Thanks for the suggestion - I wish that was the fix. I've done the column eval/auto leak check routine with each column installation, to no avail. 

This is exactly the type of oversight that I am trying to eliminate as the culprit, so it would be great to know if anyone else has recently replaced a similar column? 

Thanks! Natalie 

On Thu, Apr 6, 2017 at 7:14 AM, Mike Elefante, Elf Analytical Services <[log in to unmask]> wrote:
Natalie,

 The Trace GC has a leak check and column characterization built in. If these are not done correctly the flow through the GC will not be correct. After any column change or maintenance both need to be done. The flow controller has no way to measure the column flow. It calculates what the pressure should be for the column last characterized. It can only measure the Spilt, Purge and Total flow. So it gives a column flow readback that is a calculation based on characterization not measured flow. I just worked on a Trace that said the flow was 1 mil/min and actual flow was measured at 5 mil/min.


Mike Elefante, President
Elf Analytical Services, Inc
 



-----Original Message-----
From: Natalie Wallsgrove <[log in to unmask]>
To: ISOGEOCHEM <[log in to unmask]>
Sent: Wed, Apr 5, 2017 9:31 pm
Subject: Re: [ISOGEOCHEM] AA CSIA peak separation problems with new column(s)

Thanks for all the response! 

This issue is with our d15N AA CSIA set up, and we use the TFAA derivatization method. I've been testing the system with a suite if 14 pure AA's that have been derivitized, and the same suite injected on the old column separated tyr/lys well, but not on the newer columns. I swap the old column back in and I am able to get decent separation again. One of the BPx5 columns has slightly better separation than any of the other columns. Seemingly endless futzing with temp program and flow rate has not yielded a satisfactory separation to date. Sge and I went back and forth for ages, and they even sent me a new column in the efforts. They assured me that nothing has changed in the manufacturing, etc. In the end they just suggested that I try something other than the BPx5 (which is when we tried the DB5). Both the BPx5 and the DB5 show the same problem. I was kind of hoping that perhaps I was doing something stupid at the installation. With one of the columns I even tried a 12 hour initial conditioning instead of the 3 hour - same results. 

There is an observed broadening of the peaks, especially compared to the nice sharp peaks we were getting when the old column was newly installed back in 2011. Perhaps its the supply chain or something. 

Thanks for all the suggestions and food for thought! 


On Wed, Apr 5, 2017 at 1:50 PM, Robert Panetta <[log in to unmask]> wrote:
Aloha Natalie,

To understand, you install a new column and it fails. So you put the old one back and it works when injecting the exact same sample, correct? Is the bad seperation just overlap, or do they exhibit broadening/tailing?

If it's overlap then the new columns probably have slightly different dimensions or backpressure characteristics so playing with temperature or flows would do the trick. If it's broadening and/or tailing, then odds are you suffer from more active sites in the new column (could be due to a new manufacturing method, a change in the supply chain, relaxed QC, bad batch, etc.). Sigma used to sell an HDMS/BSA/TMSI cocktail for recovering packed columns, but I'm not sure if it worked for capillary columns. You could make your own cocktail and pray it fixes things, but I'd put more research into that one. Restek has the Rxi, which they advert as the "most inert column in the market". Could be worth a shot if it's an active site issue that's been introduced.

What does SGE have to say about the change in performance, anyway?

Malama pono,
Robert

p.s. chromforum.org/ is a great resource for this one





Robert

On Wed, Apr 5, 2017 at 5:08 PM, Natalie Wallsgrove <[log in to unmask]> wrote:
Aloha All, 
I was wondering if anyone else has observed trouble with separation between tyrosine and lysine on newly purchased columns? Our aging column in one of our Trace GCs is starting to show its age, but we've tried several new columns and consistently get bad separation between tyr and lys (all other AAs look fine). We have historically used SGE BPx5 (60m x 0.32mm x 1.0um) columns with success, but we've tried multiple new BPx5 columns as well as an Agilent DB5 (also 60m x 0.32mm x 1.0 um) with the same results. 

Has anyone else had this trouble? The old columns were purchased in 2011/2012, and all the troublesome columns have been purchased since late 2015. 

Thanks! Natalie 

--


Isotope Biogeochemistry Laboratory
University of Hawai'i at Mānoa
Lab Manager




--


Isotope Biogeochemistry Laboratory
University of Hawai'i at Mānoa
Lab Manager



--


Isotope Biogeochemistry Laboratory
University of Hawai'i at Mānoa
Lab Manager
[log in to unmask]
(808) 956 5362
http://www.soest.hawaii.edu/GG/isotope_biogeochem/