Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
Z41;Z~
This invention reIates to microporous membrane filters
composed of polyesters derived from aromatic dibasic acids,
especially polytetramethylene terephthalate and to a new
pxocess for forming microporous polymeric membrane filters.
The new process, referred to as a polymer assisted phase
inversion, obviates the necessity for a surfactant in the
membrane and represents a new approach to the formation of
microporous membrane filters.
The technology of synthetic polymeric membranes is quite
well developed in many areas and the literature on the subject
is profuse~ Probably the best and most comprehen~ive treat-
ment of this subject is the recent work by the present invent- ;
or, Dr. Robert Kesting, entitled SYNTHETIC POLYMERIC
MEMBRANES, McGraw-Hill Book Company, 1971, Library of Congress
catalog card No. 72-132345. The reader is referred in parti-
cular to Chapter 5 of this work entitled "Porous Phase-
Inversion Membranes" for a discussion of the principles of
membrane formation.
Polyester resins and the terephthalate polymers are, in
.
genaral, well known. Polyesters were described by Carothers
and Arvin, Journal of the Americal Chemical Society, Volume ~
51, pages 2560-70 in 1929. See also collected papers of ; '
W.H. Carothers on High Polymeric Substances, ~olume 1 HIGH
POLYMERS, INTERSCIENCE PUBLISHERS INC., New York, 1940. Poly~ ;
ethylene terephthalate films have been marketed for several
years under the trademark MYLAR by DuPont and other polyester ;~
materials in various forms have been available for several '~
years.
Until now, however, microporous membrane filters of the
polyesters have not been available. One of the principle
:"",,;
', `
. , . .. . ,,.. . , ~ . .
.. -. - , . . - .. .. :: . ..
~01~2~Z - :
features of this invention involves' microporous membrane
filters composed of polyesters 'derived from aromatic dibasic , '
acids. One of the more'specific features' of the invention is
the discovery and preparation of microporous membrane filters
composed of polytetramethylene terephthalate.
.. .. , . ~ ...
Another principal feature of the present invention is the '
provision of a dry process for forming microporous membrane ~
filters of aromatic dibasic acid polyesters in which membrane ~,
formation is accomplished without the necessity of the usual ~ ,'
. :
low molecular weight swelling agents and wetting agents. '' '
:':
::
~ ~ :
,' ,
;;
: .
`' ~',', '. ~', ~
;~.` ~,,:'
.
''., ''
3~ ;,,' ;
~ ~ '
-la~
-: :
~1624~2
the membrane formation being assisted by the inclusion of
a separate and distinct polymeric compound. This dry process
is called a polymer assisted phase inversion process for
forming microporous membrane filters of aromatic dibasic acid
polyesters and is particularly applicable to polytetramethyl-
ene terephthalate and equivalent polyester resins.
Within these broad and unique principal features of the
invention lie other significant features which will be dis-
cussed in detail in the specification hereinafter.
10Polyesters Derived from Aromatic Dibasic Acids
Polyesters which are derived from aromatic dibasic acids,
especially those derived from terephthalic acid, have excel-
lent chemical resistance, including resistance to acid and
base hydrolysis and resistance to a great number of organic
solvents.
In addition, these resins exhibit other outstanding `
properties, such as high tensile strength, flexibility, and
thermal stabiIity. For these reasons, this class of resins ;
represents an excellent a priori candidate for the formation -
of microporous membrane filters. Until the present invention,
however, membranes of these materials which exhibit the
structure of the "typical" microporous membrane filters, such
as those prepared from nitrate and acetate esters of cellulose,
and certain polycarbonates (see Canadian Patent 991,776, etc.),
have eluded workers in the field. ',~
Polyesters prepared from the condensation of dibasic acids `
and diols are well known as resins and possess the following ;~
general formula:
t1) ~~ RC-O-Rl-O-~ ~~n
wherein R is aryl and R' is typically alkyl, wherein n is an
~ -2- ~`
: .
'` J'` ~
~ Z~12 ;:
integer which may be as low as a few hundred to as high as
several thousand.
The only polyesters of interest in the present invention
are those in which the dibasic acid moiety is aromatic, e.g.
the benzene nucleus. The specific dibasic acid of the greatest
. . .
interest in the present invention is terephthalic acid. ~
. .:
. ,
~i'
i.
,~' .
,
-2a-
.;",,:
` ~06;~:41;Z ~ ~:
. .
Ortho- and Isophthalic acid derivecl polymer~ are regarded a~ -
equivalent onlv in the very broadest sense and only marginally suitable as
membrane formers in the present invention.
Polyesters of the class under consideration derived from tere-
phthalic acid have, within the context of the present invention, such unique
properties as to set terephthalate resins apart from other resins of this
clas s .
Diols which are suitable for the present invention are the aL'cyl diols
such as trimethylene glycol, tetramethylene glycol, pentamethylene glycol,
10 etc. Ethylene glycol derived polyester~ of the class under consideration are
regarded as only very marginally suitable, and equivalent to tetramethylene
.;`'.. ', ..
glycol derived polymer~ only in a broad sense, which would include usable, ;-;
but much less satisfactory membrane forming systems.
Generally, then, the most suitable polyesters are those having the
:i.
formula: ~
O O ' ,'
(2) ~(~)-C-O-(CH2)a -O-C ~n
wherein a is an integer from 3 to S, most preferably 4, and marginally as
low as 2 or as high as 6 or 7, and wherein n is an integer greater than about
20 45 (MW _ 10,000~ and preferably greater than 90 (MW -- Z0,000). The most
suitable available range of polymers of this class thus far proved ha~re an
n value within n = 90 (MW ~ 20,000) to n = 225 (MW ~--50,000), but the use `
of such polymers having an n value of up to 450 ~MW _ 100,000) or higher
will predictably permit preparation of High Void Volume membranes by the
dry process here disclosed, as well as by the process described in my co- I
pending patent application referred to previously. ~
.:,.
The range of effective equivalents, for purposes of this invention,
must, to some degree, be determined empirically, since a full range of
potential equivalents is not now available for evaluation and may not become ~ `
available soon enough to avoid undue delay in making this technology available
to the indu3try. r~'
1'''
I '~, '.
~ ,'~' .
)6;24~
... ..
Some guidline~ to potential equivalents are possible, however.
These are outlined briefly below.
~ :
Polyesters derived from ortho- or isophthalic acid are regarded as ~ `
~. .
potential, but significantly inferior, marginal equivalents suitable for some
appllcations if the solvating power of the ca~ting solution solvent is reduced ~
and carefully controlled. (For use in the dry process herein disclosed, the ~'t `'
solvent must also dissolve the assisting polymer, thus making solvent
selection critlcal and unpredictable). ! ~
Polyesters of the class under consideration having the general ~`
10 formula:
s O ,,
(3 ) ~ ~- C - O- ( CH2 ) a - 0- C
j, , .
wherein a and n are the integers as previously defined and wherein s is a
substituent which does not prevent polymerization of the polyester, e. g. a
methyl or other lower alkyl group (1-4 carbon atoms), are also potential
equivalents and should permit the more effective use of diols having 2 carbon ; `
., ~: . . ~ ~, .
atoms. Selection of a solvent having a lower solvating power, as discussed ` ;,~
previously is indicated for casting solutions of such polyesters, especially
where larger, or plural, alkyl substituents occupy one or more positions on ~ `~
.: . ... .. .
zo the aromatic dibasic acid nucleus.
Polyesters derived from fused polycyclic dibasic acids, for example - `
O ~ :~' '-
~4) HOC ~C OH
are not good candidates as equivalents, although the dry process described
herein could be used if a suitable solvent could be found.
Polyesters of the class under consideration derived from diols con- `
taining etheric linkage, for example
(5) Ho-c~I2cH2-o-cH2cH2
and
(6) HO-CH2CH2-0--~--0-CH2CHz-OH,
- 4 -
.-, i , . ..
.....
- ~6Z~2 ~
and from diols of high molecular weight, for example
(7) HO-(GH2) al ~OH
wherein a' is an integer of 7 or 8 or higher, are also not good candidates as
equivalents to the extent that membranes formed from such materials would
be less stable, less resistant and generally inferior and would require a
different solvent sy~tem. The dry process disclosed herein could be used,
however, if a suitable solvent could be found.
Obviously, not all criteria for equivalence are set forth, and not all
parameters are known. I do not, however, wi~h to be limited to those
materials specifically set forth in the examples, nor do I wish to ignore the
fact that one particular combination of materials (polytetramethylene tere- `
phthalate polyester, polyvinyl alcohol assisting polymer, and hexafluoro-
isopropanol) poæsesses such unexpected advantages as to be a distinct
discovery within the broader framework of this invention. `~
Until quite recently, the only commercially available dibasic acid
derived polyesters of the type under consideration here were the polyethylene
terephthalate polyesters prepared by the condensation of dimethyl tere-
phthalate and ethylene glycol. (Reference to dibasic acids and diols includes
ZO polymeri~able derivatives thereof.
Polymers of thi~ type are sold in various forms under several
trademarks including Mylar and Dacron. `
I had postulated that if the aLlcyl chain were longer than that pro-
vided by polyethylene terephthalate, polyesters of the type under consideration
would be excellent candidates for microporou~ membrane filter fabrication.
I have now proved my postulate correct and discovered that polytetramethylene
terephthalate prepared from the condensation of dimethyl terephthalate and `
1 ,4,tetramethylene glycol does indeed form microporous membrane filters
having excellent chemical and physical characteristics.
- 30 A sufficiently high molecular weight polytetramethylene terephthalate
i8 available from Eastman Chemical Company under the trademark Tenite
polyterephthalate,~rom General Electric Company under the trademark
_
-- 5 --
6Z4~2
ilalox / and ~rom the Celanese Corporation under the trademark
Celanex.
My familiarity with the process variables, the characteristics of
_____ :
the product~ obtained and the chemical and physical propertie~ of the
materials involved indicate with relative certainty that polyesters derived
from aromatic dibasic acids ~or polymeri~able derivatives, e. g. esters,
thereof) having 3 to 6 or 7 methylene groups in the aLkyl moiety of the
polymer can be fabricated into good to excellent quality microporous mem-
brane filters using the techniques and processes of this invention as described
hereinafter .
10:[n general, any polyester derived from an aromatic dibasic acid
which i9 insoluble in water, possesses acceptable physical properties (is - `
self-supporting at ambient temperature and is sufficiently dimensionally
stable), and is soluble within the range of 2% to 20% (by weight) in combina-
tion with an assisting soluble polymer which also has ~olubility in the 2% to
20% (by weight) range in a solvent which has a boiling point in the ambient
to approximately 150C range must be regarded as an equivalent in the
broadest sen~e, but not in the narrow sense of being interchangeable with
polytetramethylene terephthalate with like results in the same ca3ting solution
system. Indeed, I regard only polytrimethylene terephthalate and poly-
20 pentamethylene terephthalate polyesters as substantially certain to be closely
equivalent and functional in this latter sense.
The preferred polyester, polytetramethylene terephthalate, sold
under the trademark Tenite Polyterephthalate 6 PRO, has outstanding
chemical propertie~ illustrated by Table I which show~ the effects of various
chemicals on this grade of polytetramethylene terephthalate. .
% Increase After One Year ~ `
Chemical In Wei~htIn Thicknes~ ~
. .
Acetic Acid (Glacial) 2.65 1.51
Acetic Acid (5%) * O . 40 - O. 29
Acetone 5. 56 Z. 57 ~`
Ammonium Hydroxide (Conc. ) O. 80 -O. 16
Ammonium Hydroxide (10%) O.49 -O.53
'."`:
.~ ~ - 6 .
106Z4~LZ
:
Aniline 6.69 4.17
Benzene 3. 26 Z. 19
Carbon Tetrachloride 0.72 0.42
Chromic Anhydride (40%) 0. I0 -0.11
Citric Acid (10%) 0.37 0.16
Cottonseed Oil 0,09 -0. 16
Alcono~)Detergent (0. 25%) 0. 37 _0.34
Diethyl ~ther 0.95 0. 45
Dimethyl Formamide 2. 48 1.56
10 Deionized Water 0.39 -0 53
Ethyl Acetate 3,30 1.93
Ethylene Dichloride 18.66 6. 16
Heptane 0. 13 -0.26
~A Hydrochloric Acid (Conc. ) 0.56 0.03
Hydrochloric Acid (10%) 0. 33 _0.05
Hydrogen Peroxide (28%) 0.41 0.00
Hydrogen Peroxide (3%) 0.38 _0.05
Isooctane 0. 08 -0. 08
Keros ene 0. 10 - 0. 16
20 Methanol 1.61 0.16
% Increase After One Year .
Chemical In Wei~ht_ In Thickness
Mineral Oil 0. 06 -0. 13
Nitric Acid (Conc. ) -disintegrated-
Nitric Acid (40%) 0.68 0.05
Nitric Acid (10%) 0. 36 -0.16
Oleic Acid (93%) 0.26 -0.18
Olive Oil 0. 06 0.11
Phenol (5% in water) 9.48 4.88
30 Sodium Chloride (10%) 0.32 0.05
Sodium Carbonate (20%) 0.31 0.05
Sodium Carbonate (2%) 0.36 0.00
-- 7 --
.`
- . ~ . , . ~........... .
~L0~;2~1Z
Soc~ium EIydroxide (10%)-1.20 -1.16
Sodium Hydroxide ~1 %)O. 37 O. 77
Sulfuric Acid (Conc . )-disintegrated-
Sulfur ic Acid (3 0 %) O . Z5- O, 52 -~
.. ~ . .. ,,~ . .
Sulfuric Acid (3%) O.36 O. 00 ~;
Turpentine 0. 27 _0. 03
Toluene 2. 35 1. 40
Sodium Hypochlorite (3.5%) ~.37 -O.11
Ethanol 0.48 0. 21
10 Ethanol (50%) 0.47 0.03
Dibutyl Sebacate 0. 07 -O. 05
.:, .
Di-Z-Ethylhexyl Phthalate 0.07 `O.08
Ivory Soap ~1%) 0. 37 0. 08
Brake Fluid . - ,~
Transformer Oil 0. 06 -0.18
.
* All percentages are concentrations in water.
These membranes are extremely tough, flexible, dimensionallr stable over ;
a broad range of temperatures, maintain their strength at both high and low ~ -
temperatures and possess other highly desirable physical characteristics. ~ ~`
The Poly_ster Membrane Product `
... .
In terms of a manufacture or product, the present invention is
directed to a polye~ter derived from a dibasic acid, as discussed herein- `;
before, which is in the form of a microporous membrane. ';,
The microporous membrane of this invention i~ composed oE a
polyester derived from a dibasic acid and may include a hydrophilic second
polymer or a second polymer which is soluble in a fluid which is a non- ~; `
solvent for the polyester.
Typically, the membrane of this invention, in finished form, is
composed of a dibasic acid derived polyester, as discussed previously, and
30 a minor proportion of the second polymer, which is hydrophilic or soluble
in a polyester non-solvent fluid, or includes, as a component, such a
membrane composition.
:.
-- 8 --
" ._, . . . . .. . .. ,-. ----- . . . . . . . .
. ~6g4~Z :ThP memhrar.es here disclosed may he used as ~emhrane filters in
microfiltration? ultraf~ltratïon, in reYerse osmo~i`q~ as reYer~e osmosis
~embranes or`as por~us ~upport substrates: for ~the~ r~verse os~osis-membranes,
the cellulose ace*ate membranes, for example, and ~n electrophoresis, curtain
electrophoresis, thin layer, gas and liqui`d chromatography and as substrates
to which enzymes; and other catalysts can be attached for solïd-liquid reactions.
The porous membrane filters in this invention may be produced in flat sheets,
tubular forms on the inside or the outside of the tube, as hollow fibers,
and in microporous bead configurations. These microporous membrane filters
are of special significance to the production of microporous tapes because
of their strength and flexibility.
Oneparticularly unique advantage of the sub~ect membrane
system results from the fact that the assisting polymer fills the incipient
voids in the unleached membrane. The surface of the unleached membrane is,
therefore, still amenable to fabrication of multilayered membrane structures,
with one membrane layer being used as the substrate for the next layer. This
permits, inter alia, production of very thick membrane sections.
Other advantages possessed by the membranes of this inven-
tion are the absence of surfactants and of a tendency to build up static
charges which, in conventional microporous membranes, tracks dust and makes
handIing difficult and results sometimes uncertain.
Other and more specific characteristics of the product
will be apparent from the discussion of the process which follows and from
the subsequent references to the product.
Thus, in accordance wlth present teachings, a process is
provided for forming membrane constructions which comprises the steps of
casting a film from a solution containing a solvent resistant polyester
,: .,
polymer and an assisting polymer soluble in a fluid which is a non-solvent
for the polyester in a mutual solvent for both of the polymers. The mutual
3a solvent is then evaporated from the film to form a membrane having inchoate
pores filled with assisting polymers. At least one additional layer of polymer
suitable for formîng a thin, dense film is cast from a solution onto the
,~,, .:.
624~2
membrane formed b.y the.~receding step.
THE POLYMER ASSISTED pHASE INYER~IO~ ~ROCES.5 :
These-membranes are .~.abri`cated ~ formïng a casting solu- . .
tion of a dîhas.i.c acid dexïved polyester, a second polymer ~hi.ch.is: soluble
in a fluid which.is non-solvent for the polyester, and a liq~id which.is a .--
solvent for both.the polyes.ter and tfie second polymer. This forms the casting
solution for forming the membranes. The castïng solution is then cast upon a ~ ..
substrate to fo~m a f~l~ of desired dimensions. The solvent is permitted to
evaporate from the film and at least a portion of the second polymer is - ' :~
', ''; '''~'',~,'
".,',,''' ~
:.
~. "".
.~ -9.a-
.~
,, ~,,~,
'. '.
-- 106Z~2
leached from the film with the fluid which is a non-solvent for a polyester.
I`his leaves a microporous membrane which has excellent chemical and
physical characteristics, the characteristics of which can be varied by
controlling the nature and amount of the 3 econd polymer in ihe membrane,
as well as by the selection of the polyester, and by control of process
variables .
Polytetramethylene terephthalate i9 considered to be the first truly
suitable aromatic dibasic acid based polyester for formation of microporous
membrane filters because of its greater solubility (compared to polyethylene
10 terephthalate) in suitably volatile solvents. This property makes it possible
to prepare true, molecularly dispersed, solutions of polytetramethylene
terephthalate in a few selected volatile sol~rents. Polytetramethylene tere-
phthalate is more soluble (i.e. exhibits a higher degree of dispersion), than
polyethylene terephthalate, probably because of the greater flexibility of the
polymer chains, owing to the presence of four methylene groups in the former,
rather than the two separating the terephthalate acid moieties in the latter.
Further, and of distinguishing importance, polytetramethylene
terephthalate represents a very close approach to, if not the actual accom-
~ ~./. i
plishment of, the ideal balance between sufficient solubility to permit fabrica- ;
20 tion into membranes, and sufficient solvent resistance, to permibutilization
of the resultant membrane in the filtration of solutions in common organic ~ `
solvents. By comparison, polyethylene terephthalate is insufficiently soluble
to permit fully satisfactory membrane fabrication. On the other hand, the
inclusion of too many or too large substituents on the aromatic nucleus of the
diacid or too long methylene chains or etheric links in the diol, or other
substitution or modification which greatly increases the solubility of the poly-
ester degrades the chemical and physical resistance of the resultant polyester
~..... .
and may require significant modification of the membrane fabrication process. ~ ;`
The first partially successful prototype membranes prepared
30 according to this invention were from polyethylene terephthalate solutions.
Such solutions are turbid, whereas polytetramethylene terephthalate solutions -i
are clear indicating a finer dispersion in the latter solutions. Fine dispersion
' : .`.;~' ": :;'
i' 10 ,,,"""",,~
, ' :'''.. , :"
`` `- 11D62~1Z
; ...
in the solution, and hen'ce'clear solutions, are'highly ~ ~
desirable and are'probably mandatory in any commercially -
feasible process, since only the'latter clear casting solu- -
tions are sufficiently stable for utilization in the dry phase '
:
inversion process on a commercial scale. ~ ~`'' ';
The reader is referred to Kesting, SYNT~IETIC POLYMERIC
MEMBRANES, referenced previously and my Canadian Patent ''
991,776 entitled INTEGRAL UNSKINNED H~GH VOID VOLUME POLY-
CARBONATE MEMBRANES AND A DRY PROCESS FOR FORMING THE SAME,
for a detailed discussion of the phase inversion process for '
.,~ .; . .
fabricating polymeric membranes. The broad principles of
phase inversion membrane formation are applicable to the
present dry process; however, the present process is believed '' '
to be the first satisfactory example of a phase inversion
process for manufacturing microporous membranes in which the
swelling agent consists of a solvated second polymer which
assists in the phase inversion in a manner similar to the
manner in which typical swelling agents of low molecular -~ '
weight assist in the phase inversion in common phase inversion
membrane formation. For this reason, the process is referred '~'
to as a Polymer Assisted Phase Inversion Process, sometimes
abbreviated "PAPI" process. ;t,, '
The procelss consists in dissolving the polyester and one ;'
or more, usually only one, second polymer~s) in a volatile '
solvent, (i.e,, a solvent which has a boiling point below ''
about 150C and preferably below about 100C) or in a combina-
tion of solvents, swelling agents and non-solvents for both ~ '
the polyester and second polymer casting the solution as a '
film, and evaporating the solvents from the solution. The `
process can be either wet or dry; that is, it may involve ' ;~
.. . :
-11- .'''- .
~ ' '.
. , : .
~062412
either partial or completa evaporation, the dry process being
preferred whenever applicable.
The overall process involved herein is described as a
polymer assisted phase inversion process but it has elements
of a leaching process as weIl. In those cases where a non-
polyester polymer component is either desir~d or does not
'
''' '
~"
~ '
.
"
i'
i';: ' ': ` .
';'' ':
, . ' ' ~
:ij" "~ ,;
,~ . `" .
~'~',`"' ,
i;: ' ':'-. - .
-lla- ~
.~; .-:
:
106241Z
:. .
interefere with the end use of the filter, these non-poly~ster
polymer components may simply be lef~ in the microporous ~ :
membrane. In cases where the second polymer, i.e., the :
non-polyester.polymer, would be objectionable,
' .'
'
~ ~ ' ''`' ' : '
..
,, ''!
,' ',` ''' " .
`, ' ', ' .
.i.,,:; ''
`~ ,,' ~ '
,
. ~
.
i.- ,
~' ~
- :.
. ~` 106;~:41Z
the second polymer may be leached out with a solvent for the qecond polymer
which is a non-solvent for the polyester.
The process may be varied to form polymeric microporous mem-
brane filters for specific end uses. For use in microfiltration, ultrafiltra-
tion and electrophoresis1 curtain electrophoresis, gas, liquid and thin ~ i
layer chromatography, an unskinned (i. e., a more or less symmetrical
membrane) is desired. In these ca3es the dried membrane masr be either
leached of its non-polyester component or components or may be used with
the second polymer in place, depending upon the specific requirements of
the end use. It is also possible where two or more second non-polyester -~
polymers have been included, to leach one or more of the polymers from the
membrane, leaving behind one or more polymers together with a polyester. ~ ;`
- Likewise, it is possible to employ in situ cross linking to permanent- ~ `
ly fix one or more otherwise leachable second polymeric components in the : t, .
membrane. `i
Included second polymers may function to improve wettability of ;
the polyester in one of several ways. They may be surfactants or they may ;;
simply possess an affinity for water, up to and including water solubility. ~
Where the included polymers are not actually surfactants an important new i`.
dimension has been added to membrane filters, since the presence of sur~
factants, present in commercially available membrane filters, is often i;
deleterious to certain biological separations, and in handling foodstuffs, waterfor drinking purposes, etc. Therefore, one important feature of the invention ";`:
is that wettability can be obtained by the inclusion o either leachable or,
when necessary non-leachable hygroscopic, but not surface active, polymers.
Two types of asymmetric, i.e., skinned, reverse osmosis mem-
branes and two separate processes for their preparation have been described, ;;
These are the Loeb-Sourirajan w0t process and the Kesting dry process.
; Both of these processes use cellulose acetate polymers in their manufacture.
Cellulose acetate is, however,- rather friable microporous membrane 1 -
configurations. Stronger and less friable reverse osmosis membranes are
ther ef or e, de s ir able .
.' ;' ":' .
- 12- ;
1~62412 ` ~ :
This difficult goal is accomplished, in the present invention, by
casting a membrane ei~her a thin dense or a thicker asymmetric membrane,
of cellulose acetate or other polymer (such as the polyester itself), suitable
for reverse osmosis upon the polyester membranes prior to leaching, so
that the polyester membrane after leaching serves as a porous backing
support for the cellulose acetate or other reverse osmosis polymeric
membrane. Adhesion of the rever~e osmosis membrane to the microporous
support can be achieved by including, as a constituent of the casting solution ~`
. . . __ .:
for the reverse osmosis membrane, a solvent for the polyester. When ;;
properly balanced ~containing the correct amount of polyester solvent), this
solvent system will sufficiently soften the surface of the polyester to permit
good adhesion without destroying the porous characteristics of the polyester ~;
membrane support. Following the application of the second layer, the ~ '
leachable assisting polymer of the substrate may be removed by washing
,;
with water, alcohol or other polyester non-solvent.
It is noteworthy also that dense polyester membranes themselves
have the ability to function as reverse osmosis membranes, as do asymmetric
polyester membranes having a dense skin on a porous substructure.
Solvents
Solvent selection is extremely critical. Indeed, to date only two
solvents, hexafluoroisopropanol and trifluoroacetic acid, are known to be
acceptable for the dry process. The wet process is less desirable but the ~ `
'~ same solvents can be used in such a process. ~-
There are no known equivalent solvents for use with polytetra-
~ methylene terephthalate polyester of suitable molecular weight for membrane
! i
fabrication, and one has little basis for making reliable predictions as to
potentially equivalent solvents. Probably only highly fluorinated substituted
low molecular weight (1-4 carbon atom) alcohols, acids, ethers, esters and
acids can now be considered as solvent candidates in the preparation of the ~
subject membranes, and it is known that not all such compounds are satis- i -
. . .:
factory. Trifluoroethanol and the hexafluoroacetone-methanol hemiacetal, ;
'~`;-! for example, are not satisfactory solvents for fabricating polytetramethylene
~ 13
... ... ... ..
~06~:4~LZ
terephthalate membranes~ Indeed, even trifluoroacetic acid
is so vastly inferior to hexafluoroisopropanol as to make the
latter a component of a distinct discovery within the broad
inventive concept here disclosed.
It may be suggested that where less solvating power is
required, as with substituted aromatic dibasic acid or longer
chain diol derived polyesters, other highly fluorinated ~
compounds of the class discussed would be suitable, but ulti :
mate solvent selection must, at present, be on an empirical
~: .
basis~ i
Any desired solvating power, within the range of the ~ .
solvents, may be obtained by mixing selected solvents. Hexa- ~~
fluoroisopropanol and trifluoroacetic acid, for example, may .
be mixed or diluted with other solvents.of the class under ..
consideration.
It must be understood that solvent selection is not alone ..
a functlon of the nature of the polyester. The nature of the `~
. : .: :
assisting polymer and, more correctly, the nature of the ~ :
combination of polyester and assisting polymer are also con : :
trolling, or at least very influential in solvent selection. `
Assistin~ Polymers
Suitable second or "assisting" polymers are very rare. :
They lnclude certain hyyroscopic polymers, polyvinyl alcohol
in particular. Polyvinyl pyrrolidone also functions, but is
much less suitable than polyvinyl alcohol. These polymers
can be softened and/or leached in water. Other polymers which
are mutually soluble together with the polyester in a solvent,
and which are soluble in a polyester non-solvent are candidates ~.:
as assisting polymers. ~ ;
In a more general sense the assisting polymers must be
soluble in a solvent which is a non-solvent for the polyesters ~ .
~ -14- :.
- ' ':
,.
16)62~1Z
and must also be soluble in-amutual solvent for themseIves
and the polyesters.
Mutual solubillty of the type discussed here~ in approx-
imately equal weights of the polyester and the assisting
polymer at the high total solids content, i.e., between
approximately 5 and 20~, necessary for the practical produc-
tion of membranes is extremely rare in most solvent systems -
and is made possible by the unusual solvent properties of the
polar fluorinated compounds which are employed as mutual ;
solvents. Because of strong
;,:
: ;
~ . .
, -14a-
. ~
~ \
" ~06Z41Z
solvation, solvated chains of dissimilar polymers are apparently much less
's;
incompatible than they would be in the absence of solvation.
The void volume of the membrane produced by the polymer assisted
phase inversion process can be controlled by varying the ratio of the 9econd
(leachable) polymer to the polyester. The higher the ratio, the higher the
void volume. For the polytetramethylene terephthalate polymers of the ~
molecular weights available today, e.g., from about 25,000 to 50,000, void
volumes above about 60% are not achievable with the maintenance of the j -
excellent physical properties found to date. It is to be predicted9 however, -
I0 as higher molecular weight polyesters become available higher void volumes
., ;. . .
will be achievable using the polymer assisted phase inversion process.
The pore size and void volume of the membranes can be controlled ; -~
by controlling the nature of the second, "assi~ting", polymer. The most
suitable assisting polymer found to date, is polyvinyl alcohol.
....::, .
Pore size is a function of the molecular weight and acetyl content -;~
of the assisting polymer. High ~37-42% by weight) acetyl content (e.g. -
Gelvatol*40-lO) polyvinyl alcohols yields coarse structures. Low (0-~%
by weight) acetyl content resins (e. g. Gelvatol * 1-30 and 1 -60) yield fine
; structures. Intermediate size structures (void and pore size ~ 0.05,1lm to `i
1 0.2,1lm) are obtained with intermediate (19.5-22.7% by weight) acetyl content i~ -
(e.g. Gelvatol* 20-30 and 20-60). Fmeness also increases with increasing ;
molecular weight although less dramatically than with decreasing acetyl `-
content. In general the best physical properties are obtained by employing j `~
resins of intermediate (20 ~ 5% by weight) acetyl content and intermediate ~ `'~ (10,000 to 14,000) molecular weight (e.g. Gelvatol* 20 30) resins.
.l Polyvinyl pyrrolidone also functions as an assisting polymer, but
"~ gives such vastly infer~or process and product result, as compared with
polyvinyl alcohol, as to make the use of the latter material a distinct and
very superior discovery within the broad concept of the invention. - `
;-
!,, . . ,~ . . '
; * Trademark. ~
~',.: . .' .. ;; ' .
~,: .; . .
- 15 -
''``;".
;, i , ... .
62~
,:
Examples
Example I ` :
A solution consisting of 5% by weight, Eastman Tenite*
Polyterephthalate 6 PRO polytetramethylene terephthalate, 5%
,
Monsanto GeIvatol* 20-30 DP polyvinyl alcohol, and 90% hexa-
fluoroisopropanol was cast on a ~lass plate to a thickness of
0~020 inches and dried to completion. The result was a 0.003
inch thick microporous membrane which was rapidly wet by
water which rapialy leached the polyvinyl alcohol from the
membrane.
The membrane after leaching was homogeneous, opaque,
strong and flexible. It exhibited essentially the same
solvent resistance as the bulk form of the material and is,
therefore, applicable to filtrations in aqueous and in non-
aqueous media. The pore sized distribution was very narrow,
with a median value of about 0.1 ~m (a bubble point of about
90 psi). The final membrane was a highly satisfactory
unskinned microporous membrane filter having physical and
chemical characteristics vastly superior to cellulosé acetate
and other membranes which have been previously reported. These
homogeneous unskinned membranes are useful in microfiltration,
ultrafiltration ànd electrophoresis, curtain electrophoresis,
and in gas, liquid and thin layer chromato~raphy. For example,
in electrophoresis a membrane produced the standard S-line
(albumin, ~1' a2~ ~ and!~) spectrum for serum proteins.
The application artifact was small and outside of the spectrum
.,
when used with the Beckman* Microzone cell (see, Microzone
Electrophoresis System Brochures of the Beckman Instrument
Company, Fullerton, California).
The bubble point of a membrane mentioned above i~ a means
'
*Trademark ~ -16-
~ .
~(~6Z49 Z
., .
`: -
of determining pore~;size. See ASTM D-2499 which discusses the
essential reIationship between pressure, surface tension of
a wetting fluid, and pore diameter. This relationship ~ ~`
predicts that a wetted membrane with applied differential air
pressure will permit no flow of pressures below a critical
level referred to as the l'bubble point". At the bubble point,
wetting fluid is forced out of the smallest pores and flow thus
begins. With pore size measured in micrometers ~10 6 meters), `~ -
pressure P, measured in pounds/square inch, and with kerosene
` 10 as the wetting fluid, the relation between the pore size and
pressure is ~ = 12.5.
;~
,,`'
. ,~ . .
~,;.'` ': ` -
; ''``
..:
., .:
~'.
','`' ' `'` ,
-16a-
',: .'
62412
Example II
A solution consisting of 5% Eastrnan Tenite * Polyterephthalate 6
PRO polytetramethylene terephthalate, S~o polyvinylpyrrolidone (molecular
weight approximately 10,000) and 90%hexafluoroisopropanol was cast to a
thickness of 0. 020 inch on a glass plate and evaporated to dryness. The
resultant membrane consisted essentially of a porous polyester substructure
whose pores were filled with polyvinylpyrrolidone. A 1% solution of
cellulose acetate in a solvent consisting of 49. 5% hexafluoroisopropanol and
49. 5% methylene chloride was cast on top of the unleached polyester mem-
10 brane, forming a thin adherent cellulose acetate film. The composite
membrane consisting of a cellulose acetate skin and a porous polyester -
support was strong and flexible and was found suitable for reverse osmosis
applications. The polyvinylpyrrolidone was leached from the polyester
membrane as in Example I. ~`
Example III
A solution consisting of 5% Eastman Tenite * Polyterephthalate 6
PRO polytetramethylene terephthalate, 5% polyvinyl alcohol Gelvatol * 20-60
(molecular weight approximately 86, 000), acetyl content approximately 20%
and 90% hexafluoroisopropanol was cast to a thickness of 0. 020 inch on a
Z0 glass plate, and forcibly evaporated with a stream of hot dry air from a hair
dryer. The resultant membrane was asymmetric as evidenced by the presence
of a glossy skin which would not wet when a drop of water was placed there~
upon and a matte finish, porous substructure which readily wet. This
membrane was suitable for reverse osmosis and yielded a product flux of
2 gfd (gallons/ft2 day) and Z8% salt rejection from a 5,000 ppm NaCl feed
solution at 800 psi and 25C.
;-. ;
Example IV
- A solution consisting of 5% Eastman Tenite '~ Polyterephthalate 6 PRO
polytetramethylene ter ephthalate, and 95 % hexafluor ois opropanol was cast and
30 allowed to evaporate completely yielding a gray dense membrane . 001 inch -
thick. This membrane was utilized in liquid permeation as follows: A charge
~.
* Trademark ~
' ' .'
- 1 7 - ~ ~
,
106;~:41;Z !: ~ :
. '.'` ~'.:
containing 95% methane and 5%n-pentane at a temperature of -150~C and a '' (',
differential pressure of psi was found to permeate the membrane at a rate
of 144 grams/ft2hr yielding a permeate,containing 98. 5% methane and 1. 5~0 '
n-pentane, Such membranes are especially useful for'this application when `'~ ' ,
formed into the hollow fiber configuration. ' ' '
Example V , ~
A solution containing 0. 5 to 1. 0 percent Eastman Tenite * Poly- , ~-, ,
terephthalate 6 PRO polytetramethylene terephthalate and 0. S to 1. 0 percent ' ,~
polyvinyl alcohol Gelvatol * 20-30 and 98 to 99% hexafluoroisopropanol was , '~' ',
sprayed from an aspirator no~izle to yield fine droplets which quickly dried ~,^ - '
to a fine powder which when leached with water yielded microporous bead~ ' "'',`, ',
suitable for u~e in column chromatograph~ for separation of, e. g., protein '~
solutions. If denser powders or powders are desired, higher concentrations , , - '
,. . .
(similar to that described in Example I) may be employed. ~`' '
Example VI " ',
The solution described in Example I was cast 0. 020 inch thick onto , ~,'
....... .
a Mylar * backing yielding, after leaching, a microporous membrane ' ','
.. .
intimately bonded to the Mylar *. This supported membrane yielded the ","
typical 5 line spectrum for serum proteins when employed in the Beckman* ' '~;, ' ' ',
20 Microzone Electrophoresis System and was al~o suitable for thin layer
chromatography.
: . .:
Example VII
The solution described in Example I was cast 0. 020 inch thick on a
glass plate and, allowed to dry completely at which time another membrane
was cast on top of it and allowed to dry completely. By repeating this
procedure three times a 0. 009 inch thick membrane was produced. This ,
process can be repeated a~ many times as necessary to produce a membrane
with any desired thickness. Moreover, a membrane of e. g. coarse porosity ` ,'' ' ''
can be coated onto one of fine porosity to produce a laminated prefilter- ' ' "
filter composite membrane in one piece. Laminates of dense membranes ; ~'
with porous membranes have been described in Example II. '''
* Trademark ,, ;,~
- 18_
. .
: ~:
96Z4~Z
General Discussion
Microporous membranes ~uitable for ultrafiltration, microfiltration `-
and electrophoresis, reverse osmosi~, liquid permeation, curtain electro-
phoresi~, gas, liquid and thin layer chromatography, are fabricated from
polyesters produced by the condensation or aromatic dibasic acid and alkyl `
diols by a polymer assisted phase inversion process.
The limitations of the polyesters are that they must have at least as ~ q '.r ` .
: .' ` `' .
great a flexibility and solubility as polyethylene terephthalate and preferably
the degree of flexibility and solubility possessed by polytetramethylene
i,
terephthalate. The polyesters must be formed from the condensation of an ~ `
aromatic dibasic acid and a diol, polyesters formed by the condensation of . ~ j
an aromatic diol and an alkyl dibasic acid being unsatisfactory, Molecular
weights in the range of 25,000 to 35,000 are quite suitable but the molecular ~ ~
weight range is not so critical as in phase inversion processes employing Iow c
` molecular weight swelling agents, The polyester must be soluble to the e~tent
of at least about 0. 5% by weight and preferably greater than 1% or 2% by ; ;; ~ `~
weight in a fluorinated compound such as hexafluoroisopropanol or trifluoro-
acetic acid.
The second (phase inversion assisting) polymer must be mutually
soluble together with the polyester in one or more of the fluorinated solvents
~,! ' and must also be soluble in the polyester non-solvent. The assisting polymer ~- s
- must be mutually soluble in one of the fluorinated solvents to at least about
~; 0. 5% by weight and preferably greater than 1% or 2% by weight in an amount
nearly equal to or greater than the weight of the polyester which i9 dis901ved
in the casting solution, The solubility of the assisting polymer in the polyester
non-solvent may be les~ or more than its solubility in the polyester solvent,
depending upon the end use characteristics. For some end uses, only gross
swelling of the second polymer is sufficient "solubility" where the second ~ ~ `
polymer is to remain in the membrane. In other instances, where the second
- 30 polymer is to be removed essentially completely from the membrane, the
second polymer must be highly soluble in the polyester non-solvent. There is
.~ ,
~ no known reason other than the improbability of compatibility why 2, 3 or even `
,.'': ' ' '
., .~
- 1 9
.~,
":'`` ~06Z4~
more "second"polymers may be included, along with the polyester, in the ~^
casting solution, although such systems have not been explored. One or
more of these may be leached out by appropriate ~election of non-solvents
for the polyester but which are ~olvents of selected strength for the second
polymer which is to be leached. The nature of the second polymer, e. g.
polyvinyl alcohol, can be varied, e. g. as to molecular weight and acetyl `
content, to innuence the pore and void size and physical properties~
The water soluble polymers, polyvinyl alcohol and polyvinyl
pyrrolidone especially the former, are of greatest interest simply because
10 of the great utility of aqueous leaching media industrially, clinically, and in
the laboratory, but the invention is not limited to the use of second polymers
which are soluble in water or other hydroxyl containing solvents, and suitable
.. , . .,, ... , ~
second polymers may be selected on the basis of solubility as set forth 'j~`
hereinbefore . .! . '
Only hexafluoroisopropanol and trifluoroacetic acid are known to be ~!
entirely satisfactory casting solution solvents, i.e., have adequate solvent :~,
strength for polyesters and for the second polymers, proper boiling point and
volatility, and result in strong coherent microporous membranes. However,
the various hydrates cf hexanuoroacetone may also be acceptable under
`, 20 certain extreme conditions~ -
The casting solution for forming the microporous membranes of ~
this invention should contain at least about 2 and preferably from 5 to ZO% by !,
, ~, .. .
weight, of the resins, i.e., the total polyester and second polymer resin
content should be, preferably, greater than 5~/o by weight of the casling
solution. Typically, no surfactant is required in the casting solution, thus
,.- , . ~
obviating many of the difficult problems which have plagued manufacturers ;
; and u~ers of microporous membranes in the past.
Extreme care and highly controlled evaporation and casting condi-
tions are required to form even moderately satisfactory microporous
30 membrane~ using polyethylene terephthalate; however, casting solutions of ~'
polytetramethylene terephthalate are highly stable and the constitution of
these casting solutions can vary over a wide range of concentration and the ,'
.~, ' : .
; - 2 0
,'- ~ ....
62~
casting can be performed under a broad range of temperatures and atmos-
pheric conditions. It i9 usually desirable for health and economic reasons
to cast the microporous membranes in an enclosed hood system to prevent
exposnre to the solvents and to permit recovery of the solvents for reuse.
Other than these fairly simple precautions, no critical casting or process
conditions have been observed.
Leaching of the second polymer can be done simply by immersing ^`
the membrane in the leaching solvent, e.g., water or an aqueous solution.
Leaching will, of course, be accelerated when the leaching solution is ~-
10 agitated to provide constant washing of the membrane. Time, temperature
and process variables do not appear to be critical in either the casting or `-
the leaching process.
Casting as used here includes the formation of any large surface
area configuration, such a~ sheets, ribbons, particles, etc.
Summary
A fundamental polymer assisted phase inversion process has been `
discovered and disclosed which includes the following steps: a solvent
resistant poiymer, for example the polyesters discussed herein, and an
assisting polymer are dissolved in a mutual solvent to a weight percent
20 concentration of from about 2% to about 20% to form a membrane casting
solution. The casting solution is dried under conditions such that a large `~
surface area is exposed, e.g. in the form of a thin film or small particles.
The mutual solvent can be extracted in other ways also, as in the well known
wet process. This results in a membrane (using the term broadly to include
small particles such as chromatographic micro-spheres) in which the inchoate
voids in the solvent resistant polymer phase are filled with the assisting
polymer. The membrane is then exposed to a solvent (in the broad sense
discu~sed previously) for the assisting polymer which is a non-solvent for
~ . .
the solvent resistant polymer. This exposure may only solvate or swell the
30 assisting polymer, but in most practical applications the assisting polymer
will be leached partially or completely from the membrane to leave finished
porous membrane. This leaching may be performed in preparation for use
or as an incident to the desired end use. -
" '
_ 2 1 -
~062~12 .
Another fundamental discovery here disclosed is a polyester mem-
brane composed of a polyester having the following general formula
O R
~ c - o- (cH2) a _ ~ C ~n
-
wherein a is an integer from 2 to 7, preferably 3 to S, and n is an integer
from about 45 to greater than 225 ~corresponding to a molecular weight of
from about 10,000 to about 50,000) and preferably between about n=90 and
n=225 (MW~20,000 to MW~50,000). ;^~`
Lying within, and exemplifying, the fundamental process discovery
is a process for forming polyester membranes which could not have been -
expected or predicted even if the more fundamental process had been known
previously. According to this process, a polyester having the formula
~_~3 C-0-(CH2) a _0-C ~n
wherein a is from 3 to 5 and n is preferably from 90 to 22S or greater and
an assisting polymer selected from the group consisting of polyvinyl alcohol
and polyvinyl pyrrolidone or mixtures of these are dissolved to a total weight
20 percent of from 2% to 20~o (preferably about 5%) in a solvent which consists
essentially of (i.e. the solvating power for the polymers of interest is not ~; i ;
substantially reduced by diluents~ hexafluoroisopropanol, trifluoroacetic acid ; !;
~, ,, .. ;:,. . .
or mixtures of these solvents. The polyester and assisting polymers are
present in approximately equal volume percent (of the dry membrane) con- `
centrations in the casting solution i. e. the polyester and as B isting polymer
are pre~ent in about S0 v/o ~ 10 v/o concentrations in the membrane as cast,
i before leaching. Generally, there will be a greater amount of assisting `
polymer in the casting solution on a volume percent basis. The process is ~ ;
carried out as in the fundamental process. It is recognized that the foregoing
30 may be the only possible examples of the fundamental process, but this has
not been established.
Within the foregoing examplification of the fundamental process is
a particular process which is so unexpectedly and strikingly superior that it
- 22 - ;
.j,, .
" ~ , ", , ~ ,, " ' ,;,:
r~ ''
06~:4~Z
constitutes a distinct major discovery which will have far reaching scientific
and industrial consequences. It has been discovered that one, and only one,
combination of constituents, when used in the pol~mer assisting phase
inversion process under consideration, gives greatly superior results with
greatly reduced proceqs problems and significantly increased reliability.
These constituents are:
polyester ~ C - O- (CH2 ) - O- C ~ '' ,~wherein a is 3 to S Ipreferably 4) and _ is 90 to 22S.
assisting polymer: polyvinyl alcohol having a molecular weight of from
about 5,000 to about 2S,000 ~10,000 to 14,000
preferred) and an acetyl content of from 0 to about
50% (15% to 25% preferred).
: ,
mutual solvent: hexafluoroisopropanol
leaching solvent: water (or any solvent for polyvinyl alcohol which i9
a non-solvent for the polyester).
Of these constituent materials, only the leaehing solvent is not critical to the
achieving of the vastly superior results obtainable by this process.
The product of the foregoing superior process, i.e. the resulting
~;~ polyester membrane, constitutes a distinctive and unique discovery not
Z0 expected and incapable of prediction, even in view of the more fundamental ;
dis cover ie s .
It is apparent, then, that this invention is both general and specific.
... . .
It is general in the sense of the more fundamental aspects of the process and
resulting polyester membrane. It is specifie with respect to the unigue,
unexpected and superior results (as compared with the more general results)
which accrue from using the particular combination of the indicated polyester,
polyvinyl alcohol and hexafluoroisopropanol in the process.
The following claims are drafted with these general, certain inter- ~;~
mediate features and the specific features of the invention in mind. -
.. ,
,,, .. , '.
, :. .
- - 23 _
: -' .~"
