Multifrequency studies of the enigmatic gamma-ray source 3EG J1835+5918

Abstract

The EGRET telescope aboard the NASA Compton Gamma-Ray Observatory (CGRO) has repeatedly detected 3EG J1835+5918, a bright and steady source of high-energy gamma-ray emission which has not yet been identified. The absence of any likely counterpart for a bright gamma-ray source located 25° off the Galactic plane initiated several attempts of deep observations at other wavelengths. We report on counterparts in X-rays on a basis of a 60-ks ROSAT HRI image. In order to conclude on the plausibility of the X-ray counterparts, we reanalysed data from EGRET at energies above 100 MeV and above 1 GeV, including data up to CGRO observation cycle 7. The gamma-ray source location represents the latest and probably the final positional assessment based on EGRET data. We especially address the question of flux and spectral variability, here discussed using the largest and most homogeneous data set available at high-energy gamma-rays for many years. The results from X-ray and gamma-ray observations were used in a follow-up optical identification campaign at the 2.2-m Guillermo Haro Telescope at Cananea, Mexico. VRI imaging has been performed at the positions of all of the X-ray counterpart candidates, and spectra were taken where applicable. The results of the multifrequency identification campaign toward this enigmatic unidentified gamma-ray source are given, especially on the one object which might be associated with the gamma-ray source 3EG J1835+5918. This object has the characteristics of an isolated neutron star and possibly of a radio-quiet pulsar.

1 Introduction

The high-energy gamma-ray source 3EG J1835+5918 (also known as GRO J1837+59, 2EG J1835+5919, and GEV J1835+5921) has been subject of considerable interest since its discovery early in the CGRO mission by EGRET (Nolan et al. 1994). Unidentified sources at high-energy gamma-ray wavelengths present significant challenges today. The large error boxes preclude any simple identification based on positional association only. The absence of gamma-ray lines eliminates the possibility of making direct redshift or source composition measurements. In most cases, the photon statistics are inadequate to carry out conclusive periodicity searches. Even long-term time variability is hard to establish unless strong flaring behaviour is seen. Multiwavelength studies have been one of the most useful identification techniques. The blazar-class active galactic nuclei (AGN) seen by EGRET, for example, are characterized by strong radio emission (typically ∼1 Jy at 5 GHz) and a spectral energy distribution (SED) with one peak at IR–UV frequencies and a second peak in the gamma-ray band. Gamma-ray pulsars have high F/F_radio and often F/F_X ratios, with a single clear SED peak in the hard X-ray to gamma-ray range. The energy requirements of gamma-ray production demand powerful sources; therefore ordinary stars and most normal galaxies can almost certainly be ruled out as gamma-ray sources.

3EG J1835+5918 is a bright gamma-ray source discovered at high Galactic latitudes by EGRET, and this source was not until recently identified with any plausible counterpart at other wavelengths. This has to be seen in respect to the observational fact that, other than 3EG J1835+5918, all of the bright gamma-ray sources at high Galactic latitudes have proved to be coincident with blazar-type AGN or flat spectrum radio quasars (FSRQ). However, even the most recent correlation of gamma-ray sources from the Third EGRET Catalogue (Hartman et al. 1999) with radio observations at 4.85 GHz from the Greenbank survey does not suggest any radio counterpart for 3EG J1835+5918 (Mattox, Hartman & Reimer 2001). A dedicated deep search for pulsed radio emission with the Greenbank telescope at 770 MHz towards 3EG 1835+5919 yielded only a 1-mJy flux density upper limit (Nice & Sawyer 1997). At the neighbouring wavelengths in gamma-rays, observations have yielded only upper limits for an object at the position of 3EG J1835+5918: the Whipple Telescope from observations made in 1993 (Kerrick et al. 1995), and COMPTEL throughout their total exposure from 7 yr of CGRO observations (Schönfelder, Romani & Kawai 2000). At the 5th Compton Symposium (Reimer et al. 1999; Carramiñana et al. 1999), we reported for the first time on X-ray counterparts for 3EG J1835+5918, suggesting an indication of a possible radio-quiet pulsar. An independent search by Mirabal et al. (2000) utilizing the same X-ray data, but separate optical observations, concluded instead that 3EG J1835+5918 was likely to be a prototype of a new type of high-energy gamma-ray source.¹ Here we report our final analysis of the EGRET, X-ray, and optical data. Our analysis reinforces our original conclusion that 3EG J1835+5918 has the characteristics of an isolated neutron star, possibly a Geminga-like radio-quiet pulsar.

