The lack of an optical counterpart down to limits of 
and
(Steinle et al. 1987; White et al. 1987) implies an
X-ray/optical flux ratio 
.
The only known classes of galactic sources which can yield such a high
value are low mass X-ray binaries (LMXBs) and
isolated neutron stars.
A neutron star accreting from a low mass companion is the most likely
explanation for 4U 0142+614.
Coherent pulsations are rarely seen in LMXBs: the only known
examples among optically identified systems, 4U 1626-67, Her X-1 and
GX 1+4, have very different X-ray properties,
companion stars and evolutionary origins (see, e.g., White, Nagase &
Parmar 1995). The spin period of 4U 0142+614 is very similar to that of
4U 1626-67 (7.7 s, Rappaport et al. 1977), and it is interesting to note
that two other optically unidentified pulsars, which are likely accreting
from low mass companions, 1E 2259+586 and 1E 1048.1-5937
(Coe & Jones 1992; Mereghetti, Caraveo & Bignami 1992),
have periods of the same order,
6.98 and 6.44 s respectively (Davies et al. 1990; Corbet & Day 1990).
In order to derive some information from the
available optical limits, some considerations on the likely distance and
interstellar reddening of 4U 0142+614 are required.
The position of 4U 0142+614 is close (;SPMlt;0.5 
)
to that of two open clusters with well determined distances and
reddening: NGC 654 (2.5 kpc and A 
=2.67), and NGC 663 (2.1 kpc and
A =2.43) (see Leisawitz, Bash & Thaddeus 1989, and references
therein). The column density N 
cm 
derived from the power law spectral fits of 4U 0142+614 (White et al.
1987) corresponds to a higher absorption, A 
7 (Gorenstein
1975), hinting to a greater distance. However, 4U 0142+614 is not
necessarily much further than these clusters, since a part of its
absorption could be intrinsic to the source or due to a local (d;SPMlt;1 kpc)
molecular
cloud which is present in this region, as clearly visible on the POSS
prints. 4U 0142+614 lies near to the edge of this cloud, which does not
significantly affect NGC 654 and NGC 663 (Leisawitz, Bash, &
Thaddeus 1989). A distance of 4 kpc
would yield a 1-10 keV luminosity of 
erg s 
,
similar to 4U 1626-67. At this distance and reddening the faint
optical counterpart of the latter source
would be fainter than the present limits for 4U 0142+614.
On the other hand, an evolved companion similar to that of GX 1+4 or
Cyg X-2 (M -1, van Paradijs 1991) would have been detected even at
10 kpc (which for this direction is well outside the Galaxy).
A companion star similar to that of 4U 1626-67, i.e. either a main
sequence star with M 
or a white dwarf of 0.02
(Verbunt, Wijers & Burm 1990), is also compatible with the
limits on 
derived, which however
also allow for more massive companions. For instance, a hydrogen main
sequence star of 0.3 , would fill the Roche lobe for an
orbital period of 
hr, requiring i 
.
4U 0142+614 is similar to the unidentified source 1E 2259+586, whose
spectrum can be described by a power law with energy index 3, plus
some possible cyclotron features suggesting a magnetic field B
(Iwasawa, Koyama & Halpern 1992).
The energy spectrum of the Aug. 1984 EXOSAT ME observation was modelled by
a two power law model with photon index 
=3.2 and =1 for
the soft component (1-3 keV) and the hard component (4-11 keV),
respectively. The intrinsic absorption was worked out to be
and 
for the soft
and hard component, respectively. In the Nov. 1985 observation only the soft
part of the spectrum was present and fit with a single power law model
( =3.2; see Fig. A.4).
The possibility that 4U 0142+614 is an
isolated neutron star is suggested by its very high ,
its ultrasoft spectrum,
and the absence of significant variability on long timescales
(of course this possibility requires that the evidence for secular
spin-up is the result of chance detections in the 1985 and 1991 data).
In principle the X-ray emission could be due to non-thermal magnetospheric
processes powered by the rotational energy, to thermal emission from the
neutron star surface, or to accretion from the interstellar medium.
While examples of the first two mechanisms are well known
(see, e.g., Mereghetti, Caraveo & Bignami 1994),
no compelling evidence for a compact object accreting from the
interstellar medium has yet been found, despite several studies show that
such sources could be relatively common
(Treves & Colpi 1991; Blaes & Madau 1993).
G),
the available rotational energy loss is too small, unless
4U 0142+614 is at a distance of a few parsecs.
To investigate the possibility of thermal emission, we fitted blackbody
spectra to the 1985 ME+LE data, when the variable hard component was absent,
obtaining kT 
keV, a low interstellar absorption
N 
2-4 
cm
and a bolometric luminosity 
erg s
(however these fits were substantially worse than those with a power law:
reduced 
of 3-6). This implies an emission
region of 
,
compatible with a hot spot on the surface of a neutron star, possibly
the magnetic polar cap heated by accretion from the interstellar medium.
In this case the accretion induced luminosity would be
,
where v is the neutron star velocity
relative to the interstellar medium of density n
(Blaes & Madau 1993).
Thus a very small distance is again required for typical values of v
( 
) and of the local
interstellar medium density ( 
0.1-1 
).
The already mentioned molecular cloud (Leisawitz, Bash & Thaddeus 1989) could
be at a distance of only a few hundred parsecs
and easily provide the density required to power the
observed luminosity.
For accretion onto the neutron star to take place,
the magnetospheric centrifugal barrier must be open,
and the low rates implied by the above luminosity therefore require a magnetic
field 
G
(see, e.g., Stella et al. 1994).
Very recently new results were obtained by investigating the properties of the 4U0142+614 using new data obtained with the ASCA observatory and archival data from the Einstein and ROSAT observatories (White et al. 1996; see §A.2 for more details).