Field

Where the atoms go

VSEPR as a repulsion minimisation rather than a table, the angles that fall out of it, and the case where five sites are not all alike.

4 sites, minimised. The arrangement of 4 points on a sphere that minimises their mutual repulsion. The angles printed were measured off the result rather than quoted, and the shape was not assumed.

VSEPR, computed

The tetrahedral angle is not 109.5 degrees because a textbook says so. It is arccos(−1/3), and it falls out of minimising the repulsion of four points on a sphere without ever being written down.

Five sites are not five of a kind. The minimised arrangement of five points, with the two axial sites marked apart from the three equatorial ones. Their neighbour angles differ, so the two kinds of position are genuinely different places — which the shape's name does not convey.

Every other common arrangement has one or two distinct angles. Five has three, because two of its positions are on an axis and three are round an equator — and a molecule built that way does something about it.

water — C2v. The molecule with the point group found from its coordinates rather than looked up: every candidate operation was applied and kept when it permuted the atoms among themselves.

Why water is bent

The standard answer is lone pair repulsion, it predicts the right direction, and it cannot predict the magnitude. A better rule can, and the heavier hydrides show where both accounts run out.

Eight points: the cube loses. The cube and the minimised arrangement of eight points on a sphere, with the repulsion energy of each computed. The minimum is a square antiprism — the cube twisted by forty-five degrees on one face — and the margin is about one part in three hundred.

The shapes above six coordination

Eight points on a sphere do not arrange themselves in a cube. They twist one face by forty-five degrees, and above six the arrangements stop being the ones anybody would name and start being the ones a minimisation finds.

sp3 hybrids. The directions the hybrids point, with the angle between them computed from the coefficients rather than quoted. The set is an orthogonal transformation of the atomic orbitals, so it describes the same space in different coordinates.

Bent's rule, against the substituent series

More electronegative substituents get more p character, so the angle between them shrinks. The rule predicts a trend, the trend is observed, and the quantity it depends on turns out not to be one quantity.

Bond angle against lone-pair weight, 2 lone and 2 bonding. The bond angle a weighted repulsion minimisation gives for 2 lone pairs and 2 bonding pairs, as the lone-pair weight runs from one to 3.2. The marked molecules are water and hydrogen sulfide, each placed at the weight that reproduces its measured angle.

What a lone pair is worth

A lone pair repels more than a bonding pair, says VSEPR, without saying how much more. Put a number on it and fit that number to water, and the same number is wrong for hydrogen sulfide by a factor of two — which means it was never a property of a lone pair.

Angles at exponents 1, 2, 3, 6, 12. The distinct angles of the minimised arrangement of 4, 5, 6, 7 points, under a repulsion going as one over r to the power 1, 2, 3, 6, 12. Where the arrangement is the maximally symmetric one the angles do not move at all; where it is not, both the angles and how many of them there are depend on the law assumed.

Which angles are symmetry and which are the model

VSEPR says electron pairs repel and never says by what law. For four, five and six domains it makes no difference whatever — change the exponent by a factor of twelve and not one angle moves. For seven it decides the answer.

$The s character a bond angle requires. The fraction of the one s orbital each of two equivalent hybrids must carry in order to meet at a given angle, from pure p at 90 degrees to half s at 180. The marks are measured bond angles — H₂O, NH₃, H₂S, PH₃ — and the column beside them is the budget each implies: what one bond hybrid takes, and what is left for the lone pairs to share. The relation is orthogonality, not a model of bonding, and nothing here is fitted.$

The angle does not fix the hybridisation

Two equivalent hybrids are orthogonal only at one relation between their s character and the angle between them. Run it backwards on measured bond angles and water's bonds come out sp³·⁹⁹, hydrogen sulfide's sp²⁷, and phosphine's lone pair takes eighty-four per cent of the one s orbital.

One dihedral, three point groups. Hydrogen peroxide built at 13 values of its dihedral angle, with the point group searched for from the coordinates at each. The group is C2v when the hydrogens eclipse, C2h when they are anti, and C2 at every angle strictly between. The rails below are what the group settles on its own: the molecule may be polar except at the anti arrangement, and it is chiral except at the two ends. No energy is computed anywhere, and the marked angle is the measured one rather than a minimum found here.

One coordinate, three point groups

Hydrogen peroxide has four atoms and one soft internal coordinate. Turning it from nought to a hundred and eighty degrees takes the molecule through C2v, C2 and C2h — so it is chiral at every angle but two, and forbidden a dipole at exactly one of them.

The moment that does not depend on where the origin is. For each molecule, the dipole and the largest quadrupole component about the centre, and the same two about an origin moved 1.6 bohr away. The lowest non-vanishing moment is unchanged in every row and the one above it moves in every row. The 4 non-polar molecules here have a quadrupole that is a property of the molecule; the polar ones have one that is a property of a choice.

What a dipole cannot tell apart

A dipole moment is three numbers extracted from a whole charge distribution, and enormously many distributions give the same three. The moment above it is not even a property of the molecule unless the one below it vanishes — which is why a quadrupole is quoted with an origin and a dipole is not.

How far two hybrids are from orthogonal. The overlap between two equivalent s–p hybrids of a stated label, against the angle between them. Each curve crosses zero at exactly one angle — sp3 at 109.47°, sp2 at 120.00°, sp at 180.00° — and a molecule whose measured angle is not that angle has hybrids that overlap. The largest here is cyclopropane at 0.63.

Hybrids that were never orthogonal

Two sp³ hybrids are orthogonal at 109.47° and nowhere else. Water's are drawn at 104.5° and overlap by 0.062; cyclopropane's are drawn at 60° and overlap by 0.625, which is not a small correction to a basis but a description that has stopped being one.

The largest angle each ring size can have. The ceiling on a bond angle in a closed ring of equal bonds, 180°(n−2)/n, which is the interior angle of the regular planar polygon and follows from a closed curve having to turn through a full circle. The line at 109.47° is the tetrahedral angle: rings of 3, 4, 5 atoms cannot reach it at any geometry whatever, and every larger ring can, by leaving the plane.

The angle a ring cannot have

A closed ring of equal bonds has to turn through a full circle, so its bond angles cannot average more than 180°(n−2)/n. Three, four and five atoms are below the tetrahedral angle at every geometry whatever; six is above it, and reaches it only by leaving the plane.

A 6-ring at 111°: the twist-boat. A closed ring of 6 equal bonds meeting at 111°, drawn from the coordinates the closure conditions produce. Its torsions are 16.9°, -63.6°, 44.2° and repeat; its puckering amplitude is 0.508 bond lengths at a phase of 344°, with q₃ exactly zero, so it lies on the equator. Turning the ring changes none of those numbers, which is what makes them a description of the shape rather than of the view.

The ring that cannot hold still

Cyclohexane's chair is rigid and its boat is not, and that is a statement about the rank of a matrix rather than about strain. Hold every bond length and every bond angle fixed and count what is left: the chair has nothing, and the boat sits on a continuous loop of shapes with the same bonds and the same angles.

The bond sum, the measurement, and what is left over. For each pyramid: the vector sum of three bond moments estimated from the electronegativity difference and the measured geometry, the measured dipole, and the difference between them. Positive is towards the lone pair. Ammonia's bonds point that way and nitrogen trifluoride's point the other, which is why the trifluoride's far more polar bonds give it a dipole six times smaller. The leftover is 0.58 D for ammonia and at least 1.48 D for the trifluoride, so it is not one lone pair's property.

The lone pair is not the missing term

Ammonia's dipole is 1.47 debye and nitrogen trifluoride's is 0.235, although the N–F bonds are far more polar than the N–H ones. The bond sums explain the reversal exactly and point in opposite directions — and the lone pair that is supposed to make up the difference has to be worth 0.58 debye in one molecule and at least 1.48 in the other.

Which site a long bond takes, and it is not the one the rule says. The energy of the axial placement minus the energy of the equatorial one, against the odd bond's length. Above the line the equatorial placement wins and below it the axial one does. The textbook sentence is that a bulkier group goes equatorial where there is more room; given a longer bond and nothing else, the repulsion model puts it AXIAL, at every length, by a margin that grows. A site pushed further out interacts with everything less, so what is left to decide is the arrangement of the four sites left behind — and three equatorial with one axial is the better set of four.

The sites are not the same size

Every arrangement in this collection puts its sites on one sphere, which is an assumption about bond lengths made silently. Give the repulsion model a bond length and it predicts that the long bond goes axial — the opposite of the rule the model is always cited for.

$The twist the ring forces on the bond. For each ring size, the largest torsion about the double bond that the ring will close on — 180° being a flat trans arrangement and 90° being a π bond broken outright. The six-ring will not close at any torsion tested. The eight-ring reaches 139.3°, against 136° measured in trans-cyclooctene by diffraction: a model with bond lengths and bond angles in it and no energy anywhere agrees with the crystal to a few degrees.$

The double bond a ring cannot hold

A trans double bond needs a ring of nine carbons to sit flat, and the eight-ring will hold one twisted by 40.7 degrees — against 136 degrees measured in trans-cyclooctene. The same eight-ring threshold, applied by counting ring sizes, sorts nine bridgehead alkenes correctly with no bridgehead anywhere in the argument.

Face to face repels, edge to face attracts. Two benzene molecules 5 Å apart, one turned against the other, with only the quadrupole interaction between them. Stacked, the two negative faces meet and the energy is +8.8 kJ/mol; perpendicular, one molecule's positive rim meets the other's negative face and it is -4.6. The sign changes on the way, so the preference is not that one arrangement is weaker — it is that they are opposite. This is a molecule with no dipole moment at all.

Zero dipole is not no interaction

Benzene's dipole moment is exactly zero at every origin, and its quadrupole moment is large enough to decide a crystal structure. Two benzenes face to face repel by nine kilojoules a mole; edge to face they attract, and the sign changes on the way between.

$water, drawn with its nuclei the size they are. The molecule with a disc round each nucleus whose radius is the computed root-mean-square displacement of that nucleus in the vibrational ground state. The hydrogens' discs are a substantial fraction of the bond length, and this is at no temperature at all.$

The atoms are not at the points

Every structure in this collection is a set of points, and every bond angle it argues about is a property of that set. Computed from the fitted force fields it already has, water's hydrogens are 0.094 Å from where they are drawn and its bond angle has a spread of 8.9 degrees — larger than the difference between 104.5 and the tetrahedral value that half the essays here are about. This is at no temperature at all.

The dipole moment, and how many bands there are. For each of five molecules: the dipole moment of the point-charge model, the number of modes whose dipole derivative does not vanish, and the largest derivative. The molecules with no dipole at all have the most active bands, which is the whole of the argument.

A dipole is not what an infrared spectrum sees

Carbon dioxide has no dipole moment at all and three of its four modes are infrared active. Methane has none and six of nine; boron trifluoride none and five of six. Water, which has the largest dipole of the five, has three modes and three bands — and its dipole predicted neither number.

Flat rings: the angles, the torsions and what is measured. For each ring from three to eight, held flat: its interior angle, how far that is from tetrahedral, the angle strain that follows, the torsional strain of having every bond eclipsed, and the measured strain energy. The five-ring is the row the essay is about — its angles are almost ideal and it is strained.

The strain that is not in the angles

Cyclopentane's flat bond angles are 108°, a degree and a half from tetrahedral, and its angle strain computed from a standard bending constant is 0.4 kJ mol⁻¹. Its measured strain is twenty-six. The missing sixty kilojoules are torsional — one ethane barrier for every bond in the ring, which no account built on bond angles mentions.

The atoms that meet across a ring

Twelve closed conformers of a ten-membered ring at one bond angle, so every one has identical angle strain by construction. Two of them differ by 0.026 kilojoules a mole in torsional energy and by 0.80 ångström in how close two atoms on opposite sides of the ring come — 2.331 against 3.131, where two carbons are in contact at about 3.4. An account built from angles and torsions calls those two structures the same.

What boron trifluoride's bands are strong in, and what they move. Every infrared-active mode of boron trifluoride, with its band strength and the root-mean-square displacement of its atoms in the zero point, each scaled to its own largest. The two do not order the modes the same way — the rank correlation between them is 0.2 — and the strongest band belongs to the mode at 719 cm⁻¹, in which 89.93 per cent of the motion is the lightest atom's. Every mode moves the same weighted amount of mass, exactly, so that is not what separates them either.

The mode that moves least radiates most

Boron trifluoride's strongest infrared band is the one in which the fluorines barely move: ninety per cent of the motion belongs to the boron, which is a fifth of the molecule's mass. The mode that moves the most mass is nine and a half times weaker. Across five molecules the rank correlation between band strength and how far the atoms actually go runs from +1 to −0.66, and every normal mode carries exactly the same weighted motion by construction.

Three bent bonds, at a hundred and one degrees to each other. A carbon–carbon triple bond in its localised description: three equivalent bent bonds, spaced by 101.54 degrees, each tilted 63.43 degrees off the axis and each carrying 0.17 of an s orbital — an sp⁵ hybrid. Its charge sits 0.32 ångström off the axis, where every canonical orbital's sits on it. The mixing that makes the three equivalent is a rotation in the three-dimensional occupied space, so the density is untouched.

Three bent bonds, and the same hybrid

A triple bond localises into three bent bonds at 101.537 degrees to one another, each an sp⁵ hybrid with exactly one sixth s character. A double bond's two bent components are sp⁵ hybrids at 101.537 degrees. The two are the same hybrid, built out of different frameworks and in different numbers, and the reason is that a third of a half is a half of a third.

An effect that cannot distort a molecule, deciding how far it distorts. The energy along one distortion coordinate, three times. With only the first-order term the minimum is at 0.6; with only the second-order term there is no minimum away from zero at all, because the gap of 1.5 is above the critical 1 for a closed shell on its own. With both, the molecule distorts to 1.06 — well past the first-order answer — and gains 0.09 more than the two separate stabilisations add up to.

Two distortions in one coordinate

A second-order Jahn–Teller effect that cannot distort a molecule by itself — its gap is half again above the critical value — nearly doubles the distortion when a first-order effect is already acting. The molecule goes to 1.075 instead of 0.600 and gains twice the energy, and it does it while the gap the second-order term divides by is opening rather than closing.

The explanation with the wrong sign

Two methyl groups on one carbon make a ring easier to close, by up to eleven thousand-fold, and the textbook reason is that they compress the ring's internal angle. Computed from a standard bending term, that compression helps a three-ring and a four-ring and hinders every ring from five up — predicting a slowing of 0.754-fold for the five-ring measured to speed up 250-fold. The account with the right sign is about rotations rather than angles, and it has a ceiling of 36.5 that the measurement is already above.

The ceiling rises and the measurements fall, so they cross. The largest acceleration the rotamer account can produce, against the ring being closed, with the measured gem-dimethyl accelerations on the same axis. Closing a bigger ring means freezing more rotations, so the ceiling rises steeply; the measurements go the other way. The five-membered ring's 250-fold acceleration is above its own ceiling of 36.5 and the six-membered ring's tenfold one is far below its 121 — so the account is refused at one size and sufficient at the next.

A ceiling that rises where the measurements fall

The rotamer account of the gem-dimethyl effect has a largest possible acceleration, which looks like a limitation. It is a prediction: the ceiling is a closed form in the number of rotations a closure freezes, it rises steeply with ring size, and the measurements fall — so the account is refuted for the five-membered ring and more than sufficient for the six.

Four tables, one answer, and a reason it could not be otherwise. The intensity–motion correlation — infrared intensity against how far the atoms move — computed with charges from four published electronegativity tables. Every molecule gives the same number on all four, to machine precision, because each is made of two elements: its charges are one number times a fixed pattern, a change of table changes only that number, and a rank correlation does not notice a rescaling. The dipole moments beside them do notice, which is the check that the tables are genuinely different.

The table that could not have mattered

Charges taken from one of four electronegativity tables invite a worry, because the tables disagree with each other. They do disagree — hydrogen cyanide's dipole runs over a factor of eleven between them — and for the molecules in question the worry could not have applied, because a molecule of two elements has charges that are one number times a fixed pattern and a rank correlation does not notice a rescaling.

How far the rotor count would have to be wrong. The ceiling against the number of rotations a closure freezes, with the two measured accelerations drawn across it. The five-membered closure freezes three and its ceiling is 36.46; the ceiling does not reach the measured 250 until 5 rotors, so the count would have to be wrong by 2 on a ring that has three rotations to freeze. The six-membered closure freezes four at a ceiling of 120.88, and stays above its measured 10 down to 2 — so the refusal is airtight and the sufficiency is comfortable.

An estimate that can be wrong by two

The ceiling on the gem-dimethyl effect is exponential in the number of rotations a closure freezes, and that number was taken as n − 2 without counting — which left the refutation at five rings probable rather than airtight. It is airtight. The ceiling does not reach the measured 250 until five rotors, on a ring that has three, and no hindering of the tether can raise it.

The two rows of a table, and the curve between them. The rotamer ceiling of a closure freezing 4 rotations, against how hard each of a stated number of them is to turn. The top curve is the free-rotor row and the right-hand end of the bottom curve is the absent-rotor row; everything else is what a real hindered tether does. The measured acceleration of 10-fold is drawn across it. It falls below that line when 3 of the 4 are hindered by 4.09 kJ/mol, or when 4 of the 4 are hindered by 2.70 kJ/mol.

The curve between two rows

A rotation a ring closure has to freeze was treated as free or as absent, and the two answers sat in adjacent rows of a table. A real hindered rotor is neither. The factor one rotor contributes runs from 3.316 to one along a curve nobody had drawn, and where it matters is between one and eight kilojoules a mole — which is a torsional barrier rather than a conformational preference.

Twelve hydrogen bonds, four tables. The bond dipole of hydrogen against each partner, in debye, on each of the four tables after all four are anchored to the same hydrogen–fluorine separation. Positive is hydrogen at the positive end. A bond whose marks straddle the axis is one the tables disagree about the direction of, and there are 3 of them.

Four tables and one molecule to disagree about

A molecule of two elements is provably safe from the choice of electronegativity table, and a series down a group should land where the tables do disagree. It does. All four agree that hydrogen iodide is the exception — and they disagree about what its dipole is by 1.245 debye, which is two and a half times the 0.448 that was measured.

The two sides of a verdict, with error bars. The computed ceiling for the six-membered closure — 10.9945 — with the band a gauche energy of 3.8 ± 0.4 kJ/mol puts on it, against the measured tenfold rate ratio with an assumed 20 per cent uncertainty. The two bands overlap over most of their length, and the nine per cent margin the verdict was decided by sits inside both of them.

A verdict inside its own error bar

The tightest comparison in the rotamer argument is a computed 10.99 against a measured 10 — a nine per cent margin, offered as a verdict. The computed side is built on one quoted energy known to ±0.4 kJ/mol, and that alone puts a band of twenty-three per cent on it. The verdict turns on four tenths of one standard deviation of a number nobody had put an error bar on.

$One number separates the disputed bonds from the agreed ones. Every bond in the collection, ordered by how big its electronegativity difference is as a fraction of each table's range. The six the tables disagree about are all below 0.0507; the 25 they agree about are all above 0.0584. Nothing lies between, and which elements a bond joins does not enter — a dispute is what happens when the difference is small enough that the tables' own disagreement about it is larger.$

A dispute is a small difference

Four electronegativity tables disagree about the direction of a quarter of a set of hydrogen bonds, and the interhalogens look like the natural test: five compounds with no hydrogen in them, to say whether the disagreement is about one awkward element or about the whole idea. All four tables agree about every interhalogen. They agree because the differences are large, and that is not a fact about halogens.

The measurement is not a horizontal line. The rotamer ceiling against temperature, with the measured tenfold drawn four ways: flat, and as a rate ratio whose two activation energies differ by 3, 5.31 and 8 kJ/mol. Flat, it is crossed at 312.1 K, which is the flat-rotor answer. At 5.31 kJ/mol the two curves are parallel and never meet. At 8 they meet on the other side, and it is cooling rather than heating that refutes the account.

One number decides which way it breaks

The six-membered verdict gives way at 312.1 K, and that is a statement about the ceiling rather than a prediction about the ratio — because a real rate ratio has a temperature dependence the model has no term for. Putting that term in moves the crossing, and at 5.308 kJ/mol it removes it entirely.

The gap closed when the other hundred and twenty-two pairs arrived. Every pair sorted by the size of its normalised electronegativity difference: the 31 common bonds on the left, all 153 pairs the four tables cover on the right. Filled marks are pairs the four tables put on different sides of zero. On the small set they are the 6 smallest with nothing in between; on the full set 27 uncontested pairs sit below the largest contested one.

A gap that was a choice of bonds

One number separates every disputed bond from every agreed one with nothing in between — on thirty-one bonds, which is enough to see a gap and not enough to know it is real. It is not real. The line is in the same place on all one hundred and fifty-three pairs the four tables cover, and twenty-seven agreed pairs now sit below the largest disputed one.

The five-membered ring would need more rotors than its chain has bonds. For each measured acceleration, the least rotor count whose ceiling reaches it, against the most rotors any convention gives that ring and against the number of bonds the chain has at all. The five-membered cases need five and eight rotors from a chain with four bonds, so they are refused by supply rather than by a margin — a refutation that needs no convention to be chosen, because every convention is below the requirement.

The count that was never written down

The rotamer ceiling for a ring closure needs a count of frozen rotations, and the counts in use disagree in direction: one gives the six-membered ring more frozen rotations than the five, another fewer. Every measured verdict survives all three defensible conventions unchanged — and the margins move by a factor of eleven, which is why a verdict decided by ten per cent was worth being nervous about.

Two quantities, and the best straight line between the two kinds of pair. Every pair the four electronegativity tables all cover, by how much the two elements differ and by how much the tables disagree about them. Filled points are pairs whose sign the tables dispute. The line is the boundary that misclassifies fewest — 3 of 153, against 6 for the best rule using the difference alone. It slopes upward, which is the mechanism: more dispute buys a larger difference and still leaves the sign in doubt.

Two numbers caught what one could not

No single threshold on the electronegativity difference separates the bond pairs whose polarity the four tables dispute from those they agree on — the best misses six of a hundred and fifty-three. Adding how much the tables disagree halves that to three and catches every disputed pair. And three is what four coin-flips produce: the rule flags nineteen pairs, an eighth of which should look agreed for no reason at all, which is 2.375 against the three observed.

NH₃: four states in a well the molecule does not sit at the bottom of. The umbrella coordinate of NH₃ — the signed distance of the N atom from the plane of its three H atoms — with the quartic well that has its minima at the measured 0.3816 ångström and its barrier at the quoted 2020 wavenumbers. The lowest 4 states are drawn at their computed energies. The lowest sits 587 wavenumbers above the bottom, which is 29.1 per cent of the way up the barrier, so the state is far from the harmonic bottom that a drawing of a pyramid implies.

A barrier is not what a splitting measures

Ammonia's inversion barrier is quoted everywhere as 2020 wavenumbers. Put that number into the simplest double well its own measured geometry allows and the ground-state splitting comes out at 1.3508 against a measured 0.7935, and the excited one at 68.37 against 35.81. Both are too large because a splitting is an area under a barrier and a height is only one of its two dimensions.

One per cent on the barrier is 3.6 per cent on the splitting. The ground inversion splitting of NH₃'s quartic well against the barrier height, both logarithmic, with the geometry and the reduced mass held at their measured values. The curve is visibly bent: its local slope is -3.56 at the published barrier and steepens either side, so a power law is a tangent to it rather than a description of it. The measured 0.7935 wavenumbers is reached at 2330, which is 15.3 per cent above the quoted 2020 — so a splitting wrong by a factor of 1.70 is a barrier wrong by a sixth. The same derivative read the other way is what makes a barrier quoted to ten per cent useless for predicting a splitting.

The exponent that runs both ways

How hard does a splitting depend on a barrier? Locally, as the power −3.5628 — and the local slope runs from −2.53 to −6.15 across the same sweep, so there is no power law. What is exact is stranger: rescaling the equation forces the mass exponent to be one below the barrier's and the geometry exponent to be twice the mass's, so the model's three sensitivities are one number and the arithmetic reproduces both identities to six decimals.

The mass is worth a factor of 1.6, and the other two 1e+4 and 9e+4. Ammonia's umbrella well, with each of phosphine's three differences substituted into it one at a time and then all together. The reduced mass is 11 per cent larger and costs a factor of 1.61. The pyramid is 2.01 times taller and costs 9.6e+3; the barrier is 6.1 times higher and costs 9.4e+4. Phosphine's own splitting is below what the arithmetic resolves, so it is drawn at that bound.

It was never the mass

Phosphine does not invert, and the reason given is that phosphorus is heavier than nitrogen. Three things about phosphine differ from ammonia. Substituting each into ammonia's own well one at a time, the reduced mass costs a factor of 1.61, the pyramid height costs 9,600 and the barrier 94,000 — and the mass is the smallest of the three by four orders of magnitude.

Three constructions of one number, spanning a factor of 5.6. The reduced mass of NH₃'s umbrella coordinate under each construction, against position along the coordinate. Two of the three are constants and the third is not: if the bonds are held at their measured length, the ligands must slide outward as the apex descends, and their radial motion adds to the mass. It runs from 2.4866 at the plane to 2.9874 at the pyramid — 20 per cent — and the coordinate itself stops existing at one bond length, which is where the curve ends.

The mass nobody chose

Every one-dimensional treatment of ammonia's inversion needs a mass, and the measurement does not supply one. Three constructions are defensible and they give 1.35075, 0.94420 and 0.0000055 wavenumbers. The honest one is not a constant at all, and it moves the answer thirty per cent towards the measurement — which means the usual choice is the wrong one.

The ceiling moves by 8.7-fold across the gauche energy's own reported range. The rotamer ceiling against butane's gauche energy, for the rotor counts each convention assigns to a five-membered and a six-membered closure, with the three measured accelerations drawn as horizontal lines. Across the reported range the ceiling moves by up to 8.70-fold, and two of the nine verdicts cross a measurement during the sweep. The natural prediction is that this input would be the smaller lever of the two; it is the larger.

The lever that was supposed to be smaller

Sweeping the rotor conventions leaves every verdict unmoved, and suggests that the other quoted input will be a smaller lever. It is a larger one. Sweeping butane's gauche energy over its reported range moves the ceiling by 8.7-fold, flips the six-membered verdict at 3.630 kilojoules a mole — inside the quoted error bar — and takes the five-membered refutation with it at 4.422.

Two sweeps, and one of them covers 28 per cent less ground than the other. What each of the two sweeps reaches, measured in the one variable the rotamer ceiling has: g/RT. The temperature sweep lies entirely inside the gauche sweep, so it explored no arrangement the other does not. Their widths are 0.581 and 0.807, their union is 0.807, and treating them as independent would have credited them with 1.387.

Two sweeps and one lever

The rotamer ceiling depends on the gauche energy and on the temperature only through their ratio, so a sweep of either traces the same curve. Measured in that one variable, the temperature sweep lies entirely inside the gauche sweep — it covered no ground the other does not, and the whole reported spread in one energy is worth a temperature swing from 196 to 353 kelvin.

The orderings in use move the splitting by 0.47 per cent between them. The change in NH₃'s ground inversion splitting under each ordering of the kinetic operator, relative to BenDaniel–Duke, with the bond-conserving mass throughout. The bars are exact solves and the ticks are first-order perturbation theory. The five span 0.469 per cent, from −0.407 to 0.060; the change from a constant mass to the bond-conserving one, in the same well and box, is 43.1 per cent, 92 times as large.

An ordering worth half a per cent

A mass that varies along a coordinate has no unique quantum kinetic energy, and the choice among the Hermitian orderings in use was the one thing left that could undo a forty-three per cent correction to ammonia's splitting. It cannot. The five orderings anybody uses span 0.47 per cent between them, a ninety-second of the correction, and the family only reaches the measurement at exponents three times larger than any of them.

Every prediction of the isotope ratio overshoots, and the mass decides nothing. The ratio of NH₃'s ground inversion splitting to ND₃'s, predicted six ways, against the measured 14.94. At the published barrier the usual mass gives 15.77 and the bond-conserving mass 17.79. With a quartic fitted to NH₃'s splitting they give 19.26 and 19.00; with a well whose shape is fitted to both of NH₃'s lines, 17.62 and 18.48. The bond-conserving mass is nearer the measurement in one of the three pairs and further in two, and the difference within any pair is smaller than the distance of either from the measurement.

Deuterium cannot tell the masses apart

A reduced mass built by holding ammonia's bonds rigid predicts a deuterated molecule differently from any constant mass, and ND₃'s splitting is measured. Run as a test, it cannot choose. Every well and every mass needs a barrier for ND₃ several per cent lower than for NH₃, every prediction of the isotope ratio from a well fitted to NH₃ overshoots by eighteen to twenty-nine per cent, and the two masses differ by less than either misses — in opposite directions in the two wells.

The term two of the four molecules have and two do not. The radial coefficient of the bond-conserving reduced mass for the four isotopologues: the sum of the ligand masses, and what is left after the asymmetric correction. Three ligands at a hundred and twenty degrees on a circle whose radius changes as the apex descends move their own horizontal centre of mass outward — unless the three masses are equal. A frame that does not translate has to subtract that motion, and the amount is half the sum of the squared mass differences over the total mass. It is zero at both ends of the series and the same number in the middle.

The two that are not on the line

Ammonia and its fully deuterated twin are two points, and two points cannot show a curve. Putting the partly deuterated molecules between them needs a term neither symmetric one has — three ligands of unequal mass move their own centre of mass sideways as the apex descends — and it moves the prediction by half a per cent, which is what a whole change of mass construction was worth.

The covariant operator is BenDaniel–Duke plus this. The difference between the Laplace–Beltrami operator — the one a one-dimensional manifold with metric μ(x) distinguishes, carried across to the flat measure by the unitary map ψ ↦ μ^(¼)ψ — and the BenDaniel–Duke ordering, divided by the function it was applied to, at forty-one positions inside the molecule's own range. The curve drawn through the marks is the two-function fit every ordering is a combination of, and it passes through them to a part in ten million.

The ordering a manifold picks

A position-dependent mass leaves the kinetic energy with no unique quantum form, and an earlier sweep of the five orderings in use found half a per cent between them. A one-dimensional reduction is a one-dimensional manifold, a manifold has a distinguished Laplacian, and carrying it to the flat measure lands on exactly one of those five — not the one with no extra potential, and not the one anybody reaches for.

A long bond goes axial at five and equatorial at seven. The energy of the odd site placed axially minus placed equatorially, against the odd bond's length relative to the others, for the trigonal bipyramid and the pentagonal bipyramid. Above zero the equatorial site is preferred. At five a longer bond goes axial and a shorter one equatorial; at seven, from 0.7 to 1.3, every sign is reversed — a shorter bond axial, a longer one equatorial. An open marker is a placement that is not a minimum: at seven the site each bond avoids is one it would slide out of.

The long bond goes to the crowded site

Given one bond longer than the others, the repulsion model puts it axial in a trigonal bipyramid — against the rule it is usually cited for. The reason is a crowding count, and at seven sites the count reverses: the pentagonal bipyramid's crowded site is equatorial, so a long bond goes equatorial and a short one axial. PF₅'s long bonds are axial and IF₇'s are equatorial. And at seven the site a bond avoids is not even a minimum.

All essays