Evaluation of dry-in-place lubricants for cold forging by using an optimal steady combined forward and backward extrusion testing method

This study evaluated dry-in-place lubricants used for cold forging. A group of isothermal compression tests with a strain rate (ε˙\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\dot \varepsilon $$\end{document}) range of 0.001–1 s−1 and temperature (T) range of 30–400 °C were completed. The flow stress (σ) curves of annealed steel S45C were obtained, and a corresponding Hensel—Spittel model was developed to support finite element (FE) simulation. The sensitivity of the steady combined forward and backward extrusion (SCFBE) test proposed in another study was improved by approximately 20% after it was optimized using the results of the FE simulations. Key parameters were identified, and the calibration curves after optimization were obtained. On the basis of the optimized test, a friction testing setup with a heating system was developed, in which the die temperature could be adjusted from room temperature (RT) to 230 °C. Three dry-in-place lubricants and conventional phosphating lubricant were tested, and the friction factors (m), forming loads, and ejection loads were measured. The surface features of the specimens after testing were also investigated. According to the testing results, of the three tested dry-in-place lubricants, the mica type was the best. In addition, the optimized friction testing design was verified as effective.


Introduction
Cold forging is an advanced manufacturing technique that is inexpensive and is widely used in the mass production of industrial parts, such as stepped shafts [1,2] and complex gears [3,4]. Although cold forging has developed considerably over the course of more than 70 years, shortening the process chain [5], reducing the number of process steps [6], and developing hybrid raw parts [7] remain areas that warrant further research. In addition, the considerable surface expansion and normal pressure at the interface of the tool and workpiece involved in cold forging lead to severe tribological conditions [8]. Because of this, tribological evaluation and advanced lubrication must be employed to ensure successful cold forging.
In cold forging, proper lubrication can prevent direct contact at the interface between the workpieces and tools, which can enable parts with high surface quality and tools with long service lives to be forged. Several methods for lubrication [8], including those involving oil-type [9], polymer-based [10], and MoS 2 -based [11] lubricants, have been proposed. The most commonly used method is traditional phosphating plus soap coating. However, although phosphate lubricating is effective, this method has negative environmental effects because of the sludge associated with such heavy metals. Therefore, environmentally benign tribological systems must be developed, in which substitutes can be used for conventional hazardous lubricants [12]. Accordingly, dry-in-place lubricating systems have gained an increasing amount of attention [13].
To support the development of new lubricants, a quantitative method for evaluating the effects of a lubricant in cold forging is required. Several extrusion-type testing methods have been proposed. Of these methods, combined forward and backward extrusion was reported to more accurately simulate the metal deformation behavior of complex forging parts than double cup extrusion [14] and backward extrusion [15]. A backward cup-forward rod extrusion test [16] was reported to be suitable for estimating the friction of non-steady state extrusion [17], and a design without a die angle (θ) on the forward extrusion die was used to quantify the friction that occurred during the cold forming of aluminum alloys [18]. In addition, a novel method for determining friction in the cold forging of complex parts by using a steady combined forward and backward extrusion (SCFBE) test was proposed [19].
In the present study, the suitability of three types of dry-in-place lubricants for cold forging was evaluated, and a traditional phosphate lubricant was used as a reference for comparison. In Section 2, the three dry-in-place lubricants are briefly introduced. Section 3 describes an analysis of the flow stress () curves of annealed steel S45C, which was conducted using isothermal compression tests within strain rates ( )  of 0.001-1 s −1 and temperatures of 30-400 °C, and the development of a Hensel-Spittel model, which served as an input material model for later finite element (FE) simulations. Section 4 details the process, through which the parameters of the FE model for combined forward and backward extrusion were set. Section 5 describes how an SCFBE test was reoptimized to a range, in which the friction factor was set to 0.03-0.15, and a group of calibration curves with higher sensitivity were obtained. Section 6 describes a friction testing setup for cold forging developed on the basis of the results of the optimized SCFBE testing method and a corresponding experiment. In Section 7, the experimental results, including those related to the friction factors measured using the optimal SCFBE test, the measured extrusion and ejection loads, and the deformation of the specimens obtained using four different lubricants, are presented and compared. The main conclusions are presented in Section 8.

Dry-in-place lubricants
New, environmentally friendly lubrication systems have been developed to replace zinc phosphate-based systems. Many of these systems employ dry-in-place lubricants. Such lubricants have attracted considerable attention because they can be used to form a lubricating film through the simple process of applying a film treatment liquid to the surface of the object and then drying it. A schematic of the treatment process of dry-in-place lubricant systems is presented in Fig. 1. First, descaling is performed to remove oxides from the workpiece surface. Common methods used for descaling include shot blasting, wet blasting, and pickling. Second, the workpiece is washed with water to remove foreign substances that may have adhered to the surface. Third, the workpiece is immersed in a lubricating liquid heated to approximately 60 °C. The lubricating liquid comprises a lubricant ingredient and a water-soluble base ingredient. While the workpiece is immersed, a thin film of treatment liquid is adhered to the metal surface. Fourth, the water in the lubricating film evaporates because of the heat of the metal workpiece. After this process is complete, a double-layer lubricating film is formed on the workpiece surface.
With dry-in-place lubricant systems, a double-layer film can be deposited on a metal workpiece through a simpler treatment process than that required for normal zinc-phosphate coating [13]. For normal polymer-based films, high lubricity polymer can reduce friction, and a base material, such as organic or inorganic salt, can prevent galling. Solid lubricants, such as MoS 2 , may be used when a film is required to have a high antigalling capacity. Organic-supported clay minerals with excellent cleavage between layers have also been synthesized and applied as solid lubricants with both high antigalling abilities and lubricity [20]. In the present study, three types of dry-in-place lubricants, namely general polymer-based, MoS 2 -including, and organic-supported (denatured) mica types of lubricants, were evaluated.

Testing steel and material modeling
A typical middle carbon steel, S45C, was used to make the testing workpiece. To achieve favorable formability, the raw material was fully annealed. To enable the characterization of the material flow behavior, a group of isothermal compression tests with constant  of 0.001, 0.01, 0.1, and 1 s −1 under the working temperatures of 30 °C, 200 °C, and 400 °C were performed. True -strain () curves were obtained and plotted (Fig. 2). The  decreased as the deformation where T is the temperature (in K).

FE model
In the SCFBE test, the ratio between the forward rod and backward cup variation with the punch stroke was used to calibrate the corresponding friction factor calibration, with calibration curves determined at different friction factors through FE simulations [19]. The design of the testing method was developed on the basis of the following parameters: a degree of deformation of backward extrusion (ε b ), a degree of deformation of forward extrusion (ε f ), θ, a taper angle of the punch (α), a punch land size (h p ), the ratio (r 2 /r 1 ), and a die land size (h d ). The initial height h 0 is two times of the initial radius r 0 of the workpiece. All dimension parameters are presented in Fig. 3(a). A sensitive design for cold forging with a normal friction factor range of 0.0-0.3 was obtained [19]. A two-dimensional FE model of the combined forward and backward extrusion process was developed using COLDFORM software (Transvalor, Fig. 3(b)).
In addition to the Hensel-Spittel model presented in Section 3, other key parameters affecting the S45C workpiece material, as listed in Table 1, were input to the FE model. The workpiece's mesh size and boundary coefficient were set at 0.5 mm and 4, respectively. Therefore, the mesh size of the outermost surface was 0.125 mm. Deformable tools made of H13 were used, and the data on the material obtained from a software material library were adopted. A constant shear friction model of = mk ( : friction stress, m: friction factor, and k: shear stress) was used to assess the friction behavior between the workpiece and tools. The heat transfer was also calculated, and the heat transfer coefficient at the interface between the workpiece and tools was set at 10,000 W/(m 2 ·K). A forging press with a crank motion was selected because such presses are used in actual tests. The parameters of the model are listed in Table 1.

Optimization of the SCFBE test
To simulate the actual requirements for testing dry-in-place lubricants, an orthogonal optimal design method was used to reoptimize the SCFBE test, with the friction factor set to a range of 0.03-0.15. The 15 is the extrusion force when the friction factor is 0.15, and F 0.03 is the extrusion force when the friction factor is 0.03. The deformation degree, which is the ratio of the cross-sectional area difference before and after deformation to the cross-sectional area before deformation, of forward extrusion was set at 50%. The range of the design variables were as follows:  Table 2. The mean values of the evaluation index for each parameter at the levels of k 1 -k 5 were calculated, and the relative range (R), which is the difference between the maximum and minimum k value, was obtained. The ε b significantly  influenced the evaluation index, whereas the influence of r 2 /r 1 and h d were small. According to the k values at the five levels, an optimized design scheme could be achieved by applying the following parameters: ε b = 60%, α = 75°, r 2 /r 1 = 0.9, θ = 40°, h p = 1.5 mm, and h d = 3.5 mm. Simulations, in which the optimal workpiece design was subjected to temperatures of 20 °C (room temperature (RT)), 50 °C, and 100 °C, were performed, and relative  values were obtained (Table 3). After optimization, the evaluation index increased from www.Springer.com/journal/40544 | Friction 15.13% to 18.23%; and the sensitivity of the optimal design was 20.49% higher than that of the previous design. Although the sensitivity decreased as the temperature increased, the sensitivity remained significantly higher when the workpiece was subjected to a temperature of 100 °C.
The calibration curves of the previous design and optimal design are presented in Fig. 4. The ordinates of the curves are the ratio between the forward rod and backward cup (h 1 /h 2 , h 1 and h 2 are the heights of the forward rod and backward cup, respectively), and the abscissas of the curves are the ratio between the stroke of the punch and the initial height of the billet (h). The sensitivity after optimization significantly improved, resulting in more scattered calibration curves. The difference of the height ratio h 1 /h 2 becomes larger within the same difference of the friction factor. These curves were used to establish a high-sensitive SCFBE (H-SCFBE) testing method to evaluate the friction conditions for cold forging. h 1 and h 2 were measured, and the calibration curves were used to determine specific friction factors. Generally, more scattered calibration curves can improve the ease, with which friction factors can be used to identify differences between lubricants.

Experiments
The obtained H-SCFBE testing method design was used to manufacture a punch and die with optimal dimensions. The punch and die were made from Japanese Industrial Standard (JIS)-SKD61 and JIS-HAP40, respectively, and were both coated with TiN ( Fig. 5). In cold forging, 90%-95% of the deformation energy and interface friction energy between the workpiece and die are converted to heat, and the workpiece can be heated to several hundred degrees Celsius when it is made from mild steel [21]. The heat from the workpiece immediately transfers to the surrounding dies [22], and the dies steadily heat up to 150-250 °C after they undergo hundreds or thousands of cycles of cold forward extrusion [23,24]. To control the die temperature, eight heater rods were inserted into the die (Fig. 5(b)). Using this heating system enabled the die temperature to be adjusted from RT to 230 °C. The punch and die were installed on a 2,000 kN Table 3 Comparison of previous design and optimal design.  servo press (H1F, Komatsu Co., Ltd.), and a friction testing setup based on the obtained H-SCFBE test design was established. During the friction tests, the press was set to a typical crank motion, and the rotation speed was set to 30 r/min. To enable the forming and ejection loads to be monitored during the testing process, two load cells were assembled. The load cell used to measure the forming load was a cylinder with a diameter of 220 mm and was attached to the upper part of the die set. A pressure plate was placed between the punch and the load cell, which enabled the force applied to the punch to be evenly transmitted to the load cell. Another cylindrical load cell with a diameter of 40 mm was used to measure the ejection load, and this load cell was placed between a pneumatic cylinder and ejector. The ejector was lifted by the load cell, which was pushed by the pneumatic cylinder during ejection. The load cell recorded the reaction force when the ejector came into contact with the workpiece.
The effectiveness of a traditional phosphate plus soap coating (Phos/soap) and wax-type, MoS 2 -type, and denatured mica-type dry-in-place lubricants was evaluated using the established H-SCFBE testing setup. In addition to 30 °C (i.e., summer RT), the die was preheated to 90, 150, and 210 °C during the friction tests. Before each test, a cylindrical specimen with a diameter of 25.0 mm and a height of 25.0 mm was heated to 60 °C. After each test, the punch and die were cleaned to remove any lubricant residue. The extruded testing parts were obtained, and the h 1 and h 2 were measured with the aid of a projector (TM-3000, Keyence). The specific friction factors under various lubricating conditions were determined using calibration curves (Fig. 6).

Analysis of experimental results
A comprehensive analysis was conducted. First, the friction factor, which served as a direct evaluation index for the different lubricants, was determined in order to experimentally validate the H-SCFBE testing method. Second, the forming loads of the punch, which were measured during the extrusion process, were compared. This extrusion force F is mainly used to counter the deformation resistance of the workpiece material and the friction resistance between the workpiece and dies. Third, the ejection load of the ejector, which was measured during the ejection process, was determined. The ejection force reflects the difficulty, with which the workpiece is released from the die and is directly influenced by the amount of lubricant residue that is present after extrusion. Fourth, the workpiece's surfaces were examined after the tests to determine the surface quality after forming were complete. If the surface of a workpiece or die is damaged during extrusion, in subsequent extrusion steps, the damaged surface can cause severe friction, which can result in an excessive ejection force. This excessive ejection force can further damage the workpiece surface, which leads to the formed parts, having poor surface quality. In the cold forging industry, multistage forging processes are regularly used to manufacture complex parts. Analyzing the ejection force and workpiece surface after tests are completed can assist in analyzing the lubricant residue, which can affect the subsequent stages of the multistage cold forging process.

Friction factors
To analyze the friction factors, the h 1 and h 2 of the www.Springer.com/journal/40544 | Friction extruded specimen were measured using a projector, and the corresponding ratio of h 1 /h 2 was calculated and then plotted with the punch stroke by using calibration curves (Fig. 7). For the conventional Phos/soap lubricant, two workpieces were subjected to each testing condition. A data point for each workpiece was plotted on the calibration curves. As indicated by the position of the points, the friction factors, when Phos/soap lubricant was applied and the material was subjected to temperatures of 30, 90, 150, and 210 °C, were 0.092, 0.066, 0.048, and 0.04, respectively.
The friction factors when the various lubricants were used are plotted in Fig. 8. The points in the plot of the H-SCFBE test results are more scattered than those of the plot of the SCFBE test results. The maximum difference between the friction factors at various die temperatures measured using the H-SCFBE test was 0.053-0.076, whereas the maximum difference measured using the SCFBE test was 0.033-0.050. This indicates that the H-SCFBE test is more sensitive in determining the friction conditions of different lubricants. As indicated in Fig. 8(b), identifying the differences between the results for the MoS 2 -type and denatured mica-type lubricants in the SCFBE testing results is difficult. This supports the premise that the optimal design described in Section 5 is necessary and effective.
The friction factors of all the lubricants decreased as the preheating temperature of the lower die increased. The main reason for this is that lubricity can more effectively be achieved when the temperature at the interface of a tool and a workpiece is elevated. As reported in Ref. [25], when the temperature of Phos/soap lubricating film exceeds its melting point (120-130 °C), the zinc stearate component melts, and the lubricity improves. The order of the friction factor values, from high to low, corresponding to the four lubricants when the die temperature was 30 and 90 °C was mica type > MoS 2 type > wax type > Phos/soap. The Phos/soap coating exhibited the most favorable lubricating properties. As the die temperature increased, the measured friction factor values for  all four lubricants decreased to less than 0.1. The mica-type dry coating melted well when the die temperature increased, and its cleavability improved [20]. Therefore, the mica-type dry coating has a more favorable lubricating effect during combined extrusion processes when the die temperature is 210 °C.
The friction factor values obtained through the H-SCFBE test were generally smaller than those obtained through the SCFBE test. To identify the cause of this, the simulation results were used to analyze the tribological conditions of the two tests. A simulation, in which an initial workpiece temperature of 60 °C, an initial die temperature of 100 °C, and a friction factor of 0.1 were applied, was run. The results, including the surface expansion, relative sliding velocity (RSV), and temperature distribution, were extracted at a relative punch stroke of 80%. As indicated in Fig. 9, the thermal distributions of the surface expansion, RSV, and temperature were nonuniform, and the values obtained from the H-SCFBE test were greater than those obtained from the SCFBE test. The load-time curves (Fig. 10) reveal that the forming load of the H-SCFBE test was larger than that of the SCFBE test, and its maximum forming load was 18% higher than that of the SCFBE test. The average contact pressure increases with the size of the forming load. This is likely the main reason for a smaller friction factor value being obtained through the H-SCFBE test.

Forming load
As indicated in Fig. 11, forming load curves can have four phases. Forward extrusion mainly occurs in Phase I, and the F in the initial section increases sharply. After the material flows through the calibrating zone of the lower die, steady-state forward extrusion and backward extrusion occur. The F decreases slightly when the extrusion stroke is applied in Phase II. In Phase III, the material flows through the calibration zone of the punch. In composite extrusion processes, materials can flow more freely, which reduces the F. Furthermore, the material simultaneously flows forward and backward, resulting in the length of the rod and the depth of the cup increasing. In Phase IV, the punch moves upward, and the F sharply decreases. In Phase III, the friction resistance resulting from the landing being calibrated by the tools becomes stable, and the material flow of the combined extrusion process becomes steady. In the present study, the deformation forces of the various lubricants exhibited slight differences (Fig. 11(b)).

www.Springer.com/journal/40544 | Friction
The F decreased as the die temperature increased ( Fig. 11(a)). The F for the specimen with the Phos/soap coating was much lower than that for the specimens with the other lubricants ( Fig. 11(b)). The maximum forming loads of the specimens with different lubricants under different die temperatures are listed in Table 4. The variation in the maximum loads was consistent with the variation in the friction factor values. The absolute error between the measured maximum loads at different die temperatures was 2.66%-17.15%. The absolute error between the measured maximum loads of the different lubricants was 0.49%-11.79%, and the minimum absolute difference was only 2 kN. The effect of the die temperature on the forming load was more significant than that of the different lubricants. The forming load is determined through deformation and friction resistance, with deformation resistance having the stronger influence. Deformation resistance, which can be directly determined through material , is influenced by the temperature of a workpiece, which was affected by the preheated die temperature in the present experiments. Therefore, the variation in friction resistance when different lubricants are applied has only a secondary effect on changes in the forming load when the material  is more sensitive to the temperature of a workpiece during the extrusion process. Because of this, evaluating lubricants on the basis of variations in the forming load is ineffective.

Ejection load
The process of ejecting a specimen involves four phases. In the first, the specimen is loosened from the die by an ejector, and for this to occur, the static friction force must be overcome. The peak ejection loads are presented in Fig. 12. In the second phase, the ejector moves upward, and the lower rod and upper cup of the specimen slide along the die land and the die cavity, respectively. The ejection load gradually decreases with the ejection stroke because of the resistance created by the friction of sliding.
In the third phase, the tip of the specimen moves through the die land. The sharp edge on the tip of the lower rod comes into contact with the die land, and this introduces another peak ejection load (Fig. 12).
In the fourth phase, the upper cup slides along the die cavity, and the specimen is pushed out of the die. The ejection load decreases to zero. Ejection loads are generally increased by forming loads in forward extrusion, and a large forming load can expand a die container and cause it to have a stronger spring-back elastic force, which enables it to hold the specimen tightly. However, a jump occurs in the first phase of the ejection load curve when the die is preheated to 210 °C, as can be observed in Fig. 12. When the lubricants are melted, a specimen that is stuck in the die can more easily be loosened by the ejector, which causes a jump load. As indicated in Fig. 12(b), the second peak values of the different lubricants were similar. When this occurred, the effect of the different lubricants on the ejection load was relatively small because a small amount of lubricant remained on the tip area. The maximum ejection loads from all experiments are listed in Table 5. For the mica-type lubricant, the maximum ejection load decreased as the die temperature increased. However, for the other lubricants, the maximum ejection load was the largest at 210 °C.

Surface observations after tests
As indicated in Fig. 13, all specimens, on which different lubricants were applied, had good surface quality, and no scratches, galling, or other surface defects were identified; as evident in Fig. 13, Phos/soap coating is gray-brown, wax-type lubricant is almost transparent and colorless, MoS 2 -type lubricant is black, and mica-type lubricant is white. All lubricants were effective in the combined extrusion process. Nevertheless, the specimens exhibited geometric differences caused by the various lubricating conditions. Generally, a steady-state temperature distribution on a cold forging die can be obtained after 100 or 1,000 cycles, depending on the forging process parameters and die construction when the die temperature is 100-300 °C [23,24]. Consideration of this indicates that the mica-type lubricant is the best of the three dry-in-place lubricants when the measured friction factors, forming loads, and ejection loads of the lubricants are compared. Further comparative analysis was conducted, in which the surfaces of the specimens with a Phos/soap coating and mica-type dry-in-place lubricant that were tested using the H-SCFBE were analyzed using the scanning electron microscopy (SEM). Three typical zones were identified on the specimens (Fig. 13(a)). The dark areas in the SEM images are the metal matrix, and the light areas are the lubricating coating. The base zone is the part of the specimen that was in contact with the die land, the tip zone is the area near the front end of the specimen, and the middle zone is the part between the base and tip zones. The SEM images of the specimens at 90 and 210 °C are included in Table 6. According to the observations of the surfaces of the tested specimens, in the middle zone, the Phos/soap coating is more uniform, which is likely due to the favorable surface ductility of Phos/soap coatings; for such coatings, the chemically reacted phosphating layer moves along the deformed metal surface because of its high ductility, and the soap coating, which is stored in the microporous structure of the phosphating layer, effectively adheres to the metal surface. By contrast, the mica-type coating was unable to move along the deformed metal surface because it adheres poorly to metal surfaces. Of the zones, the tip zones had the least lubricant residue. In the extrusion direction, the small radius of the cylindrical workpiece acted as an inner core and came into contact with the surface of the die, and the lubricant coating around the surface of the end of the workpiece was nearly extruded away because of the severe deformation that occurred in that zone. This led to a band with no or little lubricant residue forming in the tip zone. In the base zones, a sliding track can be observed on the Phos/soap-coated specimens, whereas on the mica-type specimens, the original steel can be observed. The larger surface expansion of the base zone and the high ductility of the Phos/soap coating resulted in the formation of a sliding track on the surface of the specimen.

Conclusions
To characterize annealed steel S45C as a testing workpiece, a group of isothermal compression tests within the  range of 0.001-1 s −1 and temperature range of 30-400 °C were performed. A step-by-step compression process was adopted to maintain the workpiece temperature. The data in the form of true - curves were collected, and a corresponding Hensel-Spittel model was developed and input to an FE model to simulate a combined extrusion process. On the basis of the FE simulations, an optimized design was obtained, in which the deformation degrees of backward extrusion were ε b = 60%, α = 75°, r 2 /r 1 = 0.9, θ = 40°, h p = 1.5 mm, and h d = 3.5 mm. The H-SCFBE test exhibited approximately 20% higher sensitivity than the previous SCFBE test did. The experimental results reveal that the optimized friction testing design was effective.
The H-SCFBE test was used to develop a friction testing setup with a capacity for die temperature adjustment. Three dry-in-place lubricants and conventional phosphating lubricant were applied to a workpiece with an initial temperature of 60 °C and tested under four different die temperatures. The effect of die temperature on the forming load was greater than that of lubricant type. Therefore, evaluating lubricants only on the basis of the forming load is ineffective.
The testing results reveal that the friction factor values of all lubricants decreased with the die temperature. This may be because lubricity can more easily be achieved at an elevated interface temperature. The friction factor values obtained using the H-SCFBE test were smaller than those obtained using the SCFBE test, which was likely because of the higher average contact pressure. The mica-type lubricant was the best of the three investigated dry-in-place lubricants when the die is pre-heated up to 210 °C. The findings of this study may assist researchers in further improving the lubricity of current lubricants or in developing new lubricants in the future.