The costs of running this huge site are paid for by ads. Please consider registering and becoming a Supporting Member for an ad-free experience. Thanks, ECF team.

Decomposition of VG to acrolein

Discussion in 'Experiments With Equipment' started by kinabaloo, Apr 26, 2009.

Thread Status:
Not open for further replies.
Image has been removed.
URL has been removed.
Email address has been removed.
Media has been removed.
  1. kinabaloo

    kinabaloo Vaping Master ECF Veteran

    Some time ago I had the idea of testing the possibility of the decomposition of VG to acrolein by vaping a juice that was only VG and distilled water. The build up of a deposit in this scenario does not prove acrolein is produced, but it strongly suggests it might be. Similarly, the absence of a deposit would not necessarily mean no acrolein was being produced as it is volatile; but as it is reactive one might expect some polymerisation or reaction products to give rise to a coil deposit after degrading on the coil.

    Hence, the test is also important as it relates to the deterioration of the atomizer's efficiency and demise; and on our understanding of how it might be cleaned.

    Some initial testing was carried out by SunVaporer (as yet unreported) and then by Exogenesis, the results of which, with photos, are posted here:

    A sizeable deposit was formed by 30ml of juice (90% VG, 10% water).

    Subsequent posts on the nature of the deposit from that thread:


    An interesting implication of this test is confirmation of the decomposition of glycerine to acrolein.

    The deposit will likely have been derived from polymerisation of some fraction of the acrolein gas immediately on formation and its subsequent degradation. This experiment shows that the amount is quite substantial; the acrid smell that is occasionally noticed is quite possibly due to released acrolein (acrolein means 'acrid oil'). What other volatile breakdown products are subsequently released during the decomposition of the various weights of polimerised acrolein is unknown. There is not much chemistry detailed for acrolein available on the web, but one interesting article is this one: Acrolein (HSG 67, 1991)

    Although polymerisation might not be occuring, and some other reactions and products implicated, that some acrolein is produced is in little doubt as the chemistry is clear. The issue is how little is it - and what might be done to minimise it (a smaller percentage of VG might help more than just the reduction of volume would suggest, as the boiling point of the mix would be lowered; there might be some chemical way to capture and render harmless any acrolein produced, although it would likely not catch it all; adding a small amount of alcohol to lower the b.p. further; etc.) Certainly acrolein is very reactive, which is why it can be dangerous. Acrolein is one of the (many) toxins produced by burning tobacco; possibly from glycerine added to keep the tobacco moist. A tiny environmental background exposure exists and it is found in some food products after cooking.

    This is basically the straightforward and definitive test that I suggested some time ago to see if glycerine would decompose in the atomizer. The experiment described above did indeed use an atomizer, but as far as i can tell one run constantly rather than intermittently, so further testing is advisable.

    The hope that VG might not decompose as it would be largely evaporated at a lower temperature by being only a part of the juice seems not to be the case; a certain amount of fractional distillation will take place and some molecules of VG will experience its normal boiling point (coincidentally also the temperature at which decomposition begins), or possibly higher; the temperature at evaporation of the molecules would follow something like a bell curve such that some part would reach decomposition temperature even though the boiling 'point' of the mix would be lower. Localised temperature variations in the coil and inner wick would also play a part.

    There is, however, no need to panic as the use of VG based juices seems generally well tolerated but it this issue nevertheless requires further study.

    In this article, the inventor mentions seven studies that found no harmful levels of by products: though this might not have been with VG-based juices.

    It is worth noting a similar bit of chemistry in high-temperature cooked foods regarding the formation of acrylamide.

    Exogenesis :

    Although I agree that it likely that some small amount of
    acrolein is being produced,
    I'm not sure, but I doubt if the bulk of the gunk is acrolein polymer based.
    I think the 'wet' gunk would have been significantly yellow & smelly
    if it was, which it wasn't.

    "Acrolein tends to polymerize when left at room temperature,
    leaving a gummy yellowish residue with a putrid odor"

    I was a little surprised that there was practically no acrid 'burning grease'
    smell to the very smoke-like 'vapour' produced, well maybe a little bit near the end.

    I agree that the 'ashed' gunk yellow colour is probably due to inorganic salts of tin,
    and / or other metals.
    I might test that to be sure, but the bits are really tiny making that difficult.

    The 'auto-vaper' rig had periodic 'off' rests, as well as occasional
    floodings, and even some short run-dry hot (not glowing) periods.
    So all-in-all I think the gunk result is pretty close to what you'd get
    doing it manually (Sun Vaporer may be able to confirm or deny that).

