This paper describes successful efforts to design, build, test, and utilize a single crystal apparatus using the Bridgman approach for directional solidification. The created instrument has been successfully tested to grow magnesium single crystals from melt. Preliminary mechanical tests carried out on these single crystals indicate unique and promising properties, which can be harnessed for biomedical applications.

During the last several years, interest in biodegradable magnesium implants has dramatically increased. In general, biodegradable metals employed as coatings or bulk materials for medical implants offer significant advantages over conventional non-degradable metals used today.1–5 Biodegradable implants for orthopedic applications dissolve in the body over time thus eliminating the need of a second surgery for their removal. They may also reduce the chance of an immune response when the implant is in the body permanently and allow placing a new implant if necessary. Biodegradable implant materials can be either natural (i.e., proteins) or synthetic. Synthetic materials offer greater advantages than natural materials in that they can be tailored to give a wider range of properties than naturally available materials. Despite the development of novel polymers that meet a number of demanding requirements,2 polymer based implants have shortcomings such as lower Young’s modulus, undesirable reactions with the surrounding biological environment, and complex as well as expensive sterilization process.2,3 In contrast to polymeric materials, metals like Mg and Fe are present in the human body as trace elements and have favorable mechanical properties, which make them promising candidates for temporary implant materials.2–4 

Magnesium and its alloys have been shown to possess attractive biological performances and are potential biodegradable materials thanks to the following characteristics: (1) Mg is biodegradable in body fluids by corrosion; (2) Mg2+ ions are harmless to human body; (3) Mg can accelerate the growth of new bone tissue; (4) the density, elastic modulus, and yield strength of Mg are closer to the bone tissue than that of the conventional implants. Overall, Mg and its alloys seem superior to any other metallic or polymeric implants used for bone repair or orthopedics.3,5,6 The most advanced clinical applications of Mg are biodegradable cardiovascular magnesium stents that have been successfully tested in animals and the first clinical human trials have been conducted.7 Magnesium alloys were also studied as bone implants and can be applied in various designs, e.g., as screws, plates, or other orthopedic fixture devices.3 Magnesium chips have been investigated for vertebral fusion in spinal surgery of sheep5 and open-porous scaffolds made of Mg alloys have been introduced as load bearing biomaterials for tissue engineering.5 

Despite many favorable properties of Mg and Mg-based alloys, their performance as biodegradable implants is often hampered by poor response when placed in body or simulated body environments. Commercially available Mg or Mg alloys are routinely machined from cast ingots and as such are an agglomerate of small crystals with random orientations. They reveal grain boundaries, surface roughness, and structural defects. These grain boundary regions, which are within a few atomic diameters, have different characteristics from the bulk crystal because of lattice irregularities accommodating the change in orientation and are more vulnerable to corrosion.8,9 The impurities tend to segregate there and the minority phases also precipitate along the boundary regions. These phenomena are influential in directing corrosion or stress-corrosion cracking along grain boundary paths.5 Galvanic corrosion can also occur in polycrystalline alloys due to differences in microstructural phases. Distinct localized anodic and catholic microstructural areas are developed due to inhomogeneity, which act as micro-electrochemical cells in the presence of a corrosive medium (electrolyte). Currently, controlling Mg properties such as strength and corrosion resistance for implant applications is often achieved by alloying pure Mg with a variety of different elements.10,11 However, when a small amount of an alloying element is added to pure Mg metal, the grains tend to segregate alloying elements along their boundaries thereby amplifying the corrosion rate.11 This often results in the enhanced evolution of hydrogen that causes local gas cavities in vivo.6,12 Many of the alloying elements employed to improve the mechanical properties of Mg such as Al, Mn, and rare earth elements are toxic at high local concentrations if released from corrosion sites, while others such as Zn and Ca may increase corrosion by segregation along grain boundaries or forming secondary phases.8–11 Surface engineering of Mg by the way of coating, chemical treatment, and mechanical processing has met with limited success. It was found that coating and chemical treatments like anodization cannot fully prevent pitting corrosion and other undesirable events like hydrogen evolution or local alkalinity.11–13,15 Mechanical treatment is promising but unproven in a biological environment.16 Lower corrosion rate is preferred because it does not affect the cell viability.14 Since corrosion is initiated at the surface, often along continuous defect-rich areas like grain boundaries, it is plausible that eliminating grain boundaries will reduce the defects available, which can serve as sites for anodic dissolution of the metal. The elimination of grain boundaries can be achieved by growing the material as one single crystal. This will also reduce the segregation of impurities, which can exacerbate the corrosion of the metal. Furthermore, the corrosion rates of three-dimensional single crystals of Mg or agglomerates of high purity single crystals (crystallites) vary with crystallographic orientations.17,18 The observed anisotropy is most pronounced for the Mg (0001) surface, which exhibits pitting corrosion. The (1010) and (1120) surfaces are more resistant to corrosion.17 This anisotropy can be exploited in a single crystal Mg implant by orienting the implant along specific crystallographic planes with the objective of slowing or acceleration the degradation rate as required.

Publications on growing Mg single crystals cannot be found lately, since most of them are clustered before or around 1950. Bridgman-Stockbarger method is frequently used for single crystal growth, including Mg single crystals. A schematic illustrating this method is shown in Figure 1. This is a directional solidification process wherein the single crystal is grown by extremely slow rate of cooling from one end of the melt. The charge or raw Mg is contained in a container or crucible with a tapered bottom. This crucible is heated in a furnace to melt the metal and then it is moved out of the hot zone furnace at a very slow speed. This gradual cooling allows the formation of a single Mg nucleus in a solid form at the tapered end of the melt. As the crucible moves out of the hot zone, the interface between the solid nucleus and the molten Mg advances further into the liquid eventually resulting in the entire melt solidifying into one single crystal. The advantage of this technique is that it is faster and cheaper than other single crystal growth techniques. It is also a simple yet versatile technique. The shape and size of the crystal can be controlled in this method.19,20 One of the earliest attempts to grow a single crystal of Mg was made by Mochalov.19 He used a horizontal mold in hydrogen (4 atm pressure). In 1952, Burke and Hibbard obtained single crystal growth of Mg in a vertical graphite mold by using in situ cooling.20 Later, Reed-Hill and Robertson also used the same approach for successful single crystal growth.21 In 1957, Long and Smith utilized a similar technique for Mg single crystal growth.22 They maintained a slow cooling rate of 5 ° C/h and the resulting single crystal was 1 in. in diameter and 4 in. in length. Nichols fabricated 25 Mg single crystals and observed that their orientations were random.23 In 1958, Labzin and Bazhenov adopted a slightly different zone melting approach to generate single crystal Mg with controlled orientation.24 

FIG. 1.

Bridgman-Stockbarger method for directional solidification and growth of single crystals.

FIG. 1.

Bridgman-Stockbarger method for directional solidification and growth of single crystals.

Close modal

The growth of Mg single crystal seems simple because of the low (650 °C) melting point of Mg. However there are a few challenges that have to overcome in order to grow single crystals for biomedical implant applications such as:

  1. The chemical reactivity of Mg especially towards oxygen, and also the high vapor pressure of Mg at elevated temperature complicate the process of crystal formation from melts. Since the process has to be carried out in an inert environment, it is necessary to place Mg inside a sealed chamber.

  2. Previous efforts using the Bridgman crystal approach achieved directional solidification by moving the charge-containing crucible in a static tube furnace manually or by using a servomotor. This is difficult to carry out if the melt is placed inside a sealed furnace chamber since moving the whole chamber through the furnace is impractical given the space and material limitations. For larger crystals such as turbine blades, the charge is kept stationary and an elaborate multi-zone furnace enables the directional solidification with each zone independently programmed to obtain a thermal gradient, which promotes solidification from one end. Designing such a system for smaller crystals is possible but potentially expensive. Instead, if the furnace is moved vertically, a shorter chamber (∼1 m long) could be used and a thermal gradient as well as directional solidification can be obtained even with a sealed furnace chamber.

In this paper, we present new data related to designing, building, and testing a vertical Bridgman apparatus with a moving furnace and computer controlled motions and temperature. Further, we share optimized recipes for growing conditions to fabricate Mg single crystals with different dimensions. Finally, characterization data related to the Mg single crystals are revealed, including XRD, metallography, optical microscopy, and mechanical testing.

The two-zone furnace has automatic closed loop temperature control to provide the necessary temperature gradients, uniformity, temperature ramps, and stable temperature profile needed for proper crystal growth. The furnace is interfaced with the microprocessor control system for setting, monitoring, graphing, and data logging of all temperatures with respect to time. Magnesium was melted in a graphite crucible, which was placed inside a steel crucible holder. This arrangement was placed inside a vertical quartz chamber supported on a quartz rod as shown in Figure 2. The chamber is a quartz tube with a 2-in. inner diameter (Figures 2 and 3). It is closed on the top and is shaped to form a flange at the bottom sealed by a Viton o-ring. This quartz tube also referred to as “quartz chamber” secures an inert Ar environment for the melting of magnesium.

FIG. 2.

(a) Picture of the two zone furnace; (b) schematic of the two zone furnace with the quartz process chamber crucible, crucible holder, and supporting quartz rod.

FIG. 2.

(a) Picture of the two zone furnace; (b) schematic of the two zone furnace with the quartz process chamber crucible, crucible holder, and supporting quartz rod.

Close modal
FIG. 3.

Picture of the computer based GUI screen for monitoring and controlling the system.

FIG. 3.

Picture of the computer based GUI screen for monitoring and controlling the system.

Close modal

1. Growth chamber

The chamber was made out of quartz. This material can withstand temperatures up to 1200 °C. Quartz has also a very low coefficient of thermal expansion and is chemically inert. Since melting and solidification of Mg that takes place at moderate temperatures, quartz was preferred as the chamber material despite the risk of magnesium evaporating and reacting with quartz to form Mg2Si precipitates.

2. Crucible

Graphite crucible was determined to be most suitable for single crystal growth of Mg. Mild steel and titanium were also considered. However, it was found that liquid magnesium wets the surface of steel leading to adhesion between the solidified rod and the steel surface. This causes difficulties in the retrieval of the rod. On the other hand, magnesium does not wet the surface of graphite and therefore retrieval of the final solid single crystal is much easier. Pictures of graphite crucibles used for growing Mg single crystals with different dimensions are displayed in Figure 4.

FIG. 4.

Pictures of graphite crucibles used for growing Mg single crystals with different dimensions: (1) iron crucible with 3.5 mm bore; (2) graphite crucible with 3.5 mm bore; (3) graphite crucible with 6.5 mm bore; (4) graphite crucible for growing screw shaped single crystals.

FIG. 4.

Pictures of graphite crucibles used for growing Mg single crystals with different dimensions: (1) iron crucible with 3.5 mm bore; (2) graphite crucible with 3.5 mm bore; (3) graphite crucible with 6.5 mm bore; (4) graphite crucible for growing screw shaped single crystals.

Close modal

3. Crucible holder

It was initially decided that graphite would be an ideal material for the crucible holder. However the initial trial experiments resulted in the formation of powdery graphite from the holder around the melted magnesium. These particles were most likely produced from the lid of the graphite crucible holder. To avoid such a side effect, a new crucible holder was made out of ferritic stainless steel (RA 446). The evolution of the materials for the crucible holders is shown in Figure 5.

FIG. 5.

Schematic of the crucible holder (1); pictures of the graphite crucible holder (2), and ferritic stainless steel (RA 446) crucible holder (3).

FIG. 5.

Schematic of the crucible holder (1); pictures of the graphite crucible holder (2), and ferritic stainless steel (RA 446) crucible holder (3).

Close modal

The growth process is conducted following the stages illustrated in Figure 6:

FIG. 6.

Different stages in the singles crystal growth run: (a) initial unloading and (b) loading cycles of raw Mg slugs; (c) loading the furnace, pump down, and purge with Ar; (d) heating to above the melting point and soaking; (e) slow movement of furnace from high temperature zone to lower temperature zone; (f) continuation of movement and beginning of cool down; (g) controlled cooling down; (h) final purge and sample unload.

FIG. 6.

Different stages in the singles crystal growth run: (a) initial unloading and (b) loading cycles of raw Mg slugs; (c) loading the furnace, pump down, and purge with Ar; (d) heating to above the melting point and soaking; (e) slow movement of furnace from high temperature zone to lower temperature zone; (f) continuation of movement and beginning of cool down; (g) controlled cooling down; (h) final purge and sample unload.

Close modal

The following processing factors enabled creating the appropriate environment for the growing of Mg single crystals:

1. Vacuum flush and Ar purging

Air was removed from the melt chamber prior to start of the heating cycle by pumping down to a vacuum of 10−4 Torr, followed by Ar purging that continued throughout the entire run.

2. Argon atmosphere

A constant argon flow rate of 1000 SCCM (standard cubic centimeters per minute) was found to be adequate to maintain purity and prevent oxidation of Mg at elevated temperatures.

3. Temperature of melt zones and temperature gradient

The furnace is designed to move vertically thereby making it possible for the melt to be surrounded by different temperature zones. Such an approach allows for the melt to pass through a temperature gradient without actually being moved. This is different from most other Bridgman systems wherein the crucible with the raw Mg is moved into the furnace, soaked for an appropriate time period, and then gradually moved out to initiate nucleation of a single crystal and its growth.

Since the furnace has two zones each containing three controllable sub-zones, maintaining the top zone and the bottom zone at different temperatures could create a temperature gradient. The top three sub-zones were maintained at 780-760 °C for good homogenization of the melt. The three sub-zones at the bottom were maintained at 600 °C. This arrangement causes the melt to pass through an approximate temperature gradient of −1 °C/mm. This small gradient was sufficient to enable good single crystal growth.

4. Soaking time

This is the time at which the melt was maintained at a temperature slightly above the Mg melting point. Soaking time is critical for getting good temperature homogenization and growing of perfect single crystals. Soaking times from 90 min, 2 h, 4 h, 6 h, 8 h, and 14 h were tested. It was found that shorter soaking times do not allow for complete homogenization and cause nucleation and growth of 2-3 grains, while with longer soaking times there was a risk of loss of material through evaporation as well as a changed growth environment which could promote the growth of polycrystalline samples. For single crystals of around 6.5 mm diameter, best growth result was obtained for a soaking time of about 6–14 h. For growing crystals with diameters of around 10-15 mm, the melt was soaked for a longer period of around 18-24 h.

5. Crucible rotation

The rotation of the crucible helped in creating homogeneous thermal environment compensating for any minor asymmetry of the heating furnace. The rotation speed was maintained at 7.5 rpm. This low rotation speed enabled uniform exposure of the crucible to the heat flux generated by the cylindrical furnace and simultaneously prevented pronounced centrifuging forces that may cause non-uniform mass distribution within the growing crystal.

6. Furnace movement

In the First Nano crystal grower, the crucible with the charge is in the upper zone of the furnace during the initial heating and soaking where it undergoes complete melting. Once the melt has been sufficiently soaked, the furnace movement begins and gradually the melt is surrounded by lower zones which are kept at 600 °C. The initial cooldown of the melt in Ar was conducted by the computer-controlled motion of the furnace. The motion rate can be varied between 0.5 and 400 mm/h in order to optimize the growth rate enabling nucleation and growth of one Mg crystal starting from the tapered point of the crucible. In the current experiments, the rate of motion was kept at 30 mm/h for crystals with diameters up to 6.5 mm. For growing crystal with larger diameters (d ≥ 10 mm), the rate of motion was slowed down to around 5-15 mm/h. The lower rate allowed for successive layers of the melt to solidify uniformly across its width.

7. Cooling rate of the grown single crystal

Optimization of this parameter is important to avoid generating structural defects after the Mg single crystal is grown. The best value was found to be 53.33 °C/min. The largest crystal (d = 15 mm) was cooled at a slower cooling rate of 40 °C/min.

Single crystal samples were prepared using the Bridgman-Stockbarger method in a two zone Easy Crystal Furnace by First-Nano (CVD Equipment Corporation). A graphite crucible was housed in a sealed quartz cylinder to maintain positive Ar atmosphere with a constant argon flow. Ultra-high purity Ar from Wright Brothers was used during melting and crystal growth processes. Graphite crucibles with different bore sizes (3 mm, 6.5 mm, 10 mm, and 15 mm) were used. The bore depth of the crucibles with smaller bore sizes (3 mm, 6.5 mm, and 10 mm) was maintained at 50 mm. The bore depths of the crucible with 15 mm bore sizes were increased to 100 mm for growing larger crystals.

1. Raw materials

The raw materials were supplied in different forms and shown in Figure 7. Initially, fine Mg powder (325 mesh) having 99.95% purity from Sigma-Aldrich was used. Later, Mg chips were also obtained for growing single crystals. Owing to the limitations of these materials, Mg slugs (99.95% purity) were obtained from Alfa Aesar in 2 different sizes: 3.0 mm diameter × 12.5 mm length and 6.5 mm diameter × 12.5 mm length. Later Mg rods of the same purity but of larger dimensions were also used to scale up the growth process and grow bigger Mg single crystals. Some magnesium rods were also machined into a screw form and used as a charge for growing screw shaped single crystals. The corresponding graphite crucible was also similarly machined to fit the screw shaped slug. This set of experiments explored the possibility of growing pre-shaped Mg single crystals for medical implants. Pre-shaped single crystal growth is designed to reduce the need for machining the as-grown Mg single crystals. The machining process often induces defects in the crystal lattice, thus increasing the sites on the implant surface that are more vulnerable to corrosion.

FIG. 7.

