Inter-quark Potential

from Nambu-Bethe-Salpeter Amplitude

based on Ikeda, Iida, arXiv:1102.2097[hep-lat](2011).

Yoichi IKEDA
(Tokyo Institute of Technology)
in collaboration with

Hideaki IIDA (Kyoto Univ.)

Contents

Inter-quark potentials
Constituent quark model from Lattice QCD
Results on Qbar-Q potentials with finite masses
Summary

Mass spectrum of charmonium system

Quark potential models well describe
mass spectra below open charm threshold

BaBar Collaboration

Godfrey, Isgur, PRD 32 (1985).

Barnes, Godfrey, Swanson, PRD 72 (2005).

Charmonium-like (X, Y, Z) states can be

expected as candidates of exotic hadrons
All NEW states reveal as resonances
above open charm threshold

Important information is T-matrix elements based on QCD

"
#
!

R
s

s
(s
)
0
i
1
#
T ( s) =
+ a(s0 ) +
ds #

# s )
s

W
2
(s

s)(s
i
0
s
+
i

(General form of amplitudes from N/D method)

Interaction parts cannot be determined within the scattering theory

Key observations are QQbar and/or meson-meson interactions

bar
Q -Q

interquark potential

Qbar-Q potential is expected having the following form :

VQQ
(r) = r

4 s
3 r

#Q
#Q + VT (r)S
12 + VLS (r)L
# S
# +
+ Vspin (r)S
S

Spin-independent

Spin-dependent

Confinement + Coulomb-like + spin-dependent force

Effective field theory approach (pNRQCD) :
Wilson loop + relativistic correction (1/mQ, v (velocity), 1/mQv expansion)
Bali, Phys. Rept. 343 (2001).
Brambilla, Pineda, Soto, Vairo, NPB 566 (2000); Rev. Mod. Phys. 77 (2005).
Koma et al., PRL 97 (2006).
Koma et al., NPB 769 (2007).

Our approach through Nambu-Bethe-Salpeter (NBS) amplitude :

We can define inter-quark potential with finite quark mass
which becomes faithful to QCD T-matrix
Lin et al., NPB 619 (2001).
Aoki, Hatsuda, Ishii, PTP 123 (2010).

Reliable input based on QCD for quark potential models can be extracted

How to define quark model on the lattice

Y.I., Iida, arXiv:1102.2097[hep-lat](2011).
Homogenous Nambu-Bethe-Salpeter equation :

E (p0 , p) = G(p; P )

d4 p! K(p, p! ; P )
E (p!0 , p! )

P : meson 4-momentum P = (E, 0) = (Mmeson, 0) at meson-rest frame

p, p : relative 4-momentum of Qbar-Q system
K(p,p;P) : irreducible kernel
G(p;P) : product of free quark propagator w/ assumption of constant quark mass mQ

Non-relativistic reduction through Levy-Klein-Macke (LKM) method :

E (p) =

1
2i

"

Reviewed in Klein, Lee, PRD 10 (1974).

E (p0 , p)
dp0 (p0 P 0 /2 + #(p) i)1 + (p0 p0 )1

-> Equal-time Nambu-Bethe-Salpeter amplitude = Nambu-Bethe-Salpeter wave function

Nambu-Bethe-Salpeter wave functions satisfy

Shrdinger-type equation :
!

E (p) =
(E 2!(p))

E (p! )
(p, p! )
dp! U

!
!(p) = m2Q + p2

Note :
Potential is non-local but energy-independent below open charm threshold
because of Krolikowski-Rsewuski relation

Krolikowski, Rzewuski, Nuovo Cimento, 4 (1956).

bar
Q -Q

interquark potential on the lattice

Aoki, Hatsuda, Ishii, PTP 123 (2010).
Ikeda, Iida, arXiv:1102.2097[hep-lat](2011).

1. Measure equal-time Nambu-Bethe-Salpeter wave function

!
E (r) =
!0|
q (x)q(x + r)|E; J P C "
Spacial correlation of 4-point function
x
G(2) (r, t tsrc ) =

x,X,Y

"0|
q (x, t)q(x + r, t) q(X, tsrc )q(Y, tsrc )

!!
x

"

En

|0#

AEn !0|
q (x)q(x + r)|En "eEn (ttsrc )

AE0 E0 (r)eE0 (ttsrc )

(E0 = M, t ! tsrc )

2. Define potential through Schrdinger-type equation

(E H0 )E (r) =

dr! U (r, r! )E (r! )

3. Velocity expansion of non-local potential

!Q
!Q + VT (r)S
12 + VLS (r)L
! S
! + )(r r! )
U (r, r! ) = (VC (r) + Vspin (r)S
S

Leading order

NLO

We examine s-wave effective LO potentials in pseudo-scalar and vector channels

LQCD setup
Y.I., Iida, arXiv:1102.2097[hep-lat](2011).
Quench QCD simulation
Plaquette gauge action & Standard Wilson quark action
=6.0 (a=0.104 fm, a-1=1.9GeV)
Box size : 323 x 48 -> L=3.3 (fm)
Four different hopping parameters (=0.1320, 0.1420, 0.1480, 0.1520)
-> MPS=2.53, 1.77, 1.27, 0.94 (GeV), MV=2.55, 1.81,1.35, 1.04 (GeV)
Nconf=100
Wall source
Coulomb gauge fixing

bar
Q -Q

wave functions

NBS wave functions (channel & quark mass dependence)

MPS=2.53, 1.77, 1.27, 0.94 (GeV), MV=2.55, 1.81,1.35, 1.04 (GeV)
!
E (r) =
!0|
q (x)q(x + r)|E; J P C "
x

Pseudoscalar channels

mq

Vector channels

mq

Channel dependence appears in light quark sector

-> spin-spin interaction is enhanced at light quark mass
(one-gluon exchange predicts Vspin proportional to 1/mq2)
Size of wave function becomes smaller as increasing mq

bar
Q -Q

potentials
Effective spin-independent & dependent forces are constructed
by linear combination of PS & V channel potentials

eff
(r) E =
Vspinindep.

1 ! 1 2 PS (r)

mq 4 PS (r)

3 V (r) "
2

4 V (r)

eff
Vspindep.
(r) E =

1
mq

PS (r)
2

PS (r)

Spin-independent forces reveal Coulomb + linear behavior

Spin-dependent forces have strong quark mass dependence
Repulsive spin-dependent forces as expected by mass spectrum

V (r)
2

V (r)

"

Fit results of
eff
(r)
Vspinindep.

E =

bar
Q -Q

1 ! 1 2 PS (r)

mq 4 PS (r)

potentials

3 2 V (r) "
4 V (r)

fit function: V (r) = r

MV (GeV)
2.55
1.87
1.35
1.04

A
r

+C

(MeV/fm) A (MeV fm)

822 (49)
200 (7)
766 (38)
228 (6)
726 (39)
269 (7)
699 (57)
324 (12)

String tension has moderate mq dependences

Naive extrapolation to infinite mass gives comparable value from Wilson loop
Coulomb coefficients increase as decreasing mq
see also, Kawanai and Sasaki, PRL 107 (2011).

Operator dependence of

bar
Q -Q

potentials

Operator dependence of inter-quark potential is studied

by using gauge invariant smearing operator

(r) =
smr.
E

!
x

!0|
q (x)L(x, r)q(x + r)|E; J P C "

Comparison with Coulomb gauge potentials

Vector channel

Pseudoscalar channel

mq (VCeff (r) E) =

2 smr.
(r)
E
smr.
(r)
E

The potentials obtained from gauge invariant smearing operators

are consistent with the Coulomb gauge potentials

Relativistic kinematics
Inter-quark potentials with relativistic kinematics are studied
(VCeff (r) E) =
H0 E (r) =

"!
dr

dp
(2)2

m2q

H0 V
E (r)
V
E (r)
+

!
p2 eip(rr )

=0.1320 : mq ~ mc
$

E (r" )

=0.1520 : mq ~ ms

Even for relativistic kinematics, Coulomb + linear potentials are obtained

Long range parts of relativistic potentials are consistent with those of N.R. potentials
For charmonium, non-relativistic kinematics is good enough
In strangeness sector, non-locality of potentials gets to large, if one employes nonrelativistic kinematics

Summary
We study inter-quark interactions with finite quark mass in
quenched QCD simulation
Effective central Qbar-Q potentials from NBS amplitudes reveal
Coulomb + linear forms
Coulomb coefficients become smaller and smaller as increasing mq
String tension also has mq dependence and is consistent with that
of Wilson loop analysis at large mq limit
Spin-spin interactions are repulsive and strongly dependent on
quark masses

Future plans : Full QCD

Studies of tensor, LS, non-locality of inter-quark potential
Interquark potential for baryon (talk by H. Iida)
Coupled channel analysis toward above open charm threshold
Investigation of exotic states (X, Y, Z)

Thank you very much for your attention

Check (I) : O(a) improvement

! We study cutoff dependence of the qbar-q potentail
by adopting O(a)-improved Wilson-clover quark action
We compare Standard Wilson quark action
with O(a) improved action (clover action)!

Small difference!

Check (II) : Volume dependence

! We study volume dependence of the qbar-q potentail by varying lattice spacing
for O(a)-improved Wilson-clover quark action

(Lattice unit)!

L=4.5fm (!=5.8, mPS=2.47GeV, clover): red

L=3.2fm (!=6.0, mPS=2.58GeV, standard): green!

Pseudoscalar channel!

! Small difference between them volume is enough

Our setup (!=6.0, a=0.1fm, standard Wilson, (3.2fm)3) seems sufficient

for the calculation of qbar-q potential (in quark mass region calculated here)!

bar
Q -Q

potential from NBS wave function

Inter-quark potential with various finite quark masses

MPS=2.53, 1.77, 1.27, 0.94 (GeV), MV=2.55, 1.81,1.35, 1.04 (GeV)

VCeff (r) E =

1 2 E (r)

mq

mq = MV /2

Vector channels

Vc(r)-E (MeV)

Pseudoscalar channels

Vc(r)-E (MeV)

E (r)

Coulomb + linear confinement forces are reproduced with finite quark masses
(solid curves representing Coulomb + linear functions)

Relativistic kinematics
Inter-quark potentials with relativistic kinematics are studied
=0.1320 (MV=2.55 GeV)
=0.1420 (MV=1.77 GeV)

=0.1480 (MV=1.27 GeV)

=0.1520 (MV=1.04 GeV)

Even for relativistic kinematics, Coulomb + linear potentials are obtained

Long range parts of relativistic potentials are consistent with those of N.R. potentials
For charmonium, non-relativistic kinematics is good enough
For strangeness sector, non-locality of potentials may get to large, if one employes
non-relativistic kinematics

bar
c -c

potential from full QCD

Kawanai, Sasaki, arXiv:1110.0888[hep-lat].

cbar-c potential from 2+1 flavor FULL QCD simulation at almost PHYSICAL POINT
generated by PACS-CS Coll. (m=156(7), mK=553(2) MeV)
Iwasaki gauge action (=1.9, a=0.091 fm) + RHQ action
-> Mave.(1S) = 3.069(2) GeV, Mhyp.=111(2) MeV
Spin-independent force

Spin-independent force shows

Coulomb + linear form
Lattice QCD potential is
consistent with NRp model
Barnes, Godfrey, Swanson, PRD 72 (2005).

Spin-dependent force

Spin-dependent force shows

short range but not point-like repulsion
see also, Kawanai and Sasaki, PRL 107 (2011).

How to define quark model on the lattice

How to define quark model on the lattice

We start with NBS equation for invariant amplitudes at meson rest frame :

M(p, p! ; P ) = K(p, p! ; P ) +

d4 kK(p, k; P )G(k; P )M(k, p! ; P )

P : meson 4-momentum P=(M, 0) at center-of-mass frame

p, p, k : relative 4-momentum of Qbar-Q system
K(p,p;P) : irreducible kernel
G(k;P) : product of free quark propagator w/ assumption of constant quark mass mQ

Non-relativistic reduction through Levy-Klein-Macke (LKM) method

Reviewed in Klein, Lee, PRD 10 (1974).

Concept of LKM method :

Replacement of free-propagator G(k;P) to non-relativistic one GN.R.(k;P)
leads to rearrangement of interaction
kernel of original NBS equation
!

M(p, p! ; P ) = I(p, p! ; P ) +

d4 kI(p, k; P )F (k; P )GN.R. (!

k, P )M(k, p! ; P )

I(p,p;P) is new kernel and satisfying I = K + K(G F GN.R. )I

"
1 ! 0
F (p; P )
(p P 0 /2 + "(#
p) i)1 + (p0 p0 )1
2i
Note :
We do not require instantaneous NBS kernel K(p,p;P) in LKM method

How to define quark model on the lattice

3-dimensional LKM equation for NBS invariant amplitude:

!
!
P M(p, p ; P ) = P I(p, p ; P ) + P

Reviewed in Klein, Lee, PRD 10 (1974).

d4 kI(p, k; P )F (k; P )GN.R. (!

k; P )M(k, p! ; P )

U HP

dp0 F (p; P )

L.H.S. of LKM equation is found as equal-time NBS wave function (positive energy)
-> Schrdinger-type equation for NBS wave function is easily derived :

E ("
(E 2!("
p))
p) =

E ("
d3 p! U ("
p, p
"! ; E)
p! )

E = P 0 = Mmeson

with non-local, energy-dependent potential U(p,p;E) satisfying

(E H0 )E ("
r) =

!
U (!
p, p
! ; E) = P I(p, p ; E) P
!

d3 r ! U ("
r, "
r ! ; E)E ("
r! )

Schrdinger-type equation for NBS wave function in r-space

Summary of LKM method :
Relativistic 3-dimensional equation are extracted from equal-time NBS wave
function -> suitable for LQCD simulation
Obtained 3-dimensional equation is Schrdinger-type equation (LKM equation)
Effective potential of LKM equation is related to irreducible kernel K(p.p;P) of
NBS equation -> potential model based on QCD can be constructed

Interquark potentials for baryons

Three-body Schrdinger-type equation :
!

2r
2

2
2

"

+ V (r, ) E (r, ) = EE (r, )

Effective two-quark potentials : integrate out spectator particle

Doi et al, (HAL QCD Coll.), arXiv:1106,2276 [hep-lat] (2011).
2q
E
(r)

dE (r, )

2r
2

"

2q
2q
2q
+ Veff
(r)

(r)
=
E
E
E (r)
.

Effective 2q potential v.s. qbar-q potential (spin independent parts)

V2q (r) = r

2 s

Vqq (r) = r

4 s

3 r
3 r
String tension would be comparable (Wilson loop analyses)
Coulomb part is different by factor two with one-gluon exchange
Three-quark potential
T.T.Takahashi et al., PRD70, 074506 (2004)

Valence light quark effect for 2Q

A.Yamamoto et al., PLB664, 129 (2008)

Interquark potentials for baryons

Lattice QCD result of effective 2q potential:
2q
Veff
. (r)

2q
1 2r E
(r)

2q
E
(r)

+E

2q
E
(r)

!0|"abc (qaT (r/2)C5 qb (r/2))qc, ()|B, J P = 1/2+ #

Effective 2q potential shows

Coulomb + linear form

fit
(r)
V2q

= r

A
r

+C

2/Ndf = 0.7 [ 0.2 < r < 0.8 fm ]

= 797 (8) [MeV / fm]
A = 78 (2) [MeV fm]

String tension of effective 2q potential is comparable with that of qbar-q potential

(c.f., qbar-q =822 [MeV / fm], Aqbar-q =200 [MeV fm])