    Kinabaloo :

    Well, acrolein is the initial decomposition product. Some will polymerize and some and then be degraded by the at times intense heat of the coil. One would not expect to see acrolein polymers, but burnt/degraded products. The only other input in this situation is the air, and perhaps also tin if that is leeched with this juice; perhaps these play a part in the reactions from acrolein to gunk. (There will be a tiny amount of airborn contaminants too, such as dust and pollen)

    Perhaps some approximate calculations can be made. We know the amount of VG passed through (50% of the juice volume?) and can estimate the amount of gunk. Then estimate how much acrolein might result in that much gunk after degradation. The idea being to get a figure for what percentage of VG is decomposing.

    In a similar way, we might estimate the amount of acrolein that escapes as gas (some lighter polymers may also escape). We can assume that the concetration is generally below smell detection which puts it at most at about 0.05mg/m3. Most of the volume inhaled is just passed through air. Taking the volume increase at phase change of the two components and their densities into account would lead to an approximate maximum for normal-operation release of acrolein.

    It would also be interesting to try to capture any acrolein or polymers or other reaction products in the vapor output, perhaps by filter, condensation or bubbling through a reactant.

    Not sure to what the extent the decomposition (essentially dehydration) is reversible, with acrolein combining with water back to glycerine; perhaps not very likely.

    Exogenesis :

    I get the feeling you're assuming some amount of acrolein is present
    in the 'vapour' or near the coil.
    This hasn't been confirmed by analysis, and the only actual (slightly ambiguous)
    manufacturer's analysis statement I've seen was 'was not detected'.

    I wish I had my own GC/MS equipment to be able to make any definitive statement
    one way or the other, but I don't, so supposition one way or the other is a bit
    fruitless, unless some other significant evidence turns up.

    Kinabaloo :

    That's the chemistry: when heated to near its boiling point glycerine decomposes by dehydration to water and acrolein. The acrolein is therefore the starting material for the formation of the gunk seen in this test with no other constituent present (other than water). There is no other explanation on the table and it would be hard to imagine one.

    Released acrolein vapor might be significant in certain circumstances (coil running dry) but is generally indetectable by nose in normal operation. IMO some acrolein will always be produced by vaping a juice containing glycerine but the amount will be greatly reduced by being only a fraction of the juice; more than that fraction less as the other juice bulk ingredients (water and alcohol, and perhaps PG) reduce the general boiling point of the mixture in flash heating. But it cannot be eliminated; however, the amount might well be insignificant and give no short-term effects such as skin, lung and eye irritation. The argument about long-term exposure remains. It might be a level only comparable to background environmental exposure.

    It would seem the majority polymerises and degardes on the coil to form the part of the deposit. Hence, although it is still worthwhile to reduce the dry-deposit level of juices by consideration of the flavorings used, a juice that does not lead to any deposit is now clearly not possible when VG is used - it makes a considerable contribution.

    My understanding is that PG alone also causes a deposit (less in quantity but harder to remove according to testing by Sun).

    Exogenesis :

    What about oxidation 'carbonisation' and heat polymerisation of the glycerine itself....
    doesn't that seem as (or more) likely?
    28000 milligrams glycerine ----> say 20 milligrams gunk, so less than 1/1000 conversion.
  2. kinabaloo

    kinabaloo Vaping Master ECF Veteran

    The 1/1000 figure seems a reasonable guess given that glycerine would only be starting to decompose at normal (soaked as opposed to running dry) atomizer temp.s, but it might be a lot higher as the gunk is degraded (with volume loss) - and some of the products will have left the attomizer as gases. Nevertheless, 1 part in 1000 is still quite high - enough to be a concern if acrolein is produced - and for atomizer life, given the level of gunk formed foran amount of juice roughly equivalent to two weeks of vaping.

    Decomposition to acrolein:

    Interestingly, according to the source below the decomposition to acrolein and water is facilitated by the presence of metals (the coil and leeched solder tin perhaps).


    Oxidation of the glycerine is possible at elevated temperature and catalyzed by certain metals such as iron and copper, chemical oxidants (not present in the test) or electrolysis (interesting) to give a variety of end products with a 3-carbon chain (with a strong oxidant such as potassium permanganate it is oxidised to CO2 and water).

    Might play a small part.

    Polymerization of the VG itself to polyglycerols would require a catalyst; those I've seen mentioned are quite unusual, not a simple metal.

    Probably plays no part (but might occur a little bit in a more complex juice).

    Interesting source on glycerine: Glycerine: a key cosmetic ingredient - Google Book Search

    Some very approximate calculations ...
    One inhale (drag/vape) could be estimated at about 10ml (inhaled air/vapor). 0.07 mg/m3 of acrolein is smell detectable. So one vape would need only 0.7micrograms.
    One vape uses about 30,000mg/3000vapes=10mg of juice. About 1 in 10,000 glycerine molecules decomposed and emitted as acrolein. Might explain the 'bad smell' that can occur when the atomizer begins to run dry (hotter).
  3. Letzin Hale

    Letzin Hale Moved On

    Dec 28, 2008
    Intellicig had their liquid tested with a specific test for acrolein and none was found. Vaporiser temp was found to be up to 130C but test was done up to 300C. Report
  4. kinabaloo

    kinabaloo Vaping Master ECF Veteran

    That's good news of course. It might be that little acrolein gets out of the atomiser, mostly being polymerizing and degrading on the coil.

    We already know that the level of acrolein emitted as part of the vapor must be very small because it is seldom detected by nose, which is pretty sensitive to the strong odor of acrolein.

    However, the test probably didn't cover the situation of 'bad smell' as the coil runs dry. That's where the main danger would be, if there is one; the bad smell itself indicates that there is someting of concern in this. That smell might not be acrolein itself but a product of degradtion of the deposit (no matter how exactly formed); it is likely also not pleasant but once again it depends on quantity too.

    A temperature limited coil would be sensible, detecting by increased resistance that the coil is running dry (hotter).

    An alternative of course is a mist system using a piezo transducer. The drawback is that a mist is noticeably cooler when inhaled. Heating the tiny amount of fluid before it is misted would be tricky (heating more would be wasteful and perhaps impossible power-wise). Thinking about this I have devised a hybrid design ...

    The VP would be filled with a juice and with distilled water. The water would be vaporised as now while the VG/PG, flavorings and nicotine (if included) would be misted and the two inhaled as one.

    Edit (add): the report cited mentions testing a sample of smoke. This might be too small amount to detect acrolein that would be measured in ppm. It would make more sense to condense the smoke for an hour or two and test the resulting liquid.

    Impurities the reason for the deposit?

    The report cited also stated the glycerine to be 99.1% pure. If the rest is not water (typically glycerine is 10% water as it is very hydroscopic), could that remainining fraction account for part of the deposit?

    Part of the USP grade testing reads: "When a small portion of glycerin is heated to boiling in an open porcelain or platinum capsule, and then gently ignited, it should burn and vaporize so as to leave not more than a dark stain (absence of dextrin and sugar), which would leave a bulky, difficultly combustible, charred mass); and on full combustion no residue whatever should be left (absence of fixed impurities)." Quoted here:

    I wonder if the impurities, perhaps sugars, might be a major source of the deposit ...
  5. ChainSmkr

    ChainSmkr Senior Member ECF Veteran

    Jan 22, 2009
    Would PG have the same results?

    "Acrolein is used in the preparation of polyester resin, polyurethane, propylene glycol, acrylic acid, acrylonitrile, and glycerol"

    quote from Wikipedia
  6. kinabaloo

    kinabaloo Vaping Master ECF Veteran

    A quick search yielded decomposition temperature for PG as >350C far higher than its boiling point of 188C so although the decomposition products are hazardous, it is extremely unlikely to occur at all in an atomizer.

    In contrast, VG begins to decompose before reaching its boiling point.
  7. jennydotz

    jennydotz Senior Member ECF Veteran

    Feb 14, 2009
    West Virginia, USA
    Can anyone put this glycerin fear in perspective? What do you think is the bottom line here? Fear of acrolein/VG is the reason I am waffling and keep going back to analogs and I can't vape PG.

    Any ideas about comparing the analog risks to the VG risks? I know it's a great unknown, but what are your thoughts? Is acrolein DEFINITELY present in VG vapor or possibly? If it is definitely in there is it a sure health risk or a potential risk?
  8. kinabaloo

    kinabaloo Vaping Master ECF Veteran

    jennydotz - it's just a possibility with vaping VG, but a certainty with analogs, not to mention the many other toxins in analogs. Analogs contain far more chemicals and are combusted; so safety-wise there is no contest. Acrolein occurs in the environment (air), foods such as wine and can be created by cooking. In all cases, as with vaping VG, the quanity is tiny. There are so many other things to worry about too, such as all the contaminants in tap water.
  9. jangler

    jangler Senior Member ECF Veteran

    Apr 2, 2009
    Las Vegas NV
    Hi, total noobie here. So please forgive my ignorance. I've been trying to follow all these posts, but some are quite long and hard to follow. Just finished reading the "cola method bringing it to the next level" thread. Very informative material.

