Creating a Tank Within a Tank: An Infallible Solution for Lining a 40-Year-Old Moisture-Laden Concrete Underground Water Tank

From June through August, the average daily temperature in Phoenix crests 100°F. To protect students, faculty and staff at the West Valley campus of Arizona State University (ASU) from these sweltering conditions, the university has employed a complex network of large underground water storage tanks. Their contents are cooled at night when energy costs are low, then circulated through the air conditioning system.

On the West Valley campus, three thermal energy storage tanks were in need of rehabilitation. The first to undergo repairs had been built 40 years ago and relined 10 years ago, but only five years after that, the new lining had begun to delaminate (Figure 1).

The peeling of this coating was only the first indication that the project would be one with many special conditions and obstacles.

Because the tank was constantly in service and full of water, the team assigned to the rehabilitation was not able to see its inside until it was drained and opened up for the rehabilitation work in the spring of 2026. Located below an open area in the center of campus, the tank is 250 feet in length by 50 feet wide by 40 feet high. It is only accessible through two concrete pill boxes that are visible above ground (Figure 2).

The applicator for the project, Professional Piping Systems (PPS), and the general contractor, Chasse Building Team, followed confined space protocols, securing permits, setting up an air handler above ground and maintaining communication with the local fire department.

Once they were set up in the tank, the crew discovered something very surprising. The failed tank lining was soft and flexible, not hard as they expected. It was blistered everywhere and was peeling off in large sheets (Figure 3). Blasting would not be effective with such a rubber-like material, so the restoration crew had a largely a manual job to remove it. Where necessary, the crew used angle cutters to score the lining before pulling it off.

PPS could not identify the failed coating, nor could the new coatings supplier for the project, Sherwin-Williams Protective & Marine, and there were no existing records of that project. Their best guess was that it was an elastomeric polyurethane system. The crew did observe, however, that the concrete surface of the tank had not been properly prepped for the lining application 10 years ago, as it was still very smooth and without a sufficient concrete surface profile (CSP) that could have given the failed coating a better chance for adherence.

Restoration of the tank took great effort to specify an appropriate lining system that would not repeat the failures of the previous system. Creating, testing and applying that lining earned the project team of ASU, PPS, Chasse and Industrial Inspection Group (IIG) top honors in the non-potable water infrastructure category for the 2026 Sherwin-Williams Impact Award program. The award recognizes challenging projects that feature high-performance coatings from Sherwin-Williams Protective & Marine.

Prep Work Begins, Then Hits an Obstacle

After removing the old coating (Figure 4), PPS blasted the surface of the walls, ceiling and floor in accordance with SSPC-SP 13, taking care not to remove too much concrete, especially around damaged areas. For example, parts of the ceiling had large bug holes and exposed aggregate (Figure 5). Throughout the structure, the crew achieved a surface texture of CSP 3-5, ensuring the new coatings would have sufficient surface area available to achieve a strong mechanical bond (Figure 6).

“We were conscious of how far we were blasting and very aware of where to direct our nozzles so we wouldn’t overblast and create more work,” said Brian Cho, Project Manager for PPS. “Still, given the condition of the concrete in spots, we had some long strips where we had to trowel on a resurfacer to cover the aggregate.”

As a matter of course, IIG, the inspector for the job, conducted pH tests on the walls, ceiling and floor. These results were acceptable.

Next, the inspection team conducted several rounds of moisture tests, including plastic sheet testing (Figure 7) and 24-hour calcium chloride testing (ASTM F1869). Moisture levels were acceptable on the walls and ceiling, but too high on the floor. To better understand the situation, IIG inspectors installed moisture probes in the floor, requiring the crew to drill holes 3/4 of an inch in diameter by 3.5 inches deep and conduct ASTM F2170 in-situ probe moisture testing. The probes showed relative humidity in the concrete at a forbidding 98%.

Moisture levels on the tank floor did not decrease over several days. Even with industrial-grade dehumidifying equipment running above ground all day (but not at night), the relative ambient humidity in the tank was 60%.

Around this time, the crew discovered that a nearby irrigation system appeared to have a leak, a situation that was improved but never completely resolved.

Next, the crew ran the dehumidifying equipment continuously, day and night, and added heaters inside the tank. That brought the ambient relative humidity down to between 14% and 25%, an acceptable level. But the probes were still reporting between 98% and 100% relative humidity in the concrete floor. There was no question the specifications for the floor coatings would need to change.

A Major Spec Change as Tension Mounts

The original spec for the walls and ceiling would remain the same: spot resurfacing with Dura-Plate® 2300; a coat of Macropoxy 5000 as a penetrating primer/sealer; followed by a thick coat of Dura-Plate 6100 at 125 mils dry film thickness (DFT). But the floor would require a moisture mitigating solution before coatings could be applied.

ASU was naturally wary, considering that the previous coating had failed within five years of the application.

“Given its flexibility, the previous lining was likely an elastomeric polyurethane system, which may not have been the best choice for the moisture-laden below-ground environment,” said Bobby Merigan, Chief Critical Infrastructure Consultant with IIG. “Polyurethanes are typically hydrophobic, so the concrete’s high moisture content may have been a factor in the delamination of the previous lining, along with the inadequate surface preparation that was revealed after peeling off the old system.”

Earlier in discussions, the university had set its sights on using a waterproofing treatment, a competing product that had performed well on new, recently constructed thermal energy tanks at a different campus. But Sherwin-Williams explained that this product was not appropriate for older concrete.

Sherwin-Williams and the application team did not have access to an “as built” specification for the 40-year-old water tank. Drawings, maintenance documents, samplings of the concrete or a record of admixtures did not seem to exist. So, the team had to arrive at a plan based on what they could observe: a concrete substrate that remained saturated with moisture.

The Sherwin-Williams Technical Service team proposed coating the entire floor with Poly-Crete® MDB, a 100% solids, cementitious urethane cement that would block out the moisture in the concrete, before applying Dura-Plate 6100 on top. PPS installers applied the self-leveling Poly-Crete MDB at 3/16 of an inch for a thick moisture barrier.

As a precaution, the team proposed mock-ups (or test sections). These would be 10-foot by 10-foot sections marked off on the tank floor with tape. Applicators would apply Poly-Crete MDB to the entire square, followed by a broadcast of fine aggregate to enhance adhesion. Lastly, they would apply Dura-Plate 6100 over half the square. This setup would allow the team to test the adhesion of Poly-Crete MDB to the substrate, as well as the adhesion of the other layers. Most importantly, the coatings would be tested on the actual substrate – and in the actual environment where the system would need to perform (Figure 8).

The contractor, applicator, inspector and Sherwin-Williams explained the plan and rationale for the mock-ups to ASU, which appreciated the thoroughness of the approach. Officials remained concerned, especially about the timeline. The project, which commenced in October 2025, had been held up with lots of planning and consultation. Now, the early spring weather was getting warmer, making for a hot environment in the tank.

The first set of pull tests (ASTM D7234) from the mock-up showed insufficient adhesion between the Poly-Crete MDB and aggregate broadcast. Because of the high temperatures in the tank, the Poly-Crete MDB was curing faster than expected. The aggregate would need to be broadcast sooner.

“Perfecting the flooring system installation required an understanding of the coating properties and how the warm enclosed environment was affecting the material’s cure time,” said Cho. “Having the Sherwin-Williams Technical Service team involved in troubleshooting this challenge was instrumental to us finetuning the process for a successful application.”

Applicators prepared a second mock-up with the adjustment in timing. A second set of pull tests showed that the coatings were adhering at up to 200 psi, the desired effect. Now all parties were convinced they had a formula that would ensure adhesion, while completely sealing the tank. In effect, they would be able to create a tank within a tank.

“Going through the trouble of preparing mock-ups was a necessity for this project, and preparing those on the actual tank substrate was the ideal way for us to confirm how the state of the substrate and the site application conditions could affect adhesion and curing,” said Merigan. “This enabled ASU to witness how the system would perform and gain confidence in the approach and outcome.”

This approach – to create a tank within a tank – carries an important form of insurance. It can be tested for pinholes or other “holidays” and finally verified as a continuous lining using high-voltage holiday detection in accordance with AMPP SP0188 and ASTM D4787 testing standards. The same tests cannot be performed on saturants or treatments that soak into the concrete.

Figure 1. The failed tank lining was able to be peeled off in sheets due to adhesion failures.

Figure 2. Entry to the tank was limited to one of two pill boxes on either end of the tank.

Figure 3. Blistering was evident throughout the failed tank lining.

Figure 4. Applicators from PPS removed the entire previously failed lining system by peeling delaminated material by hand and using cleaning tools to remove the rest.

Figure 5. Isolated areas of damage on the tank’s ceiling required resurfacing with Dura-Plate 2300.

Figure 6. PPS blasted the concrete to achieve a uniform texture of CSP 3-5 on the surface of the walls, ceiling and floor.

Figure 7. Plastic sheeting tests on the floor showed continued presence of moisture.

Figure 8. Mock-ups for the tank’s new floor lining consisted of 10-foot by 10-foot testing areas. Applicators coated the entire area in Poly-Crete MDB before broadcasting an aggregate over the area. Then, they coated half the area in Dura-Plate 6100. For the inspection pull tests, dollies were installed on both halves of the mock-up.

Figure 9. Applicators from PPS raked and backrolled the self-leveling Poly-Crete MDB material on the tank’s floor.

Figure 10. The completed lining provides a monolithic surface across the entire tank to prevent water from working behind the coating and causing the same blistering and delamination that occurred with the previous lining.

Figure 11. PPS was able to repair pinholes (circled on the columns here) following high-voltage holiday detection testing to ensure a monolithic lining surface.