Particle physics with the computer

Authors:
Eliane Epple [email protected],
Kirill Lapidus [email protected]
Tel: 089/35831-7136
25.05.2011

Excellence Cluster Universe

Chair E12 for Experimental Physics, Physics Department, TU Munich

Contents
1 Introduction

2 Basics of the analyzed experiment

3 Particle identification
3.1 dE/dx vs Momentum . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2 Beta vs Momentum . . . . . . . . . . . . . . . . . . . . . . . . . . . .

7
7
8

4 Analyzing Events
4.1 Invariant mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2 Missing mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

9
9
10

5 Kinematics of elastic collisions

12
5.1 Selecting elastic collision events in an experiment . . . . . . . . . . . 15
6 Reaction cross section

16

7 Shell commands

17

8 Basic C-operators

19

9 ROOT: a class packet for data analysis

21
9.1 Working with 4-vectors . . . . . . . . . . . . . . . . . . . . . . . . . . 21
9.2 Working with histograms . . . . . . . . . . . . . . . . . . . . . . . . . 22
10 Data for the analysis

24

11 Tasks
11.1 Creation of 4-vectors . . . . . . . . . . . . . . . . . . . . . . . .
11.2 Analysis of proton-proton elastic collisions . . . . . . . . . . . .
11.3 Analysis of a p + p p + p + 0 reaction . . . . . . . . . . . . .
11.4 Estimation of a cross-section of the p + p p + p + 0 reaction

25
25
25
26
26

A p + p elastic cross sections

.
.
.
.

.
.
.
.

.
.
.
.

28

Introduction

As particle physics was evolving in the middle of the last century and more and more
experimental data were collected, data analysis with the help of computers became
important. Nowadays it is impossible to imagine a high-energy experiment, where
each recorded event is analyzed separately by hand, like it was done in the early
times of particle physics. There are several reasons for this. First of all, physicists
are usually looking for rare processes employing beams of high intensity and high
rates of data acquisition (up to 50 million events per second in modern experiments).
Combined with typical beam times of several weeks or even months one can easily end
up with a recorded event statistics of several billion events. Moreover, data delivered
by contemporary detectors are complex and have to be processed in several steps.

Figure 1: Left row: The Big European bubble chamber (BEBC) at CERN with a
height of 4m. It was used with a filling of liquid hydrogen and operated in 1973 1984. Lower Figure: A typical bubble chamber picture showing a few particle tracks.
Right row: The ALICE detector at CERN (2010) with a height of 16 m and a length
of 26 m. Lower Figure: A simulated event in the ALICE detector of a lead lead
collision.
With specialized software, the recorded digital signals are translated into interaction
3

points of particles with the detector (so-called hits). All interaction points in one
recorded event are combined and complex computing algorithms reconstruct out of
some thousand points real particle tracks giving the physicist high-level information like momentum, time-of-flight, energy, etc. of the detected particle. From this
point the work of data analysis starts:
to obtain interesting results
from a huge amount of collected information
with help of elaborated algorithms
implemented in computer programs.
With this practicum we want to give an insight into the field of Data Analysis in
particle physics experiments.

Basics of the analyzed experiment

The data which you will analyze within this lab were collected with the High
Acceptance Di-Electron Spectrometer (HADES) see fig 2. It is a fixed target

Figure 2: On the left side an exploded view of the HADES spectrometer is shown.
On the right side you see a side view of the spectrometer with its many sub-detectors.
experiment located at the GSI research center in Darmstadt and currently performs
measurements at the SIS18 accelerator with typical beam energies of a few GeV per
nucleon. As follows from its name, the detector has a high acceptance, covering
85% of the azimuthal angle and polar angles from 15 to 85 degrees. This is a large
spectrometer consisting out of many sub-detectors. Here is a short overview:
The target. For the analyzed experiment a liquid hydrogen (LH2 ) target was
used. It had a length of 5 cm and was cooled down to a temperature of 20K.
The Ring Imaging Cherenkov detector (RICH) for lepton identification. This
detector is filled with a gas in which only leptons create Cherenkov light when
traversing this detector. In the energy regime where the spectrometer records
events this detector is completely blind for hadrons.
The 4 planes of multi wire drift chambers (MDCs). These drift chambers
are employed for particle tracking and momentum reconstruction. The chambers are filled with a gas mixture and contain six layers of wire planes where
avalanche signals caused by particles traversing the chambers can be registered.
5

 The Magnet. Between the MDC planes I, II and III, IV the superconducting
magnet is located. It provides a toroidal magnetic field and is used for the
momentum determination of the particles. By traversing the magnetic field
each particle is bended according to its momentum. The change in flight direction can then be measured by the MDC chambers and the momentum of the
particle can be reconstructed.
The time-of-flight Detectors. These scintillating detectors are located at the
very end of the spectrometer. Their quality is a very good time resolution.
They are employed to measure the time which particles need to fly from the
target, through the spectrometer, to the TOF detectors. Together with the
measured trajectory length, the particle velocity can be reconstructed.
In summary the purpose of the Spectrometer is to reconstruct the four-momentum
of a particle with the help of sub-detectors.
Let us remind how a four-vector looks like:

E
px

p=
py
pz

(1)

The three-momentum components are directly measured by the spectrometer.

To
p
determine the energy of a particle, which can be also written like E = p2 + m2 ,
one needs to know additionally its mass. The mass of a particle is uniquely defined
by its type: a proton, a pion, a kaon etc. Therefore, to obtain the full 4-vector one
has to perform a so called particle identification.

3
3.1

Particle identification
dE/dx vs Momentum

If particles traverse material they loose energy. The energy loss can be hereby described analytically by the Bethe-Bloch formula.
4nz 2
dE
=

dx
me c2 2

e2
40


2  

2me c2 2
2

ln
I (1 2 )

(2)

with = v/c, v the velocity of the particle, E the energy of the particle, x the
distance traveled by the particle, c speed of light, z e the particle charge, e the
charge of the electron, me the rest mass of the electron, n the electron density of the
traversed material and I the mean excitation potential of the material.
As the energy loss is independent of the mass of the traversing particle one obtains
different energy loss curves for different particle species, when plotting the dE/dx vs
the measured momentum of the particle.
In case of the HADES spectrometer the energy loss of particles can be measured in
the MDC chambers and the TOF detectors. In fig. 3 you see the energy loss of the
particles in the MDC chambers plotted against their momentum.

20

MDC dE/dx [arb. units]

18

104

16
14

103

12
10
102

8
6

10

4
2
0

-2000

-1000

1000

2000

3000

polarity * momentum [MeV/c]

Figure 3: The energy loss (dE/dx) plotted against the particles momentum times
their charge.
7

Your Task: Can you name the different particles you see on fig. 3?

3.2

Beta vs Momentum

Another way to identify particles is to measure not the energy loss but the time
of flight of the particles. With help of the time of flight and the path length of
the particle one can calculate the particles velocity (). We can write the particles
momentum as a function of the velocity as follows:
mv
p= p
,
1 2

(3)

m
,
p/c = p
1 2

(4)

or

Then one can easily derive:

p
m=
c

1
1
2

(5)

The last formula shows that in order to identify the mass of a particle, it is enough
to measure simultaneously its momentum and velocity.

Analyzing Events

In April 2007 the HADES experiment collected 1.2109 events of proton-proton collisions with the proton-beam having a kinetic energy of 3.5 GeV. For the lab a subset
of the collected statistics will be provided.
As a result of the collision new particles are created out of the energy the beam
proton imposes on the target. For example:
p+pp+p
p + p p + p + 0
p + p 0 + + + +
p + p p + K+ +
p + p p + n + e+
p+pp+p+
p + p + + K 0 + p
p + p + K + + p + +

(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)

Your Task: Which of these reactions are not possible and why? Think of conservation laws which have to be fulfilled.
Some of the particles appearing in reactions 6-13 are short-living particles which
have lifetimes in the order of ns and decay lengths of few fm1 . It is impossible to
measure such particles directly. However, in order to reconstruct these particles,
one can employ two methods which are based on the laws of energy and momentum
conservation.

4.1

Invariant mass

Two variables, commonly used in high-energy physics are invariant and missing
masses. The invariant mass of two particles is defined as the sum of the particles
4-vectors squared:
MInv = (P1 + P2 )2 .
The invariant mass can be also expressed as:
MInv = m21 + m22 + 2 (E1 E2 |p1 | |p2 | cos ) .
1

fm stands for femtometer (or, an obsolete name, fermi). It is a common unit of length in
particle physics. 1 fm = 1015 m

Since the invariant mass is a square of the 4-vector, this variableas its name
suggestsis invariant under Lorenz transformations. Short-living particles can be
reconstructed with this simple method.
Note, that the invariant mass can be constructed for any amount of particles:
n
MInv
= (P1 + P2 + . . . + Pn )2 .

This definition is used for example when studying 3- and 4-body decays.
Example:
The decay of p . The has a lifetime of 0.26 ns. It is thus able to fly only
8 cm before decaying. In our Spectrometer we will therefore see no direct evidence
for the particle and it can be only reconstructed by its decay products.
The decay of + 0 . In this case one even needs to detect three particles and
build the invariant mass of their Lorentz Vectors to obtain an signal.

4.2

Missing mass

In case of elementary reactions (p + p, + p, ...) it is also possible to reconstruct

the event topology by means of the missing mass method. For this method one uses
the Lorentz vectors of the incoming beam and the target. If one reconstructs in an
event all but one particles in the final state one can identify the missing particle with
the following method:
Mmissed = (PBeam + PT arget P1 P2 . . . Pn )2 .
Thus a missing mass technique can be employed for the reconstruction of particles
that were not detected.
Example:
Assuming we have measured the reaction
+ p + KS0 .
and the experiment allows to reconstruct the K 0 via its decay into a + -pair. We
can try to reconstruct the also by its decay products or apply the missing mass
technique. A 4-momentum conservation for this reaction reads as follows
P + Pp = P + PKS0 .
10

In this case, the missing mass of the reconstructed K 0 will be:

0

K
= (P + Pp P P+ )2
Mmiss

=
(P )2 = M .

Your Task: Why in case of ion collisions, it is complicated, or even impossible to

apply the missing mass technique?

11

Kinematics of elastic collisions

An elastic collision is a special type of reaction in which no new particles are produced
and the colliding particles do not change their type. Examples are collisions of two
billiard balls or proton-proton elastic scattering: p + p p + p. In case the physicist
wants to identify elastic events he or she can exploit the special kinematics of the
reaction.
Let us consider a collision of a beam particle PB with a target particle PT and two
particles in the final state.
PB + PT P1 + P2

(14)

Since only two particles in the final state are produced, there are 8 kinematic degrees
of freedom represented by the components of the two four-vectors. Two components
are fixed by the masses of the particles, three by momentum conservation and one by
energy conservation. Therefore, only two degrees of freedom remain. This means in
the case of an elastic collision one only needs to measure two observables and deduce
all components of the four-vector.
Example:
We measure p~1,x and p~1,y . Using momentum conservation we can write:
0 = p~1,x + p~2,x
0 = p~1,y + p~2,y

(15)
(16)

p~B,z = p~1,z + p~2,z

(17)

We also know that

And (from energy conservation)

q
q
2
2
2
2
EB,Kin + 2 mp = p~1,x + p~1,y + p~1,z + mp + p~22,x + p~22,y + p~22,z + m2p

(18)

As in an experiment EB,Kin and p~B,z are normally known, it is simple to derive the
whole kinematics from measuring p~1,x and p~1,y .
But we can also derive the elastic kinematics by measuring the deflection angles
of the outgoing particles. A very natural frame for studying elastic events is the
Center of Mass frame (CM, noted with *) in which target and beam vectors sum
up to zero. (The system in which the target is at rest is called Target System,
12

