Making the Right Connections

August 22, 2016

The Column

Volume 12, Issue 15
Page Number: 9–11

Columns | <b>Column: Incognito</b>

Incognito ruminates on the woes associated with connecting liquid chromatography (LC) and gas chromatography (GC) columns correctly, and ponders a future where everything is connecting differently.

Incognito ruminates on the woes associated with connecting liquid chromatography (LC) and gas chromatography (GC) columns correctly, and ponders a future where everything is connecting differently.

I have just finished teaching a class of bright young chromatographers. The session lasted about two hours and it was the most important thing that they will learn during their induction into our industry. It taught them about the beating heart of chromatography, the fundamental nature of separation science, and the daily struggle with the forces that work to overcome everything we strive for in terms of data quality and productivity. The lesson title was “How to Connect the Analytical Column”. I know that you know how to make good connections and it’s everyone else who does it wrong, but take a good strong gulp of pride and read on, maybe this discussion will be a good “refresher”.
Oh, and by the way – the fact that making a simple reliable connection between the instrument and the column isn’t possible, lies not with you my friend, but those who choose not to focus on the real issues in a chromatographer’s life…

In the many long years that I have been a chromatographer, this is one of the areas that has seen little progress in quality, ingenuity, or technology, and remains a key barrier to the success of many separations. Just take nano-high performance liquid chromatography (n-HPLC) for example. A technology that exists but is used in a limited number of application areas despite having many benefits, including improved sensitivity, increased throughput, and reduced solvent consumption, could be used in a whole lot more, except for the fact that making a true zero dead volume between the injector, column, and detector isn’t possible. Cutting a capillary, inserting into a connector without damaging it, and then tightening into the fitting without crushing it or leaving a dead volume isn’t easy - in fact it’s so difficult it’s a major barrier for the adoption of this otherwise enabling technology. I fully appreciate that there are “chip-type” columns that are sold in easy to handle “housings” with integrated connections, but these tend to be system specific and therefore of less interest if I want to explore the selectivity afforded by more than a single vendor, or if I want to use existing equipment that isn’t compatible with the chip fittings.

I can say with certainty that gas chromatography (GC) connections fall into the same “difficult to work with” category and form a major part of my induction class. But so does making an HPLC column connection of any sort - ultrahigh-pressure LC (UHPLC) or HPLC - and whether this is due to the effect of a poorly made connection on the quality of the separation or the possibility of running the instrument at all (that is, the connection leaks!), the basic ability to make a good reliable connection is fundamental to you being a chromatographer. I’ve seen bad connections being made by everyone from first timers to old timers and the discussions below are meant as gentle reminders to practitioners of the key pitfalls in column installation as well as a challenge to anyone out there who thinks they can help us to do this better. The one operation that chromatographers do on a regular basis that significantly affects our effectiveness has been sorely neglected by the technology innovators - and that simply won’t do.

HPLC column hardware is made by a range of different manufacturers with different internal geometry/ thread length/internal dimeter. While the diameter of the threaded part of the end fitting is common (these days!), the stem length of tubing from the tip of the ferrule and the shape into which the ferrule deforms on tightening is not. Subsequently, if you don’t make the connection properly, or if you use the wrong type of end fitting, dead volumes and even leaks might occur. Ask yourself the following question: What end fitting is the right one for each of the different column manufacturers that I use?

Now I’m not here to tell you how to make a good connection. There are many articles out here which tell you how to do that (1,2,3). I simply want to point out some pitfalls and possibilities for improving the technology.

For regular HPLC column connections, many of you may still be using PEEK (polyether ether ketone) tubing with PEEK one- or two-piece end fittings. Do you know how to cut the tubing to make a perfectly flat tube end at 90° to the wall of the tube? Do you know how many times you can make and break the PEEK nut and ferrule before it is not “plastic” enough to properly deform into the shape of the next column end fitting being used? Do you realize that simply applying a set of pliers to tighten up a leaking PEEK fitting is really doing more harm than good?


For those of us using UHPLC, we will be using metal tubing with swaged ends, which use metal nuts and ferrules. What is the recommended number of connections that should be made with these fittings? What column end fittings are these nuts and ferrules designed for? When did you last use a different fitting for a column with a different internal geometry to the one you just removed?

For UHPLC, it is possible to obtain non‑permanently swaging nuts with PEEK or polymer ferrules which “deform” into the end fitting. Check how many make and break cycles these are recommended for and check the logs for the number of cycles for each fitting to see if everyone is adhering to the recommendation. Oh that’s right - you don’t have a log - because like every other laboratory in the world, you use them until they don’t function anymore! How tight do you tighten these fittings? Do you have a laboratory policy on “finger tight” or the number of turns of the wrench past finger tight?

Recently there have been some nice advances in commercially available HPLC fittings that have sprung capillaries and polymer ferrules, which do a great job of reducing the dead volume and are re-usable for many hundreds of connections with just finger tightness - so well done you guys! However, there has not been a click and connect format which takes us to the next level of convenience or connection quality. We need to look at the column end fitting and instrument design and change them - fundamentally! For those manufacturers who may offer alternative solutions to those mentioned at the beginning of this paragraph - I apologize - I’m just not aware of your technology yet.

For GC, do you know how to cut the silica tubing so that the tube end is perfectly smooth and again at 90° to the tubing wall? Are you aware of the often catastrophic consequences on your peak shape of not doing this properly? Do you know which ferrule material should be used with your application or instrument? Do you know the correct insertion depth of the column into the inlet and detector? I could go on and on with these questions - but I think you get the point. The quality of the connection in GC is, in my opinion, much more important than in HPLC and the technology even more backwards. There have been many “revolutionary” connectors designed over
the years - but none have caused much
of a stir and most of the world are still juggling with a heavy nut and ferrule perched on the end of a very fragile and springy capillary in a hot oven, trying not to burn their fingers whilst hopefully the column remains at the correct insertion depth as they use a very small wrench to tighten up the nut. Wow!

Why do any of our column connections need to have nuts and ferrules? We have quick connect/release couplings on so many pneumatic and hydraulic lines in the laboratory - why can these not be adapted for use in our systems? Why is it not possible to simply click the column in place within the system and we are ready to go? Even under circumstances that require the column to be trimmed, as with GC columns, we should still expect that the capillary could be simply pushed into the fitting, prior to clicking into place. Or better still - find another, more durable material that is easier to prepare than silica coated with polyimide - but that’s another story! For an excellent example of the use of advanced connection technology being used to simplify and miniaturize complex instruments, I would urge you to take a look at the CARPEDIEM dialysis instrument recently developed for neonatal treatment (4). Where there is a will, there is a way!


  1. John W. Dolan, LCGC North America24(10), 1078–1084 (2006).
  2. John V. Hinshaw, LCGC North America 24(7), 670–67 (2006).

Contact author: Incognito