PSF90−80
CHAPTER 6 Conclusions andfuture scope
NMP as solvent, PEG of 35 kD molecular weight as additive, aqueous NMP solution as bore
liquid, and water as coagulation bath. Particular fbcus was placed on NMP concentration of
bore liquid and AG conditions on membrane structure in order to obtain the gradient structure.
Characterization ofthe obtained membranes was perfbrmed by measuring pure water
permeate flux, tensile strength at break, and tensile elongation at break, and by analyzing
SEM images ofhollow一且ber cross sections, outer sur飴ces, and inner surfaces.
The required flow rate fbr bore liquid was highest in the case of using water as bore
liquid, and decreased with higher bore liquid NMP concentration. These results indicated that
there was more difffUsion of water from the bore liquid into the inside of nascent hollow fiber
form the inner surface toward the outer surface as water concentration of bore liquid became、
higheL It was clear that the ratio of the stnucture which is formed by bore liquid decreases
with lower water concentration of bore liquid, and that the structure is mostly formed from
outer surface in the case ofusing 70 wt%NMP solution as bore liquid.
Moreover, as AG distance shortened from 400 mm to 200 mm with the same 70 Wt%
NMP bore liquid concentration, the ratio of membrane fbrmed by bore liquid decreased
触heL It was clear that membrane was mostly formed from the outer surface in the case of
AG distance of 50㎜or less. A廿empts to use 75槻%N鯉solutions were unsuccess釦l
when AG distance was 400㎜, as nascent hollow丘ber would break apa曲ithin the AG
However, when AG distance was 50 mm or less, stable spinning of hollow fiber could be
achieved.
I The membrane made with O wt%NMP solution as bore liquid had a flux of 143
1・m−2・h}1,and that made with 60 wt%of NMP solution as 1)ore liquid had almost the same
CH41)7ER 6
Co刀cleny io nsα〃dfuture 50qρθ
flux at 1551・m−2・h−1, within the normal range fbr UF perfbrmance. Flux of the membranes
made・with・65 Wt%of NMP solution and with 70 wt%NMP solution j umped sharPly to 4,821
1・m 2・h−1and 7,8821・m−2・h 1, respectively, within the n・rmal range f・r MF performance.
SEM analysis revealed that pores were not observed in the inner surface ofPSFOO−400
and PSF60−400,0btained by using O wt% and 60 wt% NMP solution of bore liquid,
respectively・However, larger pores were fbrmed on the inner surface with bore liquid
concentration of 65 wt%NMP or more. The pore size of PSF65−4000btained by 65 wt%
NMP solution of bore liquid was bigger than that of PSF70−4000btained by 70 wt%NMP
solution of bore liquid. We considered this to be the result of a greater degree of ripPing and
pulling corresponding to a lesser extent of membrane fbrmation by bore liquid.
Pure water permeate且ux of the obtained hollow−fiber membranes with 70 wt%NMP
solution as bore liquid decreased dramatically when AG distance was shortened from 400 mm
to 50㎜. This phenomenon was explained as phase separation of outer surface of nascent
hollow fiber discharged from the spinneret being initiated by adsorption of water from the AG
and nucleation and growth of polymer lean phase progressing in accordance with the length of
AG residence time.
When AG distance was shortened from 400 mm to 200 mm, membrane formation
occurred almost completely from the outer surface. This meant that because the residence
time in the AG decreased by half, coagulation from the outer surface began earlier, so that the
ratio ofmembrane formation from the inner surface decreased.
The smallest pore layers obtained with each set of spiming conditions were located
somewhat below the outer surface・The distange from this layer to the outer surface increased
CHAPTER 6 Conclusions andfuture scope
with longer AG distance. The boundary between the region where NrPS did not occur and the
region where NPS occurred is the region where phase separation and coagulation occurred
after immersed in the coagulation bath, and this region became the smallest pore layeL As
fbrming the de sired gradient structure requires the smallest pore layer to be proximate to the
outer surface, it is apparently necessary to optimize AG humidity.
ln order to fabricate a P SF hollow−fiber MF membrane having a丘lter layer in the
region of the outer surface and gradually increasing pore size across the thickness ofthe
membrane wall toward the inner layer, it was necessary to raise the propo而on 6f membrane
formation from the outer surface with respect to membrane formation from the inner surface.
It was possible to fabricate above P SF hollow−fiber MF having gradient structure by selecting
bore liquid NMP concentration of 70 wt%or more and using AG distance of 50 mm or
shorter. Although the outer surface pore size of each species was too large for MF application,
the smallest pore size layer was located in the neighborhood of the outer surface.
Chapter 5 fbcused on the fabrication method of practical hollow fiber MF membrane
having a gradient structure, more speci丘cally, a filter layer in the region of the outer surface
and gradually increasing pore size across the thickness of the membrane wall toward the inner
surface. A 3−component dope composition of PSF/NMP/PEG was used with 35 kD Mw of
PEG as additive. NMP solution was selected as bore liquid. The necessary NMP
concentration of bore liquid and AG temperature in order to obtain a practical PSF hollowfiber MF membrane having this gradient structure were studied.
All the membranes made with AG temperature of 80°C, AG distance of 50㎜, and
bore liquid NMP concentration changing from 70 wt%to 95 wt%at 5 wt%intervals, show
CHA1)TER 6
Conclusions andfuture scope
similar water permeate flux at around 5,0001・m−2・h−1, in the normal range fbr MF
perfbrmance. Pure water permeate flux of hollow−fiber membranes obtained with 95 wt%
NMP solution as bore liquid at the same 50 mm AG distance and various AG temperatures
from 40°C to 80°C was affected by A G temperature, with flux increasing sharPly with higher
AG temperature.
The inner surface structure transfbrmed dramatically as NMP concentration of bore
liquid was changed. These inner surface structures can be classified into three categories
based on their appearance. The first of these categories is that of a lacerated appearance of
pore structures as seen with 70 wt%and 75 wt%NMP bore liquid, the second is that of a
circular pore structure as seen with 80 wt%,85 wt%, and 90 wt%NMP bore liquid, and the
third is that of a netWork structure as seen with 95 Wt%NMP bore liquid.
Formation of the lacerated appearing pore structure is considered to have been by the
mechanism described by Ohya et al.[1], being that the se structures result from the formation
of a very thin layer by water of the bore liquid after discharge from the spirmeret, this layer
then being partially ripped and pulled apart in the AG under drawing tension.
Formation of the circular pore structure is considered to be the result of a weak
coagulation power of bore liquid with NMI concentration of 80 Wt%or more, which was too
weak to solidify the inner surface, so that only nucleation and groWth of polymer lean phase
by liquid−liquid phase separation occurred during residence in the AG
The network structure as seen with 95 wt%N]Vll} bore liquid is considered to be the
result of membrane formation occurring only from the outer surface because the bore liquid i s
CH 4PTER 6
Conclasions andfuture scope
in the solvent neutral state, neither血duch19 Phase separation and solidification nor dissolving
the dope.
Regarding the outer surface, it was revealed that outer surface pore fbrmation was
affected by both AG temperature and NMP concentration of the bore liquid. Circular pores
were observed with 70 wt%NMP as bore liquid at AG temperature of 80°C. However, an
irregular, lacerated structure was observed with 75 wt%and 80 wt%NMP as bore liquid, with
the degree of laceration becoming larger as NMI concentration of bore liquid increased from
75Wt%to 80 Wt%.
This phenomenon is analogous to the mechanism of an irregular, lacerated structure
forming on the i皿er surfac e. Whether a lacerated structure will be formed on the i㎜er
surface or on the outer surface depends on whether a thin layer forms first on one or the other
surfac e in the Aq and whether this thin layer has become so inductile that it is torn apart as
the nascent fiber extends under drawing tension.
Pore structure of the outer surface also transfbrmed dramatically as AG temperature
was changed. This phenomenon is considered to result丘om the progress of phase separation
varying with the amount of water vapor in the AG as AG temperature is changed. It was
fbund that a solvent neutral state was attainable by selecting 95 wt%NMP solution as bore
liquid with our temary dope system, and that it was possible to obtain the desired gradient
structure with membrane fomiation only from the outer surface. Pore size of the且lter Iater in
the region of outer surface could be controlled by a(恥sting AG temperature and/or AG
humidit)乙hl particular, MF−level pores were obtained by setting the AG temperature to 60°C
or more.
CHAPTER 6
Conclus ions and/inure s cope
Figure 6.1. shows typical characteristics of polysulfbne hollow fiber MF membrane
in this study・The hollow−fiber MF membrane obtained in this study had a gradient structure
with an average pore size of O.23μm on outer surface. Pure water permeate flux was about
4,9001・m−2・h−1at 100 kPa. Retention of T500 was O%and that of O.137μm latex 1)eads was
almost 100%. Tensile strength at break was over 4.O MPa, and tensile elongation at break was
over 80%. These characteristics and performance figures are sufficient for use in MI? at the
practical level, and such membranes would be well suited fbr application in water treatment
and other fields.
Figure 6.2. shows the relationship betweell the newly developed PSF hollow−fiber
membrane based on this research and other membranes based on conventional technology.
This method led to a new development in MF membrane technology which outstrip s the water
permeability of conventional hollow−fiber MF membrane.
CHAPTER 6
Con(rlusions andプ勧 z〃「e scoρe
Inner su rface
●
●
●
●
●
●
●
Cross section
Inner di訊meter Outer diameter
Pure water permeate flUX
Rejection of Dextram T500(500 kDa)
Rejection of O.1μm latex bead Tensi童e strength
Tensile elongation at break
Oute「surface
=0.65mm
=0.97mm
=4952・LMH@100 k Pa,25°C
=0%
=97%
=4.2MPa
= 87%
Figure 6.1 PSF hollow−fiber MIF membranes having gradient structure achieved in this study.
CHA1)TER 6
Conclusわnsα刀61ノ諺∫z〃「θscoρe
(蓋母蕪∴)蓑慧
10000
1000
100
10
1 10 10e 1000
Pore size (nWt》
Figure 6.1 Newly developed PSF「hoHow−fiber(HF)membrane based on伽s research and other membranes・based・on・conventiona瓢technology.
CH 4PTER 6
Conclusions andfuture scope
6.2Future scope
ln this study the temperature of the coagulation bath was a(ljusted to the same
temperature as the AG Therefbre, no distinction was made between the influence on
membrane formation of coagulation bath temperature, AG temperature, and AG humidity.
More precise understanding of membrane formation mechanism in the AG particularly the
pore fbrmation mechanism on the outer surface, requires more precise control of the
temperature and the humidity of the AG independently under a certain constant coagulation
temperature. Moreover, a quantitative experiment conceming the amount of water vapor
absorbed by the nascent hollow fiber in the AG is necessary.
List of ublications
lnternatiOnal j OUrnal
No. Title
Joumal Page
Y6ar Co−Authors1 Fabrication study of polysulfbne hollow−fiber
Jo㎜al of
293−3022009
S.Shiki microfiltration membranes:Optimal dopeMembrane
且.Kawakamiviscosity fbr nucleation and growth Science
2 Fabrication of polysulfbne hollow−fiber
Joumal of
in S.Shild{microfiltration membranes:Optimal spir血ng
Membrane
submissionH.Kawakami
coIldition fbr gradier岐structure, part 1 Science
3 Fabrication study ofpolysulfbne hollow−fiber
Joumal of
in S.Shiki microfiltration membranes:Optimal sph血ngMembrane
submissionH.Kawakami
condition fbr gradient structure, part 2 Science
SupPlementary Publications
No. Title
Joumal Page
Year Co−Authors1 ポリスルホン中空糸精密ろ過膜
高分子
P4522008
Intemational confbrence
No. Title Cor[飴rence
Page
Year Co−Authors1 Fabrication study of polysulfbne hollow」fiber
高奄モ窒盾?奄撃狽窒≠狽奄盾氏@membranes
Nor血American lembrane
rociety
P.98
2007
S.Shikig.Kawakami
2 Fabrication study of polysulfbne hollow−fiber
高奄モ窒盾?奄撃狽窒≠狽奄盾氏@membranes
Intemational
lembrane
rcience and sec㎞ology bonfbrence
P.71
2007
S.Shikig.Kawakami
3 Fabrication study of polysulfbne hollow−fiber
高奄モ窒盾?奄撃狽窒≠狽奄盾氏@membranes
Ir岐emational bongress on lembranes and
lembrane
orocesses
P.524
2007
S.Shikig.Kawakami
Patent
No. Title Patent No. Y6ar Co−Authors
1 ポリスルホン系樹脂膜、及びその製造方法 JP
n3,046,086 1,991
上坂正利
2 ポリフッ化ビニリデン樹脂、及びその製造方法 JP
n3,093,811 1,991
渡辺幸平
3
親水性耐熱膜及びその製造方法
JPn3,200,095 1,991
4
新規な構造を有する中空糸膜、及びその製造方法
JPn3,167,370 1,991
大野香子
5
高透水半透膜の製造方法
JPn3,169,404
1991 上坂正利
6 中空糸状高性能精密ろ過膜 、 JP
n3,217,842
1992 上坂正利
7 Hollow fiber type filtration membrane
US
U,165,363 2,000 T.Ohishi
8
Manufacture of composite film fbr separation of 唐oec置ic components, involves contacting porous 唐浮垂垂盾窒煤@having zeohte microp articles, with synthetic 唐盾撃浮狽奄盾氏@containing raw material of zeohte, and
?b窒高奄獅〟@zeohte crystal」[ilm
WO Q007JP58272A
2,007H.Shj−ataki
y.Wang j.Aoki
9
Porous membrane used fbr filtration in manufacture 盾?@drinking water, and ajヱpurj五cation, contains
垂盾撃凾魔奄獅凾?р?獅?獅浮盾窒奄р?C and has preset value with 窒?唐垂?モ煤@to product of crystalhnity degree of polynler
≠獅п@specific surface area
WO Q007JP316250A 2007
、
r.Shiki m.1(ubota l.且attori
Ackn o wledgem en t
This dissertation draws together my study results at Asahi Chemical Ihdustry Co. Ltd.,
(now Asahi Kasei CorP.)in 1990, with all data reconsidered and reexamined under the
guidance of Prof Hiroyoshi Kawakami of Tokyo Metropolitan Universit)1. Given that I was
transferred to work in my present position around the time that I enrolled in the doctoral
course at Tokyo Metropolitan University(TMU)as an adult student, my work assignment at
the company changed dramatically. This made it a particular challenge to simultaneously
satisfy both my academic requirements and my workirユg responsibilities. It is thanks to the
strong leadership of Prof Kawakami that I have been able to make three presentations at
international conferences, submit three fUlI research papers for j ournal publication, and丘nally
to complete this dissertation despite these challenging circumstances. I ca㎜.ot overstate the
deep debt of gratitude I owe to Prof Kawakami.1 am also deeply gratefU1 to Prof. Kato and
Prof Yamato, also of TMU, fbr the invaluable advice they provided as I prepared this
dissertation.
Ialso owe my sincere thanks to the many people at Asahi Kasei who helped make it
possible fbr me to complete my dissertation。 I cannot possibly name them all here;however, I
must express my particular thanks to the fbllowing:
Mr. Kohei Watanabe, who gave me the time to conduct a basic study of pore−size
control technology fbr polysulfbne hollow−fiber membrane, the fbundation fbr my research,
when I completed the first stage of development of a polysulfbne−membrane artificial kidney,
which was my first job at the company. Moreover, he set me on my path of professional
development in the company as a speciaIi.st engineer in hollow−fiber spirming, and apProved
my study in a doctoral program at the University of Califbmia, Los Angeles(UCLA)fbr two
years fヒom 1996. I extend a special thank you to Mr. Watanabe.
1 am fortunate to have been able to leam so very much from Dr. Noboru Kubota, who I
have looked up to as a mentor since j oining the company. Notably, I was greatly inspired by
his earning a graduate degree based on his own research at the company. My profbund thanks
to Dr. Kubota.
Ithank ML Teruhiko Ohishi, who took over my job when I went to UCLA, and who
made foreign patent application based on the original study. It is thanks to his effbrts which
resulted in patent issuance that I was able to disclose my series of studies.
Adeep thanks to Mr. Ybshihiko Mori, who was my supervisor in 2005 and who
willingly gave his blessing to my idea of pursuing a doctoral degree on hollow−fiber spinning
technology at TMU as an adult student.
Iam profbundly gratefUl to Mr. Ybusuke Koizumi, my current supervisor, who made
generous allowance fbr my devoting quite a bit of time working toward my degree despite
pressing duties related to the start−up of a new plant, and who also gave me insightfUl advice
on the preparation ofthis dissertation.
Awa㎜thank you to ML S atoshi Shiki, who is a former subordinate of mine and a co−
researcher in this study. He cheerfUlly acquiesced to my many requests to run experiments in
my stead to obtain supplementary data and to conf rm reproducibility.
Ialso thank Mr. Charles Marken for transforming my crude English into elegant prose
and fbr finding and correcting mistakes血great detail. Without him, I could not have
completed this dissertation on such a hectic schedule.