First, a brief note: Quartz press fit type unions and Y splitters are now available on eBay (search "Innovaquartz") at prices lower than any catalog.

Now, a discussion of the performance of these type connectors:

Press fit type unions and splitters rely upon the polyimide coating of capillary columns to form a seal: when inserted, the capillary makes contact with the conical bore of the connector and the polyimide is compressed between the cone wall and the column glass, forming the seal. The cone angle in the connector must be low to prevent glass-to-glass contact and provide a sufficient area of polyimide compression to form a leak-tight seal, even where there is a bit of capillary ovality. The lower the cone angle, the better the seal (and the lower the area of polyimide exposed to carrier gas and analytes).

Torch made borosilicate unions and Y splitters came first but they were column diameter specific, not very inert and tended to leak due to the difference in cte between silica and borosilicate glasses. In the late 1980s came "universal" unions in fused quartz, made with a hydrogen/oxygen torch, theoretically capable of making a seal on any of the common capillary column diameters, but due to non-linearity in the conical bore angle, some column diameters sealed better than others. (The torch flame produced a curvature within the quartz tube, not a cone, such that 0.25mm columns sealed on cone angles of 2 degrees or less, but 0.53mm columns found their mating diameter on angles as high as 5 degrees or more.) Snap-cut connector ends were flame-polished with a hydrogen/oxygen torch, resulting in a slight lip in the bore of the openings that applied excessive stress to the column at the opening.

In the early 1990s, a method of producing constant cone angles (linear tapers) with a laser was patented by InnovaQuartz, producing true universal unions and Y splitters in quartz where the contact angle between column and union wall was always less than 2 degrees. Laser cut ends had less of a lip in the bore, but excess stress remained an issue in some installations.

Very little changed for two decades, but after exiting the business for a few years, InnovaQuartz returned after developing improvements to the laser forming process. Pres2fit (tm) unions and Y splitters are now made to dimensional tolerances less than 1/2 of the prior generation, thereby reducing the restriction of the conical bore minimum to less than 1/2 of prior products and cone angles are now less than 1 degree. The Pres2fit bores are also far more inert than before due to substantive changes in production processes -- orders less activity and reactivity. All connectors now have insertion cones (bore chamfers) to aid in guiding columns into the connectors without damaging the clean-cut ends and eliminate stresses due to flame-polish bore ridges. And Y splitters are produced with less than 1/3 of prior dead volume and with superior balanced flows.

Does any of this really matter?

It does if you have ever had trouble with leaking unions or Ys. (By far, the most common cause of leaks is column ovality. Low cone angles provide greater compression of the polyimide layer, overcoming more ovality than possible with higher angle products.)

It matters if you wear reading glasses. (Threading a capillary into a standard union without damaging the glass or the polyimide is difficult enough for younger chemists. For those of us with failing near vision, it is almost impossible. The lead-in cone of new products greatly facilitates damage-free insertion.)

It does matter if you have experienced target compound degradation or peak tailing when using unions or Ys. (Prior products' production required hydrothermal etching in a Parr bomb to remove deposited silica vapor in the connector bores, a process that saturated the bore surface with active silanols and introduced heavy metal ion contaminants.)

It does if you see peak broadening when using press fit splitters. (While dead volumes in prior products were very low, sample volumes are being driven ever lower so even tiny dead volume contributions can be problematic.)

It matters if you have ever had a column break at a connector opening. (Competitive "laser cut and polished" connectors continue to present small lips -- restrictions in the bore diameter -- at the connector openings and these are often of sufficient height to wear completely through the protective polyimide layers, resulting in failure of the column.)

It also matters if you have not experienced any of these problems with unions and Ys, because if you continue to use inferior products, you will have one or more of these experiences, and at lab direct pricing, you'll save money, too.