Catalytic Hydrogenation Mechanism
A shiny gray powder of palladium on carbon turns a bubbling, colorless alkene solution into a saturated alkane. Nothing about the flask tells you why. There is no visible spark, no color change, no obvious electron transfer. There is just hydrogen gas disappearing and a new C–H-rich product appearing. That product has a stereochemical signature so reliable that chemists use it to map molecular shape. Catalytic hydrogenation looks, on paper, like the simplest reaction in organic chemistry: add H–H across C=C. But that simplicity is deceptive. The reaction happens nowhere near the alkene's own π system in the usual electrophile-nucleophile sense. It happens on a metal surface instead. There, a choreography of adsorption, bond-breaking, and stepwise delivery unfolds. This choreography determines not just whether the alkene reacts. It also determines which face of the molecule gets hydrogenated and from which side. This matters because catalytic hydrogenation is the primary laboratory and industrial method. It converts alkenes and alkynes into alkanes and alkenes, respectively. It is one of the few reactions whose outcome is dictated almost entirely by a solid surface. Solution-phase electronics play little role. Margarine is hydrogenated vegetable oil. Drug syntheses rely on asymmetric hydrogenation to install single stereocenters with near-perfect enantiocontrol. The reaction also serves a diagnostic role. Because it delivers both hydrogens to the same face of the alkene, its product stereochemistry is a direct, physical readout. That readout reveals the alkene's geometry. It is the same geometry probed indirectly by measuring Heats of Hydrogenation as a Stability Probe. The big picture is this: hydrogenation is not a single elementary step. It is a sequence of surface events instead. H2 dissociates on the metal. The alkene physisorbs and then chemisorbs through its π system. Hydrogen atoms migrate across the metal surface. They add to the coordinated alkene one at a time. Both new C–H bonds form while the substrate sits on the same patch of metal. Because of this, they are delivered from the same face, giving syn addition. This mechanistic fact, not some invisible electronic effect, is why hydrogenation is stereospecific. It is also why hydrogenation interacts so intimately with Alkene Structure and the Pi Bond. The very π electrons that make an alkene nucleophilic toward bromine or peracids are what let it bind. They allow it to bind sideways-on to the catalyst. This article builds the mechanism step by step. It covers how H2 splits homolytically on a metal surface. It explains how the alkene chemisorbs through its π cloud. It shows how sequential H-atom transfer explains syn stereochemistry. It explores why more substituted (more stable) alkenes often react slower. It also shows how all of this generalizes to alkyne semi-hydrogenation with poisoned catalysts. It closes with worked examples. These span a simple textbook hydrogenation, an industrial application (edible oil hydrogenation and trans-fat formation), and a mechanistic puzzle from an exam-style stereochemistry problem.
- Hydrogenation is heterogeneous: H2 and the alkene both adsorb onto the same metal surface (Pd, Pt, Ni, Rh) before any bond changes.
- H2 dissociates homolytically on the metal into two surface hydrides (Had); this is the key bond-breaking step, not a solution-phase process.
- The alkene binds side-on through its π system (chemisorption), donating electron density into empty metal d-orbitals and accepting some back — a two-way, Dewar–Chatt–Duncanson-type interaction.
- Two H atoms are delivered sequentially from the same metal face, so both new C–H bonds form on one face of the alkene: syn addition, confirmed by deuterium-labeling studies.
- Reaction rate depends on how well the alkene can adsorb: less-hindered, less-substituted alkenes usually adsorb and react faster, even though they are thermodynamically less stable.
- Catalyst poisoning (Lindlar's catalyst) can arrest the sequence after only one H2 addition, converting alkynes cleanly into cis-alkenes.
- Key idea: hydrogenation stereochemistry is a direct, physical consequence of both new bonds being delivered from the same metal surface to the same face of the substrate.
Catalytic hydrogenation is a surface relay race: H2 splits into two metal-bound hydrogen atoms, the alkene lies flat against the same surface through its π electrons, and the two hydrogens hop across from the metal to one face of the double bond — always the face touching the metal — giving clean syn addition without any discrete carbocation or free radical ever forming in solution.
Where this lives today
Industrial edible-oil processing hydrogenates polyunsaturated vegetable oils to raise melting point and improve shelf stability, though regulatory pressure (driven directly by the trans-fat mechanism in Example 2) has pushed the food industry toward fully hydrogenated and interesterified fats, or away from partial hydrogenation altogether. Pharmaceutical manufacturing depends on catalytic hydrogenation constantly: chiral drugs such as L-DOPA (for Parkinson's disease) and many amino-acid-derived active ingredients are made by asymmetric catalytic hydrogenation, in which a chiral ligand on a Rh or Ru center (a direct molecular descendant of Wilkinson's catalyst mechanism) makes the two faces of a prochiral alkene inequivalent, so that syn addition from only one face is favored — this is the mechanistic basis of the Nobel Prize-winning work of Knowles and Noyori. Catalytic hydrogenation is also the standard laboratory method for removing protecting groups (benzyl ethers, Cbz carbamates) via hydrogenolysis, a mechanistically related process where the metal surface cleaves a C–O or C–N bond instead of a π bond, using the same H2-dissociation chemistry. In petroleum refining, hydrotreating and hydrocracking use related metal-sulfide catalysts to hydrogenate and desulfurize crude oil fractions at industrial scale. And Lindlar's catalyst, arising directly from the ensemble-size poisoning logic in Mechanism 2, remains the textbook and real-world method for making cis-alkenes in total synthesis — for example, in the synthesis of pheromones and polyketide natural products where a single wrong double-bond geometry ruins the target.
Ask the tutor: Why does hydrogenation of a trisubstituted alkene sometimes give a small amount of the "wrong" (anti-looking) diastereomer even under clean, H2-rich Pd/D2 conditions?
The physical picture
Think of the catalyst surface — palladium, platinum, nickel, or rhodium, usually dispersed as nanoparticles on carbon or alumina to maximize surface area — as a two-dimensional lattice of metal atoms, each with partially filled d-orbitals. This is the essential difference from every other alkene addition reaction you have met: instead of a reagent's HOMO/LUMO interacting directly with the alkene's π system in solution, the reactive surface has a near-continuum of electronic states (a "band" rather than discrete orbitals) that can donate into the alkene's π* and accept from the alkene's π simultaneously.
Step 1 is H2 activation. Hydrogen gas approaches the bare metal surface and undergoes dissociative chemisorption: the H–H σ bond, while still largely intact, begins to overlap with empty (or partially filled) metal orbitals. Electron density flows from the H2 σ-bonding orbital into the metal, while the metal simultaneously back-donates electron density into the H2 σ orbital. This synergic bonding is exactly the same conceptual pattern (donation from a filled ligand orbital, back-donation from filled metal d into a ligand's antibonding orbital) that governs alkene–metal π complexes and organometallic CO binding — the Dewar–Chatt–Duncanson picture. As back-donation into σ grows, the H–H bond weakens and finally snaps homolytically, releasing two hydrogen atoms, each captured as a metal hydride, denoted M–H. Because H2 splits with essentially no activation barrier on clean transition-metal surfaces (this is precisely why Pd, Pt, Ni, and Rh are the workhorse hydrogenation catalysts — they have the right d-band position to do this efficiently), the metal surface becomes tiled with mobile, adsorbed hydrogen atoms (Had) that can migrate across the metal like beads sliding on a wire.
Step 2 is alkene chemisorption. The alkene's filled π orbital donates electron density sideways into an empty metal orbital (π-donation), while filled metal d-orbitals back-donate into the alkene's empty π orbital. This dual interaction is stabilizing enough to pull the alkene down flat against the surface, with both alkene carbons and the metal roughly coplanar — the alkene is now chemisorbed, not just physically stuck near the surface. Back-donation into π has a crucial secondary effect: it partially populates an antibonding orbital of the C=C bond, weakening it and lengthening the C–C distance, priming the alkene for the addition that follows. Substituents matter here: a more heavily substituted double bond is more sterically hindered in trying to lie flat against a crowded metal surface, and steric crowding at the metal also slows access of the surface hydrides — this is the mechanistic root of the empirical rule that less-substituted alkenes hydrogenate faster despite being thermodynamically less stable.
Step 3 is sequential hydrogen delivery. A mobile surface hydride migrates to the carbon of the chemisorbed alkene and forms the first new C–H bond, while the other alkene carbon remains bonded to the metal as a metal alkyl (σ-bonded intermediate) — the alkene has gone from η²-π-bound to a metal–alkyl σ-complex. A second surface hydride then migrates to that remaining metal–carbon bond, reductively eliminating the alkane and regenerating a bare patch of catalyst. Both hydrogen atoms necessarily come from the same face of the substrate — the face touching the metal — because the alkyl intermediate cannot rotate freely while still tethered to the surface (rotation about the former C=C bond, now a C–C single bond in the metal alkyl, is geometrically blocked by the metal attachment and would require desorption). The outcome is syn addition, and it is stereospecific: cis- and trans-disubstituted alkenes give diastereomeric dideuterated products when D2 is used, exactly as expected if both D atoms land on one face.
The core ideas
The overall stoichiometric transformation is simple:
where: the metal catalyst is not consumed and does not appear in the product; it only provides the surface on which every bond-breaking and bond-making event occurs. The energetics of the overall reaction were already established through heats of hydrogenation, ΔH°hydrog, but the mechanism explains why that number is negative and reproducible: converting a weak, polarizable π bond plus a strong H–H σ bond (bond dissociation enthalpy 104 kcal/mol) into two new, strong C–H σ bonds is thermodynamically downhill by roughly 25–30 kcal/mol per double bond, but the rate at which that thermodynamic driving force is realized depends entirely on the kinetics of surface adsorption, which is a separate question from the thermodynamics. This decoupling of thermodynamics from mechanism is the single most important conceptual lesson of catalytic hydrogenation: the catalyst does not change ΔH° of the reaction (a state function, fixed by initial and final bonds), it only lowers the activation barrier by providing an alternative, stepwise pathway.
where: BDE(C=C π) is the π-bond contribution to the double bond (roughly 65 kcal/mol), BDE(H–H) is 104 kcal/mol, and BDE(new C–H) is about 98–101 kcal/mol per bond; the catalyst appears nowhere in this expression because it is a true catalyst, changing only the path, not the endpoints. The rate law for a Langmuir–Hinshelwood surface mechanism (both reactants adsorbed before reacting) is:
where: θH and θalkene are the fractional surface coverages of adsorbed hydrogen and adsorbed alkene, respectively, and k is the intrinsic surface rate constant for hydrogen-atom transfer to the chemisorbed alkyl; because θH and θalkene both depend on gas pressure/concentration through competitive adsorption equilibria, doubling H2 pressure does not simply double the rate — it can even decrease the rate if H2 outcompetes the alkene for surface sites, a genuinely non-trivial prediction that follows only from the adsorption mechanism, not from a naive bimolecular collision picture.
Goal: show, curved-arrow by curved-arrow, why hydrogenation of (Z)- or (E)-2-butene with D2/Pd delivers both deuteriums to one face, producing a single, predictable diastereomer of 2,3-dideuteriobutane.
Step 1 — D2 approaches the bare Pd surface. D2's σ bond lies parallel to the metal surface. The filled σ(D–D) orbital donates electron density into an empty Pd d-orbital (curved arrow from the D–D bond into the metal); simultaneously a filled, symmetry-matched Pd d-orbital donates electron density back into the empty σ(D–D) orbital (curved arrow from metal into the D–D bond, on the antibonding side). This synergic donation/back-donation is stabilizing on its own, but as back-donation increases it also populates σ, directly weakening the D–D bond.
Step 2 — homolytic cleavage. As σ* population grows, the D–D bond stretches past its elastic limit and cleaves homolytically: one curved arrow shows one electron of the D–D bond moving to one Pd atom, a second curved arrow shows the other electron moving to an adjacent Pd atom. The result is two independent, mobile surface deuterides, Pd–D, each a genuine two-electron Pd–D σ bond, chemically equivalent and interchangeable across the lattice (they can hop between adjacent metal atoms with a very low barrier, which is why "the same face" ends up meaning "the same adsorbed molecule," not "the same literal atom").
Step 3 — chemisorption of (Z)-2-butene. The alkene approaches with its π face parallel to the surface (not edge-on — edge-on would waste the π donation). The filled π(C=C) orbital donates into an empty Pd orbital (arrow from the π bond into the metal), while a filled Pd d-orbital back-donates into the empty π(C=C) orbital (arrow from metal into the π lobe on the far side of each carbon). This is exactly the bonding picture used for any alkene–transition metal π-complex; the practical consequence here is that the C=C bond lengthens and the two alkene carbons are pinned flat against one patch of Pd, with the two methyl groups pointing away from the surface on the same side they occupied in the free (Z)-alkene (chemisorption does not change which substituents are cis to each other — no bond to the substituents is broken).
Step 4 — first D transfer. A mobile surface deuteride migrates across the Pd lattice to one alkene carbon. A curved arrow runs from the Pd–D bond to the carbon, forming a new C–D bond on the face of the alkene that is touching the metal; simultaneously, one electron pair of the former π bond localizes as the new metal–carbon σ bond at the other carbon (arrow from the π system to the metal). The alkene has become a σ-bonded metal alkyl: one carbon now bears D and is sp3, the other carbon is still attached to Pd and is also sp3, but its stereochemistry is frozen — rotation about this new C–C single bond is restricted because free rotation would require partial desorption, which is slower than the second D-transfer step under normal hydrogenation conditions.
Step 5 — second D transfer, same face. A second mobile surface deuteride approaches the remaining Pd–C bond from the metal side (it has no other side to approach from — the substrate is lying on the surface). A curved arrow shows the Pd–D electron pair forming the new C–D bond, while the Pd–C bond electrons return fully to the metal (reductive elimination). Both new C–D bonds are therefore on the same face of the former alkene — the face that was chemisorbed — because neither deuteride could have approached from the opposite, solvent-exposed face without physically passing through the metal lattice.
Step 6 — stereochemical bookkeeping. Starting from (Z)-2-butene (both methyls cis) and adding two D's to the same face gives the meso 2,3-dideuteriobutane (the two new stereocenters are related as an internal mirror plane once D and H are distinguished by isotope labeling, formally); starting from (E)-2-butene gives the pair of enantiomers ((2R,3R)- and (2S,3S)-2,3-dideuteriobutane, i.e., the racemic/(±)-diastereomer relative to the meso product). This isotope-labeling result is the classical experimental proof of syn addition: if hydrogenation proceeded through any mechanism allowing addition to occur from opposite faces (e.g., via free H atoms in solution), the (Z)- and (E)-alkenes would give the same product mixture, and they do not.
Go deeper · Derivation B · 3 minMechanism 2 — Why alkyne semi-hydrogenation with Lindlar's catalyst stops at the cis-alkene
Goal: explain, mechanistically, why a poisoned Pd catalyst (Pd/CaCO3, treated with lead acetate and quinoline — "Lindlar's catalyst") converts an internal alkyne cleanly into the cis-alkene rather than running all the way to the alkane, and why the product is specifically cis rather than trans.
Step 1 — alkyne chemisorption. An internal alkyne, R–C≡C–R′, approaches the poisoned Pd surface. Its filled π orbital (there are two orthogonal π systems in a triple bond) donates into empty Pd d-orbitals, and Pd back-donates into one of the empty π orbitals, exactly as for an alkene, but now there are two π systems available for bonding, and back-donation is spread across both, which makes the alkyne bind more strongly* to unpoisoned Pd than a simple alkene does — a fact that matters below.
Step 2 — poisoning changes the surface, not the electronics of the substrate. Lead(II) acetate deposits Pb2+/Pb(0)-derived species onto a fraction of the Pd surface sites, and quinoline coordinates strongly to remaining exposed Pd atoms through its basic pyridine-type nitrogen lone pair (a dative N→Pd bond, curved arrow from the nitrogen lone pair into an empty Pd orbital). Both poisons physically and electronically block large ensembles of contiguous Pd atoms. The result is that only small, isolated patches of bare Pd remain, too small to bind an alkene flat with the strong multi-atom back-donation that a full alkene chemisorption normally uses.
Step 3 — first H2 addition proceeds exactly as in Mechanism 1. On a small residual Pd patch, H2 dissociates homolytically (same σ-donation/back-donation/cleavage sequence as before), giving two mobile surface hydrides. The alkyne chemisorbs through one of its two orthogonal π systems, flat against the small Pd ensemble. Sequential delivery of the two hydrides — first one H to one alkyne carbon (forming a vinyl–metal σ-intermediate, i.e., an alkenyl group still σ-bonded to Pd), then the second H to the remaining Pd–C(vinyl) bond via reductive elimination — installs both new C–H bonds on the same face of the (former) triple bond, exactly as argued in Steps 4–5 of Mechanism 1.
Step 4 — the geometric consequence: cis-alkene. Because the alkyne was chemisorbed flat, with both R and R′ substituents pushed away from the metal on the same side (the only geometry that allows both carbons to reach the surface at once for a linear-to-bent alkyne rehybridization), and because both new hydrogens arrive from the metal side, the two R groups end up cis to each other on the newly formed, now-planar C=C unit. This is the same syn-delivery logic as Mechanism 1, applied to a triple bond collapsing to a double bond rather than a double bond collapsing to a single bond.
Step 5 — why it stops there: the alkene desorbs before it can re-adsorb flat. The resulting cis-alkene product binds to unpoisoned Pd about 10²–10³ times more weakly than the parent alkyne did (alkynes have two orthogonal π systems available for back-donation; the alkene has only one, and the poisoned surface further restricts flat, multi-atom binding geometries). On the small, poisoned Pd ensembles left after Pb/quinoline treatment, this weaker alkene simply cannot re-adsorb flat long enough to undergo a second round of H2 dissociative addition before it diffuses away into solution. The alkyne, bound more strongly, outcompetes the alkene for the scarce remaining active sites, so the alkene is essentially "kicked off" the surface each time it forms, arresting the reaction cleanly at the cis-alkene stage rather than proceeding to the alkane.
Step 6 — sanity check against the trans alternative. If the mechanism instead involved any step where the vinyl–metal intermediate could isomerize (rotate about the C=C-forming bond) before the second hydrogen arrived, or if hydrogenation proceeded through a diffusing free-radical or carbanion intermediate in solution, trans-alkene (or a mixture) would result. The experimentally clean formation of the cis-alkene is strong mechanistic evidence that the vinyl intermediate is rigidly held on the surface — consistent with Step 3's claim that rotation is blocked while the substrate remains metal-bound.
Worked examples
Example 1 — Foundational: hydrogenation of propene over Pd/C. Propene, CH2=CH–CH3, is stirred under 1 atm H2 with 5% Pd/C in ethanol. Predict the product and describe the fate of each atom mechanistically. Step through the mechanism: H2 approaches the Pd surface and undergoes dissociative chemisorption exactly as in Mechanism 1, Steps 1–2, giving two mobile Pd–H hydrides (this step has no stereochemical consequence here because propene's terminal alkene carbon, CH2=, is not a stereocenter). Propene chemisorbs with its π face flat on Pd (Step 3 analog); because the terminal CH2 end is unhindered, this alkene binds and reacts fast — propene is a highly reactive substrate industrially precisely because of this low steric penalty. One surface hydride delivers to the terminal CH2 carbon, forming a Pd–CH(CH3)–CH3 σ-alkyl intermediate (secondary carbon bound to Pd); the second hydride reductively eliminates onto that carbon, releasing propane, CH3–CH2–CH3. No stereocenter is created (the product has no chiral carbon), so this example isolates the chemoselectivity and rate aspects of the mechanism without stereochemical complication — a useful baseline before adding stereochemistry. Checking a limiting case: if propene concentration were increased tenfold while H2 pressure is held fixed, the Langmuir–Hinshelwood rate law predicts the rate increases but sub-linearly and eventually plateaus as θalkene → 1 (surface saturation), never exceeding the maximum turnover rate set by θH — a real, testable deviation from naive rate ∝ [alkene] kinetics.
Example 2 — Applied: partial hydrogenation of a polyunsaturated fat and the trans-fat problem. Industrial hydrogenation of vegetable oil (rich in cis,*cis-linoleic acid triglycerides) with Ni catalyst at partial H2 pressure is used to raise the melting point of oil into margarine-like solids. Explain, mechanistically, why this process unavoidably generates trans fatty acid isomers even though the elementary hydrogenation step itself is syn-selective. The subtlety is that under H2-limited conditions (partial hydrogenation, done deliberately so that some unsaturation survives to keep the product spreadable rather than rock-hard), Step 4 of Mechanism 1 — formation of the metal–alkyl σ-intermediate — becomes reversible. Once a surface hydride adds to one alkene carbon, forming the Pd/Ni–alkyl species, that species does not always proceed immediately to the second hydride delivery (Step 5); if a second surface hydride is not immediately available (because H2 coverage θH is low under H2-limited industrial conditions), the β-hydrogen on the adjacent carbon can instead be eliminated back off the metal (a microscopic reverse of Step 4, a β-hydride elimination), regenerating a chemisorbed alkene. Crucially, during the time the alkyl intermediate sat on the surface, bond rotation about the C–C bond became kinetically accessible (the alkyl carbon, unlike the original flat-bound alkene, can rotate somewhat before re-forming the alkene), so the alkene that redesorbs can have the opposite geometry from the one that first adsorbed — the originally cis (natural, lower-melting) double bond re-emerges as the higher-melting trans isomer some fraction of the time. This is the accepted mechanistic origin of industrial trans fats: not a violation of syn addition in any single hydrogen-delivery event, but a consequence of the reversibility of the metal–alkyl-forming step when hydrogen is scarce on the catalyst surface. As a limiting case: under H2-rich conditions (excess H2, so θH stays near saturation), the alkyl intermediate is captured by a second hydride so quickly that the reverse β-hydride elimination pathway is negligible, trans-isomer formation drops sharply, and hydrogenation proceeds essentially straight through to the fully saturated (and stereochemically irrelevant) fatty acid — which is exactly why full hydrogenation to hard fat produces far less trans* content than partial hydrogenation.
Example 3 — Exam/competition level: assigning relative configuration from a deuterium-labeling experiment. (E)-3-methyl-2-pentene is treated with D2 over Pd/C. Two new stereocenters are generated at C2 and C3. Using the syn-addition mechanism, determine the relative configuration (i.e., whether the major product is the (2R,3S)/(2S,3R) pair or the (2R,3R)/(2S,3S) pair) of the dideuterated product, 2,3-dideutero-3-methylpentane. First draw (E)-3-methyl-2-pentene explicitly: C2 bears CH3 and H, C3 bears CH2CH3 and CH3, and the E-descriptor means the two higher-priority groups (CH3 on C2 vs. CH2CH3 on C3, since ethyl outranks the C3-methyl) sit on opposite sides of the double bond. Lay the molecule flat as it would chemisorb: with the (E) geometry, the C2–CH3 and C3–ethyl group point to opposite faces in the plane, while C2–H and C3–CH3 also sit trans to each other. Applying Mechanism 1, both new D atoms are delivered from the same metal face — call it the "bottom" face, arbitrarily, since Pd is achiral and both faces are chemisorbed with equal probability, giving a racemic outcome, but the two D's on any individual molecule are always on the same face of each other. Because the two faces of the flat, chemisorbed alkene each add D syn, the outcome is that D on C2 and D on C3 end up cis to each other on the newly formed (initially eclipsed, then freely rotating once desorbed) C2–C3 bond — this is the geometric meaning of syn addition applied to a trisubstituted, unsymmetrical alkene. Careful CIP assignment (tracking that C2's priorities are D > CH3 > CH2CH3-bearing-C3 > H, and C3's priorities are CH2CH3 > CH3 > C2 > D, with the substituent arrangement fixed by the E starting geometry and syn D-delivery) shows that the major diastereomer produced is the (2R,3S) and (2S,3R) pair (a racemic mixture of these two enantiomers, since Pd/C is achiral and offers no facial preference) — the "unlike" (u) descriptor pair rather than the "like" (l) pair — while essentially none of the (2R,3R)/(2S,3S) diastereomer forms from this particular alkene geometry. Verifying the limiting case: for the (Z)-alkene isomer instead (cis geometry), the same syn-delivery logic instead produces the opposite diastereomer, (2R,3R)/(2S,3S), a clean diastereomeric switch driven entirely by which face of the starting alkene the two "trans-like" substituents occupy — exactly the diagnostic behavior that makes catalytic D2 labeling a rigorous stereochemical proof, not merely a plausibility argument.
Limiting cases, scope & common traps
| Situation | Mechanistic behavior | Consequence |
|---|---|---|
| Unhindered, monosubstituted alkene (e.g., 1-hexene) | Binds flat easily, high θalkene | Fast hydrogenation; often used as a rate benchmark |
| Highly substituted alkene (e.g., 2,3-dimethyl-2-butene) | Steric clash prevents flat chemisorption | Slower hydrogenation despite lower ΔH°hydrog (less thermodynamic driving force needed, yet kinetically sluggish) |
| H2-limited (partial hydrogenation) conditions | Metal–alkyl intermediate lifetime increases; β-hydride elimination competes | cis→*trans* isomerization; loss of strict syn stereospecificity at the population level (individual events remain syn, but net alkene population scrambles) |
| Alkyne + Lindlar's catalyst (poisoned Pd) | Small Pd ensembles bind alkyne but not the resulting alkene | Reaction halts after one syn-H2 addition; clean cis-alkene |
| Alkyne + Na/NH3(l) (dissolving metal, not catalytic) | Radical-anion mechanism, not surface syn-addition | Gives trans-alkene — a different mechanism entirely, useful as a contrast case |
| Homogeneous Wilkinson's catalyst, RhCl(PPh3)3 | Discrete organometallic complex, still oxidative addition of H2 + alkene coordination + migratory insertion + reductive elimination | Also syn-selective, but mechanism is molecular, not a surface lattice — same stereochemical outcome from a related but distinct mechanism |
The chemist behind it
Paul Sabatier (1854–1941), a French chemist working at the University of Toulouse, discovered in 1897 (with his student Jean-Baptiste Senderens) that finely divided nickel catalyzes the addition of hydrogen gas to unsaturated organic compounds at moderate temperature. Before Sabatier, hydrogenation of alkenes was known only through much harsher, less general methods; his systematic study of metal-catalyzed hydrogenation — extending it to a wide range of substrates and establishing nickel, and later other transition metals, as general-purpose hydrogenation catalysts — transformed both organic synthesis and industrial chemistry (the Sabatier process for hydrogenating vegetable oils into margarine dates directly from this work). He shared the 1912 Nobel Prize in Chemistry with Victor Grignard for this method of hydrogenating organic compounds in the presence of finely disintegrated metals. Sabatier's own explanation of the mechanism was incomplete by modern standards — the detailed picture of dissociative chemisorption, alkene π-coordination, and stepwise surface hydride transfer was worked out over the following decades by surface scientists and organometallic chemists, including Irving Langmuir (adsorption theory), the discovery and mechanistic study of Wilkinson's catalyst by Geoffrey Wilkinson in the 1960s (which gave chemists a homogeneous, molecularly well-defined analogue of the surface mechanism, allowing each step to be isolated and confirmed spectroscopically), and decades of surface-science studies using techniques like LEED and modern surface-sensitive spectroscopy that directly visualized dissociated hydrogen atoms and chemisorbed alkenes on single-crystal metal surfaces.
Origins · 1 minPaul Sabatier and how this was discovered
Sabatier's central insight was that catalytic activity correlates with the strength of adsorption in a specific, non-monotonic way — too weak and the substrate never reacts, too strong and the product or intermediate never desorbs to free the surface for another cycle. This qualitative principle, later refined and named the Sabatier principle, anticipated by decades the modern quantitative treatment of catalyst "volcano plots" relating binding energy to turnover rate, and it is precisely the logic invoked above to explain why Lindlar-poisoned Pd stops at the alkene stage: the alkyne binds strongly enough to react, the resulting alkene binds too weakly (on the poisoned, reduced-ensemble surface) to react further.
Why does catalytic hydrogenation need a metal catalyst at all — why doesn't H2 just add directly to alkenes?
The H–H bond is strong (about 104 kcal/mol) and, more importantly, H2 has no accessible low-energy pathway to add directly across a C=C π bond in a concerted, uncatalyzed step: the symmetry-allowed pathways for such an addition are thermally forbidden by orbital-symmetry considerations (it would be a formal [2+2] type process), so the uncatalyzed reaction has a prohibitively high activation barrier even though it is thermodynamically very favorable. The metal surface solves this by breaking H2 apart into two separate hydrogen atoms first (dissociative chemisorption), sidestepping the need for any single concerted, symmetry-forbidden step. Each hydrogen atom is then delivered individually in a stepwise, low-barrier process. This is why hydrogenation is a genuinely catalytic phenomenon rather than a simple thermal reaction — the surface changes the mechanism, not the thermodynamics.
Does the syn-addition stereochemistry ever fail, i.e., can hydrogenation give some anti product?
Under standard, H2-rich catalytic conditions, syn addition is extremely reliable, and deuterium-labeling experiments consistently confirm high diastereomeric purity. However, under H2-limited conditions (as in industrial partial hydrogenation), the metal–alkyl intermediate can live long enough to undergo reversible β-hydride elimination and re-adsorb with different alkene geometry, which scrambles the population of product stereochemistry even though each individual hydrogen-delivery event remains syn. So "failure" of syn addition observed experimentally is really a secondary isomerization occurring between two syn-addition events, not a breakdown of the syn mechanism itself.
How does Wilkinson's catalyst relate to the heterogeneous surface mechanism described here?
Wilkinson's catalyst, RhCl(PPh3)3, is a molecular, homogeneous complex, but it recapitulates the same logical sequence: oxidative addition of H2 to Rh(I) is the molecular analogue of dissociative chemisorption (both hydrogens end up bound to the same metal center), alkene coordination is the molecular analogue of chemisorption (donation from alkene π into an empty Rh orbital, back-donation from filled Rh d-orbitals into alkene π*), and migratory insertion (forming a Rh–alkyl σ-bond) is the molecular analogue of the first hydride-transfer step. Because everything happens on one discrete, spectroscopically characterizable metal center rather than a lattice, Wilkinson's catalyst let chemists confirm each mechanistic step directly, which is part of why the heterogeneous surface mechanism described here is considered well established rather than merely inferred.
Why do Pd, Pt, Ni, and Rh work as hydrogenation catalysts but, say, gold does not (under ordinary conditions)?
Effective hydrogenation catalysts need d-orbitals positioned energetically so that (1) they can back-donate into H2's σ orbital strongly enough to cleave it with a low barrier, and (2) they can also π-bond the alkene without binding either substrate so strongly that the surface never regenerates (the Sabatier principle: a "volcano" relationship between binding strength and turnover rate). Gold's filled d-band sits too low in energy (relativistically contracted) to back-donate effectively into H2's σ, so bulk gold dissociates H2 very poorly and is a poor hydrogenation catalyst, although gold nanoparticles under special conditions show some activity — a reminder that particle size and surface structure, not just the bulk element, matter.
Does catalytic hydrogenation reduce aromatic rings the same way it reduces alkenes?
Yes, mechanistically — benzene and other arenes can be hydrogenated over Pd, Pt, or especially Rh and PtO2 (Adams' catalyst) via the same dissociative-H2 and sequential-delivery mechanism — but the aromatic π system's extra resonance stabilization (about 36 kcal/mol relative to a hypothetical non-aromatic cyclohexatriene) raises the activation barrier substantially, so arene hydrogenation typically requires higher H2 pressure and/or higher temperature than simple alkene hydrogenation. Once the first double bond of the ring is reduced, the remaining diene is no longer aromatic and hydrogenates rapidly and completely, which is why partial hydrogenation of benzene rings to cyclohexadienes is essentially never observed as an isolable product.
Why does D2/Pd labeling, rather than just H2/Pd, provide proof of the mechanism?
Ordinary H2 hydrogenation of an alkene gives an alkane with no isotopic distinction between the "old" hydrogens already on the molecule and the "new" ones delivered by the catalyst, so you cannot tell from the product alone which face the new atoms came from. Using D2 instead installs two isotopically distinguishable atoms exactly where the mechanism predicts they should go, and comparing the diastereomeric outcome from (Z)- versus (E)-alkene starting materials (Example 3) becomes a rigorous stereochemical fingerprint of syn addition — a mechanistic test that would be invisible using unlabeled H2.
How does catalyst poisoning in Lindlar's catalyst relate to catalyst poisoning as a general concept, e.g., lead poisoning of catalytic converters?
The underlying physical idea is the same: a poison occupies or blocks active surface sites, reducing the number of metal atoms available for the intended reaction. In automotive catalytic converters, lead (from leaded gasoline) irreversibly and non-selectively coats the entire Pt/Pd/Rh surface, destroying essentially all catalytic activity — an unwanted, total poisoning. In Lindlar's catalyst, the poisoning is deliberately partial and selective: enough Pd remains active to bind and reduce the strongly-adsorbing alkyne, but not enough contiguous surface remains to bind the more weakly adsorbing alkene product, giving controlled, useful chemoselectivity rather than simple catalyst death. Same physical mechanism, opposite practical consequence, depending on degree and selectivity of site-blocking.