FDA Issues Second Warning on Methanol Based Hand Sanitizers

The U.S. Food and Drug Administration has placed methanol containing hand sanitizers on an import alert (1). These products do not list this ingredient and some are incorrectly labeled “FDA approved.” Currently, 87 hand sanitizers have been found to contain methanol (2). Methanol (methyl alcohol Cas# 67-56-1), known as wood alcohol, is commonly found in solvents, and windshield washer fluid (3). Ingestion of 10 mL can cause permanent blindness and 15 mL is considered a lethal dose (4,5). Transdermal methanol poisoning has been well documented and may result in optic nerve necrosis with permanent eye damage (6,7,8,9,10,11). Denatured alcohol is ethanol mixed with other alcohols and can contain 50% methanol. Other additives include isopropyl alcohol, acetone, methyl ethyl ketone, ethyl acetate and methyl isobutyl ketone.

Figure 1: Pro EZGC Chromatogram Model of a denatured alcohol sample illustrating the ability of this column / conditions to determine percent levels of methanol in samples, specifically hand sanitizers. Measured retention times are also indicated on the far-right column as a comparison to the predicted retention times.


This is the first part in a series addressing the different aspects of hand sanitizers. Our first goal is to determine a suitable column that can resolve methanol from water and ethanol. We accessed the safety data sheet (SDS) for our denatured alcohol sample and entered those compounds into our Pro EZGC Chromatogram Modeler (12). The program provides 5 different solutions / columns. Our column choice will be discussed in future blogs. Figure 1 is a model of the compounds with the expected retention time compared to the measured retention time using a mass spectrometer. While the model suggests resolution of the compounds, the sample has been diluted 50:1 in distilled water and the solvent peak width is not calculated by Pro EZGC.

Figure 2: Denatured alcohol sample analyzed using a Rtx-VMS by GC/MS with an adjusted scan range to accommodate water and methanol detection. This method is suitable for the analysis of hand sanitizers containing methanol.

We followed USP 611 Method II (13) as a starting point. The MS scan range started at m/z 10 to detect water (m/z 18) and methanol (m/z 31). We also changed the GC program from the conditions suggested by Pro EZGC by adding a 3-minute hold time for better solvent focusing and slowed the initial ramp rate down to 3°C a minute. Future work will rely on our method translator to determine an equivalent solution using a Flame Ionization Detector (FID). Figure 2 is a total ion chromatogram (TIC) showing good resolution between water and methanol. One way to optimize the resolution of early eluting compounds from the solvent peak in the split mode is the use of the 4.0mm ID Precision Inlet Liner w/ Wool, based on Linx Waclaski’s work (14). Using a rinse solvent other than water could result in extraneous peaks, ghost peaks and carryover. The challenge with using 100% water as a rinse solvent is the potential that residue will build up in the syringe barrel. My colleague Corby Hilliard developed a method that incorporated a cosolvent to prevent syringe damage (15). In this case we used a prewash with 100% water and three post washes with 90% water and 10% n-propanol. This method is suitable for measuring percent levels of methanol in samples. Our next blog will present gel and liquid hand sanitizers using this method.

  1. https://www.accessdata.fda.gov/cms_ia/importalert_1166.html
  2. https://www.fda.gov/news-events/press-announcements/coronavirus-covid-19-update-fda-reiterates-warning-about-dangerous-alcohol-based-hand-sanitizers
  3. https://www.vercounty.org/MSDS/SDS-EMA/88-RainEx%20Deicer%20Washer%20Fluid_SDS.pdf
  4. https://www.sciencedirect.com/science/article/pii/S0379073818303037
  5. https://www.msdsonline.com/2014/07/22/methanol-safety-tips-from-msds-experts/
  6. https://pubmed.ncbi.nlm.nih.gov/19628396/
  7. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3617539/
  8. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5292112/
  9. https://pubmed.ncbi.nlm.nih.gov/26196361/
  10. https://pubmed.ncbi.nlm.nih.gov/29908646/
  11. https://www.sciencedirect.com/science/article/pii/S2173579420300530
  12. http://archpdfs.lps.org/Chemicals/Denatured-Alcohol_Ace.pdf
  13. http://www.uspbpep.com/usp29/v29240/usp29nf24s0_c611.html
  14. https://blog.restek.com/gc-inlet-liner-selection-part-ii-split-liners/
  15. http://www.chromatographyonline.com/injecting-water-gc-column-solving-mystery-poor-chromatography-0?pageID=3

LPGC – Fast way to your pesticide analysis!

Throughput is one of the most important parameters in the lab. The more samples we can analyze in a day, the sooner we can get home. Enter Low Pressure GC (LPGC) – this is an invention from our brilliant Jaap de Zeeuw [1-2], where a relatively short analytical column (10 – 15 m) with large ID and thick film (e.g. 0.53 mm and 1.0 µm, respectively) where the flow is restricted with a narrow guard column (e.g. 5 m x 0.18 mm). The restrictor (guard column) allows a normal head pressure at the inlet, while the analytical column is operated under near-vacuum conditions. The low pressure inside the 0.53mm column, shifts the optimum linear velocity about a factor 7 higher, which allows for faster analysis without a total loss of efficiency. The wider ID and thicker film provides also higher capacity, robustness and inertness. In addition, an integrated transfer line adds additional robustness to the method as the absence of phase in the heated transfer line to the MS, helps to reduce background and make the system stabilize faster. Figure 1 shows the LPGC system schematics.

Read the rest of this entry »

Water, can’t live without it so how can we deal with it (and adsorbent columns)

Often the gas samples contain some water vapor. Although we are usually not interested in the amount of moisture present in the sample, if we don’t dry it, the water will be injected onto the column with our sample. Alan, in his blog, has previously discussed the effects of water on capillary columns with liquid stationary phases and adsorbent (PLOT) columns. But as they say: a picture is worth a thousand words, so I want to share a few examples I gathered while working with adsorbent columns and the impact of  water/water vapor.

Molecular Sieve 5A

Molecular Sieves 5A are zeolites that are part of the family of hydrated aluminosilicate minerals. They are known adsorbents with a unique, tunnel-like crystalline structure and a well-defined pore size. The analytes separate on molecular sieves based on two adsorption mechanisms. First, how well the molecules fit into the pores of the material, a separation based on the size of the molecules. Second, the physical interactions between the molecules and the MSieve 5A crystal, a separation based on the polarity.  For example, nitrogen and oxygen are both small enough to fit in the pores of 5A mineral, but oxygen, a smaller molecule than nitrogen, will navigate through the “tunnels” faster because it has less interactions with the pore surface and elutes from the column before nitrogen. In the case of carbon monoxide, polarity plays a more important role and the MSieve 5A strongly retains these molecules. Adsorption on molecular sieves is reversible and the adsorption/desorption process is easily regulated with the temperature.
We know that we can’t analyze water using the MSieve 5A columns because of the high temperature required to desorb polar water from the pores of the molecular sieve. What happens if we are injecting water onto the column? How much water can we inject before it’s “too much”? How does “too much” look like? Can we regenerate the MSieve 5A column? And how long does it take to regenerate it to its original performance?

To find the answers, I have performed a quick study where I injected water onto a Rt-MSieve 5A column, 30m x 0.53mm x 50µm (Cat#19723) in between the injections of permanent gases. I kept the system at 40°C the entire time, simulating isothermal analysis of gases. Right after 1µl injection of water I noticed carbon monoxide’s retention time shifted, eluting earlier (Figure 1).


Figure 1: Overlay of chromatograms of permanent gases, black overlay – initial analysis, blue chromatogram – analysis after injecting 1µl of water onto the column.

I finally stopped injecting water after carbon monoxide’s capacity factor (k’) was at half of its original value. Until then, 100µl of water was injected onto the column under the described conditions. Analysis of permanent gases at that point shows that adsorption sites are packed with water, analytes start to tail indicating sample loading capacity is decreased, and gases elute faster or even co-elute (Figure 2, Chromatogram B). 100µl of water on the column may not seem like that much until we put it into perspective. Let’s say our sample has 50% relative humidity. With every 1ml injection of gas, we are introducing ~0.01µl of water onto the column. That means, to lose 50% of retention for carbon monoxide we can make 10,000 injections.


Figure 2: Analysis of permanent gases (in order of elution: Argon, Oxygen, Nitrogen, Methane, Carbon Monoxide, concentration 3-5mol%, 100µl split 60:1 injection)), 40°C Isothermal, flow He 4 ml/min Chromatogram A: Analysis using a new column, Chromatogram B: Analysis of permanent gases after 200 injections of 0.5µl water at 40°C Isothermal, Chromatogram C: Analysis of gases after fast 20 min conditioning, Chromatogram D: After 2h conditioning, the water was removed from the column.

To regenerate the column the water can be desorbed by conditioning at the column’s maximum temperature, 300°C.  I wanted to track the time required for the column to recover by monitoring the capacity factor of carbon monoxide every 20 min.  After 20 min of conditioning, the water “distributed” through the column, and the peak shape of the analytes improved (Figure 2, Chromatogram C). The original column performance was restored after 2 h of conditioning (Figure 2, Chromatogram D and Chart Figure 3).


Figure 3: Graph of water desorption process relative to the k’ of carbon monoxide. The orange plot is the capacity factor of CO at the beginning, and the blue plot is the capacity factor of CO after conditioning at 300°C relative to the time.

Continuous injections of water on the MSieve column without elevating the temperature of the analysis will affect the chromatography, resulting in peak tailing, loss of retention, and resolution. Just remember, even when 50% of the column performance is lost the column can be restored to the original performance by conditioning it at the maximum temperature.

Porous Polymer Columns (Rt-Q, QS, S and U BOND)

Water does not affect the porous polymers and will elute from the column as a peak. Depending on the polarity of the polymer water will be more or less adsorbed and elution time will change accordingly. Figure 4 are chromatograms of permanent and hydrocarbon gases containing water vapor. On the nonpolar divinylbenzene Rt-Q BOND column, water elutes faster, and naturally, water is more retained on the most polar column, Rt-U- BOND. Bear in mind that water will not be detected with the FID detector.


Figure 4: Analysis of permanent and hydrocarbon gases on 30m x 0.53 mm x 20µm Rt-Q and Rt-U BOND column (Carrier gas: He@5ml/min, Oven: 40°C (3min) then 10°C/min to 190°C, Detector: TCD)

ShinCarbon Column

ShinCarbon column can adsorb a very limited amount of water. The amount of adsorbed water is dependent on the column dimensions (amount of the material in the column). The adsorbed water has no impact on the retention times of the compounds. However, eventually, if the water is not conditioned out of the column or if the injected concentration of the water is high, it will show up on the chromatogram as an overloaded peak with an almost “never-ending” tail and interfere with the integration of the analytes. Water will always show as “overloaded”. Therefore, depending on the injected concentration retention time of the water peak will shift.


Figure 5: Overlay of three chromatograms on a 2m x 1mm 100/120mesh ShinCarbon column– 30µg (green), 80µg (red) of water injection, and permanent gas with C1-C2 hydrocarbon standard (black) – showing where the water will elute from the column. (Carrier gas: He@10ml/min, Oven: 40°C (2min) then 20°C/min to 200°C, Detector: TCD)

Moisture present in the carrier gas will over time have a similar effect on the column performance. Mainly MSieve and ShinCarbon columns act as a carrier gas scrubber, trapping the moisture present in the carrier gas (Figure 6). Note that these materials are hygroscopic and will also attract water when the column is not sealed upon storage.


Figure 6: Overlay of the first and second instrument blank injection after the instrument has set idle with the ShinCarbon column installed at 40°C for 48 h. (I think it was time to change my carrier gas filter.)

Don’t forget, carrier gas filters, which will remove the moisture and other impurities, are essential accessory with this type of analysis.

Thermo Trace 1310 Inlet Temperature Profile vs Agilent 7890 for Split/Splitless Injectors

Several years ago, my colleague Scott Grossman wrote an excellent article entitled “It’s a Matter of Degrees, but Do Degrees Really Matter?”  He measured the temperature profile across various Agilent inlets, demonstrating different gradients in temperature exist across inlets, depending on the type of inlet and even among the same inlet type.  One consistent finding is the top and bottom of the inlet are always cooler than the middle. The middle is closest to the actual temperature setpoint.  Depending on the inlet, some had more significant drops at the top or bottom vs others and this may affect chromatography.

I wanted to repeat this experiment on a Thermo Trace 1310 split/splitless inlet to see how it compares to an Agilent 7890 split/splitless inlet.  Please note that this was not to prove that one is better or worse, but rather to understand differences. This information will help when making decisions about inlet temperature setpoints, ultimately affecting things like vaporization potential, compound degradation, septa temperature, etc.  This is especially important in a lab operating both instruments, running high molecular weight compounds.  The fix in this case could be as simple as running one instrument at a slightly higher inlet temperature, since a difference in temperature at the bottom of the inlet will affect performance.

To obtain this data, I took a thermocouple probe and inserted it through the septum until it reached the bottom of the inlet.  After recording this temperature, I would mark and pull up the thermocouple probe in 5 mm increments, recording the temperature at each location, until I reached the top of the inlet.  I plotted the inlet temperature gradient from bottom to top for each inlet (see below).  For these experiments, I used an inlet temperature setpoint of 250 ⁰C, as this is a relatively common setting.  I also checked the Thermo inlet profile at 300 and found the same relative percent error at each inlet location compared to the 250-degree experiments.

Here’s what I found:

Comparison of inlet temperature profiles of Thermo Trace 1310 split/splitless inlet with Agilent 7890 split/splitless inlet.

What you may immediately notice is that the top of the Thermo Trace 1310 inlet is significantly cooler than Agilent’s 7890 inlet.  There is a drastic drop in temperature over the top 20 mm.  This should not affect your compound volatilization, since the sample is injected towards the center, but it will affect the temperature of the septum.  For the Thermo Trace 1300/1310 inlet, I recommend Thermolite Plus septa, which are softer than BTO septa.  BTO septa are ideal for high temperatures and will be less pliable at lower temperatures and also more susceptible to coring.  Because of the lower temperature of the septum, bleed should not be a major issue, either.

You may also notice that the bottom of the Thermo inlet is slightly cooler than the Agilent inlet.  This could affect the vaporization of analytes.  If you are using both instruments in your lab, operate the Thermo Trace 1300/1310 at a higher inlet temperature compared to the Agilent to get comparable results.  This would be most evident with response of compounds with high boiling points.

As you can see, different instruments may have different temperature gradients across their inlets.  This does not make one instrument superior to another, but can affect performance and lead to perceived differences between the instruments, if operated under the same conditions.  Also keep in mind that this data was acquired on just one inlet from each manufacturer.  It’s possible for variation to exist between inlets from the same manufacturer, as Scott Grossman demonstrated in his work.

Did you know you can manage jagged bleed with a controlled cooling program?

GC column bleed sounds like one of those old problems people used to have. The XLB was developed to lower detection limits by minimizing bleed. Now, virtually all polysiloxane GC columns are low bleed, with thick film and high polarity phases being the exception. Modern GC-MS instruments are so sensitive that I’m rarely concerned about bleed interfering with my analysis. Sure, if your quant ion shares the same m/z as a major bleed ion, the baseline may be elevated, but it is rarely a problem.

Bleed becomes a problem for GC-MS analysis when you are looking at low levels of analyte and the baseline is jagged. Stationary phases with elevated phenyl content are more prone to jagged bleed, like this Rtx-PCB shown in Figure 1. Acquiring selected ion monitoring (SIM) data, as I have done for all these examples, exacerbates the effect of the jagged bleed if some of your target ions happen to share the same mass as a bleed ion.

Figure 1 – Jagged bleed on the Rtx-PCB column taken up to 340 ⁰C

You may be wondering, “What causes jagged bleed?”

I believe column bleed condensing unevenly throughout the column at the end of a GC run when the oven cools in an uncontrolled manner causes jagged bleed. The vents on the back of the oven open and hot air is blown out, drawing cool ambient air into the oven. This rapid uncontrolled cooling causes random sections of the column to cool faster than others, abruptly condensing the column bleed unevenly in discrete bands throughout the GC column. During the subsequent run, each discrete band of condensed column bleed reenters the mobile phase and chromatographs like any other analyte. The result is random spikes in the baseline with varying heights and widths, each one attributable to the nature of a discrete band of condensed column bleed.

Jaap de Zeeuw wrote an article on ghost peaks for SeparationScience that covers this same topic, but for high cyanopropyl- phases installed in an FID (2013, Volume 5, Issue 7); He also condensed his entire series on ghost peaks into a poster presentation which can be found here on our blog. He came to the same conclusion as me on the root cause, though I think the temperature gradient situation in the oven during uncontrolled cooling is much more chaotic than what he describes as a difference in temperature between the top and bottom of the oven. We know that the four corners of a GC oven are cooler than the center, and that there is a temperature gradient from back to front, but once the oven vent flap opens and the high speed fan turns on, the oven becomes filled with air currents of varying temperature.

Assuming we are right about the cause of jagged bleed, the solution should be controlling the cooling rate of the oven so the GC column cools more evenly, reducing the amount of uneven bleed condensation. Figure 2 shows this is precisely what happened when we added a 20 ⁰C/min cooling program to the end of the oven program for the second run (turquoise). The subsequent run (purple trace) had a much smoother baseline.

Figure 2 – Three consecutive runs showing the effect of controlled cooling on the jagged bleed. The first run (black trace) ended with uncontrolled oven cooling. The second run (turquoise trace) shows the jagged bleed resulting from the first run’s uncontrolled cooling, and the descending baseline a result of 20 ⁰C/min cooling from 340 ⁰C to 200 ⁰C added to the end of the GC oven program. The third run (purple trace) shows a much smoother baseline resulting from the previous run’s 20 ⁰C/min cooling.

Jaap’s example implemented a 10 ⁰C/min cooling rate from 260 ⁰C all the way down to 60 ⁰C, but the cyanopropyl phases he was dealing with (1701 and 1301) experienced the region of jagged bleed at a much lower temperature range than the Rtx-PCB.

Figure 3 highlights the difference in data quality for low level Aroclor analysis. Injecting 20 pg of Aroclor 1268 on column, each of the individual congeners are at much lower level. The blue trace highlights how difficult it can be to discern real peaks from the jagged baseline. The black trace, which follows a run with a controlled oven-cooling program, has a much smoother baseline, and the PCB congeners are much more obvious and easier to integrate.

Figure 3 – 20 pg Aroclor 1268 on a 60 m x 0.25mm x 0.25 µm Rtx-PCB column. The overlay shows a relatively smooth baseline in the run following a controlled 20 C/min cooldown to 200C (black trace), and jagged bleed in the run following an uncontrolled oven cooling (turquoise trace).

The effect of the controlled cooling rate added to end of the GC oven program is very reproducible, allowing you to turn a jagged baseline on and off (Figure 4). There are even fingerprint regions where the jagged bleed looks the same in every run with jagged bleed, supporting Jaap’s claim about the influence of positional temperature gradients.

Figure 4 – Sequential Run on an Rtx-PCB illustrating the immediate effect a controlled oven-cooling rate has on jagged bleed. The first run after controlled cooling (3) has a significantly improved baseline. Ending the controlled cooling causes the jagged bleed to return in the following run (6).

I am currently working on a low level PCB congener application by GC-QqQ using the Rtx-PCB. I do not expect jagged bleed to be an issue because of the addition selectivity provided by the collision cell and second mass analyzer, but I will update either way.

The New U.S. EPA Method TO-15A Blog Series-Part4: Clean Lines for Clean Air

In previous blogs (Part 1, Part 2, Part 3) we’ve covered the need to use air instead of nitrogen as a fill gas, the new lower blank requirements, and how these increase the importance for very clean lab air for canister filling. We’ve shown that using high quality gas filters or generators will get you the clean air required, but getting it to the canister ends up being more complicated than one might expect.

When we first started testing the new TO-15A blank requirements, we wanted to get the best baseline possible for our air quality. To do this we hooked our fill gas line up directly to our preconcentrator, allowing us to test the dry air without worrying about canister cleanliness. Note that TO-15A recommends testing your gas directly if possible in section 9.3. We found that while we were able to meet the 20 pptv requirements for most compounds, some of them were giving us trouble. Note that while we did test for the full 75 compound list for TO-15A and NJ mix I’m only showing the problem compounds for simplicity’s sake.

Compound pptv
n-Pentane ND
Acetonitrile 27
Carbon disulfide 72
Isopropyl alcohol 24
Methylene chloride 27
Acetone 48
Hexane 344
Tertiary butanol 44
Tetrahydrofuran 71
2-Butanone (MEK) 32
Toluene ND

Table 1: Selected TO-15A blank results. Average of 3 tests.

We were confident in the quality of our incoming air, so where was the contamination coming from? We were using a setup that let us blend dry and humid air to a desired relative humidity, and while it’s very handy it also complicates the flow path with extra tubing, fittings, valves, and gauges. It had been in use for quite a while, and even low levels of contamination in the air and water could have potentially built up over time. So to test that we bypassed the system, which gave us much improved results.

Compound pptv
n-Pentane ND
Acetonitrile 23
Carbon disulfide 14
Isopropyl alcohol ND
Methylene chloride ND
Acetone ND
Hexane ND
Tertiary butanol ND
Tetrahydrofuran ND
2-Butanone (MEK) ND
Toluene 21 

Table 2: Selected TO-15A blank results after bypassing mixing system. Average of 3 tests

Acetonitrile is still slightly high, and toluene has increased as well. The air line used was next to our SPME prep area which had some small bottles of solvents, including acetonitrile. To better isolate the air from this we installed a new air line across the bench, about 15 feet away.

Compound pptv
n-Pentane ND
Acetonitrile 16
Carbon disulfide ND
Isopropyl alcohol ND
Methylene chloride 20
Acetone ND
Hexane ND
Tertiary butanol ND
Tetrahydrofuran 10
2-Butanone (MEK) ND
Toluene ND

Table 3: Selected TO-15A blank results after moving air line. Average of 3 tests.

The new line was a bit higher in methylene chloride and tetrahydrofuran, but finally got us to 20pptv or less for the other compounds. From all of this I can give a few key pieces of advice.
• Test your air with the shortest, simplest flow path to get baseline cleanliness
• Compare any extra systems, such as mixing and humidification chambers, to this baseline to determine their contribution to your blanks. Over time these systems may need to be cleaned or replaced to meet the 20pptv limits.
• Isolate your systems from solvents as much as possible. Having your air lab (and volatiles lab if you run 8260 or similar methods) completely isolated from any semivolatile GC, GC prep, or LC instruments is ideal, but if they can’t be in separate rooms at least keep them as distant as possible.

Once you are confident in the cleanliness of your air, the next complication is humidification. The next blog in this series will address that, so stay tuned.

How I clean a GC injection port

Recently I have been asked by several customers how to clean an injection port.  My initial answer is always the same, you should clean the injection port according to the manufacturer’s recommendations.  But what instructions do I provide when the injection port was manufactured by Restek?  I usually provide the same response but then I provide instructions on how I cleaned an injection port when I was working in the lab.

I hope this post will help those who may not have a set cleaning procedure or are using an instrument where instructions are not provided by a manufacturer.  I would also like to hear from you if you have a different or unique way to clean an injection port.  After all, blog posts are about sharing information.

  1. Cool the injection port.
  2. Turn off the carrier gas.
  3. Remove the GC column.
  4. Remove all injection port consumables (including items such as the injection port liner, inlet seals, septa, fittings, etc).
  5. If possible, choose at least one polar solvent and one non-polar solvent which will not contaminate your instrument. Or look for a solvent which has a polarity somewhere in between.  In other words, make sure the solvents you choose are not target compounds and will not contaminate your detector.  For example, do not use methylene chloride if your detector is an ECD.  Several of my go-to solvents were acetone (moderately polar), methylene chloride (slightly polar), methanol (polar), acetonitrile (polar), cyclohexane (non-polar), toluene (non-polar) and hexane (non-polar).
  6. Use wooden-handled cotton swabs dipped in one or more of the solvents listed above and begin swabbing the interior walls of the injection port. I usually began with a non-polar solvent and while working my way up the polarity scale I finished with a very polar, low boiling point solvent like methanol.  Never finish with water because the injection port may rust.
  7. If possible, blow dry with clean, dry N2. Never use a compressed gas which contains propellants since they may contaminate the injection port.
  8. When all surfaces are dry, install new/clean consumables and an injection port liner. Install a new injection port septa.  Re-install the GC column.  Double-check to make sure everything was properly assembled and that nothing was left inside the GC oven before you close the oven door.
  9. Do not heat the injection port yet. Turn on clean and dry carrier gas and set the flow rate that whatever it was before you cleaned the injection port.  Purge for at least 30 minutes.  Leak check using an electronic leak detector.
  10. Heat the injection port to 100°C.  Make sure no fittings are loose and that the nut for the injection port septa is not too tight.  Increase the temperature to 150C and once again check the fittings (to make sure nothing is loose) and that the nut holding the injection port septa is not too tight.  Repeat every 50°C until the set point is reached.  Leak check often.
  11. Analyze several solvent blanks to make sure the baseline is free from any contamination.  If unknown peaks are detected, the injection port may still not be clean, or these peaks may be from something else.  If needed, review this post to help locate the source of these ghost peaks. Are your ghost peaks coming from the GC column, or something else?
  12. If you are still having activity issues or ghost peaks, consider installing a Uniliner.  Because a fused-silica GC column connects directly to this injection port liner, the inside of the injection port is prevented from contacting the sample/standard.  This should help confirm if the injection port is the issue.

If the instructions above did not remove the contamination sufficiently, you may need to try something a little more aggressive.  Several suggestions are below.

A.  Try using a nylon brush with a cleaning solution like Detergent 8. This solution (when diluted according to the manufacturer’s instructions) may help remove additional contamination. Make sure you completely rinse the injection port when finished using deionized water and then quickly rinse with a fast-drying solvent like methanol (to prevent any rust from forming).  Please note – if you decide to use a nylon brush with solvents or alcohols, always check its compatibility with nylon.  Please note – because of the high pH, do not use Detergent 8 on an injection port which has been treated with Siltek.

B.  You may want to try using a metal brush instead of a cotton swab or nylon brush if the contamination remains. Choose a brush that will not scratch the walls of the injection port and/or be damaged by solvents.  I usually used brass brushes because I knew they would not scratch the stainless steel injection port.   Please note – never use a metal brush of any kind if the injection port is treated with Siltek or some other inert coating.

C.  If a metal brush and solvents do not remove the contamination, it may be time to replace the injection port.

D.  Finally, you should clean/flush or replace the injection port weldment/septa nut and split vent line to prevent contaminating a clean (or new) injection port.


For additional information on troubleshooting injection ports, you may want to review the posts below. Thanks for reading.

Split Injections…The Good, The Bad and the Ugly

Contamination of Injection System Split-Vent-Lines: A Maintenance item not to Underestimate

Troubleshooting Injection Port Septa

The new U.S. EPA Method TO-15A blog series – Part 3: Use CLEAN air on a CLEAN analytical system!

From the start of this blog series, I have been teasing you with how to take your canister blanks from the following red trace down to the blue trace:

In this post, I shall finally stop dragging it out. BUT FIRST… (Come on now, you had to have known this was coming) I shall digress a little to point out that last time we hopefully convinced you to use AIR when analyzing AIR, and how this specific detail may impact your canister blank concentrations. This time we are going to talk about how we MUST (yes, not optional in my opinion) use CLEAN air on a CLEAN analytical system (i.e., preconcentrator, autosampler rack, etc.) before we worry about anything else. You will notice how we have not even begun to talk about canister cleaning details like cycles, temperatures, sweep gas, storage time, etc. That is all coming in due time (i.e., future posts), but none of that matters until we address the following. Last time we mentioned a prevailing theme of “garbage in = garbage out.” While we keep that in mind for today, let us forget about the zero air challenge for section 9.4.2 (page 32). In fact, let us back up a couple of pages in TO-15A and go to section 9.3 (page 30). Like I have said before, customer comments/inquiries make for the best posts. And after the last post, Melanie M. drops a comment in our blog, which includes the following excerpt:

“My question to you is do we need to test the autosampler? Section 9.3 My understanding is this is only for New autosamplers?”

It is as if we paid Melanie to drop that comment when she did, as the content and timing was perfect. YES Melanie, you need to test the autosampler in my opinion. Not just upon initial setup. In fact, you set me up for the following soapbox speech:

To kick off your journey to 20 pptv blanks, I suggest you take your fill gas (air obviously) and connect it to one of the autosampler ports and sample your nominal volume of air (e.g., 250 mL) through the entire analytical system (i.e., autosampler, preconcentrator, transfer lines, GC-MS, etc.) just like a canister blank. Now, section 9.3 of TO-15A says:

“This procedure should be conducted at installation prior to initial use of the instrument. This basic evaluation does not require establishing calibration or determining quantitative results to assess potential positive bias.”

I have the following to say about this:

  1. This procedure need not be limited to post-installation and prior to initial use. I advocate using this as the first diagnostic evaluation on your road to 20 pptv clean canister blanks. It also makes for a nice routine evaluation of your air source and analytical system.
  2. I could not disagree more on the non-calibrated evaluation. In fact, I had to soften the wording of this sentiment for the published version of this post. I advocate that you calibrate the system first, which for the record may require a secondary or tertiary calibration throughout this process, as you may have to make changes as you track down leaks/contamination. I could see a strong argument to start out qualitative, but then work your way towards a calibration. But regardless, calibrate the system for Pete’s sake. Since you are reading this post, it is safe to assume you know 20 pptv represents a significant challenge for canister blank cleanliness. In particular there is not a lot of room for error here. So, to start making cleanliness determinations on our air source and analytical system based on a qualitative assessment seems nonsensical.
  3. All of the above will also force you to ensure you have adequate analytical sensitivity to verify cleanliness down to 20 pptv. There is not a lot of logic in running air/system evaluations on an analytical system/method capable of only seeing 50 pptv, is there?

Excuse the digression. So by carrying out the exercise outlined in section 9.3 you can verify the cleanliness of your fill gas and entire analytical system. You also probably just knocked out section 10.1.1 of TO-15A as well. Talk about killing two birds with one stone! You literally knocked out almost half of the canister blank contributions we touched on last time. Until you can demonstrate 20 pptv or less for target analytes by way of the above exercise, you need not worry about how clean your canisters are. In fact, that is why section 9.3 comes before 9.4. Although obvious for many, I feel compelled to point this out, as so much of the buzz appears to be focused on sections 9.4’s zero air challenge. I am concerned that this obvious, but hugely vital step found in section 9.3 may be overlooked. So thank you for your question Melanie!

Now about that teaser c-gram. What you see in the above c-gram represents the following scenario. I was looking at some canister blanks and having some higher than expected concentrations, which were not allowing me to achieve the 200 pptv blank cleanliness level outlined in TO-15 (notice, no “A”). After chasing my tail around with canisters, the cleaning oven, the transfer lines, etc… I finally performed the same exercise from section 9.3 and I generated the red trace in the above c-gram, which obviously was less than stellar. After going up one side and down the other on the autosampler, preconcentrator, and GC-MS all roads lead to my source gas of all things. I am still not sure what happened to this date, but I do know the fix. So without any further teasing, we put our lab air on the following Parker Zero Air Generator.

BAM! We went from red trace to blue trace overnight. And no, this is not a shameless plug to sell the above unit, as we were able to achieve comparable results with other means. For example, we used a Molecular Sieve S-Trap to achieve a comparable cleanup as with the generator. The problem is that the more economical approaches like the S-Trap do not handle the flow and volume desired for our whole laboratory operation. In addition, the restriction is too much for the higher flows we desired. And lastly, the volume is limited, so that you have to regenerate the trap more often than most care to do (at least me). So, we landed on the generator, which clearly works well. Note that the above c-gram represents a 400 mL sample of house air pre-generator (red trace) and post-generator (blue trace). It is also important to note that this is a TIC, which means the following:

  1. It will show compounds not included in your target analyte list. Which can be of real value, as we will find out in another follow-up post in this series.
  2. It is not representative of the true results. So, we have included the appropriate concentrations in the table below:

I am sure you have already made the observation that we only cleaned up our house air and verified our analytical system down to the old TO-15 level of 200 pptv and even missed the mark on 3 of the compounds. And you would be correct, as this work was conducted back in 2017. The goal of the current post was to demonstrate the importance of a clean air source and a clean analytical system. For the record, we did indeed track down the source of contamination for those 3 compounds (i.e., post gas generator), so they were no longer problematic. However, at this point in time I am going to hand over the virtual baton to Jason Hoisington (yes, same first name and darn close last name, just to keep you on your toes), where he walks us through similar exercises as section 9.3 and 10.1.1 to get us from TO-15 (200 pptv) down to TO-15A (20 pptv). After all, that is what you really came here for in the first place. I shall return with more discussions around storage time, sweep gas, etc… so stay tuned.

In the meantime, keep in mind that “garbage in = garbage out” and until you verified a CLEAN air source on a CLEAN analytical system, NOTHING ELSE MATTERS for achieving clean canister blanks!




Accounting for Atmospheric Pressure When Using EZGC Method Translator and Flow Calculator and the ProEZGC Chromatogram Modeler

Modern GC’s are equipped with advanced pneumatics controls, which allow for accurate control of flow rates and pressures.  In order to accurately control or calculate flow rates, atmospheric pressure must be measured by the GC, since this determines the outlet pressure of non-vacuum detectors.

Restek’s free tool, the EZGC Method Translator and Flow Calculator, has many beneficial uses to chromatographers, including translating methods, calculating column length, calculating splitless hold time, calculating average linear velocity, etc.  In order to make accurate calculations, outlet pressure must be entered. Setting an accurate outlet pressure is also important when using the ProEZGC Chromatogram Modeler, as this will lead to the best accuracy of retention time predictions.

Both the EZGC Method Translator and Flow Calculator and the ProEZGC Chromatogram Modeler allow you to select your outlet pressure as either vacuum or atmospheric (Atm), based on the detector (Figure 1).  For a mass spectrometer, you should select vacuum outlet pressure whereas, for a detector such as FID, select atmospheric outlet pressure.  Note, that when you do this, a default value of 14.70 psi (equivalent to 101.33 kPa, 1.01 bar, or 1 atm, depending on your chosen units) is auto-filled.  This equates to standard pressure at sea level.

Figure 1: Outlet pressure is found under the “Control Parameters” section. If you select “Atm” for an atmospheric pressure outlet, a default value of 14.7 psi will be automatically entered. You can manually change this value to account for elevation above sea level. Also note that pressure units can be changed using the drop-down menu located next to “Inlet Pressure”.

Increases in elevation lead to a decrease in atmospheric pressure.  For example, Denver, Colorado is at an elevation of 5,280 feet.  Here the atmospheric pressure is 12 psi, compared to the 14.70 psi found at sea level.  While this may not seem significant, it is enough of a difference to affect the accuracy of flow calculations and retention time predictions.  If you were to enter in the same inlet pressure as a lab at sea level, your average linear velocity would be higher since the pressure at the outlet of the column is lower.

Figure 2 illustrates the difference outlet pressure can have on flow and therefore analyte retention.  In the left chromatogram, the outlet pressure is set at sea level (14.70 psi) and inlet pressure is fixed at 15 psi.  In the right chromatogram, the outlet pressure is entered for an elevation of 10,000 feet (10.1 psi), say if you were in Breckenridge, Colorado, also with a fixed inlet pressure of 15 psi.  While both figures have the same inlet pressure and temperature program, the difference in outlet pressure leads to a difference in average linear velocity of carrier gas, causing the retention time of deltamethrin to vary by 45 seconds between the two chromatograms.

Figure 2: ProEZGC Chromatogram Modeler chromatograms demonstrating the difference that outlet pressure can have on retention time. The model on the left illustrates the retention time of deltramethrin at sea level, when inlet pressure is set to 15 psi. The model on the right illustrates the retention of deltramethrin at 10,000 feet, with the same conditions, except for outlet pressure, which is 10.1 psi, due to the elevation difference. Deltramethrin has a retention time difference of 45 seconds between these two models.

If your lab is at a higher elevation, you should manually enter an accurate value for outlet pressure when performing calculations in both the EZGC Flow Calculator, as well as the ProEZGC Chromatogram Modeler, rather than use the default “standard pressure” value.  Table 1 gives some example values for atmospheric pressure at various elevations.

Table 1: Atmospheric pressure vs altitude.

Keep in mind that modern GC’s should measure and account for atmospheric pressure.  Some GC’s will even display the atmospheric pressure, which you can then use for accurate calculations with the EZGC software.



Hydrocarbon Ranges of Natural Gas and Petroleum Products

Sometimes customers need to analyze natural gas and petroleum products, whether it is to confirm the purity of the product by looking for the impurities which may be present, or for an environmental reason such as a spill or leak.  In order to choose the correct GC column and reference standards, one needs to know the hydrocarbon ranges of these products.

Since many of these products are analyzed following ASTM International methods to determine their purity and to confirm that they meet certain specifications, I highly suggest you review the information contained within this link that was created by the petroleum experts at Restek.   It will guide you to an appropriate ASTM method, GC column(s) and reference standard(s).

ASTM Petrochemical Method Chromatography Product Guide


If the analysis is for environmental reasons and not product quality, the links below should help guide you on a variety of topics.   Although I included primarily US EPA methods, I also included a few state specific methods.

For information on underground storage tanks, you may wish to review the links below.


Underground Storage Tank Monitoring (UST)



For environmental analytical methods, including contamination via spills and leaks, you may wish to consider EPA 8015/8260/8270.

EPA Method 8015D

EPA Method 8260D

EPA Method 8270E


If you are looking to achieve a more specific and/or comprehensive analysis, consider one of these state methods shown below.

TNRCC Method 1005


Determination of oil and grease and total petroleum hydrocarbons (TPH) in wastewater


Please remember that, as natural products generally produced by either natural gas processing or by the distillation of crude oil, these ranges can vary, which is why I stated that they are average (approximate).  I also included their average/approximate boiling point range (where applicable) and links showing Restek chromatograms & reference standards and/or information (mostly from Wikipedia) on the product.  I hope you find this information useful for your next natural gas or petroleum product analysis.


Natural Gas

Natural Gas Chromatograms

Natural Gas Reference Standards


Raw Gas:  Primarily C1 (methane), lesser amounts of C2 hydrocarbons through C4 hydrocarbons + impurities such as light sulfur gases, fixed gases and moisture.

Other condensates may also be included (C5+).

Processed Natural Gas/Liquefied Natural Gas (LNG):  Primarily C1 (methane), low levels of C2 (ethane) and very low levels of impurities.


LPG (Liquified Petroleum Gas)


Primarily propane (C3) and butane (C4) with low levels of other light hydrocarbon gases.



Gasoline Chromatograms

Gasoline Reference Standards


Average hydrocarbon range of C6 through C12.  Average boiling point range 70°C to 220°C


Jet Fuel/Kerosene

Jet Fuel Chromatograms

Jet Fuel Standards


Kerosene Chromatograms

Kerosene Reference Standards


Average hydrocarbon range of C8 through C18.  Average boiling point range 130°C to 320°C


Diesel/Fuel Oil #2

Diesel Chromatograms

Diesel Reference Standards


Average hydrocarbon range of C10 through C28.  Average boiling point range 170°C to 430°C


Motor Oil

Motor Oil Chromatograms

Motor Oil Reference Standards


Average hydrocarbon range of C22 through C40.  Average boiling point range 370°C to 520°C