1: \documentclass[11pt,dvips]{article}
2: \usepackage[dvips]{graphicx}
3: \usepackage{color}
4: \usepackage{amsmath,amssymb,amsfonts}
5:
6: \newcommand{\s}{\,{\rm s}}
7: \newcommand{\GeV}{\,{\rm GeV}}
8: \newcommand{\keV}{\,{\rm keV}}
9: \newcommand{\eV}{\,{\rm eV}}
10: \newcommand{\Mpc}{\,{\rm Mpc}}
11: \newcommand{\fex}{{\it e.g.} }
12: \newcommand{\comp}{{\it cf.} }
13: \newcommand{\ppp}[1]{\begin{pmatrix} #1
14: \end{pmatrix}}
15:
16: \title{\vspace{-1.5cm}
17: {\normalsize\rightline{DESY 08-128}\rightline{TUM-HEP 699/08}}\ \vskip 1cm
18: \bf\boldmath
19: Hidden gauginos of an unbroken $U(1)$: Cosmological constraints and phenomenological prospects
20: \vspace{11mm}}
21:
22: \author{A.~Ibarra$^{a,b}$, A.~Ringwald$^{b}$, C.~Weniger$^{b}$\\[4mm]
23: {\normalsize\it a Physik Department T30, Technische Universit\"at M\"unchen,}\\[-0.05cm]
24: {\it\normalsize James-Franck-Strasse, 85748 Garching, Germany.}\\[2mm]
25: {\normalsize\it b Deutsches Elektronen-Synchrotron DESY,}\\
26: {\it\normalsize Notkestrasse 85, 22607 Hamburg, Germany.}\\[2mm]
27: \date{\empty}
28: }
29:
30: \begin{document}
31: \begin{titlepage}
32: \maketitle
33: \begin{abstract}
34: We study supersymmetric scenarios where the dark matter is the gaugino of an
35: unbroken hidden $U(1)$ which interacts with the visible world only via a
36: small kinetic mixing with the hypercharge. Strong constraints on the
37: parameter space can be derived from avoiding overclosure of the Universe and
38: from requiring successful Big Bang Nucleosynthesis and structure formation.
39: We find that for typical values of the mixing parameter, scenarios with
40: neutralino NLSP are excluded, while scenarios with slepton NLSP are allowed
41: when the mixing parameter lies in the range $\chi\sim\mathcal{O}
42: (10^{-13}-10^{-10})$. We also show that if the gravitino is the LSP and the
43: hidden $U(1)$ gaugino the NLSP, the bounds on the reheating temperature from
44: long lived charged MSSM relics can be considerably relaxed and we comment on
45: the signatures of these scenarios at future colliders. Finally, we discuss
46: the case of an anomalously small mixing, $\chi\ll 10^{-16}$, where the
47: neutralino becomes a decaying dark matter candidate, and derive constraints
48: from gamma ray experiments.
49: \end{abstract}
50:
51: \thispagestyle{empty}
52: \end{titlepage}
53: \newpage \setcounter{page}{2}
54:
55: \section{Introduction}
56: The existence of dark matter in the Universe is perhaps the most solid
57: indication for physics beyond the Standard Model of particle
58: physics~\cite{BHS05}. Among the many extensions of the Standard Model that have
59: been proposed in recent years, supersymmetry (SUSY) arguably remains as the
60: most popular. Among other merits, supersymmetric scenarios provide a very
61: promising candidate for the dark matter: the lightest supersymmetric particle
62: (LSP)~\cite{EHN+84,JKG96}. With the particle content of the Minimal
63: Supersymmetric Standard Model (MSSM), the LSP can be, over a large range of
64: parameters, either the lightest neutralino, the lightest sneutrino or the
65: lightest stau. Among these, only the lightest neutralino is still allowed by
66: present experiments as a viable dark matter candidate, provided $R$-parity is
67: almost exactly conserved. Furthermore, if supersymmetry is promoted to a local
68: symmetry, the particle content of the MSSM also includes the gravitational
69: supermultiplet, of which the spin $3/2$ component, the gravitino, is also a
70: viable dark matter candidate~\cite{PP82}, even if $R$-parity is slightly
71: violated~\cite{TY00,BCH+07}.
72:
73: On the other hand, many extensions of the MSSM contemplate the possibility of a
74: hidden sector, consisting of superfields which are singlets under the Standard
75: Model gauge group. Hidden sector superfields usually couple very weakly to our
76: observable sector, thus constituting a very natural arena for finding dark
77: matter candidates. Generically, hidden sector particles couple to our
78: observable sector only via non-renormalizable operators, presumably suppressed
79: by a large mass scale, with a structure that is strongly model dependent. In
80: consequence, deriving implications of the hidden sector dark matter on the
81: thermal history of the observable Universe and for future collider experiments
82: is hindered by our complete ignorance of the strength and the structure of the
83: hidden sector interactions with our observable sector.
84:
85: There are however three instances where the hidden sector particles can couple
86: to the MSSM particles via renormalizable operators, with a structure which is
87: well defined by the Lorentz and gauge symmetries. Firstly, a hidden sector
88: chiral superfield, $S$, could couple to the lepton and up-type Higgs doublet
89: superfields via the Yukawa coupling $S H_u L$ in the superpotential, or to the
90: two MSSM Higgs doublets, via $S H_u H_d$, provided these terms are also
91: invariant under the hidden sector gauge group (as well as possible discrete and
92: global symmetries of the theory). Secondly, if one of the
93: MSSM chiral superfields is charged under a hidden sector gauge group, it will
94: interact with the corresponding hidden sector gauge superfield, and in turn
95: with other hidden sector chiral superfields via the $D$-term. Finally, a
96: hidden sector abelian vector superfield, $X$, may couple to the hypercharge
97: vector superfield through a kinetic mixing term, which is always allowed by the
98: gauge symmetries \cite{Holdom86, FH91}.
99:
100: In this paper we will concentrate on the last situation and we will study
101: scenarios where the hidden $U(1)$ gauge group is unbroken at low energies. If
102: this is the case, the corresponding hidden $U(1)$ gaugino will have a mass
103: comparable to the typical soft SUSY breaking masses of the observable sector
104: particles and, in some instances, smaller. Therefore, the hidden $U(1)$ gaugino
105: will not only be the lightest supersymmetric particle of the hidden sector, but
106: also the lightest among all the supersymmetric particles.
107:
108: Our motivation to consider an unbroken hidden sector $U(1)$ group is twofold.
109: First, in string theory compactifications hidden sector $U(1)$ groups are
110: ubiquitous and some of them could remain unbroken at low energies, in complete
111: analogy to the familiar electromagnetic $U(1)$ of our observable sector.
112: Secondly, the case of the hidden sector unbroken $U(1)$ is particularly
113: intriguing, since this situation is practically unconstrained by present
114: experiments. Indeed, it was shown long ago by Holdom that, in a
115: non-supersymmetric world, the hidden $U(1)$ gauge boson (the ``paraphoton''
116: \cite{Okun82}) completely decouples from the observable sector \cite{Holdom91}.
117: This result can be also generalized to a supersymmetric theory. Let us consider
118: the SUSY invariant part of the Lagrangian,
119: \begin{eqnarray}
120: \nonumber
121: {\cal L}&=&\int d^2\theta \, \left(\hat{W}_B^\alpha \hat{W}_{B\,\alpha}+
122: \hat{W}_X^\alpha \hat{W}_{X\,\alpha}+2\chi \hat{W}_B^\alpha
123: \hat{W}_{X\,\alpha}\right)+ \text{h.c.}+\\
124: &&+\int d^2\theta d^2\bar \theta \, \left(\Phi^\dagger e^{Y g_Y \hat{B}}
125: \Phi+ h^\dagger e^{q g_X \hat{X}} h\right)\;,
126: \label{eqn:Lagrangian1}
127: \end{eqnarray}
128: where the field strength superfield is defined as
129: $\hat{W}_V^\alpha=-\frac{1}{4} {\bar D}{\bar D} D^\alpha \hat{V}$,
130: $\hat{V}=\hat{B}, \hat{X}$ being the hypercharge or the hidden $U(1)$ vector
131: superfield, while $\Phi$ and $h$ denote, respectively, any Standard Model or
132: hidden sector chiral superfield. Finally, $\chi$ is the kinetic mixing
133: parameter, which is induced through quantum effects by chiral superfields
134: charged under both gauge groups. Without additional symmetries, values around
135: $\chi\sim 10^{-3} - 10^{-4}$ are naturally obtained. However, \fex in
136: compactifications of heterotic \cite{DKM97, BMW06} and type II \cite{LS07,
137: AS04, BHK05, AJK+08, AGJ+08} strings, much smaller mixings are possible. A
138: lower bound around $\chi \gtrsim 10^{-16}$ was argued to hold in cases of gauge
139: mediated supersymmetry breaking in heterotic string models \cite{DKM97},
140: whereas in type II models with warped extra dimensions the kinetic mixing
141: parameter can be exponentially small \cite{AGJ+08}.
142:
143: The gauge kinetic terms in Eq.~\eqref{eqn:Lagrangian1} can be made canonical by
144: introducing shifted vector superfields,
145: \begin{eqnarray}
146: X&=& \hat{X}+ \chi \hat{B}\;, \\
147: B&=& \sqrt{1- \chi^2} \hat{B}\;,
148: \end{eqnarray}
149: leading to
150: \begin{eqnarray}
151: {\cal L}&=&\int d^2\theta \, \left(W_B^\alpha W_{B\,\alpha}+
152: W_X^\alpha W_{X\,\alpha}\right) + \text{h.c.}+ \\\nonumber
153: &&+\int d^2\theta d^2\bar \theta \, \left(\Phi^\dagger
154: e^{Y g^\prime_Y B} \Phi+
155: h^\dagger e^{q g_X X- q g^\prime_X B} h\right)\;,
156: \end{eqnarray}
157: where $g^\prime_Y= g_Y/\sqrt{1-\chi^2}$ and $g^\prime_X= \chi
158: g_X/\sqrt{1-\chi^2}$. Therefore, the canonical normalization of the kinetic
159: terms produces an unobservable shift of the hypercharge gauge coupling and the
160: generation of a ``minihypercharge'' for the hidden sector chiral superfields
161: \cite{Holdom86}. Different astrophysical observations and laboratory
162: experiments constrain the possible values of the minihypercharge and the masses
163: of the hidden sector particles. For instance, one obtains $\chi \lesssim
164: 10^{-13}$ for masses below $10\keV$ from plasmon decay in red giants
165: \cite{DP94, DHR00} (see also Ref.~\cite{MPS07}). Nevertheless these bounds can
166: be automatically avoided if the masses are large. In this case, as long as
167: supersymmetry remains unbroken, the hypercharge vector superfield completely
168: decouples from the observable sector and is not subject to any experimental
169: constraint.
170:
171: The breaking of supersymmetry changes dramatically the previous picture.
172: Although the hidden $U(1)$ gauge boson remains decoupled from the observable
173: sector, we will show that in the presence of SUSY breaking effects a mixing
174: between the hidden $U(1)$ gaugino and the MSSM neutralinos is induced. Then,
175: the unbroken hidden $U(1)$ might produce observable effects in the cosmological
176: evolution of the Universe, at collider experiments or in cosmic ray fluxes.
177:
178: Models with a hidden $U(1)$ extension of the SM or MSSM have been extensively
179: studied in the literature (see \fex \cite{CNW06}, for a recent review see
180: Ref.~\cite{Langacker08}). Some of these works take into account kinetic- and
181: mass-mixing \cite{FKN07, FLN07, KN05, KW06, KN04, CY07, PRV08a}, but it is
182: typically assumed that the gauge symmetry of the additional $U(1)$ is broken by
183: a Higgs or Stueckelberg mechanism.\footnote{Exceptions are \fex
184: Ref.~\cite{Dobrescu05, BBS08}, where constraints on, and consequences of,
185: higher dimensional operators that couple hidden and observable sector are
186: studied. In Ref.~\cite{FTY08} the authors study BBN and CMB constraints on the
187: particle content of a completely decoupled hidden sector which may contain
188: unbroken $U(1)$s. Furthermore, see Ref.~\cite{Redondo08a} for a short
189: discussion about gauge coupling unification in the presence of kinetic mixing.}
190: In this case, it is possible to derive bounds from high precision LEPI data if
191: the hidden $U(1)$ gauge boson mass is large (namely, $\chi\lesssim 0.05$ for
192: masses around $200\GeV$ \cite{FLN07}), and from different astronomical
193: observations and laboratory experiments if the masses are small (namely, for
194: masses around $100\eV$, the observed lifetime of the sun translates into a
195: bound $\chi\lesssim 10^{-13}$ \cite{Redondo08}, see Ref.~\cite{JRR08} for
196: bounds from considerations of the hidden CMB).
197:
198: The purpose of this paper is to consider in some detail cosmological
199: constraints and phenomenological properties of a kinetically mixed hidden
200: $U(1)$ extension of the MSSM with unbroken gauge symmetry. We will assume that
201: the hidden $U(1)$ gaugino is the LSP in most of the paper, and concentrate on
202: its prospect of being dark matter. In section 2, we will describe the model in
203: the component formalism and discuss typical values for mass- and
204: mixing-parameters in scenarios with gauge and gravity mediation of
205: supersymmetry breaking. In section 3, we derive a bound on the mixing
206: parameter from thermal overproduction of the hidden $U(1)$ gaugino. Section 4
207: considers bounds from primordial nucleosynthesis for the cases of stau and
208: neutralino NLSPs. There, we also analyze the effects of a possibly light
209: gravitino and show how bounds on the reheating temperature can be relaxed. In
210: section 5 we briefly discuss the collider phenomenology of this scenario. In
211: section 6 we study the case with an anomalously small kinetic mixing, where the
212: neutralino NLSP could become a viable, though unstable, dark matter candidate.
213: Finally, we present our conclusions and an outlook in section 7.
214:
215: \section{Model}
216: \label{sec:Model}
217: In this section, we will briefly describe the model under consideration in the
218: more familiar component formalism. In the Wess-Zumino gauge, a vector
219: superfield can be expanded in component fields as $V=-\theta \sigma^\mu {\bar
220: \theta} V_\mu +i \theta\theta {\bar \theta} {\bar \lambda}-i {\bar \theta}
221: {\bar \theta}\theta \lambda+ \frac{1}{2} \theta\theta {\bar \theta}{\bar
222: \theta} D$. Then, the pure gauge part of Eq.~\eqref{eqn:Lagrangian1}, including
223: supersymmetry breaking effects, reads:
224: \begin{eqnarray}
225: \label{eqn:Lagrangian}
226: \mathcal{L}_{gauge}&
227: =&-\frac{1}{4} \ppp{\hat{X}_{\mu\nu}&\hat{B}_{\mu\nu}} \mathcal{K}
228: \ppp{\hat{X}^{\mu\nu}\\ \hat{B}^{\mu\nu}}
229: -i\ppp{ \hat{\lambda}_X^\dagger & \hat{\lambda}_B^\dagger}\mathcal{K}
230: \bar{\sigma}^\mu\partial_\mu\ppp{\hat{\lambda}_{X}\\\hat{\lambda}_{B}}
231: \\\nonumber&&
232: +\frac{1}{2}\ppp{\hat{D}^*_X &
233: \hat{D}^*_B}\mathcal{K}\ppp{\hat{D}_{X}\\\hat{D}_{B}}
234: -\frac{1}{2}\ppp{\hat{\lambda}_X &
235: \hat{\lambda}_B}\hat{\mathcal{M}}
236: \ppp{\hat{\lambda}_{X}\\\hat{\lambda}_{B}} + \text{h.c.}\;,
237: \end{eqnarray}
238: where $\mathcal{K}$ and $\hat{\mathcal{M}}$ denote, respectively, the kinetic
239: and mass mixing matrices:
240: \begin{eqnarray}
241: \mathcal{K}=\ppp{1&\chi\\\chi&1}\hspace{0.5cm}\text{and}\hspace{0.5cm}
242: \hat{\mathcal{M}}=\ppp{\hat{M}_X&\delta\hat{M}\\\delta\hat{M}&\hat{M}_B}\;.
243: \label{eqn:MixingMatrix}
244: \end{eqnarray}
245:
246: It is convenient to work in the basis where the kinetic terms are canonical.
247: To this end, we will redefine the vector superfields according to
248: \begin{eqnarray}
249: \ppp{\hat{X}\\\hat{B}}=
250: \ppp{1&-\frac{\chi}{\sqrt{1-\chi^2}}\\0&\frac{1}{\sqrt{1-\chi^2}}}\ppp{X\\B}\;.
251: \label{eqn:GL2Transformation}
252: \end{eqnarray}
253: As explained in the introduction, the canonically normalized fields
254: $X_{\mu\nu}$ and $D_X$ completely decouple from the observable sector. On the
255: other hand, a mixing between the hidden gaugino and the bino remains, unless
256: $\delta\hat{M} \simeq \chi \hat{M}_X$. More precisely, in the basis where the
257: kinetic terms are canonical, the extended neutralino $(5\times5)$ mass matrix
258: reads, to lowest order in $\chi$,
259: \begin{eqnarray}
260: \mathcal{M}_\text{N}=\ppp{ M_X & \delta M & 0 & 0 & 0 \\
261: \delta M & M_B & 0 & -M_Z c_\beta s_W & M_Z s_\beta s_W \\
262: 0 & 0 & M_W & M_Z c_\beta c_W & -M_Z s_\beta c_W \\
263: 0 & -M_Z c_\beta s_W & M_Z c_\beta c_W & 0 & -\mu \\
264: 0 & M_Z s_\beta s_W & -M_Z s_\beta c_W & -\mu & 0 }\;,
265: \label{eqn:NeutralinoMassMatrix}
266: \end{eqnarray}
267: where, $\delta M \simeq \delta\hat{M} -\chi \hat{M}_X$, $M_X\simeq\hat{M}_X$
268: and $M_B\simeq\hat{M}_B$. Here, $\mu$ denotes the MSSM $\mu$-term, $M_Z$ the
269: mass of the $Z^0$ gauge boson, $s_W$ the sine of the Weinberg angle and
270: $s_\beta$ is related to the ratio of the two Higgs VEVs.
271:
272: Lastly, as was mentioned in the introduction, hidden sector matter particles
273: charged under the hidden $U(1)$ will acquire a minihypercharge, as a
274: consequence of the canonical normalization of the kinetic terms. For simplicity
275: we will assume throughout the paper that hidden sector matter particles are too
276: heavy to be produced during reheating, and thus do not influence the evolution
277: of the Universe.
278:
279: The size of the gaugino mass terms $\hat {\cal M}$ in the Lagrangian,
280: Eq.(\ref{eqn:Lagrangian}), is very model dependent. For example, in models
281: with gravity mediation, gaugino masses would arise from the term
282: \begin{eqnarray}
283: \mathcal{L}\supset\int d\theta^2 d{\bar \theta}^2 \frac{1}{M_P}
284: (\hat{Z}^\dagger_1 \hat{W}_B \hat{W}_B +
285: 2 \chi \hat{Z}^\dagger_2 \hat{W}_B \hat{W}_X + \hat{Z}^\dagger_3
286: \hat{W}_X \hat{W}_X). + {\rm h.c.}\;,
287: \label{eqn:LagrangianSuperFields}
288: \end{eqnarray}
289: where coefficients of order one have been dropped. In this equation $\hat Z_i$
290: are three spurion superfields, which will eventually acquire a vacuum
291: expectation value, $\langle \hat{Z}_i \rangle= M_P+\theta^2 F_i$, by some
292: unspecified mechanism\footnote{We assume that breaking of supersymmetry takes
293: place in a sector that is distinct from the hidden $U(1)$ under consideration,
294: hence the $D$-terms are zero and kinetic mixing has no effect on the MSSM mass
295: spectrum (like \fex in Ref.~\cite{DKM97}).}, yielding $\hat M_B=F_1/M_P$,
296: $\delta \hat M =\chi F_2/M_P$, $\hat M_X=F_3/M_P$, where $M_P\simeq
297: 2.4\times10^{18}\GeV$ denotes the reduced Planck mass.
298:
299: The simplest case consists on assuming just one spurion superfield which
300: couples universally to all vector superfields, namely $Z_i\equiv Z$ for all
301: $i$. If this is the case, at the scale $M_P$, the different soft terms satisfy
302: the relations $\hat{M}_B=\hat{M}_X=\sqrt{3} m_{3/2}$ and
303: $\delta\hat{M}=\sqrt{3} \chi m_{3/2}$, $m_{3/2}$ being the gravitino mass.
304: Therefore, after the canonical normalization of the gaugino kinetic terms, the
305: mass mixing term $\delta M $ in Eq.~\eqref{eqn:NeutralinoMassMatrix} vanishes.
306: However, the universality of the gaugino mass terms is broken by quantum
307: effects, thus inducing a non-vanishing $\delta M $ at low energies. Using the
308: renormalization group equations for the gaugino mass matrix in
309: appendix~\ref{app:RG}, we find that the mass mixing at the electroweak scale
310: reads
311: \begin{eqnarray}
312: \delta M_\text{EW}\simeq\frac{1}{8\pi^2}g_X^2 B_{XX} \ln
313: \left( \frac{M_P}{M_\text{hid}} \right) \chi_\text{EW} M_X\;,
314: \label{eqn:MassMixing}
315: \end{eqnarray}
316: where $g_X$ is the gauge coupling of the hidden $U(1)$ gauge group,
317: $M_\text{hid}$ is the mass scale of hidden sector matter, and $B_{XX}=\sum_i
318: Q^2_i$ is the sum, for all the particles in the theory, of all the hidden
319: $U(1)$ charges squared. On the other hand, the renormalization group evolution
320: will also make the hidden gaugino mass smaller at low energies. If there is an
321: enough number of (heavy) matter states in the hidden sector, the hidden gaugino
322: will become the lightest supersymmetric particle.
323:
324: In more general scenarios with gravity mediation, where several spurion fields
325: contribute to the breaking of supersymmetry, a tree level mixing between the
326: bino and the hidden gaugino will usually remain after canonical normalization
327: of the kinetic terms. On the other hand, the nature of the lightest
328: supersymmetric particle, whether it is the gravitino, the bino or the hidden
329: gaugino, depends on the details of the model.
330:
331: The most natural scenarios with a light hidden $U(1)$ gaugino are probably
332: those with gauge mediated supersymmetry breaking. Assuming that the messenger
333: fields are not charged under the hidden $U(1)$ gauge group, the soft mass of
334: the hidden gaugino will entirely come from gravity effects. If this is the
335: case, the gravitino and the hidden $U(1)$ gaugino will acquire similar masses,
336: which are naturally much smaller than the masses of the supersymmetric
337: particles of the MSSM. Again, whether the LSP is the gravitino or the hidden
338: gaugino depends on the particle content of the hidden sector, which will drive
339: the hidden gaugino mass to smaller values at low energies through radiative
340: effects.
341:
342: An important quantity in our analysis will be the mixing angle $\Theta$ between
343: the bino and the hidden $U(1)$ gaugino mass and interaction
344: eigenstates\footnote{Note that we always assume the mass hierarchy $M_B < M_W
345: <\mu$ in this paper.}. The approximate diagonalization of
346: Eq.~\eqref{eqn:NeutralinoMassMatrix} yields
347: \begin{eqnarray}
348: \Theta \simeq \frac{\delta M^\text{EW}}{M_B^\text{EW}-M_X^\text{EW}}\;,
349: \label{eqn:ChiEffective2}
350: \end{eqnarray}
351: where we have emphasized that all the quantities should be evaluated at the
352: electroweak scale. In the scenarios of supersymmetry breaking proposed in this
353: section with hidden $U(1)$ gaugino as LSP, the mixing angle $\Theta$ can be
354: written as
355: \begin{eqnarray}
356: \Theta \simeq C \cdot \chi_\text{EW}
357: \frac{M_X^\text{EW}}{M_B^\text{EW}-M_X^\text{EW}}\;,
358: \label{eqn:ChiEffective}
359: \end{eqnarray}
360: where $C$ is a constant that can roughly vary between $10^{-2} \lesssim C
361: \lesssim 1$, depending on whether $\delta M^\text{EW}$ is already present at
362: tree level or is generated radiatively, and on the particle content of the
363: hidden sector. For definiteness we will choose $C=1$ in the subsequent
364: analysis.
365:
366: Following the above discussion, we will first study the phenomenology of a
367: hidden $U(1)$ gaugino LSP, and second the case where both, the hidden $U(1)$
368: gaugino and the gravitino, constitute the lightest supersymmetric particles.
369:
370: \section{Overproduction constraints}
371: \label{sec:OP}
372: The least model dependent bound on the hidden $U(1)$ gaugino scenario comes
373: from considerations of the thermal history of the universe. After reheating of
374: the visible sector, hidden $U(1)$ gauginos can be produced in the hot primeval
375: plasma through the mixing with the MSSM neutralinos.
376:
377: This mechanism is similar to the thermal production of gravitinos or axions
378: (see \fex Ref.~\cite{BBB01,PS07,RS07}). However, in contrast to these cases,
379: the final abundance of the hidden $U(1)$ gaugino does not depend on the
380: reheating temperature, $T_R$. This follows from the fact that mass- and
381: kinetic-mixings appear at the renormalizable level of the Lagrangian, whereas
382: gravitino and axion couplings are mass-suppressed. Actually, the production is
383: most efficient at low temperatures, and this makes an exact calculation of the
384: hidden $U(1)$ gaugino abundance extremely complicated, since at temperatures
385: around $T\sim100\GeV$ the QCD coupling constant $g_s$ is of the order of one,
386: which precludes a sharp separation between hard, $T$, and soft, $ g_s T$,
387: momenta as required for the proper treatment of hard thermal loops and
388: screening effects~\cite{BY91}. How to deal with this drawback is an open
389: problem even in the milder case of gravitino production (see, however,
390: Ref.~\cite{RS07}).
391:
392: For the purposes of this paper it is enough to derive an order of magnitude
393: estimate on the hidden $U(1)$ gaugino abundance. To this end, we have
394: calculated the collision integral in a relativistic QCD plasma, taking into
395: account $2\rightarrow2$ scattering in the QCD and hypercharge sector with a
396: bino in the final state. Since binos mix with hidden $U(1)$ gauginos with the
397: mixing angle $\Theta$, the production rate of hidden $U(1)$ gauginos can be
398: straightforwardly calculated by multiplying the result for the binos by
399: $\Theta^2$. The relevant scattering processes and the corresponding squared
400: matrix elements are listed in Tab.~\ref{tab:MatrixElements}.
401: \begin{table}
402: \centering
403: \begin{tabular}{|l|c|}
404: \hline
405: Process & $\overline{|\mathcal{M}|^2} / \Theta^2 (g' g_s)^2 |T_{ba}^j|^2 Y_{L,R}^2$ \\\hline\hline
406: 1) $qg\rightarrow \tilde{q}\lambda_X$ & $-2\frac{u}{s}$ \\\hline
407: 2) $\tilde{q}\tilde{g}\rightarrow \tilde{q}\lambda_X$ & $-2\left( \frac{t}{s}+\frac{s}{t} \right)$ \\\hline
408: 3) $\tilde{q}g\rightarrow q\lambda_X$ & $2\frac{u}{t}$ \\\hline
409: 4) $q\tilde{g}\rightarrow q\lambda_X$ & $4$ \\\hline
410: 5) $q\bar{\tilde{q}}\rightarrow g\lambda_X$ & $-4\frac{s}{u}$ \\\hline
411: 6) $q\bar{q}\rightarrow \tilde{g}\lambda_X$ & $8$ \\\hline
412: 7) $\tilde{q}\bar{\tilde{q}}\rightarrow \tilde{g}\lambda_X$ & $4\left( \frac{t}{u}+\frac{u}{t} \right)$ \\\hline
413: \end{tabular}
414: \caption{Squared matrix elements for inelastic two-body scatterings with a
415: hidden $U(1)$ gaugino in the final state. Here, $T_{ba}^i$ denotes the
416: generators of the fundamental representation of the $SU(3)$ gauge group,
417: $Y_{L,R}$ the hypercharges of the (s)quarks, and $g'$ and $g_s$ the gauge
418: couplings of the electroweak and strong interactions, respectively. We
419: average over initial and sum over final spins.}
420: \label{tab:MatrixElements}
421: \end{table}
422:
423: Summing over all spins and particles, the collision term of hidden $U(1)$
424: gauginos in the relativistic MSSM plasma reads\footnote{We use Boltzmann
425: distribution functions for simplicity.}
426: \begin{eqnarray}
427: \gamma_\text{QCD}\equiv\frac{d^4n_X}{dVdt}&=&
428: \frac{308}{3\pi^3}\alpha' \alpha_s \Theta(T)^2
429: \left( 1-\frac{4}{7}\gamma_E-\frac{4}{7}\ln\frac{k^\ast}{T} \right)T^4
430: \nonumber\\
431: &\simeq& 3\times 10^{-3} \Theta(T)^2 T^4 \;,
432: \label{eqn:PrductionRate}
433: \end{eqnarray}
434: where one has in general to take into account the temperature dependence of the
435: mixing angle $\Theta$, which stems from the thermal mass of the bino (see
436: Eq.~\eqref{eqn:ChiEffective}). Here, we simply take $M_B(T)\simeq M_B$, since
437: most of the production is expected to happen when the particles become
438: non-relativistic. In this equation, $\alpha'$ and $\alpha_s$ denote the
439: hypercharge and QCD couplings, respectively, and $\gamma_E\simeq0.577$.
440: Furthermore, $k^\ast$ denotes the cut-off of intermediate three-momenta in
441: $t$-channel processes. Only processes with quarks in the intermediate state
442: contribute to these divergences, and we identify $k^\ast$ with their thermal
443: mass (at $T\sim 100\GeV$ the thermal quark masses lie between $0.63T$ and
444: $0.84T$ \cite{RS07}). Finally, solving the Boltzmann equation yields the
445: thermal abundance of hidden $U(1)$ gauginos,
446: \begin{eqnarray}
447: \Omega_X h^2 \approx 5.5\times 10^7\left( \frac{M_X}{100\GeV} \right)
448: \int_{T_0}^{T_R} dT
449: \frac{M_P}{T^2}\frac{\gamma_\text{QCD}(T)\Theta^2(T)}{T^4}\;.
450: \label{eqn:ThermalAbundance}
451: \end{eqnarray}
452: Taking into account thermal masses could actually lead to a further order-one
453: enhancement due to particle decay in the QCD plasma \cite{RS07}. Contributions
454: from electroweak interactions and Yukawa couplings are also expected to give
455: sizable corrections, again of order one.
456:
457: Requiring that $\Omega_X h^2 \lesssim 0.1$ gives the overproduction bound
458: \begin{eqnarray}
459: \Theta\lesssim 5\times10^{-12}\left( \frac{M_X}{M_B} \right)^{-1/2} \;,
460: \label{eqn:OverProductionBound}
461: \end{eqnarray}
462: where we have used $T_0\simeq M_B$ as a cutoff in
463: Eq.~\eqref{eqn:ThermalAbundance}, which corresponds to squark- and gluino
464: masses around $3M_B$. For other values the bound scales roughly like
465: $\sqrt{T_0}$.
466:
467: In Fig.~\ref{fig:boundsSlepton} (for $M_B=180\GeV$) and
468: Fig.~\ref{fig:boundsBino} (for $M_B=150\GeV$) we show the region in the
469: $(\chi,M_X)$ parameter space where the hidden $U(1)$ gaugino is overproduced.
470: Remarkably, a large part of the parameter space suggested by string theory is
471: excluded by the constraint Eq.~\eqref{eqn:OverProductionBound}.\footnote{A
472: sizable annihilation rate between hidden $U(1)$ gauginos would allow to
473: circumvent this bound. If the mixing angle $\Theta$ is small, this would
474: require additional light matter states in the hidden sector, which do not exist
475: in our scenario.}
476:
477: \begin{figure}[h]
478: \begin{center}
479: \includegraphics[width=0.9\linewidth]{pics/Slepton.eps}
480: \vspace{-0.5cm}
481: \end{center}
482: \caption{Summary of bounds on the hidden $U(1)$ gaugino parameter space for
483: the case of a slepton NLSP. We use $M_{\tilde{l}}=150\GeV$ and $M_B=180\GeV$.
484: The upper dark blue region is excluded by thermal overproduction. Below this
485: region, the hidden $U(1)$ gaugino is dominantly produced via late decaying
486: sleptons. The light blue region is excluded by energy injection during BBN
487: \cite{KKM05}, whereas the light green region is excluded by catalysis of
488: $^6\text{Li}$ production \cite{Pospelov07}. We also show the region which
489: would be excluded solely by free streaming arguments. The dotted lines show
490: the slepton lifetime. In the presence of a gravitino with $M_{\tilde{G}} =
491: 100\GeV$ the slepton would dominantly decay into the hidden $U(1)$ gaugino,
492: except in the red lower region. The dashed lines show the region that is
493: potentially excluded by bounds on $\Lambda$CWDM models in scenarios where the
494: thermally produced hidden $U(1)$ gaugino decays into a gravitino LSP with
495: large free streaming length (see Fig.~\ref{fig:FSXandGravitino}). The lines
496: correspond to a dark matter fraction $f=0.02$ and $f=0.2$ which is allowed to
497: be warm.}
498: \label{fig:boundsSlepton}
499: \end{figure}
500:
501: \begin{figure}[h]
502: \begin{center}
503: \includegraphics[width=0.9\linewidth]{pics/Bino.eps}
504: \vspace{-0.5cm}
505: \end{center}
506: \caption{Summary of bounds on the hidden $U(1)$ gaugino parameter space for
507: the case of a bino-like neutralino NLSP. We take $M_B=150\GeV$. The dark
508: blue region is excluded by thermal overproduction. Below this region, the
509: hidden $U(1)$ gaugino would be dominantly produced via late decaying
510: neutralinos. This scenario is totally excluded by BBN \cite{KKM05} (light
511: blue region). The bound actually strongly overlaps with the overproduction
512: region (dashed line). We also show the region (in yellow) that would be
513: solely excluded by free streaming arguments. The dotted lines show the
514: lifetime of the neutralino. (We used $M_\text{sf}=400\GeV$ and $\mu=300\GeV$
515: for the branching ratios).}
516: \label{fig:boundsBino}
517: \end{figure}
518:
519: \section{Bounds from primordial nucleosynthesis and \\ structure formation}
520: \label{sec:BBN}
521: In general, dark matter can be produced thermally and/or in the late decay of
522: some other relic particle $\xi$. The abundance of dark matter today is then
523: given by $\Omega_{\text{DM}}=\Omega_{\text{DM}}^{\text{th.}}+ (M_{\text{DM}} /
524: M_\xi) \Omega_\xi^{\text{th.}}$, where $\Omega_\xi^{\text{th.}}$ refers to the
525: thermal abundance of the particle $\xi$. Typical candidates for $\xi$, like
526: the stau or a neutralino, naturally freeze out with an abundance of the right
527: order of magnitude to account for the observed dark matter abundance, and a
528: given dark matter candidate can inherit this property as long as its direct
529: thermal production is negligible and provided the masses of the two particles
530: are comparable.
531:
532: These so-called superWIMP scenarios (super-weakly interacting massive
533: particles, see Ref.~\cite{FRT03}), where the dark matter superWIMPs are
534: produced in the late decay of other particles, are subject to strong
535: constraints due to their potential impact on primordial nucleosynthesis and
536: structure formation. If the production takes place during or after Big Bang
537: Nucleosynthesis (BBN), the hadronic and electromagnetic energy released in the
538: decay can destroy the successful predictions of the standard scenario (see \fex
539: Ref.~\cite{KKM05, HHK+07}). In cases where the decaying particle is
540: electrically charged, it could additionally form bound states with $^4$He,
541: triggering the catalytic production of $^6$Li and $^9$Be \cite{Pospelov07,
542: PPS08}. Furthermore, the superWIMP itself is potentially produced with a
543: momentum that is large enough to wash out density fluctuations on small scales
544: with impact on structure formation, similar to scenarios with warm dark matter
545: (for a discussion in the context of superWIMPs see Ref.~\cite{CFR+05}).
546:
547: Below, we will apply these bounds to the case of a hidden $U(1)$ gaugino
548: superWIMP, concentrating on the four different scenarios shown in
549: Tab.~\ref{tab:Spectra}. Following the superWIMP paradigm, the relic abundance
550: of the neutralino or the slepton is fixed by the requirement that the LSP
551: constitutes all dark matter today.
552:
553: \begin{table}
554: \centering
555: \begin{tabular}{|c|c|c|c|c|}
556: \hline
557: & $i)$ & $ii)$ & $iii)$ & $iv)$ \\\hline\hline
558: NNLSP & $\cdot$ & $\cdot$ & slepton & slepton \\
559: NLSP & slepton & neutralino & gravitino & $\lambda_X$ \\
560: LSP & $\lambda_X$ & $\lambda_X$ & $\lambda_X$ & gravitino \\
561: \hline
562: \end{tabular}
563: \caption{The different scenarios that we will investigate in this work.}
564: \label{tab:Spectra}
565: \end{table}
566:
567: \paragraph{Scenario i)}
568: After freeze out, the slepton NLSP decays predominantly through the channel
569: $\tilde{l}\rightarrow\lambda_X l$ into a lepton and the hidden $U(1)$ gaugino
570: LSP. The corresponding decay width is given by
571: \begin{eqnarray}
572: \Gamma_{\tilde{l}\rightarrow \lambda_X l} \simeq \frac{g'^2}{8\pi}\Theta^2
573: Y_{\tilde{l}}^2 M_{\tilde{l}} \left( 1- \frac{M_X^2}{M_{\tilde{l}}^2} \right)^2\;,
574: \label{eqn:DecayWidthSleptonApprox}
575: \end{eqnarray}
576: where $Y_{\tilde{l}}$ and $M_{\tilde{l}}$ denote, respectively, the hypercharge
577: and the mass of the slepton NLSP.
578:
579: The decay induces electromagnetic cascades in the plasma, the branching
580: fraction into electromagnetic energy being close to one,
581: $B_\text{EM}^{\tilde{l}}\simeq1$. The corresponding energy release is given by
582: $\epsilon^{\tilde{l}}_{\text{EM}}\approx(M^2_{\tilde{l}}-M_X^2) /(2
583: M_{\tilde{l}})$ (see Ref.~\cite{FST04} for an exhaustive discussion).
584:
585: Contributions to the hadronic energy release stem, if kinematically allowed,
586: from three-body decays into $Z^0$, $W^{\pm}$ and Higgs bosons. The
587: corresponding branching ratios are all of the order of, or smaller than,
588: $\alpha'/(16\pi s_\text{w}^2 )\sim 10^{-2}$, whereas the hadronic branching
589: fractions of the subsequent decay processes are all of the order of one. If
590: these decay channels are kinematically closed, the dominant contribution to
591: hadronic energy release comes from the strongly suppressed four-body decay
592: $\tilde{l}\rightarrow\lambda_X lq\bar{q}$, with a branching ratio of the order
593: of $10^{-6}$. Hence, the overall hadronic branching fraction is
594: $B_\text{had}^{\tilde{l}}\sim 10^{-6}$ for large masses of the hidden $U(1)$
595: gaugino ($M_X\gtrsim M_{\tilde{l}}-M_Z$) and $B_\text{had}^{\tilde{l}}\sim
596: 10^{-2}$ for small masses ($M_X\lesssim M_{\tilde{l}}-M_Z$). Variations in the
597: corresponding energy release $\epsilon_\text{had}^{\tilde{l}}$ are subdominant
598: because $B_\text{had}^{\tilde{l}}$ varies by several orders of magnitude, and
599: we simply take $\epsilon_\text{had}^{\tilde{l}}=(1/3)(M_{\tilde{l}}-M_X)$
600: \cite{FST04}.
601:
602: Applying the BBN bounds from Ref.~\cite{KKM05} to this scenario, a considerable
603: part of the hidden $U(1)$ gaugino parameter space can be excluded (see
604: Fig.~\ref{fig:boundsSlepton}).\footnote{In Ref.~\cite{KKM05} the bounds are
605: shown for the cases $B_\text{had}=1$, $10^{-3}$ and zero. We just interpolate
606: between $B_\text{had}=1$ and $B_\text{had}=10^{-3}$ linearly in
607: $\log_{10}(B_\text{had})$ and use $B_\text{had}=0$ as a cutoff.} We also show
608: the region that is excluded by the catalytic enhancement of $^6\text{Li}$ and
609: $^9\text{Be}$ production. These effects essentially require that the lifetime
610: of the slepton is smaller than around $2\times10^{3}\s$ \cite{Pospelov07}.
611:
612: The impact of superWIMP scenarios on structure formation is encoded in the free
613: streaming length $\lambda_\text{FS}$, defined as the comoving distance traveled
614: by a particle since it was produced:
615: \begin{eqnarray}
616: \lambda_\text{FS}=\int_0^{z_\text{prod.}} dz \frac{v(z)}{H(z)}\;.
617: \label{eqn:DefFS}
618: \end{eqnarray}
619: On scales below the free streaming length of dark matter particles, density
620: fluctuations are washed out before structure formation begins (see \fex
621: Ref.~\cite{KT90}). The non-observation of this effect in the latest
622: Lyman-$\alpha$ Forest data from the Sloan Digital Sky Survey \cite{M06}
623: excludes a free streaming length larger than roughly $\lambda_\text{FS}\gtrsim
624: 0.5\Mpc$.\footnote{See Ref.~\cite{SKB07}, where the power spectrum of
625: non-thermally produced dark matter is confronted with lower bounds on the mass
626: of sterile neutrino dark matter \cite{SMM+06, VLH+06}.} The region in the
627: parameter space of the hidden $U(1)$ gaugino that is excluded by this
628: requirement is shown in Fig.~\ref{fig:boundsSlepton} and lies completely in the
629: region already ruled out by BBN.
630:
631: \paragraph{Scenario ii)}
632: Two-body decays of a neutralino NLSP lead to final states where the hidden
633: $U(1)$ gaugino LSP is accompanied either by a $Z^0$ or Higgs boson or a
634: photon. However, if the decay into $Z^0$ or Higgs bosons is kinematically
635: forbidden, neutralinos predominantly decay via $\lambda_B\rightarrow\lambda_X
636: f\bar{f}$, since the radiative two-body decay into photons is one-loop
637: suppressed and negligible in most cases. We have summarized all relevant decay
638: widths in appendix~\ref{apx:decayWidth}, particularized to the case of a
639: bino-like neutralino. The hadronic branching fraction for the decay of a
640: bino-like neutralino is of the order of one for hidden $U(1)$ gaugino masses
641: that allow the decay into $Z^0$ bosons, and can range between $10^{-2}$ and
642: order one otherwise (depending on the squark masses). For definiteness we choose
643: $B_\text{had}^{\lambda_B}\sim\mathcal{O}(1)$ for all masses $M_X$, since
644: smaller values at large $M_X$ will not change our conclusions. For the
645: corresponding energy release we simply take $\epsilon_\text{had} = 1/3
646: (M_B-M_X)$. We apply the BBN bounds of Ref.~\cite{KKM05} and show the
647: corresponding excluded region in Fig.~\ref{fig:boundsBino}. Together with the
648: overproduction bounds, this scenario is excluded in the whole $(\chi, M_X)$
649: parameter space suggested by string theory, $\chi\sim 10^{-2}-10^{-16}$.
650:
651: \begin{figure}[h]
652: \begin{center}
653: \includegraphics{pics/Spectra.eps}
654: \hspace{6cm}a\hspace{6cm}b\hspace{5cm}
655: \end{center}
656: \caption{Spectra for (a) gravitino NLSP and (b) gravitino LSP (\comp last two
657: cases in Tab.~\ref{tab:Spectra}). We show the widths for the different decay
658: processes for typical particle masses. The mixing parameter is assumed to
659: lie in the allowed region of Fig.~\ref{fig:boundsSlepton}.}
660: \label{fig:SpectrumWithGravitinos}
661: \end{figure}
662:
663: \begin{figure}[h]
664: \begin{center}
665: \includegraphics[width=0.7\linewidth]{pics/FreeStreaming.eps}
666: \end{center}
667: \caption{Contour plot of free streaming lengths in units of $\text{Mpc}$. The
668: upper left (lower right) corner shows the free streaming length of gravitinos
669: (hidden $U(1)$ gauginos) that stem from the late decay of hidden $U(1)$
670: gauginos (gravitinos).}
671: \label{fig:FSXandGravitino}
672: \end{figure}
673:
674: \begin{figure}[h]
675: \begin{center}
676: \includegraphics[width=0.7\linewidth]{pics/Reheating.eps}
677: \end{center}
678: \caption{Bounds on the reheating temperature as function of the gravitino
679: mass, using Eq.~\eqref{eqn:GravitinoProduction} with $m_{\tilde{g}}=800\GeV$.
680: The mass of the hidden $U(1)$ gaugino is fixed to $M_X=120\GeV$. If the
681: gravitino is the LSP (left part), the reheating temperature is only bounded
682: by overproduction arguments (dark red region), which are only slightly
683: strengthened when, say, 20\% of the gravitino abundance is due to non-thermal
684: production (light red region). A gravitino NLSP (right part) would late decay
685: into the hidden $U(1)$ gaugino, yielding a warm dark matter component. If
686: only a fraction of $20\%$ or $2\%$ of dark matter is allowed to be warm (with
687: free streaming lengths as shown in Fig.~\ref{fig:FSXandGravitino}), the
688: corresponding blue regions are excluded.}
689: \label{fig:FSGravitinoNLSP}
690: \end{figure}
691:
692: \paragraph{Scenario iii)}
693: We will now discuss the effects of an additional light gravitino, concentrating
694: on the spectra shown in Fig.~\ref{fig:SpectrumWithGravitinos}, where we also
695: show typical values for the relevant decay widths.
696:
697: If the hidden $U(1)$ gaugino is the LSP (see
698: Fig.~\ref{fig:SpectrumWithGravitinos}a), it can be produced in three different
699: ways: thermally, via slepton decay and via gravitino decay. For mixings and
700: masses that are allowed in Fig.~\ref{fig:boundsSlepton}, the decay of sleptons
701: into gravitinos is strongly suppressed, and the decay of gravitinos into hidden
702: $U(1)$ gauginos is only accompanied by hidden $U(1)$ gauge bosons and invisible
703: to MSSM particles. Hence, bounds from BBN and overproduction are essentially
704: the same than in case $i)$.
705:
706: However, this does not hold for the free streaming bounds, since, in contrast
707: to hidden $U(1)$ gauginos which are produced thermally or via slepton decay,
708: hidden $U(1)$ gauginos that stem from the decay of gravitinos can have a quite
709: large free streaming length, of the order of several $\text{Mpc}$, as shown in
710: Fig.~\ref{fig:FSXandGravitino}. This leads to dark matter with a cold and a
711: warm component, so-called $\Lambda$CWDM models.
712:
713: Constraints on $\Lambda$CWDM models where recently studied in
714: Refs.~\cite{VLH+05, PCS+07, IMM08}. A warm dark matter component would induce a
715: step in the power spectrum of density fluctuations, with a size which depends
716: on the fraction of the dark matter that is warm and a position which is roughly
717: given by the corresponding free streaming length as $k_\text{FS}\sim
718: 2\pi/\lambda_\text{FS}$. Although models with mixed cold/warm dark matter arise
719: quite naturally in many situations, a full general analysis with all the latest
720: data is still lacking.\footnote{In Ref.~\cite{IMM08} a consideration of the
721: uncertainty of the power spectrum from WMAP 1-yr data \cite{S03} bounds the
722: fraction $f$ of dark matter that is allowed to have $\lambda_\text{FS} \simeq
723: 6\Mpc$ to be $f\lesssim 0.2 - 0.4$. This seems to be consistent with a
724: hydrodynamical analysis in Ref.~\cite{VLH+05}, where the bound on the fraction
725: of dark matter that can be made out of thermal light gravitinos with a mass of
726: $ m_{\tilde{G}} \approx 16\eV$ is stated as $f\lesssim 0.12$, since the
727: corresponding free streaming length (using Eq.~\eqref{eqn:DefFS}) is around
728: $\lambda_\text{FS}\sim 40\Mpc$. However, both analyses do not take into
729: account the latest Lyman-$\alpha$-forest data \cite{M06}. An analysis in
730: Ref.~\cite{PCS+07} which incorporates these data, and where lower bounds on the
731: mass of sterile neutrino dark matter from Ref.~\cite{SMM+06} are rescaled for
732: the case of $\Lambda$CWDM models, suggests that even a component with a free
733: streaming lengths around $\lambda_\text{FS}\sim 4\Mpc$ is bounded strongly with
734: $f\lesssim 0.1$.} Bearing in mind this limitation, we will simply show the
735: corresponding bounds on the reheating temperature and mixing parameter for the
736: two reference values $f=0.2$ and $f=0.02$, where $f$ denotes the fraction of
737: dark matter that is warm, $\Omega_\text{DM}=f
738: \Omega_\text{WDM}+(1-f)\Omega_\text{CDM}$; a complete analysis of the mixed
739: cold/warm dark matter scenario is beyond the scope of this paper.
740:
741: As discussed above, the warm component of hidden $U(1)$ gaugino DM comes from
742: the late decay of gravitinos. The thermal abundance of gravitinos directly
743: depends on the reheating temperature $T_R$ like \cite{BBB01,PS07,RS07}
744: \begin{eqnarray}
745: \Omega_{\tilde{G}}h^2\simeq0.27 \left( \frac{T_R}{10^{10}\GeV} \right)
746: \left( \frac{100\GeV}{m_{\tilde{G}}} \right) \left(
747: \frac{m_{\tilde{g}}}{1\text{TeV}} \right)^2\;,
748: \label{eqn:GravitinoProduction}
749: \end{eqnarray}
750: where $m_{\tilde{g}}$ denotes the gluino mass. Hence, the bounds on
751: $\Lambda$CWDM models translate into a bound on the reheating temperature as
752: shown in the right part of Fig.~\ref{fig:FSGravitinoNLSP}.
753:
754: \paragraph{Scenario iv)}
755: Scenarios with a gravitino LSP and a slepton (usually a stau) NLSP are known to
756: be strongly constrained due to the catalytic production of $^6\text{Li}$ and
757: $^9\text{Be}$ in the presence of long-lived charged particles during BBN (see
758: above discussion). The resulting bound on the lifetime of staus translates into
759: an upper bound on the gravitino mass of around $m_{\tilde{G}}\lesssim 1\GeV$.
760: At the same time, requiring that gravitinos with $m_{\tilde{G}}\lesssim 1\GeV$
761: are not overproduced forbids reheating temperatures as high as
762: $T_R\sim10^9\GeV$ (see left part of Fig.~\ref{fig:FSGravitinoNLSP} and
763: Eq.~\eqref{eqn:GravitinoProduction}). However, these high reheating temperatures
764: are favored in scenarios with leptogenesis as the source of the baryon
765: asymmetry of the universe \cite{FY86, BDP05}.
766:
767: One way to relax the tension between leptogenesis and gravitino DM is to assume
768: that the production of entropy between stau freeze out and BBN dilutes the stau
769: abundance sufficiently to evade the bounds \cite{BHI+06, PS07a}. Another way is
770: to impose a small violation of $R$-parity \cite{BCH+07}, which allows the stau
771: to decay fast into standard model particles. In this case, the gravitino
772: becomes unstable and decays with a lifetime which is typically longer than the
773: age of the universe. The gravitino decay products could be detected as an
774: anomalous contribution to the cosmic ray fluxes \cite{IT08, IT08a, IMM08a},
775: opening the possibility to probe this scenario.
776:
777: Interestingly, a similar mechanism can work in the presence of a hidden $U(1)$
778: gaugino for spectra of the form (slepton, $\lambda_X$, gravitino), as depicted
779: in Fig.~\ref{fig:SpectrumWithGravitinos}b. For mixings and masses of the hidden
780: $U(1)$ gaugino that lie in the allowed region of Fig.~\ref{fig:boundsSlepton},
781: the lifetime of the stau is compatible with all BBN bounds. In this case, the
782: reheating temperature can be as large as $T_R\sim10^9\GeV$, as shown in
783: Fig.~\ref{fig:FSGravitinoNLSP}, since the gravitino mass and the stau lifetime
784: are decoupled.
785:
786: The warm component of gravitino dark matter, with a free streaming length of
787: the order of several $\text{Mpc}$ (see Fig.~\ref{fig:FSXandGravitino}), comes
788: now from the late decay of hidden $U(1)$ gauginos. Hence, bounds on
789: $\Lambda$CWDM models constrain the sum over the thermal and non-thermal
790: abundance of hidden $U(1)$ gauginos in this case. Firstly, this shrinks the
791: allowed region of the hidden $U(1)$ gaugino parameter space as shown in
792: Fig.~\ref{fig:boundsSlepton} (dashed lines). Secondly, in cases where the
793: thermal production of hidden $U(1)$ gauginos is negligible, the stau abundance
794: is bounded like $Y_\text{stau} \lesssim (9\times10^{-14} - 9\times10^{-13})
795: (M_{\tilde{G}}/100\GeV)^{-1}$ if we allow a fraction $f=0.02 - 0.2$ of dark
796: matter to be warm. Thus, the $^9$Be and $^6$Li bounds on the abundance of
797: long-lived staus, which are roughly given by $Y_\text{stau}\lesssim 10^{-15}$
798: and $Y_\text{stau}\lesssim 10^{-16}$, respectively (see Ref.~\cite{PPS08}), can
799: be weakened by two or three orders of magnitude.
800:
801: In the CMSSM with gravitino dark matter the typical relic abundance of stau
802: NLSPs ranges between $Y_\text{stau} \sim10^{-14}$ and $Y_\text{stau}
803: \sim10^{-13}$ (although considerably smaller values are possible due to
804: resonant effects or coannihilation \cite{PS08}, or \fex in models with
805: non-universal Higgs masses \cite{RSW08}). In these kind of models, the
806: phenomenologically allowed parameter space will be greatly enlarged by the
807: above mechanism. However, a detailed analysis of the influence of a weakly
808: interacting hidden sector on constraints on the reheating temperature would
809: require reliable bounds for general $\Lambda$CWDM models.
810:
811: \section{Collider phenomenology}
812: At the LHC, cascade decays of squarks and gluinos can produce of the order of
813: $10^6$ NLSPs per year if the sparticle masses are close to the current
814: experimental limits~\cite{BHS+97}, thus offering a unique opportunity to test
815: scenarios with a light hidden $U(1)$ gaugino. It was shown in the previous
816: sections that in this class of scenarios the neutralino as NLSP is excluded,
817: whereas the stau as NLSP is allowed only if it has a lifetime that ranges
818: between $\tau\sim 10^{-2}\s$ and $\tau\sim 10^{3}\s$. Therefore, the stau will
819: typically traverse the detector leaving a distinctive heavily ionizing charged
820: track~\cite{DT90}. Moreover, if the stau is slow enough, it might get trapped
821: in the detector, decaying at late times and thus allowing a measurement of the
822: lifetime and, correspondingly, of the mixing parameter $\chi$. The number of
823: staus trapped can be incremented by placing a 1-10 kton massive material around
824: the LHC detectors, which will allow to collect ${\cal O}(10^3-10^4)$
825: staus~\cite{HKN+04}. If the stopper material is simultaneously an active
826: real-time detector, the decay products and their angular distribution could be
827: studied in detail. This will allow to distinguish this scenario from the
828: scenario of gravitino dark matter, which yields very similar signatures at
829: colliders~\cite{BHR+04}.
830:
831: \section{The case with an anomalously small $\chi$}
832: We have shown in sections~\ref{sec:OP} and \ref{sec:BBN} that constraints from
833: overproduction, BBN and structure formation exclude the scenario with a
834: neutralino NLSP and hidden $U(1)$ gaugino for natural values of the mixing
835: parameter $\chi\sim 10^{-2}-10^{-16}$ (see Fig.~\ref{fig:boundsBino}).
836: However, these strong constraints could be evaded in more elaborated models
837: yielding a smaller $\chi$. Our interest in exploring the region with small
838: $\chi$ is that the neutralino lifetime could become larger than the age of the
839: Universe, thus constituting a viable dark matter candidate
840: itself.\footnote{Clearly, the scenario with charged slepton NLSP is excluded
841: for very small values of $\chi$, since it would yield an abundance of
842: anomalously heavy hydrogen in conflict with the experimental constraints.}
843: Nevertheless, in contrast to the standard neutralino dark matter scenario, in
844: this scenario the neutralino NLSP is unstable and decays with very long
845: lifetimes into the hidden gaugino and Standard Model particles.
846:
847: The neutralino dark matter particle (that for definiteness we assume to be a
848: pure bino) can decay into different channels:
849: \begin{eqnarray}
850: \lambda_B& \rightarrow& Z\; \lambda_X\;, \nonumber \\
851: \lambda_B& \rightarrow& \gamma\; \lambda_X\;, \nonumber \\
852: \lambda_B&\rightarrow& h\; \lambda_X\;, \nonumber \\
853: \lambda_B& \rightarrow& f \bar f\; \lambda_X\;,
854: \end{eqnarray}
855: with decay rates listed in appendix~\ref{app:DR}. The fragmentation of the
856: $Z^0$ and the Higgs boson produces a continuous spectrum of stable particles,
857: such as photons, positrons or antiprotons which contribute to the primary
858: fluxes of cosmic rays. The experimental constraints on these fluxes will then
859: translate into constraints on the parameters of the model.
860:
861: On the other hand, it has been pointed out recently that the decay of a dark
862: matter particle with a lifetime $\tau\sim 10^{26}$s into weak gauge bosons with
863: a momentum $\sim 50$ GeV could simultaneously explain the EGRET anomaly in the
864: extragalactic gamma-ray background and the HEAT excess in the positron
865: fraction \cite{IT08a, IMM08a}. Interestingly, the scenario we are
866: considering here, where the neutralino dark matter particle decays into a $Z^0$
867: gauge boson, constitutes a promising candidate of this kind. Moreover, in
868: contrast to other candidates for decaying dark matter that have been proposed,
869: such as the gravitino \cite{BCH+07} or a hidden $U(1)$ gauge boson
870: \cite{CTY08}, the unstable neutralino could be detected in direct dark matter
871: searches.
872:
873: Although the simplest models with hidden gaugino LSP and neutralino NLSP
874: predict a mixing parameter $\chi$ ranging between $10^{-2}$ and $10^{-16}$, and
875: thus a neutralino lifetime ranging between $\tau\sim10^{-14}\s$ and
876: $\tau\sim10^{14}\s$, it is not difficult to construct plausible models where
877: the neutralino lifetime can be orders of magnitude longer than the age of the
878: Universe. To this end, let us consider a model with two hidden $U(1)$ gauge
879: groups, $U(1)_{X}$ and $U(1)_{X^\prime}$, and vector superfields denoted by
880: $\hat X$ and $\hat X^\prime$, respectively. We will assume that the hidden
881: sector particles that generate the kinetic mixing are charged either under
882: $U(1)_Y$ and $U(1)_{X^\prime}$, or under $U(1)_X$ and $U(1)_{X^\prime}$, but
883: not under $U(1)_Y$ and $U(1)_{X}$ simultaneously. If this is the case, the
884: kinetic mixings $\hat{W}_B^\alpha\hat{W}_{X^\prime\,\alpha}$ and
885: $\hat{W}_{X}^\alpha\hat{W}_{X^\prime\alpha}$ will be generated, but not
886: $\hat{W}_B^\alpha\hat{W}_{X\,\alpha}$. Therefore, the SUSY invariant part of
887: the Lagrangian reads
888: \begin{eqnarray}
889: \label{eqn:twohiddenU1sSUSY}
890: {\cal L}_{\rm SUSY} &=&\int d^2\theta \, \left(\hat{W}_B^\alpha \hat{W}_{B\,\alpha}+
891: \hat{W}_{X}^\alpha \hat{W}_{X\,\alpha}+
892: \hat{W}_{X^\prime}^\alpha \hat{W}_{X^\prime\,\alpha}+\right. \nonumber \\
893: && \left.2\chi_1 \hat{W}_B^\alpha\hat{W}_{X^\prime\,\alpha}+
894: 2\chi_2 \hat{W}_{X}^\alpha\hat{W}_{X^\prime\alpha}\right)
895: +{\rm h.c.} \;.
896: \end{eqnarray}
897: In addition, there exists a soft SUSY breaking Lagrangian:
898: \begin{equation}
899: {\cal L}_{{\rm SUSY}\hspace{-0.75cm}\line(4,1){20}}=-\frac{1}{2}
900: \ppp{\hat{\lambda}_X &\hat{\lambda}_{X^\prime} & \hat{\lambda}_B}\hat{\mathcal{M}}
901: \ppp{\hat{\lambda}_{X}\\\hat{\lambda}_{X^\prime}\\\hat{\lambda}_{B}} +
902: \text{h.c.}\;,
903: \label{eq:twohiddenU1ssoft}
904: \end{equation}
905: where
906: \begin{equation}
907: \hat{\mathcal{M}}=\ppp{\hat{M}_X&\delta\hat{M_2}& 0\\
908: \delta\hat{M_2}&\hat{M}_{X^\prime}&\delta\hat{M_1}\\
909: 0&\delta\hat{M_1}&\hat{M_B}}\;.
910: \end{equation}
911: This structure for the soft mass matrix can be straightforwardly derived from
912: Eq.~\eqref{eqn:twohiddenU1sSUSY} introducing five arbitrary spurion superfields
913: as in Eq.~\eqref{eqn:LagrangianSuperFields}. Note that $\delta\hat{M_1}$ and
914: $\delta\hat{M_2}$ are suppressed by $\chi_1$ and $\chi_2$, respectively.
915:
916: In order to canonically normalize the vector superfields, we define
917: \begin{eqnarray}
918: B&=&\hat B \;, \nonumber \\
919: X&=&\hat X-\chi_1\, \chi_2 \hat B \;,\nonumber \\
920: X^\prime &=& \hat X^\prime +\chi_1 \hat B+\chi_2 \hat X \;,
921: \end{eqnarray}
922: where we have assumed $\chi_1,\chi_2\ll 1$. Then, the SUSY invariant part of
923: the Lagrangian reads:
924: \begin{eqnarray}
925: {\cal L}_{\rm SUSY}&=&\int d^2\theta \, \left(W_B^\alpha W_{B\,\alpha}+
926: W_{X}^\alpha W_{X\,\alpha}+
927: W_{X^\prime}^\alpha W_{X^\prime\,\alpha}\right)
928: +{\rm h.c.} \;,
929: \end{eqnarray}
930: while the SUSY breaking Lagrangian reads:
931: \begin{equation}
932: {\cal L}_{{\rm SUSY}\hspace{-0.75cm}\line(4,1){20}}=-\frac{1}{2}
933: \ppp{\lambda_X &\lambda_{X^\prime} & \lambda_B}\mathcal{M}
934: \ppp{\lambda_{X}\\ \lambda_{X^\prime}\\\lambda_{B}} + {\rm h.c.}\;,
935: \end{equation}
936: with
937: \begin{equation}
938: \mathcal{M}\simeq \ppp{ M_X&\delta M_2& \delta M_{12}\\
939: \delta M_2&M_{X^\prime}&\delta M_1\\
940: \delta M_{12}&\delta M_1&M_B}\;.
941: \end{equation}
942: In this equation, $M_X\simeq \hat M_X$, $M_{X^\prime}\simeq \hat M_{X^\prime}$
943: and $M_B\simeq \hat M_B$, while $\delta M_2\simeq \delta\hat{M_2}-\chi_2\,
944: M_{X^\prime}$ and $\delta M_1\simeq \delta\hat{M_1}-\chi_1\, M_{X}$. More
945: importantly, after the canonical normalization of the kinetic terms, a mixing
946: term between the bino and the hidden gaugino $\lambda_X$ has been generated,
947: $\delta M_{12}\simeq \chi_1\, \chi_2\,(M_X+M_{X^\prime}) - \chi_2\,
948: \delta\hat{M_1} - \chi_1\,\delta\hat{M_2}$, which is doubly suppressed by
949: $\chi_1$ and $\chi_2$.
950:
951: If $\lambda_X$ is the LSP and $\lambda_B$ is the NLSP, the decay rate of the
952: bino NLSP is very strongly suppressed, thus yielding very long lifetimes for
953: the neutralino dark matter. For instance, assuming that the kinetic mixing
954: between the observable sector and the hidden sector is $\chi_1\sim 10^{-16}$,
955: as suggested by string theory, a mixing between the two hidden sector $U(1)$s
956: of $\chi_2 \sim 10^{-7}$ would be necessary in order to render a neutralino
957: lifetime of the order of $10^{26}\s$.
958:
959: The signatures in cosmic ray experiments of the scenario of neutralino dark
960: matter which decays into hidden gauginos are fairly model dependent and will be
961: shown elsewhere \cite{prep}. In this paper we will just derive
962: a conservative bound on the mixing parameter $\chi$ from requiring a gamma ray
963: flux from neutralino decay in agreement with the EGRET observations. To this
964: end, we will just consider the contribution to the gamma ray flux from the two
965: body decay $\lambda_B\rightarrow \gamma\, \lambda_X$.
966:
967: The total gamma ray flux receives several contributions. First, we expect a
968: gamma ray flux from the center of the Galaxy produced by neutralino
969: annihilations. This contribution does not depend on whether the neutralino is
970: absolutely stable or on the contrary decays at late times. Hence, it does not
971: constrain our scenario and will not be further discussed. The interested reader
972: can find in the vast existing literature the prospects of detecting such a
973: signal from neutralino annihilation coming from the center of the Galaxy
974: (see \fex Ref.~\cite{BHS05}).
975:
976: In addition, we expect a diffuse flux produced in the decay of neutralinos at
977: cosmological distances and in the Milky Way halo. Following Ref.~\cite{BCH+07,
978: BBC+07}, we will average the halo gamma ray signal of the $\lambda_B\rightarrow
979: \gamma \lambda_X$ channel over the whole sky, excluding a band of $\pm
980: 10^\circ$ around the Galactic disk in order to compare our results with the
981: EGRET data~\cite{SMR04}. To be conservative, we will require that the
982: corresponding peak in the energy spectrum, after convolving with the energy
983: resolution of 15\% quoted by EGRET, remains below the 2$\sigma$ band of the
984: EGRET spectrum obtained by Sreekumar {\it et al.} in~\cite{S98},
985: \begin{equation}\label{sreekpowerlaw}
986: \frac{dJ_{\rm EGRET}}{dE} =
987: (7.32 \pm 0.34) \times 10^{-6}
988: ({\rm cm}^2~{\rm s}~{\rm str}\ {\rm GeV})^{-1}
989: \left(\frac{E}{0.451 \ {\rm GeV}}\right)^{-2.1\pm 0.03}\; .
990: \end{equation}
991: The result is shown in Fig.~\ref{fig:EGRETbounds}, where we used a reference
992: bino mass of $M_B=150\GeV$. For hidden $U(1)$ gaugino masses around
993: $M_X\sim75\GeV$, mixings down to $\chi\sim 10^{-21}$ can be excluded
994: from the EGRET measurements of the gamma ray flux.
995:
996: \begin{figure}[h]
997: \begin{center}
998: \includegraphics[width=0.9\linewidth]{pics/EGRET.eps}
999: \end{center}
1000: \caption{Region of the hidden $U(1)$ gaugino parameter space which is
1001: excluded by confronting the EGRET measurements \cite{S98} with the
1002: extragalactic gamma ray flux produced in the decay
1003: $\lambda_B\rightarrow\gamma\lambda_X$ (blue region).
1004: The bino mass is fixed to $M_B=150\GeV$. We also show
1005: for reference the isocurves for the neutralino life-time.}
1006: \label{fig:EGRETbounds}
1007: \end{figure}
1008:
1009: \section{Conclusions}
1010: An unbroken hidden $U(1)$ that interacts with the Standard Model only via
1011: kinetic mixing with hypercharge decouples completely from the visible world.
1012: However, in the supersymmetric version of this scenario this is no longer the
1013: case. We have shown that a mass mixing between the bino and the hidden $U(1)$
1014: gaugino is always generated via radiative effects, although this mixing can be
1015: generated already at tree level in some well motivated scenarios. Moreover,
1016: we have discussed different scenarios of supersymmetry breaking in which the
1017: hidden $U(1)$ gaugino is the lightest supersymmetric particle.
1018:
1019: We have mostly concentrated on this possibility and we have derived
1020: cosmological bounds on this scenario from precluding overproduction of hidden
1021: $U(1)$ gauginos and from the requirements of successful Big Bang
1022: nucleosynthesis and structure formation. The combination of these constraints
1023: excludes a neutralino NLSP except for extremely small mixings (see
1024: Fig.~\ref{fig:boundsBino}). On the other hand, when the NLSP is a stau, an
1025: allowed window for mixings around $\chi\sim 10^{-10} - 10^{-13}$ remains (see
1026: Fig.~\ref{fig:boundsSlepton}). In this window, the stau has a lifetime larger
1027: than $10^{-1}\s$ and thus might be detected at future colliders as a heavily
1028: ionizing charged track.
1029:
1030: The reheating temperature in scenarios with gravitino dark matter and stau
1031: NLSPs is known to be strongly constrained, and we have shown that these
1032: constraints relax considerably in the presence of a hidden $U(1)$ gaugino (see
1033: Fig.~\ref{fig:FSGravitinoNLSP}). This might be a rather general effect of very
1034: weakly interacting hidden sectors and deserves further attention.
1035:
1036: Finally, we have discussed the case of an anomalously small mixing parameter,
1037: $\chi\ll 10^{-16}$. For these small mixings, the neutralino NLSP can become
1038: long lived enough to constitute the dark matter of the Universe. We have
1039: constructed a simple model with two hidden $U(1)$s where a tiny mixing can be
1040: naturally obtained. Even though the neutralino is very long lived, it
1041: eventually decays into the hidden gaugino and standard model particles, which
1042: might be detected as an anomalous contribution to the cosmic ray fluxes. Using
1043: the EGRET measurement of the extragalactic gamma ray flux, we have derived a
1044: conservative bound on the mixing parameter $\chi\lesssim 10^{-20-21}$ (see
1045: Fig.~\ref{fig:EGRETbounds}).
1046:
1047: \section*{Acknowledgments}
1048: C.W. thanks Chlo\'e Papineau, James Wells, and especially Javier Redondo for
1049: valuable discussions.
1050:
1051: \begin{appendix}
1052: \section{Decay rates}
1053: \label{app:DR}
1054: For convenience, we summarize the neutralino, slepton and gravitino decay
1055: rates that we used in this paper.
1056:
1057: The decay widths for bino-like neutralinos that decay into hidden $U(1)$
1058: gauginos can be derived from the results in Ref.~\cite{HW89, BFM86}. They are
1059: \label{apx:decayWidth}
1060: \begin{eqnarray}
1061: \Gamma(\lambda_B\rightarrow Z^0 \lambda_X)&\simeq&
1062: \frac{1}{128\pi}g'^2 s_W^2 \Theta^2 M_B
1063: \lambda^{1/2}\left( 1, \frac{M_X^2}{M_B^2},\frac{M_Z^2}{M_B^2}\right)\times
1064: \nonumber\\&&
1065: \hspace{-3.0cm}\times
1066: \frac{M_Z^6}{\mu^4 M_B^2}
1067: \left( \frac{M_B^2}{M_Z^2} + \frac{M_X^2}{M_Z^2}-2
1068: +\left( \frac{M_B^2}{M_Z^2} - \frac{M_X^2}{M_Z^2}\right)^2 +
1069: 6 \frac{M_X}{M_Z}\frac{M_B}{M_Z}\right)\;,
1070: \label{eqn:GammaBinoToZX}\\
1071: %**********************************************************
1072: \Gamma(\lambda_B\rightarrow \gamma \lambda_X)&\simeq&
1073: \frac{1}{128\pi}\left( \frac{15}{32\pi^2} \right)^2 e^2 g'^4 \Theta^2
1074: M_B\times \nonumber\\
1075: &&\times
1076: \frac{M_B^4}{M_\text{sf}^4} \left( 1-\frac{M_X^2}{M_B^2} \right)^3 \left(
1077: 1-\frac{M_X}{M_B} \right)^2\;,
1078: \label{eqn:GammaBinoTogammaX}\\
1079: %**********************************************************
1080: \Gamma(\lambda_B\rightarrow h \lambda_X)&\simeq&
1081: \frac{1}{32\pi} g'^2 \Theta^2 M_B
1082: \frac{M_Z^2 s_W^2}{\mu^2}
1083: \lambda^{1/2}\left( 1, \frac{M_X^2}{M_B^2},\frac{M_h^2}{M_B^2}\right)\times
1084: \nonumber\\&&\times
1085: \left( 1+\frac{M_X^2}{M_B^2} - \frac{M_h^2}{M_B^2}
1086: +2\frac{M_X}{M_B} \right)\;,
1087: \label{eqn:GammaBinoToHiggsX}\\
1088: %**********************************************************
1089: \Gamma(\lambda_B\rightarrow f\bar{f} \lambda_X)&\simeq&
1090: %\frac{9}{2048\pi^3}
1091: 1.4\times10^{-4} g'^4 \Theta^2 M_B
1092: \frac{M_B^4}{M_\text{sf}^4}
1093: \left(1-\frac{M_X^2}{M_B^2}\right)^5\;,
1094: \label{eqn:GammaBinoToLeptons}
1095: \end{eqnarray}
1096: where we have used the function
1097: \begin{eqnarray}
1098: \lambda(a^2,b^2,c^2)=(a^2-(b+c)^2)(a^2-(b-c)^2)\;.
1099: \nonumber
1100: \end{eqnarray}
1101: The approximations that we use for the neutralino mixing angles become exact
1102: in the limit $M_X \ll M_B \ll M_W \ll \mu$. Note that the two-body decay
1103: into a hidden $U(1)$ gaugino and a photon, Eq.~\eqref{eqn:GammaBinoTogammaX},
1104: is one-loop suppressed, and we sum over all (s)leptons in the loop, assuming
1105: that they have a common mass $M_\text{sf}$. The final state in
1106: Eq.~\eqref{eqn:GammaBinoToLeptons} incorporates neutrinos and charged
1107: leptons, and we took into account sfermions and $Z^0$ gauge bosons in the
1108: intermediate state, where the later are subdominant.
1109:
1110: The relevant decay widths with gravitinos in the initial or final state can
1111: be found in Ref.~\cite{FST04,KMY06} and are
1112: \begin{eqnarray}
1113: \Gamma(\lambda_X\rightarrow \tilde{G}X) &=& \frac{1}{48\pi M_\text{pl}^2}
1114: \frac{M_X^5}{M_{\tilde{G}}^2}
1115: \left( 1-\frac{M_{\tilde{G}}^2}{M_X^2} \right)^3
1116: \left( 1+3\frac{M_{\tilde{G}}^2}{M_X^2} \right)\;, \\
1117: %***
1118: \Gamma(\tilde{G}\rightarrow \lambda_X X) &=& \frac{1}{32\pi M_\text{pl}^2}
1119: M_{\tilde{G}}^3
1120: \left( 1-\frac{M_X^2}{M_{\tilde{G}}^2} \right)^3
1121: \left( 1+\frac{1}{3}\frac{M_X^2}{M_{\tilde{G}}^2} \right)\;,\\
1122: %***
1123: \Gamma(\tilde{l}\rightarrow \tilde{G}l) &=& \frac{1}{48\pi M_\text{pl}^2}
1124: \frac{M_{\tilde{l}}^5}{M_{\tilde{G}}^2}
1125: \left( 1-\frac{M_{\tilde{G}}^2}{M_{\tilde{l}}^2} \right)^4\;.
1126: \end{eqnarray}
1127:
1128: \section{Renormalization group equations}
1129: \label{app:RG}
1130: The RGEs for multiple $U(1)$ models with kinetic mixing where first published
1131: in Ref.~\cite{ACQ88}. One-loop RGEs for gaugino masses and their mixings can
1132: be found in Ref.~\cite{Suematsu99}.\footnote{Note that our result for the
1133: RGEs of the gaugino mass matrix in Eq.~\eqref{eqn:mRGplain} differs from the
1134: one given in Ref.~\cite{Suematsu99} by an additional symmetrization.}
1135:
1136: The RGEs acquire their simplest form in the basis where gauge bosons and
1137: gauginos have a canonical kinetic term. The matrix of the gauge couplings
1138: $\bar{g}_{ij}$ is defined according to the term $\mathcal{L}\supset
1139: \bar{g}_{ij}j^{\mu}_i A_\mu^j$ in the Lagrangian, where the indices $i=1,2$
1140: and $j=1,2$ run over the charged currents of the two sectors and over the
1141: gauge boson states, respectively. In this representation, the one-loop RGEs
1142: for the coupling constants are
1143: \begin{eqnarray}
1144: \frac{d}{dt}\bar{g}=\frac{1}{16\pi^2}\bar{g}\bar{g}^T B \bar{g}\;,
1145: \label{eqn:gRGplain}
1146: \end{eqnarray}
1147: where we have used the charge matrix $B_{ij}=tr(Q_i Q_j)$, and $t=\ln(Q/Q_0)$
1148: with $Q$ as RG scale. The trace in $B_{ij}$ runs over all chiral
1149: supermultiplets.
1150:
1151: The RGEs for the gaugino mass matrix $\bar{\mathcal{M}}$ take a similar form
1152: \begin{eqnarray}
1153: \frac{d}{dt}\bar{\mathcal{M}}=\frac{1}{16\pi^2}
1154: \left( \bar{\mathcal{M}}\bar{g}^T B \bar{g}+\bar{g}^T B \bar{g}
1155: \bar{\mathcal{M}} \right)\;.
1156: \label{eqn:mRGplain}
1157: \end{eqnarray}
1158:
1159: Note that only three of the four entries of $\bar{g}$ are physical because
1160: the basis of the gauge bosons is only fixed up to a rotation. It is
1161: convenient to state the RGEs for the case where the non-diagonal elements in
1162: the couplings $g_{ij}$ are chosen to vanish, and where their third free
1163: component is absorbed in the kinetic mixing term like in
1164: Eq.~\eqref{eqn:Lagrangian}. We obtain
1165: \begin{eqnarray}
1166: \frac{d}{dt}g_{X,B}=\frac{1}{16\pi^2}g_{X,B}^3 B_{XX,BB}\;,
1167: \label{eqn:gRGint}
1168: \end{eqnarray}
1169: \begin{eqnarray}
1170: \frac{d}{dt}\chi=-\frac{1}{8\pi^2}g_X g_B B_{XB}
1171: +\frac{1}{16\pi^2}\chi\left( g_X^2 B_{XX} + g_B^2 B_{BB} \right)
1172: +\mathcal{O}(\chi^2)\;,
1173: \label{eqn:chiRGint}
1174: \end{eqnarray}
1175: \begin{eqnarray}
1176: \frac{d}{dt}\hat{M}_{X,B}=\frac{1}{8\pi^2}g_{X,B}^2 B_{XX,BB} \hat{M}_{X,B}\;,
1177: \label{eqn:mRGint}
1178: \end{eqnarray}
1179: \begin{eqnarray}
1180: \frac{d}{dt}\delta\hat{M}= \frac{1}{16\pi^2}\left( g_X^2 B_{XX} + g_B^2 B_{BB}
1181: \right)\delta\hat{M}+\mathcal{O}(\chi^2)\;.
1182: \label{eqn:mmixRGint}
1183: \end{eqnarray}
1184: \end{appendix}
1185: \frenchspacing
1186: \bibliographystyle{h-physrev3}
1187: \bibliography{hp.bib}
1188: \end{document}
1189: