Phthalates

[natali.g17]

Does anyone have experience with method development for the migration of phthalates from plastic packaging? It's kinda urgent.

[rb6banjo]

A little. I worry about migration of all kinds of things from food to packaging and vice versa.

Phthalates are tough because they're pretty much everywhere so it's really difficult to get a good "blank".

[natali.g17]

I'm trying to develop a method for migration from plastic cup into 3% acetic acid. Volume of the cup is 200 ml, it's filled with 170 ml of acetic acid solution. It stays in the incubator for 2h at 70°C. Then the solution is tranferred to 200 ml volumetric flask, homogenized in ultrasonic bath for up to 1h. After homogenization 10 ml of the solution are transferred to 50 ml centrifuge tube, 5 ml of hexane is added and its attached to a rotating mixer for about 1h so the liquid liquid exctraction can take place. Finally, 1 ml of hexane is transferred into a GC vial and analysed on GC-MS. Calibration is made by diluting the 100 ppm and 10 ppm mixtures with acetic acid solution in 10 ml volumetric flasks. The whole volume is transferred to a centrifuge tube and further processed the same way as the sample. Calibration is kinda fine, I ordered some internal standards in hopes it will be even better, but I'm still waiting for them to arrive. The problem is in the spiked sample, the areas are smaller than those for the lowest point of the calibration curve, I suspect it's the homogenization and the part where 10 ml is taken from 170 ml, I don't know what else it could be. Any suggestions are welcome. Thanks in advance.

P.S. I tried doing solid phase extraction on C18 cartridges but calibration wasn't great. At least not when it was eluted at the speed of 1 drop/sec, it was better when it was eluted faster, but I was told that it's not supposed to be faster.

[Hollow]

GC is not my business but just some re-questions that came to mind reading your post.


- When did you spike?
- How and how much did you spike?
if spiking is done in the cup, it may be possible, that some of the phthalates adsorbs on the cups surface.
Try spiking different levels.
- also consider the uncertainty of the measurements.
How is the precision of multiple spiking experiments?
- Did you consider the dilution of the 170 ml to 200 ml in the volumetric flask?
(I assume you dilute to the mark, else why using a volumetric flask)
- Why the homogenisation for 1h in ultrasonic?
If the phthalate are in solution, then mixing solutions (by filling up to the mark) should be ok after several inversions and shaking of the volumetric flask.
- to locate the loss on the spiking, try spiking at different stages of your sample preparation and compare the findings

[rb6banjo]

I agree with all that Hollow mentioned.

What is the concentration of your lowest standard? Where should the spiked sample have fallen on your calibration curve?

Phthalates are easy to remove from water using SPE because they are so insoluble. You might try to use something more polar than C18. Perhaps C8 or phenyl stationary phase. In general, slower flow is better when applying the sample to the stationary phase packing. Too fast and some analytes can run right through there. Once it's on the extraction phase, as long as the elution solvent is strong enough (hexane should do it), then the speed isn't as critical. I like to go kind of slowly because recovery of the analyte is a competition between the stationary phase and the extracting phase. So, I still wouldn't blow it through there at 100 mL/min! Using this approach, your 0.1 ppm in the liquid, extracting 50 mL of sample into 1 mL of hexane, you end up with 5 ppm in the hexane. You could easily evaporate the 1 mL to 0.1 to get 50 ppm in your extract.

As I mentioned, you have to run blanks because there are phthalates everywhere and you likely won't get a zero blank. Then you have to deal with the whole "is the amount I measure for the sample significantly larger (from a statistical standpoint) than what I measure in a typical blank.

Do you have an expectation of what should/could be coming from the cup? if this is testing that is trying to say that migration is negligible, perhaps not being able to detect/recover a micro amount of the phthalates is not a bad thing. All methods reach their limit...............

If the concentration of the phthalate (I assume it's a particular one) is 0.1 ppm (µg/mL) in the 170 mL aliquot. Then, according to your sample prep, it's going to be about 1 ppm in the hexane extract (100% recovery). You're at the ragged edge of what I would say is good limit of detection for a direct injection GC method. Maybe if you are using SIM GCMS it'd be ok but if not, then in my opinion, you need more in your extract.

Seems to me that you need a method that will verify what you want to verify. Not one that necessarily needs to be the best method for determination of phthalates ever developed.

I found an EPA document that says 0.02 mg/Kg/day (Kg body mass) for DEHP provides no appreciable risk. Let's say that's the limit you can have. For a 150 lb person (68 Kg), one needs to ingest 1,360 µg of DEHP in his/her diet per day. Let's say the migration is 0.1 ppm into the 170 mL serving (phthalates are quite insoluble in water, happier to stay in the plastic). That's 17 µg per serving. That's 80 servings per day required to reach the limit. Is it reasonable?

[natali.g17]

Yes, spiking is done in the cup after filling it with 170 ml of 3% acetic acid solution. I fear that possibility too, but the idea comes from the method for migration of primary aromatic amines. The whole method is meant to mimic that one for amines actually. I don't know much about them, whether they can be adsorbed on plastics too and if they are more soluble in water than phthalates. My boss told me that they homogenized the solution on ultrasonic after heating and cooling (volumetric flask because it can be inverted and shaked for better homogenization I suppose) and then just took 1 ml of the solution and transferred it into a vial. The analysis for amines is done on LC-MS so no solvent exchange is needed. I tried spiking at a higher level, areas are slightly bigger, but the chromatogram in general is not much better, for example the last two analytes DINP and DIDP are barely present, one or two peaks only. The part that I'm most worried about is taking the 10 ml aliquot, since the phthalates don't seem to be very soluble in water, I'm not sure how good the homogenization can be. On the other hand, for calibration, the whole volume is used for liquid liquid extraction so I guess that's why the peaks are better. Calibration is in the range 0.25 mg/l - 5 mg/l. Spike on loq level is supposed to be 0.34 mg/l but it is below 0.25 mg/l. Only once, while doing solid phase extraction, I got 0.28 mg/l for DIBP and 0.27 mg/l for DBP, all the rest of them were below 0.25 mg/l. What also worries me when it comes to solid phase extraction is that both cartridges and the manifold are made of plastic, I can run blanks but I don't know how to check for possible adsorption. I asked my boss about this but they told me that phthalates can only come out of plastics, not stick to it... As I mentioned already, this method is mainly based on the method for amines, but also on ISO 18856 where phthalates are extracted from water using C18 cartridges, and ISO 15318 where the idea of "taking an aliquot" comes from. So yes, it is a method that "needs to verify what it's expected to verify".

[natali.g17]

As for the limits, we need to follow EU regulations, required limits can be found in Table 1 on the following link: https://measurlabs.com/blog/phthalates- ... gulations/

[rb6banjo]

Not correct. Phthalates can and will dissolve into plastics - thankfully. Polyvinylchloride is a much more useful plastic because phthalates dissolve in it readily. Phthalates are much happier in the plastic than in the water but if the plasticizer is ~50% of the plastic formulation (a soft chew toy or for gaskets for instance), you're bound to have some leaching out. It's sort of like the analogy that even rocks have solubility (Ksp constants). They're just not very soluble.
Modern analytical technology has allowed for us to detect lower and lower concentrations of phthalates and other leachables so now everyone is up in arms about them.

I would think that you might not have too much adsorption/absorption into the SPE tubes because it's hard polypropylene and migration takes time. You're probably having more trouble getting them off the C18 packing material once they're isolated there. That's why I suggested C8. Might be easier to isolate them after passing your sample through the tube.

The way to check that would be to inject a known amount of your standard directly to a tube. I would condition my tube and inject it into the liquid that remains in the tube after your conditioning step. In your case it should be water. You know 100% of it is on the cartridge. There's nowhere else for it to go. Work it up and see how it compares to what you expect. If you get 100%, all is well. If you don't you likely have an issue with your analytes getting hung up in there somewhere.

[MichaelVW]

Your calibration standards are run through the entire process, right? (Except they don't go in a plastic cup with acid?)

I would consider doing a trial calibration - just to see how it looks - by simply diluting standards in hexane and running them on the GC/MS. That way you can be sure you know what the linear range of your setup is.

Since you are doing the processing steps on your calibration standards, it will introduce variation. I'd probably want to do a few replicates on a single standard to make sure the results are consistent. This isn't my wheelhouse though so I'm just speculating. I wonder if this is a situation where you'd do multiple replicates of each calibration point and use the average.

[natali.g17]

Not correct. Phthalates can and will dissolve into plastics - thankfully. Polyvinylchloride is a much more useful plastic because phthalates dissolve in it readily. Phthalates are much happier in the plastic than in the water but if the plasticizer is ~50% of the plastic formulation (a soft chew toy or for gaskets for instance), you're bound to have some leaching out. It's sort of like the analogy that even rocks have solubility (Ksp constants). They're just not very soluble.
Modern analytical technology has allowed for us to detect lower and lower concentrations of phthalates and other leachables so now everyone is up in arms about them.

I would think that you might not have too much adsorption/absorption into the SPE tubes because it's hard polypropylene and migration takes time. You're probably having more trouble getting them off the C18 packing material once they're isolated there. That's why I suggested C8. Might be easier to isolate them after passing your sample through the tube.

The way to check that would be to inject a known amount of your standard directly to a tube. I would condition my tube and inject it into the liquid that remains in the tube after your conditioning step. In your case it should be water. You know 100% of it is on the cartridge. There's nowhere else for it to go. Work it up and see how it compares to what you expect. If you get 100%, all is well. If you don't you likely have an issue with your analytes getting hung up in there somewhere.

So it's possible that if they are spiked into the cup some of them will adsorb on its walls? I can't really get C8 right now, I pretty much have to work with what I've got and make it work somehow. I tried spiking the acetic acid solution in a glass beaker instead of a plastic cup and passed the whole volume through the tube (I read somewhere that it's not advisable to take aliquots since they are poorly soluble in water), eluted with 10 ml of hexane, transferred them to 50 ml volumetric flask and filled up to the mark (final concentration of about 0.57 mg/l). Got pretty much the same results; recoveries for the first 3 were okay, but for the last 4, they were still too low. I'm guessing internal standards might help but they won't arrive any time soon. So the only thing I've got left is liquid-liquid extraction, but then I would have to reduce the volume of the acidic solution again because the tubes are only 50 ml in volume. The only thing that comes to mind is to switch to a smaller volume container and spike in some glassware after heating (though I'm not sure it's the best possible way but I don't know how to eliminate the possibilty of adsorption otherwise).

[natali.g17]

Your calibration standards are run through the entire process, right? (Except they don't go in a plastic cup with acid?)

I would consider doing a trial calibration - just to see how it looks - by simply diluting standards in hexane and running them on the GC/MS. That way you can be sure you know what the linear range of your setup is.

Since you are doing the processing steps on your calibration standards, it will introduce variation. I'd probably want to do a few replicates on a single standard to make sure the results are consistent. This isn't my wheelhouse though so I'm just speculating. I wonder if this is a situation where you'd do multiple replicates of each calibration point and use the average.

I tried something like that. I did one point through the entire process, and one just diluted; in theory, they were supposed to be the same concentration, but the areas of the first one were much smaller. So I kinda have to run them through the whole process, but I see what you want to say.

[rb6banjo]

Ugh! This is the nature of the beast for these types of things. I'll think about it some more but I'm running out of suggestions...........................

[natali.g17]

Ugh! This is the nature of the beast for these types of things. I'll think about it some more but I'm running out of suggestions...........................

I switched to a smaller volume container (46 ml) and tried both spiking directly into the cup and after heating and transferring the solution to the centrifugation tube. The whole volume was transferred in both cases but the results were much better when spiking was done after transferring so I guess they really do adsorb on the surface of the cup.

[rb6banjo]

I went back and did some theoretical calculations for this problem. Let's say the "partitioning mechanism" is in play here. Analyte is soluble in both the extracting medium (the water in this case) and the solid (the cup) and that at equilibrium the amounts in each phase are dictated by K and the phase ratio (amount of solid relative to liquid). The model allows for 2-way migration (there could be analyte in the liquid phase, as would be the case in your overspike experiments). Depending on the conditions, x can be positive or negative.

https://i.postimg.cc/Qtgmj2bz/Partition ... -Calcs.jpg

Let's say you have 0.1% phthalate in the solid (1,000 ppm)and 0 ppm in the liquid at t=0. VL = 170 mL and M (mass of cup) is 1.00 g. A value of K=0.0001 (very small) gives 0.1 ppm leached into the water from the mass of cup. If you keep all of the other parameters constant and simply add 0.1 ppm of analyte to the liquid, most of what you add ends up in the cup and the concentration of the analyte in the liquid is essentially unchanged - certainly within the error of the measurement.

This is a reasonable explanation as to why the overspikes are not working out as you'd expect. You have to add quite a bit more to the water to move the needle. 1 ppm addition gives a liquid concentration at equilibrium of 115 ppb (compared to 100 ppb with no analyte in the water).

[natali.g17]

I went back and did some theoretical calculations for this problem. Let's say the "partitioning mechanism" is in play here. Analyte is soluble in both the extracting medium (the water in this case) and the solid (the cup) and that at equilibrium the amounts in each phase are dictated by K and the phase ratio (amount of solid relative to liquid). The model allows for 2-way migration (there could be analyte in the liquid phase, as would be the case in your overspike experiments). Depending on the conditions, x can be positive or negative.

https://i.postimg.cc/Qtgmj2bz/Partition ... -Calcs.jpg

Let's say you have 0.1% phthalate in the solid (1,000 ppm)and 0 ppm in the liquid at t=0. VL = 170 mL and M (mass of cup) is 1.00 g. A value of K=0.0001 (very small) gives 0.1 ppm leached into the water from the mass of cup. If you keep all of the other parameters constant and simply add 0.1 ppm of analyte to the liquid, most of what you add ends up in the cup and the concentration of the analyte in the liquid is essentially unchanged - certainly within the error of the measurement.

This is a reasonable explanation as to why the overspikes are not working out as you'd expect. You have to add quite a bit more to the water to move the needle. 1 ppm addition gives a liquid concentration at equilibrium of 115 ppb (compared to 100 ppb with no analyte in the water).

Now I tried switching to 50% ethanol as simulator. Did pretty much everything else the same and this time there is no big difference when I spike into the cup or after transferring to the centrifuge tube but the spikes are still too small compared to the calibration. I really don't know what to do anymore, I feel like I'm going crazy with this.