HSP Application note #8
Guideline for Residual Solvents Q3C(R3) with Hansen Solubility Parameter (HSP)2009.9.23
HSPiP Team Senior Developer, Dr. Hiroshi Yamamoto
Alternative Solvents Search for the Pharmacy Chemicals
In the organic syntheses area, a lot kinds of solvents are used, but when you use that solvents for the Pharmacy Chemical synthesis, you need to pay attention for safety of that solvents. Because, if the solvents are remained in medicine, it means direct dosage to human body. And “Guideline for Residual Solvents Q3C(R3)” is published by ICH (International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use). In this guideline, they categorize solvents into 3 class plus one.
Class 1: solvents that should be avoided.
Solvents in Class 1 should not be employed in the manufacture of drug substances, excipients, and drug products because of their unacceptable toxicity or their deleterious environmental effect. (Benzene, Carbon tetrachloride, 1,2-Dichloroethane, 1,1-Dichloroethene, 1,1,1-Trichloroethane)
Class 2: Solvents to Be Limited.
Solvents in this class should be limited in pharmaceutical products because of their inherent toxicity. PDEs are given to the nearest 0.1 mg/day, and concentrations are given to the nearest 10 ppm. ( Acetonitrile, Chloroform, Methanol, Pyridine, Toluene etc.)
Class 3: Solvents with Low Toxic Potential
Solvents in Class 3 may be regarded as less toxic and of lower risk to human health. Class 3 includes no solvent known as a human health hazard at levels normally accepted in pharmaceuticals. (Acetone, Ethanol, Ethyl acetate, 2-Propanol etc.)
(Class 4: Solvents for which No Adequate Toxicological Data was found
The following solvents may also be of interest to manufacturers of excipients, drug substances, or drug products. However, no adequate toxicological data on which to base a PDE was found. (1,1-Dimethoxymethane, Isopropyl ether, Trichloroacetic acid etc.))
You would be better not to use Class 1,2 solvents.
But sometime it is inevitable to use solvent.
The reasons why we use solvents, are “want dissolve something”.
Once dissolve, higher yield, higher selectivity, easy to heat/cool, easy to separate(re-crystallize or extract), we can expect a lot.
Then how we can expect one solvent dissolve the target molecule?
We propose you to use Hansen Solubility parameter.
Have you ever heard of “Solubility Parameter”?
It is very popular for Polymer research area, but maybe not so much in Pharmacy Chemical area. The basic concept of solubility parameter was developed by Hildebrand and Scott. (“The Solubility of None-electrolytes.1964)
The free energy of mixing is written as
ΔG=ΔH-ΔTS
and if ΔG is less than 0, at that time mixing will happen.
ΔH=φ1φ2V(σ1-σ2）2 φ：volume ratio、σ：SP value
The SP (solubility parameter) value was first appearing in this scheme.
If the SP values are very near, the ΔH becomes 0, so ΔG becomes easily less than 0.
The concept of “likes dissolve likes” comes from this idea.
σ={(ΔH-RT）/V}0.5
A lot of σ are calculated from heat of vaporization and listed on handbook or database.
Pirika JAVA Demo Applet calculate HSP. HSPLight is available here.
The Hildebrand SP is based on just Heat of vaporization, so if the Heat of vaporization is very near even though chemical structure is so different, Hildebrand SP become same. For example, the Hildebrand SP of Ethanol and Nitroethane is almost same. But every chemist feel strange if they hear solubility of Ethanol and Nitroethane is same.
Hansen divided this heat of vaporization energy into 3 parts.
δD: Dispersion energy
δP: Polarity energy
δH: Hydrogen bonding energy
And use this 3 parameter as Vectors, [dD, dP, dH].
We show you 3-dimensional chart of Hansen Solubility Parameter of Ethanol and Nitroethane.
Hansen Solubility Parameter and Hildebrand SP has relationship of following scheme.
Hildebrand SP^2= dD^2 + dP^2 + dH^2
So Hildebrand SP is the length of Hansen SP vector’s length.
The chemist easily understand the difference of solubility Ethanol and Nitroethane with this 3D chart.
Solubility parameter basic concept is “likes dissolve likes”. So Hildebrand SP is not adequate to search solvents. Our concept is “likes vectors dissolve likes vectors”.
For the vectors likeness, we use HSP distance(Ra).
HSP distance(Ra)={4*(dD1-dD2)2 + (dP1-dP2)2 +(dH1-dH2)2 }0.5 Eq.(1)
In this article, we apply Hansen Solubility Parameter(HSP) to in real world.
For the first example, let’s think about Abietic Acid (CAS# 514-10-3)
Abietic Acid was known as one of the principal constituent of pine resin. The Pine resin was used as flux for Solder and to remove flux, CFC-113 was used. The CFC-113 was phase out at 1996 because CFC causes to destroy Ozone Layer. And a lot of alternative solvents for CFC-113 are studied with testing solubility of Abietic Acid. Can you predict solubility of Abietic Acid from it chemical structure? If you can, then it will time waste to read this article, so please skip.
We show you the result of solubility(g/100g solvent).
| Name | Solubility | dD | dP | dH | Vol | Ra |
| Acetone | 23.1 | 15.5 | 10.4 | 7 | 73.8 | 12.3 |
| methyl ethyl ketone | 35 | 16 | 9 | 5.1 | 90.2 | 10.7 |
| diethyl ether | 81 | 14.5 | 2.9 | 4.6 | 104.7 | 11.7 |
| toluene | 91 | 18 | 1.4 | 2 | 106.6 | 6.4 |
| dichloromethane | 94 | 17 | 7.3 | 7.1 | 64.4 | 8.0 |
| chloroform | 116 | 17.8 | 3.1 | 5.7 | 80.5 | 5.0 |
| 1,1,1-trichloroethane | 71 | 16.8 | 4.3 | 2 | 99.3 | 8.4 |
| 1,1,2-trichloroethane | 45 | 18.2 | 5.3 | 6.8 | 92.9 | 4.9 |
| 1,2-dichloroethane | 35 | 18 | 7.4 | 4.1 | 79.4 | 6.9 |
| trichloroethylene | 122 | 18 | 3.1 | 5.3 | 90.1 | 4.7 |
| tetrachloroethylene | 71 | 18.3 | 5.7 | 0 | 102.8 | 8.0 |
| CFC-113 | 0.8 | 14.7 | 1.6 | 0 | 119.2 | 12.9 |
| HCFC-225cb | 0.7 | 13.1 | 2.9 | 1 | 165 | 15.3 |
CFC-113 and HCFC-225 dissolve abietic acid 0.7-0.8g/100g Solvent. This result means CFC or HCFC will dissolve the flux but will not dissolve other plastic like poly vinyl chloride using on the electronic substrate.
Why some solvents are good solvents and some are poor solvents?
Let’s examine with using HSP. The basic concept of HSP is “Likes Vector dissolve likes vector”. The HSP values of solvents are listed on table. You can get these vales from Hand Book, database or the Internet but those values are very old so we will not recommend using. And HSP of Abietic Acid is hard to obtain, and for such compounds you need to use HSPiP software to estimate HSP.
If you use HSPiP, you can get newest HSP for solvents, and if there is no data on database, you can estimate HSP from it chemical structure using Y-MB. For example, the Smiles notation of Abietic Acid is,
C[C@](C(CCC(C(C)C)=C3)C3=CC1)(CCC2)C1[C@@]2(C)C(O)=O
And, you copy this Smiles text and paste into text field in Y-MB, and push calculate button, Y-MB automatically break molecule into functional groups, then calculate HSP and other properties with using Neural Networking technology.
Smiles(Simplified Molecular Input Line Entry Syntax) SMILES is a string obtained by printing the symbol nodes encountered in a depth-first tree traversal of a chemical graph. Pirika JAVA Demo Applet getting Smiles. Draw2Smiles is available here. |
Y-MB Properties EstimationY-MB break Smiles into correspponding Functional Groups and Estimate various Properties. These estimation schemes are come from Pirika technologies.
Pirika JAVA Demo Applet calculate Properties. PirikaLight is available here. |
The Y-MB result for abietic acid becomes [20.3, 2.8, 6.3] and molecular volume is 302.5.
Then calculate HSP distance (Ra) from solvents using equation (1).
HSP DistanceTo calculate the distance (Ra) between Hansen parameters in Hansen space the following formula is used:
HSP distance(Ra)={4*(dD1-dD2)2 + (dP1-dP2)2 +(dH1-dH2)2 }0.5 |
If the distance is short, the vectors of abietic acid and solvent are alike, and that solvent should dissolve abietic acid very well.
There are 3 exceptions, but roughly speaking, HSP distance are lower than 8, that solvent dissolve very well. And the distance are larger than 10 it will not dissolve so much. So if you want to search good/poor solvents, checking HSP distance may help you so much.
Hansen SphereTo determine if the parameters of two molecules (usually a solvent and a polymer) are within range a value called interaction radius (R0) is given to the substance being dissolved. This value determines the radius of the sphere in Hansen space and its center is the three Hansen parameters.
From version 3.1.X, Double Spheres function is available. Pirika provide JAVA 3D Demo Applet to browse the Sphere(s). |
Let’s examine the other example.
Oleic Acid (CAS# 112-80-1) is aso studied it solubility intensively because FAME(Fatty Acid Methyl Esters) solvents are begun to use because of it enviromental safty properties.
2011.4.11
Drag=Rotate, Drag+Shift=Larger/Smaller, Drag+Alt or Command(Window key)=Translate. If you are using HTML5 enable browser such as Chrome, Safari or FireFox (IE9 is out of support), you will see the Canvas. If you pick solvent, solvent name will appear.
|
2011.6.11
|
| name | Solubility | dD | dP | dH | Volume | Ra | Ra-Alcohol |
| Acetonitrile | 1.1 | 15.3 | 18 | 6.1 | 52.9 | 15.3 | |
| Acetone | 27.4 | 15.5 | 10.4 | 7 | 73.8 | 7.7 | |
| isopropyl alcohol | 55 | 15.8 | 6.1 | 16.4 | 76.9 | 10.7 | |
| ethyl acetate | 44 | 15.8 | 5.3 | 7.2 | 98.6 | 2.7 | |
| butyl acetate | 48 | 15.8 | 3.7 | 6.3 | 132.6 | 1.0 | |
| diethyl ether | 60 | 14.5 | 2.9 | 4.6 | 104.7 | 3.4 | |
| 1,2-dichloroethane | 26.1 | 18 | 7.4 | 4.1 | 79.4 | 6.4 | |
| nitroethane | 2.2 | 16 | 15.5 | 4.5 | 72 | 12.8 | |
| nitromethane | 0.6 | 15.8 | 18.8 | 5.1 | 54.1 | 16.0 | |
| furfural | 1.3 | 18.6 | 14.9 | 5.1 | 83.2 | 13.2 | |
| butanol | 56.5 | 16 | 5.7 | 15.8 | 92 | 10.0 | |
| methyl ethyl ketone | 33.5 | 16 | 9 | 5.1 | 90.2 | 6.3 | |
| hexane | 44.4 | 14.9 | 0 | 0 | 131.4 | 7.1 | |
| methyl alcohol | 31.6 | 14.7 | 12.3 | 22.3 | 40.6 | 18.9 |
With this case also need to use HSPiP software to get HSP[16.2, 3.4, 6.3] of Oleic Acid, because there is no HSP listed on DB.
HSPiP(Hansen Solubility Parameters in Practice)The first edition of HSPiP that appeared in November, 2008, greatly enhanced the usefulness of the Hansen solubility parameters (HSP). The HSP values of over 1200++ chemicals and 500 polymers are provided in convenient electronic format and have been revised and updated using the latest data sources in the second edition (March, 2009). A third edition of the HSPiP appeared in March, 2010. There are now 10,000 compounds in the HSP file which also includes data on density, melting point, boiling point, critical parameters, Antoine constants and much more. The user is able to carry out many different sorts of optimisations of solubility, evaporation, diffusion, adhesion, create their own datasets (automatically if required) and explore the huge range of applications for HSP in coatings, paints, nanoparticles, cosmetics, pharma, organic photovoltaics and much more. The 3rd Edition v3.1 was released on 12 December 2010. Current users can upgrade free (now v3.1.09) by downloading the latest .msi installer from http://hansen-solubility.com The 4th Edition v4.0.x was released on 2 Jan. 2013. The Current users can upgrade with free charge. 2013.1.28 The Visual How to manual of HSPiP. You can understand what HSPiP can do. |
The result is very interesting. There are 2 lines. The Alcohol compounds have very long HSP distance but it dissolve Oleic Acid very well. For other solvents, it can be said, “The likes vector dissolve likes vector”.
Why Alcohol compounds dissolve oleic acid so well? The alcohol compounds have very large dH value and this make HSP distance so large. But sometime this dH will not affect solvency for some solutes.
Maybe Hydroxyl group of Alcohol make cluster with hydrogen bonding force, and stretch out the alkyl groups. So hydroxyl group will not affect to solvency. In that case, if you plot solubility to dP, then this plot explain both alcohol and other solvent’s solvency.
The Compounds that shows specific solubility to Alcohol, are likely have carboxyl group. Maybe such compounds make hydrophobic cluster like below figure.
Medicine or source of medicine, may have carboxyl group, so you need pay attention for HSP distance result for that compounds.
We have been using this HSP distance theory to find polymer solvent. In that area, we found not only pure solvent but also find a lot of mixture of solvent.
Mixture of solvents are not easily used by other solubility theories. HSP can handle
mixture of solvents just like “Addition of Vectors”.
HSP of Solvents Mixture[dDm, dPm, dHm]=[(a*dD1+b*dD2), (a*dP1+b*dP2),(a*dH1+b*dH2)]/(a+b)
Volume base ratio. Pirika Java demo applet design solvents mixture. GSD is available here. |
There are a lot of examples where two non-solvents create a good solvent when mixed. For example, PMMA[17.7, 9.1, 7.1] (Green) is insoluble in butanol[16.0, 5.7,15.8] (Blue) and insoluble in diethyl sulfide[16.8, 3.1, 2.0] (Red). You can easily find out that the Blue vector and Red vector are very far from the Green sphere center. But with a 50:50 mixture of butanol and diethyl sulfide, the new vector [16.4, 4.4, 8.8] (Purple) is very near to center of Green sphere and this blend dissolves PMMA!
If we can apply this theory to Q3C, what we want to do is design class 1, 2 solvents with “Addition of Vectors” of class 3 solvents. This concept is accepted by polymer scientist, but we are not sure in pharmacy chemistry area. Please let us know the results after you get patents.
We will explain how to search mixture of solvents whose HSP is almost same with Acetnitrile. The HSP of Acetnitrile is [15.3, 18, 6.1] and the large dP value is specific to this molecule. As listed in table, γ-Butyrolactone, 1-nitoropropane have large dP and near to Acetnitrile’s HSP but shortest HSP distance is 5.7, so it is hard to find alternative of Acetnitrile.
| 10 Acetonitrile | 15.3 | 18 | 6.1 | |
| 115 gamma-Butyrolactone (GBL) | 18 | 16.6 | 7.4 | 5.73 |
| 535 1-Nitropropane | 16.6 | 12.3 | 5.5 | 6.29 |
| 608 Sulfolane (Tetramethylene Sulfone) | 18 | 18 | 9.9 | 6.60 |
| 363 Ethylene Carbonate | 18 | 21.7 | 5.1 | 6.62 |
| 303 Dimethyl Sulfoxide (DMSO) | 18.4 | 16.4 | 10.2 | 7.60 |
| 7 Acetone | 15.5 | 10.4 | 7 | 7.66 |
Then how about mixture of solvent? If you are accustom to write Script and have HSP DB values, you can get mixture solvents HSP very easily.
This table shows the result of shortest HSP distance mixtures. We calculate each HSP distance with varying composition by 0.1 and extract shortest results.
| A | A ratio | B | B ratio | dDm | dPm | dHm | Ra |
| 7:Acetone | 0.5 | 363:Ethylene Carbonate | 0.5 | 16.75 | 16.05 | 6.05 | 3.49 |
| 363:Ethylene Carbonate | 0.7 | 963:t-Butyl Acetate | 0.3 | 17.1 | 17.05 | 5.43 | 3.78 |
| 363:Ethylene Carbonate | 0.6 | 369:Butyl Glycol Acetate | 0.4 | 16.92 | 16.02 | 5.78 | 3.81 |
| 363:Ethylene Carbonate | 0.7 | 666:Texanol | 0.3 | 17.13 | 17.02 | 6.51 | 3.81 |
| 363:Ethylene Carbonate | 0.7 | 440:Iso-Propyl Acetate | 0.3 | 17.07 | 16.54 | 6.0297 | 3.83 |
The all mixtures are based with Ethylene Carbonate, and the HSP distance becomes below 4.
If you do this by hand, it will take 1 day to complete, but using HSPiP “Solvent Optimizer” it take just few seconds. And if you have the Flash point data, you can set priority for these mixtures. Acetone(FP= -18℃), Tert-Butyl Acetate (FP= 15.5℃), Iso-propyl Acetate (FP=2℃) are very low Flash Point. BGA(Butyl Glycol Acetate)(FP=71℃), Texanol(FP=120℃) are very easy to use.
So, the first candidate of alternative of Acetnitrile is Ethylene Carbonate:BGA = 60:40 mixture.
If you use HSPiP software, you get not only HSP best match mixture but also properties of solvent that help you choose for your system, such as Boiling Point, density, Antoine Parameter, Flash Point, Melting Point so on.
Now, Professor Abbott add 2 files to HSPiP, so user can available “Solvent Optimizer” function with only class 3, 4 solvents. You can load Q3C(R) 3 and 4.sof. The extension of sof((Solvent Optimization File) is to use for regostration of solvents that can use for optimization. If you think you do not want to use class 4 solvents as sof solvents, you just move them to Q3C(R3) 1 and 2.ssd. You can load ssd(Sphere Solvent Data) file for some specific purpose.
We explains how to use them.
After you run HSPiP, then please open the Q3C(R) 1 and 2.ssd file from Open menu.
Then please select 1,2-Dichloroethane, then push “O: open the Solvent Optimizer” button on the right of pain. Then OptimizerSolvents.sof will open as offset file.
Then HSP value of 1,2-Dichloroethane will copy to target text field. You can change sof file by pushing the button of “Open an Optimizer Solvent Set”. Please select Q3C(R3) 3 and 4.sof.
And only you have to do is push “2: Find the Best 2 Solvents” button. Cumene:DMSO=62:38 will be proposed.
In this case, HSP distance of Cumene-1,2-Dichloroethane = 6.85，DMSO-1,2-Dichloroethane = 10.90, so both the solvents are far from 1,2-Dichloroethane. But the HSP of Cumene:DMSO=62:38 will become [18.2, 7.0, 4.6] and HSP distance from 1,2-Dichloroethane [18.0, 7.4, 4.1] is just 0.8.
So it may become good example of poor solvent mixtures become good solvent in Pharmacy Chemical area!
But remember that, you need not search substitution of class 1, 2 solvent. You would be better to calculate your target molecule’s HSP and search best fit solvents from class 3 or 4 solvents. If you want to calculate your molecule’s HSP, you just prepare Smiles of you compounds and put the Smiles into Y-MB.
I put Abietic Acid [20.3 2.8 6.3] as test case, “Solvents Optimizer” suggest to use TriChloroAcetic Acid as one of the components. If I do not want to use Class 4 solvents, I just Ctrl-Click that solvent. Then that solvent marked gray and be omitted from optimization. And final proposal was Anisole:Cumene=73:27. Toluene and Ether compounds are very good solvents for Abietic Acid, so this proposal is very sensed.
If you want to synthesis something with MicroWave, please read this article.
After you make something, you may want to separate target compound with liquid-liquid extraction. please refer to liquid-liquid extraction article.
Other topic of Bio Medical
GC data of class 1, class 2 solvents in Q3C:
HSP and Tamiflu: Solubility parameter of Tamiflu or other H1N1 antiviral compounds
HSP for Rabbit: How to design eau de Cologne for rabbit. what LD50(skin, rabbit) means.
HSP and logP: logP, logKow, it is just HSP volume.
HSP and Carcinogenicity: SOM(self organization map) analysis of Poly-chlorinated compounds
HSP and Endocrine Disruptor: categorize by SOM.
HSP and AntimicroBial. Sulfa Drugs and other kind of Drugs.
Gall stone solubilizer: How to dissolve Cholesterol base Gall Stone.
Caco-2 cell monolayer apparent Permeability:SOM analysis
