Medical Physics 1A03

Physics in Medicine and Biology

Lecture 9: X-Ray and CT

Useful Constants
h is Planck constant = 4.135667696...×10−15 eV⋅Hz−1

ℏ is the reduced Planck constant = h/2π = 6.582119569...×10−16 eV⋅s

ℎ
is the Compton wavelength of the electron = 2.43 x 10-12 m
𝑚𝑒 𝑐

r0 is the classical electron radius = 2.8179x10-15 m

me is the electron mass = 9.1093837 × 10-31 kg

c is the speed of light = 299,792,458 m/s

mec2 = 511 keV

2
Useful Equations
𝐸0
E= h f ->EM wave energy 𝐸𝑠𝑐 = 𝐸0 -> Compton scatter
1+ 1−cos 𝜃
𝑚𝑒 𝑐2
ℎ𝑐
𝜆𝑚𝑖𝑛 = 𝑒𝑉 -> xray production

𝐼 = 𝐼0 exp(−μx) -> attenuation

𝐼𝑠 𝑟02
= 1 + cos 2 𝜃 -> Thomson Scatter
𝐼𝑝 2𝑟 2

ℎ
Δ𝜆 = 𝑚 𝑐 (1 − cos 𝜃) -> Compton Shift
𝑒

2
1 𝛼 2 1 − cos 𝜃 2
𝐹𝐾𝑁 = 1+
1 + 𝛼(1 − 𝑐𝑜𝑠𝜃) 1 + 𝛼(1 − 𝑐𝑜𝑠𝜃) (1 + cos2𝜃)

ℎ𝑓 𝐸𝑝ℎ𝑜𝑡𝑜𝑛
𝛼= = 2
𝑚𝑒 𝑐 2 𝑚𝑎𝑠𝑠𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑛 ∙ 𝑣𝑙𝑖𝑔ℎ𝑡
3
X-Rays
• They are a form of electromagnetic radiation, in the same manner as visible light but
with a very short wavelength, around 0.25 - 25 Å
• As EM, X-rays have a specific energy.
• The characteristic range is defined by soft versus hard X-rays.
• Soft X-rays cover the range from hundreds of eV to a few keV,
• Hard X-rays have an energy range from a few keV up to around 100 keV

4
Light as a Particle - Review
• Light and other electromagnetic radiation behave as if composed of small packets of
quantized energy
• These packets of energy are called photons
• For a given frequency f of the radiation, each photon has a fixed amount of energy E
which is:
E= h f
• Where h is Planck’s constant and is equal to 6.55×10−34 J⋅s.

5
Photoelectric Effect - Review
• Light impacting a metal surface causes the
ejection of electrons from that metal
• Led to the understanding of light as
quantized packets of energy
• If the wavelength of the incident light was
too high (e.g. red), then no electrons are
ejected, even if the intensity of the light
was very high or it was shone onto the
surface for a long time

6
Photoelectric Effect
• As the wavelength decreases beyond the
threshold, the ejected electrons simply move
faster (increased kinetic energy)
• This feature is exploited to be able to produce
x-rays
• To better understand the process of x-ray
production, we need to understand the Bohr
model of the atom

7
Bohr’s Model of the Atom
• Compares an atom to a planetary system
• We have a central nucleus of protons and
neutrons, with orbiting electrons
• The electrons orbit in ‘stable shells’ that have
discrete distances from the nucleus

8
Bohr’s Model of the Atom
• Orbit distances are where the angular momentum of the
electron is an integer multiple of the reduced Planck constant
𝑚𝑒 𝑣𝑟 = 𝑛ℏ
• Where me is the electron mass (9.1093837 × 10-31 kg), v is
tangential velocity and r is the orbit distance, and n = 1, 2, 3,
… which is the principal quantum number
• The reduced Planck’s constant (h/2π) is:
• 1.054571817...×10−34 J⋅s; or
• 6.582119569...×10−16 eV⋅s

9
Bohr’s Model of the Atom
• The last postulate of the model was that electrons only gain
or lose energy by switching between orbit shells
• This is done by either emitting or absorbing EM radiation
• The energy is related to the frequency of the emitted photon
by Planck’s constant (not reduced!):
Δ𝐸 = 𝐸2 − 𝐸1 = ℎ𝑓
• Where h is:
• 6.62607015×10−34 J⋅Hz−1; or
• 4.135667696...×10−15 eV⋅Hz−1

10
X-Ray Production
• X-rays are commonly produced in X-ray tubes by accelerating electrons through a
potential difference (a voltage drop) and directing them onto a target material (i.e.
tungsten).
• Incoming electrons release X-rays as they slowdown in the target (braking radiation or
bremsstrahlung)
• Photons produced in this manner range in energy from near zero up to the energy of
the electrons

11
X-Ray Production
• An incoming electron may also collide with an atom in the target, kicking out an
electron and leaving a vacancy in one of the atom’s electron shells.
• Another electron may fill the vacancy and in so doing release an X-ray photon of a
specific energy (a characteristic X-ray)

