astro-ph0005365/nu.tex
1: 
2: 
3: % add refs to Tritium and beta decay for direct mass limits
4: % explain that neutrinos heavier than cosmologically allowed range 
5: % are too heavy to be HDM
6: %Also add delta m^2 refs 
7: % mention lyman alpha forest, DLAS, sigma_8 constraints
8: % ~30eV is cosmological bound for pure HDM closing universe
9: 
10: 
11: 
12: \section{Is Cold+Hot Dark Matter Compatible with a Cosmological Constant?}  
13: 
14: 
15: The conflict between the $\Lambda$CDM model favored by direct 
16: observations of cosmological parameters (especially 
17: the recent Type Ia supernovae results) and the current set of 
18: structure formation observations that we have found motivates us to 
19: explore whether there is a simple way to reconcile $\Lambda$CDM with 
20: those observations.  
21: We start with a version of $\Lambda$CDM which is in 
22: good agreement with all direct parameter observations, with $\Omega_m=0.4$, 
23: $\Omega_b=0.04$, and $h=0.7$.   
24: Figure \ref{fig:lchdm00} 
25: shows the $\Lambda$CDM power spectrum compared 
26: with observations of Large-Scale Structure and CMB anisotropy.
27: The problem of the peak location and shape of the matter power spectrum 
28: versus the data is the same as that identified in the preceding chapter;
29: the era of matter-radiation equality occurs later in this $\Lambda$CDM 
30: model than the data appears to prefer, and improving this requires increasing
31: $\Omega_m$ which is the opposite of what direct parameter observations 
32: recommend.
33: 
34: 
35: One can pose the following question: does adding a hot component improve the  
36: marginally acceptable LSS fit?  
37: We find that as 
38: HDM is added, the combined fit to CMB and LSS deteriorates
39: (see Figures \ref{fig:lchdm05} and 
40: \ref{fig:lchdm10}).  
41: This occurs because adding HDM reduces the power on physical 
42: scales shorter than the neutrino free-streaming length, 
43: which 
44: exacerbates
45: the mismatch in peak locations that occurs for pure CDM low-density models.  
46: A blue tilt of the 
47: primordial power spectrum ($n=1.3$) 
48: is necessary to counteract the damping 
49: of small-scale perturbations by free-streaming of the 
50: massive neutrinos, which makes the peak of the model fall 
51: even farther below that of the data unless $n>1$.
52: Even with this best-fit value of $n$, the 
53: fit to the data is worse than with no HDM, because 
54: CMB observations disfavor such a high value of $n$.  
55: For a higher HDM fraction, an even higher value of $n$ is required 
56: ($n=1.5$ for the $\Omega_\nu=0.10$ model of Figure \ref{fig:lchdm10}), 
57: leading to an even worse fit to the data.  
58: $\Lambda$CHDM has also been explored by 
59: \citet{valdarninikn98} and \citet{primackg98} with different analysis 
60: methods and significantly smaller data compilations. 
61: 
62: 
63: \section{Limits on the Neutrino Mass}
64: 
65: 
66: Neutrino masses imprint a distinct signature on $P(k)$, 
67: the reduction in power on scales larger 
68: than the free-streaming length of 
69: 41 Mpc ($m_\nu$/30eV)$^{-1}$ \citep{bondes80}.  
70: Present cosmological bounds on the mass of a light neutrino are 
71: stricter than those from laboratory experiments; a 30eV neutrino 
72: would lead to $\Omega_\nu=1$, so for a universe at less than 
73: critical density the neutrinos must all be lighter than this.  
74: The exception to this is if the neutrino is so massive 
75: that it was non-relativistic during matter-radiation equality, i.e. 
76: Cold Dark Matter.   
77: For massive CDM neutrinos, the abundance drops enough to no longer overclose 
78: the universe; however, 
79: laboratory limits rule out the possibility of an electron 
80: neutrino more massive than 15eV \citep{primackg98}.   
81: The shape of the radiation power spectrum of CMB anisotropies 
82: \citep{dodelsongs96} 
83: and of 
84: the matter power spectrum from large-scale structure are both sensitive 
85: to the mass of neutrinos, and 
86: the LSS probe may potentially
87: be more powerful \citep{huet98}.
88:  There is more dynamical
89: range available in probing $P(k)$ with 
90: LSS on the neutrino free streaming scale, 
91: where the primary signature should be present.  
92: Even current data is sensitive to a neutrino 
93: mass of around an eV; 
94: we find that the fit changes significantly between 0.1 and 1 eV.  