Figure 4: Angles of an elastic collision

in the CM-System

Figure 5: Angles of an elastic collision

in the Target-System

TS or Laboratory System). Moreover one would consider here instead of Cartesian

coordinates spherical coordinates.
q
r = p2x + p2y + p2z
(19)
 
px
= arctan
(20)
py
 
pz
= arccos
(21)
|~p|
By energy and momentum conservation in the CM system r is fixed. The angles
and are the two free parameters. The momentum conservation in this case reads
as follows:
1 = 2
1 = 2

(22)
(23)

The big advantage of choosing this coordinate system is that one does not need
to measure the particles momentum. However, in order to use these relations in
practice, we would have to transform the measured four-vectors from the TS into
the CM system. For this we would need to know the momentum components. Instead
we transform the relations (22) and (23) to the TS frame. As the beam is (by the
definition of the coordinate system) collinear with the z-axis one only has to perform
a Lorentz boost in z-direction for a transformation from TS to CMS. This means
that the transverse component (~p ) of the vector does not change under this Lorentz
transformation:
p~ =|p1 | sin 1 = |p1 | sin 1
13

(24)

The only reason why |p | and obtain new values in the TS frame is the fact, that
the momentum component parallel to the beam direction (~p|| ) changes. This can be
expressed as follows:
p~|| =|p1 | cos 1 = pp |p1 | cos 1 pp pp E1

(25)

The transformation is determined by pp and pp which are the velocity and the
Lorentz factor of the initial p+p system. To obtain a transformation of the angle
from the CM into the TS frame without a dependency on p1 we divide (24) by (25)
and obtain:
tan 1 =

sin 1
E

pp (cos 1 pp |p1| )

(26)

tan 1 =

sin 1
pp (cos 1

(27)

pp
)
1

If we use relation (22) that connects 1 and 2 we obtain:

tan 2 =
tan 2 =
The value for

pp
1

sin( 1 )
pp (cos( 1 )

pp
)
1

sin 1
pp ( cos 1

pp
)
1

(28)
(29)

can be derived using the following relations:

m2pp + m21 m22
2mpp
2
[(mpp (m1 m2 )2 )(m2pp (m1 + m2 )2 )]1/2

|~p1 | =
2mpp
E1 =

(30)
(31)

and

m2pp m2B + m2T

2mT
2
[(mT (mpp mB )2 )(m2T (mpp + mB )2 )]1/2
|~ppp | =
2mT
Epp =

14

(32)
(33)

m1 and m2 are the masses of the two outgoing particles. mB and |~pB | are the mass
and momentum of the beam particle. mT is the mass of the target particle. mpp and
Epp are here the mass and the energy of the combined vector of the target and beam
proton. As the relativistic velocity is defined as Ep we obtain pp =1.
1
Your Task: Derive equations (30) and (32). Can you connect tan 1 and tan 2
without any dependence on angles? What is the value for for a p+p reaction with
a beam kinetic energy of 3.5 GeV?

5.1

Selecting elastic collision events in an experiment

We have learned so far that two measured observables in a reaction are enough to
determine the full kinematics of elastic events. So applying cuts on tan 1 , tan 2 and
1 , 2 would be sufficient to determine completely the four-vectors of both protons
for elastic events. When the event was however, a three body production instead of
a two body reaction we have 3 more degrees of freedom.
Note that in this case we need to measure 5 observables to explicitly determine all
components of the four-vectors. In case we explicitly measure 3 particles with our
detector there is no doubt that it was not an elastic scattering reaction but if we have
one or more missing particles we could see if for example the px and py components
of our two protons add up to zero. There are however cases where this still could
add up to roughly zero but is not a sign of a two body reaction.
This means that asking for two constraints in our data analysis is a necessary constraint but not a sufficient one. An easy way to check if there was another particle
produced is to investigate the missing mass of the two protons. We expect to see the
mass of the third particle with help of this observable.

