proceedings 2006 - RECENT TRENDS 2016



proceedings 2006 - RECENT TRENDS 2016
Recent Trends
in Charged Particle Optics and
Surface Physics Instrumentation
o f th e 10th In te rn a tio n a l S e m in ar,
held in S k a lsk y d v u r n e a r B rno, C z e c h R ep u b lic,
fro m M ay 22 to 26, 2 0 0 6 ,
o rg a n iz e d by
th e In stitu te o f S cien tific In s tru m e n ts A S C R an d
th e C z ec h o slo v a k M ic ro sco p y Society.
Edited by Ilona Miillerova
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 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 of 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 80-239-6285-X
'XpoqÁJDAD j o j in j j ir u j p ire
lURstíojd oq [ [ia \ j n a p áíjsir^ s; s q j ir j u s d s sÁRp sqj suiij siqj osp: jBqj SASipq o j 3 iu M opy
s^ B S jq s S u i j s s u j s q j § u u n p s u p s s n o s ip a/v\ j a q ¿ s j b s A
u i ureSE s n u p u o a o j ju b m sa \ o q
| s a s [ j u 3 [ p o x s X jiB u o ijip u jj s q j u o s ju E d p ijjE d j o Ájq
- u j s s s b s q j S u u s q j E § u i p s p s s a o n s SEq b u o [ i u i b S b s a u o ‘ i p m
( p a i u r e p sá ea \ | e A sq j jBqM)
s u o juRSRSjd s j in b b sr tu o q i o j u m o u >[ u s a s s a B jd b o j s j s i j u s p s p sp E o p sA O a q i S u ij j s S qjiM
s u i s j q o j d u jo jj J s j j n s s j n o o>[i[ s S u i j s s u i .lo p r u iis ‘¡p M s a i a j ü s oj u i3 3 s s o sso j § uoo j s s S jrj
s q j o p i [ . \ \ s z i s s jB J s p o u i R j o s d n o j S o ju ¡ X[jRUOSJsd g u u s q j E S J o j puR S u q p A E jj jo j u is e is
- n q j u s j p q j j o [u s p p o o S b js o [ /íp B n p E jS SABq s j s i j u s p s jB qj Xea \ e q o n s u i p s S u E q a o s p j SEq
‘B ip s u i p u ti s p o j u o ij B a iu n u iu io a u i q jM O jS u i o o j q s n u i j o s o u s n g u i j s p u n oqÁRiu ‘u o ijb iijis
s iq j p u y
s s n § B S i p a u S p j o j j p q j A'q p s a B j jB qj q jiM jE a iju s p i Á p A ijE jijE tib js b sj jb u o ijb
- n jis s q j p s j s j u s 3At;q s j s i j u s p s puR 3 3 U 3 p s q a s z Q s q j u s s A \j s q u [ JE [naijJE d u i s j s i j u s p s
s a E j j n s puE s u E p ij d o u o j j a s p jo j p s z i u n S j o s S u i j s s u i j o puB j b j s u s S u i S u p u s j s j u o a a y ij u s
- p s j o s j s q d s o u ij R s q j S u i p n p x s j o u ‘S u iS u E q a jds>[ si;q S u iq jX jS A S s s iM J s q j o ‘j s a s m c j j
-3A 3 Á E p s j n q x s q i sr [p M sr s n a s q j e q j r e - u s d o s / R p s a n j s q i ‘u u ; a ij s t u a r a o s u i j s u u i p r
qjiA \ p s s o p d u j gui>pB A \ s u o o u j s j j e X E p s s u p s M 3 MX qaE 3 s s j n u iu i á j s u i u j o s> [30iq u s j j o
P3JSISU03 SÁRA\|R 3|n p 3 q O S 9 q j U3A3 pUR (X jJ^ OJ UOIJBdpiJiR d 3 q j SJIUljl /([JÜUOIJIPCJJ q o iq M
JO 3ZIS 3 q u UlOOi 31URS 3 q j UI p 3 J 3 q jB S 3J3M S]U R d pi)JR d 3 q l p u y
3 p is 3>|R| B JE u o i s n p a s
u i s p u B jq S iq s q ; j s p iu i p s jE n jis ‘s j j u s d g u q o o q a s puR p j o q s iu r s 3 ip u i 33R [d >|ooi j u s a s
a q i u j n j p j iq j sji u io j j S u ij j r is ‘s j u B d p ij J c d 3 ip j o s u o i s s s j d u i i 3[qR )s 3A J3S3jd o j j s p j o
u ¡ 'S u iu io a 3J3A\ s u o s j s d j q S p o j d n j o s d n o j g u s a s puR s s u i i j Xu b u i p s u s d d B q /C[[B3J s i q x
• s ju s p n js jB JO jaop j p q i q jiA \ j s q j s S o j s u i o o o j r n s q j S u ij j o d d n s ‘3 3 u s p s j o q o u B jq u 3 A i§ s q j
- s i p X [3 q q n d j o p E 3 js u ¡ 'u o i j b s iu e § j o j o s j u s u i s p iB n s n u n js q j E j o ju i j p s j i p a j o a fo jd u o ij e u
- p s s p s j d P3JE3J3 OS ' s s u i p u n o j i n s |U BSB3|d UI SÁBp 331U JBJ3A3S J S q js S o j p u s d s OJ ‘JSB3J jo u
X [3 jn s jn q js b | jb pu B ‘s j3 ,w s u e p3np3|M O U >|3B A p jR iu s u io iu jn o q B u B q j s u o p s í n h p s A jo s u n
jn o q B 3 .10LU jn q ‘3 S j n o 3 j o ‘s S u i u e s u i a S u B q o x a o j ‘X jiu n u iu io o 3 q j o j u i u ia q j s j b ij iu i o j j s p j o
u i qjiAV q jn oX p s o i s i u a /([m bu S u u q o j 'X[.iR[nS3.i j s q j o qoB 3 33S o j ju b m q a iq M j o s j s q u i s u i
s q j X p u iB j b u i sb j s n f s u i3 [ q o j d j B p u iis q jiM S u i[E 3 p a s o q j js q j B g o j sb m u o ij u s j u i tq jJ iq
s ji j o p o u s d a q j S u i j j g s q s o u s q puB |B u o ijip E jju n s e a \ S u i j s s u i a q j S u iX p s p u n B 3pi s q x
3 3 tip 0 jd XJJBUIJOU 3JIJ U BUinq 3 q j JO SJR3Á U33JU3A3S JBqM UEqj
3JOUI X p j n s - s p u ir n j n o p a ja jp i 3ABq q o n u i M oq p u y ¡jEqj 3 3 u is p s á u R q n SBq p [jo,w [b do[
jn o qan ui a \o h
s s u o j u s u i u i o j d ,\|[E 3 J jn q pBOjqB u .is js s m a q j u j o j j s j s s n S 3A q UBqj s j o u i
JOU qjIM puB UOIjn[OA3J J3A py\ 3 q j 3 J O p q S>[33A\ U3J ‘SUOIJBJ33dx3 qjIA\ p 3|m J3A O 3 J 3 q d s
-OUIJB UB UI paZIUBSjO SBM JBUIUI3S JSJIJ 3 q X :p u ;u i j n o SSOJO SJ3JJBUI 3JOUI U3A3 s p u s j x
JU333^J „ 0 1 JO 3SB3 3 q i U] 'EJ3J33 J3 ‘3UIIJ UI JS3J3JUI S iq j JO J U 3U ld o p A 3 p pUB UOIJipBJJ S iq j
S u in u ij u o o u i X j;u n u iu io D ju E A s p j s q j j o j s s j s j u i ‘j u s a s jB |n 3 ijjB d s iq j j o u o ij ip B jj ‘S u ij s s u i
JO pu i> | USAlS B JO SSSUJtlJSSn S>[Il SStlSSI Su iJE [dui3JU 03 puB guiJJIS SUO SSJJEUI X p j n s U0ISSS3
-3tlS UI q ju a j XpB3J|B S u i j s s u i D giJU S p S B JO s g u i p s s a o j d 3 q i OJ SDBJSjd B 3JIJM OJ >|SBJ 3 q x
a o v ja a d
J. E. B arth and M. S. B ronsgeest
A. B otm an, J. J. L. M ulders, R. W eem aes and S. M en tin k
P. Č ernoch and J. Jirák
A. D elong and P. Štěpán
L. D upák and M. Z obač
M. Escher, M. M erkel, J. R. R ubio-Z uazo and G. R. C astro
L. Frank
P. H orák and P. S chauer
M. H ovorka, L. F rank, D. V aldaitsev, S. A. N epijko, H. J. E lm ers, G. Schonhense
H. Q. H oang, J. W u, M ans and A. K hursheed
M . Jacka and A. P ratt
P. Jánský
J. Jirák, J. L in h art and V. N eděla
fA0J3[[nj/\[ ] pue jojpuB^w y
w a s a o v n o A m o i h h i n i s n o î h d h i h i v n o i s a o N O iiD g ia a
1 (nj:>{ vi put; uarauiojj in:a g ‘iraSSnjg ura f p\[
S N O iiv y a a a v s n 3 3 0 h d v w s i x v - a a o a o NOiiDaa^iOD
:a D îm o s P M v a a in n w v n i .a o iD s y y o D a io a . v a o a s a
3SU3l[lIOipS 0
w a a d a a A a o s a a - a m i Aa s N o u v a a o s o N o w s v id -H D v a m s
ö n v s c n a ia y v a N a v D iid O ‘S3 a v m N ids a o N O iiV A îiasao
s a o i D a i a a N o a iD a ia
w a s y o a s a v i s A î o n o l l v t il l n id s a o s o u a N r a AVDaa
^ubjj q pue cujoîjod 7
B5I?HPcy 1
s w v a a PM oyiDaaa ö n v n o i n i N o iiD v y a iN i aiM oanoD
SNO>ii3 Haa h u m s a i v x s a o
A iiS N a a i v D o a a m a o d n id v w i iD a y i a a o s a o H ia w
BAOJ3[inj\¡ I puc 5[EAO{vJ q
a o i o a i a a A sa N H O H i-iy v H a a A a 3 H i n i i v n o i s a o O NisssDO iid
EjBjjny y puB EXEJM f ‘Elapse A
SN O iivD iaddv ö n v N o is a a a H i
- h w a s v n O v SdODSOtfDiiM N o a iD s a a o n i n n v d s a v iN S W N o a iA N S
nn H PUB asnoy f ‘Subm q ‘exiling g
N o is a a a v D iid o
N o^ iiD aaa îio a a o H ia w D iv n a a o a v a v iiN a îia a a ia
BAoaa[|np\| j
adODSOiiDiiM N O îiiD a a a d n i n n v d s s h i n i a a o w S N 3 i a a o H iv D
IBuipiz f puB BAoausq g
IM3S A 9tf3N 3 M 03 V NI SNO>IlD3a3
s s m i v a a iM Víiooyd ( N o is s a a v D iid o N O iiiD aaa) a o 3
[11.10 IM Pub EAODuaq g
N o is s a a v o i i d o N o a iD a a a a o a siM v a o o y d
nA O PU« P33qsjnq>i y
( w a a d o i) ad O D S O iom N o is s iw a
N o^ iiD aaa i H o n a - a o - a w u v a o 3DNVF\raoayad a a i v a n w i s
J.E. Barth, M.S. Bronsgeest
Delft University o f Technology
Lorentzweg 1, 2628 CJ, Delft, N etherlands
[email protected]
To characterize electron sources, usually the reduced differential brightness
B.lfr = — lim li m --------w
y -.oSASO
( 1)
is given, with current SI, area SA solid angle S O , and extraction voltage V.
However, when electron sources are used for probe formation, the differential brightness is
not the relevant parameter:
In a scanning electron microscope the virtual source is imaged in the probe. The virtual source
has a bell-shaped intensity profile. The differential brightness, being the brightness at the top
o f the source intensity profile, is therefore not applicable.
W hen sources are used for probe formation the only brightness with a real physical meaning,
is the reduced brightness, as Crewe has pointed out [1]:
B=Hrv-d l
where / ’ is the angular current density, which is taken to be constant over a finite solid angle,
and dv is the diam eter o f the virtual source. However, Crewe does not give a definition o f that
‘diam eter’. In order to have a practical brightness a practical definition is needed. A measure
that assumes no particular shape o f the source intensity profile is tlie diam eter containing a
fraction FC o f the current / 'f l coming from the virtual source as seen from the extractor.
To complete the definition a choice has to be made for the value o f FC. For a number o f
reasons we propose a fraction o f 0.5. We consider this value most likely to be generally
acceptable to the charged particle optics community.
W e have derived the relation o f the practical brightness to the differential brightness for the
particular case o f the Schottky source.
For Schottky sources the reduced differential brightness is (e.g. [2])
b t = ——
V n kT
3 H 3
‘. U ° I j O
S I e i7c 0 ( 9 0 0 c ) 6 6
[ d d y T ‘q P B g ' 3 7 p u B u s j m z s g ' n
688 (6 9 6 1 ) Z a
‘l i n - O i
's ^ M d T 'J a i s J O M .
P 3 \ S 3 | j d o o p i l a d p a S J t J i l J j o > [ o o q p ir e H , u i o i
[£ ]
' d ‘( ¿ 6 6 1 ) S S 3 J d
q j ‘s m s j j
a 'V
[ f ] • su o p B [n o iB 3 s z i s s q c u d j o j p s s n
aq UB3 s s a u iq S u q
lE iju a is jj ip
a q j )E q j u o ij d u m s s B
]B n sn s q j u io j j â jju b o ij iu S is s j s j j ip
sjq x
' I3 W \ = 8
s; s s s i q q S u q ]B O | p B jd s q x
' 3 J n iB J 3 d u i 3 j j o j £
[ j ] s p j c u d X } is u 3 } u i U E is s r r e Q b S E q s o j n o s ]B tiy ;A
p ir e ‘j u e j s u o s u u E U i z j j o g
s q i jy ‘X j i s u s p
ju su n D
s p o q jB D s q j f q j iM
A. Botman, J. J. L. Mulders, R. W eemaes and S. Mentink
(A.B., S.M.) Philips Research Laboratories, High Tech Campus 34, 5656AE Eindhoven, The
Netherlands. Electronic Mail: aurelien.botm [email protected] .
(R.W .) Philips Research Laboratories, High Tech Campus 11, 5656AE Eindhoven, The
(J.M.) FEI Electron Optics, Achtseweg Noord 5, 5651GG Eindhoven, The Netherlands.
The technique o f electron-beam-induced deposition (EBID) [1], when performed with organic
precursors, typically results in a relatively low metal content due to the partial decomposition
o f the organic precursor, leaving carbon rich remnants in the deposition [2]. Solutions
investigated elsewhere have included annealing in-situ after [3] or during [4] deposition, or
deposition in reactive environments [5], These two approaches have apparently not yet been
We described a method applied to noble-metal structures deposited using EBID, which
consisted o f a post-treatm ent step o f heating in a reactive atmosphere o f oxygen, whereby the
amount o f carbon in the structure is strongly reduced [6]. As a result we have been able to
increase the purity o f platinum deposits from 15 at% to nearly 70 at%, and gold similarly
from 8 at% to nearly 60 at%.
The resistivity o f these structures has also been improved by up to 4 orders o f magnitude, to
achieve (1.4 ± 0.2) x 104 // in the case o f platinum. Furthermore, STEM analysis of
the treated structures has shown formation o f crystalline material (Figure 1).
Y u o ij o . w d 3 . id u i i d u o s n u D U i ) ‘^ u o q is o d s p p s s n p u j - u iE s q - u o j p s p j s ij b s s j n p r u j s p [o 8
p u e U inilIlB |d JO UOIlBDIJUnj,,
‘ ■)¡3 7 's ú i j d
‘->JUI1U3]/VI 'S pUE S3BUI33yV\
[ 9]
(S 66 I ) ‘0 8 0 £ ;( 9 l ) 9 9
7 d d y ‘^ ju su iu o jiA u a s a ij o b s j b u i s s jn p r u js o u B U p j» S j o u o p is o d s p ureaq
U()jp.T['J„ ‘U()iqÎÎIJ\\ •§ >JJR^\j puB ‘SJ3P<] [| J0l[doiSUl[3 ‘BpBfaj, JSlABf ‘qO[OJ yaqiv [si
'(9661) ‘S0It’:(9)t’X ‘9 ~!ollUJSl PS DDA f ‘„s-iruBJsduisj poseaioui SuiABq ssjdures
uo uoijisodsp psonpui uieaq-uojpap ,\q psonpojd sjajjiura uojpap ppij joj sdiyndns
puB ‘S3JIAY ‘sjop 3Aipnpuo3 „ ‘JsqsM JAI PUE ‘BÁB» y ‘j3[ss3oqos 3 ‘sdoo>| j /\\ j{ [^|
'(SOOS) ‘IT(Z)£Z a '¡outfoaj, PS ™A T ‘„uopisodsp I(|uiB3q-uojp3[S
p33npU[-lUB3q UOJP3J3 3JE3S-J3PUIOUBU JO JIUiq UO|)I1[OS3J 3q) SuiqSBOjddy,, ‘J3IZOJ3
■y J3 P d puB ‘)inj>| J3J3IJ ‘UsSbJJ ~/\\ S[[3UJ03 ‘U3J3UIOS UBA q o g ‘djOQ UBA J IU3|[I/W
•(S003) ‘6 Ç i z - i s i r :(s)ez
‘ff lOUljJdJ P S DVyi f ‘^UinUIJBld JO uoijjsodsp JOdBA |B3IUI3q3 UIB3q U0JP3]3 uo
S3|qBUBA SS330jd JO 33U3tl[|U],, ‘À3>|3BJ ’f '/V\ pUB 'SUB/W
2 ‘§UIQ S u O ^ ‘nSIJtlBSQ ( ]
Àq S 3 p O jp 3 |9 JB ]|ld d s S UIU
(S 0 0 3 )
‘¿ o e i-c o e i :(Z.)S
‘„ u o p is o d s p
9001 SPU^ 1 luaoa#
P. Č ernoch2, J. Jirák1' 2
'institute o f Scientific Instruments ASCR, Královopolská 147, 612 64 Brno, Czech Republic
■^Department o f Electrotechnology FEEC BUT, Údolní 53, 602 00 Brno, Czech Republic
[email protected]
The gaseous ionization detector [1] is often utilized for signal detection at a higher pressure in
the specimen chamber o f the environmental scanning electron microscope. We study the
properties o f halved segmental ionization detector with an electrode system consisting o f four
concentric electrodes divided into left and right halves, as seen in Fig. 1.
Each electrode o f the detector can be connected to one o f the three following options: a
potential o f 0 V to 500 V with signal detection, a potential o f -500 V to 500 V w ithout signal
detection or a ground. The electrode system connection at signal detection by the left smallest
electrode (Ai ) at a voltage o f 350 V while the right smallest electrode (A r) is at a voltage o f 200 V and all other electrodes are grounded is illustrated in Fig. 2. For this electrode system
connection, electron trajectories in the electrostatic field in a vacuum simulated via program
Simion 3D 7.0 is shown in Fig. 3a and the image o f the sem iconductor specimen is shown in
Fig. 4a. Likewise, the electron trajectories simulation and the specimen image at the electrode
system connection at signal detection by electrode A l at a voltage o f 350 V while electrode
A r is at a voltage o f 350 V and all other electrodes are grounded are shown in Fig. 3b, 4b.
In the case o f a negative voltage on the non-detection electrode, signal electrons emitted from
the specimen are repelled by the non-detection electrode (A R) and attracted by the detection
electrode (A l). This effect is observed as a brighter specimen image w ith sym metrical shadow
and diffusion contrasts in Fig 4a. In the case o f sym metrical voltages on the detection and the
non-detection electrodes, signal electrons emitted from the specimen are divided
sym metrically to the both electrodes A L and A r , therefore, the angle o f the signal electrons
detection by the detection electrode (A L) manifest itself, as seen in 4b. The simulations
correspond to the obtained specimen images, as illustrated in Fig. 3a and 3b.
According to the electron trajectories simulations and the specimen images, it is seen that the
effect o f change o f contrast, shadow and diffusion, is possible to achieve not only by a change
o f detection electrode position but also by a change o f voltages on non-detection electrodes
influencing secondary electron trajectories.
This research is supported by project G ACR No. 102/05/0886 and project MSM 0021630516.
is jip a g
o '¿ 0 £
6661 ‘3 3 1 ‘°MBPI IXAVa
uoTUJTS :X jojB.ioqBg jB ju a u iu o J iA u g p u e S u u s s u i S u g j b u o ij e n oq B p j
K)81 - 96¿ I d t 00 c ‘Se iOA
S O ts X q j J O [B U JnO f 'J O lIO B d tQ SB£) p 0 1 R ip R J J [ - lU B 3 g - U O jp 3 [ 3 UR Ul U ()¡lRU iqU K )30>[
u o |- u o a p 3 |g j o uoijB oijiJU E nf) :']A¡ V ‘P l^ u o a ‘' 1 ’0 ‘[ 3 i q i ‘ y
G ‘s [s ; u e c [
-q jo x
[z ]
6 - 8 /. 917 I 0 -C I -0 N 9 S I ‘ d £01 ‘0661 ssaJd o iiu sp B D y ‘X su p K s s d o o s o j o i w u o a j3 3 |g
S u iu u b o s [B ju su iu o a iA u g o ip u i s o i a s q j o j o s j s q sn o a sB Q j o X a o sq x : G
X000C '^Blu ‘Bd 0 ? ¿ ‘1U9UJUOJIAU3 sjodBA
A 0 S £ J0 A 00c" J °
s 38 bj| oa
:¿\ 0Ç£ j o
jsjba \
O ‘s o jb jiu b q
[ []
íp sp u n o jS S 3 p 0 jp 3 |3 J3l[JO ‘^ y 3 p 0 jp 3 |3 UO
s S bj | oa je
l y sp o a jo sp Xq u o ijo sjsp [eu S | s íajBijsqns
U03IJIS UO UOIJBZ1JJBJ3UI UiniUIUmpj JO u s u ip s d s UO SJSBJJU03 UOTSnjJip pUB AVOpBqS -q ‘Bf? 'S lJ
1 \ i I
i j
I 1
' r
. '
av u°a ose ‘"'v uoa ose(q)
S 3 3 jg 3 p 081 ° l 0 SUOJJ33J3 JO 3[3uB UOIJDSfojd ¡A3 S XSJ3U3 UOJJ33J3
0?£ J0 A 00Z" JO S3SbJ[0A put;
:p3purlO jS U 3lU I33ds PUB S3pOJJ03[S J3l[JO í « y 3p0.1)33[0 UO a
' y 3 p o jj3 3 [3 u o a
0 S £ J ° s S b j[ o a j o j umriDEA u; u o ijB [n u iis s s u o j s s f B j j u o j p s j g
ä y
uo A 0 ǣ
3 p O JJD 3 [3
uo a 0 S£
(q )
A 003-
PU B n y
u a w io a d s
q ‘B£ S ig
s p e u p s p
3 p O JJ3 3 [3
3 p O J J 3 3 |3
UO'A 0 S£
0 0
uauiiD ads
S3pOJJD3[3 jsqjo [[B í*ly 3p0jj03[3 uo A 00Z~J°
s3bj|oa ! a 0S£ jo 38bj[oa jb i y 3poaj33|3 Xq uoijosjsp
J O J D 3 J 3 P U O l J B Z |U O |
[B ju a a iS ss p s a | b h ' 1 'S | J
A. Delong, P. Štěpán
Delong instruments, a. s.
Bulharská 48
612 00 Brno, Czech Republic
Understanding the three- dim ensional structure o f proteins and other particles (viruses,
ribosomes, enzymes, DNA) is critical in determining how they perform their biological
functions. The usually used technique o f X-ray diffraction requires perfect crystals with
approximately 101- molecules. Electron microscopy should be superior to X-ray diffraction
and 200-400kV transmission electron microscopes (TEM ) are used for obtaining satisfactory
information. But today’s m icroscopes must overcom e various obstacles connected with the
properties o f biological objects. It is first o f all very low contrast and sensibility to radiation
damage. The result o f this fact is the necessity o f elaboration o f high quantity o f very noisy
pictures ( 1- 100k) to get satisfactory results.
It seems that the low voltage transm ission electron microscope developed specifically for
application in biology will be more suitable for solving from above introduced task first o f all
due to a much higher contrast. We shall compare the main parameters o f two different TEMs200kV TEM dedicated for biological applications and a 5kV low voltage TEM.
The usual objects are so thin that all impinging electrons in both microscopes are transmitted.
We will now compare how the standard object (lOnm thick specimen com posed from material
with R= lg/cm and a diam eter greater than the resolution o f the microscope) influences the
trajectories and the energy o f transm itted electrons. The results o f calculations are presented
on the figure 1.[ I ]. There are in principle three kinds o f electrons in the transmitted beam
electrons which were transm itted w ithout any interaction
electrons which were elastically scattered. Their trajectories were changed due to deflection in
electrical field o f the atomic nucleus and its electron shell. Electrons falling upon the rim o f a
physical aperture are excluded from the beam and appear to have been absorbed. This
electrons produce contrast
Some electrons are deflected by direct interaction with atomic electrons. A measurable
amount o f energy is transferred, and the resulting deflection together with energy loss is
called inelastic scattering. If we compare the results o f calculationsTn both microscopes we
see that the contrast and the radiation damage are more favourable in the LVTEM.
The resolution o f 200kV TEM is 0.27nm approximately. The resolution o f LVTEM is 2nm.
