Benzotrichloride, Benzoyl Chloride, and Phthalyl Chloride - Illustrated Practical Guide

This article is part of a series of Illustrated Key Syntheses in Chemistry which I intend to post, with the purposes and guidelines outlined in the thread of the same name in the General Chemistry section. I shall attampt to ensure that all the relevant information is contained in a single post. If sufficient reason arises for something to be changed or edited, I shall edit the entire post.

Aim

As has been mentioned in the article on the synthesis of thionyl chloride, simple methods for manufacturing alkyl and acyl chlorides open the door to the synthesis of a large range of organic compounds. Although thionyl chloride is an excellent reagent for this purpose its main disadvantage so far has been the laborious nature of its preparation, in particular the need to handle large amounts of SO3 and the potential danger this poses should accidents occur. However, it is well known that acid chlorides can be converted to one another by acting on with the corresponding acid, the equilibrium being progressed if the more volatile chloride is continuously removed.

The road to acid chlorides for the amateur chemist is provided by the well known property of benzotrichloride to generate acid chlorides by its action on water, organic acids, or acid anhydrides (for the earliest work see Deutsch Patent 11494, 1879). It can also generate acid anhydrides by acting on organic acid salts, this however is not the preferred way for the synthesis of anything but benzoic acid anhydride due to the formation of mixed anhydrides as a byproduct.

I have already demonstrated that benzotrichloride itself can be easily prepared in good yield by blue-light mediated chlorination of toluene in the article on benzaldehyde synthesis. In that work, benzotrichloride was evidenced only by its ~10% contribution at 629 cm-1 in the IR spectrum of the isolated benzal chloride, and was thus an undesired impurity. The ability to continue the chlorination to actually achieve a good yield of this compound has up to this point been subject to severe doubt, due to the existance of quite contradictory statements on this point in the literature. Two antigonistic viewpoints to the success of this synthesis raised the following doubts

1. Free-radical chlorination becomes incredibly inefficient in the last stages of chlorination, requiring such exceedingly large amounts of excess chlorine as to make the process unviable in a batch amateur setup.
   
2. By-products in the form of ring-chlorinated compounds and intractable polymerisation tars form in greater and greater amounts as chlorination is continued. The latter can arise due to the continuous increase of the b.p. of the chlorinated mixture and the increase in polymerization tendency of chlorinated toluenes with temperature, the former due to the decrease in efficiency of free radical chlorination as a consequence of the reaction rate constant for the third chlorination being a factor of ~25 times smaller than for the first chlorination, while the reactant concentration also continuosly drops. These factors can potentially make the yield unperspective.
   

Other sources, such as the Kirk-Othmer Encyclopedia of Chemical Technology mention that chlorination is carried out to almost 100% benzotrichloride in industry, indicating that such an outcome is at-least possible, although no details of the conditions under which such conversions obtain are mentioned.

Here I have definitively found that the chlorination of toluene is essentially complete, high-yielding, and free of any IR detectable impurity. The excess of chlorine required for the entire procedure is 100% (this is based on TCCA used and not on chlorine generated, so this excess can be decreased at the expense of longer process time) and these results can be obtained using simple equipment in a batch process. The chlorination time from the benzal chloride to benzotrichloride, is not excessive, being only about 1/3 of the time needed to reach the benzal chloride maximum. If the aim is benzoyl chloride, this clearly puts this method ahead of the chlorination of benzaldehyde route, which according to the literature requires a dozen or so hours.
Moreover I have found the conversion of the trichloride to acid chlorides, and the interconversion of acid chlorides to be very pure and high-yielding processes. These processes, outlined in 30's literature (H.Brown JACS Vol.60 p. 1326, and L.P.Kyrides JACS Vol. 59 p. 206), have given me equivalent yields as for the original authors and in some instances cleaner products.