    I'm actually considering to order one of the mods (well I ordered one but won't get it 'til august) but from reading the first post, are these mods then not such a good idea due to it's possibility to run hotter? I do notice when I cut with VG I get a kind of waxy coating in my mouth. Is that what you mean by "polymerization" (to a small degree maybe)? If this is so would it be "safer" to use straight PG when using these mods.

    I understand that analogs are 100 times worse than ecig and I don't intend to go back but as the previous poster said, it would be good to put this in perspective. Information such at this also serves to "ground" me somewhat... and not have a false sense that this is absolutely safe and throw caution to the wind...
  10. kinabaloo

    kinabaloo Vaping Master ECF Veteran

    jangler - the taste you mention is probably just recondensed VG. Whether running at a higher voltage might increase the chance of acrolein being produced is hard to say; on the face of it one would think so, but it depends on a number of factors. For now, I would say don't worry about it unless you detect a burning taste more often or moreso than with a standard e-cig.

    It might be wise to keep the VG content under 50%, at least not more than 50% as this should reduce the chances of decomposition.
  11. jangler

    jangler Senior Member ECF Veteran

    Apr 2, 2009
    Las Vegas NV
    Thanks Kinabaloo, so one of the warning signs would be the burning taste? I have experienced that, especially with primer liquid found in new atty and every now and then during when the atty starts to feel hot, but that could just be the batting. (not that this is any safer)…

    I don’t cut more that 50% VG, more like 10% just for increased vapor. So I guess I’m mixing well within safe parameters? I wonder if there are other forum members who could share their experience with their mods using VP, whether they experience burnt taste more, less or nothing out of the ordinary. I’m assuming the mods do run hotter?

    Thank you for guidance and everybody’s contribution. This forum is priceless.
  12. kinabaloo

    kinabaloo Vaping Master ECF Veteran

    Well, I'd say an analog tastes much worse than the 'bad smell' ;) If not a menthol anyway.

    Whatever the cause, the bad smell is not a good sign; but it doesn't keep me awake at night.

    The mods primarily aim to address longer battery life. Then style. Then sometimes more vapor.

    I know there are acrylamides formed whem i toast a piece of bread, but I still eat it. Maybe I should eat something else instead, but I am going to die anyway one day and I can't prevent it no matter what, so I don't want to spend the time I have worrying too much, especially as stress is a certainly bad for health.

    Everything has its pros and cons - even spinach (or anything else one could mention, even exercise); it's all about moderation and variety ...
  13. Sun Vaporer

    Sun Vaporer Moved On ECF Veteran

    Jan 2, 2009

    Letzin--I saw this as well and it confirms for me that there is no issue with acrolein---Sun
  14. exogenesis

    exogenesis Super Member ECF Veteran

    Mar 1, 2009
    While not unduly concerned about acrolein during 'normal' vaping of VG,
    the 'burnt taste' due to a dry hot coil is something to think about.

    My problem with these reports is they are attempting to detect a very
    small amount of acrolein in a vast amount of potentialy obscuring bulk
    liquid/vapour (VG).

    i.e. the far easier test for a small amount of just acrolein
    (in say just air or water) would probably give definitive quantitative results.

    I think that leads to the slight ambiguity of phrasing like:
    Which does not exactly state that there were no acrolein peaks seen
    in GC/MS & other test-methods,
    just that possibly they could not 'detect' acrolein in the melee of signal
    from the bulk liquid (i.e. signal noise may swamp the small target peak).

    If that could be clarified by any of the testers, it would be somewhat useful.

    There is a colorimetric acrolein-specific 'wet test' I've seen,
    but not sure how sensitive it is,
    I possibly might get the reagents & investigate a bit further.
  15. surbitonPete

    surbitonPete Ultra Member ECF Veteran

    Jan 25, 2009
    North Yorkshire UK
    I don't have the greatest sense of smell anyway but for some reason I am convinced it's probably something to do with acronolyn when I get that really horrible chip fat 'taste' and that usually only happens with a gunky coil that I have vaped a bit too dry. It never happens with a fairly new coil. I doubt if their tests recreated those conditions.
  16. ChainSmkr

    ChainSmkr Senior Member ECF Veteran

    Jan 22, 2009
    Oh yeah! I think I do remember that from another thread. Thanks kinabaloo.
  17. Sun Vaporer

    Sun Vaporer Moved On ECF Veteran

    Jan 2, 2009

    I get that same taste with PG--so where is the corellation to Acrolein? Or are you trying to say that PG will create acolein as well. I simply do not see it. Again the report just out showed no trace of acrolein while using VG so I would like to see a report to the contray before I would give this contention merit IMO-------Sun
  18. exogenesis

    exogenesis Super Member ECF Veteran

    Mar 1, 2009
    No, I'm not in the 'assuming there is acrolein' camp,
    & PG won't decompose to acrolein as far as I know.

    Good point though about the burnt taste with PG, but are you talking about:
    pure PG (as in your testing),
    or a commercial PG-only e-liquid,
    or a 'normal' commercial PG e-liquid (that's got quite a lot of VG in it) ?

    Actually I'd assume that you'd expect some 'burnt' taste with practically any organic
    liquid passed over a very hot coil, really the taste is not the important factor.
    But it obviously has connotations here :
    - perceived risk & all that (whether erroneous or not),
    especially with all this communities 'attention' on acrolein,
    even though other components (in less 'pure' e-liquids) might be
    just as worthy.

    It's very good that these tests are being done, so there's no
    'take the word of the manufacturer' distrust going on,
    kudos to them for having the tests done.
    But with that comes the whole reason for the testing
    - to attempt to clearly show absence/presence/amounts of the components being tested for.
    Otherwise what's the point?

    This testing was done aiming at the major ingredient components (glycerine/nicotine/flavour),
    but there's no specific test to confirm/deny acrolein presence in there.

    In the report it would have been good to see a GC/MS
    calibration-standard for acrolein, so you could point to a place on the
    'Smoked Cartridge' ' test's GC (& MS) traces & be able to say 'definately no acrolein peak'.

    Unfortunately that 'Smoked Cartridge' GC trace was possibly somewhat
    too noisy & complex for that anyway?
    But it shows there was certainly a substantial signal for components between
    the clear argon and glycerol/glycerol-like peaks.

    This signal could be obscuring, or possibly even be acrolein
    or other acrolein-like decomposition products.
    And the MS traces for this sample are not given, either MS was not done on
    this sample, or they were not presented due to being too confused/difficult to interpret.

    Realisticaly testing aimed at specifically measuring acrolein would be needed before you
    can dismiss (or confirm) this contention. It's still a possibility.
    Personally I'd prefer the facts rather than supposition either way.

    But would the manufacturer release that info, if it was confirmed?

    Tough choice - pay for an expensive test to potentially show a problem
    with your product.

    Then again if that testing showed zero, or below acceptable levels,
    then everyone's happy.
  19. kinabaloo

    kinabaloo Vaping Master ECF Veteran

    Acetaldehyde is another potential toxic product of either VG or PG; unlikely but possible. This one actually has a nice(ish) smell (something like apple). Although harmful (possibly carcinogenic even) it also can be found in trace amounts in food when it begins to go bad, and is a metabolite of alcohol in the body (the chemical responsible for the hangover, apart from general dehydration).

    My concern is not only about acrolein. And I'm not in a 'camp' or have any agenda; just scientific curiosity.

    This interesting paper concerns additives of PG and VG to tobacco. Although tobacco is combusted in a cigarette rather than heated, much of the PG or VG would undergo evaporation as the heat approached it, before being combusted. It found no acrolein with PG, but both did lead to nasties such as acetaldehyde and formaldehyde (though possibly by combustion and be irrelevant in vaping; not sure):

    It gives a figure of 1% conversion of VG to acrolein in the cigarette; a pretty high value. I would expect the figure to be (perhaps far) lower in the vaping scenario.

    If pure PG plus water leaves a deposit, as apparently it does, this is quite a puzzle. Seems UPC grade VG and PG would contain nothing other than water measurable in quanties over ppm so that explanation looks unlikely. Understanding PG in this regard might shed an alternative light on the deposit formation of VG, but at present the mechanism is really obscure if the acrolein route is not the primary route.

    All I can suggest for PG (decomposition/reaction) leaving a deposit is that this is catalysed by solder-leeched tin, or by the chromium or nickel of the coil, making something that would not normally be expected occur.


    ps: Comparison of grades of glycerine
  20. exogenesis

    exogenesis Super Member ECF Veteran

    Mar 1, 2009
    Re-read that acrolein from glycerol in tobacco smoke,
    (& even though it's from Phillip Morris) it's quite enlightening.

    Having second thoughts about
    'glycerol just being like any burning organic liquid on a hot coil',
    specifically probably similar to PG or other carriers...

    cos just now observing pure PG vaping on a fresh exposed coil (550mA)
    & it seem's to be producing no burnt gunk whatsoever.

    We'll see what it's like after 20 to 30 ml.

    Did SunVaporer post his VG vs PG results?
Thread Status:
Not open for further replies.

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice