VFA speciation

[adammay11]

Hi,

I am wondering if anyone can help me on this problem . I am 90% of the way to a solution but I am not 100% satisfied .

What I am running is an analysis of VFA (Acetic-Hexanoic) from waste water anaerobic digestion processes simultanerously . I run a Shimadzu 2010Ultra SE GC-MS with a carbowax column and Hta 200HT headspace autosampler . The program runs between 70-190oC and the the program runs just short of 6 minutes .

The samples are centrifuged and filtered and acified with Nitric acid to help fix the sample and prevent changes in the specimen . We then make up the sample in a 20ml standard headspace vial with a 20% NaHSO4 salt to encourage the vapours out of the liquor . The total sample size is 6 ml leaving a 14 ml headspace and we incubate at around 80oC for 30 minutes .

The problem is some times I get good reproducibility on each of the acids (<10% CV) concentrations then for seeming no reason I will get one of the acids of those tested will be 30-40% CV difference between repeat samples .

Each of the repeat injections are from spearate sample tubes (no double penetration here - ) and all samples are made up from the same bulk . Concentrations are calculated against a 7 point cailbration curve (n=5) the r values for the calibration curves are all pretty much

Can anyone sugest anything that I can try so I can improve reproducibility . Also the peak shapes are a problem . No matter what I do there is a tailing edge on the peaks and the peaks seem quite broad . I am wondering if the tailing peaks are due to interactions on the liner and if derivitising with something like octaflouropentyl chloroformate would help .

If anyone has any thoughts or sujestions I would be very greatful .

Thanks

Adam

[Peter Apps]

Hi Adam

Welcome to the forum.

It's in the details:

Column dimensions, gas flow rates, inlet liner, inlet temperature, split ratio (if any) and how are you connecting the headspace transfer line to the inlet (or is it a syringe-based headpsace sampler)?

Do you check the pH after the nitric acid is added, and how much are you adding ?

Around 80C - sample temperature is about the most impotant parameter in static headspace analysis, is it repeatbaly 80C ?

I think that you are talking about repeatability, not reproducibility - you are doing replicate injections in one batch ?

The puzzle is that only one acid is affected. Is it consistently the same one ?

Is there a possibility that you have esters in some of the samples ? - acid hydrolysis would liberate the free acid with time and temperature.

What is the n=5 in your calibration ?, 5 replicates at each of seven levels ? If so what is the repeatability of the replicates ?. What is the matrix for the standards ?

Peter

[adammay11]

It's in the details:

Column dimensions, gas flow rates, inlet liner, inlet temperature, split ratio (if any) and how are you connecting the headspace transfer line to the inlet (or is it a syringe-based headspace sampler)?

The Column is a Shim Wax column (30m x 0.25mm X 0.25um), the inlet liner is a straight glass liner, temperature 200oC, Split ratio – Running in splitless mode though we have had the same problems with split (1:20, 1:15) we are connecting the headspace via a syringe (1.7 ml draw up 1.5ml injected) high pressure injection system is off head pressure is 23.2psi but run under flow control with a total flow 44.3ml/min column flow 2.58ml/min with linear velocity 58.9cms-1

Do you check the pH after the nitric acid is added, and how much are you adding?

Yes we do check the pH and the drop Is to pH2 ideally. The nitric has been a recent addition and is something that seems to have no real affect on the results (we were having this problem before adding the acid)and is added to try and fix the samples . We are adding 2% Nitric to the specimen bulk where repeated samples are taken from the specimen.

Around 80C - sample temperature is about the most important parameter in static headspace analysis, is it repeatable 80C?

It is set as 80oC in the sampler . I believe that is repeatable .

I think that you are talking about repeatability, not reproducibility - you are doing replicate injections in one batch ?

Yes it probably is repeatability I am concerned with we are doing replicates on a single bulk specimen with discrete samples being drawn from that specimen bulk.

The puzzle is that only one acid is affected. Is it consistently the same one ?

No, although it seams to be the more volatile acids are affected more commonly (acetic and propionic) which is a problem as those are the components we are most interested in .

Is there a possibility that you have esters in some of the samples? - acid hydrolysis would liberate the free acid with time and temperature.

It might be an issue but given we are drawing off the same bulk, and the incubation period and temperature conditions are all automated, I would not have thought that the presence of esters would affect specimens in a similar way . But it is something to consider.

What is the n=5 in your calibration?, 5 replicates at each of seven levels ? If so what is the repeatability of the replicates?. What is the matrix for the standards?

The n=5 is replicates at each of the levels and the %cv is around <10% for each of the standards . The residuals however for the line fit are around 25% . It is not fantastic.

If there is anything else that I should let people know about let me know .

Thanks

Adam

[Peter Apps]

It's in the details:

Column dimensions, gas flow rates, inlet liner, inlet temperature, split ratio (if any) and how are you connecting the headspace transfer line to the inlet (or is it a syringe-based headspace sampler)?

The Column is a Shim Wax column (30m x 0.25mm X 0.25um), the inlet liner is a straight glass liner, temperature 200oC, Split ratio – Running in splitless mode though we have had the same problems with split (1:20, 1:15) we are connecting the headspace via a syringe (1.7 ml draw up 1.5ml injected) high pressure injection system is off head pressure is 23.2psi but run under flow control with a total flow 44.3ml/min column flow 2.58ml/min with linear velocity 58.9cms-1I do not know this type of heaspacer, but onthe face of it you are injecting 1.5 ml splitless with a flow to the column of 2.6 ml/min - which will give you peaks more than 30s wide before they even start moving down the column. Split injections should give you narrower peaks, and the height will stay about the same until you get to high ratios. What is the speed of the injection ? (in ul/s). Your volume flow rate is more than double the optimum (assuming that you are using helium) - is there a specific reason for that ? - this needs a high inlet pressure, which compresses the gas in the headspace syringe when the needle pierces the septum and might lead to poor injection repeatability

Do you check the pH after the nitric acid is added, and how much are you adding?

Yes we do check the pH and the drop Is to pH2 ideally what is the actual pH ?, not the ideal pH, for the stronger organic acids I would aim for a pH of 1. The nitric has been a recent addition and is something that seems to have no real affect on the results surprising to say the least, what is the pH of the samples before you add the acid ?, and please tell me the result of an actual measurement, not what it is supposed or expected to be ! (we were having this problem before adding the acid)and is added to try and fix the samples . We are adding 2% Nitric to the specimen bulk where repeated samples are taken from the specimen.

Around 80C - sample temperature is about the most important parameter in static headspace analysis, is it repeatable 80C?

It is set as 80oC in the sampler . I believe that is repeatable .

I think that you are talking about repeatability, not reproducibility - you are doing replicate injections in one batch ?

Yes it probably is repeatability I am concerned with we are doing replicates on a single bulk specimen with discrete samples being drawn from that specimen bulk.

The puzzle is that only one acid is affected. Is it consistently the same one ?

No, although it seams to be the more volatile acids are affected more commonly (acetic and propionic) which is a problem as those are the components we are most interested in . these need the lowest pH

Is there a possibility that you have esters in some of the samples? - acid hydrolysis would liberate the free acid with time and temperature.

It might be an issue but given we are drawing off the same bulk, and the incubation period and temperature conditions are all automated, I would not have thought that the presence of esters would affect specimens in a similar way . But it is something to consider. If I understand correctly you acidify the bulk, then take out the specimens for headspace, so different specimens have been acidified for different times by the time they come to be run

What is the n=5 in your calibration?, 5 replicates at each of seven levels ? If so what is the repeatability of the replicates?. What is the matrix for the standards?

The n=5 is replicates at each of the levels and the %cv is around <10% for each of the standards . The residuals however for the line fit are around 25% . It is not fantastic. What is the matrix for your standards ? - what are they dissolved in ?

If there is anything else that I should let people know about let me know .

Thanks

Adam

Peter

[adammay11]

Hi Peter,

Yep the massive injetion size is a BIG problem but until you pointed that out it really had not struck me how much of a problem that was . The issue is the VFAs are in small amounts <20mg/l and to get into the gas needs quite a big injection size .


I have talked this over with the chromatorgraphy expert of Shimadzu and he had increased the column flow I think to try and improve the peak shape for the chromatography, but I get your point it is helium and it shoudl be around 1.5ml/min . The head pressure is also quite high at 23psi. The injection speed is 20ml/min.

Although I said ideally it should be around 2.0 infact the pH adjustments have been to pH 1 and slightly less (0.97) was the lowest after correction and prior to correction was ph 7.4-8.04 .

The samples have been acidified for different times but the difference in the times is in the region of 6 minutes . There is a conditioning period but the autosampler is set up to run the conditioning periods concurrently where possible so the actual delay between each injection is minimised .

Thanks again for your help ... it is appreciated

[Peter Apps]

Hi Peter,

Yep the massive injetion size is a BIG problem but until you pointed that out it really had not struck me how much of a problem that was . The issue is the VFAs are in small amounts <20mg/l and to get into the gas needs quite a big injection size . But you say that you have the same problem with repeatability with split injections - which implies that with a split you can still integrate the peaks, so you are not working all that close to your LOD, and could afford to inject a smaller volume, or the same volume with a split


I have talked this over with the chromatorgraphy expert of Shimadzu and he had increased the column flow I think to try and improve the peak shape for the chromatography, but I get your point it is helium and it shoudl be around 1.5ml/min . The head pressure is also quite high at 23psi. The injection speed is 20ml/min. Jacking up the gas flow to make the peaks look narrower by moving them through the detector quicker has probably been counter-productive, because it will have made it more difficult to get repeatbale injections against the higher inlet pressure. With an injection speed of 20 ml/ min you are trying to increase the flow by nearly 8 times, this will cause a pressure pulse in the inlet which will send acid laden gas in all sorts of unwanted directions. I think that you need to take several steps back, first try to get sharp peaks (a wider peak is no easier to detect than a sharp one of the same height) - put the inlet pressure down to whatever gives 1 - 1.5 ml/min volume flow through the column, set a split ratio of 20:1, inject 1 ml at 20 ml/min - how do the peaks look ?

Although I said ideally it should be around 2.0 infact the pH adjustments have been to pH 1 and slightly less (0.97) was the lowest after correction and prior to correction was ph 7.4-8.04 . This should be OK, you still have not told me what you standards are made up in ?

The samples have been acidified for different times but the difference in the times is in the region of 6 minutes . There is a conditioning period but the autosampler is set up to run the conditioning periods concurrently where possible so the actual delay between each injection is minimised .

Thanks again for your help ... it is appreciated

Peter

[adammay11]

Standards are made up in pH adjusted deionised water that with 20% NaHSO4 . THe water is pH adjusted to pH 2.0 but I am wondering if that should be lower .

Thanks

Adam

[Peter Apps]

Hi Adam

I would lower the pH of the standards - it must be the same as the samples.

Peter

[Peter Apps]

Hi Adam

I forgot to ask - what temperature is the syringe ?

Peter