xraytools.

Moseley Plot and Absorption Edges

Powder or single crystal

Where an element’s characteristic X-ray lines fall, from the one relation that connects them all: the square root of the line energy is linear in atomic number. The straight line is fitted to this site’s own six anodes, so you can see what it is worth before you use it.

You supply
An element symbol (Ni) or an atomic number (28). Nothing else — a characteristic line is a property of the element alone.
Reading it
A predicted line is not a measured one: the fit is good to about a per cent from roughly Z = 20 to 50 and drifts several per cent by tungsten. Where this site holds the measurement it is printed beside the prediction. Moseley’s law fits the characteristic lines and predicts no absorption edge: the edge energies here are tabulated, and the bracket between two elements is a check on them rather than a substitute for them.

Worked examples: nickel, the filter for copper · zirconium, the filter for molybdenum · tungsten, far outside the fit

Earlier on the path: X-ray Tube Spectrum Next on the path: Absorption Coefficient Calculator On Choosing the radiation, step 2 of 4

See also: X-ray Tube Spectrum · Absorption Coefficient Calculator

Teaching with this page
Objective
After this page a learner can name the filter for an anode and say what makes that element and no other the right one.
Start from
this worked example
Ask first
Nickel is the filter for a copper tube. Would nickel also filter a molybdenum tube?
Watch for
“Yes — nickel absorbs X-rays whatever produced them”
Then
Absorption Coefficient Calculator
Check yourself: Nickel is the filter for a copper tube. Would nickel also filter a molybdenum tube?

No — a filter is chosen for one anode, not for X-rays in general Yes — nickel absorbs X-rays whatever produced them

A filter works because its K absorption edge falls between the anode’s two lines. Nickel’s edge is at 8.33 keV, which sits between copper’s Kβ at 8.91 and Kα at 8.04 — so Kβ is on the strongly absorbed side and Kα is not. Molybdenum’s lines are at 19.61 and 17.44 keV, both far above nickel’s edge, where nickel absorbs the two almost equally and improves nothing. Molybdenum’s filter is zirconium, edge 18.00 keV.

Input

Element

Either form: Ni or 28. The six elements this site carries as anodes — Cr, Fe, Co, Cu, Mo, Ag — also print their measured wavelengths, so you can see what the fit is worth.

Predicted lines for Cu (Z = 29)

Lines predicted for this element, against the wavelength measured for it. CSV
Line E / keV λ / Å measured λ / Å difference
Kα 8.05950 1.53836 1.541838 −0.23%
Kα1 8.06803 1.53673 1.540562 −0.25%
Kα2 8.04250 1.54161 1.544390 −0.18%
Kβ 8.93469 1.38767 1.392218 −0.33%

The straight line is fitted to this site’s own six anodes, Z = 24 to 47, and reproduces all of their measured wavelengths to better than 0.16% in E. It stays inside about 1% from roughly Z = 20 to Z = 50. Further out it is an extrapolation and the error grows: E against Z is not exactly straight, because screening is incomplete and relativistic corrections rise as Z4, so by tungsten (Z = 74) the predicted Kα is several per cent low. A predicted line is not a measured one; where this site has the measurement it is printed beside the prediction.

Moseley’s law

12340Cr24Fe26Co27Cu29Mo42Ag47Cr Z=24, √E = 2.3263 √keVFe Z=26, √E = 2.5298 √keVCo Z=27, √E = 2.6316 √keVCu Z=29, √E = 2.8357 √keVMo Z=42, √E = 4.1767 √keVAg Z=47, √E = 4.7017 √keVZ=29, predicted √E = 2.8389 √keVatomic number Z√(E / keV)

Filled points are the measured Kα energies of the six anodes; the dashed line is the least-squares fit through them.

Every anode in the fit, measured against what the law predicts. CSV
Anode Z measured λ / Å E/keV fitted E residual
Cr 24 2.29100 2.32633 2.32268 0.157%
Fe 26 1.937355 2.52976 2.52918 0.023%
Co 27 1.790260 2.63163 2.63243 0.030%
Cu 29 1.541838 2.83572 2.83893 0.113%
Mo 42 0.710730 4.17668 4.18116 0.107%
Ag 47 0.560871 4.70166 4.69741 0.091%

Bohr’s model predicts m = R∞⁢h⁢c×3/4 = 0.10102 keV and σ = 1 for Kα — one 1s electron screening the other. Fitted over the four light anodes alone (Cr to Cu), Kα gives m = 0.10188 and σ = 1.17, close to the N − 1 Moseley wrote. Including Mo and Ag bends the line and pulls σ to 1.50 — the value the fit above this table used. Nothing is wrong with either fit: the law is an excellent approximation over a limited span of Z, not an identity. Moseley’s own K series ran from aluminium to silver, his L series from zirconium to gold (Phil. Mag. 27, 703, 1914), and he noted that the lines of very short wavelength “gradually diverge” from the straight line — the bend that Mo and Ag put into the fit above.

How this is calculated

Moseley’s law: the square root of the photon energy of a characteristic line is linear in atomic number.E=m⁢(Z−σ)

Fitted by least squares to the six anodes this site carries, for the Kα line:m=0.103249keV,σ=1.50407

E=0.103249⁢(29−1.50407)=2.83893keV

E=(2.83893)2=8.05950keV

λ=h⁢cE=12.3984keVÅ8.05950keV=1.53836Å

The K absorption edge

Walk up the periodic table at a fixed wavelength and μ/ρ climbs steadily — more electrons, more absorption. At each of these six Kα1 wavelengths it does that everywhere except at one place, where it falls by a factor near seven between neighbouring elements. That is the K absorption edge, and the direction is worth stating in the variable that is moving. Walking up Z at a fixed wavelength: for the lighter element the edge lies below the photon’s energy, so the photon can eject a 1s electron and the K shell contributes; one element heavier the edge has risen past it, the photon can no longer reach the 1s electron, and the K shell stops absorbing altogether. Said in energy at a fixed element, which is how an edge is usually quoted, that is the same fact the other way round: a photon above the edge energy ionises the K shell and one below it cannot.

Each radiation’s Kα1 line, and the two K edges its energy falls between. CSV
Radiation Z Kα1 / Å E / keV edge lies between their K edges / keV μ/ρ either side drop
Ag Kα1 47 0.5594075 22.163 Ru (44) and Rh (45) 22.117 < 22.163 < 23.220 66.0996 → 11.9334 ×5.5
Mo Kα1 42 0.709300 17.480 Y (39) and Zr (40) 17.038 < 17.480 < 17.998 97.2508 → 15.57 ×6.2
Cu Kα1 29 1.540562 8.048 Co (27) and Ni (28) 7.709 < 8.048 < 8.333 322.096 → 46.925 ×6.9
Co Kα1 27 1.788965 6.930 Mn (25) and Fe (26) 6.539 < 6.930 < 7.112 392.229 → 53.8726 ×7.3
Fe Kα1 26 1.936042 6.404 Cr (24) and Mn (25) 5.989 < 6.404 < 6.539 443.942 → 57.5063 ×7.7
Cr Kα1 24 2.28970 5.415 Ti (22) and V (23) 4.966 < 5.415 < 5.465 557.253 → 70.0046 ×8.0

The drop is found in the coefficients and the edge energies are tabulated separately, so the last two columns are a check rather than a restatement: the Kα1 photon energy has to fall between the two edges named, and it does, on every row. The rows are drawn for Kα1 and not for the weighted mean of the doublet because the mean is an energy no photon has, and Kα1 is the more energetic component: an edge above it is above Kα2 as well. Kα2 lies between the same two edges on every row except Ag, where the K edge of Ru at 22.117 keV lies between Kα2 (21.991 keV) and Kα1 (22.163 keV) — so that element absorbs one component of the doublet through its K shell and not the other. Nothing here is predicted — the Moseley fit above is over emission lines, and it is not continued to an absorption edge.

A Kβ filter is a foil whose own K edge lies between the anode’s Kα and Kβ: Kβ is the more energetic, so it is absorbed by ejecting the foil’s 1s electrons, while Kα falls short of the edge and passes much less attenuated. The jump above names the lightest element whose edge clears Kα1, and with it the whole doublet. The conventional filter is that element or the next one heavier — nickel for copper, zirconium for molybdenum, rhodium or palladium for silver. Ag is the case that shows why the rule has to be stated for Kα1 and not for the mean of the doublet: Ru’s K edge is 126 eV above Kα2 and 46 eV below Kα1, so a foil of it would pass the weaker component and absorb the stronger. The lightest element that clears both is Rh.

Where this comes from