and the Nucleon Strange Electric Form Factor

  • Published on

  • View

  • Download

Embed Size (px)


  • Isospin Mixing in the Nucleon and 4He and the Nucleon Strange Electric Form Factor

    M. Viviani,1 R. Schiavilla,2,3 B. Kubis,4 R. Lewis,5 L. Girlanda,1 A. Kievsky,1 L. E. Marcucci,1 and S. Rosati11INFN, Sezione di Pisa, and Department of Physics, University of Pisa, I-56127 Pisa, Italy

    2Jefferson Lab, Newport News, Virginia 23606, USA3Department of Physics, Old Dominion University, Norfolk, Virginia 23529, USA4HISKP (Theorie), Universitat Bonn, Nussallee 14-16, D-53115 Bonn, Germany

    5Department of Physics, University of Regina, Regina, Saskatchewan, Canada, S4S 0A2(Received 15 March 2007; published 14 September 2007)

    In order to isolate the contribution of the nucleon strange electric form factor to the parity-violatingasymmetry measured in 4He ~e; e04He experiments, it is crucial to have a reliable estimate of the mag-nitude of isospin-symmetry-breaking (ISB) corrections in both the nucleon and 4He. We examine thisissue in the present Letter. Isospin admixtures in the nucleon are determined in chiral perturbation theory,while those in 4He are derived from nuclear interactions, including explicit ISB terms. A careful analysisof the model dependence in the resulting predictions for the nucleon and nuclear ISB contributions to theasymmetry is carried out. We conclude that, at the low momentum transfers of interest in recent mea-surements reported by the HAPPEX Collaboration at Jefferson Lab, these contributions are of comparablemagnitude to those associated with strangeness components in the nucleon electric form factor.

    DOI: 10.1103/PhysRevLett.99.112002 PACS numbers: 14.20.Dh, 12.15.Ji, 25.30.Bf

    One of the challenges of modern hadronic physics is todetermine, at a quantitative level, the role that quark-antiquark pairs, and in particular ss pairs, play in thestructure of the nucleon. Parity-violating (PV) electronscattering from nucleons and nuclei offers the opportunityto investigate this issue experimentally. The PVasymmetry(APV) arises from interference between the amplitudes dueto exchange of photons and Z-bosons, which couple, re-spectively, to the electromagnetic (EM) and weak neutral(NC) currents. These currents involve different combina-tions of quark flavors, and therefore measurements of APV,in combination with electromagnetic form factor data forthe nucleon, allow one to isolate, in principle, the electricand magnetic form factors GsE and G

    sM, associated with the

    strange-quark content of the nucleon.Experimental determinations of these form factors have

    been reported recently by the Jefferson Lab HAPPEX [1]and G0 [2] Collaborations, Mainz A4 Collaboration [3],and MIT-Bates SAMPLE Collaboration [4]. These experi-ments have scattered polarized electrons from either un-polarized protons at forward angles [13] or unpolarizedprotons and deuterons at backward angles [4]. The result-ing PVasymmetries are sensitive to different linear combi-nations of GsE and G

    sM as well as the nucleon axial-vector

    form factor GZA. However, no robust evidence has emergedso far for the presence of strange-quark effects in thenucleon.

    Last year, the HAPPEX Collaboration [5,6] at JeffersonLab reported on measurements of the PV asymmetry inelastic electron scattering from 4He at four-momentumtransfers of 0:091 GeV=c2 and 0:077 GeV=c2.Because of the J 0 spin-parity assignments of thisnucleus, transitions induced by magnetic and axial-vectorcurrents are forbidden, and therefore these measurementscan lead to a direct determination of the strangeness elec-

    tric form factor GsE [7,8], provided that isospin-symmetry-breaking (ISB) effects in both the nucleon and 4He, andrelativistic and meson-exchange (collectively denoted withMEC) contributions to the nuclear EM and weak vectorcharge operators, are negligible. A realistic calculation ofthese latter contributions [8] found that they are in fact tinyat low momentum transfers. The goal of the present Letteris to provide a quantitative estimate of ISB corrections tothe PV asymmetry.

    In the following analysis, we only need to consider thetime components of the EM current and vector part of theweak NC currentthe weak vector charge referred toabove [8]. We account for isospin symmetry breaking inboth the nucleon and -particle. We first discuss it in thenucleon.

    Ignoring radiative corrections, the EM and weak vectorcharge operators can be decomposed as j0EM j0 j1and j0NC 4s2Wj0 2 4s2Wj1 js, where j0and j1 are, respectively, the isoscalar and isovector com-ponents of the EM charge operators, js is the (isoscalar)component due to strange-quark contributions, and s2W sin2W contains the Weinberg mixing angle. In terms ofquark fields u, d, s, these charge operators read: j0 u0u d0d 2s0s=6, j1 u0u d0d=2, andjs s0s. In a notation similar to that adopted by theauthors of Ref. [9], we introduce form factors correspond-ing to the following matrix elements of j0 and j1 be-tween proton (p) and neutron (n) states:

    hpjj0jpi ! G0EQ2 G;EQ2; (1)

    hnjj0jni ! G0EQ2 G;EQ2; (2)

    hpjj1jpi ! G1EQ2 G16EQ2; (3)

    PRL 99, 112002 (2007) P H Y S I C A L R E V I E W L E T T E R Sweek ending

    14 SEPTEMBER 2007

    0031-9007=07=99(11)=112002(4) 112002-1 2007 The American Physical Society

  • hnjj1jni ! G1EQ2 G16EQ2; (4)where the arrow indicates that only leading contributionsare listed in the nonrelativistic limit of these matrix ele-ments. While higher-order corrections associated with theDarwin-Foldy and spin-orbit terms are not displayed ex-plicitly in the equations above, they are in fact retained inthe calculations discussed later in the present work. Theform factors G;EQ2 and G16EQ2 parameterize ISB effectsin the nucleon states. We also introduce the strange formfactor via

    hpjjsjpi hnjjsjni ! GsEQ2; (5)where here ISB terms in the p, n states are neglected.Contributions from sea quarks heavier than strange arealso ignored.

    In terms of the experimental proton and neutron electricform factors, derived from the matrix elementshpjj0EM jpi ! GpEQ2 and hnjj0EM jni ! GnEQ2, we ob-tain

    G0E GpE GnE=2G16E; (6)

    G1E GpE GnE=2G;E; (7)where the Q2 dependence in these and the following twoequations is understood. In the limit in which the p, n statesform an isospin doublet, the form factors G;E and G


    vanish, and G0E and G1E reduce to the standard isoscalar

    and isovector combinations of the proton and neutronelectric form factors. The proton and neutron vector NCform factors follow from the expression for j0NC givenearlier, i.e.,

    Gp;ZE 1 4s2WGpE GnE 2G16E G;E GsE; (8)

    Gn;ZE 1 4s2WGnE GpE 2G16E G;E GsE: (9)We now turn to the nuclear charge operator. At low

    momentum transfer, it is simply given by [8]


    eiqrk GnEQ2XA

    kZ1eiqrk ; (10)

    where Z is the number of protons, A Z the number ofneutrons, and q is the three-momentum transfer. An equa-tion similar to Eq. (10) holds for the weak vector chargeoperator, but with GpE and G

    nE being replaced, respectively,

    by Gp;ZE and Gn;ZE . It is also convenient to define the charge


    0q GpE GnE2


    eiqrk ; (11)

    1q GpE GnE2


    eiqrk XA


    ; (12)

    from which

    EMq 0q 1q; (13)

    NCq 4s2WEMq 2G16E GsE

    GpE GnE=20q

    21q 2G;E

    GpE GnE=21q; (14)

    where again the Q2 dependence of the nucleon form factorshas been suppressed here and in the following for brevity.The relations above lead to the definition of the followingnuclear form factors: h4Hejaqj4Hei=Z Faq witha EM, NC, 0, 1, having the normalizations FEM0 F00 1 and F10 0. The form factor F1q isvery small because 4He is predominantly an isoscalar state.Using standard techniques, it is possible to show thatAPV FNCq=FEMq (see Ref. [7], for example).Thus, ignoring second order terms like G;F1q, weobtain for the PV asymmetry measured in ( ~e, e0) elasticscattering from 4He:

    APV GQ



    p4s2W 2


    2G16E GsEGpE GnE=2

    ; (15)

    where G is the Fermi constant as determined from muondecays, and here s2W is taken to incorporate radiative cor-rections. The terms G16E and F

    1q=F0q are the contri-butions to APV, associated with the violation of isospinsymmetry at the nucleon and nuclear level, respectively.

    The most accurate measurement of the PV asymmetry,recently reported in Ref. [6] at Q2 0:077 GeV=c2,gives APV 6:40 0:23stat 0:12syst ppm,from which, after inserting the values for G 1:166 37 105 GeV2, 1=137:036, and s2W 0:2286 (including its radiative corrections [7]) inEq. (15), one obtains


    F0q 2G16E GsE

    GpE GnE=2 0:010 0:038 (16)

    at Q20:077 GeV=c2. This result is consistent with zero.In the following, we discuss the estimates for the ISBcorrections first in the nucleon and then in 4He, respec-tively G16EQ2 and F1q, at Q2 0:077 GeV=c2 (cor-responding to q 1:4 fm1).

    For G16EQ2 we use the estimate obtained in Ref. [9]adapted to our conventions, combining a leading-ordercalculation in chiral perturbation theory with estimatesfor low-energy constants using resonance saturation.Collecting the various pieces, we find

    PRL 99, 112002 (2007) P H Y S I C A L R E V I E W L E T T E R Sweek ending

    14 SEPTEMBER 2007


  • G16EQ2g2AmNm



    0Q24 3Q2




    mN 0Q25 3Q2









    ; (17)

    where the loop functions , 0=3 are given explicitly inRef. [9], along with the precise definitions of the variouscoupling constants. The chiral loop contributions inEq. (17) scale with the neutron-proton mass differencem, while the resonance part is proportional to the ! mixing angle !. We refer to Ref. [9] for a detaileddiscussion of the range of numerical values for the vector-meson coupling constants and only show the resulting bandfor G16EQ2 in Fig. 1. At the specific kinematical point ofinterest Q2 0:077 GeV=c2, we find G16EQ2 0:0017 0:0006, and with GpEQ2 0:799 andGnEQ2 0:027 [10], we obtain 2G16E=GpE GnE=2 0:008 0:003 at Q2 0:077 GeV=c2.

    We now turn to the nuclear ISB corrections. An approxi-mate calculation of the ratio F1q=F0q was carriedout more than a decade ago [11], by (i) taking into accountonly the isospin admixtures induced by the Coulomb in-teraction, (ii) constructing a T 1 J 0 breathingmode excitation based on a plausible ansatz, and (iii)generating the relevant T 1 component in the 4Heground state in first order perturbation theory. The calcu-lated value was found to be rather small, and it produced aless than 1% correction with respect to the 4s2W term inEq. (15) at low Q2.

    Since that pioneering study, significant progress hasoccurred on several fronts. First, there now exist a numberof accurate models of nucleon-nucleon (NN) potentials[1216] which include explicit ISB induced by both thestrong and electromagnetic interactions. These ISB termshave been constrained by fitting pp and np elastic scatter-ing data. It is now an established fact that a realistic studyof 4He, and in fact light nuclei [17], requires the inclusionof three-nucleon (NNN) potentials in the Hamiltonian.While these are still not well known, the models mostcommonly used in the literature [1720] do not containISB terms. The strength of the latter, however, is expectedto be tiny.

    Second, several accurate methods have been developedto compute 4He wave functions starting from a givenrealistic nuclear Hamiltonian [21]. In these calculations,T > 0 components are generated nonperturbatively. TheT 1 percentage in the 4He wave function is typicallyfound to be of the order of 0.001%.

    In this Letter, we use the hyperspherical harmonic (HH)expansion method to compute the 4He wave function [2224]. In order to have an estimate of the model dependence,we consider a variety of Hamiltonian models, including(i) the Argonne v18 NN potential [13] (AV18), (ii) theAV18 plus Urbana-IX NNN potential [18] (AV18/UIX),(iii) the CD Bonn [14] NN plus Urbana-IXb NNN poten-tials (CDB/UIXb), and (iv) the chiral N3LO [15] NNpotential (N3LO). The Urbana UIXb NNN potential is aslightly modified version of the Urbana UIX (in the UIXb,the parameter U0 of the central repulsive term has beenreduced by the factor 0.812), designed to reproduce, whenused in combination with the CD Bonn potential, theexperimental binding energy of 3H. The binding energiesB and PT1 percent probabilities obtained with the AV18,AV18/UIX, CDB/UIXb, and N3LO are, respectively,B 24:21; 28:47; 28:30; 25:38 MeV (to be comparedwith an experimental value of 28.30 MeV) and PT1 0:0028; 0:0025; 0:0020; 0:0035. These results are inagreement with those obtained with other methods (for acomparison, see Ref. [23]).

    The form factors F0q and F1q, calculated with theAV18/UIX Hamiltonian model, are displayed in Fig. 2. Thedashed (solid) curves represent the results of calculationsincluding the one-body (one-body plus MEC) EM chargeoperators (note that ISB corrections in the nucleon formfactors entering the two-body EM charge operators, listedexplicitly in Ref. [8], are neglected). Similar results (notshown in Fig. 2 to reduce clutter) are obtained with theother Hamiltonian models. In particular, the model depen-dence in the calculated F0q form factor is found to beweak, although the change of sign in the predictions cor-responding to the N3LO model occurs at a slightly lowervalue of momentum transfer than in those corresponding tothe other models, which are in excellent agreement withthe experimental data from Refs. [25]. From the figure it isevident that for q 1:5 fm1, the effect of MEC in bothF0q and F1q is negligible.

    0.0 0.1 0.2 0.3

    Q2 [(GeV/c)









    (Q2 )

    FIG. 1. The isospin-violating nucleon form factor G16EQ2.The band comprises a range of values for various vector-mesoncoupling constants, as well as an estimate of higher-order chiralcorrections. For details, see Ref. [9].

    PRL 99, 112002 (2007) P H Y S I C A L R E V I E W L E T T E R Sweek ending

    14 SEPTEMBER 2007


  • In the inset of Fig. 2, we show the model dependence ofthe ratio jF1q=F0qj (all calculations include MEC).The various Hamiltonian models give predictions quiteclose to each other, although the value for the N3LO issomewhat larger than for the other models, reflecting thelarger percentage of T 1 admixtures in the 4He groundstate, predicted by the N3LO potential. The calculatedratios F1q=F0q at Q2 0:077 GeV=c2 are of theorder of 0:002. The inclusion of NNN potentials tends toreduce the magnitude of F1=F0, while ignoring MECcontributions, at this value of Q2, would lead, at the most,to 1.5% decrease of this magnitude.

    Note that the value estimated in Ref. [11] wasjF1=F0j 0:0014 at Q2 0:077 GeV=c2, althoughit was computed in first order perturbation theory by onlykeeping the ISB corrections due to the Coulomb potential.However, the latter only account for roughly 50% of thePT1 probability in the 4He ground state [23], and, as-suming the ratio above to scale with


    p, one would

    have expected a smaller value for it than actually ob-tained (0:0014) in Ref. [11]. Therefore, at Q2 0:077 GeV=c2, both contributions F1=F0 and G16E arefound of the same order of magnitude as the central valueof in...