Separation of Palladium and Rhodium from the Spent Metal-Honeycomb Catalysts by Pulsed Discharge without Chemical Additives

Chiharu Tokoro; Yuto Imaizumi; Taketoshi Koita; Akiko Kubota; Yutaro Takaya; Keishi Oyama; Md. Mijanur Rahman

doi:10.2320/matertrans.M-M2024808

Abstract

One of the main causes of atmospheric pollution comes from mobile sources that emit noxious gases from internal combustion engines. For the suppression of toxic exhaust gases, a catalytic converter is used as an anti-pollution device, because it catalyzes or accelerates the degradation of emissions making them less harmful. In the catalytic converter, platinum (Pt), palladium (Pd) and rhodium (Rh) catalysts are normally used for their excellent performances. However, these metals are expensive, rare, and scarce in the earth’s crust. These precious metals thus emphasized the importance of developing efficient recycling practices. For the recycling of these precious metals, a very fast, easy, economical, environmentally friendly and high-safety electric pulse discharge method was used to separate the Pd and Rh from the spent metal-honeycomb (MH) catalytic converter consisting of p as the catalyst carrier. To standardize the precious metals recycling process and attain the highest level of separation of Pd and Rh, electric pulse shots ranging from 120–240 and standoff distances (gap width between the positive electrode and the sample surface) ranging from 5–10 mm were applied. During this process, an electrical explosion occurred within the honeycomb structures through the electric shock. After the electrical explosion, the particles are collected, then sieved and physical characterizations are performed. Scanning electron microscope (SEM) and energy dispersive spectroscopy (EDX) analyses revealed that the most separated particles are highly pure, and dispersed homogeneously without destroying particle structures. In this article, we first report about 87.91% of Pd and 90.77% of Rh are separated from each catalytic converter using the electric pulse discharge method that can overcome challenges and achieve ambitious recycling targets in the recycling industry. The effects of the discharge energies and the shock energy determined by varying electric pulse shots and standoff distance, respectively, are discussed. These findings provide insight into the recovery of precious metals for electric pulse discharge and play a role in developing a very effective recycling method not only for the catalytic converter but also for other devices.

1. Introduction

Catalytic converters with honeycomb structures are widely used in internal combustion engines, such as automobiles, diesel-electric, marine vessels, agriculture and construction machinery, due to control or reduction of carbon monoxide (CO) is a poison for any air-breathing animals, nitrogen oxides (NO_X) lead to acid rain, and hydrocarbons produce smog in the atmosphere [1–3]. In order to achieve a low-carbon society, the reduction of toxic emissions through the use of a catalytic converter is considered a key solution. The catalytic converter is a device that promotes oxidation and reduction (redox) reactions at lower temperatures to convert harmful compounds into harmless ones. To reduce exhaust emissions to more acceptable levels, high-performance catalytic converters are needed. For high-performance catalytic converters, platinum (Pt), palladium (Pd), and rhodium (Rh) known as noble metals or precious metals are used as catalysts because of their unique properties such as high melting point, heat resistance, high electric conductivity, and high resistance to corrosion and oxidation [4–6]. The unique chemical and physical properties of these catalysts [7, 8] are summarized in Table 1. It is noteworthy that these precious metals are also used in electric components and semiconductors, renewable energy, defense industries, and other industrial applications, and their utilization is increasing rapidly. However, precious metals are very expensive than other base metals, and have acute scarcity in the earth’s crust, shrinking day by day. Secure access to precious metals is necessary to support our modern lifestyle, and therefore the development of effective recycling practices is crucial for a recycling-oriented society.

Table 1 Chemical and physical properties of Pt, Pd and Rh.

Meanwhile, many research groups have devoted enormous efforts to efficiently extracting rare earth metals/precious metals from the spent catalytic converters through commonly available recycling methods such as chemical methods [9–11], mechanical methods [12, 13], physical methods [14, 15], and so on. Although some progress has been made in recycling research, a number of challenges remain for the large-scale commercially viable, quick and simple recycling of precious metals. Le et al. [16], recovered spent Pd catalysts by using hydrometallurgy as the main method which is limited due to the high volume of toxic and costly reagents required, and the subsequent production of hazardous waste such as nitric oxides. Therefore, it is considered inefficient for removing metals from waste solutions at low concentrations, leading to waste storage problems and reducing profits. Harjanto et al. [17], employed the leaching method based on the formation of their chloro-complexes in various concentrations of acid solution. Recovery of Pt, Pd, and Rh was carried out from the samples after hydrogen reduction pretreatment in the leaching process. In this process, large amounts of waste effluent are generated, which are generally highly toxic. Electrodeposition is considered a practical and convenient method of metal recovery due to its operational feasibility with controlled deposition. Although higher recoveries were observed in this method by increasing mass transport by using the flow of the plating solution, the overall efficiency of the recycling is not necessarily easily compared [18]. Therefore, a more environmentally sustainable and direct method without using additives or even toxic chemicals is necessary for recovering precious metals.

In this study, we applied an electric pulse discharge method with low energy discharges to separate the precious metals of Pd and Rh from the spent metal-honeycomb (MH) catalytic converters with high recovery rates. Here, recovery refers to the state of the particles obtained by filtration and drying. The details of the MH catalytic converter are discussed in the experimental section, 2.1. In previous studies discussed elsewhere [19, 20], it is anticipated that the discharge energy may lead to the destruction of the adhesive layer, which results in a rapid recovery of adherend materials at a low cost. For experiments, electric pulse shots ranging from 120–240 and standoff distances (gap width between the positive electrode and the sample surface) ranging from 5–10 mm were applied in order to demonstrate the effect of discharge energies and shock energy on the recovery rate of precious metals, respectively. The liberation of precious metals from the MH catalytic converters could be achieved by controlling the discharge energy, voltage, and some other parameters (not discussed in detail here, see Refs. [15, 19, 20]) such as current waveform, and location where heat and shock waves are generated. The separated particles were further evaluated quantitatively and qualitatively in terms of their recovery.

2. Experimental

2.1 Samples

In this study, the samples were spent MH catalytic converters used in typical and industrial products which were supplied by the domestic company. Figure 1 shows the photographs of the sample (spent MH catalytic converter) and its cross-sectional structure. The catalytic converter consists of a honeycomb-structure body in which a large number of through holes are placed side by side oriented in the longitudinal direction of the structure. The partitions between holes are covered with catalysts of precious metals like Pd and Rh, on the ceramic materials Al₂O₃, CeO₂ and TiO₂ used from the standpoint of strength, heat resistance, corrosion resistance, etc. The honeycomb structure provides a high catalyst surface area, which maximizes the contact between the catalysts and the pollutants in the hot exhaust gases. The toxic gases pass through the body of the honeycomb structure which acts as a catalyst carrier. Outside the honeycomb structure, MH catalytic converters have a cylindrical casing and a holding material prepared from steel use stainless (SUS 440), which contains Iron (Fe), Chromium (Cr), Aluminium (Al), Nickel (Ni), Zirconium (Zr), and Titanium (Ti).

Fig. 1

Schematic illustration of the spent MH catalytic converter; (a) the side view, and (b) the top view.

2.2 Experimental setup

The schematic illustration of the circuit diagram and the experimental setup of the electrical pulsed discharge is shown in the supplementary Fig. A1(a). The circuit consists of a charging and discharging section. A 1 kΩ resistor (TE1500B1K0J, TE Connectivity) was connected between the DC power supply (152A-40KV-POS, TDK Lambda Corp.) and the parallel capacitor (FL40W804KWFAAA, Shizuki Electric Co., Inc.). The capacitor was charged with the high-voltage DC power supply when the mechanical switch (E60-DT-80-1-15-BD, Ross Engineering) was placed on the charging side. Subsequently, the mechanical switch was placed on the discharging side to discharge the energy stored in the capacitor once the charge was completed. The charging current was limited by a charging resistor to protect the DC regulator during discharging. The discharging section was the RLC circuit consistent with resistance (R), an inductor (L), and a capacitor (C). The voltages across the capacitor and the current that flowed through the samples were measured by using a voltage probe (HV-P60A, Iwatsu Electric Co., Ltd.) and a current transformer (110A-EOR, Pearson Electronics), respectively. The voltage and current waveforms in the pulsed discharge were recorded on a digital oscilloscope with a frequency band of 1 MHz (HDO4104A, Teledyne LeCroy).

A positive electrode rod and a negative electrode plate of the pulsed discharge were made from stainless steel. The sample was set on the negative plate and the positive electrode was placed upward to the sample in the water as shown in the supplementary Fig. A1(b). The previous study clarified that the pressure of the shock wave generated by the pulsed discharge was increased as the gap distance of electrodes increased [21]. Since the pulsed discharge occurs in the standoff distance, the standoff distance corresponds to the gap distance (i.e., gap width between the positive electrode and the sample surface) in this study. The standoff distance was considered to examine the impact of shock energy with the wave pressure on the separation of precious metals from the spent MH catalytic converter in the pulsed discharge. To consider the effect of standoff distance on the separation, the distance was changed to 5 mm and 10 mm.

2.3 Experimental procedure

The sample was immersed in distilled water (resistivity approx. 18 Ω cm) in the cylindrical chamber. The chamber was filled with 0.6 L of distilled water to sufficiently submerge the sample. All experiments were carried out in distilled water. The charged energy in the capacitor at a single pulse shot, E_C, was calculated by eq. (1) as follows [22].

\begin{equation} E_{\text{C}} = 1/2 CV^{2} \end{equation}

(1)

where C is the capacitance of the capacitor, C = 0.8 mF, and V is the charging voltage in the capacitor, V = 36 kV. Here, we considered lower-charged energy in order to eliminate unexpected contamination and save energy from an economic point of view. Based on the experimental conditions, three identical samples were used in this study referred to as samples 1, 2, and 3. Pulsed discharges were performed on samples 1, 2, and 3 under conditions of 120 discharge shots at a distance of 5 mm, 120 shots at a distance of 10 mm, and 240 shots at a distance of 10 mm, respectively. The details of experimental conditions for each sample are listed in Table 2. The water temperatures at the pulsed discharge for the sample 1, 2 and 3 were 21.9, 21.3 and 21.0 degree. The repeated trial number for these samples applied by the discharge with the conditions shown in Table 2 were once, respectively. To investigate the impact of electrical pulses on the extraction of precious metals from samples, these conditions were applied and the desired materials were collected for further analysis. The power consumption generated in pulsed discharge conditions for each sample was measure by the power meter, and these are presented in Table 2. The electric power of pulsed discharge occurred only in the DC power supply installed in the discharge apparatus which generates the voltage in eq. (1). The power consumption measured in the charging voltage of 36 kV per one discharge shot was 0.397 Wh. The total power consumption in 120 and 240 shots were 47.7 Wh and 95.3 Wh, respectively.

Table 2 Pulse discharge conditions for each sample.

2.4 Particles collection and analysis

After each pulse discharge measurement, the water in the experimental chamber was filtered through the 0.45 µm mesh glass filter paper. Then, the filtrates were collected and air-dried overnight at ca. 60°C in a tube furnace. Subsequently, the filtrates were sieved and analyzed as separate particles of less than 32 µm to more than 2.36 mm. The elemental distributions as well as the elemental analysis and chemical composition of the particles were performed using a scanning electron microscope and energy-dispersive X-ray spectrometer (EDX: Miniscope^®TM4000 PlusII, Hitachi High-Tech Fielding Corp., Japan). Further analysis of the elements and the quantification of the concentration of the recovered particles by each size group were evaluated using energy-dispersive X-ray fluorescence (XRF) spectrometer (EDX-700HS EDX-700HS2 Rayny Energy Dispersive X-Ray Spectrometer, Shimadzu Corp., Japan) after sieving.

3. Results and Discussions

3.1 Observation of sample after pulsed discharge

Figure 2 shows the photographs of MH catalytic converters before and after electrical pulsed discharge. After the discharge, the destroyed honeycomb structures of the catalytic converters and different sizes of blackish particles were observed visually in the water chamber, where the samples were set up for experiments. Tokoro et al. [23], reported that the higher discharge energy results in a higher recovery rate via sample crushes due to the rapidly rising temperature near the electrodes. However, extreme crushing of the sample due to higher discharge energy may contaminate the recovered particles. During pulsed discharge, the result of thermal expansion also generates shock waves which may cause the plasma near the discharge electrodes. Consequently, the volume of the active layer (i.e., honeycomb structures with catalyst layers) widened and was also destroyed. In the case of 120 pulse shots at a distance of 5 mm between the electrode and the upper surface of the sample, the honeycomb structures of the sample are mildly destroyed as shown in Fig. 2(a). For 120 pulse shots at a distance of 10 mm, the honeycomb structures of the sample are moderately destroyed, but honeycomb structures are severely destroyed with 240 pulse shots at a distance of 10 mm, as illustrated in Fig. 2(b) and (c), respectively. These different patterns of crushing of the samples could be attributed to the different discharge energies used for the experiments. The shock wave which was induced by the pulsed discharge in the standoff distance between the positive discharge electrode and the sample surface crushed also the sample in this electrical disintegration method [25, 26]. The pressure of shock wave causing the crush is depended on the standoff distance which is the current passing length. The pressure increases as the distance increases in the same conditions of discharge voltage and times because the energy emitting the shock wave pressure which is generated in the length increases as the length increases. So that, it is possible to consider the sample 2 applied by the discharge with the distance of 10 mm was crushed harder than the sample 1 with the distance of 5 mm due to the increased shock wave with the two times distance.

Fig. 2

Photographs of the spent MH catalytic converters after electric discharge measurements for (a) 120 pulse shots at a distance of 5 mm, (b) 120 pulse shots at a distance of 10 mm, and (c) 240 pulse shots at a distance of 10 mm.

3.2 Distribution and compositional analysis of recovered particles

To further analysis of the particles, the crushed materials were collected followed by filtration. After drying, the particles were separated into groups of less than 30 µm to more than 2.36 mm by sieve analysis, as shown in the supplementary Table A1–A3. From these tables and the XRF analysis, the particle size distribution of Pd and Rh in each particle size fraction was estimated, and a comparison was made over Pd and Rh for samples 1, 2, and 3. Figure 3(a) shows the cumulative mass ratio for all materials recovered from the samples. The distribution of sample 3 tended to be different from the particle distribution of sample 1 and sample 2. For samples 1 and 2, approximately 40% of the particles with particle sizes smaller than 0.150 mm to 0.032 mm pass cumulatively, while approximately 20% of the particles pass cumulatively for sample 3. This means that in comparison with the overall particle distribution, the particle size distribution in samples 1 and 2 is finer and the particle size distribution in sample 3 is coarser. The coarse particles in sample 3 are predominantly Pd and Rh particles that can be distinguished from Fig. 3(b) and (c). It is assumed that fine particles are formed by the agglomeration of nanoparticles generated by the electric explosion. On the other hand, larger particles are generated by partially breaking the samples due to the elevated temperature of the explosion [24]. Although the overall distribution pattern of Pd and Rh particles for all samples is almost identical, the distribution of these particles for sample 3 is quite different and is coarser as discussed above. The SUS plates in the honeycomb structures were clashed by the discharge in the sample 3 because of the increased discharge shot as seen in Fig. 2(c). Splinters of the plate were contaminated in the recovered materials. So that it was suitable considered that the difference of particle size distribution of sample 3 was caused by the plates contamination in the particle sizes greater than 1 mm due to the increasing weight of recovered that size as shown in Fig. 2(c).

Fig. 3

Particle size distribution of (a) recovered materials from the samples, (b) Pd component, and (c) Rh component for all samples.

The weight distribution of the recovered Pd and Rh with size groups of 0.032–0.075 mm and less than 0.032 mm is higher in each sample, and the recovery (particles obtained by filtration and drying) rates of Pd and Rh are also higher. The distribution of Pd and Rh in each group was estimated from samples 1, 2 and 3 based on XRF analysis. It is clear that the particle size group higher than 0.032–0.075 has no significant concentration for every sample, which means that the majority of the Pd and Rh particles exist in the small size group. It implies that the recovered Pd and Rh particle sizes are about 0.032 mm or lower. The element distributions did not change for each particle size group, meaning that the composition of the recovered materials remained unchanged through electric pulsed discharge measurement. The percent recovery rate of Pd and Rh in each size group for a sample can be calculated as follows.

\begin{equation} R_{\text{(Pd/Rh)_size gr.}}\ (\%) = mR_{\text{(Pd/Rh)_size gr.}}/mT_{\text{(Pd/Rh)_sample}} \times 100 \end{equation}

(2)

where mR_{(Pd/Rh)_size gr.} is the weight of Pd/Rh recovered from each size group for a sample, and mT_{(Pd/Rh)_sample} is the initial total weight of Pd/Rh in each sample. Similarly, the percent recovery rate of Pd and Rh for a sample can be calculated as follows.

\begin{equation} R_{\text{(Pd/Rh)_sample}}\ (\%) = mR_{\text{(Pd/Rh)_sample}}/mT_{\text{(Pd/Rh)_sample}} \times 100 \end{equation}

(3)

where mR_{(Pd/Rh)_sample} is the total weight of Pd/Rh recovered from each sample. The recovery rate for Pd and Rh was calculated from all samples using eqs. (2) and (3), as shown in Table 3. The calculated average recovery rate of Pd for samples 1, 2, and 3 is 78.66%, 79.60%, and 87.91%, respectively. And the calculated average recovery rate of Rh in samples 1, 2, and 3 is 87.22%, 88.36%, and 90.77%, respectively. It is clear that the recovery rates for Pd and Rh increased for samples 1, 2 and 3. Comparing the measurement conditions for samples 1 and 2, it may be clear that the distance between the electrode and sample played an important role in increasing the recovery rate of Pd and Rh particles under that diameter of 0.075 mm as seen in Fig. 3. Note that the distance from the electrode to the top of the sample is 5 mm for sample 1, and 10 mm for sample 2, respectively. Therefore, it may be ascribed that the shock wave which in turn the shock impact has a significant effect on material recovery [19, 20, 23]. By comparing the measurement conditions of samples 2 and 3, the recovery rate of Pd and Rh particles were also increased under that diameter of 0.075 mm as the number of pulse shots increased from 120 to 240. This means that a higher discharge energy has been applied to sample 3, resulting in a strong electrical blast that generates massive sample destruction. Based on the above results and discussion, it can be concluded that the 240 pulses have a significant effect on the recovery of Pd and Rh.

Table 3 The percentage recovery rate for Pd and Rh in all samples.

3.3 Elemental analysis of recovered particles

The EDX measurement was performed for elementary analysis of the recovered particles. For EDX measurements, the particle size of approximately 0.032 mm for all samples was used due to the presence of high concentrations of Pd and Rh confirmed by the XRF measurements as discussed earlier. The EDX spectrum analysis indicated that C and O were the dominant components, followed by Al among the commonly found recovered materials in all samples. In addition, Zr, Fe, Ce, Ti, Ni, Cr, Pd, and Rh are also present in the samples because of SUS 440 alloy and AlO₂CeO₂TiO₂ ceramic materials. In this study, C and O are not taken into account in the estimation of elemental composition. From the EDX spectrum analysis, the average composition of each sample is estimated by examining the changes in atomic ratios for each sample. For estimation, the percentage of atomic concentration of Al, Fe, Zr, Ce, Ti, Ni, Cr, Pd, and Rh is normalized considering 100 in total for each sample. The amount of Pd element lies in the range of 0.53 to 0.57 atom%, while Rh lies in the range of 0.32 to 0.36 atom% depending on the samples. Compared to samples 2 and 3 with sample 1, the amount of Pd element decreased slightly, while the Rh element is vice versa. This could be attributed to the effect of shock energy and discharge energy on the recovery of Pd and Rh. However, it should be noted that no drastic changes in the concentrations are observed considering Pd and Rh elements as well as other elemental impurities are not detected in the samples, as illustrated in Fig. A2(a)–(c).

To further investigate the recovered particles, EDX mapping images have been acquired. EDX mapping images of Al, Fe, Zr, Ce, Ti, Ni, Cr, Pd, and Rh are shown in Fig. 4(a1)–(a12), (b1)–(b12), and (c1)–(c12). From the EDX mapping images, one can see that the Pd and Rh are uniformly dispersed. As no agglomeration or alloying nature of the recovered particles is confirmed, this means that the particles are fully liberated.

Fig. 4

SEM image and corresponding EDX mapping images of (a1)–(a12) sample 1, (b1)–(b12) sample 2, and (c1)–(c12) sample 3.

4. Conclusion

In summary, we have successfully separated the precious metals, in particular Pd and Rh from the spent MH catalytic converter by using a high-voltage pulsed discharge and characterized by XRF and EDX elemental mapping. Different physical properties such as particle distribution, weight distribution, and percent recovery rate for both Pd and Rh were determined. Furthermore, the elemental composition and distribution of the recovered Pd and Rh were determined for all samples. All samples are found to have similar physical and structural properties, although Pd and Rh recovery rates are different. Finally, Pd and Rh recovery via sample-electrode distance and pulse times were investigated. The calculated average recovery rate of Pd and Rh is approximately 78.66% and 87.22%, 79.60% and 88.36%, and 87.91% and 90.77% for samples 1, 2, and 3, respectively. Although not optimized to recover the maximum amount of Pd and Rh from samples, it was found that by comparing the recovery of Pd and Rh from samples 1 and 2, the increased standoff distance between the positive electrode and the upper surface of sample from 5 mm to 10 mm increased the recovery from 78.66% to 79.6% for Pd, and 87.22% to 88.36% for Rh, respectively. By comparing the Pd and Rh recovery rates on samples 2 and 3, the increase in pulse shots from 120 to 240 increased the Rh recovery rates from 79.60% to 87.91% and Rh that rates from 88.36% to 90.77%. This demonstrates the potential of the pulsed discharge as a useful recovery method for precious metals, and further work will extend to the recovery of other deposited metals.

Acknowledgments

This work was supported by the JST-Mirai Program Grant Number JPMJMI19C7, Japan. We are also grateful to Tanaka Kikinzoku Kogyo Co., Ltd. Japan for providing samples and useful assistance for our research.

REFERENCES

Appendix

Fig. A1

The schematic illustration of (a) the circuit diagram, and (b) the placement of positive and negative electrodes of pulsed discharge at distances of ca. 5 mm and 10 mm from the top of the samples, respectively.

Table A1 The recovered particles of the catalyst converter for 120 pulse shots at a distance of 5 mm.

Table A2 The recovered particles of the catalyst converter for 120 pulse shots at a distance of 10 mm.

Table A3 The recovered particles of the catalyst converter for 240 pulse shots at a distance of 10 mm.

Fig. A2

SEM image and corresponding EDX spectra of (a) sample 1, (b) sample 2, and (c) sample 3.

Corresponding author

Register with J-STAGE for free!