Effect of carbon nanotubes intensified coolant on the grinding performance of carburizing and quenching 12Cr2Ni4A steel

12Cr2Ni4A steel is a high-quality alloy structural steel. This article explores the surface/sub-surface condition, surface roughness, and micromorphology of carburizing and quenching 12Cr2Ni4A steel after carbon nanotubes (CNTs) fluid casting grinding, and surface quality and burn condition of the workpiece are characterized by analyzing the microhardness and the micro-morphology of the surface/sub-surface of the workpiece. The test results indicate that the high heat transfer capacity of the CNTs fluid can prevent the workpiece from being burned; further, the nanofluid casting grinding obtains the lowest grinding force and surface roughness compared to dry grinding and traditional casting grinding. The surface roughness is reduced by 19.1% at the highest because the nanoparticles play a “lubricating effect” in the grinding area, and the surface quality of the workpiece is therefore improved. The microhardness of the cross-section of the workpiece indicates that a softened layer appears on the workpiece surface ground by dry grinding, and the maximum softening layer thickness for dry grinding is about 100 µm. Tempering burns and quenching burns appear on the subsurface of the workpieces of dry grinding and traditional casting grinding, and the thicknesses of the affected layers of the tempering and quenching burns are ~ 98 and 35 µm, respectively; for the nanofluid casting, no obvious burns were observed on the sub-surface of the workpiece.


Introduction
CNC machine tools play a key role in the machining process [1]. The machining accuracy of CNC machine tools can be affected by many factors, including the structure of CNC machine tools and the connection of different modules [2]. Among them, machining accuracy is not only the main parameter to evaluate the working performance of CNC machine tools but also the main parameter to ultimately measure the working performance of CNC machine tools [3]. The machining accuracy of the CNC machine tool has impacts on the surface quality of the workpiece after grinding. The higher the machining accuracy is, the less the impact will have. Because of the high requirements of grinding processing on the machining accuracy of the machine tool and the surface quality of the workpiece after processing, CNC machine tools are widely used for precision and ultra-precision processing [4]. Especially in the field of grinding, it is used to machine the materials that are difficult to be machined [5]. However, in the process of grinding, the energy of the grinding interface is transformed into grinding heat, which is transmitted to the workpiece; this energy then causes a rapid increase in the temperature of the contact area between the workpiece and the grinding wheel, which results in the retempering of the workpiece surface [6], and a decrease in the physical properties of the workpiece surface [7]. For example, important industrial materials such as bearing steel, alloy structural steel, and ultra-high-strength steel easily burn during the grinding process, and the degree of the grinding burn considerably affects the service performance and service life of the metal parts. When the grinding conditions are not met, the temperature of the workpiece in the grinding zone increases beyond the tempering temperature of the workpiece material, which lead to grinding burns and microstructure changes on the surface and subsurface of the workpiece, Further, it can lead to a decrease in hardness and strength [8], and it may introduce unfavorable residual stress [9]. Javaroni et al. [10] reported that heat generated in the grinding zone was transferred to grinding wheel, workpiece, coolant, and grinding debris. Pande and Lal [11] said that a higher specific energy demand in grinding could lead to higher heat generation, which resulted in a higher grinding zone temperature; this could adversely affect the quality of grinding products. Therefore, it is necessary to use coolants and lubricants to reduce the grinding temperature and the transfer of grinding heat to the workpiece [12]; further, reducing the transfer of the grinding heat to the workpiece will improve the quality of the machined surface [13]. A large-capacity pouring cooling delivery system or a highpressure coolant delivery system was used to reduce the temperature of the grinding area and help reducing the tangential grinding force. Paul et al. [14] compared with dry processing and traditional casting processing to reduce the temperature of the grinding area by pouring a large amount of coolant, and this increased the processing speed by 30-40% [15]. Bhatt et al. [16] showed that water had high heat and excellent cooling capacities. Nanoparticles (such as carbon nanotubes) have good thermal conductivity and good anti-wear and anti-friction properties; adding CNTs to the coolant will help reducing the grinding temperature effectively.
12Cr2Ni4A steel is a high-quality alloy structural steel, and it has good strength, toughness, and hardenability. After carburizing and quenching, the hardness and wear resistance of the surface layer are found to be relatively high; it is suitable for alternating stress and high-load environments such as gears, shafts, worm gears, and worms. However, 12Cr2Ni4A steel suffers from grinding burn during the grinding process after carburizing and quenching. This paper proposes a method of casting grinding with a nanofluid to grind 12Cr2Ni4A steel for reducing the grinding burn; further, the results of the method are compared with those of dry grinding and traditional casting grinding. This thesis characterizes the surface quality through analyzing the roughness of the workpiece surface and the micromorphology of the workpiece surface/sub-surface and the burn condition of the workpiece by the microhardness of the cross section of the workpiece and the subsurface micromorphology of the workpiece.

