Freeze drying (or lyophilization) is a process whereby a sample (i. e. a material) is frozen and then the solvent (normally water) within the sample is removed by sublimation.

Currently, many materials are freeze dried. These include a wide range of pharmaceuticals such as vaccines, antibiotics, antitoxins, hormones, therapeutic proteins, liposomes, biological standards etc. With some types of biological material, viability of the material is retained following rehydration of freeze dried material. Examples of such cellular materials include starter cultures for various brewing and food processes and cultures of viruses, bacteria and yeasts for diagnostic purposes. Additionally, freeze drying is used as a processing tool to produce a textured matrix for use in tissue engineering and similar applications.

Shelf freeze dryers are commonly used to freeze dry products, this equipment is generally operated in a batch manner. The product to be freeze dried is placed upon shelves within the freeze drier, the temperature of these shelves is controlled by any appropriate means, commonly silicon oil is circulated through the shelves and the temperature of this oil is controlled. The shelf temperature is reduced and initial freezing of the product occurs, the temperature of the product is further reduced and then the pressure within the freeze drier is reduced and sublimation of ice from the product occurs.

The materials to be freeze dried may either be contained in single doses within vials, ampoules, blister packs or other suitable containers or alternatively be in bulk quantities in trays.

Following initial ice formation in an aqueous solution, all solutes and suspended materials, including cells, become localised into freeze concentrated compartments.

During subsequent reduction in temperature, solutes and cells are exposed to increasingly concentrated solutions; this process continues until the freeze concentrated solution solidifies. The freezing process influences the size and shape of the ice crystals formed in the sample and, therefore, the structure of the porous solid product that remains after drying. The structure of the ice crystals determines the characteristics of the sublimation step, and so the freezing process is also a major determinant of the time required to complete the drying stages (Luyet, B. J. Annals, New York Academy of Science 125,502, 1965; MacKenzie, A. P. Transplantation Proceedings, 8,181, 1976).

It is known to be problematic that during the freezing step, ice nucleation occurs in separate vials over a wide range of temperatures below the melting point of the samples.

While some samples freeze at or just below their melting point, others do not freeze until the temperature of the sample has reached up to 30°C below their melting point. The temperature at which ice nucleation occurs in any particular sample is randomly distributed within this range, with the result that over a large number of vials there is a range of freezing temperatures. The range of temperatures is particularly wide if the samples are filtered, as such samples are especially likely to be prone to undercooling (supercooling) before they freeze. Another factor that influences the range of freezing temperatures is the volume of sample in each container-a smaller volume resulting in a larger range of freezing temperatures.

It is therefore desirable to control nucleation to promote controlled freezing of samples. In some cases, it may be particularly desirable to freeze a sample at or near its melting point resulting in substantially uni-directional ice crystals in the frozen sample.

It is known to add chemical nucleators of ice to samples, but this method has the disadvantage of being invasive and of being unacceptable because the sample is contaminated by the chemical nucleators remaining in the freeze dried product. It is also known to inject water vapour into the evacuated freeze-drying chamber. The water vapour freezes into a fine snow like form which then drifts down into open vials.

However, this method is difficult to control and it is expensive to adapt existing shelf system freeze driers to use this method.

During freezing the concentration of the freeze concentrated phase increases and easily exceeds saturation, resulting in a non-equilibrium metastable state (Franks F. The properties of aqueous solutions at subzero temperatures. In Water. A Cornprehensive Treatise ed. F. Franks. Plenum Press. New York, vol 7 pp 215-338). The freeze concentrated solution essentially becomes supercooled, this is also problematical. For example, solutions containing mannitol are known to supercool extensively and the mannitol does not crystallise during cooling. However during the drying process the sample temperature is increased and crystallisation into mannitol hydrate may occur, this results in an expansion of the product leading to a fracturing of glass vials. This problem may to some extent be controlled by tempering the product during the initial freezing step, however this process is time consuming, lengthens the freeze drying cycle time and is not totally reliable. In other systems the undercooling of the freeze concentrated matrix will be expected to result in sample to sample variation of the freeze dried sample.

Ultrasound is known to induce ice nucleation in undercooled liquids. It is assumed that the effects of ultrasound are mediated by the phenomenon of cavitation.

Small bubbles of the vapour phase or dissolved gas form in the liquid as the pressure wave passes through the liquid. These bubbles collapse after the wave has passed, causing a large pressure change which in turn induces nucleation in the liquid. However, this technique has apparently not hithertofore been applied in the area of lyophilization and no apparatus has previously existed by which ultrasound can be applied to samples within freeze dryers.

It is an object of the present invention to mitigate or overcome the aforementioned problems associated with random nucleation of ice and a freeze concentrated matrix.

A first aspect of the present invention provides a method according to the appended independent method claim 1. A method comprising further novel and advantageous features may be provided as defined in the appended dependent method claims 2-16.

A second aspect of the present invention provides a freeze dried material prepared in accordance with a method as claimed in any of the appended method claims 1-17.

A third aspect of the present invention provides apparatus for freeze drying material in accordance with a method as claimed in any of claims 1-17.

A fourth aspect of the present invention provides apparatus according to the appended independent apparatus claim 20. Apparatus comprising further novel and advantageous features may be provided as defined in the appended dependent apparatus claims 21-24.

A further aspect of the present invention provides apparatus for receiving a material to be freeze dried and for locating in a shelf-free drier, the apparatus comprising means for vibrating material received thereon. Preferably, the vibrating means is adapted to vibrate said material at a frequency of between 10 kHz and 10 mHz, and preferably at a frequency of between 20 kHz and 100 kHz. Ideally, said vibrating means comprises at least one transducer.

Embodiments of the present invention will now be described with reference to the accompanying drawings, in which: FIGURE 1 is a schematic side view of a freeze drier apparatus according to the present invention; FIGURE 2 is an electron micrograph of a freeze dried mannitol solution in which spontaneous nucleation has occurred during conventional freezing at-12. 5°C ; and FIGURE 3 is an electron micrograph of a freeze dried mannitol solution in which nucleation has occurred in accordance with the present invention.

In a first embodiment of the present invention, a vial of sample is freeze dried with a method comprising the steps of : 1. Reducing the temperature of the sample to below its melting point ; 2. Subjecting the sample to vibration to induce ice nucleation in the sample ; 3. Further reducing the temperature of the sample below its melting point; and 4. Subjecting the sample to a reduced pressure to sublimate solvent from the sample.

The further reduction of the temperature ideally reduces the temperature of the sample to substantially below its melting point.

Also, the vibration is preferably mechanical, sonic or ultrasonic induced vibration, most preferably sonic or ultrasonic. Furthermore, the vibration may be generated by electromagnetic, electromechanical, piezoelectric, electrostrictive or magnetostrictive means.

In a further embodiment of the invention, a vial of sample to be dry frozen is placed in a dry freezer and the temperature of the sample is reduced to between its melting point and 10°C below its melting point, preferably between its melting point and 5°C below its melting point, at which temperature the sample is subjected to a pulse or pulses of vibration. The temperature profile within the sample when nucleation is induced may be manipulated in order to achieve a particular structure and set of properties. In this respect, a sample may be cooled by contact with the cold shelf within freeze drier. During conventional operation of the dry freezer equipment, this leads to large vertical thermal gradients within the sample. The bottom of the sample can be below the melting point and the top above the melting point. Nucleation of a vial in this configuration leads to ice crystals nucleating at the bottom of the vial which form large vertical crystals. Samples frozen in this manner have a very high porosity and sublimate rapidly. The size of the thermal gradient is determined by a number of factors including, for example, the rate of temperature reduction of the shelf, the initial temperature of the sample, the vertical height of the sample, and viscosity of the sample. In order to achieve the fastest rate of drying for a sample, any of these variables may be exploited to achieve large vertical gradients before nucleation.

Nucleation of a sample with a large vertical temperature gradient may result in segregation of the solute during the formation of large ice crystals. This may be detrimental to the recovery of some materials and may also lead to the formation of a "skin"at the top of the vial. The process conditions may be adjusted such that the temperature of the vial contents are essentially at a uniform subzero temperature. For example, one simple way to achieve this would be to maintain the shelf temperature isothermally for a period to allow equilibrium to occur. Nucleation of vials at an isothermal temperature, near to the melting point, will result in large ice crystals being formed, but will reduce the segregation of the solution.

In a further configuration, it is possible to produce samples which are colder at the top than at the bottom. This may be achieved by producing samples which are isothermal by any manner as described above. These are then heated at the bottom. In this configuration it is possible to induce ice nucleation at top of the sample. The ice crystals then grow to the bottom of the sample when heat is removed via the shelf. In this configuration, large ice crystals are formed at the top of the sample and the initial rate of drying is rapid.

In a preferred embodiment of the invention the duration of the vibration is up to five seconds, most preferably up to one second.

Where the vibration is ultrasonic, sound waves having a frequency of between 16 kHz and 10 MHz, most preferably between 20 kHz and 100 kHz are employed.

Following the nucleation step, the temperature of the sample is further reduced below its melting point (preferably to substantially below its melting point) prior to being subjected to a reduced pressure, and, whilst temperature is being so further reduced, the sample is subjected to continuous or pulses of vibration (preferably ultrasonic vibration) to promote formation of a preferred crystal form (nucleation) within the residual unfrozen fraction.

The freeze drying process may be carried out on samples placed on one or more shelves within a freezer-drier wherein the vibration is applied to the samples by means of transducers. The transducers may be mounted in a number of configurations. For example, the transducers may be attached to the underside of the or each shelf, with the vibration transmitted through the bottom surface of the shelf, the coolant and finally the top surface of the shelf. This configuration is potentially simple to fit to existing equipment so that they may be used in accordance with the invention. In order to increase the efficiency of transmission of the ultrasound to the samples, the coolant in the shelves should be de-gassed as efficiently as possible.

In an alternative arrangement the transducers are sealed through the bottom surface of the shelf and directly attached to the top surface, optimally with no direct contact with the coolant. Also, where baffles, fins or supports are connected to the top wall, the transducers may be attached to these. In yet further alternative arrangements, the transducers may be attached to the top surface of the shelf or attached to one or more of the edges of the shelf. These transducer mounting arrangements can be used in combination with one another.

As an alternative, the vials may be held in racks and the vibration applied via the racks rather than via the shelves.

As another alternative, the vibration may be applied to the samples by a metal mesh or plate which is driven at ultrasonic frequencies and may be resonant at these frequencies. This device can be placed between the samples and the top surface of the cooling shelf. The transducers may be mounted in a number of configurations such as: i) Attached to the top or bottom surface of the device (i. e. of the metal mesh/plate). ii) Attached to one or more edges of the device (i. e. of the metal mesh/plate).

The attachment of the transducers may be direct or indirect or via appropriate conductors of ultrasonic energy.

Where avoidance of standing waves is important, the driving signal may be via transducers of differing frequencies and/or phase.

To ensure ultrasonic coupling between the vials and the shelf or said device the vials may be held against the shelf or device by any suitable means.

To increase the efficiency of ultrasonic nucleation, the likelihood of cavitation within the liquid may be increased. This may be achieved by elevating the level of dissolved gas in the sample, by the addition of a volatile, miscible liquid or by reducing the external pressure, prior to subjecting the undercooled liquid sample to ultrasound.

The solvent gas may be any gas or mixture of gases, but preferably a highly soluble gas such as carbon dioxide or nitrous oxide. To maximise the amount of gas in solution the gas should be introduced at as low a temperature as possible. Alternatively a volatile liquid may be added, for example a low molecular weight alcohol may be advantageously added to aqueous solutions.

In freeze driers operated in a conventional manner, the process conditions during the sublimation phase (temperature of the shelf, chamber pressure) are selected in an empirical manner. Cooling in a standard freeze drier results in ice nucleation at a range of temperatures and the large variation in sample porosity results in the samples drying at different rates. Process conditions are selected to dry the samples with the lowest porosity, ie those which have nucleated at the maximum supercooling. If drying conditions are applied which are too"aggressive" (rapid) then the samples with the lowest porosity will melt during drying. Because of the range of variation within the samples there are no available methods available to allow the process conditions to be directly controlled by monitoring the samples. Following controlled nucleation, all samples will have a very similar porosity and it will be possible to control the drying conditions either by monitoring drying in a single vial by any appropriate technique (temperature measurements, microbalance etc) or by monitoring the vapour pressure within the drier using BTM measurements.

A further aspect of the invention provides apparatus 2 for freeze drying a sample, the apparatus comprising means 4 for reducing the temperature of the sample to below its melting point, means 6 for subjecting the sample to vibration to induce a controlled amount of nucleation in the sample, means for further reducing the temperature of the sample and means for subjecting the sample to a reduced pressure to sublimate solvent from the sample.

There are many advantages in the present invention over known methods of freeze drying. These advantages include: uniformity of freeze dried products, and the creation of large ice crystals in a sample which leads to reduced drying time and thereby reduced overall freeze drying time. Samples with large ice crystals also rehydrate rapidly.

A number of example experiments conducted in accordance with the present invention will now be discussed.

Example 1: Ultrasonic Nucleation at kHz Frequencies 1.1 Description of the system A conventional freeze drier was modified such that ultrasound could be applied during the initial cooling phase of freeze drying. In this regard, six transducers operating at 35 kHz were bonded under a stainless steel bowl (165 mm id) to serve as a modified shelf. These transducers were matched to two 100 watt generators with controllable power output. The modified shelf was placed within the condenser unit of a Heto freeze drier (Heto, Birkerod, Denmark, Model CD1. 5).

1.2 Experiment Freeze drying vials, each 14mm diameter (Adelphi Tubes ltd, Haywards Heath, Sussex) containing 1.0 ml of an aqueous solution of sucrose (0. 8M), were placed on the modified shelf. Each vial had a thermocouple (28 SWG) placed within the sample liquid and sample temperatures were recorded on a data logger (Grant Instruments, Barrington, Cambridge). When all the vials had nucleated, the experiment was stopped and the data analysed to determine the nucleation temperature within each vial. This was evident as a rapid increase in temperature as the sample returned to its melting point. The amount of undercooling for each sample was recorded.

1.3 Results Nucleation during normal operation, with no power supplied to the ultrasonic transducers, resulted in a very large scatter observed in nucleation temperatures among samples. Vials generally nucleated at the bottom, with subsequent ice growth up the vial.

However, following extreme undercooling, nucleation was also observed to propagate from the meniscus or within the body of the liquid.

In contrast, when samples were cooled below their melting point and then subjected to ultrasound, all vials nucleated within 5°C of the melting point. Under these conditions, nucleation invariably occurred at the bottom of the vial.

Example 2: Ultrasonic Nucleation at MHz Frequency The experiment described in Example 1 was repeated in all respects except that undercooled vials were irradiated with ultrasound on the face of a 1 MHz probe operating at 2 watt cm-2 (Ultrason, Medical Equipment Services Ltd, Nottingham). All vials undercooled by 5°C were nucleated by this treatment.

Example 3 : Effect of Controlled Ice Nucleation on Sublimation Ampoules containing 10ml of Dobutrex (dobutamine hydrochloride 28 mg/ml, mannitol 25 mg/ml) were nucleated close to their melting point using ultrasound at 35 kHz as described in Example 1. Values of porosity were determined by direct measurement of sublimation rate in a quasi isothermal system. The porosity of samples nucleated near to the melting point was 7.2 x 10-8 (kg s-l mol Pa-1), whilst the corresponding value for samples which had undergone extensive undercooling was 1.6 x 10-8 (kg s-l M71 Pa-1).

The effect of increasing the value of porosity, by nucleating near to the melting point, was to reduce the primary drying time (with the bulk ice temperature maintained at 295K) from 16 hours to 8 hours.

Example 4: Ultrasonic Nucleation via a Driven Plate 4.1 Small circular plate driven from centre A small circular steel plate of diameter 120mm and thickness 1. 3mm was driven directly by a 36kHz ultrasonic transducer which was attached to the plate at its centre via an M8 bolt through the plate.

Vials of diameter 24mm containing 10ml of distilled water were cooled, along with the circular plate, to-5°C. A two second pulse of ultrasound was then applied to the plate. Nucleation occurred in all those vials that were pressed onto the plate when the ultrasonic pulse was applied. As the vials were positioned randomly on the plate it was apparent that the distribution of ultrasonic power around the plate is suitable to induce nucleation irrespective of the vial position.

4.2 Rectangular plate driven near the edge A rectangular steel plate of length 400mm, breadth 360mm and thickness 1. 5mm was driven directly by a 20kHz ultrasonic transducer attached to the upper surface of the plate via an M12 bolt situated 20mm in from the centre of the 360mm long side. This ultrasonic transducer provided approximately 200W of energy when the plate was fully loaded.

Vials of diameters 24mm and 40mm, containing 10ml and 25ml of distilled water respectively, were cooled to-3. 5°C in a low temperature alcohol bath before being placed on the plate. The plate was not pre-cooled and so, in order to prevent the vials from warming up, the two second ultrasonic pulse used to induce nucleation had to be applied within 10 seconds of the vials being placed on the plate. All vials were nucleated by this treatment.

4.3 Cooled rectangular plate driven near the edge The rectangular plate described in example 4.2 was placed within an insulated chamber. This chamber was then cooled to-3. 5°C via a copper cooling coil through which cold alcohol was circulated.

Glass vials of 24 mm diameter and 40 mm diameter, containing 10 or 25 ml of distilled water respectively, were cooled to-3. 5°C in a low temperature bath and then placed on the plate. The vials were left on the plate for up to 2 minutes and, following a two second ultrasonic pulse, all vials were observed to nucleate.

Example 5: Driving the cooling shelf, mesh or plate.

A further technique for driving a cooling shelf, mesh or plate within a conventional shelf freeze drier is as follows. In order to achieve appropriate lateral distribution of the applied ultrasound, each transducer is attached to the plate indirectly via a rectangular block, preferably slotted, such as would be used as a welding tool for plastics. The driving transducer may be attached to the block via an exponential or other suitable horn. The longitudinal distance from the driving position on the block to the radiating surface of the block is typically equal to half the wavelength of the longitudinal waves in the horn at the driving frequency. To make the lateral amplitude distribution more uniform across the radiating surface of the block, additional passive vibratory elements such as"wave shaped horns" [Adachi K. & Veha S. , J. Acoust. Soc. Am. 87 (1), January 1990, pp 208-214] may be added.

The driving transducers may be of any type. To reduce reflections from the far end of the plate and so reduce standing waves, an absorber may be attached to the opposite edge of the rectangular plate. The absorber may be a dissipative device such as a dashpot, a 'motored' transducer or an active absorber tuned to the phase of the travelling wave.

Example 6: Implementation within a Commercial Freeze Drier.

A single shelf within a conventional freeze drier (Model LyoGamma 15 Special, Telstar Industrial, Terrasa, Spain) was modified to allow the method of the invention. Twenty three 50W transducers (40kHz) were bonded to the bottom of a standard shelf (454 x 458 mm). These transducers were connected to an external generator. The temperature of the shelves was controlled by the circulation of silicon oil, which was degassed by vacuum. The freeze drier also contained an unmodified shelf which allowed the benefits of the method to be examined in the same experimental run. Typical results are outlined below: Fifty glass vials (16 mm in diameter) containing 2 ml of a 10% w/v aqueous solution of mannitol were loaded per shelf (control shelf and modified shelf). Each shelf was held at 0°C and the vials were equilibrated for 30 mins. The temperature of the vials were monitored by T type thermocouples connected to a data logger. The plate was then cooled at 1°C nain-I to-14'C and held at that temperature until the liquid at the bottom of a vial was measured by a thermocouple to be at (-4°C to-6°C). The top of the vial would be expected to be at 0°C. Ultrasound was applied for 3 seconds and all (50) of the vials were observed to nucleate. In respect of the control shelf, spontaneous nucleation was observed to occur when the temperature at the bottom of the vial was between-11°C and - 15°C. Samples were then cooled to-60°C at a rate of 0. 7°C min-1. Drying was carried out at 0. 2mbar at a plate temperature of-14°C, drying time for the nucleated samples were 12-16 hours depending on the position on the plate and 28 to 39 hours for the "spontaneous"samples. The structure of the freeze dried samples was determined by scanning electron microscopy, the nucleated material had a larger pore size (SEM available) than the spontaneous material (SEM available).

Shattering of the vials-presumably associated with the crystallisation of mannitol hydrate-occurred in 5/50 of the nucleated vials and 9/50 of the control vials.

Scanning electron micrographs of a freeze dried mannitol solution (2 ml samples freeze dried in conventional glass vials in a shelf freeze drier) are shown in Figures 2 and 3. Figure 2 shows a sample in which spontaneous nucleation has occurred during freezing at-12. 5°C. The ice crystal structure is very fine and non-directional, seen in the freeze dried sample as pores. The freeze dried mannitol has a dense uniform structure and is mechanically robust. Figure 3 shows a sample in which ice has been nucleated in accordance with the invention, with a temperature at the base of the vial of-4°C. Large directional ice crystals have formed-seen as cavities in the SEM. The freeze dried mannitol formed in large sheets and the sample fi-actured very easily. Both micrographs are at the same magnification.

Example 7-Demonstration of the nucleation of the freeze concentrated matrix The nucleation of a supercooled freeze concentrated matrix is illustrated here with the well defined system water: glycerol: NaCl.

An aqueous solution of glycerol (10% v/v) = 0.15 m NaCl was cooled in a Euctimeter (Telstart Industrial, Terrasa, Spain). The solidification temperature for the freeze concentrated matrix for a solution of this composition is measured to be-64°C by differential scanning calorimetry. During cooling the freeze concentrated matrix was observed to supercool by at least 10°C, solidification as measured by a step change in resistance did not occur until the temperature was at least-75°C. During warming melting of the freeze concentrated matrix, measured by a decrease in resistance, occurred at temperatures close to-64°C.

A solution was cooled to-70°C, at this temperature the freeze concentrated matrix was supercooled, ultrasound was applied and a step increase in resistance occurred clearly indicating that nucleation of the material had been initiated.

The present invention is not limited to the specific embodiments described above.

Alternative arrangements and suitable materials will be apparent to a reader skilled in the art.