The corrosion of steel or other metallic materials under thermal insulation initiates due to the presence of water, oxygen and corrosive atmospheric contaminants such as SO₂ (in the form of H₂SO₄) and chlorides (See Figure 1). The main contribution of the thermal insulation of the corrosion process is to provide an annular space for retention and accumulation of water from rainfall and/or condensation. The insulation may also wick or absorb moisture further prolonging the periods of wetness. In some cases, it may also contribute additional corrosive species (such as chlorides and sulfur species) resulting from it chemical constituents.

An additional factor essential for CUI is the temperature. In hot applications, the metal in close proximity with the insulation is a heat transfer surface. It exhibits increased corrosion rate with increasing temperature. One problem in studying CUI is that, as shown in Figure 2, actual in-plant incidents of CUI have typically reported corrosion rates that were higher than those observed in aqueous corrosion tests conducted in either open or closed systems. [1]

CUI has plagued the chemical and petroleum industries for many years. As far back as 1971 an ASTM Committee C.16 on thermal insulation materials issued a standard for testing insulation and its potential to course chloride stress corrosion cracking of stainless steel alloys. (ASTM C 692)

CUI plant problems were also being discussed in various NACE technical committee forums. In 1983, a joint group activity held an international symposium on CUI. The industry was now awakening to the fact that CUI was wide spread and causing problems, and action was required.

A NACE Technical Probates Committee in the Chemical Process was formed to become an open forum for CUI problems and solutions. This committee T5-A30 later designated a task group that brought together various corrosion control board inspections on technologies that were being used so that they could be available to all plant operators.

This standard recommended practice RP0198-98 – "The Control of Corrosion Under Thermal Insulation and Fire-Proofing Materials – A System Approach." was published in February 1998. Unlike typical standards, this document brings together current consensus information about the Corrosion Mechanism Under Insulation, Insulation Design and Materials Selection, Protective Coolings, and Insulation Inspection and Maintenance.

The two main themes that run through the Standard are: (1) the corrosion culprit is water – all other factors only affect the rate at which the corrosion occurs, (2) the only true prevention action against CUI is to apply a high quality protective cooling to the metal.

The special test cell (CUI-Cell) shown in Figures 3a and 3b was developed by InterCorr which simulates many of the aspects common to in-service CUI. The cell consists of segmented ring specimens machined from a nominal 2-inch diameter pipe section. These specimens are electrically isolated using non-conductive spacers to allow them to be used as both corrosion coupons and for performing real-time electrochemical monitoring for determination of instantaneous corrosion rates versus time. Typically, the CUI-cell is divided into two cells separated by a dam so that independent measurements of baseline conditions and various modified conditions (e.g. surface treatments, thermal insulation, inhibitors, coatings) can be evaluated simultaneously. These procedures are currently under standardization by ASTM Committee G1.11.09. Practice A uses mass loss procedures for assessment of corrosion rates while Practice B uses electrochemical techniques for assessment of instantaneous corrosion rates during the tests. It also includes provisions for conducting laboratory simulation of CUI exposure conditions using isothermal, cyclic temperature and wet/dry .

Typical mass loss corrosion rate data from CUI-Cell tests conducted in 100 ppm chloride solutions at pH 6.0 adjusted with sulfuric acid indicate corrosion rates that are much higher than those from conventional aqueous beaker tests and more in line with CUI service data. For bare steel under isothermal conditions at 150 F, the corrosion rate was 79 mpy. Bare steel and pre-rusted steel under cyclic temperature at 150 F / 250 F had mass loss corrosion rates of 137 mpy and 77.2 mpy, respectively. During initial trials, specimens coated with a new proprietary treatment produced by Elisha Technologies showed mass loss corrosion rates that were 20 to 25 percent of that determined for non-protected steel in the CUI-Cell.

Electrochemical data from the same tests discussed above are shown in Figures 4 and 5. During isothermal exposure at 150 F, both the corrosion rates for bare and treated steel decreased with time reaching a steady state value after about 24 hours. However, during the cyclic temperature, wet/dry tests, two important physical aspects of CUI were realized:

1. The corrosion rate of steel under CUI conditions increased with each wet/dry cycle during the 72 hours period.
2. The corrosion rate reaches peak levels just prior to drying and again immediately following re-hydration.

The increasing corrosion rate with time is most likely related to the build-up of corrodants on the exposed metal surface with time due to evaporation. The spikes in the corrosion rate with time before and after drying also appear to relate to the increase in concentration of corrodants during drying. However, it may also be exacerbated by cracks that develop in the corrosion produces during the drying process. These cracks act as a path for the corrosive environment to make better contact with the metal surface particularly during rehydration when the water is added after drying. Neither of these phenomena are involved in conventional beaker tests and are why CUI can not be easily simulated with those types of tests.

The cyclic temperature wet/dry data in Figure 5 also shows a completely different behavior for the treated specimens than described above for the non-treated steel. The treated specimens show a decreasing corrosion rate with each wet/dry cycle. This indicates a surface treatment that is becoming more effective with time versus the non-treated steel where the corrosion rate increases with time.

Based on the initial trials described above, a more extensive, large scale and longer term evaluation was conducted to evaluate the performance of recently developed surface treatments under accelerated and severe CUI conditions for an extended period of exposure. The treatment was applied to ring specimens in a CUI-Cell and two large pipe sections and the test conditions included cyclic temperature and multiple wet/dry periods using an aggressive test solution consisting of 5 percent NaCl acidified to pH 6 with sulfuric acid. Following completion of the exposure, the coated and non-coated surfaces were examined in two large areas having different initial pipe surface conditions: (1) a pre-rusted surface and (2) bare steel (glass bead blasted).

These new surface treatments have been developed to utilize an innovative new process to mineralize the metal surface. Mineralization is the ability to grow a very thin mineral on metal surfaces for useful purposes, including protection against corrosion. These engineered surfaces are formed when mineral-forming compounds are delivered to the surface of a metal substrate. As shown schematically in Figure 6, the substrate material contributes donor ions to react and/or interact with delivered compounds, forming a very thin layer of mineral that is chemically bonded to the surface of the substrate and measured to be only a few monolayers thick [2]. Patent [3] and patent applications cover the delivery of novel formulations was well as the mineralized layer that is formed as a result of the interactions between the applied coating and the underlying substrate. The presence and uniqueness of the mineralized layer has been confirmed by conventional analytical surface methods such as X-ray photoelectron spectroscopy (XPS).

Figure 7 shows the test apparatus used to perform the large scale CUI tests. It involves two test sections of nominal 2-inch O.D. pipe used for the large scale evaluation. Each pipe section had approximately one foot lengths of (1) abrasive cleaning followed by epoxy coating, (2) surface treated steel and (3) bare steel. Each region was separated by an inert tape on the O.D. surface of the pipe. Cotton gauze was placed around the test sections of the pipe specimens to more uniformly wick the test solution.

The proposed ASTM CUI-Cell was placed between the two pipe sections. The CUI test cell was used to monitor corrosivity of the CUI conditions on carbon steel and to obtain an indication of relative performance of two different coating formulations. Intentional defects were placed on the exposed surfaces of the in the coated specimens to concentrate the corrosive attack and to evaluate the response of the coating materials to severe service induced damage. Cotton string was placed in the intentionally defected areas of the coated coupons and around the bare steel coupons to wick the test solution on to the surface.

The test solution was chosen to accelerate corrosive attack and most likely represent conditions that would be more severe than characteristically found in service. The test solution consisted of 1000 ppm solution of NaCl in distilled water acidified to pH 6.0 with sulfuric acid (H₂SO₄). This solution was pumped into the annular space between the insulation and the pipe O.D. The test temperature was 150 F as measured on the O.D. surface of the pipe. To simulate wet/dry conditions, the pumping of solution was stopped and the temperature elevated to 230 F and the pipe and CUI-Cell allowed to dry. The point of dry conditions was assessed using the LPR capabilities of the CUI-Cell. Dry conditions were considered to occur at the point were the corrosion rate of the bare steel coupons gave an instantaneous corrosion rate of zero during the drying cycle.

Figure 8 shows the LPR corrosion rate data versus time for the total test duration which was just short of 1000 hours. Three curves are indicated: (1) bare steel, (2) surface treatment #1 and (3) surface treatment #2.

The corrosion rate of bare steel in the CUI tests ranged from approximately 5 mpy to 28 mpy for the wet cycle with transient corrosion rates during wet/dry cycling ranging from 0 to about 55 mpy. The trend in the data for the non-treated steel with time showed an increase in corrosion with each wet/dry cycle characteristic of CUI conditions. By comparison, the corrosion rates in the defected areas for the coupons treated with both surface treatments were significantly less than that of the bare steel. The corrosion rates for the two treatments during the wet cycle were between 0.25 to about 5.0 mpy with transient corrosion rates of 0 to 7.5 mpy during the drying cycle. These corrosion rates should be considered as worst case values due to the presence of intentional defects in the coatings and the severity of the test solution. Based on the LPR corrosion rates the protection efficiency of the two coatings near the completion of the test were 82 percent and 89 percent , respectively.

The mass loss corrosion rates for the same three material conditions obtained from the CUI-Test Cell working electrodes were as shown in Table 1.

Table 1 - Mass Loss Corrosion Data from CUI-Cell  Top

The average mass loss corrosion rate for the bare steel coupon over the exposure period was 20.9 mpy. This was in good agreement with the LPR corrosion rates taking into account the natural change in corrosivity of the CUI conditions with exposure time. Corrosion rates for the two surface treatments were 4.5 mpy and 2.9 which were significantly lower than that for the bare steel coupons and comparable to the corrosion rates determined by LPR measurements. Consequently, the protection efficiencies for the coatings shown above were also similar to those determined from the LPR data. As in the case of the LPR data, the corrosion rates presented should be considered worst case values due to the presence of intentional coating defects and the severity of the test solution. Visual examination of the coupons indicated that corrosion was essentially limited to the regions were the coatings were removed that contained the wicks. Areas of intact coating on the coupons showed no signs of corrosion.

Following approximately 950 hours of cumulative CUI exposure, the pipe samples were removed from test and examined for deterioration. A detailed photographic record was made of the tests on the pipe sections and the performance of the surface treatments. Based on the visual examination, the following observations were made:

The layer produced on the O.D. surface of the pipe by the treatment did not change or deteriorate except for one region where local boiling of the water had disbonded the coating from the relatively smooth surface of the bead blasted metal. Subsequent tests conducted in boiling water showed this same effect on smooth bead blasted specimens. However, specimens given a rougher surface by abrasive blasting showed no disbonding of the coatings under the same conditions. By comparison, the pipe surface tested with the intentionally pre-rusted condition showed no evidence of disbonding treated layer at the bottom of the pipe. The coatings had excellent adhesion to the rusted metal surface which had been conditioned only by wire brushing prior to the treatment.

The portion of the pipe that was abrasive blasted and coated with a conventionally used epoxy coating was in good condition. However, the coating was starting to disbond and peel along the edge of the coating adjacent to a scribe mark placed in the coating prior to testing and rusting was observed on the metal in the scribe mark. It is a normal deterioration mode for epoxy coatings which will tend to disbond at local defects or imperfections in the coating. By comparison, defected areas in the coating, after treated showed no evidence of disbonding.

Following exposure, removal of the treated layer on the bead blasted pipe at locations in the middle and ends of the treated section revealed complete protection. No localized corrosion or under coating attack was noted in the region covered with the treated layer. Furthermore, removal of the layer on the pre-rusted surface also showed no change in appearance of the pre-rusted surface when compared to the original rusted surface.

Based on the results of the studies presented herein, it was observed that laboratory simulation of CUI is possible. The CUI-Cell simulation procedures resulted in similar corrosion morphology and corrosion rates on steel specimens as observed in piping exhibiting CUI in actual plant and field settings. Typically, corrosion rates of steel increase with each wet/dry cycle which is an important aspect of the severity and circumstances surrounding CUI.

Innovative, new surface treatments resulted in a significant reduction in the corrosive attack under accelerated CUI conditions. Worst case steady state corrosion rates in the defected areas on coated specimens were less than 5 mpy. Protection efficiencies for these treatments were between 78 and 89 percent using bare steel as the baseline condition. Based on a nominal 0.1 inch corrosion allowance this would result in an expected surface life of about 20 years under CUI conditions versus about 3 years for the untreated bare steel. The treatment showed good performance over large surface areas of steel pipe (both bead blasted and pre-rusted) when properly applied. Coupons coated with an epoxy coating were also found to be protected when in contact with the treatment material, showing no tendency to disbond around defects and imperfections as it would in the non-treated condition.

1. D. Abayarathna, W.G. Ashbaugh, R.D. Kane, N.G McGowen and R.L. Heimann, "Measurement of Corrosion Under Insulation and Effectiveness of Protective Coatings", Corrosion/97, Paper No. 266, NACE International, Houston, Texas March 1997.
2. J.J. Hahn, N.G. McGowen and R.L. Heimann, "Modification and Characterization of Mineralization Surface for Corrosion Protection", Surface and Coatings Technology, Submitted for Publication, 1998
3. Heimann, R. H. Corrosion Resistant Buffers for Metal Systems. U.S. Patent 5,714,093, 1998

Figure 1 – Example of corrosion under insulation in a chemical plant.

Figure 2 – Actual CUI corrosion rates (open circles) versus those in typical laboratory corrosion tests (closed pressurized system – solid line; open atmospheric system – dashed line).

Figure 3a - Schematic of proposed ASTM CUI-Test Cell.

Figure 3b – Photo of CUI-Cell in operation.

Figure 5 - Electrochemical corrosion rate data versus time for cyclic temperature, wet/dry CUI test.

Figure 6 – Schematic of new Elisha EDC surface mineralization treatments.

Figure 7 – Large scale CUI test cell used for extended duration tests.

Figure 8 – Electrochemical corrosion rate data for long term CUI test.