# Selectivity in organic reactions

I am finding great difficulty in understanding the differences between stereospecific, stereoselective, regiospecific and regioselective nature of reactions. Can you explain it in simple terms without mixing them with each other?

• Wikipedia has a page describing the difference between stereospecific and stereoselective. The website Mastering Organic Chemistry also has a section describing the difference between specific and selective in simple terms. – Tyberius Apr 17 '17 at 0:52

Yours is a common source of confusion. Reactions are can be selective but they are not specific. What's specific is a mechanism for the reaction. If the reaction only proceeds via one mechanism, then it is extremely selective.

1. Stereospecific. The mechanism provides a specific stereochemical outcome or proceeds from a specific stereochemical configuration. Examples:

a. Pd-catalyzed hydrogenation and hydroboration proceed via syn addition. These mechanisms are stereospecific because they indicate always adding to the same side of an olefin.

b. $\mathrm{E}_{2}$ elimination processes require that the hydrogen and leaving group are anti-periplanar. This mechanism is stereospecific.

c. Acid-catalyzed hydration proceeds via a planar carbocation intermediate where previous stereochemistry might be lost. This mechanism is not stereospecific.

2. Stereoselective. The reaction provides more of a product with a specific stereochemistry than any of the other possibilities.

a. Allylic alcohol directed epoxidation reactions frequently yield different amounts of the two product epoxides that result from reacting on the different faces of the olefin. This reaction is selective.

b. Asymmetric reactions that rely on pathways through diastereomeric transitions states are stereoselective.

c. Acid-catalyzed hydration could be stereoselective because the molecule can have other stereocenters that provide diastereoselectivity, even though the base mechanism disregards stereochemistry.

3. Regiospecific. The mechanism indicates a certain connectivity that should be obtained in the product.

a. Exo- versus endo- intramolecular reactions have regiospecific mechanisms.

b. A bit of a stretch, but carbonyl addition happens at the carbon. The mechanism is regiospecific.

4. Regioselective. The reaction favors one connectivity pattern over the other.

a. Hydrobromination favors the Markovnikov product, while hydrobromination in the presence of peroxides favor the anti-Markovnikov product.

b. In intramolecular ring closing reactions, 5-exo-tet products are preferred over 6-endo-tet products.

c. Carbonyl addition happens at the carbon. The only mechanism in operation is the one mentioned above, so the reactions are regioselective.

Importantly, reactions should not be stereospecific or regiospecific, even if they give only one product. This is a concrete case of a reaction proceeding exclusively (or almost exclusively) via a single mechanism that is stereospecific or regiospecific, respectively. The reaction itself is, respectively, stereoselective or regioselective.

In short, "xyz selective" is the more general case with one preferred outcome of the reaction, over the other one. On contrast, "xyz specific" describes a subset of the former, where only one -- but not the other -- outcome occurs, and consequently is the only observed.

Regarding the regioselectivity, you may encounter for example the reactions of nucleophiles on $\alpha,\beta$ insaturated ketones. A hard nucleophilic reagent (in terms of HSAB) will show preference to attact on the carbonyl carbon, a soft one in $\beta$-position. If there is a preference for a reaction to occur on one site over other sites, it is regioselective.

The beauty (or, the hassle) with organometallic reagents is that depending on the hardness / softness of the metal deployed, you may tune the regioselectivity of your reagents. (Other conditions like additional spectator ligands, Lewis basicity of the solvent, temperature etc., have some influence, too. Take these two examples with a grain of salt, please.)

To account for NotBaran's comment, IUPAC' Gold Book specifies regioselectivity as this:

regioselectivity (regioselective)

A regioselective reaction is one in which one direction of bond making or breaking occurs preferentially over all other possible directions. Reactions are termed completely (100%) regioselective if the discrimination is complete, or partially (x%), if the product of reaction at one site predominates over the product of reaction at other sites. The discrimination may also semi-quantitatively be referred to as high or low regioselectivity. (Originally the term was restricted to addition reactions of unsymmetrical reagents to unsymmetrical alkenes.) In the past, the term 'regiospecificity' was proposed for 100% regioselectivity. This terminology is not recommended owing to inconsistency with the terms stereoselectivity and stereospecificity.

Source: PAC, 1994, 66, 1077 (Glossary of terms used in physical organic chemistry (IUPAC Recommendations 1994)) on page 1160

Regarding stereoselectivity, consider as an example a reduction of a ketone. It may be that a reagent is performing this reaction like lithium borohydride, and the product is a mixture of both stereoisomers possible; in this case (absence of Cram conditions) even to a racemate. With help of a suitable chiral borane (derived from $\alpha$-pinene, as one example), however the reaction may yield preferentially one stereoisomer, over the other. (Here, the substrate already is prochiral.) Hence, it is a stereoselective reaction.

Note: I do not remember if in the instance of the last reaction of acetophenone the ($R$)- or the ($S$)-configurated enantiomer is the correct one. The point to retain is, yielding one (or the other) product makes a difference, energetically / kinetically speaking.

Again, to account for @NotBaran's comment, IUPAC Gold Book describes the current definition here:

stereospecificity (stereospecific)

1. A reaction is termed stereospecific if starting materials differing only in their configuration are converted into stereoisomeric products. According to this definition, a stereospecific process is necessarily stereoselective but not all stereoselective processes are stereospecific. Stereospecificity may be total (100%) or partial. The term is also applied to situations where reaction can be performed with only one stereoisomer. For example, the exclusive formation of trans-1,2-dibromocyclohexane upon bromination of cyclohexene is a stereospecific process, although the analogous reaction with (E)-cyclohexene has not been performed.

2. The term has also been applied to describe a reaction of very high stereoselectivity, but this usage is unnecessary and is discouraged.

Source: PAC, 1994, 66, 1077 (Glossary of terms used in physical organic chemistry (IUPAC Recommendations 1994)) on page 1167
PAC, 1996, 68, 2193 (Basic terminology of stereochemistry (IUPAC Recommendations 1996)) on page 2219

Both the references of IUPAC's Gold Book, as their references in Pure and Applied Chemistry may be accessed freely:

• "Again, if your reaction yields predictably only one of the two possible enantiomers, it were stereospecific." This isn't strictly true. A stereospecific reaction is something that mechanistically must give a specific product. Many asymmetric reactions give a predictable major product, but this is merely a high degree of stereoselectivity, not specificity (very few things are truly stereospecific). Essentially stereoselectivity refers to an energy barrier between two TS's (i.e. R or S in a chiral reduction), but its not mechanistically possible to form both, one is just energetically bad. – NotEvans. Apr 16 '17 at 21:45
• @NotCorey I do not see a contradiction between "yielding predictably only one of two possible outcomes" by mine and the "it must give (by mechanism) one and only one outcome" by you. For me, "yielding predictably only one of two possible outcomes" is not preference of one over the other (leading into a situation like "A was obtained in 99% ee", including there is a tiny bit of B, too). I understand it as "only A, but not B is the product". That specific reactions are lesser in number than selective ones in the flask, well, ok. – Buttonwood Apr 16 '17 at 21:57
• "yielding predictably only one of two possible outcomes" by mine and the "it must give (by mechanism) one and only one outcome" That is the exact contradiction. The terms are well defined in textbooks and the literature- just because a highly selective reaction appears specific does not make it so based on the accepted nomenclature. Just because there is a logical correlation doesn't make it acceptable to interchange the terms (sorry to be pedantic, but you cant change definitions just because they appear to fit) – NotEvans. Apr 16 '17 at 22:30