Quick & Painless Email Subscription

Use the “Subscribe” link at right. It just takes a few seconds!

Sample loading capacity for PAHs, part 3: 30m x 0.32mm x 0.25µm Rxi-5ms

My last post, “Does a thicker film GC column increase sample loading capacity for PAHs? 30m x 0.25mm x 0.50µm Rxi-5ms”, generated some interesting internal discussions between me, Jaap de Zeeuw, Chris English, and Chris Rattray, as to what constitutes overload of a GC column.  Jaap really focused on peak shape and contested my assessment that there was not a “substantial” difference between the 0.25mm x 0.25µm and 0.25mm x 0.50µm columns as regards overload, particularly for the 50 and 25 ng each PAH chromatograms.  So in addition to presenting PAH overload results for a 30m x 0.32mm x 0.25µm Rxi-5ms, which is also supposed to have increased sample loading capacity because of its wider bore, I zoomed in on the benzofluoranthene peaks for 50 and 25 ng amounts on column at Jaap’s suggestion for better inspection of the peak shapes.  So check out the first two sets of chromatograms below, which include 3 chromatograms each for 30m x 0.25mm x 0.25µm Rxi-5ms, 30m x 0.25mm x 0.50µm Rxi-5ms, and 30m x 0.32mm x 0.25µm Rxi-5ms columns.  The peak shapes do not look that much different to me, except for what appears to be a slight proportional narrowing of peaks for the 0.32mm x 0.25µm column with 25 ng PAHs, but perhaps readers can offer their opinions.

The next four sets of chromatograms demonstrate to me that there is still no dramatic increase in sample loading capacity for these PAHS with either a thicker film (0.25mm x 0.50µm) or a wider bore (0.32mm x 0.25µm) column. Essentially, these larger ring PAHs are just not “soluble” in the 5% diphenyl stationary phase, leading to overload, except at lower levels.  So up to this level of data collection, I still recommend the 30m x 0.25mm x 0.25µm Rxi-5ms GC column for semivolatiles/EPA PAH work because of faster run times, better separations, and narrower peak widths (which leads to better detectability).

As previously, the 30m x 0.32mm x 0.25µm method was translated from a previous method using the EZGC Method Translator.

Overload 3 Fig 1

Overload 3 Fig 2

Overload 3 Fig 4

Overload 3 Fig 5

Overload 3 Fig 6

Think twice before purchasing a 3/16” OD packed column

So you may be asking yourself, why the cautionary statement? Well, it’s very simple, many of our customers who purchase 3/16” OD (outside diameter) packed columns have a difficult time installing them into their GC oven.  Why?  Sometimes finding the appropriate fitting/ferrule is difficult.

Let’s say you have a packed column instrument. If your injection port was designed for 1/4” OD columns, then installation is usually not an issue.  Most customers can use a 1/4” x 3/16” OD reducing ferrule, like Restek 20258. To center the column properly (which is recommended if you plan to inject your sample into the packed column) and if you have an Agilent GC with a packed column injection port, you can use our kit 21650, which resembles what is shown in the photo below.  The silver thing that looks like a GC injection port liner is called a “centering sleeve”, and as you may guess, its purpose is to center the packed column in the injection port.

21650

 

Now let’s look at a few scenarios where installing a 3/16” OD packed column is not so simple. Many modern GC’s have injection ports designed for 1/8” packed columns.  So how does one install the larger 3/16” OD packed column into this smaller injection port?  You will need a short piece of 1/8” OD tubing and a 3/16” to 1/8” reducing compression (straight) union like Swagelok SS-300-6-2. Unfortunately, we do not sell this fitting.  However, we do sell a 1/4” x 1/8” reducing union, Restek 23170.  Simply replace the 1/4” metal ferrule set with ferrule 20258.

What if you have a GC designed for only capillary columns? Normally we would recommend a “pigtail” set-up, like described in this FAQ.  How do I install a packed column into a capillary GC injection port?  However, you would need a 3/16” x 1/16” reducing (straight) union, like Swagelok SS-300-6-1. Unfortunately, we do not sell this fitting either.  However, we do sell a 1/4” x 1/16” reducing union, Restek 23169.  Simply replace the 1/4” metal ferrule set with ferrule 20258.

In summary, before purchasing a 3/16” OD packed column, make sure you have the proper fittings/ferrules available for installation. I briefly mentioned this in my post titled Things to Consider Before Ordering a Packed Column.  To save yourself frustration, do a little homework first to determine if a 3/16” OD column is really your best choice, and if not, consider an alternative OD based upon your particular instrument.

 

Does a thicker film GC column increase sample loading capacity for PAHs? 30m x 0.25mm x 0.50µm Rxi-5ms

After my post, “Sample loading capacity for PAHs on a 30m x 0.25mm x 0.25µm Rxi-5ms GC column”, a ChromaBLOGraphy reader posed the question about how a 0.50µm column would work.  It’s a great question, because column vendors routinely advertise that thicker films have more sample loading capacity.  But do they really, and if so, how much more?  Let’s test it!

First I used the EZGC Method Translator to get the GC oven program for the 30m x 0.25mm x 0.50µm Rxi-5ms.  Ultimately, the translation results in a slower oven program rate to try and keep the elution temperatures about the same for the compounds of interest so that no compounds suddenly coelute, or worse, switch elution order.  As you can see from the chromatograms below, the 0.25mm x 0.50µm Rxi-5ms does not show a substantial sample loading capacity increase versus the 0.25mm x 0.25µm for the PAHs of interest.  Given that the run times are longer for the thicker film column, and the peak widths are broader (leading to reduced sensitivity), there seems to be no dramatic advantage to using the thicker film column in this case.

25_50 PAH Xlate

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

25_50_PAH 50

 

25_50_PAH 25

 

25_50_PAH 12_5

 

25_50_PAH 6_25

 

25_50_PAH 3_13

Pesticides/Fungicide cocktails responsible for Death of Bees: analysis via GC-MS

 

2011-jaap-pasfoto4-smallResearchers at the University of Maryland and the US Department of agriculture may have uncovered the mystery of high mortality in honeybees. These insects are responsible for pollinating over 130 different crops in the US alone. Loss of this pollinator species could cost the US $15 Billion annually. The EU (European Union) issued a temporary ban against the neonicotinoid pesticides. Honeybee deaths may be caused by  a ‘witches brew’ of pesticides and antifungals.

 

The scientists are narrowing down which pesticides and antifungals (fungicides) impairs the bees by investigating pollen of different crops (almonds, apple, cranberry, watermelon, pumpkin, cucumber and blueberries). Since a different cocktail of pesticides is used on varying crops researchers have the ability to determine which pesticides have the greatest effect on bee mortality.

On average crops were sprayed with nine different pesticides and one crop tested had 21 different pesticides present. The combination of these chemicals reduces the bees’ resistance to the parasite Nosema ceranae. Bees with compromised resistance die. Some of the most notable pesticides were chlorothalonil, pyraclostrobin (fungicides) and fluvalinate, amitraz (insecticides). This research was published in the scientific journal PLOS One.

Chlorothalonil is analyzed by GC and can be performed on a non-polar Rxi-5ms type column or a stabilized Rxi-5Sil MS. For a list of applications see:
http://www.restek.com/chromatogram/search?s=type:GC::chlorothalonil

Insecticide Fluvalinate and pyraclostrobin fungicides, are a bit more challenging:
http://www.restek.com/chromatogram/search?s=type:GC::chlorothalonil::fluvalinate

We have done this one only using comprehensive (GCxGC-TOF). Here Chlorothanil is also determined. Fig 1 shows a typical chromatogram. In the second dimension the Rtx-200 was chosen showing very good separation of peaks of interest over the whole separation space.

fig. 1 pesticides GCxGC

Fig 1: separation of pesticides using GCxGC-TOF, 30m x 0.25mm Rxi-5SilMS in first dimension and Rtx-200 in second dimension

Benzo[b]fluoranthene as part of the EFSA PAH4

I’ve already had one reader of my last post, Is separation of benzo[b]fluoranthene and benzo[k]fluoranthene on 5% phenyl-type GC columns really that important for environmental analyses? ask why I had benzo[b]fluoranthene in red text on the chromatograms.  I had planned to include something about it being on the EFSA PAH4 list in that post, but I’m an old man, so I forgot.  Here goes, in more detail…

The European Food Safety Authority (EFSA) in their Polycyclic Aromatic Hydrocarbons in Food – Scientific Opinion of the Panel on Contaminants in the Food Chain (Adopted on 9 June 2008), issued the following statement:

“The CONTAM Panel concluded that benzo[a]pyrene is not a suitable indicator for the occurrence of PAHs in food. Based on the currently available data relating to occurrence and toxicity, the CONTAM Panel concluded that PAH4 and PAH8 are the most suitable indicators of PAHs in food, with PAH8 not providing much added value compared to PAH4.”

The PAH4 are:

Benz[a]anthracene, Chrysene, Benzo[b]fluoranthene, Benzo[a]pyrene

The PAH8 are:

Benz[a]anthracene, Chrysene, Benzo[b]fluoranthene, Benzo[k]fluoranthene, Benzo[a]pyrene, Indeno[1,2,3-cd]pyrene, Dibenz[a,h]anthracene, Benzo[ghi]perylene

As you can see, Benzo[b]fluoranthene is a member of the PAH4, possibly the most significant group in estimating carcinogenic potential of food due to PAH content. Importantly, this doesn’t mean that benzo[j]fluoranthene and benzo[k]fluoranthenes are NOT present in food; they are, anytime benzo[b]fluoranthene is present.

The point I was going to make is, that in a pinch, an Rxi-5ms might be a better choice for screening of the EFSA PAH4 in food, because it would only have the absolute coelution of triphenylene and chrysene (at m/z 228). But a better choice is an Rxi-PAH GC column that will separate:

Triphenylene and Chrysene (m/z 228)

Benzo[b]fluoranthene, Benzo[k]fluoranthene, Benzo[j]fluoranthene, and Benzo[a]fluoranthene (m/z 252)

Benzo[e]pyrene and Benzo[a]pyrene (m/z 252)

Dibenz[a,j]anthracene, Dibenz[a,c]anthracene, and Dibenz[a,h]anthracene (m/z 278)

Is the Rxi-PAH only suitable for food PAH analyses? Absolutely not.  It will work very well for environmental PAH analysis where more comprehensive profiling is desired (e.g., environmental forensics, such as oil spill and other fingerprinting work), including as a primary column in GCxGC-TOFMS.  But keep in mind since it was designed with a thin film to more easily elute involatile PAHs like coronene and dibenzopyrenes, you need to inject less material on it or trim the front of the column more frequently for maintenance purposes.

Now where did I put those car keys?

Is separation of benzo[b]fluoranthene and benzo[k]fluoranthene on 5% phenyl-type GC columns really that important for environmental analyses?

Some gas chromatographers might say so because of the following statement in US EPA Method 8270D (Semivolatile Organic Compounds by Gas Chromatography – Mass Spectrometry (GC-MS):

“11.6.1.4 Structural isomers that produce very similar mass spectra should be identified as individual isomers if they have sufficiently different GC retention times.  Sufficient GC resolution is achieved if the height of the valley between two isomer peaks is less than 50% of the average of the two peak heights.  Otherwise, structural isomers are identified as isomeric pairs.  The resolution should be verified on the mid-point concentration of the initial calibration as well as the laboratory designated continuing calibration verification level if closely eluting isomers are to be reported (e.g., benzo[b]fluoranthene and benzo[k]fluoranthene).”

The language above ignores the fact that benzo[j]fluoranthene is a polycyclic aromatic hydrocarbon (PAH) isomer consistently found at significant levels in environmental (and food) samples, and more importantly, that it coelutes with either, or both, benzo[b]fluoranthene and benzo[k]fluoranthene on 5% phenyl-type GC columns. In this blog post, I will define the extent of the coelution on 30m x 0.25mm x 0.25µm Rxi-5Sil MS (5% phenyl as silphenylene, similar to DB-5MS) and 30m x 0.25mm x 0.25µm Rxi-5ms (5% phenyl as diphenyl, similar to DB-5).

PAHs were first analyzed using hydrogen efficiency-optimized flow (EOF), and optimal heating rate (OHR), with an Agilent 6890 GC-FID. Injections of PAH standards (prepared from SV Calibration Mix #5 / 610 PAH Mix and a custom benzo[j]fluoranthene standard) at a split ratio of 100:1 into a 4mm Precision split liner with wool served to minimize injection band widths to keep overall system efficiency high.  In addition to OHR, chromatograms for each column were generated at 1.5 x OHR and 0.5 x OHR to force different elution temperatures for the benzofluoranthenes to see what effect that had on their separation.

As you can see in Figure 1, the Rxi-5Sil MS shows an intractable coelution for benzo[b]fluoranthene and benzo[j]fluoranthene under all experimental conditions, with elution temperatures around 278, 293, and 257°C for OHR, 1.5 x OHR, and 0.5 x OHR, respectively.  The Rxi-5ms has a different coelution, that of benzo[j]fluoranthene and benzo[k]fluoranthene for OHR and 1.5 x OHR, but interestingly is able to partially resolve the benzofluoranthenes under 0.5 x OHR conditions (Figure 2).  Elution temperatures in this case were approximately 280, 295, and 258°C (OHR, 1.5 x OHR, 0.5 x OHR).

What does this coelution situation mean for environmental analysts? First, all quantitative values reported for benzo[b]fluoranthene and benzo[k]fluoranthene are likely inaccurate for at least one of the isomers depending on the GC column choice if using a “5-type” stationary phase.  Second, technically speaking, analysts should not be reporting the benzofluoranthenes as “individual isomers” according to the language of EPA Method 8270D when using a “5-type” GC column.

If an analyst needs to report benzofluoranthenes as “individual isomers”, a 30m x 0.25mm x 0.25µm Rxi-17Sil MS (or even a 15m version) will work well.  If the challenge is to separate the benzofluoranthenes and, triphenylene and chrysene (228 m/z coelution on 5-type columns), consider the Rxi-PAH GC column.

BF Fig 1

 

BF Fig 2

Sample loading capacity for PAHs on a 30m x 0.25mm x 0.25µm Rxi-5ms GC column

Sample loading capacity for a GC column (also known as “column capacity” and “sample capacity”) is essentially the amount of non-active compound that can be put on a GC column and chromatographed at some set of conditions where the peak shape is symmetrical. Conversely, if a GC column is overloaded with a component amount, the peak will exhibit the classic “shark fin” shape.  Most vendors estimate the sample loading capacity for a 0.25mm x 0.25µm GC column to be around 50-100 ng per analyte.  But is this even close to accurate?  Perhaps, but it depends on the compound of interest.

Environmental analysts already know that polycyclic aromatic hydrocarbons (PAHs) are prone to overload on GC columns, especially the typical “five” type columns used for semivolatile organic compound analysis in EPA methods, e.g., 8270 and CLP. Although the CLP semivolatiles method states that a GC column should be able to “accept up to 160 ng of each compound listed in Exhibit C (Semivolatiles), without becoming overloaded”, I could find no quantitative data indicating when overload occurs for Exhibit C PAHs on the commonly used 0.25mm x 0.25µm GC columns.  That’s my lead-in to say, “I’ll test it myself!”  With the follow-up… “And post about it on ChromaBLOGraphy”.

All work was done with a 30m x 0.25mm x 0.25µm Rxi-5ms utilizing hydrogen efficiency-optimized flow, and optimal heating rate, with an Agilent 6890 GC-FID.  Injections of PAH standards (prepared from SV Calibration Mix #5 / 610 PAH Mix) at a split ratio of 10:1 into a 4mm Precision split liner with wool served to minimize injection band widths to keep any peak deformations associated with the column, and not the inlet.  As you can see in Figure 1 by the “shark fin” shape of the peaks and the unacceptable resolution between benzo[b]fluoranthene and benzo[k]fluoranthene, and indeno[123-cd]pyrene and dibenzo[ah]anthracene, 200 ng of each of these PAHs greatly exceeds the sample loading capacity of this column. Figure 2, at 50 ng each compound, also shows gross overloading of PAHs, and although the separation starts to improve between critical pairs, the retention times are still shifted later based on peak deformation.  Finally, at 12.5 ng (Figure 3), the chromatogram looks good in separations and retention times, comparing relatively nicely to Figure 4 at 3.13 ng, where we would not expect overload.

Even at 25 ng of later eluting PAHs (Figure 5), we have overload, which puts the maximum sample loading capacity of the 30m x 0.25mm x 0.25µm Rxi-5ms GC column under these operational conditions between 6.25 and 25 ng each PAH, an order-of-magnitude less than what is suggested to be possible by the CLP method.  Can we tolerate overload and still get our work done?  That depends on your data quality objectives, which might include:

Separation between critical, often isobaric, pairs.

Overall required peak capacity for the chromatogram (the 200 ng overload cut the peak capacity by almost 3).

Capacity of the detector to handle overload (MS systems are very sensitive these days).

I prefer the peaks look more like those seen in Figure 7, and given the sensitivity of today’s MS systems, we can inject less (maybe via “Shoot-and-Dilute”), see almost as much, and probably keep our systems up longer due to less involatile “dirt” put on the GC column.

Figure 1 OL

 

Figure 2 OL x

 

 

Figure 3 OL

 

Figure 4 OL

 

Figure 5 OL

 

Figure 6 OL

 

Figure 7 OL

 

 

 

What are these O-rings for that I received with my baseplate trap?

Some of you have noticed that included with several of our baseplate traps are a package of two small O-rings. Technical Service has been informed that these are the O-rings which fit onto the check valves on your baseplate (see red arrows below).  Does this mean that you need to replace these O-rings?  Not necessarily.  Carefully inspect the O-rings on the check valves, and if you do not see any signs of wear (cracks, deformation, etc…), you probably don’t need to replace them.  However, if they do exhibit signs of wear, go ahead and replace them.

.

 Baseplate 4

.

If you also need the larger O-rings that are located on the baseplate, at the base of the check valves (as identified by the green arrows), you can obtain those by purchasing Restek catalog number 22023.

22023

Should I use a 2.7 or 5 µm Raptor™ column?

crossroads-man

“When you come to a fork in the road, take it.” Maybe that advice worked for Yogi Berra, but does not help much in decisions about your lab equipment.

If you are not certain whether you should use a Raptor™ column, please click on the title below to read my previous blog post before proceeding.

What is SPP and when should I use a Raptor™ column?

Once you have made the decision to take advantage of the benefits of SPP, your next question might be which particle size to use, the 2.7 or 5 µm Raptor™.   The 2.7 µm column provides higher efficiencies, but the column must be matched with an instrument that is appropriate, in terms of system pressure and internal dwell volume. These considerations are described below:

Primary Consideration- System Pressure

As is the case with a 3 µm fully porous particle, the 2.7 µm Raptor™ column will operate at backpressures a little higher than the corresponding column with a 5 µm particle size.  Pressure is a function of particle size, as well as column dimensions, flow rate and solvent properties, but is not a function of porosity. The following is a good example of how you can expect flow rate and particle size to affect the column back pressure. Please keep in mind that methanol and water both have higher surface tension, which results in higher backpressure. The data shown here was generating using acetonitrile/ water mobile phase.

Use a 2 Figure 2

Although the earlier blog post Building Up Pressure on HPLC? was written for fully porous particles, you can still get a pretty good idea of how pressures are affected by column dimensions, particle size, and solvent content by looking at the table shown in the post. The optimal flow rates will be a little different for SPP columns. Though the “optimal” flow rate for a SPP column is not much different versus a fully porous column, the SPP columns maintain their efficiency over a broader range of flow rates. Here is a good illustration, using a Van Deemter plot.

 usea2 figure 3

 

Secondary consideration-Dwell Volume

There is another factor to consider when transferring your method to a 2. 7 µm Raptor™, and that is the system dwell volume. In other words, dead volume in your LC system may negate the advantages of choosing a 2.7 µm SPP column over a 5 µm particle (either fully porous or SPP). This is of highest concern when using columns of smaller dimensions also. A symptom of excess dwell volume is be band broadening, and can be confirmed by running the same column on a system with less dwell volume. Although there are some system updates that can be made to reduce dead volume, the options are limited and it is simpler to start with a well-matched system and column from the beginning. Fortunately, LC systems with maximum pressures of about 600 bar (8700 psi) are usually also designed with a lower internal dwell volume relative to the lower pressure systems. This actually makes our choices a little easier. As a result of both considerations, we recommend using systems with a 600 bar maximum operating pressure for the 2.7 µm Raptor™ columns.

Here are examples of some LC Systems with a maximum pressure of at least 600 bar (8700 psi). Again, we recommend using a system with a max pressure of 600 bar, or higher, for use with a 2.7 µm Raptor™ column.

  • Agilent 1200 Series- RRLC models only
  • Agilent 1220 & 1260 Infinity
  • Shimadzu Prominence UFLC-XR models only
  • Thermo Ultimate 3000 -Basic Manual, Basic Automated, Binary Analytical, x2 Dual analytical, and Quaternary Analytical models
  • Thermo Ultimate 3000 -RSLC nano, RSLC Binary, and RSLC X2 Dual (~700 Bar limit)
  • Perkin Elmer Flexar -FX-10 models only
  • ABSciex/Eksigent – microLC 200, nanoLC-Ultra and nanoLC 400 (~700 Bar limit)

You can also use a 2.7 µm Raptor™ column with a UHPLC system, which typically has a maximum operating pressure limit of 1200 bar/1700psi.

 

When to use the 5 µm Raptor™ 

Your next question might be when or why you should use a 5 µm Raptor™column. The 5 µm particle size for this SPP column allows you to obtain results that resemble that of a 3 µm particle fully porous particle column on an HPLC system that has a maximum pressure of 400 bar/ 5800 psi or less (which is usually too low to use a column with a 2.7 or 3 µm particles). There are also other advantages described in our article The Effects of LC Particle Choice on Column Performance 2.7 vs. 5 µm Diameter Superficially Porous particles (SPP), which I highly recommend reading, by the way. Here are just a few examples of some LC systems that could be optimized by using them with a 5 µm Raptor™column:

  • Agilent 1100 & 1200 Series (except models listed above)
  • Varian 920-LC & 940-LC
  • Waters Alliance and Breeze models
  • Jasco 2000 Series
  • Perkin Elmer Series 200 & 275

The above list is definitely not all inclusive and there are many I have not shown. Generally, the older HPLC systems are more likely to have a lower pressure rating and many of them have 400 bar limits. The best way to find out is to consult your instrument manual or contact the manufacturer of the instrument.

I hope this has been helpful. Many thanks to our Innovations Chemists, Ty Kahler and Sharon Lupo for the data provided. Thank you for reading.

Inject more fish extracts before GC maintenance is needed using shoot-and-dilute GC

Persistent organic pollutants (POPs) like dioxins, furans, polychlorinated biphenyls (PCBs), and polybrominated diphenyl ether (PBDEs) are lipophilic and therefore may be found in fatty matrices like fish.  In order to remove the high lipid coextractives, lengthy cleanup procedures that require multiple cleanup cartridges are often used.  I recently presented work at the Dioxin meeting where we took a shortcut and analyzed PBDEs and other halogenated flame retardants (HFRs) in fish using a quick QuEChERS type extraction, and a PSA pass through cleanup (Get the poster here!).  While the PSA pass through does a good job at removing a large amount of fat (we reduced the fat by 50-70%), there is still a considerable amount of nonvolatile residue remaining in the final extract.  This is especially evident in the mackerel sample that we analyzed where 38 mg/mL of nonvolatile residue remains after the PSA cleanup.

In order to combat the negative effects of injecting so much nonvolatile material onto the column, such as decreased response of high MW compounds and increased inlet activity, we used a 10:1 split injection.  A split injection (aka shoot-and-dilute GC) takes advantage of the sensitive detectors we have (ECD, MS/MS, HRMS) and keeps the system up at least twice as long when compared to a splitless injection.  The flow through the liner is much faster when compared to a splitless injection and the majority of the sample is swept into the split vent and not onto the analytical column.  Just remember to clean that split vent once in a while (it can get pretty gross)!

The chromatograms below highlight the advantage of using a split injection for fatty samples like fish.  In all of the experiments I used a Sky Precision liner with wool and a 15m x 0.25mm x 0.10µm Rtx-1614 column.  The sequence was the following: split injection of HFR standard, splitless injection of HFR standard, split injection of mackerel sample, split injection of HFR standard, splitless injection of HFR standard, 2 x split injection of mackerel sample, followed by the standards again.  I repeated this sequence with increasing numbers of the mackerel sample injections in between standard injections until the system failed (I could no longer quantify my results).  As you can see the response for the last two eluting peaks, syn and anti-Dechlorane Plus, are dramatically decreased after only three mackerel injections.  The split injection standard did not fail until 15 sample injections.

The increased ruggedness of using shoot-and-dilute GC was also highlighted by Jack Cochran in this blog post where he made repeated injections of used motor oil.  So if you have dirty samples and sensitivity to spare, try the shoot-and-dilute GC approach and think about all those extra things you could be doing instead of GC maintenance!

Halogenated flame retardant standard injected before and after multiple mackerel sample injections.  Response of late eluting flame retardants (syn and anti dechlorane plus) is reduced after only 3 sample injections.

Halogenated flame retardant standard injected before and after multiple mackerel sample injections. Response of late eluting flame retardants (syn and anti dechlorane plus) are reduced after only 3 sample injections.

 

HFR standard injected before and after multiple mackerel sample injections.  A split injection allows more samples to be analyzed before system maintenance (liner change, column trimming) needs to be performed.

HFR standard injected before and after multiple mackerel sample injections. A split injection allows more samples to be analyzed before system maintenance (liner change, column trimming) needs to be performed.