Effect of Minor Elements on Clinker and Cement by Javed I. Bhatty

Senior Scientist, CTLGroup, 5400 Old Orchard Road, Skokie, Illinois 60077 847.972.3082, jbhatty@ctlgroup.com

The effect of minor elements on the production and performance of portland cement have been presented in this report. The elements studied are lead, molybdenum, antimony, copper, cadmium, beryllium, titanium, vanadium, chromium, manganese, nickel, and zinc. The elements are added at 0.5% and 0.25% by mass as single metals and multiple metals to the raw mix. To simulate practical scenarios, industrial by-products with known metal concentration, such as copper slag, sinter mix, and catalyst fines have also been studied.

Clinkers produced with single metal exhibited increased hardness. The clinkers did not exhibit any burnability problems; distribution of C3S, C2S, C3A, and C4AF were within the normal range without any free-lime increase. However, titanium and antimony enhanced the formation of C3A at the expense of C4AF, whereas barium and beryllium increased C4AF and reduced C3A. Alite crystals in most clinkers were decomposed and clustered as opposed to isolated, individual crystals. The belites were rounded but did form ragged edges. A majority of the metals were found concentrated into the aluminoferrite phases except vanadium, and possibly chromium and lead, which were equally distributed in the aluminoferrite and silicate phases. Zinc, copper, titanium, and beryllium delayed cement hydration and caused longer setting times that affected the strengths. Beryllium at 0.5% addition was a severe retarder that gave long setting times and very low strengths.

Clinkers made with multiple metals were also harder to grind. The distribution of major phases – C3S, C2S, C3A, and C4AF was appropriate with negligible free lime; the exception being nickel-vanadium-antimony that enhanced C3A at the expense of C4AF. Nickel-vanadium-antimony also formed irregular belite with wide lamellae. Other multiple metals decomposed the alite crystals, and produced slightly irregular and ragged belite, except copper-nickel-zinc which had very little effect on both alite and belite formation. Multiple metals imparted varying effects on the early hydration of cement, but all produced high early strengths except for nickel-vanadium-antimony, which delayed setting and reduced cement strength; copper-lead-zinc accelerated the setting.

Addition of industrial by-products such as copper slag, sinter mix, and catalyst fines affected the crystallinity of clinker. Alite crystals were decomposed and cannibalized, and belite crystals were irregular with ragged edges. The clinkers were harder to grind. Paste strengths with all by-products were notably reduced.

Reference:: Bhatty, Javed I., Effect of Minor Elements on Clinker and Cement Performance: A Laboratory Analysis, RD130, Portland Cement Association, Skokie, Illinois, U.S.A., 2006, 99 pages.

EXECUTIVE SUMMARY

In this report, the effects of minor elements on laboratory-scale production and performance of portland cement have been presented. Emphasis is given to elements that are commonly found in naturally occurring raw materials, fossil fuels, and industrial by-products used in the manu-facturing of cement. The elements studied are beryllium, titanium, lead, antimony, vanadium, chromium, molybdenum, manganese, nickel, copper, zinc, and cadmium. In order to avoid interferences, pure oxides of these elements were used as a raw feed component at levels of 0.5% and some 0.25% by mass. The effects were studied when the trace metals were added as single metals as well as multiple metals to the raw mixes. Type I clinkers (general-purpose clinker) were prepared for these studies. The effects of metals on clinker formation, phase modi-fication, and metal distribution in the phases were studied using XRD, XRF, microscopy, and chemical extraction techniques; the grindabilities of clinkers were determined by a CTL-devised method. The cements prepared from these clinkers were studied for their setting behavior, early stiffening, paste strength, mortar strength, and flow behavior. To simulate practical scenarios, selected industrial by-products commonly used in the cement industry with known trace metal contents—copper slag, sinter mix, and catalyst fines were also studied as raw feed components. Their effects on clinker properties and cement performance were investigated.

The following conclusions were drawn from these investigations: SINGLE-METAL ADDITIONS

In general, clinkers produced from the single-metal additions exhibited reduced grind-ability; the hardness generally increases with increasing metal concentrations. However, manganese, zinc, and molybdenum have minimal effects.

It also may be inferred from the XRD and XRF data that the clinkers produced with single metal additions did not show any unusual burnability problem; the contents of major phases, i.e., C3S, C2S, C3A, and C4AF were well within the normal range, and there was no unusual increase in free lime. However, upon comparing the XRD patterns to that of the control clinker, titanium and antimony appeared to have enhanced the formation of C3A at the expense of C4AF, whereas barium and beryllium increased C4AF at the expense of C3A. It also might be mentioned that the XRD tests were qualitative, and the Bogue analyses did not indicate any unusual variation in the respective C3A and C4AF phases.

At higher concentration, although zinc delays the hydration and causes longer setting time, the overall paste and mortar strengths are unaffected. Beryllium is a severe retarder and consequently gives extremely long setting times and very low strengths for both pastes and mortars; it also is an early stiffener (but not a paste/mortar setter) of the pastes that unusually affected the early hydration. Titanium at higher concentration also adversely affects the early and late strengths of both paste and mortar.

The alite crystals in most clinkers made with single metal additions are decomposed and cannibalized, i.e., clustered together as opposed to isolated, individual crystals. The

Effect of Minor Elements on Clinker and Cement Performance: A Laboratory Analysis

The alite crystals in most clinkers made with single metal additions are decomposed and cannibalized, i.e., clustered together as opposed to isolated, individual crystals. The belites are rounded but do form ragged edges. The bulk of clinker and cement properties, such as grindability and early hydration (described above), are related to alite and belite crystal habits.

A majority of the metals concentrate into the aluminoferrite phases except for vanadium, and possibly chromium, and lead, which are roughly equally distributed in the alumino-ferrite and silicate phases.

MULTIPLE METAL ADDITIONS

As is the case with single metals, the clinkers made with multiple metals are harder to grind. They are Type I clinkers with appropriate C3S, C2S, C3A, and C4AF contents and negligible free lime, the exception being the addition of Ni-V-Sb that enhanced C3A at the expense of C4AF.

It also appears that the addition of copper-nickel-zinc does very little to the formation of alite and belite, whereas nickel-vanadium-antimony only affects the belite formation by making the crystals somewhat irregular with wide lamellae. All other multiple metal additions result in decomposed and cannibalized alite, and slightly irregular and ragged belite crystals.

In general, the multiple metal additions have little effect on the early hydration of pastes, but they tend to give high early strength to both paste and mortars. Their 28-day strengths are close to those of the control pastes, but exceed them in the case of mortars. However, the addition of copper-lead-zinc accelerates the paste hydration and enhances early strengths of both paste and mortar; intriguingly, though, the 28-day strength of paste is lower, and that of mortar is higher, than the respective controls.

INDUSTRIAL BY-PRODUCT ADDITIONS

The by-product materials were added at twice the amount of that typically used in cement plants to enhance any subtleties in the effects on clinker and cement performance.

Copper Slag: A 3% addition of copper slag did not affect the relative contents of C3S and C2S contents. However, the crystallinity of the clinker was affected; in particular, the alite crystals were decomposed and cannibalized, and the C3A size was small. The clinker was harder to grind. Cement pastes showed an overall reduction in strength, but the 28-day mortar strength was higher than the control.

Sinter Mix: A 8% addition of sinter mix affected the crystallinity of clinker; the alite crystals were decomposed and cannibalized, whereas belites were slightly irregular with some ragged edges. The clinker was harder to grind. Cement prepared from the clinker showed mild early stiffening that gave lower strength pastes, but the mortar strengths were better

mild early stiffening that gave lower strength pastes, but the mortar strengths were better than the control.

Catalyst Fines: An 8% addition of catalyst fines also produced slightly decomposed and cannibalized alite; belite crystals were slightly irregular with some ragged edges. Clinker was hard to grind. The cement prepared from the clinker showed low early strength, but the 28-day strength was very close to that of the control. The mortar showed slight early stiff-ening tendency (low flow), but this did not affect its strength behavior as the mortar strengths were comparable to the control.

It must be pointed out that for all the by-products studied, the rates of strength gain for pastes (from 3 to 28 days) were significantly higher than those of the mortars.

The following summary Tables I through VII show the effects of minor elements on the production and performance of portland cement. For comparison, the effects are shown as a percentage of the control, where the control is 100.

INTRODUCTION

As a consequence of decreasing availability of quality raw materials in the manufacturing of cement, the use of alternative raw materials and secondary fuels derived from industrial by-products is on the rise. The aim of such practices is to conserve the raw material resources and reduce environmental stress by controlling waste accumulation without affecting the production and quality of cement. However, the use of such materials, or for that matter even marginal natural sources, may cause concerns regarding the incorporation of trace elements into clinker and their potential effects on the performance of cement. These effects are, to a large extent, dependent upon: a) the type of trace elements, b) their concentrations in the kiln feed and in the fuel, c) their behavior at high temperature (such as volatility), and d) the degree of retention in clinker. The type and concentrations of metals present also can affect the burnability, early hydration, setting, and strength of cement.

A brief overview on the source of minor elements in cement manufacturing is given here.

SOURCE OF TRACE METALS

Trace metals in cement come from both the primary and secondary raw materials, and fuels used. Metals found in the primary raw materials, i.e., limestone and clay/shale, are given in Table 1.1.

Trace metals also come from widely used auxiliary raw materials such as blast furnace slag, fly ash, silica sand, iron oxide, bauxite, and spent catalysts. Some specific examples of trace elements (as oxides) found in blast furnace slag and fly ash are given in Table 1.2.

A secondary but important source of minor elements is industrial by-products substituted for a portion of the primary fuel. Trace elements found in a conventional kiln fuel (coal), along with secondary fuels (petroleum coke, used oil, and scrap tires), are listed in Table 1.3.

Petroleum coke frequently contains up to 4% sulfur and 0.1% vanadium oxide (V2O3), and can contribute potentially significant levels of sulfur and vanadium to clinker when used as a supplement to coal. Scrap tires have a zinc content of 1.0% to 1.6%. However, if tires replace 10% of the primary fuel, the resulting zinc contents in clinker are increased by only 0.1% to 0.16%.

Sources of chromium are the kiln refractory and grinding media such as liners and grinding balls during the milling operation. A damaged chrome refractory lining can enter into the incoming raw mix and incorporate a detectable amount of chromium into the clinker. For this reason, the use of chrome-containing bricks is being phased out in many parts of the world.

The level of trace elements in portland cement produced in North America varies from plant to plant. The concentration ranges of various minor elements found in portland cement from North America are given in Table 1.4. No distinction is made here between the plants using secondary materials and/or alternative fuels and those which are not.

Industrial By-Products as Raw Feed Components

Blast furnace slag has been used as an auxiliary raw material in quantities up to 30%. The level of use may be reduced due to its magnesium oxide (MgO) content, particularly if the MgO level is already high in the raw materials. Ostrovlyanchik, et al. (1986) reported that the use of fly ash from coal power plants as raw material, in place of argillaceous material, was effective for improving the wet process kiln output with savings in fuel consumption. Wastes from the soda ash industry, mixed with fly ash and limestone, have been used as components of raw feed (Patel, 1989). Phosphogypsum has been used as a source of lime in kiln feed (Toit, 1988). In separate studies, spent clays from lubricating oil refining, also have been tested as raw feed components in clinker production (Midlam, 1985). Bhatty, et al. (1985) produced an ASTM C 150 Type I cement from copper-nickel and taconite mine tailings used as a partial substitute for cement raw feed. The resulting cement developed comparable microstructure and strength properties to that produced with the conventional feed. Sewage sludge as a partial kiln fuel was reported by Obrist (1987). Heavy metals in the sludge were permanently withdrawn from the biosphere with little toxic emission. De Zorzi (1988) reported Italian experience on the use of municipal solid waste as a partial fuel substitute in cement manufacturing. The chemical and physical characteristics of clinker and cement were comparable to those produced conventionally.

Industrial By-Products as Supplemental Fuel

Huhta (1990) surveyed some 20 North American cement plants operating with by-products as supplement fuel and noted that the predominant material being used was waste oil, followed by solvent-derived fuel and scrap automobile tires; wood chips and fluid coke also were mentioned. However, tire-derived fuel may be most advantageous because of its availability and easy handling, and good fuel value, typically 33,500 kJ/kg (15,000 Btu/lb) compared to 26,800 kJ/kg (12,000 Btu/lb) for coal.

Weatherhead and Blumenthal (1992), of the Scrap Tire Management Council, confirmed from a recent field study that used tires can be used successfully as an alternative fuel in cement kilns. Use of scrap tires and lubricating oil, as alternative fuel in cement manufac-turing, has resulted in substantial reduction in manufacturing costs, improvement in waste minimizing, advances in resource recovery, extended life of existing landfills, and better environmental control (McGrath, 1993; and Heron, 1993).

Hansen (1993) has strongly advocated the use of solid wastes in cement manufacturing, emphasizing potential environmental and material advantages. It was suggested that using wastes as fuel has two-fold environmental benefits: 1) conserving fossil fuel and its transportation, and 2) reducing emissions that would have resulted by disposing of these materials by treatment or landfilling.

Although there are opportunities for beneficially using by-products in cement production, their total substitution in the industry is still in the experimental stages. One recommen-dation has been to limit the use of such materials to 5% by mass of raw feed (Vogel, et al., 1987). Huhta (1990) and von Seebach, et al. (1990) have projected using 20% to 30% or even more of the by-products as fuel replacement in cement kilns. At least one kiln in the U.S.A. has used 100% waste-derived fuel. Nonetheless, the level of application and degree of success largely depends upon the waste composition in terms of the fuel value, type, and concentration of trace elements.

SCOPE OF PRESENT INVESTIGATIONS

In order to fully understand the impacts of using by-products in cement manufacturing, it is important to critically study the effects of trace elements found in these materials before a rational conclusion can be deduced as to the pros and cons of waste utilization. Trace elements found in commonly used by-products were selected for study. The elements are: beryllium, tin, lead, antimony, vanadium, chromium, molybdenum, manganese, nickel, copper, zinc, and cadmium. The elements were blended as oxides in laboratory-prepared raw feed using reagent grade materials. The raw feeds were fired in a laboratory furnace to prepare clinkers. The clinkers were characterized for their burnability, grindability, and phase formation, and the cements made from them were tested for their setting, hydration, and strength characteristics. This battery of tests was aimed at determining the effects of minor elements on clinker making, clinker properties, and the properties of the resulting cement. Type I clinker were prepared and evaluated for these investigations.

The studies have been divided into the following major tasks:

Task 1: Conduct a literature review on trace elements-clinker interaction

Task 2: Produce and evaluate clinkers containing single elements at 0.5% by weight

Task 3: Produce and evaluate clinkers containing single elements at 0.25% by

weight of selected single elements that demonstrated strong effects in Task 2 Task 4: Produce and evaluate clinkers with raw feed containing multiple metals Task 5: Produce and evaluate clinkers using selected industrial by-products in raw feed

Each task addresses a specific issue on the subject of trace element versus clinker performance. Details on these tasks, including the material, methodology, experimen-tation, results, and discussion are given in the following sections of the report.

TASK 1 A LITERATURE REVIEW ON THE INTERACTION OF TRACE METALS AND CLINKER

A literature survey on the role of trace elements in cement manufacturing and use was carried out. The search was published in 1995 (Bhatty, 1995). The report identified the knowledge gaps and brought together the currently available data on the interaction of minor elements and clinker.

The review indicates that although some selected elements have been studied in detail for their role in the clinkering process and cement properties, other elements, including some found in raw materials, alternative materials, by-products, fuels, or secondary fuels, have been studied far less.

It appears that elements such as chromium, vanadium, titanium, nickel, barium, manganese, copper, lead, zinc, cadmium, molybdenum, antimony, cobalt, and beryllium, by virtue of their frequent presence in the by-products, auxiliary, and even natural raw materials used in cement manufacturing, will have potential significance. Metals like antimony, cobalt, nickel, beryllium, and titanium may have environmental impact because of potential toxicity, whereas cadmium and lead also might have greater emission rates.

Some of these elements, particularly those frequently found in by-products, can be investi-gated in single and multiple modes to better understand their effects in cement manufacture. Subsequently, some surrogate by-products can be simulated from different combinations of selected elements and used as partial/total raw feed or fuel in cement making.

In the subsequent tasks, these metals have been added in the raw mix as single metals and multiple metals to prepare clinkers for evaluation of their role on clinker and cement properties. Details of these tasks are given in the following sections.

TASK 2 PRODUCTION AND EVALUATION OF CLINKERS AND CEMENTS MADE WITH SINGLE METALS AT 0.5% BY MASS OF CLINKER

Trace elements frequently encountered in cement manufacturing, i.e., those found in typical raw materials and commonly used by-products, were selected for study of their effects on clinker production and properties of cement prepared. Included were:

Antimony           Copper                Titanium

Barium                Lead                     Vanadium

Beryllium            Manganese          Zinc

Cadmium           Molybdenum Control (no metal addition)

Chromium          Nickel

MATERIALS AND METHODS

Raw Feed Preparation

Each metal was added as an oxide to separate raw mix at 0.5% by mass of clinker. Addition of 0.5% metals was deliberate to simulate extreme case scenarios. Reagent grade materials, including powdered CaCO3, SiO2 (high purity, ground quartz), Al2O3 (calcined alumina), Fe2O3, and MgCO3, were used to aim for Type I clinker. The following oxides were used for each metal.

Amounts of various components needed to prepare raw feed for a Type I clinker are given in Table 2.1, along with the amounts of CaCO3, SiO2, Al2O3, Fe2O3, and MgCO3 needed to produce 1 kg of clinker. Raw mix for the control clinker (i.e., containing no metal) also is included in the table. The amounts of metal used in each raw mix are given in Table 2.2.

Each raw feed was hand-mixed in a large bowl and later blended by intergrinding in a ball mill for one hour. The blended material was removed from the ball mill and compacted into 50-38 mm diameter pellets using about 22 MPa pressure. Each pellet weighed roughly 150 g. The pelletizing was necessary to bring the reactants into close contact to promote the clinkering reaction.

Firing of Clinker

The pellets were loaded in platinum crucibles and fired for one hour in a muffle furnace pre-heated to 1450°C under ambient atmosphere. While the sample is loaded, the furnace temperature would drop to 1350°C but would rise back quickly to 1450°C within 2 to 3 minutes, when the firing time of one hour is recorded. Thereafter, the clinkers were taken out of the furnace, quickly removed from the crucible, and tossed into a water-cooled metal bowl for rapid air-cooling. Two burns of six pellets each were made to prepare about one kilogram of clinker. The clinkers were kept in sealed bags and later evaluated.

EVALUATION OF CLINKERS

Evaluation was carried out to quantify the effects of 0.5% metal on both the physical and chemical properties of these clinkers. The tests included:

X-ray diffraction (XRD) to determine major phases: C3S, C2S, C3A, C4AF, and free lime

Free lime (spot check on selected clinkers) by Franke free lime method

X-ray fluorescence (XRF) for the oxide analysis

Bogue calculations

Grindability (to determine ease in grinding)

Microscopic analysis (reflected light microscopy) to identify phase modifications

Phase extraction and AA analyses to determine preferred metal distribution

XRD Analysis

X-Ray diffraction (XRD) analysis on each of the ground clinker samples was done to verify the formation of major phases such as alite, belite, tricalcium aluminate, and tetra-aluminoferrite. Any abnormalities found in the peaks were related to the presence of minor elements. The XRD patterns for the clinkers containing metals are shown in Fig. 2.1. Major peaks indicate the presence of C3S*, C2S, C3A, and C4AF, and show negli-gible free lime. An XRD pattern of the control clinker (with no metal addition) is shown in Fig. 2.1 for comparison. Qualitative information obtained from the major XRD peaks is given in Table 2.3.

Free Lime Analysis

The semiquantitative assessment of free lime in clinkers by XRD showed negligible free lime in all clinkers. Spot checks on the quantitative determination of free-lime were done on selected clinkers by the Franke Free Lime method described in the ASTM C 114-92 specification. Data on the major peaks from the XRD patterns and free lime contents thus determined are shown in Table 2.3.

XRF Analysis

The elemental analyses (as oxides) of each clinker were determined by the XRF to confirm the target clinker composition. If the clinker compositions were found significantly off target, fresh raw mix was prepared and the tests repeated. The metal content of each clinker was also determined. Table 2.4 gives the XRF analyses of selected clinkers.

As can be seen, the levels of lead, cadmium, and molybdenum in clinkers decreased because of their volatile nature, whereas in most of the remaining clinkers the metal contents were close to their starting values.

Bogue Calculations

The XRF analyses also were used to quantitatively determine the major phases by Bogue’s method. The XRF analyses of the major oxides and the corresponding Bogue calculations of C3S, C2S, C3A, and C4AF for each clinker are given in Table 2.5.

The analyses indicate that the clinkers produced were of Type I composition. With the possible exceptions of molybdenum and beryllium, the amount of C3S, C2S, C3A, and C4AF were close to those of the control. Both molybdenum and beryllium slightly reduced C3S with an increase of C2S.

Grindability Test

The effect of trace elements on the grindability of clinker was determined by a grind-ability test developed at the CTL. Twenty grams of –50/+100 sieve size fraction from each clinker was ground for 3 hours in a ceramic ball mill using a ceramic jar and ceramic grinding media. At the end of grinding, Blaine finenesses of the ground clinkers were determined as the “grindability indices.” Lower numbers than the control indicates clinkers are harder to grind, whereas larger numbers indicate softer clinkers. Details of the grindability tests are given in Appendix A1. The grindability indices of the clinkers containing 0.5% metals are given in Table 2.6.

When compared to the control, the grindability indices indicate that chromium, vanadium, titanium, nickel, barium, copper, antimony, and beryllium adversely affect the grindability of clinkers by rendering them harder to grind. Addition of manganese made the clinkers somewhat softer as compared to the control. The metals in order of increasing effect on grindability are:

nickel>titanium>chromium>vanadium>antimony>beryllium>barium> copper>cadmium>lead>zinc> molybdenum>control>manganese

Microscopical Studies

Clinker samples were polished and studied under a reflected light microscope to determine the effects of 0.5% metals on the phase formation. Alite and belite crystals in particular were observed for their color, size, shape, crystallinity, lamellae formation, relative abundance, and decomposition, if any. The interstitial phases of aluminate and ferrite also were observed for distribution with respect to the other phases along with any abnormalities in structure. The data are given in Table 2.7. Micrographs of clinkers containing individual trace elements are shown in Fig. 2.2.

The microscopic observations indicate that, generally, the addition of metals at 0.5% levels produces decomposed alite with cannibalistic features; the belites are slightly irregular with some ragged edges. The size of interstitial phases in most cases remained unaffected; in some cases it changed by 50%.

Selective Dissolution Phase Analysis

The distribution of metals in the major clinker phases, i.e., between silicates and alumi-nates, was determined by extracting the phases and analyzing for the respective metals. The silicates (combination of tricalcium silicate and dicalcium silicate) and aluminates (combination of tricalcium aluminates and tetracalcium aluminoferrite) were extracted chemically. The extractions of silicates were done by the potassium hydroxide-sugar solution (KOHS) method, and the aluminates were extracted by the salicylic acid-methanol (SAMX) method (details described in Appendix A2.). The silicates and alumi-nates were analyzed for metals by the atomic absorption (AA) method to determine any preferential partitioning. The distribution of metals in the silicates and aluminates for each clinker is given in Table 2.8.

Table 2.8. Selective Phase Analyses of Clinkers Containing 0.5% Metals, mass %

The data indicate that with the exception of vanadium, and possibly chromium and cadmium, most metals are concentrated in the aluminoferrite phases. Most probably the metals are concentrated in the ferrite phase because of the affinity of the transition metals for iron. Lead is equally distributed in the silicate and aluminoferrite phases.

EVALUATION OF CEMENTS

Cements were prepared from the clinkers by intergrinding with the appropriate amount of gypsum. The gypsum addition was calculated on the basis of the SO3/Al2O3 molar ratio (usually done for Type I/II low alkali clinkers). The Al2O3 content in clinker, as determined by XRF analysis, is used for calculating the amount of gypsum used. The calculation is given in Appendix A3.

The amount of gypsum added to each clinker is given in Table 2.9.

Table 2.9. Gypsum Added to Clinker Containing 0.5% Metals, mass %

The cements were ground to Blaine fineness between 300 to 350 m2/kg as determined by the ASTM C 204-96 procedure. The cements were tested both as pastes and mortars. The following tests were conducted.

PASTE TESTS

These tests included preparing cement pastes using w/c = 0.484 (as in ASTM C 109 mortars) at a steady shear of 13,000 rpm in a water-cooled WaringTM blender and testing for initial-setting time, workability, and early stiffening behavior by the mini-slump cone test, and compressive strengths. A 500-g fraction of cement was placed in the WaringTM blender and 242 ml of deionized water was added. The mixing was done at 13,000 rpm for a mixing sequence of 1 minute mixing—2 minutes rest—1 minute mixing.

Initial Setting by Modified Vicat Test

Soon after the mixing was complete, a portion of paste was poured in a 100-ml plastic container and the initial setting time was determined by a modified version of the ASTM C 191-92 Vicat Method, as described in Appendix A4. The time of initial setting was recorded from the initial contact of cement with water at the start of mixing. The data is given in Table 2.10 as follows. Paste setting times are usually longer than those for the mortars.

Table 2.10. Initial Setting of Cement Pastes Containing 0.5% Metals

Chromium, nickel, barium, zinc, lead, cadmium, and molybdenum appear to accelerate the paste setting. The effects with zinc and molybdenum are more pronounced. The addition of copper and beryllium retarded the setting. Beryllium appears to be a severe retarder as the paste did not set even after 40 hours. In fact, as can be seen from the cube strengths (Table 2.12), the paste was still plastic even after 28 days; even the 34-day strength was almost 3 times lower than the 3-day old cement pastes made with other metals. The list of metals in order of decreasing time of set is as follows:

zinc<molybdenum<lead<chromium<nickel<barium<cadmium<manganese <vanadium<control<antimony<titanium<copper<beryllium

Mini-Slump Tests for Stiffening

Soon after mixing at 13,000 rpm in the WaringTM blender, the neat paste is tested for early stiffening behavior by a CTL-developed mini-slump test. Details of the test method are given in Appendix A5. The paste is filled in a mini-cone on a plastic platform. The cone is then pulled up and the paste is allowed to collapse as a pat. Any abnormal reduction in the pat area as a function of time indicates abnormal stiffening behavior of the paste. The pat areas at increasing time and the percentage reduction between the initial pat area and that at 30 minutes revealed the data shown in Table 2.11.

Table 2.11. Mini-slump Area cm2 and Degree of Stiffening of Pastes Containing 0.5% Metals, cm2

When compared to that of the control, the difference in pat areas of more than 20% between 2.5 minutes and 30 minutes suggests abnormally stiffening tendencies. Manganese, zinc, cadmium, molybdenum, antimony, and beryllium abnormally stiffened the fresh pastes. The effect was most pronounced with beryllium. Chromium, vanadium, barium, and copper made the paste softer. They rendered the paste fluid even at 30 minutes after mixing.

Compressive Strength

The paste remaining after the mini-slump cone tests was cast as 25-mm (1-in.) cubes and tested for compressive strength. Details of the cube preparation and curing are given in Appendix A6. The cubes were kept in a moist room at 100% relative humidity (R.H.) and 23oC, and their compressive strengths were tested at 3, 7, and 28 days. Duplicate cubes were tested for each age and average value reported. The data are given in Table 2.12.

Table 2.12. Compressive Strength of Paste Cubes Containing 0.5% Metals, MPa (psi)

The metals chromium, titanium, nickel, manganese, copper, cadmium, and beryllium adversely affected the paste strength. The effect of beryllium appears to be of a strong retarder as the paste did not set for an extremely long time, nor gain appreciable strength even after 28 days.

MORTAR TESTS

Separate mortar batches were prepared from the cements in a Hobart mixer using the ASTM C 305-94 procedure. The mortars were tested for flow behavior and compressive strength. As specified in the ASTM C 109/109M procedure, the following proportions of cement, sand, and water were used:

Cement                                         500 g

Sand (Ottawa)                              1375 g

Water (deionized)                        242 mL

Mortar Flow Measurement

The mortars also were tested for flow behavior in accordance with the ASTM C 109 procedure. The results for each mortar are given in Table 2.13.

Table 2.13. Flow of Mortar Containing 0.5% Metals

As compared to the flow of the control, nickel, barium, manganese, lead, and molyb-denum gave harsher mixes, whereas zinc and antimony give more flowable or lean mortars. Chromium, titanium, copper, and cadmium rendered the flow behavior of mortars close to that of the control.

Compressive Strength

The mortars were cast as 50-mm (2-in.) cubes, according to ASTM C 109/109M. The molds were covered and placed in a moist room at 100% humidity and 21°C ± 1°C (70°F ± 3°F). The cubes were demolded the next day, and cured in lime-saturated water. They were tested for compressive strength at 3, 7, and 28 days. Replicate cubes were tested at each age and their average values were recorded in Table 2.14.

Table 2.14. Compressive Strength of Mortar Cubes Containing 0.5% Metals, MPa (psi)

When compared to the 28-day strength of the control mortar, chromium, titanium, nickel, manganese, copper, molybdenum, and beryllium produced weaker mortars, whereas cadmium and antimony gave stronger mortars. Beryllium rendered the mortar so severely retarded that it did not gain any appreciable strength even after 28 days.

RESULTS AND DISCUSSION CLINKERS

Relative Volatility of Metals

It appears that two metals, lead and cadmium, which have low boiling points, do volatilize under experimental conditions so that their final contents in clinkers were lower than the starting levels. The rest of the metals are retained in clinkers close to their original contents.

Clinker Color

Several metals appeared to impart color to cement clinkers. The color of pastes and mortars was affected likewise. The strongest colors have been noted with the following metals:

Beryllium                        Black

Chromium                       Dark green

Manganese                      Dark brown

Copper                            Black to tan

Nickel                             Dark brown

Vanadium                       Black to tan

XRF and Bogue Analyses

The XRF and Bogue analyses confirm that the clinkers produced with the trace metals were Type I. The amounts of C3S, C2S, C3A, and C4AF were close to those of the control. The only exceptions were the addition of molybdenum and beryllium, which decreased the level of C3S somewhat and increased C2S and C3A contents. The C4AF contents were the same as that of the control.

XRD Analysis

Although XRD is a semiquantitative method, the data have confirmed the presence of the major C3S, C2S, C3A, and C4AF phases and have shown negligible free lime in all clinkers. However, it appears that the formation of the aluminate and ferrite phases have been notably affected with the additions of metals. This indicates that the metals have primarily reacted with these phases, as is also confirmed by the chemical analyses (discussed later in this section). Some of the observations have been noted as follows:

• Beryllium enhanced the formation of ferrite content at the expense of the aluminate. It also enhanced the alite content.
• Antimony significantly decreased the formation of ferrite, but improved the aluminate content, and enhanced the alite content.
• Titanium, nickel, barium, manganese, copper, and molybdenum also registered some effects on the formation of the aluminate and/or ferrite phase.

Microscopical Studies

Effect on Alite Formation: In general, all metals have resulted in some degree of fusion, also known as cannibalism, and partial decomposition in alite crystals. The most decom-position was noted with titanium, whereas moderate decomposition was observed with beryllium. The addition of copper and chromium enlarged the alite size, whereas lead, cadmium, and manganese reduced it as compared to the control.

Effect on Belite Formation: Almost all metal additions produced round to slightly irregular belite crystals. Chromium, vanadium, titanium, nickel, and antimony produced

belite crystals with ragged edges. Barium, vanadium, manganese, and copper produced larger belite than the control. Antimony reduced the belite size.

Effect on the Interstitial Phases: The formation of interstitial phases appears largely unaffected. However, the addition of titanium, nickel, and copper showed some enlargement, and manganese showed some reduction in the size of interstitial crystals.

Grindability

Although the effect of crystal formation on clinker grindability is not clearly defined yet, the following general observations from the literature have been made (Hills, 1995):

Alite:

Of the major clinker phases, alite is the easiest to grind An increase in alite content improves grindability Small alite crystals improve grindability

Less defined alite crystals improve grindability Agglomeration of alite adversely affects grindability

Belite:

Belite crystals are the most difficult to grind

A decrease in belite content improves grindability Small belite crystals improve grindability

Poorly defined belite crystals improve grindability Agglomeration of belite adversely affects grindability

Interstitials:

The effect of interstitial phases on clinker grindability is inconclusive

Porosity:

Porosity improves grindability

Larger pores may only improve coarse grinding

The data indicate that the grindabilities of clinkers containing chromium, vanadium, titanium, and nickel are strongly affected, as the clinkers become harder to grind than the control. The clinker hardness may be attributed to the formation of somewhat larger belite crystals than the control. Belite, being hard in nature, causes difficulty in grinding

The grindabilities of other clinkers such as those containing copper, barium, cadmium, beryllium, and antimony are moderately affected. These effects can be related to the clinker crystallography; for instance, in the case of copper- and barium-containing clinkers, the alite crystals are particularly large (50 mm diameter), which can affect the grindabilities as noted by their moderate grindabilities indices.

The clinker containing manganese is somewhat easier to grind than the others because of the small size alite crystals, and highly dendritic belite crystals.

The grindability indices of the clinkers containing lead, zinc, and molybdenum are identical to those of the control, consistent with the fact that the alite crystals are of similar size.The proportions of interstitial phases in clinkers are not very different from the control, therefore their effect on clinker grindability does not appear significantly different from the control.

Selective Dissolution Phase Analysis

Most metals are concentrated in the aluminoferrite phases except vanadium which is concentrated in the silicate phases. As mentioned earlier, this observation agrees with the XRD data confirming favorable interaction of most metals with the aluminate and ferrite phases. These findings are in agreement with those of Hornain (1971), who reported that the transition metals primarily stay in the aluminoferrite phase (preferably the ferrite); the next most preferred phase for these metals is belite. Chromium is distributed roughly equally in the aluminoferrite and silicate phases. Lead and cadmium have low contents in both the silicates and aluminoferrite phases because they readily volatilize.

What is the mechanism for metal substitution in the aluminoferrite phase? In the cases of barium, manganese, and beryllium, the metal oxides are substituting for Fe2O3 in the ferrite phase and enlarging these peaks. Barium and beryllium are unlikely to substitute for iron the ferrite phase; instead, they may substitute for calcium in the ferrite phase. Copper and lead appeared to have contributed to both the aluminate and ferrite phases. Some metals also appear to have substituted for Al2O3 in the aluminate phase and enlarging its peak as in the cases of titanium, nickel, and antimony metals.

CEMENTS

Paste Properties

Initial Setting: The reduced initial setting times, as determined by the Vicat method, suggest that chromium, nickel, zinc, lead, and molybdenum accelerate the cement paste hydration. The effects are pronounced with zinc and molybdenum; the setting time is reduced by nearly two hours. The accelerating effects may be caused by the release of metals from the aluminoferrite into the aqueous phase, enhancing the nucleation of CH to

promote hydration. Both the aluminate and ferrite, being highly reactive phases, readily release the ions upon contact with water.

Copper and beryllium impart retarding effects on the cement hydration. Beryllium is a very severe retarder as its paste did not set even after 40 hours. It appears that the 0.5% addition of beryllium was extremely high for clinker to have an adverse effect on the early hydration properties. It is probable that an addition of 0.5% beryllium was much higher than would normally be expected with real materials. The effects in these studies were much more severe than would be anticipated in actual raw materials.

Mechanistically, the released metal ions can cause delayed setting by one of the following reasons:

– Delay the nucleation of CH by competing with Ca ions and pH imbalance, – Poison the growth of C-S-H by adsorption on the crystal face, or

– both of the above

Early Stiffening: The stiffening properties of the pastes, as determined by the mini-slump cone tests, suggest that manganese, cadmium, molybdenum, and beryllium impart early stiffening tendencies. The effect is very pronounced with beryllium as the paste area is reduced by nearly 41% within 30 minutes. A decrease in pat area close to 20% is considered normal. Any abnormal stiffening is primarily attributed to the imbalance between the amount and reactivity of the aluminate phase and the availability of sulfate ions for interaction.

Chromium, vanadium, barium, and copper decrease the paste rate of early stiffening as they appeared flowable even at 30 minutes after mixing.

Mechanistically, the early stiffening behavior should be related to the balance between C3A and sulfate ions available in the paste. It appears probable that the ions released from the above metals cause interference in the C3A-sulfate balance and result in abnormal stiffening of the paste.

The early setting and the abnormal stiffening behaviors of the paste do not appear to be directly related.

Compressive Strength: On average, the 28-day strengths of pastes with almost all of the metals were somewhat lower than the control. Beryllium, titanium and cadmium reduced the 28-day strength more than 14% compared to the control. In the case of beryllium, the paste did not set for a very long time (more than a month); it only showed a meager

strength of 1270 psi after 37 days. The average reduction in the paste strength, excluding beryllium, was 9.24%.

Early Strength: Paste with vanadium, titanium, copper, and antimony gave low early strengths (3-day strengths), but (except for titanium) the 28-day strengths were close to that of the control. The noticeably low strength of titanium may be attributed to the formation of its highly decomposed alite crystals in clinker that affected its hydration and strength characteristics. The abnormality in the hydration, setting, and strength of pastes with beryllium might be attributed to a moderate degree of alite decomposition in the clinker.

Barium and zinc showed high early strengths (3-day strengths), but the 28-day strength was reasonably close to the control.

Mortar Properties

Flow Behavior: As per ASTM C 109 requirements, the water-cement ratio was kept at 0.485 for all mortars tested in the investigations, and their flow was measured using the flow table. Any deviation from the specified flow value of 110±5 will reflect the degree of workability of a given mortar.

In the present studies, except for molybdenum, the flow of all other metal containing mortars ranged between 102 and 122. The paste with cement doped with molybdenum showed excessive reduction in flow behavior (flow was 93). Titanium, zinc, cadmium, and antimony exhibited higher flow than the normal (normal range being 105 to 115), whereas nickel, barium, and manganese made the mortars less flowable than the normal.

The flow properties of the mortars did not materially affect their strength characteristics.

Compressive Strength: The 28-day strength with 0.5% metals showed mixed behavior. However mortars made with a majority of metals (chromium, titanium, nickel, manganese, copper, molybdenum, and beryllium) showed reduced strengths, whereas zinc, lead, cadmium, and antimony gave increased strengths. Mortar strengths with barium and vanadium did not change from that of the control.

The effect of beryllium on mortar strength was extremely severe; in fact, the mortar did not set for 28 days, and therefore no strength was recorded. Titanium and copper reduced the mortar strength close to 15%. Up to 10% reduction in 28-day strength was recorded for nickel, manganese, and molybdenum. Notwithstanding that a 10% difference in strength may be within test variability, the data generally suggest a trend that the metals contribute to loss in strength.

Early Strength: A significant drop in the early strength (nearly 20% drop in 3-day strength) also was recorded for titanium, whereas a nearly 10% drop in strength also was recorded for vanadium, manganese, copper, and chromium. Again, the low strength of titanium may be attributed to its decomposed nature of alite crystals that adversely affected its hydration and strength properties. The same also may be true for beryllium— containing clinker that has alite crystals with a moderate degree of decomposition— enough to adversely affect cement hydration, setting, and ultimately the strength of mortar.

Cadmium and antimony have shown high early strengths (3-day strengths) as well as 28- day strengths, compared to the control.

SUMMARY OF RESULTS

A summary of the results discussed above is given in Table 2.15, in which metals that have led to abnormal properties of clinkers, paste, and mortars as compared to the respective controls have been listed.

Table 2.15. Summary of Effects Shown on Clinker and Cements by 0.5% Metal Addition

Based on the properties of clinkers, pastes, and mortars cited above, the metals showing most noticeable changes have been considered for testing at a lower level of addition (0.25%) as listed below and discussed in Task 4.

Antimony           Copper                        Nickel

Beryllium            Manganese                 Titanium

Chromium         Molybdenum              Vanadium Control (no metal addition)

TASK 3 PRODUCTION AND EVALUATION OF SELECTED CLINKERS AND CEMENTS MADE WITH SINGLE METALS AT 0.25% BY MASS OF CLINKER

MATERIALS AND METHODS

From the Task 2 investigations on the addition of 0.5% metals, and the summary of the results given in Table 2.15, metals that showed the most noticeable effects in at least four categories of tests performed were retained for testing at the 0.25% addition level. Thus, the metals barium, cadmium, lead, and zinc were omitted, and the following metals were included in the task:

Antimony                         Molybdenum

Beryllium                         Nickel

Chromium                        Titanium

Copper                             Vanadium

Manganese                       Control (no metal addition)

Table 3.1. Raw Mix Composition for Making 1 Kg Clinker with 0.25% Metal Addition, g

Table 3.2. Amount of Metal Used in Raw Mix to Produce 1 Kg Clinker with 0.25% Metal Addition, g

These metals were incorporated at the 0.25% level in raw mix; clinkers and cements were prepared and their performance evaluated as in Task 2.

RAW FEED AND CLINKER PREPARATION

Raw mix compositions for clinkers with each metal are given in Tables 3.1 and 3.2. The preparation of raw feed, pelletizing, and firing to make clinker was done as described earlier in Task 2.

The mix design for control clinker with no metal addition is also given in Table 3.1 for comparison. The clinkers were made using the same procedure described in Task 2 for 0.5% metal clinkers.

EVALUATION OF CLINKERS

The evaluation of clinkers containing 0.25% metals and the cements made from them were carried out as in Task 2. The clinkers and cements prepared above were evaluated by carrying out the following tests:

• X-ray diffraction (XRD) for C3S, C2 C3A, C4AF, and free lime identification
• X-ray fluorescence (XRF) for the oxide analysis
• Bogue calculations
• Grindability tests

XRD Analysis

X-ray diffraction (XRD) patterns for clinkers made with the individual metals (0.25%) listed above are shown in Fig. 3.1. Data obtained from the XRD patterns on the major peaks are given in Table 3.3.

Table 3.3. XRD Peaks of Major Phases in Clinkers Containing 0.25% Metals

* – same as control; > greater than control; >> significantly greater; < smaller than control; << significantly small

Fig. 3.1 X-ray diffraction patterns of clinkers made with single metal additions at a level of 0.25% by mass: (a) antimony, (b) beryllium, (c) chromium, (d) copper, (e) manganese, (f) molybdenum, (g) nickel, (h) titanium, and (i) vanadium.

The XRD patterns and the appropriate peaks confirmed the formation of the major phases, i.e., C3S, C2S, C3A, or C4AF. Again, the free lime peaks were negligible in all the clinkers to suggest a complete conversion of lime to the respective phases. Therefore, only spot checks of lime content were done on selected clinkers, such as those containing chromium and vanadium, by the Franke Free Lime method. The values, given in

Table 3.3, confirmed very low free lime (0.19% and 0.16% respectively).

XRF and Bogue Analyses

As in the Task 2, the clinkers were tested for the elemental analysis (as oxides) using the XRF method to confirm the target clinker composition. The XRF analyses of the clinkers and the corresponding Bogue calculations of C3S, C2S, C3A, and C4AF phases are give in Table 3.4.

Table 3.4. Oxide Composition and Bogue Analyses of Clinkers Containing 0.25% Metals, mass %

The analyses confirm that the clinker produced were of Type I.

Grindability Test

Similar to Task 2, the effects of trace elements on the grindability of clinker was deter-mined by the CTL-developed grindability test. The grindability indices (in cm2/g) of clinkers containing 0.25% by mass of the trace elements are given in Table 3.5.

Table 3.5. Grindability Indices of Clinkers Containing 0.25% Metal, cm2/g

The grindability indices indicate that even with 0.25% metal additions, the clinkers are still significantly harder to grind.

EVALUATION OF CEMENTS

Cements were prepared from these clinkers by mixing with an appropriate amount of gypsum, calculated per Eq. 2.1, and grinding to Blaine fineness between 300 to 350 m2/kg. The pastes and mortars made from theses cements were evaluated as in Task 2.

Paste Tests

The pastes were tested for initial-setting time, mini-slump cone test, and compressive strengths.

Initial Setting by Modified Vicat Test

The paste mixing was done in the WaringTM blender as per the procedure described in Task 2. The initial setting time of paste was determined immediately after each mixing by the Vicat Method as described in the ASTM C 191-92 procedure. The time of initial setting was recorded from the initial contact of cement with water at the start of mixing in the WaringTM blender. The data is given in Table 3.6 as follows:

Table 3.6 Initial Setting of Cement Pastes Containing 0.25% Metal

The prolonged initial setting time of 26 hours indicates the 0.25% addition of beryllium still significantly affects the setting behavior of the paste.

Mini-Slump Tests for Early Stiffening

The pastes prepared in the WaringTM blender were tested for mini-slump to assess early stiffening behavior as per the procedure adopted in Task 2. A sharp reduction in the area as a function of time indicates a stiffening behavior of the paste. Data on pat areas at increasing time and the percentage reduction between the initial pat area and that at 30 minutes is given in Table 3.7.

Table 3.7. Mini-Slump Areas and Degree of Stiffening of Pastes Containing 0.25% Metals, cm2

Change in pat area suggests that 0.25% beryllium stiffens the paste, whereas chromium softens it.

Compressive Strength

One-in. cubes were made from these pastes, cured in a moist room, and tested for compressive strength at 3, 7, and 28 days, as described in Task 2. The results are given in Table 3.8.

Table 3.8. Compressive Strength of Paste Cubes Containing 0.25% Metals, MPa (psi)

The paste strengths with 0.25% metals appear closer to the control as compared to 0.5% metal additions. Addition of 0.25% beryllium caused retardation and resulted in low early strengths (3- and 7-day strengths); however, the 28-day strength was comparable to that of the control.

MORTAR TESTS

As in Task 2, separate mortar batches were prepared from the above cements using the ASTM C 109/109M procedure.

Mortar Flow Measurement

Again, these mortars were tested for their flow using the procedure described in ASTM Method C 230-90. The spread of the mortar was recorded as in Table 3.9.

Table 3.9. Flow of Mortars Containing 0.25% Metals

Table 3.10. Compressive Strength of Mortar Cubes Containing 0.25% Metals, MPa (psi)

Compressive Strength

The mortars were cast as 50-mm (2-in.) cubes the stored at 100% relative humidity and 21°C ± 1°C (70°F ± 3°F). They were demolded the next day, cured in lime water, and tested for compressive strength at 3, 7, and 28 days. The average strengths recorded are given in Table 3.10 .

The low 3-day strength of beryllium-containing mortars suggests the retarding effect of beryllium as noted earlier in the initial setting and early stiffening tests. The 28-day strength was, however, closer to that of the control. The high 28-day strength of molyb-denum- and antimony-containing mortars is also noteworthy.

RESULTS AND DISCUSSION CLINKERS

Generally, the effects of 0.25% metal additions were less noticeable on almost all prop-erties of clinkers, cements, pastes, and mortars tested in these investigations. Overall, the properties of these materials were close to those of the control.

XRF and Bogue Analyses

The XRF and Bogue analyses confirmed that Type I clinkers were produced. The C3S, C2S, C3A, and C4AF contents were close to those of the control. The only exception in this case appears to be the addition of molybdenum which somewhat decreased C3S and increased C2S. The C3A and C4AF contents for all the clinkers were almost the same as that of the control.

XRD Analysis

Again, the effect of metals on XRD peaks is substantially reduced when added at the 0.25% level instead of 0.5%. However, antimony and beryllium are still exceptions. Antimony increases aluminate at the expense of ferrite, whereas beryllium increases ferrite at the expense of aluminate. All other metals register from slight to moderate effects on either the aluminate or ferrite phase formations.

Grindability

A somewhat mixed trend in grindability is observed with 0.25% metal additions. Whereas the grindabilities with chromium, vanadium, titanium, nickel, and copper improved and moved closer to that of the control, the grindabilities of molybdenum, antimony, and beryllium were reduced unexpectedly. The grindability with manganese also decreased more than expected.

CEMENTS

Paste Properties

Initial Setting: The initial setting times with 0.25% metal addition were closer to that of the control than those prepared with 0.5% metal addition. However, molybdenum was still an accelerator and beryllium still a strong retarder as can be seen from the short and long setting times of 4:43 and 26 hours respectively. The release of molybdenum from the aluminoferrite into the aqueous phase may enhance the nucleation of CH and promote hydration, whereas beryllium either causes delay in the nucleation of CH or poisons the growth of CH by adsorption, leading to a delayed setting.

Early Stiffening: The early stiffening properties of the pastes were far less evident with 0.25% metal addition as compared to those prepared with 0.5% metal. However, stiffening behavior with beryllium, nickel, and antimony was still observed, and likewise the softness of paste with chromium was also evident.

Compressive Strength: The 28-day strength at 0.25% addition of metals is slightly higher than the control, but appeared closer to the control as compared to the 0.5% metal addition. On average, the variation from the control was 7.8%. The only metals showing low strengths were molybdenum and beryllium, with 9450 and 9650 psi respectively after 28 days.

Beryllium again gave low early strengths but at 0.25% addition the cement did set and gave higher strengths as compared to 0.5% addition. Early strength with antimony was also low, but was slightly better when compared to that of 0.5% addition; the corre-sponding 28-day strength was closer to the control when compared with 0.5% addition.

Some unexpected behaviors from nickel (high early strength), manganese (high 28-day strength), and molybdenum (low 7-day strength) also were noticed.

Mortar Properties

Flow Behavior: The effect of 0.25% metal addition on mortar flow is insignificant. The mortar flow with all ranged between 105 and 115. Even the only exception, molybdenum, with a flow of 117.5, was very close to 115 of the control.

Compressive Strength: In general, the reduction of metal level from 0.5% to 0.25% improved mortar strengths close to the control. The only exception was antimony, which increased the mortar strength instead of decreasing it. In the cases of nickel and molyb-denum, the mortar strengths not only improved but also exceeded the control.

A significant improvement in the 3-day strengths for almost all metals was realized with 0.25% metal addition. Their mortar strengths were much closer to the control than those prepared with 0.5% metal additions. Beryllium still appears to be affecting the early setting strength behavior of the mortars.

TASK 4 PRODUCTION AND EVALUATION OF CLINKERS AND CEMENTS MADE WITH 0.5% MULTIPLE METALS BY MASS OF CLINKER

The following raw feed mixes were used for making clinkers with 0.5% additions (on metal basis) of multiple metals. The combinations of multiple metals was based on their frequency of occurrence in the raw feed and/or in fuel. The following combinations of multiple metals were tested:

Ni-V-Sb                          Cu-Cr-Ni-Zn

Cu-Ni-Zn                       Cu-Pb-Ni-Zn

Cu-Pb-Zn

MATERIALS AND METHODS

Raw Feed and Clinker Preparations

The preparation of raw feed, pelletizing, and firing for making clinker was done as described in Task 2. The following mix proportions of raw materials (Table 4.1) were used in order to prepare 1 kg of clinker.

Table 4.1. Raw Materials for Preparing 1 kg Clinker Containing 0.5% Multiple Metals, g

Table 4.2. Metal Oxides (Based On 0.5% Metal) Used in Raw Feeds for Preparing 1 kg of Clinker

EVALUATION OF CLINKERS

Evaluation of clinkers containing 0.5% metals was carried out to quantify both their physical and chemical properties. The tests included:

X-ray diffraction (XRD) for determining major phase: C3S, C2S, C3A, C4AF, and free lime X-ray fluorescence (XRF) for the oxide analysis, and Bogue calculations

Grindability

Microscopic analysis (reflected light microscopy) to identify phase modifications Selective phase analysis (chemical extraction followed by AA analysis of metals)

XRD Analysis

The XRD patterns for all the clinkers containing multiple metals are shown in Fig. 4.1. The peaks corresponding to the major phases suggest the presence of appropriate amounts in the major phases C3S, C2S, C3A, and C4AF, corresponding to the formation of Type I cement clinkers. Free lime peaks were negligible in all the cases to suggest a complete conversion of lime to the respective phases. Information obtained from the XRD patterns is given in Table 4.3.

Table 4.3. XRD Peaks Of Major Phases in Clinkers Containing 0.5% Multiple Metals

– = same as control; > = greater than control; < = smaller than control

(f) control

Fig. 4.1 X-ray diffraction patterns of clinkers made with multiple metal additions at a level of 0.5% by mass: (a) Ni-V-Sb, (b) Cu-Ni-Zn, (c) Cu-Pb-Zn, (d) Cu-Pb-Ni-Zn, (e) Cu-Cr-Ni-Zn, and (f) the control clinker.

XRF and Bogue Analyses

The clinkers were analyzed for metals (as oxides) to confirm the target clinker compo-sition. If any composition was off target, the test was repeated with fresh mix. Table 4.4 gives the XRF analyses of clinkers and the corresponding Bogue calculations of C3S, C2S, C3A, and C4AF.

Table 4.4. Oxide Composition and Bogue Analyses of Clinkers Containing 0.5% Multiple Metals, mass %

The XRF and Bogue analyses confirm that Type I clinkers were produced from these raw mixes. The amount of C3S, C2S, C3A, and C4AF were close to those of the control.

Grindability Test

The effect of multiple trace elements on the grindability of clinker was determined by the CTL-developed grindability test as used in Task 2. The “grindability indices” (cm2/g) for each clinker containing multiple trace elements are given in Table 4.5.

Table 4.5. Grindability Indices of Clinkers Containing 0.5% Multiple Metals, cm2/g

In general, the grindability indices indicate that the addition of multiple metals adversely affects the grindability of clinkers by rendering them harder to grind. The low grind-ability index of Cu-Cr-Ni-Zn-containing clinker could well be due to the larger size of C3A crystals (see Table 4.6).

Table 4.6. Microscopical Observations of Clinkers Containing 0.5% Multiple Metals

Microscopical Studies

Representative clinker samples were polished and observed under a reflected light micro-scope. Alite and belite crystals in particular were observed for their size, crystallinity, relative abundance, and decomposition. The interstitial phases of the tricalcium alumi-nates and tetracalcium aluminoferrite were observed for distribution with respect to the

other phases along with any abnormalities in their formation. Table 4.5 shows the crystal size, decomposition, lamellae formation, and size of the interstitial phase. The micro-graphs of these clinkers are shown in Fig. 4.2. The microscopic study indicates that except for Cu-Ni-Zn, all other multiple metal additions lead to partially decomposed and canni-balized alite, and slightly irregular and ragged belites.

• control

Fig. 4.2 Optical micrographs of clinkers made with multiple metal additions at a level of 0.5% by mass: (a) Ni-V-Sb, (b) Cu-Ni-Zn, (c) Cu-Pb-Zn, (d) Cu-Pb-Ni-Zn, (e) Cu-Cr-Ni-Zn, and (f) the control clinker.

Selective Dissolution Phase Analysis

The distribution of metals in the clinker phases was determined by extracting and analyzing the phases. The major phases—silicates (combination of tricalcium silicate and dicalcium silicate) and aluminates (combination of tricalcium aluminates and tetracalcium alumino-ferrite) were separated by the chemical extraction described earlier in Task 2. The silicate and aluminate phases were then analyzed by wet chemical analysis to determine the partitions of metals. The distribution of metals in the silicates and aluminates is given in Table 4.7.

Table 4.7. Distribution of Metals in Selective Clinker Phases Containing 0.5% Multiple Metals

The data confirm the findings of Task 2 on 0.5% single metal additions and indicates that except for vanadium and possibly chromium, all other metals are concentrated in the aluminoferrite phases.

EVALUATION OF CEMENTS

Cements were prepared from these clinkers by intergrinding with the appropriate amount of gypsum, calculated by Eq. 3.1 on the basis of aluminate contents in each clinker as determined by XRF. The fineness of the ground cement was close to 300 to 350 m2/kg. The cements were evaluated by testing as pastes and mortars as in Task 2.

PASTE TESTS

Pastes were mixed in a WaringTM blender at a steady shear of 13,000 rpm and evaluated by the modified Vicat, mini-slump cone, and compressive strength tests as in Task 2.

Initial Setting by Modified Vicat Test

The initial setting time was determined by the modified Vicat method described in Task 2. At the completion of mixing, the time of initial setting was recorded from the initial contact of cement with water. The data are given in Table 4.8.

Table 4.8. Initial Setting of Cement Pastes Containing 0.5% Multiple Metals

Multiple metal addition does not seem to affect the initial setting of cement pastes. The only exception could be Cu-Pb-Zn, which mildly accelerates the hydration and reduces the setting time of the paste.

Mini-Slump Tests for Early Stiffening

The mini-slump cone tests on the pastes were conducted using the procedure followed in Task 2. The pat areas at 2.5-, 5-, 15-, and 30-minute intervals and the percentage reduction between 2.5 and 30 minutes are shown in Table 4.9.

Table 4.9. Mini-Slump Areas and Degree of Stiffening of Pastes with 0.5% Multiple Metals, cm2

The data on the difference in pat area between 2.5 and 30 minutes suggest very little effects of metals on paste stiffening when multiple metals are added.

Compressive Strength

The pastes were cast in 1-in. cube molds. The molds were placed in a moist room main-tained at 100% relative humidity and 21°C ± 1°C (70°F ± 3°F). The cubes were demolded the next day, labeled, and kept in perforated plastic trays (to avoid moisture pooling) in the moist room for strength testing at 3, 7, and 28 days. The results are given in Table 4.10 as follows:

Table 4.10. Compressive Strength of Paste Cubes Containing 0.5% Multiple Metals, MPa (psi)

The effects of multiple-metal additions on paste strength do not appear very significant except for Cu-Pb-Zn and Cu-Pb-Ni-Zn, which reduce the 28-day strengths.

MORTAR TESTS

Separate mortar batches were prepared from the same cements according to ASTM

C 109/109M. The proportions of cement, sand, and water were the same as in Task 2.

Mortar Flow Measurement

These mortars were tested for their flow behavior in accordance with the ASTM C 230-90. The spread of the mortar, as given in Table 4.11, is determined after the mixing.

Table 4.11. Flow of Mortars Containing 0.5% Multiple Metals

The mortar flow appears close to the control, suggesting very little effect of multiple metals.

Compressive Strength

The mortars were cast as 50-mm (2-in.) cubes. They were covered and placed in a moist room maintained at 100% relative humidity and 21°C ± 1°C (70°F ± 3°F). The cubes were demolded the next day, labeled, and cured in lime water and later tested for compressive strength at 3, 7, and 28 days. Two mortar cubes were tested at one age and the average value was recorded. The average strength values recorded were given in Table 4.12 as follows:

Table 4.12. Compressive Strength of Mortar Cubes Containing 0.5% Metals, MPa (psi)

The addition of multiple metals appears to have enhanced the 28-day mortar strengths, which is somewhat contrary to their effects on the pastes. Some possible contributory factors to this are given in the discussion section.

RESULTS AND DISCUSSION

It appears from the above data that the multiple additions of metals affects the properties of clinkers, cements, pastes, and mortars differently than the same metals added individually.

CLINKERS

XRF and Bogue Analyses

The XRF and Bogue analyses confirmed that the clinkers produced with all the multiple-metal additions at 0.5% level were of the Type I composition. The C3S, C2S, C3A, and C4AF contents were close to those of the control. A small reduction of C3S with a corre-sponding increase in C2S is noted with Cu-Pb-Zn addition; the C3A, and C4AF contents were identical to those of the control.

XRD Analysis

The XRD tests of the clinkers also confirmed the formation of major C3S, C2S, C3A, and C4AF phases, while the free lime was shown to be negligible. It appeared, though, that the addition of multiple metals has, by interacting with any of these phases, noticeably affected the formation of aluminate and/or ferrite phases. Some of the observations have been noted as follows:

Both Ni-V-Sb and Cu-Ni-Zn addition enhanced the formation of C3A, in the case of Ni-V-Sb at the expense of C4AF.

Addition of Cu-Pb-Zn, Cu-Pb-Ni-Zn, and Cu-Cr-Ni-Zn produced well-defined C4AF crystal structure as indicated by the peaks’ sharpness, while Cu-Pb-Zn also enhanced C3A formation.

The effects of multiple metals on the formation of alite and belites were less noticeable in XRD, but some general observations noted under microscope are given in the following section.

Microscopical Studies

Effect on Alite Formation: In general, all multiple-metal additions have resulted in some degree of cannibalism and partial decomposition in alites. The alite size was also slightly enlarged as compared to the control.

Effect on Belite Formation: Again, all multiple-metal additions produced larger belite crystal than the control. The crystals were generally round to slightly irregular, except for Cu-Cr-Ni-Zn, which produced round, less ragged, very large belite crystals.

Effect on the Interstitial Phases: Generally, the addition of multiple metals appears to have reduced the size of interstitial phases when compared with the control. The addition of Cu-Cr-Ni-Zn, however, enlarges the interstitial size.

Grindability

The data indicate that all multiple metals have adversely affected the grindabilities. The clinkers became harder than the control. The strong effect has been observed with addition of Cu-Cr-Ni-Zn. This may be attributed to the formation of larger belite and aluminate crystals, which are difficult to grind.

The low grindability indices also may be related to the reduced interstitial phases in these clinkers with multiple-metal additions. Although the literature data on this subject is inconclusive, a weak inverse-relationship between the interstitial phases and the hardness of clinker may not be ruled out from the present evidence.

Selective Dissolution Phase Analysis

The distribution trend of metals appears similar to that in the single-metal addition (at 0.5%) in Task 2. Even though the additions have been made as multiple metals, the majority of the metals were concentrated in the aluminoferrite phases. Again, similar to the findings in Task 2, the only exception was vanadium, which was concentrated more in silicates, and to some extent chromium, which was slightly more concentrated in silicates than aluminoferrites. Lead in both additions (i.e., Cu-Pb-Zn and Cu-Pb-Ni-Zn) gave low concentrations in the silicates and aluminoferrite because of its volatility.

CEMENTS

PASTE PROPERTIES

Initial Setting

With the exception of the setting behaviors of Ni-V-Sb and Cu-Pb-Ni-Zn, which are closer to the control than the others, the multiple metal additions have accelerated the paste setting. It might be noted that the majority of the metals used in these combinations, such as nickel, zinc, and chromium, are accelerators by nature; copper is a retarder, whereas vanadium and antimony are neither accelerators nor retarders (refer to the initial setting times for these metals in Table 2.10). The cumulative accelerating action of most metals (nickel, zinc, and chromium), appears to supersede the retarding action of a singular metal (copper).

Early Stiffening

The additions of multiple metals generally stiffened the paste more than the individual metals added separately. The stiffening with the additions of Cu-Ni-Zn and Cu- Pb- Ni-Zn was more pronounced than the rest of the multiple-metal additions.

Paste Strength

The 28-day strength for pastes with multiple-metal additions was fairly close to that of the control. The only real exception appears to be Cu-Pb-Zn, which exhibited a reduction of almost 17% strength. The average variation (reduction) in the paste strength, excluding Cu-Pb-Zn, was less than 2%.

More than 10% increase over the control was noted in the early strengths (3-day strengths) for all multiple metal additions (including Cu-Pb-Zn), except for Ni-V-Sb, which reduced the strength by more than 10% of the control.

MORTAR PROPERTIES

Flow Behavior

The addition of multiple metals generally improves the mortar flow over the single metal addition as in Task 2. In the case of multiple metals the flow was closer to that of the control when compared to the single metals.

Compressive Strength

Multiple-metal additions gave higher mortar strengths over the control than the single-metal additions. The increase in strength over the control raged from 5% to 15%. The most prominent increase was noted with Cu-Pb-Zn and Cu-Pb-Ni-Zn addition.

A significant increase in the early strength (between 9% and 23% increase in 3-day strengths) also was recorded for all multiple-metal additions (including Cu-Pb-Zn), except for Ni-V-Sb, which reduced the early strength by nearly 10%.

TASK 5 PRODUCTION AND EVALUATION OF CLINKERS AND CEMENTS PREPARED WITH ADDITIONS OF SELECTED INDUSTRIAL BY-PRODUCTS

This task consisted of producing clinkers from a number of commonly used by-products as partial kiln feed in cement manufacturing. The clinkers (and subsequently the cements) were prepared and evaluated as in Task 2. The following industrial by-products were studied:

Sinter mix Catalyst fines Copper slag

The chemical composition of these by-products is shown in Table 5.1. The trace metal concentrations are shown at the bottom of the sections for each material in the table.

MATERIALS AND METHODS

Raw Feed and Clinker Preparation

Typically, copper slag 4%, sinter mix 4%, and catalyst fines 5% are added by weight of the raw feeds. But in the present studies these levels were doubled, i.e., 8%, 8%, and 10% respectively, to determine their effects on clinker and cement performance.

As per the by-product chemical analyses, each raw mix was formulated as shown in Table 5.1. The materials were ground in a disc-pulverizer prior to intergrinding for one hour in a ball mill with the other ingredients shown in Table 5.1. The catalyst fines, being fine in nature, did not require initial grinding and were directly interground with the raw ingredients. The remaining preparation of raw feed, pelletizing, and firing to make clinker was done as described in Task 2.

Table 5.1. Raw Mix Formulations for Making 1 kg Clinker with Selected By-product Additions

Copper Sla

Sinter Mix

Catalyst Fines

Trace metal concentrations in these by-products as determined by the XRF analyses are given in Table 5.2 as follows:

It may be pointed out that 8% copper slag will result in 0.35% additional total trace

Table 5.2. Concentrations of Trace Metal (Wt. %) in the By-products Used

metals; 8% sinter mix results in 0.11% additional trace metals; and 10% catalyst fines will give 0.07% additional total trace metals in the raw feed.

EVALUATION OF CLINKERS

The clinkers were evaluated as follows:

X-ray diffraction (XRD) for C3S, C2S, C3A, C4AF, and free lime identification X-ray fluorescence (XRF) for the oxide analysis, and Bogue calculations

Grindability

Microscopic analysis to identify phase modifications, if any

Chemical extraction of major phases and AA analysis for elemental distribution

XRD Analysis

X-ray diffraction (XRD) patterns obtained to evaluate the formation of major compounds, i.e., C3S, C2S, C3A, C4AF, and the presence of any free lime. These are shown in Fig. 5.1. Information on compounds based on the peak intensities in the XRD patterns are given in Table 5.3.

Table 5.3. XRD Peaks of Major Phases in Clinkers Containing Industrial By-products

*< = smaller than control; – = same as control; > = greater than control

The XRD patterns confirmed the formation of the major phases in all clinkers containing these by-products. The free lime peaks were negligible, suggesting complete conversion of lime to the respective phases. However, copper slag and sinter mix enhanced C4AF at the expense of C3A.

• control

Fig. 5.1 X-ray diffraction patterns of clinkers made with industrial by-products: (a) copper slag, (b) sinter mix, (c) catalyst fines, and (d) the control clinker with no by-products.

XRF and Bogue Analyses

The XRF analyses and the Bogue calculations for these clinkers are given in Table 5.4.

Table 5.4. Oxide Composition and Bogue Analyses of Clinkers Containing By-products

The analyses indicate that mostly Type I clinkers were produced from these mixes.

Grindability

The CTL-developed grindability test on these clinkers was conducted as in Task 2. The grindability indices thus obtained are shown in Table 5.5.

Table 5.5. Grindability Indices of Clinkers Containing Industrial By-products, cm2/g

The lower grindability indices than that of the control suggest that the addition of by-products made clinkers harder to grind. The low grindability indices can be attributed to the presence of large alite crystals formed in clinkers.

Microscopical Studies

The clinker specimens examined under microscope revealed the following information on the formation/modifications of the major phases. The major parameters examined were the crystal size, decomposition, lamellae formation, and the interstitial phase formation. The micrographs of these clinkers are shown in Fig. 5.2. The data are shown in Table 5.6.

Table 5.6. Microscopic Observations of Clinkers Containing Industrial By-products

The microscopical observations indicate that the by-products resulted in a partial decompo-sition and cannibalization of alite and also formed slightly irregular belite with ragged edges.

• control

Fig. 5.2 Optical micrographs of clinkers made with (a) copper slag, (b) sinter mix, (c) catalyst fines, and (d) the control clinker. In each case the field width is approximately 235 μm.

EVALUATION OF CEMENTS

Cements were prepared from these clinkers by mixing with appropriate amounts of gypsum as calculated by Eq. 2.1 and grinding to a Blaine fineness of 300 to 350 m2/kg. The cements were then tested both as pastes and as mortars.

PASTE TESTS

The pastes (w/c=0.485) were mixed in a WaringTM blender at a steady shear rate of 13,000 rpm and tested for initial setting, mini-slump cone test, and compression strength as in Task 2.

Initial Setting by Modified Vicat Test

Again as in Task 2, the initial setting time of the paste was determined by the modified Vicat method. The results are given in Table 5.7.

Table 5.7. Initial Setting of Cement Pastes Containing Industrial By-products

It appears that the addition of by-products does not impart any noticeable effect on the initial setting of pastes.

Mini-slump Tests for Early Stiffening

Mini-slump tests were conducted per procedure described in Task 2. The pat areas were measured at 2.5-, 5-, 15-, and 30-minute intervals as shown in Table 5.8.

Table 5.8. Mini-Slump Areas and Degree of Stiffening of Cement Pastes Containing Industrial By-products, cm2

The data on the difference in pat area between 2.5-minute and 30-minute tests do not suggest any clear-cut effect on the stiffening of the pastes.

Compressive Strength

The pastes were cast as 1-in. cube molds and tested for compressive strength at 3, 7, and 28 days, as explained in Task 2. The results are given in Tables 5.9 as follows:

Table 5.9. Compressive Strength of Paste Cubes Containing Industrial By-products, MPa (psi)

Additions of these by-products reduce the overall strength, except for the 28-day strength of catalyst-fines-doped cement pastes, which is close to that of the control.

MORTAR TESTS

Mortar batches were prepared from these cements using the proportions of cement, sand, and water in the ASTM C 109 procedure followed in Task 2.

Mortar Flow Measurement

The mortars were tested for their flow behavior as described in the ASTM procedure. The flow was recorded in Table 5.10 as follows:

Table 5.10. Flow of Mortars Containing Industrial By-products

When compared to the control, catalyst fines reduced the mortar flow. The other by-products did not impart any noticeable effects.

Compressive Strength

The 50-mm (2-in.) cubes cast from these mortars were cured in lime water and later tested for compressive strength at 3, 7, and 28 days. The average values are given in Table 5.11 as follows:

Table 5.11. Compressive Strength of Mortar Cubes Containing Industrial By-products, MPa (psi)

Copper slag and sinter mix improved the 28-day strength, whereas catalyst fines mortar strength is close to that of the control.

RESULTS AND DISCUSSION

The data suggest that the addition of by-products have some effects on clinker and cement properties. However, it is not certain to what degree these effects are attributed to the metals present in the by-products because, except for copper slag, the total metals (Ba+Cr+Cu+Mo +V+Zn+Mn) in the by-products are not significantly high (see Table 5.12). The total metals content based on clinker formation as shown below made with copper slag is 0.35%, of which copper is 0.21%. Ideally the effects of 0.21% copper should be similar to those with 0.25% copper as in Task 2, but the presence of other metals might have caused interference.

Table 5.12. Concentration of Total Metals (Wt. %) in Clinker Made with By-products

The overall effects of the by-products on clinkers and cements properties are discussed as follows:

CLINKERS

Clinker Color

These by-products did not significantly change clinker color from that of the control; their colors are tan/black, very similar to the control clinker.

XRF and Bogue Analyses

The XRF and Bogue analyses confirmed that Type I clinkers were produced from these by-products. Catalyst fines imparted an increase in C3S and corresponding decreases in C2S. The C3S and C2S for the copper slag and sinter mix were close to that of the control, except that their C3A contents were lower and C4AF higher than those of the control.

XRD Analysis

The C3S peaks for catalyst fines clinker are sharper and larger than the other by-products. This could be due to the fine particle distributions of the catalyst fines causing easy combinability of lime and silicates to form alite crystals. The C2S peaks are respectively smaller than the rest. Then again, with the exception of catalyst fines, other by-products have shown modification of both the C3A and/or C4AF peaks, which indicates that the trace metals in the by-product might have interacted with these phases. In all clinkers produced in this study, only a negligible amount of free lime has been noticed.

Microscopical Studies

Effect on Alite Formation: In general, all by-products have resulted in some degree of cannibalism and partial decomposition in alite crystals. Additionally, the by-product enlarged the average size of alite crystals.

Effect on Belite Formation: All other by-products produced round to slightly irregular belite crystals with ragged edges. The size of belite was close to that of the control.

Effect on the Interstitial Phases: All by-products produced smaller size C3A crystals than the control. In the case of catalyst fines, the porosity of clinker was also high.

Grindability

The grindability data indicate that clinkers containing these by-products became harder to grind than the control. The clinker hardness may be attributed to the cannibalistic nature of alite crystals which form large alite agglomerates (at the expense of belite crystals).

Although the effect of interstitial phases on clinker grindability is not defined, the proportion of interstitial phases appears less than the control with most of these by-products, except for that of catalyst fines, which is very close to the control.

CEMENTS

Paste Properties

Initial Setting: The initial setting times for cement pastes are fairly close to those of the control. Slightly prolonged initial setting times for sinter mix could be related to the increased belite formation (as suggested by the XRD tests). Small belite particles, on fringes for example, hinder the access of water to alites and retard hydration.

Early Stiffening: The addition of by-products has not imparted any unusual stiffening properties to the fresh cement pastes. The early stiffening of paste is controlled by the amount of C3A present and its interaction with gypsum. Very little gypsum appears to have been left after the early reaction with C3A, and the low C3S content in clinker might have been contributing to the stiffening of the paste structure.

Somewhat higher degrees of stiffening sinter mix are related to slightly prolonged initial setting times, most probably for the reason given above: that small belite crystals hinder the water access to alite crystals and retard hydration. Or perhaps the pastes are under-going some physical shrinking causing structural cohesiveness.

Compressive Strength: Copper slag and sinter mix decreased the overall strengths of pastes. However, for the catalyst fines, the 28-day strength is comparable to the control. The lower than the control early strength of pastes made with the by-product could be due to the increased C2S contents that contribute little to the early strengths. It also might be related to the decomposition of alite crystals in some cases which imparts abnormality to hydration, setting, and strength of paste.

Mortar Properties

Flow Behavior: As mentioned earlier in Task 2, in the present investigations, the water-cement ratio was kept at 0.485 in all the mortars and the flow was measured. Any devi-ation from the specified flow value of 110±5 will reflect the degree of workability of a given mortar. Catalyst fines appeared to have stiffened the mortars and negatively affected the compressive strengths when compared to the other by-products.

Compressive Strength: The 7- and 28-day strengths of mortars made with these by-products are at least comparable to or higher than the control. This is somewhat contrary to the strengths obtained on the respective cement pastes. It must be noted, however, that the rate of strength gain from 3 to 28 days for paste is significantly higher than those of the mortars.

This strength discrepancy between pastes and mortars might have resulted for the following reasons. Firstly, a mortar mix is different than the paste as it contains a substantial portion of sand, which creates a secondary discrete phase within the system. Therefore, a change in mortar strength because of the weaker interface between sand and paste can be expected. Secondly, because of the lower intensity of mixing in mortar, the cement in mortar may not be as well dispersed as in the paste. Thus, a variation in strength can be expected as compared to the pastes.

CONCLUDING REMARKS

Based on the data obtained and the observations made, it may be concluded that addi-tions of up to 0.5% by weight of single metals and multiple metals do not significantly affect the properties of clinkers or the cement prepared. The clinkers do not have any burnability problem as the free lime content is negligible and the C3S, C2S, C3A, and C4AF phases have appropriate concentration levels. Addition of single and multiple metals do not abnormally affect the paste hydration and strength development. In general, multiple metals tend to give high early strength of both pastes and mortars. However, the clinkers produced from the incorporation of single or multiple metals are harder to grind. Most clinkers have shown decomposed and cannibalized alites, whereas the belites are rounded with ragged edges. Mostly the metals are partitioned into the aluminoferrite phases, probably in ferrite due to metals’ affinity for iron.

One exception to the above is beryllium, which unusually affects the early hydration of fresh pastes. It is a severe retarder that delays the setting times and gives low strengths. It also abnormally stiffens the paste. At higher concentrations, zinc acts as an accelerator; it gives short setting time and generally high early strengths. Titanium adversely affects the early and late strengths. The addition of Cu-Pb-Zn accelerates the paste hydration, and enhances early strengths of both paste and mortar.

According to the XRD findings, titanium and antimony enhanced C3A at the expense of C4AF, whereas barium and beryllium increased C4AF at the expense of C3A. On the other hand, the Bogue analyses did not indicate any abnormality in the respective C3A and C4AF phases. Addition of Ni-V-Sb enhanced C3A at the expense of C4AF.

In cases where industrial by-products are used, it is critical that the material additions are made according to their composition and the level of metal they contain. Additions of by-products used in this study—copper slag, sinter mix, and catalyst fines—affected the crystallinity of both alite and belite; the alite crystals were slightly decomposed and cannibalized, whereas those of belite were irregular with some ragged edges which can adversely affect the grindability. Cements prepared from the by-products clinkers showed reduced paste strength but enhanced mortar strength at 28 days. However, the rate of strength gain for paste is higher than those of the mortars.

ACKNOWLEDGEMENTS

The research reported in this paper (PCA R&D Serial No. 2147) was conducted by Construction Technology Laboratories, Inc., with the sponsorship of the Portland Cement Association (PCA Project Index No. 93-01). The contents of this paper reflect the views of the author, who is responsible for the facts and accuracy of the data presented. The contents do not necessarily reflect the views of the Portland Cement Association.

REFERENCES

Bhatty, Javed I.; Marijnissen, J.; and Reid, K. J., “Portland Cement Production Using Mineral Wastes,” Cement and Concrete Research, Vol. 15, No. 3, pages 501 to 510, 1985.

Clark, C.; Meardon, K.; and Russel, D., “Burning Tires for Fuel and Tire Pyrolysis: Air Implications,” Section 4, Tire and TDF Use in Portland Cement Kilns, EPA-450/3-91-024, U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, December 1991.

De Zorzi, E., “Burning of Municipal Solid Waste in Cement Kilns,” 30th IEEE Cement Industry Technical Conference, Quebec, Canada, pages 461 to 476, May 24 to 26, 1988.

Gutteridge, W., “On the Dissolution of Interstitial Phases in Portland Cement,” Cement and Concrete Research, Vol. 9, pages 319 to 324, 1979.

Hansen, E., “Burning of Solid Waste in Cement Kiln,” World Cement, London, U.K., Vol. 24, No. 3, pages 15 to 18, 1993.

Heron, V., “Inland Cement Alternative Fuel Program,” World Cement, London, U.K., Vol. 24, No. 3, pages 41 to 46, 1993.

Hills, L., The Effect of Clinker Microstructure on Grindability: Literature Review Database, RP331, Portland Cement Association, Skokie, Illinois, 1995.

Hornain, H., “The Distribution of Transition Elements and Their Influences on Some Properties of Clinker and Cement,” Revue des Materiaux de Construction, Lafayette, Paris, France, No. 671-72, pages 203 to 213, August/September 1971.

Huhta, R. S., “Waste as a Fuel: A Survey Report,” Rock Products, Chicago, Illinois, pages 92 to 100, May 1990.

McGrath, B., “Used Lubricating Oil as a Fuel for Cement Kilns,” World Cement, London, U.K., Vol. 24, No. 3, pages 19 to 22, 1993.

Midlam, E. W., “Use of Solid Waste Material as Raw Material in Cement Process,” Report DOE/CS/15074-T2; Chemical Abstract 102,66401, 1985.

Moir, G. K., and Glasser, F. P., “Mineralizer, Modifiers and Activators in the Clinkering Process,” 9th International Congress of Chemistry of Cement, Delhi, India, Vol. 1, pages 125 to 152, 1992.

Obrist, A., “Burning Sewage Sludge in Cement Kilns,” World Cement, London, U.K., Vol. 18, No. 2, pages 47 to 59, 1987.

Patel, M., “Cement From Chloride Containing Wastes,” Silicates Industriels, Mons, Belgium, Vol. 3-4, pages 55 to 60, 1989.

PCA, An Analysis of Select Trace Metals in Cement and Kiln Dust, SP109, Portland Cement Association, Skokie, Illinois, 56 pages, 1992.

Smith, R. D.; Campbell, J. A.; and Nielson, K. K., “Concentration Dependence Upon Particle Size of Volatilized Elements in Fly Ash,” Environmental Science and Technology, Washington, D.C., Vol. 13, No. 5, pages 553 to 565, May 1979.

Effect of Minor Elements on Clinker and Cement Performance: A Laboratory Analysis

Sprung, S., “Technological Problems in Pyro-Processing Cement Clinkers: Cause and Solution,” Translation by Brodek, T. V., Beton-Verlag GmbH, Duesseldorf, Germany, 1985.

Takashima, S., “Systematic Dissolution of Calcium Silicate in Commercial Portland Cement by Organic Acid Solution,” Review of 12th General Meeting, Cement Association of Japan, Tokyo, pages 12 to 13, 1958.

du Toit, P., “Ono Evaluation of a Cement Clinker Produced Using By-Product Phosphogypsum as a Source of CaO,” Proceeding of the 10th International Conference on Cement Microscopy, San Antonio, Texas, pages 202 to 209, April 11 to 14, 1988.

Vogel, G. A.; Goldfarb, A. S.; Zier, R. E.; and Jewell, A., “Incinerator and Cement Kiln Capacity for Hazardous Waste Treatment,” Nuclear and Chemical Waste Management, Elmsford, New York, Vol. 7, No. 1, pages 53 to 57, 1987.

von Seebach, M., and Tompkins, J. B., “The Behavior of Metals in Cement Kilns,” Rock Products’ 26th International Cement Seminar, New Orleans, Louisiana, December 5, 1990.

Weatherhead, E. C., and Blumenthal, M. H., The Use of Scrap Tires in Rotary Cement Kilns, Scrap Tire Management Council, Washington, DC, 2005.

Weisweiler, W., and Krcmar, W., “Arsenic and Antimony Balances of a Cement Kiln Plant With Grate Preheater,” Zement-Kalk-Gips, Bauverlg GMBH/Maclean Hunter, Wiesbaden, Germany, Vol. 3, pages 133 to 135, 1989.

APPENDIX

Descriptions of some of the test procedures used for material characterization in the project are as follows:

• GRINDABILITY TEST

The laboratory grindability test developed at CTL entailed the following: 20 g clinker of 50 to 100 sieve size fraction (300 to 150 mm) is loaded into a jar ball with a fixed 200-g load of high density alumina balls. The balls have the following size range: 50% = 12.5-mm (1/2-in.) diameter, 30% = 10-mm (3/8-in.) diameter, and 20% = 6-mm (1/4-in.) diameter. The clinker is ground for 3 hours. The grinding time of 3 hours has been chosen because it gave a Blaine fineness of 300 to 350 m2/kg for the control clinker, which is comparable to the fineness of ordinary portland cement in routine grinding. At the end of grinding, the ground clinker is examined for its Blaine, which is termed as its “grindability index.”

• SELECTIVE DISSOLUTION TECHNIQUES FOR CLINKER PHASES

Salicylic Acid-Methanol Extraction [modification of Takashima method (1958)]: This disso-lution technique primarily removes silicates, i.e., alite, belite, and free lime from clinker and leaves a residue mainly composed of aluminates (C3A) and ferrite (C4AF).

In this method, a 20 g sample of ground clinker (-200 mesh) is dissolved in 1 L of 12% salicylic acid/methanol solution and stirred for 1 hour. Thereafter, the solution is filtered and washed. The residue is air-dried and stored for further analyses.

KOH-Sugar Extraction [modification of Gutteridge method (1979)]: This extraction primarily removes the aluminates and ferrites, leaving a residue mainly composed of sili-cates, i.e., alite and belite.

In this extraction method, a 5g sample of ground clinker (-200 mesh) is dissolved in 150 mL hot 13% KOH/13% sucrose solution and stirred for about 5 minutes during which the solution typically turns from gray to creamy color. The solution is filtered on Whatman GF/A glass fiber filter and the residue is washed with hot water. The residue is air-dried and stored for further analyses.

• CALCULATION OF GYPSUM ADDITION TO CLINKER

Given that the clinkers are Type I low alkali, gypsum addition was calculated per SO3/Al2O3 ratio. Value of Al2O3 was taken from clinker XRF. Since a typical suggested level of SO3 is equivalent to 0.6 molar of Al2O3 in clinker, calculation for gypsum was as follows:

Amount of gypsum used = Al2O3 content X [0.6/101.96] X [80/0.461]

Where,            101.96 = mol. wt. Al2O3

80 = mol. wt. SO3

0.461 = mol. ratio SO3/natural gypsum used (for pure gypsum 80/172 is 0.466)

• MODIFIED VICAT INITIAL SETTING TEST

The initial setting times of the pastes were determined by a modification of ASTM C 191, Standard Test Method for Time of Setting of Hydraulic Cement by Vicat Needle. This method measures the penetration by a 1-mm (1/32-in.) needle into cement paste contained in a 100-mL (3.4-oz) plastic cup.

A portion of the paste prepared in the WaringTM blender is loaded in the 100-mL plastic cup near to the top and the surface is leveled. The cup is then placed on a vibrating table to remove air traps. The cup is then removed from the vibrating table and place on the Vicar apparatus plate form and covered with a glass dish. Penetration of the Vicat needle in the paste is measured at different time intervals by allowing the needle to drop into the paste until the penetration is no more than 10 mm. This determines the initial setting time of cement paste.

• MINI-SLUMP TEST

Pastes (mixture of cement and water) were prepared in a WaringTM blender at a water-cement ratio of 0.485. The mixing was done at 13,000 rpm to simulate the mixing conditions of a concrete mixer. A clean mini-cone is placed on a smooth dust-free plastic board. The fresh paste is steadily poured into the cone until it is completely filled. The cone is gently ridded with a spatula to remove trapped air and ensure complete filling. Any additional paste is removed from the top and the surface is leveled. Then the cone is quickly lifted upward, allowing the paste to flow out and collapse under gravity. The paste typically flows out as a nearly circular pat. The test is done at 2.5-, 5-, 15-, and 30-minute intervals using fresh portions of the paste each time. The paste in the blender is kept agitated slowly to avoid bleeding. The areas of the pats are measured with a plan meter and recorded as an indication of the stiffening or flowing behavior of the paste. Generally, larger pat area indi-cates greater workability and less early stiffening. However, a significant decrease (typically exceeding 20% is typical) in pat area with time (i.e., a decrease from the 2.5-minute to the 30-minute pat area) indicates early stiffening of the paste.

• PASTE COMPRESSIVE STRENGTH

Twenty five-mm (1-in.) cubes were cast in stainless steel molds. The molds are thoroughly cleaned, then lightly coated with oil and placed on a small vibrating table. The paste is loaded in the vibrating molds and each cube cavity is nearly filled. A flat laboratory

Appendix

spatula is then used with an up-and-down motion, first on the sides and then in the corners of the molds, to remove bubbles and consolidate the paste. The molds are then slightly overfilled with paste to compensate for shrinkage, and the spatula again used sparingly with the same motion to consolidate the paste. The molds are placed in a moist room at 100% relative humidity (R.H.) and 23°C. The cubes were demolded the next day, and kept in the moist room for curing. The compressive strength tests were done after 3, 7, and 28 days of curing under these conditions.