Skip to content

Preventing Crevice Corrosion in New and Existing Steel Structures

Steel structures are often fabricated by bolting (or riveting) structural components together. Bolted joints result in multiple crevices between the bolted members and fasteners. These crevices are typically more susceptible to corrosion than flat surfaces because they tend to retain water/contaminants, they are difficult to properly coat and crevice geometry tends to support electrochemical phenomena that accelerate corrosion. This article will evaluate the effectiveness of various coating practices at mitigating corrosion around these joints.

As seen in the January 2018 Edition of the Journal of Protective Coatings & Linings (JPCL) and on

For new structures, the state-of-the-art approach is to apply zinc-rich primer to mating surfaces prior to assembly and to use galvanized fasteners during assembly. Unfortunately, there are cases where this is not always possible. Obviously, it is typically unrealistic to disassemble, prime and re-assemble existing structures. During maintenance painting, stripe coats and caulking are often used to provide added protection to crevice areas. Though effective, caulking can be expensive and may not be necessary on all surfaces.

The authors will discuss the results of a laboratory study investigating the degree of crevice corrosion protection provided to bolted joints using different coating schemes. Forty different combinations of surface preparation, fastener coating, caulking extent and coating sequence were evaluated on aged and new steel assemblies. The test assemblies were exposed to cyclic accelerated corrosion testing and evaluated for rust staining, blistering and pitting within the crevice area. The data presented will help quantify the benefits of alternative approaches for corrosion mitigation in fabricated steel joints. 


Panel Design

Complex panels were assembled to replicate various geometries and include various materials that are seen on steel structures. A steel panel 6-inches-by-12-inches-by-1/8 inch had three 4-inch-by-2.5-inch-by-1/8-inch-steel plates fastened to them with ½-inch nuts and bolts (10 black-oxide and 10 galvanized).  A 4-inch-by-2-inch-by-1/8-inch steel angle was also bolted to the panel. The panels were made in such a way that they had both vertical and horizontal water travel paths and a representation of bolted bridge structural geometries that contain crevices. 

There were a total of 20 panels for this project. To compare the effects of new steel versus aged steel, half of the test panels were weathered prior to surface preparation and coating application. The following four surface preparation methods were evaluated in this project.

  • New Steel – Assembled then abrasive blasted per SSPC-SP 10/NACE No. 2, “Near-White Blast Cleaning.”
  • New Steel – As abrasive blasted per SSPC-SP 10/NACE No. 2 as individual parts, zinc-primed then assembled
  • Weathered Steel – As abrasive blasted per SSPC-SP 10/NACE No. 2 prior to coating.
  • Weathered Steel – Power tool cleaned per SP 11, “Bare Metal Power Tool Cleaning” prior to coating.

Table 1: Study Test Matrix.

Table 1 shows the test matrix for the study. The existing steel, new steel and caulking methods are described in the following sections.

Weathered Steel

Prior to exposure, panels were coated with a light layer of epoxy coating (1 to 2 mils) over bare steel with no surface profile. Panels were exposed for 150 hours of salt-spray exposure per ASTM B117, “Standard Practice for Operating Salt Spray (Fog) Apparatus” and nine months of outdoor exposure in an industrial setting. Figure 1 shows the panels after weathering but prior to surface preparation procedures.

After weathering, the panel surfaces were prepared for coating  by one of two methods. Most panels were abrasive blasted per SSPC-SP 10/NACE No. 2 using 30-grit aluminum oxide to achieve a nominal 1-to-2 mil profile. One set of weathered steel was prepared to SSPC-SP 11 using a combination of needle gun and angle grinder to achieve a nominal 1 mil profile. Figure 2 shows representative weathered steel panels prepared by each scenario.  

New Steel

The second set of the test panels represented new steel assemblies. Half of these panels were assembled prior to abrasive blasting and priming with a zinc-rich primer. The other half was abrasive blasted as individual components, primed using the same zinc-rich primer and then assembled. Figure 3 shows representative new steel panels after assembly and preparation. All abrasive blasting was SSPC SP-10/NACE No. 2 blast cleaned using 30-grit aluminum oxide to achieve a nominal 1-to-2 mil profile. The fasteners on the panels were primed prior to assembly hand-sanded prior to additional painting.

Fig. 1: Front (left) and back (right) of panels after weathering. All figures courtesy of the authors

Fig. 2: Surface preparation after weathering (SSPC-SP 10 left, SSPC-SP 11 right).

Fig. 3: New steel panels (assembled prior to blast, left; post-SSPC-SP 10 blast, center; and SSPC-SP 10 blast and primed, then assembled, right).

Coating Application

The coating system used for this project was a standard three-coat system currently used on structural bridges today. One panel from each scenario did not receive stripe coats and two of each surface preparation scenario did not receive caulking. The coating application sequence was as follows.

  • Perform designated surface preparation (e.g., SSPC-SP 10/NACE No. 2 or SSPC-SP).
  • SSPC-SP 1, Solvent Cleaning (using isopropyl alcohol).
  • Zinc primer (3 to 5 mils).
  • Zinc stripe (3 to 5 mils).
  • Intermediate coat (3 to 5 mils).
  • Intermediate stripe (3 to 5 mils).
  • Caulk application.
  • Finish stripe (3 to 5 mils).
  • Finish coat (3 to 5 mils).
  • Caulking Methods

Fig. 4: Caulking scenarios, side view (top method, top left; 3/4 method, top right; and full method, bottom).

Three panels from each surface preparation scenario had caulk applied to determine its effectiveness at preventing crevice corrosion. With each of the three panels, the caulk was applied in a different way to determine the amount needed to prevent crevice corrosion. Figure 4 illustrates the various caulking scenarios.

Corrosion Testing

After coating application and following a full seven-day cure, the panels were exposed to 120 cycles of GMW 14872 Cyclic Corrosion Laboratory Test (Fig. 5).

Inspections were carried out at 20, 40, 80 and 120 cycles. Each panel was inspected for rust through (using ASTM D610, “Standard Practice for Evaluating Degree of Rusting on Painted Steel Surfaces”), blistering (ASTM D714, “Standard Test Method for Evaluating Degree of Blistering of Paints”), crevice corrosion and the percentage of bolts with corrosion. Following 120 cycles of corrosion testing, the panels were disassembled to determine the extent of pitting in each crevice.


Figure 6 shows the new steel panels that were primed after assembly. Overall corrosion of these panels at each cycle was very minimal with the exception of rust bleed from the crevices in three scenarios. Crevice corrosion could be seen along the sides and top of the L-channel and built-up assembly throughout the exposure testing for three conditions: no stripe or caulk, stripe with no caulk, and stripe with top caulk.

Figure 7 shows the new steel panels that were primed before assembly. The black-oxide bolts on these panels showed the most corrosion of the conditions evaluated and the panel that received no stripe coating showed the greatest bolt corrosion. These bolts were only protected by the epoxy intermediate coat. The remaining systems received a stripe coat of zinc-rich primer on the black-oxide bolts but still experienced corrosion much earlier than any of the other types of surface preparation systems evaluated. Other than the black-oxide bolts, crevice corrosion was less evident on the panels that were primed before assembly.

Fig. 5 (left): One cycle of GMW 14872.
Fig. 6 (left): New steel primed after assembly. Fig. 7 (right): New steel primed before assembly.
Fig. 8 (left): Weathered steel, SSPC-SP 10 remediation. Fig. 9 (right): Weathered steel, SSPC-SP 11 remediation.

Figure 8 shows the abrasive-blasted panels and Figure 9 shows the power-tool-cleaned panels from the weathered steel scenarios. Not surprisingly, the abrasive-blasted repair system had less bolt and crevice corrosion than the power-tool-cleaned system. With both types of surface preparation, the best overall performing system was that of the full caulking; running rust was not evident after corrosion testing. Varying degrees of running rust can be seen from the crevices on the remaining panels. As expected, crevice corrosion was reduced with additional caulking.  Interestingly, stripe coating alone had a minimal benefit for reducing crevice corrosion.

Crevice Corrosion

The extent of crevice corrosion was noted at each inspection during the exposure testing. The top, sides and bottom of the three-coupon assembly, along with the top, bottom and sides of the angle, were inspected for evidence of running rust or rust through of the coating at these areas. Table 2 shows the extent of crevice corrosion after 120 cycles of exposure testing (red indicates that corrosion was present).

Table 2: Crevice Corrosion Inspection at 120 Cycles of GMW 14872. *Red indicates corrosion was present.

The only panels that did not display some form of crevice corrosion were those from the 3/4 and full-caulking scenarios. It was interesting to note that the new steel panels that were primed after assembly displayed less crevice corrosion than those that were primed before assembly though this may be influenced by running rust from the poorly coated black-oxide bolts.

Bolt Corrosion

To explore the effects of the coating techniques on galvanized and black-oxide bolts, the percentage of bolts experiencing corrosion was documented at each inspection cycle. Table 3 shows the percentage of black-oxide bolts displaying corrosion at each inspection. Table 4 shows the same data for the galvanized bolts. A comparison of these tables clearly shows the value of galvanized bolts in any scenario. 

Table 3: Corrosion of Black-Oxide Bolts on Panels.

The panels displaying the greatest bolt corrosion were panels primed prior to assembly. The black-oxide bolts that did not receive a zinc primer or zinc stripe coat displayed corrosion at the first inspection (cycle 20). Epoxy and polyurethane do not sufficiently protect the black-oxide fasteners for even a short period of time. If black-oxide bolts must be used, they should receive full and stripe coats of zinc primer, epoxy intermediate and finish coat.

Corrosion of the weathered black-oxide bolts was less evident on abrasive-blasted panels than on the power-tool-cleaned panels. This is not surprising, as existing rust in crevices was more thoroughly removed using abrasive blasting than power tools.

When comparing Tables 3 and 4, it is worth noting that the extent of corrosion on individual fasteners was less on the weathered galvanized bolts than the weathered black-oxide bolts. Even though galvanized bolts had significantly less corrosion than the black-oxide bolts, Table 4 shows the clear benefits of stripe coating the galvanized bolts, especially on the threaded end of the bolt.

Table 4: Percentage of Galvanized Bolts Displaying Corrosion per Panel.


After accelerated corrosion testing, panels were disassembled for analysis of corrosion in the crevices. Figure 10 shows each panel dissembled.

Figure 13 shows that 3/4 caulking applied to new steel that is primed after assembly is the best way to prevent corrosion in these crevices. In one of the two crevice areas, the full caulking appears to hold moisture within the crevice of the new steel panels. As expected, the panels with neither a stripe coat nor caulking experienced the most crevice corrosion. The remaining three scenarios (stripe coat, top caulk and full caulk) had visually similar crevice corrosion on the new steel assemblies.

Fig. 10: Panels disassembled after application.

For weathered steel under repair conditions, the benefits of a full caulk system are evident. The remaining caulking schemes visually appear better than the schemes without caulking.

Pit Depth

In an attempt to better quantify the crevice corrosion, pit depths were measured on the surface of the angles that were mounted onto the larger panel. After disassembling the panels, all corrosion products were removed from the crevice surface of the steel angle using abrasive glass-bead blasting. Figure 13 shows the crevice surface of an angle before and after glass bead blasting. 

Once the surface was cleaned, the ten deepest pits were measured using a pit- and crack-depth gauge. The instrument has a measuring range of 0 to 500 mils and a resolution of 0.5 mils. The minimum, average and maximum measurement are used for the analysis presented in this article. Figure 11 shows the data for the panels representing painting of new steel and Figure 12 shows the data for the panels representing maintenance painting of weathered steel.

Comparing the two graphs, it is clear that deeper pits were measured on the panels representing maintenance painting of weathered steel. However, a certain portion of that pitting occurred when the test panels were being weathered (i.e., pre-rusted). For any given panel, it is not possible to know how much of the pitting occurred prior to the maintenance painting. Analysis of the remaining crevice surfaces may provide sufficient statistical data to get additional insight, however this article focuses on the two data sets individually.

Fig. 11: Minimum, average and maximum pits measured on the crevice face of the angle for panels representing new steel coating systems.

Fig. 12: Minimum, average and maximum pits measured on the crevice face of the angle for panels representing weathered steel coating systems.

L – Channel after disassembly

L – Channel after blastFig. 13: L channel before and after glass-bead blasting.

For the new steel painting scenarios, the pit-depth data corroborates the visual observations (Fig. 12). Pitting is clearly reduced by striping and caulking. The caulking scenarios that left at least one edge un-caulked had the least amount of pitting. Priming prior to assembly was also beneficial in these scenarios. Interestingly, priming prior to assembly led to greater pitting than priming after assembly for the scenarios that did not have a stripe coat or were fully caulked. Standard industry practice of priming surfaces before assembly and incorporating stripe coats was among the best performing scenarios, though the practice would benefit by adding limited caulking to reduce water ingress.

For the weathered steel maintenance painting scenarios, the observed pitting is probably dominated by what occurred during the pre-weathering exposure (Table 4). However, by comparison to the “nothing” set of data, there is some evidence that the following four scenarios will result in less pitting.

  • Stripe and full caulk for abrasive-blasted panels.
  • Stripe and full caulk for power-tool-cleaned panels.
  • Stripe and 3/4 caulk for abrasive-blasted panels.
  • Stripe and top-caulk for power-tool-cleaned panels.

The benefit of all caulking scenarios is corroborated by the visual observations. The panel geometry contains 11 unique crevice surfaces. Analysis of the remaining crevice surfaces may provide sufficient statistical data to obtain additional insight.


The following conclusions were formed based on the results of the study.

Galvanized bolts perform better over time than do black-oxide bolts.

  • If black-oxide bolts are utilized, abrasive blasting, application of a zinc-rich primer, and stripe coats for each coating in the system will help prevent corrosion from occurring.
  • Stripe coats and caulking of crevices directly exposed to water or moisture will help prevent crevice corrosion on new steel.
  • When caulking newly applied steel, consider leaving the bottom crevice uncaulked to allow moisture to escape.

When working with weathered steel, full stripe coats and caulking of all crevices provide the best results in regards to reducing crevice corrosion and pitting.
As a best practice, mating steel surfaces should receive a primer coating prior to assembly.


Eric Shoyer has worked with Elzly Technology Corporation in various aspects of corrosion control and materials engineering for over eight years. He is an active member of several technical societies including SSPC and NACE International and is a registered NACE CIP Level 2. Shoyer holds Bachelor of Science degrees in civil engineering and structural engineering from Drexel University.

J. Peter Ault has been actively involved in various aspects of corrosion control and materials engineering for over 25 years. Since 2006, he has been a principal of Elzly Technology Corporation. Ault is an active member of several technical societies including SSPC, ASTM, NACE International, ASNE, SNAME and NPSE. He is a registered professional engineer in New York and New Jersey and holds coatings specialist certifications from both SSPC and NACE. Ault has a B.S. degree in mechanical engineering and an MBA from Drexel University.

Peter McDonagh is the manager of the Triborough Bridge and Tunnel Authority’s (MTA Bridges and Tunnels) Bridge Painting program. He has over 20 years of experience in various aspects of corrosion control, coatings inspection,  program management, and materials engineering. He is a registered NACE-Certified Coating Inspector.  Peter holds a Bachelor of Science degree in environmental science from the State University Of New York, College of Environmental Science and Forestry.

Brian Prazenka has worked with Triborough Bridge and Tunnel Authority (MTA Bridges and Tunnels) in various aspects of corrosion control and materials engineering for over four years, and in the engineering field for over eight years. He is a registered NACE CIP Level 2. Prazenka holds a Bachelor of Science degree in civil engineering from Manhattan College.


This material is based upon work supported by Triborough Bridge and Tunnel Authority (New York, N.Y.) and performed by Elzly Technology. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the Triborough Bridge and Tunnel Authority.

Leave a Reply

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