1: %%
2: %% TANG Sumin 03/23/2005
3: %%
4: %% The first piece of markup in an AASTeX v5.x document
5: %% is the \documentclass command. LaTeX will ignore
6: %% any data that comes before this command.
7:
8: %% The command below calls the preprint style
9: %% which will produce a one-column, single-spaced document.
10: %% Examples of commands for other substyles follow. Use
11: %% whichever is most appropriate for your purposes.
12: %%
13: \documentclass[12pt,preprint]{aastex}
14:
15: %% manuscript produces a one-column, double-spaced document:
16:
17: %% \documentclass[manuscript]{aastex}
18:
19: %% preprint2 produces a double-column, single-spaced document:
20:
21: %%\documentclass[preprint2]{aastex}
22:
23: %% Sometimes a paper's abstract is too long to fit on the
24: %% title page in preprint2 mode. When that is the case,
25: %% use the longabstract style option.
26:
27: %% \documentclass[preprint2,longabstract]{aastex}
28:
29: %% If you want to create your own macros, you can do so
30: %% using \newcommand. Your macros should appear before
31: %% the \begin{document} command.
32: %%
33: %% If you are submitting to a journal that translates manuscripts
34: %% into SGML, you need to follow certain guidelines when preparing
35: %% your macros. See the AASTeX v5.x Author Guide
36: %% for information.
37:
38: \newcommand{\vdag}{(v)^\dagger}
39: \newcommand{\myemail}{skywalker@galaxy.far.far.away}
40:
41: %% You can insert a short comment on the title page using the command below.
42:
43: %% \slugcomment{Not to appear in Nonlearned J., 45.}
44:
45: %% If you wish, you may supply running head information, although
46: %% this information may be modified by the editorial offices.
47: %% The left head contains a list of authors,
48: %% usually a maximum of three (otherwise use et al.). The right
49: %% head is a modified title of up to roughly 44 characters.
50: %% Running heads will not print in the manuscript style.
51:
52: \shorttitle{No QSO Redshift Periodicities or Association with
53: Galaxies}
54: \shortauthors{TANG $\&$ ZHANG}
55:
56: %% This is the end of the preamble. Indicate the beginning of the
57: %% paper itself with \begin{document}.
58:
59: \begin{document}
60:
61: %% LaTeX will automatically break titles if they run longer than
62: %% one line. However, you may use \\ to force a line break if
63: %% you desire.
64:
65: \title{Critical Examinations of QSO Redshift Periodicities and Associations with Galaxies in Sloan Digital Sky Survey Data}
66:
67: %% Use \author, \affil, and the \and command to format
68: %% author and affiliation information.
69: %% Note that \email has replaced the old \authoremail command
70: %% from AASTeX v4.0. You can use \email to mark an email address
71: %% anywhere in the paper, not just in the front matter.
72: %% As in the title, use \\ to force line breaks.
73:
74: \author{Su Min Tang\altaffilmark{1} and Shuang Nan Zhang\altaffilmark{1,2,3,4}}
75: \email{tangsm99@mails.tsinghua.edu.cn; zhangsn@tsinghua.edu.cn}
76:
77:
78: \altaffiltext{1}{Department of Physics and Center for
79: Astrophysics, Tsinghua University, Beijing 100084, China}
80: \altaffiltext{2}{Department of Physics, University of Alabama in
81: Huntsville, Optics Building 201C, Huntsville, AL 35899}
82: \altaffiltext{3}{Space Science Laboratory, NASA Marshall Space
83: Flight Center, SD50, Huntsville, AL 35812}
84: \altaffiltext{4}{Institute of High Energy Physics, Chinese Academy
85: of Sciences, P.O. Box 918-3, Beijing 100039, China}
86:
87:
88: \begin{abstract}
89: We have used the publicly available data from the Sloan Digital
90: Sky Survey and 2dF QSO Redshift Survey to test the hypothesis that
91: QSOs are ejected from active galaxies with periodic
92: non-cosmological redshifts. For two different intrinsic redshift
93: models, namely the Karlsson $\log(1+z)$ model and Bell's
94: decreasing intrinsic redshift (DIR) model, we do two tests
95: respectively. First, using different criteria, we generate four
96: sets of QSO-galaxy pairs and find there is no evidence for a
97: periodicity at the predicted frequency in $\log(1+z)$, or at any
98: other frequency. We then check the relationship between high
99: redshift QSOs and nearby active galaxies, and we find the
100: distribution of projected distance between high redshift QSOs and
101: nearby active galaxies and the distribution of redshifts of those
102: active galaxies are consistent with a distribution of simulated
103: random pairs, completely different from Bell's previous
104: conclusion. We also analyze the periodicity in redshifts of QSOs,
105: and no periodicity is found in high completeness samples, contrary
106: to the DIR model. These results support that QSOs are not ejected
107: from active galaxies.
108: \end{abstract}
109:
110:
111: \keywords{quasars: general --- galaxies: active --- large scale
112: structure of universe}
113:
114:
115: \section{Introduction}
116:
117: The debate on whether QSOs are ejected from the nuclei of
118: low-redshift galaxies with a periodic non-cosmological
119: ``intrinsic" redshift has been going on for many years. Some
120: evidence has been claimed to suggest such an intrinsic redshift
121: hypothesis, in which QSOs have redshifts that are much larger than
122: their parent galaxies and the excess of redshift is assumed to
123: represent an always redshifted intrinsic component \citep{bb67,
124: arp90, kar90, chu98, bn01, bell04b, arp05}.
125:
126: Two models have been discussed in the literature which predict
127: exact values for the preferred redshifts. One of these is the
128: Karlsson formula which suggests a periodicity of $\triangle
129: \log(1+z_{eff})=0.089$ with peaks lying at 0.061, 0.30, 0.60,
130: 0.96, 1.14, 1.96 and so on (Karlsson 1977, 1990; Arp et al. 1990,
131: 2005; Burbidge $\&$ Napier 2001, 2003), where $z_{eff}$ is the
132: redshift of the QSO measured relative to the nearby galaxy, called
133: effective redshift, which is defined as:
134: \begin{equation}
135: 1+z_{eff}=(1+z_Q)/(1+z_G)
136: \end{equation}
137: where $z_Q$ is the observed quasar redshift and $z_G$ is the
138: redshift of the associated galaxy which is assumed to be the
139: ejecting galaxy. To explain such a periodicity, they claimed that
140: quasars are ejected by active galaxies and the putative parent
141: galaxies are generally much brighter than their quasar off-springs
142: \citep{arp05}. As claimed by Burbidge $\&$ Napier (2001, 2003),
143: the typical projected association separation is about 200 kpc.
144:
145: Another model, namely decreasing intrinsic redshift model (DIR
146: model), was proposed by Bell (2004), where the QSO intrinsic
147: redshift equation is given by the relation:
148: \begin{equation}
149: z_{iQ}=z_f(N-M_N)
150: \end{equation}
151: where $z_f=0.62\pm0.01$ is the intrinsic redshift constant, $N$ is
152: an integer, and $M_N$ varies with $N$ and is a function of a
153: second quantum number $n$. In the DIR model, galaxies are produced
154: continuously through the entire age of the universe, and QSOs are
155: assumed to be ejected from the nuclei of active galaxies and
156: represent the first very short lived stage ($10^7-10^8$ yr) in the
157: evolution of galaxies \citep{bell04b}.
158:
159: The above an intrinsic redshift hypothesis, if true, will have
160: far-reaching consequences for cosmology and the nature of QSOs.
161: Most of those previous studies on the Karlsson formula used rather
162: small samples (except for Arp et al. 2005), and have been
163: suspected that the claimed peaks were due to artifacts associated
164: with selection effects \citep{basu05}. To avoid such a
165: heterogeneous selection manner as well as personal prejudice,
166: Hawkins et al. (2002) tested the periodicity in $\log(1+z_{qso})$
167: with 2dF redshift survey data with 67291 nearby galaxies and 10410
168: QSOs; it was found that there is no periodicity in
169: $\log(1+z_{qso})$. However, Napier $\&$ Burbidge (2003) argued
170: that in order to use the 2dF sample to properly test the original
171: hypothesis, it is necessary to establish for each pair that the
172: galaxy is at least a late-type active spiral system. Arp et al.
173: (2005) also re-examined the 2dF sample and claimed that they found
174: that the redshifts of brighter QSOs in the QSO density contours
175: fit very exactly the long standing Karlsson formula and confirm
176: the existence of preferred values in the distribution of quasar
177: redshifts.
178:
179:
180:
181: In an attempt to resolve these issues, we turn to the Sloan
182: Digital Sky Survey (SDSS) (and also 2dF QSO Redshift Survey (2QZ)
183: occasionally) to carry out this study, which have the largest
184: homogeneous sample of data as well as the spectroscopic
185: sub-classification of galaxies. In section 2, to test whether
186: there is a periodicity existing in $\log(1+z)$, we construct four
187: sets of QSO-galaxy pairs with different QSOs and galaxies, with
188: all QSOs projected within 200 kpc from nearby galaxies at these
189: galaxies' distances. QSO density contours are also presented to
190: show that there is no periodicity in SDSS QSOs under such
191: analysis. In section 3, we examine the relationship between high-z
192: QSOs and nearby active galaxies to show that such QSOs are not
193: likely to be ejected by nearby active galaxies. In section 4, we
194: analyze the redshift distribution of QSOs in SDSS DR3 and 2QZ to
195: show that there is no evidence for non-artificial periodicity in
196: redshifts of QSOs, contrary to the DIR model. Discussion and
197: conclusion are described in Section 5.
198:
199: \section{No Periodicity in $\log(1+z)$}
200:
201: \subsection{The SDSS Data and Pair Selection}
202: In this section, we use the SDSS DR1 QSO catalog \citep{sch03} and
203: the New York University Value-Added Galaxy Catalog (NYU-VAGC)
204: \citep{bla04}. For reliability in the derived redshifts, we
205: consider only those QSOs in the range of $z>0.4$, and galaxies in
206: the range of $0.01<z<0.2$ with the highest plate quality labeled
207: as ``good" and with no redshift warning. This quality control
208: leaves a total of 190591 galaxies and 15747 QSOs in the sample.
209:
210: Three issues need to be carefully addressed when we analyze the
211: relation between foreground galaxies and QSOs, as well as the
212: redshift distribution of QSOs, since due to the survey strategies
213: and the instrumental limitations, the selections of galaxies and
214: QSOs are not entirely independent, and the selection of QSOs in
215: SDSS is also dependent on redshift. First, due to the mechanical
216: constraint in SDSS that spectroscopic fibers must be separated by
217: $55''$ on a given plate (Blanton et al. 2003), consequently some
218: QSO-galaxy pairs would be missing from the spectroscopic sample.
219: However, this issue would have little effect on the results for
220: two reasons: (1) $55''$ corresponds to an angular distance of $40$
221: kpc for a galaxy at $z=0.04$, which is the typical value of
222: redshift in our galaxy sample, only few pairs (about $5\%$) would
223: be missing in a given separation of 200 kpc for randomly
224: distributed QSOs and galaxies, which is also shown in Fig.~7 where
225: the distribution of data pairs are consistent with randomly
226: distributed pairs; (2) such fiber constraint is independent of
227: redshifts of galaxies or QSOs, therefore its redshift distribution
228: will not be biased, although some pairs are missing in the sample.
229: Second, the magnitude limits of the SDSS galaxy and QSO
230: spectroscopic surveys are quite different, i.e., $i<19.1$ for
231: $z<3$ QSOs, $i<20.2$ for $z>3$ QSOs, $r<17.77$ for most sampled
232: galaxies and $r<19.5$ for luminous red galaxies (Richards et al.
233: 2002; Strauss et al. 2002), hence the magnitudes of QSOs are
234: mostly higher than galaxies. However, since in the ejection
235: hypothesis, the parent galaxies are generally much brighter than
236: their QSO off-springs (Bell 2004b; Arp et al. 2005), the pair
237: making process are not likely to be affected by the magnitude
238: differences between QSOs and galaxies, which is also shown in
239: Fig.~12 and Fig.~13. Moreover, the fact that the completeness of
240: the spectroscopic selection varies with redshift (Richards et al.
241: 2002) will consequentially affect the redshift distribution of
242: QSOs and might cause artificial periodicities into data, as will
243: be discussed extensively in Section 4. However, since low-z
244: ($z<2.5$) QSOs which have flat and high completeness level
245: ($>90\%$; Richards et al. 2002) occupied a very large fraction of
246: all QSOs (about $90\%$), such selection dependence in redshift
247: would not smear out intrinsic periodicities in QSO redshift if
248: they do exist.
249:
250:
251: It has been suggested that quasars with bright apparent magnitude
252: will generally be nearby and the redshifts of these quasars will
253: require little or no correction for the periodicity effects to be
254: manifested, and those low redshift galaxies with which such
255: quasars appear to be preferentially associated, tend to be
256: morphologically disturbed active galaxies \citep{arp05}.
257: Therefore, to make our results more compelling, we select a
258: sub-sample of 3724 QSOs, called bright QSOs, which have i-band
259: magnitudes less than 18.5, and a sub-sample of 77426 galaxies,
260: called active galaxies, which are labeled as starforming,
261: starburst, starforming broadline or starburst broadline galaxies
262: in the spectroscopic sub-classification. Then we construct four
263: sets of QSO-galaxy samples by intercrossing them, and get four
264: sets of pairs in which a QSO is projected within 200 kpc from a
265: galaxy: 4572 pairs for QSO-nearby galaxies, 3216 pairs for
266: QSO-active nearby galaxies, 1129 pairs for bright QSO-nearby
267: galaxies and 791 pairs for bright QSO-active nearby galaxies. When
268: there is more than one galaxy within the 200 kpc projected
269: distance limit of the QSO, we take the closest galaxy in projected
270: distance to make up the pair.
271:
272: \subsection{Analysis and Results}
273:
274: We make power spectrum analysis to investigate the periodicity
275: hypothesis of Karlsson (1977). The power $I$ is defined as in
276: Burbidge $\&$ Napier (2001):
277: \begin{equation}
278: I(\nu)=2R^2/N,
279: \end{equation}
280: where
281: \begin{equation}
282: R^2=S^2+C^2,
283: \end{equation}
284: with
285: \begin{equation}
286: S=\sum^{N}_{i=1} w_i \sin(2\pi x_i/P), C=\sum^{N}_{i=1} w_i
287: \cos(2\pi x_i/P),
288: \end{equation}
289: with $\nu=1/P$ and $x_i=\log(1+z_i)$. Here $w_i$ is a weighting
290: function, and $w_i\equiv 1$ except in section 4, as in the
291: analysis of Burbidge $\&$ Napier (2001). For randomly and
292: uniformly distributed redshifts, $\overline {I}=2$.
293:
294: To test our code developed for this study, we first re-analyze the
295: 290 QSOs in Karlsson and Napier $\&$ Burbidge's data sets
296: \citep{kar90,bn01,bn03}, as shown in Fig.~1. Errors on $I(P)$ are
297: given by using the bootstrap methods \citep{efr79} in the
298: following steps: (1) we take the non-zero number in each bin in
299: the upper histogram as the expectation value of a Poisson
300: distribution; (2) we re-sample each bin following the Poisson
301: distribution to re-produce 1000 new sets of data, repeat the power
302: spectral analysis on these re-samplings, and finally calculate the
303: standard deviations in the derived values of $I$ at different
304: periods $P$. Clearly the periodicity at around $\triangle
305: \log(1+z_{eff})=0.089$ is highly significant at above 3.5$\sigma$
306: level.
307:
308: In Figs.~2-5 we show histograms of the effective redshifts of QSOs
309: paired with galaxies and their unwindowed power spectra with
310: standard deviations calculated in the same way as for Fig.~1.
311: Pairs in these four figures are for QSO-nearby galaxies,
312: QSO-active nearby galaxies, bright QSO-nearby galaxies and bright
313: QSO-active nearby galaxies, as described in Sec. 2.1. Our results
314: show that for these significantly larger samples than that in
315: Fig.~1, all peaks appeared in the power spectra are consistent
316: with Poisssonian fluctuations, i.e., there is no evidence for a
317: periodicity at the predicted frequency in $\log(1+z)$, or at any
318: other frequency.
319:
320: \subsection{QSO Density Contours}
321: After the work of Hawkins et al. (2002) on 2dF data, Arp et al.
322: (2005) argued that the predicted periodic redshifts are apparent
323: in the brighter 2dF quasars in the QSO density contours. We
324: therefore use SDSS DR1 data to construct the contours defined by
325: Arp et al. (2005), where the whole region is divided into boxes
326: $\bigtriangleup z \times \bigtriangleup B =0.075 \times 0.3$ in
327: the redshift/apparent magnitude plane, then the number of quasars
328: in each box is counted. To show whether the predicted periodic
329: redshifts are obscured by our coarse grid sizes, a contour with
330: $\bigtriangleup z \times \bigtriangleup B =0.05 \times 0.2$ is
331: also presented for comparison. As shown in Fig.~6, the peak
332: positions are consistent in the two contour plots, and there is no
333: evidence for redshift peaks at the predicted positions.
334:
335:
336:
337:
338:
339: \section{No Strong Connection between Active Galaxies and Bell's High-z QSOs}
340:
341: In Bell (2004b), a high-redshift QSO sample from SDSS and a
342: low-redshift QSO and QSO-like object sample from Hewitt $\&$
343: Burbidge (1993) were presented. Though the dips at redshifts of
344: 2.7 and 3.5 have been explained as being caused by the lower
345: efficiency of the selection algorithm at these redshifts
346: \citep{ric02}, Bell (2004b) nevertheless claimed that the
347: corresponding redshift peaks at 3.1 and 3.7 in the high-z SDSS
348: QSOs come from the intrinsic redshift broadening by Doppler
349: ejection and Hubble flow components, which is in favor of the DIR
350: model. Through analysis of the profiles of such peaks, Bell
351: (2004b) derived a mean cosmological components to be $z_c \sim
352: 0.066$ for the high-z sample.
353:
354:
355: In the DIR model, galaxies are produced continuously through the
356: entire age of the universe, and QSOs are assumed to be ejected
357: from the nuclei of active galaxies and represent the first very
358: short lived stage ($10^7-10^8$ yr) in the evolution of galaxies.
359: If this is true, there must be some connection between foreground
360: active galaxies and high-z QSOs beyond gravitational lensing. Here
361: we examine the high-redshift samples taken from the SDSS data. We
362: test the relationship between 2691 QSOs with redshifts in
363: 2.4$\sim$4.8 and 77426 nearby active galaxies with redshifts in
364: 0.01$\sim$0.2 from NYU-VAGC, all of which have the highest plate
365: quality labeled as ``good" and with no redshift warning. We
366: inter-compare these two data sets to find all QSO-galaxy pairs
367: within an angular separation corresponding to less than a given
368: distance $D_{separation}$ from several kpc to 1 Mpc at the
369: distance of the galaxy. In some cases, there is more than one
370: galaxy within the $D_{separation}$ projected distance limit of the
371: QSO; for these objects we take the closest galaxy in projected
372: distance to make up the pair. Since it is suggested that all QSOs
373: are born out of active galaxies and QSOs should be significantly
374: fainter than their parent galaxies (Bell 2004b), we would not miss
375: a considerable fraction of parent active galaxies for high-z QSOs
376: if the DIR model is right.
377:
378: The distribution of projected separation distance for all pairs is
379: shown in Fig.~7, and the redshift distribution of active galaxies
380: in pairs with QSOs is shown in Fig.~8. Both of them are consistent
381: with random distributions, but totally different from the
382: distribution from the ejection simulation with a ejection velocity
383: of 11,000 km s$^{-1}$ which was given by Bell (2004b) as typical
384: values, and the mean redshift of these galaxies is $z\sim0.044$,
385: also significantly different from Bell's result of 0.066. Here the
386: random distributions are obtained by moving the positions of all
387: galaxies by 1 degree in random directions; thus these galaxies
388: should be completely unrelated to background QSOs. The ejection
389: simulation is done by ejecting all QSOs from randomly selected
390: active galaxies with three ejection velocities: 11,000 km
391: s$^{-1}$, 40,000 km s$^{-1}$ and 80,000 km s$^{-1}$, and with a
392: uniformly distributed age of $0 \sim 10^8$ yr which is given by
393: Bell (2004b) as a typical value.
394:
395: To quantitatively show the differences between simulations and
396: `true' QSO-galaxy pairs, i.e., pairs found in the data but not
397: necessarily physical pairs, results of chi-squared tests are given
398: in Table 1. In $3\sigma$ confidence level for both distributions
399: of projected separation distance and redshift distribution of
400: active galaxies in pairs with QSOs, the `true' QSO-galaxy pairs
401: are consistent with random distributions, but inconsistent with
402: ejection hypothesis with ejection velocity up to 80,000 km
403: s$^{-1}$.
404:
405: \begin{table}
406: \begin{center}
407: \caption{Results of chi-squared tests for the distribution of
408: projected separation distance and the redshift distribution of
409: active galaxies in pairs with high-z QSOs between `true' pairs and
410: simulations. \label{tbl}}
411:
412:
413: \begin{tabular}{c|ccc|ccc}
414: \tableline\tableline
415:
416: simulations & $\chi^2$ \tablenotemark{a}& $\chi^2$/N \tablenotemark{a}& p \tablenotemark{a} & $\chi^2$ \tablenotemark{b}& $\chi^2$/N \tablenotemark{b}& p \tablenotemark{b}\\
417: \tableline
418: randomly distributed galaxies & 38.456 & 1.6720 & 0.023 & 18.818 & 0.9904 & 0.47 \\
419: ejected QSOs, v=11,000 km/s & 2560.3 & 111.3174 & $<10^{-10}$ & 994.10 & 52.3211 & $<10^{-10}$ \\
420: ejected QSOs, v=40,000 km/s & 227.12 & 9.8748 & $<10^{-10}$ & 265.31 & 13.9637 & $<10^{-10}$ \\
421: ejected QSOs, v=80,000 km/s & 85.496 & 3.7172 & $4.0\times10^{-9}$ & 144.37 & 7.5984 & $<10^{-10}$ \\
422: \tableline
423: \end{tabular}
424:
425: %% Any table notes must follow the \end{tabular} command.
426:
427: \tablenotetext{a}{Column 2-3 are for the distribution of projected
428: separation distance. Pairs with projected separation distance less
429: than 60 kpc (first and second points in Fig.~7) are not used in
430: the tests to avoid the SDSS 55'' fiber constraint, and all
431: simulations are normalized to get a minimum $\chi ^2$. The degrees
432: of freedom are $N=23$.}
433:
434: \tablenotetext{b}{ Column 4-7 are for the redshift distribution of
435: active galaxies in pairs with high-z QSOs. All the first zero
436: point in Fig.~8 is not used in the tests and all simulations are
437: normalized to get a minimum $\chi ^2$. The degrees of freedom is
438: $N=19$.}
439:
440:
441: \end{center}
442: \end{table}
443:
444: \section{No Periodicity in $z$}
445:
446: We also analyze the periodicity in redshifts of QSOs in SDSS DR3
447: (Schneider et al. 2005) and 2dF (Croom et al. 2004) to investigate
448: in larger database of QSOs whether there is a periodicity of
449: $\Delta z = 0.67 \pm 0.05$, predicted by the DIR model. Six data
450: sets are used in this section: all 46,420 QSOs in SDSS DR3
451: (Fig.~9), 22,497 QSOs with the highest quality flag in 2dF
452: (Fig.~10), a high completeness (close to $100\%$) sub-sample
453: containing 23,109 QSOs with Galactic-extinction corrected $i$-band
454: magnitude ($m_i$) less than 19 and redshift less than 2 in SDSS
455: DR3 (Richards et al. 2002) (Fig.~11(a)), and three sub-samples
456: containing QSOs in low completeness (less than $50\%$) regions in
457: SDSS DR3: 15,696 QSOs with $m_i>19$ and $z<2.4$ (Fig.~11(b)),
458: 19,064 QSOs with $m_i>19$ in all redshifts (Fig.~11(c)), and 9,763
459: QSOs with $z>2$ (Fig.~11(d)). To reduce the edge effect produced
460: by the truncated redshift distribution which has a lower redshifts
461: cut-off due to the small space volume sampled and a higher
462: redshifts cut-off due to the observational flux limit (see e.g.
463: Hawkins et al. 2002), we follow Hawkins et al. to use the Hann
464: function as a weighting in equation 5,
465:
466: \begin{equation}
467: w_i=\frac{1}{2}[1-\cos(\frac{2\pi x_i}{L})],
468: \end{equation}
469: where $L$ is chosen to cover the range of redshifts. Here $L=5.4$
470: for the full SDSS sample, $L=3.1$ for the 2dF sample, and $L=1.95,
471: 2.1, 5.1$ and 3.4 for the four SDSS sub-samples respectively.
472:
473: After smoothing off the sharp edges in the lowest and highest
474: redshifts, a periodicity around $\Delta z = 0.67$ is detected in
475: the full sample of SDSS QSOs, as shown in Fig.~9; however a
476: periodicity of $\Delta z = 0.67 \pm 0.05$ or any other frequency
477: is not found in the 2dF QSOs, as shown in Fig.~10. Such a
478: difference between these two surveys is not surprising since the
479: redshift-dependent spectroscopic completeness is relatively flat
480: in 2dF (Croom et al. 2004), while in SDSS the spectroscopic
481: completeness varies drastically at some redshifts (Richards et al.
482: 2002). It is therefore improper to use all QSO redshifts in SDSS
483: to probe any intrinsic periodicity without addressing selection
484: bias. To further investigate whether such a periodicity around
485: $\Delta z = 0.67$ in SDSS QSOs is spuriously produced by various
486: incompleteness as function of redshift, we select a
487: high-completeness sub-sample of 23,109 QSOs with $m_i<19$ and
488: $z<2$ in SDSS DR3, and three sub-samples containing QSOs in
489: low-completeness regions. As shown in Fig.~11, no periodicity is
490: found in the high-completeness sample where the power spectrum is
491: consistent with a continuously ascending curve due to the low
492: frequency component of the redshift distribution, whereas in
493: different low-completeness samples, strong periodicity always
494: appears, but with different peak locations (0.88 in (b), 0.67 in
495: (c) and 0.74 in (d)). This should be a strong indicator that the
496: peaks in low-completeness samples are caused by different
497: selection effects in different samples. In sum, there is no
498: evidence for intrinsic periodicity in redshifts of QSOs.
499:
500:
501: \section{Discussion and Conclusion}
502:
503: However, one might ask whether it is because we have some paired QSOs with wrong parent
504: galaxies so that not only the effective redshifts of QSOs show no periodicity, but also
505: high-z QSOs and nearby active galaxies show no strong connection. The wrong-pairing
506: indeed could happen that when there is more than one galaxy within the $D_{separation}$
507: projected distance limit of the QSO and we take the closest galaxy in projected
508: distance to make up the pair. In the following we quantitatively examine this
509: possibility and its effect.
510:
511: For the pair making process in section 2, the $D_{separation}$ is
512: 200 kpc, which is less than the average projection distance
513: between QSOs and galaxies, and the percentage of cases in which
514: there are two or more galaxies within the projected distance is
515: 27$\%$ for QSO-nearby galaxies, 19$\%$ for QSO-active nearby
516: galaxies, 27$\%$ for bright QSO-nearby galaxies and 18$\%$ for
517: bright QSO-active nearby galaxies, respectively. This means that
518: for a majority of paired QSOs ($>73 \%$), there is only one galaxy
519: within the given projected distance and would not be paired
520: incorrectly, hence the claimed periodicity should have been
521: detected in our larger samples if they did exist.
522:
523: For the pair making process in section 3, the largest $D_{separation}$ is 1 Mpc which
524: is larger than the average projection distance between QSOs and galaxies ($\sim$400
525: kpc), therefore wrong-pairing may occur more frequently here. It will be even worse if
526: the typical ejection distance is larger than the mean projection separation of QSOs and
527: active galaxies. So would such wrong pairs result in the good agreement between ejected
528: model and randomly generated pairs? We answer this question by making the following
529: test. Suppose that all QSOs are ejected by randomly selected active galaxies with a
530: given ejection velocity (11,000 km s$^{-1}$, 40,000 km s$^{-1}$ and 80,000 km
531: s$^{-1}$), and with a uniformly distributed age of $0 \sim
532: 10^8$ yr which is given by Bell (2004) as typical values, we get
533: 200 sets of false pairs. As shown in Fig.~7-8 and Table 1, the
534: distribution of such simulated ejection QSO-galaxy pairs are
535: totally different from random distribution. We therefore conclude
536: that the random-like distribution of QSO-active galaxy pairs could
537: not be produced by the ejection model.
538:
539: Another question is that whether we miss periodicities of QSOs by
540: setting a lower limit of $z=0.01$ for galaxies and no constraint
541: in QSO-galaxy magnitude relation. Though the lower limit of
542: $z=0.01$, the same as in Hawkins et al. (2002), is chosen to have
543: confidence in the derived angular distance, as well as avoid large
544: projection effect of very nearby galaxies, we re-analyze SDSS DR1
545: QSOs and galaxies again with no redshift limits on galaxies and
546: set magnitude constraint that all paired QSOs should be at least 5
547: or 3 magnitudes fainter than the paired galaxy. As shown in
548: Fig.~11 and Fig.~12, similar to our results in section 2.2, there
549: is no evidence for a periodicity at the predicted frequency in
550: log$(1+z)$, or at any other frequency.
551:
552: In summary, using samples from SDSS and 2QZ, we demonstrate that
553: not only there is no periodicity at the predicted frequency in
554: $\log(1+z)$ and $z$, or at any other frequency, but also there is
555: no strong connection between foreground active galaxies and high
556: redshift QSOs. These results are against the hypothesis that QSOs
557: are ejected from active galaxies or have periodic intrinsic
558: non-cosmological redshifts.
559:
560:
561: \acknowledgments
562:
563: We thank the anonymous referee and Dr. Bell for valuable
564: suggestions that have significantly improved this paper. This
565: study is supported in part by the Special Funds for Major State
566: Basic Research Projects and by the National Natural Science
567: Foundation and the Ministry of Education of China. SNZ also
568: acknowledges NASA for partial financial support through several
569: research grants.
570:
571: %% To help institutions obtain information on the effectiveness of their
572: %% telescopes, the AAS Journals has created a group of keywords for telescope
573: %% facilities. A common set of keywords will make these types of searches
574: %% significantly easier and more accurate. In addition, they will also be
575: %% useful in linking papers together which utilize the same telescopes
576: %% within the framework of the National Virtual Observatory.
577: %% See the AASTeX Web site at http://www.journals.uchicago.edu/AAS/AASTeX
578: %% for information on obtaining the facility keywords.
579:
580: %% After the acknowledgments section, use the following syntax and the
581: %% \facility{} macro to list the keywords of facilities used in the research
582: %% for the paper. Each keyword will be checked against the master list during
583: %% copy editing. Individual instruments can be provided in parentheses,
584: %% after the keyword, but they will not be verified.
585:
586: % Facilities: \facility{Nickel}, \facility{HST(STIS)}, \facility{CXO(ASIS)}.
587:
588:
589: \begin{thebibliography}{}
590: \bibitem[Abazajian et al.(2003)]{aba03} Abazajian, k., et al. 2003, \aj, 126, 2081
591: \bibitem[Arp et al.(1990)]{arp90} Arp, H., Bi, H. G., Chu, Y., \& Zhu, X. 1990,
592: \aap, 239, 33
593: \bibitem[Arp et al.(2005)]{arp05} Arp, H., Roscoe, D., \& Fulton, C. 2005, astro-ph/0501090
594: \bibitem[Basu(2005)]{basu05} Basu, D. 2005, \apjl, 618, L71
595: \bibitem[Bell \& Comeau(2003)]{bell03} Bell, M. B. \& Comeau, S. P.
596: 2003, submitted to ApJ (astro-ph/0305060)
597: \bibitem[Bell(2004a)]{bell04a} Bell, M. B. 2004a, submitted to ApJ (astro-ph/0403089)
598: \bibitem[Bell(2004b)]{bell04b} Bell, M. B. 2004b, \apj, 616, 738
599: \bibitem[Blanton et al.(2003)]{bla03} Blanton, M. R., et al. 2003,
600: AJ, 125, 2276
601: \bibitem[Blanton et al.(2005)]{bla04} Blanton, M. R., et al. 2005, accepted by AJ (astro-ph/0410166)
602: \bibitem[Burbidge \& Burbidge(1967)]{bb67} Burbidge, G. R. \& Burbidge, E. M. 1967, \apjl, 148,
603: L107
604: \bibitem[Burbidge \& Napier(2001)]{bn01} Burbidge, G. R. \& Napier, W. M. 2001, \apj, 121,
605: 21
606: \bibitem[Burbidge(2003)]{bur03} Burbidge, G. R. 2003, \apj, 585,
607: 112
608: \bibitem[Chu et al.(1998)]{chu98} Chu, Y., Wei, J., Hu, J., Zhu, X., \& Arp, H. 1998, \apj, 500,
609: 596
610: \bibitem[Croom et al.(2004)]{cro04} Croom, S. M., et al. 2004,
611: \mnras, 349, 1397
612: \bibitem[Efron(1979)]{efr79} Efron, B. 1979, Ann. Stat., 7, 1
613: \bibitem[Gaztanaga(2003)]{gaz03} Gaztanaga, E. 2003, \apj, 589, 82
614: \bibitem[Hawkins et al.(2002)]{haw03} Hawkins, E., Maddox, S., \& Merrifield, M. 2002, \mnras, 336,
615: L13
616: \bibitem[Hewitt \& Burbidge(1993)]{hb93} Hewitt, A. \& Burbidge, G. 1993, \apjs, 87,
617: 451
618: \bibitem[Karlsson(1977)]{kar77} Karlsson, K. G. 1977, \aap, 58,
619: 237
620: \bibitem[Karlsson(1990)]{kar90} Karlsson, K. G. 1990, \aap, 239,
621: 50
622: \bibitem[Napier \& Burbidge(2003)]{bn03} Napier, W. M. \& Burbidge, G. 2003,
623: \mnras, 342, 601
624: \bibitem[Richards et al.(2002)]{ric02} Richards, G. T., et al. 2002, \aj, 123,
625: 2945
626: \bibitem[Schneider et al.(2003)]{sch03} Schneider, D. P., et al. 2003, \aj, 126,
627: 2579
628: \bibitem[Schneider et al.(2005)]{sch05} Schneider, D. P., et al. 2005, accepted
629: by AJ (astro-ph/0503679)
630: \bibitem[Strauss et al.(2002)]{str02} Strauss, M. A., et al. 2002, \aj, 124,
631: 1810
632:
633: \end{thebibliography}
634:
635: \clearpage
636:
637:
638: \begin{figure}
639: \epsscale{.80} \plotone{f1.eps} \caption{Combined QSO data
640: set from Karlsson (1990) and Burbidge $\&$ Napier (2001,
641: 2003). Upper panel: histogram of redshifts of these QSOs. Peaks predicted by Karlsson's
642: formula are indicated by dotted vertical lines. Lower panel:
643: unwindowed power spectra of $100\log(1+z)$ (solid line) with 1$\sigma$ (dash lines)
644: and 2$\sigma$ (dash-dotted lines) deviations given from 1000
645: bootstrap simulations.\label{fig1}}
646: \end{figure}
647: \clearpage
648:
649: \begin{figure}
650: \epsscale{.80} \plotone{f2.eps} \caption{Effective redshifts of
651: 4572 QSOs paired with nearby galaxies in a projection
652: distance less than 200 kpc. Upper panel: histogram of redshifts of these QSOs. Peaks predicted by Karlsson's
653: formula are indicated by dotted vertical lines. Lower panel:
654: unwindowed power spectra of $100\log(1+z_{eff})$ (solid line) with 1$\sigma$ (dash lines)
655: and 2$\sigma$ (dash-dotted lines) deviations given from 1000
656: bootstrap simulations.\label{fig2}}
657: \end{figure}
658: \clearpage
659:
660: \begin{figure}
661: \epsscale{.80} \plotone{f3.eps} \caption{Effective redshifts of
662: 3216 QSOs paired with nearby active galaxies (starforming or
663: starburst galaxies) in a projection
664: distance less than 200 kpc. Upper panel: histogram of redshifts of these QSOs. Peaks predicted by Karlsson's
665: formula are indicated by dotted vertical lines. Lower panel:
666: unwindowed power spectra of $100\log(1+z_{eff})$ (solid line) with 1$\sigma$ (dash lines)
667: and 2$\sigma$ (dash-dotted lines) deviations given from 1000
668: bootstrap simulations.\label{fig3}}
669: \end{figure}
670: \clearpage
671:
672: \begin{figure}
673: \epsscale{.80} \plotone{f4.eps} \caption{Effective redshifts of
674: 958 bright QSOs ($i<18.5$) paired with nearby galaxies in a
675: projection
676: distance less than 200 kpc. Upper panel: histogram of redshifts of these QSOs. Peaks predicted by Karlsson's
677: formula are indicated by dotted vertical lines. Lower panel:
678: unwindowed power spectra of $100\log(1+z_{eff})$ (solid line) with 11$\sigma$ (dash lines)
679: and 2$\sigma$ (dash-dotted lines) deviations given from 1000
680: bootstrap simulations.\label{fig4}}
681: \end{figure}
682: \clearpage
683:
684: \begin{figure}
685: \epsscale{.80} \plotone{f5.eps} \caption{Effective redshifts of
686: 671 bright QSOs ($i<18.5$) paired with nearby active galaxies
687: (starforming or starburst galaxies) in a projection
688: distance less than 200 kpc. Upper panel: histogram of redshifts of these QSOs. Peaks predicted by Karlsson's
689: formula are indicated by dotted vertical lines. Lower panel:
690: unwindowed power spectra of $100\log(1+z_{eff})$ (solid line) with 1$\sigma$ (dash lines)
691: and 2$\sigma$ (dash-dotted lines) deviations given from 1000
692: bootstrap simulations.\label{fig5}}
693: \end{figure}
694: \clearpage
695:
696: \begin{figure}
697: \plottwo{f6a.eps}{f6b.eps} \caption{Apparent magnitude vs measured
698: redshift plot
699: for QSOs in the SDSS DR1 catalog. In the left one, the whole region is divided into boxes $\bigtriangleup z \times
700: \bigtriangleup B =0.075 \times 0.3$ in the redshift/apparent
701: magnitude plane, while $\bigtriangleup z \times \bigtriangleup B
702: =0.05 \times 0.2$ in the right one. The contours represent
703: QSO density in steps of 180, 150, 120, 90, 60 and 30 in the left one,
704: while in the right one they represent 75, 60, 45, 30 and 15, from the
705: innermost (high density) to outermost (low density). The predicted
706: Karlsson peaks at z=0.30, 0.60, 0.96, 1.41, 1.96 and 2.64 are
707: shown by vertical lines.\label{fig6}}
708: \end{figure}
709: \clearpage
710:
711: \begin{figure}
712: \epsscale{.80} \plotone{f7.eps} \caption{Distribution of
713: projected distance between
714: 2604 high-z QSOs ($2.4<z<4.8$) and their paired nearby active galaxies
715: in NYU-VAGC. The circles represent `true' pairs, i.e., pairs found in the data, but not necessarily physical pairs.
716: The solid line with error bars
717: is the average of 200 simulations of QSOs and random distributed
718: galaxies.
719: Averages of 200 simulations of randomly ejected QSOs and active galaxies are also presented,
720: where QSOs are produced by ejection from randomly chosen galaxies
721: with a uniformly distributed age in $0-10^8$ yr and three different velocities:
722: 11,000 km/s for dash line with error bars, 40,000 km/s for dotted line with points and
723: 80,000 km/s for dash-dotted line with points. \label{fig7}}
724: \end{figure}
725: \clearpage
726:
727: \begin{figure}
728: \epsscale{.80} \plotone{f8.eps} \caption{Distribution of
729: 2604 foreground active galaxies which have
730: at least one high-z QSO behind within a projected distance 1 Mpc in NYU-VAGC.
731: The circles represent `true' pairs. Others
732: are the same as in Fig.7.\label{fig8}}
733: \end{figure}
734: \clearpage
735:
736:
737: \begin{figure}
738: \epsscale{.80} \plotone{f9.eps} \caption{Redshifts of 46,420 QSOs
739: in SDSS DR3. Upper panel: histogram of redshifts of these QSOs.
740: Lower panel: power spectra of z (solid line) weighted using a Hann
741: function with 1$\sigma$ (dash lines) and 2$\sigma$ (dash-dotted
742: lines) deviations given from 1000 bootstrap simulations.
743: \label{fig9}}
744: \end{figure}
745: \clearpage
746:
747:
748: \begin{figure}
749: \epsscale{.80} \plotone{f10.eps} \caption{Redshifts of 22,497 QSOs
750: with the highest quality flag in 2dF. Upper panel: histogram of
751: redshifts of these QSOs. Lower panel: power spectra of z (solid
752: line) weighted using a Hann function with 1$\sigma$ (dash lines)
753: and 2$\sigma$ (dash-dotted lines) deviations given from 1000
754: bootstrap simulations. \label{fig10}}
755: \end{figure}
756: \clearpage
757:
758:
759: \begin{figure}
760: \epsscale{.80} \plotone{f11.ps} \caption{Power spectra of redshift
761: of four sub-samples from SDSS DR3. (a) is for the
762: high-completeness sample containing 23,109 QSOs with $m_i<19$ and
763: $z<2$, and others are for samples containing QSOs in
764: low-completeness regions: (b) is for 15,696 QSOs with $m_i>19$ and
765: $z<2.4$, (c) is for 19,064 QSOs with $m_i>19$, and (d) is for
766: 9,763 QSOs with $z>2$. Power spectra of z (solid line) is weighted
767: using a Hann function with 1$\sigma$ (dash lines) and 2$\sigma$
768: (dash-dotted lines) deviations given from 1000 bootstrap
769: simulations. \label{fig11}}
770: \end{figure}
771: \clearpage
772:
773:
774:
775: \begin{figure}
776: \epsscale{.80} \plotone{f12.eps} \caption{Effective redshifts of
777: 82 QSOs paired with galaxies in a projection
778: distance less than 200 kpc, which are at least 5 magnitudes fainter than the paired galaxies in $i$-band.
779: Upper panel: histogram of redshifts of these QSOs. Peaks predicted by Karlsson's
780: formula are indicated by dotted vertical lines. Lower panel:
781: unwindowed power spectra of $100\log(1+z_{eff})$ (solid line) with 1$\sigma$ (dash lines)
782: and 2$\sigma$ (dash-dotted lines) deviations given from 1000
783: bootstrap simulations. \label{fig12}}
784: \end{figure}
785: \clearpage
786:
787: \begin{figure}
788: \epsscale{.80} \plotone{f13.eps} \caption{Effective redshifts of
789: 1459 QSOs paired with galaxies in a projection
790: distance less than 200 kpc, which are at least 3 magnitudes fainter than the paired galaxies in $i$-band.
791: Upper panel: histogram of redshifts of these QSOs. Peaks predicted by Karlsson's
792: formula are indicated by dotted vertical lines. Lower panel:
793: unwindowed power spectra of $100\log(1+z_{eff})$ (solid line) with 1$\sigma$ (dash lines)
794: and 2$\sigma$ (dash-dotted lines) deviations given from 1000
795: bootstrap simulations.
796: \label{fig13}}
797: \end{figure}
798: \clearpage
799:
800:
801:
802: \end{document}
803:
804: %%
805: %% End of file `sample.tex'.
806: