Three dimensional freeze printing (3DFP) combines the advantages of freeze casting and additive manufacturing to fabricate multifunctional aerogels. Freeze casting is a cost-effective, efficient, and versatile method capable of fabricating micro-scale porous structures inside the aerogels for many different applications. The 3DFP provided the capability of fabricating highly customized geometries with controlled microporous structures as well. However, there are still many unexplained phenomena and features because of the complexity of post-processes and indirect observation methods. This study demonstrates the design and construction of the in situ imaging systems, which use the x-ray synchrotron radiography to observe freeze casting and 3DFP processes. With the advantages provided by the in situ x-ray imaging techniques, the ice crystal growth with its unique lamellar structures can be observed during the freeze casting process. The movement of freeze front, material deposition, and growth of ice crystals can also be visualized during the inkjet-based 3DFP process.
I. INTRODUCTION
Freeze casting is a well-known method for developing porous structures, including ceramics,1–5 metals,6–10 carbon-based materials,11–13 and nanocomposites,14–17 which showed promising properties for many different applications, such as energy storage and conversion,18–20 photocatalysis,21 liquid chromatography,22 sensors,23–25 and bioengineering.26,27 The freeze casting method has a simple procedure, an environmental-friendly nature, and the ability to tailor the microstructure of the final product.28 In this method, the precursor slurry, which is composed of a liquid solvent and solute particles, is solidified (frozen) inside a thermally insulating mold to obtain the desired shape. Then, the solidified solvent is sublimated under low pressure and low-temperature conditions (freeze-drying).29–32 During freeze casting, the solute particles are rejected by growing crystals of the solvent, which results in a tightly packed network of solute particles. Once the solvent crystals are sublimated by freeze-drying, a porous structure whose morphology is a replica of the solvent ice crystals is obtained.29
Even though the freeze casting provides a variety of tools (e.g., solidification rate, solid content in the precursor slurry, and using chemical additives in the formulation of precursor slurry) to manipulate the microstructure of the object, the macrostructure of the fabricated body is highly limited by the shape of the mold during the freeze casting process. To eliminate the dependency on the mold and achieve porous structures with complex macro-geometries, our group developed the Three Dimensional Freeze Printing (3DFP) technique which combined the advantages of the freeze casting process and additive manufacturing techniques. The 3DFP technique provides a way to fabricate objects with both desired micro-/macro-structures. Accordingly, Song et al. fabricated a 3D hydroxyapatite (HAP) scaffold which was used for bone cell culture and Dasyam et al. proposed the 3DFP of Cellulose Nano-Crystal (CNC) aerogels as ultra-lightweight sound absorbers.33,34 The ink was synthesized and extruded through a microneedle by the applied pressure. A cold plate was used as a substrate for deposition of the ink layer-by-layer and to freeze the ink in order to maintain the 3D structure. Once the complete structure was formed, the final structure was obtained by subsequent freeze-drying and sintering processes. Instead of using extrusion for depositing the ink, the inkjet method can be used to deposit the ink with a very low concentration solution, which can fabricate ultra-light and multifunctional aerogels, such as nanocellulose,35 MXene,36 and graphene composite.37
The microstructure of 3D freeze printed objects can be manipulated by different freeze casting conditions; however, it is quite difficult to directly observe such a process. Current approaches, such as optical microscopy, computed tomography (CT) scan, or Scanning Electron Microscopy (SEM), only provide indirect or partial observations. In order to observe the internal structure, the 3D freeze-printed samples need post-processing after printing. First, the 3D freeze-printed samples need to be stored at very low-temperature conditions (−30 to −70 °C) after printing. Then, they need to be freeze-dried to sublimate the ice; some metal or ceramic samples need an additional sintering process for consolidation. During the freeze-drying or sintering process, the internal structure may change due to shrinkage. Such observations are therefore indirect and cannot provide kinetic information, such as the real size of ice crystals and their growth speed.
In recent years, x-ray synchrotron radiography has been used due to its in situ, real-time, and direct observation features. Compared to the normal laboratory-based x-ray CT imaging, synchrotron-based x-ray has 3–5 orders of magnitude higher photon flux, which allows for shorter imaging time. Therefore, in situ x-ray imaging techniques have been used to investigate the process dynamics of several 3D printing processes, including laser powder bed fusion,38–40 powder-blown laser-additive manufacturing,41 and binder jetting-additive manufacturing.41 Since 3DFP is a hybrid process, understanding the relation between the substrate temperature and the ice crystal growth mechanism is very crucial for the printing quality.
In order to observe the ice crystals’ growth behavior during the 3DFP process, we proposed a real-time x-ray imaging system that combined x-ray synchrotron-based radiography and the 3DFP technique. Due to the limited availabilities of the x-ray synchrotron facilities, equipment, and materials, we performed our experiments in two separate stages. In the first stage, we performed an initial experiment which focused on only the freeze casting process, since the 3DPF is a hybrid process which combines the freeze casting and additive manufacturing techniques. We developed the method utilizing synchrotron-based radiography to observe the ice crystallization of the HAP slurry under the freeze casting process. The ice crystal growth behavior was tracked during the unidirectional freeze casting process. This experiment, which required a relatively simple experimental setup and less control, was conducted at the Advanced Photon Source (APS) at the Argonne National Laboratory. With the gained experience from the initial study, we further performed experiments for the in situ x-ray imaging of the inkjet 3D freeze printing process in the Stanford Synchrotron Radiation Lightsource (SSRL) at the SLAC National Accelerator Laboratory. The inkjet deposition of colloidal silica was monitored by x-ray imaging during the 3DFP process. Considering the complexity of the experimental setup, achieving a remote control on the experimental setup required serious effort and time. With the advantages provided by the x-ray imaging techniques, we were able to observe the freezing front, deposited material, and growing ice crystals simultaneously.
II. SYSTEM DESIGN AND EXPERIMENTAL RESULTS
A. In situ x-ray imaging of the freeze casting process
1. Material preparation
The material for the freeze casting experiment was hydroxyapatite (HAP) slurry. It was prepared by mixing commercial HAP powders (d50 = 1–3 μm, Trans-Tech, Adamstown, MD) in deionized (DI) water at a concentration of 60 wt. %. DARVAN C-N (Vanderbilt Minerals, Norwalk, CT) was added by 0.8 wt. % based on the HAP solid to disperse the HAP powder. The suspension was mixed homogeneously for 5 h by magnetic stirring. Thereafter, hydroxypropyl methyl cellulose (H7509, Sigma-Aldrich, Saint Louis, MO), as a viscosifier, was added by 9 mg/ml in the DI water, followed by stirring for 12 h to dissolve fully. During this step, 1-octanol (112615, Sigma-Aldrich, Saint Louis, MO) was also added by 0.5 vol. % of the distilled water as a defoamer.
2. X-ray synchrotron radiography and freeze casting setup
Synchrotron absorption micro-radiography measurements were performed at the beamline 2-BM of the Advanced Photon Source (APS) at the Argonne National Laboratory. The schematic experiment setup is shown in Fig. 1. The x-ray beam at 2-BM is from a bending magnet source. 2-BM can operate in the monochromatic beam mode in which x-ray energy is selectable between 10 and 45 keV, pink beam mode in which there is only an x-ray mirror in the beam to cut off high-energy x rays, and white beam mode in which there is no x-ray optics in the beam.42 In this study, we used the pink beam mode with a 1.5 mrad incident angle to the beamline mirror. The x-ray cutoff energy was about 45 keV. 1-mm carbon and 2-mm Si filters were used to reduce the low-energy x rays. The x-ray beam delivered to the sample had its average energy around 35–40 keV. The freeze casting equipment was placed in the middle of the x-ray beam such that the transmitted beam reached the high-resolution imaging camera which was used to record the radiography images. The detector consisted of a 100-μm-thick LuAG:Ce scintillator, which converted the x-ray absorption signal into a visible light image. Then, the image was magnified by a 2× Mitutoyo long working distance optic lens and captured by a high-speed CMOS camera (Adimec Q-12A180, Woburn, MA). The image size was 3256 × 2200 µm2, and the effective pixel size in the acquired images was 2.75 × 2.75 µm2.
The schematics of the freeze casting setup are given in Fig. 1. As shown in Fig. 1(a), the well mixed HAP slurry was loaded into a polymer mold with 10-mm length, 21-mm height, and 2-mm thickness. The mold was directly placed on the cold plate (TP294, Sigma Systems, El Cajon, CA), which was set at a constant temperature of −30 °C. The controller of the cold plate used a proportional integral derivative (PID) algorithm to adjust the heating power and cooling pumping flow to maintain the constant temperature. The liquid nitrogen cooling system was integrated with the cold plate, which consists of a pump, a Dewar, and a lid with a stoppered port for refilling during operation. Since there was only one temperature gradient which was along the vertical direction, the ice crystals only propagated along the vertical direction, as the schematic shown in Fig. 1(b). The x-ray beam was perpendicular to the freeze casting sample, which left a rectangular region to be captured during the freeze casting process.
3. X-ray imaging results
In order to generate the unidirectional temperature gradients, the polymer mold was directly attached to the −30 °C cold plate. The ice crystals only grew along the vertical direction, which is also the temperature gradient direction. Since the freezing speed was fast at the bottom region and the field of view was limited, the images captured were located 7 mm above the top surface of the cold plate. The x-ray images of unidirectional freeze casting are shown in Fig. 2. Since the imaged region was smaller than the freezing sample, only part of the sample was able to be visualized (3.256 × 2.2 mm2). We can see some circular objects (circles with black outlines) within the captured region, which are air bubbles. They were generated during slurry preparation and loading of the slurry to the plastic mold. Since the viscosity of the slurry was quite high, air bubbles were trapped and not easy to move. As the sample was continuously exposed to the x rays, the size of some air bubbles expanded due to the heating effect, which is marked by blue circles in Figs. 2(a) and 2(b). We also noticed that the air bubble was able to be pushed by the propagation of ice front and combined with other air bubbles [as marked by yellow ellipses in Figs. 2(b) and 2(c)]. This phenomenon can also result in the formation of larger air bubbles.
The ice front propagating along the temperature gradient (vertical direction) is shown in Figs. 2(a)–2(d). The ice front can be clearly identified during the freeze casting process. The unfrozen slurry is located above the ice front. Since the slurry was well mixed, no contrast can be identified within the unfrozen slurry. Below the ice front, the well-oriented and well-defined lamellar structures can be observed. These structures indicate the typical ice crystal growth during the unidirectional freeze casting process. Due to the unidirectional ice crystal growth, the HAP particles were pushed in between the growing ice crystals, which formed the lamellar structures. We also noticed that the ice crystals at the center part of the sample were growing faster than those growing at the boundary region [ice crystals are higher at the left side, which can be observed in Fig. 2(d)]. This phenomenon was due to the low conductivity of the mold, which generated certain heat effects to the slurry located at the boundary region.
B. In situ x-ray imaging of the inkjet 3DFP process
1. 3DFP process and experimental setup
3DFP is a hybrid process based on drop-on-demand (DOD) inkjet printing and unidirectional freeze casting. Most commonly, water-based Newtonian inks with low-viscosity values are used as the starting material. Using a commercial micro-dispenser (The Lee Co., CT, USA), spherical droplets are generated. For the satellite-free droplet formation, the pressure inside the syringe barrel on which the micro-dispenser is attached needs to be well-adjusted. For that purpose, a pneumatic fluid dispenser (Nordson EFD, RI, USA) is used. The spherical droplets are deposited onto a pre-cooled substrate by a liquid nitrogen operated cold plate (Instec, CO, USA). After the impact of the droplets on the substrate, they immediately freeze and protect their shapes. As in the unidirectional freeze casting process, ice crystals nucleate on the substrate surface and grow along the temperature gradient from bottom to top. The position and the velocity of the micro-dispenser are controlled by a customized three-axis motion stage made from commercially available ball screw driven linear actuators (Panowin Technologies, Shanghai, China). By adjusting the distance and time lapse between successive droplets, coalescence of the droplets is achieved, and uniform lines are obtained. Depositing uniform lines layer-by-layer, frozen structures with complex 3D shapes are obtained. Following, to obtain the 3D printed aerogel, the frozen structure is freeze-dried at −35 °C and 0.2 mbar using a benchtop freeze dryer (Labconco, MO, USA) for at least 48 h. Freeze-drying sublimates the ice crystal in the frozen structure and yields to aerogels whose porosity is a replica of the sublimated ice crystal. Depending on the nature of the material used in the ink formulation, further thermal or chemical treatments can be applied to enhance the integrity and/or provide the functionality to the 3D printed aerogel.
Since the 3DFP process requires post-printing processes, such as freeze-drying and thermal/chemical treatment (when required), investigation of final products thereafter the complete process does not provide quality information regarding the “3DFP” part of the process. To understand the effects of the substrate temperature and deposition rate of the material on the quality of the printed lines, we prepared the setup provided in Fig. 3(a) and performed the in situ investigation of the process using x-ray imaging techniques at bending magnet beamline 2–2 of the Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory, CA, USA. More detailed information about the experimental setup and conclusions drawn from the experiments can be found elsewhere.43 To be able to observe the solidification and material deposition simultaneously, we used a commercially available colloidal silica suspension (15 wt. %) with an average particle size of 4–6 nm (Nyacol, MA, USA). We investigated separate droplets and uniform lines obtained after coalescence of separate droplets and tested different substrate temperature, inter-droplet distance, and time lapse values to perform a systematic study for developing a fundamental understanding of the process dynamics. We further deposited three consecutive lines on top of each other to investigate the adhesion and the ice crystal growth mechanism along the successive layers [see Fig. 3(b)]. To collect the x-ray images, we used a scintillator-based optical system with a 100-µm YAG:Ce scintillator crystal (Crytur Ltd.) coated with 120 nm of Al on the upstream side. A high reflectance mirror (Thorlabs) was used to bend the visible light 90° off-axis to the x rays into a 4× long working distance infinity-corrected objective lens (Nikon), an infinity-corrected tube lens (Thorlabs), and a high-speed CMOS camera pco.dimax S4 (PCO). Using this optical setup, we obtained an effective pixel size of 2.4 µm for a field of view of 4.8 × 2.4 mm2, and images were captured at 500 Hz. As presented in Fig. 4, we were able to observe the deposited material along with the ice (solidification) front simultaneously. This capability of in situ x-ray imaging helped us to understand the effect of 3D printing parameters, such as printing speed, material jetting frequency, and substrate temperature on the final product quality. Furthermore, with the help of the different x-ray absorptivity of silica and water (ice), we observed the growth of the ice crystals, which showed the freezing direction and interlayer fusion of the deposited material.43
III. CONCLUSION
The water-to-ice phase transformation is probably one of the trickiest phenomena to investigate among solidification studies. The microstructural characteristics of freeze casting materials are determined by the morphology of the frozen fluid, which depends largely on the nature of the interactions between the particles within the suspension and the solid/liquid interface. As a result of the complex and interdependent relationships in the freeze casting process, material properties can vary widely even within similar systems. In addition, the 3D freeze printing combines additive manufacturing and freeze casting processes, which enables the dynamic deposition of the material during the freeze casting, making the system even more complex. Therefore, a theoretical understanding is necessary to gain predictive control of such a complex system.
In this study, we developed a method that utilizes in situ x-ray synchrotron-based radiography to observe the freeze casting and 3DFP processes. The following conclusions are obtained from this study:
The system we built provided a new method to in situ observe the formation of internal structures during the freeze casting and 3DFP processes.
The ice crystal growth with its unique, well-oriented, and well-defined lamellar structures can be observed during the freeze casting process.
With the advantages provided by the rapid x-ray imaging, we were able to observe the freeze front, deposited material, and growing ice crystals simultaneously during the inkjet 3DFP process.
With the help of rapid x-ray synchrotron radiography/tomography and the development of imaging techniques, we will be able to generate a more fundamental understanding of the physics behind the 3D freeze printing process in the future.
ACKNOWLEDGMENTS
Professor Lin acknowledges support from the National Science Foundation (Award No. 1943445), the NASA EPSCoR CAN Grant (Award No. 80NSSC19M0153), and the Johnson Cancer Center. This research used the resources of the Advanced Photon Source (APS), a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by the Argonne National Laboratory(Contract No. DE-AC02-06CH11357). Use of the Stanford Synchrotron Radiation Lightsource (SSRL), SLAC National Accelerator Laboratory, was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences (Contract No. DE-AC02-76SF00515).
AUTHOR DECLARATIONS
Conflict of Interest
The authors declare no conflicts of interest.
Author Contributions
G.Y. and H.T. contributed equally to this work.
DATA AVAILABILITY
The data that support the findings of this study are available within the article.