Ligand Substitution Mechanisms

Drop a crystal of $\ce{[Ni(H2O)6]^2+}$ into concentrated ammonia. The color changes from pale green to deep blue in a flash. This happens faster than you can mix the solution. Do the same with $\ce{[Cr(H2O)6]^3+}$ or $\ce{[Co(NH3)6]^3+}$ . Nothing visible happens for hours. This occurs even though thermodynamics strongly favor the ammine or hydroxide product. Both reactions are ligand substitutions. One ligand leaves the metal's inner coordination sphere and another takes its place. Yet their rates differ by ten orders of magnitude. Explaining that gulf, and the periodic and structural patterns behind it, is the job of this article. Ligand substitution is arguably the single most useful reaction class in coordination chemistry. It is how complexes are synthesized and how catalytic cycles turn over. It is also how metalloenzymes exchange substrate and product. It is how metal-based drugs like cisplatin find their biological targets. Every such reaction must answer one mechanistic question. Does the incoming ligand attack before, after, or simultaneously with the departure of the outgoing ligand? The answer depends on the metal's oxidation state, $d$ -electron count, coordination number, and geometry. That is why substitution mechanisms are the natural bridge between structural/electronic descriptions of complexes and their reactivity. Those descriptions are built on Crystal Field Theory and d-Orbital Splitting. The big picture is a spectrum, not a binary choice. At one extreme, a five-coordinate intermediate forms first. The incoming ligand attaches afterward — a dissociative (D) pathway. At the other extreme, a seven-coordinate intermediate forms with the incoming ligand already bonded. Only then does the leaving group depart — an associative (A) pathway. Most real reactions sit between these limits. These are interchange mechanisms ( $I_d$ , dissociative-interchange, or $I_a$ , associative-interchange). In these, bond-breaking and bond-making are concerted but unequal in extent. The mechanistic assignment is not a matter of taste. It is diagnosed rigorously from rate laws, volumes of activation, and the sensitivity of rate to the incoming ligand's identity. This article builds the associative/dissociative/interchange framework from first principles. It covers activated-complex energetics and the kinetic signatures that distinguish the pathways. It also covers the structural correlations — coordination number, $d$ -count, ligand field stabilization. These correlations explain why octahedral $d^3$ and low-spin $d^6$ complexes are inert. They also explain why square-planar $d^8$ complexes are almost universally associative. It sets up the machinery needed to understand The Trans Effect and Crystal Field Activation Energy. Both of these are refinements of the ideas developed here.

Key takeaways

Ligand substitution mechanisms span a continuum from dissociative (D, rate-limiting bond breaking) to associative (A, rate-limiting bond making), with interchange ( $I_d$ , $I_a$ ) mechanisms in between.
The mechanism is diagnosed from the rate law's dependence on $[\text{incoming ligand}]$ , from the sign and magnitude of the volume of activation $\Delta V^\ddagger$ , and from steric/electronic trends across metals.
Octahedral $d^3$ and low-spin $d^6$ complexes are kinetically inert because both associative and dissociative pathways cost large crystal-field activation energy; $e_g^1$ configurations ( $d^4$ high-spin, $d^9$ ) are labile via Jahn–Teller-assisted pathways.
Square-planar $d^8$ complexes (Pt(II), Pd(II), Ni(II)) react almost exclusively by an associative mechanism because the vacant axial sites make five-coordinate trigonal-bipyramidal intermediates cheap.
The classic two-term rate law $k_{obs} = k_1 + k_2[\text{Y}]$ for square-planar substitution captures a solvent-assisted (or genuinely associative) pathway running in parallel with direct attack by the incoming ligand.
Key idea: substitution rate and mechanism are controlled by the energy of the transition state relative to the ground state, and that energy is set by how much (or little) crystal-field stabilization energy is sacrificed in going to five- or seven-coordinate geometry.

The core insight

Ligand substitution mechanisms form a continuum defined by the degree of bond-making versus bond-breaking in the transition state; the position of a given metal–ligand system on that continuum is fixed by the change in crystal-field/ligand-field stabilization energy between the octahedral (or square-planar) ground state and the five- or seven-coordinate transition state.

Where this lives today

The associative mechanism of square-planar substitution is the mechanistic backbone of platinum-based chemotherapy. Cisplatin, $\ce{cis-[PtCl2(NH3)2]}$ , must first undergo aquation — chloride ligands substituted by water, an associative process at Pt(II) — before the resulting aqua/ammine complex can attack guanine N7 sites on DNA to form the intrastrand crosslinks responsible for its anticancer activity. The rate and selectivity of that initial substitution, tuned by ligand and leaving-group choice, is precisely why newer platinum drugs (carboplatin, oxaliplatin) use different leaving groups: carboplatin's chelating dicarboxylate leaving group substitutes more slowly than chloride, reducing systemic toxicity while preserving the same associative aquation mechanism. Industrial homogeneous catalysis leans on the same framework in reverse. Wilkinson's catalyst, $\ce{RhCl(PPh3)3}$ , initiates alkene hydrogenation by dissociative loss of a phosphine ligand to open a coordination site — a $d^8$ square-planar rhodium(I) center behaving, at this step, dissociatively rather than associatively, illustrating that the square-planar-implies-associative rule of thumb is a strong statistical trend, not an ironclad law, and depends on the specific ligand set and steric bulk of the phosphines involved. Catalyst design routinely exploits controlled lability: labile ligands (solvento complexes, weakly bound alkenes) are deliberately built into precatalysts so that substitution by substrate is fast and reversible, while spectator ligands are chosen to be inert enough to survive the catalytic cycle intact. In bioinorganic chemistry, the kinetic inertness of low-spin $d^6$ cobalt(III) is exploited experimentally: inert $\ce{Co(III)}$ complexes are used as structural and mechanistic mimics for labile, catalytically active $\ce{Co(II)}$ or other first-row centers precisely because their substitution chemistry is slow enough to allow isolation and crystallographic characterization of intermediates that would otherwise exchange too fast to trap. Water-exchange rate constants and their associated $\Delta V^\ddagger$ values, first tabulated systematically by Taube, Merbach, and coworkers, remain the standard reference data used to validate computational (DFT, ab initio molecular dynamics) predictions of ligand-exchange mechanisms for lanthanide and actinide complexes relevant to nuclear waste separations and MRI contrast agent design, where the rate of water exchange at the metal center directly controls the contrast agent's relaxivity and hence image quality.

:::quiz[A rate law $k_{obs} = k_1 + k_2[\ce{Y}]$ is found for substitution at a square-planar Pt(II) complex. What is the best interpretation of the $k_1$ term?]

Pure unimolecular dissociation of the leaving group with no ligand participation at all
A solvent-assisted associative pathway, not evidence for a dissociative mechanism ✓
Proof that the reaction is zero order overall
An artifact that should be subtracted out before interpreting the mechanism

:::faq[Why is [Co(NH3)6]3+ inert even though it is thermodynamically unstable in acid?] Kinetic inertness and thermodynamic stability are independent properties. $\ce{[Co(NH3)6]^3+}$ is low-spin $d^6$ , meaning its octahedral ground state already enjoys the maximum possible crystal-field stabilization energy ( $-2.4\,\Delta_o$ before pairing-energy corrections). Any five- or seven-coordinate transition-state geometry loses much of that stabilization, creating a large crystal-field activation energy that must be overcome regardless of how favorable the products are thermodynamically. The complex is "waiting" behind a large kinetic barrier even though it is rolling downhill thermodynamically toward hydrolysis products.

Ask the tutor: How does the trans effect quantitatively modify the associative transition-state energy at square-planar centers, and why do π-acceptor ligands accelerate substitution more than pure σ-donors?

The physical picture

Consider an octahedral complex $\ce{ML6}$ exchanging one ligand $\ce{L}$ for an incoming ligand $\ce{Y}$ :

\ce{ML6 + Y -> ML5Y + L}

Three limiting mechanistic pictures are possible for how this bond reorganization happens, and they are defined not by loose language but by the molecularity of the rate-determining transition state.

Dissociative (D). The $\ce{M-L}$ bond breaks completely before $\ce{Y}$ approaches. A genuine five-coordinate intermediate, $\ce{ML5}$ , exists with a finite (if short) lifetime, long enough to be captured and to discriminate between different possible incoming ligands. This is mechanistically analogous to $S_N1$ in organic substitution: the rate depends on $[\ce{ML6}]$ only, not on $[\ce{Y}]$ , because the ligand that captures the intermediate is chosen after the rate-determining step.

Associative (A). The $\ce{M-Y}$ bond forms completely before $\ce{M-L}$ breaks. A genuine seven-coordinate (or, for square-planar starting materials, five-coordinate) intermediate forms, with a finite lifetime and its own characterizable structure, before ejecting $\ce{L}$ . This is the analogue of $S_N2$ -with-stable-intermediate: the rate law is first order in $[\ce{Y}]$ because $\ce{Y}$ participates in the rate-determining step.

Interchange (I). Bond-breaking and bond-making occur in a single concerted step through one transition state, with no intermediate of extended lifetime. Interchange mechanisms are further subdivided by which bond-order change dominates the transition-state energy: $I_d$ (dissociative interchange) resembles D in that $\ce{M-L}$ breaking is more advanced than $\ce{M-Y}$ making, while $I_a$ (associative interchange) resembles A in the reverse sense. Interchange, not the clean limiting cases, is how the overwhelming majority of real substitutions at octahedral centers proceed — most water-exchange reactions of first-row transition-metal aqua ions are $I_d$ or $I_a$ , rarely pure D or A. The energetic origin of these preferences is ligand-field stabilization. An octahedral $d^n$ ion loses some fraction of its crystal-field stabilization energy (CFSE) when it distorts toward a five-coordinate square pyramid (dissociative pathway) or a seven-coordinate pentagonal bipyramid (associative pathway), because the $d$ -orbital splitting pattern in these lower-symmetry geometries is less favorable. The size of that CFSE loss — the crystal-field activation energy, CFAE — differs enormously between configurations. $d^3$ ( $\ce{Cr^3+}$ ) and low-spin $d^6$ ( $\ce{Co^3+}$ , $\ce{Rh^3+}$ , $\ce{Pt^4+}$ ) lose the most stabilization on either pathway and are correspondingly inert; $d^0$ , high-spin $d^5$ , and $d^{10}$ lose essentially nothing (no CFSE to lose) and are labile.

Figure

Dissociative vs. associative vs. interchange reaction coordinates

The kinetic fingerprints that distinguish these pathways are what make the mechanistic assignment testable rather than speculative. Three diagnostics dominate practice: (i) the dependence of the observed rate on $[\ce{Y}]$ (zero-order for pure D, first-order for pure A, a two-term law for mixed/associative-interchange behavior at square-planar centers); (ii) the sign of the volume of activation $\Delta V^\ddagger$ , obtained from the pressure dependence of the rate constant; and (iii) the sensitivity of rate to the nature of the incoming ligand — associative pathways are sensitive to the nucleophilicity and steric bulk of $\ce{Y}$ , dissociative pathways are almost insensitive to $\ce{Y}$ because $\ce{Y}$ is not involved until after the rate-determining step.

Figure

Volume of activation as a mechanistic probe

The core ideas

The rate law is the primary experimental handle, and it is built from the elementary-step mechanism using standard steady-state approximation treatment applied to the intermediate.

k_{obs} = k_1 + k_2[\ce{Y}]

where: $k_{obs}$ is the observed pseudo-first-order rate constant for disappearance of $\ce{ML6}$ (or $\ce{ML5X}$ in the classic square-planar case); $k_1$ is the rate constant for the ligand-independent (solvent-assisted or dissociative) pathway; $k_2$ is the rate constant for the bimolecular, ligand-dependent (associative) pathway; $[\ce{Y}]$ is the concentration of the incoming nucleophile.

Reading: a nonzero intercept $k_1$ signals a pathway that proceeds without $\ce{Y}$ in the rate-determining step (dissociative character, or associative attack by solvent that is then displaced by $\ce{Y}$ in a fast step); the slope $k_2$ signals genuine associative attack by $\ce{Y}$ itself. Plotting $k_{obs}$ against $[\ce{Y}]$ and finding a straight line with positive slope and positive intercept is the standard signature of square-planar substitution.

\Delta V^\ddagger = V^\ddagger - V_{\text{reactants}} = -RT\left(\frac{\partial \ln k}{\partial P}\right)_T

where: $\Delta V^\ddagger$ is the activation volume; $V^\ddagger$ is the (hypothetical) partial molar volume of the transition state; $R$ is the gas constant; $T$ is temperature; $P$ is pressure; $k$ is the rate constant measured as a function of applied pressure.

Reading: a negative $\Delta V^\ddagger$ means the system contracts on the way to the transition state — consistent with bond formation dominating (associative character); a positive $\Delta V^\ddagger$ means expansion — consistent with bond breaking dominating (dissociative character). Because $\Delta V^\ddagger$ is measured directly from high-pressure kinetics without needing to isolate any intermediate, it is one of the cleanest mechanistic diagnostics available for exchange reactions in solution.

\text{CFAE} = \text{CFSE}(\text{TS geometry}) - \text{CFSE}(\text{octahedral, ground state})

where: CFAE is the crystal-field activation energy, the penalty (usually a loss of stabilization, so CFAE $>0$ in cost) incurred on distorting from octahedral toward the transition-state geometry (square pyramidal for D, pentagonal bipyramidal for A); CFSE values are tabulated in units of $\Delta_o$ for each $d^n$ configuration and geometry.

Reading: a large CFAE means the metal center strongly resists forming either a five- or seven-coordinate transition state, making it substitution-inert regardless of thermodynamic driving force; a small or negative CFAE (some geometries stabilize particular $d^n$ configurations relative to octahedral) predicts high lability. This is the electronic underpinning of the classic inert/labile classification and is treated quantitatively as its own topic under Crystal Field Activation Energy.

Derivation AThe two-term rate law for square-planar substitution

Goal: derive $k_{obs} = k_1 + k_2[\ce{Y}]$ for substitution at square-planar $d^8$ centers (archetypally Pt(II)) from a mechanism with two parallel pathways, and show why both terms are, at the elementary-step level, associative.

Step 1 — write the parallel pathways. Square-planar $\ce{ML3X}$ reacting with incoming ligand $\ce{Y}$ in a coordinating solvent $\ce{S}$ follows two channels operating simultaneously. Pathway (a), the solvent path: solvent $\ce{S}$ attacks the vacant axial site of the square plane to give a five-coordinate intermediate $\ce{ML3XS}$ , which then loses $\ce{X}$ to give $\ce{ML3S}$ , which is rapidly attacked by $\ce{Y}$ to displace $\ce{S}$ and give the product $\ce{ML3Y}$ . Pathway (b), the direct path: $\ce{Y}$ itself attacks the vacant axial site directly, forming a five-coordinate intermediate $\ce{ML3XY}$ that loses $\ce{X}$ to give the same product $\ce{ML3Y}$ .

Step 2 — apply the steady-state (or pre-equilibrium) treatment to each channel. For the solvent path, since $[\ce{S}]$ (the solvent) is constant and enormous relative to solute concentrations, the pseudo-first-order rate constant for that channel, $k_1$ , absorbs $k_a[\ce{S}]$ into a single empirical constant that shows no explicit $[\ce{Y}]$ dependence:

\text{rate}_{(a)} = k_a[\ce{S}][\ce{ML3X}] \equiv k_1[\ce{ML3X}]

For the direct path, $\ce{Y}$ appears explicitly in the rate-determining attack step:

\text{rate}_{(b)} = k_2[\ce{Y}][\ce{ML3X}]

Step 3 — sum the parallel rates. Because the two channels operate independently and both consume $\ce{ML3X}$ , the total rate of disappearance of starting material is the sum:

-\frac{d[\ce{ML3X}]}{dt} = (k_1 + k_2[\ce{Y}])[\ce{ML3X}]

so that $k_{obs} = k_1 + k_2[\ce{Y}]$ , exactly the empirical law found for essentially all Pt(II) and Pd(II) substitutions studied since the 1950s.

Step 4 — interpret the mechanistic content of each term. Both channels are associative at the elementary-step level (a nucleophile occupies the vacant axial site before $\ce{X}$ leaves), which is why square-planar $d^8$ chemistry is the textbook home of clean $S_N2$ -like behavior in inorganic kinetics — the vacant axial positions above and below the square plane make a trigonal-bipyramidal five-coordinate transition state/intermediate cheap to form, unlike in octahedral complexes where all six sites are already occupied. The $k_1$ term is not evidence for a dissociative mechanism; it is evidence for a different associative pathway (solvent-assisted) that happens to be zero order in $\ce{Y}$ because solvent is in vast excess and does not vary during the run.

Step 5 — limiting case check. As $[\ce{Y}] \to 0$ (e.g., extrapolating a kinetic run to negligible added nucleophile), $k_{obs} \to k_1$ , the pure solvolysis rate constant, which is exactly what is measured as the "spontaneous" hydrolysis rate constant for platinum(II) anticancer complexes such as cisplatin in water. As $[\ce{Y}]$ becomes very large (large excess of a strong, non-hindered nucleophile like $\ce{I^-}$ or thiourea), the $k_2[\ce{Y}]$ term dominates and the reaction becomes cleanly first-order in $\ce{Y}$ , exposing the direct associative channel.

Go deeper · Derivation B · 2 minCrystal-field activation energy discriminates inert from labile octahedral centers

Derivation BCrystal-field activation energy discriminates inert from labile octahedral centers

Goal: show, using CFSE bookkeeping in units of $\Delta_o$ , why $d^3$ and low-spin $d^6$ octahedral complexes resist both associative and dissociative substitution pathways, while $d^1$ , high-spin $d^5$ , and $d^{10}$ do not.

Step 1 — tabulate octahedral CFSE. For a weak-to-moderate field (or, for $d^6$ , both high- and low-spin cases), standard CFSE values in units of $\Delta_o$ are: $d^0$ : 0; $d^1$ : $-0.4$ ; $d^2$ : $-0.8$ ; $d^3$ : $-1.2$ ; high-spin $d^5$ : 0; low-spin $d^6$ : $-2.4$ ; high-spin $d^6$ : $-0.4$ ; $d^8$ : $-1.2$ ; $d^{10}$ : 0. (Pairing energy contributions are omitted here since we compare geometries at fixed spin state.)

Step 2 — tabulate the transition-state/intermediate CFSE for the dissociative (square-pyramidal, five-coordinate) and associative (pentagonal-bipyramidal, seven-coordinate) geometries. These lower-symmetry ligand fields split the five $d$ -orbitals into more levels than $O_h$ does, and detailed angular-overlap calculations (standard tabulated results) give, for a square pyramid built by removing one ligand along $z$ from an octahedron, stabilization energies that are systematically smaller in magnitude than the parent octahedral CFSE for $d^3$ and low-spin $d^6$ , because the geometry no longer supports the especially favorable $t_{2g}^3$ or $t_{2g}^6$ arrangement. Numerically (illustrative values used across standard treatments, in units of $\Delta_o$ ): square-pyramidal CFSE for $d^3$ is about $-0.80$ to $-0.93$ (versus $-1.2$ octahedral), and for low-spin $d^6$ about $-1.68$ to $-2.00$ (versus $-2.4$ octahedral). Pentagonal-bipyramidal CFSE for these configurations is likewise reduced relative to the octahedral value.

Step 3 — form the CFAE as a difference. For $d^3$ (dissociative pathway, illustrative):

\text{CFAE}_{d^3,\,D} = (-0.93\,\Delta_o) - (-1.2\,\Delta_o) = +0.27\,\Delta_o

For low-spin $d^6$ (dissociative pathway, illustrative):

\text{CFAE}_{d^6_{LS},\,D} = (-1.8\,\Delta_o) - (-2.4\,\Delta_o) = +0.6\,\Delta_o

Both are substantial positive penalties — energy that must be supplied to reach the transition state on top of whatever bond-reorganization energy is already required. Since $\Delta_o$ for these ions is large (both are typically strong-field, highly charged or heavy ions: $\ce{Cr^3+}$ , $\ce{Co^3+}$ , $\ce{Rh^3+}$ , $\ce{Pt^4+}$ ), the absolute CFAE in kJ/mol is large enough to place these complexes firmly in the kinetically inert category (half-lives of hours to years for water exchange or substitution), even though many of them are thermodynamically quite reactive (unstable) with respect to substitution.

Step 4 — contrast with a labile configuration. For $d^0$ , high-spin $d^5$ , or $d^{10}$ , the octahedral CFSE is already zero, and the transition-state CFSE is also (to good approximation) zero, so:

\text{CFAE}_{d^0 \text{ or } d^5_{HS} \text{ or } d^{10}} \approx 0

There is no crystal-field barrier at all; substitution rates for these ions (e.g., $\ce{[Ca(H2O)6]^2+}$ , $\ce{[Mn(H2O)6]^2+}$ , $\ce{[Zn(H2O)6]^2+}$ ) are set almost entirely by simple electrostatics and ionic radius, and they exchange water on the nanosecond timescale.

Step 5 — limiting-case check with an intermediate configuration. Take high-spin $d^4$ ( $\ce{Cr^2+}$ , $\ce{Mn^3+}$ ), a strongly Jahn–Teller active configuration in $O_h$ symmetry (one electron in the doubly degenerate $e_g$ set). The octahedral ground state is already tetragonally distorted and only weakly stabilized net; distortion toward a five-coordinate transition-state geometry costs very little additional CFSE because the ion is already primed for asymmetric coordination. Correspondingly, $\ce{Cr^2+}$ and $\ce{Mn^3+}$ aqua ions are amongst the most labile of all first-row transition-metal ions — orders of magnitude faster than $\ce{Cr^3+}$ ( $d^3$ ) even though they sit right next to it in the periodic table, which is exactly the qualitative prediction the CFAE analysis makes and a clean check that the reasoning, not just the arithmetic, is doing the work.

Worked examples

Example 1 (foundational): reading a two-term rate law. A kinetics study of $\ce{trans-[Pt(py)2Cl2] + SCN^- -> trans-[Pt(py)2Cl(SCN)] + Cl^-}$ in methanol at 25 °C gives the following $k_{obs}$ values at varying $[\ce{SCN^-}]$ :

$[\ce{SCN^-}]$ / mol dm⁻³	$k_{obs}$ / s⁻¹
0.010	$3.4\times10^{-5}$
0.040	$6.8\times10^{-5}$
0.100	$1.19\times10^{-4}$

Assume $k_{obs} = k_1 + k_2[\ce{SCN^-}]$ . Using the first and third points to solve simultaneously:

k_2 = \frac{(1.19\times10^{-4}) - (3.4\times10^{-5})}{0.100 - 0.010} = \frac{8.5\times10^{-5}}{0.090} = 9.4\times10^{-4}\ \text{dm}^3\,\text{mol}^{-1}\,\text{s}^{-1}

Substituting back:

k_1 = 3.4\times10^{-5} - (9.4\times10^{-4})(0.010) = 3.4\times10^{-5} - 9.4\times10^{-6} = 2.5\times10^{-5}\ \text{s}^{-1}

Checking against the middle point: $k_{obs}(0.040) = 2.5\times10^{-5} + (9.4\times10^{-4})(0.040) = 2.5\times10^{-5} + 3.76\times10^{-5} = 6.26\times10^{-5}\ \text{s}^{-1}$ , close to the observed $6.8\times10^{-5}\ \text{s}^{-1}$ (the small discrepancy is typical of real noisy data and does not change the mechanistic conclusion). Both $k_1$ (solvent path) and $k_2$ (direct attack by $\ce{SCN^-}$ ) are nonzero and positive, confirming the standard associative two-channel mechanism expected for square-planar Pt(II), and showing that even the "spontaneous" pathway is mechanistically associative (solvent-assisted), not a genuine unimolecular dissociation.

Example 2 (applied): predicting relative lability across a first-row series. Rank the following high-spin octahedral aqua ions in order of increasing water-exchange rate constant, using CFSE/CFAE reasoning alone: $\ce{[Cr(H2O)6]^3+}$ ( $d^3$ ), $\ce{[Fe(H2O)6]^2+}$ (high-spin $d^6$ ), $\ce{[Mn(H2O)6]^2+}$ (high-spin $d^5$ ), $\ce{[Cu(H2O)6]^2+}$ ( $d^9$ ).

$\ce{Mn^2+}$ ( $d^5$ , high-spin): octahedral CFSE $= 0\,\Delta_o$ ; CFAE $\approx 0$ . Predicted: extremely labile.
$\ce{Fe^2+}$ (high-spin $d^6$ ): octahedral CFSE $=-0.4\,\Delta_o$ , a modest stabilization only from one $t_{2g}$ -favored electron beyond the half-filled shell; CFAE is small. Predicted: labile, only slightly slower than $\ce{Mn^2+}$ .
$\ce{Cu^2+}$ ( $d^9$ ): strongly Jahn–Teller distorted already in the ground state (analogous reasoning to $d^4$ above — one hole in $e_g$ ); distortion toward a five-coordinate transition state costs little extra CFSE. Predicted: extremely labile (in practice, among the fastest-exchanging first-row ions, exchange rate constants around $10^8$ – $10^9$ s⁻¹).
$\ce{Cr^3+}$ ( $d^3$ ): octahedral CFSE $=-1.2\,\Delta_o$ , the maximum stabilization available among weak-field high-spin configurations, and CFAE is large and positive as derived above. Predicted: by far the slowest.

The experimentally measured water-exchange rate constants (s⁻¹, 25 °C) confirm this ordering essentially exactly: $\ce{Cu^2+}\ (\sim4\times10^9) > \ce{Mn^2+}\ (\sim2\times10^7) > \ce{Fe^2+}\ (\sim4\times10^6) \gg \ce{Cr^3+}\ (\sim2\times10^{-6})$ . The span from $\ce{Cu^2+}$ to $\ce{Cr^3+}$ is roughly fifteen orders of magnitude, driven almost entirely by CFAE and Jahn–Teller effects rather than by charge or size, since $\ce{Cr^3+}$ and $\ce{Fe^2+}$ / $\ce{Mn^2+}$ have comparable ionic radii.

Example 3 (exam/competition level): extracting $\Delta V^\ddagger$ and assigning mechanism from pressure-dependent kinetics. The rate constant for water exchange on $\ce{[V(H2O)6]^3+}$ is measured at 25 °C as a function of applied pressure:

$P$ / MPa	$k$ / s⁻¹
0.1	$5.0\times10^{1}$
100	$6.6\times10^{1}$
200	$8.7\times10^{1}$

Determine $\Delta V^\ddagger$ and assign the mechanism.

Step 1. Convert to $\ln k$ : $\ln(50.0) = 3.912$ ; $\ln(66.0) = 4.190$ ; $\ln(87.0) = 4.466$ .

Step 2. Estimate the slope $\partial \ln k/\partial P$ using the outer two points (linear regression on all three gives essentially the same slope since the data are closely linear):

\frac{\partial \ln k}{\partial P} \approx \frac{4.466 - 3.912}{(200 - 0.1)\ \text{MPa}} = \frac{0.554}{199.9\ \text{MPa}} = 2.77\times10^{-3}\ \text{MPa}^{-1}

Step 3. Convert MPa⁻¹ to the SI-consistent volume unit. Since $\Delta V^\ddagger = -RT(\partial \ln k/\partial P)$ , and $1\ \text{MPa} = 10^6\ \text{Pa} = 10^6\ \text{J m}^{-3}$ , work in J and m³ then convert to cm³ mol⁻¹:

\Delta V^\ddagger = -(8.314\ \text{J mol}^{-1}\text{K}^{-1})(298\ \text{K})(2.77\times10^{-3}\ \text{MPa}^{-1})

\Delta V^\ddagger = -(2478\ \text{J mol}^{-1})(2.77\times10^{-9}\ \text{Pa}^{-1}) = -6.86\times10^{-6}\ \text{m}^3\,\text{mol}^{-1}

\Delta V^\ddagger = -6.86\times10^{-6}\ \text{m}^3\text{mol}^{-1} \times 10^6\ \text{cm}^3\text{m}^{-3} = -6.9\ \text{cm}^3\,\text{mol}^{-1}

Step 4 — assign mechanism. The negative sign means the activation process contracts the system's volume — bond formation to the incoming water molecule dominates the transition state. This is the signature of an associative or associative-interchange ( $I_a$ ) pathway, consistent with the accepted assignment for $\ce{V^3+}$ ( $d^2$ ) water exchange, where the modest $d$ -electron count and moderate charge leave enough "room" and driving force for the incoming water to engage the metal center before the leaving water has fully departed. Note the contrast with $\ce{[Ni(H2O)6]^2+}$ water exchange, where $\Delta V^\ddagger \approx +7.2\ \text{cm}^3\,\text{mol}^{-1}$ (positive, bond-breaking dominated, $I_d$ ) despite both ions being kinetically labile overall — lability and mechanism are separate axes, not the same question.

Limiting-case check. As $\Delta V^\ddagger \to 0$ , the mechanism is a "pure" interchange with essentially equal, compensating amounts of bond making and bond breaking at the transition state (often reported for $\ce{[Cr(H2O)6]^3+}$ 's isotopic self-exchange region among the $I$ classification, though that ion's overall rate is CFAE-dominated and slow regardless of the near-zero $\Delta V^\ddagger$ ).

Limiting cases, scope & common traps

Situation	What happens	Why
Octahedral, $d^0$ , high-spin $d^5$ , $d^{10}$	Extremely labile, near-diffusion-controlled exchange	CFAE ≈ 0; no crystal-field barrier on either pathway
Octahedral, $d^3$ or low-spin $d^6$	Kinetically inert (hours–years)	Large, unavoidable CFAE on both associative and dissociative routes
Octahedral, $d^4$ (HS) or $d^9$	Anomalously labile despite intermediate CFSE	Jahn–Teller-distorted ground state already "primed" for asymmetric distortion; low CFAE
Square-planar $d^8$ (Pt²⁺, Pd²⁺, Ni²⁺, Au³⁺)	Almost always associative, two-term rate law	Vacant axial sites make trigonal-bipyramidal 5-coordinate TS cheap; strong metal preference for square-planar geometry disfavors dissociative loss of a ligand
High pressure kinetics, $\Delta V^\ddagger < 0$	Associative/ $I_a$ character	Bond-making dominates volume change at TS
High pressure kinetics, $\Delta V^\ddagger > 0$	Dissociative/ $I_d$ character	Bond-breaking dominates volume change at TS
Rate independent of $[\ce{Y}]$ identity/concentration	D or strongly $I_d$	$\ce{Y}$ not involved in (or only weakly involved in) the rate-determining step
Rate strongly dependent on nucleophilicity/sterics of $\ce{Y}$	A or strongly $I_a$	$\ce{Y}$ directly engaged in bond formation at the transition state

Common trapCommon misconception: "inert" means "thermodynamically stable," and "labile" means "unstable."

Inertness and lability are kinetic terms describing how fast a complex substitutes; stability and instability are thermodynamic terms describing the sign and magnitude of $\Delta G$ for a substitution reaction. $\ce{[Co(NH3)6]^3+}$ is thermodynamically unstable with respect to hydrolysis in acid (large equilibrium constant favoring aquo/ammonium products) yet kinetically inert — it survives for days because low-spin $d^6$ has a huge crystal-field activation energy. Conversely $\ce{[Ni(CN)4]^2-}$ can be thermodynamically very stable (huge formation constant) yet still undergo rapid ligand exchange. Never infer mechanism or rate from $K_{eq}$ alone.

The chemist behind it

Origins · 1 minHenry Taube and how this was discovered

Henry Taube

1915–2005 · Physical inorganic chemistry, Stanford University and University of Chicago

Henry Taube built the entire modern framework for classifying substitution and electron-transfer reactions of transition-metal complexes by rate. Working first at Chicago and then at Stanford from the 1960s onward, Taube systematically measured water-exchange rates across the periodic table using isotopic labeling ( $\ce{^{18}O}$ and later NMR line-broadening techniques), establishing the correlation between $d$ -electron configuration, ionic charge/radius, and substitution lability that underlies the CFAE analysis used throughout this article. He coined the operational distinction between labile complexes (equilibrate within about a minute at room temperature, roughly $10^{-2}$ M concentrations) and inert complexes (react more slowly), a definition explicitly kinetic rather than thermodynamic — precisely the distinction most often blurred by students and precisely the one Taube insisted on. His parallel work on inner-sphere versus outer-sphere electron-transfer mechanisms, for which he received the 1983 Nobel Prize in Chemistry, is mechanistically inseparable from ligand substitution: an inner-sphere electron-transfer pathway is, at its core, a bridging-ligand substitution event that temporarily links two metal centers. Taube's insistence on quantitative kinetic data, activation parameters, and mechanistic tests (rather than qualitative reasoning from stability alone) set the methodological standard that the associative/dissociative/interchange classification still follows today.

History

- 1952 — Taube publishes the first systematic classification of labile vs. inert complexes based on measured substitution half-lives, correlating rates with

d

-electron configuration.

