Characteristics of waste catalyst reused as latent hydraulic materials

Abstract

This study investigated the pozzolanic reactions and engineering properties of waste catalyst (WC)-blended cements in relation to various replacement ratios (0–40%). The latent hydraulic characteristics were evaluated at various curing ages by examining the hydration products and compressive strength of the waste catalyst blended cement (WCBC) pastes. The results indicate that the WC contained Al₂O₃ (59 w/w %), SiO₂ (39 w/w %), and CaO (0.1 w/w %) and the pozzolanic activity index of the WC was 117%. The heavy metal leaching concentrations all met the Taiwan EPA regulatory limits. The early compressive strength of the resultant blended cement pastes (1–28 days) was lower than that of the ordinary Portland cement (OPC); however, WCBC pastes with a replacement of cement by up to 20% showed a later compressive strength (60–90 days) similar to that of OPC paste. The hydrated WCBC pastes were similar to OPC pastes in terms of hydration product composition. The main hydration products in the WCBC pastes were Ca(OH)₂ and C-S-H (Tobermorite). The results indicate that it is indeed feasible to use WC as one of the mineral components for cement. © 2011 American Institute of Chemical Engineers Environ Prog, 32: 94–98, 2013.

INTRODUCTION

Large quantities of catalysts are used in the petroleum refining industry to purify otherwise improve various petroleum streams and residues. The catalysts deactivate over time. When the activity of a catalyst falls below an acceptable level, it is generally regenerated and reused. However, regeneration is not always possible [1, 2], and over a few cycles of regeneration and reuse, the catalyst activity may decline to very low levels and further regeneration may not be economically feasible. Such waste catalysts (WCs) are discarded as solid wastes [3, 4]. The quantities of WCs that are discharged from various processing units depend strongly on the amount of fresh catalysts used, their lifetime, and the amount of deposits that formed on them during their use in the reactors. In most refineries, a large fraction of the WC is from the residue hydrotreatment and hydroprocessing units, because the catalysts that are used in these processes are rapidly deactivated by coke and metal (V and Ni) deposits, and have a short lifetime [5-9]. The annual production of catalysts by the petroleum refining industry in Taiwan is 13,419 tonne. Sanitary landfills are commonly used to dispose these catalysts, but rapid urbanization has made locating suitable landfill sites increasingly difficult [10]. Additionally, environmental laws concerning WC disposal have become increasingly strict in recent years. Increasing demand for natural resources and a scarcity of environmentally acceptable solid waste disposal sites are motivating numerous municipalities in Taiwan to consider resource recovery as an alternative.

The recycling of WCs as a cement constituent may solve the disposal problem. Considering the landfill capacity limitation for the disposal of WCs and the energy consumption during the cement production process, it is without a doubt that the recycling of WCs as partial substitution for cement can be an economically beneficial alternative. Therefore, from a viewpoint of life cycle analysis, the recycling of WCs as substitute for cement would seem economically feasible.

The major components in WCs, such compounds, which are common in ordinary Portland cement (OPC), may function as binders or exhibit self-hardening characteristics. This indicates that the WCs, by nature, a latent pozzolan and can be used to partially replace cement in cement paste or concrete [11]. The properties of WC are close to those of the common mineral pozzolan, such as blast furnace slag and coal fly ash. The work presented here is concerned with the preparation of waste catalyst blended cement (WCBC). The hydraulic activity, compressive strength development, microstructure variations, and the heavy metal leachability of the WCBC pastes are all studied.

MATERIALS AND METHODS

Materials

The WCs used was collected from the refinery plants located in Taoyuan City, Taiwan. The waste brick was homogenized, oven dried at 105°C for 24 h and then the chemical composition was characterized. American Society for Testing and Materials (ASTM) Type I Portland cement (OPC) from the Taiwan Cement Company was used in this study. It had a specific gravity of 3.15 and its physical–chemical properties met the requirements of ASTM C150 [12]. The major components of the OPC are shown in Table 1.

Table 1. Chemical composition of OPC and WC

Chemical composition	Cement	Waste catalysts
CaO (%)	64.5	0.1
SiO2 (%)	21.6	39.2
Al2O3 (%)	5.3	59.4
Fe2O3 (%)	3.4	—
Na2O (%)	0.01	—
K2O (%)	0.01	—
MgO (%)	2.25	—
Pozzolanic activity index (%)		117

Preparation of WC

The WC was then further pulverized in a ball mill until the particles could pass through a 200 mesh (74–104 μm) screen. It can be observed that 29% (wt) of the particles in the WC have a diameter of greater than 74 μm and 71.1% of particles have a diameter of less than 74 μm. The WC was ground to a fineness value (on Blaine) of approximately 500 m² kg⁻¹, with a specific gravity of 1.86. The resultant pulverized WC was desiccated before being tested. The pozzolanic activity of waste brick cured for 28 days was analyzed according to ASTM C 311 [13], and the results are also presented in Table 1.

Preparation of WC Blended Cement Paste Specimens

WCBC pastes were produced by homogeneously mixing cement and water in a mixer. The levels of WC substitution in the blended cement were between 10, 20, 30, and 40% by weight of the cement. The water/binder (w/b) ratio of the WCBC paste was kept constant at 0.4. WCBC paste cubes were prepared according to ASTM C109 [14]. The WCBC paste was poured into rectangular moulds (5 cm × 5 cm × 5 cm), which were kept under ambient conditions (25°C with a relative humidity greater than 98%) for 24 h before being de-moulded. The de-moulded specimens were then cured in a climate chamber maintained at 25°C, with a relative humidity greater than 98%, for periods ranging from 7 day to 90 days. After curing for 7, 14, 28, 60, and 90 days, the samples were subsequently crushed. The hydration reactions of the WCBC specimens were terminated with absolute alcohol. A laboratory study was also performed for the feasibility of dewatering samples by vacuum filtration. Furthermore, WCBC specimens were washed several times with acetone and then dried completely. Finally, the prepared samples were subjected to X-ray diffraction (XRD), gel/space ratios, and degree of hydration analyses.

Analysis Methodology

The major analyses performed on the WCBC pastes and the cubic specimens included the following:

1. Pozzolanic activity index: The test was performed according to ASTM C 311 [13].
2. Setting time: The setting times of the cement mixes were determined according to ASTM C191 [15] using a Vicat apparatus at room temperature.
3. Unconfined compressive strength: The test cubes were prepared according to ASTM C109, followed by a moulding process (ASTM C31-69). The specimens were then demoulded and cured in a container at 95% humidity, and 25°C for 3–90 days. At each testing age, four specimens were taken out of the moist room. The surfaces of the specimen were polished by gypsum to make the two bearing surfaces flat and parallel. Three specimens were used for the compressive strength tests and one for the microstructural examination. The average strength value of the three specimens is presented. The coefficient of variation of these results was less than 10%.
4. Toxic characteristic leaching procedure (TCLP) [[16]]: The extraction procedure requires the preliminary evaluation of the pH characteristic of the sample to determine the proper extraction fluid necessary for the experiment. Upon testing, extraction fluid no. B (pH 2.88 ± 0.05) was used for the TCLP analysis. This fluid was prepared by adding 5.7 mL acid to 500 mL double distilled water, and diluting to a volume of 1 L. A 25-g sample was placed in a 1-L Erlenmeyer flask, and 500-mL extraction fluid was added to each Erlenmeyer flask. These samples were agitated for 18 h using an electric vibrator. The slurry was filtered by 6–8 μm pore size Millipore filter paper. The leachates were preserved in 2% HNO₃.
5. Leaching concentration: Additionally, inorganic toxicity was determined by following method: SW-846, Method 1311, analyzing the presence of the following elements: Cd (SW 846–7131A), Pb (SW 846–7421), Zn (SW 846–7951), Cu (SW846–7211), and Cr (SW 846–7191).
6. Mineralogy: The XRD analyses were carried out by a Siemens D-5000 X-ray diffractometer with CuKα radiation and 2θ scanning, ranging between 5° and 70°. The XRD scans were run at 0.05° increments, with a 1 s counting time. A detailed JCPDS search/match was performed to identify the major crystalline phases that are consistent with various peaks in the diffraction pattern.
7. Gel/space ratios and degree of hydration: The degree of hydration of the WCBC pastes was determined by thermal analysis. Thermogravimetric analysis instrument was employed to determine the degree of hydration in the WCBC paste samples using the ignition method. The hydration degree of the WCBC pastes was then calculated as follows [17, 18]:
   • (1)
   • (2)
   where α: degree of hydration, %, X: gel/space ratio, %, n: evaporated water in the completely hydrated specimen; n is equal to 0.24 for OPC paste;,W₁₀₅, W₅₈₀, W₁₀₀₇: sample weights (g) at 105, 580, and 1007°C, respectively, 0.41: mass ratio was 1 mol H₂O to 1 mol CO₂, W: Amounts of water, %, C: Amounts of cement, %, P: Amounts of WC, %.

RESULTS AND DISCUSSION

Characteristics of the WC

Table 1 presents chemical composition analysis of the WC done by X-ray fluorescence (XRF). WC samples were primarily composed of SiO₂ (39%), Al₂O₃ (59%), and CaO (0.1%). Chemical analysis (Table 1) demonstrates that WC samples had insignificant amount of CaO and substantial amounts of SiO₂ and Al₂O₃ compared to those in the cement. Figure 1 shows XRD patterns of WC. The WC was composed of SiO₂and Al₂O₃. The pH of WC was 5.6. The specific surface area determined by the Blaine air-permeability method was found to be 500–580 m² kg⁻¹. The WC had a pozzolanic strength activity index of 117% at 28 days. In terms of strength, WC can be considered a good pozzolanic material. The TCLP leaching concentrations of WC all met current regulatory thresholds established by Taiwan's Environmental Protection Agency (EPA) (Table 2).

Figure 1. XRD patterns of WC.

Table 2. Total metals and TCLP leaching concentrations of WC

Heavy metal	Total metal (mg/kg)	TCLP (mg/L)	Regulatory thresholds
(mean ± standard deviation), N = 3. ^a Zn < 0.015 mg/L. ^b Pb < 0.014 mg/L. ^c Cd < 0.016 mg/L.
Zn	46.8 ± 1.12	NDa	—
Cr	9.8 ± 0.43	0.18 ± 0.03	5
Cu	9.6 ± 0.07	0.08 ± 0.01	15
Pb	21.5 ± 1.03	NDb	5
Cd	0.2 ± 0.01	NDc	1

Setting Behavior of WCBC Pastes

As the cementitious hydraulic reactions progress, the cement specimen gets hardened with time and develops strength. Basically, the setting behavior is an action of the cement grains dispersing and hydrating in water, followed by gradually forming a solid/liquid suspension of the various hydrates. As the process goes on, the inner structure of the cement specimen is reinforced and forms a network structure which makes the cement specimen set and gain strength. Figure 2 lists the setting times for WCBC pastes. The initial setting times and final setting times of OPC were 3 h 45 min. and 5 h 20 min, respectively. The setting behavior of WCBC pastes varied widely—initial setting times were 4 h 35 min to 5 h 3 min, and final setting times were 5 h 38 min to 6 h 12 min. The initial and final setting times of the WCBC pastes were longer than those for OPC pastes. Setting times for WCBC pastes increased when the amount of WC in the pastes increased. This implies that the longer setting times of the WCBC pastes may be due to the presence of relatively higher amounts of SiO₂ and Al₂O₃ compared to OPC pastes.

Figure 2. Setting time of WCBC pastes.

Compressive Strength Development of WCBC Pastes

Figure 3 shows the compressive strengths of WCBC pastes values at different times (up to 90 days). It can be seen that amount of WC blended into the pastes affected the strength. The strengths of WCBC pastes containing 10% and 20% WC were similar to those of OPC paste at 90 days. The WCBC pastes containing 40% WC had the lowest compressive strengths at all ages. The presence of SiO₂ and Al₂O₃ (up to the levels in blended cements) may also enhance compressive strength. Notably, the volume change in SiO₂ in a pozzolanic reaction is based on the C-S-H generated during the late stages of the pozzolanic reaction. The high compressive strengths of WCBC pastes (containing up to 10% and 20% WC) were consistent with the high amounts of C-S-H, C-A-H, and C-A-S-H in hydration products. The poor strengths of the WCBC pastes containing 40% WC alone can be attributed to a reduced amount of Ca(OH)₂ to activate the WC.

Figure 3. Compressive strength development of WCBC pastes.

Hydration Products in WCBC Pastes

Figures 4 and 5 show the X-ray patterns of WCBC pastes at 28 and 90 days curing time, respectively. The phases of OPC pastes were calcium hydroxide (Ca(OH)₂, 2θ = 18.1°, 47.25°, and 50.90°) and some unreacted C₃S (2θ = 29.75°) and C₂S (2θ = 31.15°). The WCBC pastes had very similar hydration products, as expected. The OPC pastes showed a higher Ca(OH)₂ peak at 28 days than the WCBC pastes; a result also supported by its lower hydration activity and strength. Various reaction products formed, including hydrates of calcium silicates (CSH, 2θ =32–34°) [19-21] and aluminosilicates (CASH, 2θ = 12.35), with consequent consumption of Ca(OH)₂. This indicates the progress of pozzolanic reactions in which active silica and alumina (Si, Al) react with cement hydrates Ca(OH)₂ to form C-S-H and CASH. This resulted in a very densified and homogeneous system, and a considerable increase in long-term strength. However, the excessive replacement of cement, up to 40%, resulted in a relative deficiency of the cement hydrate, Ca(OH)₂, thus decreasing the total volume of calcium silicate hydrates (C-S-H), their deposition density, and thus the compressive strength of the WCBC pastes.

Figure 4. XRD patterns of the WCBC pastes after 28 days curing time.

Figure 5. XRD patterns of the WCBC pastes after 90 days curing time.

Gel/Space Ratio in WCBC Pastes

It is well-known that the compressive strength of concrete depends on the gel/space ratio determined from the degree of cement hydration and the w/c ratio [22]. In other words, the volume of gel is 2.2 times the original volume of the cement. This increase in volume implies that paste porosity decreases as the hydration proceeds [23]. The gel/space ratio is defined as the ratio of the volume of hydrated cement to the sum of the volumes of hydrated cement and of capillary pores [24]. Figure 6 shows the gel/space ratios for WCBC pastes. The gel/space ratio increased as curing time progressed due to the progression of hydration, which caused partial filling of pores. The gel/space ratio of WCBC pastes containing 10% WB was similar to that of OPC pastes at 90 days. This is likely partially due to the low density of pozzolanic hydration products, and may indicate that pozzolanic reaction products are very effective in filling pores. At 40% WC, the gel/space ratio decreased significantly. The gel/space ratio data are strongly correlated with compressive strength. Increasing the gel/space ratio, increased the amount of hydration products, and the consequent synergistic effect of hydration products in pores of cement pastes enhanced compressive strength.

Figure 6. Gel/space ratio in WCBC pastes.

Degree of Hydration in WCBC Pastes

Various reaction products including hydrates of calcium silicates (CSH), aluminates (CAH), and aluminosilicates (CASH), would be expected to form, with consequent consumption of Ca(OH)₂ [25]. Figure 7 shows the amounts of hydration products in WCBC pastes. At different curing ages, the degree of WC reaction depends on the WC content. The degree of hydration for WCBC pastes increased as curing time increased. An increase in hydration of WCBC pastes decreased Ca(OH)₂ content, indicating that an active pozzolanic reaction occurred. Evidence of the WC reaction was also demonstrated by reduced Ca(OH)₂ content. At 90 days, the WC showed some etching of SiO₂ and Al₂O₃ materials. Notably, Ca(OH)₂ on WC particles undergoes redissolution and react with Ca(OH)₂. As the replacement ratio increased from 10% to 40%, hydration decreased. Hydration of WCBC pastes with high WC content decreased at 28 days. The hydration of WCBC pastes decreased at 90 days, particularly for the paste with 40% WC. Generally, the WCBC pastes with high percentages of WC had low levels of hydration. This may be due to the low concentration of Ca²⁺ ions in pore solutions of the pastes. These analytical results are consistent with compressive strength test results, and likely indicate that the hydration enhancement effect in WCBC pastes is significant. A 10% WC addition had little effect on the relative Ca(OH)₂ content in WCBC pastes. As both Ca(OH)₂ and CSH precipitate on the surfaces of WC particles, pozzolanic reaction at later ages strengthens the contact between cement and WC and between WC particles. This leads to a more densified and more homogeneous system, and a considerable long-term strength increase.

Figure 7. Degree of hydration in WCBC pastes.

CONCLUSIONS

This study investigated pozzolanic reactions and strength properties of WCBC with various replacement ratios (0–40%). The following conclusions are based on experimental results. WC samples were primarily composed of SiO₂ (39%), Al₂O₃ (59%), and CaO (0.1%) and had a pozzolanic activity index of 117%. The TCLP results demonstrated that the heavy metal content met EPA regulatory limits. The initial and final setting times of the WCBC pastes were longer than those for OPC pastes. Setting times for WCBC pastes increased when the amount of WC in the pastes increased. The compressive strength of WCBC specimens developed slowly in the beginning comparatively, while increased relatively rapid in the later ages. Various reaction products formed, including hydrates of calcium silicates (CSH, 2θ =32–34°) and aluminosilicates (CASH, 2θ = 12.35), with consequent consumption of Ca(OH)₂. The compressive strength of WCBC specimens can be enhanced by increasing the gel/space ratio, the degree of hydration, and the amount of hydration products which come from the consequent synergistic effect in the pores of the cement pastes. As both Ca(OH)₂ and CSH precipitate on the surfaces of WC particles, pozzolanic reaction at later ages strengthens the contact between cement and WC and between WC particles. This leads to a more densified and more homogeneous system, and a considerable long-term strength increase. These experimental results indicate that using WC as a cementitious material is feasible. The results indicated that it is feasible to use WC to replace up to 20% of the OPC.