Eliminating The Helium Penalty
We wanted to preface this article with a statement from Dr. Simon Mitchell about the message of this post. We feel it summarizes the important takeaways from this article and captures where Shearwater stands on this issue:
“The obvious question is whether this means that we can all tell our computers that we are diving air diluent when we are really diving trimix. The study does not answer that question unfortunately. Although it suggests that there is no evidence for treating helium and nitrogen differently (as reported by Doolette in animals) it is conceivable that the long held belief that helium needs more decompression has compensated for underestimation of required decompression (to achieve acceptable levels of risk) by the decompression algorithms we all use, and that we are doing the right amount of deco but probably for the wrong reason."
- Dr. Simon Mitchell
Our understanding of decompression science has undergone dramatic revision in the last two years thanks to decompression physiologist David Doolette, Ph.D. and the team at the Navy Experimental Diving Unit (NEDU). Their report, “Decompression from He-N2-O2 (TRIMIX) Bounce Dives Is Not More Efficient Than He-O2 (HELIOX) Bounce Dives,” published May, 2015, dispels a belief about mixed gas diving that has persisted since the NEDU developed the first working heliox tables nearly 80 years ago.
Their findings: A bounce diver’s decompression requirements depend solely upon the time, depth and level of oxygen (PO2) over the course of the dive regardless of the fraction of helium and or nitrogen used in the breathing mix. In other words the so-called “helium penalty,” i.e. the extra stops and decompression time required when breathing helium mixes on a surface-to-surface bounce dive, does not exist.
The corollary? Virtually all of the existing tables, dive computers and deco software, used by technical divers today, particularly the Buhlmann ZH-L16 algorithm which fixes nitrogen halftimes as 2.65 times longer than those of helium, have a legacy bias that adds increasing decompression time the higher the fraction of helium in the mix.
The trouble is that while this helium penalty is fictitious, the “depth penalty,” i.e. most algorithms prescribe increasing risky schedules the deeper and longer the dive, is not. However there is no reliable way, a priori, to determine which is which, that is to know which schedules can be reduced and by how much.
How Did We Get Here?
The extra decompression time calculated by various algorithms when breathing a helium mix is a consequence of the long held belief that helium, which is lighter than air, enjoys faster uptake by the body than nitrogen (in the case of the Buhlmann algorithm 2.65 times faster). As a result, current algorithms prescribe a deeper first stop (to clear the helium supersaturation in fast compartments representing tissue like brain or spinal cord) resulting in continued gas uptake in slower compartments and consequently increased total decompression time.
However, there have been some indications that the helium penalty was fictitious. Over the last two decades Global Underwater Explorers’ (GUE) divers have conducted record long cave dives at Wakulla Springs and its connecting systems as part of the Woodsville Karst Plain Project (WKPP). CEO and founder Jarod Jablonski told me they discovered that their required decompression was largely independent of the helium fraction used. Other tech divers have also noticed the apparent discrepancy; some trick their dive computers by inputting a lower helium fraction than that in the trimix they are actually breathing in order to reduce their decompression time. Note that this is not a recommended practice for the reasons cited above.
But the real clue was an animal study that Doolette, a cave explorer and member of the WKPP, conducted in the early 2000s at the University of Adelaide prior to coming to NEDU. The resulting paper, “Altering blood flow does not reveal differences between nitrogen and helium kinetics in brain or in skeletal miracle in sheep,” which wasn’t published until December 2014, found that there were no differences in nitrogen and helium kinetics i.e. uptake and elimination in the ranges studied. Note that the typo in the title, “skeletal miracle,” which should have been “skeletal muscle,” was discovered miraculously after the paper’s publication much to the chagrin of the authors. The paper concluded that it was inappropriate to assign different time constants for nitrogen and helium in all but the slowest compartments used in a decompression algorithm.
Putting Helium To The Test
In recent years, the Canadian Navy and others created trimix research programs, in part due to increasing helium costs, and invited the U.S. Navy to participate. The U.S. Navy has always used heliox as a diluent for mixed gas rebreather diving; it does not conduct open circuit mixed gas scuba diving. Doolette, his colleague Wayne Gerth, Ph.D., the head of NEDU’s decompression team, and Keith Gault. convinced their sponsors that a trimix program would only make sense if trimix offered significantly reduced decompression times over heliox, a claim that had never been tested.
The few human experiments comparing nitrox and heliox dives to the same depth were conflicting and did not provide compelling evidence of a difference in decompression requirements. There have been other studies involving trimix, but none were designed to demonstrate a real difference in efficiency between trimix and heliox. Doolette and company designed their experiment accordingly.
Using their Linear Exponential Multigas (LEM) probabilistic decompression model that NEDU had developed for assessing decompression sickness (DCS) risk, the team selected a profile—200 fsw, 40 minute bottom time—that represented the largest difference in estimated probabilities of DCS between trimix and heliox among a range of candidate profiles that were both practical to man-test and operationally relevant i.e. from about 20-60 minutes of bottom time in the 150-300 fsw range. The experiment, which was conducted in NEDU’s Ocean Simulation Facility (OSF), was carried out as follows.
Each of the volunteer divers in the experiment was equipped with a Navy MK-16 mixed gas rebreather and full-face mask. The rebreathers were charged with either trimix 12/44 (12% oxygen, 44% helium, 44% nitrogen) or heliox 12/88 (12% oxygen, 88% helium) diluent gases. The divers then submerged and were pressed to 200 fsw, where they completed the 40-minute wet-pot dive pedaling cycle ergometers.
At the end of 40 minutes, the divers were decompressed for 119 minutes according to the Mk-16 constant PO2 trimix table (see attached), which was not only 15 minutes shorter than the corresponding schedule for heliox but allowed an ascent to the first decompression stop at 70 feet in contrast to 90 feet for heliox. After surfacing the divers were monitored for signs and symptoms of decompression sickness (DCS).
According to the LEM and its associated he8n25 database which consists of more than four thousand air, nitrox, heliox and a few trimix dives, divers with the trimix diluent had a 2.14% probability of getting bent, while the heliox breathers had a 5.56% chance, more than double that of trimix. If their decompression model was correct, that is, if decompression from trimix was more “efficient” than heliox for bounce dives, Doolette et al would expect to see a higher incidence of DCS on the heliox dives in the study compared with trimix dives.
The results? Over about a ten-week period, a total of 32 volunteers conducted 50 heliox dives without incident, and 46 trimix dives with two diagnosed cases of DCS, at which point the scientists stopped the experiment. Statistically that means that researchers’ Null hypothesis was retained: trimix decompression is not more efficient than heliox. It also means that existing algorithms, including the U.S. Navy’s have a legacy bias and need revision.
According to Doolette, though they only tested a single profile, at least in part due to the expense, it is unlikely that testing another profile in the relevant range would have yielded a different outcome because of the way decompression algorithms work. The results should also apply to open circuit diving.
The NEDU is now working to revise the Navy’s algorithms to reflect their new findings. This will likely involve adjusting model halftimes for nitrogen and helium for compartments out to 100-200 minutes or more, which Doolette said was a slow process. “All of decompression science is like this. There is rarely a chance to self-correct quickly as seen in this case. It’s taken a long time for Buhlmann’s gas concepts to be corrected,” he said.
So what does this all mean for technical divers?
In the big picture it means that technical divers will likely adopt heliox, which is now arguably superior to trimix, as the preferred diluent for rebreather diving, while deep open circuit dives (Is that still a thing?) will likely stick with trimix because of the cost of helium. Interestingly, it’s a prediction that prescient explorer cum engineer Bill Stone, Ph.D. offered in his January 1992 article, “The Case for Heliox: A Matter of Narcosis and Economics,” that appeared in aquaCORPS #4 MIX.
“In summary,” Stone wrote, “the heliox versus trimix debate presently rests with the cost trade-offs inherent in the inefficiencies of open circuit scuba at depth. Eventually as [rebreather] technology becomes widely available, the debate will come to an end and heliox will become the bottom mix of choice for deep diving.” And that was before the helium penalty was found to be nonexistent. At that time, helium cost about $.22/cf3 compared to about $0.85-$1.85/cf3 today depending on location.
Of course, the more pressing issue is, “What algorithm will tekkies use to make their heliox decompression calculations?” The Navy’s current Mk-16 MOD 1 tables, which are published in the U.S. Navy diving manual, are one possibility, but as noted above the current version contains a helium penalty. So tekkies will have to wait for NEDU or others like Duke University to publish a corrected algorithm, and then wait for dive computer manufacturers to implement them.
Unfortunately, that won’t happen overnight. “Because of funding cycles, it will likely take two to three years, for this work to hit the public domain,” Doolette explained.—Michael Menduno
 In contrast, it has been proven that decompression from a heliox saturation dive requires less time than from a nitrox saturation dive because nitrogen takes longer to outgas from very slow tissue compartments that control decompression in saturation dives.
 Note that heliox is superior to trimix in other ways for rebreather diving i.e. there’s zero narcosis, lower density/work of breathing and helium’s chilling thermal properties are not an issue in rebreather diving. One possible reason to use nitrogen in one’s diluent is to ameliorate possible High Pressure Nervous Syndrome (HPNS) for very deep dives, generally beyond 500 f/150 m.
 The U.S. Navy MK-16 MOD 1 rebreather tables are ISO-risk tables meaning that the probability of getting bent is the same for all schedules and is set at about 2%. In most tables, and dive computer schedules, the risk, which is not explicitly identified increases proportionally to the decompression times. In addition most tables have a “sweet spot” where the risk is minimal.
 Stone’s team dived open circuit heliox during his paradigm shifting 1987 Wakulla Springs project, which was sponsored by Air Products.
Leave a comment
Upcoming Dive Shows
Shearwater Research will be at the following dive shows:
Really interesting,as a matter of facts I’ve been asking myself since I started diving deep with my Hammerhead reb the reason why I had to take a bit of narcosis due to the nitrogen in my mix ( typically 10/70) since my dive were max to 120 m .
I think it’s important to fully understand the conclusion and the further hypotheses the conclusion might bring. The results found that trimix decompression for this particular profile was “not more efficient than” heliox. Logically, that means either they are equally as good (or equally as bad) at decompression for this profile, or heliox is better at decompression than trimix for this profile. If the latter is true then the helium penalty” might be bigger than we all thought it was initially because it might actually introduce less deco time on heliox than trimix would and not the same amount of deco time as trimix, and certainly not more deco time as trimix as it does today.
I am a old hard hat diver and it amazes me how far we have improved in the study and use of mixture breathing and the reading of “Eliminating of the Helium Penalty” and how well it is explained. Most of my diving career we used Professor Haldane’s tables.
Nice one M2.
😉 Thanks Chris!
Another reference suggesting decompression on heliox might be better than on air: “Effect of hypobaric air, oxygen, heliox (50:50), or heliox (80:20) breathing on air bubbles in adipose tissue.” https://www.ncbi.nlm.nih.gov/pubmed/17600159
Apologies if stupid question; but is there not still some benefit in having some n2 in the mix for ICD issues when bailing out or surfacing? Great article by the way.
There are no stupid questions. However, I am not exactly sure what you mean. On OC, switching from trimix or helix mix to a Nitrox mix for deco maximizes the He gradient and therefore speeds up off gassing of He. is that what you are asking about?
So, what about OC trimix diving? Does this mean that we would have safe, but shorter, ascents when diving with trimix by setting our computers to the correct FO2, but leaving the FHe set to 0, even if we’re actually breathing trimix?
While it appears that one could “assume” the article (very informative and well written btw) suggests this, I do not think that it was intended to state the assumption, as such. Unless I misunderstood the article, I believe the intended profile and body type would still be variables that differentiate deco obligation relative to the H2 percentage. A longer time with a higher body mass index would theoretically increase the relative saturation level in the slowest of tissue compartments of H2 when compared to a diver with low body mass index. While the assumption may prove valid for a broad range of profile time & depths, I think it would be a prudent approach to wait on some additional experiments to verify that H2 percentage can be substituded as N2 in deco planning and/or trimix computers or at least determine what profile brackets this can be applied to. This of course is just MHO
That’s an interesting point RJC. You are suggesting different body types will respond differently to a given deco profile. i.e. that someday, we might be able to set deco based on the individual body type and state. The science is not there yet as far as I know. BTW, it’s He and not H2. Helium molecules are made up of single atom unlike oxygen molecules formed by 2 O atoms.
That’s the right idea, however, you don’t know if that extra time is needed for the dive (not because of helium but because of depth/time. That’s what I meant by depth penalty.
The fact that they had 2 DCS cases in 46 dives with trimix and 0 DCS cases in 50 dives with triox is an interesting hint that perhaps there needs to be a “helium accelleration” of deco rather than a penalty. That fits anecdotes I know of where 21% divers have suffered DCS when diving with 21/35 divers on the same profile. Their results are clearly not statistically significant but it would have been interesting if they would have been able to continue the study until they reached the point where it was.
Interesting food for thought. I look forward to additional researchers publishing their findings.
Beautifully written, Michael
Several years ago on CCREXPLORERS reminded Jones of the decompression model.
Quote from “The Basics of Diving Pathophysiology” Script for Students and Physicians, Military Academy of Medicine, Department of Marine Medicine 1971 y. Page 134, 3, para:
“To confirm that diffusion is not a limiting factor for tissue saturation, Jones has presented research results that indicate the rate of leaching from other human gases, namely helium, krypton, or radioactive xenon.
The results compared with the values obtained for nitrogen indicate that the flushing rate of these gases is equal to half-life saturation with respect to water tissues. The half-life of saturated fat is different due to the different values of the partition coefficient.
The results of the study published by Kety would also support the validity of such a position, at least for well perfused tissues. “
I believe the misconception of a helium penalty came from the speed with which helium was absorbed into the tissues (also why decompression is most likely faster than with nitrogen). I understand it as requiring less bottom time/time on bottom gas to gain additional decompression (which some find to be a penalty).
This supports the hypothesis of having N2 during decompression at a point where helium leaves the tissue faster than nitrogen is absorbed, oxygen still poses a toxicity problem, and narcosis is controlled while less task saturated.
The Theoretical Diver aka Robert Helling also wrote about this, investigating deeper where the Helium penalty in Buhlmann’s model actually comes from: https://thetheoreticaldiver.org/wordpress/index.php/2017/11/02/n2-vs-he-whats-the-difference/
I found that post to be a very interesting follow-up read.
To quote one part from his post:
> To me it sounds as if it is this combining of the gases is actually the part of the Bühlmann model with the weakest justification from a physics point of view. And my interpretation of the NEDU study is that it supports doubts that this is the right thing to do.
Interestingly, VPM is quite different here. All the compartments in VPM use the same parameters and limit values, making the effect of the Helium penalty less severe. In VPM, if I remember correctly, Helium and Nitrogen partial pressures are still being combined within one compartment in the bubble calculation, and just like Buhlmann, VPM assumes that for a given compartment, Helium on- and off-gassing rates are significantly faster than those for Nitrogen. So with VPM, there can be a Helium penalty in some cases, but the interaction is a lot less obvious. I wouldn’t be surprised if some profiles would generate shorter decompression schedules with Heliox than with Trimix in the VPM.
Thank you, Michael, for the interesting article. It has also been quite interesting to follow the thoughts given here in – thank you, all.
Reading all above, makes me thinking of if there are more usefull studies/models/tables for using heliox as bottom gas, turning to nitrox during ascent/decompression, and finally using pure oxygen at the end, for an accelerated outgasing?
Is there any experiance of it here in?
Thanks Daniel! Sorry I didn’t reply earlier. I just saw this. I will read the article with interest. Thanks!!
Thanks Thomas! I find this stuff fascinating. Of course the US Navy uses heliox in their rebreathers (and with helmet divers, who use nitrox blends (air, 50/50) and O2 for deco. Check out this story on commercial mix dives, which I believe are similar to those conducted by USN: http://www.alertdiver.com/Commercial_Mixed-Gas_Dives
As far as tekkies, what you propose is what the ever prescient Bill Stone did at Wakulla 1987. He wrote an article for aquaCORPS titled, “The Case for heliox” back in 1992-I could send it or post it, arguing that we forgo trimix and just get on with heliox. Interestingly, given tech divers super rapid descents (compared to commercial divers), they may need trimix (N2) to buffer them from High pressure Nervous Syndrome (HPNS) for dives in the 330-500 ft plus range. (Note Descent rates likely much faster on wreck dives than caves-depending on the system of course.) Some deep divers, like Richard Pyle, have told me they are very sensitive to pure heliox vs trimix. It makes him feel too edgy down deep. Feel free to reach out if your interested in Stone article. My address is: email@example.com
Really nice article Michael. Different body types would have different body fat vs muscle ratio, and of course hydration levels(tissue water content). Most decompression models would cater for an average, wouldn’t they? Would that be available for individuals to make a personal comparison?
It is similar to the commercial companies use of Heliox, (substituting 20/80 for Air), USN Table 6 treatment tables for DCI after surfacing after Saturation. But they don’t appear to advocate any extensions at either the 18 msw (60 fsw) or 9 msw (30 fsw) stages, whether that is due to the worry of a Helium penalty, or I think more likely the thought “if not improving try a deeper table”, I don’t know.