LearnChem · Part II — Why do atoms stick together?
Ruby red and emerald green from the same guest atom. A violet so intense one crystal stains a bucket. And iron that turns nitrogen into fertiliser for years without being spent. One shelf of electrons — part-filled, split by a cage — explains the lot.
“A ruby is red and an emerald is green. Why do both get their colour from the same metal?”
Last time, the family photograph: columns read as families, the resemblance caused — size against pull — and even the member who broke the rules was obeying them. But the tour walked straight past the photograph's wide middle: the low block where the table pauses for ten columns of metals that all seem, at first glance, simply… metal. Yet these are the elements that colour glass and gems, change colour with every change of charge, and run the world's reactions for a living. This module is why that middle block is different.
Stand at a jeweller's window with the question at the door. The ruby and the emerald in the tray owe their colours to the same impurity: a scatter of chromium ions, Cr³⁺, sitting as guests inside two different colourless minerals. Same guest, same charge, same electron count — red stone, green stone. Whatever makes colour, it cannot be a fixed property of the atom alone. The surroundings must have a vote.
Hold that puzzle, and collect the clue module 05 left lying about. Module 05's floors filled in a strange order: after argon, the table pauses for ten columns — scandium to zinc — while an inner floor, the d shell, fills up with its ten places. These are the transition metals: elements whose working electrons include a part-filled inner shelf. Ten places, some taken, some free. Everything in this module — the colours, the colour-changes, the catalysis — is that part-filled shelf at work. (One bookkeeping rule for later: when one of these metals becomes an ion, its two outermost electrons leave first; only then does the shelf pay.) And the shelf explains the jeweller's window too, once one more idea arrives: what a cage of neighbours does to the shelf. That idea is §2's bench.
Free in space, the five rooms of a d shell share one rent: an electron in any of them holds the same energy. Now give the ion six neighbours, the way six water molecules cage a dissolved metal ion. Module 07's pushing-apart rule, played with six pairs, lands them at the corners of a shape that module never needed: the octahedron, the tetrahedron's six-cornered cousin. Each neighbour points a pair of electrons inward. Two of the five d rooms happen to point straight at the incoming pairs; the other three point between them. Rooms staring down an electron pair charge more rent than rooms looking into the gaps — module 06's cost of crowding, one more time. The one shelf splits into two: three rooms lower, two rooms higher, separated by a gap chemists write as Δ.
A gap is a price that light can pay. Module 01 priced a flash of green light at a couple of electron-volts (eV) — its energy coin; each such packet of light is called a photon, and the energy each photon carries is set by its colour. λ, the wavelength, is the colour's other name. If a photon's energy matches Δ, an electron in a lower room can absorb it and jump to an upper one. That photon is subtracted from the light. Every other photon sails through unmatched. White light goes in; white-minus-one-slice comes out — and white-minus-a-slice is a colour.
Copper sulphate in water: blue. Zinc sulphate — copper's next-door neighbour on the table — in water: colourless. Commit — what does zinc's ion lack?
Commit before you look. First answers are counted anonymously, never named.
Zinc's ion holds ten d electrons — a full shelf. A jump needs both a jumper and a vacancy, and Zn²⁺ has no vacancy; scandium's ion has the opposite problem, a bare shelf with no jumper. Both are colourless for the same reason read from opposite ends: no usable gap, no subtraction, no colour. Copper's ion, nine electrons and one vacancy, sits in between — and is blue. The rule earns its keep at the edges of the middle block: colour lives where the shelf is part-filled, exactly the transition metals' territory. Now watch the arithmetic that turns a gap into a colour:
Set titanium's cited gap and read the arithmetic: Δ ≈ 20,300 cm⁻¹ means λ ≈ 493 nanometres — a green photon — at a price of about 243 kJ/mol. Notice the units: bond-scale money, module 06's own currency. Visible light is exactly energetic enough to work chemistry's till — which is why there is such a thing as looking at chemistry and seeing it.
Quick check — Bench 1 prices titanium's gap at about 243 kJ/mol: a photon from the green, near 500 nm. What colour does the solution look?