15

Reaction cross section

A value used in particle physics to characterize the probability of the particles interaction is a cross section. The reaction cross section is defined as
=

Rreac
,

where Rreac is a reaction rate for one target particle (number of reactions per unit
time), is a particle flux (amount of bombarding particles per unit area and per
unit time, cm12 s ).
The unit of a cross section is defined as barn:
1 b = 1028 m2 .
For particle reactions one normally uses milli- and microbarns, 103 b and 106 b,
respectively. Fig. 6 shows the energy dependence of the proton-proton cross section.
Data on elastic cross section can be found in Appendix A.

Figure 6: Total and elastic cross section of proton-proton collisions as a function of

the proton momentum in the laboratory reference frame [PDG]

16

Shell commands

In the course of the practicum, a minimal knowledge of Linux console (shell) commands is required.

Figure 7: Left: The symbol on your desktop to open the Terminal. Right: an open
terminal where you can type your commands.
Here is a list of the most important commands. To execute a command, write it in
the console and press enter.
Show the current directory. This is so to say the folder in which you work at the
current moment:
pwd
Change the current directory to directory named dirname (with this command
change between folders):
cd dirname
Show a list of the files in the current directory:
ls
To open C files you can use the text editor fte or any other editor you know and like.
You have to type in this case:
fte program_name.cc
To compile the program (make an executable file out of the program code):
make clean build install
17

Run the (executable) program with name program_name.

./program_name
A very comfortable command is the tab option. If you start to write a directory
path or a macro name you can type the first letters and press the tab button on your
keyboard. The complete name will then be auto-filled if this is the only possibility if
however several possibilities exist how the name could be completed this option will
print you all possibilities in your shell.

18

Basic C-operators

Comments
In C and C++ comments are marked with //. Any command written after this sign
will be ignored.
Variable creation
To create a variable one need to specify its type (integer, float or other) and its name:
Int_t var1;
Float_t var2;

//integer type: 5, 100, -10

//float type: 3.4, 3.14156

Note, that in C and C++ lines of code (different commands) are always separated
with a semicolon. A forgotten semicolon is a common mistake while programming.
Conditional statement
in C, a conditional statement has a following form:
if (condition) {
perform actions
}
for example, we fill a histogram, only if the condition that
is fulfilled

p 2
(px + p2y ) is greater 25.5

if ( sqrt(px*px+py*py) > 25.5) {

h1->Fill(var1);
}
To build a logical condition there is a set of logical operators. Most of them have
intuitive meaning:
C-operator
a>b
a<b
a==b
a!=b
a&&b
a||b

Meaning
a is greater than b
a is less than b
a is equal to b
a is not equal to b
logical AND
logical OR

Table 1: Logical operators in C

19

Arithmetic operators
C-operator
+
*
/
sqrt(a)
pow(a,b)
fabs(a)

Meaning
add
subtract
multiply
divide
square root of a
ab
|a|

Table 2: Arithmetical operators in C

for-loop
C supports a number of loop types. For our purposes the so called for-loop is
enough. In such a loop a variable (in this case i) is defined. Each loop iteration
the variable is increased by a value of 1 (i++), until it reaches its maximum allowed
valued (i<100).
for (Int_t i = 0; i < 100; i++) {
h1->Fill(i);
//filling the histogram
... //any commands on this place
... //any commands on this place
}
In this case, any operators on the place of ... will be repeated exactly 100 times
(integer variable i will take values from 0 to 99).
Also very useful is the cout command.
cout<<"Hello world"<<endl;
will print you the word Hello world on your screen when executed. But you can
also display variables in the code.
cout<<"This is the value of beta : "<<beta1<<endl;
will print you the actual value of beta1 on your screen. For example
This is the value of beta : 0.72

20

ROOT: a class packet for data analysis

ROOT is a specialized software package that was developed at CERN to serve the
needs of the particle physics community, especially for the data processing of the
experimental set-ups located at the LHC.
It is a C++ library that contains a number of useful classes. Here a basic knowledge
about working with ROOT will be given.

9.1

Working with 4-vectors

A dedicated class TLorentzVector to work with 4-vectors.

