Week II:
Introduction to
Deep Inelastic Scattering
26 September 2006
What is DIS ?
Let's start with elastic lepton scattering
on a spinless target at E 10 - 100 MeV
e(E,k)
e(E',k')
*
-q
2
= Q
2
d c
d D
=
d c
d D
point
F(q)
2
F(q) is the Fourier transform of the
spatial charge distribution
increasing the Q
2
, results e- seeing
the proton excited states and a continuum
At high energies, elastic collisions become very unlikely
- elastic scattering form factor decreases rapidly with q
2
The regime of Deep Inelastic Scattering
Instead: the proton breaks up into hadron fragments
Probe a proton structure to q
2
up to 10000 GeV
2
- (i.e down to 10
-18
m)
e p e X
Cross-section for inelastic e()p scattering
Inclusive cross-section (we observe the final lepton and
sum over all possible hadronic final states):
E
'
Electron energy transfer = E E'
q(,q)
E
p
p
1
p
2
p
3
p
4
p
2
+q=p
4
p
2
q =( M , 0)(+ , q)
q
2
+M
2
+2 p
2
q=W
2
p
2
q =M +
q
2
= -2M
Furthermore, E' and are related:
q
2
= -4EE' sin
2
/2 = -2M(E E')
Invariant mass W
Only have to measure one of them , e.g. ()
This defines your scattering angle, from which you can deduce
information and calculate the cross-section.
if M = W
d c
o
2
Q
4
L
j+
W
j+
(q , p)
starting from our lepton-hadron scattering cross section:
In the regime of DIS, mass of system W is not fixed
q
2
and are not related <=> E' and are not related
- have to measure both E' and
The two functions W
1
and W
2
contain information on the structure of the proton.
Final state contains at least one baryon :
W
2
> M
2
=> q
2
< 2M (q
2
< 0)
Most general form of DIS cross-section is:
which in the lab frame gives:
(
d
2
c
dE ' d D
)
LAB
=
o
2
4 E
2
sin
4
0
2
(
F
2
(+ , Q
2
)
+
cos
2
0
2
+
2 F
1
(+ , Q
2
)
M
sin
2
0
2
)
(
d
2
c
dE ' d D
)
LAB
=
o
2
4 E
2
sin
4
0
2
(
W
2
(+ , Q
2
) cos
2
0
2
+2 W
1
(+ , Q
2
) sin
2
0
2
)
Bjorken scaling
Functions F
1
and F
2
are known as the structure functions
They are not directly related to F.T of charge
distribution (at q
2
= 0) because of energy transfer .
Give info on the number and properties of quark
and gluon constituents in the proton
- quark-parton model
Cannot (yet) be predicted from first principles
(QCD)
- must be measured
In 1969, Bjorken proposed the following scaling properties:
In the limit Q
2
, W
2
, x is fixed; x =
Q
2
2 p. q
At fixed x (scaling variable), the scattering is independent of q
2
This suggests that the probing virtual photon scatters against something pointlike
Cross-section independent of both and Q
2
- a point is a point irrespective of wavelength
whereas for sub-sctructure with length scale 1/q
0
- cross-section would depend on q
2
/q
0
2
\
2n
q
The Quark parton model
1969: the point like constituents of the nucleon were termed
partons by Feynman well before quarks and gluons became
established
At high Q
2
the electron sees pointlike objects called partons and makes
at elastic e parton scattering:
q(, p)
xp
p
=
In the infinite momentun frame p >> M
p
, m
q
, due to time dilation:
the virtual photon * sees free partons. This is the limit
of asymptotic freedom very small values of QCD coupling:
DIS ep scattering is the sum of elastic e-parton scatterings
t
parton
t
partonparton
Asymptotic freedom & Quark parton model
pertubative QCD approach to hadronic physics applies to inclusive
hard-scattering processes
- based on
{
Asymptotic freedon (AF)
Parton model
1) Hard scattering: at leat one momentum scale Q >> M
hadron
1 GeV
- At this scale QCD coupling (Q
2
) can be sufficiently small
o
s
( Q
2
)=
4 n
0
ln ( Q
2
/ A
QCD
2
)
=
4 n
(112/ 3 N
f
) ln ( Q
2
/ A
2
)
(Q
2
)
Q
2
2) Inclusive: Factorization of long and short distance physics
~ (f
1
f
2
)
hard
d + O(1/p
)
p
), where
d = final-state decay (fragmentation) function; O( p
)
is a QCD power correction (higher-order )
as afunction of transverse mom.
