Statement of Novelty

Alumina waste was used for synthesis of adsorbents. Removal of chlorine containing compounds. Positive influence of titania deposition by ALD.

Introduction

The world is changing rapidly due to an increase in the population and a corresponding increase in production of various products. Against the backdrop of progressive pollution of the atmosphere [1] and natural waters [2], along with depletion of natural raw materials [3], ensuring environmental safety and rational use of man-made waste are becoming one of the priority areas of scientific research and activities of specific industries [4, 5].

At the same time, tightening of environmental legislation requires an increasing attention to preventing emission of toxic organic and inorganic pollutants into air and water environment from transportation, metallurgical, oil refining, chemical and other industries. An important role in mitigating a negative effect of emissions on the environment is played by adsorption and catalytic processes. Due to increasing shortage of drinking and process water, efficient adsorption technologies for removal of toxic chlorine-containing substances from wastewater have been developed [6]. Halogenated hydrocarbons, like other carcinogens, can accumulate in the environment acting as a trigger for oncological diseases [7]. Biotechnological approaches, using anaerobic bacteria, have been developed to control the soil and groundwater contamination with chlorinated hydrocarbons [8]. Photodegradation of organic pollutants in contact with mesoporous microspheres Mo-TiO2 and Al2O3-TiO2, produced by spray pyrolysis [9, 10], should be noted among the recent development.

Nanostructured titania, as a promising n-type semiconductor with a wide band gap (3.2 eV) in the ultraviolet light, along with high electric conductivity and resistance in corrosive media, is widely used for various applications. The photocatalytic and catalytic properties of TiO2 depend on many factors: specific surface area, heat treatment conditions [11] and the synergistic effect with other materials. For example, doping TiO2 nanoparticles with FeCl3 reduced the band gap and improved the photocatalytic activity during the photodegradation of rhodamine B to 93.8% [12]. The use of TiO2 with a high specific surface area of 102 m2/g in the synthesis of the Pt/C electrocatalyst made it possible to increase stability of Pt/TiO2 + C fuel cells by preventing oxidation of the carbon carrier [13]. TiO2-graphene nanocomposites are used in dye-sensitized solar cells photoanodes [14], while anatase-based supports for silver nanocatalysts per se or promoted with SeO2 [15, 16], or rutile with addition of CuO/SiO2/Al2O3 [17] allow a decrease in the amount of soot, hydrocarbons, CO and NOx, in modern gasoline engines. High activity of metal nanoparticles, deposited on TiO2 with a developed specific surface area and a porous structure, along with high mechanical strength, made it possible to use them in various reactions. Replacing aluminum oxide, a widely used traditional carrier for hydrotreating catalysts, with titanium dioxide contributed to an increase in metal dispersion and catalytic activity of NiRu bimetallic particles in deoxygenation [18]. TiO2 is considered an alternative material that stabilizes heterogeneous metal catalysts in reduction and oxidation reactions at low temperatures and pressures [19]. One of the hazardous wastes generated in large quantities during the production of propylene is SAA, deactivated with organometallic compounds during purification of the waste cyclohexane solution (the reaction medium for polymerization of ethylene into polyethylene).

One of the hazardous waste products, formed in large quantities in propylene production is spent alumina adsorbent (hereinafter referred to as SAA), deactivated by organometallic compounds in the process of the cyclohexane solution treatment.

Previously, the authors have already considered various aspects of the use of adsorbent waste for the needs of the oil and gas refining industry in the production of environmentally friendly fuels and oils, including some tests at the industrial scale.

  1. (a)

    The adsorbent of halide compounds Uz-AD-1 was obtained when impurities were first removed from SAA granules by successive washing with an excess of 1% NaOH solution, deionized water and heat treatment at about 400 °C. Then the thermochemical activation was carried out in an 8–9% NaOH solution at 40 °C, followed by drying at 150–180 °C. The pilot batch of the adsorbent operated in the process of cleaning exhaust gases with HCl concentration varying in the range of 10 to 50 ppm. The gases formed during regeneration of the bifunctional Pt reforming catalyst CR 401 with C2Cl4 also contained unidentified organochlorine and unsaturated compounds. “HCl slip” of 1 ppm was recorded after 12 months of operation, and the mechanical strength decreased by 15% [20]. The positive results of the trial run served as the basis for loading of Uz-AD-1 material into the industrial adsorber of the Bukhara Oil Refinery.

  2. (b)

    Addition of 35% SAA with a particle size of 0.06–0.1 mm to a mixture of industrial aluminum hydroxide and 12% kaolin during synthesis of the carrier made it possible to form a bimodal porous structure and improve activity of the Co–Ni-Mo catalyst in the hydrodesulfurization reaction [21]. When forming the support for the Ni-Mo catalyst for hydrogenation of polyaromatic hydrocarbons, in addition to 35% SAA, another 10% SA-Ti (15.0) was added to slow down migration of Ni2+ ions into the Al2O3 lattice with the formation of a slightly active NiAl2O4 phase [22].

  3. (c)

    From fresh pseudoboehmite obtained by processing SAA and commercial magnesia, low-temperature magnesium–aluminum spinel with a bimodal porous structure was synthesized being a suitable support for nickel catalysts for the steam reforming of light hydrocarbons to CO and H2 [23].

It was established that during processing of SAA waste, organometallic compounds VC16O5H14, TiC16O4H14, AlC24O6H21 (deactivation products of catalysts in the ethylene polymerization reaction) chemisorbed by the fresh adsorbent from the waste cyclohexane solution were only partially removed during the thermochemical activation of SAA granules with a sodium hydroxide solution. In the form of the corresponding oxides, 0.22–1.1% V, 0.08–0.29% Ti, 4.5–5.5% Na, were a part of the Uz-AD-1 adsorbent, which was intended for purification of waste gases of catalytic reforming. It is widely known that any impurities in the composition of adsorbents affect to a certain degree the strength of the granules, the specific surface area and especially the adsorption efficiency of specific adsorbates. At the same time, the patent literature reports [24] that the certain oxides of vanadium, titanium and sodium as second active metals were specially introduced into the composition of adsorbents, where AgO, AuO and CuO in the metallic state acted as the first active metal. According to [24], sodium vanadate, sodium hydroxide, tetrachloride or titanium dioxide, are preferred for introducing second metals into adsorbents. It is noted, that the use of such adsorbents is economically feasible only in the preparation of gases for the Fischer–Tropsch type processes, where the catalysts are extremely sensitive to halogen compounds at concentrations below 1 ppm, down to 1 ppb.

One of the main aims of this work, focused on maximizing utilization of the aluminium-containing waste for preparation of environment-friendly adsorbents suitable in oil refining, was, therefore, an assessment of the role of vanadium and titanium compounds on the properties of such adsorbents. Subsequently the detailed physico-chemical characterization was performed for the adsorbents which were then tested for the removal of water, tetrachloroethylene and HCl.

Experimental

Materials

The following materials have been used for the synthesis of adsorbents: commercial alumina AN-2 (Novomichurinsk, Russia); waste aluminium oxide adsorbent (SAA) from the Shurtan gas chemical complex; kaolin from the Angren quarry—“Angren Kaolin” LLC. HNO3 was purchased from JSC “Electrokhimprom”, Chirchik; while the waste from production of the catalyst for total oxidation of CO, toluene and other toxic organic compounds emissions, containing defective granules of the alumina- titania carrier SA-Ti(15.0) from JSC “Khimeks”. Hereinafter, the numbers in parentheses after the symbol of the chemical element in the samples code denote the actual content of this element in wt%.

The following materials have been used for modification of adsorbents: NaOH—AO Navoiyazot, TiCl4 99.9%—SKT Taurus, Na3VO4.12H2O, chemically pure–Chem Craft.

Composition of the SAA after drying at 100 °C according to chemical analysis, depending on the degree of inactivation, varied over a wide range (wt%): Al(39.7–45.4)2O3, V(0.28–1.4), Ti (0.18–0.72), Na (0.74–1.25), Ca (0.35–0.92), Cl (0.8–2.5), with combustible organic matter constituting the rest. A loss on drying at the temperature up to 100 °C was 3–9% by weight and loss at ignition, due to organic sediments burnout, was 12–19% by weight. Specific surface area of the SAA varied from 80 to 280 m2/g, while the total volume of pores for moisture absorption was changing from 0.14 to 0.46 cm3/g. Kaolin contains oxides (in wt%) of the following elements: Al (11.9–19.6), Si (21.9–29.9), Ca (0–0.64), Ti (0.3–0.8), Mg (0–0.06), K (0–0.83), Fe (0.2–1.4) and H2O between 6 and 15%. The phase composition of the SA-Ti(15) in wt% is γ-Al2O3 50–52, TiO2 22.4–24.7, AlOOH 21.5–23.2 with the rest being physically adsorbed water.

Synthesis of Adsorbents

Pseudoboehmite, produced by processing of the industrial waste, has been used in the synthesis of the first series of samples, which included the following steps: (1) dissolution of crushed SAA waste was carried out by heating for 7–8 h at 100–105 °C in a solution of nitric acid with a density of 1.27 g/cm3. To reduce the release of toxic nitrogen oxides and acid consumption during dissolution, a reflux condenser was used; (2) precipitation of AlOOH from the resulting solution of the aluminium nitrate, using ammonia solution at pH = 7–8; (3) maturation of the precipitate at 20–27 °C; (4) precipitation filtration and washing; (5) drying of the pseudoboehmite AlOOH powder. The SAA dissolution parameters (temperature, HNO3 concentration and time) were selected based on the results of preliminary experiments, taking into account the acid resistance of γ-Al2O3 and α-Al2O3 phases in its composition [21]. The SA material has been formed in cylindrical granules by extrusion of the AlOOH powder peptised with 3% nitric acid solution. The total content of residual impurities in the SA composition (V, Ti, Na, Ca and Cl) was less than 0.02% by weight. Adsorbents of the first series SA-Ti(5.0), SA-Ti(2.0) and SA-Ti(0.15) adsorbents with the medium and the lowest content of titanium were obtained in a similar way, but with the addition of crushed SA-Ti(15.0) to the paste (particle size—from 0.04 to 0.1 mm). The ground SAA (30% by weight with the particle size between 0.04 and 0.1 mm) and kaolin (10% by weight) were added to AlOOH powder during preparation of the material coined AKB, when peptization of the mixture of dry ingredients was carried out by stepwise addition of 3 wt% nitric acid and 2.5 wt% boric acid solutions. Ground SA-Ti(15.0) (particle size from 0.04 to 0.1 mm) was added during AКB-Ti(4.5) synthesis instead of SAA. The final heat treatment of the adsorbents of the first series was done at 550 °C for 3 h. The above particle size of SAA and SA-Ti(15.0), as well as their quantity in the molding paste, were selected based on a preliminary assessment of the texture and strength of the resulting granules. The use of particles smaller than 0.04 mm reduced the number of transport pores, and when particles larger than 0.1 mm were added, the granules lost their mechanical strength.

Modifications of Adsorbents

Adsorbents of the second series: SA-Ti(0.2), AКB-Ti(0.18), AN-Ti(0.2) with the lowest amounts of titanium were obtained by ALD of SA, AКB and AN-2 granules with the vapours of TiCl4, respectively. ALD is a method based on the sequential use of a gas-phase chemical process, in particular alternating reactions of TiCl4 with surface hydroxyl groups of Al2O3 and hydrolysis, leading to deposition of a thin film. The methodology of this work is similar to that previously described for metal oxide powders [25]. The extent of ALD was controlled by IR spectroscopy. First, the granules placed in a quartz reactor, were purged with dry nitrogen at 480–490 °C. As a result of the molecularly adsorbed water removal in the near infrared region for samples SA and AN-2, only bands of OH groups, associated with aluminium atoms at 4500 and 7300 cm–1 have appeared. After this, the temperature in the reactor was decreased to 200 °C, the surface of the granules was saturated with the vapors of titanium tetrachloride in dry nitrogen for 1.2 h, with effective exhaust ventilation in the experimental area.

After a purge of TiCl4 unreacted with hydroxyl groups on the surface with dry nitrogen the intensity of the absorption bands at 4500 and 7300 cm– 1 decreased sharply due to the interactions of the surface hydroxyl groups with TiCl4. Thereafter, water vapours were fed through the sample to hydrolyze the Ti–Cl bonds followed by exposure again to dry nitrogen at 200 °C to remove the excess of water vapours. Formation of Ti–OH bonds was accompanied by an increase in the intensity of the bands at 4500 and 7300 cm–1.

Completeness of HCl removal during the experiment was monitored by the outlet gas analysis, after a quantitative capture of HCl vapours with distilled water and nephelometrical determination of the chlorine ions concentration. All the above mentioned operations were repeated to deposit the subsequent layers of titania. Formation of the titania layer on the surface of the granules was controlled by appearance in the Raman spectrum of an intense band at ca. 142 cm–1, characteristic for the anatase phase [26] and not present in the spectra of the supports. By comparing changes in the concentration of Ti on the cut of the granules with scanning electron microscopy, it was found that the main part of titania was located on the surface of the granules having 3.5 mm in diameter, with the penetration depth of titania being close to 0.63 mm.

Spherical AN-2 granules with the specific surface area of 290 m2/g and containing no modifiers were subjected to alkaline modification giving the second series of samples. The materials in the first and the second series were exposed to an excess of 20 wt% sodium hydroxide solution. By setting the AN-2 granules exposure in the impregnation solution to 15, 40 and 120 min, the adsorbents AN-Na(1.8), AN-Na(4.3) and AN-Na(6.1)* were obtained with their surface area being 260, 223 and 131 m2/g, respectively. Similarly, SA-Ti(14.3)Na(4.9), SA-Ti(1.9)Na(5.2), SA-Ti(0.14)Na(5.5), SA-Ti(14.2)Na(5.7)*, AN-Ti(0.19)Na(5,2)*, SA-Ti(0.19)Na(5.1)*, SA-Ti(0.14)Na(6.2)*, SA-Na(6,6)* and others materials were prepared by alkali treatment of the respective supports. For synthesis of AN-Na(4.1)V(0.3), 2.5 g of Na3VO4.12 H2O has been added to 100 g of 20 wt% NaOH. AN-Na(0.3)V(0.4) adsorbent was prepared by impregnation of AN-2 with an excess of 2.5 wt% of the aqueous solution of Na3VO4.12.

Upon impregnation with alkaline solutions, the granules were filtered, washed with water, dried in air at ambient temperature and then at 200 °C for 5 h. The granules, marked with an asterisk (*), have not been washed prior to heat treatment.

Analysis and Measurements

The elemental analysis of the samples has been performed with scanning electron microscope (SEM) EVOMA 10 combined with EDX. The total content of titanium, vanadium, nickel and molybdenum in the obtained samples was also determined by photocolorimetry method, and of sodium by the flame photometry method. The textural properties including the surface are and the pore size have been determined by mercury porosimetry (Carlo Erba) operating up to 200 MPa. The relative content of chlorinated oligomers in the deposits on the granules has been evaluated by appearance of the band in the range of 1270–1300 cm−1 in the IR-spectrum using the partial internal reflection mode of Nicolet S50 spectrometer with the Fourier transform.

X-ray diffraction (XRD) analysis was carried out using an Empyrean diffractometer, tube voltage 30 kV, current 20 mA using CuKά filtered radiation. Diffraction patterns were interpreted using the 2019 International Center for Diffraction Data database.

The adsorption capacity of the prepared materials was measured by gravimetry exposing the granules (0.6–0.7 g), pre-dried at 200 °C, to water vapours, as well as tetrachloroethylene or hydrochloric acid (37.75 wt%). For evaluation of the adsorbents ability to form “green oil”, that is, chlorinated oligomers with an increased molecular weight, the solid sorbents were exposed to a model mixture analysing the corresponding reaction products.

The set-up for these measurements is illustrated in Fig. 1.

Fig. 1
figure 1

Set-up for assessing the ability of adsorbents to form “green oils”. 1—adsorber; 2—granules of the studied adsorbent; 3—container for collecting condensate, 4—circulation pump; 5—a trap for saturating the nitrogen flow with vapors of the model mixture; 6—a model mixture: tetrachlorethylene, hexene-1, hydrochloric acid; 7—container for water with ice; 8—rotameter; 9—thermostat; 10—stirrer; 11—thermometer. Relative weight of condensate from adsorbents: AN-Na(0.3)V(0.4) after 50 h—A, after 315 h—B; SA-Ti(0.14)Na(6.2) after 315 h—C

Full size image

For the purpose of measuring formation of “green oil” the granules of a studied adsorbent (2) were placed into an adsorber heated to 30–40 °C, typical for the operation of an industrial adsorber. A container (3) for collecting the liquid reaction products and a circulation pump (4) were connected to a glass tube in the lower part of the adsorber (1). A model mixture (6), which contained tetrachlorethylene, 1-hexene, and 37% hydrochloric acid, was poured into the drum trap (5). Then, the drum trap (5) with the model mixture was placed in a thermostat (9), where the temperature was maintained at 45–50 °C (necessary and sufficient to obtain a gas mixture of the planned composition). A rotameter (8) was used to control the flow of the gas mixture. After filling the set-up with nitrogen to displace atmospheric air, the circulation pump (4) was turned on. Nitrogen, after being saturated with tetrachlorethylene, hexene-1, HCl and water vapors in a bubble trap (5), passed through a layer of granules of the studied adsorbent (2). After 50 h of circulation of saturated vapors of the model mixture through the adsorbent bed in a container (3) cooled with water with ice cubes (7), a colorless condensate accumulated (70–80 ml). The greenish color of the granules in the lower part of the adsorber in the case of adsorbents SA, AN-2 and AKB appeared after 8 h. After 8–10 days, drops of green liquid flowed down the walls of the container (3) being the condensate of vapors of the reaction products. The experiment was stopped 13 days after accumulating enough of the green condensate for analysis. The initial model mixture from the drum trap (5) and the condensate from the container (3) were analyzed.

One sample of the liquid condensate was analyzed on a Nicolet S50 IR spectrometer, while 1 µl of another sample was injected into Khrom 5 gas chromatograph with a thermal conductivity detector. A 4 m long stainless-steel column was filled with Polysorb. The temperature of the column and injector of 280 °C, as well as other experimental conditions, were chosen based on the preliminary calibrations with the initial model mixtures, also taking into account the expected molecular weight and boiling points of the oligomerization products. Determination of the unreacted HCl was carried out by nephelometry after dissolving a sample of the condensate in distilled water.

Nitrogen at the flow rate of 15–18 l/h and at 30–40 °C containing 9 g/m3 HCl, 3 g/m3 tetrachloroethylene and 3 g/m3 hexene-1 was fed through 100 cm3 of the adsorbent for 315 h in the circulation mode.

Results and Discussion

Physico-Chemical Properties of Materials

Textural Properties

The textural characteristics of SA, produced from pseudo-boehmite, confirm that it has a more developed specific surface area and a larger fraction of smaller pores (Table 1) than the AKB carrier.

Table 1 Properties of the adsorbents
Full size table

Modification of carriers with titanium dioxide by adding crushed SA-Ti(15.0) to the suspension for extrusion led to their uniform distribution throughout the volume of granules, but was accompanied by a decrease in the specific surface area of calcined adsorbents. The specific surface area of SA-Ti(5.0) was only 60% of SA, SA-Ti(0.15) was ca. 75% (243 m2/g) and the determination of this parameter for mechanical mixtures of SA with SA-Ti(15.0) at calculated titanium contents of 5.0% and 0.15% led to similar results. Utilization of ALD had a less prominent effect on the textural characteristics, reducing the specific surface area of SA-Ti(0.2) by 83% due to a slight increase in the size of micro- and mesopores while decreasing their volume. The specific surface area of AKB-Ti(4.5) was 64% of AKB, in contrast to AKB-Ti(0.18) obtained by the ALD method, which as 82%.

XRD

The diffraction pattern of SA-Ti(15.0) is presented in Fig. 2, where the narrow most intense peaks were attributed to the polymorphic crystalline modification of TiO2—metastable anatase (d = 3.52; 2.431; 2.378; 38.46; 1.892; 1.699; 1.667; 1.481; 1.338 and 1.66 Ǻ) and peaks of moderate intensity to its stable polymorph—rutile (d = 3.25; 2.486; 2.297; 2.188; 2.054; 1.6874; 1.6327; 1.4797; 1.4528; 1.3598 and 1.3465 Ǻ). The main phase by mass, γ-Al2O3, was represented by two broad peaks with d = 1.98 and 1.39 Ǻ, of relatively weak intensity. The intensity of narrow peaks decreased in the order: SA-Ti(15.0) > SA-Ti(5.0) > AKB-Ti(4.5) > SA-Ti(2.0). The intensity ratio of the main anatase/rutile peaks varied from 3.2/1 in the SA-Ti(15.0) diffraction pattern to 2.4/1 for the adsorbents of the first series. This indicated partial transformation of anatase into a stable rutile phase during heat treatment of extrudates and explained the substantial decrease in the specific surface area of the modified adsorbents of the first series. Calcination of SA-Ti(15.0) also resulted in a decrease in the main anatase peak with d = 3.52 Å and an increase in the peak with d = 3.25 Å from a denser rutile structure with a lower specific surface area. After 3 h of heat treatment at 550 °C, the anatase/rutile peak ratio was 2.7, after 10 h it decreased to 2.1, and the specific surface area was 109 m2/g. Specific TiO2 peaks did not appear in the diffractograms of SA-Ti(0.15), SA-Ti(0.2), AKB-Ti(0.18) and AN-Ti(0.2) being are identical to the diffractograms of the original supports SA, AKB and AN-2 (Fig. 3) exhibiting broad intense peaks of γ-Al2O3 (d = 1.98 and 1.39 Ǻ) and a halo in the region of interplane distances, characteristic for AlOOH (d = 6.1; 3.16; 2.33; 1.98; 1.85; 1.75; 1.45; 1.39; 1.31 Ǻ). In the diffractograms of AKB and AKB-Ti(0.18), in addition to the peaks of γ-Al2O3, some narrow peaks of quartz were observed (SiO2 with d = 3.34 and 1.82 Ǻ).

Fig. 2
figure 2

Diffractograms of the initial adsorbents: SA-Ti(15.0)—1, SA-Ti(14.3)Na(4.9)—2 and upon exposure to the saturated vapours of the model mixture HCl + C2Cl4 + C6H12: SA-Ti(15.0)—3, SA-Ti(14.3)Na(4.9)—4. A anatase, P rutile, γ Al2O3, Filled rectangle bayerite, Filled circle AlOOH, Filled rhombus NaAlCO3(OH2)2, Filled ellipse Na2CO3, T NaHCO3, Up arrow NaTi2Al5O12, Filled star AlCl3, Filled four pointed star AlCl3.6H2O, Filled square NaCl, filled six pointed star-Ti(AlCl4)2

Full size image
Fig. 3
figure 3

Diffractograms of the adsorbents AN-2—1; AN-Na(4.3)—2; AN-Na(4.1)V(0.3)—3; AN-Na(4.3), ignited at 1100 °C—4. γ–γ Al2O3, Filled rectangle bayerite, Filled circle AlOOH, Yellow filled rectangle gibbsite, α–α Al2O3, Orange filled rectangle NaAlCO3(OH2)2, Filled rhombus NaAl11O17, Filled down pointed triangle 2NaAlO2.3H2O, Up arrow NaVO3, V Na3VO4

Full size image

After ALD of titania onto the carriers at 200 °C, the specific surface area of the latter decreased by only 17–18%, due to clogging of the smallest pores by anatase nanoparticles exclusively on the outer surface of the granules. According to SEM data, when scanning SA-Ti(0.2), AKB-Ti(0.18) and AN-Ti(0.2) granules from the periphery to the center, titania was located near the outer layer. On average, for the six studied spots on the surface (100 × 100 μm), the amount of Ti was in the range of 50–58%, while oxygen ranged from 41 to 52%. Only in two surface spots the presence of aluminum was confirmed for AN-Ti(0.2), while silicon was also detected on the surface of AKB-Ti(0.18).

According to SEM chlorine was absent in the samples, produced by ALD.

High alkaline medium (pH ≥ 13) during treatment caused a gradual dissolution of AlOOH in the supports, with accumulation in the liquid phase of aluminate ions [Al(OH)4]−, [AlO(OH)2], AlO2− and ion-associates Na+Al(OH)4− [27]. A high concentration of aluminate ions in the pores facilitated precipitation and growth of aluminium trihydroxide crystals, along with formation of the layer of amorphous sodium aluminates, chemisorbed on the pore walls. Phases of Al(OH)3 in the form of bayerite (d = 4.71; 4.37; 3.21; 2.224 Ǻ) and/or gibbsite (d = 4.83; 4.35; 4.28; 3.31; 3.16; 2.43; 2.37 Ǻ), were manifested most clearly in the diffractograms of AN-Na(4.3) and AN-Na(4.1)V(0.3), when sodium compounds not attached to the supports were removed by washing with water (Fig. 3). Using SA-Ti(14.2)Na(5.7)* and AN-Na(6.1)* as examples it was demonstrated that, in 120 min of impregnation, the alkaline solution penetrated through the pore network inside the granules with formation of amorphous sodium aluminates and NaOH and NaOH.H2O precipitates, while sodium carbonates and NaAlCO3(OH)2 were located predominantly on the granule surface. The shape of diffractograms with a wide halo in the area of low angles makes it possible to suggest formation of poorly crystalline phases of sodium aluminates and sodium carboaluminate, as a result of reaction with CO2 with the corresponding peaks d = 11.3; 5.63; 4.45; 2.79; 2.68; 2.51; 2.42; 2.41, 2.245; 2.136; 2.307; 1.615; 1.59; 1.484; 1.4125; 1.399 Ǻ—NaAl11O17, d = 5.30; 4.32; 2.82; 2.55; 2.30; 1.72 Ǻ—2NaAlO23H2O and d = 5.67; 4.94; 4.51; 3.38; 3.094; 2.784; 2.60; 2.50; 2.22; 2.15; 1.988; 1.728; 1.689; 1.567; 1.541; 1.47; 1.39; 1.449; 1.288 and 1.256 Ǻ—NaAlCO3(OH)2.

In diffractograms of AN-Na(6.1)*, SA-Ti(14.2)Na(5.7)*, SA-Ti(1.9)Na(5.8)*, SA-Ti(0.14)Na(6.2)* and AN-Na(5.9)V(0.5)*, exposed to heat treatment without any preliminary washing with water, narrow peaks of the crystalline phases, with d = 5.91; 4.84; 3.48; 3.082; 3.06; 2.956; 2.936; 2.68; 2.60 Ǻ—NaHCO3 and d = 3.22; 2.96; 2.70; 2.60; 2.54; 2.36; 2.25; 2.18 Ǻ corresponding to sodium carbonate were observed.

SEM

Chemical interactions of the alkaline modifier with the support resulted in an increase of intensity of the halo from sodium aluminates in diffractograms. This was in line with the chemical formulae of the presumably formed structures, which were calculated based on the of ratio of the chemical elements (sodium, aluminium, carbon and oxygen), detected by SEM–EDX for AN-Na(1.8), AN-Na(4.3) and AN-Na(6.1)*, as well as by presence of the crystalline phases NaAl11O17, 2NaAlO2.3H2O and NaAlCO3(OH)2 on the diffractogram AN-Na(4.3) upon calcination at 1100 °C (Fig. 3). Appearance of the peaks at d = 7.84; 7.76; 5.90; 4.49; 3.88; 2.95 Ǻ from NaTi2Al5O12, partially overlapping with more intense ones from NaAlCO3(OH)2, Na2CO3 and NaHCO3 (Fig. 2, diffractogram 2) has been attributed to chemical interactions of NaOH with inclusions of the titania particles, surrounded by alumina in the composition of SA-Ti(15.0).

When comparing the images of the surfaces (Fig. 4), changes in the granules morphology for SA-Ti(15.0) can be seen depending on the conditions of alkaline modification. Impregnation with the alkaline solution for 30–40 min did not guarantee a uniform distribution of the modifier across the granules because of the viscosity gradients slowing the diffusion into the material. As the impregnation time increased, the degree of the granules surface etching with the alkaline solution and the number of carbonate structures also increased. In the case of SA-Ti(15.0) long-time impregnation, the elemental composition of various areas on the surface of the granules was changing insignificantly (C—12.1–13.15; O—54.43–55.24; Na—25.99–27.56; Al—3.93–4.6; Ti—0.98–1.65% by weight), even if surface morphology varied resembling an intensively weathered rock (Fig. 4, centre). Smoothed dense fragments were observed on the surface of the granules of SA-Ti(14.2) Na(5.7)*, along with the plate-like structures with rugged edges, layered elongated formations and bright light areas of a rounded shape with the size of 5–10 µm. Based on the ratio of the elements on the surface of the granules SA-Ti(14.2)Na(5.7)*, the most probable chemical composition was calculated as: 82.87–89.0% NaHCO3; 4.5–9.3% NaTi2Al5O12; 0–12.9% NaAlO2; 0–3.3% NaTiO2; 0–3.95% Na2O; 0–1% NaOH; 0–11% NaAlCO3(OH)2; 0–2.27% TiO2; 0–3% AlOOH; 0–7% Al2O3. On the contrary in the granules volume Al2O3 (9.8–67.6%); AlOOH (0–37.6%); Al(OH)3 (0–40%); TiO2 (13.5–16.8%) were predominant. Sodium compounds included: Na2CO3 (4.1–8.1%); NaHCO3 (0–7.6%); sodium aluminates (NaAl5O8, NaAlO2, Na7Al3O8) not exceeding 6%, and NaTi2Al5O12 (3.2–9.7%). Upon washing and calcination at 200 °C, SA–Ti(14.3)Na(4.9) instead of clearly defined edges of an irregular shape in the size of 10–20 µm (Fig. 5—on the left), characteristic for the surface of SA-Ti(15.0), exhibited more blurred edges, while dark areas in the size of less than 20 µm became coated with amorphous sodium aluminates chemically bound to the support (Fig. 4 right). According to SEM, AN-Na(4.1)V(0.4) and AN-Na(5.9)V(0.5)* have a uniform distribution of NaAl11O17, NaAlCO3(OH)2 and sodium vanadates in the interior of the granules with sodium carbonates predominance on the surface (Fig. 5 left).

Fig. 4
figure 4

The SEM images and spectra of the areas, isolated on the surface of adsorbent granules: SA-Ti(15.0)—on the left, SA-Ti(14.2)Na(5.7)* modified with NaOH during 120 min without washing before heat-treatment—in the centre, the same sample, but washed before heat-treatment SA-Ti(14.3)Na(4.9)—on the left

Full size image
Fig. 5
figure 5

SEM images and spectra of areas, isolated on the surface of the adsorbent granules: left AN-Na(5.9)V(0.5)*, centre SA-Ti(0.19)Na(5.1)*, SA-Ti(0.07)Na(0.5)—on the right

Full size image

By comparing X-ray Fluorescence (XRF), which is an analytical method based on the interactions of X-rays with a material to determine its elemental composition, and SEM for the samples, heat treated without water washing, or alternatively washed, it is found that from 10 to 30% of the sodium, mainly in the form of carbonates, has been removed from the granules of the modified materials in the first series when washing with water was applied.

In the case of washing of AN-Ti(0.19)Na(5.2)* material having AN-Ti(0.2) as the support and prepared by ALD with TiO2, after heat treatment a minimum content of the alkaline modifier has been recorded in AN-Ti(0.07)Na(0.5). In the evaporated washing water besides NaHCO3 and Na2CO3, significant amounts of aluminium were measured along with titanium at the detection limit of SEM–EDX. According to SEM of the several spots on the surface of the second series of materials SA-Ti(0.19)Na(5.1)* and AN-Ti(0.19)Na(5.2)*, the average content of elements is: Na—36.3; C—11.8; O—48.7; Ti—1.8 in wt% (Fig. 5 centre). The calculated surface contributions (wt%) of different chemical compounds are, 66.1—NaHCO3; 7.36—NaOHxH2O; 4.17—NaAlO2; 3.74—Na2CO3; 3.63—NaTi2Al5O12; 1.77—Na3AlO4; 1.33—Na7Al3O8; 1.22—NaAlCO3(OH)2; 0.72—NaTiO2, 0.6– AlOOH, with adsorbed water being the rest. The elemental composition in the cut of the inner part of the granules is (in wt%), %: Na—0.3, C—0.16; Al—48.8; O—50.04, while the mean composition is 57.12—Al2O3, 41.03—AlOH and 1.85—NaAlCO3(OH)2 in wt%. Thus, a film formed after ALD of titania prevents the alkaline solution penetration into the volume of the supports AN-Ti(0.2) and SA-Ti(0.2), since the alkali treatment resulted in almost complete deposition of sodium carbonate, bicarbonate and hydroxide on the outer surface of the granules.

The crystalline phases of Na2CO3 and NaHCO3 were clearly manifested in the diffractograms SA-Ti(0.19)Na(5.1)* and AN-Ti(0.19)Na(5.2*) with the background representing AlOOH and γ-Al2O3. Weak peaks did not, however, allow to unequivocally identify NaTi2Al5O12 and NaAlCO3(OH)2. It follows from Fig. 5, right, that after removal of crystalline, numerous craters with uneven edges remained on the surface of the granules with the ratio of elements suggesting a substantial destruction of the titania film, exposure of alumina and formation of the surface sodium aluminate. In general, the titanium content in SA-Ti(0.07)Na(0.5) decreased three fold, and the sodium concentration was did not exceed 0.5 wt%.

Adsorption Properties of the Materials

The adsorption kinetic data are demonstrated in Fig. 6.

Fig. 6
figure 6

Adsorption kinetics for water (left), tetrachloroethylene (centre) and hydrochloric acid (right) vapours at 25 °C. 1-AN-2; 2- SA-Ti(15.0); 3-AN-Ti(0.2), 4-AN-Na(4.3); 5-SA-Ti(14.3)Na(4.9); 6-SA-Ti(0.14)Na(5.5); 7-AKB; 8-AN-Na(5.9)V(0.5)*, 9-AN-Na(6.1)*; 10-SA-Ti(0.19) Na(5.1)*

Full size image

The first order kinetics with a saturation at prolonged reaction time was characteristic in the case of water adsorption only for the adsorbents non modified with sodium hydroxide. The amount of the absorbed water at complete saturation [g H2O/100 g of adsorbent], correlates well with the specific surface area of adsorbents as illustrated in Fig. 7.

Fig. 7
figure 7

Correlation between the amount of adsorbed water and the specific surface area. №1—SA-Ti(15,0); №2—SA-Ti(5,0); №3—AN-Ti(0,2); №4—SA -Ti(0,2); №5—SA-Ti(2,0); №6—AN-2; №7—SA-Ti(0,15); №8—SA. Filled rhombus 2 days and filled square 13 days of exposure to water vapours

Full size image

No direct relationship between the anatase/rutile phase ratio and water adsorption could be found. The SA-Ti (15.0) sample with the ratio of 2.1 adsorbed in 13 days: 4.8 g H2O/100 g; 7.8 g C2Cl4/100 g and chemisorbed 16.4 g HCl/100 g, slightly less than noncalcined SA-Ti(15.0). In a similar way, calcination affected the specific surface area and water adsorption of AN-Ti(0.2) and SA-Ti(0.2) adsorbents, which probably contained an anatase modification of TiO2. At the same time, these parameters did not change for AN-2 and SA.

Water adsorption curve knee points were observed for the adsorbents, unwashed with water after the alkaline modification: AN-Na(5.9)V(0.5)*, AN-Na(6.1)*, AN-Ti(0.19)Na(5.1)* and SA-Ti(14,2)Na(5.7)*, where the access to adsorption centra in the pores of the surface layer of the granules was limited by the non-hygroscopic sodium carbonate crystals. In the case of a prolonged contact with the saturated water vapour, a slow increase in the water uptake can be related to hydration of hygroscopic NaOH crystals, aggregated in the inner pores of the granules. Based on the nature of the adsorption curves for SA-Ti(0.2) and AN-Ti(0.2), obtained by ALD of titania, it can be assumed that the surface film of titania has predominantly narrow pores, accessible for small water molecules contrary to larger molecules of tetrachloroethylene (Fig. 6, left and centre). Adsorption of tetrachlorethylene retained on the surface of AN-2 and SA can be attributed to the weak Van der Waals forces, reaching 60 and 72 g C2Cl4/100 g, as complete desorption of physically adsorbed tetrachlorethylene was observed after exposure to a flow of dry nitrogen. Chemisorption of hydrochloric acid vapours, very slow initially, is accompanied by a gradual destruction of the titania layer, followed by acceleration of the whole process because of the chemical interactions of HCl with AlOOH inside AN-Ti granules (0.2) (Fig. 6, curve 3—on the right). The same behavior was observed during chemisorption of hydrochloric acid vapors or dry HCl by alumina adsorbents SA (112 g HCl + H2O/100 g) and AN-2 (87 g HCl + H2O/100 g), where formation of AlCl3.6H2O and AlCl3 phases after 30 days of adsorption experiments was determined by XRD.

During the initial period of time, rapid chemisorption of hydrochloric acid or dry HCl on the alkali-modified adsorbents took place at the basic centra with pKa ≥  + 8.3 (NaOH, NaAl11O17, NaAl5 O8, NaAlCO3(OH)2, Na2CO3 and NaHCO3), with sodium chloride formation, which is clearly manifested in the diffractograms. Simultaneously, the strong basic centra were transformed into weak acid centra with pKa ≈ + 6.8, and then into strong acid centra—aluminium chloride. The flat parts of the adsorption curves are due to less intense interactions of HCl with AlOOH. With accumulation of hygroscopic AlCl3, an increase in the uptake of hydrochloric acid vapours was connected to mainly adsorption of water present in these vapours.

One of the environmental concerns in the production of high-quality gasoline is purification by adsorption of the process gases from organic and inorganic chlorine compounds, including chlorinated oligomers, i.e. so-called green oil. Otherwise, the service life of the adsorbent is diminished due to clogging.

IR Analysis

An ability of prepared adsorbents to catalyze oligomerization reactions was tested using the model mixture, simulating industrial gases of the straight-run gasoline reforming unit, with the concentration being 55-fold higher than in actual exhaust gases. The IR spectra of the components of the model mixture coincided with the spectra of the upper (hexene-1) and lower (tetrachlorethylene) layers of the condensate collected in a cooled trap downstream the reactor for 50 h of the experiment and passing 150 g of adsorbate through the bed. The bands at 3088, 2958, 2860 cm–1 were attributed to the stretching vibrations of C–H bonds in C=CH, –CH3 and –CH2 of hexene-1. The overtone appeared at 1847 cm–1, a band of moderate intensity at 1640 cm–1 is characteristic of C=C stretching vibrations in non-conjugated olefins. The water molecules correspond to bands in the region of 3300 and 1640 cm–1. A series of bands at 1460, 1360, 975, 910 and 746 cm–1 are due to bending vibrations of methyl and methylene groups in hexane-1. Tetrachlorethylene was identified by the C–Cl band at 773–780 cm–1, characteristic of polysubstituted chlorinated hydrocarbons, in addition to the more intense bands at 905 and 492 cm–1 observed in the IR spectrum of commercial C2Cl4. The IR spectra of the initial model mixture and the condensate of its vapors after passing through a bed of granules for each adsorbents did not confirm presence of interactions between HCl, H2O, hexene-1 and tetrachlorethylene in the initial period of the experiment. In the chromatograms of the upper and lower layers of the condensate, substances practically insoluble in hydrochloric acid, only a slight change in the ratio of intensity of peaks from hexane-1 and tetrachlorethylene was observed, while nephelometrically recorded a decrease in the concentration of HCl in the middle layer of hydrochloric acid (Fig. 1). The weight of substances relative to distilled water was not different from the components of the model mixture. From XRD, SEM and IR spectroscopy, it was evident that a larger amount of aluminum chloride was formed in the composition of adsorbents AN-2, SA-Ti(0.15), SA-Ti(0.2), AN-Ti(0.2), AN-Ti(0.07)Na(0.5) and AN-Na(0.3)V(0.4) after 5–6 days of contact with the model mixture vapors. Manifestation of the chemical interactions of HCl with adsorbents comes from identification of the bands at 2415, 1638 and 1150 cm−1, observed in the IR spectra of commercial AlCl3.6H2O (Fig. 8, spectrum 8), as well as the bands at 480 and 620 cm−1 originating from Al–O vibrations.

Fig. 8
figure 8

IR spectra of the liquid-phase condensed products downstream the adsorbent bed (100 cm3). Nitrogen flow of 15–18 l/h, contained 9 g/m3 HCl, 3 g/m3 tetrachloroethylene and 3 g/m3 hexene-130–40 °C, duration 315 h, flow circulation mode: 1—AN-Na(0.3)V(0.4); 2—SA-Ti(2.0); 3—AN-Ti(0.19)Na(5.1)*. Adsorbents after testing: 4—SA-Ti(15.0); 5—SA-Ti(14.3)Na(4.9); 6—AN-2; 7—SA-Ti(0.14) Na(6.2)*; 8—AN-2 after passing HCl

Full size image

In the spectra of SA-Ti(5.0) and SA-Ti(15.0), the bands were also recorded at 490 and 506 cm–1, corresponding to the Ti–Cl bond (Fig. 8, spectrum 4). Aluminum chlorides, and to a lesser extent titanium, belong to the category of acid catalysts. The strong acidic sites (pKa < 2) of aluminum chloride catalyze oligomerization and hydrochlorination reactions. As a result, products with a higher molecular weight were formed, which condensed and formed a lower layer of the green liquid along with unreacted tetrachlorethylene. In the spectra of the green liquid, strong absorption bands were observed at 466, 517 and 555 nm and weaker ones at 634, 757, 762, 826 and 883 nm (spectra not shown).

In the IR spectrum (Fig. 8, spectrum 1) of the lower layer of the condensate (Fig. 1B), against the background of a new broad absorption band, narrow bands were observed at 1270 and 1160 cm−1 from vibrations of the CH2 group associated with the chlorine atom, which, in combination with the characteristic band—CH2Cl at 747 cm−1 and the band at 610 cm−1 from the C–Cl bond in aliphatic hydrocarbons such as chlorohexane indicated formation of chlorinated oligomers of hexene-1. An increase in the intensity of the band at 1468 cm−1 and a set of bands at 1070, 990, 890 cm−1 could be due to the presence of benzene monochloride derivatives. This did not contradict with the results of GC analysis of the lower layer, which exceeds the density of hexene-1 and chlorododecane. Formation of 2-chlorohexane and the oligomerization product of hexane-1, dodecane, was proven, in addition to a tetrachlorethylene peak of moderate intensity. The chromatogram also showed a series of poorly resolved peaks from unidentified substances with a boiling point above 220 °C.

The products of oligomerization and hydrochlorination reactions also appeared in the IR spectra of the adsorbents after completion of the experiment, where a set of bands 1455–1475, 1300, 1275 and 1070 cm−1 were present (Fig. 8, spectrum 6). Thus, to identify the compounds that make up the “green oil”, the most important observation is the decrease in the intensity of the C=C bond band (1640 cm−1), appearance of a band in the region of 1096–1039 cm−1 corresponding to C–Cl bond in benzene chloride derivatives, as well as bands from bending vibrations of methyl (1375 cm−1) and methylene (1468, 1350–1150, 747–720 cm−1) groups in saturated hydrocarbons.

Formation of aluminum chlorides on adsorbents containing more than 4% Na under the influence of HCl vapor or a model mixture was recorded only after 10–12 days. Therefore, the degree of oligomerization and hydrochlorination of hexene-1 was significantly less (Fig. 8, spectra 3 and 7). In the IR spectra of adsorbents AN-Na(6.5)*, SA-Ti(0.19)Na(5.1)*, AN-Ti(0.19)Na(5.2)*, AN-Ti (0.07)Na(0.5) and SA—Ti(0.14)Na(6.2)* (Fig. 8, spectrum 7) as a result of adsorption of model mixture vapors, the bands at 1396 and 696 cm–1 from vibrations disappeared CO32− ions in sodium bicarbonate. Formation of crystalline NaCl was identified by XRD. The AN-Na(6.5)* adsorbent most efficiently purified a gas stream with a concentration of unsaturated hydrocarbons and chlorine-containing compounds many times higher than their amount in the actual exhaust gases from naphtha reforming.

The presence of large amounts of TiO2 in SA-Ti(2.0), SA-Ti(5.0) and SA-Ti(15.0) (Fig. 8, spectra 2, 4 and 5) also suppressed oligomerization of unsaturated hydrocarbons with formation of green oil in the condensate, but to a lesser extent compared to the alkaline modification. Absorption in the region of 1270–1300 cm–1 of chlorinated oligomers decreased monotonically with an increasing sodium content in the adsorbents (Fig. 8, curves 5–7). A similar picture was noticeable for samples with the addition of crystalline titanium dioxide, but with an obvious deterioration in the adsorption capacity for water, hydrochloric acid and tetrachlorethylene. Modification of SA-Ti(0.2), AN-Ti(0.2) and SA-Ti(0.07)Na(0.5) by ALD showed the same behavior. An overview of all materials, illustrating an almost linear dependence of the absorption intensity at 1270 cm−1, corresponding to chlorinated isomers, on the sodium content in various materials, is presented in Fig. 9.

Fig. 9
figure 9

Dependence of the absorption intensity at 1270 cm−1 on the sodium content for all tested adsorbents

Full size image

The amount of green condensate (green oil) was at maximum on AKB, as well as on adsorbents with a minimum concentration of sodium and transition metals giving the following values: AKB—35 ml (102 g); SA—34 ml (98.6 g); AKB—Ti(0.18) —35(98.3); SA-Ti(0.2) —34 ml (97.5 g); AN-2—34 ml (96 g); AN-Ti(0.2)—33 ml (94.8 g); AN-Na(0.3)V(0.4)—32 ml (91.8 g). Upon addition of waste SA-Ti(15.0) to the adsorbent, the amount of green oil condensate slightly changed decreasing the relative weight as follows: SA-Ti(0.15)—33 ml (90.6 g); SA-Ti(2.0)—32 ml (86.7 g); SA-Ti(5.0)—31 ml (85.2 g); SA-Ti(15.0)—32 ml (80.2 g). Modification of adsorbents with sodium hydroxide was much more effective in suppressing the acid sites of alumina and, accordingly, formation of oligomers with a high relative weight: AN-Na(1.8)—26 ml (67,6 g); AN-Na(4.3)—25 (62.3 g); SA-Ti(14.3)Na(4.9)—22 ml (53.5 g); AN-Na(6.1)*—17 ml (40 g); SA-Ti(0.14)Na(6.2)*—16 ml (39 g); SA-Ti(14.2)Na(5.7)*—16 ml (37 g); SA-Na(6.6)*—15 ml (33 g).

Conclusions

The processing of aluminum-containing waste into pseudoboehmite showed that residual impurities of vanadium and titanium in the SA adsorbent did not have a significant effect on adsorption. Adsorbents based on technical alumina and pseudoboehmite obtained from waste after alkaline modification exhibited the same adsorption properties when purifying hydrogen-containing gases from HCl, C2Cl4 and H2O. The efficiency of water vapor adsorption by SA adsorbent obtained by processing of SA waste increased by 12% compared to the commercial AN-2 adsorbent, the physical adsorption of tetrachlorethylene increased by 19%, and chemisorption of hydrochloric acid vapor by 28%.

ALD modification of SA and AN-2 decreased the adsorption capacity by 92%, 93% and 54% for water vapors, tetrachlorethylene and HCl respectively, not preventing, however, oligomerization of unsaturated hydrocarbons in the presence of inorganic and organic chlorine compounds. Titanium dioxide introduced into the adsorbent in the form of a waste aluminum-titanium carrier inhibited formation of “green oil” with a high relative mass by 8–19% when simulating the process of purifying catalytic reforming gases from chlorine-containing compounds. When titania was introduced using ALD the positive effect was much less pronounced (ca. 0.3–1.25% decrease of “green oil”). ALD technology for modifying the surface of SA and AKB granules with a TiO2 nanolayer, as well as SA-Ti(15.0) additive, after additional research, can be useful in synthesis of hydrodesulfurization catalysts diminishing the extent of interactions of the active components with the carrier.

Impregnation of AN-2 with a solution of sodium vanadate improved this descriptor by ca. 4%, modification of AN-2 with sodium hydroxide from ca. 30 to ca. 58%, SA-Ti(15.0) by ca. 54%, and SA-Na (6.6)* by 66%. A rational use of toxic waste Al2O3 will reduce the costs of its storage and decrease the amount of fresh adsorbents, also improving the environment by reducing the slippage of hydrogen chloride to less than 1 mg/m3.