Additive manufacturing tool with laser induced breakdown spectroscopy

Abstract

The present invention relates to an additive manufacturing tool. The tool comprises: a beam source (1) for creating a beam to produce a solid part by melting or fusing a powder feedstock in a build region (2) that is heated by the beam, and a laser induced breakdown spectroscopy, LIBS, subsystem (200). The LIBS subsystem (200) comprises a LIBS laser (7) to illuminate a vapour originating from the build region (2), and produce a plasma in a detection region (16), a detector configured to detect light in the detection region (16) emitted by the plasma, and a processor configured to receive spectroscopic data from the detector, the spectroscopic data being indicative of a spectrum of the light emitted by the plasma, wherein the detection region is spaced apart from the build region.

Description

[0001] ADDITIVE MANUFACTURING TOOE

[0002] Field of Invention

[0003] The invention relates to an additive manufacturing tool, and in particular, analysing the composition of vapours produced during additive manufacturing.

[0004] Background

[0005] Additive manufacturing (AM) is the construction of a three-dimensional (3D) object from a digital model, such as a computer-aided design (CAD) model. AM is typically a computer-controlled process that creates the 3D objects by depositing materials, usually in layers. The layers of material are fused or joined together to form the objects.

[0006] There are many subsets of AM. Directed Energy Deposition, also known as laser cladding or laser metal deposition, involves the feeding of a stream of metallic powder or wire into a molten pool The molten pool is generated by a heat source, such as a laser beam, an electron beam, arc, and plasma, as it scans across the target surface, depositing a coating of the chosen material. Powder Bed Fusion (PBF), another subset of AM techniques, involves building components using layer-wise melting of feedstock powder (e.g. a metal powder) using a laser or an electron beam. An AM toolpath may be defined based on slicing up a CAD model, and then raster scanning across each slice. AM has a wide range of product applications, including automotive, aerospace and biomedical, which may have stringent quality requirements.

[0007] To effectively meet such quality requirements, any defect formation mechanisms associated with AM must be understood. One potential source of defects is vaporisation during printing, which may result in compositional changes and the formation of imperfections. Vapours may also form a plume that attenuates, reflects, or refracts the laser beam. The vapour constituents may also result in a hazardous environment.

[0008] In general, a vapour plume is generated during AM when a melt pool exceeds its boiling point. The melt pool being the region of molten metal in proximity to the interface between the laser and material). The extent of plume generation may vary as a function of melt pool temperature. The vaporisation behaviour may change owing to environmental factors, such as chamber pressure.

[0009] During processing, the metallic vapours strongly mix with the shielding gas within the AM chamber, creating turbulent fluid flows, vapour-induced porosity, spatter and denudation, which can lead to lack-of-fusion defects. To minimise these vapour plume effects, a betterunderstanding of the volume and composition of vapour under different melting regimes is needed.

[0010] Optical Emission Spectroscopy (OES) and X-ray fluorescence spectroscopy (XRF) have been used to examine the chemical composition of vapour plumes released during AM processes. A typical OES apparatus may qualitatively analyse a limited number of elements (typically 1 - 2 elements, such as Fe and Cr) at an acquisition speed of up to 100 kHz but at low resolution. A typical XRF apparatus has a low data acquisition speed (utilising cumulative measurements over 163 s, i.e., 0.006 Hz) and cannot detect light elements with an atomic mass, Z < 11.

[0011] Although progress has been made to observe and quantify the vapours released during AM processes, considerable room for improvement remains.

[0012] Summary

[0013] According to an aspect of the present invention, there is provided an additive manufacturing tool, comprising: a beam source for creating a beam that is operable to produce a solid part by melting or fusing a feedstock in a build region that is heated by the beam; and a laser induced breakdown spectroscopy (LIBS) subsystem, with the LIBS subsystem comprising: a LIBS laser configured to generate a light beam to illuminate a vapour originating from the build region, and produce a plasma in a detection region; a detector configured to detect light in the detection region emitted by the plasma; and a processor configured to receive spectroscopic data from the detector, wherein the spectroscopic data is indicative of a spectrum of the light emitted by the plasma in the detection region, wherein the detection region is spaced apart from the build region.

[0014] This may allow for the composition of the vapour produced by the additive manufacturing process to be accurately determined by analysing the light emitted from the plasma ignited in the vapour by the LIBS laser. This may allow the efficiency of the additive manufacturing process to be monitored, and for any harmful fumes that are released by the process to be detected. The plasma is produced within the vapour, away from the source of the build region which may reduce any interference associated with the heating of the feedstock and the AM melting / fusing process. The AM beam source may be a laser configured to generate a beam of electromagnetic waves, or it may be an electron beam source configured to generate an electron beam.

[0015] The additive manufacturing tool may further comprise a build chamber housing the build region, wherein the detection region may be outside of the build chamber. The detectionregion, in which the detector is configured to detect the light emitted from the plasma, may be placed outside of the build chamber such that any light emitted from the melting or fusing of the feedstock does not interfere with the light emitted from the plasma and detected by the detector. The build chamber ensures that the vapour produced by the melting or fusing of the powder feedstock is not released into the atmosphere, such that, any harmful gases produced by the process are contained within the chamber. A line of sight between the detection region and the build region may be blocked by the configuration of the build chamber. The build chamber may comprise a baffle arrangement to block the light of sight.

[0016] The additive manufacturing tool may further comprise a detection chamber, secondary to the build chamber, housing the detection region.

[0017] Preferably, the additive manufacturing tool may further comprise an exhaust configured to direct the vapour from the build region to an outlet, wherein the detection region may be positioned in the exhaust. In this way, the detection region is placed within the exhaust away from the build region, which may reduce interference from the melting or fusing occurring in the build region is detected by the detector. This may improve the spectroscopic data detected by the detector.

[0018] Preferably, the additive manufacturing tool may further comprise a build chamber housing the build region, wherein the exhaust may be a manifold coupled to an opening of the build chamber and is configured to direct the vapour out from the build chamber through the opening.

[0019] Preferably, the detection region is at least 2cm from the build region.

[0020] Preferably, the LIBS subsystem may further comprise a lens and an optical fibre, wherein the lens is configured to collect the light emitted by the plasma, and the optical fibre provides at least a portion of a detection optical path to the detector.

[0021] Preferably, the LIBS subsystem may further comprise a mirror configured to reflect the collected light from the lens towards the optical fibre.

[0022] Preferably, the additive manufacturing tool may further comprise a focussing lens configured to focus the light beam from the LIBS laser into the detection region. The focussing lens may comprise more than one lens element.Preferably, the processor may be configured to determine a composition of the vapour by analysis of a spectrum of light obtained from the spectroscopic data. The analysis of the spectrum of light may include, for example, a hybrid DCSP plot calibration technique that identifies the composition of an unknown vapour from its known solid feedstock composition.

[0023] Preferably, the processor may be further configured to: i) deconvolve the spectrum into peaks; and ii) analyse the constituents of the vapour with reference to the peaks .

[0024] Preferably, step i) may comprise fitting the one or more peaks to the spectrum. The fitting of the one or more peaks to the spectrum may comprise fitting a Gaussian, Lorentzian, Voigt, or combination of different peak fitting functions. The fitting of one or more peaks to the spectrum may comprise fitting one or more peaks at wavelengths that are expected based on an anticipated composition of the vapour. Step ii) may comprise determining a concentration of at least one component of the vapour.

[0025] Preferably, step i) may comprise: dividing the spectrum into a plurality of spectral windows, each spectral window spanning a different portion of the spectrum; and fitting one or more peaks to each spectral window. This action may allow for emission peaks to be more easily detected and fitted, which may improve the deconvolution time of the spectrum. Each of the windows may span overlapping portions of the spectrum or each window may span distinct, separate portions of the spectrum.

[0026] The additive manufacturing tool may further comprise: a build chamber housing the build region; a detection chamber housing the detection region; and a first exhaust part comprising a first end a second end, wherein the first end of the exhaust part is coupled to the build chamber and the second end of the first exhaust part is coupled to the detection chamber, and wherein the first exhaust part is configured to direct vapour from the build chamber to the detection chamber.

[0027] The additive manufacturing tool may further comprise a second exhaust part coupled to the detection chamber and configured to direct vapour from the second chamber to an outlet.

[0028] Preferably, the additive manufacturing tool may further comprise: a camera directed towards the detection region and configured to capture an image of the plasma. This may allow for the plasma to be observed, such as to determine the lifetime of the plasma. The high-speed camera may capture images at a minimum of lOOOfps. The camera may be configured to capture one or more singular images or a video comprising a plurality of images. The additivemanufacturing tool may further comprise a camera directed towards the build region configured to capture an image of the melting or fusing of the feedstock. The camera may be part of a sperate subsystem used in conjunction with he LIBS subsystem.

[0029] Preferably, the additive manufacturing tool may be a laser powder bed fusion (LPBF) tool or a direct metal deposition tool. The additive manufacturing tool may be a machine that builds up components in a layer-wise manner, using a beam source to melt and re -solidify feedstock materials, e.g. , but not limited to, laser metal deposition, electron beam powder bed fusion or a laser cladding tool.

[0030] Brief Description of Drawings

[0031] Figure 1 is a schematic diagram of an additive manufacturing (AM) tool with LIBS subsystem according to the present disclosure;

[0032] Figure 2 shows a perspective view of the AM tool shown in Figure 1;

[0033] Figure 3 shows a side view of a laser-induced breakdown spectroscopy (LIBS) subsystem of the AM tool shown in Figures 1 to 2;

[0034] Figure 4 shows an example triggering diagram;

[0035] Figure 5 shows a background (argon only) and a IN625 vapour emission spectrum;

[0036] Figure 6 shows a result of a spectrum peak fitting estimation alongside the peak elemental identifications, using the expected peak locations from a database as the fitting parameter;

[0037] Figure 7 shows Saha-Boltzmann plot results for solid samples and vapour samples, over NiCoCr alloy and IN625 alloy;

[0038] Figure 8 shows the calibration plots used to scale the unknown vapour integrals against the elemental integrals from a known alloy material of similar composition (both NiCoCr alloy and IN625 alloy);

[0039] Figure 9 shows a comparison of emission peaks in vapour sample and in a solid standard reference material (NiCoCr alloy);Figure 10 shows LIBS results of a sample standard reference material welding experiment for major alloying components (Ni, Cr and Co), compared to the changes in the weld bulk composition as measured by Electron Dispersive Spectroscopy (EDS); and

[0040] Figure 11 shows LIBS results of LPBF vapour for large 3 -layer builds.

[0041] Detailed Description

[0042] Additive manufacturing tool

[0043] Figure 1 shows an additive manufacturing tool 100 according to the present disclosure. The additive manufacturing tool 100 includes a beam source 1, a build chamber 4 and a build region 2. The build region 2 is housed within the build chamber 4.

[0044] The AM tool 100 is a laser powder bed fusion (LPBF) tool. A powder feedstock is placed in the build region 2 of the build chamber 4. The powder feedstock is a powdered material in which the additive manufacturing tool 100 uses to fuse together and form a 3D object. The powder feedstock may be, e.g. , metal alloy powder, such as Inconel 625.

[0045] The beam source 1 is a laser that generates a light beam. The laser 1 is configured to generate the beam to melt or fuse the powder feedstock placed in the build region 2. The tool 100 includes a mirror 10 (e.g. one or more galvo -mirrors) configured to reflect the beam generated by the laser 1 towards the build region 2 and onto the powder feedstock. The mirror 10 may be arranged to scan the beam over the build region 2, under the control of processor 12. The laser 1 may be, for example, a 1070nm fibre laser, and the laser beam focused onto the building region 2. Other types of beam source can be used, e.g., a different type of laser.

[0046] Prior to generating the beam and fusing the powder feedstock, the build chamber 4 may be purged with a shielding gas flow for a predetermined period of time. A pressure inside the build chamber 4 is also preferably maintained at a constant level, e.g., + 10 kPa, prior to the building process.

[0047] The tool 100 includes a high-speed coaxial optical imaging camera (not shown). The highspeed camera is configured to capture images and videos of the melting and fusing process occurring as the beam generated by the laser 1 strikes the feedstock powder. This can be used separately to observe the plasma light but is not essential.

[0048] The fusing or melting of the powder feedstock by the beam 1 creates a molten pool, which solidifies to produce a solid region. Multiple layers of solid regions are fused together, eachlayer at a time, by the laser beam 1 to construct a 3D object in the build chamber 4. However, the fusing or melting of the powder feedstock may produce a vapour from the build region 2 where the powder is fused together. Vapours generated during the AM process will be removed from the build chamber 4, carried by the shielding gas, through an exhaust 3.

[0049] The AM tool 100 includes an exhaust 3. The exhaust 3 is a manifold coupled to an opening 13 in a wall 14 of the build chamber 4. The exhaust 3 directs the vapour and shielding gas flow from the build region 2, out of the build chamber 4, towards an outlet.

[0050] The exhaust 3 directs the vapour produced by the fusing of the powder feedstock to a laser -induced breakdown spectroscopy (LIBS) subsystem. The LIBS subsystem 200 analyses the vapour produced in the build region 2, as will be explained in detail below.

[0051] The LIBS subsystem 200 of the AM tool 100 includes a LIBS laser 7, a mirror 15, optical elements 5, and a detection chamber 8. The LIBS laser 7 is a different laser, secondary to the laser 1 used in the AM process to fuse the powder feedstock. The LIBS laser 7 is a pulsed laser and is configured to generate a light beam which generates a plasma from the vapour in the detection chamber 8 at the laser focal point. The LIBS laser 7 in an example is operated at an energy of 3.12 mJ, and a repetition rate of 1000 Hz (but any suitable energy capable of generating a plasma from the vapour can be used). The mirror 15 reflects the light beam towards the detection chamber 8 and may provide for a more convenient layout of the various components, but is not essential.

[0052] The LIBS subsystem 200 includes a beam dump 11, which absorbs any excess light from the LIBS laser beam 7. The AM tool 100 is also equipped with a signal generator, that controls trigger signals to the AM laser 1 and the LIBS laser 7. In this example, the signal generator is comprised in processor 12.

[0053] The optical elements 5 in this example includes two-plano-convex lenses positioned at the centre of the detection chamber 8, but any suitable optics can be used (e.g. , convex mirrors etc). The two lenses in this example correct for the beam quality to form a desired beam focal spot within the detection region 16.

[0054] The detection region 16 is a region housed by the detection chamber 8, where the LIBS laser beam interacts with the vapour. The vapour has travelled from the build region 2, where it has been produced by the melting or fusing of the powder feedstock, through the exhaust 3and into the detection region 16. The detection region 16 is spaced apart from the build region 2. The detection region 16 may be, e.g. , at least 2cm away from the build region 2.

[0055] The detection region 16 is outside of the build chamber 4. The exhaust 3 acts as a passageway between the build chamber 4 and the detection chamber 8 for the vapour released by the AM process in the build region 2 to reach the detection region 16.

[0056] The exhaust 3 includes a first part 17 and a second part 18. The first part 17 of the exhaust 3 includes a first end 19 and a second end 20. The first end 19 of the first part 17 is coupled to the opening 13 of the build chamber 4, and the second end 20 is coupled to a first opening 21 in the detection chamber 8. The second part of the exhaust 3 includes a first end 22 coupled to a second opening 23 in the detection chamber 8. The second part of the exhaust 3 directs vapour from the detection region 16, through second part 18, to an outlet (e.g. a vapour extract system / exhaust stack / HEPA filter). In this way, the detection region 8, and the LIBS subsystem 200, is located in the exhaust 3, with parts 17, 18 of the exhaust 3 directing vapour into the detection region 8 and out of the detection region 8 respectively.

[0057] As the vapour is directed out from the build chamber 4 and into the detection chamber 8, the laser beam generated by the LIBS laser 7 produces a plasma from the vapour. The plasma vapour comprises ions and free electrons produced from ionisation of the vapour, but may not be entirely ionised. Immediately after the LIBS laser 7 pulse ends, the plasma will emit a light as a blackbody emitter, before cooling and emitting light at discrete wavelengths characteristic of the elements present. The plasma light is collected by detector 9 via a series of optical elements. The detector 9 can compromise a camera or spectrometer. When detector 9 comprises of a gated spectrometer, the signal will comprise of a spectrum showing discrete and characteristic peaks. The LIBS subsystem 200 determines the chemical composition of the plasma, and thus the vapour, by analysing light emitted by the plasma, as will be described in detail below.

