Download pdf - Chiral effects of quenched η′ loops

ELSEVIER Nuclear Physics B (Proc. Suppl.) 83-84 (2000) 215-217

PROCEEDINGS SUPPLEMENTS

www.elsevier.nl/locate/npe

Chiral Effects of Quenched Loops *

W. Bardeen~,A. Duncanb,E. Eichten~,and H. Thacker ¢

Fermilab, PO Box 500, Batavia, IL60510

bDept, of Physics and Astronomy, University of Pittsburgh, Pittsburgh, PA 15260

~Dept. of Physics, University of Virginia, Charlottesville, VA 22901

Preliminary results of a study of quenched chiral logarithms at /~ = 5.7 are presented. Four independent determinations of the quenched chiral log parameter 5 are obtained. Two of these are from estimates of the ~' mass, one from the residue of the hairpin diagram and the other from the topological susceptibility combined with the Witten-Veneziano formula. The other two determinations of 5 are from measurement of virtual ~/' loop

2 effects in m~ vs. quark mass and in the chiral behavior of the pseudoscalar decay constant. All of our results are consistent with ~ = .080(15). The expected absence of quenched chiral logs in the axial-vector decay constant is also observed.

1. Q u e n c h e d C h i r a l Logs

One of the most distinctive physical effects of light quark loops in QCD is the screening of topological charge, which is responsible for the large mass of the ~1 meson. The absence of screening in the quenched approximation gives rise to singular chiral behavior (quenched chiral logs) [1] arising from soft ~' loops. In full QCD ~ loops are finite in the chiral limit, while in quenched QCD they produce logarithmic singularities which ex- ponentiate to give anomalous power behavior in

2 O. The first evidence of quenched the limit m r --+ chiral logs in the pion mass was reported last year[2,3]. Here, and in a forthcoming paper [4}, we present the results of a more detailed study of this effect.

In the context of the effective chiral Lagrangian for QCD, the V(3) × U(3) chiral field V =

• ~ i = 0 s A exhibits singular chiral be- exp

havior in the quenched theory arising from loga- rithmically divergent ~1' = S0 loops:

U--~ exp [-(¢02)/2f~] U : ~ A2 ~ ' ~ " \ j (1)

where U is finite in the chiral limit, and A is an upper cutoff on the r/' loop integral. (We con-

*Talk p resen ted by H. Thacker

0920-5632/00/$ - see front matter © 2000 Elsevier Science B.V. PII S0920-5632(00)00231-0

sider, for simplicity, the case of three equal mass light quarks.) The anomalous exponent 5 is de- termined by the ~1 hairpin mass insertion m20,

5 - 48~r2f ~ (2)

(Here f~ is normalized to a phenomenological value of 95 MeV.) From the chiral behavior (1) of the field U, we can infer the singularities expected in the matrix elements of various quark bilinears. Matrix elements of operators which flip chirality, such as the pseudoscalar charge ~75~b c< U - U t should exhibit a chirally singular factor (m2) -~. By contrast, operators which preserve chirality, such as the axial vector current ~bTsT~'~b o¢ iU-XO~'U + h.c are expected to be finite in the chiral limit. From PCAC, it follows that

1 2 mW (3)

In the data presented here, we verify all of these predictions by studying the pion mass and the vacuum-to-one-particle matrix elements

(0l~)-rs~plTr(p)) = f p (4)

(0[~)-rs-r~'¢lTr(p)} = lAP • (5)

The modified quenched approximation (MQA) [5] provides a practical method for resolving the

All rights reserved.

216 W. Bardeen et al./Nuclear Physics B (Proc. Suppl.) 83--84 (2000) 215-217

0.16

0.14

0.12

m 0.1

~ 0 . 0 8 ,:"

~ 0.06 .¥" t

a .

0.04

0.02 " "

0 i a i i i i i

0 5 10 15 20 25 30 35 40 bare quark mass (latxlOA3)

Figure 1. Chiral log effect in m 2 vs. mq. Solid line is a perturbative (quadratic) fit to the five largest masses (four are not shown). Dotted line is fit of all masses to Eq. (3).

problem of exceptional configurations, and allows an accurate investigation of the chiral behavior of quenched QCD with Wilson-Dirac fermions. The pole-shifting prescription for constructing improved quark propagators is designed to re- move the displacement of real poles in the quark propagator, which is a lattice artifact, while re- taining the contribution of these poles to con- tiuum physics. We have used the MQA procedure to study the masses and matrix elements of flavor singlet and octet pseudoscalar mesons in the chiral limit of quenched QCD. A complete discussion of these results will be presented in a forthcoming publication. Here we present results from 300 gauge configurations at ~ = 5.7 on a 123 × 24 lattice, with clover-improved quarks (Cs~ -- 1.57) at nine quark mass values covering a range of pion masses from .2387(53) to .5998(17).

In addition to observing ~/' loop effects, we also present two direct estimates of the chiral log paramter 5, one from a calculation of the ~?' hairpin diagram, and the other from a calculation of the topological susceptibility combined with the Witten-Veneziano relation. The results for 5 agree well with each other and with the expo- nents extracted from f P / f A and m 2, thus giv- ing four independent and consistent determina-

tions of 5. All four results fall within one stan- dard deviation of 5 -- .080(15). This is about a factor of two smaller than the result 5 = .17 expected from the phenomenological values of m0 ~ 850 MeV and f,~ -- 95 MeV. The agreement between the four determinations of 5 indicates that relations imposed by chiral symmetry and the Witten-Venezian formula are approximately valid, even at ~3 = 5.7. Possible reasons for the overall suppression of the exponent 5 compared to phenomenological expectations will be discussed in Ref. [4].

2. The ~' hairpin mass insert ion

Following Ref. [6] we calculate ~5 quark loops using quark propagators with a source given by unit color-spin vectors on all sites. The statisti- cal errors are dramatically improved by the MQA procedure, allowing a detailed study of the time- dependence even at the lightest masses. We de- termine the value of m02 with a one-parameter fit to the overall magnitude of the hairpin correla- tor, assuming a pure double-Goldstone pole form. In the chiral limit, we find moa = .601(30), or m0 = 709(35) MeV using a -1 = 1.18 GeV. (For unimproved C8~ -- 0 fermions, we get a much smaller value of 464(24) MeV.) Using Eq. (2), we obtain the values for the chiral log parameter 5. Extrapolating to the chiral limit, this gives 5 = .068(8) if we use the unrenormalized lattice value for f~. If a tadpole improved renormaliza- tion factor is included, this becomes 5 = .095(8).

3. Topological suscept ib i l i ty

The fermionic method for calculating the topological susceptibility[7] can be implemented with the same allsource propagators used for the hairpin calculation. For each configuration, we com- pute the integrated pseudoscalar charge Qs- The winding number v = - imqQ5 is then obtained and the ensemble average (v 2) is calculated. In the chiral limit, this gives Xt = (188 MeV) a. (Here we have used a -1 -- 1.18 GeV.) Using the Witten-Veneziano formula (with unrenormalized f,~) to obtain m0, the chiral log parameter 5 -- .065(8) is found.

W.. Bardeen et al./Nuclear Physics B (Proc. Suppl.) 83-84 (2000) 215-217 217

0.48

0.46

m 0.44 I t

--' ~" 0.420.4 t

0.38

0.36 0.65 0:1 0.15 0:2 0.25 0:3 0.35 0.4 pion massA2 Oat)

Figure 2. fp vs. m 2 in lattice units. Solid line is a linear fit to the four largest masses.

0.22 0.21 0.2

0.19 0 . 1 6

~0.17 0.16 0.15 0.14 0.13 0.12 0 0.05 0:1 0.15 0:2 0.25 0:3 0.35 0.4

pion massA2 Oat) 2 in lattice units. Solid line is Figure 3. fA vs. m,~

a linear fit to the four largest masses.

4. Q u e n c h e d ch i ra l logs in t h e p ion m a s s

The pion masses are obtained for nine kappa values over a range of hopping parameters from .1400 to .1428, with C8~ = 1.57, corresponding to quark masses from roughly the strange quark mass down to about four times the up and down quark average (i.e. a pion mass of ~-, 270 MeV).

2 A perturbative (linear+quadratic) fit of m~ as a function of mq works well for the five heaviest masses, up to ~ -- .1420. The value of m~ for the lighter quark masses deviates significantly below this perturbative fit (see Fig. 1) showing clear evidence of a quenched chiral log effect. Fit t ing to the formula (3) gives an anomalous exponent

-- .079(8).

5. P s e u d o s c a l a r a n d ax ia l -vec to r matr ix el- e m e n t s

By a combined fit of smeared-local pseudoscalar and axial-vector propagators, we obtain values for the decay constants fp and fA defined in (4)-(5). The behavior of these two constants as a function of pion mass squared is shown in Figs. 2 and 3. We see that there is a very signif- icant chiral log enhancement at light masses for the pseudoscalar constant, but the axial-vector shows no chiral log effect, just as theoretical ar- guments predicted. From the phenomenology of

O(p 4) terms in the chiral Lagrangian, it can be argued that the perturbative slopes of fp and fA in full QCD should be approximately equal. In- deed, if a chiral log factor is removed from fp, the remaining slopes seen in our lattice data are equal within errors. To obtain an estimate of 5, we fit the ratio fP/fA to a pure chiral log factor:

fP const. × (m2) -~ /A

This fit gives 5 = .080(7).

R E F E R E N C E S

(6)

1. S. Sharpe, Phys. Rev. D46 (1992) 3146; C. Bernard and M. Golterman, Phys. Rev. D46 (1992) 853.

2. W. Bardeen, et al., Nucl. Phys. B (Proc. Suppl.) 73 (1999) 243.

3. S. Aoki, et al., Nucl. Phys. B (Proc. Suppl.) 73 (1999) 189.

4. W. Bardeen, et al. (in preparation). 5. W. Bardeen, A. Duncan, E. Eichten, and

H. Thacker, Phys. Rev. D57 (1998) 1633. 6. Y. Kuramashi, et al., Phys. Rev. Lett. 72

(1994) 3448. 7. J. Smit and J. Vink, Nucl. Phys. B286

(1987)485.