Recrossing and Variational Transition-State Theory

Introduction

A conventional transition-state calculation often places its dividing surface at a potential-energy saddle. That is a sensible starting point, but a saddle geometry is not always the best boundary between reactants and products at finite temperature. Trajectories can cross it and return, especially when motion is multidimensional or the environment relaxes slowly. Variational transition-state theory improves the estimate by moving the dividing surface to find a tighter kinetic bottleneck.

Core explanation

The conventional TST rate counts positive crossings through a selected dividing surface and assumes no return crossing. If a molecule crosses and returns to reactants, the counted flux exceeds net product formation. On the same model and within classical assumptions, a correction can be represented by a transmission coefficient κ below one. A poor coordinate or misplaced surface can create much recrossing; a better surface can make a crossing more predictive of commitment.

Variational transition-state theory (VTST) evaluates candidate dividing surfaces at different positions along a path and chooses the one giving the lowest calculated TST rate . IUPAC describes this minimisation explicitly. Each surface can be thought of as a gate; the gate with the least forward equilibrium flux is the strongest bottleneck among the tested gates. In a one-dimensional cartoon, the chosen location may coincide with the largest free-energy barrier rather than exactly with the highest electronic-energy point.

Why minimise the rate rather than maximise a barrier? In TST, a surface's estimated flux incorporates its accessible configurations and crossing velocities. A narrow, entropically restricted region can be more rate limiting than the tallest point on a potential-energy line. The canonical variational calculation performs this comparison at a specified temperature. A microcanonical variant holds total energy fixed and can be important for isolated energized molecules.

The variational result remains an approximation. It searches a family of candidate surfaces, usually connected to a chosen reaction path or coordinate. The globally ideal no-recrossing boundary might depend on additional coordinates and momenta not in that family. Residual recrossing can therefore remain after optimisation, and direct trajectories can estimate it. Quantum tunnelling can also allow reaction without classical passage over the selected barrier; a tunnelling correction is conceptually distinct from reducing recrossing.

For association or barrierless reactions, the electronic potential may descend monotonically while entropy produces a free-energy bottleneck at a relatively long separation. A fixed saddle-based rule may fail because there is no ordinary first-order saddle in the entrance channel. Variable reaction-coordinate approaches place candidate surfaces at different separations and orientations. Such methods are useful for radical recombination and complex formation, but their accuracy depends on long-range interactions and rotational state treatment.

It is also essential to preserve rate-law and standard-state conventions. Moving a dividing surface does not repair an incorrect molecularity, wrong concentration standard or omitted conformer ensemble. VTST determines the best gate within its model; a credible rate still needs correct reactant populations, electronic energies and potential dynamical corrections. Comparing the variational rate with measured temperature or pressure dependence is an informative test.

Step-by-step reasoning

Define reactant and product basins and construct a reaction path or collective coordinate. Place several candidate dividing surfaces across the bottleneck region. At each temperature, compute the TST forward flux relative to the same reactant population and standard state. Select the minimum estimated rate among the surfaces. Then launch trajectories, if feasible, to measure remaining recrossing and assess whether an omitted coordinate or tunnelling matters. Report the surface family and assumptions with the final rate.

Visual explanation

Draw a reaction free-energy curve with several vertical gates A, B and C. Above each write its TST rate estimate, for example 12, 5 and 9 s⁻¹. Circle B as the variational bottleneck even if a potential-energy saddle marker sits elsewhere. Add arrows that cross A and return to show why an easy gate overcounts reactions. A second panel can show a broad barrierless association entrance with an entropy bottleneck.

Real-world analogy

To estimate how many hikers complete a route, count crossings at the narrowest checkpoint rather than at an easy trail segment where many hikers turn around. Moving the checkpoint can reduce false predictions of completion. Molecular gates differ because they live in a high-dimensional space and the chance of returning depends on velocity, solvent motion and quantum behavior.

Real-world example

In radical association, two fragments may attract and form a bond without a distinct electronic saddle. Yet only certain separations and orientations lead to capture, and the loss of relative translational and rotational freedom creates a bottleneck in accessible states. A variational dividing surface can be located by comparing flux estimates at several fragment separations. Collisional stabilisation may then affect whether the newly formed energized adduct persists.

Why?

Why can the best surface differ from a saddle? Free-energy and dynamical bottlenecks include entropy and motion, not only electronic potential height. Why choose the minimum calculated rate? It is the tightest bottleneck in the tested surface family and generally reduces positive-flux overcounting. Why still run trajectories? Surface optimisation within a limited coordinate family may leave recrossing. Why repeat at each temperature? Entropic restrictions and populations vary with T.

Common misconception

VTST does not make a rate exact simply by moving a line on an energy plot. The candidate surfaces and statistical model can omit important dimensions. It is also incorrect to interpret a lower variational rate as a slower chemical mechanism discovered by itself; the method refines the estimate for a specified pathway and reactant state.

Worked example

Question: Three candidate dividing surfaces for one unimolecular reaction yield TST estimates of 8.0, 5.5 and 6.2 s⁻¹ at 300 K. Which does a simple variational procedure select? If trajectory analysis of that selected surface finds a net-success fraction of 0.80, what is the corrected estimate?

Reasoning: Select the smallest TST value, 5.5 s⁻¹, since it is the tightest sampled gate. A separate recrossing correction gives 0.80 × 5.5 = 4.4 s⁻¹. This is still conditional on the chosen surface family, reactant distribution and trajectory sampling. The three gates should not be treated as three independent reaction channels.

Answer: Select the 5.5 s⁻¹ surface; the illustrative corrected rate is 4.4 s⁻¹.

Quick check

1. Does variational transition-state theory select the largest or smallest TST rate among tested surfaces? Answer: It selects the smallest estimated TST rate, the tightest bottleneck in that candidate family.

Exam focus

Connect recrossing with overcounted instantaneous flux. State the variational rule exactly, including fixed temperature and common reactant reference. Explain why the chosen surface need not pass through the electronic saddle and why a further dynamical correction can remain necessary.

Advanced insight

In strict classical phase-space language, an ideal nonrecrossing surface depends on position and momentum. Practical VTST often varies geometry-based surfaces and so cannot guarantee zero recrossing. For reactions with multiple channels, a single simple gate may have several disconnected facets; variable reaction-coordinate formulations can address such geometry. Quantum effects complicate a classical upper-bound interpretation. Thus a variational estimate and a dynamical correction should be documented as distinct modelling steps.

Summary

Recrossing causes ordinary TST to count crossings that do not become products. VTST moves the dividing surface and chooses the lowest estimated rate among a defined set of candidates, often finding a better bottleneck than the potential-energy saddle. The improvement is conditional on the surface family and equilibrium model. Trajectories, tunnelling corrections and experimental comparisons remain important.

Practice questions

1. What does recrossing do to a no-recrossing TST estimate in a classical model? Answer: It makes the positive-crossing estimate too high relative to net product formation.

2. Why may an entropic bottleneck exist with no electronic saddle? Answer: A restricted set of orientations or arrangements can limit flux even along a downhill potential-energy path.

3. Does minimising TST flux guarantee zero recrossing? Answer: No. It only optimises within the tested dividing-surface family and may omit coordinates or momentum dependence.

4. Why must all candidate-surface rates share the same reactant reference? Answer: Otherwise their numerical differences could reflect changed populations or standards rather than surface placement.

Sources: IUPAC Gold Book, variational transition-state theory; IUPAC Gold Book, dividing surface; Journal of Physical Chemistry B, moving dividing surfaces.