[0058] The LIBS subsystem 200 may include a high-speed coaxial optical imaging camera (not shown), that captures images and videos of the plasma formed in the detection region 16. Such a camera can be useful for monitoring the lifetime of the plasma, which is 10 - 15 ps for example. In information gained using the camera is also used to optimise the trigger settings of the LIBS subsystem 200 for different materials. This is useful in the context of research, but may not be necessary in a product, in which suitable trigger settings may have been determined by experiment for the typical conditions of operation.The LIBS subsystem 200 further includes detection apparatus, herein referred to as a detector, to detect light in the detection region 16 that is emitted by the plasma. The detector includes an optical fibre 6 and a spectrometer 9. The optic fibre 6 provides an optical path for the light emitted in by the plasma in the detection region 16 to reach the spectrometer 9 positioned outside of the detection chamber 8. The processor 12 is configured to receive spectroscopic data from the spectrometer 9.

[0059] The detector includes a collimating lens (not shown) placed at a predetermined distance from the plasma interaction zone in the build region 2. The predetermined distance for the collimating lens may be 25mm, for example, from the centre of the plasma interaction zone, or a centre of the build region 2. The collimating lens collects and directs the light emitted by the plasma towards the acceptance region of the optical fibre. The nature of the optics for the LIBS subsystem may be different than this, depending on the LIBS laser that is used and the requirements of the system.

[0060] The detector may include a mirror (not shown) that reflects the collected light from the collimating lens towards an end 25 of the fibre optic cable 6. Optical elements forming the detection path (from the detection region to the spectrometer 9) may comprise one or more reflectors and lenses positioned within the detection chamber 8. The end 25 of the fibre optic cable 6 may be fastened to the detection chamber 8, and the rest of the fibre optic cable 6 outside of the detection chamber 8, as shown in Figure 1. The spectrometer 9 is positioned outside of the detection chamber 8.

[0061] In an example embodiment, a collimating lens is used, formed from a UV -grade fused silica (f = 25 mm, 0=6 mm) lens, with an aluminium mirror (with < 90 % reflectance over 200 -1100 nm) that reflects the light at 90°. The optical fibre 6 in the example embodiment is a 400 pm core fibre optic (pure silica core, coated in doped -fluorine silica cladding with polyamide buffer, 12.7° acceptance angle). The spectrometer 9 is calibrated over a spectral range of 187 - 872 nm using an Ar-Hg lamp.

[0062] The spectrometer 9 is configured to record spectroscopic data of the light emitted by the plasma and send the data to processor 12. The processor 12 receives the data and analyses it, as will be explained below in relation to Figures 5-11.

[0063] The spectroscopic data recorded by the spectrometer 9 includes spectra of the light emitted. The spectrometer 9 records an integration of the light emitted in the detection region 16 over a predetermined period of time. The predetermined time may be lOps, for example. Thespectrometer 9 records spectra of the emitted light over the lOps with a suitable gate delay (in this example 1.5 ps) after each LIBS laser 7 pulse, to avoid collecting the blackbody emission signal produced at the start of the plasma lifetime. The spectrometer 9 used in the example embodiment has a 21 ns trigger jitter and gate edge response of 585.85 ns.

[0064] The specifics of the LIBS subsystem 200 may be different in other embodiments - for example at least one of the LIBS laser, illumination optical path, detection optical path, optical elements and spectrometer may be different in other embodiments.

[0065] Figure 2 shows a perspective view of the example AM tool 100 described with reference in Figure 1. Figure 3 shows a side view of the LIBS subsystem 200 of the AM tool 100, as shown in Figures 1 to 2. The AM tool 100 includes the build chamber 4 with build region 2 and the LIBS subsystem 200 positioned within the exhaust 3.

[0066] As shown in Figure 2, the AM tool 100 further includes a powder hopper 26. The powder hopper 26 contains the powder feedstock and spreads a layer of powder onto the build region 2, each time a layer is to be added to the object being built by the AM tool 100.

[0067] Figures 2 and 3 also show the direction of flow of the vapour produced in the build region 2 as it exits the build chamber 4 through the exhaust 3. Arrow 201 shows the direction of movement of the vapour from the build region 2, where it is produced, towards the opening 13 in the build chamber 4. Arrow 202 shows the vapour exiting the build chamber 4 and moving through the exhaust 3 towards the LIBS subsystem 200. Arrow 203 shows the vapour leaving the LIBS subsystem 200 directed towards an outlet.

[0068] Figure 3 further shows an opening 27 in the detection chamber 8 through which the collimating lens apparatus and fibre optic end 25 is received into the detection chamber 8.

[0069] As shown in Figure 1 to 3 and described above, the LIBS subsystem 200 is spaced apart from the fusing or melting of the powder feedstock, and thus is spaced apart from where the vapour from the AM process is produced. Therefore, as the LIBS subsystem 200 analyses the chemical composition of the vapour (by illuminating the vapour to ignite light -emitting plasma), it does so away from the melting of the powder feedstock.

[0070] The vapour is sampled by the LIBS subsystem 200 without interfering with, or interference from, the build region 2. Therefore, the effect of the background radiation and light emitted by the process of LPBF laser 1 melting the substrate and powder feedstock on the spectrarecorded by the spectrometer 9 is significantly reduced. This may improve the clarity of emission peaks for elemental detection of the vapour, since emission spectra from the AM process will be detected to a lesser extent.

[0071] Figure 4 shows a triggering diagram for the LIBS system. The triggering diagram shown in Figure 4 is for a delay time of 5 = 1.5 ps and an integration time of a = 10 ps.

[0072] Example of using the LIBS subsystem for metal vapour analysis

[0073] In an example, the LIBS subsystem 200 can be used to monitor vapour plumes emitted during the welding of a NiCoCr NIST standard reference material alloy using the AM tool 100. The welding process is achieved by processing NiCoCr under the AM tool 100 without spreading a powder layer. The spectra can be analysed using a hybrid technique that couples One -Point Calibration (OPC) procedures with calibration curves.

[0074] Low-power High-power

[0075] Experiment Materials (360 / m-1) (800 / m-1)

[0076] LPBF-LIBS during bi- p = 400 W

[0077] x NiCoCr hatch welding v = 0.5 m s1

[0078] LPBF-LIBS during bi- P = 180VF P = 400 W

[0079] IN625 hatch 3 -layer LPBFTv = 0.5 m s1v = 0.5 m s1

[0080] Table 1 : The experiments detailed for proof -of-concept for the LIBS subsystem 200.

[0081] An experiment using the AM tool 100 and the LIBS subsystem 200 was executed on a solid sample of the NiCoCr alloy, of dimension 10 mm x 10 mm x 2 mm, placed at the centre of the detection chamber at a 45° angle, as shown in 301, from a direction of propagation of the LIBS laser in the detection region 8. A further experiment was also performed on the vapour generated during laser welding of the NiCoCr SRM sample within the build region 2. The welding laser 1 was operated at a 400 W power and 0.5 m s1scan speed. This laser energy density was set intentionally high for LPBF and welding applications in order to generate a measurable vapour plume. In both cases, the spectra of light emitted by the plasma in the detection region 16 were recorded under an Argon (Ar) atmosphere and with the LIBS laser 7 having a pulse energy of 3.12 mJ and 1000 Hz repetition rate. LIBS quantitative results after data processing are shown in Figure 11.

[0082] In a second example, the LIBS subsystem 200 can be used to monitor vapour plumes emitted during the LPBF of an Inconel 625 (IN625) alloy, selected due to its composition high-Zelements (Ni, Cr, Fe, Mo) and frequent use in commercial purposes. For the further experiments, a 60 im layer of IN625 powder, with a particle size distribution of 15 — 45 rm, was spread and fused atop of IN625 substrates using the AM tool 100. The LPBF printing experiment produced geometries of 4 x 4 mm bidirectional hatched tracks with a 80 pm hatch spacing. The hatches were processed under two linear energy densities, LED = P v~ for laser power, P, and scan speed, v, to represent two different melting modes: conduction (LED = 360 / m-1) and keyhole (LED = 800 / -1). In each experiment, the vapour composition in the detection region 8 of the LIBS subsystem 200 was continuously monitored for 5 s after the print was initiated. LIBS quantitative results after data processing are shown in Figure 12.

[0083] Data processing of LIBS

[0084]

[0085] A processing pipeline was used for data pre-processing and spectral deconvolution, as illustrated in Figures 5 and 6. This is implemented using any suitable processor, such as processor 12 in the example of Figure 1.

[0086] For the data pre-processing, a baseline removal is applied to each spectrum. Any suitable method can be used in principle. In this example a local -minima linear approximation is used for every 70 points. The filtered spectrum is smoothed by a Savitsky -Golav filter, before being normalised by dividing the spectrum by its sum integral. The pre-processing pipeline classifies each spectrum as a ‘background’ (such as argon only), where no metal vapour is contained within the sample volume, as shown in spectrum 502 of Figure 5, or a vapour (the metal emission peaks) spectrum, as shown in spectrum 502 of Figure 5.

[0087] To improve computational efficiency, only vapour spectra, such as that shown in spectrum 502, are retained for peak deconvolution. Each vapour spectrum is further reduced into small wavelength ranges of overlapping peaks, herein referred to as ‘windows’. Splitting each vapour spectrum into windows improves the deconvolution time. It will be appreciated that any suitable peak deconvolution method can be used, and the method used here is merely by way of example.

[0088] For each window, the expected emission peak locations are identified for every alloying element in that wavelength range from a database. The database comprises known wavelength emission peaks for a set of predetermined elements, which may be obtained from a suitable reference such as the NIST database. Using a curve fitting algorithm, such as a Lorentzianfit for each identified database peak is defined, and the summed spectral window (Smax) is calculated as shown in Equation 1 :

[0089] l=L

[0090] Smax =X n L2+ (x - c)

[0091] 12)

[0092] =0(1)

[0093] for number of Lorentz-shaped peaks, L. Each peak is characterised by the peak amplitude, a, the peak centre, c, and the full-width half-maximum of the peak, 2o\ An example 600 of the Lorentz fitting result is shown in Figure 6.

[0094] To calculate the deconvoluted window for the whole dataset in a computationally efficient manner, an Ordinary Least Squares approach is used, and the algorithm solves Equation 2 as:

[0095] Sn— x■ max (2)

[0096] For each spectrum in the dataset, Sn, the Normal Equation is solved to find vector x, where x =

[0097]

[0098] The integrals of the peaks in each spectral window are now described by the solution Snand the approach is repeated over all windows. This approach affords good computational efficiency.

[0099] For qualitative analysis, the deconvoluted data is matched against the NIST LIBS database for each element (e.g., Ni, Cr, Mo, Fe for Inconel 625), and a closest matching element is returned based on each peak wavelength. Peak height thresholding is applied to each NIST elemental database spectrum, to select the significant spectral peaks. These database peaks are compared to the experimental data, allowing identification of the elements present within the plasma volume, and thus the elements present in the vapour.

[0100] An example 600 showing a comparison of the peak deconvolution fit 601 using known database peaks 605 against the IN625 experimental data 602 from the AM tool 100 is shown in Figure 6. The example 604 (z.e., a window 603 of the whole spectrum 600 detected) shows Lorentzian fit for peaks of Ni 605, Cr 606, Fe 607 and Mo 608. The deconvoluted for 609 matches the experimental data 610 with a high R2value of 0.958.

[0101] LIBS calibration plot construction and quantitative analysis

[0102] The spectra can be analysed using a hybrid technique that couples One -Point Calibration (OPC) procedures with calibration curves. This allows for the observation of the relative quantitative composition changes compared to the solid feedstock alloy. For a given material,this is determined using the sampling of the solid-state feedstock inside detection chamber 8, using the LIBS subsystem 200. Using a One-Point Calibration (OPC) approach, the plasma density from the LIBS sampling of the solid feedstock alloy is calculated, and used to correct the sampled mass in a single -point calibration curve.

[0103] To account for an unknown wavelength-dependent experimental efficiency in the spectra, the OPC method calculates correction coefficients, Pj, using the LIBS spectra of a known sample - in this example, the solid NiCoCr SRM and solid IN625 alloy. This is achieved in the following steps:1

[0104] 1. Generate a Saha-Boltzmann plot for the raw data 701 (NiCoCr) and 703 (IN625), by plotting for each peak (generated from transition zj):

[0105]

[0106] Where I is peak intensity, Atj is transition probability, gj is upper level degeneracy, Ej is upper level energy, kBis the Boltzmann constant, T is plasma temperature, N is plasma density, US(T) is the partition function. Figure 7 shows Saha-Boltzmann plot results for solid samples and vapour samples, over NiCoCr alloy and IN625 alloy. We determine the gradient of the most linear elemental species (typically Cr I or Fe I), to estimate plasma temperature T from the raw data.

[0107] 3. Use plasma temperature T to find US(T), and the number of electrons, Ne, as:

[0108] > 1 x 1016(FWHM)

[0109]

[0110] In this example, we selected FWHM of the Ni I 341.476 nm. which has an estimated electron impact factor, m ~ 0.0029.2

[0111] Calculate the total concentration, Ccert, of an element in the standard as:

[0112]

[0113] For the number of singly-ionised species C / and number of doubly-ionised species C" . Note, C / and Csare determined using the Saha equation:

[0114] "

[0115]

[0116] Where Eionis the ionisation energy of the element in eV.

[0117] 5. We identify the y-intercepts, qs, of the SB plot 701 (NiCoCr) and 703 (IN625) to get the un-corrected concentration of each species, C - which will stray from the known composition due to the unknown wavelength-dependent experimental parameters. 6. We calculate the correction values, Pj for such offsets by:

[0118]

[0119]

[0120] 7. We then re-plot the Saha-Boltzmann graph with In (I / Pt) on the y-axis 702 (NiCoCr) and 704 (IN625), to give slope and intercepts of the corrected values.

[0121] The results of the OPC calibration for IN625 and NiCoCr are shown in Figure 8. The corrected plasma density can be extracted from the y-intercepts of 702 and 704 for each species.

[0122] Using the plasma density, the sampled mass of each element, mi for element i, in the alloy is calculated as:

[0123]

[0124] Where mi is the mass of element i in femtograms,

[0125]

[0126] are the number of species of the I and II state of element i derived from the Saha-Boltzmann y-intercepts (in mm-3), MR iis the molar mass of element i in g mole1and NAis Avogadro’s number.

[0127] The density-corrected single-point calibration plots (DCSP plot), shown in Figure 8, are constructed as a single-point linear regression of mi against the peak integrals found in the solid-state LIBS spectra. To account for differences between solid 901 and vapour 901 LIBS responses, shown in Figure 9, the DCSP plot uses a sum of all non -overlapping integrals ([peaks) f°reach element. In this example, Ipeas for IN625 were selected as: {Ni I: 229.0, 231.9, 234.7, 352.9}, {Fe II: 249.1}, {Cr II: 288.0}, {Mo I: 390.7}. This technique allows for quantitative analysis of metal alloy vapours, which would otherwise be very difficult to achieve using purely OPC or purely calibration curve approaches.

[0128] In this example, this methodology was used to create DCSP plots for the main constituents of NiCoCr (Ni 801, Cr 802, Co 803), and the main constituents of IN625 (Ni 804, Cr 805, Fe 806, Mo 807). The DCSP plots were used to generate the quantitative results for the NiCoCr and IN625 LPBF-LIBS experiments.

[0129] Using the intercept and slope of the DCSDP plot.

[0130] NiCoCr Ni Cr Co

[0131] Slope 3.40 % 10“32.03 x 10“21.40 % 10“3

[0132] intercept 0 0 0

[0133] IN625 Ni Cr Fe Mo Slope 5.83 % 10“55.56 % 10“55.49 % 10“52.85 % 10“5Intercept 0 0 0 0 Table 2: The fitting coefficients of the DCSP plots created using the hybrid calibration technique .

[0134] To extract the quantitative results from the deconvoluted spectra for vapour samples of unknown composition, the integrals of all peaks that match the Ipeaks list f°reach element are summed, and calculated the masses are calculated using the coefficients in Table 3, by:

[0135] Upeaks ~ intercept)

[0136] mass = - - - slope

[0137] The above method may be repeated under different materials to create calibration plots for multi-component alloy systems applicable to the AM tool 100 and the LIBS subsystem 200. The method is as follows:

[0138] 1. Identify a known composition - either an SRM material or feedstock alloy with a specified composition.

[0139] 2. Carry out a solid-state LIBS experiment on the known composition alloy.

[0140] 3. Using the NIST LIBS database, identify regions of non -overlapping peaks for the elements of interest. Note the database peak locations and the discrete transitions that form those peaks.

[0141] 4. Perform the OPC corrected Saha-Boltzmann plot analysis for the solid-LIBS.

[0142] 5. Create the calibration plot of each element by finding mband apply the coefficients to quantitatively analysis vapour datasets.

[0143] This methodology is suitable for any pure metal or metal alloy; NiCoCr and IN625 are used as examples due to their high-Z alloying elements, which provide a good visibility under Electron Dispersive Spectroscopy (EDS) to confirm the findings.

[0144] Ex situ chemical analysis for confirmation of calibration approach efficacy

[0145] The chemical composition of IN625 powder, substrates, and printed tracks were measured using EDS. The weight percentages of the major alloying constituents in the feedstock substrate and powder materials are provided in Table 5. The EDS maps were constructed with a secondary electron (SEI) beam at an accelerated voltage of 20 keV and a 54 — 60 pm spot size over a 1-hour live acquisition time. All EDS measurements were re-scaled using calibration curves constructed using EDS measurements of the NiCoCr NIST SRM materials.Sample Ni [Wt %] Cr [Wt %] Co [Wt %]

[0146] Substrate ( .2- 0.14 2L87 0.09 13.89 O. / O

[0147] Weld 64.01 ± 0.14 22.05 ± 0.09 13.94 IN625 Ni [Wt %] Cr [Wt %] Fe [W

[0148]

[0149] Powder 66.65 ± 0.78 22.35 ± 0.45 2.88 ± 0.24 8.12 ± 0.63 Substrate 65.89 ± 0.48 21.31 ± 0.27 .3 ± 0.18 8.49 ± 0.39 Avg. Feedstock 66.27 ± 0.63 21.83 ± 0.50 3.60 ± 0.21 8.31 ± 0.51 180W LPBF 64.96 ± 0.49 21.75 ± 0.29 3.23 ± 0.16 10.07 ± 0.41 400W LPBF 65.96 ± 0.20 21.60 ± 0.72 3.26 ± 0.04 9A ± 0.20 Table 3: EDS measurements for NiCoCr pre -weld and post-weld, and EDS measurements for the LPBF feedstock materials, and printed areas for 180W and 400W printing. Avg. Feedstock refers to the average of the powder and substrate composition. Compositions have been normalised to include only the mam constituents.

[0150] LPBF-LIBS Results

[0151] Figure 10 provides the change in bulk composition of the welded area compared to the original composition 1002, as measured by EDS, and the quantitative LIBS results of the welding vapour 1001. All EDS data is listed in Table 4. During LPBF-LIBS, the analysis of the data produced by the LIBS subsystem, after processing, shows the vapour composition to be dominated by Ni. Similarly, the EDS shows a drop in the weight percentage (wt %) of Ni within the welded area of the alloy compared to the feedstock.

[0152] A large Ni presence in the vapour would be expected, as the NiCoCr alloys consists of 58.782 % Ni. As such, we define the Loss Ratio for the LIBS results, as:

[0153]

[0154] Where Mvapouris the mass of element i in the vapour, Mfeedstockis the mass of element i in the feedstock, and the Loss Ratio is found by normalising MLover all elements of the alloy. Therefore, the magnitude of loss of elements during NiCoCr welding is Ni > Co > Cr. This trend is reflected in the EDS results, which show Ni to have the largest loss in wt %, and Co to have the largest gain in wt %.

[0155] Figure 11 demonstrates the LIBS results of the LPBF vapour for large 3 -layer builds 1101. A shown in Figure 1101, the results demonstrate that Ni, Cr and Fe were readily detected in both the low-power and high-power LPBF vapour, alongside the presence of trace levels of Mo. An increased mass of vapour is observed under higher-power processing conditions,which is a well-reported feature of LPBF.3The Loss Ratio in both cases is Ni ~ Fe > Cr >> Mo; and a similar trend is reported in the EDS data 1102. The large errors due to small composition changes in EDS measurements show that measuring the bulk is not a suitable way to infer vaporisation losses; but, the LIBS and EDS results agree with previous work that report compositional gains in Mo after additive manufacturing processing of Invar 36 and SS316L.4

[0156] From reading the present disclosure, other variations and modifications will be apparent to the skilled person. Such variations and modifications may involve equivalent and other features which are already known in the art, and which may be used instead of, or in addition to, features already described herein.

[0157] Although it is described that the AM tool 100 of the present disclosure is a laser powder bed fusion (LPBF) tool, in alternative arrangements, the AM tool 100 may be any other type of AM tool, such as a type of directed energy deposition tool, and is not limited only to LPBF.

[0158] Although it is described in relation to Figure 1 that the laser 1 of the AM tool 100 produces a beam source to melt or fuse the powder feedstock, in other arrangements, the laser 1 may be an electron beam laser configured to generate an electron beam directed at the build region 2, to melt or fuse the powder feedstock together.

[0159] Although it is shown in Figures 1 and 2, that the AM tool 100 includes an exhaust 3 with two parts, first part 17 and second part 18, and the that the LIBS subsystem 200 is positioned within the exhaust 3, in alternative arrangements the LIBS subsystem 200 may be positioned at the end of the exhaust 3.

[0160] Although the appended claims are directed to particular combinations of features, it should be understood that the scope of the present disclosure also includes any novel feature or any novel combination of features disclosed herein either explicitly or implicitly or any generalisations thereof, whether or not it relates to the same subject matter as presently claimed in any claim and whether or not it mitigates any or all of the same technical problems as does the present disclosure.

[0161] Features which are described in the context of separate embodiments may also be provided in combination in a single embodiment. Conversely, various features which are, for brevity, described in the context of a single embodiment, may also be provided separately or in any suitable sub-combination. The applicant hereby gives notice that new claims may beformulated to such features and / or combinations of such features during the prosecution of the present application or of any further applications derived therefrom.

[0162] For the sake of completeness, it is also stated that the term “comprising” does not exclude other elements or steps, the term “a” or “an” does not exclude a plurality, a single processor or other unit may fulfil the functions of several means recited in the claims and reference signs in the claims shall not be construed as limiting the scope of the claims.

[0163] References:

[0164] 1. Hu, Z., Nie, J., Ouyang, Z., Zhang, D., Liu, Y., Chu, Y., & Guo, L., ‘Self-absorption correction method for one -point calibration laser-induced breakdown spectroscopy’, Opt. Lett. 48, 1-4 (2023).

[0165] 2. Konjevic, N., Lesage, A., Fuhr, J.R., & Wiese, W.L., ‘Experimental Stark Widths and Shifts for Spectral Lines of Neutral and Ionised Atoms’, J. Phys. Chem. Ref. Data, 31, 3 (2002).

[0166] 3. Liu, J. &Wen, P. ‘Metal vaporization and its influence during laser powder bed fusion process’, Materials and Design vol. 215 (2022).

[0167] 4. Yakout, M., Elbestawi, M. A. & Veldhuis, S. C. ‘A study of thermal expansion coefficients and microstructure during selective laser melting of Invar 36 and stainless steel 316L’, Addit Manuf. 24, (2018).

Claims

CLAIMS1. An additive manufacturing tool, comprising:a beam source for creating a beam that is operable to produce a solid part by melting or fusing a feedstock in a build region that is heated by the beam; anda laser induced breakdown spectroscopy, LIBS, subsystem, the LIBS subsystem comprising:a LIBS laser configured to generate a light beam to illuminate a vapour originating from the build region, and produce a plasma in a detection region;a detector configured to detect light in the detection region emitted by the plasma; anda processor configured to receive spectroscopic data from the detector, wherein the spectroscopic data is indicative of a spectrum of the light emitted by the plasma in the detection region,wherein the detection region is spaced apart from the build region.

2. The additive manufacturing tool according to claim 1, further comprising:a build chamber housing the build region, wherein the detection region is outside of the build chamber.

3. The additive manufacturing tool according to any preceding claim, further comprising:an exhaust configured to direct the vapour from the build region to an outlet, wherein the detection region is positioned in the exhaust.

4. The additive manufacturing tool according to claim 3, further comprising:a build chamber housing the build region, wherein the exhaust is a manifold coupled to an opening of the build chamber and configured to direct the vapour out from the build chamber through the opening.

5. The additive manufacturing tool according to any preceding claim, wherein the detection region is at least 2cm from the build region.

6. The additive manufacturing tool according to any preceding claim, wherein the LIBS subsystem further comprises a lens and an optical fibre, wherein the lens is configured to collect the light emitted by the plasma, and the optical fibre provides at least a portion of a detection optical path to the detector.

7. The additive manufacturing tool according to claim 5, wherein the LIBS subsystem further comprises:a mirror configured to reflect the collected light from the lens towards the optical fibre.

8. The additive manufacturing tool according to any preceding claim, further comprising:a focusing lens configured to focus the light beam from the LIBS laser into the detection region.

9. The additive manufacturing tool according to any preceding claim, wherein the processor is configured to determine a composition of the vapour by analysis of a spectrum of light obtained from the spectroscopic data.

10. The additive manufacturing tool according to claim 9, wherein the processor is further configured to:i) deconvolve the spectrum into one or more peaks; andii) analyse the constituents of the vapour with reference to the peaks .

11. The additive manufacturing tool according to claim 10, wherein step i) comprises:fitting the one or more peaks to the spectrum.

12. The additive manufacturing tool according to claim 11, wherein step i) comprises:dividing the spectrum into a plurality of spectral windows, each spectral window spanning a different portion of the spectrum; andfitting one or more peaks to each spectral window.

13. The additive manufacturing tool according to any preceding claim, further comprising:a build chamber housing the build region;a detection chamber housing the detection region; anda first exhaust part comprising a first end and a second end, wherein the first end of the first exhaust part is coupled to the build chamber and the second end of the first exhaust part is coupled to the detection chamber, and wherein the first exhaust part is configured to direct vapour from the build chamber to the detection chamber.

14. The additive manufacturing tool according to any preceding claim, further comprising:a camera directed towards the detection region and configured to capture an image of the plasma.

15. The additive manufacturing tool according to any preceding claim, wherein the additive manufacturing tool is a powder bed fusion tool or a direct metal deposition tool.

Images