As stated the problem is that these filters crack. The reason is were they are placed in the optical path. When you screw them into the eyepiece they at or very near the focal plane of the telescope and of course that is were all the energy is concentrated. Many times this depends on the design of the eyepiece. If you place the filter away from the focal plane like on the front of the star diagonal then the enegery is spread out and heating becomes very much less of an issue. We all know that you can light a piece of paper on fire when placed at the focus of the telescope but move it an inch inside or outside of focus and all you do is make the paper warm.
Also one has to understand that the size of the solar image the telescope produces also is a big factor. The size of the solar image is about 0.0092 x the focal length. So you have the relationship between size and energy. The smaller the image the more energy per unit area. So a long focus scope produces a bigger image that is less of a strain on the filter. You we have ever seen pictures of the solar osberving room at Mt Wilson, it shows a white light image of the Sun that is a number of feet in diameter, yet the table that the image is project onto doesn't burst into flame.
As for the Criterion S-4 solar eyepiece, it is very well engineered unit. First it uses a barlow lens with one of the concave surfaces metallized. So the Sun light hitting this surface is divergent and can't come to focus inside the telescope. Since it's a barlow, it needs to be placed well inside the focal plane of the telescope, and not near the focal plane so it does not see any concentrated energy. Next there is a green filter similar to the ones we are discussing here. It is placed inside the focal plane of the barlow lens. So it can never see concentrated sunlight light. The green color is around 540 nm which is suppose to be the best for "white light" solar observing and this is wavelength that Baader uses in their "Continuum Filter"
If you want the best in white light solar observing, you build an instrument designed exactly for it. Here is a picture of my 4.25" f/10 Solar Newt, which uses uncoated optics and a built in Herschel Wedge.