1: %& This is the latest version 18.05.04
2: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
3: \begin{filecontents}{leer.eps}
4: %!PS-Adobe-2.0 EPSF-2.0
5: %%CreationDate: Mon Jul 13 16:51:17 1992
6: %%DocumentFonts: (atend)
7: %%Pages: 0 1
8: %%BoundingBox: 72 31 601 342
9: %%EndComments
10:
11: gsave
12: 72 31 moveto
13: 72 342 lineto
14: 601 342 lineto
15: 601 31 lineto
16: 72 31 lineto
17: showpage
18: grestore
19: %%Trailer
20: %%DocumentFonts: Helvetica
21: \end{filecontents}
22: %
23: %\documentclass[epj,referee]{svjour}
24: \documentclass[epj]{svjour}
25: % Remove option referee for final version
26: \usepackage{psfig}
27: \begin{document}
28: %\hspace{9.8 cm}FZJ--IKP(TH)--2003--14
29:
30: \title{The $\eta{-}^3$He scattering length revisited}
31: \author{A. Sibirtsev\inst{^1}, J. Haidenbauer\inst{^1},
32: C. Hanhart\inst{^1} \and J.A. Niskanen\inst{^2}}
33:
34: \institute{
35: Institut f\"ur Kernphysik, Forschungszentrum J\"ulich,
36: D-52425 J\"ulich, Germany \and
37: Department of Physical Sciences, PO Box 64, FIN-00014 University of
38: Helsinki, Finland }
39:
40: %
41: \date{Received: date / Revised version: date}
42:
43: \abstract{
44: The possible existence of $\eta$-mesic nuclei poses an interesting
45: and still open issue of research.
46: Since the occurence of such $\eta$-nucleus bound states is
47: reflected in the corresponding $\eta$-nucleus scattering length, we
48: critically review the present knowledge for the $\eta$$^3$He system.
49: Specifically, we scrutinize the available experimental information
50: for the reaction $p{+}d \to\, \eta {+} ^3{\rm He}$ which is commonly used to
51: extract the $\eta$$^3$He scattering length. We point out several striking
52: discrepancies between the various measurements. Subject to those
53: inconsistencies we deduce a value of
54: $a{=}|4.3{\pm}0.3|{+}i(0.5{\pm}0.5)$ fm.
55: %
56: \PACS{
57: {12.38.Bx} { } \and
58: {12.40.Nn} { } \and
59: {13.60.Le} { } \and
60: {14.40.Lb} { } \and
61: {14.65.Dw} { }
62: }}
63: \maketitle
64:
65: \section{Introduction}
66: The possible formation of $\eta$-nucleus quasibound states has been an
67: interesting topic for a long time. However, so far no such states have
68: been directly observed. It is also an open and heavily debated question
69: what might be the lightest nuclei
70: for which such a bound state can occur.
71: For instance, investigations~\cite{Haider1,Liu,Chiang} based on
72: optical models indicate carbon as the lower limit for
73: nuclei able to bind an $\eta$ meson. Most recently~\cite{Haider2} this
74: limit was lowered to the $^4$He nucleus. In contrast, even formation of a
75: bound $\eta^3{\rm He}$ system is supported by other and different
76: model calculations~\cite{Wycech,Belyaev1,Belyaev2,Rakityansky,Fix1}.
77: %
78: To clarify the situation experimental studies of this
79: system are proposed at GSI \cite{GSI1,GSI2} and COSY \cite{COSY1,COSY2,COSY3}.
80: It is obvious that such experiments are very delicate and their design requires
81: good estimates for the relevant binding energies and widths of
82: $\eta$-mesic nuclei. Such estimates would dictate, for
83: instance, the necessary resolution of the detector and the
84: required beam luminosity.
85:
86: In this context very light nuclei are particularly interesting
87: because such systems are accessible to a microscopic treatment
88: whereas for heavier systems approximations have to be introduced
89: whose effects are difficult to quantify and, accordingly, might lead
90: to large uncertainties in the achieved results.
91: In particular the $\eta^3{\rm He}$ system is very appealing
92: because it can be studied within the well-established
93: Alt-Grassberger-Sandhas~\cite{Grassberger} and
94: Faddeev -Yakubovsky~\cite{Faddeev} theories.
95: The only but still very crucial ambiguity here is caused by our poor
96: knowledge of the elementary $\eta$-nucleon ($\eta N$) interaction, which
97: obviously enters any
98: microscopic calculation as an input. A compilation of values for the
99: $\eta{N}$ scattering length, obtained from different $\eta{N}$ model
100: analyses, shows that its real part ranges from 0.20 to 1.05~fm, while the
101: imaginary part varies between 0.16 and 0.49~fm~\cite{Sibirtsev1}.
102: Since the elementary $\eta{N}$ interaction is not fixed, any $\eta$-nucleus
103: calculation \cite{Wycech,Belyaev1,Belyaev2,Rakityansky,Fix1} can only
104: provide a range of results for the $\eta$-nucleus scattering lengths
105: rather than a concrete prediction.
106:
107: Under these circumstances it seems to be more promising to investigate
108: a quantity closely related to the properties of the bound state, namely
109: the $\eta$-nucleus scattering length \cite{Newton}. It is well-known
110: that in case of bound states the (real part of the) scattering length
111: should be relatively large and negative. (We adopt here the sign
112: convention of Goldberger and Watson \cite{watson} common in meson physics.)
113: Thus, studies of the $\eta$-nucleus interaction near
114: threshold can be used to determine the $\eta$-nucleus scattering
115: length, and then, in turn, would permit conclusions on the
116: existence of such $\eta$-nucleus bound states. Information
117: on the $\eta$-nucleus interaction can be deduced from analysing
118: the energy dependence of $\eta$ production reactions such as
119: $pd{\to}\eta^3{\rm He}$, $dd{\to}\eta^4{\rm He}$, etc.
120: Certainly, the energy dependence of the production cross section
121: of those reactions itself is not sensitive to the sign of the real part
122: of the scattering length, but only to its magnitude, and therefore
123: cannot provide direct evidence for the existence of a bound state.
124: (See, however, Ref. \cite{mix} for a possible experiment to determine
125: the sign of the real part.)
126: But even a good quantitative knowledge of the magnitude of the scattering
127: length could already provide a strong hint for the existence of a
128: $\eta$-mesic bound state and, more importantly, it would allow concrete
129: estimations for the energy range that should be scanned in dedicated
130: experiments.
131:
132: In the present paper we
133: provide a systematic overview of the experimental
134: information available for the reaction $p{+}d{\to}^3{\rm He}{+}\eta$
135: \cite{Berger,Banaigs,Berthet,Mayer,Bilger,Betigeri,Kirchner}.
136: In particular, we critically compare the results of the various measurements,
137: which were partly performed for different kinematical conditions, in order
138: to investigate the consistency of the data sets.
139: Special emphasis will be put on the data near the reaction
140: threshold which are commonly used to extract information about the
141: $\eta ^3{\rm He}$ scattering length. We discuss also results for the
142: $\eta ^3{\rm He}$ scattering length that can be found in the literature.
143: %
144: The aim of the paper twofold. First we want to derive an new estimate for
145: the $\eta ^3{\rm He}$ scattering length taking into account all available
146: low-energy data on the reaction $p{+}d{\to}^3{\rm He}{+}\eta$
147: and, secondly, we want to specify which further measurements
148: are necessary in order to significantly improve the present situation.
149:
150: \section{Treatment of the final state interaction}
151:
152: If a production reaction is governed by a strong $s$-wave interaction in
153: the final state then, according to Watson and Migdal \cite{Watson,Migdal},
154: the energy dependence of the reaction amplitude
155: is basically determined by the
156: on-shell scattering amplitude of the final state, i.e. by
157: \begin{equation}
158: T(q) = \frac{1}{q \cot\delta -i q}\, .
159: \label{tfsi}
160: \end{equation}
161: where $q$ is the center-of-mass momentum of the strongly interacting particles
162: in the final state and $\delta$ is the corresponding ($s$ wave) phase shift.
163: %
164: Close to threshold the phase shift $\delta$ can be approximated
165: by the effective range expansion
166: \begin{equation}
167: q \cot\delta = \frac{1}{a}+ \frac{r_0}{2}\, q^2\, ,
168: \label{efran}
169: \end{equation}
170: where for the $\eta^3{\rm He}$ interaction, of course, both the scattering
171: length $a$ and the effective range $r_0$ are complex.
172:
173: \begin{figure}[b]
174: \vspace*{-2mm}
175: \hspace*{-0.5mm}\psfig{file=etan6a.ps,width=9.3cm,height=9.cm}
176: \vspace*{-6mm}
177: \caption{Spin averaged $p{+}d{\to}^3{\rm He}{+}\eta$ transition
178: amplitude for forward (solid symbols) and backward (open symbols)
179: $\eta$-meson production as a function of the final momentum $q$ in the
180: c.m. system (lower axis) or excess energy $\epsilon$ (upper axis).
181: The experiments are taken from Refs. \cite{Berger,Banaigs,Berthet}.
182: In some cases \cite{Banaigs,Bilger,Betigeri,Kirchner} extrapolated
183: results from a fit to the measured $\eta$-meson angular spectra are
184: shown, cf. text.
185: %
186: The solid lines are the fits of Ref.~\cite{Berger} to their
187: $p{+}d{\to}^3{\rm He}{+}\eta$ data by Eq.~(\ref{fit1}) for
188: $\vartheta{=}0^o$ and $\vartheta{=}180^o$.
189: The shaded area indicates results based on the correlation
190: Eq.~(\ref{wilk}) reported in Ref.~\cite{Wilkin1}
191: for the extreme limits given by $\Re a{=}0$ and $\Im a{=}0$.
192: Here $|f|^2$ was obtained by using Eq.~(\ref{fitn}).}
193: \label{etan6a}
194: \end{figure}
195:
196: With the above sign convention a commonly quoted necessary
197: condition for the existence of a quasibound state is that $\Re a \ {<} \ 0$.
198: However, for having a quasibound state there is an additional
199: requirement, namely that the energy $E$ corresponding to the zero in
200: the denominator of Eq. (\ref{tfsi}) fulfils the relation
201: $\Re E \ {<} \ 0$. This, in turn implies that
202: \begin{equation}
203: \Re [a^3\, (a^\ast - r_0^\ast)] > 0
204: \end{equation}
205: in the two lowest orders in $r_0/a$. In the absence of the
206: effective range term this reduces to the condition that
207: $|\Re a|{>}\Im a{>}0$, given, e.g., in Ref.~\cite{Haider2}.
208:
209: Neglecting terms of higher order
210: than $q^2$, the squared reaction amplitude, $|f|^2$, can be written as
211: \cite{Watson,Migdal}
212: \begin{eqnarray}
213: |f|^2&=& |f_p|^2 \cdot |T(q)|^2
214: \nonumber \\
215: &\approx&\frac{|f_p|^2}{(1{+}\Im a q)^2{+}(\Re a q)^2{+}\Re a\Re r_0q^2
216: {-}\Im a\Im r_0q^2},
217: \label{fsqr}
218: \end{eqnarray}
219: where $f_p$ is the $s$-wave production operator, assumed to be independent
220: of the final momentum near the reaction threshold.
221:
222: While the coefficient of the term linear in $q$ is given by the
223: imaginary part of the scattering length alone, the $q^2$ term contains
224: {\it both the complex scattering length and effective range}.
225: In case of the two-nucleon system $|a|{\gg}r_0$ and therefore
226: a further approximation is reasonable. It consists in
227: the neglect of the effective range,
228: i.e. of the second term in Eq. (\ref{efran}), so that
229: the reaction amplitude is simply given by
230: \begin{equation}
231: f=\frac{f_p}{1-iaq}\, .
232: \label{fitm}
233: \end{equation}
234: Then Eq. (\ref{fsqr}) reduces to the form
235: \begin{equation}
236: |f|^2=\frac{|f_p|^2}{1{+}2 \Im a q {+}|a|^2 q^2}\, .
237: \label{fitn}
238: \end{equation}
239:
240: However, for the $\eta^3{\rm He}$ system $a$ and $r_o$ are
241: expected to be of the same order of magnitude so that the above
242: approximation is not really justified. Thus, here $|a|^2$ as determined
243: from Eq. (\ref{fitn}) can only be considered as an effective quantity
244: rather than the modulus of the physical scattering length. Clearly
245: separating the scattering length and effective range is only possible by
246: making further assumptions or within specific model calculations, which
247: means in a model-dependent way.
248: On the other hand, in the region very close to threshold where the
249: term linear in $q$ should dominate, in principle, there
250: is a possibility to determine the imaginary part of the
251: scattering length from the momentum dependence of the reaction
252: amplitude $f(q)$. It should be feasible in the momentum range
253: $q \le 1/2a$.
254:
255: \section{Data at low energies}
256:
257: General information on the data base discussed in the present
258: paper is summarized in Table~\ref{exp}.
259:
260: The application of the formalism described in the last section
261: is only sensible if two requirements are fulfilled:
262: (i) the production data show a significant momentum
263: dependence near threshold; (ii) the production occurs
264: predominantly in s-waves.
265: %
266: A strong momentum dependence of the
267: spin averaged squared $p{+}d{\to}^3{\rm He}{+}\eta$ reaction amplitude
268: defined as
269: \begin{equation}
270: |f(\vartheta)|^2 := \frac{k}{q}\, \frac{d\sigma}{d\Omega},
271: \label{averm}
272: \end{equation}
273: was indeed seen in the first reported near-threshold measurement in 1988
274: \cite{Berger}.
275: Here $k$ and $q$ are the initial and final particle momenta
276: in the center of mass system and $d\sigma{/}d\Omega$ stands for the
277: cms differential cross section. The measurements
278: were done only at the $\eta$-meson production angles $\vartheta{=}0^o$ and
279: $\vartheta{=}180^o$ in the cms and are presented in Fig.~\ref{etan6a}
280: by full and open squares. The open circles (open crosses) in
281: Fig.~\ref{etan6a} show earlier data of Banaigs {\it et al.}~\cite{Banaigs}
282: (Berthet {\it et al.}~\cite{Berthet}) for $\eta$-meson production at
283: $\vartheta{=}180^o$ at somewhat higher energies.
284:
285: \begin{table}[t]
286: \caption{Data on the reaction $p{+}d{\to}^3{\rm He}{+}\eta$ discussed in
287: the present paper. $q$ is the cms momentum in the (final) $\eta ^3{\rm He}$
288: system and $\varepsilon$ is the corresponding excess energy. We use
289: $m_{^3{\rm He}}$ = 2809.414 MeV and $m_\eta$ = 547.3 MeV.}
290: \label{etadata}
291: \begin{tabular}{l|c|c|c|c}
292: \hline\noalign{\smallskip}
293: & Ref. & Observable
294: & $q$ (MeV/c) & $\varepsilon$ (MeV) \\
295: \noalign{\smallskip}\hline\noalign{\smallskip}
296: Berger & \cite{Berger} & $\!\!\!\!$
297: $ \ \sigma (0^o)$, $\sigma (180^o)$ $\!\!\!\!$
298: & 7{-}136 & 0.054{-}20 \\
299: Banaigs & \cite{Banaigs} & $\sigma (180^o)$ & 265{-}406 & 73{-}163 \\
300: Banaigs & \cite{Banaigs} & $\sigma (\vartheta)$ & 283 & 83 \\
301: Berthet & \cite{Berthet} & $\sigma (180^o)$ & 118{-}955 & 15{-}711 \\
302: Mayer & \cite{Mayer} & $\sigma_{tot}$, $A_{cm}$ & 11{-}75 & 0.13{-}6.11 \\
303: Bilger & \cite{Bilger} & $\sigma_{tot}$,
304: $\sigma (\vartheta)$ & 138{-}334 & 21{-}114 \\
305: Betigeri & \cite{Betigeri} & $\sigma_{tot}$, $\sigma (\vartheta)$ & 214 & 49 \\
306: Kirchner & \cite{Kirchner} & $\sigma_{tot}$, $\sigma (\vartheta)$ & 568 & 298 \\
307: \hline
308: \end{tabular}
309: \label{exp}
310: \end{table}
311:
312: Despite the discrepancies between the
313: data~\cite{Banaigs,Berthet} on backward
314: $\eta$-meson production in the range $250{\le}q{\le}500$ MeV/c
315: it is clear from Fig.~\ref{etan6a} that there is a strong $q$ dependence
316: of $|f(\vartheta)|^2$ up to rather high energies.
317: %
318: Furthermore, from the data of Berger et al.~\cite{Berger},
319: which are available for both $\vartheta{=}0^o$ and $\vartheta{=}180^o$,
320: one can conclude that the angular dependence is small for
321: final momenta up to $q \approx 65$~MeV/c which suggests that
322: the reaction amplitude should be dominated by the $s$-wave in this
323: momentum range.
324:
325: Further data on the reaction $p{+}d{\to}^3{\rm He}{+}\eta$
326: near threshold, obtained with the SPES2 spectrometer at Saclay,
327: were reported in 1996 \cite{Mayer}. This experiment provided
328: data on the total reaction cross section $\sigma_{\rm tot}$ and a
329: forward-backward asymmetry in the cm system $A_{\rm cm}$ defined as
330: \begin{equation}
331: \frac{d\sigma}{d\Omega} = \frac{\sigma_{\rm tot}}{4\pi}[1 +
332: A_{\rm cm}\cos\vartheta] \ .
333: \label{assym}
334: \end{equation}
335:
336: The latter observable is
337: shown in Fig.~\ref{etan7a} as a function of the final momentum.
338: The full circles and squares indicate experimental results
339: obtained under different criteria for data analysis.
340: Evidently, within 5\% accuracy in the amplitude the asymmetry is
341: consistent with zero. Thus, this measurement confirms that for
342: $q{\le}70$~MeV the reaction $p{+}d{\to}^3{\rm He}{+}\eta$ is dominated by
343: the $s$-wave. Therefore, we will use the data on
344: $p{+}d{\to}^3{\rm He}{+}\eta$ in this momentum range
345: for investigating effects of the final state interaction
346: (FSI) for the $\eta^3{\rm He}$ system.
347:
348: Assuming that the total reaction cross section is governed by the
349: $s$ wave one can compute the average reaction amplitude squared,
350: $|f|^2$, from the data of Ref. \cite{Mayer} by means of Eq.~(\ref{averm}).
351: Corresponding results (now for the spin and angle averaged reaction
352: amplitude) are shown in Fig.~\ref{etan3b}. This figure contains also
353: available data~\cite{Nikulin,Cameron,Roessle} for the reaction
354: $p{+}d{\to}^3{\rm He}{+}\pi^0$. They are shown here in order to
355: illustrate the strong momentum dependence of the $\eta^3{\rm He}$
356: channel, which is due to the corresponding FSI.
357:
358: \begin{figure}[b]
359: \vspace*{-6mm}
360: \hspace*{-1.mm}\psfig{file=etan7a.ps,width=9.3cm,height=9.cm}
361: \vspace*{-7mm}
362: \caption{Data on the $\eta$-meson forward-backward
363: asymmetry $A_{\rm cm}$ as a function of final momentum $q$
364: from Ref. \cite{Mayer}. The different symbols show the experimental
365: results obtained for different analyzing criteria \cite{Mayer}.}
366: \label{etan7a}
367: \end{figure}
368:
369: \section{The $\eta^3{\rm He}$ scattering length}
370: Berger {\it et al.} \cite{Berger} did not attempt to extract the
371: $\eta^3{\rm He}$ scattering length from their data. However, they fitted
372: the data with the function
373: \begin{equation}
374: |f|^2=\frac{x}{(1 - yq\cos\vartheta + zq^2)^2} \ .
375: \label{fit1}
376: \end{equation}
377: The corresponding results are shown by solid lines in Fig.~\ref{etan6a} for
378: $\cos\vartheta{=}{\pm}1$. Note that Eq.~(\ref{fit1}) is
379: not the FSI correction to the production amplitude that
380: follows from the Watson-Migdal approximation~\cite{Watson,Migdal}.
381: However, it can be matched to Eq. (\ref{fitn}) to order $q^2$ after
382: averaging over the angle dependence. The explicit value for the modulus of
383: the scattering length extracted in this way amounts to $|a|{=}3.4{\pm}0.1$~fm,
384: using only the errors given in Ref.~\cite{Berger}.
385:
386: Employing Eq. (\ref{fitn}) Wilkin~\cite{Wilkin1} analysed
387: the preliminary Saclay data on the $p{+}d{\to}^3{\rm He}{+}\eta$ total
388: cross section \cite{Kessler}. He reported a correlation between the
389: real and imaginary parts of the $\eta^3{\rm He}$ scattering length
390: in the form
391: \begin{equation}
392: (\Re a)^2=21.44 - 0.449 (\Im a)^2 - 4.509 \Im a,
393: \label{wilk}
394: \end{equation}
395: as outcome of a $\chi^2$ minimization.
396: This correlation is shown in Fig.~\ref{etan9a} by the dashed line. Note
397: that Eq.~(\ref{wilk}) does not contain information about the
398: standard $\chi^2{+}1$ uncertainty of the fit and, because of the
399: unitarity condition, should be applied only for $\Im{a}{\ge}0$.
400:
401: \begin{figure}[b]
402: \vspace*{-6mm}
403: \hspace*{-1.mm}\psfig{file=etan3b.ps,width=9.3cm,height=9.cm}
404: \vspace*{-5mm}
405: \caption{Spin and angle averaged transition amplitudes
406: $|f|^2$ extracted from
407: $p{+}d{\to}^3{\rm He}{+}\pi^0$ \cite{Nikulin,Cameron,Roessle} and
408: $p{+}d{\to}^3{\rm He}{+}\eta$
409: \cite{Berger,Banaigs,Mayer,Bilger,Betigeri,Kirchner} data
410: on total reaction cross sections as functions of the final momentum
411: $q$ in the c.m. system. The dotted line shows the fit to
412: the reaction $p{+}d{\to}^3{\rm He}{+}\pi^0$ from Ref.~\cite{Nikulin}.
413: The
414: solid line is our overall fit by Eq.~(\ref{fitm}) to low energy data published
415: by Mayer {\it et al.}~\cite{Mayer} and Berger {\it et al.}~\cite{Berger},
416: while the dashed line shows our fit to the data from
417: Mayer {\it et al.}~\cite{Mayer} alone.}
418: \label{etan3b}
419: \end{figure}
420:
421: We took the $\eta^3{\rm He}$ scattering lengths constrained by
422: Eq.~(\ref{wilk}) and employed Eq.~(\ref{fitn}) to calculate the average
423: squared reaction amplitude as a function of the final momentum.
424: Corresponding results are shown in Fig. \ref{etan6a}, where the
425: shaded area indicates the spread of $|f|^2$ with the limiting
426: scattering lengths of $a{\simeq}0{+}i3.51$ fm and $a{\simeq}4.63{+}i0$ fm
427: as given in Eq.~(\ref{wilk}). It is worth mentioning
428: that Wilkin did not include the data of Berger {\it et al.} in
429: his fit, though they were already available.
430:
431: Mayer {\it et al.} used also only their own data~\cite{Mayer} when
432: they extracted the $\eta^3{\rm He}$ scattering length by utilizing
433: Eq. (\ref{fitm}). Their result,
434: \begin{equation}
435: a{=}|3.8{\pm}0.6|{+}i(1.6{\pm}1.1) \ \mbox{fm} \ ,
436: \label{sc1}
437: \end{equation}
438: is shown in Fig.~\ref{etan9a} by the shaded boxes. Since, as mentioned,
439: the sign of the real part of the $\eta^3{\rm He}$ scattering length cannot
440: be inferred from a fit to the cross section data alone we include here
441: boxes corresponding to $\pm {\Re} a$, with $a$ given by Eq. (\ref{sc1}).
442:
443: \begin{figure}[t]
444: \vspace*{-3mm}
445: \hspace*{-1.mm}\psfig{file=etan9a.ps,width=9.3cm,height=9.cm}
446: \vspace*{-7mm}
447: \caption{Real versus imaginary part of the $\eta^3{\rm He}$
448: scattering length. The shaded boxes indicate the value
449: given by Mayer {\it et al.}~\cite{Mayer}.
450: The solid contour lines show the result of our fit to the
451: data of Mayer {\it et al.}~\cite{Mayer}
452: for $\chi^2{+}0.5$, $\chi^2{+}1$ and $\chi^2{+}4$ confidence levels,
453: respectively.
454: The dashed line is the parameterization of Eq.~(\ref{wilk}) from
455: Ref.~\cite{Wilkin1}.
456: The symbols show results of various model calculations,
457: taken from Refs. \cite{Wycech} (inverse triangles), \cite{Rakityansky}
458: (squares), \cite{Fix1} (circles), \cite{Wilkin1} (triangle) and
459: \cite{Green} (star).
460: }
461: \label{etan9a}
462: \end{figure}
463:
464: The symbols in Fig.~\ref{etan9a} represent results of various
465: model calculations~\cite{Wycech,Rakityansky,Fix1,Wilkin1,Green}
466: based on different approaches and different elementary $\eta N$ amplitudes.
467: For convenience, selected results are also listed in
468: Table~\ref{allowed} together with the elementary $\eta{N}$ scattering
469: length that is employed in those model calculations. Evidently,
470: only the result from Ref.~\cite{Green} is in agreement with the
471: $\eta^3{\rm He}$ scattering length extracted by Mayer
472: {\it et al.}~\cite{Mayer}.
473: In the course of our study we have refitted the data from Ref. \cite{Mayer}.
474: Our result is
475: indicated by the solid contour lines in Fig.~\ref{etan9a} for $\chi^2{+}0.5$,
476: $\chi^2{+}1$ and $\chi^2{+}4$ confidence levels. Apparently,
477: it differs somewhat from the one published in Ref. \cite{Mayer}.
478: Specifically, one can see that the error correlation matrix is not
479: symmetric and that now several model predictions
480: from the Refs.\cite{Wilkin1,Wycech,Rakityansky,Fix1,Green} lie within the
481: $\chi^2{+}1$ confidence level.
482:
483: We want to point out in this context that the value of the total $\chi^2$
484: at the minimum that results from our fit is $\chi^2{=}$0.16, which we find
485: to be much too low. Indeed for a statistically uncorrelated set
486: of data points one would expect a value of
487: $\chi^2$ = $N_{df} \pm \sqrt{2 N_{df}}$, where $N_{df}$ is the number
488: of degrees of freedom \cite{Nijmegen} -- which in this particular case
489: would be 5. The error bars of the Saclay data are dominated by the
490: statistical error \cite{Mayer} and, therefore, they cannot be the
491: origin of this small $\chi^2$. Rather it seems to us that the published
492: data points are simply not independent.
493:
494: \begin{table}[t]
495: \caption{Model calculations of the $\eta ^3{\rm He}$ scattering length
496: that lie within the $\chi^2{+}1$ confidence level in Fig.~\ref{etan9a}.
497: The employed values of the $\eta N$ scattering length and the used
498: approach is also specified.}
499: \label{allowed}
500:
501: \begin{tabular}{l|c|c|l}
502: \hline\noalign{\smallskip}
503: Ref. & $ a(\eta ^3{\rm He})$ (fm) & $ a(\eta N) $ (fm) & \,\,Comment \\
504: \noalign{\smallskip}\hline\noalign{\smallskip}
505: \cite{Wycech} & $1.99{+}i2.86$ & $0.48{+}i0.28$ & Multiple scattering \\
506: \cite{Wycech} & $0.92{+}i3.07$ & $0.43{+}i0.39$ & Multiple scattering \\
507: \cite{Rakityansky} & $-1.96{+}i2.86$ & $0.62{+}i0.30$ &
508: Finite-rank approx. \\
509: \cite{Rakityansky} & $-2.66{+}i3.31$ & $0.67{+}i0.30$ &
510: Finite-rank approx. \\
511: \cite{Fix1} & $2.23{+}i3.00$ & $0.57{+}i0.39$ & Faddeev-Yakubovsky \\
512: \cite{Wilkin1} & $-2.31{+}i2.57$ & $0.55{+}i0.30$ & Optical potential \\
513: \hline
514: \end{tabular}
515: \end{table}
516:
517:
518: \begin{figure}[b]
519: \vspace*{-6mm}
520: \hspace*{-1.mm}\psfig{file=etan9.ps,width=9.3cm,height=9.cm}
521: \vspace*{-7mm}
522: \caption{
523: Real versus imaginary part of the $\eta^3{\rm He}$
524: scattering length.
525: The solid contour lines show the result of our fit to the
526: combined data of Mayer {\it et al.}~\cite{Mayer}
527: and Berger {\it et al.}~\cite{Berger}
528: for $\chi^2{+}0.5$, $\chi^2{+}1$ and $\chi^2{+}4$ confidence levels,
529: respectively.
530: The symbols show results of various model calculations,
531: taken from Refs. \cite{Wycech} (inverse triangles), \cite{Rakityansky}
532: (squares), \cite{Fix1} (circles),\cite{Wilkin1} (triangle) and
533: \cite{Green} (star).}
534: \label{etan9}
535: \end{figure}
536:
537: Considering this certainly to some extent strange feature of the
538: Saclay data one might {\it a priori} expect that an evaluation of the
539: $\eta^3{\rm He}$ scattering length from a combined data analysis
540: is very uncertain. Nevertheless we combine the data from Mayer
541: {\it et al.}~\cite{Mayer} and Berger {\it et al.}~\cite{Berger} to fit them
542: by Eq.~(\ref{fitm}). Corresponding result are presented in Figs.~\ref{etan3b}
543: and \ref{etan9}.
544: Besides yielding a substantially larger total $\chi^2{=}$57
545: also the confidence contours are different for the combined fit
546: as can be seen by comparing Fig.~\ref{etan9} with Fig.~\ref{etan9a}.
547: (We show again the correlation between the real and imaginary part of
548: the $\eta^3{\rm He}$ scattering length for the $\chi^2{+}0.5$,
549: $\chi^2{+}1$, $\chi^2{+}2$, $\chi^2{+}3$ and $\chi^2{+}4$ confidence
550: levels.) Obviously, the combined analysis allows for a more definite
551: determination of the $\eta^3{\rm He}$ scattering length.
552: In particular now the model predictions~\cite{Wycech,Rakityansky,Fix1,Wilkin1}
553: lie outside of the $\chi^2{+}1$ confidence level, except of the
554: most recent result from Ref.\cite{Green}. On the other hand,
555: the fact that the resulting $\chi^2$ minimum points to a
556: $\eta^3{\rm He}$ scattering length with vanishing (or even slightly negative)
557: imaginary part is definitely a reason to worry and is presumably a signal that
558: the near-threshold data base is internally inconsistent and/or afflicted with
559: errors. Evidently, for further progress in the determination of the
560: $\eta^3{\rm He}$ scattering length new measurements at final momenta
561: $q{<}100$~MeV/c are required.
562:
563: \section{Estimates for the imaginary part of the scattering length}
564: In principle, the imaginary part of the $\eta^3{\rm He}$
565: scattering length could be obtained from the total $\eta^3{+}{\rm He}$
566: interaction cross section $\sigma_{tot}$ by utilizing the optical theorem
567: in the limit $q \ \to \ 0$:
568: \begin{equation}
569: \Im a = \lim_{q \to 0} \Im f_{tot}(\vartheta = 0^o) =
570: \lim_{q \to 0} \frac{q}{4\pi}\, \sigma_{tot} \ .
571: \label{optic}
572: \end{equation}
573: Although $\sigma_{tot}$ is not accessible experimentally, one can use
574: at least experimental information on partial $\eta^3{+}{\rm He}$ reaction cross
575: sections in order to deduce lower bounds on $\Im a$. This procedure works very
576: well for the $\eta N$ case where the magnitude of the imaginary part of the
577: scattering length is strongly constrained by the data on the
578: $\pi^- p \to \eta n$ transition cross section \cite{Wilkin1,Sibirtsev1}.
579:
580: \begin{figure}[t]
581: \vspace*{-3mm}
582: \hspace*{0.mm}\psfig{file=etan3a.ps,width=9.3cm,height=9.cm}
583: \vspace*{-5mm}
584: \caption{Bounds on the imaginary part of the $\eta^3{\rm He}$ forward
585: scattering amplitude extracted from experimental results available for the
586: $p{+}d{\to}^3{\rm He}{+}\eta$
587: \cite{Berger,Banaigs,Mayer,Bilger,Betigeri,Kirchner} and
588: $\pi^-{+}^3{\rm He}{\to}t{+}\eta$ \cite{Peng1,Peng2} reactions.
589: The solid line shows our estimate based on an overall fit to low
590: energy $p{+}d{\to}^3{\rm He}{+}\eta$ data.}
591: \label{etan3a}
592: \end{figure}
593:
594: Using detailed balance
595: the $\eta^3{+}{\rm He}{\to}p{+}d$ cross section can be related to the data
596: available for the inverse reaction by
597: \begin{equation}
598: \sigma (\eta{+}^3{\rm He}{\to}p{+}d) =\frac{3k^2}{q^2} \,
599: \sigma (p{+}d{\to}^3{\rm He}{+}\eta) \ .
600: \label{bal1}
601: \end{equation}
602: In Fig.~\ref{etan3a} we show $\Im f(\vartheta = 0^o)$
603: obtained via Eqs.~(\ref{optic}) and (\ref{bal1}) from experimental
604: results~\cite{Mayer,Berger,Banaigs,Bilger,Betigeri,Kirchner}
605: available for the reaction $p{+}d{\to}^3{\rm He}{+}\eta$. The solid
606: line in Fig.~\ref{etan3a} shows the estimate based on our overall fit
607: to the low energy data published by
608: Mayer {\it et al.}~\cite{Mayer} and Berger {\it et al.}~\cite{Berger}.
609:
610: One can also evaluate the partial cross sections for
611: $\eta{+}^3{\rm He}{\to}^3{\rm He}{+}\pi^0$ and
612: $\eta{+}^3{\rm He}\to {\rm t}{+}\pi^+$
613: from the data available for the reaction $\pi^-{+}^3{\rm He}{\to}t{+}\eta$
614: \cite{Peng1,Peng2}. Taking into account the
615: isotopical relations between the different reaction channels given by
616: \begin{eqnarray}
617: \sigma (\eta{+}{\rm t}{\to}^3{\rm He}{+}\pi^-)=
618: \sigma (\eta{+}^3{\rm He}{\to}{\rm t}{+}\pi^+)\nonumber \\
619: =2\sigma (\eta{+}^3{\rm He}{\to}^3{\rm He}{+}\pi^0)\, ,
620: \end{eqnarray}
621: the corresponding value of $\Im f(\vartheta = 0^o)$ can be estimated.
622: It is also included in Fig.~\ref{etan3a}.
623:
624: \begin{figure}[t]
625: \vspace*{-8mm}
626: \hspace*{-2.5mm}\psfig{file=etan2.ps,width=10cm,height=10.cm}
627: \vspace*{-7mm}
628: \caption{Angular spectra of $\eta$-mesons
629: produced in $p{+}d{\to}^3{\rm He}{+}\eta$ reaction at different
630: final momenta $q$. The data are from Ref. \cite{Bilger} where different
631: symbols show results obtained with different analyzing criteria.
632: The solid lines indicate the fit given in Ref.~\cite{Bilger}.}
633: \label{etan2}
634: \end{figure}
635:
636: Unfortunately,
637: the lower bounds for the imaginary part of the $\eta^3{\rm He}$
638: scattering length extracted from those reaction channels turn out to be rather
639: small, i.e. $\Im a{>}10^{-2}$~fm, and, therefore, are not very useful.
640: Presumably the bulk of the inelastic cross section comes from the reaction
641: $\eta + ^3{\rm He} \to ppn$ which is, of course, not accessible experimentally.
642: In any case,
643: intuitively one would expect that $\Im a$ should be at least 3 times the
644: imaginary part of the elementary $\eta N$ scattering length. Indeed all
645: results of microscopic calculations in the literature
646: \cite{Wycech,Belyaev1,Belyaev2,Rakityansky,Fix1} (cf. also
647: Table~\ref{allowed})
648: seem to be consistent with this hypothesis. Then, based on
649: the lower bound, $\Im a_{\eta N}{\approx}0.28$ fm, estimated by using the
650: optical theorem \cite{Wilkin1} one would arrive at $\Im a{\ge}0.84$ fm
651: which might be a reasonable guess.
652:
653:
654: \begin{figure}[b]
655: \vspace*{-9mm}
656: \hspace*{-0.2mm}\psfig{file=etan7.ps,width=9.3cm,height=10.cm}
657: \vspace*{-8mm}
658: \caption{Angular spectra of $\eta$-mesons
659: produced in the reaction $p{+}d{\to}^3{\rm He}{+}\eta$ at different
660: final momenta $q$. The data are from Refs. \cite{Banaigs,Betigeri,Kirchner}.
661: The different symbols in (a) show results obtained with different analyzing
662: criteria~\cite{Betigeri}. The solid lines indicate our fit.}
663: \label{etan7}
664: \end{figure}
665:
666: \section{Data at higher energies}
667:
668: Angular spectra for the reaction $p{+}d{\to}^3{\rm He}{+}\eta$
669: at momenta $q{>}$100~MeV/c \cite{Bilger,Betigeri,Banaigs,Kirchner}
670: are shown in Figs.~\ref{etan2} and \ref{etan7}. These data exhibit
671: already a strong asymmetry. Thus, it is clear that in this energy
672: region the reaction is dominated by higher partial waves.
673: Note that the solid lines in
674: Fig.~\ref{etan2} are taken from the original work while the curves in
675: Figs.~\ref{etan7} show our own fit to the experimental results using
676: Legendre polynomials.
677:
678: From those fits one can again compute the squared spin and
679: angle averaged transition amplitude and the corresponding results
680: are included in Fig.~\ref{etan3b}. One can also extrapolate $|f|^2$
681: to very forward and backward angles and the corresponding values
682: are shown in Fig.~\ref{etan6a}.
683: We detect substantial discrepancies between the
684: $p{+}d{\to}^3{\rm He}{+}\eta$ forward cross sections
685: extrapolated from the data of
686: Bilger {\it et al.}~\cite{Bilger}, Banaigs {\it et al.}~\cite{Banaigs} and
687: Betigeri {\it et al.}~\cite{Betigeri}. The extrapolated data on backward
688: $\eta$-meson production~\cite{Bilger,Betigeri,Banaigs,Kirchner} are in
689: rough agreement with other published results~\cite{Berger,Berthet}
690: for $q{<}$300~MeV/c, taking into account that the data from Ref.~\cite{Bilger}
691: have large uncertainties.
692:
693: \begin{figure}[t]
694: \vspace*{3mm}
695: \hspace*{-2.mm}\psfig{file=etan1.ps,width=9.3cm,height=9.cm}
696: \vspace*{-7mm}
697: \caption{The $p{+}d{\to}^3{\rm He}{+}\eta$ total reaction cross
698: section as a function of the final momentum $q$ in the c.m. system (lower axis)
699: or excess energy $\epsilon$ (upper axis).
700: The data are from Refs. ~\cite{Berger,Banaigs,Mayer,Bilger,Betigeri,Kirchner}.
701: The
702: solid line is our overall fit by Eq.~(\ref{fitm}) to low energy data published
703: by Mayer {\it et al.}~\cite{Mayer} and Berger {\it et al.}~\cite{Berger},
704: while the dashed line shows our fit to the data from
705: Mayer {\it et al.}~\cite{Mayer} alone.}
706: \label{etan1}
707: \end{figure}
708:
709: The discrepancies between the available data are also reflected in
710: Fig.~\ref{etan1}, where the experimental results
711: \cite{Berger,Banaigs,Mayer,Bilger,Betigeri,Kirchner}
712: on the $p{+}d{\to}^3{\rm He}{+}\eta$ total reaction cross
713: section are shown as a function of the final momentum $q$ in the c.m. system
714: and the excess energy $\epsilon$. Here with open circles we also present
715: total reaction cross sections for the data of
716: Berger {\it et al.}~\cite{Berger}, derived from their forward
717: and backward $\eta$-meson production cross sections via Eq.~(\ref{assym}).
718: %
719: Evidently, there is not much consistency between the various
720: data sets -- neither for low nor for higher energies.
721:
722: \section{Summary}
723: We have critically reviewed the presently available data for the
724: reaction $p{+}d{\to}^3{\rm He}{+}\eta$ with the aim of
725: extracting the $\eta^3{\rm He}$ scattering
726: length. The experimental information on angular spectra clearly
727: shows that the reaction is dominated by the $s$-wave up to final
728: momenta of around 70~MeV/c and, therefore, we have used all data in this
729: energy range for the evaluation of the $\eta^3{\rm He}$
730: scattering length.
731:
732: The analysis provides strong indications that
733: the low energy data published by
734: Berger {\it et al}~\cite{Berger} and Mayer {\it et al.}~\cite{Mayer}
735: are not consistent with each other.
736: The overall fit to all low energy data~\cite{Berger,Mayer}
737: results in a large total $\chi^2$ of 57, however clearly locates the
738: scattering length
739: as $\Im a{=}0.5{\pm}0.5$~fm and $\Re a{=}4.3{\pm}0.3$~fm. The fit to the
740: data from Mayer {\it et al.}~\cite{Mayer} alone results in a rather
741: small total $\chi^2$ of only 0.16, but yields an $\eta^3{\rm He}$ scattering
742: length with much too large statistical uncertainty.
743:
744: Further progress
745: in the determination of the $\eta^3{\rm He}$ scattering length requires
746: new measurements at final momenta $q{\le}70$~MeV/c in order to settle the
747: ambiguities exhibited by the present data base. In particular, it would be
748: nice to obtain information about the $p{+}d{\to}^3{\rm He}{+}\eta$
749: reaction cross section very close to threshold, because this would provide
750: us with more stringent constraints for the imaginary part of the
751: scattering length.
752: Furthermore, measurements of the angular spectrum of the $\eta$ meson
753: at energies corresponding to final momenta around $q = 70$~MeV/c
754: would be very useful. Such data would allow to examine whether the reaction
755: is still dominated by the $s$-wave up to this energy -- as we assumed in
756: our analysis. The mentioned
757: experiments could be done at accelerator facilities like COSY
758: and CELSIUS \cite{COSY2,Khoukaz}.
759:
760: Our systematical analysis shows that the
761: $p{+}d{\to}^3{\rm He}{+}\eta$ data available at larger final momenta,
762: $q{>}$100~MeV/c, are also not consistent with each other. Obviously this
763: situation constitutes a substantial difficulty in the comparison between the
764: experimental results and theoretical calculations. Here too the problem can be
765: solved only by improving and expanding the data base for the reaction
766: $p{+}d{\to}^3{\rm He}{+}\eta$.
767:
768: \begin{acknowledgement}
769: This work was performed in part under the contract No. DE-FG02-93ER40756
770: with the University of Helsinki. Financial support for this work was
771: also provided in part by the
772: international exchange program between DAAD (Germany, Project No.
773: 313-SF-PPP-8) and the Academy of Finland (Project Nos. 41926
774: and 54038).
775: \end{acknowledgement}
776:
777: \begin{thebibliography}{100}
778: \bibitem{Haider1}
779: Q. Haider and L.C. Liu, Phys. Lett. B {\bf 172}, 257 (1986).
780: \bibitem{Liu}
781: L.C. Liu and Q. Haider, Phys. Rev. C {\bf 34}, 1845 (1986).
782: \bibitem{Chiang}
783: H.C. Chiang, E. Oset, and L.C. Liu, Phys. Rev. C {\bf 44}, 738 (1991).
784: \bibitem{Haider2}
785: Q. Haider and L.C. Liu, Phys. Rev. C {\bf 66}, 045208 (2002).
786: \bibitem{Wycech}
787: S. Wycech, A.M. Green, and J.A. Niskanen, Phys. Rev.
788: C {\bf 52}, 544 (1995).
789: \bibitem{Belyaev1}
790: S.A. Rakityansky, S.A. Sofianos, W. Sandhas,
791: and V.B. Belyaev, Phys. Lett. B {\bf 359}, 33 (1995).
792: \bibitem{Belyaev2}
793: V.B. Belyaev, S.A. Rakityansky, S.A. Sofianos, M. Braun, and
794: W. Sandhas, Few. Body. Syst. Suppl. {\bf 8}, 309 (1995).
795: \bibitem{Rakityansky}
796: S.A. Rakityansky, S.A. Sofianos, M. Braun, V.B. Belyaev
797: and W. Sandhas, Phys. Rev. C {\bf 53}, R2043 (1996).
798: \bibitem{Fix1}
799: A. Fix and H. Arenh\"ovel, Phys. Rev. C {\bf 66}, 024002 (2002).
800: \bibitem{Wilkin1}
801: C. Wilkin, Phys. Rev. C. {\bf 47}, R938 (1993).
802: \bibitem{Green}
803: A.M. Green and S. Wycech, Phys. Rev. C {\bf 68}, 061601 (2003).
804: \bibitem{GSI1}
805: R.S. Hayano et al., GSI/SIS Proposal S214, 1997.
806: \bibitem{GSI2}
807: R.S. Hayano, S. Hirenzaki, and A. Gillitzer,
808: Eur. Phys. J. A {\bf 6}, 99 (1999).
809: \bibitem{COSY1}
810: H. Machner et al., COSY Proposal no. 50.1, 2000.
811: \bibitem{COSY2}
812: A. Khoukaz et al., COSY Proposal no. 62.2, 2000.
813: \bibitem{COSY3}
814: A. Gillitzer et al., COSY Proposal no. 102, 2001.
815: \bibitem{Grassberger}
816: P. Grassberger and W. Sandhas, Nucl. Phys. B {\bf 2}, 181 (1967);
817: E.O. Alt, P. Grassberger and W. Sandhas, Phys. Rev. C {\bf 1},
818: 85 (1970).
819: \bibitem{Faddeev}
820: O.A. Yakubovsky, Sov. J. Nucl. Phys. {\bf 5}, 937 (1967).
821: \bibitem{Sibirtsev1}
822: A. Sibirtsev, S. Schneider, Ch. Elster, J. Haidenbauer,
823: S. Krewald and J. Speth, Phys. Rev. C {\bf 65}, 044007 (2002).
824: \bibitem{Newton}
825: R.G. Newton, {\it Scattering Theory of Waves and Particles}
826: (Springer-Verlag, New York, 1982).
827: \bibitem{watson}
828: M. Goldberger and K.M. Watson, {\it Collision Theory}
829: (Wiley, New York, 1964).
830: \bibitem{mix}
831: V. Baru, J. Haidenbauer, C. Hanhart, and J.A. Niskanen,
832: Phys. Rev. C {\bf 68}, 035203 (2003).
833: \bibitem{Berger}
834: J. Berger {\it et al.}, Phys. Rev. Lett. {\bf 61}, 919 (1988).
835: \bibitem{Banaigs}
836: J. Banaigs {\it et al.}, Phys. Lett. B {\bf 45}, 394 (1973).
837: \bibitem{Berthet}
838: P. Berthet {\it et al.}, Nucl. Phys. A {\bf 443}, 589 (1985).
839: \bibitem{Mayer}
840: B. Mayer {\it et al.}, Phys. Rev. C {\bf 53}, 2068 (1996).
841: \bibitem{Bilger}
842: R. Bilger {\it et al.}, Phys. Rev. C {\bf 65}, 044608 (2002).
843: \bibitem{Betigeri}
844: M. Betigeri et al, Phys. Lett. B {\bf 472}, 267 (2000).
845: \bibitem{Kirchner}
846: T. Kirchner, PhD. Thesis, Institute de Physique Nucleaire,
847: Orsay (1993).
848: \bibitem{Watson}
849: K.M. Watson, Phys. Rev. {\bf 88}, 1163 (1952).
850: \bibitem{Migdal}
851: A.B. Migdal, JETP {\bf 1}, 2 (1955).
852: \bibitem{Nikulin}
853: V.N. Nikulin {\it et al.}, Phys. Rev. C {\bf 54}, 1732 (1996).
854: \bibitem{Cameron}
855: J.M. Cameron {\it et al.}, Nucl. Phys. A {\bf 472}, 718 (1987).
856: \bibitem{Roessle}
857: E. Roessle {\it et al.}, Proceedings of the Conference on Pion
858: Production and Absorption in
859: Nuclei, AIP Conf. Proc. {\bf 79}, 171 (1982).
860: \bibitem{Kessler}
861: R. Kessler, Ph.D. thesis, UCLA (1992).
862: \bibitem{Nijmegen}
863: J.R. Bergervoet, P.C. van Campen, W.A. van der Sanden,
864: and J.J. de Swart, Phys. Rev. C {\bf 38}, 15 (1988).
865: \bibitem{Peng1}
866: J.C. Peng {\it et al.}, Phys. Rev. Lett. {\bf 58}, 2027 (1987).
867: \bibitem{Peng2}
868: J.C. Peng {\it et al.}, Phys. Rev. Lett. {\bf 63}, 2353 (1989).
869: \bibitem{Khoukaz}
870: A. Khoukaz {\it et al.}, in preparation.
871: \end{thebibliography}
872: \end{document}
873:
874: