A spray-dried press granule is a temporary structure. It exists to carry a fine ceramic powder into a die evenly, and then to disappear under the punch. Whether granulation gets that right is decided mostly before the dryer is switched on. In the published work, whether granules come out solid or hollow depends first on how well the slurry is dispersed, and the better-dispersed slurry is the one that makes hollow granules. The rest of the problem is on the certificate of analysis (COA): granule size, bulk density, flow and "binder" all depend on how they were measured, and several standards buyers cite were written for metal powders.
How to tell a spray-dried granule from a plasma bead is covered in Two Faces of Round SiC Powder. Granules sintered to strength for thermal spray are covered in YSZ powder for thermal barrier coatings. This article is about the press granule itself.
What the dryer does, stage by stage
The slurry is fine powder in water with a dispersant (deflocculant) and a binder, sometimes with a plasticizer. Published formulations vary widely. Walker, Reed and Verma (1999) granulated alumina at 30 or 40 vol% solids with 0.35–1.00 wt% ammonium polyacrylate and a polyethylene glycol (PEG) binder. Naglieri and co-workers (2013) prepared alumina–zirconia slurries at 50, 63 and 75 wt% solids for one of their routes, but took only the 50 wt% slurry to the dryer, because destabilizing the richer ones produced macroscopic flocs; their other route was spray dried at 63 wt%. Some papers quote solids in vol% and others in wt%. The conversion depends on powder density, so a specification has to say which unit it means.
The slurry is then atomized, usually by a rotary wheel, a single-fluid pressure nozzle or a two-fluid (pneumatic) nozzle. For rotary wheels, GEA's product page, which we saw only as a search rendering, says ceramic particle size increases as wheel speed decreases. Fraunhofer IKTS quotes 10–150 µm for its ready-to-press granules and runs both two-fluid nozzles in a cocurrent fountain configuration and rotary atomizers.
Drying conditions are the least documented part. The only ceramic paper we read in full that gives temperatures is Naglieri 2013, a lab-scale ultrasonic unit at 180 °C inlet and about 85 °C outlet. We found no peer-reviewed source that measures how co-current, counter-current or fountain flow changes ceramic granule properties. After drying, the product is screened. Fiven's SiC ready-to-press sheet gives a 250 µm maximum alongside a 100 µm average, so the product has a top cut.
The granule is much larger than the powder in it. A 2004 Tosoh zirconia specification, read from a third-party mirror and possibly superseded, lists a primary particle D50 (median diameter) of 0.6 µm for its 3Y grades and a granule D50 of 55–60 µm. That is a ratio of about 100 in diameter, and the two are measured by different methods. A D50 on a COA that does not say which of the two it is could be either.
Why granules come out hollow: mostly the slurry
Hollow or dimpled granules are usually explained by solids loading and drying rate; our own post on identifying SiC granules says as much. In the peer-reviewed alumina and zirconia work we read, the main variable is neither. It is the slurry's dispersion state, measured as its yield stress.
Walker, Reed and Verma found that alumina granules made with 0.35 wt% deflocculant, which gave a high slurry yield stress, were solid, while higher deflocculant levels gave "hollow granules that contain a single large open pore or crater." Their abstract explains that at low yield stress "a crater may form from the inward collapse of the surface of a forming granule when the particle packing density in a droplet continues to increase after the droplet size becomes fixed by the formation of a rigid shell", leaving, in their words, "an internal void with internal pressure lower than that of the surrounding atmosphere". The droplet sets a rigid skin, the particles inside keep packing, and the surface caves in. A flocculated slurry resists that movement and dries solid.
Naglieri 2013 found the same in alumina–zirconia. Slurries with yield stress around 0.5 and 2.2 Pa gave "a large, central void", intermediate values gave hollow granules with thicker shells, and solid granules needed a yield stress on the order of 10–20 Pa in that study. A 2022 review of silicon nitride granulation summarizes other work the same way. For morphology, then, a better-dispersed slurry is not a better slurry. Controlled flocculation gives solid granules.
Solids loading matters too. Walker 1999 ties packing density inside the granule to solids loading, and a 2019 TiO₂–PVA study went further, calling solids content and binder dose the dominant factors for that system and finding that raising both reduced granule defects. In the alumina and zirconia work, though, dispersion state is the switch between solid and hollow. Drying rate we could not confirm. The claims that fast drying causes hollow or doughnut granules came only from a kaolin paper we saw as a snippet and from US patents. Walker's mechanism is consistent with drying rate playing a part, but that is our inference, not a finding.
The binder also moves while the droplet dries. Baklouti and co-workers reported that in alumina with 3 wt% PVA only a small fraction of the binder adsorbed on the powder, and the rest migrated to form a polymer-rich surface layer; we saw this only in a search rendering. Tanaka, Chiu and Uematsu (2006) found that PVA segregated to the granule surface and subsurface while poly(acrylic acid) stayed evenly distributed, and the poly(acrylic acid) granules pressed to more uniform compacts and sintered to 580 MPa against 485 MPa. The silicon nitride review calls a binder-saturated surface "more rigid and poorly deformable." A granule hard at the skin is harder to erase in the die.
Freeze granulation, which sprays the slurry into liquid nitrogen and sublimes the ice, is sold as avoiding binder migration and hollow shells, and a 2011 study on nanometric zirconia found freeze-granulated granules crushable where spray-dried ones were hard. It is a different process to qualify, not a dryer setting.
Why these defects survive pressing
Walker (KONA, 2003) describes what happens when granules are not erased: "Artifacts of the granule structure may persist as pores and laminations after compaction, and remain as defects in the sintered microstructure." Some intergranular pores are too large for sintering to remove. X-ray CT work published in 2019 traced several types of strength-limiting defects back to the packing structure of the granules.
The most useful result for a buyer is from Naglieri 2013. Solid and hollow granules pressed the same way (uniaxially, then isostatically at 350 MPa) both reached about 56.5% of theoretical green density. After sintering, the solid granules gave defect-free samples and the hollow ones "several, large defects." A green density on target does not show the granules were sound. Checking morphology by SEM, ideally on a cross-section, is a separate test.
Moisture, plasticizer and the binder's glass transition temperature (Tg) change how easily granules deform. Nies and Messing (1984) found densification improves when binder Tg is below the pressing temperature, though more plasticizer lowered green strength. Frey and Halloran (1984) found the degree of plasticization shifts the pressure at which granules begin to crush. Shinohara and co-workers (1999) found the strength of dry-pressed alumina varied by season, and that "high temperature and humidity contributed to an increase in the deformability of granules, reducing defect sizes in summer." Granules can press differently in summer and winter. Moisture has to be specified and controlled, not simply driven as low as possible.
Why the COA numbers mean less than they look
Granule size has to be measured dry. JIS R 1639-1 (1999), the Japanese standard for ceramic granule size distribution, allows sieving or laser diffraction, but the laser instrument must measure in the dry state, agree with manual sieving within a tolerance the parties set, and receive the sample "by gentle free fall so that the granules are not broken" (our translation). Malvern Panalytical's 2013 note on dry method development uses a well-dispersed wet measurement as the reference at which agglomerates are fully broken down. For a granule that is the wrong reference: by that logic a wet result reports something closer to the primary particles, and the binders that hold a granule together, PVA and PEG among them, are water-soluble. This is why our spray-dried SiC granules page states that granule size is measured dry and the method is named on the COA.
Bulk density is a small fraction of sintered density. Tosoh lists bulk density of 1.1–1.35 g/cm³ against a sintered density of 6.05 g/cm³, roughly 18–22%. Naglieri's lab granules had tap densities of 28–33% of theoretical. The two are not like-for-like, but both show that spherical does not mean dense. Tosoh measures by filling a cup through a vibrating sieve on a Hosokawa tester, not by the ASTM B212 Hall funnel, so comparing two suppliers' bulk densities means comparing their methods too.
Flow times do not transfer between sheets. Fiven reports Hall flow of 45 and 55 s per 25 g for its two SiC grades. GNPGraystar reports "Flow Time (Sec) 100" with the funnel and mass unstated.
LOI is not binder content. Tosoh's loss on ignition is measured at 1000 °C, and even its binder-free base grades, TZ-3Y-E and TZ-3YS-E, are allowed up to 1.2 wt%; the three binder grades on the same sheet are specified at a nominal 3.3–5.5 wt%, each with a tolerance. For SiC the figure is harder to read. Fiven reports "binder content (total organic)" of about 4.8% and 9.5%, and in the 9.5% grade the carbon sintering aid is supplied as resin, so part of that figure is the carbon source. Firing SiC in air would also burn free carbon. We found no standard method for binder content of non-oxide granules, so a SiC specification has to define what the figure includes and how it is measured.
Granule strength has a standard, but a narrow one. JIS R 1639-5 (2007) covers single spherical granules that fracture in a brittle way, and its foreword noted no corresponding international standard. Crushing force at a given strength scales with the square of diameter: a granule twice as large needs four times the force. Soft, binder-rich granules that yield plastically may fall outside its scope. The more common industrial measure is an apparent yield pressure read off the compaction curve. We give no typical value because our sources do not support one.
Which standards were written for ceramic granules
This does not make the ASTM methods wrong for ceramics. Fiven reports Hall flow on SiC granules, and thermal-spray datasheets reference B212, B213 and B822, as noted in the TBC post above. It does mean a specification should name its method, and that ceramic-specific methods exist. ISO 787, sometimes suggested, covers pigments and extenders. In Korea the KS L 1618 granule series has been withdrawn; KS L 1621 (bulk density) and KS L 1626 (flowability) for fine ceramic powders are current. We found no KS adoption of ISO 14629 or ISO 23145.
What a transfer to a toll spray dryer has to pin down
Toll spray drying of ceramics is an established service, offered by Fraunhofer IKTS among others. We found no published checklist or standard for transferring a formulation between dryers. The list below follows from the variables above. A buyer who starts from non-granulated powder, such as the raw 3Y grade on our 3Y-TZP powder page, and granulates in-house or by toll, controls each of them.
- Solids loading, with the unit. vol% or wt%, not a bare number.
- Dispersant chemistry and dose, and slurry yield stress or pH. These decide solid versus hollow. Sodium and ammonium polyacrylates are both sold; the ammonium grade is the one marketed for electrical ceramics. We found no data on sodium pickup from a sodium dispersant.
- Binder and plasticizer, type and dose. They set binder migration and how easily granules crush.
- Atomizer type and setting, inlet and outlet temperature, sieve cut. Vendors caution that changing atomizer or chamber on scale-up changes droplet size and drying history.
- Acceptance tests with the method named: dry granule size, bulk density, flow with funnel and mass, moisture, LOI with its temperature, and green density at a stated pressure. Suppliers differ here: Tosoh uses 70 MPa uniaxial, Fiven 125, 150 and 225 MPa, Martoxid 100 MPa.
- Morphology by SEM, because green density does not catch hollow granules.
How many lots, split how, and measured by whose laboratory is a separate question, covered in what three lots can and cannot prove. A binder-bearing SiC granule may also classify differently at customs from plain powder, covered in the SiC import guide.
Frequently Asked Questions
Why are some spray-dried granules hollow?
In the peer-reviewed work we read, the main cause is a well-dispersed slurry with low yield stress: a rigid shell forms on the droplet while particles inside keep packing, and the surface collapses. A flocculated slurry with higher yield stress dries solid. Solids loading also has an effect; drying rate we could not confirm from a peer-reviewed full text.
Can granule size be measured by laser diffraction?
Yes, if dry. JIS R 1639-1 allows laser diffraction only with dry dispersion, calibrated against manual sieving, with the sample fed by gentle free fall. A wet measurement is meant to break agglomerates apart, and a granule is an agglomerate held by water-soluble binder.
Is LOI the same as binder content?
No. Tosoh's binder-free zirconia grades carry a loss-on-ignition (LOI) limit of 1.2 wt% at 1000 °C with no binder in them, so LOI includes moisture and other volatiles. On SiC granules a "total organic" figure can include a resin carbon source. State what the figure includes and how it is measured.
If green density meets target, can the granules still be bad?
Yes. In one alumina–zirconia study, solid and hollow granules pressed to the same green density of about 56.5% of theoretical, but only the hollow granules left large sintered defects.
Are ASTM B212 and B213 the tests for ceramic granules?
They were written for metal powders, though ceramic suppliers use them. Ceramic-specific methods are ISO 14629, ISO 23145-1 and -2, and the JIS R 1639 series. Whichever you use, name it in the specification.
References (Public Sources)
- Walker, "Persistence of Granular Structure during Compaction Processes", KONA 21, 2003, and Naglieri et al., Materials 6(11), 2013, on alumina–zirconia slurries for spray drying. These are the only two papers we read in full. Naglieri states its critical yield stress in two inconsistent ways, so we give only the order of magnitude.
- JIS R 1639-1 (1999) and R 1639-5 (2007), read in full; R 1639-2 to -4, scope only. Malvern Panalytical's technical note on dry powder method development, 22 February 2013, read.
- Grigoriev et al., silicon nitride spray-drying review, Materials, 2022, read through an HTML summary; the PDF was blocked.
- Abstracts only: Walker, Reed and Verma, J. Am. Ceram. Soc., 1999; Tanaka, Chiu and Uematsu, 2006; Nies and Messing, and Frey and Halloran, 1984; Shinohara et al., 1999; Okuma et al., Sci. Rep., 2019; Avcıoğlu and Özkal, J. Korean Ceram. Soc., 2019; Raghupathy and Binner, J. Am. Ceram. Soc., 2011. Seen only as search renderings, and hedged above: Baklouti, Chartier and Baumard, J. Eur. Ceram. Soc., 1998; the kaolin paper and the US patents on drying rate.
- Supplier documents, read in full, each figure one supplier's typical value: Tosoh zirconia specification valid August 2004, from a third-party mirror and possibly superseded; Fiven SIKA Densitec 13/15 (2020); GNPGraystar green SiC ready-to-press (Rev. 02/2020); Huber/Martinswerk Martoxid (Rev I, August 2022), from a distributor's site. Web pages read: Fraunhofer IKTS, Vanderbilt DARVAN. GEA's spray dryer page and freeze-granulation vendor pages were seen only as search renderings.
- Standards: we did not read the ASTM texts, and astm.org and iso.org both blocked us. ASTM scopes come from ASTM and ANSI listings; ISO 3923-1 and 3953 from official preview pages; ISO 4490, 14629 and 23145 from listings. KS status is from the e-나라표준인증 portal, which gives no withdrawal date for KS L 1618.
We found no verifiable information on Korean toll spray-drying practice, so none appears above.
Nami Tech Solutions supplies spray-dried SiC and 3Y-TZP press granules and arranges toll spray drying to a buyer's own formulation, project by project. Most of that work is writing the variables above into the order, with units and methods, and reading each lot's COA against it line by line.
