detailed study of parton energy loss via measurement of fractional momentum loss of high p t hadrons...
TRANSCRIPT
Detailed study of parton energy loss via measurement of fractional momentum loss of high pT hadrons in heavy ion collisions
Takao Sakaguchi
Brookhaven National Laboratory
for the PHENIX Collaboration
Hard scattering as densimeter Parton may change its momentum in the medium.
– Energy loss through Gluon radiation, etc..
Effect is path-length and parton density dependent ~a densimeter~.
Look at leading particles of jet as a measure of jet energy.
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France2 X.-N., Wang, PRC 58 (1998)2321
Energy loss =Yield suppress
0 without energy loss
0 with energy loss
Yie
ld [
GeV
-1c]
pT [GeV/c]
p+p Au+Au
Cross section of A+B collisions AB p+p collisions
High pT hadron RAA
Measured in 200GeV Au+Au collisions.
p0 and h are very consistent.
Results have been solid over years.
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France3
PRC82, 011902(R) (2010)
MinBias, Black: p0, Blue: h
RAA (1/Nevt
AA )(dN AA /dpT )
TAA (d pp /dpT )
Black: Year-2004, Red: Year-2007
PRC87, 034911 (2013)
MinBias
Black- 2004Red - 2007
0-10%
20-30% 40-50%
60-70% 80-93%
Collision energy dependence of RAA p0 in Cu+Cu @ 200, 62.4, 22.4GeV (RHIC Year-5)
– 200GeV and 62GeV show suppression, and 22.4GeV shows enhancement (Cronin)
p0 in Au+Au @ 200, 62.4, 39GeV (RHIC Year-4, 7, 10)– Similar suppression at higher pT
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France4
PRL 101, 162301(2008)PRL109, 152301 (2012)
Centrality dependence of RAA
Centrality dependence of integrated p0 RAA was investigated in Cu+Cu and Au+Au collisions
– As a function of cms energy
Top: Cu+Cu:– 2.5<pT<3.5GeV/c
Bottom: Au+Au– pT>6GeV/c
Indication of a switch of suppression and enhancement below 39GeV?
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France5
PRL 101, 162301(2008)
PRL109, 152301 (2012)
Similar suppression for RHIC and LHC?
RHIC data from PHENIX publication
– p0 in Au+Au @ 0.2TeV
LHC data from ALICE publication
– Charged hadrons in Pb+Pb @ 2.76TeV
Center-of-mass energies differ by a factor of 14!
Similar RAA doesn’t mean similar energy loss
– RAA is not a good quantity?– Power of the spectra is different
energy loss will be different
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France6
PRC87, 034911 (2013)
Looking at suppressions from a different angle
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France7
New quantity ~momentum loss~ Coming back to the original idea: measuring energy loss of partons
Statistically measure the fractional momentum loss (dpT/pT) of high pT hadrons instead of RAA.
2014-09-128 T. Sakaguchi, HPT2014@Nantes, France
pT
dpT)()(
)(:
AAppppp
pppp
TTT
TT
(p+p) denotes p+p yield scaled by number of binary nucleon-nucleon collisions (Ncoll).
In 2005, statistics was not enough, so.. We assumed that the spectra at high pT follows power-law function
With this assumption, we wrote a formula to obtain dpT/pT from RAA
– Fractional momentum shift: S(pT)/pT=S0
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France9
]/)(1[
'
)('/)()1(22
TT
T
TTTpp
AA
TTAAevtAA
dppdS
dydp
pSppd
T
dydppNdN
]/)(1[]/)(1[
]/)(1[)]([
)(
1
1
1
TTn
TT
TTnT
nTT
TAA
dppdSppS
dppdSp
pSppR
nTpdp
dE
3
3
0
)2/(1
20
1
1)(
,)1()(
SpR
SpR
nTAA
nTAA
)2/(1
0
)(1
)1/(11))(/()(
n
TAA
TTTloss
pR
SpSppSS
Note pT is pT in Au+Au PHENIX, PRC76, 034904 (2007)
Sloss in 200GeV Au+Au collisions
Sloss (dpT/pT), where pT is pT at a Au+Au point
Lines: Sloss Npart2/3
– Sloss increases approximately like Npart2/3
– GLV and PQM model suggested 2/3
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France10PHENIX, PRC76, 034904 (2007)
Path-length dependence of RAA
PHENIX measured the RAA as a function of emission angle with respect to event planes for several centralities
– Gives handle of path-length that partons traverse
Le is path from the center of ellipse to the edge
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France11PHENIX, PRC76, 034904 (2007)
cyan, 60–70%mauve, 50–60%blue, 40–50%green, 30–40%red, 20–30%black, 0–10%
Path-length dependence of Sloss
Converted RAA to Sloss and plotted against Le
Data points nicely aligned for 5<pT<8GeV/c
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France12PHENIX, PRC76, 034904 (2007)
cyan, 60–70%mauve, 50–60%blue, 40–50%green, 30–40%red, 20–30%black, 0–10%
ALICE recently performed similar study
Using Pb+Pb 2.76TeV data (ALICE). L should be equivalent to the Le in the previous slide
L is not very good at higher pT?
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France13
P. Christiansen, K. Tywoniuk, and V. Vislavicius, PRC 89, 034912 (2014)
With better statistics..
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France14
Direct calculation is possible
Direct calculation of fractional momentum loss (dpT/pT) became possible.
One doesn’t need to assume the function form of the spectra.
2014-09-1215 T. Sakaguchi, HPT2014@Nantes, France
pT
dpT
Direct calculation is possible
Direct calculation of fractional momentum loss (dpT/pT) became possible.
One doesn’t need to assume the function form of the spectra.
2014-09-1216 T. Sakaguchi, HPT2014@Nantes, France
p+p p0 spectra Au+Au p0 spectra
Fractional momentum loss (dpT/pT) of p0
Measured fractional momentum loss (dpT/pT) instead of RAA
– In Au+Au collisions
dpT/pT=0.2 in 0-10% centrality, =0.02 in 70-80% centrality
2014-09-1217 T. Sakaguchi, HPT2014@Nantes, France
dp
T/p
T
PRC87, 034911 (2013)
MinBias 0-10%
20-30% 40-50%
60-70% 80-93%
Energy dependence of dpT/pT (I) dpT/pT decreases significantly going from 200 to 62, 39GeV.
Significantly different dpT/pT while the RAA is similar.
2014-09-1218 T. Sakaguchi, HPT2014@Nantes, France
RA
A
PRL109, 152301 (2012)
Energy dependence of dpT/pT (II)
dpT/pT increase by a factor of 1.5 from 200GeV to 2.76TeV.
~0.2 for RHIC and ~0.3 for LHC
2014-09-1219 T. Sakaguchi, HPT2014@Nantes, France
LHC
RHIC
PRC87, 034911 (2013)
Energy dependence of dpT/pT (III)
dpT/pT from 62GeV to 2.76TeV: by a factor of 6 change!
2014-09-1220 T. Sakaguchi, HPT2014@Nantes, FrancePRL109, 152301 (2012)
LHC
RHIC
PRC87, 034911 (2013)
Systematic studies across systems
We tried to plot dpT/pT against universal variables
Number of participant nucleons (Npart)
Number of quark participants (Nqp)
Charged multiplicity (dNch/dh), as a measure of the energy density
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France21
Number of quark participants (Nqp)? Number of quark participants, Nqp,
is estimated using a Glauber model.– Nucleons are distributed using a Woods-
Saxon distribution.– Quarks are distributed around the
nucleon centers following: r(r)= r0proton e-
ar, where a = 4.27 fm-1.
Quarks interact when their distance, d, satisfies the condition of:
Quark-quark inelastic cross section (qq
inel) is estimated by reproducing the n-n inelastic cross section of:
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France22
Ratios of Nqp/Npart as a function of Npart for 62, 130 and 200GeV Au+Au collisions
PHENIX, PRC89, 044905 (2014)
Is Nqp a good scaling variable? Recently, PHENIX found that the dET/dh scales better with Nqp than Npart
dET/dh affects the energy loss of partons
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France23
dET/dh/(0.5 Npart) dET/dh/(0.5 Nqp)
PHENIX, PRC89, 044905 (2014)
dpT/pT over collision systems (vs Npart)
Plotting against Npart
– Npart are obtained in given centrality at given cms energy.
– dpT/pT‘s at pT=7GeV/c of p+p are plotted.
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France24
dpT/pT over collision systems (vs Nqp)
Plotting against Nqp
– Nqp are obtained in given centrality at given cms energy.
– dpT/pT‘s at pT=7GeV/c of p+p are plotted.
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France25
dpT/pT over collision systems (vs dNch/dh)
Plotting against dNch/dh (as a measure of energy density)– dNch/dh are obtained in given centrality at given cms energy.
– dpT/pT‘s at pT=7GeV/c of p+p are plotted.