12
X-Ray Production Hardware
• For medical imaging, X-rays are generated by an X-ray
tube, a vacuum tube that uses a high voltage to
accelerate the electrons released by a hot cathode to a
high velocity.
• The high velocity electrons collide with a metal target,
the anode, creating the X-rays
• In medical X-ray tubes the target is usually tungsten or a
more crack-resistant alloy of rhenium (5%) and tungsten
(95%), but sometimes molybdenum for more specialized
applications, such as when softer X-rays are needed as in
mammography.
• Highly inefficient – about 1% of energy is converted to x-
rays
13
X-Ray Production
• Maximum energy of the produced X-ray photon is
limited by the energy of the incident electron
• This is equal to the voltage on the tube times the
electron charge, so an 80 kV tube cannot create X-
rays with an energy greater than 80 keV.
• When the electrons hit the target, X-rays are created
by two different atomic processes:
1. Characteristic X-rays
2. Bremsstrahlung

14
Characteristic X-ray emission
• If the electron has enough energy, it can knock an
orbital electron out of the inner electron shell of
the target atom.
• After that, electrons from higher energy levels fill
the vacancies, and X-ray photons are emitted.
• This process produces an emission spectrum of X-
rays at a few discrete frequencies, sometimes
referred to as spectral lines.

15
Characteristic X-ray emission
• Usually, these are transitions from the upper shells to
the K shell (called K lines), to the L shell (called L lines)
and so on.
• If the transition is from 2p to 1s, it is called Kα, while if it
is from 3p to 1s it is Kβ.
• The frequencies of these lines depend on the material of
the target and are therefore called characteristic lines.
• The different target atoms have different nuclear
structures which support different orbital shells

16
Characteristic X-ray emission
• We can consider the different possible characteristic peaks that
we might expect to see from tungsten
• Assuming the incident photon had E ≥ 69.525 keV, then a K shell
electron could be ejected
• If it is replaced by a LIII shell electron, the Kα1 photon would
have energy:
• 69.525-10.204 = 59.321 keV
• For reasons beyond this course, not all transitions are possible.
• In this case LI -> is not possible so you only will see a Kα1 + Kα2
doublet peak

17
Characteristic X-ray emission
• Here we can see how the characteristic peaks
change with the accelerating voltage
• The probability of getting characteristic x-rays
increases with photon energy
• Why are there no L-shell lines?
• X-ray tubes have aluminum filtering to block
these low energy x-rays as they increase
radiation dose but aren’t productive for
imaging

18
Characteristic X-ray emission
If you had a non-medical target without the
filters, the spectrum would look different.

19
iClicker Question
An electron making which of the following transitions would produce the longest
wavelength photon
A) L -> K
B) M -> K
C) M -> L
D) N -> K
E) N -> M

20
Bremsstrahlung Radiation
• We would normally expect to only detect the characteristic photon
energies from atoms, but we end up seeing a continuous spectrum
• Bremsstrahlung or braking radiation is any radiation produced due to
the acceleration (positive or negative) as photons are scattered by the
strong electric field near the nuclei
• The photon flux for each energy or wavelength is approximately
described by Kramers' law
𝜆 1
𝐼 𝜆 𝑑𝜆 = 𝐾 − 1 2 𝑑𝜆
𝜆𝑚𝑖𝑛 𝜆
• The constant K is proportional to the atomic number of the target
element, and λmin is the minimum wavelength

21
Duane–Hunt law
• Gives the maximum frequency of X-rays that can be emitted
by Bremsstrahlung in an X-ray tube by accelerating electrons
into a metal target
𝑒𝑉
𝑓𝑚𝑎𝑥 =
ℎ
• Where h is Planck constant, e is the electron charge
(1.602176634×10−19 C) and V is the accelerating voltage
• We can use this to get the minimum wavelength
ℎ𝑐
𝜆𝑚𝑖𝑛 =
𝑒𝑉
• Where c is the speed of light

22
Bremsstrahlung Radiation
• To understand this distribution of x-rays first have to look at what would happen if we
had a very thin target material
• If electrons of energy E1 hit a thin target, the intensity of the radiation emitted in any
energy interval from 0 to E1 is constant – there is a continuous distribution of photon
energies
• If the energy of the electron beam is reduced to E2 the intensity would be constant
from interval 0 to E2.

23
Bremsstrahlung Radiation
• Consider a thick target as the superposition of several thin targets
• Electrons of initial energy E1 after passing through a thin layer of the target will have
an energy of E2 and the corresponding spectrum with it.
• After passing through the second thin layer they will have energy E3 etc.
• The total spectrum for the thick target can therefore be superposition of all the thin
layer target spectra

24
Bremsstrahlung Radiation
• The straight line can be represented by:
𝐼 𝐸 = 𝐾 ∙ 𝑍 ∙ (𝐸𝑚𝑎𝑥 − 𝐸)
• Emax is the energy of the electrons hitting the target and is also the maximum emitted
photon energy (in this case = E1)
• These photons occur when the electron is brought to a complete rest in one
interaction and is a very unlikely event.
• K is a constant and Z is the atomic number of the target material

25
Bremsstrahlung Radiation
• The lower energy photons have low penetrating depth (are absorbed quickly), and
might not make it out of the target
• This modifies the distribution of emitted photons
• The high energy photons escape the target, and that part of the curve is unchanged
• As lower energies the likelihood of absorption increases until at some energy, no
photons escape and become heat energy in the target

26
X-Ray Production
• So, the resulting output of a tube consists of a continuous Bremsstrahlung spectrum
falling off to zero at the tube voltage, plus several spikes at the characteristic lines.
• The voltages used in diagnostic X-ray tubes range from roughly 20 kV to 150 kV and
thus the highest energies of the X-ray photons range from roughly 20 keV to 150 keV.
• Both of these X-ray production processes are inefficient, with only about one percent
of the electrical energy used by the tube converted into X-rays, and thus most of the
electric power consumed by the tube is released as waste heat.

