Crystal Growth Technologies
Sapphire is the ideal material – hard, strong and inert, with amazing optical properties and a high melting temperature – for thousands of applications in every industry on earth. Trans- forming raw sapphire into finished parts is also a process subject to the craftsmanlike expertise of those who do it. S&D has forged strategic partnerships with the world's leading crystal growers and sapphire fabricators and supplies parts of all sizs and shapes – current designs and new ones – for markets including semiconductor manufacturing, energy, optics, aerospace, education & research, industrial equipment, military, government and more. Several methods are used to grow single crystal sapphire; what follows is an overview of technological development in this field.
VERNEUIL FLAME-FUSION GROWTH METHOD - Melt and drip
In 1877, while he was working a the Museum of Natural History in Paris, Edmund Fremy (1814-1894) and his assistant, Charles Feil, published a paper in which they described how they synthetically obtained small, ruby crystals by holding crucibles filled with seed material at red heat for 20 days. Ruby is cousin to sapphire, with the addition of chromium giving the material its red hue. After Feil died in 1876, Fremy's new assistant was Auguste V. L. Verneuil (1856-1913).
1n 1885, rubies selling for $1000 – 2500 per carat from an unknown source appeared in Geneva were purported to be natural gems. investigation at the Sorbonne in Paris and at Tiffany in New York found microscopic gas bubbles in the material suggesting a high-temperature synthetic process had produced them. The French Syndicate of Diamonds and Precious Stones ruled the crystals had to be marked 'artificial', which reduced their price to $25 – 40 per carat. Modern investigation of the ‘Geneva Rubies' suggests a 3-stage flame fusion synthetic process. One of the people who examined them in 1886 was mineralogist P. M. E. Jannettaz at the Museum of Natural History in Paris. He concluded that fusion was involved in the creation of the Geneva rubies and he discussed his findings with his colleague, Verneuil. Following this introduction, Verneuil - who had gone on to earn his bachelors, masters and doctoral degrees while working with Fremy - together with G.A Terreil, used a hydrogen-oxygen torch to fuse alumina powder with added chromium salt. They obtained tiny particles which Jannettaz identified as rubies, and from this experiment Verneuil began to develop flame fusion growth of crystalline substances.
The method of crystal growth Verneuil pioneered was built on the foundation of knowledge constructed in the previous century: Gauden synthesizing ruby crystals in a crucible; Senarmont observing microscopic rhombohedrons forming when aluminum chloride was heated; Gay-Lussac discovering how to create pure Al2O3; Rose finding that SiO2 contamination came from the agate mortar used for grinding; Gaudin using a torch on fused alumina plus Cr salt to produce ‘hard red particles'; and finally, Fremy's experiments with adding compounds to obtain ruby crystals, trials on which the young Verneuil assisted; and the anonymous alchemist who created the mysterious Geneva rubies.
Method or Distinction
The process that became the crystal growth method bearing his name was simple: a tower was built with the source material beginning its gravity-assisted journey from the top to the bottom. Feed material was placed in a hopper; to control the flow of material, a mechanical hammer tapped the powder down into a tube; along the way, oxygen, fuel and heat were introduced into the tube, and the raw alumina feed was brought to 1000 - 1200oC so the material would coalesce. Near the bottom, a downward pointing hydrogen-oxygen torch burned inside a ceramic muffle insulator (crucible) which sat upon an alumina pedestal. The feed material (powder by this point) falling through the hottest part of the flame was heated to just below melting point; the initial powder built up a conical pile in which the hot particles sintered together to form a rigid ‘sinter cone'. The upper tip of the cone became molten when the cone tip achieved sufficient sharpness, and fresh powder landing on the molten ball would melt and enlarge the crystal. Only the uppermost layer of the growing crystal would remain molten, and as the crystal grew vertically, the pedestal was lowered with a crank to keep the top of the crystal at the correct location in the flame; during crystal growth, the flame temperature was regulated by controlling the flow of oxygen. At the end of the growth run the flame was shut and, if there was not too much strain introduced, the crystal would not shatter. Even so, there was typically so much strain that a light hammer tap on the end would split the crystal lengthwise into two halves. Verneuil formally published his method in 1904.
Material and Applications
Material grown by the Verneuil method is limited dimensionally by physics and typically has curved growth striations throughout, so it is generally less suited for optical. The primary use for Verneuil-grown sapphire and ruby today is still for synthetic sapphire and ruby gemstones, watch jewels, watch windows, and the process is virtually unchanged from the original design.
1932, Soviet scientist S.K. Popov (1903-1953) began improving the Verneuil crystal growth process, introducing semi-automated apparatus and optimizing feed and heating conditions to produce more uniform and higher-quality corundum crystals. By 1938, he had developed equipment which enabled the manufacturing of long thin rods suitable for making jewels for watches and instruments. From then until his death, he worked at the Institute of Crystallography of the Academy of Sciences in the USSR, improving equipment, solving problems, developing new uses for sapphire such as a durable fiber guide in the textile industry, and devising equipment to bend corundum rods into shapes. It is now widely held that by the 1950's, Popov's equipment was more advanced and the crystal he produced of higher quality than that available anywhere else in the world.
There is still significant demand for Verneuil-grown material because the method is still the least expensive way to make sapphire and ruby adequate for many applications. However, higher crystal quality demanded by electronic, optical and other applications, and the need for sizes and shapes not possible by flame-fusion, impelled the development of other ways to make sapphire.
Due to the physics involved, the size and shape of parts that can be produced from material grown by the Verneuil method have relatively fixed dimensional limitations; the largest boules obtained today reach sizes 100 mm in diameter and 250mm in length. Gemstone Buzz.com says that boules are now grown 12in dia using the ‘flame-fusion method'.
A typical growth time of two (2) hours produced a cylinder with a diameter of 6mm and a length of 25mm. In late 1980's, a Verneuil sapphire and ruby grower in Switzerland had as many as 2200 Verneuil torches running simultaneously producing upwards of the 100 tons annually.
CZOCHRALSKI (CZ) METHOD - Pull from the melt
The Czochralski (CZ) technique of single crystal growth was developed in 1916 by Jan Czochralski as the serendipitous consequence of an accident and careful, astute observation by Czochralski himself. On an evening in July, he left a crucible with molten tin for a period of time and returned to writing his crystallization study notes. At some point, absorbed by his thoughts, instead of dipping his pen in the inkwell, he dipped it in the crucible and quickly drew it out. A thin thread of solidified metal hung from the tip and the discovery was made. The slot in which crystallization was initiated was eventually replaced by a narrow custom capillary and, in certain instances, by a seed of the growing crystal itself. Czochralski found later that the crystallized wire was in fact a single crystal.
The application of the CZ method exclusively as a technique for obtaining single crystals is due to W. von Wartenberg. At its simplest, it was a method of producing large single crystals by inserting a small seed crystal into a crucible filled with molten material, then slowly pulling the seed up from the melt in concert with its simultaneous rotation. Subsequent modifications of this method were applied to guide the basic theory into the realm of practical application.
Blurb: In his own investigations Czochralski used the Bridgman method to grow his single crystals.
The Czochralski method was essentially forgotten after end of World War II, but as demand for semiconductor electronic materials surged in the early 1950s, Americans G.K. Teal and J.B. Little from Bell Telephone Laboratories rediscovered this growth method, giving it worldwide recognition as the Czochralski method for growing large single crystals on an industrial scale. Today, still, it is accepted that no other crystal growth method can match its combination of crystal quality, dimensional flexibility and limited machining costs for overall value.
As with all sapphire crystal growth methods, it involved the crystalline solidification of Al2O3 molecules from a liquid phase at the interface. Fundamentally, the CZ process is the same today as it was first developed in the 1950s.
CZ crystal growth consists of the following general steps: seed material of a precise charge amount is loaded into a molybdenum crucible inside a custom growth chamber, and all gases inside the chamber are evacuated. The chamber is backfilled with an inert gas to inhibit the introduction of atmospheric gases into the melt during crystal growth. The charge material inside the chamber is then melted. A slim seed of sapphire crystal with precise orientation and tolerances is introduced into the molten sapphire. The seed crystal is then withdrawn at a very controlled rate; the seed crystal and crucible are rotated in opposite directions while withdrawal occurs. This step is repeated, with additional crystal layers added each cycle, until the desired or maximum size and shape are achieved.
The diameter of the boule is controlled by the rate of withdrawal of the seed from the melt.
Gravity dictates the maximum size as each successive crystal layer adds corresponding weight to the growing boule. Careful, continuous power and monitoring are required throughout the growth process which can last up to eight weeks.
Material and Application
In the early stages, it was determined that the optical quality of material grown by the Verneuil flame-fusion method was not adequate for lasers, and CZ grown sapphire was recruited in part because of its high optical properties. These are some of its uses today:
Boules, or ingots, of single crystal sapphire grown using the Czochralski growth technique achieve sizes in the range of 150 mm (6 in) in diameter by 250 mm (10 in) in length or height.
Seed is rotated at up to 30 rpm and is pulled out of the melt at 6 - 25 mm/hr for a growth time of up to 8-10 weeks.
KYROPOLUS METHOD - Dip and turn
In 1926, Spyro Kyropulos replaced the process of pulling the crystal pulling from the melt with the direct crystallization of the melt by decreasing the boule's temperature while it’s still in the crucible. Colleague M.I. Musatov contributed the solution to the boule-removal challenge by suggesting the crystal be pulled for the size excluding the single crystal layer contacting the crucible walls at the final point of solidification.
Kyropoulos first proposed his technique at The Physical Institute in Gottingen, Germany as a way of producing large single crystals of alkali halides that were free from cracks and any damage due to their restricted containment. Until then, the only method to grow large crystals was the Bridgman technique. The Czochralski method yielded material in thin crystal filaments, and Verneuil boules had intrinsic dimensional limitations. The Bridgman growth method was unsuitable for Kyropoulos’ goal of producing material for precision optics applications because of the cracks and flaws caused to the crystal by the container mechanism during cooling.
Method or Distinction
The basis for the Kyropoulos method is that the sapphire crystal is formed deep under the surface of the molten alumina ‘melt' and as it solidifies, it takes on the cylindrical shape of the crucible through the action of a 'shrinkage cavity' forming around the crystal. The physics behind this phenomenon are a function of the difference in density between the liquid and solid (crystallized) sapphire, 3 and 4g/cm3 respectively.
Pure alumina powder is placed in a crucible and brought to melting temperature (2050oC) turning it to molten liquid, or ‘melt’, through resistive heating. The seed crystal, attached at its top to the end of a vertical water-cooling rod, is lowered into the melt. Thermal gradient controls the process so that only the crystal layer at the solid-liquid interface remains molten, and as the seed crystal is slowly drawn back out of the melt – the pace of the draw remains less than that of crystallization - the added crystal layers increase the size of the solidified boule.
Material and Applications
In general, this crystal growth technique is ideal for materials with low thermal conductivity and a high degree of thermal expansion, the combination of which can make crystal material vulnerable to slip and fracture unless grown - and cooled - in a low-stress environment.
The Kyropoulos method, with its highly controlled thermal-gradient, produces large-diameter boules of very high optical quality due to its high purity and low dislocation density. These solid boules can be cut to any crystallographic orientation or plane. Kyropolous grown sapphire is ideal for many optical applications including:
Boules, or ingots, of single crystal sapphire grown using the Kyropoulos growth technique achieve sizes in the range of 150 mm (6in) in diameter by 250 mm (10in) in length or height.
Kyropoulos grown boules are typically 25kg taking up to fourteen days, when controlled automatically.
HORIZONTAL MOVING GROWTH METHODS incl: Bridgman - Stockbarger, Horizontally Directed Solidification Method (HDSM), Bagdasarov/HDC - horizontal slab growth
While Czochralski worked on his growth technology, Bridgman proposed a method for single crystal growth based on moving a conical crucible through a fixed melting zone. His philosophy built on Tammann's idea back in Gottengen of what is now known as a 'gradient-freeze' method, in which crystals can be grown directionally in the temperature gradient region of the furnace, adding crystallized mass as the temperature is gradually reduced. Bridgman added crucible movement; D. Stockbarger further improved the technique when, after discovering that crystal quality increased with the rise of the temperature axial gradient at the solid-liquid interface, he created a special thermal diaphragm in the crystallization zone which helped increase the temperature levels achieved.
In 1964, Kh S. Bagdasarov proposed a new element in the moving 'gradient-freeze method' which the crucible containing raw material and the single crystal seed moving in a horizontal direction.
Method or Distinction
In this method, crystals are gown from the melt in a horizontal boat-shaped container at a growth rate of 8-10mm per hour. This method makes it possible to produce large slabs with nearly perfect edges of any crystal orientation: C-plane, R-plane (optical) etc.
Material and Application
This method produces crystal in a slab form, enabling very thick windows and components possessing very high optical quality (Grade 2 or 3), nearly perfect edges in several crystallographic orientations: C-, M-, or R-plane. It is ideal for manufacturing sapphire ingots, blanks and windows with diameters from 2 to 8 in. Material gown this way also enjoys widespread use in blue LEDs, with annual volume in this area forecasted to increase.
Crystal material grown by this method takes the form of relatively thick rectangles with sizes in the range of 170 x 200 x 35-40mm.
'DIE' METHODS: EDGE-DEFINED FILM-FED GROWTH (EFG)/STEPANOV - Pull through die
In 1965, Harold Labelle, a skilled technician with only a high school education, was tagged by Tyco Industries in Waltham, MA to begin work on an Air Force contract to develop a process for growing sapphire fibers as reinforcement for metal-matrix compounds. Labelle drew on previous experience in Tyco's semiconductor crystal growth experiments with gradient freeze and 'traveling solvent' methods. Through this work, hegained both theoretical insight into, and practical experience at, controlling the solid-liquid interface, an elemental concept of crystal growth.
Labelle found that cold tungsten plunged into molten alumina formed small crystalline alumina dendrites on the tungsten, limited to 3 mm in length. Reasoning that temperature control might help, he next melted the alumina in a tungsten troth or ‘boat’, inside a vacuum, and dipped a tungsten wire into the melt. When it was carefully drawn out, a sapphire crystal grew on the wire ‘seed’. He had taken a basic Czochralski technique and introduced crystal shape control. By adding fixed dies with orifices specific to the shape and size of the desired crystal, the design of a self-filling tube achieved mechanical, and therefore thermal, stability. The resulting crystal material was 3-4 times stronger, with more consistent surfaces, than anything he had obtained previously. In 1967, using a design exploiting the thermocapillary properties of molten alumina, and a series of concentric molybdenum tubes, LaBelle produced the first as-grown sapphire tubes.
Observationally talented, Labelle noticed that the diameter of the sapphire crystals often grew larger than the aperture through which they’d been drawn; the delta depended on the angle of the upper surface of the orifice; if the top surface was flat, the molten material spread out to the edge and stopped. The diameter of the die, not of the melt column, dictated the diameter of the filament produced. When he realized this growth technique differed from existing ones, Labelle wanted to call it 'continuous shaped-film propagation', but his boss, Ed Mlavsky strongly suggested ‘edge-defined film-fed growth’ so that, or EFG, is how its known.
This attribute - what exactly defines the edge of the newly-formed crystal - is one of the key differences between EFG and the concurrently-developed Stepanov method of crystal growth.
In the USSR during the 1960’s, working independently of Labelle, famed Soviet scientist A.V. Stepanov was applying his insatiable curiosity, analysis and endless output of ideas to the novel field of shaped crystal growth. The basis of the Stepanov method lay in the distinction that the shape of the grown formed crystal is determined primarily by the aid of a fixture shaping the melt column. The shaped liquid-melt column is then transformed into solid material via temperature control of growth rate, shaping the crystal as it grows.
Like Newton with the falling apple, Stepanov's 'aha moment' came as he observed a daddy long-legs spider traveling along a water's surface by virtue of the forces of surface tension. It is this force that forms the basis of capillary shaping method of crystal growth. The difference between EFG and Stepanov methods is in what exactly shapes the crystal; in EFG, it is the shape of the die, while with Stepanov, wettable or non-wettable aids shape the melt column, which in turn shapes the crystal.
Method or Distinction
The EFG technique of crystal growing as it is used today is still as simple and straightforward as the process Labelle ultimately arrived at. Al2O3 (alumina) is melted in a molybdenum crucible. The melt 'wets' the surface of molybdenum die and moves up by capillary attraction. A sapphire 'seed crystal' of desired crystallinity is dipped into the melt on top of the die and 'pulled' or drawn out, crystallizing the Al2O3 into solid sapphire, in a shape – rod, tube or sheet (ribbon) –determined by the die. Crystal orientation can be tightly controlled – any axis or plane can be produced using proper controls during growth.
The two primary advantages of EFG crystal growth method are the ability to produce various shapes beyond what is possible with other technologies, and significant cost savings allowed by near-net growth which necessitates less machining and other costs of finishing parts.
One drawback of the EFG process is the time and cost of producing the dies needed to create the shaped crystals. Depending on several factors, this can take weeks and cost several thousand dollars, which may be absorbed by the grower, or passed on to the customer as an upfront tooling fee or a premium distributed across an initial quantity of parts like a deductible.
Material and Applications
Crystal material resulting from this method is typically low to medium optical quality (Grade 3 - 6) and can be precisely produced - by design - in different crystallographic orientations (A, C, random). It is commonly used for industrial, mechanical, and low grade optical applications where high end optical quality is not critical.
Large diameter, specific shape requirements are other areas in which the strengths of EFG and Stepanov grown sapphire make it the only solution available.
Uses for die-grown sapphire include:
EFG sapphire is also employed extensively for high-usage applications in semiconductor manufacturing equipment including:
- Ribbon or sheet with usable thickness of 0.5 in up to 12 in wide and nearly 20 in long
- Tubes with OD from 0.125 up to 3.5 in (3mm to 80+ mm); Lengths of up to 40 in or more (1.5m)
- Rods up to a meter long with diameters up to 8 mm
When existing or standard tooling can be employed, growth to completion can range from four to eight weeks; if a new die must be designed and manufactured, total time to finished product can be as long as eight to twelve weeks.