Homogeneous Hydrogenation Mechanisms

A styrene molecule drifts through a solution containing a dissolved rhodium complex and hydrogen gas. Within milliseconds it becomes ethylbenzene. There is no palladium black, no autoclave surface, no heterogeneous mystery. The entire transformation happens on a single, well-characterized metal center. This center binds H₂, binds the alkene, shuffles atoms internally, and lets go of product. It is then ready to do it again a million times over. This is homogeneous hydrogenation, and it is the founding case study of modern organometallic catalysis. Every step is a textbook elementary reaction — Oxidative Addition, alkene coordination, Migratory Insertion, Reductive Elimination — assembled into a closed catalytic loop. The subject matters because it marks a turning point. Mechanistic organometallic chemistry stopped being a collection of stoichiometric curiosities. It became an industrial and pharmaceutical technology. Wilkinson's catalyst, RhCl(PPh₃)₃, discovered in 1965, was the first well-understood homogeneous hydrogenation catalyst. It was the proving ground for the entire vocabulary of catalytic cycles. Its chiral descendants were built by replacing achiral phosphines with chelating chiral diphosphines. They deliver enantiomeric excesses above 99% in industrial synthesis. This includes L-DOPA, menthol, and countless chiral pharmaceutical intermediates. The 2001 Nobel Prize in Chemistry (Knowles, Noyori, Sharpless) was awarded substantially for this chemistry. The big picture is that a catalytic cycle is a closed loop of elementary steps. Each step individually obeys rules already established for oxidative addition, insertion, and reductive elimination. The art lies in seeing how substrate binding, steric bulk, and electronic tuning steer the cycle. Together, these factors direct it toward one product, one turnover rate, and — in the asymmetric case — one enantiomer. Rhodium(I), a d⁸ square-planar center, is perfectly poised for this role. It is cheap enough in loading and kinetically labile enough to bind alkene and H₂ reversibly. Yet it remains robust enough to survive many oxidation-state excursions between Rh(I) and Rh(III). This article delivers the full Wilkinson cycle, constructed step by step. It covers the resting-state and Halpern mechanistic studies that determined which pathway is kinetically dominant. It also examines the unsaturated versus dihydride mechanistic dichotomy. From there, it extends to asymmetric hydrogenation via the Halpern "anti-lock-and-key" model. In this model, the minor, less populated diastereomeric alkene complex reacts faster. This minor pathway controls the stereochemical outcome. It closes by pointing toward Hydroformylation (Oxo Process). That process reuses the same oxidative-addition/insertion toolkit, but with CO and H₂ instead of H₂ alone.

Key takeaways

Wilkinson's catalyst, RhCl(PPh₃)₃, hydrogenates unhindered alkenes via a closed Rh(I)/Rh(III) cycle: phosphine loss, H₂ oxidative addition, alkene binding, migratory insertion, reductive elimination.
The catalytic resting state is RhCl(PPh₃)₂(solvent); a third PPh₃ must dissociate to open a coordination site before H₂ or alkene can bind.
Halpern's kinetic studies showed the "unsaturated route" (alkene binds before H₂ oxidative addition) dominates for Wilkinson's catalyst, not the "hydride route."
Steric bulk on the alkene is why Wilkinson's catalyst hydrogenates monosubstituted and disubstituted alkenes but is sluggish with tetrasubstituted ones — a direct steric-clash argument.
Asymmetric hydrogenation replaces PPh₃ with C₂-symmetric chiral diphosphines (DIPAMP, BINAP); the minor alkene diastereomer, bound less favorably, reacts through the faster oxidative addition and sets the product's configuration.
Key idea: a catalytic cycle is a sequence of ordinary organometallic elementary steps closed into a loop, and the minor resting-state species, not the major one, often controls the selectivity of the whole cycle.

The core insight

In Wilkinson-type asymmetric hydrogenation, the catalytically dominant intermediate is the less stable diastereomeric alkene–Rh complex, because it undergoes the turnover-limiting oxidative addition of H₂ far faster than its more populated diastereomer — selectivity is kinetic, not thermodynamic, and is set by relative activation barriers, not relative ground-state populations.

Where this lives today

Industrial asymmetric hydrogenation built directly on this mechanistic foundation is the backbone of the Monsanto L-DOPA process, in which a DIPAMP–Rh catalyst hydrogenates a prochiral enamide to give the Parkinson's drug precursor in high ee at multi-ton scale — historically the first large-scale industrial application of homogeneous asymmetric catalysis. Related Rh- and Ru-BINAP catalysts (Noyori chemistry) hydrogenate β-keto esters and enamides to make menthol, naproxen intermediates, and numerous statins and antibiotic side chains, with catalyst loadings often below 0.01 mol %, a direct consequence of the fast, well-defined turnover cycle analyzed here. Wilkinson-type catalysts and their achiral variants remain standard tools in synthetic laboratories for chemoselective partial hydrogenation, exploiting exactly the steric selectivity between mono-, di-, and trisubstituted alkenes worked through above. Mechanistically, the same oxidative-addition/insertion/reductive-elimination vocabulary extends directly into Hydroformylation (Oxo Process), where a Rh or Co phosphine catalyst binds CO and H₂ alongside an alkene, inserts the alkene into a Rh–H bond exactly as in this cycle, then migratory-inserts CO before a final hydrogenolytic reductive elimination releases an aldehyde — the oxo process is, structurally, the Wilkinson cycle with an added carbonylation step. Understanding the Rh(I)/Rh(III) hydrogenation cycle in full mechanistic detail is therefore the direct prerequisite for understanding industrial hydroformylation, asymmetric transfer hydrogenation (Noyori–Ikariya Ru catalysts), and modern earth-abundant-metal (Fe, Co, Mn) hydrogenation catalysts that attempt to replicate rhodium's mechanistic logic without precious metals.

