THE EXTREME ULTRAVIOLET EXPLORER BRIGHT SOURCE LIST

 
Roger F. Malina, Herman L. Marshall (1), Behram Antia, Carol A. Christian,
  Carl A. Dobson (2), David S. Finley, Antonella Fruscione, Forrest R.
  Girouard, Isabel Hawkins, Patrick Jelinsky, James W. Lewis, J. S. McDonald
  (3), Kelley McDonald, Robert J. Patterer (4), Vincent W. Saba, Martin M.
  Sirk, Brett A. Stroozas, John V. Vallerga, Peter W. Vedder (5), Alexandria
  Wiercigroch, and Stuart Bowyer (6)
Center for EUV Astrophysics, 2150 Kittredge Street, University of California,
  Berkeley, CA 94720

(1) currently at the Center for Space Research, MIT, 77 Massachusetts Ave. Cambridge, MA 02139
(2) currently with Dobson Software Innovations
(3) currently with the Department of Astronomy, San Diego State University, San Diego, CA 92182
(4) currently at Hughes STX, Goddard Space Flight Center
(5) currently at NASA headquarters, Code SZ, Washington, CD 20546
(6) also with the Astronomy Department, University of CA, Berkeley, CA 94720


ABSTRACT 

   Initial results from the analysis of the Extreme Ultraviolet Explorer (EUVE)
all-sky survey (58-740 A) and deep survey (67-364 A) are presented through the
Bright Source List (BSL).  The BSL contains 356 confirmed extreme ultraviolet
(EUV) point sources with supporting information, including positions, observed
EUV count rates, and the identification of possible optical counterparts. One
hundred twenty-six sources have been detected longward of 200 A.


1. INTRODUCTION

   The primary scientific goal of the Extreme Ultraviolet Explorer (EUVE)
mission (Bowyer & Malina, 1991) was to conduct a photometric survey of the
entire sky over the whole extreme ultraviolet (EUV) band, as well as to conduct
a more sensitive survey along the ecliptic.  EUVE was launched on 7 June 1992
on a USAF Delta II rocket and was placed in a circular orbit of 550 km altitude
with an inclination of 28 deg.  A six-week-long in-orbit checkout of the
instrument showed that the four telescopes and seven detectors were working
as planned.  Over the following six months, two distinct surveys, the all-sky
survey and the deep survey, were conducted with the four telescopes. The
all-sky survey was carried out with four distinct filters that cover the
wavelength region 58-740 A and that are named by the materials of which
they are composed:  Lexan/B, Al/Ti/C, Ti/Sb/Al (or "Dagwood"), and Sn/SiO
(or "tin").  Table 1 shows the ~10% instrument transmission wavelengths for
all the bandpasses on the EUVE telescopes.  From 22 July 1992 to 21 January
1993, the scanners were used to map about 80% of the sky with exposures
varying from 400 s on the ecliptic equator to approximately 20000 s at the
ecliptic poles.  Concurrently, the fourth telescope, the Deep Survey/Spec-
trometer (DS/S), was used to carry out the "deep" survey (more sensitive
by a factor of ten) of a 2 deg by 180 deg strip of the sky along the ecliptic
in two filters (Lexan/B and Al/C) covering the wavelength region 67-364 A.
   The survey was interrupted on a monthly basis to conduct calibration
pointings at selected bright EUV sources, resulting in survey gaps which had
to be filled in during the second six months of the mission, from January
to July 1993.  The data corresponding to the ~20% of the sky in the survey
gaps are not included in this paper.  Since the end of the survey phase of
the mission, the three spectrometers that share the DS/S telescope are being
pointed at targets chosen by guest observers selected by NASA.  Details of the
spectrometers and the  Guest Observer (GO) program can be found in the EUVE
GO Program Handbook, which is included in the NASA Research Announcement (1993).
   In this paper we present the EUVE Bright Source List (BSL), which consists
of bright EUV point sources detected during the 22 July 1992 to 21 January
1993 survey phase of the EUVE mission, as well as bright sources observed
during the in-orbit calibration phase.  Included in the list are source po-
sition, count rates for each wavelength band at which the source was detected,
and a possible optical identification of the source.  The list is not meant
to be complete; rather it is the first release of the sources bright enough 
that their existence as EUV sources is not in doubt.  The first EUVE catalog
is in preparation (Bowyer et al., 1993), which will present the analysis after
reprocessing of the initial sky survey data as well as analysis of the gap
areas, which are not considered in this paper.  [The Wide Field Camera (WFC)
on the Roentgen Satellite (ROSAT) has produced a list of EUV sources at the
shortest EUV wavelengths (Pounds et al., 1993).  The two filter bands employed
in the WFC survey, S1 (60-140 A) and S2 (110-200 A), do not extend as far
into the longer EUV wavelengths as do the  filters, and they are less sen-
sitive.  The Lexan/B filter on the  scanners is similar in bandpass to the
S1 filter of the WFC; the surveys in the Al/Ti/C, Dagwood, and tin filter
bandpasses of EUVE are the first all-sky surveys made at these wavelengths.]


2. THE EUVE INSTRUMENTS
   The sky survey was conducted with three "scanning" telescopes, which were
coaligned and pointed 90 deg away from the satellite spin axis during the
survey.  The satellite spun at three rotations per orbit so that, during an
orbital night, the telescopes' scan path mapped a great circle on the sky.
The DS/S telescope was aligned with the axis of rotation, which was kept
pointed in the anti-sun direction, and therefore advanced along the ecliptic
at 0.986 deg per day.  Two of the scanners (Scanners A and B) are essentially
duplicates of each other, incorporating a grazing incidence Wolter-Schwarz-
schild Type I telescope with gold electroplated mirror surfaces. They have
a field of view of about 5 deg and a focal length of approximately 56 cm.
The third scanner (Scanner C) is a Wolter-Schwarzschild Type II telescope with
the surface coated with electroless nickel.  The field of view is slightly
smaller (4 deg) and the focal length is 70 cm (Finley et al., 1988).
   All seven detectors on  consist of a Z-stack of microchannel plates (MCPs)
that are read out by a wedge-and-strip anode (Siegmund et al., 1986).  The
detectors are identical except for the photocathode deposited on the front
surface to increase the quantum efficiency and the thin film filters used to
define the bandpass.  A magnesium fluoride photocathode was used on Scanners
A and B and the Deep Survey detectors, whereas the Scanner C detector was
operated without a photocathode (to decrease the sensitivity to the strong
geocoronal line at 584 A).  The quantum efficiency of the detectors did not
degrade significantly from the time of deposition of the photocathodes to
the last calibration of the instrument on the ground, and there has not been
any evidence of sensitivity variations of the instruments in orbit (with a few
percent upper limit to any change).
   The EUVE thin film filters were chosen to cover, as much as possible, the
full EUV bandpass, while remaining spectrally distinct and reducing sensitivity
to the strong geocoronal lines at 304, 584, and 1216 A.  The history of the
filter development and optimization can be found in Vallerga et al. (1992),
and the flight calibration results are given in Vedder et al. (1992).  The
filters above the scanner MCP detectors are broken into four quadrants, with
Scanners A and B both containing Lexan/B and Al/Ti/C.  The longer wavelength
filters required a different telescope with a larger grazing incidence angle
to reduce the X-ray  throughput since these filters were known to have sub-
stantial transmission in the soft X-rays (Finley et al., 1988).  The Lexan/B
filters have a low residual transmission at the UV wavelengths (> 2300 A)
and though the solar-blind detectors' response is decreasing exponentially,
there is still enough throughput (~1e-10) that very bright FUV stars such as
O and B stars can be detected.  The instrument was carefully designed to limit
detection of O and B stars to only the 100 brightest in the sky with m_V <= 5.
The Al/Ti/C filter also has a very slight soft X-ray leak to light with wave-
lengths shorter than 70 A, evident in a slightly significant detection of the
brightest soft X-ray source in the sky, Scorpius X-1.
   The prelaunch calibration of  was extensive.  Absolute calibration trace-
able to primary standards provided by the National Institute of Standards
and Technology (NIST) was performed on every instrument over the wavelength
range of 23-2537 A using monochromatic line sources of radiation (Welsh et
al., 1990).  Pencil beams were used to sample and map the response of the
telescope aperture and the entire detector surface.  These results were com-
bined to give the effective area curves shown in Figure 1.  For the bandpasses
that are on both Scanner A and B, these effective areas represent the average
of the two instrument effective areas (which differ by less than 10%).  In-
orbit calibration was carried out once a month to verify that there was no
significant change in instrument performance.
   The background count rates for each detector are due to many sources, in-
cluding intrinsic MCP background; geocoronal scattered solar radiation, pre-
dominantly 304 A (HeII), 584 A (HeI), and 1216 A (HI); primary cosmic rays;
low energy electrons in the Earth's radiation belts that can penetrate the
magnetic brooms that cover the telescope aperture; and any possible cosmic
diffuse EUV radiation.  The Lexan/B bandpass background is dominated by the
intrinsic background of the detectors due to the decay of K^(40) in the MCP
glass.  The Al/Ti/C and the Dagwood bandpasses include the 304 A line, which
thus dominates as a source of background, whereas the tin filter is strongly
dominated by the 584 A line.  The background is variable owing to the spatial
and temporal variability of the geocorona.  The detectors are also strongly
affected by passages through the South Atlantic Anomaly and by solar-induced
geomagnetic storms, where the count rates can exceed 2000 cps, which causes
an automatic shutdown of the detectors for health and safety reasons.  However,
the typical values of the background during a quiet orbit are very low, 1.7e-4
cts s^(-1) sq-arcmin^(-2) for the Lexan/B, 2.1e-4 cts s^(-1) sq-arcmin^(-2)
for the Al/Ti/C band, 5.0e-4 cts s^(-1) sq-arcmin^(-2) for the Dagwood band,
and 1.7e-3 cts s^(-1) sq-arcmin^(-2) for the Sn/SiO band.  The total background
is typically less than our telemetry bandwidth of ~450 cps, so except for the
brightest sources or high background events, corrections for deadtime due to
the electronics or telemetry allocation are usually less than 10%.
   The point spread functions (PSFs) of the instruments are a strong function
of the off-axis position of the source in the field of view because of the
inherent off-axis aberrations the telescopes.  However, for the survey, where
most sources were sampled in the entire field of view, an average survey PSF
can be determined for each bandpass.  The half energy radius for sources within
1.0 deg off-axis is 1 arcmin for Scanners A and B, and 0.5 arcmin for Scanner
C.  An average PSF cannot be determined for the Deep Survey, as sources with
different ecliptic latitudes sampled different parts of the detector field of
view and were effectively observed with differing instrument PSFs.


3. THE BRIGHT SOURCE LIST DEVELOPMENT

3.1 Science Data products

   The major science data products created by the EUVE data processing soft-
ware are exposure maps, photon skymaps, and "pigeonholes" (for a detailed
description of the EUVE software see Lewis et al. 1993; Antia 1993).  Exposure
maps (represented in Fig. 2 for each of the scanning and deep survey tele-
scopes) contain information about the amount of time each region of the sky
was surveyed.  Raw photon skymaps represent the binned distribution of the
EUV photons over the sky with a resolution of ~1.3 arcmin/pixels.  Pigeonholes
are files containing detailed photon data for a circular area of sky (of ~10
arcmin radius) centered on a selected position. The pigeonhole files also
store additional precise information not included in the skymaps, such as
the photons' arrival times and their location on the detector, and also
preserve a higher resolution of ~10 arcsec/pixel.  Before launch, a master
pigeonhole catalog containing ~6000 positions on the sky was compiled, cor-
responding to the best candidate EUV sources.  Then detailed studies from
the pigeonhole data for the most interesting objects was possible immediately
after the first processing of the data, independent of the overall source
detection process.  The source detections were derived from analysis of the
skymaps, whereas in most cases the pigeonholes, which allow more detailed
study, were used in the source verification and in the count rate measurements.

3.2 All-Sky Survey Point Source Detection Approach

   The scanner skymaps were searched for possible sources using a preliminary
version of the EUVE source detection software (Lewis 1993).  This code uses
a two-pass method to produce the final list of detections.  In the first pass,
the skymaps are convolved with the appropriate PSF.  The convolved data are
treated as if they consisted of samples from a normal distribution; a threshold
is applied to the peaks of the convolution, and significant deviations above
the local background level are reported.  The result of this pass is a list
of "interesting" pixels, each 1.2 by 1.2 arcmin in size. The first-pass thre-
shold is chosen so that real sources would be reported with very high proba-
bility while keeping the spurious detections to a manageable level.  The
second pass uses a more rigorous approach to examine each suspected source
and calculate a more reliable detection statistic.  The patch of sky around
a suspected source is modeled as a locally flat background plus a source
contribution folded through the PSF.  The background level is estimated from
a 15-24 arcmin annulus centered on the suspected source.  A likelihood ratio
test using this model is applied to the data, yielding a detection statistic
distributed as chi-squared.  All sources with a detection statistic above 10
form the candidate list.  Because of the variable exposure time in different
areas of the sky and the variations in the background, the minimum detectable
flux of an object varies from area to area of the sky.

3.3 All-Sky Survey Point Source Measurement

   The second stage in the compilation of the BSL consisted of a more critical
analysis.  The pigeonholes for each source in the BSL were analyzed with
software based on a likelihood method developed independently of the source
detection program (see Cruddace et al., 1988 for a similar application of
this method to ROSAT X-ray data).  The detailed time, position, and detector
location data for all events in the pigeonholes were used to estimate simul-
taneously the source position, count rate, and background, assuming Poisson
statistics. The model incorporates the detector vignetting map (which also
includes blockage by the filter frame) and the telescope PSFs. The PSFs vary
considerably with off-axis angle, so they have been computed for eight
annuli in detector coordinates.  As with the first-stage detection software,
a likelihood statistic is formed and maximized for each source.  The likeli-
hood difference between the best value and the value obtained when the source
count rate is fixed at zero is a measure of source existence and is distribu-
ted approximately as chi-squared with one degree of freedom.  This existence
test is more critical than the original detection statistic because more
information is used:  the PSFs vary with detector location and the event data
are not binned as is the case for the skymaps.  Therefore, although a source
may appear to be significant to the detection program, it may be eliminated
at this step.
   The existence threshold was set at a value of 36, which for uniform Poisson
background arrivals ought to yield a 10% probability of a single false source
detection for the entire BSL spanning four independent all-sky maps.
   The positions determined for bright, easily identified sources were com-
pared to the higher accuracy optical positions in order to estimate systematic
errors.  Positions in the BSL are derived from the short wavelength telescope
data (when available) for which a boresight offset has been estimated at 40
arcsec; all positions in the BSL have been corrected for this offset.  Al-
though some small systematic uncertainties still exist, all positions should
be accurate to 60 arcsec at about 90% confidence.  No boresight correction is
yet available for the long wavelength telescope, and therefore the positions
of the few sources detected only in the Dagwood or tin filters have not been
corrected for this offset. The position of these sources should be accurate 
to ~90 arcsec at about 90% confidence.
   The intensity of the sources was computed  and the survey count rates for
calibration sources were compared to their count rates calculated from the
pointed calibration observation.  In general the calibration and survey count
rates agreed to within 30%.  Again, systematic uncertainties in the calibra-
tion data and the vignetting maps dominate for bright sources.

3.4 Deep Survey Sources

   An objective detection algorithm for the deep survey is difficult because
the PSF and exposure are strong functions of ecliptic latitude.  The exposure
can vary with ecliptic latitude by as much as a factor of five over one de-
gree.  Therefore, the sensitivity of the deep survey is not constant over
ecliptic latitude, and sources more than 0.5 deg from the ecliptic have to
be very bright in order to be detected.  New algorithms for analyzing the
deep survey data are being developed and will be used to produce the EUVE
catalog (Bowyer et al., 1993).
   In order to include sources from the EUVE deep survey in the BSL, we used
a simple (but subjective) visual detection process on the raw deep survey
skymaps, binned at the 1.3 arcmin scale.  Strong enhancements that had the
spatial morphology of sources were included in the BSL if they had two or
more adjacent pixels approximately 4 sigma above the surrounding local
background.  Preliminary count rates for the deep survey sources were derived
from the deep survey skymap using a simple scheme wherein the source plus
background counts are determined within a circle of ~2 arcmin radius around
the source position, and the background is estimated in an annulus around
the source.  For the sources far from the ecliptic and therefore farther
off-axis, larger apertures were used.  Errors in the DS count rate should be
within 50%.

3.5 Bright Source List Selection Criteria

   A major effort has been made to eliminate possible spurious sources from
the BSL, and a visual verification test of the data was added to the source
selection process as a further check, independent of the source detection and
analysis software.  Each candidate pigeonhole was convolved with the telescope
PSF for each filter and independently examined by several researchers.  Any
sources with highly irregular images were set aside and removed from the BSL;
further studies of these sources are under way, and the sources will be in-
cluded in the EUVE catalog if they are confirmed as definite detections.  An
example of the raw and PSF-convolved pigeonhole data that were examined is
given in Figure 3 for the source EUVE J0623-37.6.
   As a result of the multiple-step process described above (detection in the
skymaps, analysis of the pigeonholes, visual inspection), a source is included
in the BSL only if the following criteria are satisfied simultaneously:

  [a.] The existence likelihood of the source (see section 3.3) is
	delta-S >= 36 in any single filter.
  [b.] No UV pinhole leak (see section 3.6) is revealed by a careful analysis
	of the scanning of the source in detector coordinates.
  [c.] The visual examination of the pigeonhole data, convolved with the tele-
	scope PSF, confirms the existence of the source, and the image shape
	is not anomalous.

In addition,

  [d.] Sources with a detection likelihood 10 <= delta-S <= 36 are included
	only if detailed examination of the source supported its detection
	via auxiliary data such as analysis of the light curve or positional
	coincidence with a known or candidate EUV or soft X-ray source.
  [e.] Sources observed during the calibration phase are included.  All these
	sources were detected in at least one filter with count rate uncer-
	tainties less than 15%.
  [f.] Sources detected visually from the deep survey skymaps are included if
	they satisfy the criteria explained in section 3.4.

   The primary goal of the BSL (also in view of its use by prospective EUVE
guest observers) was to be error-free for the sources included. In this con-
servative process, a number of real sources may have been omitted (for example,
some 200 sources with significance between 10 and 36, with no clear counterpart
and with no subsidiary studies to confirm the reality of the source, have been
excluded from the BSL at this time).  Readers should also note that some bright
EUV sources do not appear in the BSL if they are located in the survey gaps.
A later EUVE catalog will be produced that is expected to contain more sources
near the limit of detectability (Bowyer et al., 1993).
   Fig. 4 is a map, in galactic coordinates, showing the locations of all the
sources listed in the  BSL.

3.6 Data Anomalies 

   There were two known filter anomalies affecting the data used for the BSL
production.  The first is due to a few tiny pinholes in the Lexan/B filter
that allow non-EUV flux to be transmitted when a bright source is scanned
across them.  These pinhole leak sources were easily eliminated from the BSL
in the visual verification process.  The second problem is the known out-of-
band residual transmission in the far UV in Lexan/B, which was inherent in
the filter design and allowed UV detection of O and B stars brighter than 
m_V ~5.  A calibration of these bright B stars is in progress to allow for
estimating and subtracting the UV flux contribution. We have excluded all B
stars from the BSL except for cases where additional, detailed analysis con-
firms the detection of an EUV signal.  We have included in the BSL the A and
B stars that were reported as extreme ultraviolet sources in the ROSAT Bright
Source Catalogue.

3.7 Diffuse and Extended Sources

   Only pointlike sources have been included in this Bright Source List.  Any
sources that appear significantly extended or diffuse have been omitted.  Ex-
tended or diffuse EUV sources that have already been reported include the Moon
(Gladstone et al., 1993), the Vela and Cygnus supernova remnants (Edelstein et
al., 1993), and a cloud shadow (Lieu et al. 1993).  In addition, we have omit-
ted extended sources detected by EUVE guest observers during GO pointings, but
not detected during the sky survey; such sources include, for example, Mars
(Chakrabarti et al., 1993) and Jupiter (Moos et al., 1993).