Create a 4-vector object with the name particle_a
TLorentzVector particle_a;
Set the components of a 4-vector:
particle_a.SetPxPyPzE(px, py, pz, E);
One can for example access easily polar and azimuthal angles , (in radians) and
rapidity y of the 4-vector:
Float_t theta = particle_a.Theta();
Float_t phi = particle_a.Phi();
Float_t y = particle_a.Rapidity();
A TLorentzVector class provides a convenient way to construct the sum of two and
more 4-vectors and calculate invariant and missing masses:
//construct a 4-vector named pairAB
//out of two 4-vectors particleA and particleB
TLorentzVector pairAB;
pairAB = particleA + particleB;
//calculate invariant mass of the 4-vector pairAB
Float_t M_inv = pairAB.M();
//construct a missing 4-vector, given initial (beam+target) 4-vector
TLorentzVector missed;

21

missed = initial - pairAB;

//access the value of the missed mass
Float_t M_missed = missed.M();

9.2

Working with histograms

A histogram is a convenient and common way to represent the statistical behavior

of one or two variables. ROOT contains a set of classes to work with histograms.
For our purposes two of them are enough: TH1F and TH2F.
To create a histogram named h1 with 500 bins in the region of an X-value from 0
to 5:
TH1F *h1 = new TH1F("h1","Histogram title",500,0,5);
To add an entry to the histogram, just use the Fill() method. The entry in this
case should be either a number:
h1->Fill(2.3);
or a variable (named, in this example, theta_1), containing a number:
h1->Fill(theta_1);
After the histogram is filled with values, one can draw it:
h1->Draw();
Lets now turn to 2-dimensional histograms. They are in particular helpful, when
one would like to analyze a correlation between two variables, for example angle
and momentum p of a particle.
To work with them is very similar to the 1-dimensional case, but now there are two
axes:
To create a two-dimensional histogram named h2 with 500 bins in the region of
X-values from 0 to 5 and 100 bins in the region of Y-values from -10 to 10 :
TH2F *h2 = new TH2F("h2","Histogram title",500,0,5,100,-10,10);
To add an entry to the two-dimensional histogram, one need to provide two values
(x,y)
22

h1->Fill(4.5, -7);
or
h1->Fill(theta_1, theta_2);
Drawing of a histogram is completely analogous to the 1-dimensional case:
h1->Draw();
Much more information about histogram usage can be found in Chapter 3 of the
ROOT Users Guide.

23

10

Data for the analysis

The data for the analysis is prepared in form of an ntuple. An ntuple can be understood as a simple table that contains different numerical values. Each row in such a
table corresponds to one event, and each column to one variable (see Table 3).
Event number
0
1
...

Variable 1
100.1
437.7
...

Variable 2
-200.5
289.3
...

...
...
...
...

Table 3: General logical structure of an ntuple

The ntuple prepared for the analysis was preselected to contain only events where
two particles were detected in the spectrometer. It contains several observables. In
each event the ntuple is filled with the extracted values:
Variable name
charge1
beta1
px1
py1
pz1
dedx1

Meaning
sign of the particle charge (+1 or 1)
= v/c, where c is the speed of light
px , x-component of the particle momentum
py , y-component of the particle momentum
pz , z-component of the particle momentum
particle energy loss in the detector (MDC chambers)

Table 4: Variables in the ntuple (for one particle) and their meaning
Since the prepared data sample always contains two reconstructed particles per event,
the ntuple contains exactly the same set of variables for particle 2 (charge2, beta2,
etc.).
Usually, a routine step in data reduction is to perform an iteration over the data
(in our case to go through each entry in the ntuple), select an interesting data
sub-sample, perform calculations when necessary, and obtain a result typically a
number, or series of numbers.

24

11

Tasks

For the completion of the practicum tasks use the provided template program. It
already contains an iteration over all events. And some skeleton of basic code.

11.1

Creation of 4-vectors

The first task is to construct out of the observables contained in the ntuple two
4-vectors (with help of the TLorentzVector class) for particle 1 and particle 2.

11.2

Analysis of proton-proton elastic collisions

After the construction of two 4-vectors you have all the needed information to select
events with proton-proton elastic collisions. The aim of this task is to develop a
strategy to select elastic events and estimate their amount.
1. What kind of observables can you use to select pp-elastic events?
2. Make histograms (one or two-dimensional) of the observables you would like
to constrain on.
3. Apply a constraint to one observable and see how the distribution of the other
observables change.
4. Introduce step by step more cuts until you think your events are pure enough
and document your analysis strategy with histograms. A general aim of a data
analysis is to select the targeted reaction with as low background as possible.
However, if the cut conditions applied are too strict you will, obviously, loose
a lot of statistic. Try to find a balance between efficiency (amount of signal)
and purity (amount of background events which remain after the cuts) of your
data analysis.
5. Estimate the total amount of elastic proton-proton collisions after your final
selection. For this select one final observable and plot this observable with
selective constrains on other observables.

25

11.3

Analysis of a p + p p + p + 0 reaction

The reconstruction of two protons in the final state is enough to select the reaction
p + p + 0.
1. Construct an observable to select events with 0 .
2. Plot the chosen observable and identify events that correspond to the p + p + 0
reaction.
3. Is there a way to purify the data sample of events containing two protons and
a 0?
4. Estimate the number of p + p + 0 events. In this case of the event selection
the signal will be small as compared to the background underneath the signal.
Take a prepared function to fit the signal together with the background and
extract with help of the fit function Np+p+0 .

11.4

Estimation of a cross-section of the p + p p + p + 0

reaction

As explained in section 6 of this description, a cross section of the reaction characterizes its probability. Let us label the number of observed elastic proton-proton
collisions as Npp and number of p + p + 0 reactions as Npp0 . Then
pp
Npp
=
,
Npp0
pp0
where pp and pp0 are the cross sections of pp-elastic and p + p + 0 reactions,
respectively.
1. Extract the number of pp elastic events Npp from 11.2 point 5.
2. Use the Appendix A to interpolate the elastic cross section at the beam momentum of the analyzed experiment.
3. Given a known cross-section of pp-elastic events and using the ratio between
the number of pp-elastic events and p + p + 0 production, calculate the cross
section of the p + p + 0 reaction.
4. Compare the obtained cross-section with the known value.
26

Here we made an important assumption, that the acceptance of the events of p + p

and p + p + 0 reactions is the same. In general, the acceptance can depend very
strongly on the reaction. In our case, this assumption is acceptable (why?). Are
there other reasons why the two results are not direct comparable?
After completion of all the task a short report has to be written, summarizing the
analysis strategy and the obtained results.

27

p + p elastic cross sections

Here is the measured data on proton-proton elastic cross section, taken from the
PDG web page. The data is organized in the following columns: 1) measurement
number, 2) beam momentum (GeV/c) in the laboratory frame, 3) cross section (mb),
4) statistical uncertainty of the measurement (mb).
NUMBER

PLAB(GEV/C) SIG(MB)

ERR

01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

2.22883
2.80000
2.80553
3.00000
3.03662
3.65000
3.67024
4.00000
4.15000
4.80000
5.00000
5.25517
5.52000
5.96493
6.00000
6.60000
6.79990
6.80000
6.92000
7.02590
7.07634
7.07634
8.10000
8.50000
8.80000
8.80000
8.90000
9.00000

0.73000
1.0000
0.48000
0.70000
3.0000
0.60000
0.76000
0.30000
2.6000
1.2000
0.50000
2.0000
0.25000
2.1000
0.50000
0.33000
0.22000
0.60000
0.50000
2.0000
1.0000
0.90000
0.40000
0.40000
0.22000
0.30000
0.50000
0.32000

19.860
16.300
19.210
17.200
17.000
15.200
15.320
13.500
11.600
14.400
12.700
10.000
11.990
10.000
11.500
11.470
11.790
10.600
11.400
8.0000
9.7000
9.8000
10.800
8.7400
11.710
9.8000
10.100
10.840

28