Gottfried sum from the ratio<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"><mml:mrow><mml:msubsup><mml:mrow><mml:mi mathvariant="italic">F</mml:mi></mml:mrow><mml:mrow><mml:mn>2</mml:mn></mml:mrow><mml:mrow><mml:mi mathvariant="italic">n</mml:mi></mml:mrow></mml:msubsup></mml:mrow></mml:math>/<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"><mml:mrow><mml:msubsup><mml:mrow><mml:mi mathvariant="italic">F</mml:mi></mml:mrow><mml:mrow><mml:mn>2</mml:mn></mml:mrow><mml:mrow><mml:mi mathvariant="italic">p</mml:mi></mml:mrow></mml:msubsup></mml:mrow></mml:math>

P. Amaudruz, M. Arneodo, A. Arvidson, B. Badełek, G. Baum, J. Beaufays, I. Bird, M. Botje, C. Broggini, W. Brückner, A. Brüll, W.j. Burger, J. Ciborowski, Robert Crittenden, R. van Dantzig, H. Döbbeling, J. Domingo, J. Drinkard, A. Dzierba, H. Engelien, M.i. Ferrero, M. L. Fluri, P. Grafström, P.a. Haas, Matthieu van der Heijden, C. A. Heusch, Q. Ingram, A. Jacholkowska, K. Janson, M. de Jong, E.m. Kabuß, R. Kaiser, T. Ketel, F. Klein, B. Korzen, U. Krüner, S. Kullander, U. Landgraf, F. Lettenström, T. Lindqvist, G.k. Mallot, C. Mariotti, G. van Middelkoop, Y. Mizuno, J. Nassalski, D. Nowotny, N. Pavel, H. Peschel, C. Peroni, Bogdan Povh, R. Rieger, K. Rith, K. Röhrich, E. Rondio, L. Ropelewski, V.d. Samoylenko, Christoph Scholz, R. A. Schumacher, U. Sennhauser, F. Sever, T.-a. Shibata, M. Siebler, A. Simon, A. Staiano, G. N. Taylor, M. Treichel, J.-l. Vuilleumier, Th. Walcher, K. Welch, R. Windmolders, Frank ZetschePhysical 1991 · DOI

Abstract

Experimental results obtained at the CERN Super Proton Synchrotron on the structure-function ratio $F_2^n/F_2^p$ in the kinematic range $0.004 < x < 0.8$ and $0.4 < Q^2 < 190 \text{ GeV}^2$, together with the structure function $F_2^d$ determined from a fit to published data, are used to derive the difference $F_2^p(x) - F_2^n(x)$. The value of the Gottfried sum $\int (F_2^p - F_2^n) dx/x = 0.240 \pm 0.016$ is below the quark-parton-model expectation of 1/3.

Introduction

The Gottfried sum rule, proposed in the 1960s, serves as a fundamental test of the Quark-Parton Model (QPM). It predicts that the integral of the difference between the proton and neutron structure functions ($F_2^p - F_2^n$) divided by the Bjorken scaling variable $x$ should equal exactly 1/3. This prediction rests on a key assumption: while protons and neutrons have different "valence" quarks, the "sea" of virtual quark-antiquark pairs produced by gluons should be flavor-symmetric (i.e., the number of $\bar{u}$ quarks equals the number of $\bar{d}$ quarks). If this symmetry holds, the sea contributions cancel out in the subtraction, leaving only the contribution of the valence quarks.

Previous experiments had hinted at a lower value, but large systematic uncertainties and limited kinematic ranges made those results inconclusive. The New Muon Collaboration (NMC) sought to provide a definitive measurement by extending the reach to very low $x$ values.

Experimental Methods and Datasets

The researchers used Deep-Inelastic Scattering (DIS), firing high-energy muons (90 and 280 GeV) at hydrogen and deuterium targets at the CERN Super Proton Synchrotron. By exposing hydrogen and deuterium targets to the beam simultaneously, they could measure the ratio of structure functions ($F_2^n/F_2^p$) directly, which cancels out many systematic errors related to beam flux and spectrometer acceptance.

The measured ratio $F_2^n/F_2^p$ as a function of $x$. The data points from different beam energies show consistent behavior across a wide kinematic range.

To determine the absolute structure function $F_2^d$, the authors performed a global fit using published data from several major experiments: SLAC, BCDMS, EMC, and CHIO. The combined dataset covered the kinematic range $0.004 < x < 0.8$ and $0.4 < Q^2 < 190 \text{ GeV}^2$. The push into the very low $x$ region (down to 0.004) was critical because the Gottfried sum integral is weighted by $1/x$, making the low-$x$ contribution highly influential.

Key Results and Insights

The core finding is that the Gottfried sum ($S_G$) is significantly lower than the theoretical prediction. The measured value was $0.240 \pm 0.016$, which is roughly five standard deviations away from the QPM prediction of 1/3 ($\approx 0.333$).

The integral $\int_x^1 (F_2^p - F_2^n) dx/x'$ (open symbols, left scale) and the difference $F_2^p - F_2^n$ (solid symbols, right scale) as a function of $x$. The QPM prediction of 1/3 is indicated at the top left.

This result implies that the assumption of $\bar{u}(x) = \bar{d}(x)$ is likely incorrect. The data suggests an excess of $\bar{d}$ quarks over $\bar{u}$ quarks in the proton sea: $$\int_0^1 (\bar{d}(x) - \bar{u}(x)) dx = 0.140 \pm 0.024$$

The observed asymmetry points toward non-perturbative processes, such as the "pion cloud" model, where a proton occasionally fluctuates into a neutron and a $\pi^+$ meson ($p \to n + \pi^+$). Since the $\pi^+$ contains a $\bar{d}$ quark, this naturally explains the $\bar{d}$ excess.

Limitations and Future Work

Since experiments cannot reach $x=0$, the authors had to extrapolate their data using Regge theory. While they argue this contribution is small, it remains a source of theoretical debate. Additionally, nuclear effects like "shadowing" (where nucleons in a nucleus shield each other) could slightly alter the results. The authors estimated this effect to be around 10%, which would increase $S_G$ but not enough to reach 1/3.

The paper opened the door for more complex models of nucleon structure, specifically those incorporating chiral symmetry breaking and meson-loop corrections. It also set the stage for future Drell-Yan experiments to directly measure the $\bar{d}/\bar{u}$ ratio.

Summary updated

Abstract

Experimental results obtained at the CERN Super Proton Synchrotron on the structure-function ratio ${\mathit{F}}_{2}^{\mathit{n}}$/${\mathit{F}}_{2}^{\mathit{p}}$ in the kinematic range 0.004x0.8 and 0.4${\mathit{Q}}^{2}$190 ${\mathrm{GeV}}^{2}$, together with the structure function ${\mathit{F}}_{2}^{\mathit{d}}$ determined from a fit to published data, are used to derive the difference ${\mathit{F}}_{2}^{\mathit{p}}$(x)-${\mathit{F}}_{2}^{\mathit{n}}$(x). The value of the Gottfried sum F(${\mathit{F}}_{2}^{\mathit{p}}$-${\mathit{F}}_{2}^{\mathit{n}}$)dx/x=0.240\ifmmode\pm\else\textpm\fi{}0.016 is below the quark-parton-model expectation of 1/3.
BibTeX
@inproceedings{amaudruz1991gottfried,
  author = {P. Amaudruz and M. Arneodo and A. Arvidson and B. Badełek and G. Baum and J. Beaufays and I. Bird and M. Botje and C. Broggini and W. Brückner and A. Brüll and W.J. Burger and J. Ciborowski and Robert Crittenden and R. van Dantzig and H. Döbbeling and J. Domingo and J. Drinkard and A. Dzierba and H. Engelien and M.I. Ferrero and M. L. Fluri and P. Grafström and P.A. Haas and Matthieu van der Heijden and C. A. Heusch and Q. Ingram and A. Jacholkowska and K. Janson and M. de Jong and E.M. Kabuß and R. Kaiser and T. Ketel and F. Klein and B. Korzen and U. Krüner and S. Kullander and U. Landgraf and F. Lettenström and T. Lindqvist and G.K. Mallot and C. Mariotti and G. van Middelkoop and Y. Mizuno and J. Nassalski and D. Nowotny and N. Pavel and H. Peschel and C. Peroni and Bogdan Povh and R. Rieger and K. Rith and K. Röhrich and E. Rondio and L. Ropelewski and V.D. Samoylenko and Christoph Scholz and R. A. Schumacher and U. Sennhauser and F. Sever and T.-A. Shibata and M. Siebler and A. Simon and A. Staiano and G. N. Taylor and M. Treichel and J.-L. Vuilleumier and Th. Walcher and K. Welch and R. Windmolders and Frank Zetsche},
  title = {{Gottfried sum from the ratio<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"><mml:mrow><mml:msubsup><mml:mrow><mml:mi mathvariant="italic">F</mml:mi></mml:mrow><mml:mrow><mml:mn>2</mml:mn></mml:mrow><mml:mrow><mml:mi mathvariant="italic">n</mml:mi></mml:mrow></mml:msubsup></mml:mrow></mml:math>/<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"><mml:mrow><mml:msubsup><mml:mrow><mml:mi mathvariant="italic">F</mml:mi></mml:mrow><mml:mrow><mml:mn>2</mml:mn></mml:mrow><mml:mrow><mml:mi mathvariant="italic">p</mml:mi></mml:mrow></mml:msubsup></mml:mrow></mml:math>}},
  booktitle = {Physical},
  year = {1991},
  doi = {10.1103/physrevlett.66.2712},
}
Figure