Case Study - Acid induced stress corrosion cracking

With a few important exceptions, polymeric materials offer excellent resistance to inorganic acids. For example, taking sulphuric acid as reasonably representative of the class, Table 5.7 lists the recommended maximum concentrations and temperatures for some common rubbers and plastics. As acids promote degradation via hydrolysis the exceptions include those polymers that are known to be the least hydrolytically stable such as thermoplastic polyesters, Nylons, polyurethanes, silicones, and polycarbonates. The least resistant polymer in common use and the one most likely to suffer from acid induced stress corrosion cracking is acetal. Cases 5.6.4, 5.6.5 and 5.6.10 illustrate the weakness.

The premature failure of GRP products is most frequently traced to the phenomenon known as stress corrosion cracking (SCC) and the fluid responsible is invariably acidic. The principal degradation interaction is between the fluid and the fibres. The composition of glass fibres is as shown in Table 5.8, a mixture of oxides dominated by silicon. E glass

Table 5.7 Maximum service conditions for common polymers in contact with sulphuric acid
Material	Max. concentration	Max. temperature
UPVC	95%	20 °C
UPVC	80%	60 °C
Polyethylene	80%	60 °C
Polypropylene	80%	80 °C
Bisphenol polyester	75%	20 °C
Natural rubber (soft)	50%	60 °C
Ebonite	70%	70 °C
Polychloroprene	50%	75 °C
Butyl rubber	75%	60 °C

	Composition %
Oxides	E Glass	C Glass	ECR Glass	S Glass
Silicon	54	66	58	65
Aluminium	14	4	12	25
Boron	7	5	0	0
Calcium	20	13	20	0
Magnesium	3	3	2	10
Zinc	0	0	3	0
Others	2	9	5	0
	24 hour weight loss (%)
Water	0.7	1.1	0.6	0.5
10% Na2CO3	2.1	24	-	2.0
10% H2SO4	39	2.2	6.2	4.1

is the general purpose grade offering good tenacity and good resistance to alkalis at modest cost. However the grade is severely attacked by acids as indicated by the rapid loss of mass due to 24 hour exposure to sulphuric acid. The oxides other than silicon are highly susceptible to extraction via leaching. This is understood to be the result of an ion exchange process. The hydrogen ion in the acid replaces the anion in the glass. This results not only in a loss of glass mass but also the creation of microvoids and surface fissures. Due to its very low fracture toughness, glass has a very low defect tolerance and therefore as a result of acid attack the material is rendered fragile.

C glass is usually reserved as the reinforcement for the resin rich acid corrosion barrier. However it should be noted that C glass veils are not appropriate for alkaline environments. ECR glass and S glass are the preferred alternatives to E glass within the structural part of the laminate for acid environments. The superiority of ECR glass has been clearly demonstrated [8]. In this study stress corrosion cracking was monitored on prenotched material in one molar HCl. The crack growth rate increased with increasing temperature and increasing stress (or stress intensity). At the same temperature the ECR glass reinforced laminate required about twice the stress of that applied to the E glass reinforced laminate to induce the same crack growth rate. At the same stress the inferior laminate at 20 °C suffered the same crack velocity as the superior laminate at 80 °C.

S glass would appear from Table 5.8 to offer even better acid resistance. However in practice this is probably not the case. In 1996 a compressed air cylinder constructed from an inner aluminium liner with overwrapped reinforcement of epoxy and continuous S glass fibres, failed at a fire station in California [9]. The internal pressure of air in the cylinder was 31 MPa (4,500 psi) and therefore the failure caused considerable damage to adjacent fire fighting appliances. As the cylinder was part of a self-contained breathing apparatus (SCBA), then had the appliance been on active duty at the time of failure the consequences would have been far more serious and possibly fatal. An investigation into its recent history revealed that 6 days prior to the failure, the cylinder had probably been exposed to an accidental spillage of aluminium cleaning fluid. This fluid contained a mixture of hydrofluoric, phosphoric, and sulphuric acids, and various organic solvents such as 2­butoxyethanol. It had a pH of less than 1.0. A chemical analysis of the fracture surfaces by thermal desorption gas chromatography/mass spectrometry confirmed the presence of 2­butoxyethanol. This, together with the fact that the fracture surfaces at initiation were macroscopically flat and free from fibre pull-out, was sufficient to confirm the diagnosis of stress corrosion cracking via contact with acids in the aluminium cleaning fluid.

The outer fibres of the composite cylinder would have been very close to the surface. Although thinly coated with a polyurethane-based paint this would not have significantly retarded the rate of acid ingress into the outer load-bearing fibres of the laminate. A similar failure involving the exposure of poorly protected fibres to acid is discussed in Case 5.6.8.

An endemic (rather than accidental) example of acid induced stress corrosion has been well researched and is included here in Case 5.6.9. It involves the failure of high tension composite suspension insulators. These have been employed in preference to traditional ceramic insulators because they offer superior resistance to impact. The composite insulators comprise a pultruded E glass fibre/polyester rod. The rod is covered by a rubber sheath to protect it against water ingress. Metal end pieces are bonded to the ends of the rod for attachment to line and pylon. A typical failure due to acid induced stress corrosion cracking is shown in Figure 5.1. It has been claimed [10] that in-service brittle failures have been a major concern of many utility companies, that the rate of failure has increased with each year of service, and that the failure statistics are rarely made available because of 'commercial pressures'.

One set of statistics reports 14 complete fractures and 200 significantly damaged insulators out of a population of 1,756 after only 4 years in service. There is no doubt that these failures are due to acid induced stress corrosion cracking but there has been doubt (and understandable curiosity) concerning the nature and source of the acid. Acid rain (mainly sulphuric) was for many years the principle suspect and there is good reason to accept that some of the failures are due to this. However more recently attention has been directed towards nitric acid. Nitrogen oxides are generated by corona discharges in air and these combine with moisture to form nitric or nitrous acid. It has been suggested that the discharges occur between the rubber sheath and the rod. Although the sheath is a good barrier to water ingress it is by no means perfect. Small quantities of oxides

Figure 5.1 a) Brittle fracture zone in a 115 kV composite suspension insulator

Figure 5.1 b) Schematic of a composite insulator (high energised end) with brittle fracture cracks

(Reprinted from Composites Science and Technology, Vol. 58, A.R. Chughtai, D.M. Smith and M.S. Kumosa, Chemical Analysis of a Field-Failed Composite Suspension Insulator, p.1641-7, Copyright © 1998, with permission from Elsevier Science.)

trapped at the surface of the rod combine with small quantities of moisture to generate small quantities of concentrated nitric acid. It has been shown that the acid at a pH of less than 3.5 is sufficient to attack the composite at the rate observed in service.

The examples of stress corrosion above involve attack on glass fibres, that are close to the exposed surface. For applications where composites are purposely exposed for long periods to acid contact (e.g. fluid transport and containment) the essential means of defence is a corrosion barrier. This may be a separate liner usually employing a thermoplastic of proven acid resistance (PVDF, PVC, PP, etc.), or an integral resin rich barrier. The integral barrier approach has been used successfully and extensively in chemical process plant for the transport and containment of very aggressive acids. Nonetheless failures of such products continue to occur.

Intuitively the barrier reduces ingress of fluids by reducing acid permeation, and therefore the thicker the resin rich barrier the better. However, the rate of uptake of 1M hydrochloric acid by thin sheets of cast polyester under zero stress [11] revealed that while water diffuses readily into the resin (diffusion coefficient of 3 x 10^-9 cm² sec^-1 at 25 °C) and quite rapidly reaches a saturation level of about 2%, the diffusion rate for the chloride ions is very low. Although very mobile, the diffusion of hydrogen ions is limited to that of the chloride ions because charge separation is energetically unfavourable. The acid concentration is diluted from 1 M to about 10^-6 M by this process of selective diffusion. Thus it was concluded that normal diffusion processes through resin rich barriers cannot account for the ingress of acid that is known to cause corrosion in these structures.

The most promising explanation is similar to that given for the environmental stress cracking of thermoplastics (see Section 6.2). Under modest stress, microcracks or microvoids are initiated in the resin. In the absence of an aggressive fluid these would remain stable under stress for many years but otherwise they would provide pathways for the bulk transport of fluids through the barrier layer.