Metal-Alkene and Metal-Alkyne Complexes

In 1827, William Zeise boiled platinum tetrachloride in ethanol. He isolated bright yellow crystals with a puzzling composition. Their formula was K[PtCl₃(C₂H₄)]·H₂O. This salt defied the bonding theories of the day. An ordinary ethylene molecule was somehow held onto a metal. Remarkably, not a single C–H bond was broken. Chemists could not explain it. Nothing in Kekulé-style valence bonding allows this. A neutral, saturated-looking alkene should not sit at a metal center. The mystery lasted over a century. It was finally solved using X-ray and electron-counting arguments. These combined with a piece of orbital reasoning. That reasoning was proposed independently by M. J. S. Dewar (1951). Joseph Chatt and Leslie Duncanson proposed it too, in 1953. Their model is now called the Dewar–Chatt–Duncanson (DCD) picture. It explains Zeise's salt clearly. By extension, it explains the entire chemistry of alkene and alkyne complexes. The explanation rests on a synergic pair of donor–acceptor interactions. These occur between the π system of the unsaturated ligand and the d orbitals of the metal. This matters because alkene and alkyne complexes are not museum curiosities. They are working intermediates in some of the largest-tonnage industrial processes in chemistry. The Wacker Process converts ethylene to acetaldehyde. It runs through a Pt(II)- and Pd(II)-bound alkene, exactly like the one in Zeise's salt. Hydroformylation passes through such intermediates too. So does olefin polymerization (Ziegler–Natta and metallocene catalysis). Alkene metathesis relies on them as well. Countless C–H activation and cross-coupling cycles do too, at some mechanistic step. Understanding why an alkene binds matters greatly. So does understanding how strongly it binds. Understanding what happens to its geometry and reactivity once bound matters just as much. This is therefore not a side note in organometallic chemistry. It is close to its central pillar. The conceptual heart of the DCD model is synergic bonding. The alkene donates electron density from its filled π orbital. This goes into an empty metal d-orbital (σ-type donation). This donation is forward-facing along the M–ligand axis, viewed in local symmetry. At the same time, the metal donates electron density back. This comes from a filled d-orbital into the alkene's empty π* orbital (π-type back-donation). This dual donor–acceptor loop follows the same electronic logic used elsewhere. It describes Isoelectronic and Related Pi-Acceptor Ligands like CO, NO⁺, and PR₃. Alkenes and alkynes are card-carrying members of the π-acid ligand family. The balance between these two components determines everything. It governs how the alkene's C=C bond lengthens. It governs how substituents pyramidalize away from the metal. It determines whether the ligand is best drawn as an $\eta^2$ -alkene or a metallacyclopropane. It also determines how susceptible the bound alkene becomes to nucleophilic attack. This article builds the DCD model from group-theoretic first principles. Zeise's salt serves as the reference structure. The article derives key structural and spectroscopic consequences. These include C=C bond lengthening and rehybridization/bending-back of substituents. They also include the trans influence and the shift of the C=C stretch. The article works through three fully quantitative examples. These span a textbook rotational-barrier problem, an applied catalytic question, and a bonding-limit case at the alkyne end of the family. It assumes familiarity with the general framework of Hapticity and the Eta Notation. It also sets up the language needed for Allyl and Polyene Complexes, Migratory Insertion, and the Wacker cycle.

Key takeaways

Zeise's salt, K[PtCl₃(C₂H₄)]·H₂O, is the historical and pedagogical archetype of a metal–alkene complex, solved structurally in 1975.
The Dewar–Chatt–Duncanson model splits the M–alkene bond into two synergic components: σ-donation (alkene π → metal dσ) and π-back-donation (metal dπ → alkene π*).
Back-donation populates the alkene π* orbital, which lengthens C=C (from ~1.34 Å free to ~1.37–1.43 Å bound) and bends substituents away from the metal (rehybridization toward sp³).
The same synergic logic extended to alkynes gives two orthogonal π* acceptor orbitals, enabling four-electron donation and a shift toward a metallacyclopropene/metallacyclopropane continuum.
Ethylene in Zeise's salt rotates about the Pt–C₂H₄ axis with a measurable barrier (~40–50 kJ/mol range order), directly diagnosing the π-component's directionality.
Strong back-donation (electron-rich metal, good π-acceptor alkene) pushes the bonding toward a metallacyclopropane; weak back-donation (electron-poor metal, alkene poor π-acceptor) leaves it as a weakly perturbed π-complex, and the two extremes are ends of one continuous bonding spectrum.
Key idea: an alkene or alkyne bound to a metal is not a static ligand but a two-way electronic conduit — σ-donation and π-back-donation act in synergic feedback, and the position of the C=C(C≡C)/metallacycle continuum is read directly from bond lengths, bend-back angles, and IR shifts.

The core insight

An $\eta^2$ -bound alkene or alkyne is held by a synergic pair of donor–acceptor interactions (π→dσ donation and dπ→π* back-donation) whose relative weight — tunable by metal oxidation state and ligand electronics — continuously interpolates the ligand between a nearly-unperturbed π-complex and a metallacyclopropane(-propene).

Where this lives today

The DCD model is not a historical footnote — it is the working conceptual tool behind several of the largest processes in the chemical industry and in modern methodology. The Wacker Process, which converts ethylene and oxygen (via a Pd(II)/Cu(II) redox cycle) into acetaldehyde on a multi-million-tonne annual scale, depends entirely on the electrophilic activation of a Pd(II)-bound alkene exactly as quantified by the back-donation/π* population argument above; the same electronic logic underlies modern Pd(II)-catalyzed Wacker-type oxidative cyclizations used throughout total synthesis to build oxygen and nitrogen heterocycles from alkenyl alcohols and amines. Ziegler–Natta and single-site metallocene polymerization catalysis for polyethylene and polypropylene production begin every monomer insertion step with an $\eta^2$ -alkene bound to a cationic, electron-poor, early-transition-metal (Ti, Zr) center — a regime deliberately tuned toward weak back-donation so that the alkene remains a reactive, loosely-bound insertion substrate rather than settling into an unreactive metallacyclopropane, feeding directly into Migratory Insertion chemistry. Grubbs- and Schrock-type olefin metathesis catalysts likewise proceed through an alkene bound side-on to a metal-carbene center, where the same DCD orbital match determines the geometry of the [2+2] metallacyclobutane-forming step. On the alkyne side, $\eta^2$ -alkyne complexes of cobalt, tungsten, and molybdenum are stoichiometric reagents in alkyne-protection and Pauson–Khand cyclization chemistry, exploiting exactly the enhanced back-donation and metallacyclopropene character available to alkynes but not alkenes. Even outside classical catalysis, the DCD bonding picture underlies the design of alkene- and alkyne-tagged bioorthogonal reagents and the interpretation of vibrational (IR/Raman) spectroscopic probes used to monitor metal-bound unsaturated substrates in operando catalytic studies, and it sets up the vocabulary — filled/empty orbital symmetry matching between a metal fragment and a π-system — needed to understand more elaborate hapticities in Allyl and Polyene Complexes.

:::quiz[Using the neutral (covalent) electron-counting convention, what is the total valence electron count of Zeise's anion, [PtCl₃(η²-C₂H₄)]⁻?]

14 electrons
16 electrons ✓
18 electrons
20 electrons

Ask the tutor: Why does replacing one chloride in Zeise's anion with a stronger σ-donor (like a phosphine) change the ethylene's rotational barrier and C=C bond length?

The physical picture

Consider ethylene approaching a metal center. Free ethylene has $D_{2h}$ symmetry with a filled π orbital (a $b_{1u}$ -type combination of the two carbon $2p_z$ orbitals, symmetric bonding combination) lying in the molecular plane's perpendicular direction, and an empty π orbital (antibonding combination) of the same symmetry type but higher energy. When ethylene coordinates side-on to a metal — the defining feature of $\eta^2$ coordination, as opposed to any $\eta^1$ mode through a single carbon — the midpoint of the C=C bond sits over the metal, and the whole ligand behaves as a single two-electron donor occupying one coordination site. Two symmetry-matched interactions become possible once we set up the local coordinate frame with the M–(midpoint of C=C) vector as the z-axis and the alkene lying along, say, the x-axis in the xz-plane: 1. σ-donation. The filled π orbital of the alkene has the correct nodal pattern to overlap end-on with an empty metal orbital of appropriate symmetry — typically an sd or $d_{z^2}$ -type hybrid pointing at the ligand. This donates electron density from ligand to metal, exactly analogous to the σ lone pair of CO donating through carbon. 2. π-back-donation. A filled metal $d_{xz}$ (or $d_{yz}$ , depending on the exact orientation convention) orbital has the correct symmetry to overlap side-on with the empty π orbital of the alkene — the orbital with a node running through the midpoint of the C–C bond perpendicular to the local ligand plane. This donates density from metal to ligand, populating an antibonding orbital of the alkene. Because π is C–C antibonding, any back-donation into it weakens and lengthens the C=C bond, shifting bond order from ~2 toward ~1.5 or lower depending on the degree of back-donation. Because π also has substituent-centered lobes bent away from the metal-facing side of the alkene, populating it favors pyramidalization of the substituents (a rehybridization from sp² toward sp³) — the well-known "bending back" seen crystallographically in almost every characterized alkene complex. The synergic character is the crucial physical point: σ-donation alone would make the metal too electron-rich and the ligand too electron-poor for the interaction to be stable or for further donation to occur; π-back-donation alone is symmetry-forbidden without the σ-donation first polarizing the system. Each component reinforces the other — σ-donation makes the metal a better π-donor by raising its d-orbital energies (increased electron density raises orbital energies via increased electron-electron repulsion, making dπ→π donation more favorable), and back-donation makes the alkene a better σ-donor by depopulating π and stabilizing the whole adduct. This is the same synergic logic that governs CO, PR₃, and NO⁺ bonding as π-acid ligands, discussed as a general class under Isoelectronic and Related Pi-Acceptor Ligands.

Figure

The Dewar–Chatt–Duncanson synergic bonding scheme for an η²-alkene

Zeise's salt is the model system because it displays every one of these features cleanly. Neutron and X-ray diffraction (refined most precisely in the 1970s) show the ethylene ligand oriented perpendicular to the PtCl₃ plane, exactly as predicted: this orientation maximizes overlap of the alkene π with the filled Pt $d_{xz}$ orbital that lies perpendicular to the square plane, while the alkene π donates into the empty in-plane hybrid trans to the strong-field chloride. The C=C bond lengthens from 1.337 Å (free ethylene) to 1.375(4) Å in the salt — a modest but real elongation consistent with partial (not complete) population of π. The hydrogens bend back away from the metal by a small but measurable angle. And the ethylene ligand shows a genuine, spectroscopically and computationally accessible barrier to rotation about the Pt–(midpoint) axis, because rotating the alkene by 90° breaks the π-back-bonding overlap (the σ-donation survives rotation to first order, since it only needs end-on overlap of the π cloud, roughly cylindrically available, but π-back-donation needs specific in-plane alignment). The extension to alkynes follows immediately from symmetry: an alkyne has two mutually perpendicular π systems (and two mutually perpendicular π systems), by virtue of its cylindrical, $D_{\infty h}$ electronic structure. This means an $\eta^2$ -alkyne can, in principle, donate through both filled π orbitals and accept back-donation into both empty π orbitals — a total of up to four electron pairs' worth of donor-acceptor character (one σ-donor π, one π-acceptor π in the "in-plane" set, and a second, weaker σ-donor/π-acceptor pair in the perpendicular set). In practice, most alkyne complexes use the alkyne predominantly as a two-electron (occasionally four-electron) donor, but the extra π acceptor orbital available to alkynes relative to alkenes is precisely why alkyne complexes bend back their substituents more than analogous alkenes, and why alkynes can act as unusually good bridging ligands across two metals (using one π/π* pair for each metal).

Figure

Alkyne vs alkene: the extra orthogonal π* acceptor orbital

The core ideas

The DCD picture is qualitative orbital reasoning, but it generates quantitative, testable structural predictions. The two central relationships are the bond-length/bond-order correlation and the metallacyclopropane limiting description.

d(\mathrm{C{=}C})_{\text{bound}} = d(\mathrm{C{=}C})_{\text{free}} + \Delta_{\pi^*}

where:

$d(\mathrm{C{=}C})_{\text{free}} = 1.337\ \text{Å}$ is the reference free-ethylene bond length,
$\Delta_{\pi^*}$ is a positive increment (typically 0.02–0.10 Å) that scales monotonically with the fractional population of the ligand π* orbital delivered by back-donation,
$d(\mathrm{C{=}C})_{\text{bound}}$ is the experimentally observed, crystallographically refined bond length in the complex.

Reading: the C=C distance is a direct, model-independent readout of how much backbonding has occurred. A value close to 1.34 Å signals a "pure π-complex" (little back-donation, dominant σ-donation, alkene barely perturbed); a value approaching 1.46–1.50 Å (typical C–C single bond territory, ~1.54 Å being the full single-bond limit) signals that the complex is better described as a metallacyclopropane — two M–C σ bonds to a genuinely sp³-rehybridized ethylene-derived fragment, essentially an oxidative addition product. The second governing relationship formalizes the electron-counting consequence of this continuum:

n(\mathrm{M})_{\text{formal}} = n(\mathrm{M})_{\pi\text{-complex}} \;\; \text{or} \;\; n(\mathrm{M})_{\pi\text{-complex}} + 2,\ \text{depending on the covalent vs. ionic counting convention chosen for the alkene as a 2-electron donor vs. metallacyclopropane as two X-type ligands}

where:

in the neutral (covalent) counting method, an $\eta^2$ -alkene is always treated as a neutral 2-electron L-type donor, regardless of how far the bonding has slid toward the metallacyclopropane limit — this is the practical convention used for oxidation-state and d-electron-count bookkeeping,
in the ionic counting method, the same ligand can alternatively be treated (at the metallacyclopropane extreme) as a dianionic bis-alkyl-type ligand contributing 2 electrons and 2 negative charges, which is chemically more honest at that limit but requires reassigning the metal's formal oxidation state upward by 2,
both conventions must give the same total valence electron count for the complex; they differ only in how the electrons are partitioned between "ligand" and "metal charge."

Reading: this equation is really a caution wrapped as a formula — the d-electron count of a metal bearing an $\eta^2$ -alkene depends on which convention you fix, and the two conventions differ by exactly 2 electrons/2 oxidation-state units, never by an arbitrary amount. Consistency (always using the neutral method, as is standard in most textbooks) is what matters, not which convention is "correct."

Derivation ASymmetry-matching the alkene π/π* set to metal d-orbitals in a square-planar ML₃(alkene) fragment

Goal. Show which specific metal d-orbitals are symmetry-allowed to donate to and accept from the alkene π system in a Zeise's-salt-type square-planar Pt(II) fragment, and thereby justify why the ethylene lies perpendicular to the PtCl₃ plane rather than in it.

Step 1 — set the local axis frame. Place Pt at the origin with the three chlorides in the xy-plane (the "PtCl₃" plane), two chlorides along ±x and one along +y (mirroring the actual $C_s$ local symmetry of the anion). The ethylene occupies the fourth square-planar site along −y, with its C=C bond oriented along z (perpendicular to the PtCl₃ plane) — this is the experimentally observed geometry to be justified.

Step 2 — identify the empty σ-acceptor orbital on Pt. In a d⁸ square-planar ML₄ fragment, the metal-based orbitals split (in an idealized $D_{4h}$ parent before lowering to $C_{2v}/C_s$ for the mixed-ligand set) into a high-lying empty $d_{x^2-y^2}$ (σ*-antibonding to all four in-plane donors) and a comparatively low-lying, largely non-bonding, empty (or partially occupied, hybridized with $s/p_z$ ) orbital directed along the fourth coordination site. For Zeise's anion, the vacant orbital available for σ-acceptance from the alkene π cloud is this in-plane-directed hybrid, pointing along −y toward the ethylene's midpoint. The alkene's filled π orbital, lying in the local xz-plane of the alkene fragment (perpendicular to the C=C axis, which is along z), has a lobe pointing directly along −y toward Pt — good head-on overlap regardless of whether the ethylene C=C axis lies in-plane or perpendicular, because σ-donation only requires the π cloud (not the C=C axis orientation) to point at the metal.

Step 3 — identify the filled π-donor orbital on Pt. The square-planar Pt(II) fragment has filled $d_{xz}$ and $d_{yz}$ orbitals (both are π-type with respect to the four-coordinate plane, i.e., perpendicular to it). Only $d_{xz}$ has the correct nodal alignment to overlap with the alkene's π orbital when the ethylene C=C bond is oriented along z: $d_{xz}$ has lobes above and below the xy-plane, in phase with the two lobes of alkene π that sit above and below the C=C axis when that axis points along z. If instead the ethylene were rotated 90° so C=C lay along x (in the PtCl₃ plane), the relevant π lobes would need to overlap with $d_{yz}$ or with an in-plane, symmetry-mismatched combination — overlap becomes far weaker because the ethylene-centered π nodal surface would then lie almost parallel to, rather than piercing through, the best-aligned filled metal d-orbital.

Step 4 — compare the two orientations energetically. With ethylene perpendicular to the PtCl₃ plane (C=C along z), $d_{xz}$ –π* overlap is maximized (large positive overlap integral), delivering strong back-donation stabilization. With ethylene rotated into the plane, the equivalent overlap is reduced roughly to zero by symmetry in the idealized limit (the surviving overlap, if any, is with a different, higher-lying, more weakly filled orbital). The energy difference between these two rotamers is the rotational barrier of the bound ethylene about the Pt–ligand axis, and it is entirely a π-back-donation effect: the σ-component barely discriminates between the two orientations because the π-donor orbital of ethylene, viewed end-on, presents a roughly cylindrical face to the metal.

Step 5 — conclude the observed geometry. Because the perpendicular orientation is stabilized by back-donation and the in-plane orientation is not, the ground-state structure has ethylene perpendicular to the PtCl₃ plane — exactly the crystallographically determined geometry of Zeise's anion. The rotational barrier calculated computationally and estimated experimentally (variable-temperature NMR line-broadening studies place it in the range of several tens of kJ/mol, consistent with the loss of a full π-back-bonding interaction upon rotation) is the direct energetic fingerprint of the π-component of the DCD model — a component that would not exist, and would generate zero barrier, in a purely σ-donor, ionic, or electrostatic bonding picture.

Go deeper · Derivation B · 2 minBond-length/bend-back relationship as a rehybridization argument

Derivation BBond-length/bend-back relationship as a rehybridization argument

Goal. Derive, from simple valence-bond/hybridization reasoning, why populating alkene π* necessarily both (a) lengthens C=C and (b) forces the substituents to bend away from the metal, and connect the two observables quantitatively through a single "back-donation extent" parameter.

Step 1 — describe free ethylene's π nodal structure. In free ethylene ( $D_{2h}$ ), the π MO is the out-of-phase combination of the two carbon $2p_z$ orbitals, $\pi^* = c_1 2p_z(\mathrm{C1}) - c_2 2p_z(\mathrm{C2})$ with $c_1 = c_2$ by symmetry. This orbital is strictly C=C antibonding: putting electron density into it directly reduces C–C bond order below 2.

Step 2 — connect π population to bond order via a linear interpolation. Treat the bound alkene's electronic structure as a two-configuration mixture: a fraction $(1-x)$ resembling the free-alkene configuration (bond order 2, π empty) and a fraction $x$ resembling the fully back-bonded, metallacyclopropane-like configuration (bond order 1, π* fully populated, both carbons rehybridized toward sp³). Then

$\text{(C–C bond order)} = 2 - x$

where $x \in [0,1]$ is the fractional degree of back-donation (π* population per available electron pair, normalized to run from 0 at "pure π-complex" to 1 at "pure metallacyclopropane").

Step 3 — convert bond order to bond length using a Pauling-type relationship. Using a simple linear (first-order) bond-order/bond-length interpolation between the known single-bond (1.54 Å, sp³ C–C) and double-bond (1.337 Å, sp² C=C) references,

$d(\mathrm{C{-}C}) = 1.54\ \text{Å} - x\,(1.54 - 1.337)\ \text{Å} = 1.54\ \text{Å} - x(0.203\ \text{Å})$

so increasing $x$ (more back-donation) smoothly increases the bond length from 1.337 Å toward 1.54 Å.

Step 4 — link the same parameter $x$ to substituent pyramidalization. As π* is populated, each carbon's local hybridization must reorganize: an sp²-hybridized alkene carbon has all three σ-substituents (H, H, and the other C) coplanar, but an sp³-hybridized alkyl carbon has them arranged tetrahedrally. Since the metal occupies the position that would be the "fourth substituent" direction on each carbon in the sp³ limit, increasing $x$ smoothly rotates the H–C–H (or R–C–R) substituent plane away from the alkene plane and back and away from the metal, by an angle $\theta$ that — by the same linear interpolation logic — runs from $\theta = 0°$ (fully planar, sp², $x=0$ ) to the full tetrahedral bend-back angle $\theta_{\max}$ (roughly 35–45° from the original sp² plane, corresponding to the true sp³ geometry, at $x=1$ ):

$\theta \approx x\,\theta_{\max}$

Step 5 — state the combined structural fingerprint. Because both the bond-length increment and the bend-back angle are driven by the same underlying parameter $x$ , they must co-vary: any complex showing a strongly lengthened C=C (approaching 1.45–1.50 Å) must also show a large bend-back angle, and any complex with only slightly lengthened C=C (1.36–1.38 Å, as in Zeise's salt) must show only a small bend-back angle. This correlated pair of observables — not either one in isolation — is the empirical signature chemists use to place a given alkene complex on the π-complex/metallacyclopropane continuum, and it is precisely what is seen experimentally: Zeise's salt (weak backbonding, $d^8$ Pt(II), good competing π-acceptor chlorides already draining metal density) shows a small increment (1.375 Å, bend-back only a few degrees), while electron-rich, low-oxidation-state, strongly π-basic fragments such as Ni(0) or Pt(0) alkene complexes (e.g., Pt(PPh₃)₂(alkene) adducts) show C=C distances stretched to 1.43–1.46 Å and correspondingly larger bend-back angles, tipping firmly toward the metallacyclopropane description.

Worked examples

Example 1 (foundational): assigning the oxidation state, d-count, and geometry of Zeise's anion. Zeise's anion is $[\mathrm{PtCl_3(\eta^2\text{-}C_2H_4)}]^-$ . Using the standard neutral (covalent) counting convention in which $\eta^2$ -ethylene is always a neutral 2-electron L-type donor and each chloride is a 1-electron X-type (covalent) donor or, equivalently, treat chlorides ionically as X-type anionic donors and assign formal charges — we work through both consistently. Ionic method (formal oxidation states). Each Cl⁻ is assigned as an anionic ligand (−1 charge, 2-electron donor in the ionic scheme), and ethylene is neutral (0 charge, 2-electron donor). The overall complex charge is −1. So:

(\text{Pt oxidation state}) + 3(-1) + 0 = -1 \implies (\text{Pt oxidation state}) = +2

Pt(II) is a $d^8$ configuration (Pt, Z=78, group 10, neutral atom configuration $[Xe]4f^{14}5d^9 6s^1$ ; removing 2 electrons for the +2 ion gives $d^8$ , the two 6s electrons and one further electron are removed first per the standard convention that s-electrons ionize before d for transition metals, giving $5d^8$ ). Electron count (total valence electron count, TVEC), ionic method:

\text{TVEC} = d^8 + 3(2\ e^-,\ \mathrm{Cl^-}) + 2\ e^-\,(\mathrm{C_2H_4}) + 1\ e^-\,(\text{charge on anion, distributed to metal})

Carrying this out numerically (each anionic Cl⁻ donates 2 electrons in the ionic/donor-pair convention, matching the covalent method's per-bond electron count once charges are reconciled):

\text{TVEC} = 8 + (3\times 2) + 2 = 8 + 6 + 2 = 16\ \text{electrons}

This is the textbook 16-electron count expected for a square-planar $d^8$ Pt(II) complex — square-planar, 16-electron, coordinatively unsaturated (one open axial-type site above and below the plane, though for $d^8$ square-planar this "unsaturation" is a stable, closed-shell, non-fluxional resting state, not a reactive vacancy in the way a 16-electron count would be destabilizing for an octahedral d⁶ system). The geometry is square planar with ethylene occupying one coordination site as a single ligand point (its midpoint, not each carbon separately, occupies the "slot"), oriented perpendicular to the PtCl₃ plane as justified by the symmetry argument above.

Limiting-case check: if the ethylene in Zeise's anion is replaced by a fourth chloride (giving the well-known $[\mathrm{PtCl_4}]^{2-}$ anion), the electron count recomputes as $8 + 4(2) = 16$ electrons also — consistent, since replacing a 2-electron π-acid donor with another 2-electron halide donor at the same site does not change the total count, only the charge (which shifts from −1 to −2 because chloride is anionic while ethylene is neutral). This checks that the counting machinery is self-consistent across ligand substitution at fixed geometry.

Example 2 (applied): estimating the shift in the C=C stretching frequency upon coordination, and connecting it to the Wacker mechanism's electrophilic alkene. Free ethylene's C=C stretch appears at $\tilde\nu = 1623\ \text{cm}^{-1}$ (IR/Raman, gas phase). In Zeise's salt, the corresponding band is shifted down to approximately $\tilde\nu \approx 1516\ \text{cm}^{-1}$ (this red shift, roughly 100 cm⁻¹, is a long-established spectroscopic hallmark of alkene coordination via back-donation into π*, entirely parallel to the CO stretch red shift used to gauge back-donation to carbonyls). Model the C=C stretch as a simple diatomic-oscillator surrogate (justified because the C=C stretching mode is, to good approximation, localized on the two carbons) with force constant $k$ related to frequency by

\tilde\nu = \frac{1}{2\pi c}\sqrt{\frac{k}{\mu}}

where $\mu$ is the reduced mass of the two carbon atoms (treating the substituents as rigidly attached, a standard simplifying approximation for this qualitative comparison), $c$ is the speed of light, and $k$ is the effective C=C stretching force constant. Using $\mu = \dfrac{m_C \cdot m_C}{m_C + m_C} = \dfrac{m_C}{2}$ with $m_C = 12.00\ \text{u} = 1.993\times10^{-26}\ \text{kg}$ , so $\mu = 9.963\times10^{-27}\ \text{kg}$ :

k_{\text{free}} = \mu\,(2\pi c\,\tilde\nu_{\text{free}})^2

Plugging in $\tilde\nu_{\text{free}} = 1623\ \text{cm}^{-1} = 1.623\times10^{5}\ \text{m}^{-1}$ and $c = 2.998\times10^{10}\ \text{cm/s}$ :

2\pi c\,\tilde\nu_{\text{free}} = 2\pi (2.998\times10^{10}\ \text{cm/s})(1623\ \text{cm}^{-1}) = 3.058\times10^{14}\ \text{s}^{-1}

k_{\text{free}} = (9.963\times10^{-27}\ \text{kg})(3.058\times10^{14}\ \text{s}^{-1})^2 = 9.31\times10^{2}\ \text{N/m}

Repeating for the bound complex with $\tilde\nu_{\text{bound}} = 1516\ \text{cm}^{-1}$ :

2\pi c\,\tilde\nu_{\text{bound}} = 2\pi (2.998\times10^{10})(1516) = 2.856\times10^{14}\ \text{s}^{-1}

k_{\text{bound}} = (9.963\times10^{-27})(2.856\times10^{14})^2 = 8.13\times10^{2}\ \text{N/m}

So the effective C=C force constant drops from about 931 N/m to 813 N/m upon coordination — a drop of roughly 13%, i.e.

\frac{\Delta k}{k_{\text{free}}} = \frac{931-813}{931} = 0.127

Chemical reading, connecting to Wacker chemistry. This 13% softening of the C=C bond is the direct vibrational signature of π* population by back-donation — the same phenomenon that lengthens the bond crystallographically. The practical payoff is mechanistic: because back-donation has already begun draining electron density out of the alkene's π system and into the metal (and, in a d⁸ Pt(II)/Pd(II) system, competing with strong-field ancillary ligands for that metal d-density), the alkene carbons become simultaneously activated toward external nucleophilic attack. This is the electronic basis for the first C–C/C–O bond-forming step of the Wacker Process: a coordinated, electronically activated alkene in a Pd(II) (or, in Zeise's-salt-type model chemistry, Pt(II)) square-planar complex is attacked by an external nucleophile (water, in the industrial Wacker cycle) from the face opposite the metal, because the alkene carbons have become partially electrophilic exactly to the degree quantified by $x$ and by the force-constant softening computed here.

Limiting-case check: if back-donation were completely absent ( $x \to 0$ , a purely electrostatic or dative σ-only complex), the C=C stretch would remain at 1623 cm⁻¹ (no softening), the alkene would remain a poor electrophile, and nucleophilic attack chemistry of the Wacker type would not be kinetically accessible — consistent with the observation that alkene complexes of hard, non-π-basic metal centers (e.g., many lanthanide or early-transition-metal alkoxide/alkene adducts, where back-donation is minimal because the relevant d-orbitals are empty or too high in energy) are poor substrates for this chemistry.

Example 3 (exam/competition level): comparing η²-alkyne four-electron donation to η²-alkene two-electron donation in a d-electron-count and bond-order problem. Consider the hypothetical (but electronically well-precedented) complex $[\mathrm{W(CO)_5(\eta^2\text{-}PhC{\equiv}CPh)}]$ , where diphenylacetylene is bound end-on-perpendicular as a four-electron donor (using both filled π orbitals to donate, a mode well documented for alkynes bound to electron-poor, coordinatively unsaturated metal fragments, in contrast to the more common two-electron alkyne-donor mode). Part (a): electron count with the alkyne as a 2-electron donor. Tungsten, group 6, in the neutral covalent counting method contributes 6 electrons as a neutral d⁶ metal atom (M(0) formalism, standard for a homoleptic-plus-CO carbonyl-type fragment). Five CO ligands each donate 2 electrons; the alkyne as an ordinary 2-electron donor contributes 2 electrons:

\text{TVEC}_{2e} = 6 + 5(2) + 2 = 6 + 10 + 2 = 18\ \text{electrons}

This is a saturated, 18-electron, coordinatively stable complex — consistent with a simple substitution product of $\mathrm{W(CO)_6}$ in which one CO is formally replaced by an alkyne acting as an ordinary two-electron ligand (charge-neutral, count-neutral swap, since both are 2-electron L-type donors). Part (b): electron count if the alkyne instead donates four electrons (both π systems). Now the alkyne contributes 4 electrons rather than 2:

\text{TVEC}_{4e} = 6 + 5(2) + 4 = 6 + 10 + 4 = 20\ \text{electrons}

A 20-electron count exceeds the 18-electron closed-shell limit for a mononuclear octahedral-type d-block complex — this configuration is not accessible without first removing a two-electron donor elsewhere in the coordination sphere (in practice, four-electron alkyne donation is observed almost exclusively in complexes that are otherwise electronically unsaturated at the 16-electron level before the alkyne binds, e.g. $\mathrm{d^4}$ Mo(0)/W(0) or early-metal $d^0$ – $d^2$ fragments with fewer ancillary ligands, so that the alkyne's second π-donor interaction brings the total up to, not beyond, 18 electrons). Part (c): identify a real system consistent with four-electron alkyne donation. Strip one CO from the hypothetical fragment to make it electronically consistent: $[\mathrm{W(CO)_4(\eta^2\text{-}PhC{\equiv}CPh)}]$ with the alkyne as a 4-electron donor:

\text{TVEC} = 6 + 4(2) + 4 = 6 + 8 + 4 = 18\ \text{electrons}

This now recovers an 18-electron closed-shell configuration, matching the well-documented class of $\mathrm{W(CO)_4(RC{\equiv}CR)}$ -type complexes (and related $\mathrm{Mo}$ analogues) in which crystallography shows a short W–C(alkyne) distance, a substantially lengthened C≡C bond (often 1.28–1.32 Å, versus 1.20 Å for a typical free internal alkyne — a much larger fractional lengthening than seen for alkenes, precisely because of the second available π acceptor orbital discussed above), and a pronounced cis*-bent geometry at each alkyne carbon (substituent bend-back angles of 140° or less from linear 180°, versus the free-alkyne linear reference), all diagnostic of strong four-electron donation with heavy metallacyclopropene-like character.

Limiting-case check: setting the alkyne's donor number to the alkene-like minimum of 2 electrons recovers ordinary two-electron alkyne coordination chemistry (bond lengthening more modest, ~1.24–1.26 Å, bend-back angles closer to 160–170°), which is the majority behavior for alkynes bound to already-saturated 18-electron fragments — showing that the "extra" donor mode of alkynes is a genuine additional option unlocked by electron deficiency at the metal, not a universal property of every bound alkyne.

Limiting cases, scope & common traps

Regime	σ-donation	π-back-donation	C=C/C≡C length	Bend-back angle	Best description
Electron-poor metal, weak π-basicity (e.g., high oxidation state, hard/early metal)	dominant	minimal	close to free value (~1.34 Å alkene)	small (few degrees)	true π-complex, weak perturbation
Zeise's-salt-type (Pt(II), competing Cl⁻ π-donors)	strong	moderate	1.375 Å	small–moderate	canonical DCD π-complex
Electron-rich, low-oxidation-state metal (Ni(0), Pt(0), Pd(0) phosphine adducts)	present	strong	1.43–1.46 Å	large (tens of degrees)	metallacyclopropane limit
Alkyne, ordinary 2e donation, saturated 18e fragment	strong (one π set)	moderate (one π* set)	modest lengthening (~1.24–1.26 Å)	~160–170°	alkyne π-complex
Alkyne, 4e donation, unsaturated fragment before binding	strong (both π sets)	strong (both π* sets)	large lengthening (1.28–1.32 Å)	≤140°	metallacyclopropene

Common trapMisconception: "the alkene is just an electron-pair donor, like a simple Lewis base."

Treating $\eta^2$ -alkene coordination as pure σ-donation (as if the alkene were behaving like ammonia or an ether) misses the entire reason alkene complexes have chemically distinctive structures and reactivities. Pure σ-donation would neither lengthen C=C nor bend back the substituents nor generate the electrophilic character exploited in the Wacker mechanism — all of these are signatures specifically of the back-donation half of the synergic pair. An alkene complex with no back-donation component would in fact bind only weakly, since π-only alkene ligands are comparatively weak Lewis bases (ethylene's π ionization energy is high, ~10.5 eV, versus true amine/phosphine lone pairs) — it is the synergic feedback loop, not σ-donation alone, that makes alkene complexes robust, isolable species.

Common trapMisconception: "more back-donation is always favorable, so the strongest-donor metal gives the most stable alkene complex."

Excess back-donation, taken to its logical extreme, converts the alkene ligand into a metallacyclopropane — at which point the "complex" is arguably better described as a dialkyl (oxidative addition) product rather than an intact alkene adduct, and the alkene has lost its identity as an unsaturated π-ligand altogether. Whether this is "more stable" depends entirely on the specific thermodynamics of the metal-carbon σ-bonds formed versus the alkene π-system's intrinsic stability; it is not simply true that maximal back-donation is always the lowest-energy outcome. The DCD continuum describes a spectrum of real, isolable structures, not a monotonic stability ranking.

The chemist behind it

Joseph Chatt (1914–1994) was a British inorganic chemist whose entire career, spent largely at Imperial Chemical Industries and later at the University of Sussex, was organized around a deceptively simple question: how do transition metals actually bond to "unusual" ligands like alkenes, nitrogen, and phosphines, at a time when Werner-style ionic/coordinate-bond thinking dominated inorganic chemistry? Chatt trained originally in organic chemistry, which gave him an unusually strong instinct for molecular orbital and substituent-effect reasoning applied to metal complexes — exactly the toolkit needed to see that an alkene bound to platinum could not be understood with electrostatics alone. Working with Leslie Duncanson, he published in 1953 the paper that gave quantitative teeth to Dewar's 1951 qualitative suggestion, using infrared spectroscopy of a series of platinum-alkene complexes (extending directly from Zeise's 1827 salt) to show that the systematic red-shift of the C=C stretching frequency upon coordination tracked exactly with the electron-donating or -withdrawing character of the metal's other ligands — direct chemical proof of the back-donation component. Chatt went on to build much of the modern chemistry of dinitrogen and metal-hydride complexes using the same synergic-bonding logic, and the Chatt–Dewar–Duncanson model remains, seven decades later, the standard first-year graduate framework for every π-acid ligand in the periodic table.

Origins · 1 minJoseph Chatt and how this was discovered

Joseph Chatt

1914–1994 · British inorganic/organometallic chemist; ICI and University of Sussex; co-developer with M. J. S. Dewar and L. A. Duncanson of the synergic bonding model for metal–alkene complexes

History

- 1827 — William Christopher Zeise isolates K[PtCl₃(C₂H₄)]·H₂O, the first characterized metal–alkene complex.

- 1951 — Michael J. S. Dewar proposes a molecular-orbital description of metal–olefin bonding, introducing the donor–acceptor synergic concept.

- 1953 — Joseph Chatt and Leslie A. Duncanson publish infrared evidence quantifying back-donation in platinum–alkene complexes, establishing the DCD model in its modern form.

- 1969–1975 — Precise X-ray and neutron diffraction structures of Zeise's anion confirm the perpendicular ethylene orientation and the predicted C=C bond lengthening.

- 1959–1970s — Extension of the DCD framework to alkyne complexes and identification of four-electron alkyne donor modes in electron-deficient early transition-metal carbonyl fragments.

Why does the alkene need to donate through its π system rather than through a single carbon's lone pair, unlike most other ligands?

An alkene has no localized lone pair — its highest-energy filled orbital is the delocalized π MO spread across both carbons. To use this orbital as a donor, the metal must approach side-on, aligning with the π cloud's symmetry, rather than end-on toward a single atom. This is exactly what defines η²-coordination (see Hapticity and the Eta Notation): the ligand donates as a whole conjugated unit, and the "bond" is best drawn from the C=C midpoint to the metal rather than as two separate M–C bonds, at least in the weak-back-donation limit. Only once back-donation becomes strong does a two-separate-σ-bond (metallacyclopropane) picture become equally or more appropriate.

Is Zeise's salt representative of most alkene complexes, or is it a special, weakly-bound edge case?

Zeise's salt sits close to the "textbook-clean π-complex" end of the DCD continuum: Pt(II) is d⁸ but its d-orbitals are somewhat stabilized (contracted, lower in energy) by the three electronegative, better-π-accepting chloride ligands competing for the same back-donating d-electrons, so ethylene only receives modest back-donation. It is representative in the sense that it displays every qualitative DCD feature (perpendicular orientation, modest lengthening, small bend-back, measurable rotational barrier) cleanly and was the first structurally solved example, which is why it remains the default teaching system. But many other alkene complexes — especially those of electron-rich, low-oxidation-state Ni(0)/Pd(0)/Pt(0) fragments — sit much further toward the metallacyclopropane end and are not "more typical," just differently positioned on the same spectrum.

Does the DCD model apply equally well to substituted alkenes like electron-poor alkenes (e.g., tetracyanoethylene) versus electron-rich ones (e.g., tetramethylethylene)?

Yes, and the model correctly predicts the trend: electron-poor alkenes (those with electron-withdrawing substituents lowering the π energy, such as tetracyanoethylene) are better π-acceptors, so they bind more strongly to electron-rich metals via enhanced back-donation, often showing larger C=C bond lengthening than ethylene itself despite being intrinsically weaker σ-donors. Electron-rich alkenes (with alkyl substituents raising π energy and lowering ionization energy of π) are better σ-donors but poorer π-acceptors. This substituent-tunability is one of the strongest pieces of evidence for the two-component DCD model, since a purely electrostatic picture would predict the opposite substituent trend.

How does the DCD picture connect to the mechanism of the Wacker process specifically?

In the Wacker cycle, ethylene first coordinates to a square-planar Pd(II) center (electronically and structurally analogous to Zeise's salt's Pt(II) fragment), undergoing the same σ-donation/π-back-donation synergic bonding. The resulting partial depletion of alkene π-electron density (quantified above via the C=C stretching frequency shift) renders the coordinated carbons measurably electrophilic. An external water molecule (or hydroxide, depending on mechanistic details still debated between outer-sphere anti attack and inner-sphere syn insertion pathways) then attacks one of these activated carbons, initiating the sequence that ultimately delivers acetaldehyde after β-hydride elimination steps related to Migratory Insertion chemistry. Without the back-donation component softening and polarizing the alkene, this nucleophilic attack step would have no kinetic driving force.

Why do alkyne complexes generally show larger fractional bond-length changes than alkene complexes upon coordination?

Because alkynes offer a second, orthogonal π-acceptor orbital unavailable to alkenes, a bound alkyne can, in principle, accept back-donation from two independent filled metal d-orbitals rather than just one. Even in ordinary two-electron-donor alkyne complexes, this second acceptor orbital is often partially engaged (asymmetric back-donation into both π sets to different extents), producing systematically more C≡C bond weakening — and hence proportionally larger elongation from the very short 1.20 Å free-alkyne reference — than is typical for the single-π*-acceptor alkene case relative to its 1.34 Å reference.

Can the same metal fragment bind an alkene in one compound and behave as a full metallacyclopropane in another, seemingly identical, compound?

Yes — this is one of the clearest demonstrations that the π-complex/metallacyclopropane distinction is a continuum rather than a binary classification. Small changes in ancillary ligand electronics (a more electron-donating phosphine, a change in halide identity, or a change in overall charge/oxidation state) can shift the same metal-alkene pairing measurably along the continuum, sometimes enough to flip the preferred structural description in crystallographic and computational studies. This sensitivity is exactly what the DCD model predicts, since both σ-donation and π-back-donation depend continuously on the electron density and orbital energies of the metal fragment.

Does the DCD model require the alkene or alkyne to be a "d-block only" phenomenon, or can main-group elements bind alkenes similarly?

The synergic donor-acceptor logic in principle requires an empty orbital of correct symmetry to accept σ-donation and a filled orbital of correct symmetry, energy, and orientation to supply back-donation — conditions most cleanly satisfied by transition-metal d-orbitals, which are well-positioned in energy to overlap with alkene π/π orbitals. Some main-group Lewis acids (certain boranes, or silylium-type cations) can bind alkenes via a dominant σ-donation-only interaction (agostic-like, weak, largely electrostatic/σ-only complexes), but true back-donation-rich DCD-type bonding with substantial C=C lengthening and bend-back is essentially a transition-metal phenomenon, because most main-group fragments lack a suitably filled, well-oriented orbital at the right energy to donate into alkene π.