Why focusing does not always remove a coloured edge

A bright detail against a dark background has a bluish or reddish fringe. Turn the focuser: one shade sharpens while another spreads out. The objective may be bringing different wavelengths to different positions. Moving an eye or sensor selects a position; it does not change the properties of the glass.

An achromatic objective combines the dispersive contributions of its elements to bring selected colour foci closer together. Apochromatic correction places stricter demands on their agreement and on other errors. These are properties of an optical system; the number of lenses or a marketing word on the cell cannot establish the result by itself. Edmund Optics: chromatic focal shift.

We will trace the mechanism with specific SCHOTT datasheet values, then distinguish it from real thicknesses, cells and tests. The numbers are calculations, not measurements of a finished telescope.

From an early objective to a documented public report

The history of the achromat resists a single “first inventor”. In a letter dated 4 June 1789, Addison Smith told Jesse Ramsden how Chester Moore Hall had shown him a compound crown-and-flint objective in 1763. According to Smith, Hall recalled making it around 1733. This is valuable primary testimony about an object the writer saw, but its early date is a later recollection, not a record written in 1733. Royal Society: Smith to Ramsden, L&P/9/139, pp. 1–3.

The Royal Society archive separately records the reading of John Dollond’s paper on different refrangibility on 8 June 1758. That date distinguishes a documented communication from earlier private work; the digits 1757 in the associated DOI do not replace the reading date. Royal Society: Dollond manuscript, L&P/3/324.

A later advance required new glasses and more precise optical design. The ZEISS archival chronology dates Ernst Abbe’s apochromatic objectives to 1886, in microscopy and collaboration with Schott. This is not the invention date of every modern astronomical design. ZEISS: Abbe inventions, 1886 entry.

Glass does not have one index for every colour

The refractive index n describes an optical property of a medium and depends on wavelength. Across the visible samples shown here, N-BK7 and N-F2 have higher n at shorter wavelengths. For unchanged geometry, a simple positive lens therefore converges those wavelengths more strongly and focuses them nearer the lens. This is longitudinal chromatic aberration.

Two glass types: index depends on wavelength
1.501.551.601.65400500600700n(λ)

Wavelength λ, nm

━ N-BK7 ┄ N-F2

h: 404.7 nm
N-BK7: 1.53024N-F2: 1.65087
g: 435.8 nm
N-BK7: 1.52668N-F2: 1.64209
F: 486.1 nm
N-BK7: 1.52238N-F2: 1.63208
e: 546.1 nm
N-BK7: 1.51872N-F2: 1.62408
d: 587.6 nm
N-BK7: 1.51680N-F2: 1.62005
C: 656.3 nm
N-BK7: 1.51432N-F2: 1.61506
r: 706.5 nm
N-BK7: 1.51289N-F2: 1.61229

Nominal indices relative to air from SCHOTT datasheets: N-BK7, 01-Dec-2023; N-F2, 01-Feb-2014. Points are shown; straight joins are reading guides, not an interpolation formula. These are not batch measurements. Index is dimensionless; the vertical scale does not start at zero.

The two curves represent particular glass types, not every “crown” or “flint”. Optical design needs more than n near the middle of a range: how n changes between wavelengths matters too. SCHOTT: N-BK7, SCHOTT: N-F2.

Branch · Glass propertiesOptional · you can skip this

What the Abbe number compares

A common definition uses the spectral lines d, F and C:

V_(d) = (n_(d) − 1)/(n_(F) − n_(C)).

d is near 587.6 nm, F at 486.1 nm and C at 656.3 nm. The number is dimensionless. A larger V_(d) means smaller principal dispersion relative to n_(d) − 1, rather than an automatically better finished lens. In the selected datasheets, N-BK7 has n_(d) = 1.51680, V_(d) = 64.17; N-F2 has n_(d) = 1.62005, V_(d) = 36.43. SCHOTT TIE-29, §2.1.

The tabulated indices are rounded to five decimal places. Subtracting those F and C values for N-BK7 gives an Abbe number near 64.12 instead of the listed 64.17: principal dispersion and V are separately reported at different precision. That limits the precision of our educational calculation; it is no reason to adjust the data to force agreement. Actual production also needs the properties of the particular glass batch.

A diverging lens can improve a converging objective

Take a positive element with lower dispersion and a negative element with higher dispersion. The first provides the overall convergence. The second subtracts some optical power, but its colour-dependent difference has the opposite sign as well. Choosing the powers can reduce the separation between foci while keeping the whole pair converging.

This takes more than putting two good glasses together. Selecting high and low n is insufficient: dispersions, powers and geometry matter. A doublet has two elements, which may be cemented or air-spaced. A triplet adds another element and design freedoms, along with surfaces and assembly requirements. Neither count alone proves the resulting correction.

Branch · Optical powerOptional · you can skip this

Two conditions for a thin contacting pair

Let φ = 1/f denote optical power; with lengths in mm, its unit is mm⁻¹. For two thin lenses in contact in air:

φ = φ_(1) + φ_(2), (φ_(1))/(V_(1)) + (φ_(2))/(V_(2)) ≈ 0.

The first condition sets the total focal length. The second compensates the F–C power difference to first order. At fixed curvatures, the power of each thin element is proportional to n − 1, so its F–C difference equals its d-line power divided by V. One power must be negative. Edmund Optics: compensation with two elements.

Take f_(d) = 500 mm and the datasheet values V_(1) = 64.17, V_(2) = 36.43. The solution is:

φ_(1) = (V_(1))/(V_(1) − V_(2))(1)/(f_(d)) ≈ 0.00462653 (mm)^( − 1), φ_(2) ≈ − 0.00262653 (mm)^( − 1).

Their sum is 0.002 (mm)^( − 1), corresponding to 500 mm. The separate focal lengths are about +216.14 and −380.73 mm. This checks a principle, rather than specifying real surface radii: thicknesses, air gaps, asphericity, cementing and spherical aberration are not calculated here.

An achromat brings selected colours closer without erasing wavelength dependence

For a single thin N-BK7 lens with a 500 mm d-line focus, our model gives about 494.66 mm at F and 502.41 mm at C. The contacting pair with the same total focal length gives about 500.25 and 500.27 mm. Their separation is greatly reduced, but the intermediate d line remains near 500.00 mm. The entire range does not share a perfectly common focus.

The colour foci move closer, with a residual difference

One thin N-BK7 lens

490495500505
F (486.1 nm)
Focus, mm: 494.66
d (587.6 nm)
Focus, mm: 500.00
C (656.3 nm)
Focus, mm: 502.41

Thin contacting N-BK7/N-F2 pair

490495500505
F (486.1 nm)
Focus, mm: 500.25
d (587.6 nm)
Focus, mm: 500.00
C (656.3 nm)
Focus, mm: 500.27

Distance x from the equivalent thin lens, mm

━ F┄ d┈ C

On-axis source at infinity; selected input heights ±10 mm. Equivalent lens at x=0 is outside the enlarged region; f_d=500 mm. Both panels share x and y scales; transverse scale is exaggerated. Paraxial rays omit diffraction and spherical aberration. Rounded datasheet n and V leave a small nonzero F/C difference; the middle d line also remains separate.

The residual colour dependence of focus after this correction is associated with the secondary spectrum. One Abbe number cannot describe the shape of the whole dispersion curve. Relative partial dispersions and materials with suitable departures from ordinary relationships become important. SCHOTT TIE-29, §2.2.

Branch · Limits of the numerical modelOptional · you can skip this

From datasheet indices to ray crossings

The model keeps curvatures fixed, so for each element:

φ_(i)(λ) = φ_(i)(d)(n_(i)(λ) − 1)/(n_(i)(d) − 1), f(λ) = (1)/(φ_(1)(λ) + φ_(2)(λ)).

For an incoming axial ray at height y₀ in the thin-lens plane, the outgoing path is y(x) = y_(0)[1 − x/f(λ)]. At x = f its height is zero: an independent check of the focal position. The diagram uses two rays at y_(0) = ± 10 mm and enlarges the region near their crossing; the physical lens remains far outside the frame to the left.

The small corrected F/C difference is left visible: datasheet V and n are rounded, and the thin-pair condition is approximate. These data do not describe spherochromatism, the change of spherical aberration with colour, nor cell defects, off-axis fields or a diffraction spot. They cannot establish the finished quality or tolerance of a particular telescope.

APO, ED and triplet answer different questions

In the classical definition, an apochromatic objective corrects axial chromatic aberration at three wavelengths and spherical aberration at two. Nikon uses this definition in its microscopy glossary. It describes a level of correction, not a property every triplet automatically possesses. For a telescope, the full working range, aperture, field and actual system characteristics matter. Nikon: apochromat objective.

ED stands for extra-low dispersion. It describes glass in a manufacturer’s terminology, rather than providing a complete objective test report. Even Celestron notes the lack of a single official ED definition in its explanation. We need the particular glass type, its dispersion and its role in the full system. Celestron: ED Glass.

Element count is not a shortcut test either. For example, Sky-Watcher describes its particular Evostar 120EDX as a two-element air-spaced apochromat using FPL 53. This verifies the manufacturer’s design description and label; we have not independently tested the product. A material with unusual partial dispersion may offer another pair more design scope than ordinary crown and flint, but the result still depends on optical design. Sky-Watcher: Evostar 120EDX construction.

The dispersion calculation must become glass and a working cell

A real objective has surface radii, thicknesses, axial gaps and element orientations. Blanks are shaped and polished; form, thickness and centring are checked, required coatings applied, and the assembled group inspected. The optical axis defined by the surfaces must be aligned appropriately with the mechanical axis. A good material pairing cannot ensure this on its own. Edmund Optics: thickness, radii and centring.

The gap and cell are part of the design too
12345
  1. Positive element
  2. Air gap and spacer near the edge
  3. Negative element
  4. Housing with supporting surfaces
  5. Retaining ring outside the working opening

Hypothetical air-spaced design, separate from the numerical contacting model above. Curvatures, thicknesses, gap and seats have no dimensional scale; no tolerances are specified. Light would travel left to right; the axis is dashed. This is not a recipe for clamping or dismantling an objective yourself.

An antireflection coating reduces unwanted reflections but does not replace dispersion compensation. The cell must maintain the required gaps and positions. Temperature changes glass and housing dimensions as well as refractive index; athermalization requires separately matching these contributions. An optical achromat is not automatically athermal. Edmund Optics: thermal defocus and housing.

Joints and additional materials must be assessed for the particular construction. For example, Edmund describes a polymer aspheric layer and a working range of −20…+80°C for its hybrid Aspherized Achromatic Lenses. This is a specified product, not a temperature rule for all doublets or evidence that every objective comes uncemented. Edmund Optics: hybrid construction and its limits.

Test the result in the relevant light and field

For broadband visual observation or imaging, the system’s simultaneous behaviour across the required colours matters. A narrow spectral band poses a different task: the separation between widely spaced wavelength foci then has less effect on that particular image. A colour-corrected centre also does not prove a flat field, proper edge performance or freedom from spherical aberration.

A useful report states wavelengths, aperture and focal ratio, field position, sensor plane and assessment method. Look beyond a coloured fringe: longitudinal focal shift, transverse colour displacement, spot size and contrast transfer describe different effects. One attractive photograph cannot replace all these conditions. Edmund Optics: aberrations and image characteristics.

Try explaining three cases: why a weaker negative lens can improve a positive objective; why close F and C foci do not guarantee the same focus at d; why ED and the number “three” cannot establish sharpness. An answer that distinguishes material, design and test result captures the mechanism.

The article on optical manufacturing connects material and precision; the comparison of Galilean and Keplerian telescopes explains lens signs. Revisit image formation in “How does a telescope work?”. The course will next use these distinctions for a documented educational build with ready-made lenses. The calculations here are not physical fabrication or testing of an achromat or apochromat.