Interfacial Bonding Properties Experimental Research of 316L Stainless Steel–Carbon Steel Clad Rebar in the Process of Intermediate and Finish Rolling

Abstract

The interfacial bonding properties of stainless steel clad (SSC) rebars determine whether they can be widely used. In the industrial production of SSC rebars, the process of intermediate and finish rolling of the microstructure evolution, element diffusion behavior, and interfacial bonding properties of bimetallic interfaces are investigated. In this paper, 316L seamless stainless steel (SS) tube and HRB400E carbon steel (CS) bar were prepared by a vacuum oxidation-free composite round billet, and the industrial emergency stopping of SSC rebars’ hot rolling was carried out. The metallographic results showed that the thicknesses of the carburized austenite zone (CAZ) varied greatly (832–238 μm) and showed a parabolic downward trend, while the thicknesses of the decarburized ferrite zone (DFZ) varied little (85–99 μm). The elemental line scans showed that Fe and Cr had the same parabolic downward trend. The intermediate-rolling had a great influence on element diffusion, and, in S6–9, the diffusion distance of Fe and Cr decreased significantly. The diffusion distances of the elements in the intermediate-rolling back stage and finishing-rolling front stage (S9–12) were basically balanced. The elemental diffusion distances and interfacial bonding strength were not consistent. Among them, the shear strength (τ) of S13 was 410.7 MPa. Compared with ordinary rebars, the yield strength (Re) and tensile strength (Rm) of finished SSC rebars were increased by 7.05% (30.9 MPa) and 7.10% (43.0 MPa), respectively. The tensile properties exceed those of mixture effects. The paper provides a theoretical basis for the improvement of the interfacial bonding strength and optimization of the rolling process system for the industrial production of SSC rebars.

1. Introduction

Bimetallic composite steel is processed with two or more metals to combine the advantages of their respective component metals to form a strong metallurgical bond [1]. In recent years, stainless steel composite (SSC) rebars, which are composed of clad stainless steel (SS) and parent carbon steel (CS), are produced by the hot rolling of composite round billets (CRB). SSC rebars have high strength, high toughness, corrosion resistance, and other excellent properties, and can be used for tunnels, bridges, ports, and other major projects with a high demand for safety and durability to cope with the compound effect of extreme heat, force, matter, chemical, and other multi-energy fields in the environment. The price is about one-third of the pure SS rebars, and they have a broad application prospect [2,3,4,5,6,7]. Some scholars have studied the preparation of SSC rebars, such as STELAX (UK) et al., which adopted the recycling iron scrap method [8,9,10], Gao et al., who adopted cold drawing-high temperature brazing technology [11], Liu et al., who adopted metal deposition and hot rolling [3,12], and WU and Yu W et al., who adopted welding-vacuum treatment [13,14]. The billets are prepared by the above methods, and bimetallic composite forming is realized by hot rolling. At present, the research of SSC rebars mainly focuses on the billet formation process, rolling pass sequence, interface microstructure, element diffusion, and tensile strength of rolled finished products. For example, William M. Cross et al. studied the corrosion resistance and mechanical properties of SSC rebars as well as the size and shape of the cladding [2]. Xie et al. used 304 austenitic SS seamless pipes and Q195 CS bars to form billets, then draw billets and carried out six-pass hot rolling, to study its mechanical properties and interfacial bonding [15]. Li et al. proposed a method for the preparation of 55#-316L rebar based on vacuum hot rolling [6].

For the cladding and parent metals, close metallurgical bonding is the key to ensure that SSC rebars obtain excellent properties during hot rolling. In the process of the hot rolling of SSC rebars, the bimetallic bonding interface presents a complex microstructural form and the element diffusion is inconsistent. In previous studies, due to the presence of a chemical formula gradient, elements such as Fe, C, Cr, and Ni would diffuse at the interface during hot rolling, forming a composition diffusion zone [16,17]. The uphill diffusion of C elements from CS to SS generated a decarburized ferrite zone (DFZ) and a carburized austenite zone (CAZ) in CS and SS, respectively. There were coarse ferrite grains in DFZ, and carbides could be discovered in CAZ grain boundaries. Carbides were mainly the fracture source of the interface under loading. The DFZ and CAZ reduced the interfacial bonding strength and corrosion resistance of bimetallic composite steels, but the sufficient diffusion of Cr and Ni elements could effectively limit the occurrence of interfacial delamination and improve the interfacial bonding strength [18].

In the process of the large-scale industrial hot rolling of SSC rebars, the changes in the microstructure and interfacial bonding properties of bimetallic interfaces of intermediate and finish passes were not obvious. In this paper, a 316L SS seamless tube and HRB400E CS were used to prepare vacuum oxidation-free CRB, and the industrial emergency stopping of SSC rebars’ hot rolling was carried out to obtain samples of intermediate and finish rolling passes. Then, the microstructure evolution, element diffusion behavior, and interfacial bond strength of the bimetallic interface were carried out, and the strengthening and toughening mechanism of interfacial bonding during intermediate and finish rolling were explained. This study can further the interfacial bonding strength enhancement and provide a theoretical basis for improving the hot rolling process.

2. Materials and Methods

2.1. Material Selecting and Fabrication Processing of CRB

The cladding material used in this paper is 316L SS seamless pipe (size of ϕ159 × 9000 mm, 7 mm of thickness), and its inner surface is cleaned until smooth and free of debris. The base material is HRB400E round rod (size of ϕ155 × 9000 mm), machined to ϕ144.90 mm by centerless lathe (WXC200B), removing the peripheral surface burrs and achieving a surface roughness of about 2.3 μm. The nominal composition of the two materials is shown in

Table 1. The materials nominal composition (in wt%, adapted from [19,20]).

Elements Wt (%)	C	Si	Mn	P	S	Ni	Cr	Mo	Fe
316L SS pipe	0.03	1.00	2.00	0.045	0.030	10.00–14.00	16.00–18.00	2.00–3.00	Bal.
HRB400E CS round bar	≤0.25	0.80	1.60	0.045	0.045	-	-	-	Bal.

. Figure 1 shows the schematic diagram of the pre-treated SS pipe and CS rod grouping CRB. SS pipe and CS rod were placed vertically by crane, gravity-nested combination was adopted, and the SS ends were welded by plasma fusion technology. Additionally, a small hole was retained, vacuum-pumped to 10⁻² Pa without oxidation, and then sealed.

2.2. Industrial Field Experiment of Hot Rolling

The experiment was carried out in the rolling line of the third rebar mill of Guangxi Liuzhou Iron and Steel Group Co., Ltd., Liuzhou, Guangxi, China. The production line mainly consists of heating furnace, rolling mills, flying shears, and cooling bed. There are 18 rolling mills in the rebar rolling line, with 6 passes each for rough, intermediate, and finish rolling, and the No.1 flying shear and No. 2 flying shear are located between the 6–7# and 12–13# mills, respectively. The experiment rolled φ28 mm straight thread SSC rebars and the CRB through a total of 14 passes, numbered S1–14, in which the intermediate-rolling 7–8# mills and the finish rolling 15–16# mills were, respectively, dumming. Among them, 1–2# mills adopted box holes and the rest adopted elliptical–circular pass sequence. Figure 2a shows the distribution diagram of SSC rebar rolling line.

The heating temperature was 1250–1280 °C, the CRB heating time was 3 h, and the primary rolling temperature was greater than 1060 °C. As shown in Figure 2b, the CRB come out the heating furnace and started rolling. CRB primary rolling temperature, S1–10 temperature, S13 temperature, and cold bed temperature were measured, and the temperature curve was shown in Figure 2c. When the central control station 5# mill receives the signal, the whole line stopped in emergency. The CRB stopped between the 6# mill and cooling bed, and the partially finished SSC rebars went into cooling bed. Immediately cut the samples on site using an oxygen-acetylene gun, labelled S6–14. As shown in Figure A1, one rough rolling sample (S6), four intermediate-rolling (S7–10), and four finish-rolling (S11–14) samples, a total of nine samples, ranging in length from 0.63–2.38 m, were air cooled to room temperature (Figure 2a).

2.3. Characterization Test

For the S6 sample shown in Figure 3a, the specimen with a thickness of 10 mm was obtained along the ND-TD section using an EDM wire cutter, which was ground, polished, etched (the etching solution was configured from 4% nitric acid and alcohol, and the etching time was about 1 min), and cleaned. The S6–14 images were captured by an image sensor in a dark box with a strip uniform light source, and the cladding uniformity and interfacial bonding were observed and analyzed. The height (H), width (B), and the cladding thickness of the ND-TD section were measured by a vernier caliper.

As shown in Figure 3a, two specimens with a length, width, and thickness of 10 × 20 × 10 mm were cut in the center of the sample ND, and their ND-TD surfaces were ground and polished. The SS side of the first specimen was corroded with austenitic SS corrosion solution (hydrofluoric acid, hydrochloric acid, nitric acid, and water = 1:2:3:4) for 4–5 min. The second specimen was corroded with 4% nitric acid and alcohol solution in the CS side for 5–8 s. The corrosion results of S6 specimen were shown on the right side of Figure 3a. The optical microscope DSX500 (Olympus Corporation, TYO, JPN) was used to observe the interface metallographic microstructures, and the Energy-Dispersive Spectrometer Ultim Max 40 (EDS, Oxford Instruments, Oxford, UK) was used to scan both sides of the ND-TD interface of the specimens by point, line, and surface, in which line scans were performed three times.

Tensile shear specimens were applied to assess the interfacial bonding strength and the fracture behavior of metallurgical bonding interface. As shown in Figure 3b, two shear tensile specimens were obtained, respectively, at S6–14 samples in the center of the reduction and diffusion direction of ND-TD along the RD direction. The interfacial bonding lap zone the length and width were 3.00 and 1.50 mm, respectively [10,15,16]. A total of 36 specimens were obtained. The tensile shear specimens were subjected to uniaxial tensile testing on a tensile machine (MTS Exceed E45) with a strain rate of 1 mm/min. To measure the tensile properties of the finished SSC rebars (S14), four specimens with length of 450 mm were selected, and the ∅28 mm ordinary HRB400E rebars rolled in the same line were selected as the control group. The uniaxial tensile tests of the finished products were carried out on a tensile machine (MTS 322) with a strain rate ranging from 4–35 mm/min. The fracture morphology of tensile shear specimens and the finished products were observed by scanning electron microscope MIRA4 LMH (SEM, TESCAN, Brno, Czech Republic).

3. Results

3.1. Sectional Images and Rolling Reduction Ratio

In order to observe and analyze the bimetal interfacial bonding, as well as determine the cladding thickness and the reduction ratio, Figure 4 showed the ND-TD cross-section image of S6–14 specimens and the ND-RD section image of S14 specimen. The bimetal interface of S6 was not found to be an unbonded phenomenon, and no obvious voids or cracks were found, indicating that the interface of S6 basically formed a strong metallurgical bonding interface. With the increase in deformation, the cladding of the subsequent passes gradually became thinner without leakage, and the thinnest cladding appeared at the root of S14 transverse ribs (thickness of 0.43 mm), which was higher than the GB1499.2–2018 [20] minimum thickness of 0.18 mm. The cladding thickness of S6–14 was thinner in the decreasing direction of the roll contact area than in the expanding direction of the roll seam. Due to the rebar forming process, the plasticity of SS was better than that of CS, and the outer surface strain was higher than the core strain [21], resulting in a large plastic deformation in the roll contact area. As shown in

Table 2. The total and relative rolling reduction (R_tot, R_rel, %).

Reduction	Direction	S6	S7	S8	S9	S10	S11	S12	S13	S14
Rtot 1	Vertical	53	-	63	-	73	-	78	-	83
	Horizontal	-	72	-	80	-	81	-	88	-
Rrel 2	Vertical	-	-	37	-	39	-	25	-	36
	Horizontal	-	43	-	40	-	23	-	39	-

¹ R_tot = (H_i − 159)/159 × 100%, (i = 6–14); ² R_rel = (H_j − H_j−1)/H_j × 100%, (i = 7–14).

, the total rolling reduction (R_tot) and the relative rolling reduction (R_rel) of S6–14 were calculated by H and B. The R_tot fluctuated, and R_tot was 53% after roughing rolling (1–6# mills) and R_tot was 83% of the finished product of S14. The R_rel ranged from 23–43%, with S7 having the largest R_rel of 43% and S11–12 having a smaller R_rel of 23, 25%, respectively. The cladding thicknesses of S11–13 varied from 3.62–1.42 mm, and the distribution of S14 cladding was uneven, with 0.64 mm at the root of the transverse rib and 1.74 mm near the longitudinal rib.

3.2. Interfacial Morphologies

In order to observe the microstructure information on both sides of the interface, Figure 5 showed the interfacial metallographic micrographs of S6 (200× and 800×), which was similar to the carbon-generated DFZ and the SS-generated CAZ at the bonding interface of the SSC bar stocks [22] and the clad plates [16,23,24,25]. The diffusion process resulted in the occurrence of the microstructure and composition gradients. As shown in Figure 5a, the 316L SS cladding was almost not corroded and there were fewer austenitic grains far from the interfacial bonding (Zone I), while a wide CAZ appeared near the interface of 316L SS (Zone II) with clear, regular, and straight grain boundaries, and smaller grains were located at the high-angle grain boundaries (HAGBs) of the larger grains. There were a wavy or jagged black band at the interface, which was wavy and obviously not smooth (Figure 5b2, red circle), indicating that it was a high-hardness band (HHB) with rapid changes in elemental composition formed during metal solid-state diffusion bonding [26]. On one side of HRB400E near the interface, there was a coarse grain layer composed of pure ferrite with a certain thickness (Zone IV). The core organization was a mixture of ferrite and pearlite. The ferrite grains were coarse, the grain boundaries were uneven and a few places were unclear (Figure 5b). As shown in Figure 5, both sides of the interface were divided into five zones [27,28]: Zone I was the 316L SS cladding (from the outer surface of 316L SS to the CAZ boundary), Zone II was the CAZ, Zone III was the HHB, Zone IV was the DFZ, and Zone V was the HRB400 substrate (from the DFZ boundary to the CS core).

Figure 6 showed the metallographic micrographs of the S7–14 interface (800×). In the process of S6–14, the HHB gradually became smooth and delicate, and the tiny gap between the interfaces was eliminated by the rolling, and the residual inclusions were further fragmented and presented a diffuse state forming a tightly bonded metallurgical interface [23]. The CAZs of S6–14 were coarse equiaxed grains, and a few fine flat twinning crystals arising within the crystals, among which the twinning crystals of S8 was showed a higher result (Figure 6b) and the grains of S14 was fine and uniform (Figure 6h). The thickness of the CAZ was relatively wide, which was different from the studies of Liu et al. and Guo et al. in which the CAZ distribution was not significant (or the CAZ width was very narrow) [23,29]. For the DFZ of S6–14, the grain size and thicknesses had little overall variation (Figure 5b2 and Figure 6a2–h2). Because it did not contain impurities such as pearlite, oxides, and carbides, the growth of grains existed less impediment and the grains was difficult to refine, so the DFZ was also the region with the largest grain size [16]. The austenitic grain sizes of zone II were significantly smaller than the ferrite grain sizes in zone IV, and the grain sizes in S14 zones II and IV were 18.1 and 58.8 μm, respectively (Figure 6h,h2).

In addition, the DFZ of S13 was significantly different from the other passes in that it was a polygonal ferrite (PF) (Figure 6g2). This was because the spacings of the 1–16# mills in the steel mill were between 2.1–5.0 m, and the spacing of the 15–16# mills was about 58 m. The 15–16# mills were dumming, so the spacing of the 14–17# mills was greater than 65 m. The temperature of CRB showed an obvious drop through the 14# mill to before the 17# mill, and the temperature of CRB rose rapidly because of the 17# mill generating the deformation heat. The temperature of small cross-sectional area of S13 dropped faster after the emergency stop. At this time, a small number of organizations caused static recovery and static recrystallisation, and the PF preferentially nucleated at the original austenitic grain boundaries. Then, a large number of deformation bands existed in the uncrystallized zone, so the PF was generated homogeneously at grain boundaries and on deformation bands. After the emergency stop, the air cooling of S13 also had a partial effect, inhibiting the phase transition of PF.

According to the rolling temperature (Figure 2c), the intermediate and finish rolling were carried out in the austenitic recrystallisation zone, and the work-hardening (WH), the softening processes of dynamic restitution (DRV), and dynamic recrystallisation (DRX) would occur simultaneously, but the DRX grains had asynchronous properties on both sides of the interface. A large number of DRX grains of the low layer dislocation energy austenitic SS preferentially were generated along the grain boundaries through strain-induced mechanism, but the bias convergence of the more solute atoms and the precipitates at grain boundaries in SS exerted drag forces on the grain boundary bowing out, slowing down the process of DRX and making the degree of DRX not uniform. The temperature of the internal CS side was higher and the precipitates of the diffusely distributed was less, so the DRX preferentially started. When the $\gamma$-Fe changed to the $\alpha$-Fe with a high layer dislocation energy after the emergency stop and temperature drop, the DRV played a dominant role.

The interfacial inclusions affect the properties of the interface. As shown in Figure 7, the black inclusions of S14 interface were detected by EDS, and the results were shown in

Table 3. The point scans of point 1–5 (in wt%).

Point	C	O	Si	Cr	Mn	Ni	Mo	Fe
1	25.74	1.66	6.97	11.75	0.78	6.11	0.79	46.19
2	2.61	1.49	1.27	16.08	0.93	9.69	1.68	66.25
3	3.54	1.15	6.82	1.2	1.39	0.92	0	84.99
4	23.06	2.26	69.16	0	0	0	0	5.51
5	12.18	1.86	53.71	0	0.6	0	0	31.65

. The elements concentrations of C, Si, and Cr at point 1 were 25.74, 6.97, and 11.75%, respectively, while the O and Mn elements were less. During billet preparation, the O elements remained between the interfaces, and, during the rolling process, the O elements would react to form metal compounds. The maximum solubility of C is known to be 2.11% in austenite and 0.0008% in ferrite. The presence of many chromium-carbide-precipitated phases (Cr7C3 and Cr23C6) in the CAZ was confirmed in the studies of Liu B.X. et al. [16] and Mas et al. [30,31]. Masahiroe Nomura et al. and Li et al.’s studies [16,32,33,34] pointed out that in high-temperature and high-pressure rolling, the trace O elements remaining between the interfaces would first react to form Fe-O, and then Si and Mn elements with fast diffusion rate would react with Fe-O to produce Mn-Si-O oxide inclusion. Therefore, it was speculated that point 1 consists of more chromium-rich carbides, and a certain amount of Si-Mn-O mixed oxides, and excess Si atoms exist in the ferrite in the form of a solid solution. The low content of C and Si elements at point 2 indicates the presence of a small amount of Si-Mn-O mixed oxides and chromium-rich carbides, with the remaining Cr and Ni dissolved in austenite. At point 3, the C and Si contents were 3.54, 6.82%, respectively, and the contents of Cr and Ni were significantly lower than at points 1 and 2, indicating the presence of Si-Mn-O mixed oxides, free Si atoms, and small amounts of chromium-rich carbides. On the CS side, points 4 and 5 have Si contents of 69.16 and 53.71%, respectively, with no detected Cr and Ni contents, suggesting that these two locations consisted of a small amount of Si-Mn-O mixed oxides and a certain amount of cementite, with most of the Si element in solid solution within the ferrite, thereby strengthening the ferrite.

3.3. The Elements Diffusion Between Interfaces

For a quantitative analysis of interfacial element diffusion, as shown in Figure 8, S6, S10, S13, and S14 were selected to conduct EDS point scanning on the interface, and the results were shown in

Table 4. The point scanning of the S6 interface (in wt%, the bold of C element peak values).

Point	Fe	C	Cr	Ni	Mo	O
1	67.51	3.31	16.10	9.49	1.73	0.55
2	69.10	3.27	14.58	9.45	1.58	0.45
3	88.79	3.21	5.79	0.20	0.16	0.40
4	92.84	3.26	1.62	0.13	0.15	0.27
5	94.88	2.84	1.41	0.03	0.10	0.23

,

Table 5. The point scanning of the S10 interface (in wt%, the bold of C element peak values).

Point	Fe	C	Cr	Ni	Mo	O
1	67.48	4.14	15.95	8.82	1.61	0.70
2	68.90	4.00	14.42	9.30	1.14	0.63
3	83.70	4.50	7.74	1.62	0.23	0.57
4	92.46	3.61	1.44	0.35	0.00	0.36
5	93.99	3.21	0.14	0.22	0.00	0.47

,

Table 6. The point scanning of the S13 interface (in wt%, the bold of C element peak values).

Point	Fe	C	Cr	Ni	Mo	O
1	67.64	3.28	16.36	9.44	1.28	0.41
2	67.49	3.13	16.04	9.41	1.99	0.49
3	83.65	3.02	8.49	2.62	0.42	0.27
4	90.85	3.51	2.96	0.44	0.00	0.46
5	93.66	3.02	0.78	0.22	0.10	0.31
6	93.91	3.42	0.06	0.33	0.00	0.33

and

Table 7. The point scanning of the S14 interface (in wt%, the bold of C element peak values, the bold of Cr, Ni, Mo element significant change).

Point	Fe	C	Cr	Ni	Mo	O
1	67.26	3.67	16.20	9.16	1.75	0.55
2	67.60	3.24	16.07	9.79	1.35	0.45
3	88.02	3.17	6.12	0.74	0.03	0.32
4	91.32	3.63	2.43	0.39	0.00	0.41
5	93.79	3.45	0.45	0.15	0.00	0.36
6	93.67	3.57	0.12	0.21	0.28	0.43

. For SS and CS, the concentrations of Fe, Cr, Ni, Mo, and C differed greatly (Table 1), and elemental diffusion led to obvious fusion in the II–IV zones [15,35]. From CS to SS, Fe had a distinct concentration gradient at the interface. From SS to CS, the concentration gradients of Cr, Ni, and Mo changed obviously; the Cr concentration gradient was higher than that of Ni and Mo, and the concentration of the CS side dropped sharply, especially the concentration of Ni and Mo. As shown in points 2–4 of the finished S14 (Table 7 bold), the Cr concentrations were 16.07–6.12–2.43%, but the Ni concentrations were 9.79–0.74–0.39%, and the Mo concentrations were 1.35–0.03–0.00%. Due to the increase in R_tot, the extension of CRB was accompanied by a decrease in the thickness of the SS cladding and in the inner diameter of the CS core rod, so a new metallurgical bonding interface was formed on the fresh metal surfaces of the interfaces, increasing the actual metallurgical bonding area of the bimetal. At the same time, the composite time during the S6–14 rolling process was short, resulting in the diffusion distance of the elements becoming thinner and the concentration gradient increasing. Additionally, the Cr element and α-Fe atoms were similar in size and was mutually soluble, resulting in the Kirkendall effect [36], mainly through vacancy migration to form substituted solid solutions. As a result, the Cr concentration at the CS side interface was relatively higher, while Ni and Mo concentrations were lower.

In Table 4, Table 5, Table 6 and Table 7, from the CS to the SS, two peaks of C atoms appeared at point 3/4 and point 1 near the interface (bold), showing the clear phenomenon of ‘uphill diffusion’, which was different from the single peak measured by B.X. Liu et al. [15]. It is assumed that the phenomenon was caused by the following processes: (1) As C was an interstitial atom, the diffusion ability of C was much larger than that of metal elements, and the C content of CS (≤0.25%) was higher than that of 316L (≤0.03%) (Table 1). The interstitial diffusion of C occurred earlier than that of Fe, Cr, and so on, and the C close to the interface of the CS side entered into the SS side to form interstitial solid solutions with γ-Fe. Moreover, although Ni did not form carbides with C, its diffusion would increase the Q value of C, thereby promoting the diffusion of C. The Cr concentration of point 1 was high on the SS side away from the interface, and C diffused to this point to form a chromium-rich carbide, making the chemical potential of C decreasing. The Cr, Ni, and other replacement diffusion occurred subsequently, and C continued to diffuse to the point 1, so the peak of C was observed at point 1. (2) At the CS side near the interface, the concentration of C decreased, and the diffusion of Cr also formed chromium-rich carbides, which reduced the chemical potential. Therefore, the C of the CS side substrate would diffuse and aggregate towards the interface of lower chemical potential until the potential balance, where another peak of C appeared.

The thicknesses of the CAZ and DFZ were represented by $x_1$ and $x_2$, respectively (Figure 5a–a2), and the average values of the five-times measurements of S6–14 were shown in Figure 9. Overall, the $x_1$ dropped sharply from 832.4 μm to 237.8 μm, showing a parabolic downward trend. The intermediate-rolling drop was 476.4, which the S6–7 suddenly dropped to 573.6 μm and the S7–10 slightly increased and then dropped to 356.0 μm. The S11–14 of finishing rolling decreased from 371.6–237.8 μm, with a reduction of 133.8, which was much smaller than that of intermediate-rolling, and the S12–14 decreased less than that of the S11–12. The $x_2$ had little overall change ranging from 84.9–98.6 μm. Among them, the S6–8 slightly increased and then decreased, S8–13 continued to rise in a small range, and S13–14 significantly decreased. The diffusion driving force of C was expressed as $F=-\partial u_i/\partial x$; when the chemical potential difference was zero (${∆u}_i=0$), the diffusion stopped, and the substrate C concentration on the CS side of the interface remained basically unchanged, resulting in an essentially unchanged thickness in the DAZ, which existed in similarity with the literature where the thickness of the DFZ was kept at a constant value [15,23]. The $x_1$ of the CAZ thickness was represented by the diffusion distance X of C, calculated by the diffusion law.

The diffusivity of the atoms, expressed as a diffusion coefficient D, follows the Arrhenius equations [37,38]:

$$D=D_0\exp \left(-{\displaystyle \frac{Q}{RT}}\right)$$

(1)

where $D_0$ is the diffusion constant (m²/s⁻¹), Q is the diffusion activation energy (J/mol), R is the gas constant (8.31 J/(mol·K)), and T is the thermodynamic temperature (K).

The $D_0$ and $Q$ of C in γ-Fe of 316L SS are 10 times and 1.67 times that of α-Fe of CS, respectively [39,40]. The rolling heating and holding stage (3 h, 1250–1280 °C) suggested that, when the composite rolling started, the C has been diffusing over a long distance to form a thicker CAZ in 316L [35], which could be confirmed by the fact that the $x_2$ of S6 still reached 832.4 μm after six passes of rough rolling. The thickness of the intermetallic phases is directly proportional to the distance of atom diffusion, which can be expressed as follows:

$$X=k\ast \sqrt{Dt}$$

(2)

where k is the coefficient, X is the elemental diffusion distance (m), and t is the diffusion time (s). Collating Equations (1) and (2), the C diffusion distance at the end of heating and holding stage and the initial rolling was expressed as follows:

$$X_0=k\sqrt{D_0\exp \left(-{\displaystyle \frac{Q}{RT}}\right)t}$$

(3)

As the rolling kept pressing down, the $X_0$ of Equation (3) decreased. At the same time after the emergency stop, the C element would still diffuse and increase the CAZ thickness during the process of air cooling. Therefore, the X of diffusion distance consisted of two parts: the retained thickness after the $X_0$ rolling thinning in the heating and holding stage, and the diffusion thickness after the emergency stop $X_{6-14}$. At the same time, the larger the cross-section area is, the longer it took to reduce the temperature to room temperature, which caused the diffusion distance on the CAZ to increase. The ratio of the cross-section equivalent diameter (d) to the effective reference diameter ($d_0$) was used to express the section area affectivity, and the ellipse–circle geometric relationship was converted to obtain the equivalent diameter: $d=\sqrt{H\ast B}$. Calculate the diffusion distance $X_1$ for S6–14:

$$X_1=X_0{\left({\displaystyle \frac{d}{d_0}}\right)}^{{\gamma }_1}+X_{6-14}=k\sqrt{D_0\exp \left(-{\displaystyle \frac{Q}{R{T}_1}}\right) t_1}{\left({\displaystyle \frac{d}{d_0}}\right)}^{{\gamma }_1}+k\sqrt{D_0\exp \left(-{\displaystyle \frac{Q}{R{T}_2}}\right) \left[t_2\ast {\left({\displaystyle \frac{d}{d_0}}\right)}^{{\gamma }_2}\right]}$$

(4)

where $k, {\gamma }_1$ and ${\gamma }_2$ are constants; when the C diffuses in γ-Fe, $D_0$ is 2.0 × 10⁻⁵ m²/s and Q is 140 × 10³ J/mol; $d_0$ is the internal diameter of the finished S14 product measured at 27.74 mm; T1 of heating and holding stage takes the intermediate value of the temperature 1538 K and $t_1$ is 3 h; and T2 of the air cooling process after the emergency stop is the measured value (Figure 2c), where S11–12 is the estimated value of 1330 K and $t_2$ is the estimated value of 0.5 h.

The $x_1$ values measured by S6, S10, and S14 were selected to solve the system of nonlinear equations, which were successfully converged and solved: k = 0.12203, ${\gamma }_1$ = 1.2391, and ${\gamma }_2$ = 1.0539. In addition, the ∆E of the error correction term was also considered, and the sources were mainly as follows: there is a warming process in the temperature of heating and holding stage, which was time-consuming, and the Cr and Mo increased the Q of C in austenite, which hampered the diffusion of the C element, all of which made the calculated value large, so ∆E was taken as −50 μm. The value of $X_1-∆E$ was calculated, and the results are shown in Figure 9 and

Table 8. The experimental measured and theoretical calculated values of CAZ.

Pass	${\mathit{x}}_1$ 1	$({\mathit{X}}_1-∆\mathit{E}$) 2	Tolerance Value (%)
S6	832.4	840.2	0.9
S7	573.6	693.5	20.9
S8	606.8	577.5	4.8
S9	363.5	465.9	28.2
S10	356.0	383.1	7.6
S11	371.6	346.7	6.7
S12	272.5	308.8	13.3
S13	249.3	252.0	1.1
S14	237.8	228.7	3.8

¹ Experimental measured; ² Theoretical calculated.

. The deviations of S7 and S9 were more than 20%, the deviation of S12 was 13.3%, and the deviations of the theoretical calculated values ($X_1-∆E$) and experimental measured values ($x_1$) of the CAZ thickness for the other passes ranged from 0.9–7.6%, which were in very good agreement.

Further, in order to more intuitively study the element diffusion distance and degree, Figure 10 showed the results of the EDS line scans of S6, S10, S13, and S14 and the elemental surface scans of Fe, Cr, and Ni. A jump or a sudden drop to the equilibrium concentration of the Fe, Cr, and Ni elemental line scans on both sides of the interface was identified as the elemental boundary, and the spacing of the boundaries on both sides was used as a measure of the elemental diffusion distance [41]. In Figure 10a–d, the diffusion distances of Fe and Cr were far away and easy to observe, and there were large non-coincident zones (NCZs) for both [37]. Among them, S6 had the largest diffusion distance and small fluctuation (Figure 10a), S10 and S13 fluctuated sharply, and S14 was less fluctuating and relatively smooth, while the curve slopes of the diffusion distance middle part of Fe and Cr increased significantly. Moreover, the diffusion distance of Ni was much smaller than that of Fe and Cr and was not easily observed. On the SS side, the NCZ of Fe with Cr and Ni fluctuated and changed. On the CS side, the NCZ of Fe and Cr decreased continuously, and the NCZ of Fe and Ni was smaller at S10 (Figure 10a–d). In Figure 10a1–d1, there was an obvious delamination phenomenon on both sides of the interface. In addition, the S10 and S13 interfaces of Fe and Cr were not smooth (Figure 10b2–b3,c2–c3), while the interface of finished S14 was smoother (Figure 10d2,d3).

The Fe, Cr, and Ni element diffusion distances between each pass are the main factor affecting the interface performance. Figure 11 showed the diffusion distances of the Fe, Cr, and Ni elements of S6–14 (the mean value of three times). In S6–14, the diffusion distances of Fe and Cr varied greatly, showing a consistent parabolic downward trend, while the diffusion distances of Ni were smaller, ranging from 11.5 to 7.3 μm. According to the phase equilibrium theory, the lower the external pressure is, the higher the vacuum is, the more easily the metal element evaporates. At an atmospheric pressure of 10⁻² pa, the Cr started to evaporate at a temperature above 1062 °C. Furthermore, the Cr diffused faster in the ferrite than in the austenite. At the same time, the Cr and Fe would form a diffusion couple, and the two diffuse farther away after 3 h of heating. However, Ni started to evaporate at the temperature higher than 1157 °C. Next, when the rolling temperature is 1180 °C, and the diffusion coefficients (D) of Ni and Cr in γ-Fe were approximately 5.7 × 10⁻¹¹ and 4.6 × 10⁻¹¹ cm² s⁻¹, respectively [27]. Although the D of Ni was high, the diffusion distance of Ni was smaller than that of Cr, which was somewhat different from Equation (3). This was mainly due to the lower concentration of Ni on the SS side than that of Cr [6]. In addition, the interstitial diffusion of C atom would also lead to a hysteresis diffusion of Ni, which could have a more pronounced effect on the inherently slower diffusion of Ni itself, but the delayed effect was not obvious for the rapid diffusion elements (Fe, Cr, Mn, etc.) [37,38]. Together, these factors made the diffusion of Ni in CS more difficult, so the Ni diffusion distance was smaller.

In the S6–8 passes, the diffusion distance of Fe and Cr decreased significantly, while Ni increased slightly (Figure 11), which were, respectively, dominated by the extension of reduction and the thermal effect of deformation. The diffusion distances of Fe and Cr were relatively large, and the decrease in the diffusion layer thicknesses caused by the CRB extension of reduction exceeded the thermal effect of deformation on diffusion. In Figure 2c, the temperature of the S7–8 increased, and the measured surface temperature of the S9 billet was 1073 °C. At this time, the CRB had a larger cross-sectional area, and the cooling rate was slow, so the Ni enriched at the interface was further diffused during rolling and cooling. According to Equation (3), its diffusion distances increased.

For S8–14, Fe, Cr, and Ni showed the same trend. The diffusion distance of the three elements in S8–9 decreases synchronously, due to the fact that the extension was larger and section area was smaller, so that the temperature drop was faster and could not provide enough driving force, making the diffusion distance decrease. For S9–12, the diffusion distances of Fe, Cr, and Ni did not change much. At this time, the decrease in the diffusion layer thickness of the extension and the increase in. the elements’ diffusion distance were in a basic equilibrium state. For S12–13, the diffusion distances rose. The reason was that the 14–17# rolling mill spacing was longer and caused the surface of CRB to cool down, and then, afterwards, the S13 passes the interfaces warmed up again. According to the law of diffusion, when other conditions are certain, the longer the diffusion time is, the element diffusion distance increases. After the finished pass sequence, the diffusion distance of S13–14 decreased rapidly, which was due to the fact that the finishing hole extended downward and sprayed cold water, making the element diffusion distance significantly reduced.

The diffusion concentrations of Fe, Cr, and Ni at both sides of the interface were different in 316L and HRB400E. The diffusion behavior of elements is unsteady and follows Fick’s second law, and the interfacial concentration can be expressed by the diffusion equation [24,42]:

$$\mathrm{In} 316\mathrm{L} \mathrm{SS}\text{:} {\displaystyle \frac{\partial \rho }{\partial t}}=D_{316L}{\displaystyle \frac{{\partial \rho }^2}{\partial x}}$$

(5)

$$\mathrm{In} \mathrm{HRB}400\mathrm{E} \mathrm{SS}\text{:} {\displaystyle \frac{\partial \rho }{\partial t}}=D_{HRB400E}{\displaystyle \frac{{\partial \rho }^2}{\partial x}}$$

(6)

where $\rho$ is the concentration of element, $D_{316L}$ and $D_{HRB400E}$ are the diffusion coefficients, respectively, and x is the diffusion distance of element.

In this paper, the bonded bimetal composite can be seen as infinite diffusion couples. The initial and boundary conditions for the equations were as follows:

$$\mathrm{Initial} \mathrm{condition}\text{:} \rho \left(x,t=0\right)={\rho }_1\left(x<0\right), {\rho }_2\left(x>0\right)$$

(7)

$$\mathrm{Boundary} \mathrm{condition}\text{:} \rho \left(x,t\right)={\rho }_1\left(x=-\infty \right), {\rho }_2\left(x=+\infty \right)$$

(8)

The diffusive flux at the interface was the same and thus obtained:

$$D_{316L}{\displaystyle \frac{{\partial {\rho }_1}^2(x=0, t)}{\partial x}}=D_{HRB400E}{\displaystyle \frac{{\partial {\rho }_2}^2(x=0, t)}{\partial x}}$$

(9)

The solution of Equation (9) was obtained:

$$\rho (x,t)=\left\{\begin{array}{c}{\displaystyle \frac{{\rho }_1\sqrt{D_{316L}}+{\rho }_2\sqrt{D_{HRB400E}}}{\sqrt{D_{316L}}+\sqrt{D_{HRB400E}}}}+{\displaystyle \frac{\sqrt{D_{316L}}({\rho }_2-{\rho }_1)}{\sqrt{D_{316L}}+\sqrt{D_{HRB400E}}}}\mathrm{erf}\left({\displaystyle \frac{x}{2\sqrt{D_{316L}t}}}\right)\left(x<0\right) \\ {\displaystyle \frac{c1\sqrt{D_{316L}}+c2\sqrt{D_{HRB400E}}}{\sqrt{D_{316L}}+\sqrt{D_{HRB400E}}}}+{\displaystyle \frac{\sqrt{D_{316L}{D}_{HRB400E}}({\rho }_2-{\rho }_1)}{D_{316L}+\sqrt{{D_{316L}D}_{HRB400E}}}}erf\left({\displaystyle \frac{x}{2\sqrt{D_{HRB400E}t}}}\right)(x>0)\end{array}\right.$$

(10)

The relationship between the element concentrations and the diffusion distance (x), and the interface contact time (t) was shown in Equation (10), which can be calculated as the element concentrations at different locations on the boundary interface.

3.4. Interfacial Tensile and Fracture Properties

3.4.1. Shear Behavior and Interfacial Characteristics of S6–13

The interfacial shear tensile property is the key evaluation index in the process of intermediate and finish rolling. Figure 12a showed the interfacial shear yield strength (τ0.2), shear tensile strength (τ), and fracture elongation (As) of S6–13 shear specimens at the room temperature uniaxial tensile testing results. The τ0.2 had an overall decreasing trend, and the τ0.2 of S6 and S13 were 271.0 and 211.2 MPa, respectively, with a 59.8 difference. The overall change of τ was not obvious but fluctuated, with values of 410.6 and 410.8 MPa in S6 and S13, respectively, and the maximum value was 424.6 MPa in S11. The τ0.2 and τ were both the smallest in S9, with values of 198.2 and 367.3 MPa, respectively. In S13, τ was much higher than the 210 MPa required in ASTM A264, and was also higher than the shear strength of 333 MPa in the previous study in which SSC rebars were obtained by using 304 seamless tubes-Q195 bar stock and the six-pass hot rolling [15], indicating a stronger metallurgical bonding interface was achieved at S13. In addition, the As values showed an increasing trend in fluctuations, with S6 being 5.62%, S6–10 rising in fluctuations, S10–12 rapidly rising after a sharp decline, and S12–13 slightly increasing to 9.32% (Figure 12a). In addition, there was some dispersion in the measured data values in Figure 12a, and the increases in S6–13 R_tot (53–88%) enhanced the elongation significantly, while the increase in interfacial shear strength (τ) was not obvious, and τ0.2 showed a decreasing trend, which did not coincide with the literature where the bond strength increased with the increase in rolling reduction for composite plates [23,24,43]. This explained that the longitudinal and transverse distribution of the mill and the elliptical–circular pass sequence in the process of intermediate and finish rolling caused instability in the interfacial bonding strength for SSC rebars. In particular, the large deformation of the elliptical curvature had a large effect on the interfacial bonding properties.

The shear tensile specimens of the S14 did not fracture at the interfacial bonding lap zone. The results showed that the yield strength (Re_316L), ultimate tensile strength (UTS, Rm_316L), and fracture elongation (A5_316L) were 255.6 MPa, 607.5 MPa, and 11.35%, respectively. This indicated that the presence of the crescent grooves in S14 pass caused a larger strain at the root of the transverse rib of SS, and the shear strength of the interfacial bonding area exceeded that of the outer thin SS position, which was also the most susceptible to leakage point of the SS after wearing.

Figure 12b showed the engineering stress–strain curves for S6–13 in the process. During the tensile process of the bonding interface, the curves went directly to the strengthening stage without the yield oscillation stage, which was similar to that of a single metal. The shear strength of bonding interface gradually decreased from S6–9, and the plasticity dropped, reaching the minimum at the S9 bonding interface. For S9–10, the plasticity and toughness of the bonding interface rapidly increased, and, for S10–11, they both further improved. Next, the plasticity obviously decreased for S12, but the toughness increased. Finally, the plasticity increased again for S13, but the toughness decreased.

In order to investigate the fracture characteristics of the key passes of the shear tensile specimens, as shown in Figure 13, the cross-section and frontal fracture morphologies of the tensile shear specimens of the reduction direction were observed by SEM in the key passes (S6, S9, S10, S13, and S14). The fracture location of S6 was in the ferrite region on the CS side, the outer layer was thicker and the warping was small on the 316L side, and there was basically no bulge phenomenon, and the fracture was consistent with the loading direction (Figure 13a). Then, the outer 316L SS of S9 became thin, the warping was obvious, and the bulge also became larger at the fracture—at the fracture at an angle of 30° to the direction of loading (Figure 13b). The outer 316L SS of S10 was not obviously thinning, but the fracture presented a certain degree of tearing ‘tip’, and the angle of the fracture and the loading direction of decreased to 10 ° (Figure 13c). The outer 316L SS of S13 was significantly thinner, the warpage became larger again, the tearing “tip” at the fracture was not obvious (Figure 13d), and the angle between the fracture and the loading direction was slightly increased compared to S10. The reason for the interfacial fracture at an angle to the loading direction was that the stress at the interface exceeded the yield strength of the 316L SS, leading to preferential plastic deformation of the SS substrate, which, in turn, caused the shear interface to deflect. The fracture location of the S14 was the root of the transverse rib, which was also the thinnest point of the outer SS layer (more or less thickness) (Figure 13e).

In Figure 13a1–d1, the fracture morphologies became smoother instead of uneven. Figure 13a2 showed the local magnification of S6: the river pattern was seen more, and the tear-like crater holes visible to the naked eyes were observed (Figure 13a3), which was a brittle fracture feature. Figure 13b2–c2 showed the local magnification of S9 and S10: the river pattern decreased, and the lower elevation basins appeared and showed uneven sliding fracture with some layering (Figure 13b3–c3), which was still a brittle fracture characteristic. The local magnification of S13 showed that the fracture indentations became fine and showed an obvious lamellar shape, and there was an obvious shear toughness fossa in the tearing area (Figure 13d2). The bond strength of S13 reached 410.8 MPa. This was because the increases in R_tot caused the inter-interfacial compounds to be broken and promoted their homogeneity in both sides of the interface, and the element diffusion existed at the same time; those reduced the bonding interface from cracking due to stress concentration when subjected to external loads, and played an important role in improving the interfacial shear strength. In addition, S14 was broken at the stainless steel of the link and was significantly different from the tensile shear specimens of S6, S9, S10, and S13. For S14, the outer layer of the 316L fracture showed a ‘tongue pattern’, and there was a ‘swirling’ cavity in the CAZ, indicating that the stress was easy to concentrate to produce cracks in the cavity (Figure 13e2).

3.4.2. Tensile Behavior and Fracture Characteristics

The tensile property of the finished SSC rebar is the key evaluation index of its application in reinforced concrete. Figure 14a showed the stress–strain curves, yield strength (Re), tensile strength (Rm), and fracture elongation (A5) of ordinary HRB400E and SSC rebars. The both Re were 438.1 and 469.0 MPa, with a 7.05% increase in SSC, and the Re of SSC is also better than 435 MPa in the literature by metal deposition and hot rolling [12]. For both, the Young’s modulus (E) was 200.9 and 189.8 GPa, respectively. The ordinary rebars followed typical strain–stress behavior. In the linear-elastic stage, the slope of the SSC rebar was a little less than that of ordinary rebar, which indicated that SSC rebars had slightly less ability to resist deformation than ordinary rebars. Then, the yield stage of ordinary rebar was longer, and a clear sawtooth shape appeared at the end, which was formed by breaking free from the pinning of the Cottrel air mass. However, the yield stage of the SSC rebar was shorter and smoother. The reason was that the yield limits of the two metals were not consistent. When HRB400E CS reached its elastic limit, the 316L SS was still in the elastic phase. Then, when the 316L SS reached its elastic limit, the CS had been already in the strengthening phase. The whole process weakened the oscillation of the yield phase. In the hardening stage, both Rm were 605.7 and 648.7 MPa, respectively, and the SSC rebars increased by 7.10%. There was a large sawtooth shape in the front part of the SSC rebar hardening stage, which was significantly different from the ordinary rebar. In addition, the A5 of SSC rebars was lower by 2.00% than that of ordinary rebars, and the measured results were more dispersed. The literature had shown that the yield strength and tensile strength means of nominal SSC rebars from testing 12 times were 430.2 MPa and 645.2 MPa, respectively [2]. In this paper, the Rm basically coincided with a previous study, but the Re improved by 38.8 MPa. The results showed that the process in this paper was better for performance enhancement. Figure 14b showed the appearance and fracture profile of ordinary and SSC rebars. The ordinary rebar was observed to have significant necking and flatter fractures with a few tears, which was a typical cup-and-cone fracture. For SSC rebars, the cladding fracture showed significant tearing, and the core of CS was an uneven fracture.

Further, the fracture morphology of the finished product was observed and analyzed. Figure 15 showed the microscopic fracture profile view of the SSC rebar. The diffuse necking was observed at the fracture with some bonding existing and part delamination at the interface (Figure 15a). As shown in Figure 15b, there was a clear transition zone existing in the two phases of the SS cladding and CAZ. As shown in the enlarged view (Figure 15b1), the fracture of SS cladding was a typical cut-off-type fracture, which was smooth and bright with small and tough nests on the surface. As displayed in the enlarged view of Figure 15c2,3, there was a significant difference between the CAZ and DFZ. The CAZ showed a large inclined long granular shape and had swirling cavities. The DFZ exhibited a ripple pattern, and the cavity of DFZ was smaller than the CAZ. The CAZ was a brittle fracture, and the DFZ was a ductile fracture. In addition, the fracture of the CS substrate exhibited a tough fracture character, where the dimples were large and deep [25,44] (Figure 15d).

4. Discussion

4.1. Strengthening and Toughening Mechanism of the Interface

Interfacial bonding strength is a key indicator with which to evaluate the properties of the bimetallic composite [45]. The grain size of different passes affects the shear strength. The relationship between the grain size and shear strength is expressed by Eshelby’s improved Hall–Petch relation, Equation (11) [46,47], and grain boundary strengthening can be calculated:

$${\tau }_g\gg {\tau }_0+k{D}^{-1/2}$$

(11)

where ${\tau }_g$ is the yield strength, ${\tau }_0$ is the friction force that the dislocation motion needs to be overcome without any obstruction, $k_y$ is a constant, and D is the grain size.

As shown as Figure 14, the shear strength is analyzed in conjunction with the grain size variations of S6–14 (Figure 5 and Figure 6) and the diffusion spacing of the elements (Figure 9, Figure 10 and Figure 11):

For S6–9, the τ0.2 and τ decrease as a whole, where the A5 of S6–8 increases first and the A5 of S8–9 decreases (Figure 12a). The CAZ grain sizes of S6–8 decrease, but the DFZ grain sizes do not change much (Figure 5b and Figure 6a,b), indicating that the fine grain strengthening is not obvious. For S6–8, we see that the reduction in the elements’ diffusion distance and the fragmentation and dispersion of the inclusions lead to the decrease in the interfacial bond strength. In addition, due to the increase in CAZ grains and the decrease in the twin crystal for S9, the intergranular resistance and the toughness of CAZ decrease (Figure 6b,c), but the brittleness increases, so the interfacial bonding strength further decreases [29].

For S9–10, the τ0.2, τ and A5 rise significantly, and the CAZ grain sizes do not change much (21.3–23.0 μm), in which the elemental diffusion spacing is basically in a stable state (Figure 11). This is mainly due to the fact that, with the deformation increase, the dislocations density inside the grain increases, and the dislocations produce a large number of cut steps and twists during the slip process. For S10–11, the τ0.2 decreases slightly, the τ rises while A5 decreases, in which the grain size of CAZ decreases significantly. When the grain sizes decrease, the dislocation blockage increases accordingly, and the dislocations obstruct each other, which leads to the difficulty of plastic deformation and enhances the shear strength. At the same time, the fragmentation and dispersion of the second-phase inclusions make the interface plasticity decrease.

For S11–12, the τ0.2 decreases slightly, and the τ obviously decreases while the A5 increases. The grain size of CAZ also obviously increases (16.9–23.9 μm), resulting in a significant decrease in interfacial bonding strength. Moreover, the second-phase particles are further fragmented and dispersed, making the plasticity further significantly increase. For S12–13, the τ0.2 continues to drop slightly, but the τ and A5 synchronously show a small increase. This is because the mills of S12–13 have a larger spacing. The reason is that the larger spacing of the mills significantly facilitates the element diffusion (Figure 11). Additionally, the grain dislocation density of S13 is higher than that of the ferrite and pearlite organization, indicating that the solid solution strengthening could promote the strength and plasticity of metallurgical interfacial bonding. At the same time, the smaller flat elliptic section of S13 is conducive to heat dissipation, and the slow cooling process of the high-temperature region becomes shorter and fails to significantly eliminate the cumulative effect of WH on the pass, so the interface properties improve.

4.2. Fracture Mechanism

Due to the difference between the high-strength CAZ and the plastic DFZ of the interfacial bonding zone, cracks and losses are more likely to penetrate and propagate into the weak material [24]. The DFZ has the lowest hardness with a ductile characteristic, and, firstly, the DFZ shows the majority of cracks (Figure 15), and then the DFZ crack may carve across the rigid HHB and propagate along the CAZ boundary during the shear tensile process. In addition, the number, size, and dispersion of oxides have been shown to impact the fracture properties in studies such as Zhu et al. [26].

Due to the Cr carbide brittle phase existing at the interface, the plasticity of the DFZ is better than the brittle phase. During the tensile process of the finished product, the deformation process cannot be coordinated on both sides of the CAZ and DFZ, so the dislocation motion is impeded to produce much dislocation accumulation and causes the relative grain dislocation source to start, resulting in the cracks sprouting at the carbide of the CAZ on interface. Moreover, the reason has been mentioned before; that is, the plasticity of the DFZ pure ferrite is better than that of the CAZ austenitic, so the cracks expand to the CAZ of 316L SS with a poor deformation ability, and the discontinuous delamination appears where the metallurgical bond is not sufficiently good. Consequently, it is considered that the fracture due to delamination should be initiated by a nonuniformity caused by an unstable deformation mode other than diffuse necking, such as interface instability [48]. When reaching the tensile strength, the specimen exhibits a distinct brittle fracture, and there is a necking fracture in the core CS.

4.3. Mixture Properties

Currently, the tensile properties of bimetallic composites are predicted using the rule of mixture [27], as shown as Equation (12):

$$P_b=P_p{V}_p+P_c{V}_c$$

(12)

where P is the material properties, $V$ is the material volume content, and b, p, and c are the composite, parent, and cladding metal material, respectively.

The tensile properties of the bimetallic composite, parent, and cladding metal materials in this paper are shown in

Table 9. Tensile properties of the parent and cladding metal, and the SSC rebar and comparison with the rule of mixture.

Type	Yield Tensile Strength (Re, MPa)	Ultimate Tensile Strength (Rm, MPa)	Young’s Modulus (E, GPa)	Fracture Elongation (A5, %)
Cladding (Figure 12 and Figure 13e)	255.5	607.5	152.9	11.35
Parent metal (ordinary HRB400E rebar)	438.1	605.7	200.9	22.85
Measured rule of mixture	404.7	606.0	192.1	20.75
SSC rebar	469.0	648.7	189.8	20.85
$k_p$	2.38	1.38	0.92	1.05

. The base circle diameter and cladding thickness of the S14 are measured, and the volume fraction of SS is calculated to be 18.27%. Then, the rule of mixture is used to calculate the tensile properties. The predicted value of E and A5 coincide with the experimental value, but the predicted value of Re and Rm are smaller than the experimental value, and the differences are 64.3 and 42.7 MPa, respectively. Therefore, in this study, the rule of mixture could not accurately predict the tensile properties of SSC rebars. By considering the SSC rebars as a whole, the crescent-shaped transverse rib of SS is bonded to the base CS of this position, presenting results that go beyond the rule of mixture. Thus, the metallurgical bonding of the bimetal interface can enhance the tensile strength of the SSC rebar.

According to the microstructure characterization and performance test results in this paper, similar to the quantitative calculation method of the yield strength, the strength of SSC rebars can be expressed by Equation (13):

$$R_b=R_p(1-V_c)+R_c{V}_c+k\ast {\tau }_b$$

(13)

where R is the composite tensile properties, k is constant and a larger $k_p$ indicates a larger increase in this property, and ${\tau }_b$ is the shear properties of the bonding interface. In this paper, the shear property of S13 is used instead of S14 to solve for the value of k, and the results are shown in Table 9.

5. Conclusions

In this paper, 316L SS-HRB400 CS vacuum oxidation-free billet was used for an industrial hot-rolling experiment of SSC rebars. The microstructure evolution, element diffusion, and tensile and fracture behaviors of bimetallic interfaces during the intermediate and finish rolling passes were investigated. The main conclusions are as follows:

(1) From S6 to S14, the sectional images show good metallurgical bonding at the interface, without obvious cracks or voids. There are obvious instances of DRV and DRX on both sides of the interface, but the DFZ is not as drastic as CAZ. The thicknesses of the CAZ vary greatly and decrease significantly (832–238 μm) with a parabolic downward trend. However, the thicknesses of the DFZ are less variable and increased in fluctuations (85–99 μm).

(2) Due to the short composite time in the process of the intermediate and finish rolling passes, Fe and Cr show a consistent parabolic downward trend, and the diffusion distance of Fe and Cr varied greatly (Fe: 51.9–17.7 μm, and Cr: 45.6–15.9 μm), while the diffusion distance of Ni is small (11.5–7.3 μm). The diffusion distance of Fe and Cr in the S6–9 intermediate-rolling decreases sharply, while the diffusion distance of Ni increases first and then decreases. The diffusion distance of elements in intermediate-rolling back stage and finishing-rolling front stage (S9–12) are basically balanced. Due to the change in the mill spacing, the diffusion distance of each element in the finish-rolling back stage (S12–14) first increases and then decreases.

(3) With R_tot increasing from 53% to 88% for S6–13, the τ0.2 shows an overall decreasing trend, the τ fluctuates but the change is not obvious, and the A5 value increases in the fluctuation. The results show that the diffusion of atoms at the interface promotes the formation of metallurgical bonding, and the fluctuation of the grain size, the fracture, and the dispersion of impurities such as oxides between the interface improve the ductility and toughness of the interface bonding process.

(4) Compared with that of ordinary rebars, the Re and Rm of SSC rebars are increased by 7.05% (30.9 MPa) and 7.10% (43.0 MPa), respectively. Although the A5 decreases slightly, it also meets the requirements. During the tensile process of the finished product, the deformation coordination of the carbide at the interface is different from that on both sides of interface. The cracks sprout at the carbide of interface and expand to the CAZ of the 316L SS with a poor deformation ability, so the fracture is triggered by delamination rather than diffuse necking.

The results of the paper can provide a theoretical basis for the optimization of the rolling process system and have important significance for the industrial production of SSC rebars.