DOI: 10.1140/epja/i2006-09-002-9 Eur. Phys. J. A 28, s01, 7–17 (2006) EPJ A direct electronic only Physics at the Thomas Jefferson National Accelerator Facility L.S. Cardmana Thomas Jefferson National Accelerator Facility, 12000 Jefferson Avenue, Newport News, VA 23606, USA and University of Virginia, Department of Physics, 382 McCormick Rd., P.O. Box 400714, Charlottesville, VA 22904-4714, USA / Published online: 31 May 2006 – c© Società Italiana di Fisica / Springer-Verlag 2006 Abstract. The Continuous Electron Accelerator Facility, CEBAF, located at the Thomas Jefferson Na- tional Accelerator Facility, is devoted to the investigation of the electromagnetic structure of mesons, nucleons, and nuclei using high energy, high duty-cycle electron and photon beams. Selected experimental results of particular interest to the MAMI community are presented. PACS. 29.17.+w Electrostatic, collective, and linear accelerators – 25.20.-x Photonuclear reactions – 25.30.Bf Elastic electron scattering – 25.30.Dh Inelastic electron scattering to specific states 1 Personal Comments It is an honor and a pleasure to be here to celebrate the achievements of MAMI and the distinguished careers of Professors Arenhövel, Backe, Drechsel, Friedrich, Kaiser, and Walcher. We are all deeply aware of the extent to which the science we do builds on the achievements of those who have gone before us, and on the insights and hard work of our colleagues working in the field today. One of my very earliest memories as a scientist, dating from the days when I was a young graduate student, is that of attending “Photonuclear Physics Boot Camp” (otherwise known as the Photonuclear Gordon Conference) and learning from (and with) many of those “retiring” today. Thomas (Walcher) was one of the very first scientists I ever knew beyond the boundaries of my own laboratory. He came to visit us (at Yale), and I went and visited him and his colleagues at Darmstadt. It has been a great plea- sure to follow his distinguished career in science, from low- Q2 electron scattering to hadronic physics at CERN and beyond, and finally to the leadership role he has played at MAMI for many years. Hartmuth (Arenhövel) has been the keeper of the flame of all knowledge about the deuteron, and a worthy succes- sor to Gregory Breit. You should know that I was a grad- uate student at Yale, and it was one of Professor Breit’s missions in life to convince any and all who would listen that the deuteron was the essence of nuclear physics, and that until we understood the deuteron, we did not un- derstand anything. I think it is fair to call Hartmuth the Gregory Breit of my generation; he has made so many contributions. a e-mail: cardman@jlab.org Dieter (Drechsel) has always been one of those peo- ple I have looked to as the source of the “big picture” in nuclear physics. He has provided us with deep insights, a sense of direction, and an understanding of what is really important. He has also been an inspiring example here at Mainz of the tremendous benefits to everyone of having a close collaboration between theory and experiment. Jörg (Friedrich) has taught us all how to analyze and interpret electron scattering data with minimal prejudice (and, therefore, maximal honesty). It is a delight to see the same rigorous approach that was so successful in the study of nuclei and their excited states now being applied to nucleon structure. Karl-Heinz (Kaiser) and his mentor, Helmut Herming- haus, taught the world how to build superb continuous- wave (cw) electron accelerators effectively and efficiently. Karl-Heinz, in particular, through the design and con- struction of the double-sided microtron, is leaving the In- stitute well positioned for another generation of superb experiments. In conclusion, on behalf of so many people I have worked with in nuclear physics, I want to thank each of you for your many contributions to our field, and to express the hope we all share that for each of you “retirement” is a formality, not a reality, and that you will continue to be active for years to come. 2 Research at Jefferson Laboratory The Thomas Jefferson National Accelerator Facility, also called Jefferson Lab (or JLab), operates the Continu- ous Electron Beam Accelerator Facility (CEBAF). CE- BAF is a cw electron accelerator capable of delivering three electron beams for simultaneous experiments in the 8 The European Physical Journal A three experimental areas. Originally designed for 4 GeV, its present maximum energy is 5.7 GeV. The CEBAF user community consists of about 2000 physicists; more than half of them are actively involved in the experimental program. In addition to its main mis- sion, JLab contributes to the development and use of Free Electron Lasers, to medical imaging, and to community outreach programs. The intellectual and technical foundations for the con- struction of CEBAF were provided by the scientific suc- cesses of earlier electron accelerators (the generation that included Saclay, MIT-Bates, NIKHEF, and, to some ex- tent, SLAC), and by the enhanced research opportunities provided by cw electron beams as demonstrated at fa- cilities such as MAMI. CEBAF is a large, international laboratory with a broad research program; it has been in operation for some seven and a half year now. What are the goals of CEBAF’s research program? Ba- sically, we aim to understand strongly-interacting matter. How are the hadrons constructed from the quarks and glu- ons of QCD, and how does the nucleon-nucleon force arise from the strong interaction? We further aim to identify the limits of our understanding of nuclear structure by us- ing the high precision attainable with the electromagnetic probe and the possibiltiy of extending investigations to very small distance scales. A specific issue that motivated the construction of CEBAF was our desire to gain insight into the question of where the description of nuclei based on nucleon and meson degrees-of-freedom fails and the un- derlying quark degrees-of-freedom must be taken into ac- count. One can ultimately characterize all of this as trying to understand QCD, not in the perturbative regime acces- sible at very high energies and very short distance scales, but in the strong interaction regime relevant to most of the visible matter in the Universe. To make progress in these areas, there are other critical issues that must be addressed, such as the mechanism of confinement, the dy- namics of the quark interaction, and how chiral symmetry breaking occurs. To provides some shape and structure to the discussion of the experiments, the CEBAF program can be organized into half a dozen broad thrusts. This presentation will concentrate on two of them: – How are the nucleons made from quarks and glue? – Where are the limits of our understanding of nuclear structure 3 How are the nucleons made from quarks and glue? Among the most interesting puzzles in physics today are: why there is this effective degree-of-freedom in QCD, the nucleon; and how something as complicated as the resid- ual QCD interaction between quarks in nucleons can be characterized by a rather simple N-N potential? To pro- vide experimental insights that will help us solve the first of these puzzles, the Jefferson Lab research community has mounted an array of investigations in three broad areas: Lung Rock Bartel Arnold Jourdan1 Jourdan2 Gao Xu Selected World Data MIT-Bates: 2 H(e → ,e’n → ) Mainz A3: H(e,e’n) NIKHEF: 2 H(e,e’n) Galster Borkowski Sill Bosted Walker Andivahis 0.4 0.6 0.8 1.2 1.4 0.4 0.6 0.8 1.0 1.2 1.4 0 0.02 0.04 0.06 0.08 0.10 G E n G M n µ n G D 0.4 0.6 0.8 1.0 1.2 1.5 1.0 0.5 0.0 µG E p G M p G M p µ p G D 0 2 4 6 0 1 2 Q2(GeV2)Q2 (GeV2) 0 1 2 3 40 1 2 3 40 10 20 30 Q2 (GeV2) Q2 (GeV2) Fig. 1. Nucleon form factor data available before the start of experiments using CEBAF. Top 2 panels: electric form factors, bottom panels: magnetic form factors. Left 2 panels: proton form factors, right panels: neutron form factors (adapted from ref. [1]). – What are the spatial distributions of the u, d, and s quarks in the hadrons? – What is the excited state spectrum of the hadrons, and what does it reveal about the underlying degrees- of-freedom? – What is the QCD basis for the spin structure of the hadrons? 3.1 What are the spatial distributions of the u, d, and s quarks in the hadrons? Elastic electron scattering has provided most of our infor- mation on the spatial distributions of the quarks in the nucleons. The data on the four electromagnetic structure functions of the nucleon, GE and GM for both the proton and the neutron, available just prior to the start of exper- iments at CEBAF is shown in fig. 1. The magnetic form factors of the proton and the neutron were known rea- sonably well, but the electric form factors were not. The electric form factor of the proton had not been determined accurately enough to distinguish between a wide range of theories based on rather different physics. First results on the electric form factor of the neutron were available from Bates, Mainz, and NIKHEF, but these data were limited to moderate momentum transfers and, therefore, not sen- sitive to the details of the distribution of charge inside the neutron. The measured form factor was consistent with the r.m.s. radius derived from neutron-electron scattering. The present status of the nucleon form factors is shown in fig. 2. The measurements of the polarization transfer from the incident electron to the elastically recoiling proton have shown that the electric and mag- netic form factors for the proton differ substantially. The systematic differences between the polarization transfer data and the Rosenbluth results for GE/GM are likely L.S. Cardman: Physics at the Thomas Jefferson National Accelerator Facility 9 0 1 2 3 4 0.4 0.6 0.8 1.2 1.4 Lung Rock Bartel Arnold Jourdan1 Jourdan2 Gao Xu 0 1 2 3 4 0.4 0.6 0.8 1.0 1.2 1.4 this measurement (with statistical errors) grey: estimated systematic error 0 10 20 30 0.4 0.6 0.8 1.0 1.2 1.5 1.0 0.5 0.0 0 2 4 6 Borkowski Sill Bosted Walker Andivahis 0 0.02 0.04 0.06 0.08 0.10 0 1 2 MIT-Bates: 2 H(e → ,e’n → ) JLab E93-026: 2 H → (e → ,e’n) Mainz A1: 3 H → e(e → ,e’n) Mainz A3: 2 H(e → ,e’n → ) Mainz A3: 3 H → e(e → ,e’n) NIKHEF: 2 H → (e → ,e’n) Schiavilla & Sick: G Q JLab E93-038: 2 H(e → ,e’n → ) Galster New Fit µG E p G M p G E n G M p µ p G D G M n µ n G D Q2 (GeV2) Q2(GeV2)Q2 (GeV2) Q2 (GeV2) Fig. 2. Present status of the nucleon form factor data including the CEBAF data (adapted from ref. [1]). due to two-photon exchange effects modifying the results. Theoretical estimates suggest that the modifications are much smaller for the polarization transfer data than for the Rosenbluth data, so the former are likely to be more directly interpretable in terms of the nucleon form factors. The electric form factor of the neutron has now been measured up to a Q2 of 1.5 (GeV/c)2 using polarization transfer techniques, and the data taken with different methods agree quite well. The theoretical interpretation of the data is summa- rized in fig. 3. The theories that describe the data reason- ably well reveal two key aspects of nucleon structure: the importance of the pion cloud, and the importance of in- corporating the relativistic motion of the quarks into the theoretical description of the nucleon. When one looks at these form factors in a phenomeno- logical way with minimum prejudice [2], what emerges is some of the clearest evidence we have for the nucleon’s pion cloud (see fig. 4). Similar results have been obtained using a different approach to model-independent analy- sis [3] of nucleon form factors. We plan to extend the proton form factor data to ∼ 9 (GeV/c)2, where we may see evidence for a diffraction minimum. The neutron form factor will also be extended to ∼ 5 (GeV/c)2. Further extensions of a factor of two are planned with the 12 GeV Upgrade. Such extensions have 0.00 0.05 0.10 0 2 4 µ p G E 0.0 0.5 1.0 1.5 Bijker (VMD) Holzwarth (soliton) Hammer (VMD + disp. rel.) Miller (rel. QCM + “bag”) 0 2 4 6 8 10 1.0 1.5 0 2 4 6 8 10 0.6 0.8 1.0 1.2 0 2 4 6 8 10 F2/F1∝ ln 2 (Q 2 /Λ 2 )/Q 2 p G M p G E n G M p µ p G D G M n µ n G D Q2 (GeV2) Q2 (GeV2) Q2 (GeV2) Q2 (GeV2) Fig. 3. Theoretical descriptions of the nucleon form factor data [1]. historically proven to be important, and we expect these data will provide further insight and sensitivity for com- pleting our understanding of how to construct nucleons from quarks and gluons. 10 The European Physical Journal A -0.02 -0.01 0.00 0.03 0.02 0.01 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 r 2 · ρ ( r; G E p p o l ) / f m -1 r / fm π+ -b p p0 b p n0 total -0.02 -0.01 0.00 0.03 0.02 0.01 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 r 2 · ρ ( r; G E n ) / fm -1 r / fm π− b n p0 (1- b n )n0 total Fig. 4. Neutron radial charge distribution (top) and the polar- ization term contribution to the proton radial charge distribu- tion (bottom) as inferred from an analysis using phenomeno- logical models [2] of the world nucleon form factor data. The green solid lines show the pion cloud contribution. Note that r2ρ(r) is plotted, emphasizing the distribution at large radii. The strange quark form factors have become an in- teresting area of study that is both analogous and com- plementary to the classical electromagnetic form factors. By using the weak component of the electro-weak interac- tion we access the weak neutral current form factor, which can be interpreted very elegantly in terms of the strange quark distribution. Because there are no valence strange quarks, this measurement provides a unique window on the sea quark distribution. The strange form factors can also be expected to provide us with interesting experi- mental insights into nucleon structure: by combining the electromagnetic and the weak neutral current form factors we should be able to separate the spatial distribution of the u, d, and s quarks. Figure 5 shows the world’s data on the strange proton form factor taken at forward angles as a function of Q2. One sees data from the A4 experiment at Mainz [6,7,8], from HAPPEx I and II (JLab Hall A) [9,10,11], and from G0 (JLab Hall C) [13]. These difficult experiments would be impossible without highly polarized electron beams from magnificently stable accelerators. The fact that the data from different laboratories lie roughly on a smooth curve gives one confidence that the experimenters are do- ing it right. The first thing that strikes you about the data is that the form factor is rather small. This is to be expected, as all of the strange quarks emerge as quark-antiquark pairs popping in and out of the vacuum, and to get a finite form factor there must be some kind of a polarizing effect A4 G0 HAPPEx Fig. 5. Nucleon strange form factor data from parity-violating electron scattering at forward angles as a function of Q2. -1 0 1 -0.1 0 0.1 E s G M s G SAMPLE with calculationAG He ('04) 4 HAPPEX- A 4 G0 (extrapolated)HAPPEX-H ('04) 95% C.L. Q2 = 0.1 GeV2 Leinweber et al. (2005) + 1 2 3 Fig. 6. Separated strange nucleon form factors at Q2 = 0.1 (GeV/c)2 [10]. separating them spatially. Even at the highest momentum transfers reached experimentally we are averaging over a distance scale that is roughly the size of the nucleon, so it is not too surprising that the result is small. There is an intriguing suggestion in the data for something that one would call vaguely pion-cloud like behavior, but it is fair to say that the statistical significance of this effect is not very high. The data taken at forward angles includes a mixture of electric and magnetic form factors. At Q2 = 0.1 (GeV/c)2 we have data at both forward and backward scattering angles, so we can separate these effects (see fig. 6). The data favor a positive value of GsM , which is at variance with most of the theoretical models. Experiments are in progress that will reduce the size of the error ellipse at this Q2 value by a about a factor of 3, and additional experiments planned at both MAMI and JLab will permit separations at other Q2 values. A broad, world-wide effort will provide the results we want. Another interesting experiment is the measurement of the pion form factor. The pion is the simplest QCD bound system, the “positronium” of QCD. One expects that the pion form factor will provide us with evidence for the L.S. Cardman: Physics at the Thomas Jefferson National Accelerator Facility 11 0 0.2 0.4 0.6 0 1 2 3 4 5 6 7 8 Q 2 (GeV/c) 2 Q 2 F π BSE+DSE QCD Sum Rule CQM Perturbative QCD Amendolia π+e elastics Previous p(e,e’π + )n JLab F π -1 Results JLab F π -2 Results PRE LIMIN ARY Fig. 7. Pion form factor data [14]. transition of the strong interaction from the perturbative (QED-like) to the strong (confinement) regime at the low- est possible momentum transfer. These data also constrain phenomenological models of the pion. Measuring the pion form factor is not simple. At low Q2, one can scatter pions off atomic electrons, but a bowl- ing ball does not transfer energy to a ping-pong ball effi- ciently, and even with very high energy pions this exper- iment cannot reach high momentum transfers. To reach higher momentum transfer in the absence of a free pion target, one must scatter electrons off virtual pions inside a proton and extrapolate the data to the pion pole. The world’s data (see fig. 7) is beginning to distinguish be- tween different theoretical approaches. With the 12 GeV Upgrade, we expect to extend the data out to a momen- tum transfer of 6 (GeV/c)2 and to be able to infer the distance scale for the onset of perturbative behavior. 3.2 What is the excited state spectrum of the hadrons? If one looks at several decades worth of data on nucleon resonances and tries to use a simple quark model to clas- sify the states in terms of the excitation in units of h̄ω and the angular momentum of the three quarks, the states that have been identified so far fit nicely into this scheme, but there are many states that have been predicted but have not been found. It is an interesting fact that one can explain all of the states that have been seen so far by assuming that the nucleon and its excited states are a diquark-quark system. Since most of the data have been obtained from pion-induced reactions, and many of the missing states are predicted not to couple to pions, it is also possible that the missing states may have been over- looked for experimental reasons. In atomic spectroscopy the line spacing is large com- pared to the line width, and measuring the complete spectrum is relatively straightforward. In nucleon spec- troscopy, the strong interaction causes the width to be comparable to the spacing. Identifying weak states and Fig. 8. W-dependence of the scattered electron rate for the p(e, e′)X reaction. CLAS data taken at 4 GeV primary beam energy. The energies of the known excited states are shown in black, while those of the states “missing” in the simple quark model description are shown in red. extracting the internal quark structure from the measured cross sections is a difficult task. The problem can be seen easily in fig. 8, which shows the inclusive electron scatter- ing cross spectrum from the proton for a 4 GeV electron beam. With a modern electron accelerator and a large ac- ceptance detector one can obtain data on the transition form factors over a large Q2-range [1 → 4 (GeV/c)2] in a single shot. There is plenty of cross section in the region where the missing states (shown in red) have been pre- dicted, but extracting their individual strengths from the data is a real challenge. The combination of cw electron beams and modern, large solid angle detectors provides important advantages for addressing this problem experimentally. If one looks at the same data set of fig. 8 but uses the information on the energy and momentum of the final state proton mea- sured in coincidence with the inelastically scattered elec- tron, it is straightforward to infer the missing mass asso- ciated with the decay of the excited state (see fig. 9). One can see clearly from the raw missing mass spectrum that the “missing” states do not couple to pions, but rather to the η and ω. With the further information on the an- gular correlations of those decay particles relative to the momentum transfer axis one finally has the information necessary to decompose the spectrum of fig. 8, and learn just what is there. This effort naturally begins with the ∆(1232), which decays predominantly into pion and nucleon. Figure 10 shows a comparison of separated structure functions from CLAS data for the p(e,e′p)πo reaction with theoretical fits and results from previous experiments. The ∆ → γ∗N-transition is characterized by three mul- tipoles: the electric quadrupole E, the magnetic dipole M, and the scalar multipole S. As we examine this transi- tion as a function of momentum transfer we expect that different aspects of the excitation will become apparent. At large distance scales (corresponding to low momentum transfers) we should see the effect of the pion cloud, while at large momentum transfer (corresponding to short dis- tances) we will eventually reach the limit given by pQCD 12 The European Physical Journal A N(1680) N(1520) ∆(1232) W (G e V ) π0 η ω M x (GeV) 2π thresh. M is s in g S ta te s Fig. 9. W vs. missing-mass Mx for the same CLAS data shown in fig. 8 [15]. Fig. 10. Angular distributions of the separated structure func- tions for the p(e, e′p)πo reaction in the Delta region [16]. where REM = E/M → 1, and we further expect that the S/M ratio RSM will become constant. Results from an early experiment at JLab and data from MAMI and Bates, all in the low-Q2 regime show the effect of the pion cloud clearly (see fig. 11). As a function of Q2, REM remains small and negative at high Q 2 with a trend toward 0 and a possible sign change. RSM continues to rise in magnitude with Q2. No trend is seen towards Q2- independence. We can only conclude that even at Q2 of 10 (GeV/c)2 we are far from the pQCD regime. Pion cloud models describe the data well (fitted to low and high-Q2 points). Unquenched Lattice QCD gives the correct signs and approximate magnitudes. One of the most interesting examples of the impact of the pion cloud and of the value of measuring the tran- sition form factors for nucleon excitation is the Roper Fig. 11. Ratios REM and RSM as a function of Q 2 for the ∆ → γ ∗N-transition [17]. -150 -100 -50 0 50 100 150 0 1 2 3 4 A 1 /2 ( 1 0 -3 G e V -1 /2 ) P 11 (1440) -60 -40 -20 0 20 40 60 80 0 1 2 3 4 S 1 /2 ( 1 0 -3 G e V -1 /2 ) P 11 (1440) Q2 (GeV2) Q2 (GeV2) Fig. 12. A 1 2 (Q2) and S 1 2 (Q2) multipoles for the P11(1440) Roper resonance [18]. resonance. According to the constituent quark model the N∗(1440)P11 state is an N = 2 radial excitation of the nu- cleon. However, the properties of this state such as its mass and photocouplings are not well described by this model. The new CLAS data (see fig. 12) seem to explain this puzzle. At low momentum transfer, what one is measur- ing is dominated by the pion cloud. As you start squeezing down the distance scale, what emerges is the underlying quark structure of the Roper, which is, in fact, roughly consistent with a radial excitation. Investigation of nucleon excitation through the mea- surement of the transition form factors is now slowly mov- ing up in excitation energy. Most of this analysis is at a preliminary stage, and what is really needed is a coher- ent study of many channels at many values of momentum transfer in a consistent (and comprehensive) analysis. It will be a long time before we have all the answers. As we search through this data, we are coming across intriguing evidence for states that have been “missing”. For example, there is evidence for a possible new N* state near 1840 MeV visible in the Λ photo- and electro- production data. In the forward hemisphere, one sees a nice peak from a known N* state at 1.7 GeV; in the L.S. Cardman: Physics at the Thomas Jefferson National Accelerator Facility 13 W (GeV) 1.6 1.7 1.8 1.9 2 2.1 b /s r µ 0.05 0.1 0.15 L σ L ∈ + T σ Λ 2 = 0.7 (GeV/c) 2 ) < 0., QKΘ-1. < Cos( known N* New N* ? Fig. 13. W-dependence of the cross section for the p(e, e′K+)Λ reaction integrated over backward-going K+ [15]. backward hemisphere (see fig. 13), one sees an additional unexpected structure. A detailed analysis shows that the angular distribution can be fit nicely with the addition of a new P11 state at 1840 MeV with a width Γ = 140 MeV to the known D13(1870) and D13(2170) states. Intriguingly, a P11 state at 1840 MeV is consistent with the symmetric quark model and SU(6) × O(3) symmetry, but is incon- sistent with diquark-quark symmetry. I feel obliged to bring you up to date on the penta- quark (or lack thereof). There was a great deal of excite- ment for a while about what appeared to be evidence for a 5-quark state. There have three experiments at JLab pushing to substantially higher statistics, both in the γp and the γd channel, and for virtual photons as well. No evidence for a 5-quark state has been found in the first analyses of these new data. 3.3 What is the QCD basis for the spin structure of the hadrons? In addition to the investigation of the spatial distributions of charge and magnetization in the nucleon and its excited state spectrum, the third important experimental focus is the nucleon’s spin structure. The first thing to look at is the spin structure function of the valence quarks at high- x. The data for the proton was reasonable; the new CLAS data with somewhat tighter error bars are confirming the old results and improving the overall accuracy. There were no data of any statistical significance for the neutron above an x of 0.3. The 3He experiment at JLab has provided three new data points (see fig. 14). The new data, when folded into a global analysis of the parton distribution functions (PDF), show that the theoretical prejudices used in earlier analyses were wrong; in particular we now know that ∆d/d stays negative at high x. One can make predictions with a minimum of theoret- ical prejudice for the integrals of the spin structure func- tions at the two extremes of distance scales. In the limit of extremely small distances (i.e. for Q2 → ∞), assuming only isospin symmetry and current algebra (or the opera- tor product expansion within QCD), Bjorken showed that 0 0.2 0.4 0.6 0.8 1 x 0 0.2 0.4 0.6 0.8 1.0 -0.2 -0.4 E142 (3He) JLab E99-117 (3He) E154 (3He) HERMES (3He)A1 n A1 p C Q M ( Is gu r) p Q C D f it w ith H H C (L e a d e r e t a l.) Statistical model (Soffer et al.) pQ C D fi t w /o H H C (L ea de r e t a l.) C h ira l S o lit o n ( W e ig e l e t a l.) B a g m o d e l (T h o m a s a t a l. ) x HF perturbed QM World Data parm Q2 = 10 GeV2 Symmetric Q Wave function Helicity 3/2 suppression Spin 3/2 suppression SU(6) pQCD 0 0.2 0.4 0.6 0.8 1.0 1.2 0 0.2 0.4 0.6 0.8 1 CLAS-EB1b Q2 = 1.4 - 4.52 GeV2 HERMES SMC SLAC - E155 SLAC - E143 Fig. 14. (Left) Spin structure function of the neutron, An1 , derived from 3He data [19]; and (Right) spin structure function of the proton A p 1 [20]. Fig. 15. GDH integral as a function of the upper limit νmax [21]. the difference between the proton and the neutron inte- grals is related to the neutron β-decay coupling constant, with a small Q2 dependent correction due to the running of the coupling constant. For very large distance scales (i.e. for Q2 → 0), there is a slightly less rigorous set of assumptions (Lorentz in- variance, gauge invariance, unitarity, and the dispersion relation applied to the forward Compton amplitude) that can be used to show that the difference between the helic- ity 3 2 and 1 2 total cross sections is related to the nucleon anomalous magnetic moment (this is the GDH sum rule). There has been a lovely set of data taken at ELSA and MAMI that have determined the GDH integral as a function of the upper photon energy integration limit (see fig. 15). The experiments were technically challeng- ing [21], requiring the combination of polarized electrons, a polarized target, and large-acceptance detectors. Theo- retical analysis and interpretation of these data show that the GDH sum rule is satisfied at the 5% level. The effort has also provided us with a better understanding of the physics of the reactions contributing to the integral. These data, and the precision with which they have defined the GDH integral at the photon point, provide the foundation for our studies of the Q2 evolution of the moment of the nucleon’s spin structure functions. As one looks at the evolution of the moment of the proton spin structure function with Q2, one expects to see 14 The European Physical Journal A Q2(GeV2) Γ 1 p ( n o e la st ic ) CLAS EG1a SLAC E143 CLAS EG1b HERMES -0.02 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0 0.5 1 1.5 2 2.5 3 Fig. 16. Integral of the spin structure function of the proton as a function of Q2 [22]. 0.15 0.20 Γ p -n 1 0.10 0.05 0 0.5 1.0 1.5 2.0 2.5 Q 2 (GeV/c) 2 3.0 Fig. 17. Bjorken integral (proton-neutron difference) as a function of Q2. The grey band shows the evolution of αeff(Q 2) predicted by pQCD [22]. the anomalous magnetic moment of the proton in the long wavelength limit, whereas at infinite Q2 the Bjorken sum rule is valid. In the regime close to the long wavelength limit, chiral perturbation theory (χPT) allows us to make predictions. The transition between the two extremes is an important piece of information on how the nucleon is put together, and how nucleon structure emerges from the parton soup. We have data now, mainly from JLab, on the evolu- tion of the structure function’s integral for the proton (see fig. 16) and the neutron approaching the GDH sum rule limit at Q2 = 0, and approaching the Bjorken limit at a surprisingly low momentum transfer of about 1 (GeV/c)2. Several experiments at JLab are investigating the region Q (GeV) α s( Q )/ π α s,g1 / π world data α s,τ / π OPAL Burkert-Ioffe pQCD evol. eq. α s,g1 / π JLab α s,F3 / π GDH constraint 0.05 0.1 0.5 1 2 10 -1 1 10 Fig. 18. Effective strong-coupling constant as a function of Q2 derived from the Q2-dependence of the Bjorken integral [23]. of very low momentum transfer with high precision to test the predictions of χPT. The first significant measurement of the Q2-depen- dence of the Bjorken integral (see fig. 17) was made for Q2 = (0.05 − 2.5) (GeV/c)2. Remarkably, pQCD-based Q2 evolution matches the data down to a Q2 of about 0.7 (GeV/c)2. Deur et al. [23] have made an interesting in- terpretation of the Q2-dependence of the Bjorken integral in terms of an effective strong-coupling constant αef f (Q 2) (see fig. 18). Again, there is evidence for a transition oc- curring around Q2 = 1 (GeV/c)2. 4 Explore the limits of our understanding of nuclear structure As described above, experiments at Jefferson Lab are pro- viding essential new insights into nucleon structure. In a very similar way, the precision, spatial resolution, and in- terpretability of experiments performed using electromag- netic probes are being used to address long-standing issues in nuclear physics, including specifically nucleon-nucleon correlations and the identification of the limits of our un- derstanding of finite nuclei. 4.1 Correlations in nuclei Nucleon-nucleon correlations have been a subject of great interest since the beginnings of the field. In his fabled “bible” on nuclear physics, Hans Bethe estimated that these correlations should be of scale a third of what one observes in nuclear physics, and indeed they are. However, finding clear, interpretable evidence for these correlations has been a real challenge to experimentalists. The previous generation of (e,e′p) experiments car- ried out at Saclay, NIKHEF, and Mainz explored the spectral function strength for low-lying shells. Only about 2/3 of the strength anticipated from a simple shell model was found. However, the interpretability of these measure- ments was limited by the uncertainties introduced by the L.S. Cardman: Physics at the Thomas Jefferson National Accelerator Facility 15 a) b) x B c) 1 2 3 1 2 3 4 2 4 6 1 1.5 2 2.5 r( 5 6 F e /3 H e ) r( 1 2 C /3 H e ) r( 4 H e /3 H e ) Fig. 19. Ratio of inelastic scattering cross sections off nuclei relative to 3He as a function of Bjorken-x [24]. corrections necessary for the final-state interactions of the knocked-out protons. A new approach to nucleon-nucleon correlations avoids this problem by comparing the ratio of inelastic electron scattering off 4He, 12C, and 56Fe to 3He in a kinematical regime where the scattering is basically from the quarks within the nucleons, and the scattering from the nucleons as coherent objects is highly suppressed. These data (see fig. 19) tell us that at any given moment the number of cor- related nucleons in 4He, 12C, and 56Fe is ∼ 0.3, ∼ 1.2, and ∼ 6.7, respectively. So about 10% of the time a nucleon is involved in a nucleon-nucleon correlation. The measure- ments further show that three-nucleon correlations are clearly present (at x > 2), and about an order of mag- nitude smaller than two-nucleon correlations. Another approach [25] to the study of correlations is to search explicitly for the strength that was identified as “missing” in the last generation of (e,e′p) experiments. We are using the (e,e′p) reaction at high momentum transfers and high missing energies, a region that was simply not accessible at the lower-energy, high duty-factor facilities previously available. The missing strength was, indeed, found (see fig. 20), and agrees rougly with the predictions of Correlated Basis Function theory (although the mo- mentum distribution is not described correctly in detail). In a third study correlated pairs have been measured directly in the 3He(e,e′pp)n reaction. In this experiment, the absorption of the virtual photon kicks out a proton, and the opening angle of the remaining pair shows a back- to-back peak. One can infer from the data the shape of the pair momentum distribution. Similar, though somewhat less direct, information can be obtained from examining the 3He(e,e′p)X reaction at very high missing momentum. Significant strength above what is predicted by PWIA has been observed (see fig. 21). The quantitative understanding of the results is work in progress. 0 0.1 0.2 0.3 E m (GeV) 1e-13 1e-12 1e-11 1e-10 1e-09 E ( S m p , m V e M [ ) 4 - r s 1 - ] p m (G eV /c) 0.250 0.410 0.570 Experiment Correlated Basis Function Theory Fig. 20. Spectral function for (e, e′p) at high momentum trans- fer and high missing energy [25]. Fig. 21. Effective nucleon density for the 3He(e, e′p)X reaction as a function of missing momentum. 2bbu stands for two-body breakup [26]. 4.2 The limits of our understanding of finite nuclei One of the key issues that motivated the construction of CEBAF was our desire to gain insight into the ques- tion of where the description of nuclei based on nucleon and meson degrees of freedom fails and the underlying quark degrees-of-freedom must be taken into account. Data on the elastic scattering from the deuteron and high- energy photodisintegration, together with accurate theo- retical calculations, are providing the answers. We begin with the elastic scattering form factors for the deuteron. The theory is in an advanced state: we use the best ab initio calculation of the structure of the deuteron with a potential VN N determined from a fit to N-N phase shifts, and then add exchange currents and rel- ativistic corrections. The data set for the deuteron elastic form factors demonstrate the technical accomplishments of modern accelerators and equipment: elastic e-D scat- tering has been measured down to cross sections charac- teristic of ν-scattering! The data for the electric and the magnetic form fac- tors, and for the tensor polarization (see fig. 22) demon- strate that conventional nuclear theory works up to Q2 of 16 The European Physical Journal A 0 2 4 6 Q2 (GeV 2 ) 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 100 previous data Hall A Hall C MMD, S2, D MMD, 0, D Forest and Schiavilla, IA Forest and Schiavilla, IA+pair A(Q 2 ) 0.0 0.5 1.0 1.5 2.0 2.5 Q2 (GeV 2) 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 B(Q 2 ) previous data MMD, S2, D MMD, 0, D Forest and Schiavilla, IA Hall A 0.0 0.5 1.0 1.5 2.0 Q 2 (GeV 2 ) -1.5 -1.0 -0.5 0.0 0.5 1.0 VEPP 1985 Bates 1984 VEPP 1990 Bates 1991 NIKHEF 1995 NIKHEF 1996 JLab/POLDER MMD, S2, D MMD, 0, D Forest and Schiavilla, IA Forest and Schiavilla, t 20 (Q 2 ) IA+pair Fig. 22. Electric and magnetic form factors of the deuteron (top panels) and the deuteron tensor polarization (lower left) together with the intrinsic shape of the deuteron inferred from these data (lower right) (adapted from [27]). about 