Wear Behavior and Multi-Technique Characterization of 3D Printed TPU Under Simulated Pharmaceutical Operating Conditions

TL;DR

This study examines how printing parameters affect the wear and properties of 3D printed TPU materials used in pharmaceutical manipulators.

Contribution

The study demonstrates how printing parameters influence the wear resistance and mechanical properties of TPU materials for pharmaceutical applications.

Findings

TPU 51A printed at 265 °C with four layers showed reduced wear and lower hardness.

TPU 60A with higher printing temperatures and layers had increased hardness but higher wear.

Fewer printed layers improved tensile properties and interlayer bonding.

Abstract

This study investigates the wear behavior and multi-technique characterization of 3D printed thermoplastic polyurethane (TPU) intended for friction layers in transmission belts used in pharmaceutical manipulators. Two flexible TPU grades—TPU 51A and TPU 60A—were printed using fused deposition modeling (FDM) with varying printing temperatures (255–265 °C for 51A; 225–235 °C for 60A) and layer counts (three or four layers). Specimens were evaluated for Shore A hardness, wear resistance (mass loss using a Baroid lubricity tester under dry sliding against carton), tensile properties, crystallinity (XRD), chemical structure (FTIR), thermal stability (TGA), and scanning electron microscopy (SEM). The results show that printing parameters significantly influence the mechanical and tribological behavior of the materials. For TPU 51A, increasing the printing temperature to 265 °C and using four layers led to a substantial reduction in cumulative mass loss, although hardness decreased. In contrast, for TPU 60A, higher printing temperature and layer count increased hardness but also resulted in higher wear. Tensile tests indicated that specimens printed with fewer layers exhibited higher yield strength and strain, indicating improved interlayer bonding. XRD analysis confirmed the predominantly amorphous nature of the printed samples, with a reduction in crystallinity compared to the raw filaments. FTIR spectra showed no significant chemical degradation during printing, while thermogravimetric analysis revealed good thermal stability up to approximately 250–260 °C. The results demonstrate that wear behavior is governed by a combination of hardness, interlayer cohesion, and microstructural organization rather than crystallinity alone. Among the investigated conditions, TPU 51A printed at 265 °C with four layers exhibited the most favorable balance between wear resistance and mechanical properties, highlighting its suitability for friction layer applications.

Figures

Full text

1. Introduction

Pharmaceutical manufacturing operates within a highly regulated environment, demanding unparalleled precision and reliability in all stages of production and packaging [1]. The conveyance and manipulation systems are fundamental elements of this process, where maintaining the highest quality and accuracy is mandatory to prevent product defects or accidental mix-ups, which carry significant economic costs for manufacturers and potentially life-threatening risks for consumers [2]. Consequently, components handling sensitive products—such as blister packs, vials, ampoules, and cartons—must ensure gentle transport and exact positioning, often at high operating speeds [1].

The regulatory framework dictates stringent material standards for all contact components, particularly transmission and conveyor belts. These components must adhere to FDA compliance, Good Manufacturing Practices (GMP), and cleanroom standards [3]. Material safety is paramount; components must be non-toxic, non-leaching, and constructed from FDA-approved polymers, such as specific grades of polyurethane (PU) or PVC [1]. Furthermore, the materials must demonstrate robust chemical resistance to withstand the aggressive cleaning solutions and sterilization processes typically employed in pharmaceutical facilities [3]. Surface integrity is equally critical, as belts must be non-porous and smooth to prevent the deposition of dust, microorganisms, or dirt, simplifying cleaning procedures and minimizing the risk of contamination [3].

Conventional transmission belts rely heavily on frictional contact, and when subjected to the severe, high-cycle demands of pharmaceutical manipulators, they inherently encounter limitations that threaten operational reliability. The repetitive travel-braking cycle imposes immense shear stress on the active friction material, inevitably accelerating wear and deterioration. This operational stress accelerates the wear rate typical of industrial transmission belts [4]. Several primary mechanical failure modes are amplified in this high-cycle dynamic loading environment. Shock loads, often induced by large starting and stopping forces that exceed 10% above operating conditions or severe shock loads during operation, can severely damage the belt’s internal tensile cords, leading to a broken belt or premature failure [4]. Furthermore, continuous operation generates heat build-up; if ventilation is inadequate, this results in rapid material deterioration and substantially reduced belt life [4]. Beyond material failure, improper installation or alignment, such as non-parallel shafts or offset pulleys, causes the belt to track unevenly. This misalignment promotes severe friction material wear, sidewall cracking, edge cord damage, and can even cause the belt to jump off the sheave [4].

The technical necessity for a component that resists cyclic softening, minimizes particle shedding, and provides peak frictional grip mandates a departure from conventional manufacturing processes. The combination of additive manufacturing (AM) and high-performance polymers offers a novel solution.

For performance-critical components like friction belts, AM is instrumental, because it allows for the design and production of complex, non-conventional geometries that are unachievable via conventional machining [5].

Recent developments in additive manufacturing have also focused on the emergence of smart and programmable polymer systems, particularly within the field of 4D printing. These materials are designed to respond to external stimuli such as temperature, magnetic fields, or mechanical loading, enabling shape transformation and functional adaptation over time. Such approaches have been explored for applications in soft robotics, wearable devices, and biomedical systems, where magnetically responsive shape-memory polymers, biodegradable nanocomposites, and multifunctional smart materials have demonstrated improved mechanical performance, shape recovery, and actuation capabilities [6,7,8]. In particular, 4D-printed smart polymers have been investigated for minimally invasive biomedical devices, drug delivery systems, and tissue regeneration, highlighting the expanding role of functional polymer materials in advanced engineering applications [9].

Thermoplastic polyurethane (TPU) is uniquely qualified for pharmaceutical automation components. It offers exceptional physical properties, including impressive flexibility, high elongation, outstanding tensile strength, and cold flexibility [10,11,12,13,14,15]. These attributes are vital for accommodating the dynamic movement and shock loading inherent in the manipulator system. Critically, TPU exhibits high abrasion resistance, making it suitable for high-friction and wear applications [10].

A key advantage for this application is the ability to use fused filament fabrication (FFF) to precisely tune TPU’s frictional response through the control of the internal structure, specifically the infill density and pattern [16]. The study [16] investigates the friction behavior of FFF printed TPU 82A with different gyroid infill densities under dry sliding, and it shows that infill architecture strongly affects friction, ranging from low (0.465) to high (1.002), while all samples exhibit excellent wear resistance with no measurable mass loss. The results demonstrate that gyroid-based 3D printing enables tuning TPU components for sustainable mobility applications.

A similar study [17] reported that harder TPU grades (e.g., SMARTFIL^®^ FLEX 98A) exhibit a decreasing coefficient of friction with increasing infill density, attributed to enhanced structural support and reduced bulk deformation under load. In contrast, softer, foamy TPU grades (e.g., FILAFLEX FOAMY 70A) demonstrated relatively stable friction across infill densities. These results highlight the importance of tailoring internal architecture to achieve the desired balance between friction and wear in 3D printed TPU components.

Surface properties and printing parameters also affect tribological performance [18,19,20,21,22,23]. For instance, the work in [20] examines the effect of 3D printing parameters on TPU surface roughness. Using flow rate, layer thickness, nozzle temperature, print speed, overlap, and fan speed as inputs, the L27 fractional Taguchi method, ANOVA, and regression analyses identified layer thickness as the most influential factor, contributing 65.11% to surface roughness. Confirmation tests validated the results, highlighting that controlling layer thickness is key to optimizing TPU surface quality in FDM 3D printing. The scientific paper [23] aimed to fabricate TPU for journal bearing applications using FDM 3D printing. Cylindrical samples were produced via central composite design to examine the effects of layer thickness, infill density, and printing speed on specific wear rate (SWR), coefficient of friction (COF), and hardness. Layer thickness was the most influential factor (34–72% contribution), while the lowest COF and SWR and highest hardness were achieved at high infill density and low printing speed.

In addition, Sharma et al. [22] analyzed the wear behavior of TPU, ASA, and TPU–ASA multi-material parts fabricated via FDM 3D printing using pin-on-disk tests; a total of 38 experiments based on central composite design evaluated the effects of process parameters on wear. A neural network model, optimized with a genetic algorithm, predicted wear behavior.

Post-processing treatments, such as cryogenic cooling, can further enhance wear resistance by modifying microstructure and increasing hardness [24]. Combined optimization of printing parameters and post-processing is therefore essential to produce TPU parts capable of withstanding mechanical stress.

This study focuses on transmission belts utilized in pharmaceutical manipulators. These elements fulfill a dual role: ensuring the movement of cardboard packaging used for capsule foils and, critically, providing the frictional force necessary to stop the packaging precisely during the insertion of the foils [25]. This function necessitates a high-friction surface for controlled product movement [1]. The resulting operational requirement is a high-cycle demand characterized by a constant, severe sequence of travel followed by instantaneous braking. Achieving superior tribological performance—specifically, a low specific wear rate—is not merely an engineering efficiency goal but a mandatory requirement to maintain GMP and cleanroom compliance.

In previous work of the authors [25], the tribological performance of different friction materials—including the original belt, 3D printed TPU 60A, and TPU 95A—was evaluated under dry conditions. The study demonstrated the influence of material type and infill percentage on coefficient of friction (COF) and wear. TPU 60A achieved the highest COF at full infill, offering superior grip but experiencing substantial wear, whereas TPU 95A exhibited lower COF while maintaining excellent wear resistance. These experiments provided a preliminary assessment of wear behavior under conditions replicating the contact between moving belts and packaging surfaces.

The long-term objective of this research is to extend additive manufacturing from the fabrication of isolated friction specimens to the production of complete transmission belts. The authors’ intention is to apply dual-filament 3D printing to simultaneously deposit two functionally distinct materials within a single manufacturing process: a stiffer material dedicated to the structural support and driving component (toothed base) and a flexible, wear-optimized TPU layer forming the active friction surface investigated in the present study. Such an approach would enable the fabrication of integrated, multi-material belts tailored to specific mechanical and tribological requirements. For components of moderate dimensions but complex geometry—particularly those incorporating toothed or gear-like profiles—3D printing represents a cost-effective and highly adaptable alternative to conventional manufacturing technologies while also allowing rapid prototyping and parameter optimization.

The paper is structured as follows: Section 2 presents the materials, printing parameters, and experimental procedures adopted for the fabrication and characterization of the TPU specimens, including hardness, wear, tensile, XRD, FTIR, thermal, and SEM analyses; Section 3 reports and discusses the experimental results, with emphasis on the influence of printing temperature and number of deposited layers on the mechanical, tribological, structural, and thermal behavior of the materials; Section 4 summarizes the principal findings of the study; and Section 5 highlights limitations and future directions for future research.

2. Materials and Methods

This article examines the transmission belts used in equipment designed for packaging coated tablets into cardboard boxes (Figure 1). The belts function to slow down the cardboard boxes, allowing the insertion of the blister foils. Subsequently, the belt advances the box along the conveyor so that the cycle can be repeated. Such a belt consists of two main components: a base component, which also serves as the driving element, made of a harder polymeric material and having the profile of a toothed belt, typically reinforced with metallic inserts to maintain dimensional stability; and a wear component, applied as a superficial layer over the base component. The present research focused on the wear component, as it was found to be primarily responsible for the belt being taken out of service. All the experiments were performed under normal environmental conditions at room temperature without controlled humidity or temperature regulation.

2.1. 3D Printed Samples

For the experimental study, two TPU type materials were used, namely TPU 51A and TPU 60A, with characteristics presented in Table 1 and Table 2 [26,27].

In order to analyze the influence of printing parameters, two printing temperatures and two layer thicknesses were considered for each material, as shown in Table 3.

The working parameters used for the Artillery Sidewinder X4 Plus 3D printer (New Silk Road Company, Pingxiang, Jiangxi, China) are summarized in Table 4.

Printing soft TPU materials is challenging due to their high flexibility and the tendency of the filament to buckle during feeding. To address this issue, a technical solution developed by the authors was implemented to increase the rigidity of the component assembly involved in the extrusion process (Figure 2). This modification improved the stability of the filament path and reduced the risk of filament deformation during feeding.

The printing was performed using a 0.4 mm nozzle, which provided a stable extrusion flow suitable for the soft TPU grade used in this study. In addition, an anti-adhesive substance was applied to the printing bed to control the adhesion of the flexible material and facilitate the removal of the printed parts after fabrication.

It should also be noted that the original drive mechanism of the printer was maintained without modifications to the gear system. The combination of the increased rigidity of the assembly, the selected nozzle diameter, and the controlled bed adhesion enabled reliable printing of the TPU specimens using a standard 3D printer.

To ensure a controlled and interpretable experimental framework, the present study was intentionally limited to two printing parameters, namely deposition temperature and the number of deposited layers, while all other process parameters were kept constant. Parameters such as infill density, raster orientation, printing speed, and flow rate, which are known to influence the mechanical and tribological performance of additively manufactured polymers, were not varied in this work. This approach allowed the isolated assessment of thermal and layer-dependent effects while avoiding confounding influences arising from an expanded parameter space. The adoption of a broader range of processing variables and larger experimental matrices is beyond the scope of the present study but is necessary to further extend the generality of the observed results.

2.2. Shore A Hardness Measurements

Shore A hardness measurements were carried out on the 3D printed TPU specimens using a Shore A durometer. For each sample, hardness was evaluated at multiple locations on the specimen surface. Specifically, six measurements were taken at different positions on each sample to ensure measurement repeatability and statistical reliability. The reported hardness values correspond to the mean of these measurements.

2.3. Wear Analysis

To evaluate the wear behavior of the investigated materials based on the mass loss of the tested samples a Baroid lubricity tester tribometer (Figure 3) (Ofite, Houston, TX, USA) was employed.

The Baroid tester operates by applying a pressing force to the test specimen, generated through the torque produced by the rotation of a lever arm under a specified load [29,30]. Under these conditions, the specimen made of carton is pressed against a rotating test roller made of the TPU material, thereby creating relative motion between the two contact surfaces (Figure 4). The wear results were reported in gravimetric units. After each testing interval, the contact surfaces were carefully cleaned with a brush to remove loose wear debris, and the mass of the specimen was measured at 10 s intervals using a precision balance. The wear results were therefore reported in gravimetric units, based on the recorded mass loss as a function of time.

The interval of 10 s was chosen in combination with the load, the value being established through successive tests so that the wear would have reasonable, measurable values—within the precision range of the weighing instruments—and, at the same time, not create excessive (destructive) degradation of the materials tested.

2.4. Tensile Testing

The tensile specimens (Figure 5a) were fabricated according to the ASTM D638 Type V standard [31], as shown in Figure 5b.

Eight groups of specimens, each consisting of three samples, were subjected to mechanical testing in accordance with the ASTM D638 [31] standard using an electro-mechanical testing machine equipped with a 2.5 kN load cell, operating at a crosshead speed of 15 mm/min.

2.5. XRD Analysis

For the analysis of the samples by X-ray diffraction (XRD), a D8 advance diffractometer (Bruker AXS GmbH, Karlsruhe, Germany) was used, equipped with a copper anode X-ray tube (Cu-Kα radiation, λ = 1.54 Å), in a θ–θ configuration and Bragg–Brentano geometry. Measurements were performed using the XRD Commander v2.6.1 software over the 2θ range of 5–60°, with the following parameters: voltage 40 kV, anode current 40 mA, step size 0.1°, and scan rate 0.1°/5.

The diffractograms were analyzed by using the Diffracplus EVA v14 software and the PDF-ICDD database. Quantitative determinations using the Rietveld refinement technique were performed with Diffracplus TOPAS 4.1, employing a pseudo-Voigt profile function for peak fitting.

The degree of crystallinity (X_C_) from the XRD spectrum was calculated using the following equation:

$$[eqn]$$

where the area of crystalline peaks is A_C_, and the total area of amorphous and crystalline peaks is A_T_ [32,33,34,35].

2.6. FTIR Analysis

The functional groups of the samples that appeared after the thermochemical treatment were highlighted using a Shimadzu IR TRACER-100 FTIR spectrometer (Kyoto, Japan) recording in the scanning range of 4000–400 cm^−1^.

2.7. Thermal Analysis

Thermogravimetric analyses were conducted using TGA/DTG equipment (TGA 2 Star System Mettler Toledo, Zurich, Switzerland), varying temperatures from 25 to 600 °C at a rate of 10 °C/min in a nitrogen atmosphere. The samples were heated in aluminum trays to assess their temperature stability.

2.8. SEM Analysis

Scanning electron microscopy (SEM) was performed using a Hitachi S-3400N variable-pressure SEM (Hitachi High-Technologies, Tokyo, Japan) equipped with a tungsten thermionic electron source, operating up to 30 kV.

SEM analysis was employed to visually document the wear/fracture surface morphology of the 3D printed rubber specimens, because secondary electron (SE) imaging provides high sensitivity to surface topography and local relief, which is critical for elastomeric surfaces shaped by service loading. In contrast, backscattered electron (BSE) imaging is less informative in this case due to the predominantly organic, low atomic number nature of the material, which limits compositional (Z-contrast) differentiation.

The SEM investigation was performed on non-conductive specimens without conductive coating using surfaces taken from tensile-tested dog-bone samples. Under these conditions, image acquisition is inherently more demanding: obtaining stable, sharp micrographs required extended optimization because of the charging phenomena, potential outgassing/instability typical for elastomeric surfaces, and the pronounced topography of fracture/worn areas. Therefore, SEM observations are presented here as morphological evidence of surface deformation and wear-related features. It should also be noted that SEM provides morphological rather than crystallinity information, and any crystallinity assessment requires complementary techniques.

3. Results and Discussion

3.1. Hardness Evaluation

The mean values of hardness for the analyzed materials are presented in Figure 6.

The Shore A hardness values obtained for the two TPU grades, 51A and 60A, demonstrate a clear dependence on printing temperature and on the number of layers deposited during fabrication. For the 51A material, hardness ranged from approximately 68.5 to 80.5, indicating a pronounced sensitivity to thermal conditions. The highest value was recorded at 255 °C with three layers, while the lowest occurred at 265 °C with four layers. Such behavior is consistent with observations reported in [36], where lower printing temperatures can yield higher hardness values (for example, at 190 °C, the hardness surpasses that measured at 220 °C and 240 °C, which is attributed to the material’s baseline hardness and the reduced thermal softening that occurs when the polymer is processed at a lower temperature).

In contrast, the material TPU 60A displayed values between approximately 65.2 and 72.9 Shore A, with hardness increasing more consistently with temperature and layer count. This indicates an improved interlayer diffusion and reduced porosity at higher printing temperatures. A similar trend was described by Albardawil et al. [37], who analyzed TPU 95A printed at various temperatures and reported hardness variations between 63.5 and 74.9 Shore A depending on printing temperature and layer height, and it was found that as the printing temperature increases, the material hardness rises, while a reduced layer height of 0.1 mm further enhances hardness by improving interlayer bonding.

3.2. Results of Wear Analysis

Figure 7 illustrates the cumulative mass loss as a function of time for specimens manufactured from the 51A and 60A materials.

As shown in Figure 7a, the material 51A exhibits an approximately linear increase in cumulative mass loss with time for all investigated processing conditions, as confirmed by the high coefficients of determination (R^2^ > 0.95). Specimens printed at 255 °C generally present higher wear rates compared to those fabricated at 265 °C, indicating a sensitivity of the softer 51A material to printing temperature. Additionally, samples produced with three layers tend to show slightly higher mass loss than those with four layers, suggesting a reduced wear resistance associated with lower layer counts.

In contrast, Figure 7b shows that the 60A material demonstrates higher overall cumulative mass loss values than 51A, regardless of the printing parameters. The linear trends remain very strong (R^2^ ≈ 0.99), highlighting a stable wear mechanism over time. For this material, both increased printing temperature and higher number of layers contribute to an increase in mass loss.

Figure 8 illustrates the relationship between cumulative mass loss and hardness (Shore A) in samples manufactured under different printing temperatures and layer configurations, highlighting the influence of these parameters on wear behavior.

Samples with lower hardness values (≈63–68 Shore A), such as 60A printed at 225 °C/three layers and 60A printed at 225 °C/four layers, exhibit the highest cumulative mass loss (≈0.07–0.095 g). This indicates reduced resistance to material removal, suggesting that lower printing temperatures combined with increased layer counts lead to weaker interlayer bonding and higher susceptibility to wear.

As hardness increases to intermediate values (≈69–73 Shore A), observed for 60A 235 °C/three layers and 60A printed at 235 °C/four layers, cumulative mass loss decreases slightly but remains significant. This trend suggests that increasing the printing temperature improves material consolidation and hardness, although wear resistance is not yet fully optimized.

The lowest cumulative mass loss (~0.025 g) is observed for the 51A sample printed 265 °C/four layers. However, this sample exhibits relatively low hardness (~68 Shore A) compared to other 51A samples, indicating that reduced wear cannot be attributed to increased hardness in this case.

Samples with the highest hardness values, such as 51A printed at 255 °C/three layers, 51A printed at 265 °C/three layers, and 51A printed at 255 °C/four layers, show low cumulative mass loss (~0.05–0.058 g). This confirms that increased hardness generally contributes to improved wear resistance.

Overall, the results reveal a general inverse relationship between hardness and cumulative mass loss, which is particularly evident when comparing the 60A and 51A materials. However, deviations from this trend confirm that printing temperature and number of layers influence hardness and wear resistance.

In previous work of the authors [25], a conventional pin-on-disk tribometer configuration was used to evaluate polymer–cardboard friction for pharmaceutical packaging applications. The friction pair consisted of a 20 mm diameter cardboard disk and a 4 mm cubic polymer counter body, under dry sliding. This study showed that the original belt material exhibits a high coefficient of friction, with a mean value of 1.182, indicating strong grip on the cardboard. Among the investigated 3D printed materials, TPU 60A at 100% infill provided the highest coefficient of friction, reaching an average value of 1.325. However, this friction performance was accompanied by the highest material wear. In the present work, the Baroid lubricity tester was intended to complement pin-on-disk systems by reproducing—in a form much closer to reality, both in terms of the way the sample that simulates the belt moves and in terms of the geometric shape of the two components of the friction coupling—the actual contact geometry and loading mode of pharmaceutical drive belts. Unlike the pin-on-disk configuration, the Baroid device operates with a rotating TPU specimen against carton under a pressing force, closely replicating the real belt–carton interface during the braking of boxes. The trends obtained with the Baroid tester are consistent with those measured using the pin-on-disk tribometer: TPU 60A compositions delivered the highest friction also experienced the greatest wear, whereas softer grades and optimized processing parameters yielded reduced wear.

In the experiments carried out, the 60A material—whose deposition was performed at temperatures within the range recommended by the manufacturer—exhibits a trend of increasing hardness with both increasing deposition temperature and an increasing number of deposited layers. However, for this material, the morphological transformations do not follow a constant trend. At lower deposition temperatures (225 °C in the present case), an increase in the number of layers leads to an increase in crystallinity, whereas at higher deposition temperatures (235 °C), the opposite behavior is observed.

For this material, the analysis of the mass loss evolution reveals the presence of different wear mechanisms depending on the deposition temperature. Thus, at low deposition temperatures, wear manifests predominantly through brittle exfoliation, while at higher deposition temperatures, it occurs mainly through plastic deformation.

A similar trend in the evolution of crystallinity is also observed for the 51A material; however, in this case, the crystallinity does not change significantly. The deposition of material 51A was carried out at temperatures outside the range recommended by the manufacturer (above the upper recommended limit).

3.3. Tensile Test Results

The values of the physical and mechanical characteristics for materials 51A and 60A are shown in Table 5.

Figure 9 presents the correlation between yield strength, percentage strain, and cumulative mass loss for the investigated samples. For clarity, the discussion is structured in ascending order of cumulative mass loss, highlighting the mechanical response as degradation increases.

The 51A sample printed at 255 °C/three layers exhibits the highest yield strength (~2.86 MPa) and maximum strain (~9.8%) while maintaining a moderate cumulative mass loss (~0.05 g). This indicates that the higher printing temperature combined with a lower number of layers (thicker layers) improves interlayer adhesion and load transfer, resulting in superior mechanical performance and reduced susceptibility to wear.

A similar trend, though less pronounced, is observed for the 51A sample printed at 265 °C/three layers, which shows relatively high yield strength (~2.3 MPa) and strain (~6.2%) at a comparable mass loss level.

Samples printed with four layers (thinner layers), such as 51A printed at 265 °C/four layers and 51A printed at 255 °C/four layers, generally show lower yield strength (~1.0–1.2 MPa) and reduced ductility, even at lower mass loss values.

For the 60A material series, an overall increase in cumulative mass loss is observed, particularly for 60A printed at 235 °C/three layers and 60A printed at 225 °C/four layers, where mass loss exceeds 0.08–0.09 g. These samples exhibit marked reductions in both yield strength (<0.3 MPa) and strain (<3.3%), confirming that higher wear-induced material loss directly compromises tensile performance. The effect is exacerbated in samples with higher layer counts, emphasizing the role of layer thickness and interfacial quality.

The relatively low yield strength values obtained are characteristic of elastomeric TPU materials (similar to work [38], where for pure TPU the yield stress obtained was about 1.5–4 MPa) and are strongly influenced by interlayer bonding rather than bulk crystallinity. Samples printed with fewer layers exhibit higher yield strength and strain, indicating improved interlayer cohesion and more efficient load transfer between deposited filaments. This mechanical response correlates with wear behavior, as specimens with higher ductility tend to undergo deformation-controlled wear, while lower yield strength is associated with brittle exfoliation and increased mass loss. XRD results confirm that all printed samples remain predominantly amorphous, indicating that tensile and wear performance are governed mainly by thermal history and interlayer diffusion rather than crystallinity variations.

3.4. XRD Results

The X-ray diffraction spectra of TPU samples (shown in Figure 10) show a broad peak (in the region of 2θ 15° to 25°) centered on 2θ = 20°, which confirms the predominantly amorphous nature of TPU. The existence/presence of peaks with more or less pronounced intensity, centered around 2θ values of 11°; 21.3°; 23.7°; 27.4°; and 29.2°, are characteristic of a semicrystalline TPU.

The individual peaks mentioned above characterize domains of higher crystallinity corresponding to the microphase structure of TPU, consisting of crystalline (ordered) and amorphous (disordered) domains, respectively. These peaks probably correspond to an ordered arrangement of domains of crystallized material that are surrounded by amorphous material as a matrix [39,40]. It can be observed that in samples that have pronounced diffraction peaks in the range 2θ = 11° to 12° the degree of crystallinity increases, indicating a higher concentration of crystallized areas in agreement with previous research [41,42,43].

The results related to degree of crystallinity are presented in Figure 11.

The phase identification was performed with the Diffracplus Basic software version 14, the Search/Match option, and the ICDD database after eliminating the background and Kα2 radiation. The quantitative determinations were performed with the Diffracplus TOPAS 4.1 software using the Rietveld method. The pseudo-Voigt profile function was used to fit the peaks. In XRD, the quality of Rietveld refinement is expressed through R-values: GOF and DW.

GOF (goodness-of-fit) represents the ratio between Rwp (R-weighted pattern) and Rexp (R-expected) and cannot be less than 1. A Rietveld refinement that can be considered acceptable will give GOF values lower than 2 (usually even lower than 1.5). The DW index provides the Durbin–Watson “d” statistic and is used to detect serial correlation of the residual errors of the model, helping to ensure that the model assumptions are met for reliable predictions, especially in time series data. In X-ray diffraction, the ideal value of the DW index is 2. A value as close to 2 as possible means that there is no serial correlation.

Table 6 presents the R-values of the Rietveld refinement—GOF and DW— and also the crystallite size (cry size) and the mean error values. The low values of the figures of merit (R-values) of the Rietveld refinement (GOF and DW) indicate a good sensitivity to structural changes.

In general, the diffraction profiles show: (i) peak positions and relative intensities consistent with JCPDS standards; (ii) the absence of foreign phases or precursors; and (iii) significant peak broadening consistent with nanoscale crystallite domains in the polymer structures. In this context, the crystallite sizes estimated by the Rietveld technique should be considered credible, correct, and representative values, because the contributions of factors such as instrumental broadening, peak shift/zero point error, Lorenz-polarization factor, and microstrain to the FWHM (full width at half maximum) of the diffraction peaks have been taken into account by choosing the Full Axial Model and the “Fundamental parameters” approach, respectively, when building the TPU diffraction model. It is also well known that broad peaks (as is the case of the diffraction spectra in Figure 10) indicate reduced crystalline domains and a higher degree of disorder.

In case of polymeric structures, the low atomic number of carbon reduces the scattering of X-rays, a situation minimized by a low scanning speed (0.1°/5 s) and a longer recording time achieved by repeated scans within the same recording session.

The crystallinity values calculated by XRD are similar to those obtained on different TPU samples in previous works [43,44] and indicate the semi-crystalline nature of the TPU samples analyzed.

The crystallinity values for both the 51A sample set and the 60A sample set, reported relative to the raw material samples—51A filament and 60A filament, respectively—show a trend of decreasing crystallinity after printing the samples. In case of both TPU sample sets (51A and 60A), increasing the number of printed layers achieves approximately the same effect as increasing the printing temperature.

The intense and well-expressed peaks in the X-ray diffraction spectra of the primary material samples—51A filament and 60A filament, respectively—indicate a high degree of ordering and concentration of the crystallized material areas confirmed by the high crystallinity values calculated (Figure 11).

The decrease in the degree of crystallinity of the samples after printing indicates an increase in the proportion of amorphous domains at the expense of crystallized domains, which provided more freedom for the rearrangement/orientation of the harder crystallized regions while improving the mechanical qualities of the material.

In the structure of a polymer such as TPU, two phases generally coexist: crystalline and amorphous, respectively, and these domains have a complex organization. For practical reasons, previous research has shown that the most appropriate is a two-phase model that describes semi-crystalline polymers in terms of two phases, an average amorphous phase and an average crystalline phase. As polymer is cooled from the melt, some of the polymer chains crystallize into small lamellae or crystallites (spheres), while the other part remains amorphous as the melt solidifies [45].

Crystallinity or the degree of crystallinity is an important parameter in the two-phase model, being closely related to the mechanical properties of the polymer. In general, a high degree of crystallinity improves the wear resistance and hardness of polymers, because ordered crystalline domains resist plastic deformation better than amorphous domains. More compact structures are characterized by a higher crystallinity, leading to an increase in surface hardness and modulus of elasticity, which makes deformation (plastic) more difficult. The ordered crystalline structure, most often composed of folded polymer chains (with a lamellar appearance), offers a better resistance to plastic deformation, which also influences the resistance to wear, abrasion, and fatigue. On the other hand, the alternating arrangement of crystalline and amorphous domains predisposes to their relative sliding, facilitating wear [46].

Polymers with high crystallinity become tougher and more durable, reducing material loss due to mechanisms such as brittle spalling, abrasion, adhesion, and fatigue. Also, stronger intermolecular forces within the crystalline domains reduce polymer chain slippage. Increasing crystallinity is usually accompanied by an increase in the degree of ordering of the crystalline domains and an increase in crystallite size. However, in some circumstances, even in polymers subjected to annealing treatments, increasing crystallite size does not necessarily lead to an increase in overall crystallinity [47,48,49].

The decrease in crystallinity of the printed samples compared to the crystallinity of the raw material (filament samples) is explained by the fact that during casting (or printing), there is a rearrangement of new crystallites, as well as ordered or quasi-ordered precursors, which coexist in the amorphous mass. In this case, within the semi-crystalline polymer arrangement, a certain type of second-order crystalline disorder appears, determined by the appearance of a transition phase (small, quasi-ordered crystallites), which coexists with the ordered, stable crystals and the amorphous phase, within a limited range of crystallinity values [47,48,49].

3.5. FTIR Analysis Results

All samples exhibit the characteristic absorption bands of TPU 51A, indicating that the polymer backbone is preserved after processing (Figure 12).

The broad absorption band at 3320 cm^−1^ is assigned to N–H stretching vibrations of urethane groups involved in hydrogen bonding, typical of segmented TPU [50]. The bands at 2900 and 2800 cm^−1^ correspond to the asymmetric and symmetric stretching vibrations of aliphatic C–H groups [51]. The strong absorption band observed in the carbonyl region (1725–1705 cm^−1^) is attributed to the C=O stretching vibration of urethane groups. The presence of a lower wavenumber contribution (~1705 cm^−1^) indicates hydrogen-bonded carbonyl groups, reflecting microphase separation between hard and soft segments [52]. The band at ~1530 cm^−1^ is assigned to the amide II vibration (N–H bending coupled with C–N stretching), while the absorption at the 1220 cm^−1^ corresponds to amide III and C–N stretching vibrations of the urethane linkage [51,53]. The absorption band at 1100–1080 cm^−1^ is assigned to C–O–C stretching vibrations, confirming the presence of ether groups in the soft segments and indicating a polyether-based TPU [54,55]. Weak bands in the 770–820 cm^−1^ region are attributed to out-of-plane aromatic ring vibrations, suggesting the use of an aromatic diisocyanate [55,56], but the peak at 800 cm^−1^ may show amorphous silica of the Si–O–Si type [57]. The absence of a band at ~2270 cm^−1^ confirms the lack of unreacted isocyanate groups and indicates that no significant chemical degradation occurred during processing [53,58].

Additionally, the FTIR spectra of the 60A material, as well as of the belts manufactured from it, are presented in Figure 13.

The FTIR spectrum of segmented TPU exhibits a broad absorption band at approximately 3320 cm^−1^, which is assigned to N–H stretching vibrations of urethane groups involved in hydrogen bonding [59]. The urethane carbonyl (C=O) stretching and amide II vibrations appear at 1720–1730 cm^−1^ and 1530–1560 cm^−1^, respectively, while the C–O/C–O–C stretching vibrations are observed at 1217 cm^−1^ [60]. The absence of an absorption band at 2270 cm^−1^ confirms the complete consumption of NCO groups and indicates the lack of significant chemical degradation during processing [61].

3.6. Thermal Analysis Results

High temperatures lead polymers to undergo degradation processes like depolymerisation, side group removal, and backbone scission. These processes significantly alter the chemical structure of polymers, increasing brittleness and changing their physical characteristics [62]. To determine the temperature ranges at which the samples can be operated without affecting their structural integrity, the thermogravimetric method was used. The thermogravimetric curves are presented in Figure 14, and it was observed that all 51A materials have T_5%_ up to 260 °C, whereas the T_5%_ values of the 60A materials are around 250 °C, indicating a very good thermal stability.

In the case of both the 60A filament and the 3D printed samples, the degradation stages consist of two major stages between 240 and 390 °C and 390 and 500 °C (Figure 15). As described in the literature, this behavior is specific to polyurethanes that usually decompose in two main stages, the first stage being due to the decomposition of the urethane groups in the main chain and the release of volatile compounds [63] and the second is attributed to the pyrolysis of condensed polyol (product of the first stage), which leads to C-C and C-O bonds breakage [64]. Although this occurs in all 60A samples, some differences were observed depending on the printing temperature and the number of layers. For example the samples printed at 225 °C in three and four layers exhibited in DTG curves a degradation trend similar to that of 60A filament, showing two degradation peaks around 360 °C and 414 °C. On the other hand, the samples printed at 235 °C in three and four layers present a small a shoulder between 374 and 390 °C, caused probably by higher printing temperature and producing an overlapping reaction of urethane dissociation from the polyol and isocyanate [65].

The 51A filament and the 3D printed samples also show, in DTG curves, two degradation stages between 250 and 380 °C and 380 and 500 °C, which are characteristic of polyurethanes’ thermal degradation. Calvo-Correas et al. [66] attribute the first degradation stage centered around 340 °C to the thermal decomposition of urethane bonds (loss of hard segment—HS) and the second, centered around 412 °C, to the thermal decomposition of the soft segment (SS).

Figure 16 shows the DSC curves of the samples. As can be seen, all the samples exhibit no crystallization exothermic peaks indicating that they are amorphous polymers [67]. However, both 51A and 60A filaments as well as 3D printed samples present a broad endothermic peak with small shoulders between 300 and 350 °C. According to Saiani et al. [68], this behavior corresponds to the melting of an ordered structure from the hard segment and to a microphase containing a mixture of soft and hard segments [69,70].

These observations are in correlation with the TGA data, which indicated the presence of the two types of segments. The DSC analysis also indicates that the melting temperature (T_m_) values for 3D printed samples are slightly higher than filament and increases as the number of layers increases. This phenomenon may be attributed to the growth of hard-segment crystallites in TPU, which increases as the number of layers increases [71]. For example, the melting temperature values for 60A 3D printed samples printed at 225 °C/three layers and at 235 °C/three layers are 331.22 °C and 332.23 °C, respectively. Meanwhile, the melting temperature of the samples printed at 225 °C/four layers and 235 °C/four layers are 333.13 and 333.17 °C, respectively.

The same behavior was observed for the melting temperature values for 51A 3D printed samples. In case of samples printed at 255 °C/three layers and 265 °C/three layers, the T_m_ is 326.50 °C and 326.88 °C, respectively, while the melting temperatures of the samples printed at 255 °C/four layers and 265 °C/four layers are 327.29 °C and 327.55 °C, respectively. It is important to note that T_m_ for 60A is 329.81 °C, while T_m_ for 51A is 325.87 °C.

In conclusion, the DSC results indicate that the sample’s usage temperature should not exceed 300 °C to preserve its structural integrity.

3.7. SEM Analysis Results

SEM observations of the worn surfaces (Figure 17) indicate pronounced surface deformation induced by the specific loading conditions experienced during belt operation. The surface morphology reveals flattened and elongated features along the loading direction, suggesting progressive plastic deformation and stretching of the printed layers under cyclic mechanical stresses.

This behavior is consistent with the working conditions of the belts, where repeated tensile and bending loads promote gradual elongation of the material. As a consequence, the effective thickness of the printed layers appears locally reduced compared to the nominal layer thickness (0.2 mm), indicating material flow and strain accumulation during service.

The observed morphological features are therefore attributed to deformation-driven wear mechanisms rather than purely abrasive removal. Such deformation and layer thinning are typical for elastomeric materials subjected to prolonged mechanical loading and confirm the strong influence of operational conditions on the wear morphology.

Figure 18a shows a heterogeneous surface characterized by flattened regions, elongated features, and locally smeared material aligned with the main loading direction.

SEM images of the belt edge (Figure 18b) reveal large surface discontinuities and pronounced height differences, indicating heterogeneous surface deformation under service loading. A distinct surface pull-off feature is also visible, consistent with localized overstressing and subsequent material detachment in the most loaded edge zones.

The SEM images from Figure 18c reveal the presence of distinct surface regions with clearly different morphological characteristics, indicating a form of surface segregation induced by operational loading. Adjacent areas exhibit contrasting topography, with compact, flattened zones alternating with regions showing cavities, detachment features, and uneven surface levels.

In case of printing parameters such as 265 °C/four layers (Figure 18d), a localized surface rupture can be observed, characterized by a distinct material detachment from the surrounding matrix. The affected area exhibits a cavity-like feature with irregular edges and stretched material remnants.

Based on direct measurements performed on the SEM micrographs, representative void diameters are in the range of approximately 2.39–3.02 µm, and locally significantly larger cavities with characteristic dimensions of up to ~27.2 µm are also observed (Figure 19a). The size and distribution of these voids suggest localized material discontinuities, which are relevant for understanding surface degradation and wear-related damage during operation.

SEM observations (Figure 19b) reveal localized deformations at the edge regions of the belt, characterized by distorted surface geometry and irregular, rounded margins. These features indicate that certain zones experienced irreversible deformation under service loading.

Image 19c shows the presence of large voids/cavities together with a distinctly non-coplanar surface, where adjacent regions appear at different height levels. This stepped morphology indicates that the surface will not wear uniformly but instead can present localized deformation and material redistribution under service conditions.

At the belt edge (Figure 19d), the SEM images reveal localized surface distortion and uneven topography, indicating that the edge region experienced concentrated loading during operation. Such edge areas are prone to irreversible deformation due to the combined action of tensile, bending, and contact stresses, which promotes non-uniform surface modification.

3.8. Overall Discussion of Experimental Results

In 51A (Figure 20), the increase in hardness is associated with an increase in the yield strength value, which leads to the idea of the appearance of a structural “hardening”. Also, the increase in hardness leads to an increase in wear, possibly as a result of the development of more fragile structures.

In correlation with Figure 18, it can be observed that the structures characterized by a large volume of voids (255 °C/three layers and 265 °C/three layers) are prone to an accentuated non-uniform wear, promoted by the tearing of material that generally exhibits a hardness higher than that of the other samples. The strongly distorted structure of the sample printed at 255 °C/four layers results in non-uniform behavior under load and therefore promotes the tearing of particles from the belt composition. This explains the highest wear value (sample 255 °C/four layers). Sample printed at 265 °C/four layers presents a structure with a lower yield strength and hardness than that of other samples. The low elongation value justifies the material discontinuities found in the SEM examination, in the sense that, in this case, the structure has the availability to retain the matrix under load without substantial particle loss (low wear) until the load-bearing capacity is reached when structural damage occurs. In correlation with Figure 11a, it can be observed that a high degree of crystallinity is unfavorable for the wear behavior of the 51A material.

As for 60A (Figure 21), the deposition parameters did not lead to very large differences in terms of hardness. In association with this finding, a random variation in the cumulative mass loss and the yield strength is observed. However, an inversely proportional variation in wear relative to plasticity characteristics is highlighted. Correlating the evolution of the characteristics in the graphs with the SEM images, it can be stated that variants printed at 225 °C/three layers and 235 °C/three layers, which present voids in the structure, benefit under load from a more efficient redistribution of the material, which favors better wear behavior. Unlike the 51A material, in this case, the deformation capacity—demonstrated by high elongation values—generates superior wear resistance. In contrast, the four layer structures (Figure 19b,d) show non-uniform changes on the surface, which favors material detachment and greater wear.

Obviously, the two situations benefit from different assumptions, as long as the deposition temperatures for 51A were experienced outside the range proposed by the manufacturer and 60A falls within the range proposed by the manufacturer.

4. Conclusions

The experimental investigation of 3D printed TPU 51A and TPU 60A revealed that printing temperature and number of deposited layers significantly influence hardness, wear resistance, tensile behavior, crystallinity, and thermal stability.

In the 51A material, increasing the number of layers generally reduced hardness. At 255 °C, this was accompanied by a slight increase in wear, whereas at 265 °C, the wear rate decreased considerably, indicating improved tribological performance under the higher temperature–four-layer conditions. In contrast, in the 60A material, increasing the number of layers led to higher hardness but also to a slight increase in wear.

No direct correlation between yield strength and wear rate was identified. The studied materials exhibited a predominantly amorphous structure, and although crystallinity showed a general influence on wear behavior, this relationship was not strictly proportional. An exception was observed in the 60A material printed at 235 °C, where increased wear occurred despite a decrease in crystallinity, indicating that additional structural factors affect wear response.

Thermal analyses confirmed that the investigated printing temperatures did not induce significant structural degradation, and both materials maintained good thermal stability. Deposition generally led to a reduction in crystallinity compared to the raw filaments.

SEM observations indicated that different wear mechanisms dominate depending on the processing conditions, while XRD and DSC analyses confirmed the predominantly amorphous nature of the printed TPU and the role of microphase organization.

Considering the overall tribological performance, TPU 51A printed at 265 °C with four layers provided the most favorable results. The study also highlights that the technological behavior of polymer materials is strongly dependent on processing parameters, and each material–process combination requires individual evaluation.

5. Limitations and Future Work

Although the present study demonstrates the potential of FDM printed TPU for friction layers in pharmaceutical transmission belts, several limitations should be acknowledged.

The experimental investigation was restricted to a narrow range of printing parameters (two temperatures and two layer counts per material) and did not explore additional variables such as infill density, infill pattern, and printing speed, which are known to significantly influence mechanical, tribological, and surface properties of FDM printed TPU parts.

Future research should expand the parameter space by systematically investigating the effects of infill density, pattern, printing speed, and raster orientation on wear behavior. Also, post-processing heat treatments (annealing) could be explored to enhance interlayer bonding, crystallinity control, and overall wear resistance while preserving flexibility.

The relevance of the present results to pharmaceutical transmission belts should be interpreted within the defined experimental scope. The tribological tests employed in this study were designed to reproduce the local contact conditions between the belt friction layer and cardboard packaging, particularly in terms of contact pressure, sliding interaction, and material pairing. However, full-scale belt testing under cyclic braking conditions, long-term fatigue loading, and contamination or particle shedding assessments were not performed. Consequently, the application to pharmaceutical transmission belts is proposed as a technologically motivated use case rather than a fully validated operational demonstration. Further investigations involving real belt prototypes and cyclic braking simulations are required to comprehensively validate the suitability of FDM printed TPU components for pharmaceutical production environments.