X VV V> Ok 1 Jl^T\ •-< 'Quo U w* v V Digitized by the Internet Archive in 2012 with funding from University of California, Davis Libraries http://archive.org/details/franciscancherti55gold FRANCISCAN CHERT IN CALIFORNIA CONCRETE AGGREGATES By HAROLD B. GOLDMAN Geologist, California Division of Mines With a table by IRA E. KLEIN Geologist, U. S. Bureau of Reclamation * Special Report 55 CALIFORNIA DIVISION OF MINES FERRY BUILDING, SAN FRANCISCO, 1959 STATE OF CALIFORNIA EDMUND G. BROWN, Governor DEPARTMENT OF NATURAL RESOURCES DeWITT NELSON, Director DIVISION OF MINES GORDON B. OAKESHOTT, Chief CONTENTS Page Abstract 5 Introduction 7 Composition of chert 7 Franciscan radiolarian chert 8 Comparison of Franciscan chert and Monterey chert__ . 10 Laboratory studies on reactivity of chert and other siliceous rocks __ 11 Laboratory tests 11 Reaction of silica 14 Field studies of concrete structures containing Franciscan chert . 17 Concrete service records and laboratory tests on cherts 23 Summary and conclusions _ 28 References _ 28 ILLUSTRATIONS Figure 1. Diagram showing hypothetical mode of transformation of amorphous silica to crystalline silica 10 2. Graph showing expansions of mortar bars made with a high-alkali cement and crushed reddish-brown Franciscan chert 13 3. Graph showing expansions of mortar bars made with a high-alkali cement and crushed light-colored Franciscan chert 13 4. Graph showing expansions of mortar bars made with a high-alkali cement and crushed stream pebbles of Franciscan chert from Ingram Creek 14 5. Graph showing expansions of mortar bars made with a medium alkali cement and crushed stream pebbles of Franciscan chert from Ingram Creek 15 6. Graph showing expansions of mortar bars made with low, medium, and high alkali cements and 50 percent Franciscan chert 17 7. Above, Diagram of unhydrated surface of a silica particle. Below, Diagram of hydrated surface of a silica particle 18 8. Map showing the locations of sand and gravel deposits sampled for petrographic analysis, status of development, and distribution of the Franciscan group 25 Frontispiece. Air view of San Francisco Bay Bridge 6 Photo 1. An outcrop of Franciscan chert 8 2. Franciscan chert pebbles 9 3. Photomicrograph of reddish-brown Franciscan chert used in U. S. Bureau of Reclamation laboratory experiments 4. Photomicrograph of Monterey chert 10 5. Photomicrograph of Monterey chert , 11 6. Photomicrograph of light-colored Franciscan chert used in U. S. Bureau of Recla- mation laboratory experiments 12 7. Photomicrograph of briquet made of crushed particles of chert pebbles selected from the alluvial fan of Ingram Creek 12 8. Scott Dam on the Eel River 16 9. Close-up of concrete in Scott Dam 18 10. Close-up of Scott Dam 19 11. Morris Dam on James Creek 20 12. Close-up of concrete in Morris Dam 20 13. Concrete in Morris Dam 20 14. Van Arsdale Dam on Eel River 21 15. Clear Lake Dam on Cache Creek 22 16. North Fork Dam on Pacheco Creek 24 17. Close-up of concrete in North Fork Dam 24 Table 1. Chemical composition of Franciscan chert 9 2. Geologic comparison of Franciscan and Monterey cherts 10 3. Comparison of physical properties of Franciscan and Monterey cherts 11 4. Concrete dams built with aggregate containing Franciscan chert 18 5. Distribution of Franciscan chert in sand and gravel deposits of selected streams with Coast Ranges drainage, by Ira E. Klein 26-27 (3) ABSTRACT Sources of concrete aggregate in the Coast Ranges have been viewed with suspicion because they contain large proportions of chert derived from the Jurassic-Cretaceous Franciscan group. The Franciscan chert has been categorized with chert from the Mio- cene Monterey group as being potentially chemically reactive with alkalies commonly found in portland cement. Comparions of the geologic occurrence, petrography, chem- istry, and origin of these two rock types reveal many differences, the most important being that the Monterey chert contains significant amounts of highly reactive opaline material; while the Franciscan chert contains less reactive chalcedonic silica. Laboratory tests demonstrate that the Franciscan chert is moderately reactive. Mortar bar tests made with crushed Franciscan chert and high alkali cement showed harmful expansions due to reactivity, when the proportion of chert exceeded 10 percent in the sand sizes. Aggregates containing more than 20 percent Franciscan chert in any gravel size or more than 10 percent in any sand size are to be considered as potentially reac- tive. Data on the distribution of Franciscan chert in deposits of the major Coast Range streams are presented to show that the proportion of chert rarely exceeds these amounts. Field examination of mass concrete structures built with sand and gravel containing varying proportions of Franciscan chert revealed no signs of alkali-aggregate reaction after 20 to 50 years. Commercial aggregates or potential materials in streams draining terrains underlain by Franciscan rocks should not on the whole be considered suspect of alkali-aggregate reaction because of their chalcedonic chert content. Only a few deposits contain the chert in proportions believed large enough to cause harmful expansion. These can be utilized — as can other deposits containing excessive amounts of other reactive in- gredients — providing the alkali content of the cement is controlled. (5) : RANCISCAN CHERT IN CALIFORNIA IONCRETE AGGREGATES \y HAROLD B. GOLDMAN Vith a table by IRA E. KLEIN INTRODUCTION Certain rooks and minerals are known to react chem- ;ally with the alkalies (sodium and potassium) in some ^ T pes of portland cement to cause a reaction known as lkali-aggregate reactivity. The reaction results in erack- lg and deterioration of concrete presumably due to smotic pressures produced by the formation and hydra- ion of silica "-els. The gels are formed through inter- ction between a susceptible mineral aggregate and the lkalies liberated by the cement during hydration (Mc- lonnell et al, 1950, p. 234). Alkali-aggregate reactivity resented a problem at the time it was first recognized ; owever, the reaction can now be controlled by the proper se of pozzolans and low-alkali cement and it is. possible o make satisfactory concrete with a chemically reactive ggregate. Opal (amorphous hydrous silica) is the most eaetive constitutent in aggregates. Other rocks and min- rals reported to be reactive are : glassy volcanic rocks f medium to high silica content (andesite and rhyolite), halcedonic rocks, certain phyllites that contain a hy- romica, the silica minerals cristobalite and tridymite, nd the zeolite henlandite. Chalcedonic chert also has een included in the list of reactive material by many workers. For example, Mather (1951, p. 218-233) de- sribes an instance of cracking in a concrete lock on the Varrior River in Alabama and concludes that the racking resulted from a chemical reaction between the lkalies of portland cement and chalcedonic chert. Evi- ence of alkali-aggregate reactivity in Parker Dam was ubstantiated by a petrographic study by Holland and Jook (1953, p. 991) who stated that pebbles of rhyolite. ndesite, siliceous limestone and chalcedonic chert were bund to be reactive. Holland and Cook in a general eview of alkali-aggregate reactivity in the western Jnited States (p. 996) state that certain sedimentary inits in California, including the Franciscan and Mon- erev groups, have contributed opaline and chalcedonic ock to the various aggregate deposits. The general im- iression is that chalcedonic cherts of the Franciscan ;roup are as reactive as the opaline cherts of the Mon- erey formation. California sand and gravel deposits lerived from Coast Range sources have in this manner teen considered suspect because of the presence of "po- entiallv reactive" Franciscan chert. Merriam (1953, >. 8), in discussing the Franciscan chert, states "Chert f the Franciscan formation is widespread in northern and central Coast Ranges, hence, it is a prominent constitutent in most gravel deposits in these areas. Al- though such gravel, when used as aggregate has satis- factory service history, laboratory tests have shown them to be reactive probably because of their content of chalcedony and possibly opal." This paper discusses the nature and extent of observ- able alkali-aggregate reactions in concrete built with aggregate containing chalcedonic chert derived from the Franciscan formation, and presents data that indicate that these aggregates are not to be considered poten- tially reactive because the Franciscan chert is not pres- ent in large enough proportions to cause harmful ex- pansions. The author wishes to acknowledge the assistance of Mr. H. Duane Woods, California Department of Water Resources, Sacramento, and Mr. Ira E. Klein, U. S. Bureau of Reclamation, Sacramento, in furnishing basic data and critically reviewing the manuscript. Mr. Cliff- ord Cortright of the Safety of Dams Section, California Department of Water Resources, was helpful in furnish- ing case histories of the concrete dams in California. Test data on expansion of mortar bars containing chert were kindly made available by the U. S. Bureau of Reclamation. COMPOSITION OF CHERT There is a considerable range in composition of rocks that have been termed chert. They may be entirely opal- ine, chalcedonic, quartzitic or composed of varying pro- portions of each. A general definition by Kemp (1940) is that chert is a rock consisting of fine-grained silica minerals, quartz, chalcedony or opal, or mixtures of these. A recent study of the texture and composition of chert was made by Folk and Weaver (1953) who con- cluded that chert is composed of two petrographic end members: microcrystalline quartz and optically fibrous chalcedonic quartz. Microcrystalline quartz is composed of minute interlocking grains in random orientation which under the electron microscope have distinct poly- hedral equant forms. The individual grains in most spec- imens average 3 to 5 microns in diameter, and the a^>rregate has "pin-point" birefringence; but in other specimens the grains are so small as to make the aggre- (7) CALIFORNIA DIVISION OF MINES [Special Report 55 gate appear nearly isotropic. Each grain in a thin section appears to possess undulose extinction, which is due in part i<> the effecl of many superimposed grains, hut in part may he also a property of the individual micro- crystals. Chaleedonie quartz under the petrographie microscope appears to he composed of radiating or sheaf-like bun- dles of fibers, each fiber not more than a few microns in diameter, hut as much as 200 microns or more in length, The libers are not physically separable, for no fibrous structure is evident on fracture surfaces studied with the electron microscope. The distinction between microcrystalline and chaleedonie quartz seems to be lather sharp although transitions do exist. In the selected samples studied by Folk and Weaver there was no evidence of any opal admixed in chalcedony or elicit. However, they stated that in some rock units, such as the Monterey, chert is associated with large, entirely isotropic, masses of opal with low index of re- fraction, but the two occupy discrete areas and are not so intimately mixed that the opal affects- the physical properties of the chaleedonie or microcrystalline quartz. There is still some controversy over the composition of chalcedony. Some of the early workers believed that chalcedony was composed of a submicroscopic mixture of fibers of quartz and interstitial amorphous silica, most likely opal. Later workers disputed that chalcedony con- tained amorphous silica or opal. Midgley (1951), by means of x-ray diffraction studies, reached the conclu- sion that chalcedony seems to be essentially microcrystal- line quartz with submicroscopic pores and that the quantity of opal, if present, must be considerably less than 10 percent. These pores or cavities, filled with water or air. are apparently the cause of the lower refractive index and density of microcrystalline quartz and chalced- ony as compared with normal quartz. The values of the physical properties change in proportion to the abun- dance of the cavities ; chalcedony and micro quartz have the same properties as ordinary quartz when free of bubbles (Folk and Weaver, p. 509). Pelto (1956, p. 32) in his study of chalcedony states that there is no sharp break between quartz and chal- cedony. In principle, chalcedony may grade impercep- tibly into quartz and therefore the classification of the material may sometimes be a matter of personal pref- erence. If either refractive index of the substance is lower than that of quartz, it may be termed chalcedony .or chaleedonie quartz provided it is not one of the other recognized forms of silica. In California, significant proportions of chert occur in the Monterey (Miocene), Franciscan (Jurassic-Creta- ceous), Amador (Jurassic), and Calaveras (late Paleo- zoic) rocks. The cherts in the Amador and Calaveras groups are considered non-reactive as they have been metamorphosed. Ordinarily composed of microcrystal- line or very fine grained quartz, they are not treated further in this paper. FRANCISCAN RADIOLARIAN CHERT Geologic Occurrence. Chert is a prominent rock type in the Franciscan group, which is widely distributed in the Coast Ranges from the Santa Barbara area north- ward to Oregon. The chert occurs typically in lens-shaped bodies com- posed of numerous thin-bedded chert layers separated by Photo 1 (below). Outcrop of thin-bedded Franciscan chert. I'holo by Salem •/. Rice. 19591 FRANCISCAN CHERT IN CALIFORNIA CONCRETE AGGREGATES Photo 2. Stream-transported Franciscan chert pebbles. Blocky form indicates the strong resistance of the chert to stream abra- sion. Scale to left is marked in half-inches. Photo by II. Dunne ^Yoods. thin shale layers. The chert layers are half an inch to 6 inches thick, but average about 2 inches. In places thicker layers of massive chert, up to 10 feet or more, are interstratified with the thin-bedded chert. The chert bodies range from a few feet to several miles in length and from a foot to several hundred feet in thickness, and commonly are enclosed in shale, sandstone, or con- glomerate. In many places the chert is closely associated with greenstone (altered volcanic rocks). The color of the chert varies from white or pale gray, to pale green, red, brownish red, chocolate brown, buff, yellow-brown, yellow and black. The most common type is the chocolate brown or reddish-brown variety. Gen- erally, the massive chert is paler in color than the enclosing thin-bedded chert, though often there is no dif- ference in color. Variation in color of the chert presum- ably is due to the difference in the abundance and state of oxidation of the iron present. Some green cherts show an occasional residual core of red in their interior ; some red cherts, a green core. Change in color frequently may take place as a result of metamorphism. In the lighter- colored chert the silica is recrystallized into a micro- crystalline chalcedonic mosaic. Petrography. Franciscan chert is characterized by remnants of the tests or hard coverings of Radiolaria, which are one-celled marine animals. The chert is dense and compact with a dull vitreous to waxy luster and a hardness of 6 to 7. Dark spots of clear silica, barely perceived by the naked eye, are radiolarian remains. Thin veins of chalcedony or quartz up to 0.1 inch thick transect the chert apparently along joints. The cherts are commonly breeciated and usually reeemented by quartz and chalcedony similar to that in the veins. Microscopic examination of the chert is difficult be- cause of its extremely fine-grained texture and therefore, the exact petrography is still obscure. X-ray and other modern techniques need to be used to ascertain the true petrographic character. In thin section, the cherts are seen to consist of a microcrystalline granular aggregate of chalcedonic silica grading into a very fine mosaic of interlocking quartzitic grains. In sonic instances, the matrix is practically opaque because of numerous specks of iron oxide. Transparent circular or elliptical areas in the matrix are the casts of originally hollow radiolarian tests. These areas are now filled with quartz or chalced- ony grains usually larger than those in the matrix. Locally, the chert is recrystallized by metamorphism to quartzite in which the silica is entirely in the form of quartz crystals. In places, higher grades of metamor- phism which may involve metasomatism result in the formation of such assemblages as quartz-crocidolite and quartz-garnet-glaucophane. The latter rocks may be mas- sive, schistose or gneissose. Chemical Composition. The common brownish-red chert contains silica ranging from 93 to 96 percent, alumina about 2 percent, and iron oxides from 1 to '.I percent. Small quantities of manganese oxide, magnesia, lime, water and the alkalies are present. Chemical analy- ses of selected cherts are presented in table 1. Table 1. Chemical composition of Franciscan chert. Brownish-red chert Red chert Red chert liaqleti Canyon Red Rock Inland, l't. Richmond, .]'lt. Diablo 1 S.F.Bay' S. F. 1 Si0 2 93.54 9r,.08 96.37 AM) : , 2.20 2.17 2.38 Fe a 3 .48 2.82 1.70 FeO .7!) MgO .66 CaO .09 Naj.0 .37 K 2 .51 H,() .93 MnO .23 99.86 100.07 100.45 'Analysis by Melville, 1891, Geol. Soc. America Bull., vol. 2. -Analysis by Davis (1918, p. 268). In 1958, semi-quantitative spectrographs analyses of a reddish-brown and a light-colored chert were made Photo .'{. Photomicrograph of reddish-brown Franciscan chert used in U. S. Bureau of Reclamation laboratory experiments. Veins are composed of quartz. Circular areas represent radiolarian re- mains. Crossed nicols. 10 CALIFORNIA DIVISION OF MINES [Special Report 55 COLLOIDAL SILICA flocculotion and solidification loss of woter and crystalization CHALCEDONY increased crystal size QUARTZ FIGURE 1. Hypothetical method of transformation of amorphous silica to crystalline silica. In general, the longer the geologic time, the more complete the transformation. by John C. Hamilton, U. S. Geological Survey. The red chert was found to contain ten times as much manganese, five times as much iron, and twice as much calcium and magnesium as the light-colored chert. Similar sam- ples of these cherts were used in the U. S. Bureau of Reclamation mortar bar tests wdiich are described on a following page. Origin. Lawson (1895) in describing the origin of Franciscan chert stated "the silica seems to have been an amorphous chemical precipitate forming in the bot- tom of the ocean in which the radiolaria thrived. The deail radiolaria dropped into this precipitate, became embedded in it, and were so preserved." Taliaferro and Hudson (1943, p. 230) believed that the cherts were chemical precipitates that were deposited in shallow marine environments. The silica forming the chert was presumed to be present originally as a colloid. According to Davis (1918, p. 376) the fact that the cherts consist largely of chalcedony supports the idea that they were originally gelatinous silica. Davis quoted Lindgren (1915, p. 233) as stated that chalcedony re- sults from the crystallization of gelatinous silica and that gelatinous silica may in becoming crystalline either turn into granular quartz or into chalcedony. The probable mode of transformation of amorphous silica to crystalline silica is diagrammed in figure 1, and presum- ably involves the following processes: transformation of the colloidal silica to opal by fiocculation and solidifica- tion; opal to chalcedony through loss of water and crys- tallization; and chalcedony to quartz through growth in crystal size. In general, the longer the geologic time, the more complete is the transformation. Taliaferro and Hudson (1943. p. 218) stated that the close and prac- tically constant association of the cherts with contempo- raneous basic lavas indicates that volcanic rocks were the chief source of silica. COMPARISON OF FRANCISCAN CHERT AND MONTEREY CHERT In California, all chert has been suspected of being potentially chemically reactive with the alkalies in port- land cement to cause harmful effects in concrete. It is appropriate at this point to compare chert of the Mon- terey group with the chert of the Franciscan group as it has been conclusively proven that the Monterey cherl is highly reactive. A comparison of the two types isi nted in tables 2 and 3 serves to demonstrate the obvious differences between them. The most signifi- uit difference, of course, so far as potential reactivity is concerned, is the high opal content of the Monterey chert (Bramlette, 1946). One explanation of the lesser amount of opal in the Franciscan group is that the opal in the Franciscan chert has had sufficient geologic time to be transformed to a crystalline state. Rocks of the Franciscan group are estimated to be 75 to 125 million years old and are characteristically highly folded and faulted ; those of the Monterey group are 25 million years old and only moderately deformed. Table 2. Geologic comparison of Franciscan and Monterey cherts (mollified from Davis, 1918, p. SOS, and Bramlette, 19J t G, p. 50). Franciscan chert Monterey chert Composition Amorphous silica not abundant Opal abundant Color Brilliant colors Dull colors, yellow to gray Foraminiferal and Fossils Radiolarian remains diatom remains Veining Common quartz veins Uncommon chalcedonic veins Luster Waxy Earthy Fracture Well developed Developed only in flin- conchoidal ty types ; conchoidal Age Jurassic-Cretaceous Miocene Origin Chemical precipitate Altered diatomaceous rocks Photo 4 (below). Photomicrograph of Monterey chert. Dark areas (o) are opal, circular areas (spherulites) with black crosses are chalcedony (c), and white areas are quartz (q). Crossed nieols, 35x. Photo by H. Duane Woods. 1959] FRANCISCAN CHERT IN CALIFORNIA CONCRETE AGGREGATES 11 Table 3. Comparison of physical properties of Franciscan ami Monterey cherts, from tcsis of Stale Division of IJiyh ways, Materials and Research Department. Specific gravity Absorption Soundness loss (Na 2 SO0 Opaline silica content General appearance Franciscan chert 2.6 - 2.8 0.92-1.01 percent 1.6 - 'i.'A percent Monterey chert 1.80 - 2.43 1.7 - 10.2 percent 2.s - 72.7 percent Primarily non-opaline Primarily opaline Usually hard, dense, massive, glossy ; gen- eral lack of minute microscopic pores or open fractures. Characteristically thin- ly laminated, many samples quite soft, full of microscopic pores. LABORATORY STUDIES ON REACTIVITY OF CHERT AND OTHER SILICEOUS ROCKS Laboratory Tests Many laboratory investigations have been conducted to determine the nature of the alkali-aggregate reaction and to delineate potentially reactive rock types. Results of these experiments, particularly those that involve the use of mortar bars, indicate that chert as well as other siliceous rock types will react with the alkalies in cement under certain conditions. Gaskin, Jones and Vivian (1955) examined several different forms of silica for reactivity with alkalies and for ability to expand mortar and found that all except megacrystalline quartz caused mortar expansion. Various forms of silica (opal, chalcedony, microcrystalline, and cryptocrystalline quartz) were crushed, sieved and cast into mortar bars which were stored at room temperature and at 100° F. Megacrystalline quartz was the only form of silica that did not react with alkalies at a measurable rate. Reduction in crystal size appeared to enhance the alkali-silica reactivity at elevated temperatures. Poly- Photo 5 (below). Photomicrograph of same sample as photo 4 under higher magnification. The large areas of opal are ex- cluded from the field of view by the magnification. The diameter of the spherulites (c) is about 0.00") inch. Each spherulite con- tains thousands of crystals of chalcedonic quartz that cannot be individually distinguished. The individual crystals of quartz (q) surrounding the spherulites are distinguishable. Crossed nicols, 290x. Photo by H. Duane Woods. crystalline particles containing microcrystalline quartz (crystals of about 10 micron diameter) reacted slightly and caused small mortar expansions at an elevated tem- perature, while at room temperature they reacted very slightly and caused negligible mortar expansion. Crypto- crystalline quartz (crystals up to 10 microns in diameter) reacted considerably at room temperature, but caused abnormal mortar expansion only at an elevated temper- ature. Reaction was even more marked with chalcedonic (fibrous) silica, and small mortar expansions were de- veloped during storage at room temperature. Opal pro- duced severe reaction and large expansions at room temperature. Alderman et al. (1947, p. 10) in their study of Aus- tralian cements and aggregates, state that in several aggegates which reacted with high alkali cement it was not possible to identify any material, such as opal, which had been found by previous workers to be reactive. These aggregates all showed "delayed" reaction which did not appear until the mortar was about a year old. The re- active material appeared to be cryptocrystalline silica, which was first thought to be chalcedony. Chalcedony has an average refractive index n = 1.533, which is lower than the corresponding index of quartz of 1.547. The refractive index of the cryptocrystalline silica in these aggregates was found to be that of quartz. Thus, this reactive material must be called cryptocrystalline quartz. Opal could conceivably be intergrown with the quartz, but very careful microscopic examination did not dis- close it. Many of the sands, sandstones and quartzites they examined contained fine quartz, mostly micro- crystalline, which, in some cases, could be called crypto- crystalline. These samples did not react. Therefore, recognition of potentially reactive material of this kind is still difficult. In 1941, the U. S. Bureau of Reclamation made mortar bar tests on a variety of natural materials including many common rocks and minerals obtained from various sources to determine their degree of alkali-aggregate reactivity. They found that chalcedonic cherts were re- active to some degree but considerablv less than opaline chert (McConnell et al., 1950, pp. 236-238). In Cali- fornia, the Bureau was faced with the problem of using aggregate that contained chalcedonic chert of the Fran- ciscan group. To establish their degree of reactivity, tests were made on three samples of Franciscan chert. The samples were collected by Ira E. Klein, geologist with the U. S. Bureau of Reclamation, Sacramento, and sent to the Bureau's concrete laboratory in Denver where mortar bar tests were conducted from 1947 to 1949. The samples were described as follows : Sample No. 1 (lab. no. 7551), 80 pounds of red chert from rock outcrop along Pope Creek, about 0.2 mile above the junction with Trout Creek, half a mile north of Samuel Springs, Napa County. This sample is the typical reddish-brown Franciscan chert. The rock is composed of opaque Cryptocrystalline silica believed to be sub-microscopic chalcedony in a matrix masked by reddish-brown iron oxide containing circular, colorless areas representing radio- lai'ian remains, now filled with quartz or chalcedony grains. Sample No. 2 (lab. No. 7552), 63 pounds of light-colored chert from the same locality as sample no. 1. This rock is composed of a mosaic of microcrystalline chalcedony grains containing isolated aggregates of larger quartz grains representing radiolarian remains. Sample No. ."'>. (lab. no. 7550). !)'.! pounds of }-lj inch chert pebbles hand sorted from (he gravel at the Westley pit, sec. .'',6. T. 4 S., R. 5 E., M. 1). The pit is in an alluvial fan of Ingram 12 CALIFORNIA DIVISION OF MINES [Special Report 55 Creek a small stream on the west side of the San Joaquin Valley, that drains an area containing Large amounts of Franciscan chert. The sample represents a complete natural assortment of the rock •is it occurs in an alluvial deposit and contains all varieties of Franciscan chert ranging from the typical reddish-brown chert with abundant radiolarian remains in a eryptocrystalhne opaque matrix, to recrystallized cherts composed solely of a microcrystal- line chalcedonic mosaic. Chert pebbles similar to samples 1 and 2 are present in abundant amounts. The samples were crushed, sieved, and incorporated into 1" x 1" x 10" mortar bars which were sealed with excess moisture and stored at 100° P. for 24 months. Measurements were made periodically to note any linear expansion due to alkali-aggregate reactivity. In the first test, aggregate in the bar was composed entirely of quartz retained in equal proportions in each screen size (if #8, #16, #30 #50, #100. The quartz was gradually replaced in subsequent bars with chert in all 5 sizes in percentages of 5, 10, 20, 50, 75, and 100. Complete suites of mortar bars were made using each sample combined with three types of cement : a low-alkali cement contain- ing 0.13 percent Na^O equivalent; a medium-alkali cement containing 0.59 percent Na- 2 equivalent ; and a high-alkali cement containing 1.38 percent Na 2 equiv- alent. ' "*■• ■ •* j-l** '■'■*■■• ' ■s&kM v * *_ . Photo 6. Photomicrograph of light-colored Franciscan chert used in C. S. Bureau of Reclamation experiments. The groundmass is composed of a mosaic of interlocking quartz grains. The larger circular areas represent radiolarian remains. Crossed nicols. Results of these tests (presented on the graphs in figures 2 through 6), indicate that no expansion occurred in any of the mortar bars made with combinations of chert and low alkali cement, and only negligible expan- sion occurred in the mortar bars made with the medivim alkali cement. When a high-alkali cement was used, ex- pansion of more than 0.1 percent * occurred at the end of 12 months in bars that contained 20 percent or more of the assorted cherts (sample 3) and the light-colored type chert (sample 2), and in those bars that contained 10 percent or more red chert (sample 1). Expansion oc- curred in all bars except those containing 75 and 100 pet-cent light-colored chert (sample 3) and 100 percent ing to tlie American Society for Testing Materials specifica- > rete aggregate < < ':;;!-. r ) IT), fine aggregates that own harmful reactions in concrete generally have pro- pansiona of 0.10 percent at 1 year when tested with a high alkali cement. red chert (sample 1). This is the usual pozzolanic effect obtained from reactive materials. The light-colored chert (composed essentially of mi- crocrystalline chalcedony grains), is less reactive than the reddish-brown chert (composed of opaque crypto- crystalline silica). All the bars made with the assorted pebbles from the alluvial deposit of Ingram Creek (sample 3) showed expansion with the high-alkali ce- ment. A more recent laboratory study was made of a Coast Ranges aggregate by the- U. S. Bureau of Reclamation in connection with the concrete used in Monticello Dam, Solano County. Aggregate for the dam was obtained from deposits near Esparto, Yolo County, on Cache Creek, which drains a portion of the northern Coast Ranges. Preliminary petrographic study revealed that Franciscan chert comprised up to 15 percent of the gravel and up to 5 percent of the sand. No other reactive ingredients were noted. From the foregoing mortar bar tests, it was deduced that the chert was not present in sufficient proportions to cause reactivity. (A complete summary of the chert content is presented in table 5.) However, in accordance with sound engineering practice, mortar bar tests were made to be certain that the aggre- gate was not reactive. At the end of 12 months only negligible expansion was noted ; the maximum expansion Photo 7. Photomicrograph of briquet made of crushed particles of Franciscan chert pebbles selected from the alluvial fan of Ingram Creek. All varieties of chert are represented in this size range of ,10 to 50 mesh. Crossed nicols. 1959] FRANCISCAN CHERT IN CALIFORNIA CONCRETE! AGGREGATES 13 24 00 50% CHERT 2-0 7« CHERT. *0°/o ^ HE.R" 1 010% Permi ssublie limit of exparosroni after one year, set by specifica- tion writers (ASTM C 33-54T) 7 5 % CHERT 5% CHERT 10 0%, QUARTZ 100% CHERT 2 4 6 8 10 12 15 18 21 24 Time, in months Figure 2. Expansion of mortar bars (ratio, cement to aggregate, 1:2) made with high alkali cement (1.38 percent Na 2 equivalent) and aggregate made of quartz and reddish brown Franciscan chert (sample number 1). The samples include aggregate composed of 100 percent quartz, 100 percent chert, and mixtures of quartz and 5, 10, 20, 50, and 75 percent chert. Tests made by the U. S. Bureau of Reclamation Laboratory, Denver. T 10 12 Time, in months o -n • f ™„rt,r hnr« (nti„ cement to aggregate, 1:2) made with high alkali cement (1.38 percent Na s O equivalent) Figure 3. Expansion mortar bars ^g 10 ^^^,^ | rancisca ' n cher t (sample number 2,. The samples include aggregate composed oi f MC ?^t q^£ M0 JSS5 chert, and mixtures of quartz and 5, 10, 20, 50, and 75 percent chert. Tests performed by the U. S. Bureau of Reclamation Laboratory, Denver. 14 CALIFORNIA DIVISION OF MINES [Special Report 55 000 50% CHERT '/. CHERT 100% CHERT e o%cHEJii ■10% mit of expansion r, set by specif i ca- ASTM C-33-54T) l 0% c HER T HE RT ____=--=^=^= = ~-* A _ RTZ 10 12 Time, in months 15 18 21 24 Figure 4. Expansion of mortar bars (ratio, cement to aggregate, 1:2) made with high alkali cement (1.38 percent Na 2 equivalent) and aggregate made of crushed stream pebbles of Franciscan chert from the alluvial fan of Ingram Creek (sample number 3). The samples 'include aggregate composed of 100 percent quartz, 100 percent chert, and mixtures of quartz and 5, 10, 20, 50, and 75 percent chert. Tests made by the U. S. Bureau of Reclamation Laboratory, Denver. with a high alkali cement was less than 0.05 percent. The test results of this experiment and data on the chemical and physical characteristics of the Cache Creek aggre- gates have been published in a recent article (Klein and Goldman, 1958). Reaction of Silica Many workers have attempted to ascertain the nature of the reaction between the alkalies in cement and a reactive rock particle. Powers and Steinour (1955) utilizing the work of P. C. Carman (1940) advance an explanation of what makes some forms of silica reactive. They propose that in all forms of silica the basic struc- tural unit is a silicon ion, Si ++++ , surrounded by four oxygen ions, = , the arrangement being that of a tetra- hedron. A particle of crystalline silica is made up of such tetrahedra linked together through their vertices, each vertex being occupied by an oxygen ion that is common to two tetrahedra. According to Carman, tetrahedra are linked to form an oriented 3-dimensional network in the various crystalline forms of silica. Similar principles should bo applied to the constitution of colloidal silica, s case silica gel. The structure of silica can be represented by a schematic diagram such as figure 7. In the diagram, the tetrahedron, composed of a silicon ion surrounded by four oxygen ions, is arbitrarily repre- sented by a square with the four oxygen ions at the corners and the silicon ion in the center. This shows how a continuous structure is built, each oxygen ion being linked to two silicon ions. The three dimensional counterpart of the structure would form a silica crystal. In the interior the valence of each ion is satisfied, four for each silicon and two for each oxygen. These ions therefore do not carry unneutralized charges. However, for the ch.em.ical composition of the whole to be SiO"2, the tetrahedra. at the surface of the crystal cannot be completed. The oxygen ion at the surface can be bonded to only one silicon ion instead of two as in the interior. Thus it is left with one unsatisfied negative charge. Similarly, each silicon ion at the surface lacks one oxy- gen ion, and correspondingly it bears one unsatisfied positive charge. If moisture is available, the charges on the surface of the silica bring about surface hydration, as illustrated in figure 7. The positive hydrogen ion from the water joins with the negative oxygen ion on the surface of the silica particle, and the hydroxyl ion, OH", from the water joins with the positive silicon ion, 1959] FRANCISCAN CHERT IN CALIFORNIA CONCRETE AGGREGATES 15 thus completing the silicon-oxygen tetrahedron. The re- sult is a surface layer of OH groups. Water held in this manner may be regarded as adsorbed water, since the quantity is proportional to the surface area, and not to the total amount of silica. The hydrogen ion held by an oxygen ion at the silica surface is bound less firmly than the OH group as a whole. Hence, ionization in an aqueous medium frees some of the hydrogen, producing free hydrogen ions. The surface of the silica particle therefore is weakly acidic. This explains why the chemical properties of silica are influenced significantly by the degree of subdivision of the silica or by the imperfections in its crystal struc- ture. If the silica particle is megascopic, having a low specific surface,* its acidic character is not noticeable because the number of hydrogen ions per mole of Si0 2 is almost vanishingly small, for the hydrogen ions are produced only at the boundaries of the particle. Obvi- ously, if smaller particles are produced by subdivision i of a megascopic particle the number of adsorbed water molecules and hence the number of free hydrogen ions i per mole of silica increases and the "reactivity" in- creases. * Specific surface is denned as the surface or area of a substance or entity per unit volume ; obtained by dividing the area by the volume, and expressed in reciprocal units of lengths. Hence, a small particle will have a higher specific surface than a large particle. The reaction that evidently takes place when caustic alkalies attack opal or vitreous silica is brought about by the aggressive effect of the OH" ion. The alkali hy- droxides are capable of more aggressive action than other hydroxides, primarily because they are more soluble and can produce higher hydroxyl ion concentrations. At high enough concentrations, sodium hydroxide not only re- moves and neutralizes the surface hydrogen ions, but also severs silicon-oxygen-silicon linkages that hold the mass together. The reaction at the broken linkages is similar to that at the surface. By this process, amorphous silica is reduced to colloidal particles and higher hy- droxide concentrations produce smaller particles. A similar line of reasoning was developed by Gaskins, Jones, and Vivian (1955, p. 83-) as the result of labora- tory experiments wherein quartz-bearing aggregate was embedded in cement pastes that contained radioactive sodium. Examination of autoradiographs revealed that megacrystalline quartz particles showed no signs of ac- cumulation of radioactive sodium, even after 3 months storage at 110°F. Opal showed a very considerable ac- cumulation throughout the particles after storage at room temperature for only 1 week. The samples con- sisting of microcrystalline and cryptocrystalline quartz showed either very slight or no signs of reactivity after 3 months storage at room temperature; at 110°F. after 3 months, all samples showed signs of surface reactivity. -00 I 5 FlGl'RE 5, 010 0-0 15 — 00 10 00 5 0000 ^^^ i-0-00 5 10 12 Time, in months Expansion of mortar bars (ratio, cement to aggregate, 1:2) made with medium alkali cement (0.59 percent Na«0 equiva- lent! and aggregate made of quartz and crushed stream pebbles of Franciscan chert from Ingram Creek. 1 he samples include aggregate composed of UK) percent quartz. 100 percent chert, and mixtures of quart/, and 5, 10, 20, and .,() percent chert. Test's made by the U. S. Bureau of Reclamation Laboratory, Denver. ■ ^£5JR i •I [959] FRANCISCAN CHERT IN CALIFORNIA CONCRETE AGGREGATES 17 Photo 8 (opposite). Sfott Dam (Lake Pillsbury) on Eel River, Mendocino County. Built in 1!)21 with aggregate contain- iig Franciscan chert. [n no instance was reaction detected in the centers of juartz aggregate particles. Gaskin et al., concluded that the lack of even slight signs of reactivity throughout the mierocrystalline and :ryptocrystalline quartz particles suggests that they con- sisted only of quartz and not of intergrowths of quartz md opal. Consequently, it was reasonable to assume that juartz reacts at a very slow rate, and that the increasing •eaction rates of micro- and cryptocrystalline forms of juartz depend on their increased surface areas. They dso concluded that there is a correlation between in- creasing reactivity and decreasing crystallinity of silica. Permeability of the aggregate particles to alkali solution is also an important factor in accelerating the extent of reaction. The reactivity of opal and perhaps that of the silica minerals cristobalite and tridymite and the crystal- line forms of quartz could be attributed in part to this factor. It is probable that the Franciscan chalcedonic chert, which contains the less reactive forms of silica and is dense and relatively impermeable, will remain sound in the presence of greater concentrations of alkali while aggregates such as the Monterey chert which con- tain opal may cause abnormal expansion if this alkali limit is exceeded. FIELD STUDIES OF CONCRETE STRUCTURES CONTAINING FRANCISCAN CHERT To ascertain whether Franciscan chert contained in- gredients, such as has been previously discussed, which 0-32 030 28 0-26 MORTAR BARS MADE WITH HIGH-ALKALI CEMENT ( TOTAL ALKALI ES 138 %) -002 10 ^ STREAM PEBBLES __F^£!1 SAMPLE N°-3 in r-o aM CREEK R_EJD mSH _BR 0_WN __CH E_RJ_ . SAMPLE N° I MGHT COLOREO-SHEJLT- ^Vmp-TTn - 2 10% PERMISSIBLE LIMIT OF EXPANSION AFTER ONE YEAR SET BY SPECIFICATION WRITERS (ASTMC33- 54 -T) MORTAR BARS MADE WITH -MEDIUM ALKALI CEMENT (TOTAL ALKALIES 0-59%) MORTAR BARS MADE WITH LOW ALKALI CEMENT TOTAL ALKALIES 0-13%) 12 I 15 _1_ I 18 I 21 24 I Time, in months Figure 6. Expansion of mortar bars (ratio, cement to aggregate, 1:2) made with aggregate composed of 50 percent different samples of chert were used for the bars: reddish-brown chert (sample number 1), light-colored C number 2) and assorted chert pebbles from Ingram Creek (sample number 3). The cement used was high (138 percent Xa-,0 equivalent), medium alkali cement (0.59 percent Na*0 equivalent) and low alkali ceme cent Xa-O equivalent). Tests performed by the U. S. Bureau of Reclamation Laboratory, Denver. chert. Three hert (sample alkali cement nt (0.13 per- 18 CALIFORNIA DIVISION OF MINES [Special Report 5 Si Si Si / \ / \ / o s X J/ \/ \/ Si Si Si '|\ / \ / \ Si Si Si / \ / \ / 0, .0 ,0 .S Si — / \ — Si — / \ \ / 0- 0- 0- \ \ \ Si* Si + Si + I / \ / \ / \l \ ./ \ / l N Si Si Si . Surface region " Interior ■** f -th ;!• ~w H H H H H H I I I I I I S 0, I l\ Si s. Si Si Si 1 V - Si- / \ Si ,0 J/ \ / \ / .Si Si Si , / \ / \ , Si Si Si Si Si Si / \ / \ / \ .0 / Surface region ( h y dr a t e d ) > Interior Figure 7. .46ore, Diagram of unhydrated surface of silica particle. From Powers and Steinour, p. 499. Below, Diagram of hydra ted surface of silica particle. From Powers and Steinour, p. 500. would react with the alkalies in cement, a survey of massive concrete structures built with aggregate contain- ing Franciscan chert was made by the author in Novem- ber 1957. The field inspection consisted of examination of concrete dams in the northern and central Coast Ranges and concrete structures in the San Francisco Bay area. Photo 9. Close-up of concrete in Scott Dam. Concrete shows r signs of alkali-aggregate reactivity. Franciscan chert pebbles ai evident on the surface of the concrete. The concrete in each structure was examined closel to affirm that chert was present in the aggregate and t note any signs of alkali-aggregate reactivity. Listed i table 4 are seven mass concrete dams that were examinee None of these showed signs of reactivity. These sam structures are periodically inspected by the Californi Department of Water Resources. The reports of their in spectors corroborate the author's findings that none o these dams, which were built with aggregate containin; Franciscan chert, exhibit any signs (such as pop-outs map pattern cracking, or exudation of gel) indicative o alkali-aggregate reaction. Most of the concrete structures in Marin County wer built with Russian River aggregate that contains up t 12 percent Franciscan chert. These structures, which ar| too numerous to list, include the Waldo Grade tunnels oi U. S. 101, the San Rafael Viaduct and the streets o Sausalito. Many concrete structures in San Francisco Count} western Contra Costa County, western Alameda Countj and northern San Mateo County were built with aggre gate from the Livermore Valley area, Alameda Countj that contains up to 14 percent Franciscan chert. Thes deposits have been, and continue to be, a principal sourc of concrete aggregate for the above-mentioned servicn areas. A list of structures built with aggregate fron these sources would be overwhelming; some of the morj prominent structures are : Table 4. Concrete dams built with aggregate containing Franciscan chert. Owner Stream Location Name of dam Sec. T. R. Year | built | 33 14-23 30 25 6 16 22 38N. 18N. 18N. 15N. 12N. IN. 10S. 13W. 10W. 10W. 12W. 6W. 7W. 6E. 1927 1 Scott (Lake Pillsbury) South Eel River . 1921 t Van Aredale-- South Eel River 1907 ' 1915 j 1914 I 1917 fl 1939 | ' Hear Lake Alpine* Nortli Fork. Pacheco Creek • Crushed rock from site used as coa rse aggregate with sand from Russian Ri ver at Healdsburg. 959] FRANCISCAN CHERT IN CALIFORNIA CONCRETE AGGREGATES 19 Photo 10. Close-up of Scott Dam, showing the good condition of the concrete. ajor portions of concrete in Golden Gate and Bay Bridges >sey Tube between Alameda and Oakland neon Annex Post Office, San Francisco >praisers Building, San Francisco iss Building, San Francisco ack at Selby Smelter S. Mint Building, San Francisco lephone Building, Oakland liversity of California Library, Berkeley ty Farm Island Bridge, Alameda neifie Gas and Electric Steam Plant, Antioch and Pittsburg ikland Library I Mateo County Tuberculosis Hospital irks Air Force Base n Jose Hospital Bjpping Center, Walnut Creek uthern Pacific Bridge, Martinez [None of the aforementioned structures show any signs iat could be attributed to alkali-aggregate reactivity. pe author's observations are corroborated by a report fide by the U. S. Army, Corps of Engineers, South- Bt Pacific Division, on the San Lorenzo River project wipublished memorandum). A portion of the summary ! the service record survey from this memorandum is noted below : "Plants in the Pleasanton, Livermore, Centerville, Niles and Sunol areas have been producing concrete aggregates since 1930. This material is used extensively in concrete construction by fed- eral, state, county and municipal agencies and by local building contractors in the San Francisco-Oakland Bay area. These sand- stone aggregates are regarded to be relatively non-reactive by the California Division of Highways. Records in this department show that to date there have been no evidences of cement-aggregate reactivity found in structures containing aggregates from tbis area. The State findings have been substantiated by the examination, in this area, of approximately 48 concrete structures of various types in the service record study phase of this report. No concrete structures inspected in this area show indications of cement- aggregate reactivity." One of the structures that was selected in the Corps of Engineers report to represent the overall condition of concrete inspected in the field study was a 5x7 foot con- crete aqueduct constructed by the San Francisco City Water Department in 1925. Water flows through this structure between Snnol and Xiles in Alameda County. The aqueduct, reportedly, was constructed with local alluvial sand and gravel. There are no pattern cracks, excessive leaks, popouts, gel exudations or other signs CALIFORNIA DIVISION OF MINES [Special Report I Photo 11. Morris Dam, on James Creek in Mendocino County, was built in 1927 with aggregate containing Franciscan chert. J%*»fc [i i 12. Close-up of concrete in Morris Dam. The concrete condition and shows no signs of alkali-aggregate ily. Photo 13. Concrete in Morris Dam showing good condition. Franciscan chert pebbles are in evidence in concrete by the pick. 1959] FRANCISCAN CHERT IN CALIFORNIA CONCRETE AGGREGATES 21 i- o - a ■3 O O o a "S o 'a ' B IB tflfl ^■rx 1959] FRANCISCAN CHERT IN CALIFORNIA CONCRETE AGGREGATES 23 normally associated with expansive or reactive aggregate in concrete. There are no data available on the alkali content of the concrete that was used in the above structures. In all probability, a good percentage of the cement used in the San Francisco Bay area in the older structures could have had a relatively high-alkali content judging from the raw materials that were used in the cement-making process. In the Central Valley, aggregate produced from depos- its bordering the Coast Ranges contains varying amounts of Franciscan chert and has been used for many years in concrete construction without record of distress due to alkali-aggregate reactivity. The extensive commer- cially developed deposits in the river plain of Cache Creek in Yolo County and on the alluvial fan of Corral Hollow Creek near Tracy, San Joaquin County, are major examples of aggregate sources containing consid- erable Franciscan chert that have been widely used in major public works and have excellent service records. The Cache Creek deposits have been the subject of a recently published report (Klein and Goldman, 1958), in which the Franciscan chert content is covered in some detail. Tests on the reactivity of this aggregate were dis- cussed earlier in this report. Aggregate from the deposit developed in the Westley pit on Ingram Creek, San Joaquin County, for use in a portion of the Delta-Mendota Canal by the U. S. Bureau of Reclamation, contains more than twice as much Fran- ciscan chert as any commercial or potential undeveloped deposit considered. The deposit has been previously referred to as the source of one of the samples (no. 3) used in the mortar bar expansion tests by the U. S. Bureau of Reclamation. This material was used with a low alkali cement in 1947, and the concrete has shown no signs of reactivity. There are some deposits of aggregate that contain Franciscan chert in addition to highly reactive opaline sedimentary rocks. These deposits are chiefly in the southern portion of the state and are noted in table 5. There are also local areas, such as in Kelsey Creek near Clear Lake, where siliceous glassy volcanic rocks occur in the aggregate which also contains some Franciscan chert. In these instances the chert is of secondary consid- eration as a reactive ingredient. CONCRETE SERVICE RECORDS AND LABORATORY TESTS ON CHERTS Laboratory tests indicate that Franciscan chert is moderately reactive ; yet concrete structures built with aggregate containing chert have good service records. This raises the question of the adequacy of the labora- tory test methods in forecasting the behavior of reactive ingredients in concrete. There are two standard methods for determining alkali-aggregate reactivity— the mortar bar test and the quick chemical test. The mortar bar test has been described already. The quick chemical test consists of digesting a pulverized sample of the aggre- gate in a sodium hydroxide solution. The amount of dissolved silica and reduction of alkalinity in the filtrate Photo 15. (opposite). Clear Lake Dam on Cache Creek, Lake County, built in 1014 with aggregate from Cache Creek contain- ing Franciscan chert. Top. downstream side of dam; bottom, up- stream side of dam; mset, close-up of concrete showing good condition (Franciscan chert pebble at tip of pencil). arc used as ;i measure of the potential reactivity. Both tests arc valuable in indicating potentially reactive in- gredients; however, there are serious shortcomings in both methods. According to Lerch (1956, p. 34:5), "the rapid tests provide valuable information that can lie used to identify the presence of reactive constituents in the aggregate, but they do not always give assurance that the reactive material is present in the proportions necessary to cause expansion." This is a critical feature of reactivity — that the reactive material he present in the necessary proportions and sizes. In conducting the mortar bar expansion tests on chert, the U. S. Bureau of Reclamation used the standard ASTM method C227-52T. In this method (ASTM 1954, p. 301) either fine aggregate or coarse aggregate crushed to size is used in the following proportions. Screen Xo. 4 percent retained 8 20 16 40 30 00 no so 100 100 The ASTM specification writers admit that coarse aggre- gate crushed to sand size may give accelerated expan- sion because of the increased surface (specific surface) exposed upon crushing the aggregate. Actually, the chert particles in a natural stream deposit have been abraded and exposed to the solution effects of the stream water and do not present freshly crushed faces to the attack of alkalis as does the chert which is crushed for use in laboratory experiments. The specifications also state that if coarse aggregate tested by this method pro- duces an excessive amount of expansion the material shall not be classed as objectionably reactive with alkali unless tests of concrete specimens confirm the findings of the tests of the mortar. In actual practice, the grading of a specification con- crete sand is not the same as the grading of the sand used in the laboratory mortar bar tests (20 percent re- tained on the No. 8, 16, 30, 50, and 100 sieves). Com- mercially processed sand ordinarily contains from 8 to 20 percent retained on the No. 100 sieve and from 2 to 10 percent on the No. 200 sieve. Generally, the sand frac- tions are in the lower end of the percentage range. As already shown, particle size is an important factor in reactivity. Vivian (1951) demonstrated in laboratory experiments on crushed and sieved opaline material that mortar bar expansion is increased by a decrease in the particle size of the reactive component. Vivian's text figures (pp. 492, 493) indicate that mortar bars made with reactive particles in the 7 to 18 mesh range showed 0.5 percent expansion, while bars made with finer size reactive particles showed expansions of 1.0 to 2.0 per- cent. Maximum expansion occurred in the size ranges less than 35 mesh (0.5 mm) when the amount of reactive material was 10 to 20 percent. The mortar bar tests of the Bureau of Reclamation showed that when crushed Franciscan chert was substituted for quartz, from 10 to 20 percent was required in each si:re range, before ex- pansion reached 0.1 percent at the end of 12 months. The most important factor is that in the deposits of natural sand and gravel now used as commercial aggre- gate, and in most large undeveloped deposits the per- centage of Franciscan chert rarely approaches the CALIFORNIA DIVISION OF MINES [Special Report 55 * a ^-^- C ii Photo 16. North Fork Dam on Pacheco Creek, Santa Clara County, built in 1039 with aggregate from Pacheco Creek containing Franciscan chert. i~ Close-up of concrete in North Fork Dam. This con- ows no signs of alkali-aggregate reactivity. Franciscan • Me at tip of pick. percentage used in laboratory experiments such as those of the Bureau of Reclamation. The percentage of chert in the gravel fraction of a natural stream deposit in California rarely exceeds 15 percent; and the chert rarely exceeds 10 percent in the coarser grades of the sand fraction. Generally the percentage of chert de- creases with grain size so that an aggregate that contains 10 to 15 percent chert in the |-1| inch size will ordi- narily contain less than 5 percent in the finer sand sizes. Since alkali-aggregate reaction is a function of the specific surface, it is probable that an aggregate could contain up to 20 percent chert in the gravel fraction before harmful expansion would occur. A deposit that contains over 20 percent Franciscan chert in the coarse fraction (over ^ inch), and/or over 10 percent in the sand fraction should be considered suspect of potential reactivity. However, it should also be borne in mind that good serviceable concrete can be made with reactive materials if proper control of the alkali content of the cement is maintained. 1959] FRANCISCAN CHERT IN CALIFORNIA CONCRETE AGGREGATES 25 EXPLA NATION FRANCISCAN GROUP DEPOSIT FROM WHICH COMMERCIAL PRODUCTION OF CONCRETE AGGREGATE HAS BEEN OBTAINED LARGE DEPOSIT USED LOCALLY FOR CONCRETE AGGREGATE NO PERMANENT PROCESSING PLANT LARGE DEPOSIT NOT DEVELOPED FOR CONCRETE AGGREGATE FRANCISCAN CHERT CONTENT IN 3 / " - I '/ " GRAVEL < 1% OR ABSENT I - 4°/. 5 - 19°/. 20 - 39% 40°/. OR MORE Geology odopted from revised version of Stote Geologic Mop (unpublished) \ \ Figure 8. Map showing distribution of rocks of the Franciscan group, location of sand and gravel deposits sampled for petro- graphic analysis, and status of development of deposits. CALIFORNIA DIVISION OF MINES [Special Report 5." 13 1 s -a rt "£ 49 £<$ V £ e V o T3 W ■Sag •S-2 o -O J* cd Is 5.2 '.S .2t3 * o .5 £.ti.ti.£ £.t£ o t- i- & c t. c C-g ft c-o-c ft.2-o_- ^ , ^ ^3 -j cj T3 i> iJ3 o oJ3 o Ou & aX-S ft *"S > - — . — > >_ ti > co 5f ee rt i- a> cfl +j _cu"0 sSeI'S'S §T§ § ° O C rt rt O O rt ;- .2 * s ea w a. be a -o w^g.2 -s - So.S-a"8 ° §|&|s.al| ° e ° S e a £>: W 6 o S *o« a 43 42 ^ J 3 a a rt -i i GQ t> V fi C o. bO F P. p. ft ft s 3 = o E *— rt *2 -a c o c 3 o w .ti i bi &S * ^* -ago Q.J5 "£ o ft •a -a a. 3 . bb o.s.s ^ ^ *^ tog E..S c - J^-o o •2?3, o rt si .2 n o I B o *2 _>» £ o o T3 bl c — bC o 3 O cd bC o eg P i. -a o. cj - . C M bD 88 .S. 9 3^ O.Q. 2 e ° ° "Jo III 1 i ° bO B d ft O a > 3 *|£* * 5 a- y * 3 £f fc > ■„ > — a; o cu cS "O ca."a-- 3 — O ft b! s ^i2 3 O Q, - o 3 1 O. -J o 5 c bb| - —■ ~ d ■^ rt ft ft >> c fc S .Ha'! ~3 be C8 1-3 bC ►J -£ 4- 1 ft a >i 5^: -a e* >t •" : OT E £ « — -y. : - - 4-C J o V c « -o-o"0-a tr. a ~> at -3 ~ — ■ s es 49 s 3 ■~ B —■ - -.1 bC i = 2. jQ o3 e c = 13 ill £?* "oi2£ *> a - 3 £S a* C — o U3 CO lO to ■^ cO~* * " O CO to 01 r- ^H CO W3 -*• CO to 0—1 1-1 % N CO CO *oas 00 CM to CO c O CO CM * CO CM CO tO CM tO o co coo — "Ccoai CO CO -f tO CO CM o ao —. CO CO CO occf 00^- CCCOO — OOC— - to CO — sr>- o cm o t^- t— co to co to -^r co 00 -h CM 00 t-- 1 COCOCM OO CM CO ■< -e.b.tilUXOXX o 3 "75 0« 1 3 =3 J II «OOH •S .2 -3 ;- ^4 aj c£ D B £ & 5 3 Pi "3 -^ ScO j o"o o"o g § 2 E EEEEs 2^5 ^ 3 333 3 js a-g-g •§ -= -^ ■* SJii Ji rt o — ~z S££ = > > >o Si = e «S g-o o j; £ > O — si £ O « 3 .1 Bj| •3. «s« ^0 = C to CO t^ QO Oi O — • CM CO T tO — . ^- — CM CM C K ^O^-J ^5-3^ CO f to CO CO CO CO CO CO CO CO 1959] FRANCISCAN CHERT IN CALIFORNIA CONCRETE AGGREGATES 27 >, fl- >> _Q ea -O T3 § T3 T3 *" -o 03 b£ .— bi^ ft -o ft c 3 3 cd ■+a -«J & _w ft c « ?! &O. '3 °" ^ eg " Q ! Ed £ bO oa Ql o a "5. [2 o « o Hi E >>* ■°3 -d 5f S 1 c oS 3 C o 1 5g I -a "^ > J ^ ft o fl o 6^ o O oJ c t3 |^ 03 -2 ■S -3 .2 5 -2 11 w 03 3 c £ = H s bC bC bC bfi bo bO cd oa 08 03 "o3 .S ft Jrf J* M M ^^ u O O o O C O c £ u " M >i >> >> >, >> >> fc cd -B 3 oj «3 ft-— -d c c g *= C ^j £ c cd_^ 5 £ £ 1 Z - s S £ a — O r- -5 -5 . ~P £ "3 ■5 03 K.5 "*. 03 ' J: _C e ■w c £'3 c „."0 a . t« j% ^5'S d a | § _c >> c > .§ ? ft "* o, rt !> g.> "oa -r ft^ "3 U ft." M £ 3 o O O O O o -J rH - CM o o CO CO - o o o o W3 CO - o o o o U3 -r ^ o o > M % "eS o ~o o •g cd £ c ed s bfl OS o o ed 1 3 C oa CO B OS 'o CO 3 ed O K. C fl CQ S 03 CO o ft 2 cS cd 03 ,c 3 3 03 OS C -ft _c -g -2 c3 c9 OS b8 ■~ PQ CQ CQ CQ rt o ►J « 3 5 3 3 < c c d C CS os cd o3 oS CO CO CO CO CO > J > > > > I. fc tS S s S o3 > 1 s C 1 c 03 03 OS CJ >H fcM >< o O 3 5 & 3 »3 CT c c G 3 ed 99 93 o3 CS CO CO CO CO CO CO CN CO <«*< ifl to ^ -r "*r T .2 S •- ^-■r « ;1 - r. — ql,-- -^ ^ =.'£■-•-■=. if = '"' .2 <^ a* -a'"" =5 «^, ._ C • - = - "5 QQ - s. S 5 « .2 £ "2 ?" = .i x - = I'M * 51 >• O S — ^ rt it „ 3 — ' Sis 1-1 - - 1 |«s1sis| g S-§ , i ~'~- '- =-? _ SO -"" y. "J -, v. - yoXU £ M * |f "3 s = r: -*-h ts sj "3 ^ 1 +_J c 0. '/. ■--a-^: OJZ r H r CQ +J >a /. x t- C3 5 >»■" .a bB <* ~ z -a 21 a o a- rt ^ "t M" c- a- T^ c a — — c 5 P & B K C Ov - ^ ^ ^ - a^ " q cs Hi K :— o ? tt: : ^-a/i||^.||||j fi jt - *" a< a- 3 a i- !■; r?--CsC!>,: •S 0, 2 o^ a ° ^ c i y. a « oj a' ' a ~ '■ "■ - '~. f^-sw-ss u : * s i- o-'a-S g.2§cg l^Z^ ca«l'= i c •= 15 - c . "S - =-. — — n. T" 1 ri. . a a ^_ a x , 5?^£§^^a-£|£-S a % r , ■ _ i a. •= a v. .- a - 1 •*. / - C js ^ a, o, a bi v t £ o , a *- o H i a c ,- ■*-. p3 S <« ** "o * '-.^ a~ a- o_|.S a-^ ^b 3 »< S *S * u ** a 5 .2 t, "^ u o t^ c r- -/ o, a ^ ,w f - oi C o t "Siia.'^- e > ■= ■£ £ ■« ^ 5aia r a- ,8..*SSj 4-fc'g-S B aj oi bi av c..a c 1^-1 g i. '— - o ^ ^ av c a x a a- a^ -r* •— a^ ^ a* ^-' aa~*^ *-" -Xc v £ ■ax- ■gss a a, *t^ _x j a 4-. ja a« a_i a I £. a a QJ 4> ™ Sh W -— ^a a^ .^ ^^ x a. r, 3 5C- b.a ..aO .£ -^ ^ o " btr 4- a- a. r^3 ~ a: — "5 a rYflat5a-'^ r:: '"a. a -'" t- ' rt * O 3 jo' S a a a. av E- 4- ,fl- ^= a^ 4J.a 4- afli"s,-flii .a 3 a- a — a j, U O •. J3 5 i H C »f ! a:a: — ?-— a x a, t, s J "S x a .S a ^ a; g'^j"" T ?* "f* x f- a; Li- ° a -a a oj BH i. a "a s a ~ x — a u a b .Sgvfl'j SS'-a.a.a K w ^' , 3t4-r a Jt^a;;"'' t -4-;fl. fl^A-iw^; . 3 fl- r ■- C fl^^-£?a 4 ->^ i- x w ^3 n, C — a - a q, x 4 - , &4 at.a_- r -'w flaa£a.a=atter*4-^. n _ w a,4^ a.,C a*_ » e fl raa.ai=fl-a-^ v a a QJ 0) A* * A t, v ^'aa-^^t-. -r,-- — T:5aO ^4-E^a— a^_a-Sflao£a C -(* , o . _* flv ^ c» h C 4J a 4j a w K-a.X'-'-- ajs c c c ='J ;:: -a""fl:o--5^ a Ea.S^fl^.i-^-fH C M SflSl3?-I-".^fl -a R.S bi a bt is a. O x C > .5 aj aj a x z a- c- 4- >-, 4-> a- x z * a, s-. S +3-2,2 5 a a- a a AH H N X 0) J - dv — J^ fl V- O A- '- a- ■ a s ^ "*"55<»- ■s » "- s n- a.--;^ a x t- -— v x fla c a- B.2 tu-flfl'a S-S^w, 3 i- — a. > o '=£ a h =a ° a |g«-a a a— x SScfl ^cix' a a S'S' 3 fl-.dr ■ a a. 4_ x a; t £ ^^ a. - ^ H §^ ~3 °.S a r— i -Ffl *J -* fc t — a- " a a- t*. a; a, - o -w a- '""" . s«, ^ i, a a bi.a ja g 5i a. x a- . 'E S » a ?• a! a ° J- •,. o a- -- - x a/ "a —a:"S o =*-- a o o 3"S S a m £s a a53: «nio5 J 4J a h r-aaafl'- - o a a K .= 5? "a fl- 4-J ^ 0; t- X cm a a) a^ t- x • a^ a a fll, — .a « -fl- 1 g 3 a> a- 4^ =4a 5 a - a< a be a. . a. ^ r5 a a , ~» a aj t< o a) =■- a. > bi as bt i O 3 - 1 - ~ a. — 'o a a: a ^ si- M flj X — ct a. ; — a^ * a - ?, a ^ be ' 1 " a "t^ a^ , ■a a. ^ p a. oi 2 a ► be i?-3^|-Ffl--- .-=■£.= "j3 So a- g 5 g ^h Ol a- l^lfl a — .a a a, - aajeaa^t. Xr*B- a S liS a St a A a— a-ja-r'S a> o> a) gr - ; a) ' ■a w ao c-i -i; s x £ 2 a a S 0, 3 .a X a- t- r,fl > bo a> I* o u - -j> u -g -a .4-- a • ■S.S a x a =_£•- ■§ ■£ a n X ° _! a c x' g fl K s: '■+-> c x a "t- 5^ a a tl S bea-fa " be^ ! a be _ aa *;; ^ a. a a o " a — •, r i i X a fl 3^5 x"3 ~o a a a o - ° ? a= a g a X.- V ^'fl — a* a a- e- x a •S 5 a '■ 9M% X a; j_, cj * _c3 2 a x a g g .2 - 5 a. S:~„ S — - I*- x a H 3 = £ M a.x-r; ; j a.a a x a a- 5.'- <, a a. i. a « £ ' 3 ii a a a a.* x a, ' — -- ^r: tS EJ *jW be'x -a be a s a; a, ce - y . j- f. a. be — A- 0. fl t- ■3'"" a a a " fl > a, B — .a x j. x a ;a be 1 *- _a> fl 2 a. ■'fl ai c *" |i • X S-, +* \f **■ o-f 4- a — "7 s 3 .a^ a 2 . a A a a> '-> ~ — A a-r- 3 m M O 2 0j be-* cj be — c 1- . k •;ie C flSjfl — "a" — a- p t- a. fl /• 4- b ■- a. a. a. fl^N^'t — o o a ca , > a -a a x ,-1 ;, -_e j. a*- 1 a- a n.5-a m _ ~ a a a ■ « a, c A d ^ —- B 4- -a - T? a y a b 3 a b - be - a. I x'- 5 — a. a x — a 3 ^>. x a. t- a S Ai r- a. . be -a-. 2 — > x ' .-= --■ fl a-iCBrBflX q-i fl a. fl a y. - -™ b^c i'x-.^ SC+;C ^4-4. y. 'r: a x be —* c a a. 3 £ o J — 3 x _ a 2 x 5 i. ._ ., a. S- a " S o £ « ^>"5 a i:" - — fl— "* c a I .. aj8|! te e- S 2 a. - a a ~3---* a, a. fl 3~ a a x t- a ■ a. x* a: a 4a a. s - - — a a. - a •-»• "- a fl- a. b a -a f a. as .- » B "-g3.2 § • ^beOflB-^'afl 'x K 4_~aj- - a - a ja x a> 3 p— 1 — _^ — -a- -a ^ e- 'SB-S'- 5-a'S- a -a — a. - a 05*5 a> « a t) *•- S<<4. aHH i-fl-4- CALIFORNIA DIVISION OP MINES [Special Report 55 SUMMARY AND CONCLUSIONS Results of laboratory tests on mortar bars made with crushed Franciscan cherl and high-alkali cement show expansion owing to alkali-aggregate reactivity. Mortar bars made with similar samples of chert and medium and low alkali cements show no expansion. The reactivity presumably is due to the chalcedonic silica content as no opal can he identified. Recent studies have shown that alraosl all forms of silica if ground finely enough and if present in large enough proportions may cause reac- tion. Concrete structures made with aggregate contain- ing Franciscan chert do not show signs of reactivity. primarily because the chert is not present in large enough proportions to cause expansion. This is verified by their excellent service record. On this basis, commercial aggregate deposits in the Coast Ranges of California that have proven service records, should not be suspected of alkali-aggregate re- activity if the only reactive rock type present is the Franciscan chalcedonic chert unless the chert content exceeds '20 percent of the gravel and 10 percent of the sand. The amount and particle size distribution of chert in undeveloped deposits is of more importance than just the fact that chert is present. A reactive condition could be induced in an innocuous aggregate if a producer attempts to manufacture a concrete sand by crushing a coarse aggregate containing significant proportions of chert. (iood, serviceable concrete can be made with.an aggre- gate containing any proportion of Franciscan chert pro- viding the alkali content of the cement is controlled. REFERENCES Alderman, A. R., Gaskin, A. J., Jones, R. H., and Vivian, H. E., T.)47, Australian aggregates and cement in Studies in cement-ag- gregate reaction : Council Sei. Indus. Research Australia, Bull. 229, chapt. 1, p. 10. American Society for Testing Materials, l!)."i4, ASTM standards on mineral aggregates, concrete and nonbituminous highway ma- terials, p. 301. Bramlette, M. N., 1040, The Monterey formation of California and the origin of its siliceous rocks: U. S. Geol. Survey Prof. Paper 212, p. 50. silica : Faraday Carman, P. C, 1!»40, Constitution of colloids Society Trans., vol. 36, M). 964-973. Davis, 10. P., 1!)18, The radiolarian cherts of the Franciscan group: California Cniv., Dept. Geology, Bull., vol. 11, pp. 235-432; Folk, R. L., and Weaver, C. E., 1052, A study of the texture and composition of chert : Am. Jour. Sei., vol. 250, pp. 498-510. Gaskin, A. J., Jones, R. H., and Vivian, H. E., 1955, Studies in cement -aggregate reaction, XXI. The reactivity of various forms of silica in relation to the expansion of mortar bars: Australian Jour. Applied Sei., vol. 6, no. 1, pp. 7S-S7. Holland. W. Y„ and Cook, R. H., 1953, Alkali reactivity of natural aggregates in western United States: Min. Eng., vol. 5, pp. !)!)l-7. Kern]), J. P., 1940, A handbook of rocks, I), van Nostrand Co., Inc., Xew York, 6th ed. Klein, I. E. and Goldman, H. B., 1958, Sand and gravel resources of Cache Creek, Lake, Yolo and Colusa counties, California: Cali- fornia Jour. Mines and Geology, vol. 54, no. 2. Lawson, A. C, 1895, The geology of the San Francisco penin- sula : U. S. Geol. Survey Ann. Rept. 15, 1893-94, pp. 401-76. Lerch, Willliam, 1956, Chemical reactions : ASTM Spec. Tech. Pub. 169, p. 343. Lindgren, W., 1915, Processes of mineralization and enrichment in the Tintic mining district: Econ. Geology, vol. 10, p. 233. Mather, Bryant, 1951, Cracking of concrete in the Tuscaloosa lock : Highway Research Board Proc, vol. 30, pp. 218-233. MeConnell, D., Mielenz, R. C, Holland, W. Y., and Greene, K. T., 1950, Petrology of concrete affected by cement aggregate reaction : Geol. Soc. America Mem., Berkey vol., p. 234. Merriam, Richard, 1953, Alkali-aggregate reaction" in California concrete aggregates: California Div. Mines Special Rept. 27, p. 8. Midgley, H. G., 1951, Chalcedony and flint: Geol. Mag., vol. 88, pp. 17S-1S4. Pelto, C. R., 1956, A study of chalcedony: Am. Jour. Sei.. vol. 254. pp. 32-50. Powers, T. C, and Steinour, H. H., 1955, An interpretation of some published researches on the alkali-aggregate reaction, Part I — The chemical reactions and mechanism of expansion : American Concrete Inst., Jour., vol. 26, no. 6, pp. 497-516. Taliaferro, N. L., 1934, Contraction phenomena in cherts: Geol. Soc. American Bull., vol. 45, pp. 189-232. Taliaferro N. L., and Hudson, F. S., 1943, Genesis of the manganese deposits of the Coast Ranges of California : California Div. Mines Bull. 125, pp. 217-275. Trask, P. D., and Pierce, W. G., 1950, Geology of the Ladd- Buckeye area : California Div. Mines Bull. 152, pp. 211-228. Vivian, H. E., 1951, Studies in cement-aggregate reaction, XIX. The effects on mortar expanision of the particle size of the re- active component in the aggregate : Australian Jour. Applied Sei., vol. 2, no. 4, pp. 488-494. L9 2-59 3,500 printed in California state printing office I 4 -*?V >fl<*^