The very low contrast o f biological objects in 200kV TEM must be increased by using the
phase contrast. The necessary defocusing decreases the resolution to the value o f 1 2nm .N evertheless the correction o f both mean aberrations (spherical and chromatic) can bring
new possibilities for LVTEM. We decided to correct the spherical aberration by means o f
hexapol corrector and the chromatic aberration by means o f inhomogeneous Wien filter. We
used the proposal published by prof. Rose 17 years ago.[2] This proposal was never
implemented, although its advantages are obvious. It is the rotational symmetry o f both
correctors and minim um o f multipoles. Another advantage is the possibility to use the Wien
filter alternatively as an electron energy loss spectrometer (EELS). The cross-section o f the
'UvZ (fr00Z)66 XdoosojoiuiBjJin : « ! W 3 'a ‘SueAT ‘Anunud'-n
16 (066l)£‘ť8 >RJdo :3so^H
d 9t?6 ‘59613(3 ‘uoi3uR|SBAVÂSo[oi|n;f|
j o a jn i|)s u [ S3DJOJ p a a u y ‘Â d o o s o jo ij^ u o j j o s j g 3aijbjijubii£) : j q ü a ‘j ' ! ) ‘j s p p z g
[ []
spuB uisp j o X ju ofeu i j o j q S n ou s sq
u eo qrm |\\ u i u £ o j o u oijrqosai 3qj qoB.u oj sjq isso d oq 01 silicios j[ •uoijeuiuinjii auoo A\o[|oq
aqj osn oj i o p u e sunj OAijoofqo oqj j o su o isu su iip asRSiosp oj si Kbav J3[duus y urejjsoun
ss jn o o j o si j[tisai sji ‘p sjB d sjd si ju su iu a d x a s q x 'uoijrqos jss q aqj s q oj su isa s ssio d ijjn u i
j o SUBSIU Xq suojjEJjaqE SA ip afq o j o u o ijo s jj o 3 'ooj u oijrqossj ss e s jo u i oj (uiu i )sp 3 fq o uiqj
ÀJ3A JO 3SE3 sq j UI U3A3 X|JUB3|JIUSIS ÄJ3A p3SB3J3UI JSBJJU03 3qj U3qM uoijbijjis žq j ui 3iSo[ sq
S u isn jnoqjiM s p s f q o [BDioO[Oiq j o
qSiq qsB3J
X jijiq issod sji qjiA\ [A I3IA 1 3ll l
u o isn p u o 3
■3jnjEJ3ji] sq j ui
p B j siqj JOJ uoijBpunoj
|i:oijo.iooqj o n T M 31A T u! P s S euii uoq w su oiu iood s uiqj j o ssijia d o jd suies aqj p a u o s q o
3AVn snoiA qo sjoui s u io s s q uoijBJJsqE 3ijBuiojq3 j o oijsusjSBJEqs sojBq sq j puE u o ijn jo ssj j o
s s o | B s; sjs q j S3[diUBS >[0!qj qjfM uoijEjJsqB oijE w ojqo j o [BOidXj 3 jb jBqj so|E q DijsusjOEJBqo
oqj A\oqs jou op puB jadreq s sjb s s a œ p iq j a\ oj je saSBUii sq j ‘À qB O ijpsds 'SuuajjBos oijsB put
3l(| Ul 3SE3J33p JUBOIJIUSIS B SI 3.I3l|J jBqj SUI33S JI (lllU()£-() | ) SU3Uipsds Uiqj À|JUOI3l|jnS JOJ
‘3|duiBX3 JOJ 'pOSBflJOap SI s SbJJCA 3qj SB JSJJip SUOIJ3BJ3JUI 3S3qj A\0q jnOqB S3I1[0 SAIJBJIJBnb
3 u in 8 u ju i u 3 ai 8 3ABq JA I3IA 1 3 lll u! P s S euii s a l u r e s j o su oijB A jssq o 'XSjaus u o jjo a p
u o q ssn b |Bju3uiBpunj y K '[g] X iu u in jQ 'j'q j o a p ijjB sq j u io jj 3 j p oj
p[noA\ j \\u j o j s j i j
's p p ij 0IJ3u8bIU ÁBJJS puB SUOIJBjqiA OJ 3AIJISU3S JOU 3J0J9J3qj puB (lUUI()8[) jjoq s
ÁJ3A si UJ3JSÄS |B3ijdo u o jjo a p sjo q s q x ' z sjn S ij sq j uo p sju ssa jd si |/^ 3 1 A 1 p s p a u o D ° 3 S3
Figure 1: Quantitative distribution o f electrons which transmitted the object with p=lg/cm 3
and thickness t= 10nm. 1 - elastically scattered ( a > a opl), 2 - not scattered and elastically
scattered ( a < a opl), 3 - inelastically scattered.
Figure 2: Cross-section o f the electron optical column o f CsCc corrected LVTEM
900Z sPugJI !“ » ) /
L. Dupák*, M. Zobač
Institute o f Scientific Instruments AS C'R, Královopolská 147, 61264 Brno, Czech Republic
* E-mail: [email protected]
Electron beam m icrom achining is a well-known technology used in various industrial
applications [1], It is based on the evaporation o f material by intense electron beam [2]. The
construction o f the e-beam m icrom achining equipm ent may be similar to an e-beam welder,
but it needs a high performance electron optical system to achieve very high power density in
the focus o f the beam (usually 100 times or higher). W e have studied ability o f a common ebeam welder to machine the materials like quartz glass, ceramics, plastics and metal sheets.
The purpose o f our experiments was to
create holes as small as possible [3],
Pulse count
In micromachining, a short-pulse beam
Depth (mm)
2.78 3.42 4.06 4.19
is used instead o f a continuous beam. If
Diameter (mm) 0.32 0.34 0.39 0.43 0.48
a continuous beam is used, a lot o f heat is
Depth/dia ratio 8.7:1 10:1
9.7:1 9.8:1
supplied to the surrounding material. This
Table 1. Dependence o f hole dimensions on pulse count
has undesirable effects like growth o f the
(beam current 1 mA).
diameter, deformation and material tension.
In our first experiment, we have followed the growth o f the hole in the material exposed to
a series o f the pulses (see Table 1). The highest growth o f the hole dimensions was observed
after the first pulse. The effect o f the following pulses was decreasing with each step.
The effect o f focus point position was determined in the next experiment. The beam pulses
were counted until the beam passed through 0.9 mm thick quartz glass slide. The lowest
number o f the pulses was needed when the beam was focused exactly on the surface. But as
seen in Fig. 2, the holes created by defocused beam had higher apical angle and lower
minimal diameter.
The position o f the minimal diam eter is not located on the surface, but a small distance
from it, sometimes more than 0.1 mm. This is probably caused by a burst o f the evaporated
material. The surface tension o f the melted material then forms the narrow neck.
We have been able to create openings o f 5 to 17 (im minimum diam eter (measured by light
microscope and scanning electron microscope - see Fig. 3). H owevei^reproducibility o f such
(UdidUÁDip ui mm p pun UiUl g)
sdjdlfds D1WDJ3J U1 SdAOOJQ
■sut 0Ç ySuai s s jn j ( z qv¿ m so)
lUdAAUD wvdq uo dsjnd djSuts v Jo siodjfd Jo doudpuaddQ ~p S j j
[q o 3 Z 3 u i] o i u g ç o o £ 'S issq j e iu o jd iQ :>|BdnQ q
• jjB g jjn js
pun u js g ‘§B|j3y\ 8ba\]]bh ují «neqsspun>i s ip siu ip sx » 3 e[j,i(\ (ZL6\) M)I ¥ 3H
"oqio>[ 11 [ sríEjg ‘usjqEjjsusuojoqd pun u suojj^sjg qojriQ SBjjqBjjojs>|j3M :jsssiy\ y
'(900Č I« d y ) /sp'qs-p[BMJS<?[3jsMMM//:djjq
■ujuj z
'lUdAjno uwaq uo suoisudwtp djoi/ Jo DDUjpuDddQ 7
1Z 0
ç'z ‘£ s js m sSujjjno sqj
JO SJSJSlUBjQ JSJSUIBip ui iulu ç jo/puB ÿ
sjsm ssjsq d s siu ie jss s q x '( 9 'S ij
6*8 8’9 0|)BJ
ss s) psrqS pue p sjjssu i si /CiBj|idEO
s r o CIO (iulu) J3J3UIÜIQ
JSSJS SSS[UIBJS [[BM-Uiqj B ‘sa o o j§ siqj
(uiui) ipdsa
t i l Z80
o j iq s jsq d s diuibjso b ui sa o o jS jBjnojio
(VUI ) s!
-3 posn 3A\ qoiqM ui uopB3qddB J s q jo
•qjSu 3 | sjí Suo]B p sjn q u jsjp KjjBinSsj s 3 |oq g | p 3 ||u p
X||njSS333nS 3/\\ '[|BM >(3iqj UIUJ ¿ Q qjlM J3J3UJB(P Uf UIIU i ‘3llO[ IUUI
091 ‘sq n j ssb | 8 zjjBnb si ssiA sp siqj j o p a d juBjjoduij 'íGjauiojjDsds
uoijdjosqB 3IUIOJB jo j JszauojB-jJinuj zjjBnb j o v‘ uoijsrujsuos
SBM X§0[0uq33j pSJUSSSjd 3qj JO SU0lJB3l[ddB [B3IJ3Bjd 3qj JO 3UO
Vd ui uirl o s i puB (ç ’S ij 33s) uirl os
qSnous qSiq
V d ) s p iu re X p d puB
(l!°‘9 /9
UI p o j o q J3J3U JB ip 3 [ 0 q JS3|(EUJS 3 q [
SBM 3 d
SU iqSE U I O) ÄSB3 3J3A \
3U 3 iX q j3 Á |0 < j 's s i j s E j d j o S u iu iq s E U i
u o jj3 3 ]3
sjnsssjd sqj dss)|
Á jq ¡q E d B 3 p s A O jd u i i q jiM j u s u i d m b s i B j u s ú l u s d x s s q j
js s j
o j A | u ib u i
‘p S U U O JJS d SBM SJBIJSJBIU 3 lJS B |d JU S JS JJip /ÍUBUI j o 8 u q | u p u i B s q - g
' ( u i u i 9 ’0 s s s u > p i q ] j s s q s ) u i m u B jíj u i
U lil
p u B UIUJ
ÇO'O S S S lD |O iqj
• p S 3 )3 E J3 S 3 [ d u i B S 3 U IO S ‘X ] 3 J B U n j J O J U Q
s |[ ij s j
U B qj SS3I J3 J3 U lB ip
p s u j 3A Bq 3jW
3 p i S J IX 3 3 q j U O U I t I 0 0 1 U B q j SS3| p U B 3 p iS
u ir i o ç I S 3 jo q ¡| u p o j 3 jq E s j s m s m
s u iq s B u i o j X s b s
'S u i u s d o s q j
J 3 3 q s ) J33JS SSS|U |BJS UI U irl
B J O S 3 |O q 3 JB 3 J3 OJ S jq B S u iS q ‘U iniU B JlJ p u B J33JS SS3|UIBJS J O
jS B d u ii u iB s q s q j u o s p m
> (3 iq j u i u i 9 ’o u [
s iu ib j s s
£O zI V
3 J ! q M j o 3 | d u iE S
3 J B ‘X j i A i j s n p u o s | B U U 3 q j m o ] J i s q j o j s n p
' o o j ‘S3SB 3J3U I s n s o j 3 q j
3Z |S lU B sq s q j 3 s n B 3 s q ‘(z
3 3 S ) p s q d d E s | j u s j j n s u iB s q j s q 3 | q u s q M p s o n p o j d
s a u o s n o i A S j d s q j U Bqj
ju a n b s s q n g
Sj e | s ]o q
s § u m s d o J sp iM
‘s s i u i b j s ^
SW®! Pu®P ' § ' d
puB j s d s s p s q x
p u B j s i j j B q s i q j q S n o j q j s> |E S jq s s | n d
b s s jb s jo
u s j j o Xj s a | b i j s j e u i p s j p u i s q j s s n B D s q
m o
s s |o q
|[ b u is
M. Escher1*. M. M erkel1, J. R. Rubio-Zuazo2'3, G. R. Castro" 1
'Focus GmbH, 65510 Hünstetten, Germany
'SpL ine, Spanish CRG beamline at ESRF, Grenoble, France
3ICMM-CSIC Cantoblanco, Madrid, Spain
* [email protected]
Recently the interest in hard X-ray photoelectron spectroscopy (HAXPES) grew with the
availability o f high brilliant synchrotron beamlines [ 1].With the longer escape depth o f the
high energetic electrons HAXPES has the capability o f prob­
ing bulk properties, buried layers or interfaces down to some
tens o f nanometers, which is not possible with the surface
sensitive standard XPS.
We developed an energy analyser, FIV-CSA (High Voltage
Cylinder Sector Analyser), for electron energies up to 15keV.
It is the first commercially available analyser for that energy
range. The analyser is based on a cylinder sector with 90° de­
flection [2], 300 mm slit-to-slit distance and an entrance lens
with 50 mm sample distance. The result is a very compact
design o f the analyser that is easily integrated into a multi pur­
pose experiment with different techniques. A low noise / low
Fig.I: Photograph o f the H igh
drift electronics is capable o f continuous energy scans from 0
V oltage C ylinder Sector
to 15keV with non-linear lens curves. The first analyser has
A nalyser HV-CSA
been recently installed at the Spanish CRG beamline at the
ESRF (Grenoble) at an end station where simultaneous surface
X-ray diffraction is possible [3]. Fig. 1 shows a photograph o f the analyser.
Entrance Lens
The entrance lens retards the high energetic electrons with a kinetic energy Ekjn up to 15keV
to the analyser’s pass energy Epass (typically
20..100eV) which results in retardation ratios
R = E hi, / Epos* UP to 1000. To accept a large
phase space at the analyser’s entrance slit
E , Mss A A n a l y s e r = E tlA a mpA
n , Ple - t h e
d e s ig n
goal was to have a large lateral magnification
M = j A : /■^sample which is also well adapted
to a focussed synchrotron beam o f typically
50)im diameter.
The entrance lens consists o f a unipotential
lens near the sample followed by a three ele­
ment retardation stage. Its properties were
calculated with direct ray-tracing using
M ognificotion
Fig. 2: R ange o f useable m agnifications and
retard rations o f the entrance lens.
ZL-V 9 (SOOc)
LPS ‘V u in jjs u i - p n u ‘o jjs b 3 - y o ‘o z e n z - o iq r r a
S6 ( Z L 6 1) efr u m jjs u ] - p s A oy 'X a js i^ f
8 £c-I (SO O t) LPS ‘V u in jjs u i - p n u u i doqs>|JOM sadXVH s q iJ 0 s S u ip o o o o jj
[f ]
[3 ]
sqjp|A\ \Sj3U3 pajnsEouj o q jjo oouopuodop AS.iouo ssBd :jqSu
s' 1,d3 juo jojjip joj un3-o a 3>ISI e uiojj su o jjo o p POJOJJBOS A||B0ijsbi0 j o E.ijoods :jjo |:j 7 3 ij
(as) Äßjaug ssed
OPZ 0ZZ 00Z 081- 09L OH 031- 001
[A3] Äßjaug oysupi
A99 0 M1PIM
ojsuuju! p a tu n sse
(0OU6J1U3 ) 6=P :s|i|S
u n ß -a 3 3 H ä luoj^ s u o j p 0|0 0|is e |e as>IS l
jasÁ ieue Á6J0U0 m6¡h :00£VSQ-AH
Linß-Q3 3 Hd Luojj suojp0|3 oijSB|aAsMSljssÁ|bubAßjaua qßiH OOSVSO'AH
■soiuojjoop ouibs oqj 3u isn UMoqs oq ppioo A 3m 0fr J ° u o ijti|o so j X3jouo josA|bub ub uoij
-bjioxo [0[j 3 u is fi 'A 3 S’0 xojddB j o un3 uojjoo[o oqj j o qjpiM /¡S.iouo oisu u ju i oqj oj pojoujsoj
ojb puB jnoiABqoq jb ou i| pojoodxo oqj m o[|o j s>[Bod oqj j o sqjpiM /¡S.iouo o q x 'p
ui UMoqs
o.iB sqjpiM >|Bod oqj piiB soiS.iouo ssBd ju ojojjip j o j pojnsBOui s>[Bod oijsB |g 'u n 3-Q ggH > l
A 3>1SI B J ° s u o jjo o p pojojjBos X|[Bo;jsB|o S u isn pouiBjqo ojom V S O ’A H 3l|l
sjjnsoj j s j i j
síin sa y
■pssn u n 3-o oqj j o suoijBjjoqB jods
puB suoijBjjoqB su o | u io jj s j p s o j su o iS o j oojqj oqj uooMjoq uoijisubjj q joou is o q x 'su ojjoo|o
0 | 1 SB|0 a 3 >I0I 3 u isn 00Č .1° uoijBpjBjoj b jo j oouopuodop u oissiu isu b jj pojnsBOUj oqj SMOqs
souBjjuo oqj Áq jno si Xjisuojui oqj suoij
jn o |A B q o q [Boijojooip p o jD o d x o oqj
-¡pin s o u q p o q sE p o q j
jo j
s u o jje o ijiu S b u i
U M o q s si 1UOJJI1D ojdujBs pojn sE o ui oqj oj
p,i|Bos X j j s u o j u i >|Eod o q x
u o ij b o ij iu 3 e u i suoj j o
U O IJO U n j E SE U O IS S IU IS U B JJ J 0 S X |B U E p O jn S E O J ^ l : £
S lJ
-boijiuSbui o3jb| j o j oouBjdoooB JB]n3uB
oqj sjiu iij oouBjjuo su o| oqj ojoijm juiod
oqj oj dn
-suBjj oq l - ^ U
qjiM o su |jiav u o issu u
zM / } l = J3a"...ü
-bub oqj Xq ftejdoooB o|3u b p q o s p o x y
oqj Áq p ojou jsoj si oosd s u oissiuisu b jj
su o;jboijiu 3b ui
jo j
3 u iju u i| OMj sBq u oijboijiu3bui oqj uo
u oissiu isu b jj su o| oqj j o oouopuodop o q i
•punoj oq pjnoo soS b j|oa suoj ojBnbopB
ooost ■§
00002 q
Z apoi/M
l 9pO(/M
j o o3ubj oqj SMOqs
'S ij '0|3ub uiBoq
puB u oijboijiuSbui
‘u oijisod
sn ooj u oai§ b oj joodsoj qjiM su oisu ou u p
oojqj ui poqjoui s^ uojm on 3u isn punoj
ojom s3uijjos o3b j|oa su o g
Q'Z. U0!U1!S
L. Frank
Institute o f Scientific Instruments ASCR, Královopolská 147, 61264 Brno, Czech Republic
[email protected]
The continuous reduction to feature dim ensions in sem iconductor structures enhances the
need for high resolution imaging o f doped patterns in plan views o f substrates or on their
cross-sections. In this way a tool is sought for checking the critical dimensions (CD) o f the
structures and, moreover, for two-dimensional profiling the spatial distribution o f active
dopants. Various surface-oriented electron microscopic methods have been used with varying
success. The contrast levels are often achieved sufficient for CD measurements but detailed
interpretation o f images keeps being unclear and hence, irrespectively o f some promising
experiments [1,2], any reliable quantification o f contrasts, enabling one to determine the local
density o f a dopant, remains for the future. This review aims at brief summarizing of
achievements in this field.
N umerous o f the attempts performed so far have relied upon imaging in the secondary
electrons (SE). Great majority o f studies reported the p-type Si brighter than the n-type and in
discussions usually three possible emission-controlling factors were mentioned. These include
the SE generation rate, influence o f electrical fields below or even above the surface, and the
passage through the surface potential barrier. All these factors are interconnected through
differences in the energy band structures, determined by the types and densities o f dopants.
Older works ascribed the decisive role to the ionization potential, i.e. the surface barrier
height, combined with above-surface “patch” fields balancing the local differences in the
inner potential [3-8].
Figure 1: SE image contrast between the ptype ( 1 0 19 cm '3) patterns and n-type ( 1 0 15
cm '3) Si substrate, plotted versus the
electron energy, as- measured (two times)
for working distances o f 16 and 8 mm;
reproduced from Ref. 15.
Landing energy [eV]
Further step was to adm it that the surface conditions act not only via surface states but
also the treatm ent and hence cleanliness play their roles. In particular, the carbon
contam ination together with silicon may form a planar junction o f properties similar to a
metal-sem iconductor junction; then the balancing electric fields rem ain below the surface but
still they provide explanation for at least a part o f observations [9 -1 1 ]. This conception was
verified by covering a p/n structure with metals o f different work functions, which led to
mutually opposite contrasts [1 1 ,1 2 ]. A s regards the surface barrier, its filtering action,
similarly as that o f the patch fields, concerns only the velocity component normal to the
surface. This brings not only unequal flows surmounting the barrier but also the angular
a q j S3S.I0A3J a a u a q p u B s u i a j j B d p a d o p ( a d X j - d ) 3 i |j u i o j j u o i p s y s j s 3 S E 3 J 3 3 p qo;qA\ ‘p A 3 |
u in n a B A s q j SA oqB s s j b j s j i r e d o p X j d u i s o j u i s u o j j a a j a j u a p i a u i j o
s i u o iiB A .io s q o o q x
u o i j E i j a u a d b i a p o u |B [ d x 3
[ 0 c ‘6 l ] 3 a b a \ u o j j a a p j u s p p u i s q j o j p 3 |d n o a s s j b j s j o X j i s u s p | b s o | s q j o j
|B U 0 IJ.1 0 d 0 jd /([3S.I3AUI 3 u i 3 q X jI A i p 3 |J 3 J U 0 J P 3 ] 3 /(S.I3U 3 MO|
3l|J JO 33U3nb3SU03
B SB p 3 J 3 S d X 3
SBM P3JJ3 S iq x A 3 0 £ jnOqB AV0[3q S 3|S .I3U 3 JB p s u u i j u o o SBM |B S J3 A 3 i JSB.IJU03 B ‘HE j o
js jij ' i
S i j u i U M o q s s j n p r u j s s q j 3 u i s n A q [ g | ] (IA1333S ‘a d o s s o j a i j A j u o j p 3 [ g X 8 j a u g /v \o g
S u iu u B S S s q j ) 1AI3S A H fl p s d d i n b s g j b ui p s i x u o j . i s d U 3 3 q s e i( [ ¿ [] A p n js 3 q x
p a ja s jS s u aq
p i n o a g s j o s s B a p j jE q j ,w o | o s s a i â j a u a jb S u u 3 jjB 3 s > p E q u o j p a p a q j s ; X j i j i q i s s o d a u o
p auijap X p s o p sjo u i jo s 3 i 3 ads [buSis sq j Su ianp ojd m siuBqaaiu SuuajjBas u o j p a p J3 |d u n s b
3 u i/fo|d u ia ui jqS n os aq jqSjui in o /Cbm a q x 'sjsm su e UEqj su op san b ajou i spiAOjd ‘aiun|OA siqj
ui a.iai|A\as|a paiuasajd ‘BjEp jssm su uaAg -/(|qB ipj jse jju o s aŽEiui aqj p ip a jd oj XjijiqB spjBŽaj
si; in jssaaan s uaaq jou aAEq sa jn p n jjs jo p n p u o a u u a s aqj SuiuiuiBxa jo j |o o j aAijsjijuBnb
p s jis s p aqj jeu Sis aSEiui g § aqj a>|Bui oj sjduiajjB pauopuaui-aAoqB aqj \is q j s 3 o j |v
'[ ř l ] J o ja B j S u i n o j j u o a js B jju o a
jaqjouB se [3A3[ uinnaBA aqj aAoqB oj pajiaxa aauo suo.ijaap j o uoijdjosqB jnoqB Bapi ut!
pajBuiSuo os|B sju a u iu a d x a a s s q x ‘J issj; u oijaunf aaB jjns-qns aqj j o jdaauoa aqj puB ‘sSuijEoa
|Bjaui qjiA\ sju a u iu a d x a pauoijuaui aAoqB aqj uaqj pajEAijoiu Suipuij s;qj_ -jaqSiq sbm jsBjjuoa
aqj uinnaBA „Äjp„ ssa[ s i *iiasjoA\„ je jEqj Xbm b qans ui juauiuojiA ua uinnaBA §U|punojans
aqj uo juapuadap paqsqqE jsa sba\ jsBjjuoa aqj M aqjjnj -uoissiura sdXj-u aqj SuisESJsap
jaXiq añ jE ip aaBds v. sajBa.ia puB a jn ja m js uoijau n f aqj sajB.ijauad jEqj p p i j g j aaBjjns
-aAoqB aqj ui jqSnos sbm uoijEUB|dxa a q x ' [ n ] ASjaua SuipuB| a q jj o y\a>| [ punojB um uiixEui
b SiuqaBaj ‘(z 'S i j aas) sdXj Ks[iuoqx-UEqj3Ag aqj j o jo ja a ja p g § [BuoijuaAuoa b i|j;a\
pajinbaB jEqj uBqj ja§jB | XpuBSijiuSis sba\ jsB.ijuoa aqj ‘ijb j o jsji j -[gi] 1AI3S UB J ° apoui ( 3 3 )
su3| apoqjBa aqj 111 sujajjEd padop aqj SuiA.iasqo uai|A\ p3j33||03 sba\ s a u s u s d x s a\3|s[
X I J 3ÎI u io ij p 33n p ojd sj tuiui i7'()
a\3ia jo p p ij a q j j o qjpiM :(jq 3 u ) /\a>| ] ASjaua SuipuBi puB yy3>] 0 1 XSjaua lUBag XjEUiud aqj
qj¡M apoui su 3| apoqjBa aqj u; puB (jja]) a 3>I I
Jojaajap g s ad/(j X3 |iuoqx-W BqJ 3 Ag aqj j o
SUB3U1 Aq p3A\3ia pjB jjsqn s ¡s ( £.ui3 s |o i ) adXj-u ue uo ( e.uia 6]o i ) smajjBd adX j-j
3 jn S |j
'[ c I ‘9 1] ( s s o p u o j j 3 3 [ 3 j o ) j u a . u n a u i E a q s q j
u i p 3 |E3 A 3 j U 3 3 q S E q J o j a B j 3 u i a i i a n | j u i - j s B i j u o a | B u o ij ip p B u y
‘u j n j
3 3 U 3 I1[JU I
q a ii| A \
‘[ t ^ T ř l ]
‘J 0 J 3 3 J 3 p
p s ij ip o m
3 l |J
^ Iiu a j a j j ip
'( [ ' S y
3 3 U E jd a 3 3 B
s u io s s q
a s s ) [ ç i ‘t » l] p a j i n b a B
JB |n§U B
u o is s iu is
qj|A \
sq j
u o ij B u iq iu o a
s u o ij n q u j s ip
below about 3 eV, which was even capable o f changing its sign in dependence o f detection
conditions (Fig. 3). The effect was found o f a dynamic character, i.e. requiring the electron
dose above a certain threshold and exhibiting a time constant in the order o f seconds.
Explanation is based on a conception o f the p-type doped area to be negatively charged to
about 1 V with respect to the surrounding substrate, which is enough to make the incident
electrons totally reflected in a narrow "signal beam”. This pencil can either escape the
detection through the central detector bore or impact against the active detecting surface (Fig.
3). The proper phenom enon is then described as dynamic filling the conduction band to an
extent sufficient for recom bination o f the majority holes, which leaves the ionized acceptors
unbalanced [17]. This immobile negative space charge is then responsible for the mirroring.
Possible quantifiability o f this contrast type needs further examination.
Figure 3: P-type patterns (10 19 cm '3) on an n-type (1 0 15 cm-3) Si substrate in the CL mode at
1.5 eV, specimen tilt 0° (a), 0.45° (b), and 0.72° (c); scale bar 20 (.im, reproduced from Ref.
Another alternative o f how to avoid problems with the SE images is to excite the low
energy electron signal with photons instead o f electrons. Preliminary PEEM study o f doped Si
wafers [21] brought several surprising results. Although the energy unfiltered photoemission,
excited with the Hg lamp radiation, gives again the p-type areas brighter, if only fastest
photoelectrons arc acquired, the image contrast inverts in both p/n and n/p patterned structures
(Fig. 4). Further, the mutual shift between photoemission spectra from p- and n-type sites
reaches only 0.12 to 0.14 eV (Fig. 5) and always the photo-threshold is larger for the p-type.
Crucial is obviously the difference in amplitudes o f the spectra. These findings clearly
indicate not the ionization energy or surface barrier or near surface electric fields being
responsible for the phenomena observed, but simply absorption o f'h o t electrons on their
trajectory toward surface. The main scattering mechanism is likely the e-h pair generation.
Figure 4: Photoem ission micrographs o f the p/n and n/p structures (both lO1*’ cm '7 1 0 15 cm '3)
taken in the total photoem ission (a,c) and filtered with the threshold -0.36 V (see Fig. 5) (b,d).
' 1830/MV303 'ou JubjS jopun uoijBpunog a s u a p s qoaz 3 aqj qjiA\ poyoddns Xpnjs [ 33 ]
'I3 £ ‘lO’SO sSuipsaDOJd is d ‘Ç00Z s o a e q ‘g u o 3 A doosojoi^ a o jj :ui ‘qB js >jUEjg g [ 13 ]
[ 6 1]
An>l -sA qj ‘s;a b q -j - \ ‘aiA appg 3 ^
[0 3 ]
'S 6 8 ( f 6 6 1) Z.S 's^Md 'J S o jj s j d s y u a n s g -g
I 6 £S (3 8 6 1 ) 9 3 8
'6 0 £ (£ 0 0 3 ) 8 c 1 so isX q j u o jp a j g puB S u i3b u i| ui saouB A py :u; ‘quBjg g ‘EAO.ia|[n[A¡ q
[g j]
'8 c (S003) 901 XdoosojoiuiBjJin ‘EAOjaniM g ‘g u E jj g [¿[]
(qoaz3 ui) /.£ ‘o iu g >DSV 1S1 >0 0 3 SQd :«! ‘quBjg -g ‘b^ ijm g [91]
'IS > 0 0 3 o iu g ‘spua.ii ju aaay ql6 :u; ‘quEjg g ‘Bî[ipM g [s i]
'£63 (£003) £8 wag
sA q j q d d y ‘^upuojAj 'PM'V'S ‘aiMOH - y ‘sX ajq d u in n
'£ 3 3 (3 0 0 3 ) £6 /C doosojoiuiBum
f ' 3 ‘u io o jg g - g ‘u q E lu oq ss j
1 ‘! ïB u io o -ia IA1 IAI ‘EA0 J3 [jn w ■]
[I7 1 ]
[e t]
' 1 0 6 (5 0 0 3 )¿£
■|Buy aoBi-jaiiq puB aoEjjng ‘u o s |;m g L.iB 3i 's ‘XpoDBÂBf 'H ‘jiio S S e z g ‘ijB iu o o q g '¡aj
[3 1 ]
'8 8 3 > 0 0 3 ) IS saaiAaQ u o jjs sjg
•SUB.1 1 3 3 3 1 ‘AposBXBf 'H ‘>[UBjg -g ‘BACuannw g ‘s |p M
y \ y i ‘ij E u io o g g
lAI'IAI [ i l ]
' I £ 6 3 ( 1003) 6 L SJajjag sX q j q d d y ‘S|ia/v\ g j ' l ‘iJB U iooq g gAI 1AI [ 0 l]
S£9 ‘3 'IoA ‘0003 o u jg ‘-aso.iaiw
'[ 3 J § u o j 'Jng m 3 1 d o jj : u i ‘EA0ja|[ny\[ g ‘^ u e jj -g ‘sjp M >I'3 ' I ‘UBuiooqg w n
'9116 (3003) 16 's^Md [d d y g ‘sXa.iqduinH f '3 ‘uio o jg g g ‘jjo jn g 'g 'S
'6 L Z (¿661 ) 69 AdoasojoiuiBJJin ‘sujEjuojEg gg ‘siAOjaj q -q ‘q ajsE j ->j -|aj
l i e (0003) 6P -asopiiAi u o jp a jg g ‘A\Bqs|jM g< ] ‘ipisE 3 g j/\[ ‘/qB3s d 3
'6 1S ‘£ 'Io A ‘0003 o u jg ‘-a so ja i^ qg u ïk io j u n g qj 3 | so.ij :ui ‘a|A\OH 'V
'39£ (8 6 6 1 ) 91 g
[3 ]
lo u q a a i g a g '3 e a f ‘s u i n o j g ' Q ‘u;Bf 'h ‘sa|qBU3A ' a
> 0 1 (S 6 6 I ) 8 S AdoosojD|uiEjJin Ï<î V aiAOjaj - q - q
'361 (S 6 6 I) 081 -asojDijM g ‘aiA\o|g 'V
'1 3 1 7 (9 6 6 1 ) P I g
lo u q a a i p s 3 BA f ‘JaqBJM iM a ‘saïqB uaA a
[ 1]
'p ÉS ig ui UA\oqs sa jn p ru js aqj u iojj B jp ad s u o iss iu ia o jo q j :ç ajnSjg
A 1--
8ß e ||0A ß u ip jE jej
A I-
[ 3 3 ] u oijiu S oa j o sajnjuaApB Suisuuoad puB 3a|PBJIJB aq oj sd aag sid o j aqi s a u s p s aiiuapBaB
aqj joj ‘jaAaA\oig X|uo sjuauiajnsBaiu q j aqj qjiM ÍuiSbubui je u iie puB ‘S uqigojd lUEdop
aqi joj |o o j aAijBjijuBnb pajjsap aqj 3uijjaS u| adoq j p q j jso[ aABq sjojBaqddB [Bujsnpui
jBqj juasajd je sa ia a s ji qžnoqj|B Su inu iju oa aq ||ia \ Â puB ‘ajE sjojan p u oaiu ias
|p M
|B juauju3dxg
P. Horák, P. Schauer
Institute o f Scientific Instruments AS CR, Brno, Czech Republic, [email protected]
Organic light emitting diodes (LEDs) have attracted much interest because o f their potential
for an efficient emission in visible region and for application to large area flat displays. LEDs
utilizing conducting polymers have many advantages in terms o f simple fabrication process o f
a thin film such as spin-coating or casting methods [ 1].
Polysilylenes (PSi) are linear silicon (Si) based organic materials whose basic building
block is a chain built up o f Si substituted by alkyl or aryl groups [2], PSi are quasi one
dimensional materials with delocalized a-conjugated electrons along the polym er backbone
chain. Their optical and electrical properties are primary ascribable to the effect o f quantum
confinement on these a-conjugated electrons. PSi exhibit a sharp photoluminescence (PL)
with a high quantum efficiency in solid state. The use o f PSi in LEDs is attractive since this
can emit light in the range from the ultraviolet to the visible regions [3], But problems o f life
o f PSi LEDs are discussed only in very few papers.
Organic LEDs generally show a degradation o f their characteristics with time, such as
increase o f the threshold voltage and decrease o f the fluorescence yield [4], However, such an
evolution depends on the polym er bulk modification and changcs o f the injection mechanism
at the polym er-electrode interface. Cathodoluminescence (CL) is an interesting tool to study
the polym er behaviour since it eliminates the injection problems.
Poly[methyl(phenyl)silylene] (PM PSi) is a typical representative o f PSi. The
absorption spectrum o f PMPSi consists o f three bands (Fig. 1) [5], The electronic transition at
338 nm arises from delocalised ct-ct transition. The PL spectrum o f PMPSi consists o f two
emission bands - a sharp UV band at about 358 nm and a broad band in the visible part o f the
spectrum (Fig. 1). The sharp UV a - a band is o f exciton nature. The exciton is localized in
the Si backbone. PMPSi was selected for our CL study. The PSi specimen preparation and CL
method o f the study were described previously [6].
a*- a
Wavelength X (nm)
Fig. 1. Optical absorption (solid lines) and photoluminescence (PL) spectra (dash line) o f PMPSi thin film [5].
The excitation wavelength for PL was set to 280 nm. Chemical structure of PMPSi is inset.
■££-[£ ‘(frOOc) JiiAQ X^sje^s ‘uoiiBjusuitujsui ssisX qj asEjjrig puB saijdo
sp ijJB j p 3gjBL[3 ui spuajx ju so sy uo jb u iu iss jsiu] qi 6 jo s o j j ‘ j JsnEqos ‘ j í[Bjoh
T tt)I-£ £ 0 I (300c) 662 SP!I°S 3u i| 1b is X jj-u o n jo p iu n o f “ § >pjndssM
•¿8'£8 (0003) SU W IA I
“ D uipjBf ‘ o sjjXsg ‘ O xnajiog ‘T sb u saeq
'L \ 6 í - n 6 í (9661) ££ 'S^qd 'd d y jo f 'UBdBf ‘ h eunío>i
‘ IM oioiui5|B>| ‘ H o u a n "X o u iq so ^ ‘ \ u o u iq o ‘'JAj Bpiqso^ ‘">1 ojouiiqsoA ‘’V u fn j
'0Il7l-6S£I ‘(6861) 68
'w ^ D ” f
‘ O í l *>IIM
'393I-19Č I (Z.66 I) S8 SPPJM suaqjuXs ‘ H BUiiTo»
"IV oiouiiífB^ ‘ H o u sn ">1 ouiqsoA ‘'A u o u iq o ‘ >j ojouiiqsoA l'H EpiqsoA " y iifn j
u re a q
u o jp a p
11! 1S ( 1 1 \ :( I J °
jn o q
U IIU JD a d S
' I )
js jje
3 1)1 J O
S u iu u iS s q
U O I l E T l |B A g
3 ip
S lJ
lu s jjn o
1U 3 J 3J J I P
Á 1IS U 31 U I q
ïS d lA ld
A ^>[
U O f lB ip B J J I
J °
0 1 .4°
U JB S q
a s B a jo a p
X S -I 3 U 3
U O jp
3 |3
a u n i a i)x
31) )
p u t:
S u ljn p
S !J
[S ] a U l ) U O IlE ip E JJI
|uiu] Lflßu8|0 AeAA
lUBoq u o jp s p u| JSIS3J B sB 3|qE3i|ddB sq oj SUI33S uoijBpBjSsp uiEsq u o jp s p oj XjqiqijdsDsns
aqj qjiM isdlAId ‘Xpnjs ^ 3 sqj oj g u ip jo ssy -XiqiqBjs [eusieui sausqus ueo sisqiosqB
jo sj 9zi|;qEjs snouEA qjiM ;SdIAIdJ° gu id o g 3JH JJoqs sjí 01 anp SQgx ui jsXe| 3aissiui3 ue sb
!SdlAIcI 3Jnd jo uoijEzqijn sqj sspsduu auiij qjiM Xjisusjui X3 sq) jo asBSJoap jsbj sq x
uoijBjpxs jo suopipuos uo guipuedsp 3[q;ssod s; uoijeuuoj
puoq pou 3>)B3A\ sqj jo gup[uqssoj 3 ‘uo issp s puoq is-JS 3M1 'uojiBUJJOjsp puoq is~!S
b jo SDUsnbosuoa b si Xjisusju; -73 sqj jo 3SE3J 33p sqj ‘snqx ’(spuoq ¡S iS ) 3uoq>[3Eq |S sqi
q )IM P 3 JB ]3 J SI U O IJE JIO X asp D - D A fl 'U O IJB ipB JJI UIESq U O j p s p g u i J t i p 3]qBJS X )IE3U SUIEUI3J
3UO 3 |q i s i A s q j S B 3J3qA \ ‘301IJ qji/A XjpidEj s s s B a j o s p u o i s s i u j s A íl 3M1 '( 1 '^!d) | e u o j b l u au iB S
s q j j o u jn jjs s d s
a q j qjiM j u s u i s a j S e pooS u i s ; (£ -g ij) [SdJAJd J ° u j n j p s d s ^ 3 sq x
■SuiuujSsq sqi qjiM guuBdmoa % Qp se mo) se si
Àjisuajuj T 3 jo asBSJoap sqj Xjisusp jusjjno jssqSiq sqj je uoijBjpxa sqj jo SuiuuiSsq sqj uioij
ssjnuiui 3A1J J s y y 'Xjisuap ju suno jsqSiq sqi je jaqSiq si uoijBpBjSap jauiXjod jo sajSap sq x
UOlJEpBjgsp [BU3JEUJ B OJ pauSlSSB SEM XJISU3JU1 X3 JO 3SE3J03p 3qX 'J3J3BJBq3 |B[JU3U0dx3
-iqnuj 3AEq ssAjns sq x 'Z '§ !J ui UMoqs s; A3>101 J °
uoijBjpxs sqi guisn uiBsq
u o jp a p jo ssijisusp JU3JJ113 jU3J3jjip aqj je ¡SdW d J ° ^Jisusjui 1 3 sqj jo 3SB3J33p sq x
QOOZ spuaux W3.JSX
M. Hovorka , L. Frank , D. Valdaitsev , S. A. N epijko , H. J. Elmers , G. Schonhense
’Institute o f Scientific Instruments, AS CR, Brno, Czech Republic
bInstitute o f Physics, Johannes Gutenberg University, Mainz, Germany
*e-mail address: [email protected] isibm
In the previous PEEM study [1, 2], focused on the origin o f contrast between differently
doped areas in silicon imaged by slow electrons, strong contrast was observed between p-type
and n-type areas with the p-type brighter. The crucial mechanism was identified in electron
absorption phenom ena that are relatively w eaker with emission from the p-type Si. The
energy-filtered images o f p- and n-type areas revealed differences in the energy distributions
o f photoem ission. For the filter adjusted so that only the fastest electrons can pass to the
detector, the contrast inverts and the n-type became brighter than p-type. The shift between pand n-type photoem ission spectra indicated larger photoem ission threshold for the p-type.
Continuation o f the study has produced results presented here. The photoemission
spectra and micrographs for two sets o f samples, n-type areas on a p-type substrate as
specimens A and p-type areas on a n-type substrate as specimens B - both with different
dopant concentrations varying from 1016 to 1019 cm '3, are shown (Fig. 1 - 3). Si (100) wafers
were prepared by standard procedures [ 1] and left in-situ uncleaned, i.e. with surfaces covered
by a few nm layer o f the native oxide. The samples were irradiated by UV light from the Hg
lamp with a line at 4.9 eV and additional radiation between 3.2 and 4.4 eV.
The contrast inversion and corresponding shift between the p- and n-type spectra is
present for all concentrations o f dopants. W hile in the total-current photoemission image the
contrast between p- and n-type area disappears when decreasing the dopant concentration, the
inverted contrast behind appropriate threshold at the energy filter keeps to be distinctive. In
the former image the surface relief is evident but in the fast-electron image the topographic
contrast is suppressed which could indicate the electron signal acquired from larger depths
(Fig. 2, 3).
The energy spectra exhibit some structure, possibly connected with multiple sources o f
photoelectrons (regular energy bands, bands o f surface states). The difference in
photoemission thresholds between p- and n-type increases up *fo 0.2 eV at highest
concentrations (Fig. 1). For the specimen B the PEEM spectra also showed the photoemission
yield increasing with the increasing p-doping level [3-5], Furthermore, some changes have
been detected in emissivity o f surfaces exposed to prolonged UV light illumination
(particularly with the specimen A).
[1] Frank L. et al.\ Origin o f the contrast in im aging o f doped areas in silicon by slow
electrons - to be published.
[2] Frank L. et a l. , Proceedings o f the 6th Dreilandertagung M icroscopy Conference (2005).
[3] Allen F. G. and Gobeli G.W ., Phys. Rev. 127(1), 150-158 (1962).
[4] Ballarotto V. W., Siegrist K., Phaneuf R. J., W illiams E. D., Mogren S., Surf. Sci. 461,
L570-L574 (2000).
'( f - 1113 9 , 0 1 O) 610 l : u i o j j o q o j d o j )
suoijEjjuaauoa juedop juajajjip j o j B j j o a d s adXj-d a q j j o uosuEduioa :uiojjog 'apis XSjaua m o |
a q j uo a p i a u i o o oj j a p j o ui pajjjqs p u B ‘auo adXj-d aqj jo jqSiaq a q j j i j oj p a z q B u u o u uirujaads
adXj-u aqj :aAjna p a q s B p í ^ u i a 8|0[ j o Xjisuap j u e d o p aqj j o j B jB p : d o x '( u o i j E j j u a a u o a
J3IJJB3 ; u i o s[o i j o ajBjjsqns adXj-u a q j uo sujajjBd padop adXj-d) g suaiuiaads a q j
uo SB3JB adXj-u p u B adXj-d aqj u i o j j s j e u S i s u o i s s i u i a o j o q d a S B ja A B j o B j j a a d s X 3 j a u g : [ " S i j
(A )
B b jio
ß u jp je ja j
(A ) 3 B
b j |o a
ß u ip je ja j
'LZZZISO/tOI 'ou lUBjS J3PUI1 >13 J ° XauaSy Jubjq aqj Xq pajjoddns s| Xpnjs aqx
■(ZOOS) S¿fr-69fr
‘( l ) l 6
s^M d
|d d y
f ‘ a
s u ib tiI !A \ p u e
jn a u s q j
js u S a js
‘ M
o jjo jB [[ B g
[$ ]
' 4 -i
Fig. 2: PEEM images o f the specimens B (p-type doped patterns on the 1015 cm '3 n-type
substrate) with various dopant concentrations (top to bottom: 1016 to 1019 cm '3). Left column:
full photoelectron emission, filter bias o f +1.64 V. Right column: filter bias o f -0.44 V.
Hung Quang Hoang, Junli Wu, Mans, Anjam Khursheed
National U niversity o f Singapore,
4 Engineering Drive 3, Singapore 1 17576.
E-mail: Tel. +6565162295.
Accurate ray tracing o f electron trajectory paths through curved magnetic sector plates is
required for a variety o f different instruments, such as dispersion free beam separators [ 1] or
energy spectrom eter designs [2], Since magnetic sector field distributions are inherently
three-dimensional in nature, and often involve curved boundaries, direct ray tracing o f
electrons through them is a non-trivial task. Numerical field solving techniques such as the
finite elem ent m ethod do not in general provide enough accuracy to extract aberration
coefficients from the trajectory paths o f focused beams in three dimensions. This is because
numerical meshes in three dim ensions that model complex shaped boundaries typically need
over a million free nodes, requiring prohibitively large am ounts o f com puter mem ory and
unmanageably long program run times. The situation is made even more difficult for
retarding sector units, w here an electric field is overlaid onto the deflecting m agnetic field in
order to slow down electrons to very low energies (a few eV). Another disadvantage o f using
a fully finite elem ent approach is that high-order interpolation methods are needed to extract
accurate field information from nodal potentials. W hile not impossible to carry out, it is quite
a challenge on curved mesh geom etries in three dim ensions, and therefore, mesh-less
methods are preferred. In this paper, a semi-analytical approach is taken, one which uses a
modified two-dim ensional finite elem ent solution in combination with a Fourier Series
expansion. This approach avoids the direct use o f a fully three-dimensional finite element
solution, and has the desirable feature o f not requiring a mesh in the region where electron
trajectories are to be plot.
A two dim ensional finite-elem ent solution capable o f modeling curved boundaries is first
performed in the plane o f the sector plates (x-y plane), as illustrated in Fig. 1. In this
example, circular sector plates have an inner sector plate, an outer sector plate, and an outer
round casing. The outer sector plate is set to unity potential, while the inner sector and
casing is set to zero potential. Fig. la shows the numerical mesh, Fig. lb shows the initial
finite elem ent solution, and Fig. lc shows a modified finite elem ent solution where central
nodal coefficient Po is adjusted by a factor o f (1 + I r /ld 1), where h is the mesh size in the x-y
plane, and d is the distance to a zero potential odd-symmetry plane in the out-of-plane
direction (z-direction). In this example, d = 1mm, and h = 0.25 mm. As expected, the
influence o f the zero-volt odd-sym m etry plane causes equipotential lines to bunch towards
the sector plate edges. A Fourier-Series expansion then integrates the potential in the plane o f
the boundary and calculates field values for any point inside the sector unit. This semianalytical solution provides reasonably high accuracy when com pared to the full three
dimensional solution, which in this case, can be calculated accurately by a rotationally
symmetric finite-elem ent program, since the in-plane geom etry is circular. Fig. 2 illustrates
the accuracy o f the sem i-analytical approach for electric field values in the x-y plane 0.1 mm
above the zero-potential odd-sym m etry plane. The accuracy is plot as a function o f the
number o f term s used in the Fourier-Series. The maximum electric field variation is divided
by the m axim um field strength value along a line running through the centre o f the sector
VVV-LZV ‘9SS V SSH 'st y d ul
'umufsuj •p n u ‘g jsq jg jso 'JM pire p33qsjnq>| y
(Z6 6 I) t7 f-ie ‘I °N ‘36 m dO ‘sB Z S 5 (i3 Jd q p u B aso y h
P un
[ |]
uim f si sdjDjdfo dUDjd 01 douvjsip ‘q = z jd si ¿üvpunoq ¡üidiuiuás
-ppO 'uiui[ 0 = 2 ¿of duvjd ú-x aifi Suojv sdnjVA p p i f oujodp f o ¿DDjnooy .7 d jn S ij
su li
.X d q m n jq
ÁvMV uiiuj pdjDDOj Áuvpunoq ÁjjduiuiÁs
¡mjudiod-OAdz áqjvdu uof lunoooü 01 uoiinjos pdifipojA¡ (o) uoynjos Q£ (q) qsdw /# Áq [g (v)
S9ivid joiods jv jn o jp f o auvjd dqi ui uoijnjos fUdtUdjs-Bjiwj .7 aunSij
X <~
^ ----------------
uiiu o č
-------------- ►
■p3JU3S3jd 3q [JIMSiqj JOS3|daiBX3
p3[iBjsp Lsjiun JOJ33S psAjno qgnojqj sqjed Xbj joaiip ajBjnosB jO|d oj pssn si qoBOjddü
s!lI.L '% TO UBLtl J31[BUJS SI JOJJ3 P|3IJ 3UJ03|3 3ï?BJU33J3d 3qj JBqj SMOqS Z 'S lJ 'l!un
Marcus Jacka, Andrew Pratt*
Email: m [email protected]
*Department o f Physics, University o f York, Heslington, York, YOlO 5DD, United Kingdom
The manipulation o f neutral atoms using resonant light gives rise to the possibility o f novel
forms o f microscopy. We demonstrate the rudiments essential for achieving this, namely
focussing and scanning, using metastable helium atoms as an example. In the second part o f
the paper we propose a novel form o f ion microprobe, this time using resonance ionisation,
which could overcome some o f the disadvantages o f conventional liquid metal focussed ion
The ease with which photons, electrons, and ions can be focussed and otherwise manipulated
is reflected in the maturity o f optical and charged particle microscopies. Sub-wavelength
resolution can now be achieved with scanning near-field optical microscopy, whilst the use o f
aberration correctors has brought transmission electron microscopy to sub-Angstrom
resolution. Neutral atoms, as their name implies, are much less susceptible to manipulation,
however they have desirable properties in terms o f surface analysis, microscopy and
nanofabrication. At thermal energies, o f tens o f millivolts, there is negligible penetration o f
the atoms below the surface, meaning that the information area is confined to the size o f the
probe. Despite their low energy, the mass o f atoms compared to electrons leads to small de
Broglie wavelengths. Two other advantages o f neutral atom beams must be mentioned. One is
the lack o f repulsive Coulomb forces between the particles and similarly the lack'of screening
which gives rise to the Boersch effect. The other is that because non-conservative (for
example velocity dependent) forces can be applied to neutral atoms with laser-cooling
techniques, the Helmholtz-Lagrange relation does not apply. As a result the energy
norm alised brightness (Acm" sr" eV" ) o f the microprobe at the sample is not limited to the
brightness o f the source.
M anipulation o f neutral atoms
Laser cooling, or manipulation using light forces, is the newest, and also the most promising
o f a number o f methods by which a beam o f neutral atoms can be deflected, focussed or
otherwise manipulated. The first methods were developed by Stem, Rabi and others in the
20s and 30s, making use o f the force exerted by an inhomogeneous magnetic field on atoms
with magnetic dipoles. Later it was shown that magnetic lenses could be created from
hexapole magnetic fields. Combined with laser cooling to first collimate and reduce the
longitudinal velocity distribution o f the atom beam, spherical and chromatic aberrations can
be significantly reduced [1], Further techniques include the focussing o f ground state helium
atoms by specular reflection from clean silicon mirrors [2], and the use o f Fresnel diffraction
lenses [3],
■[¿] uojjnqujsip Ägjaus mojjeu SABq puB Äjisuap qSiq sje ‘psonpojd ¿q X|iSB3 UE3 siuBsq
uiojE |Ejjnau ojuosjsdns jo SAisnjjg 'sojnos uo; sssujqSuq-qSiq e jo uoijb3J3 sqj aq p|H03
juauirujsui SOl/V 3L1J u! P3sn 1J0S 3lP J ° SJ3SB| qjpiMSuq mojjeu sqj jo uoijeoqddB jsq jjn j y
o o i uoi passnaoj uoijbsiuoi aaueuosay
su 3 [ Di}do-oi3 uSeui aqi uo pssoduiuadns sjíoo lieds
dhou3euiojp3|3 äujsn iiiesq psssnooj aqí jo Suiuueos jqSu-ysq (D) ’(uiojjoq) uienq psssnDoj pl[i pue
"(3[ppiiu) uieaq p3jBiii|[|OD OLfi ‘(doi) uoijBiuj||oo ajojaq apjojd ureaq aqx (q) sjusuoduioo sis\[eub
SDEjjns pue SuisstiDOf ‘SU}JBUI11|0D uieiu oq] SuiMoqs oijeuraqos («) jusuiiujsui SQIAI V°A 3lll
•uoijosjíp j B q j u ; j s q j j n j psdopA sp s q u b d p u s ‘a d o D s o j o i u i e j o j / Ü B S s a o s u s3|d;ouud
siuos ssjEjjsuoiusp j i ( u i u i j - j o jsjsiuEip SBq u i B s q p a s s n o o j a q j ) adoosojoiui b j o u s i
s i q j js|iq/v\ '[9] (SQIAI) A d o o s o j j o s d s u o i j E j p x a - s p s i q B j s E j s u i u u o j j a d o j ‘s u i o j b u i n q s q § j
3 | q B j s B j a a i j o L U E s q b U B 3 S puB s n a o j Á p p r u o o j psjorujsuoo u s s q S B q j u s a i n j j s u i u b > | J o ^ J V
nqj jo
•[f] SDjjsujouEu jo SU3J uiqjiM oj suiBsq luojE jo sábjjb puB siuojB 3|Suis jo Suiuoijisod
osioojd sqj a\o|[e spoqjaui pajBoijsiqdos sjoj/\[ ajqissod sje (s p p y oijauäEiu snoausSouioqui
jo uoijeoqddB Aq) ssdjoj juspusdop A|[BijBds puB (jo s jp J3|dd0Q sqj oj snp) soaioj
juspuodop Äj|oo|9A qjoq LsqjpiM3uq oiuiojb jo sssum ojjbu s q j jo ssnEoag -suojoqd jubuosíj
-isEnb puB luojB ub u33A\jaq sssoojd uoissiuis/uoijdjosqe sqj ui poŽUEqoxs umjuaiuoiu
oqj luojj 3uij|ns3j (sojoj snoauBjuods aqj se umou>[) sojoj oqj jo ssn ss>|bui poqjoui
js3|duus sq x pajoaiip s; JSPESJ aqj qojqM oj (u p jsq j sa o u a p jsj puB ‘[-(7] 8s sss) smcmaju
puB s>[00q ju3||33\3 XuBiu 3JE sjsq j - jq g u umo sji ui p|3ij Suipuedxs ÁJptdBJ B siuoosq
sbl| ji saijqSis-piiu sqj souis puB ‘ssnbiuqosj jo aSuBj pBOjq b ssssBduioous Suijood jssEg
Ionisation can be performed efficiently and in a highly localised fashion with two intersecting
focussed laser beams. One raises the atom from its ground state to an excited state, the
second, possibly o f the same wavelength, ionises the excited atom. The ions can then be
extracted and focussed with a simple electrostatic lens. This process o f resonance ionisation
is well established in other fields, particularly mass spectroscopy, and schemes have been
developed for every element except helium [8], It is suggested that a resonance ionisation
focussed ion microprobe would have spatial resolution equal to current liquid metal ion
sources, be able to operate at significantly lower energy (the energy spread o f an LMIS is
approximately 3eV, so has to be operated at high energy to minimise chromatic aberrations)
and could be developed for virtually any atomic species.
S o u r c e d ia m e t e r ( m i c r o r s )
Modelling o f ion trajectories for a resonance ionisation focused ion microprobe. It is predicted that a
laser interaction volume o f less that 25pm diameter would produce a focussed ion beam less than
lOnm in diameter.
One o f the encouraging aspects o f this area o f research is that current techniques for
manipulating neutral atoms are very much in their infancy. In contrast, conventional charged
particle optics is a mature subject, and advances are necessarily more incremental in nature.
Its practitioners would do well to follow the progress o f its younger rival.
11] WG Kaenders, F Lison, A Richter, R Wynands, D Meschede, Nature 375, 214 (1995)
[2] B Holst and W Allison, Nature Nature 390, 244 (1997)
[3] O Carnal, M Sigel, T sleator H Takuma and J Mlynek, Phys. Rev. Lett. 67, 3231 (1991)
[4] H M etcalf and P van der Straten: Laser cooling and trapping. Springer (1999)
[5] JJ McClelland, RE Schölten, EC Palm, and RJ Celotta, Science 262, 877 (1993)
16 1 A. Pratt, A. Roskoss, H. M enard and M. Jacka, Rev. Sei. Instruin. 76, 053102 (2004).
17] NF Ramsey: Molecular beams. Oxford University Press, (1956)
18] GS Hurst and M G Payne: Principles and applications o f resonance ionisation
spectroscopy. Adam Hilger, (1988)
SP U!>J J IU 3 3 3 X
P. Jánský
Institute o f Scientific Instruments ASCR, Královopolská 147, 612 64 Brno, Czech Republic
[email protected]
Kinetic energy o f electrons incident on the surface is changing into the thermal energy. The
beam with high current density can be used for melting or vaporizing materials and for
m odifications o f the material surface. Beams with small diameters can be used for drilling
very small holes and creating sieves. Magnetic deflection o f the beam enables very high
productivity o f thousands holes per second.
The electron beam -welding machine [1] in ISI Brno is intended for joining components
and materials used in electron optical instrumentation and vacuum engineering, but in the last
years its capability was extended for experiments in micromachining [2]. As an electron
source a thermionic cathode with tungsten hairpin filament is used. Typically the beam energy
is 50 keV, the maximum pow er is about 1.5 kW. The optical part consists o f four main parts
(fig 1): electron gun, centering and stigm ator coils, a magnetic lens and a deflection system.
Fig. 1: Scheme o f the optical system o f the electron-beam welding machine: I) cathode,
2) electron beam, 3) wehnelt electrode, 4) anode, 5) centring coils and stigmator,
6) magnetic focusing coil, 7) deflecting coils .
To achieve higher current densities and smaller beam diameters, we started with
numerical simulations and optimizations o f the electron-optical system. The beam properties
depend mostly on the electron source, therefore we started with the simulations o f the electron
gun, where the space charge effect is significant [3].
For the simulations we use program EOD [4], Field calculation is based on the finite
elem ent method. Space charge distribution is iteratively evaluated from the ray tracing. From
the field intensity and potential distribution near the cathode surface a new value o f the
emission current limited by space charge is established. Then the space charge distribution is
recalculated for the new emission current until this process converges. The electron gun is
simulated as a rotationally symmetric problem. Emitting surface o f the hairpin filament
cathode is approximated by a blunted cone with a spherical cap. The shape o f the cathode
makes the left border o f the mesh. In our simulation (fig. 2) w e set the filam ent on 0 V and the
anode is on 50 kV.
We are testing and improving our new algorithms for space charge limited thermionic
emission and analyzing current electron gun [1], In the next months we want start with the
design o f a new electron gun and focusing optics designed with respect to requirements o f
301-101 ‘I IoA ‘o u jg ‘0003
■oojj ssijdo u o jp a p u; uSisap sqj joj uiEjSojd q v 3
: f jbuibjz puB g baoous'j [fz]
'8t7-9fr (£003)
Z ‘OS is3p d o PUB s3iuBq33]A¡ s u ij) B>|i}do b B^iireqoaui Buuisf '(qoaz^ ui) suiipBui
S ll|p |3 A \ U JB 3q-U O JJ03[3 UB J O S o ijd o JO S U O lJB Z I|B U II)d o " f lBlu?IZ ‘ g BAOOU3q ‘‘J /q S U B f [ f ]
’3 1-01 (9OO3) I ‘IS (S3!ld o pus soiUBqDSjAj s u ij) E>[ijdo b E5[raBip3m Buiusf ’(qD3Z3
ui) urenq u o jp o p ,\q zjjEnb ojui s q o q ouij jo U01JEOJ3 :•[\: pBqoz ‘7 >|i:dn(j ‘ -j >[BdnQ [3]
> 9 1“6 5 1 (1003) 39 uinnoEA -sjusuodiuoo
I|euis jo Su|p|3M p3[[0jju03-j3jndiu03 jo j unS uojjo3|3
jE qoz ‘ 1 >(30|A “ f >l?dnQ [ |]
patliwa st tum q uo.tpajd 3t¡¡ dna ¡ojuat/ds o ijii.w a u o j ps)un¡q d ifiiM paimutxo.iddo
apoifinn paiodif ai/i lu o u j q q ¿¡ ui unS uouiodp stuoiw jaifi ai/t Jo uoi¡D¡niutg
'S i j
J Jirák1'“, J. Linhart1'2 and V. Neděla 1
1Institute of Scientific Instruments AS CR, Královopolská 147, Brno, Czech Republic
2 Faculty of Electrical Engineering and Communication, BUT, Údolní 53, Brno, Czech Republic
[email protected]
Scintillation detectors are frequently used for the detection of secondary electrons in scanning
electron microscopes. In order to achieve an efficient scintillation in these detectors, sufficient
energy must be added to secondary electrons by an accelerating electrostatic field. This field
is usually created by connecting a voltage of 10 kV to the thin conductive layer prepared on
the surface of the scintillator. To add such a voltage to the scintillator of the detector may be a
problem in the case when the detector with the scintillator is placed in a specimen chamber of
a variable pressure microscope at a pressure of several hundreds Pa because of electric
discharges in gas. In this case, the connection of a high voltage to the scintillator requires
placing the scintillator in a separate room where the pressure does not exceed several Pa.
Simultaneously, it is necessary to permit secondary electrons to travel from the specimen
chamber to the room where the scintillator is located.
In case of the proposed detector, separation of the room with scintillator from the
specimen chamber is accomplished by a system of two pressure limiting apertures with
central openings diameters of several hundred pm and by a separate pumping of the rooms by
a rotary pump and by a turbomolecular pump as is seen in Owing to the vacuum
system the pressure of several Pa can be maintained in the scintillator room while the pressure
in the specimen chamber reaches 2000 Pa,
by rotary pump
. __ ____
Evacuated by
P r e s s u r e in s p e c im e n c h a m b e r [Pa]
Fig. I. a) Schematic drawing of detector.
b) Dependence o f pressure in scintillator room on pressure in specimen chamber.
A voltage of several hundred volts is connected to the pressure limiting apertures, and
an electrostatic lens is created which together with voltages on other detector electrodes
enables the secondary electrons to pass through from the specimen chamber to the scintillator
room. In the scintillation room, the secondary electrons are accelerated by the electrostatic
field with a voltage of 10 kV to the scintillator. Trajectories of secondary electrons simulated
for one version of the detector electrode system are shown in Fig.2a. Trajectories were
simulated by Simion Ver.7 program [1].
•91 eO£9I cOO
lAISIAI p alo jd Xq pus 9880/S0/S01 '°N ÎI3V O pafoid Áq psyoddns sem >[jom. siqx
'0002 ‘Ç 'A3J £0t70/í6-13NI ‘l ^ Q 'V P^BQ ‘pmuBW s.jasn 0'L u o isjsa q £ NOIJMIS
'x 0001 (q ‘x 0 0 S O
uoijboijiu S bui
‘b j 00c ainssaid je jo p a p p pssodojd Xq puS jaddoo psppS jo uopeAJSsqo '£'^!J
uauipsds aqi u; e<j 006 P oqe jo sjnsssjd b oj dn susm pads jo uoijBAjasqo joj uoijBjíojdxa
sjí jo Xjqiqissod sqj ujjijuoo jo p o p p siqj qjiM sjusiuusdxa Ájopnpojjuj
ui sjb jo p o p p
siqj Sujsn puS jgddoo pspjrô b jo suoije/ussqo jo sjp ssj sujos '\ OS/. = ran ‘A OSt7 = £aí"l
‘A 8" = zan ‘A 01 = lan ssp o jp a p uo soSejjoa jo j q j'3 ij ui uMoqs si sopuopuodop pojnsBoui
aqi jo auo 'sap o jp sp |BnpiAipui uo s38b)|oa SuiSuBqo joj psjnsBsuj 3jsm jsqiueqo usuipads
aqj ui ajnssaid sqi uo jioj uapjoS aqj uiojj p a p sp p |buSis sqj jo sapuspuadsp aqx
'A OS/. = H n ‘A OSfr = £3n ‘A 8- = zan ‘A Oi = ,an s a p o jp a p uo ssS b ijo a Jaqmeqo
u siu p sd s ui sjnsssjd uo [ioj u ap | 0§ uiojj jbuS;s p o p s p p jo souspuadsp psjnsB 3p\[ (q
0 [ jo)B[[ijups uo sSbj[oa ‘a 0S¿ = 4,3fl ‘A 00S = eaH ‘A 0 1 = zafl ‘A Z = lan ssp o jp sp
uo ss3b)|Oa p p y oijBjsojps[a jo p a p p ui a 3 Çjo su o jp sp /ÜBpuooss jo ssuopsfBJx (b'2'3i¿
[Bd] ja q iu e q o u a iu jo e d s u; a j n s s s j j
Anjam Khursheed and Ding Yu
National University o f Singapore, 4 Engineering Drive 3, Singapore 117576.
[email protected]
Over the last four years, proposals for time-of-flight electron emission microscopes
(TOFEEMs) have been made, both with and without methods to dynamically correct for
chromatic aberration [1, 2, 3]. Correction of aberrations in time is an important alternative to
the more widely discussed tetrode mirror method [4, 5]. This is because the tetrode mirror
method has a relatively complicated column design, where its photoelectrons are designed to
trace trajectory paths around a multiply curved axis, requiring the use o f special alignment
strategies for its beam separator [6 ], The TOFEEM column in comparison, is relatively
simple. Fig. 1 depicts the basic layout o f a dynamic chromatic aberration corrected
TOFEEM. Apart from the complication o f pulsing the source irradiation with pulse widths in
the sub-nanosecond range and modulating the potential on a lens electrode at a rate of
several volts per nanosecond (indicated by Vc(t) in Fig. 1), the TOFEEM has a single
straight electron optical axis where photoelectrons are successively focused and magnified
using standard projection principles.. The TOFEEM also has the added advantage of
inherently being an energy spectrometer, which can readily band-pass filter the
photoelectrons that make up the final image by simply capturing them over a series of
different time windows. In the tetrode mirror method, an additional energy spectrometer must
be used.
In this paper, the possible advantages of TOFEEM, both with and without dynamic
chromatic aberration will be examined. In particular, the residual resolution, dominated by
spherical aberration, will be estimated for two different objective lens designs, and presented
in a form that can be compared with predictions already made for the conventional PEEM as
well as the tetrode mirror form o f aberration correction. Wan et al have predicted that the
tetrode mirror method will be able to provide 4 nm resolution at 2% transmission, as opposed
to 20 nm at 1% for conventional PEEMs [5]. In this work, it will be shown that TOFEEM is
predicted to provide better image resolution than a conventional PEEM, and that dynamic
correction of chromatic aberration through TOFEEM is expected to achieve a similar
performance as that predicted for the tetrode mirror form of aberration correction.
[6 ]
A. Khursheed, Optik 113, No. 11 (2002), p505-9
G. Schonhense, A. Oelsner, O. Schmidt, G. H. Fecher, V. Mergel, O. Jagutzki, and
H. Schmidt-Bocking, Surface Science 480, (2001), p 180-7
G. Schonhense, J. Vac. Sci. B 20(6), (2002), p2526-2534
P. Hartel, D. Preikszas, R. Spehr, H. Muller, and H. Rose, Advances in Imaging and
Electron Physics, 120, pp. 41-132, (2002)
Wan, J. Feng, H. A. Padmore and D. S. Robin, Nucl. Instrum, and Methods in Phys.
Res. A, Vol. 519, Iss 1-2, (2004), 222-9
P. Schmid, J. Feng, H. Padmore, D. Robin, H. Rose, R. Schlueter and W. Wan, Rev.
Sci. Instrum. 76, 023302, (2005)
1A133J O I pspsJJOO u oiiß xioq e oiiBuicuqo oiureuAp
900Z SPUÍ>->1 IU909^
b joj
o ijE u isip s : i a in S ij
B. Lencová, M. Oral
Institute of Scientific Instruments AS CR, Královopolská 147, 612 64 Brno, Czech Republic
1. Introduction
The computations in particle optics can be divided into “serious” ones, where the program
packages provide a solution to a number of electron optical problems and express the final
results in terms understood in electron optics like the aberration coefficients, and to
approaches coming from outside of the electron optical community, which do not or cannot
provide such answers. The basis of serious programs is an accurate computation of 2D
electrostatic and magnetic focusing and deflection fields and programs/modules for the
computation of optical properties from aberration theory or from accurate ray tracing [1-3].
Saturated magnetic electron lenses represent a very important class of problems, and the FEM
is the only method which can analyze them (saturated lenses cannot be solved with CDM [4]).
The advantage o f CDM is, however, the ease with which electrostatic systems like FEGs with
orders of magnitude difference in dimensions can be computed easily. Other problems like the
deflection systems and their aberrations are not studied in 2D [5], Computation of magnetic
lenses with FDM/CDM is not very promising [6 ], 3D computations in SIMION have reached
quite an advanced level [7]; this was the first Windows program, although now it looks a bit
old fashioned, and its biggest advantage is the low cost. When applied to electron optical
problems, sometimes quite curiously looking results can be reached and the aberrations are
not easy to dig out [8 ].
Electron optical design is done nowadays by commercial firms, so hardly any interesting lens
geometries are found in recent literature. What is also difficult to find are accuracy and test
cases or benchmarks, the only exception being an attempt in [9] which concentrated on
electrostatic problems suited rather to CPO programs. Nowadays, more and more complex
problems can be and are being solved. From the user point o f view it is important to have
simple but accurate and easy to use tools for the computation working under Windows XP.
The software is not intended to people without some preliminary knowledge of electron
optics, although quite often it is possible to meet a large degree of ignorance.
In this paper, attention will be given first to questions o f accuracy of FOFEM as an example
o f computation method, followed by a discussion of the requirement of accuracy for the
evaluation of optical parameters. Finally interesting recent result will be shown.
2. The FOFEM
Before starting a brief overview of the accuracy of FOFEM, let us briefly mention the user
aspect of our software. The older 2D FOFEM packages were replaced with a new EOD
program, in which we can analyze quite complex systems with many elements, lenses,
deflectors and multipoles. Optimization procedures are used to adjust the image position, in
particular with the accurate ray tracing, and space charge module allows the computation of
intensive electron and ion sources, as discussed separately in these proceedings. Even if we
want to improve the input and output of ray tracing, extraction of aberration coefficients from
ray tracing results, and analysis of spot profiles and emittance diagrams, in its current form
the program is stable, user friendly and easy to use, tested by a number of users as well as on
u i s j o j a a y a p a i j a u S B i u j o s a i u o u u B q a a j q j js a A \o [ a q j
jo j
p a u iE jq o a q u s a p p y
| b ix b a j E jn a a E
‘([B IJU S Jo d JB |B 3S u ; SB SIXB UO UOIJOajJOD 3 q j S u i p n p u i ) SJU 3I0IJJ3O D [ / \ [ 3 3 0 3 J ° u o ; j E j § a j u ;
ja d o jd
S u i u u o j j a d X q p u E u,(j) j o
p E a j s u ; iuJ / lu(i) [ B ij u a jo d
je je d s
p a a n p a j a q j g u ijn d u io a X g
s 9 |o d i ) |n i u p m ; s j o p a i j a Q
■S3SUS| DipuSeui JOJ 3|qB]lBAB 3JB
SJS3J J)JBlUq3U3q 0|sl 'MB[ 3J3duiy JO XjipqEA aqj SuiMOqS J|nS3J |EIAUJ JSOU1 aqj USAS 8upj33qa
ion Xq pus LS3 /un 3 uoijezijsuSeui 3uoja\ ‘sauq qsaui suij msj ooj qjiM qsaui auy Suojm
•uoijipuos XjBpunoq Suojm puE japjoq aqj jo uojjisod Suojm ‘qsaui ssjbos §uoja\ Suiuijap Xq
J3sn oi(i X|uo usqj si sjojjs jo ssjnos [ B n p B aqx '% [0'0 J° XaBjnaaB ue oj pajnduioa aq ubd
Xjisuap xny |E¡xb aqj jBqj Su;jEJjsuoiuap ‘XjjauioaS xajdiuoa qjiA\ sasua| aAjjaafqo pajEjnjES
X|q3iq joj U3A3 juiod sixe-jjo je 3 |qE[iEAE os|E si ji jnq Xjjsuap xnjj [eixe joj puBjsjapun oj
XsBa si J0JJ3 aqx -[g [] saijEuiaqjEui u; spoqpuj qsaui ui pasn ajnpaaojd ajduiis b qjiM pajpaqa
aq ubd sjjnsaj aqj jo XaEJnaaB aqj ‘uib§v '„sijEjjid,, pajjodaj XjsnoiAajd aqj ||B uo >|joa\ Xaqj
jnq ‘sauo snoiAsad aqj usqj pajEaqduioa ajoui qoniu ooj jou ajE SB|nuuoj A\au a q i 'uoijEjSaju;
jadojd E uuojjad oj puE jEijuajod jojaaA jo pBajsui xny asn oj XpuiBu ‘j/\i3 j0 d u! uoijnjos
b pajsaSSns 3AEq sm 3SB3 siqj ui U3A3 -[o i] saqsaui ||Buis ui suia|qojd oijiosds joj sjjnsaj
pojjoo sapiAOjd J3A3U IA13JOJ u! [Eijuajod jojaaA jBqj jsEd aqj u; X|aA|suajxa passuasip sbm ji
S3SU3| OI43u8l!l\l
‘[ f l ] u> n n
‘u b x ‘jsijuaias a s a u iq j jsqjouB j o s q n s s j P3XI03 3qj oi pgjBdiuoo %Sc0'0 J ° S!XB u0 Joxia
p[oij qjiA\ pajndiuoo si |Bijuaiod [bixb s ijj ¡JA13J qiiM
z [jO £ uo s <¿ u| jsn f paau pjnoA\
?q s]ii|od [8 X [ü 9 l qj|A\ q ssu i jsS jb] aqj jo j jnq ‘sixb uo sju |o d 0091 sp ssu aq sa jjja d o id
|B3ijdo j o uojjBjndiuoa aqj j o j ’[Aiad m !M sjuiod qsaui 0 c x 00t7 MJ!M psjn d iu o a jBijuajod
aqj pBq aq ‘( [ g |] u; j a jn S ij aas ‘uoijBja|aaoB xg jo j d £ 6 '£ JB aSBuii puB Q £- jb jaafqo ‘a r o
dß3) su3| aqnj-oM j b jo j ‘sasuaj 3ijBjsojjoa|a ^iBsidXj,, pazXjBUB [j/j] n i^ uaqA\ ‘ajdtuBxa j o j
uoijEjndiuoa aqj jo j qsaui a u ij jo asjBoa aqj SuijBjauaS ui s jo jja jasn
jo sjuiod qsaiu j o ja q u in u a\0[ j o asn aqj oj pauSissB aq jaqjBj ubd 1A13JOJ M1!M suiajqojd
ja u u o j aqj ‘jjo q s u | '[ f [ ] Suiqsaiu u| s jo jja jasn oj juiod oj pasn si ajnp ao o jd [BbijDBjd jn o
0 0 3 ui ‘jaAoajo|/\[ ’[g] ui Suiobjj puB sp ja y b jo j qjoq pajsaj X nnjssaoons uaaq SBq XaBjnooB
lAIHHOd ’[c I] 3UIBS aqj ajB saqsaui jB|nSuBj3aj u; sixb uo suojjßnba IM 3 J 0 J 3M1 uoijaajjo a
I|Eius b jajjB ‘an jj jou s; s iq x '[ | |] Xpuaoaj ajinb uaAa pajBadaj sba\ rniBp siqj puE - paA|OAU|
si juiod sixE -jjo jsajBau puE sjuiod jbixb qjoq je [Eijuajod aqj ajaqM ‘9 /(clo,A*t7+ll|ii!JA+,plA)
an|EA ja a jjo a puB ja jja q b S3A|3
juiod-g aqj SEajsqM ç/(do,A*£+ll|a!JA+lplA)
uoijBnba a a u a ja jjip aqj juiod (Bjjuaa e jo j sixb uo sap|AOJd 1A13JOJ 3MJ qsaiu
cuEiibs b uj ajBjnaoBui si J ^ 3 J 0 J S!XB uo asnsaaq spjaij qans qS nojqj p a a sjj aq jaAau ubo
s Xej puB IAJ3JOJ M1!M pajnduioo aq jouuBa sua[ aijBjsojjaa[a jEqj p a u ire p sba\ j| jssd aqj u]
S3SU3[ 3I4BJSOJ439I3
p su o iju aiu X y su q sq ||¡ a\ jBqj sqjXuj ju sjsisjsd am os [||js 3JE a ja q x
■puoqajdiuoa oj q n s ijjip sjoui sauiijauios os|E jnq ‘sajSuBuj qjiA\ iai3 J pJEpuEjs oj pajBdiuoa
jaisBa puE jajsBj uoijBjndiuoa aqj sa^Biu qsaui aqj j o XSo|odoj ajduiis a q x •qsaui a u y puB
asjBoa aqj j o uoijBjauaS ja a jjo a aqj joj ju au ia jjn b aj aqj si u ia |q o jd jsaSSiq aqj ‘saqsaui jE |n§aj
X|[EaiSo|odoj aqj Suisn ‘[£] saijdo u o jja a |a u| pasn se j ^ 3 j u[ -juBAa|ajji auioaaq 3ABq [o L]
ui p a ja a y a j sb suoijEjnduioa
jnoqE suoissnasip s tio u a jd j o jsoiu os ‘sjuiod qsaui 0000S
Xbs uei| j ssa| qjiA\ saqsaui ui uoijBjndiuoa p jay aqj u u o jja d oj jaAaosjBqM asuas ou sa>(EUi
ji jBqj qSiq os sXBpBMOu s; paads uoijBjndiuoa a q x sjuapnjs ajBnpEjSjsod puB ajBnpBJïùapun
air, which can be compared with analytical expressions even for tapered coils [17]. EOD
computes the axial multipole function up to m= 8 . No accuracy tests are available as well,
except for misleading information that only SOFEM provides the hexapole and decapole field
function for deflectors [3], For multipole systems the exact shape o f the field is actually less
critical (quadrupole systems are often sufficiently characterized with SCOFF, even deflection
aberrations do not strongly depend on the accurate shape of the deflection field [18]).
3. The optical properties
What accuracy o f computation we need? With what accuracy do we need to know the position
of the focus, the value o f the axial aberration coefficient of the third and fifth order? We have
stated that for the fields on axis the accuracy can be very high, below 0 .01 % (actually in the
past a much stronger requirement was that the field is smooth so that it can be correctly
interpolated). As a matter of fact, the focal distances and aberrations (3rd order geometrical
and Ist order chromatic) can be computed often with the same or better accuracy than the field
by solving the paraxial trajectory equation and by computing aberration integrals.
The low number of mesh points was actually the main reason why tracing could not be done
properly or easily, the 2D data are more difficult to interpolate than ID axial result, and so
they may produce uncontrollable local errors. The error estimate brings information about
possible mesh errors [13, 16]. However, it is not so trivial to get the aberrations from the
direct ray tracing by solving the equation of motion, otherwise one of the first published
attempts to get aberrations of a simple dipole mirror would not fail [3, 19], Unfortunately Liu
[14], who analyzed “typical” electrostatic lenses to up to 5th order geometrical aberrations
with his own aberration formulas and with DA program, used just very weak electrostatic
lenses, which are not a proper example to check our results against his results.
4. Aberrations from ray tracing and spot profiles
One of the difficult tasks is also the evaluation of spot profiles. For such a purpose, in order to
get smooth representative curves, it is desirable to know the positions of a very large number
o f particles, say a few billions, in a given observation plane. We can then take into account
the spot profile in the object plane, the energy width of the beam, the aperture position and
diameter. The best procedure we use for this purpose is to get from a representative set of
particles, say a thousand or so, computed by accurate ray tracing, the aberration coefficients
o f any required type and order [20], The coefficients are obtained by regression, so that the
quality of the fit is easily estimated.
As an example of a difficult system to analyze we show here the'preliminary results of
computations of probe profiles of a very low voltage scanning electron microscope. The
system consists of an electrostatic lens followed by a cathode lens, slowing down the beam of
5 keV electrons to just 1 eV energy. The figure at the left shows the beam profile after this
system in the Gaussian image plane, demonstrating a good resolution of such an arrangement,
even if the beam energy width is also 1 eV. After we switch the Wien filter on [21] (the filter
is placed in the crossover above the lens, it acts as a weak electrostatic lens and an energy
dispersion element. Without changing the lens settings, the beam size becomes larger, but 10
micrometers towards the lens the beam shape changes, making the spot size smaller and
astigmatic (see Figure 1).
5. The future
EOD is a modular program and it makes a good basis for further improvements. We hope to
present more computation o f difficult problems in electron optics such as mirror and
‘Ç00Z3IM 3 0 J J -> (3 ?E JO || PM 'BAOOUO'I g ‘5 (3 0 [A I
(p sq sq q n d sq oj) 9002 o iu g ‘S |S 3qi a q j ‘[Rio K
'SS ‘9661 o u jg ‘(Jbuiuiss qlç) spuoJi iu333>| o o j j ‘b a o 3 i» i g
’LZ I ‘( 8 8 6 1 ) 8¿ > l!J d o ‘b a o s t o i g ‘s u s g ^
"60t7ť ‘(¿ 6 6 1) 89 sju su iru jsu i s i j p u s p s j o mo;a3>| ‘BAosusg g ‘o u sg pv
6PI ‘(K)0Z) 6IS V
sA qj u; qpjM ijsu] p n ^ ‘b a o o u s i g
'9002 o iu g '(jBuiiuns q I ) sp u s jj }U333>] o o j j ‘RAonurv] g
OZZ ‘(9003) 901 -uiBJJin ‘n n
' IM ‘(t-OOZ) 6 IS V S3>I sX qj ui qpjAi mjsui - p n ^ ‘o u sg ^ ‘b a o o u s t
£68 ‘(9861) III VRd ‘98 Xdo3s o i3i[A¿ u o j p 3[g Suiuubos ‘o u sg -jm ‘çAoouaq
•(1003) 911 s¿ q d u o j p s i a -3buii A p y MadsB»
6Z£ ‘(6661) LZP V
'^ M d u i q p j M Jjsui - p n ^ ‘b a o o u s i
[g [ ]
[ç | ]
[£ []
[1 []
[o I ]
‘¿S e ‘(6661)
:I ‘BAOsuaq g ‘o ijq n o q
> 11 ‘(S003) t>£S V s s y sXqd ui qpjAj j;su [ p n ^ ‘u e S o q -yy[ ‘n [ n JAI ‘3S}S 'O
LZP V S3>I sXqa ui qpjAj J}su[ p n ^ ‘[tuiiB[/ f ‘p e s y h
£661 ‘iBnuBjA i s .w s n ‘O L N 01SM 3A Clř N O M I S ‘IM«a V a
6P ‘(POOZ) 61S V 'so>I 'SÁqd ui qpjM Jjsu[ p n ^ ‘q P M P'A y ‘zsuijjbjm o ‘» p s g >J
iuoo'SO|idouojpD[3'M MM//:diiq
e e i ‘(ťOOZ) 61S V 'S3H s¿ q d ui qppM -JJSU] p n ^ -baoou3i g
¿661 ‘uoiBy Boog ‘SS3J J 3 ^ 3 -so ijd o s p i i J e j p sS je q 3 j o >|ooqpuEH -pa ‘JJO [jo T
6 6 6 1 ‘J3A\n[>| -sojido s p i j ie d psS jeqo ui p o q p u i jirau rap sjra ij s q x ‘p s s q s jn q ^ ' y
'9P-ZP ‘POOZ O iug ‘(JBUIUI3S ql6) spuojl IU333}[ DOJd ‘BAODusg g
[ []
:s 3 3 u a ja ja y
3 U B |d sS B U lj 3 q j 3 A 0 q B S JS p U IO JO jlU 0 I
j o j s p | o q j q 3 u s q i u o s p j o j d s q j p i r e ‘u o p s q o jiM S s i i s i j i j
s q j J31JB s a n S y
j b j j u 3D
‘S3SU3J 3 p o q jB 3 p u B n i j b j s o j ) o o | o o q j p a i q o q o [ i j o j d LURoq a q i s i s p i s p u B q j j 3 | a q j u o
[m u]x
[ JE ] /\¡ g S A l B J ° 3 U B |d
o o i-
[lxiu] X
t 01
snoojsp ujrl oi
1 ™ ]»
s S b UII
00 1 -
sq j ui
A S jsu a
UBISSIIBQ S q j UI X jISU Sp J U S i i n g • l 3 J t l 3 l J
[uiu] A
s. 0 l X
J 9ll!d
>13 § y jo i IS0S90ČZ0AV UB[d qoiB3S3>j [Buopnjijsui sqj Áq psjjoddng :i u oui oíí p.1[\\ o u >|oy
•SU0IPBJ3JUI qiuo[no3 oijsBqsojs Á[[t:niU3A0 puB ‘suisjsás psuSqBSiui jo sisÁ|BUB sqj ‘saisisXs
U0ip3|J3p X3BJtl33B qSiq JO SUOlJEXISqB JO SU0|133J-I03 DIUIBuXp JO S|SÁ[EUB 3qj 0 0 3 OJUl
oj uB[d 3AV ajnjnj 3qi u¡ suoipxisqB jspjo jsq3iq jo uoijBn[BA3 pire sjo p sjjo s 3[odi}iniu
B. LencováJ'b and J. Zlámalb
J Institute of Scientific Instruments AS CR, Královopolská 147, 612 64 Brno, Czech Republic
h Institute of Physical Engineering, Faculty of Mechanical Engineering, Brno University of
Technology, Technická 2, 619 69 Brno, Czech Republic
From the point of view of a user of electron optical programs for the design in microscopy it
is important to have simple and easy to use tools for the computation. Standard for most CAD
programs is that they work under Windows XP, use a mouse, on-line help, and provide simple
guidelines and examples for working with them.
Based on 15 years of experience with 2D FOFEM packages that were developed in the course
o f many years, partly at ISI Brno and TU Delft, and that were equipped with a graphical input
and display interfaces in DOS [1], we considered that the whole system has to be transferred
into Windows. The first but only a preliminary version we showed in 2000 at EUREM in
Brno [2], All FEM programs for lenses and deflectors were integrated in a single program,
their interfaces were programmed, and the modules for the computation of optical properties
and ray tracing have been written completely from scratch in order to introduce a number of
improvements in comparison to the old software. EOD is a modular program, compiled in
Visual Fortran, allowing programming o f all Windows graphics; use a mouse, pull-down
menus, dialog windows, and on-line Help (see Figure 1). EOD uses a clearly defined project
structure to analyze complex systems with many elements, lenses, deflectors and multipoles.
Optimization procedures allow that the excitations are adjusted to achieve given image
position, spot size or similar. Space charge module allows the computation of intensive
electron and ion sources [3].
Field computation by the first order FEM is no more a critical issue; by using a large number
of points (one minute computation is needed for around 700000 point meshes on 3 GHz P4
PC), error of field can be quite small and it can be simply identified and estimated. The size of
the coarse and fine meshes for FEM is virtually unlimited because of dynamic memory
allocation. FEM meshes with a graded step are used and there is more freedom in automatic
line mesh generation. Old type of data for FEM or tracing can be imported. In the background
the geometry of other data can be shown as well as CAD DXF files, to which the coarse mesh
can snap. The program displays axial potential or field and/or color-tnded saturation of iron
or map of potentials or fields [4].
Recently we concentrated on the computation of optical properties from paraxial trajectories
and aberration integrals for systems with lenses and deflectors. Let us stress that, in accurately
computed fields, the ray tracing of suitably chosen set of rays provides paraxial properties as
well as aberrations of any order. The number of data produced while ray tracing is quite large
and there is a need to establish safe and easy to use procedures for the display of results and
processing of ray tracing results for a given purpose such as the evaluation of given set of
Future development o f EOD will be aimed not only at improving further the input and output
o f ray tracing, to extracting aberration coefficients from ray tracing results, analyzing spot
profiles and emittance diagrams. For example, the analysis of complex systems should allow
•suoijexisqß pire saiyadojd
[BixBJEd jo uoiiEindujoD pire sauopsÍBjj jo S u io e ji - sjBijuajodinbs pire p[9¡J jo indjno
- AjpiuoaS :suai sqnj-j e szA jeue o j p3ÁE|dsip s/v\opuiM s q j jo auitfs jo 3|diuExg : [ s j n S i j
'HD SV JO 11S0S90ČZ0AV '°N m \d qsJESsay jEuoiinjpsui sqj Aq p sy o d d n s
•8 ‘soaeq ‘Ç00231M '30Jd ‘puiB jz T ‘baooiisi g
' ÍZ-ZZ 'U3PS3JQ ‘£0033JM 'M y ‘I®IU?IZ 'f ‘BAO0U31 g ‘/qsiref ¿
301-101 ‘I '[°A ‘o u jg ‘0003 r a f f i a '30Jd ‘[BU1BIZ -f ‘BAOOU31 g
'9Ÿ-ÎV (P00Z o u jg ‘(jbuiums mi6) sp u s« juaosy o o jj ‘eao3U31 g
[ç] ' sb 3 qjiM
ureaq aqj jo suoipEjajui jo s u o ip e j s j u ; qiuo[noj ousEqoojs U3A3 «JE! X[(biiju3A3 pue ‘siuajsÄs
psuSqESim jo s is A|eue sqj pire suisjsXs uoipsgap Ad b jm o e qSiqjo suoijBjjaqB jo suoipsuoo
DiureuXp jo sisX|Bire aqj ‘sjopsxioo 3|odij|niu pus j o j j iu i j o soijdo jo uoijEinduioo suijnoj b
900Z spus-il lu a js y
F. Mika and L. Frank
Institute of Scientific Instruments ASCR, Královopolská 147, 612 64 Brno, Czech Republic
fumici (w
One of the crucial parameters in characterization of semiconductor devices is the dopant
distribution profile. Dimensions of the doped structures have decreased to the 102 nm level, so
details in both lateral and in-depth distributions of dopants fall at least in the same order of
magnitude. Consequently, the device characterization becomes a non-trivial task. Twodimensional dopant profiling in this study was made in the cathode lens equipped low energy
scanning electron microscope (LESEM). In addition to non-destructive imaging and
availability o f multiple signal modes, advantages of this type of the LESEM include very
efficient signal collection and hence high speed of data acquisition. The dopant profile is
determined upon magnitude of the secondary electron (SE) signal [1], namely by measuring
the image contrast between doped areas and the substrate.
The contrast mechanism has been interpreted in terms of electric fields above [2] or
below [3,4] the surface, variations in energy and angular distributions of the SE emission
[4,5], and recently the primary electron dose delivered to the surface was taken into account,
too [6,7], This study aims at examining the contrast behaviour as a function o f the angular and
electron dose factors on two series of samples. Set A was with the p-type doped patterns
(1x10 to 1x10 cm of boron atoms) made on an Si (100) n-type substrate (1x10 cm"-* of
phosphorus atoms), and the set B had the opposite conduction types. Surface topography was
checked in the AFM device, which revealed the steps on boundaries between the doped areas
and substrate to be only 6.3 nm in height for the set A, and 36 nm for the set B.
Both sets of the semiconductor wafers, exposed to the same history of aging, were
observed in the as-inserted status in a Vega TS 5130MM microscope. The landing electron
energy was 1 keV and the image signal was collected with a standard Everhart-Thornley
detector (ET) and with the cathode-lens detector (CL). At low magnifications and fast scan
rates the p-type silicon appears brighter in the SE emission than the n-type Si (Fig. 1). As
usually in standard-vacuum microscopes, any increase in magnification or decrease in the
scan speed lead to more intensive contamination of the field of view with carbon, and also the
local charging is more apparent. In the same time, decrease or even reversal of the contrast
between differently doped areas is observed. With the set 5, the contrast reversal has been
found dependent on the electron dose (Fig. 2) while for the set A only some decrease in
contrast was established. No clear explanation of the contrast reversal is available yet. Let us
add that the samples freshly cleaned in HF exhibit a monotone contrast development without
Imaging in the cathode lens mode indicated the contrast behaviour possibly connected
with differences in the angular distribution of the signal emission. When comparing a
micrograph acquired with the CL detector well axially aligned and hence not collecting the
near-axis emission that escapes through the central hole (Fig. 3a) with the identical field of
view as recorded with the detector laterally misaligned by about 100 jtm, which enlarges
collection of electrons emitted near the axis (Fig. 3b), we meet the contrast inversion again.
Information collected so far obviously extends the scope of factors, connected with
technology of preparation o f the structure, its treatment before observation, environment in
the microscope chamber, up to properties and intensity of the illuminating beam, which have
to be taken into account if the image contrast o f doped areas should be reliably (and even
quantitatively) predicted and also utilized in diagnostic tasks.
3dX)-u ‘jo p s p p psyiqs s i x b - j j o (q) tsaniDid sqj jo aquas sqi ui u o i s í s a u ; ou ‘s ix b sip uo
pouSi|E jo p a p p (B ) :a3>1 [ JB jo p s p p 3 3 aqj qjiM psgEiui 3[duiES u íd h lQ l adXj-^ :£ ajn S y
(A3>1 i JB jo p sp p 1 3 ‘Emo 6101
§u|dop sdXj-u) „.oiojrl i i 8 3S0P ‘sSbuii jsejjuos psjjsAu; sqj jo |iEpp (p) u isjd op\, (o)
uiojd phi (q) ‘._w33d
(B):ssop u o jp sp sqij° uoipunj e sb [b s j s a 3 j j s b j j u o j :j ajngij
q jiA \
s jB jjs q n s
sqj ‘(q) ssdX} uoipnpuoo sjísoddo 3qj qji/v\ sjnpiujs
u b u o s m s ijE d
( u»
‘( b ) (
ř-uio s i () | x
0 1x I ) psdop sdXj-d pagBiui jo p s p p 1 3 : [ sanSij
'¿c£t/S 0/c0I VD ou JUBjS J3P™ y 3 J ° Xousgy Juejq sqj Xq psjjoddns s; 5[jom sq x
' 106 (SOOc) L i 'P W 3DvfA3iiq f i n s [B p ^ ^ ‘ijbuioo-[3
'OS ( SOOc) 10-Ç0 sS u ip a jjo .ij ¡Scj '3 ‘^uBaj ‘ 3
'£ 6 c (£ 00 č) £8 SJ3U3J sm /j '¡ddy je p 3 ‘uqBfuoips
'ÍZZ (Č00z) £6 ■'d03S0.imuD.iiif] '1 ‘>(UEjj ‘ jai iai ‘ije u io o íS "I ‘BAOJautM
'88c (t^OOČ) I S
110.113313 sun/X 3331 je p jm jaj ‘p eiu o o iS
' I 1£ (OOOt) 617 -OSOJDlfM UOUÍ33J3 f J E p ¿ J ‘X[E 3 S
>01 (S 6 6 I) 8S ¿ d 0 3 S 0 .l3 tlU D .n in J E p Q ‘OIAOJSd
[9 ]
I. Miillerová
Institute of Scientific Instruments ASCR, Královopolská 147, 612 64 Brno, Czech Republic,
e-mail: [email protected]
The Cathode Lens (CL) mode of the Scanning Electron Microscope (SEM) seems to start
attracting attention o f developers as well as users of the microscopes. Based upon our nearly
15 years experience with this experimental principle, mostly summarized in [ 1], a brief
recapitulation of usefulness of this system is presented here.
Electron optical calculations show [2] that the only way o f obtaining the high
resolution in an SEM at very low energies (below 100 eV) is to use the CL principle, in which
the sample is immersed in strong electrostatic field and forms the cathode of this electrostatic
lens. When adapting a conventional SEM, a feasible solution consists in employing a
sequential arrangement of the electrostatic field of the CL and magnetic field of the focussing
lens with only weak magnetic field appearing in the specimen plane. The CL anode serves as
a single-channel dark-field detector made from the YAG crystal with a small central opening
(say, 0.3 mm) enabling passage of the primary beam. This detector is inserted between the
specimen and polepiece of the objective lens, requiring hence a relatively large working
distance (usually above 5 mm) as its main disadvantage. Calculated and measured spot size
versus landing energy of electrons for one example of this simple adaptation of the SEM to
the SLEEM (Scanning Low Energy Electron Microscope) mode is shown in Figure 1. Here
the spot size was calculated by using approximate analytical equations [3] and aberrations of
the combination of CL and focusing lens resulted from analytical equations [4],
electron enerav. keV
Figure 1: The energy dependences of
the spot size calculated for aberration
coefficients of the magnetic focusing
lens Cs = 10 mm and Cc = 12 mm;
primary beam current w as'5 pA, gun
brightness 105 A cm'2 s r 1, primary
beam energy 10 keV and energy
spread 2 eV; the fixed aperture was 8
measured as a 25/75 edge width on
the standard A^t/C testing sample in
the Hitachi S-3500H SEM adapted to
the SLEEM mode, is marked by
In modem high-resolution SEMs the specimen is immersed in strong magnetic field in
order to secure the ultimate resolution, and detectors are mostly placed above the objective
lens so that the working distance is preserved small. Should the CL mode be introduced, the
specimen finds itself in both magnetic and electrostatic fields, namely strong and overlapped.
Some data about aberration coefficients and probe diameters for overlapped as well as
sequential configuration of the fields have been published [5]. We have measured the image
resolution for the overlapped fields and arrived at 9 nm for lOeV [6 ], The overlapped fields
surely promise better resolution than sequential ones but signal electron trajectories become
more complicated so explanation of the contrast formation is more difficult.
sqj Sumy pus jsddn sjí oj psjBJspssE si suoijssp psjjiuis jo uinxpads XSjsus 3j3[dui03 sqj
spoui 3 3 sqj u; jei|j jsquisuisj sn j 3 | ‘jxsjuos siqj u| [^i] ssodjnd siqj joj papašu s; jojosjsp
sixB-sqj-jjo pue sus]-3qj-SAoqB UB jo uSissp psjojiEj y 'Sjoq jejju33 sqj qSnoiqj uoijosjsp
sqi SuidEoss ssuojssfEjj u o jp sp |eu3is psjBuiq|03 AjSuojjs sqj Suuq ppiOM ‘s3i8jsu3
8 uipuB] mo[ 3M ‘soijbj uoisjsuiuii q8 iq oj uoijBoqddB jnq p3j33|[03 sjb S3(8ub JB|od jss 8je[
jspun 3u;jeu|8uo ssuojssfBjj usas ‘Ä[snoiAqo '[/.] ssuojosíbjj psjE|tJ3[E3 qj|M jaqpSoj ‘3jE|d
-puuBqs-ojaiui B puiqsq pssE|d \iojss|[od psujsjjßd sqj SMoqs £ ajn 3 |j -s3uej X3jsus umipsui
sqj ui jssq s^jom osjb ‘jojssjsp sqj SAoqB puB sus[ SAipafqo sqj jo sssid sjod puB usuipsds
ussMjsq psjjssui spouB 3 3 sqj qjiM psu 8 issp ‘suojjssp psjjiius jo uoijnqujsip jB|n3uB sqj
jo jojssjsp puuBqo-ijjmu sjduiis psuoijusw-SAoqE sqx suoijipuoo iunnoBA-q8 iq-Ejj[n jspun
S3|duiBS psjBSjj nijs-ui JO/puB saojB[nsu| jo uoijEAJSsqo joj '8 's |njssn sq ueo poqprn siqx
pssnpsj A|3uojjs sq ues 3u ;3jB q o sqj jsbs| je jo [ f | ] s a t u r e s psjBOSun SAijsnpuos-uou jo
XiuofBui b jo SuiSbuii psSjEqo-uou joj punoj sq ubo X3jsus SuipuE] uinujijdo ub ‘S|duiExs jo j
A3>1 1 punojB ABS ‘ssiSjsus uinipsui jb usas [tyjiruj á j s a jA jg s oi|i ui spoui 3 3 3Ml psMoqs
SBq soususdxs ‘jsasm oh ' [ j i ] jusiurujsui j/\[333 s q j s e ( [ | [ * o |] '8 's sss) s u is iu B q s s u i
( [9 ] u i o j j p s s n p o j d s y )
(m o j u io jjo q )
u j u i z 'Q p u B ( m o j j s d d n ) u i u i | o * 3 P
p p ij sq j jo
q jp i M
‘A 3>| 0 1
X S jsu s
u iB o q X jB U Ju d í s p p i j |E i j u s n b s s : m o j
l u o j j o q ‘s p p i j p s d d B j j s A o : m o j i s d d ß
a 3 o z (3 ‘ A 3 o z í (q ‘A 3 o so e (B
ISSlSjSUS 8 uipUB[ 3qj JB US>|BJ (169 IS
JESy) U3UJ|33ds SUIJS3J U0qjB3/p|0S
aqj jo sqdBjSojsijAj :j ajnSij
*3uop s q o j s ju s u iu a d x a puE
S U 0 ¡JE jn D |B 3 |E U O IJippB 3 j m b 3 j ||I M
U O I lB p jd ia jU I p 3 ] |B P Q
u 3 u ii3 3 d s
s q j jn q
a q j jn o q B
p s A js s s jd
p 3 d d B |j3 A O s q i u |
U O U Ejn3lJU O O P [3 IJ p a d d B |J 3 A 0 3 q i Ul p 3 J ;n b 3 B SISEJJU 03
'UOIJBUIJOJUI X S J3U 3 3 q i OJ À JB JJU 03 ‘J]3M /C[3A |1B|3J pS JJO dsU B JJ
u o i j b u i j o j u i jB |n 3 u B
SS3( j o
s jo u i si js b jju o s
s q j s 3 } B 3 ip u ;
s iq x
'j s o |
s¡ js b jj u o s
i E 3 |q d e j 3 o d o j s q j S 3 i§ J 3 U 3 m o [ á j s a
[ b u s jb u i
s p p ij
[ 9 ] s p | 3 i j |E |jU 3 n b 3 S p u E p 3 d d B |j3 A O q ji M q j o q ‘S 3 i3 j3 U 3 S u ip u B | SnOUBA
j e s p o u i 1 A I 3 3 1 S 3 q i u i u s i u i s s d s 3 u i j s 3 j - u o i j n | o s 3j u o q j E 3- u o - p | o 3 s q i S M o q s 3 s j n S i j
[ 5 ] SUIJBUIUJOP s d 3 3 > |
3IJB JSO JJ33J3 3 q j S 3lJO J33fB JJ U 0 JJ3 3 J3 UO UOIJ3B S p jB § 3 J SB ‘J3 A 3 M O H 'S 3 lS j3 U 3 S u ip U E ]
JS3M O ] p jB M O j S u i m o j S s p p i j S u i p u n o j j n s - u a u i p s d s q j o q s q j s s q d u u
B UI s p o l u 3 3
s q j ‘s n q i ’[ | ]
] /\ [ a S u o ; j r q o s 3 J - q S i q
UOIJEJISXS SU3J 3 A IJ3 3 fq O s q j S3SB 3J3U ¡ U3A3 S u i s n s o j s j (3 § J B |
o o j j o u ) À JBSS333U s q j ‘y \3 o i o j y\3> | o í l u o j j 'S '3 ‘s p n j i u S B u i j o s j s p j o A q i B i j u s j o d 3 3 s q j
BIA S U O J J3 3 P J O X ŽJSU S S u ip U B ] 3 q j S u |X jB A U S q ^ ' s p p i j ]E IJU S n b 3 S UI SE X |JE 1|U IIS p s u i B j q o
s q UE3 JSE JJU 03 U O lJO B JJJip 3 q j J B q j OS S 3 IJ0 J3 3 ÍB JJ IB u S jS S q j J O S 3 |§ U B (BIJIUI J O 3§U B J p E O jq
B |[S M Á p A I J B p j S 3 IÍJ S U 3 S U |P U B [ M O[ X jSA JE SJJO dsU BJJ ‘ 3 3
3 q j pUB SU3[ S 3 3 |d - S |O d S iS u iS
E Xq p s i u j o j ‘s u s ) u o i s j s u i u i j s q j jB q j p s M o q s [ g ] s u o !J B |ii3 ] E 3 X jB U |ix iip jd j n o ‘| [ i j s - j j n s i j j i p
SJOUI S 3 U I0 3 S q >|SBJ s q j s p p i j O IJS uS bU I pU E 3 IJ B JS 0 J J 3 S P p 3 d d B [J 3 A O s q j u ¡ ‘s j S u e J E p d p u B
/O j3 U 3 [BIJIUI q jl M 3 3 U B p j0 3 3 E UI ‘p p i j 3 IJB JS O JJ3 3 p 3 3 3 q j X q SIXB |B O Ijd o 3 q j OJ p S JB U II(I0 3
s jS u E iB q jn u iiz B j p q j d s s > [ s u o j j 3 3 p p s j j í m s s q j 3 s n B 3 s q [ ¿ ] X [d x u is j s q j B j p s u S i s s p
UB3 J 0 J 3 3 J 3 P 3A IJIS U 3S-U O |J|S O d p U U B q S -IJ in iU
B s p p ij
[ B l j u s n b s s JO 3SE3 s q j U|
primary beam energy, so very high efficiency of the scintillator or semiconductor detectors is
secured at low landing energies.
pola r angle o f emission, deg
Figure 3: Left: eight-channel collector of a multi-channel detector. Right: the radial coordinate
in the detector plane versus initial polar angle of the signal trajectory; primary energy 10 keV,
landing energy 2 keV, distance between the cathode (specimen) and anode (detector) is 10
mm, curves labelled with the emission energies, detector segments indicated as the grey
strips. (Reproduced from [7].)
Further indisputable advantage of the low, but also of the medium energy range is the
interaction volume proportionally diminishing, opening hence possibilities for a kind of layer
by layer in-depth tomography [15]. The SLEEM method was used for investigation of the
cube phase in an Al-Mg-Si alloy [16], In these experiments, visibility of nanometre-sized
precipitates without any surface relief was very clear thanks to small spread of electrons,
advantageous mixture secondary and backscattered electrons, and enhanced crystallinity
contrast. Consequently, multiple types of precipitates have been recognized in these
micrographs (Figure 4), as distinct from the standard SEM.
Figure 4: Surface of an agehardened Al-1.0 mass%Mg2Si
alloy with 0.4% excess Mg,
observed in (a) SE at 10 keV,
(b) BSE at 10 keV, and (c)
SLEEM at 1600 eV landing
energy and 10 keV primary
energy; the scale bar is 1 pm
long. (Reproduced from [16].)
The contrast interpretation issues can be illustrated on doped areas in semiconductors
[18]. In the example in Figure 5 two pictures are taken by so-called lower and upper SE
detector [19] with the specimen immersed in the magnetic field while the other two snaps
show the same field of view in the CL mode, i.e. with the specimen situated in both fields in
an overlapped arrangement. Thus, we have four images, all taken at 1 keV landing energy but
acquired with four different detectors. As regards the CL mode, difference was only in
diameter of the central bore and consequently in the angular range acquired.
sjnprujs jopnpuoojuiss oqi jo uouiioods oqj pspiAOjd Ajpupj oujg isi jo
'¿Z£Z/S0/Z0I 011 JUBjg jspun uoippunoj aouaps q o szj aqj Xq payoddns si 410M sqx
•¿t7 -^ (cO O Z )(l)a¿t's^ 3 N T O ar:'H uoiunzB>i
'ÍPZ-ZZZ (ZOOZ) £6 Adoosojoiuramn :'3
Pub 'W W 11BUIOOI3 ‘-I BAOJsi|t)]Ai
•PZÍ-ÍZ£ (17002)
Ä doosojD ijA i j o j X p p o s u a i S p g ‘a S s i q ‘IA I3 3 ql£ I
'9 Lí~ I Li
asB q j
(SOOZ) s u o ije u u o js u B J X
p ip s
o j- p i[ o s
3Cud : u j : '3 > [ u b jj puB
| BAOJ3j]np\[
[¿ i]
|E u o ;s ry jiQ : n ° A 's[B uoic[a] d iu b S jo u j u i s u o ijb u u o js u b jx
B p iq si ‘ >| BpnsjBj/\i
[9 []
'OVZ-iíZ (1003) £Z SumuEOS "3 -iampy puB j g n sy
'9L-ÇL (M)0Z) >13 SV ISI ‘°u jg ‘ 1 ?AOJ3[|nw
Xq pa ‘uoipjuaiurujsu] soisXqj aoBjjng puB sopdo s p ip E j ps§jBq3 ui spusjx
luooo>[ JBUIUI3S [EUO|JBUJ01U| q}6 30Jd :uI ”IA[ 5l33?JOH PUB '9 baoduox ‘ j >[30]A
0S-9£ (I00Z) i l Suiuubos : [ baojouow puB IAI [|?BJpB2 ‘ 1 >|ubjj
' 8 £6 'S 68 (fr6 6 1) ¿S 'sXqd '^ o j j day : g jarreg
'9£-8Z (S00Z) 901 /(doosojoiuiBjJin :'I BAOi3 [[n ^ puB 3 >|ubjj
£1-1 (6661 ) IZ Suiuueos "3 JatiEg puB y uBipiEj ‘3 baoj3||0im ‘-3 íjubjj
>(ubj..[ ‘j
B \o j3[ [ n ^
[g i ]
fť I ]
[[ |]
> 9 - £ 9 ( c 6 6 1 ) u |B d § E p B U B jQ
‘|B p
s o i-a y
X q - p a ‘ [ j o a ‘p \ [ 3 3 3
'LLVÍLX ( 3 6 6 1 ) ZI
01133 ‘3 B A O i a n o w
01133 P u b | BAOJ3|]n iA i
o c u j : u i : f j s s i j p u B -jai
h d d n s ] B p y B D iu iiq a o j^ i^ i
[ ]
juud ui
‘o§Bo;q3 ‘900 Z sisX|bueojoija¡ puB Xdoosojoij/\[ o o jj :ui :3 bui[bauo>[ pus 1 BAOi3|jnj/\[
> Z-81 (P00Z) SZ SumuBDS :'3 >(uej3 pus 1 baoj31[im
'LL-L9 (Z00Z) £11 W dO :'V paoqsjnq-»
>01'£01 (S66l)
>|bao|s :bab[si}bjq ‘j o d u i B i o ’3 Xq -p a ‘[A]3 )[aj o o jj :u [ :-jaj ous3
‘601-101 (9661) 101 >|!jdo:dJ!nJ>lpuB- 3 fqiJBg
'61Z-S0Z (6661) 8ť XdoosojoijAi u o jp a|g f q B\()JO[|n[Aj puB 3 >[uejj
£frt7-60£ (£ 0 0 Z ) J3IA3SI3 :oSsia ubs ‘(s3>[mbh’AV d
■ps) jj£l so|sXq(| UOJP3J3 puB SuiSbui] ui ssouBApy :u[ "3 íjubjj pus | BAOi3||ni/\[
s s a j j X u ia p B o y
[9 ]
[0Z] 'sppiA | bjoj 3J31Ujo PB3JSUI suoptiqujsip
UOISSIUI3 sqj ui psuiiusuBJ) si uoijBuuojui 3JOUI sspoui XSJ3U3 mo( puB lunipstti 3qi ui jnq
1AI3S 3m J0J X]iBJ3U3§ spjoq 3[nj siqx 'Suissaoojd | bu8 is juanhosqns sqj jo ÄouaiDijja 3i[) s;
pqM PUE p3p3[[03 SI SUOJP3J3 p3}llUI3 3qj JO UOlJtiqiJJSip XSj3U3 pUB aBJIlSUB 3q[ JO JlBd JBl[A\
|IBpp UI SUIA\0U>[ S3Jinb3J SJSBilUOD JO UOIJBPJCÜ3JUI 3|qBl[3J IBqj spnjouoo UBO 3M
(■ [/.[] Uio-U p s o n p o j d s » )
'A 3>1 I uibSb spoui IMHH'IS
sqj UI XSJ3U3 SuipUB| ‘A3>[ 01
XŽJ3U3 UIB3q iÚBUIlid ÍjapUIBip UI
uiui £ 0 (P) PUB 111111 L'\ (°) J ° MOCl
[BJJU30 j o s j o p a p p IM3 3 3 S aMl
‘j o p s p p 3 S p jB p u B is j s d d n ( q ) p uE
J3 A \0 | (b ) : s p o u i 3 3 s q j OJ poidB pB
1M3S J00£9 3 0 3 f sqj u| sjo p sp p
inoj jo SUB3UI Aq XSj3U3 8uipuE[
A 3>| 1 j e psSBui; PjBjjsqns ( [ 11) !S
oju) suoisnjjip poujojiBj :ç 3jn3|3
Eric Munro, Liping Wang, John Rouse and Haoning Liu
Munro's Electron Beam Software Ltd, 14 Cornwall Gardens, London SW7 4AN, England
e-mail: [email protected]
Tel & Fax: +44 20 7581 4479
The Differential Algebraic (“DA” ) Method [1, 2] is a powerful and elegant technique for
computing the aberrations of electron optical systems. It is far more powerful than traditional
methods based on evaluation o f aberration integrals. It is particularly useful in the following
situations: (1) Evaluation of high order lens aberrations (5th, 7lh order, etc.), (2) Aberrations
of multipole systems, including Wien filters, (3) Curved axis systems, such as prisms,
spectrometers and energy filters, (4) Electron mirrors, (5) Evaluation of asymmetry
aberrations, and ( 6 ) Systems with high beam energy, where relativistic effects are significant.
The DA method is conceptually quite simple. We define a Differential Algebraic Quantity
("DA Quantity'), of degree n in v variables, using the notation "„Q". The quantity ,,QVis a
vector whose elements are the coefficients of a polynomial of degree n in the v variables. We
also define a set of Algebraic Operations on pairs of DA quantities. Addition, C=A+B, is
defined as the addition of the two polynomials. Multiplication, C=AxB, is defined as the
multiplication of the two polynomials, discarding terms of degree greater than n. Subtraction,
Division, Square Roots, etc., are also defined, to form a complete Algebra. In software terms,
DA quantities form a new Data Type, which we call DA, with its own storage structure and
operations, which we have programmed as a C++ Class. We can then write statements in our
program like F = ((A+B)*C - sqrt(2.0*D)/E, where the variables A - F are DA Quantities.
These DA quantities dramatically simplify our software for computing aberrations. To
obtain the aberrations, we need the electron’s final position (x„v,) and final slope (.v,',v/) at the
image plane
as a power series in powers of its initial position (.v,„v„), initial slope (x„ ,vQ')
and initial energy deviation (50) at the object plane za. To express the aberrations as a power
series in these 5 variables - (x0, y 0, -v,/, v,/, Si>) - we use DA Quantities in 5 variables, i.e. v =
5. The degree n of the DA Quantities is chosen as the maximum aberration order we wish to
compute. For example, if we require the aberrations to 7th order, we choose degree n = 7.
We write out the differential equations of motion, which describe an electron’s position
(x(z), t’(;)) and slope (jc'(z), / ( - ) ) as functions of its axial position z, ftnder the influence of the
electrostatic and magnetic fields. These equations are written in curvilinear coordinates so we
can handle general curved axis systems. In these equations, we replace the physical position
and slope by their equivalent DA Quantities - X, Y, X ', Y '. We also replace the physical
clcctric and magnetic field components by DA Quantities. We fit the numerically computed
axial field functions with sets of Hermite functions [3], which are continuously differentiable
analytic functions. This is a crucial step, as the DA Method requires the fields be analytically
delined and differentiable. We now solve the equations of motion, with a single raytrace,
from object plane z„ to image plane z„ using a Runge-Kutta method. The code is a standard
Runge-Kutta solver, except that all simple variables are replaced by DA Quantities. On
completion of the single raytrace, the DA Quantities .V, Y, X ’, Y ' contain all the position and
slope aberration coefficients at the image plane. This is all achieved without any derivation
of aberration integral formulae. This makes the software very robust and reliable.
'96-06 (1?002)
'6"l (¿661) I Âdooso.i3tyq uoupaj'j fo jm u tio f ‘a so y -j-¡ pue SB zs^iajj (j
£L Su3 omojpspou3ijA¡ ‘n q z x pus ojunjAj 'a ‘n iq H ‘s s n o ^ [ ‘S ubm q
'68c-S 8c (0002) 111
'S u a q 3 g puE Subx i ‘S u e m j
> 2 1-601 (6861) PZ suoivudpoDy a p i u o j ‘z ja g 'W
aqj qjivi pajB|im iis
u o ij E U jq ii io a j 0 j j | u j - u i s u d - s u a |
q jiA \ ‘s u a | a A ij o a f q o
v, j o ’ ) p u E s3 a q j j o a j j o o ubo u i s u d p u B j o j j i u i a p o j j a j B M o q § u iA \o q s sq n s sj j u a s a i d [|im s / \ \
j o j j i i u d p o A ¡ d i Jo s d n j ü A 33
99986L L S itc e 60 l I 06 £+
S IS ¿ [( l i l L il
( s i|O A )
00 £
t - S 'O S I -
( U llU )
d jq v j
i ¿ te i -
9 ť '6 S |9 ¿’89[ —
(sjioa )
s^ ) p 9 j n d w o ¡ j
jo jjiiu
dpojjdf j v o i j D D j d
'£ d jn S ij
ueo su a j aA ijoafqo a q j j o sq a q j os ‘baijeSbu oj aAijisod m o jj
ubo sj 's u a | aA ijoafqo ub jo
a q j jo a jjo o ueo ji os ‘aAijEŽau
■paipouBo a q os]E
‘aS irej opiM E J3AO p ajsnfpB a q
sX ea\|b si ' 3 -(z 9 |q E jJ S u ijja s a8Ej|OA qoBa l o j p a jn d u io o ajaM
pni.iBA sba\
puE s3
's n o o j a q j jsnfpE oj
|Buij a q j CA J° S u ijja s qoB3 joj puB ‘s j j o a 09 1®PPM SBM l'A 3 § e j |o a
A3>| 01 j o X 3 jau a u reaq ju a p io u i ub qjiA\ ‘u m í 081“ = ” JB P3X!J ids?) aJaAV sauE jd aŽEiui p u s
loafq o a q x 3Jba\jjos y q j n o qjiM p3jB ]nuqs ‘jojjiui s p o j js j (EOijOEjd b jo a |d u re x a ub SA\oqs
Z o j n ň i j [^] jojjiiu a p o jja j e S u isn p a jo a jjo o a q ubo sua] aA ijoafqo ub jo
a q j q jo g
(E iju ajo d
(siiufi ¡s) siudiDiffdOD uoijvjjdqv pdjn d w o j ¡ d¡qvj[
/ 6 0 £ 'č 1 0 8 ‘t , c
! LOÍ'Z + 0 8 't? ü
u i o j i {3
|B ix y
8 1 9 Z—
8 I9 T -
! ZPÍ9 - 9 Z ÍPP-
/ IK 9 -S £ £ W ^ -
u o iy o js iQ
! ¿ I£ l ■ ť lS ť
lu s iiB iu S i js y
! 1¿ 6 9 + 6 'S X ť -
+ L\ÍV
Í 8 S 6 9 + l'9St-
íin s o íj v a
le is
J in s a y 3 i; Á | B u y
• q y |B 0 U 9 q d s
/ 9 1£ I
A j n j p íš i¿
BUJ0 3
u o ijB jja q y
% 2 0 > IIe 3 J o a \ ( [ 3 |q B x ) s q n s a j y q
0 J 3 M BJEp 3UIBS o q x
JOJJÍVUUOJfDdjd dpoip IfllM. }Sdl Dl)Á]DUy [ BjnSij
L U O il{3 SUBJX
a q j u; s jo jja a q x
p o q ja u i Y Q s q j q J!M p a j n d u i o o u a q j
'3 JJ U 0 0 SU3[ o q j JO JJ 0 [ o q j o j UllU c 8 9 ' ¿ 0 l Ie s j u i o d o S b u ii p u E jo o f q o p u E
LA ° 2 9 6 6 9 . 1 ° X S jo u o |B ij iu i u b i |j i a \ s u o j j o o |3 j o j ‘s p | 0 i j o ijX [B u e a q j q S n o j q j o o b jj X e j b u i o j j
POJOBJJXO S J3 M SJU3I01JJOOO u o ij B jjo q E X i E u i i j d o q j p u B s o i j o s p s s o g - j o u n o j q jiA \ p a j n d m o o
a jo A \ s p | 3 i j a q x
dB § lu u i g p u E ja ja u iB ip u iu i p p j o
s a o a id a jo d
p u B ‘d s S
u iu i o i
j a j a u r e i p u i i u Q p j o s a p o j j o a ] a j E o u p u q X o z qii/vs. j o j j i i u a p o i p o i j a u S B U i - o ijE js o jjo a ja p a u i q u i o o
b jo j a [d u re x a ub SM oqs [ a jn S ij
' s p | a y |a p o u i o [ jX ]b u b S u i s n p a > [ o a q o u a a q s s q X o e ju o o b a q x
japjo pajinbaj Xub
•sju a io ijja o o u o ijE jjaq B a q j jnoqE u o o b iu jo ju ; a q j |[B uiE juoa <: au E |d aSE iui a q j j e Z ‘ á ‘ X ‘Z
*,{ ‘X
saijiju E n £ ) Y Q
3ltl ‘30bjjX bj a]S uis b J a jjY
's j tu y
ajaqAvXjaAa ajB sju a u o d iu o o
X)iao|aA pire ssjB u ip jo o o s ^ u o jj s s p a q j ao u is ‘ju io d S u iu jn j a q j j e sa iju B jn S u is sa jE u ;u u |a s iq x
' Z ' Á ‘ X ‘Z LÁ 'X - s a ijiju B n ö Y O JuaiEA inba jia q j Xq s ju a u o d iu o o X jioo|3a puB sajB U |pjooo
|EO|sXqd a q j aoE [daj puE V a u iij j o su o ijo u n j se ( ( i ) :
p u E ( ( / ) : '( j). \ ‘( i ) x ) sa jB u ip jo o o s^ u o jjo a ja a q j j o j u o ijo u i j o
's iq j j o j
s ju a u o d u io o Xjioo[3A
suo ijB n b a a q j u io jj j j e j s sm
s jo jjiiu U O JJ33J3 puB S 3SU 3| spoqjEO síp ireq o j p a ijip o iu a q X jisbs ubo p o q ja u i a q x
V. Neděla1, J. Maxa2, R. Autrata 1
'institute of Scientific Instruments AS CR, Královopolská 147, Brno, Czech Republic,
[email protected]
“The Faculty of Electrical Engineering and Communication BUT, Údolní 53, Brno,
[email protected]
Environmental scanning electron microscope (ESEM) creates new possibilities in the
examination of various types of specimens and their phases. Experimental environmental
scanning electron microscope AQUASEM II was built in ISI AS CR in cooperation with the
TESC’AN Ltd. The main task of this microscope is the investigation and development of
special detector systems working in high pressure conditions and the study o f „in situ“
dynamic experiments.
The microscope is equipped with hairpin tungsten filament. The accelerating voltage
(AV) can be adjusted from 0.5kV to 30kV and the probe current from IpA to 2nA. The
resolution (high vacuum mode, AV=30kV, working distance =3mm) is 3.5nm, and for ESEM
mode (AV=30kV, WD=3mm, specimen chamber pressure lOOOPa) is approximately lOnm.
The vacuum system of AQUASEM II consists of two rotary pumps and one turbo-molecular
pump. The microscope is equipped with Peltier cooled specimen holder with mechanical
movement system in X,Y and Z axis. The pressure values are measured independently of the
gas type by the capacitance pressure gauges in the specimen chamber and in the differential
The differential pumping chamber with detector unit (pressure up to 25Pa) are the
main special parts of the microscope that enable to maintain high pressure (up to 2000Pa) in
the specimen chamber and vacuum (0.001 Pa) in the optical column of the microscope. These
parts make additional pumped space situated between the objective lens and the specimen
chamber, shown in Figure 1. In the small circular hole (diameter approx. 7mm), situated in the
centre o f the detector unit, is placed with the yttrium aluminium garnet (YAG) single crystal
scintillator, the main part of backscattered electron detector. The scintillator is also used as a
pressure limiting aperture; a hole with diameter 0.5mm is bored into its centre. The
differential chamber, the detector unit and the YAG single crystal are vacuum sealed with
“0 ”-rings. The circular hole through the YAG single crystal must bý*centred on the optical
axis by shifting the detector unit in direction of X,Y axis in the high vacuum mode, see Figure
2. The scintillator is used as the secondary electron detector (SED) [1] for variable pressure
and ESEM mode. Secondary electrons are swept by the positive potential around 350V
toward the electrode system deposited on the surface of YAG and switched to the current
ESEM AQUASEM II is also equipped with a special temperature isolated hydration
system situated outside the specimen chamber, see Figure 3. This system consists o f water
reservoir, water level meter, water vapour temperature meter, heating system with precise
temperature controller and precise flow calibrated micrometric valve. This system ensures an
efficient replacement of air in the specimen chamber by the water vapour from the hydration
system. It also enables to maintain stable high humidity environment respecting the physical
dependence of pressure of saturated water vapour on temperature.
'9 8 8 0 /S 0 /Č 0 1 ° N lURJg ‘o iiq n d s ^ q o s z j s q j j o Ä o u s S y I u b jq Xq puB 3 0 9 0 5 9 0 0 c
° N Ju e j3 L3 i|q n d ffy i p o z j o qi j o s o o u o p g j o X u io p B sy ^ q p s y o d d n s sem >)jom s iq x
9 1 ? - I I 9 ( £ 0 0 č ) Z s|BU3iBai s j i i j e n 'S|iïudib^\i g u p B jn su i puB p ^ \
S u iS b u ij j o j Á do3soj3ij/\] u o j p s j g S uiuuB og (B ju su iu o jiA u g j o ospi o ijx :'1A['V P IB u o q
X 01 ‘(0 6 6 1 ) ‘8 L ‘sa is X q j u o j p s j g puB s o i u o j p s i g ui sssu B A p y -A doosojoiui u o j p s p
SlUUUEOS |B)U3UlUOJ|AU3 3q] UI 33IA3p J O p s p p SIIOSSEÍ? 3 qj JO /Ü03l[J :'Q 'Q S01E[IUBQ
•u u m |0 3 a d o o s o jo iu i sq j j o jjE d e p u s ja q u iE q o S u id u in d [E iju a ia jjip J iu n
j o p s p p ‘jsqureqD u s u i p a d s (A \opq u io jj) : || lA ig S Y flÖ V IAI3S3 J ° A' ;|!A X b m e - jtq ' [ ' § 1 3
■[£] UOIPB3J |B0|UI3ip pUB UOIjn|OSSip ‘UOIJBDlJipiJOS ‘S u ijp u j JO UO|lEAJ9SqO
3|duiBS oqi j o Ájipium q jo ajnjB jaduisj ‘j o aSuBqo aiuij oqj sb qons [3] sju o u iu sd x s
jijis ui„ |C3iiuBuXp
jo j
3sn Xq u e 3 3do 3s o j3iui s q i '3SEqd ui ssSuBqs Jisq i qSnojqj
osEqd p inbq Jioqj ui su o u ip o d s ‘s u s u ip s d s oA ipnpuoo-uou jo ciAipnpuoo
L. Novak and I. Mullerova
Institute of Scientific Instruments, ASCR, Kralovopolska 147, 612 64 Brno, Czech Republic
e-mail: [email protected]
The Everhart-Thornley detector has been used for detection of secondary electrons since I960
[1], in order to increase the signal to noise ratio (SNR) in the resulting signal many
configurations o f this detection system have been developed. The collection efficiency
corresponding to a full chamber configuration has become possible to simulate [2 ] as well as
to compute and experimentally test the properties of the scintillator and the light guide [3]. So
far description of individual parts of the detection route has been based on signal averages.
This approach enables us to analyze the transfer of the signal and simultaneously to assess the
noise characteristics o f the detection route. As a matter o f fact, this method disregards
processes participating in the signal formation and besides, it is not applicable to low currents.
A new approach to the above-mentioned problem has been felt necessary in order to
incorporate measurement of very small currents (>0.01 pA) coming into the detector. For
experimental study, a special chamber equipped with two E.-T. detectors was fabricated. One
of them has the standard position (on a side of the chamber) while the other is directly
exposed to the impact of primary electrons. The primary current can be measured by means of
a movable Faraday cup. In fact, three parts of the detection route were examined. Using the
chamber described, the section starting with scintillator and ending with preamplifier was
thoroughly investigated. In order to identify the influence of photomultiplier, its input was
excitcd with a LED signal. Fig. 1 shows a pulse generated by one detected electron at the
output of the preamplifier. Finally, the signal-processing electronics was tested by means of a
voltage generator.
Two detectors of two different manufacturers have been examined. The main features
observable in detection routes of these detectors proved themselves quite similar.
Measurements exhibit an exponential decrease in the number o f high pulses of Ijght coming
out of the scintillator. In the photomultiplier/preamplifier section of the detector the shape of
the tranported signal changes in two significant ways. The photomultiplier adds noise pulses
whose number is proportional to voltage on the dynodes. In comparison with these pulses the
dark current of the photocathode becomes negligible. The preamplifier smoothens the signal
according to frequency of the filter chosen. As a result o f this, SNR [4] in the output signal
can rise but at the same time information about fast changes in the signal (edges, phase
contrasts) becomes lost. In terms of the measurement accuracy* the signal-processing
electronics does not increase the signal noise, introducing quantization noise only, as a result
o f the digitalization.
Taking into account the above-mentioned properties of the detection route, a computer
simulation routine was created in Matlab. This is based on generation of single pulses that
represent light pulses leaving the scintillator. They occur randomly but their heights and
widths are correlated with the measured values. The artificial signal created in this way is
then processed by means of individual operations, such as filtration, addition of noise pulses
etc., in line with what actually happens in the detector. Comparison between the real pulses
from the detector and the generated ones is pictured in Fig. 2. By dint of the simulation we
obtain an image in which the SNR and blurring of edges at fast signal changes can be
determined. This makes it possible for us to evaluate the features of the detection route in its
totality. The line-scans across real and simulated images are contrasted in Fig. 3. The SNR of
the simulated image and o f those obtained by means of E.-T. detectors are then portrayed in
'm c /W t/cO I ou lUEj3 ‘uoijEpunoj
oouspS ip o z 3 oqj Xq puE ‘ pjg ‘onqnday q o s z j ]gg Xq psuoddns s; >|jom oqx
'£££ (9661) 81 SuiuuEog ‘je jo Xof J a
'I I
l OA ‘ 8 ¿ 6 I
J u n J js u j - p s : g
sX qd
f ‘üdBA ^
f p u B |id t!A > j 7 M a n B q o s 'd ‘ B j B j j n y - y
[£ ]
801 (£00c) (£ q d d n s ) 6 qBUBOJOipM -osojoijai ‘b a o jo u o w 1 puB BuqBAUo^ q
'9PZ (0961) Lí u in jjs u j -p s f ‘X o |u jo q x m j y puB UEqjoAg g x
(\I9 T J Ijd lU B 3 jd
3J3M sj9j[ij
‘V^Ol lU9JjnD papapa)
p3iB[nuiis aqi uiojj
- S3SSOJ3 adXj 'X_'H sip jo saopspp
juaiajjip omj jo subsui Aq psjmboB ssSbuii sqi
S 3 |§ U B P 9 J pUB S 3 p j l ; }
* § IJ
[3 ]
(pauisosip 3 q ubd sibu§is qjoq jo adBqs
sqj aapjo ui psyiqs Á||BioijpjB 3J9m ssuiq) ubos
-9 U i[ 31 UBS sqj jo uoijBjnuiis J9jndiuoo - auq psqsBQ
•(pssn Sui^usjq jsbj ‘z^piSZ J3JUJ ‘% 0 Í ÍSBjjuoo
‘srl [ 9UJI1 \\9JAp ‘y d o i= ,3pI ‘A>10l=MBn ) Jo p ap p
JL-'H UB Suisn pajinboB ubos-suij - 9ui| pijOS £ *S|j[
[zh>1] j3^!|duieajd ¿o jajiij
[ s r l] a i u j i
(ZHIAÍ £ _ JSijqduiBSjd aqj ui J 9 j [ i ¿ |)
js b jju o d
p s jB jn iu is
sqj jo Suijunoo-
‘y d i -o = ,api
‘A ^ 0 l = 3Bn )
3U i|
pijos (%0£
J3 iJ Í¡d u iB 3 jd
sq j
j o jndjno sq; jb sssjnd jo Suijunoo- S9SSOJ3 z ‘Sjtf
[ / \ lu ]
( r z3 - ) d x 3 * v l D = ( 0 n
p ijo g
u o ip u n j
S u ijjij
a u ij
a 9 i j i |d i u B 3 j d s q j j o j n d j n o i b p s a ^ n b o B u o j p s p
p s p s p p 3 u o u io jj S u is u b s s jn d s q j - s s s s o j
’S i j
[su ] a u iji
6 u j ) u n o o ¿ o |aA a~ i
p u n o j s q ueo o jn o j u o i p o p p a q j j o u o p E j n S iju o o u i n u iijd o u b ‘s ja p u iE J E d
aso q j
s u o ijE u iq u jo o
p a y is s E p
p o s s o o o jd
3uiA B [g
'X ub o iio X o
p a u u q jjs d
ueo u o ijB in a n s ‘j o p o j a p s q j j o u o ijE jn S iju o o S u iq u o s o p s jo p iu E J B d j o p s u o a iS b j o j ' p 'Sig
Z. Pokorná and L. Frank
Institute of Scientific Instruments, Academy of Sciences, Královopolská 147, Brno
[email protected]
The electronic properties of modem nanodeviees are determined by their electronic structure,
the local density of states (LDOS) being one of the key characteristics. Various methods exist
for the LDOS mapping with lateral resolutions ranging from macroscopic scale
(Photoemission Spectroscopy) to the atomic one (Scanning Tunnelling Microscopy). Taking
nanodevices as those having at least one dimension under 100 nm, we require a method
probing in the scale of at least tens of nm. Another requirement is that the LDOS should
directly depend on the quantity measured with the method in question.
Methods for mapping the LDOS are the following:
Photoemission (and Inverse Photoemission) Spectroscopy are macroscopic techniques
for studying the band structure of occupied as well as empty electronic states. In the latter
case, electrons of a well-defined energy are incident on the surface and so injected into empty
electronic states above the vacuum level. From there they are de-excited into lower empty
energetic states and the de-excitation energy is released as an ultraviolet photon. The energy
of the lower state is determined by the difference between the incident electron energy and the
energy o f the released photon. The UV intensity versus UV photon energy plot gives
a qualitative image of the distribution of unoccupied electronic states above the Fermi level.
The angle-resolved version of this technique allows the 3D band structure E(k) to be mapped.
The tunnelling current in the Scanning Tunneling Microscope (STM) is influenced by
the local density of states. According to the Tersoff-Hamann model [1], LDOS of the sample
is directly proportional to the measured differential conductivity dI/dV(V), obtained by
modulating the tip voltage. The model holds rather exactly at low voltages with respect to
work functions of the tip and the sample (about 4 eV). However, the model generally requires
s-like tip states, and although quantitatively correct concerning the spatial distribution of the
LDOS, it is only qualitatively correct concerning the energy dependence. The energy
resolution is inversely proportional to the temperature, which is why very low temperatures
are used in the experiment.
Under certain specific conditions, the Scanning Near-field Optical Microscopy
(SNOM) has also been used for the LDOS mapping [2],
Another approach to determination of electronic bands above the vacuum level is to
measure the electron reflectivity R(E) by means of the Very-Low-Energy Electron Diffraction
(VLEED), or the electron transmission T(E) by means of a closely related method of the
Target Current Spectroscopy (TCS). Both methods employ an electron beam of kinetic
energy E in the range of 0 to 30 eV, directed onto a sample surface. In VLEED the elastically
reflected current is measured as a function of E, while in TCS the absorbed (transmitted)
current is measured. The inelastic contribution varies very little in this energy range, so both
R(E) and T(E) contain information about elastically scattered electrons, thus revealing band
structure features [3], Analysis of the data is based on the matching approach of the LEED
theory [4,5], The periodicity of the surface implies that the impinging electrons will only
couplc to the Bloch states having the same reduced wave-vector component k g parallel to the
surface. These Bloch states determine, through the matching on the surface, the amplitude of
the reflected wave. For example, if for given E and Kg of the impinging electron wave there
' I 830/M)/303 'ou lUBjS J3Pun >13 J ° -<3U3§ v j u b j q sip Xq psjjoddns s; Xpnjs sq x
> 0 1 1 ( U 6 l) 3 3 'S ^ d T ‘ I ls o j j b 9 puE g ‘X>pipA
T 9£ (6961) £1 P S J-mS ‘ O ‘UBdEj
'( P L 6 I LU 0 p U 0 1 ‘S S S J J S llU S p E S y ) U O IlJD .lJJlQ UO.lf¿>3¡g X S .L lU q M O 7 : g
f ‘X jp U S J
'6SZ.8 (S661 ) c i a Asy sX iy ‘t H ‘3 jsq u .ieis puE 'N 'A ‘a o s o jjs
'Z Q P L ( 3 0 0 c ) 8 8 ' » 3 1
‘a in iB O iiQ
[3 ]
S08 (S 8 6 I) 1£ a Asy sXi(d ‘ y a ‘uuBiuBH puß f ‘J J o s js x
A s y ' s X q j ‘ |B j s
• p s .i s p i s u o o s .i s q 0 1 | s p o u i S u i i u j o j
js e jju o s
o ip
u o i/ j c u J J iQ
o jin b s j
u o . ip a ig
1A13 u o i s s i u i s u B j j S u m i m o s
uw ag
j io S .w a u o j
sq j
jB p u i
sq j ui
sso q j
p s X o |d u i s
s |q u i s s s j
s n b iu q s s j ( 0 3 8 3 )
s u o ijip u o s
3 S 3 i[ x
■M3|A J O P |3 I J 3 q i J3A O SUIUUBSS U IE S q q jlA \ pSJSSU U O S §U|>|OOJ 3 q j JO U 3A 3 pUB U lB Sq J U S p p U I
sip j o s.mjjsdB jE|n3uE s[qiS i]3su-uou sq j j o
sq j j o 3;doj aqj si S O Û 1 J ° SuiddBiu p 3 j;p
s p n p u i ssnssi u i b u i 3i(x ’Xpnjs ju sssjd
p o q p u i IAI333S 3lll J ° X ji|iqE si[ddy
S u iS e u e u i
jo j
UIBiS 3l|l |B3A3J 01 psqsjs pUE p3qS|[0d USqj pUB ‘SZIS U|BJ§ p3J|S3p Sqj SASIl|0E OJ pSJBSUUB
pun p3||Oi sba\ 3[diuBS sq x 'XiSAijosdss.! ‘a s 61 PUB A3 63 s u o jp s p ju sp p u i jo XSasus ‘|s>piu
ui suibjS psju su o X jjusjsjjip ussM pq ssSuBqs jsejjuos sip jo uoijEAJSsqo Jsjij : [ sjn S ij
■( i -S |j) s j n p r u i s s i u o . i j s s p j u s j s j j i p j p q i uiojj S in n s s i ‘S[duiBS [B p u i s u ;i|e js X j3 X |o d b
in uoijbjusuo |ejsXjo j u s j s j j i p j o suibjS u s s M p q s s g u E q s jsejjuos M o q s ’S 's u s i p u e o
sSubj A 3 ()£-() 311J j 3 a o s u o j p s p j u s p p u i s q i j o X3jsus s q j S u i u n x 'sbsje X j i A i j s s j j s j - u o j j s s p
- q S iq p u e -a\o| j o sSeuii ub g u p s p u s j ‘ssejjiis sjduiBS s q j jsa o ubss ‘a s O ř A \o p q XSjsus
q jiM uie3b ‘u iB sq u o j p s | s XSjsus-a\o] s i p s p [ j [ - s p o q js iu 0 3 3 3 A PUB S 3 I 3lHJ ° J U E p u ss s sp
j s s j i p b S| p o q p u i (IAI333S) ,ldojsoADiyj uoupajg ,(3.i3itg-Moj Suuiuojg s q x
'[ 9] u isije iu jo j u o ip u n j u s s j q ssB jjns sq i ui je s |3 ss u io s sq X jiA ipsysj sq j puE
s a n p n jjs puBq sq j ussM jsq u o |jb [sj sq j ‘X[SA¡jBj¡juBnO -psjss[Ssu sq ubs ssssssoad sijsbjsuj
)Bi[j pspiAOjd ‘ssB jjns sq j Xq >|0Bq p s js s y s j sq ][ia\ su 0j j s s |s |[b ‘s|qE|iBAB ssjbjs s s jj ou sjb
T. Radlička
Institute of Scientific Instruments, AS CR Královopolská 147, 612 64 Brno, Czech Republic
[email protected]
The development of nanotechnologies in present days is often connected with the use of high
density ion or electron beams. Unfortunately the physical description of such beams leads into
great difficulties, because in high density beams the particle-particle interactions become
significant. We can proceed in two main ways, the space-eharge method or discrete Coulomb
The spacc-charge method was implemented into computer programs [1,2] and is used mainly
in high density beams, where the effect of Coulomb interactions may be averaged into an
additional field. A disadvantage of the method is that it is not, in principle, possible to
describe changes in the energy distribution. When we decide to calculate the Coulomb
interactions directly from Coulomb's law as a many-particles problem [3], the energy
distribution can be calculated very easily but the calculation time will be much longer.
The first Monte-Carlo calculations of particle-particle interactions in beams were started in
the late eighties in TU Delft by Jansen [3], He described electron beams in free space or in a
uniform electrostatic field and used Monte-Carlo method to calculate the beam properties like
energy distribution and trajectory displacement. The calculation time even in this simple case
was very long. The computers have become much faster in the last 20 years and we can now
easily calculate more problematic systems like ion sources. Such calculations also take a long
time, but it is possible to do it in several days or week on a PC. The Monte-Carlo method was
implemented into a computer code by Munro [4] but it cannot be used for particles in the
vicinity of the source.
The simulations of any system consists of four steps
• Calculation of the field in the system. We used FEM programs [5,6],
• Assignment of initial conditions to all particles (position, velocity and time of
emission) so that the macroscopic properties of the simulated beam correspond to the
properties of a real beam - Monte Carlo methods [7]
• Ray racing of the particle beam in the field including the particle-particle interactions.
During calculation the position and velocity of particles-at the transition through
selected planes are stored for later use.
Analysis of the beam properties from calculated data.
The ray tracing is the longest and the most problematic part of the calculation. The equation
o f motion for the ilh particle reads
m — (yvl) = qEj + qv x B + q - 2_,
j±i Ť*—p.
We use the non-relativistic approximation ( y = 1) for ion systems.
The length of calculation is affected by two properties. The first one concerns the regions
where the field is very strong. Even if they may be very small, it becomes a great problem if
we consider particle-particle interactions, because all particles must be calculated
u; s s d p s y [B o u siu n N :A js u u b u
(9 8 6 1 ) SS3-ld 'f l 3 § p u q u JB 3 ‘u e j j j o j
‘3 u i [ j o j p a 'l'M . ‘/ q s i o j p s i V S ‘SS3Jd 'H 'M
061-S 8I ‘( S 6 6 I ) 6 LA ' s so js iw f :g b a o s u s i
/iuoo-s>iJOMqjBui-MA\A\//:djiq ‘5-9 g y i lV I A I
/>[TT03'sq3urMMA\//:djjq * sjea\jjo s uiBsg u o jjs s jg oiunj/\[
(8861) U P O D I ” s i s s q i a q j ‘sujB sg s p ijJ B j ui su o ijse jsju ] q m o |n o 3 :ussuBf H '3
(5 0 0 c) 8 'd ‘SOABQ ‘SOOc SSUSJSJU0 3 Ad0SS0JSIJ/\[ " f (BUIE\Z ‘g EAOOUSJ
(M)()c) K )c “9 6 l ‘6 IS V 'S3>I s¿Md
’qjsiAJ Pub jsu i -3njs[ :S uuM og N ‘pE sy J
(3Djnos s q j uiojj lu iu g [
03UBJSIP B IB) ?uB[d s jn y s d B a q j ui S IIA n BD u ! uJBaq s q i j o u o iin q u is ip A S jai» puB S 3 |ijo jd uiBsq p3)B|H3|B3
J paqjnyadun — 1
peqjnjjad ---1
s u o i j o s p A 3>(Sl j o uiB sq psdB qs sjqBUBA
b 40 u o ije u u o j s q j u o s u o ijs e js ju i s p ijJ E d - s p ijJ B d j o j s s j j s s q j psjB[ns]B3 3A\ suiqoB ui g g g
s q j j o 3SE3 s q j u j 3 zis s o jn o s p n jjiA 3 q i puB 3UB|d s in jjs d E s q j jb (PMHAVd) qipiM A S jsu s
s q j pszA|BUB 3M jE p s ijJ E d u j -ssp ijJB d j s p i o q 0 0 c LH!M (s s p ijJ B d 0 0 S ) sJ->Bd
sb a\ u o ; j h |o a s s q i ’(ju3§JSA ip A jsa si uiB sq s q j) s o jn o s s q j u io jj u iu i j o ° l d n p sjE psiE O
3J3M u o ijs e js ju i S[0 |jJB d -sp ijJE d s q j ‘S11A11E9 J ° 3sb3 3qj u j ’uisjsA s ( g g g ) A qdE jS oqjq UIBSq
u o j j o s p u b puB ( s i l A n E9 ) s s jn o s u oi [e jsu j p in b q a in q iB á s q j ‘siusjsA s o a \j p sjE jn u iis
• s p i y E d j s p j o q j s e [ s q j puiqsq pus s p i j j B d
jsp joq js jij aqj jo ju o jj ui js 8 [ | ia\ s s p ijJ B d uresq ou jE q j os ussoqo s i s s p i j J B d isp io q jo
junouiB s q i ' sis Ajeub ui psjspisuos jou s je jB q j s s p i j J E d jsp jo q Aq sspis qjoq u io jj pspunoq
sq uEO sdnojg s s s q i sspijJBd 000c °J dn sdnojfS Aq uresq sqj sjB]ns|EO oj si Abm puosss
s q i - s s p i jJ B d 000c °1 dn s S e js a e uo s q | | ia\ jjB d qoBS u; jB q j os A|juspusdspu| psjnduios
sjb jB qj sjjBd [[buis oju i u is js A s s q j jns ubo s u o J! qj[A\ |B sp oj A\oq sAem s[qissod oa\ j sje
s j s q i -ssp ijJB d 000c UBlIl 3JOlu J0J ^qiSusj ooj si qsiqA \ ‘uiqjuoSiE (_N)0 UB ^IpnjSB si uins
s p ijjB d qui 0|n 03 s q i ’uisjsAs sqj ui ss|s¡jJBd jo jsquinu sqj si Ajjsdojd 3 u ijiu ji [ puosss s q i
•u isjsA s s q j j o j s s j s q j u io a j S u o jjs s i p p i j s q j s js q M s u o iS s j s q j u i u o ijo u i s p i j j B d
s q j j o u o i j d u o s s p s q j ojB JB dss o j sjq E siA p B s i j; s s u iijs u io § J|B U is A jsa s u ib u is j ji os ‘p p i j
3 u o .ijs b u i s j e jBqj s s p i j J B d u iE s q s q j
11b j o u o ijiijo a s s q j usqA \ j s a s m o h
Aq p s u i u u s j s p s i s z i s d s js s q j ‘p sjB ]n s[B 3 s i s s p ijJ B d
d s j s u o ije jS s ju i S uo] p p i j >]esa\ b ui si jB qj s p i j j B d qoBS
j o j s s n | | i a\ p o q j s u i u o ije jS s ju i s q j s p i j j B d s |3 u i s b j o u o ijE p o iE S j o s s e s u j - A |s n o s u B j|n u iis
P. Schauer
Institute of Scientific Instruments. AS CR, Brno, Czech Republic, [email protected]
1. Introduction
The principal quantities of image quality in SEM are contrast, spatial resolution, and
noise. However, to quantify the overall performance of an imaging system, the detective
quantum efficiency (DQE) is a better tool as it includes both the modulation transfer function
and the noise power spectrum. This means that for a detector to have high DQE, it should
possess not only high efficiency and low noise, but also good kinetic properties. A study of
the decay kinetics of some single crystal scintillators for SEM is presented in this paper.
2. Cathodolum inescence decay m easurement
The pulse mode utilizing a blanking system and 10 keV electrons for the excitation and a
sampling oscilloscope for the cathodoluminescence (CL) detection were used for the
measurement of decay characteristics [1], The measurement was controlled using the
IEEE-488 interface and processed by software written in Microsoft Visual Basic. Single
crystals of cerium-activated yttrium aluminum garnet (YAG:Ce), cerium-activated yttrium
aluminum perovskite (YAP:Ce), cerium-activated yttrium silicate (Y 2Si0 5:Ce3+), and
europium-activated calcium fluoride (CaF 2:Eu2+) were the scintillators studied as the most
interesting ones for SEM.
Some typical CL decay characteristics of single crystals for SEM measured in our
laboratory are shown in Fig. 1. The best decay characteristic belongs to the P47 single crystal,
whose decay time is 34 ns. YAP:Ce single crystal, having the decay time of 38 ns, is also a
very good solution. It has, however, a multi-exponential decay characteristic, so it shows the
afterglow of 1 % (measured 5 ps after the end of excitation). If a scintillation detection system
in SEM is to be able to operate at TV rates, its decay time should be smaller than 100 ns.
Hence, the decay characteristic of YAG:Ce is a little problematic. Having the (Jecay time of
1 10 ns, it shows the afterglow as high as 2 %. Eu-activated calcium fluoride has the decay
time of approximately 1.2 ps. and it is applicable only in slow scan rates.
It can be seen in Fig. 2 that the short-term decay component (decay time) of both YAG:Ce
and YAP:Ce single crystals depends only negligibly on the duration of excitation. On the
contrary, the long-term component of the decay characteristic depends strongly on the
duration of excitation. Therefore, at a very short excitation, the afterglows of YAG:Ce and
YAP:Ce are one and two orders smaller, respectively, comparing wifh a long excitation.
Kinetic model
It is evident, after a deconvolution of the decay characteristics from Fig. 2, that not only
emission from the Ce activator is present in the CL recombination processes in YAG:Ce and
YAP:Ce single crystals. Using the correction for the decay constant of the measuring device,
the fastest decay constants at whichever duration of excitation are about 60 ns and 20 ns for
the YAG:C'e and YAP:Ce, respectively. This corresponds to the single-exponential decay at
the photoluminescence measurement [2], Therefore, the fastest decay constant of the multi­
exponential decay characteristic can be ascribed to the emission from the activator. The other
decay constants can be ascribed to both radiative and nonradiative transitions originating from
unwanted defect and impurity centers. Utilizing the presented results and other CL
measurements (first of all dependences of intensities and decays of spectral peaks on the
activator concentration), a schematic kinetic model of radiative and nonradiative transitions in
the YAG:Ce single crystal has been created in Fig. 3.
S 6 f ‘(6 6 6 1 ) S6 'V
sÁ qd e p y "JB P í'PM ‘S jo q u u o !'V ‘!>P!qn>I ■ [ ‘b ^ s m o z jb q
' l ‘(9 6 6 1 ) 6 '[ddn§ -osojoipM 3 u iu u b o s r j u o n e ip s pue y ‘BjBjjny
e o i u o |o j
•oijqnds^i ipaz^ aqijo XousSy iubjq sqi jo p p \ZIWIZÜX ‘°N }ubj§ aqi Áq psyoddns sem >[jom siqx
Ä |9 A IJ 0 3 d S 3 J ‘SJ31U3D J O JB A ip B pUB Á JU tld lU l ‘p 9 J 3 p 3 l|J UIOJJ S9JBJ (U OISSIUIS) SA lJB ipB J 3 q i
3JB vf p u B ‘> f ÍQJ p u B ‘S J3JU 90 j o j b a i j o b s q j o j A j u n d u u a q j u i o j j 9JBJ j s j s u b j j s q ; s i VIM. í Á p A i p s d s a i ‘s s i j u n d u i i
PUB s p s j a p
‘S3JBJS 30IJJB J 9q} UIOJJ S9JBJ SAIJBipBJUO U 9 q j 9JB rS pUB >aS ‘í Í Á |3 A ip 9 d S 3 J ‘S J9 ;U 9 0 JO JB A ip B pUB
Ä j i j n d u i i ‘p 9 j 3 p s q j o ; s s j b j a jn jd B O s q i 3JB
t9JB J U O ;iB J9U 9§ UOJIDX9 3 q i S JU 9 S 3 Jd 3 J 5
p u B ‘O ÍQo í s j a j u s D p s j s p a q j u i o j j s ; b j u o i j B z í j B i u i s q i s q j s i
^ D V Á
9 m i ° 93U 3D S 3U IU iniO pO qjB O s q j j o p p o u i o i p u p j 3 q x ’£ #§ i j
s : o i SBM q i p i M s s j n d
33 :dVÄ PUB 93 :OVÄ J° Aßoap 30U3DS3U|Uin|opoqjBD
u o p B ip x s s q j j
l ^ 3 S JOJ s j o j b [ [ i j u i o s j b j s ä j o S jS lI I S j o
sqj uo qipiM ssjnd uoijB ipxa sq i j o so u sn iju i 'z *Sij
s o ijs u s p B J B q o
(sri) a l u n
s o u a o s s u iu in io p o q jB ;}
(s ri) a m u
’S i j
Johannes Gutenberg-Universitat, Institut fur Physik, D-55099 Mainz, Germany
(e-mai I: [email protected])
1. Introduction
Fast processes on the sub-nanosecond time scale and transient states in nano-scale electronic
systems are attracting high interest in basic and applied research. Full-field PEEM
(photoemission electron microscopy) imaging with high lateral resolution provides an
experimental tool to look into such phenomena. Owing to the excellent time structure of
Synchrotron radiation and pulsed laser sources, high time resolution can be achieved in
stroboscopic PEEM imaging. In the picosecond to nanosecond range remagnetisation processes
and transient domain states [ 1] or magnetic high-frequency eigenmodes in confined magnetic
structures [2-4] attract much attention due to challenging applications, e.g. in spintronics. In the
femtosecond range the lifetime contrast of “hot electrons" [5], the properties of plasmonenhanced optical near fields [6 ] or the dynamics of surface plasmons are directly accessible by
time-resolved PEEM. In a Mach-Zehnder like all-optical pump probe set-up, a temporal
precision of 50 attoseconds has been achieved by Kubo et al. [7],
In this contribution the state of the art of time-resolved PEEM is discussed with emphasis on
two classes of applications, i.e. imaging of fast magnetization processes and femtosecond
electron dynamics. In particular, we focus on stroboscopic imaging of precessional magnetic
excitations via XMCD-PEEM exploiting the time-structure of Synchrotron radiation (magnetic
field pulse pump - X-ray probe). The method is illustrated in Fig.l. In a special bunchcompression mode at the storage ring BESSY (Berlin), a time resolution of about 15 ps has been
obtained. Further, we discuss recent PEEM experiments using femtosecond laser excitation. A
current overview on time-resolved PEEM can be found in a forthcoming review [8 ].
CCD camera
Fig. 1. Schematic view of the experimental
set-up (left) and principle of stroboscopic
imaging (right). Hp denotes the field pulse
Field pulse pump - XMCD probe
(pump), hv the photon pulse (probe).
•(£ - g y 33S) s3pnji|diuB qSnoua 9 3 jb [ jo j s s q s iu B A |[ba\ p aj\[ sq i X|[buij puB p p ij S u ijp x s
apniqduiE guisE aioui qiiA\ ssseoioui y iq s s i q x 'iqSu sq j oj u o u iso d iBOupuiuiXs sjt j o
3 i[l j o
ino ||B,vi |bjiu33 3i(j j o y iq s UB3UJ b
3 q i 01
LJ3A3A\0H 'Z 'SiJ u;
3 A J3 sq o 3M
3|IJ|| Xj9A X [U0 3A0UJ [{EM |3 3 N - Q0 8 1
J3A\SUB |EOlUIBuXp 3qi
U A \o q s s s u s s sq j in oq Sn ojq i
3 lU ‘X|JE[110llJBd ‘pUB S ||B A \ UIBUlOp 3qX -UOlJBipXS
3A J3SqO
3A\ pUB
‘J01B [[j3S 0 U 3A U p B
S3|qUI3S3J UI31SÁS 3qj
*pB3isu | Abm |b ia u i b ui 3 S |n d p | 3 i j sq i a\o[|oj iou sso p X|snoiAqo 0£ ' S y u i p sp id a p sb ssu o d sa i
[ESIUlBUÁp 3 q i J3[31B |d 3q} JO s p i s 3 u O | 3 q i
|3|[BJBd p 3 J 3 3 J ip SIXB p p ij 3AI]33JJ3 3 q l pUtlOJB
U 0 lS S 3 3 3 jd B JO 3pO U I UOIJB1I3X3 lU B U IU IO pS jd 3 q i § U IJE 3ipU I
S31JBA SUjBUJOp 3 3 jB |
0 A\1 3 q i Ul XlISU31Ul a q x
siMBoiop s3jb[ oa\i sq i ssiBJBdss ||b a \ p s N - ,,0 8 1 UV
SUlEU lO p 0 M 1 sq i 3 [iq M ‘> p E |q p u B 3JiqA\ JB a d d B
'3i j
U3>]B1 3JE S 3 3 eU II
u o p ip d a i
sq i ui saqounq u o ip s p
Ol p U B IB d llU B p u B p[]BJBd paiUSUO SUfBUlOp
f\j u oiiezipuS eui
b bia pszm ojqouX s 9jb
sjqB U B A
p 3 A [0 S 3 J -3 lU |X
'q č
'§ !X
a q i j o uoprqoAS a u i u s q j j o
s a s p i d aqojd sq i p u B lusxino SuiAup
q iiM ( a p o r n
p s a n p o jd s s s jn d X e j-x
a q i oiui lu a u n o
q l|M
jBsddu qioq j o i jE jnoipusdisd p siu s u o
u i UA\oqs s j e p p jE jd X o q B u u s d sq i u i
sioqsdBUS X sp p o i u o j p s p
' ( l q 3 u ) pjE A \U A \ 0 p p u B ( J J 3 |) p iB A \d n f X
‘s d £ = whmh/)
3 [duiB S s q i j o
UA\OL|S SB p p ij
3 u u uoiiojqouX s
u o i} E u iu in ( i! a i d o o s o q o a i s
O ip u S B U I
g U |iB J3 U 3 §
.w X ousnbaij-qgiq b s p s ju ; joibj3us8 s s p d saem ojoiui y
([ l] UIOJJ E1EQ ) (p)
3 S |lld P |3 1 J
[sd] aun;
OOOt' 000£ OOOZ 0001
( j _j )
u iru p sd s
X o u 3 n b 3 jj
1 |3 | 3 q i Ul J01.13A U0I1EZI13U§BUI 3 1 [1 J0
‘o s u o d s s j D ip u g e u io jo iu i
:(b )
1U3UI3J0UI s d o o i
ib U3>[E1 S3SBUII
m i«
:( 3 )
sd OOP‘
U0I1E10J S'I
p [3 ij d ip u S bui Su;ipx9
sS pa
u o u d io s q B
W33d-0DWX J° sousnbss "i - S i j
(ZH 9)/ ^3U3tlb3JJ
20 "
oo r
ro “
sdooil * -
'iap iB |d sqi j o spis y o q s sq i Suo[B iusuoduioo
UOUBZjlSuSBUI 3UB[d-UI 3l|l 01 3Ajl|SU3S 1S01U 3JB 3A\ 03USppU| ŽUlZBiS l y 'apinSsABM JBUB|dOO B
u o p o 3 B [d ‘s s 3 u > |D iq i i u u o i p u B ‘q j 8 u 3 | u ir i
ao | s ijn s s j SM oqs z
‘qipiAV u ir l 9 1 q i|A \ s i s p i B [ d ( ° ‘ 3 J 08iJs[) X o j j B u u a j
d ■f \ l °1 I B U o i y o d o j d s i q s iq A \ ‘p x i d q o B 3 ib X jis u iu i X s b a q i 8 u iiE in D iB 3
X i i o q s q u o i o q d a i i s o d d o q iiM U 3 >[B] s s S b u i ! o a \ i i u o j j p s u i B i q o 3 j b s s S b u i i ( a j p \ I X ) u i s i o i q o i p
J B |tl3 J I3 3I10u8bIXI X bJ~ X
-I0133A U O |lB ZIJB |O d ( jB jn C Ijp ) pUB
3A|iB|3J sqi qiiM sauBA ppiX u o j p s p sq i J u a iu s p 3;i3u3bui b j o sagps uojidiosqE Eq 3qj 01
psuni si suoioqd pszuB jod X pB jnajp j o XSjsus sqi u sq ,\\ 's u o j p s p XjBpuooas j o ppiX u o j p s p
b ia
Xbj-x s q i j o u o u n q u isip [BiiBds sqi s p a p p lA133d 3lll I '3! 3 J ° d n -p s aqi uj
suiajsXs pauyuoa ui sapoui-uaSp sabav uids jo SuiSbuii p 3A|0 S3j-au i|x 7
Fig. 3 Left: Sketch and snapshots of magnetic domain patterns at the time t = 0 excited with increasing
field amplitudes I, 11 and III. The bottom panel shows comparison of an analytical model and
experimentally observed mean shift of the central domain wall. The inset shows an analog mechanical
model. Right: Within a closed system, entropy maximization increases disorder. An open system with
constant energy dissipation allows for an increase of local order. (Data from [2])
The magnetization distribution adapts itself to increase the energy dissipation and thus causes
an overall increase of local order. The near-resonance spin wave mode causes an effective force
perpendicular to the 180°-Neel wall that is balanced by the restoring force of the stray field
energy. The system is excited with a significant oscillating field component of 1 GFlz (cf. Fig.
2d), i.e., just below the resonance frequency of the free running system of about 1.25 GHz. If the
domain wall shifts to the right, the resonance frequency will decrease in the left domain;
therefore the amplitude (and the total energy) of the precession will increase in the left domain.
3. Observation of optical near fields and plasmon eigen-oscillations
In experiments using all-optical fs-laser excitation the delay between pump and probe pulse can
be generated in a Mach-Zehnder interferometer like element. Thus, any electronic jitter is
avoided and the time resolution is finally limited by the properties of the laser pulses. In phaseaveraging experiments the resolution is of the order of the pulse lengths (about ten to several
tens offs). In interferometric experiments the precision can be driven down to 50 attoseconds [7]
(the exciting and probing waves still consist of several periods). Kubo et al. have clearly
demonstrated changes of interferometric time-resolved PEEM images on a time scale of 133
attoseconds. Femtosecond-laser excited PEEM also gives access to “hot-electron" dynamics [5]
and to plasmon-enhanced optical near fields [6 ], An example is shown in Fig. 4. Upon excitation
with femtosecond laser pulses at 400 nm, a Ag crescent fabricated on a Si surface does not
appear in its true shape, but instead the spatial distribution of the optical near field becomes
visible (b). A simulation reveals strong enhancement of the optical near field at 400 nm (e), i.e.
the crescent acts as a resonant “nano-antenna” tailored for maximum response at 400 nm
excitation. In contrast to SNOM, the PEEM technique does not influence the near field structure.
oos9N£o j m a -w PaPuní
Ijaftuj •juoddns awju/oaq pooS uof ffojs XSS39 3V °l SD IPM sv uoiivuadooo lUdjpoxB uof
(zuwp\¡ ‘ijjAD3S3}¡ A3WÁJOJ .to/ ajnjiisuj youojj XDjAj) uaipuyjq pun ‘(ifoijnf z j ) jsppui/og 'WO
‘(zuw¡/\[ 'MUfi) js u s p o y ‘o y fid a fjy s 'u a a / j '/ " / / ‘ih31{duij -ja¡ ‘ynúsvuyy o; dnp 3jd syunm
‘s3 > ]m b h d Aq 'p a ‘^ sais/C q d M jo ajg puB S u iS e u ii u i s a a u B A p y „ u i ‘j b ja a s u a q u o q a s '9
£311 (S003) S S J3 JP T OUBN ‘ |E p o q ir f l y
I09Z.W) (S003) S6 WH Aay 'sAqd ‘je p ijjaqou^ PM [9 ]
£33 (3003) XL Q s^Md |ddy ‘ [b P jpiuiqos 'O [s]
PQZLXZ (S003) t>6 '«31 Aay -sAqd ‘-[B p aqEEy 7 [*,]
03fr (17003) ť0£ 33U3PS ‘'1®13 soq 3 g - s [£]
103¿03 (S003) S6 »31 Aay sAqd ‘ |e p >[nXsBj>i -y [3]
31-1 (9003) 9PZ a M131AI -Jjsui p n ^ “|B p asuaquoqas D [l]
JU3lU§p3|MOU5)3y / S33U3J3jay
( a p jdsD uoD -aD B jins'M M M ) OMj j o j o p B j b jn o q B Áq p a A O jd u u a q h im puB
s d 0 3 1 jn o q B X p u a s a jd si u o ijr q o s a j a u iij s j[ X jo u is u i u iE jS o js iq ( j ‘X‘x ) q £ b u i p a jo js s j b jB qj
SJU3A3 u o x p a p a jS u is s a j m b a e ji a s r re a a q j n ja s n X p E in a ija e d si j o j a a j a p a u i|X E p p a q x
aíimiu 3uiA¡ossu-diu¡t
S u is n qoB O jddB u b si S u iS b u ii a id o a s o q o jjs
o j a A ijB iu ajje S u is iu io jd
X |q 8 ;i| y - s a in o s u o jo q d a q j j o X jqB tib a q j u o X ijE iju a ssa s p u a d a p jB qj u o ijr q o s a j a u iij b qjiM
puB |B |ju a jo d q g ;q A js a q jiM a n b iu q o a j 8 u i8 e u ji p A o u b s i J M g g d p a A jo s a j- a u iij ‘u o i s n p u o a u j
([9] u io jj) uoijB[niuis 9i[j ui ajqisiA si (p) uiu
oj pajEdiuoa se (a) uiu oofr 1B PPL! JEau [Eoijdo aqj
j o juauiaauEqua S u o jjs '(a) a§Eiui p ^ g s ai|j ui UMoqs (ig uo) juaasaaa 8 y E jo MaiA j o p p ij |EDijuapi aqj
j o (q) aSsuu [Ajggj-uojoqd-OMj p a jp x a jasE|-puoaasojuiaj e i|J[m pajEduioa (e) sSeuii w g g j - A il
'P ’^!j
spiaij JE0U|BC|jdo jo uo!)e|nui|S
M. J. van Bruggen, B. van Someren, P. Kruit
Delft University of Technology, Faculty of Applied Sciences, Particle Optics Group
Lorentzweg 1, 2628 CJ Delft, The Netherlands
We present a irmltibeam source (MBS) based on splitting the broad beam of a Schottky Field
Emission source into many sub-beams with a microlens array (see fig. 1). To avoid problems
associated with skewed incidence in the microlenses we use an electrostatic collimator lens to
illuminate the microlens array with a parallel beam. However, this collimator lens introduces
unacceptably high deflection aberrations. We show that it is possible to create a negative lens
in that MBS with negative spherical and chromatic aberration coefficients. This lens can be
used to correct the off-axis aberrations of the collimator lens. The MBS will be mounted in a
Scanning Electron Microscope (SEM) to produce multiple 1 nm beams. This multibeam SEM
will be used as an Electron Beam Induced Deposition nanolithography machine [1],
In 1947 Scherzer published his ideas about a foil corrector that could be used to correct for
both spherical and chromatic aberration in electron optical systems [2], However, the
scattering of the electrons in the foil was a serious problem. That difficulty can be
circumvented by using a silicon wafer with holes in it that are much smaller than the diameter
of the macro electrode in front of it (see fig.2). In our case, we use the aperture lens effect in
these holes to focus each sub-beam individually, while the negative macrolens can be used to
compensate for the off-axis aberrations introduced in the collimator lens. This silicon
microlens array, electrode number 3 in fig. 2, is fabricated with Micro Electro Mechanical
System (MEMS) technology with sub-micron accuracy.
The configuration in fig. 1 is modeled using a thin-lens approximation with only one first and
third order coefficient for each macrolens, describing the deflection angle at that lens:
A a = a ih + a ih ' + 0 ( h 5)
Here aj is the lens strength and a_s the coefficient describing the third order geometrical
aberration from the perfect lens effect; h is the height of incidence in the lens. In fig. 3 it is
shown that the aj coefficient of the negative lens can be negative, depending on the distance
between electrode 3 and 4.
With this model, expressions were found for the shifts in the two perpendicular directions in
the paraxial image plane, separated in powers of h. This results in the Seidel aberrations with
their coefficients expressed in terms of the single third order coefficient as- We expect that
this approximation is valid as long as we use the model to see if there is an optimum setting at
which both the third order geometrical aberration and the first order chromatic aberration, also
taken into account in our model, o f the two macrolenses exactly cancel. A full 3D field
calculation and ray tracing of the configuration at the optimum setting will be necessary to
check the validity o f the model.
'LP61 >11=3
.mpiauo'x 3q3Si}BUiojqo pun a q o su B q d g i o ‘js z ja q o g
Ç00Z asa/AON ‘(9)ZZ
‘g |ouqo 3i Tg 'p s dbá T ‘uoiiisodsp psonpu; uiBsq uoapap uiu oi-qns j o j 3Djnos
uiB3q-uoaj03p-ij[niu b j o jusuidopASQ i'd ‘}|nj->| !uba g ‘uaisuios íuba 'f ja ‘uaSSnjg
‘> ]ijd o
(l u l u )
[3 ]
(LULU p SI p 3p O Jp 3 |9
P'P = •’A
‘A>l O Z = EA) P Pub £ spoapap
J3J3UIB|P 3qi PUB A>1
-ounj B SB SU3[OJDBUl 3ApßS3U 3qj
JO fB }U3piJJ303 J3pjO-£ '£ 3jn§¡ j
'S3pOJJ33]3 3AIJ33dS3J
B S3qSI]qBJS3 p pUB £ SpOJJOSp U33MJ3q
p p ij s q x 'J0jbuii[|03 sq j j o p p ij su3[ sq)
o i o j j ssjn jjsd B S u u |u ii|-ju 3 jjn 3 sq j s p p iq s
pue (g ) Xbub su 3]ojoiu j puB sjn jisd B
S u i}iu ii[-ju 3 iin o p su iq u io o sq j sb (Bpuajod
3UIBS aqj u o si 1 joq u in u ‘uresq pBOjq
sq j 3jbuji[|00 oj p ssn s; [ ju 3 iu 3 |g sgiA I
nqi j o u ojjE inSijuoo a p o jp a j g *j 3 J n 8 |j
JO } B U Il[|0 0
q ) IA \
uiBoq iqtitti posodojd sq) jo
A\aiAJ3A0 (BoqBuisqos •j a jn S ij
e r
P. Wandrol and I. Mullerova
Institute of Scientific Instruments ASCR, Brno, Czech Republic; [email protected]
In the low voltage SEM (LV SEM) specific approach to the signal detection is needed. As
usually, secondary electrons (SE) can be detected by the Everhart-Thornley detector situated
in the specimen chamber or inside the final lens. However, energy of BSE is too low to
produce a high quality image with standard scintillation or solid state detectors, and it is too
high to allow for efficient extraction with a lateral electric field. Some o f newest in-lens BSE
detectors employ transformation to SE3 [1] or acceleration toward a scintillation detector [2].
Most efficient BSE detector for the primary beam energy 5 keV and higher is the
YAG single crystal scintillation detector [3]. Low energy BSEs do not generate the number of
photons sufficient for quality image - before they hit the scintillator, acceleration of them to at
least 5 keV is necessary. Acceleration is carried out toward a positively biased electrode,
which is formed by a thin layer of the indium-tin oxide (ITO), deposited on the scintillator.
This layer is transparent for BSE and represents a high-voltage electrode at a potential up to 8
kV. Naturally, the electric field accelerates the secondary electrons, too, so an energy filter at
a negative potential of hundred volts is placed between the specimen and scintillator. A
grounded tube inserted in the scintillator bore shields the primary electron beam, and the
entire scintillator is surrounded with a grounded shielding as well. Schematic view of the
improved BSE detector for LV SEM applications is shown in Fig. 1.
The detector has been installed in the specimen chamber of a field emission SEM with
an immersion objective lens. In this microscope, secondary electrons move on helical
trajectories, being collimated into the final lens. This has been revealed by computer
simulation of the spatial distribution of fields as well as of the signal electron trajectories [4],
performed by means of software packages MLD, ELD and Trasys [5]. We have found that not
all SEs reach the final lens and 10 to 20 % o f them can hit the scintillator of BSE detector.
The proportion depends on the SE energy and emission angle. Computer simulations have
been confirmed by experimental results that showed essential the application of the energy
filter. Difference between signals acquired with the energy filter grounded or negatively
biased is demonstrated by images of a blast furnace slag sample in Figs. 2 and 3. With energy
filter grounded, a mixture of SE and BSE is collected (Fig. 2). Combined topographical and
compositional contrast with dominance of the topographical one is obvious here because the
SE yield at low energies is several times higher than the BSE yield. Only BSEs are collected
when applying -200 V bias on the energy filter. Then the distribution of metal oxides in the
slag is clearly visible (Fig. 3). In both cases the high voltage electrode on the scintillator is
biased to +4.5 kV. Fig. 4 presents the signal levels measured on a gold surface - the
characteristic energy dependences of SE and BSE yields are here appearing as limiting cases.
The detector can be retracted off the optical axis in order to suppress the impact of
BSE and to detect preferably the secondary electrons. For efficient collection of SE, the
energy filter must have a positive bias of hundreds volts (Fig. 5).
The above-described detector offers a solution to the BSE problem in the LV SEM
mode for every scanning electron microscope. Signal o f this detector is approximately three
times larger in comparison with a standard YAG-crystal BSE detector at accelerating voltages
from 0.5 to 3 kV. Besides detection of low energy BSE, this type enables one to detect also
the SE signal and an SE+BSE signal mixture.
' 10S0S900c3f>l ou l UBJ3 jspun y 3 SV J ° XousSy Ju ejq si|) Xq psyoddns s¡ >[aoM sq x
'95 (0 6 6 1) 86ZV MJ31M uinjjsui - p n ^ ‘-9 >(U|pss|M ‘ g çAoousg
'LL (frOOS) POQZ spuajx Jusos» aoaj in • y EiBjjny ' ¿ [ojpuEM
¿ c l (¿661 ) PI Suiuubds “ IA1 J3|iniM " ä uubiuj3h " y BJBJjny
9Z (£003) 11 ^ BP °1 XdoDSOPiw ‘ q q •^ p p A u sS p js
|IU JI|'S 3 J0 U
l|0 3 J p iU 3 ~ p |~ 3 § B d ~ 3 d X j~ J 3 p B O |3 S B d /U J O O -B jq
- jq D B } iq 'M M M //:d j} q 33S ‘8 g 'OU P S q S ‘BJEQ |E 3 IU l[3 3 X )U 3U in jJS U ] D IJ!JU 3 |3 S ll |3 B l|H
'A>I ç > IB SI 3p01j03[3
[9 ]
[ |]
s S b JJOA
q3iq aqj ‘a 00£+ s! J31|LJ X8j 3U3 sqi jo SEig
j o p s p p g s g sm Xq psjinboi: saisqds xajB]
3U3jXjsX|od psjB osun jo sSbuii 3 s u y -ç -813
’A5! Ç'P Ie 3p o jp 3[3 3§B}]OA q8 iq
sq j ‘A 00S" oi pssBiq -13}[IJ X8jsu3 sq j iuiEsq
/ÜBiuud a 3>( I 1® 38bui| g § g srux £ '813
U O IJEO I|ddB 3 3 e )|O A
[Bijusjod SuijBJspoDB sq j snsaoA
p sjjo jd ‘jo p s p p sqj u io jj puSis in d jn o > '813
'A >1 Ç > JB 3 p O j p 3 [ 3
s S b jio a
q8 iq
s q j i p g p u n o j S J 3) [ i j X S j s u a s q j
qjiAV S |B u 8 i s g s g
00 S + O» 0 ) J3H!i A ß ja u a
(A>l 8 <U dn)
puB g § j o
M n jx ijM
8 13
6 u jp | 3 ji| s p u e
p a p u n o jB
0 p o j } 3 0 | 3 a ß e } | O A L|BjL|
U O |JB J3[333B M O] JOJ
jo p s p p
U O |]B ||U U |3 S
a p m B ) i { 6 ;|
'8 |J
s o a jd 3 |od
J 0 } e ||! ; u p s

Podobné dokumenty

proceedings 2004 - RECENT TRENDS 2016

proceedings 2004 - RECENT TRENDS 2016 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 tr...


proceedings 2002 - RECENT TRENDS 2016

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