top of page

8.16 Medusa Diamond & Fast Crash Syndrome


Solvent Direct Butane, Dry Ice, Ethanol, Carbon Chemistry, Media Bros, Lustermax, Molecular Sieve, THC-A Diamonds

Medusa Diamond & Fast Crash Syndrome

An ongoing discussion of Team Medusa’s diamond and fast crash syndrome remediation efforts. To be updated as new discoveries are revealed:

Alas, Buddha observed that life was painful because of desire. Pain from not having what we desire and from having what we don’t desire, but the most painful of all is having what we do desire, because everything is fixed in time and given enough time, we will lose it.

Medusa diamonds and fast crash phenomena is a case in point. We were all comfortable with our LPG extraction and processing techniques and suddenly some of them didn’t work anymore.

Gorgeous monocrystals of THC-A and CBD started turning to chalk overnight, and sometimes the sauce crashed so fast it only had time to precipitate out sand size crystals

We still don’t know for sure what causes it, but using column chromatography experimenters have removed whatever it is from the LPG and the problem goes away.

An analysis of the column removing what ever it is causing the problem, demonstrated enrichment of the oxygenate and nitrogen adsorbing elements in the column, suggesting it is something in the LPG and most probably methanol and diethanolamine.

I say most probably because that is what the discoverer reported in his sample, but reviewing papers on methanol and amine addition to sour crude, there are other amines in use as well.

Ethanolamine is used about 20% of the time, diethanolamine about 20 to 25%, Methyldiethanolamine about 30 to 55%, and diglycolamine about 50% of the time to remove H2S and CO2.

Another line of thought believes the issue is terpenes in the LPG from extraction and it is hard to discount the effects of the mono and sesquiterpenes, which are all alcohols, ethers, ketones, aldehydes, esters, etc, et al.

If so, ostensibly more of a buildup of terpenes in the LPG, as they have always been present in “the sauce” from the LPG extracting plant material and Medusa is something that reared its ugly head recently.

And if so, seemingly the same remediation used to remove amines, is also effective on terpenes, so that will hopefully give us enough breathing room to sort out all the players.

A group of us formed a team to address Medusa issues, composed of LPG suppliers, media suppliers, extraction labs, and consultants, to share information and promote a solution.

Available papers on methanol and amine removal in the gas industry reveal that they are both added to sour crude to remove the CO2 and sulfur compounds and are subsequently mostly removed using Liquid Liquid Extraction (LLE) and are polished using column chromatography, typically utilizing mol sieve beads or ion exchange beads.

I sent out inquiries to the major zeolite (Aluminosilicate) Molecular Sieve and Ion Exchange Bead suppliers for their recommendations, and the results varied from no response to good advice, with some in the middle with bad advice, suggesting that in some cases tech support and sales are one and the same.

Since the proper mol sieve bead for removing constituents depend on matching the molecule size to the bead pore size, I developed the following numbers to facilitate that search, using the following linki:

Molecular weight to size calculator:

Solvent Direct Butane, Dry Ice, Ethanol, Carbon Chemistry, Media Bros, Lustermax, Molecular Sieve, THC-A Diamonds

A review of the manufacturer’s specification sheets for standard mol sieve list the following applications:

3A has a 3-angstrom nominal pore and shows typical applications as:

(a) Drying of unsaturated hydrocarbons ((e.g. ethylene, propylene, butadiene).

(b) Cracked gas drying.

(c) Drying of natural gas, if COS minimization is essential, or a minimum co-adsorption of hydrocarbons is required.

(d) Drying of highly polar compounds, such as methanol and ethanol.

(e) Drying of liquid alcohol, (f) Static, (non-regenerative) dehydration of insulation glass units, whether air filled or gas filled.

(g) Drying of DNG.

4A has a 4 Angstrom nominal pore size and list typical applications as:

(a) Drying and removing of CO2 from natural gas, LPG, air, inert and atmospheric gases, etc).

(b) Removal of hydrocarbons, ammonia, and methanol from gas streams (ammonia svn gas treating).

(c) Special types are used in the air brake units of buses, trucks, and locomotives.

(d) Packed in small bags, it may be used simply as a packaging desiccant.

5A has a 5 Angstrom nominal pore size and lists the following applications:

(a) The strong ionic forces of the divalent calcium ion make it an excellent adsorbent to remove water, CO2, H2S from sour natural gas streams, while mini missing CDS formation, Light mercaptans are also adsorbed.

(b) Separation of normal and iso paraffin’s.

(c) Production of high purity N2, O2, H2, and inert gases from mixed gas streams.

(d) Static, (non-regenerative) dehydration of insulating glass units, whether air filled or gas filled.

13X offers a 10 Angstrom nominal pore size and lists the following applications:

(a) Removal of CO2 and moisture from air (air pre-purification) and other gases.

(b) Separating enriched oxygen from air.

(c) Removal of n-chained compositions from aromatics.

(d) Removal of R SH and H2S from hydrocarbon liquid streams (LPG, butane, etc).

(e) Catalyst protection, removal of oxygenates from hydrocarbons (olefin streams) (f) Production of bulk oxygen in PSA units.

Note that the last column in the above chart is the Dielectric Constants and you will note that aside from Methanol, they are all low polarity, with DEA about twice the DEC of Butane.

That leads us to the ph and polarity of the media. The simple Alkanes fully saturated and have a very low dielectric constant, about half that of diethanolamine, so is more easily attracted or held by polar media or water with a dielectric constant of 80.

PH is also how one paper on amine removal measured amine content, because they are a strong base (caustic).

I reviewed a paper on the subject published in the Royal Society of Chemistry forwarded to me by a team member, that compares 13X Mol Sieve, Silica Gel, and NKA-9 ionic beads in removing Methyl-diethanolamine from light Naphtha.

13X Mol sieve is a zeolite, an amorphous and porous form of Aluminosilicate, Silica Gel is similar based on Silicon Dioxide, and NKA-9 is a strongly acidic cation exchange resin.

In that experiment, 13X Mol Sieve out performed NKA-9, 20% and Silica Gel by 55%.

It was also most easily regenerated, using steam. First regeneration achieved 95% of original capacity, the second 90%, and the third 85%.

Subzero Scientific reports good results removing Medusa effects, using 13X in conjunction with a 0.5-micron filter and activated alumina beads, suggesting that compound filter media segments might further increase effectiveness. Activated alumina is AL2O3 and commonly used as a desiccant to dry air and gas streams, as well as desulfurizing natural gas using the Claus process.

Carbon Chemistries research revealed that Mordenite zeolite has 7 Angstrom pores, which is ostensibly about perfect, but has yet to be tested. Interestingly, one of the common uses of Mordenite zeolite is cat litter.

LLE Water Wash-Moving to water wash, we have three successful LLE LPG systems of different design and in different stages of development by team members and the process itself looks like a win.

Some input suggesting that the problem became more severe with recycle, suggesting a terpene buildup may have also had effect on Medusa. A good side benefit of LLE water wash is that it will also remove mono and sesquiterpenes.

I’ve spent some quality time green-field-dreaming on my 2002 Auto Cad, which is only 32-bit so I have to run it on an Oracle virtual program, which is problematic. I did manage to both put my latest brain farts on paper, but fully exercised my repertoire of colorful language, to the chagrin of my poor wife and dog.

Attached is green-field conceptual number 8 with process flow sheet and conceptual 9 that agitates and settles in the same vessel, then uses nitrogen pressure to transfer the LPG through the polishing filters.

The clarifier being researched would replace gas/water separator in conceptual 8 and the water trap in 9.

I’ve also included a conceptual for a column for using mol sieve beads, including post filtering out the mol sieve fines using DE and felted glass filters. The same multi-chamber and filter plate concept would work with other media than shown.

The gas industry uses 13X for removing amines and 5A for H2S and mercaptans. It also uses 3A and 4A for drying, so you will note in my conceptual that the final polish after LLE is 3 columns containing all four.

Currently researching liquid/liquid clarifiers to lower the water content of the washed LPG before it reaches the mol sieve desiccant drying cartridges, to prevent serious heat rise from the exothermic reaction, as well as excessive media consumption and regeneration costs.

In perspective, the test of condition of the mol sieve used in window insulating panels, is to put a measured amount of mol sieve in a measured amount of water and measure the significant heat rise.

In summary, it is of note that the LPG refineries use LLE to remove methanol and the amines, and then polish using column chromatography.

That prima facia evidence is supported by our own Medusa team member’s actual experience, though they have also found refinements improving efficiency.

Methanol and diethanolamine may or may not be the evil spirit, or the only evil spirit in LPG inducing medusa effects, but direct experimentation by Carbon Chemistry and others has demonstrated that column chromatography removes whatever it is.

In addition, several players have demonstrated successful Liquid/Liquid Extraction systems that also remediated Medusa effects and others are building test units. One shared their equipment designs and operating procedure to public domain as one solution.

Progress is being made on several fronts, so the question would now seem to be, not is there a fix, but how soon readily available and at what cost?

The cheapest solution from an initial capital equipment perspective of course, is to find a bulk supplier who can provide you with clean gas.

The viability of that solution depends heavily on the premium added for that pristine LPG, vis a vis what the extraction labs market base will bear.

Conversely, the amount of time and money that the supplier can invest to further purify the LPG is similarly and ultimately governed by whether the market for the diamonds will bear the cost.

Column chromatography would ostensibly be the next lowest initial capital cost, but media regeneration and replacement costs are not insignificant and ostensibly higher than LLE.

We have one media supply team member researching chromatography media, several building their own carts, and one equipment manufacturer researching cart design.

13X molecular sieve performed the best in amine removal studies, but in addition to the cost of the media and cart, consideration must be given to regeneration. If steam, you need a boiler….

An opening for a business renting and servicing carts??

The fish trap exists only because of the fish, and there are a bunch of different thoughts on LLE fish trap design, but they all work on the general principle that you mix the water and LPG and agitate in a manner that mixes it thoroughly, so that the water-soluble elements are removed from the LPG and retained in the water, which is then allowed to separate.

Once separated the water is drained off and the LPG dried of retained moisture using coalescers, refrigeration, and/or desiccants.

How elaborate the system required is to a great extent how great the scale. Batch processing small quantities for a single lab can be done simpler than systems for an LPG bulk plant.

Anyone considering building their own systems, should be vigilant with regard to LPG fill. It and the water should never fill more than 80% of the pressure vessels. Vaporizing 70:30 LPG exerts about 158 psi head pressure, but liquid LPG expanding in a confined space exert hydraulic pressures high enough to split 3,000 psi hydraulic tubing.

For that reason, besides the 20% head space requirement in the pressure vessels, any plumbing that can be full of LPG and has a valve on both ends that can be simultaneously closed requires a pressure relief valve/device (PRV), as do all vessels larger than 6”, which are considered pressure vessels, instead of 6” and under, which are pressure piping and obey different rules.

Simple minded solutions: Starting simple and considering ways for the LPG and the mol sieve beads to spend quality residence time together, I was taking an inventory of my surplus hardware and my eyes drifted across venerable old Mk III Terpenator, SN-0001.

Hee, hee, hee, It occurred to me that if you replaced the column with a larger one, and filled it with mol sieve, you could flood the column with LPG and let it soak for an extended residence time, before blowing or pumping it to a clean LPG storage vessel to be used in diamond growing processes.

If you left the vent valve open so that the lower Terpenator tank served as the safety headspace, you could stop bottom flooding as soon as the column over flowed and after the desired residence time, you could pressurize the column with nitrogen and push the cleaned LPG to a clean storage vessel.

What LPG that does find its way to the lower tank, could be recovered using the traditional Terpenator refrigerant recovery pump.

To make the run worthwhile, the largest possible column should be used, so doing a little math, the lower pot on the Mk III has an internal capacity of 170 cubic inches, so it could provide 20% headspace to 170/.20 = 850 cubic inches of column.

In perspective, a 4” X 48” column has a capacity of about 603 cubic inches.

With a six-inch extension on the lower tank, it has 340 cubic inches, so would support 340/.20 = 1700 cubic inches. In perspective, a 6” X 48” column has 1357 cubic inches volume.

The LPG and the beads are sharing space in the column, so I’m not sure how much space will actually be occupied by LPG, but if it were 50%, 1357 X .5 X (.0361 X .601) = 14.7 lbs of n-Butane or about 13.4lbs 70:30 mix.

Here is the simple minded schematic and the operating procedure follows:

Diamond Extraction, Butane Extraction, BHO, Solvent Direct Butane, Dry Ice, Ethanol, Carbon Chemistry, Media Bros, Lustermax, Molecular Sieve, THC-A Diamonds

Mk III Terpenator modified to filter LPG

Terpenator Medusa Filtration Unit Operating Procedure

  • Assemble the system and open valves 9 and 12. Start with valve 10 and in either position and valve 16 in nitrogen backfill position.

  • Start vacuum pump 4 and open valve 11.

  • When the system has evacuated to -29.5” HG, switch valves 10 and 18 to their other position and evacuate until you are again at -29,5” Hg.

  • Close valves 9 and 12.

  • Open valve 14 and flood column until it overflows the column vent valve 18. You can tell when that occurs by simply laying your hand on the discharge line and you will feel when it instantly turns much colder, when the liquid overflows.

  • Close valve 14 and allow the system to soak “N” minutes.

  • At the end of the soak, turn valve 10 to recover position and valve 8 to nitrogen push position.

  • Open valve 12 and 13.

  • Start recovery pump and evacuate lower tank to -22” Hg.

  • Close valve 12 and turn off recovery pump.

  • Open valve 15 and 16 to push the LPG in the column into the clean storage vessel.

  • When transfer is complete, close valve 13, 15 and 16.

  • Bleed off any nitrogen in the clean storage tank using valve 17.

Moving on with simpler solutions, Evolved Extraction offers the following practical LLE solution and are also proposing to do custom design-build for people that need a properly certified ASME or PE stamped system or systems for commercial applications that are built for their specific application parameters such as batch sizes, production rates, wash times, filtrations etc.

Next come the green field brain fart conceptuals:

The following demonstrate process and straw man solutions on a more elaborate scale. I’ve done about nine so far, with more coming as the individual pieces of equipment and actual process is identified:

Multi chambered Chromatography Column conceptual with filter plate for membranes and a final glass filter for dust. Flow is from left to right and the first segment uses 13X mol sieve and the second diatomaceous earth with glass filter to filter out any fine dust from the mol sieve, but there can be multiple chambers filled with different media to target different evil spirits.

This design uses easy to access 4” Sanitary fittings and will meet ASME Section 8 LPG pressure piping requirements of 350 psi. The 0.083 wall 4” sanitary tubing is rated at 3,000 psi, while the 13 MHP clamp is rated at 800 psi at 70F and the SSH is rated at 870 at 450F. The 1 ½” 13MHP clamps are rated at 1500 psi and the SSH at 1450 @ 450F:

Diamond Extraction, Butane Extraction, BHO, Solvent Direct Butane, Dry Ice, Ethanol, Carbon Chemistry, Media Bros, Lustermax, Molecular Sieve, THC-A Diamonds

Diamond Extraction, Butane Extraction, BHO, Solvent Direct Butane, Dry Ice, Ethanol, Carbon Chemistry, Media Bros, Lustermax, Molecular Sieve, THC-A Diamonds

Medusa LPG LLE Conceptual #8 Component and Operational Details

Equipment by number from Medusa LPG LLE Conceptual #5:

High pressure Nitrogen cylinder

LPG Crude tank

LLE Reactor vessel

Vapor/LPG/water separator

Vaccon HVP Series 300 venturi vacuum pump or equivalent NEMA 7 Class 1 D 1 pump.

Epoxy lined 316SS Distilled/RO water tank

PH Probe Digital Analysis Corporation #192V757SD-020BB or Rosemount.

Reagent metering pump by LMI or Rosemount

RO Filtration system

Gas coalescing filter

13X Mol Sieve column

5X Mol Sieve column

Mixed 3A and 4A Mol Sieve columns

LPG/Water agitation pump

Sight Glass

Tank sight glass

High level sensor

One batch transferred sensor

Low level sensor

3-Way water tank pressurize and vent valve

High level tank shut off

Vacuum pump isolation valve

Vac/Pressurize line isolation valve

Nitrogen isolation valve

Pressure regulator

Nitrogen tank valve

Pressure regulator

LPG Tank pressurizing valve

LPG Tank supply valve

Vacuum pump air inlet valve

Water tank pressurization valve

Reactor vessel Inductor isolation valve

Reactor vessel pressurization valve

NC Emergency shutoff solenoid valve.

Reactor vessel water fill valve

Reactor vessel Inductor isolation valve 2

3-Way transfer valve

Gas/LPG/water separator pressurization valve

Coalesing filter isolation valve

13X Column isolation valve

5A Column isolation valve

3A/4A Column isolation valve

Reactor vessel pump isolation valve

Reactor vessel drain valve

3-Way Gas/LPG/Water separator drain and pump inlet valve.

13X Column discharge valve

5A Column discharge valve

3A/4A Column discharge valve

PH adjustment reagent

Mixing Inducer

Compound pressure gauge

Compound pressure gauge

Reactor tank sight window.

2.0 Procedure:

2.1 Startup and fill:

2.1.1 Turn on RO Filtration system and fill RO Water Tank, adjusting the PH using PH Probe 7 and Reagent Pump 8. When filled to mid-level in the upper sight glass, shut off RO Filtration system. High Level Sensor 18 will also deenergize Normally Closed Solenoid Valve 22 to close it.

2.1.2 Verify that Valves 23, 24, 34, 40, 41, 42, 43, 47, 48 are opened and 3-Way Valves 38 and 46 are positioned to circulate through the pump.

2.1.3 Open Valve 31 and pump system down until Compound Pressure Gauge 53 on the Reactor Vessel reads -29.5” Hg vacuum.

2.1.4 Close Valve 34 and open Valve 39 to vacuum down Gas/LPG/Water Separator # 4.

2.1.5 Allow system to rest 5 minutes and observe pressure gauge for vacuum decay. Remediate any leaks.

2.1.6 Close Valves 24, 33, 34, 39.

2.1.7 Turn water tank vent Valve 21 to vent.

2.1.8 Open water tank drain valve 36 and fill the Reactor Vessel with water, until the level in the tank reaches the second sight glass and middle level sensor.

Note that if the water tank runs dry, there is danger that the system could ingest atmosphere, so do not operate with a water level below the lower sight glass or low-level sensor.

The low-level sensor will also deenergize (close) Normally Closed Emergency shutoff solenoid Valve 35.

2.1.9 Open Valve 30 and fill Reactor Vessel with LPG by weight, as verified by it reaching Sight Window 55, then close Valve 30.

2.1.10 Verify that Valves 33, 37, and 44 are open and that Valve 39 is in circulate position.

2.1.11 Start pump 14 and circulate tank contents for “N” Minutes.

2.1.12 While the Reactor Vessel is circulating, check Compound Pressure Gauge 54 to verify that -29.5 Hg vacuum level has been reached. When reached, close Gas/LPG/water separator pressurization Valve 39. Then close Vacuum pump isolation Valve 23 and then Vacuum Pump air inlet Valve 31. Hold under vacuum for 5 minutes to check for vacuum decay on Compound Pressure gauge 54.

2.1.13 At the end of the circulation cycle, and when Step is complete, close Valve 2. Turn 3-Way Valve 38 to transfer position and pump the contents of the Reactor Vessel into the Gas/LPG/Water Separator.

2.1.14 When the transfer is complete, turn off Pump 14, and turn 3-Way Transfer Valve 38 back to circulate position.

2.1.15 Allow the contents to settle and separate for “N” Minutes.

2.1.16 Once the contents has separated, turn Three- Way Valve #46, located on the bottom of the separator, to drain position and drain off the water to waste treatment, using the sight glass to see where to cut off.

2.1.17 Turn the Separator drain Valve 46 to discharge position and then open 3A/4A Column discharge valve #49.

2.1.18 Turn 3-Way Valve 46 to transfer position and open valves 41, 42, 43, 47, 48, and 49.

2.1.19 Pressurize the Nitrogen system by opening Nitrogen tank Valves 24, 25, 27, and 39 to push the remaining LPG through the mol sieve columns.

3.0 Nitrogen pressurization and backfill.

3.1 Various parts of the system can be pressurized using the Nitrogen tank 1 to facilitate transfer, to backfill with inert gas, and to purge the system of liquid before opening.

3.1.1 To pressurize the system, open Nitrogen tank Valves 24, 25, and 27.

3.1.2 To pressurize Water Tank 6, open Water Tank Pressurization Valve 32.

3.1.3 To pressurize the Reactor Vessel, open Reactor Vessel Pressurization Valve 34.

3.1.4 To Pressurize the Gas/LPG/Water Separator, open Gas/LPG/Water Separator Pressurization Valve 39.


Conceptual Number 9 settles in the reactor tank and uses only nitrogen pressure to expel the water and to transfer, for those pumps that can’t gracefully tolerate cavitation.

It also shows a water trap where current research strives to install a more effective liquid/liquid clarifier.

Diamond Extraction, Butane Extraction, BHO, Solvent Direct Butane, Dry Ice, Ethanol, Carbon Chemistry, Media Bros, Lustermax, Molecular Sieve, THC-A Diamonds

Conceptual 9

While it is of note that the evil spirts giving us grief are in the PPB level, with typical quality tests in the PPM range. We have discussed several different approaches under way to scrub the LPG, but for Six Sigma or greater quality, the rest of the supply chain issues must be addressed and one of the elephants in the room is tank cleanliness.

There doesn’t appear to be a universally accepted standard for cleaning returned tanks before refilling and tanks typically get mixed between customers.

In discussing tank cleaning with different LPG suppliers, I found a spectrum ranging from never venting any residual VOC and just topping, to cleaning tanks that failed the “sniff test”, to cleaning 100% of the tanks, with other in between standards.

It goes without saying that if contaminants at the PPB level are of concern, a regimented cleaning process is required. I have yet to tour any tank wash operations, but in aerospace the exotic alloys must be meticulously cleaned prior to heat treat or welding, and the various investment casting configurations have numerous passageways. The method typically used for the purpose is a hot alkaline wash bath, followed by a 10 to 15KSI high pressure hot water wash.

They are then dried in a hot air oven. It should be easy to semi automate the high-pressure wash with a large pump and multiple nozzles while rotating the tank on its axis.

Final inspection could be by borescopic cameras and recorded.

Regardless of how it’s done, tank quality must be insured to maintain quality levels in the PPG levels for that LPG negatively affected by Medusa, so most certainly those tanks must be adequately maintained.

As the pharmaceutical extraction industry is but a small portion of the overall LPG market, and those negatively affected by Medusa smaller still, we need solutions that address our issues, without inconveniencing and driving up the price for the balance of the industry.

One solution might be color code those tanks used for that purpose and for the gas supplier to give them more stringent attention than the balance of the tanks in circulation. That would lessen the impact on gas suppliers not cleaning each and every tank and ensure higher quality to the affected consumer.

The consumer could add another layer of protection by buying or leasing their own tanks, so that they were never in general circulation.

Updates as they arrive. Please feel free to comment.

If you or anyone you know is experiencing Medusa, fast crash, or unusual bar formations please reach out to Solvent Direct at 1-833-PURE-GAS or by contacting us online at



bottom of page