close

Вход

Забыли?

вход по аккаунту

?

Asymmetric Membranes Preparation and Applications.

код для вставкиСкачать
[I91 C. Niisalei~tand B. Heyden, Biochem. Biophys. Res. Commun. 47. 282
(1972).
[20] H . K i i u t x l and K . P. Schu/cr, Nature New Biol. 231, 265 (1971).
[2l] M . J. Chamherliu, J. M d r a t h , and L. Wualell, Nature 22X, 227 (1970).
[22] E. K . F. Baut;, FEBS Lett. 36, 123 (1973).
1231 A . L. Lehniiiger: Bioenergetics. Benjamin, New York 1965, p. 210.
1241 J. P. Rkhardson. J. Mol. Biol. 49. 235 (1970).
[25] R. W Datis and R. W Hjwan, Cold Spring Harbor Symp. Quant.
Biol. 3.5, 269 ( 1 970).
[26] H. Bremer and D. firan. J. Mol. Biol. 3X, 163 (1968).
[27] H. D. Manor, D. Goodman, and G . Srmt, J. Mol. Biol. 39, 1 (1969).
1281 U . Maitra and J. Huwit;, Proc. Nat. Acad. Sci. USA 54, 815 (1965).
[29] D. A. Goldthwair, D. D. Anthony, and C . W Wu, LePetit Colloq. Biol.
Med. I. 10 (1970).
[30] E . N . Brorly and E. P. Geiduschek, Biochemistry 9, 1300 (1970).
1311 H . Bremrr. M . Kourud, and R. Bruner, J. Mol. Biol. 16, 104 (1966).
[32] E. P. Geiduschek, 7: Nokainoto, and S . B. Weiss. Proc. Nat. Acad. Sci.
USA 47, 1405(1961).
[33] H. Brrnier and M . W Kortrad. Proc. Nat. Acad. Sci. USA 51, 801
( 1964).
[34] M . Hayushi, Proc. Nat. Acad. Sci. USA 54, I736 ( I 965).
1351 M . Chumberlirt and P. Berg, J. Mol. Biol. 8, 297 (1964).
[36] !b L. Florentier and V. I . Iiwnoc, Nature 228. 519 (1970).
[37] G. S. Stent, Advan. Virus Res. 5. 95 (1958).
[38] P. A. Riley, Nature 228, 522 (1970).
[39] J. M . Saucier and J. C. Wang, Nature New Biol. 339, 167 (1972).
[40] M. J. Chamberlin, Annu. Rev. Biochem. 43,721 (1974).
[41] R. R. Burgess and A. A. Eaoers, Fed. Proc. 29, 1 164 ( I 970).
[42] R. Fnkuda, E Iwakura, and A. Ishihama, J. Mol. Biol. 83, 353 (1974).
1431 E Inakura, R. Fukuda, and A. Ishihama, J. Mol. Biol. 83, 369 (1974).
[44] R. G. Shorenstein and R. Losick, J. Biol. Chem. 248, 6170 (1973).
[45] W Epstein and J. Beckwith, Annu. Rev. Biochem. 37, 41 1 (1968).
[46] B. Aller and R. Solem, J. Mol. Biol. 85,475 (1974).
1471 R. Maurer, 7: Maniatis, and M . Ptashne, Nature 249, 221 (1974).
1481 7: Maniatis, M . Ptashne, B. G. Barrell, and J . Donrlson, Nature 250,
394 (1974).
[49] B. Allet, R. J . Roberts, R . F . Gesteland, and R. Solem, Nature 249,
217 (1974).
[SO] B. Mijller-Hill, Angew. Chem. 83, I95 (1971); Angew. Chem. internat.
Edit. 10, 160 (1971).
[51] S. Nakanishi, S. Adhya, M . Gattesman, and I . Pastan, Proc. Nat. Acad.
Sci. USA 70, 334 (1973).
[52] M . Ptashne in A. D. Hershey: The Bacteriophage Lambda. Cold Spring
Harbor Laboratory, 1971, p. 221.
[53] B. de Crombrugghe, B. Chen, W Anderson, P. Nisslry, M . Gottesman,
I . Pastan, and R. Perlman, Nature New Biol. 231, 139 (1971).
[54] P. Chadwick, !t Pirrotta, R . Steinberg, and M . Ptashne, Cold Spring
Harbor Symp. Quant. Biol. 35, 283 (1970).
[55] A. M . Wu, S . Ghash, and H . Echols, J. Mol. Biol. 67,423 (1972).
1561
[57]
[58]
1591
N . M . Muizels, Proc. Nat. Acad. Sci. USA 70, 3585 (1973).
W Gilbert and A. Maram, Proc. Nat. Acad. Sci. USA 70, 3581 (1973).
J. Greenblart and R. SchlelL Nature New Biol. 233, 166 (1971).
G. Wilcot, K . Clenrcmon, D. Santi, and E. Englesberg, Proc. Nat. Acad.
Sci. USA 68, 2145 (1971).
[60] M! Anderson, A. Schneider, M . Emmer, R. Perlman, and I . Pastan, J.
Biol. Chem. 246, 5959 (1970).
[61] A. D. Riggs, G. Reiness, and G. Zubay. Proc. Nat. Acad. Sci. USA
68, 1222 (1971).
[62] A . Stcwns, Proc. Nat. Acad. Sci. USA 69,603 (1972).
1631 H. Horcitz, Nature New Biol. 244, 137 (1973).
[64] P. Leibowitz, S . M . Weissmun, and C . M . Radding, J. Biol. Chem.
246, 5120 (1971).
1651 J. E. Dahlberg and F. R . Blattnur, Fed. Proc. 32, 664 Abs (1973).
[66] J. W Roberts. Nature 224. I168 (1969).
[67] J. P. Richardson, Cold Spring Harbor Symp. Quant. Biol. 35, 127
( 1970).
[68] C. Lo\r.ery-Goldhammer and J. P. Richardson, Proc. Nat. Acad. Sci.
USA 71. 2003 (1974).
1691 G . Galluppi, C. Lowrry-Goldhammer, and 1. P. Richardson, unpublished
observations.
[70] B. de Cromhruyghe, S . Adhya. M . Gottesman, and I . Pastan, Nature
New Biol. 241, 260 (1973).
[71] G. Attardi and F. Amaldi, Annu. Rev. Biochem. 39, 183 (1970).
[72] N . R. Pace, Bacteriol. Rev. 37, 562 (1973).
[73] N. Nikalaer, L. Silengo, and D. Schlessinger, Proc. Nat. Acad. Sci.
USA 70, 3361 (1973).
[74] N. Nikolart, D. Schlessingur, and D. K . Wellauer, J. Mol. Biol. 86,
74 I ( 1 974).
[75] A. Ghysen and J . F. Celis, Nature 249, 418 (1974).
[76] C. Gurhrie, J. G. Seidman, S . Altman, B. G . Burrell, J. D. Smith, and
W H . McClain, Nature New Biol. 246. 6 (1973).
[77] D. P. Nierlirh, H . Lamfrom, A. Sarabhai, and J . W Ahelson, Proc.
Nat. Acad. Sci. USA 70, 179 (1973).
1781 S. Alrman and H . D. Robertson, Mol. Cell. Biochem. 1, 83 (1973).
[79] P. Schedl and P. Primakoff, Proc. Nat. Acad. Sci. USA 70, 2091 (1973).
[SO] R. A. Weinburg, Annu. Rev. Biochem. 42, 329 (1973).
[Sl] J. J. Dunn and F . W Studier, Proc. Nat. Acad. Sci. USA 70, 1559
( 1973).
[82] J. J. Dirnn and F . W Studier, Proc. Nat. Acad. Sci. USA 3296 (1973).
[83] D. Sol!, Science 173, 293 (1971).
[84] U . 2. Littauer and H . Inouw, Annu. Rev. Biochem. 42, 439 (1973).
[ 8 5 ] R. P. Perr!. and D. E . K e l l e j , Cell I , 37 (1974).
[86] C. E . Singer, G . R. Smith, R. Cortese, and B. N . Ames, Nature New
Biol. 238, 72 ( 1972).
1871 K. P. Schaeffur, S. Alteran, and D. Soll, Proc. Nat. Acad. Sci. USA
70, 3626 (1973).
1881 M . P. Deutscher, Progr. Nucl. Acid Res. Mol. Biol. 13, 51 (1973).
1891 S. Altrnan and J . D. Smith, Nature New Biol. 233, 35 (1971).
Asymmetric Membranes : Preparation and Applications
By H.-D. Saier and H.
Strathmann[*l
The separation of molecular mixtures by semipermeable membranes under the driving force
of hydrostatic pressure has assumed major importance in recent years. Among other factors
the development of membranes with asymmetric structure which, while having the same separating properties, afford a filtration output many times greater than that of the previously known
symmetrical membranes, was decisive for this method. In the present progress report the
structures of asymmetric membranes are discussed, as well as their preparation from various
polymers and their application to the separation of molecular mixtures.
1. Introduction
Selective mass transport through membranes has been the
subject of numerous scientific investigations for more than
a century. In this work attention was first focused on biologi-
[*I
Dr. H.-D. Saier [ +] and Dr. H. Strathmann
Forschungsinstitut Berghof GmbH
74 Tubingen-Lustnau, Berghof (Germany)
[+] To whom correspondence should be addressed.
452
cal membranes, the importance of which for metabolic processes in living organisms was recognized very early, but more
recently synthetic membranes in particular and their application to separation problems in chemistry and industry have
acquired economic significance. For example, membrane filtration is nowadays used to extract drinking water from the sea,
for fractionating, concentrating, and purifying macromolecular solutions, for the treatment of industrial waste waters,
and for the recovery of valuable constituents. Membrane filtraAngeir. Chem. internat. Edit. I Vol. 14 (197.5)
No. 7
tion has proved particularly useful in the treatment of thermally sensitive substances in biochemistry, pharmacology, and
the food industry[’-”! Here and in many other problems
of separation it is often greatly superior to the more conventional distillation, crystallizaition, extraction, and other processes, for it can often be carried out faster, more economically,
and under milder conditions.
In view of the wealth of possible applications of membrane
filtration and their economy it is surprising that its industrial
exploitation became possible only a few years ago, even though
the principle of the process and its industrial potential have
been known for almost a century from Pfeffer’s investigations’”’ and uun’t Hojf’s theoretical studied’ ‘ I on osmosis.
However, the industrial application of the method was held
back not least by the difficulty of developing membranes with
sufficient filtration efficiency. Only the studies of Loeb and
Sourirajan“ ”1, which led to the integral-asymmetric membranes, have made industrial scale membrane filtration feasible,
and this stimulus was followed by some very hectic development, particularly in the desalination of seawater and brackish
waters. Recent progress in polymer chemistry considerably
extended the scope of membrane filtration.
Asymmetric membranes consist, in principle, of a relatively
thick, highly porous carrier and the actual membrane, the
latter being only 0.1 to 0.5pm thick. They are characterized
primarily by a greatly increased rate of filtration.
3. Phenomenological Description of Membranes and
Membrane Transport Processes
Since the membrane itself is decisive for the success of
membrane filtration, this section will be devoted to a short
discussion of membrane construction and of mass transport
through membranes.
Quite generally, a membrane is defined as a liquid or solid
intermediate phase separating two homogeneous phases and
offering different resistances to the passage of different chemical
substances. The mass transport through this intermediate
phase can be described by the following phenomenological
relationship[’ 81:
“
J,
=
k= 1
LikXk
i = 1 , 2 , 3 ,... n
(1)
~
2. The Principle of Membrane Filtration
Membrane filtration is based fundamentally on a simple
principle’‘’! A molecular mixture is brought by convection
onto the surfaceofa membrane. While some of the components
pass through the membrane under the influence of hydrostatic
pressure, others are retained more or less completely. Viewed
superficially, membrane filtration differs from conventional
filtration only by the size of the particles to be separated.
The real difference, however, is that a conventional filter acts
asa sieve and separates a mixture of substances simply according to particle sizes, whereas in membrane filtration specific
interactions between the mixture components and the membrane matrix are, or can also be, responsible for the separation.
It is thus possible not only to carry out separations in the
molecular range but also to separate substances having the
same or nearly the same molecular dimensions.
Terminologically, a distinction is made in membrane filtration between reverse osmosis and ~ltrafiltration[~.
l7].
The
process is called reverse osmosis when the osmotic pressure
difference between the starting solution and the filtrate is
not negligible compared with the hydrostatic pressure applied;
this is the case when low-molecular compounds are to be
separated from a solution, e. g . in the desalination of seawater.
When, however, the aim is to separate macromolecular substances with molecular weights of about 10OO0, the osmotic
pressure difference between the starting solution and the
filtrate is negligibly small compared with the hydrostatic
pressure applied, and the process is one of ultrafiltration.
Although at first sight this differentiation may appear to
be very arbitrary, it has a degree of justification in respect
of the membranes used and the hydrostatic pressure required,
and there are also genuine differences between the techniques
used for reverse osmosis and ultrafiltration (see Section 7).
where J i is the material flow of the component i, Lik is a
phenomenological coefficient, and X k is a generalized driving
force provided by the chemical and electrochemical potential
gradients of the component k in the membrane. The applicability of eq. (1) is restricted by boundary conditions set by
the limited validity of linear laws in thermodynamics of irreversible processes ia discontinuous system^[^^-^^^. Although eq.
( 1 ) describes the transport process in membranes in such
a way that it includes all the interactions between the particle
flows as well as chemical reactions, its value for the description
of membrane filtration processes is limited because the phenomenological coefficients are mathematically difficult to
obtain and because the equation is strictly valid only in a state
close to equilibrium.
The generalized definition of a membrane is also of little
value when membranes are to be developed for special purposes, because this phenomenological description, like eq.
(I), does not take into account the membrane structure or
the molecular interactions between the transported particles
and the membrane matrix. For the development of membranes
in practice it is therefore often more advantageous to work
with membrane models.
4. Membrane Models
There are two models for the description of filtration membranes, each representing an idealized limiting case: the “pore
membrane” and “the solubility membrane”.
The ideal pore membrane comes nearer in its structure
to a conventional filter. Here mass transport takes place
through well-defined pores and the selectivity of the membrane
is determined by the pore diameter. Pore membranes can
be used only for the separation of substances that differ in
molecular dimensions.
In contrast, the ideal solubility membrane is a homogeneous
polymer layer through which all the components are transported independently of one another by molecular diffusion.
The selectivity of the solubility membrane depends on differences between the diffusion coefficients and concentrations
of individual components in the membrane matrix.
4.1. Model of a Pore Membrane
The mass transport during filtration through an ideal pore
membrane depends on volume flow through the pores and
453
can be described by the Hagen-Poiseuille equation. For a
system consisting of a solvent and a dissolved component
the volume flow density J , is given by the relation[231:
J,=L,(A~-cJAR)
(2)
Here L, is the hydrodynamic permeability, A p the specified
hydrostatic pressure, A n the difference between the osmotic
pressures in the original solution and the filtrate, and CJ a
so-called reflection coefficient[241that is a measure of the
separating characteristics of the membrane. For a strictly
semipermeable membrane C J =1 L 2 1 1 .
The hydrostatic permeability L , is given by:
Here (0 is the porosity, r the mean pore radius, q the viscosity,
i. a correction factor for pore length, and A x the thickness
of the membrane. For dilute solutions and membranes with
good selectivity the flow density JI of the solvent is equal,
in the first approximation, to the volume flow density J,.
For the dissolved component the flow density J , is given
by the following
J,
=
J,c?
-
D?(dC'/d.x)
v,
For description of the filtration flow density of the dissolved
component J , in a dilute solution the pressure term R A p
can become negligible in comparison with the concentration
term RTAc,/c,L2h1;
we then have to a first approximation:
(4)
Here cF and D? are, respectively, the concentration and the
diffusion coefficient of the dissolved component in the membrane, and dcy/d.x is the concentration gradient of the dissolved component in the membrane. Eq. (4) consists of two
additive terms: the first describes convective transport of the
dissolved component with the volume flow, and the second
diffusion of this component in the volume flow. The second
term is negligible in membranes with high filtration efficiency[22!
The two partial flows of solvent and solute yield a further
parameter that can in general be used to characterize a membrane. This is the membrane's retention capacity R for the
dissolved component 5l :
Putting J,= J , d we obtain:
Here cz and c{ are the concentrations of the dissolved component in the original solution (0) and the filtrate (f). For
a strictly semipermeable membrane J , and thus cfJ,=O and
R = 1, whereas for a membrane with no separating capacitv
R =O.
4.2. Model of a Solubility Membrane
In an ideal solubility membrane the transport of the components depends on molecular diffusion in which coupling
of the material flows is neglected. The flow density of any
one component through the membrane can be described by
the following equationI2.
454
Here, J , is the mass flow of a component i through the membrane, DY the diffusion coefficient in the membrane, k , the
distribution coefficient between the membrane and the adjoining solution, c, the concentration in the adjoining solution,
and A c , the concentration difference of component i between
the two sides of the membrane.
is the partial molar volume
of component i. A p I S a specified pressure difference, R is
the gas constant, T is the absolute temperature, and Ax is
the membrane thickness.
If the filtration flow density is to be derived for a relatively
dilute solution, the concentration term RTAc,/c, in eq. ( 7 )
can be expressed as P A R . The term An is the osmotic pressure
difference between the original solution and the filtrate. The
filtration flow density for the solvent J , is given by eq. (8):
5. Production of Asymmetric Membranes
Equations (2)-(4), and (8), (9) illustrate the difference
between a pore membrane and a solubility membrane. While
a pore membrane merely differentiates between the particle
geometries, i. e. acts only as a molecular sieve. the selectivity
of a solubility membrane depends mainly on the distribution
coefficients of the components between the membrane and
the external phases. Because of this, a pore membrane can
separate only materials that differ considerably in molecular
size. while a solubility membrane can also separate materials
having the same-or nearly the same-molecular size if their
solubilities in the membrane phase are sufficiently different.
A further difference between solubility and pore membranes
lies in their different filtration flow densities. Owing to the
diffusion process the filtration flow density of a solubility
membrane is 1-2 orders of magnitude lower than that of
a pore membrane under the same driving force. In both cases,
however, as follows from the transport equations, the filtration
flow density is inversely proportional to the membrane
thickness.
Nevertheless, for economic reasons the filtration flow density
should be as large as possible, i. e. the membrane thickness
should be minimized. In the present state of the art it is
not possible to produce perfect, self-supporting films thinner
than about 10 pm on a large scale, whereas to achieve economically justifiable filtration flow densities the membranes must
not beappreciably more than 0.1 to0.5 pm thick. These difficultiesare overcome by the use of asymmetric membranes. Figure
1 shows side-by-side scanning electron micrographs of an
asymmetric and a symmetrical membrane. The total thickness
of the asymmetric membrane amounts to 0.1 to 0.5mm. It
is constructed of a relatively thick, very porous support carrying an extremely thin skin measuring 0.1 to 0.5pm. This
skin is the actual semipermeable membrane, while the coarse,
highly porous support has no selective properties and does
5.2 Production of Composite Asymmetric Membranes
Fig. 3. Scanning electron micrographs of sections through asymmetric
membranes with a foam structure (left) and finger structure (right).
allowing high filtration flow densities even at low hydrostatic
pressures; these have no ability to retain salts. At pressures
above 10 bar the structure collapses and the membranes
lose their good filtration properties; such structures are therefore usually reserved for ultrafiltration, i. e. for the separation
of macromolecular substances at low hydrostatic pressures.
5%
15 %
10%
18 %
Comparatively recently, a “composite” method of construction was devised for the preparation of asymmetric memb r a n e ~ [ ~ ’ -that
~ ~ had
] extremely high filtration flow densities
and could be used especiallyfor reverse osmosis. For composite
membranes a highly porous, readily permeable, and mechanically very strong support is prepared, on which is placed
a homogeneous separating layer 200 to ca. IOOOA thick. For
this purpose a very dilute polymer solution is poured-once
or several times-onto the support layer and the solvent
is evaporated. To prevent the polymer solution from penetrating the support, the latter is usually protected by an inert
intermediate layer which is removed in subsequent washing
processes.
Present-day composite membranes consist of two materials,
since the support must not be attacked by the solvent of
the separating layer added. Cellulose acetate, cellulose nitrate
mixed polymers, and polysulfones are the main materials used
for the highly porous carriers. The pore diameter of the carrier
material must be smaller than the thickness of the active
film, in order to prevent rupture of the separating layer. Any
polymer can be considered for the homogeneous separating
layer, provided that the solvent used does not also attack
or dissolve the material of the support. The range of materials
that can be used is limited by the conditions that can be
withstood by both, but this disadvantage is compensated by
the very high filtration flow densities.
Other methods of preparing asymmetric membranes are
still in the experimental stage, for example
of
polymers on a liquid foundation which is then applied directly
to the porous supporting structure, or coating the porous
supports in a plasma stream of monomers which polymerize
on the s ~ p p o r t [ ~ ~ - ~ ~ l .
Very good filtration results for special applications have
561,
been achieved with dynamically formed
which are not true composite membranes although they behave
as such. Almost any ultrafiltration membrane can be used
for dynamically formed membranes if its pore diameter is
not more than about 0.2pm. During the filtration a very
small amount of macromolecular or colloidal substance is
added to the initial solution, this additive being retained by
the pores of the support and thereby forming a film that
can separate even salt from water.
6. Choice of Polymers Suitable for Membrane Production
20 %
22 %
Fig. 4. Scanning electron micrographs of membrane section in dependence
on polymer content of the initial solution (figures in wt-%).
These two membrane types are the limiting cases; in practice
a continuous transition from one to the other can be achieved
by variation of the preparation parameters, and thereby a
continuous transition from dense desalting membranes to
coarse-pored ultrafiltration membranes (Fig. 4).
456
Membrane materials are required to have high mechanical,
chemical, and thermal stability as well as resistance to microbiological degradation. While in pore membranes the material
responsible for the separation plays a subordinate role, in
solubility membranes it exerts a decisive influence on the
separating characteristics. Solubility membranes are always
used when molecules of approximately equal size are to be
separated, the basis for this being special interactions between
the molecules and the membrane matrix.
Transport ofwater and salt has been the subject of numerous
studies,particularly through cellulose acetate but also through
other polymers[**2 8 * 571. Water transport is associated with
the number of polar groups present in the polymer matrix.
Angew. Chem. internat. Edit.
1 Vol. 14 (1975) 1 No. 7
These polar groups serve as adsorption centers with which
the penetrating water can interact by hydrogen bonding.
According to this view the water molecules are transported
through the homogeneous matrix by activated boundary layer
diffusion along the polar groups. A prerequisite for this to
happen is that the polar groups should be distributed in
the polymer matrix as close together and as regularly as
possible['" "I.
t
L
AFig. 5. Relation between the salt flow IS), water flow (WI. salt retention
capacity (R), and water absorption IA). schematic.
Exclusion of the salt depends on the inability of its unhydrated ions to form hydrogen bonds to any large extent with
the polar groups. Hydrated alkali metal and halide ions are,
however, too large to penetrate the polymer matrix. Polymers
with desalination properties can usually absorb up to about
20 wt-'%, of water; at high water uptakes water clusters are
increasingly formed in the homogeneous polymer, and the
hydrated ions of the salt can penetrate, so that the retention
capacity is affected. The distribution of water within the
polymer matrix can be determined with the aid of a function
devised by Zimm and Lundberg[601.
Clusters can also be present
at very small water uptakes-much below 20 wt-'%,-arising
from theall-powerful water-water interactions; such a polymer
would be unsuitable for desalination purposes. Figure 5 shows
the relation between the salt flow, water flow, salt-retention
capacity, and water abs~rption["~.
7. Problems of Membrane Filtration in Industrial Practice
7.1. Concentration Polarization
surface and, owing to the concentration gradient, diffuse back
into the original solution: after some time a stationary state
is produced, with a constant concentration profile: convective
transport of the dissolved particles to the membrane surface
is compensated by a diffusion current directed back into the
original solution.
Concentration polarization considerably impairs the economics of membrane filtration. With substances of low molecular weight, e.y. salts, the osmotic pressure of the original solution, which must be overcome by the hydrostatic pressure,
increases with increasing concentration at the membrane surface. Moreover, since in general filtration membranes are not
strictly semipermeable, and the amount of material passing
through the membrane is also directly proportional to the
concentration on the membrane surface, the quality of the
filtrate is impaired by the concentration polarization. The
separation capacity of the membrane is thus reduced.
During filtration of macromolecular substances it is very
often found that, as a result of concentration polarization,
the solubility limit is exceeded and a covering film is formed
on the membrane surface. Such films act as secondary membranes, which not only drastically reduce the filtration flow
densities but can also completely alter the separation characteristics of the original membrane.
Although concentration polarization cannot be wholly
avoided, it can be kept in check by suitably arranged flow
parallel to the membrane surface, as has been shown by
penetrating investigations of transport processes at the laminar
boundary of membrane surfaces.
7.2. Development of Membrane Filtration Systems
In large industrial filtration systems the disadvantageous
effects of concentration polarization must be kept as small as
possible by suitable arrangements for the flow of the original
solution and lay-out of the installation. In addition, the filtration unit should be distinguished by a low cost and long
service life, and should not occupy much space. Three systems
have passed into industrial use (Sections 7.2.1 to 7.2.3).
7.2.1. The Tube Bundle System
The tube bundle system is represented schematically in
Figure 6rh7].
The tube-shaped asymmetric polymer membranes
are placed in porous tubes made of metal or plastic. T o
maintain the concentration increase at a minimum level, the
flow through the tubes must be turbulent, and the relatively
large tube diameter (ca.2.5 cm) therefore requires a high pump
output. The ratio of the installed membrane surface to the
volume of the apparatus, i. e. the packing density, is low;
the relatively large costs of sealing the ends, support of the
Even when an optimally suitable membrane is available
for a given problem of material separation, its practical industrial application may run into difficulties seriously detracting
from economics of the process. One of the most important
problems in technical processes is concentration polarization[hl -661 . Th'IS appears because, on membrane filtration,
a solution is brought to the surface of a semipermeable membrane by convecrion. While the solvent permeates the membrane under the driving force of hydrostatic pressure, the
dissolved components are more or less completely retained;
they concentrate in the boundary layer on the membrane
Fig. 6. Construction of the tube bundle module. schematic
457
membranes, and changing of the membranes are also unfavorable. However, the possibility of cleaning the membrane surfaces mechanically with small foam balls constitutes an advantage of this system.
7.2.2. The Roga Module
The Roga module[681consists of two asymmetric membranes
separated by an incompressiblebut porous intermediate layer.
Flow is introduced through a lattice on each side of the
membrane. Figure 7 gives a schematic representation of the
construction. The whole arrangement is rolled up, made watertight, and placed in a pressure tube. This mode of construction
gives a filtration module with favorable cost and compactness.
However, here too problems arise with controlling the concentration polarization, particularly when colloidal or macromolecular material has to be separated.
Fig. 8. Schematic representation of filtration through hollow fibers with
an inner active layer.
\
\
direction o t t l o w
All hollow-fiber filtration systems give not only purified
filtrates but also highly concentrated solutions. The latter
can be further evaporated and finally dumped or, simply
burnt, if their content of organic matter is high.
8. Present and Future Applications of Asymmetric
Membranes
Fig. 7. Construction of the Roga module, schematic.
7.2.3. The Hollow Fiber System
The membranes developed by D u P ~ n t [701~ ~are
, in the
form of hollow fibers of asymmetric structure; the actual
separating layer is on the outside of the fibers. The fibers
are bundled together, bent into a U-shape, inserted in a pressure tube, and sealed up. The packing density of such a module
is extremely high. The fibers are selfsupporting, so that for
work at pressures up to about 60 bar only a pressure-tight
outer tube of steel, aluminum, or fiberglass-reinforced plastic
is required. Filtration units can then be constructed from
hollow fiber modules distinguished by very small bulk and
favorable cost characteristics, which have proved valuable
above all for desalination. They appear to be unsuitable for
the separation of macromolecular substances because precipitates are formed on the membrane. The precipitates formed
on the outside of the hollow fibers can only be partially
removed by the insufficientflow running over the fiber surfaces,
so that a very large part of the membrane surface becomes
unavailable for filtration.
have developed
The A m i ~ o n [and
~ ~ Berghof
]
a different hollow-fiber concept, in which the active separating
layer is in the interior of the hollow fibers. This filtration
from the inside to the outside, even with low flow rates in
the laminar flow region, can reduce the concentration polarization to economically acceptable levels. Figure 8 is a schematic
representation of the filtration process. This new hollow-fiber
concept combines the advantages of the DuPont module and
the tube bundle module. So far it has only been used for
ultrafiltration, since the resistance of the fibers to pressure
does not yet satisfy the demands of reverse osmosis.
458
The main applications of asymmetric membranes are in
the field of desalination of sea water, brackish waters, and
river waters. Desalination plants have been built especially
in the USA, in Israel, and recently in the Federal Republic
of Germany. The water obtained from these plants is sterile,
desalted, and completely free from particles.
The recent progress in polymer chemistry now makes it
possible to purify the sometimes extremely corrosive industrial
waste waters. Here membrane filtration has the edge over
other methods of separation in that its mode of operation
conserves energy. It has proved its value particularly in the
treatment of waste water from dairy plants, from dipping
tanks for electrophoretic coating in the automobile industry,
and for the separation of oil emulsions. In this way valuable
components such as proteins, varnishes, and oils can be recovered. Other fields of application are in the purification of
the waste waters from breweries and from factories dealing
with fish, starch, dyeing, and coffee. In the pharmaceutical
and clinical industry asymmetric membranes are used principally to provide sterile liquids free from microparticles. It
would appear advantageous also to use them as artificial
kidneys; it might be possible to replace the slow and relatively
unspecificdialysis by membrane filtration of blood that would
be faster and give considerably sharper separation.
The industrial application of membrane processes is certainly still in its infancy[73*
74]. The development of asymmetric
membranes is being directed toward better selectivity and
faster rates of filtration; above all, membranes with still greater
chemical, mechanical, and thermal resistance are being sought.
In the development of membranes and membrane systems
Nature itself can be taken as a model. In Nature membranes
are of decisive importance for life, as they fulfill a variety
of functions in plant and animal metabolism. Natural membranes are extremely selective,have variable separating properAngew. Chem. internat. Edit.
Vol. 1 4 ( 1 9 7 5 ) 1 No. 7
ties. serve as information carriers, give extremely fast transport
rates, and take active part in mass transport. A very large
number of new uses will follow if membrane research can
incorporate just a few of these properties in synthetic membranes.
Received: December 2. 1974 [A 65 IE]
German version: Angew. Chem. 87.476 (1975)
Translated by Express Translation Service, London
[I]
[2]
[3]
[4]
[5]
[6]
171
[XI
191
[lo]
[I I]
[ 121
[ 131
[ 141
[I51
[I61
[ 171
[IX]
[I91
[20]
[?I]
[22]
12-11
1241
[2S]
1261
[27]
[ZX]
1291
R . E . Krstiity; Synthetic Polymeric Membranes. McGraw-Hill, New
York 1971.
L:. M e r r r i i . Desalination by Reverse Osmosis. MIT Press, Cambridge.
Mass. 1966.
S. B. 7iii1%idw.Diffusion and Membrane Technology. Reinhold. New
York 1962.
R . E . L ~ i w and
r
S. Lorh’ Industrial Processing with Membranes. Wiley,
New York 1972.
W:F . Bkrrr. B. G. H I I X J I SI . M . Rohiitsoii. and E. M. Zipilirciii. Nature
216. 51 I (1967).
R . W B d o . . J. Appl. Polym. Sct. 13. 369 (1969).
A . S. M i d i o e l s in E . S . PrrrI. Advances in Separation and Purification.
Wiley. New York 1968.
H . Str[irhrii~inn.H.-D. Sairr. W W i . s u d i . and U . Mullrr, Lebensmttteltechnik X. 3 ( 1973).
H . - D . Suio. and U. I-. My/itr.\, Deut. Milchwirt. I ! . 307 ( 1974).
W Pl;.//n.: Osmotische Untersuchungen. Leipzig 1x77.
J . w i ‘ r HofL Z . Phys. Chem. 1. 481 ( 1887).
S. LoPh and S . Soiwirrrjoii. J . Appl. Polym. Sci. I . 33 (1959).
S. Loch and S. S o i i r i r u j m : Advan. Chem. Ser. 3X. I17 (1962).
S Loeh and S. Soiirirujciit: Univ. of California. Los Angeles. Report
60 -~-6O
(1960).
S. Lorh and S. S o i i r i r u j m . US-Pat. 3 I33 132 and 3 133 137 (1964).
4 . G . Hortwrh. US-Pat. 1825631 (1931).
R . 4 . Cross and H . Str(ithiii(iiiii in B. L. K q l r r . L. S. Siij.i/c,r. and
C . Hort’oth: An Introduction t o Separation Science. Wfley. New York
1973.
S.R . dc G r o o r . Thermodynamics of Irreversible Processes. North Holland. Amsterdam 1959.
0. K ~ d n i iand 4. K ~ i t c h u l s k j ~Trans.
.
Faraday Soc. 59. 1918, 1931.
1941 (1963).
H . Podull, OSW-R. & D. Rep. 303 (196X).
R . S [ . / i / i i g / :Stomtransport durch Membranen. Steinkopf. Darmstadt
1964.
H . S r r ~ i r / i i i i m i i . R . W Boker. and F . R . E i r i c h 1. Phys. Chem. 76,
-.
73X (1972).
0. Kcdriii and K . S. S p i q / / r r . Desalination I . 31 I (1966).
4. J . Stcrwrriion, Trans. Faraday Soc. 4X. 176 (1952).
J . S. J O / ~ I I . S OL.I ~Drrsito..
.
and K . A. Krmi.\. in K . S. Spirgqln.: Principles
of Desalination. Academic Press. New York 1966
W E. Clrirk, Science I3X. 148 (1963).
S. Loch and S. M [ i i i j i L i u i i , Ind. Eng. Chem. Process Des. Develop.
4 . 207 ( 1965).
C. E . R r i d and E. J . B r c ~ o i i J. . Appl. Polym. Sci. 1. 133 (1956).
M . .4. F ~ ( J I W I I W
1. . F r i i r w . 0. Kcdriit. and R . B1oc.h. Desa~ination 7.
39.1 (1970).
[30] R . Mar:. Desalination 1 1 . 207 119721.
[31] R . Mor:. Hydronautics Incorporated R. & D. Report 7005-4 t o OSW.
April 1970.
[32] R . Mor:. Proc. 41st. Annual Meeting Israel Chem. Soc.. Oct. 1971.
[33] H . S r r ~ i r h i m i i t i i . H . - D . Soirr. and R . U! B d c r . 4th Internat. Symp. on
Fresh Water from the Sea. Heidelberg 1973.
[34] H . S t r o r h m t i r and If.-D. Siiier. Chem. I n g Tech. 46, 109 (1974). Mikrofiche MS 014 74.
[35] H . S r r ~ i r 1 i i i i ~ i i t i iP.
. Schrihlr. and R . W Biikrr. J. Appl. Polym. Sci.
15. 811 (1971).
1361 R . E . Kzatiit(/. J. Appl. Polym. Sci. 17. 1771 (1973).
.
C R . Lwrt.>. J. Appl. Polym. Sci.
[37] R . L. R i l q . H . K . L o i i . ~ [ / d rand
15, 1267 (1971).
r381 R . L. Rilur. H . K . Luiisdulr, C. R . L p i s . and L. LaGroiigr. OSWR. & D. Rep. 386 11969).
[39] M . Cohrii and B. Rig~g~lnit~iit.
OSW-R. & D. Rep. 400 (1969).
[40] L. 7: R ~ : d / e .J . E . C ~ i d o r t c .and B. R . NP/.SOII.
OSW-R. & D. Rep.
725 (1971).
[41] L. Z Roze11r. Polym. Prepr. 12. 261 (1971).
[42] B. R . Ne1,soii. L. T R o x / / ~ , J . E . C d o t r r . and E. M. SCo/tc~rg/O~Jd.
Environmental Protection Agency Program No. l 2 O l O D. R. H. I ’
S. Government Printing Office. Washington. Nov. 1971.
[43] P. S. Frmc.i.\ and J . E . C d m r . OSW-R. & D. Rep. 247 11967).
[44] P. S. F r m ~ ~ rri s 111.. OSW-R. & D. Rep. 177 (1964).
1451 H . Ybstidii. US-Pat. 3 775 308 ( 1974).
Appl. Polym. Symp. 22. 241 (1973).
1461 H . K I S I I ~ K
[47] J . R . Ho//di~iit.
Science 179. 500 (1973).
[48] C . Cuiicpa. S. Mtiiiari. C. Rossi. and R . Myo. Desalination 13. 159
11973).
[49] G . k i i i i y and J . Jogiir-Clod:iii.ski. Desalination 13. 53 (1973).
[SO] E. Ho//2r and 0.Kc&iit. Desalination 2. 25 ( 1967).
[51] .4. E . M o r c i i i L o i v . ~ k I .K . .A. Krmrs. H . 0.Phillip.\. J . S. J ~ ~ h i i . \ ~and
it.
A . J . Shor. J. Amer. Chem. Soc. XX. 5744 (1966): US-Pat. 3449245
(1969).
[52] J . R . F r w r . J . L. Htirchisoii. and R . Prrrrsori. Nature 726. 149 (1970).
1531 ./. J O ~ I I I W I 4.
I . S/IW and K , KI.LIII\.Desalination 2. 243 11967).
[54] C . E\.l”llld l’r [ I / . . US-Pdt. 2022187.
1551 C.. U: So//ort.sro// and H . F . S k i t r ~ ,OSW-R. & D. Rep. 730 ( 197 I).
[56] S. €3. S o ~ hand H . K . Loitshk. J . Appl. Polym. Sci. I S . 797 (19711.
[57] H . S r r d i i i i o i v t , OSW-R. & D. Rep. 726 11971).
[SX] T A . Joirtesoii and 7: H . M r l r x r . J . Appl. Polym. Sci. 13. 267 ( l Y 7 0 ) .
[59] A . ) ‘ l i . s ~ i d i i and C . E. L~iiiinx.OSW-R. & D. Rep. 281 11967).
[M)] B. H . Ziiiirit. and J . L. Liriidhrr(/. J . Phys. Chem. 60. 425 (1956).
1611 H . Sfr[ithiii[iitii and R . W BaLar. J. Appl. Polym Sci. 14. I197 11970).
[62] H . S t r t i r / i r i i m i t . Chem. Ing. Tech. 4 4 . I160 (1973).
[63] H . SrrLir/iitimii, Chem. Inp. Tech. 4.5. 825 (1973).
[64] H . Sttxirhimiitii and B. Kriliii. Desalination 6. 179 (1969).
[65] A. J . Shor, K . A . K r ~ i i i s .J . S. J o l i r i s o i i . and W Z Sriiit/i. Ind. Eng.
Chem. Fundam. 7. 44 (1968).
1661 L‘. Mn.trii. Ind. Eng. Chem. Fundam. 2. 229 11963).
[67] Water Waste Treat. Y. 147 (1964).
[68] OSW-R. & D. Rep. 313. 341 11968).
[69] Chem. Eng. 7X. 54(l971).
[70] H . 1. Mdi~~ii.
Nat. Acad. Sci. Publ. Y47. 345 (1963). Proc. Desalination
Res. Conf.
[71] W F. Bhrt. International Laboratory. p. 14. Nov.’Dec. 1972.
1721 German Pat. Appl. P 2318346.2. Firma Berghof (1974).
[73] L‘. F . Fr~inck,Lecture. Achema 1973.
[74] 4 . S . M i c h ~ i ~ 1 . Chem.
s.
Eng. Progr. 64. 31 11968).
L
-
459
Документ
Категория
Без категории
Просмотров
1
Размер файла
956 Кб
Теги
preparation, asymmetric, application, membranes
1/--страниц
Пожаловаться на содержимое документа