95: While there are 
96: considerable systematic uncertainties in this approach, it is promising as a 
97: complement to the direct evidence for mass difference between neutrino 
98: species from SuperKamiokande \citep{fukudaetal98} and the solar neutrino 
99: problem \citep{bahcallks98}, 
100: and is already beginning to conflict with results from 
101: LSND 
102: \citep{athanassopoulosetal96} that require a large mass difference.  
103: 
104: 
105: 
106: We have assumed here that  
107: $\Lambda$CDM is the correct model of structure formation and 
108: that the primordial power spectrum is well-described by a power-law. 
109: Our limits on the neutrino mass are based upon an attempt to search 
110: the reasonable parameter space around this fiducial model to 
111: produce the best fit possible to the data for a given neutrino 
112: mass.  Since disagreement with CMB data is the main problem once 
113: a blue tilt is considered, we have tried to alleviate this by 
114: adding a significant tensor component or early reionization.  Each of these 
115: effects reduces the small-scale CMB power relative to COBE scales, which 
116: helps to reconcile $n>1$ with the CMB data.  However, no parameter 
117: combination helps enough to make $\Lambda$CHDM a better fit than 
118: the fiducial $\Lambda$CDM model, and this allows us to set 
119: upper limits on the neutrino mass.  
120: For $n=1$ we find that $\Omega_\nu<0.05$ i.e. 
121: the mass of the most massive neutrino must be 2 eV or less (the limit is 
122: tighter, of course, if there are at least 2 massive neutrinos with nearly 
123: equal masses).  For a scale-free primordial power spectrum, there is more 
124: freedom to increase $n$ to counteract the effect of HDM but this leads to 
125: conflict with CMB anisotropy observations, limiting $\Omega_\nu<0.1$ i.e. 
126: the mass of the most massive neutrino must be 4 eV or less.  This is 
127: compatible with the recent claim by 
128: \citet{crofthd99} 
129: that the Lyman $\alpha$ 
130: forest power spectrum limits the neutrino mass to 3 eV or less, but our method 
131: appears more robust as the normalization of the Lyman $\alpha$ 
132: forest power spectrum is quite uncertain and it covers a more narrow 
133: range of scales than our large-scale structure compilation.  Future constraints
134: from combining CMB and large-scale structure are discussed by \citet{huet98} 
135: and a more speculative method using weak gravitational lensing surveys 
136: is presented by 
137: \citet{cooray99}
138: .  
139: 
140: \cleardoublepage
141: \begin{figure}[htb]
142: %\figurenum{1}
143: %\epsffile{lchdm00_ebb.ps}
144: \centerline{\psfig{file=lchdm00.ps,width=6in}}
145: \caption{Constraints from LSS and CMB on $\Lambda$CDM model.} 
146: \label{fig:lchdm00}
147: \end{figure}
148: 
149: \cleardoublepage
150: \begin{figure}[htb]
151: %\figurenum{2}
152: \centerline{\psfig{file=lchdm05.ps,width=6in}}
153: \caption{$\Lambda$CDM model with $\Omega_\nu=0.05$.}
154: \label{fig:lchdm05}
155: \end{figure}
156: 
157: \cleardoublepage
158: \begin{figure}[htb]
159: %\figurenum{3}
160: \centerline{\psfig{file=lchdm10.ps,width=6in}}
161: \caption{$\Lambda$CDM model with $\Omega_\nu=0.10$ and $n=1.5$.} 
162: \label{fig:lchdm10}
163: \end{figure}
164: 
165: 
166: 
167: %\cleardoublepage
168: %\begin{figure}[htb]
169: %\figurenum{1}
170: %\epsffile{lchdm00_ebb.ps}
171: %\centerline{\psfig{file=tlchdm_cmb.ps,width=6in}}
172: %\caption{Constraints from CMB on $\Lambda$CHDM model with tensor 
173: %perturbations.} 
174: %\label{fig:tlchdm_cmb}
175: %\end{figure}
176: 
177: %\cleardoublepage
178: %\begin{figure}[htb]
179: %%\figurenum{1}
180: %%\epsffile{lchdm00_ebb.ps}
181: %\centerline{\psfig{file=n15tlchdm10.ps,width=6in}}
182: %\caption{Constraints from LSS and CMB on $\Lambda$CHDM model with 
183: %tensor perturbations.} 
184: %\label{fig:tlchdm10}
185: %\end{figure}
186: 
187: %%\cleardoublepage
188: %\begin{figure}[htb]
189: %%\figurenum{1}
190: %%\epsffile{lchdm00_ebb.ps}
191: %\centerline{\psfig{file=tau05n15lchdm10.ps,width=6in}}
192: %\caption{Constraints from LSS and CMB on reionized $\Lambda$CHDM model} 
193: %\label{fig:taulchdm10}
194: %\end{figure}
195: 
196: 
197: