PHYS 352 Energy Loss by Electrons
PHYS 352 Energy Loss by Electrons
• two things happen: a) the particle loses energy traversing matter and b)
particle is deflected from its initial direction
• two main processes cause this: 1) inelastic collisions with atomic electrons in
the material and 2) elastic scattering off nuclei
• other processes also cause energy loss: 3) bremsstrahlung, 4) emission of
Cherenkov radiation (relative of bremsstrahlung), 5) nuclear reactions (rare,
lower probability)
• makes sense to separate the consideration of heavy charged particles and
light charged particles (i.e. electrons)
• heavy particles don’t undergo 3) and 4); 5) is rare; 2) is again less common
compared to 1)...and heavy particles don’t deflect much off electrons
• basically need only consider inelastic collision with atomic electrons for
energy loss of heavy charged particles
Differences between Electrons and
Heavy Charged Particles
• previous slide accounted for the possibility of large energy loss in a single
collision and the effect of indistinguishability on the scattering cross section !"#$%&$'()*
and kinematics in the modified Bethe-Bloch formula
o +,-.'-.'.,/01'2/34'456789:'-1'1458';36<,.'=1/84'8,68'8,4.4'634'
>/3'494783/1.'0,-7,'6.'04'0-99'.44'?49/0'634'6'.<47-69'76.4'/>'
• the straggling for electrons is observed to be “larger” than for heavy particles
7,63;4@'<638-794A'BC8'8,4'7C3D4.'-99C.83684'.836;;9-1;'D43:'0499EF'
because each scattering event could have a significant fractional energy loss
• for thin absorbers, still can '
describe
GFG with!" #$%&''$(
Landau distribution
$) &*" + ,-.*" /01)2" &0' #$)3" 40 5361" 7$1" 8 9:; <=>>9? ;@:AB9>
'
O,0'+# (1 $/+ 7(#$ 6%(&,&0+ +)+%35 0(##E !!6 E 0"7"$,$"()# (1 $/+ $/+(%+$"*,0 7+$/(!# +760(5+!
2+%+ (&$,")+! 1%(7 $/+ @()$+ P,%0( !,$, ") $/"# 2(%<- ./+ 6%+#+)$ ,03(%"$/7 "# #'\*"+)$05
IBFE:FL- ./+ !"Q+%+)*+ #++) ") $/+ 7(#$ 6%(&,&0+ 3+)+%,0 $( &+ ,660"+! $( O,%"('# ")$+%!"#*"60"),%5
+)+%35 0(## "# %+0,$"O+05 #7,00 2"$/") $/+ ,**'%,*5Particles 8+0!#E1- 20 .doc %,!",$"() $/+%,65E -2/+)+O+%
")*0'!")3 13 -
(1 $/+ +>6+%"7+)$,0 !,$, 1(% 7(#$ (1 $/+ %+#'0$# +0+*$%() $%,)#6(%$ ") 7,$$+% (% $"##'+ "# +)*(')4
#/(2) ") .,&0+ ?- ./+ *(76'$+! O,0'+# (1 !!6 $+%+!-
Radiative Energy Loss for Electrons
Cherenkov Radiation
• EM “shock wave” occurs when vparticle > c/n; with emission at cos θ = 1/βn
• the electric field of the particle interacts with atoms in the material and the
particle is “right at” the atom before the field gets there
fiction
fact
σ ∝ ( mce 2 )2
2
• emission probability (bremsstrahlung cross section):
• bremsstrahlung cross section dσ/dν can be calculated from first principles
but there is screening of the nuclear charge
by the surrounding atomic electrons energy loss for e– and p in Cu
2
dσ 2⎛ e ⎞ 1
2
= 4α Z ⎜ {screening plus…}
dν ⎝ me c 2 ⎟⎠ ν 1 MeV
ν0
1 dσ
Φ rad =
E0 ∫ hν dν (E ,ν )dν
0
0
⎛ dE ⎞
− ⎜ ⎟ = N E0 Φ rad
⎝ dx ⎠ rad
ν 0 = E0 / h
• bremsstrahlung cross
2
section depends on mostly these terms
dσ ⎛ e2 ⎞ 1
= 4α Z 2 ⎜ {screening plus…}
dν ⎝ me c 2 ⎟⎠ ν
• with only a weak dependence on E0, the incident electron energy which is in
the { } in the equation above
• the fine structure constant α=1/137
• ν is the frequency of the emitted bremsstrahlung photon
• the total energy loss by bremsstrahlung radiation would be the number of
nuclear targets in the slice dx times
ν
the integral of the cross section times the
energy per photon − ⎛ dE ⎞ d σ 0
⎜⎝ = N hν
⎟
dx ⎠ rad ∫
(E , ν )dν
0
dν
0
• and since the cross section goes approximately as 1/ν, inside the integral there
is no ν dependence and so integrates to terms times hν0 = E0
• thus, we can write ν
⎛ dE ⎞ 1 0
dσ
−⎜
⎝ dx ⎟⎠ rad
= N E0 Φ rad ; Φ rad =
E0 ∫ hν dν (E ,ν )dν
0
0
• which is just a way to say that dE/dX depends on E0 and an integral that
doesn’t because it is written so that E0 cancels (approximately)
Bremsstrahlung versus Ionization Energy Loss
dE
• for E>Ec, radiation loss dominates and we re-write: − = N Φ rad dx
E
• defines the quantity radiation length Lrad: ⎛ −x ⎞
E(x) = E0 exp ⎜
⎝ Lrad ⎟⎠
Multiple Scattering
• multiple scattering causes the direction to change (and the total path length
travelled to be much longer than the penetration depth) and can be significant
for electrons
• normally dE/dx and scattering treated independently; but we know that’s not
true as large angle scattering comes from large momentum (energy) transfer
and a small angle scatter was a small energy transfer
• there are Gaussian approximation formulae describing angular distributions
no real point writing them out...
functions o f target atomic number Z. Values of eight constants to the equation is about 7 ~ .
ired. a function of Z only. Several equations to express the
s the relation between q and Z have been formulated through 1. I n t r o d u c t i o n energy region from about 10 to 100 keV, the coeffi
the simple theoretical treatments or semiempirical ap- In the detection and utilization of electrons, accurate fl is almost independent of T, and can be expresse
r of proaches 26-3°) (a brief review of previous works knowledge
is of backscattering is frequently required. a function of Z only. Several equations to express
s in- given in the appendix). At higher energies, the equa- One of the quantities featuring this phenomenon is the relation between q and Z have been formulated thro
ation
ange
Backscattering
tion for q should express the dependence on both Z
and T. Although such equations have also been pro-
backscattering coefficient defined as the ratio of the
number of backscattered electrons to the number of
simple theoretical treatments or semiempirical
proaches 26-3°) (a brief review of previous work
incident electrons. The backscattering coefficients in- given in the appendix). At higher energies, the e
cient posed,a, ~9,2~,3~), their regions of validity have rather
crease with increasing target thickness until saturation tion for q should express the dependence on bot
im- inconvenient limitations. The present paper describes and T. Although such equations have also been
is reached at a thickness around half the practical range
nce, a•new
because
empiricalthey are valid
equation so light,
in widerthe backscattering
regions of the incidentprobability
of Z is significant
electrons. The backscattering for
coefficient posed,a, ~9,2~,3~), their regions of validity have ra
been andelectrons and not so significant for heavy charged particles
T than the previous ones. r/ at saturation of the monoenergetic electrons im- inconvenient limitations. The present paper desc
pinging normally on the target is of general importance, a new empirical equation valid in wider regions
2. Formulation
trary • either from one large deflection orand multiple scattered
a good deal of experimental
When experimental data for r/ are plotted as reported
a i - 25),.
data for q have been and T than the previous ones.
netic
function of T on a semilogarithmic paper, the data for In order to obtain the value of r/ for the arbitrary 2 . F o r m u l a t i o n
o use
• this is an important consideration foratomic
T > 50 keV form S-shaped curves as shown in fig. 1.
electron
numberdetectors;
Z of the target the backscattered
and the incident kinetic When experimental data for r/ are plotted
most
the electron deposits only a fraction of its
These curves suggest that the relation between ~1and T energy
energy T of in
the the detector
electrons, it would be convenient to use function of T on a semilogarithmic paper, the data
the empirical equation which well reproduces the most T > 50 keV form S-shaped curves as shown in fi
may be expressed by an equation of the form
• backscatter fraction η depends on target thickness, reaching saturation at may be expressed by an equation of the form
probable values given by the existing data. In the These curves suggest that the relation between ~1an
T~ q = !al {1 - tgh [!(lna2 + a31nz)] },
half the practical range
~
* For energies below 50 keV, see the references listed in ref. 1. o.3 ~
0.6;7q-~--~7
0.5 ~ - ~ ~
Au I
.
i irqqqT~
,. "
r r ~ jlllll -.
j f iiiiiT----r ~ i Ill ir ! rT~ q = !al {1 - tgh [!(lna2 + a31nz)] },
* For energies below 50 keV, see the references listed in r
0.6;7q-~--~7 i irqqqT~ r r ~ jlllll
0.5 ~ - ~ ~ ,.
Au I
0.2 ~ . "
o 4 ! Agt " ~ ~ - \\
o 4 ! Agt " ~ ~
0:5
02 "" O.I 0:5
02 ""
0, I -- 0, I --
0 LLt_iO.I
l] I ! I[llrl 0 LLt_iO.I
l] I ! I[llrl
T (MeV) T T (MeV)
I IO o.I I lo I IO
(MeV)
(b) (a) (b)
Fig. I. Backscattering coefficient q of electrons as a function of incident energy 7". (a): C, AI, and Cu; (b): Ag and Au.
ion of incident energy 7". (a): C, AI, and Cu; (b): Ag and Au. Solid
circle: experiment 2 ~1); cross: M o n t e Carlo calculationa:3); solid line: present empirical equation. The dashed lines in (b) in
; solid line: present empirical equation. The dashed lines in (b) indicate
the lower limit T~ to the energy region considered in the least-squares fit. Experimental data around the curves of Ag an
e least-squares fit. Experimental data around the curves of Ag and Au at T > 15 MeV are for Cd and Pb, respectively.
e for Cd and Pb, respectively.
509
509