Cross Reference to Related Applications
The present provisional patent application is related to
U.S. Provisional Patent Application No. 60/725,238, filed on October 11, 2005
U.S. Provisional Application 60/802,457 filed on May 22, 2006
both of which are hereby incorporated by reference in their entirety.
Field of the Invention
The invention relates generally to the field of validation
testing. In specific embodiments the invention relates to integrity testing of porous
Background of the Invention
Porous materials play a significant role in a wide variety
of industrial applications including processing, e.g. filtering, packaging, containing,
and transporting manufactured goods and raw materials. The industrial settings in
which they are used include the pharmaceutical and biotechnology industries; the
oil and gas industries and the food processing and packaging industries, to name
but a few.
In several of these industries such as the pharmaceutical
and biotechnology industries and the food processing industry porous materials,
e.g. membranes, may be used as filtration devices to eliminate undesirable and potentially
harmful contaminants from marketable end products. Quality control and quality assurance
requires that these filtration devices comply with desired performance criteria.
Integrity testing provides a means for ensuring that a particular device meets its
desired performance criteria. Typically, in the case of membranes, integrity testing
ensures that the membrane is free of defects, e.g. breaches in the membrane exceeding
a desired size limitation, which would impair the membrane function and thus allow
the end product to become contaminated with harmful or undesirable material.
A variety of integrity tests suitable for ensuring the
performance criteria of membranes, e.g., filtration devices, have been previously
described. These include the particle challenge test, the liquid-liquid porometry
test, bubble point test, the air-water diffusion test and diffusion tests measuring
tracer components (see, e.g.,
U.S. Patent Nos. 6,983,505
Phillips and DiLeo, 1996, Biologicals 24:243
Knight and Badenhop, 1990, 8th Annual Membrane Planning Conference, Newton
Badenhop; Meltzer and Jorritz, 1998, Filtration in the Biopharmaceutical Industry,
Marcel Dekkar, Inc., New York, N.Y
.). A number of devices suitable for testing the integrity of a membrane
have also been described (see, e.g.,
U.S. Patent Nos.: 4,701,861
The previously described integrity tests have significant
shortcomings. The particle challenge test, for example, is destructive and thus
can only be performed once on a given specimen. Although it can be used for post-use
integrity testing, it is not suitable for pre-use validation, except for validating
the performance of a production lot. Lot validation, however, provides little assurance
regarding the integrity of individual membranes within a production lot. Moreover,
the test procedures and analysis can be difficult and complex. Flow based tests
such as the liquid -liquid porometry test and the bubble point test do not provide
a direct universal measurement of membrane retentive performance, but instead assess
performance based on a correlation between integrity testing data, e.g. gas or liquid
diffusion, and membrane retentive performance. Some flow based tests are also limited
in their sensitivity, e.g. size detection limit of membrane defects. Additionally
flow based tests are limited to single layer membrane devices, thus defects which
are present in only one layer of a multi-layered device will not be detectable using
A need therefore exists for an integrity test that is suitable
for any porous material, including, for example, both single layered and multi-layered
devices, e.g. devices comprised of membranes and which provides a non-correlative,
universal standard for assessing material performance. The test should be fast,
sensitive, non-destructive, inexpensive and easy to execute. It would also be useful
to be able to characterize a defect, e.g. by size or density, to determine if a
desired performance criteria of the porous material has been compromised as a result
of the defect or if the defect is inconsequential in terms of performance criteria.
A need also exists for a device and system which can implement such a test.
SUMMARY OF THE INVENTION
Certain embodiments of the invention provide a method,
e.g., a mixed gas test, for evaluating the integrity of a porous material that is
fast, sensitive, non-destructive, inexpensive and easy to execute, and also provides
a universal criteria for assessing the performance integrity of a porous material.
The porous material may comprise a single layered or multi-layered membrane device.
Universal criteria, as used herein, means that the test result provides a direct
measurement of performance criteria that is not dependent on correlation or extrapolation
of porous material properties. The resulting value obtained from the test is thus
independent of these properties. Thus in some embodiments the invention provides
a method of integrity testing of porous materials that is based on the concentration
of one or more gases in the permeate of a porous material. In certain embodiments
the test is a binary gas test, i.e. dependent on two gases, however more than 2
gases are also contemplated. The test may be independent of flow properties through
the porous material. Other embodiments of the invention provide a method for characterizing
a defect in a porous material, e.g. by size or density, to determine if a desired
performance criteria of the porous material has been compromised as a result of
the defect or if the defect is inconsequential in terms of performance criteria.
Still other embodiments provide a device and a system which can implement these
In one embodiment the invention provides a method of assessing
the integrity of a porous material comprising a) wetting the porous material with
a liquid; b) contacting a first surface of a porous material with a mixture comprising
two or more gases where at least one of the gases has a different permeability in
the liquid when compared to the other gases in the mixture; c) applying pressure
to the first surface of the porous material; d) assessing the concentration of at
least one of the gases in an area proximal to a second surface of the porous material.
The method may optionally further comprise e) comparing the assessed concentration
in d) with a predetermined concentration, wherein a difference in the assessed concentration
in d) and the predetermined concentration indicates the porous material is not integral.
The predetermined concentration may be, for example, the
concentration of gas calculated to diffuse through the integral, wetted porous material
at a given temperature and pressure. Integral, when referring herein to a porous
material, means non-defective. The given temperature and pressure may be the temperature
and pressure under which the test is conducted.
In another embodiment the invention provides a method of
assessing the integrity of a porous membrane comprising a) wetting the porous material
with water; b) contacting first surface of the membrane with CO2; c)
contacting the first surface of the membrane with a hexafluoroethane ; d) applying
pressure to the first surface of the porous material; e) assessing the concentration
of the hexafluoroethane in an area proximal to a second surface of the membrane;
and f) comparing the assessed concentration in e) with a predetermined concentration
of the hexafluoroethane, wherein an assessed concentration of hexafluoroethane exceeding
the predetermined concentration indicates the membrane is not integral.
In still another embodiment the invention provides a method
of assessing the integrity of a porous material comprising at least one defect,
wherein the method comprises a) wetting the porous material with a liquid; b) contacting
a first surface of a porous material with a mixture comprising two or more gases
where at least one of the gases has a different permeability in the liquid when
compared to the other gases in the mixture; c) applying pressure to the first surface
of the porous material; d) increasing the concentration of pressure applied in c)
over time; e) assessing the concentration of at least one of the gases in an area
proximal to a second surface of the porous material; g) calculating the defect density;
h) calculating the defect diameter; i) determining a defect size distribution; and
j) comparing the defect size distribution with a predetermined retention value for
the porous material, where a defect size distribution greater than the predetermined
retention value indicates that the porous material is not integral. The retention
value may be for example, the log retention value (LRV).
In yet another embodiment the invention provides a method
for finding at least one defect in a porous material comprising a) wetting the porous
material with a liquid; b) contacting a first surface of a porous material with
a mixture comprising two or more gases where at least one of the gases has a different
permeability in the liquid when compared to the other gases in the mixture; c) applying
pressure to the first surface of the porous material; d) assessing the concentration
of at least one of the gases in an area proximal to a second surface of the porous
material; and e) comparing the assessed concentration in d) with a predetermined
concentration, wherein a difference in the assessed concentration in d) and the
predetermined concentration indicates the porous material has at least one defect.
In a further embodiment the invention provides a method
of characterizing a defect in a porous material comprising a) wetting the porous
material with a liquid; b) contacting a first surface of a first layer of porous
material with a mixture comprising two or more gases where at least one of the gases
has a different permeability when compared to the other gases in the mixture; c)
applying pressure to the first surface of the porous material; d) increasing the
pressure applied in c) over time; e) assessing the concentration of at least one
of the gases in an area proximal to a second surface of the porous material; and
f) calculating the defect density in the porous material thereby characterizing
the defect in the porous material.
In another embodiment the invention provides a method of
characterizing a defect in a porous material comprising a) wetting the porous material
with a liquid; b) contacting a first surface of a first layer of porous material
with a mixture comprising two or more gases where at least one of the gases has
a different permeability when compared to the other gases in the mixture; c) applying
pressure to the first surface of the porous material; d) increasing the pressure
applied in c) over time; e) assessing the concentration of at least one of the gases
in an area proximal to a second surface of the porous material; and f) calculating
the diameter of the defect in the porous material thereby characterizing the defect
in the porous material.
In yet another embodiment the invention provides an apparatus
for assessing the integrity of a porous material comprising a) a gas source; b)
a gas feed pressure regulator; c) a porous material sample contained in a feed chamber;
and d) a permeate sampling port. The apparatus may optionally further comprise at
least one of the following: e) a feed sampling port; f) a permeate pressure measuring
device; g) a feed pressure measuring device h) a gas-liquid contactor for saturating
the feed gas; i) a purge valve on a the feed chamber; j) a permeate gas flow meter;
k) a device for measuring the purge gas flow rate; I) a device for measuring the
feed gas flow rate; m) and a thermometer for measuring the permeate gas stream temperature
and water temperature.
In still another embodiment the invention provides a system
for assessing the integrity of a porous material comprising a) a gas source; b)
a gas feed pressure regulator; c) a porous material sample contained in a feed chamber;
d) a first and second gas; e) a liquid and f) a device for measuring the concentration
of at least one gas. The system may optionally further comprise at least one of
the following: g) a feed sampling port; h) a permeate pressure measuring device;
i) a feed pressure measuring device j) a gas-liquid contactor for saturating the
feed gas; k) a purge valve on a the feed chamber; a permeate gas flow meter; I)
a device for measuring the purge gas flow rate; m) a device for measuring the feed
gas flow rate; and a thermometer for measuring the permeate gas stream temperature.
In further embodiments the invention provides a method
of assessing the integrity of a multi-layered device comprising more than one layer
of porous material, wherein each layer is comprised of a first and a second surface,
and wherein a sample applied to the device will flow from the first surface of the
porous material through the porous material to the second surface and where the
method comprises a) wetting the porous material with a liquid; b) contacting a first
surface of a first layer of porous material with a mixture comprising two or more
gases where at least one of the gases has a different permeability when compared
to the other gases in the mixture; d) applying pressure to the first surface of
the first layer of the porous material; e) assessing the concentration of at least
one of the gases in an area proximal to a second surface of a last layer of porous
material; and f) comparing the assessed concentration in e) with a predetermined
concentration, wherein a difference in the assessed concentration in e) and the
predetermined concentration indicates the porous material is not integral.
Additional objects and advantages of the invention will
be set forth in part in the description which follows, and in part will be obvious
from the description, or may be learned by practice of the invention. The objects
and advantages of the invention will be realized and attained by means of the elements
and combinations particularly pointed out in the appended claims.
It is to be understood that both the foregoing general
description and the following detailed description are exemplary and explanatory
only and are not restrictive of the invention, as claimed.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 is a graph showing the hexafluoroethane gas exit
concentration versus the flow ratio.
Figures 2A, 2B and 2C show examples of a multi-layered
membrane used for filtration and integrity testing.
Figure 3 is a schematic diagram of a device suitable for
performing integrity testing of a porous material.
Figure 4 is a graph showing defect flow rate versus pressure.
Figure 5 is a graph showing defect size distribution.
Figure 6 is a graph showing the results of an air water
Figure 7 is a graph showing flow ratio versus pressure.
Figure 8 is a graph showing the estimated defect size distribution
of 3 inch PES cartridges.
Figure 9 is a graph comparing binary gas composition using
different feed compositions versus defect size.
Figure 10 is a graph showing the log retention value (LRV)
versus permeate composition for a 10 inch single layer cartridge.
Figure 11 is a graph showing the defect size distribution
for V180 samples.
Figure 12 is a graph comparing hexafluoroethane permeate
concentration, as a function of time, for a single layered membrane device and a
multi-layered membrane device.
Figure 13 shows integral and defective single and multi-layered
membrane devices. Hatched areas represent pore defects.
Figure 14 is graph showing retention versus permeate concentration
of various single and multi-layered membrane devices which are either integral or
comprised of one or more defects.
Figure 15 a and b shows permeabilities of various gases.
DESCRIPTION OF THE EMBODIMENTS
Methods of the Invention
Certain embodiments of the invention provide a method,
e.g., a mixed gas method, such as the binary gas test, for assessing the integrity
of a porous material. The test relies on measuring the concentration of at least
one of the gases in the permeate of a porous material. The binary gas test uses
2 gases with differing permeabilities in a liquid used to wet a porous material.
Other embodiments provide for a test which uses a plurality of gases where at least
two of the gases have differing permeabilities in a liquid used to wet a porous
material. As an example, two of gases may be used for directly determining the integrity
of the porous material, while a third gas may be used as an internal standard. Other
embodiments of the invention provide a system and an apparatus to practice the methods
The conditions under which the methods of the invention
are practiced may be chosen by the skilled artisan. As an example, the methods of
the invention may be practiced at a temperature ranging from about 0°C to about
100°C. In one embodiment the invention is practiced at a temperature of about
20°C. In another embodiment the invention may be practiced at a temperature
of about 4°C. The methods of the invention may be practiced at a pressure,
e.g., a feed pressure, ranging from about 1 PSI to about 100 PSI. In another embodiment
the methods of the invention may be practiced at a pressure of about 30-50 PSI.
In one embodiment the methods of the invention may be practiced at a pressure of
about 50 PSI. In another embodiment the methods of the invention may be practiced
at a pressure of about 30 PSI. In yet another embodiment the methods of the invention
may be practiced at a pressure of about 15 PSI. In a further embodiment the invention
may be practiced at a pressure that is just below the bubble point of the porous
material. In still other embodiments the pressure may be ramped up, e.g., slowly
increased by small increments while measuring flow rate and concentration. In yet
other embodiments the pressure may be ramped down, e.g., slowly decreased by small
increments while measuring flow rate and concentration. The pressure can be ramped
up or down in stepwise increments. The stepwise increments can range from 0.5 psi
to 100 psi; or from 1 psi to 25 psi; or from 5 psi to 10 psi.
Where a plurality of gases is used the percentage of each
gas in the mixture may be chosen by the skilled artisan. As an example, where 2
gases are used the first gas may be used at a percentage volume ranging from about
0.001 % to about 99.999%, and the second gas may be present at a percentage volume
ranging from about 0.001 % to about 99.999%.
1. improvements Provided by Certain Embodiments of the Invention
In certain embodiments the invention provides a method
of assessing the integrity of a porous material comprising filling, e.g., saturating,
the porous material with a liquid and challenging the porous material by increasing
the pressure of a multicomponent gas feed while measuring the steady state gas composition
of the permeate. In other embodiments, e.g., where the device is a multi-layered
device, a steady state may not be reached, but a quasi steady state may be attained.
In a quasi steady state the concentration changes slowly with respect to time thus
permitting measurement of the gas composition concentration. Data points may be
obtained at a single pressure point once steady state is achieved or at multiple
pressure ramping points as described below. The skilled artisan will understand
that in practicing the invention test variables such as operating pressure, solvent,
gas species and gas composition may be varied to meet material requirements, sensitivity
limits, and operator convenience. It will also be understood by the skilled artisan
that functional properties, e.g. retention of a target species such as a virus,
may be correlated with integrity measurements made in accordance with the methods
In some embodiments of the invention a wetted porous material
is contacted with a first and second gas where the first and second gas each has
a different permeability in the liquid used to wet the porous material. Knowing
the composition of the first and second gas and the liquid, a skilled artisan can
readily predict the composition of the gas mixture that would permeate through the
wetted porous material assuming the material is integral. Pressure may then be applied
to the porous material such that potential defects in the porous material are no
longer occupied by liquid thus permitting a rapid influx of the gas mixture and
a change in the permeate gas mixture compared to the predicted value of the composition
of the gas mixture. In some embodiments the steady-state concentration of the less
permeable gas may be observed to be greater than its predicted value. Thus, the
methods of the invention are not dependent on the resident time a gas spends in
a porous material nor are the methods of the invention dependent on flow properties
of the porous material. In some embodiments the composition of the gas mixture found
in the permeate may be used to assess the integrity of the porous material. Consequently,
the test is insensitive to each of the following: small variations in operating
pressure, physical properties of the porous material; volumetric changes in the
membrane or test housing. In certain embodiments the invention provides a method
of assessing integrity of a porous material, e.g. a membrane that is independent
of specific characteristics of the membrane such as porosity, tortuosity, and thickness
of the wetting fluid. The invention thus provides, in some embodiments, a universal
a priori standard for assessing integrity of a porous material.
In certain embodiments the invention provides a method
of assessing the integrity of a porous material, e.g., a membrane, which is simple,
rapid, repeatable and non-destructive. The method may be performed before or after
the porous material is used for its intended purpose and may be repeated more than
once, if desired. The method allows the artisan to choose the combination of liquids,
gases and porous materials depending on individual need. Moreover, the method increases
the sensitivity of assessing integrity compared to previously described methods.
Thus certain embodiments of the invention provide a method of detecting defects
that are 1-100, 2-50, 10-50 times smaller than the defects detected by air-water
For many porous materials, e.g., membranes, filter devices,
it may be useful to assess the effect of defects on the retentive properties of
the material. When it is desirable to quantify defect size and/or defect population
density the method of the invention may be practiced in pressure ramping mode such
that a plurality of data points are measured as pressure is increased, including
for example permeate flow rate. Additional embodiments provide methods for quantifying
defect as a function of size. Thus certain embodiments of the invention provide
a method of identifying and quantifying defects in the range of 200 nm-2,000 nm,
200 nm-10,000 nm, 10 nm-10,000 nm. The skilled artisan will appreciate that detection
ranges will be influenced by filter area, choice of gases, wetting fluid, test pressure
and type of detector used.
2. The Universal Standard
As discussed above, one advantage of the methods of the
invention, e.g., the mixed gas test, over the air-water diffusion test is its invariance
to many test and material properties. As a result, it can provide a universal standard
for assessing membrane integrity. For a saturated material, the gas components diffuse
through the liquid-filled pores from the high-pressure feed side to the low-pressure
permeate side. The diffusive molar flux ni for component i is given by
Fick's law, which for a symmetric membrane is
where &egr; is the porosity; Di and Si are the diffusion
coefficient and solubility coefficient, respectively, for gas component i in the
liquid filling the material pores; yi is the mole fraction of gas component
i; subscripts f and p refer to the feed and permeate streams, respectively; &tgr;
is the tortuosity of the pores; and t is the thickness of the fluid layer through
which components must diffuse. Note that the thickness of the fluid layer is not
always equal to the thickness of the material. For example, a pleated membrane may
have a liquid meniscus between pleats creating a fluid layer thickness that is greater
than the membrane thickness.
Using the equation above for molar flux, it is straightforward
to show that the composition of the permeate gas (i.e. the ratio of the fluxes),
is independent of porosity, tortuosity, and thickness of the water layer for a defect-free
membrane. It is also apparent to one skilled in the art that the molar flow rate
(i.e. the product of molar flux and area) will depend on each of these variables.
Molar flux, which is the basis for the air-water diffusion test, shows variations
with time as a result of pleat deformation and movement, water drainage from membrane
pores, and other factors that influence molar flux, but are independent of the inherent
integrity of the filter. The air water diffusion test results will also vary with
changes in the membrane properties such as porosity and tortuosity. In contrast,
the methods described herein, e.g., the mixed gas test, provide results based on
the gas composition. The results therefore are invariant with respect to variables
for an integral membrane. Consequently one advantage of the methods of the invention
is that they provide a single-point measurement to assess porous material integrity
which is universal for all materials, and is invariant with fluid drainage issues.
This fact can greatly simplify the testing and certification of material integrity.
3. Quantifying Defect Size and Density Distribution
As discussed above, it may be desirable in certain situations
to be able to characterize a defect or defects in a porous material beyond merely
noting its presence or absence. Certain embodiments of the invention provide a method
of calculating defect diameter and distribution density, both of which may be useful
in assessing a material's integrity, particularly as it relates to retention.
Gas flow through a defect is due to convective rather than
diffusive transport. Several researchers have modeled gas flow in defects assuming
the Hagen-Poiseuille equation applies. However, one skilled in the art will recognize
that this equation is valid only at the limit of very low pressure differentials
across the membrane (
R. Prud'homme, T. Chapman, and J. Bowen, 1986, Applied Scientific Research,
.). At typical integrity test conditions, e.g., generally exceeding 20
psi (pounds per square inch), the flow through a defect more closely follows choke
flow, particularly if the defect diameter is large relative to the thickness of
the retentive zone within the membrane. In general, the transition from Hagen-Poiseuille
flow to turbulent flow to choke flow is a function of the ratio of the permeate
pressure to the feed pressure. The transition to choke flow, when frictional losses
are negligible, occurs when the critical pressure ratio is reached, which depends
on the parameter k, the ratios of specific heats, and is a property of the gas components:
For common gases used in integrity testing of porous materials,
the transition to choke flow occurs when the feed pressure exceeds about 15 psig
and the downstream pressure is ambient. Consequently defect flow discussed by some
previous authors describing integrity testing is likely in the choke-flow regime.
It is recognized in the art that defects in a liquid-filled
porous material will open when gas pressure forces exceed the capillary force holding
the liquid in the pores. The relationship between the defect diameter and the pressure
differential across the material is typically modeled via the Laplace equation:
where d is the defect diameter, &ggr; is the interfacial tension for the
gas and liquid filling the membrane pores, and &thgr; is the contact angle. As
a result, defects of different sizes can be opened by varying the operating pressure
of the test. This feature is useful since the retention of a species depends on
its size relative to the defect size. One can practice the method described herein
at a fixed pressure where the pressure is adequate to open all defects larger than
the retained species, thereby assessing the impact of defects on retention. Alternatively
the test can be run at multiple pressures, allowing defects of different sizes to
Assuming steady state, uniform upstream and downstream
fluid properties, ideal gas, and Henry's law, the composition of the gas at the
exit is a function of the operating variables and the flow rate of gas through defect(s)
in the membrane device. To simplify the formulation, it is convenient to define
the following ratios:
fr= flow ratio = gas flow rate through defects/total gas flow rate
- Pr = pressure ratio = permeate pressure/feed pressure
- &PHgr; = permeability of gas component i/permeability of gas component j
For a binary gas mixture, these test variables are related
by the following quadratic equation:
By measuring the inlet and outlet gas compositions, it
is possible to solve equation (5) to determine the flow ratio. By definition, a
membrane with a flow ratio of zero is integral. As noted above, the exit gas composition
for an integral membrane (i.e. fr = 0) is invariant with membrane properties,
and depends on the choice of operating pressures and gas composition. A skilled
artisan will recognize that the presence of a defect (i.e, fr greater than zero)
will cause the exit concentration to change from the value for an integral membrane.
Consequently the composition measurement alone is sufficient to determine if a membrane
To determine the defect density equation (5) is solved
for flow ratio. The Defect flow rate = fr*permeate flow rate. To determine the defect
density, it is necessary to use a model for the flow in the defects. As noted above,
defect flow can be described as choke flow in many instances for membranes. Differentiating
the equations for defect flow with respect to pressure, assuming choke flow, yields
defect density according to equation (6):
Where Nj/A is the number of pores of size j
per area that open as the feed pressure in incrementally increased from Pj
to Pj+&Dgr;P; R is the gas constant; T is the temperature; MW is the
molecular weight; and other symbols and subscripts are as previously described.
Equation (3) can then be used to calculate the defect diameter.
The pressure ratio (Pr) is also an important variable.
Figure 1 shows that the exit concentration varies more rapidly with the flow ratio
as the pressure ratio decreases. As a result, the test can detect smaller flow ratios
(assuming all defects are opened at the test pressure) as the pressure ratio decreases.
The test pressure ratio may be above the critical pressure for the gas, or may be
set so that the transmembrane pressure differential is just below the bubble point
of the membrane while the pressure ratio is above the critical pressure.
A porous material, such as a membrane, may contain "defects"
that do not impact its retention performance where size exclusion is the primary
separation mechanism for the membrane. There are several possible reasons why a
"defect" would not impact retention. As an example, the defect may be smaller than
the species to be retained. Thus the defect not does allow passage of the species.
As another example, the defect may be larger than the species to be retained, but
the population of defects is too small to impact integrity. Porous materials such
as membranes, including filters comprised of membranes, are often designed to remove
target species to a specified degree. One standard commonly applied to membranes
and filters comprised of membranes is the log removal value (LRV):
where C is the concentration of the target species to be retained by the membrane.
The defect may reduce the LRV, but still allow the LRV to be within the specified
range for the membrane. For example, a virus filter may have a viral clearance guarantee
of 4 LRV. Methods of the invention such as mixed gas testing may indicate the presence
of defects in the 200-400 nm range. However, if the intrinsic retention of the integral
filter is 5 LRV, the defects may only reduce the actual retention to 4.5 LRV, which
may still be acceptable. Another advantage of the methods described herein over
previously described integrity tests is the ability to quantify the defect concentration
as a function of size, so that the impact of a defect(s) on retention can be independently
assessed. This allows more discrimination among porous materials that have defects,
so that serviceable materials are not erroneously rejected by the integrity test.
The integrity of any porous material may be assessed using
the methods, devices and systems of the invention. As an example, but not as a limitation,
the porous material may take the form of a container, a bottle, a cap, a cylinder,
a tube, a hose, a cassette, a column, a chip, a bead, a plate, a sheet, or a monolith.
The porous material may be comprised of an organic or inorganic
molecules or a combination of organic and inorganic molecules. The porous material
may be comprised of a hydrophilic compound, a hydrophobic compound, an oleophobic
compound, an oleophilic compound or any combination thereof. The porous material
may be comprised of a polymer or a copolymer. The polymers may be crosslinked.
The porous material may be comprised of any suitable material,
including, but not limited to polyether sulfone, polyamide, e.g., nylon, cellulose,
polytetrafluoroethylene, polysulfone, polyester, polyvinylidene fluoride, polypropylene,
a fluorocarbon, e.g. poly (tetrafluoroethylene-co-perfluoro(alkyl vinyl ether)),
poly carbonate, polyethylene, glass fiber, polycarbonate, ceramic, and metals. The
porous material may be in the form of a single or multilayered membrane. The porous
material may be, for example, a hollow fiber, a tubular format, a flat plate, or
In certain embodiments the porous material may be a membrane,
e.g., a filter or filtration device comprising a membrane. The porous material may
be capable of excluding solutes based on one or more properties of the solutes,
e.g., the size of the solutes. As an example the pores of the material may be too
small to allow the passage of a particle of a specific size, e.g., diameter or a
particular molecular weight.
The membrane may be contained in a housing e.g., a cylinder,
a cassette. The membrane may be a single layered membrane or a multi-layered membrane.
The membrane may be a flat sheet, a multi-layered sheet, a pleated sheet or any
combination thereof. The membrane pore structure may be symmetric or asymmetric.
The membrane may be used for filtration of unwanted materials including contaminants
such as infectious organisms and viruses, as well as environmental toxins and pollutants.
In some embodiments, where the porous material is comprised of more than one layer,
an outlet or port may be provided to obtain samples from the interstitial space
Multi-layered Membrane Devices
The invention also provides methods, systems and apparatuses
for performing integrity testing of multi-layered devices. Multi-layered devices
include devices comprised of more than one layer of porous material, e.g. membranes,
which in some embodiments may be configured or contained within a housing or cartridge.
The multi-layered device may be comprised of 2, 3, 4, 5 or more layers of porous
material. The first layer of the multi-layered device may be the layer which is
first contacted by a sample entering the device. The last layer of the multi-layered
device may be the layer from which a sample exits the device.
Each layer of the porous material may be comprised of a
first and second surface. The first surface may be designated as the surface which
is first contacted by a sample entering the porous material and the second surface
may be designated as the surface from which the material exits the porous material.
In some embodiments the multi-layered device may be comprised of a spacer placed
between adjacent or stacked layers of porous materials and which may facilitate
integrity testing of the multi-layered device. The spacer may be for example a porous
In other embodiments the multi-layered device is not comprised
of a spacer between the multiple layers of porous material, e.g. membranes. In some
embodiments, the porous material may be stacked in layers such that the layers are
in close proximity to each neighboring layer. In some embodiments the stacked layers
may be contiguous with the neighboring layer. Air or gas pockets may spontaneously
form between the layers in certain embodiments. In other embodiments, e.g. where
at least one layer of the device comprises an asymmetric membrane, air or gas pockets
may form within at least one layer of the multilayer device. The air or gas pocket
may form in a membrane which is highly porous, such as a microfiltratrion membrane.
In certain embodiments multiple layers stacked contiguously with the neighboring
layer may advantageously serve to maintain the retentive capability of the device.
For example a breach or defect in one layer of a device where the material layers
are in close proximity may have minimal effect on retentive capability of the device.
In some embodiments the invention provides a method of
integrity testing separately each individual layer of a multi-layered device comprised
of porous material. The method may include performing the mixed gas test described
herein, e.g., the binary gas test. Multi-layered devices, e.g., comprising multiple
membranes, which allow integrity testing of individual layers, is described in a
co-pending patent application entitled "Integrity Testable Multi-layered Filter
Device" filed this day by Rautio et al. A brief description of integrity testing
of individual layers of multi-layered devices is provided below.
An example of a multi-layered device is shown in Figure
2a which demonstrates normal flow through the device. Fluid enters the inlet 6 into
opening 18 of the first layer and then into the core 20. Fluid then passes through
the filter element 14 leaving behind any contaminant that the filter is designed
to remove by such well-known processes as size exclusion, adsorption, philicity/phobicity
or charge repellation. Fluid exits the first element and enters the inner bore of
the housing 30. It then enters the second filter layer 14B passing through to the
core 20B out through the opening 18B and into the outlet 8 by which it leaves the
housing 4. It is understood that 50 represents an impermeable barrier. As with the
first layer, fluid passing through the filter element 14B leaves behind any contaminant
that the filter is designed to remove by such well-known processes as size exclusion,
adsorption, philicity/phobicity or charge repellation. The filter may be the same
as the first layer or if desired it may be different in size exclusion characteristics,
adsorptive capabilities and the like.
To integrity test the first layer, the set up of Figure
2B is used. Here the first filter layer 14 is wetted with a suitable liquid for
the gas or gases to be used. The outlet 8 is then closed as shown by cap 7B although
other means such as a valve (not shown) or the like may be used. The vent 10 is
opened and connected to a suitable detection device (not shown). One or more selected
gases are flowed through the inlet 6 at a predetermined pressure or series of pressures
and the change in flow or gas concentration may be measured by a detection device
that has been coupled to the vent 10.
To test the integrity of the second layer 14B, the set
up of Figure 2C is used. Here the second filter layer 14B is wetted with a suitable
liquid for the gas or gases to be used. The inlet 6 is then closed (as shown by
cap 7 although other means such as a valve (not shown) or the like may be used)
and the vent 10 is opened and connected to a suitable detection device (not shown).
One or more selected gases are flowed through the outlet 8 at a predetermined pressure
or series of pressures and the change in flow or gas concentration is measured by
a detection device that has been coupled to the vent 10.
The skilled artisan will appreciate that the device depicted
in Figure 2 may be adapted to provide a sweep gas by the addition of one or more
ports and/or tubing.
The invention also provides a method of integrity testing
a multi-layered device, e.g. comprised of more than one layer of porous material,
e.g., membranes, as a whole unit, i.e. without the need for individually testing
each material layer comprising the multi-layered device. Testing a multi-layered
device as a whole unit, compared to testing individual layers, allows for a simplified
design of the multi-layered device because it does not require special engineering
to facilitate integrity testing of each individual layer comprising the multi-layered
Surprisingly, it has been discovered that the sensitivity
of the mixed gas test is increased when a multi-layered membrane device is tested
as a unit compared to a single layered membrane tested under identical conditions
because smaller amounts of the slower, less permeable gas, are able to penetrate
all the layers of the multi-layered device (Figure 12 and Example 8, infra.).
For example, a mixture of gas comprising 90/10 CO2/C2F6,
or the like, may be used to test a multi-layered device according to the invention.
Because less C2F6 is present, small changes in C2F6
concentration indicative of smaller or fewer membrane defects will be more readily
detected compared to a single layer device. The sensitivity of the mixed gas test
is thereby increased and the impact of a defect is more easily discerned. With the
binary gas test, a defect is detected because a portion of the feed gas flows into
the permeate gas via the defect, in effect contaminating the permeate gas and causing
a concentration change. The sensitivity of the test is related to the difference
in concentration between the feed gas and the permeate gas when the membranes are
integral. Since this difference is accentuated when the membrane is binary gas tested
in multi-layer form, the sensitivity of the test in detecting defects is also increased.
As a hypothetical example, consider a situation where the permeate flow rate of
a 10/90 hexafluoroethane/CO2 gas mixture through an integral single layer
membrane is 100 cc/min and the permeate hexafluoroethane concentration is 200 ppmv.
If a leak is developed such that 0.01 cc/min of feed gas flows into the permeate,
the permeate concentration will increase to 210 ppmv, representing a 5% increase
over the integral value. For an integral double layer membrane, in which the Freon
concentration is measured to be 50 ppmv (lower than the single layer due, perhaps,
to the staging effect due to the gas layer between layers), and which will have
a permeate flow rate of about 50 cc/min (half that of a single layer), the same
0.01 cc/min leak of feed gas into the permeate will result in a permeate concentration
of 70 ppmv, representing a 40% increase over the integral value. This result is
surprising because the permeate composition is independent of the thickness of the
membrane material. Without being bound by any particular theory, it is believed
the air or gaseous composition separating the layers may contribute to the lower
hexafluoroethane levels found in the permeate of the multi-layered membrane device
because the gas separation, i.e. of mixed gases used in the integrity test, becomes
a multi-stage separation process increasing the extent of separation of the gases
in the test mixture.
The methods of the invention provide for the use of any
suitable liquid to be used as a wetting agent for the porous material. Selection
of a wetting agent is within the skill of the artisan and may be determined based
on chemical and physical properties of the porous material. Porous materials vary
in terms of their wettability, which is often expressed in terms of the contact
angle &thgr;. The methods of the invention, e.g., the mixed gas test, can be adapted
for hydrophobic membranes, for example, by selecting non-aqueous solvents or prewetting
it with low surface tension fluids (such as a mixture of 30% isopropyl alcohol and
70% water) and exchanging the low surface tension fluid with water. The operating
pressure can be adjusted by selecting fluids with the appropriate surface tension
&ggr;, which generally range form about 74 dyne/cm for water to about 10 for perfluorinated
solvents. A skilled artisan will thus understand that a liquid may be selected by
considering the chemical properties of the porous material to be tested. As an example
where the porous material is comprised of a hydrophilic material a suitable liquid
includes water or a solution comprised of water. The solution may be, for example,
aqueous solutions containing salts and oxygenated hydrocarbons such as aldehydes
or alcohols or neat alcohols such as isopropyl alcohol. Where the porous material
is a comprised of a hydrophobic material a suitable liquid may include any organic
solvent such as dodecane, perfluorinated compounds, carbon tetrafluoride, hexane,
acetone, benzene, and toluene.
The invention provides for flexibility with regard to choices
of liquid and gas components and compositions. In certain embodiments it is desirable
to choose gases which have differing permeabilities in the liquid chosen to wet
the porous material to be tested. In some embodiments a plurality of gases may be
used. Typically the gas which is most permeable in the liquid may be considered
the carrier gas. A tracer gas may be used to detect the presence of defects. The
tracer gas may be any gas which is less permeable in the liquid than the carrier
gas. The test sensitivity can be optimized by selecting gas pairs (in some embodiments)
and liquids with proper &PHgr; in the feed composition. In the limit of using a
dilute tracer gas, the sensitivity of the gas measurement is a function of the feed
composition and &PHgr;.
In general it is useful to choose gas pairs with large
differences in permeability and gas compositions that have one species in trace
concentration and the other present as the bulk species. For example, &PHgr; can
vary from approximately 0.001 to 1 for binary gas mixtures using common species
such as nitrogen, oxygen, carbon dioxide, helium, hydrogen, and hexafluoroethane,
with water as the pore-filling liquid. For tests with hydrophobic liquids, such
as dodecane, gas pairs could include high permeability gases such as ethane, propane,
and butane paired with low-permeability gases such as He, H2, and N2.
In some embodiments at least one of the gases may be Freon, e.g., hexafluoroethane.
In other embodiments at least one of the gases is a noble gas. In still other embodiments
at least one of the gases is CO2. In further embodiments at least one
of the gases is comprised of a mixture of gases. Where the gases are provided as
a mixture of more than one gas, the mixture may be premixed before contacting the
porous material. Wide ranges of gas composition are available; for example feed
gas mixtures of hexafluoroethane in CO2 can vary from less than 0.1 %
to more than 99.9%. The skilled artisan will be able to choose appropriate gases
and gas mixtures based upon known properties such as permeability (Figure 15 a and
Apparatus and Systems
An example of an apparatus suitable for use in the methods
of the invention is shown in Figure 3. The apparatus may comprise a gas source (1)
and feed gas pressure regulator (2). Depending on the volatility of the pore-filling
solution, it may be desirable to optionally saturate the feed gas in a gas-liquid
contactor (3) to prevent premature evaporation of the solution from the membrane
sample (4). A feed pressure-measuring device (5) and a permeate pressure-measuring
device (6) are optionally provided, and may be useful if permeate pressure is not
at atmospheric pressure. The feed (7) and permeate (8) gas compositions are measured
at their respective sample points. Depending on the test duration, the surface-to-volume
ratio of the test apparatus, and the permeabilities of the gases, it may be advantageous
to include a purge valve (9) on the feed chamber to ensure the feed concentration
remains constant during the test. If a purge is used, the feed gas sample point
may be in the purge stream. If it is desired to calculate the pore density, a permeate
gas flowmeter (10) may be used. As an option, the purge gas flowrate (11) and feed
gas flowrate (12) may also be measured, or the permeate flow rate can be calculated
by measuring the composition of the feed, purge, and permeate gases and any one
of the feed gas or purge gas flow rates. The temperature (12) of the filter device
should be measured, e.g., in the permeate gas stream using a thermometer.
The invention also provides a system for assessing the
integrity of a porous material. The system may comprise the apparatus described
above and further comprise a plurality of gases and a sensor device, e.g., a device
to sample and/or analyze permeate flow. Choosing a sensor device is well within
the capability of the skilled artisan. Suitable sensor devices may include a mass
spectrometer, a gas chromatography column, infrared detector, an ultra-violet detector,
a Fourier transform infrared detector, a volumetric bubbler/titrator. Since the
gas composition can vary over 4 orders of magnitude, it is desirable to use a detector
that has a wide operating range. The system may optionally include a computer, e.g.
a personal computer. The computer may be used to control automation of the test
and may also be used to store and/or analyze data.
The system may optionally include a device suitable for
assessing the integrity of a housing which is used to contain a porous material.
Housing defects do not necessarily impact the retentive properties of the porous
material, e.g. the filter. However, they can result in process fluid leaks, and
compromise the overall sterility of the process by providing an ingress route for
adventitious contamination. Incorporating a gas detector exterior to the porous
material housing facilitates concurrent gas detection for integrity and housing
leaks, saving time and equipment An example of a procedure for performing an integrity
test of the housing may include the following steps :
Example 1: Binary gas test as a universal, a priori criterion for integrity.
- 1. Saturate the membrane with the pore-filling fluid, and then drain excess
- 2. Pressurize the system with feed gas at the minimum test pressure. Note that
if test is run at only one pressure, the feed pressure should be set to open all
pores large enough to impact retention.
- 3. Set the purge rate as required to ensure constant feed composition.
- 4. Measure the steady-state feed gas composition and pressure.
- 5. Measure the steady-state permeate gas composition, pressure, temperature,
and flow rate, as required.
- 6. Increase the pressure and repeat steps 3-5.
- 7. Stop the flow, and flush the system to remove gas-saturated fluid.
Several membranes were tested using a feed gas containing
10+/-3% hexafluoroethane in CO2. The tests were run at ambient temperature,
with a feed pressure of 30+/-5 psig, and a permeate pressure of 0+/-0.5 psig. The
feed and exit gas composition was measured by a Cirrus mass spectrometer (MKS, Methuen,
MA). The hexafluoroethane concentration in the permeate gas for the membranes is
listed in Table 1. The integrity of the samples was verified by independent tests.
The membranes were all made by Millipore (Bedford, MA)
and include 0.22 micron Durapore®, a symmetric membrane made from polyvinylidene
fluoride (PVDF), tested in 15 pleated 10-inch cartridge (CVGL); Viresolve®
180, an asymmetric ultrafiltration membrane made from PVDF and tested in 2 flat
sheet samples; and an asymmetric ultrafiltration membrane made from polyether sulfone
(PES) and tested in 5 flat sheet samples.
These membranes have significantly different structural
features such as degree of asymmetry, pore size distributions, thickness, and porosity;
permeability; and materials of construction. As predicted by theory, the permeate
hexafluoroethane concentrations for different integral membranes all fall within
a very narrow range, and is close to the theoretical value predicted based on literature
values for hexafluoroethane and CO2 diffusivities and solubilities, with
no adjustable parameters. Table 1: hexafluoroethane permeate concentrations for
integral membranes and filters
hexafluoroethane concentration, ppm
0.22 micron Durapore®, 10-inch cartridge
V-180, flat sheet
PES ultrafiltration, flat sheet
Theoretical concentration for range of test conditions
The methods of the invention described herein, such as
the mixed gas test can establish a universal, a priori criterion for membrane
integrity. Other integrity tests, such as the air-water test, CorrTest™ and
transient measurements of Betjlich rely on correlations between the test measurements
and independent retention tests to establish the integrity criterion for the test.
The precision of the correlation depends on the inherent variability of the test
and membrane materials, and must be revalidated whenever significant changes are
made to the membrane or test methods, materials, hardware, etc. With the mixed gas
test, the criteria for absolute integrity can be established independent of any
specifics regarding the membrane structure, retention test methods, etc. The factors
that determine the criterion of membrane integrity are the gas composition, choice
of liquid, and pressure ratio.
Example 2: Determination of Defect Size Distribution and impact on Retention
The presence of a defect resulting in a permeate concentration
that differs from the predicted value for an integral membrane may not adversely
affect the membrane performance. The mixed gas test allows defects to be quantified
in terms of their size and population (number per unit area) and is illustrated
in the pressure-ramping method operating mode described in this example.
Two asymmetric ultrafiltration membranes made from PES
were cast at conditions that yielded the same pore size distribution, as measured
by liquid-liquid porometry. However, the casting conditions varied so that one membrane
(201) had defects, while the other membrane (205) was integral. As a result, the
two membranes would be expected to have the same virus retention, with the exception
of the influence of the defects.
The mixed gas test was performed with water as the pore-filling
fluid and 10% hexafluoroethane in CO2 as the feed gas. During the course
of the test the pressure was increased from about 20 psi to about 90 psi. The permeate
concentration was measured by mass spectrometer, and the permeate flow rate was
measured by water displacement. Based on the results, fr was calculated according
to Equation 5, and the defect flow rate calculated from Equation (6).
The results of the test for the membrane with defects are
shown in table 2 below.
The defect flow rate as a function of pressure is shown in Figure 4 for the membrane
with defects; the other membrane was integral and had a defect flow rate less than
10-2 cc/min. As expected for this type of asymmetric membrane, the defects
appear at pressures greater than about 60 psi, suggesting the defects are present
in the thin ultrafiltration layer, and terminate in the underlying microfiltration
support structure. The defect flow increases exponentially with pressure, suggesting
additional defects are continuing to open as the pressure increases.
Feed pressure (psig)
Permeate Freon Concentration (mole fraction)
Permeate Flowrate (cc/min)
flow ratio (calc)
Defect flow (cc/min)
The defect density, or number per area, as a function of
pressure is calculated from Equation (6). Equation (3) was used to calculate defect
diameter as a function of pressure. The defect size distribution (defect density
vs. defect diameter) is obtained by combining the results of Equations (3) and (6),
and is shown in Figure 5.
The two membranes were challenged with a buffer solution
containing of bacteria phage viruses, &PHgr;X-174 (nominally 28 nm in diameter)
and &PHgr;-6 (nominally 90 nm in diameter). The results are shown in table 3 below.
The results demonstrate that the defects present in 201 reduce the membrane's effectiveness,
but that the membrane is still fit for its intended use if the target clearance
is 4 LRV for &PHgr;-6. Consequently the fact that the mixed gas test can provide
a defect size distribution rather than just a pass/fail result allows it to differentiate
among filters with defects.
The impact of the defects on retention can be calculated
a priori using the measured defect size distribution. For virus filtration,
where retention is primarily due to size exclusion, the LRV is related to the defect
size distribution by the following:
where LRV* is the intrinsic retention of the integral membrane and r is the membrane
hydraulic resistance, d is the diameter of the defect, and other symbols are as
Equation (9) is useful because it shows that once the defect
size distribution is known, its impact on retention is independent of the solution
viscosity, concentration, temperature, etc. Consequently the results from the mixed
gas test can be directly applied to a variety of membrane integrity applications
where size exclusion is the primary separation mode.
Assuming the retention measured for the 205 membrane is
the intrinsic LRV*, the LRV for the 201 membrane is calculated via Equation 9. The
results in Table 3 show that the defect distribution has a minimal impact on the
retention of &PHgr; X-174, but does affect the LRV of the more highly-retained
&PHgr;-6. The calculated results are in good quantitative agreement with the measured
results, showing that the mixed gas test can provide quantitative assessment of
the impact of defects on retention.
Example 3: Comparison of binary gas versus air-water diffusion test
The results of this example demonstrate that the mixed
gas test has greater sensitivity, and is less susceptible to extraneous test variables,
than the air-water diffusion test. Three single layer 3-inch asymmetric PES pleated
ultrafiltration filters were made from a single roll of membrane. The filter fabrication
technique may introduce defects into the filters. Consequently, the filters would
be expected to have the same LRV, with any difference due to random defects introduced
during module fabrication. The three filters were wetted with water and tested at
three pressures following the air-water diffusion test. The results, shown in the
Figure 6, demonstrated that all three filters had the same air flow rate.
The three filters were then run with the mixed gas test
using 10% hexafluoroethane in CO2 as the feed gas. The flow ratio as
a function of feed pressure is shown in Figure 7. Two filters, 110-PI-1 and 110-PI-2
showed an increase in flow ratio fr above 40 psig, suggesting defects in
the ultrafiltration layer. Filter 110-PI-1 had the most defects, while 110-PI-3
had the least. Consequently the mixed gas diffusion test was able to differentiate
among the filters, whereas the air-water diffusion test could not (Figures 6 and
Following the method of Example 2, the defect size distribution
was determined for the three filters. The results, shown in the Figure 8, demonstrated
that 110-PI-1 has about 50% more defects than 110-PI-2, which in turn has 50X more
defects than 110-PI-3.
The three filters, and duplicate flat sheet samples of
the membrane that were used to make the membrane, were challenged with buffer solution
containing IgG and &PHgr;X-174. The retention data is shown below in Table 4. The
results demonstrate that the defect reduced the retention of all three filters,
compared to the retention of the presumptively integral flat sheet sample.
75% fouled LRV
This example illustrates the sensitivity of the binary
gas test. It was able to differentiate the three filters which were indistinguishable
by the air-water test (Figure 6). Further, the binary gas test was able to quantify
the defect distribution, showing that defects in filter 110-PI-3 should have a significantly
lower impact on LRV than the other filters (Figure 8).
Example 4: Comparison of gas blends
The flexibility of the mixed gas test can be illustrated
by evaluating the sensitivity of the test using alternative gas blends. A symmetric
membrane with an intrinsic LRV for two different viruses was modeled using the above
equations, and assuming that only a single defect was present. The impact of defect
size on permeate concentration is shown in the Figure 9 for 10% hexafluoroethane/90%
CO2 and 10% SF6/90% CO2 at feed pressure of 90
psig and exit pressure of 0 psig. The results show that both trace species will
increase in concentration with increasing defect size, although hexafluoroethane
is more sensitive for measuring the smallest defects.
Example 5: Correlation between viral retention and permeate composition
The utility of the mixed gas test for establishing a correlation
between test results (concentration or flow ratio) and retention can be shown with
a calculation for a virus filter (Figure 10). The worst-case scenario for a filter
is a single defect of a given size, since the impact of a single defect on retention
is worse than the impact of several defects leaking at the same volumetric flow
rate. Consequently modeling the mixed gas test with the assumption of a single defect
gives the most conservative estimate of the impact on retention. In the calculations,
the filter of Example 5 is assumed to have an intrinsic (i.e. defect-free) LRV*
of 6, 4, and 2.5 for viruses with diameters 80, 40, and 30 nm, respectively. The
permeate composition varies as a result of opening a single defect in the size range
of 100 to 2000 nm. At 90 psig feed, the hexafluoroethane concentration for an integral
membrane is about 135 ppm. As hexafluoroethane concentration increases to 300 ppm,
the LRV of the 80 nm virus species begins to decrease rapidly. The LRV of the 40
nm virus decreases once hexafluoroethane concentration reaches about 1200 ppm. The
LRV of the least-retained 30 nm virus is not impacted until hexafluoroethane concentrations
exceed 3000 ppm. Using these results, it is possible to construct a "worst-case"
correlation between the hexafluoroethane concentration in the permeate and the retention
of virus for the filter. One skilled in the art will recognize that similar calculations
can be conducted for systems using different liquids (e.g., for hydrophobic membranes);
different gases; and different applications (e.g., virus retention, sterilizing
Example 6: Determination of Defect Size Distribution and impact on Retention
for Viresolve® Membrane
The binary gas test method and virus challenge test were
conducted on a series of Viresolve® 180 PVDF membranes following the method
of Example 2. The membranes were cast under conditions that yielded the same pore
size distribution as measured by liquid-liquid porometry, but with varying amounts
of defects in the membranes. This membrane is a composite membrane, with an approximately
110 micron layer of microfiltration membrane supporting a thin ultrafiltration layer,
less than 5 microns, which accomplishes virus removal. Table 5 (below) shows results
for several membranes with increasing number of defects. All samples show a general
trend of hexafluoroethane concentration increasing with pressure, indicating defects
are opening as pressure increases. A second indication of defects is the increase
in the flow ratio, showing a greater proportion of permeate gas is flowing through
defects compared to gas diffusing through the integral portion of the membrane.
The results show that hexafluoroethane concentration, flow ratio, and defect flow
rate correlate with virus retention for operating pressures high enough to open
defects (i.e. greater than about 60 psig). In general, the defect flow rate is very
low below about 50 psi, and then increases at higher pressure. This result is consistent
with defects in the thin ultrafiltration layer. If defects had been present in the
microfiltration and ultrafiltration layers, the hexafluoroethane concentration,
flow ratio, and defect flow rates would have increased at lower pressures. Consequently
the instant technique can provide diagnostic information regarding the location
of the defects in the structure. Table 5 follows:
Phi-X Retention (LRV)
Phi-6 Retention (LRV)
Feed pressure (psig)
Permeate Freon concentration (mol fraction)
Defect flow rate (cc/min)
The defect density as a function of size for the membranes
is shown in Figure 11. The change in defect density correlates to the loss in retention
of the &PHgr;X-174 and &PHgr;-6 viruses. The results show that the mixed gas technique
can provide both a qualitative ranking for the membranes and a quantitative measure
of the defect size and surface population. Further, the ability to define the defect
distribution allows discrimination among membranes. For example, Sample 1 has an
LRV greater than 5 for &PHgr;-6 despite the presence of defects, although retention
is lower for the other samples with more and larger defects. Samples 1, 2, and 3
have LRV greater than 2 for &PHgr;X-174, although the higher defect populations
in Samples 4 and 5 decrease their LRV below 2.
Example 7: Integrity test for Hollow Fiber Modules
The integrity test can be run in different membrane module
configurations, including a hollow fiber. A hollow fiber device comprising 9 1.5mm
ID fibers with a nominal pore size of 0.2 microns and a total area of 100 cm2
was tested with a feed gas of 10% hexafluoroethane and 90% CO2 at pressures
of 11.5, 24.5, and 30.5 psig. At all conditions the flow ratio was greater than
0.5, indicating that the device was not integral.
A second hollow fiber device, model number CFP-2-E-3MA,
manufactured by Amersham Bioscience (Piscataway, NJ), was also tested at pressures
between 10 and 23 psig. The device was certified by the manufacturer as integral,
with a bubble point of 18-30 psi using a 50:50 ethanol-water mixture. The flow ratio
at each pressure was less than 0.005, confirming that the membrane was substantially
defect free. The module was then intentionally damaged to introduce a defect, and
retested at 10 psig. The permeate concentration and flow ratio both increased dramatically,
confirming that the device was no longer integral.
Example 8: Binary Gas Test Comparing Single and Multi-layered Device
A single and double layered polyethersulfone membrane (293
mm diameter disc) were tested using the binary gas test described herein. The double
layered membrane did not have a spacer between the layers. Both membranes were prewetted
with water and then contacted with a gas mixture comprising 90/10 mole percent CO2
/C2F6 at 50 pounds per square inch gauge (PSIG). To maintain
a constant gas composition on the feed side of the membrane, the integrity test
was operated in tangential flow filtration mode with a retentate flow rate of about
four times the permeate flow rate. Based on the measured operating conditions and
solubilities and diffusivities of the test gases in water, the theoretical permeate
concentration of Freon, e.g. hexafluoroethane was calculated to be about 175 ppmv.
For the single layer device, the measured concentration was consistent with the
theoretical value, however, a lower concentration of Freon was observed for the
double layered membrane (Figure 12).
Since the permeate composition is in theory independent
of the liquid thickness, and the membrane layers were adjacent to each other, this
result was not consistent with permeation through a continuous liquid path. Thus
a gas pocket in between the membrane layers may form, and the gas separation may
be divided into a two stage process, resulting in an enhancement of the gas separation.
It should be noted that because there was no retentate or removal of gas in between
layers, the very low concentration achieved in the permeate was transient. But because
of the slow permeation rate of hexafluoroethane, its interlayer concentration buildup
was very slow relative to the time scale of the measurements (5-20 minutes). Therefore,
in practice the measured concentration of hexafluoroethane in the second layer permeate
was several times lower than the value obtained with a single layer and thus provided
a means of integrity testing multi-layered devices without the need of engineering
septum or sampling ports in between layers.
Example 9: Integrity and Rentention Testing of Single Layered and Multi-layered
A series of tests were conducted on a panel of membrane
constructs to compare virus retention performance and binary gas test values of
double and single layer devices, with and without defects. The membranes tested
were suitable for viral retention applications. Using 90 mm discs (47 cm2
effective surface area), a set of integral and defect containing devices were prepared
as shown in Figure 13. The multi-layered membranes were stacked on top of each other.
No physical spacer was used to separate the layers, however it is believed a small
air spaces spontaneously formed either within a layer or between layers. Figure
13(a) shows a two layered integral, i.e. without defects, membrane. Figure 13(b)
shows a two layered membrane having a defect in the top layer (hatched area). Figure
13(c) shows a two layered membrane having a defect in the bottom layer (hatched
area). Figure 13(d) shows a two layered membrane having coinicidental defects in
both the top and bottom layers (hatched areas). Figure 13(e) shows a two layered
membrane having offset defects in both the top and bottom layers (hatched areas).
Figure 13(f) shows a single layered membrane having a defect. Figure 13(g) shows
a single layered integral membrane
The defects were created using a 1000 µm needle (large
enough to cause an essentially complete loss of virus retention) and, except for
combination (e) (offsetting defects), centrally located within the disc. For the
offsetting defect case, the defects were located about 10 mm from the outside perimeter
of the disk and were 180° apart. All the discs were cut from the same PES ultrafiltration
membrane material and prepared in duplicate. For this set of devices, two membrane
layers were carefully assembled one on top of the other
The membrane that was used in these experiments consisted
of two sections: a thin ultrafiltration section (commonly referred to as the skin
side) and a thicker microfiltration section. The two sections formed a continuous
gradient. For the binary gas test, the membranes were tested in both the skin up
(ultrafiltration section upstream) and skin down orientations. For the retention
test, the membranes were oriented in the skin down direction which is often the
preferred orientation for optimum filtration efficiency.
Before retention testing, each device was binary gas tested
using a 10/90 Freon/CO2 gas mixture as the test gas. Gas compositions
were measured using an MKS model Cirrus LM99 mass spectrometer (MKS, Wilmington,
MA). The devices were tested at 50 PSIG, with a sweep/permeate flow rate ratio of
4:1. At each test condition, feed pressure, permeate gas flow rate, retentate gas
flow rate and permeate gas composition were recorded. The retention test of the
devices consisted of permeating 250 ml of a buffer solution containing the bacteriophage
&PHgr;X-174 (approximate diameter of 28 nm) at a concentration of from about 1x106
to 1x108 pfu/ml through the membranes at a constant pressure of 30 PSIG.
Assays of the challenge and effluent streams were performed to determine the virus
log reduction value (LRV).
The results are presented in Figure 14. Letters in Figure
14 correspond to the constructs described above for Figure 13. As can be seen from
Figure 14, with the use of the binary gas test, an integral double layer device
can be differentiated from an integral single layer device and also from a double
layer device in which a defect is present in only one of the layers. Furthermore,
in the cases where defects were present in only one layer, or where defects were
present in both layers but did not overlap or coincide, the impact of defects on
virus retention was significantly greater on single layered membranes compared to
double layered membranes. This results because when double-layered membranes are
tested, the adjoining layer acts as at least a partial blockade of the flow through
Example 10: Comparison of binary gas to air-water diffusion test for a multi-layered
The results of this example demonstrated that the mixed
gas test has greater sensitivity, and is less susceptible to extraneous test variables,
than the air-water diffusion test. Three double layer asymmetric PES flat sheet
ultrafiltration filters, containing either 900 or 1800 cm2 of membrane
area, were made from a single roll of membrane. The filter fabrication technique
may introduce defects into the filters. Consequently, the filters would be expected
to have the same LRV, with any difference due to random defects introduced during
Each of the three filters were wetted with water and tested
at 30 psig pressure using the air-water diffusion test. The three filters were then
run with the mixed gas test using 10% hexafluoroethane in CO2 at 30 psig
as the feed gas and with a purge gas to permeate gas flow ratio of 4:1. These three
filters, along with three 90-mm diameter disc samples (46 cm2 membrane
area) of the membrane that were used to make the devices, were challenged with a
buffer solution containing IgG and &PHgr;X-174. Retention values were measured
after the membranes had been fouled to the extent that flux had declined by 75%
from the initial non-fouled value.
The air diffusion, binary gas, and retention data are shown
below in Table 6. The LRV of device no. 2 was not significantly different from the
average of the control 90 mm disc samples and is therefore considered integral.
Device no. 3 exhibited an LRV that was 0.9 lower than the control discs and device
no. 4 showed an LRV that was 0.3 lower than control discs. As shown in Table 6,
the air-water diffusion test could not distinguish among the three devices, as all
three were measured to have air-water diffusion values in close proximity to each
other. The binary test gas values indicated in Table 6 are the measured permeate
gas concentrations of hexafluoroethane in parts per million. In contrast to the
air-water diffusion test, the binary gas test was able to clearly identify the two
devices where the LRV was lower than the control discs. Furthermore, as indicated
from the data in Table 6, the binary gas test value showed a clear correlation between
the binary gas test value and the deviation from intrinsic membrane LRV of the devices.
Air-Water Flux (cm3/min-m2)
Binary Gas Test Value
75% Fouled LRV
90 mm no. 1
90 mm no. 2
90 mm no. 3
90 mm Average
Device no. 2
Device no. 3
Device no. 4
Many modifications and variations of this invention can
be made without departing from its spirit and scope, as will be apparent to those
skilled in the art. The specific embodiments described herein are offered by way
of example only and are not meant to be limiting in any way. It is intended that
the specification and examples be considered as exemplary only, with a true scope
and spirit of the invention being indicated by the following claims