Beruflich Dokumente
Kultur Dokumente
Handbook of Modern
Ion Beam Materials Analysis
Second Edition
EDITORS:
IMIRlsl
Materials Research Society
Warrendale, Pennsylvania
Published by:
Materials Research Society
506 Keystone Drive
Warrendale, PA 15086
Telephone (724) 779-3003
Fax (724) 779-8313
Web site: http://www.mrs.orgl
Library of Congress Cataloging-in-Publication Data
Handbook of modem ion beam materials analysis I editors, Yongqiang Wang and Michael A Nastasi. -2nded.
p.cm.
Includes bibliographical references and index.
ISBN 978-1-60511-215-2 (alk. paper)
1. Ion bombardrnent--Handbooks, manuals, etc. 2. Solids-Effect of radiation onHandbooks, manuals, etc. 3. Materials-Analysis-Handbooks, manuals, etc. I. Wang,
Yongqiang. II. Nastasi, Michael Anthony
QC702. 7.B65H36 2009
620.1'127--dc22
2009045714
Disclaimer of Liability
Neither the editors, authors, contributors, their affiliated institutions, nor the
Materials Research Society makes any warranty or representation, express or
implied, including the warranties or merchantability and fitness for a particular
purpose, or assumes any legal liability or responsibility for the accuracy,
completeness, or usefulness of any information, apparatus, product, or process
disclosed, or represents that its use would not infringe privately owned rights.
CONTENTS
ACKNOWLEDGEMENTS ...................................................................................... iv
CHAPTERS
1.
Introduction ........................................................................................................ 1
2.
3.
4.
5.
6.
7.
8.
9.
10.
iii
Acknowledgements
The development of this Handbook wat; enthusiastiacally
supported by our immediate supervisors, Anna Zurek and Toni
Taylor. It would not exist without the support of the Department
of Energy and in particular the Materials Sciences and
Engineering Division, Office of Basic Energy Sciences. We also
want to thank our families for their constant support during the
development of this Handbook.
We arc grateful.to the Materials Research Society (MRS) for
undertaking the formidable task of publishing this handbook.
The Director of Publications and Marketing, Gail Oarc, and her
staff offered valuable assistance in guiding the book to
completion.
We gratefully acknowledge the editors and authors of the first
Y.Q. Wang
M. Nastasi
Boston University
Soumendra N, Basu
CNRS-CERl, Orleans, France
Gilbert Blondiaux
Jean-Luc Debrun
Institute 0/ PhYSics and Power
Engineering, Russia
Alex Gurbich
Instituto Tecno16gico e Nuclear,
Sacavem, Portugal
Centra de Fisica Nuclear da
Universidade de Lisboa
Nuno Barradas
Juha-Pekka Hirvonen
Texas A&M University
Lin Sbao
McMaster University
John A. Davies
Paolo Rossi
University o/Wisconsin
Paul M. DeLuca, Jr,
Vanderbilt University
Robert A. Weller
iv
CHAPTER
D
INTRODUCTION
Y. Q. Wang and M. Nastasi
Los Alamos National Laboratory, Los Alamos, New Mexico, USA
Chapter 1
Chapter 16.
Finally, we urge that caution be observed when using beam
energies exceeding the Coulomb barriers of the target elements
or of the particles of the beam. Under these conditions, nuclear
reactions, whether wanted or unwanted, will occur, and there can
be associated radiation hazards from both prompt and induced
radiation. The extent of these hazards depends on the beam, beam
Introduction
REFERENCES
CHAPTER
ENERGY
Loss
J. F. Ziegler
United States Naval Academy, Annapolis, MD, USA
CONTENTS
2.1INTRODUCTION .................................................................................................................. 7
2.1.1 Basic concepts and deimitions:.................................................................................. 8
2.1.2 Units .......................................................:.................................;................................ 9
2.1.3 Compilations and programs .................................................................................... 10
2.2 THEORY AND MODELS OF ENERGY LOSS .............................................................. 10
2.2.1 Basic physics ..............................................................................................................10
2.2.2 The scaling rule .........................................................................................................11
2.2.2.1 Effective charge .................................................................................................. 12
2.2.3 Computer calculations of stopping powers ........................................................... 12
2.2.3.1 Comparisons of stopping programs to data .................................................... 13
2.2.4 Special interests for stopping data in silicon ......................................................... 14
2.2.4.1 Comparisons of helium stopping data in Si among differeut sources .......... 14
2.2.4.2 Comparisons of helium stopping data in Si:
crystalline versus amorphous ........................................................................................ 15
2.2.5 Elemental targets versus mixtures and compounds ............................................. 16
2.2.5.1 Bragg's rule ......................................................................................................... 16
Chapter 2
........................................................................................................................ 22
Chapter 2
2.1. INTRODUCTION
Figme 2.1
compositional analysis.
observed, most data points fall within 5-10% of the curve (from
http://www.SRlM.org).
...
00
...
.......
OIl
S
tJ
600
I.
I:
.,
>~
iI
---
lUI 400
....~
...0
(/l
S 200
:3
....~
S
::I
...<
10 2
10 3
value in units of keV/amu by ion mass in amu for ion energy in keY; e.g., 103 keV/amu for 4He corresponds to a 4He ion energy of
4000 keY.)
Chapter 2
t'
QJ
.c:
r,..
'+'
1.2
QJ
....S
s..
QJ
1.0
~
~
r:1
I I
....~
.,
..
0.8
10%
+'
CJ)
10 1
10 3
_.-
10 4
FIG. 2.2. Comparison between data and fit for the data ofFig. 2.1 (from http://www.SRIM.arg).
~..
and others.
.: .. t"
1rnn:Snitttea .
~artkI~
E:.i:AF!
:Zi
Zj
,MI
MJ
2.1.2. Units
Atomic mass units are used throughout this chapter. M J ;;;;: Ml[amu]
and M, ~ M,[amu] thus stand for the masse, of the ion and target
medium, respectively, in atomic mass units. For a compound or
mixture, take M2 to mean the average atomic mass; for example,
Sl02
= (MSi
+ 2lVlo)/3
p[glcm3]. Vo
medium is
is the Bohr velocity for atomic electrons.
Unit conversions relevant to energy loss calculations are presented
in Table 2.1.
,= (l1N)(dEldx) or ,= (l/p)(dEldx).
Units: [] = eVJ(1015 atoms/cm2) orkeV/(mglcm2), etc., as N is the
volume density (e.g., in atoms/em3) and p is the mass density (e.g.,
in g/cm'J.
Table 2.1. Unit conversions for energy, velocity, thickness, areal density, stopping power, and stopping cross section.
Multiply units
by
for units
Example
MeV
---
MeV/amu
vIvo
(MeV/amuJ,n
M,[amu]
0.1581
1.389 x 10'
1.661 x 1O'M,[ amu]
(MeV/amu)'"
mI,
om
om
lOJ~ atomslem2
MeV/(mg/cm')
keV/)!m
1015 atoms/em2
p[g/cm']
10
)!g/cm'
p[g/em']
10'
Jlg/cm
1.66IM,[amu]
2
eV em 1I 0 1:'i atoms
1.66 I M,[amu]
eV cm21101:'i atoms
10'p[g/cm'J
1.66 I M,[amu]
Chapter 2
Printed compilations:
Whaling (1958)
Ziegler(1977,1980)
Janni (1982)
Bichsel (1972)
Brice (1975)
[CRU (1993)
Winterbon (1975)
Ziegler (2004)
ICRU (2005)
Ziegler e/ al. (2008)
the
A short desaiption of
programs above is given in Section
A3.1 of Appendix 3. Parameters of some oftbe compilations,
tables, and programs are listed in Section A3.2 of Appendix 3.
10
dE
dx
(,,(E)
NZ, Z,e ) f ,
M,
(2.1)
Writing the stopping cross sections from Eq. (2.1) for two
different projectiles a and b having the same velocity v in a
medium ~. we obtain
:~
~
" . , 1 "",m
.-., .
:l!
.....r-' i
t , '.
..
,)
."
Lao
'J
0.001
,I
(2.2)
j
I
,,
I
equation:
,, " ,
Z",rr(V,Z,)e = r2,e.
I,
~
J..
."1
\ II
,I
I,
,
,,I
1.
".,Dll
!ntt",,;d.
[.~
,.,
(EI.h) I :MiN
"
(2.3)
1000
(2.4)
(2.5)
11
Chapter 2
Yu =A\I-exp[ -(B+C)]1
J'}
The program PSTAR gives <H(285, 715, and 1430 keV)" 0.489,
0.282, and 0.177 MeV cm'/mg, and thus, ed2. 5. and 10 MeV)"
4.40,2.54, and 1.60 MeV cm'/mg.
Ziegler (1980) gives 1 ror heavy ion scaling (Z, > 3) from proton
stopping in the energy range 200 keV/amu < E < 22 MeV/amu as
More
recent
Z,-
and
Z,-dependent
effective-charge
(2.6)
12
The scatter between the bulk. of the data is between 5% and 1C~.
.
'"
......
60
N~
I
>
40
.!i"
g:
.s
'"
.."
20
0.8
f-+--+-I-I----f--f.-f,H-
+--H-I-I--~ft~;.'j~t11
'"
10~
10.t.
( keV/amu )
Ion
FIG: ~.5. Measured stopping power data for 4He ions in aluminum from 43 data sets and 421 data points for 'He ions (top) and
deVlatton of the data from the semiempirical fit (bonom) (from hnp:llwww.SRIM.org). (Multiply units oHeV/amu by 4 for He energy
mkeV.)
\3
Chapter 2
14
Table 2.2. Experimental surface RBS yield (Ho) for Si as a function of energy and calculations using different semiempirical fits (rrGm
Bianconi et aI., 2000. RUMP refers to http://www.genptot.comi, SRlM2000 to ~ttp:l/www.srim.orgl and KKKNS refers to Kome et
al., (1998a)).
1.5
MeV
2.0
MeV
2.5
MeV
3.0
Mev
Experiment
58.5
28.8
18.3
12.7
9.47
RUMP
57.2
27.7
17.1
12.0
9.06
SR1M2000
57.6
28.1
17.3
12.1
9.09
KKKNS
60.5
29.5
18.3
12.8
9.69
Table 2.3. Stopping cross sections from several recent publications compared with serniempirical fits (from Barradas et al., (2002),
with the stopping cross sections from AST AR and SRIM-2008 added. "This work" refers to Barradas et al., (2002), Lulli to Lulli et
aI., (2000), Hoshino to Hoshino el 01., (2000), ZBL85 to Ziegler el 01.. (1985), KKKNS to Konac el aI., (1998a)). The measurelllent:;
of Barradas et ai. (2002) agree with those of Bianconi et al. (2000) within about 3%, and the calculations also reproduce the data to
within about 3%,
Sa ~eV/IO]5 at/emIl
Eolli'!e~
0.5
I.!
1.5
1.6
2.1
2.6
2.5
This work
70.9 3.9
62.10 41
60.43 39
54.20 34
52.75 32
47.5\ 37
46.32 40
42.0 33
4L0763
37.6573
This 'Norka
68.1 1.4
62.31::1:: 31
60:69 25
54.33 15
52.81L7
47.33 24
46.08 27
41.51 15
40_62 36
37.0644
68.76
61.15
59.35
-52.88
51.46
46.48
45.39
41.55
40.7
37.66
55.3 1.2
41.59
4504 1.0
50.3 1.1
ZBLS5 [23J
71.30
63.4
61.5
54.9
53.5
4&.4
47.3
43.3
42.4
39.20
KKKNS [7J
67.41
59.95
58.19
51.84
50.45
45.57
44.5
40.74
39.9
36.92
ASTAR
66.20
6004
59
53.5
52.3
47.8
43.1
39.30
61.2
59.5
53.2
51.8
45.9
42
38.10
SRIM-200&
Section AJ.5 in Appendix 3 shows figures ofhdium ion energy loss data in silicon and fits to these data from semiempirical calculations.
IS
Chapter 2
Table 2.4. Ratio of the energy Joss for "He ions in the axial direction to the energy loss in amorphous Si. The values have been read from
published graphs.
Energr~eV)
Sto~~ing
0.96
1.0
1.0
1.0
1.0
Axis
(100)
(100)
(100)
(100)
(100)
0.78
0.80
0.84'
0.88
0.89
2.0
2.0
2.0
2.0
2.0
2.0
(100)
(100)
(100)
(100)
(100)
(100)
0.71
0.90'
0.85'
0.74
0.64
0.76
(110)
0.54
0.57
0.53
0.52
0.48
0.84
0.72
0.71
1.0
1.2
1.5
2.0
2.0
1.0
1.2
1.5
(I 10)
(I 10)
. (110)
(llO)
(111)
(111)
~lll~
ratio
Reference
Absolute stopping data given; ratio calculated assuming KKKNS stopping (Konac et at, 1998a, 1998b).
>I<
as
16
(2.9)
where m + n is normalized to unity. Note that the appropriate units
should be eV cm2/atom, as this equation docs not hold for stopping
power in other tmits such as keV/!Jlll
EXAMPLE 2.6. 2.0 MeV 4He ion stopping in silicon oxide Si0 2
SRlM-2006 gives ,Si(2.0 MeV) ~ 46.88 eV cm'I10 1l atoms and
sO(2.0 MeV) = 38.36 eV cm2/101S atoms. For Si02, we then have
SiO
a = (lssi + 2(P)13 = 4l.02 eV cm2/101S atoms.
D,E=
Eo
J(dE/dx)dx
o
or
x= J(dEldxtdE.
(2.10)
The model works best for cases where the physical state of the
target has little effect and for target; similar to those in the original
experiments (see Ziegler and Manoyan, 1988). The largest
differences between the CAB calculations and simple Bragg's
additivity rule predictions are found around the stopping
maximwn. The differences reduce with increasing energy,
disappearing for ion velocities above lOvo. The average accuracy
of the calculation is better than 2% when compared to data from
severnl hydrocarbon targets.
[e
,Iii!.
17
Chapter 2
EXAMPLE 2.7. ProIOn depth scale in carbon. What is the 2.0 MeV
proton energy lost in a carbon target for depths of (a) 1000 nm and (b)
20~?
The
swface
energy
approximation
dE
AE = dx (Eo) &.: or 8E = E(Eo)Nl This
also be used to obtain the mean energy E... =
which
mean
energy
approximation,
yields
approximation
can
Eo -
dE
AE= dx (E av )&':or8E=E..)Nl
2.2.6.2. Thick targets
Numerical integration of Eqs. (2.10) proceeds as follows:
Estimate hE (or Llx or Nt) as above. Divide the target into n slabs.
Use the surface energy approximation in each of these slabs; that
is, calculate the energy lost in the ith slab, .6.Ei. and the total ~ in
the target ao;
t.E =
(e;_,)(Nt);
(2.11)
i:: t.E;,
i-=l
...
where
dE .......-.,
dx. (Ei_d
and
18
-'--- ----._------_.
dE
t.E I =-(E
dx ,- ,)&.:.I or
t.E; =
The thicknesses COrrespond to areal densities (see Table 2.1 for unit
conversions) of 17.6 x lOll and 353 )( 10'8 C atomslcm2, respectively.
2.2.7. Isotopes
The electronic stopping for a given ion X with an atomic
nwnber Z, at an energy EJA is the same regardless of ion mass
number A. The stopping of 2 MeV 2H ions is thus equal to the
stopping of 1 MeV IH ions.
M!v)'
flE,X)
Eo
2.3. STRAGGLING
When a beam of charged particles penetrates matter, the
slowing is accompanied by a spreading of the energy distribution
of the particles. This phenomenon is called energy straggling. It is
due to statistical fluctuations in the number of collision processes.
In materials analysis with ion bearns, straggling broadens the
measured 'energy distributions and resonances and impairs depth
and mass resolutiuns (Fig. 2.8).
X=R
2.3.1. Models
~(E[MeV/amU])'
Z,
E-
(2.12)
ZI
Q' _ {O.5L(X),
_
Qn'
(2.13)
1,
(2.14)
where fiB 2 is the Bohr value for the variance (standard deviation)
of the average energy loss fluctuation (full width at half-maximum
heigbt (FWHM) ~ 2.355 rlj.
E[keV/amu]
(2.15)
25Z,
19
Chapter 2
exists.
~
'
(1976) model. Using the effective charge and scaling approach for
energy straggling and the Chu model and considering correlation
effects and charge change effects, they obtained an empirical
furmula fitting the data for heavier ion straggling.
Figures 2.9 illustrate He, Li, C, 0, CI, and I ion struggling in
carbon, silicon, germanium, tin and gold as calculated from the
..
'.r'
NABt
NAt
AB
Noting that mN
N.t
ions.
ofSi02.
We now have A.nBo = SilOlJ NS,it = 0.33 x NSiOz t = 0.33
atoms/cm and Not = 0.67 x NsiO::!
(2.16)
t=
10 18
0.67 x 10 18 atomslcm2.
---
..... _-----------
(2.17)
20
2.4. RANGE
The definitions of the range parameters vary in the literatw'e.
mops
R(E)=
dE
IN[S,(E~+S,(E~]'
o.
~ Range(~XHA.'/A).
The range of2 MeV 2H ions is thus twice the range of 1 MeV lH
ions.
21
Chapter 2
Bichsel, H., and Saxon, R.P. (1975), Phys. Rev. All, 1286.
M!vJM:av), where Mav is the average mass of the medium with the
natural isotopic composition.
REFERENCES:
Adesida, L., and Karapiperis, L. (1982), Radiat. Elf. 61, 223.
AhIen, S.P. (I 980). Rev. Mod Phys. 52, 121.
Andersen, H.H., and Ziegler, J.F. (1977), Hydrogen Stopping
Azevedo, G. de M., Dias, J.F., Behar. M., Grande, P.L., and dos
Santos, IH.R., (2001), Nuc!. Instrum. Methods B174,
407.
BillTIldas, N.P., Jeynes, C., Webb, R.P., and Wendler, E. (2002),
Nucl. Ins!rum. Methods 8194,15.
Bauer, P. (1987), Nud. instrum. Methods B27, 301.
Bauer, P. (! 990), Nuel. ins!rum. Methods B45, 673.
Berger, MJ., and Paul, H. (1995), "Stopping Powers, Ranges and
Straggling", in Atomic and Molecular Data for
Radiotherapy and Radiation Research, report IAEATECDOC-799, lAEA, Vienna, Austria, chap. 7, pp. 415545.
- ._.. Bernardi, F., Araujo, L.L., Behar, M., and Dias, J.F. (2006), Nuel.
Ins!rom. Methods B249, 69.
Besenbach.r, F., Andersen, J.U., and Bonderup, E. (I980), Nuel.
I/'IS!rum. Methods 168, I.
Bethe, H.A. (1930), Ann. Phys. 5, 325.
Bethe, H.A. (1932), Z Phys. 76, 293.
Bethe, H.A., and Heider, W. (1934), Proc. R. Soc. London A146,
83.
Betz, H.D. (1972). Rev. Mod Phys. 44, 465 ..
Cui, F.-Z., and Li, H.-D. (1985), Nuel. I/'IStrum. Methods B718,
650.
dos Santos, I.H.R. Grande, P.L., Boudinov, H., Behar, M., Stoll,
R., Klan, Chr., KaIbitzer, S. (1995), Nucl Instrom.
Methods B106, 51.
Fano, U. (1963), Annu. Rev. Nuel. Sci. 13, I.
Fixsov, O.B. (1959), Zh. Elcsp. Teor. Fiz. 36, 1517. See also:
Firsov, O.B. (1957), Zh. Eksp. Teor. Fiz. 32, 1464;
(1957), Zh. Elcsp. Tear. Fiz. 33, 696; (1958), Zh. Eksp.
Teor. Fiz. 34, 447.
Bianconi, M., Abel, F., Banks, J.e., Climent Font, A, Cohen, C.,
Doyle, B.L., Lotti, R., Lulli, G., Nipoti, R., Viclaidge, I.,
Walsh, D., and Wendler, E. (2000), Nucl. Instrum.
Methods BI61-163, 293.
Forster, J.S., Ward, D., Andrews, H.R., Bail, G.c., CostO, G.J.,
Davies, W.G., and Mitchell, LV. (1976), Nuel. I/'IStrum.
Methods 136, 349.
22
KiIrth, E., and Wedell, R. (1981), Phys. Lell. 86A, 54. See also:
Kiirth, E., and Wedell, R (1983) Phys. Status &>lidi
B116, 585. KfirJi, E., WedeU, R., Semrad, D., and Bauer,
P. (1985) Phys. Status Solidi B127, 633.
Gupta, S.K., and Bhattacharya, P.K (1984), Phys. Rev. B29, 2449.
Herau1~ J., Bimho!, R., Gauvin, H., Kubica, B., Anne, R., Bastin,
Lindbard, J., Scharff, M., and ScbWtt, H.E. (1963), Mat. Fys.
Medd. K Dan. Vidensk Se/sk. 33 (14). See also:
Lindbard, J. (1954), Mat. Fys. Medd. K Dan. Vidensk
Selsk 28, 8. Lindbard, 1., and Winther, A. (1964), Mat.
Fys. Medd. K Dan. Videnslt. Selsk 34, 4. Lindhard, J.,
Nielsen, v., and Schartl; M. (1968), Mat. Fys. Medd. K
Dan. Vidensk. Selsk. 26, to. LindhaTd, J., Nielsen, V.,
Scharff, M., and Thomsen, P.v. (1968), Mat. Fys. Medd.
K Dan. Vidensk. Se/sk 33, 10.
Nipoti~ R.,
Ianni, J.F. (1982), Atomic Data and Nuclear Data Tables 27, 147.
Stanford, CA.
Kido, Y., Konomi, I., Kakeno, M., Yamada, K.., Dolurui.e, K, and
Kawamoto, J. (1986), Nud. Instrum. Methods B15, 42.
Oddershede, J., and Sabin, J.R (1983), Phys. Rev. Len. 51, 1332.
See also: Oddershede, I, and Sabin, l.R. (1984), At. Data
Nucl. Data Tables 31, 275. Oddershede, 1., and Sabin,
I.R. (1987), Phys. Rev. A35, 3283. Sabin, l.R., and
Oddernhede, J. (1982), Phys. Rev. 26,3209. Sabin, J.R.,
and Oddershede, J. (1984), Phys. Rev. A29, 1757. Sabin,
23
Chapler 2
Sigmund, P. (1982), Phys. Rev. Al6, 2497. See also: Sigmund, P.,
and Winterbon, I<. (1985), Nucl. Instrum. Methods B12,
1.
Springer, Berlin.
Paul, H, and Schinner, A. (2003b), At. Data Nucl. Data Tables 85,
377.
Sigmund, P., and Fu, D.-I. (1982), Phys. Rev. AlS, 1450.
Sigmund, P., and Hangemp.U. (1986), Phys. Rev. A34, 892.
MA.
Wilken, B., and Fritz, T.A. (1976), Nud. Instrum. Methods 138,
331.
24
WinterbOll, K.B., Sigmund, P., and Sanders, I.B. (1970), Mat. Fys.
MeM K Dan. Vidensk. Se/sk. 37 (14).
Winterbon, K.B. (1975), "Low Incident Ion Energies", vol. 2 of
Ion Implantation Range and Energy Deposition
Distributions, Plenum Press, New York.
Yamamoto, Y., Ikeda, A., Yoneda, T., Kajiyama, K., and Kido, Y.
(1999), Nucl. Instrum. Methods B153, 10.
Yang, Q., (1994), Phys.'Rev. M9, 1089.
Yang, Q., O'Connor, D.J., and Wang, Z. (1991), .'Iue/. Ill.'Itrum
Methods B61, 149.
Yang, Q., and MacDonald, R.J., (1993), .'Iuel. Iostrum. Methods
B83,303.
Ziegler, 1.F., and Chu, W.K. (1974), At. Dolo .'Iuel. Dala Tables
13,463.
Ziegler, IF. (1977), Helium Stopping Powers and Ranges ir-.:2 All
Elemental Matter, Pergamon, New York.
Ziegler, J.F. (1980), Handbook of Stopping Cross-Sections i"
Energetic Ions in All Elements, Pergamon, New Y011:::.....
Ziegler, 1.F., Biersack, J.P., and Littmark, U. (1985), The Slop~ing
and Range of Ions in Solids, Pergamon, New yo~ vol.
1.
Ziegler, 1.F., and Manoyan, I.M. (1988), Nue/. Jnsrrum. M,".hods
B35,215.
Ziegler, 1.F. (2004), Nucl. Instrum. Methods B219-220, 102.
Ziegler, J.F., Biersack, J.P., and Ziegler, M.D. (2008), The
Stopping and Range of Ions in Matter, LuLu Press,
Morrisville, .
NC.
See:
http://www.srim.org/SRIM%20Book.hlrn.
25
CHAPTER
CONTENTS
3.1 INTRODUCTION ............................................................................................................ 28
3.2 FORMAL THEORY OF ELASTIC SCATTERlNG ...................................................28
3.3 MECHANISMS OF THE PROJECTILE-NUCLEUS INTERACTION ..................32
3.4 PHENOMENOLOGY ..................................................................................................... 35
3.5 EVALUATION OF CROSS SECTIONS ......................................................................37
3.6 KINEMATICS OF NUCLEAR REACTIONS .............................................................37
3.7 INTERNET -BASED NUCLEAR PHYSICS RESOURCES RELEVANT TO rnA .39
3.8 EXAMPLES
3.8.1 EXAMPLE 3.1. Identification of peaks in the d + 27AI spectrum ................... 40
3.8.2 EXAMPLE 3.2. Resonances in the '7AI(p,y)'8 Si reaction .................................40
9
3.8.3 EXAMPLE 3.3. Resonances in the nuclear reaction IHC F,ay)"O and
19F(p,ay)160........................................................................................................... 41
3.8.4 EXAMPLE 3.4. Kinematic transformations for the 'He + d --+ p + 4He +
18.35 MeV reaction ............................................................................................. .41
3.8.5 EXAMPLE 3.5. Conversions between 3Ue(d,p)4Hc and 'U('He,p)'He ...........42
REFERENCES ................................................................................................................... .42
27
Chapler 3
3.1 INTRODUCTION
Pioneering achievements in ion beam analysis (lBA) were
made by nuclear physicists who developed a promising
application of existing instruments and knowledge to materials
investigations. Currently, IBA is a routine technique in materials
analysis, and few people using it in analytical work have even an
undergraduate education in nuclear physics. Nevertheless,
except for Rutherford backseattering spectrometry (RBS) and
particle-induced X-ray emission (PIXE) spectroscopy, all other
IBA methods are based on nuclear interactions, so at least a
basic knowledge of the underlying physics is needed to
competently employ these methods in practice.
Zze'
Vc(r)=
~'( r')
- - 3--
2R
R'
~R
for
for
r.:s:; R
(3.2)
10
T .. ---... ----
>QJ
15
r, fm
-10
-30
40
FIG. 3.1. The combined Coulomb and nucIear
potential of a nucleus. The classical turning
points for a panicle with kinetic energy T are
located at rl and r2.
28
Zze'
=-
(3.3)
dN
da= - = 27tbdb
N
(L6)
(3.4)
where j.l = Mml(M + m) is the reduced mass and rl and r2 are the
classical twning points (see Fig. 3.1).
da = n(Zze
dQ
(1. 7)
mv' 4sin4!!.
2
The conversion of this equation into a more practical farm
in the laboratory frame of reference is discussed in Appendix 4
(EXAMPLE A4.1).
'
iJ
.. i<'
+~ ____ ~-l\-=~~~~ _________ .
+Ze
2m
~'V+h'(E-V)\jI=O
FIG 3.2. Scheme of projectile scattering by
the electric field of the nucleus.
(3.8)
tan e = 2Zze'
mV'b'
k=-=-
'V = e. ik:,
(3.5)
P
tz
29
(3.9)
);.
(3.10)
Chapter 3
Here,
2m
tI'
L'l1jI + -EIjI =
(3.11)
1(9)~
(3.12)
(3.13)
(3.14)
r'v
LA,P'(cos9)Rk/(r)
,.,
(3.17)
(3.18)
Thus, the plane wave expressed through an expansion over
Legendre polynomials has the form
1jI=
[e -{
Taking into account the facts that the velocity does not
change in elastic scattering and that sin ede dcp = dn, it is
apparent that the differential cross section is expressed through
the square of the modulus of the amplitude of the scattered
wave:
dN
do=
N
(3.15)
ih
~(2/+1)i'
[{a-,~)'-e +-I~)l
.
~(cos9) e
'
e = 1..;,
,...
30
21kr
(3.19)
V(r), 'l'(r)
olk
,
------j'
..... -.
..,'/
(3.20)
(3.25)
The phose shifts o[ are defined by the boundary conditions
and are functions of k and 1but do not depend on the scattering
angle.
(3.21)
where 0, is called the phase shift.
Rk/(r) ~ e{ kr- 1
i+
25r
-e -{ kr-/i)
0'1
(3.22)
(3.23)
31
Chapter 3
st
and s~
number.
The only unknown~ in the cross-section relations are the
phase shifts 51. and as soon as they are determined, the cross
section can be -calculated. The difference between classical and
quantum mechanical results for the scattering process is striking:
In the classical approach, a particle having zero impact parameter
scatters straight back toward its point of origin, whereas the
angular distribution for the corresponding partial wave with zero
angular momentwn is isotropic (see Fig. 3.4).
E,
E,
..
JD
:=:=:=--:-~='\
'\
",t)
E,
E,
Light
Heavy
nucleus
nucleus
FIG. 3.6. Diagrams of the energy levels for light and
heavy nuclei and the corresponding cross section
behavior.
always highly excited because the absoroed particle brings both its
Even though the excited levels overlap for the compound
nucleus formed in the interaction of protons with a mediumweight or heavy nucleus, there are still some strong sharp
resonances in the cross sections. These isobaric analogue
resonances (lARs) are caused by the population in the
compound nucleus of so-called isobaric analogue states.
Except for electrical forces, the n-n, n-p, and p-p interactions
are similar. Therefore, the energy levels of isobaric (equal-A)
nuclei are relatively insensitive to the interchange of a protoD
and a neutron. The isobaric analogue states in the (Z+I,N)
kinetic energy and the bond energy of the absorbed nucleons into
the nucleus produced. The second stage is the decay of the
compound nucleus with the emission of some particle. The
original particle can always be emitted, in whicb case the original
nucleus is re-formed. A typical lifetime for a compound nucleus is
14
_10- 5, which is very long compared -with the time of direct
interaction defined as the time (10-23_10-21 s) needed for the
bombarding particle to pass through the region occupied by the
potential well of the nucleus.
32
1000
750
500
~
250
Potential Scattering
C 750
0
500
"'"
250
0
750
500
250
0
'"'"
Ie
()
J"
Breit-Wigner Resonances
Z,N+l
~"
Total
r~
0.5
1.0
1.4
2.0
1.5
Energy, MeV
1.2
1.1
13
0.9
"C
0.8
~-
0.
I'1A
1'1
1'11'1
natcr(p,po)
0.7
0.6
B'ab =141
0.5 L -_ _ _
2.65
i
o~ 1.0 F--zf'-'::"""'--':""=';""~4..4~""
464
1.3
~~-
2.70
_ _-~-~--~
2.75
2.80
Energy, MeV
FIG. 3.9. Isobaric analogue resonances in proton
elastic scattering from chromium (Nud. Instr. and
Meth. B217 (2004) 183).
respectively. The precise values for nuclear masses can be
found in Audi et al. (2003). Q-value calculators based on these
data
are
available
on
the
Internet
at
http://www.nndc.bnLgoviqcalc/ and http://nucleardata.nuclear.
Iu. se/database/masses/ .
(3.30)
where Mj, M2, M3, and M4 are the masses of the projectile,
target nucleus, outgoing particle, and residual nucleus,
33
Chapter 3
E, ,....,..:.:M:;,,- Eo +Q,
M,+M 2
where
(3.31)
(3.32)
r
.'.
A,Z
34
3.4 PHENOMENOLOGY
The interaction of a projectile with an atomic nucleus Car::::::l, in
principle, be descnbed theoretically as the sum of eleme .:tary
interactions between an incoming particle and each of the
nucleons comprising the nucleus. However, even assuming that
the potential of the scattering of ODe nucleon by the ~ther
nucleon were known (which is not actually the case), such a
description would encounter the insoluble many-body problex:n.
A,Z
Uso
(3.3 3)
where U c is the Coulomb potential defined by Eq. (3.2)
d/,(r)
dr
(3.35)
--
=( ~)' V
! d/"
d
I.s
/R(r)=[l+exp(r:~,
Jr
mnc
R , =rAlfl
,
35
(3.34)
so
(3.36)
(337)
(3.38)
Chapter 3
:a
~
400
160
9=1700
'
,,
200
s:
~exP(2il.:)[e'p(-2~:)+exP(2i~.)
where J..~
+ iJl~
ir. 1
E, - E--jr
2
l'
,,'. '.
, '
<J)
<J)
<J)
Ol
,/
'\
E 300
c
-;
100
.'
Me~-o-
0-
R
o~~~~~~~~~~~~~~
3.0
3.5
4.0
4.5
5.0
5.5
6.0
Energy, MeV
FIG. 3.12. Sensitivity of single-particle resonances to the
potential parameters (from Nuc!. {mfr. and Meth. 8261
(2007) 4ll1) ..
unrealistic for lower energies. Several attempts have been made to
develop a global set for the low-energy region. However for a
close reproduction of the measured angular distribution or
excitation function, the parameters of the model should be fitted to
experimental data specific to that energy range.
(3.39)
The optical model is also often used for the case of elastic
scattering of complex particles such as alpha particles. There are
quantities Eo, r, and rp are the energy, total width, and partial
elastic width, respectively. The subscript I is the relative angular
momentum of the proton and the target in units of 11. The plus
sign refers to the case when J = I + Vz, and the minus sign
corresponds to the case when J = 1 - Y2. The quantity tf1, is the
resonance phase shift. An example of the corresponding
calculations is presented in Fig. 3.9 of Section 3.2.
Generally. the results obtained using the optical model for the
scattering of nucleons from mediwn-weight and heavy nuclei are
quite reliable in the energy range of about 10-100 MeV. However,
at energies in the vicinity of the Coulomb barrier and in the subCoulomb region, some anomalies have been found in the
systematic dependence of the optical potential parameters on mass
number and energy. Therefore, the results obtained with global
sets acquired through optimization based on a wide collection of
experimental data measured at higher energies appear to be
36
2,.,
"AI(p,p/,AI
8=170'
~
~
:0'"
.s
~"
""6
"0
160(u,uO)"O
2000
-'"
~
100
Ll
'"
1000
E,,,7.6 MeV
1500
1000
-Theory
John et ai, e,.m.=169.2'
C"
1200
1400
1600
.800
2000
Energy (keY)
"0
""6
Hunt et ai,
eom =163.8'
500
"0
o~~~~~~~--~
100
12{)
140
160
160
eom.
37
200
Chapl.r 3
150
100
50
0
~
"-
(/)
......
.0
50
'6
'tJ
(3.42)
100
'-"
"
E=(M,+M,)V; E
,
2
+",
150
elab=150
M, E
M,+M, '
E~,
300
250
(3.43)
200
150
100
50
elab =110
(3.44)
0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Energy (MeV)
tane=
sin eo
y, + cos 8,
(3.45)
(3.46)
where
(3.40)
where Xl and X2 are the projectile and target coordinates,
respectively, is the center of mass (eM) of the system
comprising these two particles. This point moves in the
laboratory reference frame with the velocity
y, = M,M,
M,M,
(3.41)
M,
)-'12 ,
Eo
(3.47)
(3.48)
the collision.
(3.49)
38
or
dO"
( 1-y,' sin'
8)
112
<In [( 1-y,'
.
sin' e )112 +y, cose ]'
(3.50)
The azimuth angles <p and tpc arc equal, and for polar angles,
(3.5::2)
,
(3.51)
Resource address
Description
http://www.nndc.bnl.gov/qcaIC/
http://ntJ<leardata.nuclear.I~.se/database/masses/
http://www.nndc.bnl.gov/ensdfJ
http://www.nndc.bnl.gov/nudat2/
NuDat 2.5: Search and plot nuclear structure and decay data
http://www.nndc.bnl.gov/masses/mass.mas03
http://www.tunl.duke.edulNuciDatai
httpl/www-nds.iaea.orglibandll
http://www-ndsjaea.orglexfor/
http:((www~nds.iaea.org!sigm.acalc/
39
Chapter 3
3.8 EXAMPLES
3.8.1 EXAMPLE 3.1.ldentificatioD of peaks in d + l7AI
spectra
Figure 3.17 shows the yield data for the 27 A1(P,y)"Si reaction.
of Appendix 4.
E'
'"
~ (E'
where the masses MAl and MSi for 27Al and 28 Si , respectively,
were taken from Audi e( ai., 2003 (available online at
http://www.nndc.bnl.gov/m assesimass.mas03).
Eo:' keY
12741
11 400
'2
:0
12Q'J~
13128
IU21
3.8.3 EXAMPLE 3.3. Resonances in the nuclear reactions IHe~,a1)160 and 1!1F(p,o.y)160
Figure 3.18 shows the energy-level diagrams of different nuclei for the case of fluorine detection, where 19p(p,ay)16 0 has been lLsed
most freqlJently (see Chapter 7). Its reverse reaction, lH(I9p,O:y) Hi O, has commonly been used for hydrogen analysis (see Chapter- 8).
Resonance parameters for IHe9FJa.y)2~e were measured for the interaction between fluorine and hydrogen far the praton-indll..ced
at
reaction 19F(p,ay)2~e. The parameters are listed in Table 2.29 of Tilley et al., 1998 (also available
http://www.tunl.duke.edulNuclDataJ). A strong resonance is observed in this reaction at a proton energy of 872.11 keY. In the ceo..terof-mass frame of reference, the resonance energy is [see Eq. (3.43)]
Eres '_ EH
em
lab
MF
H+
For fluorine as the projectile, the resonance energy in the laboratory frame of reference is
It is worth noting that, if mass TIlimbers are used in the calculations instead of atomic masses, the error in the resonance energy is as
large as 130 keV.
The resonance width of 4.30 keY is given in Table 2.29 of Tilley et al. (1998) in the center-of-mass system. It is transfonned into the
laboratory frame in a similar manner:
4.30 1.007825+18.998403
1.007825
85.36 keY.
The resonance cross section for gamma production should be identical for 1~(p, ay) lliO at the resonance energy ofEp= 872.11 keY
and for IHe~, cry)160 at the resonance energy ofEl~F= 16.44 MeV.
position of the detector of recoiled alpha particles. The angle
trimsformation is made with help of Eq. (3.46). For registering
alpha particles at 130, the calculations give 138.5 e in the eM
system, with the values for the particle masses and for Qof the
reaction
being
taken
from
http://nucleardata.nuclear.
lu.se/databaselmassesl. Because the velocities of the outgoing
proton and alpha particle are aligned in opposite directions in the
eM frame of reference, the angle Scm for the outgoing proton is
180 0 _ 138.5 0 = 41.5 0 This value has to be converted into the
laboratory system, which gives Slab = 40.1 e. The cross section
0,(9) ~ 66.8 mb!sr for the 'He(d,p)'He reaction at this angle
(40.1) is taken from IBANDUEXFOR. Finally, the Crosssection 04($) = 47.7 mb/sr for detecting alpha particles at the
laboratory angle ~ ~ 130' is computed according to Eq. (A4.20)
of Appendix 4.
41
Chapter 3
"Ne
".,,1115
"l'
Ep
IkeVJ
SI'Oble
"Il
Stable!
Sf'obl!
'H('He,p)'He
!. . .
~.020
('~ .w.
"Ne
S'Jl
484
,3100
22'
. alJlCl"'i1 _
r-
12.81<
8,'15
~
6.06
In the eM frame of reference, the cross sections for the direct and
reverse reactions at any given energy are the same. The
transfonnation of the 3He{d,p)"He cross section into the CM
system is performed according to Eq. (A4.18) (see Appendix 4),
with the quantities A, B, C, and D being calculated assuming M I =
Md and M2 = M JHC ' To determine the result for the 2HeHe,ptHe
121
4969
14 248
'--
11.612
19FIP.aJ"O
1 .
19Ff p,n}"Nt
"Flp.yJ"N,
REFERENCES
Audi, G., Wapstra, A.H., and Thibault, C. (2003), Nuel. Phys.
A728,337.
Lane, A.M., and Thomas, R.G. (1958), Rev. Mod. Phys. 30, 257.
16,183.
Tilley, D.R., Cbeves, C.M., Kelley, J.H., Raman, S., and Weller,
H.R. (1998), Nuel. Phys. A636, 247.
Firestone, R.B., Shirley, V.S., Baglin, CM., Zipkin, J., and Chu,
S.Y.F. (1996), Table o[Isotopes, 8th Ed., Wiley-Interscience,
Berkeley, CA..
42
CHAPTER
Backscattering Spectrometry
J. A. Leavitt and L. C. Mcintyre, Jr.
University of Arizona, Tucson, Arizona, USA
M.R. Weller
Middle Tennessee State University, Murfreesboro, Tennessee, USA
CONTENTS
4.1 INTRODUCTION ......................................................................................................... .45
4.2 FUNDAMENTALS ........................................................................................................ .46
4.2.1 The kinematic factor and mass resolution ........................................................... 46
4.2.2 Elastic scattering cross sections ............................................................................ 46
4.2.2.1 Definition ofthe differential cross section ................................................... 47
4.2.2.2 Rutherford cross sections .............................................................................. 47
4.2.2.3 Non-Rutherford cross sections .....................................................................47
4.2.3 Experimental geometry ...... ..................................................................................49
4.2.4 Effects of energy loss of ions in solids ..................................................................49
4.2.4.1 Defmitions ................................................................................................. ,.....49
4.2.4.2 Depth scale ......................................................................................................50
4.2.4.3 Depth resolution .............................................................................................51
4.2.4.4 Surface spectrum beigbt. ...............................................................................52
4.2.4.5 Peak widtbs for tbin fIlms .............................................................................52
4.2.4.6 Mean energy in tbin fIlms ............................................................................. 52
4.3 SAMPLE ANALySIS ......................................................................................................53
4.3.1 Tbin-fIlm analysis ................................................................................................ .54
4.3.1.1 The peak integration method ........................................................................ 54
Multilayer films ..................................................................................................... 57
Use of elastic scattering resonance....................................................................... .58
43
Chapter 4
'.4
:1
.
.1
21
t
3
:~
"
~,.
.j,
oj
.s;ti.,
.'.~
-?
.f,!
.~
44
.~
.@
Backscattering Spectrometry
Here, Ni is the atomic density (atoms per unit volume) oftb.<:: ith
element, and t is the physical film thickness. If the scattering is
Rutherford (pure Coulomb) scattering, then 0;(E.9) c:tr::l. be
calculated from
4.1 INTRODUCTION
Backscattering spectrometry Using ion beams with energies in
the megaelectronvolt range has been used extensively for
accurate determination of stoichiometry, elemental areal density,
and impurity distributions in thin films. Measurement of the
number and energy distribution of ions backscattered from
atoms in the near-surface region of solid materials allows
identification of the atomic masses and determination of the
distribution of target elements as a function of depth below the
surface.
(zze')'
0'1l(E,9)= ~
energies (E~ , E~) of the high-energy sides of the peaks into the
expression
Ml sin 9(
Ml' -M,"
sin)9"
(4-A)
where ~ and ~ are the atomic numbers of the incident and target
ions. respectively. This equation is given in cgs units. A u.seful
number in evaluating this equation is e2 =:: 1.44 X 10-13 MeV- em.
For very thin films, E, the energy of the analysis ion immediately
before scattering can be taken as Eu. For thicker films, the Q1ean
energy of the analysis ions in the :film should be used for E (see
Section 4.2.4.6).
Q ions al Eo
I
,I
Detector
(4.1)
2000
E~ E~ Eo
III
j j
1000
(4.2)
(.)
500
10DO
Channel number
FIG. 4.1. Basic backscattering spectrometry. Experimental
geometry (top). Backscattering spectrum (bottom) for a hvoelement compound (AmBJ film of uniform composition on a
low-mass substrate.
n _ NB
AB 0" A (E, 9)
m - NA ~ AA . O"B(E,e)
(4.3)
45
(4.5)
Chapter 4
N oIJI ;mpASNO
- - andN AS ; npABN.
-_
A
M~
B
M~
(4.6)
(4.7)
to calculate t. Here, No is Avogadro's number, and MAB = mMA +
nlvts is the molecular weight of compound AmBn.
The
dE, =E.(=JdEM,.
(4.8)
ligh' ions ('H, 'He). higher-energy heavy iollS ("0, "CI), and
even lower-energy ions (He. 12C) for backscattering analyses,
1iM2 - EO ( dK )
<1M2
46
(4.9)
1i
,,
'.
Backscattering Spectrometry
(9 E) ~
es,
(.2..J
dQ(E) _1_
Nt
Q nee)
(4. I(J)
2000~~~~~~~~~~~
cr.(E,e)~002073(Z~~, J[sin~m-2(~:J]
1800
1600
1400
~1200
~ 1000
~ 800
600
400
20~~~~~~~::====~:J
o
CiR
[4.11)
20 40 60 80 100120140160
M2
FIG. 4.2. Plots of(dK/dMlfl versus target mass Mz for several
analysis beams (M 1 = ion mass). Units ofMI and Mz are u. The
alaR
=I
0049Z Z"
.
I
1
(4.12)
E~
alaR =1
O.0326Z Z'"
I
(4.13)
Eo<
47
Chapter 4
E:
where
16
E:
(4.14)
7U
14
12
MeV) at which the backscattering cross section (for 160 0 < SLab
< 180) deviates from Rutherford values by 4% for a target atom
of atomic number ~. Note that cross sections for backscattering
of lH at 1 MeV are non-Rutherford for Z2 :5 15, and cross
.. sections for 'He at 2 MeV are Rutherford for Z,:> 6. At present,
no practical method exists for rapid and accurate calculation of
these high-energy non-Rutherford cross sections; they must be
measured.
10
~
>
'He
~8
a:.Q
zj
W
4
2
0
10
20
30
40
50
48
Backscattering Spectromehy
d.
and by a factor of more than 100 for 4.27 MeV 4He. In another
d.
instance, 4He ions with energies of 8.1-9.1 MeV have been used
in superconducting
films by a factor of -25 over Rutherford (Martin et al., 1988;
Barbour el al., 1988).
to enhance the sensitivity for detection of 0
x
casal
x
cosS,
d.
d_
(IBM)
cos9 l
(4.15)
x
(Cornell)
cos 91 , cos 9 2
the
IBM
geometry
(4.16)
(atoms/cm3).
and
49
L....
is defined as
dE
'-=NE
dx
9=1t-19, 92 1
E:,
the energy loss per atom per unit area (areal density) of material
Chapter 4
compmmd AmB, by
2 3
(:!:)Al 0 =
(4.17)
=46xIO
--46 eV
A
(4.18)
dE
aE=J-dx
dx
= pN~
4--)(6)(10
11
molecules
=_-,cm~_'____....!!mlSo!!.l_
102~
mol
= 2.5xlO
3
12
All 0 J molecules
em'
0
dE)Al2 3
( dx
(4.19)
NAl.O,
em
0
8"'4 , =
seV
=N Al 20 3 e Al203
=2.35 X 1022 x 193 X 10-15
=46 eV
and
N~201 =3x2.35xlOl.l =7.1xl0 I2 0atoms/cml.
50
Backscattering Spectrometry
(4.20)
1+ (dE)
-[ (dE)
dx
dx
[5]= K -
in cos 9 1
1]
liE
out cos 9 2
[SI
(4.21)
(4.22)
(Kl:. _1_
cosS
H.
_1_)
(4.23)
cosS 2
(4.24)
[]:-' =(K,~_1_H::..._1_)
cos
9(
(4.26)
8x=-.
--
(4.25)
cos9l
AllO J
A110)
out,O
44
x 10-1!;
~t2 =
x E~.A1 +
x ~t,Al =
x 51
15
X 10-
Because both the energy loss factor and the stopping cross
section factor require the evaluation of energy-dependent
parameters, an integral over the relevant energies should be
perfonned. In many cases, use of the incident beam energy
e~t.~J
>=I
51
Chapter 4
AI,O, _
[ EO'<'1
]
-
"'1,0,
Ajt.:;.
_1_
cosS)
"', _l_
+E-.,...J
cosS]
approximation.
eVem'
and
yJO
[610"
"0
1
A10
1
=K s'''''
--+8 " - _
0 ill
cosS
..... 0 cosO,
+ 252 X 10-1
AIZO)
... =( ev)
A
.... (76.6-ev)
A XX
82.3- xx
'"
Of o =['j"N"x=
.0
where .c.EAl and .6.Eo are the energies of ions scattered from AI and
0, respectively, relative to ions scattered at the surface and x is the
depth in the target where the scattering took place.
l)
Then calculate the mean energy of the ions in the film. EO).
using
-(1)
AESEA
=EO---m2
(4.30)
(4.27)
(4.3) with E = E(1); then .6Ef~1 and E(2) are calculated using
(4.31)
(4.29)
i~l
where (; is the energy width per channel and [Eo] is the surfaceenergy approximation of the stopping cross section factor
[defined in Eq. (4.23)] evaluated at the incident energy. The
remaining symbols are those defined in Eq. (4.3). This equation
is obtained from Eq. (4.3), where the areal density (Nt)
contributing to one channel of the spectrum at the surface is
&1[EoJ [see Eq. (4.22)].
HA,O =
the energy loss of the ions passing through the film, AE~nEA,
using
AE~EA = ti(Eo)(Nt)~EA
Ho ~ ,,(Eo)OQ&I([Eo)COS
(4.32)
(4.28)
52
Backscattering Spectrometry
= crf!Q25
(O]~,O,
A1,0
24
15
Experimental parameter
76 cts.
_ crf!Q31i
0,0 -
eo
lAI,O,
0
24
326 X 10 '
~ 43 cis (or particle:;.lchannel).
Vacuum
PumE!-down time
calculated using Eq, (4.3) with E ~ E<O, where E'" is the final
mean energy obtained from the iteration process. If the scattering is
Rutherford and the film is not too thick, then
(Nt)ff) ~
1
-m
E
( Eo
Values
M.V
mmxmm
1.05.0
nA
10200
~C
5-100
k.V
min
Torr
min
1.5 x 1.5
15
5-10
2x10-'
15
)2 (Nt)SEA
1
Units
(4.33)
53
Chapter 4
supplied by the PHA); this factor was usuaUy :s 1.02 [or the
analyses discussed. in the follOwing subsections. Other methods
for determining the dead-time correction factor are discussed in
Chapter 15.
&, E'
1:;(9)
cr;(E,e)
.;(E)
Q'no'R (E,e)(~)
OR
Quantity
Symbol
Quantity
Q'
DTR
ni
Channel number (at half-maximum) of the highenergy edge oftbe signal due to scattering from
the itb target element at the sample swfa.ce
nj"
Channel number (at half-maximum) of the lowenergy edge of the backscattering peak due to the
ith target element
Hi.o
54
Backscattering Spectrometry
E:
E: - n,'+E' = (9101)(3.742O.OO5)+(83)
= (3413 7) keY
K =
E~ = (284IH) =O.752O.OO2
Eo
(37765)
K = E; = (34137) ;;O.904O.002
... Eo (37765)
- 0.2469 X 10-24 ~
sr
OCd (E ,170') _
R
21.53 X 10-24
(3.776)2
2
-1.510 X lO-z, ~
sr
(hom Appendix 6)
21
(0.049)(2)(26)'/3
(....)
ACd
0.998
3776
OR Cd
3776
(757 I)
n'=(660 I)
atom,/c:m 2
Then, the mean energy of the 4He ion in the film, .,(1), is
calculated (to first order) using Eq. (4.29) for the first-order energy
loss, <'>E~EA ,of the ions in the film and Eq. (4.30)35 follows:
Q'~20.QJ j.lC
n.-
DTR-l.008
n, - (91O I)
nA~(812
I)
55
<'>Em
2
Chapter 4
bilayer with the same areal densities for Gd and Fe. The
elemental layers are assumed to have elemental bulk densities. The
thickness of this replacement film serves as the estimate of the
physical thickness of the original film. The values of the elemental
bulk densities are N Fe = 8.44 x lOll atoms/em) and NOd ""' 3.02 x
10" atomslcm'(Appendix I). Hence, from Eq. (4.7), we have
2.55 XlO
191
=368lkeV
tGd
Fe
3776
1018
3.02 X 10 22
an =223nm
525 DID .
NFe =3.780.02
NGd .
1018 atoms/cm2
NGd - AGd
0.674
tCdF. -
Fe
(2.55 0.08)
NFe _ APe .
em = 302 DID
8.44 x loT'
=191keV
E(2) =3776 _
l8
(..Q...)
erR Fe
The film thickness is about 0.5 J.UIl. Note that the uncertainty in
the ratio NFeiNOd is about 0.5%, whereas the (Nt); values are
uncertain by about 3%. This is a typical result and reflects the:
cancellation of the hard-to-measure quantities Q and n from the
ratio.
56
..
Backscattering Spectrometry
have the nearly flat portions on top but will be sharp peaks; in
this iru;tance, the channel nwnbers of the peak centers should be
used in Eqs. (4.1) and (4.34) ror element identification (see
Appendix C in Chu.' al., 1978).
Mullilayer films
The peak. integration method can also be applied to the
(4.36)
and the measured integrated ANife peak counts (with A = Ni, B =
Fc), thus providing the Ai values for use with the peak
integration method. As indicated, the analysis of multilayer films
that produce spectra with non-overlapping peaks can be carried
out by hand, but the procedure is tedious and bears a strong risk
of numerical error in the calculations. Computer assistance with
FIG. 4.4. Tbe 1.9 MeV 4He backscattering spectrum of a trilayer film on a carbon substrate. The backscattering signals from the three
layers are clearly separated. The Ni and Fe peaks are not resolved.
57
Chapter 4
4000
CI)
<:
AJ,O,
C'
I~FAk-
::J
0
(.)
i:se
=101
9u.a
=170
Be
2000
G.
I
",.
I
I
I
I
I
.
......... ......_......
500
420
Channel number
440
FIG. 4.5. The 2.525 MeV lH backscattering spectrum of a bilayer film on a sapphire substrate. The fact that the
is about 60 times the Rutherford value at this energy allows the Be backscattering signal to be observed. section . .It is clear that
considerable work remains to be done concerning measurement of non-Rutherford cross sections. Non-Rutherford resonance crosssection data are compiled in Appendix 7.
Overlapping peaks
Figure 4.6 shows the spectrum of a thin film with
backscattering peaks that overlap. The film consist.. of 50
bilayers of Fe-Mo with a total estimated thickness of about 360
om. The individual layers are so thin that the peaks from
58
. ',"I'
t...
,<
.~
Backscattering Spectrometry
A,.
N,oiNMo ~
16
4000
Q' = 25!IC
&LAS
= 170"
1/1
C
o'"
U2000
500
1000
Channel number
FIG. 4.6. The 1.9 MeV 4He backscattering spectrum ofa 360 nm Fe-Mo multilayer (50 bilayers) film on a sapphire substrate. The
backscattering peaks for Fe and Mo overlap.
M
where the explicit expression for [E.];,' [from Eqs. (425) and
(4.17)]
m _ NA
---~-,;Jf"-'-xB
NB
(4.37)
(4.38)
59
L ..
Chapter 4
and m and n.
Note that the ratio [eo]:' /[eoCI contains the ratio min, which
is, of course, the object oftbe calculation.
3000
2500
2000
-8
III
. 1:
~
1500
.',
- ....
~.~
,,,,;:::;;.,,,,
~
.,
Mo
1000
','
Fe
500
o
460
560
660
760
Channel
860
960
'.\
FIG. 4.7. The least-squares-fit of specified peak shapes (two half gaussians and a trapezoid) to the data (dots) of Fig. 4.6 provides
integrated peak. counts, Apt and A Mo for use with the peak integration analysis procedure. This fitting procedure is applicable to films
that bave uniform composition.
This (t:o] ratio seldom departs from 1.00 by more than a few
percent. An itemtive procedure can be used to obtain the min
ratio. For the first iteration, min is calculated from appropriate
H:, H~,
(fA.
and O'B values using Eg. (4.37) with the [eo] ratio
60
,
:/.
Backscattering Spectrometry
4000
Q'
= 50 JJC
NF.
NMo =1.236O.OO6
~::;,
o
2000
o
AI
I..----_-..J(
o
500
1000
Channel number
FIG. 4.8. The 3.8 MeV 4He backscattering spectrum of the film whose 1.9 MeV spectrum is shown in Fig. 4.6. The backscattering
peaks for Fe and Mo are separated.
(4.42)
(4.39)
where AEA and AEB are the energy widths (full width at half-
min. This ratio can be evaluated using a combination of stepheight and peak-width data. From Eq. (4.39) or (4.24), we have
[1:1" _ABA
[Elii' - ABB
(4.40)
n
NB H3 CJA AB.
---_______ .______
(4.41)
avoids much of the labor associated with the peak width method
as applied in EXAMPLE 4.6 is a combination of the peak width
61
Chapter 4
N
Calculation of the stoichiometric ratio, ~;
N.,
For the first iteration, use of Eq. (4.37), with the ['0] ratio taken
as 1,QO, provides
9 = 170'
8,=10'
n = 0.78 rnsr .
K,,(I 70') = 0.7520
(from Appendix 5)
K" x Eo ~ 1423 keY
~ = (1.821 0.004) keV/chanoel
cr~"(EI'1700)J =3.521
( "~'(E,,170')
'" 9.251
(~J
OR
=0.996
Fe
1.21 08.
(1.21)
C El = (1.05 0.03)
"",(170) = 0.8471
l""o
[EO]~~
(from Appendix 6)
(~J
erR
=0-'lM~of.,-
+[(0.8471)(93.0)+
[(1.21)[(07520)(72.0)+
.
=0.992
Mo
;:~~] :
+[(0.7520)(93.0)+_1_02_J
0.985
=(1.03O.OS) .
N pe
N Mo
(1.210.OS)
(103 0.05)
(117010)
.
H;:;
Q'=20.01 ~C
nF" = (777 I)
n,; ~ (549 I)
TIM, ~ (878 I)
OM,' = (643 I)
AF,M, ~ 564380 - (8 . 3)(371) = 561412 1343 ets
62
.~
Backscattering Spectrometry
=(4153) keV.
AE"" =(4283) keV.
~=
NMo
=(l.l70.08).
Calculation of areal densities (Nt)Fe and (Nt)Mo:
In the surface-energy approximation. use ofEqs. (4.38) and
(4.39) yields
6E Ee
(Nt)SEA
Fe
[Eo Jj:eMo
e /m
(415 3)
(278 14)
1013
X
(Nt)5EA
Mo
.:IEMo
(E lFe.YoIn
o Mo
(428 3) lO'S
(347 17) x
NF.
Anatysls method
(NtlF,
(Nt)!to
142:4
142 5
fl44 1.24.0.02
1447
136 7
116::t: 5
1.17:t. 0.08
1.17 0.10
NMo
Stoil:::hiometry by spectrum
4.5, Fig. 4.6
hej~ht, Exam~le
To first order, the energy loss of the analysis ions in the film is
[Eq. (4.29)]
.~
With the data from Fig 4.6 listed in Example 4.5, Eq. (4.42)
yields
[Eo];' and [Eo t,' faclors by Eq. (4.38) to obtain more accurate
AFe
[e(E(1) )]~O
m
1+
Me
(H~~8
)( AEMo )
HFeMo
AEF
e
FerO.
561412 1343
- 1+ (152040).
C4283)
(700 40)
(415 3)
eMO
[ (-W)t
AFeMo
63
Chapter 4
The symbols are defined in Tables 4.2 and 4.3. Note that the
ratio [E,l~n I[e,l:n in the generalized form of Eq. (4.37)
contains' the ratios min, ... , mlp, which are, of course, the objects
of the calculation. As before, these [~] ratios seldom depart
from 1.00 by more than a few percent. An iterative procedure
can be used to obtain the min, ... , mlp ratios. For the first
iteration, min is calculated from Eq. (4.37) with the ['o) ratio
taken as 1.00. These ratios can then be substituted into Eq. (4.43)
to calc':!late the [so] ratios, which are then used in Eq. (4.37) to
obtain more accurate values of min, and so on.
(Nt)~
(Nt)~
values (obtained from the SPectrum) into Eq. (437) with the ['o)
ratio taken as 1.00 yields the stoichiometry listed in Table 4.5.
Atomic fractions can be obtained from the stoichiometric ratios
provided by Eq. (4.37). For example, the 0 atomic fraction, F o,
is givcn by
(4.44)
where,
for
example,
Table 4.5. Summary of analysis results obtained from the spectrum
of the ~ramic glass sample shown in Fig. 4.9.
(4.43)
dR (1 MeV, 170')
(x to-" .m'lsr)
Atomic
perce'nt
Element
(counts)
0
Na
SI
365:s: 30
0.2965
63.8:s: 1.6
125 t5
0.5993
10.8,. 1.0
410::1:: 15
0.9905
1.861
21.5:t.1.3
707
Zn
175 6
Cd
64
III':B .. D
1,0
9,.,
4.701
12.09
2.0 O2
1.9O.1
0.039 0.004
J
,J
:?
,,
,I
Backscattering Spectrometry
I
I
1.9 MeV 4 He+
9 LAe = 1700
Q'=50pC
~NEl1,S~K~Cdo.04
2000
Sc
'0"
(J
1000
Zn
HSl,oI
0
j Cd
HZn,o
1000
500
Channel number
FIG. 4.9. The 1.9 MeV 4He backscanerlng spectrum ofa ceramic glass. The indicated stoichiometry was determined from the step
heights. For uncertainties in the subscripts (atomic percentages), sec Table 4.5.
4,3.2.2 A deficiency method
A deficiency method can be used to analyze near-surface
two-element materials even when the steps for both elements are
not ,,;sible on the spectrum, provided that the identity of the
lighter element is knovro. The spectrum for the he-dvier pure
element is compared with that for the two-element material. The
deficiency in the heavy-element signal caused by the presence of
the light element is noted and used to obtain the material
stoichiometry at the sample surface. Application of this method
H~~x-
1)
(4.45)
cos 91
+ EB (KAE o )
cos 92
(4.46)
960 into Eq. (4.3) to calculate (Nt). and then using Eq. (4.7) to
estimate t. Of course, bulk density is assumed for the W0 3 film;
consequently, no uncertainty is quoted for the thickness
estimate.
65
Chapter 4
Q'=5j.C
::
-'"
2000
________
I:
=3.G4~O.17
/"reW
------------L ,
--~
:s
(.)
I
I
1000
\
I
I
'" T
W.lnWOx
I HWOx
t---'43oA
800
900
Channel number
of
--J \
W,o
1000
66
Backscattering Spectrometry
5000 r---~------~
4
4000
MoISt
~ 3000
170 nm
MO'\.
t.,
"""""',----
02000
SI
1000
0
~1
200
1'\
400
i.
600
800 1000
Channel
(a)
r---------------~60
200
. b)
50
-li
40
IJJ
100
50
s~~e
'00
'00
30~
..
o
::~::
20.2
/_
.....
.:'.':... ,.
.~,::~,...:.............. \... .
8.
Depth (hm)
-e
c
150
:,:.. .
10
,
~.,
OL---........:;:..c.....:=:.::Jt.:.~............~h;:.lO
200
250
300
350
400
Channel
(b)
Element
Mo
Si
0
67
Areal density
(x 1015 atoms/Cm2)
percent
406 19
531 .25
161 .12
37.0.0.4
48.4 0.4
14.7.0.5
Atomic
t
(nm)
63
106
1
Chapter4
5000.--------------------.
a) 1.9 MeV 4He Q' =50 JC 9t.Ae = 170
5000 ,----------------------,
b) 3.035 MeV"He Q';:: 50 pO &tAB =170
4000
terms of 0
.sA
~3ooo
'"
o
U 2000
1000
1000
i
_. _ 0 ~=_'="~......,.:'7_'__'_'_':"~~-'--"-_:_J
-- -- .._- .- -_.- 0 . -- 200 400
600 800 1000
SI
~_ ~
--~
200
400
600
Channel
Channel
(a)
(b)
800 1000
FIG. 4.12. (.) The 1.9 MeV 'He Rutherford backscattering spectrum of Sample 46A, a 170-nm MoiSi multilayer film on a Si substrate.
(b) The 3.035 MeV 'He backscattering spectrum of Sample 46A. The enhanced o peak in (b) is due to the strong non-Rutherford
resonance in the 4He-O cross section at the .(He laboratory energy of 3.04 MeV.
68
Backscattering Spectrometry
-onSIC46A)
____ OnC(46B)
"" 10,000
.S:
.l!l
<::
20
--He-O
x-section
8.
c
10 o'!
.!:!
::r
o><
5000
o L~,---,_..,......:~==::::;:::",~~-,-_-,-~--.J 0 ~
3100
3080
3060
3040
3020
3000
E'
2000
1000
Energy (MeV)
.0.5
3000
1.0
1.5
2500
-""..
0
(.)
2000
1500
1000
500
0
0
200
400
600
800
1000
Channel
FIG. 4.14. A RTJ1\..fP simulation (smooth line) of the spectrum shown in Fig. 4.6 (rough line). The experimental parameters given in
EXAMPLE 4.5 and the Fe and Mo areal densities given in tile first row of Table 4.4 were input to the program
69
Chapter 4
4.4.1.1 Sensitivity
The most obvious factor affecting the senSitivity of a given
back.scattering measurement is the scattering cross section, which
clearly varies with the mass, charge, and energy of the incoming
ion. However, the ultimate sensitivity is also influenced by the
level of counts from other sources (for example, pulse pileup or
heavyatom, thick-target scattering) underlying the region of
interest An additional consideration is the practical limit on the
time available to do the experiment. If a different beam provides
a higher cross section but is o'n.ly available at (or limited to) a
much lower beam current, the achievable sensitivity might be
less than that obtainable using ions with a smaller scattering cross
section. Acceptable beam currents might be limited by detector
considerations. Finally, sensitivity is limited by beam effects on
the target such as beam heating and sputtering, which can
destroy fragile targets.
4.4.1.2
'.1
,
~.
Mass resolution
1,
.J,
70
Backscatterlng Spectrometry
smaller. The energy loss factor, [8], depends on the incoming ion
greatly enhanced for 3.05 MeV 4He scattering, where the elastic
cross section is resonant. For scattering with heavier ions,. the
mass resolution. improves, with optimum results obtained for a
22 MeV 12C analysis beam, where the high-energy edges
resulting from scattering from the two i~otopes of eu are dearly
separated. Resolution worsens for higher beam masses, Depth
resolution for high-mass target atoms also improves with
increasing beam mass, with the 12C spectrum distinctly
indicating an increase in Y concentration with depth.
and its energy, the target composition, and thc orientation of the
target. The achievable depth resolution ox for a measurement [Eq.
(4.26)] is also proportional to SE, so Sx dcpends on the detector
resolution, the energy stability of the beam, and energy
straggling in the target, as well as the parameters that determine
[8].
4.4.2 Ion beams and ion energies
Originally, backscattering analyses were performed almost
exclusively using protons or alpha particles with energies in the
Rutherford scattering regime for most target elements. Although
this analytical technique is convenient, it does not provide the mass
resolution, sensitivity, or depth resolution required for some
measurements. Light ion beams of both higher and lower
energies are currently being used in specialized applications that
require high sensitivity and bigh depth resolution. Heavy ion
beams have also been shown to offer enhanced results over
conventional techniques for some applications. The main features
of the various techniques are compared in Table 4.7.
71
Table 4.1, A comparison of some of the main fealures of somo common backacaHering techniques,
Conventional
backscetterlng
Medium energy
bBCk8canartng
backSclittering
backscatterlng
BaamJenargyl
detectors
Cr08s section
Ruthorford 'or higher ma33SS. NonRUlherford lor some IIghlerlarget specles. Exp9rfmental/y determined cross
sections must ~9 used. Negligible
screening.
SenslUvlty
lH
;j
Mass
reaolutlon
be"",.
b",,",,,,,
Depth
praflllng
other
consIderations
<,' .
',.':'!~
Backscattering Spectrometry
V1Ba2Cu3Or.,.IA1203
~
I: 5
:I
~4
~3
:e2
CO
-1
"
~ 0 E - - - - L - -_ _' - _ _ _'-~_L..LY
200
400
600
800
Channel number
FIG. 4.15. Comparison ofbackscattering spectra obtained at various beam energies for !H and 4He analysis ofY IBa2CU~Orx. Taken
from Rauhala et at. (1991).
Y,S""Cu,O,.xlAJ2O,
28 MeV 2BsJ
energy beams.
200
400
600
800
Channel number
73
Chapter 4
.
')
atomslcm 2),
scattered from the same target also at 150 into a swface barrier
detector having a resolution of 15 keY. The natures of the
spectra are different in that the former is a plot of yield versus
flight time whereas the latter shows yield versus energy, but it is
clear that comparable mass resolutions are obtained for the two
cases.
0
,'i
.:;
500
Cu
400
300
;'
~ ..
200
Au
100
Ag
'.
"
200
100
300
400
12C
4.4.3 Cautions
As with any experimental procedure, there are a number of
possible sources of problems both in setting up the experiment
and in analyzing the result". Many of these problems are
discussed in some detail in Chapter 15, but we mention a few
briefly here as well.
74
~.:
Backscattering Spectrometry
Energy (MeV)
1.3
1.4
1.5
1.6
1.7
1.8
10
rl,r----.,--~~I-----,_,----._,----~I
0.8
:E
CIl
>.
0.6 -
"C
1!iii
E
0.4
o
t-
Ag
0.2 -
0.0
700
750
800
11
850
Channel
FIG. 4.18. Simulated energy spectrum forbackscattering analysis using 2 MeV 'H
was th e_~M~.
Ii F 4 17
Au
900
950
1000
1-"
. h .. -I::
-"
.
e ana YSIS IOns Wit a s"-tJ.ace bamer detector. Target
4.4.3.2 Channeling
Experimental parameters such as the beam energy and the
target orientation can lead to spectra that are difficult to
interpret. For crystalline targets, channeling (Chapter 12) can
occur if the beam is incident normal to the target surface,
resulting in surface peaks and reduced subsurface yields in the
backscattering spectrum. Although channeling can also occur for
other angles depending on the material. it is possible to orient the
target in a non-channeling position relative to the beam. Rotation
of the target during bombardment is also used to avoid
channeling.
4.4.4 Examples
OD
75
extraneous
peaks
in
the
Chapter 4
3000
-'"
8000
6000
TI
2000
::>
0
4000
2 keV
1000
2000
o~-L------
______~~~__-A~O
100
200
Energy (keV)
FIG. 4.19. Backscattering specmun obtained using 270 keY 7Lt ions with a TOF analyzer. The superposed spectra correspond to a
titanium foil target and a titanium silicide film. Taken from MendenhaU and Weller (1991 b).
..
76
Backscattering Spectrometry
-~-
Channel #
FIG. 4.20. RES spectra of iridium-implanted silicon for three different implant doses, obtained using 1.8 MeV alpha backscattering and
surface barrier detectors (Yu ef al., 1991) .
. --200
..
~.~
..
Ag
100
Cu
Au
10
12
Energy (MeV)
FIG. 4.21. RBS spectrum of sputter-deposited Ag and Au on an aluminum foil substrate, obtained by scatering 15 MeV 16 0 ions into a
surface barrier detector. Taken from Weller el al. (1990).
obtained by scattering into (a) a surface barrier detector and (b) a
TOF detector. 'Whereas little information about the layered
structure can be obtained from the surface barrier spectrum, the
TOF spectrum clearly resolves the first eight layers of the
structure. Energy dispersion, which worsens progressively with
depth into the target, limits the analyzable depth in this case.
77
.."l
Chapter 4
100
a}RDS
80
~
:::>
60
40
~o~----------
__
~.-
____
20
0
400
300
200
100
Depth (nm)
15
b)TaFRBS
~
:::>
10
ot.....~~_-J..._~-J...~.;..-.L.JJ........
300
200
100
Depth (nm)
FIG. 4.22. RBS spectra obtained from a target consisting of 10
layers, 8 nm thick, of lno.18Gau.82As alternating with 20-nm-thick
layers of GaAs. Only In region of spectra shown for 25 MeV
3:5Cl ions scattered into (a) a surface barrier detector and (b) a
TOF detector. Taken from Dobeli et at. (1991)
78
'
Backscattering Spectrometrt
Energy (MeV)
1,s
1,0
~5
~''''''''~~~'''~t
t.
Research and Industry", McDaniel, F,D" Doyle, B.1" BenItzhak, L, Hamm, RW" Johnson, RM_, Sisterson, J.M"
200
"'"
Channel
soo
.00
(1-2), l-l224,
"Application of Accelerators in Research and Industry:
Twentieth International Conference", McDaniel, F.D.,
0_'
i4
~
Ii'
Energy (MeV)
1,0
t.s
2.S
2.0
--
~.
,
.-
'"
>=
-;\~
'"\
--- - -
II
100
200
300
400
500
Channel
Dobeli, M"
FIG, 4.25. Analysis of the same target used in Fig. 4.24 with a
3.254 MeV alpha beam. Taken from Vizkelethy et ai, (1990),
REFERENCES
HR., Besenbaeher, E, Loftager, Po, and Moller, W,
(1980), Phys, Rev, A21, 1891.
Andersen.
79
Chapter 4
Petrov, I., Braun, M., Fried, T., and S.therolom (1983), J. Appl.
Phys. 54, 1358.
Vizkelethy, G., Revesz, P., Li, J., Matienzo, L.J., Emmi, F., and
Mayer, J.W. (1990), in High Energy and Heavy Ion Beams
in Materials Analysis (Tesmer. 1.R. Maggiore, C.J.,
Nastasi, M., Barbour, 1.C., and Mayer, J.W., eds.), Materials
Research Society, Pittsburgh, PA, p. 235.
Weller, M.R. (1983), Nucl. Jnstrum. Methods 212, 419.
Weller, M.R, Mendenhall, M.H., Haubert, P.C., D6beli, M.,
and, Tombrello, T.A. (1990), in High Energy and Heavy Ion
Beams in Materials Analysis (Tesmer, J.R, Maggiore, C.J.,
Germany.
Yu, K.M., Katz, B., Wu, I.C., and Brown, I.G. (1991), Nuc!.
Instrom. Methods B58, 27.
Ziegler, J.F., Lever, RF., and Hirvonen, J.K. (1976), in Ion
Beam Surface Layer Analysis (Meyer, 0., Linker, G., and
Kappler, F., eds.), Plenum Press, New York, p. 457.
Ziegler, J.F., Biersock, J.P., and Littmark, U. (1985), "ne
80
ELASTIC RECOIL
DETECTION: ERD
G. Dollinger and A. Bergmaier
Universitat der Bundcswehr Munchen, Neubiberg, Germany
CONTENTS
5.1 INTRODUCTION ....................................................................................................83
5.2 FUNDAMENTALS ...................................................................................................84
5.2.1 Kinematics ........................................................................................................84
5.2.2 ERD cross sections ........................................................................................ 84
- -- --
81
-,
- --- ._--
Chapter 5
82
5.1 INTRODUCTION
ERD (elastic recoil detection) or ERDA (elastic recoil
detection analysis) (Ecuyer et al., 1976; Doyle and Peerey, 1979;
Arnold Bik and Habraken. 1993; Assmann et aI., 1996),
sometimes called PRES (forward recoil spectrometry) (Mills et
aI., 1984), uses elastic scattering as the main interaction between
a smft projectile and a target atom for elemental and depth
profile analysis as it does Rutherford backscattering
spectrometry (RES) [Fig. 5. I (a)]. The main difference from RBS
(Chapter 4) is the detection of the recoil atom and not the
scattered projectile. The yield of recoil ions of a certain element
is directly proportional to the elemental content in thc layer that
is investigated. Matrix effects of surrounding atoms have only
minor effects on the scattering cross section because the
scattering cross section is, as in RBS, dominated by the large
interaction potentials between the atoms at short distances
(Coulomb or strong interaction potentials). Thus, ERD has the
potential to perform quantitative elemental analysis. The only
influence of the matrix on the scattering yield arises through
cbanneling effects, thus allOwing for structural analysis (Chapter
12).
I'"
-1
i, EI
"',
i ~
-,
.r=,
.,
' -... '
83
,1
Chapter 5
5.2 FUNDAMENTALS
5.2.1 Kinematics
In an ERD experiment, recoil ions from binary elastic
collisions are analyzed (Fig. 5.1). As a good ~pproximation. the
recoil ions are at rest before scattering, and they are emitted at
angles <p with cp < 90" in the laboratory frame because of twobody kinematics. A projectile of energy Ep transfers the amount
of energy Ei to a recoil atom of type i in the sample during a
binary collision, where
rp~
.M
= arCSlllM.
(5.3)
(5.1)
do,
(d.Q1... -K
Tbe parameter ~ is the kinematic factor ofERD. and M" and Mi
denote the masses of the projectiles and the recoil atom,
respectively. Deviations from Eq. (5.1) occur as a result of the
thermal and zero-point motion of the target atoms before
scattering, which arc treated in terms of a ''Doppler effecf' in
Section 5.5.2. Other deviations can be introduced by some
inelastic processes in the electronic shells. In general, inelastic
nuclear scattering has to be avoided in order to prevent the
ambiguity problems in elemental separation and depth scales that
arise in RBS analysis.
( 'J'(
Z,Z,e
4m:,E,
M, + M;
M,
)'
COS' '1"
(5.4)
where z" and ~ are the atomic numbers of the projectiles and the
scattered target atom of type i, respectively.
84
1+
K = [ 1+ v.
I,Lab
E,
V;,ub)'
2E,
elements.
(5.4.)
(2EpCOS~J
V.LLab
']'
M +M'x48.73xZ Z.Z
~ 213 +z211
V.1,Lab ='
M.
pip
,.
Mp
JZ,Z,e'
ETb= = ( 1+- - - - .
M, 41t8oRo
(S.4b)
(:5.6)
The cross sections can deviate by less than about I % from the
Rutherford values when a safe minimum distance of closest
approach
of
Ro
is
used
the equation
=2.074XIO-Mcm'[~J'(A,+A'l'_3 .(5.5)
(dcr,)
d.Q u'
E,[MeVj
lA,
cos ~
;..
---
2
cr ex: Z, (M" + MJ2 ex: Z:
20
40
prnjecflte alOmi~.numher-
85
Chapter 5
Table 5.1. References for elastic cross-sectioll data for helium ERD on protons. pQ-le,p)He, in the laboratory system.
Angle (laboratory system) (deg)
60.00
Reference
Bogdanovic Radovic and Benka
55.00
2.50-4.50
50.00
2.50-4.50
45.00
2.50-4.50
(2001)
(2001)
41.50
41.50
40.00
2.43-2.97
2.44-2.98
2.50-4.50
31.30
31.30
30.00
1.95-3.05
1.95-3.05
2.50-4.50
30.00
30.00
30.00
30.00
21.50
21.50
20.00
20.00
1.0()-2.50
5,03-5.96
9.5()-11.6
1.0()-12.5
2.0()-2.92
10.00
10-40
2.0()-Z.92
1.0()-2.5
1.51-1.95
1.0()-2.50
0.6()-5.00
.~
retrieved from the SigmaCa1c Web site (bttp:llwwwnds.iaea.orglsigmacalcl#). These values might be the best choice
for evaluating helium ERD for hydrogen analYSis. A compilation
of data is shown in Fig. 5.3 for some angles over a wide range of
energies.
~;
1,
86
I~_=~oor"
-;j~,250 '.
,:~.
" ~'1~99,
~'-"
"' ...
5
6
7
'He energy [ MeV I
I ql': 30
10
.. -.
- - ........
5
6
'He energy [ MeV I
1.=35'
10
"
eo 400
..
'
'200
'"
is
,,
"
600
tl
is
.' .. "'.
~E
800
:5
250
S
6
7
"'He energy [MeV]
....... 9
'0
10
.,
200
.. -.. ..
3
-. ..
_10 _
FIG. 5.3. Elastic scattering cross sections, He-p, as calculated using SigrnaCaic (http://www-nds.iaea.orglsigmacalcl#) (solid line)
compared to the Rutherford cross section (dashed line) at various scattering angles.
Table 5.2. References for elastic cross-section data for helium ERD on deuterium, d(He,d)He, in the laboratory system.
Angle (laboratory system) (deg)
41.50
40.00
35.00
31.30
30.00
30.00
30.00
25.00
21.50
20.00
20.00
10.00
87
Reference
Nagata et al., 1985
Kellock and Baglin, 1993
Chapter 5
"
(5,6.)
Table 5.3. Fit parameters for experimental scattering cross-section data (laboratory frame) for d(He,d)He according to Eq. (5.6a)
(Kellockand Baglin, 1993):
(:)Ub
'P (deg)
a (1IMeV)
20
25
30
35
-0.143
0.451
0,133
0,097
b
6,043
2,849
4.383
4.511
c (MeV)
-0,898
4,738
2,196
2,588
d (MeV')
1,476
-U03
-0,042
-0.278
Table 5.4. References for elastic cross-section data for 1Li ERD on protons p(,U,p) 1Li in the laboratory system.
Angle (laboratory system) (deg)
45.00
30.00
Reference
2.28-5.70
2.28-5.70
Siketic, 2005
Siketic, 2005
- -..:::"-.:.....- -----
88
...
"
. cO ij~Ki-'
o ... ~
:.:....::-:.~ft
~~In:.
-',.,;.~
:,."flibt:n:
"'-"c;'_.".,.;;'",~",~...;.
'0
..,......-..
~,.:.;.-
--".'
.'
~~~7C-""",",,,,,,~.,J;;1 ~o
3,0 '.
'4fteene.!'U'i"(v.eV)
1"'!
1.S .
.:L.O
. ~~G~~Yl
.'
E;IG.S.17.
,'-'
..
FIO.:S,lS..
-_.
o
o ... lcr"
-PatyDa!II.IIni
... ,....
- ~.
~ .'':'~'-''''-'-'~ .:.
....
i.ii.
1.0
..
'a . '," ; .
:ss-
.... F&n-bd
",
-.~-.-~....
>v
...
.,/
Cl:!:Do~._~_
. . N . . . . . . . .
'L'!A~~--~"~~~~2D7"----~..
~----~
..
.~,~.~~--,~.~~~..
"==----"
..~--~..
"He enerwJMeV)
flG. 5.19.
,flG.520.
._.w-
-~ftl
.~--~
'';.0
FIG. 5.4. Experimental elastic scattering data and an empirical fit according to elastic scattering cross sections He-d compared to
Rutherford cross section (dashed line) according to Eq. (5,6.) and Rutherford crosssection data (Kellock and B.glin, 1993),
5.3 METHODOLOGY
The aim of an ERD experiment is to perfonn elemental
analysis by detecting the recoil ions in forward directions. There
are two main tasks to be solved: The recoil ions have to be
identified in terms of their nuclear charge or mass, and their total
energy has to be measured in order to obtain the depth
information.
89
Chapter 5
(5.7)
N
-.!!l!.. " cot'
cp.
from both the ERD and RBS signals. In this case, it is desired to
accept the scattered projectiles in the ERD spectrum. However,
the huge nwnber of projectiles scattered from heavy matrix
atoms should be prevented from reaching the detector in order to
attain a sufficient count rate of small elemental fractions at low
background when sensitive analysis. of light elements is the aim
of an ERD experiment.
In the foHowing sections, we summarize the various
methods used to identify the nuclear charge or mass of the recoil
ions independently from depth analYSis.
with thin entrance and exit foils. Such a setup allows a uniform
adjustment of the effective thickness of the stopper medium by
adjusting the pressure inside the volume in order to obtain
optimum thickness as a range medium.
(5.8)
NERD
NRBS
NERD
--=
(5.9)
Figures 5.5(b) and 5.5(c) show the sample setup for ERD
using 2.8 MeV He ions at a scattering angle of 30 and the
resulting energy spectrum. The sample was a 400-nm
polystyrene layer on top of deuterate<! poly(methyl methacrylate)
(PMMA) tha~ in tum, was on 400 nm of poly(vinyl chloride)
(PVC). The recoil ions were transmitted through a l2_1.J.!m
aluminized Mylar foiL Both the deuterium and the hydrogen
recoils appear in one energy spectrum_ The depth profile
90
Recoils
Hanc;l~
400nm PS
D-(PS-FiMMA)
400 nmpVC
(bl
Energy (MeV)
Depth (rim)
1.2
0.6
0.8
1.0
6000 . - - - , - - - - , - - - - - , - - - - - . - - - - ,
5000
1000
800
600
400200
6~-r~~rr~~rT~~rT"~~~,",&
N-
d)
"Q
5"
"b
c)
4000
x 4
E
~
c
-;;-3
3000
11
3 .:
c
2000
'"-- -I
...........,
,,,
1000
~
<:2
B
C
~-
2 c
~
U
0 ,
I!
, .,J'"
.0
80
100
120
140
":160
180
100
Channel
80
60
...
40
Depth (Vgrnlcnf)
20
FIG. 5.5. (a) Principle of a range-foiVsolid-state detector combination for the analysis of light elements by ERD. The projectiles have
to be much heavier than the element of interest, so that the projectiles scattered from any type of matrix element will be stopped in the
range foil. If thick enough, the range foil transmits only the light recoil elements. Such an arrangement is sometimes sufficient for the
identification of light recoil ions. This method is widely used for analyzing hydrogen in thin films. (b) Layered sample that is analysed
for hydrogen isotopes by utilising 2.8 MeV He-beam and range foil (!2~m aluminised mylar) method (Green, 1994). (c) Energy
spectrum of hydrogen and deuterium recoil ions from the sample of (b) collected at a scattering angle cp = 30 and an angle of
incidence a = 15 Q The detector covered a solid angle of detection Ml = 3.6 msr. The H and D recoil signals are almost completely
separated with their approximate sUlface peak positions as labelled (Green, 1994). (d) Concentration depth profIles for H and 0
determined from the energy spectrum of (c). The concentration and depth scales are shown in densiry-exclusive units along the left and
bottom axes, respectively. The right and top axes were determined by assuming the density of the sample is 1.04 g1cnr (Barbour,
1995).
The conditions for ERD analysis were a 4 MeV He beam, a 20"un-thick A1 range foil in front ofa silicon detector, an incidence
angle of Il ~ 20, and a scattering angle of ~ ~ 40 [Fig. 5 .6(all
91
Chapter 5
a)"" r-~.,-,-~~~_~~~~~~
b)
liLllJ._e--1 X1 ,
.,_'~~~~~~~~~~lL~.=_.~~~_~O
;Z
-:iQa
.' ~
, om',)
.....
......
.f:ne:I'UY ~V)
Energy(....eV}
1000
r----'i-----i4- - - - . . : ; ._ _--.,
200
oL---,----,r----r~~~~~
a
100
400
500
92
s. 0:: Z2M In Ei
E[ .
(5.1 0)
Heavier recoil ions can have recoil energies such that the
energy loss is less than the energy-loss maximum, in which case
the scaling of Eq. (5.10) no longer holds true. Elemental
separation from the energy-loss signal is hardly possible at low
ion energies (Fig. 5.9) when the energy-loss differences for the
various elements are smaller than the 6E resolution. The
important exceptions are hydrogen isotopes,. which can be well
separated far below the energy-loss maximum.
5.3.211E-E analysis
The full potential of ERD analysis is obtained when the
elemental analysis is performed independently from the depth
analysis. In addition to the energy of the recoil ions, one needs a
second independent signal to separate the elemental information
from the depth information. A quite simple solution is to use a
.6E-E serup. where an energy-loss signal is gained from the
recoil ions using a thin transmission detector in coincidence with
the signal of a second detector that is placed directly behind it in
order to measure the residual energy (Stoquert et al., 1989;
Gebauer et aI., 1990; Assmann et al., 1994; Bergmaier el al.,
1998). Adding the energy-loss and residual-energy signals gives
the total energy of the detected ions. If the .6.E signal is plotted as
a function of the total energy, the different elements are
separated from each other because of their different energy-loss
signals for a given total energy. An example of a dE-E
measurement using a detector where the 8E part consists of a gas
ionization chamber and the residual energy, Erell, is measured by
a silicon detector is shown in Fig. 5.8. The analyzed sample was
an approximately 200-nm-tbick Al.Gal_. .N layer deposited on
AhO). It was analyzed using a t70 MeV 1271 beam, a scattering
angle of q> = 40.2, and an incident angle of a = 15, The energy
loss, tt.E, is plotted versus the total energy, Bill" which is the sum
93
"-
""
~]
.,
Chapter 5
~:
. 10000
'1000
1 oil
, 1
'cij
a
o
,f
c
.Q
0,1
aluminum
O,Ol,~~~,,",,~~=~~~=~~~~
0,1
10
100
1000
0,0~,L"1~~"4~~~10;-'---U.~100!;;--'~-';1~OOO
ion energy [MeV I
.!
,5
"
94
. "
.~.~-
.'
.80
'"''
"':>'"
:4'. ..
20
:'
lpri ene
"
,'~-~o
. -- c
~~a
t-ii""""""
Etot [MeV]
'.:.~~ .
............ ,........ .
......
"~"
Mylar (H.C",O,)
1,1
- 10
..........
100
10
Ior1
.....
allergy
.......
1000
[MeVI
--. ~
.' ,..,,~.
0,
... ......
'
". .
....,
1'0.0
....
CF.(gas)
'Ii ';"-'...~
He ..
. (p=tol3l:10'Pa..T"'293K}
";,
ogo)J,,.~---"-~=~~~~~lka-c....-'"--'''''jJ'OO"--~+-'-~U;;,aao
. ion enerii/ I MeV j
95
Chapter 5
energy loss
"
AE
ions
(5-10 Jll11)
silicon
stop detector
40mm
top view
~~~~sl:TJ
dx=O-7rnm
~gasout
cathode
./1
impinging - - ' /
~"s
. e("ltfan~ foils
position sensitive
Si PIN diode
I~
.... ,~ri.~ ~~d. ,
.AE.1.
ai"iode
gas In
FIG. 5.10. Various t.E-E detector setups for elemental analysis in ERD analysis. The ions enter from the left, deposit a fraction of
their energy in the t section, and deposit the residual cnergy in the second section of the detector. (a) Schematic for an solid-state
detectors for .1.E and residual energy. (b) Electrode configuration of an ~E-E ionization chamber (Assmann et at., 1994). The Frish
grid makes the anode signal independent of the position of the ion track. The cathode is structured in a backgammon geometry in order
to obtain additional position information. (c) Hybrid gas (fOf .1.E) and solid-state (fOf residual energy) ERD detectof. The hybrid
detector has larger dynamics in energy and type of particle analysis. The anode is divided into two stripes to obtain twil:e as much AE
information for redundant elemcntal information.
96
~
MCP2
'1Op 12
... ,
500
"0
4dO
.I:
U
;:
300
.2'
Li:
0
III
200
i=
100
Vi
(5.11)
100
12 MeV
200
300
400
:;00
Recoil Energy I ch
- FIG. 5. \ \. Principle of TOF-E methods: (a). Start-stop
configuration for TOF measurement that allows bigh TOF
resolution. Secondary electrons that are generated in thin foils
(usually carbon) are used to start and stop signals at the
beginning and end of a flight path L. The energy of the ions is
measured in an independent energy detector. (b) Example of a
TOF-E ERD analysis, showing the potential of separating the
masses of all light and heavy elements in the sample to obtain
background-free elemental spectra. Depth profiles can be
generated from energy spectra that are separately projected for
each resolved mass.
Although the principle seems to be relati vel y simple, the
TOF-E method requires sophisticated detectors to achieve its
full potential. The first problem to be solved is to measure the
97
---
----------- ----Chaptet 5
when the pulse width is small enough and fast stop detectors are
used or when long flight paths are utilized.
SEil
&., =
M,
(0,)'
+2.(OIOF,)'
E,
TOF,
5.12
98
'Jrid 1kV
,,/<>:.~;bon f~iiOV
,.' ..... :
ion~s~~~~~~_;'f'~~'~~~'
grid "2kV.. " ", 0
9ri~ {~V'
.sii:fe ~I~'~'
. . .. : '
'
' . '
HatH:um cyclotrori
____~____~
.'
time detector
~~\
""
\1
anode
4-5 kV
2-3 MCI'
detectors
god
2kV
carbon foil
OkV
FIG. 5.12. Various configwations to deflect secondary electrons out of the ion path at constant electron flight paths to the electron
detector for fast timing: (a) 180 0 magnetic field deflector, (b) electrostatic mirror configuration consisting of electrostatic grids, (c)
tilted configuration with 00 electron detection.
.1TOF
TOF
(5.13)
99
Chapter 5
oE; = 2 o(TOP)
E;
TOF;
(5.14)
M:v
.j2M;E;
r = - ' - ' = -'----'--'.'
q;B
q;B'
(5.15)
100
1m
. focal plane
ion
ho,~nv'
High Hesolution
8RD-Spe-cLrQmeter
1m
FIG. 5.14. (a) Q3D magnetic spectrograph used for particle identification and high resolution depth profile analysis (Dollinger el aI.,
1992). (b) Electrostatic spectrometer for high-resolution ERD analysis (Enders et aI., 1992).
101
Chapter 5
or
Elastic
Recoil
Coincident
(scatiered
10115\
M,
(b)
(5.16)
IM
deiectcH 2
(reeDlled afGl'I'lsl
. ;:\y::
ion beam
. ,-,~~-, .~~---
I\{I
del~Clof
(scaitered Ions)
'
Ep
I'
+M/
..
".
(5.17)
".
.. !: :. -t .
102
. ~ f:'I"
.-
'.:
,...
",
!,
..
:... ,
' .. ,: .~.~'~.
,:
(5.19 )
5.6.3).
If all relevant elements contributing to the layer or foil are
N. do
N do
Y =N - ' E.-' Ml=N n . - E . -i L'lQ.
I
PAidD,
piAlclQ
measured
area A on the target (with the areal density NlA). Thus, Il; ~ N/N
is the atomic fraction of element i, da/dn is the differential
recoil scattering cross section at scattering angle <p, and .Ml is
the solid angle of the detector. .Ej is the efficiency for the
detection of element i in the recoil-ion detector including dead
(5.20)
(5.1S)
time effects.
k=l
(Ln
o j=1
sma.
()
The first term calculates the recoil energy directly after the
scattering event by using the projectile energy reduced by the
energy loss of the projectile in the sample up to thickne.o;s x and
the kinematic factor k j of Eq. (5.1). The second term de:scribes
the energy los5 of the recoil ion on its way back to the surface,
j:1
smi3
fs~ex')dx'.
(5.21)
and the third term is the contribution of the dead layer or range
in front of the detector. The stopping force,
foil
S~(x) = dE~ Idx , indicates the energy loss per path length dx
in the passive layer (denoted as layer 0). The calculation uses
Bragg"s rule that the stopping force in a layer at depth x' is the
103
~,
Chapter 5
A,(x) 8(E:"')
n;(x)
YJEf"'(x)]
N N dO"; "'Oe.
A 'dO
'
..
..
(5.24)
AA
n;(x) =
= Y,[Ef"'(x)] PWA.~
.6.e.
N --"'
dcr;
dO
(5.22)
The yield Y, [Ef'" (x)] issummcd over an energy bin lie; of the
ion energy, E~et J depends on Dj(X) and vice versa. The solution
N =
"'e;
p(x)A,
8(E"") ,
A,(x)--'-
(5.23)
Ox
,.
104
Nam<: of program
GrSA
TyPe of analysis
Analytical simulation
RUMP
RBX
NDF (Data Furnace)
Analytical simulation
Analytical simulation
Analytical simulation
Analytical simulation
SIMNRA
MCERD
Analytical simulation
Monte Carlo simulation
CORTEO
KONZERD
SRIM
DEPTH
..
I , D;"'I
Reference(s)
Raubala, 1984; Saarilahti and Rauhala,
1992
Dootittie, 1985,1986
K6tai, 1994, 1997
Szilagyi et al., 1995; Szilagyi, 2000
Barradas'et al., 1997, 1998; Jeynes et
01,,2003
Mayer, 1999,2002
Sajavaara et al., 2000; Arstila et al.,
2001
Schiettekatte, 2008
Bergmaier et at., 1995
http://www.srim.org/;Zieglerelal.
1985
105
Chapter 5
Ox=
BE
Oct
dE, Idx
(5.25)
cr,
values, the energy shift dE~ Idx for a certain layer thickness dx
10
20
30
40
Ox
,_~
:~::,,::,,,::,::,:c"!:'::::::~'
(dE./dx)
E.
+ (dE;ldx)
sin a
E;
sinJ3
(5.26)
10"
106
tiro
J= 10.2
EiC
Q)
E
Q)
:a;
____ _
_ _ _ _-I
Al
"'~':-
Ga
to
c
........
\
\
10'3 \ \
.0
I
I
I
I
J
\'-~-~"f--""'''-:
1o '--,-'-...,.-.o-,---_;>--'-----:--;--~~
a
10 '.
20
30
.40
~!
1;
-';
';1
.:,;,1
1
The spread of the relative energy loss is the sum of several
contributions. Assuming statistically independent Gaussian-type
contributions to a first approximation, the total energy spread
SE/Ei for recoil ions of type i is the quadratic sum of the 1\vhm
of each contribution
A
:~i
:~
iJ
.,,
.~
:~
I
(5.27)
which can be cut from any intense ion beam using a lrighresolution analysis magnet and its slit system.
semp and do not change with analysis depth. They consist of the
energy spread introduced through the phase-space volume being
occupied by the beam, the intrinsic detector resolution, and a
Doppler-effect contribution, respectively. In addition, there are
The
(oE.,IE,~,
codes, where detailed energy distributions '2.5 well as nonGaussian contributions are treated in detail (Table 5.5). Surface
and interface roughness or other lateral structures in the sample
affect the energy profiles and thmi increase the measured energy
spread. The influence of roughness on the depth profiles has
been included in some of the newer evaluation programs. The
problem is that different structures lead to similar energy profiles
that cannot easily be separated from energy spread effects
according to Eq. (5.27). Careful treatment of all contributions
must be implemented to ensure correct interpretation of the data
and to obtain additional information on roughness.
(~-~)E,a,COth(E"",).
( BE,)
E. dop =J8ii2
cas m
E. EP
kT
(5.28)
't'
(5.29)
107
'.',
..
... - ...---------~-'-'---'-------
Chapler 5
OEi)
( E.
= _.J8_ln_2
dOP
COscp
( Ei
BE(X)]
[
...,
E,
(5.30)
(5.31)
108
,l
~.
.,l,
:~
1
I
"{
{
~
!'
recpiUons :
(pasztielai.,1991)
41toE p a
Small-angle scattering has - large cross sections. These-multiple scattering events lead to a continuous angular
distribution close to a Gaussian angular distribution and reduce
the depth resolution already below the flfSt atomic layers of the
analyzed sample. larger scattering angles occur a second time
after the ERD scattering event with only a small probability.
Such plural larger-angle scattering leads to tails in the energy
and thus in the depth profiles. Here, we discuss mainly multiple
scattering and its influence on depth resolution. Plural scattering
limits sensitivity and is discussed in Section 5.6.2.3. Analytical
simulation codes already include multiple scattering effects, as
well as plural scattering to some extent, whereas the full effects
can be calculated accurately only by Monte Carlo-type codes. A
compilation of available simulation codes is listed in Table 5.5.
{"a'
[_1_ xp(x)A,
sinet A,(x)
In Eq. (5.32), Z, ~
In;Z;
]}D
(5.32)
xp(x)A,
sin ~ A, (x)
' _ 41toE;a
]}D
O.885a o
a
(
Z,
2/3
10
x 10- m.
109
L
~ C 2Z Z,e'
d<p
,.
Chapter 5
(E~E)
tio,p ""-
2Z,Z,,' [ , I
Ep tan'P C 4m:oE pa
1tEl
<P(X)A.]D
sino. Al(X)
(533)
(534)
-'tcltalertlltOY~
1.'
1.' ~
~'1O"'
mulliple~[prcftttiIe)
dE.
= 0.56
( 'E)
E JIIIb.i E,
dX' _X_cot a
Ei sin~
(5.36)
sin~ Al(x)
0.'
110
0'OO~~~~~~~~'~5~~,,~,;~~~~;:~:::
15,
,. g.
,. i
1.0 ~
0.'
scale with the square root of the thickness x, the path length
scales with
0.
uruts,
1995).
5.5.5 Detector resolution
In most cases, the detector resolution is the limiting
contribution to depth resolution close to the surface, where
energy-loss straggling and small-angle scattering effects are
small (see Fig. 5.18). Energy-dispersive detectors such as silicon
detectors and ionization chambers are limited in energy
resolution because of electronic noise and the statistics of charge
creation. Although silicon detectors can reach an energy
resolution of close to 10 keY for 1 MeV protons, the energy
resolution decreases for heavier ions as a result of energy-loss
straggling in the dead layer of the detectors; inhomogeneities of
the dead layer; and as sometimes postulated, by plasma density
effects. The relative energy resolution, SEIEt, has been shown to
002-layers of graphite
(a)
o
(b)
0.5
hlllE [%]
FIG. 5.19. (a) Schematic drawing of graphite analyzed b) highresolution ERD. (b) Experimental data for the ERD analysis of
graphite by 60 MeV 1271 ions. Several single atomic layers an be
resolved, as shown by the deconvolution model (dasht:d lines)
(Dollinger e/ aI., 1998b).
III
'.p
Chapter 5
*"
112
".
:...J-
'/
"
37
,~ '37;5
'g>
"
'l'
:':C
"
.~
.. 5. .. ~
In
. _.;
. 39
'9.51"-
39.5
.0
d)
soo
200
u
m
150
400
-c
>=
:- 300
>-
100
200
AI
50
.AD
100
o :-O'"'-~f.'0"""~-:2'='0~'"'":3~O"-'-"-..
f:;-.J...-:50::'""oL,:."o""...J
0a
10
20
30
40
JAl
50
60
-FIG. 5.20. (a) Plot of scattering angle versus recoil-ion energy measured from a 20-nm-thick ~~bon foil cov.ered by~a thiii'~of aAl 0
.,.--
~.
