1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2: 

3: %             version 31/01/2007

4: 

5: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

6: 

7: \documentclass[fleqn,preprintnumbers]{elsart}

8: 

9: \usepackage{graphicx}% Include figure files

10: \usepackage{dcolumn}% Align table columns on decimal point

11: \usepackage{bm}% bold math

12: \usepackage{amssymb}

13: \usepackage{indentfirst}

14: %\usepackage{psfig,color}

15: \usepackage{epsfig}

16: \usepackage{epsf}

17: \usepackage{graphicx}

18: 

19: 

20: \usepackage{amssymb}

21: 

22: \usepackage{indentfirst}

23: 

24: %\usepackage{psfig,color}

25: 

26: \usepackage{epsfig}

27: 

28: 

29: \usepackage{epsf}

30: 

31: \usepackage{graphicx}

32: 

33: 

34: 

35: \def\VEC#1{\hbox{\boldmath$#1$\unboldmath}}

36: 

37: \def\Vec#1{{\bf #1}}

38: 

39: \def\slash{/\!\!\!}

40: 

41: \def\one{1 \!\! 1}

42: 

43: \def\D{{\mathrm d}}

44: 

45: \def\E{{\mathrm e}}

46: 

47: \def\I{{\mathrm i}}

48: 

49: 

50: 

51: \newcommand{\ie}{{\it i.e.}}

52: 

53: \newcommand{\kp}{\mathbf{k}_\perp}

54: 

55: \newcommand{\p}{\perp}

56: 

57: \newcommand{\ssh}{\!\!\!/}

58: 

59: \renewcommand{\d}{{\mathrm d}}

60: 

61: 

62: 

63: %\journal{Physics Letters B}

64: 

65: 

66: 

67: \begin{document}

68: 

69: \begin{flushright}

70: USM-TH-177\\

71: \end{flushright}

72: \bigskip\bigskip

73: 

74: \begin{frontmatter}

75: 

76: \title{Flavor separation of the Boer-Mulders function from unpolarized $\pi^- p$ and $\pi^- D$ Drell-Yan processes}

77: 

78: \author[utfsm]{Zhun Lu},

79: \author[pku]{Bo-Qiang Ma\corauthref{cor}} \ead{mabq@phy.pku.edu.cn}

80: \corauth[cor]{Corresponding authors.},

81: \author[utfsm]{Ivan Schmidt\corauthref{cor}} \ead{ivan.schmidt@usm.cl}

82: %\corauth[cor]{Corresponding authors.}

83: \address[utfsm]{Departamento de F\'\i sica, Universidad T\'ecnica Federico

84: Santa Mar\'\i a, Casilla 110-V, Valpara\'\i so, Chile}

85: \address[pku]{School of Physics, Peking University, Beijing 100871, China}

86: %\address[it]{Di.S.T.A., Universit\`a del Piemonte Orientale ``A. Avogadro'', \\

87: %and INFN, Gruppo Collegato di Alessandria, 15100 Alessandria, Italy}

88: 

89: 

90: \begin{abstract}

91: We show that measuring the $\cos 2\phi $ angular dependence in

92: unpolarized Drell-Yan processes with $\pi^-$ beams colliding on

93: proton and deuteron targets can determine the ratio of the

94: Boer-Mulders functions for $d$ and $u$ quarks inside the proton

95: $h_1^{\perp,d}/h_1^{\perp,u}$, which is still lack of theoretical

96: constraint. The comparison of the $\cos 2 \phi$ asymmetries

97: measured in unpolarized $\pi^- p$ and $\pi^- D$ Drell-Yan

98: processes, which are accessible at CERN by the COMPASS

99: collaboration, can help to discriminate whether $h_1^\perp$

100: effects or QCD vacuum effects are preferred by data.

101: 

102: 

103: 

104: 

105: \end{abstract}

106: 

107: \begin{keyword} Drell-Yan process \sep azimuthal asymmetry \sep $\Vec

108: k_T$-dependent distribution functions \sep transverse spin

109: 

110: 

111: \PACS 13.85.-t, 12.39.-x, 13.88.+e, 14.20.Dh

112: \end{keyword}

113: 

114: \end{frontmatter}

115: 

116: 

117: 

118: 

119: \vspace{1cm}

120: 

121: %\section{Introduction}

122: 

123:  The transverse spin phenomena appearing in high energy

124: scattering processes~\cite{bdr} is among the most interesting issues

125: of spin physics. Substantial single-spin asymmetries (SSA) in

126: semi-inclusive deeply inelastic scattering

127: (SIDIS)~\cite{smc,Airapetian:2004tw,compass,hermes05}, with one

128: incoming nucleon transversely polarized, have been measured by

129: several experiments. The interpretation of these asymmetries

130: provides new insights into QCD and nucleon

131: structure~\cite{sivers,collins93,anselmino95,bm,bhs02,collins02,belitsky}.

132: One of the mechanisms which can account for such SSA is the Sivers

133: effect~\cite{sivers}, related to a spin and $\Vec k_T$-dependent

134: function~\cite{levelt,kotzinian,mulders} named as Sivers function

135: $f_{1T}^\perp(x,\Vec k_T^2)$. It arises from the correlation between

136: the nucleon transverse spin and quark transverse momentum. Despite

137: its naively $T$-odd structure, $f_{1T}^\perp$ has been shown to be

138: non-zero~\cite{bhs02} due to its special gauge-link

139: structure~\cite{collins02,belitsky}. In the case of unpolarized

140: collisions, large $\cos 2 \phi$ asymmetries have been observed

141: ~\cite{na10,conway} in $\pi^- N$ Drell-Yan dilepton production

142: processes. A parton interpretation of this asymmetry has been

143: proposed in~\cite{boer} in terms of another leading twist $T$-odd

144: distribution function, the Boer-Mulders function $h_{1}^\perp(x,\Vec

145: k_T^2)$~\cite{bm}, which has the same QCD origin~\cite{collins02} as

146: the Sivers function. It describes the transverse polarization

147: distribution of the quark inside the unpolarized hadron due to the

148: correlation between the quark transverse spin and transverse

149: momentum. The model calculations of the Boer-Mulders function have

150: been performed in Refs.~\cite{gg02,bbh03,yuan,bsy04} for the proton

151: and in Ref.~\cite{lm04} for the pion, based on the gauge-link

152: structure of the distribution. The resulting asymmetries have been

153: estimated for unpolarized $p\bar{p}$~\cite{bbh03,gg05,blm06} and

154: $\pi^- N$ ~\cite{lm05} Drell-Yan processes. The role of Boer-Mulders

155: functions in unpolarized SIDIS process has also been

156: discussed~\cite{gg03,blm05}. One of the significant features of the

157: Boer-Mulders function relies on that fact that the spin structure of

158: hadrons can be also studied in physical processes without invoking

159: beam or target polarization.

160: 

161: The large anomalous Drell-Yan asymmetry implies a significant

162: non-zero size of the Boer-Mulders function. However its extraction

163: from experimental data is relatively difficult comparing to that

164: of the Sivers function. In the latter case the Sivers function is

165: convoluted with ordinary distribution/fragmentation functions,

166: which are well-known. Several sets of the proton Sivers functions

167: for both $u$ and $d$ quark have been

168: extracted~\cite{anselmino05c,vy05,efremov,anselmino05b} recently

169: from HERMES~\cite{Airapetian:2004tw,hermes05} and

170: COMPASS~\cite{compass} data on single-transverse spin asymmetries

171: in SIDIS. %, and show good agreement with the data.

172: In the process in which the Boer-Mulders function contributes, it

173: is always convoluted with itself or other chiral-odd functions,

174: such as the transversity or the Collins fragmentation

175: function~\cite{collins93}, which are also not well known at

176: present. Different models or theoretical

177: considerations~\cite{yuan,bsy04,pobylista} predict very different

178: flavor dependence of $h_1^\perp$, and even the relative sign

179: between $h_1^{\perp,u}$ and $h_1^{\perp,d}$ is not known. In this

180: paper we suggest to use unpolarized $\pi^- p$ and $\pi^- D$

181:  Drell-Yan processes to access the flavor

182: dependence of $h_1^\perp$, especially the ratio

183: $h_1^{\perp,d}/h_1^{\perp,u}$. The hadron program by the COMPASS

184: collaboration will start in 2007 at CERN, in which a $\pi^-$ beam

185: colliding with both proton and deuteron targets are going to be

186: available. We show that by comparing the $\cos 2 \phi$ angular

187: dependent part of the cross sections of unpolarized $\pi^- p$ and

188: $\pi^- D$ Drell-Yan processes, one can determine the ratio

189: $h_1^{\perp,d}/h_1^{\perp,u}$. By comparing the $\cos 2 \phi$

190: asymmetries in these processes, one can also discriminate whether

191: $h_1^\perp$ effects or QCD vacuum effects~\cite{bnm93} are preferred

192: by data. The later effects can also explain the observed asymmetry

193: in a quite different way. The idea is that the non-perturbative

194: vacuum structure in QCD can lead to fluctuating chromomagnetic

195: vacuum fields. When the $q$ and $\bar{q}$ travel through these

196: fields the spin orientations of the quark pair might be

197: correlated~\cite{nr84} before they annihilate to $\gamma^*$. In

198: Ref.~\cite{bnm93} a factorization violating spin correlation is

199: proposed and used to fit the data of the NA10 experiments. According

200: to the color interaction nature of these vacuum effects, the

201: ``correlation strengths" for different quark flavors are the same.

202: Therefore vacuum effects imply no flavor dependence of the

203: asymmetries in different processes, which is different from

204: $h_1^\perp$ effects.

205: 

206: 

207: The general form of the angular differential cross section for

208: unpolarized Drell-Yan process is

209: \begin{eqnarray}

210: \frac{1}{\sigma}\frac{d\sigma}{d\Omega}&=&\frac{3}{4\pi}\frac{1}{\lambda+3}

211: \left

212: (1+\lambda\textmd{cos}^2\theta+\mu\textmd{sin}2\theta\textmd{cos}\phi

213: +\frac{\nu}{2}\textmd{sin}^2\theta\textmd{cos}2\phi\right ).

214: \end{eqnarray}

215: 

216: 

217: 

218: The $\cos2\phi$ angular dependence of the lepton pair can be

219: analyzed in the Collins-Soper frame, in which the unpolarized

220: differential cross section is expressed as

221: \begin{eqnarray}

222: &&\hspace{-1cm}\frac{d\sigma(h_Ah_B\rightarrow l\bar{l}X)}{d\Omega

223: dx_1dx_2d^2\mathbf{q}_\perp}=

224: \frac{\alpha^2}{3Q^2}\sum_ae_a^2\Bigg{\{}

225: A(y)\mathcal{F}[f_{1a/A}(x_1,\mathbf{p}_\perp^2)f_{1\bar{a}/B}(x_2,\mathbf{k}_\perp^2)] +B(y)\textmd{cos}2\phi\nonumber\\

226: &&\hspace{-1cm}\times\mathcal{F}\left [(2\hat{\mathbf{h}}\cdot

227: \mathbf{p}_\perp\hat{\mathbf{h}}\cdot \mathbf{k}_\perp

228: -\mathbf{p}_\perp\cdot

229: \kp)\frac{h_{1a/A}^\perp(x_1,\mathbf{p}_\perp^2)

230: h_{1\bar{a}/B}^\perp(x_2,\mathbf{k}_\perp^2)}{M_AM_B}\right

231: ]\Bigg{\}}+(a \leftrightarrow \bar{a}), \label{cs}

232: \end{eqnarray}

233: where the notation

234: \begin{eqnarray}

235: \mathcal{F}[\cdots]&=&\int d^2\mathbf{p}_\perp

236: d^2\kp\delta^2(\mathbf{p}_\perp+\kp-\mathbf{q}_\perp)\times\{\cdots\}

237: \end{eqnarray}

238: shows the convolution of transverse momenta, and

239: \begin{eqnarray}

240: A(y)&=&\left ( \frac{1}{2}-y+y^2 \right )=\frac{1}{4}(1+\cos^2\theta),\\

241: B(y)&=&y(1-y)=\frac{1}{4}\sin^2\theta

242: \end{eqnarray}

243: in the c.s. frame of the lepton pair. From Eq.~(\ref{cs}) we can

244: write down the expression for the $\cos2\phi$ asymmetry

245: coefficient ($\lambda=1$, $\mu=0$)

246: \begin{eqnarray}

247: \nu=\frac{2\mathcal{F}\left [(2\hat{\mathbf{h}}\cdot

248: \mathbf{p}_\perp\hat{\mathbf{h}}\cdot \mathbf{k}_\perp

249: -\mathbf{p}_\perp\cdot \kp){h_{1a/A}^\perp

250: h_{1\bar{a}/B}^\perp}\right

251: ]}{{M_AM_B}\mathcal{F}[f_{1a/A}f_{1\bar{a}/B}]}.

252: \end{eqnarray}

253: 

254: The azimuthal angel dependent part can be picked up by a weighting

255: procedure~\cite{km97}:

256: \begin{eqnarray}

257: \left \langle \frac{q_T^2 \cos 2 \phi}{4M_A M_B} \right \rangle &

258: = & \int d\phi d^2q_T \frac{d\sigma(h_Ah_B\rightarrow

259: l\bar{l}X)}{d\Omega dx_1dx_2d^2\mathbf{q}_\perp} \cdot \frac{q_T^2

260: \cos 2 \phi}{4M_A M_B}\label{weight}\\

261: &=&\frac{\alpha^2}{3Q^2}B(y)\sum_ae_a^2h_{1a/A}^{\perp(1)}(x_1)h_{1\bar{a}/B}^{\perp(1)}(x_2)+(a

262: \leftrightarrow \bar{a}),

263: \end{eqnarray}

264: in which

265: \begin{equation}

266: h_{1a/H}^{\perp(1)}(x)=\int d^2k_\perp \frac{k^2_\perp}{2M^2_H}

267: h_{1a/H}^\perp(x,k^2_\perp)

268: \end{equation}

269: is the first $k^2_\perp$-moment of $h_{1a/H}^\perp(x,k^2_\perp)$.

270: 

271: 

272: 

273: \begin{figure}

274: 

275: \begin{center}

276: \scalebox{0.9}{\includegraphics{bmud.eps}} \caption{\small

277: $xh_1^{\perp(1),u}(x)$ and $xh_1^{\perp(1),d}(x)$ from different

278: models or theoretical considerations: (a) MIT bag model, (b)

279: spectator model with axial-vector diquark, and (c) large-$N_c$

280: limit.} \label{bmud}

281: \end{center}

282: 

283: \end{figure}

284: 

285: 

286: Because of the simple structure of the pion compared to that of

287: the nucleon, it is advantageous to use a pion beam to unravel the

288: quark structure of the nucleon in the Drell-Yan process. In

289: leading twist there are only two distribution functions for the

290: pion, the usual momentum distribution function and the

291: Boer-Mulders function. In Refs.~\cite{lm04,lm05} calculations

292: based on a quark spectator antiquark model showed that the

293: Boer-Mulders functions of the two valence quarks inside the pion

294: (such as $\bar{u}$ and $d$ inside $\pi^-$) are equal. Actually

295: this relation can be obtained from model independent principles:

296: charge conjugation symmetry and isospin symmetry. In the following

297: we will use $h_{1\pi}^{\perp}$ and $\bar{h}_{1\pi}^{\perp}$ to

298: denote the Boer-Mulders functions of the valence and sea quarks

299: inside the pion respectively, and $h_{1}^{\perp,{q(\bar{q})}}$ for

300: the Boer-Mulders functions of the quark flavor $q(\bar{q})$ inside

301: the proton. Thus the weighted cross section of the unpolarized

302: $\pi^- p$ Drell-Yan process is ($W=Q_T^2 \cos 2 \phi/4M_A M_B$)

303: \begin{eqnarray}

304: \left \langle W\right \rangle_{\pi^-

305: p}(x_1,x_2)&=&\frac{\alpha^2}{3Q^2}B(y)\left

306: \{h_{1\pi}^{\perp(1)}(x_1)\left [e_u^2h_{1}^{\perp(1),u}(x_2)

307: +e_d^2h_{1}^{\perp(1),\bar{d}}(x_2))\right ]\right.\nonumber\\

308: &&\left.+\bar{h}_{1\pi}^{\perp(1)}(x_1)\left

309: [e_u^2h_{1}^{\perp(1),\bar{u}}(x_2)

310: +e_d^2h_{1}^{\perp(1),d}(x_2))\right ]\right \}.

311: \end{eqnarray}

312: 

313: 

314: We are interested in the region were $x_{1/2}$ are not so small,

315: where the sea quark contribution to the $\cos2\phi$ angular

316: dependent cross section is expected to be small. Then we can use

317: the approximation

318: \begin{eqnarray}

319: \left \langle W\right \rangle_{\pi^-

320: p}(x_1,x_2)\approx\frac{4\alpha^2}{3Q^2}B(y)e_u^2h_{1\pi}^{\perp(1)}(x_1)h_{1}^{\perp(1),u}(x_2).

321: \end{eqnarray}

322: 

323: 

324: The cross section for the $\pi^- D$ Drell-Yan process can be

325: approximated as $\sigma_{\pi^- D} = \sigma_{\pi^- p}+\sigma_{\pi^-

326: n}$, thus

327: \begin{eqnarray}

328: \langle W \rangle_{\pi^- D}(x_1,x_2)=\langle W \rangle_{\pi^-

329: p}(x_1,x_2)+\langle W \rangle_{\pi^- n}(x_1,x_2).

330: \end{eqnarray}

331: The Boer-Mulders functions of quarks inside the neutron can be

332: related to those in the proton: $h_{1u/n}^{\perp}=h_1^{\perp,d}$,

333: $h_{1d/n}^{\perp}=h_1^{\perp,u}$. Therefore

334: \begin{eqnarray}

335:  \left \langle W\right \rangle_{\pi^-

336: n}(x_1,x_2)&=&\frac{\alpha^2}{3Q^2}B(y)\left

337: \{h_{1\pi}^{\perp(1)}(x_1)\left [e_u^2h_{1}^{\perp(1),d}(x_2)

338: +e_d^2h_{1}^{\perp(1),\bar{u}}(x_2))\right ]\right.\nonumber\\

339: &&\left.+\bar{h}_{1\pi}^{\perp(1)}(x_1)\left

340: [e_u^2h_{1}^{\perp(1),\bar{d}}(x_2)

341: +e_d^2h_{1}^{\perp(1),u}(x_2))\right ]\right \}.

342: \end{eqnarray}

343: Ignoring the sea quark contribution, we have

344: \begin{eqnarray}

345: \left \langle W\right \rangle_{\pi^-

346: D}(x_1,x_2)\approx\frac{\alpha^2}{3Q^2}B(y)e_u^2h_{1\pi}^{\perp(1)}(x_1)

347: (h_{1}^{\perp(1),u}(x_2)+h_{1}^{\perp(1),d}(x_2)).

348: \end{eqnarray}

349: Then the ratio of the weighted cross sections for unpolarized

350: $\pi^- p$ and $\pi^- D$ Drell-Yan processes is

351: \begin{eqnarray}

352: \frac{\langle W \rangle_{\pi^- D}(x_1,x_2)}{2 \langle W

353: \rangle_{\pi^- p}(x_1,x_2)}=\frac{1}{2} \left (

354: 1+\frac{h_{1}^{\perp(1),d}(x_2)}{h_{1}^{\perp(1),u}(x_2)} \right

355: ).\label{Doverp}

356: \end{eqnarray}

357: 

358: The Drell-Yan process at COMPASS can explore the kinematics region

359: $s=400\, \textrm{GeV}^2$ and $Q^2=20 \, \textrm{GeV}^2$, where

360: $x_1 x_2=Q^2/s=0.05$. Therefore the Drell-Yan experiments at

361: COMPASS can access the quark structure in the valence regime $0.1

362: \leq x_{1/2} \leq 0.5$, where Eq.~(\ref{Doverp}) is a good

363: approximation. Then the ratio of the Boer-Mulders functions of $d$

364: and $u$ quarks is

365: \begin{eqnarray}

366: \frac{h_{1}^{\perp(1),d}(x_2)}{h_{1}^{\perp(1),u}(x_2)}=\frac{\langle

367: W \rangle_{\pi^- D}(x_1,x_2)}{\langle W \rangle_{\pi^-

368: p}(x_1,x_2)}-1.\label{doveru}

369: \end{eqnarray}

370: 

371: Another possibility is to use a $\pi^+$ beam to collide on the

372: proton, which process is sensitive to the $d$ quark distribution

373: of the proton:

374: \begin{eqnarray}

375: \frac{\langle W \rangle_{\pi^+ p}(x_1,x_2)}{\langle W

376: \rangle_{\pi^-

377: p}(x_1,x_2)}=\frac{h_{1}^{\perp(1),d}(x_2)}{4h_{1}^{\perp(1),u}(x_2)}.\label{plovermi}

378: \end{eqnarray}

379: This relation can serve as cross check for Eq.~(\ref{doveru}).

380: 

381: 

382: knowledge of the Boer-Mulders functions is very poor, including

383: its size and shape, and the relative size between $h_1^{\perp,u}$

384: and $h_1^{\perp,d}$. In the last case different models or

385: theoretical considerations predict quite different values for the

386: ratio $h_1^{\perp,d}/h_1^{\perp,u}$. A calculation based on the

387: MIT bag model gave~\cite{yuan}

388: $h_1^{\perp,d}=\frac{1}{2}h_1^{\perp,u}$. The spectator model with

389: axial-vector diquarks predicted~\cite{bsy04} that $h_1^{\perp,d}$

390: has a opposite sign and a rather larger size compared with

391: $h_1^{\perp,u}$, while large-$N_c$ showed that

392: $h_1^{\perp,d}=h_1^{\perp,u}$~\cite{pobylista} modulo $1/N_c$

393: corrections. In Fig.~\ref{bmud} we show the curves for

394: $h_1^{\perp(1),u}(x)$ and $h_1^{\perp(1),d}(x)$ from these models

395: or theoretical considerations. In the case of large-$N_c$ since

396: there is no detail calculation on the size of $h_1^\perp$, we just

397: naively assume $h_1^{\perp(1),u}=f_{1T}^{\perp(1),u}$ (these two

398: function are closely related and expected to have similar size),

399: and for $f_{1T}^{\perp(1),u}$ we adopt the

400: parametrization~\cite{efremov} extracted from HERMES experimental

401: data based on large-$N_c$. In the case of MIT bag model the ratio

402: in Eq.~(\ref{Doverp}) is $0.75$ and in large-$N_c$ consideration

403: is 1, while in axial-vector diquark model the ratio is mostly

404: negative. These predictions can be tested by measuring the ratio

405: $\langle W \rangle_{\pi^- D}/(2\langle W \rangle_{\pi^- p})$. Vice

406: versa, according to Eq.~(\ref{doveru}), from the measured ratio

407: $\langle W \rangle_{\pi^- D}/(2\langle W \rangle_{\pi^- p})$ the

408: size of $h_1^{\perp,d}/h_1^{\perp,u}$ can be determined.

409: 

410: The analysis can be easily extended to the single (transverse)

411: polarized Drell-Yan processes $\pi p^\uparrow(D^\uparrow)

412: \rightarrow \mu^+\mu^- X$ at COMPASS. The measurement of the

413: single-spin asymmetries in this process is also important for the

414: more precise extraction of the Sivers function (through

415: $f_{1\pi}\otimes f_{1T}^\perp$), and to test the QCD prediction

416: $f_{1T}^\perp|_{\small \textrm{SIDIS}}=-f_{1T}^\perp|_{\small

417: \textrm{DY}}$~\cite{collins02,efremov2}, and can provide

418: alternative access to transversity (through $h_{1\pi}^\perp\otimes

419: h_1$)~\cite{boer}. In single polarized Drell-Yan processes the

420: relations similar to Eqs.~(\ref{Doverp}) and (\ref{plovermi}) can

421: be deduced for the Sivers function and for transversity.

422: 

423: \begin{figure}

424: 

425: \begin{center}

426: \scalebox{0.82}{\includegraphics{weiasy.eps}} \caption{\small The

427: weighted $\cos 2 \phi$ asymmetries in unpolarized $\pi ^- p$

428: (solid line) and $\pi^- D$ (dashed line) Drell-Yan processes. (a)

429: The MIT Bag model result. The two asymmetries are equal since

430: $f_1^d/f_1^u=1/2$ in the lowest bag mode in which case is used in

431: calculation. (b) The axial-vector diquark model result. (c) The

432: result from large-$N_c$.} \label{weiasy}

433: \end{center}

434: 

435: \end{figure}

436: 

437: It is also interesting to compare the $\cos 2 \phi$ asymmetries in

438: unpolarized $\pi^- p$ and $\pi^- D$ processes. The anomalous

439: Drell-Yan asymmetry might be produced by $h_1^\perp$, which is

440: labelled as hadronic effect. Nevertheless, there is another effect

441: that could also account for this asymmetry, the so called QCD

442: vacuum effect~\cite{bnm93}. A significant difference between these

443: two effects~\cite{bbnu} is that the first one is flavor dependent,

444: while the latter is flavor blind. Thus comparing the asymmetries

445: in different processes can discriminate which effect is really

446: responsible for the asymmetry. In the case of the hadronic effect,

447: the weighted $\cos 2\phi$ asymmetries of these two processes can

448: be calculated from following expressions:

449: \begin{eqnarray}

450: \hat{\nu}_p&=& a_{UU}\frac{\langle W \rangle_{\pi^- p}}{\langle\,

451: 1 \rangle_{\pi^- p}} =

452: \frac{h_{1\pi}^{\perp(1)}(x_1)h_{1}^{\perp(1),u}(x_2)}{f_{1\pi}(x_1)f_1^u(x_2)},\label{asyp}\\

453: \hat{\nu }_D&=& a_{UU}\frac{\langle W \rangle_{\pi^- D}}{\langle

454: \, 1\rangle_{\pi^- D}}=

455: \frac{}{}\frac{h_{1\pi}^{\perp(1)}(x_1)(h_{1}^{\perp(1),u}(x_2)+

456: h_{1}^{\perp(1),d}(x_2))}{f_{1\pi}(x_1)(f_1^u(x_2)+f_1^d(x_2))},\label{asyd}

457: \end{eqnarray}

458: where $a_{UU}=(1+\cos^2\theta)/\sin^2\theta$, and $\langle 1

459: \rangle$ is calculated from Eq.~(\ref{weight}) when $W$ is

460: replaced by $1$. Again we neglect the sea contribution here. The

461: ratio of the asymmetries in these two processes is then

462: \begin{eqnarray}

463: \frac{\hat{\nu}_D}{\hat{\nu

464: }_p}=\frac{1+\frac{h_1^{\perp(1),d}(x_2)}{h_1^{\perp(1),u}(x_2)}}{1+\frac{f_1^d(x_2)}{f_1^u(x_2)}}.

465: \end{eqnarray}

466: The QCD vacuum effect will yield $\hat{\nu}_p=\hat{\nu}_D$. In

467: Fig.~\ref{weiasy} we show the asymmetries calculated from

468: Eq.~(\ref{asyp}) and (\ref{asyd}) using the three options of

469: Boer-Mulders functions given in Fig.~\ref{bmud}. For the

470: unpolarized distributions, which appear in the denominators in

471: Eq.~(\ref{asyp}) and (\ref{asyd}) we adopt well known model

472: results. We do not aim at a precise estimate of the asymmetries,

473: only to a rough comparison among them. In the case that

474: $h_1^{\perp(1),u}/f_1^u$ is different from

475: $h_1^{\perp(1),d}/f_1^d$ (equivalently

476: $h_1^{\perp(1),d}/h_1^{\perp(1),u}$ different from $f_1^d/f_1^u$),

477: different asymmetries will be observed in these two processes, as

478: shown in Fig.~\ref{weiasy}b and Fig.~\ref{weiasy}c. Therefore the

479: difference appearing in the $\cos 2 \phi$ asymmetries in

480: unpolarized $\pi^- p$ and $\pi^- D$ Drell-Yan processes can

481: provide evidence whether hadronic effects or QCD vacuum effects

482: are preferred by data.

483: 

484: In summary, we have showed that by comparing the $\cos 2 \phi$

485: angular dependent part of the cross sections of the unpolarized

486: $\pi^- p$ and $\pi^- D$ Drell-Yan processes, which will be

487: accessible in future COMPASS experiments, one can determine the

488: ratio $h_1^{\perp,d}/h_1^{\perp,u}$ in the valence region, which

489: still lacks theoretical input. By checking the difference of the

490: $\cos 2 \phi$ asymmetries in these two processes, one can also

491: provide a discrimination whether hadronic effects or QCD vacuum

492: effects are preferred by data. The investigation will lead to

493: better understanding about the role of the quark transverse motion

494: and transverse spin in hadronic reactions.

495: 

496: {\bf Acknowledgements.} This work is partially supported by

497: National Natural Science Foundation of China (Nos.~10421503,

498: 10575003, 10505001, 10528510), by the Key Grant Project of Chinese

499: Ministry of Education (No.~305001), by the Research Fund for the

500: Doctoral Program of Higher Education (China), by Fondecyt (Chile)

501: under Project No.~3050047 and No.~1030355.

502: 

503: \begin{thebibliography}{99}

504: 

505: \bibitem{bdr}

506: 

507: For a review on tranverse polarization phenomena, see V.~Barone,

508: A.~Drago, P.G.~Ratcliffe, Phys. Rep. 359 1 (2002).

509: 

510: \bibitem{smc} A. Bravar et al., SMC Collaboration, Nucl. Phys. A 666 (2000) 314.

511: 

512: \bibitem{Airapetian:2004tw}

513: 

514: A.~Airapetian et al., HERMES Collaboration, Phys. Rev. Lett. 94

515: (2005) 012002.

516: 

517: 

518: 

519: \bibitem{compass}

520: 

521: V.Yu.~Alexakhin et al., COMPASS Collaboration, Phys. Rev. Lett. 94

522: (2005) 202002.

523: 

524: \bibitem{hermes05} M. Diefenthaler, HERMES Collaboration, in Proceedings of DIS 2005, Madison, Wisconsin (USA), hep-ex/0507013.

525: 

526: %\bibitem{abm95} M. Anselmino, M. Boglione, F. Murgia,

527: 

528: %Phys. Lett. {\bf B362} (1995) 164.

529: 

530: 

531: \bibitem{sivers} D. Sivers, Phys. Rev. D 41 (1990) 83;\\

532: D. Sivers, Phys. Rev. D 43 (1991) 261.

533: 

534: 

535: \bibitem{collins93} J.C.~Collins, Nucl. Phys. B 396 (1993) 161.

536: 

537: \bibitem{anselmino95}

538: 

539: M. Anselmino, M. Boglione, F. Murgia, Phys. Lett. B 362

540: 164 (1995). \\

541:  M. Anselmino, F. Murgia, Phys. Lett. B 442, 470

542: (1998).

543: 

544: 

545: \bibitem{bm}

546: 

547: D.~Boer, P.J.~Mulders, Phys. Rev. D 57 (1998) 5780.

548: 

549: \bibitem{bhs02} S.J.~Brodsky, D.S.~Hwang,

550:  I.~Schmidt, Phys. Lett. B 530 (2002) 99;\\

551: S.J.~Brodsky, D.S.~Hwang, I.~Schmidt, Nucl. Phys. B 642 (2002)

552: 344.

553: 

554: \bibitem{collins02} J.C. Collins,  Phys. Lett. B 536 (2002) 43.

555: 

556: 

557: 

558: \bibitem{belitsky}

559: X.~Ji and F.~Yuan, Phys. Lett. B 543

560: (2002) 66;\\

561: A.V.~Belitsky, X.~Ji, F.~Yuan, Nucl. Phys. B 656 (2003) 165; \\

562: D. Boer, P.J. Mulders, F. Pijlman, Nucl. Phys. B 667 (2003) 201.

563: 

564: 

565: \bibitem{levelt}

566: 

567: J.~Levelt and P.J.~Mulders, Phys. Rev. D 49 (1994) 96.

568: 

569: 

570: 

571: \bibitem{kotzinian}

572: 

573: A.~Kotzinian, Nucl. Phys. B 441 (1995) 234.

574: 

575: 

576: 

577: \bibitem{mulders}

578: 

579: P.J.~Mulders and R.D.~Tangerman, Nucl. Phys. B 461 (1996)

580: 

581: 

582: 

583: \bibitem{na10} S.~Falciano et al. NA10 Collaboration,

584: Z. Phys. C 31 (1986) 513;\\

585: M.~Guanziroli et al. NA10 Collaboration, Z. Phys. C 37 (1988) 545.

586: 

587: \bibitem{conway} J.S.~Conway et al.

588: Phys. Rev. D 39 (1989) 92.

589: 

590: \bibitem{boer}

591: 

592: D.~Boer, Phys. Rev. D 60 (1999) 014012.

593: 

594: 

595: 

596: 

597: \bibitem{gg02}  G.R. Goldstein, L. Gamberg, Talk given at 31st International

598: Conference on High Energy Physics (ICHEP 2002), Amsterdam, The

599: Netherlands, 24-31 July 2002, hep-ph/0209085.

600: 

601: \bibitem{bbh03} D.~Boer, S.J.~Brodsky, D.S.~Hwang, Phys.

602: Rev. D 67 (2003) 054003.

603: 

604: 

605: 

606: 

607: %\bibitem{Kre01}

608: %S.~Kretzer, E.~Leader, and E.~Christova, Eur. Phys. J. {\bf C22}

609: %(2001) 269.

610: 

611: 

612: 

613: \bibitem{yuan}

614: F. Yuan, Phys. Lett. B 575 (2003) 45.

615: 

616: \bibitem{bsy04} A.~Bacchetta, A.~Sch\"{a}fer,

617:  J.-J.~Yang, Phys. Lett. B 578 (2004) 109.

618: 

619: %\bibitem{lm04a} Z.~L{\"u} and B.-Q.~Ma, Nucl. Phys. {\bf A741}

620: 

621: %(2004) 200.

622: 

623: 

624: 

625: \bibitem{lm04}

626: Z.~Lu, B.-Q.~Ma, Phys. Rev. D 70 (2004) 094044.

627: 

628: \bibitem{gg05} G. R. Goldstein, L. Gamberg, hep-ph/0506127.

629: 

630: \bibitem{blm06} V. Barone, Z.~Lu, B.-Q.~Ma, hep-ph/0612350, to appear in European Physical Journal C.

631: 

632: \bibitem{lm05}

633: Z.~Lu, B.-Q.~Ma, Phys. Lett. B 615 (2005) 200.

634: 

635: 

636: \bibitem{gg03} L.P.~Gamberg, G.R.~Goldstein, K.A.~Oganessyan, Phys. Rev. D 67 (2003) 071504;\\

637:  L.P.~Gamberg, G.R.~Goldstein

638: , K.A.~Oganessyan, Phys. Rev. D 68 (2003) 051501(R).

639: 

640: 

641: 

642: 

643: \bibitem{blm05} V. Barone, Z. Lu, B.-Q. Ma, Phys. Lett. B 632

644: (2006) 277.

645: 

646: \bibitem{anselmino05c} M. Anselmino, M. Boglione, U. D'lesio, A. Kotzinian, F. Murgia,

647:  A. Prokudin, Phys. Rev. D 72 (2005) 094007;\\

648:  M. Anselmino, M. Boglione, U. D'lesio, A. Kotzinian, F. Murgia,

649:  A. Prokudin, Phys. Rev. D 72 (2005) 099903, Erratum.

650: 

651: \bibitem{vy05} W. Vogelsang , F. Yuan, Phys. Rev. D 72 (2005) 054028.

652: 

653: 

654: \bibitem{efremov}

655: J.C. Collins, A.V. Efremov, K. Goeke, S. Menzel, A. Metz, P.

656: Schweitzer,  Phys. Rev. D73 (2006) 014021;\\

657: J.C. Collins, A. V. Efremov, K. Goeke, M. Grosse Perdekamp, S.

658: Menzel, B. Meredith, A. Metz, P. Schweitzer, hep-ph/0510342.

659: 

660: 

661: 

662: \bibitem{anselmino05b} For a

663: comparison to this extraction, see M. Anselmino et al.,

664: hep-ph/0511017.

665: 

666: \bibitem{pobylista} P.V. Pobylitsa, hep-ph/0301236.

667: 

668: \bibitem{bnm93} A. Brandenburg, O. Nachtmann, E. Mirkes, Z. Phys. C 60 (1993) 697.

669: 

670: \bibitem{nr84} O. Nachtmann, A. Reiter, Z. Phys. C 24 (1984) 283.

671: 

672: \bibitem{km97}

673: 

674: A.~Kotzinian, P.J. Mulders, Phys. Letts. B 406 (1997) 373.

675: 

676: \bibitem{efremov2} A.V. Efremov, K. Goeke, S. Menzel, A. Metz, P. Schweitzer,

677: Phys. Lett. B 612 (2005) 233.

678: 

679: \bibitem{bbnu} D. Boer, A. Brandenburg, O. Nachtmann,  Utermann, Eur.

680: Phys. J. C 40 (05) 55.

681: 

682: 

683: 

684: 

685: 

686: 

687: 

688: 

689: 

690: 

691: 

692: \end{thebibliography}

693: 

694: \end{document}

695: