NEAR-CONTEMPORANEOUS OPTICAL SPECTROSCOPIC AND INFRARED PHOTOMETRIC OBSERVATIONS OF CANDIDATE HERBIG Ae/Be STARS IN THE MAGELLANIC CLOUDS

We observed 27 of the 28 EROS SMC HAeBe Candidates (ESHC) and EROS LMC HAeBe candidates (ELHC; Lamers et al. 1999; Beaulieu et al. 2001; de Wit et al. 2002, 2003) in the J (1.25 μm), H (1.635 μm), and Ks (2.150 μm) bands. The observations were made on 2005 January 9, using the Infrared Side Port Imager (ISPI) instrument (van der Bliek et al. 2004) on the Cerro Tololo Inter-American Observatory (CTIO) 4 m telescope (Table 1). The data were recorded on a Hawaii-2 array with a pixel scale of 0.3 arcsec pixel⁻¹, yielding a field of view of 1025 × 1025. The typical seeing during the night varied between 06 and 12. Observations of our standard star and science fields were obtained using a Fowler sample of 1, using the number of co-adds and integration per co-add listed in Table 2. We dithered the telescope after each integration to facilitate more accurate sky subtraction. Flat fielding was achieved by subtracting a series of dome flats with an illuminating lamp turned off from a series with the lamp turned on.

These data were reduced using cirred, a custom CTIO IR reduction package for IRAF.⁵ Our data were first divided by average flat-field images, constructed via the method described above. To perform sky subtraction, we computed the average background count of each image with all stars masked, and then replaced the pixel value of each star with the image's average background. All images in a dither set were then averaged to create a master sky-subtraction image for the group, and this master sky image was then subtracted from each image. A World Coordinate System (WCS) was assigned to each image using SExtractor (Bertin & Arnouts 1996) to identify stellar centroid positions and WCSTools (Mink 1997) to correlate these with the Two Micron All Sky Survey (2MASS) database. The high-order distortions of ISPI were computed and corrected using WCSTools and the IRAF routine ccmap. Finally, dithered images were registered and averaged using the program SWARP (Bertin et al. 2002).

We first performed aperture photometry on all stars in our data using a 36 (12 pixel) radius aperture and an annulus of radial size 36–60 (12–20 pixels) for background count determination. Photometric calibration to an absolute scale was achieved by comparing ∼20 stars per field of view, distributed across the entire array, to published 2MASS photometry (Skrutskie et al. 2006). The average difference between the instrumental photometry and archival 2MASS photometry of these stars defined the zero point for each of our fields of view (Table 2), and the standard deviation of this average defined the respective error in each zero point (Table 2). The accuracy of this method was checked by comparing our observations of select near-IR photometric standards from the Persson et al. (1998) catalog to published 2MASS values, and as seen in Table 3 these results are consistent to within the error of the measurements. To mitigate source confusion problems, we next computed an aperture correction to a 12 (4 pixel) radius aperture and extracted aperture photometry for all sources using this smaller aperture. The photometry reported in this paper utilizes this smaller, 12 aperture size.

We observed 21 of the 28 ESHC and ELHC stars on 2005 January 21 using Hydra, the 138 fiber multi-object spectrograph on the 4 m telescope at CTIO. The observations, summarized in Table 6 were made with the KPGL3 grating and 200 μm slit plate, providing coverage from ∼3900 to 6700 Å at R ∼ 2200. The width of the science fibers was 20 in diameter. Ten to fifteen blank-sky fibers were assigned per fiber configuration to measure the nebular sky emission near our targets. We used 3600 s integration times, yielding the signal-to-noise (S/N) per pixel, measured in a line-free continuum-normalized region near 6500 Å, shown in Table 6. Calibration data including bias frames, dome flats, and HeNeArXe "penray" lamp exposures were obtained, along with sky flats to measure fiber-to-fiber throughput, projector flats to trace the location of each fiber orientation, and milk flats to create bad pixel masks.

The data were reduced in IRAF using the dohydra package and followed standard practices for multi-object spectroscopy. Sky subtraction was performed in an interactive manner, using different nearby sky fibers to best subtract the nebular emission in each science fiber. Our choice of continuum placement is the dominant source of error in our quoted equivalent width (EW) in Table 6; we estimate this uncertainty to be <5% by computing EWs for a number of sources using a range of continuum definitions.