Ask the tutor: Why does the Curtin–Hammett regime require fast diastereomer interconversion relative to the turnover-limiting step, and what would change in the ee prediction if that assumption broke down?

The physical picture

Wilkinson's catalyst is chlorotris(triphenylphosphine)rhodium(I), RhCl(PPh₃)₃, a square-planar d⁸ complex, isoelectronic in bonding pattern with Vaska's complex and Pd(0)/Pt(0) phosphine catalysts. Square-planar d⁸ Rh(I) is coordinatively unsaturated (16 electrons) and thermodynamically primed to add two more ligands and become octahedral d⁶ Rh(III) (18 electrons) via oxidative addition — exactly the electron-counting logic that makes Oxidative Addition favorable at d⁸ square-planar centers. The catalytic cycle threads together five elementary steps, each independently well precedented: 1. Phosphine dissociation — RhCl(PPh₃)₃ loses one PPh₃ to generate the 14-electron T-shaped fragment RhCl(PPh₃)₂, rapidly trapped by solvent to give the true resting state RhCl(PPh₃)₂(solv). 2. Oxidative addition of H₂ — H–H adds across the vacant coordination sites to give an octahedral Rh(III) cis-dihydride, RhClH₂(PPh₃)₂(solv). 3. Alkene coordination — the alkene displaces solvent, binding side-on to give an 18-electron Rh(III) alkene-dihydride. 4. Migratory insertion — one hydride migrates onto the coordinated alkene, forming a Rh(III) alkyl hydride (16-electron, one open site) — the textbook syn-insertion geometry from Migratory Insertion. 5. Reductive elimination — the alkyl and remaining hydride couple to eject the alkane product and regenerate RhCl(PPh₃)₂(solv), closing the loop. The mechanistic subtlety, resolved by Halpern's kinetic and NMR studies in the late 1970s and early 1980s, is the order of steps 2 and 3. Two limiting sequences are conceivable:

Hydride route: H₂ oxidative addition happens first (on RhCl(PPh₃)₂(solv)), giving the dihydride, and alkene then binds and inserts.
Unsaturated route: alkene binds first (on RhCl(PPh₃)₂(solv), displacing solvent), giving an alkene complex, and H₂ then oxidatively adds to that alkene complex.

Both routes converge on the same alkene-dihydride intermediate before insertion; the branch point is purely about which small molecule — H₂ or alkene — reaches the 14-electron rhodium center first. Kinetics (turnover rate is zero order in alkene concentration at reasonably high [alkene], but shows saturation behavior in some solvents) combined with low-temperature ³¹P and ¹H NMR observation of resting-state species established that for Wilkinson's catalyst in the standard case (terminal, unhindered alkenes, benzene solvent) the unsaturated route dominates: alkene coordinates to RhCl(PPh₃)₂(solv) before H₂ oxidative addition, and it is this oxidative addition step — H₂ adding to the alkene-bound Rh(I) complex — that is turnover-limiting.

Figure

Wilkinson's catalyst: the closed Rh(I)/Rh(III) hydrogenation cycle

The second important physical picture is steric selection. Rhodium's coordination sphere in RhCl(PPh₃)₂ is crowded with two bulky PPh₃ ligands (cone angle ≈ 145°) plus chloride; a tetrasubstituted alkene experiences severe steric clash approaching the metal, so oxidative addition of H₂ to its complex, and insertion of hydride into it, are both slowed dramatically. This is why Wilkinson's catalyst hydrogenates terminal and cis-disubstituted alkenes efficiently but is essentially inert toward tetrasubstituted and often trisubstituted alkenes — a purely steric selectivity rule, useful synthetically for selective partial reduction of polyenes.

Figure

Steric origin of chemoselectivity: mono- vs. tetrasubstituted alkene approach

The core ideas

The catalytic cycle obeys standard steady-state (Michaelis–Menten-like) kinetics once a resting state and a turnover-limiting step are identified. Writing $K_1$ for the phosphine pre-equilibrium, $K_2$ for alkene binding, and $k_3$ for the turnover-limiting oxidative addition of H₂, the observed rate law for the unsaturated route is:

\text{rate} = \frac{k_3 K_2 [\text{Rh}]_{\text{tot}}[\text{alkene}][\mathrm{H_2}]}{1 + K_2[\text{alkene}]}

where: — $[\text{Rh}]_{\text{tot}}$ is total rhodium concentration (catalyst loading), — $K_2$ is the alkene-binding equilibrium constant on RhCl(PPh₃)₂(solv), — $k_3$ is the rate constant for H₂ oxidative addition to the alkene complex, — $[\text{alkene}]$ , $[\mathrm{H_2}]$ are substrate and dihydrogen concentrations, — PPh₃ dissociation ( $K_1$ ) is treated as a fast pre-equilibrium and folded into $[\text{Rh}]_{\text{tot}}$ via the free-PPh₃ concentration (catalysis is inhibited by added PPh₃, consistent with a dissociative pathway).

Reading: at high alkene concentration ( $K_2[\text{alkene}] \gg 1$ ) the rate saturates and becomes first order in H₂ only — the alkene-bound resting state dominates and further alkene cannot accelerate the turnover-limiting oxidative addition step. At low alkene concentration the rate is first order in both alkene and H₂. Added free PPh₃ appears as an inhibition term, $1/(1+K_1'[\text{PPh}_3])$ , multiplying the whole expression, because it competes with alkene/solvent for the open site on the 14-electron rhodium fragment — this inhibition is one of the classic diagnostic experiments for a dissociative, associative-substrate mechanism. The second governing relationship is the electron-counting bookkeeping that makes each step legal:

d^{n}\,(\text{ML}_x, \text{CN}=c) \;\xrightarrow{\text{oxidative addition}}\; d^{\,n-2}\,(\text{ML}_{x+2}, \text{CN}=c+2)

where: — $d^n$ is the metal's formal d-electron count before the step, — CN is coordination number, — oxidative addition of H₂ raises the formal oxidation state by two and the coordination number by two, converting a 16-electron Rh(I) (d⁸, square planar, CN 4) fragment into an 18-electron Rh(III) (d⁶, octahedral, CN 6) center.

Reading: this is the same electron/oxidation-state accounting used throughout Oxidative Addition chemistry; the reverse step, Reductive Elimination, restores Rh(I) d⁸ and drops the coordination number by two, expelling the alkane and regenerating the open coordination sites needed for the next turnover.

Derivation ASteady-state rate law for the unsaturated (alkene-first) Wilkinson cycle

Goal. Derive the rate law in the core-ideas box from the elementary steps, and show explicitly how the alkene-saturation behavior and phosphine inhibition emerge, using asymmetric (non-round) rate and equilibrium constants so every algebraic step is checkable.

Step 1 — Define the species and steps. Let $\mathrm{L} = \text{PPh}_3$ , $\mathrm{S}=$ solvent. The resting state is $\mathrm{RhClL_2S}$ , call it A (14-electron fragment once S is displaced, effectively). Alkene binds reversibly:

\mathrm{A} + \text{alkene} \;\underset{k_{-2}}{\overset{k_2}{\rightleftharpoons}}\; \mathrm{B}

where B = RhCl(L)₂(alkene), the 16-electron alkene complex, with $K_2 = k_2/k_{-2}$ .

Step 2 — Turnover-limiting oxidative addition. H₂ adds irreversibly (under catalytic turnover, forward flux dominates because insertion/reductive elimination downstream continuously drains product):

\mathrm{B} + \mathrm{H_2} \xrightarrow{k_3} \mathrm{C}

where C = RhClH₂L₂(alkene), the 18-electron Rh(III) dihydride-alkene complex. All downstream steps (insertion, reductive elimination) are assumed fast relative to $k_3$ , consistent with Halpern's kinetic isotope and low-temperature NMR evidence that H₂ addition, not insertion or reductive elimination, is turnover-limiting for typical terminal alkenes.

Step 3 — Steady state on B. Under catalytic conditions the concentration of the transient alkene complex B is small and quasi-stationary:

\frac{d[\mathrm{B}]}{dt} = k_2[\mathrm{A}][\text{alkene}] - k_{-2}[\mathrm{B}] - k_3[\mathrm{B}][\mathrm{H_2}] \approx 0

Because in the unsaturated regime the H₂ addition is turnover-limiting (slow) relative to alkene binding/release (fast pre-equilibrium), we may further approximate $k_3[\mathrm{H_2}] \ll k_{-2}$ , so B is essentially at pre-equilibrium with A:

[\mathrm{B}] = K_2[\mathrm{A}][\text{alkene}], \qquad K_2 = k_2/k_{-2}

Step 4 — Mass balance. Total rhodium partitions between free A and bound B (all other intermediates are kinetically insignificant in concentration because they turn over rapidly):

[\mathrm{Rh}]_{\text{tot}} = [\mathrm{A}] + [\mathrm{B}] = [\mathrm{A}]\bigl(1 + K_2[\text{alkene}]\bigr)

\Rightarrow\quad [\mathrm{A}] = \frac{[\mathrm{Rh}]_{\text{tot}}}{1+K_2[\text{alkene}]}

Step 5 — Rate expression. The overall turnover rate equals the flux through the rate-determining step, $k_3[\mathrm{B}][\mathrm{H_2}]$ :

\text{rate} = k_3[\mathrm{H_2}]\cdot K_2[\text{alkene}]\cdot[\mathrm{A}] = \frac{k_3 K_2[\text{alkene}][\mathrm{H_2}][\mathrm{Rh}]_{\text{tot}}}{1+K_2[\text{alkene}]}

which reproduces the rate law quoted above.

Step 6 — Phosphine inhibition. If free PPh₃ (concentration $[\mathrm{L}]$ ) competes for the same open site, replace $[\mathrm{Rh}]_{\text{tot}}$ 's available fraction by including a competitive-binding equilibrium $K_1'$ for L rebinding to the 14-electron fragment; carrying this through the same mass balance multiplies the whole rate by $1/(1+K_1'[\mathrm{L}])$ . Numerically: with $K_2 = 40\ \text{M}^{-1}$ , $k_3 = 0.62\ \text{s}^{-1}\text{M}^{-1}$ , $[\text{alkene}]=0.15\ \text{M}$ , $[\mathrm{H_2}]=4.1\times10^{-3}\ \text{M}$ , $[\mathrm{Rh}]_{\text{tot}}=2.0\times10^{-4}\ \text{M}$ , and no added phosphine:

K_2[\text{alkene}] = 40 \times 0.15 = 6.0

\text{rate} = \frac{(0.62)(6.0)(4.1\times10^{-3})(2.0\times10^{-4})}{1+6.0}\ \text{M s}^{-1}

= \frac{3.05\times10^{-6}}{7.0}\ \text{M s}^{-1} = 4.36\times10^{-7}\ \text{M s}^{-1}

Limiting case: as $[\text{alkene}]\to\infty$ , $K_2[\text{alkene}] \gg 1$ and the rate saturates to $k_3[\mathrm{H_2}][\mathrm{Rh}]_{\text{tot}}$ , independent of alkene concentration — the hallmark of a pre-saturated resting state, confirmed experimentally for Wilkinson's catalyst at high olefin loading.

Go deeper · Derivation B · 2 minHalpern's diastereomer-selection model for asymmetric hydrogenation (Curtin–Hammett kinetic control)

Derivation BHalpern's diastereomer-selection model for asymmetric hydrogenation (Curtin–Hammett kinetic control)

Goal. Show quantitatively why the minor diastereomeric alkene complex, not the major one, can dictate the enantiomeric excess of the hydrogenation product — the "anti lock-and-key" principle central to Rh/chiral-diphosphine asymmetric hydrogenation (e.g., DIPAMP-Rh catalysis of enamide substrates).

Step 1 — Set up the two diastereomeric resting states. A prochiral enamide substrate binds a chiral Rh(I) diphosphine fragment [Rh(chiral-P₂)]⁺ in two diastereomeric ways, forming complexes major (M) and minor (m), related by which face of the alkene coordinates. Because the phosphine is chiral but the substrate binding modes are diastereomeric (not enantiomeric), M and m have different free energies:

\Delta G^{\circ}_{\mathrm{m}\to\mathrm{M}} = G^{\circ}(\mathrm M) - G^{\circ}(\mathrm m) < 0

so M is thermodynamically preferred and dominates the equilibrium population; typical ratios are $[\mathrm M]:[\mathrm m] \approx 10:1$ to $50:1$ , observable directly by low-temperature ³¹P NMR.

Step 2 — Both diastereomers can react, but with different barriers. Each diastereomeric alkene complex undergoes the turnover-limiting oxidative addition of H₂ (as established for the achiral Wilkinson system) with its own rate constant, $k_{\mathrm M}$ and $k_{\mathrm m}$ :

\mathrm M + \mathrm H_2 \xrightarrow{k_{\mathrm M}} \text{(R)-product}, \qquad \mathrm m + \mathrm H_2 \xrightarrow{k_{\mathrm m}} \text{(S)-product}

(the two diastereomeric transition states funnel to opposite product enantiomers because the two alkene faces are enantiotopic on the achiral alkene but diastereotopic once bound to the chiral catalyst).

Step 3 — Curtin–Hammett regime. Because M and m interconvert rapidly (fast alkene dissociation/reassociation) relative to the slow, turnover-limiting H₂ addition, the system is under Curtin–Hammett control: product ratio depends on the relative free energies of the two transition states, not on the ground-state population ratio of M and m. The enantiomeric ratio is:

\frac{[\text{R-product}]}{[\text{S-product}]} = \frac{k_{\mathrm M}[\mathrm M]}{k_{\mathrm m}[\mathrm m]} = \exp\!\left(-\frac{\Delta G^{\ddagger}_{\mathrm M} - \Delta G^{\ddagger}_{\mathrm m}}{RT}\right)

where $\Delta G^{\ddagger}_{\mathrm M}$ , $\Delta G^{\ddagger}_{\mathrm m}$ are the overall free-energy barriers to product measured from a common reference state (so they already include the ground-state energy difference between M and m plus each one's intrinsic activation energy for H₂ addition).

Step 4 — The key experimental discovery. For DIPAMP-Rh/enamide systems, Halpern showed the minor diastereomer m, though only ~1% to 10% populated, reacts with H₂ some 10²–10⁴ times faster than the major diastereomer M, because in M the substrate's ester/amide carbonyl and the alkene are oriented so the incoming H₂ approach is sterically blocked by the chiral phosphine's aryl rings, while in m the same substituents rotate the alkene into an open quadrant. The overall selectivity is therefore set by:

\Delta G^{\ddagger}_{\mathrm M} - \Delta G^{\ddagger}_{\mathrm m} = \Delta G^{\circ}_{\mathrm m \to \mathrm M} + \bigl(\Delta G^{\ddagger}_{\text{intrinsic},\mathrm M} - \Delta G^{\ddagger}_{\text{intrinsic},\mathrm m}\bigr)

and the intrinsic-barrier term is large and negative (favoring m) enough to overcome the unfavorable population term, flipping the overall kinetic preference to the minor complex.

Step 5 — Worked numeric check. Suppose at 298 K the major/minor ground-state ratio is $[\mathrm M]/[\mathrm m] = 17$ (i.e., $\Delta G^{\circ}_{\mathrm m\to \mathrm M} = -RT\ln(17) = -7.0\ \text{kJ mol}^{-1}$ , since $\ln 17 = 2.833$ and $RT = 2.478\ \text{kJ mol}^{-1}$ at 298 K):

\Delta G^{\circ}_{\mathrm m \to \mathrm M} = -(2.478)(2.833)\ \text{kJ mol}^{-1} = -7.02\ \text{kJ mol}^{-1}

and suppose the intrinsic barrier for H₂ addition to m is lower than to M by $18.5\ \text{kJ mol}^{-1}$ (a modest, chemically reasonable steric difference). The net difference in overall barriers, taking m as reference:

\Delta G^{\ddagger}_{\mathrm M} - \Delta G^{\ddagger}_{\mathrm m} = 7.02 + 18.5 = 25.5\ \text{kJ mol}^{-1} \ \text{(favoring m-derived product)}

\frac{[\text{S-product}]}{[\text{R-product}]} = \exp\!\left(\frac{25.5}{2.478}\right) = \exp(10.29) \approx 2.9\times10^{4}

corresponding to an enantiomeric excess of

\text{ee} = \frac{2.9\times10^4 - 1}{2.9\times10^4+1}\times100\% \approx 99.99\%

Limiting case: if the intrinsic barrier difference vanished ( $\Delta G^{\ddagger}_{\text{intrinsic},\mathrm M} = \Delta G^{\ddagger}_{\text{intrinsic},\mathrm m}$ ), the product ratio would simply mirror the ground-state population, $[\mathrm M]/[\mathrm m] = 17$ , giving only ~89% ee and reacting through the major complex — the ordinary "lock and key" expectation. The experimentally observed near-quantitative ee for DIPAMP systems is direct evidence that the anti-lock-and-key, minor-complex-dominant pathway is operative, not naive population-weighted selectivity.

Worked examples

Example 1 — Foundational: electron counts and formal oxidation states around the Wilkinson cycle. Track Rh's d-electron count, formal oxidation state, and total electron count at each labeled stage of the cycle for hydrogenation of ethylene (as the simplest alkene). Start: RhCl(PPh₃)₃. Rh(I), d⁸. Square-planar-ish (T-shaped after dissociation), count as 16 electrons (Rh(I) d⁸ contributes 8, three neutral 2-electron PPh₃ donors give 6, Cl⁻ as X-type gives 2 → 8+6+2 = 16). This 16e resting complex first loses PPh�3 to generate the true 14-electron catalytic entry point:

\text{RhCl(PPh}_3)_3 \ (16e^-) \;\longrightarrow\; \text{RhCl(PPh}_3)_2 \ (14e^-) + \text{PPh}_3

RhCl(PPh₃)₂ picks up solvent (2e L-type donor) to give the 16e resting state RhCl(PPh₃)₂(solv) — Rh(I), d⁸, 16e. Ethylene binds (displacing solvent, alkene as 2e donor via its π system): RhCl(PPh₃)₂(C₂H₄), still Rh(I) d⁸, 16e. H₂ oxidative addition: Rh goes from Rh(I) d⁸ to Rh(III) d⁶; two new Rh–H bonds add 4 electrons total (2 X-type ligands, 2e each) to the count:

16e^- + 2e^- (\text{H}) + 2e^-(\text{H}) = 18e^-

giving RhClH₂(PPh₃)₂(C₂H₄), Rh(III), d⁶, 18e, octahedral. Migratory insertion: one hydride and the coordinated ethylene combine into an ethyl group (an X-type 2e ligand replacing what had been an L-type 2e alkene plus X-type 2e hydride — net the site count contracts by one coordination position since alkene + H become one Rh–Et bond):

18e^- \;\longrightarrow\; 16e^-

giving RhClH(PPh₃)₂(C₂H₅), Rh(III), d⁶, 16e (square pyramidal / one open site). Reductive elimination: Rh–H and Rh–Et couple, releasing ethane and dropping Rh back to Rh(I) d⁸:

16e^- \;\longrightarrow\; \text{RhCl(PPh}_3)_2(\text{solv}) \ (16e^-) + \text{C}_2\text{H}_6

The oxidation state sequence Rh(I) → Rh(I) → Rh(III) → Rh(III) → Rh(I) and the electron count sequence 16 → 16 → 18 → 16 → 16 (with the transient 14e dissociation step) together confirm every step is a legitimate 2-electron elementary process, and the cycle closes exactly where it started.

Example 2 — Applied: predicting relative rates from steric and electronic substrate variation. Wilkinson's catalyst hydrogenates 1-hexene, cyclohexene, and 2-methyl-2-butene at very different rates. Rank them and justify mechanistically. 1-Hexene is monosubstituted (terminal alkene, CH₂=CH–C₄H₉): minimal steric hindrance approaching the crowded RhCl(PPh₃)₂ fragment in the alkene-binding step and in the subsequent hydride insertion; this alkene reacts fastest, with relative rate normalized to 1.00. Cyclohexene is a cis-disubstituted (cyclic, 1,2-disubstituted) alkene: modestly more hindered than a terminal alkene because both alkene carbons carry ring substituents, but the cis geometry keeps the substituents pointed away from the developing Rh–C bonds. Literature relative rates for Wilkinson's catalyst place cyclohexene roughly threefold to fivefold slower than 1-hexene (relative rate ≈ 0.25–0.35), consistent with a steric argument scaled by substitution pattern rather than a change in mechanism. 2-Methyl-2-butene, (CH₃)₂C=CHCH₃, is trisubstituted: three alkyl groups crowd the alkene face that must approach the rhodium center for both binding and insertion. Wilkinson's catalyst reduces this alkene extremely slowly (relative rate often reported below 0.02, i.e., more than fifty-fold slower than 1-hexene), and tetrasubstituted alkenes are for practical purposes unreactive under standard Wilkinson conditions. The ranking 1-hexene ≫ cyclohexene > 2-methyl-2-butene ≈ (unreactive tetrasubstituted alkenes) follows directly from steric clash in the alkene-binding and hydride-migratory-insertion steps around the bulky bis(triphenylphosphine)rhodium fragment — precisely the steric picture in the second figure — and is the practical basis for using Wilkinson's catalyst to selectively hydrogenate the less hindered alkene in a polyunsaturated substrate (e.g., reducing a terminal vinyl group while leaving a trisubstituted alkene untouched, a strategy used routinely in total synthesis).

Example 3 — Exam/competition level: computing ee from a measured rate ratio and checking self-consistency with NMR-observed diastereomer populations. An asymmetric hydrogenation of methyl (Z)-α-acetamidocinnamate with a chiral Rh(diphosphine)⁺ catalyst gives 96.0% ee of the (S)-product. Independent low-temperature ³¹P NMR shows the major/minor diastereomeric alkene-complex ratio is 12.5 : 1, with the major diastereomer known independently (by X-ray/NOE analogy) to lead to the (R)-product upon H₂ addition. Determine (a) which diastereomer is kinetically dominant, (b) the ratio of intrinsic oxidative-addition rate constants $k_{\text{minor}}/k_{\text{major}}$ , and (c) the corresponding difference in intrinsic activation free energies at 298 K. (a) Which diastereomer dominates kinetically? The product is 96.0% ee favoring (S); the major ground-state diastereomer is stated to give (R). Since (S) dominates the product despite the major complex producing (R), the minor diastereomer must be the kinetically dominant pathway — a direct experimental signature of the anti-lock-and-key mechanism. (b) Ratio of overall rate constants. Convert 96.0% ee to a product ratio:

\frac{[S]}{[R]} = \frac{1+0.960}{1-0.960} = \frac{1.960}{0.040} = 49.0

Since $[S]$ arises entirely from the minor complex and $[R]$ entirely from the major complex, and since (under Curtin–Hammett conditions) product ratio equals $(k_{\text{minor}}[\text{minor}])/(k_{\text{major}}[\text{major}])$ :

\frac{k_{\text{minor}}}{k_{\text{major}}} = \frac{[S]/[R]}{[\text{minor}]^{-1}[\text{major}]} = \frac{[S]}{[R]}\times\frac{[\text{major}]}{[\text{minor}]}

With $[\text{major}]/[\text{minor}] = 12.5$ :

\frac{k_{\text{minor}}}{k_{\text{major}}} = 49.0 \times 12.5 = 612.5

So the minor diastereomer's H₂-addition rate constant is about 613-fold larger than the major diastereomer's — a striking but chemically realistic figure for these catalysts (literature values for related DIPAMP systems span roughly $10^2$ – $10^4$ ). (c) Intrinsic activation free-energy difference. Using $\Delta\Delta G^{\ddagger} = -RT\ln(k_{\text{minor}}/k_{\text{major}})$ with $R = 8.314\ \text{J mol}^{-1}\text{K}^{-1}$ , $T=298\ \text{K}$ , $RT = 2477\ \text{J mol}^{-1} = 2.477\ \text{kJ mol}^{-1}$ :

\ln(612.5) = 6.418

\Delta\Delta G^{\ddagger} = -(2.477\ \text{kJ mol}^{-1})(6.418) = -15.90\ \text{kJ mol}^{-1}

The negative sign (by the convention $\Delta\Delta G^{\ddagger} = G^{\ddagger}_{\text{minor}}-G^{\ddagger}_{\text{major}}$ ) confirms the minor diastereomer's transition state lies about 15.9 kJ mol⁻¹ below the major diastereomer's, more than offsetting the $-RT\ln(12.5) = -6.32\ \text{kJ mol}^{-1}$ ground-state preference for the major complex — the same qualitative pattern derived symbolically in the second derivation, now pinned to specific experimental numbers.

Limiting-case check: if the measured ee had been exactly 0% (racemic product), the calculation would require $[S]/[R]=1$ , forcing $k_{\text{minor}}/k_{\text{major}} = 1/12.5 = 0.08$ — i.e., the minor complex would need to be slower, not faster, exactly compensating its lower population, which is the naive lock-and-key limit recovered when no differential steric activation exists between diastereomers.

Limiting cases, scope & common traps

Regime / variant	What changes mechanistically	Consequence
Terminal, unhindered alkenes (1-hexene, styrene)	Unsaturated route dominates; alkene binds before H₂	Fast hydrogenation, first-order in H₂ at saturating alkene
Tri-/tetrasubstituted alkenes	Alkene binding and insertion sterically blocked at bulky RhCl(PPh₃)₂ center	Very slow or no reaction; basis for chemoselective partial hydrogenation
Excess free PPh₃ added	Competes for the open coordination site on the 14e fragment	Rate inhibition — diagnostic of a dissociative mechanism
High H₂ pressure, low alkene concentration	Hydride route (H₂ adds before alkene) can become competitive or dominant	Different (though convergent) resting-state speciation; same product
Chiral diphosphine (DIPAMP, BINAP) replacing PPh₃	Two diastereomeric alkene complexes form; Curtin–Hammett kinetics apply	Enantioselective hydrogenation; ee set by minor-complex kinetics, not populations
Very bulky chiral ligands / hindered enamides	Both diastereomers' oxidative additions slow; population effects can dominate if intrinsic barriers converge	Lower ee, more "lock-and-key"-like (population-controlled) behavior
Crabtree's catalyst (Ir, PF₆⁻ salt, no dissociating phosphine needed)	Cationic Ir(I), different resting state, tolerates tri/tetrasubstituted and unfunctionalized alkenes	Complements Wilkinson's substrate scope; different rate law (no phosphine inhibition)

Common trapTwo persistent misconceptions

First, it is tempting to assume Wilkinson's catalyst hydrogenates alkenes by first oxidatively adding H₂ and only then finding the alkene, by direct analogy to how oxidative addition is usually taught in isolation. Halpern's kinetics show that for the parent Wilkinson system with ordinary alkenes, the alkene binds first (the unsaturated route); H₂ addition to the alkene-bound species is the slow, turnover-limiting step. The "hydride route" is a real, competing pathway (favored at very high H₂ pressure and low alkene concentration) but is not the dominant one under typical synthetic conditions. Second, in asymmetric hydrogenation it is easy to assume the major, most populated (most stable) diastereomeric alkene complex is responsible for the observed product, since that is the "obvious" resting state seen by NMR. The anti-lock-and-key result — that the minor complex reacts orders of magnitude faster and dictates the ee — is a genuine, well-documented exception to naive thermodynamic reasoning and is a textbook demonstration of Curtin–Hammett kinetic control operating inside a catalytic cycle.

The chemist behind it

Origins · 1 minGeoffrey Wilkinson and how this was discovered

Geoffrey Wilkinson

1921-1996 · Organometallic and inorganic chemistry, Imperial College London

Geoffrey Wilkinson shared the 1973 Nobel Prize in Chemistry with Ernst Otto Fischer for pioneering work on the chemistry of organometallic sandwich compounds, having independently proposed the correct sandwich structure of ferrocene in 1952. He subsequently developed RhCl(PPh₃)₃ as a soluble, well-defined homogeneous hydrogenation catalyst, published with R. S. Coffey and coworkers in the mid-1960s, establishing rhodium phosphine chemistry as a template for essentially all subsequent homogeneous catalyst design. His insistence on rigorous, small-molecule stoichiometric characterization of each intermediate — rather than treating catalysis as a black box — set the methodological standard that Halpern and later mechanistic chemists built on.

History

- 1952 — Wilkinson (and independently Fischer) establishes the correct sandwich structure of ferrocene, founding modern organometallic chemistry.

- 1965 — Wilkinson and coworkers report RhCl(PPh₃)₃ as an efficient homogeneous hydrogenation catalyst for alkenes under mild conditions.

- 1968 — Knowles and Horner independently report the first asymmetric hydrogenations using chiral phosphine–Rh catalysts, achieving modest ee.

- 1977 — Halpern's detailed kinetic and NMR studies establish the unsaturated-route mechanism and the turnover-limiting oxidative addition step for Wilkinson's catalyst.

- 1980 — Halpern's kinetic studies on DIPAMP–Rh enamide hydrogenation reveal the anti-lock-and-key, minor-diastereomer-dominant mechanism.

- 2001 — Knowles, Noyori, and Sharpless share the Nobel Prize in Chemistry for catalytic asymmetric synthesis, including asymmetric hydrogenation.

Why is Wilkinson's catalyst considered square planar and 16-electron, given that Rh(I) is d⁸?

Rh(I) is d⁸, and d⁸ metal centers strongly prefer square-planar geometry because that geometry places the single highest-energy $d_{x^2-y^2}$ orbital empty and unoccupied, minimizing ligand–electron repulsion — the same crystal-field argument used for Pt(II), Pd(II), and Ir(I) d⁸ complexes. Four two-electron donor sets around a d⁸ (8-electron) center give a total count of 8 + 4×2 = 16 electrons, one short of the 18-electron closed shell. This deliberate unsaturation is the whole point: it leaves the metal primed to add two more two-electron ligands (such as H–H in oxidative addition) and complete an 18-electron, saturated, octahedral Rh(III) center. Coordinative unsaturation is therefore not a defect but the functional feature that permits catalysis.

Why does the reaction need one PPh₃ to dissociate before catalysis can begin?

The isolated solid RhCl(PPh₃)₃ is already a 16-electron, coordinatively saturated-in-practice square-planar complex once you count Cl, three phosphines, and the d⁸ metal — there is no open site for alkene or H₂ to bind directly without displacing something. Kinetic studies (rate inhibition by added free PPh₃, and the observation of RhCl(PPh₃)₂(solvent) by variable-temperature ³¹P NMR) show that one phosphine ligand reversibly dissociates, opening a site that is immediately filled by solvent, alkene, or H₂. This dissociative pre-equilibrium is fast and is the origin of the phosphine-inhibition term seen in the rate law; it is analogous to ligand dissociation pre-equilibria seen generally before Oxidative Addition at otherwise saturated d⁸ centers.

Is the "unsaturated route" the only mechanism ever operating for Wilkinson's catalyst?

No. Under conditions of very high H₂ pressure and comparatively low alkene concentration, the alternative "hydride route" — H₂ oxidative addition occurring on RhCl(PPh₃)₂(solv) before any alkene is bound — becomes kinetically competitive and can even dominate. Both routes converge on the same alkene-dihydride intermediate before migratory insertion, so the product is identical either way; only the order of small-molecule binding, and hence the exact rate law's dependence on [alkene] and [H₂], differs. Halpern's original mechanistic assignment specifically concerns the dominant pathway under standard synthetic conditions (moderate H₂ pressure, excess alkene, benzene solvent), not an exclusive, universal mechanism.

What physically causes the anti-lock-and-key selectivity in asymmetric hydrogenation?

In the major diastereomeric alkene complex, the substrate's directing group (commonly an amide or ester carbonyl coordinating weakly to rhodium) and the chiral phosphine's aryl rings arrange so that the alkene face is shielded from the incoming H₂ molecule by bulky substituents — the same steric-clash logic that slows hindered alkenes on achiral Wilkinson catalyst, but now applied diastereoselectively. In the minor diastereomer, the same groups rotate into a geometry that leaves an open quadrant for H₂ approach, so oxidative addition proceeds with a much lower barrier. The energetic penalty for populating the minor complex (a few kJ/mol) is more than repaid by a much larger kinetic advantage (tens of kJ/mol) in its subsequent oxidative addition step, which is why the minor pathway wins overall.

How does this mechanism connect to hydroformylation?

Hydroformylation (the oxo process) uses a closely related Rh (or Co) phosphine/phosphite catalyst under synergistic CO/H₂ pressure. The alkene still coordinates to an unsaturated metal fragment and still undergoes migratory insertion into a Rh–H bond, exactly as in Wilkinson-type hydrogenation, but the resulting alkyl intermediate then undergoes an additional migratory insertion of CO (forming an acyl group) before a final oxidative addition of H₂ and reductive elimination release the aldehyde product. The shared vocabulary — oxidative addition, alkene binding, migratory insertion, reductive elimination — is exactly why mastering the Wilkinson cycle is the natural gateway to understanding Hydroformylation (Oxo Process).

Why is rhodium, rather than a cheaper metal, used in Wilkinson's and most asymmetric hydrogenation catalysts?

Rhodium(I) sits in a sweet spot of the periodic table: its d⁸ square-planar complexes are coordinatively unsaturated enough to bind substrates reversibly, yet the Rh(I)/Rh(III) oxidation-state couple has a moderate enough barrier that oxidative addition and reductive elimination both proceed at practical rates near room temperature, unlike lighter 3d metals (where oxidative addition is often thermodynamically unfavorable) or heavier 5d metals like iridium and platinum (where intermediates can become too kinetically inert, slowing turnover, though iridium catalysts such as Crabtree's find their own complementary niche). Cost has driven substantial modern research into iron-, cobalt-, and manganese-based hydrogenation catalysts that try to reproduce rhodium's mechanistic behavior using cheaper, more abundant metals, generally through redesigned ligand frameworks that assist the metal electronically.

Does the alkene insert with a particular regiochemistry, and does that matter for the product?

Yes — migratory insertion is governed by the same steric and electronic rules covered under Migratory Insertion: the metal and hydrogen add across the alkene in a syn fashion, and for unsymmetrical alkenes the regiochemistry (which carbon bonds to Rh, which one gets the new C–H) determines whether a linear or branched alkyl intermediate forms. For simple hydrogenation the two regiochemical outcomes usually converge on the same saturated product after the second C–H bond forms by reductive elimination, so regiochemistry is often mechanistically interesting but synthetically silent; in hydroformylation, by contrast, the analogous regiochemical choice directly controls whether the product aldehyde is linear or branched, which is why regiocontrol becomes a central design problem there.