1: \documentstyle[prd,aps,preprint,tighten,epsfig]{revtex}
2:
3: \begin{document}
4:
5: \draft
6:
7: \title{Complete Parameter Space of Quark Mass Matrices \\
8: with Four Texture Zeros}
9: \author{{\bf Zhi-zhong Xing} ~ and ~ {\bf He Zhang}}
10: \address{Institute of High Energy Physics,
11: Chinese Academy of Sciences, \\
12: P.O. Box 918 (4), Beijing 100039, China \\
13: ({\it Electronic address: xingzz@mail.ihep.ac.cn}) }
14: \maketitle
15:
16: \begin{abstract}
17: The full parameter space of Hermitian quark mass matrices
18: with four texture zeros is explored by using current experimental
19: data. We find that all ten free parameters of the four-zero
20: quark mass matrices can well be constrained. In particular,
21: only one of the two phase parameters plays an important role
22: in CP violation. The structural features of this specific pattern
23: of quark mass matrices are also discussed in detail.
24: \end{abstract}
25:
26: \pacs{PACS number(s): 12.15.Ff, 12.10.Kt}
27:
28: \framebox{\Large\bf 1} ~
29: The texture of quark mass matrices, which can significantly impact on
30: the pattern of quark flavor mixing, is completely unknown in the standard
31: electroweak model. A theory more fundamental than the standard model is
32: expected to allow us to determine the concrete structure of quark mass
33: matrices, from which six quark masses, three flavor mixing angles and
34: one CP-violating phase can fully be calculated. Attempts in this
35: direction (e.g., those starting from supersymmetric grand unification
36: theories and from superstring theories) are encouraging but have not
37: proved to be very successful. Phenomenologically, a very common approach
38: is to devise simple textures of quark mass matrices that can predict some
39: self-consistent and experimentally-favored relations between quark masses
40: and flavor mixing parameters \cite{Review}. The flavor symmetries hidden
41: in such textures might finally provide us with useful hints about the
42: underlying dynamics responsible for the generation of quark masses and
43: the origin of CP violation.
44:
45: Without loss of generality, the quark mass matrices $M_{\rm u}$
46: (up-type) and $M_{\rm d}$ (down-type) can always be taken to be
47: Hermitian in the standard model or its extensions which have no
48: flavor-changing right-handed currents \cite{F79}. Physics is invariant
49: under a common unitary transformation of Hermitian $M_{\rm u}$ and
50: $M_{\rm d}$ (i.e., $M_{\rm u,d} \rightarrow S M_{\rm u,d} S^\dagger$
51: with $S$ being an arbitrary unitary matrix). This freedom allows a
52: further arrangement of the structures of quark mass matrices, such that
53: \begin{equation}
54: M_{\rm u} \; = \; \left ( \matrix{
55: {\bf 0} & C_{\rm u} & {\bf 0} \cr
56: C^*_{\rm u} & \tilde{B}_{\rm u} & B_{\rm u} \cr
57: {\bf 0} & B^*_{\rm u} & A_{\rm u} \cr} \right ) \;
58: % (1)
59: \end{equation}
60: and
61: \begin{equation}
62: M_{\rm d} \; = \; \left ( \matrix{
63: D_{\rm d} & C_{\rm d} & {\bf 0} \cr
64: C^*_{\rm d} & \tilde{B}_{\rm d} & B_{\rm d} \cr
65: {\bf 0} & B^*_{\rm d} & A_{\rm d} \cr} \right ) \;
66: % (2)
67: \end{equation}
68: or
69: \begin{equation}
70: M'_{\rm d} \; = \; \left ( \matrix{
71: {\bf 0} & C_{\rm d} & D'_{\rm d} \cr
72: C^*_{\rm d} & \tilde{B}_{\rm d} & B_{\rm d} \cr
73: {D'}^*_{\rm d} & B^*_{\rm d} & A_{\rm d} \cr} \right ) \;
74: % (3)
75: \end{equation}
76: hold \cite{FX99}. We see that $M_{\rm u}$ has two texture zeros and
77: $M_{\rm d}$ or $M'_{\rm d}$ has one texture zero \cite{Note}.
78: Because the texture zeros of quark mass matrices in Eqs. (1)--(3)
79: result from some proper transformations of the flavor basis
80: under which the gauge currents keep diagonal and real, there
81: is no loss of any physical content for quark masses and flavor
82: mixing. But it is impossible to further obtain
83: $D_{\rm d} = 0$ for $M_{\rm d}$ or $D'_{\rm d} =0$ for
84: $M'_{\rm d}$ via a new physics-irrelevant transformation of
85: the flavor basis \cite{FX99}.
86: In other words, $D_{\rm d} = D'_{\rm d} = 0$ can only be a
87: physical assumption. This assumption leads to the well-known
88: four-zero texture of Hermitian quark mass matrices, which has the
89: up-down parallelism and respects the chiral evolution of quark
90: masses \cite{F87}.
91:
92: Although a lot of interest has been paid to the four-zero texture
93: of Hermitian quark mass matrices \cite{4zero,FX03},
94: a complete analysis of its parameter space has been lacking.
95: This unsatisfactory situation is partly due to the fact that many
96: authors prefer to make instructive analytical approximations in
97: analyzing the consequences of $M_{\rm u}$
98: and $M_{\rm d}$ on flavor mixing and CP violation. Some non-trivial
99: parts of the whole parameter space of $M_{\rm u,d}$ were
100: unfortunately missed or ignored in such analytical approximations
101: for a long time, as pointed out in Ref. \cite{FX03}.
102:
103: The purpose of this short note is to make use of current experimental
104: data to explore the complete parameter space of the four-zero
105: quark mass matrices $M_{\rm u}$ and $M_{\rm d}$. We find that
106: the ten free parameters of $M_{\rm u}$ and $M_{\rm d}$ can well be
107: constrained. In particular, only one of the two phase parameters
108: plays an important role in CP violation. We shall also discuss the
109: structural features of $M_{\rm u}$ and $M_{\rm d}$ in detail.
110:
111: \framebox{\Large\bf 2} ~
112: Let us concentrate on the four-zero texture of Hermitian quark mass
113: matrices given in Eqs. (1) and (2) with $D_{\rm d}=0$.
114: The observed hierarchy of quark masses
115: ($m_u \ll m_c \ll m_t$ and $m_d \ll m_s \ll m_b$) implies that
116: $|A_{\rm q}| > |\tilde{B}_{\rm q}|, |B_{\rm q}| > |C_{\rm q}|$
117: (for q = u or d) should in general hold \cite{F87}. Note that
118: $M_{\rm q}$ can be decomposed into
119: $M_{\rm q} = P^\dagger_{\rm q} \overline{M}_{\rm q} P_{\rm q}$, where
120: \begin{equation}
121: \overline{M}_{\rm q} \; = \; \left ( \matrix{
122: {\bf 0} & |C_{\rm q}| & {\bf 0} \cr
123: |C_{\rm q}| & \tilde{B}_{\rm q} & |B_{\rm q}| \cr
124: {\bf 0} & |B_{\rm q}| & A_{\rm q} \cr} \right ) \;
125: % (4)
126: \end{equation}
127: and $P_{\rm q} = {\rm Diag} \{1, \exp(i\phi^{~}_{C_{\rm q}}),
128: \exp(i\phi^{~}_{B_{\rm q}} + i\phi^{~}_{C_{\rm q}}) \}$ with
129: $\phi^{~}_{B_{\rm q}} \equiv \arg (B_{\rm q})$ and
130: $\phi^{~}_{C_{\rm q}} \equiv \arg (C_{\rm q})$.
131: For simplicity, we neglect the subscript ``q'' in the following,
132: whenever it is unnecessary to distinguish between the up and down quark
133: sectors. The real symmetric mass matrix $\overline{M}$ can be diagonalized
134: by use of the orthogonal transformation
135: \begin{equation}
136: O^T \overline{M} O \; = \; \left ( \matrix{
137: \lambda_1 & 0 & 0 \cr
138: 0 & \lambda_2 & 0 \cr
139: 0 & 0 & \lambda_3 \cr} \right ) \; ,
140: % (5)
141: \end{equation}
142: where $\lambda_i$ (for $i=1,2,3$) are quark mass eigenvalues.
143: Without loss of generality, we take $\lambda_3 >0$ and $A >0$.
144: Then ${\rm Det} (\overline{M}) = -A|C|^2 < 0$ implies that
145: $\lambda_1 \lambda_2 < 0$ is required. It is easy to find that
146: $\tilde{B}$, $|B|$ and $|C|$ can be expressed in terms of
147: $\lambda_i$ and $A$ as
148: \begin{eqnarray}
149: \tilde{B} & = & \lambda_1 + \lambda_2 + \lambda_3 - A \; ,
150: \nonumber \\
151: |B| & = & \sqrt{\frac{(A -\lambda_1) (A -\lambda_2)
152: (\lambda_3 -A)}{A}} \;\; ,
153: \nonumber \\
154: |C| & = & \sqrt{\frac{-\lambda_1 \lambda_2 \lambda_3}{A}} \;\; .
