Table 4 contains the fundamental rate coefficient data. In this section, we discuss the general form of the data, describe the format of each entry, discuss particular reactions or class of reactions and describe major differences since the 1990 release.
The reactions and associated rate coefficients can be divided up into a number of categories, each of which has been ordered in terms of the mass number of the first reactant, and for a specific first reactant in terms of increasing mass number for the second reactant. The ratefile is organised as follows:
Table 4: Reactions and their rate coefficients
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Table 4: - continued
Each entry in Table 4 has the following form:
I, R1, R2, P1, P2, P3, P4, 
, Note
where I is the reaction number, R1 and R2 are reactants - R2 can be a
cosmic-ray proton (CRP), an interstellar photon (PHOTON), or a
cosmic-ray-induced photon (CRPHOT) - and P1 to P4 are reaction products, with
the entry format given by:
1X,A4,1X,4(1A7,1X),A3,1X,A3,1PE8.2,1X,0PF5.2,1X,F8.1,A9
For each reaction, 
and 
are used to calculate the
rate coefficient by:
for two-body reactions, where T is the gas temperature,
for direct cosmic-ray ionisation (R2 = CRP),
for interstellar photoreactions (R2 = PHOTON), where 
represents
the rate in the unshielded interstellar ultraviolet radiation field, 
is the extinction at visible wavelengths caused by interstellar dust,
is the parameter used to take into account the increased extinction
of dust at ultraviolet wavelengths, and
for cosmic-ray-induced photoreactions (R2 = CRPHOT), where 
is the
grain albedo in the far ultraviolet, typically 0.6 at 150 nm, 
the cosmic-ray ionisation rate and 
is the probability per cosmic-ray ionisation that the appropriate
photoreaction takes place.
Grain surface reactions are not included in this database since there is,
as yet, little consensus on the magnitude of the appropriate rate coefficients
nor, indeed, on the mode of reaction. One important omission is the formation
of 
which is known to take place on the surfaces of interstellar grains
and for which a reasonable estimate of the rate can be deduced both from
theory (Hollenbach & Salpeter 1972) and
observation (, Jura 1975a,b).
`Note' is a nine-column entry which gives information on the type and source of the data and has the form:-
Neutral-neutral reactions are usually studied experimentally at room
temperature and above and therefore application of laboratory-determined
rate coefficients to the low temperature environments of interstellar
clouds is fraught with uncertainty. For example, it is possible that
several reactions listed in the ratefile as not possessing an activation
energy do, in fact, have small barriers (
) which are not evident
in measurements done at room temperature and above.
are the only neutral products; CR, cosmic-ray ionisation;
CT, charge transfer;
DR, dissociative recombination with electrons; HA, hydrogen abstraction;
MN, mutual neutralisation;
NA, neutral radiative association; NE, neutral exchange; NI, negative ion-neutral;
PD, photodissociation;
PI, photoionisation; PM, photodetachment of electron; PN, positive ion-neutral;
PT, proton transfer; RA, radiative association between a positive ion and a
neutral molecule; RM, radiative electron (minus ion) attachment;
RR, radiative recombination with electrons.
and 
, which have
rapid reactions with 
. The low abundance of all other species X
relative to 
in many astronomical regions, means that all ion-X
reactions can be excluded. This criterion is only useful if 
is
the dominant form of hydrogen. If a reaction should be included in any
comprehensive model it is labelled with an `A'.
Major changes to the original database include:-
and 
.
and 
.
These reactions are taken from Herbst & Leung (1989).
. 
and 
denote the cyclic
species
and its proton and 
denotes the linear cumulene,
, and
its protonated form.
, the so-called Prasad-Tarafdar mechanism
(Prasad & Tarafdar 1983), with rates taken from Gredel et al. (1987, 1989).
(Herbst & Leung
1990), and a more detailed chemistry for vinyl cyanide,
, based
on the laboratory work of Petrie et al. (1992).
It should be noted that while the reaction set is closed, i.e. all species
have at least one formation and destruction reaction, several of the hot core
species are probably formed by grain surface reactions since they do not
have efficient gas-phase routes to their syntheses. As a result, the
species 
, and 
and their ions should not
be included in a purely gas-phase scheme, since they have no
gas-phase formation routes. In hot molecular core models, they are
assumed to be present with some initial abundance determined by grain
surface reactions.
The original rate file (Millar et al. 1991b) was ordered slightly differently in that the first reaction block listed those reactions possessing activation energies. The rationale was that in a particular low-temperature application, this block of reactions could be edited out. However, it turns out that a number of reactions previously thought to possess activation energies do not and vice versa. In addition, the speed of modern computers means that it is not much more costly to compute a chemical model with all reactions in the rate file. We have therefore gathered all neutral-neutral reactions in the first block.
Additional reactions to this section include those associated with
phosphorus chemistry, e.g. 
(n = 1-4), taken from
Millar (1991)
and large carbon-chain chemistry, in particular reactions involving
atomic oxygen (Herbst & Leung 1989) and atomic carbon ( Haider &
Husain 1993a,b; Clary et al. 1994),
the chemistry of 
(Marston et al. 1989), and low-temperature reactions
of CN with several neutral molecules ( Sims et al. 1992, 1993a,b). Several of the
reactions
involve non-conservation of spin and may have rate coefficients smaller
than tabulated or possess activation energy barriers. In some systems,
the presence of low-lying electronic states may alter the products and
rate coefficients.
Much effort has gone in to evaluating the neutral-neutral rate coefficients and
in trying to include as many experimental determinations as possible. We
have therefore searched the National Institute of Standards and Technology
(NIST) Chemical Kinetics Database - Version 6 (Mallard et al. 1994) and
included what we believe to be the most appropriate form of the rate
coefficient. It is important to remember that neutral-neutral reactions are
usually studied at room temperature and above so that application of
laboratory-determined rate coefficients to low-temperature interstellar
cloud models is often insecure. For example, it is possible that several
reactions listed as having no activation energy do, in fact, have small
barriers (
) which are not evident in measurements performed at
room temperature. In addition, some reactions in the NIST Database
are best characterised by a negative activation energy barrier since their
rate coefficients increase with decreasing temperature. In order to prevent
a serious over-estimate of these rate coefficients at 10 K, we have generally
preferred to adopt an alternative, although still accurate, form for the
rate coefficients, for example, one involving a power-law dependence on
temperature. Normally, the NIST Database contains such alternative
formulations. Roughly one half of the neutral-neutral reactions have
been studied in the laboratory.
Finally, it is important to note that rate coefficients for the collisional dissociation (CD) reactions are dependent on both density and temperature and reference to the appropriate values to use in particular circumstances should be made to the original papers (Roberge & Dalgarno 1982, Dove & Mandy 1986).
This block includes the reactions of some negative ions as well as the more usual positive ion-neutral reactions. Around one-third of the 2800 reactions have experimentally determined rate coefficients, with several tens of rate coefficients studied at temperatures less than 100 K. The accuracy of the remaining rate coefficients is fairly high since theoretical methods for determining these depend chiefly on long-range forces and are very reliable for exothermic systems. At temperatures less than about 50 K, the presence of a neutral having a large electric dipole increases the rate coefficient (Adams et al. 1985). As mentioned above, there are various approximations which can be used to derive the low-temperature rate coefficients for such systems.
There are some limitations on the data presented here. For example, the
radiative association reactions are labelled `M', for measured, although in
fact it is their three-body analogues which have been studied experimentally. The
association rate coefficients are derived theoretically from these rates and
are parameterised in a form which is valid only for the 
temperature
range in general. At higher temperatures, different values and different
temperature dependencies can apply. Furthermore, some important reactions,
in particular the 
and 
reactions, have rate
coefficients not easily approximated by the format in Table 4. Such reactions
are discussed in detail by Millar et al. (1991b).
This block includes electron attachment (of H, C, O and S) as well as
radiative and dissociative recombination of positive ions. An increasing
number of dissociative recombination reactions have been studied in the
laboratory (Herd et al. 1990; Amano
1990; Adams et al. 1991; Mitchell 1990;
Canosa et al. 1991) and reviews have been given by Adams (1992) and several in
a conference proceedings (Rowe et al. 1993). Branching
ratios still remain an area of uncertainty in most cases. The dissociative
recombination rate coefficient of 
remains a matter of some debate
(Amano 1990; Canosa et al. 1991;
Smith & Spanel 1993; Sundstrom et al.
1994) although there
has been a significant narrowing of the differences between rate coefficients
measured in several laboratories. In general, calculated molecular abundances
are not very sensitive to the total dissociative recombination rate
coefficient but are sensitive to the adopted branching ratios.
We have added a number of photoreactions taken from the compilation of
Roberge et al. (1991), scaled to the
interstellar radiation field determined by Draine (1978). Photorates should
be used with care for a number of reasons. Some species, most importantly
and CO, dissociate via absorption of line radiation and thus self-shield.
This is not included in the rate file and approaches such as those discussed
by van Dishoeck & Black (1988), Lee et
al. (1996) and Warin et
al. (1996) need to be incorporated into chemical models.
In addition, the intensity of ultraviolet radiation at any point in a cloud
is determined by the properties of the small interstellar dust grains which
control the transfer of radiation. Both the pre-exponential (
) and
the exponential (
) factors can be different for different grain
populations (see Roberge et al. 1991). In some cases, a bi-exponential
formula is to be preferred (see van Dishoeck 1988 for a discussion of this
point). Finally, the rates are given for the standard interstellar
radiation field incident on a slab and need to be re-evaluated
when the field in a
particular application is not a simple scaling of the interstellar field
(Spaans et al. 1995) or when the cloud structure is clumpy (Boissé 1991).
The Prasad-Tarafdar mechanism generates an internal source of UV photons
in interstellar clouds. This source of photons becomes important in regions
in which the extinction of the external field is large. We have taken the
appropriate probabilities, (
), of reaction per cosmic-ray ionisation from
Gredel et al. (1987, 1989) and Rawlings (1992) who has calculated
photoionisation rates for some metal ions. Note that the value of 
for
CO is dependent on the details of the cloud model (temperature and line-width).
The form adopted here is a fit to the temperature-dependent values listed
by Gredel et al. (1987).
