Tetrahedral Crystal Field Splitting

Tetrahedral crystal field splitting is not a historical curiosity confined to problem sets — it governs real, current chemistry across several domains.

Cobalt(II) colorimetry and qualitative analysis. The characteristic intense blue of $\text{CoCl}_4^{2-}$ (versus the pale pink of octahedral $\text{Co(H}_2\text{O)}_6^{2+}$) is a direct consequence of the smaller $\Delta_t$ pushing d–d transitions into the visible with much higher molar absorptivity (tetrahedral complexes lack a center of symmetry, so d–d transitions are only Laporte-forbidden in a weaker sense, giving more intense color than the centrosymmetric octahedral case). This is used practically in classic qualitative tests for water content (cobalt chloride paper/silica gel indicators change from blue, anhydrous/tetrahedral-like environment, to pink, hydrated/octahedral, upon exposure to moisture).

Zinc and cadmium metalloenzyme active sites. Zinc(II), $d^{10}$, has zero crystal field stabilization energy in any geometry, so it shows no strong geometric preference from CFSE and readily adopts tetrahedral coordination in enzyme active sites (carbonic anhydrase, carboxypeptidase, many zinc-finger transcription factor domains), where tetrahedral geometry is instead selected by ligand sterics, coordination number preferences, and the biological scaffold, precisely because there is no crystal-field energy penalty standing in the way (unlike a $d^8$ ion such as Ni(II), where CFSE strongly disfavors tetrahedral coordination, as shown quantitatively in Worked Example 2).

Semiconductor and materials chemistry. Many II–VI and III–V semiconductors (ZnS, ZnSe, GaAs) and spinel-structure oxide materials feature tetrahedrally coordinated transition-metal or main-group cation sites; the tetrahedral crystal field splitting parameter directly affects dopant transition-metal ion spectroscopy in materials like ZnS:Mn or spinel ferrites, relevant to phosphors and magnetic materials design.

Homogeneous catalysis. Many first-row transition-metal catalytic intermediates (particularly in cross-coupling and polymerization catalysis using Ni, Co, and Fe complexes with bulky phosphine or NHC ligands) adopt tetrahedral or pseudo-tetrahedral geometries at key steps, where the weak-field, always-high-spin character directly affects reactivity, radical character, and spin-state changes along the catalytic cycle — understanding $\Delta_t$ helps rationalize why these intermediates behave so differently from octahedral analogues in terms of accessible spin states and reaction barriers.

Foreshadowing further splitting. The $e$/$t_2$ pattern derived here is the simplest case of ligand-field symmetry lowering; distorting a tetrahedron (compressing or elongating it) or moving to four-coordinate square-planar geometry splits these sets further, connecting directly to the ideas behind Square-Planar Crystal Field Splitting and to the general energetic bookkeeping used in Crystal Field Stabilization Energy calculations across all coordination geometries.

Ask the tutor: How does the $e$/$t_2$ splitting pattern change further when a tetrahedral complex is distorted toward square-planar geometry, and why does that distortion favor low-spin $d^8$ configurations?

Set up a tetrahedron the standard way: place the metal at the origin and the four ligands at alternating corners of a cube, at coordinates $(1,1,1)$, $(1,-1,-1)$, $(-1,1,-1)$, $(-1,-1,1)$ (in units of the cube half-edge). This choice is not arbitrary — it lets you directly compare tetrahedral and octahedral geometries using the same coordinate axes, since an octahedron's ligands sit on $(\pm1,0,0)$, $(0,\pm1,0)$, $(0,0,\pm1)$, the face centers of that same cube. Now ask which d orbitals point most directly at the tetrahedral ligand positions. The $d_{z^2}$ and $d_{x^2-y^2}$ orbitals have their lobes along the Cartesian axes and in between them along $x$, $y$, $z$-bisecting planes; none of the four cube-corner ligand vectors lies along an axis, so these two orbitals are the ones that see the least direct ligand electron density along their lobe directions — they become the lower-energy set, called $e$ (no "g" subscript, because $T_d$ has no inversion center, so orbitals cannot be labeled gerade/ungerade). The $d_{xy}$, $d_{xz}$, $d_{yz}$ orbitals, by contrast, have their lobes pointing directly between pairs of axes — precisely toward the cube-corner directions where the ligands sit. These three orbitals are pushed to higher energy, forming the $t_2$ set. This is the exact inversion of the octahedral result, and the reason is geometric, not some new physics: in the octahedron the axial orbitals ($d_{z^2}$, $d_{x^2-y^2}$) point straight at ligands and are destabilized ($e_g$, high), while the off-axis orbitals ($t_{2g}$, low) avoid them. In the tetrahedron, none of the orbitals point exactly at a ligand (since no ligand sits on a Cartesian axis), but the $t_2$ orbitals come geometrically closer to the ligand directions than the $e$ orbitals do, so the roles of "closer to ligands = higher energy" persist, just redistributed onto the opposite orbital set.

The second essential piece of the picture is why $\Delta_t$ is smaller than $\Delta_o$ even before you account for having fewer ligands. In the octahedron, the two $e_g$ orbitals point directly down the metal–ligand bond axes, producing maximal $\sigma$-type destabilization; the $t_2$ orbitals in the tetrahedron, even though they are the higher-energy set, never achieve that direct head-on alignment — the angle between a $t_2$ lobe and the nearest ligand vector is about $54.7°$, not $0°$. This purely geometric mismatch (no orbital in $T_d$ gets the full "aimed straight at a ligand" overlap that $e_g$ orbitals get in $O_h$) reduces the destabilization achievable in the tetrahedral case relative to the octahedral one, layered on top of the reduction from four ligands instead of six. The third piece is spin: the practical consequence of both effects. Typical pairing energies for first-row transition metal ions are on the order of $15{,}000$–$20{,}000\ \text{cm}^{-1}$. Typical octahedral splittings $\Delta_o$ for reasonably strong-field ligands can approach or exceed that scale, making low-spin octahedral complexes achievable (e.g., $[\text{Fe(CN)}_6]^{4-}$). But because $\Delta_t \approx \tfrac{4}{9}\Delta_o$, even the strongest tetrahedral ligand fields rarely produce $\Delta_t$ larger than about $6{,}000$–$9{,}000\ \text{cm}^{-1}$ — comfortably below any reasonable pairing energy. The result is qualitative, not marginal: tetrahedral 3d complexes are essentially always high-spin, and this is a structural, not accidental, feature of the geometry.

The tetrahedral crystal field splitting energy is written, exactly as in the octahedral case, as a single parameter $\Delta_t$ that separates the two symmetry sets:

where: — $E(t_2)$, $E(e)$ are the energies of the $t_2$ (three-fold, $d_{xy}, d_{xz}, d_{yz}$-derived) and $e$ (two-fold, $d_{z^2}, d_{x^2-y^2}$-derived) orbital sets in the tetrahedral ligand field, — $\Delta_t$ (sometimes written $10Dq_t$ by analogy) is the total tetrahedral splitting, defined positive with $t_2$ above $e$. Because the barycenter (average energy, weighted by degeneracy) of the d orbitals must stay fixed relative to the free-ion energy (crystal field theory conserves total energy — it only redistributes it among orbitals), the individual orbital energies relative to that barycenter are fixed by the degeneracies:

where: — the coefficients 2 and 3 are the degeneracies of $e$ and $t_2$, — energies are measured relative to the barycenter (weighted average = 0 by definition). Solving the two equations simultaneously gives the now-familiar bookkeeping fractions:

where: — $-\tfrac{3}{5}\Delta_t$ is the stabilization of each of the two $e$ orbitals, — $+\tfrac{2}{5}\Delta_t$ is the destabilization of each of the three $t_2$ orbitals, — note these are the mirror image of the octahedral fractions $-\tfrac{2}{5}\Delta_o$ ($t_{2g}$) and $+\tfrac{3}{5}\Delta_o$ ($e_g$): the "2" and "3" have swapped which set they attach to, because in $T_d$ the doubly-degenerate set is stabilized and the triply-degenerate set is destabilized — exactly the reverse assignment from $O_h$. Reading: this pair of equations is pure electrostatic/orbital bookkeeping, not a new physical postulate — it is required by the fact that the total change in metal d-orbital energy summed over all five orbitals must vanish (spherical average preserved), for any point-group splitting. The central quantitative relationship of this article compares $\Delta_t$ for a tetrahedral complex to $\Delta_o$ for an octahedral complex of the same metal, same ligand, same metal–ligand bond distance:

where: — $\Delta_o$ is the octahedral splitting for the same metal/ligand/distance, — the factor $\tfrac{4}{9} \approx 0.444$ combines a ligand-count factor of $\tfrac{4}{6} = \tfrac{2}{3}$ and a geometric overlap factor of $\tfrac{2}{3}$, multiplying to $\tfrac{4}{9}$, — this is an approximation (point-charge/angular-overlap estimate), accurate to perhaps 10–20%, not an exact theorem, because real bonding also involves π-effects and orbital mixing not captured by the simple electrostatic model. Reading: this single number is the reason essentially no first-row tetrahedral complex is low-spin — even a full order-of-magnitude change in ligand field strength within the tetrahedral family still lands well under typical pairing energies.

Example 1 — Foundational: orbital energies for a tetrahedral $d^7$ ion. Consider a hypothetical tetrahedral $\text{Co}^{2+}$ complex ($d^7$) with $\Delta_t = 4{,}700\ \text{cm}^{-1}$ (a typical experimental value for $\text{CoCl}_4^{2-}$-type species). Using the barycenter fractions derived above,

For $d^7$ in a tetrahedral field, since $\Delta_t \ll P$ we fill by Hund's rule with maximum spin, giving the ground configuration $e^4 t_2^3$ (each of the two $e$ orbitals doubly occupied, each of the three $t_2$ orbitals singly occupied first, and then one $t_2$ orbital gets a second electron to accommodate the seventh electron), consistent with a high-spin $S = \tfrac{3}{2}$ ground state ($\text{Co}^{2+}$ tetrahedral complexes are the classic high-spin $S=3/2$ paramagnets, magnetic moment near $4.7$–$4.9\ \mu_B$, elevated above the spin-only value of $3.87\ \mu_B$ by substantial orbital contribution — itself a hallmark of the $t_2$ ground term in tetrahedral $d^7$). As a limiting-case check: setting $\Delta_t = 0$ recovers $E(e) = E(t_2) = 0$, the free-ion (spherically symmetric) limit, as required.

Example 2 — Applied: comparing CFSE between octahedral and tetrahedral Ni(II). $\text{Ni}^{2+}$ is $d^8$. In an octahedral field, $\text{Ni(H}_2\text{O)}_6^{2+}$ has $\Delta_o \approx 8{,}500\ \text{cm}^{-1}$ and a $t_{2g}^6 e_g^2$ high-spin configuration (there is no low-spin option for octahedral $d^8$ regardless of field strength, since $t_{2g}^6 e_g^2$ is the only configuration consistent with an even electron count filling $t_{2g}$ first). Its CFSE is

Numerically, $\text{CFSE(oct)} = -\tfrac{6}{5}(8500\ \text{cm}^{-1}) = -10{,}200\ \text{cm}^{-1}$. Now compare a tetrahedral $\text{NiCl}_4^{2-}$ complex, also $d^8$, with $\Delta_t \approx \tfrac{4}{9}(8500\ \text{cm}^{-1}) \approx 3{,}780\ \text{cm}^{-1}$ (using the same ligand-type/metal comparison as an estimate; actual chloride vs. water field strengths differ slightly, but this isolates the geometric effect). High-spin tetrahedral $d^8$ fills as $e^4 t_2^4$ (both $e$ orbitals filled, then each $t_2$ singly occupied before pairing, giving one $t_2$ orbital doubly occupied to reach 8 electrons total: $e^4 t_2^4$ means 4 in $e$, 4 in $t_2$, distributed as $e^4$ fully paired, $t_2^4$ as three singly + one doubly occupied). The CFSE is

Numerically, $\text{CFSE(tet)} = -\tfrac{4}{5}(3780\ \text{cm}^{-1}) = -3{,}024\ \text{cm}^{-1}$. The octahedral CFSE ($-10{,}200\ \text{cm}^{-1}$) is far more stabilizing than the tetrahedral CFSE ($-3{,}024\ \text{cm}^{-1}$) for $d^8$, which is exactly why nickel(II) strongly prefers octahedral (or square-planar) coordination and tetrahedral $\text{Ni}^{2+}$ complexes, while known (e.g. $\text{NiCl}_4^{2-}$, $\text{NiBr}_4^{2-}$), are comparatively rare and require bulky or weak-field ligands and often coexist with square-planar isomers in solution equilibria. As a limiting-case check: if $\Delta_t \to 0$, $\text{CFSE(tet)} \to 0$, correctly recovering the no-splitting limit.

Example 3 — Exam/competition level: does any first-row tetrahedral ion ever go low-spin, and what constraint on $\Delta_t/P$ would be required? Consider tetrahedral $d^6$ (e.g., a hypothetical tetrahedral $\text{Fe}^{2+}$ complex). High-spin $d^6$ fills as $e^3 t_2^3$ giving $S=2$ (four unpaired electrons, since only after all five orbitals are singly occupied does the sixth electron pair up, and by the ordering $e$ is filled preferentially since it is lower in energy — giving $e^{3}$, i.e. one $e$ orbital doubly occupied and the other singly, plus $t_2^3$ all singly occupied — total unpaired $= 1(e) + 3(t_2) = 4$, $S=2$). A hypothetical low-spin $d^6$ tetrahedral configuration would instead fill the two lower $e$ orbitals completely ($e^4$) before touching $t_2$, giving $e^4 t_2^2$ with the two $t_2$ electrons unpaired by Hund's rule within $t_2$ (since $t_2$ has three orbitals, two electrons singly occupy two of them), for $S=1$. Compare the total energies (electronic CFSE plus pairing energy cost $P$ per extra pairing beyond the free-ion reference) of the two configurations. High-spin $e^3 t_2^3$ has one forced pairing (in one $e$ orbital, to fit 3 electrons into 2 orbitals) beyond the "all singly occupied where possible" reference, contributing $+P$, plus

so total energy relative to unperturbed, unpaired reference is $-\tfrac{3}{5}\Delta_t + P$. Low-spin $e^4 t_2^2$ has two forced pairings (one extra pairing to fill $e^4$ completely, since the free-ion reference has each orbital singly occupied where possible — actually for $d^6$ there is always at least one forced pairing overall since 6 electrons into 5 orbitals requires one pair regardless of spin state; the difference between HS and LS is in which orbital carries the extra pairing and whether a second pairing is forced). Careful electron counting: $e^4$ requires both $e$ orbitals doubly occupied (2 pairings), $t_2^2$ has two singly-occupied orbitals (0 pairings) — total 2 pairings, cost $2P$, while

giving total energy $-\tfrac{8}{5}\Delta_t + 2P$. Low-spin is favored energetically when

This is the general condition, structurally identical in form to the octahedral high-spin/low-spin crossover condition, but here the crucial fact is numerical: since $\Delta_t \approx \tfrac{4}{9}\Delta_o$ and even the strongest-field tetrahedral ligands rarely exceed $\Delta_t \sim 6{,}000$–$9{,}000\ \text{cm}^{-1}$, while $P$ for Fe(II) is around $17{,}000\ \text{cm}^{-1}$, the inequality $\Delta_t > P$ is never satisfied for any known tetrahedral first-row complex. As the limiting check: even taking an unrealistically strong hypothetical tetrahedral field with $\Delta_t = 12{,}000\ \text{cm}^{-1}$ (well beyond any real ligand achieves in $T_d$ geometry for Fe(II)), this still falls short of $P \approx 17{,}000\ \text{cm}^{-1}$, confirming that low-spin tetrahedral $d^6$ remains energetically disfavored across the entire physically accessible range — which is the rigorous version of the textbook statement "tetrahedral complexes are always high-spin."

The crystal field framework used here descends from the electrostatic theory developed initially for the interpretation of transition-metal ion spectra and magnetism in crystals, and later adapted for coordination complexes.