Everything Material in Material Science

Science. Tech. Business.

Join the Conversation

Advanced Powder Characterization For Additive Manufacturing

February 27, 2019 5:56 pm

Among the various additive manufacturing (AM) processes either in use or under development, e.g., metal powder bed fusion (PBF) and ceramic binder jetting, several employ powder feedstocks. As a result of manufacturing artifacts and inter-particle forces, powder materials exhibit properties that depart from those of bulk materials in other forms. Manufacturing conditions can affect all aspects of individual particle properties, which in turn affect attractive particle interactions such as adhesion and non-bonded van der Waals forces[1,2]. Differences in these properties manifest themselves in the dynamics of powder flow and subsequent particle fusion, which must be optimal to achieve fully densified parts.

In some cases, conventional powder characterization techniques provide useful information when adapting powders to AM processes. Yet there are many cases where the techniques fail to adequately distinguish enough information to enable process optimization[2,3]. The most common example is the use of static flow characterization methods (such as Hall or Carney funnel flowrate tests) to quantify powder flowability or spreadability. Additive manufacturers have found that these methods often provide little or no useful information to explain observed differences in flow behavior between virgin and recycled powders or powders with varying levels of moisture content, for example.

A recent study from the Massachusetts Institute of Technology[2] presented a dramatic difference in the surface energy of three Ti-6Al-4V PBF powders (0.1 mJ/m2 ) compared to bulk metals (40-50 mJ/m2 ). Results were attributed to parameters such as surface roughness, surface chemistry, and surface oxidation, and the study explained why conventional metrics such as angle of repose fail to properly represent bulk behavior in metal AM powders. Therefore, a suite of advanced characterization techniques is now being deployed across the full range of the AM powder supply chain. In most cases, these are well-established methods that have found common use in fields such as catalysis, mining, pharmaceuticals, and glass.

A joint effort between ASTM and ISO is working toward standardization of powder characterization techniques for AM, borrowing from existing standards in other fields as well as developing new standards unique to AM. The Additive Manufacturing Consortium operated by EWI is concurrently providing data to support standardization through a varied portfolio of R&D projects. Nonetheless, several notable gaps in AM-specific powder characterization still exist: sample conditioning, particle morphology, moisture analysis, and powder flow characterization. In this article, a few of the myriad characterization options available to particle scientists will be discussed.

It is useful to describe individual powder properties with respect to the scale at which they are manifested—the core material level, the formed particle level, or the bulk powder level. As an example, a few of the most common powder characterization techniques are categorized using a matrix of material scale and four property categories (Table 1).


Note the inclusion of sampling and conditioning methods at the top of the matrix in Table 1. Although not characterization techniques per se, it is critical that these are not overlooked or discounted when dealing with powder samples. Proper powder sampling techniques must be used in order to produce representative samples for subsequent laboratory analysis. The unique nature of powders can also lead to self-segregation during conveyance, storage, packing, and handling[4]. In one example, a 2014 NIST study revealed significant variability in chemical composition among four sub-samples taken from a single lot of a cobalt-chromium AM alloy, as well as two sieve fractions within a recycled stainless steel powder[5]

Standard practices for representative powder sampling at all scales are well established[6,7]. At the laboratory level (typically <1 kg sample), chute splitters and rotary rifflers are necessary to produce smaller sub-samples for testing. Similarly, due to high surface area-to-volume ratios, powder materials often tend to absorb moisture more readily than in the bulk state[3]. Therefore, in addition to testing as-received materials, it is often important to condition powders (via humidification or drying) prior to analysis


The fundamental properties of materials are independent of form, at least hypothetically, so these will only be mentioned briefly. The most common chemical analysis includes multielemental inductively-coupled plasma (ICP) spectroscopy, inert gas fusion, and combustion methods for C, S, O, N, and H elemental analysis, as well as conventional wet chemistry techniques for certain species. There is also a general trend toward faster, direct powder analysis via x-ray fluorescence (XRF) spectroscopy, whenever possible.


Particle-level characterization includes chemical analyses, size distribution measurements, and morphological analysis. In cases where particle chemistry differs from bulk material chemistry, SEM-EDS analysis of whole or cross-sectioned particles can provide detailed elemental distributions on the surface and interior of particles. Variable pressure FE-SEM instruments are best suited to EDS analysis, as conductive coatings are not necessary. To avoid the small population sampling inherent in microscopic analyses, XPS can be used to analyze larger powder samples, although it is limited to the particle surface.

Despite a vast difference in measurement basis, sieve analysis and laser scattering techniques remain the most common methods for characterizing particle size distributions in bulk powders. Whichever technique is used, it is important to understand the basis for how the particle sizes are derived. As shown in Table 2, the various size analysis techniques fall into one of the three categories listed here. Depending on how the size data is to be used, one type of method may provide more meaningful results than another. However, in most cases the most useful results are obtained from a combination of methods, often including a microscopic technique coupled with an ensemble or separation method.

Obtaining particle morphology information that is both representative and useful has long been a challenge in powder technology. Table 3 shows one way to categorize particle shape and texture parameters along with typical techniques employed to derive them.


Scale Composition Density and porosity Size and morphology Flow behavior
Sampling and conditioning Drying/humidification, splitting, riffling
Material properties s Spectroscopy, inert gas fusion, gravimetric SEM, He pycnometry
Particle properties SEM-EDS N2 porosimetry Kr/N2 adsorption, sieve, laser, microscopy
Bulk properties Gravimetry, TGA/DSC, halogen IR He pycnometry, loose/apparent, tapping/packing Dynamic rheometry, flow meters


Method type Ensemble methods Counting methods Separation methods
Particles are
analyzed simultaneously.
Particles are
analyzed individually.
These rely on the
segregation of particles.
Primary methods Laser diffraction,
dynamic light scattering
Optical imaging, microscopy,
Sieve analysis,
X-ray sedimentation
Specialty methods Air permeability,
ultrasonic attenuation
Light obscuration,
time-of-flight counting
Disc centrifuge,
capillary fractionation,
flow field fractionation,
scanning mobility analysis

One drawback of many of these approaches is the necessity to simplify complex or irregular shapes into equivalent ideal geometries, analogous to how laser scattering particle size data is fit to equivalent spherical diameters. Another problem with some of these techniques is the limitation of analyzing particles with 2D imaging, where a particle outline or shadow is used to represent an entire particle. A third challenge in microscopy techniques is the limited number of particles that can be analyzed in each image, often on the order of 100 particles or fewer. Nonetheless, dynamic image analysis instrumentation is evolving rapidly and could soon become a benchmark technique.

In contrast to these approaches, the use of well-established gas adsorption techniques provides a way to probe bulk quantities of a powder—with particle quantities ranging from thousands to hundreds of thousands. The BET (Brunauer-Emmett-Teller) method uses condensation of krypton or nitrogen gas on a powder surface at cryogenic conditions, measuring the resulting gas pressure change to quantify the exposed surface area of powder. Well-established in powder metallurgy[8], the method is also used extensively in ceramics, pharmaceuticals, and geology to characterize particle morphology[9-12]. The specific surface area is reflective of the particle size, shape, texture, and porosity. It can therefore also serve as a way to quantify defects in a large ensemble of particles.

A rigorous analysis of the relative contributions of particle size and morphology to BET surface area was conducted by Corning using XRD crystallography and electron microscopy to demonstrate the strong correlation between these properties in talc powders[13]. For routine analyses, gas adsorption is therefore an excellent way to conduct true 3D, bulk morphological analysis much more efficiently and precisely than other methods. When investigating specific morphology issues, it is useful to couple this with microscopic analyses.


One of the traits of the energetics of fine powder surfaces described above is that these forms of materials tend to absorb ambient moisture much more readily than they do in other forms[3]. Both metals and ceramics exhibit this behavior, impacting powder flow and final part chemistry and density[14-16]. However, depending on the material, the absolute amount of moisture can still be very low and is often less than 0.5 wt% for metal powders. Therefore, conventional oven-based


Particle scale Surface/textural scale
Macroshape Mesoshape
Descriptors Sphericity,
elongation, etc.
angularity, etc.
Roughness Surface area
Nature 3D 2D 2D 3D
Basis Equivalent dimensions
or volume
Equivalent area or
Geometric analysis
(ratios, etc.)
Techniques Dynamic or static digital imaging,
optical or electron microscopy,
tomography, laser scattering
Gas adsorption
Microscopy (static imaging): ≈ Hundreds of particles
Laser or dynamic Imaging: Tens of thousands of particles
Tens of thousands
of particles



Non-specific Water-specific
Technique Limitations Technique Limitations
Chemical Karl Fischer
Calcium carbide
Spectroscopic TPD/TDS Precision TPD-MS/TDMS
Near IR
Sample prep
Thermogravimetric Oven/vacuum
Halogen IR
2% H2O
RH sensor
Sample size, method

(loss-on-drying) techniques are usually too imprecise. While the Karl Fischer titration method is water-specific, it is comparatively complex and laborious. The very small sample sizes used in TGA methods often lead to poor precision. Halogen IR methods (already standardized for determining moisture in plastics) are perhaps the most promising technique, especially because they can accommodate fairly large sample sizes (up to 200 g) and are simple to operate. Nonetheless, the low moisture levels in metal powders require the most advanced (and expensive) versions of these instruments, with some users still reporting insufficient sensitivity. See Table 4.

Bulk powders are often characterized according to one or more types of density. Each type may be useful in analyzing powder impacts on the various stages of material densification that occur in powder AM processes, from powder spreading to final sintering. Along with apparent (or bulk) density (both loose and tapped), skeletal density continues to be an important powder property for optimizing final AM part density. This parameter is measured through the well-established technique of gas displacement pycnometry, most often using a helium pycnometer with microcell chambers to improve precision in smaller sample volumes. An attractive aspect of this technique is that the same instrument can be used to measure the true (or absolute) density of a sample from a final part.

Perhaps the most active area of development in powder characterization for AM is the field of powder flow analysis. Table 5 summarizes the various static, quasi-static, and dynamic techniques commonly in use, along with the author’s opinions on their application to AM powders.

Dynamic flow analysis techniques are of particular relevance to powder AM applications, where recoater spreading is an example of a dynamic environment under a low stress state. In many cases, when conventional static funnel testing cannot distinguish between different powders (or when the powders will not flow through the funnels), avalanche/drum type rheometers will consistently differentiate between samples. Further, the instruments allow for parametric evaluation to identify “sweet spots” in powder flow behavior, correlating with recoater roller transversal speed or rotation speed/direction, for example.

The ASTM powder metallurgy subcommittee recently sponsored a round robin test to evaluate the ability of three commercial instruments (one quasistatic and two dynamic) to distinguish between six different AM powder samples. Dynamic avalanche-type instruments were best able to differentiate samples based on material type and history (virgin, recycled, or blended). Data presented in the ASTM study, as well as the author’s own experience using an avalanche rheometer, shows that most parameters can achieve good reproducibility.


As powder-based AM processes continue to mature, various combinations of the powder characterization techniques discussed here will be key to optimizing AM processes and final part performance. At the same time, some of these methods can provide rapid feedback for manufacturing control, powder qualification, and yield optimization. It is important for those involved in the powder supply chain to appreciate the unique nature of powders, which results in their distinct behaviors—making proper powder sampling crucial. Well-established advanced characterization techniques such as krypton BET adsorption, helium pycnometry, and powder XRF can provide more meaningful data to solve manufacturing challenges. In the near future, new and improved techniques like halogen IR moisture analysis, dynamic image analysis, and dynamic rheometry promise to provide much improved data streams. ~AM&P


Powder Flow Analysis Techniques
Static Quasi-static Dynamic
Methods Hall, Carney, etc. Shear cell, impeller Avalanche/drum
Qualitative Yes/No Flow regimes
Semi-quantitative Flowability
Flow rate
Angle of repose
Stress curves
Flowability energy
Flow function
Wall friction angle
Surface fractal
Avalanche energy
Cohesive index
Avalanche angle
Rest angle
Quantitative Carr index
Hausner ratio
Volume changes Volume change
Volume memory
Apparent density Conditioned density
Consolidation index
Dynamic density
Vibrated density
Stress state Low Moderate to high Low
Ease of operation Simple Complex Simple
Data interpretation Simple Complex Values
Parametric analysis No Yes Yes
Standardization ASTM, ISO ASTM (plastics) No

For more information: Dave van der Wiel, analytical operations manager, NSL Analytical Services Inc., 4450 Cranwood Pkwy., Cleveland, OH 44128, 216.438.5215, dvanderwiel@nslanalytical.com, www.nslanalytical.com.


  1. A. Castellanos, Adv. Phys., Vol 54,
    p 263-376, 2005.
  2. C. Meier, et al., Powder Technol., Vol
    343, p 855-866, 2019.
  3. J.A. Muñiz-Lerma, et al., Materials,
    Vol 11 (12), p 2386, 2018.
  4. J.M. Nicholas and T. Gray, Annu. Rev.
    Fluid Mech., Vol 50, p 407-433, 2018.
  5. J.A. Slotwinski, et al., J. Res. Natl.
    Inst. Stand. Technol., Vol 119, p 460,
  6. A. Jillavenkatesa, et al., NIST
    Special Publication 960-1, 2001.
  7. ASTM Standard Practice B215-15.
  8. ASTM Standard Test Method
  9. D. Faroongsarng and G.E. Peck,
    Drug Dev. Ind. Pharm., Vol 20 (15),
    p 2353-2367, 1994.
  10. C. Papelis, et al., Coll. Surf. A Phys.
    Eng. Asp., Vol 215 (1–3), p 221-239, 2003.
  11. O.P. Mills and W.I. Rose, Geosphere,
    Vol 6 (6), p 805-811, 2010.
  12. T. Theivasanthi and M. Alagar, arXiv:1307.1091, 2013.
  13. H.J. Holland and M.J. Murtagh, Adv.
    X-ray Anal., Vol 42, p 421-428, 2000.
  14. G. Matei, et al., Mod Dev in Powder
    Metall., Vol 8, p 5-11, 1974.
  15. N. Vluttert, bachelor’s degree thesis, University of Twente, 2016.
  16. P. Mellin, et al., Met. Powder Rep., at
    press, 2019.

Recent Posts

The Exponential Future of Metal Printing: New Technologies and Needs for Materials Science

Metal additive manufacturing has improved productivity by orders

Introduction to X-Ray Diffraction (XRD)

Different phases of materials have different benefits to

Moore’s Law for Metal Printing: Where We Are and How Do We Stay on Track?

In 1965, Gordon Moore set the path for