hep-lat0109028/ss.tex
1: \documentclass[fleqn,twoside]{article}
2: \usepackage{espcrc2}
3: 
4: \usepackage{graphicx}
5: 
6: \newcommand{\be}{\begin{equation}}
7: \newcommand{\ee}{\end{equation}}
8: %
9: \title{The strangeness content of the nucleon }
10: 
11: \author{UKQCD Collaboration, 
12:         C.~Michael\address{
13:  Theoretical Physics Division, Dept. of Math. Sci., \\
14:           University of Liverpool, Liverpool L69 3BX, UK},
15:  C. McNeile\addressmark and
16:  D. Hepburn\address{
17:  Dept. of Physics, University of Edinburgh, Edinburgh EH9 3JZ, UK}%
18:  } 
19: 
20: \begin{document}
21: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
22: \begin{abstract}
23:  We evaluate the matrix element of $\bar{q} q$ in hadron states on a
24: lattice. We find substantial mixing of the connected and disconnected
25: contributions so that the lattice result that the disconnected
26: contribution to the nucleon is large does not imply that the $\bar{s} s$
27: content is large. This has implications for dark matter searches. 
28: \vspace{1pc}
29: \end{abstract} 
30: % 
31: \maketitle
32: 
33: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
34: 
35: \section{INTRODUCTION}
36: 
37: 
38:  An important challenge in physics and cosmology is to understand the
39: nature of dark matter.  One plausible candidate is for this dark matter
40: to be the lightest supersymmetric  particle: the neutralino. In this
41: case the dark matter can be detected by scattering from nuclear targets,
42:  and experimental explorations are currently under way. To extract a
43: physical  flux from such experiments, one needs the appropriate cross
44: section for scattering  of a neutralino off a nucleon. This has been
45: evaluated~\cite{Bottino:1999ei} and depends on  MSSM parameters and on
46: the QCD matrix elements of the scalar quark current: $ \langle N |
47: \bar{q} q | N \rangle$.  
48:  The Higgs exchange terms are dominant and hence the scalar current 
49: enters multiplied by the relevant quark mass. For this reason, the
50: strange quark contribution is expected to  be especially important.
51: Moreover the $u$ and $d$ quark contributions are related to the  $\pi N$
52: $\sigma$ term and are relatively well known whereas the strange
53: contribution  is unknown phenomenologically to within a factor of 3. 
54: 
55:  The usual way to parametrise the contribution of the strange quark is by 
56:  \be
57:  y = {2 \langle N | \bar{s} s | N \rangle  \over 
58:         \langle N | \bar{u} u + \bar{d} d | N \rangle }
59:  \ee
60:  Estimates~\cite{Borasoy:1996zx} using chiral perturbation theory
61: suggest $y = 0.2(2) $. 
62: 
63: 
64:  This ratio of quark matrix elements can be calculated in principle
65: using lattice methods (see~\cite{Gusken:1999te} for a review). The
66: required  correlations are illustrated in fig.~1. The disconnected
67: contribution $D_3$ (numerator of $y$) can either be calculated by 
68: evaluating the three-point correlator with a disconnected loop or using
69: the  lattice equivalent of the Feynman-Hellman theorem  to relate it to
70: the derivative of the  nucleon mass with respect to the sea-quark
71: hopping parameter (see ref.~\cite{Foster:1998vw} for a discussion  of
72: this).  Likewise the connected contribution $C_3$ can be evaluated
73: either as a three-point correlation or as a derivative of the nucleon
74: mass  with respect to the valence-quark hopping parameter. 
75:  Then, in this approach we expect $y = D_3/(C_3+D_3)$.
76: 
77: 
78: 
79: \begin{figure}[htb]
80: %\vspace{-0.7cm}
81: %\epsfxsize=7.5cm\epsfbox{c2.ps}
82: \includegraphics[width=7.5cm]{c2.ps}
83: %\vspace{-1cm}
84:  \caption{Connected and disconnected diagrams}
85:  \label{fif.c2}
86: \end{figure}
87: %\vspace{-0.3cm} 
88: 
89: 
90:  Using $N_f=2$ flavours of sea-quark, SESAM~\cite{Gusken:1998wy} obtain
91: $y=0.59(13)$ and, using the  same method, we would obtain a similar
92: result. Using the  three-point correlator approach, it is possible to
93: estimate $y$ in quenched  studies and results of around 0.6 were
94: obtained~\cite{Fukugita:1995ba} although the  Kentucky group 
95: argued~\cite{Dong:1996ec} that renormalisation effects should reduce
96: this to around  0.36(3) (but see criticism of this approach in
97: ref~\cite{Gusken:1998wy}).
98: 
99: 
100:  This large value of $y$ obtained from lattice studies is surprising 
101: and it also has major implications for the analysis of dark matter
102: scattering experiments.  Here we discuss the status of these lattice
103: determinations critically and we conclude that $y$ is consistent with
104: zero.
105: 
106: 
107: 
108: \section{LATTICE ANALYSIS}
109: 
110: 
111: 
112: 
113:  The Feynman-Hellman theorem relates matrix elements of scalar quark
114: currents  in a nucleon to derivatives of the nucleon mass with respect
115: to the quark mass. These identities can be derived both in the continuum
116: and on the lattice.  Here we consider Wilson-like
117: lattice fermion formulations. 
118:  The lattice equivalent of the Feynman-Hellman theorem is
119: that~\cite{Foster:1998vw} the following lattice observables (here we
120: define $m^b \equiv 1/(2 \kappa)$) are related:
121: 
122:  \begin{equation}
123: \frac{ \partial (aM_N)} { \partial m_{val}^b }
124: = 
125: \lim_{ t_1 , (t-t_1) \rightarrow \infty} 
126: \frac{ C_3(t_1,t) }{ C(t) }
127:  \end{equation}
128: 
129:  \begin{equation}
130: \frac{ \partial (aM_N)} { \partial m_{sea}^b }
131: = - N_f
132: \lim_{ t_1 , (t-t_1) \rightarrow \infty} 
133: \frac{ D_3(t_1,t) }{ C(t) }
134:  \end{equation}
135: 
136: 
137: 
138: 
139: The  input quark mass parameters  can be written in terms of the
140: physical bare quark masses, with  an additive mass renormalisation
141: ($m_A$).
142: 
143: \be
144: m_{val}^{b}  =  m_{A} + a m_{val} \label{eq.val} 
145: \ee
146: \be
147: m_{sea}^{b}  =  m_{A} + a m_{sea} \label{eq.sea}
148: \ee
149: 
150:  Now  $m_A$ may be determined by varying the valence quark mass
151: and  determining the critical hopping parameter at which the pion mass
152: (and hence $a m_{val}$)  becomes zero, or equivalently by finding the
153: critical valence hopping parameter at which the PCAC mass becomes zero. 
154: These extrapolations to determine $m_A$ are at fixed
155: $m_{sea}^b$ and  hence $m_A$ will depend on $m_{sea}^b$. 
156:  The lattice spacing $a$ also depends on the bare sea-quark 
157: mass parameter $m_{sea}^b$. 
158:  These effects can be summarised by the following derivatives
159:  \be 
160:  X={d m_A \over d m_{sea}^b}
161:  {\rm \ \ \ \ and \ \ \ \ \ }
162:  B = {d \log a \over d m_{sea}^b}
163:  \ee
164: 
165: 
166: 
167: We are interested in the  disconnected  scalar matrix element which is
168: related in the continuum to the derivative with  respect to the sea
169: quark mass.  On a lattice this   disconnected scalar matrix element is
170: related to the derivative of the  nucleon mass with respect to the
171: sea-quark mass  parameter at fixed  valence-quark hopping parameter and
172: fixed $\beta$ by eq.~\ref{eq.sea}. The key observation is then that as
173: the sea-quark hopping parameter is varied on the lattice, both  the
174: valence quark mass and the lattice spacing also change. Thus one needs
175: to  correct for these  changes to obtain the derivative of the nucleon
176: mass at fixed valence-quark mass which  is required.  
177: 
178: 
179: 
180:  For  the valence dependence, the required derivative is 
181: directly given by the lattice observable
182:  \be
183: \frac{ \partial M_N} { \partial m_{val} }
184: =\frac{ \partial (aM_N)} { \partial m_{val}^b }
185:  \ee
186:  For the sea-quark derivative, however, there will be several other
187: factors involved as discussed above:
188:  \begin{eqnarray}
189: \frac{ \partial M_N} { \partial m_{sea} } (1- a m_{sea} B - X)  
190: =   \ \ \ \ \ \  & & \nonumber \\
191:   \frac{ \partial (aM_N)} { \partial m_{sea}^b }  
192:  + ( a m_{val} B + X)\frac{ \partial (aM_N)} { \partial m_{val}^b } 
193:  - M_N B & &
194:  \label{eq.full}
195:  \end{eqnarray}
196: 
197:  This shows that the lattice connected and disconnected contributions
198: are mixed  when related to the more physical derivatives at fixed bare
199: quark mass and  fixed scale $a$. There will  also be additional
200: perturbative matching  contributions to take into account to have a
201: precise link between lattice observables and the continuum expressions.,
202: but we do not discuss these further here. 
203: 
204: 
205:  We evaluate the above expressions using UKQCD
206: data~\cite{Allton:1998gi,Allton:2001sk}. We mainly use data from 
207: $\beta=5.2$ on $16^3 32$ lattices with $N_f=2$ flavours of sea quark
208: with $\kappa=0.1355$ or $0.1350$ and using a  NP-clover formalism with
209: $C_{SW}=2.0171$. These hopping parameter values correspond to quark masses
210: around the strange quark (since the  $\pi / \rho$ ratio is 0.58 and
211: 0.70 respectively). From the $r_0$ values~\cite{Allton:2001sk}, we obtain
212:  $B=4.4(8)$ while from extrapolating the PCAC masses to obtain the
213: critical hopping parameters  at these two sea quark masses we obtain
214: $X=-0.66(4)$. This implies that there will be substantial mixing of the
215: lattice disconnected  and connected contributions in evaluating the
216: derivative with respect to the sea-quark mass. 
217: 
218:  Setting $B=X=0$ in eq.~\ref{eq.full} gives the naive lattice ratio of
219: $y=0.53(12)$  while including the full mixing gives $y=-0.28(33)$.
220: 
221: 
222:  Since we are using finite differences to evaluate the derivatives in 
223: eq.~\ref{eq.full}, we may instead use the  value of  the nucleon mass in
224: physical  units for the four cases ($\kappa_{sea}=0.1355,\ 0.1350$; 
225: $\kappa_{val}=0.1355,\ 0.1350$) and evaluate the quark mass in each case
226: from  the bare quark mass (eqs.~{\ref{eq.val},\ref{eq.sea}). We prefer
227: to use the  bare quark mass here because of the lattice Feynman-Hellman
228: theorem, but it would be possible to use the lattice PCAC quark mass,
229: extrapolate to the continuum and  then use the Feynman-Hellman relation in
230: the continuum. The four combinations of sea and valence quark mass will 
231: then not lie at the corners of a rectangle, but at corners of some
232: quadrilateral.  We can then use the nucleon masses at these four values
233: to evaluate the required derivatives with respect  to the sea quark mass
234: at constant valence mass and vice versa. The result from  this approach
235: is $y=-0.30(34)$ which is consistent with the value quoted above. 
236: 
237:  Note that  $y$ is expected to be positive, so the negative value
238: obtained  is just a reflection of the large statistical error. We have
239: also attempted to  measure the disconnected lattice correlator using a
240: three point function approach. The  result was compatible with using
241: derivatives of the nucleon mass  but the statistical errors were larger.
242:  
243: \section{DISCUSSION}
244: 
245:  To obtain a reliable lattice determination of $y$, one should use 
246: $N_f=2$ flavours of light sea quark plus a heavier (strange mass) sea
247: quark.  Then the required disconnected diagram can be obtained either as
248: a derivative of the nucleon mass  with respect to the strange sea quark
249: mass or by evaluating the $D_3$  diagram with strange quarks in the
250: disconnected loop. In practice one will need to extrapolate the sea and
251: valence masses  to  to the physical $u$ and $d$ masses.  This 
252: extrapolation is known to be non-trivial for the  extraction of the $\pi
253: N$ sigma term~\cite{Leinweber:2000sa}. A continuum limit of the lattice
254: observables  should also be taken as well as building in perturbative
255: matching. 
256: 
257: 
258: 
259:  Instead we use $N_f=2$ flavours of sea-quark which should be a good
260: approximation for the sea. Extracting the disconnected diagram as a
261: derivative is only possible if the quarks considered in the disconnected
262: loop are the sea-quarks.  It is possible to go beyond this by explicitly
263: evaluating $D_3$ with different mass  quarks in the disconnected loop
264: and this is in progress.  We are also exploring ways to  reduce the 
265: large statistical errors we find.
266: 
267: 
268:  Our main conclusion is that the current lattice data are unable to 
269: give a determination with any precision of $y$, but that the naive
270: lattice  ratio of $y \approx 0.6$ is not appropriate and the lattice
271: result  is indeed compatible with $y=0$.  That $y$ is compatible with
272: zero  implies that there is no evidence for any dependence of the 
273: nucleon mass  on the sea quark mass  and this conclusion was also 
274: reached in an earlier lattice study~\cite{Foster:1998vw}  of the sea
275: quark dependence of the meson (pseudoscalar and vector) masses. 
276: 
277: 
278: 
279: 
280: % mod for spires tags, hep-lat etc
281: 
282: %%\bibliographystyle{h-elsevier2}
283: %%\bibliography{conferences,wimp} 
284: \begin{thebibliography}{10}
285: 
286: \bibitem{Bottino:1999ei}
287: A. Bottino et~al.,
288: \newblock Astropart. Phys. 13 (2000) 215, hep-ph/9909228,
289: \newblock %%CITATION = APHYE,13,215;%%.
290: 
291: \bibitem{Borasoy:1996zx}
292: B. Borasoy and U.G. Meissner,
293: \newblock Phys. Lett. B365 (1996) 285, hep-ph/9508354,
294: \newblock %%CITATION = PHLTA,B365,285;%%.
295: 
296: \bibitem{Gusken:1999te}
297: S. Gusken,
298: \newblock Flavor singlet phenomena in lattice qcd,
299: \newblock hep-lat/9906034.
300: 
301: \bibitem{Foster:1998vw}
302: UKQCD, M. Foster and C. Michael,
303: \newblock Phys. Rev. D59 (1999) 074503, hep-lat/9810021,
304: \newblock %%CITATION = PHRVA,D59,074503;%%.
305: 
306: \bibitem{Gusken:1998wy}
307: SESAM, S. Gusken et~al.,
308: \newblock Phys. Rev. D59 (1999) 054504, hep-lat/9809066,
309: \newblock %%CITATION = PHRVA,D59,054504;%%.
310: 
311: \bibitem{Fukugita:1995ba}
312: M. Fukugita et~al.,
313: \newblock Phys. Rev. D51 (1995) 5319, hep-lat/9408002,
314: \newblock %%CITATION = PHRVA,D51,5319;%%.
315: 
316: \bibitem{Dong:1996ec}
317: S.J. Dong, J.F. Lagae and K.F. Liu,
318: \newblock Phys. Rev. D54 (1996) 5496, hep-ph/9602259,
319: \newblock %%CITATION = PHRVA,D54,5496;%%.
320: 
321: \bibitem{Allton:1998gi}
322: UKQCD, C.R. Allton et~al.,
323: \newblock Phys. Rev. D60 (1999) 034507, hep-lat/9808016,
324: \newblock %%CITATION = PHRVA,D60,034507;%%.
325: 
326: \bibitem{Allton:2001sk}
327: UKQCD, C.R. Allton et~al.,
328: %%\newblock Effects of non-perturbatively improved dynamical fermions in qcd at
329: %%  fixed lattice spacing,
330: \newblock hep-lat/0107021.
331: 
332: \bibitem{Leinweber:2000sa}
333: D.B. Leinweber, A.W. Thomas and S.V. Wright,
334: \newblock Phys. Lett. B482 (2000) 109, hep-lat/0001007,
335: \newblock %%CITATION = PHLTA,B482,109;%%.
336: 
337: \end{thebibliography}
338: 
339: 
340: 
341: \end{document}
342: 
343: