1: \documentclass{elsart}

2: \usepackage{amssymb}

3: \usepackage{amsmath}

4: \usepackage{graphicx}

5: \newcommand{\bmath}[1]{\mbox{\boldmath ${#1}$}}

6: \newcommand{\dd}{\mbox{\rm d}}

7: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

8: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

9: \begin{document}

10: \begin{frontmatter}

11: 

12: \title{\large{\bf Model-independent analysis of the neutron-proton

13: final-state interaction region in the $\bmath{pp\to pn\,\pi^+}$ reaction}}

14: 

15: \author[Dubna,Almaty]

16: {Yuri N.\ Uzikov}~\thanks{E-mail:~uzikov@nusun.jinr.dubna.su}

17: and\ \ \

18: \author[UCL]{Colin Wilkin}~\thanks{Corresponding author:

19: C.~Wilkin, Physics and Astronomy Department,

20: UCL, Gower Street, London WC1E 6BT, UK.\ \ \ 

21: E-mail:~cw@hep.ucl.ac.uk}

22: 

23: \address[Dubna]{JINR, LNP, Dubna, 141980 Russia}

24: \address[Almaty]{Kazakh State University, Almaty, 480121 Kazakhstan}

25: \address[UCL]{University College London, London, WC1E 6BT, UK}

26: 

27: \begin{abstract}

28: Experimental data on the $pp\to pn\,\pi^+$ reaction measured in an

29: exclusive two-arm experiment at 800~MeV show a narrow peak arising

30: from the strong proton-neutron final-state interaction. It was claimed,

31: within the framework of a certain model, that this peak contained up

32: to a 25\% spin-singlet final state contribution. By comparing the data 

33: with those of $pp\to d\,\pi^+$ in a largely model-independent way, 

34: it is here demonstrated that at all the angles measured the whole of 

35: the peak could be explained as being due to spin-triplet final states, 

36: with the spin-singlet being at most a few percent. Good qualitative 

37: agreement with the measured proton analysing power is also found 

38: within this approach.

39: \end{abstract}

40: 

41: \begin{keyword}

42: pion production, final state interactions

43: \begin{PACS}

44: 13.75.Cs, 25.40.Qa

45: \end{PACS}

46: \end{keyword}

47: \end{frontmatter} 

48: 

49: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

50: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

51: \newpage

52: 

53: Inclusive data on the $pp\to \pi^+X$ reaction in the 400~MeV to

54: 1~GeV range generally show two structures. There is firstly a broad peak

55: corresponding to the quasi-free production of the $\Delta$-isobar. In 

56: addition, there is an enhancement near the edge of phase space arising

57: from the strong neutron-proton final-state interaction (fsi) in either the 

58: spin-triplet or singlet $S$-wave which comes when the $np$ excitation

59: energy, $E_{np}$, is only a few MeV. The details of the enhancement 

60: region are hard to investigate in a single-arm experiment, where 

61: only the $\pi^+$ is measured, because of

62: contamination from the much larger two-body $pp\to d\,\pi^+$ 

63: reaction~\cite{Falk}. The simplest way to overcome this background

64: is by measuring in coincidence the final proton and pion in

65: the exclusive $pp\to pn\,\pi^+$ reaction. The most complete examples

66: of such an experiment were carried out at LAMPF, where both the

67: five-fold differential cross section~\cite{HG} and the proton

68: analysing power~\cite{Hancock} were studied at a proton beam

69: energy of $T_p=800$~MeV.

70: 

71: LAMPF data taken at one pair of proton/pion angles are shown in Fig.~1

72: as a function of the detected proton momentum~\cite{HG}. The large peak

73: on the right is associated with protons and pions formed in the 

74: $pp\to \Delta^{++}n/\Delta^{+}p$ reactions and its magnitude and width

75: can be explained in different versions of

76: one-meson-exchange models~\cite{HG,Yuri}.

77: At the maximum of the smaller peak, $E_{np}$ is about 1~MeV and the

78: cross section is strongly influenced by the $np$ fsi. The form of this

79: peak is normally parameterised by Watson final-state interaction

80: factors~\cite{Watson,GW} which take into account the nearby poles in the

81: scattering amplitudes due to the deuteron bound state in the

82: spin-triplet case and the anti-bound state for the spin-singlet.

83: Since the latter is closer to $E_{np}=0$, the singlet fsi peak

84: is expected to be narrower than that of the triplet. Though this difference

85: in shape can, in principle, be used to extract the relative amount of the

86: $np$ spin-triplet and singlet in the final-state peak, the limited number

87: of points in the peak and the modest resolution in $E_{np}$ makes this

88: impractical for the LAMPF data.

89: 

90: In analysing their experiment, the authors of Ref.~\cite{HG} made the

91: \textit{ad hoc} assumption that the one-meson-exchange model for

92: quasi-free $\Delta$ production was valid for $E_{np}$ above 10~MeV

93: and that, below this excitation energy, the prediction could be smoothly

94: joined onto Watson fsi factors. The relative spin-singlet strength depends, of

95: course, upon the kinematics and their analysis suggested that the 

96: fraction was about 15-25\% of the total for the different angle pairs 

97: measured. This is close to a statistical mixture of 25\% and

98: is in complete contrast to measurements at lower energies, where the

99: singlet fraction is typically 10\%~\cite{FW2}. Alternative

100: model-dependent analyses of the LAMPF data do lead to smaller singlet

101: production, but they depend upon other assumptions made~\cite{Dubach}.

102: In view of these differing conclusions, it is worthwhile to seek a

103: different way of deducing the spin-singlet contribution from these data.

104: 

105: When the square of the low energy $np$ triplet $S$-state scattering

106: wave function at energy $E_{np}= k^2/m_N$, where $m_N$ is twice the

107: $n$-$p$ reduced mass, is analytically continued to negative energy,

108: it manifests the deuteron pole at $k^2=-\alpha_t^2$, where

109: $\alpha_t=0.232$~fm$^{-1}$. It has recently been shown~\cite{FW1}

110: that the relative

111: normalisation of the scattering, $\Psi_k(r)$, and bound-state,

112: $\Psi_{\alpha_t}(r)$, wave function depends purely upon the deuteron

113: binding energy and is independent of the $np$ potential~\cite{FW1}.

114: Using real boundary conditions, it follows that at short distances

115: \begin{equation}

116: \label{FW}

117: \left[\Psi_k(r)\right]^2 \approx

118: \frac{2\pi}{\alpha_t(k^2+\alpha_t^2)}\,

119: \left[\Psi_{\alpha_t}(r)\right]^2\:.

120: \end{equation}

121: Apart from questions associated with the $D$-state contributions which

122: are small at low $k^2$, this relation becomes exact as $k^2\to -\alpha_t^2$.

123: In the scattering region where $E_{np} \geq 0$, it allow one to

124: estimate the spin-triplet amplitude for $pp\to pn\,\pi^+$

125: in terms of that for $pp\to d\,\pi^+$. Thus, for any value of the

126: initial ($\sigma_i$) and final spin-triplet projection $\lambda$, the

127: matrix elements are related by

128: \begin{equation}

129: \label{ME}

130: M_{\sigma_1\sigma_2}^{\lambda}(pp\to\{np\}_t\,\pi^+)

131: \approx f(k^2)\,M_{\sigma_1\sigma_2}^{\lambda}(pp\to d\,\pi^+)\:,

132: \end{equation}

133: where

134: \begin{equation}

135: \label{fsi}

136: f^2(k^2) = \frac{2\pi\,m_N}{\alpha_t(k^2+\alpha_t^2)}\:\cdot

137: \end{equation}

138: We are here using a normalisation where the unpolarised

139: $pp\to d\,\pi^+$ cross section is given by

140: \begin{equation}

141: \label{pid}

142: \frac{\dd\sigma}{\dd\Omega^*}= \frac{1}{64\pi^2s_{pp}}\,

143: \frac{q_{\pi d}^*}{q_{p}^*}\,\frac{1}{4}\,\sum_{\sigma_1\sigma_2\lambda}

144: \mid M_{\sigma_1\sigma_2}^{\lambda}(pp\to d\,\pi^+)\mid^2\:,

145: \end{equation}

146: $s_{pp}$ is the square of the centre-of-mass (cm) energy, and $q_p^*$ and

147: $q_{\pi d}^*$ are the cm momenta in the initial and final states

148: respectively. At low $E_{np}$ the approximation of eq.~(\ref{ME})

149: reproduces very well

150: the results of single-arm experiments where only the pion is

151: detected~\cite{FW2}.

152: 

153: We now extend this approach to treat two-arm experiments. Starting from

154: eq.~(\ref{ME}), the triplet contribution to the laboratory five-fold 

155: differential cross section for the detection of a pion at an angle 

156: $\theta_{\pi}$ and a proton at an angle $\theta_p$ becomes

157: \begin{equation}

158: \label{yuri}

159: \frac{\dd^5\sigma_t(pp\to pn\pi^+)}{\dd p_p\,\dd\Omega_p\,\dd\Omega_\pi}

160: = \frac{1}{16 \pi^3}\, s_{pp} \frac{q_{p}^*}{q_{\pi d}^*}\,\Phi \,f^2(k^2)\,

161: \frac{\dd\sigma}{\dd\Omega_{\pi}^*}(pp\to d\pi^+)\:,

162: \end{equation}

163: where the phase-space factor is

164: \begin{equation}

165: \label{PS}

166: \Phi=\frac{p_p^2 p_\pi^3}{ p_0\, m_p\, E_p\,

167: |p_\pi^2\, E_n -{\bf p}_\pi\cdot {\bf p}_n E_\pi|}\:\cdot

168: \end{equation}

169: Here $p_0$ is the beam momentum, $E_i$ and $p_i$ ($i=p,n,\pi$) are the

170: laboratory energy and momentum of the $i$-th particle in the final state.

171: 

172: The pion production angles in the laboratory ($\theta_\pi$) and  cm

173: ($\theta_{\pi}^*$) systems are related by

174: \begin{equation}

175: \label{theta}

176: E_0\,E_\pi - p_0\,p_\pi \cos{\theta_\pi}=

177: \varepsilon_0\,\varepsilon_\pi-q_p^*\, q_\pi^*

178: \cos{\theta_\pi^*}\:,

179: \end{equation}

180: where $\varepsilon_\pi (\varepsilon_0)$  and $q_\pi^*$  are the energy 

181: of the pion (incident proton) and 3-momentum of the

182: pion in the overall cm system and $E_0$ is the total laboratory energy of the

183: incident  proton. Due to the difference between the effective mass of 

184: the final $pn$ state and that of the deuteron, 

185: $q_{\pi}^*\not =q_{\pi d}^*$. However, this effect is quite small at 

186: low $E_{np}$.

187: 

188: The results of eq.~(\ref{yuri}), which should be valid at low relative

189: energies $E_{np}$, do not depend upon the details of the pion production

190: dynamics and automatically include the final state interaction in the

191: triplet $pn$ system.

192: 

193: The input $pp\to d\,\pi^+$ cross sections are taken from the SAID

194: SP96 parameterisation~\cite{SAID} and this procedure should involve 

195: errors that are smaller than those of an individual experiment.

196: The predictions of our approach for

197: the triplet $pp\to pn\,\pi^+$ cross section are shown at one

198: angle pair in Fig.~1 together with an evaluation of quasi-free

199: $\Delta$-production~\cite{Yuri}. The values of the $np$ excitation energies

200: are indicated and from this it is seen that when $E_{np}$ is below

201: about 5~MeV the magnitude and shape of the fsi peak are both well

202: reproduced. For comparison the fsi parameterisation of Goldberger 

203: and Watson~\cite{GW}, multiplied by phase space and normalised to the 

204: peak value, is also illustrated.

205: 

206: Results at different angles in the fsi region are shown in Fig.~2.

207: Larger deviations from the predictions are to be found when the minimum

208: value of the excitation energy, $E_{np}^{min}$, is increased. This is the

209: case at $(\theta_p,\theta_{\pi}) = (25^{\circ},40^{\circ})$ and

210: $(30^{\circ},28^{\circ})$ where $E_{np}^{min}$ are respectively

211: 13 and 28~MeV. Even in this last case, the changes induced by the

212: modified kinematics are quite small, as can be judged from the dot-dash

213: curve shown in the figure. A potentially more serious effect arises from

214: averaging the estimates over the experimental angular acceptance. As

215: shown in Fig.~2, this tends to reduce a little the predictions when 

216: $E_{np}^{min}\leq 1$~MeV but can increase them otherwise. After

217: smearing, there seems to be some underprediction in Fig.~2b, though it 

218: should be stressed that these data were taken from the polarisation

219: experiment~\cite{Hancock} where the consistency with the earlier 

220: unpolarised cross section run was found only on the 15\% level. 

221: 

222: Since the predictions of the F\"aldt-Wilkin extrapolation

223: theorem~\cite{FW1} reproduce most of the magnitude and angular dependence of

224: the cross section leading to the fsi peaks, this is strong evidence

225: that the vast bulk of the LAMPF data corresponds to $np$ triplet final

226: states. Extra confirmation of this interpretation is found from the

227: proton analysing power which was measured at two pairs of angles in

228: the fsi region~\cite{Hancock}. In the fsi peak we expect $A_y$ to be

229: essentially constant at the value corresponding to that of

230: $pp\to d\,\pi^+$ at the appropriate pion angle. These values are

231: shown with the experimental data in Fig.~3. Though at

232: $(14.5^{\circ},21^{\circ})$ the data seem to fall a little above

233: the prediction, deviations of this size are not unknown in polarisation

234: measurements.

235: 

236: Although we have reproduced well the LAMPF data at low $E_{np}$ with

237: just triplet terms,

238: it may be helpful to try to estimate an upper bound on the possible

239: spin-singlet contribution from the areas under the peaks. 

240: The overall systematic error due to beam

241: normalisation and detector efficiency in the measurement was estimated

242: to be about 7\%~\cite{HG}, to which must be compounded some error

243: coming from the $pp\to d\,\pi^+$ input~\cite{SAID}. The error arising 

244: from using the extrapolation theorem in the scattering domain is likely to be

245: rather smaller than this, with variations in the wave functions at short

246: distances being on the 1--2\% level for $E_{np}<10$~MeV~\cite{FW3}.

247: Further work is needed to include the $D$-state effects more consistently,

248: though it has been shown that the extrapolation theorem is valid in

249: this coupled-channel case~\cite{Smirnov}. Under the present conditions,

250: we would estimate the accuracy of eq.~(\ref{yuri}) to be better than 5\%

251: for $E_{np}\leq 3$~MeV~\cite{smuzikov}. We therefore conclude that the 

252: singlet contribution to the unpolarised LAMPF data at low $E_{np}$ 

253: in Fig.~1 is below about 10\%. This upper bound is about a factor of two

254: less than the average quoted in the experimental paper~\cite{HG},

255: though it must be stressed that this involved considerable model

256: dependence, including the choice of a 10 MeV matching point.

257: 

258: There is a 15\% normalisation uncertainty between the first

259: measurements of the unpolarised cross section at LAMPF~\cite{HG} and their

260: later experiment, where the primary purpose was the determination of the

261: analysing power~\cite{Hancock}. If, nevertheless, we take seriously

262: the disagreement with the latter data shown in fig.~2b, it is possible that

263: this is due to a spin-singlet contribution. Assuming that

264: the singlet cross section to be of the form

265: \begin{equation}

266: \label{singlet}

267: \frac{\dd^5\sigma_s(pp\to pn\pi^+)}{\dd p_p\,\dd\Omega_p\,\dd\Omega_\pi}

268: =\zeta\,\frac{(k^2+\alpha_t^{\,2})}{(k^2+\alpha_s^{\,2})}\times

269: \frac{\dd^5\sigma_t(pp\to pn\pi^+)}{\dd p_p\,\dd\Omega_p\,\dd\Omega_\pi}\:,

270: \end{equation}

271: where $\alpha_s=-0.04$~fm$^{-1}$ then, after averaging over

272: acceptance, a value of $\zeta=0.03$ restores the agreement with the

273: data. This would correspond to a singlet fraction integrated over the 

274: range $0\leq E_{np}\leq 3$~MeV of about 10\%. However, such a fraction

275: would make the agreement worse in fig.~2c, where the data were taken

276: at a rather similar cm angle but in the other hemisphere. 

277: 

278: To improve the sensitivity to the singlet/triplet ratio using the

279: extrapolation theorem, the ratio of the $pp\to pn\,\pi^+$ and

280: $pp\to d\,\pi^+$ cross sections has to be established better by

281: measuring both reactions in the same experiment. Data with better

282: resolution on $E_{np}$ would also allow one to investigate the

283: singlet/triplet ratio from the shape of the fsi peak. A new measurement

284: of the $pp\to pn\,\pi^+$ reaction at the ANKE spectrometer of the

285: proton synchrotron COSY-J\"ulich, where both protons

286: and pions were detected near the forward direction at $T_p=492$~MeV,

287: involved both these improvements and can therefore put more stringent

288: limits on the spin ratio~\cite{VK}.

289: 

290: In conclusion, we have generalised the use of the extrapolation theorem

291: linking $np$ scattering and bound-state wave functions to describe

292: five-fold differential cross sections. In so doing, we have

293: shown that the LAMPF $pp\to pn\,\pi^+$ cross section data at 800~MeV

294: are consistent with there being no final-state spin-singlet

295: contribution. This smallness should not come as a

296: complete surprise. If one assumes that pion 

297: production passes through an intermediate $\Delta$ 

298: isobar~\cite{HG,Yuri,Dubach}, then the $S$-wave $\Delta\,N$ 

299: intermediate state cannot lead to singlet $np$ final states. These

300: must be produced through higher partial waves or from non-resonant 

301: pion production.

302: 

303: This work was initiated within the framework of the ANKE collaboration

304: at COSY and the authors would like to thank all their colleagues, especially

305: V.~Abaev, V.I.~Komarov, V.~Koptev, and V.~Kurbatov, for many fruitful 

306: discussions. Financial support from the Forschungszentrum J\"ulich and

307: BMBF (WTZ grant KAZ 99/001) and the generous hospitality of

308: Institut f\"ur Kernphysik, where the work was carried out, are gratefully

309: acknowledged.

310: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

311: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

312: \newpage

313: \begin{thebibliography}{99}

314: %

315: \bibitem{Falk} See for example

316: W.R.~Falk \textit{et al.}, Phys.\ Rev.\ C~\textbf{32}

317: (1985) 1972; R.G.~Pleydon \textit{et al.}, Phys.\ Rev.\ C~\textbf{59}

318: (1999) 3208.

319: %

320: \bibitem{HG} J.~Hudomalj-Gabitzch \textit{et al.}, Phys.\ Rev.\

321: C~\textbf{18} (1978) 2666.

322: %

323: \bibitem{Hancock} A.D.~Hancock \textit{et al.}, Phys.\ Rev.\

324: C~\textbf{27} (1983) 2742.

325: %

326: \bibitem{Yuri} O.~Imambekov and Yu.N.~Uzikov, Sov.\ J.\ Nucl.\ Phys.\

327: \textbf{47} (1988) 695.

328: %

329: \bibitem{Watson} K.M.~Watson, Phys.\ Rev.\ \textbf{88} (1952) 1163.

330: %

331: \bibitem{GW}

332: M.L.~Goldberger and K.M.~Watson, \textit{Collision Theory}

333: (John Wiley, N.Y.) 1964. 

334: %

335: \bibitem{FW2} A.~Boudard, G.~F\"aldt and C.~Wilkin, 

336: Phys.\ Lett.\ {\bf B389} (1996) 440.

337: %

338: \bibitem{Dubach} J.~Dubach, W.M.~Kloet, and R.R.~Silbar, Phys.\ Rev.\

339: C~\textbf{33} (1986) 373.

340: %

341: \bibitem{FW1} G.~F\"aldt and C.~Wilkin, Physica Scripta \textbf{56} (1997)

342: 566.

343: %

344: \bibitem{SAID} R.A. Arndt {\it et al.}, Phys.\ Rev.\ C~{\bf 48} (1993) 1926;\\

345:  \verb+http://said.phys.vt.edu/said_branch.html/+

346: %

347: \bibitem{FW3} G.~F\"aldt and C.~Wilkin, Phys.\ Rev.\ C~{\bf 56} (1997) 2067.

348: %

349: \bibitem{Smirnov} A.V.~Smirnov, Communication 34 of the Lebedev

350: Physical Institute, Moscow, Russia, 1997.

351: %

352: \bibitem{smuzikov} A.V.~Smirnov and  Yu.N.~Uzikov,

353: Yad.\ Fiz.\ \textbf{61} (1998) 421; Phys.\ At.\ Nuc.\ \textbf{61}

354: (1998) 361.

355: %

356: \bibitem{VK} V.~Kurbatov (private communication, 2001) and

357: \textit{publication in preparation}.

358: 

359: \end{thebibliography}

360: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

361: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

362: \input epsf

363: \begin{figure}[t]

364: \begin{center}

365: \mbox{\epsfxsize=4in \epsfbox{fig1.eps}}

366: \caption{Differential cross section for the $pp\to pn\,\pi^+$

367: reaction at 800~MeV at fixed proton and pion laboratory angles of

368: ($\theta_p,\theta_{\pi})=(15^{\circ},20.8^{\circ}$) as a function

369: of the measured proton momentum~\protect\cite{HG}.

370: The $pp\to \Delta N$ peak on the right can be described

371: in a one-meson-exchange model (dashed curve)~\protect\cite{Yuri}.

372: The peak on the left is a reflection of the strong $np$ fsi,

373: which can be calculated (solid curve) from the $pp\to d\,\pi^+$

374: cross section, as discussed in the text. A scale is given showing the 

375: $np$ excitation energy at the geometry corresponding to the centres 

376: of the counters. For comparison, the triplet fsi factor in the

377: Goldberger and Watson form~\protect\cite{GW}, multiplied by the

378: phase space factor of eq.~(\protect\ref{PS}) and normalised to the 

379: peak, is shown as the dotted curve.}

380: \end{center}

381: \label{fig1}

382: \end{figure}

383: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

384: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

385: \begin{figure}[t]

386: \begin{center}

387: \vspace*{-4cm}

388: \mbox{\epsfxsize=5.5in \epsfbox{fig2.eps}}

389: \caption{Differential cross section for the $pp\to pn\,\pi^+$

390: reaction at 800~MeV in the neighbourhood of the $np$ fsi region for

391: different angular positions, 

392: a)~$(\theta_p,\theta_{\pi})=(15^{\circ},20.8^{\circ})$, 

393: b)~$(14.5^{\circ},42^{\circ})$,

394: c)~$(14.5^{\circ},21^{\circ})$,

395: d)~$(25^{\circ},40^{\circ})$,

396: e)~$(20^{\circ},22^{\circ})$,

397: f)~$(30^{\circ},28^{\circ})$.

398: The $np$ excitation energies at the counter centres are indicated;

399: the number with the decimal corresponds to the minimum value of

400: $E_{np}$. The experimental data are taken from Refs.~\protect\cite{HG} 

401: (circles) and \protect\cite{Hancock} (triangles) and it should be

402: noted that these may differ in normalisation by 15\%. The data

403: are compared to the spin-triplet final-state predictions of

404: eq.~(\protect\ref{yuri}) without acceptance corrections (solid curve)

405: and with (broken curve). An attempt to resolve the apparent

406: discrepancy in b), through the introduction of a singlet contribution

407: by eq.~(\protect\ref{singlet}) with $\zeta=0.03$ makes the situation

408: worse in c) (dotted curves). The dot-dash curve in f) was

409: obtained using the deuteron mass rather than the invariant mass of

410: the $np$ system when evaluating the pion production angle

411: $\theta_{\pi}^*$ in eq.~(\ref{theta}).}

412: \end{center}

413: \label{fig2}

414: \end{figure}

415: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

416: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

417: %\input epsf

418: \begin{figure}[ht]

419: \begin{center}

420: \mbox{\epsfxsize=4in \epsfbox{fig3.eps}}

421: \caption{Proton analysing power for the $\vec{p}p\to pn\,\pi^+$

422: reaction at 800~MeV in the neighbourhood of the $np$ fsi region.

423: The experimental data of Refs.~\protect\cite{Hancock} are compared

424: to the SAID predictions of $A_y$ (horizontal lines) for the

425: $\vec{p}p\to d\,\pi^+$ reaction~\protect\cite{SAID}. It should be noted

426: that SAID uses the opposite convention for the sign of $A_y$.}

427: \end{center}

428: \label{fig3}

429: \end{figure}

430: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

431: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

432: \end{document}

433: