Coal samples used in this study were obtained from a mine located in Southern Illinois, USA. The as-received run-of-the-mine coal was first crushed in a jaw crusher (2 × 6 model, Sturtevant Inc., Hanover, MA, USA) and then further crushed in a roll crusher (8 × 5 model Sturtevant Inc., Hanover, MA, USA) to a size of 850 μm. Finally, a laboratory ball mill was used to grind the coal samples to a finer size. After crushing and grinding, the coal was screened to −75, +38 µm to be used as a feed in all flotation experiments. The particle size distribution of the flotation feed showed a P₈₀ value of 45 um. The proximate analysis of the −75, +38 µm coal (flotation feed) is shown in Table 1. Proximate analysis was conducted as per ASTM D3172-13 standards. All chemicals used in the flotation experiments were purchased from Fisher Scientific (USA). This includes kerosene, which was used as a coal collector; methyl isobutyl carbinol (MIBC), which was used as a frothing agent; sodium metasilicate, which was used as a dispersant in synergistic studies; and polyacrylamide polymer, which was used in comparative studies. Sodium hydroxide (NaOH) and hydrochloric acid (HCl) were used to adjust the pH of the flotation pulp and were also purchased from Fisher Scientific (USA).

The hybrid polyacrylamide Al-PAM was synthesized in-house according to a procedure described elsewhere [14] . Al-PAM was synthesized by the polymerization of acrylamide monomers in an Al(OH)₃ colloidal suspension using (NH₄)₂S₂O₈/NaHSO₃ as a redox initiator. To prepare the Al(OH)₃ colloid, approximately 0.33 g of anhydrous aluminumchloride (AlCl₃, >99%) were dissolved in distilled water to make 25 g of a 0.01 M AlCl₃ solution. Approximately 0.48 g of ammonium carbonate ((NH₄)₂CO₃) were dissolved in distilled water in a different beaker to make 50 g of 0.1 M (NH₄)₂CO₃. A Master FLEX mini pump was used to gradually add the (NH₄)₂CO₃ solution to the AlCl₃

solution at a rate of 0.5 g/min. After adding approximately 36 - 37 g of the (NH₄)₂CO₃ solution into 25 g of AlCl₃ solution, the pump was stopped and the solution was gently stirred at a rate of 300 rpm for approximately 30 min to complete the reaction. The prepared Al(OH)₃ suspension has an approximate particle size of 30 - 50 nm, and the measured zeta potential value was +27, −30 mV. The particle size and zeta potential studies were measured using a Zetasizer Nano ZS (Malvern, USA) The next step was the polymerization of acrylamide in an Al(OH)₃ colloidal suspension to synthesize the Al-PAM polymer. The polymerization process was achieved according to the following procedure: approximately 4.5 g of acrylamide monomer was dissolved in 25.5 mL of previously prepared Al(OH)₃ colloidal suspensions. The suspension was stirred and heated in an oil bath to 40˚C under nitrogen flow for nearly 30 min. Two milliliters of 1000 mg/L stock solutions of redox initiators (1 mL each) were slowly added to the prepared Al(OH)₃ suspension containing acrylamide monomer to initiate the polymerization of acrylamide onto colloidal aluminum hydroxide particles. The polymerization reaction was kept running overnight. The formed Al-PAM gel was dissolved in MilliQ water to 10 wt%. The aqueous polymer colloidal suspension was then introduced into acetone drop-wise to precipitate Al-PAM. The resulting solid precipitates were dried under vacuum at 55˚C for 6 - 8 hours. Freshly prepared 1000 mg/L stock solutions of Al-PAM polymers in deionized water were used in all of the flotation experiments.

The aluminum content by weight in Al-PAM polymer was measured using the PerkinElmer Inductively coupled plasma system 2000 DV instrument equipped with Optical Emission Spectrophotometer and WinLab32 for ICP version software for measurement of Aluminum at wavelength 396.153 nm. The RF power was used 1500 watts, Plasma flow was 15 L/min, Auxiliary Flow was kept at 0.2 L/min, and Nebulizer Flow was kept at 0.8 L/min, Pump Rate was kept at 2 ml/min. The calibration curve plotted using 7 different concentrations of solutions diluted in 1% HNO₃ (0 mg/L, 0.1 mg/L, 0.5 mg/L, 1 mg/L, 5 mg/L, 10 mg/L, and 25 mg/L) made by using HIGH-PURITY STANDARDS purchased from Fisher-Scientific. The correlation coefficient was 0.9999. Sample measurements were repeated in 3 times. The sample contains 0.14 wt% of aluminum.

Molecular weight of Al-PAM was determined using a Zetasizer Nano (Malvern Instruments Ltd., USA) by employing static light scattering (SLS). Zetasizer Nano software measures the intensities of the scattered light of the polymer sample and automatically calculates the molecular weight by applying the Rayleigh Equation (1) below [15] .

where Rθ is Rayleigh ratio which is the ratio of scattered light to incident light; M_W is the molecular weight; A₂ is the 2^nd Virial Coefficient which represents the interaction strength among the polymer chains and has been correlated with solubility [16] ; C is the concentration of polymer solution; P(θ) is the angular dependence of the sample scattering intensity and K is the optical constant as defined below in Equation (2).

where N_A is Avogadro’s constant; l_o is the wavelength of laser used; n_o is the refractive index of the solvent and dn/dc is the differential refractive index increment which is the change in refractive index as a function of the change in concentration. Therefore, a plot of KC/Rθ versus C is expected to be linear with an intercept equivalent to 1/M_W and a slope equal to the second virial coefficient A₂. This plot is known as “Debye plot”. To establish Debye plot of Al-PAM, samples of different concentration; 15, 13.5, 10.5 and 7.5 mg/ml were prepared. The calibration was made by using pure (>99.5%) toluene. All measurements were conducted at 25˚C. The Debye plot for the Al-PAM data is shown in Figure 1, which indicates a molecular weight of 6.180 × 10⁶ Da and a 2nd virial coefficient of 0.000126 ml∙mol/g².

An XRD analysis was used to reveal qualitative information on the mineral composition of raw coal samples (−75 µm +35 µm) used in this study. The XRD analysis was performed using a PANalytical X’Pert Pro Multi-Purpose Diffractometer (MPD; PANalytical, Inc., MA, USA) system with a Cu (k-alpha)-source at a tube voltage of 40 kV. Table 2 and Figure 2 show the quantitative and qualitative analyses, respectively, of the mineral matter composition in the coal samples obtained from the XRD analysis. The proportion of amorphous material to crystalline material in the coal sample was determined to be 76.2% by weight, which is in a good agreement with ash contents determined

by proximate analysis. The major crystalline phases observed were FeS₂ (pyrite), quartz, kaolinite Al₂(Si₂O₅)(OH)₄, corundum Al₂O₃ and small amounts of calcite Ca(CO₃).

An RA is the analogous counterpart in coal froth flotation to the float and sink methods in coal gravity concentration. The major objective of an RA is to obtain the best possible separation performance by any froth flotation process [17] . The flotation RA was carried out in a conventional laboratory flotation cell. Figure 3 shows the timed RA result for a coal flotation feed sample: combustible recovery vs. product ash. At a product ash content of 8.2%, the combustible recovery of coal is approximately 58% - 59%. The timed RA data indicated that the coal sample used in this study had poor floatability and that further cleaning of this coal sample to produce a product ash of 5% - 6% would be difficult. The combustible recovery of coal decreases sharply as the ash products decreases to less than 6%. RA was also used to obtain a clean coal product (<4% ash) for comparative zeta potential measurements.

The zeta potential of the coal samples were measured using a Zetasizer Nano ZS instrument (Malvern Instruments, Inc., Westborough, MA, USA). Data on the interactions of coal samples with ash-forming minerals in the flotation pulp were obtained by measuring the zeta potential of the raw coal samples and the froth obtained from the polymer-assisted flotation process and comparing the results to the zeta potential of clean coal obtained from the RA. The experiments were conducted using a 0.01 M KCI electrolyte solution at natural pH. Zeta potential measurements were performed for raw coal (i.e., the flotation feed), the first concentrate produced from the RA (herein referred to as clean coal with an ash content of 3.91%) and froth collected from the polymer-assisted flotation. Sample measurements were repeated 3 times. Isoelectric graphs of the clean and raw coal were obtained using MPT-2 autotitrator which is part of the Zetasizer Nano. The solution pH was adjusted using HCl and NaOH. Coal suspensions were prepared at 1 wt.% coal in a 0.01 M KCI solution. Each prepared coal suspension was agitated using an IKA RW 20 mechanical stirrer for approximately 30 min at a constant agitation rate of 300 rpm. The suspension was then allowed to settle for 5 - 10 min. The upper portion of the supernatant was considered for the zeta potential distribution measurements.

Batch flotation experiments were conducted to study the effects of novel organic/inorganic (hybrid) polyacrylamide polymers in enhancing the combustible recovery and ash rejection of coal flotation under various operating parameters. All flotation experiments were conducted 3 times as indicated by the error bars in the figures. In the study, various parameters, such as polymer dosage, impeller speed, pH and conditioning time, were assessed individually. All of the flotation experiments were run in a 5-L D12-Denver flotation laboratory cell at natural pH unless otherwise stated. The air flow rate and solids concentration were kept constant at 6 L pm and 5%, respectively, in all experiments. In a typical flotation test, 253 g of coal and 4800 mL of tap water were loaded in the 5-L cell, and the pulp was conditioned to allow wetting of the coal for 5 min prior to any reagent addition. Collector (i.e., kerosene) was then added at a predetermined dosage, and the suspension was conditioned for an additional 3 min. A desired dosage of Al-PAM was added after the slurry was conditioned with the collector, and the pulp was agitated for another 3 min. When the dispersant was used, it was first added before adding the collector. The polymer and dispersant dosages are expressed in reference to the total volume of the feed slurry (i.e., the combined volume of coal and water). The frothing agent (MIBC) was added at a fixed amount of 200 μL/ton on a mass basis relative to the dry feed mass. The pulp was further conditioned for 2 min before air was introduced. The resultant froth was collected at 1-min time intervals. The concentrate fractions were washed, filtered and dried in an oven overnight at 80˚C. After drying, the ash contents of the concentrates were determined according to ASTM D3174-73 [18] .

Flotation kinetic tests were conducted to determine the effects of Al-PAM on the flotation rate constants. All experiments were conducted in the presence or absence of Al-PAM at the optimum dosages for comparison purposes. In each test, the froth product was collected at 1-min time intervals; 5 fractions were collected. The collected fractions were washed, filtered and dried in an oven overnight at 80˚C. After drying, the ash contents of the concentrates were determined according to ASTM D3174-73.

Experiments were conducted to evaluate the use of a dual Al-PAM-dispersant system to improve the combustible recovery of coal and ash depression during coal flotation. First, floatation experiments were conducted using Al-PAM alone at various dosages. These experiments were mainly conducted to determine the optimum dosage of Al-PAM, as shown in Figure 4. Then, flotation experiments were conducted at the optimum dosage Al-PAM with varying dosages of sodium metasilicate dispersant. The Al-PAM dosage was kept constant at 0.25 mg/L throughout the experiments. In each experiment, the dispersant, collector, Al-PAM and frothing agents were added sequentially.

As shown in Figure 12, the overall combustible recovery of coal was improved using the dual Al-PAM-dispersant system. The combustible recovery of coal increased to 63.80% at sodium metasilicate dosages between 0.2 and 0.8 mg/L. A slight decrease was observed in the coal recovery when the dual polymer-dispersant system was used compared with polymer alone. This depression may be due to the adsorption of sodium metasilicate on the surface of the coal particles, imparting hydrophilicity, prior to polymer addition. However, unlike combustible recovery, the overall product ash was greatly improved when the dual Al-PAM-dispersant system was used compared with the use of dispersant or Al-PAM alone. A product ash with an improved ash content of 7.4% was obtained when the dual Al-PAM-dispersant system was used when compared with the ash content of 8.11% observed for a system using dispersant alone. When sodium metasilicate was added to the suspension prior to adding Al-PAM, the silicate dissociates to produce highly electronegative anionic species, which then electrostatically adsorb onto the ash-forming minerals (and to a lesser extent, on coal particles) and causes the charge density in an electric double layer around the coal and ash-forming mineral particles to increase. This results in increased electrical repulsion and thereby disperses the coal and ash-forming mineral particles in the suspension. The dispersion effect of sodium metasilicate allows Al-PAM to easily adsorb on the surface of high-ash-forming minerals (slimes), facilitates their aggregation and decreases the entrainment of ash particles in the froth layer.

To further investigate the dual use of the Al-PAM-dispersant system, flotation experiments were conducted at a slightly acidic pH of 5 and an alkaline pH of 10. In this set of experiments, the polymer dosage, dispersant dosage, impeller speed and polymer conditioning time were kept constant at 0.25 mg/L, 0.8 mg/L, 1800 rpm and 6 min, respectively. Figure 13(a) and Figure 13(b) show the effect of pH on the clean coal combustible recovery and product ash in the presence of dispersant alone or with the dual Al-PAM-dispersant system. As seen from Figure 13(a), the use of dispersant or the dual use of the Al-PAM-dispersant system did not improve the combustible recovery of coal at pH 5 or 10. The observed decrease in the combustible recovery at the acidic pH was due to better ability of sodium metasilicate to dissociate to form colloidal silica gels. These colloidal silica gels are negatively charged at acidic pH and will readily adsorb onto the positively charged coal surfaces, rendering them hydrophilic and allowing for high-ash-forming minerals to float. These results correspond well with the results reported in the literature [25] . As shown in Figure 13(b), higher product ash content was obtained when a dispersant was used alone at either pH. However, a cleaner product ash of 7.4% was obtained when a dual Al-PAM-dispersant system was used at pH 7.8.

The role of Al-PAM in fine coal flotation was fundamentally explored by examining the surface properties of raw coal and froth after polymer-assisted flotation. Zeta potential measurements were conducted for clean coal (4% ash), raw coal and concentrates from froth flotation experiments in the absence or presence of Al-PAM. Froth products obtained from the use of the dual Al-PAM-dispersant system were also tested for comparison. The zeta potential values of clean coal (4% ash) and raw coal as functions of increasing and decreasing pH were determined and are shown in Figure 14. Clean coal

(4% ash) exhibited a positive charge at pH levels below 3, whereas raw coal has a negative charge over the pH range of 2 - 11. As the pH increased from 3 to 11, the zeta potential of clean coal (4% ash) became more negative. The isoelectric point of clean coal is approximately pH 3.5, which indicates that as pH decreased below 3, the amount of adsorbed hydronium (H⁺) ions increased on the surface of the coal, positively charging the surface of the coal. Conversely, when the pH increased above 3, the amount of hydroxyl (OH⁻) ions increased and adsorbed on the surface of the coal particles, replacing the hydronium ions and rendering the surface negatively charged. This indicates that changing the concentration of the hydronium or hydroxyl ions changes the magnitude and sign of the zeta potential. As the pH increased from 2 to 11, the zeta potential of the clean coal (4% ash) was the most negative, whereas that of the raw coal was the least negative. The less negative surface charge of raw coal (compared with clean coal) under increasing pH is attributed to the adsorption of positively charged metal ions, such as Mg²⁺, Ca²⁺ and Fe²⁺, on the negatively charged surface of raw coal particles [26] .

Results for the zeta potential distribution peaks for clean coal, raw coal, concentrates obtained from froth flotation experiments using Al-PAM, the dual Al-PAM-dispersant system and collector or frothing agent alone are shown in Figure 9. The zeta potential distribution peaks for clean coal, raw coal and froth concentrate obtained from Al-PAM, the dual Al-PAM-dispersant system and collector or frothing agent alone were −3.5, −46.59, −21.5 −10.04 and −32.9, respectively. The shift of the peaks to more positive zeta potential values after flotation indicates a significant removal of ash-forming minerals from the coal surface by Al-PAM. For illustrative purposes, the interaction between ash-forming mineral and coal is shown in Figure 9 by black and white circles representing mineral matter (slime coating) and coal particles, respectively. For raw coal, coal particles are fully covered by the ash-forming minerals (slime coating), as shown schematically in the figure inset. This was confirmed by the measured zeta potential of raw coal. The zeta potential peak shifted left (more negative) and is dependent on the ratio of the components, illustrating the reduced combustible recovery of coal in flotation experiments.