NUINT04 - Italy Arie Bodek, University of Rochester Un-Ki Yang, University of Chicago Update on low Q2 Corrections to LO-GRV98 PDFs Work in 2004: Note this is a Pseudo Leading Order Model

1. Vector PDFs: Extrat the low Q2 corrections to d and u valence quarks separately now by including all inelastic electron-hydrogen, electron-deuterium (including Jlab), photo-production on hydrogen and photo-production on deuterium in the fits (all fits include the c-cbar photongluon fusion contribution at high energy - this cross section is zero in leading order) 2. Compare to photo-production and Jlab electro-production Data in Resonance region (data not included in fit yet) 3. Compare to neutrino CCCFR-Fe, CDHS-Fe, CHORUS-Pb differential cross section (without c-cbar boson-fusion in

yet - to be added next since it is high energy data) assume V=A 4. We aave a model for Axial low Q2 PDFs, but need to compare to low energy neutrino data to get exact 1 For applications to Neutrino Oscillations at Low Energy (down to Q2=0) the best approach is to use a LO PDF analysis (including a more sophisticated

target mass analysis) and include the missing QCD higher order terms in the form of Empirical Higher Twist Corrections. Reason: For Q2>1 both Current Algebra exact sum rules (e.g. Adler sum rule) and QCD sum rules (e.g. momentum sum rule) are satisfied. This is why duality works in the resonance region (so use NNLO QCD analysis) For Q2<1, QCD corrections diverge, and all QCD sum rules (e.g momentum sum rule) break down, and duality breaks down in

the resonance region. In contrast, Current Algebra Sum rules e,g, Adle sum rule which is related to the Number of (U minus D) Valence quarks) are valid. 2 Modified LO = Pseudo NNLO approach for low energies Applications to Jlab and Neutrino Oscillations q

P=M mf=M* (final state interaction) Resonance, higher twist, and TM Q2+mf2+O(mf2-mi2) +A w= M (1+(1+Q2/2) )1/2 +B

A : initial binding/target mass effect plus higher order terms B: final state mass mf2 , mand photo- production limit (Q2 =0) Xbj= Q2 /2 M Original approach

(NNLO QCD+TM) was to explain the nonperturbative QCD effects at low Q2, but now we reverse the approach: Use LO PDFs and effective target mass and final state masses to account for initial target mass, final target mass, and

missing higher orders K factor to PDF, Q2/[Q2+C] 3 Initial quark mass m I and final mass ,mF=m * bound in a proton of mass M Summary: INCLUDE quark initial Pt) Get scaling (not x=Q2/2M ) for a general parton Model q=q3,q0 Is the correct variable which is

P = P ,P ,mF=m* P = P ,P ,mI Invariant in any frame : q3 and P in P= P + P ,M opposite directions 0 I F

0 PI0 +PI 3 = 0 3 PP +PP PI ,P0 q3,q0 quark

photon W = 2 F Q +m + A {M [1 + (1 +Q 2 / 2 )] + B}

Most General Case: w= F 0 F

3 F 3 Special cases: (1) Bjorken x, xBJ=Q2/2M, -> x 2, m F 2 : Slow Rescaling

For m F 2 = m I 2 =0 (q + PI )2 = PF2 q 2 + 2 PI q + PI2 = mF2 2 3 I (m I2 , Pt = 0)

and High (2) Numerator as in charm production (3) Denominator: Target mass term =Nachtman Variable =Light Cone Variable =Georgi Politzer Target

Mass var. (all the same ) (Derivation in Appendix) [Q2 +B] / [ M (1+(1+Q2/2) ) 1/2 +A] (with A=0, B=0) where 2Q2 = [Q2+ m F 2 - m I 2 ] + { ( Q2+m F 2 - m I 2 ) 2 + 4Q2 (m I 2 +P2t) }1/2 Bodek-Yang: Add B and A to account for effects of additional m2 from NLO and NNLO (up to infinite order) QCD effects. For case w with P24t =0

http://web.pas.rochester.edu/~icpark/MINERvA/ 5 Correct for Nuclear Effects measured in e/muon expt. Comparison of Fe/D F2 data

In resonance region (JLAB) Versus DIS SLAC/NMC data In TM (C. Keppel 2002). 6 7 8

9 10 11 12 Note: All data has been radiatively corrected.

Neutrino Experiments MUST apply radiative corrections - Tables can be provided (no consistent PDF type analysis can be done without radiative corrections - e.g. Bardin) -----------------------------------------------------Work in Progress: Now working on the axial structure functions and next plan to work on resonance fits. Next: JUPITER at Jlab (Bodek,Keppel) will provided electron-Carbon (also e-H and e-D and other nuclei such as e-Fe) in resonance region.

Next: MINERvA at FNAL (McFarland, Morfin) will provide Neutrino-Carbon data at low energies. 13 Comparison of LO+HT to neutrino data on Iron [CCFR] (not used in this fit V=A) Last years fits Construction

Apply nuclear corrections using e/m scattering data. (Next slide) Calculate F2 and xF3 from the modified PDFs with w Use R=Rworld fit to get

2xF1 from F2 Implement charm mass effect through w slow rescaling algorithm, for F2 2xF1, and XF3 w PDFs GRV98 modified ---- GRV98 (x,Q2) unmodified Left neutrino, Right antineutrino

The modified GRV98 LO PDFs with a new scaling variable, w describe CCFR diff. Cross sect. (E=30300 GeV) well (except at the lowest x) E= 55 GeV is shown14 Compare Last Year to this Year (this year also use R1998 instead of R world) V=A, lowest x no c-cbar contribution w PDFs GRV98 modified Very low X needs work ---- GRV98 (x,Q2) unmodified

(maybe PDF f(x) needed) Left neutrino, Right antineutrino Last year 15 16 17

18 End CCFR neutrino and antineutrino Begin CDHS Neutrino Note, further studies needed. Are the data reliable at low x (rad corrections, resolution, flux). If R is smaller than R 1998 (nuclear shadowing) (waiting for Jlab data). What about c-cbar contribution (currently being calculated). Or make sure that differential cross section code has no bugs. With that caveat we show some data

versus model. We we will be better agreement between data and model if R is smaller than R 1998, and if we include c-cbar NLO terms. However, Is the axial F2 larger? Whatabout shadowing for axial versus vector especially for iron and lead. All under investigation 19

20 21 22 23 24

25 Extra comparisons 26 27 28

29 30 31 End CCFR neutrino and antineutrino

Begin CDHS Neutrino 32