Author: Marshall Schott
The primary purpose of each xBmt is to determine whether certain variables impact a beer such that blind participants are capable of reliably distinguishing a difference. To this end, we utilize the triangle test, a sensory analysis tool whereby blind participants are served 3 samples of beer, 2 from one batch and 1 from the other, then asked to select the one that is different. Because participants are unaware of the nature of the xBmt and asked to perform the same task, the triangle test is perfect for the type of “citizen science” we regularly perform; however, it is designed only to measure the perceptions of participants, whether they’re capable of distinguishing a difference, it cannot tell us whether there is actually a difference. Accomplishing such a task would require certain lab equipment and knowledge, neither of which I possess.
But I recently met someone who does.
A few days after I shared the results of the Pils malt boil length xBmt, I was contacted by a dude named Daniel Hillesheim who said he had the equipment and know-how to “quantify the concentration of DMS” in the 2 xBmt beers. Exciting! We’d already determined tasters were incapable of reliably distinguishing a Pils malt based beer boiled for 90 minutes from one boiled for only 30 minutes, but that left me wondering if the level of DMS in the short boiled beer may simply have been below the threshold of perception, or if there really wasn’t any present in either sample. After a brief back-and-forth, I sent samples from each batch to Daniel’s lab in Tennessee, packaging them in different sanitized yeast vials to avoid biasing him.
What follows is an explanation of Daniel’s process and the results he found, the details of which may be difficult for some to understand, but we wanted to ensure the methodology was made available for scrutiny. A simpler summary can be found at the end of this article for folks like me who get headaches trying to make sense of this stuff. Get your thinking caps on and prepare for a trip down chemistry lane, shit’s about to get deep.
| METHOD |
A 1µL sample was injected into an Agilent 7890 gas chromatograph (GC) equipped with a 30 m DB-Wax column and flame ionization detector (FID). Injector port temperature, oven ramp profile, and detector temperature were picked to optimize separation of the key analyte, dimethyl sulfide (DMS). Initial conditions were based on a Restek guide called Analyzing Alcoholic Beverages by Gas Chromatography. The injector temperature was set to 200°C, the detector temperature was set to 250°C, the oven profile began at 40°C for 1 minute, ramped to 150°C at 4°C/min, then finally ramped to 220°C at 10°C/min. The carrier gas is He at 45 cm/s column velocity. A split injection method was employed at a ratio of 5:1. Reagent grade DMS and ethanol (EtOH) were purchased from Sigma Aldrich. HPLC grade water was purchased from Fisher Scientific.
| RESULTS |
A standard consisting of 5µL of DMS in 5mL of HPLC grade water was prepared and injected to determine the retention time of DMS (1.72 minutes). Similarly a standard of EtOH was prepared and the retention time identified as 3.63 minutes. Data collection on the beer samples followed. Figure 1 shows a selected section of the graphs of the beer samples and the DMS sample in overlay. The large peak at ~4 minutes is EtOH. No detectable DMS was present in the beers.
To confirm that no matrix effects were responsible for shifting the DMS signal in the beer, 10µL of the DMS standard mixture was added to 5mL of the 90 minute boil sample, and that mixture was injected. Figure 2 clearly shows the appearance of a new peak with the same retention time as the DMS standard not present in the unadulterated 90 minute boil sample confirming that the peaks near DMS are other compounds.
| CONCLUSION |
Two beers of the same recipe consisting of 93% German Pils malt were produced simultaneously utilizing the exact same process with a single exception– one was boiled for 30 minutes while the other was boiled for 90 minutes. The impetus for this xBmt was to test the conventional wisdom regarding increased risk of dimethyl sulfide (DMS) in beers produced using a large portion of lower kilned Pils malt and the oft recommended longer boil for such worts. Results of the triangle test failed to achieve significance, indicating a general inability for participants to reliably distinguish between the different beers. However, as is common with most xBmt findings, the question of whether there was an actual difference in levels of DMS remained. A sample from each batch was sent to a lab for objective analysis, the results of which validated the results of the original xBmt: neither the 30 minute boil nor 90 minute boil samples contained measurable levels of DMS.
As was previously discussed, it’s possible the lack of DMS in these beers is a function of the high modification of the Pils malt used (Best Malz) and that less modified malts, such as those that are floor malted, might produce higher levels of DMS. But, it’s just as possible our understanding of the relationship between DMS and boil length is simply lacking, that our access to modern technology, higher quality ingredients, and better knowledge about brewing processes has reduced the likelihood of problems brewers of yore had to worry about. It’s these types of ideas I hope provoke the modern homebrewer to humbly question convention and embrace the thrill of treading new ground.
Support Brülosophy In Style!
All designs are available in various colors and sizes on Amazon!
Follow Brülosophy on:
FACEBOOK | TWITTER | INSTAGRAM
| Read More |
18 Ideas to Help Simplify Your Brew Day
7 Considerations for Making Better Homebrew
List of completed exBEERiments
How-to: Harvest yeast from starters
How-to: Make a lager in less than a month
| Good Deals |
Brand New 5 gallon ball lock kegs discounted to $75 at Adventures in Homebrewing
ThermoWorks Super-Fast Pocket Thermometer On Sale for $19 – $10 discount
Sale and Clearance Items at MoreBeer.com
If you enjoy this stuff and feel compelled to support Brulosophy.com, please check out the Support Us page for details on how you can very easily do so. Thanks!
41 thoughts on “Update: Lab Data on Pils Malt Boil Length exBEERiment”
awesome! i can do IBU analyses for you on some beers sometime if you want that.
Aaron, I’d love to talk to you about some IBU testing and a few methods and common lore I’d like to examine.
Please PM me or Marshall.
Please let me know how I can get that done as well. I’d like to compare the differences when using a brew bag or not.
Please let me know how I can get that done as well. I’d like to compare the differences when using a brew bag or not.
sorry, i missed this comment somehow. please get my info. from Marshall and we’ll see what we can do.
Great stuff! I’d love to see this xBmt tried with a less highly modified malt, or in an American lager with a high percentage of adjuncts.
And…my wife will thank you for taking 30 min out of my upcoming brewdays! From henceforth all pilsner based grain bills will be boiled for 60 min vice 90 (so I can get the hop additions in). Appreciate the exbeeriment and the scientific data supporting your hypothesis.
Prost,
Michael
Great data!!!
Also use Bestmalz. Kolbach index of Pilsen malt is 36-45%… moderato or high modified malt?
cheers
luis
I would say the understanding of how DMS is created is a pretty good one, it’s just that many people in homebrewing have never even bothered to look at the scientific literature and journals. There’s a lot to learn just from reading a Kunze or Narziss, or from going through the archives of scientific brewing journals. Regarding DMS production, there’s a pretty good article about it in the Journal of the Institute of Brewing: http://onlinelibrary.wiley.com/doi/10.1002/j.2050-0416.1977.tb03799.x/pdf (and probably many others)
It would be nice to know how much SMM was in the wort prior to the boil and after the boil. The amount of DMS in the beer could be a factor of how you boil and for how long.
it could also be how fast he chills, which is super fast.
Yeah, I wonder if this has a significant effect.
Fast chill vs. no chill xbmt?
Definitely on the list. I’ll probably knock it out with a standard 60 min boil first, just to be sure, but I’ll get to the combined variable soon.
I’d also be interested in how vigorous your boil was. I have a feeling the boil intensity goes a long way toward insuring you’re blasting it out of your kettle and not letting it accumulate.
Twice as much DMS standard was added to the beer spike sample as compared to the blank water sample, but the response appears to be much less. Doesn’t that suggest significant matrix interference and, therefore, a potential low bias in the result?
You’re not reading it or calculating it out. The DMS standard (DMS in water) was 1uL/mL. The “spiked” beer sample was 10 uL of the DMS standard (DMS in water) and is thus ~2 nanoL/mL, so it is actually 500X more concentrated DMS in the water than in the beer “spike.”
Sorry, I meant to say, “The ‘spiked’ beer sample was 10 uL of the DMS standard (DMS in water) in 5.010 mL total volume and is thus ~2 nanoL/ml…”
However, I would like to know two more pieces of info. What is the flavor perception threshold for DMS in beer and what is the detection limit of this analytical method. If the method cannot detect at the flavor perception or lower, then it is not useful. I think DMS can be detected much below the spike level introduced into the 90 minute beer sample in this analysis.
The perception threshold for DMS is reported as 30-50 ug/L per aroxa.com and their flavor standards.
BeerSmith reports 10-150 ppb.
One of my favorite sites is – Beer Sensory Science and he lists aprx 35 ppb.
https://beersensoryscience.wordpress/2011/06/02/corn/
Yeah I gotcha. Had to set up a spreadsheet to figure that one out.
Is there a potential for matrix interference to decrease the reported concentration in a beer sample? If so, don’t you need to know what the relative percent difference is between the actual spike concentration and the detected spike concentration in order to know if the non-detection is real or not?
“What is the flavor perception threshold for DMS in beer and what is the detection limit of this analytical method” – Good point.
OK, let’s say it’s 50 ppb. That is 2 orders of magnitude lower than the level of the DMS spike in the 90 minute beer sample! (unless I f’d up my calc’s)
Ralph, it’s best to dilute the sample, find another way to ovecome the matrix, or add in correction calculations to eliminate any matrix effects. It’s also possible to make a standard curve using standards prepped in a DMS-free beer matrix (though it can contain DMS but will raise the detection limit.) You are correct that any valid method must show there is not a high likelihood of matrix effects (suppression or enhancement of signal due to things other than DMS.)
I would just like to see a detection limit and a sample spike near the level of the flavor threshold to confirm that this method is capable of even telling the difference between a beer with detectable (by flavor) DMS and one without detectable (by flavor) DMS. The flavor spike should tell us whether the matrix is causing problems. It’s been a long time since I’ve done GCMS, so this may be trivial, and the author may have already thought of all of this and just didn’t post it.
The way I read it, the 1:1000 “DMS standard mixture” was prepared and run to get the reference peak – then 10uL of that prepared mixture was added to 5mL of the beer sample, so the peak would be ~0.2% of the size of the original. There’s no vertical scales on the graphs but eyeballing it – that looks plausible.
So if we take this data and the date from the Mash length Xbeerment, we’re looking at potentially shaving up to an hour of time off the brew process? I’m on board with that.
Maybe, maybe not. Boil length is a factor in hopping and in efficiency. I’m probably going to use 60 minute boils for a lot of beers for those reasons, as I almost always want 15-20 IBUs but don’t always want hop flavor or aroma and I get great boil-off at 7,000 feet above sea level.
Great point. Perhaps another Xbeerment to add to the list is to see if a 60min boil is required for complete bittering/isomerization. Unless it’s already known, it could very well be that you’ve fully isomerized and extracted all the bittering you’re going to get well before 60min. In which case you’d have to re-think (or at least re-scale) how you set up your hop schedule.
Wow! This blows the turkey wide open. What about boiling with the lid on? I don’t know why you’d want to, but if your busting myths anyways….
It’s on the list!
Anecdotal but I brewed the first 5 or so batches on my new system with the lid partially on (covering ~75% of the kettle) and there was a significant DMS taste. As soon as I pinpointed that was the off flavour I took the lid fully off and haven’t had the taste since. There were no other changes in my brewing process that I can think of.
My electric stove really sucks so I split my wort into two pots for 5 gallons batches and STILL can’t get a proper boil so I leave the lid on, let the boil start getting really active, then take the lid off letting built up steam out and condensation on the lid drip down on the counter and then put the lid back on as the boil dies down. Haven’t tasted anything like DMS yet.
Be sure to check out OldFart’s experiments on the brewboard eon’s ago – didn’t mess with pils but went so far as to condensate chiller on the brew point to try and keep everything in:
http://www.brewboard.com/index.php?showtopic=84807
I would also love to see whether doing a 30 or 60 minute whirlpool at the end of a 50-100% Pils grist would result in detectable DMS. I’ve been toying with using Pils in some IPAs/APAs to get more character but am worried my whirlpool will result in a DMS bomb.
But the human sensory threshold for DMS is in the range of 20ppb. So yes the FID could.not detect the presence of DMS in the sample but that doesn’t mean it isn’t there at a level that is perceptible. I believe an FID has about a 1 ppm sensitivity limit for DMS. The standard used was in the hundreds of PPM range.
Yes, I’m afraid these results are not very useful if we are trying to say that the levels can be quantified at the threshold level. I think a sulfur detector would be required on the GC in order to get into the ppb level detection limits:
http://www.perkinelmer.com/PDFs/downloads/FAR_GCBrewingQCApplications.pdf
My apologies if I have overlooked anything that makes my statements about the analysis untrue.
As mentioned above, without knowing the DMS precursor and free DMS levels in the wort pre-boil, the data presented here is not sufficient to say that a 30min boil will volatilize DMS to sub detection limits for all grain bills. (the DMS precursor being specifically a result of the malting/kilning process and malt variety.)
Also, to the question of the hopstand and DMS – without a boil to drive off DMS it will reform in solution, possibly to above detection thresholds. An approach to mitigate the DMS formation is to pre-chill the wort to 89C before the whirlpool.
If you can get access; the papers linked below, while not explicitly related to boil time and DMS, do have some relevant information regarding pre/post DMS levels as well as data regarding the pre-chill of the wort before whirlpool.
http://www.mbaa.com/publications/tq/tqPastIssues/2003/Abstracts/0226-01.htm
http://www.mbaa.com/publications/tq/tqPastIssues/2002/Abstracts/0916-06.htm
http://www.mbaa.com/publications/tq/tqPastIssues/1980/Abstracts/tq80ab17.htm
Jason.
You say “5 µL of DMS in 5 mL H2O”. What was the concentration of the DMS standard? Did you mean to say 5µg mL-1 (ie 5 ppm) in 5 mL H2O? That’s a very important detail. Injecting 1 µL of this versus 1 µL of unknown sample could be why there’s no response by the detector if the unknown is at or near taste thresholds- which are reported to be as low as about 20 ppb. This would produce a response at the detector by as much as 50 x lower! Cool stuff! Maybe ask the analyst to use matrix match standards for better precision/accuracy next analysis. Thanks for the info!
I’m guessing the DMS standard was near 100% or it is a very confusing analysis 🙂
Wow, these are perhaps the most sciency comments this simpleton’s mind has ever tried to process. I have read them a few times and still am unclear if the comment conclusions match the exbeeriment conclusions. I have also not figured out where that funny looking “u” is at on my keyboard or what the hell it means. I am guessing I was taught the meaning at one point (probably at the same time I learned that moles are not always on your skin) but I must have pickled those brain cells.
I think I need a beer….
Beer helps.
This is gonna sit in my craw for a while. Jamil claims the half life of DMS in congress mashes is 100 mins, Brad Smith says 40 min on his website, but you got rid of it in 30 min (or it never had much SMM to begin with, I suppose). Maybe Best has figured out how to kiln away SMM without over-kilning the husk. Are you going to repeat this with pilsner from other maltsers?
If you enjoyed and/or were surprised by this, you’re going to live some of the stuff we’ve got coming 🙂
Interesting. Did you also analyze for DMSO? Could it be possible that all the DMS was driven off but other sulfidic compounds remain behind?
Also, isn’t 5 µL in 5 mL 1 ppm (v/v)? Isn’t the odor threshold of DMS 1 ppb (w/w)? That’s 1000x the threshold, so I think you’re overloading the detector. Isn’t DMS traditionally analyzed with a headspace analyzer?
We recently finished a MTF wiki article that aimed at explaining why no-boil Berliner Weisse beers have mostly been reported to not have DMS issues. It might explain the results you got as well: http://www.milkthefunk.com/wiki/Dimethyl_Sulfide