HI Naizhong--

Is the problem worse if the bellows have different pressures in them?  That could indicate a problem in linearity starting to affect your D47 signals. 

Also, have you checked for leaking across your valves?  Perhaps even a transient leak to atmosphere? 

good luck

take care,


For the newbies reading this, to check for leaks fill up both bellows with gas, shut off both valves to the bellows, then pump out everything else, including capillary lines.  Once everything else in pumped out completely (rough, then turbo for a while), close all valves.  Compress bellows a bit.  Then wait.....

You shouldn't see any loss of pressure in the bellows--though pressure reading may fluctuate with temperature.  But you may still see some gas leaking.

Open up valve to roughing pump, with the readback from the pressure gauge.  Note any change and any fluctuation.  Then open up the different sections to the roughing pump and look for changes in pressure readings.  Do this one at a time, and close the previous section if possible.  If you find a section that causes a pressure jump, say reading go from 10-3 to 10-2, note that.  The cross section between bellows and external gas connection may show evidence of a leak, but you don't know which valve is leaking--it could be atmosphere leaking in.  Cap off the external connection, pump it out, and check for leaks again. 

Sometimes the leaks will go away if you just reseat the valve, i.e. open and close it for a while.  Certain leaks may not be too bad--e.g. if the valve to an unused and capped off internal port is leaking a little, but the cap is fine, just leave that valve open all the time if that increase in dead volume is acceptable.  Then fix it next time instrument in down.

On 1/26/2016 2:52 AM, ZhangNaizhong wrote:
[log in to unmask]" type="cite">
Hi Andy and Albert,

Thank you so much for the suggestions. Unfortunately, we have not solved this problem yet.

To Andy,
d13C and d18O are quite well. To the D47, we got the results of ca. 0.04‰ in last December; this month we have changed the capillaries between sample side and standard side without shut down the MAT253, now the results become ca. 0.09‰ and keep constant. At the same time, we measured D47 of left vs. left and right vs. right, both of them can give the results of ca. 0 (<0.01‰). If the problem are the bellows, do you have any suggestion to clean them? We will shut down and clean the ion source/change the filament the day after tomorrow.

To Albert,
Yes, we are using the PBL correction methods suggested by He et al. (2012). Actually Bo is my friend and he helped us to apply this method in 2013. And we have also compared the results between PBL correction and no PBL correction. There is almost no difference to the reference gas. By the way, our PBL of m/z 47 is around -15mV when we want to get a m/z 44 of 16V.

Thanks a lot!

Best wishes,

From: Stable Isotope Geochemistry on behalf of "Albert S. Colman"
Reply-To: Stable Isotope Geochemistry
Date: Thursday, January 21, 2016 at 11:54 PM
To: <[log in to unmask]>
Subject: Re: [ISOGEOCHEM] Problems of MAT253 on clumped carbonate measurement

Hi Naizhong,

Do you monitor and correct for your "pressure baseline" (He et al., 2012.  Rapid Comm. Mass Spectr., 26, 2837-2853.)?  This shift in the m/z 47 baseline upon introduction of sample CO2 into the source will scale with the size of the m/z 44 ion beam (combination of secondary electrons and main peak pressure broadening).  A mismatch in pressures between reference and sample side will lead to different baselines for each.

When this pressure baseline is sizable (10s of mV on m/z 47 for a working ~16V signal on m/z 44), then a small mismatch in reference v. sample pressure balancing can lead to self-zero tests that give apparent non-zero D47 values in the range that you observe.  That effect will change with different source tunings.  It will also drift through time, just as many labs have noted that the slopes of uncorrected HG and EG curves drift through time.

Even if you haven't been monitoring the pressure baseline explicitly, the slope of your heated gas/equilibrated gas lines is an indication of the magnitude of this baseline effect.  As we derive in the supplemental information associated with He et al. (2012), the slope of the heated and equilibrated gas lines is approximately equal to the PBL_47/V_47,ref.  You could apply this correction for post processing of the data you've already collected to see if this resolves your issue. 

Albert S. Colman
Assistant Professor
Dept. of the Geophysical Sciences
University of Chicago
5734 S. Ellis Ave.
Chicago, IL  60637
[log in to unmask]

From: Stable Isotope Geochemistry [[log in to unmask]] on behalf of ZhangNaizhong [[log in to unmask]]
Sent: Thursday, January 21, 2016 3:32 AM
To: [log in to unmask]
Subject: Re: [ISOGEOCHEM] Problems of MAT253 on clumped carbonate measurement

Hi Andy and everyone,

Thank you for your suggestions. We have repaired the problem of sensitivity decrease with the method suggested by Hagit. 

