A Study of Fluidised Bed Granulation

I\I: ei:peril1lr~n:al s+;udy 01' f l L!idis2d ~eci 0ranulatiu.l is pre~ent2d. Preliminury ex~e;ililentc, est?~,;lisn the mCl i 1-, variabl'_s in the granulation pr'ocess and er.able a systemc.tic cxperimcr.tal programme to be Jevised. In this programrre, f';ither glass p(jwdel~ Ol~ po,ous alum~na are used as the bed rrateY·ial. The ef'ft~cts of the f:.mdame!1tal fluiciised bed granulation parameters 011 rarticle growth are established from a series of batch e):periments conducted in a 0·15 m diameter; glass-v/alled bed. Togethp.r wi th rneas urerrli:~nts l' f the phys 'ica 1 properti es of the !Jroduct granules and the binder soluticns (carbowax or be~zGic acid, in m2thanol) and ':Jith supporting stUl:iies of the bed st:~uctlJre (in the form of X~Tay photography and m2v.surem~nt of temperature profiles), this ir.format-ion is also used co Pi'cpose a t:2cha.nism of particle yrovJth. Two types of produst granule ar~~ identified; agglomerates \'Jhich consist of tvw OJmore, c',ld usually sevsral, initi0.1 particles; and layered granules, which consist of single in~tial pn.i,·ticles VJith driea feed material adhering to the surface. The effects of \/arying thE: excess fluidising gas velocity. the binder concentration and the init'11 particle size ate quantified. A bed can be preventpd from quenching (a defi ui di sa ti on phenomencn lead; ng to the fan uY'e of the process) by increasing the fluldising gas rate. For a given bed 9article / blnder combination, successivelY higher excess gas velocities allow an otherwise quenching bed to be operated firstly so as to produce agglomerates ar.d sub~eo,uently layered gre,riJl~s. Simi12.~' effects a~'e observed \'!it~: increases in initial bed p~rticle size and decreases in binder concentr<1tion. fl. particle growth mechanism is proposed in which the initiv.l stages of both desirable particle growth (irrespect'iv2 of the type of gran 'Jle) and of bed quenching, are considered to be exactly the same. Beyond the i nit i a 1 formati on of 1 i q ui d bonds be tv:oen a dj acen t bed pa rt i c les, the strength of the inter-particle bridges (which is a function of the binder material and relates to growth) and the exteilt of fl uid drag ar:d inertial forces on particles (which are functions of gas velocity and particle size respectively, and relate to granu12 brGJkdow~) dE:teY'mine the equilibrium granu12 form and she. HO\'1ever, if the particles are porous the abOVE mechanism may b:"eak dovm since the liquid c~n enter the pores and not be available for the initial form~­ tion of liquid bonds. X-ray photography sholtiS that the mode of entry of atomisin;) air into the bed is by periodic bubbling. rather tha!1 by continuous issue from a jet of air. thus throwing considerable doubt upon the traditionally pY'oposerl mechanism of gr0vJth by particle circulation through a high voicage lone. Neverthe1ess. the temperflture measurefT'Gnt:; ir.dicate that 0. substantial pOition of the bed belJw t.he atcmising nozzle is at a significantly lovier' t~mp2ratllre than the l"emaind2r_ The experimental particle gtowth data for both types of ~roc!uct granule fit simp12 geometrici3.1 models. In 1:he cuse of ag~lorreratior:, a re lati onshi fJ is estab 1 i s(:·;:j DeblE'en average gl~anule s~ ze and bi nd2i" conten i~.