Labile vs Inert Complexes

The labile/inert distinction is not a historical curiosity; it is load-bearing in modern chemistry, biology, and technology. In medicinal inorganic chemistry, the entire design logic of platinum anticancer drugs (cisplatin, carboplatin, oxaliplatin) exploits the associative, moderately slow substitution kinetics of square-planar Pt(II) — fast enough to react with DNA on a therapeutically useful timescale, slow enough (especially for carboplatin's chelating dicarboxylate leaving group) to survive transport through the bloodstream, illustrating exactly the Ligand Substitution Mechanisms that this article has only sketched. Bioinorganic chemistry relies on inertness for structural metal sites: the substitution-inert $\text{Co(III)}$ and $\text{Cr(III)}$ centers are classic tools for making stable, isolable model complexes whose stereochemistry doesn't scramble during workup, while labile $\text{Zn}^{2+}$ and $\text{Ca}^{2+}$ sites in enzymes (carbonic anhydrase, calmodulin) are chosen by evolution precisely because rapid ligand exchange is mechanistically essential for catalysis and signaling. In radiopharmaceutical and MRI contrast agent chemistry, chelate design for $\text{Gd}^{3+}$ (a labile, high-spin $f^7$-like ion with essentially no CFAE-type penalty) must compensate for intrinsic lability using high-denticity macrocyclic ligands (DOTA, DTPA) to kinetically trap the metal and prevent toxic release of free $\text{Gd}^{3+}$ in vivo — a direct engineering response to the fact that the bare aqua ion would otherwise exchange far too readily to be safe. Analytical and industrial catalysis exploits inert complexes as robust, isolable, and stereochemically well-defined pre-catalysts (many Rh(III), Ir(III), Ru(II)/(III) complexes used in homogeneous catalysis are deliberately chosen as inert scaffolds that only become catalytically active after a controlled, rate-limiting ligand-dissociation step), while labile complexes (many first-row divalent aqua and ammine species) serve as fast pre-equilibrium reservoirs in ligand-exchange-driven sensing and separation schemes. The quantitative CFAE machinery developed here also feeds forward directly into the Crystal Field Activation Energy treatment of reaction-rate periodic trends and into full mechanistic classification (D, $I_d$, $I_a$, A) covered under Ligand Substitution Mechanisms.

:::quiz[In the synthesis of cisplatin from [PtCl₄]²⁻ plus two equivalents of NH₃, why does the second NH₃ enter cis to the first, giving the therapeutically active isomer?]

• NH₃ has a much higher trans effect than Cl⁻, directing substitution trans to itself
• Cl⁻ has a higher trans effect than NH₃, so the second NH₃ substitutes trans to a remaining chloride ✓
• The reaction is purely statistical with no directing effect at all
• Square-planar substitution always gives the trans product preferentially

Ask the tutor: How does the labile/inert classification change when a complex has a strong π-donor ligand (like OH⁻ or F⁻) instead of a π-acceptor, and how does that interact with the trans effect series?

Substitution kinetics is fundamentally a statement about the path between two minima on a potential energy surface, not about the depth of either minimum. The reactant $\text{[ML}_6\text{]}$ and product $\text{[ML}_5\text{X]}$ may differ in energy by a large, favorable $\Delta G$, but the rate is controlled by $\Delta G^{\ddagger}$, the barrier to the transition state that sits between them. For octahedral $d$-block complexes, decades of kinetic study (isotopic water exchange, NMR line-broadening, stopped-flow spectrophotometry) converge on a consistent mechanistic picture: substitution is dissociative in character — bond-breaking to the leaving ligand is more advanced than bond-making to the incoming ligand at the transition state. This is denoted $\text{D}$ (fully dissociative, discrete 5-coordinate intermediate) or $\text{I}_d$ (dissociative interchange, concerted but asymmetric), as distinct from associative $\text{A}$ or $\text{I}_a$ pathways. Why should octahedral complexes prefer dissociative character? Geometrically, an octahedron is already a comfortable, closed-shell-like coordination environment; adding a seventh ligand to reach a 7-coordinate associative transition state is sterically and electronically costly for most first-row transition metals, whereas removing one ligand to pass through (or near) a 5-coordinate transition state is comparatively easy. The transition-state geometry is usually modeled as trigonal bipyramidal (TBP) or, less often, square pyramidal (SP), because these are the natural low-energy shapes for 5-coordination and interconvert easily (Berry pseudorotation). The key insight — due largely to Taube and formalized quantitatively by Basolo and Pearson — is that the d-orbitals reorganize on going from octahedral to trigonal-bipyramidal coordination, and this reorganization has an energetic cost or benefit that depends entirely on which orbitals are occupied. In an octahedral field the five d-orbitals split into $t_{2g}$ (lower, 3-fold) and $e_g$ (upper, 2-fold), separated by $\Delta_o$. In a trigonal-bipyramidal field the same five orbitals split differently — into a set of comparable but distinctly lower total splitting spread — because there are fewer, differently-arranged ligands exerting the crystal field. If a configuration places electrons preferentially in orbitals that were low-energy in the octahedron but become comparatively higher in the TBP transition state, energy must be spent to reach the transition state: this energy is the crystal field activation energy.

The middle panel of the picture — orbital correlation between octahedral and TBP — is the mechanistic bridge between a static bonding diagram and a kinetic rate. It converts a bonding question ("where are the electrons?") into a kinetics question ("how much does it cost to get to the transition state?"). This is exactly analogous to how, in organic mechanism, orbital symmetry constraints (Woodward–Hoffmann) convert electron count into an allowed/forbidden verdict — here, d-electron count converts into a fast/slow verdict. A second, complementary picture concerns square-planar $d^8$ complexes (Pt(II), Pd(II), Ni(II), Au(III)), which substitute by an associative mechanism because the vacant axial positions above and below the square plane offer a low-energy site for the incoming ligand to attack, forming a genuine 5-coordinate trigonal-bipyramidal intermediate with 18 electrons (16e + 2e from the incoming ligand). Kinetics here are second order (rate depends on [substrate][incoming ligand]) and dominated by the trans effect — the ability of a ligand to labilize the position directly opposite itself, controlled by both σ-donation (ground-state weakening of the trans M–X bond) and π-acceptance (transition-state stabilization by spreading incoming electron density into empty ligand π* orbitals).

Together these two pictures — dissociative-character substitution at octahedral centers governed by CFAE, and associative substitution at square-planar centers governed by the trans effect — account for essentially all the qualitative rate trends encountered in coordination chemistry courses.

The central quantitative construct is the crystal field activation energy:

where:

• $\text{LFSE}_{\text{octahedron}}$ is the ligand field stabilization energy of the ground-state $d^n$ configuration in $O_h$ symmetry, in units of $\Delta_o$ (with $t_{2g}$ at $-0.4\Delta_o$ and $e_g$ at $+0.6\Delta_o$),
• $\text{LFSE}_{\text{TS geometry}}$ is the LFSE of the same $d^n$ configuration re-distributed among the orbital energies appropriate to the 5-coordinate transition state (commonly modeled as square pyramidal, SP, or trigonal bipyramidal, TBP), also in units of the same $\Delta_o$ scale (converted via standard tabulated fractional splittings),
• CFAE is necessarily $\ge 0$ (the electrons are placed optimally in each geometry separately, and the TS spread is intrinsically less favorable in the worst cases), and represents an extra kinetic barrier on top of whatever intrinsic bond-breaking energy is required.

Reading: a large positive CFAE means the complex must sacrifice substantial stabilization energy just to distort toward the transition state, over and above the cost of stretching the M–X bond itself — this is an additional, purely electronic contribution to $\Delta G^{\ddagger}$, and configurations with CFAE near zero are labile while those with CFAE of several tenths of $\Delta_o$ or more are inert. The companion relationship is the standard Eyring/transition-state theory link between rate and barrier height, which converts CFAE (an energy) into an observable rate constant:

where:

• $k_B, h, R$ are Boltzmann's constant, Planck's constant, and the gas constant,
• $\Delta G^{\ddagger}_{\text{intrinsic}}$ is the barrier that would exist even for a $d^0$ or $d^{10}$ ion with no CFAE penalty (electrostatics, size, solvent reorganization),
• CFAE adds directly to the exponent's numerator, so even a modest CFAE of a few $\Delta_o$-percent, multiplied through $\Delta_o$ values of order $100$–$300\ \text{kJ mol}^{-1}$, produces an enormous rate suppression because the barrier sits inside an exponential.

Reading: because $\Delta G^{\ddagger}$ appears in an exponential, differences of only $10$–$40\ \text{kJ mol}^{-1}$ in CFAE between two configurations translate into rate ratios of $10^{2}$–$10^{7}$, which is exactly the observed spread between labile and inert octahedral complexes.

Example 1 (foundational): rank the water-exchange rates of $\text{[V(H}_2\text{O)}_6\text{]}^{2+}$ ($d^3$), $\text{[Fe(H}_2\text{O)}_6\text{]}^{2+}$ (hs $d^6$), and $\text{[Ni(H}_2\text{O)}_6\text{]}^{2+}$ ($d^8$). Compute octahedral LFSE for each (weak-field aqua ions, high spin throughout since $\text{H}_2\text{O}$ is a weak/intermediate field ligand): $d^3$ ($\text{V}^{2+}$): $t_{2g}^3 e_g^0$, $\text{LFSE} = 3(-0.4\Delta_o) = -1.2\,\Delta_o$. This is the maximum possible LFSE for any $d^n$ configuration without pairing, and — critically — the transition state (SP or TBP) offers no comparably favorable orbital arrangement for exactly half-filling $t_{2g}$-like orbitals with one electron each; removing a ligand can only raise the energy of at least one occupied orbital. CFAE for $d^3$ is large (empirically among the highest of any high-spin configuration), and indeed $\text{V(H}_2\text{O)}_6^{2+}$ is markedly more inert than its high-spin $d^n$ neighbors, with water-exchange rate constant $k_{ex} \approx 90\ \text{s}^{-1}$ at 25°C — slow for a "labile-category" divalent ion. hs $d^6$ ($\text{Fe}^{2+}$): $t_{2g}^4 e_g^2$, $\text{LFSE} = 4(-0.4) + 2(0.6) = -1.6 + 1.2 = -0.4\,\Delta_o$. Modest stabilization, easily redistributed with little penalty in a 5-coordinate TS. CFAE is small (a few percent of $\Delta_o$), and indeed $\text{Fe(H}_2\text{O)}_6^{2+}$ is one of the most labile of all divalent aqua ions, $k_{ex} \approx 4.4\times10^6\ \text{s}^{-1}$. $d^8$ ($\text{Ni}^{2+}$): $t_{2g}^6 e_g^2$, $\text{LFSE} = 6(-0.4)+2(0.6) = -2.4+1.2 = -1.2\,\Delta_o$. Numerically identical LFSE to $d^3$, yet in the transition state, $d^8$ can accommodate its two $e_g$ electrons into the lower-lying SP/TBP orbitals fairly efficiently once one axial ligand departs (there is no forced electron sitting in the specific orbital that is destabilized on ligand removal, because both antibonding-derived electrons can slide into orbitals that remain comparatively stabilized). The CFAE for $d^8$ is smaller than for $d^3$, and experimentally $\text{Ni(H}_2\text{O)}_6^{2+}$, $k_{ex} \approx 3.2\times10^4\ \text{s}^{-1}$, is faster than $\text{V}^{2+}$ but slower than high-spin $\text{Fe}^{2+}$. Ranking by rate: $\text{Fe}^{2+}(hs\ d^6) > \text{Ni}^{2+}(d^8) > \text{V}^{2+}(d^3)$, spanning nearly 5 orders of magnitude in $k_{ex}$ despite all three being formally "labile" divalent first-row aqua ions — illustrating that even within the labile regime, CFAE still modulates rate systematically, it just never becomes large enough to push any of the three into true inertness. As a limiting-case check: setting the ligand field to the weak-field limit ($\Delta_o \to 0$, e.g. an extremely weak donor) drives CFAE $\to 0$ for every configuration, and all three ions would be predicted (correctly) to converge toward comparably fast, purely electrostatically-controlled exchange rates — CFAE differences only matter once $\Delta_o$ is large enough to be chemically meaningful.

Example 2 (applied): why does cisplatin survive in blood plasma but hydrolyze rapidly inside a cell? Cisplatin, $\text{cis-[PtCl}_2\text{(NH}_3)_2]$, is square-planar $d^8$ Pt(II) — kinetically the substitution rate is governed by chloride concentration via the associative mechanism, since $\text{Cl}^-$ (or $\text{H}_2\text{O}$) attacks the vacant axial site. In blood plasma, $[\text{Cl}^-] \approx 100\ \text{mM}$, which by Le Chatelier-type mass action strongly suppresses the forward hydrolysis (aquation) equilibrium

because the reverse reaction (chloride re-attacking to displace water) is fast and chloride is abundant — this is a thermodynamic/equilibrium suppression of the product concentration, not a kinetic rate suppression, but it has the practical effect that cisplatin persists largely intact in the bloodstream. Once cisplatin crosses into the cytoplasm, where $[\text{Cl}^-] \approx 4\ \text{mM}$ (roughly 25-fold lower), the equilibrium in the forward (aquation) direction is no longer suppressed, and the resulting aqua/hydroxo species — now bearing a good, kinetically labile leaving group ($\text{H}_2\text{O}$, itself weakly trans-directing) — reacts with the N7 position of guanine in DNA to form the crosslinks responsible for cisplatin's cytotoxicity. The intrinsic rate constant for associative substitution barely changes between blood and cytoplasm (both are governed by the same Pt(II) $d^8$ associative mechanism and comparable pH); what changes is which side of the chloride/aquation equilibrium is thermodynamically populated, which is precisely why the labile-vs-inert kinetic classification must be kept conceptually separate from thermodynamic stability — cisplatin is kinetically substitution-labile (associative $d^8$ Pt(II) reacts on a timescale of hours, not the instantaneous timescale of true labile ions, but far faster than genuinely inert Co(III) or Cr(III)) throughout, but its equilibrium position is what the chloride gradient controls.

Example 3 (exam/competition level): rank the following six octahedral species by expected inertness, and justify with explicit CFAE reasoning:

Work through each: (a) $\text{Cr}^{3+}$, $d^3$: $\text{LFSE}(O_h) = -1.2\Delta_o$ (three unpaired electrons, exactly half-filling $t_{2g}$). As in Example 1, this is a maximally-stabilized, symmetric $t_{2g}^3$ configuration that cannot be reproduced in any 5-coordinate geometry without promoting at least one electron to a higher orbital — CFAE is large. Combined with a sizeable $\Delta_o$ for a 3+ ion ($\Delta_o \approx 208\ \text{kJ mol}^{-1}$ from the UV-vis spectrum of $\text{[Cr(NH}_3)_6]^{3+}$), $\text{Cr(III)}$ complexes are proverbially inert — ligand-exchange half-lives of days. (b) $\text{Co}^{3+}$, ls $d^6$: as derived above, CFAE $\approx 0.74\Delta_o$, the single largest CFAE of any configuration, combined with a very large $\Delta_o$ (~265–275 kJ/mol for strong-field N-donor ligands) — this is the textbook most-inert octahedral configuration; substitution half-lives of years are typical, and virtually all classic ligand-substitution mechanistic studies (Taube, Ingold, Basolo, Pearson) were performed on $\text{Co(NH}_3)_5\text{X}$ complexes precisely because the reaction is slow enough to follow conveniently with a stopwatch or simple UV-vis kinetics. (c) $\text{Mn}^{2+}$, hs $d^5$: $t_{2g}^3e_g^2$, $\text{LFSE}(O_h) = 3(-0.4)+2(0.6) = -1.2+1.2 = 0$. High-spin $d^5$ is the unique configuration with zero octahedral LFSE (it is a symmetric half-filled shell, spherically averaged in the ligand field sense) — there is nothing to lose on distortion, so CFAE $= 0$ regardless of TS geometry. This predicts $\text{Mn(H}_2\text{O)}_6^{2+}$ should be exceptionally labile, and indeed it has one of the fastest measured water-exchange rate constants of any transition-metal aqua ion, $k_{ex} \approx 2.1\times10^7\ \text{s}^{-1}$, comparable to or faster than the labile main-group/lanthanide ions. (d) $\text{Cu}^{2+}$, $d^9$: strongly Jahn–Teller distorted even in the ground state (tetragonal elongation), which means the complex is already partway toward a lower-symmetry, lower-coordination-number-like geometry; the axial Cu–O bonds are already long and weak, so very little additional CFAE is required to lose an axial ligand entirely. $\text{Cu(H}_2\text{O)}_6^{2+}$ is exceptionally labile, among the fastest-exchanging divalent ions known ($k_{ex} \approx 4\times10^9\ \text{s}^{-1}$, close to the diffusion limit), precisely because Jahn–Teller distortion has already "pre-paid" much of the reorganization energy in the ground state. (e) $\text{Zn}^{2+}$, $d^{10}$: fully filled shell, $\text{LFSE}(O_h) = 0$ identically in any geometry (a filled shell is spherically symmetric and contributes no orientational stabilization at all) — CFAE $=0$ exactly, by the same limiting-case logic as $d^0$. $\text{Zn(H}_2\text{O)}_6^{2+}$ is highly labile, $k_{ex} \approx 3\times10^7\ \text{s}^{-1}$. (f) $\text{[Ni(CN)}_4]^{2-}$, square-planar low-spin $d^8$: this is governed by the associative mechanism and trans-effect logic (Derivation 2), not the octahedral CFAE framework — $\text{CN}^-$ is an extremely strong σ-donor/π-acceptor, giving both strong trans-directing character and a robust 18-electron TBP transition state pathway. Despite the strong Ni–CN bonds, ligand exchange in $\text{[Ni(CN)}_4]^{2-}$ is in fact quite fast on the NMR timescale (labile by associative substitution) — a useful reminder that the dissociative octahedral CFAE ranking (a)–(e) does not directly transfer to square-planar systems, which follow their own associative rate law.

Final ranking, most inert → most labile: