Octahedral, Tetrahedral, and Square-Planar Geometries

Square-planar platinum(II) chemistry underlies an entire generation of chemotherapy drugs — cisplatin, carboplatin, and oxaliplatin all rely on the cis square-planar arrangement to form the DNA crosslinks that block cancer cell replication; the trans isomer of any of these drugs is pharmacologically inert or actively different in mechanism, a direct clinical consequence of the geometric analysis above. Square-planar rhodium(I) and iridium(I) complexes — Wilkinson's catalyst $RhCl(PPh_3)_3$ and Vaska's complex $IrCl(CO)(PPh_3)_2$ — are the working platforms for homogeneous hydrogenation and oxidative-addition catalysis throughout the fine-chemical and pharmaceutical industry, exploiting exactly the d⁸ square-planar-to-d⁶ octahedral cycle worked through in Example 3. Square-planar nickel and palladium complexes are central to cross-coupling catalysis (Suzuki, Negishi, Stille couplings), where the Pd(0)/Pd(II) cycle repeatedly interconverts a low-coordinate species with square-planar Pd(II) intermediates during oxidative addition and reductive elimination steps that build carbon–carbon bonds in the synthesis of pharmaceuticals and agrochemicals. Tetrahedral coordination geometry governs the active sites of numerous metalloenzymes, including carbonic anhydrase (tetrahedral Zn²⁺, d¹⁰, coordinated by three histidines and a water/hydroxide, whose Lewis acidity activates water for CO₂ hydration) and many zinc-finger DNA-binding domains, where a tetrahedral Zn²⁺ center (again d¹⁰, hence geometrically unconstrained by CFSE and free to adopt the sterically preferred tetrahedral shape) organizes a protein fold that reads specific DNA sequences. Octahedral geometry is the dominant motif in bioinorganic electron-transfer proteins (the heme iron of cytochromes toggles between five- and six-coordinate, low-spin octahedral states) and in the vast majority of solid-state transition-metal oxides and coordination polymers, where octahedral $MO_6$ or $MX_6$ units are the basic structural building block of everything from rock-salt-type oxides to perovskite solar-cell absorbers and lithium-ion battery cathode materials such as $LiCoO_2$, whose layered structure is built from edge-sharing octahedral $CoO_6$ units.

:::quiz[Which geometry is predicted for [ZnCl4]2- and why?]

• Square planar, because Zn2+ is d8
• Tetrahedral, because Zn2+ is d10 with no CFSE to gain from square planar ✓
• Octahedral, because coordination number is always 6 for Zn2+
• Trigonal bipyramidal, because Cl- is a weak-field ligand

Ask the tutor: How would the CFSE comparison change for a d⁶ or d⁷ ion choosing between tetrahedral and square-planar coordination — would square planar still win?

Start with the electrostatic baseline. If a metal cation is treated as a point positive charge and its ligands as point negative charges (or dipoles) repelling one another on the surface of a sphere, the optimal arrangement for a given coordination number is the one that maximizes the smallest ligand-ligand angle — exactly the logic of VSEPR, extended from lone-pair/bond-pair repulsion in main-group chemistry to ligand–ligand repulsion in coordination compounds. For six points on a sphere, the electrostatically optimal arrangement is the octahedron (all ligand-ligand angles are either 90° or 180°, and every ligand is symmetry-equivalent). For four points, the optimum is the tetrahedron (all six ligand-ligand angles equal to 109.47°). Square-planar is a worse electrostatic solution for four ligands: it has four 90° contacts and two 180° contacts, which packs ligands more closely on average than the tetrahedron does. Point-charge repulsion alone therefore predicts that four-coordinate complexes should always be tetrahedral — and for d⁰, d¹⁰, and high-spin d⁵ metal ions (where the d-subshell contributes no net directional preference), this prediction holds extremely well: TiCl₄, MnCl₄²⁻, and Zn(NH₃)₄²⁺ are all tetrahedral. The picture changes completely once the metal has a partially filled d-shell with an asymmetric population — and here we anticipate crystal field theory only enough to explain the geometry choice, not to derive splitting diagrams in full. In an octahedral field, the five d orbitals split into a lower $t_{2g}$ set (three orbitals, $d_{xy}, d_{xz}, d_{yz}$) and an upper $e_g$ set (two orbitals, $d_{z^2}, d_{x^2-y^2}$). If a d⁸ ion such as Ni²⁺ sits in an octahedral field with strong-field ligands (CN⁻, or in practice for Pd²⁺/Pt²⁺ even weaker-field ligands, because 4d/5d orbitals are more diffuse and splitting is intrinsically larger), the eight d electrons fill $t_{2g}^6 e_{g}^2$, leaving one electron in each $e_g$ orbital. Removing the two ligands that lie along the $z$-axis — flattening the octahedron into a square plane — selectively destabilizes the $d_{z^2}$ orbital much less than it destabilizes $d_{x^2-y^2}$ is stabilized, because $d_{z^2}$ has no ligands left to repel it along $z$. The net effect is that the single electron which would have occupied $d_{x^2-y^2}$ can instead drop to a much lower energy, while the electron that was in $d_{z^2}$ barely changes energy or drops as well. The total electronic stabilization from this rearrangement outweighs the modest loss of ligand-field stabilization from removing two ligands, provided the field is strong enough (large ligand-field splitting parameter, big $\Delta_o$). This is why essentially all strong-field d⁸ complexes — Ni(CN)₄²⁻, PdCl₄²⁻, PtCl₄²⁻, AuCl₄⁻, RhCl(PPh₃)₃-derived Rh(I) species, Vaska's complex IrCl(CO)(PPh₃)₂ — are square planar rather than tetrahedral, even though a naive point-charge model says tetrahedral should win.

The third geometry requiring careful attention is the relationship between all three shapes and their point groups, which govern spectroscopic selection rules and vibrational analysis. A homoleptic octahedral complex $[MA_6]$ belongs to $O_h$; a homoleptic tetrahedral complex $[MA_4]$ belongs to $T_d$; a homoleptic square-planar complex $[MA_4]$ belongs to $D_{4h}$. These three point groups are genuinely distinct — $T_d$ has no inversion center and no $C_4$ axis through the metal, $O_h$ has both, and $D_{4h}$ has an inversion center and a $C_4$ axis but lacks the full octahedral symmetry (no $C_3$ axes relating all four ligand positions to two axial ones, since there are no axial ligands). This symmetry distinction is exactly what makes octahedral d–d transitions formally Laporte-forbidden while square-planar and tetrahedral complexes can show more intensity — a foreshadowing point for spectroscopy that is picked up properly under crystal field theory.

The idealized bond angles for each geometry are geometric facts, not fitted parameters, and they set the baseline against which real, distorted structures are measured.

where: $\theta_{\text{oct}}$ is the ligand–metal–ligand angle in a perfect octahedron; every ligand has four cis neighbors at 90° and one trans neighbor at 180°.

Reading: an ideal octahedron has exactly two distinct angles, and every one of the six vertices is symmetry-equivalent — there is no "special" ligand position, which is why $MA_6$ is always a single structure with no positional isomers.

where: $\theta_{\text{tet}}$ is the ligand–metal–ligand angle in a perfect tetrahedron, derived from the geometry of four points maximally separated on a sphere (equivalently, four vertices of a regular tetrahedron viewed from its centroid).

Reading: every angle in $MA_4$ tetrahedral is identical (109.47°), and all four positions are symmetry-equivalent, so — like the octahedron — there is only one possible arrangement for $MA_4$, $MA_3B$, or even $MA_2B_2$: no cis/trans isomerism is possible in a tetrahedron, because there is no "opposite" vertex.

where: $\theta_{\text{sp}}$ describes the square-planar case, geometrically a "flattened octahedron" with the two axial ligands removed; the four remaining ligands lie in a plane containing the metal.

Reading: square planar retains the cis/*trans distinction of the octahedron (since it is literally an octahedron with two ligands deleted), which is the structural root of cis-platin/trans*-platin isomerism — a direct link to Stereoisomerism: Geometric Isomers.

Example 1 (foundational): count the number of distinct ligand environments in $[Co(NH_3)_6]^{3+}$, $[NiCl_4]^{2-}$, and $[PtCl_4]^{2-}$, and state their point groups. $[Co(NH_3)_6]^{3+}$ is octahedral, $O_h$. All six NH₃ ligands sit at equivalent vertices — each has four cis neighbors at 90° and one trans partner at 180°. Since all six ligands are identical, there is exactly one ligand environment; the complex has no geometric isomers. $[NiCl_4]^{2-}$ is tetrahedral, $T_d$ (Ni²⁺ is d⁸, but Cl⁻ is a weak-field ligand, so $\Delta_o$ is small and the electrostatic/steric preference for tetrahedral wins; this is the textbook high-spin, paramagnetic, tetrahedral Ni²⁺ complex, confirmed by its magnetic moment near the spin-only value for two unpaired electrons, $\mu \approx 2.83\,\mu_B$, since tetrahedral d⁸ is $e^4t_2^4$ with two unpaired electrons in $t_2$). All four Cl⁻ ligands are equivalent by the $T_d$ symmetry — one ligand environment, all L–M–L angles 109.47°. $[PtCl_4]^{2-}$ is square planar, $D_{4h}$ (Pt²⁺ is d⁸, and the large $4d$–$5d$-derived $\Delta_o$ from the third-row metal makes square planar the strong-field winner even with only moderately strong-field Cl⁻). All four Cl⁻ ligands are equivalent by $D_{4h}$ symmetry (related by the $C_4$ axis) — one ligand environment, but now there are two distinct types of L–M–L angle present simultaneously in the molecule: four cis angles of 90° and two trans angles of 180°, exactly as in an octahedron, because square planar is literally an octahedron missing its two axial vertices. This example already shows the central lesson: coordination number alone (four, in both the second and third cases) does not fix the geometry — the d-electron count together with ligand field strength does.

Example 2 (applied): rationalize why cis-diamminedichloroplatinum(II) (cisplatin) and its trans isomer have different biological activity, using geometry alone. Cisplatin, $[PtCl_2(NH_3)_2]$, is square planar, $Pt^{2+}$ being d⁸ with strong-field-inducing behavior from the heavy third-row metal center (large $\Delta_o$ favors square planar regardless of whether the individual ligands, NH₃ and Cl⁻, are each only moderately strong-field). In a square-planar $MA_2B_2$ complex there are exactly two geometric arrangements: cis, where the two Cl⁻ ligands occupy adjacent (90°-related) positions, and trans, where they occupy opposite (180°-related) positions.

Biologically, the two chloride ligands are the leaving groups: they are slowly hydrolyzed inside the cell (low chloride concentration intracellularly shifts the equilibrium toward aquo complexes), and the resulting electrophilic platinum center then binds to the N7 positions of two adjacent guanine bases on DNA. Only the cis isomer geometrically permits the two vacated sites to bind two DNA bases close enough together to form the characteristic intrastrand 1,2-d(GpG) crosslink that kinks the DNA helix and blocks replication — the trans isomer's two leaving-group sites point in opposite directions (180° apart) and cannot reach two bases on the same strand in a productive geometry. This is a direct, load-bearing illustration of why geometric (cis/trans) isomerism, only possible because the underlying geometry is square planar rather than tetrahedral, has a real pharmacological consequence — precisely the topic taken further in Stereoisomerism: Geometric Isomers.

Example 3 (exam/competition-level): a d⁸ metal complex $[MCl_2(PR_3)_2]$ shows only one ³¹P NMR signal for its two phosphine ligands and a single set of Cl environments, is diamagnetic, and reacts rapidly with H₂ via oxidative addition to give a six-coordinate product. Determine the most likely geometry, the point group, and predict the number of ³¹P NMR signals in the oxidative-addition product, justifying every step. Step 1 — read off the electron count and coordination number. $[MCl_2(PR_3)_2]$ has coordination number 4. Diamagnetism strongly implicates a d⁸ metal center in a strong-field, low-spin state, since high-spin tetrahedral d⁸ ($e^4t_2^4$) is paramagnetic (two unpaired electrons) while square-planar d⁸ is diamagnetic (all orbitals fully paired except the empty, high-lying $d_{x^2-y^2}$). Since the complex is diamagnetic, it must be square planar, $D_{4h}$-type framework for the $MCl_2P_2$ core (though substituents on P lower the true symmetry, the coordination geometry around the metal is square planar). Step 2 — assign the isomer from the single ³¹P and single Cl environment. A square-planar $MA_2B_2$ complex has two possible arrangements. If the two phosphines were cis to each other and the two chlorides cis to each other (i.e., overall cis isomer), the two P atoms would be chemically equivalent to each other (related by the mirror plane), consistent with one ³¹P signal; the two Cl would likewise be equivalent. If instead the isomer were trans-P,P (phosphines opposite each other, chlorides opposite each other), the two P atoms are also equivalent by symmetry, and so are the two Cl. Both isomers give a single ³¹P signal and single Cl environment by symmetry alone — so ³¹P count cannot by itself distinguish cis from trans here; a full assignment would need $^2J_{P-P}$ coupling patterns or ³⁵Cl-sensitive methods (or, in practice, X-ray/IR ν(M–Cl) counting: cis isomers show two IR-active M–Cl stretches, trans isomers show one, by group theory on the $C_{2v}$ vs. $D_{2h}$-reduced local symmetry). This is itself worth noting as an exam trap: NMR equivalence does not always distinguish cis from trans isomers. Step 3 — track the oxidative addition. Oxidative addition of H₂ to a square-planar d⁸ ML₄ complex is the textbook mechanism for going from four-coordinate, 16-electron, d⁸ square planar to six-coordinate, 18-electron, d⁶ octahedral (this is precisely the reaction Vaska's complex, $IrCl(CO)(PPh_3)_2$, undergoes with H₂). The metal's oxidation state increases by two, and the H–H bond breaks heterolytically/concertedly to place one H⁻-equivalent ligand cis to the other, since H₂ oxidative addition is a concerted three-center process that necessarily delivers the two hydrides to adjacent (cis) coordination sites — they cannot end up trans, because the H–H bond approaches and adds across one face of the square plane. Step 4 — assign the product geometry and NMR count. The product is six-coordinate, octahedral ($O_h$-derived local symmetry, distorted by the different ligands), $M(H)_2Cl_2(PR_3)_2$. Given the starting cis-square-planar arrangement of the two phosphines (taking the more common outcome for bulky phosphines, which prefer to remain mutually trans across the oxidative addition to minimize steric clash — this is the well-established outcome for Vaska-type chemistry: the two phosphines remain mutually trans, while the newly added H atoms occupy two of the cis sites created by the oxidative addition, and the two Cl ligands end up cis to each other). In this most common product geometry, the two phosphorus atoms remain related by a mirror plane (the molecule retains a $C_{2v}$-type local symmetry with P–M–P = 180°), so they remain chemically equivalent and the product still shows a single ³¹P NMR signal, now with a characteristically different $^1J_{P-H}$ or $^2J_{P-H}$ coupling pattern to the two new hydride ligands, which is exactly the diagnostic NMR signature reported for such Vaska-type oxidative additions.