propane tank

The Case of the Compromised Propane Tank Coating

The exterior coating system on six propane tanks was delaminating after only one year of being stored under cover in Mexico. Could the system be repaired? Or was it a total loss?

A gas transmission company had all six of the propane tanks fabricated and coated in Mexico for a U.S. facility. The specified surface preparation was Near-White blast cleaning in accordance with SSPC-SP 10/NACE 2, followed by application of a 100%-solids polyurethane coating system. The coating thickness was specified to be a minimum of 20 mils. Although the coating could have been applied in one coat, preferably by plural-component spray, to meet the specification, the polyurethane could also have been applied in more than one layer if done within a limited recoat window. After completion, the tanks were reportedly stored under cover until shipment to the facility in the northeastern U.S. nearly a year later. After the tanks had arrived on site, problems with the coatings were observed when the tanks were installed on concrete footings. The coatings appeared to be delaminating between layers or between what appeared to be two coating layers and the steel substrate. High-voltage and low-voltage holiday testing performed by the facility owner revealed numerous holidays.

The investigator visited the site to examine the coatings and to determine what repairs were necessary. Additionally, the intended painting contractor for repair work was onsite to “sweep blast” test sections on the tank exterior to help determine the viability of this method as part of a repair procedure.

The Site Visit

The exterior tank coatings were visually examined, the total coating thickness was measured, and the adhesion and hardness of the coatings were assessed. The evaluation focused on one tank where sweep blasting of test sections was scheduled to be performed.

Visual Assessment

The black coating had a slightly rough or textured appearance over the tank exterior surfaces. Numerous areas on the tank had been previously marked where the coating was delaminating or holidays had been detected.

The marked areas showed where a thick, top coating layer had delaminated from an underlying layer. All layers were black; there was a clean separation between layers where delamination occurred (Fig. 1). The top coating layer surrounding these areas could be further removed (to the underlying layer) with moderate ease by scraping with a utility knife. When adhesion was evaluated by making a cut through the coating (as further described below), the thicker top coating layer separated from the underlying layer in the same manner. An odor characteristic of a polyurethane was also noted when the coating layer was cut into.

Fig. 1: Intercoat delamination Photos courtesy of KTA-Tator, Inc.

There were also areas of coating delamination that appeared to nearly reach the substrate, as evidenced by rust on the surface of the remaining film. The substrate was partially visible through a non-continuous remaining coating layer (Fig. 2). A relatively large area on one tank (approximately 1 square foot) showed this condition. The coating surrounding the delaminated area was irregular and appeared overly thick with sagging present. Other irregularities in the coating film were present along the side of the tank where craters and voids had formed in the top coating layer (Fig. 3). The top coating layer was easily removed when scraped with a knife.

 

Fig. 2: Failing coating

Coating Thickness

 

The total coating thickness was determined over various tank surfaces using an

electronic coating thickness gage. The total coating thickness on the initial tank examined ranged from 20 to 46 mils, with an average of 34 mils. Most readings were over 30 mils, and only one area near the end of the tank had readings consistently below 30 mils. Thickness measurements on one of the larger areas where the top layer of coating had delaminated from the underlying layer showed that the underlying (or first) layer ranged from 9 to 12 mils.

Fig. 3: Craters in surface

 

Thickness measurements for another tank ranged from 23 to 35 mils, with an average of 29 mils. Measurements over a large area of delamination where rust was present ranged from minimal (below 0.5 mils) to approximately 10 mils. Thickness measurements on a third tank ranged from 25 to 30 mils, with an average of 27 mils.

Coating Adhesion

Coating adhesion was evaluated in accordance with ASTM D6677, Standard Test Method for Evaluating Adhesion by Knife. This method involves making two intersecting cuts through the coating to the substrate, with the smaller angle of the cuts between 30 degrees and 45 degrees. The point of the knife is then employed at the vertex of the angle to attempt lifting up the coating from the substrate or underlying coating. The result is rated from 0 to 10 according to the descriptions given by the method. A rating of 0 represents a coating that can be easily peeled from the substrate to a length greater than ¼ inch, while a rating of 10 represents a coating that is extremely difficult to remove.

The adhesion evaluation of the tank coatings resulted in ratings of 2 to 6, with most tests rated as 4 (Fig. 4). The loss of adhesion was typically between the thicker top coating layer and an underlying layer. Adhesion was also determined by subjectively probing the coating with a utility knife. Subjective probing utilizes the user’s experience with the amount of force required to disbond the coating film. Adhesion evaluated in this manner was rated as poor to fair, with the same mode of forced failure occurring between coats.

Fig. 4: Poor adhesion

Coating Hardness

Coating hardness was measured using a Type D durometer. Hardness values are determined by pressing the instrument firmly down on the coating surface. The instrument has a pointed indentor at the bottom that contacts the coating surface and is depressed when the test is performed. The movement of the indentor triggers a dial scale reading from 0 to 100, with 0 indicating a softer surface and 100 a harder surface. Measurements on the tank coatings ranged from 45 to 75, with most readings between 50 and 60.

Test Sections

Three test sections on the first tank, each approximately 2 feet by 2 feet, were prepared by abrasive blast cleaning. The intent was to “sweep blast” the exterior surface and remove the poorly adhered top coating layer. Two areas were prepared near the middle of the tank, where the coating was generally thicker than 30 mils. In both areas, the top coating layer was easily removed, leaving a pock-marked coating layer with nearly bare steel visible in portions of the prepared surface (Fig. 5). The coating that remained was adhered and intact when scraped with a knife, and was uneven or irregular in appearance. Coating thickness measurements generally revealed 10 mils or less of coating remaining.

Fig. 5: Close-up after blasting

The third test section was prepared near the end of another tank, where the total coating thickness was typically less than 30 mils. When this area was sweep blasted, little coating was removed compared to the coating removed from the first two test sections. The remaining coating was intact when scraped with a knife and was uneven or irregular in appearance, particularly near the top of the area where more coating had been removed. Coating thickness measurements revealed 20 to 25 mils of coating remaining. Representative samples of the exterior tank coatings were obtained during the visit for laboratory analysis.

Analysis of Problems

Due to the observation of significant voids throughout the coating, the samples were evaluated under a digital microscope. The examination revealed that voids were present throughout coating cross-sections and on the back side of delaminated coating samples.

The voids present throughout the coating layers, the noticeable odor when cutting into the coating, and low hardness readings indicated a likely problem with application or curing of the tank coatings. The coating product data sheet stated that the cured coating should have a minimum durometer hardness value of 70.

The voids were characteristic of “foaming” that can occur when a fast cure (i.e., plural-component) coating is applied at the incorrect ratio of A and B components. Thinning of the coating, which was not recommended (the product data sheet stated “do not thin”), could also have contributed to the problem. The rough textured surface suggested that application may have been by roller, which is allowed but not recommended by the coating manufacturer.

The consistent separation between coating layers, as revealed by adhesion testing, also indicated that the recoat time may have been exceeded when the top coating layer was applied.

The product data sheet stated that, “a second coat may be applied over the first, if it is applied within the recoat window. Otherwise, roughening of the surface will be necessary to ensure good intercoat adhesion.”

While further laboratory testing could have been performed to verify curing issues or improper mixing, there already appeared to be clear evidence that coating integrity was compromised, based on the relatively poor adhesion and voids present throughout the coating layers. Time constraints on performing coating repairs and placing the tanks into service also limited the extent of laboratory testing.

The preparation of test sections by sweep blasting generally revealed that the top coating layer was easily removed. Although the remaining coating was intact, the coating surface was uneven and irregular. Further examination of photos of the test section surface showed that the remaining coating layer had widespread voids over the surface.

The presence of voids indicated that the existing coating layers, including intact coating that would remain after sweep blasting, would have compromised integrity. Given the criticality of the coating application (underground storage tanks with cathodic protection), there was substantial risk of future coating failure in attempting to sweep blast and overcoat the tanks with a similar coating.

Based on the site evaluation and laboratory results, the existing coatings were recommended to be completely removed by abrasive blast cleaning in accordance with SSPC-SP 10 and replaced with a new 100%-solids polyurethane coating. Application of a new coating was recommended to be in strict accordance with the manufacturer’s recommendations, including by plural-component spray.

 Jayson Helsel, P.E., PCS, a senior coatings consultant with KTA-Tator, Inc., manages failure investigations and coatings projects and is involved with coatings surveys and inspection of industrial structures. He holds an MS in chemical engineering from the University of Michigan, is a registered professional engineer, and is a NACE-Certified Coatings Inspector. He has been published in the past inJPCL and in the Journal of Architectural Coatings, which featured his monthly column, “Getting It Right.”

 

Richard Burgess KTA-Tator, Inc., Series Editor

 

1 comment

Leave a Reply

Your email address will not be published. Required fields are marked *