Case Studies in Graphitic Corrosion of Cast Iron Pipe

Abstract

Throughout many parts of the world, cast iron pipe has been and continues to be used for utility piping (water and sewage) and gaseous media piping. Gray cast iron is known to have a limited life in service but is often left in place beyond its reasonable service life. Therefore, end of life is then left to be defined as a rupture or a crack and a subsequent leak, often at an unpleasant time with unpleasant results. The controlling mechanism involved is time dependent and known as graphitic corrosion. This paper provides a review of the phenomenon of graphitic corrosion in the form of key characteristics. Several references are cited that further explore the mechanism involved and corrosion prevention/mitigation. Two case studies are presented that explore graphitic corrosion in gray cast iron pipe from a failure analysis point-of-view. Both case studies feature conditions that further aid in the interpretation of graphitic corrosion and an appreciation for when and how this phenomenon occurs.

Introduction

Graphitic corrosion is a recognized corrosion mechanism [1,2,3,4,5,6] most well-known when involving cast iron pipe. Although a recognized condition, graphitic corrosion has received limited attention in the failure analysis literature and has not otherwise gained wide familiarity within the failure analysis community. The two case studies provided here are presented with the objective of furthering familiarity and thereby advancing the knowledge base in interpreting, confronting, and preventing graphitic corrosion failures

Background Information

It is first useful to present background concerning this phenomenon. In this regard, [1,2,3,4,5,6] already cited are representative sources that serve here as a basis in part for the information provided as follows:

• 1. It is useful to recall that gray cast iron is primarily a mixture of primary graphite in a eutectic pearlitic matrix. The two case studies presented in this paper include photomicrographs of the microstructures of the parent gray cast iron materials involved. Figures 5, 6, 15 and 16 show two variations of flake graphite formation in gray cast iron.

• 2. Graphitic corrosion is a galvanic corrosion mechanism where, in the presence of an electrolyte (in most situations, water of varying degrees of contamination and therefore aggressiveness), steel in the graphite–steel mixture is anodic to the graphite. In effect, the steel gives itself up to the reaction forming a predominant iron oxide matrix as a replacement for the intended steel matrix.

• 3. The phenomenon is often referred to as “leaching.” Leaching as a descriptive term, but the mechanism is galvanic corrosion.

• 4. A word of caution here is appropriate. In many reference sources, graphitic corrosion is misnamed as “graphitization.” Graphitization is a different phenomenon altogether—having to do with steels placed in service at elevated temperatures. The graphitization mechanism is also time dependent. Diffusion is involved but corrosion is not. In graphitization, the carbon in the steel, in place at the onset as carbon intermetallic, transforms by elevated temperature enhanced diffusion into a graphitic (carbon) phase, weakening the strength of the steel of which the component was made.

• 5. A feature of graphitic corrosion is that, once the process sets in, the corroded surface of the component part being attacked tends, unless physically disturbed, to retain its original shape and, to some extent, its original external appearance. Even though the part often appears undisturbed to the naked eye, there now exists a mixture of graphite and corrosion product. In that which has corroded, there is little, or no strength left. The result is a part with reduced load bearing capacity. When subject to abrasion or to being cut, the corroded portion of the part is removed with little effort.

• 6. Graphitic corrosion appears to have been first realized having to do with buried gray cast iron pipe. Here the electrolyte is water, with various soil types and soil contaminants serving to enhance the aggressiveness of the wetted soil as an electrolyte.

• 7. The first use of gray cast iron for utility piping in the USA appears to have been in the 1810s [7,8,9]. In Europe, an early installation of gray cast iron pipe was at Versailles in 1664 [10].

• 8. The initial pipe casting process was horizontal, later changing to a vertical casting method known as pit casting [8, 9]. Pit casting of pipe began in the 1850s and continued into the early 1940s. There remains gray cast iron pipe that was pit cast still in service to this day.

• 9. In the 1920s, Dimitri Sensaud de Lavaud invented a method for producing gray cast iron pipe by centrifugal casting, often referred to as spin casting [10]. This process prevails today [11].

• 10. In 1955, pipe began being made of centrifugal cast, ductile iron [11]. Ductile cast iron remains as the cast iron pipe material to this day. Ductile cast iron, as opposed to gray cast iron, is reviewed further in a later section of this paper. However, although gray cast iron pipe is no longer produced, there remains many hundreds of kilometers (miles) of gray cast iron pipe in service—both pit cast and centrifugal cast.

• 11. As to the identification of graphitic corrosion as a failure mechanism, it is of curious historical interest that a review of gray cast iron pipe failures published in 1922 identified one cause as being the result of “demoralization” of the metal [12].

Case Studies

The two case studies presented here concern gray cast iron pipe. Selection of pipe as case studies is not to imply that it is pipe alone that is susceptible to graphitic corrosion. Many component parts made of gray cast iron, plumbing related or otherwise, are candidates [5, 13]. Also, selection of utility pipe is not to imply that plumbing systems alone are susceptible. For example, gray cast iron piping has been used for distribution of gaseous media such as natural gas for many years [5].

Case A—Buried Water Pipe Trunk Line

On an early Saturday morning, a 15.2-cm (6-in.)-diameter water pipe trunk line fractured. The pipe was located on a side street in a small town. Before the break was discovered, the pipe had been discharging an estimated 204,000 L/h (54,000 gallons/h) for several hours, sufficient to result in water flow-related damage downhill from the break location.

Figure 1 shows a photograph of the street after the break was repaired. The photograph shows several patches to the street. Figure 2 shows a map of the street where the break of concern took place as well as where other breaks in this same trunk line had taken place. Overlaid on the map is a list showing multiple breaks in the line, some of which took place earlier in time. One line break took place a day prior to the break of concern.

Fig. 1

Roadway patch locations at and in the vicinity of the line break of concern; yellow arrows point to location of line breaks

Fig. 2

Topographical view of roadway showing locations of line breaks, including line break of concern

There is no doubt that the line involved was in trouble. The pipe was in such a condition that repairing and recharging the line the day before provided a sufficient transient rise in pressure to create the next break (the break of concern) within one day.

The best records available set the installation of the waterline just after 1906 when the town was founded. The lines installed at that time would have been pit cast (vertical cast) gray cast iron.

As often happens, the pipe section where the break occurred was discarded as part of the repair. However, a piece of pipe was retained when a valve was installed a short distance away a brief time later. Figures 3 and 4 show this piece of pipe, as received for examination. A buildup of scale around the inner wall is evident. The pipe wall thickness was measured to be 12.7 mm (0.5 in.). There is no indication that the pipe was ever lined. Arrows placed on Fig. 4 point to corrosive attack at both the inner and outer surfaces. As described earlier, it is characteristic of graphitic corrosion that an undisturbed corroded object remains at original or near original dimension. For the most part, such is the situation here.

Fig. 3

Section of waterline removed from location near line break of concern

Fig. 4

Closer view of pipe wall in cross section; arrows point to corrosive attack at inner and outer surfaces

The pipe was subjected to a metallurgical examination to confirm that the mechanism of corrosive attack was as suspected—graphitic corrosion, and to communicate the examination results to all interested parties. Table 1 shows the results of a Fourier-transform infrared spectroscopic (FTIR) analysis of the inner wall scale deposit conducted as part of the analysis. The results are unremarkable and not unexpected for a gray cast iron pipe conveying municipal-treated water. The deposit buildup at the inner wall reflects the length of time in service—100 years.

Table 1 Results of FTIR analysis of pipe inner wall deposit (Case A)

Figures 5 and 6 show the parent microstructure in both the etched and unetched conditions. Recall that this pipe would have been manufactured prior to 1910. Nevertheless, the Type VII, Type C graphite flake formation and flake size are indicative of reasonable practice recognized today as acceptable [14]. Figures 7 and 8 show the microstructures in the unetched condition at the inner and outer surfaces. Figure 9 shows the microstructure, as etched, at the same inner surface location as shown in Fig. 7, but at higher magnification. The microstructure displays a collection of corrosion product with graphite flakes. Also shown as isolated patches of white microconstituent, are islands of steadite [15] that resisted corrosive attack. The presence of corrosion product surrounding identifiable graphite flakes verifies that the pipe has been a victim of graphitic corrosion attack.

Fig. 5

Parent microstructure of water pipe; unetched

Fig. 6

Parent microstructure of water pipe; etched with 2% nital

Fig. 7

Microstructure near pipe inner surface; unetched; indications of corrosive attack surrounding graphite flakes

Fig. 8

Microstructure in vicinity of pipe outer surface; unetched; indication of corrosive attack surrounding graphite flakes

Fig. 9

Microstructure in vicinity of pipe inner surface; etched with 2% nital; well-developed corrosion product surrounding intact graphite flake; patches of white microconstituent are steadite formations that resisted corrosive attack

Additional interpretation of the microstructural condition was gained by examination of the as-prepared metallurgical specimen using a scanning electron microscope (SEM). Figure 10 shows the microstructure at a location where corrosive attack is progressing into parent material. Corrosion product is more pervasive to the left on the photomicrograph but can be seen enveloping graphite flakes in greater progression from left to right.

Fig. 10

SEM photomicrograph of microstructure at location where corrosive attack observed progressing into parent metal, corrosion product appearing as darker shaded microconstituent surrounding graphite flakes

Figure 11 shows the results of an examination of the various microconstituents observed in the photomicrograph, Fig. 10. These locations were probed using energy-dispersive spectroscopy (EDS). The EDS results for chemical composition (semi-quantitative) are displayed in Fig. 11. The three regions probed are original eutectoid steel matrix material, Point 1; corrosion product, Point 2; and graphite flake, Point 3.

Fig. 11

Results of EDS elemental analysis of microconstituents at location of progression of corrosive attack: Point 1: parent metal matrix material. Point 2: corrosion product. Point 3: graphite flake

These results, combined with the recent history at the site of the line break of concern, served to confirm to all interested parties that the trunk line at this location (and by inference elsewhere in the town) was in a deteriorated condition due to graphitic corrosion. It remained for the town to address this condition with respect to moving forward. Attention in this regard is the subject of the final section of this paper.

Case B—Vertical Drainpipe

In a home built in 1939, an assembly of gray cast iron pipe draining a second-floor bathtub split along an estimated 1.5 m (5 ft) of its length. A photograph of a piece of the pipe as removed and set aside for laboratory examination is provided in Fig. 12. The intent of the laboratory examination was to characterize the conditions that led to the crack formation.

Fig. 12

Portion of cracked drainpipe, as-received and as-cut to remove cross section for examination

The pipe dimensions are 57.2 mm (2.25 in.) outside diameter with a 6.4 mm (0.25 in.) wall thickness. These dimensions comply with a now obsolete standard ANSI A21.12 [16, 17] for gray cast iron pipe centrifugally cast to these dimensions.

Figure 13 shows the pipe cut in cross section. The coating on the outside wall is shown. There is obvious corrosive attack along the inside wall, with thinning of the wall the greatest at the location where the crack formed—reducing wall thickness of parent metal to 2.1 mm (0.083 in.). Figure 14 shows a laminated collection of deposit on the inside wall surface and corrosion penetration into the pipe wall.

Fig. 13

Pipe cross section showing scale formation and corrosive attack along inside wall

Fig. 14

Corrosion penetration into pipe wall material and deposit collected as laminations along inside wall

Graphitic corrosion was suspected to have been involved, and a metallurgical examination was conducted to confirm this suspected mechanism. Table 2 shows the results of chemical analysis of the parent metal, having a carbon content of 3.80 weight percent and a carbon equivalent weight percent of 4.76. Even though the pipe would have been cast prior to 1939, the chemical composition is consistent with expected composition of gray cast iron pipe manufactured to current requirements for Class A (regular gray cast iron) [18]. Phosphorous content is 0.79% compared to a required maximum content of 0.75%. Sulfur content is 0.057% compared to a required maximum content of 0.15%. Silicon and manganese contents are also as expected.

Table 2 Results of chemical analysis of pipe material (Case B)

An FTIR analysis was conducted of the deposit collected along the inside wall. A summary of the results is provided as Table 3. The deposit is as expected for a bathtub drain that had been in service over 70 years.

Table 3 Summary of results of FTIR analysis of pipe inner wall deposit (Case B)

Figures 15 and 16 show the parent microstructure in the unetched and etched conditions. Given the date of installation and the dimensions of the pipe, the pipe would have been centrifugally cast. Figure 15 (unetched) shows a Type III graphite formation with Type B rosettes. Figure 16 (etched and at a higher magnification) shows a rosette core displaying hypoeutectic gray cast iron with a ferrite–pearlitic matrix. The rosette cores show hypoeutectic gray cast iron with a ferrite–pearlitic matrix. Of interest, again referencing Fig. 16, the rosette core displays a localized Type D formation. The details of the casting process are insufficient to further explore factors affecting these formations. It is known that centrifugal casting of pipe involves water cooling, promoting rapid cooling of the outside surface during casting [10]. Rapid cooling along with a carbon equivalent of 4.76%, a modest hypoeutectic composition, would have promoted the Type B rosette graphic flake formation observed [14].

Fig. 15

Pipe wall material, parent microstructure, unetched

Fig. 16

Pipe wall material, parent microstructure, etched with 2% natal

Figures 17 and 18 show the microstructure at the inside surface in the etched condition. As was observed in Case A, the microstructure is a collection of corrosion product and original graphite flake formation. The lighter appearing microconstituent is steel matrix that has not yet been attacked. Here again, corrosion product enveloping distinguishable graphite flakes indicates that the mechanism involved was graphitic corrosion. Also observed is penetration into the pipe wall of surface deposit material, appearing to envelope the surface as graphitic corrosion nears completion.

Fig. 17

Graphic corrosion underway at inside wall; graphite flakes surrounded by corrosion product (black); etched with 2% natal

Fig. 18

Same location as Fig. 17 but at increased magnification

Further analysis involved SEM and EDS examinations. Figure 19 shows a SEM micrograph, once again showing corrosive attack progressing into the parent metal (from bottom to top in the figure). EDS probes were conducted to determine each microconstituent where graphitic corrosion was observed: matrix material, corrosion product, and graphite flake. The results are displayed at the bottom of Fig. 20.

Fig. 19

Scanning electron photomicrograph showing corrosive attack at pipe inside wall

Fig. 20

Results of EDS elemental analysis at various locations along inside wall of pipe: Point 1: parent metal matrix material. Point 2: developed corrosion with bathtub drain debris. Point 3: localized graphitic corrosion product. Point 4: graphite flake

There are two points of interest concerning Case B relative to the underground water trunk line examined in Case A. First, the inside wall of the pipe would have, given that it drained a lavatory shower/bathtub, experienced intermittent exposure to water. Yet graphitic corrosion proceeded, nevertheless. A wet-dry cycle accompanied by associated expansion and contractions of the deposit is suspected as having been a factor. Figure 14 shows the laminated nature of the deposit and is supportive in this regard. The entrapment of moisture in the laminated deposit is suspected here to have offered ample time for electrolytic corrosive conditions to have persisted in spite of intermittent periods of exposure.

A second point of interest is the extent of the separation of the pipe crack surfaces, measured to be a gap of 2.1 mm (0.083 in.). The opening of the crack line is an indication of a presence of residual stress. Recall that centrifugal casting of pipe involves water cooling at the outside surface. As the pipe material cooled during manufacture in a progressive manner from the outside surface inward, the wall material favoring the outside surface would have been placed in residual tension [19]. The material favoring the inside surface would also have been in residual compression. Now, as the inside wall material was lost to corrosion over time, the balancing residual compressive stress in the pipe wall would have been lost in the same progression. That is, as corrosion proceeded, the extent of tensile residual stress in the pipe wall was increased. The eventual result was the crack and the gap observed along the crack line. This condition was addressed using finite element analysis (Fig. 21). Given the residual wall thickness at the crack location and the size of the crack opening, the estimated average residual circumferential tensile stress in the pipe was 148 MPa (21,584 psi). A stress of this magnitude, assuming a minimal tensile strength of the pipe material to be 138 MPa (20,000 psi) (ASTM Class 20), was sufficient to result in the cracking of the pipe.

Fig. 21

Results of finite element analysis

In other words, the intermittent moisture exposure of the deposit on the inside wall of the pipe was a condition sufficient to promote graphitic corrosion. The consequent progression of corrosion reducing structural pipe wall material at the inside wall of the pipe resulted in the crack in the pipe even in the absence of internal pressurization. Residual stress in the pipe wall was sufficient to crack the pipe with its now corrosion-induced, thinned wall.

Remedial Measures—A Review

The existing literature concerning cast iron pipe installations, and in particular, gray cast iron installations, confirms that losses involving existing pipelines present a concern [20]. There are today, gray cast iron pipelines in place that are more than 50 years old and, in some cases, are in excess of 100 years old. In one recent article published in the Journal of The American Water Works Association, the situation is likened to an epidemic with the condition of corrosion as a disease becoming manifest with an increasing rate of occurrences over an existing pipeline population [21].

The corrosion problem is not limited to utility water and sewer lines. An estimated 52,000 meters (32,400 miles) of natural gas trunk lines remain, as of 2012, in service in the USA. These lines are predominantly in the states of New York, New Jersey, Massachusetts, and Pennsylvania [5]. Prior to the 1950–1960 time period, most of such installations were gray cast iron pipe.

Also, the corrosion conditions involved go beyond graphitic corrosion such as: chemical attack (alkaline soils), microbial attack and induced electrical current. These various conditions are enhanced by mechanisms associated with corrosion such as: erosion corrosion, stress corrosion, corrosion fatigue, intergranular corrosion, and crevice corrosion.

In other words, an increasing rate of loss involving cast iron pipelines, be it gray cast iron or ductile cast iron pipe used in the transport of water, sewer, natural gas or other fluids, is expected to be prevalent in locations throughout the world. This situation begs the question as to what is, therefore, now being done and what is yet to be done. In other words, to what extent have assessment and remedial measures been considered, let alone undertaken. In this regard, there are two considerations. The first is the consideration of fitness-for-service of existing cast iron pipelines. The second consideration concerns the mitigation of corrosion potential in the installation of new cast iron pipe, otherwise known as corrosion control. This latter consideration has two aspects. Treatment of the pipe exterior surfaces and treatment of the pipe interior surfaces.

Each of these considerations is addressed below. The literature, including published studies, is beyond exhaustive treatment in this paper. The approach here is to identify specific references with brief descriptions of each. The references identify additional reference material. Basis for further study is thereby provided.

• 1. Fit-for-Service

  • 1.1 “Characterization and Fitness for Service of Corroded Cast Iron Pipe” [5].

    This report was sponsored by the US Department of Transportation, Office of Pipeline Safety. A thorough review of gray cast iron metallurgy is presented along with a detailed study of graphitic corrosion with additional case studies. The end-product is a step-by-step procedure for rating fit-for-service conditions of existing gray cast iron pipe.

  • 1.2 “Deteriorating Buried Infrastructure Management Challenges and Strategies” [22].

    This report was sponsored by the US Environmental Protection Agency, Office of Ground Water and Drinking Water. Various conditions of corrosive attack are reviewed. The emphasis is on corrosion relative to pipe inside surfaces and effects on the quality of potable water. There is also an emphasis on rehabilitation and renewal of existing pipelines, describing different lining technologies.

  • 1.3 “A Methodology to Estimate Remaining Service Life of Grey Cast Iron Water Mains” [23].

    This paper aids concerning decisions to repair, renew or replace existing pipe. The emphasis is on measurable pit depth and surrounding soil condition. A flowchart, as a guide in decision making, is provided.

• 2. New Installation, Corrosion Mitigation—Exterior Surface

  • 2.1 “Corrosion Considerations for Buried Metallic Water Pipe” [24].

    The subject area is new installation of ductile cast iron pipe and corrosion mitigation systems. This report was produced by the United States Department of the Interior, Bureau of Reclamation. Levels of mitigation protocols were proposed, to a greatest extent based on the soils in which the pipe is to be buried. These requirements proved controversial within the cast iron pipe industry, see 2.2 through 2.4.

  • 2.2 “Corrosion and Corrosion Control of Iron Pipe: 75 Years of Research” [25].

    This paper reviews various means for mitigating corrosion of cast iron pipe, with an emphasis on polyethylene encasement. Corrosion rate comparisons are made between gray cast iron pipe (phased out by 1965) and ductile cast iron pipe as the modern-day cast iron pipe replacement.

  • 2.3 “Corrosion Control-Polyethylene Encasement” [26].

    Based on history and development records of the Ductile Iron Pipe Research Association, a case is made for the sufficiency of polyethylene encasement as a corrosion mitigation means for installations of ductile cast iron pipe. This publication also describes a ten-point system for evaluating the corrosivity of installation soils. In this regard, see also [27].

  • 2.4 “Review of the Bureau of Reclamation’s Corrosion Prevention Standards for Ductile Iron Pipe” [28].

    The controversy over the Bureau of Reclamation requirements for buried ductile cast iron pipe resulted in a publication by the United States National Research Council. Together with the Sources 2.1 through 2.3 identified above, the subject of mitigating corrosion of new installation of ductile cast iron pipe is given thorough treatment.

  • 2.5 “Design Decision Model for Corrosion of Ductile Iron Pipelines” [29, 30]

    These publications concern a point system-based model for use by entities in deciding how to mitigate ductile cast iron corrosion as a function of installation location characteristics. The system is referred to as a “Design Decision Model.” The model is based on experience gained from test installations and field investigations.

• 3. New Installation, Corrosion Mitigation—Interior Surface

  The conventional method employed to mitigate corrosion at the internal surfaces of cast iron pipe has been cement-mortar linings, an application in the USA dating back to the first such applications in the 1920s.

  • 3.1 “Cement-Mortar Linings for Ductile Iron Pipe.” [31]

    This paper reviews the history of cement-mortar linings of cast iron pipe as well as the development and manufacturing processes over time. Performance service is also reviewed.

  • 3.2 “Cement Mortar Linings in Cast and Ductile Iron Pipes: Life Expectancy and Dependence Upon Water Chemistry” [32]

    This paper addresses a study of the long-term performance of cement-mortar lining, focusing on factors influencing life expectancy and sub-lining attack of the pipe material. The work is based on examinations of buried pipe in service as a function of time.

Conclusion

Presented have been two case studies concerning the topic of graphitic corrosion, for the purpose of renewing awareness of this corrosion mechanism. Effort was taken to dispel what myths/misunderstandings may still linger concerning this mechanism—for example, graphitic corrosion versus graphitization. Of particular interest has been emphasis on the potential for yet more failures of grey cast iron piping systems that remain in the ground with increasing susceptibility to graphitic corrosion. In this latter regard, effort was taken to provide the reader with a resource base for future failure assessments and prevention.