27
X-Ray Production
• When producing a usable flux of X-rays, the X-
ray tube must be designed to dissipate the
excess heat.
• For this reason, the target metal is usually
attached to a motor that allows it to spin, so no
one area of the target is continuously under the
electron beam

28
X-Ray

29
X-Ray Contrast

• As the x-rays pass through the tissue, some x-

rays will interact with the tissue and others will
pass straight through
• Differences in atomic make-up of tissue varies
the likelihood of interaction
• This results in changes in the intensity of the
transmitted beam
• Different tissues have differing molecular
makeups that changes this interaction strength

30
X-Ray Contrast comes from Attenuation
• We can formulate this as thinking about a change in intensity will
depend on the initial intensity and the material thickness:
Δ𝐼 ∝ −𝐼 Δ𝑥
• We need a proportionality constant that describes the probably of
interaction per unit length of a material
• This is called the linear attenuation coefficient (LAC) and is represented
by μ
• LAC represents the fraction of photons that interact per unit thickness
of a given material (1/length)
• This expands our equation to:
Δ𝐼 = −μI𝛥x

31
X-Ray Contrast comes from Attenuation
• When you see Δ, you should think about it as a differentiable equation
𝑑𝐼 = −μI𝑑x
• Bring intensity to one side and then solve:
1
𝑑𝐼 = −μ𝑑x
𝐼
• Integrate left side by dI, and right by dx:
ln 𝐼 = −μx + C
• C is the integration constant. We want x= 0 to be I0, so C = ln(I0)
𝐼0
ln 𝐼 − ln 𝐼0 = −μx =ln
𝐼
𝐼
= exp(−μx)
𝐼0
32
X-Ray Contrast comes from Attenuation
𝐼 = 𝐼0 exp(−μx)
• Object contrast comes from:
• Difference in transmitted intensity
• Which is a difference in attenuation integrated
along pathlength
• Which is a difference in μx
• This equation only holds for a monochromatic beam,
otherwise the LAC is not constant
• It also requires narrow beam geometry so we can
ignore scatter

33
Radiation and ex
• Many radiation effects decay exponentially
• Euler’s number (2.718…) and the associated exponential function (ex) come up
frequently, so get used to seeing it!
• It has many useful properties, such as being the derivative of itself, and it can be
used to ‘spin’ complex numbers, we also saw it in the Fourier transform
• In the next set of lecture notes, we will see how it is used to model radioactive
decay

34
X-Ray Attenuation
• The LAC will be dependent on the material but also the photon energy – some
examples given below
• LAC generally increases with increasing atomic number of the absorbing material,
but interaction events can change that trend (e.g. lead)

35
iClicker Question
Based on the data on the previous slide, which of the following would you want to
shield you from x-rays?
A) Aluminum
B) Copper
C) Lead

36
X-Ray Attenuation

37
Mass Attenuation Coefficient
• A variation is the mass attenuation
coefficient which normalizes the LAC to
the density of a material
• This provides a value that is per unit
mass rather than per unit distance
• Characterizes how easily a mass of
material can be penetrated
• For x-rays, this is typically expressed as
cm2/g (but the SI unit it m2/kg for
attenuation of other waves such as
sound or light)

38
iClicker Question
If a material of thickness X attenuated N% of the x-rays incident on it, how many x-rays
would be attenuated if that same material was compressed to a thickness of X/2
A) 2N %
B) N %
C) ½ N %

39
Mass Attenuation Coefficient
• The mass attenuation coefficient is:
𝜇𝑙𝑖𝑛𝑒𝑎𝑟 𝑚2
𝜇𝑚𝑎𝑠𝑠 =
𝜌 𝑘𝑔
• Simply – it is the LAC normalized for density, which removes effects of stretching or
compression of material on attenuation
• Some of the photon interactions are absorptions, while others are changes in energy
from bremsstrahlung radiation
• We can then divide this further into transfer and absorption effects:
𝜇𝑡𝑟𝑎𝑛𝑠𝑓𝑒𝑟 𝜇𝑎𝑏𝑠𝑜𝑟𝑏
𝜇𝑚𝑎𝑠𝑠 = +
𝜌 𝜌

40
Half Value Layer
• A commonly referenced property of an object meant for radiation shielding is the
Half Value Layer (HVL)
• To find it, we want to find the thickness x, where the transmitted intensity It is 0.5 I0
• Let’s start with our previously derived equation for linear attenuation:
𝐼𝑡 = 𝐼0 exp(−μx)
• Make the substitution and solve for x
0.5𝐼0 = 𝐼0 exp(−μx)
0.5 = exp(−μx)
ln 0.5 = −0.693 = −μx
𝑥 = 0.693/μ

