Discussion of Lyophilization or Freeze Drying in Trays by Thomas A. Jennings, Ph.D and Professor Ana Bacaoanu, Ph.D†. †Department of Chemical Engineering, “Gh. Asachi” Technical University of Iasi, Romania No doubt that trays played a major role in the early days of lyophilization or freeze drying of products and continues to this day to be a major vessel in conducting these drying processes. Advances in freeze drying equipment, such as automatic loading, has to some degree reduced the need for the use of a tray. But this is only for products that are contained in vials or bottles and these advance systems are generally not used to freeze dry bulk materials. So for bulk products, we still must rely on a large container that we generally classify as a tray. We intend in this INSIGHT to generally discuss the various materials used in fabricating trays that have been and continue to be used in the lyophilization or freeze drying of products. We will certainly point out the advantages, disadvantages and perhaps pose some new questions that we need to address. The trays that will be considered range from the traditional metal (stainless steel) to the more recently introduced Lyoguard(R) . It should be clear to the reader that we have NO preference nor bias for or against any given type of tray but just wish to present an objective evaluation and make it clear that the selection of the tray will no doubt be very product dependent. So what may work for one product may not be applicable for another.
Composition: Trays are generally constructed out of three basic materials: metal, glass and plastic. As a result of the composition, each material will offer its own unique advantage and disadvantage. Cold Roll Steel: To my knowledge, we know of no application today where this material is used in the fabrication of trays. Certainly, they would not be applicable in the food or health care industry because of the potential for rust contamination. However, we do not see any objections for their use in the manufacture of flowers and especially if they are either galvanized or electroplated. They would perhaps be easier to construct and may represent the least expensive of all the trays. Stainless Steel: Trays made from stainless steel are perhaps the most commonly used trays in the lyophilization or freeze drying of products. They have numerous advantages which would include and not be limited to (a), they are corrosion resistant: (b), such trays can be depyrogenated at temperatures of 250 oC; (c), these trays are also durable and require no special handling and (d), they are reusable which in time can off-set their initial higher cost.
Aluminum: Aluminum trays are used in the freeze drying of bulk materials. They are certainly light weight when compared to the cold roll and stainless steel trays and are corrosion resistance. In addition, they can be anodized to give them a black surface which will increase their emissivity aid some what to the energy transfer resulting from radiant energy. They can be steam sterilized but we do not have any knowledge where they have been depyrogenated at 250 oC. Because aluminum is a relatively soft metal, one must be concerned with the effects of “galling” (the removal of a surface layer generally in the form of a dust) should the product have to be removed by scraping. Glass: Not all bulk products that can be lyophilized or freeze dried come into contact with a metal surface. As a result of their composition, these trays have similar advantages as those constructed from stainless steel, except they are by their very nature fragile and are subject to breakage that may occur over time from stress strains resulting from age and thermal shock. Plastic Trays: Molded trays are useful in those applications not requiring high temperature depyrogenation. They tend to be (a), light weight when compared to the trays made from the previous materials,; (b), easy to clean and (c), quite durable and reusable. Lyoguard(R) Trays: While these trays can also be generally classified as plastic trays, they do have unique features and we felt they deserved their own category in this discussion. The basic reason being that they are covered with a membrane barrier that will allow a formulation prepared in an aseptic environment to be safely lyophilized or freeze dried in a non aseptic environment . This feature is of major importance to those working with small freeze dryers which were not designed to withstand the harsh rigors of steam sterilization. However, given such an advantage there are also some disadvantages in that they tend not to be reusable and as a result of their unique feature to be more costly than those trays fabricate from the above materials.
Heat Transfer Properties: We have discussed the impact and need to understand heat transfer properties in previous INSIGHTs, e.g. (Vol. 5 No. 12 and Vol. 6 No. 1) and references [2-3]. In this INSIGHT, we shall just consider the role that composition and configuration of the tray can have on the heat transfer properties. In making such a comparison, allow me to make a number of assumptions and conditions so as to avoid possible confusion. •
All the trays have the same general rectangular configuration and the only variables will be their total mass and thickness of bottom and walls.
That all trays contain the same volume of water.
In each case, the shelf-surface temperature is uniform and chamber pressures are equal.
The heat transfer rate (Qs) from shelf through the bottom wall of the tray is given by the Fourier`s law:
lambda - the thermal conductivity of the composition of the tray, W/ (m K) or W/m oC. delta - the thickness of the bottom of the tray, m; Ts - the temperature of the shelf surface, given in oC; Tp - the temperature of the internal bottom surface of the tray, oC; t - time, sec. Neglecting the contact resistance between shelf and tray, the external surface temperature of the bottom of the tray is equal to that of the shelf. One can consider that the temperature gradient has a single direction, perpendicular on the surface of the bottom of the tray. In the regions with a good contact area of the bottom of the tray with shelf surface there is also a good heat transfer. In the regions where the contact is not perfect it appears a new thermal impedance to heat transfer between shelf and bottom of the tray. This impedance is given by the existing medium (gas at very low pressure and low density), between shelf and bottom of the tray. and will increase as the pressure is lowered. Consequently, the heat transfer is reduced in these regions. But the heat transfer is more complicated. To the normal
temperature gradient (perpendicular on shelf) new gradients can be added on the other directions, anywhere a temperature difference exists between different points (position). Cold Roll Steel: For purposes of this INSIGHT we shall assume that the thickness of the walls and bottom are 0.051 inches (1.2 mm). The thermal conductivity (lambda) cold roll steel is about 45.6 W/(m oC) and the specific heat over the normal temperature range for lyophilization or freeze drying is 0.4609 kJ( kg· oC) . Now it is well known that it is rather difficult to fabricate a steel tray in which the bottom surface is in full thermal contact with the shelf surface. Thus one can only expect a fraction of the bottom surface to be in direct contact with the shelf surface. In the worse cases, contact may be only at three corners or rest on a small area in the middle. For our comparative model ( which we will use for all trays except for the Lyoguard(R) trays) we will assume only 10% of the bottom surface will be in contact with the shelf. Given that (Ts - Tp), is 1 oC, the value of Qs now becomes 3.875·A (in kW). One can see that as A increases, the thermal conductivity across the tray becomes more of a real factor. Energy will flow from the edges towards the center so one can expect a temperature gradient across the bottom of the tray as the temperature of the shelf is varied. This temperature gradient will be directly proportional to the thickness of the tray and inversely proportional to the length of the conduction path so the configuration of the tray is an important consideration. If the height of the walls of the tray are 1 inch (2.54 cm) and the density of the steel is taken as 7850 kg/m3 , the mass of the tray is defined as “abd”, where “a” is the width of the tray and “b” is the tray length in m such that the mass (M) can be expressed as: M = Vm · 7850 (kg)
where Vm is defined as the volume of metal used to fabricate the tray. The amount of energy (Q) required just to change the tray temperature by 1 oC (DT =1) can be expressed as : Q = M· cp · Delta T = M ·0.460911 ·1 = 0.46091· M(kJ)(5) Stainless Steel: Given the same thickness as the cold roll steel, the thermal conductivity (lambda) is given as 15.11 W/(m oC). A comparison of the thermal conductivity shows that the cold roll steel is three (3) times more thermally conductive than that of the stainless steel. For a tray of the same dimensions as that of the cold roll steel and with only 10% of the bottom of the tray in contact with the shelves, the heat transfer rate (Q) between the shelf surface and the tray when (Ts Tp) is 1 oC will be 1.259·A (kW). As a result, for a tray constructed out of stainless steel, it will only transfer 33 % of the energy of that for the cold rolled tray.
The specific heat of the 304 stainless steel is 0.5028 kJ/(kg oC), the amount of energy one  while the density is given as being 8020 kg/m3 . The mass (M) of the tray having the same dimensions as above, will then be given as: M = Vm ·8020 ( kg)(6) The amount of energy (Q) required just to change the tray temperature by 1 oC can be expressed as: Q = M· cp · DeltaT = M ·0.5028 ·1 (kJ) (7) Since the mass of the tray will be greater than that of the cold roll steel, the amount of energy to change the temperature of the tray by 1 oC will be more than that for the cold roll steel. Aluminum: Aluminum trays (alloy 2017) having the same thickness as the cold roll steel, the thermal conductivity (lambda) is given as about 104 W/(m oC). A comparison of the thermal conductivity shows that this alloy of aluminum is 2.23 times more thermally conductive than that of the cold roll steel and 6.88 times more thermally conductive than the 304 stainless steel tray. For a tray of the same dimensions as that of the cold roll steel and with only 10% of the bottom of the tray in contact with the shelves, the heat transfer rate (Qs) between the shelf surface and the tray when (Ts - Tp) is 1 oC will be 8.66·A (kW). Thus the heat transfer rate for an aluminum tray would be 2.23 twice that for a cold roll steel tray and more than 6.88 times that for one constructed out of 304 stainless steel. The specific heat of the of aluminum alloys is 0.8799 kJ/(kg oC) , while the density is given as being 2800 kg/m3 . The mass (M) of the tray having the same dimensions as above, will then be given as: MAl = Vm ·2800
The amount of energy (Q) required just to change the tray temperature by 1 oC can be expressed as QAl = MAl· cp · DeltaT = MAl· 0.8799·1
QAl = Vm ·2800 · 0.8799 = 2463.72 ·Vm
While for cold roll steel we have: Q = Vm ·7850 ·0.460911 ·1= 3618 ·Vm
Although the specific heat for aluminum is about twice that for the cold roll steel and the stainless steel, it will require less energy to increase its temperature 1 ºC, in fact, when compared to stainless steel it will only require 68 % as much energy.
Glass: Because glass trays are more fragile and subject to breakage the tray thickness will be greater and be of the order of 0.20 inches (5 mm) or about four times thicker than the metal trays. For a borosilicate glass the thermal conductivity is about 0.83 W/(m ºC)  while the specific heat is 0.75 kJ/(kg ºC) . The heat transfer rate (Qs) for this type of tray will be 0.166 ·A (kW). The density of borosilicate glass is about 2300 kg/m3. The mass (M) of the tray having the same dimensions as above, will then be given as: Mglass = Vmglass ·r = Vm ·2300
The amount of energy (Q) required just to change the tray temperature by 1 oC can be expressed as: Qglass = Mglass· cp · DeltaT = Mglass · 0.75· 1 = 0.75 Mglass (kJ) (11) Even though it will require less energy to increase the temperature of the tray by 1 oC, because of the poor thermal conductivity, one could expect at times large temperature differentials that can exist across such trays as a result of the much lower thermal conductivity. Plastic Trays: Assume that the plastic material that we have used to fabricate the tray is PETG  and has a gauge thickness of 18 (1.2 mm) and so will be similar in dimensions to that of the above metal trays. The thermal conductivity of the PETG is 0.1675 W/moC and the heat transfer rate (Qs) for the tray when there is a 1 oC differential would be just 0,166 ·A (kW). This “Qs” value is considerably lower than that for a thicker layer of glass which is surprising. The specific heat of the PETG plastic is 1.089 kJ/(kgoC) , while the density is given as being 1270 kg/m3 . The mass (M) of the tray having the same dimensions as above, will then be given as: MPETG = VmPETG ·1270 (kg)(12) The amount of energy (QPETG) required just changing the tray temperature by 1 oC can be expressed as: QPETG = MPETG · cp · DeltaT = 1.089·1· MPETG (kJ)(13) While it will require little energy to change the temperature of the tray, the poor thermal conductivity could, at times, result in large temperature differentials across such trays during the lyophilization or freeze drying process. Lyoguard(R) Trays: The thickness of the Lyoguard(R) tray is 0.008 inches (0.2 mm) and according to Steve Delrosso of W. L. Gore & Associates, Inc., the material is flexible enough to conform to the shelf surface. In view of this information, we shall assume that at least
90% of the bottom of the tray is in contact with the shelf surface. Mr. Delrosso also informed me that the bottom of the tray was fabricated from polyethylene. The thermal conductivity (lambda) of polyethylene ranges from 0.32 - 0.5028 W/(m oC). As a result the heat transfer rate (Qs) (assuming the l for this tray is estimated to be 0.5028 W/(m oC) will be 2.514·A (kW) for a temperature differential of 1 oC. A comparison of the (Qs) for this tray shows that its heat transfer rate will be 15 times to 18 times higher than that expected for either the glass or plastic tray. However, when compared to that of the metal trays it is considerably lower, e.g., the q will be 1/6 th that for the stainless steel tray. The specific heat of the polyethylene is 1.88 kJ/(kg oC) , while the density is given as being 940 kg/m3 . The mass (M) of the bottom of the tray will be: MLyoguard= VmLyo ·940
The amount of energy (Cp) required just to change the tray temperature by 1 oC can be expressed as: QLyoguard = MLyoguard· cp · DeltaT = 1.88·MLyoguard (kJ) (15) Because of the relative low mass of the bottom of the tray, the amount of energy required to increase the temperature of the polyethylene by 1 oC is considered to be insignificant. Summary The above calculations are presented in the following table:
Based on our model, it is clear that the heat transfer rate for the metal trays is significantly greater than those trays fabricated from glass or a plastic. However, the reader is cautioned not to be too hasty in coming to any conclusion that metal trays are far superior. There is no doubt in my mind that some trays may make far less than 10% contact with the shelf. It is quite conceivable that only 1 % or even less of surface area may be in actual contact with the shelf surface. In those cases, the heat transfer rate of the Lyoguard tray could exceed that for the 304 stainless steel tray. It is doubtful if the heat transfer rate for the Lyoguard tray would exceed a tray fabricated from cold roll steel or aluminum. Thus the lesson here is that the area of contact of the metal tray and the shelf is a critical parameter.
Effect on Freezing Process:
How a product is frozen can have a major impact on how it dries . From the above we have seen how the design and composition of the tray can affect the heat transfer rate. Except for the LyoguardďŁ¨ tray, it was assumed that thermal conductivity took place along the edges of the tray; however, it should be noted that this is not always the case. Except for the LyoguardďŁ¨ tray, freezing will occur along the edges of the tray first and then move towards the center of the tray. Unless, the formulation has a high degree of supercooling (see INSIGHT Vol. 5 No. 5) or the fill height is small, there will be a tendency to form a non-uniform cake. A mathematical model for the freezing of each of the above trays is beyond the scope of this INSIGHT and would make a good paper for the ISL-FD Inc. Conference in Chicago, IL in September 2003. But one can get some idea of what such a paper would reveal by considering the freezing of just one liter of water at say 20 oC to a temperature of -40 oC. The total energy that one would have to remove would be of the order of 501.54 kJ. For a 5 oC differential across the shelf the time required to freeze the water would be dependent on the rate at which the shelves could be refrigerated. But for the glass or plastic tray freezing would be limited by the thermal properties of the materials and less on contact area with the shelf. It will be interesting to see a more thorough mathematical analysis combined with supporting experimental data of this problem which would increase our knowledge of the freezing mechanism that takes place in the various types of trays.
Primary Drying: We have already considered the primary drying with an INSIGHT (Vol. 5 No. 2) and in depth in ref , so the only interest here is to again determine the effect that the configuration and composition will have on this process. Metal Trays: Given that contact with the shelf surface will generally occur along the edges of the tray, one would anticipate that the primary drying will not be uniform across the tray and that the area near the walls of the tray will tend to have a higher drying rate. As a result placement of temperature sensors in the product near the edge of the tray may indicate a false high product temperature (Tp ) whereas placement of the sensor probe in the middle of the tray can produce false low values of Tp . It is interesting to note that such temperature variations in the tray can occur even in the presence of a very uniform shelf surface temperature. This wide variation in Tp will be further enhanced as the chamber pressure is decreased and the amount of energy transferred as a result of gas thermal conductivity is reduced. Consequently, operating at higher pressure can help to reduce the variation in the drying rate across the tray but cannot completely eliminate it. Glass and Plastic Trays: In order to achieve a similar drying rate as that in the metal trays, for a given chamber pressure, one would have to significantly increase the shelf surface temperature. It would come as no surprise that a more thorough mathematical analysis of primary drying in these trays will reveal an even greater temperature (Tp)
variation across such trays. A real concern in using these trays is that the shelf temperature is not excessive so that a portion or perhaps the entire cake may be subject to excessive Tp temperatures at the end of the primary drying. Of course, a variation in Tp in these trays, like the metal trays, will likewise be subject to chamber pressure. If there is a positive feature to be gained in these trays it is that they tend not to be subject to changes in configuration as a result of repeated usage. Thus careful selection of these trays based on their temperature variation during primary drying could provide a more reproducible drying process. Lyoguard(R) Tray: As stated previously, this type of tray is unique because of the presence of a porous cover that will permit the flow of water but will prevent biological contamination of the product even if the product is dried in a non-sterile system. However, the presence of this cover will offer some impedance to the flow of water vapor from the tray. Thus given the same chamber pressure as that used for the other trays, the pressure in the LyoguardďŁ¨ tray will have to be greater in order to conduct the primary drying process. Because of the higher pressure, the Tp for the product will also be higher. Just how much the membrane barrier will impede the flow will be dependent on the value of the pressure and nature of the membrane. For further information on the nature of gas flow at low pressures the reader is referred to [2, 9]. No doubt this is a very complex issue and well beyond the scope of this INSIGHT. It would in itself be worthy of a technical paper.
Secondary Drying: We have already considered the role of pressure in the secondary drying process in INSIGHT Vol. 4 No. 6 and reference . But neither of these publications take into detailed account the impact that the composition of the tray will play in the drying process. It is therefore of interest to consider this topic briefly at this time. Metal Trays: Because of their higher thermal conductivity the metal trays will tend to complete secondary drying faster than those trays fabricated from glass or plastic. However, there still exists this haunting question regarding the impact that the contact area of the bottom of the tray with the shelf surface will play during this process. In our model, where there is 10% of the bottom surface area of the tray is in contact with the shelf and it is uniform along the edges of the tray, the secondary drying will tend to proceed similar to that which had occurred during the primary drying process. This would mean that the residual moisture content would tend to be higher in the middle of the tray than along the edges. In the case of drying proteins, one would have to be careful of over drying and hence denaturing of the protein (change in the configuration of the protein) along the edges of the tray. This raises the question as to configuration of the tray. Perhaps the worse case could be one in which the tray is the same dimensions of the shelf for a large production dryer. This would not only pose a problem for the distribution of moisture in the product during secondary drying but also have a similar impact during primary drying.
Glass and Plastic Trays: Since there is no sharp demarcation point for primary and secondary drying in that secondary drying is taking place during primary drying , the low thermal conductivity of these trays can present a problem with regards to the distribution of the residual moisture in the final product and one would not at all be surprised to find that secondary drying in these trays may require a longer drying time. Especially if there are variations in cake structure resulting from the freezing process. One would expect that the surface area in contact with the shelf surface for these trays will be more reproducible than that of the metal trays; however, one would still anticipate that the dimensions of the trays will play a significant factor in the drying process. Lyoguard(R) tray: While this type of tray does allow for a larger percentage of the bottom surface area to be in direct thermal contact with shelf surface, none the less there is still the question of the impact of the membrane barrier on the residual moisture content in the product. When Tp = Ts, the pressure in the tray will be equal to that in the chamber so water vapor will be transported by diffusion rather than flow. Drying by diffusion is a considerably slower process than when flow is present. So a real problem with this type of tray is knowing when the product is within the desired range of residual moistures values. Opening the tray to sample the product without destroying the integrity of the tray could prove difficult. If the cake has a good self-supporting structure, one could reduce the residual moisture by â€œpurgingâ€? the system with a dry gas such as nitrogen  INSIGHT Vol. 4 No. 6. However, once the tray is removed the product can adsorb moisture from the humidity in the room. Here the presence of membrane barrier may prove to be an advantage and slow the absorption rate of moisture in the trays until it can be safely stored to protect it from the ambient humidity. The LyoguardďŁ¨ tray without question has some real advantages but it also has some limitations or potential problem areas.
Summary: This INSIGHT has shown that we do need to pay more attention to lyophilization or freeze drying in trays. One must pay particular attention to the composition of the tray with regards to its thermal properties. Such thermal properties can have a major impact on the nature of the cake resulting from the freezing process, the duration of the primary drying and secondary drying. It was also shown that one must pay careful attention to the construction of the dryer. No doubt the metal trays are excellent when it comes to heat transfer but leave a lot to be desired when it comes to having good reproducible contact with the shelves. No doubt others will have some additional information or techniques regarding these trays which we will most warmly welcome.
References: 1. W. L. Gore & Associates, Inc. at http://www.gore.com/lyoguard/aboutlyoguard.html
2. T. A. Jennings, Lyophilization - Introduction and Basic Principles, Interpharm Press, Buffalo Grove, IL 1999. 3. T. A. Jennings and Henry Duan "Calorimetric Monitoring of Lyophilization", Journal of Parenteral Science and Technology, Vol. 49, No. 6 pp. 272-282 (1995). 4. Handbook of Chemistry and Physics 65th edition (Robert C. Weast, ed), CRC Press, Inc., Boca Taton, Florida 1984. 5. http://www.azom.com 6.http://www.matweb.com 7. http://www.oriel.com 8. http://members.aol.com/newimageplastics/PETG.html 9. S. Dushman, Scientific Foundations of Vacuum Technique, second edition (J. M. Lafferty, ed.) John Wiley & Sons, Inc. New York, 1962. Volume 6 No. 5