Sublimation Science HOME

Introduction to Lyophilization

Lyophilization vs Freeze Drying

For popular purposes of language, and at this website, lyophilization and freeze-drying are functionally synonymous. Some authors would like to differentiate between lyophilization and freeze-drying by saying that lyophilization is the sublimation event occurring when all components (tissues, interstitial matrix, water, etc.) are solidified and maintained at a temperature below their glass transition temperature or crystal melt temperature, whichever is lower. Freeze-drying, according to these authors, would be sublimation events that occur primarily to water in a substance while other components might (or not) be above their glass transition temperatures.

There are two common instances of freeze drying which deserve special mention. The first is the disappearance of snow despite the temperature being below freezing and the second is the preservation of certain food items by the Incas in the Andes mountains. The Incas use of freeze-drying to preserve food between 1300 and 1500 AD marks the first recognized use of the tool. The current application of lyophilization implies four standard conditions as follows.

  1. Freezing of the product.
  2. Application of a vacuum to assure the phase diagram region where ice sublimes directly into gas.
  3. Addition of heat to promote the sublimation
  4. Condensation of the removed water separate from the original product.
While these conditions are descriptive of the current application of the tool, they are not totally necessary to achieve the freeze-drying result. Letís look at them individually and apply them to the two special cases presented here.

The first condition is freezing. Both snow and the Incas food was frozen and by definition we will assume that they stay frozen through the process. So we can accept number one. Freezing of the product is a valid condition.

The second condition is application of a vacuum. Although the Andes are tall, the vacuum produced as a consequence of altitude below 30,000 feet is negligible compared to pressures used in commercial lyophilizers. Current commercial applications would use 100 milli-Torr or 0.1333 mBar, clearly far away from Andesí air pressure. Also, at last check, upper Michigan was still recording daily pressure variations quite close to 1 atmosphere and snow was continuing to sublime. From these data, it would appear that condition two is busted. We may use a vacuum, but it isnít necessarily needed.

Air Pressure at Altitude
English Metric
Feet Torr Meters mBar
0 760.0 0 1013.3
2000 704.1 609.6938.8
4000 652.4 1219.2 869.8
6000 604.4 1828.8 805.9
8000 560.0 2438.4 746.6
10000 518.9 3048 691.8
12000 480.7 3657.6 640.9
14000 445.4 4267.2 593.8
16000 412.7 4876.8 550.2
18000 382.3 5486.4 509.7
20000 354.2 6096 472.3
22000 328.2 6705.6 437.6
24000 304.1 7315.2 405.4
26000 281.7 7924.8 375.6
28000 261.0 8534.4 348.0
30000 241.8 9144 322.4

Another misconception about sublimation and vacuum comes about from undergraduate chemistry courses and textbooks which show phase diagrams for water with the ice/gas line ending at 4.579 torr. That is, at a pressure above 4.579 torr, and above 0°C ice will melt into water rather than subliming into steam (water vapor). That interpretation is a misunderstanding of the phase diagram. It is a diagram for pure water no other gases allowed. Furthermore, Daltonís Law of partial pressures says that other gases donít matter. At 0°C the 100% relative humidity point is 5.3 torr (7.067 mBar), and the relative humidity has only to be below 86.4% at 0°C for sublimation to begin. In light of this information, the second condition should be modified to read ďa reduction in the partial pressure of water such that it is substantially less than the saturation pressure at any given temperatureĒ. While this condition is usually met by lowering total pressure, it is only necessary to lower water vapor pressure to meet the requirement.

The third condition is the addition of heat to promote the sublimation. The third condition is true by intent, although the word heat should be changed to energy. Indeed, in the special cases above, it is radiant energy from the sun and not heat that causes the sublimation. Although the sun can certainly heat substances, our cases require the maintenance of temperature below zero degrees C because above zero, water melts and canít sublime.

Condition four calls for condensation of the removed water vapor in a location other than the original ice. Conventionally the condition is met by use of a cold trap but any mechanism that assures the water vapor will not return to the ice is sufficient. Indeed, there is no requirement that the vapor be condensed. With a traditional oil vacuum pump, if the vapor isnít condensed onto a cold surface, it will condense in the pump oil. However, it is theoretically possible to remove the water vapor through a vacuum pump that pushes it into the atmosphere and avoid condensation as a part of the system. Lyophilization does not require condensation, only that the water vapor be removed from the system.

As an aside, I have heard it stated that the condensers are a ďdriving forceĒ for the lyophilization. In an evacuated system that is wrong. One can, however, make the case that a very cold condenser chamber will necessarily reduce the water vapor pressure to a value commensurate with the chamber temperature and thereby cause lyophilization in a manner identical to the evaporation of snow. Compared with the use of a vacuum, this is a very slow way to sublime ice.

Elements of Lyophilization: 1

What is really necessary for lyophilization?  First, we should understand that we are talking principally about the sublimation of water whenever that water is mixed (or dissolved) with other chemical compounds.  To merely achieve sublimation of the water, or at least most of it, we must be be below 0°C.  That is, the water component must be frozen. In freezing, the water will crystallize, which is a separation step and as the water crystallizes, the other components will necessarily concentrate.  For pharmaceutical lyophilization, and in particular the lyophilization of active biological proteins, it will also be necessary to lower the temperature such that all components of the system are frozen.  While we can discuss the meaning of "frozen" in great depth, suffice it to say that it means solidified in the usual sense of the word.  Compounds in the frozen state can be subjected to DSC (differential scanning calorimeter) analysis to assure that they are lower in temperature than the glass transition temperature (Tg') for the compound. In some cases, that may be colder than needed, but it is adequate.  Rubber tires are above their Tg' and thus pliable.  Brittle plastics are colder than their Tg' and thus solidified in the sense meant here.  Being cold enough to have all components solidified will result in a so called "elegant" lyophilization cake.  What that means is that the other components do not settle during the sublimation.  As a consequence of not moving around, the resultant cake will have exactly the volume of the original frozen solution.

After freezing, but before pulling a vacuum, it is normal to chill condenser piping (cold trap) to at least about -65°C and to direct the vapor path across the trap in order to protect the vacuum pump from incoming water vapor. If there were several liters of water in a product and it all found its way to the entrance of a rotary vane vacuum pump, the pump would quit working, since the vapor would condense back into water as it contacted the warm oil.

After freezing and chilling condensers in a vapor path, it is common to pull a vacuum on the product.  In the context of the discussion above, a vacuum isn't needed but what is necessary is that the vapor pressure of water be lowered to less than 4.579 torr.  In fact, to achieve a reasonable sublimation rate, the water vapor pressure needs to be below about 200 mTorr (0.2 Torr). During sublimation in an evacuated chamber, the entire (100%) vapor pressure is caused by water molecules. Thus, if the pressure is reading 200 mTorr on a capacitance manometer, then the water vapor pressure is 200 mTorr. That is only about 22 fold less than the water vapor pressure at atmospheric.  Most modern lyophilizers have the ability to bleed in gas (usually N2 or air) and maintain a stable temperature. If observed during the bulk of primary drying, it can be seen that these bleed valves hardly open. The exact chamber pressure that is chosen depends on how cold the product ice must be. That is quite a complex topic and is discussed further in a more advanced context.

Once the vapor pressure is lowered, heat must be added and then sublimation proceeds to completion and determining the endpoint is a principle topic. Heat addition is normally achieved by warming shelves inside the lyophilizer, although other methods have been used, and everything works.  Development studies as well as instrumentation are used to assure that the mass of water has been removed from the product during the sublimation phase - known as "primary drying".  At the end of primary, moisture content will be less than about 15% and commonly less than 6%.  As a way of understanding how dry that is, cotton clothing in a closet will have a moisture content of about 10%.  So at the end of primary, products will look and feel dry.  No further sublimation happens because there are no remaining ice crystals.

The final step of lyophilization is desorption - commonly called "secondary drying".  Here, the moisture content is reduced from less than 15% to less than 1%.  Water molecules are layered around the product in shells - not to be confused with "waters of hydration".  :The removal of water from these shells is a surface phenomenon and is modeled by surface chemistry.

The figure above shows a really simple model for Langmuir adsorption.  However, adsorption is the opposite of desorption and a model for one is a model for the other.  It turns out that although desorption is assisted by low pressure (less than about 200 mTorr), it is mostly driven by temperature and temperatures in excess of 20 degrees are needed to cause it to proceed at a reasonable rate.  Commonly, a temperature of at least 30° to 40°C is used for a time of about 4 hours.  That time and temperature is sufficient for vialed pharmaceutical products where the cake heights are less than 1 cm.  The Langmuir model is the simplest mathematical model for desorption and it isn't simple.  A further discussion, with math, of the model will be presented on the website is another section.  Other models, suitable for multiple layers of water molecules exist, and authors are proud when they get a solution, whether exact or numerical.

In summary, The common steps of lyophilization are listed below.

  1. Freeze the product
  2. Chill a water vapor condensation path
  3. Lower the pressure in the product chamber to less than 200 mTorr
  4. Add heat to the product
  5. Detect the endpoint of sublimation
  6. Raise the temperature to remove water in secondary

Commercial lyophilizers mostly control pressure, shelf temperature, and time. The relationship among these controllable variables is the topic of the section here called Intermediate Lyophilization.