Limits of a One-Descriptor Volcano

Introduction

A one-descriptor volcano is a helpful screen, but a catalyst is more than one binding energy. A different alloy may expose a new site, high reactant pressure may crowd the surface, and a porous particle may be limited by diffusion. In these cases a plotted rate can depart from the simple trend even if the original curve was useful for the first family. The scientific response is to diagnose the departure, not force every point onto one smooth line.

Core explanation

A simple volcano commonly assumes a shared catalytic mechanism across its candidate set. If all catalysts transform A through the same sequence, a descriptor tied to A stability may capture much of the trend. But an oxide might use lattice oxygen, a metal–support boundary might perform a bifunctional step, or a ligand-bearing molecular catalyst might follow a different cycle. Their rates can depend on variables not represented by the original descriptor. A catalyst near the predicted peak can be slow if the supposed active site is absent under working conditions.

Surface coverage is another source of failure. A low-coverage adsorption energy may be calculated for one isolated molecule on a clean surface. Under operating pressure, several adsorbates can compete for sites. Lateral interactions may change the energetic cost of further adsorption or reaction. Product inhibition can block sites; a poison in trace feed can dominate occupancy. An ACS account of theory–experiment comparison shows why coverage-dependent calculations and microkinetic treatment can matter when linking atomic energies to actual rates.

Measured activity may be shaped by external mass transfer, pore diffusion or heat transfer rather than intrinsic catalytic chemistry. If transport limits supply, a very active material can appear no better than a less active one. Particle size, stirring, flow and dilution tests can reveal this problem. In an exothermic reaction, local hot spots can make the observed rate look unusually high or alter selectivity. Normalising by catalyst mass does not remove these effects.

Even within a shared mechanism, one descriptor may omit independent degrees of freedom. Two surfaces can bind an intermediate equally while stabilising its transition state differently. A product may desorb readily on one surface because of geometric arrangement rather than the chosen intermediate's adsorption energy. A second descriptor such as site geometry, proton-transfer ability or work function may improve prediction, provided it is mechanistically motivated and validated with new data.

The ACS volcano-construction study treats such diagrams as relationships to construct and validate, not universal laws. Report the catalyst family, operating window, site model, rate normalization and uncertainty. A curve built from three points should be a hypothesis with broad uncertainty, especially outside the measured range.

Step-by-step reasoning

1. Identify the assumptions behind the original descriptor–activity trend. 2. Check whether a new catalyst retains the same active surface and mechanism. 3. Compare low- and high-coverage conditions and look for inhibition or interactions. 4. Test diffusion and heat effects before calling a rate intrinsic. 5. Add a second variable or new mechanism only when independent data support it.

Visual explanation

Draw a volcano with three catalyst families as different symbols. One family follows the hill. A second family sits below it because product coverage blocks sites, and a third falls above it because an interface offers a new route. Add an inset showing measured rate plateauing with catalyst activity when reactant transport becomes limiting. The figure warns that a smooth fit to one group need not rank all materials.

Real-world analogy

A chart relating engine horsepower to travel time might work for similar cars on an open track. It fails in city traffic where roads limit speed, and it fails for a train using a different route. Horsepower is still meaningful, but the context has changed. Binding energy can likewise remain chemically informative while ceasing to predict observed throughput under different mechanisms or transport conditions.

Real-world example

A metal powder looks slower than expected from a favourable computed adsorption energy. Increasing stirring improves its measured rate, while changing catalyst mass does not change rate proportionally. This points toward external transport rather than an incorrect adsorption calculation alone. A second powder does not respond to stirring but shows strong product inhibition. Both deviate from the original volcano for different reasons, demanding different fixes.

Why?

Why does a transport-limited catalyst sometimes appear to have a low apparent activation energy? Increasing temperature may accelerate chemical steps more than diffusion, so transport becomes the bottleneck and the observed rate changes less with temperature than intrinsic chemistry would. A misleading activation energy can result if the kinetic regime is not checked.

Common misconception

“An outlier invalidates every data point on the volcano” is excessive; the trend may have a clear domain. “A rate plateau proves an optimum binding energy” could instead reflect supply limitation. “Calculated clean-surface adsorption energy is the operating-site energy” ignores adsorbates, reconstruction and solvent. “More descriptors always improve prediction” can merely overfit limited data.

Worked example

An illustrative volcano model predicts rates of 10 and 8 mmol/h for catalysts A and B under one condition. Experiments give 4 and 7 mmol/h. If A's observed rate rises from 4 to 9 mmol/h when stirring speed doubles, while B remains near 7, A was likely transport constrained in the first test; its chemical activity may still accord with the prediction. Suppose an independent spectroscopy test shows that A becomes covered by product at higher pressure and its rate drops again. The same material has different apparent rank depending on transport and coverage. A fair comparison requires kinetic-regime control and defined pressure, not selective use of one convenient measurement.

Quick check

1. What does a strong increase in measured rate when stirring is increased suggest? Answer: External mass-transfer limitation may have affected the original rate, though further checks are needed.

Exam focus

List three ways a one-descriptor ranking can fail: mechanism/site change, coverage interactions and transport are central examples. State a diagnostic experiment for each. Explain why a new descriptor should be mechanistically justified and tested on data beyond the fitting set.

Advanced insight

Microkinetic models couple adsorption, surface reactions, desorption and site balances. They can predict which coverages and elementary steps govern rate under specific conditions. Yet their output is only as reliable as the assumed active-site model and energetic parameters. A more elaborate model does not automatically repair incorrect site identity or unmeasured transport limits.

Summary

A one-descriptor volcano is conditional on shared chemistry and controlled measurement. Mechanism changes, adsorbate coverage and transport can shift or obscure the trend. Outliers are opportunities to test the model's domain and identify missing physics.

Practice questions

1. Why might a clean-surface adsorption energy fail at high pressure? Answer: High coverage and lateral interactions can change adsorption and reaction energetics. 2. What experimental variable can help diagnose external liquid-phase mass transfer? Answer: Stirring speed is one useful variable; flow and particle-size tests can help too. 3. A metal–support interface creates a new pathway. Should it automatically lie on a metal-only volcano? Answer: No. Its mechanism and active site may differ from the family used to establish the curve. 4. What is the risk of fitting many descriptors to only a few catalysts? Answer: Overfitting can produce an attractive retrospective fit that predicts new catalysts poorly.