Findings

1. 86% yiled of benzotrichloride based on toluene in a two-step procedure
   
2. Benzotrichloride contains no foreign IR peaks (purity > 95%)
   
3. First chlorination of 2.2mol of toluene was carried out at b.p. in 1L flask, lasted 7hrs with 1kW high-pressure Hg illumination at 8cm with fan cooling, used 700grams TCCA and 875ml 30% HCl, produced 30gms (~15%) loss of product to polymerised tars, and 164gms weight gain, corresponding to a composition of Ph-C-Cl2.54-H0.44 (with account for polymerisation loss), which was supported by Beers law analysis of IR spectra giving a benzal chloride/benzotrichloride ratio of 1:1 based on peaks at 586 and 629 cm-1. The chlorine excess based on this composition and the amount of TCCA used is therefore 90% (190% of the required amount).
   
4. Second chlorination of 273gms of the distilled product from the first step carried out at a gradualy decreasing temperature range from 180C to 150C, took 3hrs, 150gms of TCCA and gave a weight gain of 37gms and a pure product based on IR. This gives, and is supported by a comparison of first step data with overall data, a yield of effectively 100% for this step (within experimental error). The chlorine excess for this stage based on TCCA is therefore 150%.
   
5. The chlorine excess for both steps combined in a 10 hour processes based on TCCA used is thus 100%. Note that this does not mean all this extra chlorine has been generated - the TCCA was replaced as soon as a 20% or so decrease in chlorine evolution was evident. This is based on the cheapness of the chlorine generating precursors. Very little chlorine was evident in the exhaust gases - except in the second step.
   
6. Benzotrichloride reacts rapidly at 100C, and effectively quantitatively with phthalic acid in the presence of zinc chloride to yield benzoyl chloride and phthalic anhydride. 145gms (0.74mole) of benzotrichloride reacted in 1 hr with 112gms of 90% phthalic acid (0.61mole) - 10% phthalic anhydride (0.08mole) in the presence of 10gms ZnCl2.(0.17 H2O) to yield 99.9gms (0.71mole) of benzoyl chloride a 96% yield - 4% lost most likely in distillation - and phthalic anhydride. Weight loss of reactants was 50gms, which assuming was all HCl amounts to 1.37mole, suggesting a few grams of benzoyl chloride was lost with the HCl.
   
7. All the benzoyl chloride distilled in a narrow range of 78-80C at pump pressure and was IR pure. The phthaloyl chloride distilled at 138-142C at pump pressure as soft white crystals, and was almost IR pure. The pump pressure was indicated by a vacuum manifold gauge to be about (atmosphere - 27inch Hg) which evaluates to about 74mm Hg - however tabulated vapour pressures for both isolated compounds suggest the manifold gauge is inaccurate by about 2inch Hg, and the actual vacuum was about 12mm Hg. (b.p. benzoyl chloride 78C at 12mm Hg, b.p. phthalic anhydride 140C at 12mm Hg). The intermediate fraction was negligible (< 1gm).
   
8. Benzotrichloride in the amount of 84.1gms (0.43mole) was reacted with 60.3gms (0.41mole) of phthalic anhydride in the presence of 3.2gms ZnCl2.(0.17 H2O) (0.004mole H2O) for 12 hours at 120-130C, as per Kyrides but with the reaction time cut in half and temperature raised by 10C. On completion of reaction 54.1gms (0.385mole) of IR pure benzoyl chloride was recovered by fractional distillation through a Hempel collumn at reduced pressure in the 80-82C temperature range. This amounts to a yield of 90%. 71.9gms (0.354mole) of IR pure phthaloyl chloride came over in the 132C-136C temperature range, with almost all the chloride coming over at 136C. This amounts to a yield of 86%. 3.4gms of substance, including some HCl, was lost as vapour in the course of the reaction, 6.4gms was collected this time as a middle fraction in the temperature range 86C-132C, while the remainder of yield remained asvpolymerization product in the flask, and some undistilled liquid in the Hempel collumn.
   
9. Any moisture in the ZnCl2 - which can not be removed by evaporation due to the formation of zinc oxychloride, zinc chloride being a moderately strong Lewis acid - is removed at about 70 - 80C by the benzotrichloride with evolution of HCl and formation of benzoyl chloride. In the above case the approximately 0.25gms H2O in the zinc chloride contributed to a 'loss' of about 2.7gms benzotrichloride in this way. The subsequently anhydrous ZnCl2 in addition to catalysing the desired reaction forms sites for polymerization of the benzotrichloride as is evidenced by a brown-black surface coating of the crystals, which increases with temperature.
   
10. Benzotrichloride was found to be a heavy slighly viscous liquid with a density of 1.36 (84.4gms in 62ml). It is initially yellow due to dissolved chlorine, however on distillation it is almost colourless with a slight hint of yellow. The latter reappears on heating probably due to formation of polymerization products and evolution of chlorine. It has a heavy sweetish smell somewhat reminiscent of chloroform, which gives one a headache on any more than a brief exposure. The liquid fumes when a bottle is opened on humid days, however on days with humidity < 50% this was not evident. Benzoyl chloride is a transparent liquid with a not-unpleasant pungent fruity smell which is also reminiscent of benzaldehyde, it is however almost immediately lachrimatory, much more so than the chlorotoluenes. This is probably due to its greater polarity and so water solubility. Phthaloyl chloride is a heavy transparent liquid with little odour, in ordinary vapour concentrations I did not have lachrymatory problems with it.
    
11. It is known, see Kyrides, that some acid chlorides can be used to synthesise thionyl chloride - which is an inorganic mixed anhydride. Since benzotrichloride has been seen to act in a certain sense as an acid chloride (it converts acids to acid chlorides) it was reacted with a rapid stream (~5 bubbles/sec) of dry SO2 for a 5 hour period at temperatures up to 165C. No traces of SOCl2 - only BzCl - was found in the exhaust gases. Raising the temperature further, as per 200C employed by Kyrides, is counterindicated by the proximity of the b.p. and decomposition of the benzotrichloride, as well as of the SOCl2. The later, rather than losses with SO2, I suggest contributed to the generally low yield of SOCl2 for this type of reaction obtained by Kyrides. It is possible that SO2 is unreactive with benzotrichloride due to the low polarizability of the latter.
    

Theory

The reaction schemes of this practical are presented below



Note that benzoyl chloride rather than phthaloyl chloride is distilled off in the second reaction above. This is in accord with the general principle seen in Brown's paper, where the equilibrium is shifted to the more volatile chloride by distillation. Moreover no mixed acid-acid choride compound of the type Cl(CO)-C6H4-COOH is formed, the later immediately decomposing to HCl and phthalic anhydride.

As outlined in Deutsch Patent 11494 by Emil Jacobsen, benzyltrichloride in the presence of a weak Lewis acid, such as ZnCl2 hydrolyzes with the stoichiometric amount of H2O according to

PhCCl3 + H2O -> Ph-(CO)-Cl + 2HCl

excess H2O leads to complete hydrolysis as in the benzaldehyde thread

Ph-(CO)-Cl + H2O -> Ph-COOH + HCl.

Moreover since the equilibrium of the latter reaction is essentially completely to the right, provided one does not err in the amount of H2O added during benzoyl chloride synthesis by more than a factor of 2, no H2O will be present in the final product.

Interconversion of acid chlorides and anhydride formation

The benzoyl chloride formed in the first reaction can be reacted with many organic acids, as per the paper by Brown, provided its chloride is more volatile, it can be distilled off. Thus

BzCl + AcOH -> BzOH + AcCl

However this is quite a disadvantageous use of benzotrichloride, and twice the yield of the acyl chloride can be obtained by by-passing Browns procedure and using the direct action of benzotrichloride, as per Emil Jacobsens patent

PhCCl3 + AcOH -> BzCl + HCl + AcCl
BzCl + AcOH -> BzOH + AcCl

overall

PhCCl3 + 2AcOH -> BzOH + 2HCl + 2AcCl

For anhydride synthesis however, all anhydrides except those of benzoic acid and anhydrides of dibasic acids such as phthalic anhydride, benzotrichloride should not act directly on the acid salt due to the following competing processes

AcONa + AcCl -> Ac-O-Ac + NaCl
BzONa + AcCl -> Bz-O-Ac + NaCl
BzONa + BzCl -> Bz-O-Bz + NaCl

all but the first yielding a byproduct. It is best to isolate the acyl chloride first using its volatility difference, and then act it on the anhydrous acid salt.

Method

All operations described below were carried out in a fume hood, wearing a gas mask, and latex gloves.

Ambient temperature in the lab 20C

Formation of benzal chloride/benzotrichloride 50% mixture - step 1

350 ml of commercial toluene was placed in a 1L flask and distilled - with the condenser jacket initially dry - with 202 grams (227ml) of an intermediate fraction being collected, after 50ml or so of toluene has passed. The first fraction containing water, and the third fraction left in the still containing xylenes and metal salts are combined and used subsequently to wash polymerization products from the reaction flask (they show little solubility in other solvents).




Chlorinating apparatus with high-pressure Hg illumination is now assembled as per the benzaldehyde thread (joint grease must not be used as the chlorotoluenes will attack it). Care is taken to locate the Hg lamp a minimum 8cm from the reaction flask (to minimise hot-spots) and slighly above it - as after the solution has darkened due to formation of condensation products, much of the chlorination will occur in the vapour phase. This has the double benefit of minimising ring chlorination. A slow air current from a desk fan is used to cool the lamp. The chlorine generator with its adjustment valve is located as far as practicable from the reaction flask, with an indespersed shield to minimise radiation exposure during adjustment. The entire fume-hood window is shielded with a heavy curtain screen during the reaction (beware that many materials will discolour). 5 moles of NaOH are placed in the scrubber vessel - sufficient to bind all the HCl generated, while forming an acidic NaCl solution at the end to indicate the reaction rate at that point.




The 1L flask is heated at full rate on a 400W mantle and the course of the reaction is followed by the rising temperature of the reflux. An HCl drip rate of 1 drop/2 sec on 200gms of pulverised TCCA is used, and as soon as the rate of temperature rise shows a decline (normally after 250 ml 15% HCl had been added) the TCCA is changed. The initial temperature rise is rapid compared to that later on, due to the progressing diminution of the temperature difference between the b.p.'s of chlorinated toluenes (110C, 179C, 205C, 223C) for (toluene, benzyl chloride, benzal chloride, benzotrichloride). A typical graph of temperature variation is shown below - the flat plateau in the middle of the first graph corresponds to TCCA change.


When a b.p. of 220C is reached the reaction products are a quite dark yellow-brown, and continuation of chlorination at this stage will lead to excessive product decomposition due to a combination of high temperature and little light penetration. At this stage, after 700gms TCCA had been used in 7+ hours, the reactor was cooled (slight disconnection of all hot joints is recommended immediately) and the product distilled into a clean fresh reactor flask.



325 grams of a clear heavy-ish liquid were collected at this stage, with 31gms of dark condensation tars left in the still. A Beers law analysis of the IR graph below, based on the individuating absorptions at 586 and 629 cm-1 showed that this is about a 50/50 mix of the di and tri chlorinated products.



Benzotrichloride - step 2

According to US patent 3816287 a substantial amount of ring chlorination and tar product formation (32%) accompanies ordinary industrial chlorination to benzotrichloride. They suggest use of several stages, as well as low reaction temperatures and dilution of Cl2 with inert gas in the final stage to achieve a product pure to the 0.2% level. They claim to use only a 30% Cl2 excess (in their case actual generated chlorine rather than TCCA based). I have repeated their procedure, but found the use of inert gas (argon in my case) unnecessary. Indeed reduction of Cl2 flow rate should achieve the same effect, with the diluting medium provided by the chlorinated toluene vapour pressure. The only disadvantage of this is a lower rate of HCl efflux, however the reaction is essentially right shifted and a build up of products has little effect on the equilibrium. Hence towards the end I turned off my Ar flux.

Chlorination, which commenced with the liquid temperature held steady at 170-180C, was accompanied by a dense white fog which the Cl2 flux generated, and the liquid phase turned a ligh yellow due to Cl2 dissolution. It never however reached a brown colour in this case. Some Cl2 was now evident in the scrubber and reflux condenser. At 2hrs the temperature was reduced to 150C. At some point the white fog disappeared, and the Cl2 now bubbled in a clear yellow liquid - this indicates the end of reaction. The reactor was cooled, and the intial charge of 273gms was found to have gained 37gms weight. The product was now light yellow, and upon exposure to the air on days with humidity > ~50% fumed.



This weight gain is slightly in excess of theoretical and is due to some dissolved Cl2, since some evolution of yellow gas was observed at the intial stages of redistillation of this product at the beginning of the next procedure. The product was IR analysed, and is seen to be effectively pure benzotrichloride.




Benzoyl chloride and phthalic anhydride synthesis from phthalic acid

At this stage my aim was to use the pure product from step 2 above in Kyride's procedure (frequently advocated by Sauron) to simultaneously synthesise benzoyl and phthlyl chlorides. I redistilled the PhCCl3 as Kyride recommended, and halved all the amounts he used in the synthesis.





The 'phthalic anhydride', and zinc chloride were purchased from a reputable distributor.

A key feature of this synthesis is that all apparatus and reagents must be absolutely dry - every gram of H2O present initially will 'destoroy' 5.4gms of PhCCl3, converting it to the unreactive benzoic acid. The ZnCl2 presents a special problem being hydroscopic, while its Lewis acid properties (which make it useful as a catalyst here) also mean it can not be well dried of moisture once absorbed. When heated beyond a certain point it releases HCl rather than H2O forming zinc oxychloride

ZnCl2.x H2O -> (1-x)ZnCl2 + x ZnOHCl + x HCl

It is possible to dry hydrated ZnCl2 in a stream of HCl (or reacting with thionyl chloride), but these methods are laborious in light of the fact that benzotrichloride will perform the same drying function itself. A key point is that not too much moisture must be present as to consume above a few percent of the benzotrichloride, else the stoichiometry will be altered. As a consequence 10gms ZnCl2.x H2O were heated till melting (this occured at 240C as opposed to 275C for the pure substance) and subsequently held at 300C for 1/2 hour (zinc chloride sublimes substantially). This resulted in a loss of 0.5 grams, assuming this to be entirely HCl (the worst case) we have x=0.17 in the above formula.

The commercial phthalic anhydride was dried for 1hr at 120C and then held over CaCl2 overnight. Quite a bit of white sublimed 'fluff' formed during this procedure, and this together with the odour (which is quite sharp for the anhydride) gave the impression the sunstance was pure. The reagents are shown in the picture below.



Since the catalyst is little soluble in benzotrichloride, and the stoichiometric anhydride also does fully dissolve, this is a two-phase reaction, and hence efficient stirring is required. Due to the substantial quantities involved, this was provided by an overhead motor. To reduce product loss (this was subsequently well justified) a reflux condenser was attached, even though the recation temperature of 120-130C is well below the b.p. of any of the reactants/products. Being moisture sensitive a CaCl2 tube was used at the exhaust. The complete setup is shown below.



Once heating was turned on, a vigorous reaction with frothing and gas-evolution set in at 70C. Large quantities of gas, which quickly condensed moisture and completely filled the fumehood with a white fog appeared. A yellow liquid started to condense and drip from the ceiling of the hood - analysis showed this to be conc. HCl. After 1hr the reaction died down, further stirring produced no visible results. The fan cleared the fumehood. Surprisingly some undissolved and unmelted flakes of what was now obviously unreacted phthalic acid remained on the wall of the flask even at 130C. I analysed the suspect anhydride, and comparing the result with the known spectrum of phthalic acid one can see them identical - except for small peaks of a minority substance, such as at 1851 cm-1. This in fact is a major absorption of phthalic anhydride. Application of Beers law shows the commecrical anhydride had hydrolised on storage to now be 90% phthalic acid and 10% anhydride - see spectra below.



After cooling a solid slurry set it the flask. The reaction mixture was now fractionally distilled at low pressure through a Hempel collumn with 5cm of glass beads, since vapour assumes so much more volume at reduced pressure and bumping was substantial. 99.9 grams of a clear liquid collected, all at 78-80C see picture




An IR analysis showed this to be essentially pure benzoyl chloride



There was (initially surprisingly for me) almost no intermediate fraction, but at about 138C a white solid started to come over. I managed to collect about 20gms of this before the condenser blocked.



Analysis of the solid showed this to be relatively pure phthalic anhydride. It clearly contained a slight acid chloride impurity as it fumed and lacrimated the eyes.

Phthalyl chloride synthesis from phthalic anhydride

Having now obtained the phthalic anhydride I thought I initially had, I ground the quantity remaining in the flask (tared brown by PhCCl3 polymerisation byproducts) mixed with the white distilled anhydride, and reacted 60.2 grams of this with 84.1gms of fresh benzotrichloride, adding 3.2gms ZnCl2. 0.17 H2O as catalyst.



The reaction setup was as before, however I used a 250ml flask due to the smaller quantities. There was some HCl evolution at 70C due to the moisture in the ZnCl2, but this released no more than a fraction of a gram, after which all obvious signs of reaction ceased. After 12hrs of stirring at 120-130C, the reaction vessel was cooled, and this time no solids formed.




Fractional distillation yielded 54.1gms of BzCl which all came over at essentially 82-84C - indicating purity, there was an intermediate fraction of 6.4gms in the 86C-110C temperature range. Almost no product came over untill 132C. The final fraction was collected in the 132C-136C range, with 90% comming over at the fixed temperature of 136C.




Unlike Kyride no phthalic anhydride crystals condensed from the liquid on cooling, and analysis showed this to be almost pure phthalyl chloride




Attempted thionyl chloride synthesis from benzotrichloride and SO2 - no reaction

As mentioned in the introduction this seemed a viable reaction. Some refernces in the literature - patent 3875226 - indicated 58% yields are possible at temperatures of order 250C and pressures of 30-50 atmospheres, with the SO2 being loaded into the reactor in liquid form below -10C. Obviously this is a far more complicated route to SOCl2 than I had already had success with and so unpromissing, especially in light of the low yields - which are to be expected as SOCl2 is unstable at high temperatures. Nonetheless Kyrides obtained 66% yields from phthalyl chloride. Due to the lower b.p. of the benzotrichloride such a temperature is unsuitable in the presnt case, however the reaction, with dried SO2 was carried out at temperatures up to 165C and with fast SO2 evolution (I passed a 4-fold excess of SO2, about 1.2 mole through 63.4gms PhCCl3).



At the end of 5 hours no condensate had gathered despite the use of a very efficient condenser. Some small amount of condensation in the connections was analysed and found to be a roughly equal mix of benzotrichloride, and benzoyl chloride, formed by hydrolysis of residual moisture in the zinc chloride.

Health issues

The major concerns for this practical are the same as those already described in the benzaldehyde thread, and the same precautions should be taken. In addition benzoyl chloride is far more lacrymatory than any of the chlorinated toluenes, but has none of the after-effects of benzyl chloride.

Conclusion

Excellelent yields and purities of benzotrichloride, benzoyl chloride, and phthalyl chloride have been obtained in this experiment, starting with easily obtainable materials. A picture of my blackboard on which I kept track of results during this prac is shown below - as a memory.



[Edited on 10-5-2008 by len1]



[Edit by mod: image restoration]

[Edited on 19-8-2022 by j_sum1]

Again, I would recommend a base bath soak!

It gets rid of most of the organic residus, even sulfide smells disappear overnight, tars, stains that seem embedded in the glass, etc. A simple copious rince with water, and your glassware is perfectly clean.

This also works for old pyrex baking dishes but you have to do that with a freshly made base bath, to avoid any exposition to metals salts, or various toxic organic products that accumulate. The condensation products from the ethanol denaturants are easily washed off, a final wash in the dish washer, and wife is happy and thankfull

You can prepare you base bath by adding IPA or EtOH to conc. NaOH/KOH solution (ratios are availble on thsi forum or elsewhere on the net), to be occasionaly topped off with a little more alcohol to compensate evaporation. I use a 30L plastic tube, half full, with a firm lid. You only need to change it every 6-12 months depending on how much you use it. It's a good thing to wash the glassware with acetone before. After a certain period, the base (NaOH or KOH) gradually turns to insoluble carbonates. The mixture quickly turns brown/black afetr preparation, from the condensation of the alcohol denaturants, but it's not a problem at all.

You need full facial protection, and long thick base-resistant gloves to fish out the glassware.

The bath itself can represent a certain hazard, so always keep it in a corner on the floor, with a tight lid, under a cupboard or similar, far from the reach of children, animals or incompetents.

I never could go back to usual scrubbing since I've done my first base bath, and I will never praise it more than enough

Awesome Write-up

I really enjoyed your post. Thanks