What controls whether a reaction is S_N1 or S_N2? The mechanism will probably be the one which has the higher rate (kinetics rather than thermodynamics because the products are the same in both cases), and there are many factors which can affect the rate of each reaction pathway; the nucleophile, the leaving group, the solvent, and the steric requirements of the reactant.

The Nucleophile

In the rate equation for an S_N1 reaction the nucleophile is not present because the RDS depends solely upon the starting material dissociating - therefore the nucleophile cannot affect the rate of the reaction.However, in S_N2 the rate depends directly on the concentration of the nucleophile, so clearly this will be affected by its nature.

Nucleophiles can be classified according to their power, or nucleophilicity - then obviously as the nucleophilicity is increased, the reaction rate will increase. Unfortunately it is not completely true to say that nucleophilicity will simply parallel the availability of the electron pair - in other words basicity does not necessarily parallel nucleophilicity.

The distinction is that basicity is a thermodynamic, or equilibrium, term whereas nucleophilicity is usually a kinetic one. This is something that needs to kept in mind, but on the whole, particularly for a given nucleophilic atom, basicity can parallel nucleophilicity, so a stronger base will be a stronger nucleophile. A good example of this is OH^- versus SO₄^2- : OH^- is a base, and a much stronger nucleophile than the delocalised conjugate base SO₄^2-.

Another area is which nucleophilicity can be distinguished is well exemplified by the halogens, whose nucleophilic order is;

I^-    >    Br^-    >    Cl^-    >>    F^-

So, as the halogen gets larger and more polarisable, its nucleophilic ability increases.

The explanation for this is that firstly as the electron cloud becomes more diffuse and less tightly bound to the nucleus, bonding can occur at greater internuclear distances, and secondly the larger ions are much less strongly solvated than the smaller ones.

Therefore, the large and weakly solvated I- has much higher nucleophilicity than the small and strongly solvated F^- to such an extent that F^- is practically never observed acting as a nucleophile.

Another way of expressing an element's size and polarisability is the hard and soft classification, so in general, soft nuclei are better nucleophiles than hard nuclei (cf Sulphur much better nucleophile than Oxygen, Phosphorus better than Nitrogen).

The Leaving Group

The nature of the departing group will affect both S_N1 and S_N2, because in both the RDS involves breaking the bond to it. Several factors could be expected to influence the breaking of the carbon-LG bond - in no particular order;
1. The stability of LG^-.
2. The solvation stabilisation of LG^-.
3. The strength of the bond.
4. The polarisability (not polarisation) of the bond. 
(Clearly the first two are related.)

The influence of these factors can be identified by analysing the reactivity of the alkyl halides for a given nucleophile (and this applies for both S_N1 and S_N2):

R-I    >    R-Br     >    R-Cl    >    R-F

This series parallels increasing reactivity with decreasing R-X bond strength and polarisability, not solvation stabilisation (the smaller halides are more solvated than the larger - this would lead to the series running the other way).

Stability of the anion is a different matter - in general this also parallels increased reactivity, so that the weaker LG^- is as a base, the better a leaving group it is. Thus the conjugate bases of strong acids make very good leaving groups and enhance the rate of nucleophilic substitution for both S_N1 and 2 - examples of very good leaving groups are;

Though not strictly part of this sub-heading, another important consideration to be made in S_N1 reactions is that of the stability of the carbocation formed in the RDS.

Although most carbocations are very reactive, some are more stable than others, and the formation of these in the course of an S_N1 reaction will enhance the rate considerably. Examples include allylic systems which are resonance stabilised, more substituted centres (inductive effects and hyperconjugation) and any other systems in which electrons can be donated to the carbocation centre.

Solvent

The characteristics can be very important when considering what mechanism occurs and why. To go further, it is the polarity of the solvent which is key. A highly polar solvent, e.g. H₂O, will give increased solvation stability to any ions present, as compared to a non-polar solvent, e.g. pentane. This has different consequences for S_N1 and S_N2.

S_N1: The RDS of the S_N1 mechanism involves formation of a pair of ions, so if the solvent is polar and can solvate the ions, the rate of this step will be increased, and thus the rate of the overall reaction will be increased.

S_N2: In contrast to S_N1, the TS involves a slight decrease in overall charge, and a definite delocalisation of it, as the negative charge is spread across the nucleophile and the leaving group. This means that increasing the solvent's polarity will cause a slight decrease in the reaction rate.

A far more marked rate effect is observed when protic (those capable of hydrogen-bonding via a hydrogen) and aprotic (those not) solvents of similar polarity are compared because the nucleophile may well hydrogen-bond to a protic solvent and dramatically reduce the rate. This effect is so large that even transferring to a much more polar yet aprotic solvent from a protic one can hugely increase the rate.

For example the rate of S_N2 by N₃^- on MeI is considerably faster in DMSO (dimethyl sulphoxide) (aprotic, but very polar) than in methanol (protic, but about 2/3 the polarity) because the N₃^- nucleophile hydrogen-bonds to methanol but not DMSO. In short, the nucleophile will be much more effective if it is not surrounded by a cage of solvent molecules.

Steric Considerations

Again it is the difference in rate determining steps which causes the rates of the two mechanisms to behave differently with steric change.

S_N1: In the rate determining step of the S_N1 mechanism, an sp² hybridised carbon centre is formed from an sp³ hybridised carbon centre. In the sp³ centre the angle between neighbouring substituents is 104.5^o whereas in the sp² centre, the same angle is 120^o.

Clearly the substituents are able to lie further away from each other in the carbocation, so if they are large this will give some degree of steric relief in the intermediate which will enhance the rate of the S_N1 process. So, in the example below, there is steric relief when the three bulky phenyl substituents are allowed to move further apart in the RDS, and thus the rate is enhanced relative to the same S_N1 process with MeI rather than Ph₃CI. (Also, the Ph₃C⁺ cation has stabilisation from the phenyl rings.)

S_N2: In the rate determining step of the S_N2 mechanism, the nucleophile has to attack the electrophilic centre at the back of the leaving group (so-called 'backside attack'). To do this it has to pass through three potential sterically inhibitive substituents - so taking the above example, there are three bulky phenyl groups which sterically protect the electrophilic centre, thus reducing the rate of S_N2.

The combination of all these factors - steric relief and cation stability in S_N1, steric hindrance in S_N2, means that the above reaction is very likely to follow S_N1. At the other end of the scale, MeI with little steric relief and no cation stability in S_N1, and little steric hindrance in S_N2 - is therefore very likely to follow S_N2.