Potential source confusion has been noted in some follow-up investigations of ELHC stars, with de Wit et al. (2005) reporting concerns with ELHC 5, 6, 8, and 18 (the latter of which exhibited two images in their near-IR imagery) and Wisniewski & Bjorkman (2006) reporting that ELHC 11 might have two components. The updated coordinates (Table 1) and finder charts (Figure 1) provided by our better resolution data should help alleviate source identification problems for future investigations.

We found clear evidence in our data that ELHC 6 exhibited two sources at the position reported by de Wit et al. (2002; see Table 4), with brightnesses corresponding to 15.66 ± 0.03 and 16.10 ± 0.03 at J, 15.42 ± 0.03 and 15.63 ± 0.03 at H, and 15.42 ± 0.03 and 15.86 ± 0.03 at Ks. The point-spread function shape of ELHC 18 was also elongated and likely comprised of two sources, confirming the confusion reported by de Wit et al. (2005). ELHC 4 and ELHC 11 also had clear evidence of nearby companions at radial distances of ∼12 and ∼15 away, respectively. Careful examination of ELHC 5 and ELHC 8, reported to have potential nearby neighbors (de Wit et al. 2005), revealed no evidence of source contamination in our data.

IR photometric variability has been previously observed in both Galactic (Eiroa et al. 2002; Bary et al. 2009; Muzerolle et al. 2009) and Magellanic Cloud (Vijh et al. 2009) pre-main-sequence disk systems. While ESHC and ELHC stars were initially identified as potential pre-main-sequence objects based on their optical photometric variability, the stability of their IR brightness has not been well constrained to date. We limit our investigation of IR photometric variability to sources which have archival ground-based near-IR observations (e.g., de Wit et al. 2005) with adequate spatial resolution to mitigate the source confusion that likely affects at least some of the quoted 2MASS measurements of these stars. Thus, we do not use our IR photometry of ESHC sources to identify variability in these objects.

We subtracted the archival photometry for 15 ELHC stars from our measurements given in Table 4 and compile the results in Table 5. Although most of the ELHC stars exhibit stable absolute photometry at the level of our uncertainties, ELHC 3, 5, 7, 12, 18, and 21 exhibit photometric variability at the >3σ level in at least one filter. Three sources exhibit strong variability in all three IR filters, ELHC 21 (∼0.5 mag difference), ELHC 7 (∼1 mag difference), and ELHC 12 (∼1 mag difference). ELHC 18 was reported to have clear source confusion issues in Section 3.1; hence, we limit our interpretation of the origin of variability originating from this set of objects. We do note a hint of a trend in our photometry as compared to de Wit et al. (2005) in Table 5, such that we tend to report more ELHCs to be brighter (11 of 15) than fainter (4 of 15) as compared to the archival values. Our standard star observations do match extremely well with published values (Table 3); moreover, the majority of our targets do not exhibit variability at the 3σ level. We are therefore confident that we have not introduced a spurious signal in our data reduction, and our reported photometry is robust.

ESHC and ELHC stars are known to exhibit two different types of photometric color variability, those that become bluer when fainter (i.e., exhibit a negative color gradient) and those that become redder when fainter (i.e., exhibit a positive color gradient). The bluer when fainter phenomenon was originally suggested to arise from nebular scattered light becoming a more dominant component when the central star itself exhibited photometric fading (Lamers et al. 1999; de Wit et al. 2002). However de Wit et al. (2005) noted that the bluer when fainter phenomenon could also arise from simple variable continuum bf-ff radiation from classical Be stars, and that this was a simpler explanation than that presented by others. The redder when fainter stars have been suggested to be young pre-main-sequence stars that are variably attenuated by inhomogeneous dust clouds in their circumstellar matter (Lamers et al. 1999; de Wit et al. 2002, 2005).

Most of the ELHC stars we identify as having 3σ near-IR variability exhibit a gray color variation (ELHC 7,12,21; Table 5). ELHC 10 only exhibits 3σ variability in H band, but if one considers its 2σ J- and Ks-band variations to be real then it also exhibits a gray color variation. The only star that exhibits a significant color change is ELHC 3 (Table 5), which exhibited significantly less Ks-band flux compared to other filters in our data when compared with archival observations (de Wit et al. 2005).

We construct an IR two-color diagram for 27 of the 28 ESHC and ELHC stars for which we report new IR photometry in this paper, shown in Figure 2. For comparison, we also plot the 2MASS IR colors of sources confirmed to be classical Be stars in the SMC and LMC by Wisniewski et al. (2007a), i.e., sources which exhibited polarization Balmer jumps or electron scattering signatures. A selection of Galactic HAeBe stars from Hillenbrand et al. (1992) is also shown. We find that the near-IR properties of SMC/LMC classical Be stars share many similarities with SMC/LMC candidate HAeBe stars and even some overlap with Galactic HAeBe stars, a phenomenon which has been similarly noted by de Wit et al. (2003, 2005). This near-IR degeneracy is likely influenced by the lower metallicity content of the SMC/LMC, which might reduce the IR excess of HAeBe stars (de Wit et al. 2003), and also lead to potentially stronger free–free IR excesses of classical Be stars in the SMC/LMC as compared to our Galaxy (Bonanos et al. 2009, 2010). Though our sample size is much smaller, we see no distinct evidence in Figure 2 that the IR excesses of the ESHCs are systematically larger than those of the ELHCs, which might be expected if the sources were classical Be stars (Bonanos et al. 2010).

Zoom In Zoom Out Reset image size

We analyzed the basic properties of the Hα and Hβ lines in our continuum-normalized spectroscopic data, taken nearly contemporaneously with our near-IR photometry, to help assess the IR variability and colors we observed. The FWHM and EW of Hα, ratio of the peak-to-continuum strength of Hα, and basic line profile shapes of Hα and Hβ are detailed in Table 6, while Hα line profiles are plotted in Figure 3. For reference, nebular features in non-sky-subtracted imagery exhibited an FWHM of ∼3 Å. When available, archival spectroscopic measurements for some of these sources are also compiled in Table 3. We comment on individual sources below.

• 1.  
  ELHC 1, 11; ESHC 3, 6, 7. Our new observations of ELHC 1, ELHC 11, ESHC 6, and ELHC 7 all exhibit strong emission in Hα, with EWs ranging from −6.8 (ESHC 3) to −34.3 (ELHC 1), FWHMs ranging from 4.0 Å (ESHC 3) to 9.1 Å (ESHC 6), and peak-to-continuum ratios ranging from 2.4 (ESHC 3) to 5.7 (ELHC 1). With the exception of ESHC 3, whose Hβ emission merely fills in the underlying photospheric absorption line, each source also exhibits emission at Hβ which rises above the continuum level. ELHC 11 was noted in Section 3.1 to suffer from source confusion owing to a star ∼15 away. Given the 20 fiber width of the Hydra spectrograph, we likely captured signal from both ELHC 11 and its nearby neighbor. Its emission-line characteristics listed in Table 6 should be viewed with extreme caution.
• 2.  
  ESHC 4, 5. Our new observations of ESHC 4 and ESHC 5 show each source exhibits double-peaked Hα emission with peak-to-peak separations of ∼4.3 Å (ESHC 4) and ∼3.3 Å (ESHC 5). While ESHC 4 exhibits evidence of weak Hβ emission which does not emerge above the continuum level, ESHC 5 exhibits no obvious signatures of emission at Hβ.
• 3.  
  ELHC 3. Although earlier epoch (1994 January) spectroscopy by Lamers et al. (1999) exhibits an Hα EW and FWHM a factor of two stronger than that found in our 2005 January 21 data, we observe a stronger ratio of the peak-to-continuum Hα emission in 2005 (see, e.g., Table 6). The resolution of the 1994 data at Hα (R ∼ 2000) is similar to the resolution of our data (R ∼ 2200), indicating these variations are not instrumental artifacts. These results indicate that the H i line core flux may have increased from 1994 to 2005, while the wing flux decreased.
• 4.  
  ELHC 4. The ratio of the Hα peak-to-continuum emission is much stronger in our 2005 January 21 data (4.4) than that we estimated for the 1994 August data (∼2.2) of Lamers et al. (1999). The resolution of the 1994 data at Hα (R ∼ 1300) is somewhat less than our data (R ∼ 2200), which could contribute to some of the lower peak-to-continuum ratio seen in the earlier epoch data. Moreover, ELHC 4 was noted in Section 3.1 to suffer from source confusion owing to a star ∼12 away. We likely captured signal from both ELHC 4 and its nearby neighbor given the 20 fiber width of the Hydra spectrograph; hence, emission-line characteristics listed in Table 6 should be viewed with extreme caution.
• 5.  
  ELHC 5. Comparison of our 2005 January 21 data and 1994 August data obtained by Lamers et al. (1999) suggests that measured H i line differences (Table 6) predominantly occurred the line wings.
• 6.  
  ELHC 6. The Hα EW strength, FWHM, and peak-to-continuum ratio between our 2001 January 21 data and archival spectral from 1994 August (Lamers et al. 1999) clearly differ. Both data were obtained at similar resolutions (R ∼ 2000 versus ∼2200), hence resolution likely plays little role in the differences observed. We caution however that ELHC 6 suffers from clear evidence of source confusion (Section 3.1); hence, emission-line characteristics listed in Table 6 should be viewed with extreme caution.
• 7.  
  ELHC 7. We observe a double-peaked Hα emission-line profile which has a similar EW and FWHM to that reported in 1994 August spectra (Lamers et al. 1999). Lamers et al. (1999) do not report the Hα profile to be double-peaked in their data; however, their lower resolution (R ∼ 1300) might contribute to this non-detection especially given the peak-to-peak separation we see in our data (∼2.3 Å).
• 8.  
  ELHC 8. de Wit et al. (2005) reported a similar Hα EW in their low-resolution (R ∼ 400) spectrum obtained on 2000 January 8 as that measured in our 2005 January 21 data. The Hα FWHM in the 2000 data (13.5 Å) appears to be below the resolving power of the spectrograph used by de Wit et al. (2005); hence, comparisons of the change in line width are not possible.
• 9.  
  ELHC 12. de Wit et al. (2005) obtained a low-resolution (R ∼ 450) spectrum of ELHC 12 on 1998 January 6, shortly before the near-IR photometry they obtained on 1998 January 19–20. Although they did not report Hα EW or FWHM measurements for these data, we did estimate the ratio of the peak-to-continuum emission of their Hα profile to be ∼3.1, which is similar to the value of 3.8 we derived from our 2005 January 21 data, especially given the difference in the resolution of the two different data sets. Based on these available data, we find no clear evidence that the Hα spectroscopic properties of ELHC 12 were substantially different between the 1998 January epoch observation and the 2005 January epoch observation.
• 10.  
  ELHC 13. Our 2005 January 21 spectrum exhibited a smaller Hα EW than that observed in a low-resolution (R ∼ 400) spectra obtained on 2000 March 9 by de Wit et al. (2005). The Hα FWHM in the 2000 data (14.4 Å) appears to be below the resolving power of the spectrograph de Wit et al. (2005) used; hence, comparisons of the change in line width are not feasible and the observed differences in the Hα peak-to-continuum ratio might be influenced by the large difference in resolution of the two data sets.
• 11.  
  ELHC 16. de Wit et al. (2005) obtained a low-resolution (R ∼ 450) spectrum on 1998 January 6 and reported weak Hα emission, which filled in the underlying photospheric absorption line. By contrast, we observe emission in Hα which peaks 1.9× above the continuum as well as weak emission in Hβ which partially fills in the absorption line in our 2005 January 21 data. The Hα FWHM of our observation (3 Å) is not broader than our instrumental resolution; however, we see no strong evidence of residual nebular emission in our spectra following sky subtraction, so we suggest we are seeing emission which is circumstellar in origin.
• 12.  
  ELHC 19. de Wit et al. (2005) observed this star at low resolution (R ∼ 450) on 1998 January 6. While they did not tabulate Hα line strengths or widths, we did estimate the ratio of the peak-to-continuum emission in their Hα profile to be ∼2, which was substantially smaller than the ratio of 6.3 we observed in our 2005 January 21 data. Although the different resolution of the data likely contributes to some of the observed differences, these results suggest ELHC 19 exhibited significantly stronger emission in 2005 as compared to 1998.
• 13.  
  ELHC 20. We only detected weak evidence of emission in the core of the Hα photospheric absorption line of ELHC 20 in our 2005 January 21 data, which qualitatively matches the filled-in absorption line observed in R ∼ 450 spectra by de Wit et al. (2005) on 1998 January 6.
• 14.  
  ELHC 21. de Wit et al. (2005) obtained an R ∼ 1000 spectrum of ELHC 21 on 1998 January 8 (nearby the epoch of their near-IR photometry of the source) and found Hα to have a slightly larger FWHM than observed in our 2005 January 21 spectrum, although the line's peak-to-continuum line ratio were similar in both epochs. The peak-to-peak separation in the double-peaked emission-line profile we observe, ∼6.7 Å, is larger than the resolution of the 1998 epoch data; hence, it is difficult to assess whether the morphology of this line has evolved as a function of time. Based on these available data, we find no clear evidence that the Hα spectroscopic properties of ELHC 21 were substantially different between the 1998 January epoch observation and the 2005 January epoch observation.
• 15.  
  ESHC 1. The Hα EW, FWHM, and peak-to-continuum strength were all larger in the spectrum of ESHC 1 obtained by Beaulieu et al. (2001) on 1998 August 18 than in our 2005 January 21 data. The overall emission strength of ESHC 1 therefore seems to have decreased from 1998 to 2005.
• 16.  
  ESHC 2. We observed ESHC 2 to have a weaker Hα EW and FWHM than seen in archival observations made on 1998 August 18 by Beaulieu et al. (2001). The peak-to-continuum line strength ratio in each data was similar however, suggesting that the line profile morphology of ESHC 2 has slightly changed between these two epochs, with emission from the line wings decreasing from 1998 to 2005.

Zoom In Zoom Out Reset image size

ELHC 6 (this work) and ELHC 18 (this work and de Wit et al. 2005) have now each been confirmed to be comprised of two sources; hence, we suggest that previous efforts to identify and characterize the optical/IR variability and IR excess of these sources should be treated with some caution. We found that ELHC 4 and ELHC 11 had secondary sources located 12–15 from each star; hence, previous photometric observations which had poorer spatial resolution (or seeing) than these values may suffer from contamination. We find no support for possible source confusion for either ELHC 5 or ELHC 8, which contrasts with the suggestions of previous works (de Wit et al. 2005; Wisniewski & Bjorkman 2006), and therefore we suggest that they continue to be classified as candidate HAeBe targets.

We detected evidence of photometric variability at the 3σ level in one IR filter for ELHC 3 (Ks band) and ELHC 5 (J band, though H and Ks filters also exhibit variability at the 2σ level). ELHC 21, ELHC 7, and ELHC 12 exhibited variability in all three filters. The large (∼1 mag) IR photometric brightening observed in both ELHC 7 and ELHC 12 is at the extreme end of IR photometric variability observed in classical Be stars. ELHC 7 exhibited a very large Hα emission line in both 1994 and 2005; hence, it is plausible that the system also exhibited significant Hα emission at the earlier epoch of IR photometric observations (1998 January). Moreover, ELHC 7 has previously been classified as a UX Ori variable (de Wit et al. 2005). Hence, the IR variability we observe is completely consistent with the classification of this star as an HAeBe star. ELHC 12 also exhibited an Hα emission line with a very similar peak-to-continuum ratio near the epoch of its 1998 IR photometric observation (∼3.1) as near the epoch of its 2005 IR photometric observation (3.8). Although we do not have information about the strength of the EW and FWHM of Hα for the archival 1998 data, we suggest that there is no clear evidence to indicate that the Hα emission line was substantially different between the two epochs of observations. We are not aware of any classical Be star model which would predict the magnitude of near-IR flux increase we see for ELHC 12, and corresponding lack of strong enhancement in the Hα profile, owing to a disk building event. We therefore suggest that the IR photometric variability we observe for ELHC 12 might indicate that the system is a UX Ori type HAeBe star.

The IR brightening we observe for ELHC 21, ∼0.5 mag, is within the range of typical variability observed at near-IR wavelengths for both HAeBe stars and classical Be stars. If ELHC 21 is a classical Be star, this level of variability could indicate a factor of ∼2–3 change in the density of the disk (Dougherty & Taylor 1994). Comparison of the ratio of the peak-to-continuum line strength of the Hα spectra obtained near the time of the 1998 epoch and 2005 epoch IR photometry reveals little evidence of substantial changes. The region of a Be star's disk responsible for producing Hα emission is known to be predominantly located at larger radii than that which produces near-IR emission (Wisniewski et al. 2007b; Carciofi 2011). Thus the ∼0.5 mag IR brightening we observe in our 2005 data could indicate that a significant, recent (inner) disk building event occurred, which had not yet manifested itself in diagnostics more sensitive to the outer disk region (Hα spectroscopy). Alternatively, we cannot rule out that the moderate-scale variability we have observed could also arise from circumstellar dust clouds which attenuate light from the central source (see, e.g., de Wit et al. 2005), and therefore indicate that the system is an HAeBe star. Higher cadence, contemporaneous optical and IR observations of ELHC 21 should be pursued to differentiate between these two scenarios.

In contrast to the observed optical variability of ELHC stars, which exhibits either bluer when fainter (negative color gradient) or redder when fainter (positive color gradient) colors, most of the IR variability we observe is gray. The color effects of nebular scattering, dust reddening, and variable bf-ff emission should all be small at near-IR wavelengths, so it is perhaps not unexpected that the variability we observe is gray. We note that the sources which exhibit the greatest IR variability (ELHC 7, 12, and 21) were found to either have positive optical color gradients (redder when fainter) or gray optical color gradients (de Wit et al. 2005). Variable dust cloud obscuration has been suggested as the most likely explanation for this optical behavior (de Wit et al. 2005), which is consistent with our previous interpretation for the origin of the IR variability we observe in these sources. Two ELHC stars that exhibit bluer when fainter optical colors were also identified to be variable in one IR filter in our data (ELHC 3 and ELHC 5); however, the single bandpass IR variability detection precludes us from drawing strong conclusions on the color variability of these systems.

The new Hα spectroscopy presented in this paper enables us to investigate the spectroscopic variability of numerous ESHC and ELHC stars which have earlier epoch spectroscopic data in the literature. By analyzing the differences in the EW, FWHM, and peak-to-continuum ratio, we have been able to identify evidence of strengthening line core and/or wing emission in ELHC 3, ELHC 16, and ELHC 19 and decreasing line core and/or wing emission in ELHC 5, ELHC 13, ESHC 1, and ESHC 2. We detect modest or minimal changes in the Hα emission strength in ELHC 7, ELHC 8, ELHC 20, and ELHC 21. As the population of ESHC and ELHC stars were initially identified on the basis of their observed optical photometric variability, it is not unexpected that we have detected a variety of changes in the appearance of each star's Hα line emission. The emission strength and line profiles of both HAeBe stars (Pogodin 1994; Reipurth et al. 1996; Vieira et al. 1999) and classical Be stars (Porter & Rivinius 2003; Stefl et al. 2009; Wisniewski et al. 2010) are known to exhibit variability owing to changes in their circumstellar environments; hence, the presence of this variability alone is not a good discriminant between the two types of stars.

We have reported near-IR photometry for ESHC 2 and ESHC 3 (Table 4). Both the (J − H) and (H − Ks) colors are close to zero or marginally blue. de Wit et al. (2003) adopted an optically thin dust scaling algorithm to estimate how the colors of Galactic HAe and HBe stars would be altered if these sources were in the dust poor SMC, and in Figure 17 of their paper they show a clear degeneracy between the IR colors of classical Be stars and HAe and HBe stars. This degeneracy incorporates the range of (J − H) and (H − Ks) colors we report for ESHC 2 and ESHC 3 (see Figure 2); hence, the colors of our new observations do not permit us to differentiate the nature of these objects. Moreover, as no pre-existing IR photometry is available, we cannot measure the variability (or lack thereof) of these sources for use as a discriminant.

We have also reported Hα spectroscopy for ELHC 1 and ESHC 3, 4, 5, 6, and 7 and tabulated Hα line strengths and properties for ELHC 11, 12, 16, 19, 20, and 21. All sources exhibited Hα in emission, although line strengths varied greatly from sources which exhibited small hints of emission in the core of their underlying photospheric absorption lines (ELHC 20) to strong emission with EWs between −33 and –34 (ELHC 1, ESHC 6). The moderate resolution (R ∼ 2200) of our data enabled us to identify evidence of double-peaked line profiles in ELHC 21, ESHC 4, and ESHC 5. These new data do not provide definitive differentiation of the classification of ESHC and ELHC stars as HAeBe versus classical Be stars. However, as seen in Section 4.2, having quality near-contemporaneous optical spectroscopic and IR photometric observations in the literature can provide invaluable insight into the time evolution and hence origin of these objects.