Workpiece materials
The workpiece material is alloy structural steel (grade, 12Cr2Ni4A; workpiece size, 20 mm × 20 mm × 20 mm). Table 1 shows 12Cr2Ni4A steel chemical elements and weight (%). Table 2 shows heat treatment process of 12Cr2Ni4A steel. Figure 1 gives the microstructure of 12Cr2Ni4A steel. A significant difference in the metallographic structure is observed between the carburized 12Cr2Ni4A steel and the un-carburized 12Cr2Ni4A steel. For example, the metallographic structure of the carburized area is an acicular martensite, and that of the non-carburized area is lamellar martensite.

Grinding tests and measurement
The grinding test was performed using a MGK7120X60 high-precision surface grinder (Hangzhou Machine Tool Factory, China). The grinding direction of the wheel was opposite to the moving direction of the worktable, i.e., the back grinding method. Before grinding, the grinding wheel is dressed by a single diamond dresser to ensure a stable state of the grinding wheel. In the grinding process, the linear speed of grinding wheel is set at 30 m/s, and the detailed grinding conditions are summarized in Table 3.
The experimental setup and schematic diagram for the grinding force measurement are shown in Fig. 2. After grinding, we first measured the surface roughness of the ground surface of workpieces and used a scanning electron microscope (SEM) to observe the micromorphology of ground surface. Then, a wire cutting machine was used to cut the workpiece along the grinding direction and vertically. Finally, the cut surface was polished to achieve the mirror effect and observed under the optical microscope.

Preparation of nanofluids
The multi-walled carbon nanotubes (MWCNTs) were purchased from Shenzhen Sui Heng Technology Co., Ltd. with a purity of over 97%, the diameter of the MWCNTs is between 3 and 15 nm, and the pipe length is between 15 and 30 µm; the SEM image of MWCNTs is shown in Fig. 3b.
Because of the small size and large specific surface area of the CNTs, there is a strong van der Waals force between the CNTs, and CNTs tangles often occur and form large particles, which weaken the stability of CNTs in water [14]. Therefore, the preparation of nanofluids employs two steps [17]; the process flow is shown in Fig. 3a. First, the MWC-NTs were dispersed by the wet grinding process, and the dispersed MWCNTs were filtered and dried. After adding a proper amount of MWCNTs (2.0% wt) into deionized water and adding polyvinylpyrrolidone (PVP, a nonionic dispersant) with half of the content of MWCNTs, the water-based nanofluid casting liquid was prepared by ultrasonic vibration equipment for 40 min. Figure 4b shows the different grinding methods for 12Cr2Ni4A alloy structural steel (normal temperature dry grinding, traditional pouring grinding, and nanofluid pouring grinding) and the grinding force per unit grinding width via grinding depth. Grinding forces (normal grinding force F' n and tangential grinding force F' t ) increase as the increasing of grinding depth. F' n and F' t of dry grinding increased from 3.6 to 10.1 N/mm and from 1.2 to 2.68 N/mm; F' n and F' t of traditional casting grinding increased from 3.3 to 9.7 N/mm and from 1.04 to 2.3 N/mm; F' n and F' t of nanofluid casting grinding increased from 2.9 to 9.0 N/mm and from 0.71 to 2.01 N/ mm, respectively. When the grinding depth is a p = 20 µm, F' n and F' t of nanofluid casting grinding are reduced by 12.1% and 16.7% compared with that of traditional casting grinding and reduced by 19.2% and 25.0% compared with that of dry grinding, respectively. When the grinding depth is a p = 30 µm, F' n and F' t of nanofluid casting grinding are reduced by 7.2% and 12.6% compared with that of traditional casting grinding and reduced by 10.9% and 22.1% compared with that of dry grinding, respectively. Based on the tangential grinding forces in Fig. 4, the specific grinding energy E e can be calculated as Further, the maximum undeformed chip thickness h gmax can be calculated according to where C denotes the average number of effective abrasive particles per unit area (C = 3.5 mm −2 in the current study); r denotes the ratio of the width to the thickness on the section, r = 2tan , and γ denotes half of the apex angle of the abrasive particles. Equation (2) indicates that when the grinding parameters are the same during the grinding process, there is a positive correlation between E e and F' t . Further, the material removal rate Z' W can be calculated as Figure 5a indicates that as Z' w (different grinding depths, Z' w is 1.67 mm 3 /mm•s, 3.34 mm 3 /mm•s, and 5.01 mm 3 / mm•s) increases, the grinding force compared with F' n and (4) Z � w = a p v w F' t first increases and then decreases. Under dry grinding, the grinding force ratio F n '/ F t ' ranges between 3.0 and 4.05 at different material removal rates. Under the traditional pouring grinding method, the grinding force ratio F' n /F' t ranges between 3.2 and 4.7 at various material removal rates. Under the nanofluid casting grinding method, the grinding force ratio is between 4.1 and 4.8 at different material removal rates.

Grinding force and specific grinding energy
The grinding force ratio of nanofluid casting grinding is the largest, and this is mainly due to the good lubrication and antifriction performance of carbon nanotubes, which is helpful for reducing the friction between the grinding wheel and the workpiece and results in the decreasing of the tangential grinding force and the grinding force ratio.  According to reports, the increasing of material removal rate will result in greater grinding load, an increase in the wear of the grinding wheel, and an increase in the force ratio [18]. However, in the current research, the observed force ratio F' n /F' t vs Z' w tend to increase first and then decrease. When Z' w is 3.34 mm 3 /mm•s, the force ratio is the largest. Godino et al. [19] studied the wear behavior of alumina grinding wheels under different grinding conditions; this scenario can be attributed to the shedding of abrasive particles and the self-sharpening effect of the alumina grinding wheel caused by the fracture of abrasive particles [20]. Figure 5b shows the change in the specific grinding energy E e with respect to the maximum undeformed chip thickness h gmax . The specific grinding energy decreases as the maximum undeformed chip thickness h gmax increases. Three different lubrication conditions, i.e., the maximum undeformed chip thickness h gmax in the range of 0.092-0.121 µm; the specific grinding energy E e of dry grinding is in the range of 21.4-21.6 J/ mm 3 . The range of the grinding energy E e is 18.08-18.88 J/ mm3 for traditional casting grinding; the range of the specific grinding energy E e for the nanofluid casting grinding is 13.4-14.03 J/mm3. Ghosh et al. [21] and Li et al. [22] observed that, in the grinding process, E e is composed of friction energy and chip formation energy. However, most grinding energy is converted into heat [23]; for any grinding parameter, the friction energy accounts for more than 50% of the total grinding energy [24]. Combining the results of Fig. 5b, it can be inferred that 12Cr2Ni4A steel has the highest frictional energy during dry grinding, and the smallest frictional energy during nanofluid casting grinding; especially when h gmax is 0.092 µm, compared with dry grinding, the specific grinding energy E e of nanofluid casting grinding is reduced by 35.04%; and the specific grinding energy E e of nanofluid casting grinding is reduced by 25.0% compared with the traditional casting grinding. Figure 6 shows the surface roughness of workpieces ground by three different grinding methods under different grinding depths. The surface roughness of the workpiece is measured three times for each workpiece, and the measured surface roughness value is averaged. In the grinding process, when the grinding depth is a p = 10 µm, it is found that the surface roughness of the nanofluid cast grinding is 19.1% lower than that of the dry grinding; when the grinding depth is a p = 30 µm, the surface roughness of the nanofluid casting grinding is 9.4% lower than that of the dry grinding. Therefore, nanofluid casting grinding is helpful for improving surface quality compared with dry grinding and traditional casting grinding. Figure 7 gives the SEM observation of the ground surface micromorphology. As shown in Fig. 7a, b, and c, when the grinding depth is at 20 µm, the surface of the nanofluid casting grinding is better than that of dry grinding and traditional casting grinding. There are many defects on the surface ground by dry grinding including overlaps of the workpiece material, tearing of the workpiece material, and adhesion of metal fragments. The surface machining defects for nanofluid casting grinding are less than those of dry grinding and traditional casting grinding. When the grinding depth a p = 30 µm, the machined surface is deteriorated as shown in Fig. 7d, e, and f; surface defects for three different grinding method, such as the workpiece material overlap, surface cavities, tearing of workpiece material, and adhesion of flake debris, are more prominent compared with that in Fig. 7a, b, and c. The ground surface by nanofluid casting grinding is better than that of dry grinding and traditional casting grinding; the reasons are that the temperature of grinding zone is higher in the dry grinding process compared with that of traditional casting grinding and nanofluid casting grinding, the surface material softening is more serious, the materials which flow laterally due to the extrusion of abrasive grains and the grinding debris are squeezed into flakes and bonded to the workpiece surface, and the above phenomenon is more obvious with the increase of grinding depth. For nanofluid casting grinding, due to the good thermal conductivity and antifriction performance of carbon nanotubes, the grinding temperature is relatively low, and the softening phenomenon in the grinding zone is relatively light. Therefore, the surface integrity is better than that of dry grinding and traditional pouring grinding.  Figure 8 shows the distribution of the microhardness measured perpendicular to the feed direction for three different grinding methods under various grinding depths. At different grinding depths, surface hardening occurs in both traditional casting grinding and nanofluid casting grinding, and with the increase of grinding depth, material softening appears on the dry grinding surface (Fig. 8c). When the grinding depth is set as 30 µm, the surface maximum Vickers hardness of the traditional casting grinding is as high as 841 HV (16.7% higher than the carburized surface before grinding), and the hardness is rapidly reduced to 712 HV; then, the hardness rises to 757 HV, and then, it slowly decreases to the hardness of workpiece matrix. The reason for this phenomenon is that the instantaneous grinding temperature of the workpiece surface exceeds the critical temperature of the metal phase transition, and then, under the action of the coolant, quenching burn appears on the ground surface. The ground surface hardness of dry grinding is 618 HV, lower than the hardness of unground surface of 700 HV, and then it rises to 688 HV; and then it slowly decreases to the workpiece matrix hardness. The reason for this phenomenon is that the instantaneous grinding temperature of the workpiece surface exceeds that of martensite. The volume transition temperature is lower than the metal phase transition critical temperature, and the cooling of the workpiece is slow; therefore, the surface hardness after grinding decreases. The thermal softening of the workpiece material is attributed to the effect of thermal energy on the mechanical plastic deformation [25]. Figure 9 shows the SEM images of the subsurface using alumina grinding wheels under different lubrication conditions and various grinding depths. Subsurface defects (cavities) are found when the grinding depth a p = 30 µm for dry grinding. The subsurface of the workpiece has tempered burns, and the layer thickness H 1 of the tempered burns is about 98 µm; the subsurface of the traditional casting grinding has quenching burns, and the layer thickness of the quenching burns H 2 is ~ 35 µm. When the grinding depth is a p = 20 µm, tempering burns appear on the subsurface of the workpiece under dry grinding, and the layer thickness H 3 of the tempering burns is about 10 µm. The subsurface of the traditional casting grinding has quenching burns. The layer thickness H 4 of the quenching burn is ~ 15 µm. When the grinding depth for dry grinding is at 30 µm, the carbides are precipitated. The energy spectrum (e) indicates that the carbon content is significantly larger than that in the energy spectrum (f).

Calculation and analysis of grinding force
Grinding force plays an important role in the grinding process because it not only directly affects the wear of the grinding wheel, grinding temperature, and surface integrity, but also has a considerable effect on the removal of materials [24]. However, the variables of this important and monitorable process largely depend on the grinding conditions and material properties. Therefore, whether the workpiece is burned and the degree of burn can be diagnosed by analyzing the grinding force. Both normal and tangential grinding forces are composed of chip forming force and friction force [23]. According to the grinding force model proposed by Lichun et al. [26] and Yao et al. [27], the grinding force component can be expressed as: where F ′ nc , F ′ ns , F ′ tc , and F ′ ts denote the normal and tangential chip formation force components, normal and tangential friction force components, respectively; H denotes the grinding force per unit area; δ represents the top area of a single working abrasive particle, i.e., the workpiece and the working grinding; the actual contact area of the grain p denotes the average contact pressure between the actual wear surface of the abrasive grain and the workpiece; A represents the proportional coefficient related to the number of static sharpening; α and β represent the exponential coefficients related to the grain distribution; c1 denotes the static sharpening density; γ denotes half of the tip angle of the  abrasive particles (generally γ = 60.); μ represents the friction coefficient; and d e represents the equivalent diameter of the grinding wheel. Since the state of the grinding wheel before grinding is the same, the exponential coefficients α and β are equal. Equations (5), (6), and (7) can be written as: The results of the dry grinding force, traditional pouring grinding force, and nanofluid pouring grinding force (Fig. 4) are substituted into the equation, respectively, and coefficients H 1 , H 2 , H 3 , and H 4 in Eq. (8) are calculated. The coefficients of H 1 and H 3 generated by chip formation are very large, which indicates that the size effect is very significant in the micro cutting process [25]. Coefficients H 2 and H 4 related to friction have relatively small changes, and p is consistent with coefficients H 2 and H 4 . Table 4 and Fig. 10 illustrate the calculation results of the grinding force components under three different grinding methods and different grinding depth conditions, i.e., chip forming force and friction force. If the grinding depth is fixed, the chip forming force is considerably lower than the friction force, which is less than 15% of the total grinding force. Thus, most of the grinding force is generated by friction, which causes a greater grinding force; the friction accounts for about 85% of the entire grinding force, and thus, a large amount of heat is generated during the grinding process. The generated heat is not dissipated in time, which causes a large amount of heat to be transferred to the workpiece. This causes burns on the processed workpiece. Because of the normal grinding force component (Fig. 10a), the ratio of F' nc /F' n slowly increases as the grinding depth a p increases, and the ratio of F' ns /F' n continues to decrease. The result of the tangential grinding force component (Fig. 10b) and the result of the normal grinding force component (Fig. 10a) have the same changing trend. This phenomenon may be attributed to the above-mentioned crystal grain shedding and fracture behavior. Although it will shorten the service life of the grinding wheel, this phenomenon may result in a self-sharpening effect of the alumina grinding wheel.

Conclusions
In this study, three different grinding methods (dry grinding, traditional casting grinding, and nanofluid casting grinding) were used to perform grinding experiments on 12Cr2Ni4A steel. The workpieces after grinding were comprehensively studied. The main conclusions are as follows: (1) The nanofluid casting grinding method can effectively reduce the grinding force, especially the tangential grinding force. The normal grinding force of nanofluid casting grinding is lower than that of dry grinding up to 19.2%, and the tangential grinding force of nanofluid  casting grinding is 25.0% lower than that of dry grinding. (2) With an increase in grinding depth, the surface roughness of nanofluid casting grinding workpiece is lower than that of traditional casting grinding and dry grinding; the surface roughness is reduced by 19.1% at the highest. The surface defects of the nanofluid casting grinding workpiece were far lower than the surface defects of the dry grinding and the traditional casting grinding (workpiece material overlap, surface cavity, tearing of workpiece material, and adhesion of metal fragments). Data availability All data generated or analyzed during this study are included in this article.

Declarations
Ethics approval This article does not contain any studies with human participants or animals performed by any of the authors.