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France26
Consistency with previous studies
PHENIX studied fractional momentum loss in two publications– PRC76, 034904 (2007), PRL101, 232301 (2008)– Assuming the spectra shape is power-law with the power “n”, we can write:
If we assume dNch/dh Npart1.16, found in PRC71, 034908(2005), we can
write the relationship as follows:
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France27PRL101, 232301 (2008)
dpT/pT over collision systems (vs dNch/dh)
Plotting against dNch/dh (as a measure of energy density)– dNch/dh are obtained in given centrality at given cms energy.
– dpT/pT‘s at pT=7GeV/c of p+p are plotted.
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France28
k=0.55,b=6.010-3
k=0.35,b=2.510-2
Review the money plot
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France29
Summary
Looking at fractional momentum loss gives more insight on actual energy loss of partons– Similar RAA does not give similar energy loss
Centrality, system, and energy dependence of fractional momentum loss is studied.– A trend is seen from 200GeV Au+Au to 2.76TeV Pb+Pb
dpT/pT vs dNch/dh for 200GeV Au+Au and Cu+Cu collisions tend to merge into the one for 2.76TeV Pb+Pb collisions at larger dNch/dh, independent of √sNN.
Within same √sNN, both Npart and Nqp scaling work very well.
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France30
Backup
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France31
Hard scattering as densimeter Parton may change its momentum in the medium.
– Energy loss through Gluon radiation, etc..
Effect is path-length and system dependent ~a densimeter~.
Look at leading particles of jet as a measure of jet energy.
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France32 X.-N., Wang, PRC 58 (1998)2321
Energy loss =Yield suppress
0 without energy loss
0 with energy loss
Yie
ld [
GeV
-1c]
pT [GeV/c] PHENIX, p0 in Au+Au, PRL. 91, 072301 (2003)
dpT/pT over collision systems (vs dNch/dh)
Plotting against dNch/dh (as a measure of energy density)– dNch/dh are obtained in given centrality at given cms energy.
– dpT/pT‘s at pT=7GeV/c of p+p are plotted.
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France33
a=0.55,b=6.010-3
a=0.35,b=2.510-2
Consistency with previous studies
PHENIX studied fractional momentum loss in two publications– PRC76, 034904 (2007), PRL101, 232301 (2008)– Assuming the spectra shape is power-law with the power “n”, we can write:
In this study, if we assume dNch/dh Npart, we can write the relationship as follows:
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France34
cGeVpT /[email protected] cGeVpT /[email protected]
PRL101, 232301 (2008)
d+Au and Au+Au system similarity?
Au+Au 60-92% and d+Au 0-20% have similar Npart, Ncoll
Ratios of all ID’ed hadron spectra are on the same curve
Common production mechanism?
If all CNM scales with Npart, ratios may mean Eloss in the medium in peripheral Au+Au
Low pT increase may rise from rapidity shift in d+Au
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France35
arXiv:1304.3410
Setup for measurement Event triggered by a coincidence
of BBC South and BBC North.– Sitting in 3.1<|h|<3.9
p0 and h measurement– EMCal(PbSc, PbGl): Energy
measurement and identification of real photons.
– Tracking(DC, PC): Veto to Charged particles.
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France36
View From Beam
View From Side
Results presented here are obtained from 0.813 nb-1 Au+Au 200GeV events recorded by PHENIX in 2007.
p0 h
EMCal
How we measure p0, h? Reconstruct hadrons via 2g invariant mass in EMCal (example is in Au+Au)
Subtract Combinatorial background– Compute Mass using gs from different events. (mixed-event technique)
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France37
PRC87, 034911 (2013) PRC82, 011902(R) (2010)
)cos1(2)()( 212
212
212 EEEEM pp
p0
h
p0 h
p0, h spectra in Au+Au
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France38
p0 h
Spectra reached to ~20GeV/c for p0 and ~22GeV/c for h
Systematic errors
Type A: point-by-point fluctuating errors
Type B: pT-correlated errors
Type C: overall normalization errors
2014-09-12 T. Sakaguchi, HPT2014@Nantes, France39
source type 5GeV 10GeV 15GeV 20GeV
peak extraction B 2 2 2 2
acceptance C 2.5 2.5 2.5 2.5
PID efficiency B 7 8 8.5 9
energy scale B 7.5 8 8 8
photon conversion C 2 2 2 2
cluster merging B 0 0 8 18
total 11 12 15 22
source type 5GeV 10GeV 15GeV
peak extraction B 4 3.5 3
acceptance C 2.5 2.5 2.5
PID efficiency B 7 8 8.5
energy scale B 11 12 12
photon conversion C 2 2 2
total 12 15 15
p0 systematic errors
h systematic errors