hep-ph0509218/u1D.tex
1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: %               Non-minimal lepton flavour violating processes in             %
3: %                 realistic string-inspired supergravity models               %
4: %                     J.G.Hayes, S.F.King, and I.N.R.Peddie                   %
5: %                                                                             %
6: %          School of Physics and Astronomy, University of Southampton         %
7: %                          Southampton, SO17 1BJ, U.K.                        %
8: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
9: 
10: 
11: %
12: %
13: 
14: \documentclass[12pt]{article}
15: 
16: \usepackage{amsmath}
17: \usepackage{amssymb}
18: \usepackage{delarray}
19: \usepackage{graphics}
20: \usepackage{color}
21: \usepackage{axodraw}
22: \input epsf
23: %
24: 
25: % These settings are as they were for Iain's thesis.
26: %\textwidth 6.0in
27: %\textheight 8.4in
28: %\topmargin -0.4in
29: %\oddsidemargin 0.27in
30: %\evensidemargin 0.27in
31: %\parskip 0.08in
32: %\def\baselinestretch{1.2}
33: %\hfuzz=9mm
34: %\baselineskip 20pt
35: 
36: % These settings are as they were for Iain's paper. Note that the left
37: % hand margin is wider for them to write in the hep-ph number.
38: 
39: \textwidth 6.0in
40: \textheight 9.0in
41: \topmargin 0.0in
42: \oddsidemargin 0.5in
43: \evensidemargin 0.5in
44: \parskip 0.08in
45: \hfuzz=9mm
46: \baselineskip 24pt
47: 
48: %%% ----------------------------------------------------------------------
49: 
50: \providecommand{\SU}[1]{}
51: \renewcommand{\SU}[1]{\ensuremath{\mathrm{SU}(#1)}}
52: \providecommand{\UI}{}
53: \renewcommand{\UI}{\ensuremath{\mathrm{U}(1)}}
54: \providecommand{\psgroup}{}
55: \renewcommand{\psgroup}{\ensuremath{\mathrm{SU}(4) \otimes \mathrm{SU}(2)_L \otimes \mathrm{SU}(2)_R } }
56: \providecommand{\smgroup}{}
57: \renewcommand{\smgroup}{\ensuremath{\mathrm{SU}(3) \otimes \mathrm{SU}(2)_L \otimes \mathrm{U}(1)_Y } }
58: \providecommand{\Dpbrane}{}
59: \renewcommand{\Dpbrane}{\ensuremath{Dp\mathrm{-brane}}}
60: \providecommand{\Dpbranes}{}
61: \renewcommand{\Dpbranes}{\ensuremath{Dp\mathrm{-branes}}}
62: \providecommand{\Dbrane}[1]{}
63: \renewcommand{\Dbrane}[1]{\ensuremath{D#1\mathrm{-brane}}}
64: \providecommand{\Dbranes}[1]{}
65: \renewcommand{\Dbranes}[1]{\ensuremath{D#1\mathrm{-branes}}}
66: \providecommand{\vev}{}
67: \renewcommand{\vev}{vev}
68: \providecommand{\vevs}{}
69: \renewcommand{\vevs}{vevs}
70: \providecommand{\Tr}{}
71: \renewcommand{\Tr}{\ensuremath{\mathrm{Tr}}}
72: \providecommand{\identity}{}
73: \renewcommand{\identity}{\ensuremath{\mathrm{I}}}
74: 
75: %%% ----------------------------------------------------------------------
76: 
77: \newcommand{\sheptitle}
78: {\Large Lepton flavour violation in realistic non-minimal supergravity models}
79: 
80: \newcommand{\shepauthor}
81: {J.~G.~Hayes, S.~F.~King, I.~N.~R.~Peddie \footnote{Address after 
82: 1st September, 2005: Physics Division, School of Technology,
83: Aristotle University of Thessaloniki, Thessaloniki 54124, Greece.}}
84: 
85: 
86: \newcommand{\shepaddress}
87: {School of Physics and Astronomy, University of Southampton, \\
88:         Southampton, SO17 1BJ, U.K.}
89: 
90: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
91: %                                                                             %
92: %                                  Abstract                                   %
93: %                                                                             %
94: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
95: 
96: \newcommand{\shepabstract} {Realistic effective supergravity models
97: have a variety of sources of lepton flavour violation (LFV) which can
98: drastically affect the predictions relative to the scenarios
99: usually considered in the literature based on minimal supergravity and
100: the supersymmetric see-saw mechanism.
101: We catalogue the additional sources of LFV which occur in realistic
102: supergravity models including the effect of D-terms arising from an Abelian
103: $U(1)$ family symmetry, non-aligned trilinear contributions from
104: scalar F-terms, as well as non-minimal supergravity contributions
105: and the effect of different Yukawa textures.  
106: In order to quantify these effects, we investigate a string inspired
107: effective supergravity model arising from intersecting D-branes 
108: supplemented by an additional $U(1)$ family symmetry. In such 
109: theories the magnitude of the D-terms is predicted, and we 
110: calculate the branching ratios for 
111: $\mu\rightarrow e\gamma$ and $\tau \rightarrow \mu \gamma$ 
112: for different benchmark points designed to isolate the 
113: different non-minimal contributions. 
114: We find that the D-term
115: contributions are generally dangerously large, but in certain cases
116: such contributions can lead to a dramatic suppression of LFV rates,
117: for example by cancelling the effect of the see-saw induced
118: LFV in $\tau \to \mu \gamma$ models with lop-sided textures.
119: In the class of string models considered here we find the surprising result
120: that the D-terms can sometimes serve to restore 
121: universality in the effective non-minimal supergravity theory.}
122: 
123: %%% ----------------------------------------------------------------------
124: 
125: \begin{document}
126: 
127: %%% ----------------------------------------------------------------------
128: 
129: \begin{titlepage}
130: 
131: %%% ----------------------------------------------------------------------
132: 
133: \begin{center}
134: {\large{\bf \sheptitle}} \\
135: \vspace{0.5in}
136: \bigskip \shepauthor \\ \mbox{} \\ {\it \shepaddress} \\
137: \vspace{0.5in}
138: \bigskip {\bf Abstract} \bigskip
139: \end{center}
140: \setcounter{page}{0}
141: \shepabstract
142: \begin{flushleft}
143: \today
144: \end{flushleft}
145: 
146: \vskip 0.1in
147: \noindent
148: 
149: %%% ----------------------------------------------------------------------
150: 
151: \end{titlepage}
152: 
153: %%% ----------------------------------------------------------------------
154: 
155: %%% ----------------------------------------------------------------------
156: 
157: \newpage 
158: \section{Introduction} 
159: \label{sec:introduction}
160: 
161: % Example of labelling sections, with the \label{sec:...} to be able to
162: % refer back to that particular section easily.
163: 
164: %%% ----------------------------------------------------------------------
165: 
166: 
167: Lepton flavour violation (LFV) is a sensitive probe of new physics
168: in supersymmetric (SUSY) models
169: \cite{Borzumati:1986qx,Gabbiani:1996hi}.
170: In SUSY models, LFV arises due to off-diagonal elements in the
171: slepton mass matrices in the `super-CKM' (SCKM) basis, in which the
172: quark and lepton mass matrices are diagonal \cite{Chung:2003fi}.
173: In supergravity (SUGRA) mediated
174: SUSY breaking the soft SUSY breaking masses are generated at the
175: Planck scale, and the low energy soft masses relevant for 
176: physical proceses such as LFV are therefore
177: subject to radiative corrections in running from the Planck scale
178: to the weak scale. Off-diagonal slepton masses in the SCKM basis
179: can arise both directly at the high energy scale
180: (due to the effective SUGRA theory which is responsible for them), or can
181: be radiatively generated by renormalisation group running
182: from the Planck scale to the weak scale, 
183: for example due to Higgs triplets in GUTs or right-handed neutrinos
184: in see-saw models which have masses intermediate between these
185: two scales.
186: 
187: Neutrino experiments
188: confirming the Large Mixing Angle (LMA) MSW solution to the
189: solar neutrino problem \cite{GoPe} taken together with the
190: atmospheric neutrino data \cite{SKamiokandeColl} shows that neutrino
191: masses are inevitable \cite{King:2003jb}.
192: The presence of right-handed neutrinos, as required by the see-saw
193: mechanism for generating neutrino masses, will lead inevitably to LFV,
194: due to running effects, even in minimal SUGRA 
195: (mSUGRA) which has no LFV at the high energy
196: GUT or Planck scale \cite{Hisano:1995cp,King:1998nv}. 
197: Therefore, merely assuming SUSY and the see-saw
198: mechanism, one expects LFV to be present. This has been studied, for
199: example, in mSUGRA models with a natural neutrino mass hierarchy
200: \cite{Blazek:2002wq}. There is a large literature on 
201: the case of minimal LFV arising from
202: mSUGRA and the see-saw mechanism \cite{huge}.
203: 
204: % Even in a best case scenario, such as mSUGRA, where there are no
205: % direct LFV effects, right handed neutrinos, required by the see-saw
206: % mechanism for generating small neutrino masses, will lead inevitably
207: % to LFV
208: % \cite{Hisano:1995cp,King:1998nv}.
209: % The recent neutrino experiments, which confirm the matter enhanced
210: % Large Mixing Angle (LMA) solution to the solar neutrino problem
211: % \cite{GoPe}, together with the atmospheric data
212: % \cite{SKamiokandeColl}, show that neutrino masses are inevitable.
213: % Hence, assuming SUSY and the see-saw mechansim, one would expect that
214: % LFV must be present. 
215: % For example, this has been studied in mSUGRA 
216: % models with a natural neutrino mass hierarchy \cite{Blazek:2002wq}.
217: % There is in fact a large literature on this subject \cite{huge}. 
218: 
219: Despite the fact that most realistic string models lead
220: to a low energy effective {\em non-minimal} SUGRA theory,
221: such theories have not been extensively studied in the literature,
222: although a general analysis of flavour changing effects
223: in the mass insertion approximation has recently been performed
224: \cite{Chankowski:2005jh}. Such effective non-minimal SUGRA models 
225: predict non-universality of the soft masses at the high energy scale,
226: dependent on the structure of the Yukawa matrices. Moreover there can be
227: additional sources of LFV which also enter the analysis.
228: For example, realistic effective SUGRA theories arising from
229: string-inspired models will also typically involve
230: some gauged family symmetry which can give an additional
231: direct (as opposed to renormalisation group
232: induced) source of LFV. This is because the D-term
233: contribution to the scalar masses generated when the family symmetry
234: spontaneously breaks adds different diagonal elements 
235: to each generation in the `theory' basis, and generates
236: non-zero off-diagonal elements in the SCKM basis, leading to LFV. This effect
237: depends on the strength of the D-term contribution, which is expected
238: to be close in size to the size of the uncorrected scalar masses.
239: There can also be a significant contribution to
240: non-universal trilinear soft masses leading to flavour violation
241: \cite{Abel:2001ur,Ross:2002mr,KP0307091} arising from the 
242: F-terms of scalars associated with the Yukawa couplings
243: (for example the flavons of Froggatt-Nielsen theories). 
244: 
245: The purpose of the present
246: paper is to catalogue and quantitatively study the importance of
247: all the different sources of LFV present
248: in a general non-minimal SUGRA framework, including the 
249: effects of gauged family symmetry.
250: Although the different effects have all been identified in the literature,
251: there has not so far been a coherent and quantitative
252: dedicated study of LFV processes, beyond the mass insertion approximation,
253: which includes all these effects within a single framework.
254: In order to quantify the importance of the different effects it is necessary
255: to investigate these disparate souces of LFV numerically, both in isolation and
256: in association with one another, within some particular SUGRA 
257: model. To be concrete we shall study the effective 
258: SUGRA models of the kind considered
259: in \cite{KP0307091} which have a sufficiently rich structure
260: to enable all of the effects to be studied within a single framework.
261: Within this class of models we shall consider specific benchmark
262: points in order to illustrate the different effects.
263: Some of these benchmark points were already previously considered
264: in \cite{KP0307091}. However in the previous study the important effect
265: of D-terms arising from the Abelian family symmetry was not
266: considered. Here we shall show that such D-terms are in fact
267: calculable within the framework of the particular model
268: considered, and can lead to significant enhancement
269: (or suppression) of LFV rates, depending on the particular 
270: model considered. 
271: 
272: The outline of this paper is as follows. In
273: Section~\ref{sec:soft-from-SUGRA} we discuss soft
274: supersymmetry breaking masses in supergravity. In
275: Section~\ref{sec:sourc-lfv} we summarise the flavour problem and
276: catalogue the distinct sources of lepton flavour violation. In
277: Section~\ref{sec:442-pati-salam}, we outline the aspects of
278: the specific models that we shall study. In Section~\ref{soft} we
279: discuss the soft SUSY-breaking sector within this class of models, 
280: parameterise the F-terms, and write down the soft terms, including the D-term
281: contribution to the scalar masses. We also discuss the D-terms
282: associated with a family symmetry that are expected to lead to large
283: lepton flavour violation. In Section~\ref{sec:numerical-etc} we
284: specify two models in this class, define benchmark points
285: and present the results of our numerical analysis of LFV for these
286: benchmark points.
287: Section~\ref{sec:conclusions} concludes the paper.
288: 
289: %%% ----------------------------------------------------------------------
290: 
291: \section{Soft terms from supergravity}
292: \label{sec:soft-from-SUGRA}
293: 
294: %%% ----------------------------------------------------------------------
295: 
296: We summarise here the standard way of getting soft SUSY breaking terms
297: from supergravity.  Supergravity is defined in terms of a K\"ahler
298: function, $G$, of chiral superfields ($\phi = h, C_a$). Taking the
299: view that supergravity arises as the low energy effective field theory
300: limit of a string theory, the hidden sector fields $h$ are taken to
301: correspond to closed string moduli states ($h = S, T_i$), and the
302: matter states $C_a$ are taken to correspond to open string states. In
303: string theory, the ends of the open string states are believed to be
304: constrained to lie on extended solitonic objects called \Dpbranes.
305: 
306: Using natural units,
307: \begin{equation}
308:   \label{eq:Kahler_function}
309:   G(\phi, \overline{\phi}) =
310:   \frac{K(\phi,\overline{\phi})}{\tilde{M}^2_{Pl}}
311:   + \ln
312:   \left(
313:     \frac{W(\phi)}{\tilde{M}^3_{Pl}}
314:   \right)
315:   + \ln
316:   \left(
317:     \frac{W^*(\overline{\phi})}{\tilde{M}^3_{Pl}}
318:   \right)
319: \end{equation}
320: 
321: $K(\phi,\overline{\phi})$ is the K\"ahler potential, a real function of
322: chiral superfields. This may be expaned in powers of $C_a$:
323: \begin{equation}
324:   \label{eq:Kahler_potential}
325:   K = 
326:   \overline{K}(h,\overline{h})
327:   + \tilde{K}_{\overline{a}{b}}(h,\overline{h}) \overline{C}_{\overline{a}} C_b
328:   + 
329:   \left[
330:     \frac{1}{2}Z_{ab}(h,\overline{h})C_a C_b + h.c. \right]
331:   + ...
332: \end{equation}
333: 
334: $\tilde{K}_{\overline{a} b}$ is the K\"ahler metric. $W(\phi)$ is the
335: superpotential, a holomorphic function of chiral superfields.
336: \begin{equation}
337:   \label{eq:superpotential}
338:   W = \hat{W}(h) + \frac{1}{2}\mu_{ab}(h)C_a C_b +
339:   \frac{1}{6}Y_{abc}C_a C_b C_c + ...
340: \end{equation}
341: 
342: We expect the supersymmetry to be broken; if it is broken, then the
343: auxilliary fields $F_\phi \ne 0$ for some $\phi$. As we lack a model of
344: SUSY breaking, we introduce Goldstino angles as parameters that will
345: enable us to explore different methods of breaking supersymmetry. We
346: introduce a matrix, $P$ that canonically normalises the K\"ahler metric,
347: $P^\dagger K_{\overline{J}I} P = 1$
348: \footnote{ The subscripts on the K\"ahler potential $K_I$ means
349: $\partial_I K$. However, the subscripts on the F-terms are just
350: labels. } \cite{P_Matrix:introduction}.  We also introduce a column
351: vector $\Theta$ which satisfies $\Theta^\dagger\Theta = 1$. We are
352: completely free to parameterise $\Theta$ in any way which satisfies
353: this constraint.
354: 
355: Then the un-normalised soft terms and trilinears appear in the soft
356: SUGRA breaking potential \cite{Brignole:1997dp}
357: \begin{equation}
358:   \label{eq:soft_potential}
359:   V_{\mathrm{soft}} = m^2_{\overline{a}b} \overline{C}_{\overline{a}} C_b
360:     + \left( \frac{1}{6}A_{abc} Y_{abc} C_a C_b C_c + h.c. \right) + ...
361: \end{equation}
362: 
363: The non-canonically normalised soft trilinears are then
364: \begin{eqnarray}
365:   \nonumber
366:   A_{abc} Y_{abc}
367:   &=&
368:   \frac{\displaystyle \hat{W}^*}{|\displaystyle \hat{W}| }
369:   e^{\overline{K}/2} F_m
370:   \left[
371:     \overline{K}_m Y_{abc} + \partial_m Y_{abc} - 
372:     \left(
373:       \left(\tilde K^{-1}\right)
374:     \right. 
375:   \right. \partial_m
376:   \tilde K_{\overline{e}a} Y_{dbc}
377:   \\
378:   \label{eq:unnomalised_trilinears}  
379:   && 
380:   \left. 
381:     \left.
382:       %% this should make the closing brackets look the correct size
383:       %% ( and it is a bit of a hack )
384:       { \color{white} \tilde K^-1_a }
385:       {} + ( a \leftrightarrow b ) 
386:       + ( a \leftrightarrow c )
387:     \right)
388:   \right]
389: \end{eqnarray}
390: 
391: In this equation, it should be noted that the index $m$ runs over $h,
392: C$. However, by definition, the hidden sector part of the K\"ahler
393: potential and the K\"ahler metrics is independent of the matter
394: fields.
395: 
396: Assuming that the terms $\partial_C Y_{abc} \ne 0$,
397: the canonically normalised equation for the trilinear is
398: \begin{equation}
399:   \label{eq:normalised_trilinear}
400:   A_{abc} = F_I 
401:   \left[
402:     \overline{K}_I - \partial_I \ln
403:     \left(
404:       \tilde K_a \tilde K_b \tilde K_c
405:     \right)
406:   \right]
407:   + F_m \partial_m \ln Y_{abc}
408: \end{equation}
409: If the Yukawa hierarchy is taken to be generated by a Froggatt-Nielsen
410: (FN) field, $\phi$, such that $Y \propto \phi^p$, then we expect $F_\phi
411: \propto m_{3/2} \phi$, and then $F_\phi \partial_\phi \ln Y \propto
412: m_{3/2}$, and so even though these fields are expected to have heavily
413: sub-dominant F-terms, they contribute to the trilinears on an equal
414: footing as the moduli.
415: 
416: If the K\"ahler metric is diagonal and non-canonical, then the
417: canonically normalised scalar mass-squareds are given by
418: \begin{equation}
419:   \label{eq:normalised_scalars}
420:   m_a^2 = m^2_{3/2} - F_{\overline{J}} F_I \partial_{\overline{J}} \partial_I 
421:   \left( \ln \tilde{K_a} \right) ,
422: \end{equation}
423: 
424: and the gaugino masses are given by
425: \begin{equation}
426:   \label{eq:normalised_gauginos}
427:   M_\alpha = \frac{1}{2 \mathrm{Re} f_\alpha } F_I \partial_I f_\alpha ,
428: \end{equation}
429: where $f_\alpha$ is the `gauge kinetic function'. $\alpha$ enumerates
430: $D$-branes in the model.  In type~I string models without twisted
431: moduli these have the form $f_9 = S\; , \; f_{5_i} = T_i$.
432: 
433: Specifically, we use a K\"ahler potential that doesn't have any
434: twisted-moduli \cite{Ibanez:1998rf}:
435: \begin{eqnarray}
436:   \nonumber K &=& -\ln \left( S + \overline{S} -
437:   \left|C_1^{5_1}\right|^2 - \left|C_2^{5_2}\right|^2 \right) -\ln
438:   \left( T_1 + \overline{T}_1 - \left|C^9_1\right|^2 -
439:   \left|C_3^{5_3}\right|^2 \right) \\ \nonumber && {} - \ln \left( T_2
440:   - \overline{T}_2 - \left|C_2^9\right|^2 - \left|C_3^{5_1}\right|^2
441:   \right) -\ln \left( T_3 - \overline{T}_3 -\left|C_3^9\right|^2 -
442:   \left|C_2^{5_1}\right|^2 - \left|C_1^{5_1}\right|^2 \right) \\
443:   \nonumber && {} + \frac{\left|C^{5_1 5_2}\right|^2
444:   }{\left(S+\overline{S}\right)^{1/2}\left(T_3+\overline{T}_3\right)^{1/2}}
445:   + \frac{\left|C^{95_1}\right|^2}{\left(T_2 +
446:   \overline{T}_2\right)^{1/2}\left(T_3 + \overline{T}_3\right)^{1/2}}
447:   \\
448:   \label{eq:62}
449:   && {} +
450:   \frac{\left|C^{95_2}\right|^2}{\left(T_1+\overline{T}_1\right)^{1/2}
451:   \left(T_3+\overline{T}_3\right)^{1/2}}
452: \end{eqnarray}
453: 
454: The notation is that the field theory scalars, the dilaton $S$ and the
455: untwisted moduli $T_i$ originate from closed strings. Open string
456: states $C^b_i$ are required to have their ends localised onto
457: \Dbranes{}. The upper index then specifies which brane(s) their ends
458: are located on, and if both ends are on the same brane, the lower
459: index specifies which pair of compacitified extra dimensions the
460: string is free to vibrate in.
461: 
462: %%%----------------------------------------------------------------------
463: 
464: \section{Sources of lepton flavour violation}
465: \label{sec:sourc-lfv}
466: 
467: %%%----------------------------------------------------------------------
468: 
469: There are two parts to the flavour problem. The first is understanding
470: the origin of the Yukawa couplings (and heavy Majorana masses for the
471: see-saw mechanism), which lead to low energy quark and lepton mixing
472: angles. In low energy SUSY, we also need to understand why flavour
473: changing and/or CP violating processes induced by SUSY loops are so small.
474: A theory of flavour must address both problems simultaneously. For a
475: full discussion of this see the review \cite{Chung:2003fi}.
476: 
477: There are two contributions that can lead to large amounts of flavour
478: violation. The first is the non-alignment of the trilinear soft
479: coupling matrices to the corresponding Yukawa matrices, due to the
480: contribution $F_m \partial_m Y$, $m = \{ H, \overline{H}, \theta,
481: \overline\theta\}$.  The reasons why this can lead to large flavour
482: violation are have been given before \cite{KP0307091}, where a
483: numerical investigation of a model very similar to those considered
484: herein finds that there is a large amount of flavour violation.
485: The second contribution can come from scalar mass matrices which are
486: not proportional to the identity in the theory basis, and lead to 
487: off-diagonal entries in the SCKM basis, resulting in flavour violation.
488: 
489: In this section we begin by defining the SCKM basis, and 
490: in the following subsections we systematically
491: discuss a number of distinct sources of Lepton Flavour Violation (LFV) in 
492: SUGRA models. As well as considering generic SUGRA models, we also allow for a
493: family symmetry, which easily lead to non-universal scalar mass matrices, and
494: non-aligned trilinear matrices\footnote{By non-aligned trilinears, we mean that
495: $\tilde{A}_{ij}/Y_{ij} \ne \mathrm{constant} \ \ (\mathrm{no\ sum}).$
496: }.
497: 
498: 
499: %%% ----------------------------------------------------------------------
500: 
501: \subsection{The SCKM Basis}
502: \label{sec:high_fv}
503: 
504: %%% ----------------------------------------------------------------------
505: 
506: The most
507: convenient basis to work in for considering flavour violating decays,
508: such as $\mu\rightarrow e\gamma$ is the super-CKM (SCKM) basis, which is the
509: basis where the Yukawa matrices are diagonal. If we define the unitary
510: rotation matrices $U_f, V_f$ by
511: \begin{equation}
512:   \label{eq:3} Y^f_{\mathrm{diag}} = U_f Y^f V^\dag_f ,
513: \end{equation} such that $Y^f_{\mathrm{diag}}$ has positive
514: eigenvalues. To convert the physical mass matrices to the SCKM basis,
515: we rotate by the relevant matrix; $U_f$ for the left-handed scalar
516: matrices and $V_f$ for the right-handed. Then flavour violation is
517: proportional to the off-diagonal elements in the SCKM basis, and is
518: suppressed by the diagonal values. The selectron mass matrix is 6 by
519: 6, and the sneutrino mass matrix is 3 by 3 \footnote{The heavy right
520: handed neutrinos cause the right-handed part of the sneutrino mass
521: matrix to decouple by the electroweak scale.}. The selectron mass
522: matrix is
523: \begin{equation}
524:   \label{eq:5}
525:   m^2_{\tilde E} =
526:   \left(
527:     \begin{array}{cc}
528:       Y^e {Y^e}^\dag v_1^2 + \frac{1}{4}v^2(g_2^2 - g_1^2)\mathbf{1} + m_{LL}^2 &
529:       -Y^e v_2 \mu + \tilde{A^e} v_1 \\
530:       {-Y^e}^\dag v_2 \mu^* + \tilde{A^e}^\dag v_1 &
531:       {Y^e}^\dag Y^e v_1^2 + \frac{1}{2}v^2 g_1^2 \mathbf{1}  + m_{ER}^2 
532:     \end{array}
533:   \right) ,
534: \end{equation}
535: where $v^2 = v_2^2 - v_1^2$. The sneutrino mass matrix is then
536: \begin{equation}
537:   \label{eq:6}
538:   m^2_{\tilde \nu} =
539:   \left(
540:     Y^\nu {Y^\nu}^\dag v_2^2 + \frac{1}{4} v^2 g_1^2\mathbf{1} + m_{LL}^2
541:   \right)
542: \end{equation}
543: 
544: Off diagonal elements in any of the 3 by 3 submatrices in the SCKM
545: basis will lead to flavour violation.  We will now consider the $LL$
546: block of $m^2_{\tilde{E}}$. The arguments follow for any other block
547: of $m^2_{\tilde{E}}$ or $m^2_{\tilde \nu}$. The transformation to the
548: SCKM basis is carried out by
549: \begin{equation}
550:   \label{eq:7}
551:   m^2_{\tilde E,LL} = U_e \left( Y^e {Y^e}^\dag v_1^2 +
552: \frac{1}{4}v^2(g_2^2 - g_1^2)\mathbf{1} + m_{LL}^2 \right) U_e^\dag  
553: \end{equation}
554: 
555: $U_e$ is unitary, and $U_e Y^e V_e^\dag$ is diagonal, so the first two
556: terms will be diagonal. Any off-diagonality must come from the third
557: term. If this is proportional to the identity at the GUT scale, it
558: will be approximately equal to the identity at the electroweak scale,
559: which is the scale we should be working at. The fact that this is only
560: approximate is due to the presence of the right handed neutrino fields
561: in the running of the soft scalar mass squared matrices. If, however,
562: the soft mass squared matrices are not proportional to the identity at
563: the GUT scale, then large off-diagonal values will be generated when
564: rotating to the SCKM basis, unless the rotation happens to be
565: small. Generally this won't be the case. Since the family D-term
566: contribution is not proportional to the identity\footnote{This
567: statement assumes that the generational charges are not the same for
568: both left- and right-handed fields. This would remove the point of
569: the family symmetry generating the fermion mass hierarchy.} this will
570: usually be the case\footnote{One can, however, imagine some model with
571: aberrant points in its parameter space where a non-universal non-zero D-term corrects a
572: non-universal base mass matrix to give a universal net mass matrix.}
573: and so we expect large flavour violation in models with Abelian family
574: symmetries when the D-terms correct the scalar mass matrices.
575: 
576: 
577: 
578: %%%----------------------------------------------------------------------
579: 
580: \subsection{The relevance of the Yukawa textures}
581: \label{sec:relev-yukawa-textures}
582: 
583: %%%----------------------------------------------------------------------
584: 
585: There is one subtlety concerning the size of the off-diagonal elements of the
586: scalar mass matrices in the SCKM basis.  This comes back to the
587: definition of the SCKM basis as the basis in which the Yukawa matrices
588: are diagonal. The larger the SCKM transformation between any `theory'
589: basis and the mass eigenstate basis for the Yukawa matrices, the
590: larger the SCKM transformation that must be performed on the 
591: scalar mass matrices in going to the SCKM basis,
592: hence the larger the off-diagonal elements of the 
593: scalar mass matrices in the SCKM basis generated from non-equal
594: diagonal elements in the `theory' basis.
595: \footnote{
596: Note that the D-terms make us sensitive to right-handed mixings in the Yukawa
597: matrices, so the non-universal family charge structure for the right-handed
598: scalar masses may lead to a non-universal generational hierarchy in
599: the right-handed scalar mass matrices.}
600: The larger the off-diagonal entries in the SCKM basis
601: compared to the diagonal ones, the greater will be 
602: the flavour violation. Also, the greater the mass difference between the
603: diagonal elements in the `theory' basis, 
604: the greater the size of the off-diagonal entries
605: produced when rotating from the 'theory' basis, hence the larger the
606: flavour violating effect. Clearly these effects are sensitive to the 
607: size of the transformation required to go to the SCKM basis,
608: which in turn is sensitive to the particular choice of Yukawa
609: textures in the `theory' basis. In this way, the choice of Yukawa 
610: texture can play an important part on controlling the magnitude
611: of flavour violation, and we shall see examples of this later.
612: 
613: 
614: %%%----------------------------------------------------------------------
615: 
616: \subsection{Running effects}
617: \label{sec:running-effects-only}
618: 
619: %%%----------------------------------------------------------------------
620: 
621: Consider the case where, at the high-energy scale, the scalar mass
622: matrices are proportional to the identity matrix and each soft
623: trilinear coupling matrix is aligned to the corresponding Yukawa
624: matrix:
625: \begin{equation}
626:   \label{eq:8} \left(m^2_{f}\right)_{ij} = m^2_{0,f} \delta_{ij}
627: \;\;\;,\;\;\; \left(\tilde{A}^f\right)_{ij} = A_0^f Y^f_{ij} .
628: \end{equation}
629: This is often referred to as mSUGRA.
630: In the quark sector, due to the quark flavour violation responsible
631: for CKM mixing, 
632: when the scalar squark 
633: mass matrices are run down to the electroweak scale, they
634: will run to non-universal scalar mass matrices and non-aligned
635: trilinear coupling matrices. If this is the case, then in the SCKM
636: basis, which is the basis where the Yukawa matrices are diagonal,
637: off-diagonal elements in the scalar squark mass matrices or the trilinear
638: squark mass matrices lead to flavour violation.
639: 
640: In the lepton sector, in the absence of neutrino masses
641: the separate lepton flavour numbers are conserved and mSUGRA will not lead to
642: any LFV induced by running the matrices down to low energy. 
643: However, in the presence of neutrino masses, 
644: with right-handed neutrino fields included 
645: to allow a see-saw explanation of neutrino masses and
646: mixing angles, the separate lepton flavour numbers will be violated and, 
647: even in the mSUGRA type scenario, running effects
648: will generate off-diagonal elements in the scalar
649: mass matrices in the SCKM basis, resulting in low energy LFV.
650: 
651: 
652: %%%----------------------------------------------------------------------
653: 
654: 
655: %\subsubsection{mSUGRA with see-saw mechanism}
656: %\label{sec:msugra-with-see}
657: 
658: %%%----------------------------------------------------------------------
659: 
660: %%%----------------------------------------------------------------------
661: 
662: \subsection{Diagonal scalar mass matrices not propotional to the unit matrix}
663: \label{sec:diagonal-scalar-mass}
664: 
665: %%%----------------------------------------------------------------------
666: 
667: %%%----------------------------------------------------------------------
668: 
669: \subsubsection{Non-minimal SUGRA}
670: \label{sec:non-minimal-sugra-dsm}
671: 
672: %%%----------------------------------------------------------------------
673: 
674: In non-minimal SUGRA the scalar mass matrices may be diagonal at the
675: high-energy scale, but not proportional to the identity. In this case,
676: there will be non-zero off-diagonal elements in the SCKM basis even
677: with no contribution to running effects, or contribution from the
678: trilinear coupling matrices.
679: 
680: One way of getting diagonal mass matrices not proportional to the
681: unit matrix is from a SUGRA model
682: corresponding to the low energy limit of a string model with
683: D-branes. If each generation from the field theory viewpoint
684: corresponds to a string attaching to different branes, then the masses
685: predicted in the SUGRA can be different. This leads to diagonal but
686: non-universal scalar mass matrices.
687: 
688: %%%----------------------------------------------------------------------
689: 
690: \subsubsection{D-term contributions from broken family gauge groups}
691: \label{sec:d-term-contributions-sbfgg}
692: 
693: %%%----------------------------------------------------------------------
694: 
695: Another way of getting diagonal mass matrices not proportional
696: to the unit matrix is by having a model
697: with a gauge family symmetry, which is broken spontaneously. When the
698: Higgs which breaks the family group, the flavon, gets a vev, it
699: contributes a squark (slepton) mass contribution through the four
700: point scalar gauge interaction which has two flavons and two  squarks
701: (sleptons).
702: 
703: To make the point more explicitly, consider a $U(1)$ family
704: group. Then the mass contribution is proportional to the charge under
705: the family symmetry. As the point of a family symmetry is to explain
706: the hierarchy of fermion masses, small quark mixing angles and large
707: neutrino mixing angles, the charges are usually different.
708: 
709: Then, even if the mass matrix starts off as a universal matrix, it
710: will be driven non-universal by the D-term contribution:
711: \begin{equation}
712:   \label{eq:9} m^2_{L_L} = \left[
713:     \begin{array}{ccc} m_0^2 \\ & m_0^2 \\ && m_0^2
714:     \end{array} \right] + D^2 \left[
715:       \begin{array}{ccc} q_{L1} \\ & q_{L2}  \\ && q_{L3}
716:       \end{array} \right] .
717: \end{equation}
718: 
719: %%% ----------------------------------------------------------------------
720: 
721: \subsection{Non-aligned trilinears}
722: \label{sec:non-aligned trilinears}
723: 
724: %%% ----------------------------------------------------------------------
725: 
726: % Primordial flavour violation (but don't call it that here).
727: 
728: \subsubsection{Non-minimal SUGRA}
729: \label{sec:non-minimal SUGRA}
730: 
731: One way of getting non-aligned trilinear matrices is by having the
732: same sort of non-minimal SUGRA setup that leads to diagonal but
733: non-universal mass matrices, as described in Section
734: \ref{sec:non-minimal-sugra-dsm}. From the supegravity equations from
735: Section \ref{sec:soft-from-SUGRA}, the trilinears that appear in the
736: soft Lagrangian, $\tilde{A}_{ij}$ will be non-aligned if the
737: trilinears predicted by the SUGRA model, $A_{ij}$ are not democratic,
738: i.e. if $A_{ij} \ne \mathrm{constant}$.  From a string-inspired/SUGRA
739: standpoint, if each generation is assigned to a different brane and
740: extra-dimensional vibrational direction, then in general we expect
741: $A_{abc}$ to be different, due to the differing values of the K\"ahler
742: metrics $\tilde{K}_a$ for the different brane assignents $C^i_j$.
743: When $\tilde{A}_{ij}$ is transformed to the SCKM basis at the
744: electroweak scale, there will then be large off-diagonal elements
745: which contribute to flavour violating processes.
746: 
747: \subsubsection{Flavon contributions from the Yukawa couplings}
748: \label{sec:FNcontributions}
749: 
750: In general 
751: when one considers a family symmetry in order to understand the origin
752: of the Yukawa couplings, the new fields arising
753: from this can develop F-term \vev{}s, and contribute to the supersymmetry
754: breaking F-terms in a non-universal way. This leads to a dangerous source
755: of flavour violating non-aligned trilinears:
756: \cite{Abel:2001ur,Ross:2002mr,KP0307091,King:2003xn},
757: \begin{equation}
758:   \label{eq:10}
759:   \Delta A  = F_\theta \partial_\theta \ln Y 
760: \end{equation}
761: where the Yukawa coupling $Y$ in Eq.~(\ref{eq:10}) arises from the
762: an effective FN operator and is a polynomial of the FN field $\theta$,
763: $Y \sim \theta^n $, leading to
764: \begin{equation}
765: \Delta A = F_\theta \partial_\theta \mathrm{ln} \theta^n = F_\theta
766: \frac{n}{\theta} .
767: \end{equation}
768: However the auxiliary field is proportional to the scalar component,
769: \begin{equation}
770: F_\theta \propto m_{3/2} \theta \Rightarrow \Delta A \propto nm_{3/2} .
771: \end{equation}
772: An example of this with an arbritary $U(1)$ family symmetry is
773: \begin{equation}
774: Y_{ij} = a_{ij} \left( \frac{\theta}{M} \right)^{p(i,j)} \Longrightarrow 
775: \Delta A_{ij} \sim m_{3/2} p(i,j)
776: \end{equation}
777: The $a_{ij}$ are arbiratry couplings, all of which should be $O(1)$
778: for the symmetry to be considered natural. The $p(i,j)$ are
779: integers appearing as a power for the $ij$-th element of the above
780: Yukawa, and it comprises the sum of the family charges for the
781: $i$th-generation left-handed field and $j$th-generation right-handed
782: field.
783: In principle, if the Yukawa texture is set up so that each power is different,
784: then each element in $A_{ij}$ will be different from each other, and the
785: physical trilinear matrix, $\tilde{A}_{ij}$ will be non-aligned to the
786: corresponding Yukawa. 
787: Due to the dependence on the charges of the different fields,
788: this contribution to the trilinears is not diagonalised when we transform
789: to the SCKM basis.
790: 
791: 
792: %%% ----------------------------------------------------------------------
793: 
794: %%% ----------------------------------------------------------------------
795: 
796: % When one transitions to the SCKM basis at the high-energy scale, any
797: % off-diagonal elements will cause flavour violation.
798: % Rather than using the explicit diagonalization matrices, one can recall
799: % that the magnitude of the flavour violating effects is related to the
800: % off-diagonal elements of the sfermion mass matrices. If these off-diagonal
801: % entries are subdominant compared to the diagonal ones, we can consider
802: % the mass insertion approximation. This uses a perturbation expansion in
803: % the off-diagonal elements of the sfermion mass matrices, normalised
804: % by a common sfermion mass, to describe the sfermion diagonalisation matrices.
805: 
806: % This method is discussed at length in \cite{Chung:2003fi}, so here we shall
807: % just indicate the structure of the mass insertion parameters:
808: 
809: % \begin{equation}
810: % (\delta_{AB})_{ij} = \frac{(\Delta_{AB})_{ij}}{\sqrt{(m^2_{AA})_{ii}
811: % (m^2_{BB})_{jj}}} ,
812: % \end{equation}
813: % where the choice of the denominator follows the convention of
814: % \cite{Gabbiani:1996hi}.
815: 
816: % The off-diagonal terms in the sfermion mass matrices are $(\Delta_{AB})_{ij}$,
817: % where $AB$ is $LL$, $RR$, $LR$, or $RL$, with $i,j = 1,2,3$ as family indices.
818: % The diagonal terms are denoted $(m^2_{AA})_{ii}$ where $AA$ can only be $LL$ or
819: % $RR$.
820: 
821: % % [NOTE:$\Delta$ has mass dimension 2, so $\delta$ is dimensionless.]
822: % % DO THIS IN MORE DETAIL IN THESIS.
823: % % [Personal  NOTE: This also leads to diagonal non-universal scalars]
824: 
825: % Only the diagonal entries of the $LL$ and $RR$ mass insertion parameters
826: % are influenced by electroweak symmetry breaking as the flavour violating
827: % entries originate purely from supersymmetry breaking. However the
828: % $(\delta_{LR})_{ij}$ and $(\delta_{RL})_{ij}$ blocks are generated after
829: % electroweak breaking, so their typical size is consequently the geometric
830: % mean of the electroweak and soft supersymmetry breaking scales. These mass
831: % insertion parameters involve linear combinations of $\tilde{A}$ and $\mu$,
832: % rotated by the matrices which diagonalise the Yukawas. The $LL$
833: % and $RR$ mass insertion parameters involve the soft mass-squared parameters
834: % $m^2_Q$ and $m^2_U$ rotated by the left- and right-handed quark
835: % diagonalisation matrices respectively.
836: 
837: 
838: %%% ----------------------------------------------------------------------
839: 
840: \section{Intersecting D-brane models with an Abelian family symmetry}
841: \label{sec:442-pati-salam}
842: 
843: %%% ----------------------------------------------------------------------
844: %%% ----------------------------------------------------------------------
845: 
846: \subsection{Symmetries and symmetry breaking}
847: \label{sec:symm and symm breaking}
848: 
849: %%% ----------------------------------------------------------------------
850: 
851: 
852: 
853: \begin{figure}
854: \epsfysize=2.5truein
855: \epsffile{const.eps}
856: \caption{\small A generic type I string construction involving two sets of
857: perpendicular D5-branes embedded within a D9-brane, where the D5-brane 
858: world-volumes intersect at the origin.  
859: Charged chiral fields appear as open strings with both ends attached to the 
860: same D-brane $C^{5_{i}}_{j}$ and $C^{9}_{j}$, or different branes 
861: $C^{5_{1} 5_{2}}$ and $C^{9 5_{i}}$.  Closed
862: strings ($S,T_{i}$) can live in the full 10d space, although orbifolding 
863: leads to closed strings (twisted moduli $Y_{k}$) localised at 4d fixed points 
864: within the $D5_{i}$-brane world-volume.}
865: \label{const}
866: \end{figure}
867: 
868: 
869: 
870: In order to study the effects of LFV elucidated in the previous 
871: section, it is necessary to specialize to a particular
872: effective non-minimal SUGRA model which addresses the 
873: question of flavour (i.e. provides a theory of the Yukawa couplings).
874: The specific model we shall discuss is 
875: defined in Table~\ref{tab:particle_content_42241}.
876: This model is an extension of the Supersymmetric
877: Pati-Salam model discussed 
878: in ref.\cite{Everett:2002pm}, based on two $D5$ branes which intersect
879: at $90$ degrees and preserve SUSY down to the TeV energy scale.
880: The generic D-brane set-up that we use is illustrated
881: in Fig.\ref{const}, where the string assignment notation
882: is defined. The gauge group of the $5_{1}$ sector is 
883: $U(4)^{(1)}\times U(2)^{(1)}_{L}\times U(2)^{(1)}_{R} $, 
884: and the gauge group of the
885: $5_{2}$  sector is \( U(4)^{(2)} \) (e.g., we assume the $U(2)_{L,R}$ of the
886: $5_2$ sector are broken).
887: The symmetry breaking pattern of this model takes place in two stages,
888: which we assume occur at very similar scales $\sim M_X$. In the first
889: stage, the $U(4)$ groups are broken to the diagonal subgroup via diagonal
890: VEV's of bifundamentals; the resulting theory is an effective   
891: Pati-Salam model (with additional $U(1)$'s) which then breaks to the MSSM
892: (and a number of additional $U(1)'s$) via the usual Higgs pair of
893: bifundamentals. The string scale is taken to be equal to the GUT scale, about
894: $3\times 10^{16}$ GeV. 
895: 
896: The symmetry breaking pattern leads to the following relations
897: among the gauge couplings of the SM gauge groups in terms of the gauge
898: couplings $g_{5_1}$ and $g_{5_2}$ associated with the gauge groups of the
899: $5_1$ and $5_2$ sectors: 
900: \begin{eqnarray}
901: \label{gaugecouplings3}
902: g_{3} & = & 
903: \frac{g_{5_{1}}g_{5_{2}}}{\sqrt{g^{2}_{5_{1}}+g^{2}_{5_{2}}}}=g_4\\
904: g_{2} & = & g_{5_{1}} =g_{2R}\\
905: g_{1} & = & \frac{\sqrt{3}g_{3}g_{2}}{\sqrt{3g^{2}_{3}+2g^{2}_{2}}}.
906: \label{gaugecouplings1}
907: \end{eqnarray}
908: 
909: \begin{table}[htbp]
910:   \centering \mbox{
911:   \begin{tabular}{|c||c|c|c|c||c|c|}
912:     \hline Field & $\mathrm{U}(4)^{(1)}$ & $\mathrm{U}(2)^{(1)}_L$ &
913:     $\mathrm{U}(2)^{(1)}_R$ & $\mathrm{U}(4)^{(2)}$ & Ends & $\UI_F$
914:     charge \\ 
915:     \hline $h$ & 1 & 2 & 2 & 1
916:     & $C_1^{5_1}$ & $0$ \\ 
917:     $F_3$ & 4 & 2 & 1 & 1 & $C_2^{5_1}$ &  $q_{L3}$  \\ 
918:     $\overline{F}_3$ & $\overline{4}$ & 1 & 2 &
919:     1 & $C_3^{5_1}$ & $q_{R3}$ \\ 
920:     $F_2$ & 1 & 2     & 1 & 4 & $C^{5_1 5_2}$ & $q_{L2}$ \\ 
921:     $\overline{F}_2$ & 1 & 1 & 2 & $\overline{4}$ & $C^{5_1 5_2}$ & $q_{R2}$  \\ 
922:     $F_1$ & 1 & 2 & 1 & 4 & $C^{5_1 5_2}$ & $q_{L1}$  \\
923:     $\overline{F}_1$ & 1 & 1 & 2 & $\overline{4}$ & $C^{5_1 5_2}$ &     $q_{R1}$ \\ 
924:     $H$ & 4 & 1 & 2 & 1 & $C_1^{5_1}$ &     $q_H$ \\ 
925:     $\overline{H}$ & $\overline{4}$ & 1 & 2 & 1 & $C_2^{5_1}$ & $\!\!\!\!\!-q_H$\\ 
926:     $\varphi_1$ & 4 & 1 & 1 & $\overline{4}$ & $C^{5_1 5_2}$ & $-$ \\
927:     $\varphi_2$ & $\overline{4}$ & 1 & 1 & 4 & $C^{5_1 5_2}$ & $-$ \\ 
928:     $D_6^{(+)}$ & 6 & 1 & 1 & 1 & $C_1^{5_1}$ & $-$  \\
929:     $D_6^{(-)}$ & 6 & 1 & 1 & 1 & $C_2^{5_2}$ & $-$  \\ 
930:     $\theta$  & 1 & 1 & 1 & 1 & $C^{5_1 5_2}$ & $\!\!\!\!\!-1$ \\
931:     $\overline{\theta}$ & 1 & 1 & 1 & 1 & All & $1$ \\
932:     \hline
933:   \end{tabular} }
934:   \caption{{\small The particle content of the 42241 model, and the brane
935: assignments of the corresponding string. Note that the string
936: assignment of $\overline{\theta}$ is allowed to be any of $\{ C^{5_1
937: 5_2}, C^{5_1}_1, C^{5_1}_2, C^{5_1}_3\}$, giving a slightly different
938: model in each case.  The $U(1)_F$ charges will change, also generating
939: a slightly different model in each case. We will return to this in
940: more detail in Section~\ref{sec:numerical-etc}. } }
941:   \label{tab:particle_content_42241}
942: \end{table}
943: 
944: The extension is to include an additional $U(1)_F$ family symmetry and
945: the FN operators responsible for the Yukawa couplings
946: as in \cite{Blazek:2003wz} (see also
947: \cite{King:OperatorAnalysis}).  
948: The charges under the Abelian symmetry $U(1)_F$ are left arbritary for
949: now. The present `42241' Model is the same
950: as the model considered in \cite{KP0307091}, with the following
951: modifications considered; firstly, we allow the Froggatt-Nielsen field
952: $\overline{\theta}$ to be either an intersection state or attached to
953: the $5_1$ brane. The location of $\overline{\theta}$ dramatically
954: changes the value of the D-term contribution to the scalar masses
955: coming from the FN sector.
956: 
957: The quark and lepton fields are contained in the representations 
958: $F, \overline{F}$ which are assigned
959: charges under $\UI_F$. In Table~\ref{tab:particle_content_42241}
960: we list the charges, string assignments and representations under the
961: string gauge group 
962: $U(4)^{(1)} \otimes U(2)^{(1)}_L \otimes U(2)^{(1)}_R \otimes U(4)^{(2)}$.
963: 
964: The field $h$ represents both Electroweak Higgs doublets that
965: we are familiar with from the MSSM.  The fields $H$ and $\overline{H}$
966: are the Pati-Salam Higgs scalars;\footnote{We will also refer to these as
967: ``Heavy Higgs''; this has nothing to do with the MSSM heavy neutral higgs
968: state $H^0$.} the bar on the second is used to note that it is in the conjugate
969: representation compared to the unbarred field.
970: 
971: 
972: The extra Abelian $U(1)_F$ gauge group is a family symmetry, and is broken at
973: the high energy scale by the \vevs{} of the FN fields \cite{Froggatt:1978nt} 
974: $\theta,\overline{\theta}$, which have charges $-1$ and $+1$ respectively
975: under $U(1)_F$. 
976: We assume that the singlet field $\theta$ 
977: arises as an intersection
978: state between the two \Dbranes{5}, transforming under the remnant
979: \UI s in the 4224 gauge structure. In general the FN fields are expected to 
980: have non-zero F-term \vev s.
981: 
982: The two $SU(4)$ gauge groups are broken to their diagonal
983: subgroup at a high scale due to the assumed \vevs{} of the
984: bifundamental
985: Higgs fields $\varphi_1$, $\varphi_2$ \cite{Everett:2002pm}.
986: The symmetry breaking at the scale $M_X$
987: \begin{equation}
988:   \label{eq:gauge_breaking_pattern}
989:   \psgroup \rightarrow \smgroup
990: \end{equation}
991: is achieved by the heavy Higgs fields $H$, $\overline{H}$
992: which are assumed to gain \vevs{} \cite{King:OperatorAnalysis}
993: \begin{equation}
994:   \label{eq:heavy_higgs_\vevs}
995:   \left< H^{\alpha b} \right> = \left< \nu_H \right> = V
996:   \delta^\alpha_4 \delta^b_2 \sim M_X \;\;\; ;\;\;\; \left<
997:   \overline{H}_{\alpha x} \right> = \left< \overline{\nu}_H \right> =
998:   \overline{V} \delta_\alpha^4 \delta_x^2 \sim M_X
999: \end{equation}
1000: This symmetry breaking splits the Higgs field $h$
1001: into two Higgs doublets, $h_1$, $h_2$. Their neutral components then gain
1002: weak-scale \vevs{}.
1003: \begin{equation}
1004:   \label{eq:mssm_higgs_\vevs}
1005:   \left< h_1^0 \right> = v_1 \;\;\; ; \;\;\;
1006:   \left< h_2^0 \right> = v_2 \;\;\; ; \;\;\;
1007:   \tan \beta = v_2 / v_1.
1008: \end{equation}
1009: The low energy limit of this model contains the MSSM with right-handed
1010: neutrinos.  We will return to the right handed neutrinos when we
1011: consider operators including the heavy Higgs fields $H$,
1012: $\overline{H}$ which lead to effective Yukawa contributions and
1013: effective Majorana mass matrices when the heavy Higgs fields gain
1014: \vevs.
1015: 
1016: %%% ----------------------------------------------------------------------
1017: 
1018: \subsection{Yukawa operators}
1019: \label{sec:yukawa-sector}
1020: 
1021: %%% ----------------------------------------------------------------------
1022: 
1023: The (effective) Yukawa couplings are generated by operators
1024: involving the FN field $\theta$ with
1025: the following structure:\footnote{The field $\overline{\theta}$ will not
1026: enter the Yukawa operators because $F_i \overline{F}_j h$ will be positive for
1027: any $i,j$.}
1028: \cite{King:OperatorAnalysis}:
1029: \begin{equation}
1030:   \label{eq:dirac_operator_n_is_n}
1031:   \mathcal{O} = F_i \overline{F}_j h \left(\frac{H
1032:   \overline{H}}{M_X^2}\right)^n
1033:   \left(\frac{\theta}{M_X}\right)^{p(i,j)}
1034: \end{equation}
1035: where the integer
1036: $p(i,j)$ is the total $U(1)_F$ charge of $F_i + \overline{F}_j + h$ and
1037: $H\overline{H}$ has a $U(1)_F$ charge of zero.
1038: The tensor structure of the operators in
1039: Eq.~(\ref{eq:dirac_operator_n_is_n}) is
1040: \begin{equation}
1041:   \label{eq:dirac_mass_tensor}
1042:   \left(\mathcal{O}\right)^{\alpha\rho y w}_{\beta\gamma x z} =
1043:   F^{\alpha a} \overline{F}_{\beta x} h^y_a \overline{H}_{\gamma z}
1044:   H^{\rho w} \theta^{p(i,j)}
1045: \end{equation}
1046: One constructs
1047: $\SU{4}_{PS}$ invariant tensors $C^{\beta\gamma}_{\alpha\rho}$ that
1048: combine $4$ and $\overline{4}$ representations of $\SU{4}_{PS}$ into
1049: \boldmath $1$, $6$, $10$, $\overline{10}$ and $15$ \unboldmath
1050: representations \cite{King:OperatorAnalysis}. Similarly we construct $\SU{2}_R$
1051: tensors $R^{xz}_{yw}$ that combine the $\mathbf{2}$ representations of \SU{2} into
1052: singlet and triplet representations. These tensors are contracted
1053: together and into $\mathcal{O}^{\alpha\rho y w}_{\beta \gamma
1054: x z}$ to create singlets of $\SU{4}_{PS}$, $\SU{2}_L$ and $\SU{2}_R$.
1055: Depending on which operators are used, different
1056: Clebsch-Gordan coefficients (CGCs) will emerge.
1057: 
1058: We will return to these in section \ref{sec:numerical-etc}, when we define
1059: the two models that we will be using for the numerical analysis.
1060: 
1061: %%% ----------------------------------------------------------------------
1062: 
1063: \subsection{Majorana operators}
1064: \label{sec:majorana-fermions}
1065: 
1066: %%% ----------------------------------------------------------------------
1067: 
1068: We are interested in Majorana fermions because they can contribute
1069: neutrino masses of the correct order of magnitude via the see-saw
1070: effect. The operators for Majorana fermions are of the form
1071: 
1072: \begin{equation}
1073:   \label{eq:majorana_operators_n_is_n}
1074:   \mathcal{O}_{ij} = \overline{F}_i \overline{F}_j \left(\frac{H H}{M_X}\right)
1075:   \left(\frac{H \overline{H}}{M_X^2}\right)^{n-1}
1076:   \left(\frac{\theta}{M_X}\right)^{q(i,j)}
1077: \end{equation}
1078: 
1079: There do not exist renormalisable elements of this infinite series of
1080: operators, so $n < 1$ Majorana operators are not defined, except in
1081: the $(3,3)$ element. We assume that a $(3,3)$ neturino mass term is allowed at leading
1082: (but non-renormalisable) order.
1083: A similar analysis goes through
1084: as for the Dirac fermions; however the structures only ever give
1085: masses to the neutrinos, not to the electrons or to the
1086: quarks.
1087: 
1088: %%% ----------------------------------------------------------------------
1089: 
1090: \section{Soft supersymmetry breaking masses}
1091: \label{soft}
1092: 
1093: %%% ----------------------------------------------------------------------
1094: 
1095: \subsection{Supersymmetry breaking F-terms}
1096: 
1097: %%% ----------------------------------------------------------------------
1098: 
1099: In \cite{Everett:2002pm} it was assumed that the Yukawas were
1100: field-independent, and hence the only $F$-vevs of importance were that
1101: of the dilaton ($S$), and the untwisted moduli ($T^i$).
1102: Here we set out the parameterisation for the F-term \vevs{},
1103: including the contributions from the FN field $\theta$ and the
1104: heavy Higgs fields $H,\overline{H}$. Note that
1105: the field dependent part follows from the assumption that the
1106: family symmetry field, $\theta$ is an intersection state.
1107: \begin{eqnarray}
1108:   \label{eq:dilaton_auxilliary_vev_42241}
1109:   F_S
1110:   &=&
1111:   \sqrt{3} m_{3/2} \left( S + \overline{S} \right)
1112:   X_S
1113:   \\
1114:   \label{eq:untwisted_mod_auxilliary_vev_42241}
1115:   F_{T_i}
1116:   &=&
1117:   \sqrt{3} m_{3/2} \left( T_i + \overline{T}_i \right)
1118:   X_{T_i}
1119:   \\
1120:   \label{eq:heavy_higgs_auxilliary_vev_42241}
1121:   F_{H^{\alpha b}} 
1122:   &=&
1123:   \sqrt{3} m_{3/2} H^{\alpha b} \left( S + \overline{S} \right)^\frac{1}{2}
1124:   X_H
1125:   \\
1126:   \label{eq:heavy_conj_higgs_auxilliary_vev_42241}
1127:   F_{\overline{H}_{\alpha x}} &=& \sqrt{3} m_{3/2}
1128:   \overline{H}_{\alpha x} \left( T_3 + \overline{T}_3
1129:   \right)^\frac{1}{2} X_{\overline{H}} \\
1130:   \label{eq:family_field_auxilliary_vev_42241}
1131:   F_\theta
1132:   &=&
1133:   \sqrt{3} m_{3/2} \theta \left(S + \overline{S}\right)^\frac{1}{4}
1134:   \left(T_3 + \overline{T}_3\right)^\frac{1}{4}
1135:   X_\theta
1136: \end{eqnarray}
1137: We introduce a shorthand notation:
1138: \begin{equation}
1139:   \label{eq:shorthand_key}
1140:   F_H H = \sum_{\alpha b} F_{H^{\alpha b} } H^{\alpha b} \; ;\;
1141:   F_{\overline{H}} \overline{H} = \sum_{\alpha x}
1142:   F_{\overline{H}_{\alpha x} } \overline{H}_{\alpha x}.
1143: \end{equation}
1144: 
1145: The F-terms above use values of $S$ and $T_i$ which 
1146: are given in terms of the gauge couplings as:
1147: \begin{equation}
1148: \mathrm{Re}(S) = \frac{4\pi}{g^2_9},\ \ 
1149: \mathrm{Re}(T_i) = \frac{4\pi}{g^2_{5_i}}.
1150: \end{equation}
1151: The gauge couplings $g_{5_1},g_{5_2}$ are given from 
1152: Eq.~(\ref{gaugecouplings3})-(\ref{gaugecouplings1}) \cite{Everett:2002pm} as
1153: \begin{equation}
1154: g_{5_1} = g_2, \ \ g_{5_2} = \frac{g_2 g_3}{\sqrt{g^2_2
1155:     -g^2_3}},
1156: \end{equation} 
1157: where we shall assume that at the scale $M_X$ we have,
1158: \begin{equation}
1159: g_2 = 0.7345, \ \ g_3 = 0.6730 .
1160: \end{equation}
1161: The values of $g_{9},g_{5_3}$ are assumed to be equal and
1162: are obtained from the string relation
1163: \begin{equation}
1164: \label{eq:string gauge couplings}
1165: 32 \pi^2 \left( \frac{M_*}{M_{Pl}} \right)^2 = g_9 g_{5_1} g_{5_2} g_{5_3} ,
1166: \end{equation}
1167: as 
1168: \begin{equation}
1169: g_9 = g_{5_3} = 0.0266 ,
1170: \end{equation}
1171: where we have taken
1172: \begin{equation}
1173: \left( \frac{M_*}{M_{Pl}} \right)^2 = 2.77\times10^{-6}.
1174: \end{equation}
1175: These rather small gauge couplings imply
1176: \begin{equation}
1177: \mathrm{Re}(S) = \mathrm{Re}(T_3) = 0.877.
1178: \end{equation}
1179: In \cite{KP0307091} the string relation was not used and 
1180: it was assumed incorrectly that 
1181: $g_9 = g_{5_3} = g_2 $ which resulted in 
1182: $\mathrm{Re}(S) = 27.7$.
1183: 
1184: %%% ----------------------------------------------------------------------
1185: 
1186: \subsection{Soft scalar masses}
1187: \label{sec:scalars_42241}
1188: 
1189: %%% ----------------------------------------------------------------------
1190: 
1191: There are two contributions to scalar mass
1192: squared matrices, coming from SUGRA and from D-terms.
1193: In this subsection we calculate the SUGRA predictions for the
1194: matrices at the GUT scale, and in the next subsection we add on the 
1195: D-term contributions.
1196: 
1197: The SUGRA contributions to soft masses are detailed in 
1198: Section~\ref{sec:soft-from-SUGRA}.
1199: 
1200: From Eq.~(\ref{eq:normalised_scalars})  we can get the family independent
1201: form for all scalars:
1202: \begin{eqnarray}
1203:   \label{eq:scalars_left}
1204:   m^2_{L} &=& m^2_{3/2}
1205:   \left[
1206:     \begin{array}{ccc}
1207:       a & & \\
1208:       & a & \\
1209:       & & b_L
1210:     \end{array}
1211:   \right] \\ \label{eq:scalars_right}
1212:   m^2_{R} &=& m^2_{3/2}
1213:   \left[
1214:     \begin{array}{ccc}
1215:       a & & \\
1216:       & a & \\
1217:       & & b_R
1218:     \end{array}
1219:   \right] \\
1220:   \label{eq:mssm_higgs_at_mx}
1221:   m^2_h &=& m^2_{3/2} ( 1 - 3 X^2_S )\\
1222:   \label{eq:ps_H_at_mx}
1223:   m^2_H &=& m^2_{3/2} ( 1 - 3 X^2_S ) \\
1224:   \label{eq:ps_Hbar_at_mx}
1225:   m^2_{\overline{H}} &=& m^2_{3/2}(1 - 3 X^2_{T_3} ) \\
1226:   \label{eq:FN soft mass}
1227:   m^2_\theta &=& m^2_{3/2} \left[ 1 - \frac{3}{2} (X^2_S + X^2_{T_3} ) \right]
1228: \end{eqnarray}
1229: where
1230: \begin{eqnarray}
1231:   \label{eq:SUGRA_a}
1232:   a &=& 1 - \frac{3}{2} (X^2_S + X^2_{T_3} ) \\
1233:   \label{eq:SUGRA_b_L}
1234:   b_L &=& 1 - 3 X^2_{T_3} \\
1235:   \label{eq:SUGRA_b_R}
1236:   b_R &=& 1 - 3 X^2_{T_2}
1237: \end{eqnarray}
1238: Here $m^2_L$ represents the left handed scalar mass squared
1239: matricies $m^2_{Q_L}$ and $m^2_{L_L}$. $m^2_R$ represents the right
1240: handed scalar mass squared matricies $m^2_{U_R}$, $m^2_{D_R}$,
1241: $m^2_{E_R}$ and $m^2_{N_R}$. A discussion of the equations for
1242: $m^2_{\overline{\theta}}$ can be found in Section~\ref{sec:theta_bar D-terms}.
1243: 
1244: %%% ----------------------------------------------------------------------
1245: 
1246: \subsection{D-term contributions}
1247: \label{sec:D-term_contributions}
1248: 
1249: %%% ----------------------------------------------------------------------
1250: 
1251: There are two D-term contributions to the scalar masses. The first is the well
1252: known \cite{King:2000vp,KP0307091} contribution from the breaking of the Pati-Salam
1253: group to the MSSM group. Note that these D-terms are different to those quoted
1254: in the references above as we now consider the D-terms generated by breaking a
1255: family symmetry. See Appendix~\ref{sec:deriv D-terms} for a full derivation.
1256: The second D-term comes solely from the breaking of the $U(1)$ family
1257: symmetry. The corrections lead to the following mass matrices:
1258: 
1259: \begin{eqnarray}
1260: \label{eq:m^2_QL}
1261: m^2_{Q_L} &=& m^2_L + {\bf 1} (g^2_4) D^2_H +
1262: \begin{pmatrix}
1263: q_{L1} & & \\ & q_{L2} & \\ & & q_{L3}
1264: \end{pmatrix}
1265: g^2_F D^2_\theta \; , \\
1266: \label{eq:m^2_LL}
1267: m^2_{L_L} &=& m^2_L -{\bf 1} (3g^2_4) D^2_H +
1268: \begin{pmatrix}
1269: q_{L1} & & \\ & q_{L2} & \\ & & q_{L3}
1270: \end{pmatrix}
1271: g^2_F D^2_\theta \; , \\
1272: \label{eq:m^2_UR}
1273: m^2_{U_R} &=&  m^2_R -{\bf 1}(g^2_4 -2g^2_{2R}) D^2_H +
1274: \begin{pmatrix}
1275: q_{R1} & & \\ & q_{R2} & \\ & & q_{R3}
1276: \end{pmatrix}
1277: g^2_F D^2_\theta \; , \\
1278: \label{eq:m^2_DR}
1279: m^2_{D_R} &=&  m^2_R -{\bf 1} (g^2_4 +2g^2_{2R}) D^2_H +
1280: \begin{pmatrix}
1281: q_{R1} & & \\ & q_{R2} & \\ & & q_{R3}
1282: \end{pmatrix}
1283: g^2_F D^2_\theta \; , \\
1284: \label{eq:m^2_ER}
1285: m^2_{E_R} &=&  m^2_R +{\bf 1} (3g^2_4 -2g^2_{2R}) D^2_H +
1286: \begin{pmatrix}
1287: q_{R1} & & \\ & q_{R2} & \\ & & q_{R3}
1288: \end{pmatrix}
1289: g^2_F D^2_\theta \; , \\
1290: \label{eq:m^2_NR}
1291: m^2_{N_R} &=&  m^2_R +{\bf 1} (3g^2_4 +2g^2_{2R}) D^2_H +
1292: \begin{pmatrix}
1293: q_{R1} & & \\ & q_{R2} & \\ & & q_{R3}
1294: \end{pmatrix}
1295: g^2_F D^2_\theta \; , \\
1296: \label{eq:m^2_hu}
1297: m^2_{h_u} &=& m^2_{h_2} - 2g^2_{2R} D^2_H \; , \\
1298: \label{eq:m^2_hd}
1299: m^2_{h_d} &=& m^2_{h_1} + 2g^2_{2R} D^2_H \; ,
1300: \end{eqnarray}
1301: 
1302: The charges $q_{Li}, q_{Rj}$ are the charges under $U(1)_F$ of $F_i$
1303: and $\overline{F}_j$ respectively, as shown in
1304: Table~\ref{tab:charges_both_models}.  The correction factors
1305: $D^2_\theta, D^2_H$ are calculated explicitly in Appendix
1306: \ref{sec:deriv D-terms} in terms of the gauge couplings and soft
1307: masses as\footnote{ $q_H$ is defined to be $-q_{R3}$, thus $q_H = \frac{5}{6}$
1308: for Model 1 and $q_H = 1$ for Model 2. See Appendix~\ref{sec:deriv D-terms} for
1309: more details.}
1310: \begin{eqnarray}
1311: \label{eq:1}
1312: D^2_H &=& \frac{1}{4 g^2_{2R} + 6 g^2_4} \left[ m^2_H -
1313: m^2_{\overline{H}} + q_H(m^2_\theta - m^2_{\overline{\theta}})\right] \\
1314: \label{eq:2}
1315: D^2_\theta &=& \frac{m^2_\theta - m^2_{\overline{\theta}}}{2 g_F^2}
1316: \end{eqnarray}
1317: 
1318: We note that the factors of $g^2_F$ appearing in the mass matrices are
1319: cancelled by the $\frac{1}{g^2_F}$ in the definition of $D^2_\theta$.
1320: 
1321: The D-terms associated with the family symmetry depend on the charges
1322: of the left-handed and right-handed matter representations
1323: $F,\overline{F}$ under the family symmetry. It is well
1324: known\footnote{For an explanation, see for example \cite{Kane:2005va}.} that
1325: for Pati-Salam, one can choose any set of charges, and there will be
1326: an equivalent, shifted set of charges that are anomaly free due to
1327: the Green-Schwartz anomaly cancellation mechanism. The charges used for
1328: the D-term calculation should be the anomaly free charges.
1329: 
1330: 
1331: The gauge couplings and mass parameters in
1332: Eqs.~(\ref{eq:1}~,~\ref{eq:2}) are predicted from the model, in terms
1333: of the $X$ parameters and $m_{3/2}$ as shown in
1334: Eqs.~(\ref{eq:ps_H_at_mx}~--~\ref{eq:FN soft mass}) and
1335: Eqs.~(\ref{eq:C^5_1 _1 D^2_theta}~--~\ref{eq:C^5_1 _3 D^2_theta}).
1336: Note that the D-terms will be zero if $X_S = X_{T_i}$, or if the
1337: $\overline\theta$ brane assignment is the same as $\theta$. Choosing
1338: the second of these conditions is useful since it gives a comparison
1339: case where there are no $U(1)_F$ D-terms; this comparison will make
1340: the D-term contribution to flavour violation immediately apparent.
1341: 
1342: %%% ----------------------------------------------------------------------
1343: 
1344: \subsection{Magnitude of $D_{\theta}$-terms for different
1345: $\overline{\theta}$ assignments}
1346: \label{sec:theta_bar D-terms}
1347: 
1348: %%% ----------------------------------------------------------------------
1349: 
1350: The main point worth emphasising is that in the string model
1351: the magnitudes of the $D$-terms are {\em calculable}.
1352: We have assumed throughout that the FN field $\theta$ is an
1353: intersection string state $C^{5_15_2}$, but have not specified the
1354: string assignment of $\overline{\theta}$.
1355: Thus $m^2_{\overline{\theta}}$ takes various values depending on the
1356: string assignment for $\overline{\theta}$.
1357: 
1358: From Eq.~(\ref{eq:2}), we see that we have calculable D-terms,
1359: \begin{equation}
1360: \label{eq:D^2_theta goes like}
1361: 2g^2_F D^2_\theta = m^2_\theta -m^2_{\overline{\theta}} \; ,
1362: \end{equation}
1363: so the value of $D^2_\theta$ depends on the choice of where the $\overline{\theta}$
1364: field lives. We use Table~\ref{tab:particle_content_42241} and
1365: Eqs.~(\ref{eq:scalars_left} - \ref{eq:SUGRA_b_R}) to quantify the
1366: $D_\theta$-term for each possible $\overline{\theta}$ string assignment.
1367: As $\theta$ always lives at the intersection, on $C^{5_1 5_2}$, our first choice of
1368: $\overline{\theta}$ on $C^{5_1 5_2}$ is trivial: it gives $D^2_\theta =0$.
1369: For $\overline{\theta}$ on $C^{5_1}_1$, $m^2_{\overline{\theta}}$ is equivalent to
1370: $m^2_h$, as this is also on $C^{5_1}_1$. So using Eq.~(\ref{eq:mssm_higgs_at_mx})
1371: for $m^2_{\overline{\theta}}$ and Eq.~(\ref{eq:FN soft mass}) for $m^2_\theta$
1372: in Eq.~(\ref{eq:D^2_theta goes like}), we have
1373: \begin{eqnarray}
1374: \label{eq:C^5_1 _1 D^2_theta}
1375: C^{5_1}_1 &:& 2g^2_F D^2_\theta = \frac{3}{2} m^2_{3/2} (X^2_S - X^2_{T_3}) .
1376: \end{eqnarray}
1377: Similarly, the other two choices yield
1378: \begin{eqnarray}
1379: \label{eq:C^5_1 _2 D^2_theta}
1380: C^{5_1}_2 &:& 2g^2_F D^2_\theta = -\frac{3}{2} m^2_{3/2} (X^2_S - X^2_{T_3}) \\
1381: \label{eq:C^5_1 _3 D^2_theta}
1382: C^{5_1}_3 &:& 2g^2_F D^2_\theta = -\frac{3}{2} m^2_{3/2} (X^2_S - X^2_{T_2}) .
1383: \end{eqnarray}
1384: 
1385: In this paper, all models use $X^2_{T_3} = X^2_{T_2} = X^2_{T_1} = X^2_T$, so
1386: Eqs.~(\ref{eq:C^5_1 _2 D^2_theta}) and (\ref{eq:C^5_1 _3 D^2_theta}) are
1387: equal to each other, and opposite in sign to Eq.~(\ref{eq:C^5_1 _1 D^2_theta}).
1388: 
1389: %%% ----------------------------------------------------------------------
1390: 
1391: 
1392: 
1393: \subsection{Soft gaugino masses}
1394: \label{sec:param-gaug}
1395: 
1396: %%% ----------------------------------------------------------------------
1397: 
1398: The soft gaugino masses are the same as in
1399: \cite{Everett:2002pm}, which we quote here for completeness.
1400: The results follow from
1401: Eq.~(\ref{eq:normalised_gauginos}) applied to the \psgroup gauginos, which
1402: then mix into the \smgroup gauginos whose masses are given by
1403: \begin{eqnarray}
1404:   \label{eq:su3_gaugino_gut_scale}
1405:   M_3 &=& \frac{\sqrt{3} m_{3/2} } {\left(T_1 + \overline{T}_1\right)
1406:   + \left(T_2 + \overline{T}_2\right) } \left[ (T_1 + \overline{T}_1)
1407:   X_{T_1} + (T_2 + \overline{T}_2 ) X_{T_2} \right] \\
1408:  \label{eq:su2_gaugino_gut_scale}
1409:   M_2 &=& \sqrt{3} m_{3/2} X_{T_1} \\
1410:   \label{eq:u1_gaugino_gut_scale}
1411:   M_1 &=& \frac{\sqrt{3} m_{3/2} } {\frac{5}{3} (T_1 + \overline{T}_1
1412:   ) + \frac{2}{3} (T_2 + \overline{T}_2 ) } \left[ \frac{5}{3}(T_1 +
1413:   \overline{T}_1) X_{T_1} + \frac{2}{3}(T_2 + \overline{T}_2 ) X_{T_2}
1414:   \right]
1415: \end{eqnarray}
1416: The values of $T_1 + \overline{T}_1$ and
1417: $T_2 + \overline{T}_2$ are proportional to the brane
1418: gauge couplings $g_{5_1}$ and $g_{5_2}$, which are related in a simple
1419: way to the MSSM couplings at the unification scale. This is discussed
1420: in \cite{Everett:2002pm}.
1421: 
1422: When we run the MSSM gauge couplings up and solve for $g_{5_1}$ and
1423: $g_{5_2}$ we find that approximate gauge coupling unification is 
1424: achieved by $T_1 +\overline{T_1} \gg T_2 +
1425: \overline{T}_2$. Then we find the simple approximate result
1426: \begin{equation}
1427:   \label{eq:63}
1428:   M_1 \approx M_3 \approx M_2 = \sqrt{3}m_{3/2} X_{T_1}.
1429: \end{equation}
1430: 
1431: %%% ----------------------------------------------------------------------
1432: 
1433: \subsection{Soft trilinear couplings}
1434: \label{sec:param_tril}
1435: 
1436: %%% ----------------------------------------------------------------------
1437: 
1438: So far we have considered the soft masses and the gaugino masses. The gaugino
1439: masses are the same as in \cite{Everett:2002pm}. The soft masses have had
1440: both D-term contributions added onto the base values from \cite{Everett:2002pm}.
1441: The contributions to the soft masses and gaugino masses from the FN and heavy Higgs
1442: auxiliary fields is completely negligible due to the small size of their F-terms.
1443: However for the soft trilinear masses 
1444: these contributions are of order $\mathcal{O}(m_{3/2})$ despite 
1445: having small F-terms, so FN and Higgs contributions will
1446: give very important additional contributions beyond those
1447: considered in \cite{Everett:2002pm}.
1448: 
1449: From Section~\ref{sec:soft-from-SUGRA} we see that
1450: the canonically normalised equation for the trilinear is
1451: \begin{equation}
1452:   \label{eq:normalised_trilinear0}
1453:   A_{abc} = F_I 
1454:   \left[
1455:     \overline{K}_I - \partial_I \ln
1456:     \left(
1457:       \tilde K_a \tilde K_b \tilde K_c
1458:     \right)
1459:   \right]
1460:   + F_m \partial_m \ln Y_{abc}
1461: \end{equation}
1462: This general form for the
1463: trilinear accounts for contributions from non-moduli F-terms. These
1464: contributions are in general expected to be of the same magnitude as
1465: the moduli contributions despite the fact that the non-moduli F-terms
1466: are much smaller \cite{Abel:2001cv}. Specifically,
1467: if the Yukawa hierarchy is taken to be generated by a FN
1468: field, $\theta$ such that $Y_{ij} \sim \theta^{p_{ij}}$, 
1469: then we expect $F_\theta \sim m_{3/2} \theta$, 
1470: and then $\Delta A_{ij}=F_\theta \partial_\theta \ln Y_{ij} \sim
1471: p_{ij}m_{3/2}$ and so even though these fields are expected to have heavily
1472: sub-dominant F-terms\footnote{In our model the FN and heavy Higgs vevs are 
1473: of order the unification scale, compared to the moduli vevs
1474: which are of order of the Planck scale.}
1475: they contribute to the trilinears at the same
1476: order $\mathcal{O}(m_{3/2})$ as the moduli, but in a flavour off-diagonal way.
1477: 
1478: Here we sum over $m$,
1479: which contains all the hidden sector fields: $S, T_i, H, \overline{H}, \theta$.
1480: 
1481: In the specific D-brane model of interest here, the general results for
1482: soft trilinear masses, including the contributions for general effective Yukawa
1483: couplings are given in Appendix \ref{sec:param-tril-42241}. 
1484: From Eqs.~(\ref{eq:dirac_operator_n_is_n}, \ref{eq:dirac_mass_tensor})
1485: we can read off the effective Yukawa couplings,
1486: \begin{equation}
1487:   \label{eq:dirac_yukawa}
1488:   Y_{h F \overline{F} } h F \overline{F} \equiv
1489:   \underbrace{(c)^{\beta\gamma}_{\alpha\rho}(r)^{xz}_{yw}\overline{H}_{\gamma
1490:   z} H^{\rho w}\theta^{p}}_{ {Y_{hF\overline{F}}}^{\beta
1491:   x}_{\alpha y}}h^y_a F^{\alpha a} \overline{F}_{\beta x}.
1492: \end{equation}
1493: Note the extra group indices that the effective Yukawa coupling
1494: ${Y_{hF\overline{F}}}^{\beta x}_{\alpha y}$ has, and 
1495: proper care must be taken of the tensor structure
1496: when deriving trilinears from a given operator.
1497: We can write down the
1498: trilinear soft masses, $A$, by substituting the operators in
1499: Eq.~(\ref{eq:operator_texture1}) into the results in 
1500: Appendix \ref{sec:param-tril-42241}.
1501: Having done this we find the result:\footnote{We assume that a $(3,3)$ Yukawa
1502: coupling appears at renormalisable order. This is why the $A_{33}$ doesn't include
1503: contributions from $d_H$ and $d_\theta$.} 
1504: \begin{equation}
1505:   \label{eq:trilinear}
1506:    A = \sqrt{3} m_{3/2}
1507:    \left[
1508:     \begin{array}{ccc}
1509:       d_1 + d_H + p(i,j)d_\theta & d_1 + d_H + p(i,j)d_\theta
1510:       & d_2 + d_H + p(i,j)d_\theta \\
1511:       d_1 + d_H + p(i,j)d_\theta & d_1 + d_H + p(i,j)d_\theta
1512:       & d_2 + d_H + p(i,j)d_\theta \\
1513:       d_3 + d_H + p(i,j)d_\theta & d_3 + d_H + p(i,j)d_\theta & d_4
1514:     \end{array}
1515:   \right]  
1516: \end{equation}
1517: where
1518: \begin{eqnarray}
1519:     d_1 &=& X_S - X_{T_1} - X_{T_2} \\ d_2 &=& \frac{1}{2} X_S -
1520:   X_{T_1} - \frac{1}{2}X_{T_2} \\ d_3 &=& \frac{1}{2} X_S - X_{T_1} -
1521:   X_{T_2} + \frac{1}{2}X_{T_3} \\ d_4 &=& -X_{T_1} \\ d_H &=&
1522:   (S+\overline{S})^\frac{1}{2} X_H + (T_3 +
1523:   \overline{T}_3)^\frac{1}{2} X_{\overline{H}} \\ d_\theta &=&
1524:   (S+\overline{S})^\frac{1}{4} (T_3 + \overline{T}_3)^\frac{1}{4}
1525:   X_\theta
1526: \end{eqnarray}
1527: 
1528: These results are independent of which string assignment we give to
1529: $\overline\theta$, since this field does not enter the Yukawa operators.
1530: 
1531: %%% ----------------------------------------------------------------------
1532: 
1533: \section{Results}
1534: \label{sec:numerical-etc}
1535: 
1536: %%% ----------------------------------------------------------------------
1537: 
1538: \subsection{Two models}
1539: \label{sec:models}
1540: 
1541: We will study two different models which have the same Yukawa
1542: textures, but different $U(1)_F$ charge structures. 
1543: The anomaly-free \cite{Kane:2005va}
1544: family charges are laid out in Table~\ref{tab:charges_both_models}.
1545: This first model is essentially the model studied in
1546: \cite{Blazek:2003wz,KP0307091}, but with an extra operator in the
1547: $(1,2)$ and $(1,3)$ Yukawa matrix element to allow a non-zero $Y^e_{12}$ 
1548: and $Y^e_{13}$. 
1549: The second model is defined such that all of the
1550: charges of left-handed matter are the same, causing the $U(1)_F$
1551: D-term to left handed scalar mass matrices to not lead to extra
1552: flavour violation. This choice is a small change to the model, since
1553: two of the `left-handed' charges are the same anyway, and is made
1554: because normally the left-handed contribution dominates over the
1555: right-handed contribution. This can be
1556: achieved by changing the order of the operators in the Yukawa textures
1557: and the arbritary couplings $a,a', ...$ to compensate the change in
1558: the charge structure, as discussed in Appendix~\ref{operatorsformodels}.
1559: Most of the results presented will be for Model 1. 
1560: 
1561: Model 2 differs from Model 1 in the charges under the Abelian
1562: family symmetry, and the compensating changes to ensure the same
1563: Yukawa textues. As discussed in Appendix~\ref{operatorsformodels}
1564: there are two expansion parameters in these models which 
1565: we take to be equal $\epsilon = \delta = 0.22$. The powers
1566: of $\epsilon$ in the first row of the Yukawa matrices
1567: are one lower, and we compensate that by
1568: increasing the powers of $\delta$ in the first row. We do not have to
1569: shift the values of the $a,a',a'',a'''$ parameters since $\delta =
1570: \epsilon$. Were this not the case, we would have to shift the values
1571: by a factor of $\frac{\epsilon}{\delta}$.  Model 2 is not meant to be
1572: a natural or realistic model, we use it as a tool to investigate the
1573: contribution to flavour violation from the $U(1)_F$ D-term correction
1574: to the left-handed scalar masses. The family charges are laid out in
1575: Table~\ref{tab:charges_both_models}.
1576: 
1577: The $U(1)_F$ charge of $H$ must be equal and opposite to
1578: $\overline{F}_3$, and $\overline{H}$ must be the negative of
1579: this. This is due to the (3,3) element of the right-handed Majorana
1580: mass being allowed at leading order, so the $U(1)_F$ charges of
1581: $\overline{F}_3$ and $H$ must conspire to cancel for the operator of
1582: the Majorana fermions to be renormalisable.
1583: For Model 1, the $U(1)_F$ charges of $H$ and $\overline{H}$ are $5/6$
1584: and $-5/6$ respectively. 
1585: For Model 2, the $U(1)_F$ charges of $H$ and $\overline{H}$ are $1$
1586: and $-1$ respectively. One can use the relevent equations above to
1587: check that the anomaly coefficients do indeed satisfy the anomaly
1588: cancellation conditions.
1589: This gives us a different $D_H$-term for Model 2, as indicated by $q_H$ in
1590: Eq.~(\ref{eq:1}) being different for Model 1 and Model 2. Note that
1591: $D^2_\theta$ is the same in both models.
1592: 
1593: 
1594: \begin{table}[htbp]
1595:   \centering
1596:   \begin{tabular}{|c|c|c|}
1597:     \hline
1598:     Field  & Model 1 & Model 2 \\ 
1599:     & Charge & Charge  \\
1600:     \hline
1601:     $F_1$ & $\frac{11}{6}$ & $1$ \\
1602:     $F_2$ & $\frac{5}{6}$ & $1$ \\
1603:     $F_3$ & $\frac{5}{6}$ & $1$ \\
1604:     \hline
1605:     $\overline{F}_1$ & $\frac{19}{6}$ & $3$ \\
1606:     $\overline{F}_2$ & $\frac{7}{6}$ & $1$ \\
1607:     $\overline{F}_3$ & $\!\!\!\!-\frac{5}{6}$ & $\!\!\!\!-1$ \\
1608:     \hline
1609:   \end{tabular}
1610:   \caption{The family charges for Model 1 and Model 2.}
1611:   \label{tab:charges_both_models}
1612: \end{table}
1613: 
1614: 
1615: 
1616: The values of the arbritary couplings are laid out in 
1617: Appendix~\ref{operatorsformodels} in
1618: Table~\ref{tab:a_ap_app_for_models}. This gives numerical values for
1619: the Yukawa elements which can be used in either model, with the relevant
1620: values of $Y^e_{12}$ and $Y^e_{13}$ inserted instead of the texture zeros:
1621: \begin{eqnarray}
1622:   \label{eq:numerical_yu_model1}
1623:   Y^{u}(M_X) &=&
1624:   \left[
1625:     \begin{array}{lll}
1626:       2.159\times10^{-06}  &  5.606\times10^{-04}  &  5.090\times10^{-03} \\
1627:       0.000    &    1.105\times10^{-03}  &   0.000    \\
1628:       0.000    &    6.733\times10^{-03}  &  5.841\times10^{-01}    \\
1629:     \end{array}
1630:   \right]
1631:   \\
1632:   \label{eq:numerical_yd_model1}
1633:   Y^{d}(M_X) &=&
1634:   \left[
1635:     \begin{array}{lll}
1636:     \!\!\!\!-1.661\times10^{-04} & \!\!\!\! -5.606\times10^{-04} &   1.018\times10^{-02} \\
1637:      7.683\times10^{-04}  & \!\!\!\!-5.343\times10^{-03}  &  1.216\times10^{-02}\\
1638:     \!\!\!\!-1.769\times10^{-04}   & 3.133\times10^{-02}   & 3.933\times10^{-01}\\
1639:     \end{array}
1640:   \right]
1641:   \\
1642:   \label{eq:numerical_ye_model1}
1643:   Y^{e}(M_X) &=&
1644:   \left[
1645:     \begin{array}{lll}
1646:     \!\!\!\!-1.246\times10^{-04} &    0.000   &      0.000   \\
1647:      1.537\times10^{-03}  &  2.432\times10^{-02}&   \!\!\!\!-3.649\times10^{-02}\\
1648:     \!\!\!\!-1.327\times10^{-04}   & 3.133\times10^{-02} &   5.469\times10^{-01}\\
1649:   \end{array}
1650: \right]
1651:   \\
1652:   \label{eq:numerical_yn_model1}
1653:   Y^{\nu}(M_X) &=&
1654:   \left[
1655:     \begin{array}{lll}
1656:      2.159\times10^{-06}&    1.525\times10^{-03}&    0.000\\
1657:       0.000     &   8.290\times10^{-04} &   3.923\times10^{-01}  \\  
1658:       0.000      &  5.050\times10^{-03}  &  5.469\times10^{-01} \\
1659:     \end{array}
1660:   \right]
1661: \end{eqnarray}
1662: 
1663: The RH Majorana neutrino mass matrix for Models 1 and 2 has the 
1664: numerical values:
1665: \begin{equation}
1666:   \label{eq:numerical_21}
1667:   \frac{M_{RR}(M_X)}{M_{33}} =
1668:   \left[
1669:     \begin{array}{lll}
1670:       3.508\times10^{8}  &       3.686\times10^{9}   &     3.345\times10^{11}\\
1671:       3.686\times10^{9}   &     8.313\times10^{10}&    5.886\times10^{12}\\
1672:      3.345\times10^{11}&    5.886\times10^{12} &   5.795\times10^{14}\\
1673:     \end{array}
1674:   \right]
1675: \end{equation}
1676: 
1677: 
1678: 
1679: 
1680: 
1681: 
1682: %%% ----------------------------------------------------------------------
1683: 
1684: %%% ----------------------------------------------------------------------
1685: 
1686: 
1687: \subsection{Benchmark points}
1688: 
1689: %%% ----------------------------------------------------------------------
1690: 
1691: Since the parameter space for the models is reasonably expansive, and
1692: the intention is to compare different sources of LFV, it is convenient
1693: to consider five benchmark points, as follows.
1694: It should be noted that for all these points, we have taken all $X_{T_i}$
1695: to be the same, $X_{T_i} = X_T$, and also $X_H = X_{\overline{H}}$.
1696: $X_{\overline{\theta}}$ is taken to be zero throughout.
1697: 
1698: \label{sec:four_points}
1699: \begin{table}[htbp]
1700:   \centering
1701:   \begin{tabular}{|c|c|c|c|c|c|c|}
1702:     \hline
1703:     Point & $X_S$ & $X_T$ & $X_H$ & $X_{\overline{H}}$ & $X_\theta$ &
1704:     $X_{\overline{\theta}}$ \\
1705:     \hline
1706:     A & 0.500 & 0.500 & 0.000 & 0.000 & 0.000 & 0.000 \\
1707:     B & 0.535 & 0.488 & 0.000 & 0.000 & 0.000 & 0.000 \\
1708:     C & 0.270 & 0.270 & 0.000 & 0.000 & 0.841 & 0.000 \\
1709:     D & 0.270 & 0.270 & 0.595 & 0.595 & 0.000 & 0.000 \\
1710:     \hline
1711:     \hline
1712:     E & 0.290 & 0.264 & 0.000 & 0.000 & 0.841 & 0.000 \\
1713:     \hline
1714:   \end{tabular}
1715:   \caption{\small 
1716: Values of the X parameters for the five benchmark points, A-E.}
1717:   \label{tab:benchmarks_defined}
1718: \end{table}
1719: 
1720: \begin{itemize}
1721: \item
1722:   Point A is referred to as ``minimum flavour violation''.  At the
1723:   point $X_S=X_T$ the scalar mass matrices $m^2$ are proportional to the
1724:   identity, and the trilinears $\tilde A$ are aligned with the
1725:   Yukawas. Also, if we look back to Eqs.~(\ref{eq:1},\ref{eq:2}),
1726:   (\ref{eq:ps_H_at_mx}) and (\ref{eq:ps_Hbar_at_mx}), for
1727:   $X_S = X_T$, which is the case for point A (and points C and D) we see
1728:   that the value of both D-term contributions is
1729:   zero.  As such, both $m^2$ and $\tilde{A}$ would be diagonal in the SCKM
1730:   basis in the absence of the RH neutrino field, but in the presence
1731:   of the see-saw mechanism off-diagonal elements are present in the
1732: SCKM basis leading to LFV. Since the four relevent
1733:   soft masses are degenerate at this point, 
1734:   $m^2_{C^{5_1 5_2}} = m^2_{C^{5_1}_1} = m^2_{C^{5_1}_2} = m^2_{C^{5_1}_3}$, both
1735:   D-term contributions are zero; $D^2_\theta = 0 = D^2_H$.
1736:   
1737: \item
1738:   Point B is referred to as ``SUGRA''. With $X_S\neq X_T$ it
1739:   represents typical flavour violation from the moduli fields,
1740:   manifested as diagonal soft mass matrices not proportional to the 
1741: unit matrix in the theory basis; this is
1742:   the amount of flavour violation that would traditionally have been
1743:   expected with no contribution from the $F_H$ or $F_\theta$ fields.
1744:   This and point E are the only benchmark points investigated where
1745:   $D_\theta \neq 0$.
1746:   
1747: \item
1748:   Point C is referred to as ``FN flavour violation''.  It
1749:   represents flavour violation from the Froggatt-Nielsen sector by
1750:   itself, arising as a non-alignment of the trilinear soft terms
1751: via the FN fields in the Yukawa operators,
1752: without any contribution to flavour violation from traditional
1753:   SUGRA effects, since $X_S=X_T$ as in point A.
1754:   
1755: \item
1756:   Point D is referred to as ``Heavy Higgs flavour violation''.
1757:   It represents flavour violation from the heavy Higgs sector, 
1758: arising as a non-alignment of the trilinear soft terms, 
1759: via the Heavy Higgs fields in the Yukawa operators,
1760: without
1761:   any contribution from either traditional SUGRA effects since
1762:   $X_S=X_T$, or from FN fields since $F_\theta=0$.
1763: 
1764: \item
1765:   Point E combines features of points B and C, resulting in
1766:   Froggatt-Nielsen flavour violation from $X_\theta$, with SUGRA flavour
1767:   violation from $X_S\neq X_T$. This is the only point where we see the
1768:   flavour violation from the Froggatt-Nielsen fields and the $U(1)$ D-terms
1769:   appearing at the same time. The numerical values for this point were
1770:   obtained by taking the ratio from $X_S$ and $X_T$ for benchmark point B
1771:   and applying it to benchmark point C.
1772: 
1773: \end{itemize}
1774: 
1775: %%% ----------------------------------------------------------------------
1776: 
1777: \subsection{Varying $Y^e_{12}$ and $Y^e_{13}$}
1778: \label{sec:varying-ye_12}
1779: 
1780: %%% ----------------------------------------------------------------------
1781: 
1782: Normally, the chargino contribution to LFV dominates. Since the Feynman
1783: diagram for this includes the left-handed sfermions, we would expect
1784: the D-term corrections to the left-handed slepton mass matrix to
1785: dominate the flavour violation. However, Model 2, as defined in
1786: Section~\ref{sec:model-2}, is set up to have universal left-handed
1787: charges, so the D-term correction from the breaking of $U(1)_F$ will
1788: not contribute to flavour violation (except that it will either add or
1789: remove some mass suppression). The D-term is limited in magnitude by
1790: the difference of $m^2_\theta$ and $m^2_{\overline{\theta}}$, and although this
1791: is not a strong correction to the soft masses, it can contribute
1792: significantly to the lepton flavour violating branching ratios.
1793: 
1794: The difference in $\mu\rightarrow e\gamma$ between Model 1 and Model
1795: 2 is negligible for $Y^e_{12} = Y^e_{13}=0$. 
1796: This should not be surprising, since
1797: the texture zero coming from $Y^e_{12} = Y^e_{13}=0$ 
1798: will yield small mixing angles,
1799: resulting in small lepton flavour violation.
1800: 
1801: In order to get a picture of how great an effect the D-term
1802: contributions could have on the soft masses, it is
1803: necessary to examine a range of different values of
1804: $Y^e_{12}$ and $Y^e_{13}$. An extra operator contribution was added to the
1805: textures in Model 1 and Model 2, when compared to the model previously
1806: studied \cite{KP0307091}, to allow for variations of the order
1807: $Y^e_{12} \approx 10^{-3}$, for example.
1808: The gives $\mathcal{O}(1)$ parameters $a''_{12}$
1809: and $a'''_{12}$ for Models 1 and 2 respectively. To be precise $Y^e_{12} =
1810: 1.5\times10^{-3}$ corresponds to $a''_{12}$ and $a'''_{12} = 3.2$, and 
1811: $Y^e_{13} = 1.5\times10^{-2}$ corresponds to $a_{13}=0.320$.
1812: 
1813: %%% ----------------------------------------------------------------------
1814: 
1815: \subsection{Varying brane assignments for $\overline\theta$}
1816: \label{sec:thetabar_brane_assignments}
1817: 
1818: %%% ----------------------------------------------------------------------
1819: 
1820: The $\theta$ field is fixed to reside on the $C^{5_1 5_2}$ brane, but we
1821: allow the brane assignment of $\overline\theta$ to
1822: vary over the possibilities $C^{5_1 5_2}$, $C^{5_1}_1$, $C^{5_1}_2$
1823: and $C^{5_1}_3$.  
1824: This gives us D-terms that are calculable in each case, rather than being
1825: free parameters.
1826: The assignment of $\overline{\theta}$ to $C^{5_1 5_2}$,
1827: which is the same assigmment as the $\theta$ field, implies that
1828: the $U(1)_F$ D-term calculated in this case is zero.
1829: The other possibilities, given in Eqs.~(\ref{eq:C^5_1 _1 D^2_theta} -
1830: \ref{eq:C^5_1 _3 D^2_theta}), will highlight the contribution of the D-terms
1831: to lepton flavour violation.
1832: 
1833: 
1834: %%% ----------------------------------------------------------------------
1835: 
1836: \subsection{Numerical procedure}
1837: 
1838: %%% ----------------------------------------------------------------------
1839: 
1840: The code used to generate all the data here was based on SOFTSUSY
1841: \cite{Allanach:2001kg}, which is a program that accurately calculates the spectrum
1842: of superparticles in the MSSM. It solves the renormalisation group equations
1843: with theoretical constraints on soft supersymmetry breaking terms provided
1844: by the user. Successful radiative electroweak symmetry breaking is used as
1845: a boundary condition, as are weak-scale
1846: gauge coupling and fermion mass data (including one-loop finite MSSM
1847: corrections). The program can also calculate a measure of fine-tuning. The 
1848: program structure has, in this case, been adapted to the extension of the
1849: MSSM considered in this paper. It is modified to include right-handed
1850: neutrino fields, and thus non-zero neutrino masses and mixing angles, 
1851: generated via the SUSY see-saw mechanism. It is also set up to include the
1852: new D-term contributions considered herein, and to run
1853: over a series of string assignments for the $\overline{\theta}$ field. We use
1854:  $\tan \beta = 50$.
1855: 
1856: Electroweak symmetry breaking provides a significant constraint on the 
1857: results. The breakdown of electroweak symmetry breaking was
1858: responsible for the `spike' feature that was shown in the plots for 
1859: benchmark points A and B in \cite{KP0307091}. 
1860: For the data above the spike,
1861: radiative electroweak symmetry breaking does not work properly, as
1862: the Z-boson mass becomes tachyonic. In the present paper
1863: such `bad' regions where electroweak symmetry breaking fails are
1864: cut-off, however there is still a remnant of
1865: the spike left, which is why one can see a slight rise at the ends of the plots
1866: for our benchmark points A and B, as can be seen in
1867: Section~\ref{sec:numerical results}.
1868: 
1869: 
1870: %%% ----------------------------------------------------------------------
1871: 
1872: \subsection{Numerical results}
1873: \label{sec:numerical results}
1874: 
1875: %%% ----------------------------------------------------------------------
1876: 
1877: We have now defined our two models, Model 1 and Model 2, and a set of
1878: five benchmark points in Table~\ref{tab:benchmarks_defined} to examine
1879: within them.  We have also set up what we will be varying apart from the
1880: gravitino mass in these models -- the values of $Y^e_{12}$, 
1881: $Y^e_{13}$ and the string 
1882: assignment of $\overline\theta$ which gives different D-terms. We are now
1883: in a position to present our results.
1884: We shall focus on the branching ratios for $\mu \to e \gamma$ and
1885: $\tau \to \mu \gamma$.
1886: The branching ratio for $\tau\to e\gamma$ is not shown here as it does
1887: not constrain us beyond those limits placed by $\mu \to e \gamma$ and
1888: $\tau \to \mu \gamma$. The experimental limit for $\tau\to e\gamma$,
1889: at $2.7\times10^{-6}$, is in fact far above the predicted rate for
1890: this process at all examined parts of the parameter space.
1891: In the following plots, we do not consider the $\overline{\theta}$
1892: assignment to $C^{5_1}_3$ as this is exactly the same as $C^{5_1}_2$,
1893: due to the degeneracy of the $X_{T_i}$. Were we to allow the $X_{T_i}$
1894: to be non-degenerate, the phenomenological results of assigning
1895: $\overline{\theta}$ to $C^{5_1}_2$ and $C^{5_1}_3$ would not be the
1896: same. The detailed spectrum
1897: will look different at each parameter point,
1898: but the general trend is for the
1899: physical masses to increase in
1900: magnitude as the gravitino mass increases. 
1901: Thus too-high gravitino masses
1902: will start to reintroduce the fine-tuning problem resulting from the gluino
1903: mass being too high \cite{Kane:1998im}, although we shall not
1904: discuss the detailed spectrum here.
1905: 
1906: 
1907: %\begin{figure}[htbp]
1908: %\input{m32_meg_B_model1and2}
1909: %\centering
1910: %\caption{\small{Plots showing the Branching Ratio for $\mu \to e \gamma$ for
1911: %benchmark point B. Panel (i) corresponds to Model 1, with $Y^e_{12} = 0$.
1912: %Panel (ii) corresponds to Model 1, with $Y^e_{12} = 1.5\times10^{-3}$.
1913: %Panel (iii) corresponds to Model 2, with $Y^e_{12} = 0$.
1914: %Panel (iv) corresponds to Model 2, with $Y^e_{12} = 1.5\times10^{-3}$.
1915: %The $\overline{\theta}$ assignments are shown with the separate lines:
1916: %$C^{5_1 5_2}$ ({\em solid}), $C^{5_1}_1$ ({\em dashed}), and $C^{5_1}_2$
1917: %({\em dot-dash}).
1918: %The 2002 experimental limit \cite{Hagiwara:fs} is also given by the horizontal line.}}
1919: %\label{fig:m32_meg_B_model1and2}
1920: %\end{figure}
1921: 
1922: 
1923: 
1924: 
1925: 
1926: 
1927: %\begin{figure}[htbp]
1928: %\input{m32_meg_B_model2_Y}
1929: %\centering
1930: %\caption{\small{Plots showing the Branching Ratio for $\mu \to e \gamma$ for
1931: %benchmark point B for Model 2 only.
1932: %Panel (i) has $Y^e_{12} = 0$ and $Y^e_{13} = 0$.
1933: %Panel (ii) has $Y^e_{12} = 1.5\times10^{-3}$ and $Y^e_{13} = 0$.
1934: %Panel (iii) has $Y^e_{12} = 0$ and $Y^e_{13} = 1.5\times10^{-2}$.
1935: %Panel (iv) has $Y^e_{12} = 1.5\times10^{-3}$ and $Y^e_{13} = 1.5\times10^{-2}$.
1936: %The $\overline{\theta}$ assignments are shown with the separate lines:
1937: %$C^{5_1 5_2}$ ({\em solid}), $C^{5_1}_1$ ({\em dashed}), and $C^{5_1}_2$
1938: %({\em dot-dash}).
1939: %The 2002 experimental limit \cite{Hagiwara:fs} is also given by the horizontal line.}}
1940: %\label{fig:m32_meg_B_model2_Y}
1941: %\end{figure}
1942: 
1943: %\begin{figure}[htbp]
1944: %\input{m32_meg_E_model1and2}
1945: %\centering
1946: %\caption{\small{Plots showing the Branching Ratio for $\mu \to e \gamma$ for
1947: %benchmark point E. Panel (i) corresponds to Model 1, with $Y^e_{12} = 0$.
1948: %Panel (ii) corresponds to Model 1, with $Y^e_{12} = 1.5\times10^{-3}$.
1949: %Panel (iii) corresponds to Model 2, with $Y^e_{12} = 0$.
1950: %Panel (iv) corresponds to Model 2, with $Y^e_{12} = 1.5\times10^{-3}$.
1951: %The $\overline{\theta}$ assignments are shown with the separate lines:
1952: %$C^{5_1 5_2}$ ({\em solid}), $C^{5_1}_1$ ({\em dashed}), and $C^{5_1}_2$
1953: %({\em dot-dash}).
1954: %The 2002 experimental limit \cite{Hagiwara:fs} is also given by the horizontal line.}}
1955: %\label{fig:m32_meg_E_model1and2}
1956: %\end{figure}
1957: 
1958: Figure~\ref{fig:m32_meg_ABCD_model1} shows numerical results for
1959: $\mathrm{BR}(\mu\to e\gamma)$ for Model 1, plotted against the
1960: gravitino mass $m_{3/2}$, where each of the four panels (i) --
1961: (iv) correspond to each of the four benchmark points A -- D. 
1962: \footnote{The solid line on each plot in
1963: Figure~\ref{fig:m32_meg_ABCD_model1} corresponds to the solid
1964: lines in Figure~1 of \cite{KP0307091}. However there were errors
1965: in the code used to generate the previous data, and the corrected
1966: rates shown here differ to the previous results by up to two
1967: orders of magnitude. Furthermore, unlike the results in
1968: \cite{KP0307091}, the results here do not exhibit a sharp spike
1969: for benchmark points A and B. For the data above the spike,
1970: radiative electroweak symmetry breaking does not work properly, as
1971: the Z-boson mass becomes tachyonic. This was not realized in the
1972: previous analysis.}
1973: Panel (i) of Figure~\ref{fig:m32_meg_ABCD_model1} refers to
1974: benchmark point A,
1975: corresponding to minimum flavour violation, where the only source of
1976: LFV is from the see-saw mechanism, which for Model 1 is well below
1977: the experimental limit, shown as the horizontal dot-dash line.
1978: Panel (ii) of
1979: Figure~\ref{fig:m32_meg_ABCD_model1} refers to benchmark point B, 
1980: with LFV arising from SUGRA, with the FN and heavy Higgs sources of LFV
1981: switched off. In this case one can clearly distinguish the
1982: additional contributions to LFV arising from the D-terms. This
1983: makes benchmark point B the most phenomenologically interesting
1984: for the purposes of this study. The differing contributions stem
1985: from the $\overline{\theta}$ string assignments, which are shown
1986: by the separate lines: $C^{5_1 5_2}$ ({\em solid}), $C^{5_1}_1$
1987: ({\em dashed}), and $C^{5_1}_2$ ({\em dot-dash}). The $C^{5_1
1988: 5_2}$ case shows the zero-D-term limit, where $\theta$ and
1989: $\overline{\theta}$ are both intersection states, and hence
1990: conspire to cancel out $D_\theta$ via their soft masses being
1991: degenerate, $m^2_\theta - m^2_{\overline{\theta}} = 0$. The other
1992: two locations for $\overline{\theta}$ then turn on the
1993: $D_\theta$-term contributions. With the D-terms switched on, Model
1994: 1 is experimentally ruled out over all parameter space shown here.
1995: Panel (iii) of
1996: Figure~\ref{fig:m32_meg_ABCD_model1}, refering to benchmark point C, is
1997: the Froggatt-Nielsen benchmark point, and for this case we see
1998: that the experimental limit is satisfied for $m_{3/2}$ over
1999: 1400~GeV. Panel (iv) of
2000: Figure~\ref{fig:m32_meg_ABCD_model1}, benchmark point D, shows
2001: the heavy Higgs point, for which the experimental limit is
2002: satisfied everywhere above 800~GeV. 
2003: \footnote{Note that the predictions in this figure are
2004: lower than the corresponding figure in \cite{KP0307091} due to the corrected 
2005: values of $S,T_3$, as discussed.}
2006: 
2007: 
2008: \begin{figure}[htbp]
2009: \input{m32_meg_ABCD_model1} \centering
2010: \caption{\small{Plots showing the Branching Ratio for $\mu \to e
2011: \gamma$ for Model 1 with $Y^e_{12} = Y^e_{13} = 0$. Panel (i)
2012: corresponds to benchmark point A, panel (ii) is for B, panel (iii) is
2013: for C and panel (iv) is for D.  The $\overline{\theta}$ assignments
2014: are shown with the separate lines: $C^{5_1 5_2}$ ({\em solid}),
2015: $C^{5_1}_1$ ({\em dashed}), and $C^{5_1}_2$ ({\em dot-dash}). 
2016: The solid curve corresponds to zero D-terms, and the other 
2017: curves correspond to different models for the D-terms. 
2018: The 2002 experimental limit \cite{Hagiwara:fs} is also given by the
2019: horizontal line.}}
2020: \label{fig:m32_meg_ABCD_model1}
2021: \end{figure}
2022: 
2023: 
2024: 
2025: \begin{figure}[htbp]
2026: \input{m32_tmg_ABCD_model1} \centering
2027: \caption{\small{Plots showing the Branching Ratio for $\tau
2028: \to\mu\gamma$ for Model 1 with $Y^e_{12} = Y^e_{13} = 0$. Panel (i)
2029: corresponds to benchmark point A, panel (ii) is for B, panel (iii) is
2030: for C and panel (iv) is for D.  The $\overline{\theta}$ assignments
2031: are shown with the separate lines: $C^{5_1 5_2}$ ({\em solid}),
2032: $C^{5_1}_1$ ({\em dashed}), and $C^{5_1}_2$ ({\em dot-dash}).  
2033: The solid curve corresponds to zero D-terms, and the other 
2034: curves correspond to different models for the D-terms. 
2035: The 2005 experimental limit \cite{Aubert:2005ye} is also given by the
2036: horizontal line.}}
2037: \label{fig:m32_tmg_ABCD_model1}
2038: \end{figure}
2039: 
2040: 
2041: Figure~\ref{fig:m32_tmg_ABCD_model1} shows analogous results for
2042: $\mathrm{BR}(\tau\rightarrow\mu\gamma)$ for Model 1, plotted against
2043: the gravitino mass $m_{3/2}$. All benchmark points come below the
2044: experimental limit for a substantial amount of the parameter
2045: space. The experimental limit here is more recent, and subsequently
2046: much more stringent than the previous limit.  For these models in
2047: which there is a large (2,3) element in the neutrino Yukawa matrix the
2048: branching ratio for $\tau\to\mu\gamma$ is essentially as constraining
2049: as that for $\mu\to e\gamma$, as first pointed out in
2050: \cite{Blazek:2002wq}.  The D-term coupling to right-handed scalars has
2051: a Yukawa mixing angle of order $\lambda^3$, compared to $\lambda^2$
2052: for $\mu\rightarrow e\gamma$.  $\lambda \approx 0.22$ is the
2053: Wolfenstein parameter, which contributes on an equal footing to
2054: $\epsilon$ and $\delta$. So the right-handed sector is of equal
2055: importance to the left-handed sector. We note that the see-saw effect
2056: enters prominently in the left-handed sector, and by considering
2057: Eqs.~(\ref{eq:m^2_QL}) and (\ref{eq:m^2_LL}) for the soft scalar
2058: masses in the $(2-3)$ sector for $\tau \to \mu \gamma$, one can show
2059: that there is little effect coming from the D-terms coupling to
2060: left-handed scalars, since we have universal family charges for the
2061: left-handed $(2-3)$ sector, $q_{L_2} = q_{L_3}$. For the right-handed
2062: scalars, however, the D-terms do play an important part and can have
2063: rather interesting and surprising effects, as we now
2064: discuss in some detail.
2065: 
2066: The solid line in panel (ii) of Figure~\ref{fig:m32_tmg_ABCD_model1}
2067: for the $C^{5_1 5_2}$ string assignment of $\overline{\theta}$
2068: has zero contribution from the $U(1)_F$ D-terms, and shows just the
2069: effect of non-minimal SUGRA. This actually suppresses the flavour
2070: violation arising from the see-saw effect alone, showing an
2071: interesting cancellation between the LFV from the see-saw
2072: mechanism and the LFV from the non-universal D-terms.
2073: On the other hand the dashed line in
2074: panel (ii) for the $C^{5_1}_1$ case is very similar to the see-saw
2075: scenario of benchmark point A shown in panel (i). This is due to the
2076: D-terms actually conspiring to restore universality in the scalar
2077: masses, turning non-minimal SUGRA back into the minimal form. 
2078: One can easily see this by applying Eq.~(\ref{eq:C^5_1 _1 D^2_theta}) to
2079: Eqs.~(\ref{eq:m^2_UR}) - (\ref{eq:m^2_NR}) for the right-handed scalar
2080: mass matrices, as the D-terms bring in an equal but opposite effect to
2081: the non-universal effects from SUGRA, and subsequently force the mass
2082: matrices to become universal. It is an amazing consequence of this
2083: string assignment for $\overline{\theta}$ that in this model the
2084: effects of the non-universal $U(1)_F$ D-terms can exactly cancel the
2085: effects of the non-universal SUGRA for the branching ratio of $\tau
2086: \to \mu \gamma$, leading to universal scalar mass matrices, even with
2087: SUGRA turned on.  This is a string effect that directly affects the
2088: amount of flavour violation predicted in this scenario. For the
2089: $C^{5_1}_2$ case shown by the dot-dash line, applying
2090: Eq.~(\ref{eq:C^5_1 _2 D^2_theta}) to the right-handed scalar mass
2091: matrices Eqs.~(\ref{eq:m^2_UR}) - (\ref{eq:m^2_NR}) shows that the
2092: D-terms in this case actually enhance the effect of non-minimal SUGRA,
2093: causing the scalar mass matrices to become even more non-universal.
2094: One can understand this purely right-handed effect by considering the
2095: different mass insertion diagrams for the left- and right-handed
2096: sectors. The left-handed sector involves charginos, whereas the
2097: right-handed sector only involves neutralinos, so the right-handed
2098: masses scale differently with the gravitino mass as compared to the
2099: left-handed masses. Thus we do not have a universal mass scaling
2100: between the left- and right-handed sectors. This leads to the observed
2101: smooth cancellation of flavour violation between the two competing
2102: contributions from the see-saw mechanism and from SUGRA with
2103: additional $U(1)_F$ D-terms.
2104: 
2105: 
2106: %Figure~\ref{fig:m32_meg_B_model1and2} shows results for $\mathrm{BR}(\mu\to
2107: %e\gamma)$ for Model 1, panels (i) and (ii), and Model 2, panels (iii) and (iv),
2108: %plotted against the gravitino mass $m_{3/2}$, at benchmark
2109: %point B only. Here we also vary the (1,2) element of the electron Yukawa,
2110: %with $Y^e_{12} = 0$, panels (i) and (iii), and
2111: %$Y^e_{12} = 1.5\times10^{-3}$, panels (ii) and (iv). The $\overline{\theta}$
2112: %assignments are again shown with the separate lines: $C^{5_1 5_2}$ ({\em solid}),
2113: %$C^{5_1}_1$ ({\em dashed}), and $C^{5_1}_2$ ({\em dot-dash}).
2114: 
2115: %There is a dramatic increase in the predicted branching ratio as $Y^e_{12}$
2116: %is turned on from zero. We can see this by comparing the solid lines in
2117: %panel (i) with (ii) for
2118: %Model 1, and the solid lines in panel (iii) with (iv) for Model 2, where 
2119: %we go from $Y^e_{12} = 0$
2120: %to $Y^e_{12} = 1.5\times10^{-3}$. The lines of $\overline\theta$ assigned
2121: %to $C^{5_1}_1$ and $C^{5_1}_2$ have a strongly different shape. For them, 
2122: %switching on $Y^e_{12}$ has a small effect; the strong effect coming from
2123: %the right-handed sector D-terms is still present, and even increasing the
2124: %$C^{5_1 5_2}$ rate doesn't change this.
2125: 
2126: %We note that
2127: %the non-zero value for $Y^e_{12}$ was chosen such that the plots would be
2128: %on the very edge of violating the experimental limit; if we turn it down
2129: %slightly, then the predicted rate is reduced to below this limit. For this
2130: %particular choice, the experimental limit is satisfied for a gravitino mass
2131: %between 1300 and 1500 GeV. For $Y^e_{12} = 0$, there is little difference
2132: %between the two models.
2133: 
2134: \begin{figure}[htbp]
2135: \input{m32_meg_B_model1_Y}
2136: \centering
2137: \caption{\small{Plots showing the Branching Ratio for $\mu \to e \gamma$ for
2138: benchmark point B for Model 1 only.
2139: Panel (i) has $Y^e_{12} = 0$ and $Y^e_{13} = 0$.
2140: Panel (ii) has $Y^e_{12} = 1.5\times10^{-3}$ and $Y^e_{13} = 0$.
2141: Panel (iii) has $Y^e_{12} = 0$ and $Y^e_{13} = 1.5\times10^{-2}$.
2142: Panel (iv) has $Y^e_{12} = 1.5\times10^{-3}$ and $Y^e_{13} = 1.5\times10^{-2}$.
2143: The $\overline{\theta}$ assignments are shown with the separate lines:
2144: $C^{5_1 5_2}$ ({\em solid}), $C^{5_1}_1$ ({\em dashed}), and $C^{5_1}_2$
2145: ({\em dot-dash}).
2146: The solid curve corresponds to zero D-terms, and the other 
2147: curves correspond to different models for the D-terms. 
2148: The 2002 experimental limit \cite{Hagiwara:fs} is also given by the horizontal line.}}
2149: \label{fig:m32_meg_B_model1_Y}
2150: \end{figure}
2151: 
2152: 
2153: Figure~\ref{fig:m32_meg_B_model1_Y} shows benchmark point B for Model
2154: 1.  The four panels show the $Y^e_{12}$ and $Y^e_{13}$ electron Yukawa
2155: elements being turned on and
2156: off. The results for 
2157: Figure~\ref{fig:m32_meg_B_model1_Y}(i) are for $Y^e_{12} = 0$ and
2158: $Y^e_{13} = 0$. This is the same as in panel (ii) of
2159: Figure~\ref{fig:m32_meg_ABCD_model1}, and is the base from which we
2160: start.  Panel (ii) of Figure~\ref{fig:m32_meg_B_model1_Y} has $Y^e_{12}
2161: = 1.5\times10^{-3}$ and $Y^e_{13} = 0$, so we can clearly see the
2162: effect of turning $Y^e_{12}$ on. It only affects the $C^{5_1 5_2}$
2163: line, as the D-terms dominate over this effect for the other two
2164: string assignments.  Panel (iii) of
2165: Figure~\ref{fig:m32_meg_B_model1_Y} uses $Y^e_{12} = 0$ and $Y^e_{13}
2166: = 1.5\times10^{-2}$, highlighting the effect of just $Y^e_{13}$ alone.
2167: Again the zero D-term line of $C^{5_1 5_2}$ is the only one that is
2168: sizably affected by this change in Yukawa texture.  Panel (iv) of
2169: Figure~\ref{fig:m32_meg_B_model1_Y} shows the effect of turning on
2170: both Yukawa elements: $Y^e_{12} = 1.5\times10^{-3}$ and $Y^e_{13} =
2171: 1.5\times10^{-2}$.  We see that the shape of the solid line is
2172: determined by both Yukawa textures -- they seem to have an equal
2173: impact on it.
2174: 
2175: 
2176: \begin{figure}[htbp]
2177: \input{m32_meg_E_model1_Y}
2178: \centering
2179: \caption{\small{Plots showing the Branching Ratio for $\mu \to e \gamma$ for
2180: benchmark point E for Model 1 only.
2181: Panel (i) has $Y^e_{12} = 0$ and $Y^e_{13} = 0$.
2182: Panel (ii) has $Y^e_{12} = 1.5\times10^{-3}$ and $Y^e_{13} = 0$.
2183: Panel (iii) has $Y^e_{12} = 0$ and $Y^e_{13} = 1.5\times10^{-2}$.
2184: Panel (iv) has $Y^e_{12} = 1.5\times10^{-3}$ and $Y^e_{13} = 1.5\times10^{-2}$.
2185: The $\overline{\theta}$ assignments are shown with the separate lines:
2186: $C^{5_1 5_2}$ ({\em solid}), $C^{5_1}_1$ ({\em dashed}), and $C^{5_1}_2$
2187: ({\em dot-dash}).
2188: The solid curve corresponds to zero D-terms, and the other 
2189: curves correspond to different models for the D-terms. 
2190: The 2002 experimental limit \cite{Hagiwara:fs} is also given by the horizontal line.}}
2191: \label{fig:m32_meg_E_model1_Y}
2192: \end{figure}
2193: 
2194: 
2195: \begin{figure}[htbp]
2196: \input{m32_meg_E_model2_Y}
2197: \centering
2198: \caption{\small{Plots showing the Branching Ratio for $\mu \to e \gamma$ for
2199: benchmark point E for Model 2 only.
2200: Panel (i) has $Y^e_{12} = 0$ and $Y^e_{13} = 0$.
2201: Panel (ii) has $Y^e_{12} = 1.5\times10^{-3}$ and $Y^e_{13} = 0$.
2202: Panel (iii) has $Y^e_{12} = 0$ and $Y^e_{13} = 1.5\times10^{-2}$.
2203: Panel (iv) has $Y^e_{12} = 1.5\times10^{-3}$ and $Y^e_{13} = 1.5\times10^{-2}$.
2204: The $\overline{\theta}$ assignments are shown with the separate lines:
2205: $C^{5_1 5_2}$ ({\em solid}), $C^{5_1}_1$ ({\em dashed}), and $C^{5_1}_2$
2206: ({\em dot-dash}).
2207: The solid curve corresponds to zero D-terms, and the other 
2208: curves correspond to different models for the D-terms. 
2209: The 2002 experimental limit \cite{Hagiwara:fs} is also given by the horizontal line.}}
2210: \label{fig:m32_meg_E_model2_Y}
2211: \end{figure}
2212: 
2213: 
2214: 
2215: Figure~\ref{fig:m32_meg_E_model1_Y} shows benchmark point E, which combines
2216: the features of benchmark points B and C, thus both $U(1)$ D-term and
2217: Froggatt-Nielsen (FN) flavour violation appear in the predicted branching
2218: ratios shown in this figure. There is some interesting
2219: interplay between the FN F-terms in the benchmark point C region of
2220: parameter space, and the D-terms in benchmark point B, and benchmark point E
2221: is designed to show this difference. In Figure~\ref{fig:m32_meg_E_model1_Y} 
2222: the results are shown for 
2223: Model 1 with both $Y^e_{12}$ and $Y^e_{13}$ Yukawa elements on
2224: and off. The shape of the curves are as in benchmark point C due to $X_\theta$
2225: being turned on, and the results are numerically similar to those of benchmark
2226: point C, showing that the dominant contribution comes from Froggatt-Nielsen
2227: flavour violation. However, as in benchmark point B, the different
2228: D-terms corresponding to different $\overline{\theta}$
2229: assignments leads to noticeable shifts in the results.
2230: In panel (i) with $Y^e_{12} = 0$ and $Y^e_{13} = 0$
2231: it is seen that the presence of non-zero D-terms actually
2232: reduces the LFV rate somewhat compared to the solid curve
2233: with zero D-terms, corresponding to a region of parameter space where there is
2234: some cancellation between the flavour violation from the Froggat-Nielsen
2235: fields and that caused by the $U(1)_F$ D-terms. 
2236: The other panels show variation of the Yukawa elements as in
2237: Figure~\ref{fig:m32_meg_B_model1_Y},
2238: for example panel (iv) corresponds to both 
2239: $Y^e_{12}$ and $Y^e_{13}$ Yukawa elements being non-zero.
2240: Note that the solid curve in panel (iv) 
2241: is slightly lower than the solid curve in panel (ii),
2242: showing that sometimes a non-zero Yukawa coupling can reduce LFV.
2243: 
2244: Figure~\ref{fig:m32_meg_E_model2_Y} shows the effects of the Yukawa
2245: elements for benchmark point E using Model 2, where Model 2
2246: has the same Yukawa structure as Model 1, but has the feature that
2247: the left-handed family charges are the same for all three families,
2248: resulting in universal D-terms, at least in the left-handed sector.
2249: Comparing Figure~\ref{fig:m32_meg_E_model2_Y} to 
2250: Figure~\ref{fig:m32_meg_E_model1_Y}, we see that 
2251: in panel (i) of both figures with $Y^e_{12} = 0$ and $Y^e_{13} = 0$
2252: there is no obserable difference between the predictions 
2253: of the two models. However comparing panels (iv) of
2254: Figure~\ref{fig:m32_meg_E_model2_Y} and Figure~\ref{fig:m32_meg_E_model1_Y}
2255: we see that with non-zero $Y^e_{12}$ and $Y^e_{13}$ Model 2 has
2256: the effect of reducing the LFV resulting from the D-terms.
2257: 
2258: 
2259: We could have presented similar plots for benchmark point D, with a new
2260: point F,\footnote{The values for the $X$ parameters for benchmark
2261: point F would be
2262: \begin{equation} \nonumber
2263: \begin{array}{cccccc}
2264: X_S & X_T & X_H & X_{\overline{H}} & X_\theta & X_{\overline{\theta}} \\
2265: 0.290 & 0.264 & 0.595 & 0.595 & 0.000 & 0.000
2266: \end{array} .
2267: \end{equation}}
2268: which is an amalgamation of points B and D, but since this
2269: produces the same kind of results as Figure~\ref{fig:m32_meg_E_model1_Y},
2270: we did not present the results.
2271: 
2272: 
2273: %%% ----------------------------------------------------------------------
2274: 
2275: \section{Conclusions} 
2276: \label{sec:conclusions} 
2277: 
2278: %%% ----------------------------------------------------------------------
2279: 
2280: We have catalogued and quantitatively studied the importance of
2281: all the different sources of LFV present
2282: in a general non-minimal SUGRA framework, including the 
2283: effects of gauged family symmetry. 
2284: We have discussed five different sources of LFV in such models:
2285: see-saw induced LFV arising from the running effects of right-handed neutrino
2286: fields; supergravity induced LFV
2287: due to the non-universal structure of the supergravity
2288: model; FN (Higgs) flavour violation, due to the F-terms associated with FN 
2289: (Higgs) fields developing \vevs{}, and contributing in a non-universal way 
2290: to the soft trilinear terms; D-term flavour violation, where the D-term mass
2291: correction from the breaking of the Abelian family symmetry drives the
2292: scalar mass matrices to be non-universal; and finally the
2293: effects of different choices of Yukawa textures on LFV.
2294: 
2295: In order to quantify the importance of the different effects
2296: we investigated these disparate souces of LFV numerically, 
2297: both in isolation and
2298: in association with one another, within a particular SUGRA 
2299: model based on a type I string-inspired Pati-Salam model 
2300: with an Abelian family symmetry, which has a sufficiently rich structure
2301: to enable all of the effects to studied within a single framework.
2302: Within this framework
2303: we derived the soft supersymmetry breaking terms, including the effect
2304: of the D-terms associated with breaking the family symmetry. 
2305: For these models the D-terms are calculable, but are
2306: model dependent, depending on a particular choice of string assignment
2307: for the FN fields, and in particular the D-terms are only non-zero
2308: for the non-universal SUGRA models. We have performed
2309: a detailed numerical analysis of the five sources of LFV using five
2310: benchmark points designed to highlight the particular effects,
2311: and we have explored the effect of the variation of Yukawa texture elements 
2312: on the results.
2313: 
2314: The most striking conclusion is how dangerously large the calculable D-term
2315: contribution to flavour violation can be, at least for the class
2316: of models studied. However it should be emphasised that
2317: while the D-terms are calculable in these models they are
2318: also model dependent, and it is always possible to 
2319: simply switch off the D-terms by selecting
2320: the $\overline{\theta}$ field to have the same string assignment
2321: as the ${\theta}$ intersection state. However
2322: other choices will lead to non-zero but calculable D-terms,
2323: which can be dangerously large, or can massively suppress flavour violation.
2324: For example the curves with non-zero D-terms 
2325: in Figure~\ref{fig:m32_meg_B_model1_Y} all exceed the
2326: experimental limit for the Branching Ratio for $\mu \to e \gamma$,
2327: showing that D-term effects have the potential to greatly exceed the
2328: other contributions from SUGRA, the see-saw mechanism, FN and Higgs, depending
2329: on the choice of Yukawa textures.
2330: However in some cases the D-terms generated by breaking the $U(1)_F$ family
2331: symmetry can also suppress the Branching Ratio for $\mu \to e \gamma$,
2332: as shown in panels (i) and (iii) of
2333: Figures~\ref{fig:m32_meg_E_model1_Y} and \ref{fig:m32_meg_E_model2_Y}.
2334: 
2335: Another notable feature is the effect of Yukawa texture on the
2336: results. 
2337: The Yukawa texture has $Y^e_{12} = Y^e_{13} = 0$, and we have shown
2338: that turning on 
2339: non-zero values of these Yukawa couplings can greatly enhance the
2340: branching ratio for $\mu\to e\gamma$ almost arbitarily.
2341: The reason is that the rotations to the SCKM basis are 
2342: controlled by these Yukawa elements and the larger these
2343: rotations the larger will be the off-diagonal soft masses in the SCKM basis.
2344: The non-zero magnitudes of $Y^e_{12}$ and $Y^e_{13}$ 
2345: were therefore chosen to be large enough
2346: to show the variations in the branching ratios of the different
2347: models, but small enough to keep within the currently
2348: experimentally allowed range. 
2349: 
2350: In this paper we have worked with
2351: a particular Yukawa texture in which there is a large (2,3) element in the
2352: neutrino Yukawa matrix leading to large see-saw induced LFV and a 
2353: branching ratio for $\tau\to\mu\gamma$ which is 
2354: as constraining as that for $\mu\to e\gamma$ \cite{Blazek:2002wq}.
2355: However we have seen that in some cases the D-terms can 
2356: lead to a large suppression of the rate for particular 
2357: values of $m_{3/2}$, as seen 
2358: in panel (ii) of Figure~\ref{fig:m32_tmg_ABCD_model1}.
2359: For other cases the effects of the non-universal $U(1)_F$ D-terms 
2360: can exactly cancel the effects of the non-universal SUGRA model 
2361: leading to universal scalar mass matrices, thereby restoring 
2362: universality even for a non-minimal SUGRA model.
2363: Such effects are only possible 
2364: in certain string set-ups and thus LFV provides 
2365: an observable signal which may discriminate between different
2366: underlying string models.
2367: 
2368: In conclusion, we have seen that within realistic non-minimal
2369: supergravity models there can be several important
2370: effects leading to much larger LFV than in the case usually
2371: considered in the literature of minimum flavour violation
2372: corresponding to just mSUGRA and the see-saw mechanism, 
2373: and considered here as benchmark point A. We find that the D-term
2374: contributions are generally dangerously large, but in certain cases
2375: such contributions can lead to a dramatic suppression of LFV rates,
2376: for example by cancelling the effect of the see-saw induced
2377: LFV in $\tau \to \mu \gamma$ models with lop-sided textures.
2378: In the class of string models considered here we find the surprising result
2379: that the D-terms can sometimes serve to restore 
2380: universality in the effective non-minimal supergravity theory.
2381: Thus D-terms can give very large and very
2382: surprising effects in LFV processes.
2383: In general there
2384: will be a panoply of different sources of LFV in realistic
2385: non-minimal SUGRA models, and we have explored the relative
2386: importance of some of them within a particular framework.
2387: The results here only serve to heighten the expectation
2388: that LFV processes such as $\mu \to e \gamma$ and
2389: $\tau\to\mu\gamma$ may be observed soon, although it is
2390: clear from our results that the precise theoretical
2391: interpretation of such signals will be more non-trivial
2392: than is apparent from many previous studies in the literature.
2393: 
2394: 
2395: %%% ----------------------------------------------------------------------
2396: 
2397: \vskip 0.1in
2398: \noindent
2399: {\large {\bf Acknowledgements}}\\
2400: J.H. thanks PPARC for a studentship.
2401: 
2402: 
2403: %%% ----------------------------------------------------------------------
2404: %%% ----------------------------------------------------------------------
2405: 
2406: \newpage
2407: \appendix
2408: 
2409: \numberwithin{equation}{section} % from p426 in my LaTeX book
2410: 
2411: %%% ----------------------------------------------------------------------
2412: %%% ----------------------------------------------------------------------
2413: 
2414: 
2415: %%% ----------------------------------------------------------------------
2416: 
2417: \section{Parameterised trilinears for the 42241 model}
2418: \label{sec:param-tril-42241}
2419: 
2420: %%% ----------------------------------------------------------------------
2421: 
2422: We here write the general form of the trilinear parameters $A_{ijk}$ assuming
2423: nothing about the form of the Yukawa matricies.
2424: \begin{eqnarray}
2425:   \nonumber
2426:   %% F_{1,2} Fbar_{1,2} h 
2427:   A_{C_1^{5_1} C^{5_1 5_2} C^{5_1 5_2}}
2428:   &=&
2429:   \sqrt{3} m_{3/2}
2430:   \left\{
2431:     X_S
2432:       \left[
2433:         1 +
2434:         \left(S + \overline{S}\right)\partial_S \ln Y_{abc} 
2435:       \right]
2436:     \right. 
2437:   \\
2438:   &
2439:   \nonumber
2440:   & 
2441:   {} + X_{T_1}
2442:     \left[
2443:       -1 + \left(T_1 + \overline{T}_1 \right)\partial_{T_1} \ln Y_{abc}
2444:     \right]
2445:   \\
2446:   &&
2447:   \nonumber
2448:   {} + X_{T_2}
2449:   \left[
2450:     -1 + \left(T_2 + \overline{T}_2 \right) \partial_{T_2} \ln Y_{abc}
2451:   \right]
2452:   \\
2453:   \nonumber
2454:   &&
2455:   {} + X_{T_3}
2456:     \left(T_3 + \overline{T}_3 \right)\partial_{T_3} \ln Y_{abc}
2457:   \\
2458:   \nonumber
2459:   &&
2460:   {} + X_H
2461:     \left(S + \overline{S}\right)^{\frac{1}{2}} H \partial_H \ln Y_{abc}
2462:     \\
2463:   &&
2464:   \nonumber
2465:   {} + X_{\overline{H}}
2466:       \left(T_3 + \overline{T}_3\right)^\frac{1}{2}\overline{H} \partial_{\overline{H}}
2467:       \ln Y_{abc}
2468:       \\
2469:   \label{eq:trilinear_one_42241}
2470:   &&
2471:   \left.
2472:     {} + X_\theta
2473:     \left(S+\overline{S}\right)^\frac{1}{4}
2474:     \left(T_3+\overline{T}_3\right)^\frac{1}{4}
2475:     \theta \partial_\theta \ln Y_{abc}
2476:   \right\}
2477: \end{eqnarray}
2478: 
2479: \begin{eqnarray}
2480:   \nonumber
2481:   %% F_{1,2} Fbar_{3} h 
2482:   A_{C_1^{5_1} C_3^{5_1} C^{5_1 5_2}}
2483:   &=&
2484:   \sqrt{3} m_{3/2}
2485:   \left\{
2486:     X_S
2487:       \left[
2488:         \frac{1}{2} +
2489:         \left(S + \overline{S}\right)\partial_S \ln Y_{abc} 
2490:       \right]
2491:     \right. 
2492:   \\
2493:   &&
2494:   \nonumber
2495:   {} + X_{T_1} 
2496:     \left[
2497:       -1 + \left(T_1 + \overline{T}_1 \right)\partial_{T_1} \ln Y_{abc}
2498:     \right]
2499:   \\
2500:   &&
2501:   \nonumber
2502:   {} + X_{T_2}
2503:     \left(T_2 + \overline{T}_2 \right) \partial_{T_2} \ln Y_{abc}
2504:   \\
2505:   &&
2506:   \nonumber
2507:   {} + X_{T_3}
2508:     \left[
2509:       - \frac{1}{2}
2510:       \left(T_3 + \overline{T}_3 \right)\partial_{T_3} \ln Y_{abc}
2511:     \right]
2512:   \\
2513:   \nonumber
2514:   &&
2515:   {} + X_H
2516:     \left(S + \overline{S}\right)^{\frac{1}{2}}H \partial_H \ln Y_{abc}
2517:   \\
2518:   \nonumber
2519:   &&
2520:   {} +
2521:     X_{\overline{H}}
2522:     \left(T_3 + \overline{T}_3\right)^\frac{1}{2}\overline{H} \partial_{\overline{H}}
2523:     \ln Y_{abc}
2524:     \\
2525:   \label{eq:trilinear_two_42241}
2526:     &&
2527:     \left.
2528:     {} + X_\theta
2529:     \left(S+\overline{S}\right)^\frac{1}{4}
2530:     \left(T_3+\overline{T}_3\right)^\frac{1}{4}
2531:     \theta \partial_\theta \ln Y_{abc}
2532:   \right\}
2533: \end{eqnarray}
2534: 
2535: \begin{eqnarray}
2536:   \nonumber
2537:   %% F_{3} Fbar_{1,2} h 
2538:   A_{C_1^{5_1} C_2^{5_1} C^{5_1 5_2}}
2539:   &=&
2540:   \sqrt{3} m_{3/2}
2541:   \left\{
2542:     X_S
2543:       \left[
2544:         \frac{1}{2} +
2545:         \left(S + \overline{S}\right)\partial_S \ln Y_{abc} 
2546:       \right]
2547:     \right. 
2548:   \\
2549:   &&
2550:   \nonumber
2551:   {} + 
2552:   X_{T_1}
2553:     \left[
2554:       -1 + \left(T_1 + \overline{T}_1 \right)\partial_{T_1} \ln Y_{abc}
2555:     \right]
2556:   \\
2557:   &&
2558:   \nonumber
2559:   {} + X_{T_2} 
2560:   \left[
2561:     -1 + \left(T_2 + \overline{T}_2 \right) \partial_{T_2} \ln Y_{abc}
2562:   \right]
2563:   \\
2564:   \nonumber
2565:   &&
2566:     {} +
2567:     X_{T_3}
2568:     \left[
2569:       \frac{1}{2}  
2570:       \left(T_3 + \overline{T}_3 \right)\partial_{T_3} \ln Y_{abc}
2571:     \right]
2572:   \\
2573:   \nonumber
2574:   &&
2575:   {} + X_H
2576:     \left(S + \overline{S}\right)^{\frac{1}{2}}H \partial_H \ln Y_{abc}
2577:   \\
2578:   &&
2579:   \nonumber
2580:     {} + X_{\overline{H}}
2581:     \left(T_3 + \overline{T}_3\right)^\frac{1}{2}\overline{H} \partial_{\overline{H}}
2582:     \ln Y_{abc}
2583:     \\
2584:   &&
2585:   \label{eq:trilinear_three_42241}
2586:   \left.
2587:     {} + X_\theta
2588:     \left(S+\overline{S}\right)^\frac{1}{4}
2589:     \left(T_3+\overline{T}_3\right)^\frac{1}{4}
2590:     \theta \partial_\theta \ln Y_{abc}
2591:   \right\}
2592: \end{eqnarray}
2593: 
2594: \begin{eqnarray}
2595:   \nonumber
2596:   %% F_{3} Fbar_{3} h 
2597:   A_{C_1^{5_1} C_2^{5_1} C_3^{5_1}}
2598:   &=&
2599:   \sqrt{3} m_{3/2}
2600:   \left\{
2601:     X_S
2602:         \left(S + \overline{S}\right)\partial_S \ln Y_{abc} 
2603:     \right. 
2604:   \\
2605:   &&
2606:   \nonumber
2607:   {} + X_{T_1}
2608:     \left[
2609:       -1 + \left(T_1 + \overline{T}_1 \right)\partial_{T_1} \ln Y_{abc}
2610:     \right]
2611:   \\
2612:   &&
2613:   \nonumber
2614:   {} + X_{T_2}
2615:     \left(T_2 + \overline{T}_2 \right) \partial_{T_2} \ln Y_{abc}
2616:   \\
2617:   \nonumber
2618:   &&
2619:   {} + X_{T_3}
2620:     \left(T_3 + \overline{T}_3 \right)\partial_{T_3} \ln Y_{abc}
2621:   \\
2622:   \nonumber
2623:   &&
2624:   {} + X_H
2625:     \left(S + \overline{S}\right)^{\frac{1}{2}}H \partial_H \ln Y_{abc}
2626:   \\
2627:   \nonumber
2628:   &&
2629:     {} +
2630:     X_{\overline{H}}
2631:     \left(T_3 + \overline{T}_3\right)^\frac{1}{2}\overline{H} \partial_{\overline{H}}
2632:     \ln Y_{abc}
2633:   \\
2634:   \label{eq:trilinear_four_42241}
2635:     &&
2636:     \left.
2637:     {} + X_\theta
2638:     \left(S+\overline{S}\right)^\frac{1}{4}
2639:     \left(T_3+\overline{T}_3\right)^\frac{1}{4}
2640:     \theta \partial_\theta \ln Y_{abc}
2641:   \right\}
2642: \end{eqnarray}
2643: 
2644: %%% ----------------------------------------------------------------------
2645: 
2646: %%% ----------------------------------------------------------------------
2647: 
2648: \section{Derivation of the D-terms}
2649: \label{sec:deriv D-terms}
2650: 
2651: %%% ----------------------------------------------------------------------
2652: 
2653: What follows is a full derivation of $D_H$ and $D_{\theta}$ from the
2654: superpotential of the $U(1)_F$ extended supersymmetric Pati-Salam 42241
2655: model, as detailed in Section~\ref{sec:442-pati-salam}.
2656: 
2657: 
2658: The relevent parts of the superpotential for the 42241 model are those
2659: concerning the Higgs and Froggatt-Nielsen fields which have different gauge
2660: singlets\footnote{Note that these must still have the same quantum numbers
2661: as they are both singlets, and therefore in the same representation of the
2662: gauge group.}.
2663: \begin{equation}
2664: W = S\lambda_S (\overline{H}H-M^2_H) + S'\lambda_{S'} (\overline{\theta}\theta
2665: -M^2_\theta) ,
2666: \end{equation}
2667: where the $M_H$ and $M_\theta$ are GUT scale masses associated with the
2668: Higgs and Froggatt-Nielsen vevs respectively. We have assumed that
2669: the heavy Higgs  develop vevs along the neutrino directions only, such that
2670: \begin{equation}
2671: \langle\overline{H}\rangle = \overline{H}_\nu = \overline{H} \quad ; \quad
2672: \langle H\rangle = H_\nu = H .
2673: \end{equation}
2674: 
2675: This is because charged objects gaining vevs would break their charge group
2676: at the GUT scale, causing problems\footnote{A colour charged object, for example,
2677: would imply that QCD is broken at the GUT scale, which would lead to a very
2678: massive gluon.} which the neutral components avoid. Similarly, the
2679: Froggatt-Nielsen vevs are concisely written as
2680: \begin{equation}
2681: \langle\overline{\theta}\rangle = \overline{\theta} \quad ; \quad
2682: \langle\theta\rangle = \theta .
2683: \end{equation}
2684: 
2685: The F-terms associated with the singlet fields are
2686: \begin{equation}
2687: \label{eq:F_S-term}
2688: |F_S|^2 = |\lambda_S(\overline{H}H -M^2_H)|^2
2689: \end{equation}
2690: and
2691: \begin{equation}
2692: \label{eq:F_S'-term}
2693: |F_{S'}|^2 = |\lambda_{S'}(\overline{\theta}\theta -M^2_\theta)|^2 .
2694: \end{equation}
2695: 
2696: We use these to form our Higgs potential,
2697: \begin{equation}
2698: V_H = V_D + V_F + V_{\mathrm{soft}} .
2699: \end{equation}
2700: 
2701: The F-term potential is trivially obtained from Eqs.~(\ref{eq:F_S-term}) and
2702: (\ref{eq:F_S'-term}), and the soft terms are simply written down with
2703: mass-squared terms for each of the soft SUSY-breaking scalar masses associated
2704: with the Higgs and FN vevs, as we will later see. The D-term potential
2705: takes a little more work, so we shall cover that here. The general form for
2706: the D-term potential is
2707: \begin{equation}
2708: \label{eq:general-V_D}
2709: V_D = \frac{1}{2} g^2_F D^1_F D^1_F +\frac{1}{2} g^2_{2R} \sum\limits^3_{a=1}
2710: D^a_{2R} D^a_{2R} + \frac{1}{2}g^2_4 \sum\limits^{15}_{m=1} D^m_4 D^m_4 ,
2711: \end{equation} % see p126 LaTeX book for the \sum\limits command.
2712: where $g_F$ is the gauge coupling for $U(1)_F$, $g_{2R}$ is for $SU(2)_R$,
2713: and $g_4$ is for $SU(4)_c$.
2714: 
2715: We focus on the $a=3$ and $m=15$ contributions to Eq.~(\ref{eq:general-V_D})
2716: which involve the
2717: \begin{equation}
2718: \label{eq:SU(2)_R-generator}
2719: \tau^3_R = \mathrm{diag}\left( \frac{1}{2} , -\frac{1}{2} \right)
2720: \end{equation}
2721: and
2722: \begin{equation}
2723: \label{eq:SU(4)-generator}
2724: T^{15}_4 = \sqrt{\frac{3}{2}}\mathrm{diag} \left( \frac{1}{6} , \frac{1}{6} , 
2725: \frac{1}{6} , -\frac{1}{2} \right) ,
2726: \end{equation}
2727: the diagonal generators of the $SU(2)_R$ and $SU(4)$ groups. There is no sum
2728: in the $U(1)$ part of Eq.~(\ref{eq:general-V_D}) because this group is
2729: the unit matrix {\bf 1}, so the only generator is
2730: \begin{equation}
2731: \label{eq:U(1)_F-generator}
2732: T^1_F = {\bf 1}q_F .
2733: \end{equation}
2734: where $q_F$ is the charge of the $U(1)_F$ group for each field. When
2735: applying Eqs.~(\ref{eq:SU(2)_R-generator}) and (\ref{eq:SU(4)-generator})
2736: to conjugate fields we have to complex conjugate the generator and
2737: multiply it by $-1$, but for the $U(1)_F$ group we just use
2738: Eq.~(\ref{eq:U(1)_F-generator}) where it is known that the $q_F$
2739: charges are different for right-handed fields.
2740: 
2741: Using Table~\ref{tab:particle_content_42241} we find that $D^1_F$, $D^3_{2R}$,
2742: and $D^{15}_4$ are given by
2743: \begin{eqnarray}
2744: \label{eq:D^1_F in terms of fields}
2745: D^1_F &=& \overline{H}^\dagger (q_F)_{\overline{H}} \overline{H}
2746: + H^\dagger (q_F)_H H + \overline{\theta}^\dagger (q_F)_{\overline{\theta}}
2747: \overline{\theta} + \theta^\dagger (q_F)_{\theta} \theta \nonumber \\ & &
2748: + \overline{F}^\dagger_i (q_F)_{\overline{F}_i} \overline{F}^j \delta^j_i
2749: + F^\dagger_i (q_F)_{F_i}F^j \delta^j_i + h^\dagger (q_F)_h h
2750: \\
2751: \label{eq:D^3_2R in terms of fields}
2752: D^3_{2R} &=& \overline{H}^\dagger (-\tau^{3*}_R) \overline{H}
2753: + H^\dagger (\tau^3_R) H + \overline{F}^\dagger (-\tau^{3*}_R) \overline{F}
2754: + h^\dagger (\tau^3_R) h
2755: \\
2756: \label{eq:D^15_4 in terms of fields}
2757: D^{15}_4 &=& \overline{H}^\dagger (-T^{15*}_4) \overline{H}
2758: + H^\dagger (T^{15}_4) H  + \overline{F}^\dagger (-T^{15*}_4) \overline{F}
2759: + F^\dagger (T^{15}_4) F.
2760: \end{eqnarray}
2761: 
2762: In Eq.~(\ref{eq:D^1_F in terms of fields}), $i,j \in \{1,2,3\}$ are family
2763: indices. We used $\delta^j_i$ to pick out the trace of the outer products
2764: $F^\dagger_i F^j$ and $\overline{F}^\dagger_i \overline{F}^j$, thereby giving us
2765: the dot product.
2766: 
2767: The scalar components of the left-handed matter superfield $F$ are $q$ and
2768: $l$. The scalar components of the right-handed matter superfield $\overline{F}$
2769: are $u^c$, $d^c$, $\nu^c$, and $e^c$. The tensorial conventions are shown in
2770: \cite{King:2000vp} for $D^3_{2R}$ and $D^{15}_4$. For $D^1_F$, all diagonal
2771: elements are equal to unity for the generator, so the tensor notation is trivial.
2772: 
2773: Now we square $D^1_F$, $D^3_{2R}$, and $D^{15}_4$ (as they are squared
2774: in the D-term potential), then consider the cross terms, as we are
2775: only interested in those terms that look like
2776: $(mass)^2\cdot(field)^2$. The $(mass)^2$ terms come from the vevs of
2777: the heavy Higgs and Froggatt-Nielsen fields $\overline{H}^2, H^2,
2778: \overline{\theta}^2, \theta^2$. The $(field)^2$ terms come from
2779: $|F_i|^2, |\overline{F}_i|^2$, where the $i$ is a family index running
2780: from 1 to 3.
2781: 
2782: In Eqs.~(\ref{eq:m^2_QL}) to (\ref{eq:m^2_hd}) we have defined $D^2_H$ and
2783: $D^2_\theta$ to be
2784: 
2785: \begin{eqnarray}
2786: \label{eq:D^2_H initial}
2787: D^2_H &=& \frac{1}{8} (\overline{H}^2 -H^2)\\ \label{eq:D^2_theta initial}
2788: D^2_\theta &=& - q_\theta (\overline{\theta}^2 -\theta^2)
2789: -q_H(\overline{H}^2 - H^2)  ,
2790: \end{eqnarray}
2791: 
2792: where $q_\theta$ is defined to be $-1$ and $q_H$ is defined to be $-q_{R3}$, thus
2793: $q_H = \frac{5}{6}$ for Model 1 and $q_H = 1$ for Model 2.
2794: 
2795: By considering the charge structure in Table~\ref{tab:charges_both_models}
2796: we can clearly see how the charge structure effects the additional $U(1)_F$
2797: D-term contributions to the sparticle mass matrices in 
2798: Eqs.~(\ref{eq:m^2_QL})-(\ref{eq:m^2_hd}).
2799: 
2800: We have followed \cite{King:2000vp} in choosing our designation of $D^2_H$,
2801: but $D^2_\theta$ is new. Working explicitly within Model 1, where
2802: $q_H = \frac{5}{6}$, we shall proceed to rewrite these D-terms as functions
2803: of the soft SUSY breaking masses and gauge couplings.
2804: So we arrive at our D-term potential, which, when put together with $V_F$ and
2805: $V_{\mathrm{soft}}$, forms our Higgs potential:
2806: \begin{eqnarray}
2807: V_{\mathrm{Higgs}} &=& \frac{1}{2} g^2_F \left[ \frac{25}{36} (\overline{H}^2
2808: -H^2)^2 +(\overline{\theta}^2 -\theta^2)^2 -\frac{5}{3}(\overline{H}^2
2809: -H^2)(\overline{\theta}^2 -\theta^2)\right] \nonumber \\ & &
2810: +\frac{1}{8} (g^2_{2R} +\frac{3}{2} g^2_4)(\overline{H}^2 -H^2)^2
2811: +\lambda^2_S(\overline{H}H -M^2_H)^2 +\lambda^2_{S'}
2812: (\overline{\theta}\theta -M^2_\theta)^2 \nonumber \\ & &
2813: +m^2_{\overline{H}} \overline{H}^2 +m^2_H H^2 +m^2_{\overline{\theta}}
2814: \overline{\theta}^2 +m^2_\theta \theta^2 ,
2815: \end{eqnarray}
2816: where $V_D$ are the terms multiplied by gauge coupling factors, $V_F$ are
2817: the terms multiplied by the dilaton lambdas, and $V_{\mathrm{soft}}$ are the
2818: last four terms multiplying the TeV scale soft SUSY breaking scalar masses.
2819: $M_H$ and $M_\theta$ are GUT scale masses.
2820: 
2821: To find the form of the D-terms, we must minimise this potential with
2822: respect to the fields $\overline{H}$, $H$, $\overline{\theta}$, $\theta$,
2823: and then set these minimisation relations equal to zero.
2824: 
2825: As these are set to zero, any linear combination of them is also zero, so
2826: taking the following combinations and rearranging them, we have two
2827: minimisation conditions
2828: \begin{eqnarray}
2829: \frac{\partial V}{\partial \overline{H}} -\frac{\partial V}{\partial H}
2830: &\Rightarrow& \left\{\frac{1}{4} \left[ \frac{25}{9} g^2_F +g^2_{2R}
2831: +\frac{3}{2} g^2_4 \right] (\overline{H} +H)^2 -\lambda^2_S (\overline{H}H
2832: -M^2_H )\right\} (\overline{H} -H) \nonumber \\ & & \label{eq:minimise H}
2833: -\frac{5}{6} g^2_F (\overline{\theta} +\theta)(\overline{\theta} -\theta)
2834: (\overline{H} +H) = -m^2_{\overline{H}} \overline{H} +m^2_H H \\
2835: \frac{\partial V}{\partial \overline{\theta}} -\frac{\partial V}{\partial
2836: \theta} &\Rightarrow& \frac{1}{6}g^2_F \left[ 6(\overline{\theta}^2 -\theta^2 )
2837: -5 (\overline{H}^2 -H^2)\right] (\overline{\theta} +\theta) \nonumber \\ & &
2838: \label{eq:minimise theta}
2839: -\lambda^2_{S'} (\overline{\theta} \theta -M^2_\theta) (\overline{\theta}
2840: -\theta) = -m^2_{\overline{\theta}} \overline{\theta} +m^2_\theta \theta .
2841: \end{eqnarray}
2842: 
2843: For $D$-flatness, it is necessary to set $\overline{H}^2 = H^2$ and
2844: $\overline{\theta}^2 = \theta^2$, which results in $V_D = 0$. For $F_S$-flatness
2845: and $F_{S'}$-flatness, it is necessary to set $\overline{H} H = M^2_H$ and
2846: $\overline{\theta} \theta = M^2_\theta$, yielding zero valued F-terms. 
2847: We wish to perturb away from these flatness conditions, so we impose
2848: \begin{eqnarray}
2849: \label{eq:perturb H}
2850: \overline{H} = M_H -\overline{m} &;& H = M_H -m \\ \label{eq:perturb theta}
2851: \overline{\theta} = M_\theta -\overline{m}' &;& \theta = M_\theta -m' ,
2852: \end{eqnarray}
2853: where $\overline{m}$, $m$, $\overline{m}'$, and $m'$ are all TeV scale masses.
2854: 
2855: Thus the two minimisation conditions, Eqs.~(\ref{eq:minimise H}) and
2856: (\ref{eq:minimise theta}), become\footnote{After rearranging and taking
2857: the leading order in the GUT scale masses $M_H$ and $M_\theta$.}
2858: \begin{eqnarray}
2859: \label{eq:min H 2}
2860: g^2_H (m-\overline{m}) M_H -\frac{10}{3} g^2_F (m' -\overline{m}') M_\theta
2861: &=& m^2_H -m^2_{\overline{H}} \\ \label{eq:min theta 2}
2862: \frac{2}{3} g^2_F \left[ 6(m' -\overline{m}' )M_\theta -5(m-\overline{m})M_H
2863: \right] &=& m^2_\theta - m^2_{\overline{\theta}}\; .
2864: \end{eqnarray}
2865: 
2866: Now putting the small perturbations, Eqs.~(\ref{eq:perturb H}) and
2867: (\ref{eq:perturb theta}), into Eqs.~(\ref{eq:D^2_H initial}) and
2868: (\ref{eq:D^2_theta initial}) for the D-terms, we have, to leading order,
2869: \begin{eqnarray}
2870: \label{eq:D^2_H intermediate}
2871: D^2_H &=& \frac{1}{4} (m-\overline{m})M_H \\ \label{eq:D^2_theta intermediate}
2872: D^2_\theta &=& \frac{1}{3} \left[ 6(m' -\overline{m}')M_\theta -5(m-
2873: \overline{m} )M_H \right] .
2874: \end{eqnarray}
2875: 
2876: So, using Eqs.~(\ref{eq:min H 2}) and (\ref{eq:min theta 2}) in the above
2877: Eqs.~(\ref{eq:D^2_H intermediate}) and (\ref{eq:D^2_theta intermediate}),
2878: we have the following expressions for our D-terms for Model 1, as functions of the
2879: soft SUSY breaking masses, GUT scale masses and gauge couplings
2880: \begin{eqnarray}
2881: \label{eq:D^2_H final}
2882: D^2_H &=& \frac{1}{4g^2_{2R}+6g^2_4} \left[ m^2_H -m^2_{\overline{H}} +\frac{5}{6}
2883: (m^2_\theta -m^2_{\overline{\theta}}) \right] \\ \label{eq:D^2_theta final}
2884: D^2_\theta &=& \frac{m^2_\theta -m^2_{\overline{\theta}}}{2g^2_F} .
2885: \end{eqnarray}
2886: Note that Eq.~(\ref{eq:D^2_theta final}) was used in obtaining
2887: Eq.~(\ref{eq:D^2_H final}). This is the form of the D-terms as used in
2888: the updated version of SOFTSUSY \cite{Allanach:2001kg} to compute
2889: the slepton mass data for our lepton flavour violating branching ratios
2890: in Model 1, using Eqs.~(\ref{eq:m^2_QL}) to (\ref{eq:m^2_hd}).
2891: 
2892: For Model 2, the derivation is very similar, with the factor of $q_H$ being
2893: the only difference, and so we obtain the form of $D^2_H$ below,
2894: with $D^2_\theta$ being the same in both models.
2895: 
2896: \begin{eqnarray}
2897: \label{eq:D^2_H final2}
2898: D^2_H &=& \frac{(m^2_H -m^2_{\overline{H}})
2899: +(m^2_\theta -m^2_{\overline{\theta}})}{4g^2_{2R}+6g^2_4} .
2900: \end{eqnarray}
2901: 
2902: In both cases, we can see that the Pati-Salam limit is obtained when the
2903: Froggatt-Nielsen scalar masses are degenerate, $m^2_\theta = m^2_{\overline{\theta}}$.
2904: This result differs from \cite{King:2000vp} due to a different derivation
2905: procedure.
2906: 
2907: %%% ----------------------------------------------------------------------
2908: 
2909: \section{Operators for the two models}
2910: \label{operatorsformodels}
2911: 
2912: In this appendix we give the operators which are responsible
2913: for generating the Yukawa matrices of Models 1 and 2 analysed in this paper.
2914: %%% ----------------------------------------------------------------------
2915: 
2916: \subsection{Model 1}
2917: \label{sec:model-1}
2918: 
2919: %%% ----------------------------------------------------------------------
2920: 
2921: \begin{table}[tp]
2922:   \centering
2923:   \begin{tabular}{|c|c|c|c|}
2924:     \hline  
2925:     \multicolumn{2}{|c|}{Model 1} &
2926:     \multicolumn{2}{c|}{Model 2} \\
2927:     \hline
2928:     $\delta$ & 0.22 & $\delta$ & 0.22 \\
2929:     $\epsilon$ & 0.22 & $\epsilon$ & 0.22 \\
2930:     \hline
2931:     $a_{33}$ & 0.55 & $a_{33}$ & 0.55 \\
2932:     \hline
2933:     $a_{11}$ & -0.92 & $a_{11}^\prime$ & -0.92 \\
2934:     $a_{12}$ & 0.33 & $a_{12}^\prime$ & 0.33 \\
2935:     $a_{13}$ & 0.00 & $a_{13}$ & 0.00 \\
2936:     $a_{21}$ & 1.67 & $a_{21}$ & 1.67 \\
2937:     $a_{22}$ & 1.12 & $a_{22}$ & 1.12 \\
2938:     $a_{23}$ & 0.89 & $a_{23}$ & 0.89 \\
2939:     $a_{31}$ & -0.21 & $a_{31}$ & -0.21 \\
2940:     $a_{32}$ & 2.08 & $a_{32}$ & 2.08 \\
2941:     \hline
2942:     $a^\prime_{12}$ & 0.77 & $a^{\prime\prime}_{12}$ & 0.77 \\
2943:     $a^\prime_{13}$ & 0.53 & $a^{\prime\prime}_{13}$ & 0.53 \\
2944:     $a^\prime_{22}$ & 0.66 & $a^\prime_{22}$ & 0.66 \\
2945:     $a^\prime_{23}$ & 0.40 & $a^\prime_{23}$ & 0.40 \\
2946:     $a^\prime_{32}$ & 1.80 & $a^\prime_{32}$ & 1.80 \\
2947:     \hline
2948:     $a^{\prime\prime}_{11}$ & 0.278 & $a^{\prime\prime\prime}_{11}$ & 0.278 \\
2949:     $a^{\prime\prime}_{12}$ & 0.000 & $a^{\prime\prime\prime}_{12}$ & 0.000 \\
2950:     $a^{\prime\prime}_{13}$ & 0.000 & $a^{\prime\prime\prime}_{13}$ & 0.000 \\
2951:     \hline
2952:     $A_{11}$ & 0.94 &     $A_{11}$ & 0.94  \\
2953:     $A_{12}$ & 0.48 &    $A_{12}$ & 0.48 \\
2954:     $A_{13}$ & 2.10 &    $A_{13}$ & 2.10 \\
2955:     $A_{22}$ & 0.52 &    $A_{22}$ & 0.52 \\
2956:     $A_{23}$ & 1.29 &    $A_{23}$ & 1.29 \\
2957:     $A_{33}$ & 1.88 &    $A_{33}$ & 1.88 \\
2958:     \hline
2959:   \end{tabular}
2960:   \caption{The $O(1)$ coefficients in Model 1 and Model 2. The values of $a_{13}$,
2961:   $a^{\prime\prime}_{12}$ (in Model 1) and $a^{\prime\prime\prime}_{12}$ ( in Model 2)
2962: will be varied, as discussed in the text.}
2963:   \label{tab:a_ap_app_for_models}
2964: \end{table}
2965: 
2966: This model is almost the model studied in
2967: \cite{Blazek:2003wz,KP0307091}, but with an extra operator in the
2968: $(1,2)$ and $(1,3)$ Yukawa matrix elements to allow a non-zero $Y^e_{12}$ 
2969: and $Y^e_{13}$. The
2970: operator texture is
2971: {\footnotesize
2972: \begin{equation}
2973:   \label{eq:operator_texture1}
2974:   \mathcal{O} 
2975:   =
2976:   \left[
2977:     \begin{array}{ccc}
2978:       ( a_{11} \mathcal{O}^{Fc}
2979:       + a^{\prime\prime}_{11} \mathcal{O}''^{Ae} )\epsilon^5 &
2980:       ( a_{12} \mathcal{O}^{Ee} + a^\prime_{12} \mathcal{O}'^{Cb}
2981:       + a^{\prime\prime}_{12}O''^{Ec} ) \epsilon^3 &
2982:       ( a_{13} \mathcal{O}^{Ec} + a^\prime_{13} \mathcal{O}'^{Cf}
2983:       + a^{\prime\prime}_{13}\mathcal{O}''^{Ee} )\epsilon \\
2984:       ( a_{21} \mathcal{O}^{Dc} ) \epsilon^4 &
2985:       ( a_{22} \mathcal{O}^{Bc} + a_{22}^\prime \mathcal{O}'^{Ff} ) \epsilon^2 &
2986:       ( a_{23} \mathcal{O}^{Ee} + a^\prime_{23} \mathcal{O}'^{Bc} )  \\
2987:       ( a_{31} \mathcal{O}^{Fc} ) \epsilon^4 &
2988:       ( a_{32} \mathcal{O}^{Ac} + a_{23}^\prime \mathcal{O}'^{Fe} ) \epsilon^2
2989:       & a_{33}
2990:     \end{array}
2991:   \right]
2992: \end{equation}
2993: } 
2994: where the operator nomenclature is defined in Appendix
2995: \ref{sec:n=1-operators} and Appendix \ref{sec:n1-operators}.
2996: This leads to the following Yukawa textures:
2997: \begin{eqnarray}
2998:   \label{eq:explicit_yu_model1}
2999:   Y^{u}(M_X) &=&
3000:   \left[
3001:     \begin{array}{ccc}
3002:     a^{\prime\prime}_{11} \sqrt{2} \delta^3 \epsilon^5 & 
3003:     a^\prime_{12} \sqrt{2} \delta^2\epsilon^3 &  
3004:     a^\prime_{13}\frac{2}{\sqrt{5}} \delta^2\epsilon  \\
3005:     0 & 
3006:     a^\prime_{22} \frac{8}{5 \sqrt{5}}\delta^2\epsilon^2 
3007:     & 0 \\
3008:     0 
3009:     & a^\prime_{32} \frac{8}{5} \delta^2 \epsilon^2 
3010:     & a_{33}
3011:     \end{array}
3012:   \right]
3013:   \\
3014:   \label{eq:explicit_yd_model1}
3015:   Y^{d}(M_X) &=&
3016:   \left[
3017:     \begin{array}{ccc}
3018:       a_{11} \frac{8}{5}\delta \epsilon^5 & -a^\prime_{12} \sqrt{2}
3019:       \delta^2\epsilon^3 & a^\prime_{13} \frac{4}{\sqrt{5}} \delta^2\epsilon\\ a_{21}
3020:       \frac{2}{\sqrt{5}}\delta \epsilon^4 & (a_{22} \sqrt{\frac{2}{5}}
3021:       \delta + a^\prime_{22} \frac{16}{5\sqrt{5}} \delta^2)\epsilon^2
3022:       & a^\prime_{23} \sqrt{\frac{2}{5}} \delta^2 \\ a_{31}
3023:       \frac{8}{5}\delta\epsilon^4 & a_{32} \sqrt{2} \delta \epsilon^2
3024:       & a_{33}
3025:     \end{array}
3026:   \right]
3027:   \\
3028:   \label{eq:explicit_ye_model1}
3029:   Y^{e}(M_X) &=&
3030:   \left[
3031:     \begin{array}{ccc}
3032:     a_{11} \frac{6}{5}\delta \epsilon^5 & a_{12}^{\prime\prime} 2\delta^2 \epsilon^3 &
3033:     a_{13} 2 \delta^2 \epsilon \\ a_{21} \frac{4}{\sqrt{5}} \delta\epsilon^4 & ( -a_{22}
3034:     3\sqrt{\frac{2}{5}} \sqrt{\frac{2}{5}} + a^\prime_{22} \delta
3035:     \frac{12}{5\sqrt{5}} ) \delta\epsilon^2 & -a^\prime_{23}
3036:     \sqrt{\frac{2}{5}} \delta^2 \\ -a_{31} \frac{6}{5}
3037:     \delta\epsilon^4 & a_{32} \sqrt{2} \delta\epsilon^2 & a_{33}
3038:   \end{array}
3039: \right]
3040:   \\
3041:   \label{eq:explicit_yn_model1}
3042:   Y^{\nu}(M_X) &=&
3043:   \left[
3044:     \begin{array}{ccc}
3045:       a^{\prime\prime}_{11} \sqrt{2}\delta^3\epsilon^5 & 
3046:       a_{12} 2 \delta\epsilon^3 & 
3047:       a^{\prime\prime}_{13} 2\delta^3\epsilon \\
3048:       0 & 
3049:       a^\prime_{22} \frac{6}{5\sqrt{5}}\delta^2\epsilon^2 & 
3050:       a_{23} 2 \delta \\
3051:       0 & 
3052:       a^\prime_{32} \frac{6}{5} \delta^2\epsilon^2 & 
3053:       a_{33}
3054:     \end{array}
3055:   \right]
3056: \end{eqnarray}
3057: 
3058: For both models we define $\epsilon$ and $\delta$ as:
3059: \begin{equation}
3060:   \label{eq:def_eps_delta} \epsilon = \left( \frac{\langle\theta\rangle}{M_X}
3061: \right) \;\;\;\; ; \;\;\;\; \delta = \left( \frac{\langle H\rangle\langle\overline
3062: H\rangle}{M^2_X} \right).
3063: \end{equation}
3064: We take $\delta = \epsilon = 0.22$.
3065: 
3066: The values of the arbritary couplings are laid out in
3067: Table~\ref{tab:a_ap_app_for_models}. This gives numerical values for
3068: the Yukawa elements which can be used in either model, with the relevant
3069: values of $Y^e_{12}$ and $Y^e_{13}$ inserted instead of the texture zeros:
3070: \begin{eqnarray}
3071:   \label{eq:numerical_yu_model1A}
3072:   Y^{u}(M_X) &=&
3073:   \left[
3074:     \begin{array}{lll}
3075:       2.159\times10^{-06}  &  5.606\times10^{-04}  &  5.090\times10^{-03} \\
3076:       0.000    &    1.105\times10^{-03}  &   0.000    \\
3077:       0.000    &    6.733\times10^{-03}  &  5.841\times10^{-01}    \\
3078:     \end{array}
3079:   \right]
3080:   \\
3081:   \label{eq:numerical_yd_model1A}
3082:   Y^{d}(M_X) &=&
3083:   \left[
3084:     \begin{array}{lll}
3085:     \!\!\!\!-1.661\times10^{-04} & \!\!\!\! -5.606\times10^{-04} &   1.018\times10^{-02} \\
3086:      7.683\times10^{-04}  & \!\!\!\!-5.343\times10^{-03}  &  1.216\times10^{-02}\\
3087:     \!\!\!\!-1.769\times10^{-04}   & 3.133\times10^{-02}   & 3.933\times10^{-01}\\
3088:     \end{array}
3089:   \right]
3090:   \\
3091:   \label{eq:numerical_ye_model1A}
3092:   Y^{e}(M_X) &=&
3093:   \left[
3094:     \begin{array}{lll}
3095:     \!\!\!\!-1.246\times10^{-04} &    0.000   &      0.000   \\
3096:      1.537\times10^{-03}  &  2.432\times10^{-02}&   \!\!\!\!-3.649\times10^{-02}\\
3097:     \!\!\!\!-1.327\times10^{-04}   & 3.133\times10^{-02} &   5.469\times10^{-01}\\
3098:   \end{array}
3099: \right]
3100:   \\
3101:   \label{eq:numerical_yn_model1A}
3102:   Y^{\nu}(M_X) &=&
3103:   \left[
3104:     \begin{array}{lll}
3105:      2.159\times10^{-06}&    1.525\times10^{-03}&    0.000\\
3106:       0.000     &   8.290\times10^{-04} &   3.923\times10^{-01}  \\  
3107:       0.000      &  5.050\times10^{-03}  &  5.469\times10^{-01} \\
3108:     \end{array}
3109:   \right]
3110: \end{eqnarray}
3111: 
3112: The RH Majorana neutrino mass matrix is:
3113: \begin{equation}
3114:   \label{eq:21}
3115:   \frac{M_{RR}(M_X)}{M_{33}} =
3116:   \left[
3117:     \begin{array}{ccc}
3118:       A_{11}\delta\epsilon^8 & A_{12}\delta\epsilon^6 &
3119:       A_{13}\delta\epsilon^4 \\ A_{12}\delta\epsilon^6 &
3120:       A_{22}\delta\epsilon^4 & A_{23}\delta\epsilon^2 \\
3121:       A_{13}\delta\epsilon^4 & A_{23}\delta\epsilon^2 & A_{33}
3122:     \end{array}
3123:   \right]
3124: \end{equation}
3125: 
3126: The numerical values for the Majorana mass matrix are
3127: \begin{equation}
3128:   \label{eq:numerical_21A}
3129:   \frac{M_{RR}(M_X)}{M_{33}} =
3130:   \left[
3131:     \begin{array}{lll}
3132:       3.508\times10^{8}  &       3.686\times10^{9}   &     3.345\times10^{11}\\
3133:       3.686\times10^{9}   &     8.313\times10^{10}&    5.886\times10^{12}\\
3134:      3.345\times10^{11}&    5.886\times10^{12} &   5.795\times10^{14}\\
3135:     \end{array}
3136:   \right]
3137: \end{equation}
3138: 
3139: 
3140: %%% ----------------------------------------------------------------------
3141: 
3142: \subsection{Model 2}
3143: \label{sec:model-2}
3144: 
3145: %%% ----------------------------------------------------------------------
3146: 
3147: The operator texture for Model 2 is:
3148: {\footnotesize
3149: \begin{equation}
3150:   \label{eq:operator_texture2}
3151:   \mathcal{O} 
3152:   =
3153:   \left[
3154:     \begin{array}{ccc}
3155:       ( a_{11}^\prime \mathcal{O}'^{Fc}
3156:       + a^{\prime\prime\prime}_{11} \mathcal{O}'''^{Ae} )\epsilon^4 &
3157:       ( a_{12}^\prime \mathcal{O}'^{Ee} + a^{\prime\prime}_{12} \mathcal{O}''^{Cb}
3158:       + a^{\prime\prime\prime}_{12}O'''^{Ec} ) \epsilon^2 &
3159:       ( a_{13} \mathcal{O}^{Ec} + a^{\prime\prime}_{13} \mathcal{O}''^{Cf}
3160:       + a^{\prime\prime\prime}_{13}\mathcal{O}'''^{Ee} )  \\
3161:       ( a_{21} \mathcal{O}^{Dc} ) \epsilon^4 &
3162:       ( a_{22} \mathcal{O}^{Bc} + a_{22}^\prime \mathcal{O}'^{Ff} ) \epsilon^2 &
3163:       ( a_{23} \mathcal{O}^{Ee} + a^\prime_{23} \mathcal{O}'^{Bc} )  \\
3164:       ( a_{31} \mathcal{O}^{Fc} ) \epsilon^4 &
3165:       ( a_{32} \mathcal{O}^{Ac} + a_{23}^\prime \mathcal{O}'^{Fe} ) \epsilon^2
3166:       & a_{33}
3167:     \end{array}
3168:   \right]
3169: \end{equation}
3170: }
3171: 
3172: The operator nomenclature is defined in Appendix \ref{sec:n=1-operators} and
3173: Appendix \ref{sec:n1-operators}. The new operator setup leads to the following
3174: Yukawa textures:
3175: \begin{eqnarray}
3176:   \label{eq:explicit_yu_model2}
3177:   Y^{u}(M_X) &=&
3178:   \left[
3179:     \begin{array}{ccc}
3180:     a^{\prime\prime\prime}_{11} \sqrt{2} \delta^4 \epsilon^4 & 
3181:     a^{\prime\prime}_{12} \sqrt{2} \delta^3\epsilon^2 &  
3182:     a^{\prime\prime}_{13}\frac{2}{\sqrt{5}} \delta^3  \\
3183:     0 & 
3184:     a^\prime_{22} \frac{8}{5 \sqrt{5}}\delta^2\epsilon^2 
3185:     & 0 \\
3186:     0 
3187:     & a^\prime_{32} \frac{8}{5} \delta^2 \epsilon^2 
3188:     & a_{33}
3189:     \end{array}
3190:   \right]
3191:   \\
3192:   \label{eq:explicit_yd_model2}
3193:   Y^{d}(M_X) &=&
3194:   \left[
3195:     \begin{array}{ccc}
3196:       a^\prime_{11} \frac{8}{5}\delta^2 \epsilon^4 & 
3197:       -a^{\prime\prime}_{12} \sqrt{2} \delta^3\epsilon^2 &
3198:       a^{\prime\prime}_{13} \frac{4}{\sqrt{5}} \delta^3\\ 
3199:       a_{21} \frac{2}{\sqrt{5}}\delta \epsilon^4 & (a_{22} \sqrt{\frac{2}{5}}
3200:       \delta + a^\prime_{22} \frac{16}{5\sqrt{5}} \delta^2)\epsilon^2
3201:       & a^\prime_{23} \sqrt{\frac{2}{5}} \delta^2 \\ a_{31}
3202:       \frac{8}{5}\delta\epsilon^4 & a_{32} \sqrt{2} \delta \epsilon^2
3203:       & a_{33}
3204:     \end{array}
3205:   \right]
3206:   \\
3207:   \label{eq:explicit_ye_model2}
3208:   Y^{e}(M_X) &=&
3209:   \left[
3210:     \begin{array}{ccc}
3211:     a_{11}^\prime \frac{6}{5}\delta^2 \epsilon^4 & a_{12}^{\prime\prime\prime} 2 \delta^3
3212:     \epsilon^2 & a_{13} 2 \delta^3 \\ a_{21}
3213:     \frac{4}{\sqrt{5}} \delta\epsilon^4 & ( -a_{22}
3214:     3\sqrt{\frac{2}{5}} \sqrt{\frac{2}{5}} + a^\prime_{22} \delta
3215:     \frac{12}{5\sqrt{5}} ) \delta\epsilon^2 & -a^\prime_{23}
3216:     \sqrt{\frac{2}{5}} \delta^2 \\ -a_{31} \frac{6}{5}
3217:     \delta\epsilon^4 & a_{32} \sqrt{2} \delta\epsilon^2 & a_{33}
3218:   \end{array}
3219: \right]
3220:   \\
3221:   \label{eq:explicit_yn_model2}
3222:   Y^{\nu}(M_X) &=&
3223:   \left[
3224:     \begin{array}{ccc}
3225:       a^{\prime\prime\prime}_{11} \sqrt{2}\delta^4\epsilon^4 & 
3226:       a^\prime_{12} 2 \delta^2\epsilon^2 & 
3227:       a^{\prime\prime\prime}_{13} 2\delta^4 \\
3228:       0 & 
3229:       a^\prime_{22} \frac{6}{5\sqrt{5}}\delta^2\epsilon^2 & 
3230:       a_{23} 2 \delta \\
3231:       0 & 
3232:       a^\prime_{32} \frac{6}{5} \delta^2\epsilon^2 & 
3233:       a_{33}
3234:     \end{array}
3235:   \right]
3236: \end{eqnarray}
3237: 
3238: The RH Majorana neutrino mass matrix is the same as in Model 1:
3239: \begin{equation}
3240: \label{eq:4}
3241: \frac{M_{RR}(M_X)}{M_{33}} =
3242:   \left[
3243:     \begin{array}{ccc}
3244:       A_{11}\delta\epsilon^8 & A_{12}\delta\epsilon^6 &
3245:       A_{13}\delta\epsilon^4 \\ A_{12}\delta\epsilon^6 &
3246:       A_{22}\delta\epsilon^4 & A_{23}\delta\epsilon^2 \\
3247:       A_{13}\delta\epsilon^4 & A_{23}\delta\epsilon^2 & A_{33}
3248:     \end{array}
3249:   \right]
3250: \end{equation}
3251: 
3252: %%% ----------------------------------------------------------------------
3253: 
3254: 
3255: \section{$n=1$ operators}
3256: \label{sec:n=1-operators}
3257: 
3258: %%% ----------------------------------------------------------------------
3259: 
3260: \begin{table}[htbp]
3261:   \centering
3262:   \mbox
3263:   {
3264:   \begin{tabular}{|c|c||c|c|c|c|}
3265:     \hline Operator Name & Operator Name in
3266:     \cite{King:OperatorAnalysis} & $Q\overline{U}h_2$ &
3267:     $Q\overline{D}h_1$ & $L \overline{E}h_1$ & $L \overline{N}h_2$ \\
3268:     \hline $O^{Aa}$ & $O^A$ & $1$ & $1$ & $1$ & $1$ \\ $O^{Ab}$ &
3269:     $O^B$ & $1$ & $-1$ & $-1$ & $1$ \\ $O^{Ac}$ & $O^M$ & $0$ &
3270:     $\sqrt{2}$ & $\sqrt{2}$ & $0$ \\ $O^{Ad}$ & $O^T$ &
3271:     $\frac{2\sqrt{2}}{5}$ & $\frac{\sqrt{2}}{5}$ &
3272:     $\frac{\sqrt{2}}{5}$ & $\frac{2\sqrt{2}}{5}$ \\ $O^{Ae}$ & $O^V$ &
3273:     $\sqrt{2}$ & $0$ & $0$ & $\sqrt{2}$ \\ $O^{Af}$ & $O^U$ &
3274:     $\frac{\sqrt{2}}{5}$ & $\frac{2\sqrt{2}}{5}$ &
3275:     $\frac{2\sqrt{2}}{5}$ & $\frac{\sqrt{2}}{5}$ \\ \rule{0mm}{4mm}
3276:     $O^{Ba}$ & $O^C$ & $\frac{1}{\sqrt{5}}$ & $\frac{1}{\sqrt{5}}$ &
3277:     $\frac{-3}{\sqrt{5}}$ & $\frac{-3}{\sqrt{5}}$ \\ \rule{0mm}{4mm}
3278:     $O^{Bb}$ & $O^D$ & $\frac{1}{\sqrt{5}}$ & $\frac{-1}{\sqrt{5}}$ &
3279:     $\frac{-3}{\sqrt{5}}$ & $\frac{3}{\sqrt{5}}$ \\ $O^{Bc}$ & $O^W$ &
3280:     $0$ & $\sqrt{\frac{2}{5}}$ & $-3\sqrt{\frac{2}{5}}$ & $0$ \\
3281:     $O^{Bd}$ & $O^X$ & $\frac{2\sqrt{2}}{5}$ & $\frac{\sqrt{2}}{5}$ &
3282:     $\frac{-3\sqrt{2}}{5}$ & $\frac{-6\sqrt{2}}{5}$ \\ $O^{Be}$ &
3283:     $O^Z$ & $\sqrt{\frac{2}{5}}$ & $0$ & $0$ & $-3\sqrt{\frac{2}{5}}$
3284:     \\ $O^{Bf}$ & $O^Y$ & $\frac{\sqrt{2}}{5}$ & $\frac{2\sqrt{2}}{5}$
3285:     & $\frac{-6\sqrt{2}}{5}$ & $\frac{-3\sqrt{2}}{5}$ \\ $O^{Ca}$ &
3286:     $O^a$ & $\sqrt{2}$ & $\sqrt{2}$ & $0$ & $0$ \\ $O^{Cb}$ & $O^F$ &
3287:     $\sqrt{2}$ & $-\sqrt{2}$ & $0$ & $0$ \\ $O^{Cc}$ & $O^E$ & $0$ &
3288:     $2$ & $0$ & $0$ \\ $O^{Cd}$ & $O^b$ & $\frac{4}{\sqrt{5}}$ &
3289:     $\frac{2}{\sqrt{5}}$ & $0$ & $0$ \\ $O^{Ce}$ & $O^N$ & $2$ & $0$ &
3290:     $0$ & $0$ \\ $O^{Cf}$ & $O^c$ & $\frac{2}{\sqrt{5}}$ &
3291:     $\frac{4}{\sqrt{5}}$ & $0$ & $0$ \\ $O^{Da}$ & $O^d$ &
3292:     $\sqrt{\frac{2}{5}}$ & $\sqrt{\frac{2}{5}}$ &
3293:     $2\sqrt{\frac{2}{5}}$ & $2\sqrt{\frac{2}{5}}$ \\ $O^{Db}$ & $O^e$
3294:     & $\sqrt{\frac{2}{5}}$ & $-\sqrt{\frac{2}{5}}$ &
3295:     $-2\sqrt{\frac{2}{5}}$ & $2\sqrt{\frac{2}{5}}$ \\ $O^{Dc}$ & $O^G$
3296:     & $0$ & $\frac{2}{\sqrt{5}}$ & $\frac{4}{\sqrt{5}}$ & $0$ \\
3297:     \rule{0mm}{4mm} $O^{Dd}$ & $O^H$ & $\frac{4}{5}$ & $\frac{2}{5}$ &
3298:     $\frac{4}{5}$ & $\frac{8}{5}$ \\ \rule{0mm}{4mm} $O^{De}$ & $O^O$
3299:     & $\frac{2}{\sqrt{5}}$ & $0$ & $0$ & $\frac{4}{\sqrt{5}}$ \\
3300:     $O^{Df}$ & $O^f$ & $\frac{2}{5}$ & $\frac{4}{5}$ & $\frac{8}{5}$ &
3301:     $\frac{4}{5}$ \\ $O^{Ea}$ & $O^g$ & $0$ & $0$ & $\sqrt{2}$ &
3302:     $\sqrt{2}$ \\ $O^{Eb}$ & $O^h$ & $0$ & $0$ & $-\sqrt{2}$ &
3303:     $\sqrt{2}$ \\ $O^{Ec}$ & $O^i$ & $0$ & $0$ & $2$ & $0$ \\ $O^{Ed}$
3304:     & $O^j$ & $0$ & $0$ & $\frac{2}{\sqrt{5}}$ & $\frac{4}{\sqrt{5}}$
3305:     \\ $O^{Ee}$ & $O^I$ & $0$ & $0$ & $0$ & $2$ \\ $O^{Ef}$ & $O^J$ &
3306:     $0$ & $0$ & $\frac{4}{\sqrt{5}}$ & $\frac{2}{\sqrt{5}}$ \\
3307:     $O^{Fa}$ & $O^P$ & $\frac{4\sqrt{2}}{5}$ & $\frac{4\sqrt{2}}{5}$ &
3308:     $\frac{3\sqrt{2}}{5}$ & $\frac{3\sqrt{2}}{5}$ \\ $O^{Fb}$ & $O^Q$
3309:     & $\frac{4\sqrt{2}}{5}$ & $\frac{-4\sqrt{2}}{5}$ &
3310:     $\frac{-3\sqrt{2}}{5}$ & $\frac{3\sqrt{2}}{5}$ \\ \rule{0mm}{4mm}
3311:     $O^{Fc}$ & $O^R$ & $0$ & $\frac{8}{5}$ & $\frac{6}{5}$ & $0$ \\
3312:     \rule{0mm}{4mm} $O^{Fd}$ & $O^L$ & $\frac{16}{5\sqrt{5}}$ &
3313:     $\frac{8}{5\sqrt{5}}$ & $\frac{6}{5\sqrt{5}}$ &
3314:     $\frac{12}{5\sqrt{5}}$ \\ $O^{Fe}$ & $O^K$ & $\frac{8}{5}$ & $0$ &
3315:     $0$ & $\frac{6}{5}$ \\ $O^{Ff}$ & $O^S$ & $\frac{8}{5\sqrt{5}}$ &
3316:     $\frac{16}{5\sqrt{5}}$ & $\frac{12}{5\sqrt{5}}$ &
3317:     $\frac{6}{5\sqrt{5}}$ \\ \hline
3318:   \end{tabular}
3319:   }
3320:   \caption{Operator names, CGCs and names in \cite{King:OperatorAnalysis}}
3321:   \label{tab:operator_names}
3322: \end{table}
3323: 
3324: 
3325: The $n=1$ Dirac operators are the complete set of all operators that can be
3326: constructed from the quintilinear $F\overline{F}h\overline{H}H$ by all possible
3327: group theoretical contractions of the indicies in
3328: \begin{equation}
3329:   \label{eq:fields_tensor_app}
3330:   \mathcal{O}^{\alpha \rho y w}_{\beta \gamma x z} =
3331:   F^{\alpha a}\overline{F}_{\beta x} h^y_a \overline{H}_{\gamma z} H^{\rho w}
3332: \end{equation}
3333: 
3334: We define some \SU{4} invariant tensors $C$ and some \SU{2} invariant
3335: tensors $R$ as follows\footnote{The subscript denotes the dimension of
3336: the representation they can create from multiplying $\mathbf{4}$ or
3337: $\overline{\mathbf{4}}$ with $\mathbf{4}$ or
3338: $\overline{\mathbf{4}}$. For example
3339: $(C_{15})^{\beta\gamma}_{\alpha\rho}\overline{\mathbf{4}}_\gamma
3340: \mathbf{4}^\rho = \mathbf{15}^\beta_\alpha$ .}:
3341: \begin{eqnarray}
3342:   \nonumber \left(C_1\right)^\alpha_\beta &=& \delta^\alpha_\beta \\
3343:   \nonumber \left(C_6\right)^{\rho\gamma}_{\alpha\beta} &=&
3344:   \epsilon_{\alpha\beta\omega\chi}^{\rho\gamma\omega\chi} \\ \nonumber
3345:   \left(C_{10}\right)^{\alpha\beta}_{\rho\gamma} &=&
3346:   \delta^\alpha_\rho \delta^\beta_\gamma + \delta^\alpha_\gamma
3347:   \delta^\beta_\rho \\ \nonumber
3348:   \left(C_{15}\right)^{\beta\gamma}_{\alpha\rho} &=& \delta^\beta_\rho
3349:   \delta^\gamma_\alpha -\frac{1}{4} \delta^\beta_\alpha
3350:   \delta^\gamma_\rho \\ \nonumber \left(R_1\right)^x_y &=& \delta^x_y
3351:   \\
3352:   \label{eq:tensors}
3353:   \left(R_3\right)^{wx}_{yz} &=& 
3354:   \delta^x_y \delta^w_z - \frac{1}{2} \delta^x_z \delta^w_y
3355: \end{eqnarray}
3356: 
3357: Then the six independent \SU{4} structures are:
3358: \begin{eqnarray}
3359:   \nonumber
3360:   \mathrm{A}. &
3361:   \left(C_1\right)^\beta_\alpha \left(C_1\right)^\gamma_\rho 
3362:   &=\;\;
3363:   \delta^\beta_\alpha \delta^\gamma_\rho
3364:   \\
3365:   \nonumber
3366:   \mathrm{B}. &
3367:   \left(C_{15}\right)^{\beta\chi}_{\alpha\sigma}
3368:   \left(C_{15}\right)^{\gamma\sigma}_{\rho\chi}
3369:   &=\;\;
3370:   \delta^\beta_\rho \delta^\gamma_\alpha - 
3371:   \frac{1}{4} \delta^\beta_\alpha \delta^\gamma_\rho
3372:   \\
3373:   \nonumber
3374:   \mathrm{C}. &
3375:   \left(C_6\right)^{\omega\chi}_{\alpha\rho}
3376:   \left(C_6\right)^{\beta\gamma}_{\omega\chi} 
3377:   &=\;\;
3378:   8(\delta^\beta_\alpha \delta^\gamma_\alpha 
3379:   - \delta^\gamma_\alpha \delta^\beta_\rho )
3380:   \\
3381:   \nonumber
3382:   \mathrm{D}. &
3383:   \left(C_{10}\right)^{\omega\chi}_{\alpha\rho}
3384:   \left(C_{10}\right)^{\beta\gamma}_{\omega\chi} 
3385:   &=\;\;
3386:   2(\delta^\beta_\alpha \delta^\gamma_\rho
3387:   + \delta^\gamma_\alpha \delta^\beta_\rho )
3388:   \\
3389:   \nonumber
3390:   \mathrm{E}. &
3391:   \left(C_1\right)^\beta_\rho
3392:   \left(C_1\right)^\gamma_\alpha 
3393:   &=\;\;
3394:   \delta^\beta_\alpha \delta^\gamma_\alpha
3395:   \\
3396:   \label{eq:su4_structures}
3397:   \mathrm{F}. &
3398:   \left(C_{15}\right)^{\gamma\chi}_{\alpha\sigma}
3399:   \left(C_{15}\right)^{\beta\sigma}_{\rho\chi}
3400:   &=\;\;
3401:   \delta^\gamma_\rho \delta^\alpha_\beta
3402:   -\frac{1}{4}\delta^\gamma_\alpha \delta^\beta_\rho
3403: \end{eqnarray}
3404: 
3405: And the six \SU{2} structures are:
3406: \begin{eqnarray}
3407:   \nonumber
3408:   \mathrm{a}. &
3409:   \left(R_1\right)^z_w
3410:   \left(R_1\right)^x_y 
3411:   &= \;\;
3412:   \delta^z_w \delta^x_y
3413:   \\
3414:   \nonumber
3415:   \mathrm{b}. &
3416:   \left(R_3\right)^{zq}_{wr}
3417:   \left(R_3\right)^{xr}_{yq} 
3418:   &= \;\;
3419:   \delta^x_w \delta^z_y - \frac{1}{2}\delta^x_y\delta^z_w
3420:   \\
3421:   \nonumber
3422:   \mathrm{c}. &
3423:   \epsilon^{xz}\epsilon_{yw} 
3424:   &= \;\;
3425:   \epsilon^{xz}\epsilon_{yw}
3426:   \\
3427:   \nonumber
3428:   \mathrm{d}. &
3429:   \epsilon_{ws}\epsilon^{xt}
3430:   \left(R_3\right)^{sq}_{yr}
3431:   \left(R_3\right)^{zr}_{tq} 
3432:   &= \;\;
3433:   \delta^x_w\delta^z_y - \frac{1}{2}\epsilon_{wy}\epsilon^{xz}
3434:   \\
3435:   \nonumber 
3436:   \mathrm{e}. &
3437:   \left(R_1\right)^z_y 
3438:   \left(R_1\right)^x_w  
3439:   &= \;\;
3440:   \delta^z_y \delta^x_w
3441:   \\
3442:   \label{eq:su2_structurres}
3443:   \mathrm{f}. &
3444:   \left(R_3\right)^{zq}_{yr}
3445:   \left(R_3\right)^{xr}_{wq} 
3446:   &= \;\;
3447:   \delta^x_y\delta^z_w - \frac{1}{2}\delta^x_w\delta^z_y
3448: \end{eqnarray}
3449: 
3450: All possible $n=1$ operators were then named $O^A ... O^Z O^a...O^j$
3451: in \cite{King:OperatorAnalysis}. We rename them here in a manner
3452: consistent with the $n>1$ operators $O^{(n')}$, so that the names are
3453: $O^{\Pi\pi}$ where $\Pi$ is the \SU{4} structure and $\pi$ is the
3454: SU{2} structure.  See Table \ref{tab:operator_names} for the
3455: translation into the names of Ref.\cite{King:OperatorAnalysis} and the
3456: CGCs.
3457: 
3458: All of these operators are operators for the case without a \UI{}
3459: family symmetry. In the case when there is, we follow the
3460: prescription
3461: \begin{equation}
3462:   \label{eq:u1isation_of_operators}
3463:   \mathcal{O}_{IJ}\rightarrow \mathcal{O}_{IJ}
3464:   \left(\frac{\theta}{M_X}\right) ^{p_{_{IJ}}} .
3465: \end{equation}
3466: 
3467: Where $p_{IJ} = |X_{\mathcal{O}_{IJ}}|$ is the modulus of the charge
3468: of the operator. If the charge of the operator is negative, then the
3469: field $\theta$ should be replaced by the field
3470: $\overline{\theta}$. The prescription makes the operator chargeless
3471: under the $\UI_F$ while simultaneously not changing the dimension.
3472: 
3473: %%% ----------------------------------------------------------------------
3474: 
3475: \section{$n>1$ operators}
3476: \label{sec:n1-operators}
3477: 
3478: %%% ----------------------------------------------------------------------
3479: 
3480: In the case that $n > 1$, there will be more indicies to contract,
3481: which allows more representations, and hence more Clebsch
3482: coefficients. To generalise the notation, it is necessary only to
3483: construct the new tensors which create the new structures. However, it
3484: will always be possible to contract the new indicies between the $H$
3485: and $\overline{H}$ fields to create a singlet $H\overline{H}$ which
3486: has a Clebsch of 1 in each sector $u,d,e,\nu$. In this case, the first
3487: structures are the same as the old structures, but with extra $\delta$
3488: symbols which construct the $H\overline{H}$ singlet.
3489: 
3490: Thus taking an $n=2$ operator, say $\mathcal{O}'^{Fb}$, which forms a
3491: representation that could have been attained by a $n=1$ operator, the
3492: Clebsch coefficients are the same. This is what we mean by
3493: $\mathcal{O}^{n\prime\;\Pi\pi}$, as we have only used $n>1$
3494: coefficients which are in the subset that have $n=1$ analogues.
3495: 
3496: 
3497: 
3498: %%% ----------------------------------------------------------------------
3499: 
3500: % \begin{thebibliography}{99}
3501: % \bibitem{Borzumati:1986qx}
3502: % F.~Borzumati and A.~Masiero,
3503: % Phys.\ Rev.\ Lett.\  {\bf 57} (1986) 961.
3504: % \bibitem{Gabbiani:1996hi}
3505: % F.~Gabbiani, E.~Gabrielli, A.~Masiero and L.~Silvestrini,
3506: % Nucl.\ Phys.\ B {\bf 477} (1996) 321
3507: % [arXiv:hep-ph/9604387].
3508: % \bibitem{Hisano:1995cp}
3509: % J.~Hisano, T.~Moroi, K.~Tobe and M.~Yamaguchi,
3510: % Phys.\ Rev.\ D {\bf 53} (1996) 2442
3511: % [arXiv:hep-ph/9510309].
3512: % \bibitem{King:1998nv}
3513: % S.~F.~King and M.~Oliveira,
3514: % Phys.\ Rev.\ D {\bf 60} (1999) 035003
3515: % [arXiv:hep-ph/9804283].
3516: % \bibitem{GoPe}
3517: % G.~L.~Fogli, E.~Lisi, A.~Marrone, D.~Montanino, A.~Palazzo and A.~M.~Rotunno,
3518: % arXiv:hep-ph/0212127;
3519: % P.~C.~de~Holanda and A.~Y.~Smirnov,
3520: % arXiv:hep-ph/0212270;
3521: % V.~Barger and D.~Marfatia,
3522: % Phys.\ Lett.\ B {\bf 555} (2003) 144
3523: % [arXiv:hep-ph/0212126];
3524: % A.~Bandyopadhyay, S.~Choubey, R.~Gandhi, S.~Goswami and D.~P.~Roy,
3525: % arXiv:hep-ph/0212146.
3526: % M.~Maltoni, T.~Schwetz and J.~W.~Valle,
3527: % arXiv:hep-ph/0212129.
3528: % \bibitem{SKamiokandeColl}
3529: %                Y. Fukuda {\it et al.}, Super-Kamiokande Collaboration,
3530: %                Phys. Lett. {\bf B433}, 9 (1998);
3531: %         {\it ibid.}\ Phys. Lett. {\bf B436}, 33 (1998);
3532: %         {\it ibid.}\ Phys. Rev. Lett. {\bf 81}, 1562 (1998).
3533: % \bibitem{Blazek:2002wq}
3534: % T.~Blazek and S.~F.~King,
3535: % arXiv:hep-ph/0211368.
3536: % \bibitem{huge}
3537: % S.~Davidson and A.~Ibarra,
3538: % JHEP {\bf 0109} (2001) 013;
3539: % J.~Hisano and D.~Nomura,
3540: % Phys.\ Rev.\ D {\bf 59} (1999) 116005;
3541: % J.~Hisano,
3542: % arXiv:hep-ph/0204100;
3543: % J.~A.~Casas and A.~Ibarra,
3544: % Nucl.\ Phys.\ B {\bf 618} (2001) 171;
3545: % W.~Buchm\"uller, D.~Delepine and F.~Vissani,
3546: % Phys.\ Lett.\ B {\bf 459} (1999) 171;
3547: % M.~E.~Gomez, G.~K.~Leontaris, S.~Lola and J.~D.~Vergados,
3548: % Phys.\ Rev.\ D {\bf 59} (1999) 116009;
3549: % J.~R.~Ellis, M.~E.~Gomez, G.~K.~Leontaris, S.~Lola and D.~V.~Nanopoulos,
3550: % Eur.\ Phys.\ J.\ C {\bf 14} (2000) 319;
3551: % W.~Buchm\"uller, D.~Delepine and L.~T.~Handoko,
3552: % Nucl.\ Phys.\ B {\bf 576} (2000) 445;
3553: % D. Carvalho, J. Ellis, M. Gomez and S. Lola,
3554: % Phys.\ Lett.\ B {\bf 515} (2001) 323;
3555: % F.~Deppisch, H.~Pas, A.~Redelbach, R.~Ruckl and Y.~Shimizu,
3556: % arXiv:hep-ph/0206122;
3557: % J.~Sato and K.~Tobe,
3558: % Phys.\ Rev.\ D {\bf 63} (2001) 116010;
3559: % J.~Hisano and K.~Tobe,
3560: % Phys.\ Lett.\ B {\bf 510} (2001) 197;
3561: % J.~R.~Ellis, J.~Hisano, M.~Raidal and Y.~Shimizu,
3562: % Phys.\ Lett.\ B {\bf 528} (2002) 86,
3563: % arXiv:hep-ph/0111324;
3564: % J.~R.~Ellis, J.~Hisano, S.~Lola and M.~Raidal,
3565: % Nucl.\ Phys.\ B {\bf 621} (2002) 208,
3566: % arXiv:hep-ph/0109125;
3567: % J.~Hisano, T.~Moroi, K.~Tobe and M.~Yamaguchi,
3568: % Phys.\ Lett.\ B {391} (1997) 341; 
3569: % [Erratum - {\it ibid.} {\bf 397}, 357 (1997)];
3570: % J.~Hisano, D.~Nomura, Y.~Okada, Y.~Shimizu and M.~Tanaka,
3571: % Phys.\ Rev.\ D {\bf 58} (1998) 116010;
3572: % J.~Hisano, D.~Nomura and T.~Yanagida,
3573: % Phys.\ Lett.\ B {\bf 437} (1998) 351;
3574: % S.~Lavignac, I.~Masina and C.~A.~Savoy,
3575: % Phys.\ Lett.\ B {\bf 520} (2001) 269
3576: % [arXiv:hep-ph/0106245];
3577: % S.~Lavignac, I.~Masina and C.~A.~Savoy,
3578: % Nucl.\ Phys.\ B {\bf 633} (2002) 139
3579: % [arXiv:hep-ph/0202086];
3580: % I.~Masina and C.~A.~Savoy,
3581: % arXiv:hep-ph/0211283.
3582: % S.~Pascoli, S.~T.~Petcov and C.~E.~Yaguna,
3583: % Phys.\ Lett.\ B {\bf 564} (2003) 241
3584: % [arXiv:hep-ph/0301095].
3585: % S.~Pascoli, S.~T.~Petcov and W.~Rodejohann,
3586: % arXiv:hep-ph/0302054.
3587: % \bibitem{Abel:2001ur}
3588: % S.~A.~Abel and G.~Servant,
3589: % Nucl.\ Phys.\ B {\bf 611} (2001) 43
3590: % [arXiv:hep-ph/0105262].
3591: % \bibitem{Ross:2002mr}
3592: % G.~G.~Ross and O.~Vives,
3593: % Phys.\ Rev.\ D {\bf 67} (2003) 095013
3594: % [arXiv:hep-ph/0211279].
3595: % \bibitem{KP0307091}
3596: % S.~F.~King, I.~N.~R.~Peddie,
3597: % Nucl.\ Phys.\ B {\bf 678} (2004) 339
3598: % [arXiv:hep-ph/0307091].
3599: % \bibitem{P_Matrix:introduction}
3600: % S.~A.~Abel, B.~C.~Allanach, F.~Quevedo, L.~Ibanez and M.~Klein,
3601: % JHEP {\bf 0012} (2000) 026
3602: % [arXiv:hep-ph/0005260].
3603: % B.~C.~Allanach, D.~Grellscheid and F.~Quevedo,
3604: % JHEP {\bf 0205} (2002) 048
3605: % [arXiv:hep-ph/0111057].
3606: % \bibitem{Brignole:1997dp}
3607: % A.~Brignole, L.~E.~Ibanez and C.~Munoz,
3608: % [arXiv:hep-ph/9707209].
3609: % \bibitem{Ibanez:1998rf}
3610: % L.~E.~Ibanez, C.~Munoz and S.~Rigolin,
3611: % Nucl.\ Phys.\ B {\bf 553} (1999) 43
3612: % [arXiv:hep-ph/9812397].
3613: % \bibitem{Chung:2003fi}
3614: % D.~J.~H.~Chung, L.~L.~Everett, G.~L.~Kane, S.~F.~King, J.~Lykken and L.~T.~Wang,
3615: % Phys.\ Rept.\  {\bf 407} (2005) 1
3616: % [arXiv:hep-ph/0312378].
3617: % \bibitem{King:2003xn}
3618: % S.~F.~King and I.~N.~R.~Peddie,
3619: % J.\ Korean Phys.\ Soc.\  {\bf 45} (2004) S443
3620: % [arXiv:hep-ph/0312235].
3621: % \bibitem{Everett:2002pm}
3622: % L.~L.~Everett, G.~L.~Kane, S.~F.~King, S.~Rigolin and L.~T.~Wang,
3623: % Phys.\ Lett.\ B {\bf 531} (2002) 263
3624: % [arXiv:hep-ph/0202100].
3625: % \bibitem{Blazek:2003wz}
3626: % T.~Blazek, S.~F.~King and J.~K.~Parry,
3627: % arXiv:hep-ph/0303192.
3628: % \bibitem{King:OperatorAnalysis}
3629: % S.~F.~King,
3630: % Phys.\ Lett.\ B {\bf 325} (1994) 129
3631: % [Erratum-ibid.\ B {\bf 325} (1994) 538];
3632: % B.~C.~Allanach and S.~F.~King,
3633: % Nucl.\ Phys.\ B {\bf 456} (1995) 57
3634: % [arXiv:hep-ph/9502219];
3635: % B.~C.~Allanach and S.~F.~King,
3636: % Nucl.\ Phys.\ B {\bf 459} (1996) 75
3637: % [arXiv:hep-ph/9509205];
3638: % B.~C.~Allanach, S.~F.~King, G.~K.~Leontaris and S.~Lola,
3639: % Phys.\ Rev.\ D {\bf 56} (1997) 2632
3640: % [arXiv:hep-ph/9610517];
3641: % S.~F.~King and M.~Oliveira,
3642: % Phys.\ Rev.\ D {\bf 63} (2001) 095004
3643: % [arXiv:hep-ph/0009287]
3644: % \bibitem{Froggatt:1978nt}
3645: % C.~D.~Froggatt and H.~B.~Nielsen,
3646: % Nucl.\ Phys.\ B {\bf 147} (1979) 277.
3647: % \bibitem{King:2000vp}
3648: % S.~F.~King and M.~Oliveira,
3649: % Phys.\ Rev.\ D {\bf 63} (2001) 015010
3650: % [arXiv:hep-ph/0008183].
3651: % \bibitem{Kane:2005va}
3652: %   G.~L.~Kane, S.~F.~King, I.~N.~R.~Peddie and L.~Velasco-Sevilla,
3653: %   arXiv:hep-ph/0504038.
3654: % \bibitem{Abel:2001cv}
3655: % S.~Abel, S.~Khalil and O.~Lebedev,
3656: % Phys.\ Rev.\ Lett.\  {\bf 89} (2002) 121601
3657: % [arXiv:hep-ph/0112260].
3658: % \bibitem{Hagiwara:fs}
3659: % K.~Hagiwara {\it et al.}  [Particle Data Group Collaboration],
3660: % Phys.\ Rev.\ D {\bf 66} (2002) 010001.
3661: % \bibitem{Allanach:2001kg}
3662: % B.~C.~Allanach,
3663: % Comput.\ Phys.\ Commun.\  {\bf 143} (2002) 305
3664: % [arXiv:hep-ph/0104145].
3665: % \end{thebibliography}
3666: 
3667: %%% Correctly ordered bibliography as generated by sort2
3668: %%% (c) I Peddie 2005. This is released under the GNU Public licence
3669: %%% See www.gnu.org for details. 
3670: 
3671: 
3672: \begin{thebibliography}{99}
3673: \bibitem{Borzumati:1986qx}
3674: F.~Borzumati and A.~Masiero,
3675: Phys.\ Rev.\ Lett.\  {\bf 57} (1986) 961.
3676: \bibitem{Gabbiani:1996hi}
3677: F.~Gabbiani, E.~Gabrielli, A.~Masiero and L.~Silvestrini,
3678: Nucl.\ Phys.\ B {\bf 477} (1996) 321
3679: [arXiv:hep-ph/9604387].
3680: \bibitem{Chung:2003fi}
3681: D.~J.~H.~Chung, L.~L.~Everett, G.~L.~Kane, S.~F.~King, J.~Lykken and L.~T.~Wang,
3682: Phys.\ Rept.\  {\bf 407} (2005) 1
3683: [arXiv:hep-ph/0312378].
3684: \bibitem{GoPe}
3685: G.~L.~Fogli, E.~Lisi, A.~Marrone, D.~Montanino, A.~Palazzo and A.~M.~Rotunno,
3686: arXiv:hep-ph/0212127;
3687: P.~C.~de~Holanda and A.~Y.~Smirnov,
3688: arXiv:hep-ph/0212270;
3689: V.~Barger and D.~Marfatia,
3690: Phys.\ Lett.\ B {\bf 555} (2003) 144
3691: [arXiv:hep-ph/0212126];
3692: A.~Bandyopadhyay, S.~Choubey, R.~Gandhi, S.~Goswami and D.~P.~Roy,
3693: arXiv:hep-ph/0212146.
3694: M.~Maltoni, T.~Schwetz and J.~W.~Valle,
3695: arXiv:hep-ph/0212129.
3696: \bibitem{SKamiokandeColl}
3697:                Y. Fukuda {\it et al.}, Super-Kamiokande Collaboration,
3698:                Phys. Lett. {\bf B433}, 9 (1998);
3699:         {\it ibid.}\ Phys. Lett. {\bf B436}, 33 (1998);
3700:         {\it ibid.}\ Phys. Rev. Lett. {\bf 81}, 1562 (1998).
3701: \bibitem{King:2003jb}
3702:   For a review see:
3703:   S.~F.~King,
3704:   Rept.\ Prog.\ Phys.\  {\bf 67}, 107 (2004)
3705:   [arXiv:hep-ph/0310204].
3706: \bibitem{Hisano:1995cp}
3707: J.~Hisano, T.~Moroi, K.~Tobe and M.~Yamaguchi,
3708: Phys.\ Rev.\ D {\bf 53} (1996) 2442
3709: [arXiv:hep-ph/9510309].
3710: \bibitem{King:1998nv}
3711: S.~F.~King and M.~Oliveira,
3712: Phys.\ Rev.\ D {\bf 60} (1999) 035003
3713: [arXiv:hep-ph/9804283].
3714: \bibitem{Blazek:2002wq}
3715: T.~Blazek and S.~F.~King,
3716: arXiv:hep-ph/0211368.
3717: \bibitem{huge}
3718: S.~Davidson and A.~Ibarra,
3719: JHEP {\bf 0109} (2001) 013;
3720: J.~Hisano and D.~Nomura,
3721: Phys.\ Rev.\ D {\bf 59} (1999) 116005;
3722: J.~Hisano,
3723: arXiv:hep-ph/0204100;
3724: J.~A.~Casas and A.~Ibarra,
3725: Nucl.\ Phys.\ B {\bf 618} (2001) 171;
3726: W.~Buchm\"uller, D.~Delepine and F.~Vissani,
3727: Phys.\ Lett.\ B {\bf 459} (1999) 171;
3728: M.~E.~Gomez, G.~K.~Leontaris, S.~Lola and J.~D.~Vergados,
3729: Phys.\ Rev.\ D {\bf 59} (1999) 116009;
3730: J.~R.~Ellis, M.~E.~Gomez, G.~K.~Leontaris, S.~Lola and D.~V.~Nanopoulos,
3731: Eur.\ Phys.\ J.\ C {\bf 14} (2000) 319;
3732: W.~Buchm\"uller, D.~Delepine and L.~T.~Handoko,
3733: Nucl.\ Phys.\ B {\bf 576} (2000) 445;
3734: D. Carvalho, J. Ellis, M. Gomez and S. Lola,
3735: Phys.\ Lett.\ B {\bf 515} (2001) 323;
3736: F.~Deppisch, H.~Pas, A.~Redelbach, R.~Ruckl and Y.~Shimizu,
3737: arXiv:hep-ph/0206122;
3738: J.~Sato and K.~Tobe,
3739: Phys.\ Rev.\ D {\bf 63} (2001) 116010;
3740: J.~Hisano and K.~Tobe,
3741: Phys.\ Lett.\ B {\bf 510} (2001) 197;
3742: J.~R.~Ellis, J.~Hisano, M.~Raidal and Y.~Shimizu,
3743: Phys.\ Lett.\ B {\bf 528} (2002) 86,
3744: arXiv:hep-ph/0111324;
3745: J.~R.~Ellis, J.~Hisano, S.~Lola and M.~Raidal,
3746: Nucl.\ Phys.\ B {\bf 621} (2002) 208,
3747: arXiv:hep-ph/0109125;
3748: J.~Hisano, T.~Moroi, K.~Tobe and M.~Yamaguchi,
3749: Phys.\ Lett.\ B {391} (1997) 341; 
3750: [Erratum - {\it ibid.} {\bf 397}, 357 (1997)];
3751: J.~Hisano, D.~Nomura, Y.~Okada, Y.~Shimizu and M.~Tanaka,
3752: Phys.\ Rev.\ D {\bf 58} (1998) 116010;
3753: J.~Hisano, D.~Nomura and T.~Yanagida,
3754: Phys.\ Lett.\ B {\bf 437} (1998) 351;
3755: S.~Lavignac, I.~Masina and C.~A.~Savoy,
3756: Phys.\ Lett.\ B {\bf 520} (2001) 269
3757: [arXiv:hep-ph/0106245];
3758: S.~Lavignac, I.~Masina and C.~A.~Savoy,
3759: Nucl.\ Phys.\ B {\bf 633} (2002) 139
3760: [arXiv:hep-ph/0202086];
3761: I.~Masina and C.~A.~Savoy,
3762: arXiv:hep-ph/0211283.
3763: S.~Pascoli, S.~T.~Petcov and C.~E.~Yaguna,
3764: Phys.\ Lett.\ B {\bf 564} (2003) 241
3765: [arXiv:hep-ph/0301095].
3766: S.~Pascoli, S.~T.~Petcov and W.~Rodejohann,
3767: arXiv:hep-ph/0302054.
3768: \bibitem{Chankowski:2005jh}
3769:   P.~H.~Chankowski, O.~Lebedev and S.~Pokorski,
3770:   Nucl.\ Phys.\ B {\bf 717} (2005) 190
3771:   [arXiv:hep-ph/0502076].
3772: \bibitem{Abel:2001ur}
3773: S.~A.~Abel and G.~Servant,
3774: Nucl.\ Phys.\ B {\bf 611} (2001) 43
3775: [arXiv:hep-ph/0105262].
3776: \bibitem{Ross:2002mr}
3777: G.~G.~Ross and O.~Vives,
3778: Phys.\ Rev.\ D {\bf 67} (2003) 095013
3779: [arXiv:hep-ph/0211279].
3780: \bibitem{KP0307091}
3781: S.~F.~King, I.~N.~R.~Peddie,
3782: Nucl.\ Phys.\ B {\bf 678} (2004) 339
3783: [arXiv:hep-ph/0307091].
3784: \bibitem{P_Matrix:introduction}
3785: S.~A.~Abel, B.~C.~Allanach, F.~Quevedo, L.~Ibanez and M.~Klein,
3786: JHEP {\bf 0012} (2000) 026
3787: [arXiv:hep-ph/0005260].
3788: B.~C.~Allanach, D.~Grellscheid and F.~Quevedo,
3789: JHEP {\bf 0205} (2002) 048
3790: [arXiv:hep-ph/0111057].
3791: \bibitem{Brignole:1997dp}
3792: A.~Brignole, L.~E.~Ibanez and C.~Munoz,
3793: [arXiv:hep-ph/9707209].
3794: \bibitem{Ibanez:1998rf}
3795: L.~E.~Ibanez, C.~Munoz and S.~Rigolin,
3796: Nucl.\ Phys.\ B {\bf 553} (1999) 43
3797: [arXiv:hep-ph/9812397].
3798: \bibitem{King:2003xn}
3799: S.~F.~King and I.~N.~R.~Peddie,
3800: J.\ Korean Phys.\ Soc.\  {\bf 45} (2004) S443
3801: [arXiv:hep-ph/0312235].
3802: \bibitem{Everett:2002pm}
3803: L.~L.~Everett, G.~L.~Kane, S.~F.~King, S.~Rigolin and L.~T.~Wang,
3804: Phys.\ Lett.\ B {\bf 531} (2002) 263
3805: [arXiv:hep-ph/0202100].
3806: \bibitem{Blazek:2003wz}
3807: T.~Blazek, S.~F.~King and J.~K.~Parry,
3808: arXiv:hep-ph/0303192.
3809: \bibitem{King:OperatorAnalysis}
3810: S.~F.~King,
3811: Phys.\ Lett.\ B {\bf 325} (1994) 129
3812: [Erratum-ibid.\ B {\bf 325} (1994) 538];
3813: B.~C.~Allanach and S.~F.~King,
3814: Nucl.\ Phys.\ B {\bf 456} (1995) 57
3815: [arXiv:hep-ph/9502219];
3816: B.~C.~Allanach and S.~F.~King,
3817: Nucl.\ Phys.\ B {\bf 459} (1996) 75
3818: [arXiv:hep-ph/9509205];
3819: B.~C.~Allanach, S.~F.~King, G.~K.~Leontaris and S.~Lola,
3820: Phys.\ Rev.\ D {\bf 56} (1997) 2632
3821: [arXiv:hep-ph/9610517];
3822: S.~F.~King and M.~Oliveira,
3823: Phys.\ Rev.\ D {\bf 63} (2001) 095004
3824: [arXiv:hep-ph/0009287]
3825: \bibitem{Froggatt:1978nt}
3826: C.~D.~Froggatt and H.~B.~Nielsen,
3827: Nucl.\ Phys.\ B {\bf 147} (1979) 277.
3828: \bibitem{King:2000vp}
3829: S.~F.~King and M.~Oliveira,
3830: Phys.\ Rev.\ D {\bf 63} (2001) 015010
3831: [arXiv:hep-ph/0008183].
3832: \bibitem{Kane:2005va}
3833:   G.~L.~Kane, S.~F.~King, I.~N.~R.~Peddie and L.~Velasco-Sevilla,
3834:   arXiv:hep-ph/0504038.
3835: \bibitem{Abel:2001cv}
3836: S.~Abel, S.~Khalil and O.~Lebedev,
3837: Phys.\ Rev.\ Lett.\  {\bf 89} (2002) 121601
3838: [arXiv:hep-ph/0112260].
3839: \bibitem{Hagiwara:fs}
3840: K.~Hagiwara {\it et al.}  [Particle Data Group Collaboration],
3841: Phys.\ Rev.\ D {\bf 66} (2002) 010001.
3842: \bibitem{Aubert:2005ye}
3843: B.~Aubert {\it et al.}  [BABAR Collaboration],
3844: Phys.\ Rev.\ Lett.\  {\bf 95} (2005) 041802
3845: [arXiv:hep-ex/0502032].
3846: \bibitem{Allanach:2001kg}
3847: B.~C.~Allanach,
3848: Comput.\ Phys.\ Commun.\  {\bf 143} (2002) 305
3849: [arXiv:hep-ph/0104145].
3850: \bibitem{Kane:1998im}
3851: G.~L.~Kane and S.~F.~King, Phys.\ Lett.\ B {\bf 451} (1999) 113
3852: [arXiv:hep-ph/9810374].
3853: 
3854: 
3855: \end{thebibliography}
3856: 
3857: 
3858: 
3859: 
3860: %%% ----------------------------------------------------------------------
3861: %%% ----------------------------------------------------------------------
3862: 
3863: \end{document}
3864: