DOE/ER/40427-28-N94 November 29, 1994 COLOR TRANSPARENCY EFFECTS IN ELECTRON DEUTERON INTERACTIONS AT INTERMEDIATE Q2 5 9 L. L. Frankfurt 9 1 School of Physics and Astronomy, Tel Aviv University, 69978, Israel n a Institute for Nuclear Physics, St. Petersburg, Russia J 6 W. R. Greenberg 1 Department of Physics, University of Pennsylvania, Philadelphia, PA 19104 1 v 9 G. A. Miller 0 0 Department of Physics, University of Washington, Seattle, WA 98195 1 0 M. M. Sargsyan 5 9 / School of Physics and Astronomy, Tel Aviv University, 69978, Israel, h t - Yerevan Physics Institute, 375036, Armenia l c u and n : v M. I. Strikman i X Department of Physics, Pennsylvania State University, University Park, PA 16802 r a Institute for Nuclear Physics, St. Petersburg, Russia Abstract High momentum transfer electrodisintegration of polarized and unpolarized deuterium targets, d(e,ep)n is studied. We show that the importance of final ′ state interactions FSI, occuring when a knocked out nucleon interacts with the other nucleon, depends strongly on the momentum p~ of the spectator nucleon. In n 1 particular, these FSI occur when the essential contributions to the scattering am- plitude arise from internucleon distances 1.5 fm. But the absorption of the high ∼ momentum γ may produce a point like configuration, which evolves with time. In ∗ this case, the final state interactions probe the point like configuration at the early stage of its evolution. If the point like configuration is still small after propagating about 1.5 fm, the FSI are suppressed. The result is that significant color trans- parency effects, which can either enhance or suppress computed cross sections, are predicted to occur for 4GeV2 Q2 10 (GeV/c)2. We suggest searching for ∼ ≥ ≤ color transparency phenomenon by examining ratios of experimentally measured quantities. Possibletheoretical uncertaintiesofthecalculations,includingthosedue to the deuteron wave function and relativistic effects, are found to be small. PACS: 25.30.-c; 25.30.Fj; 25.45.-z 1 Introduction Color transparency (CT) and color coherent effects have been recently under intense experimental and theoretical investigation. The (p,pp) experiment of Carroll et al.[1] found indications of color transparency while the NE18 (e,e’p) experiment [2] found no such indications. The appearance of color transparency depends on formation of a point-like configuration (PLC) by hard scattering. The Q2 of the NE18 experiment (1 Q2 7GeV2) seem to be large enough to form a small color singlet object. The reason ≤ ≤ why significant colortransparency is notobserved istherapidexpansion ofPLCtonearly normal size (a nearly normal absorption) at the relatively low momenta of the ejected protons[3, 4]. Thus models of color transparency which reproduce the (p,2p) data and include expansion effects predicted small CT effects for the NE-18 kinematics, consistent with their findings, see the discussion in Ref.[5]. Therefore a new strategy of using light nuclei and double scattering as a means to investigate whether small objects are produced at intermediate Q2 was suggested in Ref.[6]. The idea was that using the lightest nuclei would allow the suppression of 2 expansion effects, while employing kinematical cuts to suppress the plane wave Born contributions and to enhance the contribution of double scattering interactions would increase the deviations from the Glauber model. That work concerned He nuclei target, in the present we use an improved formalism to investigate the use of a deuteron target. The deuteron is the best understood nuclear system, with a wave function determined experimentally in a wide momentum range [7]. Therefore finding a method to use the deuteron to investigate color coherent -color transparency effects is both challenging and potentially rewarding. The average separation between the nucleons in the deuteron is large, so the deviations from the plane wave impulse approximation (PWIA) computed within the ordinary Glauber approximation are usually very small, much smaller than for a He target. However a recent analysis of hard quasielastic d(p,2p) reactions within the framework ofconventionalGlauberapproximation[8]hasfoundsubstantialGlauberscreening effects for transverse spectator momenta less than 300 MeV/c and small longitudinal spectator momentum. In particular, the interference between the Born term and rescattering amplitudeinwhichtheknockedoutnucleoninteractswiththespectatornucleon, causesa reductioninthecrosssection. Thisiscalledascreening effect. Including thesquareofthe rescattering amplitude, which we call the double scattering term, increases the computed cross section. Color transparency leads to a suppression of the rescattering amplitude, predicting specific upward and downward changes in the cross section, depending on the kinematics which control the relative sizes of the screening and double scattering terms. Thus observing such effects which occur for different values of the spectator momenta seems to be an effective method to search for color coherent effects using the deuteron. Substantial screening has been observed also in the recent calculation [9] of the pro- cesses d(e,ep) within the nonrelativistic Glauber approximation. The present paper uses ′ formulae of the Glauber approximation deduced from the Feynman diagrams. As a re- sult the predicted angular dependence of the cross section at higher spectator momenta is noticeably different from that in Ref.[9]. In particular, the position in the spectator 3 angular distribution, corresponding to the maximal final state interaction depends on the momentum of the spectator nucleon. At Q2 1(GeV/c)2 our calculations agree ≈ well with calculations of Ref.[10, 11] in the kinematics where the double scattering term dominates. According to Ref.[10] in this kinematics the contribution of meson exchange currents and isobars in the intermediate is small. This is another example that the Glauber approximation is a good starting point in the hunt for the CT phenomenon. We stress that exclusive processes of high momentum transfer electrodisintegration of the polarized and unpolarized deuteron - d(e,ep)n have an important advantage. ′ One may choose special kinematical conditions for the spectator nucleon to control the essential internucleon distances ( 1.5 fm) for final state interactions (FSI). Therefore it ∼ is possible to probe the PLC at the early stage of its evolution, where expansion effects are minimal. The chosen kinematics also allow the separate investigation of the different implications of CT for the screening and double scattering terms. Another limitation on the kinematics is the requirement of reliability of the nonrelativistic description of deuteron and smallness of competing nuclear effects which might diminish CT effects. The major price for fixing kinematics is the drop in the cross section, which should not cause a problem for high intensity accelerators. The paper is organized as follows. In section 2 we calculate the cross section of the d(e,ep)n reaction within the framework of the conventional Glauber approximation[12]. ′ Polarized and unpolarized targets are studied. In section 3 color transparency effects in the d(e,ep)n reaction are calculated within ′ the framework of quantum diffusion model[3] and three resonance model[13]. The three resonance model is solved in two ways: the Green function method of Ref.[6] and the modified Glauber approximation method, used previously to calculate nuclear vector meson production (see Refs.[14, 15, 16, 17]). Section 4 contains the numerical results. We first use the Glauber approximation to investigate the kinematic requirements necessary to ensure the most favorable conditions for studying color coherent effects. This analysis allows us to introduce new experimental 4 observables which are free from theoretical normalizations and are more sensitive to the effects of final state interactions. Using the models, discussed in section 3, we calculate color transparency effects for the kinematic range accessible to the present generation of electron accelerators. Section 5 considers theoretical uncertainties and the reliability of obtained results. The main results of present paper are summarized in Section 6. 2 Glauber Approximation (GA) We consider the d(e,ep)n reaction for kinematics in which the proton carries almost all ′ of the momentum of of the virtual photon and the neutron momentum p is in the range n 200-400 MeV/c where deuteron wave functions have been determined experimentally. 1 The amplitude for d(e,ep)n processes can be expressed as the matrix element of the ′ scattering operator - Tˆ between the deuteron wave function and the wave function of the final state consisting of two outgoing nucleons: = p~ ,s ;~p ;s Tˆ(q) d,~s d3r d3r . (1) p p n n p n M h | | i Z Z Here d,~s is the position space deuteron wave function, p~ , ~r , s and ~p , ~r , s are the p p p n n n | i momenta, coordinate and spin of proton and neutron in the final state. Within the conventional assumptions on factorization of hard - electromagnetic and soft - FSI amplitudes i.e. within the Glauber approximation the nuclear scattering op- erator has the coordinate space form: Tˆ(q,r ,r ) = T (p ,p ,r ,r ) Tem(Q2) eiq~~rp, (2) p n S p n p n · H · · where Tem(Q2) - is the one-body electromagnetic current operator, ~q is the space com- H ponent of the photon four-momentum which we choose in the z-direction. We note that 1Obviously, kinematics in which the neutron is produced forward and proton sidewise is as good for our purposes. In fact, such a setup may have certain experimental advantages since in this case one detects neutron with the same momentum for different Q2. We thank B.Mecking for emphasizing this point. 5 factorization is violated when off-shell effects are taken into account. So we will focus on CT effects at small spectator nucleon momenta or in the rescattering region where only small momenta in the loop integral over internal nucleon momenta are important. We neglect the contributions of meson currents like that at Fig.1c since such contributions rapidly decrease with Q2 for x 1 (cf. discussion in [18] and sec. 5.5 bellow). T (p ,p ) S p n ∼ describes soft FSI of knocked-out proton with spectator neutron. We are studying situ- ations in which the momentum of knocked - out proton is larger than 1 GeV/c. So T S can be calculated within the Glauber approximation [12, 19, 20] 2: T (p ,p ,r ,r ) = 1 ΓN(b b ) Θ(z z ) e i∆0(zp zn), (3) S p n p n p n n p − − − − · − · ~ where ~r ~r(z,b). Here the z direction is defined by the direction of the struck proton ≡ momentum p~ . However within the considered kinematics, where proton carries almost p all momentum of virtual photon, we choose for z the direction of ~q. The profile function ΓN(b) is expressed via the nucleon -nucleon scattering amplitude (fNN) as follows: 1 d2k ΓN(b) = exp(i~k ~b) fNN(~k ) t , (4) t t 2i · · (2π)2 Z where the NN scattering amplitude normalized as ImfNN(k = 0) = σ . The addi- t tot tional factor e i∆0(zp zn) in eq.(3), where − − m +q ∆0 = (E m) d 0, (5) n − · ~q | | is a kinematical factor. In eq.(5) E - is the spectator energy, m is the nucleon mass, n m - is the mass of deuteron and q is the transfered energy. The factor ∆0 is obtained d 0 by evaluating Feynman diagrams (cf. section 5.4, eq.(58)) and accounts for the nucleon recoil usually neglected within the conventional non-relativistic Glauber approximation. This factor accounts for the fact that the variable k0 kz but not k (fig.1b) is conserved m− z in two-body high energy collisions (cf. [21]). 2Tosimplifythe discussionweneglectsmallcorrectionsduetothechargeexchangeprocessnp pn. → Since the charge exchange amplitude is predominantly real and of spin flip character (which itself is small), it practically does not interfere with the main amplitude. 6 Insertingeqs.(3)and(4)intoeq.(1),weobtaintheamplitudeforthed(e,ep)nreaction ′ as: = Tem(Q) d3r d3r eiq~~rp ~r ,s ;~r ,s d,~s M H p n · h p p n n | i Z Z 1 Γ(b b )e i∆0(zp zn)Θ(z z ) e ip~p~rpe ip~n~rn p n − − n p − · − · × − − − h i = Tem(Q) d3r d3r e ip~i~rpe ip~n~rn ~r ,s ;~r ,s d,~s H p n − · − · h p p n n | i Z Z Tem(Q) d3r d3r e ip~i~rpe ip~n~rn ~r ,s ; ~r ,s d,~s − H p n − · − · h p p · n n | i Z zZp 1 d2k ei~kt(~bp ~bn) fNN(~k ) t e i∆0(zp zn)Θ(z z ), (6) ×2i · − · t (2π)2 − − n − p Z where we define the momentum of initial proton in the deuteron state as ~p = p~ ~q. i p − To separate center-of-mass motion we introduce the center-of-mass (R ) and relative cm distance (r) coordinates: 1 ~r = R~ + ~r p cm 2 1 ~r = R~ ~r. (7) n cm − 2 Using Eq.(7) allows us to simplify Eq.(6) as: M = Tpem(Q) d3RcmeiR~cm·(p~d−p~i−p~n) d3rh~rp,sp;~rn,sn | d,~sie−2ip~i·~re2ip~n·~r Z (cid:26)Z 1 d2k d3rΘ( z) ~rp,sp;~rn,sn d,~s e−2ip~i·~re2ip~n~r ei~kt·~b−i∆0z fNN(~kt) t − − h | i· 2i · (2π)2) Z Z = Tem(Q) d3r ~r ,s ;~r ,s d,~s e ip~i~r p h p p n n | i − (cid:26)Z 1 d2k d3rΘ( z) ~r ,s ;~r ,s d,~s e i(p~it ~kt)~b i(piz+∆0)zfNN(~k ) t ,(8) p p n n − − · − · t − 2i − h | i· (2π)2) Z Z The integration over R ensures the momentum conservation: ~p = p~ = ~q p~ . In cm n i p − − the last part of eq.(8) we omit the factor (2π)3δ(p~ p~ p~ ), which will be included d p n − − in the definition of cross section. (To simplify formulae we use the Laboratory frame, in which the momentum of of the deuteron, p~ is zero.) d The integration over the relative coordinate r leads to the result: = (2π)32 Tem(Q) p ,s ,s d,~s 1 p′,s ,s d,~s fNN(~k ) d2kt , (9) M · p (h i p n | i− 4i h i p n | i t (2π)2) Z 7 where p~′i ≡ (piz + ∆0,pit − kt) and hp′i,sp,sn | d,~si - is the deuteron wave function in momentum space representation. The additional factor 1 in the second term is a 2 consequence of Θ( z) functions in eq.(8). (The accuracy of last replacement is discussed − in sec.5.4). Taking the square of the modulus of the amplitude and summing over the proton and neutron polarizations we obtain: 2 † || M || ≡ MM sp,sXn,s′p,s′n 1 d2k = (2π)3 Fe.m.(Q2) 2 ρ~s(p ,p ) e t ρ~s(p ,p )fNN (k ) ·| p | ·( d i i −R 2i (2π)2 d ′i i ∗ t Z 1 d2k d2k + t1 t2ρ~s(p ,p )fNN(k )fNN (k ) , (10) 16 (2π)4 d i1 i2 t1 ∗ t2 ) Z Z where p~ (p +∆0,p k ), p~ (p +∆0,p k ). i1 iz it t1 i2 iz it t2 ≡ − ≡ − In eq.(10) we introduced the deuteron density function: ρ~s(k ,k ) = k ,s ,s d,~s d,~s k ,s′,s′ . (11) 1 2 h 2 p n | ih | 1 p ni sp,sX′p,sn,s′n In this Glauber or distorted wave impulse approximation (DWIA) the d(e,ep)n cross ′ section can be expressed as follows: dσ = σ D (q,p ,p ) δ(q M E E ). (12) ep d p n o d p n dE dΩ d3p · · − − − e e p ′ ′ Here σ is the cross section of the electron scattering off a bound proton (up to the en flux and proton recoil factor) 3 and the decay function D (q,p ,p ) represents the joint d p n probability for the initial proton in the deuteron having Fermi momentum - p and for i the final state having a proton and neutron with a momentum p and p (for more details p n see [22]). 3In principle, one should use the light-cone quantum mechanics of the deuteron [21] to calculate the cross section. However for the effects dominated by the contribution of small nucleon momenta in the deuteronof interesthere, the difference betweenpredictions of light-cone andnonrelativistic formalisms is small (see Sec.5.1). 8 If the deuteron is polarized, the cross section depends on D~s(q,p ,p ) with: d p n 1 d2k D~s(q,p ,p ) = ρ~s(p ,p ) e ρ~s(p ,p ) fNN(k ) t d p n d i i −R 2i d i ′i · t · (2π)2 Z 1 d2k d2k + ρ~s(p ,p ) fNN(k ) fNN (k ) t1 t2. (13) 16 d i1 i2 · t1 · ∗ t2 · (2π)2(2π)2 Z The polarized density matrices of deuteron can be expressed via the s and d - wave components introducing the polarization vector ~a according to Ref.[23]: 3 k a 2 u(k )w(k ) 3 k a 2 u(k )w(k ) ρ~a(k ,k ) = u(k )u(k )+ 1 | 2 · | 1 2 + 1 | 1 · | 2 1 d 1 2 1 2 " − k22 # √2 " − k12 # √2 9(k a)(k a) (k k ) 3 k a 2 3 k a 2 1 1 2 ∗ 1 2 1 2 + · · · | · | | · | + w(k )w(k ), 2 k12k22 − 2 k12 − 2 k22 2! 1 2 (14) where the components of ~a are defined through the deuteron spin wave function as: i i ψ10 = i a , ψ11 = (a +ia ), ψ1 1 = (a ia ), (15) · z −√2 x y − √2 x − y where ψ1µ is theprojection of the deuteron’s spin onthe the µ direction. The unpolarized deuteron density matrix follows from Eq.(14) as: 1 3(~k ~k )2 1 ρ (k ,k ) = ρa(k ,k ) = u(k )u(k )+w(k )w(k ) 1 · 2 . (16) d 1 2 3 d 1 2 1 2 1 2 ·2 k2k2 − 2 a 1 2 X   ′ 3 Coherent Effects in d(e, e p)n Scattering Theoretical analysis shows that in realistic models the absorption of a hard photon leads to the formation of a point like configuration, which undergoes a reduced interaction with other hadrons, because of its small size and its color neutrality [24]. To estimate the expected effects of color coherence we consider two different models which account for the formation of the PLC and their evolution to the normal hadronic state: quantum diffusion model of Ref.[3] and the three state model of Ref.[13]. 9 3.1 Quantum Diffusion model (QDM) The reduced interaction between the PLC and the spectator neutron can be accounted for by introducing the dependence of the scattering amplitude on the transverse size of PLC.However weconsider energiesthatarefarfromasymptotic, sotheexpansionofPLC should be important. This feature is included by allowing the rescattering amplitude to depend on the distance from the photon absorption point. Including these effects leads to the modified deuteron decay function DS(q,p ,p ): d p n D~s(q,p ,p ) = ρs(u(p ),w(p ),u(p ),w(p )) d p n d i i i i 1 d2k e ρ~s(u(p ),w(p ),u˜(p ) ,w˜(p ) )fNN(k ) t −R i d i i ′i ∗ ′i ∗ t (2π)2 Z 1 d2k d2k + ρs(u˜(p ),w˜(p ),u˜(p ),w˜(p )) t1fNN(k )fNN (k ) t2, 4 d i1 i2 ′i ′i (2π)2 t1 ∗ t2 (2π)2 Z (17) with the same form of ρ~s as in Eq.(14) but with modified radial wave functions: d 1 fPLC,N(z,k ,Q2) u˜(k) = u(r)Y (rˆ) t Θ( z)e i~k~rd3r (2π)23 Z o · fNN(kt) − − · 1 fPLC,N(z,k ,Q2) w˜(k) = w(r)Y (rˆ) t Θ( z)e i~k~rd3r. (18) (2π)23 Z 20 · fNN(kt) − − · Eq.(17) is derived in the same way as the Glauber formulae in the previous section. The onlydifference is toinclude thecoordinatedependence ofthe rescattering amplitude. To calculate the deuteron decay function in the quantum diffusion model we model the amplitude of the PLC N scattering in a form[6] consistent with the optical theorem: − G (t σ (z,Q2)/σ ) fPLC,N(z,k ,Q2) = iσ (z,Q2) e2bt N · tot tot , (19) t tot · · G (t) N where b/2 is the slope of elastic NN amplitude, G (t) ( (1 t/0.71)2) is the Sachs N ≈ − form factor and t = k2. The last factor in eq.(19) accounts for the difference between − t elastic scattering of PLC and average configurations, using the observation that the t dependence of dσh+N h+N/dt is roughly that of G2(t) G2 (t) for not very large values → ∼ h · N of t and that G2(t) exp(R2t/3). h ≈ h 10