As designers and engineers, we are often interested in the mechanical behavior of our parts and when it comes to 3D printed parts, there are certain rules of thumb that can be used as guides when we design and print parts. For example, it is well established that a tensile specimen printed in the upright build orientation has a lower strength than the same specimen printed in the flat or on-edge build orientations. See our previous post related to that topic.
In this post we will look at how layer height influences the interlayer (aka z-direction) strength of an FFF part to see if there is a rule of thumb for interlayer strength and layer height. As we will see, a review of the literature yields contradicting data. Some results indicate that the strength increases as the layer height increases, while data from other groups shows that the strength decreases as the layer height increases. Let’s dive in and discuss results from 2 different papers.
A paper by Chacón et al. (1) was released in 2017 which studied the effects of the build orientation, layer thickness, and feed rate on mechanical behavior of FFF parts made from PLA. They noted that no standards for testing tensile and flexural properties of FFF parts exist, so they based their test specimens and methods on the ASTM D638 and D790 standards. Adapting existing standards is a common approach amongst AM research groups and in this case resulted in solid (100% infill) dogbone shaped tensile specimens and rectangular shaped flexural specimens. Their paper has a lot of data, but we will just focus on a plot of the tensile strength vs. layer height for specimens printed in the upright (standing) build orientation as shown in Figure 1. These specimens are well-suited for evaluating the interlayer strength since the upright build orientation produces layers that are stacked in the direction of load application. The trend in Figure 1 is clear: the interlayer bond strength increases as the layer height increases. Even adjusting the feed rate did not affect the trend. The authors offer an explanation of this trend by stating, “This effect can be explained by considering that with increased layer thickness, fewer layers were needed for a given total thickness and, therefore the number of layer bonds was reduced and strength increased.” It is true that the number of layer bonds is reduced with thicker layers, but their argument linking the number of layer bonds to the interlayer strength doesn’t quite make sense. Statistically, your odds of having a poorly bonded layer increase as the number of layers increases but the number of layers does not directly correlate to the strength of an individual bond.
The results from their upright flexural specimens also showed that the strength increases with layer height, although the flexural trend wasn’t as clear.
Now, consider some data that contradicts the results of the Chacón paper.
Few papers are dedicated solely to studying the interlayer bond strength as a function of layer height and layer width but Kuznetsov et al. did just that in (2). In their paper, they also note that there is no official standard for measuring the bond strengths, however, instead of adapting existing standards, they take a novel approach by fabricating tubes with a rectangular cross-section and 0% infill, printed upright, and tested in 3-point bending. Since they printed upright and tested in bending, the interlayer bond strength is the driving factor in the specimen strength. Like Chacón, they used the FFF method and PLA filament. They varied both the nozzle diameter, which is effectively the layer width, and the layer height and their results show a clear trend. As shown in Figure 2, the flexural strength of the specimen increases as the layer width to layer height ratio increases. In other words, as the layer height increases, the strength decreases.
The authors make the argument that, “the most important factor that defined the strength of the resulting part was the interlayer contact surface area”. They provide a very useful figure (Figure 3) with images of the printed cross-sections that illustrates how the layer height and layer width affect the printed geometry. In short, as the layer height increases, so does the void volume fraction. This is clear when comparing the left and right images in Figure 3. As the void volume fraction increases, the bonded area between layers decreases and thus the interlayer strength of the part decreases.
This argument makes sense because a part with no voids would have a z-direction strength that would be close to the bulk material strength. On the other hand, a part with voids will have reduced contact areas between layers and therefore the interlayer strength will be less than the bulk material strength.
So which trend is correct? Does the interlayer strength increase or decrease as the layer height increases? We are afraid the answer is, “it depends”. We have no reason to doubt the methods from either group, so we must assume that the results from both groups are credible and repeatable. The question then becomes: What is causing the difference in trends? We believe the answer lies in the vast number of variables, aside from layer height, that may have an influence on the bond strength. In general, these variables can be grouped into 2 categories: geometric and thermal. There are certain parameters that affect the as-printed geometry and there are certain parameters that affect the thermal history of regions within the part. We know, at the very least, that the 2 groups mentioned here did not use the same 3D printer to manufacture their specimens and we know that the specimen geometry used by each group was different. Further, we can safely assume that several, if not many, of the print parameters used by each group were not the same.
We tend to agree with the Kuznetsov group when they argue that the biggest factor in determining interlayer strength is the bonded area between beads which is another way of saying the as-printed geometry is the driving factor. Is the as-printed geometry the only factor driving reductions in z-direction strength? It is highly unlikely that the geometry is the only factor influencing the bonded strength of FFF parts, but it surely is a significant factor. Undoubtedly other factors, such as nozzle temperature, cooling rates, and polymer type, influence the bonded strength, but to what degree? It is uncertain at this point and that brings up a larger issue. There are a lot of print parameters and a lot of 3D printers. This means that there are many variables that could potentially influence the mechanical behavior of an FFF part. Being able to isolate and identify parameters that are responsible for a given mechanical property becomes quite a challenge. While the community has been generating experimental data for some time it still feels like we are in the discovery phase of learning how all the pieces fit together. Fortunately, the rate of research output in this field is increasing, making it more likely that we eventually will have definitive answers for many of the interesting questions posed by FFF parts.
One thing that is long overdue is the development and release of official test standards for additive manufacturing. Such standards will be a welcome step in the right direction by eliminating many of the aforementioned variables and allowing for more direct comparisons between data sets. The good news is that ASTM and ISO have announced a framework for additive manufacturing standards, so it would appear that standards are at least in the works. We are certainly looking forward to them.
1. Chacón JM, Caminero MA, García-Plaza E, and Núñez PJ. Additive Manufacturing of PLA Structures using Fused Deposition Modelling: Effect of Process Parameters on Mechanical Properties and their Optimal Selection. Materials and Design, 2017, Volume 124, pp 143-157.
2. Kuznetsov VE, Solonin AN, Urzhumtsev OD, Schilling R, and Tavitov AG. Strength of PLA Components Fabricated with Fused Deposition Technology Using a Desktop 3D Printer as a Function of Geometrical Parameters of the Process. Polymers, 2018, Volume 10, Issue 3, 313.