The scaling variable x
BJ
What is the physical meaning of x (x
BJ
)?
q
p
e,
zp
zp+q
m
i
2
= (q + zp)
2
0
(zp)
2
Q
2
+ 2zpq = 0
z = Q
2
/2pq x
Defn: scaling variable x fraction of proton momentum carried by the
interacting parton (0 < x < 1).
Can measure quarks' momentum from scattered leptons alone: (E', )
Partons were later indentified with quaks (Gell-Mann in 1964): They had a
more mathematical model at the time.
Structure functions F
1
and F
2
Extracted by comparing with the general expression for cross-section
F
2
+
=( 2 z
u
2
+z
d
2
)
|
++
q
2
2 Mx
2 F
1
M
=( 2 z
u
2
+z
d
2
)
|
q
2
2 M
2
x
2
6
|
++
q
2
2 Mx
Summing over uud quarks in the proton
F
2
/ +
2 F
1
/ M
=
2 M
2
x
2
q
2
=
Mx
+
F
2
=2 xF
1
Callan-Gross relation
F
2
=
2
x
F
1
x
m/M
0
1
Elastic scattering (e
-
q e
-
q) cross-section in QPM
(
d
2
c
dE ' d D
)
LAB
=
d
2
c
dx dQ
2
=
4 no
2
xQ
4
( y
2
xF
1
( x )+(1 y) F
2
( x ))
where the variables x and Q
2
are
Q
2
= q
2
= (k k')
2
and x =
Q
2
2 p.q
y=
2 p.q
p.k
inelasticity, is the fraction of lepton energy transferred to the
proton in its rest frame
Q
2
=xys
Can also show that
x =
m
M
, where m and M are parton and proton
masses, respectively
Using the Feynman hypothesis:
d
2
c
dx dQ
2
=
i
dx f
i
( x )
d
2
c
i
dxdQ
2
d
2
i
/dxdQ
2
is the cross-section for elastic electron-quark scattering
Putting it all together we get:
F
2
=2 xF
1
=
i
e
i
2
xf
i
( x )
What is F
1
(x) ?
From the data:
F
2
(x) - 2xF
1
(x) F
L
L
Where
L
= cross-section for absorption of longitudinally polarised
photons
F
L
= longitudinal structure function
Due to helicity conservation; only objects with spin can collide
head on with photons of helicity +1 or -1, objects with spin 0 cannot
absorb a photon
- quarks have spin 1/2
Experimental proof of Callan-Gross relation
What we've learnt thus far about the QPM
Bjorken scaling:
F
1
and F
2
are functions of one variable, not two, because the
underlying scattering is elastic and pointlike
Callan-Gross relation:
However, in reality these structure
functions are not delta functions
- must take into account quark-gluon coupling!
x
1/3
1
F
2
What is F
2
(x)?
What is F
2
(x) (cont'd)?
- can use this to look at proton structure in detail
u
p
(x) = probability that a u quark in a proton has momentum fraction x
d
p
(x) = probability that a d quark in a proton has momentum fraction x
0
1
u
p
( x) dx =2 and
0
1
d
p
( x) dx =1, normalisation
F
2
ep
( x)=
0
1
|
z
u
2
xu
p
( x)+z
d
2
xd
p
( x)
6
|
x +
q
2
2 M +
=
4
9
xu
p
( x)+
1
9
xd
p
( x)
Similarly neutron structure can be worked out:
F
2
e n
( x)=
4
9
xu
n
( x)+
1
9
xd
n
( x)
For isospin invariance:
(p=uud; n=udd )
&
{
F
2
ep
( x)=
4
9
xu
p
( x)+
1
9
xd
p
( x)
F
2
e n
( x)=
4
9
xd
p
( x)+
1
9
xu
p
( x)
{
u
n
( x )=d
p
( x )u( x )
d
n
( x )=u
p
( x )d ( x )
From now on we drop the suffix and include the sea quark contribution
F
2
ep
( x)=x
|
4
9
u( x)+
1
9
d ( x)+
4
9
u( x)+
1
9
d ( x)+
1
9
s( x)+
1
9
s( x)+...
F
2
e n
( x)=x
|
4
9
d ( x)+
4
9
d ( x)+
1
9
u( x)+
1
9
u( x)+
1
9
s( x)+
1
9
s( x)+...
