Automated Computational Screening of Temperature-Dependent Cannabinoid Solubility in Industrially Relevant Processing Solvents and Lipid Carrier Oils

Imagine you are trying to dissolve a stubborn piece of candy in a glass of water. You know it won't work well, so you try swapping the water for something else, like warm soda or a splash of oil. In the world of chemistry, this is the daily struggle of working with cannabinoids—the active compounds found in cannabis plants. These molecules are like oil droplets that hate water; they are "lipophilic," meaning they love fats but repel water. This makes them incredibly difficult to extract from plants, purify, or turn into medicines that your body can actually absorb. Scientists call this the "solubility problem." To solve it, researchers usually have to mix and match different liquids (solvents) and heat them up, a process that can take forever and cost a fortune if you have to test every single combination by hand in a lab.

This is where a new kind of digital detective work comes in. Instead of mixing chemicals in beakers, scientists can now use powerful computer programs to simulate how these molecules behave. Think of it like a video game physics engine: you tell the computer, "Here is a cannabinoid molecule, and here is a bottle of oil or alcohol. What happens if we heat it up?" The computer crunches the numbers based on the laws of thermodynamics to predict if the molecule will dissolve, how much of it will dissolve, and how temperature changes the game. This paper is all about running that simulation on a massive scale to see which liquids are the best "dissolving partners" for different types of cannabinoids.

---

The Digital Solubility Hunt

In this study, a team of researchers from Standard Seed Corporation and their collaborators decided to skip the messy lab work and go straight to the computer. They wanted to know exactly how well 19 different cannabinoids would dissolve in 18 different liquids, ranging from common industrial solvents to the oils you might find in your kitchen. They didn't just look at room temperature; they simulated the process at three specific temperatures: 0°C (273.15 K), 25°C (298.15 K), and 50°C (323.15 K).

To do this, they used a sophisticated tool called SolProp, developed by MIT. Imagine SolProp as a super-smart chef who has tasted every possible combination of ingredients in history. You give it the "recipe" (the chemical structure of the cannabinoid and the solvent), and it predicts the "taste" (the solubility) without you ever having to cook a single dish. The team ran this simulation 1,026 times to generate a massive dataset of predictions.

What the Computer Found

The results were surprisingly clear, almost like a map showing the best routes for a road trip.

1. The Heat Factor
The most consistent finding was that heat is the key. In every single scenario the computer simulated, raising the temperature made the cannabinoids dissolve better. Whether it was a drop of alcohol or a spoonful of oil, the molecules were more willing to mix when things were warm. This suggests that the process of dissolving these compounds "eats" heat (an endothermic process), which aligns with what we already know about physics. For example, the simulation predicted that the solubility of CBD in MCT oil would jump from -0.05 at freezing temperatures to +1.18 at room temperature, and all the way up to +2.23 at 50°C.

2. The Solvent Showdown
Not all liquids were created equal. The computer identified a clear hierarchy:

• The Winners: Polar solvents like DMSO, acetone, and alcohols (ethanol, isopropanol, methanol) were the champions. They broke down the cannabinoids most effectively. DMSO, in particular, showed the highest predicted solubility values, reaching as high as +4.27 for certain compounds at 50°C.
• The Losers: Non-polar hydrocarbons like hexane, heptane, and butane were much less effective. In many cases, the computer predicted negative solubility values, meaning the cannabinoids barely wanted to dissolve in them at all.
• The Kitchen Oils: The results for carrier oils were a mixed bag. MCT oil, coconut oil, and butter showed medium-to-good solubility, which matches what the cannabis industry already uses in real life. However, the simulation hit a wall with olive oil and sesame oil. The computer gave them wild, unrealistic numbers (like a solubility of +28.74 for CBD in olive oil or -151.21 in sesame oil). The authors admit this is a glitch in the model, likely because these complex oils are too tricky for the current software to handle accurately.

3. Acid vs. Neutral
The study also looked at the shape of the molecules. They found that "acidic" cannabinoids (those with a carboxylic acid group attached, like THCA) dissolved slightly better than their "neutral" cousins (like regular THC or CBD). The computer suggests this is because the acid group acts like a magnet, helping the molecule stick to polar solvents a bit more strongly.

4. The Twin Mystery
One of the most interesting findings involved HHC (hexahydrocannabinol), a semi-synthetic compound. There are different versions, or "stereoisomers," of HHC (9(R)-HHC, 9(S)-HHC, and iso-HHC). You might think their tiny differences in 3D shape would change how they dissolve, but the simulation showed they were practically identical. In acetone at 0°C, their predicted solubility values were +1.62, +1.62, and +1.54. The computer suggests that for solubility, the tiny twists in the molecule's shape don't matter as much as the chemical groups attached to it.

Why This Matters (and What It Doesn't)

This paper doesn't claim to have solved the solubility problem forever. The authors are careful to point out that these are simulations, not physical experiments. They didn't mix chemicals in a lab to prove these numbers are 100% real. However, the results line up well with what scientists have seen in previous experiments with common solvents like DMSO and alcohols.

The study suggests that this kind of "automated screening" is a powerful tool for the future. Instead of spending months testing every possible liquid in a lab, companies can use these computer models to narrow down the best candidates first. It helps them figure out which solvents are best for extracting cannabinoids, which ones are good for making medicines, and how temperature affects the whole process.

While the model isn't perfect yet—especially when it comes to complex kitchen oils like olive oil—it provides a massive, temperature-dependent dataset that didn't exist before. It's a digital roadmap that helps researchers navigate the tricky world of cannabinoid chemistry, saving time and resources before they ever step foot in a laboratory.