Life prediction and experimental validation of a 3D printed dental Titanium Ti64ELI implant

Fatigue analysis plays a vital role in determining the structural integrity and life of a dental implant. With the use of such implants on the rise, there is a corresponding increase in the number of implant failures. As such, the aim of this research paper is to investigate the life of 3D-printed dental implants. The dental implants considered in this study were 3D printed according to the direct metal laser sintering (DMLS) method. Additionally, a finite element model was developed to study their performance, while fatigue life was predicted using Fe-Safe software®. The model was validated experimentally by performing fatigue tests. The life of the dental implants was analysed based on Normal strain and the Brown-Miller with Morrow mean correction factor algorithm. The model revealed that there was a strong correlation between the FEA and the experimental results. The clinical success of 3D-printed dental implant experimentally is 20.51 years and computationally under Normal strain is 19.89 years and Brown-Miller with Morrow mean correction factor is 26.82 years.


Introduction
Dental diseases have a considerable impact not only on people's self-esteem but also on their eating ability, nutrition and health [1]. Problems concerning various aspects related to dental implants always remain relevant and aesthetics is the main reason why people often choose implants instead of other methods [2]. Dental implants are used to retain and support both fixed and removable dental prostheses [3] and have been reported as the preferred treatment modality for completely and partially edentulous patients [4].
Recently, three-dimensional printing (3D printing) technology has found its way into orthodontic surgery [5], [6]. Additive manufacturing allows a variety of materials to be used in 3D printing production. This printing technology facilitates the printing of complex geometries and allows for the creation of porous structures [7], [8]. There are different techniques by which 3D printing can be undertaken, with direct metal laser sintering (DMLS) being the most advanced of the available techniques [9], [10].
In dentistry, three-dimensional finite-element (FE) analysis has become an increasingly useful tool for the prediction of the effects of stress on the implant and its surrounding bone. A number of studies have been carried out to investigate the fatigue life of dental implants [11]- [13]. A numerical study was carried out to investigate the stress distribution between bone and dental implant and it was found that the stress is Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 18 August 2021 doi:10.20944/preprints202108.0379.v1 higher in a two-piece diameter implant as compared to a one-piece diameter implant [14].
Experimentally fractured dental implants were validated using FEA analysis with the results showing that FEA-predicted lifetime was within the 95 % confidence interval of lifetime estimated by experimental results. This suggests that FEA prediction is accurate for this implant system [15]. Fatigue testing of narrow and extremely narrow dental implants demonstrated that the von-Mises stress at the dental implant and abutment has high reliability (up to 97.5 %) at 50 and 100 N, with decreased reliability observed for both groups at 150 and 180 N (ranging from 0 to 82.3 %) [16].
The life prediction model plays a critical role in minimising technical problems relating to both the prosthesis and the dental implant components, including screw loosening or screw fractures, abutment fractures and implant fixture fractures. The use of fatigue life prediction models is well documented [17], [18]. The Morrow concept is suitable for low stress cycle fatigue and it was, furthermore, predicted that lifetime at loading F = 357 N is equal to 7 390 092 cycles.
Weibull evaluations can obtain satisfactory results, but are subject to a large evaluation error at the boundary loading [19]. 3D-printed dental implants as specimens.

Methodology
This section outlines the methodology that was previously used [11], [20]. The first section describes the stress analysis of the 3D printed dental implant model where the oblique loading is applied to simulate a worst-case scenario. The stress analysis results are then exported to Fe-Safe software® to perform fatigue analysis. The finite element analysis model is then validated experimentally by performing fatigue tests [19], [21].

Finite element modelling
Dental implants were designed using Abaqus CAE software with Fe-Safe software employed to perform fatigue analysis. The 3D model had three components, i.e., the crown, dental implant and specimen holder (see Figure 1). They were then connected as one body as shown in Figure 1 (B & C). The dental implant and crown were modelled in titanium to mimic the experimental model.

Material properties
All materials used in the model are considered to be isotropic, homogeneous, and linearly elastic. The elastic properties were taken from the literature, as seen in Table 1. Table 1: Material properties utilised to perform finite elements analysis.

2.1.2.Finite element analysis mesh study
Linear tetrahedron element (C3D4) was considered in the current study and this element type were previously used successfully [15], [23], [27]. The mesh convergence study was conducted and the results are presented in Figure 2. The total linear tetrahedron element (C3D4) was 223794. The mesh size were 1 mm around the specimen holder and 0.2 mm at the threaded interface. To minimize distortion, finer mesh were applied around the threads of dental implant and specimen holder. The mesh size was achieved by gradually reducing the default mesh size of the implant until the stress curve started to flatten out and a constant result was obtained.

Interface conditions and constraints
The specimen holder-3D printed dental implant interface was assumed to be perfect simulating complete osseointegration [25], [28]. Therefore, connections between 3D printed dental implant, specimen holder and crown were designed to be bonded. Surface to surface constraints were applied between 3D printed dental implant, specimen holder and crown. The bottom of the specimen holder was fixed and thus other faces were free of the condition (see Figure 3). While chewing forces are dynamic, to simplify the problem, a static analysis is done in the majority of studies [29]. The specimen holder, crown and dental implant are inclined at 30 0 to the horizontal while the force is applied vertically downwards at the centre of the crown.

Fatigue life model
The fatigue life prediction models are developed subsequent to the FEA model and are based mainly on the cumulative damage to the materials under cyclic loadings. The models are formulated according to the correlation between the cyclic numbers and the local stresses and strains. A number of models have been developed in the literature [17], [30], [31]. In this project, Normal strain life and the Brown Miller with Morrow mean correction method are used -a method first discovered by Wohler -consisting of a plot of alternating stress (S) and cycle to failure (N) [15]. The strain-life method is based on the observation that, in many components, the response of the material in critical locations is strain dependent.

Equations 1 -3, below, represent strain life, Normal-strain life and Brown-Miller with
Morrow mean correction factor models used to compute fatigue analysis [15], [32] n a a  is the mean normal stress on the critical plane, where p  is the plastic strain. The fatigue strength exponent and coefficient are taken from Basquin's law: where e  is the elastic component of the cyclic strain amplitude, and a  is the cyclic stress amplitude. The material properties are approximated using Seeger's method with the help of the re-scaling conventional monotonic ultimate tensile stress [15].
Several methods have been developed to determine the mean stress. Mean stress, defined as the mean value of peak tensile stress and compressive one, has shown a significant effect on both deformation and fatigue behaviour [17]. Typically, Morrow's concept is suitable for low stress cycle fatigue and can be represented by the below equation: where a  is the cyclic stress amplitude, m n.

Experimental setup
In this research, the EOS M290 3D printer -a direct metal laser sintering (DMLS) additive manufacturing system (specifications listed in Table 3) -is used for sample preparation on a Titanium Ti64ELI powder. The physical specimens were tested (n =3, 10 Hz, R = 0.1) using an MTS Acumen fatigue machine until fracture occurred (see Figure 4). The specimens were fixed according to the International Standard of dynamic testing of single-post endosseous dental implants, which were replicated in the FEA model [33]. It was reported that under ISO protocol testing in air and normal saline solution are equivalent in terms of likelihood of fracture versus runouts [34]. As such, the current study consider air as testing environment. The experimental models were tested at 80 % of the maximum load, as recommended by ISO 14801, and this maximum load was derived from FEA model [15], [33]. Dental implants with a diameter of 3.4 mm were 3D printed and the crown and onepiece dental implant were printed as one unit, as illustrated in Figure 6. Showing 3D-printed Dental implants after heat-treatment. The heat treatment were performed for 2 hours hold at 800 0 C in protective Argon atmosphere. [A] [B]

Results
The 3D model was simulated successfully and had good similarity with the physical specimens. The performance is predicted at 80 % of the maximum loading (810.5 N) and the clinical success of the specimens was evaluated by computing the number of years that the dental implants will likely survive. The contour plot results are presented in two parts, the maximum loading and 80 % of the maximum load, as shown in the Figure 6.  [B]

Finite element analysis (FEA) Results
The ISO14801 recommends that minimum of two threads be exposed when testing the implants, see Figure 6(B). The contour plot in Figure 7 shows the higher stresses that are located in the contact area between implant and specimen holder. Miller with Morrow mean correction factor, with a figure of 6.991 x 10 -6 . Surface roughness between the dental implant and the mandibular bone plays a critical role in determining a life of the implant and initial stability [35]. The surface roughness of 1.6 < Ra < 4 mm was selected based on previous studies [36], [37].  The failure of the 3D printed dental implant occurred in the second and third threads, which is in agreement with FEA results (see Figure 10). The cycle target was set for 5×10 6 cycles but, with a loading of 80 %, it was observed that a minimum of 137 433 and maximum of 262 142 cycles respectively could be reached before failure occurs. The clinical success of the 3D printed dental implants is 20.51 years respectively when masticatory loading of 648.4 N is applied. Figure 10: Fractured 3D-printed dental implants after experimental study according to ISO 14801 standard's fatigue test.

Statistical modelling of 3D-printed dental implant
The model validation is a useful tool in further quantifying that the FEA fatigue results are correct. The experimental models were 3D-printed and fatigue testing was performed in dry air conditions. The experimental results were reported in  factors that are noted to affect the process, based on the priority confidence intervals in the model [39].

Discussion
The results of this research are presented and the study relies on the production method of the dental implants, which is 3D printing. The dental implants were 3D-printed and computational models were developed to analyse stress distribution and predict the life of the 3D-printed dental implant. The results presented herein are not limited to the implants themselves but apply more generally to the material they are made of, namely Ti64V ELI and the production method.
Two vital parameters that lead to success with dental implants are design and insertion technique [40]. The design of dental implant plays an important role in the stress distribution that occurs primarily where bone is in contact with dental implant [41]. The clinical reports showed dental implant body fracture was more frequently observed in reduced-diameter implants as compared to regular-diameter implants [31]. The choice of pitch size was conceded because, the finer the pitch, the more threads on the dental implant body and the more complex the 3D-printing process will be. In the present study, the pitch size lies between 0.6 -0.8 mm and the dental implant is 3.4 mm in diameter.
The combination of experimental and finite element analysis models for stress analysis is capable of providing reliable results [41]- [43]. However, FEA studies of dental implants with validation experiments are relatively rare [44]. The maximum stress levels of the implant-bone boundary influence biological reactions, including bone resorption and remodelling [45], [46]. Orthodontics is gradually changing from opinion-based practice to evidence-based practice [47]. In the present study, a specimen holder was used to fix the 3D-printed dental implant during the experimental test and FEA analysis. The same methodology was previously applied successfully [15], [21], [48].
Oblique loads generate higher stress and displacement that is much greater than that produced by axial loading [23], [49], [50]. In the current study, we simulated a masticatory  Figure 3). In our results, 30 0 of loading direction showed the highest stress concentration between the second and third threads and this was due to worst-case scenario masticatory simulation. This is in agreement with previous work [20], [51].
Generally, Titanium materials have a low elastic modulus, which can eliminate bone resorption problems and minimise height contact stress between the bone and dental implant [52], [53].
The clinical evaluation of dental implant stress distribution is important because it is possible to predict where the fracture or failure will occur [11], [23], [27], [54]. Fatigue may cause such implants to break, with serious consequences from a clinical standpoint [55].
The fatigue algorithms used in the current study are Normal strain and Brown-Miller with Morrow mean correction factor [17]. The results show that Normal strain led to 169565.5 cycles and that Brown-Miller with Morrow mean correction factor led to 143040.7 cycles ( strain. In addition, it was predicted by FEA analysis that stress concentration was at the root of the implant body screw thread adjacent to the simulated bone level (Figure 7). This result correlates well with the consistent failure mode of implant body fracture among all of the tested implants ( Figure 10).
The brittle nature of titanium (Ti64V ELI) tested in air media was observed in all samples and exhibited the shear stress failure (see Figure 10). Similar to the findings of other studies, oblique forces were used to develop FEA models and the results were then exported to the Fe-Safe software for fatigue analysis [11], [18], [38]. The phenomenon of fatigue in orthodontics is of vital importance and, as the global number of individuals receiving dental implants increases, bone loss, fractured dental implants and peri-implant diseases are growing problems in clinical dentistry [53], [56].
The models were then validated experimentally by performing fatigue experimental tests and ANOVA was used to analysis the results. If the ratio of the F-value is close to one, it is unlikely any factor has a significant effect on the Normal strain. However, in Brown-miller was 0.8484. Thus, this confirms that there are no issues with the data or the model. This study can provide a starting point for analysis of 3D-printed dental implants and the possibility of using these in design of dental implant. The importance of the results is stressed, given the small size of 3D-printed dental implants, which can present problems of scale when they are manufactured by the additive manufacturing technique.

Limitations of this study
Small sample number (n = 10) reduces the predictive ability of the study from a statistical point of view. Future studies will seek to overcome this shortcoming by using a larger sample number. The test is limited to dry air condition testing environment. Single angle is considered for worst case scenario. The study considered single body dental implant, this is a limitation of the methodology due to the difficulties in 3D-printing of dental implants.

Conclusion
The possibility of using finite element analysis in predicting the performance of 3D printed dental implants was investigated. The life prediction differences between two algorithms were reported and it was found that the 3D-printed dental implants in the current study exhibit their highest performance under the Normal strain method, as years. The ANOVA analysis of variance shows the relationship between FEA and the experiment, and the results show that the p-value of the normal strain is less than 0.05, which means that the model term is significant. Therefore, the correlation coefficient is less than 1, indicating that there is no problem between FEA and experimental data.
Through comparison, it can be seen that the model prediction and the actual experimental data obey the normal distribution.

• Acknowledgments
The research is supported by the Tshwane University of Technology and the University of South Africa (UNISA). The experiments were performed in biomechanics laboratory facilities at University of South Africa's Science campus in Johannesburg, South Africa.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.