White minus its green slice is red-violet — the colour titanium's solutions do show. The colour is the gap, worn inside-out. Slide Δ smaller and the absorbed slice glides toward the red end — and past it. Press nickel's preset and the bench meets an honest limit: nickel's Δ works out beyond the visible strip altogether, in the infrared, yet nickel solutions are green. That green comes from other, higher jumps inside its nine-electron shelf — real, but past this module's one-gap arithmetic — and the bench says so rather than inventing a slice.
Back to the jeweller's window, carrying the machine from §2. Both stones hold Cr³⁺ with six oxygen neighbours at the corners of an octahedron — the same cage shape, module 07's geometry. If the gap decides the colour, and the same guest shows two colours, then the two hosts must set two different gaps. Commit to the mechanism:
A ruby and an emerald get their colour from the very same guest — Cr³⁺, in a cage of six oxygens in both stones. Commit — what changes the colour?
The cage. Corundum — aluminium oxide, ruby's host — is a tight crystal that holds the guest's six oxygens close; close neighbours press hard on the staring rooms, and the gap is wide. Beryl — emerald's host — sets its oxygens in a roomier silicate frame, a touch farther out. The press is gentler; the gap narrows. Estimates put the two gaps near 2.23 and 2.05 eV — popular-science figures, flagged on Bench 2, though GIA's measured bands agree with them. A wide gap eats the green-and-violet end and the stone glows red. A narrower gap eats a lower slice — the orange-yellow — and what survives is green. Same guest, different cage, different rent — different colour. Two panes of one stained-glass window, the same pigment set in different leading. The bench draws both stones' cited absorption bands; the toggle is the whole answer to the question at the door.
One resident of the palette breaks the rule outright. Permanganate — MnO₄⁻, the violet so intense that one crystal stains a bucket — holds manganese in its +7 oxidation state: the charge an atom is left carrying once electrons are handed away, a number this course will often call its costume. Seven electrons' worth handed away leaves the d shelf holding zero electrons. No jumper, no d-to-d jump, and by §2's rule it should be colourless as water. Instead it is the most stained-glass violet in the stockroom:
Quick check — permanganate's manganese is d⁰, no d electrons at all. Where does the violet come from?
The intensity is the tell. A d-to-d jump absorbs only weakly — the two kinds of room overlap awkwardly, so the jump is seldom made — and the aqua-ion colours run pale: pale green Fe²⁺, pale pink Mn²⁺, sky-blue Cu²⁺. A charge transfer — a whole electron thrown from an oxygen's room to the metal's — absorbs hugely, which is why one crystal outstains a bucketful of any aqua ion. Two mechanisms, one giveaway: paleness is usually a d hop, ferocity a transfer.
Iron the metal is module 09's story: a sea, answering all light more or less alike — grey. Rust is iron after the transaction — Fe³⁺ ions caged by oxygens in hematite, and the strong red-brown is mostly the ferocious mechanism, not the polite one: oxygen-to-iron charge transfer, with the d-to-d hops adding shading (Sherman & Waite 1985). The grey was the sea; the red is the cage. Same element, opposite side of the electron ledger — and when module 21 builds the battery out of exactly this transaction, the colour will be your bookkeeping aid: grey means electrons pooled, red-brown means electrons handed away.
The palette's rule — colour lives on a part-filled shelf — now walks straight into the largest colour on Earth. Grass. Leaves. The green of half the living world. Chlorophyll, the molecule behind all of it, is built round a single metal atom, and the metal is magnesium: module 05's second column, an ion with a closed shell and no usable d gap at all. The rule says no colour. Every field outside says otherwise. Commit:
Quick check — chlorophyll's metal is magnesium, with no usable d gap. Why is grass still green?
The one great colour of the living world is not a transition metal's. Chlorophyll is built round a metal, but the metal is magnesium — module 05's closed-shell column, no usable d gap at all. The absorber is the ring: a wide, flat molecule whose shared electrons spread over dozens of atoms, and spread-out electrons have closely spaced levels — gaps priced for visible light with no d shelf needed. The ring takes red and blue and lets the middle of the spectrum pass: green. Magnesium's job is to sit in the ring's centre and hold its shape. In autumn the order of demolition answers the second question: chlorophyll is dismantled first, and the carotenoids — humbler chain-molecules that absorb only the blue — are left holding the leaf: yellow (the Harvard Forest measurements track exactly this sequence). Three colour machines now: the d gap, the transfer, the ring — and a leaf runs on the third.
A sticker's powder is strontium aluminate salted with two guest ions, europium and dysprosium (the recipe dates to a 1996 paper — Matsuzawa and colleagues). Daylight promotes electrons much as in §2, here inside europium's own ion — and some of the promoted electrons are caught, on the way down, in traps: defect sites about 0.65 eV deep, too deep to escape instantly, shallow enough for the room's warmth to shake them loose a few at a time (Van den Eeckhout's review prices the traps). Each late escape releases its stored light. The glow fades as the traps empty — and glows brighter again after a minute under a lamp, because you have refilled them. It is §2's jump, run in slow motion through a tollbooth.
Module 09: a metal's sea answers nearly all visible light, throwing it back — grey-silver. Gold's sea absorbs the blue instead of returning it, and what it hands back is the warm remainder. Copper's red needs no exotic help — its absorption sits in the visible on ordinary grounds. Gold's should not, and here is module 10's mercury confession over again: gold's inner electrons move near the speed of light, relativity reshapes their floors, and an absorption that belongs in the ultraviolet slides down into the visible blue (Pyykkö's reviews; Norrby's classic paper covers gold and mercury together). Chemistry hands this one to LearnPhysics with its compliments — and keeps the observation: even colour, the most chemical of properties, sometimes has physics all the way down.
The cage does one more thing besides setting the rent: it fixes addresses. Six corners are six seats, and once two kinds of guest sit around one metal — say four of one molecule and two of another — the question which seats becomes real chemistry. Two arrangements that cannot be rotated into each other are two different substances: same formula, different compound. Chemists call them isomers, and for decades the surest sign of them was a row of differently coloured sample jars. Before the bench does your counting, commit:
Quick check first — six corners, three guests of one kind and three of another: MA₃B₃. How many genuinely different arrangements are there?
Four-and-two gives two substances: the odd pair adjacent (chemists say cis, Latin for “on this side”) or opposite (trans, “across”). Three-and-three also gives exactly two: a triangle sharing one face (fac) or a band round the equator (mer). Twenty placements collapse to two under rotation — the octahedron's symmetry does the pruning, and the bench did it by brute turning, the way you can with a marked die in your hand.
This counting founded a science. Between the 1890s and 1907, Alfred Werner and his rivals prepared cobalt compounds of identical formula that came out of the flask as different colours — the laboratory names survive: violeo salts (violet) and praseo salts (leek-green), a cis–trans pair of exactly Bench 3's four-and-two kind (Ernst et al. 2011 re-examined Werner's original samples). The counting was strong evidence: the number of distinct compounds he could isolate matched the two an octahedron predicts, and not the counts rival shapes predict. Resolving mirror-image complexes in 1911 clinched the case, and the 1913 Nobel Prize followed. Werner counted arrangements the way the bench just did, decades before anyone could see a molecule. There is a whole module of such tricks waiting at the course's end — module 30.
One family trait remains, and it is the one the world economy leans on. A part-filled shelf makes costume changes cheap: with rooms at similar rent always open, a transition metal can hold +2, +3, +4 — electrons taken or lent a few at a time, each oxidation state its own colour. Manganese wears the whole wardrobe: +7 in permanganate's violet, +4 in the brown solid MnO₂, +2 in the palest pink of Mn²⁺ (Flinn's demonstration chart has the full six-costume wardrobe). A titration flask that snaps from violet toward colourless is manganese changing costume before your eyes — the shelf's bookkeeping made visible.
Now the payoff. Cheap costume changes mean a transition metal can take part in a reaction and undo its own participation — accept electrons from one partner, hand them to another, and end the day in the costume it woke in. Add a surface built of such atoms and you have chemistry's most valuable furniture:
An iron catalyst sits in an ammonia reactor for years, turning nitrogen and hydrogen into ammonia by the tonne — and comes out iron. Commit — what is the metal doing in there?
The reaction is N2(g) + 3 H2(g) → 2 NH3(g) — run both directions at once in the real reactor, which is module 17's story — and its hard step is prying open nitrogen's triple grip, one of the most expensive bonds in nature. On the iron surface that price is never paid in one go. The nitrogen settles onto the metal, and the part-filled shelf lends electron room into the bond while taking a little from elsewhere; the triple grip loosens in stages. A half-broken nitrogen resting on iron is a genuine resting stop — far cheaper than a free, fully torn atom — so one impossible climb becomes several affordable ones. The pieces meet at bench height and leave as ammonia; the iron ends every cycle as iron. A catalyst changes no ledger and shifts no balance; it only makes the expensive step cheaper to attempt, in both directions equally — the full pricing of that sentence is modules 16 to 18. The same lending, different bookkeeping: in the sulphuric-acid works, vanadium's oxide turns sulphur dioxide to trioxide — 2 SO2(g) + O2(g) → 2 SO3(g) — and there the loan is kept in whole costumes, the vanadium's oxidation state swinging down and back up each cycle. The platinum-family honeycomb in every car's exhaust finishes the burning the engine left half-done: 2 CO(g) + O2(g) → 2 CO2(g). Fertiliser, acid, breathable streets: the part-filled shelf, hired out three ways.
These check themselves, and “New numbers” deals a fresh set — there is nothing to memorise. Each stem says how exact to be.
1. Module 05's floors, revisited: how many d electrons does hold? (A whole number.)
2. Bench 1's first conversion: a gap of cm⁻¹ absorbs at what wavelength? λ = 10⁷ ÷ Δ, in nanometres. (To ±2.)
3. And its price: that same gap in kJ/mol — read it off Bench 1, or use its constants. (To ±2.)
4. Module 09's box, revisited: copper's face-centred cell — how many atoms does the cell really contain? (A whole number.)
Do the window's sum yourself. Ruby's measured ligand-field energy is about 2.23 eV; the handy conversion is λ = 1239.84 ÷ E(eV) nanometres. Work out the wavelength ruby's main band absorbs, in nm, to the nearest whole number — then check yourself below, and do emerald's 2.05 eV the same way to see the whole answer to the question at the door.
Worked through. Ruby: 1239.84 ÷ 2.23 = 556 nm — the green-yellow slice, gone; red survives, and the number agrees with GIA's measured band near 560. Emerald: 1239.84 ÷ 2.05 = 605 nm — the orange slice, gone; green survives. A 0.18 eV difference in rent — less than a tenth of the gap itself — is the entire distance between a ruby and an emerald. (The two energies are from a popular-science source and carry that flag on Bench 2; the GIA bands are the measured anchor.)
Three colours from your own street: rust bleeding down a white wall; turmeric's yellow on a steel plate; the green-blue edge you see looking through a windowpane side-on. This module built three colour machines — the d gap in a cage, the charge transfer, the chlorophyll-style ring — and §4 added two oddities besides. Assign each stain to a machine, and for one of them propose an observation that would test your assignment. Write your three verdicts before looking.
The discussion. Rust: mostly the transfer — §3's box: oxygen-to-iron charge transfer does the heavy staining, d hops add shading; the tell is its strength, far past any pale aqua ion. Turmeric: the ring — no metal in it at all; curcumin is a long, flat molecule whose spread-out electrons take the blue, like chlorophyll's machine tuned one slice over — and a kitchen test exists: a drop of soap solution shifts its colour to red, because changing the molecule's charge retunes the ring's gap (module 19 meets exactly this trick in indicators). The window edge: the d gap — window glass carries a trace of Fe²⁺, and metres of glass edge-on stack that pale green tint into a visible one; the tell is the paleness. One street, three machines — and you now read colour the way module 09 read heat: as a fingerprint of the mechanism underneath.
A part-filled d shell gives colour, and a cheap place for a reaction to rest.