Exploring the benefits of real-time particle sizing for powder metallurgy and additive manufacturing

Introduction

The growing use of Powder Metallurgy (PM) and Additive Manufacturing (AM) processes is directly attributable to the opportunities they offer to reduce costs and/or enable the manufacture of components with properties that cannot be accessed via conventional subtractive techniques. However, successful manufacture is directly dependent on the quality of the metal powders used, with particle size and shape critical to performance. Learning how to produce high-quality powders as cost-efficiently as possible, and manage them during processing, is crucial to further improve the economics of PM and AM processes and extend their application. 

Defining particle size and shape requirements

The optimal particle size of metal powders for Powder Metallurgy (PM) and Additive Manufacturing (AM) depends largely on the technology being deployed (see Figure 1). Particle size influences both the properties of the finished component and process performance by directly impacting:

• Packing behavior 
• Flowability  
• Response to heat
• Surface finish

Figure 1: Typical particle size ranges for a variety of advanced PM and AM manufacturing technologies

Generally speaking, larger particles pack less efficiently than smaller particles of equivalent shape, while melting/sintering occurs more easily/rapidly for smaller particles. However, a powder with a range of particle sizes is often preferred, as smaller particles will fill the voids left by adjacent larger particles, densifying packing and potentially improving heat transfer. Larger particles are advantageous for flowability because the forces of attraction between particles increase with decreasing particle size. So, there is a compromise depending on the component manufacturing process being employed. 

Figure 2: A schematic showing key stages of a MIM process

Contrasting the requirements of powders for Metal Injection Molding (MIM) and AM illustrates the relative importance of these different properties, and how this influences the specification of metal powders for alternative processes. Figure 2 shows a typical MIM process which includes: mixing/blending of the metal powder with a binder, followed by subsequent granulation; injection molding to produce a green part; de-binding; and a final sintering step. 

In this process, the higher packing densities associated with finer particles enable high particle loading in the feedstock, thereby minimizing binder usage and, as a direct result, part shrinkage during sintering. The melting characteristics of the powder are less critical than additive manufacturing processes such as selective laser melting (SLM) as sintering involves heating to just below the melting point of the metal, often with the simultaneous application of pressure. But smaller particle size enhances the sintering process and is preferable. For these reasons powders for MIM are some of the finest for all PM processes, typically lying in the sub-38 µm range (see Figure 1).

Figure 3: Schematic of a Powder Bed Fusion AM Process with inset image showing how targeted melting fuses one powder layer to the next to progressively construct the component.

Figure 3 shows a Powder Bed Fusion AM process, an example of a widely used AM technology. Powder is spread by a roller across a build platform in a layer just tens of microns thick. Selective laser fusion joins successive layers to progressively form the component as the build platform is gradually retracted. Here, flowability of the as supplied powder is crucial, since achieving a smooth even layer with no air voids, rapidly, is essential for cost-effective processing. Set against this is the need for relatively fine particles to achieve the necessary layer thickness, and rapid melting at an acceptable power input. High density packing is required to produce components of consistent quality with few flaws. Balancing all these demands leads to a closely defined particle size specification in the region of 15 – 45 µm for selective laser melting processes, and 45 – 106 µm for electron beam melting (see Figure 1). For powder bed AM processes such as Binder Jetting which involve binding parts of the powder bed with adhesive and post-sintering, particle size requirements are more akin to MIM. Referring to the well know proverb ‘Horses for Courses’ we could also say ‘Powders for Processes’ when referring to AM and PM. 

Making metal powders

The production of metal powders predates many PM processes, which have increased demand and intensified pressure on specifications. Metal powder manufacturers meeting the needs of the PM industry are working to much closer size tolerances than previously, driving a need for rigorous process optimization. The primary processes used to manufacture metal powders are attrition milling and gas/water atomization. Gas atomization is the most widely employed process for producing powders for AM and MIM because it gives spherical particles in the appropriate size range. But other novel processes are being developed and employed as companies look to optimize their products and processes.

Key optimization goals are to:

• Control morphology – particle size and particle shape are critical for the end-application so there is a need to get this right on a consistent basis.
• Maximize yield – where yield is defined as the output rate of in-specification product.
• Minimize production costs – all size reduction processes are energy intensive; cutting energy consumption and waste are the key goals.