155: % (6)
156: \end{eqnarray}
157: The exact expression of $O$ turns out to be \cite{FX03}
158: \small
159: \begin{equation}
160: O \; =\; \left ( \matrix{ \cr
161: \displaystyle
162: \sqrt{\frac{\lambda_2 \lambda_3 (A-\lambda_1)}{A (\lambda_2 - \lambda_1)
163: (\lambda_3 - \lambda_1)}}
164: & \displaystyle
165: \eta \sqrt{\frac{\lambda_1 \lambda_3 (\lambda_2 -A)}
166: {A (\lambda_2 - \lambda_1) (\lambda_3 - \lambda_2)}}
167: & \displaystyle
168: \sqrt{\frac{\lambda_1 \lambda_2 (A -\lambda_3)}
169: {A (\lambda_3 - \lambda_1) (\lambda_3 - \lambda_2)}} \cr\cr\cr
170: \displaystyle
171: - \eta \sqrt{\frac{\lambda_1 (\lambda_1 -A)}{(\lambda_2 - \lambda_1)
172: (\lambda_3 - \lambda_1)}}
173: & \displaystyle
174: \sqrt{\frac{\lambda_2 (A-\lambda_2)}
175: {(\lambda_2 - \lambda_1) (\lambda_3 - \lambda_2)}}
176: & \displaystyle
177: \sqrt{\frac{\lambda_3 (\lambda_3 -A)}
178: {(\lambda_3 - \lambda_1) (\lambda_3 - \lambda_2)}} \cr\cr\cr
179: \displaystyle
180: \eta \sqrt{\frac{\lambda_1 (A -\lambda_2) (A -\lambda_3)}
181: {A (\lambda_2 - \lambda_1) (\lambda_3 - \lambda_1)}}
182: & \displaystyle
183: - \sqrt{\frac{\lambda_2 (A-\lambda_1) (\lambda_3 -A)}
184: {A (\lambda_2 - \lambda_1) (\lambda_3 - \lambda_2)}}
185: & \displaystyle
186: \sqrt{\frac{\lambda_3 (A -\lambda_1) (A -\lambda_2)}
187: {A (\lambda_3 - \lambda_1) (\lambda_3 - \lambda_2)}} \cr\cr}
188: \right ) \; ,
189: % (7)
190: \end{equation}
191: \normalsize
192: in which $\eta \equiv \lambda_2/m_2 = +1$ or $-1$ corresponding to
193: the possibility $(\lambda_1, \lambda_2) = (-m_1, +m_2)$ or
194: $(+m_1, -m_2)$. The Cabibbo-Kobayashi-Maskawa (CKM) flavor mixing
195: matrix \cite{CKM}, which measures the non-trivial mismatch between
196: diagonalizations of $M_{\rm u}$ and $M_{\rm d}$, is given by
197: $V \equiv O^T_{\rm u} (P_{\rm u} P^\dagger_{\rm d}) O_{\rm d}$.
198: Explicitly, we have
199: \begin{equation}
200: V_{i\alpha} \; =\; O^{\rm u}_{1i} O^{\rm d}_{1\alpha} +
201: O^{\rm u}_{2i} O^{\rm d}_{2\alpha} e^{i\phi_1} +
202: O^{\rm u}_{3i} O^{\rm d}_{3\alpha} e^{i(\phi_1 + \phi_2)} \; ,
203: % (8)
204: \end{equation}
205: where the subscripts $i$ and $\alpha$ run respectively over
206: $(u,c,t)$ and $(d,s,b)$, and two phases are defined as
207: $\phi_1 \equiv \phi^{~}_{C_{\rm u}} - \phi^{~}_{C_{\rm d}}$ and
208: $\phi_2 \equiv \phi^{~}_{B_{\rm u}} - \phi^{~}_{B_{\rm d}}$.
209:
210: It is well known that nine elements of the CKM matrix $V$ have
211: six orthogonal relations, corresponding to six triangles in the
212: complex plane \cite{Review}. Among them, the unitarity triangle
213: defined by $V^*_{ub}V_{ud} + V^*_{cb}V_{cd} + V^*_{tb}V_{td} =0$
214: is of particular interest for the study of CP violation at
215: $B$-meson factories \cite{PDG}. Three inner angles of this triangle
216: are commonly denoted as
217: \begin{eqnarray}
218: \alpha & = & \arg \left ( - \frac{V^*_{tb}V_{td}}{V^*_{ub}V_{ud}}
219: \right ) \;\; , \nonumber \\
220: \beta & = & \arg \left ( -\frac{V^*_{cb}V_{cd}}{V^*_{tb}V_{td}}
221: \right ) \;\; , \nonumber \\
222: \gamma & = & \arg \left ( -\frac{V^*_{ub}V_{ud}}{V^*_{cb}V_{cd}}
223: \right ) \;\; .
224: % (9)
225: \end{eqnarray}
226: So far the angle $\beta$ has unambiguously been measured from
227: the CP-violating asymmetry in $B^0_d$ vs
228: $\bar{B}^0_d\rightarrow J/\psi K_{\rm S}$ decays \cite{Browder}.
229: The angles $\alpha$ and $\gamma$ are expected to be detected at
230: KEK and SLAC $B$-meson factories in the near future. Given the
231: four-zero texture of quark mass matrices, these three angles depend
232: on the CP-violating phases $\phi_1$ and $\phi_2$. We shall examine
233: the explicit dependence of $(\alpha, \beta, \gamma)$ on
234: $(\phi_1, \phi_2)$ in the following.
235:
236: \framebox{\Large\bf 3} ~
237: Now we explore the whole parameter space of $M_{\rm u}$ and
238: $M_{\rm d}$ with the help of current experimental data. There
239: are totally ten free parameters associated with $M_{\rm u,d}$:
240: $A_{\rm u,d}$, $|B_{\rm u,d}|$, $\tilde{B}_{\rm u,d}$,
241: $|C_{\rm u,d}|$ and $\phi_{1,2}$. In comparison, there are
242: also ten observables which can be derived from the four-zero
243: texture of quark mass matrices: six quark masses
244: ($m_u, m_c, m_t$ and $m_d, m_s, m_b$) and four independent
245: parameters of quark flavor mixing (typically, $|V_{us}|$,
246: $|V_{cb}|$, $|V_{ub}/V_{cb}|$ and $\sin 2\beta$). Thus there is
247: no problem to determine the complete parameter space of $M_{\rm u,d}$.
248:
249: (1) The first step of our numerical calculations is to find out the
250: allowed ranges of $A_{\rm u}/m_t$, $A_{\rm d}/m_b$, $\phi_1$ and
251: $\phi_2$ by using Eqs. (6)-(9). For this purpose, we adopt the
252: following reasonable and generous values of quark mass ratios at the
253: electroweak scale $\mu = M_Z$ \cite{Review,PDG}:
254: \begin{eqnarray}
255: \frac{m_c}{m_u} & = & 270 - 350 \; , ~~~~~
256: \frac{m_s}{m_d} \; = \; 17 - 25 \; ;
257: \nonumber \\
258: \frac{m_t}{m_c} & = & 260 - 320 \; , ~~~~~
259: \frac{m_b}{m_s} \; = \; 35 - 45 \; .
260: % (10)
261: \end{eqnarray}
262: The predictions of $M_{\rm u}$ and $M_{\rm d}$ for the CKM
263: matrix elements are required to agree with current experimental
264: data \cite{Browder,Buras}:
265: \begin{eqnarray}
266: |V_{us}| & = & 0.2240 \pm 0.0036 \; , ~~~~~~~~~
267: \left | \frac{V_{ub}}{V_{cb}} \right | \; =\; 0.086 \pm 0.008 \; ,
268: \nonumber \\
269: |V_{cb}| & = & (41.5 \pm 0.8) \times 10^{-3} \; , ~~~~~
270: \sin 2\beta \; =\; 0.736 \pm 0.049 \; .
271: % (11)
272: \end{eqnarray}
273: Note that the results of $V$ may involve a four-fold ambiguity
274: arising from four possible values of $(\eta_{\rm u}, \eta_{\rm d})$
275: in $O_{\rm u}$ and $O_{\rm d}$, as one can see from Eq. (7). To be
276: specific, we first choose $\eta_{\rm u} = \eta_{\rm d} =+1$ in
277: our numerical analysis and then discuss the other three possibilities.
278:
279: The numerical results for $A_{\rm u}/m_t$ vs $A_{\rm d}/m_b$
280: and $\phi_1$ vs $\phi_2$ are illustrated in Fig. 1. We see that the
281: most favorable values of these four quantities are
282: $A_{\rm u}/m_t \sim 0.94$, $A_{\rm d}/m_b \sim 0.94$,
283: $\phi_1 \sim 0.5\pi$ and $\phi_2 \sim 1.96\pi$. Fig. 1(a) confirms
284: that the (3,3) elements (i.e., $A_{\rm u} \sim m_t$ and
285: $A_{\rm d} \sim m_b$) are the dominant matrix elements in $M_{\rm u}$
286: and $M_{\rm d}$. The strong constraint on $\phi_1$ comes from
287: the experimental data on $|V_{us}|$ and $\sin 2\beta$; while the
288: tight restriction on $\phi_2$ results from current data on
289: $|V_{cb}|$. Because of $\sin \phi_1 \gg |\sin \phi_2|$ as shown in
290: Fig. 1(b), the strength of CP violation in the CKM matrix is mainly
291: governed by $\phi_1$. In many analytical approximations,
292: $\phi_1 \approx 0.5\pi$ and $\phi_2 =0$ have typically been
293: taken \cite{4zero,FX03}.
294:
295: We find that both $A_{\rm u}/m_t$ and $A_{\rm d}/m_b$ are
296: insensitive to the signs of $\eta_{\rm u}$ and $\eta_{\rm d}$.
297: In other words, the allowed ranges of $A_{\rm u}/m_t$ and
298: $A_{\rm d}/m_b$ are essentially the same in
299: $(\eta_{\rm u}, \eta_{\rm d}) = (\pm 1, \pm1)$ and
300: $(\pm 1, \mp1)$ cases. While $\phi_1$ is sensitive
301: to the signs of $\eta_{\rm u}$ and $\eta_{\rm d}$,
302: $\phi_2$ is not. To be explicit, we have
303: \begin{eqnarray}
304: (\eta_{\rm u}, \eta_{\rm d}) & = & (+1, +1): ~~~~~
305: \phi_1 \sim 0.5\pi \; , ~~~ \phi_2 \lesssim 2\pi \; ,
306: \nonumber \\
307: (\eta_{\rm u}, \eta_{\rm d}) & = & (+1, -1): ~~~~~
308: \phi_1 \sim 1.5\pi \; , ~~~ \phi_2 \gtrsim 0 \; ,
309: \nonumber \\
310: (\eta_{\rm u}, \eta_{\rm d}) & = & (-1, +1): ~~~~~
311: \phi_1 \sim 1.5\pi \; , ~~~ \phi_2 \lesssim 2\pi \; ,
312: \nonumber \\
313: (\eta_{\rm u}, \eta_{\rm d}) & = & (-1, -1): ~~~~~
314: \phi_1 \sim 0.5\pi \; , ~~~ \phi_2 \gtrsim 0 \; .
315: % (12)
316: \end{eqnarray}
317: The dependence of $\phi_1$ on $\eta_{\rm u}$ and $\eta_{\rm d}$
318: can easily be understood. Indeed,
319: $\tan\beta \propto \eta_{\rm u}\eta_{\rm d} \sin\phi_1$ holds
320: in the leading-order analytical approximation with $|\sin\phi_2| \ll 1$.
321: Thus the positiveness of $\tan\beta$ requires that $\sin\phi_1$ and
322: $\eta_{\rm u}\eta_{\rm d}$ have the same sign.
323:
324: (2) The second step of our numerical analysis is to determine
325: the relative magnitudes of four non-zero matrix elements of
326: $M_{\rm u,d}$ by using Eq. (6) and the results for
327: $A_{\rm u}/m_t$ and $A_{\rm d}/m_b$. The numerical results for
328: $|B_{\rm u}|/A_{\rm u}$ vs $|B_{\rm d}|/A_{\rm d}$,
329: $\tilde{B}_{\rm u}/|B_{\rm u}|$ vs $\tilde{B}_{\rm d}/|B_{\rm d}|$
330: and $|C_{\rm u}|/\tilde{B}_{\rm u}$ vs $|C_{\rm d}|/\tilde{B}_{\rm d}$
331: are shown in Fig. 2. One can see that the most favorable values of
332: these six quantities are
333: $|B_{\rm u}|/A_{\rm u} \sim 0.25$,
334: $\tilde{B}_{\rm u}/|B_{\rm u}| \sim 0.3$,
335: $|C_{\rm u}|/\tilde{B}_{\rm u} \sim 0.003$ and
336: $|B_{\rm d}|/A_{\rm d} \sim 0.25$,
337: $\tilde{B}_{\rm d}/|B_{\rm d}| \sim 0.4$,
338: $|C_{\rm d}|/\tilde{B}_{\rm d} \sim 0.06$.
339: A remarkable feature of our typical results is that $A_{\rm q}$,
340: $|B_{\rm q}|$ and $\tilde{B}_{\rm q}$ (for q = u or d) roughly
341: satisfy a geometric relation:
342: $|B_{\rm q}|^2 \sim A_{\rm q} \tilde{B}_{\rm q}$ \cite{FX03}.
343: In addition, $|B_{\rm u}| \gg m_c$ and $|B_{\rm d}| \gg m_s$ hold.
344: While a very strong hierarchy exists between (1,2) and (2,2) elements
345: of $M_{\rm u,d}$, there is only a weak hierarchy among (2,2), (2,3)
346: and (3,3) elements of $M_{\rm u,d}$. Such a structural property of
347: quark mass matrices must be taken into account in model building.
348:
349: To be more explicit, let us illustrate the texture of
350: $\overline{M}_{\rm u,d}$ by choosing $m_c/m_u = 320$, $m_t/m_c = 290$,
351: $m_s/m_d =21$ and $m_b/m_s = 40$. We obtain
352: \begin{eqnarray}
353: \overline{M}_{\rm u} & \approx & A_{\rm u} \left ( \matrix{
354: {\bf 0} & 0.0002 & {\bf 0} \cr
355: 0.0002 & 0.067 & 0.24 \cr
356: {\bf 0} & 0.24 & {\bf 1} \cr} \right ) \sim \;
357: A_{\rm u} \left ( \matrix{
358: {\bf 0} & \varepsilon^6 & {\bf 0} \cr
359: \varepsilon^6 & \varepsilon^2 & \varepsilon \cr
360: {\bf 0} & \varepsilon & {\bf 1} \cr} \right ) \; ,
361: \nonumber \\
362: \overline{M}_{\rm d} & \approx & A_{\rm d} \left ( \matrix{
363: {\bf 0} & 0.0059 & {\bf 0} \cr
364: 0.0059 & 0.089 & 0.24 \cr
365: {\bf 0} & 0.24 & {\bf 1} \cr} \right ) \sim \;
366: A_{\rm d} \left ( \matrix{
367: {\bf 0} & \varepsilon^4 & {\bf 0} \cr
368: \varepsilon^4 & \varepsilon^2 & \varepsilon \cr
369: {\bf 0} & \varepsilon & {\bf 1} \cr} \right ) \; ,
370: % (13)
371: \end{eqnarray}
372: where $\varepsilon \approx 0.24$ in this special case \cite{Note2}.
373: Such a four-zero pattern of quark mass matrices depends on a
374: small expansion parameter and is quite suggestive for model building.
375: For example, one may speculate that $\overline{M}_{\rm u}$ and
376: $\overline{M}_{\rm d}$ in Eq. (13)
377: could naturally result from a string-inspired model of
378: quark mass generation \cite{Ibanez} or from a horizontal U(1) family
379: symmetry and its perturbative breaking \cite{Flavor}.
380:
381: Note that the numerical results in Fig. 2 and Eq. (13) have been
382: obtained by taking $\eta_{\rm u} = \eta_{\rm d} =+1$. A careful
383: analysis shows that $\tilde{B}_{\rm q}$ (for q = u or d) is
384: sensitive to the sign of $\eta_{\rm q}$, but $|B_{\rm q}|$ and
385: $|C_{\rm q}|$ are not. In view of Eq. (6), we find that the sign
386: of $\eta = \lambda_2/m_2$ may significantly affect the size
387: of $\tilde{B}$ if its $\lambda_2$ and $\lambda_3 -A$ terms are
388: comparable in magnitude. In contrast, the dependence of $|B|$ on
389: $\eta$ is negligible due to $A \gg m_2$; and $|C|$ is completely
390: independent of the sign of $\eta$.
391:
392: (3) The final step of our numerical analysis is to examine the
393: outputs of three CP-violating angles $(\alpha, \beta, \gamma)$ and
394: the ratio $|V_{ub}/V_{cb}|$ constrained by the four-zero texture
395: of quark mass matrices. We plot the result for $|V_{ub}/V_{cb}|$
396: vs $\sin 2\beta$ in Fig. 3(a) and that for $\alpha$ vs $\gamma$ in
397: Fig. 3(b). The correlation between $\alpha$ and $\gamma$ is quite
398: obvious, as a result of $\alpha + \beta + \gamma = \pi$. Typically,
399: $\alpha \sim 0.5\pi$ holds. The possibility $\alpha \approx \gamma$,
400: implying that the unitarity triangle is approximately an
401: isoceless triangle \cite{FH}, is also allowed by current data
402: and quark mass matrices with four texture zeros. Note that the
403: size of $\sin 2\beta$ increases with $|V_{ub}/V_{cb}|$. This feature
404: can easily be understood: in the unitarity triangle with three
405: sides rescaled by $|V_{cb}|$, the inner angle $\beta$ corresponds
406: to the side proportional to $|V_{ub}/V_{cb}|$.
407:
408: Finally we mention that the outputs of $|V_{ub}/V_{cb}|$ and
409: $(\alpha, \beta, \gamma)$ are completely insensitive to the sign
410: ambiguity of $\eta_{\rm u}$ and $\eta_{\rm d}$.
411:
412: \framebox{\Large\bf 4} ~
413: In summary, we have analyzed the complete parameter space of
414: Hermitian quark mass matrices with four texure zeros by using
415: current experimental data. It is clear that the four-zero
416: pattern of quark mass matrices can survive current experimental
417: tests and its parameter space gets well constrained. We find
418: that only one of the two phase parameters plays a crucial role
419: in CP violation. The (2,2), (2,3) and (3,3) elements of the
420: up- or down-type quark mass matrix have a relatively weak
421: hierarchy, although their magnitudes are considerably larger
422: than the magnitude of the (1,2) element. Such a structural
423: feature of the four-zero quark mass matrices might serve as
424: a useful starting point of view for model building.
425:
426: We remark that the phenomenological consequences of quark mass
427: matrices depend both on the number of their texture zeros and
428: on the hierarchy of their non-vanishing entries. The former
429: are in general not preserved to all orders or at any energy scales
430: in the unspecified interactions which generate quark masses
431: and flavor mixing \cite{X03}. But an experimentally-favored
432: texture of quark mass matrices at low energy scales (such as the
433: one under discussion) is possible to shed some light on the
434: underlying flavor symmetry and its breaking mechanism responsible
435: for fermion mass generation and CP violation at high energy scales.
436:
437: \vspace{0.4cm}
438:
439: This work was supported in part by National Natural Science
440: Foundation of China.
441:
442: \vspace{0.4cm}
443:
444: \hspace{-0.8cm} {\bf Note added}: While our paper was being
445: completed, we received a preprint by Zhou \cite{Zhou}, in which
446: all possible four-zero textures of quark mass matrices are
447: classified and computed. The analyses, results and discussions
448: in these two papers have little overlap.
449:
450: \begin{thebibliography}{99}
451:
452: \bibitem{Review} For a recent review with extensive references, see:
453: H. Fritzsch and Z.Z. Xing, Prog. Part. Nucl. Phys. {\bf 45}, 1 (2000).
454:
455: \bibitem{F79} H. Fritzsch,
456: Nucl. Phys. B {\bf 155}, 189 (1979).
457:
458: \bibitem{FX99} H. Fritzsch and Z.Z. Xing,
459: Nucl. Phys. B {\bf 556}, 49 (1999);
460: G.C. Branco, D. Emmanuel-Costa, and R. Gonz$\rm\acute{a}$lez Felipe,
461: Phys. Lett. B {\bf 477}, 147 (2000).
462:
463: \bibitem{Note} Because $M_{\rm u}$ and $M_{\rm d}$ (or $M'_{\rm d}$)
464: are Hermitian, a pair of off-diagonal texture zeros in each
465: mass matrix have been counted as one zero.
466:
467: \bibitem{F87} H. Fritzsch, Phys. Lett. B {\bf 184}, 391 (1987);
468: H. Fritzsch and Z.Z. Xing, Phys. Lett. B {\bf 413}, 396 (1997);
469: Phys. Rev. D {\bf 57}, 594 (1998).
470:
471: \bibitem{4zero} See, e.g.,
472: D. Du and Z.Z. Xing, Phys. Rev. D {\bf 48}, 2349 (1993);
473: L.J. Hall and A. Rasin, Phys. Lett. B {\bf 315}, 164 (1993);
474: H. Fritzsch and D. Holtmansp$\rm\ddot{o}$tter,
475: Phys. Lett. B {\bf 338}, 290 (1994);
476: H. Fritzsch and Z.Z. Xing, Phys. Lett. B {\bf 353}, 114 (1995);
477: P.S. Gill and M. Gupta, J. Phys. G {\bf 21}, 1 (1995);
478: Phys. Rev. D {\bf 56}, 3143 (1997);
479: H. Lehmann, C. Newton, and T.T. Wu,
480: Phys. Lett. B {\bf 384}, 249 (1996);
481: Z.Z. Xing, J. Phys. G {\bf 23}, 1563 (1997);
482: K. Kang and S.K. Kang, Phys. Rev. D {\bf 56}, 1511 (1997);
483: T. Kobayashi and Z.Z. Xing, Mod. Phys. Lett. A {\bf 12}, 561 (1997);
484: Int. J. Mod. Phys. A {\bf 13}, 2201 (1998);
485: J.L. Chkareuli and C.D. Froggatt, Phys. Lett. B {\bf 450}, 158 (1999);
486: Nucl. Phys. B {\bf 626}, 307 (2002);
487: A. Mondrag$\rm\acute{o}$n and E. Rodriguez-J$\rm\acute{a}$uregui,
488: Phys. Rev. D {\bf 59}, 093009 (1999);
489: H. Nishiura, K. Matsuda, and T. Fukuyama,
490: Phys. Rev. D {\bf 60}, 013006 (1999);
491: G.C. Branco, D. Emmanuel-Costa, and R. Gonz$\rm\acute{a}$lez Felipe,
492: in Ref. \cite{FX99};
493: S.H. Chiu, T.K. Kuo, and G.H. Wu,
494: Phys. Rev. D {\bf 62}, 053014 (2000);
495: H. Fritzsch and Z.Z. Xing, Phys. Rev. D {\bf 61}, 073016 (2000);
496: Phys. Lett. B {\bf 506}, 109 (2001);
497: R. Rosenfeld and J.L. Rosner, Phys. Lett. B {\bf 516}, 408 (2001);
498: R.G. Roberts, A. Romanino, G.G. Ross, and L. Velasco-Sevilla,
499: Nucl. Phys. B {\bf 615}, 358 (2001).
500:
501: \bibitem{FX03} H. Fritzsch and Z.Z. Xing,
502: Phys. Lett. B {\bf 555}, 63 (2003).
503:
504: \bibitem{CKM} N. Cabibbo, Phys. Rev. Lett. {\bf 10}, 531 (1963);
505: M. Kobayashi and T. Maskawa, Prog. Theor. Phys. {\bf 49}, 652 (1973).
506:
507: \bibitem{PDG} Particle Data Group, K. Hagiwara {\it et al.},
508: Phys. Rev. D {\bf 66}, 010001 (2002).
509:
510: \bibitem{Browder} T. Browder, talk given at the 21st International
511: Symposium on Lepton and Photon Interactions at High Energies
512: (LP 03), August 2003, Batavia, Illinois, USA.
513:
514: \bibitem{Buras} A.J. Buras, hep-ph/0307203; and references therein.
515:
516: \bibitem{Note2} A different expansion of $\overline{M}_{\rm u,d}$
517: has been given in Ref. \cite{FX03} in terms of two different small
518: parameters.
519:
520: \bibitem{Ibanez} D. Cremades, L.E. Ib$\rm\acute{a}\tilde{n}$ez,
521: and F. Marchesano, hep-ph/0212064; and references therein.
522:
523: \bibitem{Flavor} C.D. Froggatt and H.B. Nielsen,
524: Nucl. Phys. B {\bf 147}, 277 (1979);
525: L.E. Ib$\rm\acute{a}\tilde{n}$ez and G.G. Ross,
526: Phys. Lett. B {\bf 332}, 100 (1994).
527:
528: \bibitem{FH} I am grateful to H. Fritzsch and C. Hamzaoui for
529: pointing out this possibility to me.
530:
531: \bibitem{X03} Z.Z. Xing, hep-ph/0307359; and references therein.
532:
533: \bibitem{Zhou} Y.F. Zhou, hep-ph/0309076.
534: \end{thebibliography}
535:
536:
537: %%%%%%%%%%%%%%%%%%%% Fig. 1 %%%%%%%%%%%%%%%%
538: \begin{figure}[t]
539: \vspace{-1cm}
540: \epsfig{file=mfig1.ps,bbllx=2.5cm,bblly=12cm,bburx=17.5cm,bbury=30cm,%
541: width=13cm,height=15cm,angle=0,clip=90}
542: \vspace{4.4cm}
543: \caption{The allowed ranges of $A_{\rm u}/m_t$ vs
544: $A_{\rm d}/m_b$ and $\phi_1$ vs $\phi_2$ for the four-zero
545: texture of quark mass matrices with
546: $\eta_{\rm u} = \eta_{\rm d} = +1$.}
547: \end{figure}
548: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
549:
550: %%%%%%%%%%%%%%%%%%%% Fig. 2 %%%%%%%%%%%%%%%%
551: \begin{figure}[t]
552: \vspace{-5.5cm}
553: \epsfig{file=mfig2.ps,bbllx=1.5cm,bblly=10cm,bburx=18.5cm,bbury=30cm,%
554: width=13cm,height=15cm,angle=0,clip=90}
555: \vspace{6.4cm}
556: \caption{The allowed ranges of $|B_{\rm u}|/A_{\rm u}$ vs
557: $|B_{\rm d}|/A_{\rm d}$, $\tilde{B}_{\rm u}/|B_{\rm u}|$ vs
558: $\tilde{B}_{\rm d}/|B_{\rm d}|$ and $|C_{\rm u}|/\tilde{B}_{\rm u}$ vs
559: $|C_{\rm d}|/\tilde{B}_{\rm d}$ for the four-zero texture of quark
560: mass matrices with $\eta_{\rm u} = \eta_{\rm d} = +1$.}
561: \end{figure}
562: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
563:
564: %%%%%%%%%%%%%%%%%%%% Fig. 3 %%%%%%%%%%%%%%%%
565: \begin{figure}[t]
566: \vspace{-3cm}
567: \epsfig{file=mfig3.ps,bbllx=2.5cm,bblly=12cm,bburx=17.5cm,bbury=30cm,%
568: width=13cm,height=15cm,angle=0,clip=90}
569: \vspace{4.4cm}
570: \caption{The allowed ranges of $|V_{ub}/V_{cb}|$ vs
571: $\sin 2\beta$ and $\alpha$ vs $\gamma$ for the four-zero
572: texture of quark mass matrices with
573: $\eta_{\rm u} = \eta_{\rm d} = +1$.}
574: \end{figure}
575: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
576: \end{document}
577:
578: