By Wendell Dubberke
Coarse aggregates used in Iowa concretes consist primarily of; (1) crushed carbonates (limestone and dolomite) from the paleozoic era. Quarries are located in Pennsylvanian, Mississippian, Devonian, Silurian, Ordovician and Cambrian age bedrock, (2) crushed quartzite (metamorphosed quartz sandstone) from quarries located in Wisconsin, South Dakota and Minnesota and (3) natural gravels containing various percentages of limestone, dolomite, quartz, granite, other igneous particles, chert (sedimentary silica), sandstone, shale and iron spal.
The Iowa sedimentary bedrocks are located in the Forest City basin which extends into the surrounding states. Most of the bedrock is covered by varying thicknesses of glacial till, loess, sand or gravel. However, some outcrops do exist and can be quarried directly. A few underground mines exist in Iowa, allowing quarrying operations with minimal disruption of the overburden. The oldest bedrock (Cambrian) is quarried in the northeastern part of Iowa. The quarried bedrock tends to be younger (Ordovician, Silurian, Devonian, Mississippian and Pennsylvanian) towards the southwest. The age of these sedimentary deposits ranges from approximately 300 to 500 million years. Some of the sedimentary rock layers are very fossiliferous, but the degree of fossilization has little relationship to rock durability for road construction purposes. Personnel in the Iowa DOT geology section maintain geologic sections for all bedrock quarries furnishing crushed stone for highway projects. The sedimentary beds are described and assigned bed numbers. Beds are combined to form ledges of varying quality. In general, the lowest quality ledges are used for roadstone. The next higher quality ledges are used for type B asphalt. The next higher quality ledges are used for type A asphalt. The highest quality ledges are used for portland cement concrete pavements (pccp). Many quarries do not have ledges meeting the higher quality standards. Quarry bed and ledge control allows for a finer degree of production quality control and is very helpful for historical record keeping purposes.
The Iowa sedimentary bedrock is generally interpreted as being: (1) All beds in the paleozoic essentially laid down flat (pancake geology) with minor disconformities and unconformities, (2) all beds then tilted towards the southwest and (3) all beds chopped off (eroded) to a semi-flat plane. An easier explanation would be that the Forest City basin was intermittently sinking with younger beds overlapping older beds. The need to force-fit, the oldest Mississippian and Devonian beds to the surface, could be eliminated by using the overlapping model. For example, in Franklin county, let: (1) the upper dolomite ledge be the Iowa Falls member, (2) the next limestone ledge be the Eagle City member, (3) the next dolomite ledge be the Mayne Creek member and (4) the next siltstone ledge be upper Devonian. Using this scenario, the older beds (Chapin, Mccraney, etc) disappear (were never there?) towards the edge of the basin.
Crushed quartzite has been used in Iowa concrete for over 60 years. Compared to crushed carbonates and gravel, the overall quantity used is small, but still significant. The service record, of pavements made with with quartzite coarse-aggregate is very good. Some of these concrete pavements, over 50 years old, are still in service and have not been resurfaced. Great stuff, dispite what some ASR activists are claiming.
Natural gravel deposits can be found in nearly all areas of the state, but the majority of the gravel pits are found in the northwestern part of Iowa where the quarrying of sedimentary rock is lacking. When Iowa gravels perform poorly in pccp, it usually relates to the non-igneous, non-quartzite fraction. In other words, the igneous and quartzite particles are always durable when used in Iowa pccp. Volcanic (rhyolite, etc) and intrusive (andesite, etc) particles are less often found in Iowa gravels. Gravel particles, containing some highly reactive cristobalite, were used in a pcc bridge deck in northwestern Iowa, causing a severe popout problem. Cristobalite is very unstable, and should not survive more than 100 million years. It is not understood as to why it occurred in this one location.
The majority of Iowa pccp joint blowups occur in concrete made with faulty gravel coarse aggregates. These gravels are faulty because the carbonate fraction is too argillaceous (shaly). Unfortunately, some quality test methods, used to evaluate gravels, allow the highly durable igneous and quartzite fractions to overshadow the faulty limestone and carbonate fractions. Pccp joint blowups occur more often in county (secondary road system)pccp. County officials are more apt to allow local, lesser quality aggregates to be used in their pccp to save money initially.
Thermogravimetric analysis (TGA) can be used to identify faulty carbonates in gravels. Some pccp designers are not aware that Los Angeles abrasion test results on Iowa gravels are quite often an average between extremely hard igneous and quartzite particles and very soft carbonate particles. If an aggregate producer wanted to blend poor quality carbonate aggregates with high quality igneous and quartzite aggregates, to meet specifications, it would never be allowed, but if this situation occurs in a natural gravel, it is acceptable as long as the combined components meet specifications.
Iowa gravels contain a considerable amount of medium/ fine-grained dolomite particles. Most likely, these dolomite particles originated in Ordovician age bedrock in Minnesota and were transported to Iowa by continental glaciation. After thousands of years, river erosion of glacial till left these dolomite particles (as well as other particle types) behind as gravel deposits. Some upland gravel deposits occur where streams wash over retreating continental glaciers. Logic would indicate that these dolomite particles should be durable in any environment after being transported hundreds of miles and being subjected to numerous freeze/thaw cycles over thousands of years. However, many of these fine/medium-grained dolomite particles are non-durable when used in pccp that is later subjected to deicing salts. Quarried, crushed fine/medium-grained dolomite can also suffer from this same problem. Coarse-grained dolomites, containing large diameter pores, perform well in pccp even when deicers are used. The Ordovician dolomite particles are over 435 million years old. Since larger crystals grow at the expense of smaller crystals, it seems strange that such a high percentage of Ordovician dolomite is fine/medium-grained. Over time, individual crystals push evenly distributed impurities to the out side edge. Is there a coating on the crystals and crystallites that inhibit crystal growth? Could these theoretical coatings contain manganese? XRF analyses shows these dolomites to contain extra manganese in the trace amount range. Are these crystal and crystallite surfaces more salt reactive? Do the relatively small crystal interlock areas become detached when brine enters the pore system? The silurian age dolomites are younger than Ordovician age dolomites, and yet, they are mostly coarse-grained and and consequently are very durable when used in Iowa pccp. It is possible the Silurian calcites were already coarse-grained before dolomitization.
When used in pccp, chert particles in Iowa gravels range from durable to extremely non-durable. Tripoli or tripolitic chert (white, fine-grained containing an extensive capillary sized pore system)is highly reactive when used in pcc. ASR and freeze/thaw damage are both associated with the use of tripolitic chert in pcc. Tripolitic chert (as well as shale and iron spal particles) cause popouts to occur on concrete surfaces. Too much tripolitic chert in a concrete mix can cause complete failure after hydration. Concrete, durable in all other respects, can tolerate a minor amount of tripolitic chert if surface aesthetics is not a problem. The Iowa DOT allows 3% chert (any kind) in crushed bedrock coarse-aggregate used in pcc mixes. Three percent of tripolitic chert in ASTM C666-B aggregate samples can cause pcc beams to fail this test method. To make ASTM C666-B test results match service record results, the tripolic chert is removed from the samples before testing. The poor correlation between ASTM C666-B results and pccp service life would even be worse if this was not done.
The chert found in Iowa natural gravels is mostly of the durable (dense, non-porous, brown or tan) variety. Apparently, transport, erosion and freeze/thaw action, taking place for thousands of years, has removed most of the tripolitic chert from Iowa gravels. Iowa DOT specifications also limit non-durable chert in gravels used in pcc mixes. Some local county officials want to use local gravel sources and consequently allow for more chert in a pcc mix. Some county pcc pavements, made with gravels containing over 20% brown chert, have performed well for over 20 years. ASR activists need to take note of these durable pcc pavements made with gravels containing a significant amount of sedimentary chert. Success analysis should be pursued just as vigorously as failure analysis.
It should also be noted that if these reactive, tripolitic chert, coarse-aggregate particles were pulverized to a size similar to flyash particles, they could be added to a concrete mix and be considered a beneficial pozzolan. The size of a particle can determine if it will be durable or non-durable in a concrete mix. A large number of extremely small silica particles can function as nucleation sites for the production of silica gel. Most concrete pore systems can easily handle the relatively small amount of silica gel generated around each very small particle. This evenly distributed silica gel is very efficient at mopping up excessive sodium and potassium (alkali) ions. When evaluating deteriorated concrete, the presence of silica gel does not automatically implicate expansive silica gel (and therefore faulty aggregates) as the primary culprit causing the deterioration. As with silica gel, the porous concrete can also handle moderate amounts of expansive ettringite without losing its integrity. The use of staining techniques to identify silica gel and/or ettringite in faulty concrete requires careful interpretation. The primary deterioration event may have little to do with the presence of excessive silica gel or ettringite. Often, the primary deterioration event is the generation of micro-shrinkage cracks, during early hydration, causing the concrete to lose its integrity, allowing moisture to more easily enter the concrete, enhancing the generation of ettringite and/or silica gel. Some cement formulations along with a low water mix are associated with the generation of micro-cracks in early hydrating concrete. The use of the ring test method is recommended for the identification of cements and mix designs susceptible to the generation of micro-cracks in concrete. More information, about shrinkage cracking and the ring test method, can be found in a 78 page report recently published by the American Concrete Institute (ACI) titled "Visible and Invisible cracking of Concrete" written by Richard Burrows. The ACI can be found on the internet.
Flyash spheres are primarily glassy silica. They should be ASR reactive, yet type F flyash is recommended for use in concrete where ASR problems are expected to occur. The relationship of reactive particle size to expansive energy in porous concrete probably explains why no individual Iowa sand sources could be related to faulty concrete whereas individual coarse-aggregate sources, producing non-durable pcc aggregates, were easily identified by the computerized pccp service record system.
The central Iowa sands contain a minor amount of shaly particles. Sands, used in concrete, must be under the 2% shale float test limit. Patio contractors do not like to use these central Iowa sands in pcc, as they can cause small popouts on the surface. A small number of popouts and stains are not a problem for highway construction. Some of these sands can fail the ASTM C 1260 ASR quick test, particularly when the pass/fail limits are lowered. These same sands pass the more stringent Canada C-14-A test which matches our service record information. Unfortunately, the C-14-A test takes one year to run, two years if flyash is used in the mix. C-14-A is now ASTM C1293.
When examining Iowa pcc pavement joint blowups, some of the fractured pieces will show partial oxidation and carbonation on the fractured surfaces, indicating the joint was already faulty prior to the blowup. In many cases, the use of non-durable coarse-aggregates in the pccp construction causes the problem. If incompressibles (sand, etc) in the joints are the cause, as many pccp experts contend, then why do these blowups occur, just as often, in curved sections of pcc pavements?
Since the 1930s, personnel at the Iowa Department of Transportation (IADOT) have maintained service record files on all portland cement concrete pavements (pccp). These files include all components used to construct the pccp. The pavements are continually monitored to determine service life. The service life is broken down into 10 year increments, 10 years or less, 10 to 20 years, 20 to 30 years and more than 30 years. Correlations between pccp service life and components are easily obtained, as all records have been computerized. These pccp service records can also be used to evaluate aggregate and concrete test methods. An evaluation of the data suggest the following: (1) Starting in the late 1970s, some pcc pavements, as young as 4 years, showed deterioration eventhough aggregates, with excellent service records, were used in their construction implicating non-aggregate mix components and/or faulty mix designs (too little water?), (2) There is no decent relationship between Iowa fine-aggregate (sand) sources and pccp service life. Either Iowa has all great sand sources (not true, I think) or the Iowa sand fraction in a pccp mix has less to do with early pccp deterioration than some researchers think, (3) There is good relationship between coarse-aggregate source and pccp service life. The pccp coarse aggregate classification system was modified in the early 1950s when the use of deicing salts began, (4) The ASTM C666-B test method, used to evaluate coarse aggregate in concrete, correlated poorly with field service records, (5) The pore index test method, developed at the Iowa DOT, has correlations to pccp service records similar to the ASTM C666-B durability test method. The pore index test method was developed with the idea of getting a quick, rough pore size distribution result, (6) The quality index number, which uses both physical parameters (pore index) and chemical parameters (XRF, XRD, TGA test results) correlated well to actual field pccp service life and (7) The sodium/magnesium sulfate test methods were evaluated during the 1940s. The test results correlated poorly with actual pccp service life. In the 1950s, with the advent of deicing salt usage, salt susceptible aggregates were identified. The sodium/magnesium sulfate test method was not re-evaluated to see if it could identify salt susceptible aggregates. Sodium chloride and calcium chloride are the most common deicing salts used. Could chloride salts be used in this test method instead of sulfate salts?
When evaluating new aggregate or pccp test methods, test results should always be correlated to actual field performance, not other, pre-existing test methods, as none of them have an acceptable correlation to actual field service life.
If an aggregate from a given source consistently fails when used in a variety of concrete mixes, there must be a physical and/or chemical parameter(s) that can be directly measured, using modern analytical equipment.
Great care needs to be exercised, when evaluating faulty concrete, to make absolutely sure that the problem is, in fact, initially aggregate related. Contrary to the opinions of many noted concrete investigators, much of the non-durable concrete, used in construction over the last 20 years, has little to do with the choice of aggregates used in the mix.
The following describes some of the aggregate or pcc test methods evaluated by the Iowa DOT. Some were implemented, others were rejected and some were replaced by test methods that are simpler, faster and more accurate.
This test is run on loose coarse-aggregate. It has has been used by the Iowa DOT for more than 50 years. It was discovered by accident, when some alcohol accidently contaminated the water used for thawing samples. A fixed weight of loose coarse-aggregate is subjected to 16 freeze/thaw cycles. The aggregate is frozen in dynamic air and thawed in water containing a small amount of methanol. The weight of the small particles, dislodged by freeze/thaw action, is determined by subtracting the weight of the washed coarse-aggregate (after 16 freeze/thaw cycles) from the initial washed weight. This weight loss, expressed in percentage of initial weight, is the test result. Excessive washing and screening of soft samples, to determine weight loss, can add material to the loss amount which relates to hardness rather than durability. Test results correlate with the amount of clay in carbonate samples and the amount of tripolitic chert in the sample. Carbonates containing too much clay (argillaceous) do not perform well in Iowa pcc when subjected to freeze/thaw cycles and deicing salts. Argillaceous concrete aggregates that do not perform well in northern states, can perform well when used in southern states concrete. In Iowa, carbonate aggregates, containing more than approximately 6% clay, are not allowed to be used in concrete.
XRF element analysis of carbonate aggregates is now being used for quality control. The amount of aluminum correlates with the amount of clay in the carbonate sample. However, there are a few locations where a very small amount of secondary feldspar can be identified in carbonate samples, which can upset direct correlations. If feldspathic carbonates relate to pccp durability is unknown. Western Iowa feldspathic sands have a contradictory service record when used in pccp. Some 40 year old pccp, using feldspathic sands, show warpage, but not general deterioration. Some newer, faulty pccp, using these same feldspatic sands, are failing due to the use of reactive (ASR) sands according to some ASR activists. The Igneous particles, in Iowa gravels, contain feldspar and are durable when used in pccp.
This test method is very similar to the alcohol freeze/thaw test method discussed above, except that it is run for 25 cycles and the aggregate is thawed in plain water. Only very poor quality aggregates will fail this test method. In Iowa, this test method is still used to evaluate aggregate for use in roadstone and type B asphalt.
During the late 1960s and early 1970s, intermediate dolomite coarse-aggregates were implicated as a faulty component in non-durable pccp. Carbonate coarse-aggregates were pulverized and analyzed using wet chemistry techniques. XRF analysis could have been used, but XRF equipment was not readily available for production control in the 1970s. Test results in the intermediate range, between limestone (calcite) and dolomite were considered suspect. Test results correlated only moderately well with field service records. The use of this test method was abandoned because too many aggregate sources with excellent service records were being implicated. One of the early, pcc-non-durable, carbonate, crushed-aggregate sources, evaluated by this test method, was the Kingston stone from Canada. This moderately shaly carbonate stone would have been eliminated, for use in Iowa pccp, by the alcohol freeze/thaw test method.
The correlation of intermediate dolomite test results to field service records may have been better if individual beds rather than ledges had been tested. Mechanical mixtures of durable limestone (calcite) beds and durable dolomite beds can produce test results in the intermediate dolomite range. Some of the working ledges in Iowa quarries are composed of durable dolomite beds and durable limestone beds, producing durable stone for use in pccp.
Using the backscattered electron (BSE) mode (sensitive to atomic weight) on the scanning electron microscope (SEM), dolomitization within individual carbonate crystals was easily observed, which means that possibly individual crystals could contain durable calcite and either durable or non-durable dolomite discussed in the next paragraph.
Another type of dolomitization came to light in the 1980s when Iowa DOT personnel began using x-ray diffraction (XRD), x-ray fluorescence (XRF) and thermogravimetric analysis (TGA) equipment to test all aggregates. XRD analysis revealed a consistent dolomite peak shift when testing dolomites associated with poor pccp field service records. The automated XRD peak search program indicated these shifted peak dolomites were in fact ankerite. At this time, a report was presented at the Transportation Research Board (TRB) meetings eventhough some questions remained. XRF (element analysis) results from split samples showing shifted dolomite peaks also showed iron content too low for the sample to be ankerite. In nearly all cases, the iron content in these shifted dolomite samples is less than 1.0%. A serendipitous encounter with a research report, involving the staged conversion of pure calcite to dolomite, under laboratory conditions, containing XRD results for each stage, fit well with XRD results from some of Iowa's natural dolomites. The research report showed that while the whole sample could be dolomitized, under some conditions, in approximaely 10% of the evenly spaced lattice positions, magnesium ions had not replaced calcium ions as would be expected in a pure dolomite. XRD analysis of this irregular substituted dolomite gave peak locations similar to ankerite eventhough the carbonate was free of iron, and therefore may suggest instability when used as aggregate in pccp. The suggested relationship of intermediate dolomite to faulty pccp may have been correct, when proposed back in the 1960s, but they should have used XRD results instead of wet chemistry results to identify the culprit. An additional observation, relating to these shifted peak, natural dolomites, is that they often contain minor amounts of pyrite inside the individual, solid, dolomite crystals.
One of the links on the home page is dedicated to the ASTM C666-B test method. Go there for more information. Some of the information here is a repeat of information on the other site.
This test is sometimes referred to as the durability factor test method. The Iowa DOT evaluated this test method in the early 1960s. 4" X 4" X 16" were fabricated rather than the more common 3" X 3" X 12" beams. A Conrad brand freezer/recorder unit was purchased in anticipation of a positive evaluation. Initially, coarse-aggregate was the only variable in the mixes used to fabricate the beams. The amount of water in the mix varied slightly to meet slump requirements.
ASTM C666-B specifies the use of dynamic air to freeze the concrete beams to zero degrees fahrenheit and the use of water, at fourty degrees fahrenheit, to thaw the beams over a 3 hour period. 300 freeze/thaw cycles constitute a complete test. Beams are removed from the freezer, prior to 300 cycles, if too much deterioration occurs. The beams are monitored twice a week for changes in length, weight and sonic modulus. A loss of more than 20% sonic modulus, in 300 cycles, is considered a failure. Beams are removed from the freezer if the sonic modulus dips below 60%. Three pcc beams are made from each aggregate sample submitted for testing. The average of the three beams is used for the final test result. The three beams are run simultaneously.
Initially, using the 14 day moist room cure (time between beam fabrication and insertion into the freezer cabinet) no beams could be made to fail this test method even when incorporating known poor performing aggregates in to the mix. Next, moist room cures of 45, 60 and 90 days were tried. ASTM C666-B test results, using the 90 day moist room cure, correlated best with field service records, but still left much to be desired. After 40 years, 3 replaced freeze/thaw units and over 3000 test results, the correlation to field service records is only 0.4r. Certainly not a "silver bullet" test method for evaluating Iowa aggregates for use in pcc.
Even with the use of the 90 day moist room cure, most dolomite coarse-aggregates pass this test regardless of their field pccp service records. Many field observations had convinced us that some dolomites are susceptible to sodium and calcium chloride deicing salt applications. Salt tracked onto secondary (county) pccp, from salted primary road pccp, could cause a 1/4 mile of deterioration on the secondary pccp. Pccp built on pugmilled calcium chloride treated base aggregates showed earlier than normal deterioration when salt reactive aggregates were used in the concrete mix. The C666-B test method was modified by soaking the coarse aggregate in NaCl brine prior to being incorporated into the mix. This modification worked well for identifying salt susceptible dolomites but did not work well for identifying salt susceptible limestones (calcites). Calcium chloride salt enhances the dedolomitization process, but was not tried in this brine soak pretreatment process.
The C666-B test method is very sensitive to the amount of air entrainment in the mix when fine-grained limestones are being evaluated. Some coarse-grained aggregates, containing a large diameter pore system, can easily pass this test with no air entrainment added to the mix, yet many concrete investigators are convinced that freeze/thaw durable concrete must contain a significant number of evenly distributed air bubbles in the pcc matrix. Initially, air entrainment of concrete was promoted as a way to reduce or eliminate surface scaling caused by deicing salt applications, which turns out to be correct. The use of air entrainment in pccp began in the early 1950s about the same time deicing salts were beginning to be used. In Iowa, there are many pcc highways built before 1950 that are still in service including some constructed with fine-grained limestone. These non-air-entrained pcc pavements are not suffering from any excessive freeze/thaw related damage. An argument could be made that air voids, in the concrete matrix, are needed to function as reservoirs for excessive silica gel and/or ettringite.
In the late 1970s, the ASTM C666-C test method should have been used to evaluate different cement brands, formulations and mix designs. Perhaps, with a dry cure instead of a moist room cure, those cements and mix designs that are prone to shrinkage micro-cracking, during initial hydration, could have been identified. There was some internal and external pressure to find fault with the aggregates and lay off of the cements.
Over the years, the ASTM C666-B test method was used to evaluate concrete containing additives such as flyash, silica fume, blast furnace slag & etc. It was also used to evaluate variable amounts of air entrainment and water in the mix as well as a wide variety of other mix design options. There are too many in-house studies, involving this test method, to be included here. A summary of all ASTM C666-B test results, with mix parameters, was available for public perusal and still may be available from the Iowa DOT.
Iowa DOT personnel evaluated these sulfate test methods in the 1940s. The test method was not adopted because of: (1) poor correlations between test results and pccp field service records and (2) Poor reproducibility within aggregate source. There are a considerable number of research reports involving the fine-tuning of this test method with the hope of obtaining better test results. This test method was never re-evaluated by DOT personnel. The soils service division in Des Moines still specifies the use of this test method and has test results available for a limited number of Iowa coarse-aggregate sources.
The idea behind this test method is to use the growth of salt crystals, obtained by subjecting a loose aggregate sample to several cycles of brine soaking and oven drying, to mimic the growth of ice crystals in the pore system of aggregate particles. Apparently, back in the 1940s, it was more economical to use ovens than freeze/thaw cabinets. The same question always remains. If information concerning an aggregate's pore system is desired, why not measure the pore system directly? There are a variety of test methods available for measuring pore systems. Why try to infer something about the pore system with indirect test methods?
In the 1970s, it became obvious that some aggregates in pcc were susceptible to deicing salt applications and were causing the concrete to deteriorate. At the time, some noted concrete researchers chastised us on our opinion, saying it is well known concrete can be durable in salty ocean water, which is true if high quality components are used in the mix. The idea that there are some aggregates that are salt susceptible is now more widely accepted. Scanning electron microscope (SEM) images show severe corrosion on the surfaces of individual carbonate crystals and crystallites when salt treated, polished sections from reactive sources are examined. Could the magnesium/sodium sulfate test method be used to identify these salt susceptible sources? Could calcium chloride or sodium chloride be used in place of magnesium/sodium sulfate?
During early evaluations of this test method, it was designated P-214. Consequently early research will refer to P-214.
Personnel at the Iowa DOT evaluated this test method in the early 1990s. The test method was not implemented as a standard test method because: (1) it correlated poorly with pccp service records, (2) it correlated poorly with alkalis in the cement, (3) it correlated poorly with the Canadian C-14-A test method and (4) it correlated well with periclase (MgO) in the cement. An ASTM test method (autoclave) for measuring expansion related to excessive periclase in cement already exists. ASTM specifications limit the amount of MgO in cements.
The initial evaluations focused on the variety of sands found across the state of Iowa. Later evaluations focused on different brands of cement, some local and some out of state. The only decent correlation to be found was P-214 test results to MgO as determined by XRF analysis. Two of the cement sources were reversed on the above correlation. When the cements were analyzed for the mineral periclase (all periclase is MgO, but not all MgO is periclase) using quantitative x-ray diffraction (QXRD) techniques, all cement sources fell into place nicely. This test method applies mild heat to the samples (much less than the Duggan test method), however even mild heat can relate to shrinkage micro-cracking in the samples. Any growth measured could be related to silica gel, ettringite, magnesium hydroxide, etc. In the U.S., the use of lithium salts are now being proposed as a way of reducing silica gel related expansion in pcc. In the U.K., lithium salts are recommended for use in pc grout to control expansion caused by ettringite. All concrete investigators, working on expansive concrete issues, need to absolutely verify causes of expansion. Conversely, it could be argued that if the use of lithium salts in concrete mixes reduces expansion, who cares what it relates to. But wouldn't it make more sense to insure that the initial hydrated concrete matrix has integrity and maintains its integrity (no shrinkage micro-cracking) rather than trying to control secondary growth mechanisms such as silica gel and ettringite.
Silica gel and ettringite can be primary expansive agents if poor quality aggregates or poor quality cement is used in the mix.
Only a limited number of samples were evaluated with the ASTM C1293 test method. Test results correlated well with pccp service life records. Only fine-aggregates (natural sands) were evaluated. A four lane pcc highway constructed in the late 1970s, in Webster county, Iowa, began deteriorating in 4 years. In trying to determine a cause, a considerable amount of time and money was expended by Iowa DOT personnel and other outside laboratory personnel. A large number of cores were obtained and analyzed by different laboratories. One laboratory, analyzing one pcc core, blamed excessive shale in the fine-aggregate for the early pccp failure. All other laboratories reported shale contents mostly under 2% and averaging closer to 1%, which meets Iowa DOT specifications. The coarse-aggregate source and the two fine-aggregate sources, had good service records. ASR activists used ASTM 1260 (ASR quick test) test results to question the quality of the fine-aggregate sources. Some pccp researchers suggested the lowering of pass/fail limits when using the ASR quick test method, so that a force-fit (blaming aggregates) could be accomplished. Since ASTM C1260 results did not agree with aggregate service life information, it was suggested that the fine-aggregate sources be evaluated using the ASTM C1293 test method. At the time, the ASTM C1293 test method was considered the best method for evaluating aggregates for alkali-silica reactivity, but the draw-back was the length of time needed to obtain test results (one year for plain pcc and two years for flyash pcc).
ASTM C1293 test results showed the fine-aggregates to be non-reactive when used in both flyash and non-flyash mixes, which matches the aggregate service life records. Before testing, the fine-aggregate samples were split. The split samples were tested (ASTM C1293) by another laboratory. Their test results were very similar to those obtained by Iowa DOT personnel.
The ASTM C1293 test method was used to implicate fine-aggregate (strained quartz?) used in faulty concrete railroad ties used in eastern U.S. A report concerning the testing of the aggregate contained confusing information. Both 8 inch and 12 inch beams were discussed in the report. Length measurement specifications require the subtraction of metal button length (3/4 inch on each end) from the overall length. Using the actual growths reported, the aggregate tested could have either passed or failed, depending upon the actual length of the beams fabricated. A request for clarification of information in the report, from the funding agency and the laboratory performing the test, was not forthcoming.
The durability history of strained quartz, used in pcc, is somewhat confusing in itself. There are research reports showing strained quartz to be anywhere from moderately durable to extremely non-durable when used in pcc. Some older research reports relate the use of strained quartz to blotched areas on a 60 year old dam in Canada. It is quite a stretch to relate this phenomenon to pcc railroad ties deteriorating in a few years.
Many researchers, working in the area of aggregate quality, have indicated that ASTM C666-B test results relate to the pore system within the aggregate particles used to fabricate the concrete beams. The C666-B test method is time consuming, labor intensive, and does not correlate well with field service records. If information about the aggregate's pore system is desired, why not measure the pore system directly? The mercury intrusion method, for measuring pore size distribution and total volume, was tried first. Being able to obtain a full pore size distribution was a big advantage, but working with mercury under very high pressure could present environmental problems in the laboratory. This test method was abandoned. A simple test method, using an environmentally safe liquid (preferably water) was needed. The test method would not have to give a full pore size distribution, but should be able to differentiate; (1) aggregates with no pore system, (2) aggregates containing an extensive capillary-sized pore system and (3) aggregates containing large diameter pores. With these parameters in mind, the Iowa Pore Index test method was devised and tested.
Using a modified concrete air meter, water is pushed into the pore system of 9000 grams of 1/2 X 3/4 inch, oven dried coarse-aggregate with air at 35 psi. A volume reading of the water intruded at one minute and at 15 minutes duration is taken. If no water is intruded during the whole 15 minutes, the sample has no pore system. If most of the water is intruded during the first minute, the sample has open, large diameter pores. If a considerable amount of water is intruded during the last 14 minutes, the sample has an extensive capillary-sized pore system.
Fluids more viscous than water (ethylene glycol and glycol) were also evaluated with this test method, but were found to be no better than water in differentiating aggregate pore systems.
Correlations to aggregate service records involving the pore index and C-666B test methods are similar as would be expected since both relate to the aggregate's pore system. Also, like the C666-B test method, the pore index test method is unable to properly classify medium/fine-grained dolomites used in deicing-salt treated pccp. Perhaps dolomites need their own pass/fail limits, but evaluating mechanical mixtures of calcite and dolomite would still present a problem.
The main advantage of the pore index test method over the C666-B test method is that test results can be obtained in a matter of minutes, rather than months, using a very simple test unit. Combining pore index test results (physical parameters) with XRF, XRD and TGA test results (chemical parameters) can give a decent correlation to pcc aggregate service records. Also, the pore index test method is non-destructive. If an error occurs, a leaking container for example, just re-dry the sample and run it again.
Over the past 20 years, the identification of factors that affect chemical reactivity have become more important for the proper evaluation of aggregates. Pore size diameters correlate well with grain size diameters and therefore grain interlock area. Fine- grained aggregates usually contain an extensive capillary-sized pore system exposing an enormous crystal surface area to any reactive fluids in the pore system, brine for example. Crystals and crystallites having a reactive surface coating (aragonite for example) can easily become unglued (lose their interlock)when reactive fluids enter the pore system. When the individual crystals lose their interlock, the aggrgate particle loses its integrity and can become susceptible to further freeze/thaw damage.
Aggregates containing an extensive capillary-sized pore system can be durable in pcc, if chemical (XRF, XRD and TGA) test results indicate high crystalline quality.
The Montour oolite ledge in Tama county always fails the ASTM C666-B test method unless excessive air entrainment is used in the mix. This ledge also fails the pore index test method. One foot diameter blocks of this oolitic material, sitting in a few inches of water on the quarry floor, will crumble when exposed to freeze/thaw cycles over a one year period, demonstrating what can happen when the wicking action fills a capillary-sized pore system that is subjected to freeze/thaw cycles. These oolitic beds would be poor candidates for high-moisture rip-rap stone. The strange thing is, this oolitic aggregate performs admirably as coarse-aggregate in concrete. It was first used in the 1940s, before the advent of air entrainment. Highway 14, south of Marshalltown, used this oolitic stone in the mix back in the late 1940s. This highway contained no air entrainment and began receiving deicing salts in the early 1950s and yet, it was still in service in the 1980s without being overlaid with asphalt. XRD, XRF and TGA analyses show this oolitic stone to be a nearly pure calcite with very little trace-element strontium. However, it does contain a minor amount of pyrite.
As stated earlier, there is a strong relationship between pore size and grain size, but there can be notable exceptions. The Waucoma ledge, in Fayette county, is a near pure calcite. Hand specimins from this ledge appear to be lithographic (extremely fine-grained). Pore index test results show the stone to be nearly non-porous. The Waucoma stone easily passes the ASTM C666-B test method and has a very good service record when used in pccp, even though severe discoloring at the pccp joints can occur. Crystal boundary etching, on polished surfaces, shows this stone to be coarse grained. The fossils in this ledge are blurred and indistinct, suggesting that the calcite in this ledge may have been dolomitized and then dedolomitized, which also would explain the large crystals and no pore system.
Discoloring at pccp joints is usually precursor sign to the development of further deterioration. Internal cracking of the matrix exposes calcium hydroxide. Moisture carries some of the calcium hydroxide to the surface where it combines with carbon dioxide, from the atmosphere, to form a gray calcite patina on the concrete surface. The cracking of the matrix can be caused by any combination of faulty cement, faulty mix design, faulty aggregate, faulty admixtures or faulty additives in the initial mix. Dick Burrows, in his ACI report, indicates that micro-shrinkage-cracked concrete can heal itself through the process of carbonization (calcium hydroxide + carbon dioxide = calcite) if all other components in the initial mix are of high quality. For this process to work, carbon dixode would need to find its way into the inside of the concrete. Discolored pccp made with durable Waucoma stone may be an example of this process.
Analyzing salt-treated, polished carbonate samples with the SEM has verified the fact that brines can severly attack crystal and crystallite boundaries. A variety of carbonate samples were treated with either sodium chloride or calcium chloride brine. The brine had only a minor effect on some carbonate samples.
The original pore index test unit, which used a modified concrete air meter, a graduated plastic tube and a stop watch, has undergone several modifications over the years.
(The following 11 paragraphs were added on 7-6-2007) An inquiry as to why 35 psi was used as the standard pressure for this test method was recieved recently. I will attempt to recall some of the events that occurred during the development of pore index test method in the following paragraphs.
The initial contraption consisted of a portable air tank, with an adjustable regulator, hooked up to the top of a graduated plastic tube connected to the lid of a standard pcc air meter. The unit sat on a wooden desk in the geology office.
When pressures higher than 35 psi were tried, water would leak past the seals in the lid even when the latches were carefully adjusted.
I do not recall of any pressures lower than 35 psi being tried, but lower psi levels should work. It would take longer to complete the test if a lower pressure was to be used.
Large grained silurian dolomites, containing large diameter pores, were used to determine when the initial tube reading should be taken. It was rounded off to one minute.
Fine grained Pennsylvanian limestones, containing an extensive capillary pore system, were used to determine when the second tube reading should be taken. Most of the fine grained stones stopped taking on water after 15 minutes, so a secondary reading at 15 minutes became the standard.
Quartzite was used to evaluate the test unit for aggregates with no pore system. There is some expansion in the test unit itself that needs to be taken into consideration.
A third tube reading at 30 minutes is suggested to insure the test unit is functioning properly. If the test unit indicates water intrusion past the 15 minute mark, the pot should be checked for leakage.
One of the topics discussed with Jim Myers and Ken Isenberger, during the early development stages of this test, concerned a 35 psi air pocket in the center of each aggregate particle at the conclusion of the test. In aggregate particles with large diameter pores, the air pocket probably would migrate upwards and escape during the time frame of the test, but with fine grained particles containing an extensive capillary pore system, the pressurized air pocket would probably remain in the particles during the entire run of the test.
The initial plexiglass tubes abtained were extruded. When Stan was threading the inside ends of a tube on the lathe, he discovered inconsistent wall thicknesses and inside diameters. Cast plexiglass tubes from Cedar Rapids solved the problem. The plexiglass tubes were designed to withstand 35 psi. Cutting inside threads in the ends would lessen the strength, but everything worked out OK.
To see if the volume of water in the tube would work for all aggregates, crushed tripolic chert was selected for evaluation. Tripolic chert contains an enormous capillary pore system. The volume of water in the tube was insufficient handle 9000 grams of crushed tripolic chert. Using 4500 grams of crushed tripolic chert and then doubling test results solved the problem. The first test unit used a 12 inch long, one inch inside diameter tube with a plastic ruler inside. The next unit used a 30" tube with the gradation ruler on the outside.
In the 1980s, the Iowa transportation research group gave a grant of $40,000 to the Iowa DOT geology section, with the purpose of obtaining as many XRF and XRD test results as possible, using primarily carbonate samples from quarries furnishing aggregate for use in Iowa construction projects. The XRD and XRF equipment was located in the Town Engineering building, on the Iowa State University campus, a short distance from the Iowa DOT headquarters. Most of the $40,000 was spent for rental time on the XRD and XRF units. Strontium, calcium, magnesium, aluminum, sulfur, silicon, sodium, potassium, manganese and iron were quantified with the XRF unit. A large number of samples were analysed over a two year period. Fortunately, the Iowa DOT has an open policy for employee access to the mainframe computer and furnishes many terminals and much software for employee use. Having to budget ahead for computer time and software would have been an added burden. The XRD and XRF data were entered and analysed by geology section personnel. Having the analysed data available, while the work was in progress, helped in giving direction to the study.
The XRF test results were correlated to pre-existing service record results. Excessive strontium (trace element range) in limestone (calcite) samples correlated with poor performance when used in pccp. Crushed limestone from the Ullin quarry in Illinois contained some of the higher levels of trace-element strontium. Illinois DOT personnel have indicated that the Ullin stone performs poorly when used in pccp.
Some concrete investigators have argued that such a small amount of strontium, in aggregates, could not cause pccp to fail. Most likely they are correct, but just because a correlation is established does not necessarily mean the relationship is direct. However, even if the correlation is indirect and not fully understood, it can still be used for classification purposes if the relationship is good enough. For example, most of the fine-grained limestone (calcite) aggregates from quarries located in the Pennsylvanian age bedrock do not perform well when used in pccp. Some of these bedrock ledges are over 97% calcite, and yet, when used in pccp, can cause failure, particularly when deicing salts are used. The individual crystals, made up of crystallites, are in the 30 micron diameter range. The pore system is in the capillary-size range and can be very extensive. What if these crystals and crystallites have an aragonite coating? Aragonite can hold up to 2% strontium and aragonite is less stable than calcite even though both are different forms of calcium carbonate (CaCO3). Under this scenario, unstable aragonite would be the actual culprit, but the amount of strontium would be used to identify its presence. When brine, from deicing salts, is introduced into the pore system, what happens? Would the extremely small crystal interlocks be disrupted? Would the aggregate particle then lose its integrity and become susceptible to further freeze/thaw damage? Indirect correlations can also be used to give direction to future research.
Excessive manganese in dolomite aggregates correlated with poor performance when used in pccp. The dolomite samples never contained much strontium, consequently it was assumed that the dolomitization process must remove it, if it was initially there in the original limestone, but later analyses of dolomite samples from a different, eastern sedimentary basin showed strontium levels similar to those from the Forest City basin limestones. Any state wishing to use this XRF technique to evaluate carbonates, would need to set up their own service record system for correlation purposes. Also, those states that do not use deicing salts or do not have their pavements subjected to freeze/thaw cycles could use less stringent parameters.
Cement companies have been using rapidly obtained XRF results, for quality control, for a number of years. The Iowa DOT is now using XRF test results for quality control purposes. Since each bed and each ledge has its own chemical fingerprint, it is easy to verify the aggregate for a certain project is coming from the proper quarry and proper ledge using this fast turn-around test method. The quarry operators need this information in a hurry too, as they do not want to put up a large stockpile of non-complying stone.
Since the XRF, XRD and TGA test methods only require a few grams of material, larger aggregate particles in hardened concrete can be removed and tested by these methods.
More pages in a few days.
