The H particle, a dibaryon with the quantum numbers of two
's in the 
state ( 
, electrically
neutral), is the most promising candidate for a deeply bound six-quark
state. The S=-2 sector plays a special role in the spectrum of
six-quark bag states, since only a six-quark system with 2u, 2d,
and 2s quarks can exist in an SU(3)-flavor singlet configuration
with spin zero, which takes maximum advantage of the attraction due to
color-magnetic interactions arising from one-gluon exchange (OGE).
Within the context of the MIT Bag Model, Jaffe was the first to
note that the H could be deeply bound with respect to the
strong decay threshold [1].
Since Jaffe's pioneering
work, there have been a number of attempts to refine the Bag Model
calculation by including SU(3) breaking, center of mass corrections,
etc. (see Ref. [31], for example). The problem of H binding has
also been treated in various versions of the Skyrme soliton
model [32, 33, 34], in lattice
QCD [35], and
in hybrid quark/gluon plus meson exchange models [36, 37].
Hybrid models in which the short
range behavior is treated perturbatively via OGE generally yield a
bound H, whereas if non-perturbative instanton effects are included
at short distances, the H is pushed above the
threshold [38]. The repulsive three-body interaction
generated by instantons only operates in the SU(3)-flavor channel,
i.e., the H, and does not enter in baryon spectroscopy or
nucleon-nucleon scattering. The most recent lattice QCD calculation, of
course a non-perturbative result,
yielded a very deeply bound H, near the mass of two neutrons [35].
Thus the various theoretical speculations range from a deeply bound H
(binding energy of order 
MeV), as in
lattice QCD calculations, to a loosely bound or even unbound object,
as in various meson exchange models with some treatment of quark/gluon
degrees of freedom at short distances. Clearly a sensitive H search
is called for, in order to shed light on the fundamental question of
the existence of strange dibaryons.
Several H searches are underway, most notably via the 
double strangeness exchange reaction at the Brookhaven AGS [39].
In one version of the experiment, the 
reaction is used to tag the production of the 
hyperon, which
is then captured at rest via the two-body process 
.
The final state neutron is detected, rather than the H. In a second
version, the 
reaction is studied.
Theoretical estimates exist for these cross
sections, which are typically of order of a few tenths of a 
for the latter reaction [40, 41].
The 
beam intensities at the AGS
Booster are sufficient to measure such cross sections, but one does
not have orders of magnitude in sensitivity to spare. In contrast,
the H dibaryon is expected to be copiously produced in high energy
heavy ion collisions, with estimated rates of order 
to
per central Si-Au collision at AGS energies, based on
several forms of the coalescence model [42, 43].
The H yield will be significantly higher in Au-Au collisions.
The experiments do not detect the H directly, so there
is no restriction on its weak decay lifetime. However, the
branching ratio is likely to be measurable only if
the H does not lie too far below the
threshold [41].
E864, on the other hand, is sensitive to
the H if its lifetime 
is of order 50 ns or longer, because
of the long flight time required. According to the lifetime estimates
of Donoghue [44], exceeds 10 ns only if the
H lies below the 
threshold (100 MeV below the
threshold). Thus E864 is sensitive to a deeply bound
H, and is nicely complementary to the H search via the reaction, which can detect a weakly bound H.
Experiment E864 can search for the H directly by using time of flight and calorimetry. The main problem is the background due to neutrons and antineutrons. The H might be expected to be produced (because of its mass) at larger transverse momenta than those characteristic of neutrons. The large acceptance and its disposition will allow the search to extend to transverse momenta of 1.5 GeV/c. An analysis of the sensitivity of the proposed experiment for the direct detection of the H is presented elsewhere.
Another approach to searching for the H in heavy ion collisions is
to look for the composites of the H which may well exist if the H
is stable against strong decay [45]. The forces between two
H's or between H and d are expected to be attractive and only a
small amount of nuclear attraction is required to bind such systems.
The 
bound state with 
, I=0, is an S=-4
analog to the 
particle. There is also the possibility that
three or more H's could be bound together.
The rate of production of an HH bound state is expected to be of the
order of 
to 
per central collision and the rate for
Hd should be greater, perhaps of the order of 
[45].
The latter can be estimated from the ratio
of t production to d production (which will be measured)
and the calculated ratio of H to
p production. If these composites exist, their production rates
will be within the discovery range of E864.
Experimentally, the H composites provide signatures which are easier to separate from the background than is the case for direct H detection. The HH state should be cleanly separated from neutron backgrounds. For example, with a typical Lorentz factor of 2.0, the energy in the calorimeter from a neutron will be 0.94 GeV while the energy in the calorimeter from the HH will be 4.0 GeV. An antineutron with the same velocity will deposit 2.81 GeV in the calorimeter.