1: \documentclass[12pt,preprint]{aastex}
2: \begin{document}
3: \title{Evolution of the Pancaking Effect in a $\Lambda$CDM Cosmology}
4: \author{Yookyung Noh and Jounghun Lee}
5: \affil{Department of Physics and Astronomy, FPRD, Seoul National
6: University, Seoul 151-747, Korea}
7: \email{ykyung@astro.snu.ac.kr,jounghun@astro.snu.ac.kr}
8:
9: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
10: \begin{abstract}
11: We explore the evolution of the large-scale anisotropy in the velocity field
12: caused by the gravitational pancaking effect assuming a $\Lambda$CDM universe.
13: The Millennium Run halo catalogs at four different redshifts, $z=0,\ 0.5,\ 1$
14: and $z=2$ are analyzed to find that the pancaking effect starts to intervene
15: the hierarchical structure formation at redshift $z=2$ when a characteristic
16: pancake scale is around $3\ h^{-1}$Mpc. It is also clearly shown how the
17: degree and scale of the pancaking effect changes with time.
18: An analytic model based on the Zel'dovich approximation is presented to
19: explain quantitatively the evolution of the velocity-pancake alignment.
20: A cosmological implication of our finding and a possibility of
21: detecting a signal in real universe are discussed.
22: \end{abstract}
23: \keywords{cosmology: theory --- large-scale structure of universe}
24:
25: \section{INTRODUCTION}
26:
27: The observed filamentary distribution of the large scale structure in the
28: universe, which are often called the cosmic web, have so far motivated
29: plenty of works. After the ingenious breakthrough made by \citet{bon-etal96},
30: a theoretical framework has been provided within which the characteristic
31: properties of the cosmic web can be explained qualitatively in terms of
32: the tidal influences in the universe dominated by the cold dark matter (CDM).
33:
34: The influence of the tidal forces on the orientations and distributions
35: of the large scale structure has been a target of many fruitful studies
36: which either numerically or observationally proved that the tidal effect
37: is strongest near the pancakes.
38:
39: A pancake refers to the two-dimensional sheet-like structure in the universe,
40: which form through gravitational collapse along the local tidal field.
41: The first order Lagrangian perturbation theory, the Zel'dovich approximation
42: \citep{zel70}, is probably the most popular model for the pancake formation
43: which was originally suggested assuming the hot dark matter (HDM) dominated
44: universe. The cosmic web theory, however, showed brilliantly that the
45: presence of pancakes are also very possible in the CDM-dominated universe
46: due to the large scale coherence of the tidal field. N-body simulations
47: of CDM particles indeed demonstrated that the first collapsed objects look
48: like pancakes \citep{sha-etal95}.
49:
50: The large-scale coherence of the tidal field that is responsible for
51: the formation of pancakes in turn causes the dark matter halos near or on
52: the pancakes to have strong spatial correlations in the orientations
53: \citep{ara-etal07,bru-etal07,hah-etal07}.
54: The reason that the degree of the tidally induced alignments is highest
55: near the pancakes can be understood given the fact that the pancakes
56: are the first collapsed objects, being still in linear regime while
57: the filaments are more severely modified by the nonlinear process.
58:
59: The formation of pancakes induces not only the spatial correlations in the
60: halo orientations but also the anisotropy in the velocity field.
61: In our previous paper \citep[][hereafter, NL06]{noh-lee06}, we have for the
62: first time detected a $5\sigma$ significant signal of the velocity anisotropy
63: induced by the pancaking effect in the high-resolution N-body simulation.
64: In spite that the signal is found to be very weak because of the non-conserved
65: nature of the velocity field, the velocity anisotropy holds a crucial key to
66: understanding the pancaking effect on the structure formation.
67:
68: In the light of the first detection of NL06, a couple of questions naturally
69: arise: When does the pancaking effect influence most dominantly the formation
70: of dark halos?; How does the pancake scale change with time?; What are the
71: implications and consequences of the pancaking effect on the structure
72: formation?; Is the standard theory capable of quantifying the velocity
73: anisotropy?. Our goal here is to provide answers to these questions.
74:
75: We organize this paper as follows. In \S 2, we introduce the analyzing
76: method to investigate the evolution of the pancaking effect in the Millennium
77: Simulation. We describe analytic modeling and compare it with the results
78: obtained in the simulation catalog in \S 3. Finally, we discuss and conclude
79: the results.
80:
81: \section{SIGNALS FROM SIMULATION}
82:
83: To investigate the evolution of local pancaking effect, we use the Millennium
84: Run simulation halo catalogs for the concordance $\Lambda$CDM cosmology
85: \citep{spr-etal05}. A periodic box of this simulation is $500\ h^{-1}$Mpc on
86: each side and the cosmological parameters are $\Omega_{m}=0.25$,
87: $\Omega_{\Lambda}=1-\Omega_{m}$, $h=0.73$, and $\sigma_{8}=0.9$. We choose four
88: different redshift catalogs, approximately $z=0,\ 0.5,\ 1$, and $2$ among 64
89: output times of the simulation. Then we analyze the alignments of dark-matter halo
90: velocities to the local pancake plane in each catalog. Basically, to find a local
91: pancake, we follow the methodology suggested by NL06.
92:
93: As a first step in measurement, we select the halos which contain more than
94: $50$ particles (the left column of Table \ref{tab:num}) since those halos in
95: Millennium Run catalog which have too low particle numbers (less than $30$)
96: are poorly sampled and severely contaminated by the numerical noise.
97: In other words, only those halos which have enough particles to define a halo
98: density profile are reliable (V. Springel in private communication).
99: Thus, we set the particle number cut-off at 50. Next, among the selected halos,
100: we pick out field halos which are still in quasi-linear regime by using the
101: method suggested by \citet{ela-pir97} and \citet{hoy-vog02}. The number of
102: identified field halos is shown in the right column of Table \ref{tab:num}.
103: The field halos are suited for our purpose because they may not be seriously
104: affected by gravitational forces of many other halos, retaining their
105: initial conditions at the moment that they collapsed.
106:
107: Using only those field halos, we determine local pancakes using the
108: practical methods proposed by NL06. First, We find two nearest field
109: halos for a field halo and define a local pancake plane enclosing all the
110: three halos. Here, the local pancakes are found by changing the criterion
111: distance $R_{c}$ from $0$ $h^{-1}$Mpc to $17$ $h^{-1}$Mpc which is a lower
112: limit of $R_{1}$, the displacement vector to the first nearest halo.
113: In other words, $R_{c}$ satisfies the following condition:
114: $R_{c} \le R_{1} \le R_{2}$ where $R_{2}$ is the displacement vector to the
115: second nearest neighbor.
116:
117: With the determined local pancakes, then we calculate the alignment angles between
118: a halo velocity vector and a vector normal to the local pancake plane, $\cos\theta$.
119: %by using equation (\ref{eq:two}) for the first method and by using equation
120: %(\ref{eq:thr}) for the second method.
121: %{\setlength\arraycolsep{2pt}
122: %\begin{eqnarray}
123: %\label{eq:two}
124: %\cos\theta &=& \frac{{\vert{\bf v}\cdot ({\bf R_{1}}\times{\bf R_{2}})\vert}}
125: %{{\vert{\bf v}\vert}{\vert{\bf R_{1}}\times {\bf R_{2}}\vert}}\\
126: %\label{eq:thr}
127: %\cos\theta&=&\frac{\vert{\bf v}\cdot\{ ({\bf R_{2}} - {\bf R_{1}})\times
128: %({\bf R_{3}}-{\bf R_{1}})\} \vert}{{\vert{\bf v}\vert}{\vert
129: %({\bf R_{2}} - {\bf R_{1}})\times ({\bf R_{3}}-{\bf R_{1}}) \vert}}
130: %\end{eqnarray}}
131: Finally, we obtain the probability density distribution of $\cos\theta$,
132: $p(\cos\theta)$, by counting the number of halos in each bin of $\cos\theta$.
133: We note that in order to examine the dependence of the particle number cut-off
134: which is used when we determine the halos, we recalculate the velocity-pancake
135: alignment for those halos which include more than $30$ particles though
136: low-particle number halos suffer from simulation noise. From this calculation,
137: we obtain almost perfectly consistent results with the results of the catalog
138: in which halos include more than $50$ particles as can be seen in
139: Fig. \ref{fig:pro} (particle number cut-off is $50$) and Fig. \ref{fig:pro30}
140: (particle number cut-off is $30$). Thus, the velocity-pancake alignment is
141: not affected by the choice of particle number cut-off.
142:
143: Figure \ref{fig:pro} shows $p(\cos\theta)$ with Poissonian error. The
144: three columns correspond to three different cases of $z=0,\ 1$, and $2$,
145: respectively, and the three rows to three different scales of $R_{c}=0,\ 3$,
146: and $6$, respectively. The horizontal dotted line in each panel represents a random
147: distribution. As can be seen in the right column, when $R_{c}$ is $0$, the halo
148: velocity tends to be slightly aligned with the normal to the local pancake plane
149: at $z=2$. This tendency becomes stronger as $R_{c}$ increases to $6 h^{-1}$Mpc.
150: It can be interpreted that the local pancaking effect arises at least around
151: $z\simeq 2$. Meanwhile, at $z=1$, for the case of $R_{c}=0$, there is no alignment.
152: Then local pancaking effect starts to be visible near $R_{c}=3 h^{-1}$Mpc.
153: Moreover, as one can see in the left column, when $R_{c}=0$ at $z=0$, the halo
154: velocity does not tend to be aligned with the normal to the local pancake plane
155: any more. Rather it tends to lie on the plane. The alignment signal begins to be
156: seen when $R_{c}$ is approximately $6 h^{-1}$Mpc. We infer from these phenomena
157: that the pancaking effect disappears since the gravitational attraction forces
158: from the close neighbor halos affect halo velocities rather than the initial
159: pancaking effect. Therefore, we find the velocity-pancake alignments are
160: strongest at $z\simeq2$, indicating that the pancaking effect starts at
161: $z\geqq 2$. We refer that we do not make more analysis of the velocity-pancake
162: alignments for higher redshift to find a redshift at which pancaking effect
163: starts since the number of halos decreases rapidly as redshift increases.
164: Even if the value of $\eta$ itself might increase, the value of $\sigma_{\eta}$
165: which is inversely proportional to the number of halos would increase.
166: Thus, it would be hard to determine precisely at which redshift both the
167: values of $\eta$ and $\sigma_{\eta}$ reach maximums.
168:
169: Figure \ref{fig:his} plots the average of $\cos\theta$ at four different $z=0,\
170: 0.5,\ 1$, and $2$ as a function of linear characteristic pancake scale, $L_{p}$.
171: We refer that $L_{p}$ is defined as the average distance between the center of
172: mass (CM) for three halos on the local pancake plane to all the three halos,
173: which was proposed by NL06. The horizontal dotted line in each panel represents
174: $\langle\cos\theta\rangle$ of random distribution of the angles. As one can see
175: in each panel, there is a specific range of $L_{p}$ where $\langle\cos\theta\rangle$
176: is relatively higher than those at the other pancake scales. This specific range
177: implies local pancaking effect significantly occurs on specific pancake scale.
178: Comparing to panels each other, overall $\langle\cos\theta\rangle$ increases, its
179: slope is steeper and the specific range shifts to small $L_{p}$ when going to high
180: redshift. This tendency can be clearly seen when we see that the regions of filled with
181: oblique lines, which represent the area consisting of a few of the highest values,
182: move to small $L_{p}$ as going to high redshift. More quantitatively, the peak value
183: is approximately $5.035\times 10^{-4}$ at $z=0$ while the one is $5.055\times
184: 10^{-4}$ at $z=2$. In addition, the peak position of histogram is approximately
185: $2\ h^{-1}$Mpc at $z=0$ whereas the one is $8\ h^{-1}$Mpc at $z=2$.
186: This gradual change with redshifts can be understood the same way as
187: Fig. \ref{fig:pro}; after pancaking effect on halo velocities arises,
188: it is gradually attenuated by gravitational force among halos, especially among
189: very close halos.
190:
191: \section{THEORETICAL ANALYSIS}
192: In order to investigate the tendency of the anisotropic distribution of halo
193: velocity theoretically, we construct an analytic model using the similar
194: methodology to \citet{lee-pen01}. We begin on the Zel'dovich
195: approximation \citep{zel70}
196: \begin{equation}
197: \label{eq:zel}
198: {\bf x}={\bf q}-D(t)\nabla\Psi({\bf q}),
199: \end{equation}
200: which intrinsically estimates the formation of pancakes. Here, ${\bf x}$ is the
201: Eulerian coordinate, ${\bf q}$ is the Lagrangian coordinate, $\Psi({\bf q})$ is
202: the linear velocity potential and $D(t)$ is the growth factor,
203: \begin{equation}
204: \label{eq:gro}
205: D=\frac{5}{2}\Omega_{m}H(a)\int_{0}^{a}\frac{da'}{[a'H(a')]^{3}}.
206: \end{equation}
207: By taking the time derivative of Zel'dovich approximation, we obtain equation
208: of the velocity of cosmic particle smoothed on pancake scale,
209: \begin{equation}
210: \label{eq:xdot}
211: v_{i}=-({\dot D}\partial_{i}\Psi + D\partial_{i}{\dot \Psi}) \propto
212: -({\dot D}q_{j}T_{ij}+Dq_{j}{\dot T}_{ij}).
213: \end{equation}
214: Equation (\ref{eq:xdot}) is appropriate to investigate pancaking effect on halo
215: velocity at early stage since equation (\ref{eq:zel}) describes initial state of
216: non-linear evolution. Also, we assume the density field is smoothed on the pancake
217: scale, which is valid in this approximation because this approximation is available
218: in the regime preceding the moment of the collapse to a pancake.
219:
220: Using a similar logic given by \citet{lee-pen01} who found an expression for the
221: alignments between the tidal field and the halo position vectors, we take into
222: account nonlinear effects into equation (\ref{eq:xdot}), and write an expression
223: for the velocity alignments with the tidal field as
224: {\setlength\arraycolsep{2pt}
225: \begin{eqnarray}
226: \label{eq:velcor}
227: \langle v_{i}v_{j}\vert {\bf T}\rangle&=&\frac{1-\eta}{3}\delta_{ij}+\eta K_{ij},
228: \\
229: \label{eq:K} K_{ij}&\equiv& {\dot D}^{2}T_{ik}T_{kj} +
230: 2D{\dot D} T_{ik}{\dot T_{kj}} + D^{2}{\dot T}_{ik}{\dot T}_{kj}
231: \end{eqnarray}}
232: where $\bf T$ and $\dot{\bf T}$ are tidal shear tensor and its time derivative,
233: respectively. $\dot D$ is time derivative of $D$, which is
234: {\setlength\arraycolsep{2pt}
235: \begin{eqnarray}
236: \label{eq:ddot}
237: & &{\dot D}=\frac{5}{2}\Omega_{m}\left[H'(a){\dot a} \int_{0}^{a}\frac{da'}
238: {\{a'H(a')\}^{3}}+ \frac{{\dot a}H(a)}{a^{3}H(a)}\right],\\
239: \label{eq:dhda}
240: \textrm{where}\quad H'(a)&=&\frac{dH(a)}{da}=-\frac{3}{2}\Omega_{m}a^{-4}(\Omega_{m}
241: a^{-3}+\Omega_{\Lambda})^{-1/2}, \quad
242: {\dot a}=aH_{0}(\Omega_{m}a^{-3}+\Omega_{\Lambda})^{1/2}.
243: \end{eqnarray}}
244: Notice that $\eta$ is a value in $[0,1]$. If $\eta$ is $1$, it represents
245: a perfect correlation between local pancaking and halo velocity while if $\eta$
246: is $0$, it means there is no correlation between them. In velocity-velocity
247: correlation, it is unlikely that the value of $\eta$ is close to unity
248: since the velocity of a halo is not a conserved quantity and initial memory of
249: local pancaking effect on velocity have been reduced.
250: In order to show that the velocities of halos are not completely random but
251: has some degree of alignments with the normal vectors to the local pancakes
252: though The value of $\eta$ is expected to be very small, we have calculated
253: the error on the value of $\eta$, that is, the standard deviation of $\eta$ for
254: the case of no alignment. The formula for the error of $\eta$, $\sigma_{\eta}$
255: is given as $(4/15N_{tot})^{1/2}$ where $N_{tot}$ is the total number of halos
256: \citep{lee-pen01}. With defined error, if $\eta$ is larger than $3\sigma_{\eta}$,
257: one can say that there is a signal of true local pancaking effect.
258:
259: We diagonalize equation (\ref{eq:velcor}) by applying the relation referred by
260: \citet{bon-etal96}
261: \begin{equation}
262: \label{eq:bond}
263: {\dot\lambda_{i}}=\frac{D\lambda_{i}}{1-D\lambda_{i}}
264: \end{equation}
265: where $\lambda_{i}$ and ${\dot \lambda_{i}}$ are an eigenvalue of the ${\bf T}$
266: and its time derivative, respectively. Thus, equation (\ref{eq:K})
267: is changed to
268: \begin{equation}
269: \label{eq:kdiag}
270: K_{ii}=\frac{\lambda_{i}^{2}[{\dot D}(1-D\lambda_{i})+D^{2}]^{2}}
271: {(1-D\lambda_{i})^{2}},\quad
272: K_{ij}=0 \; (i\ne j).
273: \end{equation}
274: The eigenvalues $\lambda_{i}$ are ordered by $\lambda_{1}>\lambda_{2}>\lambda_{3}$.
275:
276: To obtain the probability distribution, we assume probability density distribution of
277: the alignment angles, $\cos\theta$, is Gaussian \citep{lee04}
278: \begin{equation}
279: \label{eq:pden}
280: p({\hat v}\vert{\check {\bf T}})=\frac{1}{\sqrt{(2\pi)^{3}\det(M)}}\int_{0}^
281: {\infty}\exp\left[-\frac{v_{i}^{T}(M^{-1})_{ij}v_{j}}{2}\right]v^{2}dv,
282: \end{equation}
283: where $M_{ij}\equiv\langle v_{i}v_{j}\vert {\bf T}\rangle$.
284: Finally, we can express the probability distribution of $\cos\theta$, $p(\cos
285: \theta)$ is,
286: \begin{equation}
287: \label{eq:prob}
288: p(\cos\theta)=\frac{1}{2\pi}\prod_{i=1}^{3}(1+\eta-3\eta\check{\lambda}_{Ki})^
289: {-\frac{1}{2}}\int_{0}^{2\pi}\left(\frac{\sin^{2}\theta\cos^{2}\phi}{1+\eta-3\eta
290: \check{\lambda}_{K1}}+\frac{\sin^{2}\theta\sin^{2}\phi}{1+\eta-3\eta
291: \check{\lambda}_{K2}}+\frac{\cos^{2}\theta}{1+\eta-3\eta\check{\lambda}_{K3}}
292: \right)^{-\frac{3}{2}}d\phi.
293: \end{equation}
294: Here, ${\check \lambda}_{Ki} = {K_{ii}}/{\sum_{i} K_{ii}^{2}}$.
295:
296: To compare with the results from simulation analysis, we use $\Lambda$CDM
297: cosmological parameters, $h=0.73$, $\sigma_{8}=0.9$, $T_{cmb,0}=2.725$K,
298: $\Omega_{b}=0.045$, $\Omega_{m}=0.25$, which are the same values as
299: the parameters used in Millennium Simulation.
300: First, we calculate $\sigma$ by using following equation:
301: \begin{equation}
302: \label{eq:sig}
303: \sigma^{2} (R,z)=D(z)^{2}\int \frac{k^{2}dk}{2\pi ^{2}}P(k)T^{2}(k)
304: W_{TH}^{2}(kR),
305: \end{equation}
306: where $D(z)$ is a growth function, $P(k)$ is primordial power spectrum that is $P(k)
307: \propto k$, $T(k)$ is transfer function \citep{bar-etal86} and $W_{TH}$ is spherical
308: top-hat window function.
309: When evaluating equation (\ref{eq:sig}), we use $L_{p}$ obtained by simulation
310: analysis for $R$. Also, the redshift at which $L_{p}$ is determined in the simulation
311: is chosen as $z$ value in equation (\ref{eq:sig}).
312:
313: Then we obtain $\lambda_{Ki}$ in equation (\ref{eq:pden}). We decide
314: $\lambda_{K1}$ as $1$ because $\lambda_{1}=1$ means the formation of the structure
315: like a pancake in the Zel'dovich approximation,
316: \begin{equation}
317: \label{eq:rho}
318: \rho = \frac{\bar{\rho}}{(1-\lambda_{1})(1-\lambda_{2})(1-\lambda_{3})},
319: \end{equation}
320: where $\bar{\rho}$ is the mean density of the universe and $\lambda_{1},\ \lambda_{2},\
321: \lambda_{3}$ ($\lambda_{1} > \lambda_{2} > \lambda_{3}$) are the eigenvalues of the
322: local tidal field, ${\bf T}$.
323: For $\lambda_{K2}$ and $\lambda_{K3}$, we choose the most probable values of the
324: following probability distribution,
325: \begin{equation}
326: \label{eq:dor}
327: p(\lambda_{2},\lambda_{3}\vert\lambda_{1}=1) = p(\lambda_{1}=1,\lambda_{2},
328: \lambda_{3})/p(\lambda_{1}=1).
329: \end{equation}
330: Here, $p(\lambda_{1},\lambda_{2},\lambda_{3})$ and $p(\lambda_{1})$ are
331: the probability distribution derived by \citet{dor70} and \citet{lee-sha98}:
332: \setlength\arraycolsep{1pt}
333: \begin{eqnarray}
334: \label{eq:pl1l2l3}
335: &p&(\lambda_{1},\ \lambda_{2},\ \lambda_{3})=
336: \frac{3375}{8\sqrt{5}\pi\sigma^{6}}(\lambda_{1}-\lambda_{2})(\lambda_{2}
337: -\lambda_{3})(\lambda_{1}-\lambda_{3})\exp\left (-\frac{3I_{1}^{2}}{\sigma^{2}}
338: +\frac{15I_{2}}{2\sigma^{2}}\right )\\
339: \label{eq:pl1}
340: &p&(\lambda_{1})=
341: \frac{\sqrt{5}}{12\pi\sigma}\Bigg[
342: 20\frac{\lambda_{1}}{\sigma}\exp\left(-\frac{9\lambda_{1}^{2}}
343: {2\sigma^{2}}\right)-\sqrt{2\pi}\exp\left(-\frac{5\lambda_{1}^{2}}{2\sigma^{2}}\right)
344: \textrm{erf}\left(\sqrt{2}\frac{\lambda_{1}}{\sigma}\right)
345: \left(1-20\frac{\lambda_{1}^{2}}{\sigma^{2}}\right) \nonumber \\
346: & &\quad\quad -\ \sqrt{2\pi}\exp\left(-\frac{5\lambda_{1}^{2}}
347: {2\sigma^{2}}\right) \left(1-20\frac{\lambda_{1}^{2}}{\sigma^{2}}\right)
348: +3\sqrt{3\pi}\exp\left(-\frac{15\lambda_{1}^{2}}{4\sigma^{2}}\right)
349: \textrm{erf}\left(\frac{\sqrt{3}\lambda_{1}}{2\sigma}\right) \nonumber \\
350: & &\quad\quad +\ 3\sqrt{3\pi}\exp\left(-\frac{15\lambda_{1}^{2}}{4\sigma^{2}}\right)\Bigg].
351: \end{eqnarray}
352: We put these eigenvalues in equation (\ref{eq:kdiag}) to obtain the normalized eigenvalues,
353: $\check{\lambda}_{Ki}$.
354:
355: Finally, we fit theoretical estimation of $p(\cos\theta)$, equation
356: (\ref{eq:prob}), to the numerical results as adjusting
357: a correlation parameter $\eta$ in equation (\ref{eq:prob}).
358: We determine $\eta$ minimizing $\chi^{2}$ distribution
359: \begin{equation}
360: \label{eq:chi}
361: \chi^{2}=\sum_{i}\frac{x_{s,i}-x_{t,i}}{\sigma_{s,i}},
362: \end{equation}
363: where $x_{s,i}$ and $x_{t,i}$ are the values from simulation analysis and from theory,
364: respectively and $\sigma_{s,i}$ is the Poissonian error from simulation analysis.
365: The comparison numerical results with theory is shown in Fig. \ref{fig:the}.
366: The filled circle points represent the numerically measured $p(\cos\theta)$ with
367: Poissonian errors and solid line represents the analytic distribution. As can be seen
368: in the figure, the simulation data points are in good agreement with our analytic
369: formulation. Also, as shown in Table \ref{tab:eta}, the strength of local pancaking
370: effect on the halo velocity depends on the redshift. At $z=2$, though halos are
371: closely located at each other like $R_{c}=3 h^{-1}$Mpc, $\eta$ is
372: $0.011$, which turns out to be $15.3\sigma_{\eta}$ (see Table \ref{tab:eta}).
373: Thus, even though the value of $\eta$ is much less than unity, it is
374: definitely a strong signal of velocity-pancake alignments, exceeding $15$ times
375: the standard deviation, $\sigma_{\eta}$. This strong signal which approximately
376: amounts to $15\sigma_{\eta}$ shows the existence of the velocity-pancake alignment.
377: Even when $R_{c}=0$ at $z=2$, $\eta$ is $0.003$ which is $4\sigma_{\eta}$
378: which indicates that a pancaking effect appears even at $R_{c}=0$. Then going to present epoch, the
379: effect is reduced by gravitational attraction. Finally, at $z=0$, when $R_{c}$ is
380: $3 h^{-1}$Mpc, $\eta$ is almost $0$ and $\sigma_{\eta}$ is also $0$. In addition,
381: as one can see the tendency in Fig. \ref{fig:the} and in Table \ref{tab:eta}, at
382: $z=2$, pancaking effect becomes stronger as $R_{c}$ is approximately $5-6 h^{-1}$Mpc
383: and then $\eta$ and $\sigma_{\eta}$ gradually decreases when $R_{c}$ is larger than
384: $6 h^{-1}$Mpc. In other words, the value of $\eta$ and $\sigma_{\eta}$ at $z=2$ is
385: the largest with $R_{c}\simeq 6$. A similar tendency also appears at different redshifts,
386: instead the range of $R_{c}$ is shifted to larger scale caused by the expansion of
387: the universe. Thus, we analytically find that there is a specific range of pancake
388: scale where pancaking significantly affects halo velocity. In addition, we show how
389: the pancaking effect changes with different redshifts.
390:
391: To find the tendency of the redshift dependence of $\eta$, we fit $\eta$ as a function
392: of redshift to quadratic equation. The coefficients are shown in Table \ref{tab:fz}.
393: Figure \ref{fig:zco} plots the quadratic fitting function (dashed line) and compares
394: it with the simulation data points (filled circles). As Fig. \ref{fig:zco} shows,
395: we note that $\eta$ is a function of redshift, in a good agreement with the quadratic
396: equation. Moreover, $\eta$ is a function of the characteristic pancake scale,
397: $L_{p}$. We also fit $\eta(L_{p})$ to quadratic equation which can be seen in
398: Fig. \ref{fig:rco}. The dashed line and the filled circles represent the fitting
399: function and the data points, respectively. We can see a good match between the
400: quadratic functional form and the simulation data. The best-fit coefficients are shown
401: in Table \ref{tab:fLp}.
402:
403: \section{DISCUSSION AND CONCLUSION}
404:
405: We analyze the evolution of the pancaking effect by using millennium simulation.
406: We obtain the signal of the strong alignment between the halo velocity and the
407: normal to the local pancake plane at $z\simeq 2$ when characteristic pancake
408: scale is approximately $3\ h^{-1}$Mpc. The alignment signal implies the first
409: collapse of the protocloud may have occurred around $z\simeq 2$ and the size
410: of the pancake may have been around $3\ h^{-1}$Mpc.
411: This result is in agreement with \citet{mo-etal05} which
412: reports pancakes form at $z\simeq 2$ and their masses is approximately $5\times
413: 10^{12} \textrm{M}_{\odot}$.
414:
415: Then as shown in Fig. \ref{fig:pro} and Fig. \ref{fig:his}, the alignment signal
416: evolves. Pancaking effect is weakened as going to low redshift, which means the
417: pancaking effect on halo velocity is gradually attenuated by the gravitational
418: attraction among the neighbor halos. In addition, pancake scale at low redshift
419: is larger than that at high redshift.
420:
421: We expect the local pancaking effect could be detected in observation.
422: It may be possible to achieve the signal of the velocity alignment using the velocity
423: data reconstructed from 2MASS redshift survey \citep{erd-etal06}.
424: We intend to measure the signal in observation for our future project.
425:
426: It is interesting to note that the pancakes are expected to have formed coincidentally
427: at $z\simeq 2$ when the star formation rate in massive galaxies is the highest
428: \citep{jun-etal05,feu-etal05}, which is one of the phenomena referred to `downsizing'.
429: In `downsizing' scenario, the stars in more massive galaxies formed at higher redshift
430: and those in less massive galaxies recently formed, where `antihierarchy' can be
431: suggested. Recently, however, \citet{nei-etal06,mou-tan06} show this phenomenon is
432: natural in hierarchical clustering process. Nevertheless, a coincidence between the
433: pancake formation epoch and the epoch of the heavily star-forming may imply the
434: structure formation is not simply hierarchical but complicated \citep{cim-etal06}.
435:
436: In conclusion, we obtain the result that is the evolution of the velocity anisotropy
437: induced by local pancake formation. This result would help to understand the unknowns
438: in the galaxy formation.
439:
440:
441: \acknowledgments
442: The Millennium Run simulation used in this paper was carried out
443: by the Virgo Supercomputing Consortium at the Computing Centre of the
444: Max-Planck Society in Garching (http://www.map-garching.mpg.de/millennium).
445: We are thankful to L. Gerard for the halo catalogs
446: of the Millennium Run simulation and to V. Springel for useful discussion.
447: We also appreciate the anonymous referee for helpful suggestions.
448: This work is supported by the research grant No. R01-2005-000-10610-0 from
449: the Basic Research Program of the Korea Science and Engineering Foundation.
450:
451:
452: \clearpage
453: \begin{thebibliography}{100}
454: \bibitem[Aragon-Calvo et al.(2007)]{ara-etal07}
455: Aragon-Calv, M. A. ,van de Weygaert, R. Jones, B. J. T., \& van der Hulst, J. M.
456: 2007, \apj, in press (astro-ph/0610249)
457: \bibitem[Bardeen et al.(1986)]{bar-etal86}
458: Bardeen, J.M., Bond, J.R., Kaiser, N., \& Szalay, A.S. 1986, \apj, 304, 15
459: \bibitem[Bond, Kofman, \& Pogosyan(1996)]{bon-etal96}
460: Bond, J., R., Kofman, L., \& Pogosyan, D. 1996, Nature, 380, 603
461: \bibitem[Brunino et al.(2007)]{bru-etal07}
462: Brunino, R., Trujillo, I., Pearce, F. R., \& Thomas, P. A. 2007,
463: \mnras, in press (astro-ph/0609629)
464: \bibitem[Cimatti et al.(2006)]{cim-etal06}
465: Cimatti, A., Daddi, E., \& Renzini, A. 2006, \aap, 453, 29
466: \bibitem[Croton et al.(2006)]{cro-etal06}
467: Croton, D. J., Springel, V., White, S. D. M., De Lucia, G., Frenk,
468: C. S., Gao, L., Jenkins, A., Kauffmann, G., Navarro, J. F., \&
469: Yoshida, N. 2006, \mnras, 367, 864
470: \bibitem[Doroshkevich(1970)]{dor70}
471: Doroshkevich, A. G. 1970, Astrofizika, 3, 175
472: \bibitem[El-Ad \& Piran(1997)]{ela-pir97}
473: El-Ad, H., \& Piran, T. 1997, \apj, 491, 421
474: \bibitem[Feulner et al.(2005)]{feu-etal05}
475: Feulner, G., Gabasch, A., Salvato, M., Drory, N., Hopp, U., \& Bender, R. 2005, \apj, 633, L9
476: \bibitem[Glazebrook et al.(2005)]{gla-etal05}
477: Glazebrook, K. et al. 2005, \nat, 430, 181
478: \bibitem[Hahn et al.(2007)]{hah-etal07}
479: Hahn, O., Porciani, C., Carollo, M. C., \& Dekel A. 2007, \mnras, 375,
480: 489
481: \bibitem[Hoyle \& Vogeley(2002)]{hoy-vog02}
482: Hoyle, F., \& Vogeley, M. S. 2002, \apj, 566, 641
483: \bibitem[Juneau et al(2006)]{jun-etal05}
484: Juneau, S. et al. 2005, \apj, 619L, 135
485: %Glazebrook, K., Crampton, D., McCarthy, P. J., Savaglio, S., Abraham, R.,
486: %Carilberg, R. G., Chen, H.-W., Le Borgne, D.Marzke, R. O., Roth, K., Jorgensen, I.,
487: %Hook, I., Murowinski, R.
488: \bibitem[Lee \& Shandarin(1998)]{lee-sha98}
489: Lee, J. \& Shandarin, S. F. 1998, \apj, 500, 14L
490: \bibitem[Lee \& Pen(2001)]{lee-pen01}
491: Lee, J., \& Pen, U. L. 2001, \apj, 555, 106
492: \bibitem[Lee \& Pen(2002)]{lee-pen02}
493: Lee, J., \& Pen, U. L. 2002, \apj, 567, L111
494: \bibitem[Lee(2004)]{lee04}
495: Lee, J. 2004, \apj, 614, L1
496: \bibitem[Mo et al.(2005)]{mo-etal05}
497: Mo, H. J., Yang, X., van den Bosch, F. C., \& Katz, N. 2005, \mnras, 363, 1155
498: \bibitem[Mouri \& Taniguchi(2006)]{mou-tan06}
499: Mouri, H., \& Taniguchi, Y. 2006 \aap, 459, 371
500: \bibitem[Noh and Lee(2006)]{noh-lee06}
501: Noh, Y., \& Lee, J. 2006, \apj, 652, L71
502: \bibitem[Neistein et al.(2006)]{nei-etal06}
503: Neistein, E., van den Bosch, F. C., \& Dekel, A. 2006, \apj, 372, 933
504: \bibitem[Pirin et al.(2006)]{erd-etal06}
505: Erdogdu, P., Lahav, O., Huchra, J. P., Colless, M., Cutri, R. M., Falco, E.,
506: George, T., Jarrett, T., Jones, D. H., Macri, L. M., Mader, J., Martimbeau, N.,
507: Pahre, M. A., Parker, Q. A., Rassat, A., \& Saunders, W. 2006, \mnras, 373, 45
508: \bibitem[Shandarin et al.(1995)]{sha-etal95}
509: Shandarin, S. F., Melott, A. L., McDavitt, K., Pauls, J. L., \& Tinker, J. 1995,
510: Phys. Rev. Lett., 75, 7
511: \bibitem[Spergel et al.(2003)]{spe-etal03}
512: Spergel, D. N. et al. 2003, \apj, 148, 175
513: \bibitem[Springel et al(2005)]{spr-etal05}
514: Springel, V. et al. 2005, \nat, 435, 629
515: \bibitem[Zel'dovich(1970)]{zel70}
516: Zel'dovich, Ya. B. 1970, \aap, 5, 84
517: \end{thebibliography}
518:
519: %%%%%%%%%%%%%%%
520: \clearpage
521: %%%%%%%%%%%%%%%
522: \begin{figure} \begin{center} \epsscale{1.0} \plotone{f1.eps}
523: \caption{Probability density distributions of the cosines of the relative
524: angles between the field halo velocities and the directions
525: normal to the local planes enclosing the two neighbor field halos
526: for the cases of the distance threshold: $R_{c}=0,\ 3,\ 6\ h^{-1}$Mpc
527: at different redshift: $z=0,\ 1,\ 2$ from Millennium Run simulation.
528: The errors are Poissonian and
529: the horizontal dotted lines represent no alignment. Note that halos which
530: contain more than $50$ particles are determined}
531: \label{fig:pro}
532: \end{center}
533: \end{figure}
534: %%%%%%%%%%%%%%%
535: \clearpage
536: %%%%%%%%%%%%%%%
537: \begin{figure} \begin{center} \epsscale{1.0} \plotone{f2.eps}
538: \caption{The same figure as Fig. \ref{fig:pro} but using the halos which
539: contain more than $30$ particles.}
540: \label{fig:pro30}
541: \end{center}
542: \end{figure}
543: %%%%%%%%%%%%%%%%
544: \clearpage
545: %%%%%%%%%%%%%%%
546: \begin{figure} \begin{center} \epsscale{1.0} \plotone{f3.eps}
547: \caption{The average of the cosines angles ($\cos\theta$) between the halo
548: velocities and the local pancake planes whose definition is the
549: same as Fig. \ref{fig:pro}, corresponding to the change of characteristic
550: linear size of the pancake, $L_{p}$ at four different redshift: $z=0,\ 0.5,\ 1,\ 2$.
551: The horizontal dotted lines represent no correlation.}
552: \label{fig:his}
553: \end{center}
554: \end{figure}
555: %%%%%%%%%%%%%%%%
556: \clearpage
557: %%%%%%%%%%%%%%%
558: \begin{figure} \begin{center} \epsscale{1.0} \plotone{f4.eps}
559: \caption{Probability density distributions of the $\cos\theta$s. Filed circle with
560: Poissonian errors represent the results from the simulation while solid line stand for
561: the analytic prediction with correlation parameter $\eta$. The horizontal dotted lines
562: correspond to no correlation.}
563: \label{fig:the}
564: \end{center}
565: \end{figure}
566: %%%%%%%%%%%%%%%%
567: \clearpage
568: %%%%%%%%%%%%%%%
569: \begin{figure} \begin{center} \epsscale{1.0} \plotone{f5.eps}
570: \caption{$\eta$ as a function of redshifts. Filed circle represents the numerical
571: measurement in Millennium Simulation. Dashed line is a quadratic fitting function
572: at different $R_{c}$ scale.}
573: \label{fig:zco}
574: \end{center}
575: \end{figure}
576: %%%%%%%%%%%%%%%%
577: \clearpage
578: %%%%%%%%%%%%%%%
579: \begin{figure} \begin{center} \epsscale{1.0} \plotone{f6.eps}
580: \caption{The same as Fig. \ref{fig:zco} but as a function of characteristic pancake
581: scale, $L_{p}$.}
582: \label{fig:rco}
583: \end{center}
584: \end{figure}
585: %%%%%%%%%%%%%%%
586: \clearpage
587: %%%%%%%%%%%%%%%
588: \begin{deluxetable}{ccc}
589: \tablewidth{0pt}
590: \setlength{\tabcolsep}{5mm}
591: \tablehead{\colhead{$z$}&\colhead{$N_{total}$}&\colhead{$N_{field}$}
592: ($\case{N_{field}}{N_{total}}$ \%)}
593: \tablecaption{{\it left}: Total number of identified halos whose contain
594: more than $50$ particles in each redshift catalog, {\it right}: The number of
595: field halos \label{tab:num}}
596: \startdata
597: $0$ & $6126689$ & $502108$ ($8.2$) \\
598: $0.5$ &$6631145$ & $542554$ ($8.2$)\\
599: $1$ & $6817225$ & $558138$ ($8.2$)\\
600: $2$ & $6222187$ & $516774$ ($8.3$)\\
601: \enddata
602: \end{deluxetable}
603: %%%%%%%%%%%%%%%
604: \clearpage
605: %%%%%%%%%%%%%%%
606: \begin{deluxetable}{ccccc}
607: \tablewidth{0pt}
608: \setlength{\tabcolsep}{5mm}
609: \tablecaption{Correlation parameter $\eta$ corresponding to $R_{c}$ and
610: redshift, $z$ with its standard deviation for the case of no alignment.
611: The formula for the error of $\eta$, $\sigma_{\eta}$ is given as $(4/15N_{tot})^{1/2}$
612: where $N_{tot}$ is the total number of halos \citep{lee-pen01}. \label{tab:eta}}
613: %\tablecolumns{5}
614: %\tablehead{\colhead{}&\colhead{}&\multicolumn{4}{c}{$z$}}
615: \tablehead{\colhead{$R_{c}$} &\multicolumn{4}{c}{z}\\
616: \cline{2-5}
617: \colhead{($h^{-1}$Mpc)} & \colhead{$0$} & \colhead{$0.5$} & \colhead{$1$} & \colhead{$2$} }
618: %\cutinhead{$z$}
619: \startdata
620: %\multicolumn{5}{c}{$z$}\\
621: %\cutinhead{$z$}
622: %\cline{3-5}
623: %\cutinhead{$R_{c}$}
624: $3$ & $0.000$($0.0\sigma_{\eta}$) & $0.003$($4.3\sigma_{\eta}$) &
625: $0.004$($5.8\sigma_{\eta}$) & $0.011$($15.3\sigma_{\eta}$) \\
626: $6$ & $0.005$($6.9\sigma_{\eta}$) & $0.008$($11.4\sigma_{\eta}$) &
627: $0.009$($13.0\sigma_{\eta}$) & $0.013$($18.1\sigma_{\eta}$) \\
628: $9$ & $0.007$($9.6\sigma_{\eta}$) & $0.009$($12.8\sigma_{\eta}$) &
629: $0.011$($15.9\sigma_{\eta}$) & $0.012$($16.7\sigma_{\eta}$) \\
630: \enddata
631: \end{deluxetable}
632: %%%%%%%%%%%%%%%
633: \clearpage
634: %%%%%%%%%%%%%%%
635: \begin{deluxetable}{cccc}
636: \tablewidth{0pt}
637: \setlength{\tabcolsep}{5mm}
638: \tablehead{\colhead{$z$}&\colhead{$a$}&\colhead{$b$}&
639: \colhead{$c$}}
640: \tablecaption{Fitting parameters for $\eta$ as a function of redshift, $z$.
641: Fitting functions are quadratic function, $\eta (z)=az^{2}+bz+c$, ($10^{-3}$)
642: \label{tab:fz}}
643: \startdata
644: $0$ & $1.36$\phd & $-1.28$ & $0.08$ \\
645: $0.5$ & $1.00$\phd & $3.30$ & $0.30$ \\
646: $1$ & $-0.36$\phd & $4.58$ & $5.22$ \\
647: $2$ & $-1.36$\phd & $5.28$ & $ 6.92$ \\
648: \enddata
649: \end{deluxetable}
650: %%%%%%%%%%%%%%%
651: \clearpage
652: %%%%%%%%%%%%%%%
653: \begin{deluxetable}{cccc}
654: \tablewidth{0pt}
655: \setlength{\tabcolsep}{5mm}
656: \tablehead{\colhead{$z$}&\colhead{$a$}&\colhead{$b$}&
657: \colhead{$c$}}
658: \tablecaption{The same as Table \ref{tab:fz} but as a function
659: of linear characteristic pancake scale, $L_{p}$, $\eta (L_{p})=aL_{p}^{2}+bL_{p}
660: +c$, ($10^{-3}$) \label{tab:fLp}}
661: \startdata
662: $0$ & $-0.31$ & $4.73$ & $-9.71$\\
663: $0.5$ & $-0.40$ & $5.36$ & $-9.26$\\
664: $1$ & $-0.37$ & $4.76$ & $-6.62$\\
665: $2$ & $-0.30$ & $3.39$ & $2.58$\\
666: \enddata
667: \end{deluxetable}
668:
669: \end{document}
670:
671: