Lecture 4

Wednesday - January 18, 2006

Written or last updated: January 18, 2006

P442 – Analytical Mechanics - II

Relativistic Collisions
Alex
c R. Dzierba

Four-vector momentum and conventions

We will discuss the subject of collisions within the context of special relativity. And we start off by reminding
you about some important results from special relativity. Figure 1 shows a starred inertial frame moving
along the z axis of an unstarred inertial frame with velocity v with respect to the unstarred frame. Four-
vectors in the two frames are related by a Lorentz transformation that maintains the magnitude of the
four-vector invariant in the same same way that rotations leave the length of three-vectors invariant. And
to continue the analogy the dot product (scalar product) of two four-vectors is a Lorentz invariant just as
the scalar product of two three-vectors is left unchanged under a rotation of coordinates.

x x*

z z*

y y*

Figure 1: Two inertial frames, one moving with constant velocity v along the z-axis with respect to the
other.

The four-vector of particular interest for us is the energy-momentum four-vector of a particle of mass m
which we write as p = (E, px , py , pz ) = (E, p). I will use p for the momentum four-vector and p for the
so-called three-momentum (px , py , pz ). The dot product of a the momentum four-vector with itself is a
Lorentz scalar and can be written as p · p = E 2 − p2 = m2 . This quantity is the same in any Lorentz frame
and m is rest mass of the particle as measured in a frame in which the particle is at rest. The rest mass
of the photon is zero so the magnitude of its four vector is always zero and E = |p| for the photon. The
product of two four-momenta, which is also a Lorentz invariant, can be written as p1 · p2 = E1 E2 − p1 · p2
p
You may be used to seeing this relationship in textbooks: E = p2 c2 + m2 c4 – but we choose units for
energy, momentum and mass that allow us to leave c out of these expressions. These units are, respectively:
eV, eV/c and eV/c2 . You will recall that 1 eV (electron volt) equals 1.6 × 10−19 J. Units that are more
appropriate for discussing masses of elementary particles and momenta and energies of current and planned
accelerators are MeV, GeV or TeV which stand for mega-electron volts (106 eV), giga-electron volts (109 eV)
and tera-electron volts (1012 eV) respectively.

1
Particle masses: In this note we will be referring to the mass of the proton (mp = 934 MeV/c2 ), neutral
pion (mπ0 = 135 MeV/c2 ), muon (mµ = 105 MeV/c2 ) and electron (me = 0.5 MeV/c2 ). We will also
sometimes approximate the mass of the proton as ≈ 1 GeV/c2 .

When is a particle relativistic? As a rough guide, when the momentum of the particle isp several percent
of a rest mass of the particle, relativistic effects start becoming important. Since γ = E/m = p2 + m2 /m if
p can be ignored compared to m then γ = 1 and when p is a few percent of m we can make the approximation
(using the binomial expansion) γ ≈ 1 + p2 /2m2 . This gives you an idea for how quickly relativistic effects
turn on as the momentum increases. For example, an electron with momentum 10 MeV/c2 is relativistic
since it has γ ≈ 20 while a proton of the same momentum can probably be treated non-relativistically since
it has a γ ≈ 1.005.

Kinetic energy Energy (E), mass (m) and kinetic energy (T ) have the following relationship: E = T +m.
It will be left as an exercise for you to show that with this definition, the non-relativistic approximation for
T is p2 /2m.

Lorentz transformations

The energy-momentum four-vector transforms like so:

    ∗ 
E γ 0 0 γβ E
 px   0 1 0 0   p∗x 
 py  =  0 0 1 0   p∗y
(1)
    

pz γβ 0 0 γ p∗z

p
where β = v/c and γ = 1/ 1 + β 2 .

Carrying out the matrix multiplication:

E = γ(E ∗ + βp∗z ); px = p∗x ; py = p∗y ; pz = γ(p∗z + βE ∗ ) (2)

with the familiar result that components of the four-vector transverse to the Lorentz boost direction are
unchanged.

Suppose that a particle of rest mass m is at rest in the starred frame. Then its momentum is zero and
E ∗ = m. In the unstarred frame the particle has energy E = γm and momentum p = βγm = βE along the
z direction. So we have this important result:

E ~= p
γ= and β (3)
m E

So if you know the momentum and energy of a particle in a given frame, the above provides you with the β
and γ of the particle.

2
Particle lifetime Many elementary particles are unstable with lifetimes that are determined by the un-
derlying force that determines the decay process – the stronger the force, the shorter the lifetime. The
electron and proton are stable particles. The experimental limits on their lifetimes are 1026 yr and 1032 yr
respectively. The lifetime of the muon (a heavier version of the electron) is 2.2 × 10−6 s. The lifetime of
the neutron, which is slightly more massive than the proton, is 885 s. The muon and electron (and other
leptons) carry a quantum number called lepton quantum number that has to be conserved in any process.
Since there is no particle lighter than the electron that carries this quantum number, it cannot decay. The
muon can decay into an electron plus other particles (actually two neutrinos) because it has enough mass
and therefore it is unstable. The proton and neutron (and other baryons) carry a quantum number called
baryon number which similarly has to be conserved. The proton is the lightest baryon, hence it is stable.

The lifetimes above are given as a mean lifetime τ . The probability that a particle will survive a time t after
being produced is e−t/τ .

The lifetime only makes sense in the frame in which the particle is at rest. In a frame in which the particle
is moving, the particle’s lifetime is dilated by the factor γ and the distance the particle travels in the frame
in which it is moving is γβcτ = pcτ /m. For example, at Femilab, muons can be produced with momenta
up 800 GeV/c. For these muons γ ≈ 8000. We can safely say that β ≈ 1 and without the time dilation
factor the muons would travel a mean distance, before decaying, of cτ = 660 m. With the dilation factor
the distance is 5,280 km. Of course, to the muon that distance is shortened by the factor 1/γ – in its frame
it travels 660 m.

Two-body scattering

Consider the scattering process 1 + 2 → 3 + 4 in the overall CMS as shown in the right-hand portion of
Figure 2.

p*3

LAB p3 CMS
∗
p1 p2 θ3 p*1 θ
θ4 β p*2
p4
p*4

Figure 2: Two-body scattering viewed in the LAB and CMS frames.

We are going to allow for the possibility that masses change so that masses after the scattering are not
necessarily the same as before the scattering. In the CMS the sum of the four vectors is given by:

p∗0 = p∗1 + p∗2 = (E1∗ + E2∗ ) − (p∗1 + p∗2 ) = E1∗ + E2∗ (4)

The three-vector momenta sum to zero in the CMS. In the above equation we introduce the total four-
momentum in the CMS – p∗0 which must also be equal to p∗3 + p∗4 . Since the square of the momentum
four-vector is the square of the mass of a particle with that four-momentum, we will call the magnitude of
p∗0 the mass m0 . You can picture the collision process in the CMS as particles 1 and 2 coming together to
momentarily form a particle of mass m0 in the CMS that then decays into particles 3 and 4. This will be a

3
useful picture when we look at a particle in its rest frame decaying into two other particles and then viewing
this decay in a frame where the particle is moving.

Obviously we have:

E1∗ + E2∗ = m0 = E3∗ + E4∗ (5)

In one of the problems below you are asked to show the following:

m20 + m21 − m22 m20 + m22 − m21

E1∗ = and E2∗ = (6)
2m0 2m0

m20 + m23 − m24 m20 + m24 − m23

E3∗ = and E4∗ = (7)
2m0 2m0

p
And it is also obvious that |p∗i | = Ei∗2 − m2i where i goes from 1 to 4 and βi∗ = |p∗i |/Ei∗ .

In the LAB frame the initial total three-momentum is p1 since p2 = 0 and the initial total energy is E1 +m2 .
In the LAB frame the CMS is moving with velocity β and it, and its associated γ are given by:

p1 E 1 + m2
β= and γ = (8)
E 1 + m2 m0

It is useful to remind yourself, that viewed from the CMS frame the LAB frame is moving away with velocity
−β. You should also remind yourself that:

(p1 + p2 )2 = (p∗1 + p∗2 )2 = m20 (9)

since the squares of four-vectors are preserved in going from one inertial frame to another and evaluating in
the LAB frame:

m20 = m21 + m22 + 2m2 E1 (10)

Fixed target versus collider experiments. In high energy experiments, especially those whose goal is
to discover massive new particles, like the Higgs particle, total available energy is important. One way to
do the experiment is to have a beam of particles with a certain energy hit target particles (like protons in
liquid hydrogen) at rest. The other technique is to bring two beams of particles of equal energy into head-on
collisions. This is the difference between fixed-target and collider experiments. From an energy point of view
the collider approach provides more energy for discovery. For example, the new Large Hadron Collider (LHC)
at CERN will bring 7 TeV protons in collision with 7 TeV protons. The p total available energy is 14 TeV. If a
single 7 TeV beam collides with protons at rest the energy available is 2mp (mp + E). Using this formula
and setting E = 7 TeV and being careful to express the proton mass in units of TeV (mp = 0.001 TeV) we
get a total available energy of 0.12 TeV, less than 1% of what we get in the collider.

4
Comparting the CMS and LAB quantities

Referring to Figure 2 and using our transformation equations 2, for particle 3:

p3x p∗3 sin θ∗

= tan θ3 = (11)
p3z γ(p3 cos θ∗ + βE3∗ )
∗

where β is the velocity of the CMS as given by equation 8. Dividing numerator and denominator of the last
term by p∗3 we get:

sin θ∗
tan θ3 = (12)
γ(cos θ∗ + β/β3∗ )

For the non-relativistic case (γ = 1) this is what we obtained in discussing collisions without relativity.
Again, invoking our Lorentz transformation:

p3 cos θ3 γ(p∗3 cos θ∗ + βE3∗ )

= β3 cos θ3 = (13)
E3 γ(E3∗ + βp∗3 cos θ∗ )

After some simplifying:

β3∗ cos θ∗ + β
β3 cos θ3 = (14)
1 + ββ3∗ cos θ∗

We assume that any angle in the CMS is possible, that is, 0 < θ∗ < π. If all the mi are equal, no particle
can go backwards in the LAB. If m1 = m3 and m2 = m4 and m2 > m1 , then in the LAB frame the heavier
particle has a maximum angle of π/2. See one of the problems below.

Decay of the neutral pion

The neutral pion (π 0 ) has a mass of 0.135 GeV/c2 and a mean lifetime of about 10−16 s. Including time
dilation, even a 1 TeV π 0 would travel less than 200 microns before decaying. We can assume, to within
a good approximation, that they travel hardly any distance before they decay. And 99% of the time they
decay into two photons. Figure 3 shows the decay viewed in the rest frame of the π 0 and also viewed in the
frame where the π 0 momentum pπ . We refer to this latter frame as the LAB frame.

Since photons are massless their energies and magnitudes of their three-momenta are equal. In the rest frame
of the π 0 we have β1∗ = β2∗ = 1 and in the LAB frame β1 = β2 = 1. Also in the rest frame, E1∗ = E2∗ = mπ /2.

The π 0 carries no spin – so in its rest frame there is no special direction. If we consider any one of the
photons, it will come off in any direction with equal probability. If we were to plot the number distribution
of photons in some interval of cos θ∗ (dN/d cos θ∗ ) from cos θ∗ = +1 to cos θ∗ = −1 it would be flat. Since
the integral of dN/d cos θ∗ over the full range of cos θ∗ is usually normalized to 1 then dN/d cos θ∗ = 1/2.
Obviously we only have to consider one of the photons because we know the other has to be collinear with
it.

5
 E*1
rest frame E1
γ LAB γ
θ∗ pπ θ1
γ
γ
E*2 E2

Figure 3: The decay π 0 → 2γ viewed in the rest frame of the π 0 and also viewed in the frame where the π 0
momentum pπ .

In the LAB frame we know how to find βπ and γπ . In that frame, the energy of one of the photons is:

mπ
E1 = γπ (E1∗ + βπ E1∗ cos θ∗ ) = γπ (1 + βπ cos θ∗ ) (15)
2

We can see that the minimum and maximum values of the photons in the LAB are:

mπ mπ
E1min = γπ (1 − βπ ) and E1max = γπ (1 + βπ ) (16)
2 2

and we also see that:

E1max − E1min = γπ mπ βπ = pπ (17)

What about the distribution of LAB energy E1 ? From equation 15 we have:

mπ
dE1 = γπ βπ d cos θ∗ (18)
2

Using the result quoted earlier – dN/d cos θ∗ = 1/2, we conclude that

dN 1 1
= = (19)
dE1 βπ γπ mπ pπ

But the range of energies allowed is pπ so the integral of dN/dE1 over that range is one – and that’s a good
thing.

Distribution in opening angle of the photons in the LAB. To find the angle of one of the photons
in the LAB, measured with respect to the line-of-flight of the pion, we can borrow equation 14 realizing that
the β = 1 for photons (in any frame).

cos θ∗ + βπ
cos θ1 = (20)
1 + βπ cos θ∗

6
From the above, if one of the photons in the CMS has cos θ∗ = +1 the other has cos θ∗ = −1 and these two
photons come off with angles 0 and π in the LAB respectively – the opening angle is π.

You can convince yourself that the minimum opening angle occurs when the photons come off at θ∗ = π/2
in the CMS and for that situation the LAB energies of the photons are equal (see equation 15).

Here is another way to look at the distribution in opening angle between the photons in the LAB. Let’s call
that angle ψ. Starting with the conservation four-momentum in the decay pπ = p1 + p2 we take the squares
of these four-vectors:

m2π = p2π = (p1 + p2 )2 = 2p1 · p2 = 2E1 E2 (1 − cos ψ) (21)

As a reminder, for the photons p21 = p22 = m2γ = 0. After a little re-arranging and a little trig:

ψ mπ
sin = √ (22)
2 2 E1 E2

For the minimum opening angle case E1 = E2 = Eπ /2 we find:

ψmin 1
sin = (23)
2 γπ

In one of the problems you are asked to find the minimum opening angle by differentiating equation 22. And
you will also be asked to find the formula for the distribution in the opening angle ψ.