Inorganic Chemistry · Part 7 of 9

Stability & Thermodynamics

Coordination Chemistry, Part 7 · 9 sections · about 26,529 words · CSIR-NET Chemical Sciences, GATE Chemistry & IIT-JAM

Every complex in this book exists only to the extent that an equilibrium constant allows it to. This part is about those constants: how they are defined, why they almost always fall as more ligands are added, how they are measured, and what controls their size. It treats the chelate and macrocyclic effects quantitatively rather than as slogans — separating the entropy term from the enthalpy term and showing where the usual textbook explanation is incomplete — and it ends in the analytical laboratory, with the conditional constants that make EDTA titrations work. Two layers on every section: a slow, hand-held beginner path and a research-grade advanced/reference path.

The 9 sections in Part 7

  • 1Stepwise constants Kn, overall constants βn, and moving between them Free below
  • 2Why K₁ > K₂ > K₃ … — and the instructive cases where it is not
  • 3Measuring stability constants — pH titration, spectrophotometry and Job’s method
  • 4Metal-ion factors — charge, size and the Irving–Williams series
  • 5Ligand factors — basicity, ring size, sterics and π-bonding
  • 6Hard and soft acids and bases, applied properly
  • 7The chelate effect quantified
  • 8The macrocyclic and cryptate effects, and preorganisation
  • 9Conditional constants, masking and the EDTA titration

Stepwise constants Kn, overall constants βn, and moving between them

Free extract

Section J.1 of Part 7, reproduced in full from the book — figures and all. No sign-in, no paywall on this section.

Beginner layer

What the two kinds of constant mean, how they are related, and the arithmetic of converting one into the other — done slowly, because everything later in this part depends on getting it automatic.

Put a metal ion in water and add a ligand. The ligands do not all arrive at once. They arrive one at a time, each in its own equilibrium, and each of those equilibria has its own constant. Consider a metal that ends up with four ligands:

M + L ⇌ ML       K₁ = [ML] / ([M][L])
ML + L ⇌ ML₂     K₂ = [ML₂] / ([ML][L])
ML₂ + L ⇌ ML₃   K₃ = [ML₃] / ([ML₂][L])
ML₃ + L ⇌ ML₄   K₄ = [ML₄] / ([ML₃][L])
Stepwise formation constant Kn: the equilibrium constant for adding the nth ligand to the complex that already carries (n − 1) of them. It describes one step only.

Now write the same chemistry a second way — as if each complex were assembled from the bare metal ion and free ligands in a single act:

M + L ⇌ ML         β₁ = [ML] / ([M][L])
M + 2L ⇌ ML₂      β₂ = [ML₂] / ([M][L]²)
M + 3L ⇌ ML₃      β₃ = [ML₃] / ([M][L]³)
M + 4L ⇌ ML₄      β₄ = [ML₄] / ([M][L]⁴)
Overall (cumulative) formation constant βn: the equilibrium constant for forming MLn directly from the free metal ion and n free ligands. Also called the cumulative or gross constant, and sometimes written β1n.

The two sets are not independent. Add the first two stepwise equilibria together — and remember that when you add equilibria you multiply their constants:

M + L ⇌ ML    (K₁)
ML + L ⇌ ML₂  (K₂)
―――――――――――――――――――
M + 2L ⇌ ML₂  (K₁K₂ = β₂)

You can see it in the algebra directly: multiply the two expressions and the [ML] cancels.

K₁K₂ = [ML][M][L] × [ML₂][ML][L] = [ML₂][M][L]² = β₂
βn = K₁ K₂ K₃ … Kn = ∏i=1n Ki
log βn = log K₁ + log K₂ + … + log Kn = ∑i=1n log Ki

And going the other way, each stepwise constant is the ratio of two consecutive overall constants:

Kn = βn / βn−1      log Kn = log βn − log βn−1   (with β₀ ≡ 1, log β₀ = 0)
Note: Multiplication in K, addition in log K. That is the whole of the conversion. Because the tabulated quantity is almost always log, in practice you are adding and subtracting a short column of numbers. If you find yourself multiplying logarithms, stop.

Two conventions you must know before you read any table

(a) The coordinated water is not written. The honest equation for the first step of ammine formation on nickel is

[Ni(H₂O)₆]²⁺ + NH₃ ⇌ [Ni(NH₃)(H₂O)₅]²⁺ + H₂O

Every table in every book compresses this to Ni²⁺ + NH₃ ⇌ [Ni(NH₃)]²⁺. The reason is that water is the solvent: it is present at about 55.5 mol L⁻¹ and its activity is taken as 1, so it drops out of the equilibrium expression. This is a convention, it is universal, and it is harmless — until J.8, where the whole argument about the chelate effect turns on exactly this choice. Remember that it was made.

(b) The constants are concentration constants at a stated ionic strength. Strictly, equilibrium constants are written in activities. In practice, stability constants are measured in a swamping inert electrolyte — typically 0.1 M or 1.0 M KNO₃, NaClO₄ or KCl — so that activity coefficients stay constant while the reagent concentrations are varied. The number that comes out is a concentration constant valid at that ionic strength, and it will differ, usually by a few tenths of a log unit, from the thermodynamic constant at I = 0.

A worked table: copper(II) and ammonia

StepEquilibriumlog KnCumulativelog βn
1Cu²⁺ + NH₃ ⇌ [Cu(NH₃)]²⁺4.15β₁ = K₁4.15
2[Cu(NH₃)]²⁺ + NH₃ ⇌ [Cu(NH₃)₂]²⁺3.50β₂ = K₁K₂7.65
3[Cu(NH₃)₂]²⁺ + NH₃ ⇌ [Cu(NH₃)₃]²⁺2.89β₃ = K₁K₂K₃10.54
4[Cu(NH₃)₃]²⁺ + NH₃ ⇌ [Cu(NH₃)₄]²⁺2.13β₄ = K₁K₂K₃K₄12.67
Cu(II)/NH₃ at 25 °C. Read the last column as a running total of the third: 4.15, then 4.15 + 3.50 = 7.65, then + 2.89 = 10.54, then + 2.13 = 12.67. That is the entire conversion. Note also that copper stops at four ammines in aqueous solution — the fifth and sixth ammonias bind to the long, weak axial positions of a Jahn–Teller-distorted d⁹ ion and cannot compete with water; the pentammine forms only in liquid ammonia.

Given log K₁ = 4.15, log K₂ = 3.50, log K₃ = 2.89 and log K₄ = 2.13 for Cu(II)/NH₃, find β₂ and β₄ as numbers, and state their units. Easy

Step 1 — add the logs, do not multiply them. log β₂ = log K₁ + log K₂ = 4.15 + 3.50 = 7.65.
Step 2 — antilog. β₂ = 107.65 = 4.5 × 10⁷.
Step 3 — repeat for β₄. log β₄ = 4.15 + 3.50 + 2.89 + 2.13 = 12.67, so β₄ = 1012.67 = 4.7 × 10¹².
Step 4 — units. β₂ = [ML₂]/([M][L]²), which is concentration⁻², i.e. M⁻² or dm⁶ mol⁻². In general βn has units M⁻ⁿ and Kn has units M⁻¹. This is why you may never compare β₂ with β₄ and call one “bigger” — they are not the same kind of quantity. Only constants with the same units may be compared directly.

For Ag(I)/NH₃ at 20 °C and I = 0.1 M, log β₂ = 7.22 and log K₁ = 3.31. Find log K₂, and comment. Easy

Step 1. log K₂ = log β₂ − log β₁ and β₁ = K₁, so log K₂ = 7.22 − 3.31 = 3.91.
Step 2 — comment. Here K₂ is larger than K₁, which reverses the usual trend. Silver(I) is a d¹⁰ ion that prefers linear two-coordination: the second ammonia completes the favoured geometry, and the accompanying reorganisation (loss of the remaining weakly held waters, rehybridisation towards sp) pays back more than the statistical and electrostatic penalties cost. J.2 collects the cases where this happens.
Step 3 — a check on your instinct. If you expected K₂ < K₁ and got a contradiction, that is the right reflex: the general rule is a strong one, and an exception is a signal that something structural changes during the step.

⚠ Common mistakes & exam traps

  • β₂ is not K₂. β₂ = K₁K₂. If a question gives you “β₂ = 10⁷·⁶₅” and asks for the constant of the second step, you must subtract log K₁ first. This single confusion accounts for more lost marks in this topic than anything else.
  • Do not compare βm with βn for m ≠ n. They have different units, so “β₆ = 10⁸·₇ is smaller than β₃ = 10¹⁸·³” is not a meaningful sentence. To compare two ligand systems, write the exchange reaction between them; its constant is dimensionless.
  • “Stability constant” unqualified usually means the formation constant. But the older literature also uses instability constant or dissociation constant, which is its reciprocal. If a value looks absurdly small (10⁻¹² for a complex that obviously forms), you are looking at a dissociation constant.
  • Water does not appear in the expression, but it did not go away. Every formation constant in water is really a substitution constant, ligand for water. That is not pedantry — it is the reason the chelate effect has a standard-state problem (J.8).
  • A large K does not mean fast. Thermodynamic stability and kinetic inertness are independent properties. Part 8 has the vocabulary; here, simply never use “stable” to mean “slow to react”.
Easy
For Cd(II)/CN⁻, log K₁ = 5.48, log K₂ = 5.12, log K₃ = 4.63 and log K₄ = 3.55. Calculate log β₄, and find the ratio [Cd(CN)₄²⁻]/[Cd²⁺] when the free cyanide concentration is held at 1.0 × 10⁻³ M.
Show solution
log β₄ = 5.48 + 5.12 + 4.63 + 3.55 = 18.78.
The ratio asked for is β₄[CN⁻]⁴ because β₄ = [Cd(CN)₄²⁻]/([Cd²⁺][CN⁻]⁴). So log(ratio) = log β₄ + 4 log[CN⁻] = 18.78 + 4(−3.00) = 18.78 − 12.00 = 6.78, i.e. a ratio of 6.0 × 10⁶. Even at millimolar free cyanide the tetracyano complex outnumbers free Cd²⁺ by six million to one — which is why cyanide is the standard masking agent for cadmium (J.9).
Easy
A compilation lists log β₁ = 4.34, log β₂ = 7.94, log β₃ = 10.80 for a 1:3 system. Extract all three stepwise constants and comment on the pattern.
Show solution
Successive differences: log K₁ = log β₁ = 4.34; log K₂ = 7.94 − 4.34 = 3.60; log K₃ = 10.80 − 7.94 = 2.86. The pattern is the normal one — a steady fall, here of about 0.74 then 0.74 log units. A fall of this size (roughly a factor of 5 per step) is larger than statistics alone predicts for a three-step system, so charge and steric effects are contributing as well (J.2).
Hard
Show that for the exchange reaction [ML₆] + 3 L′′ ⇌ [M(L′′)₃] + 6 L, where L is monodentate and L′′ is bidentate, the equilibrium constant is β₃(L′′)/β₆(L) and that this quantity is dimensionless.
Show solution
Write both formation equilibria from the common starting point, the free metal ion:
M + 6 L ⇌ ML₆ with β₆(L), and M + 3 L′′ ⇌ M(L′′)₃ with β₃(L′′).
Subtract the first from the second (i.e. reverse the first and add): ML₆ + 3 L′′ ⇌ M(L′′)₃ + 6 L, Kex = β₃(L′′) / β₆(L).
Units: β₃ carries M⁻³, β₆ carries M⁻⁶; the quotient carries M⁻³⁺⁶ = M³. That is not dimensionless — and this is exactly the point. Check the reaction itself: 4 particles on the left, 7 on the right, so Δn = +3 and the constant must carry M³. The claim in the question is false as stated. It becomes true only for a particle-conserving exchange. This is the seed of the standard-state argument in J.8, and being able to spot a units-inconsistent comparison is worth more than memorising any table.
Advanced / reference layer

Activity constants, mixed constants, the species-distribution picture, and the free-energy bookkeeping that connects all of it to thermodynamics.

From constant to free energy

Every stability constant is a disguised free energy, and switching between the two languages is what allows the enthalpy/entropy dissection of J.7 to happen at all.

ΔG° = −RT ln K = −2.303 RT log K

At 298.15 K, 2.303 RT = 2.303 × 8.314 × 298.15 = 5708 J mol⁻¹, so:

ΔG° / kJ mol⁻¹ = −5.708 × log K   (at 25 °C)
Note: Commit 5.708 kJ mol−1 per log unit at 25 °C to memory. One log unit is about 5.7 kJ mol⁻¹; ten log units is about 57 kJ mol⁻¹, comparable with a weak covalent bond. It converts every stability statement into an energy statement instantly, and examiners like asking for ΔG° from log β.

Because free energies add when reactions add, and log K adds when reactions add, the two bookkeeping systems are the same system. The relation log βn = ∑ log Ki is nothing more than ΔG°overall = ∑ ΔG°step, which is Hess’s law.

Thermodynamic, stoichiometric and mixed constants

Three kinds of constant appear in the literature, and a serious candidate should be able to tell which one a table is quoting.

KindWritten inHow obtainedSymbol you may see
Thermodynamic (activity) constantactivities; valid at I = 0measured at several ionic strengths and extrapolated to I = 0, or corrected with an activity-coefficient model (Davies, SIT, Pitzer)K°, TK
Stoichiometric (concentration) constantconcentrations, at a stated constant ionic strengthmeasured in a swamping inert electrolyte, e.g. 0.1 M KNO₃K, Kc
Mixed (Brønsted) constantconcentrations for everything except H⁺, which enters as activityarises automatically when a glass electrode (which reads aH) is combined with analytical concentrations of everything elseK′, mixedK
The mixed constant is the one you get without trying, because pH meters are calibrated in activity while burettes deliver concentration. Most published potentiometric stability constants are mixed constants unless the authors say otherwise. The differences are typically 0.1–0.4 log units — irrelevant for exam arithmetic, decisive for speciation modelling.

The ionic-strength dependence itself is not mysterious. For a reaction between ions, the Debye–Hückel treatment gives the leading term in the activity-coefficient correction as proportional to the change in the sum of squared charges:

log K(I) = log K° + A Δ(z²) √I / (1 + √I) + (linear term in I)

The practical consequence: constants for reactions between highly charged ions are the most ionic-strength sensitive. Ca²⁺ + Y⁴⁻ ⇌ CaY²⁻ has Δ(z²) = 4 − (4 + 16) = −16, so log K for EDTA complexes moves substantially between I = 0 and I = 0.1 M. This is why the analytical literature quotes EDTA constants at a stated ionic strength, usually 0.1 M.

Species distribution — what the constants actually predict

A set of β values is a complete description of the system: given the free-ligand concentration, every species concentration follows. The fraction of total metal present as MLn is

αn = [MLn] / CM = βn[L]n / (1 + ∑i=1N βi[L]i)

The denominator — 1 + β₁[L] + β₂[L]² + … — recurs constantly and deserves its own name. In the analytical literature it is written αM(L) and it is the side-reaction coefficient that J.9 is built on. Meeting it here, as a normalising denominator, makes it much less mysterious when it reappears as a masking correction.

αM(L) = CM/[M] = 1 + β₁[L] + β₂[L]² + … + βN[L]N
Note: Note the structure. This is identical in form to the denominator that appears in polyprotic acid–base chemistry, 1 + [H]/Ka + [H]²/Ka1Ka2 + …, and for exactly the same reason: in both cases you are asking what fraction of a total analytical concentration sits in one particular form. If you can do α calculations for a triprotic acid, you can do them for a metal — it is the same algebra with M in place of A⁻ and L in place of H⁺.

Polynuclear and protonated species — when MLn is not enough

Real systems are often not a clean ML₁…MLN series. Two extra families appear frequently enough that their notation should be recognised:

  • Polynuclear complexes MmLn, whose constants carry two subscripts: βmn = [MmLn]/([M]m[L]n). The commonest examples are hydroxo-bridged species such as [Fe₂(OH)₂]⁴⁺ and [Cu₂(OH)₂]²⁺, which is why iron(III) speciation in water is so much harder than a simple stepwise series suggests.
  • Protonated and hydroxo complexes MHjLn and M(OH)kLn, written with a third index, e.g. β1 1 1 for MHL. These matter enormously in EDTA chemistry, where MHY⁻ and MOHY³⁻ species exist at the extremes of pH.

The general convention, which is worth recognising even if you never use it, writes the cumulative constant of the species MpLqHr as βpqr, with negative r denoting hydroxo species (a deprotonated water). So β1 0 −1 is the constant for M + H₂O ⇌ MOH + H⁺, the first hydrolysis constant.

Med
For a metal forming ML, ML₂ and ML₃ with log β₁ = 5.0, log β₂ = 9.0 and log β₃ = 12.0, calculate the fraction of total metal present as ML₂ when [L] = 1.0 × 10⁻⁴ M.
Show solution
Compute each term βi[L]i in logarithms first.
β₁[L]: 5.0 + (−4.0) = 1.0 → 101.0 = 10
β₂[L]²: 9.0 + (−8.0) = 1.0 → 101.0 = 10
β₃[L]³: 12.0 + (−12.0) = 0.0 → 100.0 = 1
Denominator = 1 + 10 + 10 + 1 = 22. So α₂ = 10/22 = 0.45, i.e. 45% of the metal is ML₂. (Free metal is 1/22 = 4.5%, ML is 45%, ML₃ is 4.5%.) Note the shortcut: adding log β and n log[L] keeps every quantity in comfortable range and avoids calculator overflow.
Med
At 25 °C, log β₄ for [Zn(NH₃)₄]²⁺ is 9.06. Calculate ΔG° for the overall formation, and for the average per-ligand step.
Show solution
ΔG° = −5.708 × 9.06 = −51.7 kJ mol⁻¹. Per ligand, on average, −51.7/4 = −12.9 kJ mol⁻¹ — about the size of a strong hydrogen bond, which is a useful calibration: individual M–NH₃ interactions in water are not strong bonds in the covalent sense, because most of the intrinsic bond energy is paid back to the water molecules being displaced. It is the accumulation over four steps, plus favourable entropy, that produces a large β.
Hard
A student measures a formation constant potentiometrically at I = 0.10 M and compares it with a tabulated value at I = 0; they differ by 0.6 log units and the student concludes the measurement is wrong. Is it?
Show solution
Not necessarily — and for a reaction between multiply charged ions a discrepancy of this size is expected. Two systematic differences are in play. First, the ionic-strength difference: for M²⁺ + L²⁻ ⇌ ML, Δ(z²) = 0 − (4 + 4) = −8, and a Davies correction at I = 0.1 M gives a shift of several tenths of a log unit in the direction of greater apparent stability at higher I (the reactant ions are stabilised less than a naive treatment suggests). Second, a glass-electrode measurement yields a mixed constant, not a stoichiometric one. The correct response is to state the medium with the value, not to discard the measurement. Comparing constants from different media without correction is a genuine error; obtaining different numbers in different media is not.

Read the rest of Part 7

The remaining 8 sections of this part — Why K₁ > K₂ > K₃ … — and the instructive cases where it is not, Measuring stability constants — pH titration, spectrophotometry and Job’s method, Metal-ion factors — charge, size and the… — and all nine parts of Coordination Chemistry are part of ChemVidya Full Access, along with the other books, 55 Study Notes and 6,000+ practice questions.

See plans Open in the app

Continue through Coordination Chemistry

Related Inorganic Study Notes