29 Hydrothermal, LPE & High pressure growth

Prof. P. N. Kotru

epgp books

 

 

29.1 Hydrothermal Growth

 

Hydrothermal crystallization may be defined as a process in which use is made of an aqueous solvent under high temperature and pressure, to increase the solubility of a material (the nutrient), which ordinarily does not dissolve, to a point where it can be crystallized on a seed crystal at an appreciable rate without excessive self–nucleation.

 

This technique is used for materials which are relatively insoluble under ordinary conditions. In this technique the material to be grown is dissolved in an aqueous solvent under high temperature and high pressure and then recrystallized.  Spezia  (1905)  grew  quartz hydrothermally for  the  first  time. However, commercial growth of quartz started in 1955, and quartz is still the only material grown by this method on a  large scale.  Example  of  some  such commercially grown quartz, using seeds cut along different orientations, are shown in figure 29.1

Figure 29.1 :Synthetic quartz crystals grown on different types of seed plates

 

 

In comparison to other techniques, this technique is mostly applied to the growth of quartz. Since this is essentially the same process by which most of quartz, agate, petrified wood etc. was formed in nature and also it provides information under which these naturally occurring materials could have grown, it is probably the reason that much more research has been performed on this technique as applied to quartz than any other crystal.

 

The need for high pressure arose because of negligible solubility of the solute (i.e., quartz). Solubility of quartz in water is so small as to be almost unmeasurable. The solubility can, however, be increased by raising the temperature of solution. To obtain a solubility of even a few tenths of one percent, temperature well above the boiling point has to be used. To prevent water from boiling away, pressure must be applied. Therefore, it is necessary in this technique to apply high pressure. The amount of solubility achieved by application of high temperature and high pressure is not sufficient enough for satisfactory growth. To meet this challenge, a mineralizer is added to the system. Mineralizer is a substance which helps quartz to dissolve, usually by forming a chemical compound with it. With the addition of mineralizer, the solubilities of several percent are possible and growth from solution becomes practical. The amount of mineralizer added to the system should, however, be small enough so that it itself does not form crystals.

 

Hydrothermal crystallization is carried out in a sealed vertical high pressure vessel, also known as “autoclave” or “bomb”. The arrangement is shown in a schematic diagram of figure 29.2.The autoclave has two parts, the lower part and the upper part. A temperature difference T is established between the upper and the lower parts of the autoclave. The upper part is cooler whereas the lower part is hotter. The lower part of the autoclave which is the hotter part is known as “dissolving zone “.

Figure 29.2 :Schematic represntation of an autoclave for growing synthetic quartz

 

The nutrient or the feed material is placed in this part, where it dissolves. A frame holding the seed crystal is placed in the upper portion called as “Crystallizing Zone” or “growth zone”. A perforated metal disc or baffle separates the two zones into two nearly isothermal regions and promotes the growth of crystals of uniform dimensions throughout the zone. The seed crystals are crystal plates of quartz cut along certain suitable crystallographic directions (Viz., X-cut, Y-cut, Z-cut, etc.) as shown in figure 29.3. In commercial plants, several seed plates are placed in racks so as to grow large number of crystals in a single run. The commercial plant is shown in a schematic diagram of figure 29.4.

 

After the vessel is charged with solvent to the percent of fill desired, it is closed and placed in a furnace that gives the desired temperature and temperature differential. The moment temperature of the vessel reaches the operating conditions, the nutrient quartz starts dissolving and saturating the solution. Circulation results from convection currents since the hotter zone is arranged to be at the lower part of the vessel. When the hot solution rises into the cold region of the vessel (upper part) , it can no longer hold all the dissolved material in solution and the excess gets deposited on the seed plate. The seed plate begins to grow in size. Replenishment is influenced by the baffle arrangement, as it transports newly saturated solution away. This continuous cycle of solution and deposition permits the growth of large crystals. The rate of growth depends on temperature, pressure, the temperature difference T between the top and bottom of the autoclave and the amount of mineralizer present. The actual conditions used for the growth of quartz are to be compromised in which various factors are balanced to provide the best possible material at a realistic production rate and cost basis.

Figure 29.3 : Model of a quartz crystal, showing the different directions of cutting to produce oscilllator plates of different characteristics and also used as seed plates for growing synthetic quartz

 

Using these conditions, single crystals of quartz have been grown in the production vessels weighing several kilograms.

Figure 29.4 : Schematic diagram of an autoclave used for commercial production of synthetic quartz

 

29.2 Mineralizers

 

The materials that are grown by hydrothermal method have either negligible solubility in water or are slightly soluble in water. To increase their solubility some substances are added. These substances are called mineralizers. The mineralizers usually used for this purpose include both alkaline as well as neutral substances. The alkaline substances are sodium hydroxide and sodium carbonate, whereas the substances which are more or less neutral include ammonium chloride (NH4Cl), potassium chloride (KCl), sodium chloride (NaCl) and so on. In the growth of certain crystals acid solutions like HCl, H2SO4 and HI are also used. Generally, increasing the concentration of the mineralizer increases the solubility of the material being grown. For alkaline mineralizers, the solubility of the material being grown is proportional to the hydroxyl ion concentration which is a function of the temperature and pressure. Some typical mineralizers under some specific conditions are given in the table 29.1:

 

The above list offers examples of alkaline, neutral and acidic mineralizers.

 

29.3     Advantages of Hydrothermal Crystallization

 

Hydrothermal crystallization offers the following advantages :

 

  1. In hydrothermal growth, one usually refers to high temperatures of crystallization. However, in reality this temperature is often rather low as compared to the melting point of the material. It is this “relatively” low crystallization temperature that makes the technique an attractive means of crystal growth.
  2. The dislocation density of a hydrothermally grown crystal is less than melt grown crystals. It is because in hydrothermal synthesis the crystals grow under less thermal strain whereas in melt growth large thermal gradients are present. This so called low temperature also permits the growth of low temperature polymorphs which are often unattainable by other means.
  3. Hydrothermal growth uses a closed system. In such a closed system the atmosphere may be controlled to produce oxidizing or reducing conditions.
  4. In this technique one can achieve comparatively rapid rate of growth. Quartz grows as fast as 0.25 inches per day on (0001) plane. This rate of growth is extremely rapid for growth from a solvent solute system. This large rate of growth is due to the fact that hydrothermal solutions have comparatively low viscosities together with a large variation of density with temperature at constant average density. This results in rapid convection and very efficient solute transport, permitting large growth rates.

 

29.4  Disadvantages of hydrothermal crystallization

 

The hydrothermal crystallization, however, suffers from certain disadvantages which may be summarised as follows:

 

  1. The need for a well designed high pressure vessel with a reliable closure that is capable of withstanding the high pressure generated at operating temperatures.
  2. The vessel should be structurally strong and also chemically inert (corrosion resistant). It is because high purity of the grown crystal is usually of paramount importance and even a slight autoclave attack is intolerable.
  3. Although hydrothermal growth is relatively rapid, experiments are still lengthy.
  4. In this technique there is no provision to observe progress during a run.

 

29.5 Liquid Phase Epitaxy

 

29.5.1   Introduction

 

Solution growth techniques are slower as compared to melt growth and generally less pure. However, there are materials which are commercially produced such as sugar, common salt, hydrated materials and several compounds which decompose before melting. For such materials normal solution growth is the only option. ADP (ammonium di hydrogen phosphate) and KDP (Potassium di hydrogen phosphate) which are required in the form of very pure and defect free crystals for electronics and optical applications are grown from aqueous solutions. Optical, electronic and high Q-value grade quartz crystals are produced commercially by hydrothermal crystallization and magnetic bubble domain oxide materials are grown by flux method. Aqueous solution methods have produced very large crystals of alum.

 

It is with other solvents, fluxes and metals that liquid phase epitaxial growth (L P E growth) has assumed great importance. In this technique, layers only a few microns thick are grown on to the parent crystal slice but it has gained its importance primarily because of its usefulness to electronics industry. Epitaxial growth of III-V compounds and alloys meant for light-emitting diodes is one of the major achievements of LPE method of growth. In normal solution growth one may or may not use a seed. However, in LPE growth, a seed which may not be of the same composition as that of depositing epitaxial layers is invariably used. Basically, LPE growth differs from solution growth in respect of some points. One difference lies in that the solution here is usually more dilute which leads to slower growth rates, to fewer spontaneous crystallites, and to more stoichiometric layers. The epitaxial growth leads to very pure layers due to use of dilute solution and favourable segregation coefficients such as to keep undesirable impurities in solution. The epitaxial layers are very thin, may be only a few microns rather than millimetres as is in the case of normal solution growth. Though there are other methods of epitaxial growth but LPE has the distinction of producing materials of high purity and perfection that are required in devices of superior performance especially including lasers and light-emitting diodes based on either AlGaAs-GaAs or InGaAsP-InP alloys, the light emitting diodes of GaP and magnetic bubble memories made from magnetic garnets.

 

29.5.2     Experimental Set-up for LPE Growth

 

Figure 29.5(a, b) show schematic diagrams of two common reactors that are used for LPE growth. The reactors are classified as either vertical or horizontal. In both the cases the experimental set-up consists of a furnace, solution and substrate holders. In the growth of III -V materials, the set-up uses a quartz tube containing a high purity H2 atmosphere. For the growth of garnets the ambient gas is air and only single layers are deposited for bubble memory applications. However, for III -V compounds which are required for optoelectronic applications more than one layer is grown. In the growth of GaAs, tipping method is used wherein during equilibration a GaAs substrate is positioned at one end of a graphite boat and the solution is positioned at the other end. When equilibration is achieved, the furnace is tipped so that the solution is able to contact the substrate and the process is followed by controlled cooling cycle.

 

On account of limitations of the tipping method, horizontal slider method is used especially for multilayer fabrication. Both the techniques involving either moving of solutions to the substrate or vice versa are used depending upon the feasibility.

 

There are several invariants of the procedure and technique of liquid phase epitaxial growth. However, amongst these there is one known as “Vertical Dipping Method” which is used mainly for single layers. In this method the substrate is immersed into the solution for growth. The holder is so designed as to break through any crust on the solution surface and of provision for scrapping off any residual solution after withdrawing the substrate. The experimental set-up is shown in a schematic diagram of figure 29.5(b).Since the substrate is used in the form of wafer, it is also grouped under horizontal dipping method. The substrate is rotated to give stirring effect as we do in case of Czochralski systems (used in growth from pure melt).

Figure 29.5 (a): Schematic diagram showing horizontal LPE growth system for the growth of

III-V  compunds (Deitch 1975)

Figure 29.5 (b) : Schematic diagram showing vertical LPE growth system used in magnetic garnet growth ( Ghez and Geiss, 1974)

 

Horizontal dipping is used for the growth of semiconducting crystals. However, there are two difficulties because of which the use of this technique is rather rare. The main problems are as follows:

 

  1. The first problem is that it is difficult to design substrate holders which provide undistorted flows in semiconductor solutions on account of limited range of compatible materials.
  2. The other problem is that the semiconductor solutions are much less stable in the super cooled state as compared to oxide-or halide-based solutions. Because of this problem spurious nucleation is very much common in semiconductor solutions.

 

29.6 Main Uses of LPE Growth.

 

The major uses of LPE growth are as follows:

 

  1. Use in growth of layers where purity is of utmost importance as required in the case of semiconducting lasers.
  2. Use in requirement of growth of those materials where practical alternative techniques are not available as in the case of magnetic bubble devices.

 

29.7 LPE Growth Techniques.

 

There are a large variety of techniques in LPE growth. However, there are only three types which, in practice, are used to any appreciable extent.

 

These include:

i)  Vertical dipping technique.

ii)  Horizontal dipping technique.

iii)  Techniques which makes use of sliding boats.

 

The basic experimental set-up of these methods have already been illustrated and described. The vertical dipping has a limited use so far as semiconductor and oxide materials are concerned. It is necessary that one should be able to get uniform layer thickness. In order to achieve this it is extremely important to resort to those means which would avoid any variation in the boundary-layer thickness. The solution lies in imposing a temperature gradient. The thickness of the boundary-layer increases as the square root of distance from the leading edge of the substrate. On account of the fact that stirring of the solution is because of convection due to difference in the density, the flows near the substrate may be either more or less depending on the density gradients arising as a result of thermal and concentration effects. It is a density driven convection. As for example, we take growth of Yettrium iron garnet (YIG) from PbO. The growing face will leave partially depleted solution around

 

it which would lead to greater density near the face than the remaining solution. It means that there will be downward flow in an isothermal system. The situation is just opposite to this in case of growth of Cadmium telluride (CdTe) from telluride (Te). Whatever be the situation, it is possible to adjust the temperature distribution to reverse the flow in either system.

 

As already explained, the horizontal dipping finds very limited use for the growth of semiconductor crystals. However, the sliding boat technique is used for semiconductors. In this case, graphite boats are used. Boats may also be made of graphite in combination with materials like sapphire and silica.

 

In any of these techniques, there are basically two procedures either of which may be followed to achieve the desired results. The two procedures are:

 

  1. To bring substrate in contact with nearly saturated solution (slightly undersaturated) followed by gradual cooling or single step cooling, enabling creation of supersaturation that may be required to produce the growth of the concerned material. In this case, since initially the solution is just slightly below supersaturation level ( i.e., slightly undersaturated ) it may be used to etch the substrate and leave its surface clean free from edges, layer fronts, scratches or any other superficial defects. After this the solution could be very gradually c ooled, a process technically known ramp cooling. The other way is to create supersaturation level of the solution by cooling at once as but in steps, a process known as step cooling. Each of these cooling steps may be abrupt but should be very sharp.
  2. The substrate and an already supersaturated solution are brought in contact.

 

29.8     Dependence of Growth Rate on Cooling Procedures.

 

Dependence of temperature T with time t on growth rate for the above said cooling procedures (including ramp cooling , step cooling and pre-cooling ) is shown in figure 29.6 (a , b, and c). The effect on the growth rate (v) with time (t) corresponding to each of the cooling procedures is shown in figure 29.6 ( a/ , b/ and c/ ) respectively. In these diagrams the time t = 0 is taken at a time when the substrate is brought into contact with the solution.

Figure 29.6: Variations of temperature T with time t and their corresponding variations of growth rate (v) with time t for different cooling procedures like (a, a/) ramp-cooling ;(b,b/) step-cooling and (c,c/)Pre-cooling .Two types of lines are drawn (a/,b/,c/) which are growth rate diagrams corresponding stirred and unstirred solutions. Full lines ( -) represents growth rate Vs. Time for stirred solutions and dotted lines (—-) stand for unstirred solutions.

 

 

In this figure, the full lines (─) represent growth rate versus time for stirred solutions whereas the dotted lines (- – -) stand for unstirred solutions. Let us discuss each of the curves of temperature T versus time t for three cooling procedures and the corresponding curves of growth rate versus time t.Ts represents saturation temperature. For the ramp – cooling case as represented by figure..(a,a/), the growth rate is almost zero which subsequently rises as the supersaturation increases but after some time it decreases . In fact, the concentration C can be shown to be related to thermodynamic quantities like enthalpy H, gas constant R and temperature T by the relation:

where C0 is a constant.

 

If the solution is cooled by ΔT, it makes available for precipitation and, hence growth, an amount which for small ΔT is given by:

 

For a given ΔT, ΔC will decrease as T decreases. There is a possibility of cooling faster as T falls to cancel this effect. However, an initial transient will always occur. If cooling is done step-wise, the same initial transient effect will occur but additionally, with the passage of time the solution gets exhausted and, for long growth periods, the growth rate dec reases. The said solution exhaustion does happen when pre-cooling is done. If the case is that of stirred system as represented by full line in the curve, the initial growth is high but falls when the solution near to the growth face gets depleted. It continues for a time which is given by the expression δ2D/D, where δD represents solute boundary layer thickness and D is the diffusion coefficient. As times get longer, the solutions which are stirred supply new nutrient to the boundary layer thickness. This fresh nutrient supply to the boundary layer thickness is at a rate which attains maximum value at a time of about L2/ν, i.e., the time for the stirring to penetrate the depth L of the melt having kinematic viscosity ν.

 

The detailed theory of this process is given in the references cited here and are suggested for any extra information on the subject matter.Brice J.C. (1976) in Crystal Growth and Materials, eds. E.Kaldis & H.J.Scheel (Amsterdam, North Holland) ch.11.3;Elwell .D (1976) in Crystal Growth & Materials, eds. E.Kaldis & H.J. Scheel (Amsterdam, North Holland) ch.11.6; Elwell, D., Scheel H.J. (1975) Crystal Growth from High Temperature Solution (Academic press, N.Y.)

 

 

29.9 High Pressure Growth

 

29.9.1  Introduction

 

High pressure growth technique has been applied for the crystallization of diamond. The first confirmed synthesis was reported in 1955 by the General Electric Company, U.S.A. Very high pressure is needed for the synthesis of diamond. Pressure alone is not adequate, since the growth rate has to be kept rapid enough by the use of high temperatures. In actual practice temperatures used are 2000⁰ C or even more and pressures above 1million lbs. per square inch. Growth can be extremely rapid, a few seconds being sufficient under some conditions. Davies (1984) collected a lot of information on diamond, including the processes involved in producing it synthetically.

 

29.9.2 Apparatus Used.

 

The equipment used for such high pressures consists of large hydraulic presses, with dies of special high strength alloys. These concentrate large pressure onto very small volumes, typically less than 1 cubic inch in volume. The natural mineral pyrophyllite is used as a gasket material, since it deforms to transmit the high pressure but does not extrude out of the high pressure region. Electric current is passed through the growth region via insulated press plungers for heating the system. A typical reaction chamber arrangement for forming diamond is shown schematically in figure 29.7.

 

Figure 29.7: Schematic diagram showing reaction chamber arrangement for the synthesis of

Diamond.

 

In this arrangement, a central graphite or carbon cylinder is capped at each end by catalyst metal cylinders. On heating such a system at very high pressure, the metal next to the graphite usually forms an alloy with the carbon and eventually melts. At this stage the catalytic powers of the metal act as a result of which a thin layer of the graphite next to the metal is converted into diamond. With the passage of time, more and more of the graphite gets converted into diamond and a film of molten catalyst sweeps through the graphite, leaving diamond in the process. Finally, the entire mass may get converted into a mixture of diamond and catalyst alloy. Boron–nitride is also prepared in small single crystals by the high pressure technique.

 

The formation of diamond does not appear to be due to simply because of precipitation from solution, but seems to require catalytic effects. It is observed that small traces of certain elements may either poison or accelerate the nucleation and growth of the diamond.

 

Diamond is such a material which is worth growing at very high temperatures of >1500K and pressures > 50kbar. We know that natural diamonds are used for jewellery and hence used by those who deal with ornaments. However, industrialists require it mostly for use as tools, abrasives and ball bearings etc. Most of the synthetic diamonds grown is in the form of fine small hard grains or polycrystalline lumps.

 

It is possible to grow diamonds directly from graphite at temperatures of something of the order of 3300K and 130kbar. A better and practical process involves the use of a solution. Some useful solvents include Nickel (temperature somew here above 1400°C), Cobalt (temperature above 1400°C and iron (temperature above 1300°C). In almost all these cases, the pressures used are in the range of 54-58 kbar and the temperature differences between the source and seed say T≈ 10-30°C. Nitrogen in small quantities in the form of a nitride is suggested to increase the growth rate.

 

Figure 29.8 is a schematic diagram showing the apparatus used to grow diamond crystals. It is a process which involves the use of metallic solutions. The sodium chloride (NaCl) and the pyrophyllite are used as pressure-transfer media, and the pyrophyllite seals are used to plug the system. Here, the growth assembly is placed asymmetrically in the heater in order to ensure the temperature difference that is necessary for the growth. In that way it resembles the temperature differential method of the solution growth with the difference that very high pressures are required in this case.

Figure 29.8: Schematic diagram of an arrangement for the growth of diamond crystals from metallic solutions.

References.

you can view video on Hydrothermal, LPE & High pressure growth

Refernces:

  1. Gilman,J.J. : “ The Art and Science of Growing Crystals”,Wiley,N.Y. 1965.
  2. Hartman,P.: “ Crystal Growth – An Introduction”,North-Holland, Amsterdam,1973.
  3. Lobachev,A.N.: “ Crystallization Processes Under Hydrothermal Conditions”,Consultants Bureau,N.Y.,1973.
  4. Brice,J.C.: “ Crystal Growth Processes”, Blakie& Son Ltd.,1986.

Suggested Reading:

  1. Laudise,R.A.: in “Crystal Growth-An Introduction”,North Holland ,Amsterdam,1973.P.162.
  2. Ballman,A.A. &Laudise,R.A.: in “ Art & Science of Growing Crystals”,Wiley,N.Y.1965, Ch.13.