Supercritical CO2 Extraction of Palladium Oxide from an Aluminosilicate-Supported Catalyst Enhanced by a Combination of Complexing Polymers and Piperidine

Precious metals, in particular Pd, have a wide range of applications in industry. Due to their scarcity, precious metals have to be recycled, preferably with green and energy-saving recycling processes. In this article, palladium extraction from an aluminosilicate-supported catalyst, containing about 2 wt% (weight%) of Pd (100% PdO), with supercritical CO2 (scCO2) assisted by complexing polymers is described. Two polymers, p(FDA)SH homopolymer and p(FDA-co-DPPS) copolymer (FDA: 1,1,2,2-tetrahydroperfluorodecyl acrylate; DPPS: 4-(diphenylphosphino)styrene), were tested with regards to their ability to extract palladium. Both polymers showed relatively low extraction conversions of approximately 18% and 30%, respectively. However, the addition of piperidine as activator for p(FDA-co-DPPS) allowed for an increase in the extraction conversion of up to 60%.


Synthesis of the Complexing Polymers
The complexing polymers, p(FDA)SH homopolymer and p(FDA-co-DPPS) copolymer, were synthesized via reversible addition-fragmentation chain transfer (RAFT) polymerization.

Synthesis of p(FDA)SH Homopolymer
FDA (40 g, 0.0771 mol), CTA (2.1407 g, 0.0084 mol), AIBN (0.4146 g, 0.0025 mol) and TFT (42 mL) were added in a Schlenk flask. The mixture was stirred magnetically, and bubbled for 40 min with N2. Afterwards, the polymerization was initiated by heating the Schlenk flask in an oil bath at 65 °C. After 2 weeks, the reaction was stopped and left to return to room temperature. The polymer was precipitated in 600 mL of pentane three times, and then dried under vacuum overnight. The precipitated polymer was aminolyzed by the addition of piperidine (5 eq.) and PPh3 (3 eq.) in TFT, with the mixture stirred magnetically and bubbled for 40 min with N2. The aminolysis reaction proceeded for 3 h. The p(FDA)SH polymer was precipitated in pentane three times and the polymer was dried under vacuum overnight. After drying, the polymer was recovered as a fine white powder (68% yield).

Synthesis of p(FDA-co-DPPS) Copolymer
FDA (42.5 g, 0.0820 mol), DPPS (7.5 g, 0.0260 mol), CTA (1.305 g, 0.0051 mol), AIBN (0.2525 g, 0.0015 mol), and TFT (54 mL) were added in a Schlenk flask. The mixture was stirred magnetically and bubbled for 40 min with N2. Afterwards, the polymerization was initiated by heating the Schlenk flask in an oil bath at 65 °C. After 96 h, the reaction was stopped and left to return to room temperature. The polymer was precipitated in 600 mL of pentane three times, and was dried under vacuum overnight. After drying, the polymer was recovered as a fine pink powder (61% yield).

Polymer Characterization
The polymer composition was determined by 1 H-NMR spectroscopy with a Bruker Avance 400 MHz spectrometer at room temperature. The spectrum was recorded by dissolving 10 mg of polymer in 0.5 mL of CFC-113 with C6D6 capillary tubes. The experimental conditions for recording 1 H-NMR spectrum were as follows: flip angle 30°, acquisition time 4 s, pulse delay 1 s, and 32 scans. The degrees of polymerization of the two different monomer units (DPFDA and DPDPPS) were calculated based on the following formula, where Hi corresponds to the integral of the protons i in the 1 H-NMR spectrum (cf. Figure S2 The degrees of polymerization of the two different monomer units (DPFDA and DPDPPS) were calculated based on the following formula, where Hi corresponds to the integral of the protons i in the 1 H-NMR spectrum (cf. Figure S3

Cloud Point Curves of Polymers in scCO2
The cloud points were measured with a polymer content of 1 wt% in CO2 with the following procedure.
Cloud-point measurements were carried out in a high pressure, variable volume view cell equipped with a sapphire window on the end for visual observations. The cell was equipped with a pressure transducer and an internal thermocouple. It was thermostated by a water/isopropanol mixture delivered by a Lauda RE206 circulating pump. CO2 was delivered by an ISCO 260D automatic syringe pump. A total of 50-55 mg of polymer was weighed and transferred to the cell along with a clean magnetic stir bar at a starting cell volume of 6.39 mL. Subsequently, the cell was fed with CO2 at about 25 °C and 10.9 MPa. Then, the cell was heated to 65 °C (taking care to adjust the volume of the cell in order to stay below a pressure of 35 MPa; safety rupture disk at 50 MPa) and then cooled by steps of 5 °C down to 25 °C. Cloud points (one-phase/two-phase transition) were obtained by decreasing the pressure of the cell by increasing the cell volume through a handdriven piston after 20 min of stirring at a given temperature. The uncertainty of the cloud point pressure was ±0.5 MPa.
As can be seen in Figures S4 -S6, all three polymers were completely soluble in scCO2 at the extraction conditions, 25 or 27 MPa and 40 or 60 °C.

Characterization of Catalyst Cat D
The catalyst characterization was reported previously [2], as the same batch of catalyst was used in the present study. The Cat D characterization is shown here again, just for reference.

SEM-EDX
The SEM-EDX analyses were done with a ZEISS EVO HD15 coupled with an EDX ATzec (Oxford instrument) apparatus. The catalyst Cat D (2 wt% Pd) was deposited as a    The average size of the catalyst was 80 micrometers, measured by SEM-EDX (cf. Figure S7). From the SEM-EDX studies performed on Cat D, it was found that about 72% of the Pd was present on the surface of the catalyst (cf. Figure S8), but the precious metal was also present in the interior of the support (cf. Figure S9

XPS
XPS measurements were carried out with a THERMO Escalab spectrometer, using focused monochromatic Al Kα radiation (hν = 1486.6 eV). Peaks were recorded with constant pass energy of 20 eV. Charge neutralization was used for all the acquisitions. The pressure in the analysis chamber was around 5 × 10 -11 MPa. Short acquisition time spectra were recorded before each experiment to check that the samples did not suffer from degradation during the measurements. The binding energy scale was calibrated using the C 1 s peak at 285.0 eV from the hydrocarbon contamination invariably present. The curves fit for core peaks were obtained using a minimum number of components.  The XPS characterization ( Figure S10, Table S1) allowed for the study of the oxidation state of the precious metal on the aluminosilicate support. For palladium, the Pd 3d spectrum was recorded. The Pd 3d spectrum corresponds to a doublet, due to the spin orbit splitting of the d orbital. Hence, the Pd has two peaks named Pd 3d 5/2 and Pd 3d 3/2. For Cat D (cf. Figure S10), the presence of a unique peak at 336.6 eV, typical for Pd(II)O species (100%) was observed [3].

TEM
TEM images were obtained with a Jeol 1200EXII transmission electron microscope at an operating voltage of 100 kV, with images captured with a Quemesa camera from Olympus Soft Imaging Solutions. Supports were crushed into powder form, and embedded into an Embed 812 resin, which was then microtomed using an Ultramicrotome Ultracut UCT from Leica Microsystems, equipped with a DiATOME ultra diamond knife, and placed on a 300-mesh copper grid for TEM analysis.  For Cat D (pristine catalyst, 100% PdO), TEM studies showed nanoparticles with an average diameter of 2.7 nm, and with a relatively low dispersity in size (cf. Figure S11).

Nitrogen Adsorption-Desorption Isotherms (BET)
The mesopore size distributions and specific surface area were determined by nitrogen adsorption-desorption isotherms (BET) using an ASAP-2020 physisorption analyzer (Micromeritics). The samples were heated at 120 °C under reduced pressure (10 -3 MPa) for 24 h before the analysis.  For catalyst Cat D, a surface area of 122 m 2 /g was measured with an average pore diameter of 21 nm ( Figure S12, Table S2), large enough to allow the fluorinated polymers (ca. 10 nm diameter micelles [4,5]) to enter the catalyst pores. Cat D pore size distribution

ICP-OES Analysis for Pd Content Determination in the Samples
The content of Pd in the different samples (pristine catalyst, bubbling water/acetone solution with Pd and polymer, catalyst recovered in the extraction cell after extraction, acetone cleaning solutions of the extraction cell and the separator, acetone cleaning solution of tubes, valve and filters, and reverse osmosis membrane after extraction) collected from the extraction experiments at the two sites was measured by inductively coupled plasma optical emission spectrometry (ICP-OES) after sample digestion using mineral acids (aqua regia mixture).
An inductively coupled plasma optical emission spectrometer (ICP-OES) OPTIMA 5300DV (Perkin Elmer, USA) was used to determine Pd at the wavelengths of 340.458 nm and 324.270 nm. The operating conditions employed for ICP-OES determination were 1300 W RF power, 15 l/min plasma flow, 2.0 l/min auxiliary flow, 0.8 l/min nebulizer flow, and 1.5 mL/min sample uptake rate. Instrument calibration was performed with standard solutions prepared from commercial Pd solutions of 1000 mg/l from Merck, Darmstadt, Germany.
Digestion of solid samples was carried out in PTFE vessels of a microwave digestion oven with temperature control (Speedwave MWS-3 from Berghof GmbH). The temperature program shown in Table S3 was applied for all samples.

Digestion of sample ExS-A (the bubbling water solution containing polymer and Pd):
Sample ExS-A (up to 140 mL) was transferred to a Berzelius glass and concentrated to approx. 25 mL on a hotplate by evaporation. Then, 24 mL of aqua regia was added to the plastic bottle to leach the Pd remaining in the bottle. It was then added to the concentrated sample on the hotplate. The solution (aqua regia + sample) was boiled for 2 h to reduce to the digested sample volume (cf. SI-Chapter 4). The sample was then cooled to room temperature and filtered on a cellulose filter (circles, diam. 125 mm; Whatman) using glass funnels into a fitting volumetric flask. The final sample was analyzed for Pd by ICP-OES, and the mass concentration of Pd in the sample (cPd (mg/l)) was obtained.

Digestion of catalyst Cat D, sample ExS-B and Exc (the catalyst recovered after extraction):
100 -200 mg of sample was weighed in PTFE vessels. Then, 15 mL of aqua regia was added and introduced in the microwave oven. Afterwards, the digestion program from Table S3 was applied. After digestion, the sample was cooled down to room temperature and filtered on a cellulose filter (circles, diam. 125 mm; Whatman) using glass funnels into a fitting volumetric flask. The sample was then diluted to the digested sample volume using ultrapure water (cf. SI-Chapter 4). The resulting solution was analyzed by ICP-OES to obtain the mass concentration of Pd in the sample (cPd [mg/l]).

Digestion of sample ExS-C, ExS-D, Exa, Exb, and Exd (acetone bubbling/cleaning solutions):
Samples were transferred to PTFE vessels. For this, the bottle sent for analysis containing the sample was washed with aqua regia (14 mL) and introduced into the microwave oven. The digestion program from Table S3 was applied and after digestion, the sample was cooled down to room temperature and filtered on a cellulose filter into a fit-ting volumetric flask. The sample was diluted to the digested sample volume with ultrapure water (cf. SI-Chapter 4). The mass concentration of Pd in the sample was analyzed by ICP-OES (cPd [mg/l]).

Digestion of sample Exe (RO-membrane):
The sample (membrane) was introduced to a PTFE vessel in the microwave oven and 14 mL of aqua regia was added. Then, the digestion program shown in Table S3 was applied. The digested sample was filtered, after cooling down to room temperature, into a fitting volumetric flask, and was diluted to the digested sample volume with ultrapure water (cf. SI-Chapter 4). Afterwards the mass concentration of Pd in the sample was analyzed by ICP-OES (cPd [mg/l]).

Calculation of Pd Mass in the Samples
The Pd ppm content, ppmPd, of each sample was calculated using Equation (S1).
where cPd is the concentration of Pd in the sample (mg/l) Vsample digested is the digested sample volume (l) msample digested is the total mass of digested sample (kg).
The weight percent (wt%Pd) of Pd in the samples was calculated by Equation (S2).
The mass of Pd, mPd, in each sample was calculated either with Equation (S3) or Equation (S4).
for catalyst Cat D, Sample Exc and Sample ExS-B (where msample is the mass of the collected sample) for Samples Exa, Exb, Exd, and Exe, as well as for Samples ExS-A, ExS-C, and ExS-D.

Calculation of Measurement Errors
Two errors were taken into account. The error of the initial mass of the catalyst, as well as the error of the ICP-OES analysis. The initial error in mass was approximated to 1 mg. The ICP-OES had a measurement error of about 6%.
The total error of the extraction conversion is denoted ΔXextraction and the total errors of the extraction yield, as well as of the Pd-Balance, with ΔYextraction and ΔPd-Balance.

Xextraction is the conversion of Pd extraction
Yextraction is the yield of the Pd extraction Pd-Balance is the Pd-Balance of the extraction experiment mPd,initial weight is the initial mass of Pd (mPd) for extraction (Pd content of inserted Cat D) ΔmPd,initial weight is the error in determination of mPd,initial weight mPd,final weight is the mass of Pd remaining on the catalyst after extraction ΔmPd,final weight is the error in determination of mPd,final weight mcatalyst,initial weight is the initial mass of catalyst D for extraction mPd,recovered is the mass of Pd recovered after extraction ΔmPd,recovered is the error in determination of mPd,recovered mPd,detected is the mass of Pd detected after extraction anywhere in the system ΔmPd,detected is the error in determination of mPd,detected Table S4 shows the reactant ratios used for the control experiments. Table S5 shows the Pd content measured for each sample after extraction, and for the pristine catalyst Cat D. Based on these data, along with Equations (3) -(5) (main article), the extraction conversions and extraction yields of the control experiments, as well as the Pd-Balances, were calculated. Table S6 shows the Pd distribution in the samples after the extraction. In Table  S7, the extraction results are shown, as well as the corresponding measurement errors, calculated with Equations S5 -S17.    Table S8 shows the reactant ratios for the PPh3 extraction experiments. Table S9 shows the Pd content measured for each sample after extraction, and for the pristine catalyst Cat D. Based on these data, along with Equations (3) -(5), the extraction conversions and extraction yields of the experiments, and the Pd-Balances, were calculated. Table S10 shows the Pd distribution in the samples after the extraction. In Table S11, the extraction results are shown, as well as the corresponding measurement errors, calculated with Equations S5 -S17.   Table S12 shows the reactant ratios used in the screening experiments for p(FDA)SH extraction experiments. Table S13 shows the Pd content measured for each sample after extraction and for the pristine catalyst Cat D. Based on these data, along with Equations (3) -(5), the conversions and extraction yields of the experiments, as well as the Pd-Balances, were calculated. Table S14 shows the Pd distribution in the samples after the extraction. In Table S15, the extraction results are shown, as well as the corresponding measurement errors, calculated with Equations S5 -S17.

Pd Extraction from Catalyst Cat D with Polymer p(FDA-co-DPPS)
Table S16 shows the reactant ratios used for p(FDA-co-DPPS) extraction tests at standard conditions (40 °C, 25 MPa). Table S17 shows the Pd content measured for each sample after extraction and for the pristine catalyst Cat D. Based on these data, along with Equations (3) -(5), the extraction conversions and extraction yields of the experiments, as well as the Pd-Balances, were calculated. Table S18 shows the Pd distribution in the samples after the extraction. In Table S19, the extraction results are shown, as well as the corresponding measurement errors, calculated with Equations S5 -S17. Parameter Screening: Table S20 shows the reactant ratios used for p(FDA-co-DPPS) extraction tests at parameter screening. Table S21 shows the Pd content measured for each sample after extraction and for the pristine catalyst Cat D. Based on these data, along with Equations (3) -(5), the extraction conversions and extraction yields of the experiments, as well as the Pd-Balances, were calculated. Table S22 shows the Pd distribution in the samples after the extraction. In Table S23, the extraction results are shown, as well as the corresponding measurement errors, calculated with Equations S5 -S17.