One‐Step Reforming of CO2 and CH4 into High‐Value Liquid Chemicals and Fuels at Room Temperature by Plasma‐Driven Catalysis

Abstract The conversion of CO2 with CH4 into liquid fuels and chemicals in a single‐step catalytic process that bypasses the production of syngas remains a challenge. In this study, liquid fuels and chemicals (e.g., acetic acid, methanol, ethanol, and formaldehyde) were synthesized in a one‐step process from CO2 and CH4 at room temperature (30 °C) and atmospheric pressure for the first time by using a novel plasma reactor with a water electrode. The total selectivity to oxygenates was approximately 50–60 %, with acetic acid being the major component at 40.2 % selectivity, the highest value reported for acetic acid thus far. Interestingly, the direct plasma synthesis of acetic acid from CH4 and CO2 is an ideal reaction with 100 % atom economy, but it is almost impossible by thermal catalysis owing to the significant thermodynamic barrier. The combination of plasma and catalyst in this process shows great potential for manipulating the distribution of liquid chemical products in a given process.


Experimental
The experiment was carried out in a coaxial dielectric barrier discharge (DBD) reactor with a novel water electrode at atmospheric pressure and room temperature (Scheme S1). The DBD reactor consisted of a pair of coaxial glass cylinders (inner and outer glass tubes) and two coaxial electrodes. The inner high-voltage electrode was a stainless-steel rod with an outer diameter (o.d.) of 2 mm, installed along the axis of the inner glass tube (10 mm o.d. × 8 mm i.d.), which also served as the dielectric material. Compared to conventional cylindrical DBD reactor design, circulating water filled the space between the inner and outer glass cylinders and acted as a ground water electrode. This novel reactor design using the water electrode could effectively remove heat generated by the discharge and maintain the reaction at around room temperature (~30 o C) for the effective synthesis of liquid oxygenates at atmospheric pressure. The discharge length was 45 mm with a discharge gap of 3 mm. The catalyst was packed into the discharge area. The flow rate of CH 4 and CO 2 was controlled by mass flow controllers with a total feed flow rate of 40 ml/min. The DBD reactor was connected to an AC high voltage power supply with a maximum peak voltage of 30 kV and a variable frequency of 7-12 kHz. In this work, the frequency was fixed at 9 kHz. The electrical signals (applied voltage, current and voltage on the external capacitor) were recorded by a four-channel digital oscilloscope (Tektronix, MDO 3024). The discharge power was calculated using the Lissajous method and was fixed at 10 W in this work [1] .
The dry reforming of methane (DRM) reaction was carried out at the same temperature (~30 o C) under three different operating modes: plasma-alone, catalysis-alone and plasma-catalysis. In the catalysisalone mode, no reaction occurred at a temperature of around 30 o C.
The gaseous products were analyzed using a gas chromatograph (Shimadzu GC-2014) equipped with a thermal conductivity detector (TCD) and a flame ionized detector. A water/ice bath (0 o C) was placed at the exit of the reactor to condense liquid products. The oxygenates were qualitatively analyzed using a gas chromatography-mass spectrometer (GC-MS, Agilent GC 7820A and Agilent MSD 5973) and quantitatively analyzed using a gas chromatograph (Agilent 7820) equipped with a FID with a DB-WAX column. The change of the gas volume before and after the reaction was measured using a soap-film flowmeter (Scheme S1).
The emission spectra of the CH 4 /CO 2 DBD were recorded using a Princeton Instruments ICCD spectrometer (SP 2758) in the range of 200-1200 nm via an optical fiber, placed near the ground electrode of the DBD reactor. The slit width of the spectrometer was fixed at 20 µm. A 300 g mm -1 grating was used. For comparison, the spectra of the DBD using pure CH 4 and pure CO 2 were also recorded.
Measurement of plasma reaction temperature: an infrared camera (FLIR A 40) was focused on a 9×9 mm window of the DBD reactor to measure the temperature in the discharge area, as shown in Scheme S2 (a). The results show that the reaction temperature in the discharge area was around 30 o C, as shown in SP1, SP2, SP3 and SP4 of Scheme S2 (b), while the temperature of the inner highvoltage electrode was slightly higher (~57 o C). A fiber optical thermal meter (OMEGA, FOB102) was also used to measure the temperature in the discharge area by placing the optical fiber in the discharge area. The measured temperature of the discharge area was also around 30 o C. Scheme S1. Schematic diagram of experimental setup. Scheme S2. Measurement of reaction temperature in the discharge area by an infrared camera.
To evaluate the performance of the dry reforming reaction, the concentration of each product in the condensate was calculated via corresponding formula of standard calibrated concentration curve (Table S1). (y denotes as concentration of sample, mol/L; x denotes as GC peak area of sample).
The conversion of CH 4 and CO 2 is defined as: (1) The selectivity of gaseous products can be calculated: Note that the change of the gas volume before and after the reaction was taken into account in the calculation of above parameters.
The selectivity of the liquid products can be calculated according to: The total selectivity of liquid products % = 100% − S #9 + S # X $ Y − ca. 10% carbon deposition The selectivity of C x H y O z can be calculated:

Catalyst preparation
All the catalysts were synthesized by incipient wetness impregnation over as-is commercially obtained γ-Al 2 O 3 (Dalian Luming Nanometer Material Co., Ltd.) and as-synthesized TS-1 using a hydrothermal method. Metal precursor solution was prepared by dissolving each metal salt in water, which is just sufficient to fill the pores of 8 g of the corresponding support. The supports were first calcined to remove the impurities (e.g., adsorbed H 2 O) in a muffle furnace at 400 o C for 5 h, then the support was added to the as-prepared precursor solution and was stirred until it was thoroughly mixed. The resulting mixture was successively kept at room temperature for 3 h, vacuum freeze-dried overnight at -50 o C and dried in air at 120 o C for 5 h. The dried sample was finally calcined in an Ar-DBD plasma at 350 o C for 3 h. Metal loading amounts of noble (Pt and Au) and non-noble metal (Cu) catalysts were ca. 1 wt.% and ca. 15 wt.%, respectively.

Catalyst characterization
The acidity of the supports was evaluated by NH 3 temperature-programmed desorption (NH 3 -TPD) using a Quantachrome ChemBET 3000 Chemisorption instrument. The sample (140 mg) was pretreated at 600 o C for 1 h in a He flow (20 ml/min) and then cooled to 150 o C. The pre-treated sample was saturated with NH 3 for 30 min, and then purged with a He flow for 1 h at 150 o C. The TPD profile was recorded, heating the sample from 150 to 600 o C at a constant heating rate of 14 o C/min in a He flow.
The metal-support interaction was studied by H 2 temperature-programmed reduction (H 2 -TPR) using the same instrument as NH 3 -TPD. The sample (100 mg) was pretreated at 500 o C for 1 h in a He flow (20 ml/min), and then cooled to 50 o C. The pre-treated sample was exposed to a H 2 /He mixture (10 vol.% H 2 ) and was heated from 150 to 800 o C at a constant heating rate of 14 o C/min to create the TPR profile.
N 2 physisorption was performed at 77 K using a Micrometrics TriStar 2020 instrument. Prior to the N 2 physisorption measurements, the samples were degassed at 350 o C for 3 h. The specific surface area of the samples was calculated using the Brunauer-Emmett-Teller (BET) equation.
X-ray diffraction (XRD) patterns were collected using a Rigaku D-Max 2400 X-ray diffractometer with Cu K α radiation. Transmission electron microscopy (TEM) was used to characterize the formation of metal particles on the catalyst surface using a JEOL 2010 with EDS of Oxford Instruments INCA energy system at an accelerating voltage of 200 kV.

Results and Discussion
In this study, we found that packing the catalysts into the discharge slightly decreased the conversion of CO 2 and CH 4 . Packing the catalyst pellets into the entire discharge area was found to change the discharge mode from a typical strong filamentary microdischarge in the gas phase to a combination of weak spatially-limited microdischarge and a predominant surface discharge on the catalyst, as shown in Figure S1. Similar phenomenon has also been found in our previous studies [2] . This physical effect (e.g. weak microdischarges) induced by the presence of the catalyst could affect the DRM reaction and lead to the decreased conversion of CH 4 and CO 2 . Similar negative effect from the integration of plasma and catalysts has also been reported from other groups [3] .

Plasma Simulation
The simulation employs a 0-dimension time-evaluated model using ZDplaskin. We assumed that no surface reactions and recirculation occurred in the DBD reactor so that all species could satisfy the conditions for solving the Boltzmann Equation (BE) (8).
Where E is the electric field，m is the electron mass, f is the electron energy distribution function (EEDF), q is the elementary charge, v is the average electron velocity and C[f] represents the change rate of the EEDF. Our model can be classified into three main blocks; 263 electron-impact reactions, including momentum transfer, excitations/de-excitations, dissociation and ionisation reactions; 348 neutral-neutral reactions; 65 ion-neutral/radical/ion reactions. All simulated species are shown in table S4. Table S4. Summary of all species included in the model.
The time evolution density of all species, N i=1…imax , can be written as equation (9). The source term Q ij is used to describe the contribution from each diverse reaction, j =1…j max , and it is defined by user's input file.
In order to provide a better understanding of our physical model, we use reaction (10) as an example. The reaction rate of this reaction can be calculated using equation (11). Therefore, the source terms will be expressed as equation (12).
Additionally, calculations of the rate constants, k j (cm 3 /s), are different if electrons are taken into account. For the neutral-neutral reactions, the rate constants can be obtained from the three-parameter Arrhenius form (13) T is gas temperature in Kelvin. The three parameters of A j , B j and E j represent pre-exponential factor, temperature factor, and activation energy respectively. All parameters used in this work can be found from NIST database.
However, for the electron-impact reactions, a special range of E/n was used to solve BE (8) to obtain the electron distribution function and mean electron temperature, while the rate constants for all electron-impact reactions were calculated using equation (14).
Where σ k is the cross-section of the target particle, F represents the EEDFs, and ɛ (ɛ =v/G) is the electron energy in volt (G = 2 / ).  Figure S3. Temporal dynamic densities of vibrational and electronic excited CO 2 in two AC cycles (discharge frequency 9 kHz, The value of E/n is up to around 120 Td, and the electron energy distribution function (mean electron energy) is calculated using the Boltzmann Equation [10] . All the cross section data used for solving BE was from the LXCAT database (Morgan database), http://www.lxcat.net, retrieved on June 12, 2016). Figure S4. Effect of CH 4 /CO 2 molar ratio on the relative intensity of different species in the CH 4 /CO 2 discharge (CH 431.4 nm, H 656.3 nm, C 2 516.5 nm, CO 519.4 nm, O 844.7 nm, total flow rate 40 ml/min, discharge power 10 W, discharge frequency 9 kHz, 2 s exposure time). Scheme S3. Possible reaction mechanisms for the formation of CH 3 COOH, CH 3 OH, C 2 H 5 OH and HCHO using the plasma-catalysis approach.
The physicochemical properties of the support and catalysts were analyzed by means of N 2physisorption, NH 3 -TPD, XRD, H 2 -TPR and TEM ( Figure S5-S8) to understand the different reaction performance of γ-Al 2 O 3 supported Cu, Au and Pt catalysts ( Figure 1). The γ-Al 2 O 3 support had a specific surface area of 114.8 m 2 /g and plenty of acid sites ( Figure S5); Cu, Au and Pt were highly dispersed on the surface of γ-Al 2 O 3 with an average nanoparticle size of approx. 10 nm, 5 nm, and 5 nm, respectively ( Figure S6); Cu existed in the form of CuO over γ-Al 2 O 3 support with a reduction temperature in the range of 150-300 o C ( Figure S7 and S8); Metallic Au (Au 0 ) and Au δ+ coexisted in the Au/γ-Al 2 O 3 catalyst, as indicated by XRD and the TPR profile of the Au/γ-Al 2 O 3 with reduction peak in the range of 310-400 o C ( Figure S7 and S8); No diffraction peaks of Pt were observed on the XRD profile of the Pt/γ-Al 2 O 3 , caused by a high dispersion and small particle size for the Pt/γ-Al 2 O 3 catalyst as confirmed by TEM ( Figure S6 and S7).