The Isomerization of Safrole: A Review
[ Back to the Chemistry Archive ]
Many ways of isomerizing safrole (3,4-Methylenedioxyallylbenzene) into isosafrole (3,4-Methylenedioxy- propenylbenzene) can be found in the literature. The isomerizations, which in all reviewed cases are catalytic, often have high yields, but differs greatly in easiness and reaction times. The availability and costs of the catalysts also differ over a broad range.
The reaction, which is general for the conversion of nearly all allylbenzenes (like safrole, apiole, dillapiole and elemicin) to their corresponding propenylbenzene isomer, is usually carried out by heating the allylbenzene together with an inorganic catalyst or catalyst mixture under reflux, with or without the use of a solvent. After the reaction is finished, the catalyst is removed by filtration, any solvent removed under vacuum, and the residue is distilled to purify the propenylbenzene from unreacted allylbenzene and polymerized material.
The isomerization always gives a mixture of both cis- and trans- isomers of the propenylbenzene (approx 30:70 using standard hydroxide ion catalyzed isomerization). The trans-propenylbenzene is thermodynamically more stable than the cis-propenylbenzene, so with longer reaction times, the ratio of the trans-isomer is increased in respect to the cis. As increasing the reaction time beyond the necessary for converting the allylbenzene will lower the yield of the product propenylbenzenes (due to side reactions), so attempts to influence this ratio is usually not made. Because the reaction produces mixture of isomers rather than a single compound, the boiling point of the propenylbenzene is never sharp. In the case of isosafrole, the trans isomer boils at approx 10°C higher than the cis, so the isosafrole isomers are usually collected as a mixture over a temperature range, rather than trying to collect each fraction separately. In spite of a certain errorneous statement in Total Synthesis II For all practical purposes, cis- and trans-isosafrole reacts just the same, so there is no real point in separating the isomers.
Physical properties of safrole and isosafrole
Using only KOH as catalyst and doing the reflux under vacuum with no solvent is preferable for several reasons. First, the vacuum helps removing the water (~15%) present in the KOH, and the lower reflux temperature prevents excessive side reactions.
Potassium hydroxide (10-12g, 1-2 mole%) was added to 800g safrole (from sassafras oil fractionally distilled twice) situated in a 2000ml round-bottomed flask (never fill a flask more than half full), equipped with an efficient reflux condenser, and with the top of the condenser attached to a vacuum source of approximately 10-20 mbar capacity. Magnetic stirring was initiated in the flask, and the suspension was brought to reflux under vacuum, which was typically attained at approximately 120°C. The reflux was continued overnight, the vacuum source disconnected, and the reflux condenser was replaced by a fractional distillation setup with a vigreux column and the isosafrole was distilled, again with good magnetic stirring. Refraction index measurements shows the reaction to be mostly completed after approximately 3 hours, but the reaction was still typically refluxed overnight. Stronger vacuum means longer reaction times, because of lower reflux temperature. Bigger reactions than 1000ml of safrole that have the tendency to bump violently, apparently magnetic stirring isn't good enough anymore. For our industrially oriented bees: using a capillary to introduce a stream of air bubbles, comparable to good old-fashioned vacuum distillation should work to circumvent the batch size limitation.
The KOH residue in the distillation flask can be reused to isomerise another batch of safrole. The vacuum limits reflux (=reaction) temp and also helps in removing water formed by the reaction of KOH with the safrole. The higher the temperature, the faster this reaction will be over. As a rule of thumb, half the vacuum means lowering the boiling point about 10°C. A substance boiling at 200°C at 1000mbar (athmospheric) will boil at 190° at 500mbar, 180° at 250mbar, 170°C at 120 mbar, 160°C at 60 mbar, 150° at 30mbar and so on. This is not exact, but can be used to estimate the expected boiling points roughly. Magnetic stirring is absolutely necessary. Don't do this with more than 1L safrole, strong bumping was experienced with bigger amounts. Use distilled safrole! When there's too much water present, reaction fails. Distill the iso with column, preferrably collecting the middle 80%. Recycle the first and last 10%.
Regulating a vacuum and keeping it constant is quite difficult. It's best to use a diaphragm pump at full vacuum. Using a rotary vane oil pump here doesn't work, because it pulls too much vacuum.
Isomerization with KOH/CaO
If calcium oxide is added to a mixture of potassium hydroxide and safrole, the reaction becomes less water-sensitive as the CaO reacts with any water present to form Ca(OH)2. The reaction is also very fast. The downside is that the CaO forms a black insoluble gunk in the reaction flask, and has to be removed by vacuum filtration with a filter aid before the isosafrole can be distilled off.
100 grams of safrole was placed in a 250 ml RB flask fitted with a water-cooled Allihn reflux condenser, and was heated with good stirring. When the solution started to boil, 1 gram of KOH pellets was added through the condenser, and the safrole immediately took on a brown color. Next, 15 grams of CaO was also added through the condenser, and the solution was allowed to reflux for 15 minutes, and the heat was turned off, and the solution was poured into a beaker and was allowed to cool. The dark brown suspension was vacuum filtered through a pad of celite, and the filter cake was washed with 2x50 ml of CH2Cl2. The CH2Cl2 was removed through distillation at atmospherical pressure, and unreacted safrole was also distilled off at 240°C (bath temperature), and amounted to about 1 ml. The residue was distilled at aspirator vacuum (~25 mmHg), and the isosafrole came over at 120-130°C as a water-white oil, weighing 70 grams (70%).
Base catalyzed isomerization using Microwave Heating
Safrole and eugenol isomerization reactions were carried out under microwave irradiation at atmospheric pressure and homogeneous medium with various alcohols used as solvents and different base concentrations. The rate reaction enhancement shown was up to 13 times faster than conventional reflux.
To study the microwave irradiation effect over isomerization reactions of safrole at atmospheric pressure and homogeniuous medium, experiments with various alcohols used as solvents and different base concentrations were carried out in an adapted microwave oven and in conventional reflux.
The alcoholic alkaline solution is prepared by the prolonged stirring of 8.8g (or 4.4g for 2M) of KOH pellets in 30ml EtOH. The KOH and the alcohols were of analytically purity grade. The alkaline solutions is placed in a RB flask provided with reflux condenser then 4.0g of safrole is added and the solution heated.
The microwave oven used for this study was a continental 2001/500W. the reflux system was pyrex and adapted through a hole in the ovens upper wall, having a diameter smaller than half wavelength to avoid microwave radiation escapes.
Phase Transfer Catalyzed Isomerization
This is a method of isomerising allyl-benzenes to propenyl-benzenes in aqueous conditions at only 70°C.
Allylbenzene was isomerized to trans- and cis-propenylbenzene. The reaction was run under varying conditions of mechanical stirring speed, PTC structure, initial catalyst concentration, temperature, catalyst counterion, and aqueous NaOH concentration. Standard reaction conditions consisted of 1.42 M allylbenzene in toluene, 5 mol% tetrabutylammonium bromide (TBAB), 40-50% aqueous NaOH, 75°C and stirring speed of 850 rpm [at least 550 rpm is needed]. Other phase transfer catalysts can be used. But ethyl-tri-octyl-ammonium undergoes extensive decomposition within 1.5 h due to the available ethyl group being convenient for Hoffman elimination [The same goes for quats containing methyl groups]. Quats below a minimum organophilicity (eg. tetra-propyl-ammonium) are ineffective. Catalyst decomposition may effectively stop the reaction at 70-98% conversion, depending on reaction conditions.
Solventless PTC variation
In this reference there is a procedure for isomerization of safrole under solvent-free conditions. The scale is low, but scaling up with some modifications shouldn't be a problem.
To 10 mmol safrole (1.62g) are added 12 mmol of solid crushed KOH (480mg) and 5 mol% (168mg) Aliquat 336 (common PTC). After shaking vigoruosly, the mixture is warmed to 80°C and held there for 5 minutes. After cooling 50ml CH2Cl2 is added and the the rxn mixture is filtered over 5g florisil. CH2Cl2 is removed under vacuum and isosafrole is obtained in 96% yield after distillation.
Transition Metal Catalyzed Isomerization
According to this patent, safrole isomerization can be effected with various transition element salts. A suspension of the specified amount of transition metal salt in 50g safrole is heated to 130°C for some time, and then the mixture is distilled to give isosafrole with a minimum (<5%) of safrole forerun. The following are the amounts used, heating time and yields with selected salts tried by the authors of the patent.
Safrole isomerization with RuCl3:
Allylbenzenes can be isomerized to propenylbenzenes either in a batchwise or continous mode by letting it react with a catalyst prepared by fusing activated alumina, sodium hydroxide and sodium metal. The catalyst can also be reused.
Addition of diethanolamine to the isomerization mixture and performing the reaction in ethylene glycol monoethyl ether drastically reduces the time needed for isomerization, as well as lowering the needed temperature, thereby lowering side-reactions occuring at higher temperatures
1000g of safrole is added to 10g of potassium hydroxide and 50g of diethanolamine in 360g ethylene glycol monoethyl ether. The temperature is raised to 160-180°C and held there for a few minutes while the isomerization takes place. After the reaction the reaction mixture is poured into 1000ml water and neutralized with 25% sulfuric acid. The organic layer is separated and the aqueous layer extracted with toluene and the extract added to the organic layer. After the toluene and water was distilled off, isosafrole distilled over in 95% yield.
Fe(CO)5 (iron pentacarbonyl) and NaOH has been shown to be a very effective catalyst pair, giving a 97% yield of isosafrole at 99% purity [2-3]. A downside with this method is that Fe(CO)5 is a both toxic and expensive reagent, but on the other hand, the reaction is very rapid and the molar consumption of the catalyst is low.
Disodium Iron Tetracarbonyl 
The reagent Na2Fe(CO)4 prepared in situ in THF by the reduction of Fe(CO)5 with sodium naphthalenide reacts with safrole in the presence of CuCl to bring about isomerization to trans-isosafrole in 87% yield. Reactions involving BrCH2CH2Br in place of CuCl give similar results.
CoCl2/Ph3P/NaBH4 System 
The catalyst generated in situ using CoCl2/Ph3P/NaBH4 in a 1/3/1 ratio in THF at -10°C isomerizes safrole into trans-isosafrole. Efforts to stop the reaction at the cis-alkene stages were unsuccessful. Furthermore, no isomerization occurs at -40°C, and above 0°C the catalyst decomposes. When more than 20 mmol (3.25g) of safrole was used, a considerable amount of starting material remained unreacted.
Propenylbenzene can also be prepared using this procedure by instead of safrole adding allylbenzene (2.36g, 20 mmol) to the reaction mixture. After distillation trans-propenylbenzene is recovered in 81% yield (1.92g, 16.2 mmol).
Potassium Hydroxide/n-Butanol 
To 91 mL (100g, 620 mmol) safrole were added 500 mL of a 3M solution of potassium hydroxide in n-butanol* and the reaction mixture was stirred under reflux for 3 h. After cooling, the mixture was poured into a solution of 30 mL of concentrated hydrochloric acid and 200 mL of ice water. After neutralization, the organic layer was washed with water and dried with anydrous Na2SO4. After removal of the excess n-butanol, the residue obtained was distilled under reduced pressure giving 95 g (95%) of pure isosafrole as a colorless oil, with spectra in agreement with the literature.
* Prepared by dissolving 84g KOH in 350ml n-butanol with stirring and slight heating, and then diluting to 500ml with more n-Butanol.
Rhodium(III)Chloride Trihydrate 
To safrole (20 g) in ethanol (5 ml) was added RhCl3·3 H20 (100 mg). The temperature was maintained at 20°C by cooling and after 2 h the mixture was poured into water and extracted with ether. The usual work-up was followed by distillation; the yield of the fraction bp 130-132°C at 15 mmHg was 93%. The minimum amount of catalyst for 20 g of safrole is 5 mg; complete transformation requires 24 h in this case. The amount of the [i]trans[/i]-isomer of safrole was about 80%.
 Osmium, communication at The Hive (1998)