Why flat complexes substitute by inviting, not by leaving
The previous guide in this rung sorted complexes into labile and inert and split substitution into two limiting routes: associative, where the incoming ligand attaches *before* the old one leaves so the metal momentarily gains a coordination number, and dissociative, where the old ligand drops off *first*. This guide zooms in on one geometry that almost always chooses the associative road: the square-planar complex, the flat four-coordinate shape adopted overwhelmingly by d8 metal centres such as platinum(II), palladium(II), nickel(II), and gold(III).
Look at the shape and the reason becomes physical. A square-planar complex is an octahedron with the top and bottom ligands stripped away, leaving the metal flat with four ligands in a plane — and two wide-open faces above and below. In d8 the five d orbitals fill so that the only empty one of high energy, dx2-y2, points right at the four in-plane ligands, while the dz2 orbital aimed at those vacant faces is comparatively low and full. An incoming ligand simply walks up the open axis toward the metal, donating its lone pair, with nothing in its way. The metal does not have to wait for a ligand to leave before it can react; the front door is already open. That is exactly the setup for an associative mechanism.
Walking the associative mechanism, step by step
Picture the workhorse reaction: [PtL3X] + Y -> [PtL3Y] + X, where Y is the new ligand and X the one being kicked out. The square-planar starting material does not snap straight to product. Instead it bulges through a crowded five-coordinate intermediate shaped like a trigonal bipyramid — a triangle of three ligands around the metal's waist with one ligand above and one below. The kinetics give the game away: the rate depends on the concentration of *both* the complex and the incoming Y (rate = k[complex][Y]), which is the signature of a step where two species come together. A purely dissociative reaction would not care how much Y you added.
- Approach. The incoming nucleophile Y drifts up one of the open axial faces of the flat complex and starts donating its lone pair into the empty orbital on the metal — the coordination number begins climbing from four toward five.
- Form the trigonal-bipyramidal intermediate. Y, the leaving group X, and the one ligand trans to X (call it T) rearrange into the equatorial triangle around the metal's waist; the other two ligands sit at the axial top and bottom. This crowded five-coordinate species is the heart of the mechanism.
- Eject the leaving group. From that same equatorial triangle, X peels off, the metal relaxes back down to four coordination, and the remaining ligands flatten into a new square plane — now carrying Y where X used to be.
The crucial geometric fact lives in step 2: X and the ligand trans to it (T) end up in the *same* equatorial plane, alongside the incoming Y. That is why the ligand sitting opposite the leaving group has so much say over how fast — and even whether — the swap happens. The associative mechanism literally seats the trans ligand at the table where the bond-breaking is decided. Hold that thought; it is the whole reason the next section matters.
The trans effect: a ligand that loosens its opposite
Here is the central discovery. Certain ligands, when they sit in a square-planar complex, dramatically speed up the replacement of whatever ligand sits directly *opposite* them — trans to them. This kinetic phenomenon is the [[trans-effect|trans effect]]: the power of a ligand to labilize its trans partner. Pin down the word 'kinetic' — the trans effect is about *rate*, about how readily the trans bond can be broken in a reaction, not about how strong that bond is at rest.
Measured over many reactions, ligands fall into a rough ordering called the trans-effect series, from weak to extremely strong: H2O ~ OH- < NH3 < pyridine < Cl- < Br- < I- ~ NO2- ~ SCN- < phosphines (PR3) ~ H- < CO ~ CN- ~ C2H4. A ligand near the right end, placed trans to X, can accelerate the loss of X by a factor of thousands or millions over a ligand from the left end. Treat this as an empirical, approximate ladder, not a law of nature — solvent and the specific metal shuffle the order a little.
Why does the series look like that? Two distinct contributions add up. The first is a ground-state, electronic effect tied to sigma-bonding strength — and it is best understood as the *trans influence*, the topic of the next section. The second, and the one that puts CO, cyanide and ethene at the very top, is a transition-state effect from pi-acceptance. A good pi-acceptor like CO pulls electron density out of the metal's d orbitals in the equatorial plane; that drains charge from exactly the region where the five-coordinate intermediate is most crowded, stabilizing that bulky trigonal-bipyramidal transition state and so lowering the barrier. Strong sigma-donors loosen the trans bond in the ground state; strong pi-acceptors cushion the crowded intermediate. The series is the sum of both.
Trans effect vs trans influence: keep them apart
These two names sound interchangeable and routinely get muddled, so it is worth being deliberate. The trans influence (sometimes the *structural* trans effect) is a purely ground-state, thermodynamic property: a strongly sigma-donating ligand pushes electron density at the metal so effectively that it weakens and *lengthens* the bond to whatever sits trans to it, even in a molecule that is just sitting on the shelf doing nothing. You measure it with a ruler and a spectrometer — longer bond lengths in a crystal structure, shifted stretching frequencies.
The trans *effect*, by contrast, is kinetic: it is about reaction rates and lives partly in the transition state. The two correlate but are not the same thing, and the mismatch is the giveaway. A pure sigma-donor like the hydride H- ranks high on *both* lists — it weakens the trans bond at rest and labilizes it in reaction. But a brilliant pi-acceptor like CO tops the *kinetic* trans-effect series while having only a modest trans *influence*, because most of its power is to stabilize the five-coordinate transition state, not to lengthen the ground-state bond. So: a ligand can speed up trans substitution far more than its mere bond-weakening would explain. Spotting that gap is how you know pi-acceptance is at work.
trans INFLUENCE (ground state, thermodynamic) ~ sigma-donor strength H- ~ PR3 > CO ~ CN- > I- > Br- > Cl- > NH3 > OH- measured as: longer trans bond, shifted IR / NMR trans EFFECT (kinetic, rate of substitution) = sigma part + pi part CO ~ CN- ~ C2H4 > PR3 ~ H- > NO2- ~ I- > Br- > Cl- > NH3 > OH- ~ H2O measured as: how fast the trans ligand is replaced the gap between the two lists (e.g. CO jumps to the top of EFFECT only) = the pi-acceptor contribution, which stabilizes the 5-coordinate TS
Cashing it in: the rational synthesis of cisplatin
Now the payoff that makes this whole topic famous. Cisplatin is cis-[Pt(NH3)2Cl2], a square-planar platinum(II) complex with the two ammonias next to each other and the two chlorides next to each other. It is one of the most successful anticancer drugs ever made: inside a cell its two chlorides are slowly replaced by nitrogen atoms on adjacent guanine bases of DNA, kinking the double helix and jamming replication — the basis of cisplatin's binding to DNA. Crucially, only the *cis* isomer works. The *trans* isomer, with the chlorides across from each other, is therapeutically useless. So the chemist's job is not just to make Pt(NH3)2Cl2 — it is to make *exactly* the cis arrangement and avoid the trans.
Reach for the series. In it, chloride sits well above ammonia: Cl- has the stronger trans effect. That single fact, applied twice, dictates the product. To make *cis*-platin, start from the tetrachloride [PtCl4]2- and add ammonia. The first NH3 can replace any of the four equivalent chlorides. Now three Cl and one NH3 remain. The second NH3 wants to go trans to a strong trans-director — and the three remaining sites are: one trans to NH3, two trans to Cl. Since Cl has the stronger trans effect, it labilizes the chloride opposite *it* most, so the second NH3 displaces a chloride that is trans to a chloride. The result: the two ammonias end up cis to each other. Geometry delivered by kinetics.
Run it the other way and you build the *trans* isomer on purpose. Start from [Pt(NH3)4]2+ and add chloride. The first Cl- replaces one ammonia, leaving three NH3 and one Cl-. Now the strongest trans-director in the molecule is that lone chloride, so the second Cl- enters trans to the first chloride — placing the two chlorides across from each other, giving trans-platin. Same two reagents, same metal, opposite isomers, chosen simply by which ligand you put on first. That is the trans effect as a synthetic tool: a deliberate ordering of additions that lets you dial in geometry instead of praying for it.
What to carry forward
Pull the thread together. A square-planar d8 complex has open faces above and below, so substitution runs by an associative mechanism through a five-coordinate trigonal-bipyramidal intermediate — and that geometry seats the leaving group beside the ligand trans to it. From that seating arrangement falls the trans effect: a kinetic ranking of how strongly each ligand labilizes its opposite number, fed by sigma-donation (which also shows up as the structural trans influence) and by pi-acceptance (which only shows up in the rate). Order your additions by that ranking and you can write a specific isomer into existence — cis or trans on demand.
A last honest caveat before the rung moves on. The trans-effect series is empirical and approximate — it is a sturdy guide, not a guarantee, and a different metal, solvent or set of ligands can nudge the order. And everything here was about *substitution*, ligands trading places while the metal keeps its oxidation state. The remaining guides in this rung turn to the other great class of reactions, electron transfer, where the metal changes oxidation state without necessarily swapping a single ligand — the outer-sphere and inner-sphere mechanisms that move charge around the inorganic world.