Talk:BP Psc

BP Psc is a YSO with extended optical jets
The following is the text of an email sent on 2007-01-18 to, , , and :

"Hi guys,

We observed BP Psc twice (2006-06-29, 2006-07-13). Now a colleague here has just told me by chance that they've discovered a huge optical jet coming from it. They have a DDT program running right now at Gemini, and they are planning a Nature or Science paper (draft circulating).

Our CCF of 2006-07-13 shows a bump on the red side - a possible faint companion. This bump is almost gone on 2006-06-29, and the RV has shifted redwards. This does indicate that this is indeed a binary, and with the newly discovered jet, we have something very interesting here! You T-Tauri/YSO guys may want to check it out.

However, to be really strong in the binary case, we need one more spectrum. The star is setting shortly after sunset, but it's not too late to catch it yet! If one of you, while in La Silla, can manage to take a HARPS spectrum, we could have a Paper Two in the makings...

Cheers, $\pi$ 16:24, 18 January 2007 (CLST)

Archival data
I uploaded to the wiki the only EFOSC2 spectrum available in the ESO archive, taken in june 2003. Access it [[Media:EFOSC2 archive BP Psc.fits.gz|here]]. Lots of fringing to the red, Halpha in emission. It might help. It has not been observed very often, actually. No image found at all in the ESO archive! Here is the output of the MAST archive website. -- 10:15, 19 January 2007 (CLST)

"old" FEROS observations exist
BP Psc was part of the SACY sample (Torres et al. 2006), which includes among the authors our very own along with Claudio Melo. They find that the star is SB1 based only on a rms spread of 2.0km/s around a measured RV = -5.8km/s (FEROS spectra). I don't have the dates of their 11 spectra, but I suspect that the variation is mostly measurement errors. The star is not mentioned in the paper directly, and hence not found by Simbad or ADS, but it's in their catalog. Note that out RV measurement gives around -14 km/s! -- π 18:26, 19 January 2007 (CLST)

Update on the old FEROS data
has asked Carlos Torres about their old RV data on BP Psc. This is the reply:

We have 11 coudé and 3 FEROS observations. It is a SB1 (I do not remenmber the period now, P=3.2d? We must ask Germano.) and the RV in the SACY is the systemic velocity (-5.8 +/- 2.0) The solution is not very good and the amplitude is low. Our extreme values are +8.8 and -15.0. Then, the HARPS value is within our observed value and it can help in the orbit determination. The low quality of our orbit is one of the reasons we have not write a paper... (And the English etc). But there is another very important reason - I am not able to classify this star! This is a very strange star. Almost sure is a giant star - or spectral type is G9III and the Li is weak, 90mA (it has a suspect spread in the determinations). The variable Halpha (from 10 to 27A) is strong for a RS CVn and in some spectra there are forbidden lines. If it would be a RS CVn (as ER Eri!!!), it will be a special case. The position on the sky is, I you know, unusual for young star... An optical jet would make the situation more complicated.

Note that our two reliable RV measurements taken two weeks apart are -12.9 and -13.5 respectively, so it may seem like it - coincidentally - have been observed near but not exactly at the same orbital phase (differences mainly in H-alpha. See below). Interesting that Torres believes it is SB1, while Zuckerman et al thinks it is not... -- π15:38, 2 November 2007 (CLST)


 * Do we know which lines they used for the RV measurements? Absorption lines or the H-alpha emission? -- 13:15, 4 November 2007 (CLST)


 * They used absorption lines. RV was calculated by cross-correlation with a spectrum of tau Cet. -- π 17:29, 5 November 2007 (CLST)

New HARPS observations?
As requested by Gaspare, and confirmed by Ivo, an attempt will be made to observe BP Psc again with HARPS on the night of 2007-01-22. Star was eventually observed on 2007-01-24. See discusion below -- π

HARPS spectrum of 2007-01-24
BP Psc was observed again on 2007-01-24. It shows strong and broad doube-peaked H-alpha emission, single peaked oxygen 6300 emission, and strong and quite sharp Na D emision. All these lines are also present in the previous spectra. Lots of other emission lines in varying degrees of strangeness can be seen too. The stellar spectrum cannot be seen due to poor S/N - the RV of -21.67 km/s is due to the solar spectrum (it was taken in twilight). is looking into it a bit and will add his thoughts later. -- π 13:38, 29 October 2007 (CLST)

Na D lines: so strong and narrow! BUT - they are in this case telluric! And the narrow absorption lines which give rise the the -21.67 km/s are in fact solar. Checking the sky fiber, you see exactly the same lines, so the CCF is just the solar spectrum being picked up. It seems that we are not seeing the stellar absorption line spectrum due to inferior S/N. The emission lines are currently being investigated by and. -- π 16:36, 1 November 2007 (CLST)

IUE spectrum
IUE observed BP Psc on 29Jun1994. The low resolution spectrum is available and I put it here as a figure. No obvious p-Cyg profile but strong MgII, indicating "activity". -- Alessandro 14:06, 30 October 2007 (CLST)

Radial velocity and line profiles
I have checked a few lines to cross-check the radial velocity variation which was observed from the CCF. The spectra have been rescaled and arbitrarily shifted. All the lines have been measured with a gaussian fit, although sometimes their shape is far from gaussian. For all the lines, in the plots, I show a vertical line which refers to the first measurment. Tipically, the first two spectra agree pretty well, while there is some difference with the fourth.

H alpha and N


The He-line which happens to fall in the Halpha region has a very non-gaussian profile (please, note the red "wing"). In the third epoch (although barely visible) it seems to be slightly different (has it developed another peak?).


 * Just a comment: This line at 6583 is not HeI, but [NII]. The "normal" HeI line is at 6678 and in this system only in absorption.
 * Another comment: The line at 6678 referred to above is a strong line of Fe I at 6677.987 - G stars don't have strong helium lines. -- π 15:22, 2 November 2007 (CLST)

The Halpha can be due either to two emission lines (a blue-shifted one and a red-shifted one) or to a disk or to an emission line with an absorption in the center. Note that the red peak changes its shape and intensity between the first and the second epoch. The absorption is very deep (and shifted towards the red?) in the last spectrum.

OI 6300-6364


The [OI] doublet at 6300-6364 shows no absorption (more or less, obviously). The emission is slightly asymmetric. The narrow emission which is seen in the third spectrum comes from the sky (as noticed by Thomas). The radial velocity has been derived using only the "broad" component.

NaD


The NaD lines show a complex profile. Whatever the cause of it, for the radial velocity, I have measured the position of the high peak.


 * Notice: The Na D sharp emission in the third spectrum is telluric. -- π 16:39, 1 November 2007 (CLST)

Radial velocity change


The radial velodities have been considered with respect to the first measure for each line. I also used an absorption line at 4957AA which seems to be present in all spectra. The best explanation that I manage to come with, is that the different things which I have measured are related with different physical parts of this object. I hope to have some more deep and clever ideas about it soon. -- Alessandro 16:24, 30 October 2007 (CLST)


 * Notice: The CCF of the last spectrum (2007-01-24) picked up only the solar spectrum from the sky background.  Both the CCF, the NaD and the 4957AA line are of Solar origin in the last spectrum! -- π 16:55, 1 November 2007 (CLST)

H-alpha radial velocities




I measured the radial velocities of the H-alpha line using the double Gaussion method (Shafter 1983): two Gaussians are fitted to the blue and red line wing. I used a FWHM of 3A and separations of the two Gaussians between 4A and 20A. Notice the stable region in the plot on the right side for separations between 7 and 16A. This fits with the H-alpha line having a width of about 14A.

Result: The H-alpha lines move with considerable velocity compared to the absorption lines or the other emission lines. This would hint towards a binary and some kind of accretion event. Note that the H-alpha profile represents a nice disc-profile at high inclination. However, assuming that we see the system at high inclination, the width of the line (14A) is rather small.

Since the velocity changes of the absortion lines are much smaller, we either by chance have observed them at a similar velocity (which can easily happen with only two measurements) or we have a very high mass ratio, in the sense that the donor has a possibly 40 times the mass of the accreting object. If the system was in contact, such a mass ratio would be kind of explosive and these systems are extremely short-lived. I don't think, they have ever been observed, soi this scenario is rather unlikely. They might however be detached and accretion happening from wind only (kind of a symbiotic star). Note that a white dwarf (WD) would probably not be visible in an optical spectrum if the donor is of early G type. However, its presence might explain the strong FUV emission.

-- Linda 10:55, 2 November 2007 (CLST)

The new spectra yield intermediate radial velocities. See the plot with all RVs against time. Double Gaussian method was used with a separation of 11A.

-- Linda 11:25, 6 November 2007 (CLST)


 * About the disk orientation and the width of H-alpha emission: Zuckerman and Seong confirms (priv.comm.) that they see a near edge-on disk in AO imaging.  -- π 21:52, 13 November 2007 (CLST)

Radial velocities from various lines


For an overview, radial velocities from several sources all in one plot. The error bars of H-alpha and CaII are less than 1km/s, the forbidden lines have a lower S/N, so the errors are probably higher. CCF is of course the best with 1-2m/s. -- Linda 16:20, 6 November 2007 (CLST)


 * just to point out that the plot is being updated as data comes in... The plot as of December 2007 looks strange for several reasons:
 * * CCF RV change is the biggest so far, accompanied by a substantial change in the line appearance - see discussion of CCFs elsewhere on this page.
 * * Large excursion in the Ca II derived RV compared to earlier. I did not check the S/N, but it seems without much doubt that Ca II emission and the nebular lines and H-alpha come from three different regions, which are all distinct from the photosphere!
 * -- π, 17:33, 21 December 2007 (CLST)

Na D lines, June-July 2006
These images show the region around the Na D lines. Line positions are indicated with expected depths for a G2 star, shifted to the CCF RV of the spectrum (left) or with a synthetic spectrum fit (right). The spectra are puzzling for several reasons:
 * The stellar sodium lines look like they are actully at the wrong wavelength (the large tick-marks) if this is a heavily rotationally broadened line. The right-hand plot shows such a model with vsini=120km/s, which tells us already that this can't be the case. Rather, we have a case of either:
 * additional fill-in of both lines, redshifted with respect to the star and the strong emission line.
 * additional absorption component, which apparently is not seen in any other lines; how can we have extra Na D lines without an extra photosphere? From the disk? wind?
 * The emission lines look single-peaked, slightly blue-shifted from the CCF RV. Where do this emission come from?
 * On 2006-06-29 the stellar sodium lines look much more flat-bottomed than two weeks later. Is this emision fill-in of some sort? From where?
 * The sharp absorptions on top of the emission is what is usually associated with cool interstellar material. Is it just a coincidence that the RV of the absorbing clouds correspond to the emission? or are they related? perhaps even the same?

The lines become even more puzzling, in 2007, so read on... -- π 18:16, 1 November 2007 (CLST) -- Revised by π on 20:22, 7 December 2007 (CLST)

Na D lines, November 2007
The Na D region of the first spectrum from November 2007 (left) show a third kind of morphology compared to the previous two high-S/N spectra. This time, it fits very well with a single G2 spectral template, broadened to 42km/s, which is about the vsini suggested by the CCF. As before, the emission peaks are slightly blue-shifted with respect to the line center. 20 days later (right) we are starting to see the Na D profile change again, looking as if some additional absorption is kicking in! The same fit (red line) is used in both plots.

I checked the sky-fiber spectrum, and the weak sky emission lines are not coincident with the emission.

Note that the Na D region presents one of the prime reasons to use HARPS: no other instrument can resolve the structure of the absorptions and at the same time provide accurate RV's. -- π 22:15, 5 November 2007 (CLST) --  Revised by π on 20:22, 7 December 2007 (CLST)

Na D lines and variable vsini
I've uploaded new versions of some of the Na D plots, including a model solar spectrum (red lines), scaled to [Fe/H]=1.0 and broadened to fit the observed spectrum. As you can see, the November 2007 spectra are fit well by a vsini=42km/s G2 template, while the previous ones are much, much broader. It would seem that the star has spun down, which is ridiculous:
 * My first idea was that the last spectra are affected by a partial eclipse of the stellar disk, leading to lower apparent vsini. That's actually a very cool thing!  That explanaton fits well with the H-alpha RVs, so would be yet another hint that the star is really a binary with an unseen and very light-weight accreting companion. However, the 120+ km/s rotational velocity required is too high, and the weaker lines elsewhere in the spectrum show no such variations. This must be something particular to the Na D lines...
 * Another suggestion is that it could be instead a contribution from a disk around the star itself, acting like kind of an extra photosphere. Or a puff of a sodium cloud at high temperature and density akin to stellar photospheres somewhere in the vicinity... I'm out of suggestions now... DDT text has been updated as well. -- π 17:25, 19 November 2007 (CLST) -- Revised by π on 20:22, 7 December 2007 (CLST)

Forbidden lines
I have found a number of forbidden lines, which would indicate some kind on nebula: 3969[NeIII], 4068[SII], 6300[OI], 6583[NII], 6716[SII], 6731[SII]. There are probably more, I only checked out some. If it's indeed a binary with a compact source (WD?), this might be an indication for a planetary nebula. However, typical PN lines like 4959[OIII] and5007[OIII] are not present (maybe there is a hint of a line at 5007A).

[OI] and [NII] were already plotted above so just find the [SII] lines below.



I checked the line profiles and radial velocity changes of these lines:

FWHM of all lines (apart from the 3969 one) is around 1A, the radial velocity changes between the first two spectra (third is too noisy) vary between -1.5 and 1.5km/s, but I believe this is the error of the measurement (just fitting a Gaussian). On average, there is no variation of the radial velocity. Note that this is kind of expected from a nebula.

The line at 3969A is rather broad (2.5A), and shows a larger variation between the two spectra (6km/s). It is possibly a blend with H-epsilon. Maybe there is no NeIII present at all, and it's just a bizarre line profile of H-epsilon. Thomas just mentioned it could be chromospheric Ca. We checked for the K-line which is also in emission, so that's probably it.

-- 11:45, 2 November 2007 (CLST)

Same behaviour in the new spectra. -- 12:15, 6 November 2007 (CLST)

CaII H+K lines
Following the remarks before, I checked the radial velocity changes in the CaII emission lines, which might be chromospheric and thus trace the movement of the star. I find:

CaII 3933.663

2006-06-30: -5.563 km/s; 2006-07-14: -10.141km/s

CaII 3968.469:

2006-06-30: -18.056 km/s; 2006-07-14.: -24.862 km/s

The third epoch was too noisy to show any emission.

-- 14:55, 2 November 2007 (CLST)

New spectra yield velocities of -6.64 km/s for CaII 3933.663 and -39.14 km/s for CaII 3968.469. Note that the red side of the latter line is truncated by an absorption feature, so this velocity is not very reliable.

All in all, the CaII lines seem to follow the H-alpha emission rather than the absorption lines.

- 12:35, 6 November 2007 (CLST)

Spectra from 2007-11-26: -28.51 km/s for CaII 3933.663 and -41.48km/s for CaII 3968.469 (truncated).

- 15:46, 21 December 2007 (CLST)

Findings of Zuckerman and Song
Via a former colleague at Gemini, π heard about the work led by Ben Zuckerman on BP Psc. Thomas also saw beautiful images of the optical jet that these authors found, but have no details about extent etc. The following is an email from Inseok Song, a member of the group working with Zuckerman:

A long ApJ paper (instead of one Nature paper with a bunch of smaller ApJ papers) was submitted several months ago. In the submitted paper, we argued that BP Psc is an unusual adolescent star (due to small Li) with a delayed accretion (from optical jets). The referee report was OK and didn't require much changes. Then, we discovered more evidences supporting BP Psc being a post-MS star. Because of this drastic change after the 1st referee report, it took us awhile to get a revised version. Now the revised version is ready for resubmission. The optical jets run NNE and SSW, and the NNW jet is more prominent. We obtained near-IR AO images from which we saw a near edge-on disk (no photosphere). So, it is possible that radial velocity changes may be related to scattered light variation of clumpy accretion disk instead of an unseen companion. I am happy to be part of your follow-up study on BP Psc and Ben Zuckerman will be happy to be involved also. He is the lead author of the big ApJ paper.

Note that they do not believe in the presence of a companion. It will be intersting to read the paper once it appears on astro-ph! -- π 15:10, 2 November 2007 (CLST)

Notice: This paper is not out yet, not even on astro-ph. It has just been submitted, or is about to be submitted. Please keep this information within VSOP! -- π 14:53, 6 November 2007 (CLST)


 * I have one of the coauthor (T. Forveille) in the office beside mine. I've asked, but I don't have details on why they don't think it is a binary. -- 10:38, 8 December 2007 (CLST)

Chandra/ACIS-S3 observations planned for 2008 October
Proposal Title: 	 BP Psc: Chandra Takes a Close Look at a Pre-Main Sequence Star-Disk-Jet System Proposal Number: 	09200153 Principal Investigator: 	Joel Kastner Abstract: Observational studies of disk-jet systems in pre-main sequence (PMS) stellar evolution provide insight into a diverse range of phenomena, including jet launching and collimation, protostellar accretion, and the formation and evolution of planetary systems. We propose to obtain Chandra/ACIS-S3 imaging spectroscopy of BP Psc, the only known actively accreting and jet-driving PMS star within 100 pc of Earth. These observations will take full advantage of this unparalleled opportunity to probe star-disk-jet interactions within ~30 AU of a PMS star, via X-rays.

NB: the observations are planned for:

Start Date: 	2008-10-24 00:00:00

-- -- 17:25, 5 November 2007 (CLST)


 * Joel Kastner is a visiting scientist in my lab. Just to say. I can discut this Chandra stuff with him, if necessary. -- 10:38, 8 December 2007 (CLST)

New HARPS spectra on 2007-11-05
A new spectrum, 1800sec exposure with good S/N was obtained on 2007-11-05. The measured RV is -12.39, i.e. very close to the two previous values. The FWHM is 44.18 km/s, i.e. practically the same as the last two ones. The forbidden (nebular) lines are practically unchanged, and specificaly are at the same RV as last year, which is kinda reassuring really. I'll leave it to to measure the H-alpha velocity.

Chance orbit alignmet three times in a row is a bit too much, so we actually took another one three hours later (only 15 min). Even in case of high mass ratio between the components, at least the emission from the disk should have moved. The S/N is lower, but the stellar spectrum is clearly visible with RV=-12.34, FWHM=44.45. Emission lines look unchanged, but I'll leave the measurements to and/or.

The emission lines could come from activity (both Na D and Ca II H+K) in a medium-high rotation star, which would imply youth. Actually it does have a distinct lithium line, contrary to previous claims(?) See plots on the right. we must measure the EW of this line, possibly from a combined spectrum.

There are still some interesting questions:


 * How can the Na D lines change their shape so much (see entry above about the Na D lines)?
 * What's with the large amplitude of the Torres et al orbit?? WTF!
 * So, if it's not a binary, then what is it? It looks like a PN shell with a dust cocoon with a PMS inside...

π 01:16, 6 November 2007 (CLST)

CCF bisector variations
I've checked the CCF and the bisector of the CCF for all four good-S/N epochs (see plots). The overall shape is fairly constant, or at least similar. The only major excursion in shape is at the top of the orange bisector curve, which corresponds to the faint "bump" that can be seen in the CCF on 2006-07-13. Curiously, this point is also where we measure the largest negative RV of H-alpha, but also the largest negative CCF RV... besides, the "bump" appears on the red side, so that does not work well with the assumption that the H-alpha emission is associated with a faint companion.

Note that the RV changes correspond very well to the change in shape of the bisector i.e. of the line. Especially the top of the line seem very consistently at the same RV, while the lower parts fluctuate quite significantly. The last November point is deviating though. It looks like a real shift this time, since the shape of the bisector is similar (although with higher noise and wiggles), but the whole line has been shifted bluewards over the course of 20 days. Such shift can perhaps be explained without a companion if the light has been reflected off part of the surrounding disk, although I can't see how it would produce a blueshift... Note also that the CCF is shallower here (veiling?) -- π 12:28, 6 November 2007 (CLST) -- revised by π on 16:30, 18 December 2007 (CLST)

Binary or not?
Thomas and I discussed the possibility that it is actually a single star with an accretion disc. Here are the arguments why this is unlikely:

The width of the H-alpha line relates to an inner disc velocity of about 650km/s. Using this value, we get for a star of 1 solar mass a distance of the inner disc from the centre of the star of 1/2 solar radius. So this does not work.

But, the distance scales linear with the mass, so twice solar mass yields about a solar radius distance, while the star only becomes 1.3 times larger. Assuming the same density, we get a possible solution for a star with 1.5 solar radii and 3 solar masses. This would be an early A type star, not a G2. If it was an evolved star the density would be way lower, so again, it would not work.

Most likely explanation seems to be a binary with a faint compact source. I am not sure, but possibly also a brown dwarf could work, anybody knows the density of these objects?

-- 17:50, 6 November 2007 (CLST)

Stupid, I made a mistake, took the full width instead of the half width for the velocity. Velocity of the inner disc will thus be 325km/s and the corresponding distance for 1 solar mass star is about 1.4 solar radii. So it would just about work and is no argument against the G2 star being the accreting object. Sorry about that! However, without a companion I can't see a way to explain the high velocity changes of H-alpha and CaII. -- 10:25, 7 November 2007 (CLST)


 * The companion has to be faint, since the new spectra show no obvious sign of an existing companion. If you compare the part of the spectrum located in the Mg triplet region with a solar spectrum, you can easily recognize all strong lines. 02:40, 7 November 2007 (CLST)

Spectral Energy Distribution
While in La Silla I decided to play a bit with the measurements which are available on SIMBAD and I have found out that there are fluxes all the way out to 100microns. So I decided to try to draw a spectral energy distribution (see figure).

I don't really know how to interpret it but I would assume that should relate with some dust. Is that the dust cocoon in which the star is forming? How does it relate with the nebula? I'll try to do some "CLOUDY"-modelling in the next few days. -- Alessandro 17:20, 14 November 2007 (CLST)

You may want to plot the y-axis in dimension-free units and overplot the dust onto the stellar photosphere (e.g. via a Kurucz model atmosphere). This would be "the standard" when discussing dust excesses in SEDs (see e.g. Malfait et al. 1998, A&A, 331, 211, Sect. 3.1 and 3.2; Dominik & Dullemond, A&A, 397, 595, Sect. 2). Some de-reddening may be necessary as well. From the dust SED, only guesses on the spatial dust distribution would be possible; it cannot replace imaging data. -- Oliver 02:26, 14 December 2007 (CLST)

Here you can find a table to get the same figures I have produced.

Alessandro 16:56, 17 December 2007 (CLST)