CA2049327A1 - Composite, porous diaphragm - Google Patents

Abstract

A composite, porous, liquid-permeable article is provided which is a multi-layer structure of discrete, bonded layers of porous, expanded polytetrafluoroethylene (EPTFE). The composite has its interior and exterior surfaces coated with a perfluoro ion exchange polymer to render the composite hydrophilic so as to resist gas locking in aqueous media. Initially, the diaphragm may also contain a water-soluble surfactant to assist in initial water penetration into the pores of the composite. An improved electrolytic cell is provided having the composite diaphragm as the porous separator in the electrolysis of alkali halide solutions. The diaphragm is also useful as an improved filter medium.

Description

~`VO 90/13593 ~ P ~ /~S90/02349 A COMPOSITE, PORO~S DIAPHRAGM
CROSS-REFERENCE TO RELATED APPLICATION
-This application is a continuation-in-part of earlier filed, penaing a?plication Serial Number 07/344,707, fileo April 28, 1989.
BACKGROUND OF THE INVEN~ION
;
1. Field of the Invention The invention relates to a porous, liquid-permeable composite article useful as a diaphragm for electrolysis or as a filtering medium.

2. Description of Related Art In the electrolysis or electrosynthesis of chemical compounds, a porous diaphragm is often used to separate the anode and cathode compartments and thereaction products while penmitting the flow of some liquid components from onecompartment to another. For e~ample, in the production of chlorine and sodiumhydro~ide from brine, the brine feed flows from the anode compartment throughthe porous diaphragm to the cathode compartment and then is discharged from thecell as illustrated in Fig. 1. A~proximately half of the sodium chloride isconverted to sodium hydroxide and chlorine in the process.
The effect of dia?hragm structure on the performance of a chlor-alkali cell is ~uite complex. The diaphragm can be described in terms of pore sizedistribution, porosity, tortuosity, thickness and resultant permeability of the~20 structure. For a given set of cell operatins conditions, these parameters, and especially their uniformity across the active area of the diaphragm, determinethe electrical energy usage of the cell. Lack of uniformity of flo~ ratesacross the surface of the diaphragm cause areas of low brine velocity, whichallows hydroxyl ion back migrztion, leadin~ to ~oor current efficiency. Thiseffect can be ameliorated by using a thicker, less porous or more tortuousstructure, leading however, to hiyher cperating voltag-e and ~reater electrical ener~y usage. The art in desigming a diaphrasm for chlor-alkali production isto properly balance the diaphragm properties to minimize overall electricalener~y usage by reduciny operatins voltage while maintaining high currentefficiency. This is most effectively done with a diaphragm whose properties arehighly uniform across its active area.
In the electrolysis of brine, the p~wer consu~ption in terms of kilowatt hour (kWh) per metric ton of so~ium hydroxide can be expressed ~y the follo~ingequation:
kwh/metric ton NaOH = 67010 x cell voltacJe % caustic current efficiency Obviously it is economically desirable to achieve as high a caustic current - efficiency as possible and require as low a volta~e as ?ossible. A desirable diaphragm will have a lo~ "k" factor, k being the slope of the voltage versus , ,. . ~ , , , : : : ---: : . .:

-~ Q~ r~Q~
WO gO/13593 ~ -2- PCI/US9O/OZ349 current density relationship. Many commercial diaphragm chlor-alkali cells operatc at a maximum current density of either 2.3 kilo-amperes per square meteror 2.8 kilo-amperes per squsre meter.
A commonly used porous diaphragm is prepared from asbestos fiber by essentially a p~per making process. Asbestos flock is slurried and deposited in place on a screen in the electrolyzer to form 8 relatively thick, stiff ~iaphr~gm which is held together by the hydroxide gels formed in the asbestos while in operstion. The asbestos diaphra~m has often been termed a "living diaphragm" in that it is constantly being changed by dissolution, erosion, redeposition and precipitation of silica snd alkali esrth hydroxides. Suspended particles and dissolved alkali earth met~ls rec~eposit preferenti~lly in th~ higher flow re~ions allowing a levelinc~ ef~ect on perrneability.
However, this '~iving" or reactive feature of the asbestos diaphragm can contribute to a relatively short life. Usually, within 6-12 months, sufficient lS chemicals are leache(i out of the asbestos and the uniformity and porosity are so degraded that current efficiency drops to unacceptable levels. For the same reason, electrical upsets or fluctuations in the system can result in a very rapid destruction Or the asDestos diaphragm. Finally, the asbestos diaphrazm has a relatively high k factor of about 0.55 volt square meter per kiloampere (Vm2/kA)such that, for example, ~3t 2.8 kA/m2, the diaphragm typically requires 3.84 volts or more for operation, resulting in sub3tantial power costs. ~`or these reasons, the ",, industry has sought more inert, more stable diaphragms, which can operate consistently at high current el`ficiency and at lower voltage and which will not be destroyed by power upsets, fluctuations or outages.
A number of modified asbesto3 diaphragms have been develcped. U.S.
Patent 3,853,720 discloses a preparation of a diaphragm for chlor-alkali serviceinvolving asbestos fiber, a second fibrous material including polytetrafluoro-ethylene (PTFE) fiber, ~nd an organic exchange resin. U.S. Patent 4,170,j37, U.S.
Patent 4,170,~38 and U.S. Patent 4,170,539 describe a diaphragm of an asbestos or polymer matrix containing inorganic zirconium or magnesium compounds and, in some cases, "Nafion'~D 601 polymer solution", a colloidal dispersion of hydrolyzed perfluorosulfonic acid polymer, which is used to impregnate the structure. All of these structures rcly, ~o so me e~ent, on the esbestos or the edded co mooun~s ' '; .

:. : ' - , : .
.. . .. .

... .
, w090/l359~ ~3~ ~ 3 ~ ~327 which generate hydroxide gels to regulate porosity and uniformlty. Accordingly, though somewhat more stable than unmodified asbestos, they su~fer from the same deficiencies described for the asbestos diaphragm described above. These patentsalso mention expanded PTF~ (EPTFE) as a possible polymer n~atrix but the pore size range specified, 0.~-50 n-icrons, preferably 5-20 microns, is relatively large so that the permeability level is controlled by the hydroxide gels formed within the diaphragm.
A number of U.S. Patents, Nos. 3,930,979, 4,113,912, 4,224,130, 4,606,805, 4,385,150, 4,666,573, 4,680,101, 4,720,334 and 4,3~1,614, describe porous PTFE
diaphragms made wettable by various means. U.S. Patent Nos. 3,930,979, 4,250,002, 4,113,912, 4,38.5,150 and 4,341,614, describe porous PTFE diaphragms prepared by combining PTFE powder or fiber with 8 sacrificial filler. The mixture is formed into a sheet and the filler is removed by dissolving it or decomposing- it with heat thus lesving the PTFE sheet porous. Tne homogeneity of the nnixture and the particle size distributions of the fil!er and PTFE severely limit the uni-formity possible in the finished diaphragm. Because a large percentage of the structure is removed to provide the necessary porosity, the finished diaphragm is inherently weak. To offset these uniformity and strength problems, the finished diaphragm must be very thick resulting in high operating voltage. A number of patents, including V.S. Patent 4,~06,805, U.S. Patent 4,666,573, U.S. Patent 4,113,912, U.S. Patent 4,680,101 and U.S. Patent 4,720,334, describe porous PTFEdiaphragms prepared by a PTFE fiber slurry deposition process. The size, shapc and size distribution of the PTFE fiber availsble leads to a large pore, inherently weak, non-uniform structure which must be made very thick to provide utility.
This trade-off results in high operating voltages. In addition, in U.S. Patent Nos.

3,g30,979, 4,113,912, 4,224,130, 4,250,002, 4,341,614 and 4,606,805, no perfluoro ion exchange polymer is involved so that an adequate level of hydrophilicity, which remains chemicslly stable in a hostile environment such as a chlor-alkali cell, is not achieved. W hen an adequate level of hydrophilicity is not rnaintained, gas bubbles generated at the cathode will accumulate in the diaphragm pores blockingboth bulk and ion flow. This reduces the effective diaphragm. area leading to anincrease in operating voltage and eventually causing system shutdown. This is called "gas locking".
U.S. Patent 3,944,477 describes a diaphragm of porous polytetrailuoro-ethylene sheet material with a microstructure characterized by nodes and fibrilsand having a multilaycr structure wherein a number of such filn-,s are bonded together. Initial wettability is achieved by treatment with acetone and water, `'`-"
- , ! . ~ ' ~:: , ' ' ' ;,' ' ' ' ; ' : ' ' ' . ' -4~

2 ; . .. 7 with no mention made of organic surfactants. More importantly, there is no disclosure of impregnatiGn with a perfluoro ion exchange polymer. While the diaphragms of the present invention have given satisfactory performance after asmuch as 421 days, the eference reports no extended runs.
U.S. Patent Nos. 4,089,758 and 4,713,163 describe porous diaphragms involving EPTFE structures ma~e hydrophilic with inorganic filler particles or or~anic surfactants. These diaphragms are susceptible to gas locking, which willblock ion and bulk fluid flow, resultiny in increasing operating voltage, decreasiny current efficiency and ultimate system shutdown.
U.S. Patent 3,940,916 describes a porous diaphrasm made from a fabric spun from an ion exchange polymer. The pore size of this structure is too large for e.'icient operation in a chlor-alkali cell. Because of this, very thick structures are required for high current efficiency operation resulting in high voltage.
U.S. Patent 4,385,]50 discloses, but does not claim, a porous asbestos or ~rFE sub~trate im~regnated with an organic solution of a fluorinate~ eopolym~r having a carboxyl functional group. The disclosure does not spccify the PTFE as havins a microstructure characterized by a series of nodes interconn æ ted ~y fibrils, nor does it specify a m~ltilayer construction. Accordingly, this PTFE
structure does not provide the uniformity of structure, porosity andpermeability necessary for sustained hish efficiency operation in a chlor-alkali c~ll .
U.S. Patent Nos. 3,692,569, 4,453,991, 4,865,925 and 4,348,310 and Japanese Patent Applications JPA-6]-246,394 and JPA-63-99,246 describe and claim porous diaphra~ms involvin~ EPTFE coateo with perfluoro ion exchange resin for use in electrochemical cells. However, the EPTFE cited is not a layered structure and will not, at a corresponding thickness, provide the small pore size and ; uniformity of st N cture necessary for efficient operation of a chlor-alkali cell, especially in thicknesses exceeding 20 mils. Moreover, U.S. Patent Nos.
4,453,991 and 4,348,310 specify impregnation of the EPTFE with perfluoro ionomerqolutions ~ith e~uivalent ~eishts exceeding 1000. The relatively large micelles of these high equivalent weight systems cannot penetrate a relatively thick, small pored ~lFE st Ncture to uniformly and thoroughly impart sufficier.thydrophilicity to the entire structure to allo~l efficient chlor-alkali diaphragm operation.
`~

~,:
j;
.
., .

.. . ~ .

-4a-W O 90/13593 P ~ /US90/02349 J rr~ ~ r!

U.S. Patent No. 4,865,925 discloses use of an EPTFE/perfluoro ion exchange resin structure for a fuel cell, ~hich is an electrochemical cell, but thestructure woulu not be suitable for a diaphragm because it is porous to yas. Achlor-alkali cell reyuires that hydrosen from the cathode side must not mix uithchlorine from the anode side because st hydroc~en/chlorine mixtures are explosive. The reference malces no disclosure of the need for thorouyh impre~nation, the need for at least four layers of EPTFE thermally bondedtocJether, or the preference for certain equivalent ~eights for the ion exchanye resin.
U.S. Patent 4,277,429 describes a metho~ for producing a porous PTFE that is asymmetric in the sense that there is a measurable difference in bubble point ' ,;

i` ' ~'.
. ~l ~''"'' .

i ~: , . . .. . :
~:. '; .' . ~ - . .~ . . , . , , .... , . .. ~, ~, . .'.,'', .,. : ' '': ' ' :: . , ,. .. . - - :
. ~ . - , . .
:,: : : , , . - , . - - . ~ , W 90~3 2 ~ L~ t~ ~; PC~JUS~0/02349 between one surface and the reverse side surface. Slightly dif~..entpermeability to isopropanol is noted in one direction than in the reversedirection. Such a structure, however, is monolithic and not layered and wouldnot provide the high level of uniformity of pore size and pore size distributionnecessary for efficient operation of a chlor-alkali cell. In addition, theinterior and exterior surfaces of this porous Pl~E are hydrophobic and would~'gas lock" in chlor-alkali production or in other electrolytic or filtration uses ~here gas entrainment is a potential problem.
International Patent Application No. PCr/US88/00237, publication number W088/05687 and its counterpart U.S. 4,863,604, describes a microporous,asymmetric, intearal, composite polyfluorocarbon membrane of two or more sheetsof microporous fluorocarhon polymer having different average pore sizes. Thesestructures, however, l~ould not be useful as diaphragms in a chlor-alkali cell.
Such sheets are not EPTFE structures ~ut rather are prepared by incorporating aparticulate, inorganic, solid, pore forming filler, removeable by leaching andheating, lnto the polytetrafluoroethylene polymer, and shaping the resultantmixture by preforming and calendering it into a self-sustaining sheet or film.The multilayer structure is assembled by starting with a sheet of PTFE/pore; forming filler which 'nas small pore forming filler particles. On top of this sheet are layered additional sheets of PTFE/pore formin~ filler which containprocJressively larger pore forming filler particles. The sheets are bonded with heat and pressure, followed by sintering. Finally, the pore forming filler isr~moved by leachiny or heat, thus leaving the multilayer PTFE sheet porous. Asdiscussed abGve, the homoyeneity of the mixture and the particle sizedistributions of the filler and PTFE severely limit the uniformity possible inmicro~Grous structures prepared by leachin~ or other~lise removing incorporated particles. This limitation is further compounded by the limitation on sharplyfractionating particle sizes for the various layers of the asymmetric structure.Also, as mentioned above, because a large percentage of the structure is removedto provide the necessary porosity, the finisnea structure is inherently weal;.
To deal with these uni~ormity and ~trength problems, the finished diaphragm mustbe very thic~, which is unoesirable in electrolytic operations because of thehiyh operatin~ voltage required. In addition, in PCr~US88/00237 (W088/05687)and in U.S. 4,863,604, no perfluoro ion exchange polymer is involved so that anade~uate level of hydrophilicity -hich remains stable in a hostile cnvironment i5 not achieved. As discussed abovc, when an ade~uate level of hydrophilicit~
is not maintained, "yas locking" -occurs reducin~ effective diaphra~m areai ` leadin~y to an increase in operating voltase.

.

,:;
~,. . . .
J ` ~
... . .

"~ ' : ' ' '\, . . ' . ' ~
~ :

W0 90/13593 -6- ~ j, ` 9 3 ~ ~ Pcr/ll590/023~9 U.S. Patent 4,385,093 describes a porous PTF}~ article prepared by layering together PTFE components followed by expanding in one or more directions. The resulting article has very hi~h interlayer bond strengths and appears uninterrrupted at the layer interfaces. The interlayer bond strength of this article WRS shown to 5 be much higher than a composite prepared by layering and sintering two ~lreadyexpanded PTFE sheets. The high interlayer bond strength is desir~ble in certain applications to prevent delamination of the layers due to gas, liquid or osmoticpressure whieh mfly build up inside the structure during use. The method of IJ.S.
Patent 4,385,093 does not take advantage of the averaging effect of layering 10 because the expansion is carried out after the layering step, The resulting product is less uniform in strueture and pore size than the article of this invention which is made by layering already expanded sheets. Further, the pure PTFE structure otU.S. Patent 4,385,093 has an inherent tendency to entrain ~as in certain aqueous, electrolytic or filtration applications and will eventua11y gas lock.
U.S. Patent 4,341,615 and U.S. Patent 4,410,638 both claim a wettable, microporous diaphrflgm for electrolysis hsving a base of fluorinnted resin, the pores of the microporous diaphragm having deposited therein a copolymer of an unsaturated carboxylic acid and a non-ionic unsaturated monomer. The structure is monolithic and not layered and does not provide the uniformity of structure 20 necessary to provide high current cfficiency and low oper~ting voltage for chlor-alkali operation. This dericiency is further compounded in that this monolithic,fluorin~ted resin construction is prep~red by leaching out calcium carbonate particulates from the fluorinated resin composite. Accordingly, ~s pointed out earlier, the homogeneity of the mixture and particle size distribution of the filler 25 and resin severely limit the uniformity of structure possible in the microporous sheet. The copolymer deposited within the pores to impart hydrophilicity is not perfluorinated snd does not provide the necessary durability in the corrosive environment of chlor-alkali service.
The present invention is a porous, multilayer construction comprising mul-tiple layers of porous EPTFE bonded together wherein the internal and external surfaces are at least pflrtially coated with a perfluoro ion exchange polymer. rwo pending U,S, patent applications, U.S.S.N, 206,884 and U.S.S.N. 278,224, in the names of some of the inventors of the present invention, involve the same starting : materials, ~.S.S.N. 206,88~ "(`omposite ~Icmbranc" discloses u thin porous expande:l PTEE whose internal and e.Yternal surfaces are coa~ed with a metal s~lt of a ~; perfluoro ion exchange polymer. That composite is porous like the present . .
. ~i ~i .: .
.

.,.. ':~ , . . ' :' :ç:: ~ . ~ , :.: . ~ . . . - . . .
.`,:: ,, - .. . .
': . .. ~ :. .

W090/i3593 ~ ,3?.'1 ~7 PCI/U590/1123~9 invention but, in contrast to the presen~ invention, it is much thlnner. In thatapplication, the perfluoro ion exch&nge polymer serves as an anchor for active metal ions which may scavenge, catalyze or otherwise reQct with fluids pAssin~, through the porous structure. The EPTI:~ component is a single layer construction, 5 not the multilayer form of the present invention.
U.S.S.N, 278,224 is a multilayer composite comprising a reinforcing fabric bonded to an expanded PTFE film which is laminated to fl continuous film of perfluoro ion exchange polymer. In contrast to the present invention, that con-struction is a non-porous composite where the EPTFE is used as an interlayer 10 between the continuous film of perfluoro ion exchange polymer and a reinforcing fabric. In addition, the EPTFE component is a single layer, not the multilayer construction of the present invention.
The carboxylic polymer~s with which the present invention is concerned have a fluorocarbon backbone chain to which are attached the functional groups or 15 pendant side chains which in turn carry the functional groups. When the polymer is in melt-fabricable form, the pendant side chains can contain, for example, -fCF~^W
L~ J t groups wherein Z is F or CF3, t is 1 to 12, and W is -COOR or -CN, wherein R
20 is lower alkyl. Prefcrably, the functional group in the side chains ot the polymer will be present in terminal r O-- -c~--W
Z t groups whereln t is 1 to 3.
as The term "fluorinated polymer", used herein for carboxylic and for sulfonic polymers, means a polymer in which, ufter loss of any R group by hydrolysis to ion exchange form, the number of F atoms is at least 90% of the total number of F, H and Cl atoms in the polymer, For chloralkali cells, perfluorinated polymers are preferred, though the R in any COOR group need not be fluorinated because it is lost during hydrolysis.
Polymers containing -(OCF2CF)mOCF2CFCN
~F3 CF3 side chains, in which m i~s 0, 1, 2, ;3 or 4, are disclosed in U.S. Patent 3,8S2,326.

,~:

.
:~ ' .: . . . -,~' .. '' ' ' , ~-: ..... : , .
: . , . ,. . . . .. .; .

wo gn/l3~93 ~ 7 PCl/US90/02349 Polymers containing ~CF2)pCOOR side chains, where p is 1 to 18, are dis-closed in U.S. Patent 3,506,635.
Polymers containing -(OCF2CF)mOCF2COOR
Z
side chains, where Z and R have the meaning defined above and m is 0, 1, or 2 (preferably 1) are disclosed in U.S. Patent 4,267,364.
Polymers containing terminal -O(CF2)VW groups, where W is defined above and v is from 2 to 12, are preferred. They are disclosed in U.S. Patent Nos.
3,641,104, 4,178,218, 4,116,888, British Patent No. 2,053,~02, EP No. 41737 and British Patent No. 1,518,387. These groups may be part of ~OCF2CF)m-O-(CF2)V-W

side chains, where Y = F, CF3 or CF2CI. Especially preferred are polymers lS containing such side chains where v is 2, which are described in U.S. Patent Nos.
4,138,'126 and ~,~87,668, and where v is 3, which are described in U.~. Patent No.
4,065,366. Among these polymers, those with m=1 and Y=CF3 are most preferred The above references describe how to make these polymers.
The sulfonyl polymers with which the present invention is concerned are fluorinated polymers with side chains containing the group -CF2CFSO2X, Rf ~-; wherein Rf is F, Cl, CF2CI or a C1 to Clo perfluoroalkyl radical, and X is F or Cl, preferably F. Ordinarily, the side chains will contain -OCF2CF2CF2S02X or ` ~` 25 -OCF2CF2SO2F groups, preferably the latter. For use in chloralkali membranes.
, perfluorinated polymers are preferred.
Polymers containing the side chain -O(CF2C~ FO)k~CF2)j-S02F, Cl3 where k is 0 or 1 and j is 3, 4, or 5, may be used. These are described in British Patent No. 2,053,902.
, Polymers containing the side chain -CF2CF2SO2X are described in U.S.
Patent No. 3,718,627.
:~
, . .

. ~ .................................. .
.'' ."'' J : . ~
. . - . - .. .
,. , - , . . . .

. -. : : ~

W O 90/13593 2 ~ ?~7 ~9~ P ~ /US90/OZ349 Preferred polymers contain the side chain -~OCF2CF)r-OCF2CFS02X, Y Rf where Rf, Y and X are defined above and r is 1, 2, or 3, and are described in U.S. Patent 3,282,875. EsFecially preferred are copolymers containina the side chain -OCF2CFOcF2cF2s02F~

Polymerization can be carried out by the methods described in the above references. Especially useful is solution polymerization using ClF2CCFCl2 solvent and (CF3CF2C00)2 initiator. Polymerization can also be carried out by a~euous granular polymerization as in U.S. Patent No. 2,393,967, or aqueous dispersion polymerization as in U.S. Patent No. 2,559,752 follo~Jed by coagulation as in U.S. Patent No. 2,593,583.
~ make the lowest eguivalent ~leight ion ex.chan~e polymers, copolymer in the melt-fabricable (for example, -S02F or -COOCH3) fcrm may be ~Ytracted as in U.S. Patent 4,360,601 and the extracted polymer isolated for use in making the diaphragm. The extract has lower ~quivalent weight than the startin~
material.
~le copolymers used herein should be of hiyh enough molecular weight to produce films which are self-supportins in both the melt-fabricable precursor form and in the hydrolyzed ion-exchanged form.
., .

:.' :,, :;
j"l . . ~
~, .

f f~ ,' ' ' ~..,.':`
~",., ~ . ~
: - .' .: '-: ' . ' .,..' ~' : ' , ` ' , ' ' ~ , ' . . .

-9a-W O 90/13593 ~ ~ ~ n~ PCT/VS90/02349 .

SUM~hRY OF THE INVENTION
A multilayer, porous, com~osite, shaped article is provided comprising multiple layers of porous, e~:panded polytetrafluoroethylene bonded toyether, the composite, shaped article having at least a portion of its ~terior surfaces and at least a portion of its interior pore surfaces coated with a perfluoro ione~chan~e 2olymer. Preferably, the com~osite article has substantially all ofits exterior surfaces and s~bstantially all of its interior pore surfaces coatedwith a ~erfluoro ion exchan~je pol~er. I~e com?osite article may contaiin a water soluble surfactant within its pores. ~le article may be in the form of asheet or a tube. ~le perflucro ion e~change polymer is a perfluorosulfonic acidpolymer of equivalcnt ~ei~ht less than 1000, a perfluorocarboxylic acid polymerof e~;uivalent wei~ht less than 1000, a mixture of perfluorosul-onic acid p~lymer and perfluorocarbo,~ylic acid polym~r of ec,uivalent weiyht less than 1000, or a copolymer containing perfluorosulfonic acid and perfluorocarboxylic acid groups,with an equivalent weight less than l000. The percentase by weight of perfluoroion exchanye polymer in the composite exceeds 2%. The sheets are relatively thic;~, having a thickness exceedin~.~ .25 millimeters, preferably betueen about0.76 millim2~ers and about '.0 millimeters. The composite article has a permeability to uater containing 0.1% tetra etllyl am~onium perfluorooctane sulfonate between about 0.01 and about 3.0 reciprocal hours at 23C under a 20 ; cm head of t~ter and a specific sravity between about 0.05 and about 1.1, - preferably between about 0.15 and about 0.7. The composite article may have an ; asymmetric fine structure "~herein at least tuo of the layers have different micro~rous ~tructures. ~1e two or more layers may have methanol bubble ~oint value5 which differ by at least 10% or have s~ecific ~ravities which differ byat least 5%.

:

. ~
. . .
''' .`.

J
.1.' .

.:

:
~ .......... . .

:, .... :
x ~ - :
~"

WO 90/13593 ~ 10- pcl /ussn/o~34g In an electrolytic cell containing anode and cathode compartments separflted by a diaphragm, an improved diaphragm is provided comprising a multilayer, porous composite diaphragm of multiple layers of porous, expanded polytetrafluoro-ethylene bonded together, the composite diaphragm having at least a portion of its exterior surfaces and at least a portion of its interior pore surfaces coated with a perfluoro ion exchange polymer. A plurality of composite diaphragms may be used to separate a plur~lity of cell compartments of an electrolytic cell. The composite diaphragm preferably has substantially Qll of its exterior surfaces and substantially all of its interior pore surfaces coated with a pertluoro ion exchange polymer. l`he diaphragm may initially contain water soluble surfactant within its pores to enhance initial wetting. The diaphragm may have an asymmetric fine structure, wherein at least two of the multiple layers have different microporous structures, wherein the two or more layers have specific gravities which differ by at least 5%. The asymmetric diaphragm prefersbly is oriented such that, in the two or more layers, the layer of lower specific grsvity is closer to the anode side of the cell and the layer of higher specific gravity is closer to the cathode side of the cell.
The multilayer EPTFE structure of this invention yields an exceptionai level of uniformity in diaphragms such that they operate at cell voltages and current efficiencies si~nificantly better than those of prior art. The perfluoro ion exchange coating on the interior and exterior surfaces of the diaphragm, the other essentiai feature of this invention, provides a level of hydrophilicity that prevents gas locking and leads to sustained operation at high current efficiency and low voltage, i 25 An improved filter medium is also provided comprising a multilayer, porous, composite, shaped article of multiple layers of poro~s, expanded polytetra-fluoroethylene bonded together, the composite having at least a portion of its exterior surfaces and at least a portion of its interior pore surfaces coated with a perfluoro ion exchan~e polymer. The filter medium may be in the form of a sheet or a tube.

':`' ':

' ' .

. . , , ~ .
`' ` ~., " ' ,,, W O 90/13593 Z ~ 7 P ~ /U590/02319 Another aspect of the present invention is a process for coatiny the exterior surfaces and at least a portion of its interior pore surfaces with aperfluoro ion exchan~e polymer. One feature of this process is tlle incorporation in the lic~id coating composition of an orsanic compounQ whichenables the composition to fully wet a horizontal surface of nonporous PTFE (asdistinguished from ex~anded PTFE) and to re~ain s~read out as the compositiondries instead of formins droplets. Another feature of the present invention isthe use of vacuum tc remove most of the air from the EPTFE be~ore the coatinycomposition is introduced from one side of the EPTFE.

j':
.
:
, ' .
, . , - . . . . .

t, . ~ ' ` , ' ' ' ' '- ~ ' ' ,' ~ ~ ' ,, - '. .
,........ . ,. : ,, , , -; . ~,~ . : . : : " . , .
.

W0 90/13593 ~ P~r/US9O/02349 BRIEF DESCRIPTION OF THE DRAWIN(:;S
Fig. l is a schematic diagrsm of an electrochemicsl cell.
Fig. 2 is a pho~omicrogrnph taken at ~5x magnification of the cross-section of a symmetric composite uccording to the invention.
Fig. 3 is a photomicrogrnph tal<en at 5000x magnification of the symmetric composite and shows the microstructure of nodes and fibrils co~ted with a perfluoro ion exchange resin.
Fig. 4 is a photomicrograph take at 50x magnification of the cross-section of an asymmetric composite according to the invention.
~ig. 5 is a ?hotomicrograph taken at 5000x magnification of the asymmetric composite and shows the microstructure of nodes and fibrils coated with a per-fluoro ion exchange resin.
DETAILED DESCRIPTION OF THE INVENTION
AND PREFERRED EMBODIMENTS WrrH
REFERENCE TO THE DRAWINGS
A composite, porou~s, liguid-perme~ble urticlc is provided which is a multi-layer structure of discrete, bonded layers of porous, expanded polytetrafluoro-ethylene (EPTFE). The composite has its interior and exterior surfaces coated with a perfluoro ion exchange polymer to render the composite hydrophilic so as to resist gas lockin~ in uqueous media. Initially, the ~liaphragm may also contnin a water soluble surfactant lo assis~ in initinJ water penetration into the pores of the composlte. An improved electrolytic cell is provided having the composite diaphragm as the porous separator in electrolysis processes, particularly electrol-ysis of alkali halide solutions. The diaphragm is also useful us an improved filter medium.
More specifically, a mechanically strong, porous, composite, liqui~per-meabl~ diaphragm is provided which is a multilayer structure of discrete bonded EPT~E layers. This relatively thick, preferably greater than 5 mil thick, layered structure provides a small pore size and uniformity of structure not attainable in monolithic EPTFE structures. By coating the interior and exterior surfaces of this stru¢ture with a perfluoro ion exchange resin of equivalent weight less than l000, hydrophilicity of the resulting composite can be greatly increased, thereby drastically reducing the composite's tendency to entrain gas in the pores. Initial wetting is assured by initially incorpornting a water soluble ~surfactant in the porcs ! ~ 35 when desirable.
In the electrolysis of brine, the porous composite of this invention provides a chemically stable, porous diaphragm with n uniform microstructure such that .;

; .

~ .. ~. , ~ . . .

W O 9D/13~93 -12~ PCTlUS9n/o~349 uniformity of flow and high current efficiency can be obtained without thereactive hydroxide gel deposits encountered with conventional diaphragms.Further, the porous composite of this invention can withstand numerouselectrical upsets, shutdowns or fluctuations encountered in normal cell roomoperations without si~nificant loss of performance. The porous multilayercomposite of this invention also provi~es a small pore si~e and uniformity ofstructure and flow not obtainable in the thick, monolithic EPTFE structures orin the PTFE structures prepared by slurry deposition of fibers or by leachingsoluble particulates from a filled PTFE sheet.
The perfluoro ion e~change polymer is a copolymer of tetrafluoroethylene with one of the fur.ctional comonomers disclosed herein. The ratio of tetrafluoroethylene to functional comonomer on a le basis is 1.5 to 5.6:1.For each co nomer, the most preferred ratio of tetrafluoroethylene tofuncticnal co nomer is determined by eYperiment.
Throush the use of licuid com~sitions of perfluoro ion exchan~e resin of e~uivalent ueight less than 1,000 and by virtue of the much smaller micelledimensions of these dispersions as com~ared to dispersions from perfluoro ionexchange resins with equivalent weight exceeding 1,000, the very small ?ore~ ofthe relatively thiclc multilayered structures of this~invention can '~e penetrated and the e~terior surfaces and interior pGre surfaces can be uniformly coatedwith perfluoro ion e~:change resin. A water soluble surfactant can also be introduced to facilitate initial wettin~ by aqueous media. For bestperformance, any surfactant present must ke washed a~y before electrolysis is started. The perfluGro ion exchanse resin, hcniever, ~-ill not wash a~.~y nor will it ke chemically d2graded by the corrosive liquors of a chlor-alkali cell. Itremains, coating the ~ores, and imparting a level of hydrophilicity such thatgas generated in the electrolytic process will not displace electrolyte in thepores of the diaphracgm. This remedies a deficiency of porous PTFE diaphragms of ~rior art where dewet or "~as locked" areas with blocked electrolyte flow oftencause voltage rise and ultimate shutdo~n. The pcrous PTFE diaphrayms of theprior art do not involve a ehemically inert polymer tc impart hydrophilicity or,if they do, they employ a higher equivalent weight polymer which, because ofmic~lle size, cannot ~enetrate to uniformly coat the interior surfaces of verysmall pores in the multilayer EPTFE composite of this invention.

,: .

t "
.. .
.
......
:

~;
: ,; , ' ,. ~ . . . :
.

WO90/13~9~ ?~ 12~- PCI`/US90/023~9 It has been found thst a representative k fsctor obtained with the porous composite of this invention is about 0.32 Vm2/kA. This results in considerable power savings over the 0.55 Vm2/kA k factor Or asbestos or the 0.48 Vm2/kA k factor of modified asbestos diaphragms in current use.
By the layered approach, with proper selection of the EPTFE component layers, coupled with coating the interior and exterior surfaces of the multilayered article with a perfluoro ion exchange resin, an asyrnmetric structure that will operate well as an electrolytic separator or filtrstion medium can be created.

, , ',` ,. :
,. - . : -.: ' . . '~ -~

, -:

-13- r~ , f ~ ~ 7 W O 90/13593 PCr/US90/02349 As used in this a?plication, the term "asymmetric structure" rneans a multilayered, composite structure in which at least two of the multiple layers have different microporous structures, i.e., at least two of the layers in the composite have specific gravities which differ by at least 5%.
Such an asymmL-tric structure has utility as a diaphragm in a chlor-alXali cell. The preferre~ method of use is to orient the diaphragm such that the larger pore size, as indicate~ by a lower methanol bubble point (ASTM F316-80), ; faces t'ne anode compartment. In this mode, a higher current efficiency is achieved. This is probably because the linear velocity of the electrolyte, in the direction of the cathode, increases as the electrolyte moves through thediaphra~m anc its effect on counterac~ing the back migration of the hydroxyl ionis corre pondinyly enhanced.
A detailed ~escription of the invention and preferred embodiments is best provided with reference to the dra~-ings and the examples which follo~.
Fiy. 1 is a schematic diagram of a chlor-alkali cell 8 containing anode ~4 and cathode 16 in operation. A multilayered composite according to the invention and useful as the ~.iaphra~m lO ir. such a cell is shot~n. Fig. 2 is aphotomicrograph talcen at 45x masnification of a cross-section taken in the thiclcness or Z-direction through the multilayered sheet 10. The individual J20 layers 12 wllich make up sheet 10 are discerniU e. Fig. 3 is a photomicrosraph taken at 5000x maynification of the sheet 10 shown in Fis. 2. Therein the microstructure of nodes and fibrils coated with a perfluoro ion e~cchal?~e resinis shown.
Figs. 4 and 5 are photomicrograp~s taken at 50x and 500~x, respectively, of an asymmetric sheet accordin(~ to the invention. The overall sheet 20 is seen in j Fig. 4 to be made u~ of thick layers 22 and relatively thinner layers 24.Layers 24 have a lower specific gravity than layers 2~. The coate~
microstructure of this cGmposite is shoun in Fig. 5 which was taken at 5000x magnification.
The examples which follo~ are intended to be illustrative of the invention but not limitative in any ~ay.
Attempts to duplicate the 95-95% current ~fficiency experiments were successful, but in some cases wherein membrane thickness was less than 90 mils, they gave a current efficiency as mNch as 4% lower.

.
, .. ' : :
: ' :

W O 90/1359~3~ , 3~?~ PCT/US90/02349 In ~eneral, in ordcr to ma1ce the co~posite articles of the inv~ntion, ~panded PTFE sheetin~ havin~ a microstructure characterized by a series of nodes interconnected hy fibrils and 'navins a Gurley air flo~ of 0.8 sec. to 27 sec., thickness betuean 0.2 mil and lO mil, methanol bubble point (ASTM F316-80)S bet~leen O . 7 p5i and ~0 psi, is wound around a cylinarical mandrel.
The EPTFE sheet suitably has a thickness of 25-125 micrometers, or l.0-~.0 mils.
The len~th and diameter of the mandrel can be varied tG give the desired dimension~ for the finished sheet. Multiple layers, ~reatèr than 4, are woundonto the mandrel and the nurnber of layers is varied to giv~ the desired thickness ana uniformity. T'ne membrane is restrained at the ends of the mandrel by mec1lanical clamps or bands.
rne layers of EPTFE she2ting are 'oonded together by immersion in a molten salt bath at a tem~erature above t'ne cystalline melt point G. EPTFE. ~he layered EPTFE com~osit~ i allowed to CGOl slcwly on the mandrel in air. Th~layered composite is cut and removed from the manorel to yielo a flat sheet.
Impregnation of the layered flat sheet is carried out by using an alcohol basea liquid composition of perfluorosulfonic acid ion exchange polym~-r.
Polymer solids loadin~; in the liSuid composition can range from 0.5% up to 10%.
U2 to 8% surfactant or surractant blend can ~e included to aid in distributionof the iO11 exchanse polym~r and in initial water ;~ettin~ of the finished product. The layer~d flat sneet is fully wet uith impreynatin~y liquid comp~sition. The impre~nant is introduced from one side so as to avoid trapping air inside the structure.
Another f~ature of the preferr2a coatin~ process is to evacuate most of the air from t~e EPTFE before the li~uid co~position is added to just one surfac~ o~the EPTFE. A suitable vacuum is an a'~solute pressure of l25 mm Hg, but the absolute pressure is not critical. It is believed that this feature means that the li~uid composition ves in just one direction through the EPTFE anà it ~30 encounters less air which coulà form air bubbles durina the coatin~ process and thus prev6nt ccatin~ of local areas of tne EPTFE.
Th~ liquid composition used to coat the exterior surfaces and at least a portion of its interior pore surfaces with 2 perfluoro ion exchange polymerpraferably contains an or3anic compound or com~ounds tnat enable tn~ compositioll to fully wet the surfac~ of a full ces;si~y E~ E UpOI1 w'nic1l it is ~Gur~d. '~he com,oour)d shall b2 sGluble in water and com~atible with t;n2 solvents use~ in the liquid composition. 'l~ne com~oun~ shall wet EPIFE ar1d th2 perfluoro ion ~xcnan~e .,- . - . - :
. . ., 1 --: - :. .
':;., : . ' ~ , . , -.:
- . . . . . .: . - .

W O 90/13593 c~ 9 Q ~'~ 7 P ~ /US90/02349 poly~er, and shall have a boiling point above th~ boiling point of water butbelow the decomposition temperature of the perfluoro ion exchanye polymer. Thiscombination of prolxrties enables the liquid composition to spread evenly anduniformly over the external surfaces and internal pore surfaces of the EPTFE andto dry without drawing up into discrete droplets which would contribute tonon-uniform coating of the EPTFE with ion e~;change resin. A suitable oryanic compound is l-methoxy-2-propanol.
Exces3 impresnating liquid is removed from the outside of the sheetins by wiping or squeegeeing. The wet sheet is then restrained at its edges to prevent shrinkage as the sheet is drie~.
Drying can be carried out in air at temperatures ranging from 15C to 120C. The preferred drying conditions are in air at 23C. A post c3rying ba'.~e can be carried out at temperatures between 30C and 150C.

q~lo relatively thick multilayer EPTFE composites were prepared by different methods. The first was prepared by layering nine full density, extruded PTFE
tapes and biaxially e,cpanding and sintering this composite accordincJ to the disclosure of U.S. Patent No. 4,187,390. The second composite was prepared by winding multiple layers of biaxially ex~anded, sintered PTFE sheeting onto a 3 inch diameter aluminum mandrel and sintering these layers together by immersion in a molten salt solution for one minute at 370C. The first sample will be referred to as "layered before e~pansion composite" and the second will be referred to as ~'layered after expansion composite", the latter being the precursor of the composite of this invention.
Specimens were cut from both of these composites and a specimen was cut from the single biaxially expanded sheet which was used to ma~e the layered after e~pansion composite. S~ecimen size was chosen for each type of sheeting to give approximately the same total weight per specimen. The biaxially expanded PTFE sheetins specimens were cut so as to produce 12" x 12" squares, the layered before expansion composi'e specimens were cut to yield 4 ~" x 4 ~"
squares and the layered after expan3ion specimens were cut to yield 3~" x 3 ~"
squares. A group of SiX specimens was taken from at least two different areas of the bulk sheet of each sample type to illustrate across-the-web variation.

.
.
`'` ;` ' ', : -.. . .

W~ 90/13~g3 -15- PC~/US90/02349 c '9J' ; `
t.~, , Testing was carried out to determine the ur;iformity of each type of sheeting with respect to a number of physical characteristics. Each specimen wasweighed to + 0.002g accur~cy, thickness was measured with a snap gauge, and air flow permeability was measured using a Gurley Densometer according to ASTM
5 D726-58. Thickness and air flow measurements were taken in at least four areasof each specimen. Density values were calculated from weight and thickness data.Results of this testing can be seen in Table l.
Range is defined as the difference between the high and low individual values for a set of data. The percent range is the range value divided by the 10 average and is a measure of the full scope of values normalized to an averagevalue of one. Similarly, percent standard deviation is the standard deviation for a set of data divided by the average and is a measure of the scatter of data normalized to an average value of one. These vslues are normalized measures of uniformity for a given physical property.
Normalizing the range and standard deviation values enabies a comparison of the levels of uniformity of the different films with respect to each physi~alproperty examined.
Referring to Table 1, percent range and percent standard deviation values for the layered after expansion composite were lower than the layered before 20 expansion composite and the biaxially expanded single layer film in thickness, density and air flow measurements. This demonstrates that the technique of layering an expanded PTFE sheet to form a thick composite, surprisingly and un-expectedly, gives a much more uniform structure than a thick sheet prepared by layering full density PTFE tapes and then expanding. The data above demonstrates~` 25 that by layering expanded PTFE sheeting, variations in the sheet are averaged, giving a much more uniform layered composite than the original single layered sheet.

-; - . ~ .
. , ........... - .-. .
:: .: : . :. . . .

" ~ ' ? 3 ~7 Biaxially Layered Before Layered After Expanded Expansion Expansion Pl`FE Composite Composite 5Thicknes~ Average 3.53 39.3 69.7 High 4.05 47.0 72.5 Low 3.00 23.0 66.7 Std. Dev. 0.223 4.65 1.91 Std. Dev. % 6.3 11.8 2.7 Range 1.05 24.0 S.8 Range % 29.7 61.1 8.3 Density Average 0.290 0.264 0.302 High 0.308 0.293 0.309 Low 0.275 0.248 0.294 lSStd. Dev. 0.010 0.011 0.005 Std. Dev. % 3.3 4.3 1.5 Range 0.033 0.045 0.015 Range % 11.4 17.0 5.0 Gurley Air Flow Avera~e9.7 68.8 94.6 20High 12.3 82.6 97.5 Low .. 5 56.2 90.4 Std. Dev. 1.1 5.6 1.9 Std. Dev. % 11.2 8.1 2.0 Range 4.7 26.4 7.1 Range % 48.9 38.4 7.S

~ . - :.
: ~ .. . . . . . .

WO 90/13593 -1~l- PCr/US90/0234s A section of expanded YTFE sheeting having an average methanol bubble point of 11.8 psi (ASTi~ F316-80), an air flcw of approximately 5.l seconds as measured by Gurley Del~someter (ASTM D726-58) and thickness of approximately 54.4 mil was wound onto an aluminum mandrel (3.5" o.d. and 9" in length). A totnl of twenty (20) layers of EPTFE sheeting were wound onto the mandrel. This sheeting was restrained by placing hose clamps around the circumference of the mandrel at each end. l`he layers of EPTFE sheet were bound together by immersing the wound mandrei in a molten salt bath at 370C for one minute. The 10EPTFE wound mandrel was then allowed to cool slowly in room temperature air.
The layered EPTFE was cut along the length of the mandrel and removed to form a flat sheet. A section of this layered EPTFE sheet was impregnate(3 with a liguid composition comprising 3.2% perfluorosulfonic acid polymer (equivalent weight 920 to 950) dcrived from n precursor copolymer oî tetrafluoroethylene and1~CF2=CF-O-CF2-CF-O-C~2CF2SO2F, 1.2% Triton X-l00 non-ionic surfactant Rohm and Haas) ~nd 0.4% Triton CF-5~ non-ionic surfactant (l~ohm and Haas) in ethyl alcohol. The wet loyered EPTYE structure was restrained to prevent shrinkage and was allowed to dry at approximately 23C overnight.
20The resulting composite diaphrHgm contained 9.096 perfluoro sulfonic acid polymer by weight.
The EPTFE/perfluoro ion exchange polymer composite was wet with a 0.196tetraethylammonium perfluorooctane sulfonate solution in water to facilitate start up.
25The sample described above was tested in a laboratory scale cell consisting of a glass anode compartment sepnrated from an acrylic cathode compartment by the diaphragm. The diaphragm wus sealed in place using EPDhl gaskets. The anode compartment consisted of an anolyte chamber containing about 500 milliliters of anolyte, A DSA~ anode obtained from Oxytech, Inc., a cell heater for 30temperature control, a brine feed line and a vertical tube connected to a chlorine outlet. This tube allowed disenga~ement of the chlorine and allowed an anolyte head of up to about ~0 cm to be formed before overflow of the anolyte occured.
The cathode compartment included a heavy gauge mild steel wire screen (of a typeused in commercial ~liaphragm cells) tack-welded to a mild steel current distributor, a hydrogen disengagement area and n catholyte discharge. The cell : . .. .. ~.. .
. . . ..

:. :: . :: , .

: . .
' : : , ~, ' WO 9O/13593 -18- PCr/US90/02349 2 ~ 9 7 had an active diaphragm area of L}5 square centimeters, was controlled at a temperature of about 85C, and was operated at a current of 11.25 amperes, resulting in a current density of 2.5 kiloamperes/square meter or 232 amperes/
square foot.
Cell voltages were measured between points near the entrances of the electrodes into the cell bodies, and current efficiencies were calculated from the ratio of caustic produced over a sixteen hour period (from the total sample weight and titration to determine caustic concentration) to the number of coulombs supplied to the cell during this time. Electrical energy consumption of the cell is reported in kilowatt-hours per metric ton of caustic produced, which is calculated from the cell voltage and caustic current efficiency (CE) by the following equation:
Electrical Energy Consumption = 67010 x ~cell voltage)/(CE) in which the cell voltage is in volts and caustic current efficiency (CE) in per-centage rather than fractional units. Another important parameter reported is k factor, which is the slope of the cell voltage versus the current density at current densities greater than one kiloampere per square meter. This normalizes data taken at different current densities, because voltage is linear with current density for all practical current densities above one kiloampere per square meter. For simplicity, the k factor was estimated by the following correlation:
k factor = (cell voltage - 2.3 volts)/~current density) in which the cell voltage is measured in volts and the current density in kilo-amperes per squar~ meter. For samples operated at different current densities, the intercept was found to always be slightly greater than 2.3 volts, implying that this estimate of k factor gives an upper limit to the true value of the slope.
Typical operating conditions include an exit caustic concentration of flpproximately 10% and salt conversion of 52-55%. Brine was fed into the cells at a rate that was controlled to produce nominally 10% (by weight) caustic in the catholyte. Two kinds of brine were used in the tests: membrane quality brine, in which calcium and magnesium levels were kept below 50 ppb total, and diflphragm quality brine, in which the total calcium and magnesium were main-tained between 0.9 and 1.8 ppm. The diaphragm quality brine csme from two sources: spiking the membrane quality brine with calcium and magnesium salts, snd filtered brine from an operating asbestos diaphragm plant.
The sample diaphragm described above was installed in a laboratory cell while wet with the water/surfactant solution. Membrane quality brine was allowedto flow through the diaphrngm overnigh~ without applied current. The sample was "

: : . .

..

, . .. : .... -.. : . . : : .
. . , . , . . ~

w090/13593 ,?~ r,~'`,7 -19- pcr/lls9olo234 removed from ~he cell for several hours for cell modifications, then replaced and the brine feed restored. After a total o~ ~hree days from the initial installation, a current of 11.25 amperes was applied. For the first thirty days of operation, the average cell voltage was 3.12 volts and caustic current efficiency was 95.4%, while 5 the cell produced l0.0~6 ca~lstic. Over the next thirty days of operation, th~ cell voltage averaged 3.07 volts and the current efficiency was 95.8% at ~n average of 10.2% caustic production. For days ~1 to 132, the average cell voltage w~s 3.14 volts and current efficiency was 95.6% while producing an ~verage of 10.296 caustic. At this time, the cell cracked and the anolyte compnrtment drained.
10 Less thAn one-third of the diaphragm remained in contact with the electrolytes.
The diaphragm was removed, placed in a mixture of brine and surfactant, then installed in a new cell after about a one-week delay. The diaphragm was operatedin the new cell for an additional 60 days with an sverage cell voltage of 3.18 volts, an average current efficiency of 95.7% and an average caustic concentration of 15 10.1%. These data are summarize<l with power consumptions snd k factors in Table 2. After the cell hydraulics stabilized, the daily anolyte head measurements averaged 23 centimeters with a standard deviation of 3.3 centimeters.

Caustic Cell kWh/ k factor Days wt. % CE~6 Volts MT NaOH V/(kA/m2) 0-30 10.0 95.4 3.12 2190 0.33 31-60 10.2 ~5.8 3.07 2150 0.31 61-132 10.2 95.G 3.14 2200 0.34 142-201 lO.1 95.7 3.18 2230 0.35 Avg. 10.1 95.6 3.13 2190 0.33 EXAMPLE 3 (Comparative) A long section of expanded PTPE sheeting, the ssme as in Example 2, with an average methanol bubble point of 11.8 psi (ASTM F316-80), sn air flow of approximately 5.1 seconds as measured by Gurley Densometer (ASTM D726-58) and thickness of approximately 4.4 mil was wound onto an sluminum mandrel (3 1/2"
o.d. and 9" in length). A total of 20 layers of EPTFE sheeting were wound onto the mandrel. This sheeting was restrained by plscing hose clamps around the circumference of the mandrel at each end. The layers of EPTFE sheeting were bonded together by immersing the wound mandrel in a molten salt bsth st 370C
for one minute. The EPTFE wound mandrel was then allowed to cool slowly in room temperature air.
. '' .

.

: .
- . . .
: : . .. ~ .

. .
:~ ' . . ' ' , , ' ' ' -WO 9O/13~93 -20- ~, 1 ~ ', 7 Pcr/usgo/o2349 The layered EPTFE was cut along the length of the msndrel and removsd to form a flat sheet. A section of this layered EPTFE sheet was wet with isopropyl alcohol and placed in a permeability testing ~pl~arstus. The high headchamber of the cell was quickly filled with 0.1% tetraethylammonium perfluoro-octane sulfonate in water and this solution was allowed to flow through the sample and displace the isopropyl alcohol. After the sample was fully wet with solution, permeability was measured.
The resulting diaphragm was 70 mils thick, had a Gurley air flow of 66 sec.
and a permeability of 0.412 reciprocal hours to 0.1% tetraethylammonium ~0 perfluorooctane slllfonate in water when measured at 23C and at 8 20 cm hend height differential. This diaphragm sample contained no perfluoro ion exchange polymer.
The diaphragm was installed in a laboratory cell while wet with the sur-factant/wnter solution and tested as described in Example 2. Membrane quality brine was allowed to flow through the diaphragm overnight without applied current.
The current was started and increased to 11.25 ampere over a ten-minute period.
The initial cell voltage at full current was 3.12 volts. After one day on load, the cell was producing 10.0% caustic at 2.76 volts with a c~Jrr2nt efficiency of 86.596.
After two days, the cell was producing 11.0% caustic at 3.48 volts with a current efficiency of 94.0%. Thc cell voltage increased nbove the equipment limits (about 10-15 volts) over the next 8 hours, causing the current to be interrupted. The cell could not be restarted without exceeding the equipment's voltage capacity.
EXAiUPLE 4 A section of expanded PTFE sheeting with an average methanol bubble point of 11.8 psi (ASTM F316-80), an air flow of approximately 5.1 seconds as measuredby Gurley Densometer (ASTM D726-58) and thickness of approximately 4.4 mil was wound onto an aluminum mandrel (6" o.d. and 9" in length). A total of twenty layers of EPTFE sheeting were woun~ onto the mandrel. This sheeting was restrained by placing hose clnmps around the circumference of the msndrel at each end. The layers of EPTFE sheeting were bonded together by immersing the wound mandrel in a molten salt bath at 370C for one minute. The EPTFE woun~l mandrel was then nllowed to cool slowly in room temperature air.
The layered EPTFE was cut along the length of the mandrel and removed to form a flat sheet. A section Or this layered EPTFE sheet was impregnated witha liquid composition comprising 3.2% perfluoro sulfonic acid polymer (equivalentweight 920 to 950), as in Example 2, 1.2% Triton X-100 non-ionic surfactant (Rohm :, '`

. -.
, :-~ :: . . -, - . . .
.

W090/13593 ~` r~ Z~ PCr/US90/0234~

and Haas), 0.4% Triton CF-54 non-ionic surfactant (Rohm snd Haas) and 0.6%
tetraethylammonium perfluorooctane sulfonate in ethyl alcohol. The wet layered EPTFE structurc was restrained to prevent shrinkage and was allowed to dry at approximately 23C overnight.
The EPTFE/perfluoro ion exchsnge polymer composite was wet with a 0.1%
tetraethylammonium perfluorooctane sulfonate solution in water to evaluate liquid permeability.
The resulting composite di~phragm contained approximately S.5% perfluoro sulfonic acid polymer by weight and hnd a permesbility of 0.085 reciprocal hoursto 0.1% tetrsethylsmmonium perfluorooctane sulfonate in water when measured at 23C snd at a 20 cm head height differential.
The diaphragm was installed in a laboratory cell while wet with the surfactant/water solution snd tested as described in Example 2. Membrane gualitybrine was allowed to flow through the diaphragm without applied current overnight.
Thc current was started and increased to 11.25 amperes over a five-minute period.
The initial cell voltage at full current was 2.76 volts. After five days on load, the cell was producing a 9.6% cnustic at 2.99 volts with a current efficiency at gO.2%.
The anolyte head was 57 centimeters. The cell test was then terminated, EXAMYLE 5 (Comparative) A section of expanded PTFE sheeting, the same as in Example 4, with an average methanol bubble point of 11.8 psi ~ASTM F316-80), an air flow Or approximately 5.1 seconds as measured by Gurley Densometer (ASTM D726-58) and thickness of approximately ~.4 rnil was wound onto un aluminum mandrel (6" o.d.
and 9" in length), A total of twenty layers of EPTFE sheeting were wound onto the mandrel, This sheeting was restrained by placing hose clamps around the circumference of the mnndrel at ench end. The layers c-f EPTFE sheeting were bonded together by immersing the wound mandrel in a molten salt bath at 370~
for one minute. 'rhe EPTFE wound mandrel was then allowed to cool slowly in room temperature air, The layered EPTFE was cut along the length of the mandrel and removed to form a flat sheet, A section of this layered EP'rFE sheet was wet with isopropyl alcohol and placcd in u permeability apparatus. The high head chumber of the cell was quickly filled with 0.1% tetraethylammonium perfluorooctane sulfonate in water and this solution ~Nas allowed to flow through the sample anddisplace the isopropyl alcohol. After the sample was fully wet with solution, permeability was mcasured.

': .;
.~. . . . i.

WO 90/i3~93 -22- PCr/US90/0~349 The resulting diaphragm was 65 mils thick, hsd a Gurley nir flow of 97 sec.
and a permeability of 0.4~8 reciprocel hours to 0,1% tetraethylammonium perfluorooctane sulfonate in water when measured at 23C and at a 20 cm head height differential. This diaphragm sample contained no perfluoro ion exch~nge 5 polymer.
The diaphragm was installed in a laboratory cell while wet with the sur-fsctant/water solution and tested as described in Example 2, Membranç quality brine was allowed to flow through the diaphr~gm overnight without applied current.
The current was started and increased to 11.25 amperes over a tw~minute period.
10 Thc initial cell voltage at ful! current was 3.01 volts. After one day on line, the cell W&S producing 8.8% caustic at 3.01 volts with a current efficiency of 90.1%.
During the next û hours, the cell overvoltaged, shutting off the current supply to the cell. The cell could not be restarted.

A section of expnnded PrFE sheeting having an average methanol bubble point of 8,42 psi (ASTM F316-80), an air flow of approximately 4 seconds as measured by Gurley Densometer (ASTi~l D726-58) and thickness of approximately 3,8 mil was wound onto an aluminum mandrel (l9" o.d. and 20" in length). A totalof eighteen luyers of EPTFE shecting were wound onto the mandrel. This sheeting 20 was restrained by placing hose clamps around the circumference of the m~ndrel at each end. The layers of EPTFE sheeting were bonded together by immersing the wound mandrel in a molten salt bath at 365C for one minute. The EPTFE wound mandrel was then nllowed to cool slowly in room temperature sir.
The layered EPTFE was cut along the length of thc mandrel and l~emoved 25 to form Q flat sheet. A section of this layered EPTFE sheet was impregnated with a liquid composition comprising 3.3% perfluoro sulfonic ac;d polymer (equivalentweight g20 to 950), a.s in Example 2, 0.4% Triton X-100 non-ionic surfactant (Rohm and ~laas), 0.1% Triton CF-54 non-ionic surfactant (Rohm and Haas) snd 0,6%
tetraethylammonium perfluorooctane sulfonate in ethyl alcohol. The wet layered EPTFE structure was rcstrained to prevent shrinkage and was allowed to dry Qt approximstely 23C ovcrnight and then baked at 100C for 7 minutes, 'rhe resulting composite diaphragm was approximately 65 mils thick, had a Gurley air flow of 85 sec. nnd a permeability of 0.653 reciprocal hours to 0.1%
tetraethylammonium perfluorooctane sulfonate in water when measured at 23C
35 and at a 20 cm head height differenti~

j , . .
. .

.
~: .
-WO 90/13~93 PCr/US90/02349 , ., i 2 ~
The diaphragm WAS instslled in a laboratory c~ll while dry and tested as described in Exsmple 2. Water was red to the anolyte compartment for four hours.The water feed was stopped and membrane qu~lity brine was allowed to flow through the diaphragm over a two-day period without applied current. During four5 days of operation on membrane quality brine, the cell produced an average of 9.5%
caustic at an average cell voltage Or 2.95 volts. The average caustic current efficiency was 91.2%, and energy consumption was 2168 kilowatt hours per metric ton of caùstic. The anolyte head was steady at about seven centimeters. The brine feed to the cell was then switched to diaphragm quality brine. Over the next 10 twenty days of operation, the cell produced an average of 9.9% caustic at an average of 3.26 volts. The caustic current efficiency averaged 94.9% and the energy consumption was 2304 kilowatt hours per metric ton of caustic during the period operated on this brine. The cell was terminated after a total of 27 days on line. The anolyte head was 62 centimeters when cell operation was terminated.
EXAI~IPLE 7 A section of expanded PTFE sheeting having an average methanol bubble point of 7.0 psi (ASTi~S F316-80~, an air flow of approximately 4 seconds as measured by Gurley Densometer (ASTM D726-58) and thickness of approximately 4 mils was wound onto an aluminum mandrel (3.5" o.d. and 9" in length). A total 20 of twenty layers of EPTFE sheeting were wound onto the mandrel. This sheetingwas restrained by placing hose clamps around the circumference of the mandrel ateach end. The layers of EPTFE sheeting were bonded together by immersin~ the wound mandrel in a molten salt bath at 370 for one minute. The EPTFE wound mandrel was then allowed to cool slowly in room temperature air.
The layered EPTFE was cut along the length of the msndrel and removed to form a flat sheet. A section of this layered EPTFE sheet was impregnated witha liguid composition comprising 3.75% perfluoro sulfonic acid polymer (equivalent weight 920 to 950), as in Example 2, 1.2% Triton X-100 non-ionic surfactant (Rohm and Haas) and 0.4% Triton CF-54 non-ionic surfactant (Rohm and H~as) in ethyl alcohol. The wet layered EPTFE structure was restrained to prevent shrinkage andwas allowed to dry at approximately 23C overnight. The composite was then baked at 80C for five minutes.
The EPTFE/perfluoro ion exchange polymer composite was wet with a 0.1%
tetraethylammonium perfluorooctane sulîonate solution in watcr to evaluate liquid permeability.
:;~

, .

.. .... . .
.: .. '.-:'' . ; .... ''-. - : .
.. . - . . ~....... . ..
, . , . . ~ , ~ :
. :~ - .. . .:- : . . . .

-2~
WO 9Q/13~93 , PCr/U~90/02349 The resulting composite diuphragm contained 13.796 perfluoro sulfonic acid polymer by weight and had H permeability of 0.565 reciprocal hours to 0.1%
tetraethylammonium perfluorooctane sulfonatc in water when measured at 23C
and at a 20 cm head height differential.
S The diaphragm was installed in a laboratory cell while wet with the surfactant/water soiution and tested as described in Example 2. Membrane quFlitybrine was allowed to flow through the diaphragm overnight without applied current.
Over one hundred and ~wenty-one days of operation on membrane quality brine, thecell produced an average of 10.2% caustic at an average cell voltage of 3.17 volts.
The average caustic currcnt efficiency was 95.1% and energy consumption was 2231 kilowatt hours per metric ton of caustic. The anolyte head was steady at about six centimeters. The brine feed was then switched to a batch of brine which was contaminated with particulates. This resulted in partial plugging of the diaphragm; cell operation degenerated and was terminated.
~5 EXAMPLE 8 A section of expanded PTFE sheeting having an average methanol bubble point of 11.8 psi (ASTM F316-80), an air flow of approximately 5.1 seconds as measured by Gurley Densometer (ASTM D726-j~) and thickness of approximutely 4.4 mils was wound onto an aluminum mandrel (3.5" o.d. and 9" in length). A total i20 of seventeen layers of EPTFE sheetin~ were wound onto the mandrel. This sheeting WQS restrained by placing hose clamps around the circumference of the mandrel at each end. The layers of EPTFE sheetin~ were bonded together by immersing the wound mandrel in a molten salt bath at 370C for one minute. The EPTFE wound mandrel was then allowed to cool slowly in room temperature air.
The layered EPTFE was cut along the length of the mandrel and removed to form a flat sheet. A section of this layered EPTFE sheet W8S impregnated witha liquid composition comprising 2.7% perfluoro sulfonic acid polymer (equivalentweight 920 to 950), as in Example 2, 0.9% Triton X-100 non-ionic surfactant (Rohm and Haas) and 0.3% Triton CF-54 non-ionic surfactant (Rohm and Haas) in ethyl 3~ ~Icohol. The wet layered EPTFE structure was restrained to prevent shrinkage nnd was allowed to dry at approximately 23C overnight. The composite was then baked at 100C for 5 m;nutes.

... . . .
: : :
, .

,. ''. ~,, :

WO 90/13593 25 PCr/US90/02349 The~PTFÉ/perfluoro ion exchange polymer composite was wet with ~ 0.1%
tetr~ethylammonium perfluorooctane sulfonate solution in water to evaluate liquid permeability.
The resulting composite diaphragm contained 6.0% perfluoro sulfonic acid S polymer by weight snd had a permeability of 0.490 reciprocal hours to 0.1%
tetraethylammonium perfluorooctane sulfonate in water when measured at 23C
and at a 20 cm head height differential.
The sample described above was installed wet in a laboratory cell and tested as described in Example 2. Membrane brine was fed to the anode compartment 10 overnight before current was applied. Over the first 35 days of operation, the cell voltage averaged 2.99 volts, the current efficiency was 95.0% and the caustic concentration was 10.3%. During the next 10 days, problems with the brine feed system caused the caustic concentratiGn to increase to 18.4%, then decrease to 8.8% before the cell was stabilized again. For the next 60 days, the cell voltage averaged 3.10 volts, the current efficiency was 95.1% and caustic concentration was 9.7%. The cell WflS ullowed to operate for a total of 186 dsys, during the last 85 of which the cell operated at an average cell voltage of 3.10 volts, an average current efficiency of 94.7% and an average csustic concentration of 9.9%.
Overall, excluding the ten day caustic excursion, the cell operated at an average 20 of 3.07 volts, 94.9% current efficiency and 9.9% caustic. These results are summarized in Table 3.

Caustic Cell kWh/k factor Days wt. % (,E% Volts MT NaOH V/(kA/m 1-35 10.3 95.0 2.99 21100.28 36-45 caustic excursion to 18.4%
46-105 9.7 95.1 3.10 21800.32 105-186 9.9 94.7 3.10 21900.32 Avg. 9.9 94.9 3.08 21700.31 30 (excluding caustic excursion) "' : :. . , , . . . :
:: - , , : . . . . ;

.

. - . . , .- : :. - .: . .. :- -- . .

WO90/13593 ~$~ PCT/US90/02~9 - EXA~PLE 9 ~
A section of expanded PTFE sheeting with an average methanol bubble point of 9.9 psi (ASTIU F316-80), ~n air flow of approxilrately 6 seconds as meflsured by Gurley Densometer (ASTiU D726-58) and thickness of approximately 4.5 mil~ was wound onto an aluminum mandrel (19" o.d. and 21" in length). A total of twenty-two l~yers of EPTFE sheeting were wound onto the mandrel. This sheeting was restrained by placing hose clamps around the circumference of the mandrel at each end. The layers of EPTFE sheeting were bonded together by immersing the wound mandrel in a molten salt bath at 370C for one minute. The EPTFE wound mandrel was then allowed to cool slowly in room temperature air.
The layered EPTFE was cut along the length of the mandrel and removed to form a flat sheet. A section of this layered EPTFE sheet was impregnated witha liquid composition comprising 3.3% perfluoro sulfonic acid polymer (equiv~lentweight 920 to 950), as in Example 2, 0.4% Triton X-100 non-ionlc surfactant (Rohm and Haas), 0.1% Triton CF-54 non-ionic surfactant (Xohm and Ha~s) and 0.6%
tetraethylammonium perfJuorooctane sulfonate (Mobay) in ethyl alcohol. The wet layered EPTFE structure was restrained to prevent shrinkage and was alloweà to dry at approximately 23C overnight. The composite was then baked at 100C for 7 minutes.
The resulting composite diflphragm was approximately 80 mils thick, h~d a Gurley air flow of 95 seconds and a permeability of 0.36~ reciprocal hours to 0.1%
tetraethylammonium perfluorooctane sulfonate in water when messured at 23C
and at a 20 cm head hei~ht differential.
The sample described above was installed dry in a laboratory cell and tested as described in Example 2. The anolyte compartment was filled with water and the water was allowed to flow through the diaphragm for severQI hours. Membrane quality brine feed was then started and Qllowed to flow for several hours beforethe current was applied. Over the first two days of operation, the average cell voltage was 2.99 volts and current efficiency was 94.0% at 10.2% caustic with a 21 centimeter anolyte head. The brine feed was then switched to diaphragm quality brine, Over the next 3d days, there was Q steady increase in anode head and voltage. The average cell voltage over this tirne period was 3.13 volts an-3current efficiency was 93.1% at 10.1% caustic production and a 32 cm anolyte head. After the head and voltase stabilized, the cell was o~erate~ for 381 more days, for a total of 4~ days of operation. During this time, it e~erienced numerous shutdo~ns, caustic excursions and other up ets. Over-all, durin~ these 381 days, the cell produced an average of 9.gYO caustic, :
:

.

.: ~
. .
; , - ' . . . . ~

W 0 9~l3593 2 '~ 3 ?,rl -27- P ~ /U590/023495 at an average cell voltage of 3.25 volts, an average current e~flciency G~'...`
99.7~ and an anolyte head of 28 centimeters. ~ese result~ ~re sumrarized in Table 4.

TA~LE 4 Caustic Cell kWh/ k f~ctor Dsys wt. % CE% Volts MT NaOH V/(kA/m2) 1-2 10.2 94.0 2.99 2130 0.28 2-40 10.1 93.1 3.13 2350 0.33 41-421 9.9 94.7 3.25 2300 0.38 A section of expanded PTFE ~heeting having an average methanol bubble point of 10.5 psl ~ASTM F316-80), an air flow of 11 seconds as measured by Gurley Densometer (ASTM D7~6-58) and a 4.5 mils thickness was wound around an aluminum mandrel (6.0" o.d. and 9" in length). Eight layers of this sheetin~ were wound onto the mandrel. rhen a section of expanded PTFE sheetin~ with un average methanol bubble point of 7.1 psi (ASTM F316-80), ~n air flow of S seconds ag measured by Gurley Densometer (AS'rM D726-58) and a thicknes~ of 4 mils was wound on top of it. Seventeen layers of this sheeting were wound over the initial eight layers.
The EPTFE sheeting was restrained by placing hose clamps ~round the oircumference ot the mandrel at esch end. The layers of EPTFE were bonded together by immersing the wound mandrel in a molten salt bath at 367C for one minute. The EPTFE wound mandrel was allowed to cool slowly in room temperature uir. The exposed outer surface of the EPTFE which had been derived from the precur or with the lower methanol bubble point is herein designated side A. The reverse side is herein designated as side B.
The layered EPTFE structure was impregnated with a liquid composition of 3.3% perfluoro sulfonic acid polymer (equivalent wei~ht 920 to 950), as in Example 2, 0.4% Triton X-100 non-ionic surfactant l}~ohm and Hass~, 0.1% Triton CF-54 non-lonic surfactant (Rohm and Haas) and 0.6% tetraethylammonium perfluoro-octane sulfonate in ethyl alcohol. The wet EPTFE structure WBS restrsined to preve,nt shrinkage and was allowed to dry at 23C for 16 hours. The restraine-i EPTFE/perfluoro ion exehan~e polymer composite was then placed into a 100C
oven for 7 minutes for final dryin~.

~':
:,' .

:'- : .. . .. :
: . . .. .
. - ~ - . . -" `.' ' ': ' '' ', - ~ ' .
.. . . .

WO 90/13593 -28- ,~ P~/US9OtO2349 A strip 0.2~ inches wide and 6 cm long and 0.089 inches thick was cut from the dried composite structure (thickness and width figures were each the avera~eof three measurements slon~ the 6 cm length). The strip was weighed and density calculated. A razor incision was then made at one end of the strip in a plane 5 parallel to the surface (i.e., perpendicular to the Z axis or thickness direction).
The separation was propagated by peeling one section from the other; the multilayer structure being delaminated ut the interface between two discrete precursor layers. The section containing the A side is herein designated sectionAS. The section containing the B side is herein designated E~S. Thickness (the 10 average of three measurements along the 6 cm length) and weight of each section was measured and the density of each section was calculated. Results are shown in Table 5.

Thickness (inches) Density (gm/cc) 15 Composite structure 0.89 .350 AS section 0.42 .308 BS section 0.47 .388 These measurements demonstrate the asymmetry with respect to density and structure which can be achieved by this invention.
A sample of EPTFE/perfluoro ion exchange polymer composite prepared as above was then wet with n solution of 0.1% tetraethylammonium per~luorooctane sulfonate in water to evaluate liquid permeability. Permeubility was measured with a 20 cm head of water containing 0.1% tetraethylammonium perfluorooctane sulfonate at 23C. No measurements were taken until excess surfactant had been f~ushed from the diaphragm as evidenced by the diaphragm becoming uniformly translucent with no opaque or hazy regions. With the sample oriented in the permeability tester with flow in the direction from the A side, through the diaphra~m towards the B side, the permeability measured was 0. 20 re~iprocal hours.
A sample of the composite described above wns instulled dry in a laboratory cell with the A side oriented towards the anode. Water was fed to the anode compartment for two hours, then the feed was changed to membrane quality brine which continued overnight bcfore power was applied. The celJ wus operuted ror 11 days producing an avcrage ol ~.~% caustic. The cell volta~e avernged 3.01 `
''`

:
:.
' -: : .
; ~ -.
. :

WO 90/13593 -29- PCr/US9O/02349 volts and the current efficiency was 94.2%, for a total electrical energy con-sumption of 2140 kilow~tt hours per metric ton of caustic. The cell was shut down for maintenance to the laborAtory area, and the diaphragm was damaged when the cell was restarted. Results are summarized in Table 6.

Caustic Cell kWh/ k factor Dayswt. % CE_ Volts MT NaOH V/(kA/m2) 1-11 9.9 94.23.01 2140 0.28 A section of expanded PTFE sheeting having an average methanol bubble point of 10.5 psi (AST;VI F316-80), an air flow of 11 seconds as measured by Gurley Densometer (ASTM D726-58) and a ~.5 mil thickness was wound around an aluminum mandrel (3 3/8" o.d. and 9" in length). Nine layers of this sheeting were wound onto the mandrel. Then a section of expanded PTFE sheeting having 15 average methanol bubble point of 6.3 psi (ASTM F316-80), an air flow of 5 seconds as measured by Gurley Densometer (ASrM D726-58~, and a thickness of 3.7 mils was wound on top of it. 'Seventeen layers of this sheeting were wound over the initial nine layers.
The EPTFE sheeting was restrained by placing hose clamps around the ` 20 circumference of the mandrel at each end. The layers of EPTFE were bonded , together by immersing the wound mandrel in a molten salt bath at 364C for one minute. The exposed outer surface of the EPTFE which had been derived from the precursor with the lower methanol bubble point is herein designated as side A. The reverse side is herein designated as side B.
The lsyered EPTFE structure was impregnated with a liquid composition of 3.3% perfluoro sulfonic acid polymer (equivalent weight 920 to 950), as in Example ' ~ 2, 0.4% Triton X-100 non-ionic surfactant (Rohm and Haas), 0.1% Triton CF-54 , non-ionic surfactant (Rohm and Haas) and 0.6% tetraethylammonium perfluoro-octane sulfonate in ethyl alcohol. The wet EPTFE structure was restrained to 30 prevent shrinkage and w~s allowed to dry at 23C ~or 16 hours. The restrainedEPTFE/perfluoro ion exchangé polymer was then placed into a 100C oven for minutes for final drying.
.' .

.
. ~ , ? ~
::
::-.`:;
: . :

,:
:: :: .

WO 90/13~93 3~ '~ J ~ Pc~/us9n/o2349 A strip 0.249 inches wide and 6 cm long and 0.071 inches thick was cut from the dried composite structure. Thickness and width fi~ures were each the averageof three measurements along the 6 cm le~gth. The strip was weighed and density calculated. A razor incision was then mflde at one end of the strip in a plane parallel to the surface (i.e., perpendicular to the Z axis or thickness direction).
The separation was propagated by peeling one section from the other, the multilayer structure being delamin~ted at the interface between the two discreteprecursor layers. The section containing the A side is herein designated sectionAS. The section containing the B side is herein designated section BS. Thickness(the average of three measurements along the 6 cm length) and weight oî each section was measured and the density of each section calculated. Results are shown in Table 7.

rhickness (inches) Density (gm!cc) Composite structure .071 .403 AS section ,024 .346 BS section .0~7 .433 The composite laminate as described was installed dry in a laboratory cell with the A side oriented towards the anode. Deionized water was slowly fed into the anolyte compartment until the cell was filled, and then was allowed to flow through the diaphrugm ror two more hours. The deionized water feed was s~opped and membrane quality brinc ~eed was started to the cell overnight before currentwas applied. After one day of operation, the cell voltage was 3.20 volts, declining to 3.12 volts on the second day. Over the first nine days of operation, the average cell voltage was 3.13 volts and average current efficiency was 96.7% at 10.2%
c~ustic. The average electrical energy consumption during this period was 2170 kilowatt hours per metric ton of caustic produced. Between the eleventh and twelfth days on line, the brine feed stopped due to a salt blockage, and the cell ran without feed for an undetermined amount of time. After the brine feed was restarted and the cell was allowed to equilibrate overnight, the cell voltage was 3.20 volts. Ovcr ~he ne.Yt seven clays. the volta~e slowly decreased to 3.11 volts.
Overall, from days thirleen to twenty, the cell produced an average of 10.1%
caustic with an avera~e cell voltage of 3.14 volts and an average current efficiency of 96.2%. 'rhis corresponds to an electrical energy consumption of 2190 kilowatt hours per metric ton caustic. ~tesults are summarized in l'~ble 8.
;

: , '' . , ~ ' ' - : .~ -~ , ' WO 90/13593 -31- Pcr/us9o/o2349 R " ~, TABLE 8 Caustic Cell kWH/ k factor Days wt. % CE96Volts MT NaOH V/(kA/m2) 1-9 10.2 96.73.13 2170 0.33 Brine feed blockage: high caustic production 13-27 10.0 96.5 3.14 2180 0.34 A section of expanded PTFE sheeting having an average methanol bubble point of 10.3 psi (ASTM F316-80), an air flow of 6 seconds as measured by GurleyDensometer (ASTM D726-58) and a 2.9 mil thickness was wound ~round ~n aluminum mandrel (3.5" o.d. flnd 9" in length). Eleven l~yers of this sheeting were wound onto the mandrel. Then a long section of expanded PTFE sheeting with an average methanol bubble point of 6.3 psi (ASTM F316-80), an air flow of 5 seconds as measured by Gurley Densometer (~STM D726-58), and a thickness of 3.7 mils was wound on top of it. Fifteen layers of this sheeting were wound over the initial eleven layers.
The EPTFE sheeting was restrained by placing hose clamps around the circumference of the mandrel at each end. The layers of EPTFE were bonded together by immersing the wound mandrel in a molten salt bath at 370C for one minute. The EPTFE wound mandrel was allowed to cool slowly in room temperature air. The exposed outer surface of the EPTFE which had been derived from the precursor with the lower methanol bubble point is herein designated as slde A. The reverse side is herein designated as side B.
The layered EPTFE structure was impregnated with a liquid composition of 3.2% perfluoro sulfonic polymer (equivalent weight 920 to ~50), as in Example 2,0.75% Triton X-100 non-ionic surfactant (Rohm and Haas), 0.25% Tritor~ (`F-5~
non-ionic surfactant (Rohm and Haas) and 1% tetraethylammonium perfluorooctsne sulfonate in ethyl alcohol. The wet EPTFE structure was restrained to prevent shrinkage and was allowed to dry at 23C for 16 hours. The restrained EPTFE/pcrfluoro ion exchange polymer composite was then placed into a 100C
oven for 7 minutes as a final treatment. The dried structure was 68 mils in thickness.
The EPTFE/perfluoro ion exchange polymer composite was wet with solution of 0.1% tetraethylammonium perfluorooctane sulfonate in water to evaluate liquid permeability. Permeability was measured with a 20 cm head of , , :, . - : -. . -.. :
, .:: : . . . :
... .

W090/13593 ~ r")rl PCr/U59n/0~349 ~J ~ "'J `
water containing 0.1% tetraethylammonium perfluorooctane sulfonate at 23C. No measurements were taken until excess surfsctsnt had been flushed from the diaphragm as evidenced by the diaphragm becoming uniformly translucent with no opaque or hazy regions; this rcquired about 1 hour. With the sample oriented in S the permeability tester with flow in the direction from A side through the diaphragm towards the B side, the permeability messured WQS 0.42 reciprocal hours.
The sample as described above was installed wet in a laboratory cell with the A side oriented towards the anode. Membrane quality brine was fed to the 10 anode compartment for 70 minutes before current was applied. Over the first two days of operation, the diaphragm performed with a cell voltage of 3.12 volts andn current efficiency of 91.8% at 9.9% caustic. The brine feed was then changed to diaphragm quality brine, and thc cell operated for 3 days before it was shut down for several hours for cell room repairs. The cell was restarted and allowed15 to operate for 27 more days, for a total of 30 days on diaphragm quality brine.
Over this time, the average cell volt~ge was 3.31 volts and the current efficiency was 93.6% while producing an average of 10.1% caustic. The average electrical energy consumption was 2370 kilowatt hours per metric ton of caustic. At the endof these 32 days, the anolyte head was 29 centimeters. The cell failed due to 20 electrical problems at 35 days on line, and the diaphragm was damaged when restart was attempted. The results are summarized in Table 9.

Caustic Cell kWh/ k factor Days wt. % CE%Volts MT NaOH V/(kA/m2) __ 25 1-2 9.9 91.83.12 2280 0.33 3-32 10.1 93.63.31 2370 0.40 A section of expanded PTFE sheeting with an average methanol bubble point of 11.8 psi (ASTM F316-80), an air llow of 5.1 seconds as measured by Gurley Densometer (ASTM D726-58) and a 4.4 mil thickness was wound around an aluminum mundre] (3.5" o.d. and 9" in length). Eleven layers of this sheeting were wound onto the mandrel.

:

:,`
' , ~ :

.. ~ , ,~., 1., ` ~ 7 PCr/US90/U~349 The EPTFE sheeting was restrained by placing hose clamps around the cir-cumference of the mandrel at each end. The layers of EPTFE were bonded together by immersing the wound mandrel in a molten salt bath at 370C for one minute. The EPTFE wound mandrel WflS allowed to cool slowly in room temperature air.
The layered EPTFE structure was impregnated with a liquid composition of 3.2% perfluoro sulfonic acid (equivalent weight 920 to 950), as in Example 2, 1.2%
Triton X-100 non-ionic surfactant (Rohm and Haas) and 0.4% Triton CF-54 non-ionic surfactflnt (Rohm and Haas) in ethyl alcohol. The wet EPTFE structure was restrained to prevent shrinkage and was allowed to dry at 23C for 16 hours. Therestrained EPTFE/perfluoro ion exchange polymer composite was then placed in a 100C oven for five minutes as f.nal treatment.
A 1.4% solution of perfluorocarboxyl ester polymer in 1,1,2-tricllloro-1,2,2-trlfluoroethane (Freon TF~', DuPont) was applied to one side of the restrained EPTFE/perfluoro ion exchange polymer composite and the composite was ailowed to dry at room temperature for one hour. The perfluorocarboxyl ester used was a copolymer of tetrafluoroethylene and CF2=CP-O-CF2CF-O-CF2CF2COOCH3 with an equivalent weight in the free acid form of approximately 650 to 750. Thetreated side is herein designsted side C. The composite was wet with a 36%
potassium hydroxide solution in a mixture of 85% water and 15% isopropanol and allowed to reside in this solution for 16 hours at room temperature to hydrolyzethe perfluoro carboxyl ester polymer to the potassium ion form of the perfluorocarboxylate polymer. The impregnated structure was then immersed in a bath of 0.1% tetraethylammonium perfluorooctane sulfonate in water for one hour to dilute and largely replace residual potassium hydroxide solution in the pores.
The resultant composite (on a dry basis~ contained about 9% perfluorosulfonic acid polymer and about 0.9% perfluorocarboxylic acid polymer, the latter being largely concentrated on side C. Permeability was then measured with a 20 cm head of water containing 0.1% tetraethylammonium perfluorooctane sulfonate at 23C. No measurements were taken until excess surractunt had been flusheà from the diaphragm as evidenced by the diaphragm becoming uniformly translucent with no opaque or hazy regions. The permeability measured was 0.23 reciprocal hours.
The diaphragm was inst~lled in a laboratory cell while wet with the sur-factant/water solution from the permeubility apparatus with the side design~ted :
:

-:: .
~'''.. -......... . . , .. : .
.. ~ ........... - . -.: .
: '.' :,`, - . , .: . .
:: . : : .
- . . - . . .

WO90/13593 ~ 7 PCI/US90/02349 ! `
side C oriented towards the cathode. Membrane quality brine W8S allowed to flow through the diaphragm overnight without spp~ied current. The curren~ was startedsnd increased to 11.25 amperes over a ten-minute period. The inltial cell voltage at full current was 2.91 volts. Over seventeen days on laad, the cell produced an average of 10.0% caustic at an average cell voltage of 2.97 volts. The average caustic current efficiency was 91.5% and electrical energy consumption was 2176 kilowatt hours per metric ton of caustic. The anolyte head was steady at about 12 centimeters.

A section of expanded PTFE sheeting having an average methanol bubble point of 8.4 psi (ASTI~I F316-80), an air flow of 4 seconds as measured by Gurley Densometer (ASTM D726-58) and a 3.8 mil thickness was wound around an aluminum mandrel (3.5" o.d. and 9" in length). Five layers of this sheeting werewound onto the mandrel. Then a section of expanded PTFE sheeting with an average methanol bubble point of 26.0 psi (ASTM F316-80), an air flow of 3.5 seconds as measured by Gurley Densometer (ASTM D726-58) and a thickness of 1.1 mils was wound around the first sheeting. Four layers of this second sheeting were wound over the initial five layers.
The EPTFE sheeting was restrained by placing hose clamps around the cir-cumference of the mandrel at each end. The lnyers of EPTFE were bonded together by immersing the wound mandrel in a molten salt bath at 365C for one minute, The EPTFE wound mandrel was allowed to cool slowly in room temperature air. The exposed outer surface of the EPTFE which had been derived from the precursor with the higher methanol bubble point is herein designated side A.
The layered EPTFE structure was impregnated with a lIquid composition of 3.3% perfluoro sulfonic acid polymer (equivalent weight 920 to 950), 0.4% TritonX-100 non-ionic surfactant (~ohm and Haas), 0.1% Triton CF-54 non-ionic surfactant (Rohm snd Haas) and 0.6% tetraethylammonium perfluorooctane sulfonate In ethyl alcohol. The wet EPTFE structure was restrained to prevent shrinkage and was al2Owed to dry at 23C for 16 hours. The testrained EPTFE/perfluoro ion exchange polymer composite was then placed into a 100C
oven for 7 minutes for final drying.
The resulting composite was approximately 17 mil thick, had a Gurley air flow of 56 seconds and a density oT 0.322g/cc.

, . . .
. ; . , - - . ,.

;. , ~ ~ : . .. :

WO 90/13593 _35_ PCl/US90/02349 o ~ r' ~. .: .. ....
A 2-inch diame~er disk of the composite W8S placed in a filter holder on a vacuum fl~sk ~nd connec~ed to a v~cuum pump. The composite disk w~s- installed so thst the A side faced the high pressure side. A 1% suspension of furned silica in deionized w~ter w~s directed to the A side of the composite and the v~cuum S pump turned on.
Approxim~tely 25 milliliters of the 1% fumed silica suspension was fi~tered.
'rhe filtr~te ~ppeared clear and free of turbidity. A single drop of ~iltrate was placed on a clean gJass slide and dried. Sim ilarly, a single drop of the 1%
suspension was placed on a clean glass slide and dried. Both of these samples were examined under high magnification (lOO,OOOx) using a sc~nning electron micro-scope. The fumed silica suspension specimen showed very sm~ pherical particles, approximately 200 Angstroms in diameter, and ~gg~omerates of these particles.
The filtrate sp~cimen was virtually free of the fumed silica particles. This demonstrates that this composite is a very effective filter for even very sm~ll 1 5 p~rticles.

A section of expanded PTFE sheeting having an average methanol bubble point of 25 psi (ASTM F316-80), an air flow of 13.4 seconds as measured by Gurley densometer (ASTM D726-58) an~ a thickness of 1.5 mils was wound onto a stainless steel mandrel (14.0" o.d. and 40.0" in length). Twe~ty layers of this sheeting were wound onto the mandrel. Then a section of expanded PTFE sheeting having an average methanol bubble point of 16.5 psi (ASTM
F316-80) an air flow of i3.5 sec. as measured by Gurley densometer (ASTM
D726-58) and a thickness of 4.8 mils was wound on top of the previous layers. Twenty-nine layers of this second type of sheeting were wound over - the initial twenty layers.
The EPTFE sheeting was restrained by placing hose clamps around the circumference of the mandrel at each end. The layers of EPTFE were bonded together by immersing the wound mandrel in a molten salt bath at 370C
for two minutes. The EPTFE wound mandrel was allowed to cool slowly in room temperature air. Then the cylindrical tube was cut longitudinally and removed from the mandrel to form a flat sheet. The exposed outer surface of the EPTFE which had been derived from the precursor with the lower methanol bubble point is herein designated side A. The reverse side ~; 35 is herein designated as side B.

:, :

: , .

, . . .
. ~ .

W O 90/13593 -35a~ PCT/US90/02349 The layered EPTFE structure was restrained in a frame and placed in a vacuum chamber with side B facing up. Vacuum was drawn in the chamber to 125 mm Hg, absolute pressure. While maintaining vacuum, a liquid composition of 3.3% perfluorosulfonic acid polymer (equivalent weight 600-700, extracted from 900-950 EW polymer using CF2ClCFC12 at reflux), 1.5% Triton X-100 non-ionic surfactant (Rohm & Haas), 0.5% Triton CF-54 non-ionic surfactant (Rohm ~ Haas) and 10% 1-methoxy-2-propanol in ethyl alcohol was introduced to the B side of the EPTFE sheet. The li~uid camposition was allo~ed to fully wet the layered EPTFE sheeting while under 125 mm Hg, absolute pressure vacuum. After impregnation the vacuum was released, excess liquid composition was removed from the EPTFE structure surface and the structure was allowed to dry (A side down) for 16 hours at 23C. The restrained EPTFE/perfluoro ion exchanye polymer composite was then placed into a 100C oven for 7 minutes for final dryiny.
The resulting composite diaphragm was approximately 121 mils in thickness, had a Gurley air flow of 612 sec. and a specific gravity of 0.61.
The diaphragm was installed in a laboratory cell while dry and tested as described in Example 2, with side B oriented toward the cathode. Water was fed to the anolyte compartment for 16 hours. The water feed was stopped and membrane quality brine was allowed to flow through the diaphragm for 5 to 6 hours before current was applied. Durin~ 56 days of operation on membrane quality brine at 85C and 2.5 kA~'m2, the cell produced an average of 10.0% caustic at an average cell voltage of 3.10 volts. This corresponds to a k-factor of 0.32 Vm2~kA. The average caustic current efficiency was 95.2%, and energy consumption was 2182 kil~att hours per metric ton of caustic. The anolyte head was steady at about 23 centimeters. Repeats of this example gave essentially identical results.

"~:

.

.. ; . . .

., . -. ~

j, :: , :
, WO 90tl3593 ~35~ PCl/USgO/0234 r~ ~ ~ ri~

Comparison of Cell Performance Between Conventional Diaphragms and the Present Invention Diaphragm production of chlorine and caustic is typ~cally carried out using a slurry-deposited diaphragm made of asbestos fiber or resin bonded ssbestos fiber ~modified asbestos). This type of asbestos diaphragm has been used in the industry for a number of years. Cell performance of a typical modified asbestos diaphragmis given by Donald L. Caldwell on page 1~0 of Chapter 2 of "Comprehensive Treatise of Electrochemistry, Volume 2: Electrochemic~l Processing", edited by Bockris, Conway, Yeager and White, Plenum Press, New York and London, 1981.
Summflrizing Caldwell's data at current densities at or around 2 kA/m2, current efficiencies ranged from 95.496 to 97.8% and k f.sctors ranged from 0.48 to 0.63Vm2/kA for different commercial cells. For operation at 2.5 kA/m~ as in our examples, the cell volt~ges could be expected to be between about 3.50 and 3.87 volts, resulting in power consumption values rsnging from 2440 to 2700 kilowatt hours per metric ton of caustic. These data were taken from cells using a version of resin bonded asbestos diaphragms.
The diaphragm of the present invention has exhibited better cell per-formance than modified asbestos diaphragms. Conservative estimates of cell performance of this diaphragm give about 95% current efficiency, a k factor of 0.32 to 0.34 Vm2/kA and a cell volt~go of 3.13volts at a current density of 2.5 kA/m2, all equating to a power consumption of 2208kilowatt hours per metric ton ~ ;............ .. ~ , , .

;: . :

.

WO 90/13~93 -36- ~ 7 Pcr/US90/02349 ~ . : ., , ., ~

of NaOH, an improvement in power consumption of between 9% and 1896 from the range ~uoted above.
~ ecause the asbestos is a fiber made of Mg(OH)2 and SiO2, it is partiallysoluble in brine depending on pH and is considered a "living" diaphragm.
Caldwell discusses this "living" feature of asbestos diaphragms in the following excerpt:
"Chrysotile asbestos is not chemically stable in chlorine cell electrolytes Mg(OH)2 is soluble in acid solutions and stable in bssic solutions; the reverse is true for SiO2, When a cell is energized, a chrysotile diaphragm will become SiO2 enriched on its acidic, anolyte face, and Mg(OH)2 enriched on its alkaline, catholyte face. The flow through the diaphragm will flush Mg2+
ions from the anolyte side toward the catholyte, where they will reprecipitate ns Mg(OH)2. This precipitate will constrict the flow channels, decreasing flow rate and efficiency and increasing voltage drop and caustic strength. The diaphragm is said to "tighten", After a period in service, the diaphragm reaches a state of equiiibrium with its surroundings and its characteristics stabilize. However, any drastic change in operating conditions will cause the dissolution-reprecipitation process to begin anew."
It is important to note that with each fluctuation in operating conditions, some portion of the diaphragm is resolubilized and lost. This leads to degradation of the diaphragm and eventual loss of performance.
Even with constant operation, the diaphragm is slowly eroded or dissolved 25 so that a steady decrease in current efficiency can be expected.
The diaphragm of this invention is made of chemically stable materials and so is insensitive to pH changes, unlike the asbestos diaphragm. This pH stabllity allows the diaphragm of this invention to operate under conditions in which electrical upsets occur and in which the current fluctuates without significant loss 30 of performance. rhis enables a plant to take advantage of power price breaks at off-peak times (load shedding) and prevents permanent loss of performance due toelectrical outages. Further, this stability allows cell operators to perform chemical treatments to this diaphragm to regenerate performance. All of these treatments or changes would be detrimental to an asbestos disphragm's performance. ExamPle 35 2 demonstrates the stability of performance of the present invention. Over the , . .
:- . . ~ . .
., : : - -: - :
. ' . -- . . .

WO 90/13~93,~ 37_ P~/US90/02349 cour~e of 421 days of operation, which were plagued wlth numerous electrical upsets, the current efficicncy remflined relatively constant. Examples 8 and 11 ~Iso Illustrate the stability of performence snd resiliency of the present invention.
An asbestos dle~phragm, even under the best of opersting conditions, would have S shown a steady decline In current efficiency over thls same period of time.
Moreover, an asbestos diaphragm would not have survived the electric~l upsets experienced by these examples of the present invention.
For the l~yered ~TFE diaphragm of U.S. Patent 3,~44,477, the best example of the reference (Example 11) gave a power consumpt;on of 2770 kilowàtt hours per metric ton of NaOH, surprisingly 25% worse than the typical 2208 kilowatt hours per metric ton oî N~OH of the present invention, even though the present Invention used a 16% higher current density, which should Increase power consumptlon.
While the invention has been disclosed hereln in connection wlth certein 15 embodiments and detailed descriptions, it will be clear to one skilled in the ert th~t modiflcations or v~riations of such deteils can be made without deviating from the gist of this invention, and such modifications or variations are conjidered to be withln the scope of the claims hereinbelow, A. . `

~' .

Claims (44)

What is claimed is:

1. A multilayer, porous, composite, shaped article comprising multiple layers of porous, expanded polytetrafluoroethylene (EPTFE) bonded together, said composite, shaped article having at least a portion of its exterior surfaces and at least a portion of its interior pore surfaces coated with a perfluoro ion exchange polymer.

2. The composite article of claim 1 having substantially all of its exterior surfaces and substantially all of its interior pores coated with a perfluoro ion exchange resin.

3. The composite article of claim 1 containing a water soluble surfactant within its pores.

4. The composite article of claim 1 in the form of a sheet.

5. The composite article of claim 1 in the form of a tube.

6. The composite article of claim 1 in which the perfluoro ion exchange polymer has a ratio of tetrafluoroethylene to functional comonomer of 1.5:1 to 5.6:1.

7. The composite article of claim 1 in which the perfluoro ion exchange polymer is a perfluorosulfonic acid polymer of equivalent weight less than 1000.

8. The composite article of claim 1 in which the perfluoro ion exchange polymer is a perfluorocarboxylic acid polymer of equivalent weight less than 1000.

9. The composite article of claim 1 in which the perfluoro ion exchange polymer is a mixture of perfluorosulfonic acid polymer and perfluorocarboxylic acid polymer of equivalent weight less than 1000.

10. The composite article of claim 1 in which the perfluoro ion exchange polymer is a copolymer containing perfluorosulfonic acid and perfluorocarboxylic acid groups, with an equivalent weight less than 1000.

11. The composite article of claim 1 in which the percentage by weight of perfluoro ion exchange polymer in the composite exceeds 2%.

12. The composite sheet of claim 4 having a thickness exceeding 0.25 millimeters.

13. The composite sheet of claim 12 having a thickness between about 0.76 millimeters and about 5.0 millimeters.

14 The composite tube of claim 5 having a wall thickness exceeding 250 micrometers.

15. The composite article of claim 1 having a permeability to water containing 0.1% tetra ethyl ammonium perfluorooctane sulfonate between about 0.01 and about 3.0 reciprocal hours at 23°C under a 20 cm head of water.

16. The composite article of claim 1 having a specific gravity between 0.05 and about 1.1.

17. The composite article of claim 16 having a specific gravity between about 0.15 and about 0.7.

18. The composite article of claim 1 having an asymmetric fine structure, wherein at least two of said multiple layers have different microporous structures.

19. The composite article of claim 18 wherein said at least two layers have methanol bubble point values which differ by at least 10%.

20. The composite article of claim 18 wherein said at least two layers have specific gravities which differ by at least 5%.

21. The composite article of claim 18 wherein said at least two layers have specific gravities which differ by at least 10%.

22. The composite article of claim 20 in the form of a tube and in which said at least two layers are oriented such that the layer having higher specific gravity is inside the layer having lower specific gravity.

23. The composite article of claim 20 in the form of a tube and in which said at least two layers are oriented such that the layer having higher specific gravity is outside the layer having lower specific gravity.

24. In an electrolytic cell containing anode and cathode compartments separated by a diaphragm, an improved diaphragm comprising a multilayer, porous composite diaphagm of multiple layers of porous, expanded polytetrafluoroethylene bonded together, said composite diaphragm having at least a portion of its exterior surfaces and at least a portion of its interior pore surfaces coated with a perfluoro ion exchange resin.

25. The electrolytic cell of claim 24 wherein at least two layers of said diaphragm have specific gravities which differ by at least 5% and wherein the higher specific gravity layer is oriented toward the cathode.

26. The improvement of claim 24 wherein a plurality of said composite diaphragms are used to separate a plurality of cell compartments of an electrolytic cell.

27. The improvement of claim 24 wherein said composite diaphragm has substantially all of its exterior surfaces and substantially all of its interior pore surfaces coated with a perfluoro ion exchange polymer.

28. The improvement of claim 24 wherein said composite diaphragm contains a water soluble surfactant within its pores.

29. The improvement of claim 24 wherein said electrolytic cell is used for the production of halogen and alkali metal hydroxide from an aqueous alkali metal halide solution.

30. The improvement of claim 24 in which the perfluoro ion exchange polymer is a perfluorosulfonic acid polymer of equivalent weight less than 1000.

31. The improvement of claim 24 in which the perfluoro ion exchange polymer is a perfluorocarboxylic acid polymer of equivalent weight less than 1000.

32. The improvement of claim 24 in which the perfluoro ion exchange polymer is a mixture of perfluorosulfonic acid polymer and perfluorocarboxylic acid polymer of equivalent weight less than 1000.

33. The improvement of claim 24 in which the perfluoro ion exchange polymer is a copolymer containing perfluorosulfonic acid and perfluorocarboxylic acid groups, with an equivalent weight less than 1000.

34. The improvement of claim 24 having an asymmetric fine structure, wherein at least two of said multiple layers have different microporous structures.

35. The improvement of claim 34 wherein said at least two layers have specific gravities which differ by at least 5%.

36. The improvement of claim 34 wherein said at least two layers have specific gravities which differ by at least 10%.

37. The improvement of claim 35 wherein said diaphragm is oriented such that, in at least two layers, the layer of lower specific gravity is closer to the anode side of said cell and the layer of higher specific gravity is closer to the cathode side of the cell.

38. In a filter, an improved filter medium comprising a multilayer, porous, composite, shaped article comprising multiple layers of porous, expanded polytetrafluoroethylene bonded together, said composite having at least a portion of its exterior surfaces and at least a portion of its interior pore surfaces coated with a perfluoro ion exchange polymer.

39. The improvement of claim 38 wherein said shaped article is in the form of a sheet.

40. The improvement of claim 38 wherein said shaped article is in the form of a tube.

41. A process for making a multilayer, porous, composite, shaped article comprising multiple layers of porous, expanded polytetrafluoroethylene (EPTFE) bonded together, said composite, shaped article having at least a portion of its exterior surfaces and at least a portion of its interior pore surfaces coated with a perfluoro ion exchange polymer in which process the perfluoro ion exchange polymer liquid composition used to coat the EPTFE contains an organic liquid selected so that said liquid composition, when poured onto a sheet of horizontal non-porous PTFE, wets substantially the entire area of the non-porous PTFE, and beads do not form during drying of the coated non-porous PTFE.

42. The process of claim 41 wherein said organic liquid is soluble in water and boils at a higher temperature than water.

43. The process of claim 41 wherein the organic liquid is 1-methoxy-2-propanol.

44. The process of claim 41 wherein the coating step is carried out by adding the coating liquid to one side of an EPTFE article from which most of the air has been removed by evacuation.