One effective test for discriminating S_N1 from S_N2 reactions (that is, for detecting the presence of an intermediate), is sensitivity of the rate to "common-ion rate suppression". A good example comes from a 1940 paper by Bateman, Hughes, and Ingold at University College, London.

They reported the rate of conversion of R-Cl to R-F in the following system:

They chose the unusual solvent SO₂, which is a little less polar than acetone, because it made it easy to monitor the reaction by the change in electrical conductivity of the solution as the F^- salt is converted into Cl^- salt. The chloride is better at conducting electric current because the positive and negative ions separate more easily. (Why?)

For reasons that will become clear the rate of reaction falls off much more rapidly than is predicted by either first- or second-order rate laws, so they measured the initial rate of the reaction and expressed it as a pseudo first-order rate constant, "k" for disappearance of RCl. That is, "k" is the initial rate of the reaction divided by [RCl].

Here are their key results:

Expt #	[RCl] (M)	[Me4N+F-]	[Me4N+Cl-]	105 "k" (/sec)
I	0.067	0.003	-	2.4
II	0.068	0.050	-	7.7
III	0.066	0.005	0.0065	0.14
IV	0.065	0.005	0.0399	0.05

First compare Expt II with Expt I:

Adding fluoride speeds the reaction. Increasing the concentration of the fluoride salt 16-fold increases "k", but by only 3-fold.

For an SN1 process, which this turns out to be, one might think that the rate should not have changed with the increase in [F^-].
If it had been a SN2 process, the rate should have increased 16-fold. How to explain this intermediate behavior?

The solvent containing more salt is more polar. The presence of other ions in the solvent makes it easier to ionize RCl, because the ions being formed can be stabilized by the presence of complementary ions in the neighborhood, just as they are stabilized by favorable orientation of a polar solvent molecule. Thus increasing the amount of salt makes the SN1 reaction faster, but not as much faster as if the F^- were a nucleophile in an SN2 reaction. This modest dependence on F^- supports an SN1 mechanism.

The conclusion we need from these experiements is that a salt concentration of 0.05 M should give a "k" of 7.7 x 10^-5/s.

Now consider Expt III and especially Expt IV:

Adding Cl^- slows the reaction; adding a lot slows it drastically.
In Expt IV, the total salt concentration (chloride + fluoride) is 0.045 M, so one might anticipate about the same "k" as in Expt II. In fact the rate is 154 times slower.
Why does Cl^- retard the reaction with F^-?

If these had been SN2 reactions, adding a second nucleophile could only accelerate reaction, by providing a new pass out of the starting material valley on the potential energy surface. It would do nothing to slow escape through the original transition state.

(In time a second reaction might slow the first reaction by depleting the starting material, but it would not change the initial rate or the rate constant for the first reaction. In this case the starting material is not depleted by the reaction with added chloride, because this reaction is just substitution of one chloride for another, so there is certainly no reason to slow an S_N2 substitution by fluoride.)

In an SN1reaction, the rate of forming intermediate is (nearly) independent of added nucleophile. (We say nearly because of the influence on solvent polarity mentioned above.) But the products from the intermediate definitely depend on what nucleophile(s) are present.

When Cl^- is present it can divert some of the intermediate to RCl, which happens to be the starting material. That is why Cl^- is called the "common ion", it is present both in the nucleophile and the starting material. This definitely slows formation of RF.

In Expt. IV only one intermediate cation in 154 makes it to RF, the other 153 react with Cl^- and return to starting RCl. Hence the reaction is slowed 154 fold. In this experiment there is 8 times as much Cl^- as there is F^-, so on an ion-for-ion basis Cl^- must be about 20 times (153/8) more reactive as a nucleophile than F^-.

Note that this common-ion rate suppression would appear as the reaction proceeds even without adding Cl^- at the outset, because Cl^- is one of the products of substitution on RCl.

If you startedwith equal concentrations of RCl and F-, you might have expected the reaction to be 1/2 as fast when the starting materials are half gone (for SN1) or 1/4 as fast (for SN2). In fact when the starting materials are half gone, there is as much Cl^- product as there is remaining F^- nucleophile, and the Cl^- ion is 20 times as reactive as F^-. The SN1 cation intermediate is being formed half as fast as initially, but it is only going to product one time in 21, so RCl disappearance is only 1/42 times as fast as it was initially.

This is why the reaction slows more rapidly than either simple SN1 or SN2 kinetics would predict.

Generally the rate of SN1 reactions is determined in the first step (cation formation) and the products are determined in the second step (cation reaction). So adding an additional nucleophile (or base) can change products without changing rate. Common-ion rate suppression is a special case, where one of the products happens to be starting material.

Remember that our derivation of ideal first-step-rate-limiting behavior required that the second step be much faster than the reverse of the first step. This condition is violated under conditions of common-ion rate suppression.

The paper was submitted April 18th from University College, London. In the weeks between submission of the manuscript and its appearance in print, Holland, Belgium, and France fell to Germany and the British Expeditionary Force was evacuated from Dunkirk. When the Blitz began in September, Hughes sent his family away and slept in the Chemistry Department. This was fortunate because one morning only one wall of his flat was found standing