1: %
2: % October 7th, 2002
3: %
4: %\documentclass[aps,preprint,epsfig]{revtex}
5: %\documentstyle[aps,preprint,epsfig]{revtex}
6: \documentstyle[preprint,aps,epsfig]{revtex}
7: \tighten
8: %
9: \pagestyle{plain}
10: %
11: \oddsidemargin=0.1cm
12: \evensidemargin=0.1cm
13: \textwidth=16.5cm
14: \topmargin=-1.5cm
15: \textheight=23cm
16:
17:
18:
19: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
20:
21:
22:
23: \def\Journal#1#2#3#4{{#1} {\bf #2}, #3 (#4)}
24:
25:
26:
27: \def\APP{{\em Acta Phys. Pol.} B}
28: \def\NCA{\em Nuovo Cimento}
29: \def\NIM{\em Nucl. Instrum. Methods}
30: \def\NIMA{{\em Nucl. Instrum. Methods} A}
31: \def\NPA{{\em Nucl. Phys.} A}
32: \def\NPB{{\em Nucl. Phys.} B}
33: \def\PLB{{\em Phys. Lett.} B}
34: \def\PRL{\em Phys. Rev. Lett.}
35: \def\PR{\em Phys. Rev.}
36: \def\PREP{\em Phys. Rep.}
37: \def\PRC{{\em Phys. Rev.} C}
38: \def\PRD{{\em Phys. Rev.} D}
39: \def\ZPC{{\em Z. Phys.} C}
40:
41:
42:
43: \topmargin=-5mm
44:
45:
46:
47: \newcommand{\checkthis}{\marginpar{$\Leftarrow$\em Check}}
48: %\def\ppi{$ pp \rightarrow pp\pi ^{\mbox{\small{o}}}$ \ }
49:
50:
51:
52: \newcommand{\boldtau}{\mbox{\boldmath $\tau$}}
53: \newcommand{\boldphi}{\mbox{\boldmath $\phi$}}
54: \newcommand{\boldsig}{\mbox{\boldmath $\sigma$}}
55: \newcommand{\he}{\mbox{$^3{\rm He}$}}
56:
57:
58:
59:
60:
61: \begin{document}
62:
63:
64:
65:
66:
67: \draft
68:
69:
70:
71:
72: \title{New parameterization of the trinucleon wavefunction and its
73: application to the $\pi\, ^3$He scattering length}
74:
75:
76: \author{V. Baru$^{1,2}$, J. Haidenbauer$^1$, C. Hanhart$^1$, and J. A.
77: Niskanen$^3$}
78:
79: \address{$^1$Institut f\"ur Kernphysik, Forschungszentrum
80: J\"ulich, D-52425 J\"ulich, Germany, \\
81: $^2$Institute of Theoretical and Experimental Physics,
82: 117259, B. Cheremushkinskaya 25,\\ Moscow, Russia \\
83: $^3$Department of Physical Sciences, PO Box 64,
84: FIN-00014 University of Helsinki, Finland}
85:
86:
87: %\author{V. Baru$^1$, J. Haidenbauer$^1$, C. Hanhart$^1$, and J. A.
88: %Niskanen$^2$}
89:
90:
91:
92: %\address{$^1$Institut f\"ur Kernphysik, Forschungszentrum
93: %J\"ulich, D-52425 J\"ulich, Germany, \\
94: %$^2$Department of Physical Sciences, PO Box 64,
95: %FIN-00014 University of Helsinki, Finland}
96:
97:
98:
99: %\date{\today}
100:
101:
102:
103: \maketitle
104:
105:
106:
107:
108:
109:
110: \begin{abstract}
111: We present a new parameterization of the trinucleon wavefunction.
112: As a novel feature a separable parameterization for the complete wavefunction
113: is given. In this way
114: any calculation that considers two-body currents only is largely
115: simplified. To demonstrate this we calculate the $\pi\, ^3$He
116: scattering length in chiral perturbation theory.
117: We find reasonable agreement with experimental values
118: inferred from data on level shifts in pionic $^3$He bound states.
119: The relevance of the
120: $\pi$-triton system for an alternative determination of
121: the $\pi N$ scattering lengths is discussed.
122: \end{abstract}
123: \vskip 1cm \noindent
124:
125:
126:
127:
128: \pacs{PACS: 25.80.Ls, 25.80.Hp, 25.10.+s, 21.45.+v}
129:
130:
131:
132: \vskip 2cm
133:
134:
135:
136: %\hfill{FZJ--IKP--TH--2002--11}
137:
138:
139:
140:
141: \newpage
142:
143:
144:
145: \section{Introduction}
146:
147:
148:
149: In spite of the importance of three- and four-nucleon systems
150: as a bridge between the deuteron and heavier nuclei, there is an
151: appalling shortage of serious theoretical work done on meson-nuclear
152: physics with these.
153: One stumbling stone for calculations of meson-nuclear
154: interactions by "intermediate energy physicists" is the need for
155: at least the trinucleon ($^3$He or $^3$H) Faddeev wave functions,
156: which normally exist in the form of numerical tables computed by a different
157: society, low energy physicists. The necessity of a more accessible
158: form for nonspecialists
159: was realized a long time ago by Hajduk {\it et al.} \cite{hgs},
160: who presented the Faddeev amplitudes of the trinucleon wave function
161: for different channels and
162: particle permutations as separable analytical forms. The aim
163: of this work is to improve the parameterization of Ref. \cite{hgs}
164: in three respects:
165: \begin{itemize}
166: \item The parameterization is given for trinucleon wave functions derived
167: from two modern nucleon--nucleon interactions, namely the
168: CD Bonn \cite{bonn} and the Paris potential \cite{paris}.
169: \item We give a separable expansion for the full antisymmetrized wave function.
170: Previous
171: works \cite{hgs} parameterized the Faddeev components only.
172: \item The trinucleon wave function is not separable in all partial waves over
173: the full momentum range. Therefore we include an additional term in the
174: parameterization that allows the inclusion of correlations. Especially for
175: the pair wave functions where the $NN$ pair is in a
176: $\, ^3d_1$ state this turns out to be crucial.
177: \end{itemize}
178:
179:
180:
181: In the early 1990's Weinberg argued, that as long as we restrict
182: ourselves to interactions with pions only, three and more body
183: interactions are suppressed. Thus, for those calculations,
184: carried out either in chiral perturbation theory or within
185: a phenomenological
186: approach, all that is needed is an appropriate parameterization of
187: the nuclear wave function, which contains the momentum distribution
188: of one active pair only. All remaining degrees of freedom can be
189: integrated/summed separately. If this parameterization is given
190: for the full antisymmetrized wavefunction, to do the actual
191: calculation of a nuclear matrix element with a two-body operator
192: will be just as complicated as the evaluation of the matrix element
193: on the deuteron. In this paper we demonstrate that this program
194: can be easily carried out for the example of the $\pi\, ^3$He scattering length.
195:
196:
197: The paper is organized in the following way:
198: Some basic information about the trinucleon wave function and
199: details of the procedure how the wave function is parameterized by means
200: of simply analytical functions are given in the next section. In the
201: third section the usefulness of the new parameterization is demonstrated
202: by an explicit calculation of the $\pi \, ^3$He scattering length within
203: chiral perturbation theory. The paper ends with a short summary.
204:
205:
206:
207:
208: \section{Parameterization of the trinucleon wave function}
209: The full antisymmetric wave function of a three-nucleon
210: system, can be presented as a sum of three different Faddeev
211: components, each corresponding to different particle permutations and
212: projected on a particular set of partial waves \cite{Gloeckle}:
213: %
214: %
215: % separated into various channels (partial waves)
216: %$|(LS)J,(l\frac{1}{2})j;\frac{1}{2}\rangle$ with different
217: %permutations of the three particles as
218: \begin{equation}
219: |\Psi \rangle = |\psi [(12)3]\rangle +
220: |\psi [(23)1]\rangle + |\psi [(31)2]\rangle .
221: \label{faddeev}
222: \end{equation}
223: The individual components read for instance in coordinate space
224: \begin{equation}
225: \psi^\nu (r_{ij},\rho_k)
226: = \langle r_{12}\rho_3\nu_{12} |\psi [(12)3]\rangle
227: = \langle r_{23}\rho_1\nu_{23} |\psi [(23)1]\rangle
228: = \langle r_{31}\rho_2\nu_{31} |\psi [(31)2]\rangle ,
229: \end{equation}
230: where $r$ and $\rho$ denote the pair and spectator coordinates,
231: respectively, and the index $\nu$ denotes the relevant quantum numbers.
232: In Ref. \cite{hgs} these Faddeev components were presented
233: in separable form $v^\nu(r_{ij})w^\nu(\rho_k)=\psi^\nu (r_{ij},\rho_k)$.
234:
235:
236:
237:
238:
239: A parameterization of the individual Faddeev components is still rather clumsy
240: to use in situations where e.g. a two-body operator should act
241: on a term with a "wrong" grouping, i.e. the operator and the
242: pair wave function $v^\nu (r_{ij})$ do not involve the same particles.
243: However, the action of a two-body operator on a completely antisymmetric
244: state does not depend on the particle identifications
245: in the operator. Therefore, we
246: parameterize the total antisymmetrized wave function
247: in terms of
248: pair and spectator coordinates (or momenta) projected on
249: different angular momentum eigenstates. We choose
250: the particle pair $(12)$ as the active pair and define
251: \begin{eqnarray}
252: \nonumber
253: v^\nu(r)w^\nu(\rho) &=& \langle r\, \rho \, \nu|\Psi\rangle \\
254: \nonumber
255: &=& \langle r \rho \nu |\psi [(12)3]\rangle
256: +\sum_{\nu_{23}}\langle r\rho\nu|r_{23}\rho_{1}\nu_{23}\rangle
257: \langle r_{23}\rho_1\nu_{23} |\psi [(23)1]\rangle \\
258: & & \hskip 2.65cm
259: +\sum_{\nu_{31}}\langle r\rho\nu|r_{31}\rho_{2}\nu_{31}\rangle
260: \langle r_{31}\rho_2\nu_{31} |\psi
261: [(31)2]\rangle \ .
262: \label{def}
263: \end{eqnarray}
264: Here the $\langle r\rho\nu|r_{ij}\rho_{k}\nu_{ij}\rangle$ denote the
265: necessary recoupling coefficients for angular momentum, spin and isospin
266: and the $r_{ij}$ and $\rho_k$ are to be expressed in terms of $r$ and $\rho$
267: \cite{Gloeckle}.
268: E.g. one finds $\vec r_{23}=-\frac{1}{2}\vec r - \vec \rho$.
269: %???In what follows the channel index on the functions
270: %$v(r)$ and $w(\rho)$ is omitted to avoid cluttering of the notation,
271: %if confusion is not expected.???
272:
273:
274:
275: Physically one
276: might expect that, if in a bound three-body system either the
277: spectator or the pair is far off-shell, then it would be less
278: likely to find also the other far off-shell. In other words,
279: there should be some kind of correlation between the momenta
280: or the corresponding coordinates, which is not present in the
281: simple product. In short: there is no reason to expect the
282: triton wave function to be separable.
283: We shall allow for such a correlation by using a fit with a sum of two
284: products. Over a large momentum range this seems to give
285: sufficiently accurate results.
286:
287:
288:
289: In order to get converged results for the trinucleon binding
290: energy and the wave function, in principle, a fairly large
291: number of three-nucleon partial waves is needed \cite{Friar}. However,
292: most of those partial waves give only a small contribution to the
293: total wave function, cf., e.g., Ref. \cite{Schadow}.
294: Therefore, in the following treatment we shall
295: restrict ourselves to the most important states, i.e. those
296: where the $NN$ pairs are in the singlet spin state
297: $^1s_0$ and in the triplet states $^3s_1$ or $^3d_1$.
298: In this case the trinucleon wave function $\psi^\nu$ has only
299: five components whose quantum numbers are summarized in
300: Table \ref{Qnum}. In line with this restriction we also
301: include only those five components in the evaluation of the
302: total wave function $|\Psi\rangle$ according to Eq.~(\ref{def}).
303: The higher partial waves induced by the antisymmetrization have much
304: less weight. We should mention, however, that they still
305: contribute to the overall normalization \cite{Afnan}.
306: But, for convenience, the total wave functions employed in the present
307: work are normalized in such a way that the sum over the five considered
308: channels adds up to unity.
309:
310:
311:
312: The procedure for the fit is the following. First the full Faddeev
313: wave function (using the CD Bonn \cite{bonn} and
314: Paris potential \cite{paris}) is obtained numerically in momentum space,
315: taking into account the permutation of the particles (c.f. Eq. (\ref{def})),
316: and then projected on the 5 considered partial waves (channels).
317: This gives the
318: "exact" wave functions $\Psi^{\nu}(p,q)$ to be fitted in each of the
319: 5 channel considered ($^1s_0S$, $^3s_1S$, $^3d_1S$,
320: $^3s_1D$ and $^3d_1D$, cf. Table \ref{Qnum})
321: as functions of the momenta
322: $p$ and $q$ for the pair and the spectator, respectively.
323: Each of these is approximated by a product
324: of functions $v^{\nu}_1(p)$ and $w^{\nu}_1(q)$ given by a
325: five-term expansion of Lorentz functions,
326: \begin{equation}
327: v^{\nu}_{\lambda}(p) = \sum_{n=1}^5
328: \frac{a^{\nu}_{n,\lambda}}{p^2+(m^{\nu}_{n,\lambda})^2}\,
329: ,\qquad
330: w^{\nu}_{\lambda}(q) = \sum_{n=1}^5
331: \frac{b^{\nu}_{n,\lambda}}{q^2+(M^{\nu}_{n,\lambda})^2} \, .
332: \end{equation}
333: Here for each $\nu$ and $\lambda$ the conditions
334: \begin{equation}
335: \sum_{n=1}^5 a^{\nu}_{n,\lambda} = \sum_{n=1}^5 b^{\nu}_{n,\lambda} = 0
336: \label{asum}
337: \end{equation}
338: are required for convergence in momentum space and
339: to guarantee regularity at the origin in coordinate space
340: (for $D$ waves more conditions are needed \cite{hgs,Lacombe}).
341: %This is similar to Ref. \cite{hgs}, except that w
342: We perform
343: a standard $\chi^2$ fit to the total antisymmetrized exact
344: (numerical) three-nucleon wave functions by minimizing
345: the function
346: \begin{equation}
347: \int_0^{\infty} dp\; dq\; p^2q^2 |\Psi^{\nu}(p,q)-v^{\nu}_1(p) w^{\nu}_1(q)|^2.
348: \end{equation}
349: Like Hajduk et al. \cite{hgs} we include the additional
350: weight factor $p^2q^2$, for this factor
351: emphasizes more the relevant momentum range, as factors $p^2$ and $q^2$ always
352: appear naturally in any matrix element as they originate from the
353: integration measure.
354: After the first fit, where only the term $v^{\nu}_1(p) w^{\nu}_1(q)$
355: is included, an additional
356: product term $v^{\nu}_2(p) w^{\nu}_2(q) $ is fitted to the remaining
357: deviation between this fit and the exact wave function.
358: Thus we have
359: \begin{equation}
360: \Psi^{\nu}(p,q)=v^{\nu}_1(p)w^{\nu}_1(q)+v^{\nu}_2(p)w^{\nu}_2(q).
361: \label{twoterms}
362: \end{equation}
363: In this work we stop after the second term. However, it should be clear
364: that every additional term included systematically improves the
365: fit without influencing the previous terms.
366:
367:
368:
369: We found that the inclusion of two terms is sufficient to reproduce
370: the "exact" wave function well over a reasonably large momentum range.
371: This is illustrated in Fig. \ref{erg}, where the one and two term
372: fits are compared to the "exact" wave function in the $^3s_1S$ as well
373: as the $^3d_1S$ channel for two different spectator momenta.
374: In particular for the $^3d_1S$ case the need for a second term is
375: striking.
376: Another way of testing the quality of the analytical
377: parameterization is offered by
378: the probability for the different components of the
379: wave function in the Blatt-Derrick representation \cite{Schadow}.
380: Corresponding results, for the CD Bonn as well as for the Paris
381: potential, are compiled in Table \ref{probab}.
382: Here the apparent non-separability of the $^3d_1S$ is reflected
383: in the $D$-state probability $P_D$. With a single-term
384: fit there is still a significant deviation of the value fitted for
385: the CD Bonn potential (Paris potential) 6.995\% (8.319\%) from
386: the exact one 7.127\% (8.428\%), while the two-term fit has
387: essentially converged to the exact result with a significantly
388: better value 7.082\% (8.386\%), cf Table \ref{probab}.
389: %
390: %In addition to the $\chi^2$ we checked the fit by calculating
391: %also the overlaps
392: %\begin{equation}
393: %I = \int p^2dp\, q^2dq\, v(p) w(q) \psi_\nu (p,q)
394: %\end{equation}
395: %similar to those of Hajduk {\it et al.} and found them to be
396: %comparable or, in $S$-waves, even better.
397:
398:
399:
400: In the coordinate space representation these approximate wave
401: functions will consist of sums of Yukawa functions and
402: (for $D$ waves) their derivatives
403: \begin{eqnarray}
404: \nonumber
405: V^{\nu}_{\lambda}(r) &=& \sqrt{\frac \pi 2}\,\sum_{n=1}^5 a^{\nu}_{n,\lambda}\,
406: \, e^{-m^{\nu}_{n,\lambda}r}\quad
407: {\rm or}\\
408: V^{\nu}_{\lambda}(r) &=& \sqrt{\frac \pi 2}\,\sum_{n=1}^5 a^{\nu}_{n,\lambda}\,
409: \, e^{-m^{\nu}_{n,\lambda}r}
410: \left( 1+\frac{3}{m^{\nu}_{n,\lambda}r}+\frac{3}{(m^{\nu}_{n,\lambda}r)^2}\right)\, ,
411: \end{eqnarray}
412: %\begin{equation}
413: %v(r) = \sqrt{\frac \pi 2}\,\sum_{n=1}^5 a_{n}\,
414: %(m_{n})^2\, e^{-m_{n}r}\quad
415: %{\rm or} \quad
416: %v(r) = \sqrt{\frac \pi 2}\,\sum_{n=1}^5 a_n\, m_n^2\, e^{-m_n r}
417: %\left( 1+\frac{3}{m_nr}+3\frac{3}{m_n^2r^2}\right)\, ,
418: %\end{equation}
419: with similar expressions for the spectator $\rho$ dependence.
420: As usual, in coordinate space an additional factor of $r$ is included
421: in the definition of the wave functions ($V$ is $r$ times the
422: Fourier transform of $v$).
423:
424:
425:
426: Table \ref{tablebonn} contains the parameters for the fit to the wave
427: functions using the CD Bonn potential. It may be noted that
428: there is some freedom in distributing the strength between the
429: pair functions $v^{\nu}_{\lambda}$ and the spectator $w^{\nu}_{\lambda}$.
430: This freedom is used to present
431: the results in such a normalization that the square of
432: the first term pair function integrates to unity, i.e.
433: $\int_0^{\infty} dp\; p^2 (v_1^{\nu})^2=1$ for each partial wave $\nu$.
434: Table \ref{tableparis} gives the parameters of the triton
435: wave function fit for the Paris potential. The parameters
436: are available from the authors by e--mail\footnote{Either from
437: J. Niskanen (Jouni.Niskanen@helsinki.fi) or from J. Haidenbauer
438: (j.haidenbauer@fz-juelich.de).}.
439:
440: As mentioned before, the parametrized wave functions $\Psi^{\nu}(p,q)$ are
441: normalized in the usual way, i.e.
442: \begin{equation}
443: \sum_{\nu}\int_0^{\infty} dp\; dq\; p^2q^2 |\Psi^{\nu}(p,q)|^2=1,
444: \label{norm}
445: \end{equation}
446: for the sum over the five channels considered. However, for completeness
447: reasons we would like to point out
448: that in a fully converged antisymmetrization of the wave function, with
449: sufficient higher partial wave included, those five
450: channels would have a relative weight of 0.954 (CD Bonn) and
451: 0.945 (Paris), respectively.
452: We would like to stress that the remaining 5 \% are saturated by a large number
453: of partial waves that are individually rather small (c.f. Ref. \cite{Schadow})
454: and thus should be of minor relevance for the calculation of observables.
455:
456:
457: Fig.
458: \ref{coordspwf} shows the probability distributions
459: (pair correlations)
460: \begin{equation}
461: P(r) \equiv \int_{0}^{\infty} |\Psi^{(^3s_1S)}(r,\rho)|^2 d\rho
462: \end{equation}
463: integrated over the spectator degrees of freedom. For
464: definiteness, these are all normalized to unity.
465: Here, for comparison with the earlier work \cite{hgs}
466: (dotted line), we also include the fits to the
467: individual Faddeev amplitude (dashed line). As one
468: might expect, these results are rather similar to
469: each other, while the antisymmetrized wave function
470: gives a significantly longer ranged distribution
471: (solid line). It is
472: interesting that the short-range node in the wave functions
473: of Ref. \cite{hgs} is also present in our single-channel fits
474: but not in the full antisymmetric wave function. For
475: completeness, Table \ref{tablesingle} gives the parameterization
476: fully analogous to Ref. \cite{hgs} used to produce the dashed
477: curve. With the exception of the dotted line, the results in this figure
478: are based on the CD Bonn potential.
479:
480:
481:
482:
483:
484:
485:
486: \section{The $\pi\, ^3$H\lowercase{e} scattering length}
487:
488:
489:
490:
491:
492:
493:
494: Now we want to demonstrate the usefulness of the new parameterization by
495: an explicit calculation of the real part of the $\pi\, ^3$He scattering length in chiral
496: perturbation theory. Since
497: all the complications of having to treat a three-nucleon problem
498: were already solved on the level of deriving the wave function
499: parameterization in terms of an active pair and a spectator
500: (c.f. Eq. (\ref{def})),
501: the remaining part of the calculation for the scattering length is
502: as complicated as the corresponding calculation for the two nucleon
503: system (cf., e.g., Refs. \cite{Ulf,bk,Ericson} where the $\pi d$
504: scattering length was
505: calculated), as long as only two nucleon operators are
506: included as is the case for the NLO calculation we are
507: going to perform here. Note, however, that with the parameterization
508: given also the explicit evaluation of three body forces is
509: largely simplified. In addition, the very simple form of the terms in the separable
510: expansion allows an analytical calculation---the final result can
511: be expressed in terms of incomplete $\Gamma$-functions.
512:
513:
514:
515:
516:
517: We will work with the so called hybrid approach, introduced by Weinberg
518: \cite{Weinberg}, namely the $\pi NNN \to \pi NNN$ transition operators,
519: evaluated using chiral perturbation theory, will be convoluted with
520: phenomenological wavefunctions using the parameterizations given in the
521: previous section. It should be stressed that recently three-body wavefunctions
522: were derived utilizing $NN$ and $3N$ forces that are
523: consistent with the chiral counting scheme \cite{Epelbaum1}.
524: As will become clear below, once a parameterization in the form presented
525: in this paper for those wavefunctions is available as well, it is a simple
526: task to extract the corresponding $\pi \, ^3$He scattering length.
527:
528:
529:
530:
531: Within the counting scheme advocated by Weinberg all that contributes to
532: $\pi\, ^3$He scattering at leading and
533: next--to--leading order are one- and two-body currents, as illustrated
534: in Fig. \ref{diafull}. Due to the Goldstone nature of the
535: pions all three-body currents are suppressed by an additional
536: factor $p^2/\Lambda^2$, where $p$ denotes the typical momentum of
537: the problem and $\Lambda$ denotes the chiral symmetry breaking scale
538: ($\simeq 1$ GeV)
539: \cite{Weinberg}.
540:
541:
542:
543:
544: In order to have a theory with predictive power it is compulsory that
545: unknown counter terms
546: are either suppressed or at least small in number.
547: Naturally,
548: in different schemes those terms appear at different places. This is not only true
549: for the three-body system itself, where in a pion--less approach three-body
550: forces appear at leading order \cite{paulo}, but also for the scattering
551: process. For example in Ref. \cite{harald} it was stressed, that in an approach
552: using perturbative pions there appear counterterms in the production
553: operator already
554: at leading order in a calculation for $\pi d$ scattering. However, in
555: this paper we will not elaborate on those subtleties any further but
556: will use the approach of Ref. \cite{Weinberg}.
557:
558:
559: It is a straightforward task to relate the
560: one body contribution to the scattering length. One gets
561: \begin{equation}
562: a^{(1b)}_{\rm He}= \kappa
563: \left( Aa^{(+)}-Q_\pi \, a^{(-)}2T_3 \right) \ ,
564: \label{finala1}
565: \end{equation}
566: where $\kappa = \left(1+\frac{m_\pi}{m_N}\right)/\left(1+\frac{m_\pi}{Am_N}\right)$.
567: Here $a^{(+)}$ ($a^{(-)}$) denote the isoscalar (isovector) $s$--wave $\pi N$
568: scattering length, $A$ is the number of nucleons in the nucleus (here
569: $A=3$), $Q_\pi$ is the charge of the pion in units of $e$ and $T_3$ denotes
570: the third component of the isospin of the nucleus.
571: Since the interaction is momentum independent, the loops occurring are
572: identical to the expression of the wave function normalization.
573: In contrast to the case of pion scattering on the deuteron, here
574: the isovector piece of the scattering length does contribute and dominates
575: the one body contribution. We thus get for the one body contribution of
576: the scattering of negatively charged pions on $\, ^3$He
577: \begin{equation}
578: a^{(1b)}_{\rm He} = (92 \pm 15)\times 10^{-3} \ m_\pi^{-1} \ .
579: \label{1b}
580: \end{equation}
581: where we used $a^{(-)}=(90.5\pm 4.2)\times 10^{-3} \ m_\pi^{-1}$ and
582: $a^{(+)}=(-2.2\pm 4.3)\times 10^{-3} \ m_\pi^{-1}$.
583: %
584: These are the purely hadronic values presented in Ref. \cite{schr}, where
585: Coulomb effects as well as isospin breaking effects are already subtracted.
586: Note, that a consistent next--to--leading order calculation would require
587: to use $\pi N$ scattering lengths extracted from a calculation for
588: $\pi N$ scattering carried
589: out to the corresponding order as given, e.g., in Ref. \cite{nadia}.
590: In addition, a careful analysis would also
591: necessitate to take into account isospin breaking effects consistently,
592: as was stressed in Ref. \cite{silas}.
593: One expects, however, that especially the latter point enlarges primarily
594: the theoretical error.
595: Anyway, here we will ignore such subtleties because we only want
596: to investigate, more qualitatively, if the $\pi \, ^3$He system
597: allows to extract information on the $\pi N$ scattering lengths.
598:
599:
600:
601: Now we need to evaluate the matrix element of the wave function
602: with the two-body scattering kernel
603: ${\cal A}(\vec p \, ',\vec p)$, where $\vec p$ ($\vec p \, '$) denote
604: the relative momentum of the active nucleon pair in the initial (final)
605: state, respectively.
606: Thus, we can write
607: \begin{equation}
608: a^{(2b)}_{\rm He}=\frac{1}{4\pi}{A \choose
609: 2}\left(\frac{1}{1+\frac{m_\pi}{Am_N}}\right)
610: \int d^3p'd^3 p d^3 q\Psi (\vec p \,', \vec q)^\dagger
611: {\cal A}(\vec p \, ',\vec p)\Psi (\vec p, \vec q) \ .
612: \label{a2b}
613: \end{equation}
614: Here we already used the property that the wave function is antisymmetrized.
615: Thus, whichever of
616: the three nucleons is the spectator, the result of the matrix element
617: is the same. Therefore, inclusion of all possible pairwise
618: interactions just leads to an overall factor of 3. This factor can
619: be easily generalized to the case of a nucleus composed of $A$ nucleons.
620: Then there is a combinatorial factor which
621: is given by the binomial coefficient ${A \choose 2}$. If we now use the
622: separable form presented in the previous section, we get
623: \begin{equation}
624: a^{(2b)}_{\rm He}=\frac{1}{4\pi}{A \choose 2}
625: \left(\frac{1}{1+\frac{m_\pi}{Am_N}}\right)
626: \sum_{\nu \nu '}\sum_{\lambda ' \lambda}{\cal W}^{\nu' \nu}_{\lambda ' \lambda}
627: \int d^3p'\, d^3 p \, v (p \, ')_{\lambda '}^{\nu '}
628: \langle \nu ' ,\hat p \, '|{\cal A}(\vec p \, ',\vec p)| \nu , \hat p \rangle v (
629: p)_\lambda^{\nu} \ ,
630: \end{equation}
631: where
632: \begin{equation}
633: {\cal W}^{\nu ' \nu}_{\lambda ' \lambda} = \int d^3 q\, w(q)_{\lambda '}^{\nu '}
634: w(q)_\lambda^{\nu}
635: \langle \nu ' ,\hat q| \nu , \hat q \rangle \ .
636: \end{equation}
637: Here $\lambda$ and $\lambda '$ denote the terms of the separable ansatz,
638: $\nu '$ as well as $\nu$ denote the partial waves for the final
639: and initial state, respectively,
640: and the different state vectors $|\nu, \hat p \rangle$
641: contain the relevant spin, isospin
642: and angular momentum components of the corresponding partial wave.
643:
644:
645:
646:
647: For the two body scattering kernel, as shown diagrammatically in Fig.
648: \ref{dia}, we use \cite{Ulf}:
649: \begin{eqnarray}
650: \nonumber
651: {\cal A}(\vec p \, ',\vec p)
652: &=& \frac{1}{(2\pi)^3}\frac{m_\pi^2}{4F_\pi^4}
653: \left\{ \frac{1}{\vec q {}\, ^2}\,
654: [2\delta^{(ac)}(\vec \tau ^{(1)}\cdot \vec \tau^{(2)})-
655: \tau ^ {(1) a}\tau ^{(2) c}
656: - \tau ^ {(1) c}\tau ^{(2) a}]\right. \\
657: &-& \left. g_A^2\frac{(\vec \sigma ^{(1)} \cdot \vec q)
658: (\vec \sigma ^{(2)} \cdot \vec q)}{(\vec q {}\, ^2+m_\pi^2)^2}
659: \, [\delta^{(ac)}(\vec \tau ^{(1)}\cdot \vec \tau^{(2)})-
660: \tau ^ {(1) a}\tau ^{(2) c}
661: - \tau ^ {(1) c}\tau ^{(2) a}]\right\} \ ,
662: \end{eqnarray}
663: where the superscripts $a$ ($c$) denote the isospin of the incoming (outgoing) pion,
664: and $\vec q = \vec p\, ' - \vec p$ denotes the pion momentum.
665: It should be stressed that the same kernel was used recently to extract
666: the isoscalar $\pi N$ scattering length from $\pi d$ scattering data \cite{Ulf}.
667: Thus---as long as we restrict ourselves to $S$--waves in the bound
668: state wave functions---all we need to evaluate are the following two
669: integrals
670: \begin{eqnarray}
671: I_1(m_i,m_j) &=& \frac{1}{4\pi} \int d^3p\, d^3p'\,
672: \frac{1}{\vec p {}\, ^2+m_i^2}\,\frac{1}{(\vec p-\vec p \, ')^2}\,
673: \frac{1}{\vec p \, ' \, ^2 +m_j^2} \\
674: I_2(m_i,m_j) &=& \frac{1}{4\pi} \int d^3p\, d^3p'\,
675: \frac{1}{\vec p {}\, ^2+m_i^2}
676: \frac{(\vec p-\vec p \, ')^2}{((\vec p-\vec p \, ')^2+m_\pi^2)^2}\,
677: \frac{1}{\vec p \, ' \, ^2 +m_j^2} \ .
678: \end{eqnarray}
679: Obviously the inclusion of $D$ waves can be done easily. Since the
680: formulas are rather lengthy we do not give them here explicitly. It
681: should be stressed that
682: the contribution due to the $D$ waves did not exceed 10\%, which is of
683: the order of the theoretical uncertainty induced by both
684: the convergence rate of the chiral expansion and the model
685: dependence of the scattering length caused by the particular employed
686: nuclear wave function. We find
687: \begin{eqnarray}
688: I_1(m_i,m_j) &=& \pi^3\Gamma ( (m_i+m_j)\epsilon,0)
689: \label{int1}
690: \\
691: I_2(m_i,m_j) &=& \pi^3\left\{ \Gamma
692: ((m_i+m_j+m_\pi)\epsilon,0)
693: -\frac{1}{2}\left(\frac{m_\pi}{m_i+m_j+m_\pi}\right) \right\} \ ,
694: \label{int2}
695: \end{eqnarray}
696: where we used the incomplete $\Gamma$-function defined through
697: $\Gamma (\lambda , n)
698: =\int_\lambda^\infty dx\, x^{n-1}\exp (-x)$ \cite{aseg}.
699: The constant $\epsilon$ was introduced to render the integrals finite.
700: Eq. (\ref{asum}) implies that the results in
701: Eqs. (\ref{int1}) and (\ref{int2}) are independent of $\epsilon$ as long as
702: $m_i \epsilon \ll 1$ for all $i$.
703: %, since
704: %we get for small $x$
705: %\begin{eqnarray}
706: %\nonumber
707: %\sum_{mn}a_m^ja_n^i e^{-x(m_m^j+m_n^i)}&=&
708: %(\sum_m a_m^j e^{-xm_m^j}) (\sum_n a_n^i e^{-xm_n^i}) \\
709: %\nonumber
710: %&\simeq& (\sum_m a_m^j(1-xm_m^j)) (\sum_n a_n^i(1-xm_n^i)) \\
711: %&=& x^2(\sum_m a_m^jm_m^j)(\sum_n a_n^im_n^i)) \ .
712: %\end{eqnarray}
713: %
714: %
715: %
716:
717:
718:
719:
720:
721: We thus get for the two-body contribution to the $\pi \, ^3$He scattering
722: length
723: \begin{eqnarray}
724: \nonumber
725: a^{(2b)}_{\rm He} &=&{A \choose 2}\left(\frac{1}{1+\frac{m_\pi}{Am_N}}\right)
726: \left(\frac{m_\pi^2}{32\pi^4F_\pi^4}\right)
727: \sum_{\nu }\sum_{\lambda ', \lambda=1}^2{\cal W}^{\nu \nu}_{\lambda ' \lambda}
728: \left(\frac{2}{3}T_{\nu}\left(T_{\nu}+1\right)-1\right) \\
729: & \times & \sum_{m,n=1}^{5} a_{m,\lambda '}^{\nu}a_{n,\lambda}^{\nu}
730: \left\{I_1(m_{m,\lambda '}^{\nu}
731: ,m_{n, \lambda}^{\nu})
732: -\frac{g_A^2}{4}\left(\frac{2}{3}S_{\nu}\left(S_{\nu}+1\right)-1\right)
733: I_2(m_{m,\lambda '}^{\nu},m_{n,\lambda}^{\nu})\right\} .
734: \label{finala2}
735: \end{eqnarray}
736: Again, only the formula for the $S$--wave part of the wavefunction
737: is given explicitly, although the effect of the $D$--waves is
738: included in the final result. Under this restriction, the partial
739: wave of the initial $NN$ pair is equal to that of the outgoing pair.
740:
741:
742: Using the parameters for the two different wave functions as listed
743: in Tables \ref{tablebonn} and \ref{tableparis},
744: we get for the two-body contributions to the real part of the scattering length
745: \begin{equation}
746: a^{(2b)}_{\rm He} = \left\{
747: \begin{array}{l} -24\times 10^{-3}\; m_\pi^{-1} \ {\rm (Paris)} \\
748: -26\times 10^{-3}\; m_\pi^{-1} \ {\rm (CD\; Bonn)} \ . \end{array} \right.
749: \end{equation}
750: Compared to the leading one-body term we thus find reasonable convergence.
751: Also the dependence on the wave function used can be regarded as a
752: higher order effect.
753: As the final result we therefore have
754: \begin{equation}
755: {\rm Re}(a_{\rm He}^{\rm (theo)}) = (67 \pm 15)\times 10^{-3} \ m_\pi^{-1} \ .
756: \label{theo}
757: \end{equation}
758: The error included is only that stemming from the error on the
759: $\pi N$ scattering lengths used (c.f. Eq. (\ref{1b})). As mentioned
760: before, for a reliable estimate of the theoretical error a
761: consistent calculation including isospin breaking effects as well
762: as using consistent wave functions would
763: be necessary. One could regard the size of the NLO contribution as
764: a rather conservative estimate for the theoretical error.
765:
766:
767:
768: Information on the $\pi \, ^3$He scattering length
769: can be inferred from experimental information on level shifts caused
770: by the strong interaction in the bound state energies of
771: $\pi^- {}\, ^3$He atoms \cite{phe1,phe2,phe3}.
772: A list of results from different experiments available
773: in the literature is given in table \ref{scltab}. From those
774: energy shifts scattering lengths
775: can be easily extracted using the so called Deser formula \cite{deser},
776: \begin{equation}
777: \epsilon_{1s}=-2(Z \alpha)^3\mu^2\, {\rm Re}\left( a_{\rm He}^{\rm (exp)} \right) \ .
778: \label{scl}
779: \end{equation}
780: where $Z$ is the charge number of the nucleus and $\mu$ is the reduced
781: mass of the $\pi\, ^3$He system.
782: Taking the arithmetic mean value for the experimental results
783: given in table \ref{scltab}, we thus get
784: \begin{equation}
785: {\rm Re}(a_{\rm He}^{\rm (exp)}) = (47 \pm 13)\times 10^{-3} \ m_\pi^{-1} \ ,
786: \label{exp}
787: \end{equation}
788: where the error includes both the spread in the experimental numbers
789: as well as an additional 10\% to account for the omission of
790: electromagnetic and isospin breaking corrections in Eq. (\ref{scl}).
791: Note that in case of pionic hydrogen those corrections turned out to be of
792: the order of 7\% \cite{gasser}. Thus, we find agreement between
793: theory and experiment within the given uncertainties.
794:
795:
796:
797: As was argued above, at the current stage we cannot give a reliable
798: error estimate for the theoretical calculation.
799: Naturally, further studies are needed to
800: draw stronger conclusions. However, it should be stressed, that
801: it would be very useful to push the calculation as well as the
802: experiment for the $\pi \, ^3$He system to as high an accuracy as those for the
803: $\pi$d system, since remaining discrepancies between theory and
804: experiment in such a combined analysis can only stem from three-body
805: currents and would therefore be an important test of chiral perturbation
806: theory in few body systems.
807:
808:
809: In this context let us also emphasize
810: that an experimental study of pionic tritium would
811: be of very high interest because a combined analysis of the $\pi \, ^3$He
812: and the $\pi t$ systems promises to provide direct access to the
813: pion--nucleon scattering lengths. From Eqs. (\ref{finala1}) and (\ref{finala2}) we
814: find
815: \begin{eqnarray}
816: a_t+ a_{\rm He} &=&
817: 6\kappa a^{(+)}+{\cal O}\left(\left(\frac{m_\pi}{M_N}\right)^3\right) \ ,
818: \nonumber \\
819: a_t- a_{\rm He} &=&
820: 2\kappa Q_\pi \, a^{(-)} + {\cal
821: O}\left(\left(\frac{m_\pi}{M_N}\right)^4\right)
822: \ ,
823: \label{asad}
824: \end{eqnarray}
825: where the constant $\kappa$ was defined in Eq. (\ref{finala1}).
826: Note, that for the extraction of $a^{(-)}$ one should expect the
827: corrections to be suppressed by one chiral order compared to that
828: of $a^{(+)}$. This is the case, because the leading corrections,
829: that in the counting advocated by Weinberg
830: appear at ${\cal O}((m_\pi /M_N)^3)$---as given
831: in Eq. (\ref{finala2}) for only $s$--waves in the bound state wavefunction---do
832: not distinguish between the $t$ and $^3$He system
833: and thus do not contribute to the difference of the corresponding scattering lengths.
834: Indeed a study of $\pi^- t$ bound states appears
835: experimentally feasible nowadays \cite{gotta}. It should be stressed,
836: that Eqs. (\ref{asad}) are general and independent of the wave functions used.
837: The error estimate only requires that the Weinberg scheme is applicable to the
838: reactions under investigation. In addition, it should be clear
839: that the $^3$He/$t$ system is an ideal system to study isospin breaking effects
840: in few nucleon systems.
841:
842:
843: It turns out that for the $\pi\, ^3$He scattering
844: the omission of the second term in the expansion of the wave function
845: (c.f. Eq. (\ref{twoterms})) leads to a change
846: in the two-body contribution of the scattering length by 2\% only.
847: Thus, in the energy range relevant
848: for $\pi$ scattering at threshold, the full wave function is well
849: described by a one-term separable form.
850: As indicated in Fig. \ref{erg} in reactions with either large momentum
851: transfer or where higher partial waves become relevant, the effect
852: of the second term is expected to be much more pronounced.
853: Indeed, this has been already seen in a recent application of the
854: wave function parametrization to $\pi$ absorption
855: on $^3$He \cite{piabs}.
856:
857:
858:
859:
860:
861: It is illuminating to compare Eq. (\ref{finala2}) with that for
862: the two-body contribution in the case of pion-deuteron scattering.
863: Obviously, in this situation it does not make sense to use two different
864: terms expanded individually (thus the summation on $\lambda$ and
865: $\lambda '$ disappears). In addition, there is no third particle (thus
866: ${\cal W}^{\nu \nu}_{\lambda ' \lambda}=1$). All the rest remains
867: unchanged. If we now use for the deuteron wave function a
868: parametrization that was provided in Ref. \cite{Machleidt}
869: for the Bonn B $NN$ model we reproduce the
870: result obtained in Ref. \cite{Ulf}.
871:
872:
873:
874:
875: In the present work we do not aim at a highly accurate determination
876: of the $\pi\, ^3$He scattering length, but rather at a demonstration of
877: the usefulness of the new parameterization. However, it should be
878: clear that it is straightforward to improve the calculation. Results
879: to order $p^4$ for the $\pi d$ scattering length are given in Ref.
880: \cite{silas} and could be used as the basis for an equally accurate
881: calculation for the $\pi \, ^3$He scattering length.
882:
883:
884: Already in Ref. \cite{Ulf} it was observed, that the contribution from
885: the first diagram on the right hand side in Fig. \ref{dia} is
886: significantly larger than that from the other two, although all three
887: formally appear at the same order in the Weinberg scheme. In Ref.
888: \cite{silas} this was traced back to the small momentum scale
889: introduced by the small binding energy of the deuteron. To account for
890: this a modified counting scheme was advocated, leading to a reordering
891: of the meson exchange diagrams: the first diagram on the right hand
892: side of Fig. \ref{dia} is now the leading two nucleon current, whereas
893: the other two are down by 4 orders in the expansion parameter (c.f.
894: table 4 in Ref. \cite{silas}). It is interesting to note, that
895: in our calculation we find a similar hierarchy of the results for
896: the diagrams shown in Fig. \ref{dia}: the result
897: for the first diagram exceeds the sum of the latter two by a factor of
898: 20 or more, depending on the wavefunction used. Note that, similarly
899: to the case of the deuteron, the typical momentum of a nucleon within
900: $\, ^3$He is significantly smaller than the pion mass.
901:
902:
903:
904: Also a calculation of the imaginary part of the scattering
905: length is, in principle, feasible. However, the contributions to the
906: imaginary part are connected
907: to the pion absorption on $^3$He, where a different counting
908: scheme needs to be applied to account for the large nucleon momenta
909: in the purely nucleonic intermediate state \cite{pwaves}.
910: However, within this scheme the imaginary part should be suppressed
911: by a factor of $(m_\pi/M_N)^{(3/2)}$ compared to Re($a^{2b}$)
912: given in Eq. (\ref{finala2}). The
913: same holds true for the so called dispersive corrections, namely the
914: contribution of the inelastic channels to the real part of the scattering length.
915:
916:
917:
918:
919:
920: \section{Summary}
921:
922:
923:
924: In this paper we have given a new parameterization for
925: two different trinucleon wave functions based on modern
926: $NN$ interaction models. Although the procedure
927: is in spirit similar to an earlier work \cite{hgs},
928: it extends the single-term separability to two terms
929: allowing some correlation between the two relative
930: momenta (or coordinates) and also approximates directly the
931: full antisymmetrized wave function projected onto different
932: partial waves. The former feature might be expected to
933: be useful in phenomena involving high momenta, whereas the
934: latter drastically simplifies calculations involving
935: two-nucleon operators only.
936:
937:
938: Apparently similar benefits are expected in
939: applications to other reactions.
940: E.g., a calculation of quasifree
941: pion absorption on nucleon pairs in $^3$He based on the
942: presented parameterization has been reported in Ref. \cite{piabs}.
943:
944:
945:
946: As a demonstration
947: the parameterization has been applied to calculate the $\pi\, ^3$He
948: scattering length. It has been seen that for such a parameterization
949: the two-body contribution becomes very easy to compute even
950: analytically. The actual calculation done in NLO chiral
951: perturbation theory yields a value of
952: ${\rm Re}(a_{\rm He}) =
953: (67 \pm 15)\times 10^{-3} \ m_\pi^{-1}$, which is in
954: reasonable agreement with experimental values
955: inferred from data on level shifts in pionic $^3$He bound states.
956: We have argued, that a combined analysis of $\pi d$, $\pi \, ^3$He and
957: $\pi t$ scattering should provide both important information on
958: the $\pi N$ $s$--wave scattering lengths as well as
959: a test of the applicability of chiral perturbation theory
960: to few-body systems.
961:
962:
963:
964: \acknowledgements{
965: We would like to thank D. Gotta for providing valuable information
966: on experiments on pionic atoms. Furthermore, we acknowledge
967: useful discussions with I.R. Afnan, H.W. Griesshammer, U.--G. Mei\ss ner,
968: and J. Speth.
969: Financial support for this work was provided in part by the
970: international exchange program between DAAD (Germany, project no.
971: 313-SF-PPP-pz) and the Academy of Finland (project no. 41926) as well as
972: by RFBR (grant no. 02-02-16465).}
973:
974:
975:
976: \begin{thebibliography}{99}
977:
978:
979:
980:
981: \bibitem{hgs} Ch. Hajduk, A. M. Green and M. E. Sainio, Nucl. Phys.
982: {\bf A337}, 13 (1980).
983: \bibitem{bonn} R. Machleidt, Phys. Rev. C {\bf 63}, 024001 (2001).
984: \bibitem{paris} M. Lacombe, B. Loiseau, J.~M. Richard, R. Vinh Mau,
985: J. C\^ot\'e, P. Pir\'es, and R. de Tourreil,
986: Phys. Rev. C {\bf 21}, 861 (1980).
987: \bibitem{Gloeckle} W. Gl\"ockle, {\it The Quantum Mechanical Few-Body
988: Problem}, (Springer, Berlin-Heidelberg-New York 1983).
989: \bibitem{Friar} J.~L. Friar, B.~F. Gibson, and G.~L. Payne
990: Phys. Rev. C {\bf 37}, 2869 (1988).
991: \bibitem{Schadow} W. Schadow, W. Sandhas, J. Haidenbauer, and A. Nogga,
992: Few-Body Systems {\bf 28}, 241 (2000).
993: \bibitem{Afnan} I.~R. Afnan and N.~D. Birrell,
994: Phys. Rev. C {\bf 16}, 823 (1977).
995: \bibitem{Lacombe} M. Lacombe, B. Loiseau, R. Vinh Mau,
996: J. C\^ot\'e, P. Pir\'es, and R. de Tourreil,
997: Phys. Lett. {\bf 101B}, 139 (1981).
998: \bibitem{Ulf}
999: S.~R.~Beane, V.~Bernard, T.-S.~H~Lee and U.~G.~Meissner,
1000: %``Isoscalar S-wave pi N scattering length a+ from pi deuteron scattering,''
1001: Phys.\ Rev.\ C {\bf 57}, 424 (1998).
1002: \bibitem{bk} V. Baru, A. Kudryavtsev, Phys. Atom. Nucl. {\bf 60}, 1475 (1997)
1003: %``$\pi N$ scattering lengths from an analysis of new data on pionic hydrogen and
1004: %deuterium atoms''
1005: \bibitem{Ericson} T. E. O. Ericson, B. Loiseau, and A. W. Thomas,
1006: {arXiv:hep-ph/0009312}; Phys. Rev. C, in print.
1007: %``Determination of the pion-nucleon coupling constant and scattering lengths''
1008: \bibitem{Weinberg}
1009: S.~Weinberg,
1010: %``Three Body Interactions Among Nucleons And Pions,''
1011: Phys.\ Lett.\ B {\bf 295}, 114 (1992).
1012: \bibitem{Epelbaum1}
1013: E.~Epelbaum, A.~Nogga, W.~Gl\"ockle, H.~Kamada, U.~G.~Meissner and H.~Witala,
1014: %``Three-nucleon forces from chiral effective field theory,''
1015: {arXiv:nucl-th/0208023}.
1016: \bibitem{paulo}
1017: P.~F.~Bedaque, H.~W.~Hammer and U.~van Kolck,
1018: %``Effective Theory of the Triton,''
1019: Nucl.\ Phys.\ {\bf A676}, 357 (2000).
1020: \bibitem{harald}
1021: B.~Borasoy and H.~W.~Griesshammer,
1022: %``The S wave pion deuteron scattering length in effective field theory,''
1023: {arXiv:nucl-th/0105048}.
1024: \bibitem{schr} H. C. Schr\"oder {\it et al.}, Eur. Phys. J. {\bf C 21},
1025: 473 (2001).
1026: %\bibitem{Ulf}
1027: %S.~R.~Beane, V.~Bernard, T.~S.~Lee and U.~G.~Meissner,
1028: %``Isoscalar S-wave pi N scattering length a+ from pi deuteron scattering,''
1029: %Phys.\ Rev.\ C {\bf 57}, 424 (1998).
1030: \bibitem{nadia}
1031: N.~Fettes and U.~G.~Meissner,
1032: %``Pion nucleon scattering in chiral perturbation theory. II: Fourth order
1033: %calculation,''
1034: Nucl.\ Phys.\ A {\bf 676} (2000) 311.
1035: \bibitem{silas}
1036: S.~R.~Beane, V.~Bernard, E.~Epelbaum, U.~G.~Meissner and D.~R.~Phillips,
1037: %``The S-Wave Pion-Nucleon Scattering Lengths from Pionic Atoms
1038: %using Effective Field Theory,''
1039: {arXiv:hep-ph/0206219}.
1040: \bibitem{aseg}
1041: see e.g. M. Abramowitz and I.~A. Stegun (Eds.),
1042: {\it Handbook of Mathematical Functions}, (Dover Publications, New York, 1964).
1043: \bibitem{phe1}
1044: R. Abela {\it et al.}, Phys. \ Lett.\ B {\bf 68}, 429 (1977).
1045: \bibitem{phe2}
1046: G.~R. Mason {\it et al.}, Nucl.\ Phys. {\bf A340}, 240 (1980).
1047: \bibitem{phe3}
1048: I. Schwanner, G. Backenstoss, W. Kowald, L. Tauscher, H.-J. Weyer,
1049: D. Gotta, and H. Ullrich, Nucl.\ Phys. {\bf A412}, 253 (1984).
1050: \bibitem{deser}
1051: S. Deser {\it et al.}, Phys. Rev. 96 (1954) 774.
1052: \bibitem{gasser}
1053: V. E. Lyubovitskij and A. Rusetsky,
1054: Phys.Lett. B{\bf 494} (2000) 9; J. Gasser et al.,
1055: {arXiv:hep-ph/0206068}, Eur. Phys. J. C, in print.
1056: \bibitem{gotta}
1057: D. Gotta, private communication.
1058: \bibitem{piabs} S. Schneider, J. Haidenbauer, C. Hanhart, and J. A. Niskanen,
1059: {arXiv:nucl-th/0209051}.
1060: \bibitem{Machleidt}
1061: R.~Machleidt, Adv. Nucl. Phys. 19, 189 (1989).
1062: \bibitem{pwaves}
1063: C.~Hanhart, U.~van Kolck and G.~A.~Miller,
1064: %``Chiral three-nucleon forces from p-wave pion production,''
1065: Phys.\ Rev.\ Lett. {\bf 85}, 2905 (2000);
1066: C.~Hanhart and N.~Kaiser, arXiv:nucl-th/0208050,
1067: Phys. Rev. C, in print.
1068: %``Complete next-to-leading order calculation for pion production in nucleon nucleon
1069: %collisions at threshold,''
1070:
1071:
1072: \end{thebibliography}
1073:
1074:
1075:
1076: \newpage
1077:
1078:
1079:
1080: \begin{table}[ht]
1081: \caption{Quantum numbers of the three-body channels.
1082: $s$, $\tau$, $l$, and $j$ refer to the spin, isospin, orbital
1083: and total angular momentum in the $NN$ subsystem and $L$ and $K$
1084: are the relative orbital angular momentum of the spectator and
1085: the so-called channel spin \protect\cite{Schadow}.}
1086: \label{Qnum}
1087: \begin{center}
1088: \begin{tabular}{ccccccccc}
1089: Channel no. & Label & Subsystem & $l$ & $s$ & $j^\pi$ & $\tau$ & $K$ & $L$ \\
1090: \hline
1091: 1 &$^1s_0S$& $^1s_0$ & $0$ & $0$ & $0^+$ & $1$ & $1/2$ & $0$ \\
1092: 2 &$^3s_1S$& $^3s_1$ & $0$ & $1$ & $1^+$ & $0$ & $1/2$ & $0$ \\
1093: 3 &$^3s_1D$& $^3s_1$ & $0$ & $1$ & $1^+$ & $0$ & $3/2$ & $2$ \\
1094: 4 &$^3d_1S$& $^3d_1$ & $2$ & $1$ & $1^+$ & $0$ & $1/2$ & $0$ \\
1095: 5 &$^3d_1D$& $^3d_1$ & $2$ & $1$ & $1^+$ & $0$ & $3/2$ & $2$ \\
1096: \end{tabular} \end{center}
1097: \end{table}
1098:
1099:
1100:
1101: \begin{table}[ht]
1102: \caption{Probabilities (in the Blatt-Derrick representation) of
1103: the trinucleon wave function components for the CD Bonn and
1104: Paris potentials. ``1 term'' and ``2 terms'' refer to the single- and
1105: two-term parameterizations described in Sect. II.
1106: }
1107: \label{probab}
1108: \begin{center}
1109: \begin{tabular}{lcccc} & P(S) & P(S') & P(P) & P(D) \\
1110: \hline
1111: \hline
1112: CD Bonn 1 term & 91.77 & 1.187 & 0.047 & 6.995 \\
1113: CD Bonn 2 terms & 91.60 & 1.274 & 0.048 & 7.082 \\
1114: \hline
1115: CD Bonn & 91.55 & 1.276 & 0.049 & 7.127 \\
1116: \hline
1117: \hline
1118: Paris 1 term & 90.08 & 1.543 & 0.062 & 8.319 \\
1119: Paris 2 terms & 89.85 & 1.660 & 0.100 & 8.386 \\
1120: \hline
1121: Paris & 89.90 & 1.610 & 0.066 & 8.428 \\
1122: \end{tabular}
1123: \end{center}
1124: \end{table}
1125:
1126:
1127:
1128: \begin{table}[tb]
1129: \caption{The parameters $a_n$, $b_n$, $m_n$ and $M_n$ of the fit
1130: for the full antisymmetric three-body wave function with the
1131: CD Bonn potential.}
1132: \label{tablebonn}
1133: \begin{tabular}{lrrrr}
1134: State/term & $a_n \ [fm^{-\frac{1}{2}}]$ & $m_n \ [fm^{-1}]$ & $b_n \
1135: [fm^{-\frac{1}{2}}]
1136: $ & $M_n \ [fm^{-1}]$ \\
1137: \hline \\
1138: $^1s_0S$ & -1.420402 & 0.408491 & -3.269355 & 0.587360\\
1139: Term 1 & -0.713402 & 0.660926 & 11.674343 & 0.787360\\
1140: & -1.500957 & 0.913361 & -33.930463 & 0.987360\\
1141: & 10.077357 & 1.165797 & 44.864561 & 1.187360\\
1142: & -6.442596 & 1.418232 & -19.339085 & 1.387360\\
1143: \hline
1144: Term 2 & 0.621726 & 0.346344 & 0.230330 & 0.384846\\
1145: & 8.073312 & 0.546344 & 5.555461 & 0.735757\\
1146: & -39.510606 & 0.746344 & -27.949487 & 1.086668\\
1147: & 52.419034 & 0.946344 & 38.650601 & 1.437579\\
1148: & -21.603467 & 1.146344 & -16.486905 & 1.788490\\
1149: \hline
1150: $^3s_1S$ & 1.308614 & 0.424997 & -4.354797 & 0.534346\\
1151: Term 1 & 1.057018 & 0.818066 & 17.513718 & 0.734346\\
1152: & -2.491967 & 1.211135 & -42.954883 & 0.934346\\
1153: & -3.044193 & 1.604204 & 49.777813 & 1.134346\\
1154: & 3.170527 & 1.997273 & -19.981851 & 1.334346\\
1155: \hline
1156: Term 2 & 0.668573 & 0.359204 & -0.141018 & 0.319402\\
1157: & 8.171864 & 0.559204 & -2.573848 & 0.702473\\
1158: & -39.501304 & 0.759204 & 14.710142 & 1.085544\\
1159: & 51.788425 & 0.959204 & -21.510088 & 1.468615\\
1160: & -21.127558 & 1.159204 & 9.514812 & 1.851686\\
1161: \hline
1162: $^3d_1S$ & 0.439080 & 0.550025 & -0.260857 & 0.431744\\
1163: Term 1 & 2.063638 & 1.214291 & 0.781535 & 0.705959\\
1164: & -26.827630 & 1.878557 & -4.200276 & 0.980175\\
1165: & 41.548562 & 2.542822 & 7.620099 & 1.254391\\
1166: & -17.223651 & 3.207088 & -3.940502 & 1.528606\\
1167: \hline
1168: Term 2 & -0.068267 & 0.556661 & 0.458816 & 0.307065\\
1169: & 0.921148 & 1.096208 & 18.886372 & 0.828691\\
1170: & -2.720463 & 1.635755 & -114.616019 & 1.350317\\
1171: & 2.869650 & 2.175302 & 176.411566 & 1.871943\\
1172: & -1.002067 & 2.714849 & -81.140735 & 2.393569\\
1173: \hline
1174: $^3s_1D$ & 7.856844 & 1.077420 & 0.008223 & 0.316062\\
1175: Term 1 & 16.344288 & 1.277420 & 0.376375 & 0.820171\\
1176: & -27.636271 & 1.477420 & -2.558819 & 1.324281\\
1177: & -26.578179 & 1.677420 & 3.621148 & 1.828390\\
1178: & 30.013317 & 1.877420 & -1.446927 & 2.332499\\
1179: \hline
1180: Term 2 & -1.089313 & 0.666635 & 28.192605 & 0.828086\\
1181: & 9.975483 & 0.866635 & -226.000462 & 1.203061\\
1182: & -26.331496 & 1.066635 & 551.617505 & 1.578036\\
1183: & 27.163776 & 1.266635 & -530.511234 & 1.953012\\
1184: & -9.718449 & 1.466635 & 176.701587 & 2.327987\\
1185: \hline
1186: $^3d_1D$ & -0.430426 & 0.345360 & 0.000745 & 0.180605\\
1187: Term 1 & 14.525808 & 0.921071 & -0.016356 & 1.497104\\
1188: & -59.378637 & 1.496783 & -0.340857 & 2.813602\\
1189: & 72.417582 & 2.072494 & 0.621826 & 4.130100\\
1190: & -27.134327 & 2.648206 & -0.265358 & 5.446598\\
1191: \hline
1192: Term 2 & 0.017189 & 0.382127 & 1.692045 & 0.222982\\
1193: & -0.110339 & 0.582127 & -33.002368 & 0.422982\\
1194: & 0.209799 & 0.782127 & 119.183936 & 0.622982\\
1195: & -0.160679 & 0.982127 & -139.556966 & 0.822982\\
1196: & 0.044030 & 1.182127 & 51.683353 & 1.022982\\
1197:
1198:
1199: \end{tabular}
1200: \end{table}
1201:
1202:
1203:
1204:
1205:
1206:
1207: \begin{table}[tb]
1208: \caption{The parameters $a_n$, $b_n$, $m_n$ and $M_n$ of the fit
1209: for the full antisymmetric three-body wave function using the
1210: Paris potential.}
1211: \label{tableparis}
1212: \begin{tabular}{lrrrr}
1213: State/term & $a_n \ [fm^{-\frac{1}{2}}]$ & $m_n \ [fm^{-1}]$ & $b_n \
1214: [fm^{-\frac{1}{2}}]
1215: $ & $M_n \ [fm^{-1}]$ \\
1216: \hline \\
1217: $^1s_0S$ & 1.185873 & 0.366138 & 2.492877 & 0.537309\\
1218: Term 1 & 1.764563 & 0.647808 & -8.176120 & 0.737309\\
1219: & -3.192238 & 0.929479 & 26.193477 & 0.937309\\
1220: & -3.305482 & 1.211149 & -37.021445 & 1.137309\\
1221: & 3.547284 & 1.492820 & 16.511211 & 1.337309\\
1222: \hline
1223: Term 2 & 0.071144 & 0.246611 & 0.130706 & 0.356517\\
1224: & 10.443707 & 0.455327 & 3.775052 & 0.687907\\
1225: & -45.143466 & 0.664043 & -18.874600 & 1.019297\\
1226: & 58.923642 & 0.872758 & 26.307399 & 1.350687\\
1227: & -24.295028 & 1.081474 & -11.338557 & 1.682077\\
1228: \hline
1229: $^3s_1S$ & 1.129463 & 0.382407 & -3.468556 & 0.482186\\
1230: Term 1 & 1.063093 & 0.803847 & 13.568797 & 0.682186\\
1231: & -1.858398 & 1.225288 & -34.549146 & 0.882186\\
1232: & -4.412162 & 1.646729 & 41.474326 & 1.082186\\
1233: & 4.078003 & 2.068170 & -17.025421 & 1.282186\\
1234: \hline
1235: Term 2 & 0.536411 & 0.308809 & -0.084904 & 0.285273\\
1236: & 9.613061 & 0.508809 & -1.853015 & 0.653089\\
1237: & -45.528006 & 0.708809 & 10.592513 & 1.020905\\
1238: & 60.473587 & 0.908809 & -15.691503 & 1.388720\\
1239: & -25.095054 & 1.108809 & 7.036910 & 1.756536\\
1240: \hline
1241: $^3d_1S$ & 0.400574 & 0.519557 & -0.238406 & 0.392032\\
1242: Term 1 & -0.136721 & 1.295772 & 0.469922 & 0.693166\\
1243: & -18.270490 & 2.071987 & -3.110884 & 0.994299\\
1244: & 31.444828 & 2.848202 & 6.387897 & 1.295433\\
1245: & -13.438191 & 3.624417 & -3.508528 & 1.596567\\
1246: \hline
1247: Term 2 & -0.050950 & 0.504112 & 0.727812 & 0.331487\\
1248: & 0.788400 & 1.071790 & 17.453956 & 0.846965\\
1249: & -2.417995 & 1.639468 & -115.129833 & 1.362444\\
1250: & 2.592890 & 2.207146 & 182.257098 & 1.877922\\
1251: & -0.912345 & 2.774824 & -85.309033 & 2.393400\\
1252: \hline
1253: $^3s_1D$ & -0.328513 & 0.432149 & 0.009219 & 0.297932\\
1254: Term 1 & 4.302698 & 0.813996 & 0.321081 & 0.804086\\
1255: & 7.984795 & 1.195843 & -2.366748 & 1.310240\\
1256: & -29.173703 & 1.577690 & 3.403634 & 1.816394\\
1257: & 17.214723 & 1.959537 & -1.367186 & 2.322548\\
1258: \hline
1259: Term 2 & -0.028375 & 0.477081 & 53.452223 & 0.722757\\
1260: & 0.521663 & 0.777081 & -512.515615 & 1.152683\\
1261: & -1.805123 & 1.077081 & 1359.198606 & 1.582609\\
1262: & 2.190226 & 1.377081 &-1367.181577 & 2.012535\\
1263: & -0.878391 & 1.677081 & 467.046364 & 2.442462\\
1264: \hline
1265: $^3d_1D$ & 0.053659 & 0.204481 & 0.000161 & 0.273758\\
1266: Term 1 & -6.461875 & 0.882230 & -0.187203 & 1.240855\\
1267: & 33.853359 & 1.559978 & 1.208634 & 2.207951\\
1268: & -44.645453 & 2.237726 & -1.685015 & 3.175047\\
1269: & 17.200310 & 2.915475 & 0.663424 & 4.142144\\
1270: \hline
1271: Term 2 & -0.065878 & 0.165941 & 0.300000 & 0.638723\\
1272: & 4.577657 & 0.797478 & -2.649250 & 1.038723\\
1273: & -17.610419 & 1.429016 & 6.528825 & 1.438723\\
1274: & 20.615853 & 2.060553 & -6.235132 & 1.838723\\
1275: & -7.517212 & 2.692091 & 2.055558 & 2.238723\\
1276:
1277:
1278: \end{tabular}
1279: \end{table}
1280:
1281:
1282:
1283:
1284: \begin{table}[tb]
1285: \caption{The parameters $a_n$, $b_n$, $m_n$ and $M_n$ of the fit
1286: for single permutation Faddeev amplitudes for the CD Bonn
1287: potential. Here the single permutation normalization integral defined
1288: in Ref. \protect\cite{hgs} is $N^2 = \langle\psi(1)|\psi(1)\rangle
1289: = 0.1597$. }
1290: \label{tablesingle}
1291: \begin{tabular}{lrrrr}
1292: State/term & $a_n \ [fm^{-\frac{1}{2}}]$ & $m_n \ [fm^{-1}]$ & $b_n \
1293: [fm^{-\frac{1}{2}}]
1294: $ & $M_n \ [fm^{-1}]$ \\
1295: \hline \\
1296: $^1s_0S$ & -1.688022 & 0.532323 & -0.455000 & 0.499291\\
1297: Term 1 & 3.387087 & 0.945460 & -0.018469 & 0.699291\\
1298: & -19.727502 & 1.358596 & -0.116043 & 0.899291\\
1299: & 38.728212 & 1.771733 & 1.839667 & 1.099291\\
1300: & -20.699775 & 2.184870 & -1.250155 & 1.299291\\
1301: \hline
1302: Term 2 & 0.085356 & 0.265782 & -0.000139 & 0.107800\\
1303: & 7.248584 & 0.676027 & 0.331163 & 0.461937\\
1304: & -38.887127 & 1.086272 & -1.640077 & 0.816075\\
1305: & 58.934705 & 1.496517 & 2.272345 & 1.170212\\
1306: & -27.381518 & 1.906762 & -0.963293 & 1.524350\\
1307: \hline
1308: $^3s_1S$ & -1.426869 & 0.510733 & 0.003479 & 0.175113\\
1309: Term 1 & 0.526004 & 0.973756 & 0.819387 & 0.436154\\
1310: & -6.707522 & 1.436779 & -0.201750 & 0.697195\\
1311: & 18.647993 & 1.899802 & -1.766824 & 0.958236\\
1312: & -11.039606 & 2.362825 & 1.145708 & 1.219277\\
1313: \hline
1314: Term 2 & 0.563686 & 0.311171 & -0.111000 & 0.341604\\
1315: & 5.367244 & 0.753186 & -0.665672 & 0.567744\\
1316: & -39.146731 & 1.195201 & 4.131617 & 0.793884\\
1317: & 63.040222 & 1.637216 & -5.800799 & 1.020025\\
1318: & -29.824421 & 2.079230 & 2.445854 & 1.246165\\
1319: \hline
1320: $^3d_1S$ & 0.278338 & 0.476215 & -0.191666 & 0.416574\\
1321: Term 1 & 1.922824 & 1.189219 & 0.132050 & 0.667231\\
1322: & -23.998840 & 1.902224 & -1.267253 & 0.917888\\
1323: & 37.142550 & 2.615229 & 2.720458 & 1.168546\\
1324: & -15.344872 & 3.328234 & -1.393589 & 1.419203\\
1325: \hline
1326: Term 2 & -0.008422 & 0.434522 & 2.864122 & 0.388367\\
1327: & 0.192188 & 1.123081 & 16.511009 & 0.717339\\
1328: & -0.653916 & 1.811640 & -116.582783 & 1.046312\\
1329: & 0.733997 & 2.500199 & 174.361137 & 1.375285\\
1330: & -0.263848 & 3.188758 & -77.153484 & 1.704258\\
1331: \hline
1332: $^3s_1D$ & 0.761443 & 0.575896 & 0.020358 & 0.389875\\
1333: Term 1 & 1.462244 & 0.775896 & -0.171640 & 0.897031\\
1334: & -0.354613 & 0.975896 & 0.119029 & 1.404187\\
1335: & 1.482841 & 1.175896 & 0.131303 & 1.911344\\
1336: & -3.351914 & 1.375896 & -0.099050 & 2.418500\\
1337: \hline
1338: Term 2 & -0.010448 & 0.610872 & 3.527742 & 0.888433\\
1339: & -0.194635 & 1.021958 & 29.340847 & 1.188433\\
1340: & 1.259114 & 1.433044 & -155.301234 & 1.488433\\
1341: & -2.105286 & 1.844130 & 201.332436 & 1.788433\\
1342: & 1.051255 & 2.255216 & -78.899791 & 2.088433\\
1343: \hline
1344: $^3d_1D$ & 0.392036 & 0.756905 & 0.004056 & 0.394786\\
1345: Term 1 & 9.293162 & 1.399407 & 0.094926 & 0.825067\\
1346: & -54.630929 & 2.041908 & -0.620590 & 1.255347\\
1347: & 74.598555 & 2.684410 & 0.869670 & 1.685627\\
1348: & -29.652824 & 3.326911 & -0.348062 & 2.115907\\
1349: \hline
1350: Term 2 & 0.018926 & 0.950962 & 34.649504 & 0.697745\\
1351: & -0.178877 & 1.526973 & -323.198246 & 1.098823\\
1352: & 0.468087 & 2.102984 & 846.220442 & 1.499901\\
1353: & -0.466509 & 2.678995 & -845.312868 & 1.900980\\
1354: & 0.158373 & 3.255006 & 287.641168 & 2.302058\\
1355: \end{tabular}
1356: \end{table}
1357:
1358:
1359:
1360:
1361: \begin{table}[ht]
1362: \caption{Results for the $\pi ^3$He scattering length.
1363: The first three entries contain measured energy level shifts of
1364: $\pi ^3$He atomic bound states together with corresponding
1365: scattering lengths extracted by using Eq. (\protect{\ref{scl}}).
1366: The averaged values includes an additional error of 10 \% to take
1367: into account possible effects from isospin breaking, etc., cf.
1368: text.}
1369: \label{scltab}
1370: \begin{center}
1371: \begin{tabular}{lcc}
1372: &$\epsilon_{1s}$ [eV] & $a_{\rm He}$ [$m_\pi^{-1}$] $\times 10^3$\\
1373: \hline
1374: R. Abela {\it et al.} \protect{\cite{phe1}} &
1375: $44 \pm 5$ & $56 \pm 6$ \\
1376: G. R. Mason {\it et al.} \protect{\cite{phe2}} &
1377: $34 \pm 4$ & $43 \pm 5$ \\
1378: I. Schwanner {\it et al.} \protect{\cite{phe3}} &
1379: $32 \pm 3$ & $41 \pm 4$ \\
1380: averaged value && $47 \pm 8 \pm 5$ \\
1381: \hline
1382: theoretical prediction && $67 \pm 15$
1383: \end{tabular} \end{center}
1384: \end{table}
1385:
1386:
1387:
1388: \newpage
1389:
1390:
1391:
1392: \begin{figure}[t!]
1393: \begin{center}
1394: \epsfig{file=wf.eps, height=8cm, angle=0}
1395: \caption{The full antisymmetric wave function
1396: for the channels $^3s_1S$ (left panel) and $^3d_1S$ (right panel) as a function
1397: of the relative pair momentum $p$ for spectator momenta $q= 0.5$
1398: (larger pair) and 1 fm$^{-1}$ (smaller pair)---the latter multiplied
1399: by a factor of three. For both momenta we show the
1400: single term fit (dashed line), the two term fit (solid line) and the exact wave
1401: function ($+$).
1402: Here we used the wave function derived from the CD Bonn potential.}
1403: \label{erg}
1404: \end{center}
1405: \end{figure}
1406:
1407:
1408:
1409:
1410: \begin{figure}[tb]
1411: \begin{center}
1412: \epsfig{file=pofr.eps,height=8cm, angle=0}
1413: \caption{The coordinate space probability distributions for
1414: the channel $^3s_1S$ with the spectator motion integrated out.
1415: Dotted curve: the
1416: representation of the Faddeev amplitude of Hajduk {\it et al.}
1417: \protect\cite{hgs}; Dashed line: the same quantity, however, showing
1418: our fit to the wave function derived from the Bonn CD potential (c.f.
1419: table \protect\ref{tablesingle});
1420: Solid line: full antisymmetrized wave function. }
1421: \label{coordspwf}
1422: \end{center}
1423: \end{figure}
1424:
1425:
1426:
1427:
1428:
1429: \begin{figure}[t!]
1430: \begin{center}
1431: \vspace{3cm}
1432: \epsfig{file=diafull.eps, width=6cm}
1433: %\special{psfile=diafull.eps, hoffset= 150 voffset=10 hscale=70 vscale=70}
1434: \caption{Leading order contributions to the $\pi\, ^3$He scattering
1435: length.}
1436: \label{diafull}
1437: \end{center}
1438: \end{figure}
1439:
1440:
1441:
1442: \begin{figure}[t!]
1443: \begin{center}
1444: \vspace{1cm}
1445: \epsfig{file=dia.eps, width=10cm}
1446: %\special{psfile=dia.eps,hoffset=107 voffset=10 hscale=70 vscale=70}
1447: \caption{Leading two-body contributions to the $\pi\, ^3$He scattering
1448: length.}
1449: \label{dia}
1450: \end{center}
1451: \end{figure}
1452:
1453:
1454: \end{document}
1455:
1456: