Reaction Energy Profiles

Reading energy profiles correctly is not a historical curiosity — it is the daily working language of mechanistic and computational chemistry. In drug metabolism, cytochrome P450 oxidation pathways are mapped as multistep energy profiles with several iron-oxo intermediates, and knowing which transition state is rate-determining tells medicinal chemists which position on a drug molecule is most vulnerable to oxidative metabolism, guiding where to block a molecule with a fluorine or a deuterium. In industrial catalysis, computational chemists routinely compute full free-energy profiles for catalytic cycles (Pd-catalyzed cross-couplings, olefin polymerization, hydroformylation) and apply exactly the energetic-span reasoning developed here to identify which single transition state, out of a dozen candidate steps, actually limits turnover — informing ligand redesign aimed at destabilizing that one peak or the resting state that precedes it. In enzymology, the entire kinetic framework of enzyme catalysis (Michaelis–Menten kinetics, transition-state-analog inhibitor design) rests on picturing the enzyme active site as lowering a specific transition-state peak relative to the uncatalyzed profile, without necessarily changing the overall thermodynamics of substrate to product. In materials and polymer chemistry, competing chain-propagation versus chain-termination pathways are compared as parallel branches on the same energy-coordinate diagram, and the branch with the lower barrier wins kinetically even if the other branch is thermodynamically preferred — a direct real-world instance of the kinetic-versus-thermodynamic competition this article sets up. Finally, ultrafast spectroscopy (femtosecond pump-probe methods) now allows some very short-lived intermediates and even transition-state-like structures to be interrogated directly in time, giving experimental teeth to features that were, for most of the twentieth century, purely inferred from the shape of a drawn curve.

Ask the tutor: In my proposed mechanism, adding a stronger-binding catalyst lowered every individual step's local activation energy, yet the overall reaction got slower — how do I show this with an energy profile and the energetic span model?

A reaction coordinate diagram is deceptively simple to draw and easy to misread. Start with the axes. The vertical axis is energy — ideally Gibbs free energy $G$, since real reactions occur at constant temperature and pressure and $\Delta G$ governs the equilibrium constant via $\Delta G^\circ = -RT\ln K_{eq}$. Many textbook diagrams plot potential energy $E$ instead, which is a legitimate simplification when entropy changes along the coordinate are small (e.g., a single bond rotation or a symmetric bond-forming/breaking event with no change in molecularity), but the two are conceptually distinct and can even disagree on whether a step is "downhill." The horizontal axis is the reaction coordinate. It is not time, and it is not any one bond length in general — it is an abstract collective coordinate, a particular linear combination of bond lengths, angles, and dihedral angles that traces the lowest-energy path across the multidimensional potential energy surface from reactant valley to product valley. For a simple $S_N2$ displacement it is dominated by the antisymmetric combination of the breaking C–X bond stretching and the forming C–Nu bond contracting. For a proton transfer it is dominated by the O–H(or N–H)···base distance. The key mathematical fact, due to Eyring's transition state theory, is that this coordinate has a well-defined maximum — the transition state — which is a saddle point: a maximum in the direction of the reaction coordinate, but a minimum in every direction perpendicular to it (bond angles relaxing, other bonds staying put). That is what makes it a fleeting geometry rather than a stable species: any small nudge along the reaction coordinate rolls it downhill toward either reactants or products, but nudges in other directions cost energy, so it is a genuine (if momentary) stationary point of the surface. Valleys between peaks are qualitatively different: they are true local minima in every direction, including along the reaction coordinate. A species sitting in such a valley is a reactive intermediate — a carbocation, radical, tetrahedral intermediate, enolate, arenium ion — with a finite, if short, lifetime governed by how deep the valley is relative to the barriers on either side. This is precisely the boundary case explored when classifying intermediates: shallow valleys (barriers of a few kcal/mol on either side) give species with lifetimes near $10^{-10}$–$10^{-12}$ s, essentially never isolable, while deeper valleys (barriers well above 15–20 kcal/mol) can give species stable enough to observe by NMR at low temperature or even isolate as salts.

Two important quantitative structures appear in this picture. First, the activation energy for any single step is measured from the resting-state minimum immediately before that step's peak — not from the very bottom of the whole diagram. In the figure, the barrier for the first step is $\Delta G^\ddagger_1$, measured from R up to TS1; the barrier for the second step, $\Delta G^\ddagger_2$, is measured from the intermediate up to TS2, and it is smaller even though TS2 sits at lower absolute energy than TS1 — because the intermediate itself is a shallow, high-energy resting state. Second, the rate-determining step (more properly, "rate-determining transition state" in the modern energetic span picture) is not simply "the highest peak on the page." It is the transition state that produces the largest energetic span back to the nearest preceding resting-state minimum, because that span is what controls the overall rate constant through the Eyring equation.

This second figure previews an idea developed fully elsewhere: the position of the transition state along the coordinate — early and reactant-like, or late and product-like — correlates with whether the step is exothermic or endothermic. That correlation is the substance of the Hammond Postulate, and it is only meaningful once you can read a profile fluently, which is the entire purpose of this article.

The bridge between the drawn curve and measurable numbers runs through two rate theories: the empirical Arrhenius equation and the more mechanistic Eyring (transition state) theory.

where: $k$ is the rate constant, $A$ is the pre-exponential (frequency) factor, $E_a$ is the Arrhenius activation energy, $R = 1.987 \times 10^{-3}\ \text{kcal mol}^{-1}\text{K}^{-1}$, and $T$ is absolute temperature. Reading this equation: the rate constant is exponentially sensitive to the barrier height, so a change of only a few kcal/mol in $E_a$ changes $k$ by orders of magnitude at room temperature.

where: $k_B$ is Boltzmann's constant, $h$ is Planck's constant, and $\Delta G^\ddagger = \Delta H^\ddagger - T\Delta S^\ddagger$ is the free energy of activation — the true vertical distance from a resting-state minimum to the transition-state peak in a free-energy profile. Reading this equation: $k_BT/h$ (about $6.2\times10^{12}\ \text{s}^{-1}$ at 298 K) is a universal attempt frequency, and the exponential term is the fraction of collisions with enough organized energy to cross the saddle point.

where: $\Delta G^\circ$ is the overall free-energy difference between product and reactant minima (the "downhill-ness" of the whole reaction, thermodynamics), while $K_{eq}$ is the equilibrium constant. Reading this equation: this number lives entirely at the far-left and far-right ends of the diagram and, crucially, carries no information about the height or number of peaks in between — a reaction can be enormously favorable thermodynamically and still be immeasurably slow because of a high kinetic barrier.

Derivation 1 — from Arrhenius data to a barrier height (and why the profile has the shape it does)

Goal. Show mechanistically/mathematically how measuring $k$ at two temperatures lets you place an experimental number on the height of a peak you can never see directly, and interpret what that number means physically.

Step 1 — collision theory intuition. Picture reactant molecules colliding with a spread of kinetic and vibrational energies described (approximately) by a Boltzmann distribution. Only collisions oriented correctly and possessing enough energy to distort bond lengths/angles all the way to the saddle-point geometry are "successful." The fraction of molecules with energy at least $E_a$ above the mean, according to Boltzmann statistics, scales as $e^{-E_a/RT}$. This is the physical origin of the exponential term — it is not an empirical curve-fit, it is a statement about populating a high-energy geometry.

Step 2 — linearize. Take the natural log of the Arrhenius equation:

This is a straight line in a plot of $\ln k$ versus $1/T$, with slope $-E_a/R$ and intercept $\ln A$. This is the "derivation" in the physical sense: two rate measurements at two temperatures pin down the two unknowns $E_a$ and $A$, exactly as two points determine a line.

Step 3 — solve for $E_a$ from two data points. If $k_1$ is measured at $T_1$ and $k_2$ at $T_2$, subtracting the two log-forms eliminates $\ln A$:

Rearranged for the unknown barrier:

Step 4 — connect back to the picture. Once $E_a$ (or, via Eyring, $\Delta H^\ddagger$ and $\Delta S^\ddagger$) is known, you can draw the vertical height of the very first peak in the profile to scale, and you have converted an invisible geometric feature of a potential energy surface into a number obtained purely from a stopwatch and a thermometer. This is the entire epistemic basis for believing energy diagrams are quantitatively real and not just cartoons.

Derivation 2 — the energetic span and identifying the true rate-determining transition state

Goal. For a multistep profile with several intermediates and several transition states, derive which single TS actually controls the observed rate, correcting the naive "just find the tallest peak" rule.

Step 1 — set up a two-step catalytic-style cycle. Consider resting state R, intermediate I, transition states TS1 (R→I) and TS2 (I→P), with energies (relative to R = 0) of $E(\text{TS1})$, $E(I)$, $E(\text{TS2})$, and overall product energy $E(P)$. The overall rate is limited by the largest energy climb the system must make between any resting state and the transition state that follows it going forward, not by absolute peak height alone.

Step 2 — the naive trap. Suppose $E(\text{TS1}) = 18.2$ kcal/mol (measured from R) and $E(\text{TS2}) = 15.8$ kcal/mol (also measured from R, i.e., in absolute terms TS2 is lower). A student might say "TS1 is the highest point, so step 1 is rate-determining." This is correct only if the intermediate I lies close in energy to R.

Step 3 — the energetic span correction. Suppose instead $E(I) = 4.0$ kcal/mol. The energy gap the system must actually climb for the second step is

while the gap for the first step is

Since $\delta_1 > \delta_2$, TS1 remains rate-determining here — the naive rule happened to work. But now suppose the intermediate is much higher in energy, say $E(I) = 13.5$ kcal/mol (a very reactive, short-lived intermediate sitting in a shallow well close to TS2). Recompute:

Now $\delta_2 \ll \delta_1$, and the first step is even more overwhelmingly rate-determining — the absolute height of TS2 was irrelevant because the intermediate sits almost as high. Conversely, engineer a case where $E(\text{TS1}) = 18.2$ kcal/mol but $E(I) = 16.0$ kcal/mol (a deep-ish but still real minimum) and $E(\text{TS2}) = 15.8$: here $\delta_2 = 15.8 - 16.0 = -0.2$, meaning the second "barrier" measured from that particular intermediate is essentially barrierless, and the only meaningful climb in the whole diagram is $\delta_1 = 18.2$. The overall observed rate constant tracks $e^{-\delta_1/RT}$ almost exactly.

Step 4 — general statement. For a cycle or cascade with resting states $R_i$ and transition states $\text{TS}_j$, the rate-determining span is

taken over all pairs where $\text{TS}_j$ occurs after $R_i$ along the forward path (not merely the single largest peak-minus-baseline). This generalizes the simple "highest hill" rule taught at introductory level into the energetic span model used in modern computational and catalytic mechanism analysis, and it is the rigorous version of what "rate-determining step" means whenever more than one intermediate is present.

Example 1 — Foundational: reading a two-step profile and locating $E_a$. A hypothetical nucleophilic addition proceeds via a tetrahedral intermediate. Relative free energies (kcal/mol, reactant = 0): TS1 = 16.4, intermediate = 6.1, TS2 = 9.8, product = $-11.3$. Step 1 — draw the mental picture. Four points: 0 → 16.4 → 6.1 → 9.8 → $-11.3$. The curve rises to a first peak, drops into a valley, rises again to a smaller second peak, then drops well below the start. Step 2 — compute the two activation energies. Measured from the immediately preceding resting state:

Step 3 — identify rate-determining TS. Using the energetic span idea, $\delta_1 = 16.4$ kcal/mol dominates over $\delta_2 = 3.7$ kcal/mol, so TS1 is rate-determining; the intermediate, once formed, collapses to product almost immediately (it is a "reactive," not "stable," intermediate — consistent with a shallow well only 3.7 kcal/mol below its exit barrier). Step 4 — overall thermodynamics. $\Delta G^\circ = -11.3 - 0 = -11.3$ kcal/mol, giving

The reaction is strongly thermodynamically favorable, but the rate is entirely set by the 16.4 kcal/mol kinetic barrier — a clean illustration that $\Delta G^\circ$ and $\Delta G^\ddagger$ are independent numbers answering independent questions. Limiting case check. If the intermediate's energy were pushed up to nearly TS2's value (e.g., intermediate = 9.5, TS2 = 9.8), the second barrier $\Delta G^\ddagger_2 = 0.3$ kcal/mol becomes negligible and the profile effectively behaves as a single-step process controlled entirely by TS1 — exactly the concerted-limit behavior expected as an intermediate's lifetime shrinks toward zero.

Example 2 — Applied: extracting $E_a$ for solvolysis of tert-butyl bromide from real-style kinetic data. Suppose the observed pseudo-first-order rate constant for $S_N1$ solvolysis of tert-butyl bromide in 80% aqueous ethanol is measured as $k_1 = 1.7\times10^{-5}\ \text{s}^{-1}$ at $T_1 = 298\ \text{K}$ and $k_2 = 8.7\times10^{-4}\ \text{s}^{-1}$ at $T_2 = 328\ \text{K}$ (representative, literature-like values). Step 1 — apply the two-point Arrhenius formula.

Step 2 — compute the temperature term.

Step 3 — solve for $E_a$.

This value is close to literature activation energies for $S_N1$ solvolysis of tertiary alkyl halides (commonly cited in the low-to-mid 20s kcal/mol in aqueous ethanol), and it corresponds mechanistically to the single rate-determining transition state for C–Br heterolysis: a peak dominated by charge separation, stabilized by the polar protic solvent, well before any nucleophile (water/ethanol) becomes covalently attached. The single-peak/single-valley profile — ionization (rate-determining, high barrier) followed by a fast, low-barrier capture of the planar carbocation by solvent — is exactly the two-step $S_N1$ picture, and this $E_a$ belongs to the first, ion-forming step only. Limiting case check. As $T_1 \to T_2$ (two measurements taken at nearly the same temperature), $\ln(k_2/k_1) \to 0$ and $1/T_2 - 1/T_1 \to 0$ simultaneously; the ratio is indeterminate $0/0$ and $E_a$ becomes numerically unreliable — physically, you cannot extract a barrier height from data with no temperature spread, which is why real Arrhenius experiments use temperature ranges of at least 20–30 K.

Example 3 — Exam/competition level: dissecting a three-step catalytic cycle with a branch point. A Lewis-acid-catalyzed aldol addition is proposed to proceed: (i) catalyst binding to substrate (TS_a, low barrier), (ii) enolization (TS_b), (iii) C–C bond-forming addition (TS_c), with energies (kcal/mol, catalyst + substrates = 0): bound complex $I_1 = -3.2$, TS_a $= 2.1$, enol intermediate $I_2 = 9.4$, TS_b $= 9.9$, TS_c $= 14.0$, product complex $= -18.6$. Step 1 — lay out all resting states and transition states in order.

(the catalyst-binding step is fast and slightly favorable, dropping to $I_1$, before the system must climb back up for enolization). Step 2 — compute each local activation energy relative to its preceding resting state.

Step 3 — apply the energetic span model across the whole cycle, not just locally. The lowest resting state overall is $I_1$ at $-3.2$ kcal/mol; the highest transition state overall is TS_c at $14.0$ kcal/mol, and TS_c occurs after $I_1$ along the forward path. The true energetic span is

This is larger than any single locally-measured barrier above ($2.1$, $13.1$, or $4.6$ kcal/mol) — because the correct comparison spans from the deepest resting state to the highest transition state anywhere downstream, even though they are not adjacent on the diagram. This is the subtlety a purely local reading misses: TS_b looks like "the enolization barrier," and locally it is only 13.1 kcal/mol above $I_1$, but the system that eventually reaches TS_c had to have started from $I_1$, so the operative kinetic bottleneck for the overall catalytic turnover is the full 17.2 kcal/mol climb from $I_1$ to TS_c, which — because $I_2$ sits substantially above $I_1$ — is dominated in practice by whichever step forces the population back up near $I_2$/TS_c. Step 4 — convert to a turnover-frequency estimate. Using $\delta_{\max} = 17.2$ kcal/mol at 298 K:

giving a rough catalytic turnover frequency on the order of $1$–$2$ per second — a plausible, moderately fast catalytic cycle, illustrating how a full multi-intermediate profile compresses into a single practically meaningful rate constant. Limiting case check. If catalyst binding were made much stronger (deepening $I_1$ to, say, $-15.0$ kcal/mol) without changing TS_c, the energetic span increases to $14.0-(-15.0)=29.0$ kcal/mol and the catalyst becomes dramatically slower — the classic "too-stable resting state" trap in catalyst design, where over-stabilizing an intermediate poisons turnover even though every individual local barrier looks small.

The modern language of activation energy and transition states traces most directly to Henry Eyring (1901–1981), a Mexican-American physical chemist who, working with Michael Polanyi's ideas on potential energy surfaces, formulated transition state theory in 1935 while at Princeton. Eyring's radical move was to treat the transition state as a quasi-equilibrium species — assuming a Boltzmann-like population of activated complexes in rapid equilibrium with reactants — and to derive a rate expression, the Eyring equation, directly from statistical mechanics rather than from empirical curve-fitting. This gave chemists, for the first time, a first-principles reason to expect the exponential temperature dependence that Svante Arrhenius had already observed empirically decades earlier, and it gave physical meaning to the pre-exponential factor as an entropy-of-activation term. Eyring went on to apply the theory across an enormous range of chemistry and biology — viscosity, diffusion, enzyme kinetics — always insisting on the same picture: a potential energy surface with a saddle point that a system must cross, however briefly, on its way from one stable arrangement of atoms to another.