41
Half Value Layer
• This only applies for narrow beam geometry – such that we can ignore scattering
effects
• Since μ is only applicable to monoenergetic beams, so it this formula
• For polyenergetic beams, you get an interesting effect since the lower energy x-rays
are typically the first to be absorbed
• Thus, the beam passing through the 1st HVL will have a higher average energy
• It generally takes more material to absorb higher energy x-rays, so for the case of
the polyenergetic beam, 2nd HVL > 1st HVL
• To better understand attenuation, we need to understand the different interactions
that can occur between x-rays and a shielding medium

42
Thomson scattering
• If a photon is incident on a free electron, the electric field will cause the electron to
oscillate at the same frequency as the incident wave
• This result in radiation energy of the same phase and frequency as the incident
photon
• If incident photon will have two perpendicular electrical components E1 and E2
• When the photon reaches the particle, the particle experiences force:
𝐹 = 𝑘𝑒𝐸1
• This force from the electric field component E1, with charge e and constant k

43
Thomson scattering
• This force will accelerate the electron (a1)
• The electron will radiate this energy in the form of an EM wave at point S, with the
new electric vector E1’
𝑘𝑒 2 𝐸1
𝐸1 ′ = 2
sin Φ
𝑚𝑒 𝑐 𝑟
• The first term is called the ‘classical electron radius’ r0 (2.8179x10-15 m), which
simplifies this expression to:
𝑟0
𝐸1 ′ = 𝐸1 sin Φ
𝑟

44
Thomson scattering
• If we take direction 1 to be the scatter direction, then E2’ is more simply:
𝑟0
𝐸2 ′ = 𝐸2
𝑟
• The wave intensity is related to the square of the electric vector. At the electron it is:
𝐴 𝐸12 + 𝐸22
• The intensity at scatter point S is:
𝑟0 2 2 2
𝐴 𝐸1 sin Φ + 𝐸22
𝑟
• We can compute the fraction of energy that was scattered to point S as:
𝐼𝑠 𝑟02
= 2 1 + cos 2 𝜃
𝐼𝑝 2𝑟
Note the change from sin to cos from change in the angle used in the equation
45
Thomson scattering
𝐼𝑠 𝑟02
= 2 1 + cos2 𝜃
𝐼𝑝 2𝑟
• In the brackets, cos2 ranges from -1 to 1,
but the +1 brings it to 0-2. The division by
2 renormalized this to be angle-dependent
multiplier between 0.5 and 1.
• So, a backwards scattered photon will have
half the intensity of a forward scattered

46
Thomson scattering
𝐼𝑠 𝑟02
= 2 1 + cos 2 𝜃
𝐼𝑝 2𝑟
• This represents the fraction of the incident energy scattered by the electron in the
solid angle dΩ =1/r2
• Which is normally written as:
𝑑𝜎0 𝑟02
= 1 + cos2 𝜃
𝑑Ω 2
• This is called the classical scattering coefficient per electron per unit solid angle at
the scatter angle θ.
• At θ=0 and θ=90, it has a value of r02 and at θ=90 a value half this

47
Brief Aside – Solid Angle
• Solid angle (Ω) is a measure of the amount of the field of view from
some particular point that a given object covers
• In other words, it is a measure of how large the object appears to
an observer looking from that point
• From the image, we have a circle with area A, drawn on a sphere
• The center of the circle is distance ‘r’ away from the center of the
sphere
• We quantify the area that this circle covers in ‘steradians’ as:
𝐴
Ω= 2
𝑟
• This is effectively the 3D analogue to the 2D radian of the arc
subtended in a unit circle
48
 Thomson scattering
𝑑𝜎0 𝑟02
= 1 + cos2 𝜃
𝑑Ω 2
• σ is the cross section, and we must integrate from angles 0 to 180 deg to get the
total scattering cross section
• This is called the Thomson cross section and is:
8𝜋 2
𝜎𝑇 = r0 = 0.665 ∙ 10−24 cm2
3
• The scattering cross-section is the equivalent area of the incident wavefront that
delivers the same power as that re-radiated by the particle following scatter
• Note the frequency is unchanged by this process, and this process is symmetric for
forward and back scattering directions

https://farside.ph.utexas.edu/teaching/jk1/Electromagnetism/node107.html 49
 Thomson scattering
• Thomson scattering is the low-energy limit for Compton scattering (to come…)
• The low energy case holds when the photon energy (hf) << the mass energy of the
particle (E=mec2)
• mec2 is a known constant of 511 keV
• Typical diagnostic x-rays range from ~15 keV to 80 keV, which is too high for this to
be probable

https://farside.ph.utexas.edu/teaching/jk1/Electromagnetism/node107.html 50
Coherent Scattering
• Coherent scattering only involves a change in direction of the photon, with none of
the photon energy converted to kinetic energy
• This is an interaction between an incident photon and the entire atom
• The photon’s EM field absorbed and causes the electrons to oscillate
• These electrons in turn emit radiation of the same wavelength as the incident
photon
• Might also see it labelled as elastic or Rayleigh scattering

51
Coherent Scattering
• Thompson scattering describes scattering from a free electron.
• If the electron is bound to an atom, the situation is more complex.
• If an EM wave is incident on an atom of atomic number Z, the amplitude of the
scattered wave will not necessarily be the amplitude of the scatter from one
electron multiplied by Z
• This is because interference effects can take place from waves emitted from all
electrons
• If each electron scatters in the forward direction they will all be in phase and
constructively interfere
• If this is in any other direction, the scatter from each electron will be out of phase to
some degree and destructive interference will occur resulting in reduced amplitude

52
Coherent Scattering
• There is no excitation or ionization and hence no energy loss by the scattered
photon (photon energy goes unchanged through this event)
• It is the constructive interference from the radiation emitted by each electron in the
atom that forms the scattered photon.
• Coherent scatter is predominantly a forward scatter event as at higher angles the
waves from each atom get out of phase and tend to destructively interfere.
• The probability of coherent scattering is proportional to Z and 1/E2
• Due to the inverse relationship with energy, this method of attenuation contributes
very little at energies above 10 keV
• This energy is typically below what is used for diagnostic imaging, so it should not
contribute much to image quality

53
Incoherent (Compton) Scattering
• Compton carried out some work that shows the particle like behavior of radiation
• He used a monochromatic beam of x-rays with a single wavelength, and he made
them fall on a target material
• He then measured the intensity of the scattered beam at various angles
• Although the incident beam had only one wavelength, the scattered x-rays had two
intensity peaks at two different wavelengths, one being the same as the incident
and one being longer (or one the same energy as the incident and one with reduced
energy).

54
 Incoherent (Compton) Scattering
• The change in wavelength ∆λ is called the Compton shift.
• The fact that there was a peak with reduced energy cannot be understood by
considering the incident x-rays as a classical electromagnetic wave, as a scattered
wave can only have the same wavelength (hence energy) as the incident wave
• We can calculate the Compton shift using:
ℎ
Δ𝜆 = (1 − cos 𝜃)
𝑚𝑒 𝑐
ℎ
• is known as the Compton wavelength of the electron = 2.43 x 10-12 m
𝑚𝑒 𝑐

• For forward scatter, there is no shift in wavelength, where for a backwards scattered
photon, the shift is twice the Compton wavelength

h is Planck constant = 4.135667696...×10−15 eV⋅Hz−1 55

Incoherent (Compton) Scattering
• Compton suggested that instead of a wave of
frequency f, the incident beam consisted of
photons each having energy
E=hf
• These photons behave as a particle in the collision
with electrons, scattering like two billiard balls.
• If the electron in the target is considered
stationary before the collision, it will gain kinetic
energy when hit by the incoming photon and
hence the incoming photon must lose energy
through the collision process

56
 Compton Scattering
• Photon interaction, that knocks an orbital electron
out of its shell
• The scattered photon loses energy from this collision
• The energy lost is based on the scatter angle θ
𝐸0
𝐸𝑠𝑐 =
𝐸0
1+ 1 − cos 𝜃
𝑚𝑒 𝑐 2
• Forward scatter becomes more probable at higher
photon energies
• Compton scattering is the most probable scattering
process for high energy x-rays

https://radiologykey.com/x-ray-imaging-fundamentals/ 57
Compton Scattering
• What is the probability of a Compton event occurring?
• Klein and Nishina showed that the differential cross section per unit solid angle
is the Thomson scattering probability multiplied by a factor FKN
𝑑𝜎𝑐𝑜𝑚𝑝 𝑟02
= 1 + cos 2 𝜃 ∙ 𝐹𝐾𝑁
𝑑Ω 2
• Where the Klein and Nishina factor is defined as:
2
1 𝛼 2 1 − cos 𝜃 2
𝐹𝐾𝑁 = 1+
1 + 𝛼(1 − 𝑐𝑜𝑠𝜃) 1 + 𝛼(1 − 𝑐𝑜𝑠𝜃) (1 + cos2𝜃)
Where:
ℎ𝑓 𝐸𝑝ℎ𝑜𝑡𝑜𝑛
𝛼= 2
= 2
𝑚𝑒 𝑐 𝑚𝑎𝑠𝑠𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑛 ∙ 𝑣𝑙𝑖𝑔ℎ𝑡
58
Photoelectric
• As previously mentioned, we also have photoelectric effects
with x-rays
• The photoelectric process is most likely to occur if the incident
photon energy is just above the binding energy of the
electron.
• If the energy is just less than the binding energy of the
electron, then it cannot be ejected
• The fluorescent yield is a parameter describing the relative
probability of emission of characteristic radiation to the
emission of an Auger electron.
• ωK is the number of K x-ray photons emitted/number of K
shell holes
59
Auger Electron
• Typically, when an inner-shall vacancy is filled
by an upper shell electron, a photon is released
• This energy could be absorbed by another
electron in the atom, causing it to be ejected
from the electron cloud of the atom
• This second ejected electron is called an Auger
electron
• The probability of this effect changes with
atomic number (Z) as seen in the plot here and
in the table on the last slide

60
Photoelectric
• The linear photoelectric coefficient, τ, depends on both the energy of the incident
photon and the atomic number of the absorbing material.
• In the diagnostic imaging energy range:
τ ∝ 1/E3
• These plots show what we might expect – bone stops more x-rays than soft tissue

61
Photoelectric
• The dependence on atomic number varies between Z3 and Z4
• For low atomic number materials (i.e. humans) it is generally taken to be:
τ ∝ Z3
• Bone has a lot of calcium (Z=20), where soft tissue is mostly water (Oxygen Z = 8)

62
Photoelectric
• This presents the photoelectric effect as a way to completely
absorb an x-ray (i.e. no photon continues on after this event)
• From the table, we see that the likelihood of complete
absorption decreases with Z
• We saw that the likelihood of complete absorption decreases
with photon energy
• The other way to absorb high energy photons is pair-
production (discussed in the next set of lecture notes)
• Multiple scattering events can bring high-energy x-rays down
to an energy level where they are more likely to produce an
auger electron with no fluorescent yield

63
iClicker Question
Which reason best describes why the metallic implant is so bright in this x-ray image?
A) It blocks less x-rays so it appears brighter on the image.
B) It appears brighter due to the Compton shift altering the wavelength of the
detected x-rays
C) The titanium implant (Z=81) has a high probability of the photoelectric effect and
coherent scattering

64
Not Understanding?
• A website with nice figures, videos and alternate text to describe these x-ray
interactions can be found here: https://howradiologyworks.com/x-ray-interactions/

65
Effects on Imaging
• Given the energy and Z dependence of the
three interactions discussed above, we can
now picture what happens when we place
a patient between a typical diagnostic x-
ray spectrum and an imaging device
• Here we see the difference in the
attenuation coefficients for soft tissue and
bone over a diagnostic imaging energy
range (note y-axes labels!)
• We also see the energy dependence of the
different attenuating effects

66
Diagnostic X-Ray Imaging
• We have covered how x-rays are produced, and
how they can be differentially blocked by tissue
(i.e. skin vs bone)
• We then need some way to detect the
transmitted x-rays to be able to assess the
tissue
• We also need to be able to localize x-rays
(typical resolution is ~0.1 mm)

67
X-Ray Detectors
• Most detectors are digital now using photostimulable
phosphor plate (PSP)
• When X-rayed, excited electrons in the phosphor
material become 'trapped' in the crystal lattice
• To obtain the x-ray detection information, a laser beam
raster scans the plate, which releases the stuck
electrons
• This stuck electron returns to its ground state,
releasing a photon
• The emitted light from the electrons returning to
ground state is collected by a PMT, and the resulting
signal is digitized
68
 X-Ray Detectors - PSP

• Interesting to note that the red laser that is

used to read the PSP is a lower energy photon
than the blue that is emitted
• This is because the electrons were already at
an elevated energy state, and the red light
‘topped them off’
• In this case, the x-ray is an inverse image of the
x-ray counts
• Black regions received a high x-ray dose, and
white regions received little

69
Image: https://en.wikipedia.org/wiki/Photostimulated_luminescence#/media/File:Computed_Radiography_Process.svg
Localization
• So far, we have only considered the signal, without much thought into locating that
signal
• The resolution of any imaging system requires knowledge of where the source signal
came from
• Each system has some underlying assumptions that must be met
• Many medical imaging techniques assume that the photons of interest travel in
straight lines

70
Localization
• This assumption is often broken if there is scatter!
• The detected photon is then attributed to have come from the wrong location
• This is often dealt with using collimators

71
 Localization
• Collimators have channels or columns that are meant to block, or attenuate
scattered photons

Collimator No Collimator

https://www.semanticscholar.org/paper/The-effect-of-collimation-on-image-quality-and-in-Koernig
72
Collimation
• It is also beneficial to narrow the window to prevent x-rays from interacting with
tissue not of interest
• This reduces the chances of detecting more scattered photons
• In the next slide, we can see by decreasing the collimator window, we and improve
the details observed in the spine
• Usually nothing comes for free – narrowing the field of view could decrease your
ability to make a proper diagnosis (the problem may be somewhere else!)

73
Collimation
• It is also beneficial to narrow the window to
prevent x-rays from interacting with tissue not of
interest
• This reduces the chances of detecting more
scattered photons
• We can see by decreasing the collimator window,
we and improve the details observed in the spin

74
iClicker Question
Collimation best describes:
A) That x-rays travel in straight lines
B) Image pixels can be thought of as columns that absorb x-ray energy
C) The removal of scattered photons that are not travelling perpendicular to the film
D) Destructive interference that is observed by looking at x-rays as moving wavefronts

75
X-Ray Imaging Expanded
Images can take three forms in medical/biological imaging:

1) Projection images – x-rays, transmission microscopy, neutrons or scene

images (photography)
2) 2D slice images – computed tomography (CT), single photon emission
tomography (SPECT), positron emission tomography (PET), ultrasound,
magnetic resonance imaging (MRI), confocal microscopy
3) 3D volumetric images – magnetic resonance imaging (MRI), deconvolution
microscopy, optical projection tomography, CT (cone beam)

76
 Computed Tomography (CT)
Translating tube and detector
collects projections.

Rotating tube and detector

acquires projections at different
angles.

77
X-Ray Computed Tomography (CT)
• The basic measurement of a CT scanner is a line
integral of the linear attenuation coefficient of tissue
• For monoenergetic x-rays through a linear distribution
of attenuation coefficients, the following holds:
d

= I 0e 0
−  ( s ) ds
I det
• where Idet is the intensity of the detected x-ray beam,
I0 is the intensity of the incident beam, μ(s) is the
distribution of attenuation coefficients along the path
of the beam

78
 X-ray Computed Tomography (CT)
• Projections are created using x-rays,
• Our imaging signal represents the x-ray attenuation coefficient μ(x,y) of the tissues
• These are given in Hounsfield units (H):

 −  water
H= 1000
 water

• These units are used rather than the actual attenuation coefficients to normalize
images made with different CT scanners

79
Substance HU
Air −1000
Hounsfield Units
Lung −500
Fat −100 to −50
Water 0
CSF 15
Kidney 30
Blood +30 to +45
Muscle +10 to +40
Grey matter +37 to +45
White matter +20 to +30
Liver +40 to +60
J-O Blast +20 to +30
 −  water
H= 1000
Soft Tissue, Contrast +100 to +300
 water
+700 (cancellous bone) to +3000 (dense
Bone
bone) 80
 CT Projection Lines y g(x)
f(x,y)

x
g ( x) =  f ( x, y )dy
y

2D image -> 1D profile

g(x)

81
CT Projection Lines
y
y’ x’
f(x,y)
θ
x

g (x ' )
82
CT Reconstruction
• Projection images are directly acquired using detector arrays or by raster scanning
with a single detector (i.e. x-rays)
• 2D slice and 3D volumetric images must be reconstructed (e.g. CT)
• In practice, CT images are reconstructed using back projection

83
Simple Back Projection

84
Simple Back Projection

85
Simple Back Projection

Integrate or
summate

86
Simple Back Projection

87
Simple Back Projection

• Let’s be honest, results

aren’t great…

88
Filtered Back Projection

• If we were to qualitatively assess the previous

result, we might be inclined to say that the
edges are poor
• We can apply a sharpening filter to the signal
before back projected to enhance edge
information
• This involves convolving the 1D projection data
with a sinc-shaped filter
• This introduces negative side-lobes into the
signal – which works to ‘erase’ unwanted signal

89
Filtered Back Projection

One Angle Two Angles

90
Filtered Back Projection
b ( x, y ) g (l ,  ) * r (l )

g (l )  r (l )

f ( x, y )
91
If we didn’t filter before back projection

g (l )

Ib
92
Simple vs Filtered Back Projection

93
Sinogram

• You may often see the projection data stored in a sinogram

• The x-axis is the x-position on detector, and y-axis is the projection angle

Object Sinogram Back projection

94
CT Geometries
Figure 8

• The so-called third generation CT geometry eliminates the need to translate the
tube/detector and leads to faster scanning
95
 Collimation Reduces Scatter
• Detection of scattered rays would reduce
resolution
• Concave collimator is used for this new
geometry
• The fan beam projection views are acquired
• These re-binned into a parallel ray view and
reconstructed using filtered back-projection

96
 CT Geometries
• Fourth generation scanners eliminate the need
Figure 10

to rotate the detectors

• The fan beam can also extend in a 3D cone
• A 2D projection is collected – this is called
“conebeam CT” and uses its own
reconstruction algorithms
• This system has an increased cost due to the
significant increase in the number of detectors

97
Multi-slice CT
• To generate multiple slices at once, “rings” of detectors in the axial direction can be
used
• More typically, data is acquired in a helical pattern
• Data can be interpolated from neighbouring projections to generate discrete slices

98
Iterative Reconstruction
• It is becoming increasingly common to perform iterative reconstructions instead of
filtered back projection
• In iterative reconstructions, μ(x,y) is estimated using numerical methods where
image property constraints and noise are taken into consideration
• Algebraic reconstruction can take additive or multiplicative formulas

99
Numerical Reconstruction Example

A. N’s are the projection data, U’s

are the pixels
B. Shows the true values
C. Given only the bottom projection,
we split the value across pixels
D. We modify these values with the
added information of the side
projection
E. Further iteration for the diagonal
F. One final update with the other
diagonal
100
Additive

5.5 4.5
5.5 4.5

6.5 5.5
4.5 3.5

5 7
6 2
101
iClicker Question
Filtered back projection …
A) Aims to blur the image
B) Aims to sharped back projected images by including negative sidelobes
C) Aims to reconstruct 1D projections into a 3D volume
D) Has improved to back project image intensities in a depth-dependent way

102
 X-Ray Dose
• We will come back to discussing
radiation dose
• In x-ray imaging, there is always a trade-
off of increasing risk of cancer from the
radiation
• This is balanced against the information
that should help your immediate
medical condition
• Lead aprons/shielding is often used to
minimize your body’s exposure to
reduce the dose

https://www.arpansa.gov.au/understanding-radiation/what-is-radiation/ionising-radiation/x-ray 103
X-Rays Beyond Medical Imaging
• I will briefly highlight how X-rays are being used in research outside of medical
imaging
• μXRF imaging – micro x-ray fluorescence imaging
• XAS -> x-ray absorption spectroscopy
• This material is just for fun!