- 1965 — Basolo and Pearson's Mechanisms of Inorganic Reactions codifies the D/A/I mechanistic vocabulary and the two-term rate law for square-planar substitution.

- 1968 — High-pressure kinetics methods mature, allowing routine determination of

\Delta V^\ddagger

for water-exchange and substitution reactions as an independent mechanistic probe.

- 1983 — Taube awarded the Nobel Prize in Chemistry for mechanistic studies of electron-transfer reactions in metal complexes, work built on the same substitution-mechanism foundations.

- 1990s–2000s — Variable-pressure and variable-temperature NMR studies (Merbach and coworkers) systematically map

\Delta V^\ddagger

across the first-, second-, and third-row transition-metal aqua ions, confirming the CFAE-based predictions with quantitative precision.

What is the difference between an interchange mechanism and a "true" associative or dissociative mechanism?

In a true associative (A) or dissociative (D) mechanism, a genuine intermediate exists with a finite lifetime long enough to have its own chemical identity — it can, in principle, be intercepted, show selectivity toward competing incoming ligands, or be directly detected. In an interchange (I) mechanism, bond-breaking and bond-making occur essentially together in one transition state, and no such long-lived intermediate ever forms. The interchange classification is further split into $I_d$ (more dissociative character, bond-breaking more advanced at the TS) and $I_a$ (more associative character), based on activation volumes and ligand-dependence data. Most octahedral substitutions in solution are interchange, not the clean textbook limits.

How does the volume of activation actually get measured experimentally?

Rate constants for the same reaction are measured at a fixed temperature but at several different applied pressures, typically from atmospheric pressure up to several hundred megapascals, often using stopped-flow or NMR line-broadening apparatus built inside a high-pressure cell. Plotting $\ln k$ against pressure gives a line (or gentle curve) whose slope is $-\Delta V^\ddagger/RT$ ; multiplying the slope by $-RT$ and converting units yields $\Delta V^\ddagger$ directly in cm³ mol⁻¹. No intermediate needs to be isolated, which is what makes this technique so valuable for interchange mechanisms where no discrete intermediate exists to characterize by other means.

Does a fast reaction always mean an associative mechanism, and a slow reaction always mean dissociative?

No — rate (labile vs. inert) and mechanism (associative vs. dissociative character) are logically separate questions, even though students frequently conflate them. $\ce{[Ni(H2O)6]^2+}$ is highly labile (fast) yet its water exchange is dissociative-interchange ( $I_d$ , positive $\Delta V^\ddagger$ ). $\ce{[Cr(H2O)6]^3+}$ is inert (slow) and is generally described as interchange with only weak associative character. Lability is set by the overall activation energy (CFAE plus other contributions); mechanism is set by which bond-order change — making or breaking — dominates the transition state's structure. A reaction can be simultaneously fast and dissociative, or slow and associative.

Why does the trans effect matter for predicting substitution products, and how does it relate to mechanism?

The trans effect describes how a ligand already bound to a square-planar metal center accelerates substitution of the ligand positioned trans to it, by stabilizing the trigonal-bipyramidal associative transition state (through both σ-donation weakening the trans M–L bond and π-acceptor ligands stabilizing the developing five-coordinate intermediate electronically). Because square-planar substitution is associative, the identity and electronic character of the spectator ligands directly control which position reacts fastest and hence which geometric isomer forms — this is the synthetic basis for directed syntheses of cis- and trans-platinum complexes. The trans effect is best understood as a refinement layered on top of the associative mechanism developed here, not a separate mechanism in its own right.

Can a single metal complex switch between associative and dissociative mechanisms depending on conditions?

Yes. The same octahedral or square-planar complex can shift its dominant pathway depending on the incoming ligand's nucleophilicity, the solvent's coordinating ability, steric bulk of spectator ligands, and applied pressure or temperature. For example, increasing steric crowding around a metal center can suppress the associative pathway (blocking approach of $\ce{Y}$ ) and push the mechanism toward more dissociative character, sometimes reflected in a measurable shift of $\Delta V^\ddagger$ from negative toward positive values as ligand bulk increases across a structurally related series. Mechanism is a property of a specific reaction under specific conditions, not an immutable property of the metal alone.

What experimental evidence rules out a genuine five- or seven-coordinate intermediate in most octahedral substitutions?

For a true A or D mechanism, the rate law and product distribution should show clean saturation kinetics (for A, rate leveling off at high $[\ce{Y}]$ as the intermediate-forming step becomes rate-limiting) or complete insensitivity to the incoming ligand's identity (for D). Most octahedral substitutions instead show rates that vary continuously and roughly linearly with ligand nucleophilicity or leaving-group ability without saturation, and activation volumes intermediate between the clean D and A limits — behavior only consistent with a single concerted interchange transition state rather than a stepwise mechanism with a genuine, discriminating intermediate.