2 The EGRET source 3EG J1835+5918

3EG J1835+5918 was first discovered at photon energies above 100 MeV by the EGRET instrument aboard the NASA Compton Gamma-Ray Observatory during regularly scheduled observations in 1991. The source was repeatedly seen whenever it was in the field of view of the EGRET instrument. However, the first observation performed with a close on-axis pointing towards 3EG J1835+5918 using EGRET was only made in CGRO observation cycle 7 in 1998. EGRET observations performed at large off-axis viewing angle are problematic due to the degradation of the instrumental point spread function (PSF), most recently noted in Esposito et al. (1999). The first report on GRO J1837+59 (Nolan et al. 1994) included only data from EGRET observations between 1991 and 1993. On the basis of six analysed viewing periods a weak indication of flux variability was given. However, 3EG J1835+5918 was only once observed at less than 20° off-axis. The authors noted that, under those circumstances, a factor of two flux variation is not a strong indication for flux variability. 3EG J1835+5919 is listed in the Second EGRET catalogue of gamma-ray point sources (Thompson et al. 1995), which utilized a similar set of viewing periods from CGRO observation cycles 1 and 2 as Nolan et al. (1994). A separate variability study including EGRET data from cycles 1 to 3 again reported the source as being variable (McLaughlin et al. 1996), but the determined value of the variability criterion lies in the range where this parameter is rather inconclusive in deciding whether or not a source is indeed variable. With the appearance of the Third EGRET catalogue and the GeV source compilations (Lamb & Macomb 1997; Reimer et al. 1997), results from a total of 12 individual observations of 3EG J1835+5919 were reported. The source remains the brightest unidentified EGRET source outside the Galactic plane. Furthermore, it has been documented that there is gamma-ray emission from 3EG J1835+5919 up to the highest energies observable by EGRET.

In order to extend the coverage of 3EG J1835+5919 to its maximum we use for the first time all gamma-ray data taken by EGRET through the CGRO mission, including the previously unpublished observations made at a small off-axis angle in observation cycle 7. Table 1 lists the EGRET observations of 3EG J1835+5918.

Throughout our analysis, we clearly distinguish between observations in which the angle between 3EG J1835+5918 and the instrument pointing direction was within or without 25°. This distinction has been recommended by the EGRET instrument team for using the standard PSF (sources within 25° of the instrumental pointing) or using the wide-angle PSF if outside. The EGRET observations from CGRO observation cycle 7 extend significantly beyond the catalogued observations. They are separated by more than 3 yr from the previous observations of 3EG J1835+5918. Both the long-term observational aspect and the quality of the observation have been improved: despite the lower efficiency of the EGRET spark chamber, the 1998 observations were the first on-axis observations, unbiased from effects which most of the earlier EGRET observations of 3EG J1835+5918 could suffer from. The narrow field-of-view mode, used in EGRET viewing periods 710 and 711, does not introduce additional problems in this respect.

In order to determine the most likely location of the gamma-ray source, we co-added the viewing periods using two different energetic thresholds , and , respectively) and analysed them individually. Because 3EG J1835+5918 is a strong emitter up to GeV-energies, the best source location was obtained from the highest energetic photons, where the instrumental PSF is significantly smaller than at lower energies. Using a likelihood method (Mattox et al. 1996), we determined the best position (>1 GeV) to be , , which is consistent with the position of 3EG J1835+5918 from our analysis above 100 MeV , , the position given in 3EG and GEV catalogues, and the elliptical fit from Mattox et al. (2001),, , , , . However, the additional data gave us positional errors of only 6 arcmin and 8 arcmin for the 68 and 95 per cent confidence region, respectively. This enabled us to perform deep X-ray and optical studies of the entire region of 3EG J1835+5919. By using the position determined above 1 GeV, we examined the individual viewing periods in order to evaluate the long-term characteristic of the gamma-ray source flux. As discussed earlier, Nolan et al. (1994) and McLaughlin et al. (1996) indicated some source variability in 3EG J1835+5919 on the basis of smaller data sets than presented here. However, the most recent variability study (Tompkins 1999) puts 3EG J1835+5918 clearly among the non-variable sources, similar to the identified gamma-ray pulsars. Tompkins made use of an algorithm especially adopted for the characteristics of the observations by EGRET, i.e. sparse data sets from individual observations, often widely separated in time and characterized by different background levels. In addition, data from individual EGRET observations up to CGRO observation cycle were used in this study. A strict data selection (only within 25° on-axis) among the gamma-ray observations was used, assuring a data set of comparable quality. We complement the flux history of 3EG J1835+5918 with the data from 1998 January 13–27, the last high-energy gamma-ray data on this source to be taken for some years.

Owing to the generally lower efficiency of the EGRET spark chamber towards the end of the EGRET mission, the early 1998 viewing periods were evaluated using adjusted normalization factors (Esposito et al. 1999). These factors were checked quantitatively by means of a similar on-axis observation of Geminga during 1998 July 7–21. Assuming that the instrumental sensitivity has not changed appreciably between these observations and that Geminga remains the stable gamma-ray emitter previously observed, this normalization for Geminga could be applied to the flux of 3EG J1835+5918 in the cycle 7 observations. The light curve of 3EG J1835+5918 displaying all relevant EGRET observations is shown in Fig. 1, fluxes are for photon energies above 100 MeV, determined at the likelihood position of the GeV source. In addition, we consider (and label) observations with up to 25° off-axis separately from observations outside 25°. The fluxes above 100 MeV and above 1 GeV appear to be linearly correlated, considering the uncertainties in individual viewing periods arising from photon statistics, especially for the sparse data of the detections above 1 GeV.

We conclude, there is no indication of flux variability after all for 3EG J1835+1918, neither above 100 MeV nor above 1 GeV. Given the differing quality of the EGRET observations within their statistical and systematical uncertainties, we confirm the result from Tompkins (1999) and find 3EG J1835+5918 compatible with a non-variable source of an average flux of .

After concluding that 3EG J1835+5918 is consistent with having constant gamma-ray flux throughout the EGRET mission, we still have to examine the issue of its spectral variability. Nolan et al. (1996) reported evidence for spectral variability between individual EGRET viewing periods. Apparently, no correlation between spectral index and flux was found. Hence, we re-examined the EGRET data on 3EG J1835+5918 for indication of spectral variability. Individual spectra in each of the relevant viewing periods were determined by simultaneously analysing likelihood excesses of 3σ detection significance and above. We derived a flux value or upper limit in each of ten energy intervals (30 MeV to 10 GeV) using a likelihood method. In cases when poor count statistics gave a spectrum dominated by upper limits, the ten energy intervals were recombined into four (30–100, 100–300, 300–1000, > 1000 MeV), followed by the appropriate determination of the spectral slope. Also, when the source position determined from likelihood analysis of an individual observation differed from the GeV-position, both positions were individually considered for consequences for the resulting spectrum. None of them introduces relevant modifications in the resulting spectral slope. Therefore, the determined individual spectra could be compared at the best level currently achievable for an unidentified high-energy gamma-ray source.

We find that the spectra of 3EG J1835+5918 determined from individual viewing periods are fully compatible within their statistical and systematic uncertainties throughout the entire EGRET mission, see Fig. 2. A single power-law spectral index of is consistent within 1σ for all individual spectra.

With the consistency of the individual spectra throughout the EGRET observations established, we co-added the data from cycles 1 to 7 in order to determine the best overall spectrum of 3EG J1835+5918. A single power-law fit appears to be inadequate for this source. The spectrum of 3EG J1835+5918 resembles the gamma-ray spectra of known gamma-ray pulsars like Vela or Geminga (Thompson et al. 1997) and the spectra of candidate gamma-ray pulsars like 3EG J0010+7309, as can be seen in Fig. 3: the hard power-law spectral index, as determined to be between 70 MeV and 4 GeV, the high-energy spectral cut-off or turnover as well as a possible spectral softening at the low energies. The restriction in energy when applying a single power law fit takes these features into account: it is based on the bins with the highest instrumental sensitivity. The upper limits from COMPTEL (Schönfelder et al. 2000) do not constrain the shape of the spectrum at lower energies. The TeV upper limits as reported by Kerrick et al. (1995) are consistent with a rollover at 4 GeV, but certainly not with a simple extrapolation of the EGRET measured power-law spectrum to the highest energies.

3 X-ray observations of EG J185+5918

The first observation towards 3EG J1835+5918 took place in 1995 February. A 9-ksec ROSAT HRI observation was performed, which reached a minimum detectibility limit of about in the 0.1-to-2.4 keV band. A longer HRI observation was taken in 1997 December/1998 January as proposed by us in the ROSAT guest observer program. It exceeded the exposure of the former HRI observation by a factor of six with a total of 61.269 kseconds. Assuming a power-law spectrum with a photon index of −2 and a Galactic hydrogen column density of , the limiting unabsorbed X-ray flux is about . Table 2 lists the detected X-ray sources in the long ROSAT HRI observation before astrometric corrections were applied. The astrometric correction is discussed in the context of reliable optical counterparts, see next section. Fig. 4 shows the ROSAT HRI image, with the individual sources marked. Overlaid is the gamma-ray source location contour, determined solely from photons with energies above 1 GeV. The contours represent the 68 and 95 per cent likelihood boundaries of the source location.

Only three of the 10 sources found in this deeper HRI observations were detected in the earlier, short HRI observation (objects 1, 3, 9). We only note this here because the one object of further interest after investigating the deep HRI image, RX J1836.2+5925, is rather close to the detection limit of the short observation. When examining X-ray variability for this particular source, it is generally hard to conclude based on two detections only, especially with one relatively close to its detection limit. However, any report on X-ray source variability would push the interpretations rather hard into some unique direction. Therefore we make no conclusions on the variability until further observations have been made and analysed.

The most recent X-ray observation took place between 1998 April 20 and 22, performed by the GIS and SIS detectors on ASCA. The SIS data did not detect any object in the vicinity of 3EG J1835+5918, and the detected sources from the stacked GIS-images are located outside the GeV-source location we consider. There are no constraints from the non-detection of X-ray sources in the SIS images compared to the sensitivity achieved from the deep HRI observation.

4 Optical identifications in the vicinity of 3EG J1835+5918

We studied the X-ray sources found in the vicinity of 3EG J1835+5918 for optical counterparts. If objects at optical wavelengths appear coincident at positions of the X-ray sources, the astrometry of the obtained ROSAT HRI image could be verified and corrected. Deep follow-up observations have been conducted in order to conclude on the nature of those optical counterparts. Starting with a general assessment on the basis of DSS-2 plates and USNO-A2.0 catalogue listings, we subsequently performed VRI imaging of the entire field of the gamma-ray source at the Observatorio Astrofísico Guillermo Haro, located in Cananea, Sonora , . We used the Faint Object Spectrograph Camera (LFOSC) of the Landessternwarte Heidelberg, specially designed for optical counterpart identifications. The instrument allows photometric BVRI imaging in a FOV and two modes of low-resolution spectroscopy. Observations in 1997 July and September, and 1998 May were devoted to VRI imaging. When the ROSAT HRI image taken at 1998 January had been delivered and thoroughly analysed, spectroscopy was made in 1999 June and October by using the low-resolution mode at 4200 to 9000 Å with a ∼8.3 Å pixel⁻¹ sampling.

Fortunately, the search for counterparts has revealed one excellent astrometrical measure in the field, a star listed in the Tycho-2 (3917 00934 1) and ACT-catalogues. The offset of the coincident X-ray object 8 (RX J1834.2+5920) is about 0.5 s in right ascension and 1.1 arcsec in declination. A similar offset has been found at the USNO-A2.0 listed object coincident with the X-ray object 5 (RX J1835.9+5923). Offsets at other object pairs differ more, significantly at the edge of the field of view. Therefore we do not average any additional, in some cases contradictory offset parameter. We apply the correction found appropriate for objects 8 and 5 throughout the entire HRI image. The following objects were studied.

• (i)

  RXJ1837.0+5934. This X-ray source is outside the 68 and 95 per cent GeV error contour, already signalling that an association is unlikely. Two USNO-A2.0 listed objects are positionally consistent with the X-ray source position. The optical spectrum of the brighter object shows a strong emission line at 7198 Å which can be identified as redshifted Hβ, allowing us to associate the emission lines at 6364 and 6012 Å with redshifted Hγ and Hδ. This quasar at is a very plausible counterpart for the X-ray source.

• (ii)

  RXJ1835.9+5926. An optical object, located 7 arcsec NE from the X-ray source position has been found at the Cananea images . Its final identification with a QSO at by Mirabal et al. (2000) is plausible.

• (iii)

  RXJ1836.2+5925. Two faint objects are at 11 arcsec NW and 14 arcsec SE of the X-ray source location, not listed in the USNO-A2.0. Given the low uncertainty on the X-ray position, an association between either of them and the ROSAT source is doubtful. Our images have a detection limit around , and no other object is found anywhere closer to the X-ray source position.

• (iv)

  RXJ1836.6+5924. An optical object , lies centred on the X-ray position. It has been identified with a QSO at by Mirabal et al. (2000).

• (v)

  RXJ1835.9+5923. The optical object nearly centred at the X-ray position , shows a strong emission line, which we identify with C iv1550, recovering C iv1909 redshifted to 5469 Å (emission doublet) and Mg 2798 redshifted to ∼8020 Å. This quasar with is very likely to be the X-ray emitting source.

• (vi)

  RXJ1834.4+5920. An optical object is more than 10 arcsec distant from the X-ray position for which we measured and . Its spectrum indicates a late type star, probably M5V. This X-ray source is 1.3 arcmin from the bright star coincident with X-ray source number 8, whose glow complicates the detection of objects fainter than magnitude 20. The potential identification of the M5V star with the X-ray source would rule out their association with the 3EG J1835+5918. If one does not follow this identification scheme due to the source location offset, we find no other optical counterpart for this object. However, this would leave the question unanswered, why we observe no X-ray emission from this M5-star down to the detection limit of the ROSAT HRI image.

• (vii)

  RXJ1836.6+5920. There is no optical object brighter than inside the error box. Mirabal et al. (2000) found an optical object at at by a UV-excess selection technique using the Hobby–Eberly–Telescope. Although, we cannot adopt their argument of rejection by positional inconsistency of RX J1836.6+5920 with the EGRET-source location contour, the optical and X-ray properties of this QSO would make it highly unlikely to be the counterpart of 3EG J1835+5918.

• (viii)

  RXJ1834.2+5920. The star Tycho 3917009341 with V ∼ 9.4 is centred at the X-ray source position. The spectrum obtained indicates a late G dwarf star, probably G7V.

• (ix)

  RXJ1835.5+5915. A magnitude object is about 5 arcsec from the X-ray position, and therefore marginally consistent with the X-ray source. Its spectrum indicates a late M-type star.

• (x)

  RXJ1836.8+5910 is outside the 68 and 95 per cent location contour of the EGRET source, and their physical association is practically rejected on positional inconsistency. A bright star (Tycho 3917018071) is coincident with the X-ray position. Our spectrum indicates a K star, probably K5V.

We reject only sources 1 and 10 primarily on the grounds of inconsistency with the GeV source location. If this argument does not absolutely disqualify both X-ray sources as candidates for an association with the gamma-ray source, their identifications probably do. Also, the identified coronal emitting stars do not qualify as likely counterparts for 3EG J1835+5918. The four QSO-identifications , 1.75, 1.87, 1.86) certainly need a closer look. First of all, these are not blazar-class AGN which constitute the vast majority of the known EGRET QSO sources. Also, their redshifts would lie at the tail of the redshift-distribution of gamma-ray loud active galactic nuclei (Mukherjee 1997). The lack of observed radio emission is an additional hint that they do not belong to the class of AGN that EGRET could actually see (Mattox et al. 2001). The obvious mismatch between one of the brightest gamma-ray sources at high Galactic latitudes and the lack of observable radio emission would exhibit 3EG J1835+5918 as a unique source among the gamma-ray sources identified with active galactic nuclei. Secondly, the intrinsic gamma-ray emission characteristics argue against a quasar counterpart. The EGRET-detected AGN are highly variable at all timescales currently observable. The average spectral index of the EGRET-detected AGN is significant softer than the one determined for 3EG J1835+5918, in fact it would put 3EG J1835+5918 as the hardest source among the more than ninety EGRET-detected AGNs. The lack of obvious source variability as well as the hard spectral index of would solely argue against a quasar identification; both arguments together do so with a rather high degree of confidence. Parameters like the optical to X-ray and/or optical to γ-ray luminosity ratio for the QSO-counterparts do not give a unique signature. The range which could be occupied by QSOs in both parameter spaces is rather wide. Also, the correlation signatures reported for 61 of the gamma-ray loud blazars (Cheng, Zhang & Zhang 2000) do not allow any additional conclusive information for the four X-ray sources identified as radio-quiet QSOs. Summarizing, the various observational facts concerning an association between each of the X-ray sources identified as quasars and the unidentified gamma-ray source 3EG J1835+5918 do not support such an interpretation. More exotic scenarios must be called in order to accept one of the quasar identifications, like hypothesized radio-quiet blazars (Mannheim 1993) or a γ-ray AGN with a shifted SED (Ghisellini 1999), where the synchrotron component will fall in the MeVs and the IC-component peaks at TeV-energies. Currently, observational constraints from COMPTEL and Whipple do not indicate that the SED peaks at any other wavelength band except where 3EG J1835+5918 is observable, bright and steady: the 30 MeV to 10 GeV band where EGRET operated for nine years.

Therefore, we have to consider further only one X-ray source in the vicinity of 3EG J1835+5918, RX J1836.2+5925 (object 3, see Table 2). The lack of an optical counterpart and therefore of the possibility to identify it by means of an optical spectrum keeps this source as the only unidentified X-ray source which could have an association with the γ-ray source 3EG J1835+5918. Using their independent optical observation, Mirabal et al. (2000) also concluded that RX J1836.2+5925 is the most probable X-ray counterpart to the gamma-ray source.

5 Discussion

3EG J1835+5918 is a persistent high-energy gamma-ray source located at high Galactic latitudes and has been observed repeatedly by EGRET. It is characterized by a hard power law and a spectral break or turn-over above 4 GeV. It appears to be a non-variable source in terms of its flux as well as its spectral shape throughout the entire EGRET mission, despite suggestions of variability from earlier analyses. Its gamma-ray properties are typical of those observed from gamma-ray pulsars like Vela or Geminga, and candidate radio-quiet pulsars like 3EG J2020+4017, 3EG J0010+7309, and 3EG J2227+6122. Our deep ROSAT HRI observation revealed several X-ray sources consistent with the location of the observed GeV-emission of 3EG J1835+5918. As a result of the identification campaigns independently carried out by Mirabal et al. (2000) and ourselves, only one of the ten X-ray sources still attracts interest to be considered further for an association with the γ-ray source. This source, RX J1836.2+5925, is characterized by an obvious lack of radio-emission, indetectibility by means of an UV-excess identification technique, lack of optical counterpart up to in the V- and B-bands, and location well inside the 68 per cent likelihood test statistic contours of 3EG J1835+5918. Our HRI observation contain no information on the X-ray spectrum of RX J1836.2+5925. Hence, assuming that this X-ray source is the most likely counterpart to 3EG J1835+5918, we are restricted to using the X-ray flux of RX J1836.2+5925 and the gamma-ray properties of 3EG J1835+5918 to investigate the characteristics of the object.

To do so, we can use its multifrequency properties to ascertain its characteristics. The high F/F_radio value seems to rule out a blazar origin. The already noted similarities in the gamma-ray characteristics with known gamma-ray pulsars (Thompson et al. 1997) or radio-quiet pulsar candidates (for a recent summary, see Brazier & Johnston 1999) definitely suggest such a radio-quiet pulsar candidate. We have re-examined the gamma-ray and X-ray fluxes for all the known and candidate pulsars, using a consistent energy range in both bands. Comparing the flux of 3EG J1835+5918 in γ-rays and RX J1836.2+5925 in X-rays , this source falls among the candidates currently considered for associations between gamma-ray sources and X-ray sources with proven or suspected neutron star origin, see Fig. 5. Nearly all of the candidate gamma-ray pulsars lie at the bottom end of the sensitivity feasible for the last generation of X-ray instruments like ROSAT, ASCA and SAX. Obviously, only deep observations could reveal counterparts at all or with features not easily explained by any other astronomical objects. The lack of optical counterparts up to V ∼ 23 mag and radio emission for RX J1836.2+5925 is a further characteristic signature for isolated, radio-quiet neutron stars (Caraveo, Bignami & Trümper 1996), and ideally demonstrated by Geminga as prototype (Bignami & Caraveo 1996).

Although many of the candidate radio-quiet pulsars beside Geminga itself are located within or near SNRs, 3EG J1835+5918 does not. Neither radio observations nor the X-ray data yield any hint of a SNR in the vicinity of this object, and the high Galactic latitude seems to rule out the possibility of obscuration that might hide one.

If 3EG J1835+5918/RX J1836.2+5925 is not of quasar origin and also not the first candidate of an hypothesized extragalactic astronomical object bright and steady in gamma-rays, faint in X-rays, and yet undetectable at optical and radio wavelengths, it will reside within our Galaxy. We therefore have to suspect an isolated radio-quiet neutron star candidate. With Geminga as the only established pulsar of a predicted class of radio-quiet pulsars, extremely well characterized with its highly resolved high-energy light curve, independent measurements of rather weak radio emission and a faint optical counterpart with noticeable proper motion, a comparison of observational parameters in analogy with 3EG J1835+5918/RX J1836.2+5925 might be appropriate. First, Geminga is three times brighter in gamma-rays and about fifty times brighter in X-rays. To extrapolate from the distance to Geminga , Caraveo et al. 1996) using the observed fluxes, 3EG J1835+5918 would lie between 250 pc (scaling from gamma-rays) and 1.1 kpc (scaling from X-rays), assuming the same beaming as Gemingas in both cases. Besides, pulsars tend to begin their life in the Galactic plane. A pulsar moving with a typical velocity of about 350 km s⁻¹ would move only 300 pc even in a lifetime of 10⁶ yr, while an object seen at would have to move more than 420 pc from the plane if it were at a distance greater than 1 kpc. As pointed out by Yadigaroglu & Romani (1995) discussing the beaming evolution of pulsars in an outer-gap model, the beaming fraction becomes rather small as the pulsars age increases. Therefore a distant but old pulsar would have to be immensely powerful or exceptionally beamed. If 3EG J1835+5918 were more distant, then its gamma-ray luminosity would exceed that of Geminga, but if it were closer, then the surface brightness in X-rays of the neutron star would have to be lower than Gemingas. This indicates, that either the efficiency of the emission mechanism is different and/or the parameter space which radio-quiet pulsar candidates could occupy is wide spread. In contrast to energetic pulsars like Vela or B1706−44, non-thermal emission or pulsar nebular features have not been observed in the case of RX J1836.2+5925 so far. Nor is it an extended source in the X-rays. The lack of an associated SNR as well as the rare chance to find a similar pulsar at such high Galactic latitude (i.e. nearby) argues against a young pulsar in the case of 3EG J1835+5918. However, the striking similarities in the gamma-ray properties between Geminga, other candidate radio-quiet pulsars and 3EG J1835+5918, the absence of a radio and optical counterpart of RX J1836.2+5925, and, finally, the arrangement of 3EG J1835+5918/RX J1836.2+5925 among the other candidate gamma-ray pulsars (Fig. 5) still leaves room for accepting a hypothesis being an older but radio-quiet pulsar.

Halpern et al. (2001) hypothesized in the case of 3EG J2227+6122/RX J2229.0+6114 a medium aged gamma-ray pulsar population, efficient enough to be comparable to older pulsars like Geminga or B1055−52, although matching the luminosity (i.e. spin-down power) constraints from gamma-ray observations. With 3EG J2227+6122/RX J2229.0+6114 rather close in its F/F_X to young pulsars with non-thermal emission and pulsar nebular features, 3EG J1835+5918/RX J1836.2+5925 lies well among the other candidate radio-quiet pulsars. We rather stress the similarities seen in the F/F_X to existing candidate gamma-ray pulsars, together with its characterization by apparently similar gamma-ray properties as identified pulsars, however faint in X-rays and no counterpart yet at optical and radio frequencies. Mirabal & Halpern (2001) have recently reached a similar conclusion.

With 3EG J1835+5918/RX J1836.2+5925 being in the range of a isolated neutron star candidate explanation, other suggestions still need to be looked at. A similarity to the widely discussed association between LSI +61°303/3EG J0241+6103 and SAX J0635+533/3EG J0634+0521 does not seem to be appropriate for various and strict reasons. Both are binary systems (Be/X-ray), characterized by different states of variability on all but the shortest timescales. Also, they show a gamma-ray spectrum significantly softer than determined in the case 3EG J1835+5918. Most severe, since only a small number of such Be/X-ray binaries is expected (Vanbeveren, de Loore &van Rensbergen 1998), the high Galactic location of 3EG J1835+5918 would indicate a rather nearby one, definitely conflicting with the lack of any optical counterpart up to 23 mag. With extremely high degree of confidence we can therefore rule out the probability of seeing a massive star/compact object binary system here.

There have been other suggestions made for 3EG J1835+5918. Plaga et al. (1999) explicitly refer to 3EG J1835+5918 as an example for a hotspot of a Galactic gamma-ray burst. The low X-ray luminosity of RX J1836.2+5925 and the upper limit from the Whipple observation are severe observational arguments against such a hypothesis. Isolated accreting black holes (Armitage & Natarajan 1999) were suggested to be energetically consistent with unidentified sources at all Galactic latitudes. However, if their high energy emission occurs via similar processes to those in AGNs, it is expected to observe variability on all timescales. This is definitely not the case for 3EG J1835+1918. The persistent nature of the detections from 3EG J1835+5918 also excludes any similarities to gamma-ray transient like GRO J1838−04.

Certainly, neither the X-ray data nor the gamma-ray data currently allow wide range period scans for pulsations without known ephemerides (Jones 1998). A search for periodicity will have to be postponed until more sensitive instruments like XMM in the X-rays or GLAST in the gamma-rays have observed 3EG J1835+5918. However, if a restrictive set of parameters can be predicted from pulsar models or if a light curve can be derived from another wavelength, the archival EGRET data will permit the discovery of pulsations in the gamma-rays. The long observational history presented here will certainly assist in any such effort. Finally, RX J1836.2+5925 might be identified as a neutron star by extremely deep optical imaging/spectroscopy. To unambiguously relate 3EG J1835+5918 to a known class of astronomical objects would be of extreme importance for any collective studies of gamma-ray sources, as well as for studies of contributors to the diffuse gamma-ray background, not to mention the gain for pulsar physics if the existence of another isolated neutron star in gamma-rays is confirmed.

Acknowledgments

A part of this work was performed while OR held a National Research Council/NASA GSFC Resident Research Fellowship. KTSB acknowledges the use of the Starlink facilities at the University of Durham, England. A part of this work was funded by CONACyT grant 25559-E (AC). We acknowledge LSW Heidelberg for the use of the LFOSC instrument an Cananea. The Digitized Sky Surveys were produced at the Space Telescope Science Institute under U. S. Government grant NAG W-2166. This work made use of the USNO-A2.0 catalogue.

References

Author notes