1: \documentstyle[aps,psfig,prl,multicol]{revtex}

2: %\documentstyle[aps,psfig,preprint,tighten,floats]{revtex}

3: %\documentstyle[aps,preprint,tighten]{revtex}

4: 

5: \begin{document}

6: %\preprint{\Large \it Second draft}

7: \vspace{1.0cm}

8: 

9: \title{Nucleon Charge and Magnetization Densities}

10: \author{James J. Kelly}

11: \address{ Department of Physics, University of Maryland, 

12:           College Park, MD 20742 }

13: \date{Revised: April 22, 2002}

14: \maketitle

15: 

16: 

17: \begin{abstract}

18: We use a relativistic prescription to extract charge and magnetization

19: densities from data for the scattering of high-energy electrons by nucleons.

20: A Fourier-Bessel analysis is used to minimize the model dependence of the

21: fitted densities.

22: We find that the neutron and proton magnetization densities are very similar, 

23: but the proton charge density is significantly softer.

24: A useful measurement of the neutron charge density is obtained, although

25: the relative uncertainty in the interior will remain substantially larger than

26: for the other densities until precise new data at higher $Q^2$ become 

27: available.

28: \end{abstract}

29: \pacs{14.20.Dh,13.40.-Gp}

30: 

31: \begin{multicols}{2}[]

32: \narrowtext

33: \vfill

34: \eject

35: 

36: The electromagnetic structure of nucleons can be investigated using the

37: electric and magnetic Sachs form factors, $G_E$ and $G_M$, measured

38: by elastic scattering of high-energy electrons.

39: Recent experiments with intense high-polarization electron beams have 

40: improved the quality of the data for nucleon elastic form factors.

41: Perhaps the most dramatic observation \cite{MKJones00,Gayou02}

42: is that the ratio between the electric and magnetic form factors for the 

43: proton decreases sharply for $1 < Q^2 < 6$ (GeV/$c$)$^2$.

44: It was suggested that these results demonstrate that the proton charge is

45: distributed over a larger volume than its magnetization, 

46: but radial densities were not obtained.  

47: At low momentum transfer, the form factors are related to Fourier 

48: transforms of the charge and magnetization densities, 

49: but this interpretation is more complicated at high momentum transfer.   

50: Nevertheless, although models of nucleon structure can often calculate

51: the form factors directly, it remains desirable to relate the form 

52: factors to spatial densities because our intuition tends to be based more

53: firmly in space than momentum transfer.

54: In this paper we use a Fourier-Bessel analysis, 

55: together with a relativistic relationship between form factors and densities,

56: to parametrize the nucleon electromagnetic form factors in terms of

57: charge and magnetization densities.

58: 

59: The nucleon electromagnetic vertex function takes the form

60: \begin{equation}

61: \Gamma^\mu = F_1(Q^2)\gamma^\mu + 

62: \kappa F_2(Q^2) \frac{i \sigma^{\mu\nu}q_\nu}{2m}

63: \end{equation}

64: where  $F_1$ and $F_2$ are known as Dirac and Pauli form factors,

65: $\kappa$ is the anomalous part of the magnetic moment, 

66: and $\gamma^\mu$ and $\sigma^{\mu\nu}$ are the usual Dirac matrices 

67: ({\it e.g.}, \cite{BjorkenDrella}).

68: This current operator appears simplest in the Breit frame where the 

69: nucleon approaches with initial momentum $-\vec{q}_{B}/2$, receives 

70: three-momentum transfer $\vec{q}_B$, and leaves with final momentum 

71: $\vec{q}_B/2$ such that the energy transfer vanishes.

72: In the Breit frame for a particular value of $Q^2$, the current  

73: separates into electric and magnetic contributions \cite{Sachs62}

74: \begin{equation}

75: \label{eq:BreitCurrent}

76: \bar{u}(p^\prime,s^\prime) \Gamma^\mu u(p,s) = 

77: \chi^\dagger_{s^\prime}

78: \left( G_E + \frac{i\vec{\sigma}\times\vec{q}_B}{2m} G_M 

79: \right) \chi_s

80: \end{equation}

81: where $\chi_s$ is a two-component Pauli spinor and where the

82: Sachs form factors are given by

83: \begin{mathletters}

84: \begin{eqnarray}

85: G_E &=& F_1 - \tau \kappa F_2 \\

86: G_M &=& F_1 + \kappa F_2

87: \end{eqnarray}

88: \end{mathletters}

89: with $\tau=(Q/2m)^2$.

90: Naively it would appear that one could obtain charge and magnetization

91: densities as Fourier transforms of the electric and magnetic form factors,

92: but each momentum transfer $Q$ specifies a different Breit frame.

93: Early experiments with modest $Q^2$ suggested that 

94: \begin{displaymath}

95: G_{Ep} \approx \frac{G_{Mp}}{\mu_p} \approx \frac{G_{Mn}}{\mu_n}

96: \approx G_D

97: \end{displaymath}

98: where $\mu$ is the magnetic moment in nuclear magnetons and

99: $G_D(Q^2) = (1 + Q^2/\Lambda^2)^{-2}$ with 

100: $\Lambda^2 = 0.71$ (GeV/$c$)$^2$ is known as the dipole form factor.

101: Similarly, data for $G_{En}$ at low $Q^2$ can be described by the 

102: Galster parametrization \cite{Galster71}

103: \begin{equation}

104: \label{eq:Galster}

105: G_{En}(Q^2) \approx -\mu_n G_D(Q^2) \frac{A \tau}{1+B\tau}

106: \end{equation}

107: where $A$ and $B$ are constants.

108: 

109: Let $\rho_{ch}(r)$ and $\rho_{m}(r)$ represent spherical intrinsic charge

110: and magnetization densities normalized according to

111: \begin{displaymath}

112: \int dr \; r^2 \rho_{ch}(r) = Z, \hspace{0.25cm}

113: \int dr \; r^2 \rho_{m}(r) = 1

114: \end{displaymath}

115: %\begin{mathletters}

116: %\begin{eqnarray}

117: %\int dr \; r^2 \rho_{ch}(r) &=& Z \\

118: %\int dr \; r^2 \rho_{m}(r) &=& 1

119: %\end{eqnarray}

120: %\end{mathletters}

121: where $Z=0,1$ is the nucleon charge.

122: An {\it intrinsic form factor} can then be defined by the

123: Fourier-Bessel transform

124: \begin{equation}

125: \tilde{\rho}(k) = \int dr \; r^2 j_0(kr) \rho(r)

126: \end{equation}

127: where $k$ is the intrinsic spatial frequency (or wave number).

128: If one knew how to obtain $\tilde{\rho}(k)$ from data for the

129: appropriate Sachs form factor, the intrinsic density could be

130: obtained simply by inverting the Fourier transform, such that

131: \begin{equation}

132: \label{eq:rhor}

133: \rho(r) = \frac{2}{\pi} \int_0^\infty dk \; k^2 j_0(kr) \tilde{\rho}(k)

134: \end{equation}

135: The naive or nonrelativistic inversion from form factor to density

136: replaces $Q$ by $k$ and $\tilde{\rho}(k)$ by $G(Q^2)$,

137: but leads to unsatisfactory densities for dipole or Galster form

138: factors --- the direct Fourier transform of a dipole form factor is 

139: an exponential density with a cusp at the origin.

140: The corresponding density for the Galster form factor is more

141: complicated but also features an unphysical cusp at the origin. 

142: 

143: Although there is no rigorous model-independent relationship between

144: the Sachs form factor and the corresponding static density in the

145: rest frame, a variety of models offer prescriptions of the form

146: \begin{mathletters}

147: \label{eq:rhok}

148: \begin{eqnarray}

149: \tilde{\rho}_{ch}(k) &=& G_E(Q^2) (1+\tau)^{\lambda_E} \\

150: \mu \tilde{\rho}_{m}(k) &=& G_M(Q^2) (1+\tau)^{\lambda_M} 

151: \end{eqnarray}

152: \end{mathletters}

153: where $\lambda$ is a model-dependent constant and

154: where the intrinsic spatial frequency $k$ is related to the invariant

155: momentum transfer by

156: \begin{equation}

157: \label{eq:k}

158: k^2 = \frac{Q^2}{1+\tau}

159: \end{equation}

160: The most important relativistic effect is the Lorentz contraction of  

161: spatial distributions in the Breit frame and the corresponding increase

162: of spatial frequency represented by the 

163: factor of $(1+\tau)$ in Eq. (\ref{eq:k}).

164: A measurement with Breit-frame momentum transfer $q_B = Q$ probes a 

165: reduced spatial frequency $k$ in the rest frame.

166: The Sachs form factor for a large invariant momentum transfer $Q^2$ is 

167: determined by a much smaller spatial frequency $k^2=Q^2/(1+\tau)$ and 

168: thus declines much less rapidly with respect to $Q^2$ than the Fourier 

169: transform of the density declines with respect to $k^2$.

170: Licht and Pagnamenta \cite{Licht70b} demonstrated that by accounting

171: for Lorentz contraction a good fit to data for $G_{Ep}$ can obtained using 

172: a Gaussian density typical of quark models.

173: Thus, this procedure provides physically reasonable densities

174: free of cusps at the origin.

175: However, according to Eq. (\ref{eq:k}), the maximum spatial frequency

176: that is accessible with spacelike $Q^2$ is $k_m=2m$.

177: The limitation  $k<k_m$ can be interpreted as a consequence

178: of relativistic position fluctuations, known as {\it zitterbewegung},

179: which obscure details finer that the nucleon Compton wavelength.

180: 

181: The first attempt to relate elastic form factors to ground-state densities 

182: was made by Licht and Pagnamenta \cite{Licht70b} using a quark cluster model 

183: in the impulse approximation and a kinematic boost.

184: They proposed a relativistic inversion method using Eq. (\ref{eq:rhok})

185: with $\lambda_E=\lambda_M=1$.

186: However, these choices for $\lambda$ do not ensure that the Sachs form

187: factors scale with $Q^{-4}$, as expected from perturbative QCD, unless

188: restrictions are placed upon $\tilde{\rho}(k_m)$.

189: Using a more symmetric treatment of the cluster model, 

190: Mitra and Kumari \cite{Mitra77} found that $\lambda_E=\lambda_M=2$ 

191: automatically satisfy the perturbative QCD scaling relations at very 

192: large $Q^2$ without constraining $\tilde{\rho}(k_m)$.

193: More recently, Ji \cite{Ji91} obtained similar relationships with 

194: $\lambda_E=0$ and $\lambda_M=1$ using a relativistic Skyrmion 

195: model based upon a Lorentz invariant Lagrangian density for which the

196: classical soliton solution can be evaluated in any frame.

197: Quantum fluctuations were then evaluated after the boost.

198: Furthermore, the recent soliton calculation by Holzwarth \cite{Holzwarth96}

199: uses this prescription and accurately predicted the decline in 

200: $G_{Ep}/G_{Mp}$.

201: In this paper we employ the values $\lambda_E=0$ and $\lambda_M=1$

202: suggested by the soliton model and in a subsequent paper \cite{Kelly02a}

203: will investigate in detail the {\it discrete ambiguity} due to the 

204: choice of $\lambda$.  

205: 

206: To extract radial densities from the nucleon form factor data we 

207: employ techniques originally developed for fitting radial distributions 

208: to data for scattering of electrons or protons from nuclei

209: \cite{Dreher74,Friar73,Kelly88a}.

210: Simple models with a small number of parameters do not offer

211: sufficient flexibility to provide a realistic estimate of the uncertainty

212: in a radial density.

213: Rather, we employ linear expansions in complete sets of basis functions 

214: that are capable of describing any plausible radial distribution without 

215: strong {\it a priori} constraints upon its shape.

216: Such expansions permit one to estimate the uncertainties in the fitted

217: density due to both the statistical quality of the data and the 

218: inevitable limitation of experimental data to a frequency range,  

219: $k \leq k_{max}$.

220: The uncertainty due to limitation of $k$ is known as 

221: {\it incompleteness error}.

222: More detailed discussion of the method may be found in Refs. 

223: \cite{Dreher74,Friar73,Kelly88a},

224: but the basic idea is to supplement the experimental data by pseudodata

225: of the form $\tilde{\rho}(k_i) = 0 \pm \delta \tilde{\rho}(k_i)$

226: whose uncertainties are based upon a reasonable model of the

227: asymptotic behavior of the form factor for $k_i > k_{max}$

228: where $k_{max}$ is the spatial frequency corresponding to the 

229: maximum measured $Q^2$. 

230: On quite general grounds one expects the asymptotic form factor for a 

231: confined system to decrease more rapidly than $k^{-4}$

232: \cite{Friar73}.

233: Therefore, we assume that

234: \begin{equation}

235: \label{eq:env}

236: \delta\tilde{\rho}(k) = \tilde{\rho}(k_{max}) 

237: \left( \frac{k_{max}}{k} \right)^4

238: \end{equation}

239: Notice that we must apply this procedure to $\tilde{\rho}(k)$ instead

240: of the Sachs form factors because, according to Eq. (\ref{eq:k}), 

241: the accessible spatial frequencies are limited to $k < 2m$ for

242: spacelike $Q^2$ while convergence of Eq. (\ref{eq:rhor}) requires a

243: constraint upon $\tilde{\rho}(k)$ for inaccessible range $k>2m$.

244: 

245: The Fourier-Bessel expansion (FBE) takes the form

246: \begin{equation}

247: \rho(r) = \sum_n a_n j_0(k_n r) \Theta(R-r)

248: \end{equation}

249: where $\Theta$ is the unit step function, $R$ is the expansion radius,

250: $k_n=n\pi/R$ are the roots of the Bessel function, and $a_n$ are the

251: coefficients to be fitted to data.

252: One advantage of the FBE is that the contribution of each term to the form

253: factor is concentrated around its $k_n$ so that a coefficient $a_n$ is

254: largely determined by data with $k \sim k_n$.

255: The larger the expansion radius $R$, the smaller the spacing between successive

256: $k_n$ and the greater the sensitivity one has to variations in the form factor.

257: One should choose $R$ to be several times the root-mean-square radius

258: but not so large that an excessive number of terms is needed to span the 

259: experimental range of momentum transfer.

260: Terms with $k_n > k_{max}$ provide an estimate of the incompleteness error.

261: We chose $R = 4.0$ fm, but the results are insensitive to its exact value.

262: Small but undesirable oscillations in fitted densities at large radius

263: were suppressed using a {\it tail bias} based upon the method

264: discussed in Ref. \cite{Kelly91b}.

265: We employed a tail function of the form $t(r) \propto e^{-\Lambda r}$,

266: based upon the successful dipole parametrization for low $Q^2$, 

267: and included in the $\chi^2$ fit a penalty for strong deviations from

268: the tail function for $r>2.0$ fm.

269: The constraint on the neutron charge was also enforced using a 

270: penalty function.

271: The tail bias improves the convergence of moments of the density 

272: but has practically no effect upon a fitted density in the region where 

273: it is large. 

274: The error band for a fitted density is computed from the covariance matrix

275: for the $\chi^2$ fit and includes the incompleteness error, 

276: but does not include discrete ambiguities due to the choices for $\lambda$ 

277: in Eq. (\ref{eq:rhok}).

278: 

279: We selected the best available data in each range of $Q^2$, with an emphasis

280: upon recent data using recoil or target polarization wherever available.

281: $G_{Mp}$ data were taken from the compilation of H\"ohler \cite{Hohler76}

282: for $Q^2 < 0.15$ (GeV/$c$)$^2$ and for larger $Q^2$ from the analysis

283: of Brash {\it et al}.\ \cite{Brash01}

284: using the recent recoil polarization data for $G_{Ep}/G_{Mp}$ from  

285: Refs.\ \cite{MKJones00,Gayou02}.

286: Cross section data from Refs.\ \cite{Price71,Simon80} were used 

287: for $Q^2 < 1$ (GeV/$c$)$^2$ but cross section data for $G_{Ep}$ were

288: excluded for larger $Q^2$.

289: Similarly, the data for $G_{En}$ were limited to recent polarization data

290: \cite{Eden94b,Passchier99,Rohe99,Herberg99,Golak01,Zhu01},

291: with corrections for nuclear structure and final-state interactions

292: whenever available,

293: plus the analysis of $t_{20}$ and $T_{20}$

294: by Schiavilla and Sick \cite{Schiavilla01}. 

295: We also included the neutron charge radius obtained by 

296: Kopecky {\it et al}.\ \cite{Kopecky97}

297: from transmission of thermal neutron through liquid $^{208}$Pb and

298: $^{209}$Bi.

299: Finally, for $G_{Mn}$ we selected polarization data from \cite{Xu00} 

300: and cross section data from 

301: \cite{Rock82,Lung93,Markowitz93,Anklin94,Bruins95,Anklin98,Kubon02}.

302: 

303: Fits to the form factor data are shown in Fig.\ \ref{fig:ff} as bands 

304: that represent the uncertainties in the fitted form factors. 

305: The widths of these bands are comparable to the experimental precision

306: where data are available, but expand for larger $Q^2$ where the 

307: uncertainties are based upon Eq. (\ref{eq:env}).

308: Notice, however, that in this figure the form factors were divided by $G_D$, 

309: which decreases rapidly with $Q^2$.

310: Therefore, the impact of uncertainties at large $Q^2$ upon the fitted 

311: densities remains modest because, with the exception of $G_{En}$, 

312: the form factors themselves and their absolute uncertainties become rather 

313: small at their largest $Q^2$.

314: Although the low-$Q^2$ data for $G_{Mn}$ have improved in recent years,

315: significant systematic discrepancies remain.

316: Recent data from Refs.\ \cite{Xu00,Anklin94,Anklin98,Kubon02}

317: with small statistical uncertainties suggest a small dip near 0.2 and a peak 

318: near 1 (GeV/$c$)$^2$.

319: For $G_{En}$ we plot the Galster model \cite{Galster71} for comparison.

320: The simple two-parameter fit Galster {\it et al.} made to the rather poor

321: data available at that time did not permit a realistic estimate of the 

322: uncertainty in the form factor or fitted density and the apparent agreement 

323: with more modern data must be judged as remarkable but fortuitous.

324: 

325: Proton charge and magnetization densities are compared in 

326: Fig.\ \ref{fig:proton}.

327: Both densities are measured very precisely.

328: The new recoil-polarization data for $G_{Ep}$ decrease more rapidly

329: than either the dipole form factor or the magnetic form factor

330: for $Q^2 > 1$ (GeV/$c$)$^2$; consequently, the charge density is

331: significantly softer than the magnetization density of the proton.

332: Neutron densities are shown in Fig. \ref{fig:neutron}.

333: We find that the magnetization density for the neutron is very similar to 

334: that for the proton,

335: although the interior precision is not as good because the range of $Q^2$ 

336: is smaller and the experimental uncertainties larger.

337: Limitations in the range and quality of the $G_{En}$ data presently

338: available result in a substantially wider error band for the neutron

339: charge density.

340: Data at higher $Q^2$ are needed to improve the interior precision,

341: but a useful measurement of the interior charge density is obtained 

342: nonetheless.

343: The positive interior density is balanced by a negative surface lobe.

344: Note that polarization measurements are sensitive to the sign of the density, 

345: but that cross section measurements are not.

346: 

347: In summary, we have developed a parametrization of nucleon electromagnetic

348: form factors based upon radial densities expanded in complete bases.

349: Although we cannot claim that the fitted densities are unique, this method 

350: does provide an intuitively appealing parametrization of the data.

351: It is necessary to account for Lorentz contraction at high momentum

352: transfer in order to obtain reasonable interior densities without a

353: cusp at the origin.

354: Using recent recoil polarization data for the proton, 

355: we find that the charge density is significantly broader than the 

356: magnetization density.

357: We also find that the neutron and proton magnetization densities

358: are similar.

359: Although the available data are not yet as precise, the neutron charge

360: density is also obtained with useful precision.

361: A more detailed discussion of the model dependence of fitted densities

362: will appear elsewhere.

363: 

364: \acknowledgements

365: We thank O. Gayou for providing a table of $G_{Ep}$ data and X. Ji and

366: C. Perdrisat for useful discussions.

367: The support of the U.S. National Science Foundation under grant PHY-9971819 

368: is gratefully acknowledged.  

369: 

370: %\bibliographystyle{/group/enp/user/jjkelly/LATEX/prsty}

371: %\bibliography{/group/enp/user/jjkelly/LATEX/references}

372: \input{prl.bbl}

373: 

374: %%%%%%%%% Figures

375: %\newpage

376: 

377: \newpage

378: \widetext

379: \begin{figure}[ht]

380: \centerline{ \strut\psfig{file=fig1.ps,width=5.0in,angle=90} }

381: \caption{The bands show Fourier-Bessel fits to selected data for

382: nucleon electromagnetic form factors.

383: For $G_{En}$ the solid line shows the Galster model.}

384: \label{fig:ff}

385: \end{figure}

386: 

387: \narrowtext

388: \vfill

389: \eject

390: 

391: \begin{figure}[htbp]

392: \centerline{ \strut\psfig{file=fig2.ps,width=3.0in,angle=90} }

393: \caption{Comparison between fitted charge ($\rho_{ch}$) and 

394: magnetization ($\rho_m$) densities for the proton.  

395: Both densities are normalized to $\int dr \; r^2 \rho(r) = 1$.}

396: \label{fig:proton}

397: \end{figure}

398: 

399: \begin{figure}[hbtp]

400: \centerline{ \strut\psfig{file=fig3.ps,width=3.0in} }

401: \caption{Charge ($\rho_{ch}$) and magnetization ($\rho_m$) densities 

402: for the neutron.}

403: \label{fig:neutron}

404: \end{figure}

405: 

406: \end{multicols}

407: \end{document}

408: 

409: %%%%%%%%%%%% old stuff %%%%%%%%%%%

410: 

411: The Sachs form factors $G_E$ and $G_M$ are determined by the charge and 

412: magnetization distributions within nucleons and have been measured by

413: numerous experiments on elastic electron scattering from the proton or 

414: quasielastic scattering from the neutron in deuterium or polarized $^3$He.

415: Early experiments with modest $Q^2$ suggested that 

416: \begin{displaymath}

417: G_{Ep} \approx \frac{G_{Mp}}{\mu_p} \approx \frac{G_{Mn}}{\mu_n}

418: \approx G_D

419: \end{displaymath}

420: where $G_D(Q^2) = (1 + Q^2/\Lambda^2)^{-2}$ with 

421: $\Lambda^2 = 0.71$ (GeV/$c$)$^2$ is known as the dipole form factor.

422: Data for $G_{Mp}$ and $G_{Mn}$ with $Q^2>1$ (GeV/$c$)$^2$ show significant

423: departures from the simple dipole parametrization, 

424: but the extraction of $G_{Ep}$ from cross section data becomes increasingly

425: difficult as $Q^2$ increases.  

426: Recent data using the recoil polarization technique 

427: \cite{MKJones00,Gayou02}

428: have shown a dramatic,

429: almost linear, decrease in $G_{Ep}/G_{Mp}$ for $Q^2 > 1$ (GeV/$c$)$^2$.

430: It was suggested that those results demonstrate that the proton charge is

431: distributed over a larger volume than its magnetization, but radial densities

432: were not obtained.  

433: In this paper we use a Fourier-Bessel analysis, 

434: together with a relativistic relationship between form factors and densities,

435: to determine the nucleon charge and magnetization densities.

436: 

437: Let $\rho_{ch}(r)$ and $\rho_{m}(r)$ represent spherical intrinsic charge

438: and magnetization densities.

439: The vector magnetization density is then expressed as 

440: $\vec{\mu}(r) = \mu \rho_m(r) \vec{\sigma}$ where $\mu$ is the 

441: magnetic moment and $\vec{\sigma}$ is the Pauli spin vector.

442: It is convenient to normalize these densities according to

443: \begin{mathletters}

444: \begin{eqnarray}

445: \int dr \; r^2 \rho_{ch} &=& Z \\

446: \int dr \; r^2 \rho_{m} &=& 1

447: \end{eqnarray}

448: \end{mathletters}

449: where $Z=0,1$ is the nucleon charge.

450: Fourier-Bessel transforms of the intrinsic densities are defined by

451: \begin{equation}

452: \tilde{\rho}(k^2) = \int dr \; r^2 j_0(kr) \rho(r)

453: \end{equation}

454: where $k^2$ is the square of the spatial frequency (or wave number).

455: 

456: The interpretation of the Sachs form factors appears simplest in the 

457: Breit frame for which the energy transfer vanishes.

458: In this frame the nucleon approaches with initial momentum

459: $-\vec{q}_{B}/2$, receives three-momentum transfer $q_B$, 

460: and leaves with final momentum $\vec{q}_B/2$.

461: The Breit frame momentum is given by

462: $q_B^2 = Q^2 = q^2 / (1+\tau)$ where $(\omega,\vec{q})$ 

463: is the momentum transfer in the laboratory, 

464: $Q^2 = q^2 - \omega^2$ is the spacelike invariant four-momentum 

465: transfer, $\tau=Q^2/4m^2$, and $m$ is the nucleon mass.

466: The Sachs form factors are then determined by charge and magnetization

467: transition form factors between states with opposite momentum 

468: \begin{mathletters}

469: \begin{eqnarray}

470: G_{E}(q_B^2) &=& \tilde{\rho}_{B,ch}(q_B^2) \\

471: G_{M}(q_B^2) &=& \mu \tilde{\rho}_{B,m}(q_B^2)

472: \end{eqnarray}

473: \end{mathletters}

474: that resemble Fourier transforms of spatial densities.

475: However, there exists no rigorous model-independent relationship between

476: these transition form factors and the static charge and magnetization 

477: densities in the nucleon ground state with identical initial and final

478: states.

479: It is difficult to construct such a relationship because the boost operator

480: for a composite system depends upon the interactions among its constituents.

481: The first attempt to relate elastic form factors to ground-state densities 

482: was made by Licht and Pagnamenta \cite{Licht70b} using a cluster model and a

483: kinematic boost that neglects interactions.

484: The transition form factors were then evaluated using the impulse 

485: approximation and neglecting relative motion.

486: Ji \cite{Ji91} made a more rigorous analysis using a relativistic Skyrmion 

487: model based upon a Lorentz invariant Lagrangian density for which the

488: classical soliton solution can be evaluated in any frame.

489: Quantum fluctuations were then evaluated after the boost.

490: Although an approximation is still required to evaluate the transition form

491: factors, it was argued that this approximation is best in the Breit frame.

492: The final results offer simple relationships between Sachs form factors

493: and static densities that take the form

494: \begin{mathletters}

495: \label{eq:boost}

496: \begin{eqnarray}

497: \tilde{\rho}_{ch}(k^2) &=&  G_{E}(Q^2)\\

498: \tilde{\rho}_{m}(k^2) &=& G_{M}(Q^2) (1+\tau) 

499: \end{eqnarray}

500: \end{mathletters}

501: where the internal spatial frequency $k$ is related to the invariant

502: momentum transfer by

503: \begin{equation}

504: \label{eq:k}

505: k^2 = \frac{Q^2}{1+\tau}

506: \end{equation}

507: 

508: The most important relativistic effect is the Lorentz contraction of  

509: spatial distributions in the Breit frame and the corresponding increase

510: of spatial frequency represented by the 

511: factor of $(1+\tau)$ in Eq. (\ref{eq:k}).

512: A measurement with Breit-frame momentum transfer $q_B = Q$ probes a 

513: reduced spatial frequency $k$ in the rest frame.

514: The Sachs form factor for a large invariant momentum transfer $Q^2$ is 

515: determined by a much smaller spatial frequency $k^2=Q^2/(1+\tau)$ and 

516: thus declines much less rapidly with respect to $Q^2$ than the Fourier 

517: transform of the density declines with respect to $k^2$.

518: In fact, the accessible spatial frequency is limited to $k \leq 2m$ 

519: such that the Sachs form factors for large $Q^2$ are determined by the 

520: Fourier transform of intrinsic densities in the immediate vicinity of the 

521: limiting frequency $k_m=2m$, which is related to the nucleon Compton 

522: wavelength.

523: The difference between the multiplicative factors for $\rho_{ch}$ and

524: $\rho_{m}$ arises from the Lorentz transformation properties of scalar 

525: and vector fields \cite{Ji91}.

526: Hence, the corresponding densities would differ even if the Sachs

527: form factors were identical.

528: 

529: 

530: Having established that it is possible to fit physically reasonable 

531: charge and magnetization densities to elastic form factor data spanning 

532: a large range of $Q^2$, 

533: it is necessary to return to the question of the uniqueness of 

534: Eq. (\ref{eq:boost}).

535: The most important relativistic feature of that relationship is the

536: identification of the spatial frequency $k^2$ with $Q^2/(1+\tau)$ due

537: to Lorentz contraction of distributions in the Breit frame and is 

538: common to most models.

539: The relationships obtained by Licht and Pagnamenta \cite{Licht70b}

540: differ from those of Ji \cite{Ji91} by application of a factor of

541: $(1+\tau)$ to $G_E$ as well as to $G_M$.

542: Alternatively, some constituent quark model calculations apply factors of

543: $(1+\tau)^{1/2}$ to both form factors.

544: Differences between these prescriptions alter the shape of the 

545: fitted density in a smooth fashion, 

546: but do not affect the qualitative relationship between

547: the quality and range of experimental data and the precision of the fitted

548: density, as represented by its error band.

549: The empirical parametrization proposed by Bosted \cite{Bosted95}

550: \begin{equation}

551: G \propto (1 + a_1 Q + a_2 Q^2 + a_3 Q^3 + a_4 Q^4)^{-1}

552: \end{equation}

553: also fits the data for large $Q^2$ well and is consistent with pQCD, 

554: but its odd powers of $Q$ are incompatible with the interpretation of

555: the form factor as the Fourier transform of a radial density and

556: with the moment expansion for small $Q^2$.

557: Conversely, although we cannot claim our proposed relationship between

558: form factors and densities is model independent, it does provide a

559: physically appealing parametrization of the form factor data and realistic

560: error bands in both spatial and momentum representations.

561: Therefore, even if the identification of the extracted densities with the

562: static densities is discounted, these densities do provide a useful

563: parametrization of the form factors nonetheless.  

564: 

565: 

566: 

567: 

568: applied the Fourier-Bessel expansion and an {\it ansatz}

569: for the relationship between densities in the nucleon rest frame to transition

570: form factors in the Breit frame to extract charge and magnetization densities 

571: with realistic estimates of their uncertainties from data for Sachs form 

572: factors.  

573: Three of the four densities are determined very accurately, but more precise

574: data at higher $Q^2$ will be needed to achieve comparable precision for the

575: neutron charge density.

576: Several new experiments using recoil or target polarization will soon provide 

577: more precise $G_{En}$ data that should greatly improve the precision of the 

578: neutron charge density. 

579: