1938 E Harney Street, Laramie, WY 82072
1-307-314-2891   | 

  • linked-in
  • twitter
  • youtube
  • Rick Dalgarno

Introduction to Anisotropy of 3D Printed Parts

As additive manufacturing of plastics continues to shift from prototyping to production there is an increasing demand from designers and engineers to better understand the behavior of 3D printed parts so that they can have confidence their parts will meet performance requirements. Compared to traditional manufacturing methods like injection molding, parts made from 3D printing methods have unique characteristics that make their behavior more unpredictable. While there are many unique traits, I will discuss anisotropy today beginning with a simple thought experiment.

Image a cube that is injection molded with PLA. Now imagine measuring the stiffness of the cube in the 3 principal directions. Would you expect the stiffnesses to be the same or different? Of course you would say the stiffness would be the same and you would be right.

Now imagine a cube that is 3D printed using the Fused Filament Fabrication (FFF) method. Do you think the 3 stiffnesses of this cube would be the same? Intuition might suggest that the stiffnesses should be the same. After all, the cube is made from an isotropic material. But intuition isn’t right in this case. The stiffnesses of the block in the 3 directions will be different. Either by a little or a lot depending on the print parameters and build orientation. Let’s take a closer look to understand one of the reasons why.

All about the voids

Most plastics are isotropic so shouldn’t a “solid” block of material that is 3D printed also be isotropic? Well, I put solid in quotations on purpose because a 3D printed solid block isn’t actually solid. There are air gaps or voids between the printed filaments, unlike an injection molded part which does not have voids and is safely described as solid. In more technical terms, a 3D printed part is heterogeneous while an injection molded part is homogeneous. An example of the voids between printed filaments is shown in Figure 1. These voids create anisotropy in “solid” 3D printed parts.

Figure 1. Cross-section of FFF part showing voids between printed filaments (1).

Consider the experimental results from Zaldivar et al. (2) who studied the effects of build orientation on the mechanical properties of dogbone shaped parts made with ULTEM® 9085 and the FFF manufacturing process. The different build orientations are shown in Figure 2. The experimental stress-strain results are given in Figure 3 and one easily sees that the build orientation affects the part stiffness, yielding, ultimate load, and strain-to-failure. A clear disparity exists between orientation C and D. Orientation C has a modulus of 2.5 GPa, a tensile strength of 71 MPa, and a strain-to-failure of 7.4% and orientation D has a modulus of 2 GPa, a tensile strength of 38 MPa, and a strain-to-failure of 2%. Relative to C, D has a 20% lower modulus, 45% lower strength, and 73% lower strain-to-failure. These are massive differences for parts that appear to be the same!

Figure 2. Schematic (2) of dogbone build orientations. The xy-plane is the plane of the build plate.

Figure 3. Experimental stress-strain curves (2) for each build orientation.

At first glance it may not be obvious why the part behavior is so dependent on build orientation because from a distance, each part looks similar. Upon closer inspection, however, we find that the parts have unique meso-structures which are formed when the filament is deposited layer by layer in the part. These meso-structures can be seen in the cross-section views of the parts shown in Figure 4. The particular pattern in which the filament gets deposited is based on instructions (g-code) from the slicing software that was used to slice each part into layers. The pattern for a given layer is dependent on the orientation of the part on the print bed of the 3D printer. These patterns have a significant effect on things like void volume fraction, void shape, filament adhesion, and more which collectively alter the mechanical response of the part - for better or worse.

Figure 4. Cross-section views (2) of the dogbone parts based on build orientation. The black regions are voids.

Returning to the comparison of orientations C and D, the primary reason why D has much lower strength and strain-to-failure compared to C is due to the layer stack direction relative to the direction of the applied load. Figure 5 shows the layer stack and applied load directions for orientations C and D. For orientation C, the layers are stacked perpendicular to the load direction and in order for the part to break into 2 pieces, the filaments must rupture. For orientation D, the layers are stacked in the same direction as the load direction and the primary failure mode is delamination between print layers. In other words, C is stronger than D because in C, the filaments must rupture but with D, only the bond between layers needs to fail.

Figure 5. Illustration (2) showings the layer stack orientations relative to the applied load direction for orientations C and D.

We’ve only scratched the surface on the topic of anisotropy in 3D printed parts in this post. My intent was to simply introduce the idea that 3D printed materials and parts are anisotropic and it is important to keep this in mind, particularly when designing parts that will have performance requirements such as functional prototypes, jigs, fixtures, and production parts.


1. Rankouhi B, et al., Failure Analysis and Mechanical Characterization of 3D Printed ABS With Respect to Layer Thickness and Orientation, Journal of Failure Analysis and Prevention, Volume 13, Issue 3, pp 467-481, 2016.

2. Zaldivar RJ, et al., Influence of processing and orientation print effects on the mechanical and thermal behavior of 3D-Printed ULTEM® Material, Additive Manufacturing, Volume 13, pp 71-80, 2017.