Area under F
2
, neglecting the strange quark sea:
0
1
F
2
ep
( x) dx =
4
9
0
1
x ( u( x)+u( x)) dx +
1
9
0
1
x ( d ( x)+d ( x)) dx =
4
9
f
u
+
1
9
f
d
0
1
F
2
e n
( x) dx =
4
9
0
1
x (du( x)+d ( x)) dx +
1
9
0
1
x ( u( x)+u( x))dx =
4
9
f
d
+
1
9
f
u
Where f
u
(f
n
) is the fraction of proton(neutron) momentum carried
by all quarks and anti-quarks
The above momentum sum rule f
u
= 0.36 and f
n
= 0.18
- quarks (anti-quarks) only carry half the momentum
first indirect evidence of GLUONS!
First estimate on parton densities were done by comparing F
2
ep
& F
2
en
- take separate contributions of the valence and the sea
u( x )=u
v
( x )+u
s
( x )
&
d ( x )=d
v
( x )+d
s
( x )
All sea components are equal
u
s
( x )=d
s
( x )=u( x )=d ( x )S ( x )
F
2
ep
( x)=x
|
4
9
u
v
( x)+
1
9
d
v
( x)+
10
9
S
F
2
e n
( x)=x
|
4
9
d
v
( x)+
1
9
u
v
( x)+
10
9
S
F
2
ep
( x)F
2
e n
( x)=x
|
1
3
u
v
1
3
d
v
0
1
u
v
( x ) dx=2
0
1
d
v
( x ) dx=1
&
...normalisation
0
1
1
x
( F
ep
F
e n
) dx=
1
3
0
1
( u
v
d
v
) dx=
1
3
0
1
1
x
( F
ep
F
en
) dx=
1
3
0
1
( u
v
d
v
) dx=
1
N
v
Gottfried sum rule
Consistent with experiment
&
Example of parameterisation (parton distributions)
What we have learned until now
Nucleons are very complicated objects
To study them we need to make deep inelastic scattering with leptons
- this probes a distance proportional to 1/Q
The differential cross-section depends on 2 structure functions F
1
and F
2
From the scaling properties of two structure funcions, the QPM model
was developed i.e. the probe sees putative objects (free pointlike)
partons inside the proton.
Scaling Violations
In late 70's: experiments on DIS at CERN and Fermilab showed that,
at higher Q
2
values, deviations from Bjorken scaling appeared.
F
2
(x) F
2
(x,Q
2
)
- F
2
grows with Q
2
at low x (sea region)
- F
2
decreases with Q
2
at high x (valence region)
Good news because we can apply pertubative QCD
- Dokshitzer-Gribov-Lipatov-Altarelli-Parisi equations.
- QPM is the 0
th
order of pertubative expansion
F
2
( x , Q
2
)=
q
xe
2
( q( x)+Aq( x , q
2
))
Aq( x , Q
2
)=
o
s
2 n
x
1
dx '
x
q( x ' ) P
q q
(
x
x
' ) ln(
Q
2
k
2
)+...
P
qq
is probability that a quark emits a gluon radiating a fraction x/x'
of its momentum, when Q
2
changes by dlnQ
2
Dokshitzer-Gribov-Lipatov-Altarelli-Parisi
equations
F
2
increases with Q
2
at small x as the number of soft gluons
increases inferred from data
F
2
decreases with Q
2
as the valence quarks emit gluons and q(x)
decreases
The proton, according to HERA measurements:
It is densely filled with quarks, antiquarks and
gluons.
New view of the proton as a result of scaling violations
- Number of quark anti-quark pairs is unexpectedly large
Scaling violations at low x (HERA)
Low Q
2
fixed target data
HERA data
F
2
increases steeply at low x
At low x, the proton is made of many partons with low momentum
- so far, no parton recombinations have been seen!
Other topics not covered in this series
Neutrino Deep Inelastic scattering
Charged current (cc) has only electro-weak contribution
(E
)
l(e,)
W
P
Neutral current (NC) has both electro-weak and
electromagnetic contributions
(E
)
l(e,)
*, Z*
P
Can use the above to look for :
new Physics (squarks, lepto-quarks and study of quark-substructure!)
Other exciting, new physics to be looked at
- Low x behaviour at LHC (2007 - )
e.g. Higgs production via gluon-gluon fusion