,-
--
2 3- -
film. The energy spread from energy loss is negligible, The angular information is obtained from a position.sensitive silicon detector,
The kinematic energy shift is visible for all elements, resulting in broad peaks in energy projection. (b) Softwarecorrected spectrum
with a constant energy-angle relationship and enhanced energy resolution.
113
..
Chapter 5
5.6.1 Accuracy
The relative statistical error, 6.n/nj, of the elemental fraction of
element i present in a thin layer is obtained from the
uncertainties, t.. YIc:o of all elements k present in the layer as can
be deduoed from Eq. (5.19):
An.
I
- ' =n, Y,
. ~ata
(a)
data
(b)
,.,
..
0'
]'
AY.
----L[dcr.
E
,=. dQ
l222ev
II
dQ
'm
(5.39)
c
c
8
dcr. ;)
(I-n,. " )AY; +(n,-,
MAP reconstruction
with confidence interval
depth [10"allcm' J
FIG. 5.21. J3e profiles of 222 eV Be implanted in pure
amorphous 12C analyzed by40 MeV Au at '" ~ 15 using a Q3D
magnetic spectrograph (Neumaier et aI., 2001). The spectra are
deconvoluted using the deptb-dependent resolution function and
sbow 10 error bands: (a) Calculated using the DEPTH code
(Szilagyi et aI., 1995). (b) Obtained using an experimental
resolution fimction as extracted from analysing a lJefl2e
multilayer. This resolution function assumes worse energy
resolution than assumed in (a), resulting in sharper profiles than
in (a). Both profiles show a type of bimodal structure in the
range profiles.
114
.0
2004].
5.6.2 Sensitivity: Trace element analysis
The elemental sensitivity is influenced by several factors.
1.0
0.8
-~
~
0.6
.a'"
~"
Q,4
<l
:;;!
02
1.0
115
".+
Chapter 5
,IN; and 2N
B.
on
Z2
Fa, which
116
P also help to
r2$ldual energy
! a.u.]
117
,...
Chapter 5
O'n
T= _ E"", -21n T~
= f
n
y
0"" COS
8n:(dEIdx)2~'.E.",
E"",
T~
E"",
1&t01
1
1~.oo
'1e-41
,-
1~
within the matrix at the Coulomb barrier energy Elhll given in the
,.....
1&-OS
10-06
10
;x:
= T"",,' Ec,..
1t
(dcr/ctn)Ml
Ml
E"",
(5.41)
2E....
lIS
(5.42)
119
. '3Z
Chapter 5
Table 5.6. Hydrogen release cross sections, OJ'l:io derived from different experimental data compared with the calculated damage cross
sections for di'iplacements, O"di5J1' and ionizations, Gioo. according to Eq. (5.42). The threshold energy was set to Edisp ~ 10 eV for
displacements and EiOll ~ 5 eV for ionizations.
arel (
Experiment (reference)
experiment)
(mb)
O"disp
5 x 10 10
3 X IOu
7):' 1013
2 x 10 14
1011
2)(
1011
2 x to'
I x 107
1 )( 107
3)( 108
3)( 108
I x 10'
7 x 10'
3 X 1013
2 X 1010
3 X 1014
4 x 10 10
4)( lOll
2 x 10"
5 x lOll
1 x 10 10
3 x 10'
6 x 10'
6 X 10 12
5)( 1011
4)( 10 10
6 x 10'
8 X 107
2)( 10'
5)( lO"
5 X 10"
2
1011
5 x lOll
7 x 10 12
j x
2 x 1011
2 X 10 13
3 x 10 10
3 l( 10 10
5 X 1011
5 x 1011
3 x 1011
2 X lO"
1 x 10 13
l( 1013.
1 x 1011
I x 10 14
4 x 10 10
4 x 10 10
3 x 10 9
2 x 1011
1013
7 x 1013
6 x 10 12
3 X 10 13
120
~&
.......
.
1 . :T
.Doi!--''-t1O.--'-~2l;;D~-l30.--'--40!n'-!50;;-'-'/'OO.-'-~7"'O--'-'.!80.--'-~80''..J'06
'fluence [1010 ions cm~2]
~.01
0.1
1~
100
FIG. 5.25. Plot of the deuterium content of a-deute rated amorphous carbon (a-C:D) layer as a function of heavy-ion beam fluence using
80 MeV Au and 210 MeV I beams (Behrisch etal., 2000): (a) linear scale, (b) logarithmic scale offluenee.
Baumann, H., Rupp, Th_, Bcthge, K., Koidl, P_, and Wild, Ch_
(1987), Eur_ Mater. Res. Soc. Can! Proc. 17,343_
Behrisch, R., Mayer, M., Jacob, W., Assmann, W., Dollinger,
G_, Bergmaier, A., Kreissig, U., Friedrich, M_, Sun, G.Y.,
Hildebrandt, D., Akbi, M., Sclmeider, W., SchleuBner, D.,
REFERENCES
42.
Bergmaier, A, Dollinger, G_, and Frey, CM. (1995), Nuel_
Instrum, Methods B99, 488,
Bergmaier, A, Dollinger, G" Faestenruum, T" Frey, CM"
(1992), Nuel. Instrum. Methods B64, 296,
Arnold Bi!<, W,M" deLaat, CTAM" and Habraken, F,H,P,M, " ____ Ferguson, M" Guttier, H, (1996), Diamond Relat. Mater_ 5,(1992), Nuel. Instrum, Methods B64, 832,
995
.
.
.
. .". _ .
:
Arnold Bi!<, W,M" and Habraken,F,H,P,M_ (1993), Rep. Prog. - Bergmaier,-A_, Dollmgei, G" and'tiey, C,M. (1998),- Nuel. .
Instrum, Methods BI36-138, 638,
Plrys, 56, 859,
Arstila, K" Sajavaara, T" and Keinonen, j, (2001), Nucl.
Bogdanovic Radovic, I., and Benka, O. (2001), Nucl. Instrum,
Instrum, Methods B174, 163.
Methods B174, 25_
Assmann, W., Hartlmg, P., Huber, H., Staat, P., Steffens, H. t and
Bohne, W., Rohrich, J., and Roschert, G. (1998), Nuc1. Instrum.
Steinhausen, Ch, (1994), Nuel. Instrum_ Methods B89, 13 L
Methods B139, 219.
Assmann, W., and Maier, HJ. Nucl. Instrum. Methods A362
Bohr, N. (1948), Mat. Fys: Medd. K Dan. Vidensk. Selsk. 8, 18.
(1995) 143,
Bergesen, p" Behrisch, R" and Scherzer, RM,U, (1982), Appl,
Assmann, W., Davies, lA., Dollinger, G., Forster, J.S., Huber,
Phys. A 27, 183.
H" Reichelt, Th" and Siegele, R. (1996), Nuel, Instrum,
Browning, J,F" Langley, R.A., Doyle, B,L., Bank", J,C" and
Methods B1l8, 242_
Wampler, W,R. (2004), Nuel, instrum, Methods B219--220,
Barbour, J.e., and Doyle, B.L (1995), in Handbook of Modern
317.
Busch, F., Pfeffer, W., Kohlmeyer, B., Scbiill, D., and Pihlhofer,
F_ (1980), Nuc/, Instrum, Methods 171, 7L
121
-e-
Chapter 5
Kim, C.S., Kim, S.K., and Choi, H.D. (1999), N"el. Instrum.
Methods B155, 229.
Kimura, K., and Nakajima, K. (2001), Appl. Phys. Leu. 78,
1679.
Kinchin, H.G., and Pease, RS. (1955), Rep. Prog. Phys. 18, 1.
Koschar, P., Kroneberger, K, Clouvas, A., Burkhard, M.,
Meckbach, W., Heil, 0., Kemm1er. 1., Rothard, H.
Groeneveld, K.O., Schramm, R., and Betz, H.-D. (1989),
Phys. Rev. A40, 3632.
K6tai, E. (1994), Nuel. Instrum. Methods B85, 588.
Kotai, E. (1997),AlP Con! Proe. 392, 631.
Kottler, C., DobeIi, M., Glaus, F., and Suter, M. (2006), Nuel.
Instrom. Methods B248, 155.
Kruse, 0., and Carstanjen, H.D. (1994), Nuc!. Instrum. Methods
t
B89,191.
Kruse, O. (1995), PhD thesis,'Universitllt Stuttgart Stuttgart,
Germany.
Liechtenstein, V.Kh., Ivkova, T.M., Olshanski, .0.,
Feigenbaum, L, Di Nardo, R., and DobeIi, M. (1997), Nuc!.
Instrum. Methods A397, 140.
L6filer, M., Scheerer. H.-I., and Vonach, H. (1973), Nuc!.
Instrum. Methods Ill, 1.
Maree, C.RM., Kleinpenning, A.. Vredenberg, AM., and
Habraken, F.H.P.M. (1996a), Nuel. Instrom. Methods B118,
301.
Maree, C.H.M., Vredenberg, A.M., and Habraken, F.H.P.M.
(1996b), Mater. Chem. Phys. 46, 198.
Mayer, M. (1999), AlP Con! Proe. 475, 54.
Mayer, M. (2002), Nuel. Instrum. Methods B194, 177.
Mills, PJ., Green, P .E., Pa1mstrom, C.l, Mayer, lW., and
Kramer, EJ. (1984), Appl. Phys. Lett. 45,958.
Nagata, S., Yamaguchi, S., Fujino, Y., Hori, Y., Sugiyama, N.,
and Kamada, K. (1985), Nuel. [nstrum. Methods B6, 533.
Neumaier, P., Dollinger, G., Bergmaier, A., Genchev, I.,
Goergens. L., Fischer, R., Ronning, c., and Hofs!ss, H.
(2001), Nuc!. Instrum. Methods B183, 48.
Nolte. H., Assmann, W., Huber, H., Karamian, S.A., and
Mie,kes, H.D. (1998), Nuel. [nstrum. Methods B13&-138,
587.
Faszti, F., Szilagyi, E., and K6tai, E. (1991), Nuc!. Instrom.
Methods B54, 507.
Pus., P., Raubala, E., Alanlm, T., and Raisnen, J. (2003), J.
Appl. Phys. 93, 6370.
Pusa, P., Rauh.la, E., Gurbich, A., and Nurmela, A. (2004),
Nuc!. Instrom. Methods B222, 686.
QuilicI, V., Abe~ F., and Schott, M. (1993), Nuel. Instrom.
Methods B83, 47.
Raubala, E. (1984),J. Appl. Phys. 56,3324.
Reichart; P., Dollinger, G., Datzmann, G., Hauptner, A., Korner,
H.-J., and Scbmebner, O. (2002), Nuc!. Instrum. Methods
B197, 134.
122
123
-_._-_._------
:,-"
CHAPTER
CONTENTS
6.1 GENERAL PRINCIPLES AND EQUIPMENT ..............................................................127
6.2 GEOMETRY AND METHODS FOR Fll..TERING UNWANTED PARTICLES ........128
6.2.1 Absorber foil technique ................................................................129 ~_. ___ .._
6.2.2
6.2.3
6.2.4
6.2.5
6.5.1
125
Chapter 6
126
"
127
Chapter 6
,~,.~ __-=;;-/F,\
3.
6.2
GEOMETRY
AND
METHODS
FOR
FILTERING UNWANTED PARTICLES
The geometry in NRA can be either backward or forward,
'"
S"",(E) <lx,
(6.1)
If(cos<<;.)
E;,(x)=E o -
S.,(E)dx,
(6.2)
128
E,,,(x) = Eoo,{x)-
(6.3 )
where X~bs and Sabs(E) are the thickness and the stopping power
of the absorber foil. respectively. In Appendix 10, the energies
of protons and alpha particles after having passed through
frequently used thicknesses of Mylar are tabulated for deuteroninduced reactions from 2H to 19F. Readers can use SRIM
129
Chapter 6
the depth resolution. The depth scale can be calculated frOID Eqs.
(6.1), (6.2), and (6.3) hy calculating the detected energy of the
reaction product for different depth values. An alternate
tcchnique in obtaining the depth scale is to use an expression
similar to that used in ion hackscattering [see Eq. (4.22) in
Chopter 4]
A = Q,Ocr(Eo)Nt ,
cos am
.
(6.5)
where N is the atomic density and []m is the nuclear reaction
stopping cross-section factor, defined as
(6.4)
(6.6)
The reaction factor n, is taken from Eq. (6.8) (also see Table
6.1) and weights the energy loss of the incident particle in the
same fashion a.s the kinematic factor weights the backscattering
stopping cross-section factor (E]in in Eq. (4.23) of Chapter 4.
Values of ein and ~I can be calculated assuming either a
surface-energy or mean-energy approximation. When either the
standard or unknown contains more than one element, the Bragg
rule must he applied toEq. (6.6). (See EXAMPLE 6.2.)
Table 6.1. Values of fitting parameters (IE and i3e: in the expression
E3 = ae: El +i3E forB = 165".
FItting
Q
Reaction
(MeV)
14N(d,o:)12C
13.579
'''c(d.p) 14C
5.947
range
~.
(MeV)
(MeV)
0.122:t 0.013
9.615:t 0.016
0.8 - 1.4
0.005
5.339 :t 0.006
0.6 - 2.0
0.571
:t
1SN(d,o.)13C
7.683
0.273 0.010
5.487:t 0.014
0.6 - 2.0
9Se(d,p)10Be
4.585
0,456
j:
0.006
3.923 :t O.OOS
0.6 2.0
'2c(d,p)''c
2.719
0.591
::!:
0.003
2.384 0.004
0.6 - 2.0
''O(d.~!'70
1.919
1.722 0.003
0.6 - 2.0
130
.~
th!S
(6.8)
The variables M" M2, M), M.., Eh El , and e in Eq. (6.8) are
defmed in Fig. A4.1 in Appendix 4. The second form ofE3 given
in Eq. (6.8) is a linear approximation of the ftrst form. The'
variables aE and ~E are specific to the reaction under study and
depend aD the detection angle 0. Values for aE. ~E. and Q are
2iven in Table 6.1 for some .specific reactions.
In this example, Mit M z, M 3 , and !>v4 are the masses of the
deuteron, 160, proton, and ]70, respectively. The energy E1 is
defined as being equal to Eo, and E) is the final proton energy,
which, in this example, is 2.352 and 1.556 MeV for Po and Ph
respectively. The energies measured after the protons have
passed through the Mylar foil can be calculated from the data in
Appendix 10. Although those energies were calculated at e =
150, the energy of the protons in our case differs only slightly
(energies in Appendix 10 are 2.36 and 1.58 MeV), so we can
assume that the protons will lose the same amount of energy.
Data from Appendix 10 show that pc loses 0.24 MeV in the
absorber foil and PI loses 0.35 MeV. Therefore, the energies of
the protons after they have passed through the absorber foil are
2.312 III1d 1.206 MeV.
Let us estimate the max.imum thickness of Si0 2 that can be
measured at this energy. The stopping power of deuterons in
Si0 2 changes slowly between 800 and 900 keY, so we can use
an average, which is 7.6 eV/A. The energy range that we can use
is approximately 30 keY, so the maximum thickness is
x ... =
S(E)
keV
7.6 ,VIA
~ 4000 A
(M,M,E, )'12
'"
cos
M,+M,
and
~= 30
M,Q+E,(M, -M,)
M,+M,
131
-ct
Chapter 6
Eq. (6.13)]
Nt
:xo"
=N
"
----. "O(d,pol"'o
"O(d.P)"O.6-15O
'"o{doP,J"O
~=6.69xlOll~::4.54xlont
31875
t,..,o, YTa,O,
- -- - -- -
17
~OO~~rr~~~"~~'T~~rr
,~~
t'o(d.P11170
2000
J!l
1500
5o
u 1000
(evem')
I-
.11
"o(d.PoInO
Catom
lISf(d.p)ZlSI
and
100
500
200
300
~)Ue
400
500
600
Ch8Elnels
(evem')
Hatom
132
..
Eo - E~scmanc~
S(E)/ cos u;. '
(6.9)
(6.10)
whcrc
E;
j=i
(6.11)
(6.12)
133
."
Chapte'6
- Y(E)]' .
"
'"g 10
'ii
"E
,
g
I-
These requirements are the same as those required by highprecision RES, so they are not discussed further here (see
Chapters 4 and 16). The frequently used standards are LiF, CaF,
(Dieumegard elal., 1980), NbN (Maurel and Amsel, 1983), TaN
(Amsel and David, 1969), and Ta,O, (Pringle,1972; Phillips and
Pringle, 1976; Amsel el 01., 1978" 1978b). The measurements
of the standard and the unknown sample should be performed
under the same conditions. namely, the same solid angle, same
incident energy, same absorber, and also the same uncertainty of
the current meastliement. In this case Cif the above require!l1ents
are ful.filled), the total amount of the measured nuclei in the
unknown sample is given by
..
,
1.0
'
t
,
'tj
"
C)
O.S ~
A Q~f
Nt = Nt ----,~'A
Q
'-
u...~u....~1..L.~'..L.'~..L.~.L..~.L....~~~...
0.0
620
625
630
635
640
64S
Proton energy [keY]
650
"'
655
(6.13)
where A and Am are the peak areas in the spectra; Qc and Q~f are
the collected charge; and Nt and Ntm are the total amounts of the
measured nuclei for the unknown sample and the reference
standard, respectively. Equation' (6.13) assumes that the
thicknesses of both the reference sample and the unknown
sample are small enough to use an energyindependent cross
section and that their respective peaks can be fully integrated. If
134
the same amount of charge is collected in both cases, the Q~f IQ<;
(d.p,)
UN
"c
Nuclear reaction
(d,Po)
0.032
(d.p,+P2)
0.078
(d,P4)
O.OSS
(d,ps)
0.45
(d,Pe)
0.057
(d.",,)
O.OOSO
(d,.a,)
0.119
(d,pp)
1.91
cr A (Eo)QuQul>u x
[l~ casCl"
(6.15)
(6.14)
[I: l::" and [I: l:c are the nuclear reaction stopping crosssection factors for element A in the unknown and the known,
respectively [as defined in Eq. (6.6)]; and the variables n, Q,
where Ntc and Nto are the absolute contents and Yc and Yo are the
counts of carbon and oxygen, respectively.
H U [el
AB
x=~~Rz
Ac
HKA,
(6.16)
[el A.,
135
-1
Chapter 6
wbere R is given by
are
[]1:
converge.
cr A (E,)OuQu CN;:"t)
(6.18)
COSU io
fLi,
I
I
I
I
I
I
I
II
I
4.15 for d., 3He, and a., respectively.- These slopes can provide a
good guideline to the appropriate energy in a given situation.
(6.19)
where
(6.20)
6.5.1.2 'Li(p,a)'He, Q = 17.347 MeV
This reaction is excellent for 'Li detection. The cross section
was measured at (3 = 90 between 1.4 MeV and 2.0 MeV and at
0= 1200 between 0.4 MeV and 2.5 MeV (Sweeney and Marion,
1969). The cross section increases very smoothly up to 3.26 mb/sT.
Detailed cross-section measurements were earned out by
136
"'
Cassagnou el al. (1962) for 0 ~ 85', 109', 119", 128', 138', 147",
15Y, and 164' between 1.5 MeV and 3.5 MeV and by Maurel (in
Mayer and Riminl, 1977) at = 150 in the energy range ofO.S2.0 MeV. The cross section has a broad maximum around 3
MeV. An example application of the reaction can be found in
Sagara et al. (1988). The (p,a.) reaction that can interfere is
1~(p,a.)160. but if one chooses the proton energy carefully, it can
be avoided.
105', and 1200 between 0.9 MeV and 1.3 MeV and at 0 = 1400
between 0.9 MeV and 3.0 MeV (Hagendom and Marion, 1957).
A broad resonance exists around 1 MeV, with a narrower one at
1.21 MeV. Cross sechoDS at ~ 120' and 0 ~ ISO' between 2.5
MeV and 5.0 MeV can be found in lausel-Hlisken and
Freiesleben (1977).lnterferenee with 180(p,a.)''N is possible.
137
0.777(90)
0.777(150)
0.840(150)
0.843190)
1.347(150)
1.354190)
1.652(150)
1.713(90)
1.842(t50)
1.901(90)
cr
(keV)
(mb/sr)
7
10
32
17.5
36
36
90
72
122
0.075
0.17
0.22
0.13
3.2
1.05
1.4
2.9
32
1.6
Chapter 6
629
1165
1766
-"
(mb/sr)
0.05
2.1
0.05
60
4.5
135
.-'
energy.
recommended.
~
!. 15
3.0
"c"
J= 10
2.5
0
"'"
--""
_.- d
,;
2.0
a
~
..""
15
>
~1.o
05
Energy of protons [MeV]
0.0
2
3
Energy or proton [MeV]
Ct,
and 3He energies are those that gave the same range as the
proton beam in the Mylar. The relationship between the proton
energy and the a., d, or 3He energy is linear, and the slopes are
1.35. 3.6, and 4.15 for d, 3He. and a., respectively.
138
139
Chapter 6
e ;: : ;
140
tao
6 J
8 ~-
'_'"
141
Chapter 6
:;G
!
-,
--- p
"< 4
"0:
..
E'
<
w
---
2
3
Energy of 'He [MeV]
3000
>'
..
!!'
_'4N(a.o:)''N
-
2500
- '4N(a,p)l'Q
-.~
"'si(a.pflp
,
,
,, ,,
, ,, ,
.!!.
1l
,
"~
,
-,,
,
0: 2000
1500
.{
;
"
~
a;
5003000
3500
4000
4500
5000
5500
5000
Appendix II.
The (a,p) reactions with negative Q values can interfere with
RBS measurements at high enough energies. This interference
could be impoI1ant in the case of elastic resonance
'4
FIG 6.11. The relationship between the energy of 3He and the
energy of the reaction products after they have passed through an
absorber foil that stops the primary beam.
-.-
reaction are above 3 MeV. and the peak of the protons coming
from the ''N(a,p)"0 reaction (Q = -1.1914 MeV) can overlap
with the peak of the elastically scattered a. particles in a certain
energy range.
142
Amsel, G., Faszti, F., Szilagyi, E., and Gyulai, J. (1992), Nud
Instrum. Melhods B63, 421.
For details on this reaction, see Endt and van der Leun
(1978). The resonance around 2.443 MeV was used for depth
profiling by Smulders (1986) and Kuper et al. (1986). McIntyre
et al. (1989) studied the depth profiling capabilities of this
reaction.
Baskhin, S., and Richards, HT. (1951), Phys. Rev. 84, 1124.
REFERENCES
Abel, F., Amsel, G., d'Artcmarc, E., Ortega, C, Siejka, J., and
Vizkelethy, G. (1990), Nuellnstrum. Methods B45, 100.
Berti, M., and Drigo, A.V. (1982), Nuel. Instrum. Methods 201,
473.
. __ . _.
Bird, J.R., Duerden, P., and Clapp, R.A. (1981), Nud. Instrum.
Melhods 191, 345.
Bonner, T.W., Conner, J.P., and Lillie, A.B. (1952), Phys Rev.
88,473.
Amsel, G., Nadai, J.P., Ortega, C., and Siejka, J. (1978b), Nuel.
Instrum. Methods 149, 713.
Chaturvedi, u.K., Steiner, U., Zak, 0., Krausch, G., Shatz, G.,
and Klein, J. (1990), Appl. Phys. Lett. 56, 1228.
143
-- -_. - - - - - - - - - -
..
"'\'1
Chapter 6
Kashy, E., Miller, PD., and Risser, J.R (1958), Phys. Rev. 112,
547.
.
Radioanalytjcal
Kashy, E., Perry, RR, and Risser, J.R (1960), Phys. Rf:v. 117,
1289.
Koenig, W., Riccato, A., Stock, R., Cuzzocrea., P., Perillo, E.,
to
Kuan, H-M., Bonner, T.w., and Risser, J.R. (1964), Nucl. Phys.
51,481.
Endt, P.M., and van der Leun, C. (1978), Nucl. Phys. A310, 1.
Lee Jr., L.L., and Schiffer, J.P. (1959), Phys. Rev. 115, 160.
Chapter 12.
L~
. 335.
Lin, Z., McIntyre Jr., L.e., Leavitt, I.A., Ashbaugh, M.D., and
Lu, X, Yuan, Y., Zbeng, Z., Jiang, W., and Liu, J. (1989), Nucl.
lnstrum. Melhods B43, 565.
Mani, G.5 .. Forsyth, P.D., and Perry, RR. (1963), Nucl. Phys.
44,625.
Manchester, U.K.
z.
Marion, lB., Weber, G., and Mozer, F.S. (1956), Phys. Rev.
104, 1402.
Phys. A283,
144
. '~
Maurel, B., Arosel, G., and Nadai, J.P. (1982), Nucl. Instrum.
Methods 197, 1.
McIntyre Jr., L.e., Leavitt, J.A., Ashbaugh, M.D., DezfoulyArjomady, B., Lin, Z., Oder, J., Farrow, R.F.C., and Parkin,
S.S.P. (1989), N"cI. Scj. Technol. 1, 56.
Mcintyre Jr., L.e., Leavitt, lA., Ashbaugh, M.D., Lin, Z., and
Stoner Jr., J.O. (1992), N"c!. Instnlm. Methods B66, 221.
Purser, K.H., and Wildenthal, B.H. (1963), Nud. Phys. 44, 22.
available
onlme
at
Monco/Ire, N., Nucl.lnstrum. Methods (1992) B66. 126.. __ ._. Qiu, Q., Arai, E., Aratani, M, Yanokura, M., Nozaki, T., Obji,
Y., and Imura, R. (1989), Nud. Instrum. Methods B44, 179.
Moller, W., Hufschmidt, M, and Kamke, D. (1977), Nu~l.
, Reichle, R~ Berisch, R., and Reith, I. (1990),' NuC!. Instrum.
lnstrum. Methods 140, 157.
Methods B50, 68.
Moller, W. (1978), Nuc! Instrum. Methods 157, 223.
Sagara, A., Kamada, K., and Yamaguchi, S. (1988), Nucl.
Moiler, W.o and Besenbacher, F. (1980), Nucl. Instrum. Methods
Instrum. Methods B34, 465.
168, 111.
Sawicki, 1.A., Davies, lA, and Jackman, T.E. (1986), Nucl.
Olivier, C, McMillan, J.W., and Pierce, T.B. (1976), J.
lnstrum. Methods 815, 530.
Radioanal. Chern. 31, 515.
Scanlon, P.J., Farell, G., Ridgway, M.C., and Valizadeh, R
Ott, W.R, and Weller, H.R. (1972), Nud. Phys. AI79, 625.
(1986), Nucl. Instrum. Methods B16, 479.
Payne, RS., Clough, A.S., Murphy, P., and Mills, P.J. (1989),
Nucl. Instrum. Methods B42, 130.
Schiffer, I.P., Bonner, T.W., Davis, R.H., and Prosser lr., F.W.
(1956), Phys. Rev. 104, 1064.
Seiler, F., Jones, c.R., Anzick, W.J., Herring, D.F., and Jones,
K. W. (1963), Nucl. Phys. 45, 647.
145
1,
Chapter 6
SigmaBase;
An
archival
site
is
available
online
at
http://www.mfa.ldki.hulsigmabasei.
France.
Vickridge, I.C. Trompetter, W.J., Browo, LW.M., and Patterson,
IE. (1995), Nucl. Instrum. Methods 899 454.
Vickridge, I.e., Trompetter, W.J., and Coote, G.E. (1996), N"cl.
lnstrum. Methods 8108367.
Sweeney Jr., W.E., and Marion, I.B. (1969), Phys. Rev. 182,
1007.
Symons, G.D., and Treacy, P.B. (1963), N"cl. Phys. 46, 93.
Szilagyi, E., Paszti, F., and Amse1, G. (1995), Nue!. lnstrum.
Methods, 8100, 103.
'.~
146
. -~
,[
I
CHAPTER
I
I
I
PARTICLE-INDUCED GAMMA
EMISSION: PIGE
J. Raisanen
University of Helsinki, Helsinki, Finland
Contributors: J.-P. Hirvonen and R. Lappalainen
CONTENTS
7.1 INTRODUCTION ......................................................................................................149
7.2 PARTICLE--GAMMA REACTIONS ........................................................................149
7.2.1 Deptb profiling with resonance reactions .......................................................... 149
7.2.1.1 Fundamentals .................................... ;..................... ,................................... 149
7.2.1.2 Choice of reaction ........................................................................................ 152 ..
7.2.1.3 Experimental considerations ...................................................................... 152
Accelerator ........................................................................................................... 152
Gamma-ray detection .......................................................................................... 153
Detector and beam energy calibration ................................................................153
7.2.1.5
7.2.1.6
7.2.1.7
7.2.1.8
7.2.1.9
147
~--.
_.. _ . _ - - - - - - - - - - - - - - -
Chapter 7
1
~'
:~
j
<
.~
..
.~
148
7.1 INTRODUCTION
The unique feature of any technique for nuclear reaction
analysis (NRA) is that it is sensitive only to specific isotopes.
This: of course, restricts its general applicability, but makes
NRA a powerful tool for those cases in which it can be utilized.
To understand the limitations and advantages of nuclear reaction
analysis with "(-ray emission and to enable the reader to decide
when these techniques are worth uSing, it is necessary to
compare this technique to other methods perfonned with
energetic ion beams.
7.2PARTICLE-GAMMAREAcnONS
I
When a beam of charged particles, such as protons, heliwn
ions, or deuterons, bits the sample surface layer, nuclear
reactions are induced, and gamma radiation is emitted. Here, the
discussion is restricted to prompt gamma emission, meaning that
radiation is detected during irradiation, in contrast to activation
techniques where the rndiation is detected after ilTadiation.
Activation analysis, which has been used widely to measure bulk
concentrations of elements with mediwn or heavy atomic
weights, is discussed in detail in Chapter 10. Activation methods
have also been used for profiling by etching the sample surface,
layer by layer.
149
Chapter 7
a+ A -+ A + a'.- A + a'+y.
(7.2)
Typical inelastic scattering reactions involved with particlegamma analysis are (p,p'y) reactions. With light (Z ,;; 20) target
nuclei, the excitation curves show resonance features on top of a
continuum. With heavier target nuclei, the inelastic scattering
occurs well below the Coulomb barrier, and the mechanism is
Coulomb excitation. In such reactions, the cross sections
increase smoothly as a function of bombarding energy. In
rearrangement reactions, particles other than the bombarding
particles are emitted. Such a reaction can be written as
a + A -+ B + b --+ B + b + 1'.
(7.3)
features.
The use of resonance reactions for light-element depth
profiling is based on bombardment of the sample with an
energy-analyzed partic1e beam and detection of the induced y_
ray emission as shown in Fig. 7.1. If the energy of the beam
corresponds to the energy of a narrow resonance Er, resonance
reaction can take place only on th.e surface. A.t higher
bombarding energies, the particles must slow down in the
sample to the depth at which Er is reached, when the reaction is
possible. The yield detected at this bombarding energy, E.,
reflects the concentration at this depth. Heo.ce, the excitation
curve of the resonance reaction YCEb). that is, the yield of the
reaction versus the bombarding energy, gives an estimate of the
concentration distribution in the surface layer (see Eq. 7.5.). As
examples of tbin- and thick-target excitation cwves, the curves
for Al are provided in Fig. 7.2.
(E-E,)'+r'/4'
E,>E,
i
J
--
I
!
,-~-~:----
..
..
(7.4)
150
"
1.0
z'
:?:
z
v AI(p,y)
992 k.V
--1>.=0.00
--1>.=0.05
_. _. 1>.=0.10
.. "1>.=0.15
- '''1>.=0.20
0.5
500
1000
1SOD
EE, (eV)
Experimental yield curve (points) from the
27 A1(p,y) 28 Si resonance in solid aluminum. (Figure adApted from
the data ofVickridge and Amsel (1996).1
FIG.
7.3.
151
1
Chapter 7
Depth (nm)
100
200
300
5
~
c
'"....
'C
Gi
's.
...>-,
?-
til
,,
----
-,
850
870
860
880
152
:;
10'
Gamma-ray detection
In particle-induced gamma emission measurements, the
radiation is generally detected either by a sodium iodide
scintillation NaJ(TI), bismuth germanate (EGO), or higb-purity
Ge (HPGe) detector. In all of these detectors, single- and doubleescape peaks are observed in addition to the full-energy peak
(photopeak) corresponding to the 'Y-ray energy. These escape
peaks are observed 0.511 MeV and 1.022 MeV below the
photopeak and are prominent and especially useful with highenergy gamma rays. The Nal scintillation counter is
conventionally used for profiling when high efficiency is needed
and energy resolution is not critical. This is normally the case
wilen Light elements are profiled in samples of high Z value
using well-isolated resonances. BGO detectors have higher
efficiency for a given size than NaI delectors and a larger
fraction of pulses in the full-energy peak. Furthermore, the BOO
detectors offer a bettcr signal-to-background ratio 3Ild a compact
setup, which are especially necessary when profiling with weak
resonances. HPGe detectors are generally used for the analysis
of total average concentrations of elements in the surface layer,
because these detectors have high enough energy resolution to
distinguish nearby gamma rays from competing reactions.
Before profiling, it is helpful to measure a gamma spectrum
from the sample with a Ge detector 10 identify .possible
interfering peaks from competing reactions or laboratory
background and to determine the suitable energy window
(region of interest) to sum the total yield. Improved efficiency
geometries have also been utilized leading to nearly 41t geometry
(piel et al., 1996).
:r7AI{p,,f'sI. 1m koV
~103
~
e
.:a 102
s:
10'
10'
0
10
12
Energy (MeV)
g(Eb,E)=[
(211:) cr b
]exP{-[(Eb-E)/crbll/2J, (7.6)
as
7.2.1.4 Cautions
Factors affecting measurements
The main factors determining the efficiency of measurements
are the strength of the resonance, the beam current, and the
153
Chapter 7
y '" w(6)nItSIE"
(7.7)
are due 10 2D'n (2614.6 keY) and <OJ( (1461 key), which are also
(7.8)
154
'}
104
lD
'"lD
~
~
."
III
N
\!!
"...m
103
"
"'"
~
I:
iO
lD "
t:!.g
W
<II N
::l
"
"C
Gi
>= 102
a-2O&n
101
b _ 214e1
c _228 Ac
d _ 212ei
,-.-~~-------
500
1500
1000
2000
2500
3000
Energy (keV)
FIG. 7.6. Typical laboratory background spectrum measured with an unshielded Ge detector. The origins of the strongest
(7.9)
155
Chapter 7
"'dE
= jS(E)'
(7.10)
x(E)= E, -E,
SeE,)
x(E) = x. +
,
,-I
2(E. -E. ,)
'
,-
SeE,) + S(E,_,) ,
(7.11)
i~2
Energy
(leoV)
Yield
(counts)
Depth
(nm)
Concentration
429.00
429.33
429.66
430.97
432.29
434.93
450.94
452.29
453.63
456.34
459.05
4B1.n
464.50
487.92
500
710
0
2.0
4.0
12.1
20.1
36.4
135.8
144.3
152.8
169.9
187.0
204.3
221.7
243.6
26.8
35.5
51.5
58.7
59.4
1190
1455
1485
1495
1510
1425
1200
945
185
25
10
0
(8l %)
59.7
60,0
57,9
51,8
43,9
11.1
1.61
0,65
sr.
~ 0.163
keVinm
a 992 keY
27A1(P;y)
28
156
the
0.'
05
..
0.'
02
990
0.'
998
1000
0.0
2.
*100
:(.
~
:
~A~.~I~
:!!
22
~:~~~::y
oL-~==='~'~'______~______~--1
OA
'"
1.
100
<a)
C.O
2M
Depth (nm)
(b)
0.'
...
18
(]
r."o..ubV
.,
1.0
.,
FIG. 7.7. (a) Experimental resonance yield step for the 992 keY 27 Al(P.y/8Si resonance. The sulid line is a fit obtained using the
Vavilov theory for straggling. (b) A Gaussian distribution and its integral.
no
-kT)V'
157
Chapter 7
with
measured yield curve using only Eq. (7.5). Extra conditions such
as smoothness, positivity, or boundary values are always needed.
fOflC<3
forl<~3
such as
Il8
600
r--------,--~---------,---.--~----,
500
51
E 400
c:
c:
o
,., 300
::s
Protons
T.
'0
Ul
CI)
a: 200
Helons
SI ././
/"
--
100
/"
Ta
Depth, nm
FIG. 7.8, Calculated depth resolution curves for 1 MeV protons and 2 MeV 4He+ ions.
--~.. ,.:....~
. _
Carbon
The 1748 keY resonance of the iJC(p,y)i'N reaction is the
159
Chapter 7
shown. In this case, the energy window for the yield curve
measurement has to be restricted to include only the 9.17 MeV
gamma-ray photopeak because of the interfering gamma rays
from Si.
Nitrogen
Both 14N and 1~ isotopes can be profiled with several
resonances. Those used most often are the 429 and 898 keV
12
1'N(p,ay) C and 1059 keY I'N(P,r)"O resonances. The best
depth resolution can be obtained with alpha particles, using the
1531 keY 1"N(Ct,y)IBp resonance. However, generally, the 429
keY (p,ny) resonance offers the best sensitivity, a longer
profiling range, and good depth resolution. Because the use of
this resonance is one of the most popular applications, the
selection of a proper energy window is discussed briefly here.
Oxygen
Oxygen profiling is possible using the IBO(p,y) resonances
and enriched samples. All resonances have complicated decay
schemes, leading to many intense gamma-ray peaks. Table
A12.1 in Appendix 12 lists only the strongest, high-energy
gamma rays. All resonances decay through the low-energy
excited states of 19p, leading to 110 and 197 keV gamma rays
that might also be due to proton inelastic scattering from 19F.
The resonance at 1167 keY has been used for high-resolution
surface studies. Figure 7.11 shows 1BO profiles in 90010 180_
enriched TazOs layers (Maurel et aI., 1982). The prominent
surface peak is due to the Lewis effect, as discussed before
(Section 7.2.1.1, Figure 7.3).
10'
!
(a)
10'
Energy (MeV)
10'
(b)
10'
Energy (MeV)
10
160
300
.
~
19p(p,cq)'50
250 i-
"--'....
.,. 200
'2
::l
.e 150
~
>=
'2c(p,-r)13 N
C
"C
Qi
i-
.::l
100 -
50
J.~
2
(a)
tSN(p,ccr) 12c
1~(p,y)'SN
.
~
...,. . ,"
..dI.
I.A
.",.
I
I? ,T~Ir
~
8
Energy (MeV)
300
250
'E
::l
.e 150
-'"
(b)
"C
]i
> 100
N
50
Energy (MeV)
FIG. 7.10. Gamma-ray spectra for carbon, nitrogen, and fluorine measured using a 500 keV proton beam with (a) a Ge detector and
(b) a Nal detector.
161
Chapter 7
100
200
300
Channels
Gz, =100 eV
Go = 101.5 eY
20
(~+D::.140eV)
r=105eV
XA\.o~ c
i/--
0,5,10,20,50,1 oo,2OCJA
..- ---
---
..I
~
(b)
,.
o L _ _ _ _~=-_=:.:._.._..___-_-_- _-L- _-_-_- =-.:-:.- :.-: ;-: ;~: ;-: ~:=-"_"~-:."': .~:.-, " ' =__
o
20
40
60
Channels
FIG. 7.11. (a) Excitation curves for 90% lBO_enricbed Ta20~ layers of increasing thickness X. The lines are best fits obtained using
stochastic simulations with the parameters given in the figure. (b) Calculated excitation curves for a thick aluminum sample covered
with an AI 20) layer of thickness X. The dotted lines represent the theoretical contributions of a 20-run-thick oxide layer and aD
aluminum sample covered by an absorber equivalent to that oxid.e layer (Maurel et a/., 1982).
"
162
,,
Fluon'ne
Fluorine is one of the most profiled elements, with (p,n'Y)
resonances and a typical interfering contaminant due to its many
strong and broad resonances. The most frequently employed
resonances are those at E, 340.5 and 872 keY. They mainly
decay emitting 6129 keY gamma rays, being typical for fluorine
contamination. Figure 7.12 illustrates the profiling of fluorine in
glass using the 340 keV resonance (Deconninck and van
Oystaeyen, 1983). The calibrated experimental resolution
0.3
>0
>.
"
iii
I:
Q)
0.2
.E
I:
:;::
1\1
'6
"''"
"
ffi
>
"
0.0
--=-ill
~
:>
"
0
"
>:!
Radiation emission
0.1
>
~
>
~
0.4
>
Glass surface
>
0.2
>
>
~
:=
'"
"
1\1
CC
0.0 I.L.......:"..L..-~c.......~~....lL...........l!...~d::............:::>.L......!.<:~....>..:_.......J
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1:6 1.8 2.0
Depth below surface (340 keV resonance)
(microns)
FIG. 7.12. Calculated yields initiated by the 340 keY resonance in fluoridated glass at different beam energies. The curve corresponds
to the experimental resolution. Interference from the next resonance at 483 ke V is also represented.
,...,,,,---,,,- -
..
-"
Neon
In principle, the Urree stable neon isotopes can be profiled
with (P,'Y) resonances, but only the 2~e (natural abundance
90.51%) and 22Ne (9.22%) resonances are sufficiently Darrow
and well isolated. The 1169 keY '''Ne(p,y)21 Na resonance is
especially well isolated, but the relatively low-energy gamma
rays (3.54 MeV) involved are easily overlapped by interfering
gamma rays. However, the 22Ne(p,y)23 Na resonances with high-energy gamma rays can be used for high-accuracy profiling even
with light backings (Lappalamen, 1986).
Aluminum
Aluminum has only one stable isotope, "AI. which has at
least 22 (p,y) resonances below 1 MeV. They are all very narrow
< 200 eV), but none of them offers a wide energy range free
of other resonances both above and below the resonance. The
two strongest resonances are those at :p = 632 and 992 keY
given in Table AI2.! (Appendix 12). Both of these resonances
decay with primary high-energy gamma rays that can be well
separated from those due to other light elements such as fluorine,
nitrogen, and carbon. In tbe case of deep profiles (up to several
microns), interfering nearby 21A1(P.'Y) 28Si resonances can be
excluded using primary high-energy gamma rays and a Ge
detector. Also, the strong (S = 250 eV) and narrow (r < 200 eY)
resonance at 1684 keY of the 27Al(p,p''Y)21Al reaction has been
shown to provide fast and accurate depth profiling of aluminum
(Mateus e/ al., 2008).
cr
Sodium
Magnesium
163
Chapter 7
Silicon
Silicon has been profiled mainly using the 620 keY
JOSi(p,y}3Ip resonance. Profiles can be measured with high depth
resolution using a Nal detector and a 7.0--8.3 MeV energy
window. Several weak, low-energy resonances for profiling 29Si
have been employed, but doing so requires improved gamma-ray
detection efficiency (piel ef al., 1996).
Heavy elements
Elements heavier than silicon are seldom profiled with (P,1)
resonances. This is because these resonances are typically broad
and not well isolated and otCW at high bombarcting energies.
Phospborus and sulfur are the main exceptions, both having
narrow resonances for profiling. The (p,y) resonances of Ti, Cr,
Ni, and Fe can be used for profiling, but generally, it is
necessary to use a HPGe detector to measure primary
characteristic gamma rays.
7.2.2 Elemental analysis using particle-Induced gamma-ray
emission (PICE)
.
In principle, any ion could be used with proper bombarding
energy to induce gamma-ray emission. Protons and alpha
particles are mostly used, but some sparse measurements have
bcen made using light ions such as deuterons, tritons, 3 He , Li, B,
N, F, andCI.
(7.15)
o
with i referring to the measured nuclide and m to the matrix. n is
the number of bombarding particles, E is the detection efficiency
including the solid angle, F" is the incident ion energy, fwi is the
weight fraction of nuclide i, 8m is the stopping power for the
matrix, and oJE) is the cross section for the specific reaction. In
addition, we assume thaI the nuclide to be measured is evenly
spread through the matrix.
General guidelines
The aim is to formulate an equation that enab1es estimation of
concentrations using thick-target gamma-ray yields of pure
elements. The main concern is finding a procedure that accounts
for the relationship between stopping power and gamma-ray yie1d
as a function of bombarding energy. The following procedure
follows the guidelines of Kenny ef al. (1980). A similar procedure
has also been presented, for example, by Ishii ef al. (1978a,b,c).
Using Eq. (7.15), the formulation for the thick-mget gammaray
yield for nuclide i in matrix m is obtained as
~~~
dE
(7.16)
where Y PlIIl' is the pure-element yield under the same
experimental conditions, and SlE) is the stopping power for the
164
I
.1
= Y B Y;" SA (E) f
wB
yB
A
pure
S (E)
wA'
(7.19) ,
(7.17)
(7.18)
Here, the energy E1I1 is the energy at which the yield for the
pure-element target bas decreased to one-half of its initial value.
that is, Y 1Mt:(E1I2) = 1/2 Y pure{Ep). The difference in the yield
values calculated from Eq. (7.16) compared to those obtained
using the geoeral guidelines [Eqs. (7.17) and (7.18)] was studied
by Kenny ef 01. (1980). They found that the guidelines provided
values within 1.5% of the values obtained using Eq. (7.16).
Use ofstandards
l.Wcts 72x]if N,U.cst 9.62 MV,(rrgcm')
Concentrations of elements distributed homogeneously in
c ~
]O~35JPI1
thick samples can be obtained by comparison to standards. Many
lillds 4.6x10' N,(j.Cs-t RISMV,(rrgan-')
.
multicomponent elemental standards exist for the analysis of
geological, biological, and medical samples [e.g., National
lrutitute of Standards and Technology (NISn, Institute for
The relevant stopping power values have been taken from SRlMReference Materials and Measurements (lRMM), International
2008. Note for possible peak overlapping.
Atomic Energy Agency (lAEA)J. In many cases, researchers can
prepare the proper standards themselves, by mixing a known -_ ..
amount of the element to be determined into a matrix similar to
. The general guideline has been applied successfully, Jor
the one being studied. For example, the detennination of carbon,
example, in the determination of Na, Mg, AL P, aDd CI in
nitrogen, and oxygeo can be carried out with organic compouods
geological samples (Olabanji et ai" 1996). It should be pointed
as standards. The known stoichiometry for C, N, and 0 of thc
out that, for accurate analysis, the relevant pure-element yields
compounds can be used directly. As a restriction preparing one's
should be measured using the appropriate geometrical setup (see
own standards, the seleoted organic compounds should not have
also Section 7.2.2.4). For an order-of-magnitude resul~ the
literature pure-element yield values can be applied.
very high concentrations of any single element (above about 40
wt %). This procedure is a consequence of the similarity of the
stopping powers for typical organic samples and the compounds
7.2.2.3 Experimental considerations
used as standards (Lapatto and Riiisiiaen, 1988).
Several variations of the experimental arrangements used for
PIGE analyses have been reported. The trend has been to
The stopping power values can be obtained, for example, using
construct multipurpose chambers for simultaneous PIGE, PIXE,
the SRlM-2008 program (www.SRlM.org). If the e~citation curve
and RBS measurements. When low concentrations are
is known as a :function of the bombarding energy, more accurate
determined,
minimal system background is essential. In choosing
calculation can be made using a simple computer program,
the construction and lining materials for the experimental setup,
although this is not generally necessary. The general guidelines
the thick-target gamma-ray yield tables of Appendix 12 can be
presented in Eqs. (7.17) and (7.18) can be employed, which, in
employed. It should be noted that the best materials of choice are
twn. simplifies the standardization procedure. In this case, the
not the same for PIXE and PIGE.
concentration of only one element (element A) in the unknown
_0
_ _ _ _ _ _ _ _ __
165
Chapter 7
In-vacuum measurements
In all experimental considerations, careful shielding and
lining are needed owing to particle scattering behavior. Tantalum
is commonly used for shielding from scattered particles. With
tantalum, low-energy gamma-ray lines occur, so better materials
are tin and lead. Tin is ideal when the proton energy is below
....2.5 MeV, and lead can be used at higher energies. Lead is often
used as an absorber in front of the detector to reduce the count
rate and improve the sensitivity. Copper is often used as a
construction material because it is a good conductor of heat. A
drawback of copper, however, is its contribution of spurious
peaks to the ganuna spectra.
External-beam measurements
External-beam PIGE can be applied to the analyses of
sensitive materials (e.g., paintings). Special attention should also
be paid to the exit-foil material. Polyimide (H"C"N,O,)
(Kaptan) is commonly used in external-beam PIXE
measurements, and it can also be employed for PIGE analyses.
The induced background is low but interference can result from
the C, N, and O. Also, exit-foil cooling might be required in some
cases. With organic exit foils and ions heavier than protons and
deuterons, a cautious approach should be adopted because of the
large energy losses involved, which can lead to rapid foil rupture.
In PIGE, high beam currents are often used, and therefore,
metallic exit foils, such as havar (a high-strength alloy), titanium,
or nicke~ are more coDvenient This is a.q important point to be
considered, especially when carbon, nitrogen, and oxygen are
analyzed The gamma-ray peaks originating from the exit foil can
be used to nonnalize the beam current of various measurements
(Raisanen, 1986). The gamma-my yields are generally
sufficiently high for such pwposes but not so high that they
provide a disruptive background signal In some cases, PIXE and
..... PIGE measurements are done simultaneously with the same
setup, although the best sensitivity is not then achieved for both
methods. A multipurpose external-beam PlXE-PIGE-RBS setup
is depicted in Fig. 7.13.
7.2.2.4 Cautions
It should be noted that, when perfonning PIGE measurements
with ion energies close to the energies of strong resonances,
caution should be taken because a small change in the
bombarding energy can lead to significant changes in the yields.
Some additional points that should be kept in mind are discussed
in this section.
Geometry
Generally, one can expect thick-target gamma-ray yields to be
isotropic. However, when the yield is mainly due to strong
resonances, noticeable deviations from isotropic distributions can
be observed (see, for example, Kenny et aJ., 1980). In such cases,
the angular distributions should be taken into account if the
general guidelines arc employed. When fixing the measurement
166
Peak broadening
In some cases, the gamma-ray peaks are significantly broader
than the resolution of an HPGe detector. When gamma rays are
emitted while the residual nucleus is moving, a shift in the
ganuna-ray energy is observed. This broadening is mainly due to
the short lifetimes of de-excitation states and the high recoil of
the light nucleus (Doppler effect) or to a large resonance width.
The Doppler effect is especially significant in the case of light
target nuclei. The Doppler-shifted energy of gamma rays is
determined by the velocity of the recoiling nucleus relative to
the observer according to the expression (up to second order)
(7.20)
where is the angle between the beam direction and the sampledetector line, v is the velocity of the excited nucleus, c is the
velocity of light in a vacuum, and E~ is the energy of the
gamma ray if the nucleus decays at rest.
The recoil nucleus must decay during flight before stopping.
In the case of solid materials, the lifetime of the excited state
should be between 5 x 10- 15 and 2 x 10-12 S in order to cause
Doppler broadening. The shape of the Doppler-shifted gammaray peak depends on the lifetime of the excited state and the
stopping power of the retarding medium. In cases where the
products of the average stopping power and mean lifetime are
the same, similar line shapes are obtained. The effect is most
prominent at 0 and is not observable at 90. Peak broadening is
evident with protons for Ey values of 478 keY (Li), 3562 keY
(Be), 429 keV (B), 4429 keV (N), and 1634 keY (Na) and with
'He+ ions for E, values of 478 keY (Lil, 4439 keV(Be), 3684
keY (B), 1275 keY (F), and 1809 keY (Na). The peak
broadening is useful in identifying the origin of the gamma rays,
especially for elements having only one gamma-ray line.
Lithium has two stable isotopes 5Li (7.5%) and 7Li (92.5%).
'Li can be easily detected using the 478 keY y-ray line, which
is slightly Doppler-broadened. For 6Li analysis, no usable y-rayemitting reactions exist. When analyzing samples with high
lithium concentrations, one should keep in mind that neutrons
are easily emitted.
167
";- 'i
Chaptel7
Carbon has two stable isotopes, IlC (98.9%) and IlC (1.1 %).
Detection of carbon at low proton energies is Dot sensitive. A
good sensitivity for "c is obtained at higber bombarding
energies employing the 12C(P,p'r)12C reaction and the 4439 keY
line. At energies below the strong 5.370 MeV resonance, the
sensitivity is poor, and above about 8 Me VI the background
usually increases significantly. The optimum energy region is
thus between 7.5 MeV and 8.0 MeV. In some cases, nitrogen can
cause interference (sec below). l3e can be determined with
moderate sensitivity at low bombarding energies.
Nitrogen has two stable isotopes, ''N (99.63%) and "N (0.37%).
The most significant 1-ray line originating from "N at low
proton energies is the 4439 keV line due to the 1~(p,ar)12C
reaction. The Doppler-broadened peak in the spectrum can be
Sulfor bas four stable isotopes, "5 (95.02%), "5 (0.75%), "s
(4.21%), and "5 (0.02%).
Sulfur can be detected employing the reaction 32S(p,p'r)12S
(E, = 2230 keV) at the level of 100 ppm (by weigbt) from thick
organic samples and at the 100 nglcm2 level from aerosol
samples (Raisanen and Lapatto, 1988). The optimum
bombarding energy is about 4.9 MeV. The reaction shows two
strong resonances at energies of 4.77 MeV and 5.10 MeV. At
168
Iron has four stable isotopes, S4Fe (5,9%), s6pe (91.72~), s7Fe
(2.1%), and "Fe (0.28%).
.
Iron has a significant line at 847 lee V that can be used for
practical purposes. It is often found as a background peak in
PIGE spectra due to scattering from stainless steel materials.
Unfortunately, the 847 keY line often overlaps with the strong
844 keY line from aluminium.
169
1
Chapter 7
DTGE can also be used for S, el, and K analyses. The detection
limits obtainable depend on the sample and the background
Low-yield elements:
The y-ray yields of the heavy elements are generally rather
low. Several such elements are employed as low-background
materials for construction purposes. With proton energies below
3 MeV, the best low-yield elements are tin (no l-ray lines) and
Low-yield elements:
The element, of Fe, Ni,
ell, Zn,
although the rather strong y-ray lines at 136 keV, 165 keV, and
302 keV increase the background stightly.
12).
Low-yield elements:
The best low-yield materials are the heavy elements; for
example, Mo, Pt, and Au can be recommended.
intensities and net counts/collected charge. The absolute gammaray yields for Ti through Zn (excluding V) induced by.5 and 10
MeV a-particles are given in Kocsonya et al. (2006).
It should be noted that. with a-particle energies below 5 MeV,
no ')'-ray lines originate from carbon. The elements that are easily
dete<:table by u-particle bombardment are Li, Be, B, and F.
When comparing the yields obtained by 2.4 MeV a-particles
with those obtained by 2.4 MeV protons, it can be concluded
that the a-particle-induced yields are higher only for beryllium.
and detecting the trompt 871 keV y-rays emitted from the first
For beIYllium, the sensitivity obtained by a-particle-induced rray emission (sometimes referred to as AlGE) is about 1000
170
Low-yield elements:
Of the light elements, carbon can be used as a low-yield
material. Of the heavier elements from Co and Ni, no y-rays
originate. Also, the yields for Y, Nb, In, and Sn are sufficiently
low for these materials to be employed as construction materials.
7Li-induced y-ray emission
Energetic 1Li ions are sometimes used in RBS measurements
for substituting 4 He ions. Lithium-ion-induced y-ray emission
can be used in connection with other ion beam analytical
techniques for simultaneous detection of the light elements.
Gamma-ray yields induced by 12 MeV and 18 MeV 7Li ions are
provided by Raisanen (1990). Elemental analysis using 7Li ions
is surveyed in Table AI2.l5 (Appendix 12).
Although the y-ray yields increase with increasing particle
energy (from 12 MeV to 18 MeV), the background in the spectra
are also elevated. As a result, the minimum detectable amount of
an element is not necessarily lower for increased particle energy.
The exact optimum bombarding energy depends on the matrix
and the element to be determined. A limitation on the use of 1Li
ions is that Li, Be, and B cannot be detected, as no significant rray lines originate from these .e)ements.
Low-yield elements:
..... ,..........- .
Lead is a good low-yield material to be used with 7Li ions.
Because of its rather high neutron yield, carbon is not usable as a
construction material in experimental setups when 7Li ions are
employed. Tungsten, hafuium and tantahun are good backing
materials, for example, for nuclear physics experiments with 7Li
beams. No significant y-ray lines originate either from
zirconium.
Heavy-ion-induced y-ray emission
He, UN, and J60 ions: Backscattering spectrometry (BS) and
elastic recoil detection analysis (ERDA) can be carried out with
16
12e, I"N, and 0 ions. In such dete~ations. the simultaneous
we of the ion-induced y-ray emission technique can be
exploited. Absolute y-ray yields induced by 22 MeV and 28
MeV 12C ions, 28 MeV 14N ions, and 28 MeV and 33 MeV 160
ions for the light elements are available in the literature (Seppala
et aI., 1998). The bombarding energies are well above the
Coulomb barriers of the light elements (2 < 9), for which several
strong y-ray lines are reported. The highest yields obtained by all
ions are for beryllium. The yields induced by 12C, l"N, and 160
ions can be compared ,with the values obtained by other ions.
With 12C, '~. and 160 ions, the values are systematically higher
for beryllium, carbon, nitrogen, and oxygen than for 2.4 MeV
protons (equivalent energy/nucleon). The "(-ray yields induced
by 2.4 MeV a-particles are systematically lower for all elements
except for beryllium. It can be coI1cluded that the 12C, '""N, and
16
0 ion-induced yields are sufficiently high for p:ractical
detection of carbon, nitrogen, and oxy~en. The relevant reactions
and yields for analyses with 12e, ~, and 160 bea::m.s are
provided in Table AI2.l6 (Appendix 12).
Low-yield elements:
The yields are low for elements heavier than titanium., and all
such elements can be used as construction materials. Also, the
neutron yields induced by the heavy ions are low.
JJCI ions
The usability of 55 MeV 3sCI+7 ions for elemental analyses
was demonstrated by Borderie el al. (1979). For all other
elements except helium, lithium, and beryllium, 55 MeV is
below the Coulomb barrier. The method was found suitable also
for analyses of heavier elements in special cases. The y-ray
yields and detection limits obtainable from a niobium backing
are listed in Table A12.17 (Appendix 12). The detection limit
values refer to one hour of irradiation with a beam current of 0.5
!lA.
Low-yield elements:
As expected, heavy elements are the best low-yield materials.
In all cases, the reactions are Coulomb excitation, and the
involved y-ray energies are low.
171
.,';
Chapter 7
REFERENCES
Adler, P.N., and Schulte, RL. (1978), Scripta Metall. 12,669.
Hunt, S.E., and Jones, W.M. (1953), Phys. Rev. 89, 1283.
Ishii, K., Sastri, C.S., Va11adon, M., Borderie, B., and Debrun,
lL. (1978a), Nuel.Instrum. Methods 153, 507.
Ishii, K., Valladon, M., Sastri, C.S., and Debrun, 1.L. (l978b),
Nucl.Instrum. Methods 153, 503.
Anttila, A., Bister, M., Fontell, A., and Winterbon, K.B. (1977),
Radiat. EfJ. 33, 13.
Aslam, Prestwich, W.V., and McNeill, F.E.
Radioanal. Nuel. Chern. 254, 533.
Ishii, K., ValladoD, M., and Debrun, J.L. (1978c), Nucl. Instrum.
Methods 150, 213.
(2002), J.
Badica, T., Besliu, C., Ene, A" Olariu, A" and Popescu, I.
(1996), Nuel.Instrum. Methods B111, 321.
Kiss, A.z., Biron, I., Calligaro, T., and Salomon, J. (1994), N.el.
Instrum. Methods B85, 118.
Kumar, S., and Raju, V.S. (2004), Nuel. Instrum. Methods B226,
631.
Distribution Profiles from Resonance Reaction GrnmmaRay Yields", in Ion Beam Surface Layer Analysis, Meyer,
yol.2, p. 851.
262.
Lapano, R., and Raisanen, 1. (1988), Phys. Med. BioI. 33, 75.
Elekes, Z., Kiss, A.Z., Biron, I., Calligaro, T., and Salomon, 1.
(2000), Nucl. Instrum. Methods BI68, 305.
Ene, A., Badica, T., Olariu, A., Popescu, LV., and Besliu, C.
(2001), Nuel. Instrum. Methods B179, 126.
Mateus, R., Jesus, A.P., and Ribeiro, J.P. (2005), Nuel. Instrum.
Mateus, R., Fonseca, M., Jesus, A.P., Luis, H., and Ribeiro, J.P.
Maurel, B., Amsel, G., and Nadai, J.P. (1982), Nuel. Instrum.
Methods 197, I.
317.
172
van Bebber. H.. Borucki, L., Farzin, K., Kiss, I\.Z., and Schulte,
173
CONTENTS
8.1
8.5
8.6 SELF-CONSISTENT ANALYSIS COMBINING NRA AND RBSDATA: TIlE NEED FOR
NRA DATA FOR PROPER RBS ANALYSIS ......................................................................... 184
8.7 SUMMARY AND CHECKLIST OF NUCLEAR REACTION
PROFILlt"IG PROCEDURES AND CONCERNS ................................................................ 185
REFERENCES
........................................................................................................................... 186
175
Chapter 8
8.1 INTRODUCTION
Megaelectronvolt ion beam analysis for hydrogen is
important for the following reasons:
Hydrogen is probably the most common elemental
contaminant, especially in thin~film materials.
The presence of hydrogen can have dramatic effects on the
electrical, mechanical, and chemical properties of some
materials.
Hydrogen is invisible in most modem analytical probes.
Of
The three resonant reactions that have been used far more
than other.; arc those induced by "N (Lanford el ul., 1976), IOF
(Leich and Tombrello, 1973), and 'Li (Adler ef 01. 1974).
Parameters for each of these reactions are given in Table A13.1
of Appendix 13 (Trocellier and Engelmann, 1986; Xiong er 01.,
1987; Barnes el 01., 1977). Each of these reactions has its own
advantages. For most applications, the I~ reaction has the
advantage of having the best combination of analytic
characteristics (depth resolution and sensitivity). The 19p
reaction bas the advantage that it can be conducted using natural
(as opposed to isotopically enriched) F in the accelerator ion
source.' The 1Lj reaction has the advantage of allowing profiling
to much greater depths in a sample than either of the other
reactions.
176
r
I
/~
"WindOii!
Beam
--75A
/ / /
Hydrogen
detection
resonance
ii;}? !
SA EXPERIMENTAL DETAILS
Figure 8.2 shows a schematic of the experimental chamber
used for H profiling at the State University of New York
(SUNY) at Albany, NY. The 1'N beam enters from the right and
bombards the samples, which are mounted on the circumference
of a rotatable sample wheel. The gamma rays from the reaction
are detected by a scintillation detector [a NaI crystal or a
177
Chapter 8
Rotatable
Photomultiplier
valve
sample
wheel
Samples
8.4.4 Backgrounds
The large scintillation detectors used to measure the gammaray yield for NRA are also sensitive to other sources of
radiation, most notably cosmic rays and natural radioisotopes
(1.46 MeV gamma rays from the decay of"'K and 2.61 MeV
gamma rays from the decay of
to 203Pb). In some
situations, background radiation can also be produced by the
accelerator. When measuring H at low levels, it is important to
measure background carefully. This can be d.one, for example,
by recording the gamma-ray counts from a "H-free" sample such
as a clean Si wafer that has a very low level of intrinsic
hydrogen.
mu
178
'f
E - Eres
ended machine
When running I5N 2+ beams from a single-ended accelerator
without a system in the terminal that selects the charge-to-mass
ratio of the particle being accelerated, there will likely also be a
one-quaner-energy "contaminant" beam. This beam comes from
the acceleration ofN 2+ ions that break up after acceleration, but
before the magnetic analysis system. This one-quarter-energy
beam and the full-energy beam follow identical trajectories in a
magnet field (they have the same value of MElq'). Separating
this beam from the full-energy beam requires electrostatic
analysis (e.g., deflecting the beam with electric steerers and
bringing it back with magnet steerers).
i
I
I
dE/dx
X=
(8.1)
y =
Qn(x)s(x)dx
(8.2)
Y =Q
(803)
a(E)=
r
aO4
(E-E
i
res
+L
y_Qnlt/2aor
dE/dx
(8.4)
(8.5)
or
n(x) = K (dE/ dx) Y(x)
(8.6)
179
Chapter 8
0.0
0.1
0.2
0.3
0.4
0.5
8
300
10
.Q
.r
6 E
.!!
"
0
tJ
:..
-"
......
u
0....
200
!!!
tJJ
4 ~
U
c:
Counts -:- 50
c:
0
tJ
C,)
100
2
- ---- -- -
0
6.4
6.5
6.6
6.7
__
B_a:~~ro~:d
6.8 6.9
0.0
~ 1000
.!!
7.0
7.1
7.2
J:
Depth In microns
0.1
0.2
0.3
0_5
0.4
0.6
Amorphous silicon
~ 0.55 microns
8"
8"
500
o~----~--
0.2
__
______
0.4
____
__
0.6
E-ERES (MeV)
FIG. 8.3. Typicall~ nuclear reaction profile data for (top) a Si sample implanted with I016H1cm 2 at 40 keY and (bottom) a thin film
of hydrogenated amorphous Si on a Ctystalline Si substrate. Both raw data (gamma-ray counts versus beam energy) and final results
(H concentration versus depth) are given. Note that the solid data point (bottom) was measured after the profile had been completed to
verify that H was not lost during analysis.
180
I
EXAMPLE 8.1. For the chamber shown in Fig. 8.2, using a
3-in.-diameter by 3-in.-long cylindrical BGO detector 2 em
behind the sample, with Y being counts per 1 ~C (2 J.lC of
1~2l of incident I~ particles, integrating all events in the
4.4 MeV gamma-ray full-energy peak and the 3.9 MeV
single-escape peak, the constant K is obtained as K = 0.45 X
10" ~mI(MeV cm'), or Eq. (8.6) becomes
n ~ 0.45 x 10" (dEJdx) x Y
(8.7)
different from that given here. Data using this setup are
discussed further in Section 8.5.1 on calibration procedures.
=11.26 MeV/(mglcm')
8.5.1 Chamber calibration: Determination of K
Although, in principle, the chamber constant K can be
determined from-the reaction cross section, resonance width, and
detector efficiency, it is dangerous to trust the literature for
absolute cross sections to the accuracy needed for analysis. The
usual procedure is to record H profile data from a sample of
known H content and determine K by comparing these data with
the known H content. There are two types of standard samples
that can be used: one bas a known amount of H per unit area
2
(cm ), and the other has a known amount of H per unit volume
(cm').
181
Chapter 8
Y'ff~' = (1.28
(8.8)
Hr')!
where I is the integrated area (in keV counts) of the surface layer
peak.
2.0
f1: Lorentz
t2: G3uss
1.5
8.5.3 Depth resolutioll
There exists some confusion about the depth resolution of
these methods in the literature. Commonly, the near-surface
depth resolution was considered to be given simply by the width
of the resonance divided by dEldx. Deeper into the sample, the
depth resolution would be dominated by straggling of the incjdent
beam. For the ''1' case, the resonaoce is so narrow (1.8 keV) that
the motion of H bonded in a solid will shift the reaction on and
off resonance' (Zinke-Allmang e/ al., 1985). As a result, the
reaction behaves as if the cross section were not Breit-Wigner
(Lorentzian) but rathcr Gaussian, with the width
.. -determined by how the H is bonded (in the range of5-15 keV). For
the I~ reaction, this Doppler effect dominates the near-surface
depth resolution.
1.0
FlG. 8.4. Plot of the width that results from the convolution of a
Gaussian with a Lorentzian (Breit-Wigner), as well as the result
for a Gaussian convoluted 'with a Gaussian (Arnsel and Maure~
1983). The results are expressed as the ratio of the total (T) width
to the width of one of the functions. See EXAMPLE 8.3.
"
.~.
..','
;~
182
I
EXAMPLE S.4. Evaluation of the Off-Resonance Yield.
Figure 8.6 shows the raw data for a hydrogen profile of a
sample that consisted of plasma-deposited silicon nitride
(very high hydrogen content). on aluminwn (low hydrogen
content), on a silicon substrate. The surface nitride gave a
very high count rate (about 5500 counts/4.8 ~C). The
(a)
~ Hydrogen-bearing layer
(b)
10000 , - - - - - - - - - - - - - - ,
.
8.S.S High-sensitivity measurements
The ('N profiling technique outlined above gives a sensitivity
on the order of hundreds of parts per million (atomic), with
1000
.
.
.
beam currents on the order of SO nA and run times on the order
.....
- ..
ofa few minutes. To measure sarilples with lower H content, it is ......... ~
necessary to (1) increase the beam current, or (2) decrease the
100
background, or (3) increase the detector efficiency, or (4) use a
reaction with a larger cross section. All of these are done. The
increase in beam current depends on the accelerator available
10
and how much beam a sample will take (with cooling and beam
rastering). The decrease in gamma-ray detector background is
1~~~~~~~~~
accomplished with material and electronic shielding of the
6.3 6.5 6.7 6.9 7.1 7.3 7.5 7.7 7.9
gamma-ray detector to reduce the number of cosmic-rdY events
Beam energy (MeV)
recorded (Damjantschitsch et o/.. 1983; Kuhn et 0/.. 1990; Horn
FIG. 8.6. Raw data for a hydrogen profile ofa sample consisting
and Lanford, 1990). New well-type BGO detectors provide a
of a thin layer of plasma silicon nitride on aluminum 00 a silicon
high detector efficiency compared to the commonly used 3-in. x
substrate. The very high-countrate surface peak (6.385-6.55
3-in. cylindrical detector (Kuhn et a/., 1991). Going to the 13.35
MeV) corresponds to the H in the silicon nitride. The low-countMeV 1~ resonance increases the gamma-ray yield by an order
rate region (6.7-7.3 MeV) corresponds to the alwninurn layer,
of magnitude (at a sacrifice in depth resolution and more
and the peak at 7.5 MeV corresponds to the Al/Si interface. The
complicated off-resonance contributions). Hence, there are many
approximately 10 counts in the Al region are attributed to the
ways to increase the sensitivity of this method, and all of these
off-resonance reaction with the surface H layer. See EXAMPLE
approaches have been explored. However, experience indicates
8.4.
that many research problems in this field can be effectively
studied without such changes.
---.,'-~.--.
!
1
183
_-_._.
Chapter 8
1500
'"
1000
soo
'\Si
o .
"
100
200
300
Channel
,00
'00
600
CJSi to re-analyze the NRA data and using the new H/Si values
to determine new c/Si values results in final compositions of
SiC, O.sHL14 and SiCl.ooHlJ.3.4
184
Repeating the RUW fit of the RES data and then the re-analys!s
of the NRA data gives new values for y and z as follows:
zeroth-order first-order
second-order
y,z
y, z
y,z
as-<iepnsited
0.65,2.00
0.88, U8
1.05, 1.24
annealed
0.89,0.33
0.41,2.00
1.00.0.34
Comments
It is worth noting that the effect of hydrogen is large in this
case because RES spectra were being analyzed making use of
only the Si RES signal to determine the C/Si ratio, that is,
determining the carbon content by subtraction. If the RBS
spectra were analyzed by measuring the RES C signal relative to
the RBS Si signal, the effect ofR would be negligible. However,
accurate measurement of the RBS C signal is difficult because of
the large and uncertain background from the Si substrate.
= (28
where the stopping powers for 6.4 MeV nitrogen in Si, C, and H
are 6.24, 8.45, and 28.4 MeVI(mglcm'), respectively.
K[(dE/dx)p/IO]Y
Si
p(6.02xlO")/gmwt
Z=-
-'-.
.. P
K(dEldx)(gmwt)Y
IOx(6.02xlO")
where K is the chamber calibration constant (0.45 x 10 19 for
Albany), dE/dx. is the stopping power of the film in
MeV/(mglcm'J, gmwt = gram molecular weight of our
molecular unit, and Y is the NRA reaction yield (in counts12 J-lC
for SUNY -Albany).
185
Chapter 8
REFERENCES
Adler, P.N., Kamykowski, E.A., and Padawer, a.M. (1974), in
Hydrogen in Metal, American Society of Metals, Materiais Park,
OH,p.623.
Amsel, G., and Maurel, B. (1983), Nud. Instrum. Methods
218,183.
Barnes, C.A., Overley, lC., Switzkowski, Z.E., and Tombrello,
T.A. (I 977), Appl. Phys. Lett. 31, 239.
Cohen, B.L., Fink, C.L., and Degnan, I.H. (1972), J. Appl Phys.
43,19.
845,256.
Xiong, F., Rauch, F., Sm, C., Zhou,' Z.,Livi, RP.; and
Tombrello, T.A. (1987), Nud. lnstrum. Methods 'Phys. Res.
B27,432.
'1
"
,',
186
CHAPTER
II
HYDROGEN ANALYSIS BY PROTONPROTON SCATTERING
P. Reichart and G. Dollinger
Universitiit der Bundeswehr Miinchen, Neubiberg, Germany
CONTENTS
9.1 INTRODUCTION .......................................................................................................... 189
9.1.1 Basics of the proton-proton scattering technique ............................................. 189
9.1.1.1 Kinematics .................................................................................................... 189
9.1.1.2 Coincidence filtering .................................................................................... 190
9.1.1.3 Energy analysis ............................................................................................. 191
9.1.1.4 Timing ........................................................................................................... 191
9.1.2 Quantification .....................................................................................................192
9.1.2.1 Loss due to angular spread .......................................................: ...:::.:..:. .:.......192
9.2 SENSITIVITY LIMIT .................................................................................................. 194
9.2.1 True events versus accidental coincidences ..................................................... 194
9.2.2 Irradiation damage ............................................................................................ 195
9.2.3 Hydrogen microscopy using a proton microprobe ......................................... 197
9.3 DEPTH PROFILING .................................................................................................... 198
9.3.1 Depth scale .......................................................................................................... 198
9.3.2 Depth resolution ................................................................................................. 199
9.3.2.1 Correlated-path-Iength effect, IiE,,.th .......................................................... 199
9.3.2.2 Small-anglc-scattering-path effect, SEa ..................................................... 199
9.3.2.3 Energy-difference method ........................................................................... 200
9.3.2.4 Sum of contributions ....................................................................................200
9.4 EXPERIMENTAL REQUIREMENTS ....................................................................... 200
9.4.1 Sample thickness ................................................................................................ 200
9.4.2 Detector thickness .............................................................................................. 202
187
Chapter 9
.................................................................................................................. 205
188
9.1 INTRODUCTION
This chapter presents a brief collection of the basic
principles, advantages, limits, and technical requirements of
pratall-jJrotan (Pp) scattering far hydragen analysis. The
principle of proton-proton scattering is illustrated in Fig. 9.1.
Two protons are scattered elastically from a reaction between an
incident proton of kinetic energy Eo and a hydrogen atom in the
target and are detected in coincidence in transmission geometry.
The benefit of coincident detection is that it filters out elastic ppscattering events from all other reactions, in particular from
nuclear reactions, giving a background-free hydrogen analysis.
Because the absolute kinetic energy minimwn (in the laboratory
system) for pion production is 286 MeV, this scattering can be
regarded as essentially elastic even up to energies of about 350
MeV (Bethe, 1999).
E,
E,+E.=E..
which is the energy of the protons before scatteri:ng. Note tJ
we use Eo for the incident beam energy before entering the tarj
but Ein for the incident energy before scattering.
<-;J
sample
..
......
,.
. ...
(9.1)
189
Chapter 9
lill- Eo - E, - Em(eos'
a - sin' 6) -
Em cos 26
(9.4)
190
21
'23
l2'
pa~ksuri'~c( ,
~lOd :H~atJcm'
"'24
front surface,
-10" n-.item'<
.buikregion
li
i
depth
resolution
-8~m
depth resolution
FWHM-15~
higb
I"",' energy
background
eoergy
SG
-100
-50
appr"ox. depth scale (pm)
FIG. 9.3. Depth profile from 25 MeV proton-proton scattering of. quartz sample that was polished to a thickness of about
130 J1ID and analyzed without further surface preparation. Hence, surface peaks are dominant but can be sepamted from the
bulk bydrogen content of about 10 at-ppm.
9.1.1.4 Timing
The scattered protons have different flight times to the
detector if 9 +- 45. For scattering on a thick sample with
thickness d, the maximum differeoce in the time of flight for
scattering is observed at z = 0 (at the froot of the sample; see
Fig. 9.1). Assuming a constant stopping f<?rce along the paths,
one can calculate, in a Doorelativistic approximation, the time
difference, 6.1.1.(9), for the two protons hitting a perpendicular
detector plane at distance D [Fig. 9.1(a)) or the time difference,
.
., 9 Sid
- smS
E sm
(
o
sine
l!Ild
191
J-'
(9.5)
Chapter 9
,
S,d
- cosS Eosin 8--.sma
-'
, (9.6)
. (a)__ ._ Detektor
:i
lBearn AXIs
D=29 rnrn'
3 MeV
d (Si) =
10)lm /
.5
2 .5
:
.
.
. ..
,; I.S
.E
.,
: 3 MeV
"
~OMeV
d=O
...
-",#'
... "
3 MeV
! d (Si) = 10 J.lm
2.5
..,,
.
I.S
30 MeV
d=O
........
.;: O.S
,.,
:
"
....
..
-R=~O
IOO,...m,'
,
'
;;:'"
d (Si) = :
..
30 MeV
d (Si) ~ 100 J-lm
...
05
"~- ~...
.##
--SJ.5""~-~-6~0""--~65""-':""''':''~70.64'~5'~'~-';'--;;~;';~~0~'''''''''''~5SO::.:''-''-'',-'''-'''~'6O"~';;:'';'''::'':':'':';~~~::-'...;d::.=~~O.
''''5'''''',:=::_SLO""_,
4
a (lab Systern)
Ei (lab system)
FIG. 9.4. Difference in time offiigbt for coincident protons scattering from a Si sample (thickness d) at different proton energies, as
labeled. The results are shown ror setups mth (a) the detector perpendicular to the beam IIXlS at distance D 20 mm [ef. Fig. 9.1 (a)J
.,and (i_the detector parn.Jlel to the beam axis at radius R = 20 mm [ef. Fig. 9.l(b)J.
9.I.Z Quantification
with the
N..
with
cr,,= J(dcr)
ill
(9.7)
n" dQ "
192
,.
~1.8Z1 ~
(9.8)
1$0
160
~eB111"""""
.. eI4.Ia~v
. e '"1eJ I.k\'
;..
.... e .. \ulobtv
'a2Ul~
.. e29AL.l4Iv
..... E. So.:UiI(lV
nr
1~--
SO"
193
.. E"1UI1$NeV
~200
Chapter 9
.)
,,
.,
direction
~.
xi@'
/~I
b)
2.0
e=
b
u
:
dded:or Ii
"
t
1$
u
'"'-. 'll
,,
f4 ..
<)
detector i .:
bcaJi,
. ..
.... ,.
1.5
"':
1.0
..
0.5
.1(1)
400
d)
EnellO' CorrelatIon
16
I.
>u
:;:
.E
~
""
I.
16
14
\-'
" ',e" I.
'lle
A'
11
ID
.=
:t:,.
~.
,-j"
,1
L'
. ,it
-.
8
J
12
- [nugy.~t~~I~orn III ~eV'
:~:
~;r.
18
I.
12
12
';
'"
dO'
20
......... 2
'~~ -
I.
. --"~Oo
11)0
~OO
300
.40~
FIG. 9.6. (a) Proton-proton scattering spectra of a sandwich sample that consists of five O.9-jJ.m-thick Mylar foils separated by.
hydrogen-free sheets of60-~m Si. The incoming proton energy is 25 MeV. (b) Correlation of the energies of coincident protons. The
slight bending appears at lower total energy as a result of path-length effects. (c) Depth profile of coincidence events with approximate
depth scale according to Eqs. (9.14) (Section 9.3.1). The depth resolution improves toward the back side ofllie sample because of the
decreasing energy spread and lower path-length effect. (d) Depth profile of integrated peak content that decreases toward the front side
ofllie sample (z "" 0) because of the increasing angular spread of the protons (true event loss).
194
ncO"
TJE
(9.9)
IDs cr p
(9,1), and
O"pp =
r (dO"/ctn) P dD
hw
15
(9.10)
195
/
Chapter 9
boou!
(leclOt) .
(ring)
Semiconductor Design S2 with 48 rings at the p-doped side and 16 sectors at the
D-
51 C.
b)
(I) The hydrogen release cross sections listed in Table 5.6 .in
Chapter 5 for various beam/target material combinations
196
proton microprobe (Fig. 9.9) and even lower for many solids,
such as diamond, where bydrogen is much more stable (Reichart
el al., 2004a).
as
sample
: -"
~f
2D~ilppitig
1
L-l!!!!!."->','
depth profile
.j
!Eo
L.j- __ _
FIG. 9.9. (a) Principle of three-dimensional hydrogen microscopy by proton-proton scattering using a nuclear microprobe. (b)
Example of hydrogen microscopy performed on a mayfly wing using 10 MeV protons.
Lateral spread can be kept low when high-energy protons
Chapter 9
J MeVprptmis .
20 MeV protons
30
sample depth (11m)
50
FIG. 9.10. SRIM2008 simulation (Ziegler el al., 2008) of (a) 3 MeV protons compared to (b) 20 MeV protons penetrating a diamond
sample.
z( 45') = E..m
- Eo + 2./2S",d
2./2S", - s,
(9.14)
(Fig. 9.11). The mean energies of the protons leaving the sample
at angles of 8 and ~ are then given by
E,(e)~(Eo
-Soz)cos
d-z
e-s,-cos 8
and
(9.11)
SID 8
and
Eo,
(9.12)
Ee + Et> is calculated as
Sa
S. (d-z).
E_(e,z,d)~Eo-Soz- --+-.( cose
(9.13)
SlOG
1
,
198
(.18)
SeE) = _ dE = c, In(c,E) ,
dx
E
(9.15)
(8 d,1'fJ\ _{ ~
s..., + S, S..., ] d-z
~""pm;z, ) - 1XIS6 ca;(e+l'fJ) sine sin(e+88) (
).
);1'
(9.19)
"
cos8
sm8
(9.17)
.
d-z
8E.(8;z,d)=2
199
)[
sine
00;8
' + s,-.-, fu,-,,== )' ,(9.21)
(~-,-fu
~
rosS
=e
sm8
sine
Chapter 9
""r
(cose sine
ae"'="12E,sin28+3(d-Z)S(E,) sin'S + cos's
J,r~
(9.22)
Emmas
d~
cosS
.(E'
4e
E')
o cos
-min
c,21n(c,E, cos' 8)
(9.24)
200
;)
,.
.,
.: pfldIl::d~in\N:iI\gi!br~.oos"";wo round)
--::-:-'P3!II~AIO,..(lm#_hy~d!~Jb-ao".) -.
-:G,J
'_,..;.'~_~-::=:_~;;..,dt~_~~._
bijO
<. .. ':
.'iJ
'u
<1':.~~~,,",",,!~I'<""~pr_ilf:.. ~b:y~p.)
-'.;', row_ In-l.urabt_l"tSlIbtIIIIiIli...nd)
".,. .
It.::vn~.spr=l~~.o,b=ndk~aP .. Q}
,Diamonii
,.,
E.:03W
. 8=450
.,
,, .
..
Diamond
Diamond
Eo =3MeV
8 = 35/55
E. = 2.0 MeV
e = 35/55
ma.ximum
thickness
".
..
'00)
maximum
tllicknCis
'"
d)
'l
Iton
. Iron
" 3 MeV
20 MeV
dS
"
FIG. 9.12. The different contributions for depth resolution/energy spread as labeled for a diamond sample (nT = 1.7 x lOll at./cmJ)
using (a) 3 MeV and (b) 20 MeV protons depending on the target thickness d. (c) and (d) show the overall depth resolution only for an
iron sample at different scattering angles (beam divergence Sf), 0 mrad; detector energy resolution per proton, 1%, with a minimum of
20 keV).
.
201
Chapter 9
.:,._-_., . .- .
.~';.';-;.~~'.:~ ";...,v\,,,, '.t" ,b
',.....:~u~
. .'. .
~IA~"'"'
~~17_
_ ..
',;,_".,.,/_,~ ,~
...
"
"
-:+-~ ...
'.
--m,1.I>
.......
__
: J
OM.IU~
.
..
. l
.)
J
,f,
,,
.,. ,
~~~;...,..:-;~.,.....~~
~'"
TltlO
J "lUI
'~j'1'~1\1
~
, , 1 ~
1I
,,/",
.
JOO"'"
JI
O.. Str
:;J\~(t .. ~.~
.., ~ III O.. {O"
31)0"
itridcmJl'OlmClOl\l)'E.~"*'
20
Eo,
30
40
Scattering Angle
b)
50
>
~ 40
50
60
,.,
o+---',''='~;-'--~~Jor-. .,..","",~.,..,4TO~~-OO"T"--';--."'~
.S"'eri!i9 Angie a .
.
. " ,
:'~ ~.
..
202
~ = arctan
-sine
]
[ y=(cos6=em -1)
(9.25)
(9.26)
for each state of total spin J, that is, J = 0 (singlet) or J = 1
(triplet) for proton-proton scattering [Fig. 9.16(b)]. In general,
the scattering cross section has contributions from all spin states.
If scattering is independent of the spin, s, of the parti<:les (no
spin flip) and no spin polarization occurs, the scattering
probability can be obtained by incoherent addition of the
different spin states according to their statistical weights a<;
(9.27)
,,,o'-~~~~~~~c.....
..
_ _ _ _ _':""_ _-,J
O'(lab~)
203
/
Chapter 9
b)
"Pm'
;,,'\' '
, . -.s
1\1,)"
'" . ,......"il4)
f7~l
I
T.2./
T
'/","'"
.
'P m =0 or'S m
'.
's
'",
'.
=0
FIG. 9.16. (a) Two different identical-particle collisions that are indistinguishable. (b) Plot of the four principle spin states of the pp_
scattering event, including the resulting lowest-order scattering states. The antiparallel adjustment of the spins is the only way to fonn,
at a 50% probability, the singlet IS state. Only this singlet contributes significantly to the s[2ttering cross section below E = 30 MeV.
The very different scattering cross sections from S and P
the spins of the protons in the beam and in the target are adjusted
before scattering. Parallel spin orientations result in a nearly zero
count rate, whereas antiparallel adjustment results in count rates
that are twice those. obatined in the unpolarized case: the
probability to obtain S-wave scattering is 50% for the
204
Ecm
=__,.;;E~i',===;
1+
(9.28)
1+~2
2mp c
rem=
E.
--m-2+1.
2mp c
(9.29)
"
1- r =2 tan 2 e
San = arccos [
2
2
l+y= tan
(9.30)
.J
8::: arctan
I-cosO C I l lI_
[ 1+cos9 Yem
cm
em
l+co,O
(9.31)
J3/2 '
(9.32)
(9.33)
Cohen, B.L., Degnan, J.H., and Fink, C.L. (1972), J. Appl. Phys.
43 (I), 19.
dcr
dO
dcr
(6,m) = ill (6=) d[l[6(6,mll or
=
'm
E. (1959), Phys.
dO
dO
ill =Y= [
Blair,
(9.34)
REFERENCES
Allred, J.C, Armstrong, A.H., Bondelid, R.O., and Rosen, L.
(1952), Phys. Rev. 88 (3), 433.
Dujrnic, D., Jaksic, M., Soic, N., Tadic, T., and Bogdanovic, I.
(1996), Nuel. instrum. Methods Bill (1-2), 126.
Green, P.F., and Doyle, B.L. (1995), "Application of ion beam
analysis techniques to polymer science", in Scattering
Methods in Polymer Science (Richards, R.W., ed.), Ellis
Horwood, London, pp. 19'1-218.
GilDther, G., Bethge, K., and Schubert, K.R. (1967). Nucl. Phys.
AlOI (2), 288.
Jannie, N_, Jett. J.H., Detch, J.L., and Hutson, RI.. (1970), Phys.
Rev. Lett. 25 (I), 34.
Bergervoet, 1.R, van Campen, P.C., van der Sanden, W.A., and
de Swart, J.J. (1988), Phys. Rev. C38 (I), 15.
Jannie, N., Jett, 1.H., Detch, 1.1.., and Hutson, RI.. (1971), Phys.
Rev. C3 (I), 10.
Bergervoet, 1.R., van Campen, P.C., Klomp, R.A.M., de Kok, J.L., Rijken, T.A., Stoks, V.G.!', and de Swart, J.J. (1990),
Phys. Rev. C41 (4), 1435.
Bethe, H..A. (1999), Rev. Mod. Phys. 71 (2), S6.
Bethge, K., GilDther, G., Schubert, K.R., and Speth, E. (1967),
Phys. A202 (I), 70.
Jannie, N., and Jett,!.H. (1974), Phys_ Rev. CIO (I), 54.
Jannie, N., Jett, J.H., and Semper, R1. (1974), Phys. Rev. C10
(5), 1748.
z.
205
/
Chapter 9
Science, Amsterdam.
Johnston, L.H., and Young, D.E. (1959), Phys. Rev. 116 (4),
989.
Ki1ruchi, S., Sanada, 1., Suwa, S., Hayashi, I., Nisimura, K., and
Fukunaga, K. (1960), J. Phys. Soc. Jpn. 15 (1), 9.
Knecht, DJ., MesseIt, S., Bernen;, E.D., and Northcliffe, L.C.
(1959), Phys. Rev. 114 (2), 550.
Richter, M., Henning, W., Komer, H.1., MillIer, R.. Rehm, K.E.,
Rother, H.P., Schaller, H., and Spieler, H. (1977), Nuel.
Phys. A278 (1),163.
Schiettekatte, F. (2008), Nud. Instrum. Methods B266, 1880.
Sigmund, P., and Winterhon, K.B. (1974), Nucl. Instrum.
Methods 119, 541.
Sjoland, K.A., Kristiansson, P, Elfinan, M., Mahnqvis~ K.G.,
Pallon, J., Utui, R.J., and Yang, C. (1997), Nuel. Instrum.
Methods B124 (4), 639.
Slobodrian, R.I., Conzet~ H.E., Shield, E., and Tivol, W.F.
(1968), Phys. Rev. 174 (4), 1122.
Slobodrian, R.I., Frois, B., Birchall, I., and Roy, R. (1976),
NucJ.Instrum. Methods 136 (3), 525.
Szihigyi, E., Paszti, F., Arosel, G. (1995), N"eI. Instrum.
Methods BIOO, 103.
Wassmer, H., and Muhry, H. (1973), Helv. Phys. Acta 46 (5),
626.
206
CHAPTER
1m
CHARGED PARTICLE
ACTIVATION ANALYSIS
G. Blondiaux and J.-L. Debrun
CNRS-CERI, Orleans, France
c.J. Maggiore
Los Alamos National Laboratory, Los Alamos, New Mexico, USA
CONTENTS
10.1
10.2
10.3
10.4
207
/
Chapter 10
............................................................................................................................ 226
208
10.1 INTRODUCTION
Charged particle activation analysis (CPAA) can, in principle,
be used to determine almost any element, in solids or liquids,
over a wide range of concentrations (from less than the part~per~
billion level to the percent level). However, taking into account
practical considerations and the availability of nwnerous other
analytical techniques, the important application of CPAA is to
anaI}'2e light elements such as C, N, and 0 at trace levels in
solids. For this application, CPAA is the best reference method
for the following reasons:
dW'I dt =
nm dx - A.7V ,
(10.1)
(10.2)
= 'A!NfJ (becque~els).
(IOJ)
10.2 THEORY
10.2.1 General prinCiple of radioactivation
The sample to be analyzed is "activated" by irradiation with
an ion beam of suitable type and energy. Artificial radioisotopes
such as in EXAMPLE 10.1 are created by nuclear reactions on
the matrix itself and o~ the impurities of Interest Qualitative
analysis is bascd on the identification of the created
radioisotopes, mainly using high~resolution gamma-ray
spectroscopy. Quantitation is based on the correspondence
between the stable isotope present in the sample and the
radioisotope produced by the beam. It depends on the direct
relationship that can be established between the concentration of
the element (isotope) of interest and the counting rate (activity)
of the corresponding radioisotope.
(lOA)
where d stands for decay.
Cross sections are functions of energy, and because the
incident ions lose energy as they traverse the sample, cr can also
be thought of as a function of depth, x. The activity for a thick
target at the end of irradiation is then given by
J,SeE)
A =ni(l-e-,-,,)E a(E) dE
o
(10.5)
209
Chapter 10
ratio of the counting rates in the standard and the unknown, the
concentration in the unknown, Dx. can be calculated if the ratio
standard. The sample and the standard are irradiated at the same
energy with appropriate total charges, Qil and the ratio of the
activities after irradiation is compared. The ratio of the
concentrations is then given by
(I_e-
,),
fa(E) dE
Q"
S,(E)
Q, "fa(E) dE
A,
.. .. , S, (E)
(10.6)
A_-',"__ -'(I-e-"')
__
_ _---'"'-1 Q
(l-e-~'),
A,
Q,
I~.
End of Irradiation
C
Saturation
100
90
70
'-.
';; 60
... ~ 50
<t 40
30
Increase of
activity during
Irradiation
20
Decay of activity
after the end of
Irradiation
A, = C(1-e"'Ij
10
0L..l.-'---'----'--''--1.-'---'----'-L..L-'''=''--..J
o 1 2 3 4 56 1234567
e"'C
= 0 184C
~.
0.0693
(1- e-o.0693xJ)
eO. 06
..
x'
=1 199C
9)' tJ C
.._
n,
!'
O.184C,(l-e"'0691"'),xO.6
-;;- = (1- e
....0693.")
xl 199C
x'
I
A.
O069),'C
(I-e "')
0.0693
(l_e-lJ.0693xll) e
where C, is the total count in the peak during the 3-min. counting
Number
of half-lives (T)
Co =
where ex is the total count in the peak of interest during the 11min. counting time. If the activation of the standard begins at 1:57
and ends at 1:58 with a beam cwrent of 0.01 ,UA, then the
integrated charge is 0.6 ftc. If the counting of the standard begins
at 2: 15 and ends at 2: 18, then the counting time is t., = 3 min., and
the delay time is" = 17 min. Equation (10.7) gives the corrected
count rate for the standard as
~ 80
~
x 132 IJ(.
C
=5.97xl0"'-' <J?.
C,
(10.7)
210
R,(EJ - R,(E,,)
R,(E,)-R,(E m)
(10.10)
R,(EJ
R,(E)
'1(.=--
S(E)
then (S)
(s)
0.575.
Em ~ (5 + 6)12 =
5.5 MeV.
R,(6)
SNYlon(5.5)
=
0.594.
(10.8)
E,
fEcr(E) dE
o
E,
(10.9)
Notice that the higher the incident energy and the higher the
reaction threshold, the smaller the error introduced by these
approx.imations. The error also is smaller when the average
mean atomic numbers of the matrix and the standard are closer.
f cr(E) dE
To a good approximation 1%), Em is matrix-independent,
and the ratio of the integrals, <.R.. is then given by
211
/
Chapter 10
0
0
20
AI V
40
Zr
60
80
Pb
~M'V
9 MeV
6 MeV
~
'#.
<J
12 MeV
9 MeV
10
6 MeV
212
Sampling
ion
Reaction
(MeV)
depth
(p, n)
(p,2n)
(p, pn)
10 -30
Incident
(p,)
d
(d, n)
(d,2n)
(d, p)
(d, a)
3 -20
(I, n)
(1, p)
(I, d)
3 -15
&3He. n)
( He,2n)
(,He, p)
3-20
(a, n)
(a,2n)
(a, 3n)
(a, p)
(a, pn)
((,3n
3He,n
,n
Z+1
p,n
p,y
d,n
3He,np
e<,np
I,n
'He,p
Original
p,pn
n,1
d,p
I,pn
101Jm-2mm
nucleus
Z-1
few
~m
p,"
Z-2
1545
I,)>
t,3 He
p:He
d,,,
p,2p
N-1
n,p
-100 ~m
(3He, a)
a,2n
Z+2
p,ap
N-2
(et, an)
"-~"-~~"-~"'~~T'~--r-;j_,
:2:_ :
:::.
~
,~ 20
.l
~ 10
O~~~~~+~LL~~~~u-~~
10
15
20
25
Energy (MeV)
FlG._IOA. Interference of Fe by the s6pe(p,2n)S5Co reaction
s
above 15 MeV, when analyzing Ni by the '3Ni(p,a/ Co reaction.
213
"'I
Chapter 10
1
'H.,p2n
10.3.3.1 Etching
To demonstrate the importance of surface contamination,
Table IO.2lists tberange of surface concentrations ofC, N, and
o for various samples after chemical etching under the best
conditions, that is, the cleanest surface that can be obtained
(Quaglia et a/., 1976). This minimum surface contamination is
activated at the same time as the trace element in the bulk and
corresponds to a minimum detectable bulk concentration. The
table also includes the equivalent bulk concentration in parts per
million (ppm) that it would represent if not removed. This
equivalent is a function of the nature of the sample, the incident
energy. and the reaction used. It is clear that etching is an
absolute necessity at low levels, at least for C and O.
104
L -__~~______~____~__~
10
20
(MeV)
30
He
Minimum
Equivalent
surface
concentration
bulk
Contaminant
(mlifcm2)
(ppm)
0.05-{J.5
<0.01
0.05-{J.5
0.1-10
<0.01
0.1-10
214
concentration
,.
:\
t
i,
Permanent magnets
Target
Beam-H~~~
Insulator
Faraday cup
FIG. 10.6. Setup for irradiation of samples in the vacuum at low power using direct earn current integration.
Beam
Water outlet
Target
vacuum
Water Inlet
FIG. 10.7. Setup for irradiation of samples at atmospheric pressure, either air or controlled atmosphere, at high power. Irradiation dose
measured by activation of a thin foil such as Havar.
Table 10.3. Semiconductor etches.
Matrix
.Bath composition
(acids are commercial
concentrated acids)
Approximate
etching speed
(I'm/min.)
Si
GaAs
13
loP
215
/
Chapter 10
216
Reaction
(MeV)
Interfering
reactions
uN(d, t) '3 N
14N(d, dn) 1~
1eO(d, an) 13N
9Be(-3He. nl
"
1OS(3He. d)l'C
I'B(3He.l)l1C
14NeHc, a.d) C
"
lliQ(lHe.2o:)l1C
Threshold
!MeV)
4.9
12
8.4
a
2.5
10
6.3
10.4.1 Carbon
The two main reactions are listed in Table 10.4. They allow
the determination of carbon down to the part-per-billion level
(10-') if a post-irradiation radiochemical separation is used.
Nondestructive analysis is possible well below the part-permillion level in matrixes with Z > 30, where it is possible to take
advantage of the Coulomb barrier effect. In the case of
radiochemical separations, the methods mostly used are based on
reducing fusion for DN and trapping N2 on hot titanium or on
oxidizing fusion for lie with absorption of 11eo2 on soda
asbestos. The usual standard is graphite.
10.4.1.1 "qd,n)"N
The 12c(d,n)l~ reaction produces a pure positron emitter
with a half-life of 9.97 min. It is free of interferences below 4.9
MeV, and in practice, it can usually be used up to the 8.4 MeV
threshold for the oxygen reaction. The reasons are that the cross
sections for the deuteron-induced reactions on nitrogen are small
and the concentration of nitrogen in the sample is usually less
than the carbon concentration. The corresponding cross sections
are shown in Appendix 15. Figure A15.l shows the cross section
versus energy for the (d,n) reaction, and Fig. All.2 shows the
ratio of iJN produced from carbon to that produced from
nitrogen as a function of deuteron energy. The thick-target yield
curves for the reaction 00 carbon and the interfering reactions on
Nand 0 are shown in Figs. AI5.3, A1S.4, and AI5.5. These
figures were taken from Krasnov ef aI. (1970). EXAMPLE 10.7
shows the use of this reaction in metals. and EXAMPLE 10.8
shows its use with a radiochemical separation.
C'O,
10.4.2.1 "N(p,n)l'O
As shown in Fig. A15.12 (Appendix 15), the cross section is
not very high (Nozaki and Iwamoto, 1981; Kuan and Risser
1964). The sensitivity is therefore limited to about
ppm/weight. The interesting points are that the reaction is
interference-free up to -22 MeV and that 140 exhibits a
characteristic high-energy gamma ray (2313 keY). It is thus
possible to analyze nitrogen without radiochemical separation
and without complex decay curve analysis of the 511 keY peak
if the matrix does not activate too much. The drawback of this
reaction is its high threshold energy of 6.3 MeV.
10.4.1.2 "q'He,a)I1C
This reaction also produces a pure positron emitter, but with a
20.3-mm. half-life. Several measured experimental cross-section
curves are shown in Fig. 10.10 (Lamb, 1969). The differences
between the curves are typical of total cross-section
measurements, but they are not important for activation analysis
using standards. Beryllium and boron interfere at all energies.
The cross-section curves for these interferences are given in
Figs. All.7 and AIl.8 (Appendix Il). The thicktarget yield
cwves for the reaction on carbon and the interfering reactioDs on
Be and B are given in Figs. AI5.9, AI45.1O, and AIl.I1
(Appendix Il).
10.4.2.2 1'N(p,a)llC
If the s'ample does not contain too much boron, this is the
most suitable reaction for nitrogen analysis at levels of I ppm or
greater. Figure A15.!3 (Appendix Il) shows the crosssection
curve for the 1"N(p.a)lIC reaction (Jacobs et al., 1974). A
detailed curve up to 7 Me V was redetermined by Kohl et al.
(1990). Figure AIl.14 (Appendix 15) shows the cross section
for the interfering "B(P,n) l1 C reaction (Anders, 1981). The
thicktarget yield curves for the 1'N(p,a)I1 C and the IIB(p,n)IIC
reaction are shown in Figs. A15.15 and A15.l6 (Appendix 15)
(Krasnov e/ al., 1970). The usc of this reaction is demonstrated
in EXAMPLE 10.9.
I
...L
10.4.2.3 l'N(d,n)l'O
This reaction is interference free up to -11 MeV. Although it
is intrinsically sensitive enough to reach the part-per-billion
level, it cannot reach this level in practice because ISO must be
radiocbemically separated. which takes time. and a large amount
of activity is lost because of the short half-life of 1'0 (122 5).
The best use of this reaction is probably for concentrations of
217
/
Chapter 10
10.4.2.4 "N(~p)''N
for access to a beam of tritons, but it bas very good potential for
the analysis of nitrogen. It is the eqwvalent of the analysis of
oxygen with 14 MeV neutrons [160(n.p)1~. except that one can
obtain a much higher flux on the sample and use of 3 MeV
tritons results in low matrix activity compared to nitrogen.
Preliminary tests (unpublished results from the group in Orleans)
7.1 s).
References for actual analysis of nitrogen in various matrixes
are listed in Table 10.7.
10.43 Oxygen
The main reactions for the analysis of oxygen are listed in
Table 10.8.
10.43.1 "O(p,a)"N
This reaction is a pure positron emitter with a lO-min. halflife. The cross section is shown in Fig. A15.19 (Appendix IS)
(Furukawa and Tanaka, 1960). The thick-target yield curve is
shown in Fig. A15.20 (Appendix IS) (Krasnov ef aI., 1970).
This reaction is not selective at any energy and should be used
only in selected cases (high fluorine concentration in the sample
and low carbon concentration). The importance of interfering
reactions is shown in Appendix IS in Fig. A15.21 for carbon and
in Fig. A15.1S for nitrogen.
ScIntillation detector
Scintillation detector
"t Sample
218
Matrix
Reference
AI
12c(d. n)13N
s;
12C(3He,o.)11C
et 81., 1980
Engelmann etaf., 1971b; Bottger
et aI., 1980; Martin and Haas, 1972;
Sann! at af., 1984; Bonger at aI.,
102~~~~~-L~~~~~
0.0
5.0
10.0
15.0
1975
20.0
Time (minutes)
FIG. 10.9. Semilogarithmic plot of activity versus time for a
radionllclide with a 3-min. half-life.
F.
12C(ct. n) \50
N;
12C(d, n)'3N
Zr
12(;(0, n)'::lN
Zircalloy
12C(d, n)13N
Nb
12C(d, n) 13N
Me
Ta
12C(a, an)I1C
Au
12C(d, n)'3N
GaAs
12C(d, n)13N
-~- ~
Ga AI As
10
15.20
25
30
35
12C(d, n)1::lN
219
(MeV)
Interfering
reactlon(s)
Threshold
Reaction
14N(p. n) 140
6.3
160(p, t) 140
21.7
14N(p, a)l'C
3.1
(MeV)
tlB(p, n)lIC
''c(p. d)"C
17.9
14N(d, n) 150
r'O(d,l)''O.
10.6
14N(I, p) 16 N
13C{t. y)'6N
170(1. a)16N
0
0
/
Chapter 10
10.4.3.2 16 0('He,p)"F
Reaellon used
'''N(p. a)"C
51
14N(p, a)l1e
li
14N(p, n) 140
AJ..TI
14N(P. n) '40
Fe
14N(d.
n)'SO
NI . _ _~~(P. a):'C.. __
Zr
Zircatkly
Nb
Ta
W
Mayolet
st al., 1972
14N<p.o.)lle
'''N(d, n}'SO
14N(ct, an)13N
14N(a, an)13N
1~(p,
n)140
''N(p.2aJ'Be
Mo
Reference
14N(d, n) 15 0
14N(p,a)"C
10.4.3.3 "O(ll,d)"F
Activation with alpha particles has been used in the past
when 3He was rare and expensive. The use' of alpha particles
should be avoided, however, because of the high value of the
fluorine interference and because thc high threshold values cause
high matrix activation. The cross-section curve (Nozaki et al.,
19740) and the thick-target yield curve (Krasnov el al., 1970) are
shown in Figs. AIS.25 and A15.26 (Appendix 15), respectively.
The thick-target yleld curve for the interference of fluorine is
shown in Fig. AIS.27 (Appendix IS).
10.4.3.4 "O(~n)"F
uN(p, n) 140
HN(p,o)'1e
14 N(p,
a)"'C
Reaction
Threshold
(MeV)
Threshold
(MeV)
''C(p.ll'3N
'"c(p, n) '3N
14N(P. d) 13N
0
3.2
8.9
19F('He, a) "F
3.8
0.4
.. 5.5
'50(3He, pi I8F
Intelfering
reactions
I6O(3He, nl '8N.
+e.C-7 'B F
'50(a, d)'8F
20.4
"Nla, nl 18f
8.1
23.2
19F(a, 1lIl)"F
12.6
'50(a,2n)I8Ne
p. "F
29.7
'"Na(a, 2an)'SF
25
2ONe(t, an)'SF
19f(t. tn)'SF
.~:-
'-'
.~.
;'.,t.
:-'r",
4
12.1
."~;.,
~~
~,t
.,'~$'
..~~
220
'.~~
reasons for this arc thc high intrinsic sensitivity at low energy
and the possibility of removing surface contamination. Also note
that accidental channeling, as in other methods, can be a source
of Jarge errors in charged particle activation analysis. This has
been demonstrated up to factors of 2 in the work of Hanna
Bakraji e/ al. (199Ia).
EXAMPLES 10.11 and 10.12 demonstrate that charged
particle activation combined with channeling can be used as a
tool to study the lattice position of some light elements at the
sub-part-per-million level, probably down to 0.1 ppm/weight.
The interest of the method is that it is applicable to a number of
as-gro'WD. pure materials, not especially doped for the pmpose of
characterization. Also, the elements that can be investigated-B,
C, and O-are most often the major residual impurities. of the
semiconductor materials (8 for Czochralski crystals prepared
under liquid B20 3). A considerable amount of work remains to
be done to fully understand the reiC'ltionship between the light
elements cited above and the properties of semiconductors.
Charged particle activation is an interesting tool for these typcs
of studies,
160(t,n)18F reaction.
The material studied was unalloyed steel. and the reaction used was
days, so the loss of activity due to the natural decay of the radioisotope
is small compared to the loss of activity due to corrosion. The
irradiation energy was 9 MeV; the corresponding activated depth was
40 JlOl, as shown in Fig. 10.13. The sensitivity was 2 x to-l l11m./year,
for a lOO-h test. Examples of results are given in Figs. 10.14 and 10.15.
The variation of the corrosion-erosion with the velocity and the
temperature of the cooling water and the effects of alloying with a small
readily
were
demonstrated.
percentage
of
chromium
221
Chapter 10
Reaetlon used
Reference
Na
AI
Matrix
Reaction used
Reference
Engelmann 1970a
CuZn
''o('He, p)'6F
160(3He, p)18F
ZnTe
160(3He, p)18F
GaP
160(3He, p)18f
Kim, 1971
lSO(t, n)ISF
GaAs
160(3He, p)18f
160(1, n)16F
GaAIAs
''orHa, p)'8F
Hanna Bakraji
Ge
160fHe, p)1BF
1978
lB()(p, n)leF,
Vialane, 1973
''o(p, a)13N,
''0(0., pn) 18F
et al.,
AIMg
160(3He, p) ls F
AI SI
''orHe, p)18F
Si
160(3He, p)1Sf
160(t, n)18F
Zr
160(a. pn)18F
Mo
16o(t, n)18f
1I30eHe, p)18f
1969
160(a., pn)18F
1991a
1968
at af., 1974b
160(a, pn)18F
lSO(p, n)laF
Ag
160(0:, pn)18F
lSO(t, n) 1sF
InP
160(1, n) 18 F
''o(,He, p)18F
HI
'60(0:, pn)'8F
'SO(t, n)18F
Ta
160rHe, p)18F
Cr
''orHe, p)18F
160(3tie, p) 18 F
Fa
160{cr., pn)tBF
160(a, pn}16F
Revel and
Co
160(0.. pn)18F
Pb
160(a, pn)16F
Ni
U'O(lHe. p)18F
PI
160(3He, p)18F
Au
16 0((1, pn}lBF
160{p, n)H'F
l60(d, n)17F
SiN
Cu
''o('He, p)18F
160(a, pn)18F
Fluoride
glasses
Diamond
160rHe,
p),aF
Albert.
1968
.,:1
~"ij
'.,:1
.'7-
?i{
,;.~
":1
;i
222
Matrix
Reference
Chemical
separat.ion
Na
Y"
AI
Y"
no
no
yes
no
Co
no
Ni
yes
Cu
Zn
Y"
Nb
no
Si
Fe
Chemical
Matrix
Biological
Tb
Dy
Ho
no
no
no
1976
Ta
no
yes
yes
I,
no
yes
no
1976
Glass
yes
ye,
Landsberger, 1984
no
y.s
Zr F4
Rh
no
Coal
Ag
no
Minerals
Environmental
yes
InP
no
p,
no
Airborne
particulate
matter
Fly ash
1976
separatlon
1976
Au
Cd Te
Referenee
223
no
~ ..
no
ye,
no
Chapter 10
'.,
e: 0.5
(J
IB 52% AI
032% AJ
o 25% AJ
,
.. m
I 987b) .
'
--~~~~
-----------------
(at/e""')
Concentration (ppmlweight)
1.2x10 16
SSx10.12
S2x10 12
S6;.; 10 12
S 10 13
2 x 10 14
S4x10 13
2>: 10 13
:S 10'2
B
C
FIG. 10.1 I. Variation of the CIR ratio of the channeled yield (C)
of 13N in the (100) direction to the yield of 13N in a random
Ti
Cr
F.
Ni
Cu
1986).
G.
Zr
10 16
3x 10U
7 x 10 1:3
S 6 x 10 12
2 x 10 13
2 x 10 14
:::; 4 x 10 1J
10 13
S 10'2
metallurgy.
C3
~4
(trtlcm")
6 IIh
OArslneflow=1411h
OArsine flow 16 Vh
C:C
ContaminatecJ
sample*
1.0 ,..--,--,----,-----,----,
0.8
2
~0.6
.;;
O~LW~~~~~~~~~~
10
20
30
40
50
60
c(
224
0.4
0.2
10
20
30
Depth (!lm)
40
50
>Ope(p,n)S6Co reaction.
1.0
.!!
ca
One phase
Twophase
.;;;0
0~
0.5
..
0
CI)
>
ca
Gi
It
0.0
2
1
Chromium content (%)
FIG. 10.14. Corrosion rate for unalloyed steel for different speeds of the circulating waler as a function of water temperature
(Bouchacourt el aI., 1989).
.
~
.c
8
<) V:1 mls
0
0
0
<:)
0
6 C-
V:3 mts
-E
CI)
V:4 rnIs
'"
.;;;
2 C-
t-~~
/~O...______0--o-------o~
t.
e
~
()
o~
),.~t.
______"_________
~,.~
~
/.0:
A V:2 mls
80
100
120
140
160
180
200
:----
0,
220
_L
240
260
Temperature (Oe)
FIG. 10.15. Relative corrosion rate versus the chromium content of steel for two experimental conditions (Bouchacourt et aI., 1989).
225
---------_.
--_._.
..~
-,
Chapter 10
10_5 CONCLUSION
Charged particle activation can, in principle, be used to
analyze ahnost any element at the part-per-billion (10-') leveL
H, He, and Si are notable exceptions. Except for C, N. and 0 at
very low levels, there are many excellent methods to achieve
trace analysis: inductively coupled plasma (ICP), rcp with mass
spectrometry (lCP-MS), secondary ion mass spectrometry
(SIMS), and glow discharge mass spectrometry (GDMS), among
H, Be, C, N, 0, F, AI, Si, P, S, el, MD, Co, Nb, Rh, Ag, In, Pr,
n, 127.
elements that are very sensitive in one method are less sensitive
in the other. Activation with other particles such as 3He and 4He
bas been very seldom used and is of little interest in practice.
REFERENCES
Chaudri, M.A., Bums, G., Reen, E., Rouse, J.L., and Spicer,
8.M. (1976), in Proceedings of the International Conference on
H.~J.,
Anden;, B. (1981),
G.,
,l'\
and
.,,
B45,105.
Benaben, P., Barrandon. J.N., and Debrun, J.L (1975), Anal.
Chirn. Acta 78, 129.
Benaben, P., Barrandon, IN., and Debrun, 1.L. (1979),
Radiochem. Radiaanal. Lett. 37,241.
~;
,.
'.~
Defebve,
,
P.,
Lerch,
P.,
and
Vandecasteele,
C.
(1976),
';I
f
,:l
1l'
'1./,0<W
:f?
~#
226
Ishii, K., VaUadon, M., Sastri, C.S., and Debrun, J.L. (1978b),
Nucl. Insrrum. Methods 153, 503.
32,117.
M., and
London.
Lett. 6, 189
Ku"", H.M., and Risser, 1.R. (1964), Nucl. Pltys. 51, 518.
c., Blondiaux,
G.,
c.,
Ali, M., Mathez, E., and Misdaq, M.A. (1989), Nud. Instrum.
Methods B4~I, 1193.
Hanna Bakraji, E., Ducouret, E., Blonrliaux, G., and Debrun, J.L.
(1991a),Nucl.lnstrum. MethodsB56-57, 819.
Huber, A., Morillo~ G., Linb, N.T., Debrun, J.L., and Valladon,
M. (1978), Nud. Instrum. Methods 149, 543.
Huber, A., Linh, N.T., Valladon, M., Debrun, J.L., Martin, G.M.,
Mitouneau, A., and Mircea, A. (1979),1. App/. Rhys. 50,4022.
227
/
Chapter 10
L827.
Shibata, S., Tanaka, S., Suzuki, T., Umezawa, H., Lo, J.G., and
Yeh, 8.J. (1979),1nt. J. Appl. Radiat. Isotopes 30,563.
Analysis 3, 171.
z.
Nozaki, T., Iwamoto, M., Usami, K., Mukai, K., and Hiraiwa, A.
(1979), J. Radioanal. Chern. 52, 449.
Nozaki, T.M., and Iwamoto, M. (1981), Radiochim. Acta 29,57.
Pars" B., and Markowitz, 8.S. (1974), Anal. Chern. 46, 186.
Peti~ 1. (1978), Mern. Sci. Rev. Metall. 75,395.
Quaglia, L., Weber, G., David, D., Van Audenhove, J., and
Pauwels, I. (1976), Ewisotop Report ITE No. 90.
Revel, G., Da Cunha Belo, M., Linck, I., and Kraw, L. (1977),
228
,
'-'1
r
I
I
Vandecasteele,
Chim.
229
CHAPTER
PARTICLE-INDUCED
X-RAY EMISSION: PIXE
J. L. Campbell
University of Guelph, Guelph, Canada
CONTENTS
11.1 INTRODUCTION ........................................................................................................ 233
11.2 FUNDAMENTALS ...................................................................................................... 234
11.2.1 Intensity--concentration relationship ............................................................... 234
11.2.2 Bremsstrahlung background ............................................................................ 234
11.2.3 PIXE database ....,............................................................................................... 235
11.3 SAMPLE TYPES AND ANALYTICAL APPROACHES:.:: ....::.:.......................;.~.236 ....
11.3.1 Thin-sample approximation .............................................................................. 236
11.3.2 Thick-sample: Known matrix ........................................................................... 236
11.3.3 Intermediate sample; Known matrix ............................................................... 236
11.3.4 Intermediate sample: Unknown matrix ........................................................... 236
11.3.5 Thick samples: Matrix concentrations unknown ............................................236
11.3.6 Layered samples ................................................................................................. 237
11.3.7 Spectrum fitting ................................................................................................. 237
11.4 EXPERIMENTAL ASPECTS .................................................................................... 237
11.4.1 Accelerator and Ion beam .................................................................................237
11.4.2 X-ray detector ..................................................................................................... 237
11.4.3 On-demand beam deflection .............................................................................238
11.4.4 X-ray absorbing fIlters ...................................................................................... 238
11.5 SPECTRUM FITTING AND CONVERSION OF PEAK
AREA TO CONCENTRATIOtl'S ................................................................................ 238
11.5.1 Spectrum fitting ...........................................................................................:..... 238
231
I
Chapter 11
..
232
-;';,
~.
11.1 INTRODUCTION
Particle-induced X-ray emission (most often effected with
proton beams) is based on the ejection of inner-shell electrons
from target atoms and the spectroscopy of the subsequently
emitted X-rays. Figure 11.1 shows the many emission lines that
i
~epo~.~ib~l~e~~~~~~~~~~~~~~~~~~
i
N
lt~5
K ,._-
l2Serlt15
(11.~."S'''',.
'''')
USOrl=
(I.u"l'O:<I'~~1.5'1J.t.
.~)
- - - - -.. ,--------
KSer\elll
(~nt.flt,~,
etc)
irl
100
tt
. I"
series
...
."' .. .,...
"
iIIl
jJ
OJ
.~
"
Go
t};
"'
io
OJ
0.
... ..
Oll.
,o' """
I!1l\
" ''.1J
V .OJ
lib
.. .,
jo
. i...,
"""'(a'i)
"
233
Chapter 11
234
11.2 FUNDAMENTALS
11.2.1Intensity--concentration relationship
For the geometry shown in Fig. 113, with a homogeneous
sarople thick enough to stop the ion beam and with the justifiable
neglect of straggling (given that the bulk of the X-ray production
arises from the earUest part of the proton track), the intensity or
yield Y(Z) of the principal characteristic X.ray (Koi or La,) of
element Z (atomic mass Ai) having concentrations Cz is
_ N.. coL b zt;;(OJ47t) N C 'fO"iE)TI(E) dE
Y(Z)_
'
S (E)
At
Eo
M
(11.1)
Tz(E)=expH*ls:~:ls~~)l
(11.2)
S"p"e<imen
updated on Ziegler's Web site (http://www.srim.org). appear to .be accurate at the level of a few percent and are well-suited to
. inclusion in PIXE analysis codes. For protons incident on the K
shell, the tabulated ECPSSR-DHS (energyloss Coulomb-
Interacthm Point
a'
'on
deflection perturbed-stationary-state relativistic Dirac-HartreeSlater) theoretical ionization cross sections of Chen and
Crasemaun (1985, 1989) and the "reference" cross sect ions of
Paul and Sacher (1989) are in agreement and appear to be
accurate to within a few percent. Except at very low atomic
DWDbers, theoretical K-shell fluorescence yields (HubbeU et aJ.,
1994) and X-ray branching frnctions (Scofield, I 974a, 1974b)
agree closely with experiment. The most modcm tabulation of
X-ray energies is that of the National Institute of Standards and
Technology (NlST) (Deslattes et 01., 2005). There are many
tabulations and calculations of X-ray attenuation coefficients;
NISI', XCOM code (Berger et al., 2009), based on theoretical
calculations. is convenient and, with some exceptions, again
provides accuracy at the level of a few percent. One ex.ception
arises when ali X-ray energy value lies just above the K
absorption edge of a major constituent. An example is analysis
of FeNi standard alloys where the Ni K X-ray is strongly
absorbed by iron; analysis based on Eq. (11.1) gives an
erroneous re~lUlt even if a secondary fluorescence correction is
included (CampbeU et ai., 1993). When the XCOM attenuation
coefficients for Ni K X-rays in Fe are replaced by near-edge
measured values from synchrotron experiments, the analysis
agrees closely with the certified concentrations.
XRay
x- ray energy
235
Chapter 11
where Y I is the theoretical X-ray yield per unit solid angle, per
unit concentration, and per unit beam charge. We have combined
the detector solid angle and any calibration factor for the
x-
(11.3)
236
aI., 2006).
11.3.6 Layered samples
The methods described in the preceding sections can be
extended to films on substrates and, indeed, to specimens
Si(Li)~
237
IChapter 11
u.s
238
SRM
1155
CI296
1\57
1261A
1155
C1296
The various examples described in the remainder of this
chapter were chosen to go beyond the conventional trace1160
element analyses in thin or thick samples that dominate PIXE
and to illustrate some of the more specialized aspects. Several _
examples, both conventional and less so, are provided to
illustrate the capabilities of the GUPIX package.
1155
11.5.2 Standardization
CI296
1160
R(Fe)
H(Ni)
H(Mo)
0.0853
. 0.0003
0.0843
0.0003
0.0838
0.0003
0.0842
0.0002
Absorber thickness ~ 350 fUll
0.0853
0.0007
0.0847
0.0007
0.0850
0.0857
0.0012 ____ 0.0001
Absorber thickness
0.0843
0.0042
0.0857
0.0023
0.0835
0.0078
0.0847
0.00t8
0.0838
0.0005
0.0849
0.00l3
0.0876
0.001I
0.084-3
0.0007
0.0847
0.0055
0.0854
0.0002 -
=500 flm
0.0830
0.0022
0.0856
0.0003
0.0850
0.0007
0.0857
0.0007
0.0858
0.0003
Care must be taken with this approach at the extremes of Xray euergy. At higher X-ray energies, the detector's intrinsic
efficiency faUs below 1.00, and the approach is vulnerable to
error in detector thickness and to inadequate treatment of
Compton interactions in the efficiency model. At energies below
about 4 keV, entrance windows, contacts, ice buildup. the escape
of photo- and Auger electrons, and the poorly understood effects
referred to as incomplete cbarge collection (ICC) all conspire to
render calculated intrinsic efficiencies inaccurate. If H is
determined over a wide energy range, deviations from constancy
can occur at these extremes; database imperfections might also
contribute and have been shown (Campbell et aI., 2000) to cause
differences between H values measured u.-,ing K and LX-rays.
~PIXE
239
/
Chapter 11
0.02 t-'li----1r--t--r-'--'l..-r--
::J: 0.015 H~
0.01
t-+"--t--if--t-"'+--t---+--+---I
0.005
E[keV]
FIG. \l.5. H values measured by Gama e/ al. (2001) using many standards. The dashed curve w", obtained with the nominal
geometric parameters, and the solid onc was obtained from a nonlinear least-squares optimization procedure. Reprinted with
pennission from Elsevier.
240
100
Ep 2.4 MeV
660 ~m Mylar abS<lrbe r
o DL : dry 4mgJcm2 ]
DL : acid digest
10D ~C
ash 1mg/cm2
Cl
~ DL:
OJ
::i
~
:!::
10
E
:::i
l:
~CD
...
Q)
0
0
LOD plots for the latter (Fig. 11.6) were valuable in showing
how different approaches to the methodology of target
preparation (drying, ashing, digestion) could improve LODs.
0.1
K Ti Mn Ni Ga Se Rb
Mo
Cd
Pb
Element
FIG. 11.6. Detection limits for biological samples prepared by
various methods. Reprinted from Maenhaut et al. (1984) with
pennission from Elsevier.
both estimates of statistical error, and the actual LOD. The last
of these is especially useful when the element's X-rays are
absent from the spectrum, as an upper limit on its cODcentration
can be a valuable piece of information.
supported on a very thin polymer film which shows no Xrays. The stoichiometry is known tobe PtlCn. Only the Pt L
X-rays are visible in the spectrum. The PtC is treated by
GUPIX as a layered specimen, in the special limit of just a
single layer. And in that layer the C is obviously an invisible
DEPTH DISTRIBUTIONS
There are two reasons that PIXE does not lend itself to
accurate depth profiling in the way that NRA and RES do. First,
the observed proton-induced X-ray yield is integrated over a
wide range of depths within the sample and, therefore, does DOt
reflect any specific depth. Second, the characteristic depth
(defined here as the depth within which 90% of observed X-rays
originate) is element-dependent. However, PIXE can provide
accurate measurements of thickness and composition of single aT
multiple homogeneous thin layers, provided that the speclrum
processing code takes full account of secondary fluorescence
bernreen the layers.
241
I
Chapter 11
values obtained by the two iteration methods were 8.47 and 8.30
11.7 M1CRO-PIXE
All of the foregoing discussion applies equally to
conventional broadbeam PIXE or to microbeam PIXE. The
majority of current PIXE work is of the microbeam variety (}lPIXE), and in this context, the attachment of a high-power
optical microscope to analysis chambers becomes important for
selection of the region to be analyzed and for observation of its
structure. High optical magnification is desirable, which
necessitates a small working distance for the objective and thus
.encowages a design that has the objective within the chamber
vacuum. A standard approach using a beam incident
perpendicular to the sample has been to pass the bearp. through a
45 mirror and to have the objective close to the mirror (i.e., the
microscope is at 90 with respect to the beam); this
accommodates a detector at 45, which is favorable in terms of
minimizing bremsstrahlung background. To achieve greater
optical magnification (300x), Perujo et al. (1988) installed a
high-power microscope within the vacuum, with its axis along
the sample no1'll'Ul1; this forced a design in which the beam
arrived at 45 to the normal and the X-rai takeoff angle was also
45. A more recent design (Ryan et 01., 1999) made the first use
of catadioptric optics (i.e., an image-forming optical system
including both lenses and curved mirrors) by introducing a
Questar reflecting microscope with a 15-cm workitig distance
viewing a mirror that transmits the microbeam through a small
aperture.
incident angle
Weber el al. (2005) studied the behavior of the X-ray yields
of iron, gold, and zinc in an archaeological artifact as the
incident angle of their proton beam was varied from 0 to
70. The iron was presumed to be in pigments at the surface;
the gold was presumed to be part of a subsurface gilding
layer; zinc was not expected to be present. The Fe Ka
intenSity increased continuously, suggesting that the iron
was located on the surface~ the Au LP intensity increased
and then decreased, confmning iliat it lay deeper, and the Zn
Ka. decreased continuously, suggesting that zinc was present
below the successive pigment and gilding layers. The
unexpected zinc was traced to a restorative procedure.
242
,.
243
Chapter 11
REFERENCES
Beck, L (2005), X-ray Speelrom. 34, 393.
Hanison, J.F., and Eldred, R.A. (1973), Adv. X-ray Anal. 17,
560.
Hubbell, J.H., Treban, P.N., Singh, N., Chand, B., Mehta, D.,
Garg, M.L., Singh, S., and Puri, S. (1994), J. Phys. Chern. Ref
Data 23, 339.
Campbell, I.L. (2003), At. Data Nud. Data Tables 85, 291.
Campbell, J.L (2009), At. Data Nud. Data Tables 95, liS.
Chichester, UK
Liu, Z., and Cipolla, S.1. (1996) Comput. Phys. Commun. 97,
31~
.
Doyle, B.L., Walsh, D.S., Kotula, P.G., Rossi, P., Schulein, T.,
and Rohde, M. (2005), X-Ray Speclrom. 34, 279.
Fedi, M.E., Chiara, M., Giuntini, L., Lucarelli, F., and Mando,
P.A. (2002) Nucl. Instrum. Methods B189, 56.
244
Orlic, I., Sow, CH., and Tang, S.M. (1994), At. Data Nucl. Data
Tables 56, 159.
Papp, T., Maxwell, lA., Papp, A., Nejedly, Z., and Campbell,
J.L. (2004), Nud. Instrum. Methods B219-220, 503.
Paul, H., and Sacher, J. (1989), At Data Nuel. Data Tables 42,
105.
Ryan, C.G., Cousens, D.R., Sie, S.H., Griffin, W.L, and Suter,
O.F. (1990), Nucl. Instrum. Methods B47, 55.
Weber, G., Martinot, L., Strivay, D., Gamir, H.P., and George,
P. (2005), X-ray Spectrom. 34,297.
292.
c.,
245
CHAPTER
lEI
ION CHANNELING
M. L. Swanson
University of North Carolina, Chapel Hill, North Carolina, USA
L. Shao
Texas A&M University, College Station, Texas, USA
CONTENTS
12.1
INTRODUCTION ..................................:.................................................................249
247
Chapter 12
...-=-'H.-.,.
..,...,.___ ~
.-..
248
Ion Channeling
12.1 INTRODUCTION
Ion channeling is a powerful tool for the characterization of
crystalline solids. The method is based on the so-<:alled
channeling effect where, if a well-collimated beam of fast
charged particles impinges on a crystal in a direction aligned
with the crystal axis, the trajectories of incident particles are
excluded from a small region around the rows of atoms. The
radii of these forbidden zones are on the order of the ThomasFermi screening dIStance (-o.l A). Consequently, the yield of
close-encounter collisions is significantly lower than the normal
yield obtained under non-channeling conditions. If the crystal
contains defects involving atoms that are displaced from the
atomic TOWS, close-encounter collisions with these displacements
will increase the yield. The yield can be used to precisely
calculate the defect densities in imperfect crystals.
oI
0
I
~d-{
d_
249
(12.1)
/
Chapter 12
where p is the distance of the ion from the string of atoms, x is the
distance traveled along the string, and d is the spacing between
atom, in the string (Fig. 12.1). If the string consists of different
atomic species, the potentials for each atom are different, and an
average potential is used.
(.,
..
f.(p/a)
= L a,K,(-Il,p/a),
(12.6)
i_I
where ai and Pi are the same eonstants as used in Eq. (12.3) and
~ is a modified Bessel function. Note that the value of fMIn: is
given by FRS of Fig. A17.2 in Appendix 17 (EXAMPLE 12.1).
EXAMPLE 12.1. Calculation of ft and fM for "He+ incident
along a (I I 0) channel in Si.
Thomao:;-Fermi fonn
Z Z e'
VCr) ~ - '-'-<p(r/a) ,
(12.2)
r
where Zl and 22 are the atomic numbers of the ion and an atom
in the string, respectively; e is the electronic charge; r is the
nuclear separation distance; ([)(r/a) is the Thomas-Fermi screening
function; and a is the screening distance. The use of such a
potential is appropriate for the intennediate impact parameters
involved in ion channeling, as discussed by Lindbard el al. (1968).
It should he noted that such a two-body potential is independent
of the ion velocity; that is, inelastic scattering is neglected.
An analytical approximation for q>(r/a), developed by Moliere
(1947), is often used in calculating continuum potentials
J
cp(r/a) =
La,
,., exp(-Il;,/a),
(12.3)
(12.4)
+ I].
(12.7)
( ')'"
2Z,Z,e
Ed
(12.8)
of atoms.
EXAMPLE 12.2. Calculation of 'Vl for 2 MeV 4He+
VT(p)
2Z,Z,e' [(pIa).
d
.1
250
Ion Channeling
Surf-CIce
Surface
o
000
000
I
0000
0000000
000
Ion
00
0 0
0 0
O~."~i'-~'''''
0
o 0 0
000
00.0000
0.0
~
000000
B. Dechanneling
from point
c
X
o ~A==::::::::=;~_
defects
Figure 12.3 illustrates the effects of crystalline
BaekseattQfOd
imperfections on channeling {B0gh, 1968; Gemmell, 1974;
energy
Swanson, 1982; Feldman et al., 1982). For channeled ions in a
FIG. 12.3. Schematic diagram illustrating ion scattering from
nearly perfect crystal [Fig. 12.3(a)}, c1ose-encounter
defects: (a) little scattering for shallow depths in a nearly perfect
probabilities are small, and backscattering yields are
crySlal, (b) dechanneling by multiple scattering from pOinl
significantly reduced. If the crystal contains defects, even
defects, and (c) direct backscattering from an amorphous surface
slightly displaced. from atomic rows, the string potential is
layer. (d) Nonnalizcd yields contributed from the above three
disturbed, and ions will be dechanneled [Fig. 12.3(b)J. The
scattering mechanisms.
dechanneled ions 'begin to see lattice atoms similarly to
nonchanneled ions. If a near-surface layer of the crystal is
The nonnalized yield, x.~-), for alignment along an axial
completely disordered (amorphized), as shown in Fig. 12.3(c),
channel (uvw), according to the continuum model (Lindhard,
backscattering yields are significantly enhanced, and the close(965), is given simply by the fraction of the channel area that is
encounter probability becomes equal to that obtained in the
blocked by the vibrating strings of atoms (EXAMPLE 12.4), or
non-channeling mode. Figure 12.3(d) shows the normalized
yields contributed from the above three scattering mechanisms.
NormalizatioD was achieved by dividing the aligned yields by -'"-...... X~uvw} = Nd1t(U 2 +a 2 ) ,
(12.9)
2
the random yield.
where N is the atomic density; a is the screening length dermed
in Eq. (12.4); u, is the lwo-dimensioDal vibration amplitude, u, =
1.414u\; and Ut is the one-dimensional vibration amplitude. This
result agrees remarkably well with experimental data for
channeling of protons in a material such as tungsten where U2 is
small, but it is generally less than the experimental values in
x~....... ),X~uvw) are the normalized yields for host and solute
~O.l 0
X~hkl) ,X!hkI f are the normalized yields for host and solute
atoms, respectively, for alignment along
channels.
XbD is the dechanneled fraction of ions.
{hkl} planar
251
/
Chapter 12
where,+, is the initial incident angle when the ions strike the crystaL
Pin is distance between the atomic row and the initial striking
(12.10)
Using the results of many numerical simulations, Barrett (1971)
position.
(12.11)
where ~ = 0.85u,/a and Fd~) is given in Fig. A17.2 in Appendix
(12.13)
The ions are excluded from an area npl close to the string. It
follows th~t the ion flux is enhanced near the center of a channel.
Under these conditions, a simple estimate of the flux distribution
within a channel can be made for a given value of '1', using
potential contours of the type shown in Fig. 12.5. For ' = 0, the
normalized flux, F i, at the jth potential contour is given
approximately by (EXAMPLE 12.4)
A,
Fj = JdAlA=ln(A,IA,j,
(12.14)
equl-potentlal como.urs
1.4
. '00
tOO
50
2D
10
5'----8-6-4-20246
Potential
minimum
12.2.2 FJux distribution and trajectories of channeled ions
Because of the steering action of cbanneling, ions are directed
toward the center of a channel, resulting in an enhanced ion flux
there. In order to calculate ion trajectories and, thus, the spatial
distributions of channeled ions within a given channel, it is
usually assumed that the transverse component, Eb of the ion
energy is conserved (energy losses are neglected). Thus. the ion
transverse energy remains constant and is given by
(12.12)
Center
of
channel
0.036 LLJ..l...L.......l_'-....L_......l_L..L-"o
0.10.512
5 10
20
50100200
.,
i
252
.4
.(
,,'
.li
Ion Channeling
(12.15)
where n is the number of atomic strings bordering the channel,
pch is the distance from a string to the center of the channel, C is
a constant taken as 3 1n , a is the screening distance, and 'VI is the
Linhard characteristic angle (Van Vliet, 1971).
,I
0.05 _?
---
12
<110> Channel
Equi-flux contours
0.1
1\
16
0.01
x
x'
"
20
... xx
x
\
;0'
"('
I[
(a)
\ ,-, /
x~x'"
"E
,..,'"
II.
iII ....
~
x
(b)
4
2
0.5
1.0
1.5
2.0
400
)
Potential
minimum
Center
01
channel
I
2.37
2.34
\.\
~~'--
0.5
0.2 0,1.05
253
- .. -=-_. "--
1200
L---.l_.L..L:'--'...L_-1_lLLl..L'::'
2.0 1.5 1.0
800
Depth (A)
, 'i
Chapter 12
expressed as
n (x)
X,(x) ; X",,(x)+[I-X",(x)]_O_,
(12.17)
where nD(x) is the defect density at a depth ofx and n is the total
atomic density. In order to extract nD(x), the decbanneled
fraction XhD must be known.
d(1-X",)
<Ix
(12.18)
(12.19)
wbere X:;W)
254
Ion Channeling
(l2.22)
%bD(X)-X",,(O) =
Jg(EL'x) dEL
(12.20)
E,
(12.23)
Ap
v,
dO
J,
(12.24)
"D
J~ dO.
Eb,
cZ l Z 2 e Np
=(
(12.21)
255
Chapter 12
(b) Channeling
1.5
Depth ()un)
1.0
0.5
8.
--"'r--..,..--~-~ ~v
Nonehenneled
Ii
....
Dfsloeatlon
(al TEM
0.8
1.0
1.2
1.4
1.6
Enl!rgy (MeV)
(e)
p. Implanted SI
3000 A
0.5
1.0
1.5
'-----'
E"2 (MeV)'12
FIG. 12.8. Analysis of pbosphorous-diffused Si by <a) transmission electron microscopy and (b) backscattering spectra for 2.5 MeV
He ions incident along {llO} planar and (Ill) axial channels and (e) energy dependence of the planar dechanneling. A misfit
dislocation network is observed at a depth of 450 nm (from Picraux et aT., 1980).
(
Fig. 12.9. The regrowth rate was lower and nonlinear for a
crystal having a (Ill) surface, and considerable residual damage
(largely in the form of twinned regions) remained after complete
recrystallization.
256
',i
I
I
I
I
I
I
I
I
I
I
Ion Channeling
l!!
<
.. 0
=.
~
1:)
150
m'ng~4
t~~~~~~'~6~'-~~~'~~'~>U~~~"~!~~~'
~~
~~~~.
S ...
amln.,;
1~ ~ \~"-.:--~
...min......
...
~-~<v...,.l
oo
.....
.i'
..
.;
~~.
mm ~f;O
~9
...In~"
...
n.-.
'""'v \I.-
min
.:
.~,~:
'...
o~~
<111> _. _
-i
.",
'I.
840
"'<"t
__~____~__~~
are provided
2002).
dE
dx
(12.25)
257
Chapter 12
till
E=c+-n,
(12.26)
,---r--,
'\.2 ; - .'-'1'--'---'!t'~:
1.. 0
0.8
0.' I0.'
1.2
".0
<100>
L:==",*=====l======l
~
.. _-...................
,0,8
,6f8
(oj)
'
.J t,."
.~~ :~...-'",.
'~';", .
,.
J'~
.~.~
'" ~q-';'l,:-f~..;
',~~-:to>~:u'
~~~~~--~--~~
1'
E~Gf'~'(MQ!O
-'..9J ..... '..
.'
258
&
Ion Channeling
~ !"'q~pr~~t~d ~i
'... 'Y'irgi~. ~i
ta)
depth
J! . ; .......... .
o~~~----.---.----.----r---~
o
-,300
600
Si Depth (tim)
.3
E'
"
i'"
jlii.
(b)
.!'!-
'"'"
'"'"
>='"
(f)
(12.28)
hD
'*" ~
"1.,
J,!
m=1
m:;;;2
m=3
--- . . n;=4
--m=10
"I
Caculated Ino
2
{I
'"
I
"r
(12.29)
. ..
"
..
:;00
soo
Si Depth (nm)
259
Chapter 12
hcp structure would show profiles of type (a) for {1120) planar
The easiest lattice site to study by channeling is a highsymmetry interstitial site such as the tetrahedral interstitial site
in a face-centered-cubic (fcc) lattice (Fig. 12.12). This position
is the e/J,}/4,lh~ site, where do is the lattice parameter. It lies
exactly at the center of (100) channels, is one-half the distance to
the center of (110) channels, and is completely shadowed in
(III) channels (that is, it lies along (Ill) strings of atoms). The
projection of tetrahedral sites into each of these channels is
shown in Fig. 12.12. The normalized yields of backscattered
ions (or other close-encounter reaction products) from solute
atoms that are exclusively in tetrahedral sites will be
r~'II)
(10 TO)
channels.
angular scans, a clear peak in the yield from solute atoms should
be seen at perfect {l00} alignment because the ion flux is high at
the center of the channel.
260
Jon Channeling
ljJiJ
o
2.4~-~-~-~-~-~--~-..,
20
2.2
<11fb. Ctlannel
'If
(a) Substitutional
2.0
Channel cross section
1.4
Host
'If
,.
'If
:!! 1.2
(e) Random
1.0
l
1
[1[1
"""
"-""
'If
(c) Large displacement
'If
X
1
0.8
0
0
'If
(f) Substitutional
plus Interstitial
o
'If
g) Substitutional
'plus small
displacement
261
1_Chapter 12
x;'''''').
atoms,
f(uvw) ::;:
X.
Ft
vw
)
is very sensitive to
{uvw)
-Xh
(12.30)
l-X~"""')
<h
true displaced fraction when the solute atoms are near the
(1) Examine the angular scans (or just the apparent displaced
fraction 'rd~U""') to estimate lattice sites from Fig. 12.13.
(1231)
where
to
the fraction of solute atom sites that project into the given
~IIVW)
In other words,
F;(m,w}
2; then f~D""')
(12.32)
262
Ion Channeling
m2 + ...
small displacement from the atom row (Tang and Zhang, 1988).
(]2.34)
The ion flux F(rtJ at rtl is contributed by only those ions with EJ.
~ ub' Under the condition that El..,i = E~ + Ui = Uh, ion flux is
given by
F(r.)=x
(12.35)
(b)
(a)
and (b) a group of 9-atom rows with a circular unit cell (Tang
and Zhang, 1988).
the Table 12.1, Ui and E.l.,i represent the potential and transverse
energy, respectively, for M. mi represents the number of areas
6A that have the same potential energy (Tang and Zhang, 1988).
u,
... U;
... ELi
... mpA
ELI
mlM
the
angular
dependence
b
U,
.E.Lb
m,,6A
of
.1..1
m.M
A
rnl
N
.c.
L
UL
E.L.L
mLM
g(E)=-'- = - ,
(12.36)
(1972).
For an emitting atom that is displaced from the atomic row,
the distribution probability in cylindrical coordinates is given by
(12.33)
p'
263
(12.37)
/
Chapter 12
(12.43)
where [0 is the upper limit of atom displacement, which is given
by mo' = (NdrJ , with N as the atomic density.
x'
, 'r d(r')
EMITTED
PARTICLE
where
(12.44)
,r,-r
+ U(r).
The obtained
distribution can be nonnalized with the value obtained for 'Ve :::;
5'Vl. Fig. 12.17 shows the results obtalned by Picraux l al. for I
MeV He .Iong the Si (110) axis. The angular width of the
channeling dip increases with increasing displacement from the
lattice row (picraux et aI., 1972).
(12.38)
(12.39)
(12.40)
cosy
(12.41)
(12.42)
264
Ion Channeling
Similarly, a beam directed toward the P wall of the channel
I~'
( \ r-,
/ 1. \
12
~ / l\ \
L
f
Le
"
:Z
Si {lW}, l'/IIK
"! /
' ...
G><::?-S~~=::::;:::;::::~l
- I I
II I '
iUfI
MID.,
I
I
,.
the (100) axial and {I OO} planar channels, but are exactly in the
center of the (110) axial and {Ill} planar channels. (See
Appendix 17 for projections of different interstitial sites into
various cbannels.)
'0 0.0
from the row for I MeV He+ along the (1l0) axis in Si at 296 K
(picraux ef aI., 1972),
265
."
Chapter 12
the Vb yields for the (100) and (lll) channels (Andersen el 01.,
1971b).
..
:!!
~1.8
c
~ 1.&
Q. 1.4
1.2
1.0
L:.2!;.-.-;:,.:-:o:-.-,0:':-:2:-'.J
Angle-
I ++
E
~
>'1.0
~~O8
j'
a: 0.6
{'OO)
1m!
T
.I
f.
II
[J
{111}
iWOli
T
V .Jl
'D
1.8
>.1.6
c
21.4
eCL1.2
~
'"
1.0
L----:;-_~~-----!----.J
0
1
~1
Angle-
266
FIG. 12.19. (100) angular scans for irradiated Al..().02 at. % In.
Backscattering yields (X) of 1 MeV He ions from Al and In
atoms are shown as a function of the angle from a (100) axis.
The sample was irradiated with 1 Me V He ions to the indicated
fJuences at 35 K and then annealed for 600 s at 220 K before
each of the angular scans. The yields were measured at 35 K for
a depth interval of 50-280 om from the surface. Because the Ai
yields were almost the same after each irradiation, only one set
of AI data is shown (from Swanson and Howe, 1983). The inset
show5 the deduced ~trahedral interstitial position of an In atom
(0), surrounded by foor vacancies (D).
Ion Channeling
~
C
:::J
[111'.:
200
[110]~
.....
[Y ''''
">.
;;
'"c
100
.!lo
III
300
-s"!,
FIG. 12.20. An example of a shadow cone behind a surface atom
bombarded by energetic ions. R is the radius of the cone at the .
next atom (Feldman e( al., 1982; Swanson, 1982).
A surface peak is seen in channeling spectra even for a
perfect, uncontaminated surface because the outennost layer of
atoms is not shadowed. An example of a surface peak. for clean'
Pt is shown in Fig. 12.21 (Davies et al., 1978). The presence of a
surface oxide will, of course, erihance the surface peak:, so that
measurements are usually done in ultrahigh vacuum. Depending
on the radius, R of the shadow cone (Fig. 12.20), more than one
atom per string might be seen. For Coulomb scattering, the
radius, R:. of the cone at the second atom is
R, ~ 2(Z,Z,e'dIE).
Surface
o,,'\o~
On.
'... "
o
atom
, '
Q.,~P.~
o 0 0""",,, strmg
00
One
o,~, 0
a-atom
0, 0 ... 0 -
o 0'-':-0... 0
0 ~''::~C string
0--
0, 0
0000'
00000""-
(e) Re.laxed
per
Two
atoms
SUrfa~::~ng
o
00000
0---
Two
atoms
OODOo-per
o
000
o 0
string
A
Energy _
450
400
350
Channel number
267
--~--------------------------------Chapter 12
well.
1979).
(Feldman
el al.,
experiments, the total areas of the surface silicon peak. and the
oxygen peak were measured as the oxide was gradually removed
by etching in a HF solution. The variation of the total area
density of Si atoms in the swface peak as a function of the
oxygen coverage (Fig. 12.24) demonstrated that the oxide was
stoichiometric SiOz. In addition, the density of Si atoms
extrapolated to zero oxygen atom density was 8.6 x lOU atoms
cm-2, which was greater than the density of unshadowed Si
atoms, 6.4 x lOIS cm-2 , in the surface region of a clean Si{IlO}
crystal. The excess 2.2 x lOIS atoms cm-2 were attributed to
displaced Si atoms at the Si-Si02 inlenace.
.1
.l
1
;
268
Ion Channeling
(a)
(d)
[ [ [[
Surface
Bulk
"
W-NI(110)
I<eV
Eo = 101
})))) 6.d
Surface
oJ
06
~~~~\;
(b)
If
t
."
"\
Oxygen
covered
Gi
:;:
Bulk
Surface
Backscattered energy
(c)
I
I
:-'~:-'~\SJ'-:""
_.:-,-- . -~-
I
fl'r-f.::::::::::-_ Surface
Bulk
,I lI
.......
'A6
i!
,
,
Bulk
'
40
43
46
49
52
Scattering angle
,
FIG. 12.23. An illustration of the method to study surface relaxation experimentally, based on channeling and blocking.
(a) The scattering plane coincides with the crystallographic plane. A displacement of the first layer results in different angular
, ,
I
.!
"a,
269
Chapter 12
dislocations.
A thin epitaxial Dverlayer of Ge on (100) Si with a capping
layer ofSi is shown schematically in Fig. 12.25 (Feldman et al.,
1987). For a lattice match in planes parallel to the surface, a
considerable strain is introduced in the normal direction. In the
case of a (100) surface, this strain should not create an increase
in X~OO) but it does cause Ge atoms to project into the (110)
24
E"OJ!leV
~tll>~
.. 20
E
~
.Ea 16
~
~
,..0 12
~
iii
6
8 10 12
Oxygen (1015 atoms/em2)
14
(12.46)
16
01
=d.+ (i+u)(d""-d,,,)
oSl
l-u
(12.47)
.~
{
1
270
Ion Channeling
[1001
150!
SI eap
Norton, 1980).
271
-"
51 (IDOl
substrate
Chapter 12
1110)
~' MeVHe.
1.6
...
Qj
1.4
6~200A
1.2
'$.
... 1.0
Q)
.!:! O.B
1.6
... 1.4
"i 1.2
S[ CAP
'$.
... 1.0
51
Ge
IS
= 0.8
iii
CD
E 0.6
E 0.6
-. .-.-.:..~-~ ~ -0.4
0.2
0.0 ~-'~.L..L.......Jlo~-LL~-'--_
47.0 46.0 45.0 44.0 43.0
Angle from normal (degrees)
1l1li1 (100)
'ciW!
~10
SMLGe
051
"tl
Gi
's:.
1.2
0.4
0.2
0.0
~--'~.........J!O~""""",~,---_
~'.8M.VH
ML Ge
1SO
1.4
A 51 CAP
1.2
:g ,
~ 1.0
1.0
"tl
.~ 0.8
c;; 0.6
=a
E
(; 0.4
0.8
0.6
'.BMeVHe.
;S1100
100ML
SI+10t
%Go
051
- Ge
136..1.51 CAP
E
(; 0.4
z 0~2
.----z .
0.2
0.0
He
200ASICAP
Ge
(a)
1.4
1.8 MeV
L---1-LL.L..!...<........:'~L.L.L-'-'--'~_
0.0
(e)
(d)
FIG. 12.26. Channeling angular scan of GelSi structures, using 1.8 MeV He ions (Feldman ef at., 1987). (a) Scan through the [110]
axial direction in the (l00) plane for an epitaxial structure consisting of Si(IOO)/Ge [6 monolayers (ML)]lSi (20 run), (b) s<:an through
the [110] axial direction in the (100) plane for the same sample as in (a), (c) scan through the [100] (crystal normal) channeling
direction for the same sample as in (a), and (d) scan through the [110] channeling direction in a sample of Si(lOO)lSio."G""" (100
ML)/Si (13.6 nm). The stereogram in the inset indicates the direction of the angular scan for (a) and (b).
272
ton Channeling
r"~~~""-_
.-Goniometer
1972).
The choice of ion energy is determined by considcra.tions of
cross section and mass resolution. The cross sect jon for
backsc~ttering generally decreases with increasing energy,
Scattered
beam
C
i
',"
,
.
"
Cryst I
Solid state
~ detector
He beam
Collimators
1982).
273
Chapter t2
cross section, the ion species, and the ion energy. Nuclear
12.8.S Sensitivity
For lattice location of impurities, the impurity concentration
must be quite high, typically between 10-3 and 10"'" atomic
fraction.
274
.:,:
Ion Channeling
[OWl, [001], and [00 T] poles. The angle between two planes
equals the angle between their nonnals or between their great
circles. These angles can be measured precisely by the distance
between the poles along a great circle connecting the poles.
100
--
'001
01 0 t----+---i.,~'-+_--_" 010
I
.1
I
-...
".....
"",,"-
100
The angles at which these planar dips occur are then plotted
on a polar net for the given tilt angle 8, as sholMl in Fig.
l2.30(c). The procedure is repeated for one or more different
tilts (e.g., ~ 5), the angles for the planar dips are plotted on
the polar diagram, and the points are connected by straight lines.
These lines represent {I!o}, {IDO}, and two {Ill} crystal
planes, and they interneet at a (110) crystallographic axis located
near the origin, 8 = 0. The angles 1jJ and 8 at this intersection
point arc the goniometer settings for (110) alignment. It is
sometimes difficult to distinguish between OlO) and (100)
crystal alignments, because in both cases, major planes intersect
at 90. However, a close inspection of Fig. A17.6 (Appendix
17), which shows the major planes: for cubic crystals having
different alignments, is helpful. In the case of a (lIO) surface
normal, the {HO} and {IOO} planes are perpendicular, but the
{Ill} planes are about 35 from the {lID} planes, as is also the
case in Fig. 12.30(c).
275
/
Chapter 12
PoInt of
o<OiOd!on
-4--------
seellgn through
JIB and PC
A~C
P'
of
first be aligned with the [Oil] axis, then rotated about S from
that axis along the (011) plane, and then tilted to scan across
276
Ion Channeling
8 10 12 1416 20 2428
42 56
112224
,,
, ,
\
\
\
\
\
\
\
1~L-~~~~~~__L-~__~~__~__
0.2Eo
0.4Ec
O.GEo
Energy (Eo
(oj
=1.0 MeV)
0.8Ec,
1.0Eo
(e)
x 5 I''''
~
..... J
....
' . ,. ...
4
.l!la::
I:
..e~2m.
1
.-
>=
~)
g
.. Aligned at 0.9: ~.
{110}
0 40 50 60 70 80 90 100 110120
Rotation angle", (degrees)
(b)
-2
o
Tilt angle 8
(d)
-1
277
Chapter 12
to note the e and $ settings for the axial alignment and for two
points in the plane on either side of the axis and then to
interpolate the intermediate e and ~ settings for the scan. If the
alignment is slightly off the axial setting, the plane will be
em::ountered on one side of the axis, giving an asymmetric scan
[Fig. 12.3 I (b)]
(ioo]
(100)
(011)
plane
plane
ACKNOWLEDGMENTS
The authors are grateful to l.A. Davies, L.e. Feldman, L.M.
Howe, S.T. Pieraw<, R.P. Sharma, J.C. Soares, and D.O. Boerma
NOTATION
a = Thomas-Fermi screening length
11
llo = Bohr radius = 5.292 x 10- m
d = atomic spacing along axial direction
do = lattice constant
~ = interplanar spacing
e 2 = electronic charge squared = 1.44 x 10-13 cm MeV
E = ion energy
dEJdx = energy loss along the ion beam path
FRS =square root of the Moliere string potential
Fps = square root of the planar potential
H "" number of counts per unit energy loss in an RBS spectrum
Mh M2 = atomic masses of ions and target atoms, respectively
N = atomic density per unit volume
Nd = defect concentration per unit volume
(a)
[100]
(b)
o'
T = crystal temperature
._._~ UI= one
<001>
,,'.
'X.~lOOf ,x.!hkl.1 "'" normalized yields for host and solute atoms,
(e)
ad
278
Ion Channeling
REFERENCES
GENERAL
Aono, M., Katayama, M., Nomura, E., Chasse, T., Choi, D., and
Kato, M. (1989), Nucl. Inslrum. Methods 37-38, 264. Coaxial
ion scattering spectroscopy (CAlCISS).
SI~fULA TIONS
Barrett, J.H. (1971), Phys. Rev. BJ, 1527. Monte Carlo computer
simulations of channeling.
Gemmell, D.S. (1974), Rev. Mod. Phys. 46, 129. Good general,
review of channeling.
Hofsass, H., and Lindner, G. (1991), Phys. Rep. 201, 121.
Emission channeling ...
Gemmell, D.S. (1974), Rev. Mod. Phys. 46, 129. Good general
_ revie~ C?f ~banneling.
.~
"_ . . __ -
Tang, J.Y., and Zhang, Z.H. (1988), Stopping Power, Range ond
Channeling Effect of Ions in Solids, Atomic Energy, Beijing,
P.R China.
279
/
Chapter 12
Morgan, D.V., and Van Vliet, D. (1972), Radiat. Eff 1Z, 203.
Computer simulations.
Nelson, RS., and Thompson, M.W. (1963), Philos. Mog. 8,
1677. Early theory.
z.
Tromp, RM., and van der Veen, I.F. (1983), Suif. Sci. 133, 137.
van der Veen, I.F. (1985), Surf Sci. Rep. 5, 199. Computer
simulations and review of surface studies.
Schiett, H.E., and Thomsen, P.V. (1972), Radiat. Eff 14, 39.
ion flux.
EXPERIMENTAL METHODS
Barrett, I.H. (1978), Nucl. Instrum. Methods 149,341. Monte
impurity atoms.
Bottiger, J., Davies, J.A., Lori, J., and Whitton, J.L. (1973),
Nucl. Instrum. Methods 109, 579. Low-temperature setup for
channeling measurements.
Chu, W.K., Mayer, I.W., and Nicolet, M.-A (1978),
Backscatterlng Spectrometry, Academic Press, New York.
Experimental setup and alignment description.
Cbu, W.K., and Wu, D.T. (1988), Nucl. Instrum. Methods B35,
518. Coincidence methods Qf Clflalysis.
280
Ion Channeling
CRYSTAL ALIGNMENT
Picraux, S.T., Chu, WK., Allen, W.R., and Ellison, J.A. (1986),
Cullity, B.D. (1967), Elements of X-ray Diffraction, AddisonWesley, London. Description of stereographic projection and
crystal orientation using Laue back reflection X-ray patterns.
DAMAGE MEASUREMENTS
Baeri, P., Campisano, S.U, Ciavola, G., and Rimini, E. (1976),
Nucl. lnstrum. Methods 132, 237. Defect displacemeu"tS from
channeling data.
Ellison, J.A., Picraux, S.T., Allen, W.R., and Chu, W.K. (1988),
Phys. Rev. B37, 7290. Also see: Picraux, S.T., Biefeld, RM..
Allen, W.R, Chu, W.K., and Ellison, J.A. (1988), Phys. Rev.
Gemmell, D.S (1974), Rev. Mod. Phys. 46, 129. Good general
review of channeling.
Howe, L.M., Swanson, M.L., and Quenneville, A.F. (1976),
Nucl. Instrum. Methods 132, 241. Channeling yields and
dechanneling experiments in AI, Si, Cu, and Ge.
on Si.
z. Naturforsch. A15,
281
Chapter 12
dislocations.
Sharma, R.P., Rebo, L.E., Baldo, P.M., and Liu, J.Z. (1988),
Phys. Rev. B38, 9287; (1989), Phys. Rev. Lett. 62, 2869.
Discontinuity of channeling widths in high-Tc superconductors
at the transition temperature.
. Bugeat, J.P., and Ligeon, E. (1979), Phys. Lett. A71, 93. Lattice
sites of H in fcc metals.
Eriksson, L., and Davies, J.A. (1969), Ark. Fys. 39, 439.
Variation of channel width with the atomic mass of constituent
elements in compounds.
J.c.
CHANNELING
AND RANDOM
STOPPING POWER
..---_.
--->.
--'~-"-
Araujo, L.L., Grande, P.L., Behar. M_. Dias, J.F., Lifschitz, A_F_,
Arista, N.R., and Schiwietz, G. (2004), Phys. Rev. A 70, 032903.
Channeling stopping power_
Gemmell, D.S. (1974), Rev. Mod. Phys. 46, 129. General review
of procedures.
Shao, L., Wang, Y.Q., Nastasi, M., and Mayer, J.W. (2006),
Nud. Instrum. Methods Phys. Res. B249, 51. Stopping power of
channeled ions.
282
Ion Channeling
PI.
Davies, J.A.~ Jackson, D.P., Norton, P.R., Posner, D.E., and
Unertl, W.N. (1980), Solid State Commun. 34, 41. Surface
relaxation ofPt
Davies, J.A., and Norton, P.R. (1980), Nuc!. Instrum. Methods
168, 611. Re(.;onstruction of Pt surfaces.
Feldman, Le., Kauffman, R.L., Silverman, P.J., Zuhr, R.A., and
Barrett, l.H. (1977), Phys. Rev. Lell. 39, 38. Effect of thennal
vibrations on channeling surface peak.
Feldman, L.C., Silverman, PJ., Williams, LS., Jackman, T.E.,
and Stensgaard, I. (1978), Phys. Rev. Lett. 41, 1396. Si-Sio,
interfaces in thin Si crystals.
"
Frenken, l.W.M., Maree, P.M.!., and van der Veen, l.F. (1986),
Phys. Rev. B34,7506. Surface melting ofPb.
283
'. '!.
Chapter 12
Jackson, D.P., and Barren, J.H. (1979), Phys. Lett. A71, 359.
van der Veen, J.F. (1985), Surf. Sci. Rep. 5, 199. Computer
peak.
van der Veen, J.F., Smeenk, R.G., Tromp, R.M., and Saris, F.W.
(19790), Surf. Sci. 79,212. Relaxation ofNi sunaces.
van def Veen, J.F., Smeenk, RG., Tromp, R.M, and Saris, F.W.
(I979b), Surf. Sci. 79, 219. Relaxation ofPtsunaces.
and
Turkenburg,
van der Veen, J.F., Tromp, R.M., Smeenk, R.G., and Saris, F.W.
(1979c), Surf. Sci. 82, 468. Position of S on Ni sunaces.
284
CHAPTER
CONTENTS
13.1 MICROPROBE OPTICS ............................................................................................ 286
13.1.1 Introduction ........................................................................................................ 286
13.1.2 High-demagnification nuclear microprobes ...................................................288
13.1.3 Comparison between megaelectronvoit ions and
kiloelectronvolt electrons ................................................................................... 289
13.1.4 Ion beam damage and ion beam modification ................................................ 289
I
I
I
I
I
I
285
Chapter 13
13.1.1 Introduction
A very brief introduction to microprobe optics is given here
because the subject is well-covered in a variety of other texts.
General accounts of quadrupole optics can be found in many
sources, sucb as Dymnikov and Yavor (1964), Septicr (1967),
and Hawkes (1966), as well as those specific to microprobe
probe-fonning systems such as Maggiore (1982), Grime and
Watt (1984), Wall and Grime (1987), Breese of al. (1996), and
Dymnikov and Osetinskij (1989).
imllge
__________""'\ibiEici _________-'00
..'
...~dlstii!tce:
disiiinct!
.-
. t
,sah~iai~s
~"~~~;;~~~~~~~~~~~:::::::llflll
.Qbjed
:slitS
COflitnlllor
-'.IM
1tj1~ge,
~~1~'
FIG. 13.1. Schematic of the major components of a microprobe, in this example using a quadruplet focusing system.
286
produce a point focus. The plots show the different effects on the
position and angle, (x,9) in the horizontal plane and (y,p) in the
vertical plane, of each quadrupole system. The features of each'
focusing system are summarized in Table 13.1, based on a
quadrupole lens length of 10 em and a bore radius of7.5:mID.
(b) Triplet
(c) Quadruplet
(d) Highdemagnification
triplet
-10,-62
25
213,378
77,-20
39
504, 525
-14, -14
14
112,274
230,-55
112
1024,1286
130,150
340,960
101, 160
314,830
-85,-99
133, 1663
-64, -61
2000,-7700
Demagnifications
(919), (~I$)
Mean demagnification D
Lem strengths
(ampere turns)
Chromatic aberration
(,.unlauad% )
(x/SO), (ylejli)
Sphericaiaberrations
(I'm/auad')
(x/Sf), (yI9'?)
*Only the most important chromatic and spherical aberration coefficients are included for brevity .
.(y,$)
-'4
-'""
6>20
."
,.W
:Z-'iJ~;"
{m)
.'"
;'.00
,
,
>-
i;g
-~ .!
~.-I,or
:~t
.
~cc
Il~~
0 D0 0
:ti.+CI
11,'50: _ (;.6p
z-"'''1~,'f''1 ..
7.~
FlG. 13.2. Ion optical trajectories of an ion with a divergence of +0.1 mrad and -0.1 mrad. respectively. in the horizontal and vertical
planes of (a) doublet:(b) bigh-c::x(;itation triplet, and (c) "Russian quadruplet" microprobes. The microprobe length of 7.1 m and the
image distance of 15 em are kept constant. In (d), the triplet in the image distance is reduced to 7 em, and the lens lengths are halved
to 5 em. Produced using the matrix code PRAM (Jamieson and Legge, 1987).
287
produce a I - m beam
spo~
the angles
e and ~.
so that the pole pieces project 15 1DD1 beyond the main lens body
(Fig. 13.3(0)]. This moves the effective exit face of the fin.llens
as close as possible to the wall of the target chamber in order to
reduce the incage distance (Fig. IJ.3(b)]. Figure 1l.2(d) shows
the focusing action of a triplet of these new lenses, with the total
microprobe length kept the same as for the other configurations.
The image distance was reduced from 15 em to 7 cm as a result
of the new lens design and a redesigned target chamber,
resulting in a maximum demagnification of -230 and a mean
demagnification of -11 0 in Table 13.1. The beam converges on
the sample surface more steeply owing to the higher
demagnification, giving a smaller depth of focus over which
regions of the samples with different heights are in focus. A
drawback of suc~. highdemagnification microprobes is that there
. is little space for detectors to be located to view the front surface
of samples, requiring better chamber design.
smallest beam spot size for a given beam brigbtness and current.
This system is called the high-excitation triplet (Grime et al.,
1982; Grime et al., 1991), because there are other possibilities of
fanning a point focus with lower lens currents, although with
lower demagnifications. The lens currents for the triplet are
higher than those for the other configurations, so the lenses tend
to become hotter and can cause thermal drift effects ifheat is not
adequately removed.
288
ions require focusing systems that are more complex a.:nd more
expensive and have higher aberrations. ~e payoff for tlJ.is extr~
complexity is that the megaelectronvolt Ion beam tends to stay
focused through many microns of the material. because the
greater ion momentum makes large-angle scattering unlikely. In
comparison, lighter electrons undergo a strong increase in the .
lateral b.eam spread with distance into the materiaL To jllustrate .
this effect, Fig. 13.5 shows Monte Carlo plots of the trajectories
of megae1ectronvoll protons and 50 keV electrons through 10
~ of the polymer poly(methyl methacrylate) (PMMA). The
vertical scales are the same on both plots, but the h.oQnzontal
scales differ by a factor of 100. This means that focused
megaelectronvolt ion beams can be used to attain higher spatial
resolution analysis than kiloelectronvolt electrons in layer
thicknesses of greater than 1 I-UD.
<00
350
joo
25'
"
e
::. .200
o
"
15'
Sigma:i6.5i.4
FWHM=34.S%3,3
289
'., ~
Chapter 13
2003).
13.2 APPLICATIONS OF NUCLEAR MlCROPROBES
13.2.1 High- and low-beam-current applications
Microprobe analytical applications can be divided into two
regimes. In the high-current regime of typically 10 pA to 10 nA,
l (3)
6 .....
"'j"
8i .. _ .. _' .. .H
"Literal
.'. , b);pli1ceiije[ll(nrn)
'., ., ..... ",... .
'
FIG. 13.5. Monte Carlo plots of the trajectories of (a) 2 MeV protons (and also the subsequent generation of secondary electrons) and
290
r
I
I
I
I
ion induced
luminescence ~
IBIL
Ion induced
l,S
$ ; Electron-hole pair
-=.:!
creation - ISIC
..
techniques
Most applications of these techniques use unfocused
megaelectronvolt ion beams, as described in other chapters of
this book. PlXE bas relatively large production cross sections
compared with R;BS and nuclear reaction analysis (NRA), so it is
quite easy to generate spatially resolved images of elemental
ctistributions at beam currents as low as 10 pA with a microprobe.
Most microprobe applications of RBS, NRA, and elastic recoil
detection analysis (ERDA) rely on recording spectra from
spatially resolved points across a surface rather than producing
images that require larger beam currents. Many microprobe
applicatioIlli of these techniques are given in Grime and Watt
(1984), Watt and Grime ([987), and Breese et al. (1996).
291
. --
Chapter 13
I~
2O!l
:C ii:o
100
. '50
2DO
250
300
350
40[,
Choiulnel
FIG. B.S. (Top) 500 x 500 m'p3rtic1e-induced X-ray emission
(PIX) images showing the Si and Ti distributions in a silicon
carbide composite sample exposed to LhTi03 for 1000 h.
(Bottom) Rutherford backscattering spectrometry (RBS) spectra
measured from the two points identified in the FIXE images.
Reprinted from Alves el 0/. (2001).
I" ~Or
400
-'-=; .
resolution of a few microns the best that can .be achieved with
microprobe analysis. Typical microprobe applications include its
use for point analysis in measuring the distribution of hydrogen
across catalyst pellets (Sofield e/ or., 1981) and in melt
inclusions (Mosbah eJ 01., 1991).
292
.1
10
293
Chapter 13
FIG. 13.10. (a) Low-magnification No~ki image of partially relaxed silicon--germani~ surface. (b}-{d) Transmission channeling
images of the 11 x 11 "",' box area shown in (a). The tilt angle with respect to the (110) planes is (b) +0.16' , (c) 0.00' , (d) ...;}.16'.
Reprinted from Breese et al. (2005).
As an exnmple, Fig. 13.11(a) shows a typical schematic of the
STIM setup (Minqin el a/., 2007). with cell samples mounted on
a thin silicon nitride window. Behind this is a charged particle
detector, which can be either a silicon surface bamer detector or
a PW diode (i.e., a diode with a nearMintrinsic semiconductor
region between regions of n- and p-type semiconductors). PIN
294
.,
ft?cl!seqion,
be"m'
~arn;!.rnatsu Pin
diode 51223
100000
(c)
19000
II
<
1000
0
0
.r:
0.
100
<e
lQ
1
500
650
950
1100
En~rgy (keV)
FIG. 13.11. (a) Scanning transmission ion microscopy (STIM) setup for imaging cells. (b) Optical micrograph of cells to be analyzed.
(c) STllvl ene"rgy spectrum obtained using a focused I MeV helium ion beam s~anning over the area shown in (b). (d) STIM. image
from the 50 x 50 ~2 box area sbown in (b). Reprinted from Minqin el al. (2007).
295
Chapter 13
P~amptifierJ
amptifieriDAQ
(a)
MeVions;
Side
. - - - - - - . - - - - - - - irradiation
with
Pr&'amprifieri
ampJifierlDAQ
MeVioris
(b)
FIG. 13.12. Geometries used for ion beam induced cbarge (IBIC) analysis of (a) devices and (b) planar semiconductor or insulator
wafers, containing grain boundaries, dislocations or other types of lattice defects.
296
increasing magnification of a region at the tip of a singIe 100).lIIl-wide electrode. The images were acquired for 20 min. each,
using a focused 2 MeV proton beam and a diamond bias ,",oltage
of 200 Von alternate electrodes. In Fig. 13.13(0), the d.etector
active area is in the lower portion. The vertically running dark
areas correspond to the locations of the metal electrodes;=- where
the collection efficiency is lowest. The vertically running white
regions correspond to the regions where the largest charge pulses
are measured from the exposed diamond surface between the
interdigitated electrodes, where an electric field gradient exists.
The highest regions of charge collection are from the tips of the
electrodes, consistent with this being the location of the highest
electric field gradient. In the higher-magnification images [Figs.
13.13(b)-{d)l, individual grains are resolved, where their widely
differing collection efficiency can be observed.
297
Chapter 13
..
298
.'.
B If
~.
Ii
p.olt~.t..
n-origat.s
and drains
~iJ.d
(a)
dnlilis
(b)
FIG. 13.15. (a) Single-event upset (SEU) cross section of a silicon-an-insulator (Sal) static random access memory (SRAM) device
from broad-beam tests carried out at Sandia National Laboratories (SNL) and Brookhaven National Laboratory (BNL). (b) SEU image
of the same device stnu;ture recorded with the nuclcar microprobe and overlaid on the circuit layout.
XAXIs
4,OXl o\!-Tr'"",,,,,,,====,,=~,.--.:--:<::;;;;:;;':=-----.
3.Sx10'5
'3.Ox.,o'~1L:=:'::===
2,:5;11;10"
-~
2,OX10:
1.5x'10
5,0;1(10'"
(c)
0.0-:_
.,
<
~tl
......
_~---<i!I
'.
'
En"tgy [EN]
FIG. 13.16. (a) Scaoning electron microscopy (SEM) image of.
diamond microdosimeter tip. Ion beam induced luminescence
(IBIL) and particle-induced X-ray emission (FIXE) images were
recorded from the 100 x 100 }illl' box area shown in (b).
Superposition of the tungsten PIXE image (dark region) and a
panchromatic IBIL im.ge (light gray region). (c) IBIL spectra of
a small region of the microdosimeter tip. Reprinted from Vittone
e/ al. (200Ib).
299
Chapter 13
13.3 SUMMARY
There are about 50 microprobe research groups around the
world, involved in a wide range of 'studies of microanalytical
work in biology, biomedicine, materials analysis, geology, and
300
w:ut
REFERENCES
Alves, L.c., Alves, E., da Silva, M.F., Soares, J.e., Paul, A., La
Barber., A., and Riccardi, B. (2001), Nuel. Instrum. Methods
BI81,377
Arteaga-Marrero, N., Pallon, I., Olsson, M.G., Auzlelyte, V.,
Elfman, M., Kristiansson, p" Malmqvist, K., Nilsson, C., and
Wegden, M. (2007), Noel. Instrum. Methods B260, 91.
Aton, T., Seitchik, I., Yang, S.-H., and Shichijo, H. (1997), Nuel.
Instrum. Methods B130, 507.
Bettiol, A.A., Nugent, KW., and Jamieson, D.N. (1997), Nuel.
instrum. Methods 8130, 734.
Doyle, B.L., and Wing, N.D. (1983), IEEE Trans Nuel. Sci. 30
1214.
'
Doyle, B.L., Walsh, D.S., and Lee, S.R. (1991), Nuel. Instrum .
Methods B54, 244.
Dymnikov, A.D., and Yavor, S.Y. (1964), SOy. Phys.-Teeh Phys.
(USA) 8 (7),639.
DymnikOY, A.D., and Osetinskij, G.M. (1989), Sov. J. Pa..-/. Nucl.
(USA) 20 (3), 293.
Dymnikov, A.D., and Glass, GA. (2004), Nucl. IMStrum.
Methods B219-220, 994.
Esquinazi, P., Spemann, D., Hohne, R, Setzer, A., Han, K.-H.,
and Butz, T. (2003), Phys. Rev. Lett. 91,227201.
Fischer, B.E., and Muhlbauer, e. (1990), Nucl. Instrum. Adethods
847,271
Fiscber, B.E., Schlagl, M., Barak, I., Adler, E., and Metzger, S.
(1997),Nucl.lnstrum. Methods B130, 478.
Folkru:d, M., Vojn~vic, B., Schettino, G., Forsberg, M., 13owey.
G., Pose, K.M., MIchael, 8., Michette, A.G., and Pfauntsch, SJ.
(1997), Nucl. Instrum. Methods Bl30, 270.
Galbiati, A., Breese, M.B.H., Knights, AP., and Sellin, P.J.
(2001), Nllc/. Instrum. Methods A466, 52.
~9~::---~;;kes:
301
L.
s~rin-;'~ Tra-';;~
Chapter 13
Jaksic, M., Gracin, D., Meduni, Z., Pastuovi, Z., Vittone, E., and
Nava, F. (2002), Nud. Instrum. Methods B188, 130.
Minqin, R., van Kan, J.A., Bettiol, A.A., Daina, L., Yee Gek, c.,
Boon Hua~ B., Whitlow, H., Osipowiez, T., and Watt, F. (2007),
Nucl. Instrum. Methods B260, 124.
Mosbah, M, Clochiatti, R., Tirira, J., Gosse~ 1., Massin~ P., and
Troeellier, P. (1991),Nucl. Instrum. Methods B54, 298.
Morris, W.G., Katz, W., Bakhru, H., aod Haberl, A.W. (1985), J
Vac. Sci. Techno!. B3 391.
Nilsson, C., Petriconi, S. Reinert, T., and Butz, T. (2007), Nucl.
lnstrum. Methods B260, 71.
Ortega, R. (2005), Nucl.lnstrum. Methods B23I, 218.
Osipowicz, T., Zmeck, M., Watt, F., Fiege, G., Balk, L.,
Niedemosfueide, F., aod Schulze, H.-I. (2001), Nuel. I.strum.
Methods B181, 3 II.
Pallon, J., Yang. C., Utui, R.I., Elfman, M., Malmqvis~ K.G.,
Kristiaosson, P., and Sj6land, KA. (1997), Nucl. Instrum.
Methods B130, 199.
Lo Giudlce, A" Pratesi, G., Olivero, P., Paolini, c., Vittone, E.,
Manfredotti, C., Sammjceli, F., and Rigato, V. (2003), Nucl.
Instrum. Methods B210, 429.
Pierce, T.B., McMillan, J.W., Peck, P.F., and Jon"" I.G. (1974),
Nuc!. Instrum. Methods lI8, 115.
Manfredotti, C., Fizzotti, F., Vittone, E., Boero, M., Poleselln, P.,
Galassini, S., Jaksic, M., Fazinic, S., and Bogdanovic, 1. (1995),
Nucl. Instrum. Methods BIOO, /33.
Rath, S., Sellin, P., Breese, M.B.H., Herman, H., Alves, L.e.,
aod Hollaod, A.H. (2003), Nuc/. Instrum. Methods A512, 427.
Rossi, P., Di Maggio, C., Egeni, G.P., Galligioni, A., Gennaro,
G., Giacomelli, L, Lo Giudice, A, Pegoraro, M., Pescarini, L.,
Rudello, V., and Vittonc, E. (2001), Nuc!. lnstrum. Methods
BI81,437.
c.,
Manfredotti,
Vittone, E., La Giudice. A., Paolini, C, Fizzotti,
F., Dinca, G., Ralchenko, V., and Nistor, S.V. (2001), Diamond
Relat. Mater. ]0,568.
302
Sakai, T., Oikawa, M., Sato, T., Nagamine, T., Moon, H.D.,
Nakazato, K., and Suzuki, K. (2005), Nuc/. lnstrum. Methods
B231, 112.
Schone, H., Walsh, D.S., Sexton, F.W., DOYle, B.L., Dodd, P.E.,
Aurand, J.F., and Wing, N. (1999), Nucl. lnstrum. Methods
BI58,424.
Sellin, PJ., Breese, M.B.H., Alves, L.e., Sussman, R.S., and
Whitehead, AJ. (2000), Appl. Phys. Lell. 77, 913.
Seng, H.L., Osipowicz, T., Zhang, J., Tok, E.S., and Watt, F.
(2005), Nucl. Ins/rum. Methods B231, 446.
Soptief, A. (od.) (1967), Focusing of Charged Particles,
Academic Press, London, vols. I and II.
Sexton, F.W., Hom, K.M., Doyle, B.L., Laird, J.S., Cholewa, M.,
Saint, A., and Legge, G.J.F. (1993), Nud. [nstrum. Methods B79,
436.
Shariff, A., Auzelyte, V., Elfman, M., Kristiansson, P.,
Malmqvist, K., Nilsson, C, Pallon, 1., and Wegden, M. (2005),
Nucl. Instrom. Methods B231, 1.
Sofield, C.1., Bridwell, L.B., and Wright, C.1. (1981). Nucl.
instrum. Methods 191, 79.
Teo, E.1 .. Breese, M.B.H.. Bettiol. A.A., Watt, F., and Alves,
L.C. (2004), J. Vac. Sci. Technol. B22, 560.
Townsend, PD., Chandler. P.1., and Zhang, L. (1994), Optical
Effects of Ion Implantation, Cambridge University Press,
Wang, W., Zhu, J., Le, A., Lu, R, Lin, S., Bao, 1., Jiang, D., She,
J., and Ji, A. (1993), Nuc!. [nstrum. Methods B77,349.
Watt, F., and Grime, G.W. (eds.) (1987), Principles and
Applications of High Energy 10n Microbeams, Adam Hilger,
Bristol, UK.
Watt, F., Choo. T.F . Lee, K.K.. Osipowicz, T.. Orlie, I., and
Tang, S.M. (l995),N"cl. [nstrum. Methods Bl04, 101.
Watt, F., van Kan, 1.A., Rajta, I., Bettiol, A.A., Choo, T.F.,
Breese, M.B.H., and Osipowicz, T. (2003), N"c/ Instrum.
Methods B210, 14.
Yang, C., Jamieson, D.N., Pakes, C.L, George, D.P., Reame,
S.M., Dzurak, A.S., Gauj., E., Stanley, F., and Clark, R.G.
(2003) Nucl. [nstrum. Methods B210, 186.
Younger, P.A., and Cookson, J.A. (1979), Nuc!. [nstrum.
.... Methods B158, 193.
London.
Udalagama, C.N.B. (2006), PhD thesis, National University of
Singapore, SingafXlre.
Visser, 1., Mous, DJ.W., Gottdang, A., and Haitsma, RG.
(2005), Nucl.Instrum. Methods B231, 32.
Vittone, E., Fizzotti, F., Gargioni, E., Lu, R., Polese\10, P., Lo
Giudice, A., Manfredotti, C., Galassini, S., and JaksiC; M. (1999), ~j.71""'~'-~"'"""""'"
Nucl. Instrum. Me/hods B158, 476.
303
I I.
IL
':."'-~ '"-----.-"."'-"'><~~..,....,."""--.~-
CHAPTER
E. Rauhala
University of Helsinki, Helsinki, Finland
CONTENTS
14.1 INTRODUCTION..~ ............................................ ! 307
14.1.1
Data analysis software .............................................:.. :.. :..:.:.: ... ,.............::.:.307
14.1.2
Scope of the chapter .....................................................................................307
14.1.3
Historical development and reviews of software programs ..................... 307
14.2 TYPES OF CODES .................................................................................................307
14.2.1
The direct calculation method .................................................................... 307
14.2.2
Simulation by successive iterations ............................................................ 308
14.2.3
Interactive versus automated ......................................................................311
14.2.4
Deterministic versus stochastic (Monte Carlo) ......................................... 311
14.3 CAPABILITIES OF CODES .................................................................................312
14.3.1
Design basis ................................................................................................... 312
14.3.1.1. Techniques implemented ............................................................. 312
14.3.1.2. Experimental conditions supported ...............................................312
14.3.1.3. Description of samples ..................................................................... 312
14.3.2
Databases implemented ...............................................................................313
14.3.2.1. Stopping power ................................................................................. 313
14.3.2.2. Scattering cross section .................................................................... 314
14.3.3
Basic physics ................................................................................................. 315
305
.-
"~~r
Chapter 14
t
!
14.3.4
+I
t
t
14.4
ACCURACY ............................................................................................................320
14.4.1
Numerical accuracy of codes .......................................................................320
14.4.2
Intrinsic accuracy of IDA experiments ......................................................320
14.4.2.1. Models and basic physical quantities .............................................320
14.4.2.2. Experimental conditions ..................................................................321
14.4.2.3. Counting statistics ............................................................................321
14.4.3
Physical effects on data analysis ................................................................. 321
14.5
14.6
14.7
14.8
ACKNOWLEDGMENTS ...................................................................................................337
.i
1
!,
,
;
";
306
14.1 INTRODUCTION
14.1.1 Data analysis software
The data analysis software in ion beam methods are computer
programs designed to extract information about samples from
the measured ion beam spectra. The desired information includes
identification of sample elements, their co~centrations, areal
densities, and layer thicknesses. At best, one spectrum can be
converted to concentration--depth distributions of all elements in
the sample, Often, however, such a full description of the sample
based on a single experiment is not possible. The analyst can
then perfonn additional experiments using different
experimental parameters such as ion energies or a different
measurement geometIy or including information from one or
more other complementary techniques.
ter
307
Chapter 14
,0000
16000
o~
1"
r-
.Si
:;;;.
12000
.i
4"
La
i~O"
..;;.;i;;
:rj . ~
2000
P,~~~~r-~~~~
~5ilo 600
to() Iioo
900 1000' hOt! . j'2(j()
4i'
.fa'\
j,
~~tt~red i.ori
.~
"
enefgy{keY)
Fig. 14.1. Backscattering spectrum (0) for 2.00 MeV ~e ions incident on a LaGaO layer on a silicon substrate.
308
....
..
.;' ..
. " .. " .'
1<00. 1450 1500 " 1550,1600 1650 :1700'1750
laoo 1850
Figure 14.1 shows the spectrwn for 2.00 MeV 4He ions incident
on a LaGaO layer on a silicon substrate. In the experiment, the
following parameters were used:
309
i:
l"'-. _
Chapter 14
subl.yers are shown as thin solid lines. The thick solid line thus
represents the sum of the thin lines of the elemental signals in
the subl.yers.
not well defmed and that the concentrations change with depth
within the thin film.
''''0
HiM
2~i50 MeV iU
PbZr(Sn)O/S(
2S[0"~t1M"""'"
2400
2000
Ph
1600
1400
1200
1lO0
'G;
1000
600
(5
z: "'0
,!,!
"
600
Si
"
:~ 600
" ..... <J ..
200
. '''''-cd+~'..
~If-""'T'"......"..,..=--='H
OiO ~o GI aD 'roo i2ii
FIG. 14.2. B.ckscattering
'.0 JO;<~",*~
"300
'.3.ii
';00 '410"
spec~~:;~~r:e~r7Li ions incident on. PbZrSnO thin'~=;I~~:!Si substrate. The thick solid f,'
light elements. The oxygen signal was not fitted; the oxygen
concentration follows from the concentrations of the heavier
elements, as it is presumed to make up the total fractional
concentration of each sublayer to 1. Bohr straggling and a scaled
stopping of 1.08 times the TRIM-91 stopping value were
assumed for the metal oxide layers.
310
:~ ~
-~
.-
k i
:f
O,6
5'
..
.~0.20
Pb
115
---ZI
f!
[5
... _
C 0.4
o.s .. '"_ .. ~ . . . . .
1.t
.,
:
gO.10
0.05
100
200
300
400
Areal
500
600
701)" 8()0
density(1015at1.cm~
900
1060
Table 14.1. Experimental parameters and concentrations of the elements and areal densities of the sublayers.
Sublayer
lOIS at/em:!:
Pb
Sn
Zr
Si
100
0.16
0.0075
0.145
100
0.173
0.0128
0.163
0.6875
0.6512
100
0.168
0.0155
0.165
0.6515
0.1
0.1
0.611
4
-:5
0.17
50
--
-50
0.11
0.019
-0
0.022
0.31-='=- 0.558
50
0.036
0.0153
0.66
50
0.023
0.977
200
0.0015
0.9985
5000
;...~
~~--
...".
.. ~
----
0.2887
311
--_._-----------,...--
'" ,,1
"":
Chapter 14
developed.
'When choosing which code to use, users must consider
whether a given code is capable of extracting the information
required and how easy it is to do so with each code. A complex
sample where interface roughnes.s and multiple scattering playa
crucial role might need a last-generation code, but to determine
the thickness of a thin film with one single heavy element on a
light substrate, a full Monte Carlo calculation is probably too
much work without extra benefit. Also, in some cases. manual
calculations with careful consideration of error sources might be
the most accurate method (see, for instance, Boudreault et aI.,
2002).
14.3.1 Design basis
The fIrst consideration is obviously whether the data to be
., ,
l
'r
analyzed are within the design basis of the code considered. The
two main points are the techniques implemented by each code
and, for each technique, the experimental conditions supported.
;<~I'
.~
For instance, jf a code can only analyze RBS data with H and
He beams, there is no point in trying to use it to analyze NRA
312
~N
,-.l
''>'"
6000
b) 45<
'10'
FIG 14.4. Data and fit (solid line) obtained for a 21-layer
ZrOz/SiO z antireflection coating with a top Au laycr, shown for
the data collected at (a) nonnal and (b) 45 0 incidence, The fitted
partial spectra for Si and 0 are also shown.
. '
.,
. For in;i;mce. ill Fig: '14.5, ;;"e show the calculated spectrum
for a 3 MeV 'Li beam impinging normally on a SilSiO:z (2 x lO l8
atJcrn')fTi (1.5 x 10" at./em') sample and detected at .160', as
obtained using 1985 stopping-power values and SRfM version
2006.02. It is clear that the very large changes in the simulation
will be reflected in the accuracy of analyses of real data, with
errors of up to 10% being made in thickness values, depending
on the stopping values used.
Large changes also occurred for light ions in some common
systems. For instance, the stopping of He in Si is now known
with an accuracy of better than 2%, and this update has been
integrated into SRIM and other databases.
313
Chapter 14
200
I
I_lOin SiO,
150
III
'2
:::l
.ri
~
ttl
100
"iii
Ti
Si bulk
---
>=
50
a
500
1000
1500
E (keV)
FIG. 1'4.5. Calculated spectra for 3 MeV 1Li beam incident on a SilSi0 2 sample. The onJy difference between the simulations is that
different stopping powers were used: values from 1985 (solid line) and the most recently available values (dashed line). for accurate
analyses, users must be aware of the stopping powers used by data analysis codes.
314
~I
I
I
I
I
I
I
I
I
Data Analysis
Softw~re
I
I
I
r
I
I
!1
15000
C 11.1 at.%
18
6.59x10 at.lcm 2
.e-!!.
::Q 10000
~
;;:
Ni
.~."\
I
I
5000
Ta
o~____~__~/__~__~~__~__~-.__~__~'~__~__-L~
1000
1500
1250
1750
E (keV)
FIG. 14.6. Spectrum for 1.75 MeV 'H backscattering off a TaNiC film on Si.
1%) and much smaller than the error in the stopping-power
values.
With the assurance that the codes (at least those that
participated in the IAEA intercomparison) are correct for simple
cases t the user might be concerned with how fast the simulations
are. With modem computers, the simulation of simple spectra,
even with straggling included, is always fast. Efficiency
315
physics~
such as
Chapter 14
and channeling.
The first comment is that, in principle, the best way of
dealing with all of these phenomena is full Monte Carlo
simulations. MCERD, for instance, is a Monte Carlo code for the
analysis of ERDA data, but it can also handle RBS. It can
include all of these effects, except for channeling. Different
"---~codes have been 'dedicated to the analysis of channeling in
specific systems.
However, Me codes are mostly used by their authors, and
calculation times can be an issue. Development of intuitive user
interfaces and continuing gains in efficiency might change this
situation, but for the moment, traditional codes are still the most
often used. The following discussion is therefore related to
traditional codes
shape
spectrum.
316
400
Experiment al
...... Smooth + plural scattering
_._._. Rou:Jh
- - ROllilh + plural scattering
300
~:l
o 200
U .....
300
400
500
600
700
800
900
Channel
FIG. 14.7. 2.5 MeV protons back.scattered from a 3.5-~ W layer on a rough carbon substrate, at normal incidence and with a
scattering angle 165'. (CiIcles) Experimeotal data, (dotted line) calculated spectrum for a smooth W layer (3.6 l'Il') on a ,mooth C
substrate including plural scattering, (dashed line) calculated spectrum for a rough W layer (3.5 Jlm, cr = 0.30~) on a rough substrate
(fwhm 20'); (,olid line) ,ame as d""hed line hut including plural ,cattering. (From Mayer, 2002.).
317
<..\~~. ..
~"
Chapter 14
318
is
319
.. -
--------'-------------....
., ,
.,
,~
Chapter 14
14.4 ACCURACY
The accuracy, that can be achieved in an IBA experiment
I'
,t "
]~:
limit to the accuracy with which layer thicknesses and depth A"
profiles can be determined. A statistical analysis by Ziegler
-1 fi,
:., ., ';,._-: .:~.
""
;:~ '>
-~ '"
:i~,
,,'~_:
~'~ ~~
~ :J
1~g"
:E~
~~~
320
~l
TIili
321
Chapter 14
FIG. 14.8. First step of a RUMP analysis: load the data and define the experimental conditions.
322
O~ta
The second step is often to configure some simulation parameters using line commands. In this case. SRIM-2006 stopping powers are
used. the Don-Rutherford scattering cross section for the 160(a,a)160 is loaded, and straggling and multiple scattering are enabled. By,
default, all isotopes are calculated separately (as opposed to making one :lingle calculation for the average mass of the target element),
and Andersen electron sCT7ening (Andersen et oZ., 1980) is included.
.
FIG. 14.9. Second step ofa RUMP analysis: Configure ;imulation parameters.
The third step is to defme the sample structure. This is done using inruitive, easy-to-Iearn commands. The corresponding first guess is
then calculated.
323
Chapter 14
Energy (MeV)
'\
\~..-.
-u
..........
__ _____ -:o
~
O~~~-r~~~~~~~~~~
500
1000
1500
FIG. 14.10. Third step ofa RUMP aoalysis: define the sample structure, perform the simulation.
324
2000
Data Analysis
Softwar~
A manual iterative procedure as described in EXAMPLE 14.1: would follow the first guess. Alternatively, RU1vlP inc ludes an
optimization algorithm. The window shows the definition of the fitting space and the final results. In most situations, [lot all -variables
would be fitted at the same time. A combination of manual iteratio.Ds and fitting of some parameters is typical.
.:...~
"
'";.':
.f,
.'
Edqyo.1cV}
OJ
1.0
Jl
"L'
....
_0,
__ .
~_..-~
_ _ --...... _e-.
--.,
,
,,
,.,
,,..
'"''
'''''
FIG. 14.l1. Final result of a RUMP analysis after manual or: automated iteration until agreement between simulation and fit is
achieved.
Chapter 14
analysis.
RU1v[P lacks many of these capabilities, which is probably
(as opposed to the many codes that are used mostly by their
author only), are elSA (Raubal., 1984; Saarilabti and RaubaIa,
1992) and REX (K6tai, 1994), which places strong emphasis on
working with sets of spectra from the same sample.
result
Barradas and co-workers publisbed algorithms to improve the
simulation of buried .resonances (Barradas et al., 2006), double
scattering at grazing angles (Barradas, 2004), the yield at very
low energies (Barradas. 2007). and the shape of surface signals
in high-resolution experiments (Barradas e/ aI., 2007a). As of
the time of publication of this handbook, these effects are
included only in NDF, but other codes can easily include them
~~
~,
~,;
\.
.
~f'J.
14.6.2 SlMNRA
.Ii'
Development of SIMNRA started in 1996. It is a Windows l'"
code with intuitive menus and windows.
.~,
,.1
326
FIG. 14.12. First step of a SlMNRA analysis: load the data and define the experimental conditions.
327
Chapter 14
The second step is to define the physics included in the calculation. In this case, all isotopes are calculated separately (as opposed to
making one single calculation for the average mass of the target element), Bohr straggling with the ChuIYang correction and including
multiple scattering is calculated, SiuM-2003 stopping powers are used, the scattering cross section for the l'O(a,a)"O is calculated
with SigmaCaic, and Andersen electron screening (Andersen et aI., 1980) is calculated for the other elements .
.~
FIG. 14.13. Second step ofa SMINRA analysis: Configure simulation parameters.
328
Finally, the sample is defmed: Layers are created, with any number of elements. The user inputs the thickness and concentration
values, as well as roughness parameters if required. Then, a simulation is generated and compared to the data. The user can then
iteratively refme the sample description until a good fit is reached or use a fitting routine to adjust the thickness and concentration of
one layer at a time.
FIG. 14.14. Third step ofa SIMNRA analysis: define the sample structure, perform the simulation. Here the final result is shown, after
manual iteration until agreement between simulation and fit is achieved.
14.6.3 NDF
!,ih _
329
Chapter 14
n
C!
be
If
"
.. FIG. 14.15. First step of a NDF analysis: load the data and define the experimental conditions, configure simulation parameters.
~
,.
<
\'
..
"
330
'.
~;
',~
.\.
,,
The second step is to define which elements exist in the sample. This is done by creating a structme file with logical elements, which
can be simple elements such as Si, molecules such as Si02 with a given density, or complex molecules with unknown stoichiometry t(J
be determined in the fit Restrictions on the depth and concentration ranges where each logical element can exist might also be given.
If many completely different samples are tobe analyzed in batch mode, one structurc flie is created for each sample. If man)' similar
samples are to be analyzed, the same structure file is used for all of them.
'ndffOl01.dat.Expt
,.- ~dff0101.d.at.~f
FIG. 14.16. Second step of a NDF analysis: defme which elements or molecules are present in the sample.
331
b......
Chapter 14
The final step is to run an automated fit. In this case, the user does not have to define a layer structure; rather, this structure will be an
output of the fit. The algorithm used, simulated annealing, tries. to find the simplest structure consistent with the data. AI; the
calculation is completely automated, the user must check the results, because it can find unwanted solutions. Alternatively, the user
can also make an iterative analysis by defining a layer structure in the usual way. This can also be the initial guess for a local search
on all parameters.
FIG. 14.17. Third step of a NDF analysis: run a fully automated fit, or, alternatively. perform a iterative analysis.
14.6.4 Issues
supported. This means tbat users <::an rely on advice and help
from the developers.
332
333
Chapter 14
Time-of-flight
1GMaV"Cu*
Tart(3OnmVSi~100nm)1Si
81191.... 3:8.&.,19.:2" .
,!
~
'J
FIG. 14.18. Time-of-flight ERDA ofa SifSiO, (100 nm)!TaN (30 nm) sample measured with a 16 MeV "eu beam.
.~
!\
;j
:J
The code GeoPIXE was developed more recently (Ryan et fi
%.1','
;~
(,
.~
.~
14.8.1 PlXE
Over the years, the PIXE community developed its own
)t~
lj
:;;l
.,
:t!
334
:~
Neoq<.
CU .....
CU slmu ~ 1m
Nnw.aZinn
-.NdJaLS3nn
-Namu.Unm
'000
...
4.G
Oe=~~~--~~~~~
1Q.Q 10.5 ".D 11.5 12.0 12.5 13.0 13.5 , .... 0
Energy (MeV)
8CIO
7IlO
~soa
11-
"Ii
c
~15OD
"
~1000
~S-
SI .....
51""" B2 rill
2IlO
...
'00
0
ao
'200
as
4.
4.5
"0
5.5
S.G
....
I.!
Energy (MeV)
16 MoV '"cu =>1IIN(un}lSlO,(11lOm1)1S1
'000
e
EmrgJ(MeV)
Trtstnu. v rm
. 0 .....
2IICI
TaupL
Ii
lam
r~
-ORInTaN
0 t05In 'Ta\!
"
8-
8--
Ii
5800
loa
50
200
Oanv.D2nm
- - OftiTTaN
Enorgy (MeV)
\.
a.75
flO
1.2!
8.!1
Energy (MeV)
FIG. 14.19. MCERD analysis ofTOF-ERDA ofa SilSi02 100nrnfTaN 30nm sample measured with a 16 MeV "Cu beam.
to GUPIX. this approach is geared toward spectra and not twodimensional maps.
between the analysis of the RBS and PIXE data, and it also does
not fully tak.e advruitage of the combination of the depth
sensitivity of RBS with the elemental and trace sensitivity of
PIXE.
335
Chapter 14
The nuclear reaction is often a (elastic or inelastic) particleparticle reaction or a particle-gamma reaction. These techniques
are often called resonant NRA and particle-induced gamma
emission (FlOE), respectively. Different codes have been
A new rnA data format, called IDF. that aspiTes to support all
forms of rnA data in a well defined simple format permitting
easy exchange between sites and analysis programs, was
presented in the 19th International Conference on Ion Beam
336
ACKNOWLEDGEMENTS
We thank Mike Thompson, Matej Mayer, and Kai Arstila for
providing the figures for the RUMP, SIMNRA, and MCERD
examples, as well as all those who commented on early drafts.
Technical contact
Current status
Distribution mode
program
Status of
source code
Operating
syslem(s)
DEPTH
Edit Szilagyi
KFKl Research Institute for Particle and
Nuclear Physics
Budapest, Hungary
szilagyi@nnki.klki.hu; www.kfki.hU/-ionhp/
Active
development
No charge;
downloadable: from
Web site
Restricted to
author/not
available
DOS and
Windows
GISA
Eero Raubala
University of Helsinki, Helsinki, Finland
and Jaakko Saarilahti, Technical Research
Center of Finland
Eero.rauhala@helsinki.fi;
laakko.Saarilahti@vtt.fi
Kai Arstila
IMEe, Leuven, Belgiwn
kai.arsti.Ja@iki.fi
In use
No charge; "",'lite to
author for copy
Restricted to
author/not
available
DOS (or
emulators)
Active
development
No charge; write to
author for copy
Source code
available
Linux
(Windows)
Nuno Barradas
Technological and Nuclear Institule
Sacavem, Portugal
nunoni@itn.pt;
www.ee.surrey.ac.uklResearchiSCRIBNndf7
Active
development
Commercially available
through Univ. of
Surrey; evaluation
copies available upon
request
Windows,
DOS,
Endre K6tai
KFK.I Research Institute for Particle and
Nuclear Physics
Budapest, Hungary
kotai@rrnki.kfki.hu
Mike Thompson
Dcpt. of Materials Science, Comen University
Ithaca, NY USA
motl@comell.edu;www.genplot.com
Active
development
No cbarge; 'Write to
author for copy
Restricted to
author/not
available;
some
modules
open source
Restricted to
author/not
'available
Active
development
Commercially available
through Computer
Graphics Service;
evaluation copies
available from Web site
Commercially _available
through MPI for
Plasma Physics;
evaluation copies
available from Web site
Source code
available
Windows,
Linux,
MCERD
NDF
(DataFumace)
RBX
RUMP
SIMNRA
Matej Mayer
MPI for Plasma Physics
Garching, Germany
MateiMayer@ipp.mpg.de;
www.rzg.mEgdel-marnl
Active
development
337
UNIX
Windows
UNIX,OS2
Restricted to
author/not
available
Windows
Chapter t4
program
Nature of
the primary
Primary simulation
Graphical
output
modes
DEPTH
interface
Interactive
Manual iteration
N/A
GlSA
or batch
Interactive
Manual iteration;
Screen only
automated parameter
search
MCERD
NDF
(DataFumace)
Batch
Batch
directed or
Manual iteration
N/A
Publication
Manual iteration
Draft
quality
quality
interactive
RBX
Interactive
RUMP
Interactive
SIMNRA
Interactive
Manual iteration;
Publication
automated parameter
search
quality
Manual iteration;
Draft
quality
automated parameter
search
Cross sections
DEPTH
ZBL'95
GISA
MCERD
NDF
(DataFurnace)
per ion/element
RBX
RUMP
SIMNRA
SRIM-1997, SRlM,orKKKNS
User-defiDed" correction per layer and ion
338
Isotope calculation
Single
i~otope
Screening calculation
Yes
Plural scattering
Multiple scattering
Geometric
stmggling
Channeling
None
Yes
No
None
None
No
No
Full Me calculation
Full Me calculation
Yes
No
No
calculation
None
Yes
Defect
calculation;
simulation of
channeled spectra
None
None
No
No
Yes
No
approximated as Gaussian)
Straggling
models
Bohr, Chu, Yang,
Tschalilr
Natural abundance
andlor specific
isotopes
Energy/angle; external
tables by users
Bohr +
Natural abundance
and/or specific
isotopes
No
Natural abundance
and/or specific
i:mlopc:)
Natural abundance
and/or specific
isotopes
Energy/angle;
Andersen and L'Ecuyer
RUMP
Natural abundance
andlor specific
isotopes
Bohr
SIMNRA
Natural abundance
andlor specific
isotopes
Energy/angle;
Andersen and L'Ecl1yer
TschaHlr
OlSA
MCERD
NDF
(DataFumace)
RBX
LindhardlScharff
!
!
Yes
TschaUir
~ter14
Analysis
program
DEPTH
All
IBM, Cornell
Pileup
correction
No
,
Stopper foils
Energy calibration
Energy-dispersive
(magnetic spectrograph
coming)
Simulated; including
Linear
Detection system(s)
inhomo~eneities
Quadratic
GISA
All
RBS
IBM
No
Energy-dispersive
N/A
MCERD
All
ERDA,RBS
IBM
No
Energy-dispersive;
TOF; different layers in
detector for gas
detectors, for example
Simulated; equivalent
treatment to sample
Linear
NDF
(DataFumace)
All
mM, Cornell,
Yes
Energy-dispersive
Simulated; equivalent
treatment to sample
Quadratic; varying by
ion species
RBX
All
RBS,ERDA,
nomesonant NRA, PIXE,
resonaot NRA (PIGE,
NRP)
RBS,ERDA,
nomesonant NRA
IBM, Cornell
Yes
Energy-dispersive
Simulated;
homogeneous foils only
Linear
RUMP
All
RBS, ERDA
Cornell, IBM,
General
Yes
Energy-dispersive,
partial TOF
Linear
SIMNRA
All
RBS,ERDA,
IBM, Cornell,
General
Yes
Energy-dispersive,
TOF, electrostatic, thin
solid-state detectors
with transmission of
particles
Simulated; equivalent
treatment to sample
Quadratic; varying by
ion species
General
<!';
nonresonant~
~"""':""""""'~""<",,"o.lro>'''';''''"''.''''~,"'''''''''''P .
.J
Table 14.7. Fitting capabilities DEPTH and RBX have no fitting capabilit~~).
Analysis
program
GISA
Starting conditions
Reasonable guess
Optimization
method
Error estimation
Xi
None returned
Searchable
experimental
parameters
Autorefinement
oflayers
None
No
One layer
at a time
Poi~~on
Yes
A11 parameters
variable, user
controls which
ones change
Poisson
No
No internal limit,
practical of 30
parameters at a
time
Poisson
E e l, charge
No
One layer at a
time, all
characteristics
Statistics
used
Limitations
minimization
NDF
(DataFurnace)
Simulated
annealing
plus grid
search
RUMP
Reasonable guess
Marquart
search
x:
SIMNRA
Reasonable guess
Simplex
search
!,
I,
Table 14.8. Sample definition and complexity handled,
Analysis
Sample description
Continuous profiles
Substrate
roughness
Slab limitations/elements
I'
, "
program
Layer
roughness
None
10 layers; 10 clements; user can
define so layer"'element is constant
No
No
No
No
Yes; approximated
as energy
broadening
Yes; approximated
as energy
broadening
None
No
No
None
No
Yes; single- or
dual-sided, all
interfaces possible
Slabs only
Yes; Lorentzian or
Gaussian angular
distribution
DEPTH
GISA
Layer definition
Layer definition or profile
function
Slabs only
Maps continuous profiles onto
slab structure'
NDF
(DataFumace)
REX
Layer definition
RUMP
SIMNRA
ii
II .
~
I,:
Chapter 14
Ion
Energy
Scattering
Target clement
Effect on spectrum
Error
Effect on analysis
al1sl~
cross section
--electron'screening--lXil---------Aii----------<is o ------------Aii-----------------D~-rease(fyiefdAll
~ 150
> I SC
Heavy
keVIZ J
All
>.500
>90'"
Heavy
--_. -Unk-nown
(iarge) ---------------------------f~ts_e ~u_a_'!t!~l!t~ye_1!t}1!!:t~i..s_ . 1%
~.2%
.. _... ____ .. ______ . ___ ... _____________If~Y-'?_l ________ . _. _________________ .. ____ . ___________________ . _... ________ . _______________________ . _. ___________________________ . ______ . _. _. _. ____________ .
nuclear effects
2:100 keY
All
>2.0 MeV
All
Light (medium
heavy)
Light
Increased or decreased
yield
stopping power
. ______ _. ______________________ ___ __________________________ . __________ _. _._. _____________________________ _. __ ___________________ ~!.l.!'ll'~!!!. _ _. _. ___ ._._. ______ .
general
uncertainties
All
All
Si
All
Li
Heavier
ions
Bragg rule violations
--ph;;i~;i;~te~ffe~tS-
energy spread,
All
4,2% average
(>10% for 13% of target atoms)
4.1% average
(>10% for 11 % of target atoms)
9%
5.1% average
(>10% for 17% oflarget atoms)
6.1% average
(> 10% for 18% of target atoms)
-Ail-- _... ----- --- -- ------ -- -... --- -- -------_. _. --- i 5%" --- ---- _... -. ----- -- -" ------- -----. _. _. _.. ---. ---------------_. _.
.______ . _. _. ______ ._._._,._._. ________ .______0!1~l'J~~C!'JO_______ . _..l:teJ!i..s_,!l)~_ ~i~.t~~ ______ ......... __ . __ .. __________ . ______ _. _._. _______________________ . _. ___ .' __ .. _. _. _.
--Lightc~-~p;~~d~-
All
Heavy
Grazing-angle
geometry
Nonna1
incidence
All
mUl1iplc scattering
surfnce roughness
Increased or decreased
yields and widths
All
All
-i~~~;s-e-d ~~'d~~~~~~d
Increased or decreased
yields and widths
Increased or decreased
edge and peak broadness
-~
o~i
:55~ 1 0%
Large
Small
All
broadne!l~
channeling
All
Incidentally
Decreased yields
Increased or decreased
;rie\ds l backsround
Increased background
aligned
gcornet!),:
pulse pileup
All
low-energy
background
All
Heavy (all)
Lew (all)
Grazing
geometry (all)
Heavy (all)
This list of web sites rele",ant for IBA software and data analysis is by no means complete. Several of the codes mentioned in this
chapter have no dedicated Web site.
Corteo
hnp:llwww.lps_umontreal.ca/~schiette/index_php?n-Recherche.Corteo
http://www.kfki.hU/-ionbp/dodprog/wdepth.htm
Windows-based program for calculating the energy ~d depth resolution of ion beam analysis methods such as
Rutherford backscattering spectrometrY (RBS) and elastic recoil detection analysis (ERDA)
FLUX
http://members.home.nllp.j.m.smuldersIFLUXlHTMll
Program for simulating the trajectories of high-energy ions in single crystals in channeling, or near-channeling,
directions
GeoPIXE
bttp://www.vrop.csiro.aulGeoPIXE.html
Software package that perfonns real-time particle-induced X-ray emission (PIXE)/synchrotron X-ray fluorescence
(SXRF) spectral deconvolution using dynamic analysis
.
GUPIXI
GUPIXWIN
bttp:llpixe.physics.uoguelpb.ca/gupixlmain!
Software packsge for fitting particle-induced X-ray emission (PIXE) spectra from thin, thick, intermediate, and layered
specimens
IBANDL
http://www-ndsjaea.orglibandli
Library of experimental data on nuclear cross sections and excitation functions relevant to ion beam analysis (IBA);
resulted from the merger Sijffi]aBase and NRABASE
IBIS
http://www.klki.hul-ionbp/
Collection of resources for the ion beam community (called the Ion Beam Infonnation System) that includes links to
several programs and other useful sites, as well as an archive of documents
IDF
bttp:l/idf.schemas.itn.pt
Defmition of the IBA data fonnat IDF including documentation and examples
LibCPIXE
-,<-.
~~----~-~-~-------
MSTAR
http://www.exphys.uni-linz.ac.atistopping/
Collection of data on stopping powers of light and heavy ions
NDF
http://www.itn.ptifacilities/lfJind'ukJfi_ndf.htm
General-purpose cD.de for the analysis of Rutherford backscattering spectrometry (RBS), elastic recoil detection
analysis (ERDA), particle-induced X-ray emission (FIXE), non-resonant nuclear reaction analysis (NRA), and neutron
depth profiling (NDP) data for any ion, any target, any geometry, and any number of,pectra
http://www.ee.surrey.ac.uklSCRIBAlndfi
Fitting code (not a simulator) that extracts elemental depth profiles from Rutherford backscattering and related ion
beam analysis spectra
DataFurnace
QXAS
http://www.iaea.org/OurWorkiSTINAlNAAlJpcilinshrrf/poiXRFdown.php
.
DOS-based software nacka.e (called Quantitative X-rav Analvsis Svsterol for auantitative X-ray fluorescence analysis
RUMP
http://www.genplot.com
Graphical data analysis and plotting package that pro-vides comprehensive analysis and simulation of Rutherford
backscatterin~ spectrometry-(RBS)-and elastic recoil detection reRD) spectra
Sigmacalc
http://www-nds.iaea_orgfsigmacalC/
.
Service providing evaluated (recommended) differential cross sections for ion beam analysis
343
Chapter 14
SIMNRA
http://www.rzg.mpg.del-mami
Windows-based program for the simulation ofbackscattering spectra for ion beam analysis with megaelectronvolt
ions, mainly intended for non-Rutherford backscattering, nuclear reaction analysis (NRA), and elastic recoil detection
anal sis ERDA
SRlM
http://www.srim.org
Group of programs that calculate the stopping and range of ions in matter using a quantwn mechanical treatment of
ion-atom collisions
Doyle, B.L., Provencio, P.P., Kotula, P.G., Antolak, AJ., Ryan,
C.G., Campbell, J.L., Barrett, K. (2006), Nucl. lnstrum. Methods
B249,828.
REFERENCES
Albertazzi, E., Bianconi, M., Lulli, G., Nipoti, R, and Cantiano,
M. (1996), Nuc!. Instrum. Methods B1I8, 128.
Alkemade, P.F.A., Habrakeu t F.R.P.M., and van der Weg, W.F.
- (1990), Nucl. Instrum. Methods B45, l39.
Blaauw, M., Campbell, 1.L., Fazinic, S., JakSic, M., Orlic, I., and
Van Espen, P. (2002), Nucl. Instrum. Methods B189, 113.
Bohr, N. (1948), Milt. Fys. Medd. K DW!. Vidensk. Selsk. 18 (8).
B0rgesen, P., Behrisch, R-, and Scherzer, B.M.U. (1982), Appl.
Phys. Al7, 183.
http://www.ee.surrey.ac.uklIBC/ndflwindfinanuaI70.pdf.
Jeynes, C., Barradas, N.P., Marriott, P K., Boudreault, G." .
Jenkin, M., Wendler, E., and Webb, R.P. (2003), 1. Phys. D: .
Appl. Phys. 36, R97.
Jeynes, C., Peng, N., Barradas, N.P., and Gwilliam, R.M. (2006),
Nud. Instrum. Methods. B249, 482.
Boudreaul!, G., Jeynes, C., Wendler, E., Nejim, A., Webb, R.P.,
and W~~en, U. (2002), Surf Interface Anal. 33, 478.
Butler, J.W. (1990), Nucl. Instrum. Methods 845,160.
Chu, W.K. (1976), Phys. Rev. A13, 2057.
344
Raubala, E., Barradas, N.P., Fazinic, S., Mayer, M., Szilagyi, E.,
and Thompson, M. (2006), Nucl. Instrum. Methods B144, 436.
B68,300.
Ryan, e.G., Cousens, D.R., Sie, S.H., Griffm, \V.L., ud Suter,
G.G. (1990), Nuc!. Instrum. Methods B47, 271.
Ryan, e.G. (2000), Int. 1. Imaging Syst. Technol.ll, 219.
Mateus, R., Jesus, A.P., and Riheiro, J.P. (2005), Nud. Instrum.
Methods B229, 302.
Maurel, B., Amsel, G., and Nadai, J.P. (1982), Nud. Instrum.
Methods 197,1.
345
CHAPTER
N. P. Barradas
Instituto Tecnologico e Nuclear, Sacavem, Portugal
Centro de Flsica Nuclear da Universidade de Lisboa, Lisbon, Portugal
Contributors: M. Dobeli, J. A. Davies, W. N. Lennard, and I. V. Mitchell
.'
15.1
_.
w __
.-
._-~
..... -
__
. _ -
-,--"
INTRODUCTION ....................................................................................................349
347
.-~
...
#_~.
Chapter 15
..................................................:............................................................ 380
348
15.1 INTRODUCTION
Ion beam analysis (iliA) IS a quantitative analytical
technique, and in this chapter, we intend to show how to avoid
many pitfalls when determining elemental depth profiles
accurately with light-ion Rutherford backscattering spectrometry
(RBS) using megaelectronvolt ion beams. IBA can, of course,
use various other beams and use these for a variety of other
purposes, including profiling of crystalline defects. The present
discussion is usually, but not always, easily generalized to these
other cases. We will therefore also cover pitfalls in a number of
other important examples.
We will, in general, be guided by Chapter 12 of the previous
edition of this handbook (Tesmer and Na~tasi, 1995). However,
that chapter contains many subtle discussions that we will avoid
or simplify, referring the advanced IBA user back to it. Today,
most laboratories tPat are starting to usc IBA have the new
generation of accelerators, eliminating many of the problems
associated with the older accelerators that the previous chapter
addressed.
A new generation of software is available that renders
obsolete many of the nwnerical examples of the previous
handbook. We will assume that users interested in accurate work
will have access to one of the codes described and evaluated in
the 2006 International Atomic Energy Agency (IAEA)
sponsored IBA software intercomparison ex:ercise (Barradas et
- .
-~~~~.'~
- .
need to know the product of the charge and the solid angle. Qfl.
This is detennined from the Si signal and the known energy loss
of the He beam in a silicon matrix. For backscattering with the
detector at an angle of 8 to the beam and normal beam incidence,
the surface yield. Yo (in counts/channel), for element A in matrix
AB is given by the equation
..
.
.--~
~g
.. ------ _ . .
-----.-~-.-
.. --~----
..
~--
- -...
~----
349
Chapter 15
5O
a
'. <:hanhei
.<50
40
..
450
t:
"o
.".
320
Channel
420
. Depthn,"
240
..
d
,-ao
FIG. 15.1. Accurate analysis of As implanted fluenee in Si. Ca) Nominal implant 5 x 10" As/cm' at 100 keY. L5 MeV He RES
spectra from two detectors collected simultaneously, and 0.5 mC collected charge, with fits. The beam is incident normally on
sample such that it channels in the substrate. and the near-surface region of silicon amorphized by the ~jon beam is clearly seen.
small 0 peak from the surface oxide is also clearly distinguishable. The As, Si, and 0 edges are marked for the detector with the
solid angle. (b) Detail of the As signal. (c) Expanded detail of the As signal, with the nonlinear pileup correction function
Pileup-corrected As signal from each detector plotted on a concentration versus depth scale (using a Si density of 5 x 'n"I,,~'\
(Reproduced from Jeynes et al., 2006.)
350
Sl
351
.""
i~
Chapter 15
for different beams and one really needs to see the spectrum to
set up the SeA correctly. Users have also found that a gate is
needed to exclude edge signals from the detector. It is possible,
but not easy, to obtain excellent precision (1 %) with such a
design.
desired particle,
Transmission detectors take a fraction of the beam, of course,
and this fraction must be measured directly with a downstream
One of the simplest ways to test for a mixture of "equalmass" beams in single-ended machines is to detect elastically
scattered projectiles from a thin self-supporting Au target. An
energy spectrum analysis with a standard surfacebanier
detector system will readily distinguish the fragments of
molecular im,Euritics, such as mixtures ofR3+, HD+, and JHe+ or
mixtures of He" and D 2", which can arise through ionsource :.;
. memory effects.
1-
.~
:'~
- - - - - - - - - - - 'j
Table 15.1: Relative particle yields and energies for several .~
~
H
0.33 MeV
0.67 MeV
'He
1.0 MeV
.~
o
i
,
";':
H,'
27
"
IID+
_ _ _ _ _ _ _ _9_ _ _ _ _ _2_._2_5________
3
He
------------------------ i
.>::->
'~~
352
The molecular interferences are not problematic for tandemtype accelerators, where there is an injection magnet at the
source, an analyzing magnet following acceleration, and a'
(collisional) stripping event at the high-voltage tenninal.
.
Wall scattering can be minimized using a series of apertures;
however, care must be exercised so as not to restrict beamline
pumping speed excessively_
ns. However, where PUR is used, any pileup events that are not
rejected are considered nearly simultaneoWi, that is, the
amplitude of the resulting sum pulse is nearly the sum of the
amplitudes of the pileup pulses. Where PUR is not used, the sum
pulse can be of (nearly) any amplitude.
It is possible to eliminate pileup almost completely by
detennining both the momentum and the energy of the detected
particles using a AE-E detector telescope (Gurbich, 1996) or
using a time-of-flight detector with a pulsed incident ion beam
(Gurbich and Komilov, 1991). However, these approaches are
significantly more complex, and the llE-E method requires very
the
353
--~
Chapter 15
Au
.C
Mg
i1
o .
.12
~~l
.~.
:.
. .. .
,.
..
..~
...
o.
512
Channel
FIG. 15.3 .. Pileup with pileup rejection. Data obtained from 707 keY 'It RBS of a AulMg multilayer on C, with a count rate of 4.1
kHz, an amplifier shaping time of 500 ns, and a PUR time resolution of 550 ns. A distinct 0 contamination peak can also be seeD. Fit
including pileup calculated with Wle1opolski and Gardner's algorithm is as a solid line.
j
,.
_. .:.~.,
>.J
.i
7.
.,~
~
,}
),'J
'.~
354
,)
. Jtk:!...........
355
Chapter 15
15.3,1 Energy
Accelerator energy calibration is critical for accurate work.
Clear pitfalls here were highlighted in the first edition of the
___ handbook (also s~e. Section 15.3.2), which pointed out that the
fust RBS round-robin experiment had vel)' disappointing 'results
partly because of very poor energy calibration (see Section
15.3.2.4). For RBS, the spectra have similar shapes for all
energies. Of course, the energy loss is sensitive to the beam
energy, but it is not nearly well enough known for use in energy
calibration. In any case, the energy loss for any given layer can
rarely be measured with a precision exceeding I %. However,
because the cross section is proportional to llE2 but linear in
gain, an energy error wilJ propagate through to errors in the
determined layer thicknesses. For elastic (non-Rutherford)
backscattering spectrometry (EBS) [and nuclear reaction
analysis (NRA), there are very often strong resonances that, in
some circumstances, can lead to very large errors consequent on
an energy calibration error. PIXE is the one case where energy is
a second-order effect, and one often flnds that PIXE laboratories
arc rather cavalier about energy calibration. This is a potentially
an)
356
tr
Table 15.2. Primary (recommended) accelerator calibration points for protons [from Marion (1966), except where
otherwise stated]. Note that, for threshhold reactions, the 2003 masses are used (Audi et aI., 2003); also see the
cODvenient calculator at www.nndc.hnl.gov/qcalc.
Energy
(keY)
340.46
872.11
991.90
1747.6
1880.60
1880.443
1880.356
3235.48
4234.43
5796.9
5803.73
5803.621
6460.47
7026.56
9193.85
9510.55
12728
14233
14230.75
19641.9
0.04
0.02
0.04
0.9
0.07'
0.020'
0.081
0.29
0.31
3.8'
0.12'
0.100
0.15
1.6
0.20
1.46
10
8
0.02
1.1
Reaction
fwhm
(key)
19F(p,ay)160
I'F(p,ay)160
27 Al(p,y)"Si
IJC(p,y)I'N
'Li(p,D)'Be
'Li(p,n)'Be
'Li(p,n)'Be
IJC(p,n)I'N
I9F (p,nt'Ne
27Al(p,n)" Si
" A1(p,n)"Si
27Al(p,ni7Si
"S(p,n)"Cl
"'Ni(p,n)6IlCu
"Fe(p,n)"C
s8Ni(p,ni 8Cu
I6O(p,p)160
l2C(p,p)12C
12C(p,p)12C
12C(p,n)12N
2.4
4.7
0.1
0.8
Comment
Marion, 1966
White ef al., 1985
Marion, 1966
White e/ 01., 1985
<2
<1
1.2
0.5
0.6
0.5
0.14
1.7
1.5
1.5
033
0.612
0.32
1.6
0.67
0.67
0.36
. Reactfon--.:"- ._ ..,,....,
"C<'l"1e,n)"0
l2C('He,n)I'0
12C ~e,n)140
"c( e,n) 14 0
'Li{'He,n)'B
'Li{'He,n)'B
'Li('He,n)'B
l'O(,He,n) 18Ne
'Li(a,n)l'll
I'N(a,n)I'F
'Li(a,n)'B
1'N(a,n)I'F
1'C(a,n)uO
l60(a,n)I'Ne
.. - -.--....."- Comment
357
.I~ ..
~ ~~
Chapter 15
Table 15.4. Secondary energy calibration points based on SigmaCalc, version 1.6. Note that uncertainties were derived from
the original literature, not from SigmaCalc. and are standard deviations, not fwhm.
Energy
(keY)
3044
3885
4265
3470
1483
1748.5
3198
1734
Reaction
Reference
(keY)
5
5
5
16O(a,a)160
"O(a,a)I'O
"C(a,a/'C
"O(p,p)I'O
"Mg(p,p)''Mg
1'N(p,p)I'N
1'N(p,p)I'N
0.2
12CSpp'tC
fwhm
10
2
27
2
0.3
9
6
45
{1
i~
instruments.
358
the
359
-.~.-
Chapter 15
.:
I
~
:1
.,
,j
~'1
:;1
;
:1
.'4
l
..
i~
~
'-."'1I
""11
t:.
fr:.
.1)
ctiw-,
.y~\'
360
-----
true
361
~ ..
Chapter 15
ble accuracy for the electronic gain, .6., was only 0.5%. Lennard
e/ al. (1999) did not make a comparable accuracy estimate
explicitly, but their work is also consistent with an accuracy in A
of about 0.5%. The extremely accurate analysis of Munnik et al.
(1996) includes a full PHD correction, and Gurbicb and Jeynes'
(2007) benchmark measurement of the ""Mg(p,p)m'Mg elastic
scattering cross sections depends on the analysis of a PHDcorrected set of spectra with a single gain across the whole
energy range (see Fig. 15.3 for one of these spectra).
necessary to account for this aspect ill detail. ThIS should be :~'i
done as a matter of course for PIXE, where the detectors have ,,~
relatively high resolutions (typically <200 eV) and the shape of
the instrumental function is well-known to be non-Gaussian. The'.'~
GUPIX program (see Blaauw e/ al., 2002), for instance, assumes X
that the instrumeotal function is specified in detail by the user.. :~
For PIXE, the instrumental function must be measured as a :~
function of X-ray energy over the whole energy range using a S1~:.
series of pure targets. If this is not done, the ratios of the vari~us '~.;.
K and L lines cannot be detennined with precision. Particle :."
scattering detectors have much lower energy resolution
(typically -15 keY), and the instrumental functions of these ,t;,
detectors are not usually measured in detail, although remarkable ..~
results can be obtained when it is. Fischer e/ of. (1997), for. ~,
example, demonstrated that they could reliably resolve the. i~
isotopes of Co using a full Bayesian deconvolution code! ,"!,~
However, none of the routine IBA codes currently have facilities~ ~~~
for a particle scattering detector resolution that is a function of t
energy, and the highest-resolution work yet reported with these :)'_
detectors (determining interface roughness of <1 nm) used 8.:::'-;
simple Gaussian (symmetrical) function (Barradas, 2002).
...""
It IS
.:ii
,::..1:.
..~l~.
~
362
J~t.'
160
Vl
C
::::J.
O
()
a
Channel
160
160
350
160
'"
E
:::l
d
290
Channel
330 290
. Channel
330 290
Channel
330
FIG. 15.4. Detector resolution and its effect on the calibration of gain. (a) Spectrum for 1.557 MeV 4 He+ RBS with scanering at
149.2". Sample is Au/Ni/SiO,/Si (l2.8Il4.8/l607 TFU). (b) Delector resolution of 25 keY, fitted gain for fixed metal thickoess =
4.457 keV/channel (fitted metal !bickoess = 28.55 TFU); (c) 24 keY, 4.472 keY/channel (28.07 TFU); (d) 26 keY, 4.460 keVlchannel
(28.96 TFU).
asslllDption is valid and how well the user can calculate the
variation of the energy resolution with depth.
15.4.1 The Rutherford cross section
How accurate is the Rutherford scattering law? This is often
the key question in achieving quantitative RES analysis, because
363
N-k
Chapter 15
.. 0
o.
If the energy is below the lower limit for RBS, then the major
part of the collision takes place before the projectile has
penetrated all of the electron shells of the target atom; that is, the
unscrecned Coulomb field assumed in Rutherford scattering
plays a minor part in the collision. With megaelectronvolt
accelerators and surface-barrier detectors, this almost never
happens, even for the heaviest target nuclei. However,
backscattering studies in the medium-energy ion scattering .'-;
(MEIS) regime involve energies well below the limit for
Rutherford scattering. For quantitative work, suitable algorithms .~
are given by Mendenhall and Weller (1991); however, they
point out that a fairly accurate (2-3%) approximation to the .. 1.
screened Coulomb cross section can still be obtained by
applying the simple screening correction of Eq. (15.3).
"
.,,~
:1
I
,
(15.3)
~.~:~
A
~~
If the energy exceeds the upper limit for RES, then we are in
the regime
of non-Rutherford elastic backscartering
spectrometry (EBS). As examples, proton scattering on C is
significantly non-Rutherford at 380 keY, and He scattering on C
is significantly non-Rutherford at 2.2 MeV. The rough estimates
364
365
I
h .. _ .
Chapter 15
SOOkoV
-"'"
"0Il0r;::=====~-~~""l
- - SlMNRA. "'"
------ SIMNRA. $lodll
--11UM.SP
10000
sooo
,
,
,,
'
-'
Jf
'"
H.:
Energy (keV)
FIG. 15.5: Double scattering effecls in 'He RBS of Au on Si.
Comparison of the experimental energy spectrum (dots) at a
polar emission angle of 15" for 0.5 MeV "He. The histogram
gives the TRlM.SP result; the dashed and solid lices represent
the single-collision model and the dual-<:ollision approximation,
respectively, both calculated with SIMNRA. The lower part of
the figure shows the background between the Si edge and the
lower Au edge in more detail. (Reproduced from Eckstein and
Mayer, 1999.)
J\
't
,-
;-:':~
.&
~t
"\j;
.:;~
366
367
-,
Chapter 15
Table 15.5. Uncertainty Budget. Coverage factor k = I. (Reproduced from Jeynes etal, 2006)
TypeAorB
Cornell
IBM detector
Comment
2.60%
0.80%
2"10
2%
0.28%
0.47%
0.08%
0.13%
ScaHering angle ~. -
0.28%
0.07%
i,
,I
From shape-fitting accuracy
Electronic gain
0.5%
0.5%
Pileup correction
0.05%
0.02%
Relative uncertainty
0.64%
0.70%
0.48%
averagel.fi
:.--
Beam energy
0.20%
0.16%
Screening correction
0.54%
Relative accuracy
0.2%
ERD
Si stopping power
0.6%
0.83%
Absolute accuracy
depth profiles.
We say that an optimal solution is a depth profile from which
a spectrum can be calculated that fits the data well and that a
spectrum is ambiguous where more than one optimal solutions
exist.
For the AF, implant of Fig. 15.1, for example, one element in
the state space is a pure As sample, another is a pure Si sample,
and there are a very large number of intermediate elements. The
state space is large, but it is not infinite because the energy
368
detector
Pileup correction
I,i
369
..
Chapter 15
(.)100
,Cb)
1
Orlglilal
80
..
'#
40
+Ir~\~v-~,~'~~-'~===:I~~
20
o
EI,,--;;1f\~.!l~~.~.="=:'j"
o
2500
ThIn FIlm
100
'5000
UnJ"
(e)
(d)
60r-~~--------------~
S0t--1~--------------~
.-
30
t"":.:.....t=-"""'''';;::-------1
: t=t;~:;~rP=--_l
__
c-
. .. l ..-...
"-,
40t-~~--------------~
o~~~~~~
-- -
C,"lJl1al
o.\oInoI
- - 8"""",, 2
~-,'
-~.1
5000
FIG. 15.6. Butler's (1990) ex~ple of an o,udiz.ed NiCrAI alloy re-analyzed (reproduced from Jeynes el 0/., 2003). Ca) The original
profile from which the spectrum was calculated; (b) spectrum (symbols) and fit (line); (c) atomic profile fitted to data assuming
molecules and complete oxidation from the surface, using two spectra at different detector angles and excluding alwnina from the
surface; (d) comparison with the original profile of the Cr profile calculated under various assumptions. Specifying only elements
barely constrains the profi1e, and even with two detectors, the profile is not recovered at intermediate depths. Using only one detector
with the assumption of molecules is also not sufficient. The molecules used .are NiO. Cr20}. AhO), and NiI9~CrHI~119.
It turns out that the Cr profile is the most sensitive to the prior
assumptions of the analysis, and Fig. 15.6(d) shows the Cr
profiles obtained under four different asswnptions. To retrieve
Butler's initial proftle unambiguously, one must specify not only
the molecules present, but also the following conditions: only
oxides are present near the surface, oxygen is excluded from the
substrate, A1 is excluded from the near-surface region. and two
independent spectra were recorded (at different scattering angles
in our example). Butler did not point out this last condition for
this example, although he noted that, in general multiple speclTa
are always a help. These particular data are very ambiguous:
many different solutions, all of them as good as that shown in
Fig. 15.6(b), can be obtained without aoy of the conditions
mentioned.
:'i
J
.~
OJ
'1
'l,
There are many alternative ways to fit any given set of~ .:;.i
If a good fit to the data bas been found, it might represent a valid .J
'~11
,1
370
.
j{
,it.
,
solution. Whether or not it is valid depends on whether correct
parameters have been used. Sometimes, this might be an
intricate question if one is interpreting the detail of a spectrum.
Even if the solution is valid, in the presence of ambiguity, it is
not necessarily true. Even if the solution is not false, it might not
be useful in the case where Occam's razor has not been wielded.
In Butler's example, the spectrum is consistent with a variety of
Cr profiles; one can discriminate among these profiles to
detennine the real one only by restricting the number of
parameters describing the system (using Occam's razor), for
example, imposing prior chemical assumptions on the system. In
this example, four molecules are used instead of four elements,
but the light element is correlated with the heavy elements so
that, in addition to the number of stoichiometric possibilities
being reduced, the insensitivity of the system to the light
elements is not allowed to seriously distort the results.
Moreover, analysts sometimes overlook assumptions that
they tacitly impose on data when interpreting the data according
to a model. For example, the As implant profile of Fig. IS.!
could be fitted with a Gaussian distribution, and many IBA
codes facilitate such a fitting. Now, there is nothing wrong with
this approach, as long as it is borne in mind that an explicit
assumption is being imposed on the data, which can lead to error
(that might be large) if the profile is not, in fact, Gaussian. A
model-free fit, imposing no prior assumption on the data, would
instead involve a number of layers of varying stoichiometry.
Occam's razor requires that the number of layers be restricted in
order to restrict the number of fining parameters. Then, there is
an explicit tradeoff between the goodness of fit and the number
of layers. Clearly, for any given spectrum, the number of fitting
371
Chapter 15
Possible
Impurities ignored
cons~guences
.... --,-,,--_._----
Use a complementary
technique that can detect the
missing element; collect
spectra under different
experimental conditions.
Collect spectra under different
experimental.conditions; test
Depth profile
QBA
-.
._--- ----_.,- -
different models.
Beam energy
Calibration of accelerator
inaccurate
Scattering angle
Angle of incidence
Energy resolution
Solid angle
372
parameter.
Beamcbarge
Stopping power
Database inaccurate
Straggling
I~o:~d_~ben relev~t
Plural scattering
--
--- - -
Multiple scattering
--
--- -
--
Unnoticed accidental
channeling overlooked in
analysiS
Pulse pileup
Acronyms: EBS. elastic backscanering spectrometry; MeA, multicharwc1 analy:zer; NRA. nuclear reaetioD analysis; RBS, Rutherford backscattering
spectrometry. Caution should be used where the suggestion is to treat a variable as a fit parameter, as this could lead to evenlarger erron!.
373
Chapler 15
simply due to lack of knowledge (for instance, that newgeneration codes can extract roughness information with
considerable ease). However, most often, undcrintcrpretation
occurs when the data are not well fitted Modem codes are now
available that facilitate excellent fitting of spectra, enabling the
analyst to put much greater weight on the results.
A major pitfall is the use of the wrong value for the product
of the charge and solid angle, QQ, where normalization of this
quantity is not easily available from the spectrum. QO is
implicitly determined by the composition of the sample (where
there is no cbanneling). so that, if the wrong value is imposed on
the data, the results can be severely distorted. An illuminating
example in which the uniqueness of Q!l was used to determine
the sample composition with great elegance was reported by Lee
et al. (2006)
Finally, an outstanding pitfall to bc avoided is the
overinteIpretatian of poorly fitted data, which must always be
treated with great caution. Given that particle scattering spectra
are not easily inverted into depth profiles, because the signal at a
particular channel can come from various elements at different
depths, if the spectrum is poorly fitted, the model could be
entirely wrong! In particular, the fact that part of the spectrum
can be fitted petfectly does not mean that the model is partly
374
,j
j
Ji"
ewe,
375
>
Chapter 15
O.l(dEIdx)n ~ __ , .
U, cos<j>
(15.4)
376
Energy (MeV)
0.3
0.4
0.5
0.6
0.7
800~--~~-----,------~------.-----~-.
~600
a;
'"
!'"
~c:
:::I
...o
400
Grounded
'C
a;
:>
200
o
100
150
200
250
300
350
Channel
FIG. 15.7. Surface charging effet:::t. Comparison of RES spectra from a quartz target using 1 MeV 4He: ungrounded and grounded
through a thin conductive surface layer of graphite by rubbing a pencil lightly across the swface (Almeida and Macauley-Newcombe.
private comm~nication 1991) ..
insulating specimens.
charging include
Furthermore,
surface charging sometimes produces
sufficiently high electric fields to interfere with the performance
of the Faraday cup system. 'Also, in certain insulators (for
example, BaTiOJ ), excessive surface charging can even cause
the target to disintegrate.
377
L>.o...... __
Chapter 15
II
Ii
.1
{I
"1
.J
.~
.'.
f'iil' .
,"Ii-
. "'fr'
.~:'
:,,;.1 .
.~,.;~
378
6.
'.~
.
. 6 .....
51
a. 4
..
.ooi.
. As dep()sited
'in'
<::
_____ .
D
~~8.
70"
b
4
'"0
:.'\]HSl
2t
Pb
..
~
....
D
o----~
~
(a)
..
."
CD 0
(b)
>.
."
<:
.;:
.,
co
.,u
...
/v
..
ID
Pb
Annealed at 280'C
51
(20 min.)
2
8Q'
1.0
IPb
1.5
Energy (MeV)
FIG. 15.8. Ambiguity in RBS between roughness and diffusion, Backscattered energy spectra for several tilting angles: (a) from a
uniform Pb layer (50 run) on a Si substrate and (b) from Ibe same layer after being annealed for 20 min. at 280'C (Campi,ano ef 01.,
1978).
The pitfall that must be avoided is forgetting Ib,t presenting
379
~.
.. _.. -
..
_-------Chapter 15
ACKNOWLEDGEMENT
REFERENCES
"
'1,'
Amsel, G., Girard, E., Vizkelethy, G., Battistig, G., Girard, Y.,
and Szilagyi, E. (1992), Nuel. Instrum. Methods B64,
811.
l
~
{,
Bianconi, M., Abel, A., Banks, le., Climent Font, A., Cuben,
C., Doyle, B.L., Lotti, R, Lulli, G., Nipoti, R., Vickridge,
I., Walsh, D., and Wendler, E. (2000),Nuel. Instrum
Methods BI61-163, 293.
..
Blaauw, M, Campbell, J.L., Fazini6, S., Jaldie, M., Orlie, I., and
Van Espen, P. (2002), Nuel. Instrum. Methods B189, 113.
'.
BAM (2009), BAM-S107, BAM-SI08, BAM-SI09, and BAMSIlO. in Cawlogue o/Certified Reference Materials,
Bundesanstalt rur Materialforschung und -priifung
(BAM), Berlin, Germany.
Boudreault, G., Claudio, G., Jeynes, C., Low, R., and Sealy, BJ.
(2004.), Nuel. Instrum. Methods B217 (I), 177.
380
!,
:!"
:,~
:~~
BlPMI1ECIIFCCIlSOIIUPACIIUPAP/OIML, identical to
Edge, R.D. (1983), IEEE Trans. Nucl. Sci. NS-30 (2), 1685.
El Bouanani, M., Pelican, P., Razpet, A., CadeZ, I., Budnar, M.,
Simcic, J., and Markelj, S. (2006), Nuel. lnstnlm.
Methods B243, 392.
Ecker, K.H., Watjeo, U., Berger, A., Persson, L., Pritzkaw, W.,
Radtke, M., and Riesemeier, H. (2002), Nucl.lnstrum.
Methods B188, 120.
Huenges, E., Rosier, H., and Vonaeh, H. (1973), Phys. Lett. 846
(3),361.
lBANDL (2009), database at www-nds.iaea.orgiibandl
maintained by AF.Gurbich.'::::- ~- -....-.---.-~-.- .:.-:.
~.
--
..
Jeynes,
Joynes, C., Jafri, Z.H., Webb, R.P., Ashwin, M.J., and Kimber,
A.C. (1997), Suif. Interface Anal. 25, 254.
Goulding, F.S., and Landis, D.A. (1982), IEEE Trans. Nuc/. Sci.
NS-29, 1125.
Measurement,
381
- ---_._-----==
Chapter 15
Kon8C, G., Kalbitzer, S., Klatt, C., Niemann, D., and Stoll, R.
Pritzkow, W., Vogl, 1., Berger, A., Ecker, K., Grotzschel, R.,
Klingbeil, P., Persson, L., Riebe, G., and W!l.tjen, U.
(2001), Fresenius J. Anal. Chem. 371,867.
Roush, M.L., West, M.A., and Marion, J.B. (1970), Nuel. Phys.
A147,235.
Mayer, M., Fischer, R., Lindig, S., von Toussaint, V., Stark,
R W., and Dose, V. (2005), Nuel. Instrum. Methods
B228,349.
Mendenhall, M.H., and Weller, RA. (1991), Nucl.lnstrum.
Methods B58, II.
Seah, M.P., David, D., Davies, lA., Jeynes, C., Ortega, C.,
Read, PM., Sofield, C.J., and Weber, G. (1988), Nucl.
Instrum. Methods B30, 140.
Sitter, C., Davies, 1.A., Jackman, T.E., and Norton P.R. (1982),
Rev. Sci. Instrum. 53,797.
Paszti, F., Manuaba, A., Hajdu, C., MelD, A.A., and Da Silva,
M.F. (1990), Nuel. instrum. Methods B47, 187.
382
van Lieshout, R., Wapstra, A.H., Ricci, R.A.. and Girgis, R.K.
(1966), "Scintillation spectra analysis", in Alpha-, betaand gamma-ray spectroscopy (Siegbabn, K., cd.), NorthHolland, Amsterdam, p. 515.
Wang, Y.Q., Liso, C.G., Yang, S.S., and Zheng, Z.H. (1990),
Nucl. Instrum. Methods B47, 427.
-- ..
383
. ..
~.~
-_..
....-.-_.--,", --
----.-----.---...-..-.....,--...........- ...
_,--"
~~-
-'-
CHAPTER
INSTRUMENTATION AND
LABORATORY PRACTICE
RA. Weller
Vanderbilt University, Nashville, Tennessee, USA
Contributor: Y. Q. Wang
CONTENTS
16.1
INTRODUCTION ....................................................................................................387
385
Chapter 16
386
j
1
16.1 INTRODUCTION
Although many aspects of mstrwnentation are important for
specific tasks that arise in connection with ion beam analyses,
five general areas are of special significance. These are ion
beam production, transport, and diagnostics; particle detectors;
pulse electronics; vacuum technology and thin-film fabrication;
and statistics and data handling. Anyone who is active in the
field of ion beam analysis can ex.pect to have to deal with some
or all of these areas routinely. This chapter consists of brief
overviews of each of these areas of technology, emphasizing
the relevant topics for ion beam analytical work and suggesting
more detailed references for further study.
Most of the technology that is in current use by materials
scientists for ion beam analysis was originally developed for
the study of nuclei and can be accessed bibliographically
through such key words as "nuclear electronics" and "radiation
detection" and in journals such as Nuclear Instruments and
387
Chapter 16
broken into
terminal.
16.2.2 Accelerators
The next stage in beam production is the accelerator.
Exhaustive discussions of the types of accelerators now used
widely for materials analysis can be found in the book by
England (1974). Voltages below approximately 500 kV can be
produced and maintained with reasonable success using only
air inswation. Above this value, it becomes necessary to
surround the terminal of the main high-voltage power supply
with a pressurized gas. SLIlfur hexafluoride. SF6 is perhaps the
best insulating gas; althougb various mixtures of CO, and N,
are also lL~ecL The vessels that contain the insulating gas give
accelerators their characteristic tanklike look. Voltages up to
about 400 kV can he generated with relatively conventional
power supplies that use voltage multiplication circuitry similar
to tbe original design of Cockcroft and Walton (England,
1974). Above about 500 kV, other techniques are often used to
generate high voltage. One of the most widely used devices for
the production of megaelectronvolt beams is the Vande Graaff
generator, named for its inventor, Robert J. Van de Graaff. An
engaging history of this ingenious device has been compiled by
Bromley (1974). We limit our further discussion of accelerators
to electrostatic accelerators of this type. Discussions of the
principles by which other accelerators operate can be found in
England's book or in the more recent monograph by Hwnphries
(1986).
., i
~'
All accelerators that are routinely used for materials analysis . ~;"
.f.
-.
;:~
at tt
".
388
'
charge state for the fmal ion. Tandem accelerators can produce
ion beams that are far more energetic than those available from
single-ended accelerators with conventional ion sources,
because n can be much larger. As a result. tandem setups are
more appropriate for techniques sucb as nuclear reaction
Single Ended
2MV
Ground
potential
(a)
Source
Accelerator tube
Wafer
Terminal
Ground
potential
Ie: ~r-i
++
r-T""O,.-r--,
Source
,.
--~-.....;.
--
--- .
Stripper
----.;.-_.
- - ....
- -/
(b)
Wafer
Terminal
FIG. 16.1. Schematic diagram of (a) single-ended and (b) tandem Van de Graaff electrostatic accelerators. Terminal voltages up to
approximately 20 MY have been achieved with tandem accelerators. Voltages of 1-3 MY are typical for machines designed for ion
beam analytical applications. Figure courtesy of National Electrostatics Corporation, :Middleton, WI.
not always the case. Magnets actually filter the ions by
momentum per unit charge or, equivalently, by the magnetic
rigidity
16.2.3 Beams
The ion beam that emerges from the accelerator typically
contains a range of energies and even a multiplicity of species.
These variations result from the imperfect vacuum of the beam
tube, poor quality of the beam injected into the accelerator. and
processes such as scattering and charge exchange that occur
during acce1eratioD. To extract a beam suitable for materials
analysis, it is necessary to reject all ions that do not have the
desired mass and energy. This is almost universally accomplished
by passing the beam into the field of a magnet, called an
analyzing magnet, whose field has been tuned to pennit only the
desired species to pass through its narrow entrance and exit slits.
Under most circumstances, this means that it is possible to select
a unique beam with the de:sired mass and energy. However, this is
"'
( q' J
Bp= 2mE
(16.1)
389
Chapter 16
one-half the charge, one-half the energy, and one-half the mass of
2
the He + ion, cannot be distinguished by a magnet. It is sometimes
bard to detect this condition by looking at backscattering spectra
alone, as backscattered protons can be interpreted as ordinary
structwe in an otherwise unremarkable He spectrum.
beams usually are the ones expected on the basis of, for example,
the composition of the source gas. By making some initial
assumptions and iteratively assigning masses to other beams, it is
possible to arrive at a consistent relationship between B and m at
a given energy. All of this is necessary because ion sources
ahnost universally produce many more beams than the uninitiated
operator would reasonably expect These arise from impurities in
the feed material and sometimes from the materials comprising
the source itself. In addition, molecules can be manufactured in
the source. For example, chlorine feed gas is especially effective
at generating volatile chlorides from solid constituents within thc
source. The process is so efficient, in fact, that it is used to
produce beams of some species that are otherwise very difficult to
vaporize.
390
a series
391
Chapter 16
16.3.3 Scintillators
applications.
Some materials emit brief pulses of light when they are Sln,_' ",1)':;
form the basis for the detection of 1-rays and X-rays and,
frequently, electrons and particles. The qualities that distin:gui,;h a'
good scintillator are a large absorption cross section
incident radiation, efficient generation of scintillation
and a long mean free path for these photons to be reabsorbed.
most common scintillators used in materials work are IIlO'l:ani(
crystals such as sodium iodide activated with thallium,
B~Ge,012'
which is
detection and extremely fast response times. They are widely used
392
Table 16.1. Properties of inorganic scintillators. Sources: Radiation Detectors Catalog, Solon Technologies, Inc.; Sakai (\988), Holl ef
al. (1988), Knoll (1989).
.
Scintillator
Den~
Ig/e )
Principal
decay
time
constant
Ips)
Pulse
rise
Hm.
1()'"
9.%
Ips)
Ught
output
(photons!
Absolute
scintillatlon
efficiency
(nm)
Index of
refractlon at
'mn
MeV)
I")
Hygroscopic?
0.23
0.5
38000
11.3
Yes
Wavelength or
Ughl
output
relative 10
Nal(TI)
emission
I")
m~.
l",~
Na(TI)
3.67
100
415
1.85
CsI(Tl)
4.51
45
550
1.80
1.0
52000
11.9
Slightly
Csl(Na)
4.51
.5
420
1.84
0.63
39000
11.4
Yes
'.8
8200
2.1
No
BI.Ge 30'2
7.13
CdWO"
7.9
CaF~Eu)
BaF 2
Fast
3.18
4.88
Slow
12-14
...0
2.15
0.'
40
540
2.2-2.3
5.0
50
435
1.44
4
20
225
310
325 nm
1.49@
0.94
0.0006
0.62
CsF
4.54
390
1.48
589 nm
0.005
UI(Eu)
4.08
35
470
1.96
1.4 n
radiation
0.6y
radiation
15000
3.8
No
24000
6.7
No
10000
4.S
No
Yes
11000
2.'
Yes
393
Chapter 16
Table 16.2. Physical constants of p1astic and liquid scintillator'S. Also included are glass scintillators, and for comparison, anthracene
and Nal(TI). Source: Table of Physical Constants of Scintillators, NE Technology, Inc., 9 Deer Park Drive, Monmouth, NJ 08852.
Used by permission.
MoWng
soflenlng
Aefrac or boiling
live
Scintillator
Plastic
NE lOlA
Plastic
P&aslic
Plastic
D,nslty
index
1.032
1.581
1.58
1.58
1.58
1.58
1.582
1.58lI
1.594
1.581
1.58
1,581
PiIoIU
PHot425
P~sUc
1.032
1.032
1.032
1.032
1.032
1.032
1.020
1.14
1.032
1.032
Plastic
1.19
1.49
NE:213
Uqu~
0.874
NE224
NE226
o.an
1508
1505
1.38
150
1.43
NE236
Uquid
Uquid
Deutemted liqllid
Oeulerated liqllid
Uquid
Uquk!
NE237
Uquk!
0.813
NE104
NE 108
NE 110
NE111A
NE114
NEIlS
NE 118
NE 125
PiIotF
Liquid
Type
NE230
NE 232
NE135
Plastic
Plastic
Plastic
Plastic
Plastic
Plastic
Plastic
1.61
0.945
0.89
0.858
0.796
1.47
1.444
1.459
point
re)
WaveHIt
length 01 Contenl (no. H
output main com- maximum of loading atoms!
emission element no.ofC
(%an ponen!
fhratene)
(ns)
(nm)
(% by wt) atoms}
Decay
Ugllt
75'
75'
75'
75'
75'
75'
75'
65
66
65
60
55
50
35
60.
55
65
99'
SO'
_ 75'.
75'
100'
or
78
80
20
141"
169'
SO'
81'
81'
350'
192'
192'
80
80
40
48
61
constant,
2.4
1."
15
3.3
to
4.0
320
3.3
2.7
21
1.36
3.7
26
3.3
3.0
4
4
3.5
3.15
423
400
545
434
370
434
395
434
425
425
391
425
.25
.25
430
425
43<l
013.8%
1.104
1.100
1.103
1.104
1.103
1.109
1.108
T." ~ 1ast n
u1!ra-fast counting
with pholodiodes
1. a. p, last n, etc.
utJatast timing
asforNE110
pIiOswidIes
1.104
high temperatures
n
T, a, p, last n
ultra fast timing . ~.
Cheren~ov detector
0.957 O/C
1.103
1.100
1.6
014.:2%
024.5%
420
425
425
Principal
applications
1.213
fasl n (1'.S.o.)
1330
1. fast n
1. Insensitive 10 n
0
0.984 O/C Spocial applicalions
1.960/c Special applicalioM
large tanks
2.0
1.92
low temperatures
high flash porn.
1.92
general
Loade<t
"lui<!
Neutron
NE311 &311A
NE314A
NE316
NE320
NE343
6 badad IIqu~
Pb loaded fiqllid
Sn loaded 6qui<!
'u loaded fiquid
GIl loaded fiquld
NE '22& 426
'L>ZnS(Ag)
ZnS(Ag) plastic
. (ZnS-type) NE 451
nnd
glass
C')"IaI
N.I(!1) .
Crystal
Crystal
0.91
1.411
65'
0.96
1.53
1.496
1.497
1.502
141"
O~
0.906
0.884
2.36
1.443
2.64
2.5
2.42
65
2'
35
148.5'
160'
168"
34
65
3.8
2.0
4.0
2.19
3.0
300
200
110
200
300
c.1200 2B 20 8 60
c.l200" 25 20858
110'
~'2
1.58
1.55
1.566
1.55
1.25
3.67
162
1775
c.1200
20
c.l200
25
217'
100
230
6SO'
18862
18855
30
230
425
425
425
425
425
450
450
395
395
399
397
447
413
65%
Pb7.5%
Snlo%
'uO.I5%
GIl 0.5%
1.701
1.261
1.411
1.428
1.360
n,p
T.X-reys
l.x.;ays
n
sIJw n
US'll.
last n
n. p
U2.3%
U6.6%
U7.S'/.
U7.7%
l,a,
.-
last n
:,;
i
1, x-rays
i,
-1
'i
.~,
,~
"
;~
\
',:i{
>g
;~
,j
394
11
.....
hotocathode
focus
120
..,"
0
~
..,.."
Dynodes
100
Photoelectric
Pair production
dominant
effect
dominant
80
60
40
Compton effect
dominant
20
0
0.01
0.5 1
0.05 0.1
5 10
50100
r Energy (MeV)
FIG. '16.3. Relative importance of the photoelectric effect,
Compton effect, and pair production as a function of the )'-ray
energy and atomic number of the absorber. (From Evans, 1955.
Copyright McGraw-Hill; used with permission.)
0,,-
SI~j.HV_1-i'VIM Anode
~
Anode Signal
PO
~hleldin9
e)lcltea states
follOwIng IoniZatIon
PhotoefeC1ron
Dynode ____--~C1"
(secondary
emlttedfrom
cathode
elec:tron
emission)
Anode
395
Chapter 16
"
28""5
23 - 92.13
29 - 4.67
3O-S.n
n-27.4
73-7.8
74-36..5
76-7.8
"-f9'm'I
.",
5.323
.96
4.41
. _ Lilearthermal expansion _
coeffide1ll (/('"1)
"""".--
"""""''''''''''
...... ""(K)
1683
mO.6
""""""""""
12
18
1.115
1.165
0.665
0.746
2.a<105
47
1350
"'"
<SO
1.1xlo'
1900
8.2xlo'i
1()1
5.9xlot'
9.SxloG
3..62
3.76
HoI appecable
2.96
0.5430
0.5658
""lK
OK
..,
nK
2.1xlO"
3.6xW
nK
nx
nK
~.2xl04
396
Specific
...... J_na ..
(ohm-c:m)
n"""
so
.....,.,-,,
Depfc1J"n
cap_llan_
{pFlmm~
~"'~
s.
..
,0-- -.
,0'
~.
". " u
,. ,..
'"
.,.
so "
. ..
..
~.o'
,. ,"
_2.~
._""
...
,,.
:- 10
"
-. .
.,
I-
.1!;!
--- 10-''''
....
m
,. ..,.
,,~
1-""
I:
--
=-'0'
.,
f- ",:
,.,. ,
---
(M"Y)
0.1 ~
B~
,0'
......"
P .uclo-encrgy
lor 1'8t\9f' ot
depletion dopth
.-~
.. ---
"
b"
'~
"
..
f- :C " .
f- - ,
..
"-'
.00 ,
...
.~
"-'
.05
.M
"
-'"- - ~-----
... -
- -
--
...
.~
1.0
~.
gJ_ comblnatJons of n=lat!vfty and deto<:tor bills that may be used to achieve that
dqth. (Shown.
t,,' .. ~ "" thO votrage ttI.!lt ........ be appl_ l<> .. 13,COO "hm-c'"
4500 oto .......... ......typo _mean .... ~ to atop" 23-6k0Y ..Ipha, II &-MaV
proton, Of'1I :e$O-k"y electrOn wfthln I:hCI depletion depth..)
p-typo _
FIG. ~6.5. S~co~ surface-barrier nomogram, showing the relationship between bias, depletion depth, resistivity, and specific
capacItaDc.e. Sunllar to the nomogram reported by Blankenship (\960). From the Ortee Catalog, copyright EG&G Ortec. Used
Wlth penmsslon
397
Chapter 16
100r--------------=~~,---~~------,
~
0.4
(a) C
Polymer
~ 80
c:
3mmThick
'u
'"
IE 60
5mmThick
UJ
c:
--'"
'"
.Q
<>
40
8pm
Cl 20
>-
l'---.13
01
~
'c:"
UJ
'------ 25
~m
~m
I I I I III
:;
u.
"1
0.1
10
100
Energy (keV)
10,000
(b)
Jt
1,000
:a
J:
:;:.
lL
> ~ ~;~
Ie'
Q)
100
:""~
1=80
,,0 '.b.J
~ ~'"
J;'
1/
[2.351ffi"(
F=O.1
L...-
V :Oe
10
0.1
..
1'0
keY
<0
~""Iedron pair at
HiD
1,000
FIG. 16.6. (a) Transmission efficiency curves of a typical Si(Li) detector for various types and thicknesses of windows. The
polymer window curve does not show the effect of the support grid on the overall efficiency. (b) SiCLi) energy resolution,
including effects of intrinsic and electronic noise, as a function of X~my energy. From the Canberra Catalog, copyright Canberra.
Used with permission.
398
Pulse-height t - - - - - - I Time-to-amplitude
converter
analyzer
+1600 V
CFD
CFD
Preamps
.." ,,
~ro.:~
.:(0 .-
c;,~ ~""
~,e.
..
,,'.., \ ' ...",'"
""""
\
->
Q:::J=I--;S~top
pulse
-800 V bias
FIG. 16.7. A time-of-flight detector system. Note: CFD stands for constant-fraction discriminator, and MFP stands for microchannel
plate.
comparable to the 20 keY being obtained using a silicon surfacebarrier detector. Assume that the tota1length between the carbon
399
L
..
Chapter 16
foil and stop detector is 1 m. The flight tirrie required for the a
particle to cover this distance at an energy of 2 Me V is about
100 ns. An energy resolution of 20 keY is equivalent to 1%, so
that a timing resolution of 0.5% is required (E - r). This means
that a timing resolution of 500 ps is needed Consider first the
implications afthis resolution for the start and stop pulses. These
pulses are produced by very fast electron multipliers, and
although there is no directly useful information in their
amplitudes, these amplitudes do vary. The pulses might also
need to be amplified by a voltage~sensitive preamplifier. As a
result, they can have durations of up to several nanoseconds and
16.4 ELECTRONICS
16.4.1 Introduction
The vast majority of all modem scientific measurements
follow the same general form. The physical quantity of interest
is related to an electrical quantity through a transducer, the
electrical quantity is processed using the standard techniques of
analog electronics; the analog signal is digitized with an
appropriate analog-to-digital converter; and the resulting values
are stored digitally by a computer, possibly after additional
digital processing. The procedures used for ion beam materials
analysis follow this scenario quife closely.
The transducers used in ion beam work are the various
detectors that were discussed in the previous section. The
pwpose of this section is to introduce the remainder of the chain.
All of the significant signal-processing and analysis tools
necessary for ion beam studies of materials are available from
commercial vendors. However, this does not imply that it is
possible to treat these t~ls as "black boxes". Quite the contrary
is true. 'If one is to obtain the best results in routine
measuremeots and have the flexibility to push into new areas, at
least the functioning of standard electronic modules must be
understood
-
400
.. .,..'
-"'~-----"-=':"
401
Chapter 16
Table 16.4. Swnmary of some commonly available NIM' modules, their typical inputs, outputs, and functions.
KIM module
Preamplifier
Input
Output
Camments
amplified pulse
Spectroscopy
amplifier
Biased amplifier
Timing amplifier
Delay-line amplifier
Summing amplifier
more channels
Delay or
delay amplifier
Single channel
ana!y2er
Generates an output pulse only if the input pulse amplituda lies between adjustable upper and lower thresholds.
Constant fraction
discriminator
Unear gate
Coincidence J
anticoincidence unit
Scaler. counter
(timer)
Aatemeter
Pulser
Current integrator
Visual
or computer output
of accumulated number
of pulses
100 mV
402
,
,
Bin
tonn~or
0 1001
034
Top
28
,,0 HOd
dc5 O
o,,0"Od d
dOd
18
DgO
T""
16
-6volts
Reserved
Re .... rved
.12vo!t5
-12 .. oIls
OS
"""
"
"
"
,.
..
4bcS~o
32
OS
"",.
Ground
guide
'""'"
GGS
..
"
42
G
28 34
'oga
'oga
19
..
.Pb,O'b'b
()"
0" t)()0
bQOb O
"
+6 volts
22
Reserved
0 0'
0"0 "0
Rear vI_
0
b10
Spare
.".~
Uc?,'
O
0
t;
,
"
"
"
"
"
"
"
" 5_
" R_""'"
'"
" "'.......
"" R-"'"
s""~
D~O
0 '(5 .
0 0
D.C.
22
"
26
Reservecl
S".~
19
41
Module conl'lllC'ttlr
Sp...
.2110vo~
16
36
F"netlon
Ru._
R........d
Fleserved
S,..~
''''~
'O?O
.:!4volts
-24 volts
S... ~
'P"~
"090 6
O"Ob
,
0
d
0.,,0
'P~
117 volts A..C. (hot)
ll_fMUmGND
R....
....
Sporn
..
J.
-,.~
1-0
0'0
"""""....
.cOottom
,~
guide
117 vaH"
High quality GND
Ground guide pin
GGP
PO".
Nates;
1. Rl!served pins 1m! for future assignment by the CCIfJ1mlttee mel shall notbe usl!d until sucf1
~,~P'~:~:-tf,
~~1B~1~
GGS
GOP
3. The POWER RETURN GROUND (bIn pin 34) Is the return b~ fOr III cle supplies.
The HIGH QUAUTY GROUND (blf\ pin 42) 131n~ud iI3 uro potential rde~". High quality
Bin conn.
Mod. conn.
gr.... nd aum:nt load by any mod,,1e ahould b<lllmlted t61 mA. AIsft, pulses 01' varying loads
ahould tie Omlled to riO mORl than 100 mieroamperes. cant mould be takefI not to
Sh.-lng guido pins Emd eeckets
capadtlvely coupll!l high quality ground to the loc:aJ grouncL Any capacitive coupling to local
grourw:l should be IsoIatecI by a ,,"1slanc1t of at least 1000 ohms. The 01ASSIS GROUND I..
fIOnna\tJ conne.::tecllo the bulldhlg grolln4. CClnnll<illoroa batw'cen the HIGH QUAUTV
GROUND bus [bIn pin G). !he powet' RETURN GROUND BUS (bIr1 pin 34), and the ehassls are
rnacle at the ground guide pin of PG1B.
""m-.
4. Slgnal3an pins 35 ancl36 st\aD UtlIlm Iha -lOgIc levelS 10f transrnlS3kln Cli digital data down to
OC- ,,' tha nlm .. peclfl...atlQllS. A 1091<;; "'" gate .. ON Ilnd RESETS.
FIG. 16.8. The pin =ignments for the NIM _dard connector (from USDOE, 1990. p. 25).
403
Chapter 16
Signal
Description
GND
Power-supply ground
GND
High-quality ground
+12
+12 V dc supply
-24
-24 V de supply
--,-_~c_.L. __ "c:~~_.,::?4_
+24 V de supply
12
+ 12V
12 Vde supply
TTL
Logic
level
Output
(must
deliverl
NIM
--I
Input
(must accept)
(1/2) ICL7667
or
(1/2) 050026
+5V
rnA
Because of the age of the NIM standard, its logic and voltage
___. levels .,do no! correspond with any standard logic family used
today or with the :15 V power supplies that are now commonly
used 10 power operational amplifiers (op amps) for analog
circuit functions. For the occasional analog construction project,
it is usually easiest to simply operate op-amp circuitry at the
TTL
NIM
TTL Gate
-12V
404
15
Detector bias
supply
+ High voltage
Spectroscopy
amplifier
MCAorADC
and other
preamplifier!
.........:~_ _ _...J' 1..-_ _---1_
signal
rocessing
.---lL
Particle
or
Photon
or
16.4.4 Preamplifiers
Preamplifiers. come in two types, charge-sensitive and
VOltage-sensitive. Energy spectroscopy systems, such as the one
in Fig. 16.10, use charge-sensitive preamplifiers. As the diagram
implies, these devices integrate the pulse from the detector. The
reason for this is thai"' the - mforniation about the radiation's
energy is encoded in the number of electron-hole pairs that are
created by the detector. By integrating the pulse from the
detector, the energy can be measured directly. The resistorcapacitor (Re) time constant in the preamplifier's feedback
circuit is chosen to be on the order of 100 ).IS, which is vastly
longer than the pulse generated by the detector. For this reason,
the preamplifier's output resembles the graph of e-l for t:> 0 and
is called a tail puLse. The information is contained in the height
of the step at t = O.
405
~-
Chapter 16
Voltage-sensitive
preamplifiers
are
used
for
timing
J--
Input
from
preamp
I~~Rf-2.~. .-R' -l
A,
Differentiator
and pole-zero
cancellation
FIG. 16.11. A simplified schematic of a spectroscopy amplifier. Adapted from the Ortec Catalog. copyright EG&G Ortec. Used with
permission.
Most modern spectroscopy amplifiers have additional circuitry
that aids in improving performance at high count rates. Three
such features are pole-zero cancellation, baseline restoration, and
pileup rejection. A pole~zero circuit is also shown in Fig. 16.11.
It derives its name from the cancellation of a pole and a zero in
the complex plane in the Laplace transfonn analysis of the
406
103
-.,
I::::. ...
I
.~
II
I:
I
I
I
I
I
I
I
a;
Ir-
II:
101
__
Pulser
.,"
102
I:
:l
0
-.,
..'"
iI
I
I
I
I
I
I
I
\."
100~~~\t---L----L____~__~____~__~
Channel nuinber
FIG. 16.12. Spectra of 55 Fe showing distortion caused by pulse pileup. From Wielopolski and Gardner (1976). Used with permission.
407
Chapter 16
8
6
J'\/ I'\.
Unipolar
/A \
/ 1/ \
J/ \
"- "-
.\
-2
-4
.'-V
3
Bipolar
4
Time (liS)
FIG. 16.13. Characteristic shapes of unipolar and bipolarNThninear pulses from a spectroscopy amplifier. The
bipolar pulse is obtained as the derivative of the unipolar pulse.
data channels, computer-based systems such as CAMAC (Clout,
1982), which was originally developed for high-energy pbysics,
are used. Here. we describe a standalone MeA with a Wilkinson
analog-to-digital converter for the sake of simplicity. Data
acquisition and handling are the most individualized aspects of
nuclear spectroscopy. The best advice that can be given is to
understand your own system, whatever its details. and not to
leave those details to someone else.
408
To
computer
Display
'-"'e"'m"o"","-' Inte~ce
Input
"~l
buffer
.A @
Analog to
I1lglhll
converter
Shaped
pulse
.rL
Busy
out
ranges from one part in 256 to one part in 8192. Figure 16.15
illustrates the principle of the Wilkinson ADe. The input pulse
is wed to charge a capacitor until the:: peak of the:: pulse is
detected (by a zero crossing of its derivative). At this time or
with a small delay, a constant current source is turned on and
begins to linearly discharge the capacitor. At the time that the
E=an+z.
=
(16.2)
409
--:1'
~
Chapter 16
iLl
~
TIme
(a) Amplifier :
output_
pulse
I
Lower level
discriminator
UI
(b) Lower
level discriminator I
output
I
(c) Signal on
rundown
capacitor
(d) Address
clock _--,_~!u
:1
(el Memory
cycle -~f--i--------,I
(I) Linear
gate
closed
~I_ _ _ _ _ _ _ _,
(g)Deadt~
gate
L
L
410
I
!
I
j
II
i
j
I
i
i
Half-life
Ex
(keV)
Ey
Isotope
(keV)
(XJy}
54Mo
312.5 d
S.47 (Kapl
834.84
0.251 (11
57Co
270.9 d
122.06
1157 (1)
0.79 (21
0.tt2 (21
6'zn
243.9 d
6.4 (K.)
7.1 (Kpl
14.4 t (yl
B.04 (Kal
1,115.5
B.9(~1
85S r
Bay
64.84 d
106.6 d
-~
0.660 (51
0.091 (2)
t3.38 (K.l
15.0 (~I
514.00
0.502 (3)
0.088 (1)
14.12 (KcJ
15.B5 (Kp)
89B.04
0.549 (7)
0.9B9 (2)
22.10 (K.)
25.0 ("pl._
8l30037
24.14 {Kal
27.4 (Kp)
391.69 ~
1.22 (4)
0_27 (1)
32.1 (t<.)
36.6 (K,I
661.66
0.067 (21
0.Ot59 (5)
33.29 (K.)
165.85
0.81 (9)
0.20 (2)
70.15 (K.l
80.7 (Kpl
411.80
0.0229 (51
0.0064 (2)
72.11 (K.)
B3.0 (Kpl
279.20
0.125 (3)
0.03-48 (B)
38.0(~)
22 (1)
4.7 (2)
411
~-.-----:=-~-
---_..
!
I~
Chapter 16
Table 16.8. Suggested y-ray calibration sources including half-lives and principaly-ray eoergies. Uncertainties in the final digits are
shown in parentheses. Sources: Helmer 0/ 01. (1979), Tuli (1985) (continued).
(k~
Source
Half-lIfe
(k7#)
Source
Half-life
'Be
53.29 d
477.605 (3)
10SmAg
127y
11(mAg
249.76 d
124Sb
60.20 d
602.730 (3)
645.855 (2)
713.781 (5)
722.786 (4)
790.712 (7)
968.201 (4)
1,045.131 (4)
1,325.512 (6)
1,388.164 (7)
1,436.563 (7)
1.690.980 (6)
2.090.942 (8) -
137CS
30.17y
661.660 (3)
13C
+ ex
22Na
a source
2.602 y
6,129.27 (5)
t,274.54 (1)
24Na
15.02 h
1,368.63 (1)
2,754.03 (1)
"50
83.83 d
889.277 (3)
1,120.545 (4)
51Cr
27.704 d
54Mn
312.5 d
320.064 (')
834.643 (6)
"Fe
44,496 d
1,099.251 (4)
1,291.596 (7)
56co
78.76 d
646.764 (6)
1,037.644 (4)
1,175.10(1)
, .238.20 (')
1,360.21 (')
, ,771.35 (2)
1,810.72 (2)
. ',963.71(')
2,015.18 (1)
2,034.76 (1)
2,113.11 (1)
2.212.92 (1)
2.598.46 (1)
3.009.60 (2)
3.201.95 (1)
3.253.42 (1)
3.273.00 (1)
3.451.15(1)
57Co
6Oeo
270.9 d
5.271 Y
144Ce
122.0614 (3)
136.4743 (5)
152.Eu
1.173.238 (4)
1.332.502 (5)
2B4.4d
13.33y
433.936 (4)
614.281 (4)
722.929 (4)
446.8tl (3)
620.360 (3)
657.762 (2)
677.623 (2)
687.015 (3)
706.682 (3)
744.277 (3)
763.944 (3)
616.031 (4)
664.685 (3)
937.493 (4)
' .364.300 (4)
1,475.788 (6)
1,505.040 (5)
1,562.302 (5)
696.5'0 (3)
1.469.160 (5)
2.165.662 (7)
121.7824 (4)
244.699 (1)
344.281 (2)
153Gd
2".6d
69.6734 (2)
97.4316 (3)
103.1807 (3)
898.042 (4)
1.636.06(1)
'1>fm
128.6 d
84.2551 (3)
64.02 d
724.199 (5)
'82'fa
114.5 d
2.03x10 4 y
702.645 (6)
871.119 (4)
67.7500 (2)
84.6808 (3)
100.1065 (3)
113.6723 (4)
SSZn
243.9 d
1.115.546(4)
say
106-54 d
"Zr
"Nb
412
I
i
Table 16.8. Suggested y-ray calibration sources including half-lives and principal y-ray energies. Uncertainties in the final digits are
shown in parenlheses. Sources: Helmer el al. (1979), TuU (1985) (continued).
Source
'''Ta
73.821 d
b....-
Half.llfe
(ConI)
2.696 d
413
411.804 (1)
675.888 (2)
1,087.691 (3)
46.6d
279.197 (1)
32.2y
569.702 (2)
1,063.662 (4)
l,nO.24 (1)
1.9131 Y
136.3434 (5)
205.7955 (5)
295.958 (1)
308.457 (1)
316.508 (1)
(k~
416.472 (1)
468.072 (1)
484.578 (1)
588.585 (2)
604.415 (2)
612.466 (2)
884.542 (2)
1921r
116.4186 (7)
152.4308 (5)
156.3B74 (5)
179.3948 (5)
19B353 (1)
222.11 0 (1)
229.322 (1)
264.076 (1)
1,121.301 (5)
1,189.050(5)
1,221.408 (5)
1.231.016 (5)
1,257.418 (5)
1,273.730 (5)
1,289.156 (5)
1,373.636 (5)
1,387.402 (4)
(ConI)
,
!
SOurce
Half-life
238.632 (2)
563.191 (2)
B60.564 (5)
B93.408 (5)
1,620.74 (1)
2,61453 (1)
-~
Chapter 16
Table 16.9. Alpba particle energy standards, including haf-lives, energies, and relative intensities. Uncertainties in the last digit of tbe
energies are sbown in parentheses. Vel)' thin 212Si and 212po sources can be f,repared by collecting lWRn on a sample surface in the
vicinity of a thorium surface. The controlling halflive is 10.64 b, that of" Ph from which mBi forms by ~ decay. Sources: Rytz
(1979), Ledereer aod Shirley (1978), Tuli (1985).
Isotope
Half-life
'''Sm
1.03x10 By
Eo (MeV)
2.470 (6)
Itt (%)
Isoto~
Half-life
100
2-4Pu
6570y
5.1682 (2)
5.1237 (2)
73.5
26.4
100
2~Am
738() y
5.350 (I)
5.2753 (5)
5.2330 (5)
0.22
87.4
210pO
138.376 d
5.3044 (1)
99+
(calculated)
1.8:><106 '1
150Gd
2.719 (8)
(calculated)
148Gd
75.:l:30y
""Tn
1.41xl0 1O y
... _. ';
.~.-.
:. .. ......_.. ..........
238U
4.468x10 9 y
23SU
7.038x10 By
236U
2.342x107 'I
3.18271 (2)
100
22BTh
1.9131 Y
4.197 (5)
4.150 (5)
n
23
5.4232 (2)
5.3403 (2)
12.7
26.7
241Am
432.2 y
5.544 (1)
5.4856 (1)
5.4429 (1)
0.34
85.2
13.1
""'Pu
87.74 y
5.4991 (2)
5.45S3 (2)
71.5
28.5
24-4Cm
18.10y
5.80482 (5)
5.76270 (5)
23.6
4.599 (2)
4.400 (2)
4.374 (4)
4.368 (2)
. 4.218 (2)
1.2
57
6.1
12.3
6.2
4.494 (3)
4.445 (5)
74
26
4.688 (2)
4.621 (2)
75.5
24.5
234U
2.45xl0 5 y
4.n5(1)
4.723 (1)
12.5
27.5
209pO
102 Y
4.881 (2)
99.52
23 1Pa
3.28xl04 y
5.059 (1)
5.028 (I)
5.014 (1)
4.952 (1)
4.736 (1)
11.0
20.0
25.4
22.B
B.4
239pu
2.412x10 4 y
5.115 (2)
5.1566 (4)
5.144 (1)
5.105 (1)
'43Cm
28.5 y
242Cm .
162.8 d
73.3
15.1
11.5
6.066 (I)
6.058 (1)
5.992 (1)
5.7850 (5)
5.7411 (5)
76.4
1.0
5.0
5.8
73.54
10.65
6.1128 (I)
6.0694 (I)
26.2
73.8
254Es
275.5 d
6.429 (1)
93.2
2S3es
20.47 d
6.63257 (5)
6.5914 (5)
90.0
6.6
212Bi
SO.55 m
6.08994 (4)
6.05083 (4)
27.2
69.9
212pO
0.2981'S
8.7842 (1)
100
414
11.0
n
23
7.S4xl04 'I
I.{%)
4.013 (3)
3.954 (8)
""rh
Eo (MeV)
100
15':;;'
~il=
p -
!>'
!='
."
Cable
Designation
(Badlo
gulde--ll
1JJ.nlversal)
conductor
styl,e
(AWa .
stranding}
RG-BIU
RG-8AJU
13(7)(21)
RG-9/U
13(7)<21)
RG-11/U'
RG-11A1U
1B (7)(26)
.g
Signal
Center
.lnsulaUon
propagation
velocity
Nominal
Nominal
Cable Maximum
Impedance
capacitance
(% of sptI&d
voltage
O.D.
(MHz,
C
(pFlm)
of light)
(mm)
Z.
(a)
(V .,)
Polyethylene 10.29
Polyethylene 10.67
Polyethylene 10.29
5,000
52
51
5,000
75
5,000
96.8
98.4
67.3
66
66
66
100
400
1000
Nominal
~.
signal
attenuation
Governing
(dBI100m)
specification
Comments
0
...,
7.2
JAN-C-17A
15.4
MIL-C-17D
contaminatIng jacket.
29.2
7.2
JAN-C-17A
100
400
1000
29.2
100
400
1000
6.6
13.8
23.3
JAN-C-17A
100
400
1000
14.8
32.8
55.8
JAN-C-17A
100
400
16.1
37.7
70.5
JAN-C-17A
15.4
MIL-C-17D
Gray. non~
contaminating
PVC
1>
~
"n
Jacket. Double
shielded.
contaminating jacket
~.
e-
~
~
AG-S8AJ
...V;
RG-58AJU
20 (solid)
20 (19 x 32)
polyethylene
Polyethylene
4.95
4.95
1.900
53.5
50
1,900
93.5
101.0
66
66
1000
RG-58C/U
4.95
1,900
50
101.0
66
Polyethylene
6.15
2,300
73
69.9
66
100
400
1000
RG-59/U
22 (solid)
100
400
1000
RG-59B/U
23 (solid)
Polyethylene
6.15
75
2,300
67.3
66
;i
RG-62IU
RG-62BJU
22 (solid)
24 (7 x 32)
Seml-sond
. polyethytene
Semisolid
polyethylene
6.04
6.15
700
700
Ii 93
93
44.3
44.3
. 64
64
100
400
1000
100
400
~.
16.1,
37.7
70.5
MIL-C-17F
11.2
23.3
. 39.4
JANC-17A
11.2
23.0
MIL-C'17D
39.4
6.9
1000
17.7
2B.5
100
400
1000
20.0
36.1
9.5
."
JAN-C17A
MIL-C-17D
"'e-.
~
~
~
~...
,
~
"""
00
00
.'"
ro
..S:
...
0
:t!.
P(")
i'l-
po
"'"2"
3
"~
0
""
"....""
0-
~
"gOJ
.
n
5';:7
&ocr "
? ~
~,
0_
00
o '
~
S'
" 0
8.//
~a.
n
~
0
....
Cable
Designation
(Badlo
g,ulde-#
IUnlversal)
Center
conductor
style
(AWe]
strandIng}
Insulation
Signal
propagation
Cable Maximum
Nominal
Nominal
velocity
0,0.
voltage Impedance capacitance (% of speed
(mm)
(Y .,)
C (pFIm)
'Zo(n)
of light)
(MHz)
90
Nominal
signal
attenuation
(dBl100m)
Governing
specification
29.2
MIL-C-170
Comments
RG-174N
26(7:<34)
Polyatl1ylane
2.56
1,500
50
101.0
66
100
400
1000
57.4
98.4
RG-17BB/U 30 (7 x 38)
Teflon
1.83
1,000
50
95.10
69.5
100
400
1000
45.9
MIL-C-170
91.9
Brown. fluorinated
ethylene-propylene
150.9
jacket High
temperatura.
"'O;"
RG-213/U
RG-2231U
13(7x21)
19 (SOlid)
Polyethylene 10.29
Polyethylene
5.38
5,000
1,900
50
' 50
101.0
101.0
66
66
7.2
100
400
1000
29.2
10'0
400
1000
14.B
30.2
53.5
MIL-C17D
15.4
MIL-C-17F
Black, noncontaminating
PVC jacket.
Black; non_
contaminating
PVC Jacket.
DotJble shielded.
RG-31atU
2077t
26 (7 x, 0067)
4 (solid)
Teflon
High molecufarwelght
2.49
38.1
1,200
300,00D
VDC
polyethylene
50
67
95.2
83.0
69.5
60
100
400
;000
34,1
54.1
101,7
MIL-C-170
Brown, fluorinated
elhylene-propvlene
jacket. High
temperature.
0
0
0
~,
eo.
~
~
'"d
~,
0
0
'6'
e>~
"
~
&
D
"
is'
OQ
&;
-'"
'"0:
0
"
~
'i'
n
Po
n
!'l-
-"
:::J~~,~,~(;~:,;~-~~:)~(';.:,{i,~\ ';'" <:i.:};~::::.i .~_;.~~~..;~\:~:;~'j$2~'i~~~~:~~J~~:~;~~~~i:~~4:~~.'.iii~~l~i\;_~-i~::~i;t~yil.Mir~~:ig1~t\;:}d~'tJ:$it.fi~~;\~;~~fil~jt~?t~~;,~~;~;$~*jAAI).~.r,ru
r'
tJ>-l
3: 2:
_J
!!l.~
P~
~::::
(fl)
Nominal
cable size
range
(Inches)
eNC
50
SHV
MHV
g-.g
rfl
Maximum
VoUage
Diolectrlc
withstanding
Coupling
style
fraquency
rallng
(VRMsl
voltage
(VRMS)
Governing
specification
0.150 - 0.250
Bayonet
500
1500
MIL-C-39012
Universal, Inexpensive.
"Baby Meill.concelman:
Not constant
0.054 - 0.440
Bayonet
3500
10,000 VDC
MIL-C-39012
NBS-ND-545
Preferred HV conneclor.
Not constant
0.054 - 0.440
Bayonet
3500
6500 VDC
Mll-C-39012
Potentially dangerous.
Nominal
Impedance
Series
Q
o "
(GHz)
Comments
s<o 01
"'...,
o 0
p~~
e~
p."
Avoid rigorously.
"
na
>0
UHF
Microdot
....
...,
Not constanl
0.199 - 0.432
Threaded
0.2
None
500
93,50,70
< 0.154
Threaded
750
1000
50eM
LEMO (sarles 00)
K-Loc
50
0.150 - 0.250
Locking
2 (typical)
500
1500
TNO
50
0.150 - 0.250
Threaded
11
500
1500
MIL-C-39012
50
0.350 - 0.450
Threaded
11
1000
2500
MIL-C-39012
HN
C
50
50
0.195 - 0.945
0.350 - 0.450
Threaded
11
Bayonet
1500
5000
1000
3000 (HV type)
3000
NBS-ND-549
MIL-C-39012
MIL-C-39012
..
"'
~a
~.
~~
sc
50
0.350 - 0.450
Threaded
11
1000
3000
MIL-C-39012
SMA
50
0.047 - 0.250
Threaded
1.2-18
, (some versions
to 27 GHz)
350 - 500
500 -1500
MIL-C-39012
~ ~
0
S" ~
~
b'
!?So C-
C-0Q
~'"
... :5
~i;g
c::~
en",
>M
en_
o '"
~
0
rtg
~ 2.
ID &l
()
~.o
's<
?
>-l
~
,.,.
0
p-
."
.~
;::
~ ~ ::
I! '!
i ,I
!'C
0
:>
M
"
~.
0>
~
"r
0>
cr
"
~.
~
Chapter 16
lazy to do the job correctly. Now having said this, we know that
you will do something like this anyway. Therefore, obtain a
small bottle and spray can of Vacseal, as well as a tube of a
solventless epoxy such as Torr-Seal from Varian while your
system is still leakfree so that, if you must cast good vacuum
practice to the wind, at least your crude patches will work. You
should also know that, if you must glue something that will be
used in an ultrahigh-vacuum system, the thennal-curing epoxies
H2ID (electrically conducting) and H72 (electrically insulating)
from Epo-tek are useful alternatives to Torr-Seal.
16.5.2 Thin fIlms
Thin filins for use in ion beam work can be produced by
electrochemistry, eIectron-beam evaporation, resistiveheater
evaporation, sputtering, ion implantation, molecular beam
epitaxy, or laserassisted deposition. They are used in
applications rul diverse as beam stripper foils for tandem
accelerators, entrance windows for gas detectors, and
occasionally even exit windows for bringing ion beams into the"
air to perform analyses on samples that are incompatible with
vacuum. The classic reference on thin films is Maissel and
GIang (1970), but more recent comprehensive reviews are also
available (Schuegraf, 1988). Much recent work h;'" been
undertaken in conjunction with the development of
semiconductor technology and is covered in conference
proceedings such as those edited by Poate ef al. (1978) and
Lucovsky (1988). The most common technique for making thin
films is by evaporation. An extensive listing of the physical
properties of compounds, including their vapor pressures, was
compiled by Graper (1990) to assist in this task. This table is
presented in Appendix 18 as Table A18.1.
Several factors must be considered in selecting a foil for a
specific purpose. These include the required thickness;
constraints on the foil's composition such as the likelihood of
nuclear reactions; the area, if the foil is self 5uppOrting; the
differential pressure that the foil must support; and the level of
radiation that the foil will experience. Beam stripper foils
mounted in the terminal of a tandem Van de Graaff accelerator
are universally composed of carbon and are usually about 3-5
~glcm2 thick. At the other extreme are vacuum-system exit
windows, which must support atmospheric pressure on ODe side
and yet permit an ion beam to penetrate. In the early days of
nuclear physics, these windows were almost always made of Ni.
Now, however, AI, Ti. and the radiation-resistant organic
material kapton are also used_ The required thickness depends,
of course, on the area of the window, but it is typically a few
mglcm 2. Ti.is also commonly used for exit windows in highcurrent electron accelerators that manufacture plastic films by
radiation-induced polymerization_ An analysis of the mechanical
behavior of thin films subjected to a differential pressure was
published by Spivack (1972).lfthe required thickness ofa fi1rn
418
Pen)
(16.3)
nl
(16.4)
n_O
Pen)
16.6 STATISTICS AND DATA HANDLING
!2a 2
(16.5)
,jZncr'
JlOO
419
-,
Chapter 16
i'if
i'if
(16.6)
dz =-dx+-dy.
Ox
iJy
it.
(16.8)
cr,,' =(Cx-(X)(y-(y)).
If (Jz2
0, then the errors in x and y are said to be
uncorrected. In charged-particle spectrometry, errors are
usually either known to be uncorrected or are assumed to be.
Some counterexamples are described in Nicholson (1974).
=;;
Table 16.12 lists the rules for propagating eITors for several
specific functional fonns of f. Notice, in particular, the form z
= x. - y, which is used for background subtraction of y counts
from a total ofx. In this case, (IE' = OJ,: 2. +
= x + y. Thus, if a
peak contains 400 total counts of which 100 are background,
0'/
(16.9)
where Sex) is the Dirac delta function and the integral covers ali
of (x,y) space.
A-A' =U',.
A'
Function
z= axy
Uncorrelated errors
cr ,
cr
cr '
-'-=-' +-'z1
Xl
yl
ax
cr
. .
. ,
1
cr
cr
cr
....!... =-'-+-'-+2--!!,
z
cr
cr
xy
=ax,t,b
xy
~ .. b'~
!.'
z= aeb:c.
(16.11)
....!...=_-+---2....!L
,
z=-
Correlated errors
A=-- or A'=--.
I-A',
1+ A,
7.5% error.
Table 16.12. Error propagation formulas for some simple
functions of the random variables < and y. The parameters a and
b are constants.
(16.10)
In(bx)
420
I!
!
iI
i
I
I1
I
I!
L
(16.14)
and
'1, = J'1,(t) f(t) dt.
(16.15)
(16.12)
(16.13)
JtP(t) dt
fp(t) dt
o
421
(16.16)
Chapter 16
REFERENCES
Althaus, R.F., and Nagel, L.W. (1972), IEEE Trans. Nucl. Sci.
NS-19 (1),520.
Betts, R.R. (1979), Nucl. Instrum. Melhods 162, 531.
Bevington, P.R. (1969), Data Reduction and Error Analysisfor
the PhySical Sciences. McGraw-Hill, New York. A standard
introductory text on the subject of data analysis.
Blankenship, J.L. (1960), Trans. Nucl. Sci. NS7 (2-3), 190.
Bromley, D.A. (1974), Nucl. Instrum. Methods 122, I. See also:
Bromley, D.A. (1984), "The Development of Heavy-Ion Noelear
Physics'\ in Treatise on Heavy-Ion Science (Bromley, D.A.,
ed.), Plenum, New York, voL I, p. 3. A higbly readable
historical article by a person with broad knowledge of both the
field and the personalities who made it.
Bromley, D.A. (ed.). (1979), "Detectors in Nuclear Science',
NucJ. lnsfnlm. Methods 162. This entire issue is devoted to
invited reviews of detector technology. The articles contain
much valuable technical and historical infonnation.
Brown, I.G. (1989), The Physics and Technology ofJon Sources,
Wiley, New York.
Bryant, J.H. (1984), IEEE Trans. Micruwave Throry Tech.
MTI-32 (9), 970. This issue also contains a number of
fascinating review articles on the history of microwave
electronics. See also: Maury, M.A. (1990), 'Microwave Coaxial
Connector Technology: A Continuing Revolution", in 1990
State of the An Reference (a special issue of MicrowuYe
Journal), Horizon House, Norwood, MA, p. 39.
Campbell, J.L., and McNelles, L.A. (1975), Nucl. Instrum.
Methods 125, 205.
Canberra. (1978), Canberra Laboralory Manual, Canberra
Industrics, Meriden, CT.
Clout, P. (1982), A CAMAC Primer, Report LA-UR-82-2718,
Los Alamos National Laboralory, Los Alamos, NM ThiS
document can also be obtained from a CAMAC module
manufacturer,
KineticSystems
Corporation,
bttp:llwww.kscorp.comIProductsicamacJ.
422
0,0;.
_'~;iil!
York, p. 712.
Holl, I., Lorenz, E., and Mageras, G. (1988), Trans. Nucl. Sci. 35
(1), 105.
text and reference for physical scientists. Readable, wideranging, and full of practical experience and insight
423
Chapter 16
technology,
Poate, I.M., Tu, K.N., and Mayer, I.W. (eds.). (1978), Thin-Film
InterdifJusion and Reactions, Wiley. New York.
Press, W.H., Teukolsky, S.A., Vetteriing, W.T., and Flannery,
B.P. (1992), Numerical Recipes in C, 2nd Ed., Cambridge
University Press, Cambridge, UK. See especially Chapter 14, p.
609, and Chapter IS, p. 656. An instant classic in the field of
424
(1974),
CHAPTER
RADIOLOGICAL SAFETY
P. M. DeLuca, Jr.
University of Wisconsin, Madison, Wisconsin, USA
J. R. Tesmer
Los Alamos National Laboratory, Los Alamos, New Mexico, USA
P. Rossi
University of Padua, Padua, Italy, and Sandia National Laboratories, Albuquerque, New Mexico, USA
CONTENTS
17.1
17.2
17.3
17.4
17.5
17.6
17.7
17.8
425
---
Chapter 17
17.1 INTRODUCfION
Although a complete discussion of the many aspects of the
radiation environment encountered near low-energy particle
accelerators is beyond the scope of this chapter, general
information is provided here to guide the user. More detailed
information is available in the references listed at the end of this
chapter.
426
Radiological Safety
D=~
dm'
(17.1)
dm'
(17.2)
.?'
",'b
[I[ICI:IJ
[I
100
10~ 103
10'
(17.3)
427
Chapter 17
Table 17.1. Mean quality factor, Q, and fluence per unit dose
equivalent for monoenergetic neutroDS.
Neutron
energy
Quality
factor*
(Me\')
(Q)
2
2
2
2
2
2
2.5
7.5
11
11
9
2.5 x 10->
1 X 10-7
1 x 10-<
1 x 10-'
1 X 10-4
1 X 10-3
1 X 10-2
1 X 10-1
5 x 10-1
1
2.5
7
10
14
20
40
60
4 xla'
4 X 102
4 x 102
4
10'
7
6.5
7.5
8
7
5.5
4
3.5
3.5
3.5
Stochastic
(whole body)
Deterministic
on leos of eye
Deterministic
on sIQn, bands, and
effects
effects
Ann. Eq-D = 150 mSv
effects
feet
Ann.: annual.
Whole body: defined as extending from the top of the head
down to just below the elbow and just below the knee
428
Radiological Safety
1993)
Occupational Dose Limits
Stochastic
(whole body)
effects
Detenninistic effects
on lens of eye
Deterministic effects
on skin and extremities
Deterministic
effects
/) =.S",Qx1.602xIO-'" Gyfs
(17.4)
skin.
4000
1000
/
800
"m
N
E 600
.. ....
400
/
/
3200
....
>
""\
"
", ,
.-
24<10
1600
200
,,
aoo
skin,
and
10
10
10'
100
/
10-1
10.2
"e
80
/
/
/
20
"'"
\
\
..
........
F1u.ln>rnlllAlnl~
0
10-2
/
/
.0
; 4<1
320
60
400
\
/
/
Eq-D
5
mSy
during entire pregnancy
,,-, \
10-1
10'
Particle Energy (MeV)
,,
240
160
",
so
0
10'
FIG. 17.3. Absorbed dose rates for protons and alpha particles
<l;S a fimction of particle energy for interactions with m~cl~
tJss~e. Values correspond to a partJ.cle flux rate of 0.6 x 10 scm .
429
.-
Chapter 17
AlplUl
102
,-
j
~
'"
c~
'"
la'
.- .-
10
10.1
,.- .-
.- ,-
,-
,-
10
Particle Energy (MeV)
,- .-
,-
,-
Y
{sr-l I!C- I }
Reaction (MeV)
'H(p,n)'He
'H(d,n)'He
2.2
0
5.93
2.26
90
3.35
0.27
'H(d,o)'He
2H~t,ntHe
0.59
0.43
0.35
0.40
17.8
E.
(mSv h-l
(MeY)
0
1.2
5.24
18.3
18.7
90
0.42
2.95
14.8
14.4
nK')
0
0.64
0.24
90
0.30
0.03
0.10
0.08
0.06
0.07
la'
430
Radiological Safety
where Q = baA is the dose, ha, averaged oyer the roof area, A; I.
is the attenuation length of neutrons in air; and r is the distance
from the source to the field point of interest. This simple
expression ignores the fact that neutrons are not
monochromatic_ and. moreover, that their spectrum is affected
by scattering off the air and even the ground, a.<; the high-energy
_neutrons are likely to survive to greater distances ...
(175)
!le,
..
_.
__ . .
.'~'~-'
---~.--~--.--~-
~~
..
431
H(r)=( __
a_)
(r+b)'
exp(~)
perskyshineneutron
A[E,]
(17.6)
-~,
Chapter 17
).
(MeV)
{fSv ml}
(rn)
(m)
~fSv m'}
1.1
1424
4.5
183 2
5.7
12.2
213 3
7.4
45
267*4
9.6
125
11.3
400
13.2
14.1
14.6
1 x 10
14.7
3 x 10
14.7
432
Radiological Safety
readings in dose equivalent units such as ferns or sieverts (Sv),
whereas others indicate in absorbed dose units such as rads or
grays (Oy). Other detectors are calibrated in counts per minute
- --------------. --_._._---
_.- -.-.-
-=--~.
.~
.-
-------
17.8 CONCLUSIONS
In today's environment, a careful evaluation of the hazards
to personnel and the environment associated with accelerator
operation is a requirement (and there certainly are other hazards
in addition to radiation). Considerations of both sample
activation and prompt radiation must be addressed. Particular
attention should be paid 10 abiding by the prevailing
regulations, orders, and laws. which seem to be changing daily.
The concept of ALARA. or as low as reasonably
achievable, must also be taken into consideration. If the
evaluation (including credible accidents) shows the hazards
from radiation to be of sufficient magnitude, then radiation
interlock systems by themselves might not be adequate, and
some type of access control or shielding might also be
necessary to keep personnel safe. It is good practice to keep a
log book recording beam species, energies, and target
combinations in addition to any prompt radiation or sample
activation that is detected. This can be used as a reference for
future analyses and for indicating when a new beam-sample
433
---'--'Chapter 17
Carlo
N-Particle
eXtended,
434
.... ..:...;.,
INDEX
A
experimental geometries, 49
heavy ion, 71
high energy, 48
kinematic factor, 50-51
:mass resolution, 46-47, 70
medium energy backscattering, 71
non-Rutherford cross sections, 47-49
resonance depth profiling, 151
Rutherford cross sections, 47
sample analysis
by deficiency method, 65-67
by elastic resonance, 58
by peak integration method. 54-57
by peak widths, 60-64
by surface spectrwn heights, 64
of bulk materials, 64
of multilayer fIlms, 57
of thin fihns, 54
stoichiometry of thin films, 57-58
with overlapping peaks, 58-60
screened cross sections, 85
Beam contamination. See Accelerators
Bobr theory. See Energy straggling
Bobr velocity, 9,
Bragg's rule, 16-17, 115, 155, 156
Breit-W iilner fonnula, ISO, 151, 182'--..:::cc:.:-.-.c.-'--__'--''--': ._~ .
Bremsstrahlung
background, 234-235
atomic, 238
quasi-free electron, 238
:secondary electron, 238
Absorber foils
for deuterium-induced reactions, 139
for ERD, 90-93
Accelerators
beam contamination
causes of, 388-389
determination of, 355
in hydrogen NRA, 179
beam energy scanning, 160
beam viewers, 390-391
energy calibration
absolute, 389-390
in energy loss, 8
inERD,103
in hydrogen analysis, 185
in nuclear theory, 49
Axial half-angle. See Channeling
Background reactions
in hydrogen NRA, 178-179
Calibration
accelerator energy, 356
detector, 153-154
electronic gain, 360-361
energy, 153,356
for hydrogen NRA, 185
solid angle, 358
system geometry in ERD, 112
Calibration sources. See Properties
Center ofmass system, 38, 192
Channeling, 318
alignment, 268
angular scans, 276-277
beam collimation, 430
blocking, 269
Backscattering spectrometry
areal density, 70
Bragg's rule, 49-50
cautions, 74
comparison of techniques, 70-71 .
computer simulation, 70
depth profiling, 61-68, 16, 78
depth resolution, 51, 70, 77-78
depth scale, 50-51
elastic resonance, 78- 79
energy loss, 49-50
energy loss factor, 51
experimental considerations, 70
435
dislocations, 254
high T, supercooductors, 375
lattice vibrations, 254
point defects, 266
minimum yield, 258. See also Channeling:
normalized yield
axial channels, 260
by backscattering, 273
by charged particle activation, 226
by nuclear reactions, 273
dechanneling
defmition of, 254
experimental methods
alignment procedure, 276
angular scans, 276
Anahna, 336
Axsia, 334
measurement of
damage, 253-254
436
Carteo, 343
Dan32,335
Depth, 366
Flatus, 336
Fthie,333
GeoPIXE, 334
GISA,105,310,319,326,337,338,340,
GUPIX,343
MCERD,335-337
Monte Carlo simulation, 333-334
MEIS, 336, 364,
NDF, 319, 326, 329-333,337, 341
RBX,337
RUMP, 337
SIMNRA, 319, 326-329, 337
SPACES, 133,336
SQUEAKIE, 307
WinRNRA, 336
Computer simulation
Concentration. See also Backscattering spectrometry:
Damage
sample radiation damage, 375-376
sputtering, 376
Data analysis
Bayesian inference, 319
counting statistics, 321
deterministic vs. stochastic (Monte Carlo)
procedure, 333-334
direct calculation method, 307
Elastic Recoil Detection Analysis (ERDA), 307
fitting, 238-239
interactive vs. automated procedure, 311
MEIS,336
NRA,336
NRP,336
PIGE,336
PlXE,336
simulation by successive iterations, 308-310
Dead time. See Electronics and Pitfalls
Dechanneling. See Channeling
Depth profile
in backscattering spectrometry, 67
in ERD, 78, 83
in hydrogen analysis, 182
Depth scale
energy loss relations, 17
in backscattering spectrometry, 50-51
Detectors
BGO (bismuth genninate), 177
channeltron, 392
dosimeters, 432
electron mult.ipliers, 391-392 _::.~ ___ _. ~ __._ .__ . ~ __ ~ _
fim; b"adges, 432
for personnel exposure, 432-433
.
Ge (Li), 397
particle telescope, 399-400
photomultipliers, 300, 392-393
pulse height defect, 396
scintillation, 396
semiconductor, 396
sunace barrier, 257-258
thermoluminescent dosimeters, 432
time-of-flight, 71, 73, Ill, 353, 399
Differential scattering cross section. See Cross sections
Dosimeters. See Detectors
Double alignment, 268. See Channeling
non-Rutherford
for backscattering spectrometry, 47-49, 59
for ERD, 78-79
optical model, 35
. particle-gamma, 149-150
potential scattering, 30
resonant reactions, 176-177
Rutherford, accuracy. 363
I
I
I
I
~.
437
fluctuations additivity, 21
Epitaxial laye,,_ See Channeling
ERD. See Elastic recoil detection
Error analysis. See Statistics
Effective charge, 12
Famdaycup
design, 3S1, 391
transmission, 351
Forward recoil spectrometry. See Elastic recoil
detection
Fractional channel numbers, 361
H
Kinematic factor
for backscattering, SO-51
[orERD,84
L
Lattice defects. See Channeling
M
Mass resolution
in backscattering spectrometry, 46-47
438
N
Neutrons, thick target yic:lds
Nonnalized yield. See Channeling
Nuclear Microprobe
Beam damage, 289-290
EDRA,292
high current applications, 290-292
Ion Beam Induced Charge (TBIC). 296-297
Ion Beam Induced Luminescence (IBIL). 298-299
Ion Induced Electron Imaging, 300
low current applications. 290-291
NRA,291
PIXE,289
Scanning transmission Ion Microscopy (STIM),
294-296
Single Event Upset (SEU), 298
thne-resolved IBIC (TRIB1C), 297-298
transmission channeling, 293
Nuclear reaction analysis: particle-gamma. See also
Hydrogen NRA
background, 149
beam energy scanning, 153
depth profiling, 159-164
depthresoiution, 157-158
y detectors, 153
particle-induced y-ray emission (pIGE), 164-171
sensitivity,157-158
standards, 156-157
thick target yields, 164 .
Nuclear reaction analysis: particle-particle
depth profiling
non-resonant, 130-l32
resonant, 133-134
depth scale
non-resonant, 130-132
resonant, \J 3
filtering methods
absorber foil, 129
coincidence, 129
electrostatic and magnetic, 129
thin detector, 129
time-of-flight, 129
general principles and equipment, 127-128
geometry, 128-129
Q values, 128
standards, 134
proton-induced, 136-138
Nuclear reactions
angular momentum effects, 32
background reactions. See Background reactions
cross sections
for hydrogen analysis, 178. See also Hydrogen
NRA
theory of, 176
Nuclear resonance reactions. See Nuclear reactions
Nuclear theory. See Chap 3. See also Nuclear reactions
o
Optics
beam energy" spread, 289
chromatic aberration, 289
cylindrical electrostatic lenses, 286
electrostatic quadrupole lenses, 286
electric rigidity, 286
magnetic quadrupole lenses, 286
object aperture. 286
demagnification, 286
collimator aperture, 286
solenoid lenses, 286
spherical aberration, 286
orthomorphic focus. 288
magnetic rigidity, 286
quadrupole optics, 286
P
Pitfalls
charge exchange, 352
charge measurement, 351-352
deadtime, 353-355
~~~- .-~=
events, 351-356
ground loops, 355-356
lost beam, 351-356
multiple scattering, 365-366
particle identification, 352-353
Pulse pileup, 353-355
surface roughness, 378
PIXE
accuracy. 241
X-ray emission, 233-234
intensity-concentration relationship, 234
depth distributions, 241-242
ionization cross section, 234
useful reactions
'He-induced,141-142
alpha-induced,142-143
deuteron-induced,139-141
439
Q
Qvalu,;,s
definition of, 33
R
Radiation damage. See Damage
Radiation detectors. See Detectors ._~ ..~,.,....... __ .
Radiological safety
absorbed dose, 427-428
accelerator safety order, 426
agreement states, 426
dose equivalen~ 428
hazards
activated samples, 431
charged partjcle beams. 428
kenna, 427
National Council of Radiation Protection, 426
Nuclear Regulatory Commission., 426
quality factor, 427
special nuclear materials (SNM), 426
Range foils. See Absorber foils
Ranges
T
Tandem accelerator. See Accelerators
Thermal vibrations. See Channeling
Thiek layer analysis
by ERD,103104
Thin fIlm production, 418-419
Thin layer analysis
by backscattering spectrometry, 54
S
Safety. See Radiological safety
Sample damage. See Damage
Scattering cross section, 314-315 See also Cross
sections
Scattering kinematics, 365
Screening. See Cross sections
440
by ERD, 103
Thomas-Fermi screening function, 250
Time-of-flight. See Detectors
Transmission ERn. See ERD
Vacuum practice
cleanliness, 418
leak repair, 418
441