104
X-Rays Production Beyond Medical Imaging
• In these cases, x-rays are typically produced
by a synchrotron
• Synchrotron radiation is generated when
particles are moving at really high velocities
and are deflected along a curved trajectory
by a magnetic field.
• The charged particles are first accelerated by
a linear accelerator (LINAC)
• Particles are accelerated in a booster ring to
nearly the speed of light

105
X-Rays Production Beyond Medical Imaging
• As they round the storage ring, they are
accelerated inwards, with this acceleration
causing the release of a photon
• X-rays with a broad spectrum of energies are
generated and emitted tangential to the
storage ring.
• Beamlines are placed tangential to the
storage ring to use the intense X-ray beams at
a wavelength that can be selected varying the
set up of the beamlines

106
 XAS
• used method to investigate atomic local
structure as well as electronic states
• An X-ray strikes an atom and excites a
core electron that can either be
promoted to an unoccupied level or
ejected from the atom.
• Both processes will create a core hole,
which is filled by an upper-level election
which will release a characteristic x-ray
• XAS therefore requires high-energy X-ray
excitation, which occurs at synchrotron
facilities

107
https://chem.libretexts.org/Bookshelves/Physical_and_Theoretical_Chemistry_Textbook_Maps/Supplemental_Modules_(Physical_and_Theoretical_Chemistry)/Spectroscopy/X-ray_Spectroscopy/XAS_-_Theory
 XAS
• An x-ray absorption spectrum is generally divided
into 4 sections:
1. Pre-edge (E < E0 )
2. X-ray absorption near edge structure (XANES),
where the energy of the incident x-ray beam is E =
E0 ± 10 eV
• XANES has a high probability of occurring,
which gives it the peak

108
https://www.ucalgary.ca/live-uc-ucalgary-site/sites/default/files/teams/269/XAS.pdf
 XAS
3. Near edge x-ray absorption fine structure
(NEXAFS), in the region between 10 eV up to 50 eV
above the edge
• NEXAFS, the ejected photoelectrons have low
kinetic energy (E-E0 is small) and experience
strong multiple scattering by the first and even
higher coordinating shells of neighbouring
atoms
4. Extended x-ray absorption fine structure (EXAFS),
which starts approximately from 50 eV and
continues up to 1000 eV above the edge
• photoelectrons have high kinetic energy (E-E0 is
large), and single scattering by the nearest
neighbouring atoms normally dominates
109
https://www.ucalgary.ca/live-uc-ucalgary-site/sites/default/files/teams/269/XAS.pdf
 XAS
• You can think of reduced scattering at higher energies, because the photoelectron
has enough energy to easily escape the atom

110
https://www.ucalgary.ca/live-uc-ucalgary-site/sites/default/files/teams/269/XAS.pdf
 XAS
• More simplistically – we can irradiate a sample with
x-rays and capture the number of photon counts
over different energy levels
• We can then match characteristic peaks to a
database of peaks (curve fitting) to determine which
atoms are present

111
https://chem.libretexts.org/Bookshelves/Physical_and_Theoretical_Chemistry_Textbook_Maps/Supplemental_Modules_(Physical_and_Theoretical_Chemistry)/Spectroscopy/X-ray_Spectroscopy/XAS_-_Theory
μXRF imaging
• Combines the goals of microscopy and XAS
• A technique to analyze a micro-area of a sample by using X-ray beams and to get
elemental distribution images on a sample by spectrum acquisition at individual
pixel position
• You can think of this as acquiring an absorption spectrum at every pixel in the image
• You can then fit each pixel knowing the characteristic peaks, to determine the
elemental composition of each pixel
• These imaging systems can permit resolutions at a fraction of a millimeter, and make
images of tissue sections of several centimeters squared

112
μXRF imaging
• Pixel-wise, or region averaged fitting
• Each box contains the atomic element (e.g. Cl – Chlorine), and the characteristic x-
ray label

113
https://www.sciencedirect.com/science/article/pii/S0969806X20301183

μXRF imaging
• Results of breast adenocarcinoma analysis by H&E
staining and by synchrotron micro-XRF.
• In the left column the optical microscope view of the 2
μm slices stained with H&E
• In the center column the phosphorous quantification by
synchrotron micro-XRF of the adjacent 30 μm slices are
shown.
CT = connective tissue
AT = active tumor
N = necrotic

114
iClicker Question
Which material is commonly used as the target in medical X-ray tubes?
A) Aluminum
B) Copper
C) Tungsten
D) Lead

115
iClicker Question
Which of the following interactions is most likely to occur at high photon energies?
A) Photoelectric effect
B) Compton scattering
C) Coherent scattering
D) Pair production

116
iClicker Question
What is the primary by-product of a Compton scatter event?
A) Free electron
B) Scattered photon
C) Characteristic photon
D) Both A and B

117
iClicker Question
During the x-ray production process, what percentage of energy is converted into x-ray
photons?
A) 99%
B) 1%
C) 25%
D) 75%

118
iClicker Question
Which of the following interactions contributes the highest proportion of scatter to the
radiographic image?
A) Compton scatter
B) Photoelectric absorption
C) Coherent scatter
D) Pair production

119
iClicker Question
Which of the following is the least common x-ray interaction?
A) Photoelectric absorption
B) Compton scatter
C) Coherent scatter
D) Pair production

120