Synthesis and Catalytic Performance of Graphene Modified CuO-ZnO-Al 2 O 3 for CO 2 Hydrogenation to Methanol

CuO-ZnO-Al 2 O 3 and graphene nanosheet (GNS) were synthesized by coprecipitation route and reduction of exfoliated graphite oxides method, respectively. GNS modified CuO-ZnO-Al 2 O 3 nanocomposites were synthesized by high energy ball milling method. The structure, morphology, and character of the synthesized materials were studied by BET, XRD, TEM, and H 2 -TPR. It was found that by high energy ball milling method the CuO-ZnO-Al 2 O 3 nanoparticles were uniformly dispersed on GNS surfaces. The catalytic performance for the methanol synthesis from CO 2 hydrogenation was also tested. It was shown experimentally that appropriate incorporation of GNS into the CuO-ZnO-Al 2 O 3 could significantly increase the catalyst activity formethanol synthesis. The 10wt.% GNS modified CuO-ZnO-Al 2 O 3 catalyst gave a methanol space time yield (STY) of 92.5% higher than that on the CuO-ZnO-Al 2 O 3 catalyst without GNS. The improved catalytic performance was attributed to the excellent promotion of GNS to dispersion of CuO and ZnO particles.


Introduction
The concentration of atmospheric CO 2 had risen from ∼280 ppm before 1860 to ∼390 ppm in 2010, which is further predicted to be ∼570 ppm by the end of the century [1].The increase in CO 2 emissions arguably causes to the increase in global temperatures and climate changes, which is called the "greenhouse effect." Hence, the comprehensive utilization of CO 2 has drawn more and more attention from the entire world.Because converting CO 2 to methanol not only could promote the efficient utilization of CO 2 but also has found a new way of methanol production, CO 2 hydrogenation to methanol has become a hot topic of academic research [2].But due to the fact that CO 2 is inert molecule and very difficult to be activated, so making an appropriate catalyst that could bring the enhancement of CO 2 conversion and methanol yield will be the key topic of this research.
A conventional CuO-ZnO-Al 2 O 3 catalyst for methanol production from synthesis gas (mixture of CO, CO 2 , and H 2 ) is now accepted [3][4][5].But it is not adequate for methanol synthesis from a H 2 /CO 2 mixture at temperatures below 250 ∘ C [6][7][8].Hence, several modifications have been used to improve conventional CuO-ZnO-Al 2 O 3 methanol synthesis catalysts.The key is to keep a high dispersion of the Cu/Zn crystallites where the active sites are as stated in [9,10].A lot of works have been done by doping other metal elements in CuO-ZnO-based catalyst for investigation, for example, Cr, Zr, V, Ce, Ti, Ga, and Pd [11][12][13].In addition, some works tried to find the suitable method to synthsis CuO-ZnO-Al 2 O 3 catalyst, such as sol-gel, hydrothermal, reverse microemulsion, and organic chelating agent to the common coprecipitation method [14,15].
In comparison to the other preparation techniques, high energy ball milling has been used as an effective and easy method for preparation of high dispersed and refining grains including nanocrystalline as well as amorphous materials [16].High energy ball milling process can change not only the size but also the structure of the material, so that makes the active species in the secondary distribution of the sample surface [17,18]; moreover, the particles undergo heavy pressure and mechanical deformation forming high-density fault structure and high concentration of lattice defects [19].As a result, it makes the catalytic activity significantly improves due to the dynamic performance of the catalyst increasing.
Graphene nanosheet (GNS) is considered an excellent catalytic carrier and additives owing to its unique physical chemical properties [20], such as, special electronic transmission characteristics, huge specific surface area, and excellent electrical and thermal conductivity.So far, the research of graphene in catalyst fields primarily appears to be on photoelectric catalysis, fuel cell, lithium ion battery, and so on.
In this paper, we doped GNS with CuO-ZnO-Al 2 O 3 composite oxides by high energy ball milling method.The prepared catalysts exhibit high dispersion of nanocomposite particles and high activity for CO 2 hydrogenation to methanol.

Experimental
2.1.Material Synthesis.Graphene nanosheet (GNS) was synthesized by reduction of graphite oxides.Firstly, we prepare graphite oxide by a modified Hummers method [21].Then, the prepared graphite oxide was dispersed in deionized water, followed by ultrasonic as graphene suspension and the hydrazine hydrate was added as the reducing agent.After that the mixture was refluxed at 100 ∘ C. In the end, GNS was obtained after filtering, washing, and drying the resulting mixture.
The mixed aqueous solution of Cu(NO 3 ) 2 , Zn(NO 3 ) 2 , and Al(NO 3 ) 3 (the mole ratio is 6 : 3 : 1) with a total concentration of 1 mol/L and the aqueous solution of 1 mol/L Na 2 CO 3 were prepared.The CuO-ZnO-Al 2 O 3 catalysts were prepared by concurrently dropping Na 2 CO 3 solution and mixed nitrates solution into an around-bottomed flask at 70 ∘ C under the water bath with vigorous stirring.The mixed nitrates solution was dropped at a rate of 5 mL/min and the pH was kept constant at 7-8.Then, the precursor was filtered and washed followed by dried overnight at 80 ∘ C. Finally the precursor was calcined at 350 ∘ C for 6 h to produce the CuO-ZnO-Al 2 O 3 catalysts.
In our experiment Fritsch 7 Pulverisette planetary ball mill was applied.The anhydrous ethanol was used as ball mill solvent and WC balls (Φ = 10 mm) were used as grinding balls.The milling rotational speed was 500 r/min.And ball mill could stop running for 15 min after working for 30 min.The cumulative milling time is 6 h.The mixture was dried at 60 ∘ C for 24 h after ball milling.The catalysts are denoted as CZA (without GNS), 5G-CZA (5 wt.% GNS), 10G-CZA (10 wt.% GNS), and 15G-CZA (15 wt.% GNS), respectively, in terms of graphene contents variations.

Material Characterization.
Panalytical X'Pert Pro polycrystal X-ray diffraction (XRD) was performed with Cu-K radiation,  = 0.15406 nm, tube voltage being 40 kV, and tube current being 40 mA in crystal structure of material characterization.The grain size of material is calculated by the Scherrer formula based on the strongest half peak width of the diffraction peak.The Brunauer-Emmett-Teller (BET) surface area of the catalyst was measured with an ASAP-2010 surface analyzer.The samples were degassed at approximately 200 ∘ C for 4 h before analysis.Hydrogen temperatureprogrammed reduction (H 2 -TPR) was conducted with a GC-7890II Gas chromatograph equipped with a thermal conductive detector (TCD) made from Shanghai Tianmei Company.The quality of the catalyst tested was 10 mg with the 5% H 2 -N 2 mixture gas as the reducing gas at a flow rate of 50 mL/min and heating rate of 10 ∘ C/min.Tecnai G2-F30 transmission electron microscopy (TEM) was used to test the morphology of catalyst with 300 kV accelerating voltage.

Catalytic Tests.
The activity tests were carried out in a stainless steel fixed bed microreactor.The catalyst (20-40 mesh size, 0.5 g) was packed into a stainless steel fixed bed microreactor (15 mm Φ) and reduced in a mixed ((H 2 )/(N 2 ) = 5/95) flow at 280 ∘ C for 6 h.After the reduction, the gas feed was switched to the reactant gas ((H 2 )/(CO 2 )/(N 2 ) = 69/23/8).The reaction was tested at a pressure of 3.0 MPa and temperature of 250 ∘ C with a space velocity of 12000 mL/h⋅g catal. .The CO 2 and CO were separated and regulated by carbon packed molecular sieve column and thermal conductivity detector (TCD).Methanol and other organic matters were separated and regulated by PoraPak Q column and hydrogen flame ionization detector (FID).

Structure Properties of Catalysts.
The XRD patterns of the pure CuO-ZnO-Al 2 O 3 samples and the CuO-ZnO-Al 2 O 3 doped by varying contents of GNS samples are compared in Figure 1.No matter how much is the amount of GNS added in the samples, the characteristic diffraction peaks of CuO appeared at 2 = 35.6 ∘ , 38.8 ∘ , and 48.9 ∘ and the ZnO special diffraction peaks appeared at 2 = 31.8∘ , 34.5 ∘ , and 36.3 ∘ in all the samples.The peak at 2 = 26 ∘ is the typical diffraction peak (002) of graphite material.With the added GNS increasing, the typical diffraction peak (002) intensity weakened a lot compared with the natural graphite.As part of the crystal structure of the incomplete oxidation of flake graphite damaged to some extent, so that its peak intensity is greatly decreased.
The physicochemical properties of CuO-ZnO-Al 2 O 3 doped with different amounts of GNS are summarized in Table 1.For comparison, the physicochemical properties of pure CuO-ZnO/Al 2 O 3 are also included.The surface areas of CZA, 5G-CZA, 10G-CZA, and 15G-CZA are 40.9,113.3, 147.8, and 186.3 m 2 /g, respectively.The surface areas of the samples increase with the increase of GNS doping amount and it proves that specific surface area of sample is proportional to the GNS addition.From Table 1, we could get the crystallite size data of CuO and ZnO calculated from their corresponding diffraction peak by Scherrer equation.And  the crystallite size of CuO particles decreased from 13.6 nm of CZA to 12.4 nm of 10G-CZA and the size of ZnO particles decreased from 18.3 nm to 14.0 nm.When milling time was constant, the minimum crystallite size of the samples is the sample powder with 10 wt.% GNS doped.For this result, there might be two reasons: (1) the CuO-ZnO/Al 2 O 3 particles adsorb on the surface of GNS, which reduce the surface energy of the particles and prevent the recombination between breaking chemical bonds, leading to the decrease of mechanical strength of the particles and the external force for dislocation extension and finally resulting in reinforcement of the crushing effect.(2) The compatibility of the powder with ethanol increases by the addition of GNS, which improves the rheological properties of the slurry.As a result, the slurry viscosity is decreased and the fluidity and dispersal ability of particles are promoted making the collision frequency between the ball mill and powder particle higher.However, for the sample with 15 wt.%GNS doped the crystallite size of CuO increased to 14.5 nm and the size of ZnO increased to 20.6 nm.This result illustrates the excessive addition of graphene reduced the collision frequency between the ball mill and particles, so that the particles were bigger for aggregating.Such a conclusion suggests that crystallite size is the smallest when GNS doped in the composite is 10 wt.% and the crushing grain effect is weakened over this doped amount.

TEM.
The morphology of CuO-ZnO/Al 2 O 3 doped with different amounts of GNS is shown in the TEM images in Figure 2. In Figure 2(a), the particle sizes are larger with obvious agglomeration of the pure CuO-ZnO-Al 2 O 3 catalysts which could be seen after ball milling.In Figures 2(b), 2(c), and 2(e), the transparent cicada's wing shape of nanographene sheets characteristic could be seen and the composite metal oxides particles are evenly distributed on the GNS.The stacks of graphene layers are clearly observed in the high-resolution image of Figure 2(d).Due to the fact that the GNS has a large specific surface area and the big  electron delocalization system, it is beneficial to the anchoring of the metal oxide nanoparticles and increases the electrostatic repulsion between the particles.GNS made by chemical reduction method has a certain amount of lattice defects and some residual oxygen functional groups making the distribution of the catalyst nanoparticles more uniform.The ball mill method made the metal oxide particles distributed evenly on GNS with the shape more regular and the size smaller.Figure 2(c) shows the CuO-ZnO-Al 2 O 3 catalysts doped with 10 wt.% GNS with the smallest size distributed on the GNS sheets most uniformly which is consistent with the change of the grain size of CuO and ZnO.

Reduction Properties of the Catalysts.
The H 2 -TPR patterns of CuO-ZnO/Al 2 O 3 doped with different amounts of GNS are given in Figure 3 and the four samples are all reduced in the temperature range of 150-290 ∘ C. For the pure CuO-ZnO/Al 2 O 3 catalyst, there is a big side-by-side peak which can be decomposed into two peaks noted as  and , respectively.Because ZnO and Al 2 O 3 are not reduced under the experimental conditions described here [22], so the two peaks appeared in the TPR profile stand for CuO existing in two different ways.The low temperature reduction peak ( peak) is assigned to reduction of CuO highly dispersed on the surface of catalysts and the high temperature reduction peak ( peak) is assigned to reduction of CuO in bulk phase.It is noticeable that there is a new small peak appearing at lower temperature in the H 2 -TPR patterns of other three samples doped with GNS.We name the lower temperature peak as  peak and assume that it is the contribution of GNS.GNS makes the shifts of reduction peak of the highly dispersed CuO toward lower temperature; that is, it makes part of CuO being more easily reduced by H 2 .This is because GNS has high adsorption and activation ability for hydrogen species [23] and implies that in the reaction GNS is not only a carrier but also a good promoter.In order to prove that the appearance of  peak has nothing to do with the reduction of GNS, we run a H 2 -TPR test in the same steps as before to pure GNS, getting a curve which is almost a straight line parallel to the baseline without any peaks.The areas of reduction peaks and their contributions to the TPR pattern over catalysts are given in Table 2.We can see that the proportion of areas of  and  peaks together is as high as 69.1% when the amounts of GNS doped is 10 wt.% indicating a higher content of highly dispersed CuO.The result is in accordance with the conclusion of aforementioned XRD, TEM characterization.

The Performance and Stability of the Catalysts.
The activity evaluation of CO 2 hydrogenation to methanol is carried out in the fixed bed microreactor and methanol and CO are the main carbonaceous products with a small amount of CH 4 .In Table 3, the copper dispersion ( Cu ) and active surface area (MSA) of the catalysts are showed by calculation from N 2 O chemisorption data.We could find that  Cu and MSA values are increased with the amount of the GNS increased till 10 wt.%, and then those are reduced when the GNS amount is up to 15 wt.%.As shown in Table 3, the CO 2 conversion and methanol space time yield (STY) are in accordance with  Cu and MSA.The 10G-CZA has the highest CO 2 conversion and STY of 14.6% and 360 mg/g⋅h, which are 71.8% and 92.5% higher than that on the CZA catalyst, respectively.These results suggest that the dispersion and particle size of active components of CuO and ZnO have a significant influence on the catalytic activity.The 10G-CZA catalyst, which has the most highly dispersion and the smallest particle size, is most conducive to the improvement of the catalytic performance.The result is in agreement with the XRD, TEM, and H 2 -TPR and other analysis conclusions.
Figure 4 shows the assay results of conversion and methanol selectivity at 230-280 ∘ C over the catalysts of CZA and 10G-CZA.The CO 2 conversion of the both catalysts went up as the temperature increased because appropriately raising the reaction temperature could promote the CO 2 reaction.On the other hand, both the methanol selectivity decreased with the temperature increased because the methanol formation from the hydrogenation of CO 2 is thermodynamically restricted within low conversion under the operating conditions [24].It can be showed that the CO 2 conversion and the methanol selectivity of 10G-CZA catalyst were always higher than that of the CZA catalyst, confirming that graphene is an excellent promoter for CuO-ZnO-Al 2 O 3 catalysts on CO 2 hydrogenation to methanol.
The 10G-CZA as the best catalyst performance was selected for the stability test, and the results are shown in Figure 5.The CO 2 conversion and methanol selectivity

Conclusions
The high energy ball milling method has been used to prepare GNS modified CuO-ZnO-Al 2 O 3 nanocomposites catalysts for methanol synthesis from CO 2 hydrogenation.GNS/CuO-ZnO-Al 2 O 3 nanocomposites catalyst with 10 wt.% GNS showed the best catalytic activity with CO 2 conversion and methanol space time yield is 71.8% and 92.5% higher separately than the catalyst without GNS addition.The GNS modified CuO-ZnO-Al 2 O 3 nanocomposites catalysts prepared by high energy ball milling method make the dispersion of active components of CuO and ZnO higher and smaller particle size and exhibit a much enhancement in catalytic performance.The ball mill operation is simple with low energy consumption, environmentally friendly, and suitable for large-scale preparation of industrialization.

Table 1 :
Physicochemical properties of CuO-ZnO-Al 2 O 3 doped with different amounts of GNS.

Table 2 :
Temperatures of reduction peaks and their contributions to the TPR pattern over catalysts. ,   , and   represent the areas of , , and  peaks, respectively.

Table 3 :
Catalytic properties and N 2 O chemisorption data of CuO-ZnO-Al 2 O 3 doped with different amounts of GNS.
2 -TPR patterns of CuO-ZnO-Al 2 O 3 doped with different amounts of GNS.