Regeneration of Exhausted Palladium-Based Membranes: Recycling Process and Economics.

The aim of the present work is the recycling treatment of tubular α-Al2O3-supported ceramic membranes with a Pd/Ag selective layer, employed in hydrogen production with integrated CO2 capture. A nitric acid leaching treatment was investigated, and recovered ceramic supports were characterized, demonstrating their suitability for the production of novel efficient membranes. The main objective was the metal dissolution that preserved the support integrity in order to allow the recovered membrane to be suitable for a new deposition of the selective layer. The conditions that obtained a satisfactory dissolution rate of the Pd/Ag layer while avoiding the support to be damaged are as follows: nitric acid 3 M, 60 °C and 3.5 h of reaction time. The efficiency of the recovered supports was determined by nitrogen permeance and surface roughness analysis, and the economic figures were analysed to evaluate the convenience of the regeneration process and the advantage of a recycled membrane over a new membrane. The experimentation carried out demonstrates the proposed process feasibility both in terms of recycling and economic results.

Solvent versus thermal treatment for glass recovery from end of life photovoltaic panels: Environmental and economic assessment.

End of life photovoltaic panels of different technologies (poly crystalline Si, amorphous Si, and CdTe) were treated mechanically in pilot scale by single shaft shredder minimizing the production of fine fractions below 0.4 mm (<18% weight). Grounded material was sieved giving: an intermediate fraction (0.4-1 mm) of directly recoverable glass (18% weight); a coarse fraction (which should be further treated for encapsulant removal), and fine fractions of low-value glass (18%), which can be treated by leaching for the removal of metal impurities. Encapsulant removal from coarse fraction was successfully performed by solvent treatment using cyclohexane at 50 °C for 1 h giving high-grade glass (52% weight), which can be reused for panel production. Experimental results of solvent treatment were compared with those from thermal treatment by economic analysis and Life Cycle Assessment, denoting in both cases the advantages of solvent treatment in recovering high-value glass.

Physical and chemical treatment of end of life panels: An integrated automatic approach viable for different photovoltaic technologies.

Different kinds of panels (Si-based panels and CdTe panels) were treated according to a common process route made up of two main steps: a physical treatment (triple crushing and thermal treatment) and a chemical treatment. After triple crushing three fractions were obtained: an intermediate fraction (0.4-1mm) of directly recoverable glass (17%w/w); a coarse fraction (>1mm) requiring further thermal treatment in order to separate EVA-glued layers in glass fragments; a fine fraction (<0.4mm) requiring chemical treatment to dissolve metals and obtain another recoverable glass fraction. Coarse fractions (62%w/w) were treated thermally giving another recoverable glass fraction (52%w/w). Fine fractions can be further sieved into two sub-fractions: <0.08mm (3%w/w) and 0.08-0.4mm (22%w/w). Chemical characterization showed that 0.08-0.4mm fractions mainly contained Fe, Al and Zn, while precious and dangerous metals (Ag, Ti, Te, Cu and Cd) are mainly present in fractions <0.08mm. Acid leaching of 0.08-0.4mm fractions allowed to obtain a third recoverable glass fraction (22%w/w). The process route allowed to treat by the same scheme of operation both Si based panels and Cd-Te panels with an overall recycling rate of 91%.

Leaching of electrodic powders from lithium ion batteries: Optimization of operating conditions and effect of physical pretreatment for waste fraction retrieval.

Experimental results of leaching tests using waste fractions obtained by mechanical pretreatment of lithium ion batteries (LIB) were reported. Two physical pretreatments were performed at pilot scale in order to recover electrodic powders: the first including crushing, milling, and sieving and the second granulation, and sieving. Recovery yield of electrodic powder was significantly influenced by the type of pretreatment. About 50% of initial LIB wastes was recovered by the first treatment (as electrodic powder with size <0.5mm, Sample 1), while only 37% of powder with size <1mm (Sample 2) can be recovered by the second treatment. Chemical digestion put in evidence the heterogeneity of recovered powders denoting different amounts of Co, Mn, and Ni. Leaching tests of both powders were performed in order to determine optimized conditions for metal extraction. Solid/liquid ratios and sulfuric acid concentrations were changed according to factorial designs at constant temperature (80°C). Optimized conditions for quantitative extraction (>99%) of Co and Li from Sample 1 are 1/10g/mL as solid/liquid ratio and +50% stoichiometric excess of acid (1.1M). Using the same solid/liquid ratio, +100% acid excess (1.2M) is necessary to extract 96% of Co and 86% of Li from Sample 2. Best conditions for leaching of Sample 2 using glucose are +200% acid excess (1.7M) and 0.05M glucose concentration. Optimized conditions found in this work are among the most effective reported in the literature in term of Co extraction and reagent consumption.

Cobalt products from real waste fractions of end of life lithium ion batteries.

An innovative process was optimized to recover Co from portable Lithium Ion Batteries (LIB). Pilot scale physical pretreatment was performed to recover electrodic powder from LIB. Co was extracted from electrodic powder by a hydrometallurgical process including the following main stages: leaching (by acid reducing conditions), primary purification (by precipitation of metal impurities), solvent extraction with D2EPHA (for removal of metal impurities), solvent extraction with Cyanex 272 (for separation of cobalt from nickel), cobalt recovery (by precipitation of cobalt carbonate). Tests were separately performed to identify the optimal operating conditions for precipitation (pH 3.8 or 4.8), solvent extraction with D2EHPA (pH 3.8; Mn/D2EHPA=4; 10% TBP; two sequential extractive steps) and solvent extraction with Cyanex 272 (pH 3.8; Cyanex/Cobalt=4, 10% TBP, one extractive step). The sequence of optimized process stages was finally performed to obtain cobalt carbonate. Products with different degree of purity were obtained depending on the performed purification steps (precipitation with or without solvent extraction). 95% purity was achieved by implementation of the process including the solvent extraction stages with D2EHPA and Cyanex 272 and final washing for sodium removal.

Automobile shredded residue valorisation by hydrometallurgical metal recovery.

The aim of this work was developing a hydrometallurgical process to recover metals from automobile shredded residue (or car fluff). Automobile shredded residue (ASR) was characterised by particle size distribution, total metal content and metal speciation in order to guide the choice of target metals and the operating conditions of leaching. Characterisation results showed that Fe is the most abundant metal in the waste, while Zn was the second abundant metal in the fraction with diameter lower than 500 µm. Sequential extractions denoted that Zn was easily extractable by weak acid attack, while Fe and Al required a strong acid attack to be removed. In order to recover zinc from <500 µm fraction leaching tests were operated using acetic acid, sulphuric acid and sodium hydroxide at different concentrations. Sulphuric acid determined the highest zinc extraction yield, while acetic acid determined the highest zinc extractive selectivity. Sodium hydroxide promoted an intermediate situation between sulphuric and acetic acid. Zn recovery by electro winning using acetic leach liquor determined 95% of Zn electro deposition yield in 1h, while using sulphuric leach liquor 40% yield in 1h and 50% yield in 2h were obtained. Simulation results showed that the sulphuric leaching process was more attractive than acetic leaching process.

Isolation and quantification of cadmium removal mechanisms in batch reactors inoculated by sulphate reducing bacteria: biosorption versus bioprecipitation.

Biosorbing properties of sulphate reducing bacteria were tested to distinguish the amount of cadmium removed by bioprecipitation from that bound onto biomass surface (biosorption). Experimental results of cadmium abatement in batch growth tests (bioprecipitation tests) were then compared with metabolism-independent binding properties of SRB cell wall surface (biosorption tests performed with dead biomass). Experimental results showed that SRB inoculum removed 59 + or - 5% of sulphates in 21 days even in presence of cadmium (0-36 mmol L(-1)), while non-monotonous kinetic effects were observed for increasing Cd concentrations. Comparison between bioprecipitation and biosorption tests denoted a significant contribution of biosorption (77%) in total Cd removal (0.40 + or - 0.01 mmol g(-1)). Characterisation of bacterial acid-base surface properties by potentiometric titrations and mechanistic modelling denoted that carboxylic, phosphate and amino groups of cell wall are the main responsible of metal removal by biosorption mechanism.

Assessment of solid reactive mixtures for the development of biological permeable reactive barriers.

Solid reactive mixtures were tested as filling material for the development of biological permeable reactive barriers for the treatment of heavy metals contaminated waters. Mixture selection was performed by taking into account the different mechanisms operating in sulphate and cadmium removal with particular attention to bioprecipitation and sorption onto the organic matrices in the mixtures. Suspensions of eight reactive mixtures were tested for sulphate removal (initial concentration 3 g L(-1)). Each mixture was made up of four main functional components: a mix of organic sources for bacterial growth, a neutralizing agent, a porous medium and zero-valent iron. The best mixture among the tested ones (M8: 6% leaves, 9% compost, 3% zero-valent iron, 30% silica sand, 30% perlite, 22% limestone) presented optimal conditions for SRB growth (pH 7.8 +/- 0.1; E(h)= -410 +/- 5 mV) and 83% sulphate removal in 22 days (25% due to bioreduction, 32% due to sorption onto compost and 20% onto leaves). M8 mixture allowed the complete abatement of cadmium with a significant contribution of sorption over bioprecipitation (6% Cd removal due to SRB activity). Sorption properties, characterised by potentiometric titrations and related modelling, were mainly due to carboxylic sites of organic components used in reactive mixtures.

Non-electrostatic surface complexation models for protons and lead(II) sorption onto single minerals and their mixture.

Potentiometric titrations and lead sorption tests were conducted using muscovite, clinochlore, hematite, goethite, quartz, and a mixture of these same minerals. Mechanistic models were developed to represent and interpret these data. The aim was isolating the specific contribution of each mineral in proton and lead binding. Acid-base properties of each single mineral as well as their mixture were represented by discrete models, which consider the dissociation of n monoprotic sites (n-site/n-K(H) models). A one-site/one-K(H) model (logK(H1) = 10.69) was chosen for quartz (dissociation of SiOH edge hydroxyl groups). Goethite and hematite (FeOH groups) were represented by the same one-site/one-K(H) model (logK(H1) = 10.35). Three-site/three-K(H) models were used for muscovite (logK(H1) = 4.18; logK(H2) = 6.65; logK(H3) = 9.67) and clinochlore (logK(H1) = 3.84; logK(H2) = 6.57; logK(H3) = 9.71) assuming that SiOH and AlOH of the aluminosilicate matrix dissociate in the acid-neutral pH range while SiOH groups of quartz inclusions dissociate in the basic range. Similarly, the mixture of these minerals was represented by a three-site/three-K(H) model (logK(H1) = 3.39; logK(H2) = 6.72; logK(H3) = 10.82). According to crossed comparisons with single minerals, the first two sites of the mixture were associated with the aluminosilicate matrix (SiOH and AlOH respectively) and the third site with iron oxides (FeOH) and quartz groups. Additivity of proton binding in the mixture was demonstrated by simulating the mixture's titration curve. A unified model for the entire set of titration curves (single minerals and mixture) was also developed introducing a three-peak distribution function for proton affinity constants. Experimental data for lead sorption onto the mixture and individual minerals in 3-5 pH range denoted the competition between protons and metallic ions. The entire set of lead isotherms (individual mineral and mixture data) was represented adequately by a unified model taking into account both monodentate and bidentate complexes with the three active sites (additivity of lead binding). Experimental data of metal distribution in solid and liquid phases were successfully simulated by implementing the protonation and the surface complexation constants into the database of a dedicated software for chemical equilibria.

Biosorption of protons and heavy metals onto olive pomace: modelling of competition effects.

Heavy metal biosorption onto solid wastes from olive oil production plants, olive pomace, has been investigated. Acid-base properties of the active sites of olive pomace were determined by potentiometric titrations and represented by a continuous model accounting for two main kinds of active sites. Competition among protons and heavy metals in solution was considered by performing biosorption tests at different equilibrium pH with single (Cu and Cd) and binary metal systems (Cu-Cd). Both Langmuir extensions and non-ideal competitive adsorption models (NICA models) can be used to represent experimental data of Cu and Cd biosorption in single metal systems at different equilibrium pH. Nevertheless only NICA models, accounting for site heterogeneity and non-ideal adsorption of the different species simultaneously present in solution, can adequately simulate the competition among Cu and Cd in binary metal systems by using the parameters fitted to single system data.

Proton binding onto soil by nonelectrostatic models: isolation and identification of mineral contributions.

In this paper a methodological approach is proposed to validate mechanistic modeling for proton binding onto active sites of mineral and soil samples by reducing the uncertainty and arbitrariness of model schematization. This approach is based on the quantitative formulation (X-ray calibration method) of a simulating mineral mixture (SMM) accounting for the main mineral phases in the soil (quartz, goethite, hematite, muscovite, clinochlore). Mineral and organic contributions were separated by comparing titration curves of river sediment and SMM. Specific mineral contributions to the acid properties of SMM were separated by comparing titration models of SMM and single minerals. Different nonelectrostatic models were used for titrations of SMM and single minerals: two-site/three-KH models (one amphoteric plus one monoprotic site) for clay minerals and SMM; one-site/two-KH models (one amphoteric site) for goethite and hematite; and a one-site/one-KH model (one monoprotic site) for quartz. Crossed-comparisons of titration models allow for identifying and quantifying the specific contributions of the distinct edge hydroxyl groups of iron oxides, clay minerals, and quartz in the different pH ranges. In particularthe amphoteric sites of aluminosilicates mainly contribute in the acid-neutral pH range, the amphoteric sites of iron oxides take part in the neutral-basic range, and finally the monoprotic edge hydroxyl groups of quartz react in the upper basic region of pH. The good simulation of the acid-base properties of SMM (according to single mineral titration models and quantitative composition by X-ray) confirms both model schematization and SMM formulation. Speciation diagrams of the active sites of the different mineral components (aluminosilicates, iron oxides, and quartz) were obtained by implementing the database of a dedicated software with the apparent equilibrium constants regressed by titration modeling of single minerals.

Modelling of the acid-base properties of natural and synthetic adsorbent materials used for heavy metal removal from aqueous solutions.

In this paper a comparison about kinetic behaviour, acid-base properties and copper removal capacities was carried out between two different adsorbent materials used for heavy metal removal from aqueous solutions: an aminodiacetic chelating resin as commercial product (Lewatit TP207) and a lyophilised bacterial biomass of Sphaerotilus natans. The acid-base properties of a S. natans cell suspension were well described by simplified mechanistic models without electrostatic corrections considering two kinds of weakly acidic active sites. In particular the introduction of two-peak distribution function for the proton affinity constants allows a better representation of the experimental data reproducing the site heterogeneity. A priori knowledge about resin functional groups (aminodiacetic groups) is the base for preliminary simulations of titration curve assuming a Donnan gel structure for the resin phase considered as a concentrated aqueous solution of aminodiacetic acid (ADA). Departures from experimental and simulated data can be interpreted by considering the heterogeneity of the functional groups and the effect of ionic concentration in the resin phase. Two-site continuous model describes adequately the experimental data. Moreover the values of apparent protonation constants (as adjustable parameters found by non-linear regression) are very near to the apparent constants evaluated by a Donnan model assuming the intrinsic constants in resin phase equal to the equilibrium constants in aqueous solution of ADA and considering the amphoteric nature of active sites for the evaluation of counter-ion concentration in the resin phase. Copper removal outlined the strong affinity of the active groups of the resin for this ion in solution compared to the S. natans biomass according to the complexation constants between aminodiacetic and mono-carboxylic groups and copper ions.

Olive mill solid residues as heavy metal sorbent material: a preliminary study.

Biosorption of heavy metals is an innovative and alternative technology to remove these pollutants from aqueous solutions using inactive and dead biomasses such as agricultural and industrial wastes, algae and bacteria. In this study olive mill solid residue was used as heavy metal adsorbent material for its wide availability as agricultural waste and also for its cellulosic matrix, rich of potential metal binding active sites. Preliminary studies concerned with the removal of different heavy metals (Hg, Pb, Cu, Zn and Cd), the effect of pre-treatments by water and n-hexane and the regeneration possibility. Olive mill solid residue resulted able to remove heavy metals from aqueous solutions with an affinity series reflecting the hydrolytic properties of the metallic ions, but also a particular affinity for copper. It can be supposed that biosorption phenomenon occur by a general ion exchange mechanism combined with a specific complexation reaction for copper ions. Water pre-treatment is sufficient to reduce COD release in the effluent according to the law limit, while n-hexane pre-treatment strongly reduces also the adsorption properties of this material. Experimental isotherms obtained under different operating conditions were fitted using a non linear regression method for the estimation of the Langmuir parameters. Moreover a simple Scatchard plot analysis was performed for a preliminary investigation of the active sites, showing the presence of two different site affinities depending on the metal concentration, according to the previous hypothesis of two kinds of uptake mechanisms for copper biosorption. Regeneration tests gave good results in terms of yield of regeneration and also concentration ratios.