1: \documentclass{elsart}

2: 

3: \usepackage{graphicx,hyperref}

4: \usepackage{epsfig,array}

5: \usepackage{booktabs}

6: %\usepackage{natbib}

7: %\preprint{ROM2F/2003/10}

8: 

9: % fb commands

10: % -----------

11: 

12: \newcommand{\MeV}{{\rm MeV}}

13: \newcommand{\MSbar}{\hbox{$\overline{MS}$\ }}

14: 

15: %\hyphenation{super-charges}

16: \newcommand{\ket}[1]{\left| {#1} \right\rangle}

17: \newcommand{\bra}[1]{\left\langle {#1}\right|}

18: \newcommand{\braket}[2]{\left\langle {#1} \right| \left. {#2} \right\rangle}

19: \newcommand{\too}{\mathop{\ \longrightarrow}\ }

20: \newcommand{\beq}{\begin{equation}}

21: \newcommand{\eeq}{\end{equation}}

22: \newcommand{\beqa}{\begin{eqnarray}}

23: \newcommand{\eeqa}{\end{eqnarray}}

24: \newcommand\eqn[1]{\label{eq:#1}} 

25: \newcommand\eq[1]{eq.~(\ref{eq:#1})} 

26: \newcommand\tab[1]{{\footnotesize {\bf Table}~[{\bf\ref{#1}}]}} 

27: \newcommand\sect[1]{sec.~\ref{sec:#1}} 

28: \newcommand{\Det}{\mathop{\rm Det}}

29: \newcommand{\CD}{{\cal D}}

30: \newcommand{\CO}{{\cal O}}

31: \newcommand{\CL}{{\cal L}}

32: \newcommand{\bfn}{{\bf n}}

33: \newcommand{\bfr}{{\bf r}}

34: \newcommand{\bfw}{{\bf w}}

35: \newcommand{\bfm}{{\bf m}}

36: \newcommand{\bfe}{{\bf  e}}

37: \newcommand{\xh}{{\bf \hat x}}

38: \newcommand{\yh}{{\bf \hat y}}

39: \newcommand{\zh}{{\bf \hat z\,}}

40: \newcommand{\Tr}{\mathop{\rm Tr}}

41: \newcommand{\sla}[1]%

42:         {\kern .25em\raise.18ex\hbox{$/$}\kern-.75em #1}% Feynman slash

43: \newcommand{\mybar}[1]%

44:         {\kern 0.8pt\overline{\kern -0.8pt#1\kern -0.8pt}\kern 0.8pt}

45: \newcommand{\gsim}{\, \raisebox{-0.8ex}{$\stackrel{\textstyle >}{\sim}$ }}

46: \newcommand{\lsim}{\, \, \raisebox{-0.8ex}{$\stackrel{\textstyle <}{\sim}$ }}

47: \newcommand{\ttbs}{\char'134}           % \backslash for \tt (Nucl.Phys. :)%

48: 

49: \setlength{\parindent}{20pt}

50: \setlength{\parskip}{5pt}

51: 

52: \newcommand\mycaption[1]{\caption{\footnotesize #1}} 

53: 

54: % tables

55: 

56: %==============================================================================================%

57: \begin{document} 

58: %==============================================================================================%

59: 

60: 

61: 

62: 

63: 

64: %=========================================================================================================================================%

65: \begin{frontmatter}

66: %=========================================================================================================================================%

67: 

68: \title{On the discretization of physical momenta in lattice QCD}

69: 

70: \author[romeII]{G.M.~de~Divitiis}

71: \author[romeII]{R.~Petronzio}

72: \author[romeII]{N.~Tantalo}

73: 

74: 

75: \address[romeII]{University of Rome ``Tor Vergata'' and INFN sez. RomaII, 

76: 	Via della Ricerca Scientifica 1, 

77: 	I-00133 Rome}

78: 

79: 

80: 

81: \begin{abstract}

82: The adoption of two distinct boundary conditions for two fermions species on a finite 

83: lattice allows to deal with arbitrary relative momentum  between the two particle 

84: species, in spite of the momentum quantization rule due to a limited physical box size.

85: We test the physical significance of this topological momentum by checking in the continuum limit 

86: the validity of the expected energy-momentum dispersion relations. 

87: \end{abstract}

88: 

89: 

90: %=========================================================================================================================================%

91: \end{frontmatter}

92: %=========================================================================================================================================%

93: 

94: 

95: 

96: %_______________________________________________________%

97: \section{Introduction}

98: \label{sec:introduction} 

99: %_______________________________________________________%

100: 

101: 

102: Among the restrictions of field theory formulations on a lattice, the finite volume

103: momentum quantization represents a severe limitation in various 

104: phenomenological applications. For example, in a two body hadron decay where the energies of the 

105: decay products, related by 4--momentum conservation to the masses 

106: of the particles involved,

107: cannot assume their physical values unless these masses are consistent 

108: with the momentum quantization rule. 

109: In this letter we propose a solution to the problem 

110: based on the use of different

111: boundary conditions for different 

112: fermion species\footnote{R.P. thanks M.~L\"uscher for drawing his attention on this point.}.

113: 

114: We test the idea in the simplest case of a flavoured quark-antiquark 

115: correlation used to determine asymptotically the energy of the corresponding meson.

116: In this case the fermion and the antifermion are the different fermion species and we show that

117: suitable different boundary conditions can propagate a meson with a momentum that 

118: can assume continuous values.

119: 

120: Section~\ref{sec:method} introduces the boundary conditions, 

121: section~\ref{sec:results} reports on the numerical 

122: results and section~\ref{sec:conclusions} summarizes the conclusions.

123: 

124: 

125: %_______________________________________________________%

126: \section{Generalized boundary conditions}

127: \label{sec:method}

128: %_______________________________________________________%

129: 

130: In order to explain the method to have continuous physical momenta

131: on a finite volume we first re--derive, for the sake of clarity, the momentum quantization rule in

132: the case of a particle with periodic boundary conditions (PBC).

133: To this end we consider a fermionic field $\psi(x)$ 

134: on a 4--dimensional finite volume of topology $T\times L^3$

135: with PBC in the spatial directions

136: %

137: %

138: \beq

139: \psi(x+\vec{e}_i\ L) = \psi(x)\;, \qquad i=1,2,3

140: \eeq

141: %

142: %

143: This condition can be re-expressed by Fourier transforming

144: both members of the previous equation

145: %

146: %

147: \beq

148: \int{d^4p\ e^{-ip(x+\vec{e}_i\ L)}\ \tilde{\psi}(p)}

149: = \int{d^4p\ e^{-ipx}\ \tilde{\psi}(p)}

150: \;, \qquad i=1,2,3

151: \label{eq:fspbc}

152: \eeq

153: %

154: %

155: It follows directly from the previous relation that, 

156: in the case of periodic boundary conditions, one has

157: %

158: %

159: \beq

160: e^{ip_iL} = 1 \quad \Longrightarrow \quad p_i = \frac{2\pi\ n_i}{L}

161: \;, \quad i=1,2,3  

162: \label{eq:pbcmom}

163: \eeq

164: %

165: %

166: where the $n_i$'s are integer numbers.

167: The authors of \cite{Jansen:1996ck} 

168: have first considered a generalized set of boundary conditions, that here

169: we call $\theta$--boundary conditions ($\theta$--BC),

170: depending upon the choice of a topological 3--vector $\vec{\theta}$

171: %

172: %

173: \beq

174: \psi(x+\vec{e}_i\ L) = e^{i\theta_i}\ \psi(x)\;, \qquad i=1,2,3

175: \label{eq:thetabc}

176: \eeq

177: %

178: %

179: The modification of the boundary conditions affects the zero of the 

180: momentum quantization rule. 

181: Indeed, by re-expressing equation~(\ref{eq:thetabc}) in Fourier space, 

182: as already done in the case of PBC in equation~(\ref{eq:fspbc}), one

183: has

184: %

185: %

186: \beq

187: e^{i(p_i-\frac{\theta_i}{L})L} = 1 \quad \Longrightarrow 

188: \quad p_i = \frac{\theta_i}{L} + \frac{2\pi\ n_i}{L}

189: \;, \quad i=1,2,3  

190: \label{eq:thetamom}

191: \eeq

192: %

193: %

194: It comes out that the spatial momenta are still quantized as for

195: PBC but shifted by an arbitrary \emph{continuous} 

196: amount ($\theta_i/L$).

197: The observation that this continuous

198: shift in the allowed momenta it is physical and can be thus profitably used in 

199: phenomenological

200: applications is the key point of the present work.  

201: The generalized $\theta$--dependent

202: boundary conditions of equation~(\ref{eq:thetabc}) can be  implemented by

203: making a unitary Abelian transformation on the fields satisfying $\theta$--BC

204: %

205: %

206: \beq

207: \psi(x) \quad \longrightarrow  \quad 

208: {\mathcal U}\ (\theta,x) \psi(x)= e^{-\frac{i \theta x}{L}}\  \psi(x)

209: \label{eq:unittransf}

210: \eeq

211: %

212: %

213: As a consequence of this transformation the resulting field satisfies

214: periodic boundary conditions but obeys a modified Dirac equation

215: %

216: %

217: \beqa

218: S[\bar{\psi},\psi] \quad

219: &\longrightarrow&  \quad \sum_{x,y}{ \bar{\psi}(x)\ {\mathcal U}(\theta,x)D(x,y) {\mathcal U}^{-1}(\theta,y)\ \psi(y)}

220: \nonumber \\

221: &=& \quad \sum_{x,y}{ \bar{\psi}(x) \ D_\theta(x,y)\ \psi(y)} 

222: \label{eq:Smod}

223: \eeqa    

224: %

225: %

226: where the $\theta$--dependent lattice Dirac operator $D_\theta(x,y)$ is obtained

227: by starting from the preferred discretization of the Dirac operator and 

228: by modifying the definition of the covariant lattice derivatives, i.e.

229: by passing from the standard forward and backward derivatives: 

230: %

231: %

232: \beqa

233: \nabla_\mu \psi(x) &=& \frac{1}{a}\left[U_\mu(x)\psi(x+a\ \hat{\mu}) - \psi(x) \right]

234: \nonumber \\

235: \nabla^\dagger_\mu \psi(x) &=& \frac{1}{a}\left[\psi(x) - U^{-1}_\mu(x-a\ \hat{\mu})\psi(x-a\ \hat{\mu}) \right]

236: \label{eq:standardder}

237: \eeqa    

238: %

239: %

240: to the $\theta$--dependent ones

241: %

242: %

243: \beqa

244: \nabla_\mu(\theta) \psi(x) &=& \frac{1}{a}\left[\lambda_\mu\ U_\mu(x)\psi(x+a\ \hat{\mu}) - \psi(x) \right]

245: \nonumber \\

246: \nabla_\mu(\theta)^\dagger \psi(x) &=& \frac{1}{a}\left[\psi(x) - \lambda_\mu^{-1}\ 

247: U^{-1}_\mu(x-a\ \hat{\mu})\psi(x-a\ \hat{\mu}) \right]

248: \label{eq:dermod}

249: \eeqa    

250: %

251: %

252: where we have introduced

253: %

254: %

255: \beq

256: \lambda_\mu = e^\frac{ia\theta_\mu}{L} \qquad \theta_0 = 0

257: \eeq

258: %

259: %

260: 

261: 

262: The authors of ref.~\cite{Bucarelli:1998mu} have considered for the first time

263: $\theta$--BC in perturbative phenomenological applications. They used the shift

264: in the momentum quantization rule, that they called a ``finite size momentum'',

265: in order to build an external source to probe the tensor structure 

266: of the Wilson operators. 

267: A similar analysis was then repeated non--perturbatively by the same group in ref.~\cite{Guagnelli:2003hw}.

268: The use of $\theta$--BC has been considered in different contexts 

269: also in~\cite{Bedaque:2004kc,Gross:1982at,Kiskis:2002gr,Kiskis:2003rd,Roberge:1986mm}.

270: 

271: In this work we point out that the term $\vec{\theta}/L$ acts as a true

272: physical momentum.

273: 

274: As a test, we calculate  the energy of a meson

275: made up by two different quarks with different  $\theta$--BC for the two 

276: flavours.

277: We work in the $O(a)$--improved Wilson--Dirac lattice formulation of the 

278: QCD within the Schr\"odinger Functional formalism~\cite{Luscher:1992an,Sint:1994un} but,

279: we want to stress that the use of $\theta$--BC in the spatial directions 

280: is completely decoupled from the choice

281: of time boundary conditions and 

282: can be profitably used outside the Schr\"odinger Functional formalism,

283: for example in the case of standard periodic time boundary conditions. 

284: Let us consider the following correlators

285: %

286: %

287: \beq

288: f_P^{ij}(\theta;x_0) = -\frac{a^6}{2}\sum_{\vec{y},\vec{z},\vec{x}}{

289: \langle\ \bar{\zeta}_i(\vec{y})\gamma_5 \zeta_j(\vec{z}) \ 

290: \bar{\psi}_j(x) \gamma_5 \psi_i(x) \ \rangle

291: } 

292: \label{eq:fp}

293: \eeq

294: %

295: %

296: where $i$ and $j$ are flavour indices, all the fields satisfy periodic boundary conditions 

297: and the two flavours obey different $\theta$--modified Dirac equations, as explained in equations~(\ref{eq:Smod}),

298: (\ref{eq:standardder})~and~(\ref{eq:dermod}). In practice it is adequate to choose the flavour $i$

299: with $\theta=0$, i.e. with ordinary PBC, and the flavour $j$ with $\theta\neq 0$.

300: After the Wick contractions the pseudoscalar correlator of

301: equation~(\ref{eq:fp}) reads

302: %

303: %

304: \beq

305: f_P^{ij}(\theta;x_0) = \frac{a^6}{2}\sum_{\vec{y},\vec{z},\vec{x}} \Tr{

306: \langle\ \gamma_5\ S_j(\theta;\vec{z},x)\ \gamma_5\ S_i(0;x,\vec{y})\ \rangle

307: } 

308: \label{eq:fpcontractions}

309: \eeq

310: %

311: %

312: where $S(\theta;x,y)$ and $S(0;x,y)$ are the inverse of the $\theta$--modified 

313: and of the standard Wilson--Dirac operators respectively.

314: Note that the projection on the momentum $\vec{\theta}/L$

315: of one of the quark legs in equation~(\ref{eq:fpcontractions}) it is not

316: realized by summing on the lattice points with an exponential factor but

317: it is encoded in the $\theta$--dependence of the modified Wilson--Dirac

318: operator and, consequently, of its inverse $S(\theta;x,y)$. 

319: 

320: This correlation is expected to decay exponentially at large times

321: as

322: %

323: %

324: \beq

325: f_P^{ij}(\theta;x_0) \qquad \stackrel{x_0\gg 1}{\longrightarrow} \qquad f_{ij}\  e^{- ax_0 E_{ij}(\theta,a)}

326: \eeq

327: %

328: %

329: where, a part from corrections proportional to the square of the lattice spacing, $E_{ij}$ is the

330: physical energy of the mesonic state

331: %

332: %

333: \beq

334: E_{ij}(\theta,a) = \sqrt{M_{ij}^2+\left(\frac{\vec{\theta}}{L}\right)^2} + O(a^2)

335: \label{eq:drlatt}

336: \eeq

337: %

338: %

339: here $M_{ij}$ is the mass of the pseudoscalar meson made of a $i$ and a $j$

340: quark anti--quark pair.

341: In the next section we will show the calculation of the meson energies for

342: different flavours and for different choices of $\vec{\theta}$. We will show that

343: after the continuum extrapolations we will find the expected relativistic

344: dispersion relations

345: %

346: %

347: \beq

348: E_{ij}^2 = M_{ij}^2+\left(\frac{\vec{\theta}}{L}\right)^2

349: \label{eq:drcont}

350: \eeq

351: %

352: %

353: 

354: 

355: 

356: 

357: 

358: %_______________________________________________________%

359: \section{Numerical tests}

360: \label{sec:results}

361: %_______________________________________________________%

362: 

363: All the results of this section are obtained in the quenched

364: approximation of the QCD.

365: We have done simulations on a physical volume 

366: of topology $T\times L^3$ with $T=2L$ and linear extension $L=3.2\ r_0$, where $r_0$

367: is a phenomenological distance parameter related

368: to the static quark anti--quark potential \cite{Sommer:1994ce}. In order

369: to extrapolate our numerical results to the continuum limit

370: we have simulated the same physical volume using three different

371: discretizations with number of points $(32\times 16^3)$, 

372: $(48\times 24^3)$ and $(64\times 32^3)$ respectively. 

