1: %\documentclass[useAMS,usenatbib]{mn2e}
2: \documentclass[usenatbib]{mn2e}
3: \input epsf
4:
5: \title[Periodicity Search of Possible X-ray Counterparts]{Periodicity Search
6: of Possible X-ray Counterparts to Radio-quiet Gamma-ray Pulsar Candidates}
7: \author[L. C.-C. Lin and H.-K. Chang]{Lupin Chun-Che Lin and Hsiang-Kuang Chang\thanks{E-mail:
8: hkchang@phys.nthu.edu.tw}\\
9: Department of Physics and Institute of Astronomy, National Tsing Hua university, Hsinchu 30013, Taiwan}
10:
11: \begin{document}
12:
13: \date{November 2007; February 2008}
14: \pagerange{\pageref{firstpage}--\pageref{lastpage}}
15: \pubyear{2008}
16: \maketitle
17:
18: \label{firstpage}
19:
20: \begin{abstract}
21: Periodicity search in gamma-ray data is usually difficult
22: because of the small number of detected photons.
23: A periodicity in the timing signal at other energy bands from
24: the counterpart to the gamma-ray source may help to
25: establish the periodicity in the gamma-ray emission
26: and strengthen the identification of the source in different energy bands.
27: It may, however, still be difficult to find the period directly
28: from X-ray data because of limited exposure.
29: We developed a procedure, by cross-checking two X-ray data sets,
30: to find candidate periods for X-ray sources which are possible
31: counterparts to gamma-ray pulsar candidates.
32: Here we report the results of this method
33: obtained with all the currently available X-ray data of 8 X-ray sources.
34: Some tempting periodicity features were found.
35: Those candidate periods can serve as the target periods
36: for future search when new data become available so that
37: a blind search with a huge number of trials can be avoided.
38: \end{abstract}
39:
40: \begin{keywords}
41: pulsars: general -- stars: neutron -- gamma-rays : observations -- X-rays: general
42: \end{keywords}
43:
44: %SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS
45: \section{Introduction}
46:
47: The electrodynamics in the magnetosphere of a neutron star
48: has not yet been well understood, although fascinating phenomena
49: as revealed in emissions across the whole electromagetic spectrum
50: from neutron stars have been observed for forty years.
51: Among the 2000 or so radio pulsars cataloged to date,
52: only 7 are detected at photon energies higher than 100 MeV
53: and 3 more are detected at a lower confidence level \citep{thompson04}.
54: These 7 include Geminga, which does not emit radio waves or emits only
55: at a very low level \citep{kuzmin99}.
56: The effort to understand gamma-ray emissions from pulsars has resulted in
57: two types of models: the slot-gap models \citep{muslimov04} and the outer-gap models
58: (e.g. \citet{cheng00,hirotani07,takata07}).
59: These models, although different in the assumed emission site and mechanisms,
60: have been successful in many aspects in explaining the pulse profiles and spectra.
61: However, they do give different estimates of the fraction of radio-quiet gamma-ray pulsars,
62: like Geminga, among all the gamma-ray pulsars \citep{harding07}.
63: A gamma-ray pulsar may be radio-quiet if its gamma-ray beam is wider than its radio beam
64: or these beams are in different directions. It is also possible that it is
65: intrinsically weak in radio emission because of unfavored spin periods and magnetic field strength.
66: Although a consistent model for radio emission from pulsars is still lacking, several empirical
67: models are available.
68: Given a radio emission model (e.g.
69: \citet{johnston06}), different models for
70: gamma-ray pulsars may be discriminated by comparing their estimates of
71: the radio-quiet gamma-ray pulsar fraction with observations, in addition to comparing
72: their prediction of high-energy emission polarization characteristics (e.g.\
73: \citet{takata07b}).
74:
75: The number of currently known gamma-ray pulsars is only 7, which is obviously too small to
76: give a meaningful fraction of the radio-quiet ones.
77: With the launch of GLAST and AGILE, the number of gamma-ray pulsars is expected to increase significantly.
78: However, radio-quiet gamma-ray pulsars may still be difficult to find, mainly because of the limited amount
79: of gamma-ray photons.
80: Just like the case of Geminga, only after its period was discovered in the ROSAT data
81: \citep{halpern92}, its pulsation in gamma-rays was then found.
82: Very often it is also difficult to find the period in X-ray data, particularly when the data is sparse and a blind search
83: over a large range of periods is performed.
84: Efforts to find new radio-quiet gamma-ray pulsars from unidentified EGRET sources in a multiwavelength approach
85: have not yet yielded definite results (e.g.\
86: \citet{reimer01,halpern02}).
87: Among the unidentified EGRET sources, many are at low galactic latitude and may be gamma-ray pulsars.
88: Continuing this effort, we conducted a periodicity search for possible X-ray counterparts to some bright unidentified
89: EGRET sources, using all available archival data with the method of cross-checking two data sets to look for candidate periods.
90: The purpose of this cross-checking procedure is to find relatively
91: significant signatures of possible pulsations occurring in two independent data sets.
92: Candidate periods and the corresponding period time derivatives
93: thus found are reported and can serve as the target
94: periods to verify in future X-ray or gamma-ray data.
95:
96: %SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS
97: \section{Source selection and data reducton}
98:
99: As an effort to understand the nature of unidentified EGRET sources,
100: \citet{roberts01} presented
101: a catalog of 2-10 keV
102: ASCA/GIS images of
103: fields containing 30 bright GeV sources.
104: These GeV sources were selected from the brightest ones in
105: the catalog compiled by
106: \citet{lamb97} using only
107: photons with energies above 1 GeV from the EGRET public
108: archives that incorporate the first 4.5 yr of CGRO/EGRET observations.
109: Among these GeV sources, some are identified pulsars and quasars,
110: some are possible counterparts to previously known objects, and
111: the others are not yet identified to sources in other energy bands.
112: Possible X-ray counterparts of these unidentified GeV sources were
113: proposed in
114: \citet{roberts01}.
115: To search for new radio-quiet gamma-ray pulsars,
116: we look for periodicity of possible X-ray counterparts to those unidentified
117: GeV sources for which the possibility of being gamma-ray pulsars has not yet been ruled out.
118: Furthermore, as explained in the next section, two data sets adequate for timing analysis
119: are needed for the cross-checking method
120: to find candidate periods.
121: 6 GeV sources among the 30 in the catalog of
122: \citet{roberts01} were thus selected.
123: One of the 6 sources (GeV J1417-6100)
124: has four potential X-ray counterparts.
125: Another one of the 6 sources (GeV J0008+7304)
126: has been investigated with the same procedure described in this paper.
127: The periodicity search result of its possible X-ray counterpart,
128: RX J0007.0+7302, was reported by
129: \citet{lin05} and will not be repeated here.
130: The 8 X-ray sources and their data used in this study are listed in Table~\ref{data}.
131: Detailed descriptions of each GeV sources can be found in
132: \citet{roberts01}.
133: %TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT
134: \begin{table*}
135: \caption{List of data employed in this study.
136: Column 1 is the name of the X-ray source with its possible gamma-ray counterpart in the parentheses.
137: Column 2 is the instrument used to obtain that data.
138: Column 3 is the data epoch (in MJD), which is the middle point in the whole data time span.
139: Column 4 and 5 are the total exposure time and the total time span of the data used. Both are in units of ks.
140: Column 6 and 7 are the target position, photon numbers after extraction, the radius of the extraction circle, and
141: the extraction energy range (in keV) for the X-ray target under study.
142: Column 8 is the sequence number to label the data used in this work for later referencing.
143: }
144: \label{data}
145: \begin{tabular}{cccccccc}
146: \hline
147: \hline
148: (1) & (2) & (3) & (4) & (5) & (6) & (7) & (8) \\
149: \hline
150: AX J1418.2--6047 & ASCA/GIS & 50317.00727 & 25.6 & 106 & ($14^h18^m15^s.8,-60^{\circ}46'44''$) & 311; $90''$; 1-10 & 01 \\
151: (GeV J1417--6100) & ASCA/GIS & 51223.68497 & 22.2 & 173 & ($14^h18^m15^s.8,-60^{\circ}46'44''$) & 126; $90''$; 1-6 & 02 \\
152: & BeppoSAX/MECS & 52289.79012 & 30.8 & 51.0 & ($14^h18^m15^s.8,-60^{\circ}46'44''$) & 105; $120''$; 2-10 & 03 \\
153: \hline
154: AX J1418.6--6045 & ASCA/GIS & 50317.00727 & 25.6 & 106 & ($14^h18^m37^s.0,-60^{\circ}45'12''$) & 271; $90''$; 1-8 & 04 \\
155: (GeV J1417--6100) & ASCA/GIS & 51223.68497 & 22.2 & 173 & ($14^h18^m37^s.0,-60^{\circ}45'12''$) & 133; $90''$; 1-7 & 05\\
156: & BeppoSAX/MECS & 52289.79013 & 30.8 & 51.0 & ($14^h18^m37^s.0,-60^{\circ}45'12''$) & 148; $120''$; 2-10 & 06 \\
157: \hline
158: AX J1418.7--6058 & ASCA/GIS & 50317.00728 & 25.6 & 106 & ($14^h18^m38^s.6,-60^{\circ}57'49''$) & 314; $90''$; 1-8 & 07 \\
159: (GeV J1417--6100) & ASCA/GIS & 51223.68496 & 22.2 & 173 & ($14^h18^m38^s.6,-60^{\circ}57'49''$) & 551; $90''$; 1-7 & 08\\
160: & BeppoSAX/MECS & 52289.79050 & 30.8 & 51.0 & ($14^h18^m38^s.6,-60^{\circ}57'49''$) & 569; $210''$; 2-10 & 09 \\
161: & XMM/pn & 52708.65941 & 26.9 & 26.9 & ($14^h18^m42^s.8,-60^{\circ}58'03''.6$) & 632; $15''$; 0.2-12 & 10 \\
162: \hline
163: AX J1420.1--6049 & ASCA/GIS & 50317.00729 & 25.6 & 106 & ($14^h20^m07^s.8,-60^{\circ}48'56''$) & 214; $90''$; 1.5-8 & 11 \\
164: (GeV J1417--6100) & ASCA/GIS & 51223.68495 & 22.2 & 173 & ($14^h20^m07^s.8,-60^{\circ}48'56''$) & 252; $90''$; 1.5-8 & 12\\
165: & BeppoSAX/MECS & 52289.79017 & 30.8 & 51.0 & ($14^h20^m07^s.8,-60^{\circ}48'56''$) & 904; $210''$; 2-10 & 13 \\
166: & Chandra/ACIS & 52534.12055 & 31.2 & 31.2 & ($14^h20^m08^s.2,-60^{\circ}48'16''.9$) & 229; $2''$; 1-7 & 14 \\
167: \hline
168: AX J1809.8--2332 & ASCA/GIS & 50526.78845 & 41.5 & 180 & ($18^h09^m48^s.6,-23^{\circ}32'09''$) & 1133; $90''$; 0.8-7 & 15 \\
169: (GeV J1809--2327) & XMM/pn & 52539.07915 & 11.1 & 15.6 & ($18^h09^m50^s.0,-23^{\circ}32'24''$) & 542; $15''$; 0.2-12 & 16 \\
170: & XMM/pn & 53280.71956 & 49.1 & 69.3 & ($18^h09^m50^s.0,-23^{\circ}32'24''$) & 4070; $15''$; 0.2-12 & 17 \\
171: \hline
172: AX J1836.2+5928 & ROSAT/HRI & 50800.00499 & 40.1 & 443 & ($18^h36^m13^s.6,+59^{\circ}25'29''$) & 224; $30''$; 0.1-2.5 & 18 \\
173: (GeV J1835+5921) & Chandra/HRC & 53430.07235 & 45.2 & 45.2 & ($18^h36^m13^s.7,+59^{\circ}25'30''$) & 278; $1''$; 0.08-10 & 19 \\
174: & Chandra/HRC & 53438.80937 & 28.1 & 28.1 & ($18^h36^m13^s.7,+59^{\circ}25'30''$) & 181; $1''$; 0.08-10 & 20 \\
175: & Chandra/HRC & 53440.60818 & 45.2 & 45.2 & ($18^h36^m13^s.7,+59^{\circ}25'30''$) & 247; $1''$; 0.08-10 & 21 \\
176: \hline
177: AX J1837.5--0610 & ASCA/GIS & 50905.53135 & 19.7 & 37.0 & ($18^h37^m29^s.0,-06^{\circ}09'38''$) & 243; $90''$; 1-9 & 22 \\
178: (GeV J1837--0610) & ASCA/GIS & 51105.10478 & 17.9 & 40.0 & ($18^h37^m29^s.0,-06^{\circ}09'38''$) & 203; $90''$; 1-9 & 23 \\
179: & BeppoSAX/MECS & 51981.25968 & 40.2 & 94.5 & ($18^h37^m32^s.5,-06^{\circ}09'49''$) & 739; $210''$; 2-10 & 24 \\
180: & BeppoSAX/MECS & 52016.65630 & 42.7 & 105 & ($18^h37^m32^s.5,-06^{\circ}09'49''$) & 770; $210''$; 2-10 & 25 \\
181: \hline
182: AX J2021.1+3651 & ASCA/GIS & 51300.06891 & 21.8 & 69.8 & ($20^h21^m07^s.8,+36^{\circ}51'19''$) & 294; $90''$; 1-7 & 26 \\
183: (GeV J2020+3658) & Chandra/ACIS & 52682.60172 & 20.8 & 20.8 & ($20^h21^m05^s.5,+36^{\circ}51'04''.8$) & 575; $1''$; 0.2-5 & 27 \\
184: \hline
185: \end{tabular}
186: \end{table*}
187: %TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT
188:
189: GeV J1417--6100 is in the direction of the `Kookaburra' radio complex
190: \citep{roberts99}.
191: A 68-ms radio pulsar, PSR J1420--6048, was found at a position consistent with
192: AX J1420.1--6049
193: \citep{damico01}, which is in the northeastern wing of the Kookaburra.
194: The pulsation at the radio period in X-rays was only
195: marginally detected with ASCA data
196: \citep{roberts01b}.
197: Chandra images of AX J1420.1--6049 clearly show a compact source at the
198: position of PSR J1420--6048, but its pulsation in X-rays at the radio
199: period cannot be confirmed with the current Chandra data
200: \citep{ng05}.
201: The identification between PSR J1420--6048 and AX J1420.1--6049
202: is not yet conclusive.
203: We therefore include AX J1420.1--6049 in our periodicity search.
204: AX J1418.7--6058 is located in the southwestern wing of the Kookaburra
205: (the `Rabbit').
206: Two compact sources denoted as `R1' and `R2' in the Rabbit
207: were identified in Chandra and XMM
208: images
209: \citep{ng05}.
210: Although `R1' might be a background source,
211: we still used it as the target in the XMM data
212: for cross-checking periodicity signatures
213: with earlier ASCA and BeppoSAX data, since `R1' is much brighter than `R2' and
214: in those earlier data
215: photons from `R1' dominate.
216: More recently, two extended TeV sources (HESS J1420--607 and HESS J1418--609),
217: spatially coincident with the two wings of the Kookaburra,
218: were discovered
219: \citep{aharonian06}.
220: It is likely that 3EG J1420--6038 and GeV J1417--6100 are two separate sources
221: associated with HESS J1420--607 and HESS J1418--609 respectively
222: \citep{aharonian06}.
223: Another two X-ray sources, AX J1418.2--6047 and AX J1418.6--6045, denoted as
224: `Src2' and `Src3' in
225: \citet{roberts01}, are both located in the 95\% error contour of GeV J1417--6100
226: and in the 99\% contour of 3EG J1420--6038
227: \citep{aharonian06}.
228:
229: AX J1809.8--2332 is the brightest X-ray source within the 99\% error box
230: of GeV J1809--2327 (3EG J1809--2328).
231: Chandra images of this region reveal a point X-ray source connected to a
232: nonthermal X-ray/radio nebula
233: \citep{braje02}.
234: XMM images show a larger, broader X-ray emission region with a ridge
235: of emission along the symmetric axis of the radio nebula
236: \citep{roberts06}.
237: The X-ray point source is very likely a pulsar moving
238: at the tip of its associated pulsar wind nebula.
239: It is therefore very much desired to find its pulsation period.
240:
241: GeV J1835+5921 and its possible X-ray counterpart AX J1836.2+5928 have
242: long been considered as the next Geminga to find.
243: Strong evidence from Chandra and HST observations indicates
244: AX J1836.2+5928 (RX J1836.2+5925) is a neutron star
245: \citep{halpern02}.
246: Using the Chandra data listed in Table~\ref{data},
247: \citet{halpern07} reported
248: unsuccessful searches for pulsations from RX J1836.2+5925
249: in the period range from 1 ms to 10 s
250: and a pulsed-fraction upper limit
251: of 35\%.
252:
253: There are fewer observations of the GeV J1837--0610 region.
254: A 96-ms radio pulsar (PSR J1837--0604) was discovered and suggested to be
255: associated with GeV J1837--0610 (3EG J1837--0606)
256: \citep{damico01}.
257: However, the position of PSR J1837--0604 is at the rim of the 95\% confidence
258: contour of GeV J1837--0610 and is inconsistent with AX J1837.5--0610,
259: which is the single X-ray point source within the 95\% confidence contour of
260: GeV J1837--0610 in ASCA images.
261:
262: Chandra observations of AX J2021.1+3651
263: \citep{hessels04} indicate that
264: this X-ray source is a pulsar wind nebula and there is an X-ray
265: point source located at the center of this nebula, where
266: a 104-ms radio pulsar (PSR J2021+3651) was found
267: \citep{roberts02}.
268: The X-ray pulsation of this point source at the radio period was
269: only marginally found in the Chandra data with an $H$-test score of 11.7
270: \citep{hessels04}.
271:
272: All the data were processed with standard procedures as outlined below.
273: Photons for timing analysis were selected within a circle around the target
274: and within a certain energy range. The circle radii and energy-range boundaries
275: were so chosen that the target contribution was clearly detected.
276: The target positions, the radii of the circles and the energy ranges adopted in this work are listed
277: in Table~\ref{data}.
278: All the photon arrival times were then corrected to the solar system barycenter.
279: The corrected time list was then used for timing analysis to look for periodicity.
280:
281: %SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS
282: \section{The cross-checking method for periodicity search}
283:
284: Periodicity search in sparse data usually employs photon-counting
285: coherent-timing methods, that is, each photon arrival time is converted to
286: a phase value with a given period and its time derivatives. The distribution of these phase values
287: is then tested against the null hypothesis with a certain statistical method
288: to assess the probability that the observed distribution is consistent with
289: a statistically flat distribution.
290: In this work, we adopted the $H$-test
291: \citep{dejager89,dejager94} as the method to assess
292: the random probability.
293:
294: The difficulty to find the period in sparse data, in addition to the limited
295: number of observations,
296: is that a blind search with a huge number of trial periods (and period time derivatives)
297: is unavoidable if an a priori period or period range is not known.
298: When the total number of independent trials is taken into account, one seldom obtains
299: a significantly low random probability.
300: A possible way to improve upon this is to cross-check two different data to look for
301: possible support for the reality of a tempting candidate period found in one data set.
302:
303: For each of the X-ray sources discussed in the previous section
304: we first performed a blind search in every available data of that source.
305: Depending on the time resolution of the data
306: different search ranges of trial periods were chosen.
307: For all the ASCA/GIS data, the searching range is from
308: 0.1 s to 100 s.
309: For all the others, it is from 0.01 s to 100 s.
310: Since each data set spans a relatively short time,
311: we constrained ourselves at this stage to consider trial periods only
312: without their time derivatives.
313: This makes the search a one-dimensional one.
314: We chose the searching step to be a small fraction (typically 1/5) of the corresponding independent Fourier spacing,
315: which is $P^2/T$ in the period domain ($1/T$ in frequency),
316: where $P$ is the trial period and $T$ is the total time span of the data.
317: A blind search usually gives several tempting features.
318: One example is shown in Fig.~\ref{blind}.
319: %FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
320: \begin{figure}
321: \epsfxsize=8.4cm
322: \epsffile{fig1.eps}
323: \caption{Examples of $H$-test blind search results.
324: Shown here are those of AX J1420.1--6049 with data sets 11 and 12 (see Table~\ref{data}).
325: The number of independent trial frequencies is about $10^6$ for each case.
326: Most of the trials yield a very low $H$-value.
327: Only $H$-values larger than 20 are plotted. }
328: \label{blind}
329: \end{figure}
330: %FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
331: For readers' convenience, the random probability for a single trial to have an $H$-value larger
332: than a certain number is plotted in
333: Fig.~\ref{htest}.
334: %FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
335: \begin{figure}
336: \epsfxsize=8.4cm
337: \epsffile{fig2.eps}
338: \caption{The $H$-test probability distribution \citep{dejager89}.
339: The plotted curve is the null hypothesis random probability for a single trial to have an $H$-value larger
340: than $H_0$.}
341: \label{htest}
342: \end{figure}
343: %FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
344:
345: From the result of a blind search in a certain data set,
346: a few trial periods with high $H$-values were picked out
347: for cross-checking. The number of such trial periods to pick out is obviously arbitrary.
348: We picked out the top five in this work.
349: Each of these five usually has a significantly low random probability if the total
350: number of independent trials in that blind search is not taken into account.
351: As mentioned above, a blind search has a huge number of independent trials,
352: typically more than $10^6$. Therefore, for each of them, we look for possible associated
353: features in other data sets to assess its reality.
354: To do this, we further set two criteria.
355: At first we look for associated features of the $H$-test results in another data set
356: only within the range that gives
357: a corresponding characteristic age ($\frac{P}{2\dot{P}}$)
358: larger than 1000 years.
359: A neutron star younger than that age probably would have a record in human history.
360: In determining that range, the uncertainty of the period is estimated based on the method described in
361: \citet{leahy87}.
362: Secondly, the most prominent features in that range are identified and the resultant random probability
363: is calculated, which is the product of the random probability of the feature in the first data set and
364: that of the feature in the second data set. When calculating the random probability of the feature in the
365: second data set, only independent trials within the range specified by the first criterion are considered.
366: Oversampling was not taken into account in our computation of the random probability.
367: We keep only those cases whose resultant random probability is less than 0.01.
368: The periods and the corresponding period time derivatives for those cases are listed in Table~\ref{pcan}.
369: The probability threshold of 0.01 is arbitrary. Of course a higher one gives more candidate periods,
370: but with less significance.
371: Two examples of such an association are shown in Fig.~\ref{asso1} and \ref{asso2}.
372: %FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
373: \begin{figure}
374: \epsfxsize=8.4cm
375: \epsffile{fig3.eps}
376: \caption{The identification of the candidate period at 159 ms for AX J1418.6--6045 using data sets 04 and 05.
377: The trial period range shown in the lower panel of this figure
378: (from data set 05) is roughly the range specified by the criterion that the
379: corresponding characteristic age is larger than 1000 years
380: if a possible period is to be associated with the one in the upper panel.
381: The most significant feature in that range is at 0.15937378 sec, and its
382: association with the one at 0.15935800 sec in the upper panel
383: yields a resultant random probability of $5.3\times 10^{-3}$.
384: We list all such cases with a random probability lower than 0.01 as
385: a candidate period in Table~\ref{pcan}. }
386: \label{asso1}
387: \end{figure}
388: %FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
389: %FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
390: \begin{figure}
391: \epsfxsize=8.4cm
392: \epsffile{fig4.eps}
393: \caption{The identification of the candidate period at 145 ms for AX J1420.1--6049 using data sets 11 and 12.
394: The same as in Figure~\ref{asso1},
395: the trial period range shown in this figure is roughly the range specified by the criterion that the
396: characteristic age is larger than 1000 years. This figure is in fact a zoom-in view of Figure~\ref{blind} around
397: 6.88564 Hz but plotted in the period domain.
398: This example, with a resultant random probability of $4.1\times 10^{-5}$, is
399: the most significant signature found in this study.}
400: \label{asso2}
401: \end{figure}
402: %FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
403:
404: %TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT
405: \begin{table*}
406: \caption{Candidate periods.
407: Column 1 is the candidate period (in seconds) obtained from cross-checking the two data sets listed in column 7.
408: The number in the parentheses is the 1-$\sigma$ uncertainty to the last digit,
409: estimated with the method described in \citet{leahy87}.
410: Those followed by a triangle $(\triangle)$
411: are candidate periods favoured by examing the consistency with
412: properties of currently known gamma-ray pulsars, as discussed in Section 4.
413: The epoch (MJD) of this period is listed in column 5, which is the epoch of the data set containing the stronger signature for pulsation
414: in the two data sets.
415: Column 2, 3 and 4 are the corresponding period time derivative (in units of $10^{-13}$ s s$^{-1}$),
416: characteristic age (in years) and spin-down power (in units of $10^{36}$ erg s$^{-1}$).
417: Column 6 is the resultant random probability as explained in the text.
418: }
419: \begin{tabular}{lcccccc}
420: \hline
421: \hline
422: (1) & (2) & (3) & (4) & (5) & (6) & (7) \\
423: \hline
424: AX J1418.2--6047 & & & & & & \\
425: \hline
426: 0.2029690(1) $\triangle$ & 2.00(4) & 16000 & 9.4(2) & 50317.00727 & 4.7$\times 10^{-3}$ & 01, 03 \\
427: 0.3097181(5) & 47.3(1) & 1000 & 6.29(1) & 51223.68497 & 2.7$\times 10^{-3}$ & 01, 02 \\
428: 0.4289840(5) & 33.6(1) & 2000 & 1.682(6) & 51223.68497 & 5.3$\times 10^{-3}$ & 01, 02\\
429: \hline
430: AX J1418.6--6045 & & & & & & \\
431: \hline
432: 0.12180376(5) & 6.65(2) & 2900 & 14.51(4) & 51223.68497 & 9.1$\times 10^{-3}$ & 05, 06 \\
433: 0.15935800(7) $\triangle$ & 2.01(1) & 12500 & 1.96(1) & 50317.00727 & 5.3$\times 10^{-3}$ & 04, 05 \\
434: 0.3311744(5) & 39.39(9) & 1300 & 4.275(9) & 50317.00727 & 2.7$\times 10^{-3}$ & 04, 05 \\
435: 0.3311744(5) & 31.04(6) & 1700 & 3.366(7) & 50317.00727 & 3.2$\times 10^{-4}$ & 04, 06 \\
436: \hline
437: AX J1418.7--6058 & & & & & & \\
438: \hline
439: 0.10139550(2) & 0.080(4) & 200000 & 0.30(2) & 50317.00728 & 9.6$\times 10^{-3}$ & 07, 09 \\
440: 0.10139550(2) & 7.757(5) & 2100 & 29.32(2) & 50317.00728 & 3.6$\times 10^{-3}$ & 07, 09 \\
441: 0.10171064(5) & 9.337(6) & 1700 & 34.95(2) & 50317.00728 & $< 3.9\times 10^{-3}$ & 07, 09 \\
442: 0.10969281(4) & 14.45(1) & 1200 & 43.03(3) & 50317.00728 & 4.2$\times 10^{-3}$ & 07, 10 \\
443: 0.1443257(1) & 18.35(2) & 1250 & 24.13(2) & 51223.68496 & 1.9$\times 10^{-4}$ & 07, 08 \\
444: 0.3274125(3) & 31.29(8) & 1650 & 3.515(9) & 51223.68496 & 9.5$\times 10^{-3}$ & 08, 09 \\
445: \hline
446: AX J1420.1--6049 & & & & & & \\
447: \hline
448: 0.014799899(6) & 0.392(3) & 6000 & 478(4) & 52534.12055 & $< 6.2\times 10^{-3}$ & 13, 14 \\
449: 0.019576788(3) & 2.616(3) & 1200 & 1376(1) & 52289.79017 & 7.2$\times 10^{-3}$ & 13, 14 \\
450: 0.033549387(8) & 0.133(7) & 40000 & 13.9(8) & 52289.79017 & 1.2$\times 10^{-3}$ & 13, 14 \\
451: 0.14526767(4) & 4.84(2) & 4750 & 6.24(2) & 51223.68495 & 4.1$\times 10^{-5}$ & 11, 12 \\
452: 0.14526767(4) $\triangle$ & 1.35(3) & 17000 & 1.74(3) & 51223.68495 & 5.0$\times 10^{-3}$ & 12, 13 \\
453: 0.1881326(4) & 10.1(3) & 2900 & 6.0(2) & 52534.12055 & 3.8$\times 10^{-3}$ & 13, 14\\
454: 0.2540350(5) & 13.17(7) & 3100 & 3.17(2) & 50317.00729 & 5.0$\times 10^{-3}$ & 11, 12 \\
455: \hline
456: AX J1809.8--2332 & & & & & & \\
457: \hline
458: 0.13247455(3) & 16.439(5) & 1300 & 27.792(8) & 50526.78845 & $< 1.9\times 10^{-2}$ & 15, 17 \\
459: 0.2649491(2) & 27.4(1) & 1500 & 5.80(2) & 50526.78845 & 7.4$\times 10^{-4}$ & 15, 16 \\
460: \hline
461: AX J1836.2+5928 & & & & & & \\
462: \hline
463: 0.022425706(5) & 0.77(9) & 5000 & 2.7(3) $\times 10^{2}$ & 53440.60818 & 6.5$\times 10^{-3}$ & 19, 21 \\
464: 0.03350910(2) & 2(2) & 2000 & 2(2) $\times 10^{2}$ & 53438.80937 & 5.3$\times 10^{-3}$ & 20, 21 \\
465: 0.034677614(1) & 0.0201(4) & 270000 & 1.90(4) & 50800.00499 & 5.5$\times 10^{-3}$ & 18, 19 \\
466: 0.1679234(4) $\triangle$ & 0.16(2) & 170000 & 0.13(1) & 53440.60818 & 9.1$\times 10^{-3}$ & 18, 21 \\
467: \hline
468: AX J1837.5--0610 & & & & & & \\
469: \hline
470: 0.1736610(3) & 18.6(5) & 1500 & 14.0(3) & 51105.10478 & 2.8$\times 10^{-3}$ & 22, 23 \\
471: \hline
472: AX J2021.1+3651 & & & & & & \\
473: \hline
474: none & & & & & & \\
475: \hline
476: \end{tabular}
477: \label{pcan}
478: \end{table*}
479: %TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT
480:
481: In data set 07, the candidate period of AX J1418.7--6058 at 0.10171064 s has an $H$-value of 56.7, which has a
482: very low random probability for a single trial.
483: With our procedure, there will be candidate periods satisfying our criteria in all the data sets 08, 09, and 10.
484: Here we only report the most significant one.
485: Its random probability listed in Table \ref{pcan} was derived using an
486: $H$-value equal to 50 instead of 56.7, since in \citet{dejager89}
487: the signle trial random probability was provided only up to $H$ equal to 50.
488: Two similar cases are the candidate period at 0.014799899 s
489: for AX J1420.1--6049, which has an $H$-value of 53.8 in data set 14,
490: and that at 0.13247455 s for AX J1809.8--2332,
491: which has an $H$-value of 59.5 in data set 15.
492: Although in our procedure only the most prominent feature
493: in the cross-checking window is picked out for association, for
494: the candidate period of AX J1418.7--6058 at 0.10139550 sec in data set 07,
495: the association with the period of the second highest
496: $H$-value in the cross-checking window of data set 09
497: is also reported here, since it is very close to the candidate period
498: in data set 07 and its $H$-value is not too different from the highest.
499: That association is shown in Figure~\ref{asso3}.
500: %FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
501: \begin{figure}
502: \epsfxsize=8.4cm
503: \epsffile{fig5.eps}
504: \caption{The identification of the candidate period at 101.4 ms for AX J1418.7--6058 using data sets 07 and 09.
505: The same as in Figure~\ref{asso1} and \ref{asso2},
506: the trial period range shown in the lower panel of this
507: figure is roughly the range specified by the criterion that the
508: characteristic age is larger than 1000 years.
509: The two possible period associations, which result in different period time derivatives, are both included in Table~\ref{pcan}.
510: }
511: \label{asso3}
512: \end{figure}
513: %FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
514: The two candidate periods of AX J1809.8--2332 listed in Table \ref{pcan} are
515: related to each other.
516: Based on their pulse profiles
517: we suggest the period at 0.13247455 s be the second harmonic of the
518: one at 0.264949 s
519: if they are real.
520: In such a case, the derived period derivative, characteristic age and
521: spin-down power for the association between periodicity signatures found in
522: data sets 15 and 17 near 0.1325 s will be 2, 1 and 0.25 times those listed in
523: Table \ref{pcan} respectively. It makes this association even more unfavoured
524: if one compares its spin-down power with that of the Crab pulsar;
525: see discussion in the next section.
526: On the other hand, the strong periodicity signature (Figure~\ref{hdis1809})
527: makes the
528: candidate period at 0.264949 s very tempting.
529: %FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
530: \begin{figure}
531: \epsfxsize=8.4cm
532: \epsffile{fig6.eps}
533: \caption{The $H$-test blind search results of AX J1809.8--2332
534: with data set 15.
535: The two highest features correspond to periods at 0.13247455 s and
536: 0.2649491 s respectively.
537: }
538: \label{hdis1809}
539: \end{figure}
540: %FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
541:
542: To examine whether the procedure described above is useful, we applied it to the case of
543: RX J1856.5--3754, whose X-ray emission is believed to be thermal emission from
544: the surface of a neutron star.
545: There were several XMM observations of that source, but no periodicity was found until
546: a period of 7.055 s with a random probability of $6\times 10^{-4}$ was reported by
547: \citet{tiengo07}.
548: That result was obtained by analysing the XMM data taken in October 2006.
549: Applying our cross-checking method as described above to previous data,
550: we found that the 7-s period was also picked out
551: and suggested as a candidate period.
552: It demonstrates, although not guarantees, the usefulness of our cross-checking method.
553:
554: %SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS
555: \section{Discussion}
556:
557: We applied the periodicity cross-checking procedure described in
558: Section 3 to the 8 X-ray sources listed in Table~\ref{data}.
559: Candidate periods obtained with this method are listed in Table~\ref{pcan}.
560: Future periodicity searches at X-rays or gamma-rays
561: may be conducted around these
562: candidate periods, instead of performing blind searches.
563: Among all these candidate periods, the one with the lowest random probability (4.1$\times10^{-5}$) is
564: 145.2297 ms for AX J1420.1-6049. Its implied characteristic age, 4800 years, is between that of the
565: Crab pulsar and of the Vela pulsar. However, its corresponding spin-down power is similar to that
566: of the Vela pulsar, which in turn is about 100 times smaller than that
567: of the Crab.
568: If there is a correlation
569: between the characteristic age and the spin-down power,
570: this candidate period is not favoured because
571: its associated spin-down power is too small for its
572: 4800-year characteristic age.
573:
574: Indeed, among the currently known gamma-ray pulsars,
575: the spin-down power decreases with increasing characteristic ages.
576: The Crab pulsar, with a characteristic age of $10^{3.1}$ years, has a spin-down power about
577: $10^{38.7}$ erg/s. That of the Vela pulsar are $10^{4.1}$ years and $10^{36.8}$ erg/s and
578: of Geminga are $10^{5.5}$ years and $10^{34.5}$ erg/s.
579: One possible way to further assess the reality of these candidate periods, therefore,
580: is to compare their corresponding
581: properties, such as the characteristic age and the spin-down power, with that of
582: known gamma-ray pulsars.
583: Along this line of thinking, the candidate periods
584: at 203 ms for AX J1418.2-6047,
585: at 159 ms for AX J1418.6-6045,
586: at 19.6 ms and 145 ms for AX J1420.1-6049,
587: and at 22.4 ms, 33.5 ms and 168 ms for AX J1836.2+5928
588: are favoured, since their corresponding characteristic ages and spin-down powers
589: are roughly consistent with that of known gamma-ray pulsars.
590:
591: Another property to consider is the X-ray to gamma-ray energy spectral index, defined in
592: \citet{roberts01}, which assumes an X-ray counterpart to those unidentified GeV sources
593: and leads to three X-ray brightness categories for the GeV sources under study.
594: In that scheme, the Crab pulsar is X-ray bright, the Vela pulsar is X-ray moderate, and
595: Geminga is X-ray faint.
596: The GeV sources associated with the aforementioned X-ray sources
597: are all X-ray moderate (similar to Vela),
598: except for GeV J1835+5921, which is X-ray faint (similar to Geminga).
599: %Some assumed X-ray counterparts in \citet{roberts01} are different from the X-ray sources studied
600: %here. We have examined the X-ray brightness category for these 6 GeV sources with the possibile X-ray
601: %counterparts we adopt in this worked and found that they all stay in the same category as determined
602: %in \citet{roberts01}.
603: Examing the characteristic ages, we note that the four candidate periods
604: at 203 ms for AX J1418.2-6047,
605: at 159 ms for AX J1418.6-6045,
606: at 145 ms for AX J1420.1-6049
607: and at 168 ms for AX J1836.2+5928
608: survive the comparison.
609: The candidate period at 127.5 ms for RX J0007.0+7302 reported
610: by \citet{lin05} is also supported by these comparisons.
611:
612: These five X-ray sources are possible counterparts to three (or four)
613: GeV sources,
614: namely, GeV J0008+7304, GeV J1417--6100 (or a probable separate source
615: 3EG J1420--6038), and GeV J1835+5921.
616: For the former two (or three),
617: if the candidate periods suggested here turn out to be true,
618: they will be more Vela-like, in terms of timing properties.
619: It would suggest that the major cause for a gamma-ray pulsar
620: to be radio quiet is
621: likely geometrical, rather than intrinsic, and the
622: gamma-ray beam is wider than the radio one so that
623: for a certain viewing angle both
624: the radio and gamma-ray emissions are observed, as for the case of Vela,
625: and in another viewing angle only the gamma-ray emission is observed.
626: Instead, if all the radio-quiet gamma-ray pulsars are similar to Geminga
627: in their timing properties, the cause of being radio quiet, although not yet
628: clear, is probably intrinsic.
629:
630: Timing properties of not-yet-discovered gamma-ray pulsars may not be similar to
631: the known ones, particularly because the number of the latter is still small.
632: All the candidate periods listed in Table~\ref{pcan}
633: may therefore still be helpful
634: for pinning down the periodicity of emissions from
635: these sources in the future when new X-ray data or gamma-ray data obtained with AGILE and GLAST
636: become available. A larger sample of gamma-ray pulsars, either radio loud or quiet, will help us to understand
637: better the radiation mechanisms in the magnetospheres of gamma-ray pulsars.
638:
639: \section*{Acknowledgments}
640:
641: This research has made use of data obtained through
642: the High Energy Astrophysics Science Archive Research Center Online Service,
643: provided by the NASA/Goddard Space Flight Center.
644: This work was supported by the National Science Council and the
645: National Space Organization of
646: the Republic of China
647: through grants NSC 96-2628-M-007-012-MY3 and
648: 96-NSPO(B)-SP-FA04-01.
649:
650:
651: \begin{thebibliography}{99}
652: \bibitem[\protect\citeauthoryear{Aharonian et al.}{2006}]{aharonian06}
653: Aharonian, F., et al., 2006, A\&A, 456, 245
654: \bibitem[\protect\citeauthoryear{Braje et al.}{2002}]{braje02}
655: Braje, T.M., Romani, R.W., Roberts, M.S.E., Kawai, N., 2002, ApJ, 565, L91
656: \bibitem[\protect\citeauthoryear{Cheng et al.}{2000}]{cheng00}
657: Cheng, K.S., Ruderman, M., Zhang, L., 2000, ApJ, 537, 964
658: \bibitem[\protect\citeauthoryear{D'Amico et al.}{2001}]{damico01}
659: D'Amico, N., et al., 2001, ApJ, 552, L45
660: \bibitem[\protect\citeauthoryear{De Jager et al.}{1989}]{dejager89}
661: De Jager, O.C., Swanepoel, J.W.H., Raubenheimer, B.C., 1989, A\&A, 221, 180
662: \bibitem[\protect\citeauthoryear{De Jager}{1994}]{dejager94}
663: De Jager, O.C., 1994, ApJ, 436, 239
664: \bibitem[\protect\citeauthoryear{Halpern et al.}{2007}]{halpern07}
665: Halpern, J.P., Camilo, F., Gotthelf, E.V., 2007, ApJ, 668, 1154
666: \bibitem[\protect\citeauthoryear{Halpern et al.}{2002}]{halpern02}
667: Halpern, J.P., Gotthelf, E.V., Mirabal, N., Camilo, F., 2002, ApJ, 573, L41
668: \bibitem[\protect\citeauthoryear{Halpern \& Holt}{1992}]{halpern92}
669: Halpern, J.P., Holt, S.S., 1992, Nature, 357, 222
670: \bibitem[\protect\citeauthoryear{Harding et al.}{2007}]{harding07}
671: Harding, A.K., Grenier, I.A., Gonthier, P.L., 2007, Ap\&SS, 309,221
672: \bibitem[\protect\citeauthoryear{Hessels et al.}{2004}]{hessels04}
673: Hessels, J.W.T., Roberts, M.S.E., Ransom, S.M., Kaspi, V.M., Romani, R.W.,
674: Ng, C.-Y., Freire, P.C.C., Gaensler, B.M., 2004, ApJ, 612, 389
675: \bibitem[\protect\citeauthoryear{Hirotani}{2007}]{hirotani07}
676: Hirotani, K., 2007, ApJ, 662, 1173
677: \bibitem[\protect\citeauthoryear{Johnston \& Weisberg}{2006}]{johnston06}
678: Johnston, S., Weisberg, J.M., 2006, MNRAS, 368, 1856
679: \bibitem[\protect\citeauthoryear{Kuz'min \& Losovskii}{1999}]{kuzmin99}
680: Kuz'min, A.D., Losovskii, B.Y., 1999, Astron. Lett., 25, 80
681: \bibitem[\protect\citeauthoryear{Lamb \& Macomb}{1997}]{lamb97}
682: Lamb, R.C., Macomb, D.J., 1997, ApJ, 488, 872
683: \bibitem[\protect\citeauthoryear{Leahy}{1987}]{leahy87}
684: Leahy, D.A., 1987, A\&A, 180, 275
685: \bibitem[\protect\citeauthoryear{Lin \& Chang}{2005}]{lin05}
686: Lin, L.C.-C., Chang, H.-K., 2005, Ap\&SS, 297, 361
687: \bibitem[\protect\citeauthoryear{Muslimov \& Harding}{2004}]{muslimov04}
688: Muslimov, A.G., Harding, A.K., 2004, ApJ, 606, 1143
689: \bibitem[\protect\citeauthoryear{Ng, Roberts \& Romani}{2005}]{ng05}
690: Ng, C.-Y., Roberts, M.S.E., Romani, 2005, ApJ, 627, 904
691: \bibitem[\protect\citeauthoryear{Reimer et al.}{2001}]{reimer01}
692: Reimer, O., Brazier, K.T.S., Carrami\~nana, A., Kanbach, G., Nolan, P.L., Thompson, D.J.,
693: 2001, MNRAS, 324, 772
694: \bibitem[\protect\citeauthoryear{Roberts et al.}{1999}]{roberts99}
695: Roberts, M.S.E., Romani, R.W., Johnston, S., Green, A.J., 1999, ApJ, 515, 712
696: \bibitem[\protect\citeauthoryear{Roberts et al.}{2001a}]{roberts01}
697: Roberts, M.S.E., Romani, R.W., Kawai, N., 2001a, ApJS, 133, 451
698: \bibitem[\protect\citeauthoryear{Roberts et al.}{2001b}]{roberts01b}
699: Roberts, M.S.E., Romani, R.W., Johnston, S., 2001b, ApJ, 561, L187
700: \bibitem[\protect\citeauthoryear{Roberts et al.}{2002}]{roberts02}
701: Roberts, M.S.E., et al., 2002, ApJ, 577, L19
702: \bibitem[\protect\citeauthoryear{Roberts et al.}{2006}]{roberts06}
703: Roberts, M.S.E., Gotthelf, E.V., Halpern, J.P., Brogan, C.L., Ransom, S.M.,
704: 2006, in Proceedings of the 363 WE-Heraeus Seminar on Neutron Stars
705: and Pulsars, eds.\ W. Becker, H.-H. Huang (MPE Report 291), 24
706: (astro-ph/0612631v1)
707: \bibitem[\protect\citeauthoryear{Takata \& Chang}{2007}]{takata07}
708: Takata, J., Chang, H.-K., 2007, ApJ, 670, 677
709: \bibitem[\protect\citeauthoryear{Takata et al.}{2007}]{takata07b}
710: Takata, J., Chang, H.-K., Cheng, K.S., 2007, ApJ, 656, 1044
711: \bibitem[\protect\citeauthoryear{Thompson}{2004}]{thompson04}
712: Thompson, D.J., 2004, in Cosmic Gamma-ray Sources, ed Cheng, K.S., Romero, G.E. (Kluwer, Dordrecht), p.149
713: \bibitem[\protect\citeauthoryear{Tiengo \& Mereghetti}{2007}]{tiengo07}
714: Tiengo, A., Mereghetti, S., 2007, ApJ, 657, L101
715: \end{thebibliography}
716:
717: \label{lastpage}
718: \end{document}
719:
720:
721: