⇦Chinese Physics Letters, 2018, Vol. 35, No. 1, Article code 018101 Deformation and Spallation of Explosive Welded Steels under Gas Gun Shock Loading * Ying Yu(余颖)1,2,3, Chao Li(李超)2,3**, Hong-Hao Ma(马宏昊)4, Mei-Lan Qi(祁美兰)1**, Sheng-Nian Luo(罗胜年)2,3** Affiliations 1School of Science, Wuhan University of Technology, Wuhan 430070 2Key Laboratory of Advanced Technologies of Materials (Ministry of Education), Southwest Jiaotong University, Chengdu 610031 3The Peac Institute of Multiscale Sciences, Chengdu 610207 4CAS Key Laboratory of Materials Behavior and Design, Department of Modern Mechanics, University of Science and Technology of China, Hefei 230027 Received 25 September 2017 *Supported by the National Basic Research Program of China under Grant No 2014CB845904, and the National Natural Science Foundation of China under Grant Nos 11627901, 11372113 and 11672110.
**Corresponding author. Email: emmy_qi@163.com; cli@pims.ac.cn; sluo@pims.ac.cn Citation Text: Yu Y, Li C, Ma H H, Qi M L and Luo S N 2018 Chin. Phys. Lett. 35 018101 Abstract We investigate deformation and spallation of explosive welded bi-steel plates under gas gun shock loading. Free surface histories are measured to obtain the Hugoniot elastic limit and spall strengths at different impact velocities. Pre- and post-shock microstructures are characterized with optical metallography, scanning electron microscopy, and electron backscatter diffraction. In addition, the Vickers hardness test is conducted. Explosive welding can result in a wavy steel/steel interface, an ultrafine grain region centered at the interface, and a neighboring high deformation region, accompanied by a hardness with the highest value at the interface. Additional shock compression induces a further increase in hardness, and shock-induced deformation occurs in the form of twinning and dislocation slip and depends on the local substructure. Spall damage nucleates and propagates along the ultrafine grain region, due to the initial cracks or weak interface bonding. Spall strengths of bimetal plates can be higher than its constituents. Plate impact offers a promising method for improving explosive welding. DOI:10.1088/0256-307X/35/1/018101 PACS:81.05.Bx, 81.20.Vj, 81.70.Bt, 62.20.F- © 2018 Chinese Physics Society Article Text Explosive welding is an excellent metallurgical bonding method via high velocity collision. An explosive welded composite possesses advantages over its components. Previous studies on such materials largely address the effects of process parameters on welding interface morphology,[1] strength tests of the welding interface by tensile, shear and bend loading,[2,3] and modification of the microstructure of welding interfaces by heat treatment.[4] However, microstructure effects on yield and tensile strengths of welding interfaces have never been investigated under high strain rate loading ($>$10$^{5}$ s$^{-1}$), such as plate impact. The weld interface of a metal composite is expected to play a key role in deformation and damage under shock loading. It is also curious and useful to know how strong a weld interface is. Shock compression waves are reflected and refracted at an interface, due to the mismatch in strength and wave velocity of constituent metals.[5,6] In subsequent fracture, the interface may also act as damage nucleation site due to its weak bonding.[7] In this work, we perform shock compression and spallation experiments with gas gun shock loading to investigate deformation and spallation of explosive welded #45/Q235 steel bilayer plates at high strain rates. Free surface histories are measured to obtain the Hugoniot elastic limit and spall strength at different impact velocities. Pre- and post-shock microstructures are characterized with optical metallography, scanning electron microscope (SEM), and electron backscatter diffraction (EBSD). In addition, the Vickers hardness test is conducted.
Table 1. Chemical composition (wt%) of #45 and Q235.
Element C Mn Si S P Fe
(wt%)
Q235 0.17 1.30 0.30 0.04 0.04 97.85
#45 0.45 0.70 0.32 0.03 0.03 98.47
The explosive welded steel/steel plate samples are provided by CAS Key Laboratory of Materials Behavior and Design, Department of Modern Mechanics, University of Science and Technology of China. The samples are manufactured from two steels: low carbon steel Q235 and middle carbon steel #45, and their chemical compositions are listed in Table 1. The explosive welding setup is illustrated schematically in Fig. 1(a). Here #45 is a flyer plate with the size of 150$\times$100$\times$2 mm$^{3}$, and Q235 is a base plate with the size of 100$\times$100$\times$16 mm$^{3}$. The stand-off distance of the flyer plate and the base plate is 6 mm. These two steels possess approximately the same mechanical properties except strength. High explosion pressure from explosive charge drives the flyer plate (#45) to impact the base plate (Q235), at an impact velocity ($V_{\rm P}$) of 700 m/s and a collision angle ($\beta$) of 11$^{\circ}$, respectively. Then the surfaces of these two metals reach the molecular spacing, and are bonded. Microstructure of the initial welding interface in the steel/steel bilayer is characterized with electron backscatter diffraction (EBSD; Figs. 1(b)–1(d)). The typical wavy welding interface is formed, with a wavelength of $\sim$300 μm and amplitude of $\sim$60 μm. At the center of the wavy interface (marked as 1 in Fig. 1(b)), the grain size is ultrafine (Fig. 1(c)) and the residual strain is negligible (Fig. 1(d)). This is a result of dynamic recrystallization under local high temperature produced by the collision of two surfaces.[8] However, ultrafine grain regions (UGR) may contain cracks (red arrows), indicating flawed bonding. Near the interface, severely plastic flow is observed in both highly deformed regions (HDR; marked as 2 and 3). The basal microstructure region (BMR; marked as 4 and 5) represents the region less affected by surface collision,[9,10] and is located $\sim$200 μm away from the interface. At this area, the plastic deformation is considerably weaker.
cpl-35-1-018101-fig1.png
Fig. 1. (a) Schematic diagram of explosive welding. A: explosive charge; B: buffer; C: flyer plate; D: base plate; $V_{\rm D}$: velocity of detonation front; $V_{\rm P}$: impact velocity of flyer plate; $\beta$: collision angle. (b) Image quality map of the interface of explosive welded Q235/#45 steels, and corresponding (c) inverse pole figure map and (d) kernel average misorientation map.
Table 2. Material parameters of individual steels and #45/Q235 bilayer plate.
Material Grain size Density Yield stress Sound velocity Bulk sound velocity
(μm) $\rho_0$ (g/cm$^{3}$) $\sigma_y$ (GPa) $C_{\rm L}$ (km/s) $C_{\rm B}$ (km/s)
#45 $\sim$20 7.80 0.35 5.95 4.56
Q235 $\sim $20 7.80 0.25 6.01 4.65
#45/Q235 7.80 6.04 4.66
cpl-35-1-018101-fig2.png
Fig. 2. Schematic setup of the gas gun flyer-plate impact experiments. (1) Polycarbonate sabot; (2) o-ring; (3) gun barrel; (4) recess for release waves; (5) flyer plate; (6) optical fibers and detectors for the optical beam block system; (7) sample; (8) sample holder; (9) optical fiber connected to the Doppler pin system (DPS) probe; (10) thin mirror; (11) soft recovery; (12) vacuum chamber.
Planar flyer plate impact experiments are conducted with a 14 mm bore, single-stage, gas gun,[11] as shown in Fig. 2. A flyer plate (5) is attached to polycarbonate sabot (1), with a recess (4) behind it to initiate a release fan for spallation. O-rings (2) are placed around the sabot to prevent gas leakage. When a solenoid valve is fired, high-press gas helium or nitrogen stored in a reservoir is released into the gun barrel (3) to accelerate the sabot-flyer plate assembly to impact the target or sample (7). The flyer plate or impact velocity ($u_{\rm imp}$) is measured with an optical beam block system (6) while the flyer plate exiting the muzzle, and a Doppler pin system (DPS) (9) is installed to measure free surface velocity ($u_{\rm fs}$) history of the target. The muzzle, target, DPS and related optics are located in a vacuum chamber (12), and the impacted targets are 'soft-recovered' with soft materials (11) for microstructure characterizations. Two parallel surfaces of a flyer plate or sample are polished to micron level or mirror finish. The disk-shaped flyer plates are made of Q235 steel and have a diameter of 14 mm. The explosive welded #45/Q235 steel samples are rectangles with the size of $9.7\times9.7$ mm$^{2}$ and a thickness of 2 mm. The wavy steel/steel interface is located in the middle of a sample along the thickness direction. The experimental parameters and the sample sizes are listed in Table 3. For all the experiments, the #45 steel is on the impact side.
Table 3. Impact conditions for #45/Q235 steel bilayer plates.
Shot 1 2 3 4
Sample thickness (mm) 2.01 1.99 2.02 2.01
Flyer thickness (mm) 0.99 1.01 1.01 4.11
Impact velocity (m/s) 110 152 397 423
Figure 3 shows the measurements of free-surface velocity profiles through a Doppler pin system (DPS). Segments AB and BC represent elastic precursor and plastic wave, respectively. The stress at the Hugoniot elastic limit (HEL; point B) is $\sigma _{\rm HEL}=\frac{1}{2}\rho _0 C_{\rm L} u_{\rm fs}|_{\rm B}$. It is about 1.21 GPa, independent of impact velocity. The HEL of the explosive welded steel bilayer is lower than that of Q235 steel.[12] A supported plastic shock achieved with a duration of $\tau$ (CD), and subsequent velocity drop (DE) is caused by the release fan from the back surface of the flyer plate. The interaction of this release fan with that initiated from the target free surface releases the mid-part of the sample from compression into tension.[13,14] In our experiment, the steel/steel interface experiences the maximum tensile stress. When this tensile stress exceeds the strength of the interface, spall will occur. Point E signals spallation, because the compression wave originated from the spall plane within the sample induces ensuing velocity pullback or re-acceleration (EF). Re-acceleration EF represents growth and coalescence of isolated cracks or voids.[15,16] Spall strength ($\sigma _{\rm sp}$) is calculated from the pullback velocity with the acoustic method, $$ \sigma _{\rm sp} \approx \rho _0 C_{\rm L} \Delta u\frac{1}{1+(C_{\rm L}/C_{\rm B})},~~ \tag {1} $$ where $\Delta u$ is the difference in $u_{\rm fs}$ between C and E in Fig. 3. The spall strengths of the explosive welded bilayer steel are 1.99 GPa and 3.08 GPa at $u_{\rm imp}=152$ m/s and 397 m/s, respectively, and higher than that of its constituent steel Q235, which are 1.6–2.2 GPa at $u_{\rm imp}=122$–427 m/s.[12]
cpl-35-1-018101-fig3.png
Fig. 3. Representative free surface velocity histories. F: flyer plate; S: sample
cpl-35-1-018101-fig4.png
Fig. 4. Micrographs of (a) as-received and (b) shock-compressed samples showing indents (black dots) in the Vickers hardness testing. (c) Hardness profiles across the weld interfaces, centered at the white lines marked in (a) and (b). (d)–(h) Inverse pole figure maps for areas marked as 1–5 in Fig. 5(b), respectively.
After shock compression, hardness tests are conducted. Metallographical micrographs and corresponding hardness profiles are shown in Figs. 4(a)–4(c), respectively. Indentation is performed along the horizontal lines at an increment of 50 μm, and the indents are displayed as black dots. The dark regions in the metallographical micrographs (Figs. 4(a) and 4(b)) are defects such as cavity, dislocations and grain boundaries, and the bright regions represent nearly perfect crystal. For the as-received sample (Fig. 4(a)), the metallographic microstructure of #45 steel (left) is different from that of Q235 (right; showing streaks). Their interface (the white dashed line in Fig. 4(a)) is much darker than the other areas, identified as ultrafine grain region (UGR). UGR possesses the highest hardness as a result of the Hall–Petch effect.[17] Within a distance of $\sim$200 μm from the interface (the highly deformed region or HDR), the color becomes lighter and hardness decreases monotonically. At distance beyond 200 μm (the 'matrix' region or MR), microstructure and hardness change slightly. The higher hardness in HDR relative to MR is attributed to high density of defects and residual strain. After shock compression, a pronounced feature of the steel/steel bilayer is darkening of the area near the interface ($ < $210 μm), accompanied by a sharp increase in hardness, both indicating an increase in defect density. This area (two boundaries marked by red lines in Fig. 4(b)) consists of UGR and HDR, which are affected considerably by shock compression. However, microstructure and hardness in MR undergo slight changes. One possible reason is that dislocation density is saturated by oblique shock compression during explosive welding. Compared with Q235 steel (zone 5, Fig. 4(b)), the flyer plate impact induces larger changes to #45 steel (zone 1). In addition to metallography, EBSD characterizations of zones 1–5 (Fig. 4) are conducted. Zones 1 (#45) and 5 (Q235) contain abundant subgrain boundaries in grain interiors, and the density of subgrain boundaries in steel #45 is higher. Subgrain boundaries contribute to blackening in optical metallography (Fig. 4(b)) and a slight increase in hardness after shock compression (Fig. 4(e)). In HDR (zones 2, 4), we observe copious lens-shaped twins and subgrains, which are responsible for the increase in hardness.[18] Twin density in steel #45 is higher than that in Q235, but secondary twinning only occurs in the latter.
cpl-35-1-018101-fig5.png
Fig. 5. Microstructure characterization of spalled samples, sectioned along the impact direction (the black arrow). (a)–(c) SEM micrographs of spallation regions for $u_{\rm imp}=110$ m/s, 152 m/s and 397 m/s, respectively. Inset in (c) shows spallation in pure Q235 steel from a separate study.[12] (d) The corresponding inverse pole figure (IPF) map for $u_{\rm imp}=152$ m/s. IPF maps collected from (e) area 1 and (f) area 2 in (c).
However, previous studies[19-21] showed that pre-strain impedes twin nucleation, because the probability of twin nucleation is inversely proportional to mobile dislocation density and the degree of uniformity of dislocation distribution. In our case, microstructure of HDR formed in explosive welding is different: inhomogeneous structure produced by severe shearing between the two surfaces upon oblique collision;[22] mixing of chemical elements during welding;[9,23] incomplete annealing; and pinning of mobile dislocations by atoms, such as C and N atoms.[24,25] Moreover, the wavy layer structure at the interface and mismatch in yield strength between the two steels also result in large shear concentration, facilitating twin nucleation. However, twins are absent in UGR (3), due to the small grain size.[26] Many dislocations appear in these recrystallized grains after shock compression (dark areas near grain boundaries in Fig. 4(f)). These high density dislocations induce an increase in hardness in UGR (Fig. 4(c)). Different damage extents are shown in Fig. 5. SEM and EBSD characterizations of the spalled samples at $u_{\rm imp}=110$ m/s (not shown in the free surface velocity histories), 152 m/s and 397 m/s are presented. Spallation at 110 m/s is identified as incipient spallation, with characteristic isolated small voids (Fig. 5(a)). These voids can grow and coalesce into wavy cracks along the center of UGR (152 m/s), and cracks are not observed in other areas. Therefore, UGR is weaker than other areas, due to initial cracks or weak interface bonding. For full spallation at 397 m/s, a main crack is formed along the interface, which is much smaller than that of pure Q235 sample spalled at a similar velocity (the inset of Fig. 5(c)). Small cracks appear on the #45 steel side $\sim$200 μm away from the interface. These areas with cracks contain many subgrain boundaries (Fig. 5(f)). Nonetheless, the area near the interface ($ < $210 μm on the #45 steel side) is free of cracks. One reason is that spallation at the interface releases tension in the area immediately adjacent to it, and another possible reason is that massive deformation twins (Fig. 5(e)) induce local hardening and resistance to damage in this area. Cracks are scarce at the area near the interface on the Q235 steel side or free surface side. Similarly, damage at the area near the spall plane on the free surface side is much smaller than that on the impact surface side for 1145 aluminum,[27] because the former experiences a shorter duration of tensile loading stress than the latter. References Effect of explosive characteristics on the explosive welding of stainless steel to carbon steel in cylindrical configurationEffect of interlayer thickness on shear deformation behavior of AA5083 aluminum alloy/SS41 steel plates manufactured by explosive weldingDeformation and fracture of explosion-welded Ti/Al plates: A synchrotron-based studyEffect of post-weld heat treatment on the interface microstructure of explosively welded titanium–stainless steel compositeAn experimental investigation of shock wave propagation in periodically layered compositesDynamic fracture of carbon nanotube/epoxy composites under high strain-rate loadingMicrostructure and mechanical properties of Ti/Al explosive claddingStudy on the microstructure and mechanical properties of explosive welded 2205/X65 bimetallic sheetStructural properties of Ti/Al clads manufactured by explosive welding and annealingSpall damage of a mild carbon steel: Effects of peak stress, strain rate and pulse durationSpall failure of aluminum materials with different microstructuresMeasurement and Analysis of Spall Characteristics of High-Pure Aluminium at One-Dimensional Strain LoadingMicromechanics of spall and damage in tantalumModel of superlattice yield stress and hardness enhancementsTwin nucleation in cold rolled low carbon steel subjected to plate impactsShock-Induced Substructural Changes in Prestrained IronThe effect of a pre-strain on the low temperature mechanical properties of a low carbon steelExamination of copper/stainless steel joints formed by explosive weldingEffects of heat treatment on microstructures and mechanical properties of a directionally solidified cobalt-base superalloyChange in the electron structure caused by C, N and H atoms in iron and its effect on their interaction with dislocationsDeformation twinning
[1] Mendes R, Ribeiroand J and Loureiro A 2013 Mater. Des. 51 182
[2] Han J H, Ahn J P and Shin M C 2003 J. Mater. Sci. 38 13
[3] E J C, Huang J Y, Bie B X, Sun T, Fezzaa K, Xiao X H, Sun W and Luo S N 2016 Mater. Sci. Eng. A 674 308
[4] Mousavi S A A A and Sartangi P F 2008 Mater. Sci. Eng. A 494 329
[5]Meyers M A 1994 Dynamic Behavior Of Materials (New York: John Wiley & Sons) p 45
[6] Zhuang S M, Ravichandrani G and Grady D E 2003 J. Mech. Phys. Solids 51 245
[7] Bie B X, Han J H, Lu L, Zhou X M, Qi M L, Zhang Z and Luo S N 2015 Compos. Part A 68 282
[8] Xia H B, Wang S G and Ben H F 2014 Mater. Des. 56 1014
[9] Zhang L J, Pei Q, Zhang J X, Bie Z Y and Li P C 2014 Mater. Des. 64 462
[10] Fronczek D M, Wojewoda-Budka J, Chulist J R, Sypien A, Korneva A, Szulc Z, Schell N and Zieba F 2016 Mater. Des. 91 80
[11]Zaretsky E B 2016 Light Gas Gun (Berlin: Springer) vol 10 p 3
[12] Li C, Li B, Huang J Y, Ma H H, Zhu M H, Zhu J and Luo S N 2016 Mater. Sci. Eng. A 660 139
[13] Wang Y G, Qi M L, He H L and Wang L L 2014 Mech. Mater. 69 270
[14] Qi M L, He H L and Yan S L 2007 Chin. Phys. Lett. 24 2338
[15]Wang Y G, Hu J D, Qi M L and He H L 2011 Acta Phys. Sin. 60 126201 (in Chinese)
[16] Zurek A K, Thissell W R, Johnson J N, Tonks D L and Hixson R 1996 J. Mater. Process. Technol. 60 261
[17] Chu X and Barnett S A 1995 J. Appl. Phys. 77 4403
[18]Salem A A, Kalidindi S R, Doherty R D and Semiatin S L Metall. Mater. Trans. A 37 259
[19] Visser W and Ghonem H 2017 Mater. Sci. Eng. A 687 28
[20] Mahajan S 1969 Phys. Status Solidi 33 291
[21] Lindley T C 1965 Acta Metall. 13 681
[22]Blazynski T Z 1983 Exploxive Welding, Forming and Compaction (New York: Springer) p 94
[23] Durgutlu A, Gulenc B and Findik F 2005 Mater. Des. 26 497
[24] Jiang W H, Guan H R and Hu Z Q 1999 Mater. Sci. Eng. A 271 101
[25] Gavriljuk V G, Shivanyuk V N and Shanina B D 2005 Acta Mater. 53 5017
[26] Christian J and Mahajan S 1995 Prog. Mater. Sci. 39 1
[27]Seaman L, Barbee Jr T W and Curran D R Dynamic Fracture Criteria of Homogeneous Materials (Menlo Park, June 1970–October 1971)