373: We have fixed the three values of the bare couplings corresponding

374: to the different discretizations using the $r_0$ scale with the

375: numerical results given in \cite{Necco:2001xg}. All the parameters

376: of the simulations are given in table~\ref{tab:simpar}.

377: %

378: %

379: %

380: \begin{table}[t]

381: \begin{center}

382: \scriptsize

383: \begin{tabular}{cccccc}

384: \toprule               

385: $\bf \beta$ &          & $\bf L/a$   &          & $\bf k$      & $\bf r_0\ m^{RGI}$ \\

386: \toprule               

387:             & $\qquad$ &             & $\qquad$ &  0.132054    & 0.645(7)      \\

388: 5.960       & $\qquad$ & 16          & $\qquad$ &  0.132609    & 0.520(6)      \\

389:             & $\qquad$ &             & $\qquad$ &  0.133315    & 0.362(5)      \\

390:             & $\qquad$ &             & $\qquad$ &  0.133725    & 0.269(4)      \\

391: \midrule               		  

392:             & $\qquad$ &             & $\qquad$ &  0.134208    & 0.655(9)      \\

393: 6.211       & $\qquad$ & 24          & $\qquad$ &  0.134540    & 0.521(7)      \\

394:             & $\qquad$ &             & $\qquad$ &  0.134954    & 0.354(6)      \\

395:             & $\qquad$ &             & $\qquad$ &  0.135209    & 0.251(5)      \\

396: \midrule               		  

397:             & $\qquad$ &             & $\qquad$ &  0.134517    & 0.676(15)     \\

398: 6.420       & $\qquad$ & 32          & $\qquad$ &  0.134764    & 0.540(12)     \\

399:             & $\qquad$ &             & $\qquad$ &  0.135082    & 0.365(10)     \\

400:             & $\qquad$ &             & $\qquad$ &  0.135269    & 0.262(9)      \\

401: \toprule

402: $\bf [\theta_x,\theta_y,\theta_z]$ & = & $[0.0,\ 0.0,\ 0.0]$ & $[1.0,\ 1.0,\ 1.0]$ & $[2.0,\ 2.0,\ 2.0]$ & $[3.0,\ 3.0,\ 3.0]$ \\

403: \toprule

404: \end{tabular}

405: \vskip 0.3cm

406: \mycaption{Parameters of the simulations. The values of the bare couplings has been chosen in order to

407: fix the extension of the physical volume $L = 3.2\ r_0$.

408: For each

409: value of the $k$ parameter we have simulated all the values of $\vec{\theta}$.}

410: \label{tab:simpar}

411: \end{center}

412: \vskip 1.5cm

413: \end{table}

414: %

415: %

416: %

417: The values of the RGI quark masses reported in table~\ref{tab:simpar} have been calculated

418: starting from the PCAC relation

419: %

420: %

421: \beq

422: m^{PCAC}_{ii} = \frac{ \tilde{\partial_0}f_A^{ii}(0;x_0) + a c_A \partial_0^\dagger \partial_0 f_P^{ii}(0;x_0)  }{2 f_P^{ii}(0;x_0)} 

423: \label{eq:PCACdiagonal}

424: \eeq

425: %

426: %

427: where $\partial_\mu$, $\partial^\dagger_\mu$ are the usual forward and backward lattice derivatives respectively while

428: $\tilde{\partial}_\mu$ is defined as  $(\partial_\mu + \partial_\mu^\dagger)/2$. The time correlator $f_P^{ij}(0;x_0)$ has 

429: already been defined in equation~(\ref{eq:fp}) while $f_A^{ij}(0;x_0)$ is defined in the following relation

430: %

431: %

432: \beq

433: f_A^{ij}(\theta;x_0) = -\frac{a^6}{2}\sum_{\vec{y},\vec{z},\vec{x}}{

434: \langle\ \bar{\zeta}_i(\vec{y})\gamma_5 \zeta_j(\vec{z}) \ 

435: \bar{\psi}_j(x) \gamma_0 \gamma_5 \psi_i(x) \ \rangle

436: } 

437: \label{eq:fa}

438: \eeq

439: %

440: %

441: The improvement coefficient $c_A$ has been computed non--perturbatively in \cite{Luscher:1997ug}.

442: The RGI quark masses are connected to the PCAC masses of equation~(\ref{eq:PCACdiagonal}) from

443: the following relation

444: %

445: %

446: \beq

447: m_{ii}^{RGI} = Z_M(g_0) \ \left[ 1 + (b_A-b_P)\ am_i \right] \ m^{PCAC}_{ii}(g_0)

448: \label{eq:rgimassii}

449: \eeq

450: %

451: %

452: where the renormalization factor $Z_M(g_0)$ has been computed non--perturbatively in \cite{Capitani:1998mq}.

453: Also the difference of the improvement coefficients $b_A$ and $b_P$ is known non--perturbatively from

454: \cite{deDivitiis:1998ka,Guagnelli:2000jw}. In (\ref{eq:rgimassii}) the masses $m_i$ are the bare ones

455: defined as

456: %

457: %

458: \beq

459: am_i = \frac{1}{2} \left[\frac{1}{k_i} - \frac{1}{k_c} \right] 

460: \label{eq:barequarkmass}

461: \eeq

462: %

463: %

464: 

465: For each value of the simulated quark masses reported in table~\ref{tab:simpar} we have

466: inverted the Wilson--Dirac operator for three non--zero values of $\vec{\theta}$.

467: Setting the lattice scale by using the physical value $r_0= 0.5$ fm, the expected values of the physical momenta

468: associated with the choices of $\vec{\theta}$ given in table~\ref{tab:simpar} are simply calculated according

469: to the following relation

470: %

471: %

472: \beq

473: |\vec{p}| \; = 

474: \; \frac{|\vec{\theta}|}{L} 

475: \; \simeq\; 0.125\ |\vec{\theta}| \; \mbox{GeV}

476: \quad=\quad \left\{

477: \begin{array}{l}

478: 0.000 \\

479: 0.217 \\

480: 0.433 \\

481: 0.650 \\

482: \end{array}  \right. \; \mbox{GeV} 

483: \qquad L \ \simeq\ 1.6\; \mbox{fm} 

484: \eeq

485: These values have to be compared with the value of the lowest physical momentum allowed

486: on this finite volume in the case of periodic boundary conditions, 

487: i.e. $|\vec{p}| \simeq 0.785$ GeV.

488: %

489: %

490: %

491: %

492: %

493: \begin{figure}[t]

494: \begin{center}

495: \epsfig{file=./pics/effmass.eps,width=12cm}

496: \mycaption{Effective energies $E_{eff}^{ij}(\theta,a;x_0)$, as defined in eq.~(\ref{eq:effenerg})

497: at fixed cut--off. The results correspond to the simulation done at $\beta=6.211$

498: with $r_0\ m_1^{RGI} = 0.655$ and $r_0\ m_2^{RGI} = 0.354$.

499: Similar figures could have been shown for other combinations of the simulated quark masses

500: and for the other values of the bare coupling.}

501: \label{fig:effmass}

502: \end{center}

503: \vskip 1.5cm

504: \end{figure}

505: %

506: %

507: %

508: %

509: %

510: 

511: At fixed cut--off, for each combination of flavour indices and for each value of $\vec{\theta}$ 

512: reported in table~\ref{tab:simpar} we have extracted the 

513: effective energy from the correlations of eq.~(\ref{eq:fp}), $f_P^{ij}(\theta;x_0)$, as follows

514: %

515: %

516: \beq

517: a\ E_{eff}^{ij}(\theta,a;x_0) \; =\; \frac{1}{2}\ \log\left( \frac{f_P^{ij}(\theta;x_0-1)}{f_P^{ij}(\theta;x_0+1)} \right)

518: \label{eq:effenerg}

519: \eeq

520: %

521: %

522: In fig.~\ref{fig:effmass} we show this quantity for the simulation performed at $\beta = 6.211$

523: corresponding to $r_0\ m_1^{RGI} = 0.655$ and $r_0\ m_2^{RGI} = 0.354$, for each simulated value

524: of $\vec{\theta}$. As can be seen the correlations with higher values of $|\vec{\theta}|$ are

525: always greater than the corresponding ones with lower values of the physical momentum

526: %

527: %

528: \beq

529: |\vec{\theta}_1| > |\vec{\theta}_2| \quad \Rightarrow \quad  E_{eff}^{ij}(\theta_1,a;x_0) > E_{eff}^{ij}(\theta_2,a;x_0)

530: \label{eq:bounds}

531: \eeq

532: %

533: %

534: a feature that will be confirmed in the continuum limit.

535: 

536: In the continuum extrapolations we have fixed the physical values of the quark masses

537: slightly interpolating the simulated sets of numerical results.

538: %

539: %

540: %

541: %

542: %

543: \begin{figure}[t]

544: \begin{center}

545: \epsfig{file=./pics/continuum.eps,width=12cm}

546: \mycaption{Continuum extrapolations of the plateau averaged effective energies $E^{ij}(\theta,a)$. 

547: The results correspond to the quark masses $r_0\ m_1^{RGI} = 0.650$ and $r_0 m_2^{RGI} = 0.350$.

548: Similar figures could have been shown for other combinations of the simulated quark masses.}

549: \label{fig:contlimit}

550: \end{center}

551: \vskip 1.5cm

552: \end{figure}

553: %

554: %

555: %

556: %

557: %

558: Being interested in the ground

559: state contribution to the correlation of eq.~(\ref{eq:fp}), we have averaged the effective energies

560: over a ground state plateau of physical length depending upon the quark flavours.

561: We call $E^{ij}(\theta,a)$ the result of the average and in fig.~\ref{fig:contlimit}

562: we show a typical continuum extrapolation of this quantity. 

563: Similar figures could have been shown for the other values of simulated quark masses.

564: %

565: %

566: %

567: %

568: %

569: \begin{figure}[t]

570: \begin{center}

571: \epsfig{file=./pics/dispersion.eps,width=12cm}

572: \mycaption{

573: Continuum dispersion relations. The data correspond to different

574: combinations of the simulated quark masses and reproduce very well

575: the expected theoretical behavior, i.e. straight lines having

576: as intercepts the meson masses and as angular coefficients one (see eq.~\ref{eq:drcont}).

577: }

578: \label{fig:disprel}

579: \end{center}

580: \vskip 1.5cm

581: \end{figure}

582: %

583: %

584: %

585: %

586: %

587: 

588: The continuum results verify very well the dispersion relations of

589: equation~(\ref{eq:drcont}) as can be clearly seen from fig.~\ref{fig:disprel} in which

590: the square of $E^{ij}(\theta)$ for various combinations of the flavour indices is plotted

591: versus the square of the physical momenta $|\vec{\theta}|/L$. 

592: The plotted lines have not been fitted but have been obtained by using as

593: intercepts the simulated meson masses and by fixing their 

594: angular coefficients to one.

595: 

596: 

597: 

598: 

599: \clearpage

600: 

601: 

602: 

603: %_______________________________________________________%

604: \section{Conclusions}

605: \label{sec:conclusions}

606: %_______________________________________________________%

607: 

608: We have argued that the limitation

609: represented by the finite volume momentum quantization rule can be overcame

610: by using different boundary conditions for different fermion species.   

611: 

612: We have supported this observation by calculating the relativistic dispersion

613: relations satisfied by a set of pseudoscalr mesons in the case of quenched

614: lattice QCD. 

615: We have shown that the physical momentum carried by these particles

616: can be varied \emph{continuously} by enforcing different $\theta$--boundary conditions

617: (see eq.~(\ref{eq:thetabc}))  for the two quarks inside the mesons.

618: 

619: The method proposed can be applied to study all the

620: quantities of phenomenological interest that would benefit from the

621: introduction of continuous physical momenta like, for example, weak matrix elements. 

622: The suggestion can be applied in quenched QCD also in the case of flavourless mesons 

623: while can be extended to full QCD in the flavoured case only.

624: 

625: 

626: 

627: %_______________________________________________________%

628: \begin{ack}

629: We warmly thank M.~L\"uscher for enlightening discussions.

630: We also thank F.~Palombi for useful remarks.

631: \end{ack}

632: %_______________________________________________________%

633: 

634: 

635: 

636: 

637: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

638: \bibliographystyle{h-elsevier} 

639: \bibliography{tt}

640: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

641: 

642: 

643: \end{document}

644: