Last revised on September 22, 2014
Impermeable pipes
Antonio Carvalho – IBCOM
Introduction: The composite pipes currently used to carry water, sewage, oil, gases and
industrial solvents have three drawbacks. While generally not considered as serious, these
drawbacks are nevertheless questioned by some end users and have, in a few cases,
developed into failures. The composite pipe manufacturers do not see this situation as a
clear danger to their market position and have, therefore, taken no action to address these
issues.
The three drawbacks that upset the composite pipes are summarized in the following
statements.
1. All composite pipes are permeable to water, gases and solvents.
2. All composite pipes are designed and tested for weep, instead of burst failure.
3. The sand core used in underground pipes are susceptible to delamination from
impact and mishandling
The above minor nuisances do not seriously affect the pipe’s performance, but should be
eliminated. The problem, however, is how to eliminate them. Their apparently unsolvable
nature has challenged the pipe fabricators for decades. Over the years the contractors and
designers of composite pipes have learned to adapt and live with this apparently unsolvable
situation.
This paper will suggest a low cost solution that would make the already good composite
pipes into even better products. This new solution may be a major step in the direction of
offering improved pipe products for water, oil, gas and solvent transmission.
1 – Permeability to water, gases and solvents. The high permeability of composites is a
cause of concern in two situations.


The fear of soil contamination from solvents carried in the pipes.
The fear of water contamination from solvents in the soil.
The risk of contamination from either the inflow or the outflow of solvents has always been
a concern for highly permeable plastic underground pipelines. In addition, the high
permeability has also placed some constraints on the applicability of such pipes in high
pressure/high temperature applications involving gases and solvents. Recently a new mode
of failure has been identified which links the permeability of composite pipes to an
anomalous type of rupture failure. This new mode of failure is so unexpected and unusual,
that we have named it as “anomalous failure”. This paper will present a brief discussion of
the anomalous failure.
The permeability of composite pipes is a property that depends on the resin matrix as well
as on the permeating material. It is obtained by multiplying the diffusivity D and the
solubility S, as in the following relation
P = DS
Where P is the permeability, D is the coefficient of diffusion and S is the solubility of the
solvent in the resin.
The diffusivity D depends on the resin matrix, the arrangement of the reinforcement in the
laminate and the penetrating material. The solubility S also varies with the penetrating
material and the type of resin. Perhaps the multiplicity of values for D and S has
discouraged research to measure them. Not many values for D and S are found in the
literature. We have a few words to say about these two parameters.
We begin with the diffusivity D.
1.1 – Diffusivity D. Not many values have been published for the parameter D. We have
been able to locate the following, from reference 1. The measurements were taken at 37
C for water, methanol and toluene in laminates made with two high performance resins.
Laminate
D for water
(𝟏𝟎− 𝟏𝟐
𝐦𝟐
)
𝐬
Type of resin
High reactivity polyester
Vinyl ester novolac
Casting
4.0
1.0
Chopped strand mat
2.0
1.0
Woven roving
2.0
0.5
The diffusivity of water is 4 times larger in polyesters than in vinyl ester novolac.
Laminate
D for methanol
(𝟏𝟎− 𝟏𝟐
𝒎𝟐
)
𝒔
Type of resin
High reactivity polyester
Vinyl ester novolac
Casting
2.0
2.0
Chopped strand mat
0.5
0.5
Woven roving
0.5
0.5
The diffusivity of methanol is the same in polyesters and in vinyl ester novolac.
Laminate
D for toluene
(𝟏𝟎− 𝟏𝟐
𝒎𝟐
)
𝒔
Type of resin
High reactivity polyester
Vinyl ester novolac
Casting
0.5
0.02
Chopped strand mat
0.1
0.01
Woven roving
0.01
0.01
The diffusivity of toluene in novolac resins is too small.
It is obvious that the inclusion of an impermeable metallic liner in the pipe would stop the
flux of solvents and reduce the diffusivity to D = zero. We have developed a low cost
aluminum foil that bonds to the laminating resin to make pipes that are absolutely
impermeable. We will have more to say about this aluminum foil as we proceed.
Having discussed the diffusivity D, we move on to the solubility S.
1.2 – Solubility S. The Underwriters Laboratories have tested and approved specific resins
for tanks used in the storage of ethanol and methanol. The UL approval is valid for nonstructural and room temperature applications. This situation changes in high temperature
and structural applications. Why is that so?
The reason is the solubility parameter S. The maximum working temperature and the
structural feasibility of laminates in solvents is controlled by the solubility parameter S.
The resin matrix may absorb large quantities of solvents, which swell the resin and strain
the composite, reducing its ability to perform in structural applications and in high
temperatures.
The decrease in operating temperature and the loss of structural capability are the bad news
regarding solvents. The good news is the solvents are not reactive with the resin and do not
cause long-term damage. The damage that they do is done in the short-term. The
deterioration ceases when saturation is reached. The maximum damage is reached as soon
as saturation is achieved.
The maximum solvent absorption at saturation, also known as solubility S, can be measured
by conducting immersion tests of resin castings in solvents. The tests are ceased when the
resin saturates. Inasmuch as composites are concerned, the organic solvents can be
classified in three categories.



Low absorption - The solvents in this category do little harm to composites and are
in fact not a deterrent to structural applications. They do, however, reduce the
maximum working temperatures. The most notorious representative of this category
is water, which saturates the resin at low levels of absorption, usually in the 1%
range. The low swelling caused by such small absorption do little structural harm to
the composite and allow its use in demanding load bearing applications, such as
high pressure pipes. The use of high pressure composite pipes in the oil industry is a
widespread application. Low absorption solvents like water, therefore, are not a
deterrent to the use of composites in load bearing structural situations.
Medium absorption - The level of swelling caused by the solvents in this category is
moderate and may allow the application of composite in non-load bearing
applications and low temperatures. There is a variety of organic solvents falling in
the category. The most notorious are, maybe, ethanol and methanol, which saturate
the resin at circa S = 10% absorption. The S = 10% saturation observed in
ethanol/methanol should be compared with the S = 1% observed in water.
High absorption - The swelling caused by the solvents in this category is too high,
leading to resin cracking and to the failure of the composite even in non-structural
applications. Acetone is a good example of a solvent in this category.
The following tables illustrate the solubility of water, ethanol and toluene in castings made
with two premium resins. The tests were conducted at 37 C. The reported values are for
resin castings.
Resin
S for water (%)
High reactivity polyester
1.0
Vinyl ester novolac
1.2
The water uptake (1.0%) is small, allowing the use of composite pipes in structural applications.
The water pickup causes a small decrease (usually 10C) in the maximum operating temperature.
Resin
S for methanol (%)
High reactivity polyester
10.0
Vinyl ester novolac
13.0
The high (10.0%) methanol uptake allows the use of composite pipes only in non-structural
applications at room temperature, as in underground storage tanks.
Resin
S for toluene (%)
High reactivity polyester
18.0
Vinyl ester novolac
10.0
The high (18.0%) uptake indicates that polyester resins may not be adequate for contact with
toluene. The vinyl ester novolac, with uptake of 10.0%, may be used in non-structural applications
at room temperatures.
In conclusion:
1. The solubility of water S = 1% is small and has no appreciable effect on the
composites, other than a minor, say 10C, decrease in the maximum allowable
temperature.
2. The solubility of the majority of the organic solvents is high enough to reduce the
working temperature and the pressure rating of composite pipes to very low values.
In some cases the solubility may be so high as to preclude the use of composites in
structural applications even at room temperatures.
3. There are solvents with solubility so high that composites should be avoided,
regardless of pressure and temperature.
It is obvious that the inclusion of an impermeable metallic barrier in the pipe liner would
bar the entry of solvents and eliminate all these difficulties. We have developed a low cost
aluminum foil that bonds to the laminating resin to make pipes that are absolutely
impermeable. We will have more to say about this aluminum liner as we proceed.
Having discussed the diffusivity and the solubility, we are ready to move on to the
permeability.
1.3 – Worked examples. The flow of material in composite laminates can be estimated by
Fick’s first equation
[flow] =
D×S×C
t
Where [flow] is the quantity of material passing across the unit area of laminate, D is the
coefficient of diffusion, S is the solubility, C is the solvent concentration in the solution
carried by the pipe and t is the wall thickness.
The solubility S and the diffusivity D parameters of various solvents were discussed in the
previous section and are supposed known. The permeability is obtained as the product P =
D.S.
Example 1 – What are the expected differences in permeability to water, methanol and
toluene in pipes made with the two premium resins mentioned in this paper?
The permeability is obtained by multiplying the diffusivity and the solubility.
P = D.S
Permeability to water:
The solubility of water is approximately the same (1.0%) in both resins
The diffusivity of water is 4 times higher in polyesters than in novolac vinyl esters.
Therefore we can expect the flow of water to be 4 times larger in polyesters than in novolac
vinyl esters.
Permeability to methanol
The solubility of methanol is approximately 30% higher in novolacs
The diffusivity is approximately the same in both resins.
Therefore we can expect the flow of methanol to be 30% lower in polyesters.
Permeability to toluene
The solubility of toluene is 80% higher in polyesters.
The diffusivity of toluene is at least 25 times higher in polyesters
Therefore we can expect the flow of toluene to be negligibly small in novolacs when
compared to polyesters.
Example 2 – Calculate the flow of water through the walls of a polyester pipe at 37 C.
Assume the following conditions
𝑚2
The wall thickness is t = 10.0 mm. 𝐷 = 12 × 10−12 𝑠 (coefficient of diffusion for water)
S = 0.01 (solubility of water in polyester) C = 1 g/cm3 (the pipe carries water 100% pure).
Entering the above in Fick’s equation
𝑚2
12 × 10−12 𝑠 × 0.01 × 1 g/cm3
[flux] =
10 mm
[flow] = 378 g/m2.year
The flow of water is very high. In 10 years it reaches almost 4 liters per square meter in
pipes that are 10.0 mm thick. The water that permeates the pipe wall dissipates into the
atmosphere. Should the pipe wall have cracks, voids or delamination, the water would
accumulate and fill the cavities before proceeding on its way. The pressure of the water in
the cavities would drop linearly from a maximum in the cavities located near the pipe’s
inner surface to a minimum in those near the outer surface.
There is nothing unusual about this flow of water. The resin in the inner surface is saturated
with water, while that on the outer surface has zero concentration. The cavities in the wall
are all filled up with water at different pressures, depending on their distance to the inner
surface.
The important point here is that the wall cavities are filled with pressurized water. Under
constant pressure this has no negative effect on the pipe. The problem arises when the
pressure inside the pipe oscillates.
Suppose the pressure drops inside the pipe. As this happens, the pressure in the water filled
cavity becomes higher than the pressure inside the pipe. This pressure mismatch causes a
slight expansion of the cavity. The cavity grows by a small amount, just enough to balance
the pressures. As the pressure inside the pipe rises, the cavity is once again filled with water
at high pressure. This is followed by a new pressure drop and a new cavity expansion,
etc…. After many cycles of pressure oscillation, the cavity may develop a considerable
enlargement. And the enlarged cavity is filled with water at high pressure. This process will
not stop and eventually the cavity develops into gigantic blisters.
The blisters grow unchecked until eventually they rupture the pipe’s inner wall. In some
cases, the pipe itself may rupture. This mode of failure was recognized just recently (mid
2012) and came as a surprise. In fact the surprise was so intense that this mode of failure
has been named “anomalous failure”, to emphasize that it has nothing to do with the ageing
processes known to affect composite materials. In particular, it has nothing to do with
osmotic pressure.
Anomalous water blisters. The photos show the cracked sand core that forms the cavity where the
anomalous crack grows. Also shown are two anomalous blisters filled with pressurized water. The
picture in the lower right hand side shows a ruptured blister.
The anomalous failure may be a disaster in pipelines carrying gases, as the gases (unlike
water) are compressible and capable of large expansions. The development of anomalous
cracks in gas lines may be very fast and very ugly. That is a strong reason to avoid the use
of composite pipelines in gases under high pressure. The anomalous failure would grow
and disintegrate the pipe in the vicinity of voids or other laminate flaws that would allow
the accumulation of gases.
The inclusion of an impermeable metallic barrier in the liner should bar the entry of water
and gases and eliminate the anomalous failure. We will have more to say about this
impermeable liner as we proceed.
2 – Weep failure. The weep failure dominates all current commercial codes, standards and
specification for composite pipes. The design requirements, as well as the qualification and
quality control tests embodied in the standards AWWA C950, API 15 HR and ISO 14692
are for the most part based on the weep mode of failure. The cost of composite pipes would
drop significantly should the requirements for weep testing be eliminated.
The weep failure comes from the flow of water through cracks that develop in the glassresin interphase when the pipe is pressurized. The length, opening and number of these
cracks depend on the magnitude of the tensile strain that is imposed by the pressure. As the
pressure increases, so do the number, length and opening of the cracks. If the pressure is
high enough, the cracks will coalesce and allow the passage of water. At this point, the pipe
will weep.
The aluminum foil will not prevent the pipes from weeping. The foil is too thin (10
microns) and not tough enough to arrest the development of cracks in the laminate. The
impermeable pipes with aluminum foil crack and weep like any other pipe. However, the
foil plays a fundamental role in reducing the cost of testing against weep.
The quality control tests listed in the commercial codes are designed to assure absence of
gross manufacturing defects that would lead to premature weeping. The pipes are known to
weep and in fact are expected to weep above the threshold weep strain. This is an accepted
fact. However, expensive tests are required to assure the absence of manufacturing defects.
By its own nature, the impermeable aluminum foil will prevent any premature failure from
manufacturing defects, and eliminate the need for the expensive weep quality control tests.
This, in itself, is a major advance in the composite pipes technology.
4 – Susceptibility to delamination cracks. The aluminum foil will not protect the pipe
from weeping. The foil is not tough enough to arrest the cracks that grow across the plies.
The pipes with aluminum foils will have the same weep performance as any other. The
good news is the aluminum foil is not affected by delamination cracks, that is, those cracks
that develop parallel to the plies.
The delamination cracks are a major concern in sand cored composite pipes. The sand core
is susceptible to delamination cracks that come from mishandling and other impact loads.
This has always been a major source of worry for contractors and pipeline owners over the
years. The cracks that develop in the sand core of pipes that have been mishandled or
impacted are the preferred sites for formation of anomalous blisters.
Have the impacts from shipping and receiving, or from that tool that dropped, caused
damage to the pipe? Have we at some point inadvertently mishandled the pipe? What is the
effect of that tiny little rock that we may have overlooked to remove from the trench
bedding in the installation? A visual inspection will indicate no apparent damage, but how
about the sand core? Has it delaminated? There are no simple answers to these questions.
The worry and the doubt will always be there.
It is obvious that the inclusion of the aluminum foil will stop all these worries. The pipes
made with aluminum foils in the liner could be handled just like any other pipe.
We will presently proceed to describe the impermeable pipes.
5 – Market acceptance - The solution that we propose is very simple and straightforward
and consists essentially in making pipes, fittings and joints having a low cost impermeable
liner to prevent the water/solvent from entering the structural wall.
Our solution is market-ready and does not require extensive and expensive testing. In many
instances the qualification routine is not really necessary, since the pipe manufacturer
would in fact be offering a “new product” that is essentially the same regular, pre-qualified
product from his portfolio, plus the additional impermeable liner.
Figures 1, 2 and 3 show the manufacturing process. The impermeable metal foil is helically
wound on a rotating mandrel prior to the lamination of the structural plies. The foil will
have perfect adhesion to the laminating resin and be an integral part of the pipe wall.
The end-users will realize that the “new impermeable pipes” are in essence the same “old
standard pipes”, with the added feature of the impermeable foil. Market acceptance is not
expected to be a major hurdle. The impermeable liner eliminates the three drawbacks that
currently affect the composite pipes.
Fig. 1 - The pipes are made in the usual way. The impermeable metal foil is helically wound on the
rotating mandrel prior to laminating the structural plies
6 - Conclusion. The inclusion of impermeable liners in composite pipes brings the
following benefits:





It allows the use of low cost resins for solvent transmission in high pressure and
high temperature situations.
Lower cost and process friendly vinyl esters can substitute for epoxies in oil + gas
pipes
Elimination of tests related to weep failure
Elimination of the anomalous failure
Elimination of need for extra careful handling of components
Fig 2 The aluminum foil is helically wrapped over the mandrel
Fig. 3 The pipe in the picture has the following layup:




Conductive inner layer. The dark color is imparted by graphite. Conductivity is a
requirement for hydrocarbon solvents.
Impermeable liner of tape wound Aluminum foil.
Filament wound ± 55 angle-plies.
Polyester terephthalic resin
Biography: Antonio Carvalho is an engineer with 45 years dedicated to composites. Past
experience includes 30 years with Owens Corning and 15 years as a full time consultant.
For direct communication please contact [email protected]