, so that the microscope can be
operated with the filter switched off. e.g. in the standard SEM mode. J h e aberrations
introduced by the filter are dem agnified by the objective lens. In order to eliminate the effect
o f energy dispersion for high-resolution imaging, an intermediate crossover m ust be placed in
the center o f the Wien filter.
In the SEM mode the objective lens works as a unipotential lens and the conical shape o f its
outer electrode allows enough space in the sample region. In the low-energy mode we put
high voltage on the insulated sample. The action o f the cathotte lens decreases the beam
landing energy on the sample w ithout much sacrifice on spot size. The values o f image-side
axial aberration coefficients on beam landing energy are given in figure 2 for the working
distance 1.5 mm. W ith the cathode lens switched on. we have Cs=0.9 mm and Cc= 0 ! 5 mm at
100 eV. at 5 keV Cs=47.3 mm and Cc=18.3 mm, respectively. The signal electrons are
accelerated into the lens and a diffraction pattern is formed in the back-focal plane. The backfocal plane is then im aged with transfer lenses on the CCD elem ent used as a fast positionsensitive detector, whose output is then used to form the SEM image [4].
I. M ullerová and L. Frank. Adv. Imag. Electron Phys. 128 (2003). p. 309.
B. Lencová and I. V lček in “Proceedings o f the 12th European Congress on Electron
M icroscopy", ed. L. Frank and F. Čiampor, (CSEM . Brno) (2000) Vol. Ill, p. 187.
M. Horáček, Review o f Scientific Instrum ents 74 (2003), p. 3379.
Supported by the Grant Agency o f the Czech Republic, grant no. 202/03/1575.
'A 3 ()() [ JE mu 11 pin; a 351 ç j e u i u ç popoclxn si uoijn josaj
sd o osojoiui ni[ [ (uiui ç" j soirejsip 8up[iOA\. ' a 0>I ç ÂSjduo ureaq X m u u d ) sus] SAipafqo
jn o jo j \S jd u 3 guipirej uo 03 pire s 3 sju sio ijjso o uoijBXisqE rtqj j o aouapiradsQ •£ ru n 3 |;[
[As] A6jaua 6uipue|
u m \ \ oqi put; siraj SAipafqo aip - IAIHSH'I p aso d o jd j o sy c d Â3 >( [ ainSi.j
uaw ioads
SU 0|
jaillj U0i/\A¿0
M 9 jA
jjo d s u e jj
P. W androl, R. Autrata
Institute o f Scientific Instrum ents AS C'R, Brno, Czech Republic, [email protected]
Nowadays the developm ent o f the image formation in scanning electron microscope
(SLM) is oriented to the use o f scanning electron microscopes with low accelerating voltage
o f the primary beam (LV SEM).
Secondary electrons (SEs) for topographical contrast observation and backscattered
electrons (BSEs) for material contrast observation are the main components o f the detected
signal in LV SEM. While the secondary electrons (SEs) can be detected either by EverhartThornley scintillation detector [1] or by the "in lens” SE detector, the detection of
backscattered electrons (BSEs) in LV SEM is an unsolved problem yet.
The scintillator-photom ultiplier detection system installed under the pole piece o f the
objective lens represents the most efficient BSE detector in SEM. However, the initial energy
o f the BSEs o f 0.5-3 keV is the energy, on w hich the light yield o f scintillators decreases.
Therefore, it is necessary to accelerate the BSEs energy by electrostatic field before they
reach the scintillator. By this electrostatic field both BSEs and SEs are accelerated to the
scintillator and the image is a mixture o f material and topographical contrasts. It is necessary
to separate SEs by an energy filter or an immersion magnetic lens for the true material
contrast observation.
The electrostatic field for the signal electrons acceleration is created by a high positive
bias o f 3-5 kV o f the conductive layer, which is applied on the scintillator. The energy filter
for the SEs separation is represented by a retarding grid with negative bias o f around -100 V.
This potential is sufficient to deflect the essential part o f SEs. Another possibility o f the SEs
separation is the installation o f the detector under the pole piece o f an im mersion magnetic
objective lens, w here the SEs are focused by a strong magnetic field into the objective lens
and only BSEs are collected by the scintillator.
The described low voltage BSE detector and its electrostatic field was simulated by the
software package Simion 3D [2] (see Figure 1). The scintillator covered by a conductive layer
is placed in a shielding casing, below the scintillator is the retarding grid and the primary
beam is shielded by a tube on a ground potential located in the hole in the centre o f the
scintillator. Figures 2 and 3 represent simulations o f low energy BSE and SE trajectories in
the electrostatic field o f the detector. It is evident that the computed trajectories certify the
theoretical predictions. W hen there is a bias o f 3 kV is on the scjntillator and a negative bias
o f -100 V on the grid, the BSEs, with their energies o f 0.5-3 keV, are not affected by the
negative bias o f the grid and continue on straight trajectories to the scintillator (see Figure 2).
SEs. with an energy o f 5 eV, are deflected from the detector by an electrostatic field o f the
grid (Figure 3).
A strongly excited m agnetic immersion objective lens is another possibility o f SEs
separation. The immersion objective lens, low voltage BSE detector and signal electron
trajectories were simulated by software packages MLD, ELD and TRASYS [3]. A magnetic
field created by this lens affects the SEs so that approximately 95% o f SEs with an energy o f
5 eV are focused on helical trajectories into the objective lens (see Figure 4) and there they
can be detected by the “in lens" SE detector. BSEs are also influenced by this strong magnetic
field, but owing to their higher energy, they are not focused into the objective lens. They are
spinning around the primary beam trajectory and they can be detected by a low voltage BSE
-ojsj; iubj3 ‘aiiqnday q o s z j sq i j o ssD u ap s j o XuispBoy sq i Xq p a p o d d n s si -sjjom siq g
99- 9Ç -d '0 6 6 1 '86ČV 'S3>I -sXqj qpjAj -uirujsui -pnjs[ -sjo p a g sp
pue s3sus[ j o uopBjndiuoD sqj joj sSE^pEd iuej3oj<j : q >[Ui[3ssi^ ‘ g BAoauaq
L\L 'd ‘S661 ‘B1UBPV 'soidox psqiV PUB XdoDsojjoadg
SSB^ uo JU0 3 SWSV pj£t7 JO sSuip3330Jd u¡ o ’9 u oisjsa q £ uoiuiiç "V ' a IW 1
'8PZ-9PZ d 'L í 'I0A ‘0961 'ujrujsui 'p s ’f s ju s jjn j u o jp s jg XSjsus
-Moq aj 3 diuB0 J 3 im - 0 J3 i^\¡ jo j jo p s p c i pueqspiyw :-ja g y Xspjjoqx " g ’I J-reqjSAg
'A0 P!J8 S u ip jB p j 'A>1 £ J o p n p u p s
sqj UO 3 p O jp 3 |X 'SU3[ UOISJ31UUJ1
sqi jo p[3 ij 0 |P u3 bui sqj ui
S3S A 3 S J ° s s u o p s fe jj :p sjn S ij
\i i
A 0 0I- P!jS SuipjBpj ‘A>[ £ JOIEIJIJUIS
aqi uo s p o jp s jg sg § a 3 ç jo sa u o p a fe jx '■£ SJnSij
3 pO JJ3 3[3
A 00I- PP 3 SuipjBpj
£ Jo p u p u is sqj
uo s p o jp s jg s g s g a 3>[ 1 jo s s u o p s fe jj : j a m S y
o 0 6 'o 0 /^
(a ®s) as
•jo p a p p psuSissp sqj jo pporn
u a u ip s d s
h o iu iis
g ajn S ij
« p a ip a i»
p u S ä u ; p J B | 3J
ja p jo q p u S
a p o jp a ja
äupiU^DDB JO;B|mupjj
sq n ;
S uisbd
■pspsau st puS SuipjBjaj s q i uo SEiq ou su s ] OAposíqo uoisjsuiiui ub qjiA\
s jA jg s A I UI 'P!JS SuipjBpj aqj uo SEiq SApESau b XjddB 01 Xjb ssso su si ji u a q j 's u s ] SApaafqo
uoisjaiuiui sqi Xq psddm bs jou sjb qaiqA\ "sjA jgs A I u í u o ip s p p g g g aqi jo j pssn aq
UBO jo p a p p psuopusui SAoqB p q i ‘pajEJisuoiusp sauopoÍEJj g § puE g s g jo Xpnjs sq g
Liping Wang. Haoning Liu, John Rouse, Eric Munro and Xieqing Zhu
M unro's Electron Beam Software Ltd. 14 Cornwall Gardens. London SW7 4AN, England
Tel & F a x :+44 20 7581 4479
Electrostatic and/or magnetic multipole elements are widely used in electron optical systems,
which, in combination with round lenses, usually act as aberration correctors. Numerical
simulation o f multipole systems is very important in designing such systems. Here we present
the differential algebraic (DA) m ethod for aberration analysis o f electrostatic and/or magnetic
multipole systems. The DA m ethod [1] is a smart and effective way to compute the
aberrations o f electron optical systems, theoretically up to arbitrary order, with very high
accuracy. A novel technique, im plem enting this method using C++ classes, has been
introduced by W ang et. al. [2] Recently, in order to simulate realistic optical systems, we have
enhanced our DA software by combining it with Hermite fitting o f the practical lens fields.
The conventional m ethod o f analyzing aberrations o f multipole systems is to first obtain the
paraxial properties and then obtain the aberrations order by order by successive
approximations. This involves enorm ous effort in deriving complicated aberration formulae,
which will become increasingly unbearable. In 2002. an extensive aberration theory, up to
secondary aberrations, for multipole focusing and deflection systems was developed by Liu et
al. [4], which can handle systems consisting o f electrostatic round lenses, electrostatic and/or
magnetic multipoles, as well as deflectors. Recently, this theory has been extended to
multipole systems including magnetic round lenses.
By contrast, the principle o f the DA method is based on a non-standard analysis methodology.
For the aberration analysis o f electron optical systems, the core o f the DA method is simply to
trace a single ray. It is similar to conventional direct ray-tracing, but instead o f on the
conventional real space R, it is carried out on a non-standard extension o f R, namely nD r , with
the initial position, slope, etc. transformed as DA variables. The space„Z)vhas many
rem arkably sim ilar properties to real space, but has some unusual and striking features, such
as the existence o f infinitesimals, which are nilpotent. [2J A fter numerically solving the
general trajectory equation, the results are DA quantities which simultaneously provide
information about the optical properties, including the aberrations, up to any required order.
A fter perform ing the single raytrace, the aberration coefficients, in their conventional form,
can be extracted from the DA quantities by simple algebraic formulae.
As an example, a multipole system including a magnetic round lens, four electrostatic
quadrupoles and two electrostatic deflectors (as shown in Figure 1), has been calculated by
the DA method, and the results have been compared with the results o f the conventional
aberration theory. In Table 1, the values o f the third order im aging aberration coefficients are
given, w hich shows that DA results agree with the conventional theory. The beauty o f the DA
method is that it is trivial to extend it to compute the higher order aberrations.
•ZOOS ‘16Z-\LZ ‘£6 ‘¿doosouaiwvjjin ‘n q z x P«® ‘ssno-fl [
'a ‘n iq H
eOOZ d as ‘sSpuquiBD T 00c 3NJM ' m i X pu® ‘oiunjAj -3 'nr[ h ‘ss n o y f ‘§ueav d ' I
ZOOZ ‘¿81-181 ‘£11 'V.l(lO "Ie ^ S u a q j f - g ' S u b í 1 - 1 - 3 u b m d - 1
'6861 > Z I-6 0 I PZ ‘s-io w ^p o o v s p ii.m j ~z.10g
( '°,nJo div2nfuoo xs/diuoD =.ig :¡‘m fo ajDSnfuoo xajdiuoD =j y
:uoijisodpiiiui x!>idiuoj=uM=g : ado¡s ¡vijiui x3¡diuoo=fM.= y :\DUdpuddap ¡DuoipunJu¡)
(s/iun /£ ui) sjuspijfioo uoiiDUuaqn S uiS duii uapuo puiqj fo uosuodiuoj 7 a/qnj
% K 6£'0
%I 6£'0
% 9(t’0
% 6 L Í'0
%LZ6000 0
sSk j s a v
9 90S0S¿ I 6I 66S'99íTI
88e9f 1t ' 0Sc8lwsťoe
io w h I‘6l08IfrSSM)£
^0 30 0
30 30 30
30 0 0
u o iy o is iQ
= .1 n B I3 A \ /
8 iC£8ť
VPZ9PČ8' 00t?£
ZO'blOis 'm s 8
3V 30 0
3V 30 30
3V 0 0
30 30 V
0 0 V
30 0 V
u is iiB u iä n s y
n . i n in A . i n )
33E J3A Y
£I9l f96C9'¿££C9'0£¿S5I8'6£c
I 8f ¿ 6c
S£'¿St8£I 6'£S£c 66811
30 3V V
3V3V a
30 3V 3V
0 V V
3V 0 V
30 V V
aSE JO A y
9C00 0801 99Č"
06000 0
St+'IS 6
3 A I)E |3 ^ J
K 0000
ra o o o o ’ o-
s j] n s 3 j [E jH a iiq u o p E J i s q y
3V 3V V
DV Oy 3V
3V V V
s u o ijB jja q v
a jn p s d y
[B u o ip u n j
suoíosifap Difvisojfosp
o,nj pun ss/odn.ipvnb o im so jp a p .m o f‘sua¡ puno.t aipiäm u v l/ iim utaisÁg 7 sutiSij

Podobné dokumenty

proceedings 2006 - RECENT TRENDS 2016

proceedings 2006 - RECENT TRENDS 2016 Published and distributed by the Institute of Scientific Instruments AS CR. '* Technical editing by Marek Frank, printed by CCB, Brno. Copyright © 2006 by the Institute of Scientific Instruments AS...


proceedings 2002 - RECENT TRENDS 2016

proceedings 2002 - RECENT TRENDS 2016 Recent Trends in Charged Particle Optics and Surface Physics Instrumentation