Thank you for visiting Nature.com.The browser version you are using has limited support for CSS.For the best experience, we recommend that you use an updated browser (or turn off compatibility mode in Internet Explorer).In the meantime, to ensure continued support, we will display the site without styles and JavaScript.
A new mechanism based on selective laser melting to control the microstructure of products in the manufacturing process is proposed.The mechanism relies on the generation of high-intensity ultrasonic waves in the molten pool by complex intensity-modulated laser irradiation.Experimental studies and numerical simulations show that this control mechanism is technically feasible and can be effectively integrated into the design of modern selective laser melting machines.
Additive manufacturing (AM) of complex-shaped parts has grown significantly in recent decades.However, despite the variety of additive manufacturing processes, including selective laser melting (SLM)1,2,3, direct laser metal deposition4,5,6, electron beam melting7,8 and others9,10, the Parts may be defective.This is mainly due to the specific characteristics of the molten pool solidification process associated with high thermal gradients, high cooling rates, and the complexity of heating cycles in melting and remelting the material 11 , which lead to epitaxial grain growth and significant porosity. 12,13 showed that it is necessary to control thermal gradients, cooling rates, and alloy composition, or apply additional physical shocks by external fields of various properties, such as ultrasound, to achieve fine equiaxed grain structures.
Numerous publications are concerned with the effect of vibration treatment on the solidification process in conventional casting processes14,15.However, applying an external field to a bulk melt does not produce the desired material microstructure.If the volume of the liquid phase is small, the situation changes dramatically.In this case, the external field significantly affects the solidification process.Intense sound fields16,17,18,19,20,21,22,23,24,25,26,27, arc stirring28 and oscillation29, electromagnetic effects during pulsed plasma arcs30,31 and other methods32 have been considered .Attach to the substrate using an external high-intensity ultrasound source (at 20 kHz).The ultrasound-induced grain refinement is attributed to the increased constitutive subcooling zone due to the reduced temperature gradient and ultrasound enhancement to generate new crystallites through cavitation.
In this work, we investigated the possibility of altering the grain structure of austenitic stainless steels by sonicating the molten pool with sound waves generated by the melting laser itself.The intensity modulation of the laser radiation incident on the light-absorbing medium results in the generation of ultrasonic waves, which alter the microstructure of the material.This intensity modulation of laser radiation can be easily integrated into existing SLM 3D printers.The experiments in this work were performed on stainless steel plates whose surfaces were exposed to intensity-modulated laser radiation.So, technically, laser surface treatment is done.However, if such a laser treatment is performed on the surface of each layer, during layer-by-layer build-up, effects on the entire volume or on selected parts of the volume are achieved.In other words, if the part is constructed layer by layer, the laser surface treatment of each layer is equivalent to “laser volume treatment”.
Whereas in ultrasonic horn-based ultrasonic therapy, the ultrasonic energy of the standing sound wave is distributed throughout the component, while the laser-induced ultrasonic intensity is highly concentrated near the point where the laser radiation is absorbed.Using a sonotrode in an SLM powder bed fusion machine is complicated because the top surface of the powder bed exposed to the laser radiation should remain stationary.In addition, there is no mechanical stress on the top surface of the part.Therefore, the acoustic stress is close to zero and the particle velocity has a maximum amplitude over the entire top surface of the part.The sound pressure inside the entire molten pool cannot exceed 0.1% of the maximum pressure generated by the welding head, because the wavelength of ultrasonic waves with a frequency of 20 kHz in stainless steel is \(\sim 0.3~\text {m}\), and the The depth is usually less than \(\sim 0.3~\text {mm}\).Therefore, the effect of ultrasound on cavitation may be small.
It should be noted that the use of intensity-modulated laser radiation in direct laser metal deposition is an active area of ​​research35,36,37,38.
The thermal effects of laser radiation incident on the medium are the basis for almost all material processing laser techniques 39, 40, such as cutting 41, welding, hardening, drilling 42, surface cleaning, surface alloying, surface polishing 43, etc.materials processing technology and summarized preliminary results in many reviews and monographs 44, 45, 46.
It should be noted that any non-stationary action on the medium, including lasing action on the absorbing medium, results in the excitation of acoustic waves in it with more or less efficiency.Initially, the main focus was on the laser excitation of waves in liquids and the various thermal excitation mechanisms of sound (thermal expansion, evaporation, volume change during phase transition, contraction, etc.) 47, 48, 49.Numerous monographs50, 51, 52 provide theoretical analyses of this process and its possible practical applications.
These issues were subsequently discussed at various conferences, and laser excitation of ultrasound has applications in both industrial applications of laser technology53 and medicine54.Therefore, it can be considered that the basic concept of the process by which pulsed laser light acts on an absorbing medium has been established.Laser ultrasonic inspection is used for defect detection of SLM-manufactured samples55,56.
The effect of laser-generated shock waves on materials is the basis of laser shock peening57,58,59, which is also used for the surface treatment of additively manufactured parts60.However, laser shock strengthening is most effective on nanosecond laser pulses and mechanically loaded surfaces (eg, with a layer of liquid)59 because mechanical loading increases peak pressure.
Experiments were conducted to investigate the possible effects of various physical fields on the microstructure of solidified materials.The functional diagram of the experimental setup is shown in Figure 1.A pulsed Nd:YAG solid-state laser operating in free-running mode (pulse duration \(\tau _L \sim 150~\upmu \text {s}\ )) was used.Each laser pulse is passed through a series of neutral density filters and a beam splitter plate system.Depending on the combination of neutral density filters, the pulse energy on the target varies from \(E_L \sim 20~\text {mJ}\) to \(E_L \sim 100~\text {mJ}\) .The laser beam reflected from the beam splitter is fed to a photodiode for simultaneous data acquisition, and two calorimeters (photodiodes with a long response time exceeding \(1~\text {ms}\)) are used to determine the incident to and reflected from the target, and two power meters (photodiodes with short response times\(<10~\text {ns}\)) to determine incident and reflected optical power.Calorimeters and power meters were calibrated to give values ​​in absolute units using a thermopile detector Gentec-EO XLP12-3S-H2-D0 and a dielectric mirror mounted at the sample location.Focus the beam onto the target using a lens (Antireflection coating at \(1.06 \upmu \text {m}\), focal length \(160~\text {mm}\)) and a beam waist at the target surface 60– \(100~\upmu\text {m}\).
Functional schematic diagram of the experimental setup: 1—laser; 2—laser beam; 3—neutral density filter; 4—synchronized photodiode; 5—beam splitter; 6—diaphragm; 7—calorimeter of incident beam ; 8 – calorimeter of reflected beam; 9 – incident beam power meter; 10 – reflected beam power meter; 11 – focusing lens; 12 – mirror; 13 – sample; 14 – broadband piezoelectric transducer; 15 – 2D converter; 16 – positioning microcontroller; 17 – synchronization unit; 18 – multi-channel digital acquisition system with various sampling rates; 19 – personal computer.
Ultrasonic treatment is carried out as follows.The laser operates in free-running mode; therefore the duration of the laser pulse is \(\tau _L \sim 150~\upmu \text {s}\), which consists of multiple durations of approximately \(1.5~\upmu \text {s } \) each.The temporal shape of the laser pulse and its spectrum consist of a low-frequency envelope and a high-frequency modulation, with an average frequency of about \(0.7~\text {MHz}\), as shown in Figure 2.- The frequency envelope provides the heating and subsequent melting and evaporation of the material, while the high frequency component provides the ultrasonic vibrations due to the photoacoustic effect.The waveform of the ultrasonic pulse generated by the laser is mainly determined by the time shape of the laser pulse intensity. It is from \(7~\text {kHz}\) to \ (2~\text {MHz}\), and the center frequency is \(~ 0.7~\text {MHz}\).Acoustic pulses due to the photoacoustic effect were recorded using broadband piezoelectric transducers made of polyvinylidene fluoride films.The recorded waveform and its spectrum are shown in Figure 2.It should be noted that the shape of the laser pulses is typical of a free-running mode laser.
Temporal distribution of laser pulse intensity (a) and sound velocity (b) on the rear surface of the sample, the spectra (blue curve) of a single laser pulse (c) and an ultrasound pulse (d) averaged over 300 laser pulses (red curve) .
We can clearly distinguish the low-frequency and high-frequency components of the acoustic treatment corresponding to the low-frequency envelope of the laser pulse and the high-frequency modulation, respectively.The wavelengths of the acoustic waves generated by the laser pulse envelope exceed \(40~\text {cm}\); therefore, the main effect of the broadband high-frequency components of the acoustic signal on the microstructure is expected.
The physical processes in SLM are complex and occur simultaneously on different spatial and temporal scales.Therefore, multi-scale methods are most suitable for theoretical analysis of SLM.Mathematical models should initially be multi-physical.The mechanics and thermophysics of a multiphase medium “solid-liquid melt” interacting with an inert gas atmosphere can then be effectively described.The characteristics of material thermal loads in SLM are as follows.
Heating and cooling rates up to \(10^6~\text {K}/\text {s}\) /\text{ due to localized laser irradiation with power densities up to \(10^{13}~\text {W} cm}^2\).
The melting-solidification cycle lasts between 1 and \(10~\text {ms}\), which contributes to the rapid solidification of the melting zone during cooling.
Rapid heating of the sample surface results in the formation of high thermoelastic stresses in the surface layer.Sufficient (up to 20%) portion of the powder layer is strongly evaporated63, which results in an additional pressure load on the surface in response to laser ablation.Consequently, the induced strain significantly distorts the part geometry, especially near supports and thin structural elements.The high heating rate in pulsed laser annealing results in the generation of ultrasonic strain waves that propagate from the surface to the substrate.In order to obtain accurate quantitative data on the local stress and strain distribution, a mesoscopic simulation of the elastic deformation problem conjugated to heat and mass transfer is performed.
The governing equations of the model include (1) unsteady heat transfer equations where thermal conductivity depends on phase state (powder, melt, polycrystalline) and temperature, (2) fluctuations in elastic deformation after continuum ablation and thermoelastic expansion equation.The boundary value problem is determined by experimental conditions.The modulated laser flux is defined on the sample surface.Convective cooling includes conductive heat exchange and evaporative flux.The mass flux is defined based on the calculation of the saturated vapor pressure of the evaporating material.The elastoplastic stress-strain relationship is used where the thermoelastic stress is proportional to the temperature difference.For nominal power \(300~\text {W}\), frequency \(10^5~\text {Hz}\), intermittent coefficient 100 and \(200~\upmu \text {m}\ ) of the effective beam diameter.
Figure 3 shows the results of numerical simulation of the molten zone using a macroscopic mathematical model.The diameter of the fusion zone is \(200~\upmu \text {m}\) (\(100~\upmu \text { m}\) radius) and \(40~\upmu \text {m}\) depth.The simulation results show that the surface temperature varies locally with time as \(100~\text {K}\) due to the high intermittent factor of the pulse modulation.The heating \(V_h\) and cooling \(V_c\) rates are on the order of \(10^7\) and \(10^6~\text {K}/\text {s}\), respectively.These values ​​are in good agreement with our previous analysis64.An order of magnitude difference between \(V_h\) and \(V_c\) results in rapid overheating of the surface layer, where thermal conduction to the substrate is insufficient to remove the heat.Therefore, at \(t=26~\upmu \text {s}\) the surface temperature peaks as high as \(4800~\text {K}\).Vigorous evaporation of the material can cause the sample surface to be subjected to excessive pressure and peel off.
Numerical simulation results of melting zone of single laser pulse annealing on 316L sample plate.The time from the beginning of the pulse to the depth of the molten pool reaching the maximum value is \(180~\upmu\text {s}\).The isotherm\(T = T_L = 1723~\text {K}\) represents the boundary between the liquid and solid phases.The isobars (yellow lines) correspond to the yield stress calculated as a function of temperature in the next section.Therefore, in the domain between the two isolines (isotherms\(T=T_L\) and isobars\(\sigma =\sigma _V(T)\)), the solid phase is subjected to strong mechanical loads , which may lead to changes in the microstructure.
This effect is further explained in Figure 4a, where the pressure level in the molten zone is plotted as a function of time and distance from the surface.First, the pressure behavior is related to the modulation of the laser pulse intensity described in Figure 2 above.A maximum pressure \text{s}\) of about \(10~\text {MPa}\) was observed at about \(t=26~\upmu).Second, the fluctuation of the local pressure at the control point has the same oscillation characteristics as the frequency of \(500~\text {kHz}\).This means that ultrasonic pressure waves are generated at the surface and then propagate into the substrate.
The calculated characteristics of the deformation zone near the melting zone are shown in Fig. 4b.Laser ablation and thermoelastic stress generate elastic deformation waves that propagate into the substrate.As can be seen from the figure, there are two stages of stress generation.During the first phase of \(t < 40~\upmu \text {s}\), the Mises stress rises to \(8~\text {MPa}\) with a modulation similar to the surface pressure.This stress occurs due to laser ablation, and no thermoelastic stress was observed in the control points because the initial heat-affected zone was too small.When heat is dissipated into the substrate, the control point generates high thermoelastic stress above \(40~\text {MPa}\).
The obtained modulated stress levels have a significant impact on the solid-liquid interface and may be the control mechanism governing the solidification path.The size of the deformation zone is 2 to 3 times larger than that of the melting zone.As shown in Figure 3, the location of the melting isotherm and the stress level equal to the yield stress are compared.This means that the pulsed laser irradiation provides high mechanical loads in localized areas with an effective diameter between 300 and \(800~\upmu \text {m}\) depending on the instantaneous time.
Therefore, the complex modulation of the pulsed laser annealing leads to the ultrasonic effect.The microstructure selection pathway is different if compared to the SLM without ultrasonic loading.Deformed unstable regions lead to periodic cycles of compression and stretching in the solid phase.Thus, the formation of new grain boundaries and subgrain boundaries becomes feasible.Therefore, the microstructural properties can be intentionally changed, as shown below.The obtained conclusions provide the possibility to design a pulse modulation-induced ultrasound-driven SLM prototype.In this case, the piezoelectric inductor 26 used elsewhere can be excluded.
(a) Pressure as a function of time, calculated at different distances from the surface 0, 20 and \(40~\upmu \text {m}\) along the axis of symmetry.(b) Time-dependent Von Mises stress calculated in a solid matrix at distances 70, 120 and \(170~\upmu \text {m}\) from the sample surface.
Experiments were performed on AISI 321H stainless steel plates with dimensions \(20\times 20\times 5~\text {mm}\).After each laser pulse, the plate moves \(50~\upmu \text {m}\), and the laser beam waist on the target surface is about \(100~\upmu \text {m}\).Up to five subsequent beam passes are performed along the same track to induce remelting of the processed material for grain refinement.In all cases, the remelted zone was sonicated, depending on the oscillatory component of the laser radiation.This results in a more than 5-fold reduction in average grain area.Figure 5 shows how the microstructure of the laser-melted region changes with the number of subsequent remelting cycles (passes).
Subplots (a,d,g,j) and (b,e,h,k) – microstructure of laser melted regions, subplots (c,f,i,l) – area distribution of colored grains. Shading represents the particles used to compute the histogram.Colors correspond to grain regions (see the color bar at the top of the histogram. Subplots (ac) correspond to untreated stainless steel, and subplots (df), (gi), (jl) correspond to 1, 3 and 5 remelts.
Since the laser pulse energy does not change between subsequent passes, the depth of the molten zone is the same.Thus, the subsequent channel completely “covers” the previous one.However, the histogram shows that the mean and median grain area decreases with increasing number of passes.This may indicate that the laser is acting on the substrate rather than the melt.
Grain refinement may be caused by rapid cooling of the molten pool65.Another set of experiments was carried out in which the surfaces of stainless steel plates (321H and 316L) were exposed to continuous wave laser radiation in atmosphere (Fig. 6) and vacuum (Fig. 7).The average laser power (300 W and 100 W, respectively) and molten pool depth are close to the experimental results of the Nd:YAG laser in free-running mode.However, a typical columnar structure was observed.
Microstructure of the laser-melted region of a continuous wave laser (300 W constant power, 200 mm/s scan speed, AISI 321H stainless steel).
(a) Microstructure and (b) electron backscatter diffraction image of the laser melting zone of vacuum continuous wave laser (constant power 100 W, scanning speed 200 mm/s, AISI 316L stainless steel) \ (\sim 2~\text {mbar }\).
Therefore, it is clearly shown that the complex modulation of the laser pulse intensity has a significant effect on the resulting microstructure.We believe that this effect is mechanical in nature and occurs due to the generation of ultrasonic vibrations propagating from the irradiated surface of the melt deep into the sample.Similar results were obtained in 13, 26, 34, 66, 67 using external piezoelectric transducers and sonotrodes providing high-intensity ultrasound in various materials including Ti-6Al-4V alloy 26 and stainless steel 34 the result of.The possible mechanism is speculated as follows.Intense ultrasound can cause acoustic cavitation, as demonstrated in ultrafast in situ synchrotron X-ray imaging.The collapse of the cavitation bubbles in turn generates shock waves in the molten material, whose front pressure reaches about \(100~\text {MPa}\)69.Such shock waves may be strong enough to promote the formation of critical-sized solid-phase nuclei in bulk liquids, disrupting the typical columnar grain structure of layer-by-layer additive manufacturing.
Here, we propose another mechanism responsible for structural modification by intense sonication.The material just after solidification is at a high temperature close to the melting point and has an extremely low yield stress.Intense ultrasonic waves can cause plastic flow to alter the grain structure of the hot material just solidified.However, reliable experimental data on the temperature dependence of yield stress are available at \(T\lesssim 1150~\text {K}\) (see Figure 8).Therefore, to test the hypothesis, we performed molecular dynamics (MD) simulations of a Fe-Cr-Ni composition similar to AISI 316 L steel in order to evaluate the yield stress behavior near the melting point.To calculate the yield stress, we used the MD shear stress relaxation technique detailed in 70, 71, 72, 73.For the interatomic interaction calculations, we used the Embedded Atomic Model (EAM) from 74.MD simulations were performed using LAMMPS codes 75,76.Details of the MD simulation will be published elsewhere.The MD calculation results of yield stress as a function of temperature are shown in Fig. 8 together with available experimental data and other evaluations77,78,79,80,81,82.
Yield stress for AISI grade 316 austenitic stainless steel and model composition versus temperature for MD simulations.Experimental measurements from references: (a) 77, (b) 78, (c) 79, (d) 80, (e) 81.refer to.(f)82 is an empirical model of yield stress-temperature dependence for in-line stress measurement during laser-assisted additive manufacturing.The large-scale MD simulation results in this study are denoted as \(\vartriangleleft\) for a defect-free infinite single crystal and \(\vartriangleright\) for finite grains taking into account the average grain size via the Hall-Petch relation Dimensions\(d = 50~\upmu \text {m}\).
It can be seen that at \(T>1500~\text {K}\) the yield stress drops below \(40~\text {MPa}\).On the other hand, estimates predict that the laser-generated ultrasonic amplitude exceeds \(40~\text {MPa}\) (see Fig. 4b), which is sufficient to induce plastic flow in the hot material just solidified.
The microstructure formation of 12Cr18Ni10Ti (AISI 321H) austenitic stainless steel during SLM was experimentally investigated using a complex intensity-modulated pulsed laser source.
Grain size reduction in the laser melting zone was found due to continuous laser remelting after 1, 3 or 5 passes.
Macroscopic modeling shows that the estimated size of the region where ultrasonic deformation may positively affect the solidification front is up to \(1~\text {mm}\).
The microscopic MD model shows that the yield strength of AISI 316 austenitic stainless steel is significantly reduced to \(40~\text {MPa}\) near the melting point.
The obtained results suggest a method for controlling the microstructure of materials using complex modulated laser processing and could serve as the basis for creating new modifications of the pulsed SLM technique.
Liu, Y. et al.Microstructural evolution and mechanical properties of in situ TiB2/AlSi10Mg composites by laser selective melting [J].J. Alloys.compound.853, 157287. https://doi.org/10.1016/j.jallcom.2020.157287 (2021).
Gao, S. et al.Recrystallization grain boundary engineering of laser selective melting of 316L stainless steel [J]. Journal of Alma Mater.200, 366–377.https://doi.org/10.1016/j.actamat.2020.09.015 (2020).
Chen, X. & Qiu, C. In situ development of sandwich microstructures with enhanced ductility by laser reheating of laser-melted titanium alloys.science.Rep. 10, 15870.https://doi.org/10.1038/s41598-020-72627-x (2020).
Azarniya, A. et al.Additive manufacturing of Ti-6Al-4V parts by laser metal deposition (LMD): process, microstructure and mechanical properties.J. Alloys.compound.804, 163–191.https://doi.org/10.1016/j.jallcom.2019.04.255 (2019).
Kumara, C. et al.Microstructural modeling of laser metal powder directed energy deposition of Alloy 718.Add to.manufacture.25, 357–364.https://doi.org/10.1016/j.addma.2018.11.024 (2019).
Busey, M. et al.Parametric Neutron Bragg Edge Imaging Study of Additively Manufactured Samples Treated by Laser Shock Peening.science.Rep. 11, 14919.https://doi.org/10.1038/s41598-021-94455-3 (2021).
Tan, X. et al.Gradient microstructure and mechanical properties of Ti-6Al-4V additively fabricated by electron beam melting.Alma Mater Journal.97, 1-16.https://doi.org/10.1016/j.actamat.2015.06.036 (2015).