CERN–PH–TH/2007088 DCPT/07/52, IPPP/07/26
UMN–TH–2607/07, FTPI–MINN–07/20
Light Heavy MSSM Higgs Bosons at Large
J. Ellis, S. Heinemeyer, K.A. Olive and G. Weiglein
TH Division, Physics Department, CERN, Geneva, Switzerland
Instituto de Fisica de Cantabria (CSICUC), Santander, Spain
William I. Fine Theoretical Physics Institute,
University of Minnesota, Minneapolis, MN 55455, USA
IPPP, University of Durham, Durham DH1 3LE, UK
Abstract
The region of MSSM Higgs parameter space currently excluded by the CDF Collaboration, based on an analysis of of integrated luminosity, is less than the expected sensitivity. We analyze the potential implications of the persistence of this discrepancy within the MSSM, assuming that the soft supersymmetrybreaking contributions to scalar masses are universal, apart from those to the Higgs masses (the NUHM model). We find that a light heavy MSSM Higgs signal in the unexcluded part of the sensitive region could indeed be accommodated in this simple model, even after taking into account other constraints from cold dark matter, electroweak precision observables and physics observables. In this case the NUHM suggests that supersymmetric signatures should also be detectable in the near future in some other measurements such as , and , and would have to be very close to the LEP exclusion limit. In addition, the dark matter candidate associated with this model should be on the verge of detection in direct detection experiments.
CERN–PH–TH/2007088
February 14, 2021
The searches for the bosons appearing in its extended Higgs sector are among the most promising ways to search for evidence of supersymmetry (SUSY) at the Tevatron collider. The CDF and D0 Collaborations have already established important limits on the (heavier) MSSM Higgs bosons, particularly at large [1, 2, 3, 4, 5]. Recently the CDF Collaboration, investigating the channel
(1) 
has been unable to improve these limits to the extent of the sensitivity expected with the analyzed integrated luminosity of [4], whereas there is no indication of any similar effect in D0 data [5].^{1}^{1}1The same analysis shows a deficit of events compared to the Standard Model (SM) expectation. Changing the luminosity by one to accomodate this ‘deficit’ would raise the observed rate of final states around slightly above the expected rate (by somewhat less than one ) [6]. Time will tell whether the CDF effect persists. Within the MSSM the channel (1) is enhanced as compared to the corresponding SM process by roughly a factor of [7], where is the ratio of the two vacuum expectation values, and includes loop corrections to the vertex (see Ref. [7] for details) and is subdominant for the final state. Correspondingly, the unexpected weakness of the CDF exclusion might be explicable within the MSSM if and and, doubtless, also within other theoretical frameworks.
In this paper we investigate whether light heavy Higgs bosons just beyond the region currently excluded by CDF could be accommodated within GUTinspired MSSM scenarios, and what the possible consequences would be. We consider the constraints imposed by other measurements, such as the limits on , , and , assuming that parity is conserved and that the lightest neutralino constitutes the astrophysical dark matter [8]. Whereas we find no solution within the constrained MSSM (CMSSM), in which all soft SUSYbreaking contributions to scalar masses are assumed to unify at the GUT scale, we find that all the constraints may be satisfied in the case that universality at the GUT scale is relaxed for the scalar Higgs mass parameters (the NUHM model [9, 10, 11]). However, we point out that any interpretation of the CDF effect within the NUHM would be tightly constrained by the other measurements. Specifically, the constraints are so tight that one or more of these measurements should display a discrepancy with the SM, either now or in the near future.
The essence of the argument is as follows. The absence of exclusion by CDF, compared to their expected sensitivity, as mentioned above, would correspond to , and a value of or greater. Since the contribution to , values of are already excluded for this value of for substantial portions of the NUHM parameter space, depending largely on the values of and . The parameter space is so constrained that, in areas which are still allowed, we expect that a SUSY signal should appear very soon, as we show below. We assume parity conservation, and restrict our attention to the NUHM with values of the relic CDM density that fall within the range favoured by WMAP and other astrophysical and cosmological observations. This restriction imposes important constraints on and , the values of and being less essential. As for , it is well known that the world average experimental value currently agrees well with the SM [12]. In the MSSM, there are two important contributions with opposite signs, due to and chargino exchanges, respectively. In order for the net MSSM contribution to be unnoticeable so far, the and chargino exchanges must cancel to a great extent, imposing a relation between the and chargino masses. Since the mass is very similar to the masses, this yields a preferred range of relatively small values of the chargino mass and hence the soft supersymmetrybreaking gaugino mass , favouring in turn a nonnegligible contribution to . The combination of a preferred value for and the WMAP constraint then limits the possible range of . The preference for relatively light sparticles (see also Ref. [13]) translates into a relatively small value for , and compatibility between the LEP limit on [14, 15] and the upper limit on [16, 17] selects a limited range of .
In preparation for our survey of the NUHM parameter space, we first recall that in the CMSSM, the electroweak vacuum conditions determine and in terms of and the input soft supersymmetrybreaking parameters. In the NUHM, these are the scalar mass (which is assumed to be universal, except for the Higgs multiplets) ^{2}^{2}2We discuss later the implications of relaxing this assumption., the gaugino mass , the trilinear coupling and the degrees of nonuniversality of the soft supersymmetrybreaking contributions to the masses of the two Higgs doublets. In our analysis, we invert the electroweak vacuum conditions, treating and as free parameters and adjusting the nonuniversal Higgs mass inputs accordingly. In the cases of interest this difference in the scalar masses at the GUT scale is no more than 50% for low values of , and of order 10% for higher values of . The values and are chosen to match the excess of signallike events observed in CDF, and we assume that so as to avoid severe problems with and . Hence our four free parameters are and . However, these are tightly constrained by other phenomenological limits, as we now discuss.
Since we assume that parity is conserved, and that the lightest neutralino constitutes the astrophysical dark matter [18], we impose the requirement that the relic neutralino density falls within the range allowed by WMAP and other observations: [8]. As we see later, requiring the relic density to fall within this narrow range effectively reduces the dimensionality of the NUHM parameter space. In order to apply the constraints on the NUHM parameter space that are provided by , , and , we use the following experimental values and theory evaluations (for more details, specifics on error treatments and an extended list of references see Ref. [19]): [12], where the theory evaluation is based on Refs. [20, 21]; [16, 17], where details about the theory evaluation can be found in Ref. [22]; [23] (for a discussion and references to other determinations, see Ref. [19]), and [14, 15], where the theory calculations have been performed with FeynHiggs [24]. These observables impose important constraints, as we now show ^{3}^{3}3Consequences for other observables such as and the boson mass are discussed further below..
We consider first the () plane shown in panel a) of Fig. 1, which has , and , as well as .^{4}^{4}4Here and elsewhere, we assume and , but our results are insensitive to the exact values of and , and take into account their current uncertainties, and . The dark (brown) shaded region at low is forbidden, because there the LSP would be the lighter stau. The WMAP cold dark matter constraint is satisfied only within the lighter (turquoise) shaded region. To the left of this region, the relic density is too small, due to channel annihilation through the Higgs pseudoscalar . As increases away from the pole, the relic density increases toward the WMAP range. However, as is increased, the neutralino acquires a larger Higgsino component and annihilations to pairs of and bosons become enhanced. To the right of this transition region, the relic density again lies below the WMAP preferred value. The shaded region here is therefore an overlap of the funnel and transition regions discussed in Ref. [10].
The constraint is satisfied between the outer black dashdotted lines, labelled , representing the current limit on that branching ratio^{5}^{5}5 A slightly more stringent upper limit of at the C.L. has been announced more recently by the D0 Collaboration [25]. However, applying this limit would have only a minor impact on our analysis. . Also shown are the contours where the branching ratio is , close to the sensitivity likely to be attainable soon by CDF and D0. Between these two contours, there is a strong cancellation between the flavorviolating contributions arising from the Higgs and chargino couplings at the oneloop level and the Wilson coefficient counterterms contributing to .
The dashdotted (red) line shows the contour corresponding to , and only the region to the right of this line is compatible with the constraint imposed by (it should be kept in mind that there is still a uncertainty in the prediction of [24]). Also shown in pink shading is the region favoured by at the two level. The one and two contours for are shown as elliptical dashed and solid black contours, respectively. The region which is compatible with the WMAP relic density and , and is also within the two experimental bound, has . The measured value of is in agreement with the theory prediction only to the left of the solid (green) region. We see, in the space between the and exclusions, a slightly diagonal allowed strip of width . in
We see that there is a narrow wedge of allowed parameter space in Fig. 1(a), which has and to 1100 GeV. The constraint is satisfied easily throughout this region, and cuts off the top of the wedge, which would otherwise have extended to . Within the allowed wedge, is very close to the LEP lower limit, and . If were much smaller (), there would be no wedge consistent simultaneously with the , and constraints (the bound, however, could then be relaxed due to a weaker coupling). At higher values of , the allowed region drifts up to higher values of , as is shown in Fig. 1(c) for for the same input values of , and .
We have also considered the potential impact of direct searches for supersymmetric cold dark matter in the NUHM. As has been pointed out previously [26], these searches are potentially important at small and large , as required in the scenario discussed here (see also Ref. [27]). However, the interpretation of the search limits is less precise than for the other constraints, for two reasons. One is the local density of supersymmetric cold dark matter, which is usually estimated as /cm. This estimate is subject to a systematic uncertainty that is itself uncertain, but might be or so. The second significant uncertainty is in the hadronic scattering matrix element of the local operator generated by the shortdistance supersymmetric physics. The dominant contribution to the scattering is spinindependent, and given by the matrix elements in the nucleons of the scalar quark densities : . These may be determined from the octet matrix element , which is estimated to be , and the N scattering matrix element . These give a range [28]. Generally speaking, the darkmatter scattering cross section increases with .
Given the matrix element uncertainties summarized above, we show in Fig. 1a) the 1 lower limit on the calculated value of the elastic cross section as compared to the CDMS upper limit [29]. In the portion of the plane to the left of the (orange) dotted line, the lower limit on the calculated spinindependent elastic cross section is smaller than the CDMS upper bound, assuming the canonical local density. Whilst we have assumed , the calculated lower limit effectively assumes zero strangeness contribution to the proton mass, i.e., . In the region of interest, the lower limit on the calculated cross section is about 80% of the CDMS upper bound, whereas with a strangeness contribution of , the cross section would exceed the CDMS bound by a factor of . Thus, if Nature has picked this corner of the NUHM parameter space, we expect direct detection of dark matter to be imminent. We note that the XENON Collaboration has recently announced a stronger upper limit on the elastic scattering cross section [30]. Consistency with this limit would further require a reduction in the local dark matter density (to its lower limit). Intriguingly, the XENON10 experiment has seen some potential signal events that are, however, interpreted as background. In the following we will show the limits obtained by CDMS, but the potentially somewhat stronger limits from XENON10 should be kept in mind.
This example was for the particular values and . We now investigate what happens if these values are varied. Panel a) of Fig. 2 explores the () plane for , values close to the lower tip of the allowed wedge in Fig. 1(a). In this case, the region allowed by the constraint is below the upper dashdotted black line, and the LEP constraint is satisfied only above the dashdotted red line. We see that only a restricted range is compatible simultaneously with the and constraints. Very large negative values of are excluded as the LSP is the lighter stau. On the () plane, the elastic scattering cross section is a rapidly decreasing function of and is almost independent of . Indeed, NUHM points excluded by CDMS (or XENON10) generally have low values of and [26]. Values of are compatible with CDMS if the strangeness contribution to the proton mass is negligible. For this choice of parameters, the entire displayed plane is compatible with and . is compatible with the dark matter constraint. This corresponds to the transition strip where the neutralino is the appropriate bino/Higgsino combination. To the left of this strip, the relic density is too small and to the right, it is too large. Only a very restricted range of
Panel b) of Fig. 2 shows what happens if is increased to 700 GeV, keeping and the other inputs the same. Compared to Fig. 2(a), we see that the WMAP strip becomes narrower and shifts to larger , and that starts to exclude a region visible at smaller . If were to be increased much further, the dark matter constraint and would no longer be compatible for this value of . We also see that, by comparison with Fig. 2(a), the constraint has moved to lower , but the constraint has dropped even further, and over the entire visible plane. The net result is a region compatible with all the constraints that extends from down to for , with a coannihilation filament extending to larger when . Once again all of the WMAP strip in this panel is compatible with CDMS.
The larger allowed area of parameter space is reflected in panel b) of Fig. 1, which has and , as well as and as before. In this case, we see that a substantial region of the WMAP strip with and a width , extending from to higher is allowed by all the other constraints. The and constraints have now moved to relatively low values of , but we still find and close to the LEP lower limit. We note that the and constraints now disfavour a bigger fraction of the parameter space with large , but their only effect here is that truncates the allowed region at . Once again, the allowed region is compatible with CDMS (to the left of the (orange) dotted line) provided the strangeness contribution to the proton mass is small.
Panels c) and d) of Fig. 1 show the constraints on the () plane for for the same set of input parameters used in panels a) and b). The dominant effects at higher are the shift to the left of the contours of and the exclusion of a much larger portion of the plane at low due to a stau LSP. Because of the strong dependence on , the constraint essentially excludes all values of and higher . At higher , the CDMS bound also becomes stronger and restricts and bounded by CDMS, , and . There, the branching ratio, is near its minimum value of due to the strong cancellation discussed above. In panel d) we are left with a significantly larger quadrilateral region. in panel d). Thus in panel c), we are forced into a small triangular region centered at in panel c) and in panel d). Nevertheless, a viable portion of parameter space remains in tact at lower in panel c) and
The analogous () planes are shown in Fig. 2c) and d). In panel c), only a small portion of the WMAP strip around is compatible with both and , while the range in panel d) is considerably larger. In both cases, the CDMS limit is stronger and begins to cut into the WMAP strip.
We have surveyed systematically the allowed region of parameters in the () plane for between 45 and 55, varying the inputs and . To summarize the regions in the NUHM parameter space which are compatible with CDM and all phenomenological constraints and a light heavy Higgs, we first recall that the values of compatible with WMAP depend on , but depend less on . Specifically, as increases, the preferred region of the () plane is a nearvertical strip that moves to larger , which is truncated at low just above the region where the lighter stau is the LSP. For , there is no WMAPcompliant region compatible with the LEP lower limit on , but a small allowed region appears for slightly below . Typical values for are at this value of and must be below . When , the WMAPcompatible strip moves to larger , and there are allowed regions of the () plane for to . As increases further, the WMAPcompatible strip moves to even larger , and the only allowed region is a small piece of coannihilation strip close to the boundary with the stau LSP region. This region is negligible for . Combining all the allowed values of and , we find that only a small portion of the () plane can ever be compatible with all the constraints. It is roughly triangular in shape, with vertices and .
We now analyse the dependence of the results on the assumption of scalarmass universality. The assumptions that all squarks with the same electroweak quantum numbers have identical input soft supersymmetrybreaking scalar masses is motivated in general by the suppression of flavourchanging neutral interactions. Specifically, the upper limit on and the agreement of with the SM would require a certain degree of finetuning in the presence of large squark mass nonuniversality. On the other hand, the relative locations of both the WMAP strip and the constraint depend on the relationship between the soft SUSYbreaking squark and slepton scalar masses. If the slepton masses are decreased relative to the squark masses (which we continue to denote by ), the lower ends of the WMAP strips in Fig. 1 will rise (due to relatively ligher staus), as will the contours, raising the preferred range of . Conversely, if the slepton masses are increased relative to the squark masses, the preferred range of will be lower. Thus, neglecting the universality between squark and slepton masses in general enlarges the allowed region of parameter space that is compatible with the experimental constraints.
We now discuss the possible phenomenological signatures of a scenario with and within the NUHM. The interplay of the various constraints in Fig. 1 implies:

The predicted value of in the allowed region is generally . Thus, this channel may offer good prospects within the near future for either supporting or contradicting the NUHM interpretation of the weaker CDF bound, as compared to the expected sensitivity.

We find that must be very close to the LEP lower limit, i.e. in the range where LEP observed a couple of Higgslike events [15]. This part of parameter space could be probed at the Tevatron with of integrated luminosity. Due to the large value of and the small mass of the boson, the rates of decays into bottom quarks and tau leptons are enhanced as compared to the SM.

The predicted value of in the allowed region is , which is about one above the current experimental value (if the experimental and theory errors are added linearly). Consequently, an improvement in the present theoretical uncertainty might enable a discrepancy to appear between and the SM value.

The discrepancy between the experimental measurement of and the SM calculation can easily be explained in this scenario, although the discrepancy should be somewhat smaller than the current central value. However, a much smaller discrepancy (corresponding to larger values) could also be accomodated.

Confronting the prediction for [31] with the measurement from Belle and BABAR [32] already yields interesting constraints on the charged Higgsboson mass as a function of , although the present experimental errors are still very large, see e.g. Ref. [33]. In the scenario considered here is predicted to be relatively low as compared to its SM value, where the ratio of MSSM/SM is . This is about one below the current central value.

The ratio of the mass difference to the SM prediction is close to unity, [31]. In view of the theoretical uncertainties, it will be difficult to establish such a small deviation.

The masses of the other SUSY particles are estimated to be –, –, –, –, –, –, –, –, –, and for the squarks , , , , each with uncertainties . The sparticle spectrum is not particularly light, and at the Tevatron no further SUSY particle discoveries could be expected. On the other hand, the stronglyinteracting sparticles should mostly be within reach of the LHC, and many weaklyinteracting sparticles should be visible in their cascade decays. At the ILC, depending on the centerofmass energy, the ligher neutralinos, charginos and staus could be produced.

Direct detection experiments should see a signal at current sensitivities. The lack of a signal in the CDMS experiment could be explained by a reduced strangeness contribution to the proton mass. The reported XENON10 limit would further require a reduction in the local halo density (to its lower limit).
Performing a fit for nine precision and physics observables along the lines of Ref. [19] yields a total value of in the allowed part of the NUHM parameter space, where even slightly smaller values can be found for ^{6}^{6}6It should be kept in mind that the lines shown in the plots are roughly the LEP 95% C.L. exclusion limits, not the 2 bound. In applying this bound the current theory uncertainty in the prediction of needs to be taken into account. .
Very likely the weaker CDF bound from the search for heavy Higgs bosons compared to its expected sensitivity is due to a statistical fluctuation that will eventually evaporate. Nevertheless, it is interesting to know whether a signal at this level could be accommodated within the MSSM. We have shown that is possible for within the NUHM, though it stretches various experimental constraints to their limits. Correspondingly, small improvements in some of these measurements, e.g., of or of the dark matter scattering rate, would either exclude such a signal in the NUHM framework, or else provide supporting evidence. One way or another, it should be possible soon to cast light on the interpretation of the CDF search.
Acknowledgements
We would like to thank Patricia Ball, Prisca Cushman, and Athanasios Dedes for useful conversations. The work of K.A.O. was partially supported by DOE grant DEFG0294ER40823. The work of S.H. was partially supported by CICYT (grant FPA2006–02315). Work supported in part by the European Community’s MarieCurie Research Training Network under contract MRTNCT2006035505 ‘Tools and Precision Calculations for Physics Discoveries at Colliders’.
References
 [1] A. Abulencia et al. [CDF Collaboration], Phys. Rev. Lett. 96 (2006) 011802, hepex/0508051.
 [2] V. Abazov et al. [D0 Collaboration], Phys. Rev. Lett. 97 (2006) 121802, hepex/0605009.
 [3] A. Abulencia et al. [CDF Collaboration], Phys. Rev. Lett. 96 (2006) 042003, hepex/0510065.

[4]
CDF Collaboration,
CDF note 8676, see:
http://wwwcdf.fnal.gov/~aa/mssm_htt_1fb/note/cdf8676.pdf. 
[5]
D0 Collaboration,
D0 Note 5331CONF, see:
http://wwwd0.fnal.gov/cgibin/d0note?5331. 
[6]
G. Landsberg,
talk given at Rencontres de Moriond:
QCD and Hadronic interactions,
La Thuile (Italy), March 1724 2007,
see:
http://moriond.in2p3.fr/QCD/2007/WednesdayMorning/Landsberg.pdf (p. 20).  [7] M. Carena, S. Heinemeyer, C. Wagner and G. Weiglein, Eur. Phys. J. C 45 (2006) 797, hepph/0511023.
 [8] D. Spergel et al. [WMAP Collaboration], astroph/0603449.
 [9] J. Ellis, K. Olive and Y. Santoso, Phys. Lett. B 539 (2002) 107, hepph/0204192.
 [10] J. Ellis, T. Falk, K. Olive and Y. Santoso, Nucl. Phys. B 652 (2003) 259, hepph/0210205.
 [11] M. Olechowski and S. Pokorski, Phys. Lett. B 344 (1995) 201, hepph/9407404; V. Berezinsky, A. Bottino, J. Ellis, N. Fornengo, G. Mignola and S. Scopel, Astropart. Phys. 5 (1996) 1, hepph/9508249; M. Drees, M. Nojiri, D. Roy and Y. Yamada, Phys. Rev. D 56 (1997) 276, [Erratumibid. D 64 (1997) 039901], hepph/9701219; M. Drees, Y. Kim, M. Nojiri, D. Toya, K. Hasuko and T. Kobayashi, Phys. Rev. D 63 (2001) 035008, hepph/0007202; P. Nath and R. Arnowitt, Phys. Rev. D 56 (1997) 2820, hepph/9701301; J. Ellis, T. Falk, G. Ganis, K. Olive and M. Schmitt, Phys. Rev. D 58 (1998) 095002, hepph/9801445; J. Ellis, T. Falk, G. Ganis and K. Olive, Phys. Rev. D 62 (2000) 075010, hepph/0004169; A. Bottino, F. Donato, N. Fornengo and S. Scopel, Phys. Rev. D 63 (2001) 125003, hepph/0010203; S. Profumo, Phys. Rev. D 68 (2003) 015006, hepph/0304071; D. Cerdeno and C. Munoz, JHEP 0410 (2004) 015, hepph/0405057; H. Baer, A. Mustafayev, S. Profumo, A. Belyaev and X. Tata, JHEP 0507 (2005) 065, hepph/0504001.

[12]
E. Barberio et al. [Heavy Flavor Averaging Group (HFAG)],
hepex/0603003,
see:
http://www.slac.stanford.edu/xorg/hfag/.  [13] J. Ellis, S. Heinemeyer, K. Olive and G. Weiglein, JHEP 0605 (2006) 005, hepph/0602220.
 [14] LEP Higgs working group, Phys. Lett. B 565 (2003) 61, hepex/0306033.
 [15] LEP Higgs working group, Eur. Phys. J. C 47 (2006) 547, hepex/0602042.

[16]
CDF Collaboration,
CDF Public Note 8176, see:
http://wwwcdf.fnal.gov/physics/new/bottom/060316.blessedbsmumu3/ . 
[17]
D0 Collaboration,
D0 Note 5344Conf,
see:
http://wwwd0.fnal.gov/Run2Physics/WWW/results/b.htm .  [18] H. Goldberg, Phys. Rev. Lett. 50 (1983) 1419; J. Ellis, J. Hagelin, D. Nanopoulos, K. Olive and M. Srednicki, Nucl. Phys. B 238 (1984) 453.
 [19] J. Ellis, S. Heinemeyer, K. Olive, A.M. Weber and G. Weiglein, arXiv:0706.0652 [hepph].
 [20] M. Misiak et al., Phys. Rev. Lett. 98 (2007) 022002, hepph/0609232.
 [21] P. Gambino and M. Misiak, Nucl. Phys. B 611 (2001) 338, hepph/0104034.
 [22] J. Ellis, K. Olive and V. Spanos, Phys. Lett. B 624 (2005) 47, hepph/0504196. J. Ellis, K. Olive, Y. Santoso and V. Spanos, JHEP 0605 (2006) 063, hepph/0603136.
 [23] M. Davier, hepph/0701163.
 [24] S. Heinemeyer, W. Hollik and G. Weiglein, Comp. Phys. Commun. 124 2000 76, hepph/9812320; Eur. Phys. J. C 9 (1999) 343, hepph/9812472; G. Degrassi, S. Heinemeyer, W. Hollik, P. Slavich and G. Weiglein, Eur. Phys. J. C 28 (2003) 133, hepph/0212020; M. Frank, T. Hahn, S. Heinemeyer, W. Hollik, H. Rzehak and G. Weiglein, JHEP 02 (2007) 047, hepph/0611326; the code is available via http://www.feynhiggs.de .

[25]
D0 Collaboration,
D0 Note 5344Conf,
see:
http://wwwd0.fnal.gov/Run2Physics/WWW/results/b.htm .  [26] J. Ellis, K. Olive, Y. Santoso and V. Spanos, Phys. Rev. D 71 (2005) 095007, hepph/0502001.
 [27] M. Carena, D. Hooper and P. Skands, Phys. Rev. Lett. 97 (2006) 051801, hepph/0603180.
 [28] J. Gasser, H. Leutwyler and M. Sainio, Phys. Lett. B 253 (1991) 252; M. Knecht, hepph/9912443; M. Sainio, PiN Newslett. 16 (2002) 138, hepph/0110413.
 [29] D. Akerib et al. [CDMS Collaboration], Phys. Rev. Lett. 96 (2006) 011302, astroph/0509259.

[30]
J. Angle et al. [XENON Collaboration],
arXiv:0706.0039 [astroph],
see also: http://xenon.astro.columbia.edu/ .  [31] G. Isidori and P. Paradisi, Phys. Lett. B 639 (2006) 499, hepph/0605012.
 [32] K. Ikado et al. [Belle Collaboration], hepex/0604018; B. Aubert et al. [BABAR Collaboration], arXiv:0705.1820 [hepex].
 [33] S. Villa, talk given at FPCP07, 5th Flavor Physics and CP Violation Conference, Bled (Slovenia), May 12–16 2007, see: http://wwwf9.ijs.si/fpcp07.
 [34] S. Heinemeyer, W. Hollik, D. Stöckinger, A.M. Weber and G. Weiglein, JHEP 08 (2006) 052, hepph/0604147.
 [35] S. Heinemeyer, W. Hollik and G. Weiglein, Phys. Rept. 425 (2006) 265, hepph/0412214.