Thanks to David for the detailed explanation of cracked tubes. I have some additional comments and a question:

Columns that do not break on cooling seem to be the exception rather than the rule on our Carlo Erba NC2500, which is fed with everything from pure amino acids, to soils and ceramic powders.

Salts in samples are the main culprit for devitrification, typically forming sodium silicates.

Biogenic or any other sample-derived silica melted onto the quartz tube likely has a different thermal expansion coefficient, which will also stress the tube on cooling.

Is it possible to program the EA for a long, slow (overnight) cool-down to decrease the probability of cracking?

Stanley H. Ambrose, Professor
Director, Environmental Isotope Paleobiogeochemistry Laboratory
Department of Anthropology
Department of Geology
Center For African Studies
Program in Ecology Evolution and Conservation Biology
Division of Nutritional Sciences
University of Illinois,
Urbana, IL 61801 USA

From: Lola Oliver <[log in to unmask]<mailto:[log in to unmask]>>
Reply-To: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>>
Date: Tuesday, March 24, 2015 at 1:22 PM
To: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>>
Subject: Re: [ISOGEOCHEM] Why do my EA (combustion side) quartz tubes keep breaking on cooling?

Thank you for the very clear explanation!
     I gave up trying to save combustion columns years ago, but wondered why they always broke where they did and sometimes merely fractured rather than broke when removed.  It was just easier to get a new column ready rather than try to cool and save the column, which inevitably fractured when taken from the oven, no matter how slowly and carefully it was cooled.
     Good to know I wasn't wasting time or materials.


On Tue, Mar 24, 2015 at 10:05 AM, David A. Mucciarone <[log in to unmask]<mailto:[log in to unmask]>> wrote:
Here's my voodoo on the issue of breaking combustion/oxidation tubes after use.

Many of us run all kinds of stuff in our EAs that have been treated with acid, not treated, in silver, in tin, glass fiber filters, soil, sediments, diatomite, etc. If it burns and is worth looking at, in it goes. Our ovens are at 1000 to 1030C, more or less, and these oven do not heat the columns evenly. Quartz is getting close to its plastic state at these temps and with the addition of O2 with tin, we get flash temps higher. Under these conditions the fused quartz is more susceptible. Some of us run inserts open at the bottom, some closed at the bottom, and others no inserts at all. All of which can produce different column degradation behavior.

So what the deal with columns breaking when we cool down our ovens so we can make column swaps? Fused quartz has some nice properties, but these properties can also cause issues. Fused silica at high temp >1000C when exposed to impurities under these more susceptible conditions can undergo a process called devitrification, but under pure conditions this process is uncommon at our EA oven temps. However, our systems are not pure given what most of us run in our EAs and these impurities can work their way into our fused quartz columns. This process can be more complicated in applications like ours (in EAs) where the columns are hung and not supported well at the bottom (hopefully most of you use a stand under your columns for support). Quartz has low thermal conductivity and can crack if heat is concentrated - the hottest part of the EA oven is near the bottom of the insert or top of the chromium. Most of this exposer is at the insert/chromium interface. After weeks or months at these high temps and many samples later, it is common the combustion tubes will be exposed to impurities of all kinds and the fused quartz begin to devit. Quartz also has low thermal expansion. Open spaces will cool differently than filled areas and these filled areas may react differently depending on the properties of the material. If you combine the normal qualities of quarts, low thermal expansion, low thermal conductivity, annealing point, the thickness of the tubes we use and how they are use, it shouldn't come as a big surprise there can be issues when cooling down the tubes.

Now comes the issue, we want to change out our columns without them breaking because it's a pain to cleanup so we cool them down so they can be handled safely. To do this, we need to cool them down slowly and evenly - a process call annealing which aids in removing the stress in the glass. Here's the catch, most EA ovens don't do this well and with different packing in the oxidation column and open space above the packing does not allow them to cool evenly, the glass can stress and crack. Annealing T for transparent quartz is about 1190C and it's distortion point around 1090C, opaque glass is lower around 1080 and 1030C respectively. So in mostly all cases we use our tubes at or close to the distortion point. Common areas are the chromium/insert area and the area between the chromium and silver cobaltic oxide. If you are careful and patient enough you should be able to cool the columns down without incident, but most of us want to get the instruments up and running quickly so the process tends to get rushed a bit. if you pull them at high T they have been used a lot, you may see the quartz delaminate or spider crack. In either case, after removal you may see a milky finish to the fused quartz. This is a sign of devitrification. If your tubes are coming out clear, this process is minimal or not present. If you look at most of the inserts coming out, they are frosted at the bottom, but not at the top (more than likely the fused quartz is of lower quality and is thinner walled), you might also notice that the oxidation column is milky in the combustion area, but not above or at the bottom where the temps are lower and the fused quartz less likely to devit.

I have been running my EA since the mid 80s and have experienced my oxidation column cracking and most times it doesn't, all depends on how the column was used. I have never experienced this on the reduction column (in fact, I re-use this column many times until I break or crack it when cleaning out the copper) which in most of our systems is at 600 to 650C which is well below the plastic state of fused quartz and unlikely to devit at this temp.

For piece of mind I have adopted a ~2000:1 policy for my EA or a 4:1. After ~2000 analyses I change out the oxidation/combustions column or after 4 reduction column changes which is ~450 to 500 samples/ea with the copper I am using. Doing the math gets you to about 2000 samples. At ~$25 to $30/ea it's not worth the risk of a column break and lose an expensive tank of helium overnight. multiple uses of the reduction column more than makes up for the cost of the 1 use oxidation column. I do know that when I used to leave the reduction columns in for 5000 or more analyses, they would crack upon removal. Now they occasionally crack after removal using less time. When I used opaque columns I would see this problem more often, this is mostly because of the different properties of the quartz compared to the transparent quartz.

To change out the reduction column, I bring the oxidation/combustion column down to 800 to 820C, reduction column to ~350C, purge the helium, pull and replace the new reduction column, close the purge, let stand and equilibrate for about 15 to 20 min, longer is fine, then bring both ovens up to operating T. Change out the water trap, load and go.

I think you can play around with using quartz wool at the insert interface with the chromium or you can add more quartz wool to the insert bottom or you can use closed bottom inserts and avoid the sample oozing. Some of us know how much fun it is to work with glass fiber filters and keep it from oozing down or even have tin fuse into the sides of the columns. These steps might help and perhaps worth pursuing. How we pack our columns may also matter since a more tightly packed column will expand differently than a loosely packed column. There are a combination of factors at work and given the inherent properties of fused quartz, and the nature we are using these tubes small changes in reagent densities, their thermal responses with respect to the properties of quartz glass, complicated by potential devitrification degrading the quartz glass I think if you haven't had a cracked column, you will sooner or later.

As for standby mode on the EAs. There is a reason why the standby temps on the EAs are the way they are, basically to avoid column breakage.

All in all, this isn't a big issue, it doesn't really impact your results, it's not a high expense issue, it's more or less a maintenance concern and perhaps one of curiosity. If it's a safety concern, put a non-flammable catch tray under the columns so the hot reagents cannot cause a burning issue. If it cracks, shut off the helium, the mass spec should already be isolated, let it cool, change and clean.

Anyway, my voodoo on the subject.  Happy running.


At 03:54 PM 3/23/2015, you wrote:
Hello all,

I vaguely recall a brief Isogeochem discussion about EA quartz reactor tubes breaking on cooling in the not too distant past.  I have been experiencing this problem for some time and I wonder if you great people have any further insights and recommendations for managing the problem.  All thoughts and ideas most welcome!

Here's the situation:

Continuous flow (100ml/minute UHP Helium) EA-IRMS analyses of d13C and d15N go smoothly.  No leaks in the reactors (confirmed by capping the gas line immediately up stream of the water trap, and monitoring of He consumption rates across several reactors and He cylinders - which by the way are about 100PSI on the He tank pressure per day and the majority of this is actually lost via the conflo and a minority lost via the EA).

Tin and silver capsules (from micro to semi-micro to macro O2 modes) - with both organic (you name it) and predominantly inorganic (i.e. muddy-sandy sediments) samples and standards - return excellent delta values and C and N yields.  The data is great and nothing goes wrong during the runs - including multiple ash tube changes.

However, on cooling the combustion side reactors unfailingly break (not talking spider cracking here, they literally break in two pieces usually right around the top of the chromium oxide just below the quartz wool beneath the ash tube - i.e. at/near the hot point of the reactor).  The breakage happens when the reactors pass between 500degrees and ambient temps - not during the stepped cooling from 900degrees to 500degrees.  Breakage happens with 100ml/minute helium flow (around 1.5 bar) and it also happens when I've experimented with venting the reactor during the 500degrees to ambient cooling stage - thus, it seems to me that the pressure in the tube is not part of the problem.

Breaking tubes is not so much of problem for getting good data - it's more a pain than anything else.  Perhaps the quality of the Cr oxide has changed?  Perhaps there's something in my samples (nb - I don't think this is the case as I've run a huge variety of sample types with various treatments and they all end up with a broken reactor)?  Perhaps the quality/purity of the quartz tubes has changed (I only take them up to 900degrees so no where near melting point of pure qtz)?

Thanks for reading and thanks for your thoughts and comments!!

Travis Horton
University of Canterbury

This email may be confidential and subject to legal privilege, it may
not reflect the views of the University of Canterbury, and it is not
guaranteed to be virus free. If you are not an intended recipient,
please notify the sender immediately and erase all copies of the message
and any attachments.

Please refer to for more

David A. Mucciarone
473 Via Ortega, Rm. 140
Stanford, CA 94305
650-723-0817<tel:650-723-0817> (O), 650-804-6403<tel:650-804-6403> (M)
[log in to unmask]<mailto:[log in to unmask]>
Office is in Green Earth Sci. Bldg. Rm. 331

Lola K. Oliver, Supervisor
e-mail [log in to unmask]<mailto:[log in to unmask]>
Telephone (907) 474-7114
Fax (907) 474-6184


For commercial freight services:                or                  For USPS only

Forest Soils Laboratory                                                   Forest Soils Laboratory
University of Alaska Fairbanks                                        PO Box 757200
335 O'Neill Building                                                         Fairbanks, AK  99775
905 Koyukuk Dr.
Fairbanks, AK  99775