However, the first problem became even worse. Recently we have changed the capillaries from left to right, and also from right to left, the values were not reversed, but deviated farther from 0 (positive as before), now the result was increasing gradually from 0.04‰ to 0.10‰. Does anyone ever faced this problem? I am now worrying about the pollutions of bellows. If so, how could we clean the bellows? Or is there any other suggestion?

By the way, does Nickel capillaries work better than stainless steel ones? We are now considering to change them to Nickel ones.


From: Stable Isotope Geochemistry on behalf of Andrew Schauer
Reply-To: Stable Isotope Geochemistry
Date: Thursday, January 7, 2016 at 5:44 AM
To: <[log in to unmask]>
Subject: Re: [ISOGEOCHEM] Problems of MAT253 on clumped carbonate measurement

Greetings Naizhong,

Are the bellows calibrated? Ours fell way off calibration frequently. Make sure both sides are actually 0% and 100% open when the software indicates so. 


On Thu, Dec 24, 2015 at 7:38 AM, ZhangNaizhong <[log in to unmask]> wrote:
Hi Hagit,

Thanks for your quick reply. You are right, the sensitivity decreased on both sides. Our box and trap currents in this Aug. are 0.60mA and 0.90mA, respectively, and graduated changed to 0.54mA and 0.96mA now. I will check the parameters on focus board. Hopefully it will be helpful.


From: Stable Isotope Geochemistry on behalf of Hagit Affek
Reply-To: Stable Isotope Geochemistry
Date: Thursday, December 24, 2015 at 6:48 PM
To: <[log in to unmask]>
Subject: Re: [ISOGEOCHEM] Problems of MAT253 on clumped carbonate measurement

Hello Naizhong

I cannot say if these 2 problems are related. they seem unrelated but the timing of both being the same is suspicious. you do not say, but  I assume sensitivity is lost on both sample and reference side. if only on one side- then it is a capillary or bellow problem that is related to the bad zero enrichment. if both side, it is probably something else.

We had similar and sudden sensitivity loss on a 253 used for clumped isotopes back in 2009. we would lose 30% of the signal suddenly, every 2 months. cleaned that source and got it back. after 1 year we could not it get it back and sensitivity was going up and down like crazy. then we realized the the problem is electronic - a faulty x-focus board. to check it, change the focus setting by very little on each parameter and see if sensitivity responds strongly. the plate that responds probably means a faulty electronic part associate with it.

how are your box and trap currents?


Hagit Affek| Associate Professor 
Institute of Earth Sciences 
The Hebrew University of Jerusalem
Edmond J. Safra Campus
Jerusalem, Israel.

Phone: +972.2.6584654 | Fax +972.2.5662581
mail address: [log in to unmask]
On 12/24/15 11:35 AM, ZhangNaizhong wrote:
Dear all,

We are managing a MAT253 for CO2 clumped isotope since 2013. Recently we have faced two problems and need your kindly suggestions.

1. When we measured the same gas (e.g. OZT std gas) at sample and reference gas sides, D47 deviated from 0 per mil (e.g. 10~20 per mil). 
We have asked this question in 2013, when we initially installed MAT253. During that time, we got many suggestions, and we follow the method to change the capillary. It worked quite well until this spring, and we could get ca. 0 per mil when we measured the same gas. However, it suddenly changed to ca. 20 per mil from this May when every other parameters were no difference except the sensitivity (see our second problems).
During the past months, we have tried to washing/baking the capillary by flame and Helium, but it didn’t work at all.

2.About the sensitivity. 
We faced the problem that the sensitivity of MAT253 decreased suddenly (when we set Mass 44 to 16V, bellow pressure changed from 65mbar to 85mbar), also in this May. 

About these two problems, we thought that they were caused by either the pollution in ion source, or the pollution in capillary, or both of them. Therefore, we would clean the ion source soon, as well as considering to change the capillaries from stainless steel to nickel. Does anyone has faced this kind of problems and may have any better suggestions?

Thank you so much for the kindly help.

Merry Christmas and happy new year.

Naizhong Zhang 
PD in Tokyo Institute of Technology