Sulfur melt data smogasbord

21 Jan
Published by RyanPakula

I've been working on the sulfur melt reaction, trying to optimize the reaction yield of the singly dehydrogenated product (DHP below).  Many studies have been conducted, and here I'll discuss some of what I've found out.
 
Many methods for separating PZQ and the dehydrogenated product were investigated, but in the end a trace contaminant (not visible by proton NMR) in the commercial PZQ is present underneath the product peak.  That is, the retention time is identical, no matter what type of method I use for analysis.  This prevented 100% accurate yields to be determined by GC, but the peak was determined to be consistent in size with respect to the PZQ, so comparisons were still viable between GC runs.  Thus, this 'quick and dirty' method was used for analyzing the PZQ sulfur melt reaction.
 
The method employed was simply a ten minute run at 320ºC on an HP-1 column.  The product (DHP) elutes from the column around 8.4 minutes, with the PZQ starting material around 7.2 minutes.  Another product - believed to be formed from the singly dehydrogenated product, and, thus, nicknamed di-dehydrogenated PZQ or DDHP) - appears around 4.2 minutes.
 

 
Thus, monitoring the reaction with this method, the following data/facts were obtained:
- Using 0.88 eq. sulfur to PZQ leads to 50.0% production of DHP, with 41.5% PZQ remaining (and 8.4% minor products, including some DDHP).
- Adding sulfur in stepwise additions does not seem to effect further production of DHP, consumption of PZQ, nor prevention of formation of DDHP or other side products.  When added in three discrete steps (0.325 eq., 0.378 eq., and 0.295 eq.; total of 0.999 eq.), DHP yield was 49.7% by the end, with 42.0% starting material remaining.
- The optimal equivalency of sulfur to PZQ is between 1 and 2, although I didn't finish honing in on the precise value.  In a study with 0.50x, 1.01x, 1.37x, 1.98x, 4.01x, and 9.80x, the 1.37x sulfur to PZQ reaction had the greatest yield of DHP while also minimizing the amount of DDHP and other side products formed.  A successive study with 1.15x, 1.27x, 1.39x, and 1.51x displayed 1.51x as the most promising result, although all these reactions had very similar results.  A neat reaction like thi sis unlikely to yield very clear results due to poor mixing during the reaction.  If we can get a solvated reaction to work well, precise optimization studies will be more fruitful.

 
Some facts that became apparent include:
- Doing the reaction neat wtih an internal standard of either 1,2-dinitrobenzene or 1,3-dinitrobenzene leads to complete loss of the internal standard.  We aren't sure how the internal standard was disappearing, considering the higher melting point of it than both reagents, unless, of course, all three components are evaporating.  We cannot think of any way in which these standards could react with the sulfur or PZQ.
- Better separation can be achieved with a ramping method, such as starting at 250ºC for 3 minutes, ramping at 5ºC/min to 320ºC and holding.
- Performing the reaction in DMSO, heating to reflux, yields no reaction, although the internal standard is preserved.
- Bubbling nitrogen through the raection mixture (i.e., when DMSO is used as solvent) does not effect further reaction (by driving off formed hydrogen sulfide gas, for example).
- Heating PZQ without sulfur in the same manner and the sulfur melt reactions does produce some minor product peaks.  Dioxygen gas may be the oxidant in this case.
 
From all of this, so far the optimized reaction seems to be using ~1.5 eq. of sulfur in a neat reaction.  Unfortunately, this mostly just confirms what we suspected before.

Comments

ndt228's picture

I would think it likely that your nitro compounds are being reduced by the molten sulphur.  Perhaps a dimethoxybenzene might be more stable ?  Or a naphthalene derivative if you need a higher boiling point ?
Also, you didn't describe your reaction conditions - what temperature is the reaction run at ?  How do you work up samples before analysis ?
Concerning the "mystery impurity" that always runs in the same palce as the product - have you considered the possibility of "chemistry" occcurring in the gc injection port ?  Depending upon what is (left in) the liner it is possible that either thermal or chemical change might be converting a small part of your SM to product, upon injection.
Best wishes
Nick