In gas atomization, molten metal is ejected through a nozzle of defined dimensions into a high-pressure gas stream (typically argon or nitrogen) to create a fine spray of metal droplets that is cooled rapidly to form particles. The size of these particles is influenced by the properties of the molten metal and nozzle geometry. But for routine process control, it is usually feed rate and/or  pressure driving the metal through the nozzle that are controlled. Post-atomization processes such as scalping – the removal of oversized particles – sieving and/or air classification enable for tailoring of the as-atomized product to meet the defined particle size specification. This is the first optimization objective.

Post-atomization processes are effectively ‘yield killers’ as they slow down overall production rates by adding in steps, processing time and cost. To maximize yield there is a need to reduce such processes to an absolute minimum through optimum process control. So, the objective here is to make as much product as possible with the right particle size and shape specifications. This is also essential to reduce waste and energy consumption and achieve the third optimization objective which is to minimize production costs, energy and waste. This raises the question of how to gain the process understanding needed to determine the optimal way of making any given product, and to exert effective control of the process.

Taking control

Traditionally, processes such as milling and atomization are controlled with reference to offline, laboratory particle size measurements. Even with an analytical technique that offers fast measurement times, this approach is associated with time delays - between production and sampling, sampling and measurement, and measurement and receipt of results – that compromise the efficiency of process control. The more modern and efficient approach is to install in-line particle sizing technology so that the output from the process is continuously monitored in real-time. A well-established technique that offers highly reliable continuous measurement with minimal manual intervention is in-line laser diffraction, such as Malvern Panalytical’s Insitec (Figure 4).

Figure 4: Insitec is an in-line laser diffraction system that can measure particle size distribution in real-time for wet and dry dispersions, and even sprays. 

As a result, such technology makes it easier to:

• Minimize production costs: when the impact of a process change is clear, rather than obscured by the time delays associated with an offline regime, the correlations between process variables and product parameters become sharper. This makes it easier to identify the most economic processing conditions for any given product.
• Optimize control: an instantly measurable response makes it easier to take effective manual control action, but more valuably, continuous monitoring provides a data stream for the implementation of fully automated control.
• Switch products: when product quality is being continually monitored, it is easy to determine when a new product specification has been reached. Real-time monitoring therefore minimizes both the time and waste associated with product changeovers, a routine task in metal powder production facilities.

The following studies illustrate the application of real-time particle sizing and the benefits that can be obtained.

Example 1: Mill operation with automated particle size control.

Figure 5 shows a control trend for a micronization process operating under automated particle size control driven by data from an in-line laser diffraction system (Insitec, Malvern Panalytical). A simple PID (Proportional Integral Derivative) loop adjusts the feed rate to the mill to maintain the Dv50 of the product at the specified set point, where Dv50 is the particle size below which 50% of the particle population lies on the basis of volume. Set point is changed a number of times during the monitored period of operation, in the range 3 to 6 µm, and in each case the process moves swiftly to new operating conditions.

Figure 5: With automated process control in place, the particle size specification is continuously and reliably met and the time taken to switch from one set point to another is minimized.

Examination of the trend highlights the following points:

• Particle size is being controlled to within just a few tenths of a micron. The automated loop results in frequent, small process changes to keep particle size within a narrow range. Manual control, in contrast, typically involves larger, less frequent changes, giving rise to a more variable product.
• The time taken to move from one size specification to another is minimal, in the order of tens of seconds. This minimizes waste and at the same time maximizes process throughput.  

In this set-up both routine analysis and control are fully automated, enabling continuous unattended operation, with minimal manual input required. Furthermore, the control loop maintains the specification regardless of variability in, for example, raw material parameters (hardness, moisture, size) or mill condition (wear over time) ensuring the very highest levels of product consistency.

Example 2: Monitoring a sieving process with laser diffraction

Figure 6 shows an automated metal powder recycling system for AM called PowderCleanse. This was designed to facilitate the re-use of excess and reclaimed metal powder in the AM process. The system combines closed-loop powder processing with in-line particle size analysis. Here excess powder from the manufacturing process is collected in a bulk hopper and pneumatically transported to an automated sieving system.  Oversize material is collected and weighed in the ‘oversize’ collection vessel while sieved powder is passed through the Insitec particle sizing instrument sitting below the sieve deck.  The sieved powder is then conveyed to a second collection hopper ready for re-use. The whole process is controlled using Malvern Panalytical’s Malvern Link ii software. 

Figure 6: The PowderCleanse system consists of an automated sieving unit with an Insitec particle sizing instrument sitting below the sieve deck for real-time analysis during the sieving process.

This closed-loop system is designed to allow for near 100% assessment of the batch rather than off-line sampling, minimising the risk of contamination and enabling full traceability to be maintained. The PowderCleanse system shown in Figure 5 utilizes Carpenter Additive hoppers with integrated sensors, and a Farleygreene Sievgen 04 for the sieving operation, although Insitec can be integrated into a range of sieving operations and with a range of process equipment. 

Figure 7: Assessment of in-line and off-line particle analysis shows that Insitec result are comparable with Mastersizer 3000 despite different sample volumes and optical configurations.

Figure 7 shows a comparison of particle size data with the in-line Insitec and off-line Mastersizer 3000. Despite differences in the powder dispersion method and optical geometry of the two systems the results are in good agreement, confirming that the Insitec is comparable to off-line testing, but has the advantage of testing an entire batch with minimal manual intervention and in real time. 

Example 3: Using laser diffraction to monitor and control atomization processes

Figure 8: Schematic of a gas atomization process for the production of metal powder and where Insitec can be used.

Figure 8 shows a schematic of a typical gas atomization process used for the production of metal powders. Metal droplets solidified in the atomizer tower are transferred to the product cyclone for further cooling and separation of fines, from the required product which is cooled further ahead of storage. In processes such as these, in-line laser diffraction analyzers can be usefully installed at points highlighted in green to continuously monitor the quality of the atomized powder by providing a real-time measurement of particle size. Figure 9 illustrates how the data from such analyzers can be used for process optimization and control.

Figure 9: Real-time measurement system makes it possible to instantly observe the effect of decreasing the speed of rotation of the atomizing disk and the time taken for the process to re-establish a steady state. (Green Dv10, blue Dv50, orange Dv90, red is transmission).

In this study the performance of a rotary atomizer was investigated. With this technology particle size is controlled through manipulation of the speed of rotation of the disk, so a key requirement is to quantify the correlation between disk rotation speed and the size of particles produced for a given feed material. Here, materials produced at different disk speeds were analysed using an Insitec on-line laser diffraction particle sizing system. As is clear from the x-axis, the trial was completed over a relatively short timescale, simply by stepping through the operating conditions of interest to determine those required to produce the target Dv50, which in this case was 55 – 65 µm. The level of fines in the product was also a concern so the ability to simultaneously track the levels of particles down to around 0.1 µm was an additional, important benefit of the laser diffraction technology.

The data show the increase in particle size associated with reducing disk speed. They also indicate that after each change the process takes a certain amount of time to reach a new steady state, highlighting the need to run for around 30 minutes to generate robust particle size information and/or ensure secure collection of a new product. The ‘sweet spot’ identified for this process occurs at a speed of rotation of 8,000 rpm, which produces particles with the required Dv50 and an acceptable level of fines as quantified by Dv10 data (see Table 1). Real-time measurement accelerates the identification of this optimal operating condition and at the same time, going forward, provides the data required for secure, consistent and potentially automated process control.

Table 1: 8,000rpm was found to be the optimum speed to achieve the target Dv50, and an acceptable level of fines.

Looking forward

Manufacturing optimal metal powders for PM and AM processes economically demands exemplary process control. Established in- and online particle sizing technologies provide continuous process monitoring and a data stream that supports complete process automation. Such technology therefore makes it easier to make highly consistent products, to a tight particle size specification at the lowest possible variable cost. At the same time, continuous particle sizing has a role to play in the management of high value powders, in particular for monitoring the reuse of AM powders, a critical issue for economic viability. With a proven track record and strong reliability, in and on-line particle sizing technology holds considerable value for metal powder manufacturers establishing robust supply chains for the PM and AM industry, and for AM processors optimizing powder management strategies.