# Binder jetting

Powder-and-binder shaping, metal densification versus printed sand tooling, shrinkage compensation and final-part qualification.

Reviewed: 2026-11-01

Tags: processes, shaping, additive-manufacturing

With assumed **18% linear shrinkage relative to the green length**, a **50.00 mm sintered target** requires **60.976 mm green length** (50.00 ÷ 0.82), or **21.95% enlargement relative to the target**. Adding 18% instead gives a **48.38 mm** sintered result. Qualify directional shrinkage and distortion for the actual powder, furnace and support route.

For metal binder jetting, specify the route from green print through cure, depowdering, thermal debinding and furnace sintering. Densification during sintering changes the dimensions. [Fraunhofer IFAM: metal binder jetting](https://www.ifam.fraunhofer.de/en/technologies/metal-binder-jetting-from-prototype-to-series-production.html)

## Identify what is being delivered

Three routes illustrate why “binder-jetted” is insufficient as a material description:

| Route | Printed object and subsequent purpose |
| --- | --- |
| Sintered metal | A powder/binder green shape becomes a densified metal component after binder removal and sintering |
| Infiltrated metal | A printed metal skeleton receives another material; the final composition differs from the starting powder |
| Sand mold or core | Bound sand tooling forms a cavity or internal passage during a separate metal-casting operation |

ExOne’s sand process selectively bonds sand layers to produce molds and cores. Its current portfolio includes furan- and phenolic-based binders and different sands; it also describes core-wash coatings and, for non-casting applications, epoxy infiltration or sealing. These are application-specific routes, not a requirement to sinter every binder-jetted object. [ExOne: printed sand molds and cores](https://www.exone.com/en/3d-printing-solutions_overview/3dp-solutions-sand-casting/)

For an infiltrated-metal example, ExOne’s **Manufacturing Guidelines for Metal Infiltrated Parts**, Revision 5, effective **February 22, 2019**, describes its 420 stainless steel/bronze material as 60% stainless steel and 40% bronze infiltrant, with the bronze identified as 90% Cu/10% Sn. Do not identify that finished matrix simply as wrought Type 420 or apply properties from a single-alloy sintered material. The percentages are the producer’s description; no mass-versus-volume basis is inferred here. [ExOne guidelines, printed p. 3](https://www.exone.com/ExOne/media/About/mfg-guidelines-for-metal-infiltrated-parts.pdf)

## Printing support is not furnace support

During printing, loose surrounding powder supports the bound geometry. Desktop Metal’s **©2023 Ultimate Guide** distinguishes this from sintering, when supports or setters may be required. Its process description also treats curing and depowdering as necessary stages before densification. Green-part handling and cleaning therefore constrain geometry independently of whether the printer can place binder at the desired location. [Desktop Metal guide, printed pp. 14–16 and 31](https://www.desktopmetal.com/uploads/2023web_DesktopMetal_UltimateGuidetoMetalBinderJetting.pdf)

A cavity must have a practical powder-removal route. ExOne’s infiltrated-metal guide explicitly requires access for loose-powder removal and explains that deeper or more complex cavities can need larger or additional openings. Those guidelines are for its identified infiltrated materials; their numerical feature limits should not be transferred to another metal, powder or machine. Agree removal access and verification for the actual internal geometry. [ExOne guidelines, printed p. 7](https://www.exone.com/ExOne/media/About/mfg-guidelines-for-metal-infiltrated-parts.pdf)

For quoting, review fragile walls, long unsupported spans, enclosed passages, handling contact points and the surfaces available for furnace support. Determine which support contacts or infiltration attachments need removal and whether their witness marks are acceptable. A successful external shape does not demonstrate that an inaccessible passage is clean.

## Shrinkage, distortion and orientation

Fraunhofer identifies powder packing and sintering activity as important influences on densification and shrinkage. A qualified route consequently includes the powder and binder system, printing settings and thermal processing, not just the nominal alloy. Different final densities can result from different sintering conditions. [Fraunhofer IFAM: process-chain development](https://www.ifam.fraunhofer.de/en/technologies/metal-binder-jetting-from-prototype-to-series-production.html)

Uniform enlargement is only one compensation tool. Desktop Metal’s Live Sinter documentation explicitly models non-linear shrinkage, gravity, friction, deformation and density variations from powder spreading. It allows different printing and sintering orientations, ceramic setters and machining offsets. These features show why an enlarged CAD model alone cannot establish final dimensional capability. Software predictions still need confirmation on the chosen geometry and route. [Desktop Metal: Live Sinter](https://www.desktopmetal.com/products/livesinter)

Keep **direction-dependent dimensional change** separate from **direction-dependent mechanical properties**. Neither an assumed equal shrinkage in X/Y/Z nor a supplier’s X/Y tensile result establishes performance in every orientation. Record build orientation, furnace orientation and test-specimen orientation as distinct information.

### Original shrinkage calculation

Assume a development trial establishes **18% linear shrinkage relative to green length** in one direction, and the desired sintered dimension before machining is **50.00 mm**. Under an ideal uniform-shrinkage model:

`final length = green length × (1 − 0.18)`

`required green length = 50.00 / 0.82 = 60.976 mm`

Applying the wrong denominator gives 59.00 mm green and the 1.62 mm shortfall shown above.

These are invented development assumptions, not a universal shrinkage factor. If actual shrinkage were 17–19%, that same 60.976 mm green dimension would produce approximately **50.610–49.390 mm** before other distortion. Applying 18% independently in three directions would leave `0.82³ = 0.5514`, or about **55.14% of the original volume**; linear shrinkage and volume change must not be confused. This volume calculation does not by itself predict density because binder removal and mass changes also matter.

## Named property example: Shop System 316L

Desktop Metal’s **©2024** two-page sheet bearing identifier **20240508_MDS_Shop_PuerSinter_316L** contains two separate processing tables. Page 1, explicitly for its **PureSinter Furnace**, lists as-sintered tensile strength **540 ± 11 MPa**, yield strength **185 ± 11 MPa**, elongation **77.3 ± 5%** and density **7.79 ± 0.09 g/cm³**. Its footnote limits tensile, yield, elongation and modulus entries to **X and Y orientations**. These are reported supplier data, not guaranteed part minima or a complete Z-direction qualification. [Desktop Metal 316L sheet, PDF p. 1](https://www.desktopmetal.com/uploads/20240508_MDS_Shop_PureSinter_316L.pdf)

Page 2 is labeled **third-party commercial furnace** and has different values. Do not merge those results with the PureSinter table. Page 1’s corrosion results concern specimens cleaned and passivated before testing; they do not establish corrosion performance of an untreated internal surface or every service fluid. The sheet also states that geometry, application and evaluation conditions affect end-use performance. Its printed identifier and ©2024 are retained here without treating the filename as proof of a separately stated revision date. [Same 316L sheet, both pages and footnotes](https://www.desktopmetal.com/uploads/20240508_MDS_Shop_PureSinter_316L.pdf)

## Acceptance and manufacturing handoff

Specify final dimensions after the required sintering, heat treatment and finishing. Define critical surfaces, machining stock, texture and internal-passage cleanliness. For structural or fluid-handling parts, agree the relevant chemistry, condition, density/porosity assessment, mechanical tests and leak or pressure acceptance rather than relying on “high density” as a complete requirement. Density, tensile behavior and leak tightness answer different questions.

For printed sand tooling, qualify the mold/core and resulting casting separately: sand/binder selection, removal and coating access, handling strength, assembly, casting geometry and downstream machining belong in the foundry route. A sand printer’s build envelope is not a guaranteed finished-casting tolerance.

A useful release package includes drawing and model revisions, intended route and final material, powder/binder identification, build and furnace orientations, approved compensation model, support/removal strategy, thermal-process record requirements, secondary operations and inspection stage. Compare cost and lead time through accepted finished parts, including depowdering, furnace capacity, finishing and yield, rather than printer cycle alone. Desktop Metal’s guide explicitly distinguishes printing time from the complete processing sequence. [Desktop Metal guide, printed p. 31](https://www.desktopmetal.com/uploads/2023web_DesktopMetal_UltimateGuidetoMetalBinderJetting.pdf)

The ExOne manufacturing guide is a historical, identified route reference, not evidence of current service availability. The cited Fraunhofer, Live Sinter and ExOne sand pages are live descriptions without a separately verified technical issue date. The worked calculation is original.

## Related reference

[Manufacturing processes](../index.md)