[phth]

http://www2.chem.rochester.edu/~nvdtest/test/pages/how-to.php?page=run_prep_tlc  <<Will save you time.  Will get you intermediates without losing them on the column, and saves time wasted doing a useless column.  Do they mention anything about lighting conditions?  Spike your crude NMR with simething that has low signal to weight ratio.  Then you can rule out if you're losing your s#*$ to the silica

[darkdevil]

Just a little update to this,
I changed my xylene to o-xylene, which is the highest boiling point member in the xylene family. The reaction seemed to work better. After 1 day, there were less side products. Last time I refluxed the reaction for 2 days in m-xylene and obtained very less product. I am now trying to reflux it in o-xylene for 2 days and see what will happend.

At the same time, since the reaction involves a condensation reaction (water elimination) as the final step, shall I remove the nitrogen purge for the condensation to go better? I am not sure if water elimination can occur in a refluxing reaction in an enclosed nitrogen atmosphere.

[phth]

We have to remove water form the solution which is why I think they chose Xylene.  It boils higher so water can be distilled because it boils at a lower temperature.  Not doing so its allowing other mechanisms to take over, and increasing the time just makes it worse.  H₂O adds to the reagent etc. Reflux keeps re-introducing the high % water distillate back into the reaction.  How would you go about modifying a reflux setup to accomplish this goal?

[discodermolide]

Try using a Dean-Stark apparatus  to remove the water. This will stop it re-entering the RM and causing problems.
https://www.google.ch/search?q=dean-stark+trap&safe=off&biw=1890&bih=918&tbm=isch&tbo=u&source=univ&sa=X&ei=sRINVZH-DM_hasCpgugK&ved=0CCQQsAQ

[orgopete]

Now that we see the product, this is helpful. In this case, dropwise is important. This is how I would run thus reaction. I'd set up the reaction with a Dean-Starke trap to separate the water formed. I'd start the reaction at reflux and slowly add the acid chloride as indicated. Id also be inclined to add a catalytic amount of pyridine to the reaction to catalyze the acylation and potentially the cyclization. Because water is formed during the reaction, I'd try to keep the acid chloride concentration low as well.