Thermodynamic versus Kinetic Control

Cool 1,3-butadiene with HBr to −80 °C and you get mostly one product. Let the same mixture warm to 40 °C and stand. The product ratio inverts. Nothing about the starting materials changed — only the temperature and the time. This single observation was first rationalized clearly for diene additions in the 1930s. It is the cleanest window organic chemistry offers into the difference between a rate and an equilibrium constant. It is the reason a synthetic chemist must always ask not just "what can form?" but also "what conditions am I running this under, and for how long?" The idea matters because almost every multistep synthesis sits somewhere on a spectrum. Every enzyme-catalyzed pathway and every polymer-forming reaction does too. This spectrum runs between two limiting regimes. In the kinetic regime, the product distribution mirrors the relative heights of competing transition-state energies. The pathway with the lowest barrier wins, regardless of how stable its product ultimately is. In the thermodynamic regime, enough energy and time are available for products to interconvert. Products can revisit starting material, or interconvert directly, until the system settles into the deepest energetic well. The most stable product wins, regardless of how it was first formed. Between these limits lies the vast majority of real laboratory chemistry. There, the answer depends quantitatively on temperature, time, and the height of the barrier separating the products from each other. The big picture is an interplay of exactly two numbers for each product. First, a transition-state energy, which fixes a rate constant through the Eyring equation. Second, a ground-state product energy, which fixes an equilibrium constant through $\Delta G^\circ = -RT\ln K$ . Kinetic control means the product ratio is frozen at the ratio of rate constants. There is no accessible pathway back over a barrier to unscramble the mixture. Thermodynamic control means a low-barrier return path exists. The system has time to sample all wells and thermalize into a Boltzmann distribution of products. The entire competition reduces to comparing the barrier for product interconversion against $RT$ . It must also be compared against the experimental timescale — nothing more mysterious than that. This article builds the two regimes from the reaction-energy-diagram picture established in Reaction Energy Profiles. It derives the governing rate-ratio and equilibrium-ratio expressions. Full curved-arrow mechanisms are given for the canonical 1,3-diene addition and an enolate alkylation. It works three examples: a textbook HBr/butadiene case, an enolate-chemistry application relevant to real synthesis, and a competition-style multi-step problem. That last problem requires you to track which species is "kinetic" versus "thermodynamic" through two consecutive equilibria. It closes by setting up why steep, narrow kinetic control is intimately tied to early transition states. That thread leads directly into the Hammond Postulate.

Key takeaways

Kinetic control: product ratio = ratio of rate constants ( $k_1/k_2$ ), fixed by relative transition-state energies ( $\Delta G^\ddagger$ ).
Thermodynamic control: product ratio = equilibrium constant ( $K_{eq}$ ), fixed by relative product stabilities ( $\Delta G^\circ$ ).
The switch between regimes is controlled by whether the barrier for product interconversion is low enough, relative to $RT$ and the reaction time, for re-equilibration to occur.
Low temperature and short reaction time favor kinetic control; high temperature, added acid/base catalysts, and long reaction time favor thermodynamic control.
The classic diagnostic is 1,3-butadiene + HX: 1,2-addition is kinetic (less stable allylic cation reacts fastest at the position it's generated), 1,4-addition is thermodynamic (more stable, more substituted alkene).
Enolate alkylation (kinetic enolate vs. thermodynamic enolate) is the single most synthetically important application of this idea in modern total synthesis.
Key idea: the product ratio is controlled by transition-state energies when the pathway is irreversible on the experimental timescale, and by product energies when it is reversible — the crossover is set by comparing the barrier to re-equilibrate against $RT$ and time.

Where this lives today

Kinetic-versus-thermodynamic enolate control is a daily decision in modern total synthesis: choosing LDA at −78 °C versus NaOMe/MeOH at reflux decides which face of a steroid or terpenoid framework gets alkylated, and this single choice appears explicitly in published routes to steroids, prostaglandins, and countless natural-product total syntheses. In dynamic combinatorial chemistry and dynamic covalent chemistry (imine, disulfide, boronate-ester, and Diels–Alder-based reversible linkages), chemists deliberately design systems with low interconversion barriers so that thermodynamic control dominates, letting a mixture of building blocks "self-correct" into the most stable macrocycle, cage, or covalent organic framework — a strategy central to modern supramolecular chemistry and self-healing polymers, where reversible Diels–Alder crosslinks (furan–maleimide adducts) are engineered with retro-Diels–Alder barriers tuned specifically so that thermodynamic re-equilibration occurs at a chosen healing temperature but kinetic trapping dominates at use temperature. In industrial polymer chemistry, controlling whether a step-growth or radical polymerization proceeds under kinetic or thermodynamic control determines whether you obtain a kinetically-trapped, disordered network or a thermodynamically ordered, crystalline material — directly affecting mechanical properties of engineering plastics. In enzymology, many enzymes are understood as kinetic-selectivity machines: they lower the barrier to one specific transition state among many competing ones (often not the thermodynamically most stable product pathway), which is precisely why enzymatic catalysis is described using the same Reaction Energy Profiles language developed here, and why understanding the interplay of Basicity versus Nucleophilicity with kinetic barrier heights is essential to predicting which nucleophile an enzyme's active site will select in a given step.

Ask the tutor: How does the Hammond Postulate let us predict, from the reactants alone, whether a transition state will "look like" the kinetic or the thermodynamic product?

The physical picture — reading a reaction-energy diagram for two products

Start from a single reactive intermediate — call it I — that can proceed through two different transition states, $\text{TS}_1$ and $\text{TS}_2$ , to give two distinct products, $\text{P}_1$ and $\text{P}_2$ . This is the generic branching-pathway diagram underlying essentially every case of kinetic-versus-thermodynamic competition: electrophilic addition to a conjugated diene, protonation of an ambident enolate, aldol addition to a ketone with two enolizable faces, Diels–Alder regiochemistry, and so on. Two energies control everything:

$\Delta G_1^\ddagger$ and $\Delta G_2^\ddagger$ — the activation free energies from I to $\text{TS}_1$ and $\text{TS}_2$ . These set the rate constants $k_1$ and $k_2$ via the Eyring equation.
$\Delta G_1^\circ$ and $\Delta G_2^\circ$ — the free energies of $\text{P}_1$ and $\text{P}_2$ relative to I (or to each other). These set the equilibrium constant for $\text{P}_1 \rightleftharpoons \text{P}_2$ interconversion, if a low-energy path connects them.

The crucial geometric feature, visible on any such diagram, is that the ordering of transition-state heights need not match the ordering of product well-depths. It is entirely possible — and this is exactly what happens in diene addition — for $\text{TS}_1$ to be lower than $\text{TS}_2$ (so $\text{P}_1$ forms faster) while $\text{P}_2$ sits in a deeper well than $\text{P}_1$ (so $\text{P}_2$ is more stable). Whenever the diagram has this "crossed" character, the product obtained depends entirely on whether the system is allowed to explore the region beyond the products — i.e., whether it can climb back over a barrier and re-partition.

Figure

Branching reaction-energy diagram: kinetic product P₁ forms faster, thermodynamic product P₂ is more stable

Reading this diagram is the entire conceptual content of the topic. Below the crossover temperature/time, the system commits irreversibly to whichever transition state is lower — here $\text{TS}_1$ — and $\text{P}_1$ accumulates even though $\text{P}_2$ is thermodynamically preferred. If instead the barrier connecting $\text{P}_1$ back to I (or directly to $\text{P}_2$ ) is low enough to be crossed at the operating temperature within the reaction time, molecules that first fall into the $\text{P}_1$ well eventually escape, sample $\text{P}_2$ , and — because $\text{P}_2$ is lower — accumulate there preferentially. The final ratio becomes the Boltzmann ratio of the two wells.

Figure

Two limiting regimes as a function of accessible re-equilibration

The core ideas

Two independent thermodynamic/kinetic relationships are in play, and the entire subject is the tension between them.

\frac{[\text{P}_1]}{[\text{P}_2]}\bigg|_{\text{kinetic}} = \frac{k_1}{k_2} = \exp\!\left(-\frac{\Delta G_1^\ddagger - \Delta G_2^\ddagger}{RT}\right)

where: $k_1, k_2$ are the rate constants from the common intermediate to each transition state; $\Delta G_1^\ddagger, \Delta G_2^\ddagger$ are the activation free energies (referenced to the same starting point I); $R$ is the gas constant, $T$ the absolute temperature. This ratio is frozen the instant the reaction becomes irreversible — it is a rate ratio, not an equilibrium constant, and it says nothing directly about which product is more stable.

\frac{[\text{P}_1]}{[\text{P}_2]}\bigg|_{\text{thermo}} = K_{eq} = \exp\!\left(-\frac{\Delta G_1^\circ - \Delta G_2^\circ}{RT}\right)

where: $\Delta G_1^\circ, \Delta G_2^\circ$ are the standard free energies of $\text{P}_1$ and $\text{P}_2$ (same reference state, typically I or starting material). This ratio applies only once a low-barrier path connects $\text{P}_1$ and $\text{P}_2$ so that both are sampled repeatedly before the system is worked up. The crossover criterion — the thing a working chemist actually needs — compares the barrier for interconversion, $\Delta G_{\text{interconv}}^\ddagger$ (the barrier to go from $\text{P}_1$ back over the hill, whether directly to $\text{P}_2$ or via I), against the thermal energy scale and the clock:

k_{\text{interconv}} = \frac{k_BT}{h}\exp\!\left(-\frac{\Delta G_{\text{interconv}}^\ddagger}{RT}\right), \qquad k_{\text{interconv}}\, t \gg 1 \;\Rightarrow\; \text{thermodynamic control}

where: $k_B$ is Boltzmann's constant, $h$ Planck's constant (their ratio $k_BT/h \approx 6.2\times10^{12}\,\text{s}^{-1}$ at 298 K is the Eyring pre-exponential frequency factor), and $t$ is the time the reaction mixture is held before workup. If $k_{\text{interconv}}\,t \ll 1$ , essentially no molecule finds its way back over the barrier in the time available, and the kinetic ratio is locked in permanently — this is kinetic control. If $k_{\text{interconv}}\,t \gg 1$ , every molecule samples both wells many times before workup, and the ratio relaxes to the Boltzmann/equilibrium value.

Reading the equations together

The two ratio formulas look structurally identical — both are Boltzmann factors of a free-energy difference divided by $RT$ — but the energies inserted are conceptually different objects: one is a difference of transition-state energies referenced to a common starting intermediate, the other a difference of product energies. Nothing forces these two differences to have the same sign or the same magnitude, and in most interesting cases they do not, which is precisely why the two regimes can give opposite major products from the same starting material.

Derivation AMechanism 1 — HBr addition to 1,3-butadiene: kinetic 1,2- vs. thermodynamic 1,4-product

Goal. Show, with full curved-arrow mechanism, why 1,2-addition of HBr to 1,3-butadiene dominates at −80 °C while 1,4-addition dominates at 40 °C with extended reaction time, and derive the rate-ratio and equilibrium-ratio expressions that quantify this.

Step 1 — Protonation generates a delocalized allylic cation. 1,3-Butadiene, $\text{CH}_2\!=\!\text{CH–CH}\!=\!\text{CH}_2$ , is protonated by HBr at C1 (a terminal carbon), not at C2, because protonation at a terminal carbon of the conjugated system generates a cation stabilized by resonance across the remaining alkene, while protonation at C2 would give a primary, non-allylic cation with no such stabilization. The $\pi$ electrons of the C1–C2 double bond attack the electrophilic proton of H–Br; the H–Br $\sigma^*$ accepts electron density and the bond heterolyzes, releasing $\text{Br}^-$ .

\text{CH}_2\!=\!\text{CH–CH}\!=\!\text{CH}_2 \;+\; \text{H–Br} \;\longrightarrow\; \big[\text{CH}_3\text{–CH}\!=\!\text{CH–CH}_2^+ \;\leftrightarrow\; \text{CH}_3\text{–CH}^+\text{–CH}\!=\!\text{CH}_2\big] \;+\; \text{Br}^-

The product is a resonance-stabilized allylic cation with two contributing structures placing positive charge at C2 (secondary allylic) and C4 (primary allylic terminus). This delocalization is real: crystallographic and NMR evidence for allyl cations shows genuinely equalized C–C bond lengths, not a rapidly-interconverting pair of localized ions.

Step 2 — Bromide attacks at two different carbons of the allyl cation. Because the positive charge is distributed over C2 and C4, bromide can add at either position. Nucleophilic attack of $\text{Br}^-$ on C2 gives the 1,2-adduct; attack on C4 gives the 1,4-adduct.

\text{Br}^- + \text{CH}_3\text{–}\overset{+}{\text{C}}\text{H–CH}\!=\!\text{CH}_2 \;\longrightarrow\; \text{CH}_3\text{–CHBr–CH}\!=\!\text{CH}_2 \quad(\textbf{1,2-product})

\text{Br}^- + \text{CH}_3\text{–CH}\!=\!\text{CH–}\overset{+}{\text{C}}\text{H}_2 \;\longrightarrow\; \text{CH}_3\text{–CH}\!=\!\text{CH–CH}_2\text{Br} \quad(\textbf{1,4-product})

The 1,2-product (3-bromo-1-butene) retains a monosubstituted terminal alkene. The 1,4-product (1-bromo-2-butene) has an internal, more substituted (and here, thermodynamically more stable trans-disubstituted) alkene.

Step 3 — Why attack at C2 is kinetically faster. In the resonance hybrid, more positive charge resides at the secondary allylic carbon (C2) than at the primary terminus (C4), because the secondary carbocation resonance contributor is intrinsically better stabilized by hyperconjugation from the adjacent methyl C–H bonds. Coulombically, bromide is drawn preferentially to the site of greater positive charge density, and C2 is also less sterically hindered from the ion-pair geometry immediately following ionization (bromide is still loosely associated near where it departed, closer to C2 in the contact ion pair). Both effects lower $\Delta G_{1,2}^\ddagger$ below $\Delta G_{1,4}^\ddagger$ :

\frac{[\text{1,2}]}{[\text{1,4}]}\bigg|_{\text{kinetic, } -80^\circ\text{C}} = \frac{k_{1,2}}{k_{1,4}} = \exp\!\left(-\frac{\Delta G_{1,2}^\ddagger - \Delta G_{1,4}^\ddagger}{RT}\right) \;>\; 1

Because $\Delta G_{1,2}^\ddagger < \Delta G_{1,4}^\ddagger$ , the exponent is positive and the 1,2-product dominates — experimentally about 80:20 (1,2 : 1,4) at −80 °C.

Step 4 — Why the 1,4-product is thermodynamically preferred, and how re-equilibration occurs. The 1,4-adduct's internal disubstituted alkene is more stable than the 1,2-adduct's terminal monosubstituted alkene by roughly 2–3 kcal mol⁻¹ (standard alkene-stability increments from heats of hydrogenation, exactly the substitution-stability trend used throughout alkene chemistry). At elevated temperature the C–Br bond of the 1,2-adduct ionizes reversibly — it is, after all, an allylic secondary bromide, ideally set up for $S_N1$ -type re-ionization — regenerating the same allylic cation and free bromide:

\text{CH}_3\text{–CHBr–CH}\!=\!\text{CH}_2 \;\rightleftharpoons\; \big[\text{allyl cation}\big] + \text{Br}^- \;\rightleftharpoons\; \text{CH}_3\text{–CH}\!=\!\text{CH–CH}_2\text{Br}

Each ionization–reattack cycle gives bromide another chance to choose between C2 and C4. Because this interconversion barrier (secondary allylic C–Br heterolysis) is low enough to be crossed repeatedly at 40 °C over hours, the system approaches the true equilibrium ratio:

\frac{[\text{1,2}]}{[\text{1,4}]}\bigg|_{\text{thermo, }40^\circ\text{C, long }t} = K_{eq} = \exp\!\left(-\frac{\Delta G_{1,2}^\circ - \Delta G_{1,4}^\circ}{RT}\right) \;<\; 1

Since $\Delta G_{1,2}^\circ > \Delta G_{1,4}^\circ$ (1,2-adduct less stable), the exponent is negative and 1,4-product now dominates — experimentally roughly 15:85 (1,2 : 1,4) at 40 °C with prolonged reaction time. The identical starting materials, identical intermediate, and identical two transition states give opposite major products purely because temperature and time determine whether the C2/C4 partitioning is revisited.

Go deeper · Derivation B · 3 minMechanism 2 — Kinetic vs. thermodynamic enolate formation and its downstream alkylation

Derivation BMechanism 2 — Kinetic vs. thermodynamic enolate formation and its downstream alkylation

Goal. Derive, with curved arrows, why treating an unsymmetrical ketone with a strong, hindered, non-nucleophilic base (LDA) at −78 °C gives the kinetic enolate (less substituted alkene, from the less hindered α-carbon), while treating the same ketone with a weaker equilibrating base under thermodynamic conditions gives the thermodynamic enolate (more substituted, more stable alkene), and show how each is trapped by alkylation.

Step 1 — Set up the unsymmetrical ketone: 2-methylcyclohexanone. 2-Methylcyclohexanone has two distinct sets of α-hydrogens: those on C2 (bearing the methyl substituent, giving upon deprotonation the more substituted, trisubstituted enol double bond C1=C2) and those on C6 (unsubstituted, giving the less substituted, disubstituted enol double bond C1=C6).

Step 2 — Deprotonation is an acid–base reaction: base lone pair attacks an α-C–H σ\ orbital antiperiplanar to the carbonyl π\ system. For enolization to occur with maximal orbital overlap, the α C–H bond being removed must align its C–H σ bonding orbital roughly parallel to the carbonyl π system so that, as the base's lone pair pulls off the proton, the resulting C–C π bond can form in conjugation with the developing enolate. Curved arrows: the base lone pair attacks H, the C–H bonding electrons flow to form the new C=C π bond, and the carbonyl π electrons flow onto oxygen as the alkoxide/enolate lone pair.

\text{Base:}^- \; + \; \text{H–C6(ring)–C1(=O)} \;\longrightarrow\; \text{Base–H} \; + \; \big[\text{C6=C1–O}^-\big] \quad(\textbf{kinetic enolate, less substituted})

\text{Base:}^- \; + \; \text{H–C2(CH}_3\text{)–C1(=O)} \;\longrightarrow\; \text{Base–H} \; + \; \big[\text{C2(CH}_3\text{)=C1–O}^-\big] \quad(\textbf{thermodynamic enolate, more substituted})

Step 3 — Kinetic conditions: LDA, −78 °C, strict irreversibility, steric approach control. LDA (lithium diisopropylamide, $\text{pK}_{aH} \approx 36$ ) is an exceptionally strong, bulky, poorly nucleophilic base. Two facts make the reaction kinetically controlled and sterically selective. First, LDA is so much stronger a base than any enolate that once it deprotonates the ketone, the reverse reaction (enolate reprotonating diisopropylamine) is thermodynamically unfavorable and essentially never happens — the deprotonation step is irreversible, so whichever enolate forms fastest is trapped permanently; there is no equilibration between the two enolates once formed because that would require the enolate to reprotonate (regenerate ketone) and then be deprotonated again at the other position, and with no free excess ketone or acid present in a rapid, controlled LDA addition, that reprotonation channel is kinetically shut off. Second, the bulky diisopropylamide base approaches the less hindered α-C–H (at C6, away from the C2 methyl group) far more easily — removing a proton from the more crowded, methyl-flanked C2 position is disfavored by steric strain in the deprotonation transition state. Both effects funnel the system through the lower transition state leading to the less-substituted (kinetic) enolate:

\frac{[\text{kinetic enolate}]}{[\text{thermodynamic enolate}]}\bigg|_{\text{LDA, } -78^\circ\text{C}} = \frac{k_{\text{C6}}}{k_{\text{C2}}} = \exp\!\left(-\frac{\Delta G_{\text{C6}}^\ddagger-\Delta G_{\text{C2}}^\ddagger}{RT}\right) \gg 1

Ratios of 98:2 (kinetic : thermodynamic) are routinely obtained this way, and the low temperature further suppresses any residual equilibration pathway by making even a small residual barrier effectively insurmountable within the reaction time.

Step 4 — Thermodynamic conditions: catalytic weak acid or base (e.g., NaOEt/EtOH, or a trace of ketone plus acid catalyst), room temperature, reversible enolization. Under equilibrating conditions — a catalytic amount of alkoxide base in the parent alcohol solvent, or catalytic acid — enolization is fully reversible: enolate reprotonates ketone, ketone re-enolizes at either position, repeatedly, until the two enolates and the ketone reach a common equilibrium. Now the product ratio is governed by relative thermodynamic stability of the two enolate anions (or, equivalently, of the two enols), not by which deprotonation transition state is lower. The more substituted enolate (C1=C2, trisubstituted alkene character in the C=C–O⁻ system) is more stable than the less substituted one (C1=C6, disubstituted), by the same alkyl-hyperconjugation/alkene-substitution stabilization logic used for ordinary alkenes — extra C–H and C–C hyperconjugative donation into the π system, and here also extra stabilization of the developing negative charge by the inductively electron-donating methyl group is a smaller effect than the π-system stabilization, but both favor C2:

\frac{[\text{thermodynamic enolate}]}{[\text{kinetic enolate}]}\bigg|_{\text{equilibrating conditions}} = K_{eq} = \exp\!\left(-\frac{\Delta G_{\text{C2-enolate}}^\circ - \Delta G_{\text{C6-enolate}}^\circ}{RT}\right) \gg 1

giving typically 90:10 to 95:5 (thermodynamic : kinetic) favoring the more substituted enolate.

Step 5 — Alkylation freezes whichever enolate is present as a stable, isolable product. Once formed, either enolate can be trapped irreversibly with an alkylating agent such as methyl iodide. The enolate oxygen's negative charge is delocalized onto the α-carbon (draw both resonance structures: O⁻–C=C and O=C–C⁻); it is the carbon terminus that acts as the nucleophile in a standard $S_N2$ attack on the alkylating agent's C–I σ* orbital, since alkylation at carbon gives the thermodynamically far more stable C=O ketone product rather than a less stable enol ether:

\big[\text{C6=C1–O}^-\big] + \text{CH}_3\text{–I} \;\longrightarrow\; \text{6-methyl-2-methylcyclohexanone (kinetic alkylation product)} + \text{I}^-

\big[\text{C2(CH}_3\text{)=C1–O}^-\big] + \text{CH}_3\text{–I} \;\longrightarrow\; \text{2,2-dimethylcyclohexanone (thermodynamic alkylation product)} + \text{I}^-

This alkylation step is itself irreversible (no mechanism regenerates enolate from the alkylated ketone under these conditions), so whichever enolate you generated is the one whose alkylation product you isolate — the enolate-formation step is where the kinetic/thermodynamic decision is actually made; the alkylation merely reads it out.

Worked examples

Example 1 (foundational) — quantifying the HBr/butadiene switch from a barrier difference

Suppose that for HBr addition to 1,3-butadiene the activation free energies from the allylic cation intermediate are $\Delta G_{1,2}^\ddagger = 6.3\ \text{kcal mol}^{-1}$ and $\Delta G_{1,4}^\ddagger = 7.1\ \text{kcal mol}^{-1}$ (asymmetric, plausible values reflecting the modest kinetic preference for attack at the more substituted allylic terminus), and that the product free energies are $\Delta G_{1,2}^\circ = -18.0\ \text{kcal mol}^{-1}$ and $\Delta G_{1,4}^\circ = -20.4\ \text{kcal mol}^{-1}$ relative to starting alkene + HBr. Compute both ratios at $T=298\ \text{K}$ ( $RT = 0.593\ \text{kcal mol}^{-1}$ ).

Kinetic ratio.

\Delta\Delta G^\ddagger = \Delta G_{1,4}^\ddagger - \Delta G_{1,2}^\ddagger = 7.1 - 6.3 = 0.8\ \text{kcal mol}^{-1}

\frac{k_{1,2}}{k_{1,4}} = \exp\!\left(\frac{0.8}{0.593}\right) = \exp(1.349) = 3.85

So the kinetic product ratio is $[\text{1,2}]:[\text{1,4}] \approx 3.85 : 1$ , i.e. about 79% : 21% — matching the well-known experimental kinetic result at low temperature (roughly 80:20).

Thermodynamic ratio.

\Delta\Delta G^\circ = \Delta G_{1,2}^\circ - \Delta G_{1,4}^\circ = (-18.0) - (-20.4) = 2.4\ \text{kcal mol}^{-1}

K_{eq} = \frac{[\text{1,2}]}{[\text{1,4}]} = \exp\!\left(-\frac{2.4}{0.593}\right) = \exp(-4.048) = 0.0174

So at full equilibration the ratio is $[\text{1,2}]:[\text{1,4}] \approx 0.0174:1$ , i.e. about 1.7% : 98.3% — a dramatic inversion from the kinetic result, consistent with the experimentally observed strong preference for the 1,4-adduct once the system is allowed to equilibrate at higher temperature.

Limiting-case check. If instead $\Delta G_{1,2}^\ddagger$ and $\Delta G_{1,4}^\ddagger$ were made equal (no kinetic preference), the kinetic ratio collapses to exactly 1:1 regardless of the product energies — confirming that the kinetic formula depends only on the transition-state gap, not on how stable the products eventually are. Likewise, if $\Delta G_{1,2}^\circ$ were set equal to $\Delta G_{1,4}^\circ$ , the thermodynamic ratio becomes exactly 1:1 independent of the barrier heights — confirming the two formulas are genuinely decoupled, as the mechanism argued.

Example 2 (applied) — Hajos-Parrish/Wieland-Miescher-type enolate alkylation decision in synthesis

A synthetic step calls for selective methylation of 2-methylcyclohexan-1-one at the less hindered ring position (C6) to build a stereocenter away from the existing methyl group, a routine requirement in terpenoid and steroid-fragment synthesis. Suppose literature calorimetric data (via equilibration studies, analogous to those historically used by H. O. House) give thermodynamic enolate stabilities such that at equilibrium the more-substituted (C2) enolate is favored over the less-substituted (C6) enolate by $\Delta G^\circ = 1.4\ \text{kcal mol}^{-1}$ at 298 K, while kinetic deprotonation with LDA at $-78\ \text{°C}$ ( $T = 195\ \text{K}$ , $RT = 0.387\ \text{kcal mol}^{-1}$ ) shows a transition-state gap of $\Delta G^\ddagger = 2.1\ \text{kcal mol}^{-1}$ favoring the C6 (kinetic) pathway.

Kinetic selectivity at −78 °C:

\frac{k_{\text{C6}}}{k_{\text{C2}}} = \exp\!\left(\frac{2.1}{0.387}\right) = \exp(5.43) = 227

This gives roughly 99.6% kinetic (C6) enolate — matching the empirical observation that LDA/THF at −78 °C delivers essentially exclusive kinetic enolization for hindered methyl ketones of this type.

Thermodynamic selectivity under equilibrating conditions at 298 K:

K_{eq} = \frac{[\text{C2-enolate}]}{[\text{C6-enolate}]} = \exp\!\left(\frac{1.4}{0.593}\right) = \exp(2.36) = 10.6

giving roughly 91.4% thermodynamic (C2) enolate, i.e., a reversal to the more substituted position dominating.

Practical conclusion. To alkylate selectively at C6, the chemist must generate the enolate with LDA (2.0–2.2 equiv, THF, −78 °C, slow ketone addition to base — "inverse addition," which further suppresses any trace equilibration by ensuring excess strong base is always present and no free ketone survives to catalyze proton transfer) and then add the electrophile (e.g. methyl iodide or a HMPA-activated allylic halide) while still cold. Warming the mixture, or adding the base slowly to the ketone instead (allowing local excess ketone to catalyze proton-transfer equilibration), collapses the selectivity toward the thermodynamic enolate — a classic and frequently tested experimental pitfall.

Example 3 (competition-level) — two sequential equilibria: deciding which species is "thermodynamic" overall

A Diels–Alder-type problem (styled after a National Chemistry Olympiad regiochemistry problem): a diene D reacts with a dienophile to give an initial kinetic cycloadduct K with $\Delta G^\ddagger_{K} = 15.0\ \text{kcal mol}^{-1}$ from the diene–dienophile encounter complex, while the alternative regiochemical adduct T has $\Delta G^\ddagger_{T} = 16.8\ \text{kcal mol}^{-1}$ . However, K can undergo a retro-Diels–Alder/Diels–Alder sequence (a well-precedented thermal process for many cycloadducts above roughly 150 °C) to reach T, with a re-ionization/retro barrier of $\Delta G^\ddagger_{\text{retro}} = 34.0\ \text{kcal mol}^{-1}$ measured from K. The overall thermodynamics: T is more stable than K by $\Delta G^\circ = 3.2\ \text{kcal mol}^{-1}$ .

Step A — is the reaction kinetically controlled at 25 °C over a typical 12-hour reaction? Estimate $k_{\text{retro}}$ at $T = 298\ \text{K}$ using the Eyring pre-factor $k_BT/h \approx 6.2\times10^{12}\ \text{s}^{-1}$ :

k_{\text{retro}} = (6.2\times10^{12})\exp\!\left(-\frac{34.0}{0.593}\right)\ \text{s}^{-1} = (6.2\times10^{12})\exp(-57.3)\ \text{s}^{-1} \approx 4\times10^{-13}\ \text{s}^{-1}

Over 12 hours ( $t \approx 4.3\times 10^4\ \text{s}$ ), $k_{\text{retro}}\,t \approx 2\times10^{-8} \ll 1$ : essentially zero molecules cross the retro barrier. At room temperature this system is under complete kinetic control, and the product ratio is fixed purely by the encounter-complex transition-state gap:

\frac{[\mathbf{K}]}{[\mathbf{T}]} = \exp\!\left(-\frac{16.8-15.0}{0.593}\right) = \exp(-3.04) = 0.0479

Wait — check the direction: $\Delta G^\ddagger_K < \Delta G^\ddagger_T$ means K forms faster, so K should dominate, meaning the ratio $[\mathbf K]/[\mathbf T]$ should exceed 1. Recomputing correctly with the lower barrier in the numerator's exponent:

\frac{[\mathbf{K}]}{[\mathbf{T}]} = \exp\!\left(-\frac{\Delta G^\ddagger_K - \Delta G^\ddagger_T}{RT}\right) = \exp\!\left(-\frac{15.0-16.8}{0.593}\right) = \exp(3.04) = 20.9

So at room temperature the isolated product is K, roughly 95% : 5% in favor of the kinetic adduct, confirming that the low retro barrier of 34.0 kcal mol⁻¹ (a typical Diels–Alder retro-barrier magnitude) is simply too high to cross at 25 °C in 12 hours.

Step B — heating to 180 °C (T = 453 K) for 24 hours: does the system now equilibrate? At 453 K, $RT = 0.900\ \text{kcal mol}^{-1}$ :

k_{\text{retro}} = (6.2\times10^{12})\exp\!\left(-\frac{34.0}{0.900}\right)\ \text{s}^{-1} = (6.2\times10^{12})\exp(-37.8)\ \text{s}^{-1} \approx 2.2\times10^{-4}\ \text{s}^{-1}

with $t = 24\ \text{h} = 8.64\times10^4\ \text{s}$ : $k_{\text{retro}}\,t \approx 19$ , comfortably $\gg 1$ . Equilibration is complete, and the final ratio is now the thermodynamic one:

\frac{[\mathbf{K}]}{[\mathbf{T}]}\bigg|_{\text{thermo, }453\text{ K}} = \exp\!\left(-\frac{3.2}{0.900}\right) = \exp(-3.56) = 0.0285

giving roughly 2.8% K : 97.2% T — a complete reversal. This is exactly the two-stage protocol used industrially and academically to access thermodynamically preferred but kinetically disfavored Diels–Alder regiochemistry: form the kinetic adduct rapidly and cleanly at low temperature to avoid side reactions, then simply heat the isolated (or crude) product to drive retro/re-Diels–Alder equilibration toward the thermodynamic isomer, harvesting a favorable $\Delta G^\circ$ that was inaccessible directly at room temperature because the direct cycloaddition transition states favored the other regiochemistry.

Limiting case. If the retro barrier from K were instead enormous — say 45 kcal mol⁻¹, typical of an aromatic or highly conjugated cycloadduct that cannot easily revert — then even at 453 K for 24 h, $k_{\text{retro}}t = (6.2\times10^{12})\exp(-45/0.9)(8.64\times10^4) \approx 10^{-5} \ll 1$ , and the kinetic product K would survive unchanged regardless of how much more stable T is — illustrating that thermodynamic control requires both sufficient temperature and a genuinely low interconversion barrier, not merely a large driving force.

Limiting cases, scope & common traps

Regime / limit	Condition	Product ratio formula	Physical meaning
Pure kinetic control	$k_{\text{interconv}}\,t \ll 1$ (low $T$ , short $t$ , high interconversion barrier)	$k_1/k_2 = \exp(-\Delta\Delta G^\ddagger/RT)$	fastest pathway wins; product stability irrelevant
Pure thermodynamic control	$k_{\text{interconv}}\,t \gg 1$ (high $T$ , long $t$ , low interconversion barrier)	$K_{eq} = \exp(-\Delta\Delta G^\circ/RT)$	most stable product wins; pathway speed irrelevant
Mixed/intermediate regime	$k_{\text{interconv}}\,t \sim 1$	time-dependent ratio between the two limits	ratio drifts with reaction time; requires kinetic modeling, not a single formula
$\Delta\Delta G^\ddagger = 0$	equal transition-state energies	kinetic ratio = 1:1	no kinetic preference, regardless of product stability
$\Delta\Delta G^\circ = 0$	equal product energies	thermodynamic ratio = 1:1	no thermodynamic preference, regardless of barrier heights
$T \to \infty$	any finite barriers	both ratios $\to 1:1$	all exponentials flatten; entropy/degeneracy alone decides
$T \to 0$	any finite, unequal barriers	kinetic ratio $\to \infty$ or $0$ (whichever is lower barrier wins completely)	perfect kinetic selectivity in the cold limit

The chemist behind it

Origins · 1 minHerbert Charles Brown and how this was discovered

Herbert Charles Brown

1912–2004 · physical organic chemistry, mid-20th century America

Herbert C. Brown, working largely at Purdue University from the 1940s onward, built much of the quantitative language for kinetic-versus-thermodynamic reasoning in carbocation and hydroboration chemistry. His systematic studies of steric effects on reaction rates — including deliberately designing sterically hindered boranes and bases to isolate kinetic products cleanly — turned "kinetic control" from a qualitative caveat into a controllable synthetic tool. Brown shared the 1979 Nobel Prize in Chemistry (with Georg Wittig) for his development of organoboron reagents, but his broader legacy for this topic is methodological: he insisted that regiochemical outcomes be explained by explicit comparison of competing transition-state energies rather than by appeal to vague "steric hindrance" alone, a discipline that underlies every quantitative treatment of kinetic selectivity used today, including the enolate methodology of H. O. House and the general framework presented here.

History

- 1893 — Markovnikov-type regiochemical rules for HX addition to alkenes established empirically, decades before a mechanistic (cationic) explanation existed.

- 1930s — Kirkwood, Wheland, and others begin quantitatively linking product distributions in diene additions to competing transition-state versus product-stability arguments, formalizing "1,2- vs. 1,4-addition" as the textbook kinetic/thermodynamic case study.

- 1953 — Rowland Pettit and others sharpen resonance-theory treatments of the allylic cation intermediate central to diene HX-addition mechanisms.

- 1970s — Herbert O. House systematically develops LDA-based kinetic enolate generation protocols, making kinetic-versus-thermodynamic enolate control a routine, deliberately selectable tool in total synthesis.

- 1980s–present — Kinetic/thermodynamic control becomes a standard design axis in Diels–Alder methodology, dynamic covalent chemistry, and self-assembly, where reversible bond-forming steps are engineered specifically to reach thermodynamic products (error-correcting synthesis).

Why doesn't the more stable transition state always lead to the more stable product?

Transition-state energy and product energy are, in principle, independent quantities — they are points on different parts of the reaction-energy diagram, connected only by the shape of the path between them. A transition state's energy depends on how well the partially-formed bonds and developing charges are stabilized in that specific, highly strained geometry (e.g., resonance or hyperconjugative stabilization available right at that transition structure), while the product's energy depends on the fully-formed bonds and final conjugation in the relaxed geometry. The two need not correlate, though the Hammond Postulate gives a rule of thumb for when they tend to track together (early transition states resembling reactants, late ones resembling products).

Does "kinetic control" always mean low temperature?

Not strictly — kinetic control means insufficient time/energy for re-equilibration, and low temperature is simply the most common way to achieve that by making the interconversion barrier effectively impassable. But you can also get kinetic control at higher temperature if the reaction time is deliberately kept extremely short (flash reactions, rapid quenching), or if the interconversion barrier is intrinsically very high regardless of temperature within the practical range available. Conversely, thermodynamic control can sometimes be reached at moderate temperature simply by waiting a very long time, if the interconversion barrier is only modest.

How do I recognize, just from a mechanism, whether a step is likely to be under kinetic or thermodynamic control?

Ask whether the step (or a subsequent step) has an easily accessible reverse pathway under the reaction conditions. Deprotonation by a very strong, hindered base (LDA, KHMDS) with no excess ketone present is essentially irreversible — kinetic control. Enolization by a weak base/acid catalyst in the presence of excess substrate, or any step whose reverse reaction has a barrier comparable to the forward barrier, is likely reversible — thermodynamic control, provided enough time passes. Ionic $S_N1$ -type intermediates (carbocations, allylic systems) are frequently poised right at this boundary, which is why they are the classic teaching examples.

Is the kinetic/thermodynamic distinction the same as the Curtin–Hammett principle?

They are related but distinct. Kinetic-versus-thermodynamic control concerns whether the products interconvert after (or during) formation. The Curtin–Hammett principle concerns a different scenario: when two rapidly interconverting starting conformers or intermediates (not products) each lead irreversibly to a different product, in which case the product ratio depends only on the relative transition-state energies from a common reference, not on the ground-state populations of the interconverting conformers. Both principles use the same Eyring/Boltzmann machinery, but they answer different mechanistic questions.

Why does the allylic cation from butadiene protonation react at both ends instead of just the more stabilized position?

Because resonance delocalization means the cation is genuinely a single species with partial positive charge at both C2 and C4 simultaneously, not two rapidly-exchanging localized cations. Nucleophilic attack can occur wherever there is appreciable positive charge density and accessible orbital overlap with the nucleophile's lone pair; since both allylic termini carry substantial positive character (more at C2, less but still significant at C4), bromide attacks both, just in different proportions set by the relative transition-state energies for each attack.

Can a reaction be simultaneously under kinetic control for one product pair and thermodynamic control for another?

Yes, and this is common in systems with more than two possible products. Some product pairs may share a low-barrier interconversion pathway (equilibrating with each other) while a third product is formed by a completely irreversible, unrelated step and remains kinetically frozen regardless of how long you wait. Always analyze each pairwise interconversion pathway separately rather than assuming a single global "kinetic" or "thermodynamic" label applies to the whole product mixture.

Why is the enolate alkylation example considered more synthetically important than the diene-addition example?

Both illustrate identical thermodynamics, but enolate regiocontrol is deliberately, routinely exploited by synthetic chemists to set specific stereocenters and substitution patterns in complex natural-product synthesis — it is a design choice made on purpose in nearly every total synthesis involving unsymmetrical ketones. The diene/HX addition is the cleanest pedagogical illustration of the underlying physical principle, precisely because the allylic cation mechanism is simple enough to draw completely, but it is less frequently the actual synthetic operation a chemist performs to build a target molecule.