Efficient simultaneous removal of diesel particulate matter and hydrocarbons from diesel exhaust gas at low temperatures over Cu–CeO₂/Al₂O₃ coupling with dielectric barrier discharge plasma

The metal oxide catalysts Cu–CeO₂, CuO and CeO₂ were prepared using the ultrasonic aerosol pyrolysis (UAP) method (figure S1). A certain amount of copper and cerium nitrates (Shanghai Macklin, China) were dissolved in deionized water in which an ultrasonic atomization unit was installed. The atomized mists were heated to 70 °C and sent into a quartz tube reactor with N₂ (0.8 L min^–1) for thermal decomposition at 600 °C. The decomposition products were collected with a water absorption bottle and separated using a centrifuge (TG16-WS, Liangyou Instruments, China) operated at 10 000 rpm for 4 min. The solid precipitates were washed with deionized water four times and ethanol once, and dried at 80 °C for 10 h.

As a Cu/Ce atomic ratio of 1 has the best performance on diesel soot combustion [21], Cu–CeO₂ with an atomic Cu/Ce ratio of 1:1 was prepared by controlling the amounts of copper and cerium nitrates.

γ-Al₂O₃ balls (diameter between 1.0 and 1.6 mm; Jiuzhou Chemicals, China) were used as the catalyst carrier. Cu–CeO₂/Al₂O₃ was made by grinding Cu–CeO₂ powder with γ-Al₂O₃ balls (mass ratio 1:10) [33]. CuO/Al₂O₃ and CeO₂/Al₂O₃ were also prepared by mixing CuO and CeO₂ catalysts with γ-Al₂O₃ balls (mass ratio 1:10) using the same grinding method.

The oxidation properties of the DPM during heating from 25 °C to 800 °C at a heating rate of 10 °C min^–1 in an Ar-balanced 10% O₂ atmosphere were tested using a thermogravimetric (TG) analyzer (LABSYS evo, SETARAM Instrumentation, France) equipped with a mass spectrometer (LC-D200, Tilon, USA).

Catalyst characterization details are provided in the Supplementary Material.

Figure 1 shows a plasma catalysis experimental system for DPM removal. The system mainly consists of DPM generation, DBD reactor and detection units. The DPM was generated by burning diesel fuel (0#, China 6 standard, SINOPEC, China) in a diesel burner fed with 1.5 L min^–1 air [34]. The exhaust gas was stabilized with a buffer tank and supplied to the DBD reactor (figure S2). The tubular furnace, pulse power supply, voltage probe, current probe, oscilloscope and definitions of discharge power and energy density (ED) were as per the literature [35]. Typical discharge waveforms are illustrated in figure S3, showing that the voltage pulse had a rise time and pulse width of 40 µs and 50 µs, respectively.

Zoom In Zoom Out Reset image size

The DPM mass concentrations in the exhaust gas were detected using a PM counter (LX-600S, Jiangsu Lingxi, China) based on laser scattering; similar measurements were reported by Ni et al [36].

An ozone analyzer (BMOZ-200T, Shengxin, China) was used to detect the ozone (O₃) concentration in the exhaust gas. A flue gas analyzer (HPC500, Mincee, China) satisfying China standard JJG 688-2017 'Vehicle Exhaust Emissions Measuring Instruments' was used to detect the concentrations of hydrocarbons (HCs, equivalent to propane), CO, CO₂, O₂ and NO in the exhaust gas. The concentration data were automatically collected every 0.54 s.

DPM removal (X) was defined as

$$X=\frac{{C}_{\mathrm{i}\mathrm{n}}-{C}_{\mathrm{o}\mathrm{u}\mathrm{t}}}{{C}_{\mathrm{i}\mathrm{n}}}\times 100\% . \tag{ 1 }$$

Here, C_in and C_out are DPM concentrations in µg m^–3 at the inlet and outlet of the DBD reactor, respectively.

DPM samples were collected from the buffer tank (figure 1), and the oxidation properties in 10% O₂ atmosphere (Ar balance, 100 mL min^–1) were analyzed using TG-MS. The T₁₀, T₅₀ and T₉₀ values (the temperatures at which 10%, 50% and 90% of DPM were removed) are 290 °C, 628 °C and 700 °C, respectively (figure 2(a)), close to those of a DPM from a real diesel engine [37]. The main oxidation products of the DPM are CO₂ and CO (figure 2(b)), proving that the major component of DPM is carbon.

Zoom In Zoom Out Reset image size

3.2.1. Synergy effect on DPM and HC plasma catalysis oxidation

Figure 3(a) shows the DPM and HC removal efficiencies with different catalysts with and without plasma at 200 °C. When γ-Al₂O₃, CeO₂/Al₂O₃, CuO/Al₂O₃ and Cu–CeO₂/Al₂O₃ are used without plasma, DPM removals are 10%, 22%, 24% and 29%, respectively; HC removals are 12%, 24%, 26% and 30%, respectively. When coupling with plasma, DPM removals increase to 27%, 75%, 55% and 90%, respectively, while HC removals increase to 25%, 70%, 60% and 92%, respectively. The DPM and HC removals on CeO₂/Al₂O₃, CuO/Al₂O₃ and Cu–CeO₂/Al₂O₃ are higher than that on γ-Al₂O₃, implying CeO₂, CuO and Cu–CeO₂ can enhance oxidation of DPM and HCs. The DPM and HC removals are only improved by 18%‒19% using Cu–CeO₂/Al₂O₃ catalyst without plasma compared with γ-Al₂O₃, and by 63%‒67% with plasma, proving that the plasma coupled with Cu–CeO₂ catalyst can obviously enhance DPM and HC removals. The DPM and HC removal on Cu–CeO₂/Al₂O₃ are obviously higher than on CeO₂/Al₂O₃ and CuO/Al₂O₃, indicating that a synergistic effect for DPM catalytic oxidation can be obtained by doping Cu into CeO₂. DPM and HCs can be removed efficiently and simultaneously over Cu–CeO₂/Al₂O₃ catalyst coupled with DBD plasma at 200 °C.

Zoom In Zoom Out Reset image size

3.2.2. Effect of reaction temperature

3.2.3. Influence of ED

3.2.4. By-products of O₃ and NO_x

As oxidative removals of DPM and HCs are driven by DBD plasma, O₃ is simultaneously generated as a by-product; this has a strong oxidation potential and is involved in combustion enhancement [38]. When the ED is 9.8 J L^–1, the O₃ concentration decreases from about 60 ppm at 60 °C to 0 at 230 °C on Al₂O₃, and from about 15 ppm at 60 °C to 0 at 180 °C on Cu–CeO₂/Al₂O₃ (figure 4). These results show that O₃ decomposition efficiency on Cu–CeO₂ catalyst is higher than that on Al₂O₃, implying that Cu–CeO₂ catalyst can improve the interaction of O₃ with DPM [39]. Cu–CeO₂ catalyst can promote O₃ decomposition and dissociative adsorption (equations (2)‒(4)), resulting in the formation of various O_x ⁻ species and oxygen vacancies (V_O) on the catalyst surface and conversion of Ce⁴⁺ to Ce³⁺ [40, 41]. Ce⁴⁺ can be reduced to Ce⁴⁺ by DPM or HC (equation (5)). The Ce⁴⁺/Ce³⁺ redox cycles may be accelerated in the presence of copper, resulting in an enhancement of DPM oxidation.

Zoom In Zoom Out Reset image size

$${\mathrm{O}}_{3}+{[\mathrm{C}\mathrm{e}}^{4+}-{\mathrm{O}}^{2-}]\to {\mathrm{C}\mathrm{e}}^{4+}+{2\mathrm{O}}^-+{\mathrm{O}}_{2} \tag{ 2 }$$

$${4\mathrm{O}}_{3}+2{[\mathrm{C}\mathrm{e}}^{4+}-{\mathrm{O}}^{2-}]\to 2{\mathrm{C}\mathrm{e}}^{4+}+{4\mathrm{O}}^-+{3\mathrm{O}}_{2} \tag{ 3 }$$

$${2\mathrm{O}}_{3}+4{[\mathrm{C}\mathrm{e}}^{3+}-{\mathrm{V}}_{\mathrm{O}}]\to {4[\mathrm{C}\mathrm{e}}^{4+}-{\mathrm{O}}^{2-}]+{\mathrm{O}}_{2} \tag{ 4 }$$

$$2[\mathrm{C}\mathrm{e}^{4+}-\mathrm{O}^{2-}]+\mathrm{C}\mathrm{\ \ }\mathrm{o}\mathrm{r}\mathrm{\ \ }\mathrm{H}\mathrm{C}\to2[\mathrm{C}\mathrm{e}^{3+}-\mathrm{V}_{\mathrm{O}}]+\mathrm{ }\mathrm{C}\mathrm{O}_2 \tag{ 5 }$$

NO_x is a notable by-product from air plasma. Figure S5 shows the concentration of NO_x as a function of ED and temperature. At 200 °C, NO_x formation can be found at an ED higher than 20 J L^–1 (figure S5(a)). When ED is fixed at 9.8 J L^–1, the NO_x concentration increases significantly at temperatures above 225 °C (figure S5(b)). Temperature is the main factor contributing to NO_x formation when using DBD plasma. Special attention should be paid to NO_x generation under discharge conditions above 225 °C.

3.2.5. Long-term stability evaluation

Figure 5 shows the results of a long-term stability evaluation of DPM removal on Cu–CeO₂/Al₂O₃. DPM concentration was reduced gradually from 1550 µg m^–3 (inlet concentration) to 220 µg m^–3 (outlet concentration) after 25 min without plasma (figure 5(a)). The decrease in DPM concentration was due to the accumulation of DPM on the catalyst surface, which resulted in high pressure of the buffer tank upstream of the DBD reactor. DPM concentration decreased rapidly from 1580 µg m^–3 to 155 µg m^–3 with a stable time duration of around 10 h with plasma (figure 5(b)). Meanwhile, HC concentration also decreased from 190 ppm to 15 ppm with a 92% reduction, but the concentrations of O₂, CO₂, CO and NO did not change obviously (figures 5(c) and (d)). During the 10 h discharge duration, DPM and HC concentrations were kept at low levels. These results showed that the DBD reactor coupled with Cu–CeO₂/Al₂O₃ could simultaneously remove DPM and HC with a stable removal performance.

Zoom In Zoom Out Reset image size

3.3.1. Morphologies

Figures 6(a)–(c) show the morphologies of the catalysts. Each catalyst is mainly composed of microspheres with hollow cavities and uniform particle sizes of 1–2 µm. Catalysts with hollow nano- and microstructures can be prepared by spray techniques [42] due to the decomposition of nitrate precursor aerosols to nanoparticle aerosols and aggregation of the nanoparticle aerosols to microspheres with hollow cavities with or without rupture. Figure 6(d) shows the scanning electron microscopy (SEM) energy dispersive x-ray spectroscopy (SEM-EDS) mapping image of Cu–CeO₂ catalyst, which displays that each element of Cu–CeO₂ is evenly distributed on the catalyst. Figure S4 shows the morphology of the Cu–CeO₂ catalyst after reaction. Hollow microspheres of Cu–CeO₂ can be found with a diameter less than 0.5 µm, but many nanoparticles can also be observed. As the Cu–CeO₂ catalyst has a stable DPM and HC removal performance, the observed nanoparticles might be generated from hollow microspheres or Al₂O₃ due to grinding of the catalyst balls to powder before SEM analysis.

Zoom In Zoom Out Reset image size

Figure 7 reveals the high-resolution transmission electron microscope (HRTEM) images of the catalysts. Three kinds of lattice fringes of 0.27, 0.31 and 0.23 nm can be ascribed to CeO₂ (200), CeO₂ (111) and CuO (111), respectively [16]. The exposed (200) crystal facet of CeO₂ in Cu–CeO₂ can display better catalytic activities than other facets in soot oxidation [20].

Zoom In Zoom Out Reset image size

3.3.2. Crystal features

Table 1 lists the crystal features of CuO, CeO₂ and Cu–CeO₂ catalysts prepared using the UAP method. The specific surface area of Cu–CeO₂ catalyst is close to that of CeO₂ but higher than that of CuO, indicating that CeO₂ functions as the base structure for Cu doping.

Table 1. Crystal features of the catalysts.

Sample	Specific surface area (m2 g–1)	Pore volume (cm3 g–1)	Pore diameter (nm)	Cu/Ce molar ratio	Ce3+/(Ce3++Ce4+) ratio	Cu+/(Cu++Cu2+) ratio	Oads content (%)
CeO2	105.7	0.114	4.68		0.21		27
CuO	16.2	0.054	5.95			0.08	51
Cu–CeO2	78.4	0.093	8.51	0.85	0.32	0.46	67

The Cu/Ce atomic ratio in the bulk phase of Cu–CeO₂ catalyst from inductively coupled plasma optical emission spectroscopy analysis is 0.85, close to the 1 of the Cu/Ce atomic ratio in their nitrates.

The x-ray photoelectron spectra (figure S6) show that Ce³⁺, Ce⁴⁺, Cu⁺, Cu²⁺, lattice oxygen (O_latt) and surface oxygen (O_ads) are present on the surface of the Cu–CeO₂ catalyst. The Cu–CeO₂ catalyst has a higher Ce³⁺/(Ce³⁺+Ce⁴⁺) ratio, Cu⁺/(Cu⁺+Cu²⁺) ratio and O_ads content than CuO and CeO₂; these are due to the synergistic effect between Cu⁺/Cu²⁺ and Ce³⁺/Ce⁴⁺ [43]. The higher ratios of Ce³⁺/(Ce³⁺+Ce⁴⁺) and Cu⁺/(Cu⁺+Cu²⁺) are conducive to oxygen vacancy generation on Cu–CeO₂ catalyst, and oxygen vacancies are also the active center for O₃ adsorption and decomposition [44]. The higher O_ads content is beneficial for catalytic DPM oxidation [45]. Thus, Cu–CeO₂ catalyst has better catalytic performance for DPM and HC removals than CuO and CeO₂.

Figure 8 describes the x-ray diffraction (XRD) patterns of CeO₂, Cu, and Cu–CeO₂ catalysts. The characteristic peaks of CeO₂ at 2θ of 28.6°, 33.1°, 47.5° and 56.3° are attributed to (111), (200), (220) and (311) planes, respectively [46]. Cu–CeO₂ has diffraction peaks mostly similar to CeO₂, and the peaks at 28.7°, 33.3°, 47.9° and 56.6° shift slightly to higher 2θ angles than those of CeO₂, suggesting that the lattice parameters decrease due to Cu²⁺ doping into the CeO₂ crystal lattice (ion radius Cu²⁺ < Ce⁴⁺) [16, 17]. Thus, Cu is successfully doped into CeO₂. In addition, Cu–CeO₂ has peaks at 2θ of 35.5° and 38.7° which are associated with CuO (002) and (111), indicating that CuO nanoparticles were generated and not all Cu is doped into the CeO₂ lattice. Cu–CeO₂ has a polycrystalline structure with components of Cu-doped CeO₂ and CuO nanoparticles, resulting in more active sites.

Zoom In Zoom Out Reset image size

Operando DRIFTS-MS was employed to investigate surface functional groups and gaseous products during DPM oxidation at 200 °C by mixing DPM with no catalyst (no cat./DPM), CuO (CuO/DPM), CeO₂ (CeO₂/DPM) and Cu–CeO₂ (Cu–CeO₂/DPM) without and with plasma.

During DPM oxidation without catalyst (figure 9(a)), no obvious spectral signal was found in the first 60 min without plasma, and some peaks appeared in the next 60 min with plasma. The peaks at about 3150, 2340, 1750, 1390 and 1050 cm⁻¹ are ascribable to ν(O–H) in hydroxyl groups, ν(CO₂), ν(C=O) in carbonyls group, ν_s (COO) and ν(C–O) in carboxylic acids and ethers, respectively [46, 47]. The adsorbed CO₂ can be gasified during plasma DPM oxidation, so its peak is negative. The heights of all the peaks increase with increasing discharge time, indicating accumulation of the surface species. The peak height variations of typical oxygen-containing groups C=O at ~1750 cm⁻¹ and C–O at 1050 cm⁻¹ are plotted in figure 10(a). The peak heights of surface carbon oxides (C=O and C‒O) are near zero without plasma (1‒60 min), indicating that the surface oxygen groups on DPM are little changed without plasma. With plasma (61‒120 min), they increase along with time, and the CO₂ MS signal intensity also becomes higher (figure S7), suggesting that the DPM is easily oxidized to surface carbon oxides and CO₂ [26, 48]. The burn-off of the nano-sized DPM particles results in a decrease in DPM particle number [49] and formation of more CO₂.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

During DPM oxidation on CuO (figure 9(b)), no strong peaks emerged in the first 60 min without plasma, indicating that little DPM is oxidized. With plasma in the next 60 min several peaks appeared. The peaks at 1540, 1380, 1060 and 835 cm⁻¹ are assigned to bidentate carbonate (b-${\mathrm{CO}}^{2-}_3$), monodentate carbonate (m-${\mathrm{CO}}^{2-}_3$), adsorbed oxygen (M⁺–${\mathrm{O}}^-_2$) (M denotes Cu) and bridging oxygen (M–O–M), respectively [50–52]. The peak heights of M–O–M and M⁺–${\mathrm{O}}^-_2$ increased with discharge time, indicating that active oxygen from plasma could be adsorbed on CuO and transformed into M–O–M and M⁺–${\mathrm{O}}^-_2$. The peak of m-${\mathrm{CO}}^{2-}_3$ is extremely prominent when with plasma (61‒120 min), but there are no strong peaks within 3800‒3000 cm⁻¹ (assigned to OH), indicating that hydroxyl and carboxyl groups could be converted into m-${\mathrm{CO}}^{2-}_3$. The peak heights of carbonates (m-${\mathrm{CO}}^{2-}_3$ and b-${\mathrm{CO}}^{2-}_3$) are plotted in figure 10(b) as functions of reaction time. The m-${\mathrm{CO}}^{2-}_3$ peak height increases with time, meaning that plasma can promote [m-${\mathrm{CO}}^{2-}_3$] accumulation, but has no effect on b-${\mathrm{CO}}^{2-}_3$ accumulation on CuO. The main surface oxygen complex (SOC) on CuO during DPM plasma oxidation is m-${\mathrm{CO}}^{2-}_3$.

During DPM oxidation on CeO₂ (figure 9(c)) peaks at 3200, 2340, 1650, 1460 and 1380 cm⁻¹ are found which, respectively, correspond to OH, CO₂, C=O, b-${\mathrm{CO}}^{2-}_3$ and m-${\mathrm{CO}}^{2-}_3$ species [53]. The peak heights of two carbonates (m-${\mathrm{CO}}^{2-}_3$ and b-${\mathrm{CO}}^{2-}_3$) as functions of reaction time are plotted in figure 10(c). Without plasma (1–60 min), the heights of the two peaks (m-${\mathrm{CO}}^{2-}_3$ and b-${\mathrm{CO}}^{2-}_3$) did not change, indicating that no DPM is oxidized. With plasma (61‒120 min), more carbonates are generated and subsequently decomposed to gaseous CO₂ (figure 10(c)) as their peak heights increase first and then decrease.

For DPM oxidation on Cu–CeO₂, peaks were found at 3240, 2340, 1700, 1620, 1490, 1360, 1060, 840 and 750 cm⁻¹ (figure 9(d)), which respectively correspond to OH, CO₂, C=O, ${\mathrm{HCO}}^-_3$, b-${\mathrm{CO}}^{2-}_3$, m-${\mathrm{CO}}^{2-}_3$, M⁺‒${\mathrm{O}}^-_2$ (M denotes Cu or Ce), terminal oxygen (M=O) and M‒O‒M [53–55]. Without plasma (1‒60 min), the peak patterns gradually appear, indicating that the activity of Cu–CeO₂ is superior to CuO and CeO₂, but little DPM was oxidized as the peaks were weak. With plasma (61‒120 min), the increases in the heights of M=O and M‒O‒M peaks are due to the increasingly stronger peak of M⁺‒${\mathrm{O}}^-_2$ because more active oxygen could be generated with plasma and adsorbed by metal oxides. The peak heights of SOCs (m-${\mathrm{CO}}^{2-}_3$, b-${\mathrm{CO}}^{2-}_3$ and ${\mathrm{HCO}}^-_3$) raised to a certain level and were maintained (figure 10(d)), indicating that their generation and consumption were balanced. The MS analysis result (figure S7) also indicates that the SOCs on Cu–CeO₂ can easily convert to gaseous CO₂. This fact implies that HC${\mathrm{O}}^-_3$ and b-${\mathrm{CO}}^{2-}_3$ are easier to decompose to gaseous CO₂.

The mechanism of DPM oxidation enhanced by plasma and Cu–CeO₂ is illustrated in figure 11. The main surface intermediates for DPM oxidation are b-${\mathrm{CO}}^{2-}_3$ and m-${\mathrm{CO}}^{2-}_3$. Although CuO and CeO₂ catalysts can promote DPM plasma oxidation, the formation of surface m-${\mathrm{CO}}^{2-}_3$ carbonate (which is difficult to convert to gaseous CO₂) inhibits the catalytic effect. Cu-doped CeO₂ catalyst can greatly improve DPM oxidation because the surface reactive oxygen species [O] (M⁺‒${\mathrm{O}}^-_2$ and M=O) can participate in DPM oxidation and enhance gasification of SOCs to CO₂.

Zoom In Zoom Out Reset image size