THE UN IVERSITY OF MICHIGAN COLLEGE OF ENGINEERING Department of Civil Engineering Progress Report EFFECT OF THE ADDITION OF FLY ASH ON THE SULFATE RESISTANCE OF CONCRETE F. E, Legg, Jro Associate Professor of Construction Materials and Materials Consultant Michigan State Highway Department R. H. Vogler Research Assistant and Physical Testing Engineer Michigan State Highway Department UMRI Project 2211 under contract with: THE DETROIT EDISON COMPANY DETROIT, MICHIGAN administered by: THE UNIVERSITY OF MICHIGAN RESEARCH INSTITUTE ATNN ARBOR December 1960

SYNOPSIS A study was undertaken by The University of Michigan to determine the effect of the addition of fly ash on the resistance of concrete to deterioration due to sulfate action. Specimens containing a variety of cement and fly ash contents, using two different fly ashes from the Detroit Edison COnipany's plants, were fabricated in 1954. Beams made from different concrete mixtures have been in three test exposures for approximately five years. One set of beams is in a solution of magnesium sulfate in the laboratory, a companion set is in a stream of "sulfur water," and a third set is exposed to a spray of raw sewage. Change in the concrete has been determined at intervals by measuring the sonic modulus of elasticity and weight of the bars. ii

INTRODUCTION Some investigators have found that the addition of fly ash to concrete reduces deterioration due to the action of sulfate waters and soils. One investigator has found regular concrete specimens to disintegrate in 18 to 24 months of exposure to strongly sulfate soils. As a result, a study was undertaken by The University of Michigan,.sponsored by the Detroit Edison Company, to determine the effect of fly ash typical of that produced by Detroit Edison when used in concrete exposed to sulfate action. In selecting mixes, it was hoped to provide sufficient variation in the fly-ash contents and cement contents to cover the range used in common practice, Limestone coarse aggregate with a good service record was used in all mixes. The fly-ash mixes were made with air-entrained concrete only. Evidence by other investigators shows air-entrained concrete to have superior resistance to sulfate attack. For comparison, concrete without fly ash was also made, with and without entrained air. MIX DESIGN The concrete mixtures were proportioned by the procedure given by the American Society for Testing Materials, "Tentative Method of Testing Air Entraining Admixtures for Concrete (ASTM Designation C 233-52T)." The only deviation was in the use of a fine aggregate slightly coarser than that specified. The fine aggregate was 41% of the absolute volume of the total aggregate in the mixes with entrained air and 45% in the mixes without purposely entrained air. Air-entrained concrete with three cement contents were used, namely, 4, 5, and 6 sacks per cubic yard. Two fly-ash contents with each of two fly ashes were used, along with no fly-ash control, for each cement content. In addition, mixes with nominal cement contents of 5, 6, and 7 sacks per cubic yard without fly ash and without air entrainment were also made, MATERIALS lo FLY ASH The fly ashes used were furnished by the Detroit Edison Company from the Trenton Channel station and from the Conners Creek station4 The Trenton ash 1

was furnished in January, 1954, and the Conners Creek ash in March, 1954, Physical tests and chemical analyses of these ashes have been previously reported but are here repeated in the Appendix, 2, PORTLAND CEMENT Three brands of cement available locally, namely, Huron, Peerless, and Wyandotte, were blended in equal weights to give an "anonymous" cemento This was done to reduce the effect of variations between different brands due to differences in physical and chemical characteristics, Chemical and physical tests of the cement are shown in the Appendix, 3, FINE AGGREGATE Natural sand from the Killins Gravel Company, located about 3 miles west of Ann Arbor, was used in all specimens. The gradation was slightly coarse compared to that recommended in ASTM Designation C 233. 4~ COARSE AGGREGATE Limestone coarse aggregate from the Inland Quarry, Manistique, Michigan, was used in the concrete, This was separated into four sizes - passing 1 inch and retained on 3/4 inch, passing 3/4 inch and retained on 1/2 inch, passing 1/2 inch and retained on 3/8 inch, and passing 3/8 inch and retained on No, 4 sieve, These were then recombined using equal weights of each size of stone, 5 AIR-ENTRAINING ADMIXTURE Neutralized vinsol resin in water solution was used as the air-entraining admixture, This was prepared from commercial powdered NVX, manufactured by the Hercules Powder Company, FABRICATION OF SPECIMENS American Society for Testing Materials methods were generally followed in the fabrication of the test specimens. The fine aggregate was air-dried before use by spreading in a thin layer on the laboratory floor. The coarse aggregate was dry as received, The us'e of dry aggregate aided in accurate determination of the water content of the mixes, The concrete was mixed in a Blystone mixer having rotating paddles on a 2

horizontal shaft. The mixer was "buttered" each day before using with a mixture of cement, sand, and water to coat the paddles and tub. The dry-concrete materials were weighed on an 800-lb-capacity Toledo scale and placed in the mixer in the following sequence: stone, sand, fly ash, and cement. The water was weighed on a 250-lb Buffalo scale. Most of the water was introduced before starting the mixer in order to reduce dusting of the dry materials. Simultaneously with starting the mixer, the vinsol resin solution was added, The mixing continued for 2 minutes during which time additional water was added to adjust the slump. Following the -minute mixing, the mixer was stopped and the concrete allowed to rest for 3 minutes. A 1-minute final mix followed the rest period. Occasionally small additions of water were made at the start of the final period. Water not used was weighed back so that the exact amount actually put in the batch could be determined, Following mixing, the concrete was dumped into a moistened flat pan and slump and air tests were conducted simultaneously. These were followed by a weight-per-cubic-foot determination, using a 1/2-foot calibrated measure. An Acme pressure air meter was used for the air content. The batch contained nominally 2.2 cubic feet of concrete, sufficient to make 6 test cylinders, 6 x 12 inches, and 6 beams, 3 x 4 x 16 inches. This furnished 2 cylinders for each age of 7, 28, and 90 days for determination of the compressive strength of the concrete, and 2 beams for each of 3 conditions of exposure to sulfate attacks Waxed cardboard molds with metal bottoms were used for casting the cylinders. Immediately after filling, they were covered with steel plates to prevent evaporation from the fresh concrete. The beams were cast in steel molds which were covered with wet burlap after the concrete had taken its initial set. The mixing was restricted to two batches per day by the number of steel beam molds available. A random sequence of making batches was followed with the one limitation that the corresponding mixes of Trenton and Conners Creek fly ashes were made on the same day. The beams were removed from the molds the day after making concrete, and together with the cylinders were placed in the moist fog room for curing. COMPRESSION-TESTING OF CONCRETE CYLINDERS When the concrete cylinders reached their designated ages for testing, namely, 7, 28, or 90 days, they were removed from the moist fog room, stripped of the cardboard molds and capped with gydrostone capping plaster on each end, Caps were cast against hardened, polished 1/2-inch-thick special steel-bearing plates. After allowing sufficient time for the capping plaster to harden, the cylinders were broken in a Riehle 300,000-lb testing machine. 5

DISCUSSION OF TEST RESULTS 1. COMPRESSIVE STRENGTH RESULTS Table I gives the mix data and the compressive-strength test results for 7, 28, and 90 days. The compressive strengths of the concrete made in this study are in good agreement with the strengths obtained in the previous studies of Trenton Channel and Conners Creek ashes in air-entrained concrete. This is somewhat surprising since the mixes in this study generally have a higher water-cement ratio than comparable mixes in the previous work. The increased water was required by the relative increase in the amount of sand, Up to twice the weight of sand per cubic yard of concrete was required by using angular instead of pebble coarse aggregate by the ASTM proportioning procedures. The additional sand required additional water to lubricate the fresh concrete, but in general a smaller slump was utilized thyn in the previous studies of Trenton and Conners Creek ashes. The reason for the low compressive strength of the 7-sack non-air-entrained concrete in relation to the corresponding 6-sack concrete is not apparent, 2, LABORATORY TESTING OF SULFATE RESISTANCE Two beams from each mix were used in the laboratory tests of the sulfate resistance of the concrete. When the concrete had reached an age of 88 or 89 days, the beams were removed from the moist fog room and were immersed in water so that they would be in the saturated condition that would be typical throughout the remainder of the testing sequence, At 90 days, the beams were weighed and the sonic modulus of elasticity was determined in accordance with ASTM Designation C 215. The beams were then immersed in a 10% magnesium sulfate Solution. Seven days after being placed in the solution and at intervals thereafter, the beams were removed from the solution, washed with fresh water to remove possible accumulation of salt, reweighed, and the sonic modulus was determined again. The concentration of the magnesium sulfate solution was checked weekly by means of a hydrometer and corrected whenever necessary. After 55 months in the magnesium sulfate solution, the beams have generally increased in sonic modulus, indicating internal strength gain, but at the same time have lost weight due to surface disintegration, as indicated in Table II. Weight losses are fairly small in 55 months for all the specimens not containing fly ash-either air-entrained or otherwise. In almost all cases, specimens containing fly ash show greater surface disintegration~ The trend

TABLE I CONCRETE MIX DATA AND COMPRESSIVE-STRENGTH TEST RESULTS Nominal th Date Fly Ash, Actual Material Proportions, Wt Fresh Pressure Vinsol Compressive Strength, Cement Content, No. Made, lb/cyd, Cement Content, lb/cyd gal/k Concrete, Air Content, i Resin, sk/cyd 1954 and Source sk/cyd Sand StoneI Net Water lb/cu ft per cent lb/cyd 7 days I 28 days I 90 days 127 10- 4 0 4.00 1330 1959 238 7.13 145.7 4.5 2.25 0.044 1820}i89 27901}0 345 1960 J 26701 3000 4 750 115 9-22 150 4.02 1268 1867 229 6.88 145.9 4.6 2.25 0.135 26152620 4010 620 Trenton 2630 4170 J 4610 1 4 116 9-22 150 4.01 1269 1869 240 7.21 146.2 4.4 2.5 0.578 2315 15 33551 Conners Ck 2315 I 3885 J 449011 124 9-30 250 4.02 1217 1791 254 7.62 145.9 4.1 2.5 0.245 1855}1 5 3 }45 0}429 Trenton 1835 305590 131 10- 7 150 5.0o4 1215 1788 2675 6.11 145.8 4.4 30. 0.147 2915}2790 3995} 5 41 5 132 10- 7 150 5 ~3 1207 1778 267 6.42 145.4 4.5 3.5 0.565 2970 12960 36g05 }75 50oo }5185 Conners Ck 2950 3605 5750 5500 122 9-27 250 4.99 1171 1724 274 6.59 144.8 4.4 3.25 0.245 245 2470 35 Trenton 24go91 3655.1 44:5 121 9-27 250 4.99 1151 1695 288 6.90 143.4 5.4 4.25 0.982 2580 2535 355 380 70 05}47 118 9-23 0 6.09 1265 1862 249 4.99 149.2 3.9 3.0 0.065 385}3530 4 }445 1}525 348o1n 495 45J 0 80 119 9-24 100 6.08 1216 1791 250 5.00 148.2 4.1 3.0 0.123 37620 4755 } } Trenton 33750 45051 5620 6 120 9-24 C Ck 5.92 1206 1776 265 5.29 143.8 6.2 3.0 0.540 3285 145 }381550 } Conners Ck 30051 4o45 130 10-6 Trenton 6.o6 1129 1662 286 5.72 144.6 4.0 3.75 0.196 2755}2870 3920 9 }5080 ~~~~~~~~~~~~~~~Trenton 2985~~ 1 LV IV ~ 40~9 0 507180 58 p9 Conr6 200 6.oo 1120 1650 302 6.o4 143.0 5-1 4.5 0.786 26580 2765 3780 }8 5~ 5 Conners Ck 288o 3995^ J }}} 5 128 10- 4 0 4.81 1519 1900 268 6.42 149.4 1.4 2.5 12405}2290 33175} 445}4140 2175 321756 6 1126 10- 1 5.89 1458 1823 254 5.09 1503 1.9 2.0 } 36950^ }^1.}22835 5440}5 90. 44355 7 125 10- 1 0 6.83 1421 1775 274 4.71 150.4 1.8 3.0 3500 5850} 538175.~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~4. 4.1.0 0.1.. 23...

TABLE II CHANGE IN SONIC MODULUS AND WEIGHT OF BEAMS AFTER 55 MONTHS IN LABORATORY MAGNESIUM SULFAE SOLUTION Cement Fly-.Ash Content, Loss in Increase in Sonic Content, lb/cyd, and Source Weight, Modulus of Elasticity, sk/cyd per cent per cent 0 2,8 17.5 2o3 (-1.2) 150 - Trenton 5o3 2,7 5o3 5,8 4 150 - Conners Creek 7.0 0.9 8.7 2o7 50 - Trenton 12.2 1.9 14 6 (-0o8) 250 ^ Conners Creek 5.2 6.7 8.8 2.6 0 4,7 4,8 5.7 6o9 150 Trenton 5o2 9o8 4,0 504 5 150 - Conners Creek 6.9 12,4 6,1 6,9 250 - Trenton 9 4 8o0 8,0 5.7 250 Conners Creek 7o5 10o6 6.7 7.3 0 2,5 5,6 2,9 4.9 100 - Trenton 4,5 (-0o6) 4~9 0.2 6 100.- Conners Creek 1 3 9,3 0.3 7.6 200 - Trenton 5 9 7 3 5,4 8.7 200 - Conners Creek 5.1 7o4 4.3 2.8 5 (Non-Air- 0 309 15,0 Entrained) 3 2 lo 2 6 (Non-Air- 0 2,0 10 4 Entrained) 1,7 8,2 7 (Non-Air- 0 2.7 7.7 Entrained) 1.9 5,8 6

of these data is perplexing in two aspects and is not in accord with that of two recent investigations-. 2 lo Weight losses for all the concretes are far less than would be expected from such a prolonged exposure. 2, Improvement in behavior is not indicated for the specimens containing fly ash. For instance, Ref. 1 (discussion, p. 1019) shows 5-sack concrete with a high C3A cement to have a 25% weight loss in only 9 months under similar exposure. Reference 2 shows superior performance for concrete in which some of the cement is replaced by a pozzolan when complete submersion was used, The effect of a pozzolan is more controversial if periods of drying are introduced It has been speculated that in the present investigation the concrete, for some reason, is inherently much more resistant to sulfate attack or that a difference in manner of handling the solution is the reason for such a disparity of results. The latter seems more plausible but is not well understood, In other investigations, the sulfate solution, generally, was completely changed once monthly. In the present tests, the solution was renewed only once, at age of 8 months, but weekly hydrometer checks were made to insure a specific gravity of close to 1.103. Either sulfate crystals or water was added, as necessary, to adjust the specific gravity, The specimens were stored upright in two drums with covers, each containing 18 bars and requiring 18 gallons of solution to submerge the specimens. This provided a somewhat greater volume of solution than specimens. Approximately 33.9 lb of epsom salts (MgS04 a 7H20) were required in each drum to provide the 10% concentration of MgS04. Approximately 300 lb of concrete were in each container so that the amount of salts (as MgS04) available to react is about 5% of the weight of concrete or, on the average, about 50% by weight of the portland cement in the bars, Neither actual specific gravity readings that were taken nor our knowledge of chemistry would lead us to believe that this amount of sulfate would be actually consumed and thus that entire replacement of solution at intervals would be necessary, Prolonged storage in the stagnant sulfate solution would probably cause leaching of the calcium hydroxide from the concrete, and it has been speculated that the presence of this lime may depress the number of available sulfate ions which cause destruction, In usual field exposures, sufficient solution would normally be available to carry off the leached lime and to furnish fresh sulfate ions. In this sense, the present laboratory exposure may not be realistic. The data actually give some support to an entirely contrary viewpoint proposed by Benton,3 who, after presenting some basic experimental background work, postulates that a pozzolan containing alumina should be less sulfate resistant when combined with a high C3A cement, Only the entirely siliceous pozzolans should improve sulfate resistance with such cements 7

3. FIELD TESTING OF SULFATE RESISTANCE Companion concrete beams were placed in two field exposures, In one case, the beams were placed in a running stream of sulfur water and in the other case, the beams were placed in a moist atmosphere in the grit chamber of a sewage treatment plant. SULFUR WATER EXPOSURE The pool of sulfur water where the beams were immersed is spring-fed and is located in the bottom of the Sibley limestone quarry at Trenton, Michigano They were completely immersed and freezing does not occur in the running water during the winter. The beams were laboratory-cured in excess of 90 days in the moist fog room. On January 5, 1955, they were placed in a tank of water where they remained for two days. The sonic modulus was then determined and the beams were individually weighed The beams were then returned to the moist room until January 19, 1955, at which time they were transferred to the sulfur water pool described above. The beams were brought to Ann Arbor on July 20, 1955, for weighing and sonic modulus determination and returned to the pool on July 22. When the beams were removed from the pool, they were covered with heavy sulfur "streamers"t and required thorough cleaning, Repeat weight and sonic modulus determinations were made on January 19 and July 10, 1956; August 5, 1957; and July 31, 1959; Table III shows the total changes in weight and sonic modulus of elasticity, expressed as per cent, that have taken place in the 54 months that the beams have been in the test exposure0 It is observed from the data that all the airentrained concrete mixes without fly ash have experienced a slight weight loss and all containing fly ash have gained weight0 The pattern of behavior is so regular as to predict positive effects from the fly ash which will be more pronounced within the next few years. The non-air-entrained concretes do not yet exhibit a positive trend but are certainly no less favorable than the airentrained at this age. Sonic modulus determinations are inconclusive since practically all specimens show increases, thus suggesting internal strength gains during the exposure periods SEWAGE SPRAY EXPOSURE The beams in the sewage spray exposure are located in a gate well at the outlet from the grit chamber at the Sewage Treatment Plant of the City of Detroit, Michigan~ There is considerable turbulence in this gate well, creating 8

TABLE III CHANGE IN SONIC MODULUS AND WEIGHT OF BEAMS AFTER 54 MOTHS IN "SULFUR WATER" EXPOSURE Cement Fly-Ash Content, Change in Increase in Sonic Content, lb/cyd, and Source Weight, Modulus of Elasticity, sk/cyd per cent per cent 0 -0.o4 8.5 -0o2 7~5 150 - Trenton +0 3 7 +0.2 6 4 150 - Conners Creek +0.1 305 +0.2 2,5 250 - Trenton +0.2 7,5 +0.2 6.5 250 - Conners Creek +0,2 6.5 +0,2 5,5 O -Oo. 0o5 -0,1 (-0.5) 150 Trenton +o-)5 4 5 +0o4 6 5 150 - Conners Creek +0.3 7 +0,5 6.5 250 - Trenton +0o 3 7 +0,2 5.5 250.- Conners Creek +0.3 8 +0.5 6.5 O -0.1 1 -0o4 1 100 - Trenton +0.2 0 +0.1 1.5 6 100 - Conners Creek +0o6 705 +0o6 6.5 200 Trenton +0, 4 3 +Oo. 4 45 200 4- Conners Creek +0 7 4 +0o7 2,5 5 (Non-Air- 0 -0.1 4.5 Entrained) 0.0 6 6 (Non-Air- 0 +0,1 5.5 Entrained) +0.1 4,5 7 (Non-Air- 0 +0.2 9 Entrained) 0o0 9 9

a very moist atmosphere. The space is enclosed so that there is no danger of freezing the beamso As in the case of the specimens in the sulfur water exposure, the beams were cured in the laboratory moist room in excess of 90 days. On January 6, 1955, they were weighed and the sonic modulus was determined. They were returned to the moist room until January 10, 1955, at which time they were transported to Detroit and placed in the test exposure. On July 19, 1955, the beams were returned to Ann Arbor for reweighing and sonic modulus tests. They were returned to the test exposure the following dayo The bars were observed to be quite dry at this examination since the summer heat reduced the moisture content of the air considerably. Tests on the bars were repeated on January 17 and July 2, 1956; August 7, 1957; and July 31, 1959. The total change in weight and in sonic modulus of elasticity after 54 months exposure, expressed as per cent, is shown in Table IV. As in the case of the sulfur water exposure, there is some tendency for the air-entrained concrete without fly ash to show a greater weight loss than that containing fly ash. Several more years of exposure may be necessary to confirm this trend positively. Again, the weight loss results from the non-air-entrained concrete without fly ash are inconclusive, All specimens show increase in sonic moduluso SUMMARY Reported herein are three series of sulfate resistance tests of concrete containing fly ash from the Detroit Edison Company, one laboratory exposure and two field exposures. The laboratory exposure, terminated after 55 months, tended to show no benefit to sulfate resistance by incorporating fly ash in the concrete. However, it should be emphasized that the laboratory exposure may have induced quite opposite effects from those that would normally be experienced in field concrete, A standard laboratory procedure was not available at the time these tests were started, Constant agitation and frequent changes of the solution would be suggested as alternative procedures if future laboratory determinations were to be conducted, Not unexpectedly, sulfate deterioration in the two field exposures has been slow, and only faint trends are discernible at 4-1/2 years. Weight-loss changes of the specimens indicate that the air-entrained concrete containing fly ash possesses superior resistance It seems possible that at least another five years of exposure in the field will be required to demonstrate positive trends, particularly to distinguish preferable fly ash or cement contents, 10

Custody of the raw data from the field exposures is being transferred to the Edison Company, which may continue making periodic measurements of the field-exposure specimenso REFERENCES 1o Eo C. Higginson and 0 Jo Glantz, "The Significance of Tests for Sulfate Resistance of Concrete." Proceedings, AST, 53, 1002-1020 (1953) 20 Mo Polivka and E, H, Brown, "Influence of Various Factors on Sulfate Resistance of Concretes Containing Pozzolan," Proceedings, ASTM, 58 10771100 (1958) 35 Jo Benton, "Cement-Pozzolan Reactions," Hihway Research Board Bulletin, No. 239, ppo 56 65 1959.

TABLE IV CHANGE IN SONIC MODULUS AND WEIGHT OF BEAMS AFTER 54 MONTHS EXPOSURE IN SEWAGE PLANT Cement Fly-Ash Content, Change in Increase in Sonic Content, ib/cyd, and Source Weight, Modulus of Elasticity, sk/cyd per cent per cent 0 -03 9 -lol 3~5 150 - Trenton -0 2 6 -0o4 6 4 150 - Conners Creek -0o8 505 -0o4 2,5 250 - Trenton +0ol 605 -0o7 2 250 - Conners Creek -0o2 9 -0.5 5 0 -lo0 5.5 -0o8 4~5 150.- Trenton 0.0 7 5 -0o. 7 5 150 - Conners Creek +0,2 10 Oo0 5.5 250 - Trenton -0o7 11.5 -Oo5* 5.5 250 - Conners Creek +0o4 10,5 -Ool 8.5 0 -0.6 5.5 -0o6 6.5 100 - Trenton -0 4 4,5 -o05 4 6 100 - Conners Creek -0ol 7.5 -0.3 6.5 200 - Trenton +0.5 9 -0.3 7 200 - Conners Creek +0 2 7 +Ool 5~5 5 (Non-Air- 0 0o0 11 Entrained) -o 9* 8 5 6 (Non-Air- 0 -0o6 705 Entrained) -1o0 6 7 (Non-Air- 0 -0o3 13 Entrained) -0 5 7 5 * Corner broken off in handlingo 12

APPENDIX

PROPERTIES OF FLY ASH Trenton Conners Channel Creek Physical Properties 54C-159 54C-295 Specific surface, air permeability test, sq cm/gram 2960 3476 Compressive Strength9 25% by weight of cement, sand replacement, machine mixing, 73 F cure, per cent of control 7 days 143 147 28 days 142 143 90 days 155 145 Water requirement, per cent of control 100 102 Drying shrinkage, 28 days, per cent 0o09 0o09 Autoclave expansion, per cent 0,02 (-)0o01 Specific gravity 2,42 2047 Chemical Properties, per cent Silicon dioxide, SiO2 46.o 40,7 Aluminum oxide, A1203 2707 21,9 Ferric oxide, Fe20s 17o0 21.9 Calcium oxide, CaO 2o6.17 Magnesium oxide, MgO 0 9 0o9 Sulfur trioxide, SOs 0o7 0.7 Loss on ignition 2o9 9o9 Moisture 0 2 0 3 14

PROPERTIES OF CEMENT (54C-158) Physical Properties Specific surface, air permeability test, sq cm/gram 3133 Autoclave expansion, per cent o08 Normal consistency, per cent 24.8 Time of set, Gillmore Initial 4 00 Final 6 o00 Compressive strength, psi Hand Mix Machine Mix 7 days 3129 3283 2950 28 days 4463 4650 4271 90 days 5050 5379 4438 Tensile strength, psi 7 days 358 28 days 462 Air in mortar, per cent 12 0 Chemical Properties Per cent by Weight Silicon dioxide, Si02 21o0 Aluminum oxide A1203 5 9 Ferric oxide, Fe203 3o0 Calcium oxide, CaO 61 9 Magnesium oxide, MgO 2,3 Sulfur trioxide, SO3 2,2 Loss on ignition 1l5 Sodium oxide, Na2O o049 Potassium oxide, K20 0o68 Tricalcium silicate, 3CaOSi02 42 Dicalcituii silicate, 2CaOSi02 29 Tricalcium aluminate, 3CaO0A120 11 Tetracalcium aluminoferrite, 4CaO Al20Oo Fe20O 9 Total alkali expressed as Na2O 0o94 15