1: \documentclass[aps,prc,groupedaddress,showpacs,twocolumn,floatfix]{revtex4}
2: %\documentclass[aps,prc,groupedaddress,showpacs,preprint,floatfix]{revtex4}
3:
4: \usepackage{graphicx}
5: \usepackage[dvips]{color}
6: \usepackage{colordvi}
7:
8: \def\micro{\mu}
9: \def\deg{^\circ}
10: \def\gtorder{\mathrel{\raise.3ex\hbox{$>$}\mkern-14mu
11: \lower0.6ex\hbox{$\sim$}}}
12: \def\ltorder{\mathrel{\raise.3ex\hbox{$<$}\mkern-14mu
13: \lower0.6ex\hbox{$\sim$}}}
14: \def\mugegm{\mu_p G_E / G_M}
15: \def\gegm{G_E / G_M}
16: \def\ge{G_E}
17: \def\gm{G_M}
18: \def\se{\sigma_E}
19: \def\sm{\sigma_M}
20:
21: \def\beq{\begin{equation}}
22: \def\eeq{\end{equation}}
23: \def\ba{\begin{eqnarray*}}
24: \def\ea{\end{eqnarray*}}
25: \newcommand{\rt}{\langle r \rangle_{(2)}}
26: \begin{document}
27:
28: \title{Proton radii and two-photon exchange}
29:
30: \author{Peter G.~Blunden}
31: \affiliation{Dept.~of Physics and Astronomy, University of Manitoba,
32: Winnipeg, MB, Canada\ \ R3T~2N2}
33:
34: \author{Ingo Sick}
35: \affiliation{Dept.~f\"{u}r Physik und Astronomie, Universit\"{a}t Basel,
36: CH4056 Basel, Switzerland}
37:
38: \date{\today}
39:
40: \begin{abstract}
41: We investigate the effect of two-photon exchange processes upon the
42: $rms$- and Zemach radii extracted from electron-proton scattering. We
43: find that the changes are small and do not help to explain the
44: discrepancy between experimental and calculated HFS in the hydrogen
45: atom.
46:
47: \end{abstract}
48:
49: \pacs{25.30.Bf, 21.20.Ft, 14.20.Dh}
50:
51: \maketitle
52:
53: \section{Introduction}
54: %
55: The recent progress in measurements of transition energies in the
56: Hydrogen atom has been remarkable. The 2p-1s transition energy and the
57: 1s hyperfine structure interval (HFS) are now known to 14 and 12
58: significant digits, respectively \cite{Niering00,Eides01}. The
59: interpretation of these energies in terms of {\em e.g.} tests of QED now
60: depends entirely on the accuracy with which the proton finite-size
61: corrections are known.
62:
63: These finite-size corrections can be determined from elastic
64: electron-proton scattering at low momentum transfer $q$. The proton
65: moment relevant for the 2p-1s energy (and the Lamb shift) is the charge
66: rms-radius $r_{rms}$ which has been extracted from the {\em world} data
67: on e-p scattering \cite{Sick03} to 0.895$\pm$0.018~fm. With this radius
68: as input, calculated and experimental 2p-1s transition energies agree
69: within the error bars. The proton moment relevant for the HFS is the
70: Zemach moment, derived from a convolution of charge- and magnetization
71: densities. This radius has also been extracted from the {\em world} e-p
72: data in Ref.~\cite{Friar04} which found $\rt=1.086\pm0.012$~fm. The
73: corresponding calculated HFS disagrees with experiment by 3.6(5)~ppm
74: \cite{Friar04} (see also \cite{Brodsky05a,Brodsky05b,Friar05}), where the uncertainty
75: is dominated by the uncertainty in $\rt$.
76:
77: This disagreement is partly explained by nuclear polarization effects in
78: the hydrogen atom, processes involving virtual excitation of the proton
79: to intermediary continuum states. Faustov {\em et al.} \cite{Faustov02}
80: have calculated this correction and find a contribution of 1.6~ppm in
81: the right direction using the experimental $g_1(q)$ and $g_2(q)$ spin
82: structure functions of the proton. The uncertainty of this correction is
83: hard to estimate as the nuclear polarization correction depends on $g_1,
84: g_2$ at very low $q$, where these functions are poorly known.
85:
86: Before assigning the remaining discrepancy to this correction (or to not
87: yet calculated higher-order QED terms affecting the HFS) one should
88: note, however, that the moments extracted from electron scattering are
89: based on an interpretation of the data in one-photon exchange (plus
90: exchange of additional soft photons responsible for the Coulomb
91: distortion of the electron waves \cite{Sick98}). Electron scattering is
92: also subject to exchange of two hard photons, which {\em e.g.} have a
93: considerable effect upon the proton charge form factor $G_E$ as
94: determined from longitudinal/transverse separations at very large
95: momentum transfers \cite{Blunden03}.
96:
97: In this Rapid Communication, we investigate the role of the two-photon
98: exchange in the determination of the proton moments, in order to find
99: out whether these corrections could be responsible for the discrepancy
100: with the HFS values.
101:
102: \section{Calculation of two-photon exchange}
103: %
104: Details of the two-photon exchange correction to elastic electron-proton
105: scattering are described elsewhere \cite{Blunden03, Blunden05}. We
106: consider the contribution to the two-photon exchange box and crossed-box
107: amplitudes with an intermediate nucleon. At the low momentum transfers
108: of interest here (up to $q=4$~fm$^{-1}$) the contribution of an
109: intermediate $\Delta$ (or higher resonances) to the two-photon exchange
110: amplitude is negligible \cite{Kondratyuk05}.
111:
112: Hadronic form factors consistent with the experimentally measured charge
113: and magnetic form factors $G_E(q^2)$ and $G_M(q^2)$ are introduced at
114: the photon-nucleon vertices. The two-photon exchange correction is
115: relatively insensitive to the particular choice of hadronic form factors
116: \cite{Blunden05}. For very low $q$, the two-photon exchange correction
117: behaves the same as that found for scattering from a point particle ({\em
118: e.g.} electron-muon scattering), and is therefore completely independent
119: of the hadronic form factors in this limit. This is a useful check on
120: our calculation.
121:
122: In order to give an impression of the calculated results, we show in
123: Fig.~\ref{tdep} the two-photon correction for a typical electron energy.
124: The two-photon exchange contribution is compared to the contribution
125: involving only the piece from a second {\em soft} photon (Coulomb
126: distortion of the electron wave), calculated according to
127: Ref.~\cite{Sick98} in second Born approximation. The same contribution,
128: but for a point-nucleus, is given for comparison. Fig.~\ref{tdep} shows
129: that at forward angles the two-photon contribution is entirely dominated
130: by Coulomb distortion, while at backward angles the exchange of two hard
131: photons contributes appreciably.
132:
133: \begin{figure}
134: \includegraphics[scale=0.45,clip]{test2ps.ps}
135: \caption{Relative contribution of two-photon exchange to elastic e-p scattering at
136: E=160~MeV. The results in second Born-approximation account for the Coulomb distortion
137: (exchange of second soft photon) only.
138: \label{tdep}}
139: \end{figure}
140:
141: \section{Analysis of world e-p data}
142: %
143: It has been previously shown in Ref.~\cite{Sick03} that for an optimal
144: determination of the proton moments a parameterization in terms of a
145: Continued Fraction (CF) expansion should be used. For this
146: parameterization, the contribution of model-dependence has been
147: investigated in detail. As in Ref.~\cite{Sick03} we use the cross
148: sections up to a maximum momentum transfer of 4~fm$^{-1}$.
149:
150: We use the {\em world} cross sections on e-p scattering (for references
151: see \cite{Sick03}). These data have been corrected for the contribution
152: of two-photon effects (omitting the second Born Coulomb corrections, which are
153: already included in the two-photon contribution), and then fitted with
154: 5-parameter CF expressions
155: for both the charge and the magnetic form factors $G_E(q^2)$ and $G_M(q^2)$.
156: The two-photon corrected longitudinal/transverse separation is thus done
157: implicitly during the fit. The charge-rms radius is obtained from the
158: slope at $q^2$=0 of $G_E(q^2)$, the Zemach moment is obtained via \ba
159: \langle r \rangle_{(2)} = -\frac{4}{\pi} \int_0^{\infty} \frac{d q}{q^2}
160: (G_E(q^2) G_M(q^2) - 1) ~~. \ea The statistical errors
161: have been determined using the error matrix, the systematic errors have
162: been obtained by changing the data sets individually by their systematic
163: errors, refitting the data and adding quadratically all the resulting
164: changes of the moments. The contribution of the model dependence is
165: accounted for as well. \\
166:
167: \section{Results and conclusion}
168: %
169: The change of $\rt$ found when removing the contribution of two-photon
170: exchange (beyond Coulomb distortion) amounts to +0.0052~fm. This change
171: is small ($\sim$40\% of the error bar) and goes in the wrong direction
172: in terms of helping to explain the HFS discrepancy. The Zemach moment
173: with two-photon effects corrected for amounts to \ =1.091$\pm$0.012~fm.
174: The change in the charge-rms radius, +0.0015~fm, is also small. The
175: radius after two-photon correction amounts to 0.897$\pm$0.018~fm.
176:
177: From these results we conclude that the discrepancy between calculated
178: and experimental HFS in the hydrogen atom cannot be attributed to
179: two-photon exchange contributions to (e,e) that could have falsified the
180: Zemach moment from e-p scattering. The origin of this discrepancy
181: presumably has to be sought in the uncertain nuclear polarization
182: correction to HFS, or potentially not yet calculated higher-order
183: contributions to the HFS.
184:
185: \begin{acknowledgments}
186: This work has been supported by NSERC (Canada) and by the Schweizerische
187: Nationalfonds.
188: %
189: \end{acknowledgments}
190:
191: \bibliographystyle{unsrt}
192: \bibliography{/usr/users/sick/sum2,/usr/users/sick/friar}
193: %\begin{thebibliography}{10}
194:
195:
196: %\end{thebibliography}
197:
198: \end{document}
199: