This article is also available in Japanese.

A few years ago the manufacturers of optical glasses started the process of reformulating their glasses to remove arsenic and lead. The driving forces for this programme were environmental and the health issues at various manufacturing stages. To most end users there are few differences between the old and new glasses [and I’m sure no photographer ever suffered from lens poisoning!]

At first glance, most of these glasses appear to be the exact equivalents of the original glasses – for example Schott N-F2 seems to be the same as F2 – after all, they do have the same catalogue glass number - 620364 – which suggests that the refractive indices at the F, d and C wavelengths are the same. Indeed they are; to about 1 or 2 in the 5th decimal place. And that is where ‘exact equivalence’ ends! However, examining the indices in the near UV and the NIR shows that the indices can differ in the 3rd decimal place.

Furthermore, many other properties of the glasses differ significantly: the new formula glasses absorb earlier at each end of the spectrum, and there are massive differences in the thermal properties between the old and new formulations: for example, the dn/dT for Schott SF14 and N-SF14, at the e-wavelength for 20 to 40°C, are 9.4 and 0.4 respectively!

I am sure there are many other properties that will be found to differ between the old and the new formula glasses.

The ZEMAX manual does carry a warning under the heading Obsolete Catalog Data that states: "Because some new glasses may have the same name as old glasses, although the exact composition may have changed, optical engineers need to be especially vigilant about checking the index data predicted by the software against the melt sheets of the glass which will actually be used." How true!

The Schott web site does make a statement along the lines that ‘ if only arsenic has been removed then the new glasses can be considered exact equivalents’. Considering that the majority of glasses have also had lead removed, in my opinion a more appropriate statement would have been ‘the majority of new glasses cannot be considered as exact equivalents of the glasses they replace’.

You will notice that ZEMAX refers to 'obsolete' glasses in the glass catalogs. However, a publication by Schott shows that several of the glasses will not necessarily become obsolete and will be melted in both old and lead-free forms. And, just to ‘help’ matters they are introducing HT (High Transmission) variants in some cases, so, for example, there will be SF6, SF6HT, N-SF6 and N-SF6HT.

Schott have produced a document entitled “Positive List of Optical Glasses – Update December 2006” This is attached to the end of this article.

So, let’s look at some consequences of replacing the old glasses with their lead-free equivalents!

The cemented achromatic doublet is known, understood and loved by all, and we really do not expect any problems by switching designs from old to new glass types. Here is a 100mm f/5 doublet designed for the F, d and C lines (equal weights) and a field of 1 degree.







This is a perfectly ordinary achromatic doublet. Let’s now change the glasses to the lead free forms. This is the wavefront plot now:



Without doing any re-optimisation, the performance is clearly unchanged. So here, yes, the glasses are ‘exact equivalents’.

Let’s now look at a design balanced in the blue and near-UV at wavelengths of 0.37, 0.4 and 0.46 microns [I started this example with a wavelength of 0.36 until I discovered that although it is valid with SF5, the reduced UV transmission means that it does not compute with N-SF5! Also, the f/no has been changed to f/6.67 so that the design has a similar performance to that of the previous example – at f/5 it was found that the shift in reference wavelength and the non-optimal glass selection had significantly increased the OPD]

Here is the design:



The aperture has been reduced slightly to produce a lens of similar performance to the previous design and the flint has been selected to produce the effect I wish to demonstrate (it is not the optimum flint for this design at this wavelength!)





Now, let us again change the glasses for the new lead free types, and without optimizing, the performance is now changed:






Clearly a significant change showing that, in this waveband, the old and new versions of F2 can no longer be considered exact equivalents. I deliberately did not mention the BK7 as the change from BK7 to N-BK7 makes no difference. The effect shown is entirely caused by the change from F2 to N-F2. [This, I believe, reinforces Schott’s comment that when only arsenic has been removed from an old type glass then the new glass is the exact equivalent.]

Re-optimizing the design does restore the original performance.

For a further example, you might like to design an f/5 NIR doublet of 100 mm focal length, corrected for 1.0, 1.5 and 2.2 microns using CaF2 (as the crown!) and SF1 as the flint - and then switch the flint to N-SF1. By now, the effect will not surprise you.

I suspect that not all the readers will be familiar with apochromatic lenses. The theory is available in several places – Herzberger’s book ‘Modern Geometric Optics’ comes to mind.

In achromatic lenses, the focus versus wavelength curve is essentially parabolic. In designs for the visible region, the foci at the blue and red wavelength will coincide with the green wavelength falling slightly short (for simple positive power lenses). The amount that the focus falls short is known as the secondary spectrum. Secondary spectrum is a ‘property’ of achromatic designs made from ‘ordinary’ glasses and where each lens group is colour corrected.

However, using anomalous-dispersion glasses such as the short flints (KzFSN4 for example), fluorite-crowns and flints in certain combinations can produce lens designs where the variation of focus with wavelength is essentially cubic - in a visible region design, for example, the red, green and blue regions will have the same focal plane with only small deviations of focus for wavelengths in between.

It was while designing an apochromatic lens that I discovered that serendipity and lens design do, just once in a while, go together! Let’s look at an ordinary apochromat. This one is a ‘split – triplet’ – i.e. one where an element has been split off to give coma control. It is easily derived from one of the ZEMAX sample files.

The design is for a very wideband apochromat of 100mm focal length, f/10 color corrected from 0.36 to 1.0 microns:



and looks like this:



With the following OPD performance and focal shift plot:





And, so far, absolutely no surprises…But now just replace F2 with N-F2, and without optimising we get:





These changes show clearly that the partial dispersion in the blue/near-uv region are very different between the old and the lead-free glass types. It is well known that apochromats are very sensitive to the exact melt data of the glasses so this large change in performance is not unexpected.

But what is unexpected is the result of re-optimising the design with the N-F2 glass:



The residual aberrations are clearly smaller. But why? We have developed a super-apochromat! Just look how well the focal shift is corrected now...



I do feel like one of the Princes of Serendip!

But will it make my fame and fortune? No, probably not. And why not? Well, I would expect the designs to be extremely sensitive to exact melt data. And at these unusual wavelengths the glass manufacturers charge a lot extra to undertake the measurements.

I leave the reader to explore further.

If anyone would care to study the effect of substituting new glasses for old in thermally stabilised lenses,  it would make an excellent Knowledge base article ;-)

I’m sure I have made the point several times already but I would like to emphasise once again that the new, lead-free glasses generally are not the exact equivalents of the old glasses that they replace.