Hello fellow MCD users,

In case this is helpful to anyone, we've had good luck (maybe emphasis on luck?) making leak-tight seals on the MCD for a while now. Here's our approach:
  • When pre-swaging the ferrule onto the capillary, I do this somewhat lightly with the fingertight tool, i.e., don't crank down as much as possible-- swage it so that the ferrule ends up in the right place, but with some "squish" remaining. 
  • Then, when attaching it inside the MCD -- after making sure it is properly seated-- crank down as much as possible with the knurled fingertight tool, never a wrench. Because there is some "squish" remaining in the ferrule, I think this helps it make a seal inside the MCD. (In my experience, tightening with a wrench can only result in leakiness because it is so easy to overdo.)
  • We then heat-cycle the GC 2 times on a dummy program that ramps up to 300 C & back down. When the temp returns back down, go back in with the fingertight tool and see if it will tighten any further. Again, as long as you do not use a wrench, there is no risk of overtightening (as far as I know-- this could be dependent on an individual's hand strength...).
  • We routinely are able to get a very tight seal with this method-- around a 2-5 mV increase in m/z 40 between backflush on & off. The seal is often retained through hundreds and hundreds of heat cycles-- both the GC column and the 3 Silcosteel capillaries (2 reactors & backflush), in our setup.
  • One caveat is that if a particular MCD has experienced a lot of overtightening, this might not work-- my impression is that overtightening slightly expands the MCD ports. At some point we started using a new MCD and only using this tightening method, and we've been very happy with the results since then.
  • When we have not been able to make a good seal using this method, sometimes we can solve the issue by using a new Siltite nut. BUT on occasion it has been the fault of the MCD itself, as Marilyn mentions. Just this week we had our most recent MCD "go bad". It's possible that it got gummed up with something(?) Overall, we've gone through 3 MCDs in 3.5 years. Each has been beautifully leak-tight once we figured out the tightening sweet spot. But they seem to not last as long as one might hope for a ~$560 part. Thankfully, based on Aaron's previous advice, we had a spare Valco cross on hand and installed that in a pinch.
  • Finally, I don't know how others cut their Silcosteel capillaries, but that might also be a factor. We score aggressively with a ceramic wafer and use two pairs of pliers to gently bend back and forth, importantly placing the pliers directly next to the score so the "good" end of the capillary does not bend. We use blunt-end, smooth-faced (non-serrated or non-marring) pliers on the "good" end of the capillary to hold it very flat. Then of course inspect the cut under magnification before cleaning & swaging.
Good to know about the variations in jig sets as well as all the other advice!


Hilary Close, Ph.D.
Assistant Professor
Department of Ocean Sciences, RSMAS
University of Miami
[log in to unmask]

On Wednesday, March 10, 2021, 01:24:03 PM EST, Linnea Heraty <[log in to unmask]> wrote:

Hi Hi

Oh SilTite, you vex me so!

After many many frustrating hours/days/weeks trying to make a good "finger tight" seal, I found success when I switched SilTite jig sets. The "bad set" worked fine for silica capillaries, but I could never get a good seal with the stainless capillaries. With the "good set" I usually get a good seal, usually. I also make sure the capillary cut end is perfect and clean so that it fits into the jig base just right. I use superfine sandpaper to square the cut end then rinse in methanol.

Not my discovery, though I honestly do not remember who mentioned to me that not all SilTite tool kits are created equal. I'm not too pleased with their quality control. Luck of the draw if you get a good set of jigs. I only had two sets because I happened to have purchased a whole new kit after destroying one base plate. I also had 2 brand new base plates with visibly bad welds! 

So much fun!

Happy Wednesday

On Tue, Mar 9, 2021 at 8:34 AM Diefendorf, Aaron (diefenan) <[log in to unmask]> wrote:
Hi all,

I love Marilyn’s comment that “luck works as good as skill”! Sounds about right with the Silflow MCD! We tried them out after the sales pitch and were never able to use them reliably.

We use the Vici 1/32” crosses (ZX.5) and tees (ZT.5) with 0.25 mm bore and either 0.4 or 0.5 mm ferrules (for glass capillary or metal capillaries, respectively). These tend to be leak tight after the first couple oven cycles and stay leak tight for weeks (don’t go tightening when they aren’t leaking!) and they aren’t crazy expensive. We typically use the tee for hydrogen and the cross for carbon (we replace the hydrogen reactor with a glass capillary and run it through the reduction reactor).

I complained at one point here about issues with the nuts on the Vici crosses/tees seizing over time. We found that after 1-2 months, the male threads on the nuts deform (or oxidize?) but the female threads on the cross/tee are ok. We replace the nuts when the threads no longer turn cleanly. We now typically get a year or so out of a cross or tee (and they fail usually b/c bad nuts get over tighten and then can’t be removed).

On a side note, I went to order combustion and reduction reactors and was shocked on the price increases (reduction is now $465). Anyone have a trick for tightening unions on homebuilt reactors in tiny ovens? I used to do this all the time in the Agilent GC, but the oven is so cramped in the Trace GC between the column and the reactors, it’s hard to get a good angle to tighten the union without snapping the ceramic tube. Does this require “luck" too?


Aaron F. Diefendorf, Ph.D.
Associate Professor of Geology
Department of Geology
University of Cincinnati
PO Box 210013
Cincinnati, Ohio 45221-0013

Office: 513-556-3787
Website: aarondiefendorf.weebly.com

On Mar 8, 2021, at 7:46 PM, Marilyn Fogel <[log in to unmask]> wrote:

Dear Isogeochemists,

These fittings were supposed to solve everything. Valco unions were easy to get leak tight, but if you needed to install a new column or reactor--forget the fitting. I had to throw them out and start over. Then the sled--as Seth mentioned, awful device with 2-holed ferrules. Folks in my lab have similar experience using 0.5 mm capillary tubing and trying (repeatedly) to get them leak tight. We've learned the Siltight "blocks" wear out--and we've replaced them when they seem to have bits of capillary stuck in them.

Finding another set of fittings defeats the purpose of the siltite! 8 unions instead of 4--ugh. This should be a message to Thermo engineers to try to fix this problem and come up with a solution. At first we thought it was sloppiness in putting them together, but we've learned that's not the case. I think luck works as good as skill. Patience is key.

How about glass press fit fittings for the 0.4 to 0.5 mm union? 

Marilyn Fogel

On Mon, Mar 8, 2021 at 4:26 PM Jack Hutchings <[log in to unmask]> wrote:
Hi all,

I have also suffered under the Silflow MCD. I can second Chris: getting the 0.4 size leak-free is *significantly* easier than getting the 0.5 leak free. I have had little luck in getting 0.5 ferrules leak free using fused silica capillary, much less with the metal capillary attached to the thermo reactors (this is almost impossible). I've been given the suggestion of pre-conditioning the ferrules by leaving them in a beaker in the GC oven, but this hasn't helped at all. Like Chris, I have resorted to using 0.4 ferrules with fused silica capillaries and using microunions to join the FS capillary to the 0.45 mm OD metal capillary attached to the reactors and the backflush. I have not found a permanently satisfactory union for this purpose - the most success I've had so far is using the Restek MXT union (Restek #21386). However, I have (literally last week) broken the nuts of one during tightening and have decided to try using fresh nuts with each install. The MXT union was leak-free for about a month and I think could probably be maintained with as-needed (careful) tightening. I have also tried the Siltight microunion (Restek #23886), but I was only able to get that leak-free with the metal reactor capillary once and have since given up on it. I've tried the Gerstel GRAPHPACK 3D/2 connector on the metal capillary (Restek #20272) without any success yet. It's worth noting that VICI sells a Valco microunion (VICI # EU.5) that I have yet to try, but uses the same exact ferrules as the Restek MXT union.

Anyone with advice on a preferred microunion for the metal 0.45 mm OD capillary coming from the reactor would be welcome, although I've tried damned near every option at this point!

Stay sane and good luck!


Jack Hutchings, Ph.D.
Manager, Konecky Lab
Department of Earth and Planetary Sciences
Washington University in St. Louis

On Mon, Mar 8, 2021 at 4:17 PM Wurster, Chris <[log in to unmask]> wrote:
Hi Ewerton,

I had asked for advice on this very problem not too long ago. The solution for us was to use capillaries (0.4 size) on all SilFlow connections instead of the silcosteel (0.5 size) that runs from the reactors and backflush, then I have to run a microunion between the reactor/backflush and the Silflow. After that, I can get the system leak tight (after a few temperature cycles and tightening the union). I know other folks successfully use the Valco as a replacement, so you can consider that if you have gotten the system leak tight using Valco?


Chris Wurster, Ph.D.
Senior Research Associate
Centre of Excellence for Australian Biodiversity and Heritage
College of Science, Technology, and Engineering
James Cook University
Cairns 4870

Office: D3-114

Fax: +61 7 4232 1583

On Mar 9, 2021, at 5:20 AM, Santos, Ewerton <[log in to unmask]> wrote:

Dear Isogeochemists,

Recently, we had a Delta V Plus installed in our lab. This equipment has installed with a Silflow platform to direct the samples to reactors. We are experiencing a constant leak coming from the Silflow, we did some tests to make sure that, exchanging the Silflow for a Valco 4-way valve, and the leaks disappear using the Valco valve. We had tightened the nuts of the Silflow using the Fingertite as mentioned by Thermo, we are using the correct ferrule dimensions, and heated the oven at high temperature for a few minutes but the leaks were coming from all nuts. Someone had experienced something similar when is working with the Silflow platform. If so, how to avoid or fix this problem? 

Thank you,

Ewerton Santos

Department of Earth, Environmental and Planetary Sciences
Brown University
324 Brook Street
Providence, Rhode Island 02912

Linnea Heraty
Environmental Isotope Geochemistry Lab
University of Delaware Dept of Geological Sciences
ISE Lab 458
221 Academy St.
Newark DE 19716