Connecting the Quenched and Unquenched Worlds Via the Large Nc World

Physics Letters B 543 (2002) 183–188 www.elsevier.com/locate/npe

Connecting the quenched and unquenched worlds via the large Nc world

Jiunn-Wei Chen

Department of Physics, University of Maryland, College Park, MD 20742, USA Received 16 July 2002; accepted 5 August 2002 Editor: H. Georgi

Abstract

In the large Nc (number of colors) limit, quenched QCD and QCD are identical. This implies that, in the effective ﬁeld theory framework, some of the low energy constants in (Nc = 3) quenched QCD and QCD are the same up to higher-order corrections in the 1/Nc expansion. Thus the calculation of the non-leptonic kaon decays relevant for the I = 1/2 rule in the quenched approximation is expected to differ from the unquenched one by an O(1/Nc) correction. However, the calculation relevant to the CP-violation parameter / would have a relatively big higher-order correction due to the large cancellation in the leading order. Some important weak matrix elements are poorly known that even constraints with 100% errors are interesting. In those cases, quenched calculations will be very useful.  2002 Elsevier Science B.V. All rights reserved.

Quantum chromodynamics (QCD) is the under- fermion determinant arising from the path integral to lying theory of strong interaction. At high energies be one. This approximation cuts down the comput- ( 1 GeV), the consequences of QCD can be studied ing time tremendously and by construction, the re- in systematic, perturbative expansions. Good agree- sults obtained are close to the unquenched ones when ment with experiments is found in the electron–hadron the quark masses are heavy ( ΛQCD ∼ 200 MeV) deep inelastic scattering and hadron–hadron inelastic so that the internal quark loops are suppressed. Thus scattering (Drell–Yan process) [1]. At low energies the quenched approximation is an efficient and reli- ( 1 GeV), however, QCD becomes non-perturbative. able tool to use as long as all the quark masses are Quantitative results not based on symmetry properties heavy. However, in the real world, the light quark along can only be calculated directly from QCD via (u, d and s) masses are less than ΛQCD so the jus- computer simulations on a Euclidean lattice. tification for quenching does not apply. In this range In lattice QCD simulations, it is common to per- of quark mass, quenching effects can be systemat- form the quenched approximation which is to drop ically studied by quenched chiral perturbation the- the internal quark loop contributions by setting the ory (QχPT) [2], an effective field theory of quenched QCD (QQCD). In QχPT, quenching could induce a double pole to the flavor singlet meson propagator and E-mail address: [email protected] (J.-W. Chen). could make quenched quantities more singular than

Fig. 1. Feynman diagrams contributing to the pion two point function. The solid lines and curly lines denote the quark and gluon propagators, respectively. the unquenched ones [2] in the infrared region. In the poorly known weak matrix elements for which even ultraviolet region, QCD and QQCD are different as 100% error constraints are interesting. well because QQCD can be considered as having in- It is known that the quenched world has some finite quark masses for the internal loops so its ultra- similarity to the large Nc world. Take a pion two- violet behavior is different from that of QCD. Those point function 0|π(x)π(0)|0, for example: there are uncontrolled effects make it hard to see what we can contributions from diagrams shown in Fig. 1. For learn about QCD from QQCD. convenience, the quarks connected to the external However, there is a known connection between sources are called “valence quarks”; the quarks not QCD and QQCD in the limit that the number of connected to the external sources are called “sea colors (Nc) becomes infinite; the large Nc limit of quarks”. In the quenched approximation, the diagrams QQCD is indistinguishable to the large Nc limit of with sea quark loops such as the one in Fig. 1(b) QCD. In this paper, we study the finite Nc quenching are zero. (Then the gluon–quark coupling in surviving errors systematically using effective field theories with diagrams such as the ones in Fig. 1(a), (c) are usually 1/Nc expansions. Generally, in the real world where rescaled to mimic the situation that the sea quark Nc = 3, some of the contributions which are formally masses are set to infinity. We will not do this rescaling higher order in the 1/Nc expansion could become at this moment but leave it for later discussion.) In numerically important for certain observables near contrast, in the large Nc limit proposed by ’t Hooft the chiral limit (zero quark mass limit). Fortunately, [7], these big contributions are typically non-analytic in 2 quark masses and can be computed in effective field Nc →∞,g→ 0,Ncg → constant, (1) theories of QCD and QQCD. After these big non- analytic contributions in QQCD are replaced by those where g is the quark gluon coupling, both the Fig. 1(b) in QCD, it is plausible that the errors which are diagram with one sea quark loop and the “non-planar” formally O(1/Nc) in the quenched approximation diagram of Fig. 1(c) with one crossing in gluon lines could imply numerically ∼ 30% errors. Sometimes are suppressed by one power of 1/Nc compared to that even smaller errors can be achieved by forming ratios of Fig. 1(a). The diagrams with more sea quark loops of observables, such that the O(1/Nc) errors from or more gluon crossings are further suppressed by the leading terms in the chiral expansion cancel. The even higher powers of 1/Nc. So, in the large Nc limit ratio of B meson decay constants fBs /fB [3,4] is of QCD, only the planar diagrams such as Fig. 1(a) an explicit example. We then study the quenched survive, while in QQCD both Fig. 1(a) and (c) survive. calculations of the non-leptonic kaon decays K → Now if we further take the large Nc limit of QQCD, ππ [5,6]. We find that the quenching error in the the non-planar diagrams vanish as well. From this we part relevant to the I = 1/2 rule is consistent conclude that QQCD and QCD have the same large Nc with O(1/Nc) ∼ 30%. However, the CP violation limit—at least in the mesonic two-point function case. parameter / calculation could have a big quenching It is easy to generalize the above observation to an error due to the large cancellation in the leading terms arbitrary n-point function even involving baryons. To in the 1/Nc expansion. Finally, for illustration, we generalize the baryons to the large Nc case, we follow identify some cases that quenched calculations can be the prescription suggested by Dashen and Manohar [8] very useful. Those are calculations of important but such that there is no valence strange quark in a proton. J.-W. Chen / Physics Letters B 543 (2002) 183–188 185

Q Q Q The identity between the large Nc QCD and large where the low energy constants B , fπ and K3 Nc QQCD imposes some constraints on the low en- have different values than those in QCD and where 2 ergy effective field theories of QCD and QQCD, which M0 is the strength of the η η coupling. In QCD, are the chiral perturbation theory (χPT) [9–12] and the 2 M0 can be iterated to all orders to develop a shifted quenched chiral perturbation theory (QχPT) [2], re- η pole such that η becomes non-degenerate with spectively. χPT and QχPT have the same symmetries the Goldstone bosons in the chiral limit. However, and symmetry breaking patterns as QCD and QQCD. in QQCD, all the sea quark loops are dropped such In both cases, low energy observables are expanded 2 that the iteration stops at one insertion of M0 .This in a power series of the small expansion parameters gives the η propagator a double pole dependence and (p, MGB)/Λχ ,wherep is the characteristic momen- makes physical quantities in QQCD more singular tum transfer in the problem, MGB is the Goldstone than those in QCD in the infrared region [2]. The boson mass and Λχ is an induced chiral perturbation quenched pion mass in Eq. (4) is an explicit example. scale. In QCD, Λ ∼ 1 GeV and the Goldstone bo- 2 χ The quenched chiral logarithm (the M0 term) is more son masses are Mπ ∼ 140 MeV, MK ∼ 500 MeV and singular than the QCD chiral logarithm in Eq. (2). As M ∼ 550MeV.IntheQCDcase(N = 3), χPT gives 2 η c Nc becomes large, M0 scales as 1/Nc while the other [10] quenched low energy constants scale the same√ way as the corresponding unquenched ones, f Q = O( N ), 1 M2 π c M2 = 2m B 1 + M2 log π BQ and KQ = O(1). Thus as N →∞, π u 2 π 2 3 c (4πfπ ) µ 2 2 Q → Q + Q + 3 1 M Mπ 2muB 1 2muK3 O mq . (5) − M2 log η 3 η µ2 SinceQCDandQQCDhavethesamelargeN c 2mu + ms limits, Eqs. (3) and (5) imply + 2muK3 + K4 3 = Q = Q →∞ B B ,K3 K3 as Nc . (6) + 3 O mq , (2) Thus some low energy constants (to be more precise, those that are leading in the 1/N expansion) of χPT where mq is the mass of the flavor q(= u, d, s) c and QχPT are identical in the large N limit. This quark and we have used md = mu. The pion decay c f = implies their differences are higher order in 1/Nc: constant π 93 MeV and the renormalization scale µ dependence is absorbed by the low energy constants 1 BQ = B 1 + O , (or counterterms) K3 and K4 defined in [10]. When √ Nc N becomes large, f = O( N ), K = O(1/N ), c π c 4 c 1 while B, quark masses and K are O(1). Thus as Q = + O 3 K3 K3 1 . (7) Nc Nc →∞, only the B and K3 terms contribute. Now we can have a clear idea about how large 2 → + + 3 Mπ 2muB(1 2muK3) O mq . (3) the quenching error is. The error could come from the quenched logarithms which might be formally Note that the µ dependence of K is O(1/N ),soK 3 c 3 higher order in the 1/N expansion but numerically becomes µ independent as N →∞. In the real world, c c significant. Fortunately, those chiral logarithms can the chiral logarithms could be numerically enhanced be calculated using effective field theories so that even though they are formally higher order in 1/N . c one can just replace the quenched logarithms by the In the QQCD (N = 3) case, QχPT gives a similar c real ones. Another source of error comes from the result for the pion mass [2] difference between the low energy constants which is Q2 an O(1/Nc) ∼ 30% effect. This is consistent with the 2 2 M Q = Q − 2 π ∼ 25% quenching error seen in simulations (see [4], Mπ 2muB 1 M0 log 3(4πf Q)2 µ2 π for example). Recently kaon weak matrix elements relevant to + Q +··· 2muK3 , (4) the K → ππ process have been simulated using the 186 J.-W. Chen / Physics Letters B 543 (2002) 183–188 quenched approximation by two groups [5,6]. In both the second term which is the kaon cloud contribution is calculations, QχPT low energy constants are extracted ∼ 2.0 n.m. This combined with the experimental data → → 0 from K π and K vacuum amplitudes in order implies the leading order low energy constant µs is to recover the desired K → ππ amplitude. Based ∼−2.0 n.m. Thus a quenched calculation would have on the above discussion that leads to Eq. (7), the a quenching error of ∼ 0.8n.m.Again,thelargecan- errors of these analyses are O(1/Nc) and naively cellation makes the quenching error relatively big and, are ∼ 30%. Indeed this is the case for the I = hence, the result is not reliable within a 100% error. 1/2 rule. The simulations give 25.3 ± 1.8[5]and The second quantity that is also interesting but even 10 ± 4 [6] compared to the experimental value 22.2. less well measured is the longest range parity violating However, for the case of CP violation / parameter, (PV) isovector pion nucleon coupling, hπ .Formany Ref. [5] obtained (−4.0 ± 2.3) × 10−4, consistent years, serious attempts have been made to measure with Ref. [6], compared to the current experimental this quantity in many-body systems where the PV ef- average of (17.2±1.8)×10−4.Inthe/ case, a large fects are enhanced. However, currently even its order accidental cancellation happens between the I = 0and of magnitude is still not well constrained (see [17] for I = 2 contributions such that the estimated ∼ 30% reviews) largely due to the theoretical uncertainties. errors from the two pieces add up to ∼ 8 × 10−4. Few-nucleon experiments, such as the new np → dγ Besides, in the / quenched calculation, there is experiment at LANSCE [18], are expected to have an additional quenched artifact that can be fixed by small theoretical uncertainties and will set a tight con- dropping some “eye diagrams” [13]. Doing this will straint on hπ . Also, recent studies in single nucleon increase the / by ∼ 4 × 10−4 [14]. Thus in this case systems in Compton scattering [19] and pion produc- the quenching error is significant. tions [20] show that the near threshold γp → π+n The above examples suggest that in principle, the process is both a theoretically clean and experimen- quenched approximation is typically expected to give tally feasible way to extract hπ [20]. Given the in- a ∼ 30% error result if there are no big cancellations tensive interests and the level of difficulty involved in between different contributions and if the quenched the experiments, a lattice QCD determination of hπ chiral logarithms (or non-analytic contributions, in with a 100% error will be very valuable. In fact, is- general) are replaced by the unquenched ones. Now sues such as chiral extrapolation has been addressed we will try to identify some matrix elements that can in [21] in the context of partially quenched approx- be reliably computed in the quenched approximation imations [22] in which sea and valence quark masses with 100% errors and are still worth computing even are different. Since one does not expect large cancella- with that size of uncertainties. tions between the first few orders—as what happened The first quantity we look at is the proton strange in the µs case—we suggest that a quenched simulation magnetic moment µs. It is measured in the electron– of hπ is enough to give a 100% error result. proton and electron–deuteron parity violation exper- Another interesting PV quantity is the longest range iments [15]. The measured value has a large uncer- PV photon–nucleon–delta (γN∆) coupling. Zhu et al. tainty suggested a model, which was inspired by the hyperon decays, to argue that there could be enhanced PV µexp = 0.01 ± 0.29(stat) ± 0.31(sys) s mixing in the initial and final state wave functions ± 0.07(theor.) n.m. (8) such that the PV γN∆coupling is 25–100 times larger This implies that the strange quark only contributes than its size from the naive dimensional analysis [23]. around −0.1±5.1% of the proton’s magnetic moment. If correct, this enhancement will introduce an easily measurable PV signal in γp → π+n at the delta peak. In χPT, the first two orders in chiral expansion for µs is [16] To check if the large PV enhancement due to the wave function mixing really takes place, one can compute M M + − µ = µ0 + N K 5D2 − 6DF + 9F 2 , (9) 3 ; 3 s s 2 the PV mixing between ∆( 2 ) and N(1520 2 ) 24πfπ + − and between the nucleon ( 1 ) and N(1535; 1 ) in where D and F are Goldstone boson–nucleon cou- 2 2 QQCD. One expects the mixing between these states plings. Using the standard values for the parameters, J.-W. Chen / Physics Letters B 543 (2002) 183–188 187 to be 5–20 times larger than their sizes from the References naive dimensional analysis according to the model of Zhu et al. This can be checked easily by a quenched [1] J. Pumplin, et al., hep-ph/0201195; calculation with a 100% error. A.D. Martin, R.G. Roberts, W.J. Stirling, R.S. Thorne, Eur. Recently, Younget al. [24] studied the lattice results Phys. J. C 23 (2002) 73; A.D. Martin, R.G. Roberts, W.J. Stirling, R.S. Thorne, Phys. of nucleon and delta masses. They suggested the dif- Lett. B 531 (2002) 216. ference between the quenched and unquenched cases [2] S.R. Sharpe, Nucl. Phys. B (Proc. Suppl.) 17 (1990) 146; is largely due to the one-loop self-energy contribu- S.R. Sharpe, Phys. Rev. D 46 (1992) 3146; tions calculated in a certain cut-off scheme, while the C. Bernard, M.F.L. Golterman, Phys. Rev. D 46 (1992) 853; difference due to the low energy constants seems to J.N. Labrenz, S.R. Sharpe, Phys. Rev. D 54 (1996) 4595; ∼ M.J. Savage, Nucl. Phys. A 700 (2002) 359. be smaller than a generic estimation of 30%. This [3] S.R. Sharpe, Y. Zhang, Phys. Rev. D 53 (1996) 5125. might be because in the real simulations, only the mass [4] C. Bernard, et al., Nucl. Phys. (Proc. Suppl.) 106 (2002) 412; ratios are computed. The absolute mass scale is set by S.M. Ryan, Nucl. Phys. (Proc. Suppl.) 106 (2002) 86. assuming zero quenching error for a certain quantity, [5] RBC Collaboration, T. Blum, et al., hep-lat/0110075. then the coupling constant and, equivalently, the lattice [6] CP-PACS Collaboration, J.-I. Noaki, et al., hep-lat/0108013. [7] G. ’t Hooft, Nucl. Phys. B 72 (1974) 461. spacing, is determined accordingly. It is conceivable [8] R.F. Dashen, E. Jenkins, A.V.Manohar, Phys. Rev. D 49 (1994) that the ratio of certain quantities can be computed 4713; more reliably because the leading O(1/Nc) errors can- R.F. Dashen, E. Jenkins, A.V.Manohar, Phys. Rev. D 51 (1995) cel. An explicit example is the ratio of B meson decay 2489, Erratum. [9] S. Weinberg, Phys. Rev. Lett. 18 (1967) 188; constants fB /fB ,whereBs (B) denotes a low-lying s R. Dashen, M. Weinstein, Phys. Rev. 183 (1969) 1261; pseudoscalar meson with the same quantum number L.-F. Li, H. Pagels, Phys. Rev. Lett. 26 (1971) 1204; as a bs(bd) state [3]. However, to argue that all the P. Langacker, H. Pagels, Phys. Rev. D 8 (1973) 4595; mass ratios have smaller than O(1/Nc) quenching er- S. Weinberg, Physica A 96 (1979) 327; rors is beyond the scape of effective field theory. To S. Coleman, J. Wess, B. Zumino, Phys. Rev. 177 (1969) 2239; see if certain error suppression really takes place, it C. Callan, S. Coleman, J. Wess, B. Zumino, Phys. Rev. 177 (1969) 2247. would be very interesting to study the Nc dependence [10] J. Gasser, H. Leutwyler, Nucl. Phys. B 250 (1985) 465; for quenched approximations by direct simulations. J. Gasser, H. Leutwyler, Ann. Phys. 158 (1984) 142. There is another class of ratios in large Nc QCD [11] E. Jenkins, A.V. Manohar, Phys. Lett. B 255 (1991) 558. O 2 [12] For a recent review see U.-G. Meißner, hep-ph/0007092, to that the finite Nc correction is (1/Nc ) to the leading order. It would be interesting to see whether the appear in: M. Shifman (Ed.), Encyclopedia of Analytic QCD, O 2 World Scientific. correction is still (1/Nc ) in QQCD. [13] M. Golterman, E. Pallante, Nucl. Phys. (Proc. Suppl.) 106 In summary, quenched QCD and QQCD have the (2002) 335. same large Nc limit, such that some of the low [14] T. Blum, private conversation. energy constants in their corresponding effective field [15] SAMPLE Collaboration, R. Hasty, et al., Science 290 (2000) 2117. theories are identical in the large Nc limit. This implies [16] T.R. Hemmert, U.-G. Meissner, S. Steininger, Phys. Lett. that some physical quantities can be determined in B 437 (1998) 184. the quenched approximation with O(1/Nc) ∼ 30% [17] E.G. Adelberger, W.C. Haxton, Annu. Rev. Nucl. Part. Sci. 35 accuracy. A few interesting applications are explored. (1985) 501; W. Haeberli, B.R. Holstein, in: W.C. Haxton, E.M. Henley (Eds.), Symmetries and Fundamental Interactions in Nuclei, World Scientific, Singapore, 1995, p. 17; W.T.H. van Oers, Int. J. Mod. Phys. E 8 (1999) 417. Acknowledgements [18] W.M. Snow, et al., Nucl. Instrum. Methods A 440 (2000) 729. [19] P.F. Bedaque, M.J. Savage, Phys. Rev. C 62 (2000) 018501; J.W. Chen, T.D. Cohen, C.W. Kao, Phys. Rev. C 64 (2001) The author thanks Tom Cohen, Xiangdong Ji, Steve 055206. [20] J.W. Chen, X. Ji, Phys. Rev. Lett. 86 (2001) 4239; Sharpe and Martin Savage for useful discussions. This J.W. Chen, X. Ji, Phys. Lett. B 501 (2001) 209. work is supported in part by the US Department of [21] S.R. Beane, M.J. Savage, nucl-th/0204046. Energy under grant No. DE-FG02-93ER-40762. [22] C.W. Bernard, M.F.L. Golterman, Phys. Rev. D 49 (1994) 486; 188 J.-W. Chen / Physics Letters B 543 (2002) 183–188

S.R. Sharpe, Phys. Rev. D 56 (1997) 7052; S.R. Sharpe, N. Shoresh, Phys. Rev. D 64 (2001) 114510. M.F.L. Golterman, K.-C. Leung, Phys. Rev. D 57 (1998) 5703; [23] S.L. Zhu, C.M. Maekawa, B.R. Holstein, M.J. Ramsey-Musolf, S.R. Sharpe, N. Shoresh, Phys. Rev. D 62 (2000) 094503; Phys. Rev. Lett. 87 (2001) 201802. S.R. Sharpe, N. Shoresh, Int. J. Mod. Phys. A 16S1C (2001) [24] R.D. Young, D.B. Leinweber, A.W. Thomas, S.V. Wright, hep- 1219; lat/0111041.