2 (GeV/c)2, i.e. the nucleon-based picture is still valid at distance scales of about one half the size of a nu- cleon. Why, we do not know; none of us expected it to work that well before the experiments were undertaken. The shape of the deuteron derived from the form fac- tor data is also shown in fig. 22; one can see clearly that the nucleon spins are aligned “end-to-end” (resulting in a “dumbell”-shaped distribution) rather than anti-parallel (which would have yielded a “donut” shape). The photodisintegration of the deuteron was one of the first experiments done in nuclear physics (at ener- gies of only a few MeV) and also one of the most re- cent ones (now at energies approaching 6 GeV). The reac- tion probes internal nucleon momenta well beyond those accessible in electron scattering because of the momen- tum mismatch between the photon and the nucleon. In a parton-based description of the reaction, one expects the cross section to scale like s−11, where s is the CM energy squared. The data (see fig. 23) demonstrate that s−11 scaling of the cross section is reached at photon en- ergies which change with the proton center-of-mass angle. The transition occurs consistently at a transverse momen- tum of about (1.0 − 1.3) GeV/c, which shows that below ∼ 0.2 fm the nucleon-meson description of the deuteron is no longer valid, and a parton-based description is more appropriate. A more recent experiment [28] using CLAS has extended these data to include angular distributions for a broad range of energies; the data is described by a quark-gluon string model. Conventional Nuclear Theory Fig. 23. Cross sections for deuteron photodisintegration. The energies associated with a transverse momentum of 1.37 GeV/c are indicated with a blue arrow in each panel [29]. 5 Summary The CEBAF accelerator at JLab is fulfilling its scientific mission to understand how hadrons are constructed from the quarks and gluons of QCD, to understand the QCD basis for the nucleon-nucleon force, and to explore the transition from the nucleon-meson to a QCD description. Its success is based on the firm foundation of exper- imental and theoretical techniques developed world-wide over the past few decades, on complementary data pro- vided by essential lower-energy facilities, such as MAMI, and on the many insights provided by the scientists we are gathered here to honor. It is a pleasure to acknowledge the assistance of Bernhard Mecking in the development of this article, and thoughtful com- ments on the manuscript from Volker Burkert, Kees de Jager, John Domingo, and Rolf Ent. This work has been supported through The Southeastern Universities Research Association, Inc., which operates the Thomas Jefferson National Accelera- tor Facility under Contract No. DE-AC05-84150 with the U.S. Department of Energy. References 1. C. Hyde-Wright, C.W. de Jager, Annu. Rev. Nucl. Part. Sci. 54, 217 (2004) and references therein. 2. J. Friedrich, T. Walcher, Eur. Phys. J. A 17, 607 (2003). 3. J.J. Kelly, AIP Conf. Proc. 698, 393 (2004). 4. T.M. Ito et al., Phys. Rev. Lett. 92, 102003 (2004). 5. D.T. Spayde et al., Phys. Lett. B 583, 79 (2004). 6. F.E. Maas et al., Phys. Rev. Lett. 94, 152001 (2005). 7. F.E. Maas et al., Phys. Rev. Lett. 95, 022002 (2004). 8. F. Maas, these proceedings. L.S. Cardman: Physics at the Thomas Jefferson National Accelerator Facility 17 9. K.A. Aniol et al., Phys. Rev. Lett. 96, 022003 (2006). 10. K.A. Aniol et al., Phys. Lett. B 635, 275 (2006). 11. K.A. Aniol et al., Phys. Lett. B 509, 211 (2001). 12. K.A. Aniol et al., Phys. Rev. C 69, 065501 (2004). 13. D.S. Armstrong et al., Phys. Rev. Lett. 95, 092001 (2005). 14. T. Horn, private communication for the Fπ collaboration. 15. V.D. Burkert, Eur. Phys. J. A 17, 303 (2003). 16. L.C. Smith, Invited Talk, Japan-US Workshop on Electro- magnetic Meson Production and Chiral Dynamics, Osaka, Japan (April 2005). 17. CLAS Collaboration (V.D. Burkert), Int. J. Mod. Phys A 20, 1531 (2005). 18. I. Aznauryan talk at N*2005, Tallahasse, FL (October 2005), to be published. 19. X. Zheng et al., Phy. Rev. Lett. 92, 012004 (2004). 20. CLAS Collaboration (K.V. Dharmawardane et al.), sub- mitted to Phys. Rev. Lett. (2006). 21. A. Thomas, these proceedings. 22. A. Deur, 13th Int. workshop on Deep Inelastic Scattering (DIS2005), AIP Conf. Proc. 792, 969 (2005). 23. A. Deur, V. Burkert, J.P. Chen, W. Korsch, arXiv:hep- ph/0509113. 24. K. Egiyan et al., Phys. Rev. Lett. 96, 082501 (2006). 25. D. Rohe et al., Phys. Rev. Lett. 93, 182501 (2004). 26. F. Benmokhtar et al., Phys. Rev. Lett. 94, 082305 (2005). 27. M. Garçon, J.W. Van Orden, Adv. Nucl. Phys. 26, 293 (2001) and references therein. 28. M. Mirazita et al., Phys. Rev. C 70, 014005 (2004). 29. E.L. Schulte et al., Phys. Rev. Lett. 87, 102302 (2001). Personal Comments Research at Jefferson Laboratory How are the nucleons made from quarks and glue? What are the spatial distributions of the ensuremath {u}, ensuremath {d}, and ensuremath {s} quarks in the hadrons? What is the excited state spectrum of the hadrons? What is the QCD basis for the spin structure of the hadrons? Explore the limits of our understanding of nuclear structure Correlations in nuclei The limits of our understanding of finite nuclei Summary