So strange the HGL is!

26 Jun 2015

Did anyone once build the HGL like this? Is that  possibly because of the re-ordering effect due to by using the uncleaned cellected gases tube? If anyone of you would like to share your opinions about this, please post it. Thank you! 

this picture URL is:http://img4.douban.com/view/photo/photo/public/p2252548179.jpg

HGL


brosenheim
Offline
Joined: Sep 15 2011
The negative slope is indeed

The negative slope is indeed very strange.  It would suggest that you are overestimating the amount of negative species supressing your mass 47 beam.  Typically, other instruments have reported a positive slope because they measure the backgrounds without a large mass 44 beam in the flight tube.  The PBL correction originally described by He et al., 2012, showed this quite well. Rosenheim et al 2013 (RCM) also showed this on an instrument with a very high positive slope, and the method described therein very slightly overcorrected for the backgrounds. 

Are you doing a PBL correction?  Are you perhaps overcorrecting?  Which instrument are you using? 

Brad E. Rosenheim

Assistant Professor

College of Marine Science

University of South Florida

Saint Petersburg, FL, U.S.A. 33701

http://www.marine.usf.edu/rosenheim/BER-Group.html

Cedric's picture
Cedric
Offline
Joined: Feb 6 2011
I agree with what Brad said:

I agree with what Brad said: I never had a negative slope on my HG line. It would indeed be useful to know what instrument you are using. The slope is large, about twice as large as my oldest instrument. Is this an isoprime? I noticed that your correlation coefficient is somewhat low too,nothing alarming, but I typically had r2 of >0.90 when I was still using HG corrections.

Perhaps a first question to consider is are you sure that your heated gases are well heated, and that all of the vials were vacuum tight? I know Hagit had issues with some earlier glasses she was using, which would start exchanging after a while (maybe she can bring additional comments).

If you can rule the gases out (i.e. if you are sure they are well prepared), then I would try to measure the off-peak background (do a scan) with gas in the source. It would be interesting to see how your backgrounds compare between a light and a heavy gas. You might also want to expend the range in d47 of your gases, just to see how your machine behaves away from the WG value, perhaps with a negative gas.

Are you using Oztech as your working gas (WG)?

Good luck!

Cedric

John Eiler
Offline
Joined: Aug 23 2013
It is possible to create a

It is possible to create a negative slope in this empirical ‘heated gas line’ space simply by tuning the source to a highly aberrant set of potentials. I’ve forgotten the details, but I know we’ve made such slopes just by playing with the extraction potential and perhaps the shield or r-plate.

 

In any event, the important thing to remember is that the heated gas slope is an empirical correction for a background that is present when gas is flowing but absent when gas flow is interrupted (sometimes referred to as a ‘pressure baseline correction’, though I have no idea where that phrase comes from!). The usual problem we all contend with is that the instruments we use have a negative background when ions are arriving at the detector array, perhaps due to negative ions that are sputtered of a detector edge or other metal surface by one of the major ion beams. Some labs observe this negative background directly and ‘peak strip’ it from the signal. The more canonical approach is to calibrate the effect out through measurement of a heated gas slope (recognizing that it will be valid only for a narrow range of instrument conditions and source pressures). One could make arguments for and against both approaches (and their variants), and all things people have tried to date can yield accurate data when done properly, so I would say it is a matter of taste.

 

 

It is less clear to me exactly how you produce a negative slope, but I have done it myself as I described, and I imagine it is due to a positive background that correlates with beam intensity (scattered positive ions? a very diffuse tail to the ion beams?). 

Yangrui Guo's picture
Yangrui Guo
Offline
Joined: Jun 26 2015
The instrument I'm using

The instrument I'm using right now is Isoprime100 coulped with DI inlet system. When we run the clumped isotope measurement, the internal precision is always around 0.015 permill on d47 and 0.013 permill on D47. Is there any possible that one can get the results higher than this by using this instrument? Of course the precision would be higher if the we increase the Mass 44 intensity. But currently we can only make the current intensity to 80 mA on Mass 44 cup. If we run the measurement with a current higher than 80mA, the Mass 46 ion beam would be saturated. As I know, the pretty much the same instrument Rosenheim et al used at 2013 has resistors configuration different from we have. But I was been told that your lab actually upgarded the old Isoprime into the new one with 100V head amp, which has the same resistors configuration as we have. Is taht so? If that so then I was wondering how the precison you generally acheived by using the new Isoprime. Well, I'm glad that you mentioned the PBL correction. But I'm quite embarrassing to talk about this. Becuase  I actually  don't know how to do it. With all my due respect I feel like the correction procedure presented in your paper(Rosenheim et al 2013) is really compilcated and impractical. It's not sraightforward I think. As for the odd HGL I built,  I think the key to the problem would be associated with the off-line system we used for prepare sampe, such as some trace water absorbed on the tube wall or contaminants. Now I'm working on this and I hope I can get the normal HGL  next few weeks.

Yangrui Guo
Yangrui Guo's picture
Yangrui Guo
Offline
Joined: Jun 26 2015
The instrument I'm using

The instrument I'm using right now is Isoprime100 coulped with DI inlet system. When we run the clumped isotope measurement, the internal precision is always around 0.015 permill on d47 and 0.013 permill on D47. Is there any possible that one can get the results higher than this by using this instrument? Of course the precision would be higher if the we increase the Mass 44 intensity. But currently we can only make the current intensity to 80 mA on Mass 44 cup. If we run the measurement with a current higher than 80mA, the Mass 46 ion beam would be saturated. As I know, the pretty much the same instrument Rosenheim et al used at 2013 has resistors configuration different from we have. But I was been told that your lab actually upgarded the old Isoprime into the new one with 100V head amp, which has the same resistors configuration as we have. Is taht so? If that so then I was wondering how the precison you generally acheived by using the new Isoprime. Well, I'm glad that you mentioned the PBL correction. But I'm quite embarrassing to talk about this. Becuase  I actually  don't know how to do it. With all my due respect I feel like the correction procedure presented in your paper(Rosenheim et al 2013) is really compilcated and impractical. It's not sraightforward I think. As for the odd HGL I built,  I think the key to the problem would be associated with the off-line system we used for prepare sampe, such as some trace water absorbed on the tube wall or contaminants. Now I'm working on this and I hope I can get the normal HGL  next few weeks.

Yangrui Guo
brosenheim
Offline
Joined: Sep 15 2011
Yangrui, You've piqued my

Yangrui,

You've piqued my interest now that I know you are using an Isoprime.  We have implemented the new head amplifier and adjustible resistors, but we have only done so recently.  We are only currently gearing up to run heated gases and establish a reference frame on the instrument with these new resistors.  Have you tried running the mass 46 cup on low gain?  We have not tried this, but we have seen the limit of 80 nA too. 

The way we ran the pressure baseline correction is quite simple, and does not add much time to the analysis.  The first step is to examine the shape of your mass 44 peak on the low energy side, and find the low point in emission.  Ours has consistently been abou 41V less than the peak center.  Keeping all tune parameters the same, you can save another tune file with the accelerating voltage set X volts lower than the optimal tune.  Set up an MS file that uses this tuning, and run that after every separate sample acquisition.  You only need a snapshot - you don't need as long an integration on this parameter. This adds about 40 minutes to sample analysis time.

Then you have to get the raw data and subtract the current reading with the PBL tuning from the raw current readings at optimal tuning.  Raw data are available in the .dat files with each acquisition or the calculations tab of the datalog.xls files for each acquisiton.  The code we posted with the RCM article uses the .xls files accessed through Matlab, but only works on windows PCs and not linux-based PCs (Macs). As you can see, pressure basline correction methods are more complex than simply the empirical corrections that John mentioned and that many users stick to, however I like this method because you end up with both sets of data and two options for constructing the reference frame and transferring your data onto the theoretical scale.

Good luck and let me know if you have any questions about the routine.  Matthieu Daeron uses a different routine that wastes less time (and sample gas) then ours does.  You may want to also contact him. 

 

Brad

Brad E. Rosenheim

Assistant Professor

College of Marine Science

University of South Florida

Saint Petersburg, FL, U.S.A. 33701

http://www.marine.usf.edu/rosenheim/BER-Group.html

hagit
Offline
Joined: Dec 21 2011
Not too unusual. On a 253,

Not too unusual.

On a 253, our HGL had a negative slope all through the first year of the instrument. we started a bit less steep then you are and gradually approached a zero and then positive slope (which is also not very steep).

hagit

 

 

Yangrui Guo's picture
Yangrui Guo
Offline
Joined: Jun 26 2015
Everyone I really appreciate

Everyone

I really appreciate your reply here. Some of your opinions are inspiring and helpful. Now I like to discuss some questions with you here.

Running the mass 46 and 45 on low gain will affect the precision of d45 and d46 without affecting that of d47. However, due to the D47 calculation requiring the d45 and d46, the low precision of both of these will eventually affect the precision of D47. Therefore, I try not to run the mass 46 or mass45 cup on low gain. Besides, I’m afraid running high current intensity of the mass 44 (say increase it beyond the limit of 80 nA) would damage the filament.This may be not the truth, but I don't want to risk it.

Alternatively, I typically set the target beam to 80 nA (on mass 44), use a changeover delay of 12 s and 20 s integration of beam intensities. Each sample use 20 acquisition blocks and each acquisition have 5 cycles (the set of on-peak sample/reference measurement cycles).By doing these, I generally can achieve a precision around 0.013 permil on D47. However, problems still remain as shown below.

 

When I deal with the sample of low gas content, the sample depletion rate will be higher than the normal condition, which eventually impacts the precision. To this issue, I usually prepare the CO2 gas sample within the quartz glass tube (1/4 in, 18cm length ) under a gas pressure near 1Bar to make sure the gas volume is sufficient to avoid the depletion problem. But what if the gas volume is too high (only by using the Pirani gauge to rule the gases out is somewhat rough), even though the data are in high precision? I think the heated gases I prepared are all sufficient to be free of the depletion problem but are on different gas pressure. Sure you're all right with this?

Yangrui Guo
Mathieu Daëron's picture
Mathieu Daëron
Offline
Joined: Sep 17 2014
Hi all, I concur that the

Hi all,

I concur that the negative slope in itself is not a problem. Our latest slope is around -5.6e-3 (R^2 = 0.97, N = 22).

For the record, I'm also using an Isoprime, with amplifiers 45 to 49 set on high gain. Main beam intesity is 40 nA, and for each replicate we measure the backgrounds in all cups on both sides (ref/sample) at the beginning and end of each acquisition, by changing the accelerating voltage value. We use eight such acquisitions, providing 32 independent background mesurements for each cup. As expected, background values are almost completely predicted by main beam intensity over the course of one replicate (3 hours time), but change slightly over longer time scales (days or weeks). Keeping track of these backgrounds as described above, gases and carbonate standards remain fully consistent for over one month, with external reproducibilities between 0.012 permil (typical for carbonates) and 0.015 permil (typical for gases).

Regarding the pressure/amount of heated gases you use, I would be careful to try and match the amount to that of your samples, because there have been reports of sample size nonlinearities related to the purification preocedure.

Best,

-- Mathieu

stefanob
Offline
Joined: Sep 18 2011
Hi everybody,    for a

Hi everybody, 

 

for a negative slope you must have positive backgrounds on your mass 47 cup,  would be interesting to see a scan of your peaks, and I would suggest you try a background correction for it.

cheers

Stefano