Magnesium raw materials: (1) magnesium powder; (2) magnesium chips; (3) magnesium slugs 3 mm diameter × 5 mm long; (4) magnesium slugs 6.5 mm diameter × 12 mm long; (5) magnesium rod 6.5 mm diameter × 50 mm long; (6) magnesium rod 15 mm diameter × 100 mm length; (7) magnesium rod machined as screw for growing screw shaped single crystal.

FIG. 7.

Magnesium raw materials: (1) magnesium powder; (2) magnesium chips; (3) magnesium slugs 3 mm diameter × 5 mm long; (4) magnesium slugs 6.5 mm diameter × 12 mm long; (5) magnesium rod 6.5 mm diameter × 50 mm long; (6) magnesium rod 15 mm diameter × 100 mm length; (7) magnesium rod machined as screw for growing screw shaped single crystal.

Close modal

The practiced steps displayed in pictures for growing Mg single crystals are illustrated in Figure 8.

FIG. 8.

Magnesium single crystal growth steps from left to right: (1) Mg raw material; (2) graphite crucible; (3) stainless steel holder; (4) loading the crystal grower; (5) furnace ready to go.

FIG. 8.

Magnesium single crystal growth steps from left to right: (1) Mg raw material; (2) graphite crucible; (3) stainless steel holder; (4) loading the crystal grower; (5) furnace ready to go.

Close modal

2. Process conditions

The different stages of single crystal growth process have been elaborated in Sec. III C. For our growth experiments, Mg was melted at 760 °C-780 °C. The optimum soaking time for crystals with diameters below 10 mm was found to be 14 h. The larger 15 mm wide crystals were soaked at 20-24 h, which enabled uniform temperature distribution within the magnesium melt. The growth rate for crystals up to 6.5 mm diameter was maintained at 30 mm/h. For wider crystals, the growth rate was slowed down and maintained between 5 and 15 mm/h. All experiments were carried out in an inert Ar environment at an ambient pressure of 720 Torr. The different types of raw materials used for single crystal growth and the corresponding operation parameters are listed in Table I.

TABLE I.

Single crystal growth parameters for different raw materials.

Raw materialSoaking temperature (°C)Soaking time (h)Furnace movement speed (mm/h)Rate of final cooling (after solidification) (°C/h)
Mg powder 760 2-8 30 53.33 
Mg chips 760 2-8 30 53.33 
Mg slugs 3 mm ϕ × 5 mm height 760 2-8 30 53.33 
Mg slugs 780 8-14 30 53.33 
6.35 mm ϕ × 12 mm height 
6.35 mm ϕ × 50 mm height 
Mg slugs 10 mm ϕ × 50 mm height 780 18 15 53.33 
Mg slugs 15 mm ϕ × 100 mm height 780 24 40 
Mg screw shaped ingot 780 14 15 53.33 
Raw materialSoaking temperature (°C)Soaking time (h)Furnace movement speed (mm/h)Rate of final cooling (after solidification) (°C/h)
Mg powder 760 2-8 30 53.33 
Mg chips 760 2-8 30 53.33 
Mg slugs 3 mm ϕ × 5 mm height 760 2-8 30 53.33 
Mg slugs 780 8-14 30 53.33 
6.35 mm ϕ × 12 mm height 
6.35 mm ϕ × 50 mm height 
Mg slugs 10 mm ϕ × 50 mm height 780 18 15 53.33 
Mg slugs 15 mm ϕ × 100 mm height 780 24 40 
Mg screw shaped ingot 780 14 15 53.33 

1. Metallography and optical microscopy

The cylindrical single crystal ingots of Mg were mounted in Beuhler Epokwick resin along the long side. They were coarse polished parallel to the cylindrical axis using Beuhler emery paper from 400, 600, 800, and 1200 grit and fine polished using diamond films having particle sizes 15, 9, 3, 1, and 0.25 μm. The polished samples were then etched for 10-15 s with nitric acid based etchant (25% nitric acid + 25% ethanol + 50% methanol) in order to reveal any grain boundaries. A polycrystalline Mg disc was also polished and etched similarly in order to observe the contrast between single crystal and polycrystalline microstructures.

2. X-ray diffraction

The single crystallinity of the obtained Mg rods was also confirmed using a Panalytical X’pert Powder X-ray diffractometer with a Cu Kα radiation (0.154 nm). By mapping the cast ingot via exposure to the X-ray beam across the whole length, the single crystallinity was confirmed. The persistence of X-ray peaks across the whole sample indicated the presence of the same crystal grain across the length. Optical micrographs obtained after chemical etching of polished surface confirmed whether the grain was indeed single crystal, or in some cases of not optimal growth conditions, a bi-crystal. The XRD parameters are as follows:

  1. Start angle: 25° (polycrystalline Mg), 15° (single crystal Mg).

  2. End angle: 110° (polycrystalline Mg), 92° (single crystal Mg).

  3. Step size: 0.08°.

  4. Time per step: 0.03 s.

  5. Scan speed: 0.266 667°/s.

  6. Voltage: 45 KV.

  7. Current: 40 mA.

3. Laue XRD

Laue XRD is often used for determining the single crystallinity and the orientation of the crystals. A single crystal rod was analyzed using the back-scattered technique in a Multiwire Laue XRD system, which is operational at the Argonne National Lab.

4. XRD pole figure analysis

XRD pole figure analysis was carried out in a Bruker AXS Discover D8 XRD instrument (Figure 9) at North Carolina Agricultural and Technological State University (NCAT). This system allows for rotation of the sample along the “chi” (χ) and “phi” (φ) angles in addition to the regular “theta” (θ) and “2theta” (2θ) angles. This enables us to perform pole figure analysis of individual families of planes in order to determine their orientation of the planes and also to evaluate whether the sample is a single crystal or composed of several crystals. The samples were cut into disc form using an Accutex SPi 300 Electric Discharge Machining (EDM) setup and further polished with SiC paper up to 1200 finish. They were mounted on an adjustable stage so that the samples could be aligned as precisely as possible. The process parameters were fed into a “script” which was utilized by the XRD computer program to run specific rotations along χ and φ planes at predetermined θ-2θ combinations in order to detect the presence and orientations of specific families of planes.

FIG. 9.

Pictures of the Bruker AXS D8 Discover XRD instrument at NCAT: (1) the XRD chamber; (2) the stage with the sample on top; (3) the stage placed between the source and the detector.

FIG. 9.

Pictures of the Bruker AXS D8 Discover XRD instrument at NCAT: (1) the XRD chamber; (2) the stage with the sample on top; (3) the stage placed between the source and the detector.

Close modal

5. Tensile testing

The single crystal rods were cut into dog-bone shaped test pieces in line with the ASTM E9 standard for tensile using EDM setup. The gauge length of the test pieces was maintained at 12 mm. The tensile pieces were tested in an MTS-EM tensile testing machine. A 500 N load cell was used for the tests.

Different forms of Mg used as raw materials yielded different results, as shown in Figure 10. The list of raw materials and the corresponding crystal size and quality are listed in Table II. The powdered Mg seemed to form a dark gray mass of agglomerated particles indicating some combination of oxidation and partial melting. The magnesium chips seemed to form a more homogenous ingot. However the size was very small indicating considerable shrinkage in the initially packed mass of chips upon melting. The magnesium slugs and rods with 6.35 mm diameter upon melting yielded cylindrical as-cast ingots, which were about 12-50 mm in length. The smaller slugs with diameter of 3.5 mm did not yield consistent rods upon melting and showed significant shrinkage. Once the single crystal growth process was established, a few experimental runs were carried out in order to grow pre-shaped single crystal screws of magnesium. The obtained screw shaped single crystal is also shown in Figure 10. Optical and XRD characterization of the as-cast ingot was conducted to confirm single crystallinity. The primary characterization procedure in this case was a combination of XRD and optical microscopy. Later Laue XRD and XRD pole analysis confirmed the results obtained from the XRD study.

FIG. 10.

Final products from different raw materials: (1) Mg powder melted in 3 mm bore and soaked for 6 h; (2) Mg chips melted in 6.5 mm bore and soaked for 6 h; (3) Mg slugs (3 mm diameter × 5 mm length) soaked for 6 h; (4) Mg slugs (6.5 × 12 mm) soaked for 6 h; (5) Mg rod (6.5 × 50 mm) soaked for 14 h; (6) increase in the size of crystals grown over the years (bottom to top: 6.5 mm ϕ × 50 mm, 10 mm ϕ × 50 mm, and 15 mm ϕ × 100 mm); (7) as-grown single crystal Mg screw.

FIG. 10.

Final products from different raw materials: (1) Mg powder melted in 3 mm bore and soaked for 6 h; (2) Mg chips melted in 6.5 mm bore and soaked for 6 h; (3) Mg slugs (3 mm diameter × 5 mm length) soaked for 6 h; (4) Mg slugs (6.5 × 12 mm) soaked for 6 h; (5) Mg rod (6.5 × 50 mm) soaked for 14 h; (6) increase in the size of crystals grown over the years (bottom to top: 6.5 mm ϕ × 50 mm, 10 mm ϕ × 50 mm, and 15 mm ϕ × 100 mm); (7) as-grown single crystal Mg screw.

Close modal
TABLE II.

Single crystal growth results for different raw materials.

Raw materialFinal single crystal size and quality
Mg powder Dark mass of particles 
Mg chips Small rounded solid ingot 
Mg slugs 3 mm ϕ × 5 mm height Shrunk rod with graphite particles adhering to the surface 
Mg slugs Solid single crystal (6.5 mm ϕ × 50 mm height) 
6.35 mm ϕ × 12 mm height 
6.35 mm ϕ × 50 mm height 
Mg slugs 10 mm ϕ × 50 mm height Solid single crystal (10 mm ϕ × 50 mm height) 
Mg slugs 15 mm ϕ × 100 mm height Solid single crystal (15 mm ϕ × 100 mm height) 
Mg screw shaped ingot Screw shaped single crystal 
Raw materialFinal single crystal size and quality
Mg powder Dark mass of particles 
Mg chips Small rounded solid ingot 
Mg slugs 3 mm ϕ × 5 mm height Shrunk rod with graphite particles adhering to the surface 
Mg slugs Solid single crystal (6.5 mm ϕ × 50 mm height) 
6.35 mm ϕ × 12 mm height 
6.35 mm ϕ × 50 mm height 
Mg slugs 10 mm ϕ × 50 mm height Solid single crystal (10 mm ϕ × 50 mm height) 
Mg slugs 15 mm ϕ × 100 mm height Solid single crystal (15 mm ϕ × 100 mm height) 
Mg screw shaped ingot Screw shaped single crystal 

The polished and etched surfaces were observed optically in a Keyence VHX-2000 digital microscope as shown in Figure 11. The etched surface of the single crystal appears relatively cleaner due to the absence of grain boundaries. The sharp parallel lines on the single crystal surface are thought to be twinned regions.

FIG. 11.

(1) Etched Mg single crystal (14-h soak), (2) polycrystalline Mg microstructure.

FIG. 11.

(1) Etched Mg single crystal (14-h soak), (2) polycrystalline Mg microstructure.

Close modal

An initial scan was carried out on the polished polycrystalline sample with 2θ ranging from 15° to 120° to get the diffraction peaks. Polycrystalline X-ray data are presented in Figure 12. They show the presence of all the allowed reflections for the Hexagonal Close Packed (HCP) polycrystalline Mg.

FIG. 12.

Polycrystalline Mg XRD data.

FIG. 12.

Polycrystalline Mg XRD data.

Close modal

XRD measurements were conducted on cast and polished ingots across the surface mapping along the axial length of the sample at different locations, as depicted in Figure 13. Both figures show the XRD patterns at corresponding locations. As can be seen, the peaks persist across the patterns confirming the existence of a single grain with identical orientations and implying single crystallinity. Figure 13 shows the XRD peaks persisting across the sample indicating same orientation. Some peaks were shorter most likely due to the curvature of the cylindrical crystals. However, the peak location remained unchanged, which indicated the single crystal nature of the rod. Similarly the screw shaped single crystal was also mapped for XRD peaks across its length as shown in Figure 14.

FIG. 13.

XRD mapping of Mg single crystal at locations 1, 2, 3. The crystal was grown with 14 h soaking time.

FIG. 13.

XRD mapping of Mg single crystal at locations 1, 2, 3. The crystal was grown with 14 h soaking time.

Close modal
FIG. 14.

XRD mapping of screw shaped Mg single crystal at 6 different locations.

FIG. 14.

XRD mapping of screw shaped Mg single crystal at 6 different locations.

Close modal

The Laue XRD was performed at different locations along the length of the crystal rods. The voltage and current were 20 KV and 15 mA, respectively. The single grains in the 6 h soaked samples (Figure 15) were represented in the Laue patterns by distinct set of points. This arrangement of points represents HCP lattice and is consistent along the length of the crystals. A control polycrystalline magnesium rod was also used to observe the difference in the patterns between single crystal and conventional polycrystalline Mg (Figure 16). As expected, the polycrystalline sample was marked by a diffuse pattern owing to multiple grains.

FIG. 15.

Laue XRD patterns for 6 hour soaked Mg crystal obtained at top, center, and bottom of the rod. The patterns are clearly dotted and are identical for the scans obtained for the tested top, bottom, and center regions.

FIG. 15.

Laue XRD patterns for 6 hour soaked Mg crystal obtained at top, center, and bottom of the rod. The patterns are clearly dotted and are identical for the scans obtained for the tested top, bottom, and center regions.

Close modal
FIG. 16.

Laue XRD pattern (left) for polycrystalline Mg rod shown (right). The patterns is diffused and in clear contrast with those obtained for the single crystal Mg rods.

FIG. 16.

Laue XRD pattern (left) for polycrystalline Mg rod shown (right). The patterns is diffused and in clear contrast with those obtained for the single crystal Mg rods.

Close modal

The XRD pole figures for the Mg single crystal rod (14 h soaked) are shown in Figure 17. The 002 plane is represented by a sharp dot on the pole figure. This indicates the presence of a single crystal. The 002 basal plane makes a chi angle of 22° with respect to the circular cross section of the crystal rod. The prismatic (101) and pyramidal (112) pole figures are in agreement with the crystallinity and orientation indicated by the pole figure of the basal plane.

FIG. 17.

XRD pole figures for 14 h soaked Mg single crystal: 002 plane (left), 101 plane (center), and 112 plane (right).

FIG. 17.

XRD pole figures for 14 h soaked Mg single crystal: 002 plane (left), 101 plane (center), and 112 plane (right).

Close modal

High ductility and superplastic behavior were revealed when conducting tensile tests on single crystal samples (Figure 18). The initial strain at zero stress is due to the movement or adjustment of the specimen before the load is applied. Despite their lower yield strength (∼60 MPa), they exhibit tendency for unusually high plastic deformation up to 50%-60% indicating potential superplastic behavior. The large area under the blue single crystal curve in Figure 18 indicates good toughness and a large safety margin that may prevent catastrophic failures. The limitations of the load cell (500 N) prevented the polycrystalline Mg specimen from being completely loaded to the point of failure.

FIG. 18.

Stress-strain curves for polycrystalline Mg and Mg single crystal.

FIG. 18.

Stress-strain curves for polycrystalline Mg and Mg single crystal.

Close modal

Bridgman pioneered the directional solidification technique for single crystal growth presented in multiple publications during the first half of the twentieth century.25 Despite the subsequent advancement of this technology in temperature and motion control, the single crystal growth process requires going through a development progression involving also trial and error attempts. The selection of material for crucible and crucible holder is important since these components are in close proximity to the molten magnesium. Graphite, i.e., carbon does not react with molten Mg and hence ensures that the melt remains relatively uncontaminated. Graphite is a soft material and relatively easy to machine. This adds to its advantage since a variety of crystal sizes and shapes can be grown by easily changing the melting cavity inside the crucible. Graphite is usually available as coarse-grained, fine-grained, and ultrafine-grained. The fine-grained versions of graphite are denser, less porous, and easy to machine. The coarse graphite crucibles used in our initial experiments were plagued with magnesium oxide possibly due to absorbed atmospheric oxygen in the pores. After we switched to fine-grained graphite (McMaster-Carr) for melting magnesium, this problem was eliminated.

The initial use of graphite as a crucible holder often resulted in its wear out upon repeated heating and cooling cycles. The resulting particles of graphite in the melting cavity are not reactive but can interfere with the growth process of the Mg single crystal. Hence ferritic stainless steel (RA 446) was used for making a crucible holder since it can withstand temperatures in excess of 1000 °C. Ferritic stainless steel is also sufficiently strong within the operating range of the furnace (from room temperature to 850 °C). It was found to be inert and resistant to any thermal wear. Once the stainless steel holder was used, the powdery graphite was not observed and the final solid magnesium product was much cleaner. The growth process was found to be more reproducible after this change of the crucible holder material.

Similarly the form of raw Mg material was also critical in the quality of output. The initial use of powdered Mg (325 mesh/44 μm) did not result in any melting. A brownish powder was obtained after the powdered Mg was subjected to a single crystal growth cycle in the furnace. The particulate morphology of this raw material indicated that it has a high surface to volume ratio. It is believed that significant surface oxidation of the particles prior to melting rendered them with an oxide barrier, which persisted even above the Mg melting point and prevented any amalgamation of the molten Mg particles. Furthermore the oxygen content may also have reacted with any available pure Mg along with carbon from graphite to form a more complex compound. This problem was avoided when Mg chips (4-30 mesh/0.6–4.7 μm) were used as raw material. The relatively lower surface to volume ratio prevented any heavy oxide formation and ensured complete melting of Mg. It was however seen that the melting process in the case of magnesium chips was accompanied by heavy shrinkage. Some of the molten Mg evaporated and settled on the wall of the quartz chamber. The vapor pressure of Mg is around 102 Torr at 800 °C which is the temperature at which the molten metal was soaked. Even though this pressure is lower than the 720 Torr at which the furnace environment was maintained, it is believed that some small Mg regions on the top may have reached higher temperatures which may allow the atoms from these regions to attain sufficient thermal energy in order escape from the molten metal as a vapor. It can also be inferred that the surface of the chips still has sufficient oxide layer which may result in overheating of the magnesium to temperatures higher than 800 °C. The latter may have caused further evaporation and loss of Mg. Using Mg ingots or rods provided further reduction in the surface to volume ratio of the raw material. Subsequent single crystal runs using these ingots resulted in a drastic reduction in the shrinkage. Thus, consistently long single crystal rods were obtained using ingots as starting material. This progression of raw material suggests that the surface morphology of the Mg raw material plays an important role in determining the efficacy of the melting process. While melting can be achieved by lowering the temperature close to the melting point, it was found that the homogenization of the molten Mg at such temperatures was insufficient as observed from the quality of the solidified crystal rods that were obtained. The uniform mixing of the molten Mg at lower furnace temperature would then require much longer soaking times with the associated energy and personnel time costs.

The confirmation of single crystallinity using XRD offered a unique challenge. While Laue XRD and the 4 angle XRD pole figure analysis provide good validation of the grown Mg crystals, these facilities may not be available on a regular basis. Additionally the XRD pole figure analysis was time consuming and destructive since it required slicing of the rod into discs. A quick and non-destructive approach that would be easily available is required for ascertaining the quality of the crystals grown. The powder XRD was readily available and by careful adjustment of the crystal rod via rotation and translation, it was found that consistent single peaks corresponding to one family of planes could be obtained for Mg single crystals along the length of the rods. Laue and XRD pole figure analysis of these crystal rods confirmed that they were single crystals. While this approach cannot detect any secondary crystals that may have nucleated along the periphery or at any end of the rod, it nevertheless offers a quicker way to screen and identify undesirable multi-crystalline Mg rods.

The Mg single crystal was found to have much lower yield strength which points to an orientation more favorable for yielding. The single crystal also showed extraordinary ductility achieving up to 60% strain at room temperature (Figure 18). Such high plasticity is unusual in metals at room temperatures and may occur due to multiple twinning and slip-induced inhibition of strain localization. Super-plasticity has often been seen in ultra-fine grained metallic alloys.26 Schmid et al. have observed the presence of this behavior in single crystals of Mg in 1931.27 Recently Molodov and Al-Samman et al. have observed similar high plasticity in single crystals of Mg under compression along certain orientations. This behavior has been attributed to deformation twinning followed by dynamic recrystallization that resulted in reorientation of the lattice into softer orientations, which promoted basal slip. It is also suggested under other orientations that the opposite trend of strengthening of the crystal and low ductility can also occur.28,29 This behavior allows for modifying the mechanical properties of Mg single crystal by changing the orientation without the need to add alloying elements. Since the Mg single crystals are intended to support load-bearing tissues such as bones and would need to function in a biological environment, the observed high ductility of Mg single crystals needs to be studied further.

  1. A vertical Bridgman crystal apparatus with a moving furnace and computer controls was successfully designed, built, and used for the growing of Mg single crystals.

  2. Graphite was found to be the most suitable crucible material because carbon has very low reactivity with Mg in an argon environment. The conical tip in the casting cavity of the crucible ensures nucleation and growth of fewer nuclei and as such promotes directional growth of a single grain along the length of the crucible.

  3. The slow crucible rotation of 7.5 rpm allows for uniform distribution of heat along the length and width of the crucible during the melting and solidification stages of the growth process. This reduces the chances of additional grain formation, which might occur due to any local temperature fluctuation.

  4. A minimum soaking time of 6 h is needed for the raw Mg to melt completely and for the molten metal to equilibrate the temperature across the melt cavity. For soaking times below 6 h, the obtained rods solidified as multiple crystals.

  5. An optimal growth rate of 30 mm/h or lower yielded good quality Mg single crystals. Laue XRD, XRD pole figure analysis, and powder XRD confirmed the single crystal nature of the obtained samples.

  6. Preliminary tensile tests show that the magnesium single crystals have a modest yield strength and ultimate tensile strength (60-70 MPa), but reveal a remarkable tendency for higher ductility (50%-60%) compared to polycrystalline Mg which is in the range of 10%. Effectively the overall toughness and energy absorption in Mg single crystal is much higher than polycrystalline Mg which makes the material attractive for biomedical device applications.

  7. Preliminary experiments to grow pre-shaped Mg single crystals were successful and yielded screw shaped crystals. More work is needed to adapt the process to grow other implant shapes such as plates and pins.

  8. Magnesium single crystal demonstrated promising properties, which can be harnessed for biodegradable medical implant applications.

The authors are grateful to the National Science Foundation for funding this research through the Engineering Research Center—Revolutionizing Metallic Biomaterials Grant (No. NSFEEC-0812348). The authors are also thankful to Dr. Stephen Rosenkranz of Argonne National Labs for providing his guidance and kind support in performing Laue XRD. Additionally, the authors acknowledge the invaluable contribution of Dr. Sergei Yarmolenko towards performing the XRD pole figure analysis.

1.
G.
Balasundaram
and
T. J.
Webster
, “
An overview of nano-polymers for orthopedic applications
,”
Macromol. Biosci.
7
,
635
642
(
2007
).
2.
R.
Zeng
,
W.
Dietzel
,
F.
Witte
,
N.
Hort
, and
C.
Blawert
, “
Progress and challenge for magnesium alloys as biomaterials
,”
Adv. Biomater.
10
,
B3
B14
(
2008
).
3.
F.
Witte
,
N.
Hort
,
C.
Vogt
,
S.
Cohen
,
K.
Kainer
,
R.
Willumeit
, and
F.
Feyerabend
, “
Degradable biomaterials based on magnesium corrosion
,”
Curr. Opin. Solid State Mater. Sci.
12
,
63
72
(
2008
).
4.
F. A.
Rodriguez-Gonzalez
,
Biomaterials in Orthopedic Surgery
(
ASM International
,
Materials Park, OH
,
2009
).
5.
F.
Witte
, “
The history of biodegradable magnesium implants: A review
,”
Acta Biomater.
6
(
5
),
1680
1692
(
2010
).
6.
G.
Song
and
S.
Song
, “
A possible biodegradable magnesium implant material
,”
Adv. Eng. Mater.
9
(
4
),
298
302
(
2007
).
7.
H.
Hermawan
,
D.
Dube
, and
D.
Mantovani
, “
Developments in metallic biodegradable stents
,”
Acta Biomater.
6
(
5
),
1693
1697
(
2010
).
8.
B.
Zberg
,
P. J.
Uggowitzer
, and
J. F.
Löffler
, “
MgZnCa glasses without clinically observable hydrogen evolution for biodegradable implants
,”
Nat. Mater.
8
,
887
891
(
2009
).
9.
Y.
Zhang
,
C.
Yan
,
F.
Wang
, and
W.
Li
, “
Electrochemical behavior of anodized Mg alloy AZ91D in chloride containing aqueous solution
,”
Corros. Sci.
47
,
2816
2831
(
2005
).
10.
Z.
Li
,
X.
Gu
,
S.
Lou
, and
Y.
Zheng
, “
The development of binary Mg-Ca alloys for use as biodegradable materials within bone
,”
Biomaterials
29
,
1329
1344
(
2008
).
11.
T.
Hassel
,
F.
Bach
,
C.
Krause
, and
P.
Wilk
, “
Corrosion protection and repassivation of magnesium alloys coated with protective magnesium fluoride layer
,”
Magnesium Technol.
485
490
(
2005
).
12.
J.
Yang
,
Y.
Jiao
,
Q.
Yin
,
Y.
Zhang
, and
T.
Zhang
, “
Calcium phosphate coating on magnesium alloy by biomimetic method: Investigation of morphology, composition and formation process
,”
Front. Mater. Sci. China
2
(
2
),
149
155
(
2008
).
13.
Y.
Xin
,
J.
Jiang
,
K.
Huo
,
G.
Tang
,
X.
Tian
, and
P. K.
Chu
, “
Corrosion resistance and cytocompatibility of biodegradable surgical magnesium alloy coated with hydrogenated amorphous silicon
,”
J. Biomed. Mater. Res.
Part A 89A
,
717
726
(
2008
).
14.
H.
Wang
,
Y.
Estrin
, and
Z.
Zúberová
, “
Bio-corrosion of a magnesium alloy with different processing histories
,”
Mater. Lett.
62
,
2476
2479
(
2008
).
15.
L.
Li
,
J.
Gao
, and
Y.
Wang
, “
Evaluation of cyto-toxicity and corrosion behavior of alkali-heat-treated magnesium in simulated body fluid
,”
Surf. Coat. Technol.
185
,
92
98
(
2004
).
16.
B.
Denkena
and
A.
Lucas
, “
Biocompatible magnesium alloys as absorbable implant materials—Adjusted surface and subsurface properties by machining processes
,”
CIRP Ann. Manuf. Technol.
56
,
113
116
(
2007
).
17.
C. R.
McCall
,
M. A.
Hill
, and
R. S.
Lillard
, “
Crystallographic pitting in magnesium single crystals
,”
Corros. Eng., Sci. Technol.
40
(
4
),
337
343
(
2005
).
18.
P.
Schmutz
,
V.
Guillaumin
,
R. S.
Lillard
,
J. A.
Lillard
, and
G. S.
Frankel
, “
Influence of dichromate ions on corrosion processes on pure magnesium
,”
J. Electrochem. Soc.
150
(
4
),
B99
B110
(
2003
).
19.
M. D.
Mochalov
, “
Procedure for Obtaining Single Crystals of Magnesium from the Melt
,”
J. Tech. Phys. (USSR)
6
,
605
(
1936
).
20.
E. C.
Burke
and
W. R.
Hibbard
, “
Plastic deformation of magnesium single crystals
,”
J. Met.
4
,
295
(
1952
).
21.
R. E.
Reed-Hill
and
W. D.
Robertson
, “
The crystallographic characteristics of fracture in magnesium single crystals
,”
Acta Metall.
5
,
728
(
1957
).
22.
T. R.
Long
and
C. S.
Smith
, “
Single crystal elastic constants of magnesium and magnesium alloys
,”
Acta Metall.
5
,
200
(
1957
).
23.
J. L.
Nichols
, “
Orientation and temperature effects on the electrical resistivity of high purity magnesium
,”
J. Appl. Phys.
26
,
470
(
1955
).
24.
V. A.
Labzin
and
V. V.
Bazhenov
,
Phys. Met. Metall. (USSR)
6
,
941
(
1958
).
25.
P. W.
Bridgman
, “
Crystals and their manufacture
,” U.S. patent 1793692 (24 February
1931
).
26.
T. G.
Nieh
,
J.
Wadsworth
, and
O. D.
Sherby
, “
Types of superplasticity
,” in
Superplasticity in Metals and Ceramics
, 1st ed. (
Cambridge University Press
,
Cambridge UK
,
1997
), Chap. 3, pp.
22
31
.
27.
E.
Schmid
and
G.
Siebel
, “
Stress-strain curves of Mg single crystals
,”
Z. Elektrochem.
37
,
447
(
1931
).
28.
K. D.
Molodov
,
T.
Al-Samman
,
D. A.
Molodov
, and
G.
Gottstein
, “
On the ductility of magnesium single-crystals at ambient temperature
,”
Metall. Mater. Trans. A
45
(
8
),
3275
3281
(
2014
).
29.
K. D.
Molodov
,
T.
Al-Samman
,
D. A.
Molodov
, and
G.
Gottstein
, “
Mechanisms of exceptional ductility of magnesium single crystal during deformation at room temperature: Multiple twinning and dynamic recrystallization
,”
Acta Mater.
76
,
314
330
(
2014
).