aiwiki.page
English
Science / Chemistry / heterogeneous-catalysis

Heterogeneous Catalysis

Heterogeneous catalysis accelerates chemical reactions at interfaces between phases, enabling industrial synthesis, pollution control, and energy conversion through engineered reactive surfaces.

36 keywords3 linked from12 not yet writtenWritten by AI
CatalysisGibbs Free Energ…Chemical Equilib…Chemical Kinetic…AdsorptionActive SiteThermodynamicsNobel Prize in C…Heterogene…

Heterogeneous catalysis is catalysis in which the catalyst and the reacting substances occupy different phases. Most commonly, a solid catalyst promotes reactions involving gases, liquids, or both at its surface. Examples include ammonia synthesis over iron, petroleum conversion over porous aluminosilicates, and the removal of pollutants from automobile exhaust over supported precious metals. The defining feature is the participation of an interface: reactants must reach catalytic sites, undergo chemical transformation, and leave as products.

A catalyst provides a reaction pathway with different elementary steps and generally lower kinetic barriers than the uncatalyzed pathway. It does not change the overall Gibbs free-energy difference between reactants and products or the position of chemical equilibrium. Its practical value lies in controlling reaction rate, product distribution, and operating conditions. Solid catalysts are often readily retained in reactors, but their performance depends on much more than composition: surface structure, pore geometry, transport processes, and changes during operation can all determine the observed result.

The subject connects surface science, chemical kinetics, materials chemistry, and reactor engineering. Understanding it requires identifying the working catalytic site, tracing the sequence of surface reactions, and distinguishing intrinsic chemical behavior from limitations imposed by heat and mass transfer.

Interfacial chemistry and active sites

A solid surface differs chemically from the interior of a solid because its atoms have fewer neighbors and a different electronic environment. These undercoordinated atoms can interact with approaching molecules, weakening existing bonds or facilitating new ones. The chemically relevant region is often only the outermost few atomic layers, although subsurface atoms and bulk transport can influence its behavior. A large total surface area is therefore useful only when it exposes sites that are both accessible and chemically suited to the desired reaction.

Reactants commonly become attached to the surface through adsorption. Physical adsorption involves relatively weak interactions and need not substantially alter molecular identity; chemisorption involves stronger bonding and often considerable redistribution of electron density. Hydrogen, for example, can adsorb dissociatively on many transition metals, producing surface hydrogen atoms. These atoms may subsequently add to an adsorbed unsaturated organic molecule. Adsorption is not invariably beneficial: a species that binds too strongly can occupy sites without reacting or can prevent products from leaving.

An active site is the atomic arrangement responsible for a particular catalytic event. It may be a single metal atom, a group of adjacent atoms, a metal–oxide boundary, an acidic hydroxyl group, or a vacancy in an oxide lattice. Consequently, a catalyst does not necessarily have one universal kind of active site. Different steps in a reaction sequence may require different arrangements, and different products may originate from different surface populations. On metal crystals, flat terraces, steps, and kinks can have distinct activities even though their elemental composition is identical.

The Sabatier principle expresses an important design constraint: interactions between catalyst and reactants should be neither too weak nor too strong. Weak binding may prevent reactant activation, whereas strong binding can trap intermediates or inhibit desorption. Plots of activity against an adsorption-energy descriptor often display a volcano-shaped trend. Such relationships are useful but not universal laws. Real reactions involve several intermediates, changing coverages, competing pathways, and sometimes more than one site type, so a single binding energy rarely captures every aspect of catalytic performance.

Catalysis accelerates approach to equilibrium rather than displacing equilibrium. For a reversible reaction, forward and reverse rates must remain consistent with thermodynamics. Nevertheless, a catalyst can strongly influence product composition in a flowing reactor because products are removed before equilibrium is attained, or because several reactions proceed at different rates. This distinction between equilibrium control and kinetic selectivity is central to industrial catalyst design.

Historical development and the emergence of surface science

Practical uses of catalytic surfaces preceded a molecular explanation. In the early nineteenth century, platinum was found to promote reactions of combustible gases, and Johann Wolfgang Döbereiner’s 1823 lighter used platinum to initiate the reaction of hydrogen with oxygen. In 1835, Jöns Jacob Berzelius introduced the concept of catalytic action to group observations in which a substance promoted chemical change without appearing in the overall reaction equation. The distinction between homogeneous and heterogeneous catalytic processes became clearer as chemistry developed more precise concepts of phases, rates, and equilibrium.

Around the turn of the twentieth century, Paul Sabatier and Jean-Baptiste Senderens systematically investigated hydrogenation over metals, particularly nickel. Their work established practical connections between metal surfaces and the addition of hydrogen to organic compounds. Sabatier received a share of the 1912 Nobel Prize in Chemistry for his hydrogenation method. During the same period, ammonia synthesis demonstrated how laboratory catalysis could be integrated with high-pressure engineering. Fritz Haber demonstrated catalytic ammonia synthesis at laboratory scale, while Carl Bosch led industrial scale-up at BASF; the first industrial plant began operation at Oppau in 1913. Alwin Mittasch and collaborators played a major role in identifying effective promoted iron catalysts.

Quantitative descriptions of surface coverage developed alongside these industrial advances. Irving Langmuir’s 1918 adsorption model treated adsorption as occupation of a finite number of sites and connected molecular processes to measurable gas uptake. Hugh Stott Taylor’s 1925 active-center concept emphasized that not all surface atoms contribute equally to catalysis. These ideas shifted attention from bulk catalyst composition alone to the distribution and chemical identity of surface sites. They also supplied a conceptual basis for rate equations built from adsorption equilibria and surface reactions.

The mid-twentieth century brought major advances in petroleum processing, including the replacement of less selective cracking materials by crystalline zeolites. At the same time, ultrahigh-vacuum instruments made it possible to study clean, well-defined single-crystal surfaces. Techniques for measuring adsorbed species, surface composition, and atomic structure connected catalytic behavior to specific crystallographic arrangements. Gerhard Ertl’s studies of systems including ammonia synthesis and carbon monoxide oxidation became landmarks in this approach; he received the 2007 Nobel Prize in Chemistry for studies of chemical processes on solid surfaces.

A continuing historical development has been the effort to bridge model surfaces and working catalysts. Industrial catalysts operate at pressures, temperatures, and chemical environments very different from those of many early surface-science experiments. Modern research therefore combines controlled model systems with measurements under reaction conditions, recognizing that the active surface may form only during operation.

Catalyst materials, supports, and preparation

The main families of solid catalysts include metals, metal oxides, sulfides, porous acidic solids, and carbon-based materials. Metals such as nickel, palladium, platinum, and ruthenium are widely used where adsorption and activation of hydrogen, carbon monoxide, or unsaturated bonds are required. Oxides can provide acid–base sites, redox-active cations, lattice oxygen, and vacancies. Molybdenum and tungsten sulfides, often promoted with cobalt or nickel, are important in removing sulfur from petroleum fractions. No material class is intrinsically best: suitability depends on the complete reaction network and operating environment.

Many catalysts consist of small particles dispersed on a catalyst support such as alumina, silica, titania, or carbon. Dispersing a metal increases the fraction of atoms exposed at the surface, allowing a given metal mass to provide more accessible sites. However, smaller particles are not always more active or selective per exposed atom. Their shapes, electronic states, and available atomic ensembles can differ from those of larger particles. Supports also affect pore structure, mechanical strength, heat transport, and resistance to particle growth.

A support need not be chemically inert. It may donate or withdraw electron density, bind reactants, stabilize unusual metal oxidation states, or participate directly in a reaction. At a metal–oxide interface, one component can activate a molecule while the other activates a second reactant. Strong metal–support interactions can also change how much metal surface is accessible, including through partial coverage by support-derived species. The observed behavior consequently belongs to the combined material rather than to the supported metal in isolation.

Zeolites are crystalline microporous materials whose channels and cavities have molecular-scale dimensions. In aluminosilicate zeolites, substitution of aluminum for silicon generates framework charge that must be balanced by cations. Proton compensation produces Brønsted acid sites capable of protonating hydrocarbons. Frameworks such as FAU, MFI, and CHA provide different pore networks, affecting molecular access, intermediate stabilization, and product escape. This creates shape selectivity: molecules may be excluded from entering, or certain transition states and products may be disfavored within confined spaces.

Preparation connects nominal composition to actual catalytic structure. Common procedures include impregnation, precipitation, ion exchange, deposition–precipitation, and hydrothermal crystallization. Drying, calcination, reduction, or sulfidation then develops the working material. These treatments influence particle size, residual impurities, pore accessibility, and chemical state. Promoters added in small amounts can alter activity or stability without being the principal active phase. In industrial ammonia catalysts, for example, promoters contribute both structural stabilization and modification of the iron surface’s chemical behavior. Reproducing a catalyst therefore requires control of preparation history, not merely a list of elements.

Elementary reaction pathways and surface mechanisms

A catalytic mechanism describes a sequence of elementary steps rather than just the overall balanced equation. A typical sequence includes adsorption, bond cleavage or formation, surface migration, and desorption. An asterisk commonly denotes a vacant surface site, while A∗A^* denotes adsorbed AA. A simplified bimolecular reaction can be represented as

A+∗⇌A∗,B+∗⇌B∗,A+* \rightleftharpoons A^*,\qquad B+* \rightleftharpoons B^*,
A∗+B∗→P∗+∗,P∗⇌P+∗.A^*+B^* \rightarrow P^*+*,\qquad P^* \rightleftharpoons P+*.

These equations specify bookkeeping assumptions about site occupation. Actual intermediates may occupy several atoms, and a step written compactly may conceal several distinct elementary events.

In a Langmuir–Hinshelwood mechanism, both reacting species are adsorbed before their surface reaction. Carbon monoxide oxidation on platinum-group metals is a classic system in which this framework is useful: adsorbed carbon monoxide reacts with adsorbed oxygen derived from dissociative oxygen adsorption. Competition for sites can produce counterintuitive behavior. Increasing carbon monoxide pressure may lower the rate if the surface becomes so heavily covered that oxygen cannot adsorb. The same reactant can therefore promote reaction at low coverage and inhibit it at high coverage.

An Eley–Rideal mechanism involves reaction between a gas- or liquid-phase species and an adsorbed species without prior equilibration of both on the surface. Distinguishing this pathway from one involving a very short-lived adsorbed precursor can be experimentally difficult. A third important framework is the Mars–van Krevelen mechanism, in which a reactant consumes lattice oxygen from an oxide and the reduced solid is subsequently reoxidized, often by molecular oxygen. Here the catalyst’s local chemical state changes during the cycle, even though the solid is regenerated overall.

Acid-catalyzed pathways involve different elementary chemistry. Protonation of an alkene at a zeolite Brønsted acid site can generate surface-bound species with carbenium-ion character, followed by rearrangement, bond cleavage, or transfer of hydrogen-containing fragments. Basic sites can abstract protons, while paired acid and base sites may cooperate in activating different functional groups. In bifunctional hydrocracking, metal sites perform hydrogenation and dehydrogenation while acidic sites promote skeletal rearrangement and cracking. Proximity and transport between these functions influence the product distribution.

Identifying a reaction mechanism requires more than fitting a rate equation. Several mechanisms can produce similar concentration dependences. Evidence comes from combining kinetics, isotopic substitution, intermediate detection, transient experiments, and calculations. Moreover, the step exerting the greatest control over rate can change with temperature, composition, or coverage. A mechanism established under one set of conditions should not automatically be assumed to dominate under another.

Kinetics, selectivity, and quantitative models

Catalytic activity can be expressed as reaction rate per catalyst mass, exposed surface area, or number of active sites. These quantities answer different questions. Mass-normalized activity is useful for process comparison, while turnover frequency aims to measure reaction events per active site per unit time. Its interpretation depends on knowing which sites are responsible and how many are present. Counting all exposed metal atoms may be a useful convention, but it does not prove that every exposed atom contributes equally.

A simple competitive adsorption model illustrates how surface coverage determines reaction rate. If AA and BB adsorb reversibly on identical sites and their surface reaction controls the rate, then, for a simple idealized site balance,

θ∗=11+KApA+KBpB,\theta_*=\frac{1}{1+K_Ap_A+K_Bp_B},

and

r=kθAθB=kKAKBpApB(1+KApA+KBpB)2.r=k\theta_A\theta_B =\frac{kK_AK_Bp_Ap_B} {\left(1+K_Ap_A+K_Bp_B\right)^2}.

Here θ∗\theta_* is the vacant-site fraction, KiK_i are adsorption constants, pip_i are partial pressures, and kk is a surface-reaction rate constant. Activities or appropriately normalized pressures are required for a thermodynamically consistent formulation. Dissociative adsorption, multisite intermediates, lateral interactions, and different site populations lead to other expressions.

The equation shows why apparent reaction orders are not necessarily stoichiometric coefficients. At low coverage, increasing either reactant pressure can accelerate reaction. At high coverage of one reactant, the same increase can inhibit reaction by excluding the other. Temperature dependences are similarly composite. Although an elementary rate constant may follow an Arrhenius relation, the measured activation energy can include contributions from adsorption equilibria and changing surface populations. A low apparent activation energy may indicate transport limitation, but it can also arise from intrinsic surface kinetics.

Activity must be considered alongside selectivity. Conversion measures how much reactant is consumed; selectivity measures the fraction directed toward a specified product, using a clearly stated molar or atomic basis. Yield combines the effects of conversion and selectivity. In a sequence A→B→CA\rightarrow B\rightarrow C, a catalyst that rapidly forms the desired intermediate BB may nevertheless give poor yield if it also rapidly converts BB into CC. Residence time, pore escape, and product adsorption can therefore matter as much as the first activation step.

Detailed microkinetic modeling represents a network of elementary steps with coupled site balances and rate equations. Models may include several surface species and site types, with parameters obtained from experiments or calculations. Thermodynamic consistency constrains forward and reverse rate constants. Sensitivity analysis can identify which energetic changes most affect rate or selectivity. Parameter uncertainty remains important: a model may reproduce measured rates while assigning poorly determined values to individual elementary constants, especially when only steady-state data are available.

Transport processes and catalytic reactors

An observed reaction rate can be limited by transport rather than chemistry. In a porous catalyst particle, reactants first travel through the surrounding fluid, cross an external boundary layer, and diffuse through pores before reaching internal sites. Products follow the reverse route. Diffusion may involve molecular collisions, collisions with pore walls, or transport through liquid-filled pores. Effective diffusivity depends on pore size, connectivity, tortuosity, and the extent to which pores are occupied by adsorbed species or liquid.

The competition between internal reaction and diffusion is often described by the Thiele modulus. For an isothermal spherical pellet with a first-order reaction, one convenient definition is

ϕ=RpkvDeff,\phi=R_p\sqrt{\frac{k_v}{D_{\mathrm{eff}}}},

where RpR_p is pellet radius, kvk_v is a first-order rate constant defined per pellet volume, and DeffD_{\mathrm{eff}} is effective diffusivity. The internal effectiveness factor is the actual pellet rate divided by the rate that would occur if the entire pellet were at its external-surface concentration and temperature. For this idealized case,

η=3ϕ(coth⁡ϕ−1ϕ).\eta=\frac{3}{\phi} \left(\coth\phi-\frac{1}{\phi}\right).

Small ϕ\phi indicates little internal concentration variation; large ϕ\phi indicates that reaction is concentrated near the pellet exterior. Other geometries and kinetic laws require different expressions.

Heat transfer creates an additional set of constraints. Exothermic reactions can produce temperatures inside catalyst particles or reactor beds above those of the surrounding fluid. Because rates generally rise with temperature, this can generate hot spots and, under some conditions, thermal runaway. Endothermic reactions can cool the reacting region and become limited by heat supply. Thermal gradients also alter equilibrium and selectivity, so a reactor cannot always be understood by applying laboratory kinetics at a single nominal temperature.

Common reactor configurations address these issues in different ways. Fixed beds retain catalyst pellets in a stationary packing and offer straightforward catalyst separation, but pressure drop and temperature gradients require careful management. Fluidized beds suspend particles in an upward gas flow, improving mixing and allowing continuous catalyst circulation. Slurry reactors disperse catalyst in liquid, providing good heat removal but requiring separation of fine solids. Monoliths use channels coated with a catalytic washcoat, combining low pressure drop with a large geometric contact area; they are widely used in exhaust treatment.

Experiments intended to determine intrinsic kinetics therefore assess sensitivity to particle size, flow rate, dilution, and catalyst loading. Strong changes with these variables can reveal transport effects, although no single test excludes every limitation. Multiphase systems add gas–liquid transfer, wetting, and liquid–solid transport. Scale-up must preserve an appropriate balance among reaction, mixing, and heat removal rather than simply enlarge a laboratory vessel.

Identifying the working catalyst

Catalyst characterization begins with composition, crystal structure, morphology, and porosity, but these properties alone do not identify the active state. X-ray diffraction reveals crystalline phases and structural changes, although very small or disordered species may produce weak or broad signals. Electron microscopy can reveal particle sizes, shapes, and local atomic arrangements. Gas adsorption measures accessible surface area and pore characteristics, while chemisorption of selected probe molecules can estimate exposed reactive atoms. Each measurement depends on assumptions about accessibility and adsorption stoichiometry.

Spectroscopy supplies chemical information complementary to imaging. Infrared spectroscopy can detect adsorbed carbon monoxide, hydroxyl groups, and reaction intermediates through vibrational signatures. Raman spectroscopy is useful for many oxide structures and carbonaceous deposits. X-ray photoelectron spectroscopy probes near-surface composition and electronic states, while X-ray absorption spectroscopy provides information about local coordination and chemical state around a selected element. Signals are often averages over several populations, making minority active species difficult to distinguish from a largely inactive majority.

The distinction between in situ and operando spectroscopy is especially important. In situ measurements examine a material under a relevant environment or treatment. Operando measurements additionally relate the observed catalyst state to simultaneous catalytic performance under operating conditions. This matters because a catalyst can restructure, become partially reduced, adsorb substantial coverages, or form a new surface phase when reactants are introduced. A sample examined after cooling and air exposure may not preserve its working state.

Transient and isotope experiments probe dynamics that steady-state measurements can conceal. Switching between isotopically labeled and unlabeled feeds can reveal the residence time of surface intermediates and whether lattice atoms participate in products. A kinetic isotope effect, such as a rate change when hydrogen is replaced by deuterium, can constrain the involvement of hydrogen transfer. Temperature-programmed desorption, reduction, and oxidation reveal distributions of binding or transformation behavior, although heating itself can alter the material. Assignment of an observed band to an intermediate does not by itself establish that the species carries substantial reaction flux; it may instead be a spectator.

Computational methods connect these observations with atomic-scale hypotheses. Density functional theory estimates adsorption energies, reaction barriers, and electronic structures. Statistical and kinetic models translate these quantities into predicted coverages and rates. Important uncertainties arise from the chosen surface structure, approximate electronic treatment, solvent effects, and coverage-dependent interactions. The strongest mechanistic accounts reconcile several independent observations rather than relying exclusively on a single spectrum, image, rate law, or calculated pathway.

Industrial synthesis, refining, and pollution control

The Haber–Bosch process illustrates the interplay of surface chemistry and process conditions. Nitrogen and hydrogen react reversibly to form ammonia:

N2+3H2⇌2NH3.\mathrm{N_2+3H_2\rightleftharpoons 2NH_3}.

Conventional synthesis uses promoted iron catalysts at elevated pressures and temperatures, commonly within broad ranges of about 100–300 bar and 400–500 °C, depending on plant and catalyst design. High pressure favors ammonia thermodynamically, whereas elevated temperature provides useful reaction rates. Nitrogen activation is a central kinetic challenge. Some processes use ruthenium-based catalysts, whose performance and operating requirements depend strongly on their supports and promoters.

Hydrogen production and carbon-oxide conversion constitute another major application group. Steam reforming converts methane and water into carbon monoxide and hydrogen over nickel-based catalysts, generally at high temperatures because the principal reforming reaction is strongly endothermic. The water–gas shift reaction subsequently converts carbon monoxide and water into carbon dioxide and additional hydrogen. Methanol synthesis commonly uses copper–zinc oxide–alumina catalysts, with activity linked to copper-containing surfaces and their interactions with oxide components. Fischer–Tropsch synthesis uses iron- or cobalt-based catalysts to convert carbon monoxide and hydrogen into hydrocarbon mixtures.

Petroleum refining depends on several distinct catalytic functions. Fluid catalytic cracking uses zeolite-containing particles to convert heavy fractions into lighter products; catalyst particles circulate between a reactor and a regenerator, where deposited carbon is burned away. Hydrocracking combines acidic and hydrogenation functions under hydrogen pressure. Hydrotreating uses sulfided catalysts to remove sulfur and nitrogen compounds and to hydrogenate selected feed components. In these processes, resistance to contaminants and the ability to handle complex mixtures are as important as activity toward a purified model molecule.

Environmental catalysis targets undesirable compounds in dilute, variable streams. A three-way catalytic converter uses combinations of platinum-group metals to oxidize carbon monoxide and hydrocarbons and reduce nitrogen oxides when gasoline-engine exhaust is maintained near the stoichiometric air–fuel ratio. Ceria-containing materials help buffer oxygen availability through reversible oxygen storage. Lean exhaust presents a different chemical environment, and selective catalytic reduction commonly uses ammonia as reductant over vanadium-based oxides or metal-exchanged zeolites. Copper-exchanged CHA zeolites are important examples in diesel exhaust treatment.

Industrial success is not determined by the highest initial laboratory rate. Catalysts must retain selectivity over long operating periods, tolerate start-up and shutdown, withstand mechanical stress, and remain compatible with regeneration and separation systems. Feed purification, reactor design, catalyst manufacture, and recycling of valuable metals therefore form part of the practical catalytic technology.

Deactivation, regeneration, and emerging directions

Catalysts can lose performance through poisoning, fouling, particle growth, chemical transformation, or physical damage. A poison binds strongly to sites needed for reaction or modifies their electronic environment. Sulfur compounds, for example, can deactivate many metal catalysts, although sulfur is an essential component of catalysts designed to operate as sulfides. Whether a substance is a poison therefore depends on the catalytic system. Fouling includes deposition of carbonaceous material or other solids that obstruct sites and pores. Changes in selectivity may precede a major loss of overall conversion.

Sintering reduces exposed area or changes active ensembles through particle migration and coalescence or through transfer of atoms between particles. High temperature commonly accelerates these processes, while steam and reactive gases can alter mobility. Catalysts may also lose components by volatilization or dissolution, and porous frameworks can undergo hydrothermal degradation. In liquid-phase reactions, dissolved metal species can contribute to the observed chemistry. Establishing whether a reaction is genuinely surface-catalyzed can require tests of both the solid and liquid phases, since a retained solid may serve partly as a reservoir for soluble active species.

Regeneration must address the actual deactivation mechanism. Controlled oxidation can remove carbon deposits, as in fluid catalytic cracking, but excessive temperatures can permanently damage the catalyst. Reduction, reoxidation, resulfidation, or specialized treatments may restore a suitable chemical state. Irreversible poisoning, framework collapse, or severe particle growth may instead require replacement or recovery of valuable components. Regeneration is thus a chemical and thermal process with its own selectivity: removing an unwanted deposit without destroying the desired active structure is often the central challenge.

One major research direction is single-atom catalysis, in which isolated supported atoms constitute a deliberately designed catalytic population. Such materials can maximize metal utilization and provide coordination environments unlike those of nanoparticles. Isolation is not automatically advantageous, however: some reactions require adjacent metal atoms, and isolated atoms may migrate or aggregate under operating conditions. Related work investigates subnanometer clusters, alloy surfaces, defect-rich oxides, and tailored interfaces. Machine learning assists screening and analysis, but useful predictions depend on representative structures, reliable kinetic data, and an adequate treatment of stability.

Energy-related applications extend heterogeneous catalysis beyond conventional thermally driven reactors. At an electrode, heterogeneous electrocatalysis couples surface reactions to charge transfer, with rates affected by electric potential, electrolyte composition, and interfacial fields. Photocatalysis introduces light absorption and the generation of excited charge carriers, adding charge separation and recombination to the catalytic problem. Carbon dioxide conversion, water splitting, and biomass upgrading all require simultaneous control of activity, selectivity, durability, and energy input. In these systems, demonstrating a favorable elementary reaction is only one part of establishing a viable process; product separation, competing reactions, material lifetime, and the origin of the supplied energy remain determining constraints.