Introduction

Bulk metallic glasses (BMGs), with their amorphous structure possessing attractive properties, such as high strength and elasticity, which is often coupled with high corrosion resistance and toughness1,2,3. Motivated by these properties and the potential to process them like thermo-plastics, BMGs are currently at the frontier of metals research4,5,6.

One important requirement for any material class to be of practical use for structural applications is the ability to join them to like and dislike materials. Particularly for BMGs, the metastable nature and resulting incompatibility with conventional processing and joining methods have become stumbling blocks for their practical applications7, 8. To address the ability to join BMGs, various techniques that have been developed specifically or tailored for BMGs based on liquid-state processing including electron beam welding9, 10, laser beam welding11, 12, gas tungsten arc (GTA) welding13, and pulse current method14, 15; or solid-state processing including friction welding16, 17, explosion welding18, ultrasonic welding19, 20, diffusion bonding21, spark welding22, and resistance spot welding23. Recently, a thermoplastic-based method24 and a liquid-solid joining method7 were introduced. The most important issue in the welding of BMGs is the avoidance of crystallization in the fusion zone (FZ) and heat-affected zone (HAZ), which requires rapid cooling and has been generally challenging for developed joining methods25, 26. This is particularly the case for joining methods where the joint region is melted, so the weld must be subsequently cooled fast enough to avoid crystallization27. Laser welding can reduce the possibility of crystallization28,29,30,31 since it results in a deep and narrow weld region with higher cooling rates that can be achieved32, 33. Avoiding crystallization in a weld joint of BMG is in general difficult and has only been realized for a small number of BMG formers with high glass forming ability under highly optimized welding parameters34, 35.

To address the requirements for the fast cooling we use pulsed laser beam welding method which provides high welding energy concentrated within a narrow zone, and a much shorter retention time, which results in higher cooling rates. The pulsed laser beam welding method was used to do bead-on-plate (BOP) experiments of Pd43Cu27Ni10P20 (at.%) BMG under different welding parameters. The weld seams were examined for their amorphous nature. We then used one of the welding parameters of BOP tests to join Pd43Cu27Ni10P20 BMG of 1 mm thickness. After joining, the amorphous nature and mechanical properties of the weld joint were investigated. We show that the pulsed laser beam welding method yields precise welds with bulk like mechanical properties.

Materials and Methods

Pd43Cu27Ni10P20 master alloy ingots were prepared by arc-melting a mixture of the high-purity (min. 99.95%) elements in a pure titanium-gettered argon atmosphere with a low oxygen level of 350 ppm. The amorphous state was achieved by rapid quenching of alloy casting-suction into a copper mould. Subsequently, thermoplastic forming (TPF)36,37,38,39 was used to essentially eliminate casting induced porosity. The amorphous nature of the BMG plates was determined by X-ray diffraction (XRD, Rigaku SmartLab, using Cu Kα radiation), and differential scanning calorimetry (Perkin Elmer, Diamond DSC) at a heating rate of 20 K min−1. First, the BOP experiments on Pd43Cu27Ni10P20 BMG were carried out using the pulsed laser beam welding setup (IPG PHOTONICS 1500) with 1500 W maximum laser beam power in the wavelength region of 1060–1100 nm. Table 1 shows the welding parameters for the bead-on-plate experiments. Effective peak power density (EPPD) determines interaction intensity of laser beam with the BMG for a given spot size considering the pulse overlap (PO) during welding process and is given by40,

$$EPPD=\Gamma \times PPD,$$
(1)

where Γ is the pulse overlapping index and PPD is the peak power density. Due to Γ = 1/(1 − PO) and PPD = P peak /[π(d/2)2], where P peak is the peak power of the laser beam, and d is the spot diameter, Eq. (1) can be rewritten as,

$$EPPD=\frac{{P}_{peak}}{\pi {(\frac{d}{2})}^{2}(1-PO)}$$
(2)

where PO = 1 − ν/df, ν is the welding speed, and f is the pulse frequency.

Table 1 Welding parameters of the pulsed laser beam welding.
Full size table

After welding, the appearance of the weld seam was observed with a VHX-500F digital optical microscope (OM). Samples were cut perpendicular to the welding direction of the weld joint for characterization.

Subsequently, in order to study the mechanical properties of the weld joint two sheets of Pd43Cu27Ni10P20 BMG were joined together using the pulsed laser beam welding method, also called butt welding (Fig. 1). The different zones of FZ, HAZ and base material (BM) in the butt joint were separated in order to further determine their glassy nature. The butt joint of Pd43Cu27Ni10P20 BMG was ground on the planar-section of the weld seam with wet abrasive papers and mechanically polished to obtain a mirror-polished test planar-section. In order to characterize the mechanical properties of the butt joint, room-temperature nanoindentation tests were performed on the polished planar-section using a MTS Nanoindenter XP (MTS, Oak Ridge, TN) with a Berkovich tip. All nanoindentation experiments were done using the CSM (Continuous Stiffness Measurement) technique41 which gives the load on the sample and the contact stiffness as a function of the displacement of the indenter into the samples. A fused silica specimen with a known modulus was used to calibrate the system. The first indent is 1.5 mm far away from the edge of the butt joint starting at the middle of the weld seam covering a distance of 5 mm. Indentations were spaced at 500 μm along the x-direction and 500 μm along the y-direction. A series of indentations were carried out in a 10 × 4 grid. A specially developed method42, 43 was used, which uses exponential loading at a constant strain rate of 0.05 s−1, corrects for thermal drift and calculates the average for the modulus and hardness from a penetration depth of 150 nm to 500 nm. All the indentations are up to a peak load of 50 mN. The output of an instrumented indentation test is the load-displacement curve during loading and unloading of the indenter, as shown in Fig. 2  44. P max is the maximum indentation load, h max is the maximum indentation depth, h f is the final depth, the elastic strain energy W elastic is the area under the unloading curve (this area represents the elastic energy associated with residual stresses caused by indenter withdrawal), and the indentation absorbed energy W plastic is the area under the loading curve (this area represents the energy dissipated during indentation due to plastic deformation, cracking and crushing processes).

Figure 1
figure 1

Schematic of the pulsed laser butt welding. A laser beam of controllable power was used to weld metallic glass pieces. The orange arrow indicates the direction of the laser beam is travelling.

Full size image
Figure 2
figure 2

Schematic P-h curve for Berkovich indentation. P max is the maximum indentation load, h max is the maximum indentation depth, h f is the final depth, the elastic strain energy W elastic is the area under the unloading curve, and the indentation absorbed energy W plastic is the area under the loading curve.

Full size image

In Fig. 2, the loading portion of the load-displacement curve is often described by Meyer’s law45,

$$P={k}_{1}{h}^{n}$$
(3)

where P is the instantaneous load, k 1 is the loading curve constant, n is the loading exponent and h is the instantaneous depth. The total energy (characterizing energy-absorbing or energy-releasing events occurring beneath an indenter), W total, is obtained by integrating Eq. (3) from zero depth to h max,

$${W}_{{\rm{total}}}={\int }_{0}^{{h}_{{\rm{\max }}}}P{\rm{d}}h=\frac{{k}_{1}{h}_{{\rm{\max }}}^{n+1}}{n+1}$$
(4)

The unloading curve is described by the following expression46,

$$P={k}_{2}{(h-{h}_{f})}^{m}$$
(5)

where k 2 is the unloading curve constant and m is the unloading exponent. W elastic is obtained by integrating Eq. (5) from h f to h max,

$${W}_{{\rm{elastic}}}={\int }_{{h}_{{\rm{f}}}}^{{h}_{{\rm{\max }}}}P{\rm{d}}h=\frac{{k}_{2}{({h}_{{\rm{\max }}}-{h}_{{\rm{f}}})}^{m+1}}{m+1}$$
(6)

Bend tests were performed on bar shaped samples with 6 mm length, 0.6 mm width and 1 mm thickness that were bent around mandrels of different radii at room temperature. The strain to failure can be calculated from ε = h/2 R, where R is the neutral radius of the bend sample and h is the sample’s thickness. The morphology of fracture surface was observed using an OM.

Results and Discussion

The morphology of the weld seams of Pd43Cu27Ni10P20 BMG after BOP experiments processed in air was examined by an OM (Fig. 3). The appearance of BOP of Pd43Cu27Ni10P20 BMG obtained by the pulsed laser welding technique is uniform, does not exhibit porous and no visible oxidation can be detected. The cross-section of the weld seam was characterized using XRD (Fig. 4). For all considered processing parameters a weld was achieved which resulted in a broad halo, typical for an amorphous sample. This finding implies that the processing conditions and particularly cooling rate of the pulsed laser beam welding method are sufficient to avoid crystallization of Pd43Cu27Ni10P20 BMG.

Figure 3
figure 3

Morphology of bead-on-plate of Pd43Cu27Ni10P20 bulk metallic glass obtained by the pulsed laser welding method under 1050 W laser beam power.

Full size image
Figure 4
figure 4

X-ray diffraction pattern of the cross-section of bead-on-plate of Pd43Cu27Ni10P20 bulk metallic glass using the pulsed laser welding method with various powers.

Full size image

Butt welding of Pd43Cu27Ni10P20 BMG was carried out using the pulsed laser beam welding technique under 1050 W laser beam power. No visible defects were observed in the butt joint, suggesting a metallurgical bond. XRD patterns of FZ, HAZ and BM are shown in Fig. 5a, indicating the existence of a completely amorphous joint. To further confirm that the butt joint of Pd43Cu27Ni10P20 BMG does not contain any crystallinity, a comparison of the DSC traces of the FZ, HAZ and BM with 20 Kmin−1 was carried out (Fig. 5b). DSC curves exhibit the similar glass transition temperature, crystallization temperature and heat of crystallization of the different zones in the butt joint of Pd43Cu27Ni10P20 BMG. The measured glass transition temperature (T g) of the FZ and HAZ samples was 582 K and 581 K, respectively. The measured onset crystallization temperature (T x) of the FZ and HAZ samples was 676 K and 679 K, respectively. The measured heat of crystallization (ΔH) of the FZ and HAZ samples was both 76 Jg−1. Such values are essentially the same as the base material (T g = 580 K, T x = 679 K and ΔH = 77 Jg−1), suggesting that all parts are fully amorphous.

Figure 5
figure 5

X-ray diffraction patterns (a) and differential scanning calorimetry thermograms (b) of the base material (BM), heat-affected zone (HAZ) and fusion zone (FZ) in butt joint of Pd43Cu27Ni10P20 bulk metallic glass under laser beam power of 1050 W.

Full size image

To study the mechanical properties of the butt joint of Pd43Cu27Ni10P20 BMG, nanoindentation tests were carried out from the middle of the weld seam to the BM along the direction perpendicular to the weld seam. Figure 6 shows the variation of indentation modulus, hardness, and W elastic/W total (W total = W elastic + W plastic), as a function of the distance from the middle of the weld seam. Indentation energies are useful parameters in analyzing the mechanical behavior of materials, and W elastic/W total was linked to the material’s deformation recovery capability and the initial unloading stiffness47. The ratio energy unload to energy load values of FZ, HAZ and BM in the butt joint of Pd43Cu27Ni10P20 BMG were essentially identical. Similarly the indentation moduli and hardness of the FZ, HAZ and BM in the weld having amorphous structures did not reveal significant differences. This indicates that the material properties of the FZ and the HAZ are comparable to the base material.

Figure 6
figure 6

Indentation modulus, hardness and W elastic/W total, vs. the distance from the middle of weld seam, plot for the butt joint of Pd43Cu27Ni10P20 bulk metallic glass obtained by the pulsed laser welding method.

Full size image

A powerful test to determine mechanical property of the weld joint is bend test48,49,50,51. Butt joint and BM samples of Pd43Cu27Ni10P20 BMG exhibited significant bending ductility with a failure bending strain of about 7.5% for samples of 1 mm thickness. These numbers are comparable with highest reported bending strains for this alloy50. The micrographs of fractures of the bend samples are shown in Fig. 7. Typical vein patterns as found in ductile fracture, were observed over the entire fracture surface of BM sample of Pd43Cu27Ni10P20 BMG (Fig. 7b). The fracture position of the butt joint sample is located at the center of the weld seam (Fig. 7a), and the microstructures show multiple shear bands formation with shear band spacing of approximately 25 μm and 60 μm for the butt joint and BM samples, respectively. This confirms that the butt joint of Pd-based BMG possesses the high ductility.

Figure 7
figure 7

Fracture morphology after bend tests of Pd43Cu27Ni10P20 bulk metallic glass: (a) weld joint; (b) base material.

Full size image

Summary

We used the pulsed laser beam welding method to join Pd43Cu27Ni10P20 BMG. For a range of welding parameters completely amorphous welds can be achieved as quantified by XRD and thermal analysis. The mechanical properties of the weld are comparable with that of the bulk material suggesting that pulsed laser beam welding is a versatile method to join BMGs. For example, it can be used for additive manufacturing to fabricate on demand geometries with superb mechanical properties.