Chromatography Forum

Hello Chromatography Forum!!

This is my very first post and I am quite a n00b when it comes to chromatography. However, I don't think I'm asking super basic questions so I chose to put it in this forum instead of the "student projects" forum. Sorry if I goofed that up, Tom.

This may be a long post, so I apologize in advance.

So as the subject states, I will be analyzing volatile compounds in wine. The company I work for just ordered a refurbished Thermo Trace/DSQ II, so I haven't had the chance to familiarize myself with the machinery or calibrate my methods. I only have a rough guideline of what my methods will be like based on the literature I've read and the forum lurking I've done here over the past couple months. I have literally zero experience with SPME and GC-MS, as I am currently a Microbiology undergrad. So once we get the GC-MS I'm sure I'll be picking your brains for more information regarding method development and any problems I may run into. But in the meantime, I figured I'd run some things by the chromatography gurus here to make sure I'm on the right path. I'm going to number these points so that the replies can be easily referenced.

1) I plan to use HS-SPME to extract wine volatiles. We are not looking for any analyte(s) in particular. Quantification is important to us. There seems to be a lot of debate about how to go about that, or if it's even feasible, for complex matrices such as wine. I'm planning on using multiple internal standards - one for each classification of volatile compound in wine. For example, only compare peak area for a particular ester to the ester internal standard; and alcohol to alcohol so on and so forth... Labelled isotopes would be cool, however, there are hundreds of analytes that may or may not be present in a particular wine sample so the benefit of using a standard with identical properties to the analyte doesn't apply here.

2) Most literature seems to agree that the tri-phase fiber (PDMS/DVB/CAR) is best suited for wine volatiles and that's what I plan to utilize. I plan to use the Restek Rxi-5ms column with dimensions 30m x .25mm x .25um

3) I'm going to be building my own odor port. Since the odor port works under atmospheric pressure and the MS requires vacuum, I plan to use a long guard column in between the MS and odor port split. I guess the longer the better for better vacuum conditions? Restek makes a 60m guard column. I just don't know how that is going to affect the MS/odor port split proportions... That may also screw up how the compounds are transferred to the MS making for a funky chromatogram. I'm going to use a ribbon potentiometer interfacing through an Arduino to use as the odor intensity gauge. Our refurbished Thermo system is packaged with Xcalibur. I'm skeptical that we would be able have Xcalibur track potentiometer position so that we can look at the chromatogram and olfactogram in parallel. We were thinking of just using something outside of Xcalibur to track position over time. However, maybe we'll try OpenChrom since we can customize it. And it's free, so why not.

4) Long or short extraction? I read a few articles that spoke of what was referred to as "true" headspace extraction. This "true" headspace extraction calls for a very short (<= 1 minute) extraction time. The argument was that it's a better representation of the true relative concentrations of volatiles since SPME fibers become displaced with less volatile and bigger compounds in long extraction times so that the low weight volatiles are underrepresented. That doesn't sit well with me... It goes the other way around as well; a short extraction time is not going to capture the less volatile compounds. I'm leaning more towards the longer/typical extraction time - 15-30 minutes (tbd upon equipment arrival and method calibration). What are your opinions? It is very important to us to capture the true relative concentrations. And that leads to my next point

5) I'm a little unsure about using NaCl or some other salt to coax volatiles out of the wine. If doing so coaxes ALL volatiles out the same amount, then this will not affect relative concentrations and will be just fine for our application.

...

Annotation to 4 & 5) Obviously I'm thinking about this as I write this out. It just clicked that if I'm going to be using internal standards for quantification that the relative peak areas aren't really of concern. It would only be an issue if we were doing a purely qualitative analysis. The beauty of critical thinking... So I plan on doing a 15-30 minute extraction and I will add NaCl. I will also use agitation and heat to coax the volatiles into the headspace.

That's really all I can think of right now that I wanted to run by you guys. I'm sure other topics will come to mind. Some topics, I'm sure, will be brought up by your replies!

Thanks for reading! I'm looking forward to having chromatography discussions with the fine people of this forum!

-Ben

Welcome, Tom.

This is not going to work:

"3) I'm going to be building my own odor port. Since the odor port works under atmospheric pressure and the MS requires vacuum, I plan to use a long guard column in between the MS and odor port split. I guess the longer the better for better vacuum conditions? Restek makes a 60m guard column. I just don't know how that is going to affect the MS/odor port split proportions... That may also screw up how the compounds are transferred to the MS making for a funky chromatogram. I'm going to use a ribbon potentiometer interfacing through an Arduino to use as the odor intensity gauge. Our refurbished Thermo system is packaged with Xcalibur. I'm skeptical that we would be able have Xcalibur track potentiometer position so that we can look at the chromatogram and olfactogram in parallel. We were thinking of just using something outside of Xcalibur to track position over time. However, maybe we'll try OpenChrom since we can customize it. And it's free, so why not. "

If there is an extra 60m of empty tube after the split, there will be a 4-5 min discrepancy between elution times at the odor port and the MS (a 60 m column would have a dead time of about 2 min, but it only gets half the flow). You need short, narrow bore transfer lines (these are not "guard columns") going to both the MS and the port. Ideally the temperature of the transfer lines needs to be independent of the GC oven, otherwise the split ratio between port and MS changes as the oven heats up. This is easy for the line to the MS , because you can put the split close to the MS column connection, but the line to the port would need an independent heater. Please don't be offended if I say that setting up this kind of plumbing is not the kind of thing that you learn in microbiology, and you should look seriously at purchasing an off the shelf splitter and port if you want to avoid spending a lot of time fiddling with little pipes. The sniffer port itself needs a heater and a source of humidified air of course.

What is it that you need to find out about the samples ? - wine is one of the most intensively researched flavour areas, and you might well find that someone else has already done the work for you.

Peter

2. I would use a Restek Rxi-624SilMS for doing volatiles analysis, it has a thicker film and is designed to better separate volatile compounds. Plus this column has a upper temperature limit at or above 300C which will allow you to bake off any high boiling compounds that might hang up on the column.

As Peter said you will find doing the split before the MS quite complicated. With the MS being under vacuum it is going to want to pull in air through your guard column instead of sending some of the column flow out. You would need to have the analytical column at a large flow rate, with the transfer line into the MS being of a much smaller diameter and the split line going to the odor detector being large diameter and not too long. It also would work better using an actual open split interface which has a micro adjustable needle valve to control the flow to the split so you can optimize the amount going to the MS and to the split line. For this you would be looking at using a 0.53mmID column so you can run 5-10ml/min helium carrier flow and send 1-2ml/min into the MS and the rest to the odor detector.

Quantification will depend on how accurate you want the results to be. For super accurate results you need to have a calibration standard with each compound of interest along with internal standards. For less accurate but still better than estimation, you can quantitate against an internal standard that is similar, for ball park figures you can just estimate relative peak areas of each compound against the others. An example of the super accurate problem is that even if you inject the same amount of benzene and benzene-d6, you will get different area responses for the peaks, sometimes up to 10% difference, if you are looking at benzene-d6 versus methyl-benzene the response difference can be even larger.

I would approach the project by first identifying the analytes of interest and estimating their relative concentrations, then make or order standards with known concentrations of those compounds to make calibration curves to calculate the concentrations accurately.

You are on the right track and you will learn much more as you progress. We all started out the same way, and give our advice because we have already done things the wrong way ourselves.

I have such a setup. I have a Gerstel ODP 3. There is a split and the length and diameter of the capillary on each side determines what portion goes to the MSD and what portion goes to the sniffer port. If you make the capillary narrower and/or longer there is more resistance and more of the gas goes the other way (path of least resistance). There is a heated transferline to the ODP port where nitrogen makeup adds to the miniscule flow from the column/split and there is also air that bubbles through a water container to humidify the gas at the olfactory port otherwise the persons nose dries out quickly hampering his/her ability to smell the compounds.

I agree with the other posters a 624 or 1701 column should be used for small alcohols and esters or you will see bad peak shape of polar compounds on nonpolar stationary phase.

The salt helps tie up the water molecules so that it is easier for the polar compounds to escape into the headspace. The water molecules are engaging in polar bonds with the salt and less with your analytes.

I've actually analyzed a lot of alcohols using static headspace but for trace compounds like sulfur compounds I'd use SPME. Also since you are splitting between the MSD and ODP you need more sample so SPME would be a good choice.

Thanks so much! I knew the fine fellows of the forum would be helpful!

Our target analytes are anything that has a perceived scent in humans and exists above the odor threshold concentration in the wine matrix. So they are numerous and variable depending on the wine. What we are doing with this data I'm not sure I can tell you yet... I just got hired as the chemist for this startup company... I originally was going to use a different username including the company name but my coworkers kinda freaked out about that so I'm going to be extra careful now. lol.

Peter, we are going to try to use others' previously published data anywhere we can.

James, making calibration curves for so many compounds does not seem feasible. However, we don't need SUPER accurate quantification. So using internal standards alone will be accurate enough. Thanks for describing my options.

To all the posters, I knew I would be way off on the odor port. And thanks for the suggestions. I know the relative transfer line lengths determines the split ratio. However I think calculating that requires some knowledge in fluid dynamics which I definitely do not have. Thanks for reminding me about the path of least resistance; I understand it that way. I also read about making the air humid. I read one paper that seemed to discredit that the air should be humidified... But the general consensus says humid is good and it's nice to hear that you agree. If I had it my way, I would use a commercially made odor port. However, money is not abundant in the stage my company is in. This all started when our autosampler sales rep said that you can build your own odor port. So since then my coworkers have been like, "you can do that, right?" Luckily, that sales rep is also going to be installing our GC/MS and the autosampler and has confirmed that he will assist me in making an odor port.

Thanks again guys. I can't wait to come to you with more n00b questions! I also look forward to telling people I meet that I'm an analytical chemist. That's bada**.

You get full marks for ambition ! GC-olfactometry of natural flavours is about the second most challenging application for GC.

"Our target analytes are anything that has a perceived scent in humans and exists above the odor threshold concentration in the wine matrix."

That means that you are interested in several hundred compounds in each sample - when you get you sniffer port going you will find that there is a a continuous stream of different odours coming out of it - the challenge will be for your nose, brain and recording to keep up with them all.

Also, at the level that corresponds with what you can smell at the port the chromatogram will be full of peaks - it will be tricky to find an empty spot for an internal standard to elute in.

Are you going to be determining odour thresholds yourselves, or using literature values ? The literature values need to be taken with a pinch of salt. Being able to smell something at the odour port does not mean that it is above the odour threshold in the wine - even a relatively inefficient sampling method like SPME selectively concentrates some components so that you get proportionately more on the GC column than was in the wine.

This should not put you off - others have done it before, so you can do it again. But you do need to think very carefully about what it is you need to know about the wine flavour, and exactly how you can find it out given the analytical capability at your disposal.

Peter

You most definitely get an "A+" for effort.

All of what you talk about is possible but setting up the system is not a trivial affair. I have a GCO that we purchased from a commercial source (Agilent GCMS and I'll make a pitch for my pals at MOCON/Microanalytics for the sniffer plumbing and control software, check them out at http://www.mdgc.com). My system is a 2-column system with a dean's switch between them so that I can do a precolumn separation (non-polar, 5-phase) first and then heartcut whatever I want to the second column of completely different polarity (wax phase). It is rare that you get a coeluter on both phases. A flame monitors what's coming out of the first column and the MSD monitors what's coming out of the second column. The system takes the effluent from the second column to the MS via some narrow bore tubing (a fixed restrictor) and the remainder of the effluent from column 2 goes to the sniff port. I actually have the ability to sniff the precolumn as well. My game is off odor identification so the precolumn sniff feature helps me to identify the correct heartcuts that I need. The sniff port is also fed with humidified air so that your nose doesn't dry out while you're "rabbit sniffing" your way through the chromatogram. It's a very powerful instrument and if you're really serious about this type of work, I would highly recommend that you get a system that has been set up by the experts. You'll save a lot of time in the long run. If your company's trying to make money, time is money.

If you're "stuck" doing it yourself, I would buy an add-on sniff port like this one:

http://www.glsciences.eu/sniffer/gc-o.pdf

SGE used to make one that I've seen people have success with but they don't sell it anymore. I'm sure that with the correct combination of narrow-bore restrictor tubing and column dimensions you can make it work. It'll probably be some trial and error however and lots of venting/pumping down your mass spectrometer.

Best of luck to you! Again, I have to give you and "A+" for effort!!

rb6banjo, that's some serious hardware you have there !

Peter

It has helped me find the root cause of a lot of aroma problems over the years. I would be seriously handicapped in my role here if I didn't have it!

Peter, you have brought up a good point about the internal standards. There may not be an empty spot. I will have to think about this more. I have spoke to my superiors and it is the goal to focus on a set number of analytes in the future, they are just waiting for the GCMS system so that those analytes can be decided upon. So hopefully its a reasonable amount for calibration curves to be made up. We are going to use literature values for odor threshold. The only thing is that different literature has different values, sometimes vastly different... I know that just because I can smell a compound in the eluent does not mean that it exists above the concentration threshold in wine. However, compounds that exist in concentrations less than the odor threshold may still be playing part with all the additive and synergistic effects going on. More to investigate.

Rb6banjo, we got an estimate for a commercial odor port (I don't remember which at the moment) and decided that these pieces of equipment were out of our price range. The PHASER looks pretty cool, though. I have asked GL Sciences for a quote. Finger crossed that I can convince my team that it will be better for us in the long run.

Thanks again!

There was a webinar I attended by Teledyne Tekmar not to long ago: http://info.teledynetekmar.com/webinar- ... d_trap_301
You might find it helpful, because it talks about taste and odor, and lists some compounds. I believe they said Trichloroanisole was a common one for wine. This one if for water, but you should be able to relate it to a different matrix. Good luck

another webinar: http://info.teledynetekmar.com/webinar- ... _headspace
open access journals: http://www.omicsonline.org/2157-7587/21 ... -2-121.php

If it is any help I built my own odour port



Essentially it is just an open split interface - a low dead volume tee piece from Agilent. The column is a 0.53mm column. From the split a short length of 0.53 column goes up to the heated sniff point. The column to the MS is something like a 0.11mm capillary of about 1.5m - 2m in length - it actually sits just inside the wide bore capillary and the MS vacuum draws a fixed flow - the rest goes up to the sniff port

The other advantage of using a wide bore capillary is that the peaks are wider. With a narrow bore capillary the peaks are so narrow you can easily miss them whilst breathing out. Keep your breathing steady and natural - otherwise you can start to hyperventilate. Every so often move your head to one side to "refresh" your nose. The sniff port must be heated otherwise you will get condensation of analytes that linger.


I tried with and without a humidifier and didn't notice any difference. For recording I would have a colleague sitting watching the monitor and noting the time when I said "now" and recording my comment (don't pause to think too much - just say the first thing that the odour reminds you of. There is a delay of 4 seconds after you detect the peak to the MS responding

Good luck - it's a fun and interesting thing to do

Cheers

Ralph