| Summary: | 碩士 === 國立清華大學 === 化學工程學系 === 85 === The objective of this work is to reduce the pore size and
improve the surface roughness of the porous plate for making
membranes with smaller pores . The sol-gel method was usually
adopted to prepared the membrane for gas separation industries.
Cracking ,however, is the major problem to affedt the quality
ofthe membrane , and the surface morphology of the support is
one of the causes .A membrane often cracks upon drying if the
pores are not uniform. TiO2 particles of 0.5um and 30nm were
dispersed in water with Darvan 7 and then dipped in sequence
onto the support and water was filtrated out by applying vacuum
on the down stream side . The first and second intermediate
membrane layer were formed with mean pore sizes 0.5um and 30nm,
respectively. The main job of this work is to investigate the
relationship amoung processing variables - microstructure -
properties of the membrane and condition for producing the
optimum products. TiO2 particles of 0.5um can be well dispersed
in 200ml D.I water with 1ml Darvan 7 (0.654 Vol%) to show about
90% particles around 0.5um from LPA analysis. However , the 30nm
particles show about 45% at 50nm indicating agglomeration
between particles . The grain growth of the sintered TiO2
particles is affected by sintering temperature and the
coordination number amoung the particles .Grain growth will
accompany with increasing pore size. The surface roughness ofthe
membranes will be greatly improved as shown from SEM , AFM and
a-Step analysis .A flat surface will be advantageous to
subsequent sol-gel process t o further reduce pore sizes.
As for the microstructure ,the membrane thickness is linear with
the quantities of TiO2. The mean pore size (0.40→0.125 um) and
overall permeability (420x10-7→160x10-7) (mol/sec-Pa-m2))
decreased , while th porosity (23→35%)is increased with
thickness . If compared with commercial products , our membrane
can offer smaller pore size , yet similar fluxes. When thporous
support is tested as a bubble device ,we found that the mass
transfer coefficent (KL.a) is higher than commercial products
made of sintered glasses , especially in the low gas flux region
(Ug < 1cm/sec) .This mainly due to the very small bubble size
associated with our porous support. When tested in crossflow
operations , both the support and the membrane show decaying
flux with time ,suggesting that particles are gradually plugging
the support. Nervertheless , the adhesive property between
support and membrane is quitesatisfactory.
|
|---|
