2. PRIMARY CASE STUDIES

2.1 I-90 Near Spearfish, South Dakota (SD-090-019)

Project Description

The South Dakota Department of Transportation (SDDOT) provided this project as a pavement with a history of durability problems. In particular, the pavement is experiencing surface or map cracking over the entire pavement surface. This project represents the primary case study site in the dry-freeze climatic region. The area receives approximately 400 mm of precipitation each year and has a freezing index of 684 °C-days.

This 14-km-long project is located on I-90 near Spearfish, South Dakota and extends from milepost 19.8 to milepost 28.5 in both directions. Table 3-2 provides a summary of the specific design features for this project.

Table 3-2. Summary of design features for SD-090-019.

The project is a four-lane divided highway, and the same design was placed in both the eastbound and westbound lanes. It is a jointed reinforced concrete pavement (JRCP) containing wire mesh reinforcement. The joints are spaced at 12.3-m intervals and contain 25-mm dowel bars. The longitudinal centerline joint is sealed with a hot-pour asphalt sealant, whereas the transverse joints are sealed with a silicone sealant. The pavement structure consists of a 200-mm JRCP, a 75-mm lime-treated gravel base, and a 150-mm lime-treated subgrade. The subgrade is a red clayey soil. There are no provisions for subsurface drainage. The inside and outside shoulders are AC-surfaced and are 1.2 and 3.0 m wide, respectively.

Field Evaluation

After an initial investigation, the survey team selected two sections-one each in the inside and outside traffic lanes-to be surveyed in order to evaluate the differences between lanes. Section 001 is located in the eastbound, inside traffic lane beginning at milepost 23.1. This section was constructed in a cut section of approximately 10 m. Section 002 is also located in the eastbound direction but in the outside traffic lane and begins at milepost 24.5. This section was constructed on approximately 3 m of fill material.

A summary of the distress survey results is provided in tables 3-3 and 3-4 for Sections 001 and 002, respectively. Overall, Section 001 appears to be in better structural condition than Section 002, as would be expected due to the lower traffic volumes on the inside traffic lane. Section 001 contains some low-severity transverse cracks, which are expected to occur on JRCP. Spalling occurs at 6 of the 14 transverse joints but only 1 has progressed to medium severity. The spalling appears to be due to the progression of MRD. Three small rigid patches, each of which is located along a transverse joint, are also present. Faulting is virtually nonexistent, averaging 0.6 and 0.7 mm measured at distances of 0.30 and 0.75 m from the slab edge.

Section 002 exhibits more distress and greater deterioration. Although low-severity transverse cracks are expected on JRCP, there are considerably more cracks as compared to Section 001. In addition, Section 002 also exhibits two medium-severity cracks and two high-severity cracks. Faulting is also much more significant, averaging 3.7 and 4.2 mm at 0.30 and 0.75 m from the slab edge. Another distress that is more significant on Section 002 is patching. A high-severity flexible patch is observed, as are 10 low-severity and 2 moderate-severity rigid patches. Some of the rigid patches are full-depth patches at transverse joints, indicating that the joints were likely badly deteriorated at one time. Five of the 14 transverse joints exhibit spalling, including 2 that have progressed to moderate severity.

MRD Field Characterization

During the field surveys, the attributes of the MRD were characterized. Although a definitive diagnosis cannot be made in the field, it is important to evaluate the attributes of the MRD as well as the effect these distresses have on pavement performance. A summary of the MRD characterization for both sections is provided in table 3-5. Figures 3-2 through 3-4 show some the typical conditions observed on the two test sections.

Table 3-3. Summary of pavement condition surveys for SD-090-019-001.

Table 3-4. Summary of pavement condition surveys for SD-090-019-002.

Table 3-5. Summary of MRD characterization for SD-090-019.

(a) Section 002	(b) Section 002

Figure 3-2. Typical distress manifestation observed on SD-090-019-002.

Figure 3-3. Typical distress manifestation observed on SD-090-019, Sections 1 and 2.

Figure 3-4. Typical distress manifestation observed on SD-090-019-002.

On both pavement sections, map cracking was observed throughout the entire area. On Section 001, the cracks appear to be confined to the upper 50 mm at the pavement surface. The majority of cracks run perpendicular to the centerline, but there are some cracks that run parallel to the centerline. The combination of cracks forms a criss-cross pattern on the surface. Although the cracking pattern is similar on Section 002, the transverse cracks on Section 001 are more pronounced and some are opened at the surface.

On Section 001, the area around the joints is discolored, showing a brownish-gray staining. However, the cracking pattern around the joints is similar to the slab interior. The MRD has progressed at a few of the slab corners and spalling has occurred. There is no exudate from the cracks on this section.

Section 002 exhibits a different cracking pattern along the joints. The cracking and staining form a semi-circular pattern, widening at the slab corners. A dark gray staining is observed around both the longitudinal and transverse joints. Unlike Section 001, exudate is observed at some cracks, particularly cracks located near a joint. The exudate is either a dark gray or white substance.

Laboratory Analysis

Core Selection/Visual Inspection

Based upon the field survey, distress was detected at joints and near slab corners. Photos of typical distresses are shown in figure 3-5. To look at concrete from more than one slab, Cores B and D were selected from Section 001 and Cores A, B, and C were selected from Section 002. All cores were cut to produce slabs for examination with stains.

Figure 3-5. Core specimens from SD-090-019.

Mix proportions were estimated by inspecting the cores visually before and after slicing. In this case, detailed construction records were unavailable to verify the mix design. The concrete was well consolidated with no apparent segregation or parallelism of the aggregates. No scaling or sub-parallel cracking was apparent on these sites. The embedded steel was at a sufficient depth to prevent corrosion and no entrapped water voids were seen under aggregates or embedded steel. Surface cracking was apparent that was not related to plastic shrinkage cracking.

Stereo Optical Microscopy

The stereo optical microscope was used to first examine polished slabs cut from each core to assess the general condition of the concrete. Typical micrographs of interesting features are presented in figures 3-6 and 3-7. The aggregate type was determined to be a natural gravel with a varied lithology including limestone, siltstone, dolomite, and rhyolite as the main rock types for the coarse aggregate. Many of the rhyolite particles had small feldspar inclusions. The fine aggregates contained the same rock types seen in the coarse aggregate in addition to shale, sandstone quartzite, and granite. Cracks passing through the paste also passed through aggregates. Reaction rims were visible along with secondary infilling in cracks and air voids. A yellow to white "soft" crumbly siltstone constituent of the coarse aggregate natural gravel is frequently cracked, with the cracks extending into the surrounding cement paste, and occasional white deposits in cracks. Aggregate particles that were volcanics or rhyolites appear to be reactive.

Figure 3-6. Stereo optical micrographs of typical cracking pattern associated with porous siltstone aggregate SD-090-019.

Figure 3-7. Stereo optical micrograph showing gel deposits in SD-090-019 aggregates.

The stereo microscope was also used to perform a modified point count in accordance with ASTM C 457. As part of the modified point count, the volume fractions of paste and aggregate were also determined to confirm mix volumetrics. The results of this analysis are given in table 3-6.

Table 3-6. Results of ASTM C 457 for concrete from SD-090-019.

The sodium cobaltinitrite/rhodamine B staining tests were applied and a number of aggregates were identified as being susceptible to alkali-silica reactivity (ASR). The phenolphthalein staining method was used to determine the depth of carbonation on freshly cut surfaces. Slabs cut from the analyzed cores were tested for depth of carbonation with no core having a depth of carbonation greater than 2 mm below the road surface. Barium chloride/potassium permanganate stain was used to identify sulfate minerals. Examples of the stained slabs are presented in figures 3-8 through 3-11.

Figure 3-8. Slab 1B stained with sodium cobaltinitrite/rhodamine B from SD-090-019-001.

Figure 3-9. Slab 1B stained with sodium cobaltinitrite/rhodamine B from SD-090-019-001.

Figure 3-10. Slab 2B stained with sodium cobaltinitrite/rhodamine B from SD-090-019-002.

Figure 3-11. Stereo optical micrographs of air voids filled with sulfate minerals stained with potassium permanganate (note differences due to polishing).

Petrographic Optical Microscopy

Based upon stereo microscope observations and staining, thin sections were prepared from the selected cores. Surfaces were sectioned from the core adjacent to stained sections to avoid contamination from the stains. The reactive coarse aggregates were primarily the siltstones and rhyolites, although others were noted as reactive. The shale was commonly associated with ASR in fine aggregate. In addition to cracking associated with ASR, other cracking of non-reacted siltstone aggregates was noted. The siltstone aggregates had a very porous microstructure as seen in thin section. These aggregates may be susceptible to aggregate freeze-thaw deterioration, leading to some of the cracking seen in the concrete. In addition to possible ASR and aggregate freeze-thaw deterioration, evidence of alkali-carbonate reactivity (ACR) was noted where densified paste regions or "halos" with a large amount of calcite were seen surrounding dolomite coarse aggregates (Spry et al. 1996). Secondary deposits within cracks and voids were identified. In addition to specific phases identified (e.g., ASR gel, calcite), ettringite was common as a secondary deposit. In addition to these diagnostic features, hydrocalumite (Friedel's salt) secondary deposits were found. Given the high chloride concentration needed to precipitate hydrocalumite, this is taken as a diagnostic feature of deicer attack. Petrographic micrographs are presented in figures 3-12 and 3-13.

Scanning Electron Microscopy (SEM)

A conventional SEM was used to identify secondary deposits seen in the petrographic microscope examination to confirm those observations. Figure 3-14 containsa backscattered electron image showing an ettringite filled void, a crack filled with hydrocalumite, and characteristic x-ray spectra from each phase illustrating their compositions. The SEM analysis confirmed the petrographic analysis with regards to the composition of the secondary deposits. The phase identified as hydrocalumite was confirmed, as were the presence of ettringite and the composition of various ASR reaction products. The results of x-ray microanalyses of the ettringite and the hyrocalumite phases are presented in tables 3-7 and 3-8, respectively. Figure 3-15 presents the ternary diagram showing the probable range of composition for the hydrocalumite deposits.

Figure 3-14. Ettringite (a) and hydrocalumite (b) infilling in void and crack, respectively. Example spectra from each phase are shown in (c) and (d), respectively.

Table 3-7. Summary of 10 analyses from ettringite deposits, compared to a calculated composition for dehydrated ettringite.

Table 3-8. Summary of 13 analyses from hydrocalumite deposits, compared to calculated compositions for the 3 dehydrated end members of the hydrocalumite solid solution series.

Figure 3-15. Ternary diagram showing the probable range of composition for the hydrocalumite deposits analyzed from SD-090-019.

Chemical Laboratory Tests

Ion chromatography was used to analyze the sulfate content of soil samples taken from the grade below the individual core holes. The complete analysis is presented in the final report. To summarize, the soil base below the test sites would be classified as a negligible sulfate exposure using the criteria set forth in ACI 201.2R-92 Guides to Durable Concrete.

Interpretation and Diagnosis

Having performed the described laboratory analyses and applied the diagnostic flowcharts as shown in figures 3-16 through 3-20, several possible MRDs were identified in SD-090-019-001, including ASR, ACR, aggregate freeze-thaw, and deicer attack. This is consistent with the visual observations of the distress reported from the field where mixtures of diagnostic features were apparent. To finalize the diagnosis, the diagnostic tables were consulted. The diagnostic features identified in the analysis processes are listed below in table 3-9 along with their associated MRD type and significance as related to this pavement. A brief discussion follows of each possible MRD identified in the laboratory analysis:

ASR - This MRD seems to be the most dominant given its extent in the sections sampled. From the standpoint of the guidelines, all diagnostic features of ASR were present with the exception of known poor performance for the aggregate used.

ACR - This MRD was identified as a possible but in the final analysis is not listed as probable as a major contributor. Although there was strong evidence of the reactivity of some dolomite aggregates, the extent and magnitude of this reaction was not great.

Aggregate Freeze-Thaw - Like ASR, this appeared to be a dominant distress in terms of extent. The likelihood or certainty of diagnosis is also very high given that, with the exception of known poor performance for the aggregate used, 75 percent of all diagnostic features for aggregate freeze-thaw were present.

Deicer Attack - This MRD is probably the most difficult to diagnose as it can often be present and hidden by other MRDs. The key diagnostic feature that makes deicer attack probable is the occurrence of hydrocalumite as an infilling material in cracks and voids.

As stated previously, it is not rare to find a pavement with diagnostic features representative of more than one distress mechanism present. In most of these cases, as with this one, the failure of the concrete cannot be attributed to one particular cause. However, in this case some general observations can be made. First, the ASR, aggregate freeze-thaw, and potential ACR distresses may not have occurred if a higher quality aggregate source was used. As is most often the case, contractors use the best possible aggregate source economically feasible but in some locations, such as central South Dakota, the possibilities are limited. The other distress mechanism identified, deicer attack, is more problematic as deicers are clearly required on this portion of the interstate system. A lower water-to-cement ratio (w/c) would likely reduce the concrete permeability and thus reduce the likelihood of a recurrence of this distress.

Recommended Treatment/Rehabilitation Alternatives

Using the procedures presented in Guideline III in Volume 2: Guidelines Description and Uses, feasible treatment and rehabilitation alternatives were selected. The two most significant MRD mechanisms found were aggregate freeze-thaw deterioration and ASR. Because the two mechanisms are acting in concert, it is difficult to rate the severity of each independently, but the level of spalling and patching at the transverse joints indicates that the severity level is likely a medium severity in Section 001 and medium to high in Section 002. The extent was at both joints and cracks and at corners. As a result, feasible treatment/rehabilitation alternatives include:

• Sealing less deteriorated joints and cracks.
• Partial- and full-depth repairs.
• The application of a high molecular weight methacrylate (HMWM).
• Overlay.

The use of patching is still feasible even though ASR was observed since most deterioration is isolated in the vicinity of joints and cracks. Further, lithium compounds are not suggested since they are ineffective in delaying aggregate freeze-thaw damage.

Ultimately, as the pavement continues to deteriorate, a reconstruction/recycling option becomes more viable. If recycling is considered, precautions must be taken to avoid aggregate freeze-thaw deterioration and/or ASR in the newly constructed pavement.

 

Figure 3-16. Flowchart for assessing the likelihood of MRD causing the observed distress in the pavement as applied to the Spearfish, South Dakota site.

Figure 3-17. Flowchart for assessing general concrete properties based on visual examination as applied to the Spearfish, South Dakota site.

Figure 3-18. Flowchart for assessing the condition of the concrete paste as applied to the Spearfish, South Dakota site.

• Heavy media separation.
• Blending with non-susceptible aggregates.
• Crushing to a non-susceptible size (potentially blending with larger non-susceptible aggregates).

• Heavy media separation.
• Blending with non-susceptible aggregates.
• Washing.

• Reducing the total mixture alkalinity to less than 3 kg/m³.
• Using a blended cement containing a pozzolan or ground slag.
• Through supplementation or addition, use pozzolans or ground slag in the mixture.
• Add lithium nitrate.