4. THE BRIGHT SOURCE LIST

   Table 2 contains information on the 356 sources that meet the selection
criteria in section 3.5. The list includes all bright sources whose existence
was established before April 1993 in the data set obtained from 7 June 1992
to 21 January 1993.  New sources continue to be found in the data and will
be included in later releases of the EUVE catalog. The table is ordered in
right ascension, and it is organized as follows:

Column (1). -- EUVE source name as specified by the International Astronomical
	Union.  Each source is identified by its J2000 coordinates in hours
	and minutes of right ascension and decimal degrees of declination.
	Errors in the preliminary BSL names for half a dozen sources have been
	corrected.

Columns (2) and (3). -- Right ascension (hh:mm:ss) and declination (dd:mm:ss)
	of the EUV source in J2000 coordinates.  Positions should be accurate
	to 60 arcsec at about 90% confidence. Less accurate positions (~90
	arcsec) are available at present for the few sources detected only in
	the Dagwood or Sn/SiO filters (due to unknown boresight errors), and
	for the sources detected in the deep survey skymaps.  The distribution
	of the BSL sources over the sky is shown in Figure 4.

Columns (4) and (5). -- Galactic longitude (l) and latitude (b) of the EUV
	source.

Columns (6), (7), (8) and (9). -- Count rate, in counts/kilosecond, for the
	sources detected in the scanners' Lexan/B, Al/Ti/C, Dagwood, and tin
	filters.  The count rates have been calculated as explained in section
	3.3 and are accurate to within 50%, although there are occasional er-
	rors of a factor of 2 (see section 3.3).  Count rates have been rounded
	to the nearest value in multiples of 10, except where the rounding
	would alter the count rate by more than 30%.  Note that, although all
	BSL sources have delta-S >= 10, the list should NOT be considered
	complete at that level.
	   Count rates of calibration sources not scanned during the all-sky
	survey (because of gaps, etc.) were measured from the calibration data,
	on an orbit-by-orbit basis.  The highest significant value (i.e., high-
	est signal-to-noise ratio [SNR] and long exposure) has been quoted in
	the BSL.  In cases when a calibration source was also scanned during
	the survey, the survey measurements have been quoted in the BSL.  The
	detections by filter break down as follows:  327 sources in Lexan/B,
	92 in Al/Ti/C, 17 in Dagwood, and 11 in tin. Table 3 lists BSL source
	detection by filter.

Columns (10) and (11). -- Count rate, in counts/kilosecond, for the sources
	detected in the Deep Survey Lexan/B (DSL) and Al/C (DSA) filters. The
	values are derived as explained in section 3.4, except for deep survey
	sources observed during the calibration phase, for which the count
	rate is derived as explained in the paragraph above.  Some 39 sources
	were detected in the Deep Survey Lexan/B filter and 6 in the Al/C
	filter. (See Table 3.)

Columns (12) and (13). -- Possible optical counterpart of the EUV source and
	alternate name. The candidate counterparts come from a detailed search
	of astronomical catalogs, including the SIMBAD (Egret et al., 1991)
	and NED databases (Helou et al., 1991).  We choose the possible coun-
	terparts among the sources most likely to emit EUV radiation, e.g.,
	white dwarfs, active late-type stars, cataclysmic variables, and active
	extragalactic objects in low hydrogen column density directions
	( ~< 2e20/cm^2).  All counterparts lie within 1.5 arcmin of the EUVE
	source position, except for the deep survey sources, where the most
	likely candidate within 3 arcmin of the detected location is listed.
	(The rate of coincidental identification at 3 arcmin for the deep sur-
	vey sources was tested by searching for candidates around random po-
	sitions:  only one "false" object out of 28, equal to the total number
	of deep survey sources, had a plausible counterpart.)
	   No detailed nomenclature is given in general for different compo-
	nents of a same system (e.g., binary systems) and the notation "NOID"
	is used for currently unidentified sources.  For the purpose of this
	BSL, we chose the most likely identification in case the EUV source had
	more than one possible optical counterpart: follow-up spectroscopy and
	photometry are in progress to identify all the "NOID" sources and to
	confirm the other identifications. A more detailed description of the
	optical identification process is given in Christian et al. (1993).

Column (14). -- Source type or spectral type of the optical candidate for the
	EUV source. Spectral types generally come from the SIMBAD database
	and follow the SIMBAD spectral type coding (see Appendix D of the
	SIMBAD User's Guide and Reference Manual).  Extragalactic source types
	are from NED.  Some of the white dwarfs spectral types were provided
	by Finley (1993, private communication).  The notation "CSPN" indicates
	the central star of a planetary nebula, "XRB" an X-ray binary system,
	and "LMXB" a low mass X-ray binary.  Table 4 is a breakdown of the BSL
	sources by spectral type.

Column (15). -- Visual magnitude of the optical counterpart.  The measurements
	generally come from the variety of catalogs included in SIMBAD and NED
	databases and therefore should be considered only as a rough indica-
	tion.  See Figure 5 for a plot of the visual magnitude of the optical
	counterpart vs. the EUV flux.  We observe that 92% of white dwarf stars
	are fainter than m_V ~13 but are the brightest EUV sources, while 96%
	of late-type stars are brighter than  but lie within the lower two
	decades of EUV flux. 

Column (16). -- Where applicable, the name of the EUV source as it appears in
	the ROSAT WFC Bright Source Catalogue (Pounds et al., 1993).  EUVE
	and ROSAT sources are considered identical if they lie within 1.5
	arcmin of one another.

Column (17). -- General notes and/or comments about specific EUV sources (e.g.,
	which sources were calibration targets); all notes are explained at
	the end of Table 2.


5. SURFACE DENSITY OF EUV SOURCES

   Given the distribution of source intensities, we have computed the sky
angular surface density of sources given an estimate of the sky coverage
function.  We have generated an approximation to the sky coverage function
using the pigeonhole data for arbitrarily placed points on the sky, without
producing a detailed map of the sky that gives the sensitivity limits ap-
propriate for the detection runs used to find the sources. 

5.1 Sky Coverage

   In order to calculate the sky coverage included in the BSL, we added
locations to the master pigeonhole catalog at every 10 deg in right ascension
and declination as well as one each at the celestial poles.  There are 614
locations in this grid.  Although the grid points are not equally spaced on
the sky, they were chosen without regard to either the survey strategy (which
places emphasis on the ecliptic pole and may have gaps along ecliptic longi-
tude lines) or to the actual distribution of sources (the grid was determined
long before the survey and before the preliminary results from the WFC survey
were available).
   The pigeonhole for each grid position was handled in exactly the same way
as for those that were expected to contain sources, except that the "source"
location was not allowed to vary.  For detected sources with total count
rate estimate R and uncertainty sigma_R, an empirical relation between the
SNR, defined as R/sigma_R, and the square root of the significance, delta-S
(see section 3.3), indicates that a source with a significance of 36 is
expected to have a SNR of about four.  For grid sources, then, the sensitivity
was set to 4*sigma_R, the uncertainty as derived by the flux measurement
software.  The declination of the grid position was used to estimate the
size of the region for which this sensitivity was effective.  The angular
size of the region, in steradians, is given by 

(1)  omega(delta) = (4 * pi / 2 * 36)(sin(delta + 5 deg) - sin(delta - 5 deg))

and at the poles the angular size of the 5 deg radius region was used.  For
each location with coordinates (alpha_i, delta_i) there is a value of R_i
and omega_i, so the sky coverage function may be formed:

(2)  Omega(R) = sum_{R_i < R} omega_i

which gives the solid angle of the sky surveyed to the sensitivity R.  Inte-
gral versions of these curves are given for each filter in Figure 6.  A
differential version of the coverage function for the Lexan/B filter is
shown in Figure 7.
   The sky coverage function for the Lexan/B filter was checked by plotting
the distribution of count rates for sources with low SNR.  The measured count
rates were normalized by 6f/(delta-S)^0.5 and each source was given an equal
fraction of the sky.  The two distributions match reasonably well (Figure 8).

5.2 Source Counts

   One may construct the integral source counts for each filter by forming

(3) N(>R) = sum_{R_i > R} 1/Omega(R_i)

This curve is computed for all sources satisfying the significance criterion
(delta-S > 36) and is shown in Figure 9 for the Lexan/B band only and in
Figure 10 for all filters.  The number count curve reproduced from Pounds et
al. (1993) is shown for comparison in Figure 9.  (The sky coverage function
was estimated graphically from Figure 7a of Pounds et al. (1993).) These two
distributions can be compared by shifting the WFC curve by a factor of 2.1
to account for the different effective areas of the two instruments (the EUVE
filter has the larger effective area).  This factor was derived by comparing
the WFC and BSL count rates for white dwarfs in both catalogs and appears
independent of average count rate.  The shifted WFC curve is also presented
in Figure 9 and shows substantial deviations (> 20%) from the corresponding
curve at count rates below about 0.05 count/s.  This deviation could be due
to incompleteness in the EUVE BSL, but it is not likely to be caused by a
defect in the estimate of the sky coverage.  Because the remaining three
bands have never been observed elsewhere, it is very difficult to estimate
completeness from external data.
   A likelihood method that uses the information in the many count rate limits
from Figure 6 was used to estimate the value of a single power law fit to
the data (see Marshall 1985).  We find sources N(> R) = 19.3 * R^(-alpha)
per sky, where alpha = 0.80 +/- 0.05.  These values are consistent with those
of Pounds et al. (1993), after applying the ratio of 2.1 to convert the WFC
count rate, C1, to EUVE count rate, R.


6. CONCLUSION 

   We have presented the list of 356 bright sources detected with the EUVE
satellite during the in-orbit calibration phase of the mission and during the
all-sky and deep surveys.  More than 100 sources have been observed in the
wavelength interval between 200 and 800 A. For each source we have listed the
maximum likelihood EUVE position and the EUVE count rate.  For 341 sources we
propose a possible optical identification and visual magnitudes. Fifteen EUV
sources do not have any plausible optical counterpart, and followup spectro-
scopy and photometry are in progress to find the new optical identifications
and to confirm the others. The two largest classes of source detected at the
EUV wavelength are the late-type and white dwarf stars.


ACKNOWLEDGEMENTS

   The authors would like to thank the entire  team at the University of
California, Berkeley; in particular we acknowledge the critical roles of
M. Abbott, D. Biroscak, J. Drake, J. Dupuis, A. Hopkins, W. Marchant,
M. Lampton, R. Lieu, and S. Vennes.  We thank NASA's Goddard Space Flight
Center (GSFC), which manages the  project:  the GSFC Project Manager, Paul
Pashby; the Project Scientist, Dr. Yoji Kondo; the Deputy Project Scientist,
Dr. Ronald Oliversen; the NASA Headquarters Program Scientist, Dr. Robert
Stachnik; the Deputy Program Scientist, Dr. Derek Buzasi; the Program Manager,
Dr. Guenter Riegler; and the GSFC Project Operations Director, Mr. Kevin
Hartnett.  We would also like to acknowledge the critical roles that Frank
Martin, Charles Pellerin, and Ed Weiler have played in ensuring the success
of EUVE.  We acknowledge Andrea Frank's editorial assistance.  This research
is supported by NASA contracts NAS5--29298 and NAS5-30180.


REFERENCES

Antia, B. 1993, JBIS, 46, 337
Bowyer, S. et al. 1993, in preparation 
Bowyer, S., & Malina, R.F. 1991, in Extreme Ultraviolet Astronomy, ed.
	R.F. Malina & S. Bowyer (New York:  Pergamon Press), 397
Chakrabarti, S., Krasnopolsky, V., Gladstone, G.R. 1993, BAAS, 25, 855
Christian, C., et al. 1993, ApJ, in preparation
Cruddace, R.G., Hasinger, G.R., & Schmitt, J.H. 1988, in Astronomy from
	Large Databases (Garching:  ESO Conference Proceedings), 177
Cully, S., Siegmund, O.H.W., Vedder, P.W., & Vallerga, J.V. 1993,
	ApJL, in press
Edelstein, J., Vedder, P.W., & Sirk, M. 1993, BAAS, 25, 863
Egret, D., Wenger, M., & Dubois, P. 1991, in Databases and On-line Data in
	Astronomy, ed. M.A. Albrecht & D. Egret (Dordrecht:  Kluwer), 79
Finley, D., Jelinsky, P., Bowyer, S., & Malina, R.F. 1988, Appl.  Opt.,
	27, 1476
Finley, D., Jelinsky, P., Dupuis, J., & Koester, D. 1993, ApJ, in press
Gladstone, G.R., McDonald, J.S., Boyd, W.T., & Bowyer, S. 1993, BAAS, 25, 862
Helou, G., Madore, B.F., Schmitz, M., Bicay, M.D., Wu, X., & Bennett, J. 1991,
	in Databases and On-line Data in Astronomy, ed. M.A. Albrecht & D.
	Egret (Dordrecht:  Kluwer), 89
Lewis, J.W. 1993, JBIS, 46, 346 
Lewis, J.W., Girouard, F.G., & McDonald, J.S. 1993, IEEE Trans. Softw.  Eng.,
	in preparation
Lieu, R., Bowyer, S., Sumner, T.J., Sidher, S. D., & Lampton, M. L. 1993,
	Nature, submitted.
Marshall, H.L. 1985, ApJ, 289, 457
Marshall, H.L., Carone, T., & Fruscione, A. 1993, ApJL, in press
Moos, H.W., et al. 1993, BAAS, 25, 828
NASA Research Announcement #93--OSSA--02.
Patterer, R.J., Vedder, P.W., Jelinsky, P., Brown, A., & Bowyer, S. 1993,
	ApJL, in press
Pounds, K.A., et al. 1993, MNRAS, 260, 77 
Siegmund, O.H.W., Lampton, M., Chakrabarti, S., Vallerga, J.V., Bowyer, S.,
	& Malina, R.F. 1986, in Instrumentation in Astronomy VI, ed.  D.L.
	Crawford, Proc. SPIE, 627, 660
Stringfellow, G., & Bowyer, S. 1993, in IAU Circ. 5803, V1974 Cygni
Vallerga, J.V., Vedder, P.W., & Siegmund, O.H.W. 1992, in Multilayer and
	Grazing Incidence X-ray/EUV Optics for Astronomy and Projection
	Lithography, ed. R.B. Hoover & A.B. Walker Jr., Proc. SPIE,  1742, 392
Vallerga, J.V., Vedder, P.W., & Welsh, B.Y. 1993, ApJL, in press
Vedder, P.W., Patterer, R.J., Jelinsky, P., Brown, A., & Bowyer, S. 1993,
	ApJL, in press
Vedder, P.W., Vallerga, J.V., Gibson, J., Stock, J., & Siegmund, O.H.W. 1992,
	in Multilayer and Grazing Incidence X-ray/EUV Optics for Astronomy
	and Projection Lithography, ed. R.B. Hoover & A.B. Walker Jr., Proc.
	SPIE, 1742, 486
Vennes, S., Dupuis, J., Rumph, T., Drake, J., Bowyer, S., Chayer, P., &
	Fontaine, G. 1993, ApJL, 410, 119
Warren J., Vallerga, J.V., Mauche, C.W., Mukai, K., & Siegmund, O.H.W. 1993,
	ApJL, in press
Welsh, B.Y., Vallerga, J.V., Jelinsky, P., Vedder, P.W., Bowyer, S., & Malina,
	R.F. 1990, Opt. Eng., 29, 752



TABLES:
                       TABLE 1:  10% Filter Bandpasses 
	----------------------------------------------------------------
	| Instrument      Filter      lambda_peak (A)     Bandpass (A) |
	----------------------------------------------------------------
	|  Deep Survey    Lexan/B           91               67-178    |
	|                  Al/C            171              157-364    |
	| Scanner A & B   Lexan/B           89               58-174    |
	|                 Al/Ti/C          171              156-234    |
	|  Scanner C      Ti/Sb/Al         405              345-605    |
	|                  Sn/SiO          555              500-740    |
	----------------------------------------------------------------




		      Table 3:  BSL Sources by Filter Detection 
		----------------------------------------------------
		!             Scanners            |   Deep Survey  |
		! Lexan/B  Al/Ti/C  Dagwood  Tin  |  Lexan/B  Al/C |
		----------------------------------------------------
		!   327       92       17     11  |    39       6  |
		----------------------------------------------------




		   Table 4:  BSL Sources by Classification
		---------------------------------------------
		!     Number     Classification             |
		---------------------------------------------
		!      172       Late-type stars            |
		!      117       White dwarfs               |
		!       14       Cataclysmic variables      |
		!       10       Extragalactic objects      |
		!        6       A stars                    |
		!        7       B stars                    |
		!       15       Other                      |
		!       15       No ID                      |
		!      356       Total BSL Sources          |
		---------------------------------------------



FIGURE CAPTIONS:

Figure 1:  EUVE effective areas for the Lexan/B (filled squares) and Al/C
	(filled circles) filters in the Deep Survey telescope, Lexan/B
	(filled squares) and Al/Ti/C (filled diamonds) filters in each one
	of the short wavelength scanners, and Dagwood (asterisks) and tin
	(filled hexagons) filters in Scanner C.  The points mark the cali-
	bration values obtained prelaunch.  For the Deep Survey and the
	short wavelength scanners, the solid line represents the best-fit
	model from in-flight measurements. No in-flight measurements are
	available yet for the scanner C filters.
 
Figure 2:  Aitoff projections in equatorial coordinates (zero hours at the
	center and increasing toward the left) of the exposure maps for
	(a) the all-sky survey Lexan/B and Al/Ti/C filters; (b) Ti/Sb/Al
	and Sn/SiO filters; and (c) deep survey Lexan/B and Al/C filters.
	The figures show contour levels of equivalent exposure times for
	(a) 500, 1000, 2000, and 5000 s; and (b) 275, 500, 1000, and
	2500 s where the highest contour for the all-sky survey exposure
	maps is the one near the ecliptic pole.  The higher exposure around
	the ecliptic poles is a consequence of the survey geometry.  In the
	deep survey exposure map, the contour shown is for 2500 s, and the
	exposure times reach up to 25000 s within the contour.  Gaps with
	no exposure result from the periods when the survey was stopped to
	carry out calibration.
 
Figure 3:  Raw (bottom) and PSF-convolved (top) images for the pigeonhole
	corresponding to the white dwarf source EUVE J0623-37.6.  The source
	is confirmed as detected in the Lexan/B, Al/Ti/C, and Dagwood filters
	from simultaneous visual inspection of the raw and convolved data.
 
Figure 4:  Aitoff projections in galactic coordinates showing the location of
	the EUV sources in the BSL over the sky.  White dwarf stars (filled
	circles), late-type stars (filled triangles), other classes of sources
	(crosses), and sources with no current possible identification (open
	squares) are plotted.
 
Figure 5:  Plot of the EUV flux vs. visual magnitude for the following
	classes of objects:  white dwarf stars (filled circles), late-type
	stars (filled triangles), cataclysmic variables (empty squares),
	extragalactic objects (empty circles), and other objects (crosses).

Figure 6:  Coverage functions for each EUVE bandpass, given as fraction
	of the sky surveyed as a function of count rate.  Curve with open
	circles:  Lexan/B band; solid stars:  Al/Ti/C band; open stars:
	Dagwood; solid circles:  tin band.

Figure 7:  Differential coverage function for the Lexan/B band, in bins of
	0.075 dex.
 
Figure 8:  Sky coverage functions for the Lexan/B band derived by two dif-
	ferent methods.  Solid curve w/ open circles:  method using grid
	sources; thin dashed curve:  method using source count rates from
	this paper.  The two are reasonably close, so the grid source method
	may be used for each bandpass, even when there are very few detected
	sources.
 
Figure 9:  Number count distribution computed using Equation 3 (bold solid
	line) and WFC (thin dashed line) surveys.  Because the instruments
	have different effective areas, the WFC curve (bold dashed line)
	has been shifted upward by a factor of 2.1 (determined from count
	rates of mutually observed white dwarfs) for comparison.  The agreement
	between the shifted WFC and the EUVE curves is very good for count
	rates greater than 0.05 count/s.
 
Figure 10:  Same as Figure 9 for all of the EUVE wavelength bandpasses.
	Lexan/B:  bold solid line; Al/C:  bold dashed line; Dagwood:
	dashed-dotted line; tin:  solid thin line.

[HomePage] [Email] [Search] [Glossary]

Page created by webmastr@cea.berkeley.edu
Last modified 10/14/98