1. Introduction
Tissue engineering is a promising tool to solve the issue of organ shortage due to supply–demand imbalance, and the goal is to assemble functional constructs that restore, maintain, or improve damaged tissues or whole organs [
1,
2]. Nevertheless, traditional tissue engineering strategies encounter challenges in the construction of multicellular complex tissues and even personalized structures [
3]. Bioprinting is an advanced tissue engineering approach, which utilizes cells, proteins, and biomaterials as raw materials, i.e., bioink, for the additive manufacturing of biological products [
4,
5,
6].
The major challenge faced by bioprinting technology is how to dispense bioink into discrete units through a material dispensing system (usually called a printhead). Therefore, bioprinting is normally categorized by dispensing technology. Extrusion-based bioprinting (EBB) is one of the most popular bioprinting technologies currently available, where bioink located in a reservoir is pushed by a driving mechanism and extruded as continuous filaments via a printhead nozzle [
7]. EBB technology is widely used mainly because it is compatible with most bioinks with viscosities ranging from 30 mPa·s to 6 × 10
7 mPa·s at a reasonable cost [
8].
In general, there are three different driving mechanisms of EBB printheads: piston-driven, pneumatic-driven, and leadscrew-driven. Since a leadscrew-driven printhead usually has relatively lower cell viability due to the larger pressure drop inside reservoir [
9], piston- and pneumatic-driven printheads are commonly used for higher cell viability (90%) [
7,
8]. Theoretically, a piston-driven printhead has even better printing precision because the output flow volume of bioink is dependent on the feed rate of the piston [
7,
10]. In contrast, in practice, trial-and-error methods are used to achieve high precision and reproductivity [
11], since the control is more complex than a pneumatic-driven printhead. One of the major reasons here is the non-Newtonian behavior of bioink, which leads to complexity. The ideal material properties of bioink are high mechanical strength with shear-thinning behavior, as well as great biocompatibility, which ensure an excellent balance of printability, structural support, and cell proliferation [
12,
13]. To be specific, to achieve higher printing resolution, bioink needs to improve its viscoelasticity to maintain the mechanical properties of the layer-by-layer structure. However, by doing this, embedded cells will experience an increase due to higher shear stress during the printing process, resulting in poor cell viability. On the other hand, bioink with lower viscosity will protect resident cells from shear stress during extrusion, but it will have unsatisfied structural fidelity [
14].
To solve the problem, large amounts of work have been carried out to develop novel bioink and crosslinking strategies. Guo developed a double-network hydrogel composed of hyaluronic acid and alginate, which featured both great injectability and mechanical strength for self-supporting bioprinting [
15]. Ouyang introduced a new in situ crosslinking method to simultaneously crosslink the photo-sensitive hydrogels during extrusion [
16]. Nevertheless, even with newly developed bioink, it will take a lot of effort to find out the proper parameters and factors required to achieve satisfactory printing results and cell viability, not to mention the fact that some of the bioinks are expensive and difficult to synthesize. Recently, the finite element method (FEM) has been used to study the printing process to help researchers gain a better understanding of the relationship among bioink properties, printing parameters, and printhead design. Juan applied the finite element method (FEM) to study the extrusion bioprinting process with different nozzle tip structures [
17], as well as bioink temperature’s influence on shear stress, pressure, and velocity during bioprinting [
18]. Hamed investigated the impacts of different velocities and viscosities on extrusion-based chitosan 3D printing [
19]. While most of the simulation studies focused on investigating different printing parameters or specific materials, few of them thrived when optimizing bioprinter structure design based on simulation studies to adapt emerging materials.
In this study, we advance the field of tissue engineering by optimizing the design of a bioprinting printhead with enhanced temperature control, crucial for maintaining bioink integrity during printing. Our innovative approach integrates thermal and fluid dynamics analyses with FEM simulations, leading to a novel printhead design featuring thermal insulators that stabilize temperature-sensitive bioinks. This development not only improves the structural fidelity and cellular viability of printed tissues but also joins theoretical predictions with practical implementations, setting new benchmarks for the use of printhead technology in bioprinting. Our findings offer substantial advancements in bioprinting precision and efficiency, significantly contributing to regenerative medicine and organ fabrication.
2. Materials and Methods
2.1. Piston-Driven Extrusion-Based Bioprinting (EBB) Printhead
In this research, a custom-build piston-driven EBB printhead was designed and built for a bioprinting system. The printhead consisted of a linear motion system, a bioink loading unit, and a temperature-controlling system, which were mainly made of aluminum alloy. A stepper motor and leadscrew-based mechanism was used for the linear motion system, and the bioink loading unit was connected to the linear motion system. A 3 mL disposable syringe (Becton Dickinson, Franklin Lakes, NJ, USA) was designed to insert into the loading unit, and a pressing unit driven by the stepper motor was mounted to exert linear force on the thumb rest of the syringe plunger during printing. Additionally, a polyimide heater film (10 W) was adhered around a copper sleeve inside the printhead to serve as heat source for the printhead, and a temperature sensor (Pt100, Allymore, Shenzhen, China) was also installed to monitor the real-time temperature of the printhead.
2.2. Material Preparation and Rheological Analysis
Since both thermal and fluid analysis were performed in this study, a temperature-dependent material was used as the bioink. To prepare the material, UV-sterilized gelatin and alginate powders were evenly mixed and dissolved in 0.9% NaCl solution (m/v) at 10% and 1% (m/v), respectively. The solution was then sterilized by heating three times in an oven (70 °C) for 30 min.
Since the bioink used in this study exhibited shear-thinning behavior. Rotational and oscillatory tests were performed using an Anton Paar MCR 302 Rheometer (Anton Paar, Ostfildern, Germany) to investigate the rheological properties and the proper gel point. The viscosity parameters were measured at 15 °C, 30 °C, and 37 °C with a shear rate of 0.1–100 s−1. The measurement parameters of the oscillatory test used were a deformation range of 0.1%, a cone plate diameter of 25 mm, and a cone angle of 2°, and the distance between the cone tip and the base plate was 99 µm. The elasticity of the bioink was represented by the storage modulus G′, and the viscosity was represented by the loss modulus G″. G′ and G″ were calculated through a temperature sweep at a frequency of 1 Hz. Before measurement, all samples were heated in a water bath at 37 °C for 5 min to ensure that the material was in a sol state, with an initial temperature of 37 °C. The sol–gel temperature transition point test was completed through temperature scanning tests, with the samples being cooled from 37 °C to 20 °C at a rate of 1 °C per minute. Three parallel experiments were conducted on each experiment for reproductivity.
2.3. Cell Culture and Cell-Laden Structure Bioprinting
The Primary Human Umbilical Vein Endothelial Cells (HUVECs) were obtained from the Cell Resource Center, IBMS, CAMS/PUMC (Beijing, China) and cultured in high-glucose Dulbecco’s modified Eagle medium (H-DMEM; Invitrogen, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS; Gibco, Waltham, MA, USA) under conditions of 37 °C and 5% CO2. The cells were passaged by 0.25% trypsin (Invitrogen) when the cell confluency reached about 90%. The culture media were changed every 2–3 days.
To prepare cell-laden bioink for bioprinting, the cells were first harvested by centrifugation at 1000 rpm for 5 min and suspended in cell culture medium to a density of 4 × 106 cells/mL. Then, the HUVECs, 20% gelatin, and 4% sodium alginate were mixed at a volume ratio of 1:2:1 to obtain the bioink for bioprinting experiment. The final bioink consisted of 10% gelatin, 1% sodium alginate, and a cell density of 1 × 106 cells/mL.
During bioprinting process, the bioink was loaded into a syringe to print a grid-like structure with geometric dimensions of 10 mm × 10 mm × 4 mm. After printing, the structure was crosslinked with 3% CaCl2 solution for 5 min. Following crosslinking, the structure was gently washed three times with phosphate-buffered saline (PBS) solution and then cultured in DMEM medium.
2.4. Cell Viability Analysis
In addition, printed cell-laden 3D structures were immediately assessed for cell survival conditions using a live/dead staining method. Specifically, the printed structures were washed three times with PBS, then incubated with a mixture of 1 μmol/L Calcein-AM (Dojindo, Kumamoto, Japan) and 2 μmol/L PI (Dojindo) at 37 °C in the dark room for 30 min. After incubation, the samples were gently washed three times with PBS and observed under a confocal microscope (C2+, Nikon, Tokyo, Japan). Three independent samples were set up as replicates, with three fields of view chosen randomly from each sample.
Cell proliferation was analyzed using the cell-counting kit-8 (CCK-8) assay kit (Dojindo) on days 1, 3 and 7. Specifically, each group of samples was washed three times with PBS, and then a mixture of 1 mL H-DMEM and 0.1 mL CCK-8 solution was added. The samples were incubated at 37 °C for 2 h. After incubation, the culture medium was transferred to a 96-well plate, and the optical density (OD) values were read at 450 nm using a microplate reader. Each group had three independent repeats at each time point.
2.5. Finite Element Analysis
The FE simulation was conducted in COMSOL Multiphysics 6.2 (COMSOL, Burlington, MA, USA) to study the temperature field of the printhead and the extrusion process of the hydrogel via the piston-driven printhead. To accelerate the simulation efficiency, simplified 2D axis-symmetric models were generated for both thermal and flow analysis based on 3D model of the lower-part printhead, as shown in
Figure 1. Since each analysis focused on different target outcomes, the geometrical features were different, but both models shared the same critical geometries. To be specific, models of thermal analysis consisted of bioink, a syringe (with nozzle tip), a heating sleeve, and a printhead structure, while models of flow analysis consisted of bioink and air. The dimensions of syringe and nozzle were considered critical geometries for both simulations. In addition, two kinds of conventional nozzle tips (both were 250 µm), a 1/2-inch stainless steel (SS) tip (
Figure 1a) and tapered tip (TT) (
Figure 1b), were applied in the simulation study.
2.5.1. Thermal Analysis of Printhead
Generally, heat transfer has three major mechanisms: thermal conduction, thermal convection, and thermal radiation. During bioprinting with gelatin-alginate bioink, the printhead was usually heated to 37 °C, which is a conventional heating temperature for gelatin-alginate based bioink with living cells. The bioink was around 4 °C before printing (as it was previously stored at fridge). The atmosphere temperature was 15 °C, which was the environment setup in the sterile cabinet at the biosafety laboratory. In this scenario, thermal conduction and convection play dominant roles in the temperature distribution of the printhead, while thermal radiation can be neglected.
Heat convection happened both inside and outside of the printhead, between the syringe and the bioink, and between the external surface of the printhead and the ambient air. The boundary condition could be described as follows:
where
is the normal vector of the boundary,
is the heat transfer coefficient (W/(m
2·K)), and
is the temperature of the surrounding medium (K).
The heat transfer coefficient inside the printhead could be assumed to be a constant value of 15 W/(m2·K). The bioink’s velocity was assumed constant, and its temperature decreased slightly.
The heat transfer coefficient with air was assumed to be a free-convection process and could be described as follows [
20]:
where
is the thermal conductivity of air,
is the typical length, and
is an empirical coefficient as a function of the incidence angle
;
is the Grashof number.
In order to study both the temperature distribution and heat transfer velocity, static and transient simulations were performed in the thermal analysis. The parameters selected for the simulation can be seen in
Table 1.
2.5.2. Flow Analysis of Extrusion Process
To investigate how temperature distribution could affect the extrusion process, a flow analysis was conducted. Since the bioink was treated as an incompressible, laminar fluid, the momentum and continuity equations by COMSOL were as follows:
where
is the density (kg/m
3),
is the velocity vector (m/s),
is the pressure (Pa),
is the dynamic viscosity (Pa·s),
is the absolute temperature (K),
is the unit tensor, and
is the surface tension force (N/m
3). In this study, the densities of both air and bioink were set as
kg/m
3.
The surface tension force can be calculated as follows:
where
is the surface tension coefficient (N/m),
is a Dirac delta function that is nonzero only at the fluid interface,
is the curvature, and
is the interface normal. The surface tension coefficient was set to 0.7 N/m, and all other functions could be calculated as follows:
Since the bioink in this research was a shear-thinning fluid, it was described by the power law (Ostwald de Waele) model to demonstrate non-Newtonian flow behavior, as shown in Equation (10):
where
is the apparent viscosity (Pa·s),
is the shear stress (Pa),
is the flow consistency factor (Pa·s
n),
is the shear rate (s
−1), and
is the flow behavior index [
21]. To find those values, Equation was transformed into logarithm form:
and values were derived from rheological data by applying linear regression fitting with the logarithm data of related shear rate and viscosity. For shear-thinning fluid, the viscosity decreased with increased shear rate, which meant that was less than 1.
In this study, two-phase flow was analyzed. To solve the two-phase flow, the level-set method was used, and the transport equation of the two-phase interface could be demonstrated as follows:
where
is the level-set function of the gas–liquid interface,
is the reinitialization parameter (m/s) used to solve the equation,
is the controlling interface thickness (m), and
is the flow velocity (m/s).
In this study, the 0.5 contour of the level-set function defines the interface, where equals 0 in air and 1 in bioink. Since determined the amount of reinitialization, a suitable value of γ was the maximum magnitude occurring in the velocity field. We used , where is the maximum mesh size.
In addition, since the density and viscosity were changing in the two-phase flow, the smoothing process was applied with the use of level-set function
:
where
is the level-set function,
is the density of air,
is the density of the bioink,
is the viscosity of air, and
is the viscosity of the bioink.
The geometrical features used for flow analysis were measured and calculated based on the homemade printhead with a 3 mL syringe, and the dimensions are listed in
Table 2.
Boundary conditions are demonstrated in
Figure 1. The inlet was on the top of the syringe and the flow rate was at a constant speed of 10 mm
3/s, which derived from conventional extrusion flow rate. The outlet was the nozzle tip with 101,325 Pa, which was 1 atm. All other boundaries were considered to be under the “no slip” condition.
4. Discussion
In this investigation, we conducted a comprehensive analysis of the influence of temperature and nozzle types on the morphological characteristics of the filaments, as well as on the pressure and shear stress fields during the bioprinting process. Our flow analysis highlighted the critical role of bioink behavior at the nozzle, emphasizing the necessity of achieving an optimal gelation state to enhance the quality of bioprinted structures. However, our thermal assessments revealed the presence of a temperature gradient across the printhead, with a noticeable reduction in temperature at the nozzle tip. Given the profound effect of temperature on the fidelity of prints made with thermally sensitive bioinks, addressing this issue becomes imperative. To this end, we devised a hardware optimization strategy with the addition of thermal insulators on the printhead within this study.
The design of thermal insulators for the nozzle tip (shown in
Figure 7) was driven by the following considerations: Initially, simulations indicated the nozzle’s susceptibility to temperature fluctuations, which led to significant temperature gradients affecting the extrusion process. To counteract this, we employed copper for the insulator due to its excellent thermal conductivity, which facilitates the maintenance of a stable temperature at the nozzle tip. This modification ensured a uniform temperature distribution within the bioink, resulting in more consistent extrusion outcomes.
Furthermore, the standard practice of utilizing transparent materials for syringe nozzle tips poses a challenge when working with photo-sensitive bioinks, as it exposes the bioink to potential premature photo-crosslinking due to light exposure. By adopting opaque materials for the insulator tips, we mitigate this risk, thereby preserving the bioink’s integrity until the desired moment of crosslinking.
Additionally, the constrained spatial environment encountered when printing onto well plates necessitates careful consideration of the insulators’ geometries to avoid physical interference. To accommodate various nozzle configurations, we designed two distinct types of insulators tailored for tapered tips and for 1/2 inch stainless steel tips, respectively.
To rigorously assess the impact of thermal insulators on temperature management during the bioprinting process, we employed both thermal simulation and empirical printing experiments.
The thermal simulation focused on analyzing the temperature distribution at the nozzle tip, with particular attention being paid to the efficacy of thermal insulation (
Figure 8a,b). The results from the static study clearly showed that printheads equipped with thermal insulators exhibited superior thermal retention and a more uniform temperature distribution, effectively mitigating heat loss at the nozzle tip. Specifically, with the printhead temperature set at 37 °C, the temperature of the bioink at the hub reached 36.89 °C for the stainless steel (SS) tip and 36.98 °C for the tapered tip (TT), while average temperatures at the nozzle tip were observed at 36.88 °C (SS) and 35.60 °C (TT), and the lowest temperature of the tip could reach 31.02 °C (SS) and 28.77 °C (TT), respectively.
Moreover, the transient study, which aimed to evaluate the velocity of heat transfer, indicated that printheads fitted with thermal insulators not only enhanced thermal efficiency but also accelerated the heat transfer process (
Figure 8c,d). This improvement could significantly reduce the operational time for users. The experimentally validated benefits of thermal insulators for maintaining optimal temperature distribution and expediting the heating process underscore their importance in achieving high-quality bioprinting outcomes, especially when working with temperature-sensitive bioinks.
Subsequent to our simulation efforts, empirical experiments were conducted to validate the simulation outcomes and assess the effectiveness of the optimized printhead design. However, it was difficult to measure the temperature of bioink directly, and the major purpose of the entire study was to achieve high fidelity structures with a stable printing process. Therefore, a porous cube model was fabricated under various thermal conditions using both traditional and modified printheads, with a particular focus on heating speed and the filament width as a critical measure of printing fidelity. For comparison with simulation results, the printing speed was set to 10 mm3/s
Firstly, the heating speed was measured for a temperature of 37 °C and 26.1 °C via a two-temperature sensor (Pt100) attached at the heating sleeve near the heating film and the nozzle tip, respectively. The experiments lasted for 180 s (3 min), and values were collected every 5 s for the first 120 s and every 10 s for the last 60 s. The results are shown in
Figure 9. It could be seen that for both types of tips, printheads with insulators were able to heat the printhead body, especially at the needle tip, to the desired target temperature. This result practically proved the conclusion of the thermal simulation results. Furthermore, it was noted that thermal insulator significantly improved the heating speed, as the printheads without insulators took about 120 s to heat up to the target temperature, whereas ones with insulators could approach the target temperature within 60 s. These test results showed that printheads with insulators were not only more accurate in terms of temperature control but also heated up faster. However, even with better-controlled thermal distribution and speed, in reality, it was difficult to precisely control the temperature within a range of 1 °C. It could be improved by applying more complex thermal control and stabler sensor installation.
Secondly, the experimental results, as illustrated in
Figure 10, highlight the performance of a 250 µm 1/2 inch stainless steel (SS) nozzle with and without a thermal insulator (
Figure 10a–d and
Figure 10e,f, respectively). Morphologically speaking, when printing without an insulator at a temperature of 37 °C, the structure was successfully printed, with an average filament of 590 µm. However, when the temperature was set to 30 °C, the printed filaments exhibited over-gel conditions. The nozzles were even clogged when the temperature was set to 26.1 °C and lower due to the over-gel bioink and high viscosity.
In comparison, when printing with an insulator at 37 °C, the bioink remained in a solution state, proving inadequate for forming the intended structures due to insufficient gelation. Conversely, at a reduced temperature of 30 °C, the bioink was successfully extruded, albeit with a resolution yet to be satisfied, which was 765 µm on average. A further decrease in temperature to 15 °C induced an over-gel state, characterized by the formation of non-smooth filaments. Optimal printing results were observed when the printhead was set to 26.1 °C, where the filament width and the overall morphology—smooth and well structured with 338 µm—suggested that the bioink was extruded at its proper gelation temperature.
Moreover,
Figure 11 demonstrated the average filament width of the simulation results and printing experiments with an insulator for the SS (
Figure 11a) and TT tips (
Figure 11b), respectively. The results showed similar trends for both tips. When the temperature was set to 26.1 °C and 30 °C, the filament widths from experiments were close to the simulation results. Notably, at a temperature of 15 °C, filaments from the experiment were significantly wider than the simulation results; the possible reason behind this result was the non-smooth filament with over-gel bioink. This condition not only facilitated the strict control of temperature by the printhead but also demonstrated the printhead’s capacity to maintain the bioink within ideal processing conditions for high-quality bioprinting outcomes.
In addition to printing with bioink, cell-laden 3D structures were also printed to investigate the cell viability with the optimized printhead. To examine the effects of cell printing results, the live/dead staining method was used to study the cell survival rate after printing, and the CCK-8 kit was applied to analyze cellular metabolic activity.
Figure 12a shows the cell survival rate calculated with live/dead staining by using different tips with and without an insulator. All groups showed relatively high cell viability (>85%), while printing without an insulator had relatively lower cell viability. Specifically, the SS tip without an insulator recorded 87.6% ± 1.9% viability, and it rose to 95.1% ± 1.5% when applying an insulator. The TT tip without an insulator had 92.9% ± 1.6% cell viability, and the TT tip with an insulator had 96.8% ± 0.6% cell viability. The results indicated that the printhead could deliver viable cells and the cell viability could improve when applying an insulator.
The cell proliferation data captured on days 1, 3, and 7, as presented in
Figure 12b, indicate a consistent increase in CCK-8 signal across all groups, confirming the ongoing proliferation of HUVECs from day 1 to day 7. Notably, while all groups exhibited a growing trend over the seven days, groups with an insulator printhead demonstrated notably higher OD values by day 7 compared to those without an insulator. This suggests a more favorable environment for cell proliferation in the insulator groups. The prior simulations suggest that printing without an insulator is more susceptible to ambient temperature variations, potentially leading to increased shear stress at the nozzle tip. The integration of an insulator into the printhead not only enhances cell proliferation but also mitigates the negative impacts of environmental factors, making it a crucial component for improving bioprinting outcomes.
Overall, these findings broadly align with our simulation predictions, underscoring the insulator’s effectiveness in minimizing heat loss through natural convection at the nozzle tip and achieving a uniform temperature distribution across the printing zone.