Nucleon-Nucleon Correlations, Short-lived Excitations, and the Quarks Within

Or Hen Massachusetts Institute of Technology, Cambridge, MA 02139 Gerald A. Miller Department of Physics, University of Washington, Seattle, WA 98195 Eli Piasetzky School of Physics and Astronomy, Tel Aviv University, Tel Aviv 69978, Israel Lawrence B. Weinstein Department of Physics, Old Dominion University, Norfolk, VA 23529

(Dated: April 17, 2017)

This article reviews our current understanding of how the internal quark structure of a nucleon bound in nuclei differs from that of a free nucleon. We focus on the interpre- tation of measurements of the EMC effect for valence quarks, a reduction in the Deep Inelastic Scattering (DIS) cross-section ratios for nuclei relative to deuterium, and its possible connection to nucleon-nucleon Short-Range Correlations (SRC) in nuclei. Our review and new analysis (involving the amplitudes of non-nucleonic configurations in the nucleus) of the available experimental and theoretical evidence shows that there is a phenomenological relation between the EMC effect and the effects of SRC that is not an accident. The influence of strongly correlated neutron-proton pairs involving highly virtual nucleons is responsible for both effects. These correlated pairs are temporary high-density fluctuations in the nucleus in which the internal structure of the nucleons is briefly modified. This conclusion needs to be solidified by the future experiments and improved theoretical analyses that are discussed herein.

Contents 1. Nucleons only 23 2. Nucleons plus pions 24 D. Beyond Conventional Nuclear Physics: Nucleon I. Introduction - Short Range Correlations (SRC) and Modification 25 Nuclear Dynamics 2 1. Mean field 26 A. The challenge of describing nuclei 4 2. Suppression of point-like configurations 27 B. The need for short range correlations/Beyond the 3. Six-quark bags and the EMC Effect 29 nuclear shell model 4 IV. The EMC - SRC Correlation 29

arXiv:1611.09748v4 [nucl-ex] 13 Apr 2017 1. Spectroscopic factors 4 2. From the NN Interaction to the Shell Model and A. Experimental Overview 29 Beyond 6 B. Theory Overview 31 3. Short-ranged two-nucleon clusters 8 1. High momentum nucleons and PLC suppression 31 2. Effective Field Theory 33 II. Hard scattering and Short-Range Correlations 11 3. The Isovector EMC Effect 34 A. Hard Reactions 11 4. Summary 34 B. Exclusive Scattering 11 C. Are the nucleons in the correlated pair really C. Inclusive Scattering 15 nucleons? 34 D. Universal Properties of Short Range Correlations in D. Determining the structure function of a free neutron 34 Nuclei 18 1. The Deuteron IMC Effect 35 2. The Free Neutron Structure Function 35 III. Deep Inelastic Scattering (DIS) and the EMC effect 19 3. The d/u ratio at large xB 36 A. DIS and nucleon structure functions 20 B. The EMC effect 21 V. Existing searches for medium-modified electromagnetic C. Why Conventional nuclear physics cannot explain the form factors 37 EMC effect 23 A. Polarization transfer in the (~e,e0~p) reaction 37 2

B. Polarization transfer in the (~e,e0~n) reaction 39 1949)) explained that nuclei with certain numbers of neu- C. The (e, e0) reaction and the Coulomb Sum Rule trons and protons (the magic numbers 2, 8, 20, 28, 50, (CSR) 39 82, 126, ...) are particularly stable because their shells VI. Future directions in nuclear deep inelastic scattering and are filled. Many studies were devoted to understanding detecting short-ranged correlations 40 how the liquid drop model, with its collective features, A. Experiment 40 could be consistent with the shell model. 1. Tagged Structure function Measurements 40 Detailed studies of nucleon-nucleon scattering indi- 2. Inclusive EMC and SRC Measurements 42 3. Semi-Inclusive and Exclusive SRC cated that their interaction contains something like a Measurements 43 hard core, making the origin of the shell model even 4. Polarized EMC Measurements 43 more mysterious than its coexistence with the liquid drop 5. Parity Violating Deep Inelastic Scattering 43 model. Brueckner and other early workers (see the ref- B. Theory 44 erences in (Gomes et al., 1958)) showed that in the VII. The way we think it is and the ways to check 45 nuclear medium, the large, short-ranged effects of the strong nucleon-nucleon potential could be summed and VIII. Acknowledgments 45 treated in terms of a smoother object, defined as a G ma- IX. Appendix 46 trix. This idea allowed much of nuclear phenomena to be A. Understanding the np relative wave function 46 understood (at least qualitatively) in terms of the fun- B. Basic terminology 47 damental nucleon-nucleon interaction. The nucleus was C. Why center-of-mass and relative coordinates made of nucleons, with the occasional evanescent meson factorize 48 existing as it propagated from nucleon to nucleon. After the single-particle shell model, the natural next I. Introduction - Short Range Correlations (SRC) and step in describing nuclei is including the effects of two- Nuclear Dynamics nucleon correlations. The strong short-ranged nucleon- nucleon force that is averaged to make the mean-field Nuclear physics is one of the oldest fields in modern G-matrix also causes a significant nucleon-nucleon corre- physics. Its history (Wong, 1998) separate from atomic lation function (see the Appendix for definitions). How- physics, can be said to start with the discovery of ra- ever, definitive experimental evidence for correlations dioactivity in 1896 by Henri Becquerel. Fifteen years had to await two kinds of high-energy reactions (Frank- later Rutherford used backward scattering of alpha par- furt and Strikman, 1981a). These are the inclusive (e, e0) ticles to discover that the nucleus is a tiny object at the scattering at values of Bjorken xB > 1 (Egiyan et al., heart of the atom. In 1932 Chadwick discovered a neu- 2003, 2006a; Fomin et al., 2012a) and exclusive reactions tral particle of about the same mass as the proton that that could isolate the effects of ground-state correlations he called the neutron. This discovery allowed scientists from the various two-body currents and final state in- to understand that the binding energy accounted for less teractions that occur in nuclear reactions (Baghdasaryan than one percent of the nuclear mass. Thus it is natural et al., 2010; Hen et al., 2014c; Korover et al., 2014; Makek to say that the nucleus is made of neutrons and protons. et al., 2016; Monaghan et al., 2014; Piasetzky et al., 2006; In 1935 Yukawa suggested a theory of the strong force to Shneor et al., 2007; Subedi et al., 2008; Tang et al., 2003). explain how the nucleus holds together. In the Yukawa Meanwhile, deep inelastic scattering on nucleons led to interaction a virtual particle, later called a meson, medi- the discovery that the nucleons are made of quarks. How- ated a force between nucleons. This force explained why ever due to the small ( 1%) nuclear binding energy and nuclei did not fall apart due to proton repulsion, and the idea of quark-gluon≤ confinement, it was thought that it also explained why the attractive strong force had a quarks had no explicit role in the nucleus and that there- shorter range than the electromagnetic proton repulsion. fore nuclei could still be described in terms of nucleons Thus we may think of the stable nucleus as a tight ball of and mesons. The simple and compelling nucleon/meson neutrons and protons (collectively called nucleons), held picture of the nucleus was shaken to its core by the 1982 together by the strong nuclear force. discovery by the European Muon Collaboration (Aubert This basic picture has been studied for many years. et al., 1983), of the non-trivial dependence of the per- Early models treated heavy nuclei, which could contain nucleon lepton deep inelastic scattering cross section on hundreds of nucleons, as classical liquid drops. The the specific nuclear target. The EMC initially reported liquid-drop model can reproduce many features of nu- incorrect results for xB < 0.15. As a result many refer clei, including the general trend of binding energy with to the EMC effect as the reduction of the cross section respect to mass number, as well as nuclear fission. per nucleon in the region 0.2 < xB < 0.7. This reduction The liquid drop idea cannot explain more detailed has been observed many times and we use the term, ‘the properties of nuclei. Quantum-mechanical effects (which EMC effect’ to refer to this region. The observation of can be described using the nuclear shell model developed this reduction, caused by the nucleus, showed that the initially by Mayer (Mayer, 1950) and Jensen (Haxel et al., quarks have a small but definite role in the nucleus. We 3 need to understand this. conventional (non-quark) nuclear physics cannot There are a number of fundamental unanswered ques- • account for the EMC effect, tions about nuclear physics. models need to include nucleon modification to ac- 1. Is the nucleus really made of nucleons and mesons • count for the EMC effect. These models can modify only? the structure of either: 2. How does the nucleus emerge from QCD, a theory – predominantly mean-field nucleons, which are of quarks and gluons? modified by momentum-independent interac- 3. How does the partonic content of the nucleus differ tions, or from that of N free neutrons plus Z free protons? – predominantly nucleons belonging to SRC No one asked such questions before the discovery of the pairs, or EMC effect. – both mean-field and SRC nucleons, At first glance there appears to be little relation be- tween nucleon-nucleon correlations and the EMC effect. there is a phenomenological connection between However, there is a strong phenomenological connection • the strength of the EMC effect and the probability between them (Weinstein et al., 2011) that occurs for that a nucleon belongs to a two-nucleon SRC pair the valence quarks that carry large momentum and that (a2(A)). This connection has also been derived us- connection is the subject of this review. Indeed, the fun- ing two completely different theories, so that it is damental challenge for current explanations of the EMC no accident, effect is to explain also the inclusive and exclusive high in contrast to previous static models of the EMC momentum transfer reactions dominated by short ranged • correlations which take up about 20% of the wave func- effect, the association with SRC implies that nucle- tion. The data suggest that the non-nucleonic admixture ons are temporarily modified only when they briefly in these correlations is at most about 10%, leading to a fluctuate into an SRC pair, 2% non-nucleonic contribution. However, the EMC effect the influence of SRC pairs can account for the is about 15%, so that one needs to find an enhancement • mechanism. EMC-SRC correlation because both effects are driven by high virtuality nucleons (p2 = M 2), We now summarize our most important conclusions for 6 the benefit of the reader: high-virtuality nucleons have an enhanced but still • there is much indirect and direct evidence for the small amplitude for non-nucleonic configurations. • existence of nucleon-nucleon short-ranged correla- Interference effects between nucleonic and non- tions (SRC), nucleonic components (linear in the amplitudes) are responsible for the EMC effect, – high energy (e, e0pN) and (p, 2pN) reactions show that two-nucleon correlations exist in nu- modified nucleons, by definition, must contain a • clei, dominate the high-momentum (k kF ) small fraction of baryons that are not nucleons. tail of the nuclear momentum distribution,≥ Amplitudes for such baryons, with effects enhanced and are dominated, at certain nucleon mo- in a coherent manner, exist in the short-ranged cor- menta, by np pairs, and relations, and are the source of the EMC effect. – high energy (e, e ) reactions at large values of 0 We aim to critically discuss the reasons for these con- x (the Bjorken scaling variable) show that all B clusions and provide enough details for the reader to ap- nuclei have similar momentum distributions preciate the progress that has been made in recent years. at large momentum, consistent with the di- The remainder of this article describes the experimental rect observation that strongly-correlated two- and theoretical evidence for the existence of two-nucleon nucleon clusters exist in the nuclear ground short range correlations and the properties thereof; the state, theoretical and experimental facts regarding deep inelas- – a consequence of the np-SRC dominance is the tic scattering, nucleon structure functions and the EMC possible inversion of the kinetic energy shar- effect; and, the need for nucleon modification to explain ing in nuclei with N > Z (i.e., that protons the EMC effect. It will then present the unexpected cor- might have more kinetic energy than neutrons relation between the strength of the EMC effect in a given in neutron-rich nuclei). nucleus and the probability that a nucleon in that nucleus – this leads to a dynamic model of nuclei where belongs to an SRC pair. The ensuing discussion presents SRC pairs are temporary large fluctuations in theoretical ideas connecting SRC and EMC physics, and the local nuclear density. explores the idea that the SRC-EMC correlation can be 4 used to determine the structure function of a free neu- derstanding the broad range of nuclear phenomena re- tron. The final sections are concerned with other evi- quires the use of many experimental tools. Since electro- dence that the nuclear medium modifies the structure of magnetic interactions are well-understood and presum- bound nucleons, and future directions for experimental ably simple, electron scattering has long been used as a and theoretical research. The Appendix presents formal tool to investigate different aspects of nuclear structure. definitions of the terms we use, and also explains some We examine the use of electron scattering to probe the equations used in the main text. Specific locations of the validity of the single-particle shell model in the next sub- various subjects are listed in the Table of Contents. section.

A. The challenge of describing nuclei B. The need for short range correlations/Beyond the nuclear shell model Nucleons bound in nuclei move under the influence of the strong interaction as effected by short-ranged two and 1. Spectroscopic factors three body potentials. Solving even the non-relativistic A-body Schroedinger equation was initially an impossi- Data from electro-induced proton knockout reactions bly daunting challenge, so that understanding the vast on nuclei, A(e, e0p), provided early evidence for the va- array of relevant experimental data required the use of lidity of the shell model (Frullani and Mougey, 1984). models. These studies complemented the use of low-energy nu- clear reactions, such as (d, p) and (p, pp). Later on, more The nuclear shell model was one of the earliest and per- detailed studies using higher energy electron beams ex- haps most powerful models. In this model, each nucleon plored the limits of the validity of the shell model. We moves independently in the average field produced by the next explain how this happened. other nucleons. This shell model provides a reasonable In the (e, e p) reaction the electron knocks out a nu- description of many nuclear properties and is the funda- 0 cleon so that an initial nuclear state i of A nucleons is mental starting point for all efforts to provide a theory of converted to a final nuclear state f |ofi A 1 nucleons. nuclei. Its explanation of the nuclear magic numbers is a The reaction can be analyzed in terms| i of− spectroscopic major accomplishment in the history of physics. Despite factors (Macfarlane and French, 1960), which are proba- this, early research involving collective degrees of free- bilities that all but one of the nucleons willl find them- dom established that the single particle picture of nuclei selves in the final state. More formally, if one considers could not be complete. More generally, corrections to the a single-particle state of quantum numbers α, the spec- shell model can be classified broadly in terms of the rele- troscopic factor S is given by the square of the overlap: vant distances needed to describe the various phenomena. α S = f b i 2, where b destroys a nucleon. If the in- There are both long-ranged ( the size of the nucleus) α α α dependent|h | particle| i| model were exact, then S would be and short-ranged ( the size of∼ the nucleon) phenomena. α ∼ unity for each occupied state α. Thus measuring S is a The strong nucleon-nucleon force is known to bind α useful way to study the nuclear wave functions and the medium and heavy nuclei, all with about the same aver- 3 limitations of the independent particle model. age central density of ρA = 0.16 nucleons/fm . Thus, the 1/3 In the Plane Wave Impulse Approximation (PWIA), average distance between nucleons is about 1/ρA = 1.8 an electron transfers a single virtual photon with mo- fm. The radius of a nucleon is about r = 0.86 fm, so mentum q and energy ν (sometimes written ω) to a sin- that most (but not all) of the time it does not overlap gle proton, then leaves the nucleus without reinteracting in space with other nucleons. The nucleon has a volume 4 3 3 and can thus be described by a plane wave (see Fig. 1). of V = 3 πr = 2.5 fm and a corresponding density of 3 In PWIA the cross section factorizes in the form (Kelly, ρN = 0.4 fm− . Thus ρN /ρA = 2.5 and the maximum 1996) nuclear density, even without nucleons overlapping, is 2.5 times the average nuclear density. dσ = KσepS(Emiss, pmiss) (1) The fact that a nucleon has about 2.5 times larger den- dνdΩedEmissdΩp sity than the nuclear central density and that nucleons where K = E p /(2π)3 is a kinematical factor, E and move in the nucleus with about a quarter of the velocity p p p p are the energy and momentum of the outgoing pro- of light opens up the possibility of large local density fluc- p ton, σ is the electron cross section (De Forest, 1983a) tuations. These also lead to large local momentum fluc- ep for scattering by a bound proton, and S is the spectral tuations via the uncertainty principle. The strong short function, the probability of finding a nucleon in the nu- range repulsive force between nucleons restrains the size cleus with momentum p and separation energy E . of these fluctuations, but since its range is smaller than a miss miss The missing momentum and missing energy are given by: fermi, the density and momentum fluctuations in nuclei can still be quite large. p = q p miss − p The diverse features described above indicate that un- Emiss = ν Tp TA 1 (2) − − − 5

FIG. 1: The A(e, e0p) reaction in the Plane Wave Impulse Approximation. A nucleus of four-momentum PA emits a nucleon of four-momentum Pmiss that absorbs a virtual photon of four-momentum q to make a nucleon of four momentum Pmiss + q, with 2 2 (Pmiss + q) = M , where M is the nucleon mass. The blob represents the in-medium electromagnetic form factors.

where Tp and TA 1 are the kinetic energies of the de- tected proton and− residual (undetected) A 1 nucleus. − However, the knocked-out proton then interacts with other nucleons as it leaves the nucleus; these final state interaction (FSI) effects have been typically calculated using either an optical model at low momenta (Kelly, 1996) or using the eikonal or Glauber approximations at higher momenta (Ryckebusch et al., 2003; Sargsian et al., 2005b). Calculations where the wave function of the knocked-out proton are distorted by FSI are referred to as distorted wave impulse approximation calculation (DWIA). [Note that FSI effects mean that pmiss is no longer equal to the initial momentum of the struck nu- cleon.] In DWIA, the (e, e0p) cross section does not ex- actly factorize as in the PWIA. However, factorization FIG. 2: (upper) The O(e, e p) cross section plotted is a good approximation at Q2 >> p2 and the cross 0 miss versus missing energy at Q2 = 0.8 GeV2 and ν = 0.439 section is approximately proportional to a distorted spec- GeV for different angles, θ , between the proton tral function SD (Kelly, 1996). Neither PWIA nor DWIA pq spectrometer and q. The curve labelled DWIA is a calculations conserve current because the initial and final distorted wave impulse approximation calculation of wave functions of the model calculations are not orthogo- s-shell knockout; the other curves are calculations of nal and because the effective NN interactions used in the two-nucleon knockout including meson exchange initial and final states are different. (Some models force currents (MEC), delta production (IC), and central current conservation by arbitrarily modifying kinematic and/or tensor correlations. Figure from (Liyanage variables such as qµ (De Forest, 1983b).) Relativisitic et al., 2001). (lower) The cross section plotted versus DWIA models were developed by Van Orden and collab- missing momentum for the 1p and 1p states. orators (Picklesimer and Van Orden, 1989; Picklesimer 1/2 3/2 Figure from (Gao et al., 2000). The curves show DWIA et al., 1985) and later elaborated by (Udias et al., 1999a, calculations. See (Fissum et al., 2004; Gao et al., 2000; 1993, 1995) and (Kelly and Wallace, 1994; Kelly, 1999). Liyanage et al., 2001) for details. Thus, (e, e0p) measurements should be sensitive to the 6

plies the existence of collective effects (long range cor- relations) and short range correlations in nuclei. The spectroscopic factors and the size of the collective ef- fects depend on momentum transfer (Frankfurt et al., 2001; Lapikas et al., 2000). In addition, the spectroscopic strength for valence proton knockout (e.g., 1p3/2 proton knockout from C) is distributed over many states and not all of these states are included when measuring the spectroscopic factor. The results in Fig. 3) cannot be directly related to the probability of short range correla- tions in nuclei due to the effects of momentum transfer- dependence, state-splitting, and collective effects. Our focus will be on the short range correlations as observed using high-momentum transfer probes. In the DWIA independent particle shell model we would expect that the spectroscopic factors are unity and that there is little cross section at large Emiss. The fact that spectroscopic factors are significantly less than unity for all nuclei, and that there is significant cross section at large missing energy indicates that this simple model picture omits important physics. This is not surprising, since the short-ranged nature of the strong nuclear forces implies that nucleons must be influenced by nearby nucle- FIG. 3: The fractional spectroscopic factors (the ratio ons. There is no fundamental one-body potential in the of measured cross sections to those calculated with the nucleus, unlike the central one-body Coulomb potential Independent Particle Shell Model) for valence nucleon that binds electrons to form the structure of the atom. knockout (e, e0p). Reproduced based on (Lapikas, 1993). Indeed, since the NN forces are short ranged, the fact that the shell model approximation has any relevance is somewhat surprising. In the early days of nuclear spectral function, i.e., to the momentum and energy dis- physics, the fundamental question of nuclear physics was: tributions of nucleons in the nucleus. Fig. 2 shows the how does the very successful shell model of the nucleus 16 2 2 O(e, e0p) cross section at Q = 0.8 GeV and ν = 0.439 emerge in spite of the strong short-ranged interactions GeV plotted versus missing energy at several different between nucleons? missing momenta and plotted versus missing momen- We next answer this fundamental question, then ex- tum for the two p-shell states. There are sharp peaks at amine the consequences of the answer. Emiss = 12 and 18 MeV, corresponding to proton knock- out from the 1p1/2 and 1p3/2 shells, a broad peak at Emiss 40 MeV corresponding to proton knockout from 2. From the NN Interaction to the Shell Model and Beyond the 1s≈shell (and other processes), and a long tail extend- ing to large Emiss, especially at the largest missing mo- How can the mean-field shell model arise from a menta. The momentum distribution calculations shown system made of nucleons interacting by strong short- in Fig. 2(lower) use an optical potential, a modern bound ranged forces? An answer to this question was pro- state wave function, and an off-shell cross section σep and vided early on by Brueckner & Goldstone, see the re- fit only the magnitude (see Ref. (Gao et al., 2000) for view by Bethe (Bethe, 1971). The strong two-nucleon details). The calculations describe the data well, except interactions encoded by the potential V , constructed to for the fact that the ratio of data to theory (the spectro- reproduce experimentally measured NN scattering ob- scopic factor) is approximately 0.7. This means that the servables and believed to include strong repulsion at short experiment only measured 70% of the expected number distance and attraction at longer ranges, are summed to of p-shell protons. form the T matrix of scattering theory and the G-matrix This depletion of the spectroscopic factor was observed for bound states. The operator G is obtained from T over a wide range of the periodic table at relatively low by modifying the propagator of the Lippmann-Schwinger momentum transfer (see Fig. 3) for both valence nu- equation to include the effects of the Pauli principle and cleon knockout using the (e, e0p) reaction (Lapikas, 1993) to use the appropriate self-consistent (single) nucleon en- and stripping using the (d,3 He) reaction (Kramer et al., ergies. The G matrix is considerably weaker than V . For 2001). Only about 60–70% of the expected valence nu- example, even if the potential is infinitely strong, the cleon strength was observed. The missing strength im- product V Ψ of the potential with the wave function ob- 7 tained from the chosen Hamiltonian would be finite and ables. See for example the review (Bedaque and van well behaved. Schematically, one has GΦ = V Ψ, where Kolck, 2002). In such theories the short distance interac- Φ is the shell-model two-nucleon wave function. Calcula- tion can be treated as a contact interaction, modified by tions show that G is reasonably smooth and can be used the inclusion of a cut-off, and the longer ranged interac- as input in higher-order calculations. tions are accounted for by one and two pion (or more) ex- The theory proceeds by forming the nuclear mean field change interactions (Machleidt and Entem, 2011). The U through the Hartree-Fock method employing the G- advantage gained is that different parts of the poten- matrix, and the first approximation to the wave function tial are divided between more easily understood long is the anti-symmetrized product of single particle wave ranged contributions and presumably unknown short- functions engendered by U. However, the complete nu- ranged contributions. clear wave function is obtained in a perturbative hole- (2) Modern first-principles calculations of nuclear spec- line expansion that includes two-particle – two-hole exci- tra have been applied to an ever increasing mass range. tations and other excitations which incorporate correla- One of the main tools is the use of soft potentials, which tions. The presence of such correlations is demanded by do not connect low-relative momentum states to those of the theory. high relative momentum. This greatly simplifies the cal- Later work formulated a relativistic version of Brueck- culations by increasing the validity of perturbation the- ner theory in which the Dirac equation replaces the ory and other approximation techniques. Schroedinger equation (Anastasio et al., 1983; Brock- The softness (involving low momentum) or hardness mann and Machleidt, 1984). There is also a light front (involving higher momentum) of the potential is deter- version (Miller and Machleidt, 1999a; Miller, 2000). mined by the value of the cutoff see e.g. (Epelbaum et al., The Brueckner theory approach described above pre- 2009; Machleidt and Entem, 2011). Such potentials in- sumes that the two-nucleon potential contains strong troduce a cutoff in momentum space at fairly low values short-distance repulsion. Early attempts to construct of momenta. Typically, the momentum-space potential soft potentials (i.e., lacking the strong repulsion) that obtained from Feynman diagrams, V (p, p0), is replaced: also reproduce scattering data did not succeed in obtain- 0 ( p )n ( p )n ing interactions that could be used perturbatively to cal- V (p, p0) V (p, p0)e− Λ e− Λ (3) culate nuclear bound states (Bethe, 1971). This failure is → now known to be caused in large measure by the omission with p = p , p0 = p0 , Λ ranges between 400 and 500 | | | | of three-body forces. Relativistic G-matrix calculations MeV and n ranges from 2 to 4. These are very strong cut- include important three-body forces (Anastasio et al., offs in momentum that introduce significant non-locality 1983; Brockmann and Machleidt, 1984; Miller and Mach- to the nucleon-nucleon interaction. This causes difficul- leidt, 1999a; Miller, 2000). There are also fundamental ties in maintaining conservation of the electromagnetic three-nucleon forces, such as those involving an interme- currents (Gross and Riska, 1987). diate ∆ resonance. In addition to true three-body forces, Another approach uses renormalization group methods induced multi-nucleon forces occur as a result of using to generate a soft NN potential from a hard interaction unitary transformations to produce soft, two-nucleon in- either by integrating out high momentum components teractions (Bogner et al., 2010). (in the case of Vlow K ), or by using the similarity renor- − Much more has been learned since Bethe’s 1971 re- malization group (Bogner et al., 2010). This potential is view. (1) Our understanding of the connection through perturbative in the sense that the Born series for scat- symmetries between the NN interaction and the under- tering converges. Furthermore, many-body perturbation lying theory of QCD is much improved. (2) Our ability theory starting from a Hartree-Fock bound state can be to make fundamental first-principles calculations of nu- applied to the nuclear bound state problem. clear energies is also much improved. (3) However, it is (3) But there is another more general issue that arises possible that improved treatments of nuclear energy lev- in trying to understand high momentum transfer nuclear els decrease our ability to understand the nuclear high- reactions. The ability to originate and predict the results momentum transfer interactions of interest in this review. of experiments that probe short-ranged correlations (as (4) We now know that 2nd order interactions of the NN was done in (Frankfurt and Strikman, 1981a, 1988b)) de- potential have a major effect on the density distribution pends on the idea that the simple impulse approximation and the correlation function in all existing approaches. is the best way to think about the relevant kinematics (1) Chiral effective field theory provides a low-energy and reaction physics. This simplicity may be lost if one version of QCD, guided by chiral symmetry, in which uses dynamics generated by the different intent of simpli- one obtains the potential as an expansion in powers of fying nuclear spectroscopy. We explain. Let us suppose (Q/Λχ) where Q is a generic external momentum or the that the renormalization group successfully eliminates pion mass, and Λχ is the chiral symmetry breaking scale matrix elements of the nucleon-nucleon (or inter-nucleon) of about 1 GeV. Such approaches have the advantage of potential connecting low and high relative momentum being systematically improvable for low-energy observ- states, leading to an accurate reproduction of nuclear 8 binding energies and spectra. This procedure would also daque and van Kolck, 2002). lead to wave functions without high-momentum compo- The effect of this on the relative s-wave function of two nents and truly short ranged-correlations. However, it nucleons in nuclei can be characterized by the effective would be necessary to consistently transform all other potential operators (Anderson et al., 2010; Neff et al., 2015) in or- 1 der to calculate observables. For high momentum trans- V = V QV , (4) 00 T E H T fer reactions, the renormalization group changes a known − 0 simple probe, described by a single-nucleon operator, into where VT is the tensor potential, the subscript 00 in- a more complicated probe describable by unknown (in dicates an s-wave to s-wave matrix element, H0 is the practice) A-nucleon operators. This could prevent the Hamiltonian in the absence of VT , and Q is a projection efficient analysis of any high momentum transfer experi- operator taking the Pauli principle into account. The op- ment. The same remark holds for chiral potentials. The erator V00 has a major effect on the density distribution use of a cutoff, as in Eq. (3), leads to the violation of and correlation function (as discussed in the Appendix). current conservation in electromagnetic interactions un- These effects occur in all existing approaches. A major less the currents are modified substantially. For example, purpose of this review is to show that the influence of one could use minimal substitution, which would intro- the correlations induced by the tensor force is manifest duce terms involving several powers of the electromag- in high momentum transfer reactions. µ netic potential A . This means that the simplicity of us- To summarize, nuclear theorists have made tremen- ing electromagnetic probes would be lost because of the dous progress in understanding the connections between need to use very complicated operators to analyze exper- NN potentials and QCD, as well as in calculating nu- iments. Again we reach the same conclusion: the use of clear energies and states. High momentum transfer ex- potentials with strong momentum-dependence is not op- periments are easier to analyze using well-defined cur- timum for the purpose of using high momentum transfer rent operators, rather than using transformed A-nucleon electromagnetic processes to understand the short-range operators with a renormalization-group-transformed po- structure of nuclei. tential. These well-defined current operators can be used It is worthwhile to put comments (1)-(3) into a broader if the effects of correlations are maintained in the nu- perspective. The goal of EFT is to obtain results that clear wave function instead of being hidden in the current are independent of the chosen cutoff. In principle, this operators through the use of the renormalization group can be done. In practice, one chooses a given scale or very soft NN potentials. However, regardless of ap- to simplify the problem at hand. The use of low mo- proach, the influence of the correlations induced by the mentum scales simplifies nuclear structure calculations, tensor force is manifest in all theoretical approaches to but complicates the currents needed to understand high- date, and, as we shall see, is manifest in high momentum momentum transfer reactions. The use of one-body cur- transfer reactions. rents of the impulse approximation simplifies the un- derstanding of high-momentum transfer nuclear reac- tions, but involves NN potentials that do not have low- 3. Short-ranged two-nucleon clusters momentum cutoffs. Bjorken scaling, (Bjorken, 1966) ob- tained via the use of the simple currents of the non- As discussed in previous Sections, in the nucleus, nu- interacting quark model (impulse approximation) offers cleons behave approximately as independent particles a useful historical example. If Bjorken had been overly in a mean field created by their average interaction concerned with issues of QCD evolution, Bjorken scal- with the other nucleons. But occasionally (20 25% ing and the existence of quarks might never have been in medium/heavy nuclei) two nucleons get close enough− discovered. Therefore, we take the experiment-based, to each other so that temporarily their singular short discovery-based view that we are using an implicit mo- range interaction cannot be well described by a mean mentum scale at which the impulse approximation offers field approximation. These are the two nucleon a reasonable first approximation to the physics at hand short-ranged correlations (2N-SRC), defined op- throughout this review. erationally in experiments as having small center (4) Second-order effects of the tensor term of the one- of mass momentum and large relative momen- pion exchange potential are common to all of these ap- tum. These pairs are predominantly neutron-proton proaches, since the beginning (Bethe, 1971; Bogner et al., pairs. Colle et al. (Colle et al., 2015) show that it is pre- 2005; Brown, 1967; Holt et al., 2013; Machleidt, 1989) dominantly nucleon-nucleon pairs in a nodeless relative- and through to the current days of effective field theory. S state of the mean-field that create these 2N-SRC. The These effects are large enough to cause convergence diffi- force between the nucleons in the pair is predominantly a culties in the application of Brueckner theory (Vary et al., tensor force which creates a pair with the quantum num- 1973), and also cause challenges in defining the power bers of the deuteron (S = 1,T = 0), a neutron-proton counting which defines any effective field theory (Be- system (Vanhalst et al., 2011). 9

A FIG. 5: Scaled two-body distribution function ρ2,1(r)/A (see Eq. (83)) for nuclei with A = 2, 3, 4. A correlation hole is seen for all of these nuclei. The two sets of curves are obtained with the AV18+UIX (left) and N2LO (right) potentials. Figure reproduced based FIG. 4: The nucleon momentum distributions n (k) 0 on (Chen et al., 2016). The meaning of R0 is discussed (dashed line) and n(k) (solid line) plotted versus in the text. 1 4 12 56 momentum in fm− for the deuteron, He, C and Fe. Adapted from (Ciofi degli Atti and Simula, 1996a). momentum distribution, which extends from about 300 to 1400 MeV/c in the AV18 potential. See Sect. IX for The two nucleons in 2N-SRC have a typical distance of an explanation. This broad shoulder is also a dominant about 1 fm which means that their local density is a few feature in the tail of the single-nucleon momentum dis- times higher than the average nuclear density. The rela- tributions computed with realistic internucleon interac- tive momentum of the two nucleons in the pair can be a tions, see Fig. 4, in particular with the AV18 potential few times the Fermi momentum, kF , which is large. SRC for A 12 (Wiringa et al., 2014b) and more effective ap- of more than two nucleons probably also exist in nuclei, proaches≤ for heavier systems (Ciofi degli Atti and Simula, and might have higher density than that of the 2N-SRC. 1996b; Ryckebusch et al., 2015). However their probability is expected to be significantly We can also consider the spatial wave function of the smaller than the probability of 2N-SRC (Bethe, 1971). nucleus. The short range part of the NN interaction The 2N-SRC are isospin-dependent fluctuations. For gives a correlation hole at small NN relative distances, example, the deuteron is the only bound two-nucleon sys- see Fig. 5. Precise definitions are given in Sect. IX. Cal- tem. We know now that density fluctuations involving culations with various bare realistic interactions show one neutron and one proton occur more often than those that, apart from a normalization factor depending upon involving like-nucleons, see Sect. IIC. Therefore we ex- the different number of pairs in different nuclei, the rela- amine the deuteron first. tive two-nucleon density ρrel(r) and its spin-isospin com- N1N2 The simplest nucleus, the deuteron, has spin S = 1, ponents ρST (r) at r 1.5 fm exhibit similar correla- isospin T = 0, and J π = 1+. The relevant quantity tion holes, generated by≤ the interplay of the short-range for electron scattering is n(k) which is the probability of repulsion and the intermediate-range tensor attraction of finding a nucleon of momentum between k and k + dk. the NN interaction, with the tensor force governing the This function is the sum of two terms, one arising from overshooting at r 1.0 fm. The correlation hole is uni- the l = 0 (s-wave), and the other from the l = 2 (d- versal, in that it is' almost independent of the mass A of wave). At momenta of interest for short range correlated the nucleus (C. Ciofi degli Atti, 2015). The depth of the pairs (i.e., p significantly greater than p 250 MeV/c, correlation hole depends on the short-distance behavior F ≈ where pF is the typical Fermi momentum for medium and of the potential. The value of R0 shown in Fig. 5 refers heavy nuclei), the otherwise-small d-wave becomes very to the cutoff on the short distance N2LO nucleon-nucleon important. This is especially true at p 400 MeV/c potential, as defined in (Gezerlis et al., 2014). A correla- ≈ where there is a minimum in the s-wave. In the Argonne tion hole is seen to occur for R0 = 1 fm, but is much less V18 potential (Wiringa et al., 2014b) the d-wave compo- deep for R0 = 1.2 fm. The use of such a soft potential is nent is due to the tensor force. The combination of d- not suitable in the present experiment-based high-scale and s-waves leads to a “broad shoulder” in the deuteron context. Furthermore, this soft potential predicts erro- 10

3 neous nucleon-nucleon phase shifts for the D1 partial A justification of this factorization is presented in wave, and also for lab energies greater than 250 MeV. Sect. IX. In momentum space, the existence of this universal cor- The coordinate-space correlation holes (Fig. 5) give relation hole translates into nucleon momentum distribu- similar NN relative (prel) momentum distributions (at tions n (p) that are significant at large momentum (p large p in all nuclei. Exact calculations with the AV18 A ≥ rel p ) and that are similar for all nuclei, n (p) n (p), potential for 4He, show that, at small p there is a min- F A ∝ d tot at these large momenta (Alvioli et al., 2013; Ciofi degli imum in prel for pp pairs at prel = 400 MeV/c. This is Atti and Simula, 1996a; Frankfurt and Strikman, 1981a, because, at small ptot, the pp pair must be in a relative 1988b). Frankfurt and Strikman realized that these could s-state which has a minimum at prel = 400 MeV/c, just be measured with hard probes (see Section II). like in deuterium. For np pairs, this minimum is filled in Ciofi degli Atti and Simula (Ciofi degli Atti and Sim- by the d-wave caused by the short range pion-exchange ula, 1996a; Ciofi degli Atti et al., 1991b) used this sim- tensor force (Wiringa et al., 2014b). ilarity to model the nucleon spectral function P (p,E) Thus, the combination of the minimum in the s-wave (the joint probability to find a nucleon in a nucleus with momentum distribution at p 400 MeV/c and the filling momentum p and removal energy E) for all nuclei in of this minimum by the d≈-wave pion-exchange tensor force, leads to the expected dominance of np correlated P (p,E) = Ψ b†(p)δ(E H)b(p)Ψ , (5) h | − i pairs over nn and pp pairs at 300 p 500 MeV/c. where Ψ represents the nuclear wave function and spin, This ratio of np to pp pairs should≤ decrease≤ at relative | i isospin and nuclear (A) labels are suppressed for simplic- momentum significantly greater than 400 MeV/c, the s- ity. The momentum density n(p) is given by wave minimum (as we will discuss Section II). Short-range correlations in light nuclei have been n(p) = dEP (p,E). (6) examined recently theoretically from several points of Z view (Ciofi degli Atti, 2015; Feldmeier et al., 2011; Rios These authors write et al., 2014; Ryckebusch et al., 2015; Vanhalst et al., 2011, P (p,E) = P0(p,E) + P1(p,E) (7) 2012; Weiss et al., 2015; Wiringa et al., 2014a). One consistent finding of such work is the dominance of np where the subscript zero refers to values of E correspond- deuteron-like pairs (ST = 10) over other pairs at high ing to low-lying intermediate excited states and the sub- momentum. script one refers to high-lying continuum states that are These facts described in this sub-section lead to an ef- caused by the short-ranged correlations. Therefore one fective description of nuclei in momentum space as having also has n(p) = n (p) + n (p), where n (p) is associated 0 1 1 two important regions: (1) a mean-field region (k p ), with the high momentum caused by short-ranged corre- F which accounts for about 80% of the nucleons,≤ where lations. n (p) is typically dominant for p < 250 MeV/c 0 the many-body dynamics result in single nucleons mov- or so and n (p) becomes dominant for larger values. Fur- 1 ing under the influence of an effective potential created thermore, n (p) is almost independent of A at p > 400 1 by the residual A 1 system and (2) a high-momentum MeV/c; they attribute this to NN correlations. See Fig. region (p p ),− which accounts for about 20% of the 4. It should be noted that for 3He Ref. (Ciofi degli Atti F nucleons (but≥ 70% of the kinetic energy (Benhar et al., et al., 1991b) showed that the proposed model spectral 1989; Polls et al., 1994)), where nucleons are predomi- function agrees with the one obtained by direct calcula- nantly in the form of pn-SRC pairs, having a very weak tion. interaction with the residual A 2 system. As noted SRC pairs are conventionally defined in momentum above, it is possible to use unitary− transformations to space as a pair of nucleons with high relative momentum derive a low-momentum effective interaction that weaken and low center of mass (c.m.) momentum, where high the strong short-ranged correlations present in the origi- and low are relative to the Fermi momentum of medium nal interactions. However, the one and two body density and heavy nuclei. Thus the most prominent effect of SRC operators also need to be transformed. It is necessary to will be to populate high-momentum states in the nuclear include three or more body effects to obtain accurate re- momentum distribution. As conventional mean-field the- sults with these soft interactions (Feldmeier et al., 2011). ories predict only a very small high-momentum tail, the This approach complicates the analyses of experiments. effect of SRCs there should be substantial. Formally, one To summarize, the high momentum nucleons in nuclei needs the two-nucleon momentum density, n(p , p ) (see 1 2 are mainly due to 2N-SRC and are therefore associated Section IX), where p = p + p and p = 1 (p p ) tot 1 2 rel 2 1 2 with high density fluctuations in the nucleus. In what are the center of mass and relative momenta of the− two follows (see Section III.D) we will examine the hypothesis nucleons. Studies of spectral functions show that at large that these temporary high density/large momentum ‘hot values of p , the two-nucleon momentum density factor- rel spots’ are the sites where the nucleon internal structure izes: is modified and the EMC effect is created. First, we n(ptot, prel) = n(ptot)n(prel). (8) will present the experimental evidence for short range 11 correlations. GeV. This allows the experiments to be sensitive to the internal structure of a proton or neutron and avoid the influence of individual nucleon resonances, which cause II. Hard scattering and Short-Range Correlations the cross section to fluctuate rapidly with W . Early studies at the high energy facilities (SLAC and A. Hard Reactions CERN) measured DIS for 5 Q2 50 GeV/c2 and ≤ ≤ found that the ratios of DIS cross sections for 0.3 xB In optics the resolving power is the minimum distance 0.7 are largely independent of Q2 (Norton, 2003).≤ The≤ at which an imaging device can separate two closely newer JLab experiments used lower lepton energies (typ- spaced objects. This is normally proportional to the ically 4 5 GeV) and therefore lower Q2, 4 Q2 6 wavelength of the light. The smaller the wavelength, the GeV2 (Seely− et al., 2009a). The higher-energy≤ SLAC≤ and better the resolution. CERN measurements required W 2 GeV. However the We often scatter particles to try to resolve the internal lower-energy JLab data required only≥ W 1.4 GeV. structure of a complex target. The sizes of the target and ≥ For inclusive (e, e ) QE scattering, there are again only its constituents define the required resolving power. For 0 two independent kinematical variables, normally chosen example, to observe the nucleus of an atom one needs to be Q2 and x . However, in addition to making sure a spacial resolution of about 10 fm, to observe nucleons B that the resolving power is sufficient, we also need to in nuclei one needs a resolution of about 1 fm, and to optimize the kinematics to select scattering from high- observe the partonic structure of a nucleon one needs momentum nucleons in the nucleus and to reduce the sub-fermi resolution. effects of non-single-nucleon currents. In order to resolve The spacial resolution of a scattering experiment is de- nucleons in SRC pairs, measurements are typically made termined by the de Broglie wave length (λ) of the probe 2 2 at Q > 1.5 (GeV/c) . Large (p > pF ) minimum ini- (scattering particle) and the momentum transfer of the tial momentum of the struck nucleon (assuming no final reaction (q). We define as ’hard’ a process that fulfills the state interactions) can be selected at Q2 > 1.5 (GeV/c)2 following conditions: λ R and qR 1, where R is the by choosing either x 1.5 or x 0.6 (see Section size of the target or the structure to be studied. In prac- B ≥ B ≤ II.C). xB 1 is preferred, so that the energy transfer is tice, we shall see that the results of measurements can smaller, inelastic≥ processes (resonance production, meson be interpreted as observing a hard reaction even though exchange currents [MEC] and isobar configurations [IC]) these kinematic conditions are not always rigorously met. are suppressed, and the reaction is more sensitive to the Another important lepton-scattering length scale is the nuclear momentum distribution. Increasing Q2 further coherence length, or Ioffe length (Gribov et al., 1966; suppresses MEC contributions. The inclusive QE scat- 2 0.4 fm Q2 Ioffe, 1969): lI = , where xB = . Here tering data discussed in Section II.C were measured at MxB ≈ xB 2Mν M is the nucleon mass, Q2 is the negative of the square of x 1.5. B ≥ the virtual exchanged photon four-momentum, and ν is In exclusive and semi-exclusive reactions, (e, e0p) and its energy. This length is the typical distance between the (e, e0pN), large initial nucleon momenta can be selected absorption and re-emission of the virtual photon. This directly and the xB restrictions can be relaxed (see Sec- length must be short enough to resolve the relevant inter- tion II.B). nucleon distance scales of the order of a fermi. Thus, we will focus on the region xB > 0.3 where valence quarks are dominant and the sea is almost invisible. B. Exclusive Scattering In this paper we are dealing with two reactions and the connection between them. Deep inelastic scattering The study of SRCs using exclusive reactions has a long (DIS) attempts to resolve the partonic structure of nucle- history that extends beyond the scope of this review. ons and quasielastic scattering (QE) attempts to resolve Here we focus only on exclusive measurements performed the nucleonic structure of nuclei. These reactions have with high energy probes and large momentum transfer different required resolutions and hence different kine- (hard reactions). See (Kelly, 1996) and references therein matical conditions to achieve them. for a review of the older measurements. We use the term For (e, e0) DIS reactions, which are typically mea- exclusive to refer to measurements in which, in addition 2 sured as a function of xB = Q /2Mν for xB < 1, to the scattered probe particle, two knocked-out nucleons there are two important parameters, the 4-momentum are measured in the final state. transfer squared of the virtual photon, Q2, and the in- In the context of SRC studies, exclusive reactions are variant mass of the virtual photon plus struck nucleon, hard processes in which a probe scatters from one nu- 2 2 2 W = M + 2Mν Q . Since xB,Q , and W are all cleon in an SRC pair and all particles emitted in the functions of the same− two variables, only two are indepen- final state (e.g., the scattered probe and both nucleons p dent. For the inelastic scattering to be considered deep of the pair) are detected. The energy of the probe and (the “D” in DIS), experiments typically require W 2 the momentum transfer must be large enough so that the ≥ 12 probe interacts with a single, high-momentum (pi > pF ) Glauber approximation to calculate to the effects of FSI nucleon in the pair. If the pair was at rest (pcm = 0) and and to select kinematics to minimize their effects, either neither nucleon rescattered as it left the nucleus, then in the measured cross sections or the kinematical distri- the struck nucleon’s correlated partner would recoil with butions. momentum p = p . This back-to-back angular cor- Specifically, at Q2 1.5 2 (GeV/c)2 and x 1 (or 2 − i B relation between the initial momentum of the knocked proton scattering experiments≥ − at t , u , s 2 GeV/c≥ 2) out nucleon and the momentum of the recoil nucleon is Glauber calculations show that| the| | outgoing| | | ≥ nucleons a clear experimental signature for exactly two nucleons predominantly rescatter from each other and not from being involved in the interaction. We note that these re- the residual A 2 system (Arrington et al., 2012b; actions can be analyzed in terms of the decay function Ciofi degli Atti and− Simula, 1996a; Frankfurt and Strik- introduced by (Frankfurt and Strikman, 1988b). man, 1981a, 1988b; Frankfurt et al., 1993). This implies However, other reaction mechanisms can also involve that certain quantities such as the total pair momen- two nucleons, leaving the residual A 2 nucleus almost at tum, pcm, and pair isospin structure are insensitive to rest. The probe can scatter from one− nucleon, which can rescattering while other quantities like the pair relative rescatter from a second (FSI), the probe can scatter from momentum, prel, are very sensitive to rescattering and a meson being exchanged between two nucleons (MEC), thus cannot be reliably extracted from the experimental or the probe can excite the first nucleon which can then data, see (Frankfurt et al., 1997; Shneor et al., 2007) for de-excite via interaction with a second nucleon (IC). Dis- details. The contribution of Meson Exchange Currents entangling these competing and interfering effects can be (MEC) and Isobar Currents (IC) are also minimized at difficult. It is important to realize that the effects of MEC high Q2 and x 1. B ≥ and IC are dramatically decreased by choosing kinemat- The first exclusive hard two nucleon knockout exper- 2 12 12 ics with xB > 1 and with larger values of Q . The effects iments, measuring the C(p, 2pn) and C(e, e0pN) re- of FSI can also be dramatically decreased by (a) choos- actions, were done at BNL and JLab, respectively (Pi- ing kinematics where the relative momentum of the two asetzky et al., 2006; Shneor et al., 2007; Subedi et al., final-state nucleons is large and (b) avoiding kinematics 2008; Tang et al., 2003). These experiments scattered 5 where the opening angle between the two outgoing nu- - 9 GeV/c protons (BNL) and electrons (JLab) off high cleons is 70 90◦. (Non-relativistically, when one billiard initial momentum (300 p 600 MeV/c) protons in − i ball scatters from a second billiard ball at rest, the open- 12C and looked for a correlated≤ ≤ recoil nucleon emitted in ing angle in the final state is 90◦.) the direction of the missing momentum. The JLab exper- The detection of the outgoing nucleons in exclusive re- iment measured both proton and neutron recoils, whereas actions provides complementary information to the in- the BNL experiment only measured recoiling neutrons. clusive reactions discussed above below. By detecting Both experiments measured at large momentum transfer the struck nucleon at large p and looking for the re- (Q2 2 (GeV/c)2), which suppressed competing reac- miss ≈ coil partner nucleons, exclusive measurements can mea- tion mechanisms and largely confined FSI to be between sure the fraction of high-momentum nucleons belonging the nucleons of the pair. to SRC pairs. They can also extract information on the The main results of the 12C measurements are shown in SRC pair isospin structure and pcm distribution, as well Figs. 6, 7 and 8. Figs. 6 and 7 show the angular correla- as their A and momentum dependence. tion between the momentum vector of the recoil nucleons This additional information however comes at the price and the reconstructed initial momentum of the knocked- of increased sensitivity to FSI. FSI can be generally split out proton. For the BNL data, the angle is shown as a into two main contributions: re-scattering between the function of the recoil neutron momentum. Two distinct nucleons of the pair, and re-scattering between the nu- regions are visible: below the Fermi momentum where no cleons of the pair and the residual A 2 system. Rescat- angular correlation is observed, and above the Fermi mo- tering between the nucleons of the− pair will alter the mentum where a clear back-to-back correlation is seen. measured relative momentum but leave pcm unchanged. The width of the recoil nucleon opening angle distribu- Rescattering between the nucleons of the pair and the tion allowed extracting the pair c.m. motion; this motion residual A 2 system will change the momentum of the can be described by a Gaussian distribution in each direc- outgoing nucleons− and “attenuate” them. The attenua- tion, with σ = 143 17 (BNL) and σ = 136 20 (JLab). ± ± tion of the nucleons as they traverse the nucleus is usually These values are also in overall agreement with theo- referred to as the ’nuclear transparency’ and limits the retical calculations (Ciofi degli Atti and Simula, 1996b; spatial region probed in the experiment to the outer part Colle et al., 2014). The electron and proton reactions are of the nucleus. It can be calculated in the Glauber ap- characterized by completely different operators and FSI proximation (for large enough nucleon momentum). The mechanisms; therefore the agreement of their c.m. mo- momentum changes also affect the measured kinematical mentum distributions validates the consistent treatment distributions. Here the use of high momentum trans- of FSI in these measurements. fer, as required for hard reactions, also allows using the For example, for proton induced reactions the effec- 13

FIG. 7: Distributions of the relative angle (γ) between the reconstructed initial momentum of the knockout 12 proton and the recoil nucleon. Results for C(e, e0pp) events from JLab at kinematics corresponding to scattering off 500 MeV/c initial momentum protons. ∼ 12 Figure from (Shneor et al., 2007). 8 4 102 0 Events 20 10 10 SRC Pair Fraction (%) Fraction Pair SRC 0 0.3 0.4 0.5 0.6 -1 -0.5 0 0.5 1 Missing Momentum [GeV/c] ��� � FIG. 8: The ratio of 12C(e, e pN) double knockout FIG. 6: Distributions of the relative angle (γ) between 0 events to 12C(e, e p) single knockout events, shown as a the reconstructed initial momentum of the knockout 0 function of the reconstructed initial (missing) proton and the recoil neutron. Results for 12C(p, 2pn) momentum of the knocked-out proton from the events from BNL, shown as a function of the 12C(e, e p) reaction. Triangles and circles mark momentum of the recoil neutron (a) and for events with 0 12C(e, e pn)and 12C(e, e pp) events, respectively. The recoiling neutron momentum greater than (b) and less 0 0 square shows the 12C(e, e pp)/12C(e, e pn) ratio. A clear than (c) k = 225 MeV/c. Note the transition from an 0 0 F dominance of 12C(e, e pn) events is observed, evidence isotropic distribution to a correlated one at about 0 of the tensor nature of the nucleon-nucleon interaction k = 225 MeV/c. Figures adapted from (Piasetzky F in the measured momentum range. The pie chart on the et al., 2006; Tang et al., 2003). right illustrates our understanding of the structure of 12C, composed of 80% mean-field nucleons and 20% SRC pairs, where the latter is composed of 90% ∼ tive nuclear density is smaller than for electron induced np-SRC pairs and 5% pp and nn SRC pairs each. reactions due to absorption effects that prefers scatter- Figure adapted from (Subedi et al., 2008). ing from the edge of the nucleus. The overall agreement between the results obtained using different high energy hadronic and leptonic probes at very different momen- ton knockout events and the ratio of proton-neutron to 2 2 tum transfer (2 GeV and 5 GeV ) strongly supports the proton-proton two-nucleon knockout events. The ratios interpretation that in these reactions the projectiles in- are all corrected for finite acceptance effects and shown as teract with one nucleon of the SRC. Note also that the a function of pmiss, the reconstructed initial momentum saturation of the recoil channels by neutron and protons of the knocked out protons for 300 pmiss 600 MeV/c. puts a strong limit on the admixture of non-nucleonic The ratio of two nucleon knockout to≤ single≤ proton knock- degrees of freedom in SRCs. out is directly related to the fraction of high-momentum Fig. 8 shows the extracted ratio of two nucleon knock- protons that are in SRC pairs. As can be seen, within out (proton-neutron and proton-proton) to single pro- statistical uncertainties of about 10%, all single nucleon 14

30 40 4He(e,e’pn) kinematics set out to better constrain the importance 30 md 30 Counts 40 4He(e,e’pn) 30 20 m 30 d 25 Counts of the tensor part of the NN interaction at short dis-

Counts 10 20 25 0 Counts 20 10 tance, and extend the experimental data to larger initial 01.8 1.85 1.9 1.95 2 2.05 Missing Mass [GeV/c2] 20 151.8 1.85 1.9 1.95 2 2.05 2 Counts 20 Missing4 Mass [GeV/c ] momenta, 400 p 800 MeV/c (Korover et al., 2014). 15 m2n He(e,e’pp) i Counts

Counts 1510 4 ≤ ≤ 15 m2n He(e,e’pp) 10 Counts 105 At these higher momenta, the scalar repulsive core of the 0 10 5 1.8 1.85 1.9 1.95 2 2.05 10 0 nucleon-nucleon interaction is expected to dominate over 5 Missing Mass [GeV/c2] 2 1.8Missing 1.85 1.9 Mass 1.95 [GeV/c 2 2.05 ] 5 Missing Mass [GeV/c2] the tensor part, increasing the fraction of pp-SRC pairs. 0 4 00 The He nucleus was chosen to further reduce FSI and −1 −0.99 −0.98 −0.97 −0.96 −0.95 −0.94 allow for comparisons with detailed ab-initio few-body -1.00−1 −0.99 -0.98−0.98 −0.97 -0.96−0.96 −0.95 -cos(0.94−0.94γ) cos(γ) calculations. The results of this measurement are shown ��� � in Figs. 9 and 10. The two-nucleon opening angle distribution for 4He FIG. 9: (color online) The distribution of the cosine of (see Fig. 9) is very similar to that for C (see Fig. 7). The the opening angle γ between pmiss and precoil for the reconstructed missing mass distribution peaks at small 4 He(e, e0pn) reaction. The solid curve is a simulation of missing mass for both pp- and np-SRC pair knockout. scattering off a moving pair with a c.m. momentum As can be seen, there is a peak at back angle, associated distribution having a width of 100 MeV/c. The insets with a breakup of 4He into a SRC pair and a residual 2N show the missing-mass distributions. Figure reproduced system with low excitation energy. As with the 12C mea- based on (Korover et al., 2014). surements, the width of the opening angle distribution is due to the c.m. motion of the SRC pairs which was found to be consistent with a Gaussian in each direction with a width of 100 20 MeV/c. 0.3 ± The extracted 4He pp/np SRC pairs ratio increases with pmiss for pmiss > 400 MeV/c (see Fig. 10). The 0.2 measured ratios are consistent with ab-initio Variational Monte Carlo (VMC) calculations of Ref. (Wiringa et al., pp / np 0.1 2014b) integrated over c.m. momentum up to about 300 MeV/c, which is consistent with the measured width of the c.m. momentum distribution. At higher c.m. mo- 0 mentum, the two body momentum distribution is dom- 0 200 400 600 800 q [MeV/c] inated by large contributions from un-correlated pairs (Weiss et al., 2016). Similar results were also obtained by different calculations (Alvioli et al., 2016; Ryckebusch FIG. 10: (color online) The measured pp to pn ratio as et al., 2015). function of the proton missing momentum (labelled q) The importance of tensor correlations was further (Korover et al., 2014) compared to calculations of the shown by measurements of the pp to pn ratio in two-nucleon momentum distribution (Wiringa et al., 3He(e, e pp)n measured using the CLAS detector at JLab 2014b) integrated over various ranges of the c.m. 0 (Baghdasaryan et al., 2010). They measured the relative momentum (Weiss et al., 2016). The data is shown as a and total momentum distribution of pp and pn pairs in function of the nucleon momentum and the calculations 3He by detecting events where the virtual photon was are shown as a function of the pair relative momentum. absorbed on one nucleon and the other two (spectator) The two are equivalent for low c.m. momentumof the nucleons were also detected. Fig. 11 shows the ratio pair but differ at large c.m. momentum. Figure adapted of pp to pn pairs in 3He as a function of the pair to- from (Weiss et al., 2016) tal (e.g., center-of-mass) momentum for two pair rela- tive momentum ranges, 300 p 500 MeV/c and ≤ rel ≤ 400 prel 600 MeV/c. The first range is centered at knockout events at 300 p 600 MeV/c were accompa- ≤ ≤ ≤ i ≤ the s-wave minimum at 400 MeV/c where the effects of nied by the emission of a recoil nucleon. The proton-to- tensor correlations are expected to dominate; the second neutron recoil ratio was found to be approximately 1:10, is not. For prel centered at 400 MeV/c, the pp to pn ratio which corresponds to 20 times more np-SRC pairs than is very small at ptot 100 MeV/c and consistent with 12 12 ≤ pp-SRC pairs in C (Subedi et al., 2008). This observed the C(e, e0pN) measurements discussed above. For prel proton-neutron pair dominance was associated with the centered at 500 MeV/c, the pp to pn ratio at ptot 100 dominance of the tensor part of the nucleon-nucleon in- MeV/c is significantly larger, consistent with the≤ ex- teraction at these initial moments (Sargsian et al., 2005a; pected decreased dominance of tensor correlated pairs Schiavilla et al., 2007). at this higher relative momentum. At large ptot, the pp 4 A follow-up measurement of He(e, e0pN) in similar to pn ratio is 0.5, consistent with simple pair counting. 15

Encouraged by these results, the latest exclusive mea- surements extended to medium and heavy nuclei (12C, 27Al, 56Fe, and 208Pb), where the persistence of np-SRC dominance was still unproven (Hen et al., 2014d). In this experiment, the A(e, e0pp) and A(e, e0p) reactions were measured at similar kinematics to the previous 4He and 12C measurements, covering a reconstructed initial pro- ton momentum range of 300 pi 600 MeV/c. The analysis assumed that, in these≤ nuclei,≤ the reaction is still dominated by scattering off SRC pairs and extracted the relative fraction of np- and pp-SRC pairs. Fig. 12 shows that SRC pairs are predominantly np-SRC pairs even in 3 FIG. 11: The ratio of pp to pn pairs in He(e, e0pp)n. heavy neutron rich nuclei. The solid and star points show the ratio for 300 prel 500 MeV/c and for 400 prel 600 MeV/c≤ respectively,≤ as a function of the≤ total≤ (e.g., C. Inclusive Scattering center-of-mass) momentum of the pair. The pp/pn ratio is much less than one for small ptot, increasing to the Inclusive quasi-elastic electron scattering allows prob- pair counting ratio of 0.5 at large ptot. The ratio at ing the momentum distribution of nucleons in the nu- small ptot is about 0.1 for 300 prel 500 MeV/c, and cleus. Elastic scattering from a nucleon at rest occurs at ≤ ≤ Q2 about 0.25 for 400 p 600 MeV/c. The solid line fixed kinematics, ν = . This corresponds to xB = 1. ≤ rel ≤ 2M shows a calculation by Golak for 300 prel 500 If all of the struck nucleons in a nucleus were at rest, the MeV/c which neglects rescattering of the≤ struck≤ nucleon cross section would show a pronounced narrow peak– the but includes the reinteraction of the two nucleons in the quasi-elastic peak. SRC pair. The dashed line (blue online) shows the 3He This peak is broadened by nucleon motion for elec- momentum distribution integrated over the tron scattering from bound nucleons. In order to study experimental acceptances. From (Baghdasaryan et al., nuclear momentum distributions, experiments typically 2010). focus on the low energy transfer side of the QE peak, or xB 1. In this case the initial momentum of the struck≥ nucleon must be in the opposite direction from the momentum transfer so that the final momentum of the struck nucleon pf = q + pmiss (in the absence of fi- nal state interactions or FSI, prmmiss = prminit) is less than the momentum transfer. As the energy transfer de- creases, the final momentum of the struck nucleon must decrease and therefore the minimum initial momentum of the struck nucleon must increase. The quasielastic inclusive electron scattering (e, e0) FIG. 12: The relative fraction of np and pp SRC pairs cross section can be written in terms of a function F 2 2 (excluding nn pairs) derived from A(e, e0p) and that depends on (Q , y) rather than (Q , ν) (Day et al., A(e, e0pp) measurements on a range of nuclei. From 1987b): (Hen et al., 2014d). 2 d σ(q, ν) 2 q = F (y, Q )(Zσp + Nσn) dνdΩ M 2 + (y + q)2 The points at 300 p 500 MeV/c are consistent (9) ≤ rel ≤ p with a calculation by Golak (Golak et al., 1995) which where σp,n are the elastic electron scattering cross sec- neglects rescattering of the struck nucleon but includes tions from a bound nucleon, the last term is the Jacobian the reinteraction of the two nucleons in the SRC pair. dy/dν, and y = y(Q2, ν) is the minimum momentum of The combined results of the 3He, 4He and 12C measure- the struck nucleon (assuming that the residual A 1 sys- − ments indicate that for 300 pi 500 MeV/c, nucleons tem is unexcited) (Arrington et al., 2012a; Day et al., are predominantly part of pn≤-SRC≤ pairs as predicted by 1990). dominance of the tensor part of the NN interaction at Non-relativistically, y is the component of the struck short distances. At higher initial momentum, the contri- nucleon’s initial momentum (pmiss) in the direction of q. bution of pp-SRC pairs seems to increase by a factor of 2 The cross section at fixed y then includes an integral over – 3, possibly due to larger contributions from the scalar the perpendicular components of pmiss. Relativistically, repulsive core of the NN interaction. it is a little more complicated. y is determined from 16

energy conservation, assuming no FSI and that the A 1 nucleus recoils with momentum y: − 2 2 1/2 2 2 1/2 ν + MA = (M + (q + y) ) + (MA 1 + y ) (10). − 2 At the QE peak, ν = Q /(2M), xB = 1, and y = 0. As ν decreases, xB increases and y decreases. By selecting 2 xB or y (at fixed Q ), we can select the minimum initial momentum of the struck nucleon (see Fig. 13). At large enough Q2 the function F (y, Q2) scales and depends only on y (Ciofi degli Atti et al., 1991a). The nucleon momen- tum distribution, n(p = y), can be calculated from the derivative of the scaling function, dF (y)/dy, at large Q2: 1 dF (p) n(p) = − . (11) 2πp dp Fig. 14 shows the deuteron momentum distribution de- rived in this manner. The original y-scaling model discussed here assumes a that the residual A 1 nucleus is in a low-lying state. This FIG. 13: The minimum momentum of the struck procedure neglects− the possibly large excitation energy of nucleon in inclusive (e, e0) scattering as a function of the residual nucleus, which is an important feature of the xB. The top panel shows the minimum momentum for spectral function. As a result, for scattering by a nucleon deuterium for a variety of momentum transfers and the in a SRC, the same internal momenta corresponds to a bottom panel shows the minimum momentum for a very different values of y for different nuclei. variety of nuclei at Q2 = 2 GeV2. The residual A 1 − Another approach is to compare the momentum distri- system is assumed to be in its ground state. Figure butions in different nuclei with reduced uncertainties by from (Egiyan et al., 2003). taking ratios of cross sections. We write the momentum density in terms of the light cone variable αtn for the interacting nucleon belonging to the correlated pair, ν q + 2M α = 2 − (1 + 1 4M 2/W 2). (12) tn − 2M − Using this variable, the cross sectionp ratios do not de- pend on Q2 in the kinematic range of the SLAC exper- iments (Day et al., 1987a). The onset of the plateaus discussed below occur for the same values of αtn but for slightly different values of xB. Then the ratios of cross sections can be expressed in terms of the light-cone spectral function at large Q2 and 1.5 < xB < 2 as (Frankfurt et al., 1993): σ (x ,Q2) ρ (α , p )d2p n (p) A1 B = A1 tn t t A . (13) σ (x ,Q2) ρ (α , p )d2p ≈ n (p) A2 B R A2 tn t t D Thus this ratio of crossR sections should be a function of 2 αtn only, which, since it is a function of (Q , xB), is di- FIG. 14: Momentum distribution of the deuteron. rectly related to y, the minimum momentum of the struck Points show the results extracted from the experimental nucleon. The approximate equality shown in Eq. (13) holds for 1.3 α 1.7 and p > p . The second scaling function F (y) at four different momentum ≤ tn ≤ F transfers (Fomin et al., 2012a). Curves show the approximate equality appearing in Eq. (13) is obtained 1 αtn calculated momentum distributions using three different using the relation p M | − | . Measured ra- √αtn(2 αtn)) | | ≈ − NN potentials Paris (Lacombe et al., 1981), tios should be less sensitive to the influence of final state Nijmegen (Stoks et al., 1994b) and Argonne interactions, as discussed below. Nevertheless, the accu- V14 (Wiringa et al., 1995a). Figure from (Fomin et al., racy of replacing cross section ratios by ratios of densities, 2012a), which uses k for momentum instead of p. as shown in Eq. (13), needs to be studied further. Fur- thermore, as yet there is no separate calculation of the 17

momentum distributions for nucleus A and for deuterium should be almost identical. This similarity should show up as a plateau in the per-nucleon cross section ratio of the two nuclei. Fig. 15 shows a cartoon of this process. The cross section ratio of nucleus A to deuterium or to 3He has been measured at SLAC (Frankfurt et al., 1993) and at Jefferson Lab (Egiyan et al., 2003, 2006a; Fomin et al., 2012a). They have all observed a plateau in the cross section ratio at Q2 > 1.4 GeV2 and from FIG. 15: A cartoon of electron quasielastic scattering 1.5 x 1.9. See Fig. 16. This corresponds to y from a nucleon in deuterium (left) and from a nucleon B p ≤ = 275≤ 25 MeV/c, which is slightly larger than≥ in a SRC pair in a heavier nucleus (right). The labels thresh the Fermi momentum± in medium and heavy nuclei. The Γ and Γ refer to the deuteron and nuclear vertex NN A value of Q2 is large enough to ensure that contributions functions respectively. from uncorrelated nucleons (with momentum governed by the size of the nucleus) are negligible. numerator term of Eq. (13), i.e., the basic nuclear cross However, in order to relate these observed plateaus to the ratio of momentum distributions in the different section for the (e, e0) reaction at large values of xB. Physics at large values of x . The next step is to nuclei, we need to take into account the final state inter- B actions (FSI) of the nucleon with its correlated partner use the inclusive (e, e0) cross section to look for the effects 2 2 of SRC pairs in nuclei by choosing kinematics where mean and with the residual system. For Q > 1 GeV and field nucleons cannot contribute to the reaction. This 0.35 < ν < 1 GeV, the space-time physics (Frankfurt et al., 2008, 1993) of the inclusive process tells us that is done by using xB > 1. Just as conservation of four- momentum ensures that x = 1 is the kinematic limit for final state interaction effects occur predominantly within B the two-nucleon correlation. Such effects are indepen- scattering from a single nucleon, xB = 2 is the kinematic limit for scattering from a cluster of two nucleons and dent of the nuclear target, and should be small for large values of ν. The relevant values of ν are large enough xB = 3 is the kinematic limit for scattering from a three- nucleon cluster. so that final state interactions within the pair are not very important (Frankfurt and Strikman, 1981a, 1988b). As a result, we can expand the (e, e0) cross section into pieces due to electrons scattering from nucleons in 2-, 3- Therefore, the effects of FSI will be approximately the and more-nucleon SRC (Frankfurt and Strikman, 1981a, same for high momentum nucleons in deuterium and in 1988b; Frankfurt et al., 1993) heavier nuclei and will predominantly cancel in the cross section ratios. A Some measurements, e.g., (Egiyan et al., 2003, 2006a), 2 2 σ(xB,Q ) = aj(A)σj(xB,Q ), (14) applied isoscalar corrections to the ratios of Eq. 15 j=2 X (i.e., they corrected for the unequal electron-proton and electron-neutron cross sections). Since the discovery of where σ (x ,Q2) = 0 for x > j and the a (A) are j B B { j } pn-dominance in SRC pairs (see Section II.B), these cor- proportional to the probability of finding a nucleon in a rections are no longer applied (Fomin et al., 2012a). j-nucleon cluster. This is analogous to treating nuclear The flatness of the cross section ratio plateau at Q2 > structure in terms of independent nucleons, independent 2 1.4 GeV and from 1.5 xB 1.9, and its approxi- nucleon pairs, etc. This expression is based on the lack mate independence of Q2≤in this≤ region where SRC ef- of interference between amplitudes arising from scatter- fects dominate indicates the similarity of the momentum ing by clusters of different nucleon number that occurs distributions in the two nuclei for p > pthresh and the because the important final states are different. Its im- validity of the expansion in Eq. 14. The onset of the portance lies in the fact that in a given kinematic region 2 2 plateau at xB = 1.5 for Q > 1.4 GeV indicates that the ratio of cross sections can be used to determine in- the momentum distributions become similar at a thresh- formation about short-ranged correlations. old momentum of pi = pthresh = 275 25 MeV/c (Egiyan If we consider only the a2 term, then we can write ± et al., 2003). The height of the plateau, a2(A), indicates 2 σ (x ,Q2) the relative probability that a nucleon in nucleus A has a (A) = A B . (15) 2 A σ (x ,Q2) high momentum (p > pthresh) relative to a nucleon in d B deuterium. This approximation should be valid for 1.5 < xB 2. In a naive model, this relative probability for a nucleon The effect of neglecting clusters of three or more nucleons≤ to have high momentum equals the relative probability has never been studied. that it belongs to an NN SRC pair. However, even if If the momentum distribution for y > pfermi is domi- all nucleons with p > pthresh belong to an NN SRC pair nated by nucleons in SRC pairs, then| | we expect that the as evident from the exclusive measurement (see Section 18

Fomin et al., 2012a; Vanhalst et al., 2012). This was found to be about a 20% effect in Fe. Thus, while the ratio of the proportion of high-momentum nucleons in Cu to deuterium is a2(A) = 5.4 0.1, the ratio of the number of SRC NN pairs in Cu to± deuterium (using the Fe correction factor) is about 20% less, R2N = 4.3 0.3 (Fomin et al., 2012a). ± Multiplying the 4% probability for a nucleon in deu- terium to have momentum p > pthresh by the measured ratios in the plateau region (a2(A)), as indicated by Eq. (13), gives us the probabilities for a nucleon to have high momentum in 4He, C, Fe/Cu and Au to be 14%, 19%, 21% and 21% respectively (Fomin et al., 2012a; Hen et al., 2012). Thus, the existence of a plateau in the measured per- nucleon cross section ratios of various nuclei to deuterium 3 2 2 or He at Q > 1.4 GeV and 1.5 xB 1.9 shows that the momentum distributions of all≤ nuclei≤ at high momen- tum are similar and are thus dominated by 2N-SRC, that the threshold for “high momentum” is pthresh = 275 25 MeV/c, and that the probabilities for nucleons in nuclei± to have high momentum range from 4% in deuterium to 21% in heavy nuclei. While the inclusive scattering cross section ratios of carbon and iron to 3He measured by Egiyan are flat for 1.5 < xB < 2, the ratios of carbon, copper and gold to deuterium measured by Fomin appear to slope upwards slightly. This is not due to the choice of nucleus in the de- nominator, since the ratio of 3He to deuterium measured by Fomin is flat. This is also probably not due to c.m. motion effects as these are simililar for 4He and 12C (Ko- rover et al., 2014; Shneor et al., 2007; Tang et al., 2003), which do show flat ratios, and are expected to be the same for heavy nuclei (Ciofi degli Atti and Simula, 1996b; Colle et al., 2014). This might be due to differences in kinematics. The Egiyan data covers 1.4 < Q2 < 2.6 GeV2 (concentrated at the lower values), while the Fomin data was taken at Q2 = 2.7 GeV2. At Q2 = 1.5 GeV2 and 1.5 xB 2, the minimum momentum of the struck nucleon≤ ranges≤ from 250 to 500 MeV/c, covering the ex- FIG. 16: Inclusive per-nucleon cross section ratios of pected region of tensor force dominance. However, at 2 2 (top) nuclei to 3He from (Egiyan et al., 2006a) at Q = 2.7 GeV , the minimum momentum of the struck 1.4 < Q2 < 2.6 GeV2 and (bottom) nuclei to deuterium nucleon ranges from 320 to 700 MeV/c, where central at Q2 = 2.7 GeV2 (Fomin et al., 2012a). Figures correlations could become important. It would be useful 2 adapted from (Egiyan et al., 2006a) (top) and from to measure the Q dependence of the cross section ratios (Fomin et al., 2012a) (bottom). in future SRC measurements.

D. Universal Properties of Short Range Correlations in II.B), we still need to consider the effects of pair motion. Nuclei The high momentum NN pair in the deuteron has cen- ter of mass momentum pcm = 0. The non-zero center of The combined results from the exclusive and inclusive mass momentum distribution of the pair in heavier nu- measurements described in Sections II.B and II.C lead to clei will smear the high-momentum tail of the nucleon a universal picture of SRC pairs in nuclei. In the conven- momentum distribution, increasing the cross section ra- tional momentum space picture, the momentum distri- tio in the plateau region (Ciofi degli Atti et al., 1991b; bution for all nuclei and nuclear matter can be divided 19

2014a), i.e., that a minority nucleon (e.g., a proton) has a higher probability of belonging to a high-momentum SRC-pair than a majority nucleon (e.g., a neutron). This effect should grow with the nuclear asymmetry and could possibly invert the kinetic energy sharing such that the minority nucleons move faster on average then the ma- jority. This asymmetry could have wide ranging im- plications for the NuTeV anomaly (Zeller et al., 2002, 2003) (see Sects III.D.1,VI.A.5), the nuclear symmetry energy and neutron star structure and cooling rates (Hen et al., 2016a, 2015c), neutrino-nucleus interactions (Ac- ciarri et al., 2014; Weinstein et al., 2016) and more. The study of the nuclear asymmetry dependence of the num- ber of SRC pairs and their isospin structure is an im- portant topic that could be studied in future high-energy FIG. 17: A qualitative sketch of the dominant features radioactive beam facilities. of the nucleon momentum distribution in nuclei. At k < kF , the nucleon momentum is balanced by that of the other A 1 nucleons and can be described by mean − III. Deep Inelastic Scattering (DIS) and the EMC effect field models. At k > kF , the nucleon belongs to a pn-SRC pair and its momentum is balanced by that of Basic models of nuclear physics describe the nucleus one other nucleon. as a collection of unmodified nucleons moving non- relativistically under the influence of two-nucleon and three-nucleon forces, which can be treated approximately into two regimes, above and below the Fermi-momentum as a mean field. In such a picture, the partonic structure (see Fig. 17). The region below the Fermi momentum functions of bound and free nucleons should be identi- accounts for about 80% of the nucleons in medium and cal. Therefore, it was generally expected that, except for heavy nuclei (i.e., A 12) and can be described using nucleon motion effects, Deep Inelastic Scattering (DIS) ≥ mean-field approximations. The region with momenta experiments which are sensitive to the partonic structure greater than the Fermi momentum accounts for about of the nucleon would give the same result for all nuclei. 20–25% of the nucleons (see the pie chart in Fig. 8) and Instead, the measurements (Arneodo, 1994; Aubert is dominated by nucleons belonging to NN-SRC, pre- et al., 1983; Frankfurt et al., 2012; Geesaman et al., 1995; dominantly pn-SRC. Hen et al., 2013a; Malace et al., 2014; Norton, 2003; The SRC dominance of the high-momentum tail im- Piller and Weise, 2000) show a reduction of the structure plies that the shape of the momentum distributions of all function of nucleons bound in nuclei relative to nucleons nuclei at high momenta is determined by the short range bound in deuterium in the valence quark region. We term part of the fundamental NN interaction. The average this reduction the EMC effect. Since its discovery, over number of SRC pairs is determined by global prperties 30 years ago, a large experimental and theoretical effort of the nucleus. has been put into understanding the origin of the effect. The specific predominance of pn-SRC over pp- and While theorists have had no difficulty in creating models nn-SRC is largely associated with the large contribu- that qualitatively reproduce nuclear DIS data by itself, tion of the tensor part of the NN interaction at short- there is no generally accepted model. This is because the distances (Alvioli et al., 2008; Sargsian et al., 2005b; Schi- models are either not consistent with or do not attempt avilla et al., 2007), implying that the high-momentum to explain other nuclear phenomena. distribution in heavier nuclei is approximately propor- The nuclear deep inelastic scattering data also show a tional to the deuteron momentum distribution. Experi- reduction in the small xB region of the structure func- mental and theoretical studies of the latter show that, for tion, known as the shadowing region. The physics of 300 k 600 MeV/c, n(k) 1/k4 (Hen et al., 2015a). shadowing has been well-reviewed (Frankfurt et al., 2012) This≤ specific≤ functional form∝ follows directly from the recently, and is not a subject of the present review. dominance of the tensor force acting in second order, see Section I.B showed that the nucleon-nucleon interac- Section IX.A for details. tion leads to the existence of Short-Range Correlated The predominance of np-SRC pairs implies that, even (SRC) pairs in nuclei and Section II showed the evidence in asymmetric nuclei, the ratio of protons to neutrons for and our knowledge of the properties of these pairs. in SRC pairs will equal 1. This, in turn, implies that This section will describe Deep Inelastic Scattering in neutron rich nuclei, a larger fraction of the protons and its relationship to nucleon parton distributions. The will be in an SRC pair (Hen et al., 2014d; Sargsian, EMC effect and the limitations of conventional nuclear 20

2 2 + + physics to explain it will then be discussed. Section IV ν + ν + Q = ν + q , q = ν q , q− q , so that will present the phenomenological relationship between Eq. (16) can be re-written| | as − | |  p the number of SRC pairs in a nucleus and the strength 2 2 + 2 Q k q k− + m of the EMC effect and use that relationship to gain new k+ = − − q . (17) insight into the origin of the EMC effect. ν + ν2 + Q2 If the quark is on its mass shellp (as is the case with light- 2 2 A. DIS and nucleon structure functions front wave functions) then k = mq. Furthermore, if the Primer- Deep Inelastic+ 2 quantity q k− Q , the numerator becomes simply Scattering-Q2. Then large one defines ⌫ a,Q dimensionless,2 Lorentz invariant variable by dividing the resulting equation by P +, so that + 2 e Electron loses energy and is scattered byk some angle: two variables:Q ⌫,Q2 + ξ, (18) Large ⌫,Q2 scattering on elementary quark.P Four-momentum≈ P +(ν + conservationν2 + Q2) ≡ q ! Q 2 dynamics depend on x = Q where2M⌫ ξ is the Nachtmannp variable. We see that the + + 0 3 x ratio of quark p to proton momentum P (P ± P P ) fraction of target momentum⌘ ± (plus-component)19. Structure is sim- functions 13 x P+ ply ξ. This explains why deep inelastic scattering shows the scaling phenomenon.Nucleon Thefrom relevantPDG dynamical vari- P + 1 1 P k NNPDF2.3 (NNLO) g/10 able, + , depends0.9 only on one specific0.9 combination of ν P xf(x,µ2=10 GeV2) xf(x,µ2=104 GeV2) 2 s and Q . Note that0.8 this description is frame-independent.0.8 g/10 FIG. 18: Deep inelastic scattering at large values of 0.7 0.7 One need not go to the infinite momentum framed to un- 2 0.6 0.6 The EMC effect involves deep u Q . A lepton (labelled ‘e’) scatters from a nucleon by v derstand scaling or the parton model. c 0.5 0.5 u emitting a space-like virtual photon with If one further takes the Bjorken limit (ν2 Q2),v then inelastic scattering from nuclei 0.4 0.4 u + 2 d  four-momentum q, which is absorbed on a single quark k Q v 0.3 0.3 b + = µ x . The dominant dependence ondv x + P 2P qµ B s B with momentum fractionEMC=xB PEuropean. Only the outgoingMuon Collaboration≡ 0.2 0.2 d is called Bjorken scaling, andu its discovery, using hydro- lepton is subsequently detected. 0.1 0.1 gen and deuteron targetsc (to obtain the neutron infor- 0 0 -3 -3 10 10-2 10-1 1 10 10-2 10-1 1 mation), was the primary4 x evidence for the existencex of We begin with a brief description of deep inelastic scat- quarks withinFigure the 19.4: nucleon.The bands are x times the unpolarized parton distributions f(x) (where f = uv,dv, u, d, s s,¯ c =¯c, b = ¯b, g) obtained in NNLO NNPDF2.3 global tering on a nucleon. See one of the many texts for details, 2≃ 2 2 4 2 2 Quarks areanalysis confined, [45] at scales soµ they=10GeV areand neverµ =10 onGeV their,with massαs(MZ)=0.118. e.g., (Close, 1979; Collins, 2013; Halzen and Martin, The analogous results obtained in the NNLO MSTW analysis [43]canbefoundin shell. The off-massRef. [62]. shell effects, however, decrease with 1984; Roberts, 1994; Thomas and Weise, 2001). The lat- increasing values of Q2 and are regarded as “higher where we have used F γ =2xF γ + F γ,nottobeconfusedwithF γ of Sec. 19.2. Complete est information is contained in the Particle Data Group 2 T L 2 twist”. Suchformulae effects are given, could for example, be important in the comprehensive at Jefferson reviewofRef.80. Lab tables (Olive et al., 2014). The inclusive deep inelas- γ 2 energies. EffectsThe hadronic of final photon state structure interactions function, F2 ,evolveswithincreasing (which depend Q from tic scattering process, (e, e ), involves a lepton scattering the ‘hadron-like’ behavior, calculable via the vector-meson-dominance model, to the 0 on the kinematicsdominating ‘point-like’ of the probing behaviour, calculable beam) inare perturb notative contained QCD. Due to the point-like γ 2 from a target, with only the final state lepton being de- coupling, the logarithmic evolution of F2 with Q has a positive slope for all values of x, in the light-frontsee Fig. 19.15. wave The ‘loss’ function of quarks at (Cosyn large x dueet to gluonal., radiation2014). is over-compensated tected. If spin variables are not observed, the process by the ‘creation’ of quarks via the point-like γ qq¯ coupling. The logarithmic evolution Suppose the struck quark is confined→ in a nucleon of depends on only two variables, which are traditionally was first predictedµ in the quark–parton model (γ∗γ qq¯)[81,82],andtheninQCDin four-momentumthe limit ofp largethatQ2 [83]. is bound The evolution within is now known a→ nucleus to NLO [84–86].of mo- The NLO data chosen to be the electron energy loss ν and negative of mentum Panalysesµ. Then to determine we have the parton densities of the photon can be found in [87–89]. the four momentum transfer from the lepton to the tar- + + 2 2 2 19.5. Diffractivek DISp (DDIS) get Q = q ν , see Fig. 18. At large enough values of = ξ. (19) 2 − Some 10% of DIS events+ are+ diffractive, γ∗p X + p, in which the slightly deflected ν and Q , conservation of momentum and energy leads proton and the clusterp XPof outgoing hadrons→ are well-separated in rapidity. Besides x and Q2, two extra variables are needed to describe a DDIS event: the fraction x + IP to the result that the dynamical information, can be en- where a nucleusof the proton’s of momentum momentum transferredP acrosscontains the rapidity anucleon gapandt,thesquareofthe of coded (at a given scale) in the structure functions q(x ), 4-momentum+ transfer of the proton. The DDIS data [90,91] are usually analyzed using B momentum of p which contains a quark of momentumD two levels of factorization. First, the diffractive structurefunctionF2 satisfies collinear which is interpreted as the fraction of the target momen- k+. This is the origin of the convolution model to be August 21, 2014 13:18 tum carried by the struck quark. discussed in Sect. III.C.1. Therefore, in order to calculate Let’s see how this arises. Four-momentum conserva- deep inelastic scattering from nuclei we need to know the tion, the idea that the quark is briefly free after absorbing nuclear wave function, expressed in light-front variables. the high-momentum photon, and ignoring the emissions More formally, one derives the expression for the mo- of gluons gives mentum distribution (the probability that a quark has + + (k + q)2 = m2 (16) a given value of k /P ), known as a quark distribution q function, by starting with the the square of the invariant where k is the four-momentum of a quark in the target, scattering amplitude. The important part of this ampli- µν and mq is the quark mass. Let the spatial momentum tude depends on the hadronic tensor W , which is a ma- of the photon lie in the negative z direction and using trix element of a commutator of electromagnetic current 0 3 the light-front momentum variables, e.g P ± P P , operators. After expanding in terms of the separation r µ ≡ ± where P is the target four-momentum, we have q− = of the spatial variable of the two current operators, the 21 momentum distribution (for a specific flavor of quark) is B. The EMC effect given in the Bjorken scaling limit (in which the variable Q2 is not explicit) by the Fourier transform (Thomas and As stated, the discovery of Bjorken scaling was made Weise, 2001) using hydrogen and deuterium targets. It occurred to many experimentalists that MeV-scale nuclear effects 1 iq+r− should be negligible at GeV-scale momentum and en- q(ξ) = dr− e P ψ† (r−)ψ (0) P , (20) 2π h | + + | ic ergy transfers and that therefore they could increase Z their experimental statistics by using nuclear targets. where P is the proton wave function, the subscript c Surprisingly, the CERN European Muon Collaboration | ic denotes a connected matrix element, ψ is the quark field- (EMC) found that the per-nucleon (e, e0) cross section operator, the subscript + denotes multiplication by the ratio of iron to deuterium was not unity (Aubert et al., 0 3 projection operator (1 + γ γ )/2, and r− is the minus- 1983), see Fig. 20. This surprising result, now called the component of the separation distance. EMC Effect, was confirmed by many groups, culminat- Parton distributions are needed for a wide variety of ing with the high-precision electron and muon scattering 2 applications in high-energy physics. q(xB,Q ) has been data from SLAC, Fermilab, NMC at CERN, and Jeffer- determined for various flavors and for a wide range of son Lab (see Fig. 21). See one of the many EMC reviews values of x and Q2. Vast amounts of data are now cod- for details (Arneodo, 1994; Geesaman et al., 1995; Hen ified as parton distributions, giving the probability as a et al., 2013a; Malace et al., 2014; Norton, 2003; Piller and function of Q2 that a given flavor of quark carries a mo- Weise, 2000). mentum fraction xB, see Fig. 19. This sub-section is concerned with nucleon targets, but (as mentioned above) we need to know how to evaluate a nuclear version of Eq. (20),which would involve nuclear wave functions expressed in terms of light front variables. This difficulty has been handled (Frankfurt and Strik- man, 1981b, 1988a; Smith and Miller, 2002),(Blunden et al., 1999; Miller and Machleidt, 1999b; Miller, 2000; Miller and Smith, 2002). One can implement light-front coordinates using a simple transformation. This works because the nucleus does not contain a significant NN¯ content.

1 1 NNPDF2.3 (NNLO) g/10 0.9 0.9 xf(x,µ2=10 GeV2) xf(x,µ2=104 GeV2) s 0.8 0.8 g/10 0.7 0.7 d 0.6 0.6 u v c 0.5 0.5 FIG. 20: Image of the EMC data as it appeared in the uv 0.4 0.4 u November 1982 issue of the CERN Courier. This image d v nearly derailed the refereed publication (Aubert et al., 0.3 0.3 b d s v 1983), as the editor argued that the data had already 0.2 0.2 d been published. u 0.1 0.1 c 0 0 -3 -3 10 10-2 10-1 1 10 10-2 10-1 1 The conclusion from the combined experimental ev- x x idence was that the effect had a universal shape, was independent of the squared four momentum transfer Q2 FIG. 19: The bands show x times the unpolarized B starting from remarkably small values of Q2 (see Fig. 22), parton distributions for the different parton flavors increased with nuclear mass number A, and increased u , d , u, d, s =s, ¯ c =c, ¯ b = ¯b, and g obtained in v v with the average nuclear density. An early study (Bick- NNLO{ NNPDF2.3 global analysis (Ball}et al., 2013), at erstaff and Miller, 1986) of the Q2 dependence of nuclear Q2 = 10 (GeV/c)2 and 104 (GeV/c)2, with effects showed that the nuclear-binding and dynamical α (M 2 ) = 0.118. From the PDG (Olive et al., 2014). s Z rescaling models predict very little variation with Q2 over Here x = x . B the range from 4 to 104 GeV2. One way to characterize the strength of the EMC effect 22

FIG. 22: The Q2 dependence of the EMC ratio for iron at various values of xB . From (Gomez et al., 1994).

FIG. 23: The slope of the EMC effect for 0.35 xB 0.7 plotted versus the average nuclear density≤ for various≤ light nuclei as measured at Jefferson Lab. From (Seely et al., 2009a).

is to measure the average slope of the cross section ratio for 0.35 xB 0.7. Plotting this slope versus the av- erage nuclear≤ density≤ for light nuclei (see Fig. 23) shows that the EMC effect does not simply depend on average density. Since 9Be can be described as a pair of tightly bound alpha particles plus one additional neutron, it has FIG. 21: The per nucleon cross section ratio of various been suggested that the local density is more important nuclei to deuterium as measured at (top) SLAC (Gomez than the average density (Seely et al., 2009a). See an et al., 1994) and (bottom) Jefferson Lab (Seely et al., early discussion in (Frankfurt and Strikman, 1981b). 2009a). The solid curves show the A-dependent fit to The immediate parton model interpretation of the data the SLAC data (Gomez et al., 1994), while the dashed at high x is that the valence quarks of a nucleon bound 12 curve is the SLAC fit to C. Figures from (Gomez in a nucleus carry less momentum than those of free nu- et al., 1994) (top) and (Seely et al., 2009a) (bottom). cleons. This notion seems uncontested, but determining the underlying origin remained an elusive goal for a long 23

evaluation of deep inelastic scattering from nuclei there- k+q fore involves knowledge of the nuclear wave function, ex- pressed in light-front variables. k k There were many attempts to explain the EMC effect without invoking medium modifications. We cite a few p p of the references (Akulinichev and Shlomo, 1990; Benhar P P et al., 1997, 1999; Benhar and Sick, 2012; Ciofi Degli Atti and Liuti, 1989; Dieperink and Miller, 1991; Jung and Miller, 1988, 1990; Marco et al., 1996), with others to be found in the reviews. FIG. 24: Deep inelastic scattering diagram. A virtual The appeal of the nucleon-only idea can be understood photon, γ∗, of momentum q is absorbed on a quark of using a simple caricature of the probability that the nu- momentum k contained in a nucleon of momentum p in cleon carries a momentum fraction y. The width of the a nucleus of momentum P . The imaginary part of this function fN (y) is determined by the Fermi momentum diagram corresponds to the hadronic tensor W µν . divided by the nucleon mass, which is small. In the ab- Figure adapted from (Miller and Smith, 2002). sence of interactions, fN (y) is peaked at y = 1. If the average separation energy S M (which for nuclear matter can be as large as 70 MeV),≡ (Benhar et al., 1997, time. The great number of models created to explain the 1999; Benhar and Sick, 2012; Dieperink and Miller, 1991) EMC effect caused one of us to write in 1988 (Miller, then fN (y) is peaked at about y = 1 . Taking for sim- 1988) that “EMC means Everyone’s Model is Cool”. plicity a zero width approximation −

fN (y) = δ(y (1 )), (24) C. Why Conventional nuclear physics cannot explain the − − EMC effect then the convolution formula (Eq. (21)) tells us that

1. Nucleons only F (x ) x 2A A F A . (25) A ≈ 2N 1  One must first try to explain the EMC effect using only  −  the simple kinematic effects of binding energy and Fermi As shown in Fig. 19 the structure function falls rapidly motion without modifying the bound nucleon structure. with increasing xB, so that a slight increase in the ar- If the nucleon structure function is not modified and is gument leads to a significant decrease in the structure the same on and off the energy shell (nucleon-only hy- function. In particular, pothesis) then evaluation of the diagram of Fig. 24 leads F2A(xA) F20N (xA) to the simple convolution formula: 1 +  1 γ, (26) AF2N(xA) ≈ F2N (xA) ≈ − A F2A(xA) γ = dyfN (y)F2N (xA/y), (21) where we have assumed F2N (xB) (1 xB) at large A x ∼ − Z A xB with 3 γ 4. ≤ ≤ where P is the total four momentum of the nucleus, and Frankfurt and Strikman (Frankfurt and Strikman, 1987), using a more detailed calculation found that a Q2A x AM x = B . (22) value of  = 0.04 was sufficient to reproduce the early A ≡ 2P q M · A EMC data. However, we will show that the ideas of with M and MA as the free nucleon and nuclear masses, shifting the value of xA based on binding energy or sep- respectively. xA can be thought of as a version of xB aration energy considerations violates rigorous (Collins, corrected for the average nucleon binding energy. The 2013) baryon and momentum sum rules, and therefore variable y = Ap+/P + is the fraction of the nuclear mo- cannot be a viable explanation of the EMC effect. Con- mentum (per nucleon) carried by a single nucleon, and sider a nuclear model in which nucleons are the only de- fN (y) is the corresponding probability distribution. The grees of freedom. There will be a conserved baryon cur- origin of the convolution formula can be understood us- rent and an energy-momentum tensor expressed in terms ing the simple terms of Sect. III.A. Suppose the struck of these constituents. This means that when expressed quark is confined in a nucleon (of four-momentum p) that in terms of the convolution approach of the previous sub- is bound within a nucleus of momentum P . Then from section we must have the momentum sum rule: Eq. (19) we have dyyfN (y) = 1, (27) ξ k+/P + = (k+/p+)(p+/P +) = x /y. (23) ≈ A Z This accounts for a nucleon in the nucleus of momentum where the factor of y represents the momentum. The use p+ that contains a quark of momentum k+. A proper of Eq. (24) in Eq. (27) leads immediately to a substantial 8

Oxygen Calcium

1.1 1.1

1 1 R R 0.9 0.9

0.8 0.8

0.7 0.7 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 24 x x Lead Nuclear Matter Thus F2A/A is well approximated by F2N and one gets no substantial EMC effect this way (Miller and Smith, 1.1 1.1 2002; Smith and Miller, 2002). This is shown as the solid

1 1 curve in Fig. 25. R R 0.9 0.9

0.8 0.8

0.7 0.7 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 x x

FIG.FIG. 5: Ratio 25: functionsThe measured for 16O, EMC40Ca and effect208Pb in showing gold (Gomez data for Carbon, Calcium and Gold, respectively, from SLAC-E139 [30].et al. The, 1994) nuclear compared matter calculation to a nucleons-only shows extrapolated calculation data [32]. of the EMC effect in lead. From (Smith and Miller, 2002). mesons carry a small amount of plus momentum [21] that vanishes as A .Theclosertoy =1thepeakisin FIG.→∞ 26: The Drell-Yan process. A quark with Fig. 2, the less pronounced the minimum in Fig. 5, all else remaining constant.momentum The second fraction effectx isfrom due to theM,which incident proton reachesviolation a minimum of the momentum at 56Fe corresponding sum rule: to a more pronounced minimum of the ratio function1 than for A<56 or A>56, keeping the scalar meson contribution constant. annihilates with an anti-quark from the nuclear target Using a Gaussian parameterizationdyyfN (y) = 1 of the. plus momentum(28) distributionwith momentum and the experimental fraction bindingx2 to form energy a time-like per virtual − + nucleon via the semi-empiricalZ mass formula, we have modeledthedependenceoftheminimumoftheratiofunction,photon which decays to a µ µ− pair. Figure adapted RFrankfurt(x 0.72), and on Strikman the number (Frankfurt of nucleons and in Strikman, the nucleus 1987) in Fig. 6. The motivationfrom for (Bickerstaff the use of Gaussianet al., 1986). plus momentumalso≃ included distributions an important is based relativistic on the expansion correction [5] known as the ’flux factor’, which significantly reduces the effects of nuclear binding. F2A(xA)=F2N (xA)+ϵxAF2′N (xA) 2 +γ[2xAF2′N (xA)+xAF2′′N (xA)] (5.5) Going beyond the zero width approximation only 2. Nucleons plus pions makes this problem worse (Millerwhere and Machleidt,ϵ 1 1999a).dyyf (y) (5.6) ≡ − N The inclusion of the effects of short-ranged correlations! Nucleons-only models fail, but it was natural to con- broadens the function fN (y) leadingγ to a valuedy(y of1) the2f (y) (5.7) ≡ − N sider the idea that the missing momentum  of Eq. (28) ratio that exceeds unity for small values of! x, an effect is carried by non-nucleonic degrees of freedom, e.g., pions Thefound Gaussian earlier parameterization in (Dieperink and uses Miller, the peak 1991). location A and viola- width, y and ( y2 y 2)1/2 respectively, from the (Ericson⟨ ⟩ and⟨ ⟩−⟨ Thomas,⟩ 1983; Llewellyn Smith, 1983). In relativistiction of the Hartree sum rule calculations by a few in percent Fig. 2, and is actually is normali azed huge to unity.this This case, allows us to obtain a plus momentum distributionviolation, because for any A thewith EMC minimal effect eff itselfort. We is only show a the 10-15% combined effect of scalar mesons and binding energy per nucleoneffect. on Thus the ratio nucleon-only function along models with are the logically effect of scalar inconsis- mesons alone using a constantP + binding= P + energy+ P + = perM nucleon. (29) of 8.5MeVindependentofA. It can be seen that a changing M with A has the most effect forN nucleiπ muchA larger tent− and therefore wrong, even if they can be arranged than Iron, but does not change the general trend that the minimum ofMany the ratio authors, function see becomes the reviews less pronounced (Arneodo, as 1994;A Frankfurt increasesto describe due tothe the data. vanishing scalar meson contribution andthepeakoftheplusmomentumdistributionapproaching unity.One This might dependence argue that of the sum binding rules eff canect on notA beis quite applied different,and both Strikman, in magnitude 1988b; and Geesaman shape, thanet the al. tren, 1995;d Piller and 1/3 + indirectly experimental to the data data summarized because of in the Ref. need [32] which to incorporate satisfies R(x Weise,0.72) 2000)A− found,sothattheminimumbecomes that using Pπ /MA = 0.04 is suffi- moreinitial pronounced and final as stateA increases. interactions. This fully Nevertheless, demonstrates in usingthe inadequacy≃cient of to∼ conve accountntional for nucleon-meson the EMC effect. dynamics However, to if nuclear the convolution formalism in the nucleon-only approxi- pions carry 4% of the nuclear momentum (in the rest mation one must use a light-front wave function of the frame the plus component of momentum is the nuclear nucleus consistent with the conservation of baryon num- mass) then there should be more nuclear sea quarks (i.e., ber and momentum, as discussed above. There is no way both quarks and anti-quarks). This enhancement should to avoid the constraints imposed by the sum rules. be observable in a nuclear Drell-Yan experiment (Bicker- Indeed, the application of sum rules and simple rea- staff et al., 1984, 1986; Ericson and Thomas, 1984). The soning shows that Eq. (21) leads to the result that the idea, see Fig. 26, is that a quark from an incident proton nucleon-only hypothesis can not explain the EMC ef- (defined by a large value of x1) annihilates an anti-quark fect. Under the Hugenholz van Hove theorem (Hugen- from the target nucleus (defined by a smaller value of holtz and van Hove, 1958; Miller and Smith, 2002; Smith x2). A significant enhancement of pions would enhance and Miller, 2002) nuclear stability (pressure balance) im- the anti-quarks and enhance the nuclear Drell-Yan re- + plies (in the rest frame) that P = P − = MA. But to an action. But no such enhancement was observed (Alde + excellent approximation P = A(MN 8 MeV). Thus et al., 1990) as shown in Fig. 27. This caused Bertsch et an average nucleon has p+ = M 8− MeV. As carica- al. (Bertsch et al., 1993) to announce “a crisis in nuclear N − tured in Eq. (24), the function fN (y) is narrowly peaked theory” because conventional theory does not work. This because the Fermi momentum is much smaller than the statement is the verification of the title of this subsection. nucleon mass. This means that the value of y in the in- The reader might ask at this stage, if the two-pion tegral of Eq. (21) is constrained to be very near unity. exchange effects discussed in the Appendix and Sects. I 25

e0 e0

e q =(⌫, q) p + q

p

PA PA 1

(Fe) (D)

FIG. 28: The A(e, e0p) reaction in the Plane Wave Impulse Approximation. A nucleus of four-momentum P emits a nucleon of four-momentum p that absorbs a virtual photon of four-momentum q to make a nucleon 2 2 x2 of four momentum p + q, with (p + q) = M , where M is the nucleon mass. The blob represents the in-medium electromagnetic form factors.

differs from that of a free nucleon. The medium modifies the nucleon. FIG. 27: Drell-Yan experimental results (Alde et al., This is not surprising, as there are evident simple ex- 1990). Ratio of Drell-Yan cross sections as a function of amples. A free neutron undergoes β decay, so it can be the momentum fraction x2 of a quark in the nucleus. thought of as having a pe−ν component. When bound The version of the rescaling model shown in this figure in a stable nucleus, the| neutroni is stable. This “medium does not reproduce the nuclear deep inelastic scattering modification” suppresses the pe−ν component. Addi- data (Bickerstaff et al., 1985, 1986). Figure adapted | i tionally, in the (e, e0p) reaction shown in Fig. 28, four- from (Alde et al., 1990). momentum conservation shows that the square of the initial four-momentum of the struck nucleon, p, cannot satisfy p2 = M 2. Thus the form factor of a nucleon & II lead to a significant pion content and an enhanced bound in the nucleus cannot be the same as that for a sea in the nucleus. Explicit calculations show that the free nucleon; it is instead the amplitude for a transition pionic content associated with the tensor potential is very between a virtual nucleon of mass p2 and a physical small (Miller, 2014). nucleon of mass M. p Subsequent work has confirmed that an intrinsic mod- Now we must ask: what is the origin of the medium ification of the nucleon structure function is needed to modification? This question is coupled to the broader explain the EMC effect (Frankfurt and Strikman, 2012; questions listed in Sect. I, and more deeply to the very Hen et al., 2013a; Kulagin and Petti, 2010, 2014, 2006b). nature of confinement. This result had been expected for some time, as stated The parton model interpretation of the large-x part explicitly “The change of the structure functions in nuclei B of the EMC effect is that the medium reduces the nu- (EMC effect) gives direct evidence for the modification clear structure functions for large x , so that there are of quark properties in the nuclear medium” (Walecka, B fewer high-momentum quarks in a nucleus than in free 2005). The following sections discuss specific proposals space. This momentum reduction leads, via the uncer- for such modifications. tainty principle, to the notion that quarks in nuclei are confined in a larger volume than that of a free nucleon. D. Beyond Conventional Nuclear Physics: Nucleon There are two general ways to realize this simple idea: Modification mean-field effects cause bound nucleons to be larger than free ones, or nucleon-nucleon interactions at close range The failure of the nucleon-only or nucleon+pion mod- cause nucleon structure to be modified, by including ei- els to explain the EMC and Drell-Yan data indicates that ther NN ∗ configurations or 6-quark configurations that the structure of a nucleon bound in a nucleus significantly are orthogonal to the two-nucleon wave functions. All 7

1.2 1.2 7 15 1.1 Li 1.1 N 1 1 0.9 0.9 Ratios Ratios 0.8 0.8 Experiment: 9Be 0.7 12 EMC Unpolarized EMC effect EMC Experiment: C 0.7 Polarized EMC effect: R(3/2 1) 0.6 Unpolarized EMC effect 2 2 As Q2 = 5 GeV2 Q = 5 GeV Polarized EMC effect: R3/23/2 Polarized EMC effect 0.6 As 0.5

0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 xx xx FIG. 6: The EMC and polarized EMC effect in 7Li. The FIG. 8: The EMC and polarized EMC effect in 15N. The 26 empirical data is from Ref. [31]. empirical data is from Ref. [31].

1.2 1.2 11 27 1.1 B 1.1 Al

1 1

0.9 0.9 Ratios Ratios

0.8 Experiment: 12C 0.8 Experiment: 27Al

EMC Unpolarized EMC effect EMC Unpolarized EMC effect 0.7 (3/2 1) 0.7 (5/2 1) Polarized EMC effect: RAs 2 2 Polarized EMC effect: RAs 2 2 3/23/2 Q = 5 GeV 5/25/2 Q = 5 GeV 0.6 Polarized EMC effect: RAs 0.6 Polarized EMC effect: RAs

0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 xx xx FIG. 7: The EMC and polarized EMC effect in 11B. The FIG. 9: The EMC and polarized EMC effect in 27Al. The empirical data is from Ref. [31]. empirical data is from Ref. [31]. FIG. 30: (color online) The measured EMC effect for FIG. 29: Evolution of nuclear physics from structureless 27 nucleons in the 1940s to independent 3-quark nucleons Al (Gomez et al., 1994) compared to QMC where we find a polarized EMC effect roughly twice thatcalculationsresults for of the both polarized the EMCregular effect and in nuclei the polarized corroborate EMC in the 1970s toof the the modified unpolarized nucleons case. of today, either effectearlier of Ref. nuclear (Cloet matteret [11, al. 32],, 2006). light nuclei Figure [33] andfrom small (Cloet modified single-nucleons (left) or modified two-nucleon x [28] studies that found large medium modifications to configurations (right). the spin structure functionet al., 2006).relative to the unpolarized IV. CONCLUSION case. In particular, we find that the fraction of the spin of the nucleon carried by the quarks is decreased in nuclei Using a relativistic formalism, where the quarks in the (see Table II). If this result is confirmed experimentally, of the papers seekingbound nucleons to explain respond the to EMC the nuclear effect environment, using wemesonitcoupling, would give QMC) important used insights the MIT into in-medium bag model quark to repre- calculated the quark distributions and structure func- dynamics and help quantify the role of quark degrees of medium modification use7 one11 15 of the two27 ideas (that are sent the three confined quarks in the proton (Guichon, tions of Li, B, Nand Al. For a spin-J target there freedom in the nuclear environment. cartooned in Fig.are 29). 2J +1independentquarkdistributionsorstructure1988; Guichon et al., 1996; Stone et al., 2016). Later 27 Since only aboutfunctions 20% inof the nucleons Bjorken belong limit. For to example, SRC pairs,Al there-work used the QMC model with more general confine- fore has six structure functions, however we find that thement mechanisms (BlundenAcknowledgments and Miller, 1996), the co- Fig. 17, five timeshigher more multipoles nucleons are suppressedwould be relative modified to the by leading mean-field effectsresult than by at by least nucleon-nucleon an order of magnitude. interactions variantW.B. NJL wishes model to (Cloet thank S.et Kumano, al., 2009a, T. Suzuki 2006, (Nihon 2016) and at close range. Therefore,We were readily if nucleons able to are describe only the modified EMC effect inthe chiralUniversity) quark and soliton B. A. Brown model for (Diakonov discussions.et We al. also, 1996; at short range, thenthese the nuclei, modifications and importantly needed obtained to explain the correct ASmith- thank and S. Miller, Kumano 2003, for the 2004, QCD evolution 2005). codeIn these [34, 35]. models dependence. Although we do not show the results, wethe attractionThis work wasneeded supported to produce by the a Australian bound state Research is gener- the EMC effectalso would determined have to the be EMC five ratiotimes for larger12C, 16 thanOand28Si Council and DOE contract DE-AC05-84150, under which if all nucleons wereandfound modified results by very mean-field similar to their effects.A 1neighbours.atedJSA by the operates exchange Jefferson of Lab, scalar and quantum by the Grant numbers in Aid for (either − A phenomenologicalIn Eq. (23) assessment we define theof this polarized idea EMC in which ratio in nu-by aScientific scalar meson Research (Guichon, of the Japanese 1988; Ministry Guichon ofet Educa- al., 1996; clei. This ratio is such that in the extreme nonrelativis-Stonetion,et al. Culture,, 2016) Sports, or by Science pairs of and pions Technology, (Smith Project and Miller, the mean-field andtic limit, SRC with related no medium origins modifications, of the EMC it effect is unity. The No. C2-16540267. were treated phenomenologically was made in Ref. (Hen 2003, 2004, 2005)) and the repulsion needed to obtain nu- et al., 2013a). The separation of the spectral function clear saturation is caused by exchange of vector mesons. Eq. (7) into terms arising from low-lying excited states Within these mean field models the exchanged mesons P and higher-energy continuum states related to short- are treated as classical static fields, as such these mesons 0 do not interact with the photon probe. ranged correlations, P1 was used. In the mean field model, a nucleus-independent modification of F2 was in- We next explain two classes of models. The chiral cluded in the contribution to the nucleon distribution quark soliton model (CQSM) is based on the instanton- function, fN (y) Eq. (21) arising from P0. In the alternate dominated nature of the vacuum (Negele, 1999). The model a much larger nucleus-independent modification of coupling of quarks to vacuum instantons spontaneously F2 was included in the contribution to fN (y) arising from generates a constituent quark mass of about 400 MeV. P1. Both approaches gave reasonably good descriptions These quarks interact with pions through an effective of the nuclear DIS data. CQSM Lagrangian. This model reproduces nucleon We next describe specific models associated with the properties well, including structure functions which van- two different mechanisms. ish at xB = 0 and 1 (Diakonov et al., 1996). Nuclei are formed by collections of such nucleons ex- changing scalar and vector mesons (Smith and Miller, 1. Mean field 2003, 2004, 2005). Excellent saturation properties were obtained. The dominant effect of the medium is a In mean-field models of nucleon modification, the in- slight broadening of the effective potential that binds the teraction between nucleons occurs by the exchange of quarks in the nucleon. The use of the medium modified mesons between quarks confined in different nucleons. wave function to compute structure functions allows one Four general models of the quarks confined in the nu- to account for the EMC effect, while still agreeing with cleon have been used for this. The earliest model (quark the Drell-Yan data. This indicates that the sea is not 27 very modified. Schematic The next model places an NJL-model nucleon in the medium (NJLMM) which is a relativistic extension of the two-component earlier QMC including the effects of spontaneous symme- + ✏ .. try breaking. Here the external scalar field enhances the nucleon model lower component of the quark’s Dirac wave function by . about 15%. This model describes the EMC effect well . (see Fig. 30). It also predicts an enhancement of the FIG. 31:. Two component nucleon model: normal-sized EMC Effect for spin structure functions (Cloet et al., 2005b) in nuclei which could be measured at Jefferson component plus point-like configuration component. Lab (see Section VI.A.4). The NJLMM predicts the effects of having different 2. Suppression of point-like configurations numbers of neutrons N and protons Z. Cloet and Thomas (Cloet et al., 2009a) explained that a neutron or proton excess in nuclei leads to an isovector-vector mean- We can also make a more general model of the nu- field which, through its coupling to the quarks in a bound cleon as a superposition of various configurations or nucleon, causes the quark distributions to be evaluated at Fock states, each with a different quark-gluon structure. a shifted value of the Bjorken scaling variable (Detmold Fig. 31 shows a two-component nucleon where one com- et al., 2006; Mineo et al., 2004). In relativistic mean-field ponent is “blob-like” (BLC) with the normal nucleon size models, the effect of a vector field is to shift the energy and the other is “point-like” (PLC). The BLC can be and therefore the value of the plus component of mo- thought of as an object that is similar to a nucleon. The mentum of the single particle state. The isovector-vector PLC is meant to represent a three-quark system of small mean field is represented by the ρ0, and in this work its size that is responsible for the high-x behavior of the dis- strength is chosen to reproduce the nuclear symmetry en- tribution function. The smaller the number of quarks, ergy. In a nucleus like 56Fe or 208Pb where N > Z, the ρ0 the more likely one can carry a large momentum frac- tion. Furthermore, because the PLC is smaller than the field causes the u-quark to feel a small additional vector 1 attraction and the d-quark to feel additional repulsion. BLC, the uncertainty principle tells one that quarks con- This effect leads to a significant correction to the NuTeV fined in the PLC have higher momentum. The small- 2 sized configuration (with its small number of qq¯ pairs) is measurement of sin ΘW (Zeller et al., 2002, 2003). The sign of this correction is largely model independent, and very different than a low lying nucleon excitation. it accounts for approximately two-thirds of the NuTeV When placed in a nucleus, the blob-like configura- anomaly. Thus the NuTeV measurement provides fur- tion feels the regular nuclear attraction and its energy ther evidence for the medium modification of the bound decreases. The point-like-configuration feels far less nucleon wave function. nuclear-attraction because the effects of gluons emit- ted by small-sized configurations are cancelled in low- Both sets of mean field models predict modification of momentum transfer processes. This effect is termed color nucleon electromagnetic form factors. The QMC model screening and has been verified in several different reac- predicts modifications to both GE and GM (Lu et al., tions (Dutta et al., 2013; Frankfurt et al., 1994). The nu- 1999), while the chiral quark soluton model only modifies clear attraction increases the energy difference between GE (Smith and Miller, 2004). Both models predict the the BLCs and the PLCs, therefore reducing the PLC same ratio GE/GM . Note that electron-nucleus quasi- probability (Frankfurt and Strikman, 1985). The PLC elastic data was used (Sick, 1985) to put a limit of be- is suppressed. Reducing the probability of PLCs in the tween 3 and 6% on the possible increase of the nucleon nucleus reduces the quark momenta, in agreement with radius in nuclear matter. None of the mean field models the EMC effect. discussed here violate this limit. This idea was studied (Frank et al., 1996) using The QCD eigenstates of a free nucleon form a complete a relativistic constituent quark model for the nu- set. Thus the medium modified nucleon can be regarded cleon (Schlumpf, 1992, 1993). A nucleon is placed in as a superposition the nucleon and all of its excited states. the nucleus and therefore subject to a mean field that vanishes for configurations in which the three quarks are Despite the general success of mean-field models it close together. The quark momentum distribution de- must be noted that none predicts significant extra high- creases for xB > 0.3, see Fig. 32. The effects of nucleon momentum strength in the nuclear momentum distribu- motion are not included, so there is no rise for large values tion. Therefore, it is very difficult to see how they could of xB, and the dip at low values of xB would be removed reproduce the plateaus observed in the cross section ra- by such effects. This model gives only a 2.5 % enhance- tios at x 1.5 seen in Section II.C. ment at x = 0.5 because the enhancing effects of large B ≥ B 28

the relevant kinematics, and our model will include this dependence explicitly. Our model is similar to the model of (Frankfurt and Strikman, 1985), with the important difference that the medium effects will enter as an ampli- tude instead of as a probability. In (Frank et al., 1996) the PLC is subject to a non-zero, but small, attractive potential that fluctuates with the nucleon configurations. The complete Hamiltonian H = H0 +H1 is now given by E U V H = B , (34) VE− | |  P  in which the attractive nature of the nuclear binding po- tential is emphasized. Then interactions with the nucleus FIG. 32: The ratio of F2 in the nucleus to the free F2 increase the energy difference between the BLC and the (the EMC ratio) in the point-like configuration PLC, which decreases the PLC probability. suppression model. From (Frank et al., 1996). The medium-modified nucleon and its excited state, N M and N ∗ M , are now (using first-order perturbation |theory)i | i virtuality discussed below were not included. The PLC model, being a modification at large values of x , does N M = B + M P (35) B | i | i | i not contradict the nuclear Drell-Yan data, Sect. III. N ∗ M = M B + P , (36) The notion that different constituents of the nucleon | i − | i | i have different sizes and therefore different interaction where strengths is directly related to medium modifications of V EB EP all kinds. The main features of this idea can be under- M = =  − (37) EB U EP EB U EP stood using a simple schematic two-component model of − | | − − | | − the nucleon with a dominant normal-sized blob-like con- so that the PLC probability in the medium is suppressed. stituent (denoted by B) and a very small point-like con- Both M and  are less than zero, so that M  > 0. The medium modified nucleon N may be− expressed stituent (denoted by P ). The Hamiltonian is given by | iM in terms of the unmodified eigenstates N , N ∗ as the matrix | i | i EB V N M N + (M ) N ∗ . (38) H0 = , (30) | i ≈ | i − | i VEP   Within this model the medium-modified nucleon contains where EP EB. Because of the hard-interaction po- a component that is an excited state of a free nucleon. tential, V , that connects the two components, the eigen- The amount of modification, M , which gives a devia- states of H are N and N ∗ rather than B and P . − 0 | i | i | i | i tion of the EMC ratio from unity, is controlled by the po- In lowest-order perturbation theory, the eigenstates are tential U. An initial pioneering qualitative description of given by the EMC effect was obtained (Frankfurt and Strikman, N = B +  P , (31) 1985) (at xB = 0.5, where effects of Fermi motion are | i | i | i small) using U = 40 MeV and EP EB 500 MeV. N ∗ =  B + P , (32) − − ∼ | i − | i | i The present treatment instead calculates the effects of with  = V/(EB EP ). We assume V EP EB, so the medium on the amplitude instead of the probability, that the nucleon− is mainly B and| its|  excited− state is so that the effects are generally larger. We will explore mainly P , and also take V| >i 0. We use the notation this further in Section IV.B. | i N ∗ to denote the state that is mainly a PLC, but the The PLC suppression model also predicts changes to | i PLC, as discussed above, does not resemble a low-lying the elastic electric and magnetic form factors GE,M . The baryon resonance. electromagnetic form factor in free-space is obtained as Now suppose the nucleon is bound to a nucleus. The 1 nucleon feels an attractive nuclear potential H : F = B J B + 2 B J P + 2 P J P ,(39) 1 1 + 2 h | | i h | | i h | | i U 0 H = (33) where momentum and spin labels have been suppressed.
1 0 0   It is instructive to examine what to expect at both to represent the idea that only the large-sized component high and low momentum transfer. At low momentum of the nucleon feels the influence of the nuclear attraction. transfer the first term dominates so that the spatial ex- The treatment of the nuclear interaction, U, as a num- tent of the nucleon and its modification in the medium ber is clearly a simplification. The interaction varies with are important. Frankfurt and Strikman (Frankfurt and 29

nucleon is part of a six-quark configuration (Pirner and Vary, 1981) (who predicted the existence of plateaus in (e, e0) cross section ratios) that is orthogonal to any two- nucleon wave function. Because a six-quark configura- tion is larger than a nucleon, the quarks are partially deconfined. Larger confinement volumes are associated with lower momenta, and therefore with a suppression of the structure function. The idea was usually imple- mented through the MIT bag model, or by guessing the FIG. 33: Medium modification of form factors. Figure related structure functions. Several reviews discuss this adapted with slight modifications. Adapted idea (Arneodo, 1994; Berger and Coester, 1987; Frank- from (Frank et al., 1996). furt and Strikman, 1988b; Geesaman et al., 1995; Miller, 1984b; Mulders, 1990; Norton, 2003; Sloan et al., 1988). It was relatively easy to use this idea to compute a wide Strikman, 1985) estimated the value of r2 . Assuming variety of nuclear phenomena (Guichon and Miller, 1984; that only the blob-like configuration B h contributesi to Koch and Miller, 1985; Miller, 1984a,b, 2014; Miller and this long-ranged observable, one finds| i Kisslinger, 1983), but the calculation of each new observ- able was accompanied by the need to incorporate an ad- B r2 B ditional free parameter. The use of 6-quark models that r2 = h | | i. (40) h i 1 + 2 describe nuclear DIS led to predictions of large effects in the nuclear Drell-Yan process discussed in Sect. III.C, In the medium the potential U acts, so the value of  but little modification was seen, Fig. 27, severely limiting is changed to  . Since  <  , immersion of the M M the applicability of six-quark bag models. In addition, in nucleon in the medium suppresses| | the| | point-like compo- some applications the necessary six-quark bag probabil- nents and increases r2 . The effect is of order ( ), M ity needed to reproduce the EMC effect is so large as to which was estimatedh toi be between 2 and 5%. − conflict with knowledge of nuclei (Farrar et al., 1988). At high momentum transfer, the term 2 B J P be- comes dominant. Then the change in the formh | factor| i is For a recent study of the possible influence of hidden- color and short-ranged correlation effects at EIC energies, of order  M , which is a larger effect. The application− of the PLC-suppression idea presented see (Miller et al., 2016). in the present two-state model is schematic: it does not distinguish between the electric, G , and magnetic, G , E M IV. The EMC - SRC Correlation form factors. A more detailed evaluation was included by (Frank A. Experimental Overview et al., 1996). Medium modifications of the proton form factors were predicted as shown in Fig. 33. The impor- While there is no obvious connection between DIS scat- tant modifications shown by the red arrow occur at larger tering from quarks in the nucleus at 0.3 x 0.7 and values of momentum transfer than currently accessible B QE scattering from nucleons in the nucleus≤ at 1.≤5 x < experimentally. Fig 33 shows fairly significant effects, B 2, analysis of world data showed a remarkable correlation≤ greater than about 10% (consistent with our present (see Fig. 34) between the magnitude of the EMC effect analysis) for the individual form factors. Experimentally in nucleus A and the probability that a nucleon in that it is easier to measure the medium modifications of the nucleus is part of a 2N-SRC pair (Hen et al., 2012; We- ratio G /G . The figure shows that since both G and E M E instein et al., 2011). GM are decreased, the change in the ratio GE/GM is expected to be smaller. The strength of the EMC effect for nucleus A is char- In addition to the medium modifications, (Frank acterized as the slope of the ratio of the per-nucleon et al., 1996) also predicted the more spectacular de- deep inelastic electron scattering cross sections of nu- crease (Gayou et al., 2002; Jones et al., 2000; Punjabi cleus A relative to deuterium, dREMC /dx, in the region 0.35 xB 0.7 (Seely et al., 2009a). This slope is et al., 2005) in the free-proton ratio GE/GM with in- ≥ ≤ creasing values of Q2. proportional to the value of the cross section ratio at xB 0.5, but is unaffected by overall normalization un- certainties≈ that merely raise or lower all of the data points 3. Six-quark bags and the EMC Effect together. Table I shows data from the xA corrected EMC data base of (Hen et al., 2013b) which used the EMC data One of the earliest attempts to understand the EMC of (Gomez et al., 1994; Seely et al., 2009a). effect (Bickerstaff et al., 1984; Carlson and Havens, 1983; The SRC scale factors were determined from the Jaffe, 1983) was to hypothesize that part of the time one isospin-corrected per-nucleon ratio of the inclusive (e, e0) 30

TABLE I: A compilation of world data on SRC scaling factors, a2(A) and EMC slopes dREMC /dx. Columns 2 through 4 show the SRC scaling factors extracted from various measurements. Column 5 shows the SRC scale factor prediction of (Weinstein et al., 2011) based on the EMC-SRC correlation. Column 6 shows the world average of the EMC effect slope as compiled by (Weinstein et al., 2011), using the data of (Gomez et al., 1994; Seely et al., 2009a). See text for details.

(Frankfurt et al., 1993) (Egiyan et al., 2006a) (Fomin et al., 2012a) (Weinstein et al., 2011) (Weinstein et al., 2011) [excluding the CM EMC-SRC Prediction EMC Slope

Nucleus a2(A) a2(A) motion correction] a2(A) [dREMC /dx] column # 2 3 4 5 6 3He 1.7 ± 0.3 1.97 ± 0.10 2.13 ± 0.04 −0.070 ± 0.029 4He 3.3 ± 0.5 3.80 ± 0.34 3.60 ± 0.10 −0.197 ± 0.026 9Be 3.91 ± 0.12 4.08 ± 0.60 −0.243 ± 0.023 12C 5.0 ± 0.5 4.75 ± 0.41 4.75 ± 0.16 −0.292 ± 0.023 56Fe(63Cu) 5.2 ± 0.9 5.58 ± 0.45 5.21 ± 0.20 −0.388 ± 0.032 197Au 4.8 ± 0.7 5.16 ± 0.22 6.19 ± 0.65 −0.409 ± 0.039 EMC-SRC slope 0.079 ± 0.006 0.084 ± 0.004 σ(n+p) |x =0.7 1.032 ± 0.004 1.034 ± 0.004 σd B χ2/ndf 0.7688/3 4.895/5 cross sections on nucleus A and 3He or deuterium. scale factor is robust (Hen et al., 2012). It applies to Columns two through four of Table I show the SRC scale both SRC data sets of (Egiyan et al., 2006b) and (Fomin factors measured by (Egiyan et al., 2006a; Fomin et al., et al., 2012b). The quality of the correlation also does 2012a; Frankfurt et al., 1993). The large uncertainties not depend on the corrections applied to the SRC data. in the SRC ratios of (Frankfurt et al., 1993) are due These corrections include isoscalar cross section correc- to extrapolating data from different experiments mea- tions, center-of-mass motion corrections, and isoscalar sured at different kinematics. The SRC ratios measured pair-counting corrections. The isoscalar correction to the by (Egiyan et al., 2006a) were used in the original EMC- SRC scale factors accounts for the different elementary SRC analysis of (Weinstein et al., 2011). The later re- electron-neutron and electron-proton cross sections. This sults of (Fomin et al., 2012a) include 63Cu rather than has a negligible effect on the fit quality and the extracted 56Fe; the SRC scaling factor of 63Cu is assumed to be fit parameter. Fomin et al. did not apply this correction, the same as that of 56Fe. The values of 9Be and 197Au in arguing that short range correlations are dominated by the fifth column are those predicted by Ref. (Weinstein np pairs. Fomin et al. also argued that the SRC scale fac- et al., 2011) based on the measured EMC effect and the tors measured the relative probability of finding a high- linear EMC-SRC correlation. These predictions are in momentum nucleon in nucleus A relative to deuterium remarkable agreement with the later results of (Fomin and that these scale factors needed to be corrected for et al., 2012a). Following (Hen et al., 2012), the (Fomin the center-of-mass (cm) motion of the pair in order to et al., 2012a) results are shown without the center of mass determine the relative probability that a nucleon in nu- motion correction (i.e., including inelastic, radiative, and cleus A belongs to an SRC pair. As shown in both (Hen coulomb corrections only). Applying the SRC-pair center et al., 2012) and (Arrington et al., 2012a), including the of mass motion correction decreases the ratios by 10% to pair c.m. motion correction improves the EMC-SRC cor- 20%. relation only slightly. The EMC effect correlates imperfectly with other A- This EMC-SRC correlation gives new insight into the dependent quantities (see (Arrington et al., 2012a; Seely origin of the EMC effect. As discussed in Sect. III, many et al., 2009b) and references therein). In general, nu- different explanations of the EMC effect have been pro- clei with A 4 fall on one straight line but deuterium posed since 1983. After accounting for the standard nu- and 3He do≥ not. This is true when the EMC effect is clear effects of binding energy and Fermi motion, expla- 1/3 plotted versus A, A− , or the average nuclear separa- nations for the EMC effect fall into two general cat- tion energy. When plotting the EMC effect versus aver- egories, those that require modifications of mean-field age nuclear density, 9Be is a clear outlier (see Fig. 23). nucleons and those that require modifications of high- This indicates that the excellent correlation with the SRC momentum (large virtuality) nucleons. scale factor is not just a trivial byproduct of their mutual The linear correlation between the strength of A-dependence. the EMC effect and the SRC scale factors indi- The correlation between the EMC effect and the SRC cates that possible modifications of nucleon struc- 31

energy of the spectator nucleus (known as the removal energy (Ciofi degli Atti and Simula, 1996a)), we find

2 2 2 2 2 v(p,E) = MA (MA∗ 1) + p p M (42) − − − −  q  which reduces to A p2 v(p,E) 2M + E , (43) ≈ − A 1 2M  −  in the non-relativistic limit. The magnitude of the vir- tuality, v(p,E) increases with both the A 1 excitation energy and the initial momentum of the struck− nucleon. (C. Ciofi degli Atti, L.L. Frankfurt, L.P. Kaptari and M.I. Strikman, 2007; Frankfurt and Strikman, 1985) ob- tained a relation between the potential U of Section III.D.2 and the virtuality v(p,E) by using the extension of the Schroedinger equation to an operator form: p2 + U = E, (44) 2Mr − FIG. 34: The slope of the EMC effect (REMC, ratio of nuclear to deuteron cross section) for 0.35 xA 0.7 where M = M(A 1)/A, and U is the interaction that ≤ ≤ r − plotted vs. a2(A), the SRC scale factor (the relative both binds the nucleon to the nucleus and modifies its probability that a nucleon belongs to an SRC NN pair) structure. The simple idea behind this equation is that, for a variety of nuclei (Hen et al., 2013b). The fit if the nucleon binding energy is fixed, then the NN in- parameter, a = 0.070 0.004 is the intercept of the teraction energy, U, must become more negative as the − ± line constrained to pass through the deuteron (and is kinetic energy becomes more positive. In this work the therefore also the negative of the slope of that line). modification of nuclear properties was found to be pro- From (Hen et al., 2013b). portional to v(p,E) for moderate values of the virtuality. It should be noted that the short ranged correlations give a dominant contribution to the average nucleon virtual- ture occurs in nucleons belonging to SRC pairs. ity, which naturally leads to an approximate proportion- This implies that the EMC effect, like short range ality of the EMC effect to a2. correlations, is a short-distance, high virtuality, Comparing this equation with Eq. (43) one finds and high density phenomenon. v(p,E) U = , (45) 2Mr B. Theory Overview so that the modification of the nucleon due to the PLC 1. High momentum nucleons and PLC suppression suppression is proportional to its virtuality. Potentially large values of the virtuality greatly enhance the differ- Next we try to use the EMC-SRC correlation to better ence between m and . understand the relationship between short-ranged corre- Now we need to understand how the structure function lations measured in the A(e, e ) reaction and deep inelas- changes in the medium. In principle one needs to calcu- 0 µν tic scattering reactions. Both processes involve a probe late the hadronic tensor W and q(x) for the medium modified nucleon of Eq. (38) by replacing the state P in that strikes a nucleon of four-momentum p in the nucleus, | i Fig. 28. It is natural to expect that the medium mod- Eq. (20) by the state N M . To leading order, the change in the structure function| i will be linear in  . The ification depends on the virtuality v(p,E) of the struck M − nucleon (C. Ciofi degli Atti, L.L. Frankfurt, L.P. Kaptari hadronic part is an off-diagonal matrix element between a and M.I. Strikman, 2007): free physical nucleon, N and a free physical state N ∗ . Thus the modification| isi the product of a coefficient| thati 2 2 2 2 v p M = (PA PA 1) M . (41) depends on the medium and a term that is independent ≡ − − − − of the medium. In the (e, e0p) reaction in PWIA (see Fig. 28), the nu- These hadronic matrix elements have not yet been cleon initial momentum opposes the A 1 recoil mo- calculated. Instead we adopt a phenomenological ap- − mentum p = PA 1. Using the recoil mass MA∗ 1 = proach, based on the suppression of point-like configura- M M + E,−where− E > 0 represents the excitation− tions (Frank et al., 1996; Frankfurt and Strikman, 1985) A − 32 where the medium modified quark structure function is given by the expression q (x) = q(x) + ( )f(x)q(x), (46) M M − with the suppression of point-like components manifest by the condition df/dx < 0. so that the ratio of structure functions is given by R(x) = qM (x)/q(x), so that

dR df = (M ) . (47) dx − dx

This expression is only meaningful for xB < 0.7 where Fermi motion effects can be ignored. Given that   > FIG. 35: (color online) α solid, U dashed for Q2 = 2.7 M | | 0 (as discussed above), Eq. (47) shows that the slope− of GeV2 and p = 0. The quantity U is presented in ⊥ | | the EMC ratio is negative, consistent with observations. units of the nucleon mass M and is proportional to the (Ciofi degli Atti et al., 2007) calculated the expected virtuality v(p,E) via Eq. (45) size of the modification of Eq. (45) using the spectral function P (p,E) of (Ciofi degli Atti and Simula, 1996b) to the Frankfurt-Strikman variable α by a = α M . Then (as discussed in Section I.B.3). The average values of Md the virtuality are quite large, as can be seen from Ta- p2 + M 2 ble II. The average kinetic and removal energies in chan- pf− = ⊥ + > 0 (48) nel 1 (high excitation final states) are much larger than aMd + q 2 2 the corresponding quantities in channel 0 (low excitation p + M ps− = ⊥ > 0. (49) final states) and the high momentum components are (1 a)Md − linked to high removal energies (Ciofi degli Atti et al., In our convention q+ < 0 so that Eq. (48) tells us that 1980). (Ciofi degli Atti et al., 2007) shows that these a > 0 and Eq. (49) tells us that a < 1. Conservation of values of the virtuality, for reasonable choices of EB and + + energy tells us that p +p− +p +p = 2(M +ν), which E , can account for the EMC effect at x 0.5. f f s s− d P B ≈ leads to a quadratic equation for a: + TABLE II: The virtualities (in MeV) for channels 0 and + Md + q 2 2 (aMd + q )(1 a)Md = (p + M ). (50) 1 (see Eq. (7)) and their sum (Ciofi degli Atti et al., − Md + q− ⊥ 2007). The condition that this equation for a has real roots leads to limits on the value of p . ⊥ A hv0(p,E)i/2M hv1(p,E)i/2M hv(p,E)i/2M Fig. (35) shows the results of a specific example using 3 He -7.15 -27.44 -34.59 Q2 = 2.7 GeV2 and p = 0. Solving Eq. (50) gives the 4 ⊥ He -26.82 -42.58 -69.40 resulting values of α as a function of xB. We see that 12 C -33.17 -49.11 -82.28 α is considerably greater than one for 1.5 < xB < 1.8, 16O -31.40 -48.28 -79.68 corresponding to the plateau region of Fig. 16. Using the 40Ca -35.00 -49.54 -84.54 displayed values of α we can calculate v(p,E = p0): 56 Fe -31.66 -50.76 -82.44 0 + 2 2 208 v(p, p ) = p p− p M (51) Pb -32.87 -59.33 -92.20 − ⊥ − 2 2 + M +p⊥ where p = aMd, p− = + q−.Then the use of aMd+q − This shows that high-momentum nucleons in nuclei can Eq. (45) gives the values shown in Fig. (35): Thus, for cause the EMC effect. Now we need to find a similar re- 1.5 < xB < 1.8, we have lation between these high-momentum nucleons and the 270 MeV < U < 600 MeV. (52) plateaus observed at high xB in inclusive (e, e0) QE scat- | | tering. We first review the kinematics. We assume that Such large values of U can only arise from hard inter- the virtual photon is absorbed by one of the baryons con- actions of two nucleons,| | i.e., at short range. tained in an interacting system of two baryons M M . 2 ≈ d Thus (e, e0) at high xB is associated with short-ranged The virtual photon hits a baryon of momentum p in a correlations. Next we relate the virtuality to the observed ‘deuteron’ of momentum P , and the second, spectator plateaus in the cross section ratios. (Ciofi degli Atti and baryon has momentum p = P p. The struck nucleon s − Simula, 1996b) showed that, for large values of p : has final momentum pf = p + q. Let the plus-component | | of p be given as aM . The light-front fraction a is related n (p) n(1)(p) a (A)n (p). (53) d A ≈ A ≈ 2 D 33

This relation is explained in Sect. IX. where αn depends on Λ but not A and is completely To summarize: there is a consistent picture in which determined by the two-nucleon system. This relation is short-ranged correlations are involved with significant valid for all n, so after an inverse Mellin transform, the modification of the nuclear quark distribution function isoscalar PDFs satisfy by suppressing the point-like configurations. The key fea- 1 ture is that larger values of the nuclear excitation energy F A(x ,Q) = F N (x ,Q) + g (A, Λ)f˜ (x , Q, Λ)(56), A 2 B B 2 2 B E, associated with the short-ranged correlations, corre- spond to larger values of virtuality and therefore to more where significant deformations of the nucleon. These very same 1 2 short-ranged correlations are also responsible for the va- g2(A, Λ) = A N †N A , (57) A | | Λ lidity of Eq. (53) for large values of momentum (where
the virtuality is large), which via the logic of (Frankfurt and f2(xB, Q, Λ) is an unknown function independent of and Strikman, 1981a, 1988a; Frankfurt et al., 1993) is re- A. This feature is similar to that of our Eq. (47). Indeed, sponsible for the cross section ratio plateaus. The spec- Equation (56) was also obtained phenomenologically in tral function P (p,E) contains the information necessary Ref. (Frankfurt and Strikman, 1981a, 1988a) using the to compute both the virtuality needed to understand the impulse approximation. The equation Eq. (56) appears DIS EMC effect and the momentum probability nA(p) also in (Hen et al., 2013a; Kulagin and Petti, 2010, 2014, needed to understand the plateaus. 2006b). Note that g2 receives dominant contributions from the single nucleon density. The factorization scale of the PDF is µf = Q, while 2. Effective Field Theory Λ is the nuclear physics “ultraviolet” cut-off that sepa- rates the high energy parton physics from lower energy It is not necessary that the suppression of point-like hadronic and nuclear effects. The two scales must be sig- configurations for off-shell nucleons be the sole origin of nificantly separated for the EFT description to be valid. the EMC effect. Indeed another dynamical idea could The second term on the right-hand side of Eq. (56) is also account for the experimental findings. For example, the nuclear modification of the structure function. The the presence of non-nucleonic 6-quark clusters (Sect. III shape of distortion, i.e., the xB dependence of f2, which D) in nuclei could be important. A more general ap- is due to physics above the scale Λ, is A independent proach, using effective field theory (EFT), which is not and hence universal among nuclei. The magnitude of specific as to the underlying mechanism of medium mod- distortion, g2, which is due to physics below the scale Λ, ification has been presented (Beane and Savage, 2005; depends only on A and Λ. Chen and Detmold, 2005; Chen et al., 2016). The au- At smaller values of Q2, the previous analysis was gen- thors (Chen et al., 2016) show that the empirical linear eralized to apply to the (e, e0) cross section at large xB, relation between the magnitude of the EMC effect in deep so that inelastic scattering on nuclei and the short range correla- tion scaling factor a2 extracted from high-energy quasi- σA/A = σN + g2(A, Λ)σ2(Λ), (58) elastic scattering at x 1 is a natural consequence of B where the E (incident electron energy), x and Q2 de- scale separation and derive≥ the relationship using effec- 0 B pendence of σ is suppressed. With σ vanishing for tive field theory. i N x > 1, for both DIS and QE, Their EFT Analysis proceeds by studying the domi- B nant (leading-twist) parton distributions determined by g2(A, Λ) target matrix elements of bilocal light-cone operators. a2(A, xB > 1) = . (59) g2(2, Λ) Applying the operator product expansion, the Mellin mo- 2 ments of the parton distributions, In principle, a2 could depend on E, xB and Q . However A the EFT factorization shows that this dependence cancels n n xB A(Q) = xBqA(xB,Q)dxB, (54) at this order yielding a plateau in a2 as observed exper- h i A Z− imentally at 1.5 < xB < 2. (The influence of Fermi mo- are determined by matrix elements of local operators. tion extends the contribution of the single nucleon PDF Each of the QCD operators is matched to hadronic opera- to xB above 1, pushing the onset of the plateau to larger tors (Chen and Detmold, 2005). The relative importance values of xB.) The function a2 was also computed us- of the hadronic operators in a nuclear matrix element can ing the Green’s Function Monte Carlo method (Carlson be systematically estimated from EFT power counting. et al., 2015) and it agrees well with the data. A N The nuclear matrix element is given by Eq. (56) and the definition R(A, xB) F2 /(AF2 ), lead to the result that ≡ n n 2 x A(Q) = x N (Q) A + αn(Λ,Q) A (N †N) A Λ , h Bi h Bi h | | i dR(A, xB) = C(xB)[a2(A) 1] , (60) h (55)i dxB − 34

N has a linear relation with a2, with C(xB) = g2(2)[f20 F2 4. Summary N 0 N 2 − f2F2 ]/[F2 + g2(2)f2] independent of A and Λ (here, f 0 = df/dxB). This means that EFT naturally accounts for the linear relation between the EMC slope and the In summary, driven by the short-range correlations be- height of the plateau. However, the sign of the EMC tween two nucleons, the strong connection between the effect is not explained. EMC effect and the plateaus observed in (e, e0) scattering at high xB is both a natural consequence of the impulse approximation of scattering theory, and also of effective 3. The Isovector EMC Effect field theory. In the impulse approximation the relevant ratio is that of momentum-space densities; in the EFT This SRC-related PLC suppression model also leads to the relevant ratio is that of coordinate space densities. an explanation of the NuTeV anomaly (Sargsian, 2014c). Sect. IX shows that ratios of these are the same as long as We have discussed the dominance of the pn SRCs, rel- large values of momenta are used in the impulse approx- ative to the pp and nn correlations, for nuclear internal imation and small values of relative distance are used in momenta between 300 and 600 MeV/c, that is caused the EFT. This means that the relation shown in Fig. 34 is by the effects of the tensor force. The pp and nn com- derived using two very different techniques. The fact that ponents of the NN SRC are strongly suppressed since using two different technical approaches, each driven by they are dominated by the central NN potential with short-range physics, leads to the same conclusion, gives relative L = 0. The resulting picture for nuclear mat- significant credence to the interpretation that the same ter consisting of protons and neutrons at densities in short-range physics accounts for both the EMC effect and which inter-nucleon distances are about 1.7 fm is rather the QE cross section plateaus. unique: it represents a system with suppressed pp and nn but enhanced pn interactions. Using this idea Sargsian The underlying mechanism of the distortion of the nu- (Sargsian, 2014b) predicted two new properties for the cleon structure is not yet established, and could occur nuclear momentum distributions for momenta between from PLC suppression or from other mechanisms. Nev- the Fermi momentum and about 600 MeV/c. There is ertheless, it is very clear that the relation shown in Fig. 34 an approximate equality of p- and n- momentum distri- is no accident. There is a true underlying cause of the butions weighted by their relative fractions in the nucleus EMC effect and the observed plateaus in ratios of (e, e0) x = Z/A and x = (A Z)/Z : scattering cross sections. p n − x nA(p) x nA(p) (61) p p ≈ n n with d3pnA(p) = 1. The probability of a proton being C. Are the nucleons in the correlated pair really nucleons? in a high momentum NN SRC is inversely proportional R to its relative fraction, xp, and can be related to the According to the logic presented here, most of the cor- momentum distribution in the deuteron nD(p): related pair consists of nucleons, but the part that is A 1 responsible for the EMC effect consists of non-nucleonic np (p) = a2(A, N)nD(p) (62) 2xp configurations. This conclusion is valid for both classes of models of the EMC effect: the mean-field based or and similarly for neutrons. The main prediction of SRC based. The non-nucleonic configurations could be a Eq. (62) is that high momentum protons and neu- medium modified single-nucleon, or NN or N N con- trons became increasingly unbalanced as the ratio (N ∗ ∗ ∗ figurations, or even more complex 6-quark configurations. Z)/(N +Z) increases. Using this equation one can calcu-− late the fraction of the protons having momenta greater than the Fermi momentum as D. Determining the structure function of a free neutron 1 3 Pp(A, N) a2(A, N) d p nD(p)Θ(p kF ), ≈ 2xp − R (63) Determining the structure function of the neutron is and similarly for neutrons. For example in Iron, Pp = challenging because a free neutron target does not ex- 23%, and Pn = 20%. ist. Experimentalists have therefore used deuteron or The energetic protons in neutron rich nuclei will result 3He targets to extract the neutron structure. This im- also in the stronger nuclear modification of u-quarks as plies that our knowledge of the neutron structure func- compared to d-quarks and the effect grows with A. The tion is intimately connected with medium effects in light predicted effects also can be checked in parity violating targets. As we shall see, medium effects in the deuteron deep inelastic scattering off heavy nuclei (Cloet et al., must be accounted for accurately if one hopes to correctly 2012; Souder, 2016) (see Section VI.A.5). understand the free neutron structure function. 35

1. The Deuteron IMC Effect

The deuteron In Medium Correction (IMC) effect 1.04 refers to the difference between the DIS cross section for

the deuteron and the sum of the DIS cross sections for a � 1.00 free proton and neutron (Frankfurt and Strikman, 1985; +

Melnitchouk et al., 1994b). The term IMC was intro- �

duced in Ref. (Weinstein et al., 2011) which showed that / 0.96

one can use the EMC-SRC correlation as a phenomeno- � logical tool to constrain the deuteron IMC effect, and 0.92 thus extract the free neutron structure function. Fol- lowing Weinstein et al. (Weinstein et al., 2011), we can 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 extrapolate the linear fit to the EMC-SRC correlation to x the limit of a2(A) 0. This is the limit of no correla- tions, which is equivalent→ to a free proton-neutron pair. FIG. 36: The deuteron IMC ratio The intersection of this limit with the y-axis is there- p Rd = F d/(F n + F ) as extracted from the BONuS fore the IMC ratio of the free proton-neutron pair to the EMC 2 2 2 data. Total systematic uncertainties are shown as a deuteron. band arbitrarily positioned at 0.91 (blue). The yellow The a2(A) 0 extrapolation to the y-axis of the EMC- band shows the CJ12 (Owens et al., 2013) limits SRC correlation→ gives dR /dx = 0.070 EMC a2(A)=0 expected from their nuclear models. The black points 0.004. Since the EMC effect is linear for 0.3 −x 0.±7 ≤ A ≤ are the combined 4- and 5-GeV data, whereas the red for all nuclei with A > 2, we assume that the EMC effect points are the 4-GeV data alone. The dashed blue line is also linear in this region for the deuteron. This implies shows the calculations of Ref. (Kulagin and Petti, that the EMC effect for the deutereon relative to a free 2006a). The solid line (black) is the fit to the black proton plus neutron can be written as: points for 0.35 < xB < 0.7. From (Griffioen et al., σd 2015). = 1 a(xB b) for 0.3 xB 0.7, σp + σn − − ≤ ≤ where σd and σp are the measured DIS cross sections 2 2 for the deuteron and free proton, σn is the free neu- Q > 1 GeV and W > 1.4 GeV, see Fig. 36 (Grif- tron DIS cross section that we want to extract, a = fioen et al., 2015). A linear fit for 0.35 < x < 0.7 yields dRd /dx = 0.1 0.05 where the uncertain- dREMC/dx a2(A)=0 = 0.070 0.004, and b = 0.34 0.02 EMC | | ± ± 1 − ± is the average value of xB where the EMC ratio is unity . ties comes from the fit. This result is consistent with This implies that σd/(σp + σn) decreases linearly from 1 the IMC prediction of 0.07. For x < 0.5 the EMC d − to 0.97 as xB increases from 0.3 to 0.7. We can then use ratios REMC agree within uncertainties with those ob- this relationship to extract the free neutron cross section tained using more stringent cuts in W . The ratio for in this xB range, as shown in the next section. xB > 0.5 continues the trend of the lower-xB data, with The uncertainty quoted above for the IMC slope is due a hint of the expected rise above xB = 0.7 as seen in A to the EMC and SRC data and to the fit. It does not REMC for heavier nuclei, but these high-xB values are include any uncertainty due to corrections applied to the more uncertain because there are fewer data points for EMC and SRC data. As stated above, if we include the resonance averaging. proposed correction for a2(A) due to the c.m. motion of the correlated pair, then the fit parameter increases by 25% and so does the free proton plus neutron EMC 2. The Free Neutron Structure Function effect. These effects are discussed in detail in (Hen et al., 2012). If the structure function F2 is proportional to the DIS Following the prediction of the IMC effect, the BONuS cross section (i.e., if the ratio of the longitudinal to trans- collaboration (Tkachenko et al., 2014) published their verse cross sections is the same for n, p and d [see discus- experimental extraction of the IMC effect, measured at sion in (Geesaman et al., 1995)]), then the free neutron n 2 structure function, F2 (xB,Q ), can also be deduced from the measured deuteron and proton structure functions and from the deuteron IMC effect: 1 The xA correction does not significantly change the slope, a , d 2 p 2 of the EMC-SRC correlation, and it increases the b parameter n 2 2F2 (xB,Q ) [1 a(xB b)]F2 (xB,Q ) F2 (xB,Q ) = − − − by less than the uncertainty reported in Ref. (Weinstein et al., [1 a(xB b)] 2011) − − (64) 36

n p which leads to The IMC based extraction of F2 /F2 , extrapolated in d the region of x < 0.3, is compared in Fig. 38 to several F (x ,Q2) B n 2 2 B F (x ,Q ) 2 F p(x ,Q2) [1 a(xB b)] other experimental and phenomenological extractions of 2 B = 2 B − − − . (65) p 2 this ratio. Also shown are several QCD predictions. F2 (xB,Q ) [1 a(xB b)] − − see (Holt, 2013; Roberts et al., 2013) for details. This is only valid for 0.35 x 0.7. ≤ B ≤

Entries 0 Mean x 0 Mean y 0 RMS x 0 RMS y 0 p 2 / F n

2 0.6 F

IMC Corrected (This Work) 0.5

Fermi motion 0.4 corrected

n p 0.3 0.4 0.5 0.6 0.7 FIG. 38: F2 /F2 as a function of xB. Results from the xB IMC and other phenomenological extractions are compared to selected theoretical predictions. From (Holt, 2013). See (Holt, 2013; Roberts et al., 2013) for details. FIG. 37: The ratio of neutron to proton structure n 2 p 2 functions, F2 (xB,Q )/F2 (xB,Q ) as extracted from the measured deuteron and proton structure functions, F d and F p. The filled symbols show F n/F p extracted 2 2 2 2 3. The d/u ratio at large xB by (Weinstein et al., 2011) from the deuteron In Medium Correction (IMC) ratio and the world data for n p The ratio of the neutron structure function, F2 , to the d 2 2 p F2 /F2 at Q = 12 GeV (Arrington et al., 2009). The proton structure function, F , is particularly interesting n p 2 open symbols show F2 /F2 extracted from the same as it can be related, within the parton model, to the data correcting only for nucleon motion in deuterium ratio of the d-quark and u-quark distributions. The lat- using a relativistic deuteron momentum density ter provides a unique opportunity for studying the flavor (Arrington et al., 2009). From (Weinstein et al., 2011). and spin dynamics of quarks in the nucleon, with the d/u quark distribution ratio in particular being very sensitive to different mechanisms of spin-flavor symmetry break- Fig. 37 shows the ratio of F n/F p extracted by (Wein- 2 2 ing (Holt and Roberts, 2010; Melnitchouk and Thomas, stein et al., 2011) using the IMC-based correction and the 1996b). Q2 = 12 GeV2 ratio F d/F p from Ref. (Arrington et al., 2 2 Historically, proton DIS data placed strong constraints 2009). Note that the ratio F d/F p is Q2-independent 2 2 on the u-quark distribution, while neutron structure from 6 Q2 20 GeV2 for 0.4 x 0.7 (Arring- B functions were used to constrain the d-quark distribu- ton et al.≤, 2009).≤ The dominant uncertainty≤ ≤ in this ex- p d tion and form the d/u ratio. Specifically, the d/u ratio traction is the uncertainty in the measured F2 /F2 . The in the valence quark dominance domain (i.e., at large xB IMC-based correction increases the extracted free neu- p ) was extracted from the F n/F structure function ratio tron structure function (relative to that extracted using 2 2 using: the deuteron momentum density (Arrington et al., 2009)) p by an amount that increases with xB. This is quali- n F2 /F2 = [1 + 4(dv/uv)]/[4 + (dv/uv)], tatively similar to the recent extraction of (Cosyn and n Sargsian, 2016). Thus, the IMC-based F2 strongly fa- where the absence of free neutron targets meant that n vors model-based extractions of F2 that include nucleon the neutron structure function was not measured di- modification in the deuteron (Melnitchouk and Thomas, rectly, but instead extracted from deuterium DIS data. 1996a). However, uncertainties in the nuclear corrections in the 37 deuteron, such as those associated with nucleon off- in agreement with the models of (Cloet et al., 2005a; Far- shell effects and the large-momentum components of the rar and Jackson, 1975) which predict intermediate values deuteron wave function, give rise to significant uncertain- of d/u between the SU(6) symmetry and scalar diquark ties in the resulting d/u ratio for xB & 0.5 (Accardi et al., dominance limits. 2011). To rectify the situation, Hen et al., (Hen et al., 2011a) used the phenomenological IMC corrected extraction of V. Existing searches for medium-modified electromagnetic n p F2 /F2 discussed above as an added constraint on the form factors extraction of the d/u ratio in the global analysis of the CTEQ-JLab collaboration (Accardi et al., 2011). We have shown that the experimental and theoretical New data on charged lepton and W boson asymme- evidence indicates that the structure of the nucleon is try measured at the Tevatron (Abazov et al., 2013, 2014, modified by its immersion in a nucleus. The only mod- 2015) are sensitive to the large-xB d/u ratio with no nu- els that account for the EMC effect, the plateaus of the clear uncertainties (Accardi et al., 2016). high xB (e,e’) reaction and the lack of a medium ef- Fig. 39 shows the d/u ratio at large-xB extracted from fect in the nuclear Drell-Yan data are those involving a global QCD analysis using DIS data without (CJ11, short-ranged correlations. Nevertheless, the task of un- (Accardi et al., 2011)) and with (CJ11+IMC, (Hen et al., derstanding the EMC and SRC effects is not complete. 2011a)) the IMC constraint and using the new asymme- The available models need to be improved (to be dis- try data with no nuclear corrections applied (CJ15, (Ac- cussed in Sect. VI.B). We need models that are suffi- cardi et al., 2016) and CT14, (Dulat et al., 2016)). As can ciently complete that they can explain both the EMC be seen, while the various extractions somewhat differ at effect, the nuclear Drell-Yan data and also predict and large-xB, the IMC constraints and the new asymmetry account for new independent phenomena. data both contrain the CJ11 analysis similarly. If the nuclear medium modifies the bound nucleon To summarize, the use of the IMC-extracted neutron structure functions (and thus their wave functions), then structure function directly constrains the d-quark PDF it almost certainly will modify their electromagnetic form for x 0.7, and indirectly for x 1. We find the d/u . factors. All of the medium modification models modifica- ratio in the limit x 1 to be 0.23→0.09 at the 90% con- tion of bound electromagnetic form factors, see Sect. III. fidence level, in overall→ agreement± with new extractions These effects could be manifest in quasi-elastic nucleon using charged lepton and W boson asymmetry data and knockout (e, e0N) cross sections and in the inclusive lon- gitudinal A(e, e0) response. The influence of nucleon modification on the nuclear elastic form factor can not be detected because the distribution of nucleons in the nucleus is imprecisely known. This section will discuss the experimental evidence for modification of bound nucleon form factors.

A. Polarization transfer in the (~e,e0~p) reaction

Polarization transfer in the H(~e,e0~p) reaction was used to measure the ratio of the free proton electromagnetic form factors GE/GM with much smaller systematic un- certainties than previous methods (Perdrisat et al., 2007). This technique was then applied to measure the ratio of bound proton electromagnetic form factors us- ing the quasielastic A(~e,e0~p) reaction (Dieterich et al., 2001; Malace et al., 2011; Paolone et al., 2010; Strauch FIG. 39: d/u ratio at Q2 = 12 GeV2 with the full et al., 2003; Strauch, 2012). The ratio of the longitudi- theoretical uncertainty from Ref. (Accardi et al., 2011) nal and transverse polarization transfers is proportional (black) and with the IMC constraint at the 90% C. L. to the ratio of G /G for the free proton, P /P (red) from (Hen et al., 2011a). Also shown for E M x0 z0 G /G (Perdrisat et al., 2007). For a bound proton,∝ comparison are recent extractions that do not include E M one must also correct for the effects of meson exchange nuclear correactions from the CJ15, (Accardi et al., currents, isobar configurations, and especially final state 2016) and CT14, (Dulat et al., 2016)) PDF extractions. interactions. After using a model to correct for these 38

2000) can only explain the data if they include medium- 1.1 Madrid RDWIA (RLF) MAMI Madrid RDWIA (RLF) + QMC modified form factors. They calculated the induced po- E93-049 Madrid RDWIA (RLF) + CQS larization and the polarization transfer ratio using the H Schiavilla (2010) E03-104 ) z 1.0 unmodified but offshell cc1 and cc2 (De Forest, 1983b) /P' x current operators and the optical potentials of (Horowitz,

/ (P' 1985; Murdock and Horowitz, 1987) to account for final He )

z 0.9 state interactions. No charge exchange effects (photon /P' x knocks out neutron, which undergoes a charge exchange

(P' reaction) were included. This unmodified calculation 0.8 agreed with the induced polarization data when using

0 1 2 3 the cc1 current operator. However, good agreement with 2 2 Q (GeV/c) the measured value of R was only achieved by including

Madrid RDWIA (RLF) either the QMC (Lu et al., 1999) or CQS (Smith and 4 3 E93-049 Figure 4: He(~e, e0p~) HMadrid polarization-transfer RDWIA (RLF) + QMC double ratio R as a function Miller, 2004) medium modified form factors. of Q2 from Mainz [33] andMadrid Je RDWIA↵erson (RLF) Lab + CQS experiments E93-049 [32] (open Schiavilla (2010) E03-104 symbols) and E03-104 [34] (filled circles). The data are compared to calcu- lations from the Madrid group [38–40], using the cc1 (lower set of curves) 0.00 Schiavilla et al. (Schiavilla et al., 2005) calculated Py and cc2 (upper set of curves) current operators, and Schiavilla et al. [37]

= 0) and R using DWIA. They computed the final state in-

as in Fig.m 3. Not shown are a relativistic Glauber model calculation by the Ghent group [41] and results from Laget [26] which give both a value of (p teractions using an optical potential that includes both y

R 1. P ⇡ spin-independent and spin-dependent charge exchange

-0.05 terms. However, they updated their calculation in 2010 cc1 yields the largest prediction for Py in absolute value and describes the with new parameters. While their calculation describes data well; possibly hinting at the importance of the lower spinor compo- both Py and R without medium modified form factors, nents in this relativistic0 calculation;1 see [40]. We2 note that these3 RDWIA calculations provide also good descriptionsQ2 (GeV/c) of,2 e.g., the induced polariza- its significance is decreased because they did not follow 12 tions as measured at Bates in the C(e,e0p~) reaction [40, 47] and of ATL the standard procedure (Austern, 1970) of independently in 16O(e, e p) as previously measured at JLab [48]. While the polarization- 40 3 4 3 Figuretransfer 3:FIG. observablesHe(e, 40: e0p~) are TheH induced expected measured polarization to be sensitiveHe( data to~e,e from possible0 Je~p)↵erson nucleonH Lab medium exper- constraining the parameters of the optical potential they polarization-transferimentmodifications, E93-049 results [32] along of the double with RDWIA results ratio calculation fromR experiment(upper including E03-104 medium-modified panel) [36]. and The used to describe the final state interactions. Thus our dataform are factors compared show onlyto calculations some small from e↵ect Schiavilla on the inducedet al. [37] polarization. and the Madrid The inducedgroupdata are [38–40] polarization also using compared the cc1 withP(lowery the(lower set results of curves) panel) of a calculation and ascc2 (upper a from function set Schiavilla of curves) ofet view is that the results of the nuclear polarization exper- Qcurrental.2:[37] open operators. (shaded symbols: band). In-medium (Dieterich That form model factors useset from variational al. the, 2001; QMC wave [6] Strauch functions(solid curve) for iments strongly indicate that medium effects do influence andthe boundCQS [10] three- (dashed and curve)four-nucleon models systems, were used non-relativistic in two of the MECMadrid and calcu- free etlations. al., 2003) Note that and the filled comparison circles: is made (Malace for missinget momentum al., 2011;p 0 electromagnetic form factors. We eagerly await new ex- m ⇡ Paoloneand thatet the al. experimental, 2010). The data havedata been are corrected compared for the to spectrometer DWIA periments with improved precision and at larger values of acceptance for this comparison. 7 calculations from (Schiavilla et al., 2005) (updated in pmiss which would confirm or rule out this interpretation. 2010) using unmodified form factors and from the Madridwere taken group during E03-104 (Caballero to studyet these al. and, 1998; help significantly Udias et reduce al., sys- Experiments performed at the Mainz Microtron 1999b;tematic Udias uncertainties and in Vignote, the extraction 2000) of Py using. The data the are cc1 compared (lower with (MAMI) using the A1 beam-line (Yaron et al., 2016) results of a relativistic distorted-wave impulse approximation (RDWIA) cal- setculation of curves) by the Madrid and cc2group (upper [38–40]. In set this of model curves) FSI are off-shell incorporated measured the polarization transfer ratio R for deuterium currentusing an updated operators version in of combination the RLF relativistic with optical unmodified potentials [42, 43] and 12C at lower Q2 (Q2 = 0.175 and 0.4 GeV2) but (blackthatdashed distort the line), final nucleon QMC wave modified function; the (red MRW solid optical line) potential and of higher virtuality than at Jefferson Lab. For deuterium, [44], used in [34], does not yield an as good description of Py as the modi- CQSfied RLF modified potential shown (red here. dashed Charge-exchange line) in termsmedium are not form taken into the ratio R decreases significantly with virtuality and is factors.account in See the Madrid text for RDWIA details. calculation; From preliminary (Strauch, studies 2012). show, how- consistent with that previously measured on 4He. This ever, that they are of small e↵ect in this model [45]. Calculations are shown for choices of cc1 and cc2 current operators as defined in [46]. The choice indicates that the effect in nuclei is due to the virtuality of the knocked-out proton and not due to the average effects, the polarization double6 ratio nuclear density. The deuteron calculations (Arenhovel et al., 2005) predict this decrease and associate most of P 0 P 0 R x / x (66) it with FSI (Yaron et al., 2016). The 10% differences ≡ P P 1 ≈  z0 A  z0  H between the data and calculations may indicate the need should be sensitive to medium modification of the form for in-medium modifications. The carbon data is still factor ratio. The induced polarization Py (measured under analysis. Other double polarization experiments in the (e, e0~p) reaction) should be more sensitive to fi- were not sensitive to the effects of nucleon modification nal state interactions and much less sensitive to medium (Mayer et al., 2017; Mihovilovic et al., 2014; Passchier modification effects. et al., 2002). 4 3 Fig. 40 shows the He(~e,e0~p) H double ratio R and the induced polarization, Py, measured at small values of Jefferson Lab experiment E12-11-002 will measure missing momentum (pmiss < 150 MeV/c) over a range of polarization-transfer observables as a function of virtu- Q2. Relativistic distorted-wave-impulse approximation ality for both 4He and 2H and will measure the proton (rDWIA) calculations by the Madrid group (Caballero recoil polarization at Q2 = 1.8 GeV2 to help us better et al., 1998; Udias et al., 1999b; Udias and Vignote, understand the effects of medium modifications and FSI. 39

B. Polarization transfer in the (~e,e0~n) reaction where H is the nuclear Hamiltonian, the ground state energy is taken as 0, and for simplicity we assume that A complementary experiment would be the mea- neutrons do not contribute. The non-relativistic formula- surement of polarization transfer to the neutron in tion is only valid when q2 Q2. Since R is proportional ≈ L quasielastic scattering in the (~e,e0~n) reaction. Clo¨et to the square of GE, its sensitivity to medium effects is et al. (Cloet et al., 2009c) studied possible in-medium greater than that of the polarization transfer measure- changes of the bound neutron electromagnetic form- ments. factor ratio with respect to the free ratio, the superratio The Coulomb sum is the integral over all values of ν 2 (GE∗ /GM∗ ) / (GE/GM ). At small values of Q this super- (including the inaccessible time-like regime where ν > ratio depends on the in-medium modifications of the neu- q ): tron magnetic moment and the effective electric and mag- | | Z netic radii. The superratio of the neutron is dominated RL(q) dνRL(ν, q) iq (ri rj ) = = A e · − A . by the expected increase of the electric charge radius G2 (q2) G2 (q2) h | | i E R E i,j=1 in the nuclear medium and is found to be greater than X (68) one. In contrast, the proton superratio is predicted to be smaller than one. A comparison of high-precision mea- Splitting Eq. (68) into terms with i = j and i = j we get: 2 4 3 6 surements of the reactions H(~e,e0~n)p and He(~e,e0~n) H RL(q) 3 3 iq (r r0) would test these predictions. = Z + Z(Z 1) d rd r0e · − ρ (r, r0), G2 (q2) − 2 However, a major drawback to nuclear polarization E Z transfer measurements, no matter whether the proton or (69) neutron is detected, is that medium modifications that where ρ is the two proton density function, see Eq. (80). affect both G and G will cancel in the ratio. See 2 E M At large enough momentum transfer the second term Fig. 33, for example. vanishes as 1/q4, so that one finds the CSR:

RL(q) C. The (e, e0) reaction and the Coulomb Sum Rule (CSR) lim = Z. (70) Q2 G2 (q2) →∞ E This sum rule (De Forest and Walecka, 1966; McVoy Since electron scattering cannot measure the cross sec- and Van Hove, 1962) states that the integral of the tion in the time-like region, the Coulomb sum is properly A(e, e0) longitudinal response function at fixed momen- defined (Cloet et al., 2016) as an integral over ν from en- tum transfer over all energy transfers should equal the ergies just above the elastic peak to q : | | total charge of the nucleus, Z. The first CSR experiment q | | R (ν, q ) (Altemus et al., 1980) observed that the sum rule was q L SL( ) = dν 2 2 | | 2 2 . (71) | | + ZG (Q ) + NG (Q ) “quenched”, i.e., they measured less than Z. This indi- Zν Ep En cated that the cross section for scattering from a bound The quantity S can be correctly be compared with the nucleon was significantly less than the free cross section. L results obtained from electron scattering. Thus, (Cloet et al., 2016) say that the first hints of QCD The initial motivation to measure the Coulomb Sum effects in nuclei came from quasielastic electron scattering Rule (De Forest and Walecka, 1966) was to learn about on nuclear targets (Altemus et al., 1980; Meziani et al., ρ . However, the recent focus has been to learn about 1984; Noble, 1981). However, later work cast doubt on 2 nucleon medium modification at large values of the mo- this result. mentum transfer where the effect of ρ is negligible. The (e, e ) inclusive cross section can be written as 2 0 (Cloet et al., 2016) discuss the interesting history of d2σ Q4 the theory. Calculations (Horikawa and Bentz, 2005; =σMott RL(ν, q ) Saito et al., 1999), in which the internal structural prop- dΩdν q 4 | | | | (67) erties of bound nucleons are self-consistently modified Q2 θ + + tan2 R (ν, q ) by the nuclear medium unsurprisingly predict significant 2 q 2 2 T | |  | |   quenching of the CSR. However, calculations that assume an unmodified nucleon electromagnetic current (Carlson where σ is the Mott cross section, R and R are Mott L T et al., 2002; Do Dang et al., 1987; Kim et al., 2006; Mi- the longitudinal and transverse response functions, and haila and Heisenberg, 2000), including the state-of-the- θ is the electron scattering angle. In the non-relativistic art Green function Monte Carlo (GFMC) result for 12C limit of the impulse approximation (Bertozzi et al., 1972; from Ref. (Lovato et al., 2013, 2016), find modest or no De Forest and Walecka, 1966) one has quenching of the CSR. Most recently (Cloet et al., 2016) Z Z used an NJL model in the medium to find a dramatic re- iq ri iq ri 2 2 R (ω, q) = A e · δ(ω H) e− · A G (q ), duction of the Coulomb sum rule for q & 0.5 GeV, driven L h | − | i E | i=1 j=1 by changes to the bound-proton Dirac form factor. X X 40

The experimental status of the CSR has been unclear. struck nucleon (the nucleon that absorbed the virtual The initial measurements found quenching of the CSR for photon), and thereby learn more about the microscopic 12C, 40Ca and 56Fe (Altemus et al., 1980; Meziani et al., origin of the EMC effect. 1984). However, a reanalysis of these data (Jourdan, The simplest example for such a process is DIS on the 1995, 1996), utilizing an alternative prescription for the deuteron. If we can detect a recoil nucleon with mo- Coulomb corrections, concluded that there is no quench- mentum p that did not interact in the DIS reaction and ing. The analysis of the Coulomb corrections in those did not have a final state interaction (i.e., a spectator), works was later challenged (Aste et al., 2005; Aste, 2008; then we know that the struck nucleon had initial momen- Wallace and Tjon, 2008). These papers support the con- tum p. We can then measure the DIS cross section for clusion that quenching of the CSR occurs as reported in scattering− from a nucleon in the nucleus as a function of Ref. (Morgenstern and Meziani, 2001). New results at its initial momentum. This will allow us to extract F2 high momentum transfer and on a variety of nuclear tar- and hence the quark distributions. In particular F2 can gets from Jefferson Lab Experiment E05-110 (Choi et al., be measured as a function of virtuality. This experiment 2005) are eagerly anticipated. Verification or disproof of thus provides an opportunity to test the importance of the CSR quenching should reveal critical aspects of nu- the effects of virtuality that are discussed above. cleon modification in nuclei. This was initially studied with 5.7 GeV electrons inci- dent on deuterium, measuring the scattered electron and the recoil proton with the CLAS spectrometer (Klimenko VI. Future directions in nuclear deep inelastic scattering et al., 2006). While this measurement did not have the and detecting short-ranged correlations kinematic reach to unambiguously measure a change in the nucleon structure function, they did show that pro- A. Experiment tons emitted at large angles, θpq > 120◦ (where θpq is There are several different experimental approaches to the angle between the proton and the virtual photon), understanding the EMC-SRC correlation and the ori- were predominantly spectators. Later theoretical works gin of the EMC effect. The most promising approach support this observation (Cosyn and Sargsian, 2011; Palli is to directly test the EMC-SRC correlation by measur- et al., 2009). ing the change in bound nucleon structure function with In practice, experiments will measure the ratio of cross nucleon momentum using tagged structure function mea- sections at fixed recoil momentum and different values of surements. xB0 where The second approach is to test other predictions of Q2 Q2 models of the EMC effect by measuring other quanti- x0 = = B 2p qµ 2[(M E )ν + p q] ties related to nucleon modification, including the bound µ d − S S · ratio of electric to magnetic elastic form factors using po- is the value of xB in the frame of the struck nucleon, larization transfer A(~e,e0~p) and the Coulomb Sum Rule. Md is the deuteron mass, and ES and pS are the energy Lastly, we can learn more about SRC and about the and momentum of the spectator nucleon. This data will EMC effect individually in several ways. The first way be used to extract (Hen et al., 2014b, 2011b) is to extend EMC and SRC inclusive measurements to more nuclei over a wider range of momentum transfer. F bound(x hi, q2, p ) We can also extend semi-exclusive and exclusive SRC 2 0B 1 S = free hi 2 measurements in a similar manner to abteined more de- F2 (xB ,Q1) tailed information, especially about the potential isospin hi 2 free low 2 σDIS(x0B ,Q1, pS) σDIS (xB ,Q2) dependence of the EMC effect, SRCs, and their correla- RFSI (72) σ (x low,Q2, p ) · σfree(xhi,Q2) · tions. We can select the nucleons we study by measuring DIS 0B 2 S DIS B 1 the polarized EMC effect and we can measure the isospin where σfree is the free-nucleon DIS cross section, R dependence of the EMC effect in asymmetric nuclei by DIS FSI is a correction factor for the effects of final state interac- measuring parity violating deep inelastic scattering. low tions, xB 0.3 where the EMC effect is very small (i.e., ≈ hi where the EMC ratio is very close to 1), and x0B > 0.45. 1. Tagged Structure function Measurements By measuring the ratio of the bound to free nucleon structure functions as a function of spectator momentum The EMC Effect is measured in inclusive (e, e0) DIS (i.e., of nucleon initial momentum), these experiments from a nucleon in a nucleus. In order to learn more about will answer the crucial question of which nucleons are the DIS reaction, we can “tag” the reaction by detecting modified and to what extent. Little momentum de- a recoiling nuclear fragment in coincidence with the scat- pendence would imply that the mean-field nucleons are tered electron. By choosing the nuclear fragment and modified and large momentum dependence would imply kinematics wisely, we can restrict the initial state of the that SRC nucleons are modified. 41

scattering angles from 160◦ to 170◦. The expected results are shown in Fig. 41a. A second category of experiments consists of measuring the tagged EMC ratio. We can “tag” different reaction mechanisms by detecting either a spectator nucleon or a recoil A 1 nucleus. The main idea is that the electron scatters− from a quark in one nucleon. If that nucleon belongs to an SRC NN pair, then its partner nucleon will leave the nucleus. If that nucleon does not belong to an NN SRC pair, then the A 1 nucleus is much more likely to recoil intact. In either− case, we will need to fully account for FSI effects proper interpertation the reasults of such measurements requires full (a) Bound proton structure via d(e, e0n )X scattering recoil understanding of many-body FSI effects that, to the best of our knowlege, so far were only studied for the deuteron. Instead of the inclusive cross section ratio, the tagged EMC ratio is

σ (e, e p )/A R = A 0 S σd(e, e0pS)/2

integrated over spectator momenta and angles. Typi- cally, backward angles, θpq > 120◦, are chosen to mini- mize FSI. If the spectator is a proton and has momentum greater than 300 MeV/c, then it almost certainly is expected to belonged to an np SRC pair. If nucleon modifica- tion is due to nucleons belonging to SRC pairs, then nu- cleon modifcation should be the same in deuterium and in the heavier nucleus and therefore the tagged EMC ra- 0 (b) Bound neutron structure via d(e, e precoil)X tio should be independent of xB and should be equal to scattering a2(A), the relative probability of finding a nucleon in an SRC pair in nucleus A relative to d. FIG. 41: The expected results from future Jefferson Lab tagged DIS measurements (Hen et al., 2014b, 2011b). The biggest Large uncertainty in interpretting these The dashed line is obtained from the color screening tagged EMC measurements is stems from the possibil- model (Frankfurt and Strikman, 1985), the dotted line ity that the fragments of the struck nucleon will break up is from the color delocalization model (Close et al., another SRC pair as they exit the nucleus, significantly 1983), and dot-dashed the off-shell model (Melnitchouk increasing the number of backward nucleons. This ef- et al., 1994a). From (Hen et al., 2014b, 2011b) fect should be smaller for light nuclei. An addi- tional A larger complication arises from the nuclear spectral function that associated high-momentum nu- cleons with large excitation energies that will be dis- There are two approved Jefferson Lab experiments tributed to the different fragments and need to be to measure this reaction. Experiment E12-11-107 (Hen taken into account. et al., 2011b) will measure neutron modification by de- tecting the scattered electrons in the Hall C magnetic If the measured spectator is an A 1 nucleus, then spectrometers and the spectator protons in a set of GEM the struck nucleon almost certainly did− not belong to an detectors and scintillators covering scattering angles from SRC pair. Assuming one can overcome the above com- 4 about 80◦ to 170◦. The expected results are shown in plications, comparing the tagged EMC ratio for He with Fig. 41b. Experiment E12-11-003A (Hen et al., 2014b) spectator (proton+deuteron) and or with spectator 3He, will measure proton modification by detecting the scat- can give further insight as to whether nucleon modifica- tered electrons in the CLAS12 forward detector and the tion depends on the struck nucleon momentum or on the spectator neutrons in a large scintillator array covering struck nucleon SRC pairing. 42

ent nuclear effects. Medium and heavy nuclei (A 10) exhibit properties of nuclear saturation and can be≥ rela- tively well described using effective theories for strongly interacting many-body Fermi systems. However, light nuclei span a wide range of nuclear densities and asym- metries, with some nuclei exhibiting a rich cluster-like substructure. In the case of the EMC effect, special care should be given to “standard” nuclear structure effects that do af- fect the DIS cross-section ratio and can therefore po- tentially mimic an isospin dependent. Therefore, any measurement of the ratio of the EMC effect in isospin asymmetric nuclei (e.g. 3He/3H, 48Ca/40Ca etc.) should be compared to a calculation of standard nuclear effects that take into account differences in the proton and neu- tron momentum distributions and pairing probabilities. These can be studies indirectly by inclusive QE (e,e’) FIG. 42: Phase-space of nuclei considered for future reactions or directly using semi-inclusive and exclusive EMC/SRC measurements at Jefferson Lab as a function (e,e’N) and (e,e’pN) reactions. In the case of light nuclei of their proton (Z) and neutron (N) numbers. The two ab-initio few-body calculations can also provide relevant inserts focus on the light and medium mass nuclei information. refims. For light nuclei one can systematicaly study a An additional advantage of light-nuclei studies is the series of symmetric nucli and the detailed effect of the ability to compare the experimental results with detailed addition of one neutorn (proton). For medium mass ab-initio nuclear structure calculations. Assuming reac- nuclei the horizontal and vertical bands mark nuclei tion mechanisms such as FSI, MEC and others are under with similar numebr of protons (neutrons) and a varing control, such comparisons of experiment and theory can number of neutrons (protons) allowing to study the offer significant insight into the underlying microscopic effect of nuclear asymmetry. physics. For heavy nuclei, such ab-initio calculations of short-range nuclear structure are still limited, but rapid progress is being made (Carlson et al., 2015; Hagen et al., 2015; Wiringa et al., 2014b). 2. Inclusive EMC and SRC Measurements In addition to extending the range of nuclei measured, it is also important to extend the measured Q2 range. The inclusive EMC and SRC measurements described This is especially important for SRC studies where the in sections III and II were performed on a limited number minimum initial momentum depends strongly on Q2 (see of nuclei and, in the case of SRC measurements and the Fig. 13). The SRC cross section ratios of (Egiyan et al., JLab EMC measurements, in a limited kinematic range. 2003) were measured at 1.4 Q2 2.6 GeV2 with Therefore, it is natural to extend both EMC and SRC most of the data at Q2 < 2≤ GeV2≤(see Fig. 16a). measurements to additional nuclei over a wider kinemat- They observed flat plateaus in the cross section ratio for ical range. 1.5 xB 1.9, which corresponds to 250 pmin 500 Fig. 42 shows nuclei that can or will be measured at MeV/c≤ which≤ is where we expect tensor correlations≤ ≤ to Jefferson Lab as a function of their proton (Z) and neu- dominate. By contrast, the SRC cross section ratios of tron (N) numbers. A wise selection of nuclei allows (Fomin et al., 2012a) which were measured at Q2 = 2.7 for a systematical experimental study of SRC and the GeV2 exhibit “plateaus” that are not quite as flat, espe- EMC effect for fixed nuclear asymmetry as a function of cially for heavier nuclei (see Fig. 16a). At Q2 = 2.7GeV2, mass number and for fixed mass number as a function 1.5 xB 1.9 corresponds to 325 pmin 700 MeV/c, of asymmetry. The planned Jefferson Lab measurements which≤ extends≤ beyond the tensor correlations≤ ≤ region into (Arrington and Day, 2006; Arrington et al., 2010; Pe- the central correlations region. Measuring the Q2 de- tratos et al., 2010; Solvignon-Slifer and Arrington, 2011) pendence of the SRC plateaus will help us quantitatively will systematically measure both the size of the EMC ef- relate the experimental results to detailed ab-initio nu- fect and the height of the SRC plateau over many nuclei clear structure calculations - a needed comparison that from 3He and 3H to 208Pb, covering a wide range of mass was not done to date. numbers and nuclear asymmetries (N/Z). Measurements The Q2-dependence of the EMC effect has been stud- with unstable nuclei at other laboratories could signifi- ied over a wide kinematical regime. However, there are cantly extend the available range of nuclear asymmetry. still several intriguing questions about higher twist effects Light and heavy nuclei can exhibit significantly differ- that should be studied systematically. The Jefferson Lab 43

6 GeV EMC effect measurements included data with in- and detecting its correlated partner, either with electron variant mass, W > 1.4 GeV, a region that is dominated or proton probes, A(e, e0pN) or A(p, 2pN). By extend- by resonance production rather than DIS. The fact that ing the range of missing momentum we can study the the measured EMC ratios agreed with the SLAC data, transition from tensor dominance (at 300 p 500 ≤ miss ≤ measured at higher W , showed that resonance contri- MeV/c) to the scalar repulsive core (at higher pmiss). butions largely cancel in the A/d ratio. By covering a By focusing on lower pmiss, we can map the transition broader kinematic range, the future 12 GeV measure- fro the mean-field to the SRC-dominated domain (the nu- ments will help quantify this issue. A review of the pos- clear “Migdal jump”). By extending the A-dependence sibility of studying the large xB at the lHC is presented of SRC pair abundancies and properties we can learn in (Freese et al., 2015). about SRC-pair quantum numbers and provide data for a quantitative theory of SRCs.

3. Semi-Inclusive and Exclusive SRC Measurements 4. Polarized EMC Measurements In the context of the EMC effect, the interpretation of planed future experiments in isospin asymmetric nuclei Motivated by open questions about the EMC effect and (e.g. 3He/3H, 48Ca/40Ca) requires understanding the the “proton spin crisis”, Jefferson Lab will perform the differences in the proton and neutron momentum distri- first measurement of the spin-dependent EMC effect uti- butions and SRC pairing probabilities, as these enter into lizing CLAS12 in Hall B with 11 GeV polarized electrons the baseline calculation of “standard” nuclear effects and and polarized targets (Kuhn and Brooks, 2014). They EMC models calculations, that are to be compared with will determine the ratio of the double-spin asymmetries the experimental data. in 7Li (using 7LiD) in which a highly polarized proton The required information about high momentum nu- is embedded in the nuclear medium, and on the proton 6 cleons and SRC in nuclei can be obtained by scattering an (using LiH). The double spin asymmetry is measured as electron or other probe from a nucleus and detecting one dσ dσ or more of the ejected nucleons. A(e, e0p) experiments A = ↓⇑ − ↑⇑ k dσ +dσ can measure the amounts of high momentum nucleons in ↓⇑ ↑⇑ different nuclei and how that changes with nuclear isospin and is approximately equal to the ratio of polarized to asymmetry. 7Li 7Li unpolarized structure functions: g1 /F1 . Many sys- The fact that the A(e, e0p) reaction is mainly sensitive tematic uncertainties will cancel in the asymmetries and to the protons in nuclei whereas the (e, e0) reaction is sen- in the ratios of asymmetries. Together with the unpo- sitive to all nucleons in nuclei make their measurements larized structure function (also to be measured at Jef- complementary and crucial to allow for a detailed study 7Li ferson Lab), they will also extract g1 and, using a so- of the dependence of SRC effects on the nuclear asym- phisticated modern wave function model, extract the in- metry. While more challenging to perform, A(e,e’n) mea- p 7Li medium proton spin structure function function g k surements can complement the other reactions and yield 1 for a proton bound in 7Li. They will cover a kinematic additional information on the role of protons and neu- range of 1 < Q2 < 15 GeV2 and 0.06 < x < 0.8. trons in asymmetric nuclei. B Mean field models of nucleon modification predict One Jefferson Lab experiment (Hen et al., 2014a) will 3 3 stronger effects than in the unpolarized structure func- measure H and He(e, e0p) as a function of pmiss in kine- tions. On the other hand, since nucleons in tensor corre- matics where FSI are small in order to determine the ra- 3 3 lations tend to have opposite spin to the overall nuclear tio of the He and H momentum distributions. In the spin, the EMC effect could be minimal or even in the op- naive SRC picture, this ratio should be two at low pmiss 3 3 posite direction. These data will provide new constraints because there are twice as many protons in He as in H on models for the EMC effect, some of which predict and it should decrease to one at high pmiss because there 3 3 that medium modifications of quark distributions depend are two pn pairs each in He and H. strongly on the quark helicities (see Fig. 43). Similar experiments in medium mass nuclei could mea- sure how the number of high momentum protons changes as you add eight neutrons from 40Ca to 48Ca and by 5. Parity Violating Deep Inelastic Scattering adding six more protons from 48Ca to 54Fe (Hen et al., 2016b). Ongoing Jefferson-Lab CLAS data-mining anal- There is some evidence that u- and d-quark distribu- ysis of A(e,e’n) and A(e,e’p) scattering off 12C, 27Al,56Fe tions are modified differently in asymmetric nuclei. The- and 208Pb is expected to provide new insight into isospin oretically, since protons move faster than neutrons in asymmetry effects on SRCs. neutron-rich nuclei, if nucleon modification depends on We can gain more information about SRC pairing in nucleon virtuality (as in the PLC model), then we expect nuclei by knocking out a high-initial momentum nucleon protons, with 2 u- and 1 d-quarks, to be more modified 44

distributions. In the symmetric nucleus limit

+ + 9 2 12 uA dA a1 4 sin θW + − + + ... ' 5 − − 25 uA + dA

where uA refers to all the up quarks in the nucleus and the superscript + refers to the sum of the quark and anti-quark distributions. Thus the parity violating asym- metry is sensitive to the difference between the u and d quark distributions in the nucleus.

B. Theory

The review of the theory presented here shows that there is a strong connection between the cause of the EMC effect and the short-ranged correlations that cause the high xB plateau in (e,e’) scattering on nuclei. Never- FIG. 43: The expected results of the polarized EMC theless, there are gaps in almost every part of the theory, effect measurement at Jefferson Lab. The ratio of the from the initial state wave function, to the modification 7 parallel double spin asymmetry A for Li(~ ~e,e0) to of nucleon structure, to the need to include the effects of k ~p(~e,e0), normalized by multiplying it with the “naive” final state interactions. We therefore present an outline unpolarized structure function ratio for 7Li over of the necessary improvements. hydrogen, plotted vs xB. The models are NNM (naive The EMC effect is a modification of nucleon struc- nuclear model with no fermi motion), SNM (standard ture functions. Obtaining an understanding of this ef- nuclear model with fermi motion and kinematical fect therefore requires a working understanding of the binding energy effects), QMC (Quark-Meson Coupling valence sector of the free nucleon wave function, so that model), and MSS (x-rescaling (Fanchiotti et al., 2014)). the effects of the medium on the relevant components Figure adapted from (Kuhn and Brooks, 2014). can be correctly included. Lattice calculations, e.g., (Ji, 2013; Lin et al., 2015), and the Dyson-Schwinger ap- proach (Cloet and Roberts, 2014) are making progress on computing free nucleon parton distributions. It also than neutrons. would be necessary to build nucleon models that are eas- Experimentally, the NuTeV experiment compared neu- ily related to the output of these Euclidean-space theo- trino and anti-neutrino DIS off an Iron target and ex- ries, e.g. (Burkardt et al., 1997; Hobbs et al., 2016). A tracted a value of the Weinberg mixing angle that differs twenty-first century calculation of medium modifications from the standard model by about 3σ (Zeller et al., 2002, cannot be made without inputs from such models. 2003). While this led to much excitement and attempts The calculation of deep inelastic scattering from nuclei to relate it to physics beyond the standard model, re- needs to be improved in several different ways. For ex- cently it was shown that an isospin dependent EMC ef- ample, the calculations using the PLC-suppression model fect that affects protons more than neutrons could resolve have been made mainly for xB = 0.5 (C. Ciofi degli Atti, the anomaly (Cloet et al., 2009b). L.L. Frankfurt, L.P. Kaptari and M.I. Strikman, 2007), A measurement of parity violation in A(e, e0) DIS where effects of Fermi motion nearly vanish. To under- would directly measure the d u difference as a func- stand the EMC ratios discussed above it is necessary to − tion of xB (Riordan et al., 2016). The difference in the be able to make accurate calculations for a range of val- left-right asymmetry for helicity +1 and 1 electrons is ues of 0.3 xB 0.7. So far this has been done (Freese ≤ ≤ proportional to the product of the photon− and Z am- et al., 2015) by assuming that no medium modification plitudes divided by the square of the photon amplitude. occurs for xB < 0.45 and linearly interpolating the region 2 3 This asymmetry will be 10 to 10 parts per million for between 0.45 < xB < 0.65. Calculations need to han- DIS scattering from a heavy nucleus: dle finite-sized nuclei without resorting to infinite nuclear matter calculations using a local density approximation. 2 2 Such a program would require computation of nuclear GF Q 1 (1 y) APV a1(x) + − − a3(x) spectral functions for finite-sized nuclei. This would in- ≈ −4√2πα 1 + (1 y)2  −  volve intensive numerical work, so it would be important to present such spectral functions in an easily accessible where y = 1 E/E0, and a and a depend on the quark manner. − 1 2 45

We have seen that only models with medium modifi- DIS from a nucleus is very different than DIS from a col- cation arising from short-ranged effects can handle both lection of free nucleons; this is the EMC effect which is the EMC effect and high xB (e, e0) scattering. How- parameterized in terms of the slope of the EMC ratio, R, ever models in which the medium modification is driven of bound to “free” cross sections. This slope cannot be by mean-field effects give an excellent description of the explained unless the internal quark structure of a bound EMC effect, see e.g. (Cloet et al., 2009a, 2005b, 2006). nucleon differs from that of a free nucleon. It would not be realistic to think that the ultimate accu- Quasi-elastic scattering, in which a nucleon is knocked rate description would make use of only one of the two out of the nucleus intact, reveals plateaus in the cross possible ideas. Therefore it is important to build models section ratios of nuclei to deuterium at large values of of medium modifications of nucleon wave functions that xB that correspond to scattering from short range cor- includes both mean-field effects and the effects of cor- related (SRC) two-nucleon pairs. Different experiments relations. The necessary model of nuclei would need to show that the slope of the EMC effect is linearly pro- be consistent with nuclear saturation properties, include portional to the height of the plateaus! Further studies non-nucleonic degrees of freedom and have those rela- showed that the two-nucleon pairs consist of a neutron tivistic effects needed to compute nuclear deep inelastic and a proton. scattering cross sections. A review of the available experimental and theoreti- Many treatments of final state interactions for exclu- cal evidence shows that the relation between the EMC sive reactions (e.g., (e, e0p) and (e, e0pN)) use complex slope, dR/dxB, and the SRC plateau height is no ac- optical potentials, which automatically violate current cident. There is an underlying cause of both effects: conservation. To fully understand spectroscopic factors the influence of strongly correlated neutron-proton pairs. and nucleon-nucleon correlations it is necessary to ensure These correlated pairs are temporary high-density fluc- that the reaction theory models conserve current. We tuations in the nucleus in which the internal structure of also need to better understand electromagnetic current the nucleons is briefly modified. This conclusion needs to operators in models of the nucleon-nucleon interactions be quantified by future experiments and improved theo- that employ low-momentum cutoffs. retical analyses that are discussed in this article. There is a need to understand higher twist effects in The connection between the EMC effect and nucleon- nuclei, so we can understand why the EMC ratios mea- nucleon correlations is very profound. Although the bind- sured at JLab are nearly the same as those measured at ing energy of a nucleon is less than a percent of its mass, much higher energies at SLAC and CERN. the fact that the nucleon is made of quarks and gluons In addition to improving our understanding of the the- is manifest in two distinct sets of phenomena, via ex- oretical underpinnings of the causes of the EMC-SRC periments that have been repeated several times. The correlation, it is necessary to explore the implications of direct influence of the quark presence in nuclei is now the EMC-SRC correlations and of pn dominance in SRC. established. The possible inversion of the kinetic energy sharing in This presence is a subtle effect as it must be, given asymmetric nuclei could significantly affect several sub- the generally small deviation of R from unity, and does fields of physics. In astrophysics the nuclear symmetry not arise via the usual low-energy, low momentum trans- energy is of fundamental importance. It describes the fer nuclear physics observables: binding energy, spec- change in energy of a nuclear system when a proton is tra, radii, electroweak transition rates, etc. Nonetheless, replaced by a neutron. np-SRC dramatically reduce the the quark presence cannot be denied. We expect that a kinetic part of the symmetry energy (Hen et al., 2015c) deeper understanding of the EMC/SRC connection will and work is ongoing to understand other effects. Addi- ultimately lead to an improved understanding of the na- tional implications of SRCs on nuclear systems include ture of confinement of light quarks. the nuclear response to neutrino scattering (Fields et al., The Jefferson-Lab 12GeV program includes a series 2013; Fiorentini et al., 2013), cooling rates of neutron of approved experiments targeted at improving our un- stars, contact interactions in Fermi systems (Frankfurt derstanding of the EMC effect, SRCs and their connec- et al., 2008; Hen et al., 2015a) and more. While the dis- tion. The forthcoming results of these experiments are cussion of these effects extends beyond the scope of this expected to shad new light on the origin of the EMC ef- review, they are extensively discussed in the literature. fect and provide stringent constraints on current of future theoretical calculations.

VII. The way we think it is and the ways to check VIII. Acknowledgments This article has focused on explaining two seemingly unrelated phenomena: lepton-nucleus deep inelastic scat- We thank our many colleagues for their efforts in ac- tering (DIS) and quasi-elastic (QE) electron-nucleus scat- complishing the research discussed in this review and also tering at large values of xB, and their surprising relation. for the many insights that they have provided through 46 discussions. This work of was supported by the U. S. De- (Haftel and Tabakin, 1970) partment of Energy Office of Science, Office of Nuclear 2 2 Physics under Award Numbers DE-FG02-97ER-41014, k Q2(z) + p Q0(z) I02(p, k) = I20(p, k) = Q1(z), DE-FG02-94ER40818 and DE-FG02-96ER-40960 and by 2pk − the Israel Science Foundation (Israel) under Grants Nos. (74)

136/12 and 1334/16. 2 2 2 with z (p + k + µ )/(2pk), and Qi are Legendre functions≡ of the second kind in the conventions of that reference. The result, Eq. (73) corrects errors in (Hen IX. Appendix et al., 2015b). The errors do not affect the qualitative statements made in the cited paper, as we now demon- A. Understanding the np relative wave function strate. We use Eq. (73) to estimate quantities of interest. The aim of this Appendix is to provide a qualtitative We note the asymptotic property: limp I02(p, k) = 1 (k2 + µ2)/p2 + . Thus the integrand→∞ of Eq. (73) is explanation that the momentum space wave function of − ··· the deuteron, a very weakly bound system, has a sig- dominated by large values of p and diverges unless there nificant high momentum k tail. Indeed one sees an ap- is a cutoff. This means that V00(k0, k) is approximately a 4 constant, independent of k and k0. This is the signature proximate k− behavior of the deuteron density for large values of k. This tail persists in nuclei because of short of a short ranged interaction. We expose this feature in ranged correlations between nucleons. more detail by assuming that for the important regions of the integral appearing in Eq. (73) by treating the vari- 4 2 A 1/k density comes from 1/k in the wave function ables k, k0 as small compared to the cutoff momentum. which can be obtained if the nucleon-nucleon interaction p2 Then I02(0, p) 2 2 , and is a delta function in coordinate space, as occurs in lead- ≈ p +µ ing order EFT or in the effective range expansion behav- 32f 4 M p2dp p2 2 ior. Such approximations are valid only at very small V (k, k0) M (1+ ), 00 ≈ − µ4π2 MB + p2 p2 + µ2 ··· values of momentum 1/re k 1/a, where there is Z0   approximate scale invariance, where a is the scattering (75) where we have cut off the linearly divergent integral for length of about 5 fm and re is the effective range of about 2 02 2 momenta p > M and represents terms of ( k +k ). 2 fm. The 1/k behavior of the wave function emerges at ··· O M 2 large values of k due to the second-order effects of the of All realistic models of the NN interaction employ some the one pion exchange (OPE) contribution to the tensor sort of a cutoff, and a mass scale of the nucleon mass is typical of one-boson exchange potentials (Machleidt, potential VT . The Schoedinger equation for the spin-one two-nucleon system, which involves S and D state com- 1989; Machleidt et al., 1987). Thus V00(k, k0) is ap- ponents, can be expressed as an equation involving the S proximately independent of its momentum arguments, the hallmark of short-ranged interactions. The use of state only by using ( B H0) ΨD = VT ΨS , where B − − | i | i Eq. (75) provides an approximate upper limit. is the binding energy of the system and H0 is the Hamil- 4 tonian excluding the tensor potential. Thus one obtains The resulting asymptotic 1/k dependence of the 1 square of the wave function can be seen by using the an effective S state potential: V00 = VT ( B H0)− VT , − − − Lippmann-Schwinger equation in the form Eq. (4), where VT connects the S and D states. The in- termediate Hamiltonian H0 is dominated by the effects of 1 3 k ψ k ( B H )− d k0V (k, k0) k0 ψ (76) the centrifugal barrier and can be approximated by the h | Si ≈ h | − − 0 00 h | Si V00(0,0) 3 V00(0,0) 3/2 kinetic energy operator (Brown and Jackson, 1976). This − k2 d k0ψS(k0) = −R k2 (2π) ψS(r = 0) ≈ B+ M B+ M second-order term is large because it contains an isospin R factor (τ τ )2 = 9, and because S2 = 8 2S . Eval- where the subscript S refers to the S state and the inte- 1 · 2 12 − 12 − uation of the S-state potential, neglecting the small ef- gral over all momenta, k0 leads to a proportionality to the fects of the central potential in the intermediate D-state, coordinate-space wave function at the origin. In terms of yields the usual S-state radial wave function u(r) we have u(r) 1 ψS(r = 0) = lim . (77) 4 2 r 0 √ 32f p dp → r 4π V00(k, k0) M 4 2 2 I02(k, p)I20(p, k0), ≈ − µ π MB + p u(r) Z Using known wave functions, we find limr 0 = (73) → r 2 3/2 where M is the nucleon mass, f 0.08 is the square (0.0267, 0.0584, 0.0792) fm− for the Nijmegen, of the πN coupling constant, µ is≈ the pion mass, and Reid93 (Stoks et al., 1994a), and Argonne V18 (Wiringa ILL0 are partial wave projections of the OPEP extensor et al., 1995b) potentials respectively. The result Eq. (75) interaction in momentum space. These are evaluated in shows the 1/k2 dependence of the wave function, with 47 overall strength determined by the detailed potential The two-body density in coordinate space is given by models. The density is the square of the wave function 1 1/k4 with an overall strength varying by a factor of 9, ρ(2)(x, y) = Ψ δ(x x )δ(y y ) Ψ . ∼ A(A 1)h | − i − j | i depending on the potential used. Thus we find a high − i=j 4 X6 momentum 1/k behavior far beyond the validity of the (80) effective range approximation. Potentials without this high-momentum density either have a very weak tensor The integral of the two-body density over x yields the force or a cutoff at low momenta. density ρ(y). The correlation function C(x, y) is the de- We may check the rough validity of these findings by viation of the two-body density from the mean field ap- computing the D state probability, PD: proximation:

1 (2) PD = ψs VT 2 VT ψS C(x, y) = ρ (x, y) ρ(x)ρ(y). (81) h | (B+H0) | i − 2 32f 4 3 2 M p2dp p2 4 2 (2π) ψ (r = 0) 2 2 2 (78) The quantity C(x, y) vanishes if the wave function Ψ µ π 0 (B+ p )2 p +µ ≈ M can be represented as a product of single-nucleon wave| i R   We evaluate PD using u(r = 0) for each of the Nijmegen, functions. Furthermore the stated normalization condi- Reid93 (Stoks et al., 1994a), and Argonne V18 (Wiringa tions lead to the result et al., 1995b) potentials respectively. Numerical evalua- tion of Eq. (78) equation yields PD = (2, 10, 18) % for dxC(x, y) = 0, dyC(x, y) = 0. (82) the three potentials repsectively. The actual value for Z Z all of these potentials is about PD = 6%. These results It is useful to also define the probability ρ2,1(r) that if a show that qualitative treatment here is adequate only for nucleon is at a given position, another one is separated rough estimates that maintain the qualitative idea that by a distance r. the iterated effects of OPEP produce the 1/k4 behavior 1 of the deuteron density. The results of this sub-section ρ2,1(r) 4πr2A Ψ i=j δ(r ri rj )Ψ ≡ h | 6 − | − | i depend on the chosen scale (M here). Choosing a suf- 3 = d Rρ2(RP+ r/2, R r/2). (83) ficiently softer scale would modify the high-momentum − dependence of the wave function.A detailed comparison where R is the centerR of mass position of the two-nucleon of the momentum dependence of known deuteron wave system. functions is presented in (Hen et al., 2015b). The same kind of analysis can be done in momen- Results similar that the relevant interaction matrix tum space. Evaluation of ρ(x) requires the square of element is approximately independent of its momen- the coordinate-space representation of Ψ , while that of | i tum arguments have been obtained previously. Mosel’s n(k) requires the momentum-space representation of the group (Konrad et al., 2005; Lehr et al., 2000, 2002) as- same wave function. The probability for a nucleon to sumed this independence and used it to help to clarify have a momentum k is given by the basic, fundamental origins of the nucleon spectral n(k) = 1 Ψ A δ(k k ) Ψ . (84) functions and the high-momentum tails. Using a con- A h | i=1 − i | i stant interaction matrix element, along with the Fermi- It is convenient to defineP a two-body density n (K, κ) gas model, and solving the relevant Dyson equation gave 2 in momentum space, which gives the probability of two high-momentum tails with a density 1/k4, and spec- nucleons having a total momentum of K and a relative tral functions essentially identical to∼ those of more de- momentum κ: tailed computations.

n2(K, κ) = 1 B. Basic terminology A(A 1) Ψ i=j δ(K/2 + κ ki)δ(K/2 κ kj) Ψ − h | 6 − − − | i P (85) We define some basic terms. The probability to find a nucleon at a coordinate x (where this notation includes Experimentalists defined a correlation as existing if the spatial position, nucleon spin and isospin) is given by system has κ K, with κ > kF and K < kF .  It is also useful to consider the integrated quantity: 1 A ρ(x) = Ψ δ(x x ) Ψ , (79) 3 Ah | − i | i n2,1(κ) d Kn2(K, κ) i=1 ≡ X 2 κ = A(A 1) Ψ i=Rj δ(ki kj 2 ) Ψ , (86) where Ψ is the relevant nuclear wave function. The − h | 6 − − | i quantity| ρi(x) is known as the density. The normalization which is the momentumP space version of Eq. (83). is dxρ(x) = 1, where the integral includes a sum over A specific model for the two-nucleon density is used nucleon spin and isospin. in the analysis of the data relevant to this review. For R 48 small relative distances r one writes the two-nucleon wave the results of the independent pair approximation appear function Ψ(R, r) in the following form as the first term in the hole-line expansion. We explain how this works. Consider two-nucleons in Ψ(R, r) = FA(R)ψD(r), r << RA, (87) nuclear matter, which interact independently of the other nucleons (except for the influence of the Pauli principle). where RA is the radius of the nucleus, and the often used assumption is that at short distances all relative wave The two-nucleon Hamiltionian, h, is given by functions are the same as that of the deuteron D. In this h = h + h (93) model 0 1 P 2 p2 h0 = 4M , h1 = M + Qv, (94) ρA (r) = d3RF 2 (R) ψ2 (r) a (A) ψ2 (r), r R 2,1 A D ≡ 2 D  A Z where P is the center-of-mass momentum operator, p is (88) the relative momentum operator, v is the two-nucleon potential, and Q is an operator that projects both nu- In momentum space cleon momenta to be greater than the Fermi momentum, kF . Since the two-nucleon Hamiltonian is a sum of two terms, h = h + h , that commute the solution to the Ψ(K, k) = F (K)ψ (k), k 1/R (89) 0 1 A D  A Schroednger equation, h ψ = E ψ , is a product: | i | i where thee tilde denotesf Fourierf transform and the mo- ψ(R, r) = FA(R)χ(r) (95) mentum variables are canonically conjugate to R and r. The one-body density nA(k1) is given by where 2 3 nA(k1) = d k2 Ψ(K, k) h0F (R) = EcmFA(R), h1χ(r) = χ(r),E = Ecm + 

2 2 (96) 3 R = d P FA (P ) ψD (ek1 P/ 2) (90) − with 2 2 2 3 R d P FA (P )fψD(k1 )f = a2(A) ψD (k1) , (91) iK R ≈ ψ(R, r) = e · χ(r), (97) R where k1 1f/RA is assumed f and the relation f in terms  where we suppress notations regarding spin and isospin of a2(A) is an example of Parceval’s theorem. to simplify the discussion. In general the function χ(r) The next step is to relate the quantities nA(k1) and contains all values of angular momentum and has both ρ2,1(r). The use of Eq. (88)and Eq. (91) leads immedi- ately to the result short ranged and long-ranged aspects. The essence of Eq. (87) is that for small values of r all relative wave A | | ρ2,1(r) nA(k) functions look like the deuteron wave function: a2(A) = D = , (r RA, k > 1/RA)(92). r2,1(r) nD(k)  lim χ(r) = γψD(r), (98) r d The early workers (Frankfurt et al., 1993) used the ratio  of momentum-space densities, and recent workers (Chen where γ represents the probability amplitude that the et al., 2016) use the coordinate space version, but both wave function χ corresponds to the deuteron quantum are the same in the leading-order approximation of each numbers. approach. It is necessary to introduce a single-particle, mean-field operator U to extend this idea to finite-sized nuclei. In that case, Eq. (97) is often replaced (see e.g. (Haxton C. Why center-of-mass and relative coordinates factorize et al., 1980) by

We provide a qualitative explanation of the factoriza- ψ(R, r) = αβ Cαβφα(r1)φβ(r2)χ(r), (99) tion inherent in Eq. (87). Start with the non-relativistic nuclear Hamiltonian with only two-nucleon forces, and where φα,β are solutionsP of the single-particle equation, consider infinite nuclear matter. The basic assumption Cαβ are coefficients computed using the shell model. The is the independent pair approximation. The idea is that single-particles vary over the size of the nucleus, while the the average separation between nucleons d = 1.7 fm, so variations of χ(r) 1 occur over the range of the nucleon- − that when one of the nucleons of the pair makes a close nucleon interaction. If the size of the nucleus is much encounter with a third particle the collision occurs under larger than this range Eq. (95) remains true. In these ap- conditions such that the original pair had no interactions plications the Miller-Spencer correlation function (Miller at all (Gomes et al., 1958). This idea was formally cod- and Spencer, 1976) has often been used to represent χ(r). ified by Bethe and co-workers (Bethe, 1971), such that 49

References ables in deuteron electrodisintegration,” Eur. Phys. J. A23, 147–190, arXiv:nucl-th/0407053 [nucl-th]. Arneodo, Michele (1994), “Nuclear effects in structure func- Abazov, Victor Mukhamedovich, et al. (D0) (2013), “Mea- tions,” Phys. Rept. 240, 301–393. surement of the muon√ charge asymmetry in pp¯ → W+X → Arrington, J, F. Coester, R.J. Holt, and T.-S. H. Lee (2009), µν + X events at s=1.96TeV,” Phys. Rev. D88, 091102, “Neutron structure functions,” J. Phys. G 36, 025005. arXiv:1309.2591 [hep-ex]. Arrington, J, A. Daniel, D. B. Day, N. Fomin, D. Gaskell, Abazov, Victor Mukhamedovich, et al. (D0) (2014), “Mea- and P. Solvignon (2012a), “Detailed study of the nuclear surement of the W Boson Production Charge√ Asymmetry dependence of the emc effect and short-range correlations,” in pp¯ → W + X → eν + X Events at s = 1.96 TeV,” Phys. Rev. C 86, 065204. Phys. Rev. Lett. 112 (15), 151803, [Erratum: Phys. Rev. Arrington, J, and D. Day (2006), “Inclusive Scattering from Lett.114,no.4,049901(2015)], arXiv:1312.2895 [hep-ex]. Nuclei at x > 1 in the quasielastic and deeply inelastic Abazov, Victor Mukhamedovich, et al. (D0) (2015), “Mea- regimes, Jefferson Lab Experiment E12-06-105,”. surement of the electron charge asymmetry in p√p¯ → Arrington, J, D. Gaskell, and A. Daniel (2010), “Detailed W + X → eν + X decays in pp¯ collisions at s = studies of the nuclear dependence of F2 in light nuclei, Jef- 1.96 TeV,” Phys. Rev. D91 (3), 032007, [Erratum: Phys. ferson Lab Experiment E12-10-008,”. Rev.D91,no.7,079901(2015)], arXiv:1412.2862 [hep-ex]. Arrington, J, D.W. Higinbotham, G. Rosner, and Accardi, A, L. T. Brady, W. Melnitchouk, J. F. Owens, M. Sargsian (2012b), “Hard probes of short-range nucleon- and N. Sato (2016), “Constraints on large-x parton dis- nucleon correlations,” Prog.Part.Nucl.Phys. 67, 898–938, tributions from new weak boson production and deep- arXiv:1104.1196 [nucl-ex]. inelastic scattering data,” Phys. Rev. D93 (11), 114017, Aste, Andreas, Cyrill von Arx, and Dirk Trautmann (2005), arXiv:1602.03154 [hep-ph]. “Coulomb distortion of relativistic electrons in the nuclear Accardi, A, W. Melnitchouk, J. F. Owens, M. E. Christy, electrostatic field,” Eur. Phys. J. A26, 167–178, arXiv:nucl- C. E. Keppel, L. Zhu, and J. G. Morf´ın(2011), “Uncertain- th/0502074 [nucl-th]. ties in determining parton distributions at large x,” Phys. Aste, Andreas W (2008), “Coulomb distortion effects in quasi- Rev. D 84, 014008. elastic (e,e’) scattering on heavy nuclei,” Nucl. Phys. A806, Acciarri, R, et al. (ArgoNeuT) (2014), “Detection of back-to- 191–215, arXiv:0710.1261 [nucl-th]. back proton pairs in charged-current neutrino interactions Ciofi degli Atti, C, L. L. Frankfurt, L. P. Kaptari, and M. I. with the ArgoNeuT detector in the NuMI low energy beam Strikman (2007), “On the dependence of the wave function line,” Phys. Rev. D90 (1), 012008, arXiv:1405.4261 [nucl- of a bound nucleon on its momentum and the EMC effect,” ex]. Phys. Rev. C76, 055206, arXiv:0706.2937 [nucl-th]. Akulinichev, S V, and S. Shlomo (1990), “Nuclear Binding Ciofi degli Atti, C, E. Pace, and G. Salme (1980), “Realistic Effect in Deep Inelastic Lepton Scattering,” Phys. Lett. nucleon-nucleon interactions and the three-body electrodis- B234, 170–174. integration of H-3,” Phys. Rev. C21, 805–815. Alde, D M, et al. (1990), “Nuclear dependence of dimuon pro- Ciofi degli Atti, C, E. Pace, and G. Salm`e(1991a), “y -scaling duction at 800-GeV. FNAL-772 experiment,” Phys. Rev. analysis of quasielastic electron scattering and nucleon mo- Lett. 64, 2479–2482. mentum distributions in few-body systems, complex nuclei, Altemus, R, A. Cafolla, D. Day, J. S. Mccarthy, R. R. Whit- and nuclear matter,” Phys. Rev. C 43, 1155–1176. ney, and J. E. Wise (1980), “LONGITUDINAL AND Ciofi degli Atti, C, and S. Simula (1996a), “Realistic model TRANSVERSE INELASTIC ELECTRON SCATTERING of the nucleon spectral function in few- and many-nucleon FROM FE-56,” Phys. Rev. Lett. 44, 965–968. systems,” Phys. Rev. C 53, 1689. Alvioli, M, C. Ciofi degli Atti, L. P. Kaptari, C. B. Mezzetti, Ciofi degli Atti, C, S. Simula, L.L. Frankfurt, and M.I. Strik- and H. Morita (2013), “Nucleon momentum distributions, man (1991b), “Two-nucleon correlations and the structure their spin-isospin dependence, and short-range correla- of the nucleon spectral function at high values of momen- tions,” Phys. Rev. C 87, 034603. tum and removal energy,” Phys. Rev. C 44, R7. Alvioli, M, C. Ciofi degli Atti, and H. Morita (2008), “Proton- Ciofi degli Atti, Claudio (2015), “In-medium short-range dy- neutron and proton-proton correlations in medium-weight namics of nucleons: Recent theoretical and experimental nuclei and the role of the tensor force,” Phys. Rev. Lett. advances,” Phys. Rept. 590, 1–85. 100, 162503. Ciofi degli Atti, Claudio, and S. Simula (1996b), “Real- Alvioli, Massimiliano, Claudio Ciofi degli Atti, and Hiko istic model of the nucleon spectral function in few and Morita (2016), “Universality of nucleon-nucleon short- many nucleon systems,” Phys. Rev. C 53, 1689, arXiv:nucl- range correlations: the factorization property of the nu- th/9507024 [nucl-th]. clear wave function, the relative and center-of-mass mo- Aubert, JJ, et al. (1983), Phys. Lett. B 123, 275. mentum distributions, and the nuclear contacts,” Phys. Austern, N (1970), Direct Nuclear Reaction Theories (Wiley- Rev. C94 (4), 044309, arXiv:1607.04103 [nucl-th]. Interscience). Anastasio, M R, L. S. Celenza, W. S. Pong, and C. M. Baghdasaryan, H, et al. (CLAS Collaboration) (2010), “Ten- 0 Shakin (1983), “RELATIVISTIC NUCLEAR STRUC- sor correlations measured in 3he(e, e pp)n,” Phys. Rev. TURE PHYSICS,” Phys. Rept. 100, 327–401. Lett. 105 (22), 222501. Anderson, E R, S. K. Bogner, R. J. Furnstahl, and R. J. Perry Ball, Richard D, et al. (2013), “Parton distributions with LHC (2010), “Operator Evolution via the Similarity Renormal- data,” Nucl. Phys. B867, 244–289, arXiv:1207.1303 [hep- ization Group I: The Deuteron,” Phys. Rev. C82, 054001, ph]. arXiv:1008.1569 [nucl-th]. Beane, Silas R, and Martin J. Savage (2005), “Dvcs- Arenhovel, Hartmuth, Winfried Leidemann, and Edward L. dissociation of the deuteron and the EMC effect,” Nucl. Tomusiak (2005), “General survey of polarization observ- 50

Phys. A761, 259–268, arXiv:nucl-th/0412025 [nucl-th]. C. Ciofi degli Atti, (2015), “In-medium short-range dynam- Bedaque, Paulo F, and Ubirajara van Kolck (2002), “Effec- ics of nucleons: Recent theoretical and experimental ad- tive field theory for few nucleon systems,” Ann. Rev. Nucl. vances,” Phys. Rep. 590, 1. Part. Sci. 52, 339–396, arXiv:nucl-th/0203055 [nucl-th]. C. Ciofi degli Atti, L.L. Frankfurt, L.P. Kaptari and M.I. Benhar, O, A. Fabrocini, and S. Fantoni (1989), “The Nu- Strikman, (2007), “On the dependence of the wave function cleon Spectral Function in Nuclear Matter,” Nucl. Phys. of a bound nucleon on its momentum and the emc effect,” A505, 267–299. Phys. Rev. C 76 (5), 055206. Benhar, O, V. R. Pandharipande, and I. Sick (1997), “Nu- Caballero, J A, T. W. Donnelly, E. Moya de Guerra, and clear binding and deep inelastic scattering,” Phys. Lett. J. M. Udias (1998), “Analysis of factorization in (e, e- B410, 79–85. prime p) reactions: A Survey of the relativistic plane Benhar, O, V. R. Pandharipande, and I. Sick (1999), “Density wave impulse approximation,” Nucl. Phys. A632, 323–362, dependence of the EMC effect,” Phys. Lett. B469, 19–24. arXiv:nucl-th/9710038 [nucl-th]. Benhar, Omar, and Ingo Sick (2012), “Nuclear binding, cor- Carlson, C E, and T. J. Havens (1983), “Quark Distributions relations and the origin of EMC effect,” arXiv:1207.4595 in Nuclei,” Phys. Rev. Lett. 51, 261. [nucl-th]. Carlson, J, S. Gandolfi, F. Pederiva, Steven C. Pieper, Berger, Edmond L, and F. Coester (1987), “Nuclear Effects R. Schiavilla, K. E. Schmidt, and R. B. Wiringa (2015), in Deep Inelastic Lepton Scattering,” Ann. Rev. Nucl. Part. “Quantum Monte Carlo methods for nuclear physics,” Rev. Sci. 37, 463–491. Mod. Phys. 87, 1067, arXiv:1412.3081 [nucl-th]. Bertozzi, W, J. Friar, J. Heisenberg, and J. W. Negele (1972), Carlson, J, J. Jourdan, R. Schiavilla, and I. Sick (2002), “Contributions of neutrons to elastic electron scattering “Longitudinal and transverse quasielastic response func- from nuclei,” Phys. Lett. B41, 408–414. tions of light nuclei,” Phys. Rev. C65, 024002, arXiv:nucl- Bertsch, G F, L. Frankfurt, and M. Strikman (1993), “Where th/0106047 [nucl-th]. are the nuclear pions?” Science 259, 773–774. Chen, Jiunn-Wei, and William Detmold (2005), “Universality Bethe, HA (1971), “Theory of nuclear matter,” of the EMC effect,” Phys. Lett. B625, 165–170, arXiv:hep- Ann.Rev.Nucl.Part.Sci. 21, 93–244. ph/0412119 [hep-ph]. Bickerstaff, R P, M. C. Birse, and G. A. Miller (1984), “DIS- Chen, Jiunn-Wei, William Detmold, Joel E. Lynn, and ENTANGLING EXPLANATIONS OF DEEP INELAS- Achim Schwenk (2016), “Short Range Correlations and the TIC LEPTON NUCLEUS SCATTERING BY LEPTON EMC Effect in Effective Field Theory,” arXiv:1607.03065 PAIR PRODUCTION,” Phys. Rev. Lett. 53, 2532–2535. [hep-ph]. Bickerstaff, R P, M. C. Birse, and G. A. Miller (1985), Choi, S, J.P. Chen, and Z.E. Meziani (2005), “Precision mea- “DYNAMICAL RESCALING AND THE SIZE OF THE surement of longitudinal and transverse response functions PION,” Phys. Lett. B161, 393–396. of quasi-elastic electron scattering in the momentum trans- Bickerstaff, R P, M. C. Birse, and G. A. Miller (1986), “PAR- fer range 0.55 gev/c ¡ q ¡ 0.9 gev/c, jefferson lab experiment TONS IN NUCLEI,” Phys. Rev. D33, 3228–3245. e05-110,”. Bickerstaff, R P, and G. A. Miller (1986), “Q2 Dependence Ciofi Degli Atti, Claudio, and S. Liuti (1989), “On the Ef- of the European Muon Collaboration Effect,” Phys. Rev. fects of Nucleon Binding and Correlations in Deep Inelastic D34, 2890–2892. Electron Scattering by Nuclei,” Phys. Lett. B225, 215–221. Bjorken, J D (1966), “Applications of the Chiral U(6) x (6) Cloet, I C, W. Bentz, and A. W. Thomas (2009a), “Isovector Algebra of Current Densities,” Phys. Rev. 148, 1467–1478. EMC effect explains the NuTeV anomaly,” Phys. Rev. Lett. Blunden, P G, M. Burkardt, and G. A. Miller (1999), “Light 102, 252301, arXiv:0901.3559 [nucl-th]. front nuclear physics: Mean field theory for finite nuclei,” Cloet, I C, W. Bentz, and A. W. Thomas (2009b), “Isovector Phys. Rev. C60, 055211, arXiv:nucl-th/9906012 [nucl-th]. EMC effect explains the NuTeV anomaly,” Phys. Rev. Lett. Blunden, Peter G, and Gerald A. Miller (1996), “Quark - 102, 252301, arXiv:0901.3559 [nucl-th]. meson coupling model for finite nuclei,” Phys. Rev. C54, Cloet, I C, W. Bentz, and A. W. Thomas (2012), “Parity- 359–370, arXiv:nucl-th/9602031 [nucl-th]. violating DIS and the flavour dependence of the EMC ef- Bogner, S K, R. J. Furnstahl, and A. Schwenk (2010), “From fect,” Phys. Rev. Lett. 109, 182301, arXiv:1202.6401 [nucl- low-momentum interactions to nuclear structure,” Prog. th]. Part. Nucl. Phys. 65, 94–147, arXiv:0912.3688 [nucl-th]. Cloet, I C, Wolfgang Bentz, and Anthony William Thomas Bogner, S K, A. Schwenk, R. J. Furnstahl, and (2005a), “Nucleon quark distributions in a covariant quark- A. Nogga (2005), “Is nuclear matter perturbative with diquark model,” Phys. Lett. B621, 246–252, arXiv:hep- low-momentum interactions?” Nucl. Phys. A763, 59–79, ph/0504229 [hep-ph]. arXiv:nucl-th/0504043 [nucl-th]. Cloet, I C, Wolfgang Bentz, and Anthony William Thomas Brockmann, R, and R. Machleidt (1984), “NUCLEAR (2005b), “Spin-dependent structure functions in nuclear SATURATION IN A RELATIVISTIC BRUCKNER- matter and the polarized EMC effect,” Phys. Rev. Lett. HARTREE-FOCK APPROACH,” Phys. Lett. B149, 283. 95, 052302, arXiv:nucl-th/0504019 [nucl-th]. Brown, G E (1967), Unified Theory of Nuclear Models and Cloet, I C, Wolfgang Bentz, and Anthony William Thomas Forces (North-Holland Publishing Company, Amsterdam). (2006), “EMC and polarized EMC effects in nuclei,” Phys. Brown, G E, and A. D. Jackson (1976), The Nucleon-Nucleon Lett. B642, 210–217, arXiv:nucl-th/0605061 [nucl-th]. Interaction (North-Holland Publishing Company, Amster- Cloet, I C, Gerald A. Miller, E. Piasetzky, and G. Ron dam). (2009c), “Neutron Properties in the Medium,” Phys. Rev. Burkardt, M, M. R. Frank, and K. L. Mitchell (1997), “On the Lett. 103, 082301, arXiv:0903.1312 [nucl-th]. calculation of hadron form-factors from Euclidean Dyson- Cloet, Ian C, Wolfgang Bentz, and Anthony W. Thomas Schwinger equations,” Phys. Rev. Lett. 78, 3059–3062, (2016), “Relativistic and Nuclear Medium Effects on the arXiv:hep-ph/9611256 [hep-ph]. Coulomb Sum Rule,” Phys. Rev. Lett. 116 (3), 032701, 51

arXiv:1506.05875 [nucl-th]. Joey Huston, Pavel Nadolsky, Jon Pumplin, Carl Schmidt, Cloet, Ian C, and Craig D. Roberts (2014), “Explanation and Daniel Stump, and C. P. Yuan (2016), “New parton distri- Prediction of Observables using Continuum Strong QCD,” bution functions from a global analysis of quantum chromo- Prog. Part. Nucl. Phys. 77, 1–69, arXiv:1310.2651 [nucl-th]. dynamics,” Phys. Rev. D93 (3), 033006, arXiv:1506.07443 Close, F E (1979), An Introduction to Quarks and Partons [hep-ph]. (Academic Press, London). Dutta, D, K. Hafidi, and M. Strikman (2013), “Color Trans- Close, F E, R. G. Roberts, and Graham G. Ross (1983), “The parency: past, present and future,” Prog. Part. Nucl. Phys. Effect of Confinement Size on Nuclear Structure Func- 69, 1–27, arXiv:1211.2826 [nucl-th]. tions,” Phys. Lett. B129, 346–350. Egiyan, K, et al. (CLAS Collaboration) (2003), “Observation 0 Colle, C, et al. (2015), “Extracting the mass dependence of nuclear scaling in the a(e, e ) reaction at xb greater than and quantum numbers of short-range correlated pairs from 1,” Phys. Rev. C 68, 014313. 0 0 a(e, e p) and a(e, e pp) scattering,” Phys. Rev. C 92, Egiyan, K, et al. (CLAS Collaboration) (2006a), “Measure- 024604. ment of 2- and 3-nucleon short range correlation probabil- Colle, Camille, Wim Cosyn, Jan Ryckebusch, and Maarten ities in nuclei,” Phys. Rev. Lett. 96, 082501. Vanhalst (2014), “Factorization of exclusive electron- Egiyan, K, et al. (CLAS Collaboration) (2006b), “Measure- induced two-nucleon knockout,” Phys. Rev. C89 (2), ment of 2- and 3-nucleon short range correlation probabil- 024603, arXiv:1311.1980 [nucl-th]. ities in nuclei,” Phys. Rev. Lett. 96, 082501. Collins, John (2013), Foundations of perturbative QCD (Cam- Epelbaum, Evgeny, Hans-Werner Hammer, and Ulf-G. Meiss- bridge University Press). ner (2009), “Modern Theory of Nuclear Forces,” Rev. Mod. Cosyn, Wim, Wally Melnitchouk, and Misak Sargsian (2014), Phys. 81, 1773–1825, arXiv:0811.1338 [nucl-th]. “Final-state interactions in inclusive deep-inelastic scat- Ericson, Magda, and Anthony William Thomas (1983), “Pi- tering from the deuteron,” Phys. Rev. C89 (1), 014612, onic Corrections and the EMC Enhancement of the Sea in arXiv:1311.3550 [nucl-th]. Iron,” Phys. Lett. B128, 112–116. Cosyn, Wim, and Misak Sargsian (2011), “Final-state in- Ericson, Magda, and Anthony William Thomas (1984), “Ev- teractions in semi-inclusive deep inelastic scattering off the idence for an Enhanced Nuclear Sea From the Proton - Deuteron,” Phys. Rev. C84, 014601, arXiv:1012.0293 [nucl- Nucleus Drell-Yan Process,” Phys. Lett. B148, 191–193. th]. Fanchiotti, Huner, Carlos A. Garc´ıaCanal, Tatiana Tarutina, Cosyn, Wim, and Misak M. Sargsian (2016), “High x Struc- and Vicente Vento (2014), “Medium effects in dis from po- ture Function of the Virtually Free Neutron,” Phys. Rev. larized nuclear targets,” The European Physical Journal A C93 (5), 055205, arXiv:1506.01067 [hep-ph]. 50 (7), 1–5. Day, D B, J. S. McCarthy, T. W. Donnelly, and I. Sick (1990), Farrar, Glennys R, and Darrell R. Jackson (1975), “Pion and Annu. Rev. Nucl. Part. Sci. 40, 357. nucleon structure functions near x = 1,” Phys. Rev. Lett. Day, D B, J. S. McCarthy, Z. E. Meziani, R. Minehart, 35, 1416–1419. R. Sealock, S. T. Thornton, J. Jourdan, I. Sick, B. W. Fil- Farrar, Glennys R, Huan Liu, Leonid L. Frankfurt, and ippone, R. D. McKeown, R. G. Milner, D. H. Potterveld, Mark I. Strikman (1988), “Transparency in nuclear and Z. Szalata (1987a), “y,” Phys. Rev. Lett. 59, 427–430. quasiexclusive processes with large momentum transfer,” Day, D B, et al. (1987b), “y scaling in electron-nucleus scat- Phys. Rev. Lett. 61, 686–689. tering,” Phys. Rev. Lett. 59 (4), 427. Feldmeier, H, W. Horiuchi, T. Neff, and Y. Suzuki (2011), De Forest, T (1983a), Nucl. Phys. A392, 232. “Universality of short-range nucleon-nucleon correlations,” De Forest, T (1983b), “Off-Shell electron Nucleon Cross- Phys. Rev. C84, 054003, arXiv:1107.4956 [nucl-th]. Sections. The Impulse Approximation,” Nucl. Phys. A392, Fields, L, et al. (MINERvA Collaboration) (2013), “Measure- 232–248. ment of muon antineutrino quasielastic scattering on a hy- De Forest, Jr, T, and J. D. Walecka (1966), “Electron scat- drocarbon target at Eν ∼ 3.5 GeV,” Phys. Rev. Lett. 111, tering and nuclear structure,” Adv. Phys. 15, 1–109. 022501. Detmold, W, G. A. Miller, and Jason Robert Smith (2006), Fiorentini, G A, et al. (MINERvA Collaboration) (2013), “Role of the nuclear vector potential in deep inelastic scat- Phys. Rev. Lett. 111, 022501. tering,” Phys. Rev. C73, 015204, arXiv:nucl-th/0509033 Fissum, KG, et al. (Jefferson Lab Hall A Collaboration) [nucl-th]. (2004), “The Dynamics of the quasielastic O-16(e, e-prime Diakonov, Dmitri, V. Petrov, P. Pobylitsa, Maxim V. p) reaction at Q**2 = 0.8 (GeV/c)**2,” Phys.Rev. C70, Polyakov, and C. Weiss (1996), “Nucleon parton distri- 034606, arXiv:nucl-ex/0401021 [nucl-ex]. butions at low normalization point in the large N(c) limit,” Fomin, N, et al. (2012a), “New measurements of high- Nucl. Phys. B480, 341–380, arXiv:hep-ph/9606314 [hep- momentum nucleons and short-range structures in nuclei,” ph]. Phys. Rev. Lett. 108, 092502. Dieperink, A E L, and G. A. Miller (1991), “Nucleon binding Fomin, N, et al. (2012b), “New measurements of high- corrections to lepton nucleus deep inelastic scattering: Use momentum nucleons and short-range structures in nuclei,” of a realistic spectral function,” Phys. Rev. C44, 866–869. Phys. Rev. Lett. 108, 092502. Dieterich, S, et al. (2001), “Polarization transfer in the Frank, M R, B. K. Jennings, and G. A. Miller (1996), “The He-4(polarized-e, e-prime polarized-p)H-3 reaction,” Phys. Role of color neutrality in nuclear physics: Modifications Lett. B500, 47–52, arXiv:nucl-ex/0011008 [nucl-ex]. of nucleonic wave functions,” Phys. Rev. C54, 920–935, Do Dang, G, M. L’Huillier, Nguyen Van Giai, and J. Wallace arXiv:nucl-th/9509030 [nucl-th]. Van Orden (1987), “Coulomb sum rules in the relativistic Frankfurt, L, V. Guzey, and M. Strikman (2012), “Leading Fermi gas model,” Phys. Rev. C35, 1637–1645. Twist Nuclear Shadowing Phenomena in Hard Processes Dulat, Sayipjamal, Tie-Jiun Hou, Jun Gao, Marco Guzzi, with Nuclei,” Phys. Rept. 512, 255–393, arXiv:1106.2091 [hep-ph]. 52

Frankfurt, L, M. Strikman, and M. Zhalov (2001), “Single arXiv:1406.0454 [nucl-th]. particle strength restoration and nuclear transparency in Golak, J, H. Kamada, H. Witala, W. Glockle, and S. Ishikawa high Q**2 exclusive (e, e-prime p) reactions,” Phys. Lett. (1995), “Electron induced pd and ppn breakup of 3he B503, 73–80, arXiv:hep-ph/0011088 [hep-ph]. with full inclusion of final-state interactions,” Phys. Rev. Frankfurt, L L, G. A. Miller, and M. Strikman (1994), C 51 (4), 1638–1647. “The Geometrical color optics of coherent high-energy pro- Gomes, L C, J. D. Walecka, and V. F. Weisskopf (1958), cesses,” Ann. Rev. Nucl. Part. Sci. 44, 501–560, arXiv:hep- “Properties of nuclear matter,” Annals Phys. 3, 241–274. ph/9407274 [hep-ph]. Gomez, J, et al. (1994), “Measurement of the a dependence of Frankfurt, L L, M. M. Sargsian, and M. I. Strikman (1997), deep-inelastic electron scattering,” Phys. Rev. D 49, 4348. “Feynman graphs and Gribov-Glauber approach to high- Gribov, V N, B. L. Ioffe, and I. Ya. Pomeranchuk (1966), energy knockout processes,” Phys. Rev. C56, 1124–1137, “What is the range of interactions at high-energies,” Sov. arXiv:nucl-th/9603018 [nucl-th]. J. Nucl. Phys. 2, 549, [Yad. Fiz.2,768(1965)]. Frankfurt, L L, and M. I. Strikman (1981a), “High-energy Griffioen, K A, et al. (2015), “Measurement of the EMC phenomena, short-range nuclear structure and qcd,” Phys. Effect in the Deuteron,” Phys. Rev. C92 (1), 015211, Rep. 76 (4), 215. arXiv:1506.00871 [hep-ph]. Frankfurt, L L, and M. I. Strikman (1981b), “High-Energy Gross, Franz, and D. O. Riska (1987), “Current Conservation Phenomena, Short Range Nuclear Structure and QCD,” and Interaction Currents in Relativistic Meson Theories,” Phys. Rept. 76, 215–347. Phys. Rev. C36, 1928. Frankfurt, L L, and M. I. Strikman (1985), “Point- Guichon, Pierre A M (1988), “A Possible Quark Mechanism like CONFIGURATIONS IN HADRONS AND NUCLEI for the Saturation of Nuclear Matter,” Phys. Lett. B200, AND DEEP INELASTIC REACTIONS WITH LEP- 235–240. TONS: EMC AND EMC LIKE EFFECTS,” Nucl. Phys. Guichon, Pierre A M, and Gerald A. Miller (1984), “Quarks B250, 143–176. and the Deuteron Asymptotic D State,” Phys. Lett. B134, Frankfurt, L L, and M. I. Strikman (1987), “On the Normal- 15–20. ization of Nucleus Spectral Function and the EMC Effect,” Guichon, Pierre A M, Koichi Saito, Evgenii N. Rodionov, Phys. Lett. B183, 254. and Anthony William Thomas (1996), “The Role of nu- Frankfurt, L L, and M. I. Strikman (1988a), “Hard Nuclear cleon structure in finite nuclei,” Nucl. Phys. A601, 349– Processes and Microscopic Nuclear Structure,” Phys. Rept. 379, arXiv:nucl-th/9509034 [nucl-th]. 160, 235–427. Haftel, Michael I, and Frank Tabakin (1970), “NU- Frankfurt, Leonid, Misak Sargsian, and Mark Strikman CLEAR SATURATION AND THE SMOOTHNESS (2008), “Recent observation of short range nucleon corre- OF NUCLEON-NUCLEON POTENTIALS,” Nucl. Phys. lations in nuclei and their implications for the structure of A158, 1–42. nuclei and neutron stars,” Int. J. Mod. Phys. A23, 2991– Hagen, G, et al. (2015), “Neutron and weak-charge distribu- 3055, arXiv:0806.4412 [nucl-th]. tions of the 48Ca nucleus,” Nature Phys. 12 (2), 186–190, Frankfurt, Leonid, and Mark Strikman (1988b), “Hard nu- arXiv:1509.07169 [nucl-th]. clear processes and microscopic nuclear structure,” Phys. Halzen, F, and Alan D. Martin (1984), QUARKS AND LEP- Rep. 160 (5-6), 235 – 427. TONS: AN INTRODUCTORY COURSE IN MODERN Frankfurt, Leonid, and Mark Strikman (2012), “QCD and PARTICLE PHYSICS (Wiley, New York). QED dynamics in the EMC effect,” Int. J. Mod. Phys. E21, Haxel, Otto, J. Hans D. Jensen, and Hans E. Suess (1949), 1230002, arXiv:1203.5278 [hep-ph]. “On the ”magic numbers” in nuclear structure,” Phys. Rev. Frankfurt, LL, M.I. Strikman, D.B. Day, and M. Sargsyan 75, 1766–1766. (1993), “Evidence for short-range correlations from high q2 Haxton, W C, B. F. Gibson, and E. M. Henley (1980), “PAR- (e,e’) reactions,” Phys. Rev. C 48, 2451. ITY NONCONSERVATION IN F-18, F-19, AND NE-21,” Freese, Adam J, Misak M. Sargsian, and Mark I. Strikman Phys. Rev. Lett. 45, 1677–1681. (2015), “Probing superfast quarks in nuclei through dijet Hen, O, A. Accardi, W. Melnitchouk, and E. Piaset- production at the LHC,” Eur. Phys. J. C75 (11), 534, zky (2011a), “Constraints on the large-x d/u ratio from arXiv:1411.6605 [hep-ph]. electron-nucleus scattering at x > 1,” Phys. Rev. D 84, Frullani, S, and J. Mougey (1984), “Single particle properties 117501. of nuclei through (e, e0p) reactions,” Adv. Nucl. Phys. 14, Hen, O, E. Piasetzky, and L. B. Weinstein (2012), “New data 1. strengthen the connection between short range correlations Gao, J, et al. (2000), “Dynamical relativistic effects in and the emc effect,” Phys. Rev. C 85, 047301. quasielastic 1p-shell proton knockout from 16o,” Phys. Rev. Hen, O, A. W. Steiner, E. Piasetzky, and L. B. Weinstein Lett. 84, 3265. (2016a), “Analysis of Neutron Stars Observations Using a Gayou, O, et al. (Jefferson Lab Hall A) (2002), “Measure- Correlated Fermi Gas Model,” arXiv:1608.00487 [nucl-ex]. ment of G(Ep) / G(Mp) in polarized-e p —¿ e polarized- Hen, O, L. B. Weinstein, S. Gilad, and W. Boeglin (2014a), p to Q**2 = 5.6-GeV**2,” Phys. Rev. Lett. 88, 092301, “Proton and Neutron Momentum Distributions in A = arXiv:nucl-ex/0111010 [nucl-ex]. 3 Asymmetric Nuclei, Jefferson Lab Experiment E12-14- Geesaman, DF, K. Saito, and A.W. Thomas (1995), “The 011,” arXiv:1410.4451 [nucl-ex]. Nuclear EMC Effect,” Ann. Rev. Nucl. and Part. Sci. 45, Hen, O, L. B. Weinstein, E. Piasetzky, G. A. Miller, M. M. 337. Sargsian, and Y. Sagi (2015a), “Correlated fermions in nu- Gezerlis, A, I. Tews, E. Epelbaum, M. Freunek, S. Gan- clei and ultracold atomic gases,” Phys. Rev. C 92, 045205. dolfi, K. Hebeler, A. Nogga, and A. Schwenk (2014), “Lo- Hen, O, L. B. Weinstein, E. Piasetzky, G. A. Miller, M. M. cal chiral effective field theory interactions and quantum Sargsian, and Y. Sagi (2015b), “Correlated fermions in Monte Carlo applications,” Phys. Rev. C90 (5), 054323, nuclei and ultracold atomic gases,” Phys. Rev. C92 (4), 53

045205, arXiv:1407.8175 [nucl-ex]. BUTION TO LEPTON NUCLEUS DEEP INELASTIC Hen, O, L.B. Weinstein, E. Cohen, and D.W. Higinbotham SCATTERING,” Phys. Lett. B200, 351–356. (2016b), “The CaFe Experiment: Short-Range Pairing Jung, H, and G. A. Miller (1990), “Pionic contributions Mechanisms in Heavy Nuclei, Jefferson Lab Proposal E12- to deep inelastic nuclear structure functions,” Phys. Rev. 16-004,”. C41, 659–664. Hen, O, L.B. Weinstein, E.I. Piasetzky, and H. Hakobyan Kelly, J J, and Stephen J. Wallace (1994), “Comparison be- (2014b), “In Medium Proton Structure Functions, SRC, tween relativistic and nonrelativistic models of the nucleon- and the EMC effect, Jefferson Lab experiment E12-11- nucleon effective interaction. 1: Normal parity isoscalar 003A,”. transitions,” Phys. Rev. C49, 1315–1336. Hen, O, L.B. Weinstein, S.A. Wood, and S. Gilad (2011b), Kelly, James J (1999), “Effects of spinor distortion and → 0 → “In Medium Nucleon Structure Functions, SRC, and the density-dependent form factors upon quasifree 16O( e , e p EMC effect, Jefferson Lab experiment E12-11-107,”. ),”. Hen, O, et al. (2013a), Int. J. Mod. Phys. E 22, 133017. Kelly, JJ (1996), “Nucleon knockout by intermediate energy Hen, O, et al. (CLAS Collaboration) (2014c), Science 346, electrons,” Adv. Nucl. Phys. 23, 75. 614. Kim, K S, B. G. Yu, and M. K. Cheoun (2006), “Coulomb Hen, O, et al. (2014d), “Momentum sharing in imbalanced sum rule in quasielastic region,” Phys. Rev. C74, 067601, Fermi systems,” Science 346, 614–617, arXiv:1412.0138 arXiv:nucl-th/0611067 [nucl-th]. [nucl-ex]. Klimenko, A V, et al. (CLAS Collaboration) (2006), “Electron Hen, Or, D. W. Higinbotham, Gerald A. Miller, Eli Piasetzky, scattering from high-momentum neutrons in deuterium,” and Lawrence B. Weinstein (2013b), “The EMC Effect and Phys. Rev. C 73 (3), 035212. High Momentum Nucleons in Nuclei,” Int. J. Mod. Phys. Koch, V, and G. A. Miller (1985), “SIX QUARK CLUSTER E22, 1330017, arXiv:1304.2813 [nucl-th]. EFFECTS AND BINDING ENERGY DIFFERENCES Hen, Or, Bao-An Li, Wen-Jun Guo, L. B. Weinstein, and Eli BETWEEN MIRROR NUCLEI,” Phys. Rev. C31, 602– Piasetzky (2015c), “Symmetry energy of nucleonic matter 612, [Erratum: Phys. Rev.C32,1106(1985)]. with tensor correlations,” Phys. Rev. C 91, 025803. Konrad, P, H. Lenske, and U. Mosel (2005), “Short range Hobbs, T J, Mary Alberg, and Gerald A. Miller (2016), correlations and spectral functions in asymmetric nuclear “A Euclidean bridge to the relativistic constituent quark matter,” Nuclear Physics A 756 (1-2), 192–212, cited By model,” arXiv:1608.07319 [nucl-th]. 8. Holt, Jeremy W, Norbert Kaiser, and Wolfram Weise (2013), Korover, I, N. Muangma, O. Hen, et al. (2014), “Probing the “Nuclear chiral dynamics and thermodynamics,” Prog. Repulsive Core of the Nucleon-Nucleon Interaction via the Part. Nucl. Phys. 73, 35–83, arXiv:1304.6350 [nucl-th]. 4He(e,e’pN) Triple-Coincidence Reaction,” Phys.Rev.Lett. Holt, Roy J (2013), “Large x physics: recent results and future 113, 022501, arXiv:1401.6138 [nucl-ex]. plans,” in Proceedings, 43rd International Symposium on Kramer, GJ, H.P. Blok, and L. Lapikas (2001), “A consis- Multiparticle Dynamics (ISMD 13), arXiv:1311.1527 [nucl- tent analysis of (e,ep) and (d,3he) experiments,” Nuclear ex]. Physics A 679 (3), 267 – 286. Holt, Roy J, and Craig D. Roberts (2010), “Nucleon and pion Kuhn, SE, and W.K. Brooks (2014), “Jefferson lab exper- distribution functions in the valence region,” Rev. Mod. iment e12-14-001: The emc effect in spin structure func- Phys. 82, 2991–3044. tions,”. Horikawa, T, and W. Bentz (2005), “Medium modifications of Kulagin, S A, and R. Petti (2010), “Structure functions for nucleon electromagnetic form-factors,” Nucl. Phys. A762, light nuclei,” Phys. Rev. C 82, 054614. 102–128, arXiv:nucl-th/0506021 [nucl-th]. Kulagin, S A, and R. Petti (2014), “Nuclear parton distri- Horowitz, C J (1985), “Relativistic Love-Franey model: Co- butions and the Drell-Yan process,” Phys. Rev. C90 (4), variant representation of the NN interaction for N-nucleus 045204, arXiv:1405.2529 [hep-ph]. scattering,” Phys. Rev. C31, 1340–1348. Kulagin, SA, and R. Petti (2006a), “Global study of nuclear Hugenholtz, N M, and L. van Hove (1958), “A theorem on structure functions,” Nuclear Physics A 765 (1), 126 – 187. the single particle energy in a Fermi gas with interaction,” Kulagin, Sergey A, and R. Petti (2006b), “Global study of Physica 24, 363–376. nuclear structure functions,” Nucl. Phys. A765, 126–187, Ioffe, B L (1969), “Space-time picture of photon and neu- arXiv:hep-ph/0412425 [hep-ph]. trino scattering and electroproduction cross-section asymp- Lacombe, M, B. Loiseau, R.Vinh Mau, J. Ct, P. Pirs, and totics,” Phys. Lett. B30, 123–125. R. de Tourreil (1981), “Parametrization of the deuteron Jaffe, R L (1983), “Quark Distributions in Nuclei,” Phys. Rev. wave function of the paris nn potential,” Physics Letters B Lett. 50, 228. 101 (3), 139 – 140. Ji, Xiangdong (2013), “Parton Physics on a Euclidean Lat- Lapikas, L (1993), “Quasi-elastic electron scattering off nu- tice,” Phys. Rev. Lett. 110, 262002, arXiv:1305.1539 [hep- clei,” Nuclear Physics A 553, 297 – 308. ph]. Lapikas, Louk, G. van der Steenhoven, L. Frankfurt, M. Strik- Jones, M K, et al. (Jefferson Lab Hall A) (2000), “G(E(p)) / man, and M. Zhalov (2000), “The Transparency of C-12 for G(M(p)) ratio by polarization transfer in polarized e p —¿ protons,” Phys. Rev. C61, 064325, arXiv:nucl-ex/9905009 e polarized p,” Phys. Rev. Lett. 84, 1398–1402, arXiv:nucl- [nucl-ex]. ex/9910005 [nucl-ex]. Lehr, J, M. Effenberger, H. Lenske, S. Leupold, and U. Mosel Jourdan, J (1995), “Longitudinal response functions: The (2000), “Transport theoretical approach to the nucleon Coulomb sum revisited,” Phys. Lett. B353, 189–195. spectral function in nuclear matter,” Physics Letters B Jourdan, J (1996), “Quasielastic response functions: The 483 (13), 324 – 330. Coulomb sum revisited,” Nucl. Phys. A603, 117–160. Lehr, J, H. Lenske, S. Leupold, and U. Mosel (2002), “Nu- Jung, H, and G. A. Miller (1988), “NUCLEONIC CONTRI- 54

clear matter spectral functions by transport theory,” Nucl. Melnitchouk, W, Andreas W. Schreiber, and An- Phys. A 703, 393. thony William Thomas (1994b), “Relativistic deuteron Lin, Huey-Wen, Jiunn-Wei Chen, Saul D. Cohen, and Xiang- structure function,” Phys. Lett. B335, 11–16, arXiv:nucl- dong Ji (2015), “Flavor Structure of the Nucleon Sea from th/9407007 [nucl-th]. Lattice QCD,” Phys. Rev. D91, 054510, arXiv:1402.1462 Melnitchouk, W, and A. W. Thomas (1996a), Phys. Lett. B [hep-ph]. 377, 11. Liyanage, N, et al. (The Jefferson Lab Hall A Collaboration) Melnitchouk, W, and A.W. Thomas (1996b), “Neutron pro- 0 (2001), “Dynamics of the 16o(e, e p) reaction at high miss- ton structure function ratio at large x,” Physics Letters B ing energies,” Phys. Rev. Lett. 86, 5670–5674. 377 (1), 11 – 17. Llewellyn Smith, C H (1983), “A Possible Explanation of the Meziani, Z E, et al. (1984), “Coulomb Sum Rule for Ca-40, Difference Between the Structure Functions of Iron and Ca-48, and Fe-56 for —q (Vector)— ¡= 550-MeV/c,” Phys. Deuterium,” Phys. Lett. B128, 107–111. Rev. Lett. 52, 2130–2133. Lovato, A, S. Gandolfi, Ralph Butler, J. Carlson, Ewing Mihaila, Bogdan, and Jochen Heisenberg (2000), “Micro- Lusk, Steven C. Pieper, and R. Schiavilla (2013), “Charge scopic calculation of the inclusive electron scattering struc- Form Factor and Sum Rules of Electromagnetic Response ture function in O-16,” Phys. Rev. Lett. 84, 1403–1406, Functions in 12C,” Phys. Rev. Lett. 111 (9), 092501, arXiv:nucl-th/9910007 [nucl-th]. arXiv:1305.6959 [nucl-th]. Mihovilovic, M, et al. (Jefferson Lab Hall A) (2014), Lovato, A, S. Gandolfi, J. Carlson, Steven C. Pieper, and “Measurement of double-polarization asymmetries in the 3 ~ 0 R. Schiavilla (2016), “Electromagnetic response of 12C: a quasielastic He(~e, e d) process,” Phys. Rev. Lett. 113 (23), first-principles calculation,” arXiv:1605.00248 [nucl-th]. 232505, arXiv:1409.2253 [nucl-ex]. Lu, Ding-Hui, Kazuo Tsushima, Anthony William Thomas, Miller, G A (1984a), “SIX QUARK CLUSTER COMPO- Anthony Gordon Williams, and K. Saito (1999), “Electro- NENTS OF NUCLEAR WAVE FUNCTIONS AND THE magnetic form-factors of the bound nucleon,” Phys. Rev. PION NUCLEUS DOUBLE CHARGE EXCHANGE RE- C60, 068201, arXiv:nucl-th/9807074 [nucl-th]. ACTION,” Phys. Rev. Lett. 53, 2008–2011. Macfarlane, M H, and J. B. French (1960), “Stripping reac- Miller, G A, and R. Machleidt (1999a), “Infinite nuclear mat- tions and the structure of light and intermediate nuclei,” ter on the light front: Nucleon-nucleon correlations,” Phys. Rev. Mod. Phys. 32, 567–691. Rev. C60, 035202, arXiv:nucl-th/9903080 [nucl-th]. Machleidt, R (1989), “The Meson theory of nuclear forces and Miller, G A, and R. Machleidt (1999b), “Light front theory nuclear structure,” Adv. Nucl. Phys. 19, 189–376. of nuclear matter,” Phys. Lett. B455, 19–24, arXiv:nucl- Machleidt, R, and D.R. Entem (2011), “Chiral effective field th/9811050 [nucl-th]. theory and nuclear forces,” Physics Reports 503 (1), 1 – Miller, Gerald A (1984b), “Building the Nucleus From 75. Quarks: The Cloudy Bag Model and the Quark Description Machleidt, R, K. Holinde, and C. Elster (1987), “The Bonn of the Nucleon-nucleon Wave Function,” Int. Rev. Nucl. Meson Exchange Model for the Nucleon Nucleon Interac- Phys. 1, 189–323. tion,” Phys. Rept. 149, 1–89. Miller, Gerald A (1988), “NUCLEAR WAVE FUNCTIONS Makek, M, et al. (2016), “Differential cross section mea- IN DEEP INELASTIC SCATTERING AND DRELL- 12 10 surement of the C(e,epp) Beg.s. reaction,” Eur. Phys. J. YAN PROCESSES,” in Workshop on Nuclear and Particle A52 (9), 298. Physics on the Light Cone Los Alamos, New Mexico, July Malace, S P, et al. (2011), “A precise extraction of the induced 18-22, 1988, pp. 0042–64. polarization in the 4He(e,e’p)3H reaction,” Phys. Rev. Lett. Miller, Gerald A (2000), “Light front quantization: A Tech- 106, 052501, arXiv:1011.4483 [nucl-ex]. nique for relativistic and realistic nuclear physics,” Prog. Malace, Simona, David Gaskell, Douglas W. Higinbotham, Part. Nucl. Phys. 45, 83–155, arXiv:nucl-th/0002059 [nucl- and Ian C. Clot (2014), “The challenge of the emc effect: th]. Existing data and future directions,” International Journal Miller, Gerald A (2014), “Pionic and Hidden-Color, Six- of Modern Physics E 23 (08), 1430013. Quark Contributions to the Deuteron b1 Structure Func- Marco, E, E. Oset, and P. Fernandez de Cordoba (1996), tion,” Phys. Rev. C89 (4), 045203, arXiv:1311.4561 [nucl- “Mesonic and binding contributions to the EMC effect in a th]. relativistic many body approach,” Nucl. Phys. A611, 484– Miller, Gerald A, and Leonard S. Kisslinger (1983), “Quark + 513, arXiv:nucl-th/9510060 [nucl-th]. Contributions to the pp ↔ dπ Reaction,” Phys. Rev. C27, Mayer, M, et al. (CLAS) (2017), “Beam-target double- 1669. spin asymmetry in quasielastic electron scattering off Miller, Gerald A, Matthew D. Sievert, and Raju Venu- the deuteron with CLAS,” Phys. Rev. C95 (2), 024005, gopalan (2016), “Probing short-range nucleon-nucleon in- arXiv:1610.06109 [nucl-ex]. teractions with an Electron-Ion Collider,” Phys. Rev. Mayer, Maria Goeppert (1950), “Nuclear configurations in C93 (4), 045202, arXiv:1512.03111 [nucl-th]. the spin-orbit coupling model. i. empirical evidence,” Phys. Miller, Gerald A, and Jason Robert Smith (2002), “Return of Rev. 78, 16–21. the EMC effect,” Phys. Rev. C65, 015211, [Erratum: Phys. McVoy, K W, and L. Van Hove (1962), “Inelastic Electron- Rev.C66,049903(2002)], arXiv:nucl-th/0107026 [nucl-th]. Nucleus Scattering and Nucleon-Nucleon Correlations,” Miller, Gerald A, and James E. Spencer (1976), “A Survey Phys. Rev. 125, 1034–1043. of Pion Charge-Exchange Reactions with Nuclei,” Annals Melnitchouk, W, Andreas W. Schreiber, and An- Phys. 100, 562. thony William Thomas (1994a), “Deep inelastic scatter- Mineo, H, Wolfgang Bentz, N. Ishii, Anthony William ing from off-shell nucleons,” Phys. Rev. D49, 1183–1198, Thomas, and K. Yazaki (2004), “Quark distributions in arXiv:nucl-th/9311008 [nucl-th]. nuclear matter and the EMC effect,” Nucl. Phys. A735, 482–514, arXiv:nucl-th/0312097 [nucl-th]. 55

Monaghan, P, et al. (2014), “Measurement of the (1994), “Energy weighted sum rules for spectral functions 12C(e, e0p)11B two-body breakup reaction at high missing in nuclear matter,” Phys. Rev. C49, 3050–3054. momentum,” J. Phys. G41, 105109, arXiv:1301.7027 [nucl- Punjabi, V, et al. (2005), “Proton elastic form-factor ratios to ex]. Q**2 = 3.5-GeV**2 by polarization transfer,” Phys. Rev. Morgenstern, J, and Z. E. Meziani (2001), “Is the Coulomb C71, 055202, [Erratum: Phys. Rev.C71,069902(2005)], sum rule violated in nuclei?” Phys. Lett. B515, 269–275, arXiv:nucl-ex/0501018 [nucl-ex]. arXiv:nucl-ex/0105016 [nucl-ex]. Riordan, S, R. Beminiwattha, and A. Arrington (2016), “”the Mulders, P J (1990), “Modifications of Nucleons in Nuclei emc pvdis experiment, a constraint on isovector-dependent and Other Consequences of the Quark Substructure,” Phys. nuclear modification effects using parity-violating deep in- Rept. 185, 83–169. elastic scattering, proposal pr12-16-006”,”. Murdock, D P, and C. J. Horowitz (1987), “Microscopic Rel- Rios, A, A. Polls, and W. H. Dickhoff (2014), “Density and ativistic Description of Proton - Nucleus Scattering,” Phys. isospin asymmetry dependence of high-momentum compo- Rev. C35, 1442–1462. nents,” Phys. Rev. C89 (4), 044303, arXiv:1312.7307 [nucl- Neff, Thomas, Hans Feldmeier, and Wataru Horiuchi (2015), th]. “Short-range correlations in nuclei with similarity renor- Roberts, Craig D, Roy J. Holt, and Sebastian M. Schmidt malization group transformations,” Phys. Rev. C92 (2), (2013), “Nucleon spin structure at very high-x,” Phys. Lett. 024003, arXiv:1506.02237 [nucl-th]. B727, 249–254, arXiv:1308.1236 [nucl-th]. Negele, John W (1999), “Instantons, the QCD vacuum, Roberts, R G (1994), The Structure of the proton: Deep in- and hadronic physics,” Lattice Field Theory. Proceedings: elastic scattering (Cambridge University Press). 16th International Symposium, Lattice ’98, Boulder, USA, Ryckebusch, J, D. Debruyne, P. Lava, S. Janssen, Jul 13-18, 1998, Nucl. Phys. Proc. Suppl. 73, 92–104, B. Van Overmeire, and T. Van Cauteren (2003), “Rel- arXiv:hep-lat/9810053 [hep-lat]. ativistic formulation of Glauber theory for A(e, e-prime Noble, J V (1981), “Modification of the nucleon’s properties p) reactions,” Nucl. Phys. A728, 226–250, arXiv:nucl- in nuclear matter,” Phys. Rev. Lett. 46, 412–415. th/0305066 [nucl-th]. Norton, P R (2003), Rep. Prog. Phys. 66, 1253. Ryckebusch, Jan, Maarten Vanhalst, and Wim Cosyn (2015), Olive, K A, et al. (Particle Data Group) (2014), “Review of “Stylized features of single-nucleon momentum distribu- Particle Physics,” Chin. Phys. C38, 090001. tions,” Journal of Physics G: Nuclear and Particle Physics Owens, J F, A. Accardi, and W. Melnitchouk (2013), “Global 42 (5), 055104. parton distributions with nuclear and finite-Q2 correc- Saito, K, Kazuo Tsushima, and Anthony William Thomas tions,” Phys. Rev. D 87, 094012. (1999), “Effect of nucleon structure variation on the lon- Palli, V, C. Ciofi degli Atti, L. P. Kaptari, C. B. Mezzetti, gitudinal response function,” Phys. Lett. B465, 27–35, and M. Alvioli (2009), “Slow Proton Production in Semi- arXiv:nucl-th/9904055 [nucl-th]. Inclusive Deep Inelastic Scattering off Deuteron and Com- Sargsian, M M, T. V. Abrahamyan, M. I. Strikman, and plex Nuclei: Hadronization and Final State Interaction Ef- L. L. Frankfurt (2005a), “Exclusive electro-disintegration fects,” Phys. Rev. C80, 054610, arXiv:0911.1377 [nucl-th]. of He-3 at high Q2. II. Decay function formalism,” Phys. Paolone, M, et al. (2010), “Polarization Transfer in the Rev. C71, 044615, arXiv:nucl-th/0501018 [nucl-th]. 4He(e,e’p)3H Reaction at Q2 = 0.8 and 1.3 (GeV/c)2,” Sargsian, M M, T. V. Abrahamyan, M. I. Strikman, and L. L. Phys. Rev. Lett. 105, 072001, arXiv:1002.2188 [nucl-ex]. Frankfurt (2005b), “Exclusive electrodisintegration of he3 Passchier, I, et al. (2002), “Spin momentum correlations at high q2. ii. decay function formalism,” Phys. Rev. C in quasielastic electron scattering from deuterium,” Phys. 71 (4), 044615. Rev. Lett. 88, 102302, arXiv:nucl-ex/0109015 [nucl-ex]. Sargsian, Misak M (2014a), “New properties of the high- Perdrisat, C F, V. Punjabi, and M. Vanderhaeghen (2007), momentum distribution of nucleons in asymmetric nuclei,” “Nucleon Electromagnetic Form Factors,” Prog. Part. Phys. Rev. C 89, 034305. Nucl. Phys. 59, 694–764, arXiv:hep-ph/0612014 [hep-ph]. Sargsian, Misak M (2014b), “New properties of the high- Petratos, GG, J. Gomez, R.J. Holt, and R.D. Ransome momentum distribution of nucleons in asymmetric nuclei,” (2010), “MeAsurement of the F2n/F2p, d/u RAtios and Phys. Rev. C89 (3), 034305, arXiv:1210.3280 [nucl-th]. A=3 EMC Effect in Deep Inelastic Electron Scattering Off Sargsian, Misak M (2014c), “Protons in High Density Neutron the Tritium and Helium MirrOr Nuclei, Jefferson Lab Ex- Matter,” Proceedings, 2nd International Symposium on the periment E12-10-103,”. Modern Physics of Compact Stars and Relativistic Gravity: Piasetzky, E, M. Sargsian, L. Frankfurt, M. Strikman, and Yerevan, Armenia, September 18-21, 2013, J. Phys. Conf. J. W. Watson (2006), “Evidence for the strong dominance Ser. 496, 012007, arXiv:1312.2263 [nucl-th]. of proton-neutron correlations in nuclei,” Phys. Rev. Lett. Schiavilla, R, O. Benhar, A. Kievsky, L. E. Marcucci, and 97, 162504, arXiv:nucl-th/0604012 [nucl-th]. M. Viviani (2005), “Polarization transfer in He-4(polarized- Picklesimer, A, and J.W. Van Orden (1989), Phys. Rev. C e, e-prime polarized-p) H-3: Is the ratio G(Ep) / G(Mp) 40, 290. modified in medium?” Phys. Rev. Lett. 94, 072303, Picklesimer, A, J.W. Van Orden, and S.J. Wallace (1985), arXiv:nucl-th/0412020 [nucl-th]. Phys. Rev. C 32, 1312. Schiavilla, R, R. B. Wiringa, Steven C. Pieper, and J. Carlson Piller, Gunther, and Wolfram Weise (2000), “Nuclear deep in- (2007), “Tensor forces and the ground-state structure of elastic lepton scattering and coherence phenomena,” Phys. nuclei,” Physical Review Letters 98 (13), 132501. Rept. 330, 1–94, arXiv:hep-ph/9908230 [hep-ph]. Schlumpf, Felix (1992), Relativistic constituent quark Pirner, Hans J, and James P. Vary (1981), “Deep Inelas- model for baryons, Ph.D. thesis (Zurich U.), arXiv:hep- tic electron Scattering and the Quark Structure of He-3,” ph/9211255 [hep-ph]. Phys. Rev. Lett. 46, 1376–1379. Schlumpf, Felix (1993), “Relativistic constituent quark model Polls, A, A. Ramos, J. Ventura, S. Amari, and W. H. Dickhoff of electroweak properties of baryons,” Phys. Rev. D47, 56

4114, [Erratum: Phys. Rev.D49,6246(1994)], arXiv:hep- Structure of the Nucleon. ph/9212250 [hep-ph]. Tkachenko, S, et al. (CLAS Collaboration) (2014), “Measure- Seely, J, et al. (2009a), “New measurements of the european ment of the structure function of the nearly free neutron 2 0 muon collaboration effect in very light nuclei,” Phys. Rev. using spectator tagging in inelastic H(e,e ps)x scattering Lett. 103, 202301. with clas,” Phys. Rev. C 89, 045206. Seely, J, et al. (2009b), “New measurements of the european Udias, J M, J. A. Caballero, E. Moya de Guerra, J. E. Amaro, muon collaboration effect in very light nuclei,” Phys. Rev. and T. W. Donnelly (1999a), “Quasielastic scattering from Lett. 103, 202301. relativistic bound nucleons: Transverse-longitudinal re- Shneor, R, et al. (2007), “Investigation of proton-proton sponse,” Phys. Rev. Lett. 83, 5451–5454. 0 short-range correlations via the 12c(e, e pp) reaction,” Phys. Udias, J M, J. A. Caballero, E. Moya de Guerra, Jose Enrique Rev. Lett. 99 (7), 072501. Amaro, and T. W. Donnelly (1999b), “Quasielastic scatter- Sick, I (1985), “How Much Do Nucleons Change in Nuclei?” ing from relativistic bound nucleons: Transverse longitudi- Nucl. Phys. A434 ( 1985) 677C-684C, Nucl. Phys. A434, nal response,” Phys. Rev. Lett. 83, 5451–5454, arXiv:nucl- 677C–684C. th/9905030 [nucl-th]. Sloan, T, R. Voss, and G. Smadja (1988), “The Quark Struc- Udias, J M, and Javier R. Vignote (2000), “Relativistic nu- ture of the Nucleon from the CERN Muon Experiments,” clear structure effects in (e,e-prime polarized-p),” Phys. Phys. Rept. 162, 45–167. Rev. C62, 034302, arXiv:nucl-th/0007047 [nucl-th]. Smith, Jason Robert, and Gerald A. Miller (2002), “Return Udias, JM, et al. (1993), Phys. Rev. C 48, 2731. of the EMC effect: Finite nuclei,” Phys. Rev. C65, 055206, Udias, JM, et al. (1995), Phys. Rev. C 51, 3246. arXiv:nucl-th/0202016 [nucl-th]. Vanhalst, Maarten, Wim Cosyn, and Jan Ryckebusch (2011), Smith, Jason Robert, and Gerald A. Miller (2003), “Chi- “Counting the amount of correlated pairs in a nucleus,” ral solitons in nuclei: Saturation, EMC effect and Drell- Phys. Rev. C84, 031302, arXiv:1105.1038 [nucl-th]. Yan experiments,” Phys. Rev. Lett. 91, 212301, [Erratum: Vanhalst, Maarten, Jan Ryckebusch, and Wim Cosyn (2012), Phys. Rev. Lett.98,099902(2007)], arXiv:nucl-th/0308048 “Quantifying short-range correlations in nuclei,” Phys. [nucl-th]. Rev. C 86, 044619. Smith, Jason Robert, and Gerald A. Miller (2004), “Chi- Vary, J P, P. U. Sauer, and C. W. Wong (1973), “Conver- ral solitons in nuclei: Electromagnetic form-factors,” Phys. gence Rate of Intermediate-State Summations in the Effec- Rev. C70, 065205, arXiv:nucl-th/0407093 [nucl-th]. tive Shell-Model Interaction,” Phys. Rev. C7, 1776–1785. Smith, Jason Robert, and Gerald A. Miller (2005), “Polarized Walecka, J D (2005), Electron scattering for nuclear and nu- quark distributions in nuclear matter,” Phys. Rev. C72, cleon structure (Cambridge University Press). 022203, arXiv:nucl-th/0505048 [nucl-th]. Wallace, S J, and J. A. Tjon (2008), “Coulomb corrections Solvignon-Slifer, P, and J. Arrington (2011), “Precision mea- in quasi-elastic scattering: Tests of the effective-momentum surement of the isospin dependence in the 2N and 3N short approximation,” Phys. Rev. C78, 044604, arXiv:0808.2029 range correlation region, Jefferson Lab Experiment E12-11- [nucl-th]. 112,”. Weinstein, L B, O. Hen, and E. Piasetzky (2016), “Ham- Souder, P A (2016), “Parity Violation in Deep Inelastic Scat- mer events, neutrino energies, and nucleon-nucleon correla- tering with the SoLID Spectrometer at JLab,” Proceed- tions,” Phys. Rev. C94 (4), 045501, arXiv:1604.02482 [hep- ings, 21st International Symposium on Spin Physics (SPIN ex]. 2014): Beijing, China, October 20-24, 2014, Int. J. Mod. Weinstein, L B, E. Piasetzky, D. W. Higinbotham, J. Gomez, Phys. Conf. Ser. 40, 1660077. O. Hen, and R. Shneor (2011), “Short range correlations Stoks, V G J, R. A. M. Klomp, C. P. F. Terheggen, and and the emc effect,” Phys. Rev. Lett. 106 (5), 052301. J. J. de Swart (1994a), “Construction of high quality N N Weiss, R, R. Cruz-Torres, N. Barnea, E. Piasetzky, and potential models,” Phys. Rev. C49, 2950–2962, arXiv:nucl- O. Hen (2016), “The nuclear contacts and short range cor- th/9406039 [nucl-th]. relations in nuclei,” arXiv:1612.00923 [nucl-th]. Stoks, V G J, R. A. M. Klomp, C. P. F. Terheggen, and Weiss, Ronen, Betzalel Bazak, and Nir Barnea (2015), “Gen- J. J. de Swart (1994b), “Construction of high-quality nn eralized nuclear contacts and momentum distributions,” potential models,” Phys. Rev. C 49, 2950–2962. Phys. Rev. C92 (5), 054311, arXiv:1503.07047 [nucl-th]. Stone, J R, P. A. M. Guichon, P. G. Reinhard, and Wiringa, R B, R. Schiavilla, Steven C. Pieper, and J. Carl- A. W. Thomas (2016), “Finite Nuclei in the Quark- son (2014a), “Nucleon and nucleon-pair momentum distri- Meson Coupling Model,” Phys. Rev. Lett. 116 (9), 092501, butions in A ≤ 12 nuclei,” Phys. Rev. C89 (2), 024305, arXiv:1601.08131 [nucl-th]. arXiv:1309.3794 [nucl-th]. Strauch, S, et al. (Jefferson Lab E93-049) (2003), “Polariza- Wiringa, R B, R. Schiavilla, Steven C. Pieper, and J. Carl- tion transfer in the He-4 (polarized-e, e-prime polarized-p) son (2014b), “Nucleon and nucleon-pair momentum distri- H-3 reaction up to Q**2 = 2.6-(GeV/c)**2,” Phys. Rev. butions in a ≤ 12,” Phys. Rev. C 89, 024305. Lett. 91, 052301, arXiv:nucl-ex/0211022 [nucl-ex]. Wiringa, R B, V. G. J. Stoks, and R. Schiavilla (1995a), “Ac- Strauch, Steffen (2012), “Hadron medium modifications,” curate nucleon-nucleon potential with charge-independence Proceedings, Resonance Workshop at UT Austin on breaking,” Phys. Rev. C 51, 38–51. Hadronic resonance production in heavy ion and elemen- Wiringa, Robert B, V. G. J. Stoks, and R. Schiavilla (1995b), tary collisions, EPJ Web Conf. 36, 00016. “An Accurate nucleon-nucleon potential with charge inde- Subedi, R, et al. (2008), “Probing Cold Dense Nuclear Mat- pendence breaking,” Phys. Rev. C51, 38–51, arXiv:nucl- ter,” Science 320, 1476–1478, arXiv:0908.1514 [nucl-ex]. th/9408016 [nucl-th]. Tang, A, et al. (2003), “n-p short-range correlations from Wong, S S M (1998), Introductory nuclear physics. (p,2p+n) measurements,” Phys. Rev. Lett. 90, 042301. Yaron, I, et al. (A1) (2016), “Polarization-transfer measure- Thomas, Anthony William, and Wolfram Weise (2001), The 57

ment to a large-virtuality bound proton in the deuteron,” Zeller, G P, et al. (2003), “Erratum: Precise determination arXiv:1602.06104 [nucl-ex]. of electroweak parameters in neutrino-nucleon scattering Zeller, G P, et al. (2002), “Precise determination of elec- [phys. rev. lett. 88 , 091802 (2002)],” Phys. Rev. Lett. 90, troweak parameters in neutrino-nucleon scattering,” Phys. 239902. Rev. Lett. 88, 091802.