Influence of Non-metallic Inclusions in 316L Stainless Steels on Machining Using Different Cutting Speeds

Abstract

This research focuses on providing a detailed characteristic of non-metallic inclusions (NMIs) in 316L stainless steels with and without Ca treatment after machining using different cutting speeds. The electrolytic extraction (EE) technique was used for three-dimensional determinations of the inclusion characteristics. Quantitative data from the fragile non-metallic inclusions (such as size, surface area, number) in chips obtained from different cutting speeds and materials were determined. The morphologies of NMIs in the chip samples were quite different compared to the original inclusions in the stainless steel samples before machining. It was proved that the deformation degree of soft inclusions such as MnS and CaO–SiO₂–Al₂O₃–MgO–TiO_x is dependent on the cutting speed as well as the temperatures and deformation degrees of the metal matrix during machining. The total surface areas of MnS inclusions increase from 2.8 to 3.8 times compared to the original total areas with an increased cutting speed. The total surface areas of soft oxide inclusions also increase from 1.1 to 3.5 times compared to the original total areas. In addition, the tool-chip contact lengths were also measured on the rake face of the tool, and the results were compared to the determined characteristics of the observed inclusions. It was found that the modification of NMIs by Ca treatment in 316L stainless steels is preferred for high cutting speeds.

1. Introduction

By improving the steel-making technologies and process, the impurities in the steel have been kept at a low-level, which generates excellent mechanical properties of steel. However, new grades of steels with extremely high qualities lead to production problems due to a decreased machining performance and a dramatically reduced tool life during the machining process, which makes up a big part of the entire manufacturing cost. Thus, the consumption of time, energy, and money have increased during the manufacture of these new grades of steels.¹⁾ Thus, how to manufacture these steels with a high efficiency and a low cost is a task for both the steel production industry and the manufacturing industry.

The machinability is affected by several factors. However, even when using the same tools to machine steels with similar mechanical properties, steels having different non-metallic inclusions (NMIs) also can show different machinabilities.^{2,3,4,5,6,7,8,9,10,11,12,13,14)} The different hardnesses and thermal expansions of NMIs in steels will play different roles (negative or positive) during the machining of these steels. Rigid inclusions (such as Al₂O₃, TiCN) will scratch the tool surface and decrease the tool life dramatically.^10,11) The existence of so-called free-cutting elements (such as S, Ca, Cu, Bi) are beneficial for the machinability by decreasing the cutting force, chip thickness, rake wear, and flank wear. This, in turn, will lead to an increased tool life and cutting surface quality.^{1,2,3,4,5,6,7,8,9)} Thus, a modification of NMIs is beneficial to balance the preferred machinability and required mechanical properties.

The behavior and role of non-metallic inclusions have been discussed in several previous studies,^{2,3,4,5,6,7,8,9,10,11,12,13,14)} as mentioned above. However, the benefit of the modification of NMIs was not the same for different cutting conditions or different reference materials, according to previous studies.^15,16,17) This is due to that the behavior of non-metallic inclusions will be different during different machining conditions. As a result, the effect of non-metallic inclusions on the machinability will be different. Though there are a few publications that have directly investigated the NMIs on the cross-sections of the chips after machining,^12,17) no previous studies have focused on a more in-depth investigation of the non-metallic inclusion characteristics after machining processes using different cutting parameters.

In the orthogonal cutting process, the chip produced in the shearing zone will move along the rake face of the tool until it curves off or breaks up.¹⁸⁾ The contact length between the chip and tool in the chip flow direction from the tip to where the chip leaves the tool is defined as the tool-chip contact length. On the one hand, the tool-chip contact length (L_t) can be affected by several factors such as cutting parameters, tool geometry, tool material, and coating material, cutting fluid, workpiece diameter, workpiece material, etc.¹⁹⁾ Several groups have studied the effect of different factors on the tool-chip contact length. For example, it was found that the contact length will increase with an increased cutting depth but decrease with an increasing speed.^19,20,21) On the other hand, the tool-chip contact length could affect and be related to several essential factors in the cutting process (such as the tool life, chip form, tool temperature, cutting forces, tool stability, surface finish, and energy consumption). Previous work has shown a clear proportional relationship between the contact length and the chip compression ratio,²⁰⁾ the radius of chip curvature,²²⁾ and tool wear.²³⁾ Thus, investigations of the tool-chip contact length will give quite much information about the effect of NMIs on the machining performance.

This study aims to investigate the characteristics of different non-metallic inclusions in stainless steel before and after machining. Ca treatment was applied at the end of ladle treatment to modify the NMIs in the steel more preferable for machining. The non-metallic inclusions in chips obtained from different cutting speeds and materials are systematically investigated and compared after the electrolytic extraction. The tool-chip contact lengths for the orthogonal cutting processes with different cutting speeds were also measured and compared. Therefore, based on the comparison results of NMIs’ behaviors and tool-chip contact lengths at different cutting speeds, the effects of different NMIs on the machinability of the given stainless steel during machining are discussed.

2. Experimental

2.1. Workpiece Materials and Cutting Tools

The workpiece materials studied in this research are two rolled bars of 316L stainless steel with a diameter of 121 mm. The reference steel (named 316R) was taken from a standard industrial trial. Furthermore, the experimental steel (named 316Ca) was produced using the same production process with an extra addition of CaSi wire at the end of the ladle refining to modify non-metallic inclusions. The contents of the main elements in investigated steels are listed in Table 1.

Table 1. Chemical compositions of the reference steel (316R) and the Ca-treated steel (316Ca).

Steel Grade	C	Si	Mn	Cr	Ni	Mo	N	Al	S	O	Ca
	[mass%]
316R	0.02	0.38	1.60	16.82	11.18	2.02	0.059	0.004	0.007	0.002	–
316Ca	0.01	0.46	1.58	16.86	11.14	2.09	0.060	0.004	0.009	0.0059	0.0028

TPUN 160304 2025 inserts with chemical vapor deposition (CVD) coating by Ti(C,N)+Al₂O₃+TiN were used for machining tests in this study.

2.2. Machining Test

The orthogonal cutting processes (Fig. 1(a)) were carried out at speeds of 100 m/min and 250 m/min without using any cutting fluid from a 120 mm diameter to an 89 mm diameter of the bar. The feed rate and cutting depth were 0.25 mm/rev and 3 mm, respectively. Since all other cutting parameters were the same in this study, the only difference was the different workpiece materials, which have similar compositions but quite different NMI characteristics.

Fig. 1.

(a) Schematic diagram of tool action; (b) photograph of tool rake face of 316R stainless steel at Vc = 100 m/min.

A typical picture of the rake face of the tool with information of tool-chip contact length (L_c) after the orthogonal cutting process is shown in Fig. 1(b). Two different regions can be observed in the tool-chip contact, a first region where a seizure and adhering process takes place and a second region where a conventional sliding friction process takes place.^24,25) There is also a stochastic contact region (the enforced contact at the backwall) found for the reference steel 316R for which the wear pattern is scattered. The length from the cutting edge to the edge of each region in the middle part was measured, and thereafter the average value was calculated.

Typical long chips (Fig. 2) with a spirally increasing radius of curvature obtained during the cutting process were collected and used to investigate the non-metallic inclusion content using the EE method. The thicknesses of chips from different materials and cutting speeds were also measured using light optical microscopy (LOM).

Fig. 2.

Collected chips from the orthogonal cutting process of different materials at different speeds: (a) 316R at V_c = 100 m/min; (b) 316Ca at V_c = 100 m/min; (c) 316R at V_c = 250 m/min; (d) 316Ca at V_c = 250 m/min.

2.3. Investigation of Inclusions in Steel and Chips

In this study, the electrolytic extraction was carried out to evaluate the characteristics of NMIs in the steel samples before machining tests and in chip samples. The electrolytes named 10% AA (10% acetylacetone – 1% tetramethylammonium chloride - methanol) was used. The electrolytic dissolution was run with the following settings of electric parameters: a voltage of 4.2–4.4 V, an electric current of 60–70 mA, and an electric charge of 1000 coulombs for the 316R steel sample, and 500 coulombs for the 316Ca steel sample; furthermore, a voltage of 2.9–3.6 V, an electric current of 32–50 mA, and an electric charge of 80 to 120 coulombs for chip samples from both steels. After extraction, the undissolved inclusions in the solution were collected by filtration on a polycarbonate (PC) membrane film filter with an open pore size of 0.4 μm. Thereafter, the characteristics (morphology, size, number, and composition) of the extracted inclusions were investigated on the surface of film filters by using a scanning electron microscope (SEM) in combination with Energy-dispersive X-ray spectroscopy (EDS).

The investigations of NMIs were carried out using six samples: two metal samples from each original metal bars and four chip pieces obtained at different cutting speeds (V_c) of these two materials. To extract the aimed zone of the chip, the side opposite to the chip-tool contacting surface of the chip sample was covered by an unconductive hot melt glue, preventing the mixing of NMIs from different chip sides. Since the secondary deformation zone (Zone II in Fig. 1(a)) is where the chips glide across the rake face of the cutting tool, the NMIs in this zone were mainly investigated in this study. Thus, the depths of the dissolved steel layer (D_dis) for chips were controlled to an approximate value of 30 μm by adjusting the electric charge value based on the chip size. The weight of the dissolved metal chip samples varied from 0.017 to 0.026 g, depending on the electric charge value. Table 2 shows detailed information about the samples, parameters, conditions of electrolytic extraction and SEM investigations.

Table 2. Main parameters used in the electrolytic extractions and SEM investigations of non-metallic inclusions (NMIs) in different steels.

Steel Grade	Sample	Vc [m/min]	Ddis [μm]	Afil [mm2]	Wdis [g]	Aobs [mm2]	Observed NMIs Number n	Size Range [μm]
316R	316R		188	1200	0.1563	0.898	435	2–98
	316R-L	100	27	80	0.0177	0.125	218	2–24
	316R-H	250	33	80	0.0199	0.045	126	2–23
316Ca	316Ca		85	1200	0.0935	0.898	180	3–124
	316Ca-L	100	31	80	0.0215	0.090	153	2–32
	316Ca-H	250	34	80	0.0258	0.056	142	2–32

The length (l), width (w), and total surface area (A_NMI) of every investigated NMI were measured on the photos obtained by SEM after EE by using the Image-J image analysis software. The aspect ratio value (AR), the total surface area of NMIs per volume of steel (A_v), and the average surface area of individual inclusions (A_ave) were calculated by using the following equations:   

$$AR=l/w$$(1)

  

$$A_{v(Group\mathrm{\hspace{0.17em} } i,\mathrm{\hspace{0.17em} } Type\mathrm{\hspace{0.17em} } j)}=\frac{\sum A_{NMI(Group\mathrm{\hspace{0.17em} } i,\mathrm{\hspace{0.17em} } Type\mathrm{\hspace{0.17em} } j)}}{\frac{W_{dis}}{\rho }\times \frac{A_{obs}}{A_{fil}}}$$(2)

  

$$A_{v(Type\mathrm{\hspace{0.17em} } \text{j})}=A_{v(Group\mathrm{\hspace{0.17em} } \text{I,}\mathrm{\hspace{0.17em} } Type\mathrm{\hspace{0.17em} } j)}+A_{v(Group\mathrm{\hspace{0.17em} } \text{II,}\mathrm{\hspace{0.17em} } Type\mathrm{\hspace{0.17em} } j)}+A_{v(Group\mathrm{\hspace{0.17em} } \text{III,}\mathrm{\hspace{0.17em} } Type\mathrm{\hspace{0.17em} } j)}$$(3)

  

$$A_{ave(Group\mathrm{\hspace{0.17em} } i,\mathrm{\hspace{0.17em} } Type\mathrm{\hspace{0.17em} } j)}=\frac{\sum A_{NMI(Group\mathrm{\hspace{0.17em} } i,\mathrm{\hspace{0.17em} } Type\mathrm{\hspace{0.17em} } j)}}{n_{(Group\mathrm{\hspace{0.17em} } i,\mathrm{\hspace{0.17em} } Type\mathrm{\hspace{0.17em} } j)}}$$(4)

where A_{NMI(Group i, Type j)} is the total surface area of a Type j non-metallic inclusion with the morphology of Group i investigated on the filter, n is the number of measured inclusions, W_dis is the weight of dissolved sample during EE, and ρ is the steel density (~0.0078 g/mm³ in this study), A_fil is the area of the film filter for filtration (80 or 1200 mm² depending on applied filters), A_obs is the total observed and analyzed area on the film filter. When calculating A_v, two times of the NMI area measured on the figure (A_NMI) were considered as the total surface area of this NMI.

3. Results and Discussions

3.1. Classification of Non-metallic Inclusions in Steel and Chips

Three main types of non-metallic inclusions were found in the original Ca-treated steel 316Ca samples: 1) MnS inclusions (including pure MnS and some oxy-sulfide inclusions with some oxides as cores), 2) soft oxides (SO, which were significantly deformed during rolling), and 3) hard oxides (HO, which were not deformed during rolling), as shown in Table 3. The reference steel 316R only contained Type 1 sulfide inclusions and Type 3 hard oxide inclusions. However, despite that the 316R and 316Ca steels contained both Type 3) hard oxides, the compositions were quite different. In the 316R sample, the oxide composition was mainly Al₂O₃–MnO–MgO–TiO_x (AMnMT). However, the oxides found in the 316Ca steel (Type 2 and Type 3) consist of CaO–SiO₂–Al₂O₃–MgO–TiO_x (CSAMT, a composition between gehlenite and anorthite) and CaO–SiO₂–Al₂O₃–MgO (CSAM, a composition close to gehlenite).

Table 3. Classification of the inclusions observed in the reference steel sample (316R) and the Ca-treated (316Ca) stainless steel samples before machining.

The morphologies of non-metallic inclusions in chips obtained from the machining tests were quite different from the original inclusions present in the steel. These morphologies were investigated after EE by using SEM. As shown in Fig. 3, the NMIs observed in the chips can mainly be divided into three different groups: 1) NMIs of Group I having quite similar morphologies as the original inclusions; 2) Group II contains stretched inclusions which were stretched out to form a quite thin film (Group II-a) or were fragmented on small connected segments containing some fissures (Group II-b); and (3) Group III are brittlely fractured inclusions, which have a similar shape as the original inclusions but contained big cracks inside the inclusions. In the observed samples, only Group I and Group II inclusions were found for MnS and SO inclusions. In contrast, only Group I and Group III inclusions were found for HO inclusions. This phenomenon could be explained by the different softnesses of these inclusions at the high temperature during cutting. MnS inclusions (Type 1) is the softest inclusion type with the lowest hardness at room temperature so that it can be easily stretched. This can even lead to the formation of a quite thin film in some high-temperature locations, for instance, in the secondary deformation zone which is in contact with the tool. SO inclusions (Type 2) are not as soft as MnS inclusions, but also tends to be stretched in the secondary deformation zone during the cutting process. HO inclusions (Type 3) are the hardest oxides of the three types, which are not deformed even at high temperatures in the cutting region. Thus, these inclusions can keep their initial original shape (Group I) or be brittlely fractured at an extreme deformation of the steel matrix during the chip formation.

Fig. 3.

Typical morphologies of NMIs observed in the chips obtained from orthogonal cutting: (a) a Group II-a stretched MnS inclusion, (b) a Group II-b stretched SO inclusion, (c) a Group III brittlely fractured HO inclusion.

3.2. The Behavior of NMIs in the Different Steels and Cutting Speeds

Based on the classification listed before, the frequency distribution of different shapes and different types of NMIs are calculated and summarized in Table 4 and Fig. 4. It can be seen that for the reference stainless steel (316R), most NMIs (96–97%) are MnS inclusions. However, in the Ca-treated steel (316Ca), the MnS inclusions only contribute to 54–64% of the total amount of the investigated NMIs. Approximately 22% of the NMIs are hard oxides containing (Ca, Si, Al, Mg)O, and 24% are soft oxides containing (Ca, Si, Al, Mg, Ti)O.

Table 4. Distribution of different non-metallic inclusions in samples obtained from different cutting processes.

Sample*	Type of NMIs	Group I: Similar	Group II: Stretched	Group III: Brittlely fractured	Sum	RII/I
316R	MnS	96%	–	–	96%
	HO	4%	–	–	4%
316R-L	MnS	33%	64%	–	97%	2.0
	HO	3%	–	–	3%
316R-H	MnS	17%	79%	–	96%	4.2
	HO	2%	–	2%	4%
316Ca	MnS	54%	–	–	54%
	HO	22%	–	–	22%
	SO	24%	–		24%
316Ca-L	MnS	26%	37%	–	63%	1.4
	HO	15%	–	8%	23%
	SO	8%	6%	–	14%	0.8
316Ca-H	MnS	10%	48%	–	58%	4.9
	HO	15%	–	7%	22%
	SO	10%	10%	–	20%	1.1

*:  L sample is chip obtained at low cutting speed, and H sample is chip obtained at high cutting speed.

Fig. 4.

Frequency distribution of different shapes and different types of NMIs in different samples.

For the 316R steel, similar MnS inclusions as that found in the original sample (Group I) decrease from 33% to 18%, whereas the stretched MnS inclusions (Group II) increase from 64% to 78% as the cutting speed increase. As a result, the ratios of the frequencies between Group II and Group I (R_II/I) are 2.0 and 4.2 at low (V_c(L) = 100 m/min) and high (V_c(H) = 250 m/min) cutting speeds, respectively. Similar results were also found for MnS inclusions in the 316Ca stainless steel samples. The R_II/I values are 1.4 and 4.9 for low and high cutting speeds, respectively. The temperature near the chip-tool contact region is dependent on the cutting speed. At the high cutting speed, the MnS inclusions are softer than that at a low cutting speed, which results in a severer deformation of NMIs. In Ca-treated samples, both the 316Ca-L (63%) and 316Ca-H (58%) samples contain a larger amount of MnS inclusions compared to the original steel sample 316Ca, which has not been machined (54%). These results indicate that in both cases, broken MnS inclusions are created.

The total frequency of HO inclusions is constant before and after cutting in both the 316R and 316Ca steel samples. There is either no apparent difference between the samples using low cutting speeds and high cutting speeds with respect to the frequency of HO inclusions.

For the Ca-treated stainless steel, the frequency of oxides (both hard and soft) is much higher compared to the 316R reference steel sample. The soft oxides also show a small tendency to become stretched at the high cutting speed. The ratio of Group II and Group I inclusions increases from 0.8 at a low cutting speed to 1.1 at a high cutting speed. However, the SO inclusions have a much smaller ratio than the MnS inclusions, which could be explained by the much larger softness of MnS inclusions at the given cutting conditions.

Except for the frequency distribution of different NMIs, the total area (A_v) and average area (A_ave) of NMIs were also estimated to compare the deformation situation in different samples, as shown in Figs. 5 and 6. It is clear that the total area of MnS inclusions for the reference steel 316R is increased from a low speed (2.8 times compared to that in the original metal sample) to a high speed (3.8 times compared to that in the original metal sample). The total area of the MnS inclusions in chips of a 316Ca steel also rises to 2.8 and 3.8 times compared to the original sample at a low and a high cutting speed, respectively. The average areas of inclusions (A_ave) of Group I MnS inclusions for both the 316R and 316Ca samples are significantly smaller than that in the original steel sample before machining. This fact indicates that large size soft inclusions were more easily deformed to become Group II inclusions, as discussed in previous studies^26,27) that the relatively plastic depends on the size of MnS inclusions. It also shows a decreasing trend of A_ave of Group I MnS inclusions with an increasing cutting speed for both steels, which indicates that the critical size of a MnS inclusion to reach a high degree of deformation decreases with an increased cutting speed. This may be caused by different temperatures of chip-tool contact zone at different cutting speeds. However, the A_ave value of Group II MnS inclusions dropped slightly or similarly at a higher cutting speed for both the 316R and 316Ca steels. In this case, the significantly increased total area of MnS inclusions in Fig. 5(a) can be explained by the decreased critical size for deformation in chips. This, in turn, leads to the formation of more small size inclusions to become Group II inclusions. Though the original 316Ca steel contains fewer MnS inclusions than the 316R steel with respect to both frequency and A_v, they showed a similar tendency in both steels at the different studied cutting speeds.

Fig. 5.

Total are of inclusions per mm³ of different kinds of NMIs depending on materials and cutting speed: (a) MnS inclusions; (b) SO inclusions.

Fig. 6.

The average area of inclusions of different kinds of NMIs depending on materials and cutting speed: (a) MnS inclusions; (b) SO inclusions.

The A_v value of the SO inclusions in 316Ca chips (Fig. 5(b)) obtained at the low cutting speed is only slightly larger (1.1 times) than that in the original sample. In contrast, the A_v value of the SO inclusions at a high cutting speed is 3.5 times that of the original steel samples before machining. Moreover, the average area (A_ave) of stretched soft oxides (Group II) increased by almost two times at the high cutting speed compared to the low cutting speed. It indicates that the SO inclusions are quite hard at low cutting speeds so that the deformation rate is low. However, when the cutting speed was set to 250m/min, the degree of deformation of SO inclusions was remarkably improved. The average area (A_ave) of Group I SO inclusions are quite similar at both cutting speeds, but both values are much smaller than that in the original 316Ca steel sample. This fact shows that the small size SO inclusions are more likely to keep the original shape or a similar shape no matter which cutting speed it is used. According to the A_v and A_ave results quantitatively determined for different groups of various types of inclusions, a qualitative picture of the deformation behavior has been obtained.

According to the predicted diagram for the different cutting feeds and cutting speeds obtained by Kara et al.²⁸⁾ by using an artificial neural network, the highest temperature in the contact zone during orthogonal turning tests with a Ti(C,N)+Al₂O₃+TiN coated tool using a 316L stainless steel as the workpiece material was predicted to be 790 and 1066°C at the cutting speeds of 100 m/min and 250 m/min, respectively. Since few previous studies directly focus on the deformation of NMIs during cutting, previous studies concerning rolling, which is also a process where NMIs are deformed due to the steel matrix deformation, were referred to in this study. The hardnesses for phases of MnS, gehlenite (2CaO·Al₂O₃·SiO₂), and anorthite (CaO·Al₂O₃·2SiO₂) at room temperature are HV170 (MnS), HV1200 (gehlenite), and HV850 (anorthite), respectively.²⁹⁾ Though as the temperature is increased, the hardness of all three phases will decrease,³⁰⁾ the relative plasticity of these inclusions (ratio between the average effective strain of the inclusion and the average effective strain of the steel matrix) has different tendencies. As reported by Segal,²⁷⁾ the relative plasticity of MnS inclusions decreases from a 400°C temperature during rolling. However, silicate inclusions will first act as a rigid particle at relatively low temperatures, but thereafter act as plastic particles after reaching a narrow transition temperature range (also called the softening period, where the start temperature differed from 700°C to 1000°C depending on the composition and rolling conditions).^26,31,32) In this study, the HO inclusions did not reach the transition temperature region at the studied cutting speeds. So they have only a brittlely fractured morphology (Group III) or the same morphology as the original ones (Group I). However, for SO inclusions that have a lower melting temperature and hardness, it seems that the temperature of chips during the cutting process reaches the transition temperature range where the inclusions have a relative plasticity value between rigid particles (like HO inclusions) and high-degree plastic particles (like MnS inclusions). Thus, the SO inclusions after machining contain some inclusions with a fragmented morphology on small connected segments containing some fissures (Group II-b) and those having a similar morphology as the original ones (Group I). The SO inclusions in chips obtained at a high-speed cutting, which is suggested to have a higher maximum temperature above 1000°C at the chip-tool surface, shows a higher degree of deformation than those SO inclusions which have been machined at a low cutting speed (the maximum temperature is estimated to be about 790°C).

3.3. Results of Tool-chip Contact Length in the Cutting Tests of the Different Steels

A summary of the tool-chip contact length of 316R and 316Ca steel at both cutting speeds are given in Table 5. Though the composition and steel production process are almost the same for the 316R and 316Ca steels, the tool-chip contact length is quite different at the same machining conditions. The results show that the 316R reference stainless steel sample gives a longer (477 μm) secondary region (L_c2) dominated by sliding friction compared to the Ca treated 316Ca stainless steel (383 μm). Also, a stochastic tool-chip contact further away from the dense secondary region up to 1429 μm is visible in the 316R reference sample. However, the 316Ca stainless steel sample has a longer first region (617 μm) than the 316R steel (485 μm), which indicates that the 316Ca steel experiences more adhering problems at the low cutting speed compared to the 316R steel. This fact could be explained by that the 316R steel samples contain 96% MnS inclusions, whereas the 316Ca samples contain 56%. During the low temperature, the MnS inclusions will be softer than oxides, which may be beneficial for the deformation and sliding of metal layers (for instance, in the secondary deformation zone) in chips during machining.

Table 5. Summary of the tool-contact length and thickness data [μm] on the tool after cutting.

Sample	Length	316R		316Ca
Speed [m/min]		100	250	100	250
First region	Lc1	485±2	546±19	617±5	515±18
Secondary region	Lc2	477	161	373	267
	Lt	962±46	707±16	990±39	782±26
Stochastic contact region	Lt2	1429±18	1386±104	Not observed
Chip thickness	t	877	922	909	814

At a high cutting speed, the total tool-contact length (L_t) decreased with an increased cutting speed for both the 316R (on 255 μm) and 316Ca (on 208 μm) sample, which agrees with previously reported results.^19,20) It is also observed that the 316Ca has a longer (267 μm) secondary region (L_c2) than the 316R steel (161 μm), but the stochastic tool-chip contact further away remains in the case of the 316R steel up to a distance of 1386 μm. Besides, at this high-speed cutting (250 m/min), L_c1 of the first region of the 316R steel is slightly larger than that of the 316Ca steel. This fact indicates that 316Ca steel performs better in the high-speed cutting region when the temperature is higher than that during cutting at a low speed. Simultaneously, the 316R steel has a longer adhering region at a high cutting speed (546 μm) than at a low cutting speed. It might be explained by the increased amount of stretched soft oxides at a higher temperature during cutting of the 316Ca steel at the high speed of 250 m/min, whereas the 316R steel only contains hard oxides.

The thicknesses of the chips were also measured and the data are presented in Table 5. It has a similar tendency as for the length of the first region, especially for the same material or at the same cutting speed. Since the chip thickness and first region length are usually directly related to the cutting force, they would be a good indicator of the cutting force applied during machining. The results show that at a low cutting speed, the 316R steel was preferred, whereas 316Ca steel was preferred when using a high cutting speed of 250m/min.

The 316Ca steel, which has been Ca-treated, did not have a clear stochastic tool-chip contact on the rake face at both the low and high cutting speeds. It indicated that the chips from the 316Ca steel were more easily to curl. Thus, they have a less stochastic contact during cutting. On the other hand, the first region in the tool-chip contact area of the 316Ca sample at a low cutting speed is longer than that of the 316R sample. Furthermore, the chip thickness is also larger, which means that a larger cutting force is needed to remove the chips from the workpieces. It shows that the Ca treatment was not beneficial for a low cutting speed with respect to the tool-chip contact length. This phenomenon could be explained by the relatively lower plasticity and greater strength of MnS inclusions at low cutting speeds, which could be beneficial for tools due to the formation of a protecting layer on the tool surface to decrease the abrasive wear. However, the SO inclusions are near the rigid state so that they cannot help to protect the tools at low cutting speeds. These results are in agreement with the results from some previous research,^14,33) showing that MnS inclusions are beneficial for machinability at low cutting speeds. However, when the materials were cut at a high cutting speed, the 316Ca sample shows a more considerable decrease than the 316R sample in the first region and of the chip thickness compared to a low cutting speed. Since the temperature at the chip-tool surface is up to about 1000°C, MnS inclusions become so soft that they easily can be torn away during cutting. On the other hand, the SO Ca-containing inclusions can be continuously sliced off when being in contact with the tool, making it possible to form a protective layer on the tool surface that leads to a decreased abrasive wear. So the cutting force to remove the chips from workpieces for a 316Ca sample is smaller than that for a 316R steel. Thus, the energy needed for cutting is also less. It could be expected that the cutting temperature of 316Ca steel will be lower than that of a 316R steel, which also has been reported by Saoubi.³⁴⁾

The role of different NMIs in machinability could be suggested to take place in two ways based on the above results and previous results:^{1,2,3,4,5,6,7,8,9,10,11,12,13,14)} 1. A formation of a protective layer on the tool surface to prevent adhesion wear and chemical wear, such as SO inclusions at a 250 m/min speed and MnS inclusions at a 100 m/min speed in this study; 2. Acting as lubricants to decrease the abrasive wear, like MnS inclusions at a 250 m/min speed in this study. However, they will not work as a protective layer due to their too low strengths.

4. Conclusion

The focus of this study was to determine the effect of non-metallic inclusions (NMIs) in two types of stainless steels on the machinability namely a regular 316L stainless steel (316R) and a Ca-treated stainless steel (316Ca). The characteristics of NMIs in steel samples and chips have been systematically investigated and compared using the electrolytic extraction (EE) method combined with SEM+EDS investigations. The main results are summarized as follows:

(1) The electrolytic extraction (EE) method was successfully applied for 3D investigations of morphologies, sizes, and surface areas of fragile extreme thin film-like or fragmented stretched inclusions. This information cannot be obtained by using conventional 2D investigations on polished metal surfaces. The quantitative information of the characteristics of NMIs in the secondary deformation zone will help to study the roles of different NMIs during machining.

(2) Soft NMIs in stainless steel were significantly deformed during machining. The total areas of sulfide inclusions in the secondary deformation zone of the chip increased significantly (up to 2.8–3.8 times) compared to the areas found in the reference steel sample before the cutting test for both steels. Moreover, at high cutting speeds, soft oxide (SO) inclusions also have a total area that is 3.5 times of the soft oxide (SO) inclusions’ total area in the steel samples before cutting. However, at low cutting speeds, the soft oxide (SO) inclusions have a similar total area and average area than that found in the steel sample before machining.

(3) The degree of deformation of MnS and soft oxide NMIs in the chips depends on the workpiece material as well as the cutting speed. The frequencies, average and total areas of these different NMIs are affected by the temperatures and deformation degrees of the metal matrix during machining. The total surface area of NMIs per volume of steel (A_v) and the average surface area of individual inclusions (A_ave) values of SO inclusions increase obviously (2–3 times) with cutting speed increase. Sulfides also have an increased A_v value (+43%) at the higher cutting speed compared to at the low cutting speed. As a result, the same inclusions can have different behaviors and roles during machining.

(4) As the results of the chip-tool contact length show, the modification of NMIs for machinability improvements is only beneficial in some machining processes. The 316R steel was preferred at low cutting speeds, whereas the 316Ca steel was preferred at high cutting speeds. An appropriate application range of cutting parameters is also essential for manufactures to reach a low cost and high efficiency.

Acknowledgments

The authors wish to thank Jan Haraldsson and Olle Sundquist of Sandvik Materials Technology for fruitful discussions during the project, and Nils Stavlid and Thomas Björk for the support to carry out this study. H. Du also acknowledges the financial support from the China Scholarship Council (CSC), Jernkontoret and Stiftelsen Prytziska Fonden.

References

• 1)   N.  Ånmark,  A.  Karasev and  P. G.  Jönsson: Materials, 8 (2015), 751. https://doi.org/10.3390/ma8020751
• 2)   E.  Uhlmann,  A. M.  Zanatta,  F.  Mahr and  J. O.  Gomes: Int. J. Adv. Manuf. Technol., 68 (2013), 2451. https://doi.org/10.1007/s00170-013-4835-y
• 3)   T.  Akasawa,  H.  Sakurai,  M.  Nakamura,  T.  Tanaka and  K.  Takano: J. Mater. Process. Technol., 143–144 (2003), 66. https://doi.org/10.1016/s0924-0136(03)00321-2
• 4)   X. D.  Fang and  D.  Zhang: Wear, 197 (1996), 169. https://doi.org/10.1016/0043-1648(96)06924-4
• 5)   Y. N.  Wang,  J.  Yang and  Y. P.  Bao: Metall. Mater. Trans. A, 46 (2015), 281. https://doi.org/10.1007/s11661-014-2596-3
• 6)   H. S.  Qi and  B.  Mills: Wear, 198 (1996), 192. https://doi.org/10.1016/0043-1648(96)80023-8
• 7)   N.  Ånmark,  T.  Björk,  A.  Ganea,  P.  Ölund,  S.  Hogmark,  A.  Karasev and  P. G.  Jönsson: Wear, 334–335 (2015), 13. https://doi.org/10.1016/j.wear.2015.04.008
• 8)   Z.  Li and  D.  Wu: J. Mater. Sci. Technol., 26 (2010), 839. https://doi.org/10.1016/s1005-0302(10)60134-x
• 9)   L.  Jiang,  K.  Cui and  H.  Hänninen: J. Mater. Process. Technol., 58 (1996), 160. https://doi.org/10.1016/0924-0136(95)02144-2
• 10)   P.  Leskovar and  J.  Grum: CIRP Ann., 35 (1986), 537. https://doi.org/10.1016/s0007-8506(07)60199-2
• 11)   G. M.  Faulring and  S.  Ramalingam: Metall. Trans. A, 10 (1979), 1781. https://doi.org/10.1007/bf02811716
• 12)   L.  Sidjanin and  P.  Kovac: Mater. Sci. Technol., 13 (1997), 439. https://doi.org/10.1179/mst.1997.13.5.439
• 13)   P.  Hoier,  A.  Malakizadi,  S.  Friebe,  U.  Klement and  P.  Krajnik: Wear, 428–429 (2019), 315. https://doi.org/10.1016/j.wear.2019.02.024
• 14)   A.  Larsson and  S.  Ruppi: Mater. Sci. Eng. A, 313 (2001), 160. https://doi.org/10.1016/s0921-5093(01)00964-9
• 15)   H.  Roelofs,  A. M. G.  Boeira,  R.  Margot,  J. T.  Gomes and  M.  Eglin: Proc. 8th Int. Conf. on Advanced Manufacturing Systems and Technology, International Centre for Mechanical Sciences, Udine, (2008), 1.
• 16)   X.  Zhang,  S. N.  Lekakh,  R. J.  O’malley and  E.  Scheid: ISIJ Int., 60 (2020), 782. https://doi.org/10.2355/isijinternational.ISIJINT-2019-476
• 17)   A.  Nordgren and  A.  Melander: Mater. Sci. Technol., 5 (1989), 940. https://doi.org/10.1179/mst.1989.5.9.940
• 18)   L. H.  Huang,  J. C.  Chen and  T.  Chang: J. Ind. Technol., 15 (1999), 1.
• 19)   M. I.  Sadik and  B.  Lindström: J. Mater. Process. Technol., 37 (1993), 613. https://doi.org/10.1016/0924-0136(93)90122-m
• 20)   S. A.  Iqbal,  P. T.  Mativenga and  M. A.  Sheikh: Int. J. Adv. Manuf. Technol., 42 (2009), 30. https://doi.org/10.1007/s00170-008-1582-6
• 21)   A. K.  Balaji,  G.  Sreeram,  I. S.  Jawahir and  E.  Lenz: CIRP Ann., 48 (1999), 33. https://doi.org/10.1016/s0007-8506(07)63126-6
• 22)   X. F.  Bi,  G.  Sutter,  G.  List and  Y. X.  Liu: Mater. Sci. Forum, 626–627 (2009), 71. https://doi.org/10.4028/www.scientific.net/MSF.626-627.71
• 23)   N. A.  Abukhshim,  P. T.  Mativenga and  M. A.  Sheikh: Proc. Inst. Mech. Eng., Part B, 218 (2004), 889. https://doi.org/10.1243/0954405041486064
• 24)   E. M.  Trent: Wear, 128 (1988), 29. https://doi.org/10.1016/0043-1648(88)90251-7
• 25)   E. M.  Trent: Wear, 128 (1988), 47. https://doi.org/10.1016/0043-1648(88)90252-9
• 26)   H.  Iwai,  B.  Tsujino,  S.  Isa and  T.  Ao: Tetsu-to-Hagané, 55 (1969), 887 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.55.10_887
• 27)   A.  Segal and  J. A.  Charles: Met. Technol., 4 (1977), 177. https://doi.org/10.1179/030716977803292187
• 28)   F.  Kara,  K.  Aslantaş and  A.  Çiçek: Appl. Soft Comput., 38 (2016), 64. https://doi.org/10.1016/j.asoc.2015.09.034
• 29)   V.  Ollilainen: Proc. Swedish Symp. on Non-Metallic Inclusions in Steel, Uddeholms AB, Hagfors, (1981), 429.
• 30)   K. B.  Gove and  J. A.  Charles: Met. Technol., 1 (1974), 279. https://doi.org/10.1179/030716974803287834
• 31)   C.  Luo and  U.  Ståhlberg: J. Mater. Process. Technol., 114 (2001), 87. https://doi.org/10.1016/s0924-0136(01)00576-3
• 32)   T. J.  Baker,  K. B.  Gave and  J. A.  Charles: Met. Technol., 3 (1976), 183. https://doi.org/10.1179/030716976803391656
• 33)   H.  Opitz and  W.  König: Advances in Machine Tool Design and Research 1967, ed. by S. A. Tobias and F. Koenigsberger, Pergamon Press, Oxford, UK, (1968), 173. https://doi.org/10.1016/B978-0-08-003491-1.50013-7
• 34)   R.  M’Saoubi and  H.  Chandrasekaran: Int. J. Adv. Manuf. Technol., 56 (2011), 865. https://doi.org/10.1007/s00170-011-3257-y