TO-15 + PAMS + TO-11A = China’s HJ759 + PAMS + HJ683

It is very fitting that I write this blog while I am in Shanghai, China. The impetus for this blog, and the blogs to follow, is that the Chinese Ministry of Ecology and Environment (formerly the Ministry of Environmental Protection) published their “VOCs Monitoring Scheme of Environmental Air Quality.” In short, this standard outlines the sampling and analysis of 117 VOCs via the following three methods: HJ759, PAMS, and HJ683. These methods are very similar to U.S. EPA Methods TO-15, PAMS, and TO-11A, respectively. They have done so, because PM2.5 is not solely responsible for hazy skies and VOCs play a critical role in atmospheric reactions, which generate ozone, smog, etc. The following picture illustrates why the People’s Republic of China (PRC) is interested in VOCs:

In the photo, those are large shipping boats on the Yangtze river. I took this photo on final approach at 4 pm, and no, it was not rainy or overcast. This was smog. And it gets better… I could actually see the sun today, which I did not see during my 13 day visit in November of 2018. So, this was a good day! Okay, you get the point.

Before we break down HJ759 and PAMS, let us quickly review up to present day: I have spent the last several years presenting at conferences, publishing blogs, and writing application notes on TO-15. However, there is also the Photochemical Assessment Monitoring Stations (PAMS) network, which utilizes a lesser-known and -discussed canister-based method; since this method is executed by government monitoring sites and laboratories. We have countless blogs, chromatograms, etc. for TO-15, but we have never blogged about PAMS and confess that we only have one PAMS chromatogram. Despite our lack of focus, PAMS is interesting. Long story short, the PAMS network was put in place for the monitoring of ozone. Therefore, this network focuses in on aliphatic, aromatic, and carbonyl VOCs. Whereas, TO-15 has some overlap, but places more of an emphasis on halogenated VOCs. Overall, the sampling and analysis is very similar between the two methods. The divergence is in the target analyte lists. For more on PAMS, be sure to check out the EPA’s website.

Now for present day: The PRC has used TO-15 as the model for HJ759. Most of the method is very similar, which makes sense. And PAMS is simply PAMS. Why reinvent the wheel, when the U.S. already tackled the same problems over 3 decades ago. However, there appears to be a new trend coming out of the PRC, which is a push for one sampling and analytical approach for all the VOCs found in To-15 (HJ759 in China), PAMS (PAMS in China), and TO-11A (HJ683 in China). I have the following mixed feelings about this movement:

  1. I love the fact that manufacturers/providers like Restek are being pushed to develop innovative approaches for meeting the PRC’s desire for one analytical method.
  2. There is a reason TO-11A requires the sampling of carbonyls by derivatization on cartridges and not by collection into canisters in the U.S. We already know this does not work well in canisters. However, despite the PRC document specifying the use of cartridges for carbonyls, some of the air sampling community in the PRC seems to think otherwise. For the record, today’s blog will only focus on combing TO-15 and PAMS, and not TO-11A. More blogs to follow on this specific topic later.

The concept of combining the TO-15 and PAMS target analyte lists into one analytical run is not new. In fact, I have been presenting a lot of my canister cleaning work at the National Environmental Monitoring Conference (NEMC), Air and Waste Management Association (A&WMA), and to the U.S. EPA since 2014; for all of which I used the following single analysis combining TO-15 and PAMS target analyte lists:

As you may see in the above chromatogram, there were 116 compounds (i.e., 111 target analytes, 3 internal standards, and BFB). There were only 2 critical coelutions (2-Methylpentane and vinyl acetate; and n-hexane and ethyl acetate), which I never bothered to work out, because this method was only used internally. Not sure why I never published this method until now.  Wait, I take that back! Here are the following reasons:

  1. I believe most laboratories do not have interest in the above, as they are conducting either TO-15 or PAMS, not both; and/or not in one analysis. This is evident since we have never received a request.
  2. This method requires on-column cooling, which labs tend to avoid.

That is of course until now! The PRC is pushing the envelope and asking for just that (i.e., the single analytical method). It would be nice to know that we just have to work out those 2 critical coelutions and have a solution. But it is not quite that simple. We need to develop further because:

  1. Preconcentrator instrument manufacturers have moved away from using liquid nitrogen to cool their traps. These makes on-column cooling that much less attractive.
  2. The C2 compounds in PAMS (ethylene, acetylene, ethane) do not respond well on a mass spec, so the desired detection limits (200 pptv) may be hard to achieve.

Okay, I have reached my personal blog size limit and jet lag remains present. So, stay tuned for the next blog where we show you a dual column GC-FID-MS analysis with all 116 TO-15 (HJ759) and PAMS compounds resolved. Oh yeah, if you happen to be in Shanghai this week, be sure to stop by our booth at the IEexpo.

Choosing Your Citral Column

Image credit: Wikipedia

The name ‘citral’ or 3,7-dimethyl-2,6-octadienal, suggests the scent of lemons so it’s not surprising that another name for this compound is lemonal. There are two isomers of this compound with the same chemical formula (C10H16O); geranial (citral A) and neral (citral B). As seen in the image to the right, the difference between these compounds is subtle1.

Geranial and neral both have a lemon scent, however, neral has a milder, and sweeter lemon odor. These compounds are used individually or together depending upon the desired scent or flavor since they are used in perfumes, candy and even soft drinks. They can also be added to enhance other flavors for artificial grapefruit, orange and lime. The most common uses are cleaning products, laundry and dishwashing detergents. Chances are good you will use citral today. Interestingly lemongrass has significantly higher amounts of citral compared to lemons1.

As you will see from the chromatograms below column choice is important. Two different columns were evaluated; the Stabilwax and the Rtx-Wax, for peak shape, resolution and bleed. The obvious column of choice is the Rtx-Wax as shown in the following chromatograms.

Shown below are chromatograms of a custom citral mix on both columns with the Rtx-Wax column showing excellent peak shapes and separation. The Stabilwax column shows a fronting effect leading up to the citrals, which can cause problems with integration/quantitation.


Figure 1: Image of Custom Citral Mix on an Rtx-Wax


Figure 2: Image of Custom Citral Mix on a Stabilwax, notice the fronting peak shape indicating interaction with the analyte and stationary phase.


Figure 3: Zoomed image of the Custom Citral Mix on the Rtx-Wax and the Stabilwax. The red trace represents Rtx-Wax whereas the blue trace represents the Stabilwax.


Figure 4: Chromatogram of lemon Oil using the Rtx-Wax column. Lemon oil contains percent levels of citrals.



Paul’s excellent questions on “Liner Selection for HS VOCs”

I had a feeling the blog I posted yesterday was sure to prompt some thought-provoking questions, as some of my peers had already been doing so. So, it came as no surprise that Paul posted the following excellent questions (in black) to which I have responded to in blue. Normally, I would just address all this in the comments section, but in my opinion the comments section tends to get lost in the weeds. In addition, the questions and my responses were all so long it justified the following blog, so I will respond to the following questions from Paul below.

Paul says:

First thanks for this blog series. I like fundamentals. I read the linked article by Jason S. Herrington and got a few questions about it.

Thank you for the thoughtful questions.

1. Why is there a difference in peak area between the 0,75 mm and 2 mm liner? The Peak width should be wider for the 2 mm liner but the Peak area should be the same. (Degradation or lost in the split?)

My alternative theory is that the 0.75 mm liner fits the fiber like a glove. So, thermal transfer to the fiber needle, fused silica, and most importantly phase is more efficient. Add in the increased velocities of the 0.75 mm liner and now we have further pushed the partitioning equilibrium in favor of desorption. This all translates into faster and more complete partitioning of VOCs out of the fiber. Honestly, although never explicitly stated, I thought this was all the “logic” behind other vendors pushing the 0.75 mm liners for SPME. But then again, I say “logic” because I certainly do not see the data to support any of the above, so it appears to be based on what I shall call logical theory. Of course, maybe it was all just in my head in the first place. Bottom line, the lack of data surrounding the subject (i.e., SPME liner dimensions) was the impetus for me to collect this data in the first place.

2. With a split of 1:5 the velocity in the inlet is much higher than with splitless injection. I can imagine with the higher velocity in the liner because of the split, the liner diameter is not that important any more.

My one colleague has been hounding me about this very point. Obviously, you are both correct about the split minimizing the significance of liner dimensions. Confession: in my previous blog, I made a serious mistake by not justifying the 5:1 split due to the chromatography looking so horrible without the split. TEASER: I ran the same type of experiments in splitless, and initial review of the data continues to say that liner dimensions do not make a Tinkers Dam for HS-SPME VOCs. Future blog to follow…

3. Why is the % RSD of the 2 mm liner seams to be better than for the 0,75 mm liner?

Ah, you saw this too!? I am not entirely sure this is a consistent trend. Perhaps we will know further in the future with a more substantial data set. With that said, I will pose the following theory to explain this observation: as stated above, the 0.75 mm liner fits around the SPME fiber like a glove. Well this means that the SPME fiber has occupied a lot of the available real estate in the liner. This could translate into one or both of the following phenomenon taking place, which may explain the observation in question:

  1. Flow has been constricted/obstructed in the 0.75 mm fiber and this impacts the split, thereby causing some inconsistencies. For the record, this may be a contributing theory to help explain the discrepancy we addressed in the first question.
  2. The SPME fiber “clogging” up the 0.75 mm liner results in turbulent flow, which also causes some inconsistencies.

Everything here is theoretical and will be hard to isolate. Is it turbulent flow, higher velocities, better thermal transfer, inaccurate splits, residence time, etc… the list goes on ad infinitum. I say it is probably the culmination of all of the above, but doubt we will truly ever know. However, it is certainly fun to hypothesize and debate some of the theories.

4. Which role does the distance of the column to the liner Play in this case? (keyword tapered liner)

I have only been running straight-walled liners and have not evaluated column distance.

I hope the community got some ideas.

I hope so too, as the data surrounding SPME liners appears to be hidden. Thanks again for your questions.

SPME Fundamentals: Liner Selection for HS VOCs

My colleague Linx Waclaski is usually the one doling out excellent liner advice, so bear with me as I take a crack at this liner stuff. Ever since we came out with traditional SPME and the SPME Arrow, a lot of customers have had concerns/questions regarding SPME liner dimensions. The short answer to most of these questions is that none of this “makes a tinker’s dam” (an old expression my dad used to say to me, so as to let me know I was focusing in on inconsequential minutia) for head space (HS) volatile organic compounds (VOCs). Look up tinker’s dam for a fun fact. If you want the long answer (with appropriate caveats), turn to page 36 of the following:

The Column (17 December 2018 Volume 14 Issue 12)

I really wrote the current blog and the above article for the new traditional SPME and/or SPME Arrow end user. In particular, to let them know they need not get wrapped around the axle when it comes to liner selection. Oh, and to challenge the unsupported claims I see some vendors make that narrow bore inlet liners are more efficient for SPME.

If you are interested in moving the needle significantly, there a far more important details surrounding SPME extraction and desorption that deserve your precious time and consideration. In fact, the SPME end-user has over a dozen extraction and desorption conditions (e.g., extraction temperature, desorption duration, etc…) they can manipulate. For example, look at Colton’s recent blog on incubation/extraction temperatures for cannabis residual solvents using SPME. In particular, look at what happens when you incubate/extract at 30 vs 80°C for o-Xylene. Yes, I know it may be hard to see with the log scale. Luckily, I happen to share an office with Colton and can tell you that we are looking at average peak area responses of 1.66 x 107 vs 5.97 x 106, respectively. That is an 89% difference in response, without having to purchase different liners. Mind you, in the LCGC article I wrote the largest statistically significant difference we saw was only 57%. I hope you see the point.

I am not saying to ignore liner selection. What, I am saying is that there is only so much time in a day and I would encourage you to invest your limited time into optimizing the extraction/desorption conditions, as illustrated in Colton’s blog. So, stay tuned for some up-coming blogs where we continue to demonstrate which SPME parameters actually make a tinker’s dam.

What dSPE works with spinach in GC-MS/MS analysis? #NationalSpinachDay

Today is a National Spinach Day! What is a better way to celebrate than to talk about spinach analysis?

During part 1 of my blog series, I discussed what dSPE was best for celery. I found that most dSPE, with the exception of dSPE containing high amounts of graphitized carbon black (GBC), showed acceptable analyte recovery and matrix removal. The same cannot be said for spinach. Spinach was more complicated although initially the data looked similar (Fig. 1). The planar pesticides have lower recoveries while the rest of the analytes are unchanged.

Figure 1: Comparison of mean responses of pesticides spiked in spinach using individual dSPE. The planar pesticides, such as Thiabendazole and Cyprodinil, have low recoveries.

Examining spinach-related pesticides is more challenging because the planar pesticides in our mix are not commonly found on spinach (Fig. 2).

Figure 2: Comparison of mean responses of selected pesticides spiked in spinach using individual dSPE

So, can we use any dSPE since the recoveries are comparable? Well, not really. Unlike celery, spinach has high levels of pigmentation. If we choose a clean-up method that doesn’t address the pigmentation, the GC-MS/MS system (e.g. liner, inlet seals, MS source), will suffer. As you can see in Figure 3, the differences in pigment removal are striking.

Figure 3: Visual comparison of spinach extract after clean-up with individual dSPE

The high GCB dSPE (#26123 and #26219) are the best at removing the pigments. Table 1 provides a list of the type and amount of sorbent contained within the dSPE evaluated for this work.

Table 1: Description of dSPE used for optimization

As I mentioned in my celery blog, high GCB can also lead to removal of planar pesticides. Therefore, either #26123 or #26219 are viable options for spinach analysis. There are ways to mitigate this effect, mainly using internal standards, such as anthracene, deuterated standards, or take advantage of the matrix-matched calibration. Although these approaches work well, take care when you add the internal/calibration standards. Matrix-matched calibration won’t correct losses in the dSPE step if the internal standard or surrogate is added after the clean-up step.
I’ve also looked into which QuEChERS salts are the best to use with spinach (Fig 4). I got the best results with AOAC 2007.1 salts (6 g MgSO4, 1.5 g NaOAc).

Figure 4: Comparison of mean responses of selected pesticides using individual QuEChERS salts

To summarize, based on the selected pesticides, the best recoveries and matrix clean-up were achieved using AOAC 2007.1 salts in combination with dSPE with high GCB content (#26123 or #26219).
Next time I’ll tackle cleanup techniques for the analysis of oranges!

You can find part 1 (celery) here!

3-MCPD blog part 2: How hot we can go with inlet and can we use regular split/splitless inlet?

In my last blog, I (hopefully!) made a case that switching to split injection for analysis of 3-MCPD esters is a viable option. Now I’d like to show that if we stick to split, we can also use much higher temperatures and if we can do that, we can also use regular split/splitless inlet.

To evaluate, I used a similar approach as last time, i.e. set of calibration standards. Standards were analyzed at 120°C and 280°C with MMI inlet and at 280°C using split/splitless inlet (Figure 1).

Figure 1: Comparison of calibration curves of 3-MCPD-d5 obtained using different inlets and inlet temperatures.

Unfortunately, the data were not normalized with an internal standard, therefore, the response of second analysis (split/splitless) was higher than for the analysis with MMI due to evaporation. Nevertheless, the calibration curves from analysis using MMI inlet are almost identical for both temperatures, showing that the temperature of the inlet has little to no effect on the analysis (Fig. 1).

This means that with split injection we can use either MMI (PTV) or regular split/splitless inlet.

Next time I’ll look into how switching extraction solvents affects the analysis!

This Valentine’s Day is all about QUICK connections…

This year, you do not have to be in a long-term, committed relationship in order to embrace Valentine’s Day. In fact, this year is dedicated to “QUICK connections.” You say WHAT???? Check out the following video and you will understand:

Choosing the right dispersive SPE for GC-MS/MS analysis of celery

I’ve recently started experimenting with QuEChERS extractions for pesticide analysis. The available options are overwhelming, especially when it comes to dispersive solid phase extraction (dSPE) for the cleanup. For starters, I’ve been looking at the smaller volume dSPE (2 mL, summarized in Table 1), because I didn’t want to waste the raw material, solvents and salts. Using 15 mL dSPE we’d be using 6-8 mL of extract, while 2 mL dSPE uses only 1 mL, which allows for more replicates per extraction.

Table 1: Summary of available dSPE

Read the rest of this entry »

What are GenX and PFBS? Why are they important in PFAS analysis?

Analysis of PFAS (per- and polyfluorinated alkyl substances) has been a hot topic for envoironmental labs for the last several years. Currently there are two C8 based PFAS compounds (PFOA and PFOS) with a health advisory level of 70 ppt in drinking water announced by the US EPA. In 2018 at the National Leadership Summit, the US EPA announced that they will work with other US federal agencies to establish human health toxicity values for two additional fluorinated compounds, hexafluoropropylene oxide dimer acid (HFPO-DA, a.k.a. GenX), and Perfluorobutanesulfonic acid (PFBS). They have been made to replace PFOA and PFOS, respectively.

(GenX and its precursor acid hydrolyze into HFPO-DA. Credit: Courtesy of Mark Strynar and Laurence Libelo/US EPA)

If your laboratory has a plan to add these compounds to your PFAS analyte list, HERE is what it looks like on our Raptor C18 column (9304552).

If you already have GenX and PFBS in your analyte panel, look what can be done faster on our Raptor C18 column (9304552). Total runtime of 4 min with often early-eluting PFBS well retained at 1.6 min for reliable quantitation.

Are you looking at even more analytes in the PFAS group? Do you want a sub-2µm particle size column solution?

I’m glad you asked. Don’t miss our 34 PFAS compound analysis with a short 8-min runtime solution below.

And did I mention why you need a PFAS Delay Column (27854)?

Stay tuned for more PFAS research updates from us.

Limitations of Alumina: What you should NOT inject on alumina / Al2O3 columns

Alumina (Al2O3) capillary PLOT columns have been available for about 35 years. They are best suited for separation of C1-C10 hydrocarbons and especially the C1-C5 range. The alumina has unique selectivity making it possible to separate all the unsaturated hydrocarbons at temperatures above ambient, see figure 1. While its an excellent choice for unsaturated hydrocarbons, special attention is required for components like pentadiene and 1,2-butadiene which can show reactivity on alumina, see


Fig 1 C1-C4 hydrocarbons separated on Al2O3 PLOT

Compounds that elute are basically all hydrocarbons up to C12 and  aromatics up to the xylenes. Permanent and noble gases like Nitrogen, Oxygen, Hydrogen, Neon, Argon, Krypton, Xenon will elute virtually without retention. Also CO elutes from alumina together with methane and the permanent gases. These gases can be separated, but the oven temperature has to be lowered to -100C or lower.

Gases like N2O are well retained on the Alumina column, and show good peak shape, see:

Several chlorofluorocarbons are retained, however the type of alumina and deactivation used will determine the (re)activity, see below.


We will discuss here the components we should NOT inject starting with the ones that have the biggest impact: water


Traces of water will deactivate the alumina and the result is that retention decreases see figure 2.  Water influences overall retention and changes the selectivity of the column.  This can be seen when analyzing polar hydrocarbons, like methyl acetylene.

What can be done?

  • Remove the water from the sample. This is difficult and adds a lot of time.
  • Elute the water after every injection. This can be done by heating the alumina at high temperature (200/250) for 8-10 minutes. The AluminaBOND MAPD can be heated up to 250C, speeding up this process.
  • If water is present at trace level, give the column a conditioning after XX analysis. Usually XX is determined when the target peak runs out of the integration window.
  • Use a Rtx-Wax pre-column (thick film) via a valve in series with the Alumina. Water will be retained on this column, while C1-C6 will elute first. When the C1-C6 have entered the alumina, the water peak is switched to vent or sent to a second detector (if it needs to me monitored)

Fig. 2 Impact of water on retention of Al2O3

POLAR COMPOUNDS (alcohols, amines, acids, amino alcohols, ammonia)

Polar compounds will not elute as a peak from the alumina column. If they elute, they will elute as a “blob” or a raised baseline and once they are baked out of the column the column will continue to perform well.

If polarity of component decreases, the impact on the alumina will also decrease. So, ketones, ethers and aldehydes will still be adsorbed but are easier to elute at higher temperatures.


Halogenated compounds may elute OK from the alumina column. We can measure several CFCs using AluminaBOND CFC columns, see figure 3. Reactivity can occur with some halogenated components, Figure 4 shows examples for 1,2 dichloroethane and 2-chloropropane, which decompose into propylene and vinylchloride while splitting off a HCl molecule.

Fig. 3 CFC separations on AluminaBOND CFC

Fig. 4 Reactivity of alumina


Carbon dioxide will be adsorbed completely, but it takes significant amounts of CO2 to impact retention times. If the retention times decrease, condition the column for a few hours at its maximum temperature limit.


Sulfur compounds will be strongly adsorbed. Hydrocarbon samples containing Sulfur compounds in the sub ppm range (COS/H2S) can be analyzed with minor changes in retention time.



Hydrocarbons C12 and above will not elute as discrete peaks and will manifest themselves as a raised baseline. The alumina column will appear to “bleed”.  If it bleeds, it is an indication that there is something adsorbed on the column. Make sure you periodically condition the column at maximum programmable temperature until the baseline is flat.


Be aware that siloxanes can also stay on the column and will produce a raised baseline. This is often caused by the septum. For light hydrocarbons, try to use a moderate injection port temperature (80-100°C). You do not need a 250°C injection port temperature for mixes that contain C1-C6 compounds.



In all cases when there are components present that are adsorbed on alumina, and that do impact your retention and separations, the best way is to use a pre-column that can be back-flushed. This way the adsorbed component do not make it onto the analytical alumina column and retention times for hydrocarbons will be reproducible. You can choose to use a short (1-2m) alumina column as a guard, and do the back flushing using valve switching or using a Deans (flow) switching. Today these systems are very easily to setup and operate.