CSR-LVSI instead of Internal Standard for DRO by GC-FID

[LALman]

Edit: I'm convinced its better to try CSR-LVSI combined with method 3511 microextraction for my DRO analysis. While this discussion starts with the merits of adding internal standards to DRO analysis. I become convinced by the other posters to forget internal standards and try CSR-LVSI and method 3511 instead.
/Edit

I'm trying to improve my DRO method from simply using external calibration with o-terphenyl, 1-Chlorooctane, and 1-Chlorooctadecane surrogates added during sample prep. I'm looking for an appropriate internal standard that is robust for doing DRO by GC-FID. My instrument is a 6890 that I upgraded to network and I am running chemstation d.02. I'm running a 50m x 0.2mm x 0.33um HP-5 column with UHP hydrogen carrier. DRO C10-C28 comes out from 3.9min to 14.0min.
4.52min 1-chlorooctane
9.96min o-terphenyl
10.82min 1-chlorooctadecane
I run the time out to 20 min and total run time is 26min/sample.
cal curve (ppm) hexane solvent for standards and samples injecting 1uL...
100K, 50K, 25K, 10K, 5K, 2.5K, 1K, 0.5K and RDL is 100 ppm
I am making up the last 4 standards to have 125, 100, 50, 20 ppm of the 1-chloro surrogates so as to calibrate the surrogates. It seems to me that the peaks will be difficult to quantitate in 10K and higher concentrations of DRO. The o-terphenyl I have been calibrating by prepping and extracting blanks for 200, 100, 50, and 25 ppm and running those spiked blanks to make my curve. I spike 1mL of o-terphenyl in acetone and 10-20uL of chloro- mix in pentane to give 50 ppm at the instrument for samples.

5-alpha-Androstane seems to be the internal standard that Google brings up. Since DRO has such a hump I was hoping for something that acted like DRO but fell on the high temperature side of the hump so that I would not have problems quantitating the internal standard. On the other hand, I suppose a valley to valley integration for the internal standard would likely take care of the DRO hump.

[Steve Reimer]

Is there a reason you want to add an internal standard to the method?
I think you have seen why it will be difficult. Looking for an internal standard that elutes after the diesel range will still leave you with a challenge when motor oil is present.

[LALman]

The main reason is to try and compensate for loss of hexane when doing seperatory funnel extractions on water samples. I am struggling to get good surrogate recoveries from 1L volumes. I get in oil field brines and hexane tends to disappear into the brine. So the small volume I recover from 4mL of hexane added to 1L of brine is the problem. Have I lost it to the brine, or did it evaporate? I thought perhaps an internal standard might let me compensate for these factors. I would use MeCl instead but its a VOC target and most of my business is VOC by P&T GC-MS. I've considered putting together a blow down setup to get my final volume precise instead of just using what remains.

When I am dealing with small sample volumes (e.g. ~40mL) I can do the closed VOA method adding all reagents and spikes through the membrane and I get very good QC. I see that for drinking water, this is working well for some labs when using this micro extraction method (3511) with 25uL injections. I suppose I would need a special inlet port to allow such a large injection.

I've been reading the posts here mentioning the closed bottle methods for DRO and PAH extractions with great interest. Because that might take care of my problem if I can develop a bottle method but without the expensive extraction instrument.

[Peter Apps]

I'll bet cash money the hexane is evaporating - is there a reason why you use so little to start with ?

Peter

[Steve Reimer]

Hexane solubility in salt water is minimal. But that small of volume of hexane is going to be tough to recover from 1 liter of water unless the water is clean. Any particulate will foul things up.

Is the problem with DCM that you don't want it in the lab? That is our solvent of choice for TPH-D extractions, although now it's just to elute disks.

[LALman]

Hexane solubility in salt water is minimal. But that small of volume of hexane is going to be tough to recover from 1 liter of water unless the water is clean. Any particulate will foul things up.

Is the problem with DCM that you don't want it in the lab? That is our solvent of choice for TPH-D extractions, although now it's just to elute disks.

Yes, I don't want to start having a DCM peak in all my 8260 samples. Otherwise I would use it.

I'll bet cash money the hexane is evaporating - is there a reason why you use so little to start with ?

Peter

I get decent recoveries from ground water samples its just these gas field brines. I have to put in 8mL to recover a couple mL. Maybe I am seeing this wrong and the brine is making evaporation worse.

The previous owner of the lab never used blow down to constant volume although I used to do that in a previous lab (circa 1990) for PCB's and pesticides after using DCL to extract. I've been trying to avoid having to buy a blow down setup. I used to add a 1mL of mineral oil that had been greatly diluted in isooctane. Then if the blow down went too far, the residue acted as a keeper and I recovered my spikes and unknown. I suppose DRO is too volatile to do that.

[James_Ball]

Mineral oil would probably interfere with the DRO pattern especially the heavy hydrocarbon end of things.

If you don't have many samples I have done blow down by using 40ml voa vials to hold the solvent and using a needle attached to a line from a regulator or compressed air line to blow into the top of the vial. Hexane and DCM with both evaporate quickly doing that. You could also sit it in a beaker on a hotplate set at about 40C. Honestly the best hotplate to use is a cheap electric griddle from Wal Mart. We used those after we had a hood fire from arcing inside a Fisher hotplate when evaporating Hexane for oil and grease analysis. The electric griddles don't have exposed contact switches which can arc and are much less expensive than the explosion proof hotplates from the scientific suppliers.

Most of the extraction methods stipulate using 60ml of solvent in triplicate when doing 1L sepfunnel extractions, so you may not be getting the recovery you need from very dirty samples. The 3511 extraction would also be another option to not need to do blow down and to save on solvent usage. A 25ul injection can be done in a normal split/splitless injection port using CSR-LVI. There are plenty of papers available on Restek where one of their chemist has been working on large volume injections using normal injection ports for quite a while now, here is one link https://blog.restek.com/?p=9670 the key is using a large enough guard column with the proper polarity deactivation layer and doing solvent focusing with the oven program. I have tried it and it works really well.

[LALman]

Hexane solubility in salt water is minimal. But that small of volume of hexane is going to be tough to recover from 1 liter of water unless the water is clean. Any particulate will foul things up.

Is the problem with DCM that you don't want it in the lab? That is our solvent of choice for TPH-D extractions, although now it's just to elute disks.

I have a 4 port setup for doing O&G using Environmental Express G5407mm UltraFlow disks. What cat# or type of disk do you use for DRO?

Mineral oil would probably interfere with the DRO pattern especially the heavy hydrocarbon end of things.

If you don't have many samples I have done blow down by using 40ml voa vials to hold the solvent and using a needle attached to a line from a regulator or compressed air line to blow into the top of the vial. Hexane and DCM with both evaporate quickly doing that. You could also sit it in a beaker on a hotplate set at about 40C. Honestly the best hotplate to use is a cheap electric griddle from Wal Mart. We used those after we had a hood fire from arcing inside a Fisher hotplate when evaporating Hexane for oil and grease analysis. The electric griddles don't have exposed contact switches which can arc and are much less expensive than the explosion proof hotplates from the scientific suppliers.

Most of the extraction methods stipulate using 60ml of solvent in triplicate when doing 1L sepfunnel extractions, so you may not be getting the recovery you need from very dirty samples. The 3511 extraction would also be another option to not need to do blow down and to save on solvent usage. A 25ul injection can be done in a normal split/splitless injection port using CSR-LVI. There are plenty of papers available on Restek where one of their chemist has been working on large volume injections using normal injection ports for quite a while now, here is one link https://blog.restek.com/?p=9670 the key is using a large enough guard column with the proper polarity deactivation layer and doing solvent focusing with the oven program. I have tried it and it works really well.

I well remember doing three 60mL DCM extractions for Pest PCB work back in the mid 1990's at my first envirionmental lab job. I "assumed" the 4mL method this lab was previously using was OK for DRO since it seemed to recover the spikes. But I want to get this lab method properly certified and produce the best possible data while using as little solvent as possible. So, I'm very interested in the 3511 extraction and larger injection volumes. I've been reading along on several others postings in the forum on 3511. I am using a 1uL injection presently. I would guess from typical CRDL's of 10 ppm for soil and 0.2 ppm for water that might I do well enough using 3511 with a 20uL injection. I usually get only a few water DRO every few months. Most of my DRO sample load is for soils and I do a 20g extraction in a 40mL VOA with 10mL of 10% MeOH, 1mL o-terp surrogate in acetone, 20uL of the other two surrogates in pentane, and finally 4mL of Hexane. I put them on a rotator to get sample disagragated, followed by periodic shaking and ultrasonication in a high powered bath for an hour. Then centifugation and transfer to vial for analysis.

Recently I ran two water samples for which I received only a 2 oz jar for DRO and 2 40mL VOAs for the volatiles. I had been reading on this forum what others were doing. So, I did the 3511 microextraction but using only 32mL of water per replicate and 2mL of hexane. I got acceptable surrogate 86-113% and DRO 88-90% (after blank subtraction) spike recoveries. My MDL was about 6 mg/L.

[James_Ball]

Certification can be a pain depending on how the auditor interprets the methods. Some want to make you stick precisely to what is written, others allow for deviation from the method as long as you show that you achieve equal or better results. It would probably be easier to do 3511 for reduced solvent volume than to try to justify 4ml per 1L in a sep funnel per 3510, though some instances can allow for using 3510 with proportionally lower sample volume, such as 100ml sample with 6ml solvent, but auditors will demand you do three replicates which will leave you with 18ml solvent to deal with.

The other thing to watch out for when trying for certification is the blank subtraction. Unless specifically allowed in the method, most auditors don't allow it.

[Steve Reimer]

We use a Baker Speedisk (8055-07) for DRO extractions

[LALman]

We use a Baker Speedisk (8055-07) for DRO extractions

Thank You! I am going to order some and give that method a try along side this large volume injection method.

Certification can be a pain depending on how the auditor interprets the methods. Some want to make you stick precisely to what is written, others allow for deviation from the method as long as you show that you achieve equal or better results. It would probably be easier to do 3511 for reduced solvent volume than to try to justify 4ml per 1L in a sep funnel per 3510, though some instances can allow for using 3510 with proportionally lower sample volume, such as 100ml sample with 6ml solvent, but auditors will demand you do three replicates which will leave you with 18ml solvent to deal with.

The other thing to watch out for when trying for certification is the blank subtraction. Unless specifically allowed in the method, most auditors don't allow it.

Agreed. I never liked blank subtraction anyway. If I have my MDL at 100 and read 50, I don't want to be subtracting 50 from my LCS that is reading 950.

I'm thinking of adding the 5m RESTEK Guard/retention column and getting the 50uL autosampler syringe and trying for 10 or 20 uL injections instead of my present 1uL. Not having to waste money on expensive hexane and evaporating it will be a huge help.

Additionally, I am interested is seeing if I can eventually add PAH analysis using the 3511 method for extraction.

[LALman]

What is your opinon...
I have an DB-5MS 30m x 0.53mm x 1.50um column that I will probably never use again now that I am using 0.2mm columns in my instruments.

To implement CSR-LVSI, would it be best to use the RESTEK 5m deactivated 0.53mm column in front of my 50m x 0.2mm x 0.33um HP-5 column, or use 5m of 0.53mm DB-5MS column? In other words is CSR-LVSI focusing better if it is deactivated silica versus having an active coating?

[James_Ball]

If using LVI and retention gap, the gap needs to match the solvent in polarity. Non-polar column with polar solvent would not work so well as the solvent would not form an even layer on the surface.

[LALman]

If using LVI and retention gap, the gap needs to match the solvent in polarity. Non-polar column with polar solvent would not work so well as the solvent would not form an even layer on the surface.

Ah, I think I understand. The deactivated guard column is really a better match for solvents in general than any particular column coating.

[James_Ball]

The 5 phase column would probably work with hexane solvent since they are both non polar, though a 1 phase would be more non-polar. But you would not want to try it with methanol since it is a polar solvent, and I have had to add an intermediate polar guard to a 5 phase column to do ethyl acetate solvent because the solvent was too polar for the phase.