A new method to improve homogeneity and oxidation stability of Cu nanoparticles for lubricant additive in liquid phase reduction process

In the preparation of Cu nanoparticles by liquid phase reduction, the traditional reagent mixing mode was changed to improve the homogeneity in Cu nanoparticle size. In addition, Cu nanoparticles were collected by adding the volatile organic solvent (benzine) to the reaction system to prevent the oxidation of them. XRD and SEM analysis confirmed that the reaction products were spherical Cu particles with relatively uniform size of 30–50 nm. EDS analysis showed that the oxygen content in the product was 36.2% for using a centrifuge and 12.6% for using benzine, indicating that the oxidation of Cu nanoparticles was relatively prevented in the preparation process. The lubricant containing Cu nanoparticles of 0.15 wt% exhibited decreased friction coefficient by 60% compared the pure lubricant. Moreover, the decreasing tendency of friction coefficient of the Cu nanoparticles prepared by the proposed method was more stable than that of Cu nanoparticles prepared by the common liquid phase reduction process.


Introduction
Copper, as a soft metal, has lower shear stress, good tribological properties. So, Cu nanoparticles have been attracted the interest as an oil additive with excellent load carrying capacity, anti-wear and friction-reduction properties [1][2][3][4].
There are several synthesis methods of nanomaterials including hydrothermal method [5], thermal decomposition [6], liquid phase reduction [7][8][9][10][11] and biological synthesis methods [12,13]. Among them, the liquid phase reduction is widely adopted because the size and morphology of Cu nanoparticles are easily controlled by varying metal salts, solvents, reducing agents, reduction temperature, concentrations and other parameters [14]. Cu nanoparticles were synthesized at different pH levels using the aqueous reduction method and reducing agents such as hydrazine hydrate, ascorbic acid and capping agents such as cetyltrimethylammonium bromide (CTAB) and polyvinyl pyrrolidone (PVP). The as-produced Cu nanoparticles have nonuniform size and morphology, and the level of oxidation also varies [15].
Khorasani et al [16] and Ayesha et al [17] prepared Cu powder of size 1 μm by liquid phase reduction with ascorbic acid as reduction agent. In [18,19], Cu powder was obtained by direct reduction of CuSO 4 solution with formaldehyde. The prepared Cu powder possessed poor particle size homogeneity and less than 100 nm in size, because of the rapid nucleation rate and very short growth time. Sulekh et al [20] reported that Cu powder with 50-100 nm was produced using hydrazine as reduction agent and gelatin as capping agent at the temperature of 70°C. But hydrazine is strong reduction agent and toxic. In [21,22], Cu nanoparticles were fabricated by the liquid phase reduction of CuSO 4 with KBH 4 . To decrease the reduction rate of Cu ions, the complex solution of Cu 2+ was formed by adding EDTA into CuSO 4 solution, and then the reduction reaction was performed by the drop-wise of KBH 4 alkali solution at the rate of 2 drops per second with stirring.
In the common liquid phase reduction process, the reducing agents and capping agents were added dropwise to copper salt solution under reduction process condition (pH, temperature). Then, nanoparticles were obtained through the reducing reaction, followed by centrifugal separation, washing with absolute alcohol and Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. drying in vacuum. Many reports have focused on optimization of several process parameters (for example pH, reaction temperature, reaction time, selection of reagents etc), however, the research on performance improvement of Cu nanoparticles by the process control is scant.
As known well, it is primarily important to control the nuclei formation-growth process of crystal in reduction process. The nuclei formation process had to be occurred homogeneously and simultaneously in all reaction system to improve the homogeneity of Cu nanoparticles. Even in case of mixing by the drop-wise addition, crystal nuclei formation-growth process could be occurred rapidly due to local supersaturating caused by rapid reduction of copper ions, resulting to the generation of inhomogeneous nanoparticles. In case of the collection by a centrifuge, the oxidation of Cu nanoparticles becomes unavoidable by contact with air. Also, it couldn't be prevented to loss some Cu nanoparticles during the collecting procedure.
The main objective of this study is to improve the homogeneity in size of Cu nanoparticles by controlling the crystal nuclei formation-growth process and to prevent their oxidization by simplifying the collecting procedure using volatile solvent addition. In addition, experimental verification on the lubricating effect of the prepared Cu nanoparticles as a lubricant additive was carried out.

Chemicals and instruments
Copper sulfate (CuSO 4 ·5H 2 O), sodium hypophosphite (NaH 2 PO 2 ·H 2 O), disodium ethylenediamine tetra acetate (C 10 H 14 N 2 O 8 Na 2 , EDTA), sodium dodecyl sulfate (SDS) and sodium dodecyl benzene sulfonate (SDBS) are all chemical reagent grade. Deionized water was used as solvent throughout all of the experiments. X-ray diffraction (XRD) (X'Pert PRO), scanning electronic microscope (SEM) equipped with energy dispersive spectrometry (EDS) (ULTRA 55 Plus, Germany) and Brunauer-Emmett-Teller (BET) surface analyser (JW-BK 122 W) were used to characterize the size and morphology of the prepared Cu nanoparticles. Friction and wear tests on the lubricating properties of the lubricant (Total Quarts SAE 5W-30) with Cu nanoparticles were performed in a friction and wear tester (MMW-1, China) (figure 1). Friction material is structural carbon steel 1045, and the hardness of specimens are HB 190 and HB 210. The test temperature, load and rotating speed are 300 N, 80°C and 800 r min −1 , respectively. Friction coefficients were recorded automatically and displayed in graph.

Experiment
First, 256 ml of 0.0715 mol l −1 CuSO 4 was prepared by adding cupric sulfate hydrate into deionized water. 0.42 g of EDTA was added into the cupric sulfate solution to stir till the color of solution changed from blue to pale blue. Next, 0.2 g of SDS and 0.22 g of SDBS were added and stirred to ensure the complete dispersion. 24 ml of 1.032 mol l −1 NaH 2 PO 2 was added into this solution, the mixed solution was stirred for 30 min following by adding H 2 SO 4 to adjust pH of the solvent to 1.5. CuSO 4 did not reduced by sodium hypophosphite at the temperature below 40°C. The mixed solution was heated with stirring at a rotation speed of 120 r min −1 . When the reaction temperature was raised to 40°C, 10 ml of benzine was added into the mixed solution. The solution temperature was kept to 50°C-55°C and stirred for 30 min. As the reaction proceeded, the color of the solution turned slowly from blue to green, purplish red (figure 2).
This color change indicates the formation of Cu nanoparticles by the reduction process. The following chemical reactions took place during the liquid reduction synthesis: After 30 min of reaction, reaction solution was cooled to room temperature with stirring the solution. By difference in specific gravity, reaction solution was delaminated into two: upper part -benzine containing Cu nanoparticles and lower part-aqueous solution ( figure 3).
The aqueous solution of the lower layer was removed using a tap funnel, added and shaken again deionized water, and leaving it alone. When the reaction solution was delaminated into two layers, the aqueous solution was removed again. This procedure was repeated 3 times. And then the benzine containing Cu nanoparticles was dried in a vacuum oven. Finally, benzine was volatilized and Cu nanoparticles were obtained. In this way, Cu nanoparticles were separated and collected from aqueous solution.
For comparison, Cu nanoparticles were prepared by the common liquid phase reduction using the drop wise addition. First, CuSO 4 solution was prepared by adding cupric sulfate hydrate into deionized water. EDTA was added into the CuSO 4 solution and stirred till the color of solution changed from blue to pale blue. 0.2 g of SDS and 0.22 g of SDBS were added and stirred to ensure the complete dispersion. Next, this solution was heated to 55°C with stirring at a rotation speed of 120 r min −1 . When the reaction temperature was raised to 55°C, 24 ml of 1.032 mol l −1 NaH 2 PO 2 solution was added drop-wise into the mixed solution with stirring for 30 min. Then, the Cu nanoparticles were collected by a centrifuge, washed with deionized water for 3 times and dried in a vacuum oven for 24 h at 60°C.

Result and discussion
As-prepared samples were analyzed by XRD and SEM. Figure 4 displays the XRD patterns of the metallic Cu nanoparticles. The measurement was conducted from 30°to 85°in angular range at room temperature. The patterns showed peaks at 43.3°(111), 50.3°(200) and 74.2°(220), which were attributed to cubic metallic Cu (JCPDS Card No. 4-0836). All peaks in the XRD patterns correspond to Cu and CuO. It proves that product was composed with Cu (91.9%) and CuO (8.1%) ( figure 3). The average size of the synthesized Cu nanoparticles was calculated using the Debye-Scherrer's formula [5]: where d denotes the moderate crystallite size (nm) and λ is the X-ray wavelength (1.5405 Å). β and θ are the full width at half maximum of the diffraction peak and half diffraction angle (degree), respectively. Scherrer's formula (equation 2) suggests that the average size of the synthesized Cu nanoparticles is about 30 nm.
Meanwhile, according to BET analysis, the specific surface area of the synthesized Cu nanoparticles is about 13.4 m 2 g −1 and the corresponding average equivalent size was calculated to be about 50 nm. This indicates that the proposed chemical reduction method enables the preparation of Cu nanoparticles with a size of 30-50 nm.
SEM images demonstrated that Cu nanoparticles collected using benzine were finer and more homogeneous spherical particles than that collected using a centrifuge ( figure 5). It was also confirmed by DLS result ( figure 6).
The EDX results confirmed the existence of 12.6% O in Cu nanoparticles collected using benzine, it decreased significantly in comparison with 36.2% O in Cu nanoparticles collected using a centrifuge (table 1). This indicates that in case of using organic solvent such as benzine, the oxidization stability of Cu nanoparticles increased significantly compared with using a centrifuge.  In this paper, Cu nanoparticles were prepared by mixing copper salt and reduction agent and heating at the room temperature, in the stead of conventional mixing method (drop-wise addition). Previous experiments confirmed that cupric sulfate did not reduced by sodium hypophosphite at the temperature below 40°C.
Mixing copper salt solution with reduction agent by drop-wise procedure had affect adversely on crystal nuclei formation. From the experiment results, the reduction rate of copper ions were very fast and the reduction reaction was completed before the addition of reduction agent, therefore, crystal nuclei were formed by the local super saturation. Some of Cu atoms reduced freshly had adhere to the existing nuclei and grew and another came to be formed into new nuclei. It is because crystal nuclei had already formed but the reduction agent had continued to add into the reaction system. Consequently, crystal nuclei had not be formed uniformly and the nuclei formation-growth process did not proceeded in sequence, and at last the as-prepared Cu nanoparticles had become inhomogeneous in size.
When all the reagents are already mixed below the reaction temperature, the copper ions were uniformly reduced throughout the reaction space and the concentration distribution of copper atoms became homogeneous. At that time, enormous crystal nuclei are formed simultaneously and homogeneously due to the supersaturation in concentration throughout the system. So, the nuclei formation-growth process proceeds in sequence, resulted in the homogeneous size of Cu nanoparticles.
In liquid phase reduction method, Cu nanoparticles are obtained in the state of dispersion. Therefore, they have to be separated and collected from the aqueous solution and this is realized by a centrifuge in the common liquid phase reduction method. Cu nanoparticles are segregated from the solution by strong centrifugal force and adhered to the vessel wall. Then, the water in the vessel was removed. Cu nanoparticles were collected by scraping at the wall after washing of 3 times in this way. During the repeated washings including several times of centrifugation, Cu nanoparticles are exposed to the atmosphere and further oxidized by oxygen in ambient. The centrifugation procedure have to repeated several times according to the vessel volume of a centrifuge. As Cu nanoparticles adhere to inner surface of vessels, the more the procedure, the more the loss of Cu nanoparticles. Moreover, use of a centrifuge results in the increase of cost and oxidization, so, this is not suitable for mass production of Cu nanoparticles for lubricant additive.
When benzine is added into the reaction system, Cu nanoparticles are collected into benzine by the emulsion effect of SDS and SDBS of capping agent, which is remained on the surface, followed to protect Cu nanoparticles and decrease the oxidization. Also, the collection of Cu nanoparticles is achieved by only a tap funnel without using of a centrifuge. Therefore, the process becomes simpler and the consumption of energy decrease. This  method has the merit of decreasing the production cost and the possibility of mass production of Cu nanoparticles.
In our experiment, the as-prepared product contained some CuO, which seems to be due to incomplete washing of Cu nanoparticles collected in benzine.
In order to verify the effectiveness of the proposed method, a comparative experiment on the lubricating effect of Cu nanoparticles was conducted. 0.075 g of our Cu nanoparticles was mixed in 50 g of lubricant using a ball mill. After 1 h of test, the friction coefficient was measured. Figure 7 and table 2 show the friction and wear test results for the lubricants with different contents of the Cu nanoparticles prepared by the proposed method. From the test results, it was found that the lubricant with the Cu nanoparticle content of 0.15 wt% has the best lubricating properties. Figure 8 shows the lubricating properties of the pure lubricant without Cu nanoparticles. At the beginning of the test, the friction coefficient decreased continuously, but increased again to 0.036 after 24 min. With the test time, the temperature of the lubricant increased and the viscosity decreased. When the viscosity of the lubricant is reduced to a certain value, it flows out between friction surfaces. Therefore, the lubricant could not lubricate, resulting the increase in the friction coefficient.
However, the friction coefficient decreased continuously to 0.012 for the lubricant containing Cu nanoparticles of 0.15 wt%. It is because even if all lubricant flows out, the Cu nanoparticles remains between friction surfaces and acts as lubricant. This indicates that the friction coefficient of lubricant is reduced by 60% when the Cu nanoparticles prepared by the proposed method are added. The wear loss of friction surfaces in two lubricants with and without the Cu nanoparticles was compared in table 3. After 5 h of test, the weight of  specimens was measured. From the data of table 3, it was found that the wear loss decreased significantly in the lubricant with the Cu nanoparticles. Figure 9 displays the variation of friction coefficients for the lubricants with Cu nanoparticles prepared by the common liquid phase reduction method and the proposed method.
The lubricant with Cu nanoparticles prepared by the common liquid phase reduction method showed a tendency to decrease with time, however, the decreasing characteristics was unstable ( figure 9(a)). Meanwhile, for the lubricant with the Cu nanoparticles prepared by the proposed method, the decreasing tendency of friction coefficient was very stable, as shown in figure 9(b). It is because the Cu nanoparticles prepared by the proposed method have smaller and more homogenous size than one prepared by the common liquid phase