An interview at the Agilent Folsom GC Column Manufacturing facility

Blog Post created by paulsalverda Employee on May 31, 2016

2016-0531_AgilentFolsomFront.jpgI had the opportunity to visit the Agilent Folsom GC column manufacturing facility.  The following is an interview with jasonellis,  Factory Customer Experience Manager,  and akv, GC Columns Product Development Chemist.  I hope that this will answer some questions you've always had!  If you have more, please reply here and we'll get some answers for you...


Paul: What can we do to help increase customer’s confidence in the use of our columns?

  • Jason: I think the issues are often around confidence in our manufacturing and test procedures, product specifications, and quality processes. We test each and every GC column manufactured and our product quality and specifications lead the industry. The results you see on the Certificate of Analysis (CofA) in the column box are the results from that specific unit when tested in Folsom. I think people don’t always understand this, or don’t think of the value that this brings to the product. We have a constant focus on product quality and continuous improvement here at the factory – we didn’t get to be the industry leaders by making poor quality products or by standing still. The factory technical and support team meets every morning to review any column test failures, product availability or delivery issues, customer returns, and proactively evaluate column QC test results to identify trending; we review each of these and respond immediately as appropriate. This is an essential part of how we ensure quality and availability of our columns for our customers.
  • Allen: Confidence also comes with the thorough understanding of the whole chromatographic process. One can’t be fully confident in any job with only bits and pieces of understanding. I remember when I was a purge and trap instrument person prior to coming to J&W Scientific, I could walk my way through the whole sample flow path up to when the sample reached the column, and then it became a “black box”. I was also an expert on the other end of the column because the company I worked for made GC detectors. I could talk traps and trap substrates, valves, nickel reaction tubes, mother boards, electrometers, and I was even pretty handy with ChemStation, but the column was just the thing in the middle of it all. When I came to J&W Scientific my eyes were opened to a whole other side of chromatography that I had not seen because I had never stopped to study or understand it – it just wasn’t my focus at the time. I also had to learn the intricacies of vaporization injection when I came to J&W. I had a great insight to headspace/vapor injection, but I had to study the split-splitless inlet when I had to start talking about it to support customers. My confidence and effectiveness improved as I gained expertise in these areas.


Paul: How can we help customers avoid immediately jumping to the conclusion that the column must be the root cause of a performance issue they are experiencing?

  • Jason: Education. Understanding the fundamental manufacturing process and chromatography in general to help guide them to what a column can and cannot do, and to understand how a column can fail and what can cause that failure. Every column we manufacture is installed into a GC and tested to verify how it performs using a challenging test mixture. Every column is shipped with its unique Certificate of Analysis report that documents its coating efficiency (plates), coating thickness (k), inertness (tailing, peak height ratio), selectivity (RI), and bleed. So we know that the column was performing well when it left the factory door, and columns don’t “go bad” while sitting on the shelf. When customers encounter a problem, it is important to follow a troubleshooting process to gather information to help find root cause. For example, if a peak shape issue arises after installation of a new column, what else might have changed during that maintenance event – installation of the column also involves cutting the ends (clean cuts are essential), selecting proper ferrules, setting proper installation distances into the inlet and detector, and often involves replacing the septum, gold seal and/or inlet liner. Every one of these variables can contribute to performance issues, so carefully walking through the process and changes can help identify what might be contributing to the performance issue.
  • Allen: Not to over generalize by saying “everybody does this,” but it is easier to simply blame what we don’t understand than to try and fix what we don’t understand. I completely appreciate that when a customer’s system goes down, time is of the essence in order to get back up-and-running samples, and sometimes this results in replacing multiple components in the system at the same time in an effort to restore performance quickly. However, ultimately the best long-term solution is for customers to learn how to effectively troubleshoot chromatographic problems, because this helps to focus in on root cause for an issue and speed up the troubleshooting process. For example, we will get complaints about column bleed and very frequently the chromatogram the customer provides shows discrete chromatographic peaks that are being claimed as “column bleed”. Chromatographic peaks mean that the column is separating compounds that were introduced to the front of it (usually inlet contamination or septum bleed). A column cannot create discrete chromatographic bleed peaks on its own – the phase simply does not do this. Column bleed is seen as a baseline rise during the oven temperature program, and the baseline should really only increase as you approach 20-30C below the column upper temperature limit. We can help the customer correctly identify that this is not column bleed by talking about how chromatographic peaks arise – this maintains confidence in the quality of their GC column – but they could solve their own problems by learning how to identify the source of typical issues, and in the long run this independence results in being able to solve these problems faster. Learning how to troubleshoot problems by half-splitting the system should be part of every operator’s education. They will be happier because they know how to solve the problem on their own (for example do better sample prep, improve their system maintenance, etc.) and their management will be happier because their system will be running billable samples sooner.


Paul: What does it take to kill a column?

  • Jason: It all depends on the application. It might take a lot to kill a column for some applications (e.g. simple TPH analysis of hydrocarbons), or it might not take much (e.g. one injection of a “bad” sample when running EPA 8270 analysis). It is tied to the level of performance needed from the column for the customer’s application.
  • Allen: We can point to a few specific “poisons” to be aware of. The biggest one is oxygen. Ask any polymer chemist and they will tell you that even one molecule of oxygen is bad at high temperature, and it just gets worse as you add more oxygen molecules to the equation. Even at low ppm levels oxygen will cause phase damage at elevated temperatures (it does not happen at room temperature) – oxygen facilitates the unzipping of the polymer chains, resulting in high bleed, phase loss and inertness problems. It is important to note that you will not see ppm levels of oxygen on your MSD air/water check, so that is not a good diagnosis tool for small leaks or poor quality gas. Use only high purity carrier gas (99.9995% or better) and be very careful about leak checking all of the plumbing connections in the lab between the cylinder and the instrument. Gas scrubbers are often a good idea as well, as they can help prevent any oxygen or other contaminants from reaching the GC – they’re like buying insurance for your system (so to speak). If you do use gas scrubbers, make sure they are on a replacement interval and/or indicating traps are in-line – traps must be changed before they expire. The only thing worse than dirty gas is clean gas that is being made dirty by going through spent gas purifiers. It is also important to understand that it is not a good strategy to try to save money by over-stretching the usage life or using the lowest-grade/cheapest available for inexpensive essential items that protect the health of the GC column and system – for GC columns this is often observed with inlet septa, ferrules, etc... You can easily kill a $500 GC column by neglecting to change a $1 to $2 inlet septum, when that septum divided by 100 injections only costs 1¢ to 2¢ per injection. Use an Agilent autosampler syringe and get 150-200 injections/septa. Change out the septum nut periodically (and don’t over tighten it) and get 250+ injections/septa.


Paul: Why change out the septum nut?

  • Allen: Our autosamplers are amazing, but they cannot put the tip of the syringe directly onto the exact middle of septum nut every single time – 100.0% accuracy is just not possible. The cone at the top of the septum retaining nut is there to guide the needle into the septum. Over time, because the syringe needle rubs along this cone, it wears and can leave extremely small metal particles on top of the septum. The needle can then carry these particles into the liner. These can cause active sites that are difficult to troubleshoot. Also, because the needle guide is not worn on a new septum nut, this helps get the needle back through the same “target” more precisely, helping the septum to reseal and giving the customer longer times between septum changes.
  • Paul: I’m not sure many know that.
  • Allen: It’s a piece of the instrument that is easy to ignore but it can have a significant impact on results.

    Getting back to things that can kill your GC column… Chemically, inorganic acids and inorganic bases are bad, and some strong organic oxidizers are not good if injected at high oven temps or allowed to be in the column at high temperatures (>100C). Derivatizing reagents aren’t bad for the column per se (assuming you are using an appropriate stationary phase), but they will make it look like your column is bleeding if you do inject them, and they also can play havoc on deactivated glass wool in liners (which can make it look like you have a dead column). We also recommend changing the inlet liner and cleaning the bottom of the injection port periodically because material collects down there, including fibers of glass wool from the liners, septum particles and sample residues that can make your system look like you have a dead or damaged column.
  • Jason: The secret to eternal column life is … If you want your column to last a long time, don’t ever install it and inject any samples. Of course this advice isn’t practical – it’s meant as a joke – but it emphasizes that “things happen” when the column is in use that contribute to aging of the column. Column lifetime is the result of cumulative experience while in use – this includes inorganic and organic content of samples and standards, carrier gas purity, plumbing and system leak rate, thermal history, and maintenance practices.
  • Allen: Still, even if you think the column’s dead there are things that can give you more life out of the column.


Paul: How do you know for sure that a column is dead?

  • Jason: At a basic level, it’s probably dead if you trim the ends and bake the column and performance isn’t restored. Trimming of the ends must be significant – don’t trim only 1 inch off the end and expect that to effectively remove all the contaminated or damaged tubing – you generally need to trim a larger piece (maybe 0.5 to 1.0 m) to see an effect, particularly if you are running dirty samples. Also, regarding column baking – there is no need to do overnight bake-outs of your column; if the contamination you are trying to remove isn’t coming out after 2-3 hours at max temp, then it probably isn’t going to come out tomorrow morning after an overnight bake. That is similar to trying to clean up a frying pan by putting it on your stove and turning the burner up high.

    You can also try reversing the direction of the column and running some test injections to see what performance looks like – this can work because in temperature program analysis the majority of analyte interaction occurs near the front of the column. Column solvent rinsing can work wonders with bonded and cross-linked columns, but I have found that few customers do this in real life. Agilent sells a column rinse kit if you do wish to try column rinsing (P/N 430-3000, about $160 USD list).
  • Allen: For how to assess if a column is dead, we like to talk about the customer doing some sort of chromatography test when the column is new and in good operating condition, then this test can be used as a convenient reference for future troubleshooting and/or column assessment. The test needs to be simple and informative – it should be meaningful for the customer’s application. Column testing can be one compound that tests for a simple parameter like peak shape, or an entire EPA method target list, but simpler is better. The chromatogram needs to be obtained with a defined set of instrument parameters and there needs to be defined acceptance criteria that are meaningful for the application. Without a way to test the system, it’s like looking at an AC outlet and guessing that the reason your light bulb isn’t very bright is because there’s only 75 VAC, not 110. You need a way to measure the problem – a volt meter, in this example. I know Jason would agree that a system bleed test when a column has just been installed and conditioned is easy and it is the best way to know later if your column is getting damaged. The bleed test is simply running the customer’s analysis temperature program without injecting anything (take the syringe out of the autosampler….but don’t forget to put it back in afterwards!) – some people call this an instrument blank or system blank. Saving that file for comparison and doing an overlay comparison will help in a diagnosis later on in the column’s life – just like the test mixture injection discussed above.

2016-0531_JE and AV_FolsomTour_Fig1.jpg

Paul: Do you have a bleed profile that shows the difference between when inlet maintenance is done properly and when it is done poorly?

  • Jason: Here are two slides that emphasize the importance of cleanliness when doing inlet maintenance (see figures at the right) – it shows how a little bit of contamination goes a long way in system background. In addition to contamination risk, there is also the possibility of a leak in the inlet after doing maintenance – this will manifest itself by high column bleed; generally there will be upward drift in baseline when holding at a high temperature (you’re seeing the damage to the phase resulting increasing column bleed over time). Of course leaks in the inlet can also cause sample introduction issues (low or variable response) and in severe cases issues with retention time stability. You can also get peak shape issues related to poor column cutting practices and/or inconsistent or incorrect installation distance.

2016-0531_JE and AV_FolsomTour_Fig2.jpg


Paul: Do you have procedures that customers can use to prove that a column is still good?

  • Jason: Some people try to replicate the column test shown on the column CofA, but it is more meaningful to the customer to have their own test mixture and conditions; something that is relevant to their application. We discussed this in one of the questions above. It can also be helpful to try to isolate the problem. For example, try moving the column to a different inlet or detector or instrument, if available. If the system has a headspace or purge-and-trap instrument on the front, try some direct injections to simplify sample introduction for troubleshooting. For ghost peak or baseline problems, sometimes a condensation test and/or a jumper-tube test are helpful in identifying issues with carrier gas or inlet cleanliness. We also recommend looking at peak shape for unretained compounds (injected in split mode) – an unretained peak should be razor sharp; if it isn’t then that’s a sign of a flow path problem. For details on these and other troubleshooting recommendations, I suggest looking at the material at the links below (especially the how-to videos):


Paul: Why do we sell HP-5, DB-5, and VF-5 columns? Aren’t they all the same?

  • Jason: There are subtle differences in manufacturing processes between the different types of columns, and these can result in subtle differences in performance for certain compounds or applications. The HP and DB columns are manufactured in Folsom, California, and the VF are manufactured in Middelburg, Netherlands. Agilent is in the unique position of having columns in the portfolio originally from three different column manufacturers: Hewlett-Packard, J&W Scientific and Varian. We continue to sell all of these columns because there are customers with these columns written into their methods, and we don’t want to disrupt their lab operations by making them change. In a situation where you aren’t sure where to start, we recommend using the HP-5msUI or DB-5msUI to help see the benefits of an Ultra Inert flow path.
  • Allen: Did you know that HP means “High Performance” not Hewlett-Packard, and DB means “Durabond”? DuraBond (DB) was the J&W Scientific prefix for all of the cross-linked and bonded stationary phases. You may not know that J&W Scientific figured out the cross-linked stationary phase before any other company. So any column that has the DB-prefix is a cross-linked and bonded column (stabilized and solvent rinse-able). Most of the popular “HP” columns are also bonded and cross-linked (for example HP-1, HP-5, HP-1ms, HP-5ms, HP-Innowax, HP-50+, etc.), but some are not (HP-101, HP-17, HP-88).


Paul: What exciting new products can you tell us about?

  • Jason: DB-WaxUI is a significant new product launch in the portfolio, and represents the first commercially available polyethylene glycol column specifically designed for Ultra Inert performance. A lot of work went into developing the chemistry processes required for DB-WaxUI, and we have received lots of positive feedback from customers who have evaluated the column so far. It’s an exciting product – its performance and stability are well above our competitors and its selectivity has been proven to be the same as DB-Wax.
  • Allen: Ultra Inert in general is always exciting because the benefits it can bring to system performance and inertness reproducibility are substantial. Agilent has set the bar for our competition with respect to inert system performance, and the results can be dramatic for many customer applications.


Paul: Where can customers ask questions about columns?

  • Allen: They can post questions on the Agilent Customer Community page ( (Ed. note: The best place for questions relating to columns and other consumables is the Consumables area. For help and tips, see How to start a Discussion. - jkurutz, Community Manager).
  • Jason: [For those who would like to use the phone], in the US and Canada, they can call 800-227-9770 option 3,3,1 and talk to a specialist.