OHcounterpBody.tex ApJ 461 L41 April 10, 1996
To: gcnews@astro.umd.edu
Date: Wed, 10 Apr 1996 21:11:48 +0200 (MET DST)
Return-Receipt-To: sjouwerm@oso.chalmers.se

Phone in Onsala: +46-31-7725506 (switch 5500, fax 5590)
Phone in Leiden: +31-71-5275837 (switch 5834, fax 5819)
X-Mailer: ELM [version 2.4 PL23]
Content-Type: text
Content-Length: 21200     

%%%%%
% Sjouwerman & Van Langevelde: OH counterparts.. revision 1.1 (last)
%
% main text, table follows
%
%%%%%%%%%%%%%%%%%%%%%
% referee lay-out:
%\documentstyle[12pt,aasms4]{article}
% one-column preprint:
\documentstyle[12pt,aaspp4]{article} 
% two-column:
%\documentstyle[12pt,aas2pp4]{article} 

\received{}
\revised{}
\accepted{}
\begin{document}

\title{OH counterparts for H$_{\bf 2}$O masers in the Galactic center: 
       evolved stars instead of signs of recent star formation}

\author{Lor\'ant O.\ Sjouwerman}
\affil{Onsala Space Observatory, 439 92 Onsala, Sweden}
\affil{Sterrewacht Leiden}
%\authoraddr{OSO, Onsala Space Observatory, 439 92 Onsala, Sweden}
\affil{sjouwerm@oso.chalmers.se}
\and
\author{Huib Jan van Langevelde\altaffilmark{1}}
\affil{Joint Institute for VLBI in Europe,
P.O. Box 2, 7990 AA Dwingeloo, the Netherlands}
\affil{huib@nfra.nl}
%{JIVE, P.O. Box 2, 7990 AA Dwingeloo, the Netherlands}
%\authoraddr{NRAO AOC, P.O. Box 0, Socorro, 87801 NM, USA}
\altaffiltext{1}{temporarily at N.R.A.O., P.O. Box 0, Socorro, 87801 NM, USA}

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\begin{abstract}

We present the detection of OH maser emission associated with the
H$_2$O masers recently found in the Galactic center by Levine et al.\
(1995) and by Yusef-Zadeh \& Mehringer (1995). The 1612 MHz OH masers
were found in high-sensitivity maps created by combining 17 VLA
observations taken by Van Langevelde et al.\ (1993) as well as in new
observations with the ATCA. Both Levine et al.\ (1995) and Yusef-Zadeh
\& Mehringer (1995) consider the H$_2$O emission to be clues for
recent massive star formation, by associating it with a supergiant and
an \ion{H}{2} region. The newly found OH masers show the typical
double peaked spectra for evolved oxygen-rich stars and do not stand
out amongst other OH/IR stars in this region, neither in H$_2$O nor OH
maser characteristics. We conclude that the H$_2$O maser detections
are associated with evolved, low to intermediate mass stars and that
these H$_2$O masers thus cannot be regarded as signposts for massive
young stars or star forming regions.

\end{abstract}

\keywords{Galaxy: center -- infrared: stars -- masers -- stars: AGB and 
          post-AGB -- stars: formation -- surveys}

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\section{Introduction}

Claimed signs of recent massive star formation in the inner few
parsecs of the Galactic center (GC) have resulted in an ongoing
debate. The total mass and luminosity of young stars found 
until recently has been
insufficient to power and drive the phenomena in the central few
parsecs of our Galaxy, giving rise to searches for signs of a central
engine, an accretion disk around a massive black hole. Recent reports
on the detection of young and massive stars, that could provide the
energy needed to explain the observations, undermine the
necessity for a massive black hole located in the center (cf. Lacy et
al, 1980; Krabbe et al., 1995; Zylka et al., 1995; Genzel et al., 1994
for a review).

From the observations presented here we cannot tell whether massive
star formation has occurred in the past couple of hundred million
years. We merely want to react upon Letters written by Levine et al.\
(1995, hereafter referred to as LFMM) and by Yusef-Zadeh and Mehringer
(1995, Y-ZM), who both used an H$_2$O maser as evidence for recent or
ongoing massive star formation in the central 10 parsecs of the GC.

LFMM associate the H$_2$O maser with a luminous and massive supergiant
that could be physically connected to the circumnuclear disk and hence
argue for recent star formation. The link to ongoing high-mass star
formation by Y-ZM is based on lacking 1612 MHz OH maser emission and
the coincidence of the H$_2$O maser with a possible shocked or
pressurized molecular cloud.

However, in two surveys to find new OH masers for dynamical studies,
we have found that both H$_2$O
maser sources coincide with 1612 MHz OH maser sources, both with a
double peaked spectrum which is typical for oxygen-rich stars
undergoing spherical mass-loss on the asymptotic giant branch (AGB).
As the OH and H$_2$O maser emission characteristics seem to indicate
that both objects are AGB stars, OH/IR stars or Mira variables, the
two H$_2$O masers found by LFMM and Y-ZM cannot be used as proper
indicators for star formation in the GC in the last couple of hundred
million years.

In the following sections we describe the method of detection and
present 1612 MHz OH spectra. The characteristics of the observed OH
and H$_2$O maser emission will be discussed and compared to masers of
previously known OH/IR stars in the GC. We discuss the near-infrared
photometry available for the LFMM source in terms of an evolved AGB
star. We argue that both objects are low to intermediate mass, evolved
stars and therefore at least several hundred million years old (Iben
\& Renzini, 1983, for a review on AGB star evolution).

\section{Observations}

In a monitoring program with the VLA\footnote{The Very Large Array
(VLA), as part of the National Radio Astronomy Observatory, is
operated by Associated Universities, Inc., under cooperative agreement
with the National Science Foundation.}, Van Langevelde et al.\ (1993)
observed the strongest OH/IR stars found by Lindqvist et al.\ (1992a)
in a survey of the GC in the 1612 MHz maser line of the OH molecule
(Van Langevelde et al., 1993, and Lindqvist et al., 1992a, for details
on the observations). We have combined 17 epochs of the monitor data
into a high-sensitivity visibility data set. OH/IR stars can vary up
to a factor of two in integrated 1612 MHz flux densities. The profile
of OH/IR stars varies with periods in the range of 400 to 2000 days
(e.g.\ Van Langevelde et al., 1993). Finding weak OH/IR stars in this
combined data set is therefore probably more efficient than searching
each epoch separately or than searching new high-sensitivity
observations at one epoch. The data reduction techniques and our final
results, including a complete list of many weak and previously unknown
OH masers in the GC, will be presented in a forthcoming paper.

In addition, we used the ATCA\footnote{The Australia Telescope Compact
Array (ATCA) is operated by the Australia Telescope National Facility,
CSIRO, as a National Research Facility.} in July 1994 to search for
high velocity OH/IR stars (Sjouwerman et al., 1996). 
Both the OH data sets have comparable
sensitivities and are about four times more sensitive than the
original Lindqvist et al.\ (1992a) survey. 

We give the OH counterpart detections and approximate integrated flux
densities in Table 1. Allowing for resolution and possible systematic
effects in the radio maps, the OH maser positions are consistent with
the H$_2$O masers.  Both sources are clearly variable in the OH maser
line, but it is difficult to interpret the integrated flux density
from the VLA data as the flux density is an average over 17 epochs,
taken in a time-span of almost three years. The 1612 MHz spectra from
the VLA data can be found in Fig.\ 1a and 1b. In Fig.\ 2a and 2b the
spectra from our ATCA survey are displayed. Note that there are
considerable difficulties when reaching such low noise levels in the
GC area; residuals of the extended OH absorption, continuum
subtraction and very strong maser sources cause the poor baselines in
both the VLA and ATCA data.  The OH masers have expansion velocities
of about 15 and 19.5 km s$^{-1}$ centered on stellar radial velocities
(line-of-sight, with respect to the LSR) of about 49 and 105 km s$^{-1}$
respectively, very well in agreement with the observations of LFMM and
Y-ZM. We therefore believe the H$_2$O and OH masers are associated
with the same object in both cases.

\section{Discussion}

Most of the previously unknown double peaked OH masers found are seen
in both the combined VLA monitor data and our new ATCA
observations. So far we have only searched for OH emission exceeding
40 mJy, which is already a factor 2.5 deeper than the Lindqvist et
al.\ (1992a) survey. The weaker intrinsic OH emission explains why
LFMM and Y-ZM unsuccessfully searched the survey of Lindqvist et al.\
(1992a) for a previous OH detection of their H$_2$O maser sources.

We assume that our new, double peaked OH masers are associated with
similar objects as the Lindqvist et al.\ (1992a) stars, i.e.\ low to
intermediate mass, evolved AGB stars.  We reach this conclusion by
noting that our newly detected OH masers compare very well with the
stronger OH masers detected by Lindqvist et al.\ (1992a) in all their
properties (sky distribution, radial velocity distribution, expansion
velocities, OH maser variability). Also, the expansion velocity
distribution ranges from about 5 to 30 km s$^{-1}$ and has a mean of
17 km s$^{-1}$ for both the known and newly found OH/IR stars in the
GC (cf.\ e.g.\ Blommaert et al., 1994).  Adopting the same
metallicity, the distribution of the expansion velocities indicates a
similar bolometric luminosity distribution for the weaker OH maser
stars (Habing et al., 1994); they appear not to be different from the
previously known OH/IR stars in any other respect than showing a
weaker OH maser.

Parallel to the observations of LFMM and Y-ZM, our group has conducted
a VLA survey for H$_2$O maser emission in the 150 OH/IR stars found
previously near the GC. The detection rate for H$_2$O in OH/IR stars
is low, partly because the H$_2$O masers are intrinsically weak. Also,
in contrast to the OH masers, the H$_2$O maser emission varies
strongly and irregularly. At the distance of the GC, as a consequence
of sensitivity limits, one will find in most cases only one dominant
H$_2$O peak near one of the OH maser peaks, a few km s$^{-1}$ closer
to the mean velocity, or H$_2$O emission at the mean velocity. The
H$_2$O masers in the GC that we and Lindqvist et al.\ (1990) did
detect in OH/IR stars, as well as the detections of LFMM and Y-ZM, all
have similar H$_2$O maser characteristics: a double peaked H$_2$O
profile with the same mean (stellar) velocity as the OH maser or one
single peak close to the mean velocity (e.g.\ Engels et al., 1986).
The combination of the LFMM and Y-ZM H$_2$O masers and our OH
detections fit the general OH/IR star maser picture as outlined
above. For supergiants and star forming regions one finds in general
much stronger H$_2$O masers with a complex spectral profile (cf.\
Engels et al., 1988; Reid et al., 1988).

In general the OH maser luminosities for evolved, low to intermediate
mass OH/IR stars are 1 -- 100 times higher than the H$_2$O maser
luminosities. Depending on the mass-loss rate, occasionally the
opposite is also found for the thinner circumstellar shells of Mira
and semi-regular variables (e.g.\ Bowers \& Hagen, 1984). Of course,
when working with variable sources one has to be careful to make
comparisons between luminosities. Nevertheless, we think we can regard
the H$_2$O luminosity for the Y-ZM source to be less than the OH
luminosity, suggesting that this object is a regular OH/IR star,
whereas the comparison of OH and H$_2$O for the LFMM maser indicates a
thinner circumstellar shell (see Table 1).

Lindqvist et al.\ (1992b) discuss the kinematics of the OH/IR stars
found in Lindqvist et al.\ (1992a). In good agreement with the results
of McGinn et al.\ (1989), Sellgren et al.\ (1990) and Rieke \& Rieke
(1988) for K and M giants, the OH/IR stars show Galactic rotational
behavior. The stars have a large dispersion in radial velocity (70 km
s$^{-1}$ or more) and the agreement of the radial velocity of the LFMM
star with the circumnuclear disk is in our view accidental. The
agreement is no strong evidence for such an object to be young, i.e.\
a couple of tens of million years.

The strongest argument by LFMM for a young age of the star associated
with the H$_2$O maser is its very high luminosity (100,000 L$_\odot$).
LFMM derive this from H, K and L band photometry by arguing that these
observations are consistent with an M5 supergiant, seen at large
visual extinction (A$_{\rm V} \approx$ 37) when applying a standard
interstellar reddening law. With the knowledge that the star must have
a considerable circumstellar shell, because it supports an OH 1612 MHz
maser, the red colors of the object can be understood readily with
less foreground extinction. Moreover, following the identification
with a circumstellar shell, the near-infrared flux is produced mostly
by this extended envelope. Hence the bolometric correction is much
more moderate, resulting in a lower bolometric luminosity for the
star.

The detailed calculation of a total luminosity or an accurate
bolometric correction is quite uncertain because only near-infrared
colors are available. Most of the luminosity is expected at M band and
beyond, where the extinction correction would be less steep as
well. Correcting the photometry with a standard extinction law (Rieke
\& Lebovsky, 1985) and a value of A$_{\rm V}$ = 30, which seems more
appropriate for the location of this object (Catchpole et al., 1990),
we find colors that are quite blue for OH/IR stars, but still
consistent with a circumstellar envelope. We estimate a bolometric
correction in L band of 4.9 magnitudes in this case (cf.\ Jones et
al., 1994).  With these uncertainties the total bolometric luminosity
could be in the range 7000 -- 10000 L$_\odot$.  In fact, the
interstellar extinction law is a possible source of uncertainty and we
favor a reddening law based on extinction curve no.\ 15 of Van de
Hulst (1949) for the GC, which leads to even a lower bolometric
luminosity of 5500 L$_\odot$ (Blommaert et al., 1996).

The estimated luminosity, based on the identification as an AGB star,
is comparable with the luminosity of the known OH/IR stars in the GC
(Jones et al., 1994; Blommaert et al., 1992). This indicates a main
sequence mass of the star in the order of 2 M$_\odot$ and an age
largely exceeding several hundred million years (e.g.\ Van der Veen \&
Habing, 1990). The near-infrared colors would then imply a (K-L)$_0$
as low as 0.56, corresponding to a mass-loss rate of about
3.5$\times$10$^{-6}$ M$_\odot$ yr$^{-1}$ (Lepine et al., 1995). This
mass-loss rate is sufficient to form an OH masering circumstellar
shell (Bowers \& Hagen, 1984), but the low value of (K-L)$_0$ also
indicates the circumstellar shell itself is optically thin in the
near-infrared, such as in the optically visible Mira variables or
proto-planetary nebulae (e.g.\ Lepine et al., 1995). The latter
however, normally have no H$_2$O masers (Lewis, 1989). A thin
circumstellar shell is also justified by the earlier remark that the
H$_2$O luminosity might be higher than the OH luminosity. We therefore
base our classification of the LFMM object as an evolved low mass Mira
type OH/IR star on the OH maser profile and the OH and H$_2$O maser
characteristics. This is entirely consistent with the near-infrared
measurements taken by LFMM.

A final remark about the object found by LFMM concerns an article
written by McGinn et al.\ (1989). They use integrated infrared
starlight to determine mean stellar velocity and velocity dispersion
in several 20\arcsec\ diameter telescope apertures to probe the
stellar kinematics and mass distribution in the GC. One of the
telescope beams (45\arcsec\ NE) is pointed towards the LFMM maser
source and McGinn et al.\ (1989) derive an average stellar velocity of
48 ($\pm$8) km s$^{-1}$ for this beam. This is exactly the velocity of
the LFMM maser source. It seems likely that the LFMM maser source has
dominated the telescope beam. The observation of an increasing
velocity dispersion towards the dynamical center seen by McGinn et
al.\ (1989) and also by Sellgren et al.\ (1990), could differ from the
constant velocity dispersion observed by Rieke \& Rieke (1988),
because of telescope beams being dominated by single stars instead of
representing true stellar averages.

% Conclusion

In summary, we conclude that the H$_2$O maser of Y-ZM is associated
with an evolved OH/IR star. We think the H$_2$O maser found by LFMM,
instead of a massive supergiant, is more likely to be a Mira type
OH/IR star. We cannot rule out recent massive star formation in the
GC, but we do conclude the H$_2$O masers found by LFMM and Y-ZM do not
support the presence of a young population of massive stars.
 
\acknowledgements

We thank Joris Blommaert and Wil van der Veen for their help with the
interpretation of the infrared results and Harm Habing, Anders
Winnberg and Michael Lindqvist for carefully reading the manuscript.
HJvL acknowledges support for this research by the European Union
under contract CHGECT920011.

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\begin{table}
\dummytable\label{1}
\end{table}

%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\begin{thebibliography}{}

\bibitem[]{}Blommaert, J.\ A.\ D.\ L., Van Langevelde, H.\ J., Habing, H.\ J., Van der Veen, W.\ E.\ C.\ J.\ \& Epchtein N.\ 1992, in Variable Stars and Galaxies, ed.\ B.\ Warner (Astron.\ Soc.\ Pac.\ Conf.\ Ser.\ Vol.\ 30), 269
\bibitem[]{}Blommaert, J.\ A.\ D.\ L., Van Langevelde, H.\ J.\ \& Michiels, W.\ F.\ P.\ 1994, \aap, 287, 479
\bibitem[]{}Blommaert, J.\ A.\ D.\ L., Van der Veen, W.\ E.\ C.\ J.\, Van Langevelde, H.\ J., Habing, H.\ J., Epchtein N.\ \& Sjouwerman, L.\ O.\ 1996, in preparation; see Blommaert, J.\ A.\ D.\ L.\ 1992, PhD thesis, Leiden University, Chapter 3 or Van Langevelde, H.\ J.\ 1992, PhD thesis, Leiden University, Chapter 7 
\bibitem[]{}Bowers, P.\ F.\ \& Hagen, W.\ 1984, \apj, 285, 637
\bibitem[]{}Catchpole, R.\ M., Whitelock, P.\ A.\ \& Glass, I.\ S.\ 1990, \mnras, 247, 479
\bibitem[]{}Engels, D., Schmid-Burgk, J.\ \& Walmsley, C.\ M.\ 1986, \aap, 167, 129
\bibitem[]{}Engels, D., Schmid-Burgk, J.\ \& Walmsley, C.\ M.\ 1988, \aap, 191, 283
\bibitem[]{}Genzel, R., Hollenbach, H.\ \& Townes, C.\ H.\ 1994, Rep.\ Prog.\ Phys.\ 57, 417
\bibitem[]{}Habing, H.\ J., Tignon, J.\ \& Tielens, A.\ G.\ G.\ M.\ 1994, \aap, 286, 523
\bibitem[]{}Iben, I.\ \& Renzini, A.\ 1983, \araa, 21, 271
\bibitem[]{}Jones, T.\ J., McGregor, P.\ J., Gehrz, R.\ D.\ \& Lawrence, G.\ F.\ 1994, \aj, 107, 1111
\bibitem[]{}Krabbe, A., Genzel, R., Eckart, A., Najarro, F., Lutz, D., Cameron, M., Kroker, H., Tacconi-Garman, L.\ E., Thatte, N., Weitzel, L., Drapatz, S., Geballe, T., Sternberg, A.\ \& Kudritzki, R.\ 1995, \apjl, 447, L95
\bibitem[]{}Lacy, J.\ H., Townes, C.\ H., Geballe, T.\ R.\ \& Hollenbach, D.\ J.\ 1980, \apj, 241, 132
\bibitem[]{}Levine, D.\ A., Figer, D.\ F., Morris, M.\ \& McLean, I.\ S.\ 1995, \apjl, 447, L101 (LFMM)
\bibitem[]{}Lepine, J.\ R.\ D, Ortiz, R\ \& Epchtein, N.\ 1995, \aap, 299, 453
\bibitem[]{}Lewis, B.\ M.\ 1989, \apj, 338, 234
\bibitem[]{}Lindqvist, M., Winnberg, A.\ \& Forster, J.\ R.\ 1990, \aap, 229, 165
\bibitem[]{}Lindqvist, M., Winnberg, A., Habing, H.\ J.\ \& Matthews, H.\ E.\ 1992a, \aaps, 92, 43
\bibitem[]{}Lindqvist, M., Habing, H.\ J.\ \& Winnberg, A.\ 1992b, \aap, 259, 118
\bibitem[]{}McGinn, M.\ T., Sellgren, K., Becklin, E.\ E.\ \& Hall, D.\ N.\ B.\ 1989, \apj, 338, 824
\bibitem[]{}Reid, M.\ J., Schneps, M.\ H., Moran, J.\ M., Gwinn, C.\ R., Genzel, R., Downes, D.\ \& R\"onn\"ang B.\ 1988, \apj, 330, 809
\bibitem[]{}Rieke, G.\ H.\ \& Lebovsky, M.\ J.\ 1985, \apj, 288, 618
\bibitem[]{}Rieke, G.\ H.\ \& Rieke, M.\ J.\ 1988, \apjl, 330, L33
%\bibitem[]{}Rieke, G.\ H., Rieke, M.\ J.\ \& Paul, A.\ E.\ 1989, \apj, 336, 752
\bibitem[]{}Sellgren, K., McGinn, M.\ T., Becklin, E.\ E.\ \& Hall, D.\ N.\ B.\ 1990, \apj, 359, 112
\bibitem[]{}Sjouwerman, L.\ O., Winnberg, A., Van Langevelde, H.\ J., Habing, H.\ J.\ \& Lindqvist, M., 1996, in preparation
\bibitem[]{}Van de Hulst, H.\ C.\ 1949, Rech. Astr. Obs. Utrecht, Vol 11, Part 2
\bibitem[]{}Van der Veen, W.\ E.\ C.\ J.\ \& Habing, H.\ J.\ 1990, \aap, 231, 404
\bibitem[]{}Van Langevelde, H.\ J., Janssens, A.\ M., Goss, W.\ M., Habing, H.\ J.\ \& Winnberg, A.\ 1993, \aaps, 101, 109
\bibitem[]{}Yusef-Zadeh, F.\ \& Mehringer, D.\ M.\ 1995, \apjl, 452, L37 (Y-ZM)
\bibitem[]{}Zylka, R., Mezger, P.\ G., Ward-Thompson, D., Duschl, W.\ J.\ \& Lesch, H.\ 1995, \aap, 297, 83

\end{thebibliography}

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

%\figcaption[a=vll.eps, b=vly.eps]
%{1612 MHz OH spectra from the VLA monitor:
%a) OH 359.956--0.050, the counterpart for the H$_2$O maser found by LFMM and 
%b) OH 359.980--0.077, the counterpart for the H$_2$O maser found by Y-ZM. Both
%spectra are an average over 17 epochs.}

%\figcaption[a=atl.eps, b=aty.eps]
%{1612 MHz OH spectra from the ATCA survey:
%a) OH 359.956--0.050 (LFMM) and b) OH 359.980--0.077 (Y-ZM)}

%%%%%%%%%%%%%%%%%%%%%%%%%%%

%\end{document}



\begin{figure}
\figurenum{1a, 1b}
\plottwo{vll.eps}{vly.eps}
\caption{1612 MHz OH spectra from the VLA monitor:
a) OH 359.956--0.050, the counterpart for the H$_2$O maser found by LFMM and 
b) OH 359.980--0.077, the counterpart for the H$_2$O maser found by Y-ZM. Both
spectra are an average over 17 epochs.}
\end{figure}

\begin{figure}
\figurenum{2a, 2b}
\plottwo{atl.eps}{aty.eps}
\caption{1612 MHz OH spectra from the ATCA survey:
a) OH 359.956--0.050 (LFMM) and b) OH 359.980--0.077 (Y-ZM)}
\end{figure}

\end{document}